UBC Faculty Research and Publications

Correlation of proteome-wide changes with social immunity behaviors provides insight into resistance… Parker, Robert; Guarna, M M; Melathopoulos, Andony P; Moon, Kyung-Mee; White, Rick; Huxter, Elizabeth; Pernal, Stephen F; Foster, Leonard J Sep 28, 2012

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


52383-13059_2012_Article_2953.pdf [ 396.6kB ]
JSON: 52383-1.0220559.json
JSON-LD: 52383-1.0220559-ld.json
RDF/XML (Pretty): 52383-1.0220559-rdf.xml
RDF/JSON: 52383-1.0220559-rdf.json
Turtle: 52383-1.0220559-turtle.txt
N-Triples: 52383-1.0220559-rdf-ntriples.txt
Original Record: 52383-1.0220559-source.json
Full Text

Full Text

RESEARCH Open AccessCorrelation of proteome-wide changes with socialimmunity behaviors provides insight intoresistance to the parasitic mite, Varroa destructor,in the honey bee (Apis mellifera)Robert Parker1, M Marta Guarna1, Andony P Melathopoulos2, Kyung-Mee Moon1, Rick White3, Elizabeth Huxter4,Stephen F Pernal2 and Leonard J Foster1*AbstractBackground: Disease is a major factor driving the evolution of many organisms. In honey bees, selection for socialbehavioral responses is the primary adaptive process facilitating disease resistance. One such process, hygienicbehavior, enables bees to resist multiple diseases, including the damaging parasitic mite Varroa destructor. Thegenetic elements and biochemical factors that drive the expression of these adaptations are currently unknown.Proteomics provides a tool to identify proteins that control behavioral processes, and these proteins can be usedas biomarkers to aid identification of disease tolerant colonies.Results: We sampled a large cohort of commercial queen lineages, recording overall mite infestation, hygiene, andthe specific hygienic response to V. destructor. We performed proteome-wide correlation analyses in larvalintegument and adult antennae, identifying several proteins highly predictive of behavior and reduced hiveinfestation. In the larva, response to wounding was identified as a key adaptive process leading to reducedinfestation, and chitin biosynthesis and immune responses appear to represent important disease resistantadaptations. The speed of hygienic behavior may be underpinned by changes in the antenna proteome, andchemosensory and neurological processes could also provide specificity for detection of V. destructor in antennae.Conclusions: Our results provide, for the first time, some insight into how complex behavioural adaptationsmanifest in the proteome of honey bees. The most important biochemical correlations provide clues as to theunderlying molecular mechanisms of social and innate immunity of honey bees. Such changes are indicative ofpotential divergence in processes controlling the hive-worker maturation.Keywords: Honey bee, Proteomics, Social immunity, Hygienic behavior, Varroa sensitive hygieneBackgroundSocial insects such as the honey bee (Apis mellifera L.)derive great benefit from living in tight-knit groups thatenable greater efficiencies in brood care, foraging anddefense against predation. However, the high populationdensities and relatedness of individuals leave colonies sus-ceptible to emerging infectious diseases [1]. Varroadestructor, an ectoparasitic mite of the honey bee [2]causes varroasis, which is a leading contributor to ongoingcolony losses in commercial apiculture worldwide [3].V. destructor feeds on the hemolymph of larval and adultbees, inflicting nutritional stress and immune suppression,as well as acting as a major vector for viral pathogentransmission [4].In solitary insects, cellular or humoral-based defensesprovide the only known system for immunity, butA. mellifera’s genome reveals that while honey bees con-tain these systems for immunity, the number of immu-nity genes is lower than that of solitary insects such as* Correspondence: foster@chibi.ubc.ca1University of British Columbia, Centre for High-Throughput Biology andDepartment of Biochemistry & Molecular Biology, 2125 East Mall, Vancouver,BC, V6T 14, CanadaFull list of author information is available at the end of the articleParker et al. Genome Biology 2012, 13:R81http://genomebiology.com/2012/13/9/R81© 2012 Parker et al.; licensee BioMed Central Ltd. This is an open access article distributed under the terms of the Creative CommonsAttribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction inany medium, provided the original work is properly cited.flies, moths and mosquitoes [5]. As an apparent com-pensation for this, social insects have evolved collectivesystems of behavior that provide defenses against diseaseand parasitism. Two related behaviors, hygienic behavior(HB) and Varroa sensitive hygiene (VSH), are highlyvariable among A. mellifera colonies and are seen asimportant traits in the development of disease and mite-resistant stock. HB is a well-documented protectivebehavior that involves nurse-aged worker bees uncap-ping brood cells and removing parasitized or diseasedpupae [6]. VSH is less well-understood but it encom-passes a suite of behaviors that ultimately suppress mitereproduction by uncapping and/or removing mite-infested pupae from sealed brood resulting in a highproportion of non-reproductive mites in the brood thatremains [7,8]. HB and VSH can be quantified using fieldassays and are heritable so, while both are now used inthe selective breeding of Varroa-resistant bees [9,10],the genetic and biochemical mechanisms that drivethem are poorly resolved.To date, most selective breeding in commercial apicul-ture focuses on traits such as honey production, color,gentleness, winter survival or other economic parameters.When combined with continual dilution of the gene poolthrough importation of susceptible stock, these selectionslimit host adaptation to pathogens. In order to improvedisease and mite tolerance, field assays for HB and VSHmust be incorporated into the stock selection process[11,12]; however, these assays are resource intensive, lacksensitivity and may require closed breeding [13], limitingtheir suitability for widespread application. To support thecreation of novel assays, a molecular-level mechanisticunderstanding of resistance traits is seen as a promisingavenue to support commercial breeding and disease pre-vention through marker-assisted selection (MAS) [14]. Todate, low-resolution microsatellite-based quantitative traitloci (QTL) for HB have been reported [15], as have someof the biochemical consequences to the host of infectionby V. destructor and associated viruses [16,17]. Transcrip-tome changes in A. mellifera and in Varroa’s natural hostA. cerana also pinpoint subtle changes in transcriptexpression for components responsible for neuronal rewir-ing, olfaction, metabolism and aspects of social behaviorthat may be critical components driving mechanisms ofVarroa tolerance [18,19].All the molecular investigations of HB and VSH haveused well-controlled colonies or individual bees withoutexamining the natural variation and distribution of boththe traits and their molecular components. Thus, here wetested the hypothesis that inter-colony variation in diseaseresistance parameters is reflected by changes in the expres-sion of specific proteins. Sampling from a large cohort ofcolonies, we measured the relative abundance of approxi-mately 1,200 proteins from two bee tissues involved ininteractions with the pathogens and correlated these withestimates for active bee behavioral phenotypes for HB andVSH, as well as host-pathogen population dynamics.Through meta-analysis of these data with other availableinformation, proteins and biochemical processes mostlikely to be responsible for the observed disease resistancetraits were identified.ResultsBehavioral phenotypes and mite-bee dynamicsTo assess the expression of disease tolerant behaviors,colony-level measurements of various metrics of HB,VSH and mite-host population dynamics were made(Figure 1), including phoretic infestation (PH, mites onadult bees), natural mite drop (ND, estimate of mitedeath rate), and levels of brood infestation (BI). HB wasestimated by observing the proportion of a defined num-ber of freeze-killed, sealed brood cells that bees firstuncapped (uncapped, U) and then removed (removed, R)at 24 and 48 hours (Figure 2a). The hygienic response tofreeze-killed brood was time dependent, with widely