Changesintheroot-associatedfungalcommunitiesalong - PDF document

Changesintheroot-associatedfungalcommunitiesalong aprimarysuccessiongradientanalysedby454 pyrosequencing RAKELBLAALID,*TORCARLSEN,*SURENDRAKUMAR,*RUNEHALVORSEN,† KARLINNEUGLAND,‡GIOVANNIFONTANA§andHA  VARDKAUSERUD* *MicrobialEvolutionResearchGroup(MERG),DepartmentofBiology,UniversityofOslo,POBox1066Blindern,N-0316 Oslo,Norway, NaturalHistoryMuseum,UniversityofOslo,Oslo,Norway, ‡ MarineBiology,DepartmentofBiology, UniversityofOslo,POBox1066Blindern,N-0316Oslo,Norway,§EcologyGroup,DepartmentofBiology,Universityof Bologna,ViaS.Alberto163,I-48123Ravenna,Italy Abstract Weinvestigatedchangesintheroot-associatedfungalcommunitiesassociatedwiththe ectomycorrhizalherb Bistortavivipara ampliconsequencing.Ourmainobjectivewastoassessthedegreeofvariationinfungal richnessandcommunitycompositionasvegetationcoverincreasesalongthechronose- quence.Sixtyrootsystemsof B.vivipara weresampledinvegetationzonesdelimitedby datedmorainesinfrontofaretreatingglacierinNorway.WeextractedDNAfromrinsed rootsystems,amplifiedtheITS1regionusingfungal-specificprimersandanalysedthe patchinessindistributionoffungalOTUsacrossrootsystems,indicatingthatstochastic processestoalargeextentstructurethefungalcommunities.However,timesince deglaciationhadimpactonthefungalcommunitystructure,asasystematicshiftinthe communitycompositionwasobservedalongthechronosequence.Ectomycorrhizal basidiomyceteswerethedominantfungiintherootsof B.vivipara, whenitcomesto  vardKauserud, Fax:+4722856339;E-mails:rakel.blaalid@bio.uio.noand haavarka@bio.uio.no  2011BlackwellPublishingLtd MolecularEcology(2012) 21 ,1897–1908doi:10.1111/j.1365-294X.2011.05214.x (Nesje&Dahl1991).Althoughcriticalremarkshavebeenmadeontheuseofchronosequenceasasubstitutefortime(Johnson&Miyanishi2008;Walkeretal.thestudybyNesje&Dahl(1991)indicatesthatthestudyareahasfollowedthesametrajectory.Theridgesaretheresultofglacierretreatsince1750,whichisknownasthepeakof‘thelittleiceage’.Thisenabledustoindirectlystudyfungalcolonizationovera250-yeartimespan.Theglacierforelandwasdividedintofourzonesbasedonthedatedmoraines(Fig.1).Weestablishedthreeparalleltransectsintheglacierfore-landacrossthedatedmorainesfromtheglacierfore-front(zone1)tothefullyestablishedalpinevegetationoutsidethe1750moraine(zone4).Thedeglaciationofzone4mighthavehappenedlongbeforethelittleiceage,becausethemorainewasdepositedwhenthegla-cierreacheditsmaximumextent.Hence,zone4canprobablybeconsideredasrepresentingtheclimaxvege-tationinthearea.Alongeachtransect,wesampledveplants(includingtheirrootsystemsembeddedinsoil)atrandomwithineachzone,totallingto60plants.SoilwassampledapproximatelyonemetreawayfromeachsampledplantfordeterminationofpHandorganicsoilcontent(measuredaslossonignition;TableS1,Sup-portinginformation).Morphometricdata,suchasshootheightandrootsize,wereobtainedfromeachsampledplant(TableS1,Supportinginformation).TopreventdegradationoffungalDNAwithintheplantrootsbeforeDNAextrac-tion,plantswerestoredat4Candtreatedwithin36haftersampling.Werstrinsedtheplantrootsystemsintapwatertoremovevisiblesoilandplantdebrisandremovedrootsnotattachedtotheplantrhizome.Subse-quently,theplantrootswererinsedthoroughlyindis-tilledautoclavedwateronetimeforatleast5min.AllrootswerethencutoffandtransferredtoEppendorftubescontainingcetyltrimethylammoniumbromide(CTAB)lysisbufferforDNAextraction.FreshrootweightswereobtainedbyweighingtheEppendorftubesbeforeandafteraddingtherootsystems.EntirerootsystemswerestoredinCTABbufferat–20CuntilDNAextraction.MolecularanalysesWeextractedtheDNAfromtheentireplantrootsys-temsusingCTABextraction(Murray&Thompson1980;Gardes&Bruns1993).Sampleswerepreparedfor454pyrosequencingbyperformingnestedPCRampli-cationusingthefungal-specicprimersITS1FandITS4etal.1990;Gardes&Bruns1993)intherststepandfusionprimersincludingITS5andITS2(Whiteetal.1990)inthenestedstep.Fusionprimerswerecon-structedbyaddingvedifferent6bpuniquetagsand454pyrosequencingadaptorsAandBtobothITS5andITS2,respectively.Tagswereatleast3bpdifferentfromeachotherinbothdirections.PCRwasperformedin20Lreactionvolumescontaining2LtemplateDNAand18Lreactionmix.FinalconcentrationsinthePCRwere0.16mdNTPmix,0.2ofeachprimerand0.4unitsFinnzymesPhusionpolymerase.TheamplicationprogrammeforbothstepsofthenestedPCRwasasfollows:30sat98C,followedby30cyclesof10sat98C,20sat55C,20sat72Candanalextensionstepat72Cfor7minbeforestorageat4PCRproductswerecleanedwithWizardSVGelandPCRClean-UpSystem(Promega,Madison,WI,USA),quantiedusingabioanalyzer(Algilent2100)andpooledinto12equimolarampliconlibraries.Notably,samplesfromdifferentzoneswereincludedinthedif-ferentlibraries.GSFLXsequencingofthetaggedampli-conswasperformedattheUltra-highThroughputSequencingPlatformattheUniversityofOslousing12lanesofone454plate.WeincludedtwonegativecontrolsduringtheentireanalysesfromtheDNAextractionstep.TherawdatahavebeenaccessionedintheEuropeanNucleotideArchive,accessionno.BioinformaticsandstatisticalanalysesWeanalysedtherawdata,including119054reads,etal.anewlydeveloped Transect 1Transect 2Zone1Zone2Zone3Zone4Transect 3 glacier Fig.1Schematicillustrationoftheglacierforelandinfrontofisenglacierdisplayingthethreetransectsalongthechrono-sequence.Themoraineridgesindicatetheboundariesbetweenthefouridentiedzones.Alongeachtransect,vereplicaterootsystemswerecollectedfromeachzone,yieldingatotalof60samples.PRIMARYSUCCESSIONOFROOT-ASSOCIATEDFUNGI2011BlackwellPublishingLtd 60rootsystems.Usinga98.5similaritycut-offforclustering,thereadsgroupedinto1663OTUsofwhichahighnumber(1193)weresingletons.However,the1193singletonsonlymadeup1.6ofthetotalnumberofreads.Clusteringusing97sequenceidentityreturnedalsoahighnumberofsingletons�(50).Afterallsingletonshadbeendiscarded,470OTUsremainedandwereusedinthefurtheranalyses.AfewOTUsaccountedforalargefractionofreads.ThetenmostcommonOTUsmadeup58ofthereads,whilemanyOTUswererepresentedbyfewreads(Fig.2a).Thefre-quencyofthe470OTUsacrossthe60rootsystemsvar-iedbetween1and47(Fig.2b).Notably,wefoundnosignicantrelationshipbetweenthenumberofOTUsandthenumberofreadsperrootsystem(�0.05).TaxonomiccoverageAllsequenceshadbestmatchestofungi,indicat-ingthattheimplementedprimerpairswerehighlyfun-galspecic.Seventy-onepercentageofthe470OTUsbelongedtoBasidiomycotaand19toAscomycota