Ecology and Evolution. 2017;1–13.  | 1www.ecolevol.org

Received:8November2016  |  Revised:10July2017  |  Accepted:3August2017DOI: 10.1002/ece3.3343

O R I G I N A L R E S E A R C H

Close and distant: Contrasting the metabolism of two closely related subspecies of Scots pine under the effects of folivory and summer drought

Albert Rivas-Ubach1,2  | Jordi Sardans2,3 | José Antonio Hódar4 |  Joan Garcia-Porta2 | Alex Guenther5 | Ljiljana Paša-Tolić1 | Michal Oravec6 |  Otmar Urban6 | Josep Peñuelas2,3

ThisisanopenaccessarticleunderthetermsoftheCreativeCommonsAttributionLicense,whichpermitsuse,distributionandreproductioninanymedium,providedtheoriginalworkisproperlycited.©2017TheAuthors.Ecology and EvolutionpublishedbyJohnWiley&SonsLtd.

1EnvironmentalMolecularSciencesDivision,PacificNorthwestNationalLaboratory,Richland,WA,USA2CREAF,Bellaterra,Barcelona,Spain3CSIC,GlobalEcologyUnitCREAF-CSIC-UAB,Bellaterra,Barcelona,Spain4GrupodeEcologíaTerrestre,DepartamentodeBiologíaAnimalyEcología,FacultaddeCiencias,UniversidaddeGranada,Granada,Spain5DepartmentofEarthSystemScience,UniversityofCalifornia,Irvine,CA,USA6GlobalChangeResearchInstitute,CzechAcademyofSciences,Bělidla986/4a,60300,Brno,CzechRepublic

CorrespondenceAlbertRivas-Ubach,EnvironmentalMolecularSciencesDivision,PacificNorthwestNationalLaboratory,Richland,WA,USA.Email:albert.rivas.ubach@gmail.com

Funding informationCSIC;SpanishGovernment,Grant/AwardNumber:CGL2016-79835-PandOAPN022/2008;CatalanGovernment,Grant/AwardNumber:SGR2014-274;EuropeanResearchCouncilSynergy,Grant/AwardNumber:SyG-2013-610028;ASCR,Grant/AwardNumber:M200871201;MSMT,Grant/AwardNumber:CZ.1.07/2.3.00/20.0246

AbstractMetabolomes,aschemicalphenotypesoforganisms,arelikelynotonlyshapedbytheenvironmentbutalsobycommonancestry.Ifthisisthecase,weexpectthatcloselyrelatedspeciesofpineswilltendtoreachsimilarmetabolomicsolutionstothesameenvironmentalstressors.WeexaminedthemetabolomesoftwosympatricsubspeciesofPinus sylvestrisinSierraNevada(southernIberianPeninsula),insummerandwinterandexposedtofolivorybythepineprocessionarymoth.Theoverallmetabolomesdif-feredbetweenthesubspeciesbutbothtendedtorespondmoresimilarlytofolivory.Themetabolomesofthesubspeciesweremoredissimilarinsummerthaninwinter,andibericatreeshadhigherconcentrationsofmetabolitesdirectlyrelatedtodroughtstress. Our results are consistent with the notion that certain plant metabolic re-sponses associatedwith folivory have been phylogenetically conserved. The largerdivergencebetweensubspeciesmetabolomesinsummerislikelyduetothewarmeranddrierconditionsthatthenorthernibericasubspeciesexperienceinSierraNevada.Ourresultsprovidecrucialinsightsintohowibericapopulationswouldrespondtothepredicted conditions of climate change under an increased defoliation in theMediterraneanBasin.

K E Y W O R D S

drought,evolutionaryprocesses,folivory,herbivorousattack,metabolomics,Pinus sylvestris,processionarymoth,sympatricsubspecies

1  | INTRODUCTION

An organism’smetabolome consists of thousands of compounds oflowmolecularweight(metabolites)presentinanorganismatagiventime(Fiehn,2002).Suchmoleculesincludethesubstratesandprod-uctsofcellularprimarymetabolism,suchassugars,aminoacids,and

nucleotides,andofsecondarymetabolismthatareinvolvedinalargevarietyofcomplexphysiologicalprocessesformaintaininghomeosta-sisandnormalfunction.Themetabolomeisthechemicalphenotypeofanorganism(Fiehn,2002)andisthefirsttorespondtobioticandabiotic stressors (Peñuelas&Sardans,2009).The recentapplicationofnewmetabolomictechniquesinthefieldsofplantphysiologyand

2  |     RIVAS- UBACH et Al.

ecology(ecometabolomics)hasallowedthedetectionoftheextremeplasticity of metabolomes under different environmental situations(Rivas-Ubach,Sardans,Pérez-Trujillo,Estiarte,&Peñuelas,2012;Rivas-Ubach,Barbeta,etal.,2016;Sardans,Peñuelas,&Rivas-Ubach,2011).However,themetabolome,asanyotheraspectofthephenotype,canalso be subject to evolutionary divergence given thatmetabolic re-sponses ultimately depend on genetic composition and expression(Riedl etal., 2012). From this point of view,we expect that closelyrelatedorganismswillhavemoresimilarmetabolomesthandistantlyrelatedorganisms.Ifthisisthecase,environmentalchangeswouldnotnecessarilyleadtoacompletereorganizationofanorganism’smetab-olome,becausegeneticandevolutionaryconstraintswoulddetermineseveralmetabolomiccharacteristics.Therefore,closelyrelatedspeciesexposed to the same environmental conditionswould reach similarmetabolicsolutionsinresponsetosimilarenvironmentalchanges,in-cludingabioticstressors(suchasdrought),abioticfluctuations(suchasseasonalvariability),andbioticstressors(suchasherbivorepressure).

Understandingtheenvironmentalandphylogeneticcontributionsofthemetabolomeiscriticalinthecurrentcontextofglobalenviron-mental change (Edwards, Still,&Donoghue,2007;González-Orozcoetal.,2016;Kuntner,Năpăruş, Li,&Coddington,2014).Ecosystemsarecurrentlyfacinganenvironmentalchangeofplanetarydimensions,including a global increase in average temperatures and someareaswithsubstantialincreasesinaridity,suchastheMediterraneanBasin(IPCC,2002).Asidefromtheseabioticeffects,climatechangealsopro-ducesincreasesinthevirulenceofpestattacksincertainareasoftheplanet(Battistietal.,2015).Organismswillhavetorespondtoallthesechanges,particularlyplantsthatdonothavethecapabilitytodrasti-callychangealtitudinalorgeographicaldistributionsovershorttimes-cales(Chen,Hill,Ohlemüller,Roy,&Thomas,2011;Gonzalez,Neilson,Lenihan,&Drapek, 2010;Meier, Lischke, Schmatz, & Zimmermann,2012;Parmesan&Yohe,2003).Differentphylogeneticandenviron-mental contributions to themetabolome can potentially induce dif-ferent sensitivities to climate change in different plant species, andtherefore,studiesfocusingonthis issuearecriticaltounderstandinghowdifferentspeciesandecosystemswillrespondtothisglobalthreat.

Wetookadvantageofasemi-experimentalsituationwheretreesofasubspeciesofScotspineweretransplantedfromtheirnaturalhabitatincentralSpaintosouthernmountains.Weexploredwhethertheme-tabolomesofcloselyrelatedspeciesrespondsimilarlytoenvironmen-talconditionsorstressors.Scotspine(Pinus sylvestrisL.)isoneofthemostimportantandwidespreadforesttreesintheHolarctic(Ceballos&Ruizde laTorre,1971;Gausen,Heywood,&Chater,1964)andisan important species both ecologically and economically (Mäkinen&Hynynen,2014).NaturalP. sylvestrispopulations inSierraNevadaNaturalPark(southernIberianPeninsula),describedasP. sylvestrisssp.nevadensis (hereafternevadensis) (Boratynski, 1991;Ceballos&RuizdelaTorre,1971;Gausenetal.,1964),wereintensivelythinnedduringthe19thcentury(Hódar,Castro,&Zamora,2003)andrepresentthemostsouthernpopulationsofScotspine.Recently, theendemicne-vadensispopulationsareconsequentlyprotected for theirecologicalimportance (Blanca, Cueto, Martínez-Lirola, & Molero-Mesa, 1998;Hódaretal.,2003),butenvironmentalpressuressuchasdefoliation

bythepineprocessionarymoth(hereafterPPM)Thaumetopoea pityo-campa(Denis&Schiffermüller)arethreateningtheirsurvival(Castro,Gómez, García, Zamora, & Hódar, 1999; Hódar & Zamora, 2004).Manyareasoriginallycoveredbynevadensisweremassivelyreforestedlater inthemid-twentiethcenturywithScotspinesfromhigher lati-tudesofSpain(Figure1),specificallyfromtheSistemaCentralmoun-tain range (Navacerrada, central IberianPeninsula, 450kmnorth ofSierraNevada)(Robledo-Arnuncio,Navascués,González-Martínez,&Gil,2009),thatbelongstoanothersubspecies,P. sylvestrisssp.iberica (hereafteriberica).Thenevadensis(native)andiberica(introduced)sub-speciesconsequentlycoexistinsomelocalitiesofSierraNevadaandSierradeBaza (Figure2).These localitiesprovideauniqueopportu-nitytostudythemetabolomesoftwotreesubspeciesunderthesameenvironmentalconditions.Thisstudythusallowsadirectcomparisonbetweennevadensis,adaptedtothesouthernconditionswithhighertemperatureandlowerrainfallinsummerthanNavacerrada(Figures2and3),andtheintroducedibericapopulationsinSierraNevada,whichmayalreadybeexperiencingtheenvironmentalconditionsprojectedforthenearfuture(IPCC,2002).

