MicroRNA

ThePlantCell,Vol.23:3185–3203,September2011,http://wendang.chazidian.comã2011AmericanSocietyofPlantBiologists.Allrightsreserved.

FunctionandEvolutionofaMicroRNAThatRegulatesaCa2+-ATPaseandTriggerstheFormationofPhasedSmallInterferingRNAsinTomatoReproductiveGrowth

YingWang,aAsukaItaya,a,1XuehuaZhong,a,2YangWu,bJianfengZhang,a,3EsthervanderKnaap,cRichardOlmstead,dYijunQi,bandBiaoDinga,e,f,4

aDepartmentbNational

ofMolecularGenetics,OhioStateUniversity,Columbus,Ohio43210InstituteofBiologicalSciences,102206Beijing,China

cDepartmentofHorticultureandCropScience,OhioStateUniversity/OhioAgriculturalResearchandDevelopmentCenter,Wooster,Ohio44691

dBurkeMuseum,DepartmentofBiology,UniversityofWashington,Seattle,Washington98195

eMolecular,Cellular,andDevelopmentalBiologyProgram,OhioStateUniversity,Columbus,Ohio43210fCenterforRNABiology,OhioStateUniversity,Columbus,Ohio43210

MicroRNAs(miRNAs)regulateawidevarietyofbiologicalprocessesinmosteukaryotes.WeinvestigatedthefunctionandevolutionofmiR4376inthefamilySolanaceae.Wereportthatthe22-nucleotidemiR4376regulatestheexpressionofanautoinhibitedCa2+-ATPase,tomato(Solanumlycopersicum)ACA10,whichplaysacriticalroleintomatoreproductivegrowth.Deepphylogeneticmappingsuggested(1)anevolutioncourseofMIR4376lociandposttranscriptionalprocessingofpre-miR4376asalikelylimitingstepfortheevolutionofmiR4376,(2)anindependentphylogeneticoriginofthemiR4376targetsiteinACA10homologs,and(3)alternativesplicingasapossiblemechanismofeliminatingsuchatargetinsomeACA10homologs.Furthermore,miR4376triggerstheformationofphasedsmallinterferingRNAs(siRNAs)fromSlACA10anditsSolanumtuberosumhomolog.Together,ourdataprovideexperimentalevidenceofmiRNA-regulatedexpressionofuniversallyimportantCa2+-ATPases.ThemiR4376-regulatedexpressionofACA10itself,andpossiblyalsotheassociatedformationofphasedsiRNAs,mayfunctionasanovellayerofmolecularmechanismsunderlyingtomatoreproductivegrowth.Finally,ourdatasuggestthatthestochasticemergenceofamiRNA-targetgenecombinationinvolvesmultiplemoleculareventsatthegenomic,transcriptional,andposttranscriptionallevelsthatmayvarydrasticallyinevencloselyrelatedspecies.

INTRODUCTION

MicroRNAs(miRNAs)playessentialrolesinregulatinggeneex-pressioninmosteukaryotes.ThebiogenesisofmiRNAsinvolvesseveralmechanisticsteps.Ingeneral,miRNAgenesaretran-scribedintomiRNAprecursorsbytheDNA-dependentRNApolymeraseII.TheseprecursorsarefurtherprocessedintomaturemiRNAsthatcarryoutspeci?cbiochemicalfunctionsingeneregulation.Inanimals,thelongprimarytranscripts,calledprimary-miRNAs,are?rstprocessedbyDroshaandassociatedprotein

address:AgricultureandAgri-FoodCanada,960Carling

Avenue,Building21,Ottawa,OntarioK1A0C6,Canada.

2Currentaddress:DepartmentofMolecular,Cell,andDevelopmentalBiology,3218LifeSciencesBuilding,UniversityofCalifornia,LosAngeles,CA90095.

3Currentaddress:DepartmentofCellandSystemsBiology,UniversityofToronto,25WillcocksStreet,Room4055,Toronto,OntarioM5S3B2,Canada.

4Addresscorrespondencetoding.35@osu.edu.

