Identification of candidate genes for phenolics accumulation in tomato fruit

(1)

ContentslistsavailableatSciVerseScienceDirect

Plant

Science

j our na l h o me p ag e :w w w . e l s e v i e r . c o m / l o c a t e / p l a n t s c i

Identification

of

candidate

genes

for

phenolics

accumulation

in

tomato

fruit

Antonio

Di

Matteo

a,1

_{, Valentino}

_Ruggieri

a,1

_{, Adriana}

_Sacco

a

_{, Maria}

_Manuela

_Rigano

a

_,

Filomena

Carriero

b

_,

_Anthony

_Bolger

c,d

_,

_Alisdair

_R.

_Fernie

c

_,

_Luigi

_Frusciante

a

_,

_Amalia

_Barone

a,∗

a_Department_of_Agricultural_Sciences,_University_of_Naples_Federico_II,_Via_Università_100,₈₀₀₅₅_Portici,_Italy

b_Metapontum_Agrobios,_SS_Jonica₁₀₆_Km_448,₂_{– 75010}_Metaponto,_MT,_Italy

c_{Max-Planck-Institute}_of_Molecular_Plant_Physiology,_Am_Mühlenberg_1,₁₄₄₇₆_Potsdam,_Golm,_Germany

d_RWTH_University_Aachen,_Worringerweg_1,₅₂₀₆₂_Aachen,_Germany

a

r

t

i

c

l

e

i

n

f

o

Articlehistory:

Received16July2012

Receivedinrevisedform31January2013

Accepted3February2013

Available online 13 February 2013 Keywords: Solanumlycopersicum Flavonoids Introgressionlines Microarray TILLING

Ethyleneresponsivefactor1

a

b

s

t

r

a

c

t

Phenolicsareantioxidantspresentintomatofruitthatconferhealthybenefitsandexhibitcrucialroles forplantmetabolismandresponsetoenvironmentalstimuli.Anapproachbasedontwogenomics plat-formswasundertakentoidentifycandidategenesassociatedtohigherphenolicscontentintomatofruit. AcomparativetranscriptomicanalysisbetweentheS.pennelliiIntrogressionLine7-3,whichproduced anaveragehigherleveloffruitphenolics,andthecultivatedvarietyM82,revealedthattheirdifferences areattributedtogenesinvolvedinphenolicsaccumulationintothevacuole.Theup-regulationofgenes codingforoneMATE-transporter,onevacuolarsortingproteinandthreeGSTssupportedthis hypoth-esis.Theobservedbalancingeffectbetweentwoethyleneresponsivefactors(ERF1andERF4)wasalso hypothesizedtodrivethetranscriptionalregulationofthesetransportgenes.Inordertoconfirmsuch modelaTILLINGplatformwasexplored.AmutantwasisolatedharbouringapointmutationintheERF1 cdsthataffectstheproteinsequenceanditsexpectedfunction.Fruitsofthemutantexhibiteda signifi-cantreducedlevelofphenolicsthanthecontrolvariety.Changesintheexpressionofgenesinvolvedin sequestrationofphenolicsinvacuolealsosupportedthehypothesizedkey-roleofERF1inorchestrating thesegenes.

1. Introduction

In the last few years, medical research has highlighted the importance of thecompositional qualityof vegetable crops for human health [1–3]. Health benefits conferred by vegetables havebeenmainlyattributedtothepresenceofhealth-promoting phytochemicalswithantioxidantproperties.Givenitsfruit chem-icalcompositionandextensiveconsumptionworldwide,tomato (Solanumlycopersicum)isanimportantsourceofantioxidantsfor humannutrition,suchasphenolics,carotenoids,tocopherols,and

Abbreviations: ERF1,ethyleneresponsivefactor1;ERF4,ethyleneresponsive

factor4;FW,freshweight;GAE,gallicacidequivalents;GSH,reducedglutathione;

GSSG,oxidizedglutathione;GSH+GSSG,totalglutathione;GST,glutathione

S-transferase;HPRG,histidineprolinerichglycoprotein;IL,introgressionline;MATE,

multidrugandtoxiccompoundextrusion;PR,pathogenrelated;QE,quercetin

equivalents;TC,tentativeconsensus;TILLING,targetinginducedlocallesionsin

genomes;VSP,vacuolarsortingprotein.

∗ Correspondingauthor.Tel.:+390812539491;fax:+390812539486.

E-mailaddresses:adimatte@unina.it(A.DiMatteo),valentino.ruggieri@unina.it

(V. Ruggieri), adriana.sacco@unina.it (A. Sacco), mrigano@unina.it (M.M.

Rig-ano),fcarriero@agrobios.it(F.Carriero),Bolger@mpimp-golm.mpg.de(A.Bolger),

Fernie@mpimp-golm.mpg.de (A.R. Fernie), fruscian@unina.it (L. Frusciante),

ambarone@unina.it(A.Barone).

1_These_authors_contributed_equally_to_this_work.

ascorbicacid.Regularconsumptionoftomatofruitshasbeen asso-ciated witha decreased incidence of chronic degenerative and proliferative diseasessuchassometypes ofcancerand cardio-vasculardiseases[4,5].Inparticular,amongphenolics,flavonoids preventheartdiseases,decreasebloodviscosity,reduce inflamma-toryresponsesandallergicreactions;inadditiontheyhavebeen reportedtohaveanti-tuberculosis,antiviralandantimalarial activ-ities[6].

Phenolicsarealargegroupofmoleculesthatalsohave impor-tant functionsin plants[7]. Amongthese,stilbenes,coumarins, andisoflavonoidsareimplicatedindefencemechanismsandare comprehensivelydefinedphytoalexins.FlavonoidsareUV protec-tants andanthocyanins arepigmentsresponsiblefor coloursof flower, fruit,vegetables,while salicylicacid(SA) isa signalling molecule involved in plant–microbe interactions and monolig-nolsarethebuildingunitsoflignin,thesecondmostprevalent biopolymer on earth after cellulose. Phenolicscontent in plant tissuesvarieswithspecies,varieties,organs,soilpropertiesand climaticconditions.Thesesecondarymetabolitesaccumulate, gen-erally,inallplantorgans(roots,stems,leaves,flowersandfruits) althoughpreferentiallyintheaerialorgans.Inthefruit,theouter tissuesarericherinflavonoidsthanothertissues,while chloro-genicacidandcoumarinesaremoreevenlydistributed.Intomato fruit,flavonoidsaretypicallypresentintheepidermaltissuesat

(2)

differentlevelsdependingonthedevelopmentalstage[8].In addi-tiontotheflavonoidnaringeninchalconethataccumulatesupto 1%dryweightofthetomatofruitcuticle,variousotherflavonoids accumulateintomatofruitsuchasrutin(quercetin-3-rutinoside), kaempferol-3-O-rutinosideandaquercetin-trisaccharide[9,10].

Processes controlling the level of phenolics in plant tissues includebiosynthesisanddegradation/utilizationofintermediates aswellasvacuolesequestration.Intomatofruitsseveralauthors haveextensivelyreportedthewide panoramaofphenolics pro-duction, withparticular attentionbeing paid to the flavonoids biosynthesis and storage [11–14].Although genes that operate in thebiosynthesis of flavonoidsare already known,additional insightsarerequired toelucidatetheregulativemechanismsof theirproductionandaccumulationintomatofruits.

Atsub-cellularlevel,thetwomainsitesofaccumulationof phe-noliccompoundsarethecellwall,wheretheligninisdeposited,and thevacuole,wheredifferentclassesofphenolicsarestored, par-ticularlyflavonoids.Sincethiscompartmentalizationisimportant forlocalmaintenanceofhighmetaboliteconcentrations, mecha-nismsforflavonoidsandanthocyaninsstoragehavebeenwidely investigatedandrecentlyreviewed[15–18].Briefly,among oth-ers,GrotewoldandDavies[19]proposedtwomodelstodescribe thevacuolar compartmentalization of flavonoids via intracellu-lar movement from cytosol, the so-called vesicular (VT) and ligandintransport(LT)models.Inthefirst,thetransportis medi-atedbyvesicletraffickinginvolvingproteincomponents,suchas vacuolarsorting andcargo proteins,whereas thesecondmodel relies onmembrane transporters,suchas MATEand MRP-type ABC transporters.The lattermediates sequestration tothe vac-uolesof glutathioneconjugates,whoseconjugation iscatalyzed byglutathione-S-transferases(GST),hencethrougha glutathione-dependentmechanism[20].IntheframeoftheLTmodel,GSTscan alsooperateascarrierproteinsinthetransportofGST-flavonoid complexesthroughaglutathione-independentmechanism[21].