(Fig.3).ThenumberofOTUsdetectedfromZygomy-cota,ChytridiomycotaandGlomeromycotaconstitutedeach.Avastmajority()ofthereadsrepre-sentedBasidiomycota,asdidalsomostoftheabundantOTUs,bothinnumbersofsequencesandoccurrencesacrossrootsystems.Agaricales,ThelephoralesandSeb-acinaleswerethemostfrequentordersencompassing,17and9oftheOTUs,respectively.Theabun-danceofthetwelvemostcommonfungalgroupsatafamilylevelinthefourzonesisshowninFig.S2(Supportinginformation).Mostgroupsweredetectedinallzones,butespecially,werefarmoreabundantinzone1comparedtotheotherzones.Ontheotherhand,werealmostabsentinzone1.Atotalof28OTUswerewidespreadbydenitionofbeingpresentinmorethan10rootsys-tems(Table1).ThemostwidespreadOTU,detectedin47of60rootsystemsandaccountingforatotalof27776reads,had95pairwisesimilaritytoaGenBankaccession(AY669673)ofCortinariusrubricosus.ThesecondmostwidespreadOTU,detectedin32of60rootsystemsandrepresentedby4601reads,showed96pairwisesimilaritytoanaccessionofHebelomamesophaeum Number of OTUsNumber of sequencesNumber of occurrences 100–11473221191716141312119876543211100 1007550 Fig.2(a)Distributionofnumberofreadswithinthe1663identiedoperationaltaxonomicunits(OTUs),singletonsincluded.(b)Numberofoccurrencesofthe470OTUsacrossthe60rootsystems. PezizalesMitosporicascomycotaOtherascomycetesAgaricalesThelephoralesSebacinalesRussulalesCantharellalesOtherbasidiomycetesGlomeromycotaZygomycotaFungiincertaesedisChytridiomycota Chaetothyriales Fig.3Taxonomicdistributionofthe470operationaltaxo-nomicunitsmainlyattheorderlevel,accordingtothetopBLASThitsagainsttheNCBIdatabase.PRIMARYSUCCESSIONOFROOT-ASSOCIATEDFUNGI2011BlackwellPublishingLtd (FJ845404).AlthoughmostECMfungiweredetected,someputativerootpathogensappeared.Forexample,onefrequentlyoccurringOTUdetectedinallzoneswithtaxonomicafnityto(Table1)couldrepre-sentasystemicpathogen.FungaldiversityalongtheprimarysuccessiongradientThenumberofOTUsincreasedsignicantlyalongtheprimarysuccessiongradientfrom163inzone1closesttotheglacierto179,205and213inzones2,3and4(chi-squaretest,=0.038),respectively.Speciesaccu-mulationcurveswerecalculatedforthefourvegetationzones,separately(datanotshown)andcombined(Fig.4).Whenthenumberofanalysedrootsystemswasextrapolatedto1000,weestimated596OTUstoappearinzone4,followedby557,489and452inzones3,2and1,respectively.RegardingaveragenumberofobservedOTUsperrootsystem,nosignicantdiffer-encewasobservedacrossthezones(=0.36).VariationinthenumberofOTUsperrootsystemswasnotexplainedeitherbyzone,pH,organicsoilcontentorrootsize(GLM,�0.05).ChangesincommunitystructurealongtheprimarysuccessiongradientOnaverage,2.8OTUsweresharedbetweenpairsofrootsystems(TableS2,Supportinginformation).How-ever,therewasasignicantdifferenceinaveragenum-berofsharedOTUsbetweenpairsofrootsystemswithinandacrossvegetationzones(0.0001).Withinthefourzones,theaveragenumberofOTUssharedbetweenpairsofrootsystemswas3.1–4.2,whilethecorrespondingnumberbetweenpairsindifferentzoneswas2.0–3.8.InspiteofthegenerallowoverlapinOTUsbetweenrootsystems,theGNMDSandDCAordinationanalyses Table1The28mostfrequentlydetectedoperationaltaxonomicunitsintermofnumberofrootsystems,theirtaxonomicafliationanddistributionTophitTaxonomicgroupCov.