dis-tributed levels of HB at 24 hours and the majority ofhives achieving the accepted 95% threshold for the pro-portion of removed cells at 48 hours [11]. Because colonyscores in the other measured parameters were distributedroughly similarly to HB at 24 hours as indicated in thekernel density plot of Figure 2b, we asked which of themwere independent measures and which were interrelated.Pearson’s product-moment correlation (PPMC) coeffi-cient revealed a statistically significant negative pair-wisedependence between estimates for hygiene and miteinfestation dynamics (PH, ND) (Figure 2b). Maximumdependence was observed between cell removal and PH(r = -0.54, P = 0.0007), confirming that colonies betterable to detect and uncap cells with affected brood areable to reduce adult infestations more efficiently. Obser-ving differences in density for temporal aspects ofhygiene and interactions between disease-tolerant beha-viors and infestation indicated we were detecting naturalvariation in the speed of HB removal that directly influ-enced the colonies’ tolerance to Varroa mites.Correlation of protein expression and behaviorWhile most of the parameters discussed above are knownto have a genetic basis, all must ultimately manifest as theresult of changes in protein expression and/or activity.To explore potential mechanisms underpinning naturalvariation in Varroa tolerance across these colonies, weexamined the protein expression profiles of two tissuesthat play a critical role in the bee-Varroa interaction:antennae of brood-nest workers (that is, mostly nursebees) and the integument from fifth instar worker larvae.Antennae were used because they are adult bees’ primarysensory organs and many of the behaviors evaluated hereParker et al. Genome Biology 2012, 13:R81http://genomebiology.com/2012/13/9/R81Page 2 of 15involve bees being able to sense the presence of either thepathogen itself or a damaged/diseased nest-mate. Integu-ment was chosen because it is the initial physical barrierto Varroa when they feed on larvae and as such theinnate processes found here may be critical componentsin the response of hygienic adults and provide directinnate mechanisms of tolerance. It is possible thatchanges in the composition of larval proteins or themetabolites produced by these proteins during infectionmay trigger HB or VSH in adult nurses.Using liquid chromatography-tandem mass spectrome-try (LC-MS/MS) to analyze three independent samples ofeach tissue, we constructed protein expression profilesfor approximately 1,200 proteins across all colonies asdescribed previously [20]. By centering and standardizingacross labels and colonies, the relative expression ratiosfrom individual LC-MS/MS experiments are convertedinto a roughly normalized distribution of protein effect,representing the expression level of each protein in eachcolony relative to the population average. These variables(response variable) were then regressed against the beha-vior and infestation estimates (predictor variable) mea-sured for that colony. The direction of each regressionwas determined by the sign of the estimated regressioncoefficient and the significance of that effect was accessedusing a mixed linear model with probability cut-off at(a) Hygienic behavior (HB)UncappedRemoved(b) Varroa sensitive hygiene (VSH)(c) Brood infestation (BI)(d) Phoretic infestation (PI)(e) Natural drop (ND)Brood cellLarvaNursehoney beeInside the hiveVarroa miteReproductive Varroa miteFigure 1 Diagram depicting honey bee disease tolerant traits and infestation dynamics. (a) Hygienic behavior (HB) is composed of twocomponent behaviors, ‘uncapping’ (uncapped, U) which involves the opening of the cell containing a dead pupa and ‘removal’ (removed, R)which involves the removal of the dead pupa from the cell after uncapping has occurred. These behaviors are not always performed by thesame bee. HB was recorded over 24 hour (rapid) and 48 hour (slow) periods. (b) Varroa sensitive hygiene (VSH) was defined by determining theproportion of Varroa-infested cells in which no reproductively viable Varroa mite daughters were produced. Increases in this measure infer thatgreater proportions of mites remaining in the brood have had their reproduction suppressed because of infertility, death, the production of onlymales, or have had their reproduction delayed preventing sexual maturation of females. (c) Brood infestation (BI) is the percentage of brood cellsinfested by one or more mites regardless of the mite’s reproductive status. (d) Phoretic infestation (PI) is an estimate of the density of mitephoresy on adult bees, and (e) natural drop (ND) is a normalized measure of the number of mites falling from the adult bees onto an adhesiveboard on the bottom board of colonies.Parker et al. Genome Biology 2012, 13:R81http://genomebiology.com/2012/13/9/R81Page 3 of 15Phoretic InfestationBrood infestationNatural drop0.6 0.2610.170.941-0.21-0.19-0.131-0.10-0.28-0.34-0.101-0.02-0.45-0.540.110.441-1-0.8-0.6-0.4- sensitive hygiene** P<0.05Correlation coefficient (Peasons)Removed 24 hUncapped 24 hRemoved 24 hUncapped 24 hRemoved 48 hUncapped 48 h Varroa sensitive hygiene30405060708090100% CellsHigh h ygienic >95 % Low hygienic <50 %1201300.00 0.050.025DensityRemoved 24 hUncapped 24 h Removed 48 h Uncapped 48 h Varroa sensitive hygieneIntermediate hygienicPhoretic infestationBrood infestationRemoved 24 hUncapped 24 hNatural dropVarroa sensitive hygiene(a)(b)** **Figure 2 Distribution of and correlation between behavioral traits. (a) Box and whisker, and density plot for disease tolerant traits forhygiene at different times after the introduction of freeze killed brood and the measure for VSH. (b) Correlation matrix showing the correlationscore and significance for estimates of rapid hygienic behavior and VSH with measures of infestation. VSH, Varros sensitive hygiene.Parker et al. Genome Biology 2012, 13:R81http://genomebiology.com/2012/13/9/R81Page 4 of 15Q <.2 adjusting for multiple comparisons or later P < .05for explorative data analysis (see Figure 3).Several proteins are highly significant predictors ofresistance to Varroa mite infestationTo adjust significance levels to account for the multiple-testing hypothesis, proteins were filtered using Q <.2cut-off; for HB one antennal protein and five larval pro-teins survived this additional filter (Tables 1, 2). In theantennae, the hypothetical protein ‘LOC552009’ ofunknown function correlated with HB at 48 hours (Q =0.09) for both ‘uncapped’ and ‘removed’ behaviors.Sequence analysis revealed that LOC552009 contains aconserved domain similar to the mammalian proteinlipid transport protein Apolipoprotein O (ApoO) [21].Figure 4a, b shows the added variable plot for this pro-tein correlating with HB (removed 48 hours), peptidesidentified and protein sequence containing the con-served domain for ApoE.In larvae, several more candidate proteins were identi-fied as strong positive and negative predictors for HB(Table 2), suggesting that events in the larvae may beable to influence HB of the adult. Further correlationanalysis of mite infestation/fertility measures (PH, BI, NDand VSH) identified the hemocyte protein-glutaminegamma-glutamyltransferase-like (a putative transglutami-nase) as highly significant (Q = 0.02) and positively corre-lated with ND (Figure 4c).To increase the specificity of our measures for infestationdynamics, we next calculated the ratio of mites observedphoretically to those found in brood cells (PH/BI). Thisadjustment enabled quantitation of the relationshipbetween two important stages in the mite life cycle, wherelong phoretic phases may be indicative of poor reproduc-tive success and the influence of adult bee behavior or lar-val attractiveness. After adjustment, several proteins (ninein antenna, four in larva) were highly significant (Q <.2) inboth tissues. Importantly, the adjusted metric was alsohighly correlated with ND (Pearsons product correlation: t= 5.2211, df = 36, P = 7.633e-06) and in larva correlatedwith increased significance with the protein Tg (Q = 0.01),supporting the role of Tg as an accurate measure of Varroa0 10 20 30 40 50 60 70  P<0.05 Q<0.2 P<0.05 Q<0.2 Antenna Larva Number of proteins HB removed 24 h  HB removed 48 h  HB uncapped 24 h  HB uncapped 48 h  Varroa sensitive hygiene  Phoretic infestation/brood infestation  Natural