发布时间 : 星期一 文章2015-6-Mandarins-TGG - 图文更新完毕开始阅读6c7fb9d4de80d4d8d05a4f00
TreeGenetics&Genomes (2015) 11:123 DOI10.1007/s11295-015-0951-1
ORIGINALARTICLE
Geneticdiversityandpopulationstructureanalysisofmandaringermplasmbynuclear,chloroplasticandmitochondrialmarkers
AndresGarcia-Lor1&Fran?oisLuro2&PatrickOllitrault1,3&LuisNavarro1Received:29July2015/Revised:13October2015/Accepted:18October2015#Springer-VerlagBerlinHeidelberg2015
AbstractThemandarinhorticulturalvarietalgroup(abasictaxonofthecultivatedcitrus(Citrusreticulata))ishighlypolymorphic.Itincludesalsogenotypesintrogressedbyotherspecies.Theprecisecontributionofancestralspeciestothemandaringroupisnotknown.Thegoalsofthisworkwere(1)tocharacterisethemandaringermplasmusingnuclear(simplesequencerepeat(SSR),indel,singlenucleotidepolymorphism(SNP)),chloroplasticandmitochondrialmarkers;(2)toeval-uategeneticdiversityanddetectredundancies;(3)todeter-minethepossiblepresenceofcitrusancestralgenomeintro-gressionsintothemandaringenome;and(4)todeterminethegeneticstructurewithinthemandaringroup.Fiftymicrosatellites(SSRs),24insertion-deletions(indels),67SNPs,8chloroplasticSSRs(cpDNA)and4mitochondrial(mtDNA)indelmarkerswereanalysedfor191genotypes,includingthe4maincitrusancestralspecies(C.reticulata,C.maxima,C.medicaandC.micrantha)andFortunella.C.maximawasthemaingenomeintrogressedinthemandarin
CommunicatedbyW.-WGuo
ThisarticleispartoftheTopicalCollectiononPopulationStructureElectronicsupplementarymaterialTheonlineversionofthisarticle(doi:10.1007/s11295-015-0951-1)containssupplementarymaterial,whichisavailabletoauthorizedusers.*LuisNavarro
lnavarro.ivia@gmail.com
1germplasm.SevenclusterswererevealedbyStructureanaly-sisatthenuclearlevel(N)withinthemandaringermplasm.Atleastfouroftheseclustersshowedaclearintrogressionfromotherancestralspecies.Moreover,mostofthemandarinsap-pearedtobecomplexmixturesofthesegroups.Thematernalindelanalysis(mtDNAandcpDNA)revealedtencytotypesinwhichmandarinswererepresentedinsevenofthem.Thisworkprovidesnewinsightsintotheorganisationoftheman-daringermplasmanditsstructureatthenuclearandcytoplas-miclevelsandwillbeusefultodesignmoreefficientbreedingprogrammesandmanagementofcitrusgermplasmcollections.
KeywordsCitrusreticulata.Molecularmarkers.Diversity.Germplasm.Genetics.Introgression
Introduction
Citrusisthemostimportantfruitcropintheworld,withaproductionofover128milliontonsandacultivatedareaof9.2millionhectares(FAOSTAT2013).Amongthecommer-cialcitrusfruits,mandarinsarethesecondmostimportantgroupinthefreshfruitmarketworldwide.‘Mandarin’isacommonnamegiventomostsmall,easy-peelingcitrusfruits.Thistermincludesinterspecifichybrids,whichmakemanda-rinsthemostgeneticallyandphenotypicallypolymorphicgroupoftrueCitrus(Nicolosietal.2000;Barkleyetal.2006;Garcia-Loretal.2012,2013a).Moreover,arecentphy-logeneticstudy(Garcia-Loretal.2013a)revealedaclosere-lationshipbetweenthegenusFortunellaandthemandaringroup.MandaringermplasmwasclassifiedasCitrusreticulataBlancobySwingleandReece(1967)andMabberley(1997).Onthecontrary,Webber(1943)classifiedmandaringenotypesintofourdifferentgroups:king,satsuma,
CentrodeProtecciónVegetalyBiotecnología,InstitutoValencianodeInvestigacionesAgrarias(IVIA),46113Moncada,Valencia,SpainUMRAGAPInstitutNationaldelaRechercheAgronomique(INRACorse),20230SanGiuliano,France
UMRAGAP,CentredecoopérationInternationaleenRechercheAgronomiquepourleDéveloppement(CIRAD),StationdeRoujol,97170Petit-Bourg,Guadeloupe