In this study,wespecificallycontrasted themetabolomesofne-vadensisandibericaunderthesameenvironmentalconditionsofSierraNevada.Inparticular,ourmainaimwastodeterminewhethertheover-allmetabolomesofbothsubspeciestendedtoconvergeordivergetotwocommonstressors:(1)attackbyPPMcaterpillars,themaininsectdefoliatorofPinus species in theMediterranean area (Battisti etal.,2015)and(2)thenaturalsummerconditionsofSierraNevada.

On theonehand,weexpect thatclosely relatedsympatric spe-cies or subspecieswill reach similarmetabolomic solutions to bothstressors due to their shared evolutionary history; however, manystudieshavereportedalowdegreeofmetabolomicconservationbe-tweenplantspeciesofthesamegenusexposedtodrought(Sánchez,Schwabe,Erban,Udvardi,&Kopka,2012)orherbivorousattack(Rivas-Ubach,Hódar, etal., 2016), suggesting thatmetabolic responses tostressorscouldbetaxonspecific.Totestthesehypothesesinnatural

F IGURE  1 ScotspineforestsofColladodeMatasverdesinSierraNevadaNationalPark.ThephotographillustratespartofthestudysitewherePinus sylvestrisssp.nevadensis(native)coexistswithP. sylvestrisssp.iberica(introduced).PhotographbyDr.JoséAntonioHódar

     |  3RIVAS- UBACH et Al.

populationsoftrees,wesampledneedlesofbothP. sylvestrissubspe-ciesinSierraNevadaNaturalParkinwinter,themainperiodofPPMattack,andinsummer,oncethetreeshadflushedtheirneedlesafterPPMattack.Thefoliarmetabolomesofattackedtrees(ATs)andnon-attackedtrees(NATs)ofbothsubspecieswereanalyzedbyliquidchro-matographycoupledtomassspectrometry(LC-MS),andtheobtaineddataweresubsequentlysubmittedtodiverseunivariateandmultivari-atestatisticalanalysestoaddressourhypotheses.ThisstudyprovidescrucialinformationofthemetabolomicflexibilityofScotspinetotheattackbyPPMsandtoclimatechange,producingkeyknowledgeforabettersustainablemanagementofpineinacontextofclimatechange.

2  | MATERIAL AND METHODS

2.1 | Study site

Foliar samples were collected in the P. sylvestris forests of ColladodeMatasverdes (37.05°N, 3.27°W; 1,900m a.s.l.) in SierraNevada

National Park (Granada, SE Spain; Figure2a,b,d,f), where nevaden-sis coexistswith iberica in the samevalley (Robledo-Arnuncioetal.,2009)(Figures1and2).Navacerrada,thenativeareaofibericatrees(Figure2a),hasameanannualtemperature(MAT)of6.4°C(Figure2c)andameanannualprecipitation(MAP)of1,330mm(Figure2e).Theclimate inSierraNevada ismoreMediterranean,withhotsummers,coldwinters,andusuallyaseveresummerdrought.TheMATis9.8°C(Figure2d),andtheMAPis945mm(Figure2f).Theminimumtemper-aturesinwinterandthemaximumtemperaturesinsummer,togetherwith precipitation, are crucial factors determining the distributionalniche of most plant species. The average minimum and maximumtemperatures and the accumulated rainfalls forwinter and summerfor Navacerrada and the study site in Sierra Nevada are shown inFigure3.Navacerradaclimaticdatawereextractedby interpolatingthedatafromthefourstationsclosesttothe ibericanaturalpopula-tionsfromthedatabaseAEMET(www.aemet.es).SierraNevadadatawere obtained from a meteorological station 1km from the studysite,manageddirectlybytheNationalParkadministrationattheLa

F IGURE  2 Native(green)andplanted(blue)populationsofScotspineacrossSpain(a)andinSierraNevadaandSierradeBaza(b).NativepopulationsinSierraNevadaandSierradeBazaarePinus sylvestrisssp.nevadensis.Mapsofmeanannualtemperature(MAT)forNavacerrada,thenativehabitatofP. sylvestrisssp.ibericaplantedinsouthernSpain,(c)andforSierraNevada,thestudysite,(d)andofmeanannualprecipitation(MAP)forNavacerrada(e)andSierraNevada(f)areprovided.NativepopulationsofP. sylvestrisssp.ibericainNavacerradaaredelimitedbytheblacklinesin(c)and(e).ThestudysiteisidentifiedbythebluecirclesinthemagnificationsofSierraNevadain(d)and(f).MapswiththedistributionofnativeandintroducedP. sylvestriswoodlandswerekindlysupportedbytheSpanishnetworkonGeneticsandConservationofForestResources(GENFORED).Climaticmapswereobtainedfromthe“AtlasClimáticoDigitaldelaPenínsulaIbérica”(Ninyerola,PonsFernandez,&RoureiNolla,2005)

4  |     RIVAS- UBACH et Al.

CortijuelaBotanicalGarden.We considered January, February, andMarchaswinterandJuly,August,andSeptemberassummer.

Foliar sampleswere collected in earlyMarch 2011 (winter) andmid-July2011(summer).ThePPMlate-instarlarvaeinflicttheheavi-estdamagetopines in latewinter (Battistietal.,2015;Hódaretal.,2003),andtheneedlesofthecurrentyeararecompletelyflushedinJulyduringthetypicaldroughtoftheMediterraneansummer.

2.2 | Experimental design and sampling of needles

Twenty-fourmature iberica andnevadensis trees,>45yearsoldand>6m inheight,wereusedasstudycases (totaln=48).Allsampledindividualswerewithinaradiusofapproximately500minthesamevalleyat thesamealtitude.Werandomlyselected12treesofeachsubspeciesattackedbyPPMs(ATs),easily identifiedbytheirwintertents(2–4pertree),and12treesofeachsubspecieswithnosignsofcaterpillarattack(NATs).NATsandATsweresampledsimultaneouslyand within a short period of time (10:30–14:30) under sunny andconstant light and temperature conditions to avoid large variationsofmetabolomesdue to circadian rhythms (Kim,Choi,&Verpoorte,2010;Rivas-Ubachetal.,2013).AsmallbranchexposedtothesunwasremovedfromtheNATswithapruningpole.Asmallnonattackedbranch(NAB)andasmallattackedbranch(AB),bothexposedtothesun,werealsoremovedfromATs(seeFig.S1).NABswerecollectedbetween3 and4m away from thePPMattack ensuring at least 1branchbetweenthefocusofattackandthesampledundamagednee-dles.These foliar sampleswillbe referredasAT.NABsandAT.ABs,

respectively. Theyoungestwell-developedneedles fromeach sam-pledbranchwerequicklyfrozeninsituinliquidnitrogenfortheme-tabolomicanalyses.

Periodic outbreaks of PPMs occur in Spain and France with areturn period of 5–9years, and infestation is more unpredictablein someother areasof theMediterraneanBasin (Hódar, Zamora,&Cayuela,2012;Li,Daudin,Piou,Robinet,&Jactel,2015;Tamburini,Marini,Hellrigl,Salvadori,&Battisti,2013).Theintensityoftheout-breaks,however,wasvariableanddefoliationcouldbepatchilydis-tributed, even at sites heavily attacked by PPMs. A mild outbreakoccurred inthestudyarea inwinter2010,duringwhichmanytreeswerenotaffectedbythePPM,andthePPMpopulationdeclinedaftertheoutbreakduringthewinterof2011,theyearwhenthesampleswerecollected.

Pineprocessionarymothattacksvarydependingonpinespecies,and preferences can alsovary from site to site (Jactel etal., 2015).Differences inattackpreferencebetweensubspeciesofpines,how-ever, have not been documented.At our study site,nevadensis andiberica were equally attacked, and any possible differences wouldnotaffectourresultsbecauseourselectionoftrees inthewildwasbasedonthepresence/absenceofnaturaldefoliation.Thisselectiondetermines that our pines were not completely randomly assignedto the different folivory levels (NAT orAT).Moths inmonospecificstands,as inourcase,however,wouldmainlyrelyonvisualcuestoattack isolatedor taller trees thatweremore likely toprovideopti-mal microclimatic conditions (high solar radiation) for egg survivalandsuccessfullarvaldevelopmentinsteadofonchemicaldifferences

F IGURE  3 Averagemaximum(soliddots)andminimum(opendots)temperatures(°C)andaccumulatedrainfall(mm)(solidsquares)forNavacerradaandSierraNevadaNaturalParkforwinterandsummer.Thetemperaturesaremeans±3SE.Theaccumulatedrainfallsaremeans±SE.DifferentlettersdenotesignificantdifferencesbetweenNavacerradaandSierraNevada(p<.05)identifiedbyattest(seeTableS3)

     |  5RIVAS- UBACH et Al.

betweenindividuals(Jacteletal.,2015).Theassignmentofattacked/unattackedtreatmentsbyfemalemothswhenovipositingcanthusbereliablyconsideredasarandomselectionofthepriorchemistryofthetrees.TheNATsthusservedascontrols,andtheAT.ABsandAT.NABswereusedtodeterminethelocalandsystemicresponsestofolivory,respectively,andtorepresentfolivorylevels(FLs).