*Sim.Zone1Zone2Zone3Zone4CortinariusrubricosusAgaricales(B)989578.346.786.7100.080.0HebelomamesophaeumAgaricales(B)1009653.340.053.353.366.7sp.Mitosp.Asco.(A)959835.026.753.340.020.0Zygomycota(Z)939831.733.333.320.040.0SpadicoidesbinaSordariomycetes(A)998028.333.326.733.320.0HydnumrufescensChantarellales(B)279426.720.020.06.760.0LactiuusgriseusRussulales(B)1009323.326.720.013.333.3TilletiabarclayanaTilletiales(B)2110023.326.76.720.040.0InocybeegenulaAgaricales(B)1009423.320.013.313.346.7Sebacinaceaesp.Sebacinales(B)998723.320.026.733.313.3Thelephoraceaesp.Thelephorales(B)1009623.320.033.320.020.0TetracladiumfurcatumMitosp.Asco.959823.313.326.720.033.3GeoporacervinaPezizales(A)839121.713.326.740.06.7LaccariamontanaAgaricales(B)1009621.720.033.36.726.7RussulapascuaRussulales(B)999921.720.06.733.326.7XenostigminazilleriCapnodiales(A)949321.720.046.713.36.7sp.Mucoromycotina909821.720.020.020.026.7TomentellabadiaThelephorales(B)999420.026.76.713.333.3TomentellaatromentariaThelephorales(B)999620.026.720.026.76.7InocybecurvipesAgaricales(B)969020.026.713.320.020.0PhialocephalafortiniiHelotiales(A)1009820.013.326.726.713.3LeohumicolaminimaLeotiomycetes(A)959420.033.333.36.76.7TyphulavariabilisAgaricales(B)1009420.013.333.320.013.3InocybeegenulaAgaricales(B)1009218.320.013.313.326.7MycenaamabilissimaAgaricales(B)359218.313.320.013.326.7ThelephoraamericanaThelephorales(B)1009618.313.36.733.320.0CladophialophoraminutissimaChaetothyriales(A)959618.36.713.313.340.0Cortinariussp.Agaricales(B)1009916.713.313.326.713.3*SequencecoverageinBLASTnsearch.SequencesimilarityinBLASTnsearch.Percentoccurrenceacrossthe60rootsystems.Percentoccurrenceacrossthe15rootsystemsineachvegetationzone.R.BLAALIDETAL.2011BlackwellPublishingLtd ingeneral,afarhighernumberofreadsperOTUcom-paredtotheascomycetesthatmightindicatethatbasid-iomycetesaredominantintermsofmycelialbiomassaswell.Thisspeculationrestsontheassumptionthatthereisacorrelationbetweennumbersofsequencereadsandtheinitialmycelialbiomasspresentinthesamples.Althoughthisrelationshipwasnottested,weconsidersuchacorrelationplausible,althoughtaxo-nomicbiasescertainlyareintroducedduringextraction,PCRand454sequencing(Feinsteinetal.2009;Belle-etal.2010;Tedersooetal.2010).AlownumberofOTUshadbestBLASTnmatchagainstZygomycota,GlomeromycotaandChytridiomycota,suggestingthatnon-dikaryafungiarepresentinlowfrequenciesintherootsystems.Freemanetal.(2009)founddominanceofchytridsinthefungalcommunitiesinhighelevatedandrecentlyde-glaciatedsoils,incontrasttothetaxo-nomicdistributionobservedinourstudy.However,etal.(2009)obtainedtheirsamplesfrompuresoil,wherechytridsmaybemoredominant.Intheory,someofthechytridsdetectedinourstudycouldbelooselyassociatedwiththeroots.TheemployedITSprimersmightalsomismatchagainstchytridfungilead-ingtoalowerrepresentationofthesefungi.AsjudgedbytheBLASTnmatches,mostofthefre-quentOTUs,intermsofbothnumberofsequencesandnumberofoccurrencesinthe60rootsystems,hadafn-itywithwell-knownectomycorrhizalfungiintheordersAgaricales,Thelephorales,SebacinalesandRussulales.TheseECMgroupshavealsoearlierbeendemonstratedtobeassociatedwithB.vivipara(e.g.Sønstebø2002;etal.2008).Amongtheascomycetes,wedetectedsomewidespreadOTUswithputativeectomy-corrhizalecology,suchas.Fur-thermore,OTUswithafnitytowidelydistributeddarkseptateendophyteslikePhialocephalafortiniidetectedfrequently.Thesefungimightalsohaveamycorrhizalandbenecialfunctionforthehostplantinthesehabitats(Jumpponen2001;Newshametal.Newsham2011).Notably,OTUswithafnitytoaquatichyphomycetes(Ingoldianfungi)likeArticulosporatetracladialadiumfurcatumwerewidespread,appearingin21and14rootsystems,respectively.Ithasearlierbeenshownthataquatichyphomycetesmighthaveaplantendo-phyticstageduringtheirlifecycle(Sridhar&Ba1992).Hence,theremightbeapossibilitythattheseIn-goldianfungihaveanendophyticstageintheB.vivi-ChangesindiversityMoststudiesondiversitypatternsalongprimarysuc-cessiongradientshavefocusedonabovegroundplantspeciesrichness(Matthews1978;Vetaas1994;Freynotetal.1998;Rafetal.2006)oranimals(Kaufmann2001;Hodkinsonetal.2004;Hagvar2010).Inseveralofthesestudies,thehighestspeciesrichnesshasbeenobservedatintermediatestagesofsuccession(Mat-thews1978;Vetaas1994;Rafetal.2006).Ageneraltrendobservedinthesestudiesisthecolonizationofnewlyexposedareasbyafewopportunisticspecies,fol-lowedbyanincreaseinspeciesrichnessassociatedwithanincreaseinaccumulatedresources(biomass),numberoftrophiclevelsandintensityofbioticinteractions.Atlaterstages,whentheecosystemapproachesamaturestage,strongcompetitorsmaydominate,resultinginadeclineinspeciesrichness.Therichnessofroot-associ-atedfungi,asobservedinthisstudy,doesnotfollowthispattern.Fungalrichnessisrelativelyhighalreadyintherecentlyexposedareas,followedbyaweak,butsignicantincreaseintotalrichnesstowardsmoreestablishedecosystems.Thisresultissimilartothend-ingsofSchuetal.(2010)wheretheyobservedthatdiversityofsoil-inhabitingbacteriainanarcticglacierforefrontwasgenerallyhighandincreasedsignicantlyalongthegradient.Competitionmightbealessimpor-tantfactorforlimitingtherichnessofmicro-organismsinaclimaxsystemcomparedtomacro-organisms,asmorenichescouldbeavailable.Ourresultsdonotsup-portearlierproposedmodelsoffungalprimarysucces-sion(Jumpponenetal.2002;Naraetal.2003a,b),whichpostulatethatonlyfewruderalandr-selectedspeciesareabletocolonizerootsystemsinthenewlyexposedareasbysporesandthatwithtime,theyarejoinedandsometimesreplacedbyadditional,morecompetitivespecies.However,thesestudieswereconductedusingfruitbodysurveys,whichgivelimitedinformationabouttheextantfungaldiversity.Moreover,ourresultssuggestthatthespeciesrichnesspatternsoffungalcom-munitiesaremoresimilartothoseofbacteria,andprobablyothermicrobes,comparedtothoseofthemac-robiota,i.e.plantsandanimals.CommunitystructureAhighpatchinessinthedistributionoffungalOTUswasobserved,whichcouldindicatethatthecoloniza-tionprocesshasastrongstochasticelement.Highpatchinessseemstobeacommonfeatureinmostinves-tigatedfungalcommunities(Horton&Bruns2001;etal.2002;Lilleskovetal.2004;Stuken-brock&Rosendahl2005;Tayloretal.2010).However,inseveralofthesestudies,itisarguedthatthepatchi-nessmaybeduetounder-sampling(Lilleskovetal.2004;Tayloretal.2010).Ourresults,includingahighnumberofsequencesperrootsystem,aswellasafairlyhighnumberofsamples(60),indicatethatthispatternR.BLAALIDETAL.2011BlackwellPublishingLtd