Mechanisms ofself-incompatibilit y and unilateral incompatibility in diploid potato {Solanum tuberosum L.) Promotor: dri rE . Jacobsen Hoogleraar ind eplantenveredeling , inhe tbijzonde r in degenetisch e variatie enreproducti e

Co-promotor: drM.S . Ramanna Universitair docent, Departement Plantenveredeling en Gewasbescherming Ronald Eijlande r

Mechanisms of self-incompatibility and unilateral incompatibility in diploid potato (Solanum tuberosum L.)

Proefschrift terverkrijgin g van degraa d van doctor op gezagva n derecto r magnificus van deLandbouwuniversitei t Wageningen dr C.M. Karssen, inhe topenbaa rt everdedige no pmaanda g 14septembe r desnamiddag s te 13.30uu r ind eAul a

'VA/ 3S'V'T Thisthesi s encompasses apar to fth e scientific researchcarrie dou t ondiploi d potato, at the former department ofPlan t Breeding, Wageningen Agricultural University. An international cooperation onself-incompatibilit y (SI)i npotat obetwee npartner si nItal y(Universit yo fSiena) ,German y(Ma x Planck Institute, Cologne) and TheNetherland s (KUN,Nijmege n and WAU, Wageningen) was focused on various aspects of SI. Thisjoin t project was supported by the European Community 'BridgeProgramme ' (BIOT-CT-900172).

CIP-DATAKONINKLIJK E BIBLIOTHEEK, DENHAA G

Eijlander ,R

Mechanisms ofself-incompatibilit y andunilatera l incompatibilityi n diploidpotat o {Solanum tuberosum L.) Thesis Wageningen Agricultural University- with references- with summaries in English and Dutch. Department of Plant Breeding, P.O. Box 386, 6700 AJ, Wageningen,N L

ISBN 90-5485-893-1

Cover: seeals oFig.5 ,pag e 66

Key words: Solanum tuberosum, Solanum verrucosum,self-incompatibility , self-compatibility, unilateral incompatibility, unilateral incongruity, 5-glycoprotein, S-RNase, sense, antisense, overexpression.

BibliographicAbstract :Thi sthesi sdescribe sth ecreatio nan dselectio no fdiploi d potato genotypes withwel ldefine d self-incompatibility (SI)reactions . Thecontributio n ofth e stylarproduct s ofth e incompatibility alleles, the ^-glycoproteins, is described for both the gametophytic self- incompatibilityreactio nan d theincomplet einterspecifi c crossingbarrie rtha texist sbetwee ndiploi d potato and its self-compatible relative, S.verrucosum. This barrier is called unilateral incompatibility or unilateral incongruity (UI). Complex interactions between incompatibility determining genes are described.

BIBLIOTHEEK LANDBOUWUNIVERSITEIT WAGENINGEN Stellingen

1 De aanscherping van de unilaterale incongruentie-hypothese door de S-locus uit te sluiten van de bijdrage aan interspecifieke incompatibilité«(Hogenboom , 1973)i sonterecht . DeU I hypothese heeft in deze vorm derhalve zijn tijd gehad (dit proefschrift).

2 DeRNase-activitei t vanS-glycoproteïne nspeel ti nhe tunilateral e incompatibiliteitssysteem bestaande tussenLycopersiconperuvianum e nL.esculentum een ondergeschikte rol (Rick, 1986; Kowyama et al., 1994; Royo et al., 1994;di t proefschrift).

3 Het "allele-specific dominant negative effect" van een RNase-defecte kopie van een werkzaam incompatibiliteitsallel zoalsbeschreve n door Mc Cubbin et al (1997) iso p zijn best co-dominantie.

4 Alle verwijzingen door de in dit proefschrift genoemde auteurs naar oudere literatuur m.b.t. de uitdrukking "wederzijdse afzwakking (mutual weakening)" betreffen integenstellin g tot wat wordt gesuggereerd niet de uitdrukking maar slechts het verschijnsel.

5 De "two-power competition" hypothese van Abdalla (1970) die beschrijft hoe S.verrucosum en S.tuberosumkruisingstechnisc h van elkaar gescheiden zijn, heeft een hoogantropopatisc h gehalte (= het vertonen van menselijke gevoelens).

6 CytoplasmatischeMannelijk e Steriliteitzoal sdi eoptreed t inhe tsystee m S.verrucosumx S.tuberosum (Abdalla &Hermsen, I97lb) isgee n probleem voor deveredelin g maar een oplossing.

7 Het gebruik van incompatibiliteit ind ehybride-rassen-producti e van zelfcompatibele Solanaceae is een achterhaald idee.

8 Detoekennin g van de soortstatus aanS.sucrense doorHawke s wordt ontkracht door de beschrijving van Hawkes (1989). (Hawkes &Hjertig, The potatoes of Bolivia. 1989).

9 Dedetecti e van 9 (negen!) verschillende incompatibiliteitsallelen in een diploïde Fl van Nicotiana glauca(Pandey , I98l) isee n sterke aanwijzing voor slaperigheid bij de referenten of hoge activiteit van specifieke paramutatie (sic). Dit pleit voor het vermelden van de referenten bij gerefereerde artikelen.

10 Demogelijkhei d vanxenotransplantati e zouvoo r het individu niveau zeer aantrekkelijk zijn, maar ongewenst voor de gemeenschap.

11 Het model zoals dat door de theoretisch natuurkunde Gerard 't Hooft werd gepresenteerd in het programmaNoorderlich t ("Kosmische anarchie",VPRO,28-12-1997 , 20:05 - 20:28) impliceert niet alleen zoals hij zelf alaangaf , predestinatie, maar ook dat de geest een functie isva n de materie.

12 De "Pensee"di e bekend staat als"He t Godsbewijs van Pascal"bewijs t alleen Pascal's opportunisme en zijn feilbaarheid in de kansrekening.

13 Slimheid verhoudt zich tot A.I.O.-schap alsS-homozygoti e tot incompatibiliteit.

Stellingen behorende bij hetproefschrif t "Mechanisms of self-incompatibility and unilateral incompatibility in diploid potato (Solanum tuberosumL) "

Ronald Eijlander Wageningen, 14septembe r 1998 Voorwoord

Ind ejare n zestig enzeventi gwer d eraa nhe t toenmalige Instituut voor Plantenveredeling van de Landbouw Hogeschool Wageningen veel onderzoek gedaan aan aardappel en aan aardappel verwante soorten, endi t gaat in feite tot op dehuidig e dag verder. In die tijd ise r veel bijzonder materiaal ontwikkeld. Een deel van dit materiaal heeft direct of indirect zijn weg gevonden naar aardappelveredelingsbedrijven, terwijl ander materiaal verder gebruikt werd op tal van andere onderzoeksinstellingen.Nakomelinge nva ndri ehee lbijzonder e aarappelplantenvonde nz ohu nwe g naar het Max Planck Instituut te Keulen, alwaar DrRichar d Thompson enmedewerker s opnieuw een aantal eigenaardigheden onder de loupe namen, maar dit keer met modernere, moleculaire technieken. De complicaties in de analyses waren aanleiding om dehul p in te roepen van zowel Prof. Dr Ir Jacobsen's onderzoeksgroep "Genetische variatie en reproductie" van de vakgroep Plantenveredeling als van de emeritus hoogleraar Prof.D rI r Hermsen, die nog steeds actief was op devakgroe p enberei d gevonden werd zijn kennis omtrent dit materiaal met anderen te delen. Dezesamenwerkin gresulteerd euiteindelij k inee nbi jd evakgroe pgeplaatst epositi evoo ree nA.I.O . binnenee nE.E.G . gefinancieerd internationaalproject .Dez epositi ewer ddoo rmi j ingevuld enhie r heb ik dan ook de afgelopenjare n met veelplezie r aan gewerkt.

Ditproefschrif t isee nweergav eva nslecht see ndee lva na lhe twer kwa ti ksame nme tvel e anderen aandi taardappelmateriaa l hebverricht .He ti svoo rsommigen ,nie ti nd elaatst eplaat svoo r mijzelf, misschien dan ook frustrerend temoete n zien dat zoveel werk niet inpublicatievor m het daglicht zalzien .D etegenslage n enachtera fonjuis t geblekenwerkhypothese n hebbenmi j enmij n studenten danoo kee nzeker e"faam "opgeleverd . Ikherinne rmi jno gd euitspraa k naar een student toe, toen een experiment wat normaal gesproken nooit mis gaat, maar nujammerlij k de mist in ging, die luidde: "ja , wat hadj e anders verwacht,j e werkt immers bij Ronald". Het heeft deze student er echter niet van weerhouden om ook zelf A.I.O. te worden. Maar het proefschrift is er dan toch gekomen.

Ik wil hierbij mijn promotor prof. dr ir. Evert Jacobsen bedanken. Evert, je hebt mij in de gelegenheid gesteld dit tochwe l erg leukeonderzoe k teverrichten . Jeheb t voortdurend de grote lijnen in het oog gehouden en je hebt er voor gezorgd dat e.e.a. hopenlijk toch nog enigszins begrijpelijk oppapie r is gezet. Onshemelsbred e verschil in stijl heeft ervoo r gezorgd datw ehe t nodige geduld met elkaarmoeste nhebben ,maa r wezij n er samentoc huitgekomen . Ikzelfbe ni n ieder geval tevreden methe tuiteindelijk e resultaat. Ikhoo p datji j end elezer s dat ook zijn. Verder wil ik mijn co-promotor Dr Ramanna bedanken voor zijn, zoals dat heet, dagelijkse begeleiding. Ik heb u, zoals al vele promovendi voor mij, leren kennen als aimabel, geduldig en onbaatzuchtig. Uvon d altijd tijd ommij n uiteenzettingen aant ehore n envanui t uw grote kennis meet edenken . Ikhe b inu niet alleen eenkundig e collega gevonden, maar ook eenvriend .

Prof. Hermsen heb ik reeds in mijn studententijd en ook nu tijdens mijn A.I.O.-schap leren waarderen alsdocen te nal smens .D ehe bme tu vele ,vel eure naa ntafe l doorgebracht, gebogen over papieren vol met afstammingen en genetische modellen, speculerend over wat er nu weer aan de handko nzijn . Uwkenni sva nhe tmateriaa l enhe tverschijnse l "incompatibiliteit" zijn voormi j van onschatbarewaard egeweest .U wbijdrage n aanmij n onderzoek hebben danoo k geresulteerd inee n co-auteurschap, waar umi j een grootplezie rme eheef t gedaan.

Dirk-Jan Huigen wil ik bedanken voor al het werk dat hij gemerkt en ongemerkt voor mij heeft verricht.Ji j coördineerde, samenme t demense nva nUnifarm , hetwer k ind ekassen .Toe n iko pd e vakgroep kwam, had jij het voorwerk al gedaan en kon ik zo aan de slag. Je hebt me veel organisatorische takenui thande n genomen enwaarschuwd ewannee r erwee r eenswa thoo i vand e overvolle vork afviel. En tot de promotie aan toe blijf je betrokken bij wat er moet gebeuren. Veertien september iswaarschijnlij k delaatst eda gdatj e nogvoo rm ei n dewee rbent . Bedankt.

Marja Schipperse nAnj a Posthuma,julli ehebbe nvoo rmi jhe tnodig ein-vitr ower kverrich t envee l werk uit handen genomen. Ditbetro f een aantal transformatie-experimenten, maar demeest e tijd isden ki ktoc hwe lgaa n zitten ind einstandhoudin g vand egigantisch e hoeveelheid basisklonen en transformanten. Hoewel niet alles heeft opgeleverd wat er van werd verwacht, zijn er toch leuke dingen uitgekomen. Een deel hiervan vindenjulli e in ditproefschrif t terug.

Voortswi l ik demense n vanTUPE Abedanken . Het lijkt erop dat ikmagisch e vingertjes hebo m PC'svas tt elate nlopen .He t inloggenonde rmij n eigennaa mwa svaa ka lvoldoend e omee nP Co p tiltt ekrijgen . Eenmailtj e aanjulli e viaee nniet-recalcitrant ecompute r was meestal voldoende om jullie te doen uitrukken. Jullie hebbenm e altijd naar tevredenheid uit debran d geholpen.

Dan heb ik nog met een groot aantal studenten samengewerkt. Hoewel demeest e van hun werk weinigteru gzulle nvinde ni ndi tproefschrift , wildi tnie tzegge nda tzi jnie tee nbelangrijk e bijdrage hebbengeleverd . Monique Mouwen wasreed sbezi gme t diverse analyses toen iko pd e vakgroep binnenkwam. Zij werd gevolgd door Mohammed Sohani, Bart Bronnenberg, Marlijn Vos, Jaap Kooyman,Este rAba dI Canter oe nWend yte rLaak .He twa spretti go mme tjulli esame nt ewerken . Ikhoo p dat dit wederzijds is. Eendee l van debelonin g zit in ieder geval in dit proefschrift.

Voortszij ndaa rd evel ecollega' sva nd elaboratori adi ei kwi lbedanken .O me ree npaa rt enoemen : Luuk Suurs,Ell yJansen ,Irm a Straatmane nMarja n Bergervoet. Ikhe bjulli e allemaal wel voor de voeten gelopen en met vragen en problemen lastig gevallen. Als goede collega's hebben jullie meegedacht enmeegeholpe n waar datnodi g was.I khe b ervee lva n opgestoken.

Binnen de vakgroep heb ik met vele mensen eenpretti g contact gehad. Dit betreft niet alleen de mensenwaarme e iko pd ekame rhe bgezete n ofgedurend e mijn schrijffase op deTer p lief en leed hebgedeeld ,maa roo kvel e anderemensen . De lijst islang , eno mhe trisico va nmense n vergeten tevermijden , wili kjulli ezonde r verdername n tenoeme n hierbij bedanken voor degezellig e tijd.

Tot slot wil ik hierbij mijn familie bedanken. Het ishaas t traditioneel, maar niet minder waar en oprecht.Mij n ouderswi li kbedanke nvoo ra lhe tmedeleve nwa tz ehebbe ngetoond .Mij n succesjes hebbenjulli e altijd meervreugd egebrach tda ni kwaar dvond ,maa rd etegenslage nbedrukte njulli e ook altijd meer dan nodig was.Kij k eens aan, eri see n proefschrift. Mij kinderen weten haast niet beter dan dat werken inhoudt dat pappa of "weg" is, of op zolder achter de computer zit, of dat het ervoor heeft gezorgd dat pappa weer "heel erg moe" is. Jullie moetenm emaa rgeloven ,he ti smeesta lhee l ergleuk ,tenminst e alsj ehe twer k doetwaa rje , zoals ik,zel fvoo rheb tkunne nkiezen .I khoo pda t ike ri nmij n volgendebaa nmee rvoo rjulli eka n zijn. Ente n slottemij n partner, Leontine.J ewee t wathe t is omt epromoveren , maar ookbi jjo u zal de vertwijfeling wel eens hebben toegeslagen, alhe bj e datnooi t zo latenblijken . Jekwa mwe l eens met de vraag "moet je nu echt weer naar het lab/overwerken/achter de computer/vannacht doorwerken/humeurig zijn/etc", maar over het algemeen heb je met een bewonderenswaardige tolerantie mijn A.I.O.-trekjes geaccepteerd. Ik hoop dat ik dezetrekje s nu achter me gelaten heb. Bedankt. Contents

Chapter 1.Introduction . 9

Chapter 2. Selection ofvigorou s and fertile S-homo-an dheterozygou s tester 21 clones from self-incompatible diploid potato,Solanum tuberosum L.

Chapter 3. Manipulation of self-incompatibility in diploid Solanum tuberosum L. 41 using sense and antisense constructs ofS-RNas egene s

Chapter 4. Expression andinheritanc e of self-compatibility and self-incompati- 55 bility after crossing diploidS. tuberosum (SI)wit hS. verrucosum (SC)

Chapter 5. Contribution ofth e S-locust oUnilatera l Incompatibility when 73 crossingS. verrucosum(SC )wit h S.tuberosum (SI)

Chapter 6. General discussion 87

Appendix 1 93

References 97

Summary 107

Samenvatting 110 Chapter 1 General introduction.

Floweringplant s canpropagat e themselves intw oways :vegetativel y and sexually. Vegetative propagation leadsnormall y toindividual s that aregeneticall y identical to the original plant. Well- known exampleso fvegetativ epropagatio n areth erunner sproduce d by strawberry (Fragaria)an d the tubers produced by potato. Variation in nature on vegetative propagation is endless, various plant organs canb e used toproduc e clonal offspring. Even seeds,normall y theresul t of a sexual process,ca nproduc eclona loffspring : forinstance ,twin-embryo s (polyembryony) inseed so fcitru s often contain cloneso fth emothe rplan t(e.g .Webber , 1948)an dapomixi si nblu e grasses (Poasp ) isanothe rgoo dexampl eo fbypassin gsyngamy .Vegetativ epropagatio ni sattractiv ewhe nimportan t characteristics havet ob ekep t together, butthi s limitsth erespons e ofth eplan t to changes inth e environment: response depends completely on the genetic information present in the genotype dealing with. Sexualpropagation , involving fertilisation andgeneti crecombination ,provide splan t speciesno t onlyanothe rmechanis mo fspreadin gi nth eenvironment ,bu tals oth epossibilit yt oadap tt ochange s in the environment ort o invade different environments. Self-fertilisation orhybridisatio n with a closerelativ e(inbreeding ) limitsth evariatio ni nth eoffsprin g andthu s limits adaptiveness. Duet o inbreeding, accumulated fitness-negative characteristics whichmigh t be recessive,hav e ahighe r chance of becoming homozygous, thus reducing the fitness of those plants. Another effect of inbreeding, that can be advantageous, is that it also purifies species from apar t of those fitness negative traits and can fix positive gene combinations. Combined with a certain amount of outcrossing,whic h ascertains also the adaptiveness, inbreedingprove d tob e asucceesfu l strategy for somespecies ,an dca nb e found inman yimportan t cultivated crops,suc h aswheat ,barley , peas andbeans , which arecalle d self-pollinators. Inbreeding is,however , for many plant species a risky strategy to rely on. There areman y mechanisms topreven t or limit selfing. Somear ebase d on floral morphology, otherso ndifferenc e ofmaturatio n time ofmal e and female reproductive organswithi n individual flowers. A number of these self-fertilisation impeding mechanisms are well visible and recognisable, based on temporal or spatial separation of male and female reproductive organs. Dioecy, which means that plants carry either male or female flowers, is such amechanism , well known from, for instance, willow (Salix).Separatio n mechanisms, such as protogyny (temporal separation, e.g. Victoria amazonica, anonymus),monoec y (male and female flowers onth e same plant, e.g.in Zea mais), and hermaphrodity (flowers are male and female at the same time, e.g. potato or cabbage) do exist that areles s strict then dioecy.

Self-incompatibility systems Someo fth evariation s onth eseparatio nmechanism s asdescribe d abovear eeffective , butcanno t alwaysavoi dhig hlevel so fself-fertilizatio n orcross-hybridizatio n withclosel y related genotypes, thus leading to inbreeding. In many cases, undesired selfing does not lead to fertilization, though, due to specific pre- 10 Chapter 1 fertilization barriers. Barriers against self-fertilisation can be found at many places, e.g., on the stigmatic surface ( e.g., Cruciferea, likeBrassica oleracea), at various places in the transmitting tract in the style (e.g., various Solanaceae, like Petuniahybrida an d Solanumtuberosum), i n the ovary (e.g., Beta vulgaris) or even in the ovulejus t before syngamy (e.g., Theobromacacao). Mango (Magnifera indica) shows fertilization even after selfing, but the resulting zygotes die approximately twoweek spost-fertilizatio n (Sharmaan d Singh, 1970),althoug h it canb e disputed whether orno t thisphenomeno n reallybelong st o incompatibility. Someo fthos e incompatibility characteristics appeart ob ecorrelate dwit haspect ssuc ha spolle nbein gbi -o rtrinucleate ,dr yo rwe t stigmasurface , etc.Thes e aspects are extensively reviewed andexemplifie d in themonograp h of DeNettancour t (1977) and recently by, for instance, DeNettancour t (1997), Kao and McCubbin (1997) and Nasrallah (1997). It is known that the sporophytic incompatibility system (to be explained later)contain sbot hdiallelli can dpoly-alleli csystems .Th eheteromorphi c system appears tocoincid ewit hdialleli csystems ,whic hmean stha tdifference s inflowe r morphologyreflec t which types are intercompatible and which are not. The homomorphic incompatibility systems (sporophytic and gametophytic) do not betray their intercrossability by their morphology. The barriers that are active in plants with dry stigmata are usually on or directly under the stigma, whereasplan t specieswit hwe t stigmatausuall ydispla yth ebarrier si nth estyl eor , less frequently, eveni nth eovaries .Mos tgametophyti c SIsystem sdispla ywe tstigmat aan dbi-nucleat epollen .Dr y stimata and tri-nucleatepolle n areusuall y found in sporophytic systems.

Sporophytic Self-Incompatibility. Inth esporophyti c self-incompatibility system (SSI)th egenotype s ofbot h thepolle n parent and thepolle n recipient (the sporophytes) determine whether acombinatio n iscompatibl e ornot . The pollen (the gametophyte) reflects the genotype of thepolle n donorbu t not the actual genotype of thepolle n itself.

