elucidation of the genetic architecture of self

HAL Id: hal-01525802https://hal.archives-ouvertes.fr/hal-01525802

Submitted on 26 May 2020

HAL is a multi-disciplinary open accessarchive for the deposit and dissemination of sci-entific research documents, whether they are pub-lished or not. The documents may come fromteaching and research institutions in France orabroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, estdestinée au dépôt et à la diffusion de documentsscientifiques de niveau recherche, publiés ou non,émanant des établissements d’enseignement et derecherche français ou étrangers, des laboratoirespublics ou privés.

Distributed under a Creative Commons Attribution| 4.0 International License

Elucidation of the genetic architecture ofself-incompatibility in olive: evolutionary consequences

and perspectives for orchard managementPierre Saumitou-Laprade, Vernet Philippe, Xavier Vekemans, Sylvain Billiard,

Gallina Sophie, Essalouh Laila, Mhaïs Ali, Moukhli Abdelmajid, El BakkaliAhmed, Barcaccia Gianni, et al.

To cite this version:Pierre Saumitou-Laprade, Vernet Philippe, Xavier Vekemans, Sylvain Billiard, Gallina Sophie, et al..Elucidation of the genetic architecture of self-incompatibility in olive: evolutionary consequences andperspectives for orchard management. Evolutionary Applications, Blackwell, 2017, 10 (9), pp.867-880.�10.1111/eva.12457�. �hal-01525802�

https://hal.archives-ouvertes.fr/hal-01525802

http://creativecommons.org/licenses/by/4.0/


https://hal.archives-ouvertes.fr

Evolutionary Applications 2017; 1–14 | 1wileyonlinelibrary.com/journal/eva

Received:4October2016 | Accepted:5January2017DOI:10.1111/eva.12457

O R I G I N A L A R T I C L E

Elucidation of the genetic architecture of self- incompatibility in olive: Evolutionary consequences and perspectives for orchard management

Pierre Saumitou-Laprade1,a | Philippe Vernet1,a | Xavier Vekemans1 | Sylvain Billiard1 | Sophie Gallina1 | Laila Essalouh2 | Ali Mhaïs2,3,4 | Abdelmajid Moukhli3 | Ahmed El Bakkali5 | Gianni Barcaccia6 | Fiammetta Alagna7,8 | Roberto Mariotti8 | Nicolò G. M. Cultrera8 | Saverio Pandolfi8 | Martina Rossi8 | Bouchaïb Khadari2,9,a | Luciana Baldoni8,a

1CNRS,UMR8198Evo-Eco-Paleo,UniversitédeLille-SciencesetTechnologies,Villeneuved’Ascq,France2MontpellierSupAgro,UMR1334AGAP,Montpellier,France3INRA,URAméliorationdesPlantes,Marrakech,Morocco4LaboratoireAgroBiotechL02B005,FacultédesSciencesetTechniquesGuéliz,UniversityCadiAyyad,Marrakech,Morocco5INRA,URAméliorationdesPlantesetConservationdesRessourcesPhytogénétiques,Meknès,Morocco6LaboratoryofGenomicsandPlantBreeding,DAFNAE-UniversityofPadova,Legnaro,PD,Italy7ResearchUnitforTableGrapesandWineGrowinginMediterraneanEnvironment,CREA,Turi,BA,Italy8CNR,InstituteofBiosciencesandBioresources,Perugia,Italy9INRA/CBNMed,UMR1334AméliorationGénétiqueetAdaptationdesPlantes(AGAP),Montpellier,France

ThisisanopenaccessarticleunderthetermsoftheCreativeCommonsAttributionLicense,whichpermitsuse,distributionandreproductioninanymedium,providedtheoriginalworkisproperlycited.©2017TheAuthors.Evolutionary ApplicationspublishedbyJohnWiley&SonsLtd

aThesefourauthorscontributedequallytothiswork.

CorrespondencePierreSaumitou-Laprade,CNRS,UMR8198Evo-Eco-Paleo, Université de Lille - Sciences etTechnologies,Villeneuved’Ascq,France.Email:[email protected]

Funding informationThisresearchwassupportedbytheProjectOLEA–GenomicsandBreedingofOlive,fundedbyMiPAAF,Italy,D.M.27011/7643/10,bytheProject“BeFOre-BioresourcesforOliviculture,”2015–2019,H2020-MSCA-RISE-MarieSkłodowska-CurieResearchandInnovationStaffExchange,GrantAgreementN.645595,bytheFrenchNationalResearchAgencythroughtheproject“TRANS”(ANR-11-BSV7-013-03),byFrench-MoroccanscientificcooperationPRAD14-03,andbyAgropolisFondation“OliveMed”N°1202-066throughthe“Investissementsd’avenir”/LabexAgroANR-10-Labex-0001-01.,managedbyFrenchNationalResearchAgency.

AbstractTheolive(Olea europaeaL.)isatypicalimportantperennialcropspeciesforwhichthegenetic determination and even functionality of self-incompatibility (SI) are stilllargelyunresolved.ItisstillnotknownwhetherSIisundergametophyticorsporo-phyticgeneticcontrol,yetfruitproductioninorchardsdependscriticallyonsuccess-fulovulefertilization.WestudiedthegeneticdeterminationofSIinoliveinlightofrecentdiscoveriesinothergeneraoftheOleaceaefamily.Usingintra-andinterspe-cificstigmatestson89genotypesrepresentativeofspecies-wideolivediversityandthecompatibility/incompatibilityreactionsofprogenyplantsfromcontrolledcrosses,we confirmed that O. europaea shares the same homomorphic diallelic self-incompatibility(DSI)systemastheonerecentlyidentifiedinPhillyrea angustifolia and Fraxinus ornus.SIissporophyticinolive.TheincompatibilityresponsediffersbetweenthetwoSIgroupsintermsofhowfarpollentubesgrowbeforegrowthisarrestedwithin stigmatissues.As a consequence of thisDSI system, the chance of cross-incompatibility between pairs of varieties in an orchard is high (50%) and fruit

http://orcid.org/0000-0001-7413-3090


mailto:[email protected]

2 | SAUMITA- SAPSADE DEI Sl

1 | INTRODUCTION

Self-incompatibility(SI),apostpollinationprezygoticmechanismpre-venting self-fertilization in simultaneous hermaphroditic individuals,isacommonfeatureinfloweringplants,occurringinaround40%ofangiospermspecies (Igic,Lande,&Kohn,2008).Geneticdetermina-tionof SI is highlyvariable,with a single locus or several loci (dial-lelicormulti-allelic)andgametophyticorsporophyticcontrolof thepollen SI phenotype (Castric & Vekemans, 2004; De Nettancourt,1977).Becausedistinct individualscanshare identicalSIgenotypes,incompatible crosses are not limited to self-pollination (Bateman,1952).Bylimitingcompatiblematings,SIcancauseadirectdecreasein seed production and can be an important demographic factor,a phenomenon known as the S-Allee effect (Leducq etal., 2010;Wagenius, Lonsdorf, & Neuhauser, 2007). This effect is especiallyimportantinpopulationswithlowgeneticdiversity(Byers&Meagher,1992;Vekemans,Schierup,&Christiansen,1998).Despite itswide-spreadoccurrence in angiosperms, thegeneticbasisofSIhasbeenidentifiedinalimitednumberofcases,andtheunderlyingmolecularmechanismhasbeenshowninonlyahandfulofplantfamilies.Theseinclude theBrassicaceae (Kitashiba&Nasrallah,2014;Tantikanjana,Rizvi,Nasrallah,&Nasrallah,2009),Papaveraceae(Eavesetal.,2014),Solanaceae,Plantaginaceae,andRosaceae(Iwano&Takayama,2012;Sijacicetal.,2004;Williams,Wu,Li,Sun,&Kao,2015).

AmongplantspeciespossessingafunctionalSIsystem,cropspe-ciesareofparticularimportancebecausetheSIsystemcaninterferewith plant production and breeding, representing a major obstaclefor constant high yield (Sassa, 2016).A reduction in genetic diver-sity incommercialvarietiesmayalsopotentially limit seedand fruitproductioninfieldconditions,withadverseeconomicconsequences(Matsumoto, 2014). This issue has stimulated active crossing pro-gramstoassessallelicdiversityattheSIlocusincropspeciesshow-ing functional SI, such as apple (Broothaerts, 2003), Japanese pear,sweetcherry,apricot(Sassa,2016;Wünsch&Hormaza,2004),cab-bage(Ockendon,1974),chicory(Gonthieretal.,2013),andsugarbeet(Larsen1977).However,despitetheobviousinterestforbreederstouse the SI system to their advantage as part of their breeding pro-grams, proper understanding of the genetic factors and molecularmechanisms involved inSI is lacking formost speciesandgenerallytechnicallydifficultforbreedingcompanies.

The mechanisms controlling SI are often conserved and sharedamong species belonging to a given plant family (Allen & Hiscock,

2008;Charlesworth,1985;Weller,Donoghue,&Charlesworth,1995).Hence,evolutionaryapproachescanhelptouncoverSImechanismsincropspeciesbasedonknowledgeof relatedspecies.AlthoughSIhasevolvedindependentlymanytimeswithinangiosperms,therateofevolutionofnewSIsystemsisthoughttobelow,andtheoccurrenceofdistinctmechanismsofSIgeneticdeterminationwithinagivenfam-ilyshouldberare(Igicetal.,2008).

Theolive(Olea europaeasubsp.europaea)istheiconictreeoftheMediterraneanarea,presentincultivated(var.europaea)andwild(var.sylvestris) forms (Green, 2002). Despite the economical, ecological,cultural, and social importance of this species, itsmating system isstilllargelycontroversialandnoconsensusmodelhasbeenaccepted.ThegeneticdeterminationandevenfunctionalityofSIarestilllargelycontroversial.Inthisspecies,eventhemostbasicbiologicaldetailsofSIareunresolvedandwedonotknowwhetherSI isundergameto-phytic (Ateyyeh, Stosser, &Qrunfleh, 2000) or sporophytic (Breton&Bervillé,2012;Collanietal.,2012)geneticcontrol.ThenumberofgenesinvolvedintheoliveSIsystemandtheirpatternoflinkageandchromosomallocationarealsounknown.