EriksenM,BjurekeKE,DhillionSS(2002)MycorrhizalplantsoftraditionallymanagedborealgrasslandsinNorway.Mycorrhiza,117–123.FeinsteinLM,SulWJ,BlackwoodCB(2009)AssessmentofbiasassociatedwithincompleteextractionofmicrobialDNAfromsoil.AppliedandEnvironmentalMicrobiology,5428–FiererN,NemergutD,KnightR,CraineJM(2010)Changesthroughtime:integratingmicroorganismsintothestudyofsuccession.ResearchinMicrobiology,635–642.FreemanKR,MartinAP,KarkiDetal.(2009)Evidencethatchytridsdominatefungalcommunitiesinhigh-elevationProceedingsoftheNationalAcademyofSciences18315–18320.FreynotY,GloaguenJC,CannvacciuoloM,BellidoA(1998)PrimarysuccessiononglacierforelandsinthesubantarcticKerguelenIslands.JournalofVegetationScience,75–84.GardesM,BrunsTD(1993)ITSprimerswithenhancedspecicityforbasidiomycetes—applicationtotheidenticationofmycorrhizaeandrusts.MolecularEcology113–118.gvarS(2010)PrimarysuccessionofSpringtales(Collembola)inaNorwegianGlacierForeland.AntarcticandAlpineResearch,422–429.HelmDJ,AllenEB,TrappeJM(1996)MycorrhizalchronosequencenearExitGlacier,Alaska.CanadianJournalof,1496–1506.HesselmannH(1900)OmmycorrhizabildingarhosartiskaBihangtillKongligaSvenskaVetenskaps-AkademiensHandlingar,1–46.HillMO(1979)DECORANA—AFORTRANProgramforDetrendedCorrespondenceAnalysisandReciprocalAveragingCornellEcologyProgramSeriesCEP-40.EcologyProgram,CornellUniversity,NewYork.Availableat:http://gbic.tamug.edu/request.htm.HillMO,GauchHGJ(1980)Detrendedcorrespondenceanalysis:animprovedordinationtechnique.HodkinsonID,CoulsonSJ,WebbNR(2004)Invertebratecommunityassemblagealongproglacialchronosequenceinthehigharctic.JournalofAnimalEcology,556–568.HortonTR,BrunsTD(2001)Themolecularrevolutioninectomycorrhizalecology:peekingintotheblack-box.MolecularEcology,1855–1871.IshidaTA,NaraK,TanakaM,KinoshitaA,HogetsuT(2008)Germinationandinfectivityofectomycorrhizalfungalsporesinrelationtotheirecologicaltraitsduringprimarysuccession.NewPhytologist,491–500.JohnsonEA,MiyanishiK(2008)Testingtheassumptionsofchronosequencesinsuccession.EcologyLetters,419–431.JumpponenA(2001)Darkseptateendophytes—aretheymycorrhizal?,207–211.JumpponenA(2003)SoilfungalcommunityassemblyinaprimarysuccessionalglacierforefrontecosystemasinferredfromrDNAsequenceanalyses.NewPhytologist,569–578.JumpponenA(2004)Soilfungalcommunitiesunderneathwillowcanopiesonaprimarysuccessionalglacierforefront:rDNAsequenceresultscanbeaffectedbyprimerselectionandchimericdata.MicrobialEcology,233–246.JumpponenA,JonesKL(2009)Massivelyparallel454sequencingindicateshyperdiversefungalcommunitiesintemperateQuercusmacrocarpaphylloshere.NewPhytologist,438–448.JumpponenA,VareH,MattsonKG,OhtonenR,TrappeJM(1999)Characterizationof‘‘safesites’’forpioneersinprimarysuccessionofrecentlydeglaciatedterrain.Journalof,98–105.JumpponenA,TrappeJM,CazaresE(2002)OccurenceofectomycorrhizalfungiontheforefrontofretreatingLymanGlacier(Washington,USA)inrelationtotimesincedeglaciation.,43–49.KaufmannR(2001)Invertebratesuccessiononanalpineglacierforeland.,2261–2278.KauserudH,KumarS,BrystingA,NordenJ,CarlsenT(2011)Highconsistencybetweenreplicate454pyrosequencinganalysesofectomycorrhizalplantrootsamples.inpressDOI:10.1007/s00572-011-0403-1.KruskalJB(1964a)Multidimensionalscalingbyoptimizinggoodnessofttoanonmetrichypothesis.KruskalJB(1964b)Nonmetricmultidimensionalscaling:anumericalmethod.Psykometrika,115–129.KumarS,CarlsenT,MevikBetal.