Sporophyticheteromorphic systems. In the heteromorphic self-incompatibility systems there exists an association between incompatibilitygroup san dflora l morphology:incompatibilit ybehaviou rdepend so nth ephenotype . Thisi sbelieve dt ob ealway s sporophytic (Pandey, 1970).Distyli c andtristyli c systems havebee n described forvariou splan tspecies .A heterodistyli c self-incompatibility systemha sbee n described, forinstance ,fo rprimros e(Primula) orHypericum . Flowerswit hlon ganther san dshor tstyle s(Pin) , canonl y fertilize plantswit h long styles and short anthers (Thrum),an dvic eversa . Offspring will segregatei na 1: 1 ratio ofPi nan dThrum . Here selfing isexcluded ,bu t full sibmatin g ispossibl e in 50 %o fth ecases . Itha sbee nreporte d for severalprimros e speciestha t linkagebetwee n S-morphology and actual crossabilityca nb ebroke no rdisrupte d (Ernst, 1932,1936,reviewe db yD eNettancourt , 1977;Shar ­ ma and Boyes, 1961). Some of thebiochemica l aspects of the system have been characterised by Heslop-Harrison et al (1981) and Shivanna eta l(1981) . Generalintroduction 11

Sporophytichomomorphic systems Sporophytic homomorphic self-incompatibility hasbee ndescribe d forth eCrucifera e anda tleas t five otherfamilie s (Bateman, 1955;Charlesworth , 1988),bu t itha sbecom e clear that sporophytic systemsca ndispla ycharacteristic stha tar enormall yfoun d ingametophyti cself-incompatibilit y sys­ tems and vice versa (see also review by De Nettancourt, 1997). SSI is, as mentioned before, characterised by the fact that the interaction between pollen and stigma/style depends on the dominancerelationship sbetwee nth e^-allele so fbot hparenta lplant s (thesporophytes) ,th epolle n (the gametophyte) carryingth einformatio n ofth epolle n donor inth epolle n coating (Stephenson etal. , 1997).Th egenotyp eo fth epolle ngrai nitsel fi so fn oimportanc e forth ereactio ni n SSI(Fig . la), thisi ncontras t with gametophytic self-incompatibility (GSI),wher eth e genotypes of the style andtha t ofth epolle ngrai n itself determine whether acombinatio n iscompatibl e or not (Fig. lb). Insuc ha sporophyti c system thefractio n ofcompatibl epolle ni n apopulation , isa function ofth e numbero fallele spresent ,a swel la sdominanc erelationship sbetwee nS-allelele si npolle nan dstyle . Polyallelyresult sthe ni nnumerou scompatibl ecombinations ,whic hi sincrease dwhe n5-allele sca n bedominan t over otherS-allele s(e.g. ,SI overS2 i nth epolle nphenotype ,a si nFig . 1,mos tright combination). Mosto fth efundamenta l researcho nth eone-locu smulti-alleli csporophyti c self-incompatibility system is done within the Brassicaceae (reviewed by Nasrallah, 1997), that belong to the Cruciferae. Brassica displays aone-locus ,multi-alleli c system, the genetics of itbein g elucidated by Bateman (1955). He described a single S-locus that segregated Mendelian. Due to the incompatibility mechanism,S-homozygote sar epossibl ebu twil lb ever yrare :plant sar ei ngenera l heterozygous.Recom - bination between different alleles would intheor y lead to arapi d increase

Fig. 1. Compatible and incompatible combinations in a sporophytic one-locus self-incompatibility (SSI) system (top) and SSI a one-locus gametophytic self-incompatibil­ ity(GSI ) system (bottom). Some dominance relationships between S-alleles combina­ tions are shown (SSI, top). The pollen grains are genetically of the genotype 5/ or S2. The pollen grains in the sporophytic -•S1S2- S2S3^ system are coated with both SI and S2 pollen: SI = S2 51 - S2 determinants. The pollen recipients are, jtyle : SI - S2 52 - S3 from left to right: S1S2, S2S3, S3S4 and S2S3 (top) or S1S2 (bottom).The gameto­ phytic incompatibility shows the effect of GSI two different alleles present in the same pollen grain, as produced by polyploids (bottom right). The interaction between different 5-alleles (known as competitive action or mutual eakening) causes self- compatibility, a phenomenon not known from the sporophytic system, but here - (inversed) dominance relationships between the 5-alleles can also bring about (self-) compatibility (top right). 12 Chapter 1 ofth enumbe ro fallele s andperhap s event oa break-dow n ofth e SIsystem .Althoug h amultitud e ofS-allele sha sbee n found (Nou etal. , 1993;Brac ee tal. , 1994),frequen t recombination onth eS- locusseem sno tt ob eth ecase .Th ereaso ni stha tth eS-locu scomple x (alsocalle dS-haplotype )ha s been shown to contain a subset of genes in close linkage (S-locus complex, Fig. 2 top), that maintainfunctiona l specificities andd ofunctio n asa se t(se eals orevie wb yNasrallah , 1997).Tw o genes,importan t forth estyla r SIreaction ,hav ebee ninvestigate d extensively, andals oothe rgene s linked to those two genes are investigated on their contribution to the pollen-style interaction, leading to either acompatibl e or anincompatibl e reaction.

•- 200 kb

K- SLG SRK Ligand f Ligand, pollen borne

ff • Cell wall •ifff • Plasma membrane AQ papillär cell

Fig. 2 The S-locus and a model for self-incompatibility inBrassica. Top: a diagram showing the genes that code for the stigmatic receptormolecule s SLG (

The two stigma expressed 5-locus genes that are required for an inhibition of self-pollen, are bothhighl ypolymorphic . Oneo fthes e genes codes for the so-called S-locusglycoprotei n (SLG), whichi sa solubl ecel lwall-localize d glycoproteinwit ha molecula rweigt ho f55-6 5k D( Nasralla h andNasrallah , 1984;Nasralla h etal. , 1985,1987),highl ypolymorphi c andusefu l for5-phenotyp e identification. Thesecon dgen ecode sfo ra n5-locu srecepto rkinas e(SRK), whic hi sa receptor-lik e kinase that spans themembran e (Stein et al.,1991,1996) .SLG sequences arehigl y homologous with sequenceso fth eextracellula r (S)domai no fSRK. I t is speculated that early inth e evolution of theS-haplotype sSLG arose from SRKb y duplication (Tantikanjana et al., 1993). Quite some (pseudo-) genesbelongin gt othi s family havebee n found tob eclustere d and linked to theS-locu s (Suzukie t al., 1997),althoug h alsounlinke d related geneswer edetecte d that might play arol e in the SI process (Luu et al., 1997). Other features of SLG and SRK led to the belief that the hypothesized S-locus encoded pollen Si-determinant is a ligand for the receptor. SRK and SLG should bind to different sites of the samepolle n ligand, thus precipitating an intracellular (Ca2+ dependent) phosphorylation cascadetha t results inth e arrest of self-type pollen (Fig.2 ,bottom) . The sense and anti-sense approach for elucidating more of the functions of SLG and SRK is seriouslyhampere db yth ehig h sequencehomolog ybetwee nthos etw ogenes .Sens ean d antisense inhibitionwil li nmos t cases affect both genes,wherea s onlyon e effect washope d for (Conner et al., 1997).Tha tbot h genespla y ake yrol ei n SI,i sundisputed , however. Although thepolle n component is still unknown, some candidate genes and products have been found. Yue ta l(1996 ) found inB. napus twogene slocate d inbetwee nSLG an dSRK, one of them (SLL,, S-locuspolle n ligand 1)bein gS-locu san d SIspecific . Its expression was only detectable inanthers . Itwa s deduced thatth eSLL l proteinwa s2 o r3 kDa ,bu tn orelate d sequences could be found inth edatabases . Stephenson et al(1997 )analyse d protein fractions from SIpolle n from B. oleracea anddetecte d watersolubl ecomponents ,wit ha M ,<. 10kDa .Fro mthis ,a basic ,cysteine - richprotei ncoul db eisolate dtha tbelong st o thefamil y ofPolle n CoatProtein s classA (PCP-A) , oneo fwhic hi sknow nt obin d to stigmatically expressed components ofth eS-locu si nBrassica. PCP-Al is regarded asa candidat e for playing arol e in SI andperhap s also in aspecifi c type of interspecific incompatibility: unilateral incompatibility. Unilateral incompatibility in Brassica might berelate d to self-incompatibility (Hiscock andDickinson , 1993).

GametophyticSelf-Incompatibility Inth egametophyti cself-incompatibilit y system(GSI )th egenotype so fth epolle n (gametophyte) and the pollen recipient (the sporophyte) determine whether orno t a combination is compatible (Fig. lb). In the SSI system the genotype of thepolle n itself is of no importance for SI,bu t the information about the pollen donor, carried at the outside of the pollen grain, is (fig. la). This difference in information supply by the pollen forms the basis of distinguishing GSI and SSI. Nevertheless,classification s areno talway s clear-cut and GSI species can have other features that aremor ecommo ni nth eSS Igrou p(e.g .rye ,Wehlin g etal , 1994) orvic eversa .Thus , also within thegametophyti c self-incompatibilty system several distinct groups canb erecognized . Based on 14 Chapter 1 thenumbe r ofloc i involved, thenumbe r ofplan t families showing one-locus GSI, is presumably largertha nthos efoun d for multi(>2 )loc iGSI .Th etw o loci system has been found so far inonl y four families (DeNettancourt , 1977).

PolygenicGSI. Withinth e heteromorphic SSIclasse s canb e identified visually,whic h facilitates the analysis ofeve nth emult i locussystem .On elocu sGS Isystem s arerelativel y easy to identify, because the number of allelecombinatio n permutations is quite limited.Nevertheless , also someo fth emor e multi locus GSI systemshav ebee n analysed. Lundqvist (1956, 1990, 1991)report s complex SI systemswit h threeo r four loci inth egener aRanunculus, Beta and Lilium. In the grasses a less complex two-loci sytem with dry stigmata has been found. In rye(Secale cereale),th etw ounlinke dloc iS an dZ (Lundqvist , 1956)hav ebee nreporte dt ocontro lth esystem . When for both loci the alleles inth e style arematche d bythos ei n thepollen , an incompatibility reactionoccurs .Unti lrecently , forry eneithe r female San dZ product sno r incompatibility-related products could be identified (Tan and Jackson, 1988).I tha sbee n shown, though, that Ca2+ and kinaseactivit ypla y arol ei nS Io fry e(Wehlin g et al., 1994) andher etoo, stigm apapillä r ligands are expected to be involved. From the grassPhalaris coerulescens pollen S-alleleswer e cloned, from which the deduced amino acid sequences shared homologies with thioredoxins (Li et al., 1994).

One-locusGSI systems with a drystigma and without S-RNases: Poppy. Fieldpopp y(Papaver rhoeas) ha sextensivel ybee n investigated onth eunderlyin gmechanism s of SI. The incompatibility reaction of poppy is determined by apolyalleli c one locus system. The incompatibilityreactio noccur so na dr ystigm a(Lawrence , 1975;Lawrenc ee tal. , 1978).Fro mthi s plant species stigmaticglycoprotein swer eisolate d and identified, andthe y cosegregated with the ^-alleles (Franklin-Tong et al., 1989). Subsequently, the cDNA of the stigma papillär SI glycoproteinwa sclone dan dsequence d(Foot ee tal. , 1994).T odate ,th esequenc ei sdifferen t from anyknow ni nth eBrassicacea e orSolanaceae .Th e^-glycoprotein sd ono tposses s RNase activity, whichth e5-glycoprotein so fth eSolanacea e dohav e (Franklin et al., 1995; McClure et al.,1989). The ^-glycoproteins, that proved to be S-specific, are believed to adhere to (yet unknown) receptors, thus eliciting Ca2+ (Franklin-Tong et al.,1993 , 1995), which results in a cascade of phosphorylation of specific proteins, in which Ca2+ dependent protein kinases and inositol triphosphate mayb e involved (Franklin-Tong et al., 1995, 1996).A n incompatible combination of ^-alleles results in an increased Ca2+dependen t phosphorylation of at least two 26 kD pollen proteins(Rud d etal. , 1996)an di tals ocause sth eslow-movin gcalciu mwave ,regulatin g thepolle n 2+ tubegrowth ,t o show"rapi d anddramati c alterationsi n [Ca ](withi n afe w secondso fchallenge" . An unusual high peak is reached, followed by abreak-dow n of the tip-focused [Ca2+]; gradient (Franklin-Tong etal. , 1997).Finally ,th epolle ntub egrowt h isinhibited . Amode l for some of the elucidated interactions involved inpolle n tube growth and inhibition ispresente d in figure 3. General introduction 15

direction of Ca wave

Fig. 3. Model for thepropagatio n ofth e Ca2+wav e inpopp y (Papaver rhoeas)polle n tubes. The normal growth of pollen tubes of P. rhoeas is regulated by a slow moving calcium wave propagated by inositol 1,4,5- triphosphate. Increases of [Ca2*],wil l activate Ca2+sensitiv e phosphoinositidase C (PIC), which will then 2+ hydrolyze specifically the membrane lipid PIP.Th e now raised Ins(l,4,5)P3wil l stimulate release of Ca 2+ from Ins(l,4,5)P3 sensitive intracellular stores. Continued slow Ca wave s can be generated. It is hypothesized that the binding of a stigmatic S-protein (S) to a membrane receptor initiates a signal transduction chain in which Ca2+ dependent phosphorylation plays an important role. The normal Ca2+ 2 gradient is disturbed. Increases of Ins(l,4,5)P3 and the consequent elevation of [Ca *]; , inhibition of

phosphoinositide (PI) turnover and inhibition of Ins(l,4,5)P3 binding to itsrecepto r lead to an inhibition of pollen tube growth. With small modifications, from Franklin-Tong et al, 1996.

One-locusGSI systems with a wet stigma and with S-RNases: Solanaceae Single-locusgametophyti c self-incompatbility isbelieve dt ob eth emos tcommo n incompatibil­ itysyste mwithi nth esolanaceou s species,althoug hther emigh texis t alsosolanaceou s speciestha t aregoverne db ya two-loc isyste m(e.g .Abdall aan dHermsen , 1971).A remarkabl efeatur e ofmos t SIsystem sis ,tha t SIseem st ob ever ystabl ean dresist st oa grea texten tth e spontaneousmutation - induced turn-over intoself-compatibility . Fromth epopulation-geneti c pointo fvie wne wS-alleles , and especially self-compatibilizing alleles,woul d relatively easily accumulate in apopulation . A large number of natural S-alleles that are found, seems to conflict with low frequencies of mutantswit h adisrupte d S-locus.Natura l selection against thosemutant s may beon ereason ,bu t difficulty inrecognizin gthos emutant sma yb eanother .Because ,i ngeneral ,importan t information canb egaine dfrom th eanalysi so fdevian tgenotypes ,spontaneou so rinduce dmutant sar eregarde d as highly valuable research material. Induced mutations facilitate a rational search for those mutants. Mutation studies (by means of chemical mutagens and irradiation), in order to create point mutations, deletions,duplication s andtranslocations ,resulte d innearl y all cases in self-compati­ bility that was either pollen-borne or style-borne. A change in specificity (Van Gastel and De Nettancourt, 1975; VanGaste lan dCarluccio ,Va nGaste l 1976;se eals omonograp hb yD eNettan ­ court, 1977)coul dno tb eshown ,o rca nb eexplaine dno wotherwis eb y applyingth e accumulated knowledge about the organisation ofth e5-locus .Thi shold s als otru e for nearly all the cases of believed-to-bepollen-par t mutations. 16 Chapter 1

S-SPECF1CITY PART STYLE PART POLLEN PART

X X X Fig. 4 Envisage of Lewis' hypothesis (1949,1960) of the WW WW WW tripartite structure of the 5-locus, and the x O * interrelationship of S-/ocus gene products. The model shown here assumes that each part of the S- — o locus encodes a different protein, and self-recog­ nition resultsfrom th e interaction of anS specificity part expressed inbot h pollen and pistil. Specificity parts form with pollen- or style activity parts the receptor-ligands. Interactions ofidentica l ^-specifi­ cities inpolle n and styleresul t inpolle n tube growth arrest (with minormodifierions , from Sing andKao , 1992).

Anelegan tmode lo fth e5-locu sa spropose db yLewi s(1949,1960) ,wa sth eso-calle d tri-partite structure,tha t could explain satisfactorily mosto fth eresult so fth e aforementioned studies. In this model (Fig.4) ,th e5-locu sconsiste d of an S-specificity- (identity), astylar - and apolle n activity part,al li ntigh t linkage.Th especificit y part gaverise t o aspecifi c proteini nbot hpolle n andstyle , whereasth estyl ean dpolle npart swer especificall y expressed instyl ean dpolle nrespectively .Th e combinationo factivit yan didentit ypart sresulte d ina specifi c receptor-ligan dpair ,causin ginhibi ­ tiono fth epolle n tubewhe npolle n and stylematche d in specificity. Therear efe w reportso na chang eo fth especificit y (changeo fS-allel especificity) , for instance after anther-culture (Ramulu, 1982) or inbreeding (Maheswaran et al., 1986; Kheyr-Pour and Pernes, 1986).Som eo fthes eresult sma yb eexplaine db yth eexpressio n of accumulated modifier genes(polygenic )tha t canals obrin gabou treduce d self-incompatibility (pseudo-compatibility) up toa leve lo f self-compatibility (e.g.Henn y andAscher , 1976;Robacke r and Ascher, 1978).Eve n theappearanc e ofmonogeni c pseudo-compatibility geneswit h strong effects cannot be excluded (see also introductions byDan a andAscher , 1986;Lied l andAnderson , 1994). Mutations of SI are, as stated, a valuable source for research. Olsder and Hermsen (1976) detected both a self-incompatible (G609) and two self-compatible dihaploid potato genotypes (G254 and B16) with a high degree of male and female fertility. In successive studies on this material (Hermsen, 1978a, 1978b, 1978c), the underlying genetics was analysed. The self- incompatibility systemwithi nth epopulatio nbase d onthos ethre eclone swa sver yreliabl e for its expression. Based on a complete diallel crossing scheme, four 5-alleles were identified. The self- compatibility was explained by a putative translocation of the pollen-part of the SI allele from chromosome 1 tochromosom e 12(Hermsen , 1978a ;Hermse ne tal. , 1973,1978b).Heterozygosit y General introduction 17 forS-allele si nth epolle n(e.g. ,i n2n-polle nan dpolle n from polyploids;fig . lb)woul dthe nbrin g about "competitive interaction" (Lewis, 1947), nowadays known as "mutual weakening". This materialwa soriginall y analysedb y classical crossing experiments;biochemica l identification of ^-groups was unsuccessful until the late eighties. Hermsen's material formed the basis for an extensive study onmolecular , biochemical andbiologica l aspects of SIi n diploid potato (see for instanceKirch , 1993; VanEldik , 1996;L ie t al, 1994). The^-glycoprotein si n solanaceous specieswer e shown to co-segregate withth eS-phenotype s (e.g.Kirch et al., 1989) and different S-classes could easily be identified. These ^-glycoproteins couldb etrace d extracellularly inth e sametissu e (stigma, transmitting tract ofth e style and even in a single cell layer) where the SI reactions occurred (Anderson et al., 1989). These tissues showed also an accumulation of corresponding mRNA (Cornish et al., 1987). Within the solanaceous species,a whol e rangeo fS-allele sha smolecularl ybee n cloned. Kirch (1992,1995) isolated, from the aforementioned diploids, molecular clones of SI and S2 alleles. The translocation hypothesis for AS7was investigated byThompso n et al(1991) ,usin g RFLP analysis ofplan tmaterial ,codin g forthi stSl withth eclone dstyla rpar to fSI. Wit hthi sapproac h itshowe d to be impossible to discriminate between plants with and without tSl and subsequently a good candidate for theS-polle npar to rotherwis e ausefu l Si-interacting toolcoul dno tb ecloned . Based onthes e data, aswel l aso n sequence data,th etri-partit e structure ofth e5-locus , aspropose d by Lewis (1961) hadt ob erejected . Thetranslocatio n hypothesis for tSl couldb emaintaine d under theassumptio n thatth etranslocatio n would have involved onlyth epolle npar tbu tnothin g ofth e analyzed stylar expressed SI-fragment. The pollen component plays an essential role in the elucidation of the SI mechanism in the Solanaceae. It was showntha tth e^-glycoprotein shav eRNas epropertie s (McClure et al., 1989) and aretherefor e alsocalle dS-RNases . Thisle dt oa rang eo f experiments dealingwit h sensean d anti-sense transformations and transformations with coding regions of ^-alleles, modified for RNase properties or for presumed identity determining stretches (see for references chapter 5). Based on some ofth e information gained,tw o mechanisms wereconsidere d for the contribution ofth eS-glycoprotein . Onemode l(Fig .5 ,left ) isbase do nth eassumptio ntha tth epolle npar tcode s for amembran erecepto r that isspecifi c inth euptak e ofth e5-RNase ,th eothe r one isbase d ona non-specific uptake combined with aspecifi c inhibition ofnon-sel fribonucleases (Fig . 5,right). Bothmodel sresul ti nonl yon especifi c ^-glycoproteinbein g activei nth epolle ntube . Identifica­ tion of the membrane receptor (model 1) or the ribonuclease inhibitor (model 2) will play an important rolei n unraveling the SImechanism . Todate ,man yfactor shav ebee nfoun dtha tpla ya rol ei npolle ntub egrowt han dpolle nviability , someo fthe mbein g essential for asuccessfu l fertilization, but noneprove d tob eth e long-sought pollenS Ifactor .Recently ,however ,Ca 2+-dependen tprotei nkinase shav ebee nisolate d frompolle n tubes ofNicotiana alatatha t seem to play arol e in the SIreaction , presumably shortly after the uptakeo fth eS-RNas e( Kun z etal , 1997).Thi sdiscover yi si naccordanc ewit hth e important role ofphosphorylatio n inth e SIrespons eo fBrassica, poppy andry e(se eabove ) andwil l presumably bever y helpful in afurthe r elucidation ofth e SIpolle n pathway. 18 Chapter 1

• 0 m 1 i 0 S2polle ntub e S2polle ntub e 0 o o- • • a= 0- 0 O • -o •" • / 1 • i S2 RNaseenter san d ƒ \ inhibits pollentub e / T T, \T*l -o o o .^_>* o - yVJ^ •" • 0 • S2 pecificity domain O S1 RNase O S1 specificity domain • S2 RNase •• ribonuclease catalytic domain =C pollen S2transmembran e receptor gr pollen S2 ribonuclease inhibitor

Fig. 5. Two models for self-incompatibility interactions inth e one-locus GSI system of the Solanaceae.