Cultivars able to produce seed by selfing are thought to exist(Farinelli, Breton, Famiani, &Bervillé, 2015;Wu,Collins,& Sedgley,2002),butthiscontentionisrarelysupportedbymolecularpaternitytests(DelaRosa,James,&Tobutt,2004;Díaz,Martın,Rallo,Barranco,&DelaRosa,2006;Díaz,Martín,Rallo,&DelaRosa,2007;Mookerjee,Guerin,Collins,Ford,&Sedgley,2005;Seifi,Guerin,Kaiser,&Sedgley,2012).Inarecentstudyinvolvingpaternityassessmentofseedprog-enies from the Koroneiki cultivar (Marchese, Marra, Costa, etal.2016), itwas shown that seeds produced on twigs protected fromoutcrosspollenwerederivedfromself-fertilization.Inaddition,openpollination resulted in11%of the seedsbeingproducedvia selfing.Together,theseresultssuggestthatSIisindeedfunctionalinthisvari-ety,althoughleaky.PollinationexperimentshavebeenperformedinseveralstudiestocharacterizeSIatthephenotypiclevelandidentifygroupsof compatibility amongvarieties.These studies scored com-patibility either at the prezygotic stage, by cytological observationsofpollentubeelongation instigmas,orat thepostzygoticstage,bymeasuring seed production (supported or not by paternity assess-ment) after application of pollen from donor plants (Breton etal.,2014;Cuevas&Polito,1997).ContradictoryconclusionsweredrawnintermsofclassifyingcultivarsintoSIgroups,aswellasintermsofthequantitative strengthof the incompatibility reaction.Someof thesediscrepanciesmaybecausedbypollencontamination,likelybecause

productionmaybelimitedbytheavailabilityofcompatiblepollen.Thediscoveryofthe DSI system in O. europaeawillundoubtedlyofferopportunitiestooptimizefruitproduction.

K E Y W O R D S

diallelicself-incompatibilitysystem,homomorphicsystem,Olea europaea L., Oleaceae, olive diversity,plantmatingsystems,sporophyticgeneticcontrol,trans-genericconservationofSIfunctionality

| 3SAUMITA- SAPSADE DEI Sl

O. europaeaproducesalargequantityofpollentypicalofmostwind-pollinatedspecies(Ferrara,Camposeo,Palasciano,&Godini,2007).

Here,westudiedthegeneticdeterminationofSI inoliveinlightof recent discoveries in other genera in the same family, Oleaceae. SIhasbeeninvestigatedinPhillyrea angustifolia L. and Fraxinus ornus L, two androdioecious species in which males and hermaphroditesco-occurinthesamepopulation.Bothspeciesshareahomomorphicsporophyticdiallelic SI (DSI) system (Saumitou-Lapradeetal., 2010;Vernet etal., 2016). Self-incompatible hermaphroditic individualsbelongtooneoftwohomomorphicSIgroups:IndividualsofagivenSIgroupcanonlysireseedsonhermaphroditesfromtheothergroup,and cross-pollination between individuals of the same group elicitsanincompatibilityresponse(Saumitou-Lapradeetal.,2010).TheDSIsystemhasbeenconserved inbothspecies,andcross-speciespolli-nationtestshavedemonstratedthattherecognitionspecificitiescur-rentlysegregatinginthetwospeciesareidentical.P. angustifolia and F. ornusbelong to twodifferent subtribeswithin theOleaceae (sub-tribe Oleinae for P. angustifolia and subtribe Fraxininae forF. ornus).Hence, it has been suggested that this DSI system originated ances-trallywithinOleaceae(Vernetetal.,2016)andthuswaspresentinthemost recent common ancestor of these two subtribes about 40 million yearsago(Mya)(Besnard,deCasas,Christin,&Vargas,2009).BecauseO. europaea and P. angustifoliabelongtothesamesubtribe(Oleinae),wehypothesizethattheysharethesameSIsystem.

Weappliedexperimentalapproachesdeveloped forP. angustifo-lia and F. ornus (Saumitou-Laprade etal., 2010;Vernet etal., 2016)tocharacterizetheSIsysteminO. europaea,bothphenotypicallyandgenetically.First,weperformedcontrolledpollinationsinafulldiallelcrossingschemebetweenhermaphroditic individualsusedaspollenrecipients and pollen donors (including self-pollination). The objec-tivewastocomparethepatternoftheincompatibilityreactionswiththose described in P. angustifolia and F. ornus. Second,we analyzedthepatternof segregationof incompatibilityphenotypesamong91offspringfromonesingleintervarietalcross(DelaRosaetal.,2003).Theresultswere inagreementwithageneticmodelconsistingofaDSI systemwith twomutually incompatible groups of hermaphro-dites.Thevalidityofthisgeneticmodelwasassessedbyperformingcontrolledpollinationswith89genotypesrepresentativeofasignifi-cantportionoftheolivediversitypresent inaworldwidecollectionagainsttwopairsoftestergenotypes,eachpairbeingcomposedoftworeciprocallycompatiblehermaphroditesphenotyped inthefirstdiallel crossingexperiment andeachassigned tooneof the twoSIgroups. Using pollen from hermaphroditic individuals of P. angus-tifolia and F. ornus, we also demonstrated that the same two allelic specificitiesaresharedamongthethreegenera, therebyconfirmingourhypothesisthattheysharethesameDSIsystem.TheresultsarepresentedinlightofpreviousattemptstocharacterizetheSIsystemin olive. It isworthmentioning that this study focused on the cul-tivated formofolive (var.europaea), analyzingvarieties representa-tiveofdomesticatedMediterraneanolivediversity (ElBakkalietal.,2013;Haouaneetal.,2011).Ourresultssuggestnewavenuesforthedevelopmentofoliveorchardmanagementpracticestooptimizefruitproduction.

2 | MATERIALS AND METHODS

2.1 | Plant material

To avoid anymisclassification of varietal clones and to allow theirauthentication by means of voucher samples, DNA was extractedfromeach individual treephenotyped forSIandwasgenotypedbyassaying 15 different polymorphic microsatellite (SSR) marker loci.Hence,weidentifiedeachindividualtreewithareferenceDNAsam-ple code, a physical position in the orchard, a genotype referencenumbercorrespondingtoaspecificmarkerallelecombinationforthe15SSRmarkerloci(TableS1),andanSIphenotype.

In2013,sixgenotypeswerechoseninItalianorchardsandtestedforcross-compatibilityusingstigmatestsinareciprocaldialleldesign(Table1).Becausetestershadtobeusedaspollenrecipientsinfuturetests, the six genotypeswere chosen among those represented byseveraltreesintheexperimentalstations,atdifferentsitesunderdif-ferent agroecological conditions favoring different flowering timesforasinglegenotype,and locatedascloseaspossibleto laboratoryfacilitiestoensurequicktransferofreceptiveflowerstothe labora-toryoverthewholestudyperiod.Fromthesesixgenotypes,fourwereselectedtoconstitutethetwopairsoftestersusedforscreeningvari-etiesfromtheolivecollections.ReceptiveflowerssampledfromthefourchosentestergenotypeswereusedtophenotypeSIin2013and2014. For phenotyping, 118 trees, corresponding to 89 genotypes,wereselectedfromdifferentex situcollections.Inparticular,64treeswerekeptfromtheworldwideOliveWorldGermplasmBank(OWGB)ofINRAMarrakech,attheexperimentalorchard(Tessaout,Morocco),45 fromthePerugiacollection ((43°04′54.4″N;12°22′56.8E), Italy),and three from theCNR—InstituteofBiosciences andBioresources(CNR-IBBR)experimentalgarden(Perugia,Italy),andsixwerederivedfromtheolivegermplasmcollectionoftheConservatoireBotaniqueNationalMéditerranéen(CBNMed)(PorquerollesIsland,France)(TableS1).Thesewereusedaspollendonors,todefinethegeneticarchitec-tureofSIandtomaximizethegeneticdiversityofsampledO. europaea (Belajetal.2012;ElBakkalietal.,2013).

ToverifysegregationoftheSIphenotypeinprogenyfromanF1cross, pollenwas collected from 91 trees (hereafter called LEDA)growingattheCNR-IBBRInstitute(TableS2)thataretheprogenyof a controlled cross between Leccino andDolceAgogiavarieties(referencedasOit27andOit15).Theirpaternitywaspreviouslycon-firmedusingRAPD,AFLP,SSR,andRFLPmarkers(DelaRosaetal.,2003).

2.2 | Genotyping of the sampled trees with microsatellite markers

Togenotype sampled trees, totalDNAwasextracted from100mgof fresh leaftissuebyGeneElutePlantGenomicDNAMiniprepKit(Sigma-Aldrich),followingmanufacturer’sinstructions,andthenquan-tifiedbyaNanodropspectrophotometer.Genotypeidentificationwasperformedbyanalyzing15informativenuclearSSRmarkers(Baldonietal., 2009; El Bakkali etal., 2013). PCR products were separated


usinganautomaticcapillarysequencer(ABI3130GeneticAnalyzer,AppliedBiosystems),andelectropherogramsweretheninvestigatedforallelecompositionacrossmarkerlociusingGenMapper3.7soft-ware(AppliedBiosystems).

Toverify the genetic representationof the selected sample set,SSRdataobtainedon the118 trees (TableS1)werecompared toacollectionof342genotypes:the309olivegenotypespresentintheOWGBcollection(ElBakkalietal.,2013)togetherwiththe33geno-typessampledinItalian(27)andFrench(6)collectionsnotpresentintheOWGB.

2.3 | Assessments of the compatibility/incompatibility reactions

2.3.1 | Incompatibility tests

To ensure that receptive stigmas were free of contaminant pol-len,branchesabout40–50cmlongandbearingseveralflowerbudswerebaggedon tester recipients at least oneweekbeforeflowersopenedandstigmasbecamereceptive(usingtwoPBS3d/50bags,anouterbagenclosinganinnerbag,eachofsize16×50×16cm;PBSInternational,Scarborough,UK).A10×25cmPVCwindowallowedus tomonitorflowersor treat themwithoutopening thebags. Forprezygotic stigma tests, twigs were collected when flowers weremature(i.e.,when5%–10%offlowerspresentonatwigwereopen).Whenperformingpostzygotictestsbycontrolledcrossesandscoringof produced seeds, to prevent pollen contaminationduring pollina-tion,theouterbagwasremovedandtheinnerbagwaspiercedwithaneedletoinjectpollenwithaspraygun,theneedleholewascarefullytapedimmediatelyafterspraying,andtheouterbagwasputbackinplace.Toensurecontinuouspollenavailability,freshlycollectedpol-lenwasstoredat−80°C(Vernetetal.,2016)until itwasappliedtorecipientstigmas;thisprocedurealsoallowedustocollectpollenonthe latestfloweringtree in2013foruse inphenotypingonstigmasin 2014.