Multiplelevelsoftranscriptionalandpost-transcriptional reg-ulationhavebeenimplicatedincontrollingthebiosynthesisand accumulationofphenolics.Inparticular,ithasbeendemonstrated that ethylene and other hormones can modulate the intensity and the direction of regulative effects on metabolism and the finallevelofphenolicsinplanttissues.Severalauthorsreported that theproduction of phenolic compounds canbe affectedby exogenoustreatmentswithethylene[22–24]and/orbybioticand abioticstresses,andthatthiseffectcouldbemediatedbyethylene responsiveelements(EREs)andMYBfactors[24].Indeed,byusing variousArabidopsismutants,ithasbeenrecentlydemonstrated that auxin and ethyleneregulate flavonol biosynthesis through MYB12-mediatedsignallingandthatethylenemodulatesflavonoid accumulationintheroots.

The accumulationof phenolics in tomato fruit is a complex traitcontrolledbyvariousQTL[25,26]and stronglyaffectedby theenvironment. In tomato the dissectionof complex traits is enabledbygeneticresourcesthathavebeenavailableforalong time,suchasintrogressionlines(ILs).Thesearehomozygouslines withsinglechromosomesegmentsubstitutionsfromonewild rel-ative[27].ThecombineduseofILsand transcriptionalprofiling [28] mightfacilitate therapid identificationof candidategenes involvedin theaccumulationof fruitphenolics.Theaimofthis workwastoidentifyregulativemechanismsandgenes control-lingtheaccumulationofphenolicsintomatofruitintheS.pennellii introgressionlineIL7-3,which harboursapositive QTLfor phe-nolics content. For this purpose, a comparative transcriptomic approachwascarriedoutonmaturefruitfromthisintrogression lineanditscultivatedparentalline(M82).In particular,among candidate genes we identified an ethylene responsive factor 1 (ERF1)genethatwasup-regulatedintheS.pennelliiIL7-3,which mightdirectlyorindirectlyregulateamoreefficienttransportof

phenolicstothevacuole.Characterizationofatomatoerf1mutant lineisolatedthroughTILLING(targetinginducedlocallesionsin genomes)technology[29,30]allowedustosuggestaroleforthe ERF1geneincoordinatingtheaccumulationofphenolicsintomato fruits.

2. Materialsandmethods 2.1. Plantmaterial

Twotomatoaccessions(LA4066andLA3475)wereusedinthis researchandseedswereprovidedbytheTomatoGeneticsResource Centre,UniversityofCalifornia(Davis,USA).TheaccessionLA4066 identifiestheintrogressionline(IL)7-3whichharboursa32cM singlehomozygouschromosomesegmentfromthewildspecies S.pennellii(LA0716)inthegenomicbackgroundoftheprocessing tomatovarietyS.lycopersicumcv.M82(LA3475).Plantsweregrown in greenhouse for four consecutive years. After seed germina-tion,sixseedlingsperaccessionweretransplantedinto20cmpots containingamixtureofmediumsandysoilandcompostand trans-ferred in a cold greenhouse of the Department of Agricultural SciencesattheUniversityofNaples(Portici,Italy).Plantswere ran-domlydistributed,fertilizedandirrigateddailyuntiltheyreached maturity.Fruitswereharvestedwhenabout75%ofthemreached thered-ripestage.Withina singletrial,threefruit samplesper each genotypewereobtainedfromindividualplantsbypooling fullyred-ripefruitscollectedfromthesameplant.Sampleswere obtainedfromwholefruitsdiscardingtheseeds,jellyparenchyma, columellaandplacentatissuesandthenwerefrozenbyliquid nitro-gen,grindedandstoredat−80◦_C.

2.2. Fruitmeasurements

Totalphenolicswereassayedusingamodifiedprocedureofthe Folin–Ciocalteu’stest[31].Inbrief,250mgoffrozenground tis-suewerehomogenizedinamortarwithpestleandextractedusing 1mlof60%methanol.Samplesweretransferredtoa1.5mltube andleftonicefor3mininthedark.Crudeextractswere trans-ferredina15mltubeandvolumewasincreasedto5mladding 60%methanol.Thesampleswerecentrifugedat3000×gfor5min; then,62.5!lofthesupernatant,62.5!lofFolin–Ciocalteu’sreagent (Sigma)and250!lofdeionisedwaterweremixedandincubated for6min;625!lof7.5%sodiumcarbonateand500!lofdeionised waterwereaddedtothesamplesandincubatedfor90minatroom temperatureinthedark.Absorbancewasmeasuredat760nm.The concentrationoftotalphenolicswasexpressedintermsofmgof gallicacidequivalents(GAE)per100goffreshweight(FW).

TheflavonoidcontentwasmeasuredasreportedbyMarinova etal.[32].Inparticular,5mlof80%met/H2Owereaddedto0.5g

ofsample.100!lofdistilledwaterand30!lof5%NaNO2 were

addedto100!lofextract.30!l10%AlCl3wereadded5minlater.

After6min,200!lof 1MNaOHwereaddedand thetotal was madeupto1mlwithdistilledwater.Absorbancewasmeasuredat 510nm.Totalflavonoidcontentwasexpressedasmgofquercetin equivalents(QE)per100gFW.

Reduced(GSH)andoxidized(GSSG)glutathionewereassessed usingGlutathioneAssayKit(BioVision–Cataloguenumber K264-100) following the procedure suggested by the manufacturer. Fluorescence(Ex/Em340/420nm)ofthesampleswasreadwitha VICTORTM_X_Multilabel_Plate_Reader_{(PerkinElmer).}_{Concentrations}

wereexpressedasmgpergofFW.

Phenotypicdatawerestatisticallyprocessedbyusingthe Sta-tisticalPackageforSocialSciences,version18(SPSSIncChicago, Illinois). UnivariateANOVAwasappliedtodeterminetheeffect ofgenotypeandtrialasfixedfactors.Theeffectofthegenotype

(3)

Table1

PrimerpairsusedforqPCRvalidationofgenesinvolvedintheaccumulationofphenolicsthataredifferentiallyexpressedinIL7-3andM82.

TIGRID Annotation Primersequence

TC170110 GlutathioneS-transferase-likeprotein fw 5′_{-AGCAACGGGAGACGAACATC-3}′

rv 5′_{-ATCTCCAAGTGCTCCCTCCA-3}′

TC170138 Ethylene-responsivetranscriptionfactor4 fw 5′_{-GTCATCCAGCGGAGAAACGG-3}′

rv 5′_{-GCAGGTGGAAGTCGTCGCCG-3}′

TC172097 Vacuolarsortingprotein fw 5′_{-TTGTGGCGACCTTATGATCC-3}′

rv 5′_{-GGCTTTATCTCCCAGGATTCAG-3}′

TC172606 GlutathioneS-transferase fw 5′_{-TCCTTATGAAAGAGCCCATGCT-3}′

rv 5′_{-TTGTTGTCCATACCAATCTCCCT-3}′

TC172943 Hydroxyproline-richglycoprotein fw 5′_{-TACCGGAAAAAGTGGTGGTTG-3}′

rv 5′_{-ACTGGTTTAGGAGGCTGAGGC-3}′

TC174512 Ethylene-responsivefactor1 fw 5′_{-CGCGTAATGGAATTAGGGTTTG-3}′

rv 5′_{-CATTGATAATGCGGCTTGATCA-3}′

TC176201 GlutathioneS-transferase fw 5′_{-TCTGCAGGTTCAACTGGAAGAA-3}′

rv 5′_{-GGAGGAGCGAAGGGAAAGAA-3}′

TC177287 O-acyltransferaseWSD1 fw 5′_{-TCCCATGCTGAGGCAACTTC-3}′

rv 5′_{-GGGCAATTCCATCTCCAAGAG-3}′

TC177388 Osmotin-likeprotein fw 5′_{-ACGTCTCAATCGAGGCCAAA-3}′

rv 5′_{-ACGTGCCATCTTAGTTCCCCT-3}′

TC178155 Chymotrypsininhibitor-2 fw 5′_{-TTTGTCCAGGTGTGACAAAGGA-3}′

rv 5′_{-GCAAACTTAGCTGGTGTTCCAAG-3}′

TC178409 MultidrugresistanceproteinmdtK fw 5′_{-GCTTTGCTTGGATATGTGGC-3}′

rv 5′_{-CTCTTCATCCCAATCTGTTTTCC-3}′

TC179186 Pathogenesis-relatedfamilyprotein fw 5′_{-CGTGGGAAGTGATCAACGTG-3}′

rv 5′_{-AAAGAAACCCCAATGCCTGA-3}′

TC185053 GlutathioneS-transferase/peroxidase fw 5′_{-GCAAAGGCTGAGTGTCCAAAG-3}′

rv 5′_{-GACTTGGCCACACTGTCCCT-3}′

TC186230 Serine/threonineproteinkinase fw 5′_{-AACCGAGGCATGTGCTTCTG-3}′

rv 5′_{-TGAAACATGACGATGCTTTTTCC-3}′

wasconfirmedwithineachtrialbyStudent’st-testcoupledwith

bootstrapusing1000samplesand˛=0.05.