drop  Figure 3 Results of proteome correlation screen for antennal and larval tissues. Bar plot giving the number of proteins found correlatingwith traits at different cut-off’s for statistical confidence.Parker et al. Genome Biology 2012, 13:R81http://genomebiology.com/2012/13/9/R81Page 5 of 15resistance (Figure 4, Tables 1, 2). Tg also provides theclearest link to phenotype. In insects, Tg is the primaryprotein facilitating the crosslinking of the clotting factors,hemolectin and Eig71Ee. Tg drives clot formation, animportant defensive process against ectoparasites such asthe Varroamite. Of the proteins correlating in the antenna,several form part of diverse metabolic pathways and areimpossible to link to function at this point. However, calcy-phosin-like protein and phenoloxidase subunit A3 areworthy of discussion. Calcyphosin (Q = 0.02) has beenfound expressed in olfactory cells in lobster and with aputative role in signal transduction in sensory cells. Pheno-loxidases (Q = 0.1) are a critical part of insect immunityduring the pathogen encapsulation response, but are alsoimportant in the ecdysteroid-dependent processes linkedto polyphenism and caste differentiation.Neuronal proteins underpin hygienic behavior and VSHin antennaeThe proteins discussed above that survive correction formultiple hypothesis testing should be excellent predictorsof HB, perhaps even usable in marker-assisted selectivebreeding. To find so many highly significant proteins in acompletely out-bred population is remarkable but therequirements that they must pass mean that it may betoo restrictive a dataset to understand fully some of themolecular mechanisms underlying the relevant behaviors.To this end, we expanded the analysis to discover pro-cesses with mechanistic relevance for HB and VSH.Those proteins with a significant correlation (P < .05) toone or more behavioral traits were explored using onto-logical classifications provided by the honey bee refseqentry in the National Center for Biotechnology Informa-tion (NCBI) or the flybase homolog. The most significantof these enrichments (P < .05, Table 3) largely trackedwith the altered distributions for estimates of HB, sug-gesting some of the molecular mechanisms that may reg-ulate this behavior. The set of proteins highly correlatedwith rapid HB (>95% removed by 24 hours) was particu-larly enriched for proteins involved in ‘sensory develop-ment’. At 24 hours, both uncapping and removal traitscorrelated with the up-regulation of the secretory pro-teins windbeutel, amphiphysin (Amph) and [RefSeq:CG6259] which encodes the homolog to human CHMP5protein. Proteins down regulated in rapid hygienic beeswere ankyrin 2, laminin A, Zasp (Z band alternativelyspliced PDZ-motif protein) and fasciclin 1 (Fas1) allinvolved in ‘cell adhesion’. Proteins correlating with bothTable 1 Name, expression and function of larval proteins found correlating with traits measured for resistance toVarroa infestation.Protein Name R48 U48 PH/BI ND Functionsimilar to CG5903-PA ↓ ↓ - - CDD:ApoE (Apolipoprotein, lipid transport)PREDICTED: polyadenylate-binding protein 1-like isoform 1 - - ↓ - mRNA surveillance pathwayPREDICTED: 60S ribosomal protein L14 - - ↓ - TranslationPREDICTED: alpha-tocopherol transfer protein-like ↓ Vitamin E distributionPREDICTED: alcohol dehydrogenase [NADP+] A-like ↓ Alcohol metabolismphenoloxidase subunit A3 ↑ Monooxygenase, phenoloxidase activityPREDICTED: dehydrogenase/reductase SDR family member 11-like - - ↑ - Oxidoreductase activity (Metabolism)PREDICTED: v-type proton ATPase subunit F 1-like ↑ Oxidative phosphorylation (Electrom transport)PREDICTED: calcyphosin-like protein-like - - ↑ - Intracellular signal transductionPREDICTED: sorbitol dehydrogenase-like isoform 2 - - ↑ - Fructose and mannose metabolismBI, brood infestation; ND, natural drop, PH phoretic infestation; R, removed; U, uncapped.Table 2 Name, expression and function of antennal proteins found correlating with traits measured for resistance toVarroa infestation.Protein name R48 U48 PH/BI ND Functioneukaryotic translation initiation factor 3 subunit A ↓ ↓ - - Mitotic spindle elongation; mitotic spindle organizationrab proteins geranylgeranyltransferase component A 1 ↓ ↓ - - Sensory transduction, intracellular transportargininosuccinate synthase-like ↓ - - Alanine, aspartate and glutamate metabolismglucose-6-phosphate 1-epimerase-like ↑ ↑ - - Glycolysis/GluconeogenesisUDP-glucose:glycoprotein glucosyltransferase ↑ - - - Protein glycosylationmitochondrial import inner membrane translocase subunit Tim8 - - ↓ - Protein import into mitochondrial inner membrane60S ribosomal protein L10 isoform 1 - - ↓ - n/a26S proteasome non-ATPase regulatory subunit 2 - - ↑ - Regulatory subunit of the 26 proteasomehemocyte protein-glutamine gamma-glutamyltransferase - - ↑ ↑ Peptide cross-linking, emolymph coagulationBI, brood infestation; ND, natural drop, PH phoretic infestation; R, removed; U, uncapped.Parker et al. Genome Biology 2012, 13:R81http://genomebiology.com/2012/13/9/R81Page 6 of 15elements of HB at 48 hours (>95% open/cleaned 48 hours) were enriched in the ontology ‘mitochondrial innermembrane’. This ontology was also identified for therapid uncapping behavior and corresponds to reductionin primary metabolic pathways.VSH and HB correlated strongly with some of the sameproteins, even though there was no apparentinterdependence between VSH and HB (Figure 5). Pro-teins that correlated with VSH and HB included Fas1,which was negatively correlated with rapid HB and VSH,while Amph and helicase 25E (Hel25E) levels were signif-icant positive predictors of both VSH and rapid HB(Figure 5). In addition, several unique proteins with rolesin synaptic function correlated only with VSH. VSH was−40 −30 −20 −10 0 10 20−1012Removed 48 hhypothetical protein ‘LOC552009’ (ApoE) å   −2−101234Natural dropTransglutaminase K (Tg)>gi|328780825|ref|XP_624392.3|_PREDICTED:_hypothetical_protein_LOC552009_isoform_1_[Apis_mellifera] MQRTLFFKKFLMPCGLCAAVPAMKPPIPEEHPAPCSNEIQGKKLIKPSELPIYSIDDGYT KQMPCIQYPSIVEDNIRKIRQTVSEIKLTIDKISRDISSSLESLKFISDYLQDQANLMPR IGAVGVGGLSGLILSLRGGIFKRLMYTTTGAAIVGCVCFPKETKETVNTMEHYGNVSYNF IYGVKPGDNKKEISFNEFPLVKSVLESEYFRMLVQFFEQKTNDTPTTTDVVLTDTTAKTE INKKKǨApoO conserved donmain Peptides identified(b)(a)(c)>gi|66510575|ref|XP_392939.2| PREDICTED: hemocyte protein-glutamine gamma-glutamyltransferase-like [Apis mellifera]MSREEPLVVEMVYLYEKENAKLHHTINYELVHLDPPAAVLRRGQSFHIALRFNREYIDEIDIVRLLFSFGPNPNVLRGTRGVNTITNRDSYLTDLEAWGVRLIGVSGVDLSAEVRSPVDSPVGMWQLNIETTIVGSKRSPNTYNYDKDIYLLFNPWLKEDLVYMEDEQLLDEYILNDVGKIWVGAWGSARGREWIFGQFDAYVLKACQLLLERSGIKANSRGDPIQMCRAISRIVNSNDDKGVVTGRWDGDYQDGTAPAAWTGSVPILEQFLETGESVKYGQCWVFAGVVTTVCRALGIPSRVVSNLVSAHDANASLSVDRYYSKENEELEYDPNNVEGEDSIWNYHVWNDVWMARPDLPKGYGGWQAIDSTPQEPSEGVYQCGPASVEAIKQGVVGYNYDVTFMLASVNADLMRWIEDPDSEMGFRKIDCNKYHIGRMILTKAPWVYDPNGDRDREDITSLYKAKEGTELERLTLYRAVRSTELAKRFYSLPSPAKEDVEFDLVDIERVNIGEPFAVIVNIKNKSNEKRTIQAILSAGSVYYKGIKAYLVKRASGDFVLEPYASEQLRLTITVDDYLDKLVEYCNMKLYSIATVVETKQTWADEDDFQVLKPNIVVKIDGEPTVGKPSIISLRFKNPLQRVLTDCKFNYAGPGLTRNKTLAFRDVDPEEDVYVEHQLIPQKAGSQKIIATFTSKELVDVTGSAVIDVLDMDE Domains Transglut (pfam00868) 4RANSGLUT?CORECL	Transglut_C ( pfam00927)Peptides identified Figure 4 Protein sequence, peptide identification coverage, and conserved domains for two proteins with highly significantcorrelation with Varroa tolerant traits. (a-b) Sequence encodes a probable ApoO type protein found to correlate with hygienic behavior at48 hours. (c-d) A transglutaminase (Tg) correlates with natural drops, an estimate of mite death rate in the colony. ApoO, apolipoprotein O.Table 3 Summary of gene enrichments for larval proteins correlating with anti-parasitic traits.Behavior Expression Enriched ontologiesR24 ↑7 Macromolecular localization (wbl, Amph, Hel25E, CG6259), protein localization, sensory organ development (wbl, Amph)↓13 tissue development, nerve growth factors (Ank2, LanA), axon guidance, cell adhesion (Ank2, LanA, Fas1, Zasp52, RpS25)U24 ↑11 Macromolecular localization, protein transport (wbl, Amph, Hel25E, CG6259), microtubule cytoskeletal organisation, cell cycle(RpS30, RpL12, Hel25E, Rab11)↓14 Cellular respiration (Gdh, ND42 CG5703 Cg6463), mitochodrial electron transport (Gdh, ND42 CG5703)R48 ↑9 Translation, mitotic spindle organization (Aats-cys, RpL21, Cctgamma)↓9 Mitochondrial inner membrane (sesB, ND42, CG5703, I(2)06225)U48 ↑12 No enrichment↓7 Mitochondrial inner membrane (ND42, CG5703)VSH ↑14 Ubiquitin-dependent protein catabolic process (Rad23, Prosalpha7)↓20 Cytokinesis (Gammacop, SNAP)R, removed; U, uncapped; VSH, Varroa sensitive hygiene.Parker et al. Genome Biology 2012, 13:R81http://genomebiology.com/2012/13/9/R81Page 7 of 15negatively correlated with expression of the mushroombody protein (Mub), soluble NSF attachment protein(SNAP) and gammaCOP. Mub is involved in temperaturepreference in Drosophila, [22] while SNAP is a key pre-synaptic protein mediating synaptic vesicle fusion andgammaCOP is involved in vesicle trafficking at synapsesand other vesicle sorting pathways [23]. The protein withgreatest change in expression with respect to VSH mea-surements was [RefSeq:LOC412768], a poorly annotatedmember of the take-out/juvenile hormone binding pro-tein (To/JHBP) superfamily involved in chemoreception.Larval proteins with cuticular and immune functioncorrelate with HB and VSHHB and VSH are considered to be specific behavioraladaptations of the adult honey bee to diseased brood.However, as our molecular understanding of these traitsimproves, it is possible that factors expressed within thelarva in response to the pathogens may influence themanifestation of the relevant behaviors. Protein/behaviorcorrelations of larval proteins revealed enrichment ofchitin-based cuticle structural proteins, particularly cuti-cular proteins 3 and 13 (CRP3 and CRP13; Table 4).Larval expression of the peptidyl-amino acid modifyingenzymes Caf1 and [RefSeq:CG6370] was also positivelycorrelated with HB. CG6370 encodes a dolichyl-dipho-sphooligosaccharide-protein glycosyltransferase involvedin N-glycan biosynthesis in the lumen of the endoplasmicreticulum and, interestingly, four other lumenal proteinswere also positively correlated. The larval protein/beha-vior correlation with the steepest slope was [RefSeq:CG1318-PA], a beta-N-acetyl-D-hexosaminidase withbroad substrate specificity ranging from N-glycans tochitooligosaccharides. Three proteins clustered under theterm carbohydrate metabolism were negatively correlatedwith VSH in larval tissue. Ecdysone-inducible gene(ImpL3), phosphoglucomutase (Pgm) and gene analogousto small peritrophins (Gasp) all process carbohydrates.Gasps specifically bind or regulate chitin structure indeveloping embryos, ImpL3 is a lactate dehydrogenaseinduced by the prohormone of 20-hydroxyecdysone thatregulates insect molting, while Pgm is a glycolyticenzyme, a process vital for the biosynthesis of chitin fromglycogen. Taken together, these observations suggest arole for larval chitin biosynthesis and/or structural regu-lation in hygienic and VSH behavior. In the proteinspositively correlated with VSH, gene enrichment identi-fied casein kinase II, alpha 1 polypeptide isoform 1(CkIIapha1 or CK2) and peptidoglycan-recognition pro-tein-SA (PGRP-SA). Both are components of the ‘cellsurface receptor linked signaling pathway’ ontology andeffectors of interferon and lipopolysaccharide (LPS)macrophage inflammatory signaling [24]. PGRP-SAdetects Lys-type peptidoglycan (PG) from gram-positivebacteria, leading to activation of the Toll receptor path-way and, ultimately, to increased expression of antimicro-bial peptides [25,26]. The larval inflammatory responsemay serve not only as an individual defense mechanismbut also as an initiator of social immunity behavior, thatis, VSH.DiscussionWe have described here the discovery of several proteinswhose expression levels may impact honey bee resistanceto infestation by the Varroa mite. Natural diversity inthese behaviors was a prerequisite to this study and weobserved that the levels of each behavior in any given col-ony were not random. As expected, there was a strongnegative correlation between mite infestation levels andHB. At the expression level, several proteins were highlysignificant predictors of HB and mite infestationdynamics. Highlighted within these proteins were theputative ApoO homolog and a putative Tg. Apolipopro-teins are called apolipophorins in insects, and they havediverse roles in lipid solubilization and the transport ofPhoretic infestationBrood infestationHB 24, removedHB 24, uncappedNatural dropVarroa sensitive hygieneCell adhesion Function0           T- Value                1wblHel25ECHMP5gammaCopSnapmubRpS25Fas1Zasp52Ank2LanAAmphMacromolecular localisationCytokinesisFigure 5 Heat map of t values for proteins with involvement inneuronal/synaptic function in antennae found to correlatesignificantly with hygienic behavior and VSH. VSH, Varroasensitive hygiene.Parker et al. Genome Biology 2012, 13:R81http://genomebiology.com/2012/13/9/R81Page 8 of 15small hydrophobic ligands [27-29]. In innate immunitythe apolipophorin ApoLp-III stimulates antimicrobialactivity in the hemolymph, acting as a pattern recogni-tion system for LPS and lipoteichoic acid (LTA) [29].Lastly, the strong correlation of Tg with both NDs andan increase in the ratio of phoretic mites to brood mitessuggests that Tg activity could provide a measure ofresistance to Varroa reproduction.V. destructor is an ecto-parasite feeding communallyand repeatedly on hemolymph of the honey bee througha bite wound in the cuticle [30-32]. In insects’ innateimmunity the cuticle provides the first line of defense;once breeched, innate defense systems of the haemocoelcavity are orchestrated by hemocytes, the fat body andhemocoel [33]. Normal wounds heal as hemocytes andplasmatocytes exocytose the clotting factors hemolectinand Eig71Ee [34]. These molecules and other plasma-based factors such as fondue are cross-linked by Tgs in aCa2+ dependent mechanism to form a primary clot. How-ever, V. destructor transmits bio-active compounds thatprevent healing and allow continued feeding to occur atthe same wound [35]. In the tick arthropod-mammalianecto-parasitic systems, 18 known bio-active suppressantstarget innate antiseptic defenses, including severalimmune cells types, inflammatory and coagulatory cas-cades [36]. In honeybees, the effect V. destructor elicitson the immune system is uncertain. Yang and Cox-Foster[37] demonstrated that Varroa parasitism increases thesusceptibility of adult bees to bacterial infection, but nomajor immunosuppressive effects were revealed by tran-scriptomic studies on specific immune genes or in globalanalyses [38,39]. More recently a study has reveled thatsalivary secretions from the Varroa mite are able todamage hemocyte aggregation in the tomato moth,(Lacanobia oleracea) [38] but no known factors of eitherpathogen or host are identified. We report here that ele-vated expression of a putative key clotting factor (Tg) isfound in the larva of Varroa resistant bee colonies. Thesedata indicate that honey bees have adapted to Varroa,increasing the clotting capacity of hemolymph in orderto limit mite reproduction.While the experiments described here were clearly ofsufficient power to permit the discovery of some correla-tions between protein expression and behavioral traits,the variability within such out-bred populations is veryhigh. This is likely a significant limitation in fully definingthe molecular mechanism of something as complex as abehavior. Practical limitations in the number of coloniesthat could be sampled and the depth to which the pro-teome could be measured across multiple samples wereinherent problems here, as with any proteomics study.Even so, an exploratory approach was seen as an impor-tant step in generating new hypotheses in a currentlypoorly understood area of biology.It is thought that the speed with which hygienic beesrespond is driven by a lower limit of olfactory detection ofthe diseased brood odor [40], which is in turn influencedby the neuromodulator octopamine [41]. In the antennallobe, octopamine concentration varies between behavioralstate, being low in nurse bees and high in foragers.Juvenile hormone and brood pheromone both modulatebehavioral responses to octopamine [42] and both areinvolved in several aspects of behavioral maturation, withthe best-understood system being the transition fromnurse to forager. This maturation invokes physiologicalchanges that are underpinned by increased neural proces-sing which is required to interpret complex visual infor-mation for flight behavior. Anatomically, expansion of themushroom body neurophil space