23 123 Page2of15mandarinandtangerine.Tanaka(1954)dividedthemandarinsintofivegroupsthatincluded36species,basedonmorpho-logicaldifferencesinthetree,leaves,flowersandfruits.Group1includedC.nobilisLour.(‘King’),C.unshiuMarc.(satsumas)andC.yatsushiroHort.exTanaka;group2includ-edC.kerajiHort.exTanaka,C.otoHort.exYuichiroandC.toragayoHort.exYuichiro;group3included14species,includingsomeofthemosteconomicallyimportantvarieties,suchasC.reticulata(‘Ponkan’),C.deliciosaTenore(‘Willowleaf’or‘commonmandarin’),C.clementinaHort.exTanaka(clementines)andC.tangerinaHort.exTanaka(‘Dancy’);group4includedC.reshniHort.exTanaka(‘Cle-opatra’),C.sunkiHort.exTanaka(‘Sunki’)andC.tachibana(Mak.)Tanaka;andgroup5includedthespeciesC.depressaHayata(‘Shekwasha’)andC.lycopersicaeformis(Lush.)Hort.exTanaka.Hodgson(1967)dividedthemandarinsintofourspecies:C.unshiu(satsumas),C.reticulata(‘Ponkan’,‘Dancy’,clementines),C.deliciosa(‘Willowleaf’)andC.nobilis(‘King’).
Noneofthesecitrusclassificationsystemsisperfect,buttheTanakasystemseemsbetteradaptedtothehorticulturalfeaturesofeachgroup,whereastheSwinglesystemsim-plifiesittotheextreme.Atpresent,C.reticulata(mandarin)isconsideredtobeoneofthefourancestralgroupsofthecultivatedcitrus(Nicolosietal.2000;BarretandRhodes1976;KruegerandNavarro2007),alongwithC.maxima(Burm.)Merr.(pummelo),C.medicaL.(citron)andC.micranthaWester(papeda).Thecentreofdiversifi-cationofC.reticulataislocatedinAsia,fromVietnamtoJapan(Tanaka1954).Thisgroupishighlypolymorphic,asrevealedbymolecularmarkers(Coletta-Filhoetal.1998;Ollitraultetal.2012a),chromosomalbandingpatterns(YamamotoandTominaga2003)andphenotypiccharacters,suchasfruitpomologyandthechemicalvariabilityofpeelandleafoils(Lotaetal.2000;Fanciullinoetal.2006),aswellastolerancetobioticandabioticstresses.Severalgerm-plasmcollectionshavebeencharacterisedbymorphologicalcharacteristicsand/ormolecularmarkers(Barkleyetal.2006;Koehler-Santosetal.2003;TapiaCamposetal.2005).Thisphenotypicandgeneticvariabilityreflectsalonghistoryofcultivation,inwhichmanymutationsandnaturalhybridisationshavegivenrisetotheexistingdiversitywithinthismainlyfacultativeapomicticgroup,includingtherecent-lypublishedintrogressionofancestralgenomes,likeC.maxima,insomemandarins(Wuetal.2014;Curketal.2014,2015).Theintraspecificorganisationofmanda-rinsandthedeterminantsofthegroup’sphenotypicdiversityremainpoorlyunderstood.
Inadditiontothetaxonomiccomplexityofthemandaringroup,thegenotypesincludedincitrusgermplasmcollectionsaresometimesofdoubtfulorigin.Theoriginofthesegeno-typescanbefromplantexplorationsinregionsofnaturalgeneticdiversity(mainlyAsiainmandarins),selectionof
TreeGenetics&Genomes (2015) 11:123
newmaterialsfromhybridisationsormutationsorbyex-changebetweengermplasmcollections(KruegerandNavarro2007).Theassignationofacultivarnameand/ormembershipinaspeciescanbedonearbitrarily,withnomolecularbasis,leadingtopossiblemistakesinassignationorduplicationofmaterial(KruegerandNavarro2007).Forthesereasons,mo-lecularstudiesareimportantforthedetectionofmisidentifi-cationsandredundancies(KruegerandRoose2003).
Toclarifytowhichgeneticgroupwereferbyusingtheterm‘mandarin’inthetextofthiswork,wehaveusedthefollowingnomenclature:C.reticulatafor‘mandarin’asatruespecies(oneofthefourancestorsofthecultivatedcitrus);C.reticulata(Sw)for‘mandarin’accordingtotheSwingleclassification;C.reticulata(Tan)foroneofthe17‘mandarin’speciesrepresentedinthisworkaccordingtotheTanakaclas-sification;and‘mandarin-like’forgenotypesthatarepheno-typicallysimilartomandarins.