Theauthorresponsiblefordistributionofmaterialsintegraltothe?ndingspresentedinthisarticleinaccordancewiththepolicydescribedintheInstructionsforAuthors(http://wendang.chazidian.com)is:BiaoDing(ding.35@osu.edu)http://wendang.chazidian.com/cgi/doi/10.1105/tpc.111.088013

1Currentfactorsintoshorterprecursor-miRNAs(pre-miRNAs)inthenu-cleus,whicharethenexportedintothecytoplasmforsubsequentcleavagebyDicerandassociatedfactorsintoduplexesofmiRNA/miRNA*(Kimetal.,2009;Okadaetal.,2009).PlantslackDroshahomologsandbothprocessingstepsarecatalyzedbyaDicer-likeenzyme(DCL1)inthenucleus(Parketal.,2002;Reinhartetal.,2002;KuriharaandWatanabe,2004;Kuriharaetal.,2006).Inmostcases,miRNAsareloadedintotheRNA-inducedsilencingcom-plextoguidecleavageoftargetmRNAsorrepressionoftransla-tion(CarthewandSontheimer,2009;Voinnet,2009).

Inplants,manymiRNAsaredeeplyconservedandabundantlyexpressed,andtheyregulateoverwhelminglytheexpressionoftranscriptionfactorscriticalfordevelopmentorstressresponses(AxtellandBowman,2008;Axtell,2008;Chen,2009).Recentbioinformaticsandhigh-throughputsequencingstudiesalsohaveuncoveredalargenumberofnonconservedmiRNAs,fromagreenalgato?oweringplants,whicharespeciesorcladespeci?c.

´retal.,2007;ExamplesincludeChlamydomonasreinhardtii(Molna

Zhaoetal.,2007),PhyscomitrellapatensandSelaginellamoellen-dorf?i(Axtelletal.,2007),Arabidopsisthaliana(Luetal.,2005;Luetal.,2006;Rajagopalanetal.,2006;Fahlgrenetal.,2007;Zhangetal.,2007;Fahlgrenetal.,2010),Arabidopsislyrata(deFelippesetal.,2008;Fahlgrenetal.,2010;Maetal.,2010),Oryzasativa(Luetal.,2008;Morinetal.,2008;Zhuetal.,2008;Johnsonetal.,

3186ThePlantCell

2009),Pinuscontorta(Morinetal.,2008),Eschscholziacalifornica(Barakatetal.,2007),Populustrichocarpa(Luetal.,2005),Medi-cagotruncatula(Szittyaetal.,2008;Lelandais-Brie

`reetal.,2009),Triticumaestivum(Weietal.,2009),Brachypodiumdistachyon(Weietal.,2009),Solanumlycopersicum(Moxonetal.,2008),andGlycinemax(Joshietal.,2010).SomemiRNAsarespeci?ctomonocotsordicots(SunkarandJagadeeswaran,2008).AcommonpictureemergingfromtheseresultsisthatthesenonconservedmiRNAsareexpressedatlowlevels,targetdiversegenes,exhibitorgan-ortissue-speci?cexpression,andmayhavespecializedfunctions.Forinstance,theA.thalianamiR824,whichisconservedinBrassicaceaebutisnotfoundoutsidethisfamily(Rajagopalanetal.,2006;Fahlgrenetal.,2007),regulatestheexpressionofAGAMOUS-LIKE16thatplaysaroleincontrollingstomataldensityanddevelopmentintheleaves(Kutteretal.,2007).ManynonconservedmiRNAs,however,havenopredictableorvalidatedtargetgenes.AcurrentmodelpostulatesthatmanynewmiRNAsevolvedtransientlyand,withoutgainingbiologicalfunctions,diedrapidlythroughmutationaldriftsduringthecourseofevolution(Fahlgrenetal.,2007,2010;Maetal.,2010;reviewedbyAxtell,2008;AxtellandBowman,2008;Cuperusetal.,2011).ThebiologicalroleofthevastmajorityofnonconservedmiRNAsremainsunknown,evenformanywhosetargetshavebeenpredictedorvalidated.