2.3. Microarrayhybridization

Duringtwoconsecutivetrials,threefruitsamplespereach

geno-typehavebeenchosenforfurthertranscriptomicanalysis.Total

RNAextractionwasobtainedusingtheCTAB

(hexadecyltrimethy-lammoniumbromide)method[33].TotalRNAwasquantifiedusing

aNanoDropND-1000Spectrophotometer.RNAswithabsorbance ratioA230/260 andA260/A280 higherthan 2.0weretested for integrityusingtheAgilent2100Bioanalyzer(Agilent).

Transcriptomic analysis was performed ona 90K TomatAr-ray1.0 microarray synthesized using theCombiMatrix platform at the Plant Functional Genomics Center of the University of Verona (http://ddlab.sci.univr.it/FunctionalGenomics/) as previ-ouslydescribed[34].Expressiondatawerenormalizedusingthe SPSSsoftwarepackageandprobeswithsignificantchange(P-value <0.01)intheirhybridizationsignalswereidentifiedbyrunninga two-factorANOVAmoduleincludedintheTIGRMultiple Experi-mentViewerSoftwarev4.0(http://www.tm4.org/).Co-regulations wereinvestigatedbyPearsonmetric.Functionalannotationof dif-ferential expressed geneswas performedby using theITAG2.3 database [35] available at the Solgenomics network website (http://solgenomics.net).Detailedinformation aboutmicroarray dataisavailableinGeneExpression Omnibus(http://www.ncbi. nlm.nih.gov/geo/)undertheseriesaccessionGSE37568.

2.4. QuantitativeReal-TimePCR(RT-qPCR)

The relative expression of candidate mRNAs for controlling accumulation of phenolics in IL7-3fruit was validated using a 7900HTFast Real-TimePCR System(AppliedBiosystems). Total RNAwasretrotranscribedbytranscriptorHigh-FidelitycDNA Syn-thesiskitaccordingtomanufacturer’sprotocol.Real-TimeRT-qPCR assaywasdonein12.5!lreactionvolumesthroughSYBRGreen PCRMasterMix(AppliedBiosystems)withthreereplicationsper eachreaction.PlateassemblingwasautomatedbyusingaFreedom

EVO1.50(TECAN)liquidhandlerstation.Quantificationwas per-formedaccordingtothe!!Ctmethod[36].Primerswerevalidated usingastandardcurveoveradilutionrange1–10−3_(R2_>_0.98;_slope

closeto−3.32).AlistofprimersequencesisprovidedinTable1. 2.5. Tentativeconsensusmapping

Sequences of the 21,550 tomato tentative consensus sequences (TCs) retrieved from TIGR database Gene Index Release 11.0 (June 21, 2006) were mapped on the tomato chromosomes. Advanced mode of the SGN BlastN tools (http://solgenomics.net/tools/blast/index.pl), were used toalign TCswithpseudomoleculesoftheTomatoWGSChromosomes(SL 2.40) database.OnlyTCsproducing alignmentwithexpectation values <10−10 _were _considered._Also, _TCs_mapping _to _the_wild

S. pennellii introgression 7-3 were identified by anchoring the sortedlistofTCsmappingtothepseudochromosome7tomarkers thatdefinethe7-3segmentintheRFLPTomato-Expen2000v.52 markers,thatareSGNMarkerT1719BandSGNMarkerTG20. 2.6. ScreeningoftomatoTILLINGmutants

The tomato TILLING platform at Metapontum Agrobios[37] wasusedforscreeningapopulationofM3mutagenizedRed Set-ter tomato families in order todetectmutations in thecoding sequenceoftheERF1gene.Ampliconswereamplifiedby nested-PCRaccordingtotheexperimentalconditionspreviouslydescribed [38].Theexternalprimerpair(fw:5′_{-TACCCACCCAAGAGTCAT-3}′

and rv:5′_{-CCCTACACCCAAACAAAC-3}′₎_produced_a ₁₄₂₈_bp

frag-ment.Internalprimers(fw:5′_{-CACTATTTCTCTTTCCTCATCC-3}′_and

rv5′_{-CAACTCATCAAAAGCTCTTC-3}′₎_were₅′ _labelled _with_IRDye

700andIRDye800dye(LI-COR®_,_Lincoln,_NE,_USA)_respectively

and produced afragmentof 994bp.Unlabeledinternal primers werealsousedforprimingsequencingreactions.Inducedpoint mutations were identified by means of the mismatch specific endonucleaseENDO1(SerialGenetics)intheexperimental con-ditions suggested by the supplier. Digested and labelled DNA fragmentswereseparatedontheLI-COR4300(LI-COR®_,_Lincoln,

(4)

Table2

LevelsoftotalphenolicsdetectedinfruitofM82andIL7-3growningreenhouseover

4consecutivetrials.Totalphenolicsareexpressedasmgofgallicacidequivalents

per100goffreshweight.Meanvalue,StandardErrors(SE),andrelativelevelin

IL7-3vs.M82(%ofcontrol)areshown.Withineachtrial,asteriskindicatessignificant

differencesbetweenIL7-3andM82atStudent’sttestcoupledwithbootstrap(1000

samplesandalpha0.05).

Trial Genotype Mean SE Relative

level(%) I M82 31.99 1.68 129 IL7-3 41.30* _3.09 II M82 31.98 4.73 171 IL7-3 54.84* _1.41 III M82 30.60 3.10 154 IL7-3 47.21* _1.98 IV M82 34.07 2.01 168 IL7-3 57.31* _6.34

NE,USA)usingdenaturingpolyacrylamidegelsandgelimageswere

analyzedusingAdobe Photoshopsoftware(Adobe SystemsInc.,

SanJosé,CA,USA).Onceidentified,mutationswerevalidatedby

sequencingtheerf1targetfragment.

Prediction analysis of the relevance of the mutation for

the protein to keep the function was achieved by using SIFT

(http://sift.jcvi.org/) [39], and I-TASSER (http://zhanglab.ccmb. med.umich.edu/I-TASSER/)a tool enabling prediction ofprotein conformationalchanges[40].

3. Results

3.1. Fruitphenotyping

Inourlaboratoryapopulationofintrogressionlinesharbouring singlehomozygoussegmentsoftheS.pennelliigenomeintheS. lycopersicumcv.M82backgroundhasbeenanalysedover consecu-tivegrowingseasonsinordertoinvestigateQTLcontrollingtomato fruitquality.Amongtheselines,IL7-3wasselectedforfurther anal-ysisbecauseitstablymaintainedoverfourtrialshigherlevelsof fruittotalphenolicscomparedtoM82(Table2).Overall,thelevelof totalphenolicswas32.16±1.51(mean±SE)and50.14±2.32mg ofGAEper100gFWinred-ripefruitsofM82andIL7-3, respec-tively.AccordingtotheUnivariateANOVAtest,thedifferencewas statisticallysignificant(P<0.001)withIL7-3expressingonaverage aleveloftotalphenolics56%higherthanM82andtheinteraction genotype×trialwasnotsignificant.Inaddition,thehigher accu-mulationoftotalphenolicsinIL7-3fruitcomparedtoM82was confirmedwithineachtrialbasedonStudent’st-test(P<0.05) reit-eratedon1000samplesgeneratedbybootstrap.Whencompared toM82,IL7-3accumulatedinthefruitahigherrelativelevel(per centofcontrol)oftotalphenolics.Inthelasttrial,totalflavonoids werealsoevaluated.Theirlevelwas8.48±0.54and5.06±0.41mg ofQEper100gFWinred-ripefruitofIL7-3andM82,respectively.

ThehighervalueinIL7-3wasstatisticallyvalidatedbyStudent’st testcoupledwithbootstrap1000samplesandalpha0.05.

Inthesecondandthirdtrials,red-ripefruitsfromIL7-3andM82 werealsoassayedforglutathionecontent(Table3).Accordingto Student’st-testonabootstrappopulationof1000samples,IL7-3 expressedastatisticallysignificanthigheraccumulation(!<0.05) oftotalglutathione(GSH+GSSG)andoxidizedglutathione(GSSG) thanM82inbothtrials,whereasthevariationsinthelevelofthe reducedglutathione(GSH)wassignificantonlyinthesecondtrial. 3.2. Transcriptomicanalysis

InordertoidentifymRNAsregulatedatdifferentlevelsinM82 andIL7-3,biologicalreplicatesoftotalRNAwereextractedfrom threesamplestakenfromeachgenotypewithinbothsecondand thirdtrials.Differentialtranscriptaccumulationwasdocumented usingsingle-colourhybridizationontheTomatArray1.0 microar-ray followed bya two-factorialANOVA test (P<0.01) with the M82transcriptomeasareference.Thecomparativetranscriptomic approachallowedidentifying260sequences(1%ofthose repre-sentedontheTomatArray1.0)thatweredifferentiallyexpressedat thered-ripestage(SupplementaryTableS1),ofwhich129(49.6%) wereup-regulatedand131(50.4%)weredown-regulatedinIL7-3 andonly2.7%ofthemdidnotmatchwithanyannotatedpredicted gene.Theyweredistributedoveralltomato chromosomeswith aproportionofTCsmappingtoeachofthetwelvechromosomes outofalldifferentiallyexpressedTCsrangingfrom2.7%(7TCs)for chromosome11to23.8%(62TCs)forchromosome7.