Left: model based on the assumption that the pollen Right:mode lbase d on the assumption that the pollen S- S-allelesencod emembran e receptors.Specifi c uptake alleles encode ribonuclease inhibitors. Here, all S- ofth eS2 RNas e (^-glycoprotein)b yth e correspond­ RNases canb e transported over the membrane.Polle n S- ing S^-receptor occurrs. SI RNase can only be allele ribonuclease inhibitorshav e twobindin g domains: transported over the membrane by an SI receptor, onebindin g indiscriminately toth eribonucleas e activity which is absent here.The S2 RNase degrades non- domain of theS-RNase s and one binding specifically to specifically therRN A and/o r themRNA ,resultin g in thespecificit y domain ofth ecorrespondin g S-RNase.A s inhibition of protein synthesis and finally in pollen a result, in this example, only the self Si-RNase is tube arrest. Pollen tubes carying other S-alleles that capable of degrading pollen RNA, whereas the (pollen SI or S2 do not transport the S-RNases over the non-self) SI-RNas e is inactivated. With minor modifica­ membrane and are hence not arrested. tions, from Kao and McCubbin (1997).

Unilateral incompatibilityin the Solanaceae It is frequently found that in interspecific hybridisation between a self-compatible and a self- incompatible species,fertilisatio n ispossibl ei non edirectio nonly .Thi sphenomeno n is, therefore, calledUnilatera l Incompatibility (UI)an di sfoun d throughout theincompatibilit y systems.I twas , according to DeNettancour t (1977), first defined byHarriso n and Darby (1955) but inth e early days described by many others (e.g. Anderson and De Winter, 1931; Mather, 1943;Lewi s and Crowe, 1958). In general, the SI species can be used successfully as a pollinator, but not as a pistillate parent whenpolle n from aS Cparen t isuse d (SC x SI- * Fl; S I x SC -* -). Because of the strongcorrelatio n between oneparen t being SC and another being SI,ther e hasbee n astron g opinion among many researchers that the 5-locus is involved in this process. Lewis and Crowe (1958) formulated the dual function hypothesis for the S-locus, and described an evolutionary General introduction 19

pathway for thedevelopmen t from SIt o SC,finally resultin g inUI .Th etwo-power s competition hypothesis,describe d forU Ibetwee n the SIdiploi dpotat o andth eS Crelativ eS.verrucosum (ver), isbase d on this ando nth e co-evolution of sympatric SCan d SIspecie s (Abdalla, 1970;Abdall a and Hermsen, 1971, 1972; Abdalla, 1974). The presence or the evolution of cytoplasmic male sterility in hybrids derived from SC x SI, plays an important role in this. In this hypothesis, additional geneswit hvariou s alleles areintroduced ,tha t areno tnecessaril y located at theS-locus . Atth e sametime ,researc hi nothe rcrop sindicate d thatth eS-locu sdi d notpla y arol e at all inUI . Hogenboom (1973) introduced, based on his work on Lycopersicon esculentum (SC) and L. peruvianum(SI) ,th eUnilatera lIncongruit yhypothesi s(als oabbreviate dUI) .H emad ea distinctio n between SI andU I and argued thatincompatibilit y andincongruit y are separate phenomena. Hermsen et al (1974)detected , in the same material thatplaye d such an important role in the research on SIi n potato, clonestha twer e "acceptor" for verpollen . Absence ofth eU I response in species crossing combinations where UI is the rule, is also called "acceptance", and "non- acceptance"stand stherefor e for UI. Geneticmodel swer e tested and similarities were found with amode lpropose db yGru nan dAuberti n (1966).Acceptanc e segregated independently from both the S-alleles and the pollen-borne SC-factor tSl, and appeared therefore to support the UI hypothesis of Hogenboom (1973). Chetelat and De Vema (1991) mapped pollen-mediated UI factors on the chromosomes 1,6 and 10.Th e factor on chromosome 1 mapped on ornea r theS- locus,thu s supportingbot hth epossibl einvolvemen t ofth eS-locu san dS-locu sindependen t acting genes onUI . Fora lon gtime ,now ,a debat ei songoin g aboutth epossibl e involvement ofth eS-locu si nUI , and, directly related to this, whether in this connection the term incongruity or incompatibility should be used. The molecular cloning of S-alleles and the construction of sense and anti-sense constructs opened thepossibilit y totes twhethe r theS-locu sha s adua l function (causing both SI and UI,o rmor e indirectly, causing SI and contributing toUI ) ornot .

General aimso fth eThesi s The diploid potato material of Hermsen, based on the clones G254, G609 and B16, was maintained over a long period by means of both vegetative and generative propagation. This resulted in material that became weak and diseased. Inbreeding isknow n to affect the reliability ofth eS Irespons e andshoul db erestricte d asmuc ha spossible.Th e S-homozygous clonestha twer e present,wer eal lbase do nth e SCfacto r tSl and of little value for SIresearch . The use ofpseudo - compatibility (brought about by inbreeding) to create ^-homozygoteswa s also not an option for the creation of fertile, well-flowering and Si-reliable clones. Some incompatibility related genes or their products had previously been isolated from this type of material. Molecular constructs, basedo nthes eclone dgenes ,ha dt ob eteste di npotat o onthei r effect. Welltransformabl e diploid potato material with the proper S-allele composition was not available yet, so this was another problem that hadt ob etackled . 20 Chapter 1

Themai n aims ofth eresearc h ofwhic hthi s thesis isa reflection , were: 1 Thecreatio n andselectio no fS-homozygou san dS-heterozygou smateria l andteste r cloneswit h areliabl e SIreaction , lackingnegativ efactor s like inbreeding depression, poor flowering, poor fertility and pseudo-compatibility. 2 Creationan dselectio no fwel ltransformabl e cloneswit ha prope rfunctionin g ofth eS Iresponse . 3 Elucidation ofmor ebiologica l aspects ofgametophyti c self-incompatibility. 4 Testingwhethe ro rno tther ei sa direc trelatio nbetwee nself-incompatibilit y andth e interspecific crossing barrier called "Unilateral incompatibility".

In Chapter 2, the development of well performing diploid potato clones with a reliable SI reactioni sdescribed . Some ofthi smateria l was used to createwell-transformabl e clones (briefly mentioned inth echapter s3 an d 6).I nChapte r2 ,th eprocedur e is alsodescribe d how the creation ofself-incompatibl e S-homozygoteswa sachieved ,withou t accompanying effects as break-down ofth e SI reaction. Thismateria l wasuse d asteste r clones,a sdescribe d in someo fth e following chapters. Inchapte r 3,th e effect ofsens e and anti-sense constructs based onth e codingregio n ofth eSI andth eS2 alleles ,o nth eincompatibilit yreactio n isdescribed . An attempt wasmad e toprov e the essential role of^-glycoprotein s(5-RNases ) inth e SIreaction . In chapter 4, the creation of male and female fertile S.tuberosum (tbr) x S.verrucosum(ver) hybridsi sdescribed .Potat oclone s(SI )tha tar eaccepto rfo r ver(SC )polle nwer eselecte d from the materialmentione di nchapte r2 .Furthermore ,th eexpressio no fself-compatibilit y ofve ri nhybrid s and in (backcross) offspring thereof isinvestigated . The contribution of thepolle n part of theS- locuso fve ro n self-(in)compatibility and unilateral incompatibility is analysed. In chapter 5, the materials and results described in the preceding chapters are used in an integrating analysis ofth erelatio n between UIan d SI.Th e loss-of function approach, asuse d in chapter 3, isapplie d inbot h tbran dtbr x verhybrids . Therol eo fth e stylarpar t of thetbr S-locu s inU I is examined and variousU I and SIphenomen a are integrated in adescriptiv e model. In chapter 6th e relevance ofth e developed material is addressed. Some of theresult s already mentioned in the experimental chapters are discussed in abroade r framework. Some significant questions,no t addressed inth echapter s2-5 ,ar epose dan dsom especulativ e approaches and ideas are highlighted. Chapter 2

Selection ofvigorou s and fertile 5-homo-an d heterozygous tester clones from

self-incompatible diploid potato,Solanum tuberosum L.

Abstract

Forth e selection ofdiploi d (2n=2x=24)potat o (Solanum tuberosum) genotypes that are useful for themolecula r andgeneti c analysiso fth ephenomeno n of gametophytic self-incompatibility, three different typeso fbasi cpopulation s were investigated. Thesepopulation s werederive d from three primarydihaploi d clones,G609 ,G25 4 andB16 ,whic hpossesse d theS-allel ecombination s S1S2, SIS3 and S3S4 respectively. In order to select highly vigorous, profusely flowering, fertile and tuberising progenies, three types of populations, derived from the above mentioned diploid genotypes,wer escreene d forperformanc e andclassifie d for theexpressio n of self-incompatibility. Although the selection for well defined S-genotypes was sometimes complicated due to the occurrence ofpseudo-compatibilit y and ofa self-compatibilisin g factor, theus eo fa combination ofcriteria ,viz. ,Is oElectri cFocusin g(IEF) ,polle ntub egrowt hi nth estyle s andth e extent ofberr y and seed set made the selection of sufficient representatives of all six types of 5-heterozygotes (S1S2, S1S3, S1S4, S2S3, S2S4 and S3S4) possible. After evaluating the strength of the self- incompatibility reaction in these heterozygotes, those with high expression were selfed, and intercrossed within their ^-allele incompatibility group through the method of counterfeit pollination. In theseprogenies ,well-performin g ^-homozygotes (S1S1;S2S2; S3S3;S4S4) for all four ^-alleleswit hhig h expression of self-incompatibility were selected. As aresult , all possible S-homo-an dheterozygou s genotypes with apredictabl e typeo f self-incompatibility are available andmaintaine dbot hvegetativel yan da sbotanica lseed .Th edevelopmen to fthi smateria lha spave d thewa y for morecritica l analysis ofmolecula r factors involved in self-incompatibility in diploid potato.

This chapter is published in a slightly modified version as:R.Eijlander, M.S.Ramanna and E.Jacobsen (1997) Selection of vigorous and fertile S-homo-and heterozygous tester clones from self-incompatible diploid potato, Solanum tuberosum L. Euphytica 97:97-111 22 Chapter2

Introduction

Thecultivate dpotato ,Solanum tuberosum (2n=4x=48),i sa self-compatibl e (SC)crop . Dihaploids (2n=2x=24)from tetraploi dcultivar sar e usuallyhighl ysterile ,les svigorou san d self-incompatible (SI). Self-incompatibility in dihaploids is expected to be similar to the one-locus, multi-allelic, gametophytic system that is found in almost all other tuberous, diploid Solanum species. This expectation wasprove d tob etru e from thegeneti c analysis of self-incompatibility inmal e fertile genotypes that occurrs rarely among dihaploids (Olsder and Hermsen, 1976; Hermsen, 1978a; Hermsene t al., 1978).Thes eauthor sanalyse d three fertile dihaploids, and thepresume d tetraploid parent oftw o of thedihaploids , the cultivar Gineke, andpostulate d thepresenc e of five S-alleles viz.,S1,S2, S3,S4 an d S5.

Through acomplet e diallel crossing scheme, the following genotypes were assigned to the three dihaploidstha twer einvestigate d(Hermsen ,1978a) :S1S2-G609; S1S3-G254 an dS3S4-B16. Thetw o latterclones ,G25 4an dB16 ,possesse d thenecessar y S-allelesfo r conferring self-incompatibility; nevertheless,the ywer e self-compatible (i.e.,se t seed after selfing). This anomalous phenomenon was explained as due to the presence of an extra SI allele (a duplication) in a presumably translocated segment ona differen t chromosome andth e authors designated this segment as "tSl" (Hermsen,1978 a; 1978b). Thishypothesi swa sinvestigate db yThompso ne ta l(1991) ,usin gRFL P analysis of plant material, coding for this tSl with the cloned stylar part of SI. The tri-partite structure of the 5-locus, as proposed by Lewis (1961) had to be rejected and the translocation hypothesis for tSl could be maintained under the assumption that the translocation would have involved only the pollen part but none of the analyzed genomic SI-fragment . In dihaploids, containing sucha tSl translocation, afractio n ofth epolle n grainscontaine d thepolle n parts oftw o different S-allelesinstea do fone ,an dhenc einhibite d self-compatibility bya competitiv e interaction (Crane and Lawrence, 1929; see alsoreview , DeNettancourt , 1977)o rmutua l weakening as, for example, in Brassica (Wallace, 1979) or in polyploids (Lewis, 1943). Besides competitive interaction, also the so-called 'pseudo-compatibility' can occasionally bring about berry development with a few seeds in a basically otherwise self-incompatible genotype (Hermsen, 1978b).

Apartfro m suchcomplication sregardin gself-incompatibility , theabov ementione d genotypeswer e valuablefo r thecharacterisatio n ofprotein stha tar eassociate d with self-incompatibility inSolanum tuberosum (Kirch et al, 1989;Peil , 1995).A sa resul t ofthi s study, itwa spossibl e to correlateSI to S4 alleles with specific bands of anumbe r ofpolypeptide s differing in their iso-electric points (Kirche t al, 1989).Thi sobservation , obviously, openedu ppossibilitie s for amor ereliabl e method Creation andselection ofbasic S-homo- and heterozygous SImaterial 23

of identification of the S-alleles through electrophoresis, and characterisation of the self- incompatibility system in diploidpotat o morecritically . Inthi s context, itwa s essential to select defined diploidpotat omaterial ,homozygou s orheterozygou s forparticula r S-alleles,tha twoul db e suitable for more critical genetic andmolecula r analysiso fth eincompatibilit y system.

Selection of diploid potato genotypes with a defined S-allele composition, e.g., S-homo- and heterozygotes,wit ha predictabl e expression ofself-incompatibilit y isdifficul t for severalreasons . Potatobein ga highl yheterozygou scrop ,inbreedin gdepressio ni sa sever eproble mbot hfo r selfing andfo rintercrossin g amongindividual swithi na smal lgrou po f(diploid )genotypes .Thi si sbecaus e theprogenie si nthes ecase sar egenerall yles svigorous ,non-flowering , highlysterile ,non-tuberisin g andfrequentl y segregatingfo r lethalan dsemi-letha lgenes .I norde rt ocircumven tthes e difficulties, a rigid selection of diploid parents based on the performance of their progenies for some of the important characteristics, includingth etypica l expression ofself-incompatibility , is essential.

The aims of the present investigation were: 1) to select diploid potato genotypes with highly vigorous, fertile, earlyan dprofusel y flowering habit, showing good tuberisation characteristics; 2) toisolat ehomo -an dheterozygou steste r stocks for different S-alleleswit hpredictabl e and reliable expression and 3) to produce sufficient plant material of each allelic class (seeds and tubers) for generative and vegetative maintenance.

Materials and methods

Selectionof the basic genotypes Two different types of populations were screened for desirable genotypes. The first of these consisted of 'basic' populations derived from crossesbetwee n three dihaploid Solanum tuberosum (2n=2x=24) clones, G609 (S1S2), G254 (S1S3) and B16 (S3S4). The origin and the indicated genotypes ofthes ebasi cclone s havebee n described earlier (Olsder andHeimsen ,1978 ; Heimsen, 1978a). The progeny used for the selection of desirable genotypes had originated (see Table 1, column 2)no t only from direct crosses (five original Fl's ) bet-ween the dihaploid clonesbu t also from intercrossing and selfing of progeny plants (14 populations from selfings and inbreds). Because the 5-genotypes of each ofth eparent s werehomozygou s for the marker 'embryo-spot' (Hermsen andVerdenius , 1967),th eseed sresultin g from counterfeit pollination couldb e separated from those resulting from the first pollination. In diallel crossing, at least 10 pollinations per combinationwer emade ,usin gflower s from twoo rmor e inflorescences in thecas eo fbot h normal and counterfeit pollination methods. 24 Chapter2

Besides normal and counterfeit pollinations, "prickle pollinations" (pollination with only IvP- pollen)wer emad ei norde rt o usethe m ascontrol s for determining theproductio n ofspontaneou s spotlessseed si nsuc hfemal eplants . Theseparticula rspotles sseed sar ebelieve dt ob enormall yth e result of diploid orhaploi d parthenogenesis, thelatte r leading to monohaploid offspring (Van Breukelen et al, 1977;Uijtewaa l et al, 1987).

Statisticalanalysis Data onth e number ofpollinations , obtained berries and number ofseed s were analysed with the computerprogramm e StatgraphicsPlu sv 7.1 . The number ofspotles s seeds after anincompatibl e cross plus counterfeit pollination with IvP pollenwa scorrecte dwithi n eachSi-grou pb y subtraction ofth enumbe r ofspotles s seedsproduce d after pollination with only IvP-pollen.

Table1 .Basi cpopulation s(colum n 1)obtaine d from SIan d SCparent s(colum n 2)wit hknow n5-genotype s(colum n 3) used for the selection of well performing SI heterozygotes and SChomozygote s (column4) . GB = G254 x B16; BG= B16 x G254;S1S3, S2S3 an d S3S4= self-incompatible tester clones selected from GB,Gx( Gx G609) and GBrespectively . Numbersbehin d brackets indicateclon enumber . 8= selfing, SI= self-incompatible, SC=Self - compatible duet o tSl. Italicsbold: last SC-clone used in across . Population Parents Description Obtained genotypes Selfings 6107 {(G254 x S1S3)20 x S1S3)8 S 5753» S1S3 + S3S3,S C 6108 {(G254xSlS3)20xSlS3)l0E S1S3B S1S1 +S1S3 + S3S3 6233 6105-6 S S2S4H S2S2 + S2S4 + S4S4* 6234 6105-8 S S2S3H S2S2 + S2S3 + S3S3 Inbreds 6101 (G254xSlS3)20xS\Sl 575/ xSI S3 S1S3,SI/SC 6102 S2S3 x {(G2S4xSlS3)20x S1S3J1 S2S3xSlS3 S1S2 + S1S3 6103 {(G254 x S1S3)20 x S1S3}1 x S2S3 SlS3xS2S3 S1S2 + S2S3 6104 S2S3 x {(G254 x S1S3)20 S }4 S2S3 x 5757 S1S2 + S1S3.SI/SC 6105 {(BG112x GB61)21 x S3S4J2 x S2S3 S3S4xS2S3 S2S3 + S2S4 6106 {(BG112xGB61)23 »/J x S2S3 S4S4x S2S3 S2S4 + S3S4, SI/SC 6206 {(G2S4 x S1S3)20 x S1S3J9 x S2S3 5753 x S2S3 S1S2 + S2S3, SI/SC 6208 S2S3 x {(G254 x S1S3)20 x S1S3}2 S2S3 x 5753 S1S2 + S1S3, SI/SC, S2S3 + S3S3.SC 6536 (GB49 xB16)\l x (GB53 x G2S4)4\ 5/53xSlS4 S1S4 + S3S4, SI/SC 6539 (GB66xGB65)ll XS1S4 5754xSlS4 S1S1 + S1S4 + S4S4, SI/SC Original F1' 5 6221 G254 xB16 5753 x 5354 S1S4 + S3S4, SI/SC, S1S3 + S3S3,S C 6222 G254 x G609 5/53xSlS2 S1S2 + S2S3, SI/SC 6223 G609xB/(J SlS2x5354 S1S3 + S1S4 + S2S3 + S2S4 SI/SC 6224 G609 x G254 SlS2x5753 S1S3 + S2S3.S1/SC 6225 B16 x G254 5354 x 5753 S1S3 + S1S4, SI/SC, S3S3 + S3S4,S C

! S4S4 not detected. Creation andselection ofbasic S-homo- and heterozygous SImaterial 25

Selections of spotless seed sampleswer e sown and the seedlings were tested for the accuracy of embryo spot detection, for plant performance and the expression of SI/PC/SC. The statistical analysis ofth eproductio n of spotless seedswas ,however ,base d onth e determination of spotless seeds and not on seedlings without nodalband . Simultaneous analysiso fpolle n andstyl e effect wasperforme d on"withi n incompatibility group" level. On total population level, the analysis of the main effects of pollinator (pollen parent) or recipient (seedparent )wa sperforme d separately.Here ,pollinato r or stylar effect meanspe r clone out of the over-all analysis were consecutively added four times as covariates in an iterative approachafte r aninitia l separateanalysis .Th eanalysi so nth eover-al l levelwa s alsoperforme d by adding within-group means ascovariate .

Classification ofgenotypes for S-alleles Four criteriawer e used for the classification ofS-allel egenotype s andthei r SIreaction : 1)th eS- alleles were identified by iso-electric focusing of stylar extracts with Polyacrylamide gel electrophoresis (PAGE)o rprecas tagaros egels ;2 )PAG Eresult swer everifie d throughtes tcrossin g and vice versa; 3)th eexten t ofpolle ntub egrowt h inth epollinate d styles wasmonitore d undera fluorescent microscope;4) berr yan dsee d setwer eevaluate d after selfing aswel l asafte r crossing withteste r genotypes.

Biochemical identification-PAGE Iso-electricfocusin g ofstyla rextract swit hPAG Ewa sperforme d asdescribe db yKirc he ta l(1989 )o r by means of pre-cast agarose gels (Hypure gel VG 1020, Isolab inc.) following the silver staining procedure based on Tungstosilicilic acid in stead of sodium permanganate, according to company specifications.

Pollentube growth instyles Pollen tube growth in styles was studied according to the modified technique of Martin (1959). Briefly, thetechniqu e was as follows: receptive styleswer epollinated ; 48h later, theywer e fixed in freshly prepared 3:1 solution of ethanol acetic acid for a day or longer; macerated with 8N sodiumhydroxid e solution at65° Cfo r atleas t 8 min.;rinsed wit hwater ; stained with 0.1% aniline blue dissolved in 0.1M potassium pyrophosphate; softened styles were mounted in glycerol and observed under a fluorescent microscope (BG12/4filte r combination inZeis smicroscope) .