WereasonedthatexperimentalanddeepphylogeneticanalysesofamiRNAanditstargetgeneinagroupoforganismswithclearphylogeneticrelationshipscouldprovideuniqueinsightsintothebiologicalfunctionandevolutionpatternofmiRNA-regulatedgeneexpression.Inthisstudy,weinvestigatedmiR4376inthefamilySolanaceae.Thisfamilyismonophyleticwithawell-understoodphylogenytree(Olmsteadetal.,2008)andincludessomeofthemostimportantfood,medicinal,andornamentalplants.Here,wepresentevidencethatthe22-nucleotidemiR4376regulatestheexpressionofanautoinhibitedCa2+-ATPase,ACA10,whichplaysaroleintomatoreproductivegrowth.DeepphylogeneticmappingrevealedtheevolutioncourseofMIR4376lociinSolanaceaeandshowedposttranscriptionalprocessingofpre-miR4376asalikelylimitingstepduringtheevolutionofmiR4376insomeplantspecies.Themappingfurthersuggestedindependentphylogeneticoriginandpossibleposttranscriptionalmodi?cationofthemiR4376targetsiteinACA10homologs.Finally,miR4376triggersformationof21-nucleotidephasedsmallinterferingRNAs(siRNAs).Together,ourstudyrevealsanexampleofmiRNA-regulatedexpressionofuni-versallyimportantCa2+-ATPasesandanovellayerofmolecularmechanismunderlyingtomatoreproductivegrowth.Itfurthersug-gestsposttranscriptionalprocessingasanimportantcontrolpointfortheevolutionaryemergenceofamaturemiRNAandfortheeliminationofitstargetsiteinsomehomologousmRNAsfromdifferentplantspecies.Thus,thestochasticemergenceofamiRNA-targetgenecombinationinvolvesmultiplemoleculareventsatthegenomic,transcriptional,andposttranscriptionallevelsthatmayvarydrasticallyincloselyrelatedplantspecies.RESULTS

Identi?cationofsly-miR4376

Wepreviouslyidenti?eda22-nucleotidetomatosmallRNA,namedSlsmR-31a(forS.lycopersicumsmallRNA-31a),by

conventionalsequencing(Itayaetal.,2008).ThesamesmallRNAalsowaslateridenti?edintomatoviadeepsequencingbyMoxonetal.(2008)andbyothers(http://smallrna.udel.edu/libraries.php)andsummarized(http://ted.bti.cornell.edu/cgi-bin/TFGD/sRNA/sRNA.cgi?sRNA_ID=S0000125).ABLASTsearchidenti?edthecorrespondingDNAsequenceinoneclone(SL1.00ct22836)fromchromosome6ofthecompletedtomatogenomesequencedraft(SOLgenomicsnetwork;http://sgn.cornell.edu,databaseofTomatoWGSContigs[SL1.00]),impli-catingitasasingle-copygene.AsshowninFigure1A,sequencealignmentrevealsthatthissmallRNAisahomologofGma-miR4376fromG.max(Joshietal.,2010).ThissmallRNAalsomeetsseveralcriteriaformiRNAdiscussedbelow;therefore,itisrenamedSly-miR4376inthisreport.First,RNAfoldingofaninsilicotranscribed91-nucleotidesequencecontainingthesly-miR4376sequencewithmfold(Zuker,2003)yieldedahairpin(Figure1B)?ttingthestructuralcriteriaofamiRNAprecursor(Meyersetal.,2008).Furthermore,bothsly-miR4376andgma-miR4376arederivedfromthesamearmsoftheirrespectiveprecursors.Incomparisonwithsly-miR4376,gma-miR4376hasa1-nucleotideshifttowardthe59endofitsprecursor,givingrisetotwogapsatboth59-and39-endsandtwoadditionalmis-matchesatpositions16and22(Figure1A).Second,wefoundthemiR4376-codingsequenceinthepartialgenomesequenceofpotato(Solanumtuberosum)(gb|AC232436.2|),acloserela-tiveoftomato.TheinsilicotranscribedpotatosequencealsofoldedintoacharacteristicplantmiRNAprecursorstructure(Figure1C),suggestingthatmiR4376isconservedamongcloserelatives.Basedontheavailablegenomesequencedata,themiR4376-codingsequenceintomato(http://wendang.chazidian.com/genomes/Solanum_lycopersicum/genome_data.pl)andpotato(http://solanaceae.plantbiology.msu.edu)lieswithinaninter-genicregion.Finally,themostrecentdeepsequencingdata-base(http://ted.bti.cornell.edu/)listssequencesmatchingthemiR4376*(tomatosmallRNAS0086007)(Figure1B)aswellasmiR4376(tomatoputativemiRNAM00690).