2.3.2 | Stigma test

Wescoredcross-compatibilityfollowingtheprotocolinVernetetal.(2016).Undertheseconditions,stigmastreatedwithpollenwerefixed16hrafterpollination,thenstainedwithanilineblue,andobservedunderaUVfluorescentmicroscope,whichallowedustodistinguishpollengrainsandpollentubesfrommaternaltissues(Figure1).Whenthepollenrecipientandthepollendonorarecompatible,severalpol-lentubesconvergethroughthestigmatictissuetowardthestyleuntilthebaseof thestigmaandentranceof thestyle (Figure1panelsband c). The absence of pollen tubes or the presence of only shortpollentubesgrowingwithinthestigmabutneverreachingthestylewas used as the criteria to score incompatibility (Figure1 panels aandd).Giventheriskofcontamination,asinglepollentubegrowingin the stigmatic tissuewas never considered a reliable criterion todeterminecompatibility.Threereplicateflowerswerepollinatedforeach cross.

2.3.3 | Interspecific stigma tests between O. europaea, P. angustifolia, and F. ornus

To (i) substantiate our conclusions on the occurrence of DSI inO. europaea, (ii) determinewhether SI recognition specificities haveremained stable in the hermaphrodite lineageof theOlea genus as itsdivergencefromthelineagecontainingtheandrodioeciousspeciesP. angustifolia,and(iii)takeadvantageofrecentknowledgeaboutthegenetic architecture of SI inP. angustifolia (Billiard etal., 2015),weperformedinterspecificstigmatestsusingP. angustifolia and F. ornus hermaphrodites assigned to the two SI groups in previous stud-ies (calledG1andG2) (Saumitou-Lapradeetal.,2010;Vernetetal.,2016).Stigmatestswereperformedusing the twopairsofO. euro-paeatesters(Oit27/Oit15andOit30/Oit26)asrecipients,withfrozenpollenfromP. angustifolia(Pa-01C.02[G1]andPa-06G.15[G2])andF. ornus(Fo-A17[G1];Pa-01N.08,Pa-13A.27,andFo-G1999-48[G2])belongingtotheG1andG2SIgroups.

TABLE 1 Resultsfromself-pollinationandreciprocalstigmatestsperformedinadiallelcrossingdesignamongsixOlea europaeagenotypes

Pollen donor

SI group G1 G2

DNA database reference Oit27 Oit26 Oit24 Oit15 Oit30 Oit28

Pollenrecipient G1 Oit27 SI 0 0 1 1 1

Oit26 0 SI 0 1 1 1

Oit24 0 0 SI 1 1 1

G2 Oit15 1 1 1 SI 0 0

Oit30 1 1 1 0 SI 0

Oit28 1 1 1 0 0 SI

SI,self-incompatibilityreactiondetected,nooronlyshortpollentubesobservedinstigmatictissueafterself-pollination;0,incompatibilityreaction,nooronlyshortpollentubesobservedinstigmatictissue(Figure1,panelaandd);1,compatibilityreaction,pollentubeswereobservedconvergingthroughthestigmatictissuetowardthestyle(seeFigure1panelbandc).Twoincompatiblegenotypeswereassignedtothesameincompatibilitygroup(eitherG1orG2);twocompatiblegenotypeswereassignedtodifferentincompatibilitygroups.DNAdatabasereferencecorrespondstovoucherspecimenaccessibleinreferencedcollections(seeTableS1).

pierre_s

Barrer


Weperformedstigmatestsbydepositingpollenfromonetestsam-pleonstigmasofapairofcross-compatibletesters(i.e.,belongingtotwodifferentSIgroups).UnderthehypothesisthatO. europaeaexhibitsDSI,weexpectedpollenfromeverysampletobecompatiblewithoneofthetwotesterlines(therebyconfirmingpollenviability)andincom-patiblewiththeother;indeed,casesinwhichpollenwouldbecompat-iblewithbothtesterlineswouldindicateeitherthepresenceofathirdSIgroup(differentfromtheserepresentedbythetesterrecipients)oranonfunctionalSIgenotype.Casesinwhichtestedpollenwasnegativewithbothreferencerecipientswerelikelycausedbyeitherlowpollenviabilityor lowstigmareceptivity.Hence,pollinationswere repeateduntilcompatibilitywasobservedonatleastonepollenrecipient.

2.4 | Postzygotic validation of SI group assignment

Tovalidatethecompatibilityversus incompatibilitystatusassessedbetween pairs of genotypes, according to pollen tube behavior inthestigmatests,wecarriedoutadditionalphenotypicassessmentsbasedonseedproductionaftercontrolledpollination.WefollowedtheprotocolestablishedandvalidatedforP. angustifolia (Saumitou-Laprade etal., 2010). Each tested genotype was used as a pollenrecipientand treatedas follows:Two40–50cm longbranchesper

treecarryingnumerousinflorescenceswereselectedaweekbeforetheopeningofthefirstflowersandcarefullyprotectedintwobagseach.One inflorescencewas pollinatedwith pollen collected frompollendonorsbelongingrespectivelytotheG1andG2incompatibil-itygroups(Oit27andOit15,respectively,seeResultsandTable1).Foreachcross,pollinationwas repeated threetimesoveraperiodof 8days, beginning when the first flowers opened (between lateMayandmid-June2014,dependingon theflowering stageof therecipient). Isolationbagswere removed several days after the endofflowering (on July10)and replacedbynetbags toprevent lossoffruitduringripening.Finally,fruitswerecollectedandcountedinmid-October,andtoconfirmpaternity,genomicDNAwasextractedfromfruitembryos(Díazetal.,2007)andfromleavesoftheparents.Parentsandoffspringweregenotypedusing10highlypolymorphicmicrosatellitemarkers (Table S3) having high exclusion probability(ElBakkalietal.,2013).PaternityassignmentswerecalculatedwithCervus3.0.3(TableS4).

2.5 | Genetic assignment of the trees phenotyped for SI group and assessment of genetic diversity

TodeterminehowoursamplingrepresentedthegeneticdiversityandthegeographicstructureoftheMediterraneanolivetree,SSRalleleswere scored in a single analysis (Tables S1 and S5) and combinedwith previous data obtained from the complete OWGB collection(ElBakkalietal.,2013).

The number of alleles per locus (Na), the observed (Ho) het-erozygosity and expected (He) heterozygosity (Nei, 1987) wereestimatedusingtheExcelMicrosatelliteToolkitv3.1(Park,2001).Principalcoordinateanalysis(PCoA),implementedintheDARWINv.5.0.137program(Perrier,Flori,&Bonnot,2003),wascarriedoutusingasimplematchingcoefficient.Toidentifythegeneticstructurewithinthestudiedsamples,incomparisonwiththeMediterraneanolivegermplasm,amodel-basedBayesianclustering implementedintheprogramStructurever.2.2(Pritchard,Stephens,&Donnelly,2000)was used. Bayesian analysiswas run under the admixturemodelfor1,000,000generationsafteraburn-inperiodof200,000,assumingcorrelationamongallelefrequencies.Analyseswererunfor values of Kbetweenoneandsixclusterswith10iterationsforeachvalue.ValidationofthemostlikelynumberofK clusters was performedusingtheΔKstatisticsdevelopedbyEvanno,Regnaut,and Goudet (2005)with the R program, and the similarity indexbetween10replicatesforthesameKclusters (H’)wascalculatedusing CLUMPP 1.1 (Greedy algorithm; (Jakobsson & Rosenberg,2007)). For each selected K value, each accession was assigned to its respective clusterwith a posteriormembership coefficient(Q>0.8).

We testedwhether theallelicdiversityobserved in the89gen-otypes representing a subsample of the core collectionwas signifi-cantlylowerthanthatoftheoverallOWGBcollectionusingaMann–WhitneyU-test(p-value>.01one-tailedtest)afterstandardizationofthedatasetusingtherarefactionmethodaccordingtoADZE(Szpiech,Jakobsson,&Rosenberg,2008).

F IGURE 1 Stigmatestsperformedtoassessself-incompatibilityin O. europaea:examplesofhermaphroditesOit26andOit28.(a)ThepollenofthehermaphroditeOit26 does not germinate on its own stigmademonstratingtheself-sterilityofthisindividual;(b)Oit26 pollengerminatesonhermaphroditeOit28 attesting to its viability; (c)thestigmafromOit26 allows germination of Oit28pollenattestingtothestigma’sfunctionalreceptivitywhenpollinatedbycompatiblepollen;(d)theOit26pollendoesnotgerminateonitsownstigmademonstratingtheself-sterilityofthisindividual.Arrowspinpointtheregioncorrespondingtothebaseofthestigmaandentranceofthestyle.M:genotypeusedasmalepollendonor;F:genotypeusedasfemalerecipient

FM

Oit26

G1Oit28

G2

Oit26

G1

Oit28

G2

(a) (b)

(c) (d)


2.6 | Statistical analysis of pollen tube length scored in incompatible crosses

Thespecificlengthofpollentubeswithinstigmatictissuewasmeas-ured for a given set of incompatible reactions (i.e., the growth thatoccurredpriortothearrestoffurthergrowth).Basedonthisgrowth,wedefinedninediscretephenotypicclasses,fromi_1toi_9(Figure2).Because an incompatible response scored in the highest phenotypicclasses(i.e.,longerpollentubes,see[i_7],[i_8],and[i_9])couldbemis-takenforacompatibleresponse,weappliedgeneralizedlinearmodel(GLM)analysestothephenotypicdata.Asubsetof86pollendonorswas crossed with the four O. europaea testers involved inthestigmatestdescribedabove(TableS6).Foreachpollendonor,weobservedfour crosses (two compatible and two incompatible), and for eachcross,wephotographedpollen tubes growingdown stigmatictissueandstylesinthreedifferentflowers.Theimageswererandomlylabeledandobservedfourtimesindependently,providingfourreadsforassign-menttoaphenotypicclass(i.e.,12independentscoresforeachcross).

