The draft genome, transcriptome, and microbiome of Dermatophagoides farinae reveal a broad spectrum

Thedraftgenome,transcriptome,andmicrobiomeofDermatophagoidesfarinaerevealabroadspectrumofdustmiteallergens

Ting-FungChan,PhD,a*Kun-MeiJi,PhD,b*AldrinKay-YuenYim,BS,a,c*Xiao-YuLiu,MSc,d*Jun-WeiZhou,PhD,e*àRui-QiLi,BSM,d*KevinYiYang,PhD,c,e*JingLi,MS,f*MengLi,MSc,dPatrickTik-WanLaw,PhD,aYu-LanWu,BS,dZe-LangCai,MSc,dHaoQin,PhD,a,cYingBao,MSc,dRossKa-KitLeung,PhD,c,ePatrickKwok-ShingNg,PhD,eJuZou,MSc,dXiao-JunZhong,MSc,dPi-XinRan,MD,fNan-ShanZhong,MSc,fZhi-GangLiu,MD,b,dandStephenKwok-WingTsui,PhDc,eHongKong,Shenzhen,andGuangzhou,ChinaBackground:Asequencedhousedustmite(HDM)genomewouldadvanceourunderstandingofHDMallergens,acommoncauseofhumanallergies.

Objective:Wesoughttoproduceanannotated

Dermatophagoidesfarinaedraftgenomeanddevelopacombinedgenomic-transcriptomic-proteomicapproachforelucidationofHDMallergens.

Methods:ADfarinaedraftgenomeandtranscriptomewereassembledwithhigh-throughputsequencing,accommodatingmicrobiomesequences.TheallergengenestructureswerevalidatedbymeansofSangersequencing.Themite’s

microbiomecompositionwasdetermined,andthepredominantgenuswasvalidatedimmunohistochemically.Theallergenicityofaubiquinol-cytochromecreductasebindingprotein

homologuewasevaluatedwithimmunoblotting,immunosorbentassays,andskinpricktests.

Results:Thefullgenestructuresof20canonicalallergensand7noncanonicalallergenhomologueswereproduced.Anovelmajorallergen,ubiquinol-cytochromecreductasebindingprotein–likeprotein,wasfoundanddesignatedDerf24.All40serasamplesfrompatientswithmiteallergyhadIgEantibodiesagainstrDerf24.Of10patientstested,5hadpositiveskinreactions.Thepredominantbacterialgenusamong100identi?edspecies

FromatheSchoolofLifeSciences,ctheHongKongBioinformaticsCentre,andetheSchoolofBiomedicalSciences,ChineseUniversityofHongKong;btheStateKeyLaboratoryofRespiratoryDiseaseforAllergyatShenzhenUniversityanddtheShenz-henKeyLaboratoryofAllergyandImmunology,SchoolofMedicine,ShenzhenUni-versity;andftheStateKeyLaboratoryofRespiratoryDisease,theFirstAf?liatedHospitalofGuangzhouMedicalUniversity.*Theseauthorscontributedequallytothiswork.

àJune-WeiZhou,PhD,iscurrentlyaf?liatedwiththeBasicMedicalCollegeofZhengzhouUniversity,Zhengzhou,China.

Disclosureofpotentialcon?ictofinterest:Theauthorsdeclarethattheyhavenorelevantcon?ictsofinterest.

ReceivedforpublicationJune29,2013;revisedSeptember3,2014;acceptedforpubli-cationSeptember25,2014.

Correspondingauthor:Nan-ShanZhong,MSc,StateKeyLaboratoryofRespiratoryDis-ease,theFirstAf?liatedHospitalofGuangzhouMedicalCollegeNo.151,YanjiangRd,Guangzhou510120,China.E-mail:nanshan@http://wendang.chazidian.com.Or:Zhi-GangLiu,MD,StateKeyLaboratoryofRespiratoryDiseaseforAllergyatShenzhenUniversityandShenzhenKeyLaboratoryofAllergyandImmunology,SchoolofMedicine,ShenzhenUniversity,NanhaiAve3688,518060Shenzhen,Guangdong,People’sRepublicofChina.E-mail:lzg@http://wendang.chazidian.com.Or:StephenKwok-WingTsui,PhD,SchoolofBiomedicalSciences,theChineseUniversityofHongKong,Shatin,N.T.,HongKong.E-mail:kwtsui@cuhk.edu.hk.0091-6749