WepreviouslyreportedavariantofmiR4376,differingbyanadditionalnucleotideatthe59-endofmiR4376(Itayaetal.,2008).ThesevariantsappearedtobederivedposttranscriptionallyfromthesameMIR4376locuswithextensionofanucleotideatthe59-endofmiR4376.ThisimpreciseprocessingisconsistentwithmanynonconservedmiRNAsfoundinA.thalianaandA.lyrata(Maetal.,2010).Furthermore,thesinglecopygeneformiR4376alsoisconsistentwiththerecentevolutionofnonconservedmiRNAs(Fahlgrenetal.,2010;Maetal.,2010).

RNAgelblotsshowedhighaccumulationofmiR4376inleavesandyoung?owerbudsbutdiminishedaccumulationinyounggreenfruitsandnearabsencefrommaturefruits(Figure1D).DetailedanalysisshoweddevelopmentalregulationofmiR4376accumulationinleaves(seeSupplementalFigure1online).CloningandIdenti?cationofthemiR4376TargetGeneInitialcomputationalpredictionofthemiR4376targetwasnotsuccessfulduetolackofthecompletetomatogenomicse-quencewhenthisRNAwas?rstidenti?ed.Therefore,weclonedthemiR4376targetgenebyconventionalmolecularapproaches(seeMethods).Brie?y,we?rstclonedapartialsequenceofthe

MicroRNAEvolutionandFunction3187

Figure1.PrecursorStructures,ExpressionPattern,andTargetValidationformiR4376.

(A)AlignmentofDNAsequencesencodingGma-miR4376andSly-miR4376fromG.maxandS.lycopersicum,respectively.TheMIR4376sequencesarehighlightedinbold,whereastheMIR4376*sequencesarehighlightedinlightgray.Thestarsindicatetheidenticalnucleotidesbetweenthetwospecies.TheGma-MIR4376–containingsequencewasretrievedfromPhytozyme(http://wendang.chazidian.com/)usingtheGma-miR4376precursorsequencefrommiRBase(http://wendang.chazidian.com/)asbait.TheCreateAlignmentfunctionintheCLCFreeWorkbenchSoftware(CLCBio)wasusedtoalignGma-MIR4376andSly-MIR4376,withminormanualadjustment.

(B)and(C)Secondarystructuresofpre-miR4376fromtomatoandpotato,respectively.ThemiR4376andmiR4376*sequencesarehighlightedinboldandlightgray,respectively.

(D)ExpressionofmiR4376anditstargetgeneACA10indifferenttomatoorgans.ThemiR4376accumulationlevelsindifferentorganswereanalyzedbyRNAgelblotsandareexpressedrelativetothatintheleaf(arbitrarilysetto1.0).RA,ratioofaccumulationlevelsobtainedfromtwobiologicalreplicates.TheexpressionofACA10mRNAwasanalyzedbyRT-PCR,withampli?cationofActinmRNAasacontrol.Threebiologicalreplicateswereperformed.L,leaf;B,?owerbud;YF,youngfruit;MF,maturefruit.

(E)Invivocleavagesiteinthe59-UTRofSlACA10mRNAasmappedbymodi?ed59-RACE.Thearrowindicatesthecleavagesiteidenti?edin11outof11clonessequenced.

(F)ComplementationofCa2+-ATPaseknockoutyeaststrainK616withtransfectionbySlACA10R,whichisamutantversionofSlACA10thatcontainsamutationrenderingitresistanttosilencing(seetextfordetailsofthisconstruct).

miR4376targetgeneby39-rapidampli?cationofcDNAends(RACE).Wedesignedseveral39RACEforwardprimersbasedonthemiR4376sequence,withvarioustruncationsattheir39endstocircumventmismatch(s)betweenthemiR4376andtargetsequence.Amongtheforwardprimerstested,aprimerwithonenucleotidedeletionatthe39-endallowedcloningofapartialcDNA.BasedonthesequenceofthispartialcDNA,newprimersweredesignedforsubsequent39-and59-RACEtoclonethefull-lengthcDNA.TheclonedcDNAcontains3246nucleotidesandshares86%similarityattheaminoacidlevelwithA.thalianaACA10,whichencodesanautoinhibitedCa2+-ATPase(seeSup-plementalFigure2online).Thesequenceofthisgenewascon-?rmedintherecenttomatogenomesequence(http://wendang.chazidian.com).WedesignatethetomatogeneasSlACA10.