in the brain and decreasein the volume of the olfactory glomeruli of the antennalTable 4 Summary of gene enrichments for antennal proteins correlating with anti-parasitic traits.Behavior Expression Larval Tissue Enriched Gene ontologiesR24 ↑18 Peptidyl-amino acid modification (caf1, Cg6370), structural constituent of chitin-based cuticle (Cpr65Av, Ccp84Ae),membrane enclosed lumen (CG1140, CG2118, Adk Caf1)↓16 Protein modification process (Rep, Art, Pp2A-29B), regulation of neuro transmitter release (Rep, beta-spec), phosphorousmetabolic process (Pp2A-29B, CG7712, PyK),U24 ↑18 Peptidyl-amino acid modification (caf1, Cg6370), cell division (Caf1, hts), cell periphery (Hexo1, NepYr)↓10 Protein modification process (Rep, Ugt, eff), regulation of neurotransmitter release (Rep, beta-spec), nervous systemdevelopment (beta-spec, eff, RpL30)R48 ↑12 Regulation of embryonic development (Hrb27C, Cg18811), oogenesis (hts, Hrb27C, nudC)↓18 Muscle tissue development (beta-Spec, alt, wupA), neurological system process (Rep, alt best-spec), cytoskeletal organization(Beta-spec, alt, cher, Rpl32, Rpl30, eIF3-S10)U48 ↑13 Regulation of embryonic development (Hrb27C, CG18811), oogenesis (nudC, hts, Hrb27C)↓17 Muscle tissue development (alt, wupA), ligase activity, forming carbon-nitrogen bonds (pug, CG1315)VSH ↑18 Cell surface receptor linked signaling pathway (PGRP-SA, CkIIapla), bristle development (ChIIalpha, CG12163) Translationelongation (Ef1beta, Ef2b)↓15 carbohydrate metabolic process (ImpL3, Pgm, Gasp)R, removed; U, uncapped; VSH, Varroa sensitive hygiene.Parker et al. Genome Biology 2012, 13:R81http://genomebiology.com/2012/13/9/R81Page 9 of 15lobes occurs during this transition [43]. Olfactory sensoryneurons from the antennae project onto the glomeruli ofthe antennal lobe via the antennal nerve, and olfactoryinformation is processed and projected to higher-orderbrain centers such as the mushroom bodies or lateralprotocerebrum.The data presented here indicates that cells (most likelyneurons with antennal axons) of bees performing rapidhygiene express different levels of proteins involved inadhesion and vesicle processing (Figure 6a), supportingthe role of octopamine and maturation as an importantcontrol of this behavior. The cell adhesion proteins identi-fied were all integrin proteins, some of which have beenreported to regulate synaptic plasticity [44]. Specifically,ankerin 2 stabilizes synaptic connections to the spectrin-actin cytoskeleton and laminin A, Zasp and Fas1 areinvolved in the assembly of functional integrin adhesionsites essential for growth cone extension in axon guidanceduring neurogensis [45]. The increased expression of vesi-cle sorting proteins in hygienic bees indicates that whileplasticity may be reduced, antennae of hygienic bees pro-vide a strong input into higher brain function. These datacould be explained by the environment of a hygienic nestbee, in which strong brood and queen-based olfactorycues are the major sensory inputs for bee development,behavior and social cohesion [43] (Figure 6d). Dimorphismin neural plasticity has been well characterized in theantennae of drones, where the antennal sensory nerves arethicker but project into a smaller number of glomerulithan in workers [46,47]. This configuration providesVarroa sensitive hygieneTo/JHBPReduced synaptic plasticityBehavioral maturationTo/JHBPOlfactionIncreased synaptic capacitywblCHMP5AmphNest beeForagerFas1 Ankerin2LamininA Zasp52Antennal lobeAntenal nerveCRP3glycosyltransferasehexosaminidasePGRP-SAChk1alpImmuneresponseIntegument signatureHygienic behaviourForaging olfactionVisual perceptionFlight controlNavigationDedicated to olfactionof brood signals(a)(c)CRP13(b)LarvaFigure 6 Summary model for proteins identified in antennae and larvae from correlation screen. For details see discussion.Parker et al. Genome Biology 2012, 13:R81http://genomebiology.com/2012/13/9/R81Page 10 of 15drones with the lower limit of detection for queen phero-mone, enabling efficient queen finding during matingflights.VSH limits mite reproductive success in the brood byspecifically detecting the presence of a post-ovipositionalmite. As part of the bees’ response, a sensitive adultuncaps and re-caps the cell, effectively inhibiting mitereproduction [48]. The signal being sensed in this processremains unknown, although it peaks between three andfive days after the cell is initially capped, leading to spec-ulation that VSH adult bees respond to temporal fluxesin pathology mediated by oviposition, wounding relatedstress responses, infections, and olfactory cues [48]. Cor-relation between VSH scores and two proteins encodingdivergent members of the To/JHBP super-family suggestthey may be functionally linked to the behavior; To/JHBPs contain a conserved ligand binding domain withdiffering affinities to small lipophilic molecules such asJH and the N-terminal signal peptide indicates that theyare probably secreted into the hemolymph where theyact as soluble receptors for their ligands [49,50]. In thehoney bee genome there are eight To/JHBP genes,located at two distinct loci, and we see one protein fromeach loci, one positively correlated with VSH and onenegatively correlated (Figure 6b). Biologically, this separa-tion in the genome suggests divergent functions and thisis further supported by their differential regulation in ourstudy. One of these proteins is completely uncharacter-ized but in Phormia regina the ortholog of the other To/JHBP is thought to be involved in chemosensation inantennal olfaction and taste [49] leading to the attractivehypothesis that it is playing a similar role in sensingbrood.That sensory and neuronal processes have a link to dis-ease tolerant behavior may be expected but, intriguingly,a class of proteins involved in larval cuticle formation/structure also emerged as likely candidates. An arthro-pod’s cuticle forms the primary physical barrier to theenvironment so while the cuticle plays an obvious role inpathogen defense, how it may contribute to social immu-nity mechanisms is less clear. Cuticular lipids differbetween bees depending upon caste and attacks by V.destructor can alter the composition in adults and larvae[51,52]. The role of the cuticle in social immunity is sup-ported by the data presented here, which indicates thatseveral proteins involved in forming and maintaining thecuticle are significantly correlated with disease tolerancebehaviors of nurse bees (Figure 6c).ConclusionsOur analysis of tissue proteomes from a large cohort ofcommercial honey bee colonies provides new clues tothe evolution of biochemical components facilitatingadaptation to disease. The control of behavior poten-tially represents the most complex paradigm in all livingcreatures so its study in natural, outbred systems isfraught with many difficulties, explaining the lack ofcoherent mechanisms describing these processes. Honeybees live in eusocial colonies and provide a scalable sys-tem for the study of developmental social biology andthe divisions of labor it defines. Our results representindications of molecular mechanisms underlying innateand social immunity behaviors in honey bees and buildupon previous work demonstrating adaption involvingneural remodeling and odorant recognition. A focusedinvestigation of the processes identified here will providean explanation of how host-pathogen interactions driveselection to generate disease tolerant colonies.Materials and methodsReagentsAll chemicals used were of analytical grade or better andall solvents were of HPLC-grade or better; all wereobtained from ThermoFisher-Scientific (St. Waltham,MA, USA).Honey bee-Varroa populations and physiologyWe established 40 genetically heterogeneous honey beecolonies at a research apiary (Grand Forks, BC, Canada)in the spring of 2009 by shaking workers into a largecage and then portioning them back into single Lang-stroth box colonies with nine frames in each. Selectedqueens were then introduced into each new colony withinitial populations of 1 kg of bees with relatively uni-form V. destructor infestation rates, varying among colo-nies from 6.2% to 7.6% per 100 adult bees. Colonieswere allowed to develop for six weeks to allow workerpopulations to turn over and be composed of the intro-duced queen’s progeny, at which point we evaluatedeach for physiological V. destructor interactions