Estimationof berry and seed set At least 10pollination s were made in order to determine whether a genotype was SC or SI.Th e genotypestha tha da hig hpercentag e ofpolle n stainability but failed to setberrie s on selfing were classified as SIan dthos etha tproduce dberrie s and seeds inhig hnumber swer econsidere d as SC. Becausea S Creactio ncoul dresul teithe r from thepresenc e oftSl orb edu et o pseudo-compatibi­ lity, in ambiguous cases progenies of PC/SC plants wereteste d in order toverif y whether those 26 Chapter2

parentswer e SCo rPC .Berr yan dsee d setwer eestimate d onth ebasi so fseed spe rberry ,seed spe r pollination and berries per pollination. In the case of counterfeit pollinations and the control pollinationswit h onlyIvP-pollen ,th e spotless seedswer e separated from thosewit h spots under a binocular microscope and counted. Pollen stainabilitywa s estimated bymountin g fresh pollen grainsfrom thre e flowers, on different dates in each case,i n adro p of 2% acetocarminesolution . On anaverage ,20 0polle n grains were counted per assessment.

Results

Performance ofth ebasi c populations

Three types of basic populations consisting of selfs, inbreds and the original Fis, that were investigated in the greenhouse for performance are described in Table 1 with indication of their parentsan do fth egenotype s obtained. In all cases,wit h the exception ofth epopulatio n 6233, the obtained5-allel egenotype s ofth eprogenie s were fully concurrent with the established genotypes ofth eparent s(Tabl e 1).I nth e exceptional population 6233, only S2S2an dS2S4 genotype s could bedetected , whereas the also expected S4S4genotyp e was absent. Therewer eclea rdifference s inperformanc e (Table2 )amon gth eprogen ypopulation s derivedfrom selfs, inbreds and the basic Fis with regard to the average scores for vigour, flowering, fertility, tuberisation and the number of cripples (plants that were tiny, weak and brittle). In general, the progenies of the basic Fis were superior to the other two categories for all the four parameters estimated. For example, the average scores inth efive basi c Fl populations (6221t o 6225,Tabl e 2) were consistently higher ( with 77 useful plants) than in the 14 populations (with 74 useful plants)derive dfrom th eself s andinbreds .Especiall yth efrequency o fusefu l plantsafte r selfing was low.Becaus eo fthes edifference s betweenth ethre epopulatio ntypes ,th eprogenie so fth e basic Fis wereno tonl ymor eusefu l for^-heterozygote sbu t alsofo rth eselectio no fth eS-homozygote susin g counterfeit pollination (seelater) .

Selection of SI and well performing ^-heterozygousplant s out ofth e basic populations

Theevaluatio n ofth ebasi cpopulation sprove dtha t amajorit y (407/548)o fth eplant s among them were unfit for selection of S-heterozygotes since they did not meet the four criteria used for selection(Tabl e2) .I na furthe r roundo fselectio n amongwel lperformin g plantso fth epopulations , a total of 31 useful individuals from different populations were evaluated for Si-expression. All these genotypes were classified on the basis of •S-allele composition through both IEF and test crossing. Creation and selection of basic S-homo- and heterozygous SI material 27

Table 2.Performanc e ofbasi cpopulation s forth eselectio no fusefu l heterozygous SI genotypes. All characters were scored on anordina l scale for eachplan t andth e averagevalue s arepresente d inth ecolumns . Ranking: 1 =lacking ,2 = presen tbu tbad ,3 =poor ,4 = insufficient , 5 =jus tsufficient , 6= sufficient , 7= satisfactory, 8= good ,9 = ver ygood ;betwee nbracket s() :# wel lperformin g self-compatible clones,- = no tsegregatin gSC-plant s

Population # plants Vigour Flowering Fertility Tuberization # useful plants Selfings SI SC

6107 10 2 3 2 1 0 (0) 6108 30 3 3 4 2 0 (-) 6233 11 4 4 4 3 1 (-) 6234 14 5 4 5 4 5 (-) Inbreds 6101 30 6 6 7 4 6 (6) 6102 30 5 5 5 5 7 (-) 6103 40 5 5 5 4 9 (-) 6104 30 6 6 8 7 11 (H) 6105 20 6 6 7 6 7 (-) 6106 40 6 6 6 6 7 (10) 6206 30 6 8 7 4 5 (4) 6208 40 7 8 8 3 2 (6) 6536 20 6 6 7 7 6 (6) 6539 20 5 5 7 6 8 (6) Basic Fl's 6221 40 8 6 8 7 13 (22) 6222 40 8 7 8 7 18 (19) 6223 40 8 7 8 7 18 (15) 6224 23 8 7 8 7 9 (11) 6225 40 8 6 8 7 19 (16)

Theexpecte d sixclasse so ffou r different alleles,i.e. ,S1S2, SIS3, S1S4, S2S3,S2S4 an d S3S4wer e found. Table 3present s for all sixexpecte d SIclasse sth e average scores ofth e selected plants for each ofth e four characters vigour, flowering, pollen shedding (scale 1-10) andpolle n fertility (% stainable), together with their Si-expression. A notable feature was that 19 out of 31 of the genotypes that showed atypica l Si-reaction (Table 3)wer e derived from the progeny of the five basicFi s (cfTabl e 1),wherea sonl y 12o fth e genotypes originatedfrom th e 10 inbredpopulation s andnon efrom th e selfings. In order to evaluate the strength of Si-expression, all six5-genotypes ,consistin g of 30plant s in total,wer eteste d forberr yan dsee dse tafte r selfing (Table3) .A stric tself-incompatibilit y reaction (no berry and seed set)wa s expected in allplants .However , testing the 30plant s sever-allyears , revealed that stillnin eo fthe m occasionally did set (self)seed, ranging from 10- 80seed spe r berry (compatible crosses givegoo dberr y formation and 150- 25 0 seedspe rberry) .Thi s could often be 28 Chapter2

attributed to pollination of young flowers in which there might still have been an incompletely developed SI barrier (cf. bud-pollination). Thisphenomeno n was considered as a less reliable SI reaction. Excludingthes egenotypes ,th estrengt ho f theSi-expressio n inal lother swa s satisfactory aswa seviden tfrom th eabsenc e ofberr y and seed set after 30t omor etha n 100selling s that were madei ndifferen t genotypes(Tabl e3) .Berr yan dsee dse ti nsom eo fth eS-heterozygoti cplant swa s an indication for the persistence (genetic transmission) of pseudocompatibility. As pointed out already inth epopulatio n section, amajorit y ofth eusefu l S-heterozygotes (viz., serial numbers: 6221, 6222, 6223, 6225) in Table 3wa s derived from three of thefive population s of basic Fis mentioned inTabl e 1.Thi swa sreflecte d inth efinal selectio n ofth eclone s tob emaintained . The populations 6221an d622 5 aswel l asth epopulation s 6222 and 6224 are theresul t of reciprocal crosses.Th eunderrepresentatio n ofth epopulation s 6224 and622 5 doesno t reflect inferiority but wasjus t amatte r ofrando m choice.

Selection of5-homozygou s SIplant s from thebasi c populations

All well performing plants in the self and inbred populations, that could have contained S- homozygous genotypes,wer einvestigated . Among thoseplants , self-incompatible aswel l as tSl- basedself-compatibl e S-homozygousgenotype scoul db epresent .Becaus eonl yth e5-homozygote s with self-incompatibility were essential, thebasi cpopulation swer e screened for such genotypes, andthos ewit h self-compatibility werediscarded . Inpopulation s 6108,6233,6234 and653 9(Tabl e 1),S-homozygote swit h a typical SI reaction were found. On the other hand, theS-homozygote s from populations 6107,6208,6221,622 5 and653 9 wereal lfoun d tob eself-compatibl e andwere , therefore, discarded. Among the self-incompatible 5-homozygotes,th epopulatio n 6108 consisted of less vigorous individuals, and 6539ha d high levels of sterility besides poor tuberization; and these wereno t suitable for thefinal selection . Only twopopulations , 6233 and 6234, gaverise t o some desirable genotypes with valuable features (Hermsen, 1978c; Hermsen and Olsder, 1974) despitehavin g arelativel ypoo r generalperformance . Botho fthes epopulation s were derived from 6105-06 and 6105-08 which wererar e cases of seed set upon selfing (Table 1).Thi s seed setwa s mostprobabl y theresul t of environment-induced PC.Th e successrat eo fthi s type of selfing was notpredictable . Theperformanc e and ^-genotypeso f six ofth eplant s selectedfrom th e 6233 and 6234population s arepresente d in Table 4.Althoug h the selected genotypes wereno t completely satisfactory inperformance , theywer etypicall y self-incompatible, andinitiall y useful astesters .A greater disadvantage ofthes e successful populations wastha t only S2S2an d S3S3 S-homozygote s wereobtained . Forth eselectio n ofmor evigorou sS2S2 andS3S3 genotype s and of SISI andS4S4 homozygotes aswell , amor e effective method ofbypassin g SI, using many genotypes within an incompatibility group,wa srequired . Thiswa sdon ethroug h counterfeit pollinations. Creation andselection ofbasic S-homo- and heterozygous SI material 29

Table 3 Performance of selected Si-expressing S-heterozygousgenotype s +ke y identifiers totabl e 8. Vigouran dflowerin g impression: 1 =extremel ybad ,2 = bad ,3 = poor ,4 =insufficient , 5 =nearl y sufficient, 6 = sufficient, 7 = satisfactory, 8= good, 9 = very good, n.d. = not determined; pollen fertility expressed as % acetocarmine stainable pollen; no additional mark = standard deviation between periods: 0 -5, * = 5 - 10, ** > 10. Si-expression: totals of seeds/berry/selfing in thefirs t two years;* = caused by pollination of very young flowers; SC= self-compatible Plant nr. Key Genotype Vigour Flowering Pollen fert. Pollen shed SI expression Seeds / berries / selfs

6102-16 I S1S2 7 7 84 6 0/ 0/ 32 6104-09 2 S1S2 8 9 93 8 0/ 0/ 71 6104-21 4 S1S2 7 8 92 9 0/ 0/ 43 6222-05 15 S1S2 8 9 96 9 0/ 0/ 39 6222-39 18 S1S2 9 9 87 7 0/ 0/ 49

6101-11 1 S1S3 7 7 90 8 0/ 0/ 73 6104-19 3 S1S3 7 8 80 7 0/ 0/ 78 6104-23 5 S1S3 8 8 97 9 80/ IV 148 6223-15 III S1S3 7 8 78 7 0/ 0/ 30 6225-05 IV S1S3 8 9 92 8 n.d (SC) 6225-15 V S1S3 8 8 80 8 0/ 0/ 22

6223-40 23 S1S4 9 9 60* 9 0/ 0/ 69 6221-01 8 S1S4 8 8 78 7 0/ 0/ 92 6221-05 9 S1S4 7 9 61* 8 4/ 27 64 6221-17 10 S1S4 8 8 64 7 50/ 2/ 50 6221-19 11 S1S4 8 8 64 8 42/ 4/ 62 6221-20 12 S1S4 7 9 75 9 0/ 0/ 107

6105-08 II S2S3 7 8 50** 9 30/ 17 32 6222-06 16 S2S3 7 9 80 7 150/ 37 51 6222-24 17 S2S3 8 9 91 8 0/ 0/ 48 6222-40 19 S2S3 9 9 74** 9 0/ 0/ 70

6105-06 6 S2S4 7 7 72 8 61 / 57 112 6105-15 7 S2S4 7 6 70 8 19/ 27 52 6223-01 20 S2S4 8 7 82 6 0/ 0/ 41 6223-29 21 S2S4 9 9 79* 8 0/ 0/ 50 6223-39 22 S2S4 9 9 67* 9 0/ 0/ 51

6536-01 24 S3S4 7 9 71* 9 32/ 17 55 6536-02 25 S3S4 8 9 42** 8 0/ 0/ 55 6536-09 26 S3S4 8 8 64 9 0/ 0/ 52 6221-32 13 S3S4 7 9 82 9 0/ 0/ 71 6221-37 14 S3S4 8 9 46** 8 0/ 0/ 54 30 Chapter2

Production of superior homozygotes from selected S-heterozygotes

All genotypes mentioned in Table 3 have been used in some way, trying to obtain more S- homozygotes. Not allth e genotype combinations thatwer emad e for thispurpos e havebee n used forth e screening and selection ofS-homozygotes ,particularl y because of suspected expression of PCi nth eoffsprin g or(expected ) inbreeding depression. Someparenta l clones that havebee n used were excluded from thestatistica l analysisbecaus eo f (re)appearance of SCi n the selected clones or temporary regrowth problems, leading to bad synchronisation of flowering. Five clones were excluded because of virus infection, resulting in too few observations to be of use for statistical analysis. Those genotypes were indicated by 'Roman numbers" in Table 3. Four of them gave useful 5-homozygotes though (Tables 4 & 8). The PC genotypes with occasional seed set after young flower pollinationwer einclude d inthi s experiment.Unlik eth e selectiono f S-heterozygotes, itwa s generally muchmor edifficul t to createS-allel ehomozygote s from genotypes that showed thetypica l SIreaction . Theproble m in thesegenotype s wast o obtain seeds from selfing, or from intercrossing withina nincompatibilit y group.Thes edifficultie s were largely overcome by making alarg enumbe r ofpollinations , followed by counterfeit pollinations. For making the counterfeit pollinations, different genotypes within each ^-incompatibility group wereselfe d andintercrossed ,followin g thegenera lcrossin gschem eexemplifie d forS2S3 genotype s (A, B and C) intabl e 5.Thi s schemewa s applied for allpossibl e six S-heterozygousgroups . One set of these crosses was carried out without counterfeit pollination (normal incompatible pollination) and the other identical set with counterfeit pollination (normal + counterfeit pollination);thi smean s that 48hr s after the incompatible cross,a secon dpollinatio n was carried outbu tno w withpolle n from theIv Pmarke rgenotype so fS.phureja. Inthi scontext , approximately 4300pollination swer emade .Thes epollination sinclude dthre egroup so fwithi ngrou pincompatibl e crossing: 1)selfin g aS Igenotyp e(Self , inTabl e 5: bold),2 )th e samegenotyp euse d assee d parent (SP, Table 5: row) with various non-self pollinators (non-bold) and 3) the same genotype used aspollen parent (PP) (Table 5: column) in non-self crosses. These pollinations yielded, approximately, atota lo f 1100berrie swit h 125000seed so fwhic h950 0wer ewithou t embryo-spot. In order to assess of the strength of the incompatibility reaction and to determine whether the number of spotlessseed sresultin g from thecounterfei t pollinations was different from the number of spotless seedsobtaine d from thecontro lpollination s (stylespollinate d onlywit hpolle n from the marked IvP clones ), a statistical analysis was performed (Table 6). The occurrence of spotless seeds from prickle pollination alone (based on LSD-values) was statistically not significantly different from zero,bu t significantly different from thecounterfei t pollination effect (Table 6).Thi s was evident regardless of the consideration ofparenta l effects. The occurrence of the number of spotless seeds after counterfeit pollination ornorma l single crosswa s calculated and analysed in Creation andselection ofbasic S-homo- and heterozygous SImaterial 31

Table 4 Performance of selected SI S-homozygous genotypes derived from selfed SI clones.

Plant Genotype Vigour Flowering Pollen fert. Pollen shed SI Seeds/ berries/ selfs

6233-11 S2S2 5 5 55** 5 0/ 0/ 70 6234-05 S2S2 5 7 75** 5 0/ 0/ 50 6234-12 S2S2 5 5 50** 5 0/ 0/ 75 6234-01 S3S3 5 5 60** 4 0/ 0/ 50 6234-08 S3S3 6 6 74** 6 0/ 0/ 90 6234-10 S3S3 6 6 41** 5 0/ 0/ 50 The scale of values is ranging from 1 (extremely bad) to 9 (very good). ** = Standard deviation > 10%. SI reaction as seeds/berries/selfed flowers (totals).

Table 5. Crossing scheme of the counterfeit pollination experiment for the production of S-homozygotes

'Seed parent • Pollen parent Normal incompatible crossing Normal +( Control ABC A IvP

A: S2S3 X X X X X X X B: S2S3 X X X X X X X C: S2S3 X X X X X X X A, B and C belong to the same S-allele combination group, e.g. S2S3. Bold X = selfing. Control with IvP = pollination with IvP-pollen as in counterfeit pollination but without an incompatible first pollination.

Table 6 A -îalysis of variance (ANOVA) in counterfeit pollination experiment using spotless seed proc uction per pollination orpe r berry. Main Eff.= Main effect; Seedp. = = seed parent; Phu =counterfeit/prickl e pollination with IvP's; d.f = degrees of freedom; S.S= Sumo f squares;F =(S.S . maineffect/df)/(S.S . residual/d.f); Y= significance evel.

Spotless seeds/pollination Spotless seeds/berry

Main Eff d.f S.S. F Y Main Eff d.f S.S. F Y Seedp. 24 4293.17 2.82 0.0000 Seedp. 24 13530.17 4.02 0.0000

Phu 1 558.09 7.68 0.0061 Phu 1 2149.44 15.32 0.0001

Residual 207 15039.48 Residual 207 29037.51 twoways ,viz. ,pe rpollinatio n andpe rberr y(Tabl e6) .Analyse swithi n5-genotyp egroup s andth e iterative approaches of the determination of the main effects gave basically the same (nearly identical) results. One conclusion was that thenumbe r of spotless seeds/berry was twice ashig h asth enumbe r of spotlessseed s/ pollination (Fig. 1). From these calculations, itwa s evident that 32 Chapter2

Figure 1.LS D analysis for set of spotless seeds after single normal incompatible pollina­ tion (S.P.) or counterfeit pollination assisted incompati­ ble crossing (C.P.). Y-axes: Left (+): spotless seeds/ pollination; Right (O) spotless seeds/berry. + and O; means. Capped error bars: 0.95% confidence intervals, pp op p p Based on 473 corrected POLLINATION TYPE means- counterfeit pollinations gaverise t omor eseed s(spotless )tha twer epotentiall y5-homozygote stha n single pollinations in incompatible genotypes. The other difference was that in the case of counterfeit pollinations moreberrie swer eobtaine d andha d tob e extracted. Despite a strong selection for typical self-incompatible genotypes, it was evident that certain individuals occasionally setsee d after selfing. Thiswa sa nindicatio n for theoccurrenc e ofpseudo - compatibility duet oth einfluenc e of either themal e orth e female parent. The genotype 6222-40, for example, was an instance of showing PC effect from the seed parent and 6222-06 from the pollenparen t (Table 7).Th eoccurrenc e ofpseudo-compatibl e genotypes occurred inth e progeny in anumbe r of cases, especially when aparenta l clone (e.g. 6502-38,paren t of 1127-14,tabl e 8) showed strong PC (data not shown), clearly indicates the genetic basis of this character. After excluding suchP Cgenotypes ,i twa sstil lpossibl e toretai n aconsiderabl e number ofgenotype s of all^-allel ehomozygote swit hhig hlevel so fvigour , flowering, pollen fertility, pollen shedding and Si-expression(Tabl e 8). Thosehomozygote swer eeithe rhybrid s(intercrosse s within incompatibil­ itygroup )o rselfings .Bot hgroup swer eobtaine dwit han dwithou tth eai do fcounterfei t pollination. Asa resul to f selection ofwel lperformin g genotypes,mos t ofth e selfing-based genotypes (many of them showing inbreeding depression),hav ebee n excluded in favour of hybrid types.Becaus e this population was still excessively large, it was narrowed down. The plants indicated by an asterisk (*)hav ebee n selected eithero nth ebasi so fperformance , scaleo ftestin g for SI expression and their value for other research topics, or have been selected at random. The SJS1 and S2S2 genotypes and to a lesser extent the S3S3 genotypes were the most important for the molecular unravelling ofth e Si-system (Kirche tal , 1989, Eijlander andFicker , inprep) .Th eS4S4 genotypes areslightl yunderrepresented ; thisma ypartl yb edu et oth elo wpriorit y of obtaining this genotype, but the number was still lower than expected. Numbers are too small, however, to draw final conclusions on5-genotype-relate d fitness or certation. Creation andselection ofbasic S-homo- and heterozygous SImaterial 33

Table 7. Strength of the SI reaction in selfings and intercrossings within incompatibility groups of selected SI- expressing S-heterozygous genotypes with or without using counterfeit pollination. SP = tested as seed parent; PP = tested as pollen parent. Ranking is from "1 = most PC genotype" to "9= most SI genotype", sc = self-compatible.