2.3 | Foliar processing for metabolomic analyses

Briefly,needlesfrozeninliquidnitrogenwerelyophilizedandstoredinplasticcansat−20°C(Rivas-Ubachetal.,2013).Samplesweregroundwith a ball mill at 1,600rpm for 8min (Mikrodismembrator-U; B.BraunBiotechInternational,Melsungen,Germany),producingafinehomogeneouspowderthatwasstoredat−80°Cuntiltheextractionofthemetabolites(Rivas-Ubachetal.,2013).

2.4 | Extraction of metabolites for liquid chromatography- mass spectrometry analyses

Wefollowedawell-establishedprotocol for theextractionofpolarandsemi-polarmetabolites(t’Kindt,DeVeylder,Storme,Deforce,&VanBocxlaer,2008)withminormodifications.First,twosetsof2-mlcentrifugetubeswerelabeled:setAfortheextractionsandsetBfortheextractsfromsetA.OnehundredmilligramofthesamplepowderwasweighedintoeachtubeofsetA,and1mlofMeOH/H2O(80:20)wasaddedasanextractant.Alltubeswerevortexedfor15min,soni-catedfor5minatroomtemperatureandthencentrifugedat23,000g for5min.Aftercentrifugation,0.6mlofthesupernatantfromeachtubeofsetAwas transferred to thecorresponding2-mlcentrifugetubesofsetB.Thisprocedurewasperformedtwicefortwoextrac-tions of each sample. The tubes of set Bwere then centrifuged at23,000gfor5min,andthesupernatantswerecollectedbyglasssy-ringes,filteredthrough0.22-μmporemicrofiltersandtransferredtoalabeledsetofHPLCvials.Thevialswerestoredat−80°CuntiltheLC-MSanalysis.

2.5 | LC- MS analyses

Themetabolomicfingerprintsofpolarandsemi-polarmetabolitesofpineleaveswereobtainedbyLC-MSanalyses.LC-MSchromatogramswereobtainedusingaDionexUltimate3000HPLCsystem(ThermoFisherScientific/DionexRSLC,Dionex,Waltham,USA)coupledtoanLTQOrbitrapXLhigh-resolutionmassspectrometer (ThermoFisherScientific) equippedwithanHESI II (heatedelectrospray ionization)source.A reversed-phaseC18Hypersilgoldcolumn (150×2.1mm,3 μmparticlesize;ThermoScientific)at30°Cwasusedforchromatog-raphy.Themobilephasesconsistedofacetonitrile(A)andwater(0.1%acetic acid) (B).Bothmobile phaseswere filtered anddegassed for10mininanultrasonicbathpriortouse.Theelutiongradientbeganat10%A(90%B)ataflowrateof0.3ml/minandwasmaintainedfor5minandthento10%B(90%A)untilminute20andheldfor5min.Theinitialproportions(10%A,90%B)weregraduallyrecoveredoverthenext5min,andthecolumnwaswashedandstabilizedfor5more

minutesbeforeinjectionofthenextsample.Theinjectionvolumeofthesampleswas5μl.Allsampleswereinjectedtwice,oncewiththeHESIoperatinginnegativeionizationmode(−H)andonceinpositiveionizationmode(+H).TheOrbitrapmassspectrometerwasoperatedinFourierTransformMassSpectrometryfull-scanmodewithamassrangeof50–1,000m/zandhigh-massresolution(60,000).Thereso-lutionandsensitivityofthespectrometerweremonitoredbyinjectingastandardofcaffeineafterevery10samples,andtheresolutionwasfurthermonitoredwithlockmasses.Blanksampleswerealsoanalyzedduringthesequence(seeRivas-Ubach,Sardans,etal.(2016)formoredetailsoftheLC-MSanalyses).

2.6 | Processing of LC- MS chromatograms

TherawdatafilesfromthespectrometerwereprocessedbyMZmine2.17(Pluskal,Castillo,Villar-Briones,&Orešič,2010).Chromatogramsfrom both positive and negative modes were separately baselinecorrected, deconvoluted and aligned before themetabolic assigna-tion. For each database generated (positive and negative),metabo-liteswereidentifiedbyexactmassandretentiontimebasedonthemeasurementsofthestandardsintheMSwithtotalexactmasswiththeautomaticassignationfunctionofthesoftware(seeTableS1fordetailsandTableS2fortheidentifiedmetabolites).Thenumericalda-tabaseswere thenexportedtoaCSVsheet.Chromatogrambuilderand deconvolution algorithms may separate diverse ions with thesamemass to ratio (m/z) into different variables due a slight shiftin retention times,dependingon theparametersestablishedduringchromatogram processing for obtaining themetabolomic data sets.All identified features corresponding to the same molecular com-poundswerethussummedtoobtainonlyonevariablepermetabolite.Furthermore,mostcarbohydratesco-elutedatverysimilarretentiontimeswithourchromatographicmethodinareversed-phaseC18col-umnwhichmakesimpossibletodifferentiatethemwhentheysharethesameexactmass.Forthisreason,someofthedetectedcarbohy-drateswerethusclassifiedintodifferentgroupsbasedontheirreten-tiontimeandm/z(“hexoses”(Hex)forfructose,glucose,galactoseandmannose;“pentoses” (Pent) forribose,xyloseandarabinose;“disac-charides” (Disacch) for saccharoseandmaltose;group1sugars (S1)fordeoxy-galactose,deoxy-glucoseandD-fucose;group2sugars(S2)forraffinoseandmaltotriose;andgroup3sugars(S3)forarabitolandxylitol).

Metabolomicvariablespresentinfewerthaneightindividualsofacellfactorwereremovedfromthedataset.Valuesofaspecificvari-able threefoldhigher thanthethirdquartileor threefold lower thanthe firstquartileofeachcell factorwereconsideredasoutliersandweresubsequentlytreatedasmissingdata.

Thenumericalvaluesof the featuresextracted fromtheLC-MCchromatogramscorrespond to theabsolutepeakareasof the chro-matogramsdetectedbythespectrometer.Theintegratedpeakareasfrom thedeconvolutedpeak chromatogramsdonot reflect the realconcentrationasunitweightofmetaboliteperunitweightofthesam-ple,but it isproportional totheconcentrationof thecorrespondingvariableandsoissuitableforcomparativeanalyses,asdemonstrated

6  |     RIVAS- UBACH et Al.

inseveralmetabolomicstudies(Lee&Fiehn,2013;Marietal.,2013;Rivas-Ubachetal.,2014;Rivas-Ubach,Sardans,etal.(2016)).Wethususethetermconcentrationwhenreferringtotherelativeconcentra-tionsofthemetabolitesamongthefactorsstudied(season,subspe-cies,andFL).

2.7 | Data analyses

TheaverageminimumandmaximumtemperaturesandaccumulatedrainfallforwinterandsummerinNavacerradaandSierraNevadawerecomparedwith a t test (Table S3) to determine if the environmen-talconditionsdifferedsignificantlybetweenlocalities ineachoftheseasons.

Forthemetabolomicdata,Shapiro–WilkandLevene’stestswereperformedoneachvariabletoassessnormalityandhomogeneityofvariances.Allknownvariablesmettheassumptionsfortheposterioranalyses of variance (ANOVAs), and any unidentified metabolomicvariablethatdidnotmeettheassumptionswasremovedfromthedataset(127variablesremovedfromtheinitialdatasetof8,492variables).Thedatasetforthisstudywasthuscomposedofthreeindependentfactors,season(winterandsummer),subspeciesofP. sylvestris(iberica andnevadensis) andFL (NATs,AT.NABs,andAT.ABs), andcontained8,365metabolomicvariables,72ofwhichwereidentifiedbyourme-tabolitelibrary(TableS2).

The metabolomic fingerprints for the P. sylvestris needles weresubjected to a permutational multivariate analysis of variance(PERMANOVA)usingEuclideandistancestotestfordifferencesintheoverallmetabolomesbetweenseasons,subspecies,andFLs.Thenum-berofpermutationswassetat10,000.