and HB.HB was measured as the proportion of sealed broodcells uncapped (U) and removed of pupae (R) within 24and 48 hours of freeze-killing defined patches broodwith liquid nitrogen; PH and BI were estimated asdescribed [53]. The proportion of uncapped cellsreferred to all cells uncapped by nurse bees includingthose where the pupae had been removed and thosewhere the pupae was still present at the time of theobservation. To estimate the ND V. destructor in eachcolony we counted the number of mites captured onscreened bottom boards over four 24-hour collectionperiods spanning a period of 10 days. The ND estimateswere normalized by colony size using the total weight ofbees to determine the number of bees in each colony.VSH was estimated as the production of sexually viablefemale offspring as described [12].Parker et al. Genome Biology 2012, 13:R81http://genomebiology.com/2012/13/9/R81Page 11 of 15Analysis matrixWe used triplex dimethylation labeling and generated aD-optimal design matrix to group the samples in blocksof three and assigned a label to each sample as described[20]. A randomized incomplete block design similar towhat we have used previously was chosen to minimizethe standard error of the estimate of the colony effect onprotein expression level [20].Sample collection and protein preparationThe antennae and larvae from colonies were sampled intriplicate. Ten pairs of antennae from nurse bees andthree fifth instar larvae were removed in situ and frozenon dry ice. Larvae were further dissected to remove thedigestive tracts and free-flowing hemolymph with stabi-lity maintained in PBS (50 mM K2HPO4, 150 mM NaCl,pH 7.4) containing complete, EDTA-free protease inhibi-tor cocktail (Roche, Mississauga, ON, Canada). Bothsamples were washed three times in PBS and preparedusing essentially the same method. Both tissues werehomogenized in 50 mM Tris-HCl, 150 mM NaCl, 1%NP-40, 20 mM dithiolthretitol in a Fast Prep bead millwith 2.8 mm ceramic beads (MP Biomedicals, Santa Ana,CA, USA) using 1 or 3 cycles of 20 s at 6.5 M/s withcooling for larvae or antennae, respectively. Tissue lysateswere clarified at 5,000 relative centrifugal force (rcf) for5 min at 4°C before ethanol/sodium acetate precipitation[54]. Proteolytic digestion of 20 μg (antennal) and 50 μg(larval) total protein was carried out as described [54])and samples were labeled by reductive dimethylationusing formaldehyde isotopologues [55] with slight modi-fications [20]. After labeling, each sample was pooled asrequired by the experimental design and each pool wasseparated into five fractions (antennae) using C18-SCX-C18 STAGE tips [56] or into six fractions (larvae) byisoelectric focusing with the OFFGEL system (AgilentTechnologies, Santa Clara, CA, USA) [57].Proteome screenQuantitative proteomic datasets were generated for anten-nae exactly as described in [20]. For larval tissue, LC-MSwas done on a 1200 Series nanoflow HPLC system (Agi-lent Technologies) interfaced with a chromatin immuno-precipitation (CHIP)-cube to a 6520 Q-TOF (AgilentTechnologies). Peptide separation was performed byreversed phase chromatography using a micro-fluidicCHIP comprised of an analytical column (75 μm ID, 150mm length with a 300 Å C18 stationary phase) and a 160nL trap column of the same phase. Peptides were loadedin 5% (v/v) acetonitrile, 0.1% (v/v), formic acid at 0.3 μL/min and then resolved at 0.3 μL/min for 90 min, duringwhich a linear gradient of acetonitrile was created from5% to 50% in 0.1% (v/v) formic acid. Mass spectrometry:Operating in auto MS/MS acquisition mode, the Q-TOFwas set up to acquire full scan data over a mass range of350 to 2,000 m/z and MS2 for the six most intense, multi-ply-charged ions. Peak lists were created using SpectrumMill extractor specifying fixed modification carboxyamido-methylation (C) and triplex dimethyl-mix (K,N-term)which accounts for all possible label moieties. Scans weremerged within +/- 45 s elution time and maximum m/zwindow of 0.5 Da (usually 20 ppm). For database search-ing, oxidation (M) was added as a variable modification,peptide tolerance was +/- 20 ppm and fragment ion toler-ance was 50 ppm, dynamic peak thresholding wasswitched on. Search results from Spectrum Mill were vali-dated using autovalidation for protein score >20; (charge2, Score >5, % SPI <50, Fwd - Rev Score >1, Rand 1-2score >1), (charge 1, Score >6, % SPI <60, Fwd - Rev Score>1, Rand 1-2 score >1), (charge 3, Score >5, % SPI <50,Fwd - Rev Score >1, Rand 1-2 score >1) and for peptidehits (charge 2, Score >12, % SPI <60, Fwd - Rev Score >2,Rand 1-2 score >2), (charge 1, Score >8, % SPI <70, Fwd -Rev Score >2, Rand 1-2 score >2), (charge 2, Score >12, %SPI <60, Fwd - Rev Score >2, Rand 1-2 score >2). Both lar-vae and antennae datasets have peptide identification falsediscovery rates (FDR) well below 1% and were compiledinto protein data arrays as described [20]. All experimentaldesign and proteomic results are listed in Additional files1, 2, 3 and 4.Data availabilityAll MS/MS data used in this study have been madeavailable in two locations: they have all been depositedinto the Honey Bee PeptideAtlas [58,59] as processedspectra and the raw files themselves are available on ourFTP site (ftp://foster.chibi.ubc.ca/Downloads/BeeBiomar-kers/).Statistical analysisLogarithms of intensities were normalized by first sub-tracting the average of the three measurements in eachblock (for each protein independently) and then center-ing and standardizing within each label (across proteins)by the median and median absolute deviation [20]. Foreach protein, a linear mixed effects model was used toestimate the effect of each predictor variable on the pro-tein expression level, adjusting for block and label fac-tors. Colony was treated as a random factor to controlfor the three repeated measures within each colony. Foreach predictor variable an estimated effect, standarderror and P-value was computed for each proteinresponse. FDRs (q-values) were computed for the set ofP-values of a given predictor over all protein responsevariables to adjust for multiple comparisons. All calcula-tions were performed in the R statistical language.Parker et al. Genome Biology 2012, 13:R81http://genomebiology.com/2012/13/9/R81Page 12 of 15Gene enrichment analysisAnalysis was performed using Exploratory Gene Associa-tion Networks (EGAN) software [60] with pre collatednetworks for Drosophila melanogaster (Dmel). A. mellifera(Amel) gene identifiers were mapped to Dmel orthologsusing the Round-Up database [61] and any unmappedAmel identifiers were assigned functions based on theirclosest homolog in D. melanogaster using BLAST-P,resulting in a total of 90% coverage for all antennal pro-teins identified. The remaining 10% were dealt with manu-ally by drawing information from several sources: 1) honeybee genes othologs implicated in immunity [61]; 2) pro-teins found significantly regulated in response to bacterialinfection by Paenibacillus larvae [63]; 3) proteins regu-lated in response to V. destructor infestation [18]; and 4)proteins specific to colony collapse disorder (CCD)affected colonies [64]. Enrichment analysis of proteinswhose expression levels correlated (P < .05) with behaviorand their direction of regulation was carried out byintegrating these nodes with gene ontology nodes forcomponent, function and process. For each node, over-representation or enrichment analysis was carried outemploying a standard one-tailed Fisher’s exact (hypergeo-metric) test using the entire gene list as background. Heatmaps representing the interaction of important genes andtheir relationships by nodes were generated for both data-sets independently.Additional materialAdditional file 1: Spreadsheet of five pages for larval dataset. Page1, P values and protein annotation. For all proteins identified andquantitated the resultant P value from the correlation analysis with eachtrait and refseq annotation. Page 2, T statistic for each correlationprovides a standardized value for the rate of change for each proteinagainst per unit of each trait measured. Page 3, EGAN mapping, honeybee protein GI mapped to NCBI fly gene id. Page 4, P values forbiological process gene enrichment for protein groups created by traitcorrelation and direction of correlation. Page 5, P values for globalbiological process gene enrichments.Additional