Ó2014TheAuthors.PublishedbyElsevierInc.onbehalfoftheAmericanAcademyofAllergy,Asthma&Immunology.ThisisanopenaccessarticleundertheCCBY-NC-NDlicense(http://wendang.chazidian.com/licenses/by-nc-nd/3.0/).http://wendang.chazidian.com/10.1016/j.jaci.2014.09.031

wasEnterobacter(63.4%).Anintronwasfoundinthe13.8-kDaDfarinaebacteriolyticenzymegene,indicatingthatitisofHDMorigin.TheKyotoEncyclopediaofGenesandGenomespathwayanalysisrevealedaphototransductionpathwayinDfarinae,aswellasthiamineandaminoacidsynthesispathways,whichissuggestiveofanendosymbioticrelationshipbetweenDfarinaeanditsmicrobiome.

Conclusion:AnHDMgenomedraftproducedfromgenomic,transcriptomic,andproteomicexperimentsrevealedallergengenesandadiverseendosymbioticmicrobiome,providingatoolforfurtheridenti?cationandcharacterizationofHDMallergensanddevelopmentofdiagnosticsand

immunotherapeuticvaccines.(JAllergyClinImmunol2014;nnn:nnn-nnn.)

Keywords:Housedustmite,allergen,genome,microbiome,transcriptome,proteome,ubiquinol-cytochromecreductasebindingprotein,Derf24,Enterobacterspecies

Allergicdiseases,whichaffect30%to40%oftheworld’spopulationandareincreasinginprevalenceinternationally,particularlyamongyoungpeople,havenegativeeffectsonpatients’workandsociallivesandhavebecomeacostlyglobalhealthproblem.1,2Housedustmites(HDMs)arepredominantsourcesofinhalantallergens,withmorethan50%ofallergicdiseasecasesbeingattributedtothem.3-5Decadesofresearchhaverevealed23HDMallergengroups,withthecanonicalgroup1and2allergensbeingthemostclinicallyimportantbecausetheypossessIgE-bindingactivityinmostseraofpatientswithmiteallergy.5-7Group1and2allergensinduceTH2immuneresponsesbyencodingcysteineproteasesandbyfacilitatingToll-likereceptor4signaling,respectively.8,9

ItremainsaperplexingquestionwhyHDMsareseeminglyteemingwithallergeniccomponents.TheidentitiesofthefullspectrumofHDMallergeniccomponentsarenotyetknown.Allergen-speci?cimmunotherapyrepresentstheonlycurrentlyavailabletherapythathaslong-lastingeffectsonallergicdiseases.10HDMallergenvaccinesaregenerallymadefromextractsofpuri?edmitebodies,whichincludecomponentsofmicrobesthatinhabitmites.11,12Itisdif?culttoensurethelot-lotconsistencyofthevaccinebecauseofitscomplexcomponents.Distinguishingtheeffectivecomponentsofvaccinesfromthosethatproducesideeffectswouldenablemorepotentandsafevaccinestobedeveloped.

HavingknowledgeoftheHDMgenomeanditsendosymbioticmicrobiomewillbepivotaltoresolvingtheaforementionedcore

1

内容需要下载文档才能查看

2CHANETAL

scienti?candclinicalissuesinthe?eldofallergy.TheclosestspeciestotheHDMforwhichagenomedrafthasbeenproducedisthespidermiteTetranychusurticae,13whichisacauseofoccupationalallergicdiseaseinagriculturalworkers.14However,despitetheirprominentroleasallergensources,thegenomesoftheHDMsDermatophagoidespteronyssinusandDermatopha-goidesfarinaehaveyettoberesolved,restrictingmorein-depthresearchonHDMallergensandthemechanismsunderlyingtheirallergenicity.Here,wecombinedgenomicandtranscriptomicapproachestoproduceaDfarinaedraftgenomethatcanprovideinsightsintotheidentitiesofthefullarrayofDfarinaeallergensandthemechanismsmediatingtheirallergenicity,includingthepotentialroleofthemicrobiome.Weappliedourdraftgenomeincombinationwithproteomicandcomparativeanalysestouncoveranovelmajorallergenandexaminethegenesunderlyingphysiologicandmetabolicprocesses.