Univariant analyses consisted in one-wayANOVAs, andTukey’sHSDposthoctestswereperformedforeachindividualidentifiedme-tabolite foreachsubspeciesandseasonseparatelywithFLascate-goricalfactor(TableS4).TheentiresetofpvaluesfromtheunivariateanalyseswassubsequentlysubmittedtoaBenjamini–Hochbergcor-rectiontesttocontrolforfalsepositives.Theresultoftheseone-wayANOVAs showedwhether therewere statistically significant differ-encesbetweentheFLs(NATsvs.AT.NABsvs.AT.ABs)(TableS4).

Thefoliarmetabolomicfingerprintsofbothpinesubspecieswerealsosubjectedtoprincipalcomponentanalysis(PCA)foreachseasonseparately.PCAsreducethedimensionalityofadatasetintotypicallytwodimensions(PC1vs.PC2),andsamplesandvariablesareprojectedonthefactorplaneconstrainedbythetwodimensions.WeusedPCAtounderstandthemetabolomictrendsofFLsandsubspeciesandtoshedlightintherelationsbetweenvariablesandstudysubjects.ThescorecoordinatesofthevariablesofthetwofirstPCAaxesweresub-jectedtoone-wayANOVAsandTukey’sHSDposthocteststoidentifystatisticaldifferencesamongthemetabolomefingerprintingofthedif-ferentgroups(subspeciesandFLs)acrossthevariabilityexplainedbythetwofirstaxesofthePCA(Rivas-Ubachetal.,2013).

Additionally, we used Euclidean distances as a proxy to deter-minethedistancebetweenthemetabolomes(metabolomicdistances)amongdifferentgroupsoftrees.TheEuclideandistancesbetweenthemetabolomic fingerprints of iberica and nevadensis were calculated

foreachindividualtreewithinthesameFL(eachiberica-NATvs.eachnevadensis-NAT, each iberica-AT.NAB vs. each nevadensis-AT.NABandeachiberica-AT.ABvs.eachnevadensis-AT.AB)(Fig.S2a–c).ThoseEuclideandistancesarereferredas“FLdistancesbetweensubspecies”alongthetext.Allthe144FLdistancesbetweensubspecies(12iberica trees×12nevadensistreesforNATs,AT.NABsandAT.ABs)weresub-sequently submitted to a one-way ANOVA considering winter andsummer together and for each season separately (6 groups: 2 sea-sons×3FLs).ANOVAswereappliedto identifyforstatisticaldiffer-encesbetweenFLdistancesbetweensubspecies.EuclideandistancesbetweenFLswithineachsubspecies(eachNATvs.eachAT.NABandeachNATvs. eachAT.AB, separately for iberica and fornevadensis)werealsocalculated foreachseason (Fig.S2d,e).Thesesecondcal-culatedEuclideandistancesarereferredas“FLdistanceswithinsub-species”alongthetextandwerealsosubmittedtoaone-wayANOVAconsideringwinterandsummertogether(8groups:2seasons×2dis-tances×2 subspecies) to identify differences betweenFL distanceswithinsubspecies.

MultivariateanalysessuchasPERMANOVAs,PCAs,andEuclidean-distance calculations were performed using the complete data set(identifiedandnonidentifiedvariables).ThePERMANOVAs,one-wayANOVAs,ttests,Tukey’sposthoctests,Shapiro-Wilktests,Levene’stests,PCAs,andEuclidean-distancecalculationswereperformedwithR(RCoreTeam,2013).TheShapiro-Wilktests,one-wayANOVAs,t tests,Benjamini-Hochbergcorrection,andEuclidean-distancecalcu-lationswereperformedwiththefunctionsshapiro.test,aov,p.adjust, t.test,anddist,respectively,inthe“stats”package(RCoreTeam,2013).Tukey’sHSDposthoctestswereperformedbytheHSD.testfunctionofthe“agricolae”package(deMendiburu,2015).Levene’stestswereperformedwith the leveneTest function in the “car” package (Fox&Weisberg,2011).ThePERMANOVAwasconductedwith theadonis functioninthepackage“vegan”(Oksanenetal.,2013).Usingthereg-ularizediterativePCAalgorithm(Josse&Husson,2013)withtheim-putePCAfunctionofthe“missMDA”package(Husson&Josse,2015),themissingvaluesofthedatasetwereimputedbeforethePCA.Oncethemissingvalueswereimputed,PCAswereperformedwiththePCA functionof the “FactoMineR” package (Husson, Josse, Le,&Mazet,2016).

3  | RESULTS

Our univariate analyses contrasting the environmental data (tem-perature and accumulated rainfall) between Sierra Nevada andNavacerradashowedmorecontrastedenvironmentalconditionsbe-tween localities insummercomparedtowinter (Figure3;TableS3).Average maximum temperatures were higher in both winter andsummer in Sierra Nevada, the native locality for nevadensis, thanNavacerrada,thenativelocalitiesforiberica.Averageminimumtem-peratureandaccumulatedrainfalldidnotdiffersignificantlybetweenthetwoareasinwinter,buttheaverageminimumtemperaturewashigher and accumulated rainfall was lower in Sierra Nevada thanNavacerradainsummer.

     |  7RIVAS- UBACH et Al.

Theoverallcompositionofthepinemetabolomeschangedsignifi-cantlybetweenseasons,subspeciesandFLs(Table1).PERMANOVAalso found significant differences in all the factor interactions (sea-son×subspecies; season×FL; subspecies×FL; season×subspe-cies×FL) (Table1) indicating that the impact of any of the studiedfactor(season,subspecies,andFL)onthestructureoftheoverallme-tabolomesofpinesalsodependsontheleveloftheothertwofactorstogetherandseparately.

ThePCAperformed inwinter,when thePPM is active, showedthatthedifferentlevelsofFLsofbothsubspeciesfollowedthesametrendalongthePC1suggestingthuscertaingradeofsimilarityonthemetabolomicresponsestoPPMattackinbothsubspecies(Figure4a).PC2 of the PCA inwinter displayed clear separation between sub-species indicating their different overall metabolome composition(Figure4a).PCAsummershoweddifferenttrendsforsubspeciesandFLs. Subspecies still clustered separately but along PC1 (Figure4c).Althoughnotas clearas in thewinterPCA,FLsofboth subspeciesalsofollowedsimilartrendsalongPC1andPC2(Figure4c).Forbothsubspecies,NATsandAT.NABswereseparatedfromtheAT.ABsalongPC1while PC2 separated the NATs from theAT.NABs andAT.ABs(Figure4c).Additionally, FLsof iberica in summerwere clearlymoresegregatedamongthemthanFLsofnevadensis(Figure4c).Thesere-sultsindicatethatthemetabolomeshiftsbetweenFLswerelargerinibericathannevadensisinsummer.

Regarding the relations between themetabolomicvariables andstudy subjects, the PCA showed that the concentrations of mostaminoacids,sugars,phenoliccompounds,andterpenestendedtobehigher innevadensis than ibericaneedles inwinter (Figure4a,b).TheAT.ABsofbothsubspecieshadthehighestconcentrationsofvitexin,catechin,carvone,disaccharides,andδ-tocopherol,and theNATsofboth subspecies had the highest concentrations of amino acids, ni-trogenousbases(adenine,guanine,andcytosine),andmostphenoliccompounds(Figure4a,b;TableS4).Nevertheless,theconcentrationsofaminoacids,phenoliccompounds,andmostsugarswerehighestinibericainsummer,especiallyfortheNATsandAT.NABs(Figure4c,d).TheNATsofbothsubspecieshadhigherconcentrationsofmostaminoacids,organicacids,sugars,nitrogenousbases,andmostphenolicsrel-ativetotheATs(Figure4c,d;TableS4).

Inwinter,theFLdistancesbetweensubspeciesdidnotdiffersig-nificantly (Figure5a). In summer, however, the distance for iberica NATsversusnevadensisNATswashighest,whereas thedistance foriberica-AT.ABs versus nevadensis-AT.ABs was lowest (Figure5a).Those results indicate that, in summer, when the environmentalconditions between Sierra Nevada andNavacerrada aremore con-trasted (Figure3), NATs between subspecies had more contrastedmetabolomes compared to AT.ABs which presented the smallestmetabolicdifferencesbetweenibericaandnevadensis.Inbothwinterand summer, FLs distances within nevadensis did not differ signifi-cantly among them (Figure5b). Differently, iberica had higherNATsversusAT.ABsdistance thanNATsversusAT.NABsdistance inbothseasons(Figure5b).Additionally,bothNATsversusAT.ABsandNATsversusAT.NABsdistancesforibericaweresignificantlyhigherinsum-mercomparedtowinter(Figure5b).Nonetheless,FLdistanceswithin nevadensiswerehigher inwinter,when thePPM ispresent, than insummer (Figure5b).Those results indicated largermetabolomic dif-ferencebetweensubspeciestoPPMattackinsummerthaninwinter.

4  | DISCUSSION

TheconclusionsofthepresentstudydirectlyrelyonthecomparisonofthemetabolicresponsesoftwocoexistingScotpinesubspeciesinSierraNevada, one native (nevadensis) and one introduced (iberica).Althoughthisstudydoesnot include ibericasamplesfromitsnativeregion (Navacerrada) to definitely corroborate our conclusions, themultivariateanalysesofthemetabolomesofbothsubspeciesclearlyindicateclosermetabolomicresponsestoPPMattackbetweensub-speciesandmoredistantmetabolomicresponsestothemarkedsea-sonalityofSierraNevada.