Inordertoexplainthehighercontentofphenolicsinthefruit of IL7-3, we directed ourattention to genesbelonging to pro-cessescontrollingthelevel of phenolicsin planttissues,which includebiosynthesisanddegradation/utilizationofintermediates aswellasvacuolesequestration.Weinitiallyanalysed transcrip-tomicdatasearchingforchangesinthesteadystatemRNAlevels ofgenesrequiredforbiosynthesisanddegradationofphenolics. Among genes implicated in controlling biosynthesis of pheno-lics, 88 TCs (corresponding to55 predictedgenes according to ITAG release 2.3, Supplementary Table S2) putatively encoding forkey enzymeswereprobedonthemicroarray slideandthey did not show differential expression between IL7-3 and M82. Consequently,we focusedourattentiononadifferentgroupof genesthatshowedhighdifferencesinexpressionlevelbetween IL7-3andM82andselectedthemforfurtheranalysis.The iden-tifiedsub-set oftranscripts(Table4)wasincludedin“ethylene signalling pathway”, “conjugation/compartmentalization path-way” and “response to stress”. Among up-regulated genes there were: one ERF1 (TC174512, Solyc05g051200), two GST (TC170110andTC185053,Solyc07g056420andSolyc07g056440, respectively), one gene coding for a vacuolar sorting protein (VSP)(TC172097, Solyc06g005080),onegene coding for a mul-tidrug resistanceproteinbelonging toMATEfamily (TC178409, Solyc10g081260) in addition to a group of genes involved in stress responses.Among the down-regulated genesthere were

Table3

Fruitcontentofreducedglutathione(GSH),totalglutathione(GSH+GSSG)andoxidizedglutathione(GSSG)inM82andIL7-3overthesecondandthirdtrials.Contentlevels

areexpressedasmgpergoffreshweight.Meanvalue,standarderrors(SE),andrelativelevelinIL7-3vs.M82(%ofM82)areshown.Withineachtrial,asteriskindicates

significantdifferencesbetweenIL7-3andM82atStudent’sttestcoupledwithbootstrap(1000samplesandalpha0.05).

Trial Genotype GSH Relativelevel(%) GSH+GSSG Relativelevel(%) GSSG Relativelevel(%)

Mean SE Mean SE Mean SE

II M82 0.049 0.005 106 2.675 0.089 112 2.629 0.085 112

IL7-3 0.052 0.005 3.002* _0.041 _2.950* _0.040

III M82 0.046 0.004 119 2.086 0.105 148 2.052 0.108 140

(5)

Table4

SubsetofdifferentiallyexpressedTCsinvolvedinphenolicsaccumulationandstressresponse.ForeachTCtheIDofITAG2.3annotation,itsfunctionaldescription,thefold

change(IL7-3vs.M82)fromthemicroarray(seealsoSupplementaryTable1)andtheRT-qPCRanalysesisreported.

Codea _TC_ID _ITAG2.3_gene_ID _Chr _Annotation _IL7-3_vs._M82 _erf1_vs._RS

Microarray RT-qPCR RT-qPCR

Mean SE Mean SE

Ethylenesignallingpathway

1 TC174512 Solyc05g051200.1.1 5 Ethylene-responsivefactor1A 3.04 1.98* _0.72 _−0.80 _0.51

2 TC170138b _{Solyc07g053740.1.1} ₇ _{Ethylene-responsive}_{transcription}_factor₄ _−7.65 _−1.04* _0.11 _1.97* _0.32

Conjugation/compartmentalizationpathway

3 TC170110b _{Solyc07g056420.2.1} ₇ _Glutathione_{S-transferase-like}_protein _3.65 _0.59* _0.03 _−8.30* _1.66

4 TC172606b _{Solyc07g056510.2.1} ₇ _Glutathione_{S-transferase} _−7.18 _−12.83* _0.20 _−1.62* _0.68

5 TC176201b _{Solyc07g056490.2.1} ₇ _Glutathione_{S-transferase} _−5.27 _2.53* _0.17 _0.80* _0.39

6 TC185053b _{Solyc07g056440.2.1} ₇ _Glutathione_{S-transferase-like}_protein _2.76 _4.28* _0.27 _−0.36 _0.58

7 TC172097 Solyc06g005080.2.1 6 Vacuolarsortingprotein 1.20 0.52* _0.03 _0.73* _0.05

8 TC178409 Solyc10g081260.1.1 10 MultidrugresistanceproteinmdtK 1.30 1.63* _0.35 _0.49 _0.32

Stressresponse

9 TC172943 Solyc02g063420.2.1 2 Hydroxyproline-richglycoprotein 2.68 2.07* _0.30 _−0.61* _0.44

10 TC177287 Solyc01g095930.2.1 1 O-acyltransferaseWSD1 2.12 1.81* _0.23 _0.93* _0.36

11 TC177388 Solyc08g080640.1.1 8 Osmotin-likeprotein 2.39 2.42* _0.70 _−2.50* _0.51

12 TC178155 Solyc08g080630.2.1 8 Chymotrypsininhibitor-2 3.40 2.79* _0.42 _−6.66* _1.26

13 TC179186 Solyc02g031950.2.1 2 Pathogenesis-relatedprotein 3.53 3.49* _1.41 _0.82* _0.24

14 TC186230 Solyc05g041420.2.1 5 Serine/threonineproteinkinase 0.68 1.20* _0.20 _−0.25 _0.13

a_The_same_codes_for_the_TCs_listed_in_column₂_are_used_in_Figs.₁_and₃_.

b_TCs_mapping_on_the_{introgression}_region_7-3.

*_{Statistically}_significant_differences_at_P_<_0.05.

ERF4(TC170138,Solyc07g053740)andthreeGSTgenes(TC171442, TC172606,andTC176201,Solyc07g056420,Solyc07g056510,and Solyc07g056490,respectively).TC170110andTC171442revealed to represent the same GST predicted gene (corresponding to Solyc07g056420),and followingdetailed sequenceanalysis(see below)onlyone(TC170110)waschosenforsubsequent investi-gations.

Sincethedown-regulationofgenesmappingintothewild intro-gressed region7-3(TC170138, TC170110,TC172606, TC176201, TC185053) might also be due to mismatches occurring in the regionwherethe35mermicroarrayprobeand/ortheqPCRprimers anneal, we searched for polymorphisms in these genes, using the sequences obtained from the ongoing S. pennellii genome sequencingproject.ForERF4(TC170138),wecouldverifythatfour nucleotidemismatchesoccurredbetweenIL7-3andM82in the regionwhere themicroarray probeannealed. Suchmismatches could conceivably affect the expression level revealed by the microarray hybridization(SupplementaryFig.S1), thus explain-ingthenegativefoldchange(−7.65 inIL7-3vs.M82)observed, sincetheprobewasdesignedonaTCcomingfromthecultivated tomato.Onlyonemismatchwasdetectedintheprobesfortwo GSTs(TC172606, TC176201) but these would not influence the hybridizationsignal,whereasmismatchesoccurredbetween IL7-3 andM82 withtheprobeTC171442,thus leadingustoselect TC170110asthespecificESTrelated toSolyc07g056420.In any case,forabetterexpressionlevelestimationofallthesegenes,the qPCRprimerpairsweredesignedonconservedregionsbetween thewildgenotypeS.pennelliiandthecultivatedS.lycopersicum. Consequently,thenegativefoldchangeestimatedbyqPCRforERF4 (−1.04)wasmorereliable,aswellasthoseestimatedfortheGSTs. qPCRconfirmedthehigherexpressionlevelofthegenescodingfor ERF1,GST,MATEandVSPproteins(Table4).Theup-regulationof agroupofstress-relatedgeneswasalsoconfirmed,aswellasthe down-regulationofERF4andoneGST(TC172606).Adiscrepancy betweenmicroarrayandReal-Timeexpressiondatawasfoundonly forTC176201.