METHODS

Mitecultureandpuritycheck

DfarinaemiteswereisolatedfromindoordustsamplesfromShenzhenCityinsouthernChina.2ThemitecultureandpuritycheckmethodsaredescribedintheMethodssectioninthisarticle’http://wendang.chazidian.com.

Genomeandtranscriptomesequencing

DfarinaegenomicDNAandRNAsampleswerepreparedasdescribedintheMethodssectioninthisarticle’sOnlineRepository.Fourpaired-endsequencinglibrarieswithinsertsizesof200,500,2000,and5000bp,respectively,wereconstructedbyusingDfarinaewholeDNAandthensequencedwithanIlluminaHiSeq2000Sequencer.Atotalof24gigabase(Gb)pairsofsequencingdataweregenerated.DfarinaecDNAsweresequencedwiththeIlluminaHiSeq2000Sequencer;5.8Gbofpaired-endsequencingdata(insertsize,approximately200bp)wasgeneratedfortranscriptomeanalysis.

Genomeassemblyandannotationusingtranscriptomedata

GenomeassemblybeganwithreconstructionbyusingSOAPdenovo,15ALLPATHS-LG,16andVelvet17(seetheMethodssectioninthisarticle’sOnlineRepository).Protein-codinggeneswerepredictedwiththeuseof2abinitiogenepredictiontools:GeneMark-ES18andGimmerHMM.19AnnotationofnoncodingRNAgeneswasdonewithtRNAscan-SE20andRNAmmer.21TranscriptomesequencingdatawereassembledbyusingTrinity,22andtheassembledtranscriptswereusedtore?netheannotationsbyusingGeneMark-ESandGlimmerHMM.SplicejunctionsandrelativeabundanceofRNAsequencingreadsweredeterminedwithTopHat,23SpliceMap,24andCuf?ink.25Finally,weevaluatedthecompletenessofourdraftgenomerelativetotheCoreEukaryoticGenesMappingApproach(CEGMA)setof248coreeukaryoticgenes(CEGs)withtheCEGMApipeline.26

JALLERGYCLINIMMUNOL

nnn2014

Microbiomeanalysis

Becauseofthepossibilityofsymbioticrelationshipsbetweenmitesandmicroorganismsthatmightprecludeentirelysterilecultureconditions,mitesequencingdatawereseparatedfrommicrobiotasequencesbymeansofmanualcurationbasedonBLASTsearchesinthemicrobialdatabase.Theassembleddraftgenomewascomparedwithmicrobialdatabases,asdescribedindetailintheMethodssectioninthisarticle’sOnlineRepository,todistinguishbetweenDfarinaeandmicrobialgenomes.Brie?y,wesearchedthemicrobialRefSeqdatabaseusinggenomicsequencingreadswitha

high-stringencycutoff(E-value_<

1e250);matchesforeachreadmustmaptothesamegenus.

ThefollowingaspectsaredescribedintheMethodssectioninthisarticle’sOnlineRepository:mitecultureandpuritycheck;KyotoEncyclopediaofGenesandGenomes(KEGG)pathwayandphylogeneticanalysesandmetaboliccomparisontoTetranychusurticae;allergengenecloningandpro-teomicidenti?cation;IgE-bindingassayandskinpricktests;andEnterobactercloacaeimmunohistochemistryandbacteriolyticenzymegenecloning.

RESULTS

Miteculturepurity

MorphologicinspectionsandPCRexperimentscon?rmedthepurityoftheDfarinaespeciesinourcultures.TherewasnocontaminationfromDpteronyssinus.Aftertheculturemediumhadbeendigestedwithnuclease,thegenomicDNAsamplewascon?rmedtobecontaminationfree(Fig1,AandB,andseeFigsE1-E3andTablesE1andE2inthisarticle’http://wendang.chazidian.com).