4.1 | Close metabolic responses to folivory attack

DefoliationbyPPMoccursduringwinter(Battistietal.,2015);interest-ingly althoughdifferent FLs presenteddifferent overallmetabolomestructureinbothsubspecies(Table1),thewintermetabolomesofthesubspeciestendedtoconvergeinresponsetolocalcaterpillarattack.

TABLE  1 FullfactorialPERMANOVAmodelofthecompletemetabolomicdatasetconsideringallfactorsandinteractions:season,subspecies,folivorylevel(FL),season×subspecies,season×FL,subspecies×FL,season×subspecies×FL,andresiduals

Degrees of freedom Sums of squares Mean squares Pseudo- F p

Season 1 3.60×1020 3.60×1020 94.75 .0001

Subspecies 1 1.37×1020 1.37×1020 35.904 .0001

Folivorylevel(FL) 2 4.39×1019 2.20×1019 5.776 .0001

Season×Subspecies 1 8.55×1019 8.55×1019 22.469 .0001

Season×FL 2 4.74×1019 2.37×1019 6.231 .0001

Subspecies×FL 2 3.09×1019 1.55×1019 4.067 .0004

Season×Subspecies×FL 2 5.52×1019 2.76×1019 7.259 .0001

Residuals 132 5.02×1020 3.80×1018 0.39784

Total 143 1.26×1021 1

8  |     RIVAS- UBACH et Al.

F IGURE  4 PC1versusPC2ofthePCAsofthefoliarmetabolomesofPinus sylvestrisssp.ibericaandP. sylvestrisssp.nevadensisforwinterandsummer:case(a)andvariable(b)plotsofthePCAforwinter,andcase(c)andvariable(d)plotsofthePCAforsummer.Thefolivorylevels(FLs)arerepresentedbydifferentcolors:blue,NATs;green,AT.NABsandred,AT.ABs.Crossesrepresentnevadensisanddotsrepresentiberica. ThecoloredarrowsindicatethecoordinateaveragesofPC1andPC2ofeachfolivorylevel(FL)foribericaandnevadensis.DifferentlettersbesidethearrowsindicatesignificantdifferencesbetweenFLsforeachsubspeciesdetectedbyTukey’sHSDposthoctests(p <.05).Differentmetabolomicfamiliesareindicatedbydifferentcolors:blue,sugars;green,aminoacids;cyan,nucleotides;orange,organicacidsassociatedwiththetricarboxylicacidcycle(TCA);violet,phenolics;brown,terpenes;darkred,othersecondarymetabolites.Unidentifiedmetabolitesarenotrepresentedinthevariableplot.Mostmetabolitesarereferencedbyabbreviations:disaccharides(Disacch),hexoses(Hex),pentoses(Pent),group1sugarsrepresentingdeoxy-glucose,deoxy-galactose,andD-fucose(S1),group2sugarsrepresentingraffinoseandmaltotriose(S2),group3sugarsrepresentingxylitolandarabitol(S3),alanine(Ala),arginine(Arg),asparagine(Asn),asparticacid(Asp),glutamicacid(Glu),glutamine(Gln),glycine(Gly),isoleucine(Ile),leucine(Leu),lysine(Lys),methionine(Met),phenylalanine(Phe),proline(Pro),serine(Ser),threonine(Thr),tryptophan(Trp),tyrosine(Tyr),valine(Val),adenine(Adeni),adenosine(Adeno),cytosine(Cyt),guanine(Gua),uridine(Uri),uracil(Ura),α-ketoglutaricacid(a.keto.ac),citricacid(Cit.ac),lacticacid(Lac.ac),malicacid(Mal.ac),oxaloaceticacid(Oxa.ac),pyruvicacid(Pyr.ac),succinicacid(Succ.ac),acacetin(Acac),apigenin(Apig),caffeicacid(Caff.ac),catechin,catechol,coumaricacid(Coum.ac),D-pinitol,epicatechin(Epicat),epigallocatechin(Epigallocat),ferulicacid(Fer.ac),galangin(Gal),kaempferol(Kaemp),myricetin(Myr),quercetin(Quer),quinicacid(Qui.ac),resveratrol(Resve),rhamnetin(Rhamn),robinetin(Rob),sodiumsalicylate(S.sal),taxifolin(Tax),vitexin(Vit),vanillicacid(Van.ac),5,7-dihydroxy-3,4,5-trimethoxyflavone(Flavone:Flav),choline,δ-tocopherol(d.tocoph),eugenol,vitaminB5(Vit.B5),shikimicacid(Shik.ac),sabinene(Sabi),carvone(Carvo),caryophyllene(Caryo),farnesol(Farn),abscisicacid(ABA),gibberellicacid1(GA1),andgibberellicacid3(GA3)

     |  9RIVAS- UBACH et Al.

PC1of thewinterPCAcaseplot (Figure4b) separated theFLs, andAT.ABsandAT.NABsofbothsubspeciesclearlyhadthesamedirec-tionrespecttotheNATsalongPC1suggestingthatbothsubspeciespresentedcertainsimilarityinthemetabolomicresponsestofolivory.Inparticular,theAT.ABsofbothsubspecieshadthehighestconcen-trationsofvitexin,catechin,carvone,disaccharides,andδ-tocopherol(Figure4a,b;TableS4),metaboliteswhichhavebeendirectlyassoci-ated with folivory. Folivory causes oxidative stress in plants (Bi &Felton,1995;Ruuhola&Yang,2006),andδ-tocopherolandflavonoidssuchasvitexinandcatechinhavebeencommonlyconsideredas im-portantantioxidantsinplants(Apel&Hirt,2004;KhorasaniEsmaeili,MatTaha,Mohajer,&Banisalam,2015;Kimetal.,2005;Ramanetal.,2016; Rice-Evans,Miller, & Paganga, 1996; Singh, Sahu, & Sharma,2017).Hernández,Alegre,VanBreusegem,andMunné-Bosch(2009),however,reviewedtheroleofflavonoidsinplantsandconcludedthattheantioxidantfunctionofflavonoidsinplantsisstillamatterofde-bateduethelackofastrongspatiotemporalcorrelationbetweenoxi-dativestressandflavonoidoxidation,eventhoughdifferentspeciesofflavonoidscanaccumulate in invitroplantsunderoxidativestress inresponse todiversebiotic and abiotic stressors.On theother hand,transgenicplantsthatover-expressthegenesthatcodeforflavonoidshaverecentlybeenusedtoproduceprogenywithimprovedantifungal

andantioxidativeproperties (Mierziaketal., 2014;Ravensdaleetal.,2014).Althoughmost studies agree that flavonoidshave strong an-tioxidantpropertiesinplants,furtherresearchisnecessarytodeeplydecipherthespatiotemporalmechanismofflavonoidswithplantoxi-dationstress.Carvonehasbeendescribedasaterpenewithrepellentandantifeedantproperties inpines (Schlyter,Smitt,Sjödin,Högberg,& Löfqvist, 2004). Some studies have demonstrated higher rates ofsucrosesecretionindamagedleavesthatcanattractmoreinsectvisi-tors (Ness,2003;Rivas-Ubachetal.,2014)andthatcanaccount forthehigherconcentrationsofdisaccharidesintheAT.ABs(Figure4a,b;Table S4). Amore detailed discussion of the functional roles of themetabolomic differences between FLs in both subspecies has beenpublishedelsewhere(Rivas-Ubach,Sardans,etal.,2016).Additionally,insummer,whenthetemperaturesanddroughtinSierraNevadaaremoreextremethaninNavacerrada(Figures2and3),theAT.ABsdis-tancebetween subspecieshave similar values respect towinter butNATsdistancebetweensubspeciesincreasedsignificantly(Figure5a).TheseresultsalsosuggestthattheAT.ABsofbothsubspecieshadmorecommonmetabolomicresponsesinsummer,likelyforcopingwiththeinjuriesfromthelastfolivory(Rivas-Ubach,Sardans,etal.,2016).