file 2: Spreadsheet of five pages for antenna dataset.Page 1, P values and protein annotation. For each protein identified theresultant P value from the correlation analysis with each trait and refseqannotation is provided. Page 2, T statistic for each correlation provides astandardized value for the rate of change for each protein against perunit of each trait measured. Page 3, EGAN mapping, honey bee proteinGI mapped to NCBI fly gene id. Page 4, P values for biological processgene enrichment for protein groups created by trait correlation anddirection of correlation. Page 5, P values for global biological processgene enrichments.Additional file 3: Spreadsheet of five pages for processed antennaLC-MS data. Page 1 gives the experimental design designating used togenerate LC-MS datasets. Each block is shown with colony numbers andthe output file name generated for each block. Pages 2 to 4 areexamples of the output from LC-MS and protein identification andquantitification result for each peptide used to generate proteins ratios.Page 5 is the result of rolling all peptide identification up into uniqueproteins using parsimony rules.Additional file 4: Spreadsheet of five pages for processed larval LC-MS data. Page 1 gives the experimental design designating used togenerate LC-MS datasets. Each block is shown with colony numbers andthe output file name generated for each block. Pages 2 to 4 areexamples of the output from LC-MS and protein identification andquantitification result for each peptide used to generate proteins ratios.Page 5 is the result of rolling all peptide identification up into uniqueproteins using parsimony rules.AbbreviationsApo: apolipoprotein; BI: brood infestation; CCD: colony collapse disorder;CHIP: chromatin immunoprecipitation; Fas1: fasciclin 1; FDR: false discoveryrate; Gasp: gene analogous to small peritrophins; HB: hygienic behavior;HPLC: high pressure liquid chromatography; JH: juvenile hormone; LC: liquidchromatography; LPS: lipopolysaccharide; LTA: lipoteichoic acid; MS: massspectrometry; ND: natural drop; PBS: phosphate buffered saline; Pgm:phosphoglucomutase; PG: petidoglycan; PGRP-SA: peptidoglycan recognitionprotein SA; PH: phoretic; PPMC: Pearson’s Product movement correlation;QTL: quantitative trait loci; RCF: relative centrifugal force; STAGE: stop and goextraction; To/JHBP: take-out/juvenile hormone binding protein; VSH: Varroasensitive hygiene; Zasp: Z band alternately spliced PDZ-motif protein.AcknowledgementsThe authors wish to thank Nikolay Stoynov for technical assistance in LC/MSanalysis, Julian Yiu and Tram Nguyen for help with sample collection, as wellas members of our respective groups for advice and fruitful discussions. Thiswork was supported by funding from Genome Canada, Genome BritishColumbia, the British Columbia Honey Producers Association through theBoone-Hodgson-Wilkinson Trust Fund, the Canadian Honey Council andCanadian Association of Professional Apiculturists through the Canadian BeeResearch Fund, the British Columbia Blueberry Council and the BritishColumbia Cranberry Marketing Association. Mass spectrometry infrastructureused in this project was supported by the Canada Foundation forInnovation, the British Columbia Knowledge Development Fund and theBritish Columbia Proteomics Network (BCPN). LJF is the Canada ResearchChair in Quantitative Proteomics.Author details1University of British Columbia, Centre for High-Throughput Biology andDepartment of Biochemistry & Molecular Biology, 2125 East Mall, Vancouver,BC, V6T 14, Canada. 2Agriculture & Agri-Food Canada, Beaverlodge ResearchFarm, PO Box 29, Beaverlodge, AB, T0H 0C0, Canada. 3University of BritishColumbia, Department of Statistics, 2207 Main Mall, Vancouver, BC, V6T 1Z4,Canada. 4Kettle Valley Queens, 4880 Well Rd., Grand Forks, BC, V0H 1H5,Canada.Authors’ contributionsAPM, SFP, MMG, LH and LJF designed the experiments. LH maintained thebees and was assisted in performing behavioral assays by APM and SFP. LJF,KM, MMG and RP dissected the bees. RP and KM performed thebiochemistry and mass spectrometry, as well as the functionalbioinformatics. LJF authored the scripts for compiling proteomic data. RWdesigned and performed the statistical analyses. RP, LJF and MMGinterpreted the results and wrote the initial version of the manuscript. Allauthors have read and approved the manuscript for publication.Competing interestsThe authors declare that they have no competing interests.Received: 14 April 2012 Revised: 29 June 2012Accepted: 28 September 2012 Published: 28 September 2012References1. Evans JD, Spivak M: Socialized medicine: individual and communaldisease barriers in honey bees. J Invertebr Pathol 2010, 103:S62-S72.2. Anderson DL, Trueman JWH: Varroa jacobsoni (Acari: Varroidae) is morethan one species. Exp Appl Acarol 2000, 24:165-189.3. Le Conte Y, Ellis M, Ritter W: Varroa mites and honey bee health: canVarroa explain part of the colony losses? Apidologie 2010, 41:353-363.Parker et al. Genome Biology 2012, 13:R81http://genomebiology.com/2012/13/9/R81Page 13 of 154. Rosenkranz P, Aumeier P, Ziegelmann B: Biology and control of Varroadestructor. J Invertebr Pathol 2010, 103:S96-S119.5. Evans JD, Aronstein K, Chen YP, Hetru C, Imler JL, Jiang H, Kanost M,Thompson GJ, Zou Z, Hultmark D: Immune pathways and defencemechanisms in honey bees Apis mellifera. Insect Mol Biol 2006, 15:645-656.6. Rothenbuhler WC: Behavior genetics of nest cleaning in honey bees. IV.Responses of Fx and backcross generations to disease-killed brood. AmZool 1964, 4:111-123.7. Harbo J, Harris J: Suppressed mite reproduction explained by thebehaviour of adult bees. J Apicultural Res 2005, 44:21-23.8. Ibrahim A, Spivak M: The relationship between hygienic behavior andsuppression of mite reproduction as honey bee (Apis mellifera)mechanisms of resistance to Varroa destructor. Apidologie 2006, 37:31-40.9. Spivak M, Reuter GS: Performance of hygienic honey bee colonies in acommercial apiary. Apidologie 1998, 29:291-302.10. Harbo JR, Harris JW: Responses to Varroa by honey bees with differentlevels of Varroa Sensitive Hygiene. J Apicultural Res 2009, 48:156-161.11. Spivak M, Downey DL: Field assays for hygienic behavior in honey bees(Hymenoptera: Apidae). J Econ Entomol 1998, 91:64-70.12. Harbo JR, Harris JW: Resistance to Varroa destructor (Mesostigmata:Varroidae) when mite-resistant queen honey bees (Hymenoptera:Apidae) were free-mated with unselected drones. J Econ Entomol 2001,94:1319-1323.13. Pernal S, Sewalem A, Melathopoulos A: Breeding for hygienic behaviour inhoneybees (Apis mellifera ) using free-mated nucleus colonies.Apidologie 2012, 43:403-424.14. Rinderer TE, Harris JW, Hunt GJ, de Guzman LI: Breeding for resistance toVarroa destructor in North America. Apidologie 2010, 41:409-424.15. Oxley PR, Spivak M, Oldroyd BP: Six quantitative trait loci influence taskthresholds for hygienic behaviour in honeybees (Apis mellifera). Mol Ecol2010, 19:1452-1461.16. Gregory PG, Evans JD, Rinderer T, de Guzman L: Conditional immune-genesuppression of honeybees parasitized by Varroa mites. J Insect Sci 2005,5:7.17. Navajas M, Migeon A, Alaux C, Martin-Magniette ML, Robinson GE,Evans JD, Cros-Arteil S, Crauser D, Le Conte Y: Differential gene expressionof the honey bee Apis mellifera associated with Varroa destructorinfection. BMC Genomics 2008, 9:301.18. Zhang Y, Liu X, Zhang W, Han R: Differential gene expression of thehoney bees Apis mellifera and A. cerana induced by Varroa destructorinfection. J Insect Physiol 2010, 56:1207-1218.19. Le Conte Y, Alaux C, Martin JF, Harbo JR, Harris JW, Dantec C, Severac D,Cros-Arteil S, Navajas M: Social immunity in honeybees (Apis mellifera):transcriptome analysis of varroa-hygienic behaviour. Insect Mol Biol 2011,20:399-408.20. Parker R, Melathopoulos AP, White R, Pernal SF, Guarna MM, Foster LJ:Ecological adaptation of diverse honey bee (Apis mellifera) populations.Plos One 2010, 5:e11096.21. Lamant M, Smih F, Harmancey R, Philip-Couderc P, Pathak A, Roncalli J,Galinier M, Collet X, Massabuau P, Senard JM, Rouet P: ApoO, a novelapolipoprotein, is an original glycoprotein up-regulated by diabetes inhuman heart. J Biol Chem 2006, 281:36289-36302.22. Hong S-T, Bang S, Hyun S, Kang J, Jeong K, Paik D, Chung J, Kim J: cAMPsignalling in mushroom bodies modulates temperature preferencebehaviour in