InsubsequentsearchesforadditionalcandidatetargetsofmiR4376usingonlinepredictionsoftware(http://wendang.chazidian.com/miRNA/miRU.htm)fromitsdefaulttomatocDNAlibrary,we

obtainedafragment(BI932149)belongingtoageneencodingaTCPfamilytranscriptionfactor.However,wewereunabletovalidatethistargetgeneexperimentally(seebelow).ExperimentalValidationofthemiR4376TargetGeneBasedontheopenreadingframepredictionbyCLCFreeWorkbenchsoftware(CLCBio),thetargetsequenceofmiR4376residesinthe59-untranslatedregion(UTR)ofbothACA10mRNAandtheTCPhomolog(http://wendang.chazidian.comingmodi?edRNAligase-mediated59-RACE(Llaveetal.,2002;Kasschauetal.,2003),wemappedthecleavagesiteinACA10atapositionoppositetothe9thnucleotidefromthe59-endofmiR4376(Figure1E).ThesedataestablishthatmiR4376isafunctionalmiRNAthatregulatesthestabilityofACA10mRNAinvivo.

TotestwhethertheclonedACA10encodesafunctionalCa2+-ATPase,weperformedacomplementationtestby

transforming

3188ThePlantCell

theyeast(Saccharomycescerevisiae)triplemutant(pm1pmc1cnb1)K616thatlacksallCa2+-ATPasesandcalcineurinsubunitB(CunninghamandFink,http://wendang.chazidian.complementationofK616growthhasbeenwidelyusedtodemonstratetheCa2+-trans-portingfunctionsofvariousplantCa2+-ATPases(Geisleretal.,2000;Szeetal.,2000).OurresultsdemonstratedthatthetomatogenecomplementedthegrowthofK616,indicatingthatitindeedencodesaCa2+-ATPase(Figure1F).

miR4376TriggeredtheFormationofPhasedsiRNAsfromACA10Transcripts

Recentstudiesshowedthat22-nucleotidemiRNA-mediatedcleavageoftargettranscriptscouldleadtotheproductionofphased21-nucleotidesiRNAs,dependingontheactivitiesofRNA-dependentRNApolymerase6andDCL4inA.thaliana(Chenetal.,2010;Cuperusetal.,2010).SuchphasedsiRNAsmayfunctionastrans-actingsiRNAstoregulatethestabilityoftheirtargettranscripts(Allenetal.,2005;Yoshikawaetal.,2005).GiventhatmiR4376is22nucleotideslong,wesearchedthedatabasecontainingdeepsequencingdataofsmallRNAsfromtomatoandotherplantspecies(http://smallrna.udel.edu/)forthepresenceofphasedsiRNAsfromdifferentACA10homologs.Oursearchusingthefull-lengthSlACA10cDNAsequenceasbaituncoveredmanysmallRNAs(sRNAs)matchingvariousregionsofthecDNAfromtheleaf,?ower,andfruitinboththesenseandantisenseorientations(seeSupplementalDataSet1online).Wethenmanually?lteredtherawdatatoidentifythosesRNAspresentin21-nucleotideregistersimmediatelydownstreamofthemiR4376-mediatedcleavagesite.AsshowninFigure2,weretrievedaseriesofphasedsRNAsfromthepooloftotalSlACA10-matchingsRNAs.ThesensestrandsRNAsaredesig-nated39P1+,39P2+,39P3+,etc.,startingdownstreamofthe39-endofSlACA10cleavagesite(Figure2A).Theminussignfollowingadesigneeindicatesantisensepolarity.WerefertothemasphasedsiRNAsinthisreport,consideringtheirsimilarpatternwith21-nucleotidephasedsiRNAsinA.thaliana.

Notably,theabundanceofphasedsiRNAswashighestintheleaf,withadecreaseinthe?owerandtheninthefruit.Thistrendisdirectlycorrelatedwiththehigh-to-lowaccumulationpatternofmiR4376intheseorgans(Figure1D).

Inanattemptto?ndthepossibletargetsofthetwomostabundantphasedsiRNAsintomato,39P2+and39P5+,weperformedcomputationalpredictionfollowingthecriteriaestab-lishedbyAllenetal.(2005).WefoundseveralputativecandidatesasshowninSupplementalDataSet2online,withonehavingthebest(butnotperfect)matchandtheothershavinglessconvinc-ingmatches.Noneoftheseputativetargetshaveknownorvalidatedfunctions.Whetheranyofthesearebona?detargetsandwhether39P2+and39P5+indeedregulatetheexpressionofanyofthesegenestocontrolreproductivegrowthorotheraspectsofplantgrowthremainstobeinvestigated.