The convergence of the foliar metabolomes of the subspeciesinthe localresponsestoPPMattack(AT.ABs) isconsistentwiththe

F IGURE  5 Metabolomicdistance(Euclideandistances)betweenibericaandnevadensismetabolomesforeachfolivorylevel(FL;NATs,AT.NABs,AT.ABs)andseason(winter,summer)(a).MetabolomicdistancesbetweenNATsversusAT.NABs(circles)andNATsversusAT.ABs(triangles)withineachsubspeciesandforeachseason(b).Solidandopencirclesandtrianglesrepresent,respectively,toibericaandnevadensis inpanelb.Valuesrepresentthedistancemeans±3SE.Differentlettersdenotesignificantdifferencesidentifiedbyone-wayANOVAsandHSDposthoctests(p<.05).TheANOVAforpanelbusedbothwinterandsummerdata

10  |     RIVAS- UBACH et Al.

notion that severalmetabolic responses inducedbyherbivorousat-tackmay be evolutionarily conserved (Carrillo-Gavilán etal., 2015).OverallmetabolomicresponsestoPPM,however,havebeenreportedtodifferinthreepinespecies,buttheresponsesweredirectlyrelatedtothephylogenyofthepines,suggestingagainparticularevolution-arilyconservedresponsestoPPMattack(Rivas-Ubach,Hódar,etal.,2016).OurstudyanalyzedtwosubspeciesofP. sylvestristhatarehis-toricallymoreclosely related than individualsbelonging todifferentspecies,sotheconvergencebetweenthesubspeciesinmetabolomicresponsestoherbivorousattack inouranalysesofthemetabolomicfingerprints also suggests a strong genetic component determiningthe responses. If metabolomes track phylogeny, we would expectthatdistantly relatedspecieswouldproducedivergentmetabolomicresponsestoherbivorousattack.Futureresearchanalyzingdifferentplant species attacked by the same folivorewould be necessary tostronglysupportthisstatement.

4.2 | Distant metabolomic responses to summer drought

Despitebelongingtothesamespecies,theseasonaloverallmetabo-lomicdifferencesbetweenibericaandnevadensis(Table1)mayhavebeenduetoboththegeneticcomponentbetweensubspeciesandtothe atypically extreme environment experienced by the introducedibericainsummer(Figures2and3;TableS3).Wefounddiverseme-tabolomicevidencethatibericainSierraNevada(introduced)experi-encedamoreextremeenvironmentinsummerrelativetonevadensis (native). First, FLs distances between subspecies did not differ sig-nificantlyamongtheminwinter,butinsummer,whenthePPMisnotpresent,NATscdistancebetweensubspecieswassignificantlyhigherthanallFLsdistancesbetweensubspeciesinwinter(Figure5a).ThisresultsuggeststhatAT.ABsofbothsubspeciesretainsomecommonresponsesinsummertoPPMattack,butNATs,whichdonotneedtocopewithfolivoryinjuries,aremetabolicallymoredifferentbetweensubspecies. This trend was not found in winter when the climaticconditionsaremoresimilarbetweenSierraNevadaandNavacerrada,butmaximumtemperaturesarehigherandrainfall is lower inSierraNevadaduringsummer,sodroughtismoresevereinSierraNevadathanNavacerrada(Figure3;TableS3).Second,thePCAforsummerindicatedthatmetabolomicvariationbetweentheFLswaslowerfornevadensis than for iberica. The metabolomes of the iberica NATs,AT.NABsandAT.ABsweremoreclearlyseparatedthantheFLsofthenevadensistreesinthemultidimensionalspaceofthePCA(Figure4a).This PCA thus indicated that the iberica responses to PPM attackwerelargerthannevadensisinsummer,whenthePPMisnotpresent.ThirdandalsosupportedbythePCA,theFLsdistanceswithinsub-species insummerpresentedlargerdifferencesbetween ibericaandnevadensisthaninwinter,whenonlytheibericaNATsversusAT.ABsdistancewassignificantlylarger(Figure5b).Thisdivergencebetweennevadensis and iberica in summer clearly supports the idea that theenvironmentwasmore influential in summer in iberica, the subspe-ciesintroducedtoSierraNevada.Fourth,independentlyoftheFL,theconcentrationsofsugars,phenolics,andaminoacidssuchasproline

(Figure4c,d) in summerwere higher in iberica thannevadensis nee-dles,alsosuggestingthatibericaexperiencedmorestressinsummer.Inpreviousecometabolomicstudies,wehavealsoobservedincreasesin sugar, amino acid, andphenolic concentrations in different plantspeciesduringthedrysummersoftheMediterraneanclimate(Rivas-Ubachetal.,2012,2014;Rivas-Ubach,Barbeta,etal.,2016).Prolineisanimportantosmoprotectantinplants(Szabados&Savouré,2010),so the higher proline concentrations in the needles of iberica com-paredtonevadensisalsosuggest that thissubspeciesmaybefacingmore intensedrought conditions in SierraNevada than those in itsnativerangeinNavacerrada(Figure3).Thehigherconcentrationsofphenolic compounds and sugars in iberica than nevadensis needleshave been also widely reported as protective mechanisms againstwaterdeficit(Hura,Hura,&Grzesiak,2008;Ingram&Bartels,1996;Rivas-Ubachetal.,2014).

Our results thus indicated thatnevadensis experienced lessme-tabolomicvariationinsummer,whenwarmtemperaturesanddroughtaremore prominent, than iberica. Ourmetabolomic results suggestthus that the native populations of Scots pine are better adaptedthan the introducedpopulations to theenvironmental conditions inSierraNevada(Herrero&Zamora,2014),anticipatingamorestressfulabioticenvironmentfor ibericapopulationsasclimaticbeltsprogresspoleward.ThePPMispresent inthe ibericapopulations,butthe in-tensityandfrequencyofdefoliatonhasbeen increasingaltitudinallyand latitudinally during recent decades (Battisti etal., 2005; Hódar&Zamora,2004).Themaximumtemperaturesareabout4°Chigher,and the accumulated rainfall is 40% lower in Sierra Nevada thanNavacerrada,sotheconditionsaresignificantlywarmeranddrierthaninthenaturalrangeofiberica(Figure3;TableS3).Theincreasesinthesynthesisofsomemetabolitesnecessarytomaintainecophysiologicalfunctionunderdroughtenvironmentalconditions(Gasparetal.,2002)indicatethatthemoreextremesummerenvironmentalconditionsinSierraNevadacomparedtoNavacerradacontributedsubstantiallytothelargeshiftsinthefoliarmetabolomeofibericapines.Additionally,planted(iberica)populationsofP. sylvestrishavedeclinedsignificantlyinsoutheasternareasofSierraNevadawheresummersareevenmoreextreme than in our study site (Cerrillo, Varo, Lanjeri, & Clemente,2007;Guada,Camarero,Sánchez-Salguero,&Cerrillo,2016),support-ing our results.Ourmetabolomic data thus support our hypothesisthat theplantedpopulationsof iberica inSierraNevadaareoutsidetheirnaturalenvironmentalniche,eventhoughibericaandnevadensis belongtothesamespecies,andthereforeneedtocopewiththemoreextremesummerconditions,forcingtreestoproducelargershifts intheir metabolomes (Shao etal., 2007).We thus predict that iberica populationswill have anuncertain future aswarmer anddrier con-ditions incombinationwithsevererdefoliationcontinuetoprogresspoleward.

5  | CONCLUSIONS

The metabolomes of the two closely related subspecies of Scotspine tended to have similar local responses to herbivorous attack,

     |  11RIVAS- UBACH et Al.

suggesting that some metabolic pathways associated with folivorymayhavebeenevolutionarilyconserved.

The environmental conditions of summer are significantly moreextremeinSierraNevadathaninNavacerrada,withsignificanthighertemperaturesandmoreseveredroughtsinsummer.Bothpinesubspe-ciesanalyzedinthisstudycoexistinthesameenvironmentinSierraNevada,butthemetabolomicdifferencesbetweenthemweremorepronounced in summer and the concentrations ofmetabolites typi-callyassociatedwithdroughtstresswerehigheriniberica(introducedsubspecies)thannevadensis(nativesubspecies).

Themetabolomicresultsofbothpinesubspeciessuggestthatthedivergencebetweenthesummeribericaandnevadensismetabolomesrelativetowinteris,inpart,associatedwiththenaturaldistributionsofthesubspecies.

Alongerperiodoflocaladaptationlikelyprovidednevadensiswithametabolismthatisbetteradaptedtodroughtconditionsthanisthemetabolismofiberica,whichissubjectedtomoredroughtandhighertemperaturesinSierraNevadathanitexperiencesinitsnaturalhab-itat.Thesemoreextremeconditions for ibericamayaccount for thelargershiftsintheirmetabolomestomaintainphysiologicalhomeosta-sis.WeanticipateanuncertainfutureforibericapopulationsinSierraNevadawith thewarmer and drier conditions expected during theforthcomingdecades.

Eco-metabolomic techniques are potential tools to understandlong-timeecologicalprocessesratherthanonlybiochemicalprocesses.

ACKNOWLEDGMENTS

The authors thank Karim Senhadji, Laia Mateu-Castell, Sara Férez,Ander Achotegui-Castells, and Joan Llusia for their field and labo-ratory support. We also thank the anonymous reviewers for theirconstructivecommentsthathelpedtoimprovethemanuscript.ARUappreciatesthefinancialsupportoftheresearchfellowship(JAE)fromtheCSIC.This researchwassupportedby theSpanishGovernmentprojects CGL2016-79835-P and OAPN 022/2008 (PROPINOL),theCatalanGovernment project SGR2014-274, and theEuropeanResearchCouncil Synergy grant SyG-2013-610028 IMBALANCE-P.MOandOUweresupportedbythegrantprojectsM200871201(ASCR),CZ.1.07/2.3.00/20.0246(MSMT),LO1415andLM2015061.

CONFLICT OF INTEREST

Nonedeclared.