First,wetestedtheeffectoftheSIgroup,replicatescoring,pol-linator genotype, recipient genotype, and individual flowers on theSIphenotypicresponse(i.e.,thelengthofpollentubegrowthbeforegrowtharrestwithinstigmatictissueinincompatiblecrosses,scoredamongninephenotypicclassesbytheexperimenter).Wethenusedgeneralized linearmixedmodels (GLMM)onthecategorizedpheno-typeoftheSIresponsetotestthefollowing:(i)whetherthepheno-typescoringbasedondigitalimageswasrepeatable,(ii)whetherthephenotypewas consistent among replicatesof the samepollinationtest, (iii)whether theSIgroupsshowedadifferentSI response,and(iv)whether the genotype of the recipient had an effect on the SIresponse.Weconsideredthefactors“flowerread,”“pollinatorgeno-type,”and“flower”asrandomeffects,and“recipientgenotype”and“SIgroup”asfixedeffects.TheSIresponsewasthedependentvariableandfollowedaPoissondistribution.Totestwhetherarandomorfixedfactorhadasignificanteffect,weperformedalikelihoodratiotestofnestedmodels,usingthepackagelme4inR(Batesetal.,2014).

3 | RESULTS

3.1 | Phenotypic characterization of self- incompatibility in O. europaea

In2013,sixaccessions(Oit27,Oit26,Oit24,Oit15,Oit30,andOit28)correspondingtosixdifferentgenotypes(TableS1)wereusedasbothpollenrecipientsandpollendonors,inareciprocaldialleldesign,forthestigmatests(Table1).Self-fertilizationwastestedonthesixgeno-types,andnopollentubesuccessfullyreachedthestyleinanyoftheobserved pistils, confirming strong SI reactions (Table1). However,thelengthofpollentubeswithinthestigmatictissuesvariedbetween

genotypes.Pollentubesdidnotgrowatall,ortheirgrowthstoppedveryearlyinthefirstlayerofthestigmacellsintheOit26genotype(Figure1, panel a). In comparison, arrest of pollen tube growth didnotoccuruntilthepollentubeshadreachedthedeeperlayersofthestigmacells,intheOit28genotype(Figure1,paneld).However,eveninthiscase,thepollentubesstoppedbeforereachingthetransmittingtissueofthestyle.

Fortheintergenotypepollinationtests,pollentube/stigmainterac-tionssuggestedtheexistenceofSIreactionsforeachofthesixdiffer-ent individualtreeswhencrossedwithspecificpartners (Table1).AnincompatibilityphenotypesimilartotheOit26self-fertilizationreaction(Figure1,panela:noorveryshortpollentubesdetected)wasobservedinthestigmatictissuesfromOit27,Oit26,andOit24whentheirpistilswerepollinatedbyoneanother.Theviabilityoftheirpollenandrecep-tivityoftheirstigmaswereverifiedincompatiblecrosseswithOit15,Oit30,andOit28.AphenotypesimilartotheSIreactionobservedwithOit28(Figure1,paneld:pollentubesofvariablelengthneverreachingthestyle)wasobservedinthestigmasfromOit15,Oit30,andOit28whenpollinatedbyoneanother.Hereagain,pollenviabilityandstig-maticreceptivityofthesamethreeindividualswerecheckedincom-patiblecrosseswithOit27,Oit26,andOit24.WeconcludedthattreesOit27,Oit26,andOit24areincompatiblewitheachotherandbelongtoasingleSIgroup,whereastreesOit15,Oit30,andOit28belongtoadifferentSIgroup.These results suggest thatO. europaea individu-alscanbeclassifiedintoatleasttwogroupsofSI,withincompatibilityreactionsbetweenindividualsbelongingtothesamegroupandcompat-iblereactionsbetweenindividualsbelongingtodifferentgroups.

3.2 | The two O. europaea SI groups are functionally homologous to those of P. angustifolia and F. ornus

Nonambiguous and repeatable incompatibility phenotypes wereobserved when P. angustifolia and F. ornusG1pollenwasdepositedon stigmas from Oit26 and Oit27 (Figure3A,B, panel a), whereascompatibility phenotypes were scored on stigmas from Oit15 andOit30 (Figure3A,B, panel b). This demonstrated the capacity oftrans-generic pollen to germinate and elicit both incompatible andcompatible responses on O. europaea stigmas. Similarly, incompat-ibility phenotypes were scored on stigmas from Oit15 and Oit30(Figure3A,B,paneld),andcompatibilityphenotypeswereobservedwith P. angustifolia and F. ornusG2pollenonstigmasfromOit26andOit27 (Figure3A,B, panel c). Therefore, we concluded that the SIsystem of O. europaea isfunctionallyhomologoustotheDSIsystempreviouslyreportedforP. angustifolia and F. ornus(Saumitou-Lapradeetal.,2010;Vernetetal.,2016).Weassigned theOit26andOit27genotypes,andalltheirincompatiblemates,totheG1SIgroup,andtheOit15andOit30genotypes,andalltheirincompatiblemates,totheG2SIgroup.

F IGURE 2 Classesofincompatibilityphenotypesobservedwithinself-incompatibilitygroupsaccordingtopollen(donor×recipient)interactions.OnstigmasbelongingtotheG1group,pollentubelengthaftergrowthwasarrestedwashomogenous:fromnulltolow(seethecasesG1:[i_1]to[i_3]).OnstigmasofG2groups,pollentubelengthafterthearrestofgrowthvariedwidelyamong(donor/recipient)pairs:fromnulltohigh(seethecasesG2:[i_1]to[i_9])


3.3 | Segregation of the self- incompatibility phenotypes in a controlled cross

Among the91LEDA full-sib trees thatflowered in2013and/or in2014,41wereincompatiblewithG1recipientsandcompatiblewithG2, indicating that they belong to theG1 compatibility group, and50were incompatible with G2 recipients and compatible with G1,indicating that they belong to G2. No offspring appeared compat-ible or incompatible with both groups of recipients in any of thetests(Table2).Theobserveddataagreewithageneticmodelassum-inga1:1segregationofthetwophenotypicgroups(Chi²teststatis-tic=0.345,df=1,ns).

3.4 | Two and only two self- incompatibility groups detected in O. europaea

The118sampledtreesfromthefourgermplasmcollections(OWGB,Perugiacollection,CNR-IBRR,andCBNMed)represented89distinctgenotypes(20genotypeswererepresentedbymorethanoneclonal

replicate,TableS1).Weperformedatotalof1,500pollinationtests,whichallowedus todetermine theSIphenotypeofeach individualtreewithameanof2.6replicatespertestedgenotype.Allreplicateswere fully concordant (Fig. S1), demonstrating the robustness andreliability of the stigma tests performed.Among the89 genotypes,42genotypeswereincompatiblewithG1recipientsandcompatiblewithG2, indicatingthattheybelongtoG1,and47genotypeswereincompatiblewithG2 recipients and compatiblewithG1, revealingthattheybelongtoG2.Noneofthegenotypeswereeithercompat-ibleorincompatiblewithbothgroupsofrecipients,provingthattwoandonlytwoSIgroupswerepresentinourextendedsample(Table3).

3.5 | Population genetic assessment of the sampling and representativeness of olive diversity

ThesampleswephenotypedforSIrepresented89distinctgenotypes.We were highly conservative in our genotype identification: Wegroupedgenotypesdefinedbyaspecificallelecombinationat15locithatdifferedbyonlyonealleleortwoalleles intoasinglegenotype

TABLE 2 Self-incompatibilityphenotypingofthe91LEDAF1treesfromthe(Oit64×Oit27)controlledcross

SI group [G1] S1S2a [G2] S1S1a [Other] SxSy

IncompatiblewithG1andcompatiblewithG2 IncompatiblewithG2andcompatiblewithG1 CompatiblewithG1andG2

Total 41 50 0

Threetypesofbehaviorwerescored.[G1],individualincompatiblewithG1testersandcompatiblewithG2testers;[G2],individualincompatiblewithG2testersandcompatiblewithG1testers;[Other],individualcompatiblewithG1andG2testersandthereforebelongingtoaSIgroupdifferentfromG1andG2TheS-locussegregatesasasinglelocuswithtwoallelesS1 and S2(withS2 dominant over S1)(Chi²test=0.345,df=1).aExpectedgenotypededucedfromgeneticanalysesinP. angustifolia(Billiardetal.,2015).

F IGURE 3 Trans-genericconservationoftheself-incompatibilityreactionbetweenOlea europaeaandtwootherOleaceaespecies:(A)Phillyrea angustifoliaand(B)Fraxinus ornus.Inthephotographspresented,stigmafromO. europaeaispollinatedbyP. angustifolia and F. ornus pollen.(a)IncompatibilityreactionbetweenstigmaofOit26andG1pollen;(b)compatibilityreactionbetweenstigmaofOit15andG1pollen;(c)compatibilityreactionbetweenstigmaofOit26andG2pollen;(d)incompatibilityreactionbetweenstigmaofOit15andG2pollen

(A)

a b a b

c d c d

(B)Olea s�gma recipient

G1 plants G2 plants

Phillyreapollendonor

G1plants

G2plants

Olea s�gma recipient

G1 plant G2 plant

F. ornuspollendonor

G1plant

G2plant


consideringthepossibilitythattheirdifferencesderivedfromsomaticmutations that occurred within old clones (Haouane etal., 2011).Withinthe89genotypes,wedetected179allelesoverthe245scoredonthecollectionof342genotypes(TableS5-A).Hence,oursamplecaptured73%ofthetotalallelicdiversityobservedinthecollectionof342genotypes.Tocheckfortherepresentativenessofolivediversityinoursubsampleof89genotypes,wecomparedallelicrichnessinthesubsamplewiththatofthecollectionof342genotypesaftercorrec-tion fordifference in samplesizesbasedon the rarefactionmethod(seeTableS5-B).Allelicrichnessofthetwosamplesetswasnotsig-nificantlydifferent(Mann–WhitneyU- test at p ≤ .01 using one- tailed test;U=89,P-value=.171;TableS5-B).Furthermore,wenotedsimi-larexpectedheterozygosityvalues (He=0.745 in the89genotypesandHe=0.749inthe342genotypescollection;TableS5-A).