Thegoalsofthisworkwere(1)tocharacterisethemandaringermplasmusingnuclear(simplesequencere-peats(SSRs),indels,singlenucleotidepolymorphisms(SNPs)),chloroplasticSSRs(cpDNA)andmitochondrialindelmarkers(mtDNA);(2)toevaluateitsgeneticdiver-sityanddetectredundancies;(3)todetectintrogressionsofothercitrusancestraltaxaintothemandaringermplasm;and(4)todeterminethegeneticstructurewithintheman-daringroup.
Inthisstudy,wehaveobservedtheintrogressionofseveralancestralgenomesinthemandaringenotypesconsideredpre-viouslyaspuremandarinsandalsothatmanymandarinsap-peartohaveaverycomplexgeneticorganisation(mixtureofmandaringenomes).
Materialsandmethods
Analysedgermplasm
Onehundredninety-onegenotypeswerestudiedtodeterminetheirnucleardiversity.Throughoutthetext,thesegenotypeswillbereferredtobyidentificationnumber(ID),showninOnlineResource1.GenotypeclassificationwasperformedaccordingtotheSwingleandReece(1967)andTanaka(1954)systems.AsummaryofthegenotypesusedisshowninTable1.PlantmaterialfortheanalysiswascollectedfromthegermplasmcollectionsoftheInstitutoValencianodeInvestigacionesAgrarias(IVIA,Valencia,Spain),mainlyob-tainedfromAmericanandMediterraneansources,andtheStationdeRecherchesAgronomiques(CIRAD-INRA,Corsi-ca,France),whichincludemanygenotypesofAsiaticorigin(China,Japan,Philippines,India,Vietnam…).ThedatabasesofbothcollectionsarebasedontheTanakasystemofclassi-fication.Thesegenotypesbelongtothefourancestralspecies(26C.reticulata(Sw,mandarins),C.indicaandC.tachibana),
TreeGenetics&Genomes (2015) 11:123 Table1
Summaryofgenotypesemployedinthestudy
SwinglesystemTanakasystemNG/SNGSAC.reticulatahybridC.amblycarpa22C.reticulataC.deliciosa111C.reticulataC.daoxianensis10C.reticulataC.depressa52C.reticulataC.erythrosa21C.indicaC.indica11C.hystrixC.hystrix11C.reticulataC.kinokuni42C.maximaC.maxima1010C.medicaC.medica66C.micranthaC.micrantha11C.reticulataC.nobilis
62C.reticulataC.paratangerina22C.reticulataC.reshni11C.reticulataC.reticulata493C.reticulataC.suavissima11C.reticulataC.succosa11C.reticulataC.suhuiensis72C.reticulataC.sunki32C.tachibanaC.tachibana11C.reticulataC.tangerina112C.sinensisC.tankan10C.reticulataC.temple20C.reticulataC.unshiu82FortunellaFortunella
44C.reticulataHybridmandarin280C.reticulataTangelo40C.reticulataTangor120?
Bintangor10C.reticulataC.clementina30?
Unknown
2
0
Genotypes’classificationbasedonSwingleandtheirclassificationwithinourdatabasesbasedontheTanakasystem
NG/Snumberofgenotypesperspecies,NGSAnumberofgenotypesfromeachspeciesincludedwithinanancestralpopulation
tenC.maxima(pummelos),sixC.medica(citrons),twoPapeda(C.hystrixD.C.andC.micrantha)andfourFortunella(kumquats:F.crassifoliaSwing.,F.hindsii(Champ.)Swing.,F.japonica(Thunb.)Swing.andF.margarita(Lour.)Swing.).The26mandaringenotypesconsideredasC.reticulatabySwingleandReece(1967)wereconsideredbyTanaka(1977)as15species.Theothergeno-types(141‘mandarin-like’accessions,intra-andinterspecifichybrids)werenotassumedinanyofthepreviouslymentionedmaintaxa,inordertodeciphertheirstructureanddeterminewhethertheirTanakaclassificationinthegermplasmbankdatawasproperlyassignedinourdatabases.SeveriniaPage3of15 123
buxifolia(Poir.)Tenorewasaddedasanout-groupforneighbour-joininganalysis.