AUTHOR CONTRIBUTIONS

AR-Uwrotethemanuscript.AR-U,JS,JAH,AG,andJPdesignedandperformedtheresearch,collectedthesamples,andinterpretedthedata.AR-U,JS,JAH,andJG-Panalyzedthedata.MOandOUobtainedthemetabolomicfingerprintsofthesamples.AR-U,JS,JAH,JG-P,AG,LP-T,MO,OU,andJPreadandapprovedthefinalversionofthemanuscriptaswellascontributedaddressingproperlyallconcernsfromreviewers.

ORCID

Albert Rivas-Ubach http://orcid.org/0000-0003-1293-7127

REFERENCES

Apel,K.,&Hirt,H. (2004).Reactiveoxygen species:Metabolism,oxida-tivestress,andsignaltransduction.Annual Review of Plant Biology,55,373–399.

Battisti,A,Avcı,M,Avtzis,DN,BenJamaa,M,Berardi, L,Berretima,W,… Zamoum, M. (2015). Natural history of the processionary moths(Thaumetopea spp):New insigths in relation to climate change. InA.Roques (Ed.),Processionary moths and climate change: An update (pp.15–79).Dordrecht:SpringerNetherlands.

Battisti,A., Stastny,M.,Netherer, S., Robinet,C., Schopf,A., Roques,A.,&Larsson,S. (2005).Expansionofgeographicrange inthepinepro-cessionarymothcausedby increasedwinter temperatures.Ecological Applications,15,2084–2096.

Bi,J.L.,&Felton,G.W.(1995).Foliaroxidativestressandinsectherbivory:Primarycompounds,secondarymetabolites,andreactiveoxygenspe-ciesascomponentsofinducedresistance.Journal of Chemical Ecology,21,1511–1530.

Blanca, G., Cueto,M.,Martínez-Lirola,M. J., &Molero-Mesa, J. (1998).ThreatenedvascularfloraofSierraNevada(SouthernSpain).Biological Conservation,85,269–285.

Boratynski,A. (1991). Range of natural distribution. InM.Giertich&C.Mátyás(Eds.),Genetics of Scots pine(pp.19–30).Budapest:AkadémiaiKiadó.

Carrillo-Gavilán, A., Moreira, X., Zas, R., Gonzalez-Voyer, A., Vila, M., &Sampedro, L. (2015). Phylogenetic and biogeographical patternsin defensive strategies and quantitative allocation to chemical de-fencesinPalaearcticandNearcticpinetrees(RLadle,Ed.).Journal of Biogeography,42,684–693.

Castro,J.,Gómez,J.M.,García,D.,Zamora,R.,&Hódar,J.A.(1999).Seedpredationanddispersal inrelictScotspineforests insouthernSpain.Plant Ecology,145,115–123.

Ceballos,L.,&Ruizde laTorre,J. (1971).Árboles y arbustos de la España peninsular(pp.1–512).Madrid:InstitutoForestaldeInvestigacionesyExperiencias.

Cerrillo, R. M. N., Varo, M. A., Lanjeri, S., & Clemente, R. H. (2007).Cartografíadedefoliaciónenlospinaresdepinosilvestre(Pinus sylves-trisL.)ypinosalgareño(Pinus nigraArnold.)enlaSierradelosFilabres.Revista Ecosistemas,16,163–171.

Chen, I.-C.,Hill,J.K.,Ohlemüller,R.,Roy,D.B.,&Thomas,C.D. (2011).Rapid range shifts of species associatedwith high levels of climatewarming.Science,333,1024–1026.

Edwards,E.J.,Still,C.J.,&Donoghue,M.J.(2007).Therelevanceofphy-logeny to studiesof global change.Trends in Ecology & Evolution,22,243–249.

Fiehn,O.(2002).Metabolomics—Thelinkbetweengenotypesandpheno-types.Plant Molecular Biology,48,155–171.

Fox,J.,&Weisberg,S.(2011).An {R} companion to applied regression (2nded.).Rpackage.ThousandsOaks,CA.Sage.

Gaspar, T., Franck, T., Bisbis, B., Kevers, C., Jouve, L., Hausman, J.-F., &Dommes,J.(2002).Conceptsinplantstressphysiology.Applicationtoplanttissuecultures.Plant Growth Regulation,37,263–285.

Gausen,H.,Heywood,V.H.,&Chater,A.O.(1964).Pinus.InT.G.Tutin,V.H.Heywood,N.A.Burges,D.H.Valentine,S.M.Walters,&D.A.Webb(Eds.),Flora Europaea,Vol.I(pp.32–35).Cambridge:CambridgeUniversityPress.

Gonzalez,P.,Neilson,R.P.,Lenihan,J.M.,&Drapek,R.J. (2010).Globalpatternsinthevulnerabilityofecosystemstovegetationshiftsduetoclimatechange.Global Ecology and Biogeography,19,755–768.

12  |     RIVAS- UBACH et Al.

González-Orozco,C.E.,Pollock,L.J.,Thornhill,A.H.,Mishler,B.D.,Knerr,N.,Laffan,S.W.,…Cruber,B.(2016).Phylogeneticapproachesrevealbiodiversity threats under climate change.Nature Climate Change,6,1110–1114.

Guada, G., Camarero, J. J., Sánchez-Salguero, R., & Cerrillo, R. M. N.(2016).Limitedgrowthrecoveryafterdrought-inducedforestdiebackinverydefoliatedtreesoftwopineSpecies.Frontiers in Plant Science,7,418.

Hernández, I.,Alegre, L.,VanBreusegem, F.,&Munné-Bosch, S. (2009).Howrelevantareflavonoidsasantioxidants inplants?Trends in Plant Science,14,125–132.

Herrero, A., & Zamora, R. (2014). Plant responses to extreme climaticevents:Afieldtestofresiliencecapacityatthesouthernrangeedge.PLoS One,9,e87842.

Hódar, J.A., Castro, J., & Zamora, R. (2003). Pine processionary caterpillarThaumetopoea pityocampaasanewthreatforrelictMediterraneanScotspineforestsunderclimaticwarming.Biological Conservation,110,123–129.

Hódar, J. A., & Zamora, R. (2004). Herbivory and climatic warming: AMediterraneanoutbreakingcaterpillarattacksarelict,borealpinespe-cies.Biodiversity and Conservation,13,493–500.

Hódar,J.A.,Zamora,R.,&Cayuela,L.(2012).Climatechangeandthein-cidenceofaforestpestinMediterraneanecosystems:CantheNorthAtlantic Oscillation be used as a predictor? Climatic Change, 113,699–711.

Hura,T.,Hura,K.,&Grzesiak,S. (2008).Contentsof totalphenolicsandferulicacid,andPALactivityduringwaterpotentialchangesinLeavesofmaizesingle-crosshybridsofdifferentdroughttolerance.Journal of Agronomy and Crop Science,194,104–112.

Husson,F.,&Josse,J.(2015).missMDA:Handlingmissingvalueswithmul-tivariatedataanalysis.Rpackageversion1.9.

Husson,F.,Josse,J., Le,S.,&Mazet,J. (2016).FactoMineR:Multivariateexploratorydataanalysisanddatamining.Rpackageversion1.32.

Ingram,J.,&Bartels,D.(1996).Themolecularbasisofdehydrationtoler-anceinplants.Annual Review of Plant Biology,47,377–403.

IPCC. (2002). Climate change and biodiversity. IPCCTechnical PaperV.Geneva,Switzerland:IPCCWorkingGroupIITechnicalSupportUnit.

Jactel,H.,Barbaro,L.,Battisti,A.,Bosc,A.,Branco,M.,Brockerhoff,E.,…Schlyter, F. (2015). Insect—Tree interactions in Thaumetopoea pityo-campa. InA.Roques (Ed.)Processionary moths and climate change: An update(pp.265–310).NewYork:Springer.

Josse, J., & Husson, F. (2013). Handling missing values in exploratorymultivariate data analysismethods. Journal de la Société Française de Statistique,153(2),79–99.

KhorasaniEsmaeili,A.,MatTaha,R.,Mohajer,S.,&Banisalam,B. (2015).Antioxidantactivityandtotalphenolicandflavonoidcontentofvarioussolventextracts from invivoand invitrogrownTrifolium pratense L.(RedClover).BioMed Research International,2015,643285.

Kim,H.K.,Choi,Y.H.,&Verpoorte,R. (2010).NMR-basedmetabolomicanalysisofplants.Nature Protocols,5,536–549.

Kim,J.H.,Lee,B.C.,Kim,J.H.,Sim,G.S.,Lee,D.H.,Lee,K.E.,…Pyo,H.B.(2005).TheisolationandantioxidativeeffectsofvitexinfromAcer palmatum. Archives of Pharmacal Research,28,195–202.

t’Kindt, R., De Veylder, L., Storme, M., Deforce, D., & Van Bocxlaer, J.(2008).LC-MSmetabolicprofilingofArabidopsis thalianaplantleavesand cell cultures:Optimization of pre-LC-MS procedure parameters.Journal of Chromatography. B, Analytical Technologies in the Biomedical and Life Sciences 871,37–43.