Finally,promoters withinthesegeneswerescreenedfor the presenceofGCC-boxes.Suchelementswerefoundatposition−360 (complementarystrand)fromthestartcodonoftheMATEgene, andatposition−79oftheVSPgene.Thechymotrypsininhibitor,

theosmotin-likeproteinandthepathogenesis-relatedprotein pro-motersalsoincludeGCCboxes,however,noGCC-boxwasfoundin thepromoterofanyoftheanalyzedGSTs.

Focusingontheselectedtranscripts,weproposethatthehigher levelofphenolicsinIL7-3istheresultofamoreefficient conju-gation/compartmentalizationofthesemetabolites,drivenbythe higherexpressionlevelofaMATE,aVSP,andoftheover-expressed GSTgenes(Fig.1).Inourmodelwealsoproposeakey-roleofERF1 asapotentialregulatorofthemechanismsthatcontrolphenolics compartmentalizationinthevacuole.Inordertovalidatetherole envisagedforthisethyleneresponsivefactor,weexploreda func-tionalTILLINGplatformavailableforthetomatocultivarRedSetter tofindmutantsfortheERF1gene.

3.3. Isolationandcharacterizationofanerf1mutant

Screeningof5221EMS-treatedM3familiesresultedinthe iden-tificationofamutantlineharbouringamutationinthesequenceof theerf1gene(TC174512).Themutationwasacytosinetothymine transitionatposition521ofthecds(Fig.2)andcausedintheERF1 proteinsequencetheaminoacidchangeatposition174from Ser-inetoPhenylalanine(S174F).SIFTalgorithmallowedtheprediction ofwhetherthisaminoacidsubstitutionaffectsproteinfunction. Thepredictionwasbasedonthedegreeofconservationofamino acidresiduesinasequencealignmentderivedfrom27ERF1 homo-loguesproteins(SupplementaryFig.S2).Accordingtothepredicted output,theSinposition174showed100%conservationandits substitution withF wasexpectedtobedeleterious(Fig.2).The impactofthemutationonthesecondarystructureoftheprotein hasbeenalsoevaluatedbyI-TASSERshowingthatthissubstitution inthemutantproteincausesthecoiled-coiltochangeinalpha-helix structure.

TenM5plantsoftheerf1mutantlinewereusedinour anal-yses. Thehomozygosity for the mutation in theerf1 gene was validated by sequencing. These plants were grown in a green-house andcharacterizedin comparison withRedSetter control plants(Table5).Onaverage,fruitsfromtheerf1mutantshoweda significantreducedleveloftotalphenolicscomparedtoRedSetter fruits.Inparticular,RedSetterfruitsshowedanaveragecontentin totalphenolicsof44.39±3.65whereasfruitsfromtheerf1mutant

(6)

ERF1 1 ETHYLENE SIGNALING PA THW AY CO N JU GA TI O N /T RA N SP O RT 7 VSP MATE 8

FLAVONOIDS

VACUOLE

3; 5; 6 GST

Fig.1. AschematicrepresentationofphenolicsvacuolaraccumulationintomatofruitoftheIL7-3.

ThemaincandidategenesinvolvedinthesequestrationofphenolicstothevacuolearereportedandnumberedaccordingtocodesindicatedinTable4.ERF1:ethylene

responsivefactor1;GST:glutathione-S-transferase;VSP:vacuolarsortingprotein;MATE:multidrugandtoxiccompoundextrusion.

showedthesignificantlowercontentof27.63±1.31(! <0.05).Also flavonoidsweresignificantlyreducedintheerf1mutantcompared toRedSetter(3.83mg±0.23vs.7.74±0.48ofQEper100gFW). ThefruitlevelsofGSH,GSH+GSSGandGSSGdidnotshowany significantvariationintheerf1mutantlinecomparedtoRedSetter.

ToconfirmtheinvolvementoftheERF1geneintheregulative mechanismscontrollingthelevelof totalphenolicsintheIL7-3 tomatofruits,weassayedbyRT-qPCRtherelativeabundanceof candidatetranscriptsintheerf1mutantcomparedtoRedSetter (Table4).Asexpected,nosignificantreductioninthetranscript

Fig.2. Analysisofmutatederf1geneisolatedbyTILLING.

Thefigureshows:(a)nucleotidesequencesoferf1geneinRedSetteranderf1mutantshowingacytosinetothymidinetransitionat521bpposition;(b)changeofthe

aminoacid174fromSerine(S)toPhenylalanine(F)intheproteinsequence(theunderlinedsequenceidentifytheAP2/EREdomain);(c)conservationintheERF1protein

assessedbyasequencecomparisonof26differenthomologuesproteins(seealsosupplementaryFig.2);(d)SIFTalgorithmoutput;(e)predictedsecondarystructureby

(7)

Table5

Fruitphenotypingoferf1mutantandRedSetter.Fruitcontentoftotalphenolics,totalflavonoids,reducedglutathione(GSH),totalglutathione(GSH+GSSG)andoxidized

glutathione(GSSG)intheerf1mutantandRedSetter.MeanvalueandStandardError(SE)areshown.Asteriskindicatessignificantdifferencesbetweenerf1mutantandRed

SetteratStudent’sttestcoupledwithbootstrap(1000samplesandalpha0.05).

RedSetter erf1mutant

Mean SE Mean SE

Totalphenolics(mgGAE× 100g−1_FW) _44.39 _3.65 _27.63* _1.31

Totalflavonoids(mgQE× 100g−1_FW) _7.74 _0.47 _3.83* _0.23

GSH(mgg−1_FW) _0.03 _0.00 _0.03 _0.00

GSH+GSSG(mgg−1_FW) _2.82 _0.11 _2.88 _0.10

GSSG(mgg−1_FW) _2.73 _0.11 _2.79 _0.10

levelof thetargeterf1(TC174512)gene wasobservedbetween

theRedSetterand theerf1mutant genotype.By contrast,ERF4

(TC170138)showeda1.97foldincreasedexpressioninthemutant

comparedtothewildtype.Asforgenesinvolvedinthe

transmem-branetransportofphenolics,redfruitsfromtheerf1mutantline

showedasignificantlowerabundanceofmRNAsrelatedtotwoGST

genes(TC170110,TC172606),whereasasimilarlevelofexpression

intherelativetranscriptabundanceoftheMATEgene(TC178409)

andahigherexpressionoftheVSPgene(TC172097).Finally,the

lowerexpressionofstress-relatedgenesexhibitedbythemutant

alsosupportedthereducedefficiencyofthemutatederf1genein

promotingtheirexpression.

4. Discussion

Inthepresentworkweinvestigatedthemolecularmechanisms

thatleadtoahighertomatophenolicscontentinaS.pennellii

intro-gressionlinebycombiningcurrentlyavailable-omicsresourcesfor

thisspeciessuchasthetranscriptomicCombimatrixarrayandthe

TILLINGgenefunctionalplatform.Thefirstapproachallowedthe

identificationofagroupofgenescontributingtotheincreaseof

vacuolarsequestrationofphenolicsandledustohypothesizea

key-roleofthegeneERF1inorchestratingthemechanismsinvolved

therein.Thesecondallowedtheidentificationandcharacterization

ofoneerf1mutantthatsupportstheroleofERF1geneincontrolling

thelevelofphenolics/flavonoidsinthefruitoftheselectedtomato

introgressionline.

Thegenotypeweanalysed(IL7-3)exhibitedanaveragelevelof

fruittotalphenolicshigherthanitscultivatedparentallineM82

over fourgreenhouse trials.The highercontent of phenolicsin

fruitsofIL7-3vs.M82wasconfirmedovertwofieldtrials(datanot

shown).Inarecentworkcarriedoutonthesameplantmaterials

usedinthepresentpaper,ithasbeenestablishedbyHPLCanalysis

thattheleveloftheflavonolrutinwassignificantlyhigherinthe

fruitofIL7-3comparedtoM82[41].Accordingly,ourpreliminary

datarevealedahigherleveloftotalflavonoidsintheIL7-3fruit. Comprehensively,thesedataallowedtheconfirmationofamajor QTLforfruitphenolicsaccumulationmappingtotheintrogressed region3ofchromosome7thatwasexpressedindependentlyofthe trialandenvironmentalconditions.Indeed,intheTomato Expres-sionDatabase(http://ted.bti.cornell.edu)[42]thislinewasalready reportedasexhibitingafruitrelativelevelhigherthanM82(139.7% ofcontrol)andaverylowlevelofcarotenoids.Therefore,inorderto usethisintrogressionlineassourceoffavourableallelesforhigher phenolicsinthefruititwouldbenecessarytodisruptthelinkage betweenthegene(s)thatpromotephenolicscontentandthosethat contributetolowercarotenoids.Nowadays,genomicstoolssuchas transcriptomicprofilingandgenomesequencing/annotationmight enabletheaccuratedissectionofthisquantitativetraitin candi-dategenestobeusedinbreedingforfruitphenolicsaccumulation. Altogether,theintrogressionlineIL7-3wasusedinacomparative transcriptomicanalysiswithrespecttoitscultivatedparentalline M82.