Forthematernalphylogeny,besidestheancestralspeciesandinterspecifichybridsdetailedbefore,eightextrageno-types(secondaryspecies)wereanalysed(twoC.sinensis(L.)Osb.,twoC.aurantiumL.,twoC.paradisiMacf.,oneC.aurantifolia(Christm.)Swing.andoneC.limonOsb.).Genotyping
FiftySSRmarkers(Kijasetal.1997;Froelicheretal.2008;Luroetal.2008;Alezaetal.2011;Cuencaetal.2011;Kamirietal.2011;Garcia-Loretal.2012,2013a,b)locatedalongtheninelinkagegroupsofthereferencegeneticmapofclemen-tine(Ollitraultetal.2012b),24indelmarkersidentifiedinadiscoverypanelrepresentativeofthegenusCitrus(Garcia-Loretal.2012,2013a)and67SNPmarkersminedin27nucleargenes(Garcia-Loretal.2013a)andinclementineBAC-ends(Ollitraultetal.2012a)wereused(OnlineResource2).Toassessthematernaloriginofthemandaringermplasm,eightchloroplasticSSRmarkers(Chengetal.2005)(ccmp1,ccmp2,ccmp4,ccmp5,ccmp6,NTCP7,NTCP9andNTCP28)andfourmitochondrialindelmarkers(Froelicheretal.2011)(nad2,nad5,nad7,rrn5/rrn18)wereused.Aftertheinitialanalysis,nad5(nopolymorphismswerefoundinourpopulation)andthreechloroplasticmarkers(NTCP28,ccmp1andccmp4)werediscarded.Inthelatter,duetobadamplificationsorduetoadifferenceinpolymorphismsofjustonebasethatcouldbemoreconfusingthanclarifying.
ForSSRandindelmarkers,amplificationsbypolymerasechainreaction(PCR)andanalyseswithacapillarygeneticfragmentanalyser(CEQ/GeXPGeneticAnalysisSystem;BeckmanCoulter,Fullerton,CA,USA)wereperformedasdescribedinGarcia-Loretal.(2012).TheGeneticAnalysisSystemsoftware(GenomeLabGeXP,v.10.0)wasusedfordatacollectionandanalysis.
SNPsweregenotypedbycompetitiveallele-specificPCRasdescribedbyGarcia-Loretal.(2013b).Dataanalysis
TheallelicdataobtainedwiththeSSR,indel,SNP,cpDNAandmtDNAmarkerswereusedtocalculategeneticdissimi-laritymatrix,atnuclear,chloroplastic,mitochondrialandcy-toplasmiclevels,usingthesimplematchingdissimilarityin-dex(di–j)betweenpairsofaccessions(units),withtheDar-win5software,version5.0.159(PerrierandJacquemond2006).Weightedneighbour-joining(NJ)analyses(SaitouandNei1987)werecomputedtodescribethepopulationdi-versityorganisation,androbustnessofbrancheswastestedusing1000bootstraps.
PopulationstructurewasinferredwiththeprogrammeStructure,v.2.3.3(http://cbsuapps.tc.cornell.edu/structure),
123 Page4of15whichimplementsamodel-basedclusteringmethodusinggenotypedata(Pritchardetal.2000;Falushetal.2003).Whenthereisaknownpopulationstructure,itallowstocalculatetheircontributiontogenomesofgenotypesofunknownori-gin.Incasesofunknownpopulationstructure,theStructureprogrammehelpstoassigntheoptimalnumberofpopulationswithinthesampledatasetunderstudy,basedontheparame-tersofEvannoetal.(2005).
F-statisticswerecalculatedwiththeprogrammeGENETIX,v.4.03(Belkhiretal.2002),basedontheparam-etersofWright(1969)andWeirandCockerham(1984).SomeothergeneticpopulationstatisticswereestimatedfromthealleledatausingtheprogrammePowerMarker,v.3.25(LiuandMuse2005).
Results
Nucleargeneticdiversityparameters
GeneticdiversitystatisticswerecalculatedforeachSSR,indelandSNPmarkerforthe191genotypesanalysed(OnlineResource3).