Mostof the89genotypesphenotyped forSIwere classified intoone of the three western, central, and eastern Mediterranean clusters detected in previous studies (El Bakkali etal., 2013; Haouane etal.,2011), with a slight underrepresentation of the eastern gene pool(TableS5-C).Thisconclusionwasfurtherconfirmedbytheirpositiononthefirst twoaxesof theprincipalcoordinateanalysis (PCoA,Fig.S3).Overall,despitethelimitednumberofeasternolivetrees,the89geno-typesweredistributedamongthethreeMediterraneangenepools,indi-catingtheywereafairrepresentationofdomesticatedolivediversity.

3.6 | The incompatibility response differs between the two self- incompatibility groups

Whenweanalyzedvariationinpollentubelengths,wefoundnosignif-icantvariationamongreplicateobservationsofthesameflowerandamongflowersofasingleindividualwhenpollinatedwithincompat-iblepollen, indicatingconsistent incompatibilityreactions.However,theSIgroupof thepollen recipienthadasignificanteffect (p-value <.0001)onthedistancethatincompatiblepollentubeswereabletogrownwithinthestigma:PlantsbelongingtoG1showedanSIphe-notypethatfellinclassesoflowvalue(shortpollentubes),whereasplantsbelonging toG2showedphenotypes that can fell inawiderpanelofvalues(fromshorttolongpollentubes).TheincompatibilityreactionbetweenG1individualsseemedtooccuralmostimmediatelyafterpollenlandedonthestigmaaseitherpollengrainsdidnotger-minate or pollen tube growth stopped shortly after germination. Incontrast,the incompatibilityreactionbetweenG2individualsseemstooccurlater:Pollengrainsgerminatedandpollentubesgrewintothestigmatictissuebeforetheirgrowthwasarrested.

Ouranalysesalsorevealedasignificanteffectoftherecipientgen-otypeonthescorevaluewithineachSIgroup(p-value <.001 for both G1andG2).WithinG2,Oit30showedahigherscorethanOit15,andwithinG1,Oit27showedahigherscorethanOit26.Thissuggestscon-sistentdifferencesamonggenotypesinthetimingoftheSIresponse(earlyorlate),whosefunctionalsignificanceremainstobedetermined.

3.7 | The SI assignment based on prezygotic stigma test validated by postzygotic genotyping

Weverifiedatthepostzygoticstagewhethercasesofincompatibilityinwhichpollen tubeswere able to germinate andgrow substantialdistancesinthestigma(thereforethemostambiguouscasesbecauseof their relative similarity to compatible phenotypes)were cases inwhich fertilizationwas not achieved. The functional incompatibilityof10differentgenotypes,belongingtoSIgroupG2andwhichscoredinthehighestphenotypicclasses[i_7],[i_8],and[i_9](Figure2),wasassessedatthepostzygoticstagethroughprogenyanalysis(TableS4),aswellascountsofthenumberofseedsproduced(Table4).All99progenyfromcrossesbetweenputativelycompatiblemates(assignedbasedontheprezygoticstigmatest),hadgenotypescompatiblewithbothparents(TableS3).Thisconfirmsthatthestigmatestisreliableandsuggeststhatourexperimentaldesignpreventspollencontami-nation. In contrast, thenumberof seedscollected following the10pollinations between parents belonging to the same SI group wasextremely low(noseedproducedinsevencrossesand2,2,and10seedsinthethreeremainingcrosses,respectively).Inaddition,noneoftheseedsharvestedinthesethreecrosseshadagenotypethatwasconsistentwith itsputativefather.Again, theseresultsconfirmthatthestigmatestisareliableproceduretopredictwhichincompatibilitygroupaplantbelongsto,even in thosecases inwhichsomepollentube growth occurswithin the stigma. Interestingly, the few seedsobtainedwereallattributedtoselfing.

4 | DISCUSSION

4.1 | Confirmation that the three genera Olea, Phillyrea, and Fraxinus share the same self- incompatibility system

TheevolutionofnewSIsystemsinplantsisthoughttobeararephe-nomenon,whichisinagreementwiththegeneralobservationthatSImechanismsaregenerallysharedamongspeciesthatexhibitSIwithin

TABLE 3 Resultofstigmatestsperformedwith89O. europaeagenotypestestedforcompatibilityandincompatibilitywithtwopairsofpollenrecipientsusedastesters

SI group [G1] [G2] [Other]

IncompatiblewithG1andcompatiblewithG2 IncompatiblewithG2andcompatiblewithG1 CompatiblewithG1andG2

Total 42 47 0

Threetypesofbehaviorwerescored(seeTable2caption).ThecultivarstestedbelongeithertoG1ortoG2,andnonebelongtoahypotheticalthirdincompatibilitygroup.Inthesampletested,wedetectedonlytwoincompatibilitygroups.


agivenplantfamily(Allen&Hiscock,2008;Charlesworth,1985;Igicetal.,2008;Welleretal.,1995)andthat lossesofSIwithinacladearemuchmore common than gains (Igic, Bohs, & Kohn, 2006). Asexpectedbasedonthesearguments,weconfirmedinO. europaea the occurrence of the same DSI discovered in P. angustifolia and F. ornus (Vernetetal.,2016),twoandrodioeciousspeciesthatbelongtotwophylogenetic branches of the same family that diverged from eachothermorethan40Mya(Besnardetal.,2009).WhileSIsystemsareoftentrans-generic, long-termstabilityofhomomorphicDSI—that isthepresenceofonlytwoallelesoveralongtime—isunexpectedfortworeasons.First,SIsystemsaresusceptibletotherapidinvasionofnewincompatibilityalleles,asaconsequenceofthestrongfrequency-dependent advantage of rare mating phenotypes (Wright, 1939).Gervaisetal.(2011)showedthatinamodelwherenewallelesarisethroughself-compatibleintermediates,selectionforallelicdiversifica-tion is inverselyrelatedtothenumberofsegregatingS-alleles, that is,moreactivediversificationwitha lownumberofalleles.Second,inhermaphroditic species, self-compatiblemutants areexpected toinvade a homomorphic DSI population regardless of the extent ofinbreeding depression (Charlesworth & Charlesworth, 1979). Thestability of DSI was recently explained in the case of androdioecywithatheoreticalmodel(VandePaer,Saumitou-Laprade,Vernet,&Billiard,2015),showingthatandrodioecyandDSIhelpmaintaineachother.DSI facilitatesthemaintenanceofmales (Billiardetal.,2015;Husse,Billiard,Lepart,Vernet,&Saumitou-Laprade,2013;Pannell&Korbecka,2010),andthefullcompatibilityofmaleshinderstheinva-sionofself-compatiblemutants(VandePaeretal.,2015).

ThesituationisquitedifferentforO. europaea.Thespeciesbelongsto the subgenus Oleawhichcontainsonlyhermaphroditespeciesandhasdivergedmorethan30Myafromthelineagecontainingandrodi-oecioustaxasuchasOsmanthus and Phillyrea (Besnardetal.,2009).TheevolutionarycausesofthemaintenanceofDSIover30Myremain

tobeidentified.MolecularcharacterizationoftheSIlocusisaprom-ising avenue of research to resolve issues related to the origin and maintenanceofhomomorphicDSI,becausethesimplestexplanationisthatthegeneticarchitectureofthesystemdoesnotallowthegen-erationofadditionalSIphenotypes (e.g.,athirdSIallele).Molecularcharacterizationwillbe facilitatedby the trans-generic functionalityof DSI that we observed among the P. angustifolia, O. europaea, and F. ornusspecies.

4.2 | Self- incompatibility in O. europaea is sporophytic

OurresultsareconsistentwithdeterminationofSIonO. europaea by asingleS-locuswithonlytwoallelespresent inallcultivatedformsof the species and demonstrate the sporophytic nature of this SI.First, the 1:1 proportion of the two parental SI groups in the con-trolledcrossprogenyexcludesthepossibilityofgametophyticgeneticcontrolof self-incompatibility (GSI) (Bateman,1952) inO. europaea. Second,withGSI,theincompatibilitygeneattheS-locusisexpressedinthehaploidpollengrainsandinteractswiththediploidtissueofthestigma. Tobe functional, aGSI system requires strict codominancebetweenS-allelesinthepistiltoavoidcompatibilityofheterozygousindividualswithhalfoftheirown(self)pollenandaminimumofthreeallelesthatdefineaminimumofthreeincompatibilitygroups(Hiscock&McInnis,2003).Incontrast,inthecaseofsporophyticgeneticcon-trolofself-incompatibility(SSI),theincompatibilitygeneisexpressedbeforemeiosis in thediploidsporophytictissue,and incompatibilityariseswithonlytwoalleles,withacompletedominanceofonealleleover the other (see reviews by (Hiscock &McInnis, 2003; Billiard,Castric,&Vekemans,2007). Inour recentgeneticstudyperformedwith P. angustifolia(Billiardetal.,2015),weshowedaSIsystemgov-erned by an S- locus with two alleles, S2 and S1 (withS2 dominant

TABLE 4 NumberofseedscollectedonG2genotypesaftercontrolledcompatibleandincompatiblecrossesperformedinJune2014andverifiedbypaternitytesting

[G2] Genotypes used as recipient

Pollen donors

[G1]: Oit27 [G2]: Oit15

Seeds produced Paternity confirmed/tested Seeds produced Paternity confirmed/tested

LEDA_222 27 NA 0 –

LEDA_262 24 NA 0 –

LEDA_282 102 20/20 0 –

LEDA_301 98 20/20 0 –

Oit28 15 NA 2 0/2a

Oit03 30 10/10 0 –

Oit55 16 12/12 10 0/10a

Oit57 17 17/17 2 0/2a

Oit36 25 10/10 0 –

Oit22 40 10/10 0 –

aSelfingcannotbeexcludedwiththe10microsatellitemarkersused(seeTablesS3forgenotypingresultsandS4forestimationofexclusionprobabilitybasedonmarkersandcalculatedusingCervusver.3.0.3.);NA,fruitsnotcollected.


over S1),whichproducedthetwoincompatibilitygroupsG1andG2(Saumitou-Lapradeetal.,2010),andwithS1S2correspondingtoG1and S1S1toG2(Billiardetal.,2015).