(2011)CLOTU:anonlinepipelineforprocessingandclusteringof454ampliconreadsintoOTUsfollowedbytaxonomicannotation.Bioinformatics,182.LesicaP,AntibusRK(1986)Mycorrhizaeofalpinefell-eldcommunitiesonsoilsderivedfromcrystallineandcalcareousparentmaterials.CanadianJournalofBotany1691–1697.LilleskovEA,BrunsTD,HortonTR,TaylorDL,GroganP(2004)Detectionofforeststand-levelspatialstructureinectomycorrhizalfungalcommunities.FEMSMicrobiology,319–332.MassicotteHB,MelvilleLH,PetersonRL,LuomaDL(1998)Anatomicalaspectsofeldectomycorrhizason(Polygonaceae)andKobresiabellardii(Cyperaceae).Mycorrhiza,287–292.MatthewsJA(1978)Plantcolonizationpatternsonagletschervorfeld,southernNorway:ameso-scalegeographicalapproachtovegetationchangeandphytometric,155–178.McArdleBH,AndersonMJ(2001)Fittingmultivariatemodelstocommunitydata:acommentondistance-basedredundancyanalysis.,290–297.MinchinPR(1987)Anevaluationoftherelativerobustnessoftechniquesforecologicalordination.,89–107.MohamedDJ,MartinyJBH(2010)Patternsoffungaldiversityandcompositionalongasalinitygradient.TheISMEJournal,379–388.hlmannO,BacherM,PeintnerU(2008)mycobiontsonanalpineprimarysuccessionalglacierforefront.Mycorrhiza,87–95.MurrayMG,ThompsonWF(1980)RapidisolationofhighweightplantDNA.NucleicAcidsResearch,4321–4325.NaraK(2006a)Ectomycorrhizalnetworkandseedlingestablishmentduringearlyprimarysuccession.Phytologist,169–178.NaraK(2006b)PioneerdwarfwillowmayfacilitatetreesuccessionbyprovidinglatecolonizerswithcompatibleectomycorrhizalfungiinaprimarysuccessionalvolcanicNewPhytologist,187–198.R.BLAALIDETAL.2011BlackwellPublishingLtd DataaccessibilityTheraw454datahavebeenaccessionedintheEuropeanNucleotideArchive,accessionno.SRP006836.1.Therepresenta-tivesequencesofthe470non-singletonOTUsareavailableinthesupportinginformationinfastaformat.SupportinginformationAdditionalsupportinginformationmaybefoundintheonlineversionofthisarticle.Fig.S1Therelationshipbetweennumbersofclustersobtainedandsequencesimilaritylevel(80-100)usedduringclusteringofsequences.Fig.S2Theabundanceofthe12mostabundantgeneraliesacrosszones,asmeasuredbypresenceabsenceofOTUsinrootsystemsandbestblasthit.Fig.S3GNMDSordinationdiagram(axes1and2)forthecompositionoffungalOTUsinthe60rootsystems,includingonlyOTUswithbestblastmatchtoectomycorrhizaltaxa.Fig.S4Frequencydistributionsofthe36mostabundantOTUsasmeasuredbypresenceabsenceinrootsystems.TableS1Characteristicsforthe60rootsamples.TableS2NumberofsharedOTUsacrossrootsystems.TableS3Representativesequencesofthe470non-singletonPleasenote:Wiley-Blackwellarenotresponsibleforthecontentorfunctionalityofanysupportinginformationsuppliedbytheauthors.Anyqueries(otherthanmissingmaterial)shouldbedirectedtothecorrespondingauthorforthearticle.R.BLAALIDETAL.2011BlackwellPublishingLtd