Kuntner,M.,Năpăruş,M.,Li,D.,&Coddington,J.A.(2014).Phylogenypre-dictsfuturehabitatshiftsduetoclimatechange.PLoS One,9,e98907.

Lee,D.Y.,&Fiehn,O. (2013).Metabolomic responseofChlamydomonas reinhardtiitotheinhibitionoftargetofrapamycin(TOR)byrapamycin.Journal of Microbiology and Biotechnology,23,923–931.

Li, S.,Daudin, J. J., Piou,D., Robinet, C., & Jactel,H. (2015). PeriodicityandsynchronyofpineprocessionarymothoutbreaksinFrance.Forest Ecology and Management,354,309–317.

Mäkinen,H.,&Hynynen,J.(2014).WooddensityandtracheidpropertiesofScotspine:Responsestorepeatedfertilizationandtimingofthefirstcommercialthinning.Forestry,87,437–447.

Mari,A., Lyon,D.,Fragner,L.,Montoro,P.,Piacente,S.,Wienkoop,S.,…Weckwerth,W. (2013). Phytochemical composition of Potentilla an-serinaL.analyzedbyanintegrativeGC-MSandLC-MSmetabolomicsplatform.Metabolomics,9,599–607.

Meier, E. S., Lischke, H., Schmatz, D. R., & Zimmermann, N. E. (2012).Climate, competition and connectivity affect future migration andranges of European trees. Global Ecology and Biogeography, 21,164–178.

deMendiburu,F. (2015).agricolae:StatisticalProceduresforAgriculturalResearch.Rpackageversion1.2-3.

Mierziak,J.,Wojtasik,W.,Kostyn,K.,Czuj,T.,Szopa,J.,&Kulma,A.(2014).Crossbreedingoftransgenicflaxplantsoverproducingflavonoidsandglucosyltransferase results in progenywith improved antifungal andantioxidativeproperties.Molecular Breeding,34,1917–1932.

Ness,J.H.(2003).ContrastingexoticSolenopsis invictaandnativeForelius pruinosusantsasmutualistswithCatalpa bignonioides,anativeplant.Ecological Entomology,28,247–251.

Ninyerola,M.,PonsFernandez,X.,&RoureiNolla,J.M.(2005).Atlascli-máticodigital de laPenínsula Ibérica:metodologíay aplicaciones enbioclimatologíaygeobotánica.pp.44.Bellaterra:UniversitatAutonomadeBarcelona.

Oksanen, J., Guillaume-Blanchet, F., Kindt, R., Legendre, P., Minchin, P.,O’Hara,R.,…Wagner,H.(2013).vegan:CommunityEcologyPackage.Rpackageversion2.3.2.

Parmesan,C.,&Yohe,G.(2003).Agloballycoherentfingerprintofclimatechangeimpactsacrossnaturalsystems.Nature,421,37–42.

Peñuelas,J.,&Sardans,J.(2009).Ecologicalmetabolomics.Chemistry and Ecology,25,305–309.

Pluskal,T.,Castillo,S.,Villar-Briones,A.,&Orešič,M. (2010).MZmine2:Modular framework for processing, visualizing, and analyzing massspectrometry-based molecular profile data. BMC Bioinformatics, 11,395.

RCoreTeam.(2013).R:Alanguageandenvironmentforstatisticalcomput-ing.RFoundationforStatisticalComputing,Vienna:Austria.

Raman,S.T.,Ganeshan,A.K.P.G.,Chen,C.,Jin,C.,Li,S.-H.,Chen,H.-J.,&Gui,Z. (2016). Invitroandinvivoantioxidantactivityofflavonoidextractedfrommulberryfruit(Morus albaL.).Pharmacognosy Magazine,12,128–133.

Ravensdale, M., Rocheleau, H., Wang, L., Nasmith, C., Ouellet, T., &Subramaniam, R. (2014). Components of priming-induced resistancetoFusariumheadblightinwheatrevealedbytwodistinctmutantsofFusarium graminearum. Molecular Plant Pathology,15(948),956.

Rice-Evans,C.A.,Miller,N.J.,&Paganga,G.(1996).Structure-antioxidantactivity relationships of flavonoids and phenolic acids. Free Radical Biology and Medicine,20,933–956.

Riedl, J., Kluender, C., Sans-Piché, F., Heilmeier, H., Altenburger, R., &Schmitt-Jansen,M.(2012).Spatialandtemporalvariationinmetabolicfingerprints of field-growingMyriophyllum spicatum. Aquatic Botany,102,34–43.

Rivas-Ubach,A.,Barbeta,A.,Sardans,J.,Guenther,A.,Ogaya,R.,Oravec,M., … Peñuelas, J. (2016).Topsoil depth substantially influences theresponsestodroughtofthefoliarmetabolomesofMediterraneanfor-ests.Perspectives in Plant Ecology, Evolution and Systematics,21,41–54.

Rivas-Ubach, A., Gargallo-Garriga, A., Sardans, J., Oravec, M., Mateu-Castell,L.,Pérez-Trujillo,M.,…Peñuelas,J.(2014).Droughtenhancesfolivory by shifting foliar metabolomes in Quercus ilex trees. New Phytologist,202,874–885.

Rivas-Ubach,A.,Hódar,J.A.,Sardans,J.,Kyle,J.,Kim,Y.-M.,Oravec,M.,… Peñuelas, J. (2016). Are the metabolomic responses to folivoryof closely related plant species linked to macroevolutionary andplant–folivore coevolutionary processes? Ecology and Evolution, 6,4372–4386.

     |  13RIVAS- UBACH et Al.

Rivas-Ubach,A.,Pérez-Trujillo,M.,Sardans,J.,Gargallo-Garriga,A.,Parella,T., & Peñuelas, J. (2013). Ecometabolomics: Optimized NMR-basedmethod.Methods in Ecology and Evolution,4,464–473.

Rivas-Ubach,A.,Sardans,J.,Hódar,J.A.,Garcia-Porta,J.,Guenther,A.,Oravec,M., … Peñuelas, J. (2016). Similar local but different sys-temicmetabolomicresponsesofcloselyrelatedpinesubspeciestofolivorybycaterpillarsoftheprocessionarymoth.Plant Biology,18,484–494.

Rivas-Ubach,A.,Sardans,J.,Pérez-Trujillo,M.,Estiarte,M.,&Peñuelas,J.(2012).Strongrelationshipbetweenelementalstoichiometryandme-tabolomeinplants.Proceedings of the National Academy of Sciences of the United States of America,109,4181–4186.

Robledo-Arnuncio,J.J.,Navascués,M.,González-Martínez,S.C.,&Gil,L. (2009). Estimating gametic introgression rates in a risk assess-ment context:A case studywith Scots pine relicts.Heredity,103,385–393.

Ruuhola, T., & Yang, S. (2006). Wound-induced oxidative responses inmountainbirchleaves.Annals of Botany,97,29–37.

Sánchez,D.H.,Schwabe,F.,Erban,A.,Udvardi,M.K.,&Kopka,J.(2012).Comparativemetabolomicsofdroughtacclimationinmodelandforagelegumes.Plant, Cell & Environment,35,136–149.

Sardans, J., Peñuelas, J., & Rivas-Ubach, A. (2011). Ecological metabo-lomics: Overview of current developments and future challenges.Chemoecology,21,191–225.

Schlyter, F., Smitt, O., Sjödin, K., Högberg, H.-E., & Löfqvist, J. (2004).Carvone and lessvolatile analogues as repellent and deterrent anti-feedants against the pineweevil,Hylobius abietis. Journal of Applied Entomology,128,610–619.

Shao,H.-B.,Guo,Q.-J.,Chu,L.-Y.,Zhao,X.-N.,Su,Z.-L.,Hu,Y.-C.,&Cheng,J.-F.(2007).Understandingmolecularmechanismofhigherplantplasticityunderabioticstress.Colloids and Surfaces B: Biointerfaces,54,37–45.

Singh,B.,Sahu,P.M.,&Sharma,R.A.(2017).FlavonoidsfromHeliotropium subulatumexudateandtheirevaluationforantioxidant,antineoplasticandcytotoxicactivitiesII.Cytotechnology,69,103–115.

Szabados,L.,&Savouré,A. (2010).Proline:Amultifunctionalaminoacid.Trends in Plant Science,15,89–97.

Tamburini, G., Marini, L., Hellrigl, K., Salvadori, C., & Battisti, A. (2013).Effects of climate and density-dependent factors on population dy-namicsofthepineprocessionarymothintheSouthernAlps.Climatic Change,121,701–712.

SUPPORTING INFORMATION

Additional Supporting Information may be found online in the supportinginformationtabforthisarticle.

How to cite this article:Rivas-UbachA,SardansJ,HódarJA,etal.Closeanddistant:ContrastingthemetabolismoftwocloselyrelatedsubspeciesofScotspineundertheeffectsoffolivoryandsummerdrought.Ecol Evol. 2017;00:1–13. https://doi.org/10.1002/ece3.3343