WewereunabletovalidateBI932149byrepeated59-RACE.Consistentwiththisobservation,wedidnot?ndanyphasedsiRNAs,butonlyone18-nucleotidesRNA(tworeads)matchingtheBI932149sequenceintheabovesRNAdatabase.Therefore,BI932149maynotbeabona?detargetgeneformiR4376.However,wecannotruleoutthepossibilitythatmiR4376medi-

atesalowlevelofcleavageofthistargetRNAwithoutproducingphasedsiRNAsorregulatesitsexpressionbyamechanismotherthancleavage.Wealsocannotruleoutthepossibilitythatunknowntechnicalreasonsaccountedforourfailuretoclonethecleavageproducts.

DisruptionofmiR4376-RegulatedSlACA10ExpressioninTransgenicTomatoSpeci?callyAlteredFlowerMorphologyandFruitYield

ToexperimentallytesttheroleofmiR4376-regulatedACA10ex-pressionintomato,wegeneratedtwotypesoftransgenictomatoplants,inwhichsuchregulationisperturbed.Onetypeoverex-pressesmiR4376(miR4376-OX)andtheotheroverexpressesamiR4376-resistantSlACA10mutant(SlACA10R-OX)withaFLAGtag,bothunderthecontrolofthecauli?owermosaicvirus(CaMV)35Spromoter.ToengineerACA10R,wealteredthemiR4376targetsequence59-TCTGGTGTCATCTCTCCTGCGA-39to59-TCTGG-TGTCATCTATCCTGCGA-39,inwhichthe“A”substitutionfortheunderlined“C”waspredictedtointerferewithrecognitionbymiR4376,therebyreducingtheef?ciencyofcleavage,basedon?ndingsfromstudiesonothermiRNAs(Malloryetal.,2004).Transgenictomatoplantsoverexpressinggreen?uorescentprotein(GFP)fromthesamevectors(GFP-OX)servedascontrols.

WeobtainedsixindependentmiR4376-OXlines,?veSlACA10R-OXlines,and?veGFP-OXlines,allveri?edtoexpressthetrans-genes.Asanexample,showninFigure3A,overexpressionofmiR4376andACA10Rwasveri?edbyRNAgelblotsandRT-PCR(inconjunctionwithsequencing),respectively.GFPexpressionintheGFP-OXlineswasveri?edbyimmunoblots(seeSupplementalFigure3Aonline).IntheGFP-OXcontrolplants,miR4376wasdetectedinsepalsandpetals,butnotinstamenorpistils.ThiswasinverselycorrelatedwiththeexpressionpatternofACA10mRNAinthese?oralparts.InmiR4376-OXplants,elevatedlevelsofmiR4376wereobservedinsepals,petals,thestamen,andpistils,whichwascorrelatedwithsigni?cantlyreducedaccumulationofACA10mRNAtonearlyundetectablelevels.InSlACA10R-OXplants,endogenousACA10hadasimilarexpressionpatternasinGFP-OXplants.ExpressionofACA10Rwashigherinsepals,petals,andstamenscomparedwiththatofendogenousACA10,asexpectedofreducedef?ciencyofmiR4376-mediatedcleavage.Expressionofthismu-tantwasadditionallyveri?edbyimmunoblotsusinganti-FLAGantibodies(seeSupplementalFigure3Bonline).TheproteinbandsizesuggeststhatthedimericformofACA10waspresent.WhetherthisisthefunctionalformofACA10remainstobedetermined;however,previousstudieshavedemonstratedthatCa2+-ATPasesexistandfunctionashomodimers(BasuandBriskin,1995;SackettandKosk-Kosicka,1996;UshimaruandFukushima,2008).

NeitherthemiR4376-OXnortheSlACA10R-OXplantsexhib-itedobviousdifferencesinseedgermination,height,orleafmorphologyincomparisonwiththeGFP-OXlines(seeSupple-mentalFigure4online).The?rstnotableandconsistentpheno-typeofmiR4376-OXandSlACA10R-OXplantswasanincreasedverticaldistancebetweentheantherandstigmacomparedwithGFP-OXplants(Figure3B).Theother?oralorgansappearednottobeaffected.Quantitativeanalysisshowedthattheincreasedanther-stigmadistancewasattributedtoelongatedstamen?laments(seeSupplementalFigure5online).