Mitegenomedraft

High-throughputsequencing(seeTableE3inthisarticle’http://wendang.chazidian.com)yielded24Gbofgenomesequencesorroughly380-foldcoverageoftheestimatedgenomesize.Afterbuildingadenovodraftassemblyandapplyinggap?lling,4sequencinglibrarieswereassembledinto554scaffolds(totallength,61Mb;N50length,197kb).Becausethesampleincludednucleicacidsattributabletothemite’smicro-biome,weexamineditsmicrobialcomposition(seeTableE4inthisarticle’http://wendang.chazidian.com).SeparationofmicrobialDNAresultedina53.5-MbDfarinaedraftgenomewith516nucleargenomescaffolds(N505187kb)anda14.3-kbmitochondrialgenome(seeTableE5inthisarticle’http://wendang.chazidian.com).ThedraftgenomewassubmittedtotheNationalCenterforBiotechnologyInformation(NCBI)BioProject(ID:PRJNA17406,accessionno.:ASGP00000000).

Ourdraftgenomeincluded242(97.58%)of248CEGs,with239(96.58%)of248completeCEGs(seeTableE6inthisarticle’http://wendang.chazidian.com),indicatinggoodcompleteness.22Weretrieved264nucleotidesequencesand189aminoacidsequencesofDfarinaefromtheNCBI(April2012)andcon?rmedthat261(98.8%)ofthenucleotidesequencesand182(96.3%)oftheaminoacidsequenceswerepresentinthedraftgenome(E-valuecutoff:1e26).

Theguanine-cytosinecontentsofthecodingDNAsequencesandwholegenomewere34.4%and29.5%,respectively.GeneMark-ES,GimmerHMMrun1(withCaenorhabditiselegans),andGimmerHMMrun2(withTurticae)predicted13,475,20,165,and14,156genemodels,respectively.Another3,265genemodelswereinferredwithTopHatandCuf?inkstranscriptomesequenceanalysis.Altogether,16,376genemodels

JALLERGYCLINIMMUNOLVOLUMEnnn,NUMBERnn

CHANETAL3

FIG1.HDMmorphologyandphylogeny.A,PhotomicrographofaliveadultfemaleDfarinaemite.B,ScanningelectronmicroscopicimageofanadultfemaleDfarinaemite.C,GOdistributionofDfarinae.D,PhylogenetictreeofDfarinaewith25otherarthropodspecies.

wereobtained,including9,142thatweresupportedbyRNAsequencingresults.GeneannotationwithBlast2GOyielded8,201geneswithsigni?canthits(E-valuecutoff:1e210)intheNCBInonredundantdatabase.Amongthem,7,348wereassignedatleast1geneontology(GO)term(Fig1,C).

UsingtRNAscan-SE,weidenti?ed65annotatedtransferRNAgeneswithintheDfarinaenucleargenome.TherRNAgenes(18S,5.8S,and28S)thatwereexperimentallyidenti?edinDfarinaepreviouslywerefoundwithinourDfarinaegenomedraft(seeTableE5).Both18Sand28SwerepredictedintheirentiretybyusingRNAmmer.

Phylogeneticanalysis

Aphylogenetictreewasconstructedfromtheproteinsequencesof101CEGsthataresharedamong25Arachnida,Branchiopoda,andInsectaspecies(seeTableE7inthisarticle’http://wendang.chazidian.com)torevealDfarinae’sphylogeneticrelationshipswithotherarthropods.AcomparisonofDfarinae’sgenomeannotationswiththoseofTurticae,towhichourmaximumlikelihoodtreeshowedDfarinaetobephylogeneticallyclose(Fig1,D),revealedsimilarGOdistributions(seeFigE4inthisarticle’http://wendang.chazidian.com).23Regardinggenesassociatedwithmetabolicprocesses,662of3,029genesinDfarinaeand476of2,492genesinTurticaedidnotcorrelatewitheachother.

TranscriptomeandKEGGpathwayanalysis

Weassembled5.8Gbofpaired-endsequencingdata,annotatingatotalof16,376genes,withonly9,142(55.8%)beingsupportedbyRNAsequencingdata,probablybecauseofdevelopmentalvariationofgeneexpressionpro?les.FunctionalannotationforGOandtheKEGG27pathwaydatabaseusingtheBLAST2GO28programmapped7,348protein-codinggenestoGOprojectcategories(Fig1,C).Weidenti?edmostconstituentsoftheKEGGphototransductionpathwayintheDfarinaeanno-tatedgenes(seeFigE5,A,inthisarticle’http://wendang.chazidian.com)withhighhomology,withtheexceptionofrhodopsin.Onecandidategene(DEFA_098690)encodinga7-transmembranedomainproteinwith83.7%similaritytotherhodopsinfamilytransmembranereceptordomain(aa84–206;XP_002430048inGenBank)ofPediculushumanuscorporiswasannotatedasaclassArhodopsin–likeGprotein–coupledreceptor,GPRadr2(seeFigE5,B).