Thedifferentialgeneexpressionevidencedinourexperiments strengthensthehypothesisthatincreasedtransportandvacuolar sequestrationofphenolicsinIL7-3ledtothehigherphenolicslevel initsfruitsratherthantoanelevatedbiosyntheticflux.Wethus assumedthatinIL7-3thesequestrationofphenolicsintovacuoles isregulatedaccordingtomechanismspreviouslydescribedinother species, suchas maize,petunia,Arabidopsis andgrape [43–48]. Indeed,thehigherexpressionintheIL7-3ofgenescodingforthree GSTs,oneVSPandoneMATEfitswiththeco-operative contribu-tionoftheseproteinsfortheaccumulationofanthocyaninsinthe vacuole,recentlyevidencedingrapevine[49].

Inparticular,themodelweproposedisbasedonthecentral roleofoneERF1genemappingonchromosome5,whichis dis-tinct fromtheTERF1mappingonchromosome3 thatenhances osmoticstresstoleranceintomato [50]. Theover-expressionof ERF1inIL7-3makesitagoodcandidatefortheregulationof pro-cessesinvolvedinfruitphenolicsaccumulation.Wereasonedthat theover-expressionofERF1mightbalancetheevidenced down-regulationof ERF4sincenopolymorphism isexpectedbetween IL7-3 and M82 in the ERF1 gene mapping on chromosome 5. Indeed,downregulationofERF4,thatmapsintheintrogression region,mightbeduetopolymorphismsevidencedinthe“wild” pennellii promoter of ERF4and/or todifferent regulators acting inthewildandthecultivatedtomatogenotypesinconcertwith theERFgenes.Thisbalancebetweenethyleneresponsivefactors wouldensureproperethylenefunctionsduringthedevelopment of IL7-3fruit through a homeostatic cellular mechanism, even thoughthesetwofactorsmightalsohavedifferentspecificroles in thefruitcells. Indeed,nospecificfunctionshavebeenuptill now assigned tothe numerous membersof tomato ERFfamily and a highfunctional redundancy hasbeenproposed forthem [51].Followingourhypothesis,inthegenotypeIL7-3thehigher expressionofERF1mightdriveanincreasedtranscriptional reg-ulationof genescodingfor someGSTs,oneMATEandone VSP that resultsin anincrease of vacuolarcompartmentalizationof phenolics,specificallyflavonoids.Hence,amongmembersofthe AP2/ERF family, ERF1 integrates signals fromethylene and jas-monicacidsignallingpathwaysandultimatelytriggersadefence response againstpathogenattackbyinducingtheexpressionof defencegenes,suchasPRproteins,HPRGsandGSTs,usuallyby bindingthespecificDNAtargetGCCbox[52–54].Here,the regula-toryactionofERF1throughtheGCC-boxbindingcouldbeassumed fortheMATEandtheVSP,aswellasforsomestress-relatedgenes thatalsodisplayedup-regulation.Unexpectedly,thiscis-element wasnotdetectedinpromotersoftheGSTsup-regulatedinIL7-3. Thismightbesimilartowhatwaspreviouslyobservedforthe tran-scriptionfactorPti4,whichappearedtoregulategeneexpression intomatobybindingGCCboxbutalsonon-GCCboxciselements ofmanygenepromoters[55].Inaddition,recentlyitwas demon-stratedinArabidopsisthattheAtERF2proteinspecificallybindsto thepromoteroftheAtGST11wherenoGCC-boxwasdetected[56]. TheseobservationscouldthussupportourhypothesisthatERF1 mightdirectlyregulateexpressionofgenesinvolvedinthe trans-portofflavonoidstothevacuole.Itisimportanttonote,however,

(8)

Fig.3. TranscriptomicandphenotypictrendanalysisofthetwodifferentgeneticsystemsIL7-3vs.M82anderf1mutantvs.RedSetter.

GeneexpressiondataarethosereferredtoqPCRanalysis(foldchange)andgenesarenumberedaccordingtocodesreportedinTable4.Phenotypicdataarealsoreported

asfoldchange.ERF1:ethyleneresponsivefactor1;ERF4:ethyleneresponsivefactor4;GST:glutathione-S-transferase;VSP:vacuolarsortingprotein;MATE:multidrug

andtoxiccompoundextrusion;HPRG:histidine–prolinerichglycoprotein;WSD1:O-acyltransferaseWSD1;Osmo:Osmotin-likeprotein;CI:Chymotrypsininhibitor-2;PR:

Pathogenesis-relatedprotein;Ser/Th:Serine/threonineproteinkinase;TotPhe:Phenolicsfruitcontent;TotFla:Flavonoidfruitcontent.

thatERF1mightalsoindirectlyinteractwithothertranscription factorsinactivatingsomepromotersassuggestedbySolanoetal. [57]whostudiedthetranscriptionalcascadecontrollingethylene signallinginArabidopsis.ImplicationsofanERF1genecontrolling phenolicsaccumulationhavebeenpreviouslyreportedin photo-synthetictissuesof Arabidopsis[54],where theinvolvement of auxinandethyleneininducingflavonoidsaccumulationhasbeen demonstrated.Inparticular,itwasproposedthattheincreaseof expressionofasubsetofgenesoftheflavonoidpathwaywas reg-ulated through the activationof ethylene responsive elements, perhapswiththeparticipationofaMYBtranscriptionalfactor.

Toourknowledge,thisisthefirsttimethattheroleofERF1 inmodulatingphenolicsaccumulationisreportedinfleshyfruit, andtheselectionofaTILLINGmutantharbouringasinglepoint mutationintheerf1cdspermittedustosupportsuchhypothesis. Unfortunately,nofunctionalmutantwasfoundintheERF4gene (datanotshown)andonlyonemutationwasdetectedintheERF1 gene.Theavailabilityofmorethanonegenotypeharbouring muta-tionsineitherERF1orERF4,orinbothgenes,wouldhaveallowed confirmingtheirinvolvementintheregulationofphenolics accu-mulationin tomato fruit.However,theserine tophenylalanine aminoacidsubstitutionobservedintheerf1mutantlineis pre-dictedtoinduceamodificationinthethree-dimensionalstructure oftheprotein(Fig.2).Moreover,thisreplacementinvolvesthe C-terminaloftheproteinthatishighlyconservedandthereforecould becrucialforitsproperfunctioning.Basedontheseconsiderations, wepresumedthatthemodifiedstructuremightaffecttheactivityof theerf1protein.Inaddition,eventhoughothermutationsaffecting theproteinactivitycouldhaveoccurredinthesamemutant geno-typeinothergenesinadditiontoerf1,accordingtothefrequency ofmutationreportedfortheRedSetterTILLINGplatform[37,58]

itisveryunlikelythatsuchmutationswouldinvolveothergenes influencingthephenolicsbiosynthesisandstorage.

The subsequent characterization of our erf1 mutant line at thetranscriptionalandphenotypiclevel supportedthe involve-mentof thisethyleneresponsiveelementinregulatinga group of genes that comprehensively might control the level of phe-nolicsintomatofruit.Firstofall,thehigherexpressionof ERF4 mightresultfromareducedfunctionalityoferf1andthiswouldbe inlinewiththebalancingbetweenthesetwoERFgenesalready observed in IL7-3. Moreover, the reduced functionality of erf1 might also be related to the lower expression, in comparison withthatevidencedin thelineIL7-3,of theMATEand ofsome stress-relatedgenes,whoseexpressionismodulatedbythe pres-ence of GCC-box in their promoter. At the phenotypic level, a reduced level of total phenolics and flavonoids in the mutant respecttothewildtypeRedSetterreinforcestheroleofERF1in orchestratingthecompartmentalizationofthesecompoundstothe vacuole.Comprehensively,whencomparingthetwosystems ana-lysedinthepresentwork(IL7-3vs.M82anderf1mutantvs.Red Setter),changesintheexpressionofgenesinvolvedinthe conju-gation/compartmentalizationofphenolicsandinstressresponse appear to follow variations of thephenolics/flavonoids content in thetomato fruit(Fig.3).Thisismost probablyexplainedby thereducedactivityof themutatederf1proteinin theTILLING mutant andbythehigherexpression ofERF1gene intheIL7-3 introgressionline(bothcompensatedinpartbythecontrasting expressionofERF4).Nevertheless,afurtherfunctionalvalidationby silencingtheseERFgenesinIL7-3orover-expressingtheminM82 willneedtobecarriedoutbeforedefinitivelyconsideringthem askey-candidatesforimprovingthephenolicscontentoftomato fruit.