UsingtheSSRmarkers,wedetected529alleles.Allelenumbersvariedbetween4(MEST107)and18(MEST192).TheaveragenumberofallelesandtheHe(expectedheterozy-gosity)valueperlocuswere10.6±0.508and0.63±0.021,re-spectively.Thewholepopulationhadanobservedheterozygos-ity(Ho)of0.60±0.024.Fw(Wrightfixationindex)valuesvar-iedfrom?0.40(CAC23)to0.48(CimCrCIR01D11).Theav-erageFwvalueoverallSSRlociwascloseto0(0.04±0.024).Wedetectedatotalof74alleleswiththeindelmarkers.Allelenumberperlocusrangedfrom2(10markers)to7(IDDFR),withanaverageof3.1±0.182.TheHevaluesrangedfrom0.01(IDINVA2)to0.66(IDDFR),withamedianvalueof0.20.Fwvaluesvariedfrom?0.52(IDF’3H)to1.00(IDINVA1).TheoverallHoandFwvaluesamongalllociwere0.19±0.033and0.29±0.067,respectively.
AllSNPmarkerswerebiallelic(134allelesidentified).Thewholepopulationhadanobserved(Ho)andexpected(He)heterozygosityof0.27±0.023and0.26±0.016,respectively.TheminimumHovaluewas0.01(4markers)andthemaxi-mum0.52(CCC1-M85).TheoverallFwvalueamongalllociwas0.12±0.059.
Geneticpopulationstatisticswithinthewholepopulation(AG,[ID1–191]),all‘mandarin-like’genotypesofunknownorsupposedhybridorigin(AM,[ID51–191])andthe28genotypesselectedfromallTanakaspeciesrepresentedinourcollections(MT,[ID1–28])(OnlineResource1)aresummarisedinTable2.Genediversity(GD)andtheHovalueswerehigheramongSSRmarkersthanSNPandindelmarkers,reflectingthehighermaximumallelefrequencies(MAF)oftheindelandSNPthanSSR.Comparingthewholepopulation
TreeGenetics&Genomes (2015) 11:123
(AG),all‘mandarin-like’genotypes(AM)andmandarinsfromTanakaspecies(MT),themeanallelenumberdecreasedateachstepforSSRandindelmarkers(SSRsAG=10.58>AM=6.84>MT=6.76;indelsAG=3.08>AM=2.25>MT=2.02)andfortheSNPs(AG=2>AM=1.90>MT=1.79).TheGDwashigherinAGthaninAMorMTforbothkindsofmarkers.ForAG,HowasslightlylowerthanHe,leadingtoslightlypositiveFwforSSR,indelandSNPmarkers.InAMandMT,HovalueswerehigherthanHe,providingnegativeFwvaluesforbothkindsofmarkers.InAM,13locididnotshowpolymorphismandinMT14.Rarealleles
The‘mandarin-like’populationincluded20genotypeswithuniquealleles(OnlineResource4),rangingfrom1(11geno-types)to10(‘Nicaragua’;[ID112])uniqueallelespergenotype.
ClassificationsbyNJanalysis
Forthewholedataset(SSR,indelandSNPmarkers),NJanalysis(Fig.1)revealedacleardifferentiationbetweenthefivemaintaxastudied,thefourancestralCitrusgroups(papeda,citron,pummeloandmandarin)andkumquat,withveryhighbootstrapsupport.ThecombinationofthedatafromSSR,indelandSNPmarkersrevealedhighintraspecificdi-versityinthemandaringroup,whichwasnotwellresolved(lowbootstrapsupportinmanybranches).Fromthewholedataset,22genotypeswerereducedto8multilocusgenotypes(MLGs;OnlineResource5).Someofthesewereoriginatedbybudsportmutations(differentcultivarsofC.unshiuorC.clementina),andothersarepossibleredundantgenotypescollectedandnameddifferentiallyindifferentlocations.Contributionoftheancestraltaxatothemandaringroupandmodernhybrids:analysiswiththeStructuresoftwareTheSSR,indelandSNPdatawereanalysedwiththeStructuresoftwaretoassessthecontributiontothemandaringermplasmofthefourancestralCitrustaxa(C.reticulata(Sw),C.maxima,C.medicaandPapeda)andFortunella,usinganadmixturemodelandtheoptionofcorrelatedallelefre-quenciesbetweenpopulations.Thedegreeofadmixturealphawasinferredfromthedata.Theburn-inperiodwassetto500,000,andMCMC(MarkovChainMonteCarlo)repetitionsweresetto1,000,000;10runsofStructurewithK=5(5pop-ulationsassumed)wereperformed.Thesepopulationswereasfollows:mandarin(Sw),C.indicaandC.tachibana(28sam-ples,representing17Tanakaspecies),pummelo(10samples),citron(6samples),papeda(2samples)andkumquat(4sam-ples).Theothersamplesanalysed(141)wereassumedtohavebeenderivedfromtheseancestralpopulations(Online