The gametophytic versus sporophytic nature of the SI systemin O. europaeahasbeenquestionedfora longtimeinthe literature,usingindirectarguments,andseveralstudiesontheSIofoliveculti-varshaveresultedinvariableandconflictingresults(forareviewseeSeifi,Guerin,Kaiser,&Sedgley,2015).Featuresrevealedbyhistolog-icalinvestigations,suchasbinucleatepollenandwetpapillaestigmaorasolidstyleandalargenumberofpollengrainsgerminatingonthestigmasurface,werereminiscentofspecieswithGSI(DeNettancourt,1997),whereasadrypapillaestigmawasalsoreported inOleaceae(Heslop-Harrison&Shivanna,1977).AdditionalargumentsbasedontheobservationofpollentubegrowthinincompatiblecrosseswereinfavorofGSI:Thewaypollentubeshaltedintheproximalareaofthestylewas interpretedas the interventionofprogrammedcelldeath,a frequent featureofGSI.OtherargumentsbasedonhistochemicallocationofkeyenzymeactivitiesinvolvedinGSIwerealsoreportedinolivetree(Serrano&Olmedilla,2012).Moreover,transcriptomeanal-yseshavebeenperformedtoscreenforconservedtranscriptstypicalofGSIinotherplantspecies(e.g.,S-ribonucleasetranscriptssuchasinSolanaceae;(McClure,2006))orSSI(e.g.,S-receptorkinasetranscripts,suchasinBrassicaceae;(Takasakietal.,2000)).TranscriptssimilartomaleandfemaleSSIdeterminantsofBrassicaceaewereidentifiedinolive(Alagnaetal.,2016;Collanietal.,2010,2012);however,thereisnoevidenceoftheirfunctionalityinSIreaction.

Here, we demonstrated that one of these numerous indirect arguments for assessing the gametophytic/sporophytic status ofSIwaswrong: ForoneSI group, the incompatibility reaction takesplaceatthestigma,whereasfortheotherSIgroup,theincompati-bilityreactionoccurs later,sometimesat theentranceof thestyle.ThisfeaturemayexplainsomeofthepastdifficultyinidentifyingthegametophyticorsporophyticnatureoftheincompatibilitysysteminO. europaea.

4.3 | Within- group incompatibility is stricter than within- individual self- incompatibility in O. europaea

Onesurprisingobservation fromourexperiments is theproductionofasmallnumberofselfedseedsbyG2individualsfollowingpollina-tionwithincompatibleoutcrosspollen.ThisistheonlyindicationofapartialbreakdownofSIinthefaceofanotherwiseverystrongSIreac-tion.Whyself-pollinationseemstobepromoted inthepresenceofincompatibleoutcrosspollenremainstobedetermined.Thisfeatureisunexpectedbecause inSIsystems, the incompatibilityphenotypeofapollengrainshouldonlydependonthepollenparentgenotypeat the S- locus, which is shared among individuals from the same SI group.Thisresultmayindicatethat,inoliveor,atleast,insomeolivegenotypes,theincompatibilityreactionmaybestrongerwithoutcrosspollenfromthesamegroupthanwithself-pollen. It isalsopossiblethat this observation results from the larger amount of self-pollendepositedonstigmasthroughautonomousself-pollination,comparedwiththeoutcrosspollentransferredexperimentally.

4.4 | Olea europaea is a true self- incompatible species in which some genotypes can produce seeds by selfing

AllgenotypestestedforSIinthepresentstudywereclassifiedasself-incompatibleaccordingtothecriteriaofourstigmatest,andallbelongtooneofthetwoSIgroupsidentifiedinthespecies.ThesestatementsconfirmthatO. europaeaisatrueself-incompatiblespecies.Theyarein agreement with conclusions of studies that tested SI in O. europaea atthepostzygoticlevel,bymeasuringseedproductionaftercontrolledcrosses or open pollination, togetherwith paternity analysis of theprogeny(DelaRosaetal.,2004;Díazetal.,2006;Marchese,Marra,Caruso,etal.2016;Marchese,Marra,Costa,etal.,2016;Mookerjeeetal.,2005).Justasinourstudy,manystudiesobservedseedspro-duced by selfing either from controlled crosses with pollen fromincompatiblegenotypes(seeOit28,Oit55,andOit57inTablesS3andS4)or incontrolledselfing(Farinellietal.,2015)oropenpollination(Marchese,Marra, Costa, etal., 2016). The self-incompatible statusofaspeciesdoesnotexcludethepossibilitythattheincompatibilityreactionmaybebrokenforself-polleninsomegenotypes.Theunder-lying mechanism allowing this remains to be studied. The occurrence ofalowrateofselfinginindividualplantswithanactiveSIsystemiscommonlyreportedandisreferredtoaspseudo-self-compatibilityorleakyself-incompatibility.LeakySI isgenerallythoughttobeacon-sequenceofenvironmentalfactorsinterferingwiththeSIreactionortotheactionofmodifiergenes(Busch&Schoen,2008;Levin,1996).

The leakySIobserved inolivehasprovidedmaterial forgeneticmappingandsequencing(Marchese,Marra,Caruso,etal.2016)andallowsanopportunity tomeasure inbreedingdepression.Forexam-ple,inthewildrelativeP. angustifolia,2%of2,000surveyedseedlingsproduced from controlled crosseswere found to have been selfed(Billiard etal., 2015). Notably, none of these selfed seedlings everflowered (unpublished results). In addition, leaky SI in olive mightexplain the gradient of results that have until nowmasked the realself-incompatibilitysystem.

5 | CONCLUSION: ADDITIONAL EVOLUTIONARY APPLICATIONS

The level of interindividual incompatibility thatweobserved in ourstigmatestwasveryhigh:Onaverage,halfofthepairsofgeneticallydistincttreesfromthesampledcollectionsweremutuallyincompat-ible.Similarly,mostof the studies checking for compatibilitywithinandamongolivevarieties,whenusingseedproductionandpaternityanalyses, detected numerous cases of cross-incompatibility (De laRosaetal.,2003;Díazetal.,2006;Mookerjeeetal.,2005;Wuetal.,2002).Incontrast,studiesinorchardsorcropsofotherdomesticatedspecieswithSI,undereitherGSI(e.g., Prunus, Malus, Pyrus, Amygdalus)orSSI (e.g.,Brassica, Cichorium),showhighnumbersofS-allelesandtherefore high levels of cross-compatibility within or between cul-tivars (Dreesen etal., 2010;Ockendon, 1982;Wünsch&Hormaza,2004). In the olive, the lownumber of elite varieties that co-occur


inanorchard,togetherwiththe50%chanceofcross-incompatibilitybetweenpairsofvarietiesaccordingtoitsDSIsystem,maylimitfruitproduction.Limitationoftheavailabilityofcompatiblepollen,aphe-nomenondescribedastheS-Alleeeffect,occursinwildpopulationsofSIspecieswithlowS-allelediversity(Leducqetal.,2010;Wageniusetal.,2007).Smallisolatedpopulationsorpopulationsthathaveexpe-riencedarecentgeneticbottleneckmayhavelimitedallelicdiversityat theS-locus, leading toan increase in theprobabilityof interindi-vidual incompatibility,whichinturncausesareductioninseedpro-duction(Byers&Meagher,1992;Vekemansetal.,1998).

The discovery of the DSI system in O. europaea will undoubtedly offeropportunitiestooptimizefruitproduction.First,ithelpstounder-standtheheretoforeunexplainedbeneficialeffectofancestralprac-ticesthatencouragetheplantingofaminimumnumberofvarietiestoensuresatisfactoryoliveproduction.Second,easy-to-usemethodsshouldbedevelopedtodeterminetheSIphenotypeofeachcultivatedvarietyofolivetohelpguidethechoiceofvarietiestobeassembledin a given orchard, especially in nontraditional olive growing areas.Finally,ecologicalmodelscanbedevelopedtoaddressthequestionoftheoptimalnumberofvarietiestobeintroducedtoensure,effectivepollinationinanorchard,regardlessofclimate.Clearly,mono-varietalorchardsmustbeavoided.InadditiontotheSIphenotype,thechoiceofvarietiesshouldtakeintoaccountotherimportantparameterssuchas flowering phenology, the direction ofwind during the floweringperiod,andtherelativepositionsofthedifferentvarietieswithintheorchard.

Inthepresentstudy,wechosetopresentvarietiesthroughtheirreference genotype and not through their variety name, to assessthestrictassociationbetweengenotypeandSIphenotype.Previousstudiessuggestedpossiblediscrepanciesbetweenvarietalnamesandgenotypes(ElBakkalietal.,2013;Haouaneetal.,2011;Trujilloetal.,2014),andduringourstudy,weobserveddifferentnamesassociatedwithasinglegenotype(TableS1)aswellasdifferentgenotypesasso-ciatedwithasinglevarietyname;indeed,inmorethan20%ofcases,thegenotypesassociatedwiththesamenameweredifferent intheItalianandOWGBcollections (datanotpresented).Therefore, thereisnostrictassociationexpectedbetweenvarietynameandSIpheno-type.Therefore,eachgenotypeofinterestforoliveproducersneedstobeassignedtooneofthetwoSIgroups.Thiswillrequirecharacter-izingthesegenotypesfortheirSIphenotypeusingthestigmatestinrigorousconditions.Lastly,aneffortshouldbedevotedtoidentifyingmolecularmarkerswithstronglinkagewiththeS-locustoprovideaneasy-to-use diagnostic molecular assay for genotyping trees at theS-locus.Weareconfident that theevolutionaryconservationof thefunctionalityoftheDSIamongtheOlea, Phillyrea, and Fraxinus genera willbeanassetforaccomplishingthistask,throughgenomicandtran-scriptomiccomparativeanalysesofthetwogroupswithinandamongthese three genera.

ACKNOWLEDGEMENTS

We warmly thank Drs. Vincent Castric, Isabelle De Cauwer, andLyndaDelphforscientificdiscussionsandhelpfulcommentsonthe

manuscript.We thank Sylvia Lochon-Menseau for collecting pollenfrom the French germplasm collection in Porquerolles Island, andSylvain Santoni, Pierre Mournet, and Ronan Rivallan for their techni-calsupportofmolecularanalysisingenotypingplatform(UMRAGAP,Montpellier).

DATA AND MATERIAL SHARING

AllrelevantdataarewithinthepaperanditsSupportingInformationfiles.

CONFLICT OF INTEREST

Theauthorshavedeclaredthatnocompetinginterestsexist.