(9)

Overall,thesedatasuggestthatamoreefficient compartmental-izationtothevacuoleunderliesthehigherfruitphenolicscontent ofIL7-3.Thisislikelyachievedbytransportmechanismsmediated byaMATE,aVSPandsomeGSTs,andtranscriptionalresultsfrom theerf1mutantstronglysuggestthattheroleofoneparticularGST (TC170110)mightbekeyinaffectingphenolics/flavonoids accu-mulation.Finally,aconcertedactionofethyleneresponsivefactors ERF1andERF4indirectingthistransporthasbeenhypothesized, designatingthemasnovelcandidategenestoincreasephenolics contentin tomatofruitviageneticmanipulation strategiesthat could enhance/reduce their fruit specificexpression levels. We mightalsoassume theputativerole ofothergenesmappingin theintrogressionregion7-3asmasterregulatorsofERF1.Indeed, thegeneticdiversitythatexistsbetweenthecultivatedandwild species in this genomicregion mightreflect differences alsoin keyregulatoryelements,whichcouldboostERF1levelsintomato fruit cells. Nowadays, thetomato genomehasbeencompletely sequenced[59] and a preliminaryannotationof genes mapped onthe region7-3 revealed thepresence of putativeregulatory elements(datanotshown).Thesegenescoulddirectlypromote thehigher expressionofERF1 inIL7-3in a trans-acting regula-torymechanismormightbeactiveinmodifyingERF4expression, thus indirectly inducinga higherexpression of ERF1through a homeostasisbalancing mechanism.Theirpotentialroleas addi-tionalcandidategenestodrivetheincreaseofERF1transcription inthefruitandtheconsequentaccumulationofphenolicsshould befurtherinvestigated.These “wild”regulatoryelementscould thenbetransferredintocultivatedvarietiesbysexualhybridization withoutadoptinggenetictransformation.

Acknowledgements

This workwas principallyfunded by the MiPAF AGRONAN-OTECHandMiURGENOPOMprograms.

AppendixA. Supplementarydata

Supplementary data associated with this article can be found,intheonlineversion,athttp://dx.doi.org/10.1016/j.plantsci. 2013.02.001.

References

[1]B.Demmig-Adams,B.B.AdamsIII,Antioxidantsinphotosynthesisandhuman

nutrition,Science298(2002)2149.

[2] H.C.Hung,K.J.Joshipura,R.Jiang,F.B.Hu,D.Hunter,S.A.Smith-Warner,G.A.

Colditz,B.Rosner,D.Spiegelman,W.C.Willett,Fruitandvegetableintakeand

riskofmajorchronicdisease,JNatl.CancerI96(2004)1577–1584.

[3]F.J.He,C.A.Nowson,G.A.MacGregor,Fruitandvegetableconsumptionand

stroke:meta-analysisofcohortstudies,Lancet367(2006)320–326.

[4]K.Canene-Adams,J.K.Campbell,S.Zaripheh,E.H.Jeffery,J.W.ErdmanJr.,The

tomatoasafunctionalfood,J.Nutr.135(2005)1226–1230.

[5]J.P.Spencer,G.G.Kuhnle,M.Hajirezaei,H.P.Mock,U.Sonnewald,C.Rice-Evans,

Thegenotypicvariationoftheantioxidantpotentialofdifferenttomato

vari-eties,FreeRadic.Res.39(2005)1005–1016.

[6]J.B.Harborne,C.A.Williams,Advancesinflavonoidresearchsince1992,

Phy-tochemistry55(2000)481–504.

[7]G.A.Cooper-Driver,M.Bhattacharya,Roleofphenolicsinplantevolution,

Phy-tochemistry45(1988)1161–1190.

[8]S.Moco,E.Capanoglu,Y.Tikunov,R.J.Bino,D.Boyacioglu,R.D.Hall,J.Vervoort,

R.C.DeVos,Tissuespecializationatthemetabolitelevelisperceivedduring

thedevelopmentoftomatofruit,J.Exp.Bot.58(2007)4131–4146.

[9] S.Moco,R.J.Bino,O.Vorst,H.A.Verhoven,J.DeGroot,T.A.VanBeek,J.Vervoort,

C.H.R.deVos,Aliquidchromatography-massspectrometry-basedmetabolome

databasefortomato,PlantPhysiol.141(2006)1205–1218.

[10]Y.Iijima,Y.Nakamura,Y.Ogata,K.Tanaka,N.Sakurai,K.Suda,T.Suzuki,H.

Suzuki,K.Okazaki,M.Kitayama,S.Kanay,K.Aoki,D.Shibata,Metabolite

anno-tationsbasedontheintegrationofmassspectralinformation,PlantJ.54(2008)

949–962.

[11] M.E.Verhoeyen,A.Bovy,G.Collins,S.Muir,S.Robinson,C.H.deVos,S.Colliver,

Increasingantioxidantlevelsintomatoesthroughmodificationoftheflavonoid

biosyntheticpathway,J.Exp.Bot.53(2002)2099–2106.

[12]S.R.Miur,G.J.Collins,S.Robinson,S.Hughes,A.Bovy,C.H.R.DeVos,A.J.

vanTunen,M.E.Verhoeyen,Overexpressionofpetuniachalconeisomerasein

tomatoresultsinfruitcontainingincreasedlevelsofflavonoids,Nat.Biotechnol.

19(2001)470–474.

[13]H.Mathews,S.K.Clendennen,C.G.Caldwell,X.LiangLiu,K.Connors,N.Matheis,

D.K.Schuster,D.J.Menasco,W.Wagoner,J.Lightner,D.R.Wagner,Activation

taggingintomatoidentifiesatranscriptionalregulatorofanthocyanin

biosyn-thesis,modificationandtransport,PlantCell15(2008)1689–1703.

[14]J.Luo,E.Butelli,L.Hill,A.Parr,R.Niggeweg,P.Bailey,B.Weisshaar,C.

Mar-tin,AtMYB12regulatescaffeoylquinicacidandflavonolsynthesisintomato:

expressioninfruitresultsinveryhighlevelsofbothtypesofpolyphenol,Plant

J.56(2008)316–326.

[15]E.Grotewold,Thechallengesofmovingchemicalswithinandoutofcells:

insights into the transportof plantnatural products, Planta219 (2004)

906–909.

[16] K.Yazaki,Transportersofsecondarymetabolites,Curr.Opin.PlantBiol.8(2005)

301–307.

[17] J.Zhao,A.D.Dixon,The‘ins’and‘outs’offlavonoidtransport,TrendsPlantSci.

15(2009)73–80.

[18] G.Agati,E.Azzarello,S.Pollastri,M.Tattini,Flavonoidasantioxidantsinplants:

locationandfunctionalsignificance,PlantSci.196(2012)67–76.

[19]E.Grotewold,K.Davies,Traffickingandsequestrationofanthocyanins,Nat.

Prod.Commun.3(2008)1251–1258.

[20]K.A.Marrs,M.R.Alfenito,A.M.Lloyd,V.Walbot,AglutathioneS-transferase

involvedinvacuolartransferencodedbythemaizegeneBronze-2,Nature375

(1995)397–400.

[21]L.A.Mueller,C.D.Goodman,R.A.Silady,V.Walbot,AN9apetunia

glutathione-S-transefraserequiredforanthocyaninsequestration,isaflavonoid-binding

protein,PlantPhysiol.123(2000)1561–1570.

[22]C.S.Buer,G.K.Muday,Thetransparenttesta4mutationpreventsflavonoid

synthesisandaltersauxintransportandtheresponseofArabidopsisrootsto

gravityandlight,PlantCell16(2004)1191–1205.

[23]C.S.Buer,P.Sukumar,G.K.Muday,Ethylenemodulatesflavonoidaccumulation

andgravitropicresponsesinrootsofArabidopsis,PlantPhysiol.140(2006)

1384–1396.

[24] D.R.Lewis,M.V.Ramirez,N.D.Miller,P.Vallabhaneni,W.K.Ray,R.F.Helm,B.S.

Winkel,G.K.Muday,Auxinandethyleneinduceflavonolaccumulationthrough

distincttranscriptionalnetworks,PlantPhysiol.156(2011)144–164.

[25]M.C.Rousseaux,C.M.Jones,D.Adams,R.Chetelat,A.Bennett,A.Powell,QTL

analysisoffruitantioxidantsintomatousingLycopersiconpennellii

introgres-sionlines,Theor.Appl.Genet.111(2005)1396–1408.

[26]B.Okmen,H.O.Sigva,N.Gurbuz,M.Ulger,A.Frary,S.Doganlar,Quantitative

traitloci(QTL)analysisforantioxidantandagronomicallyimportanttraitsin

tomato(Lycopersiconesculentum),Turk.J.Agric.For.35(2011)501–514.

[27] Y.Eshed,D.Zamir,AnintrogressionlinepopulationofLycopersiconpennellii

inthecultivatedtomatoenablestheidentificationandfinemappingofyield

associatedQTL,Genetics141(1995)1147–1162.