Normal ino Normal + Counterfeit pollination Plant Key Self SP PP Self SP PP 6104-09 2 9 6104-21 4 9 6222-05 15 8 6222-39 18 9

6101-11 1 9 6104-19 3 9 6104-23 5 8

6221-01 8 8 6221-05 9 7 7 6221-17 10 sc sc 6221-19 11 7 6221-20 12 9 6223-40 23 9

6222-06 16 4 6222-24 17 7 6222-40 19 9

6105-06 6 8 6105-15 7 8 6223-01 20 8 6223-29 21 7 6223-39 22 8

6536-01 24 8 6536-02 25 9 6536-09 26 9 6221-32 13 9 6221-37 14 9 34 Chapter 2

Table 8. Selected SI and SC homozygous clones derived from SI and SC clones respectively after (in)compatible pollinations using within incompatibility group pollinations or selfings in combination with or without counterfeit pollination Italic-bold :SC-genotypes . Method= obtained by counterfeit pollination (c f ), selfing (se), non counterfeit pollination aided intra incompatibility classpollinatio n (ii) orb y aSC-base d compatible cross(co) .Fl . =flowering , P.F. =polle n fertility expressed as %aceti c carmine stainablepollen , P.Sh. =polle n shed. Scalesrangin g from 1=extremel y bad to 9= very good. SI-expr. = self-incompatibility expression expressed as seeds/berries/selfed flowers, n.d = not determined. A = vegetatively maintained

Plant Parents Method Genotype Vigour Fl. P.F. P.Sh. StexpressionjsJQtals...of Seeds/ Berries/ Flowers

6496-01A IVx3 cf S1S1 8 9 57** 9 0/ 0/ 60 6496-04A IVx3 cf S1S1 7 8 71** 8 0/ 0/ 64 6499-04A 111x3 cf S1S1 7 8 96 7 0/ 0/ 73 1127-14A 6502-38 x 6496-01 ii S1S1 8 8 88 7 0/ 0/ 58 1130-03 5x1 cf S1S1 7 8 92* 8 0/ 0/ 22 1136-01 12x23 cf S1S1 8 7 65** 7 n.d. 1136-02 12x23 cf S1S1 7 8 74* 6 n.d. 1136-05 12x23 cf S1S1 9 8 77* 7 0/ 0/ 33 1181-02 (7254x6496-1 cf S1S1 8 8 55** 8 0/ 0/ 25 1138-07 16x 16 cf, se S2S2 7 7 62** 6 n.d./ 1/ 12 1138-08 16x16 cf, se S2S2 7 7 69++ 7 0/ 0/ 20 1139-03 17x19 cf S2S2 9 7 63** 7 0/ 0/ 30 1139-05 17x19 cf S2S2 8 8 70* 7 0/ 0/ 30 1140-01 6x22 cf S2S2 8 8 55** 4 0/ 0/ 20 1140-02A 6x22 cf S2S2 7 7 60** 7 0/ 0/ 54 1140-05A 6x22 cf S2S2 7 7 70** 6 0/ 0/ 48 1146-02 16x 19 cf S2S2 8 9 90 5 0/ 0/ 22 6499-01A 111x3 cf S3S3 8 9 88 8 0/ 0/ 55 1130-01 5x1 cf S3S 3 7 3 50** 3 0/ 0/ 4 1138-04 16x16 cf, se S3S3 7 6 64** 7 52/ 2/ 5 1138-08 16x16 cf, se S3S3 7 7 72* 7 12/ 1/ 32 1142-02A 25x26 cf S3S3 8 7 79 7 0/ 0/ 26 1171-01 24x24 cf, se S3S 3 7 7 66* 7 0/ 0/ 23 1095-04 23x11 ii S4S4 7 7 69** 6 0/ 0/ 67 1095-06 23x11 ii S4S4 7 7 75** 7 0/ 0/ 71 1134-01 11x11 se S4S4 6 7 78* 8 n.d/ 8/ 18 1147-04 23 x 11 cf S4S4 7 6 71 6 0/ 0/ 12 6539-10* see table 1 ii S1S1-SC 8 9 99 8 2204/ 10/ 12 1132-07 12x70 co S1S1-SC 5 5 51** 7 n.d 1132-20 12x70 CO S1S1-SC 7 6 75* 5 n.d Creation andselection ofbasic S-homo- and heterozygous SImaterial 35

Self-compatibility inS-homozygote s

As mentioned before (Table 1), self-compatible clones were found some of which were S- homozygous. The S2S2, S3S3 and S4S4 self-compatible clones have already been obtained on a routine basis, as is partly shown in Table 1. This was not the case with SISI. The crosses IV (=6225-05) x 3 (=6104-19) and G254 x 6496-1, (in italics,bold : SC-clones; Table 8)wer e made with the secondary aim of testing the validity of the assumption made by Olsder and Hermsen (1976)tha t tSl does not cause mutual weakening when together in a (monohaploid) pollen grain withth ecomplet eSI -allele .Offsprin g populations ofthes ecrosse s did not contain self-compatible S1S1 plants.Progen yo fthes eS1S1 plantsprove dth eabsenc eo fth e SC-factor tSl. Plants6539-10 , 1132-07an d 1132-20,however ,prove db yIE F (Fig.2 )an dtes tcrossin gt ob eSI SI homozygotes, but they were self-compatible (table 8) and capable of fertilizing other SI-expressin g plants.The presenceo fth etSl -base dSC-clone s 1132-07an d 1132-20(tw oSI SI SCclone sou to f2 0S Cplants , P(k(n=20 ,p=l/4 ) < 2)= 0.09 )prove stha t evenS 1-tSl polleni sno tcompletel ysupersede db yS4- tSl pollen, although acertativ e disadvantage isver y likely.

Figure 2 .Iso-Electric-Focusin g -^ — ^- o ov n - -a (IEF) pattern of stylar extracts Os r^l e*l E T3- <-^ CNI I ^D <0 *0 t>0 after silver staining of 3 S- homozygotes(6233-11 ,6539-19 , 6499-03), 5 5-heterozygotesan d a control sample. Approximately 15-25^g protein was added to each slot. SKI and SK2 are style specific proteins but not 5-locus related.. SI, S2, S3 and S4 are stylar expressed 5-allele specific glycoproteins. S4 gives facul­ tatively asecondar y band. 36 Chapter2

Maintenance ofth eselecte d genotypesb ymean so fi nvitr opreservation ,see dtuber s andtru e seedproduction .

Mosto fth eselecte dmateria l ismaintaine d invitro ,an dth egenotype smentione d inTable s3, 4 and 8 areals oavailabl e inth e form ofseed s andtubers . Seedswer eproduce d by crossing S-heterozy- goteswit hS-homozygote si norde rt oproduc e 'single-class'^-heterozygotes ,i.e. ,SI SI xS1S2 give s rise to asingle-clas s S1S2heterozygote . Suchgenotype sar eavailabl efo r allsi x combinations with thefou r S-alleles.Beside sthese ,a limite damoun to fsingle-clas shomozygote s (intru e seed form), derived from counterfeit pollination, is available for SISI, S2S2 and S3S3. One group of S- homozygotes,S4S4, wasrecentl ylos tbecaus eo fvira linfectio n andi sno wonl ypresen t intru esee d form. The number of S-homozygous seeds from S-homozygotes with self-incompatibility was relatively small because of the difficulty of producing seed from their well functioning SI, even when counterfeit pollination wasuse d (Table8) .

Discussion

In spite of the presence of a well defined monogenic, multiallelic, gametophytic type of self- incompatibility indiploi d potato,th e systemwa sno t amenable to acritica l genetic and molecular analysis in the past for the following main reasons: 1) criteria for the classification of S-allele genotypes were less well defined; 2) well performing genotypes, or testers, of S-homo- and/or heterozygotes were not available and 3) potato being a highly heterozygous crop, inbreeding depression and the expression of deleterious recessive genes in the progeny were serious impediments for analyses.I nth epresen t investigation, anattemp twa smad et oovercom eth eabov e mentioned drawbacks byth e selection ofwel lperformin g andwel l defined SIgenotypes .

The classification of the genotypes of S-alleles in potato is complicated by the fact that self- compatibility often occurseithe rdu et oth e so-called pseudo-compatibility or self-compatibilising factors, such astSl, in anotherwis e self-incompatible genotype. Classification of such genotypes inafor ementione d caseso nth ebasi so fberr y andsee dset ,togethe rwit hpolle ntub egrowt h studies (Hermsen, 1978 a, b), were relatively subjective in some cases. The identification of S-alleles through IEF (Kirch et al., 1989)combine d with studies onpolle n tube growth in styles aswel l as on test crossing was a step forward for amor ereliabl e classification of theS-genotypes . Using a combination of the three criteria, well performing genotypes of both homo- and heterozygous genotypes for four different S-alleles have been selected in the present investigation. Accurate determination of S-allele genotypes through IEF was especially useful for the selection of self- Creation andselection ofbasic S-homo- and heterozygous SImaterial 37

incompatiblehomozygote s aswel l asheterozygote s and for gaining more insight into the probable inheritance of pseudo-compatibility.

Olsder and Hermsen (1976) found a complete absence of self-compatible S1S1 homozygotes. Segregation ratios displayed skewness and pollen certation or absence of mutual weakening between theSI allele andth ehypothesise d tSl was oneo fth e explanations. Mutual weakening between two identical S2 alleles does not occur (crosses were made with a tetraploidised versiono fclon e 1140-2,- S2S2S2S2, not givinga compatibl e reaction pattern inS2- containing styles,dat ano t shown).Thi smean s thatb y applying themutua l weakening hypothesis on the ocurrence of ?*S7-based SC S1S1 genotypes the self-compatibilizing factor tSl is not expressing the pollen part of the SI allele but probably an independent gene. This was already suggested by the results obtained by Thompson et al (1991), although they did not rule out the possibility that onlyth epollen-par t wastranslocated . Re-evaluation ofol dmateria l ofOlsde r and Hermsen(1976 )b yIE Fprove dth epresenc eo fa previousl yundetecte d self-compatible SISI plant. SI tSl pollenha sprobabl y acertativ edisadvantag ecompare d with theothe rthre eS Cpolle ntypes . This means that their tô7-hypothesis is not valid anymore, and one of the other six available hypothesis has to be accepted that was previously rejected because of the absence of self- compatibleSlSl-tSl genotypes.Th eexpressio n tSl istherefor e actually anincorrec t one. Wehav e strong indications that self-compatibilizing factors likethi s "tSl" dooccu rmuc hmor e frequently than isgenerall y believed.

Forproducin gteste r genotypes,suc ha sth e5-homozygotes ,showin g typical Si-expression, it was essential to selfth egenotype stha t showed strong Si-expression. Incertai n cases, such as 6105-06 and6105-0 8(Tabl e 1),i twa srarel ypossibl et oobtai nberr y set and alimite d amount of seedsan d progenies.Th eperformanc e ofthes eprogenie swit hregar dt ovigour , flowering andpolle n fertility (Table 4), however, did frequently not reach the acceptable levels observed in the progenies generated from the basic Fis (Tables 2 and 3).Moreover , the number of really well performing progeny genotypes in the case of 6105-6 and -8 was very low. Obviously, it was necessary to produce moreprogenie s after selfing genotypes that werewel l performing and showing a strong SIreaction . Inothe rplan t species sucha sPetunia andNicotiana , theso-calle dbu dpollination s are practised for producing seeds and progenies from SI genotypes (Pandey, 1963; Shivanna and Rangaswamy, 1969;Clarke tal , 1990).Thi smethod ,however ,wa sno tapplicabl ei npotato ,becaus e the stigma becomes receptive only during anthesis when the exudate becomes available on the stigmaticsurface . Self-pollinations, usingth epseudo-compatibl e genotypes forproducin g seedsan d progenies, could be another option for obtainingS-homozygotes . 38 Chapter2

Flaschenriem andAsche r(1979) ,Dan aan dAsche r(1985,1986a,b )an dLied lan dAnderso n (1994) investigated aspectso fP Ci nPetunia hybrida. This speciestrace sbac kt o interspecific hybridsan d displays SC or ahig h level of PC. They often found strong PC at levels difficult to distinguish from SC.Expressio n ofP Ccoul db ea teithe rth epolle n orth estyla rside .Nevertheless ,thi s system mayhav emor ei ncommo nwit hth ehybri d system ofS.tbr xS.ver, (like S-locuslinke d SC-factors and segregation of Unilateral Incompatibility factors) than with the SI system in diploid potato (Eijlander et al, in prep.). Complications and deviations from the normal SI system in hybrids is discussed by Trognitz and Schmiediche (1993). These authors tried to integrate the incongruity hypothesis ofHoogenboo m (1973)wit h thenorma l gametophytic self-incompatibility hypothesis. Becauseth etyp e ofpseudo-compatibilit y investigated here seems tob e aheritabl e character, both from the male and the female side, the Si-expression in the progenies might be weaker and the selectionswil lb emor e frequently unreliable. Itha sbee n observed as likely for several crops that PC may be polygenic and heritable ( Mather, 1943;Takahashi , 1973;Henn y and Ascher, 1976; Litzow and Ascher, 1983)an doffer s agoo d explanation why inbreeding can lead to an increased level of PC. Invie w ofthis ,i ti sessentia lt o avoidP Cthroug hcarefu l progenytestin g in such genotypes sotha t completely predictable types of self-incompatible genotypes are selected. The problem with bypassingth e SIreactio n inth e stylei stha t theremigh t bea constan t selection for PC expression onth epolle n side.Althoug h thisi sunavoidable , itmigh tb ereduce d init seffect . Forreducin g this selection, stylarP Cclone sca nb e usedwhe n onlystron gS Ia tth epolle n sidei srequire d (and vice versa). The use ofP C is out of the question when pollen and style of thedesire d ^-homozygotes areuse d intes t crosses,unles s alarg eprogen y canb escreene d forreliabl e Si-clones. Additionally, the use of young flowers and at least one strongly SI parent is preferred for the production and selectiono fS I5-homozygotes .Her eth e counterfeit pollination withpolle n from appropriate clones like IvP 35, 48 and 101 have proven to be of great help for obtaining otherwise extremely rare genotypes.Thi swa sals oobserve d fordifficul t interploidyan dinterspecifi c crosseslik eS.tbr x S.acl andS.sto xS.tbr (Iwanag a etal , 1991; Singsit andHanneman , 1991;Brow n andAdiwalaga , 1991). Aswa s alreadydetecte d in someo fth e

reaction against pollen from diploid relatives, expressing an identical 5-allele. A S1S2S3S6 genotypewa s completely incompatiblewit hSI, S2 an dS3 polle n (datano t shown),thu s indicating thatther eca nb ea wid erang efo r^-glycoprotei nconten ti nth estyl eand/o rth eglycoprotei n content needed for areliabl e SIreaction . However, differences between expression levels ofth eS-allele s may reduce the efficacy of some of the ^-alleles when the weakest are down-regulated, as is suggested byresult so fKirc h eta l(1989 )an dEijlande r eta l(i nprep ,se echapte r3) .I tals oindicate s that S-homozygousdiploid s canb e quite useful intestin g tetraploids. Theter mpseudo-compatibilit y hasofte n beenuse di nothe rplan t species although the definition of this expression isno tclea rcut .I na broa d sense,i tha sbee nconsidere d as 'leakage' of a functional incompatibility system. The criteria for considering a genotype to be pseudo-compatible are generally arbitrary.Th ebasi sis ,however ,th eleve lo fsee dse ti na self-incompatibl e genotype after self-pollination ascompare d withth emea n seed seto fth epopulation , expressed as seeds/berry or thenumbe r of seed bearing berries/pollination. Such seed set can also result from a system where aself-compatibilisin g factori soperativ e(fo r discussion, see,Rowlands , 1964;Olsde ran dHermsen , 1976;Hermsen , 1978a and 1978b).Pseudo-compatibilit y has alsobee n observed in several crops after some cycles of inbreeding of self-incompatible genotypes (De Nettancourt, 1977). The mechanism of the origin of pseudo-compatibility in these crops is not clear yet. In the present investigation, aswa s alsoreporte d earlier(Olsde r andHermsen , 1976;Hermsen , 1978aan d 1978b), therewer egenotype s thatwer edifficul t tob eclassifie d either asP C or SC. Theywer e considered to be PC. Genotype 6221-17 (Table 3) is a typical example of the fact that even SC may be unreliable init sexpression . Iti squit e possible that even thispolle n expressed factor is influenced by modifier genes as has been observed for a comparable S.verrucosum-derivedself-com - patibilizing factor (Eijlander, unpublished). The level of seed set upon selfing was initially considered asa nindicatio n for aputativel y useful level ofP C in order to easily obtainS-homozy - gotes but this proved to be incorrect because of its inheritable character. In any case, all those genotypes that showed seed set upon selfing were eliminated and, as aconsequenc e of this, only typical heterozygous SI genotypes werepresumabl y selected. Such acarefu l selection was indeed effective as is evident from the fact that alarg e majority ofth e genotypes (20 out of 29, Table 3) was strictly self-incompatible after several rounds of selection during different years.

Becauseo fth eimportanc e ofplan t vigour, fertility, avoidance of lethal genes andth e high degree ofheterozygosit yrequire d inth eprogen yplants ,th eperformanc e ofth ebasi cpopulation s used in thisinvestigatio n deserve attention.Th ethre eorigina ldihaploi d clones,G254 ,G60 9 and B16 were known to be vigorous, profusely flowering and fertile (Olsder and Hermsen, 1976). The Fl progenies of these clones, on average, performed much better than the progenies obtained from 40 Chapter2 selfings or inbreds (Table 2). Furthermore, the number of ^-heterozygotes that were selected originated predominantly (19ou t of 31) from theprogenie s ofth eorigina l Fis (Table 3)whic h is evenmor estrikin g whenth erati o of SI/S C is considered. This clearly indicates that even within a restricted number of genotypes that wereuse d in this investigation, competent SIparent s could beselected , givingrise t o desirableprogenies .Fro m thepoin t ofvie w ofth e good performance of the progenies of the basic Fis and some of the inbred lines, it should be concluded that the establishment of inbred lines ofdiploi d potato,comparabl e tothos e ofmaize ,migh t bepossible .

Invie w ofth erecen tmolecula r approachest oelucidat eth ephenomeno n ofself - incompatibility in diploidpotat o( suc ha sgai nan dlos so ffunctio n analysisi ngeneticall ymodifie d plants),geneticall y well defined plant material is essential. Part of the material selected in this investigation, which includes both S-homo- and heterozygous genotypes, expressing typical SI reaction, was highly valuable for thisresearch .Th etw omos timportan t factors for using suchselecte d clones as testers areth e absenceo f PC anda goo dpolle nfertilit y when used asa pollinator . This stresses thenee d for an extensive screeningprocedur e asdescribre d here. Chapter 3

Manipulation of self-incompatibility in diploid Solanum tuberosumL using

sensean d antisenseconstruct s ofS-RNas egene s

Abstract

Diploidpotat o (S.tuberosum) expresses aon elocu sgametophyti c self-incompatibility system. The so-called ^-glycoproteins are style specific and arehel d responsible for the stylar part inth e self- incompatibility (SI) interaction between pollen and style. Thepotat o genescodin g for SI and S2 glycoproteins havebee n isolated molecularly anduse d forth econstructio n ofvariou s homologous and heterologous sense and anti-sense constructs. Six different diploid potato clones, expressing either SI orS2, hav ebee ntransforme d with theseconstructs . Theanti-sens e approachwa smos t successful whenth e35S promote rwa sused , asoppose d to anti- senseversion sdrive nb yth e52-RNas eo rSK2 promoters . Transformation of genotypes displaying SI orS2 incompatibilit y reactionsresulte d incompatibilit ywit hth ecorrespondin g SI orS2 pollen , that gaveincompatibilit y reactions inth enon-transforme d genotypes. The sense approach confirmed the finding that the^-glycoprotein s aredirectl y involved in the SI reaction,becaus eth eintroductio no f stronglyS2 expressin gconstruct s resulted inth epredicte d S2 pollen inhibition. The constructs based on the SK2 promoter were much more efficient in this respect, than those driven by the S2 promoter. Introduction of S2 driven by the SK2 promoter resulted not only in gain-of function, but in some cases also in an efficient down-regulation of endogenous alleles likeS3 o r570 . Thus, the anti-sense approach gave a specific suppression ofth e target alleles,wherea s the sense approach couldno t onlyad d ane w incompatibility group,bu t could also simultaneously suppress allothe r^-alleles .Thi spossibl e effect should betake n into consideration whenever thesetype s of constructs will be used for theproductio n ofhybrid s inbreedin g programmes.

This chapter is submitted for publication in aslightl y modified version as:Ronald Eijlander, Michael Ficker, Ester Abad ICantero ,Munikot e S.Ramanna ,Richar dD .Thompsonan d EvertJacobsen .Manipulatio n ofself-incompatibilit y in diploid Solanumtuberosum Lusin g sense and antisense constructs ofS-RNas e genes 42 Chapter3

Introduction

Thephenomeno no f self-incompatibility (SI)occur si nalmos tal ldiploi d tuberousSolanum species, including thediploi d forms (dihaploids,2n=2x=2 4) o fth ecultivate d (tetraploid, 2n=4x=48)potato , Solanum tuberosum L{tbr). Asi nth ecas eo fothe rSolanum species,i ndiploid-tó r SIi s genetically controlledb yth egametophyti c systembase d ona singl e locus,th eS-locus ,wit h multipleS-alleles . Such5-allele swer edetecte d inthre edihaploid s oftbr throug hdiallel ecrossin g and were identified asSI, S2, S3 an dS4 b yHermse n(197 8a,b) .Althoug hth edihaploid stha twer eheterozygou s for S- alleles (e.g.,S1S2) showed typical Si-reactions, there were also similar genotypes that were self- compatible (SC).Th e SCreactio n was causedb y apolle n expressed factor, called tSl, which was believed tob e atranslocatio n ofth epolle npar t ofth eSI allele (Hermsen, 1978a;Thompso n et al, 1991).Recently ,usin gth esam ebasi cmaterial ,al lpossibl eS-allel eheterozygote s (viz.,S1S2, S1S3, S1S4, S2S3, S2S4 and S3S4) with well defined SI reactions have been selected. In addition, homozygous genotypes for most of the^-allele shav ebee nproduce d and arebein g maintained as tester stocks (Eijlander etal. , 1997).

Besidesth etraditiona lmethod so fdetectio nan dclassificatio n ofS-genotype s( Olsde ran dHermsen , 1976; Hermsen, 1978a,b), gene products corresponding to the four ^-alleles have also been molecularly characterised inth e abovementione d plantmateria l (Kirch et al, 1989). By analysing protein extracts from the styles of defined S-allele genotypes through two dimensional gel electrophoresis,th epresenc eo fa groupo fbasi cglycoprotein swa sestablished .I twa sfurthe r shown that each of the four S-alleles was associated with the presence of polypeptides differing in their isoelectricpoint san dwit hth ehel po fthes eSi-associate d proteins(th e^-glycoprotein so rS-RNases ) the S-genotypes could be clearly distinguished. A comparison of sequence homologies of S- associated glycoproteins oftbr reveale d similarities with those ofothe r solanaceous plants such as Nicotiana alata and Lycopersicon peruvianum (Kirch et al., 1989; Peil, 1995). Within the Solanaceae,th ehighl ybasi cglycoprotein s havebee n shownt oposses sRNas e activity and,becaus e ofthei rspecifi c associationwit hth eS'-locu sthe yar ecalle dS-RNase s(Cornis he tal. , 1987;McClur e etal. , 1989;Clark ean dNewbigin , 1993;Newbigi n etal. , 1993;Sims , 1993;Kowyamae t al., 1994; Royo etal. , 1994).I nadditio n to theseS-RNases , twomor e abundant proteins, that are designated asSKI andSK2, wer eals oconstantl ypresen t inth estyle s ofmos t ofth e genotypes and these were non-S-linked pistil specific proteins. Of the two non-S-linked pistil specific proteins, the most abundant SK2 polypeptide has been shown to be specifically located (through a immuno- cytochemical method)i nth e styles andprove dt ob ea nendochitinas e (Wemmere t al., 1991;1994) , showing homologies with thetomat o ChiPgene .