AUTHORS’ CONTRIBUTIONS

All authors contributed significantly to thework presented in thismanuscript.P.S.-L.andP.V.jointlydesignedandcarriedthesamplingdesignandphenotypingstrategieswithB.K.andL.B..P.S.-L.andP.V.performed phenotyping and crosses togetherwith F.A., R.M., S.P.,andM.R..B.K.organizedthecollectofpollen intheOWGBcollec-tion inMarrakech (Morocco) and at the CBNMed in Porquerolles(France)togetherwithA.MhandA.Mo.L.B.organizedthecollectofpollen in thePerugiacollection (Italy)and in theCNR-IBBRcol-lections (including the LEDA F1 progeny). L. E. andN.G.M.C. per-formedDNAextractionandgenotyping.B.K.andA.E.B.performedpopulation genetic structure analyses. R.M. performed paternityanalyses.S.B.performedstatisticalanalysesofthevariationinpollentubelengths.S.G.createdbioinformaticstoolsallowingmanagement,comparison,andsharingamongpartnersofthethousandsofpicturesproducedfortheSIphenotyping.G.B.providedexpertiseonoliveSIandstronglycontributedtoinitiatetheproject.X.V.providedexper-tiseontheSIandpopulationgeneticsanalysis.P.S.-L.,P.V.,andX.V.wrotethepaper.

REFERENCES

Alagna, F., Cirilli, M., Galla, G., Carbone, F., Daddiego, L., Facella, P., …Perrotta, G. (2016). Transcript analysis and regulative events duringflowerdevelopmentinolive(Olea europaeaL.).PLoS One, 11(4),1–32.

Allen,A.M.,&Hiscock,C.J. (2008).Evolutionandphylogenyofself-in-compatibilitysystemsinangiosperms.InV.E.Franklin-Tong(Ed.),Self-incompatibility in flowering plants – Evolution, diversity, and mechanisms. BerlinHeidelberg:Springer-Verlag.

Ateyyeh,A. F., Stosser, R., &Qrunfleh,M. (2000). Reproductive biologyoftheolive(Olea europaea L.)cultivar’NabaliBaladi’.Journal of Applied Botany, 74(5–6),255–270.

Baldoni,L.,Cultrera,N.,Mariotti,R.,Ricciolini,C.,Arcioni,S.,Vendramin,G.,…Testolin,R.(2009).Aconsensuslistofmicrosatellitemarkersforolivegenotyping.Molecular Breeding, 24(3),213–231.

Bateman, A. J. (1952). Self-incompatibility systems in angiosperms.Heredity, 6(3).

Bates, D., Maechler, M., Bolker, B., Walker, S., Christensen, R. H. B.,Singmann, H., … LinkingTo Rcpp (2014). Package ‘lme4’. Vienna: R FoundationforStatisticalComputing.


Belaj,A.,Dominguez-García,M.delC.,Atienza,S.G.,Urdíroz,N.M.,DelaRosa,R.,Satovic,Z.,Martín,A.,Kilian,A.,Trujillo,I.,&Valpuesta,V.(2012).Developingacorecollectionofolive(Olea europaeaL.)basedonmolecularmarkers (DArTs,SSRs,SNPs)andagronomic traits.Tree Genetics & Genomes, 8(2),365–378.

Besnard,G.,deCasas,R.R.,Christin,P.-A.,&Vargas,P.(2009).Phylogeneticsof Olea (Oleaceae) based on plastid and nuclear ribosomal DNAsequences: Tertiary climatic shifts and lineage differentiation times.Annals of Botany, 104(1),143–160.

Billiard,S.,Castric,V.,&Vekemans,X.(2007).Ageneralmodeltoexplorecomplexdominancepatternsinplantsporophyticself-incompatibilitysystems. Genetics, 175(3),1351–1369.

Billiard, S., Husse, L., Lepercq, P., Godé, C., Bourceaux, A., Lepart, J., …Saumitou-Laprade,P.(2015).Selfishmale-determiningelementfavorsthe transition from hermaphroditism to androdioecy. Evolution, 69, 683–693.

Breton, C. M., & Bervillé, A. (2012). New hypothesis elucidates self-incompatibilityintheolivetreeregardingS-allelesdominancerelation-ships as in the sporophyticmodel.Comptes Rendus Biologies, 335(9),563–572.

Breton,C.M.,Farinelli,D.,Shafiq,S.,Heslop-Harrison,J.S.,Sedgley,M.,&Bervillé,A.J. (2014).Theself-incompatibilitymatingsystemoftheolive (Olea europaea L.) functionswithdominancebetweenS-alleles.Tree Genetics & Genomes, 10(4),1055–1067.

Broothaerts,W.(2003).NewfindingsinappleS-genotypeanalysisresolveprevious confusion and request the re-numbering of some S-alleles.Theoretical and Applied Genetics, 106(4),703–714.

Busch,J.W.,&Schoen,D.J.(2008).Theevolutionofself-incompatibilitywhenmatesarelimiting.Trends in Plant Science, 13(3),128–136.

Byers,D.L.,&Meagher,T.R.(1992).Mateavailabilityinsmallpopulationsof plant specieswith homomorphic sporophytic self-incompatibility.Heredity, 68(4),353–359.

Castric, V., & Vekemans, X. (2004). Plant self-incompatibility in naturalpopulations:Acriticalassessmentofrecenttheoreticalandempiricaladvances. Molecular Ecology, 13,2873–2889.

Charlesworth, D. (1985). Distribution of dioecy and self-incompatibil-ity in angiosperms. In P. J. Greenwood, P. H. Harvey, & M. Slatkin(Eds.),Evolution—essays in honour of John Maynard Smith. Cambridge: Cambridge University Press.

Charlesworth,D.,&Charlesworth,B.(1979).Theevolutionandbreakdownof S- allele systems. Heredity, 43,41–55.

Collani,S.,Alagna,F.,Caceres,E.M.,Galla,G.,Ramina,A.,Baldoni,L.,…Barcaccia, G. (2012). Self-incompatibility in olive: A new hypothe-sis on the S-locus genes controlling pollen-pistil interaction. Acta Horticulturae, 967,133–140.

Collani,S.,Moretto,F.,Galla,G.,Alagna,F.,Baldoni,L.,&Muleo,R.(2010).Anewhypothesisonthemechanismofself-incompatibilityoccurringin olive (Olea europaea L.): Isolation, characterization and expressionstudies of SLG and SRK genes as candidates for a sporophytic self-incompatibilitysystem.Journal of Biotechnology, 150,502.

Cuevas,J.,&Polito,V.S.(1997).Compatibilityrelationshipsin`Manzanillo’olive. HortScience, 32(6),1056–1058.

DelaRosa,R.,Angiolillo,A.,Guerrero,C.,Pellegrini,M.,Rallo,L.,Besnard,G.,…Baldoni,L.(2003).Afirstlinkagemapofolive(Olea europaeaL.)cultivars using RAPD,AFLP, RFLP and SSRmarkers. Theoretical and Applied Genetics, 106(7),1273–1282.

DelaRosa,R.,James,C.M.,&Tobutt,K.R. (2004).Usingmicrosatellitesforpaternitytestinginoliveprogenies.HortScience, 39(2),351–354.

De Nettancourt, D. (1977). Incompatibility in angiosperms. Berlin:Springer-Verlag.

De Nettancourt, D. (1997). Incompatibility in angiosperms. Sexual Plant Reproduction, 10(4),185–199.

Díaz,A.,Martın,A., Rallo, P.,Barranco,D.,&De laRosa,R. (2006). Self-incompatibilityof‘Arbequina’and‘Picual’oliveassessedbySSRmarkers.Journal of the American Society for Horticultural Science, 131(2),250–255.

Díaz,A.,Martín,A.,Rallo,P.,&De laRosa,R. (2007).Cross-compatibilityoftheparentsasthemainfactorforsuccessfulolivebreedingcrosses.Journal of the American Society for Horticultural Science, 132(6),830–835.

Dreesen,R.S.G.,Vanholme,B.T.M.,Luyten,K.,VanWynsberghe,L.,Fazio,G.,Roldán-Ruiz,I.,&Keulemans,J.(2010).AnalysisofMalusS-RNasegenediversitybasedonacomparativestudyofoldandmodernapplecultivarsandEuropeanwildapple.Molecular Breeding, 26(4),693–709.

Eaves,D.J.,Flores-Ortiz,C.,Haque,T.,Lin,Z.,Teng,N.,&Franklin-Tong,V.E.(2014).Self-incompatibilityinPapaver:Advancesinintegratingthesignallingnetwork.Biochemical Society Transactions, 42(2),370–376.

El Bakkali, A., Haouane, H., Moukhli, A., Costes, E., Van Damme, P., &Khadari,B.(2013).Constructionofcorecollectionssuitableforasso-ciationmappingtooptimizeuseofMediterraneanolive(Olea europaea L.)geneticresources.PLoS One, 8(5),e61265.

Evanno,G.,Regnaut,S.,&Goudet,J.(2005).Detectingthenumberofclus-tersofindividualsusingthesoftwareSTRUCTURE:Asimulationstudy.Molecular Ecology, 14(8),2611–2620.

Farinelli,D.,Breton,C.M.,Famiani,F.,&Bervillé,A. (2015).Specificfea-tures in the olive self-incompatibility system:Amethod to decipherS-allelepairsbasedonfruitsettings.Scientia Horticulturae, 181,62–75.

Ferrara,G.,Camposeo,S.,Palasciano,M.,&Godini,A.(2007).Productionof total and stainable pollen grains inOlea europaea L. Grana, 46(2),85–90.

Gervais,C.E.,Castric,V.,Ressayre,A.,&Billiard,S.(2011).Originanddiver-sificationdynamicsofself-incompatibilityhaplotypes.Genetics, 188(3),625–636.

Gonthier,L.,Blassiau,C.,Moerchen,M.,Cadalen,T.,Poiret,M.,Hendriks,T.,&Quillet,M.-C.(2013).High-densitygeneticmapsforlociinvolvedin nuclearmale sterility (NMS1) and sporophytic self-incompatibility(S-locus) in chicory (Cichorium intybus L.,Asteraceae).Theoretical and Applied Genetics, 126(8),2103–2121.

Green, P. S. (2002).A revisionofOlea L. (Oleaceae).Kew Bulletin, 57(1),91–140.