MicroRNAEvolutionandFunction3189

Figure2.PhasedsiRNAsDerivedfromACA10inTomatoBasedonSearchesinthesRNADeepSequencingDatabase(http://smallrna.udel.edu/).(A)Phasedpatternof21-nucleotidesiRNAsalongtheSlACA10sequence.ThesensestrandsofphasedsiRNAsaredesignated39P1+,39P2+,39P3+,etc.,startingfromthevalidatedmiR4376-cleavagesitethatismarkedbyanarrowhead.TwoantisensesiRNAs(39P6-and39P11-)areshownabovetheACA10sequence(indicatedbyredlines).Thenucleotidepositionsareindicatedabovethesequence,withthenucleotideimmediatelydownstreamofthecleavagesitedesignatedas“1”(notshownforimageclarity).GreenindicatessiRNAsfoundintheabovedeepsequencingdatabase.TherestofpredicatedphasedsiRNAswerenotfoundforanyorgansinthedatabase.ThereadnumberofeachsiRNAintheleaf/?ower/fruitispresentedinparenthesis.TheIDofeachsiRNA(orsRNA)intheTomatoFunctionalGenomicsDatabase(http://ted.bti.cornell.edu/)isgiven.

(B)DistributionofACA10-matchingsRNAsofvariouslengthsintheleaf,?ower,andfruitoftomato.Thenumbersinthe“21nttotal”columnrepresentthetotalreadsforeachofthetomatoorgans,whereasthoseinthe“21ntphased”columnrepresentsubsetsfromthe21-nucleotidetotalreads.

Themostsigni?cantimpactofmiR4376orSlACA10Rover-expressionwasdrasticallyreducedyieldofmaturefruits(Figure3C).ThesixmiR4376-OXlineseachproducedasimilarnumberof?owersasthe?veGFP-OXcontrollines(Table1).However,theyproducedzerotosixmaturefruits,incontrastwiththeGFP-OXcontrollinesthatproduced20to42maturefruits(exceptforonelinethatproducedsix)(Table1).SlACA10R-OXplantsalsohadsimilarnumberof?owersastheGFP-OXplantsbutreducednumberofmaturefruitslikemiR4376-OXplants(Table1).Thematurefruitsfromthetransgeniclinesdidnotexhibitdifferencesinsizeormorphology.CloseexaminationshowedmanysmallgreenfruitsthatfailedtogrowintomaturefruitsinbothmiR4376-OXandSlACA10R-OXtransgenicplants(seeSupplementalFigure6online).Forexample,weobserved36and40suchsmallfruitsintwomiR4376-OXplantsand28and34suchfruitsintwoSlACA10R-OXplants.

ItwassurprisingthatmiR4376-OXandSlACA10R-OXplantsexhibitedsimilar,ratherthanopposite,phenotypes.AsimpleexplanationisthataparticularlevelofACA10expressionintomato?owersandfruitsiscriticalforreproductivegrowthandthatalteredexpressiontolevelslowerorhigherthantheoptimalwouldcausesimilarphenotypes.Therearepriorexampleswhereup-anddownregulatedexpressionofagenecanleadtoasimilarphenotype.TheseincludethegenesencodingBRI1-ASSOCI-ATEDRECEPTORKINASE1(BAK1)inO.sativa(Wangetal.,2007a;Lietal.,2009),N-METHYL-D-ASPARTASE(NMDA),asubtypeofGlureceptors,inanimals(reviewedinHardinghamandBading,2003),andCALCINEURIN(CN)inCaenorhabditiselegans(Ahnetal.,2006).BAK1isinvolvedinbrassinosteroidsignaling,andplantsexhibitedasimilardwarfphenotypewhenBAK1wasupregulatedordownregulated.ItisnoteworthythatNMDAreceptorsandCNareprobablyinvolvedinCa2+signaling.Up-ordownregulationofNMDAreceptorsistoxictoneurons,andup-ordownregulationofCNinC.elegansledtosimilarslowgrowthandareductioninreproduction.

ItisimportanttonotethatthetransgenicexpressionofmiR159underthecontroloftheCaMV35Spromoterintomatoledtopleiotropicphenotypes,includingdevelopmentaldefectsinleaves,?owers,andfruits(Buxdorfetal.,2010).Bycontrast,therewerenoobservablephenotypesforvegetativegrowthin

our