Genomic and cDNA clones, corresponding to pistil specific proteins, have been isolated and characterised inpotat o (Kaufman et al., 1991;Kirch , 1992;L i et al., 1994;Wemme r et al., 1994; Senseand antisense effectsofS-RNase gene constructson SI 43

Kirch etal. , 1995; Peil, 1995; Fickere tal , 1998a,b).Thes e included two alleles,SI andS2, ofth e 5-locus(Kaufma n etal. , 1991)an dSK2 o fa non-S-linke dgen e(Wemme re tal. , 1994).A functional analysisha sbee ncarrie d out forSI- an d52-RNas epromoter s aswel l asth epromote r OÎSK2gen e by using GUS as reporter (Ficker et al., 1998; and unpublished results). These analyses have indicated that the expression patterns of these genes may be strongly dependent on the type of promoteran dth ehos tplan tint owhic hthe yar eintroduced . Similarfunctiona l analyseso nS-RNase s (promoters and especially coding regions) in other solanaceous plants like Tobacco and Petunia have established that S-RNases are indeed responsible for SIreactio n of the styles (Huang et al., 1994; Lee et al., 1994; Murfett et al., 1994). RNase activity was shown to be essential for a functional inhibition and gradually more information has become available about the identity determinants inth ehype r variableregion s ofth eS-Rnase s(M eCubbi n et al., 1997;Matto n et al., 1997).

In view of the available functional information, together with the cloned genes and defined plant material,i twa srelevan tt otes twhethe rth ebiologica l activity(i.e. ,Si-reaction ) indiploi dpotat oca n bemanipulated , asearlie r described forPetunia andNicotiana, throughth eintroductio n ofS-allele s intoappropriat eplan tgenotypes .I nthi scontext ,antisens eversion so fhomologou san dheterologou s constructs of SI and S2 alleles as well as a sense version of the S2 allele were introduced into defined genotypeso fpotat othroug hgeneti ctransformation . Theresult so fth etransgeni c expression of5-allele si n different types oftransformant s are described and discussed inthi sarticle .

Material and methods

Basicplant material Two groups of diploid potato (Solanum tuberosum, 2n=2x=24) genotypes were used for genetic transformation. The first group consisted of two self-incompatible (SI) genotypes, S1S4 (code: 195/5, Kirch et al., 1989) and S3S10(code : 6618-10-IV, El-Kharbotly et al., 1995). The second group consisted offou r self-compatible (SC)genotypes , viz.,SI S3: 6486-0 4 (R2);SI S3: 6486-19 (R5); S1S10: 6486-09(R3 )an dS2S10: 6487-09 (V).Th e latter group ofclone spossesse d pollen- based SC (homologous to the so-called tSl-like reaction, data not shown) ,bu t expressed stylar specific SI reliably and was related to two interrelated well-transformable genotypes, A16 (El- Kharbotly et al., 1995) and 1024-02 (Jacobsen et al., 1989),se eals o Appendix 2.I n addition, six diploidhomo -an dheterozygote s for 5-alleleswer euse d asteste r pollinators in ordert o verify the SIreactio n inth etransformants . Amongthes e sixgroups ,thre ewer e SI-homozygotes,SI SI (6496- 01, 6496-04,6499-04) ,S2S2 (6233-12 ,6234-05 , 1140-02)an dS3S3 (6499-01);an dth e other three were S-heterozygotes:S1S2 (6222-39) , S2S3(6222-40 ) (Eijlander et al., 1997) andS2S10 (4002 - 04)(El-Kharbottl y etal. , 1996). 44 Chapter3

Plasmidconstructions Allth e constructs used (without showing vectors) are shown in figure 1. Plasmidscontainin gthos einsert sar eaddresse d byth einser tnames .Th econstructio n ofinsert san d Plasmidsi s described below.

P35S - S1AS L P CaMV DSS 1 cDNA

P35S - S2AS | P CaMv" ^«^ESSTCaMV

(P35S - S2AS)2 | P CaMV ~jf^ s?rnNApiBB B p CaMV ^GfWWlA TCaMV

S2 genomic | / / P S2 9.9k b )| S2 %M

P(0.7)S2 - S2 P S2 0.7 kb S2

PSK2- S2 "7 / P SK2 1 kb )T S2

PSK2 n - S2 | / / P SK2 1 kb >n| S2~

Fig. 1. Schematicrepresentatio n ofth e inserts inth ebinar y vectors pGDW32 (P35S-S/AS ) and pBin 19 (from P35S- S2AS toPSK2CÏ-S2 ).Externa l arrow headedpolygons :promoter s (P) ofCaM V and of the style specific genes S2 and SK2. Promoters larger than 0.7 kb are indicatedb ybroke n polygons, ii - enhancer fragment. Internally arrow headed boxes: sense (>) and antisense (<) orientated SI cDNA,S2 cDNA or the intron (striped box) containing genomic S2 coding region. Black boxes: polyadenylation sequence (T = terminator) of the CaMV or 52-RNase gene.

Anti-sense (AS)S1&S2 Type: P35S - SIAS. Plasmid pGDW57AS was constructed by cloning into the EcoRI site of pGDW32 (Wing et al., 1989) a partial EcoRI digest fragment (P35S - SIAS) of 1,48 kb of pAP57AS. pAPS7ASwa sconstructe d by insertiono fa 0. 7k b Sall/BamHI fragment ofSI cDNA intoa 35S-NO Scassett eo fvecto rpA P(Kirch ;Pereira ,unpublished).Type :P35 S -S2 AS. Plasmid p35S-52ASwa s constructed byclonin g into theNcol andBamHl sit e of pRT104GUS the 320 bp BgUVNcol fragment of the 52-RNase cDNA (plasmid pHK22, unpublished, genomic clonepublishe d byKaufman n et al, 1991). Type: P52 - 52AS. Plasmid P52-52AS was constructed by cloning into the NcoVHindm siteso f plasmidp522- 2(Ficke re tal.,199 8b )a fusio n ofa 32 0b pBglWNcol fragment, extendingfrom b p 127 to44 7o fth e52-RNas e cDNAan da 25 0b pBamMJHindm fragmen t ofpRT104GU S containingth e CaMVterminato r(Töpfe r etal. ,1993) . Type:P5 0 -52AS .Plasmi dpSK2/l contains thepromote r ofth e style-specific endochitinaseSK2 (Wemmer et al., 1994)an dha spreviousl ybee n described by Ficker et al (1998a, inpress) .Plasmi d pSK2-S2AS was constructed by cloning intoth eNcol andHindlll site so fplasmi dp5r^2/ l a 570b p NcoVHindlll fragment ofplasmi d p52-52AS, containing a fusion of the52-RNas e coding region in antisenseorientatio n andth eCaM Vterminator . Senseand antisense effects ofS-RNase gene constructson SI 45

SenseS2 Type: S2genomic .Plasmi dpBinS 2wa sconstructe d by theinsertio n of a 12,9k bSaä fragment ofa genomicclon eo fth eS2 RNas e(startin ga tapproximatel y 9.8k bupstrea m fromth estar tcodon )ou to f 1G131/1(Kaufmann , 1991)a tth eSaK sit eo fpBinl 9 (Bevan,1984) . Type: S2 genomic. Plasmid p(Q.l)S2-S2 was constructed by replacing the GUS-CaMV-terminator- fragment ofpS24 (Ficke re tal. , 1998b)b ya 1.8genomi cNcoV Saä fragmen t of52 ,th eNcol restrictio n sitecontainin gth estar tcodon . Type: VSK2-S2. PlasmidpSK2-S2 wa sconstructe db yclonin gint oth eNcol andHindUl siteso fpSK2/\ a 1.7k bNcoVHindUl fragmen t ofplasmi dpLAT52S 2containin gth eS^-RNas ecodin gregio nan dS2- RNase3 ' flanking sequences(Kirc he tal. , 1995). Type:PSK2 -S2, Q-enhanced .Plasmi dpSK2DS2 wa sconstructe db ycuttin gpSK2S2 wit hNcol an d removingth enucleotid eoverhan gwit hSI nucleasefollowe db ya nHindUl digest .Th eQ-sequenc ewa s constructedb y annealingpartiall yoverlappin gnucleotides ,Kleno wfill i nan dcuttin gwit hNcol. This Q sequencei sblun ta tth e5 'en dan dcontain sa Ncol sit ea tth e3 'end .Th eQ. sequencewa s fused with a 1.7k bNcoVHindUl fragment of'pSK2S2 andth eresultin g fusion wasclone dint opSK2S2 processe d asdescribe dabove .Th e6 8b pQ sequenc ecorrespond st oth eleade rsequenc eo fth eTM VRN Astrai n Ul andact sa sa translationa lenhance r(Wilso ne tal. ,1993) .Th eoligonucleotide suse d for constructing theQ. sequencewere : sense 5'GTATTTTACAACAATTACCAACAACAACAAACAACAAACAACATTACAA3' antisense 5' CCCCATGGTAATTGTAAATAGTAATTGTAATGTTGTTTGTTGTTTGTTGT 3'

The afore mentioned inserts of plasmids pSK2Q-S2,pSK2-S2 and pSK2-S2ASwer e cloned into pBIN19a sEcoRVHindUl fragments, p(0J)S2-S2a sa SalUHindUl, p35S-S2A Sa sa HindUl fragment and pS2-S2AS as aKpNllHindUl fragment. Atande m insertion ofP35 S -S2AS resulted in (P35S- S2AS)2, with a mutated, non-cleavable HindUl site. All plasmid constructions were checked by restrictionmapping .

DNA methodology DNAisolation , subcloning, restriction analysis and screening ofth egenomi c librarywer e carried outusin g standardprocedure s (Sambrook et al., 1989).

DNA sequencing DNA sequencing wasperforme d with anautomate d DNA sequencer (Applied Biosystems model 373A) using the Ready Reaction DyeDeoxy Terminator Cycle Sequencing Kit (Applied Biosystems) accordingt oth emanufacturer s instructions.

Transformationprocedures andvectors For plant transformation the plasmids were introduced intoAgrobacterium tumefaciens LBA4404 46 Chapter3

(Hoekemae tal. , 1983)b y eitherelectroporatio n(Wen-ju n andForde , 1989) ordirec t transformation ofcompeten t cellsaccordin gt oHöfge n andWillmitze r (1988). Potatotransformatio n wascarrie dou t intw oways : 1) asdescribe db yVisse r(1991 )o r 2)accordin gt oFlips ee tal , (1994).Transgeni ccall i obtained by the method of Visser were selected on MS-medium (Murashige and Skoog, 1962) supplemented with 10g/ 1 sucrose, 2 mg/1 zeatin, 0.01 mg/1 NAA, and 0.1 mg/l GA3. Shoots were inducedo nMS-mediu mcontainin g3 0g/ 1sucrose ,0.2 5mg/ 1benzyladenin ean d0. 1 mg/ lGA3 .Expiant s inoculated accordingt oFlips ee ta l(1994 )wer e transferred twoday safte r inoculation onto selective media(kanamyci n10 0mg/ 1o rhygromyci n2 5mg/1 )wit hM Smediu msupplemente dwit h20g/ lsucrose , lmg zeatin andbot h 200m gcefotaxim e andvancomycin . Hygromycin resistancewa steste dunde ra monthlyrecurren t selectioncycl eo ftw oweek so f2 5mg/ 1 hygromycinan dtw oweek so fn o selection pressure.Selecte dshoot swer etransferre d tohormone-fre e MS-medium. Allmedi awer esupplemente d with20 0mg/ 1cefotaxim e and5 0mg/ 1kanamyci no r 10 mg/1hygromycin .

Proteingel electrophoresis Up to 50 mg of plant tissue was ground in an Eppendorf tube with 20-100 ß\ 5m M potassium phosphatep H 6.0, 2.5 %(w/v )sucrose , 0.1% (v/v )b-mercaptoethanol ,usin ga ground-glas spestle . Singlestyl eextract swer emad ei na volum eo f2 5(A extractionbuffer.Tota l antherextract swer emad e by collecting all anthers of a flower and grinding them in lOOfA buffer. After centrifugation of the homogenatea t 14000gfo r 15 min,th esupernatan twa sfractionated o nhorizonta l thin-layer isoelectric focussing (IEF)poly-acryl-amid egel s(pH3.5-10 )(Schmidt-Stohn , 1979)o ragaros egel s(Hypur ege l VG1020 ,Isola binc) .Separate dprotein swer eelectr otransferre dt onitrocellulos efilter susin ga 'semidry ' procedure(Kyhse-Anderson , 1984)wit ha Sartoblo tII S(Sartorius ,Göttingen) .

Detectionof IEF-separatedand electroblotted proteins Immunodetection was carried out using the ECL Western blotting detection system (Amersham Buchler, Braunschweig) basically as described by Kaufmann et al. (1991). Silver staining was performed asdescribe db yKirc he ta l(1989 )fo rth ePolyacrylamid e gelsan daccordin g tocompan y specifications for the agarosegels .

Monitoringof incompatibility Using transformed and non-transformed plants (controls) as pistillate parents, test crosses were madeb yusin gpolle nfrom th ethre eclasse so fS-allel ehomozygotes ,SI SI, S2S2an d S3S3 orwit h thepolle n of the heterozygotes S1S2 (compatible with all transformants), S2S3,S2S10 orS3S10 depending upon the genotype andth ereactio n thatwa st ob e monitored. Styles were harvested 48 hours after pollination andfixed an d stained according to the modified technique of Martin (1959) and pollen tube penetration was observed under a fluorescence microscope (Eijlander et al., 1997). Senseand antisense effectsofS-RNase gene constructson SI 4?

Results

Antisense suppression

Antisenseeffect of the SI alleleconstruct in S1S1 0 transformants After transformation ofth ediploi d potato clone,6486-0 9 (SIS10, pollen-expressed SC),wit h the SJ antisense construct pGDW

Aminimu m oftw ocopie s ofth econstruc t waspresen t in all selected transformants except for the clone R3-S35S1AS-S that had asingl e copy.I n agreement withth e expectations areductio n in SI reaction, as compared to the control, was observed in tube growth ofpolle n from SIS J homozy­ gotes,wherea s inth ecas eo fth econtro lplan t therewa s complete inhibition ofpolle ntub e growth of SI pollen. The transformants permitted different degrees of pollen tube penetration, with full compatibility being observed inR3-P35S7AS-2 4(Tabl e 1&2).I norde rt overif y whether the styles ofal lth etransformant s confined toth eexpecte dnor m ofth eS Ireaction ,the y werepollinate d with S2 pollen and all were found to be fully compatible (Table 2). With regard to the style-specific proteins,IE Freveale d thatther ewa s no strong reduction ofth eSI glycoprotein detectable in four

Table 1. Expression of antisense SI by P35-S7AS in five transformants of theSIS10 - clone 6486-09 (R3). The plants were analysed for minimal copynumbe r ofpGDWS/a s (insert: P35S-57AS), the pollen tube ingrowth ofSI andS2 pollen fromS-homozygote s (SISI: 6496-01, 6499-04 ;S2S2: 1140-02) in the styles by fluorescence microscopy (C= compatible, PC = Pseudo compatible, I= incompatible) and for the banding pattern after silver staining of single style extracts on IEF gels (— = absent, -= barely visible, + = clearly visible, ++ = apparently normal level, +++= higher).

STreactio n against SI and S2 TF.F-partarn SI S2 Untransformed 0 I c 57-H-, SW ++, SK1++ R3-P3557as-3 >2 I-C c S1++, S10 ++, SK1++ R3-P35S7as-8 >1 I-PC c S1++, SI0++, SKI ++ R3-P35S7as-24 >2 c c SI +/++, S10+++, SK1 + R3-SP351as-29 >2 I-C c SI ++, SI0++, SKI -^ R3-P35S7as-32 >2 I-PC c SI ++, SI0++, SKI ++ 48 Chapter3 outo ffive o f thetransformants . Asa comparison , thestyl eexpresse dSKI, togetherwit hS10, th e levels of which were expected not to be altered, are also indicated in table 1. This revealed that transformant R3-P35S7AS-24, which was fully compatible with SI pollen, showed a detectable relativereductio n ofSI, wit h an estimated 50%whe n compared to either itsS10 ban d or SI in the control plant.

Antisenseeffect of S2allele constructs in S2S1 0 transformants Forth eantisens eS2 approach ,tw oconstruct s wereuse dtha t werebase d on a32 0b p S2 fragment. The constructs F35S2AS and (P535S-52AS)2, were used for the genetic transformation of the diploidpotat oclon e6487-0 9 (V);th elatte rconstruc t contained atande m duplication ofth e former insertwithi n the vectorpBinl9 . CloneV ha dth e genotype OÎS2S10an d showed like R2, R3an d R5, apollen-factor-base d self-compatible reaction. A total of 30 transformants, 20 based on the P35-S2AS construct and 10 on the (P35-S2AS)2 construct,wa steste dfo rthei rS Ireaction ,b ymonitorin gpolle ntub egrowth ,usin gpolle nfrom S2S2 homozygotes instead of selfing them.Th eclone sha d insert copynumber srangin g from 1 to3 .

»WWwWWWW

S10- S2 •

SK1- SK2- m. ^jLinm Control: S2-inc. as-eff: S2 comp. Sense-eff: S2inc . V66666665

Figure 2.(Left) . Fluorescence microscopypicture s ofpolle n tube Figure 3. (Right) Silver stained Iso Electric penetration in styles. Focusing pattern of style extracts of A: Control pollination with 52-pollen on clone V (S2S1Ö): untransformed (V) and (P35S-S2AS)2 incompatible. transformed clones 5 and 6. Single style B: Pollination with S2 pollen on clone (P35S-52AS)2-V-6, extracts show a strong reduction of S2 only in showing antisense 52 effect: compatible. clone (P35S-52AS)2-V-6 but not in transgenic C:Pollinatio n withS2 polle n on clone PSK2Q-S2-VI-2, showing clone 5 or untransformed clone V. 52 and 570 sense S2 effect: incompatible. Clone VI is S3S10. Pollination of are S-glycoproteins. SKI and SK2 are other clone VI with S2 pollen reacts as inpane l B. style specific proteins. Senseand antisense effects ofS-RNase geneconstructs onSI 49

A majority (23) showed a stable SI reaction. However, in some cases there was a tendency for pseudo-compatibility whenth eplant swer estil lyoung ,bu t atlate r stagestypica l SIwa s evident in all of them. The remaining seven transformants showed clearly antisense effects. Among those seven,thre ewer ephenotypicall y unstableb y showing full SCan dS Ireaction si ndifferen t flowers of one and the sameplan t when pollinated by S2S2plants . In theremainin g four transformants, P35-S2AS-V-5&8,(P35-S2AS)2-V-4& 6(cop y numbers 3, 1, 1 and 2respectively) , astron g and stable antisense effect was observed in all flowers. S2polle nwa s compatible and full pollen tube penetration wasvisible ,indistinguishabl e from compatible control pollinations. Accompanying thechang ei nS Ireactio n(Figure s 2A andB) , IEF showed that the52-RNas eban d wasgreatl yreduce d inintensit y (Fig.3 :arrowhead) ,thu scontrastin gwit hth e limited effect shown earlier for the antisense SI constructs (Table 1). There was clear correspondence between the reduction ofS2 protei n and thecompatibl epolle ntub e growth of S2S2polle n inth e styles.

Sense expression studies.

Expressionof genomic S2 constructs. pBin£2 was used to introduce a 12kB genomic S2 clone into two SIS3 genotypes (with pollen- mediated SC).R 2 (6486-04) gave riset o 38, andR 5 (6486-19) yielded 32flowerin g transgenics. These 70plant swer eteste db y IEF forS2 expressio n instyle s andleave s and innon e of them was the 52-RNase detected. Some test crosses with S2S2 pollinator genotypes were made, and , as expected from absence ofth eS^-RNase , allo fthe mreacte d compatibly. Twodiploi d transgenics ( pBinS2-R2-l andpBiaS'2-R5-35 ) with unknown T-DNA copy number were selfed (<8>) and 35 transgenic offspring plants from each were tested for S2 expression. The two offspring plants pBmS2-R5-35-33 &-3 9showe da fain t52-ban dwhe nstyla rextract swer esilver-staine d afterIEF , in intensity comparable withth e S2-bands shown in figure 3fo r the antisense S2clones . Thecut-bac kpromote rversio no fpBin£2 ,~P(0.7)S2-S2, wa stransforme d intoSI S3 clon eR 2 .On e out of2 2 flowering transgenicplant s showed awea k S2band , asdescribe d above. Asexpecte d from thelo w levelo fexpression ,non eo fthes eweakl yS2 expressin gtransgenic swer e incompatible with ^-pollen.

Gainof SI function byheterologous sense S2 constructs. The two heterologous constructs ?SK2-S2an dVSK2Q-S2, containing theSK2 promote r and the coding region of52 , differed only by the absenceo rpresenc e ofth e Q-enhancer fragment, which is a translational enhancer. They were expected to be style specific and to give rise to ahig h S2 expression (Ficker et al., 1998, in press). The YSK2S2 construct was introduced into clone VI (S3S10) andPSK2Q. S2int oth eclone sV I(S3S10) and 195/5 (S1S4).Th epresenc e of T-DNA was confirmed for 14régénérant so fclon e 195/5b yPC R analysis and for all 11teste drégénérant so f 50 Chapter3

Fig. 4. Expression ofth e pSK2 based heterologous constructs, driving S2 Panel A(to p left). IEFpattern s of style extracts ofPSK2C1-S2 transformed clone 195/5(SJS4) andcontrol s (S1S2 S1S4 and S2S2).