Haouane,H.,ElBakkali,A.,Moukhli,A.,Tollon,C.,Santoni,S.,Oukabli,A.,…Khadari,B.(2011).GeneticstructureandcorecollectionoftheWorldOlive Germplasm Bank of Marrakech: Towards the optimised man-agementanduseofMediterraneanolivegeneticresources.Genetica, 139(9),1083–1094.

Heslop-Harrison,Y.,&Shivanna,K.R.(1977).Thereceptivesurfaceoftheangiospermstigma.Annals of Botany, 41(6),1233–1258.

Hiscock,S.J.,&McInnis,M.(2003).PollenRecognitionandrejectionduringthe sporophytic self-incompatibility response: Brassica and beyond. Trends in Plant Science, 8(12),606–613.

Husse,L.,Billiard,S.,Lepart,J.,Vernet,P.,&Saumitou-Laprade,P.(2013).A one-locus model of androdioecy with two homomorphic self-incompatibility groups: Expected vs. observed male frequencies.Journal of Evolutionary Biology, 26(6),1269–1280.

Igic,B.,Bohs,L.,&Kohn,J.R.(2006).Ancientpolymorphismrevealsunidi-rectionalbreedingsystemshifts.Proceedings of the National Academy of Sciences of the United States of America, 103(5),1359–1363.

Igic, B., Lande, R., & Kohn, J. R. (2008). Loss of self-incompatibility andits evolutionary consequences. International Journal of Plant Sciences, 169(1),93–104.

Iwano,M., &Takayama, S. (2012). Self/non-self discrimination in angio-sperm self-incompatibility. Current Opinion in Plant Biology, 15(1),78–83.

Jakobsson,M.,&Rosenberg,N.A. (2007).CLUMPP:A clustermatchingandpermutationprogramfordealingwith labelswitchingandmulti-modality in analysis of population structure. Bioinformatics, 23(14),1801–1806.

Kitashiba,H.,&Nasrallah,J.B.(2014).Self-incompatibilityinBrassicaceaecrops:Lessonsforinterspecificincompatibility.Breeding Science, 64(1),23–37.

Larsen,K. (1977). Self-incompatibility inBeta vulgaris L. Hereditas, 85(2),227–248.


Leducq,J.-B.,Gosset,C.C.,Poiret,M.,Hendoux,F.,Vekemans,X.,&Billiard,S. (2010). An experimental study of the S-Allee effect in the self-incompatibleplantBiscutellaneustriaca.Conservation Genetics, 11(2),497–508.

Levin,D.A. (1996).Theevolutionarysignificanceofpseudo-self-fertility.American Naturalist, 148(2),321–332.

Marchese,A.,Marra,F.P.,Caruso,T.,Mhelembe,K.,Costa,F.,Fretto,S.,&Sargent,D.J.(2016).Thefirsthigh-densitysequencecharacterizedSNP-based linkage map of olive (Olea europaea L. subsp. europaea)developedusinggenotypingbysequencing.AJCS, 10(6),857–863.

Marchese,A.,Marra,F.P.,Costa,F.,Quartararo,A.,Fretto,S.,&Caruso,T.(2016).Aninvestigationoftheself-andinter-incompatibilityoftheolivecultivars‘Arbequina’and‘Koroneiki’intheMediterraneanclimateof Sicily. Australian Journal of Crop Science, 10(1),88–93.

Matsumoto,S.(2014).Applepollinationbiologyforstableandnovelfruitproduction: Search system for apple cultivar combination showingincompatibility,semicompatibility,andfull-compatibilitybasedontheS- RNase allele database. International Journal of Agronomy, 2014(9),138271.

McClure, B. (2006). New views of S-RNase-based self-incompatibility.Current Opinion in Plant Biology, 9(6),639–646.

Mookerjee, S., Guerin, J., Collins, G., Ford, C., & Sedgley, M. (2005).Paternityanalysisusingmicrosatellitemarkerstoidentifypollendonorsin an olive grove. Theoretical and Applied Genetics, 111(6),1174–1182.

Nei, M. (1987). Molecular evolutionary genetics. New York: ColumbiaUniversity Press.

Ockendon, D. J. (1974). Distribution of self-incompatibility alleles andbreeding structure of open-pollinated cultivars of Brussels sprouts.Heredity, 33(2),159–171.

Ockendon,D.J. (1982).AnS-allelesurveyofcabbage (Brassicaoleraceavar.capitata).Euphytica, 31(2),325–331.

Pannell,J.R.,&Korbecka,G.(2010).Mating-SystemEvolution:RiseoftheIrresistibleMales.Current Biology, 20(11),R482–R484.

Park,S.D.E.(2001).TrypanotoleranceinWestAfricancattleandthepopu-lationgeneticeffectsofselection.,Dublin.

Perrier, X., Flori, A., & Bonnot, F. (2003). Data analysis methods. In P.Hamon,M.Seguin,X.Perrier,&J.-C.Glazmann(Eds.),Genetic diversity of cultivated tropical plants.Plymouth,UK:EnfieldSciencePublishers.

Pritchard,J.K.,Stephens,M.,&Donnelly,P.(2000).InferenceofPopulationStructure Using Multilocus Genotype Data. Genetics, 155(2), 945–959.

Sassa, H. (2016). Molecular mechanism of the S-RNase-based gameto-phyticself-incompatibilityinfruittreesofRosaceae.Breeding Science, 66(1),116.

Saumitou-Laprade,P.,Vernet,P.,Vassiliadis,C.,Hoareau,Y.,deMagny,G.,Dommée,B.,&Lepart,J.(2010).Aself-incompatibilitysystemexplainshighmalefrequenciesinanandrodioeciousplant.Science, 327(5973),1648–1650.

Seifi,E.,Guerin,J.,Kaiser,B.,&Sedgley,M. (2012).Sexual compatibilityoftheolivecultivar”Kalamata”assessedbypaternityanalysis.Spanish Journal of Agricultural Research, 3,731–740.

Seifi,E.,Guerin,J.,Kaiser,B.,&Sedgley,M.(2015).Floweringandfruitsetinolive:Areview.Journal of Plant Physiology, 5(2),1263–1272.

Serrano, I.,&Olmedilla,A. (2012).Histochemical locationofkeyenzymeactivities involved in receptivity and self-incompatibility in the olivetree(Olea europaeaL.).Plant Science, 197, 40–49.

Sijacic,P.,Wang,X.,Skirpan,A.L.,Wang,Y.Y.,Dowd,P.E.,McCubbin,A.G., … Kao, T.-H. (2004). Identification of the pollen determinant ofS-RNase-mediatedself-incompatibility.Nature, 429(6989),302–305.

Szpiech,Z.A.,Jakobsson,M.,&Rosenberg,N.A.(2008).ADZE:Ararefac-tionapproachforcountingallelesprivatetocombinationsofpopula-tions.Bioinformatics, 24(21),2498–2504.

Takasaki,T.,Hatakeyama,K.,Suzuki,G.,Watanabe,M.,Isogai,A.,&Hinata,K. (2000). The S receptor kinase determines self-incompatibility inBrassicastigma.Nature, 403(6772),913–916.

Tantikanjana,T.,Rizvi,N.,Nasrallah,M.E.,&Nasrallah,J.B.(2009).ADualRolefortheS-LocusReceptorKinaseinSelf-IncompatibilityandPistilDevelopment Revealed by anArabidopsis rdr6Mutation. Plant Cell, 21(9),2642–2654.

Trujillo,I.,Ojeda,M.A.,Urdiroz,N.M.,Potter,D.,Barranco,D.,Rallo,L.,&Diez,C.M.(2014). IdentificationoftheWorldwideOliveGermplasmBank ofCórdoba (Spain) using SSR andmorphologicalmarkers.Tree Genetics & Genomes, 10(1),141–155.

VandePaer,C.,Saumitou-Laprade,P.,Vernet,P.,&Billiard,S.(2015).Thejointevolutionandmaintenanceofself-incompatibilitywithgynodio-ecy or androdioecy. Journal of Theoretical Biology, 371, 90–101.

Vekemans,X.,Schierup,M.H.,&Christiansen,F.B.(1998).Mateavailabil-ityandfecundityselectioninmulti-allelicself-incompatibilitysystemsinplants.Evolution, 52, 19–29.

Vernet,P.,Lepercq,P.,Billiard,S.,Bourceaux,A.,Lepart,J.,Dommée,B.,&Saumitou-LapradeP.(2016).Evidenceforthelong-termmaintenanceofarareself-incompatibilitysysteminOleaceae.New Phytologist, 210, 1408–1417.

Wagenius,S.,Lonsdorf,E.,&Neuhauser,C. (2007).PatchAgingand theS-AlleeEffect:BreedingSystemEffectsontheDemographicResponseof Plants to Habitat Fragmentation. The American Naturalist, 169(3),383–397.

Weller,S.G.,Donoghue,M.J.,&Charlesworth,D. (1995).Theevolutionof self-incompatibility in flowering plants: A phylogenetic approach.Experimental and Molecular Approaches to Plant Biosystematics. St. Louis, Mo.: Missouri Botanical Garden ((Monographs in Systematic Botany, 53, 355–382.

Williams, J. S., Wu, L., Li, S., Sun, P., & Kao, T.-H. (2015). Insight intoS-RNase-based self-incompatibility in Petunia: Recent findings andfuturedirections.Frontiers in Plant Science, 6, 41.

Wright,S.I.(1939).Thedistributionofself-sterilityallelesinpopulations.Genetics, 24,538–552.

Wu,S.B.,Collins,G.,&Sedgley,M.(2002).Sexualcompatibilitywithinandbetweenolivecultivars.Journal of Hortical Sciences and Biotechnology, 77(6),665–673.

Wünsch, A., & Hormaza, J. I. (2004). Genetic and molecular analysis inCristobalina sweet cherry, a spontaneous self-compatible mutant.Sexual Plant Reproduction, 17(4),203–210.

SUPPORTING INFORMATION

AdditionalSupportingInformationmaybefoundonlineinthesupport-inginformationtabforthisarticle.

How to cite this article:Saumitou-LapradeP,VernetP,VekemansX,etal.Elucidationofthegeneticarchitectureofself-incompatibilityinolive:Evolutionaryconsequencesandperspectivesfororchardmanagement.Evol Appl. 2017;00: 1–14. https://doi.org/10.1111/eva.12457

https://doi.org/10.1111/eva.12457

elucidation of the genetic architecture of self

Documents