[28] A.Barone,A.DiMatteo,D.Carputo,L.Frusciante,High-throughputgenomics

enhancestomatobreedingefficiency,Curr.Genomics10(2009)1–9.

[29]C.M.McCallum,L.Comai,E.A.Greene,S.Henikoff,Targetinginducedlocal

lesionsingenomes(TILLING)forplantfunctionalgenomics,PlantPhysiol.123

(2000)439–442.

[30] B.J.Till,T.Colbert,R.Tompa,L.C.Enns,C.A.Codomo,J.E.Johnson,S.H.Reynolds,

J.G.Henikoff,E.A.Greene,M.N.Steine,L.Comai,S.Henikoff,High-throughput

TILLINGforfunctionalgenomics,MethodsMol.Biol.236(2003)205–220.

[31]V.L. Singleton, J.A. Rossi, Colorimetry of total phenolics with

phosphomolybdic–phosphotungstic acid reagents, Am. J. Enol. Viticult.

16(1965)144–158.

[32]D.Marinova,F.Ribarova,Atanassova,Totalphenolicsandtotalflavonoids

inBulgarianfruitsandvegetables,J.Univ.Chem.Technol.Metall.40(2005)

255–260.

[33]A.Griffiths,C.Barry,A.G.Alpuche-Solis,D.Grierson,Ethyleneand

develop-mentalsignalsregulateexpressionoflipoxygenasegenesduringtomatofruit

ripening,J.Exp.Bot.50(1999)793–798.

[34]A.DiMatteo,A.Sacco,M.Anacleria,M.Pezzotti,M.Delledonne,A.Ferrarini,L.

Frusciante,A.Barone,Theascorbicacidcontentoftomatofruitsisassociated

withtheexpressionofgenesinvolvedinpectindegradation,BMCPlantBiol.

10(2010)163.

[35]L.A.Mueller,R.KleinLankhorst,S.D.Tanksley,J.J.Giovannoni,R.White,J.

Vre-balov,Z.Fei,J.vanEck,R.Buels,A.A.Mills,etal.,Asnapshotoftheemerging

tomatogenomesequence,PlantGenome2(2009)78–92.

[36]K.J.Livak,T.D.Schmittgen,Analysisofrelativegeneexpressiondatausing

realtime quantitative PCRand the 2-DDCTmethod, Methods 25(2001)

402–408.

[37]S.Minoia,A.Petrozza,O.D’Onofrio,F.Piron,G.Mosca,G.Sozio,F.Cellini,

A.Bendahmane,F.Carriero,Anewmutantgeneticresourcefortomatocrop

improvementbyTILLINGtechnology,BMCRes.Notes3(2010)69.

[38]M.Dalmais,J.Schmidt,C.LeSignor,F.Moussy,J.Burstin,V.Savois,G.Aubert,V.

Brunaud,Y.deOliveira,C.Giuchard,R.Thompson,A.Bendahmane,UTILLdba

Pisumsativuminsilicoforwardandreversegeneticstool,GenomeBiol.9(2008)

R43.

[39]P.Kumar,S.Henikoff,P.C.Ng,Predictingtheeffectsofcodingnon-synonymous

variantsonproteinfunctionusingtheSIFTalgorithm,Nat.Protoc.4(7)(2009)

1073–1081.

[40] A.Roy,A.Kucukural,Y.Zhang,I-TASSER:aunifiedplatformforautomated

(10)

[41]M.Minutolo,C.Amalfitano,A.Evidente,L.Frusciante,A.Errico,Polyphemol

distributioninplantorgansoftomatointrogressionlines,Nat.Prod.Res.(2012),

http://dx.doi.org/10.1080/14786419.2012.704371.

[42]Z.Fei,J.-G.Joung,X.Tang,Y.Zheng,M.Huang,J.M.Lee,R.McQuinn,D.M.

Tie-man,R.Alba,H.J.Klee,J.J.Giovannoni,TomatoFunctionalGenomicsDatabase:a

comprehensiveresourceandanalysispackagefortomatofunctionalgenomics,

NucleicAcidsRes.39(2011)D1156–D1163.

[43] C.D.Goodman,P.Casati,V.Walbot,Amultidrugresistance-associatedprotein

involvedinanthocyanintransportinZeamays,PlantCell16(2004)1812–1826.

[44] H.Zhang,L.Wang,S.Deroles,R.Bennett,K.Davies,Newinsightintothevesicles

andformationofanthocyanicvacuolarinclusionsinflowerpetals,BMCPlant

Biol.6(2006)29.

[45]K.Marinova,L.Pourcel,B.Weder,M.Schwarz,D.Barron,J.M.Routaboul,I.

Debeaujon,M.Klein,TheArabidopsisMATEtransporterTT12actsasavacuolar

flavonoid/H++antiporteactiveinproanthocyanidinaccumulatingcellsofthe

seedcoat,PlantCell19(2007)2023–2038.

[46]S.Conn,C.Curtin,A.Bezier,C.Franco,W.Zhang,Purification,molecularcloning,

andcharacterizationofglutathioneS-transferases(GSTs)frompigmentedVitis

viniferaL.cellsuspensionculturesasputativeanthocyanintransportproteins,

J.Exp.Bot.59(2008)3621–3634.

[47] C.Gomez,N.Terrier,L.Torregrosa,S.Vialet,A.Fournier-Level,C.Verriès,J.-M.

Souquet,J.-P.Mazauric,M.Klein,V.Cheynier,A.Ageorges,Grapevine

MATE-typeproteinsactasvacuolarH+dependentacylatedanthocyanintransporters,

PlantPhysiol.150(2009)402–415.

[48]L.Pourcel,N.G.Irani,Y.H.Lu,K.Riedl,S.Schwartz,E.Grotewold,Theformation

ofanthocyanicvacuolarinclusionsinArabidopsisthalianaandimplicationsfor

thesequestrationofanthocyaninpigments,Mol.Plant3(2010)78–90.

[49]C.Gomez,G.Conejero,L.Torregrosa,V.Cheynier,N.Terrier,A.Ageorges,Invivo

grapevineanthocyanintransportinvolvesvesicle-mediatedtraffickingandthe

contributionofanthoMATEtransportersandGST,PlantJ.67(2011)960–970.

[50]Z.Huang,Z.Zhang,X.Zhang,H.Zhang,D.Huang,R.Huang,TomatoTERF1

mod-ulatesethyleneresponseandenhancesosmoticstresstolerancebyactivating

expressionofdownstreamgenes,FEBSLett.573(2004)110–116.

[51]H.J.Klee,J.J.Giovannoni,Geneticsandcontroloftomatofruitripeningand

qualityattributes,Annu.Rev.Genet.45(2011)41–59.

[52]H.Guo,J.R.Ecker,Theethylenesignalingpathway:newinsights,Curr.Opin.

PlantBiol.7(2004)40–49.

[53] T. Nakano, K.Suzuki, T. Fujimura, H. Shinshi, Genome-wide analysisof

the ERF genefamily in Arabidopsisand rice, Plant Physiol. 140 (2006)

411–432.

[54]B.Dombrecht,G.P.Xue,S.J.Sprague,J.A.Kirkegaard,J.J.Ross,J.B.Reid,G.P.Fitt,N.

Sewelam,P.M.Schenk,J.M.Manners,K.Kazan,MYC2differentiallymodulates

diversejasmonate-dependentfunctionsinArabidopsis,PlantCell19(2007)

2225–2245.

[55]S.Chakravarthy,R.P.Tuori,M.D.D’Ascenzo,P.R.Fobert,C.Després,G.B.Martin,

ThetomatotranscriptionfactorPti4regulatesdefence-relatedgene

expres-sionviaGCCboxandnon-GCCboxciselements,PlantPhysiol.15(2003)

3033–3050.

[56]T.Kouno,B.Ezaki,MultipleregulationofArabidopsisAtGST11gene

expres-sionbyfourtranscriptionfactorsunderabioticstresses,Physiol.Plant.(2012),

http://dx.doi.org/10.1111/j.1399-3054,2012.01699.x.

[57]R.Solano,A.Stepanova,Q.Chao,J.R.Ecker,Nuclereventsinethylenesignaling:

atranscriptionalcascademediated byethylene-insenitiev3and

ethylene-response-factor1,GenDev.12(2012)3703–3714.

[58]M.F.Silletti,A.Petrozza,A.L.Stigliani,G.Giorio,F.Cellini,C.D’Ambrosio,F.

Carriero,Anincreaseoflicopenecontentintomatofruitisassociatedwitha

novelCyc-BalleleisolatedthroughTILLINGtechnology,Mol.Breeding(2012),

http://dx.doi.org/10.1007/s11032-012-9824-6.

[59]TheTomatoGenome Consortium,Thetomatogenomesequenceprovides