Panel B (middle left). Western blot with monoclonal anti bodies against SI, as a control on panel C, specially detecting the presence of SI as detected by silver staining in panel A. PanelC (bottom left). Western blot with monoclonal anti bodies against S2, confirming thepresenc e and identity of S2 in transgenics of panel A.

PanelD (top right). Over-expression ofS2 b yPSK2-S2 an dVSK2Q-S2 i nclon e V(S2SI0) and VI (S3S10). V40 shows down-regulation for SIOan dVI 8fo r bothS3 andSIO. VI 6 showsnorma lgai n of function {S3S10 ->S2S3S10). V and VI areuntransforme d controls.

theclon eV Ib y Southernblotting . Southernblottin g showed acop y number ranging from 1 (eg FSK2-S2-VI-2 andSK2Q «-VI- 1) t o 6 ÇPSK2ÇI-S2-VI-2). Ten out of 14 transformants of clone 195/5 showed IEF detectable «-bands (Fig. 4A), up to endogenous levelso fth eothe rS-RNase so rhigher ,an ds odi dnearl yal l(1 6ou to f 18) ofth etrans ­ formants of clone VI (Fig. 4D). Two transgenics (?SK2nS2 -VI-2&8) showed in several cases much higher«-content stha nth eendogenou s level.Th eidentit y ofth eIEF-detecte d bands onth e «-position as «-RNase was confirmed by Western blotting (Fig.4B,C). The strength of the SI reaction coincided with the amount of« glycoprotein present. Plants with high levels of« expression showed strong incompatibility reactions and plants with normal levels allowed for a deeperpolle ntub epenetratio n (Fig.2C )befor e thearres t wascomplete .Man yplant s were unstable Senseand antisense effectsofS-RNase gene constructson SI 51 inth eexpressio n ofth etransgen ei nth e course oftime .Hig h temperatures reduced the level of5 2 in most of the plants, causing a shift towards compatibility with S2 pollen. The most stable genotypes (1/3o fth etransformants ) inV I and 195/5wer e incompatible withS2 unde r all circum­ stances.Som egenotype s showed abarel y detectable amounto fth eS2 glycoprotei n inentir e anther extracts. Pollen from these plants did not show a shift from SI towards SC on non-transformed plants of the same genotype.

S2-sense inhibitionand -over-expression. Of the selected 21 ?(0.7)S2-S2(sense ) transformed plants (clone V, S2S10 SC),non e showed a clear-cutinhibitio n effect onth eendogenou sS2 incompatibilit yreaction , whichwa si n accordance withth e apparent normal levels ofstyla r52-RNas ewhe n analysed by IEF. The constructs VSK2ÇÏ-S2 and PSK2-S2wer e also introduced into genotype 6487-9 (clone V, S2S10). Transformants wereteste d for SIb ypollinatio n with tester clones.Non e of the2 8 tested plants showed abreakdow n of the SIreactio n against S2pollen . IEF showed S2 levels at least as high as the endogenous concentration and under moderate climatic conditions often an enhanced ^-content was found when compared with S10. Two clones, FSK2-S2-V-8& -40, synthesised (much)mor eS2-RNase ,tha ntha tderive dfro m theendogenou sS10 allele . Evenexces so fcomplet e single style extracts (thus not standardised for total protein content) did show little or no S10 glycoproteinwhe n silver-stained but onlySK2 an dS2-bands ,a si fth eclone swer eiS'2-homozygous . This phenomenon was also incidentally observed for two ,S2-sense transformed S3S10 plants, PSK2Q.-S2-VI-2 &- 8(Fig .4D) .Pollination s with S2S2,S3S3 an dS2S10 plant sreveale d a losso f bothth eS3- an dS10- incompatibilit yreaction ,wherea sth ereactio nagains tS2 wa smaintained ,thu s reconfirming theke yrol eo fspecifi c S-RNasesi nth eincompatibilit y reaction.Th eindicate d effects ofth e constructs used, are summarized intabl e2 .

Discussion.

S2 genomic clones: sense expression and sense inhibition. Long ,S2-RNasepromote r fragments fail to direct high-level style-specific expression of reporter genes intransgeni c potato and tobacco (Kirch, 1992;Kirc h et al., 1995;Murfet t et al., 1995) and distally located cis acting regulatory elements have been postulated for an optimal level of expression. Ficker eta l(199 8a ,i npress )conducte d afunctiona l analysiso nmiddle-lon g and short versionso fth eS2 promote r andhypothesise d thatther ei sa ninteractio nbetwee nth e coding region ofth e^-allel e andit spromote rt oregulat ea prope rtissue-specifi c activity.Thi shypothesi s argues fairly against the expected result. Butw e (arguably) tested thishypothesi s bytransformatio n with aconstruc t having alon gpromote r andon ewit h ashor tone .Becaus eal lconstruct suse di nthi san d the previous studies apparently lacked those distally located regulatory elements, high-level 52 Chapter3

Because the cut-back versions of the S2 promoter showed nearly the same expression levels in activity studies (Ficker et al, 1998b)whe n compared with the largerpromote r versions (up to 9.8 kb),littl esystemati c effect wast ob eexpecte d from our transformants based on either the small or thelarg epromoters .Onl y afortuitou s integration ofth ehomologou s constructs in an "activating" siteo fth epotat ogenome ,containin g suchregulator yelements ,migh thav egive na goo d expression of S2. Passing through a sexual cycle, which also involves recombination, can bring about expression oftransgene stha twer eno t expressedbefor e (see also reviewb y Stam et al., 1997).Fo r thatreason , someoffsprin g wasteste do ntransgen eactivit ya swell .Th elo wleve lo fS2 expression

Table 2. Overview of maximal effects detected inth e antisense, sense and over-expression studies. Blancs:no t determined; -= presenc eno t detected /polle n tube growth arrested; + = presence detected / full pollen tube penetration; blank = not tested; ++ = high contents, < = slightly reduced; <« strongly reduced, barely noticeable.

Construct Host plant IEFdetecte d glycoproteins Pollen tube growth in styles

Clone 5-genotype SI S2 S3 S4 S10 SI S2 53 S4 S10

R2 SJS3 + - + - - - + - + none R3 S1S10 + - - - + - + + (control) R5 S1S3 + - + - - - + - +

V S2S10 - + - - + + - + + -

VI S3S10 - - + - + + + - + -

195/5 SIS4 + - - + - - + + -

P35 SIAS R3 SIS10 < - - - + + + +

P35 S2AS V S2S10 - <« - - + + + + -

(P35 S2AS)2 V S2S10 - <« - - + + + + -

S2 genomic R2 S1S3 + +* + - - - _i_** - +

R5 S1S3 + +* + - - - _!_** - +

P(0.7)S2 S2 R2 S1S3 + + + - - - + -

VSK2 S2 V S2SI0 - ++ - - - + - + + +

VI S3S10 - ++ < - < + - + + +

VSK2C1 S2 V S2S10 - ++ - - - + - + + +

VI S3S10 - ++ - - - + - + + +

195/5 S1S4 + + - + - - - - *) = in offspring after selfing; **) = also in offspring after selling; <«: in offspring Senseand antisense effectsofS-RNase gene constructs onSI 53 in expressing ofsspring was comparable to that ofth e single expressorplan t containing the0.7k b promoter versiono fpBinS 2an dthi swa s in accordance with the findings ofFicke r et al., (1998 b) after transformation of acomparabl e construct into tobacco. Itwas ,du et o the quantitative nature ofth eS Iresponse ,no tsufficien t topromot ea functiona l SIreactio nagains t^-pollen . Thedetectio n of sense-inhibition is therefore highly unlikely if this depends on the transcription level of an inserted construct.

AntisenseSI andS2 effects. Theantisens eeffec t forS2 i sindisputable ,althoug hth ephenomeno nwa sno texpresse d inal lplant s stably. It was shown that the P35 based antisense ^-constructs were transmitted through the gametes. Offspring clones like 1184-01 (= S2S2 +(P35-S2AS)2, derived from crossing (P35- S2AS)2-V-6wit hth eS2S2 clon e 1140-02)showe dth eexpecte d antisense effect (datano t shown). Thehig hpercentag e (80%)o f transformants showingsom eantisens e effect forSI (bypGDWS 7AS , Table 1) is most likely due to the selection procedure,becaus e only a few stably Hyg+ responding transgenicswer eobtained ,wherea sKan +base dconstruct swoul dhav eyielde dman ymor etransgenics . The reduction of the SI-glycoprotei n has even in clone R3-P35-57AS-24no t been so dramatic as observedfo rS2 i nvariou stransgenic sderive dfrom clon eV .Th eS1 0allel ebelong sprobabl yt oth eSU S3!SIR family (Kirche tal. , 1989) becauseSI andS3 specifi c primersallowe d forPCR - amplification andSI probesstrongl yhybridise dwit hS10 i nRFL Panalyse s(dat ano tshown) .Compariso n ofth eSI signalwit htha to f57 0can ,therefore , givea fals e impressiono fabsenc eo fantisens eeffec t onprotei n levelwhe nS10 i sreduce da swell . Antisenseaffectin g allelesha sals obee nreporte db yLe ee ta l(1994 ) forth eS2 an dS3 allele si nPetunia inflata. I.E.Fbase dcompariso nwit hth eSK2 signalma yb e difficult, becausethi ssigna li si ngenera lver ystrong ,i nthi swa ya 50 %reductio no fSI mayremai nundetected , whereasSKI isinappropriat efo rthi sdu et oit sunpredictabl e levelo fexpression .Ther eremaine donl y onetransgeni cwit hsom edemonstrabl ereductio no fth eS-RNases ,bu tthi sreductio nwa sno t asstron g asdetecte d for someantisens eS2 transgenics .Nevertheless ,clon eR3-P35-S7AS-2 4wa sconstan tan d reliablei nthi sacquire dSI compatibility.I ti spossible ,tha tth eS7-RNas e contenti nstyle so fclon eR 5 wasinitiall ylowe rtha nth e.S^-RNas ei nstyle so fclon eV ,an di tmigh t alsob etha ti nthes eclone sth e activity ofSI andS2 RNas e differs, thusexplainin gwh ya nantisens einduce dreductio no fS7-RNas e contentwit h 50% ismuc hmor eeffectiv e thana comparabl ereductio no fS2-RNase . Therewa sn osee dse to nth establ eantisens eSI clone,becaus eth echromosom enumbe ro fthi splan t wasspontaneousl ydouble ddurin gtransformation .Thi swa sals otru efo rsom eo fth eS2 antisens eplants . The diploid transgenic clone PS2-S2AS-V-6, showing such astron gS2AS effect (Fig.2B ,3) ,ha d a disturbedfemal e fertility andse tonl ylimite d seedi nal lpollinatio ntypes .Th emal efertility , however, wasnearl yunaffected . Duet othi san dt oth epresenc eo fth epollen-expresse dself-compatibilizin gfactor , the aforementioned S2S2- S2AS clone 1184-01coul d be obtained, and this allowed testing of the heritabilityo fth eAS-effect . Theweake ro rles sstabl eexpressio no f(P3 5-5'2AS) 2i nthi sclon emigh t bedu et oa reductio ni ntransgen ecop ynumbe rfrom tw ot oon eb y meiotic-recombination. Inadditio n tothis ,o rinstea do fthis ,th ehomozygosit yo fth e S2-allelecoul daccoun t for aweake rAS-effec t than 54 Chapter3

inth eparenta l clone.Thi sdoubl econtributio n ina nendogenou sS-homozygou splan t is,accordin gt o MeCubbi ne ta l(1997) ,als oth emos tlikel yexplanatio nwh ythei rintroduce dmutate d.S-protei nwa s less effective inbreakin gdow nth estyla rpar t inS I(probabl yb ycompetition/multime r formation withth e homologous endogenousglycoprotein )tha nwhe npresen t ina n^-heterozygou splant .

Senseexpressio no fheterologou sconstructs . Thesense-expressio no fS2 drive nb yth e( Q ^SK^-promote ri si naccordanc ewit hth estron gan dtissue - specific expressiono fth ereporte rconstruct s(Ficke re tal. , 1998 a,i npress )an di sals osupporte db yth e resultso fMurfet t eta l(1994) ,usin gconstruct sdrive nb yth epromote ro fth eSK2 tomatohomologu e ChiP. 52-RNaseconten ti nstyle so f S-alleletransgenic sca nsurpas stha to fth eendogenou sS-allele s The studies by Ficker et al (1998 a,b)pointe d out that there might be a specific interaction between promoter, coding region and postulated distally located regulatory elements for a tissue specific expressiono fth e5-alleles .Th eSO-promote rwa si nthi srespec tdifferen tfrom th e^-promoter .Th e down-regulation ofendogenou s^-allele smight ,therefore , bea logica lresul to fover-expressio no f5 2 when driven by the SK2promoter . Theheterologou s construct mayb e less sensitive to some down- regulatingmechanism so rmigh teve nlac ksom eo fthos eregulator yelements .Th eSK2 promote rbein g thiseffectiv e isi naccordanc ewit hth esucces sreporte d forth ehomeologou spromote r ChiP. Matzke eta l(1989 )reporte depistati csuppressio nb yreversibl emethylation ,whic hmigh tb eth ecas eher etoo , becauseth eapparen tsuppressio no fS10 i nth etransforme d Van dV Iclone sdisappeare dan dreappeare d withth efluctuatio n ofth egreenhous etemperatures .Th esuppressio n of570 ,whic hwa sno tdetecte di n all52-expressin gtransformants , stressesth enee dt ocombin eH5 F withth eutilisatio no f5-homozygou s tester lines, for test crosses with S2S10pollinator s would have given conflicting results: high S2 expression andnevertheles scompatibilit y intes tcrosses . Fickere ta l(199 8a , inpress )reporte dtha t Gt/5-expressiono fth eheterologou s?SK2-GUS construct s could alsob efoun d inanthers ,an dthi swa si naccordanc ewit hth eobservation s(Murfet t etal. , 1994) onth eheterologou sconstruct sbase do nth eChiP promote r(Chi2;l encodes,a sstate d earlier, atomat o homologueo fSK2). Thisca nexplai nwh ysom eo fth etransformant s showeda lo wbu tdetectabl eleve l ofS2 i ntota lanthe rextract sdurin gth eperiod so fhig hstyla rexpression .I twa sno tdetermine d whether the52-RNas ewa spresen ti npolle nonl yo rtha ti tha da sporophyti corigin .Th epolle nfertilit y appeared not to be affected by the presence of 52-RNase, which is supported by observations of Kirch et al (1995). Iti sno wconfirme d forpotat ob ysense- ,antisense -an dover-expressio no fth e52-allel etha tth estyle - specific5-RNase sar eth eke yfacto ri nth e stylarcontributio nt oth egametophyti c self-incompatibility reaction.I tha s alsobecom eclea rtha t areliabl e gainan dlos so f specific SIreaction sb y amolecula r approach is not so evident, but manipulation of the stylar expression of SI in potato is shown tob e possiblenow . Chapter4 Expression andinheritanc e ofself-compatibilit y and self-incompatibility after crossing diploid S.tuberosum(SI )wit hS.verrucosum (SC )

Abstract

Diploid potato, Solanumtuberosum (tbr), andS.verrucosum (ver)ca nhybridis e when thelatte r species isuse d asth epistillat eparen tbu tno twhe nuse d asth estaminat eparent .Thi sphenomenon , calledunilatera l incompatibility (UI),i sfrequentl y observed when aself-compatibl e (SC) species is intercrossed with aself-incompatibl e (SI)one .S.verrucosum issuc h aself-compatibl e species. Verx tbrhybrid s display cytoplasmic male sterility (CMS) andare ,therefore , not suitable for geneticanalysi so fothe rcrossin gbarriers .Previously , specific diploid tbrclones , called acceptors, weredetecte dtha t showedbilatera lcompatibilit ywit hver instea d ofUI .Thes e selected cloneswer e used to realize the reciprocal crosses in order to circumvent CMSan dt o create, by repeated backcrossing, verwit h tbrcytoplasm . The resulting Fl's werebot h male andfemal e fertile. This "acceptance" for ver-polle n is based on the presence of a dominant gene A (acceptance), in combination with theabsenc e ofa n inhibitor geneI. The Fl's showed only expression ofth eS- allele that wasderive d from thetbr parent . Itwa sshow n that this verdoe s not produce style- specific ^-glycoproteins, ^-glycoproteins areresponsibl e for the stylar contribution to the self- incompatibility reaction in potato. TheFl-population s investigated here, were SC,bu t skewed segregation ratios for this trait, and disappearance andre-appearanc e of SC showed up inth e following offspring generations.Thes edeviation sfrom th e expected behaviour couldb e explained bypostulatin g amor ecomple xinteractio no fth eacceptanc e(o fve rpollen )determinin ggene sA an d

I,th einvolvemen t ofS Igovernin gS-allele sfro m tbr, astyla rnon-activ eS VCIallel e(SV ) anda weakl y

S-locus-linkedpollen-expresse d SCfacto r (SCver)from ver, resulting in"S Ib yU Ibase d inhibition of SC", is explained hereafter. The presence of the stylar non-active Svtr -allele allowed forth e penetration ofver-polle ni nstyle so fhybrid swhe nth erecipien t was ofth e genotypeA* iian d for anytbr polle ntha t didno t expressstyle-activ e tbr-S-alleles .Th e latterbehaviou r isnorma l in any gametophytic SI reaction. Pollen containing simultaneously an active tbr»S-allel e andth e SCVC, pollen factorwa sno teffectiv e incausin gS Cwhe nth erecipien twa so fth egenotyp eaal*. Itcaused , however, theexpecte d SCreactio n onan yothe r genotype,irrespectiv e ofth etbr ^-allele s activei n bothparents .Thus ,aal* non-accepto r genotypes,containin g SCmT, areS Ib yU Ibase d inhibitiono f SC. Tetraploidhybri d genotypes,obtaine dfrom doublin g anS Inon-accepto r diploid hybrid, produced

This chapter is submitted forpublicatio n ina slightl y modified version as:Ronal d Eijlander, Wendyte rLaak , JanG.Th . Hermsen, Munikote S. Ramanna and Evert Jacobsen. Occurrence of Self-Compatibility, Self-Incompatibility and Unilateral Incompatibility after crossingS.tuberosum (SI ) withS.verrucosum (SC): I Expression andinheritanc e ofSelf - Compatibility 56 Chapter 4 pollentha t showedmutua lweakening .Thi swa scause db yi ninteractio n of thepollen-activ e tbrS- allele and the, apparently pollen-active, Svtr allele from ver. Styles of non-acceptor recipients showed for thistyp eo fpolle n acomplet ebreak-dow n ofth e SIan dU Ireactio ni nth euppe r part of the style, whereas in the lower part inhibition reactions reappeared, thus indicating at least two different factors inUI ,whic hma y coincide with "a" and "/". Thus,i t was concluded that at least four different loci are involved inth e expression ofUI : the acceptor locusAla, Hi for inhibition, thecompatibilizin g factor SCm and thepolle npar t ofth eregula rS-locus .

Introduction.