Identi?cationofHDMallergengenes

Weretrievedcompletegenesequencesandstructuresof20reportedHDMallergens,8,9includingDerf1toDerf23,exceptforDerf17,Group12,andGroup19,fromourassembledgenomewithsupportfromtranscriptomicanalysis(TableIandseeFigE6inthisarticle’http://wendang.chazidian.com)anddeterminedtheirrelativeexpressionlevelsinadultD

farinae

4CHANETAL

JALLERGYCLINIMMUNOL

nnn2014

FIG1.(Continued).

mitesintermsoffragmentsperkilobaseoftranscriptpermillionmappedreads(TableI).Derf4isreportedhereforthe?rsttime.Wealsoidenti?edthecompletesequencesof7noncanonicalallergencandidates(pro?lin,Alta6,a-tubulin1,cathepsin,Malas6,aldehydedehydrogenase,andenolasehomologues)withaminoacidsequencehomology(approximately41.5%to90.5%)toexperimentallyvalidatedallergensinotherspecies(seeFigE7andTablesE8andE9inthisarticle’http://wendang.chazidian.com).Thestructuresofthese20clonedcanonicaland7noncanonicalallergengenesweremapped(seeFigsE6andE7)andtheirsequenceswerecon?rmedtobeidenticaltosequencesinourassembledDfarinaegenome,con?rmingtheywereofmite,ratherthanmicrobial,origin.Matrix-assistedlaserdesorption/ionizationtime-of-?ightmassspectrometryanalysisofproteinspotsreactivetopooledserafrom20patientswithHDMallergywithourintegrated-omicsapproach(circledspotsinFig2,AandB,andseeFigE8,AandB,inthisarticle’http://wendang.chazidian.com)revealed4knowncanonicalallergens(Derf1,Derf2,Derf11,andDerf14)and12otherproteins(seeTableE10inthisarticle’http://wendang.chazidian.com).Thegenesequencesofthese12homologueswerecon?rmedbyusingtheSangermethodtobeidenticaltosequenceswithinourassembledDfarinaegenome.Thesesequencingdatacon?rmedthatthesegeneswerefromthemite

genome.

JALLERGYCLINIMMUNOLVOLUMEnnn,NUMBERnn

CHANETAL5

TABLEI.GenestructuresofcanonicalDfarinaeallergenscon?rmedbymeansofSangersequencing

Deducedno.ofHomologueyDerDerDerDerDerDerDerDerDerDerDerDerDerDerDerDerDerDerffffffffffffffffff34567891011131415161820212223DEFA_036500DEFA_092370DEFA_009370DEFA_160240DEFA_012670DEFA_112610DEFA_108510DEFA_012620DEFA_029610DEFA_016640DEFA_023480DEFA_127470DEFA_053360DEFA_042810DEFA_122350DEFA_009360DEFA_072800DEFA_123860Trypsina-Amylase

StructuralproteinChymotrypsinUnknown

GlutathionetransferaseSerineproteaseTropomyosinParamyosin

Fattyacidbinding

Vitellogenin:eggyolkstorageChitinase

Gelsolin:actinbindingChitinase

ArgininekinaseStructuralprotein

MD-2–relatedlipidrecognitionChitin-bindingdomaintype2232322451126473522325952613227921322127228487613116665564804623561361559136.7912.99410.2924.83481.87353.394.371532.23321.771772.53130.8440.83147.9188.28342.86943.53820.9831.67P49275(99%)

AAD38942(88%)àABO84970(100%)ABG23667(100%)ACK76298(99%)AAP35080(96%)àAAP57077(92%)Q23939(99%)AAO73464(98%)2A0A_A(100%)AAM21322(88%)àAAD52672(96%)AAM64112(99%)AAM19082(100%)AAP57094(99%)AAX34048(100%)ABG35122(100%)ACB46292.1(84%)

FPKM,Fragmentsperkilobaseoftranscriptpermillionmappedreads.*LocustagsareincludedinourassembledDfarinaegenome.