Drosophila. Nature 2008, 454:771-775.23. Yanay C, Morpurgo N, Linial M: Evolution of insect proteomes: insightsinto synapse organization and synaptic vesicle life cycle. Genome Biol2008, 9:R27.24. Jung D-H, Park H-J, Byun H-E, Park Y-M, Kim T-W, Kim B-O, Um S-H, Pyo S:Diosgenin inhibits macrophage-derived inflammatory mediators throughdownregulation of CK2, JNK, NF-kappa B and AP-1 activation. IntImmunopharmacol 2010, 10:1047-1054.25. Bischoff V, Vignal C, Boneca IG, Michel T, Hoffmann JA, Royet J: Function ofthe drosophila pattern-recognition receptor PGRP-SD in the detection ofGram-positive bacteria. Nat Immunol 2004, 5:1175-1180.26. Michel T, Reichhart JM, Hoffmann JA, Royet J: Drosophila Toll is activatedby Gram-positive bacteria through a circulating peptidoglycanrecognition protein. Nature 2001, 414:756-759.27. van der Horst DJ, van Hoof D, van Marrewijk WJ, Rodenburg KW:Alternative lipid mobilization: the insect shuttle system. Mol Cell Biochem2002, 239:113-119.28. Li WH, Tanimura M, Luo CC, Datta S, Chan L: The apolipoproteinmultigene family: biosynthesis, structure, structure-functionrelationships, and evolution. J Lipid Res 1988, 29:245-271.29. Whitten MM, Tew IF, Lee BL, Ratcliffe NA: A novel role for an insectapolipoprotein (apolipophorin III) in beta-1,3-glucan pattern recognitionand cellular encapsulation reactions. J Immunol 2004, 172:2177-2185.30. Garedew A, Schmolz E, Lamprecht I: The energy and nutritional demand ofthe parasitic life of the mite Varroa destructor. Apidologie 2004, 35:419-430.31. de D’Aubeterre JP, Myrold DD, Royce LA, Rossignol PA: A scientific note ofan application of isotope ratio mass spectrometry to feeding by themite, Varroa jacobsoni Oudemans, on the honeybee, Apis mellifera L.Apidologie 1999, 30:351-352.32. Tewarson NC, Singh A, Engels W: Reproduction of Varroa-Jacobsoni incolonies of Apis cerana-indica under natural and experimentalconditions. Apidologie 1992, 23:161-171.33. Feldhaar H, Gross R: Immune reactions of insects on bacterial pathogensand mutualists. Microbes Infect 2008, 10:1082-1088.34. Dushay MS: Insect hemolymph clotting. Cell Mol Life Sci 2009,66:2643-2650.35. Kanbar G, Engels W: Ultrastructure and bacterial infection of wounds inhoney bee (Apis mellifera) pupae punctured by Varroa mites. ParasitolRes 2003, 90:349-354.36. Jensen K, de Miranda Santos IK, Glass EJ: Using genomic approaches tounravel livestock (host)-tick-pathogen interactions. Trends Parasitol 2007,23:439-444.37. Yang X, Cox-Foster DL: Impact of an ectoparasite on the immunity andpathology of an invertebrate: evidence for host immunosuppressionand viral amplification. Proc Natl Acad Sci USA 2005, 102:7470-7475.38. Richards EH, Jones B, Bowman A: Salivary secretions from the honeybeemite, Varroa destructor: effects on insect haemocytes and preliminarybiochemical characterization. Parasitology 2011, 138:602-608.39. Navajas M, Migeon A, Alaux C, Martin-Magniette M, Robinson G, Evans J,Cros-Arteil S, Crauser D, Le Conte Y: Differential gene expression of thehoney bee Apis mellifera associated with Varroa destructor infection.BMC Genomics 2008, 9:301.40. Swanson JAI, Torto B, Kells SA, Mesce KA, Tumlinson JH, Spivak M:Odorants that induce hygienic behavior in honeybees: identification ofvolatile compounds in chalkbrood-infected honeybee larvae. J Chem Ecol2009, 35:1108-1116.41. Spivak M, Masterman R, Ross R, Mesce KA: Hygienic behavior in the honeybee (Apis mellifera L.) and the modulatory role of octopamine. JNeurobiol 2003, 55:341-354.42. Schulz DJ, Barron AB, Robinson GE: A role for octopamine in honey beedivision of labor. Brain Behav Evol 2002, 60:350-359.43. Fahrbach SE, Robinson GE: Juvenile hormone, behavioral maturation, andbrain structure in the honey bee. Dev Neurosci 1996, 18:102-114.44. Koch I, Schwarz H, Beuchle D, Goellner B, Langegger M, Aberle H:Drosophila ankyrin 2 is required for synaptic stability. Neuron 2008,58:210-222.45. Dent EW, Gupton SL, Gertler FB: The growth cone cytoskeleton in axonoutgrowth and guidance. Cold Spring Harb Perspect Biol 2011, doi: 10.1101/cshperspect.a001800.46. Arnold G, Masson C, Budharugsa S: Comparative study of the antennallobes and their afferent pathway in the worker bee and the drone (Apismellifera). Cell Tissue Res 1985, 242:593-605.47. Brockmann A, Bruckner D: Structural differences in the drone olfactorysystem of two phylogenetically distant Apis species, A-florea and A-mellifera. Naturwissenschaften 2001, 88:78-81.48. Harris JW: Bees with Varroa Sensitive Hygiene preferentially remove miteinfested pupae aged <= five days post capping. J Apicultural Res 2007,46:134-139.49. Fujikawa K, Seno K, Ozaki M: A novel Takeout-like protein expressed inthe taste and olfactory organs of the blowfly, Phormia regina. FEBS J2006, 273:4311-4321.50. Hagai T, Cohen M, Bloch G: Genes encoding putative Takeout/juvenilehormone binding proteins in the honeybee (Apis mellifera) andmodulation by age and juvenile hormone of the takeout-like geneGB19811. Insect Biochem Mol Biol 2007, 37:689-701.51. Del Piccolo F, Nazzi F, Della Vedova G, Milani N: Selection of Apis melliferaworkers by the parasitic mite Varroa destructor using host cuticularhydrocarbons. Parasitology 2010, 137:967-973.Parker et al. Genome Biology 2012, 13:R81http://genomebiology.com/2012/13/9/R81Page 14 of 1552. Salvy M, Martin C, Bagneres AG, Provost E, Roux M, Le Conte Y, Clement JL:Modifications of the cuticular hydrocarbon profile of Apis melliferaworker bees in the presence of the ectoparasitic mite Varroa jacobsoniin brood cells. Parasitology 2001, 122:145-159.53. Harris JW, Harbo JR, Villa JD, Danka RG: Variable population growth ofVarroa destructor (Mesostigmata: Varroidae) in colonies of honey bees(Hymenoptera: Apidae) during a 10-year period. Environ Entomol 2003,32:1305-1312.54. Foster LJ, de Hoog CL, Mann M: Unbiased quantitative proteomics oflipid rafts reveals high specificity for signaling factors. Proc Natl Acad SciUSA 2003, 100:5813-5818.55. Boersema PJ, Aye TT, van Veen TAB, Heck AJR, Mohammed S: Triplexprotein quantification based on stable isotope labeling by peptidedimethylation applied to cell and tissue lysates. Proteomics 2008,8:4624-4632.56. Ishihama Y, Rappsilber J, Mann M: Modular stop and go extraction tipswith stacked disks for parallel and multidimensional peptidefractionation in proteomics. J Proteome Res 2006, 5:988-994.57. Rogers LD, Fang Y, Foster LJ: An integrated global strategy for cell lysis,fractionation, enrichment and mass spectrometric analysis ofphosphorylated peptides. Mol Biosyst 2010, 6:822-829.58. Peptide Atlas. [http://www.peptideatlas.org/builds/honeybee/index.php].59. Chan QWT, Parker P, Sun Z, Deutsch EW, Foster LJ: A honey bee (Apismellifera L.) PeptideAtlas crossing castes and tissues. BMC Genomics2011, 12:290.60. Paquette J, Tokuyasu T: EGAN: exploratory gene association networks.Bioinformatics 2010, 26:285-286.61. DeLuca TF, Wu IH, Pu J, Monaghan T, Peshkin L, Singh S, Wall DP:Roundup: a multi-genome repository of orthologs and evolutionarydistances. Bioinformatics 2006, 22:2044-2046.62. Rannikko K, Ortutay C, Vihinen M: Immunity genes and their orthologs: amulti-species database. Int Immunol 2007, 19:1361-1370.63. Chan QWT, Melathopoulos AP, Pernal SF, Foster LJ: The innate immuneand systemic response in honey bees to a bacterial pathogen,Paenibacillus larvae. BMC Genomics 2009, 10:387.64. Johnson RM, Evans JD, Robinson GE, Berenbaum MR: Changes in transcriptabundance relating to colony collapse disorder in honey bees (Apismellifera). Proc Natl Acad Sci USA 2009, 106:14790-14795.doi:10.1186/gb-2012-13-9-r81Cite this article as: Parker et al.: Correlation of proteome-wide changeswith social immunity behaviors provides insight into resistance to theparasitic mite, Varroa destructor, in the honey bee (Apis mellifera).Genome Biology 2012 13:R81.Submit your next manuscript to BioMed Centraland take full advantage of: • Convenient online submission• Thorough peer review• No space constraints or color figure charges• Immediate publication on acceptance• Inclusion in PubMed, CAS, Scopus and Google Scholar• Research which is freely available for redistributionSubmit your manuscript at www.biomedcentral.com/submitParker et al. Genome Biology 2012, 13:R81http://genomebiology.com/2012/13/9/R81Page 15 of 15


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items