 AnalyzedrelativetotheNCBIdatabase(byBLASTalgorithm);GenBankaccessionnumbersarelisted.àPartial.

Onthebasisofproteinsizeandsignalintensity,weselected6ofthese12proteinsforproductionofrecombinantproteinsforprobingofallergenicity:ubiquinol-cytochromecreductasebind-ingprotein(UQCRB)–likeprotein;myosinalkalilight-chainprotein;secretedinorganicpyrophosphatase;DFP2;co?lin;andferritinheavy-chain-likeprotein.OnlyrecombinantUQCRB-likeprotein(seeFigE9inthisarticle’http://wendang.chazidian.com)wasstronglyboundbyIgEinseparateserumsam-plesfrom18of18patientswithHDMallergyandnotboundbyIgEantibodiesinseraobtainedfromanyof18nonallergichealthycontrolsubjectsand7patientswithpollenallergy(Fig2,E,andseeFigE10andTableE11inthisarticle’http://wendang.chazidian.com).Theother2recombinantproteins(co?linandsecretedinorganicpyrophosphatase)yieldedveryweakpos-itiveresponses,whereastheremaining3recombinantproteinsdidnotshowanyIgE-bindingactivities(datanotshown).ThegeneencodingtheUQCRB-likeproteinwascloned,anditsgenestruc-turewasmapped(seeFigE11,A,inthisarticle’http://wendang.chazidian.com).AnIgE-ELISAdemonstratedstrongIgEbindingtorecombinantUQCRB-likeprotein(Z55.6702,P<.0001)in22(100%)of22serafrompatientswithmiteallergy(Fig2,F).Invivotestingshowedskinreactivityin5of10patientswithmiteallergy(seeTableE12inthisarticle’http://wendang.chazidian.com).ThisnovelmajorallergenwasdesignatedasDerf24bytheWorldHealthOrganization/InternationalUnionofImmunologicalSocietiesAllergenNomenclatureSub-committee(http://wendang.chazidian.com/viewallergen.php?aid5772).

predominantlytoEcloacaeandEnterobacterhormaechei;Staph-ylococcus(17.8%)andEscherichia(4.9%)specieswerethenextmostpredominantgenera(Fig3,A,andseeTableE5).Bartonellaspeciesaccountedforonly1.7%ofthereads.Ourimmunohisto-chemistryexperimentcon?rmedtheabundantpresenceofenter-obacteriainthegutsofDfarinae(Fig3,B).

WeexaminedwhetherthegenesencodingKEGGpathwayenzymeswerepresentintheDfarinaegenomeoritsmicrobiomefractionandidenti?ed3pathwayswhereinthemajorityofgeneswerefromthemicrobiome:thiaminebiosynthesis(seeFigE12inthisarticle’http://wendang.chazidian.com)andaro-maticandaliphaticaminoacidbiosynthesis(seeFigsE13andE14inthisarticle’http://wendang.chazidian.com).Additionally,bycombiningtheDNAandRNAsequenceinfor-mation,wedeterminedfromourdraftgenomethatanintronispresentinthe13.8-kDaDfarinaebacteriolyticenzyme(Fig3,C).

Microbiome

Oftheapproximately112,000genomicsequencingreadsthatwelinkedto100microbialspecies(Fig3,A,andseeTableE5),71,000(63.4%)mappeduniquelytoEnterobacterspecies,most

DISCUSSION

Weproduceda53.5-MbDfarinaedraftgenomewith516scaf-foldsandacomplete14.3-kbmitochondrialgenome.ItscompletenesswasassessedbymeansofapplicationofCEGMAtoidentifytheexistenceof248CEGsthatarepresentinawiderangeoftaxa.26AsshowninTableE6,AandB,thenumberofcompleteproteinsencodedwassimilartothatofTurticae,13withbothreachingmorethan95%completeness.Notably,weidenti?ed3additionalcompleteproteinsinourdraftgenome(seeTableE6).

Adraftgenomeprovidesaframeworkforanorganism’sDNAfragmentassemblyandallowsupto95%ofgenestobeidenti?ed.However,comparedwithacompletegenome,itdoeshavelimitations.The?rstlimitationisthepotentialomissionofrepetitivesequencesinsomeintronsorintergenicregions.Second,theremightbealackofcorrelationamong

scaffolds.