Manyplan t specieshav edevelope d systems againstinbreeding .The yca nb ebased , for instance,o n morphological featuresb ywhic hself-pollinatio n isprevented , oro ndifference s inmaturatio n time ofpolle n and style.I n some speciesheteromorphi c characteristics are linked with incompatibility genescontrollin gpollen-pisti linteractio n(e.g .Primula vulgaris, Richards, 1986),bu thomomorphi c self-incompatibility ismor e common. Based onth etyp e of interaction between pollen andpistil , twomajo r self-incompatibility (SI)system sca nb e distinguished: sporophytic and gametophytic. Inth esporophyti c systemth eS Ireactio ni sdetermine db yth eexpressio no fth eparenta l sporophytic genotypes in thepolle n and inth epistil . Inth e gametophytic system iti sbase d on the expression ofth e individual allelespresen t inth epolle n (the gametophyte) itself and inth estyle . Interspecific and higher order crossing barriers exist innature . Quite often, related species canb e intercrossedreciprocally . Sometimes intercrossing is possible inon edirectio n only. Whenthi s is based onprocesse sbetwee npolle n landingan dfertilization , itca nb ecalle d unilateral incompatibi­ lity(UI) , althoughth eexpressio n "incongruity"i sals oofte n used instead ofincompatibility , when notreferrin g to SI(Hogenboom , 1973).U I ismos t frequently found incrosse sbetwee n SC and SI species. UI following the SI/SC rule can be found throughout the two major incompatibility systems.Fo r an extensive review on (unilateral) incompatibility mayb ereferre d to a monograph by de Nettancourt (1977), which is still informative. More recently, Mutschler and Liedl (1994) gave a good overview on interspecific crossing barriers in Lycopersicon,an d they favoured the opinion that SI andU I are discrete barriers, although they admit that theremigh t exist systems in which SI contributes somehow to UI. A more refined and extensive theoretical approach on unilateral incongruity by Trognitz and Schmiediche (1993) deals with all kinds of interactions, involving alimite d seto fnecessar y geneswit h alimite d number of essential alleles,bu t evenwit h this approach it remains difficult to draw conclusions about the correctness of Hogenbooms incongruity hypothesis (1973). Oneo fth edebate stha ti songoin g overth epas t 40years ,i swhethe r orno t theS-locu si s involved in theU I reaction. It has alwaysbee n tempting to assume such ageneti c relation, because UI and SIofte n have several aspectsi ncommon ,an da stron gcorrelatio n mayb e observed between UIo n Expressionof SC and SI in S.tbr-S.ve r offspring 57 the one hand and parental speciesbein g SI and SC on the other; this suggests to be more than a coincidence. Of course, many arguments have been countered by exceptions and erratic results (Mutschleran dLiedl , 1994).Nevertheless ,ther eha sbee na naccumulatio no fexperimenta l evidence that SI and UI are frequently associated (Lewis and Crowe, 1958; Chetelat and DeVerna, 1991; Hiscock andDickinson , 1993).Latel y Murfett et al(1996 ) carried out transformation experiments with S'-genebase d constructs inNicotiana an d demonstrated that 5'-glycoproteins(also called S- RNases, e.g., McClure et al., 1993) cancontribut e to at least apar t ofth eU I reaction. Oneo fth ereason swh yth edebat ei sstil longoing ,migh tb etha tdifferen t taxonsma ysho wU Iwit h different strengths andreactio npatterns ,base do ndifferen t numbers andtype s ofgene s and alleles. Theextrapolatio n ofresult s from one speciest o anotherrelativ e may,therefore , be inappropriate. Potato, Solanum tuberosum L.(tbr) isa self-compatibl e crop,bu tthi sca nb eattribute d toit sploid y level(2n=4x=48) .Mutua lweakenin gbetwee ndifferen t 5-allelesi nth esam epolle ngrai nmake stha t the speciesi sself-compatible . Potato dihaploids(2n=2x=24 )ar eusuall y SI,althoug h exceptions to thisrul ed ooccu r(Olsde r andHermsen , 1977;Hermsen , 1978a,b,c) . These dihaploids display UI whencrosse dwit hth e self-compatible speciesS. verrucosum (ver). This species isclosel y related to potato and evenbelong s to the same series (Tuberosa). The appearance of SCi n aputativ e SI potato plant does not necessarily cause a shift in the UI/SI relationship. The SC diploid potato clones G254 and B16 (Olsder and Hermsen, 1977) did not cause a break-down of UI when pollinated byver, andremaine d compatible asstaminat eparen twit h allothe rdiploi dpotat oclones . Hybrid offspring ofthes e two clones segregated, however, into acceptor and non-acceptor clones for ver pollen. Acceptance is the exception to UI, thus non-acceptance is equivalent to UI. The simplest andbes t fitting hypothesis isbase do ntw ostyla ractiv e genes,th e acceptor geneA and its inhibitor geneI. Bot h genes aredominant , butI isepistati c overA. So,onl yAM tbrgenotype s are acceptor (Hermsen et al, 1974). Analysis of other types of plants, including S.andigena based dihaploids and other accessions of ver,reveale d differential behaviour ofpollinato r and recipient (Hermsen etal , 1977;Hermse nan dSawicka , 1979),thu srefinin g themodel .Th epossibilit y ofth e involvement ofmor eallele sand/o rmor egene swa sdiscusse dhere ,an dgene-for-gen e relationships, pollen penetration capacity, differential reactions to foreign pollen and unilateral incongruity as proposed by Hogenboom (1973)wer e included,bu tth eresult s wereno t conclusive. Thereaso nwh y aspecie s is self-compatible canpla ya nimportan trol ei n answeringth equestio n which factors arereall y involved inth eU Ireactio nbetwee n SCan d SI species.Fro m vern o active ^-alleleproduct s are known (Kaufmann et al, 1992).The y claimed that aver x tbrhybri d (mal e sterile), expressing the 57 allele from tbr,wa s unexpectedly compatible with SI pollen from the latter species,thu sgivin gris et oSI SI homozygotes. Theypostulated , therefore, the existence ofa verborn e style-expressed suppressor ofth eSI function. Theexistenc e ofth eaccepto r linesa sdescribe db yHermse n et al (1977) andEijlande r eta l (1997) allowed for more detailed analysis of male and female fertile backcross progenies. Now the 58 Chapter 4 segregation of SCrelate d factors, S-alleles andU I could beteste d for their interrelationship. Here theresult s of thoseexperiment s arepresente d and itwil lb e discussed which factors play arol e in the SCbehaviou ro fve ran dho wthe y arerelate dt oth eUI , SIan d SCfactor s asreporte d for the tbr material used here.

Materials and Methods

All diploid potato clones used here, with the exception of clone 1024-02, are offspring from the three basic tbr clones G254, G609 and B16, and were used earlier as basic material for the investigation of the gametophytic Si-system and UI (Hermsen et al, 1974;Olsde r and Hermsen, 1976;Hermsen , 1978,a,b,c ; Eijlander etal , 1997).I nthi smateria l genotypes were found that were acceptor (AAiio rAaii) for verpollen , and non-acceptor clones were based on the absence of the acceptor geneA (thusaa**) o ro nth e expression ofth e dominant inhibitor gene/(thu s **ƒ*). Theoffsprin g clone 6233-11= S2S2 an dnon-accepto r (NA)o fth etyp e aaii, 6234-08 =S3S3 an d acceptor of the typeAAii, clone 6536-01 = S3S4 and acceptor of the type AAii and full sib clone 6536-02 =S3S4 and acceptor ofth e typeAaii. These clones and all other diploid tbrteste r clones ( 1140-02, 1140-05,6104-21 , 6221-17,6221-32 ,6221-30 , 6221-39, 6223-39, 6223-40) used here, havebee n assesses earlier for their SIreactio n by Eijlander et al (1997). Clone 1024-02 (Kuipers etal , 1994)ha sth egenotyp eS2S10, isnon-accepto r and,a smentione d before, not related to G254, G609o r B16.Th eclon ewa suse d aspistillat e parent in sometes tcrosses . Allver polle nparent swer efül l sibsfro m veraccessio n PI 195172,o roffsprin g ofthos e sibs.Thi s accession was chosenbecaus e among the ver accessions tested onvariou s acceptor tbrclones ,P I 195172ha d thehighes t pollen penetration capacity (Hermsen et al, 1974). Thebackcros s population VTV= {verP I 195172x G254) x verP I 195172, segregates for SI, S3 andSv, al lclone sbein g 100%cytoplasmi c male sterile.Th eF l VT(ver P I 195172 -27 x tbrS2S3) is also CMS and contains S2Svan dS3Sv genotypes . TV5= tbr (acceptor)-ve rhybrid ,backrosse d 4time swit hvariou sver accession s(mainl y anda tleas t the last time with verP I 195172) =97 %ver . It contains 5-alleles from tbr.Al l plants were male fertile andself-compatible . TV6=TV 5x ve rP I 195172.TV 5an dTV 6wer euse d inbackcrosse s with non-acceptor tbrgenotypes . The population 6484 =TV 6x tbr6233-1 1 (SvSv x S2S2,aalT) consisted of fourteen plants.These Fl hybrids arenearl ypur ever-tbr hybrids ,mal e and female fertile and non-acceptor (S2Sv, a*I*); the full sib clone 6251-19 = (TV5x tbr S2S3 aaii) x (TV5x tbr S2S3 aaii) =S3Sv, NA and SI; population 1173(1 1plants ) = tbr6234-0 8 x 6251-19 (S3S3,accepto r x S3*,NA). Thepopulatio n 6541(2 5plants )= tbr 6234-0 8 x verP I 195172-2 7 (=S3S3, accepto r x SvSv,ver) = Fl hybrid. Clones from population 6541 (all expected to be S3Sv, acceptor) were randomly Expressionof SC and SI inS.tbr-S.ve r offspring 59 picked and backcrossed with tbr 6536-01 or 6536-02 (both S3S4,acceptor) : 1051 (35 plants) = 6541-03x 6536-01; 1052(3 0plants) = 6541-03x 6536-02; 1053(6 5plants) = 6541-06 x 6536-01; 1054 (30plants) = 6541-06 x 6536-02 and 1055 (30plants) = 6541-11x 6536-01. Populations 1061 to 1066(u p to tenplant s each) were obtained by selfing ofrespectivel y the Fl clones 6541-01, -02,-03 , -06,-11 ,-25 . Pollinations weremanuall yperformed . Flowers,use d for controlledpollinations ,wer e emasculated prior toanthesis .Compatibilit ywa steste db yobservin gberr yan dsee d set after pollinating at least 5 flowers. In case of doubt pollen tube ingrowth was monitored by means of fluorescence microscopy asdescribe db yEijlande r eta l(1997) .Penetratio n capacity was classified based on the amount ofpolle n tubes capable ofpenetratin g upper and lower stylar sections.5-allele s and some otherstyl especifi c proteins (likeSKI andSK2; Kirc he ta l(1989)) ,wer edetermine db ymakin gtes t crosseswit hteste r lines(Eijlande r et al, 1997)and/o rb ymean so f Iso-ElectricFocusin g (IEF)o f stylar extractso npoly-acry l amideo ragaros egels .PAG Ewa sperforme d asdescribe d byKirc h et al (1989) or by means of pre-cast agarose gels (Hypure gel VG 1020,Isola b inc.) following the silverstainin gprocedur ebase do nTungstosilicili c acidi nstea d ofsodiu mpermanganate , according to company specifications.

Results.

Expression studies ontbr ^-allele s in vercytoplasm .

Kaufmann et al (1992) detected in a ver x tbr hybrid {SISv) that some offspring plants after backcrossing withSI S3 genotype s didno t show the expected S3allele ,bu t only theSI allele;th e postulated ver borne allele Sv does not produce a detectable basic glycoprotein. Two likely explanations for thistyp eo fprogen y areparthenogenesi s (Abdalla, 1970)o r suppression ofth etbr «S-allele(SI) in the hybrid. Presence of such phenomena could impede research on SC/SI, non- acceptance (NA, here equivalent to UI)an d thepossibl e relation between UI and SIi n ourhybri d system. Therefore, the (verx tbr)x verBC 1 (TVT,i nS-alleles :SvSv x SlS3)x SvSv)an d thever x tbrF l hybrid

Table 1. Incompatibility reaction of VT Fl and VTV BC1plant s inrespons e toSI, S2 and S3 pollen. VT= verP I 195172 -27 x tbr S2S3 ;VT V = (ver PI 195172x tbr SIS3 )x ver PI 195172.S-allele s were detected by IEF followed by silver staining. Sv = postulated S-allele from ver, noban d visible;S1/S2/S3: S-alleles derived from the tbrclones . Pollinations were made with SIS1, S2S2 and S3S3 homozygotes (Eijlander et al, 1997). Berry set isindicate d by + or -,based on 5t o 10pollination s ;blan k = + expected, but not tested. Average seed set in case of berry formation ranged from 31 to 107.

Clone S-alleles Berry set or polle tube ingrowth after pollination with:

SI -pollen <52-pollen •Si-pollen

VT-4 S3 + -

VT-6 S2 - +

VT-7 S2 - +

VTV4-1 Sv +

VTV4-5 SI - +

VTV5-8 S3 + -

VTV6-2 Sv + +

VTV7-1 S3 + -

VTV7-4 Sv + +

Fig 1. IEF pattern of stylar extracts after silver staining of a population of nearly-ver TV6 (left panel) and lit ver. Two bands of basic proteins are visible: Kl (upper) and K2 (lower). Segregation of the bands of TV6 and ver can be explained by assuming thatTV 5=K2K2 was crossed with ver =K1K2o r vice versa and that the ver population tested here, oroginated from the cross verl x ver2 =KIKl x K1K2. Expression ofSC and SI inS.tbr-S.ve roffspring 61

Segregation andallelis mo fstyl especifi c proteins in "S.verrucosum"an dit shybri d offspring.

For ver 15 full sib clones and for TV6 eight clones were tested for SC and the presence of stylar proteins.Al l2 3genotype swere ,a sexpected , SC.Bot hpopulation s segregated fortw oprotein s(Fig . 1),abundantl y present inth emos tbasi c (pH>7)par t ofth e IEFge l (silver stained).Thos e proteins wereno tvisibl ei n extractso flea fan dste mtissue .Becaus eo fthei r localisation andthei r presumed non-S-allelenature ,the ywer e designatedKl andK2, analogous to the style specific non-S-linked tbrpolypeptide sSKI (presumably anRNase : Leee t al, 1992;Thompso n etal , 1995)an dSK2 (a n endochitinase: Wemmer et al, 1994).Kl focused approximately at the SKI place, but could be identified by amor eredis h colouring after silver staining.K2 focused evenmor ebasi c than SK2. Basedo nobservation s asshow ni nth efigure s 1an d2 ,th erankin g from acid tobasi c ofth e stylar proteins appears tob e as follows: S4 (1stband) ,,S7(no tshow nhere) ,S4 (mainband) , S2, S3, Kl, SKI, SK2, K2. The ver (selfing) population segregated in 7 Kl : 8 (K1+K2) plants and TV6 segregated into 4 (K1+K2) :4 K2 plants (Fig. 1). Segregation analysis of offspring plants from crossesbase do nK-band s( Kl®; K2®\ (K1+K2)® ;K1 xKl; Klx K2;(K1+K2) x K2 an d(K1+K2) x(SK1+SK2)) showedbot hnorma l andskewe d segregations,bu tn ogenotype swer efoun d missing simultaneously bothKl andK2. This strongly indicates allelism ofKl andK2. Ifthi s assumption iscorrect , then theS-allele so fve r (Sv) produce apparently littleo rn obasi c5-glycoproteins .

00 00 as O •* t-~ 00 OS o O o o o o —' V) in i/-> in «n »o io O o O o O O IT) o o o CS o o O O O t "•* #> » 4 4* , rIT m •1 W * m

S4

-SK2

Fig.2 IEFpatter n of stylar extracts after silver staining of (left panel) tbr parent 6234-08 (1),ver paren t PI 195172-27 (m), theF l clone 6541-06 (r)an d (rightpanel ) some offspring BC1 clones ofpopulatio n 1053,derive db y backcrossing clone 6541-06 with an S3S4 tbr clone. Visible bands:S4 (upper two),S3, Kl and SK2. SKI is faintly visible in some lanes between Kl and SK2. 62 Chapter 4

TheF l population 6541 camefrom th e cross tbr6234-0 8 x verP I 195172-27(=S3S3, acceptorx SvSv).Th ever clon euse dhere ,wa spreviousl y tested for absence oftbr- 5-allelesuppressio n (see above). Twelve 6541-clones were tested for SC and IEF-band composition. All plants were, as expected, SCan d allwer e showing theS3 an dSK2 ban d from the tbrparen t and also theKl band from the verpolle n parent (Fig. 2).Seve n randomly selected plants were tested for acceptance of verpolle n and proved tob e acceptors.Al l seven were aspolle nparen t compatible on theparenta l acceptor clone 6234-08 (S3S3). Functional activity of the S3 allele in 6541 plants was directly confirmed by fluorescence microscopy performed on styles of test crosses, showing a proper incompatibility reaction against S3 pollen (Fig.3). Thus, SI against S3 pollen was still active, acceptance was maintained and SCcharacteristic s of verwer e expressed inth e 6541 population.

TheF l 6541 clonestha twer eteste dfo r acceptance (all S3Svan d acceptor)wer ebackcrosse d with the S3S4 acceptor lines 6536-01 and 6536-02 as pollen parents. Five BC1 populations (1051 to 1055), with a total of 143 flowering plants, were investigated in more detail. All these clones showedth e54-allel eo fth eS3S4 pollinato rclone s(Fig .2) ,thu sexcludin gsellin go rfertilisatio n by anS3 polle n grain and confirming that theS3 allel ewa s fully functional and effective inth e style and not suppressed by some unknown ver factor. If there had been a stylar Si-suppressor, the penetration ratio (certation)betwee nS4 an dS3 woul dhav ebee n at least 98: 2 (at y =0,05) , sothi s possibility canb erule d out. Itwa s postulated that vercontaine d an5-locu sbu t that its stylar expression could not be detected throughIE Fbecaus ei tlacke da n5-specifi c band. Therati o "presence ofS3" to "absence ofth eS3 (=presenc eo f"Sv") "wa spoole d overfiv epopulations ,an dsegregate d into7 0S3 :7 3S V ( fits1: 1; 2 2 X = 0,06). The Kl band segregated in arati o of 65 present and 78 absent (X = 0,59; P):1=0.8), which is also reflected inth e ratio ofK1+S3 : Kl+Sv : S3 : Sv (32:33:38:40),whic h fits a 1:1:1:1 segregation (X2= 1,04; ?,.,.,.,= 0.8), thusprovin g the independent assortment of theS-allele san d Kl (fig.2). That K2 (presumably allelic with Kl) was not allelic with the tbr ^-alleles, was confirmed by banding patterns of clone 6251-22, a sibling from clone 6251-19, that had the genotypeS2S3 K2(fig . 4).So ,S3S4K1 andS2S3K2 plant swer efound , indicating, asexpected , the non-allelism of eitherKl orK2 wit h the tbr5-alleles . Thesi xpopulation s 1061t o 1066(654 1clone s selfed) allowedfo r 30plant s aprope r identification of Kl and SK2 bands. Because these genes proved to be non-allelic with S-alleles, segregation analysis ofth e^-allele swa sno t consideredhere . Sevenplant s showed onlyth eKl band, 13bot h the Kl andSK2 band s and 10showe donl yth eSK2 band .Whe nKl andSK2 woul db eallelic ,the n theF l population (6541)wa smos t likelyderive dfrom th ecros sSK2SK2 xK1K1, resultin g in654 1 withth eSK2K1 genotype.I nth ecas etha tSK2 andKl areno t allelicbu t independent genes,th e Fl would then have been hemizygote: SK2—Kl—, selfing would give different segregation ratios. Testing for these ratios gave a x2o f 1,13 for allelism (P=0.57) and a% 2 of 6,50 for the hypothesis ofunlinke d genes(P=0.09). HenceKl andK2 aremos t likely allelic withSK2. Expression ofSC and SI inS.tbr-S.ve r offspring 63

—S2 —S3 —S4 —SKI —SK2 —K2

Fig.3 (left) Incompatibility for S3-pollen as monitored in styles from the S3Sv clones from the Fl population 6541 after fluorescence microscopy on anilin-blue stained styles. Top: stigma; bottom: ovary with ovules. Fig.4 (top) IEFpatter n of stylar extracts after silver staining of tbr-ver offspring clones. From top to bottom: S2, S3, Kl, SKI, SK2 and K2. Most left lane: 6251-19, expresses S3Sv SK1K2. Most right lane: 6251-22 , expresses S2, S3, SKI, SK2 and K2.

Self-compatibility basedo n(differential ) acceptanceo fSv an da pollen-expressed factor,SC vt

The populations 1052 and 1054 (= 6541-03 & 6541-06 x 6536-02 = S3Sv, acceptor x S3S4, acceptor) segregated into acceptor lines (AAiian dAaii) an dnon-accepto r lines (aaii)whe n tested with the hereafter mentioned tester clones,i n apoole d segregation ratio of 32acceptor s : 14non - acceptors( = 3:1; X2 =0,72) .Th epopulation s 1051,1053an d 1055( =6541-03 , 6541-06 &6541-1 1 x 6536-0K S3S4,acceptor> ) showed onlyaccepto r genotypes (AAii orAaii) whe n pollinated with the tester clones verP I 195172(clone22 xclone27)-31&-32 ,bu t showed in about aquarte r of the casesa nU I(thu snon-acceptor )reactio nwhe nteste db yanothe rver genotype ,(clon e3 5x clon e37) - 1 (data not shown). No clones were found that were acceptor for the latter ver clone and non- acceptor for the former two verclones . Itwa s clear from these observations that inthi smateria l a differential non-acceptance reaction against certain types of verpolle n was found, as previously reported for this system (Hermsene t al., 1977;Hermse nan d Sawicka, 1979). 64 Chapter 4

Table 2. Segregation ofS-allele s and SC in the acceptor populations 1051 and 1053. The populations are BCl's ofth e type TVT: 6541-06 x 6536-01 & 6536-02; 6541-06 = tbr 6234-08 x ver. All are acceptor for verpolle n of tester lines 31 and3 2(F l of PI 195172).Number s are pooled. S-allelesdetecte d by IEF:S3 and S4 were derived from the tbrclone s 6234-08 and 6536-01; Sv =postulate d stylar non-active S-allelefro m ver","detected"b y absence

ofth e S3 tbr allele. SI= self-incompatible, SC = self-compatible; 5Cver -= single SC factor from ver,- - = n o 5Cver

factor. The class S4Sv, SC contains S4Sv plants with and without theSC m factor. •S-alleles S3S4 S4Sv

compatibility SI SC SI SC

number found 33 9 2* 33

postulated genotypes S3S4, - - S3S4,SC„- S4Sv, -- S4Sv,--/SCm- * showing differential reaction againstver. Self-compatibility waspredominantl y found inth eS4Sv genotype s (table 2). The SCo fth eS4Sv genotypesca nsimpl yb eexplaine db yacceptanc efo rth eSv sel fpolle ntype .Th etw oS IS4Sv clone s showed adifferentia l reaction againstpolle n ofth ever teste r clones,bu tth eres t ofth eS4Sv clone s with a differential reaction were still SC. Here aquarte r of the S4Svplant s appeared to be of the differential type. This SC/non-acceptor discrepancy was also detected in the 1052 and 1054 populations.Becaus enin eS3S4 plant s were SC,no t attributable toSv (Sv i salleli c with the tbrS- alleles,thu sno tpresen t indiploi dS3S4 genotypes),a n additional pollen expressed SC-factor must bepostulated , codedSC va. The skewed segregations inth eS3S4 an dS4Sv subgroup s indicatetha t

SCveri slikel yt ob elinke dwit hth eS-locus .I nth emothe rclon e6541-0 6 iti si ncouplin gphas ewit h SVan dthu si nrepulsio nphas ewit hS3. Fo rth eS3S4 par to fth epopulatio nther ewa sa recombinan t fraction of9/42 =0.21 4 andfo r theS4Sv subpopulation itwa s (2/35)/ (1/4)= 0.229 , so on average there appeared tob e 22%recombinatio n betweenSv andth eSC vtr factor, with aconfidenc e interval of 13%- 33 % (y =0.05) .S4Sv plant scan ,therefore , beS Cdu et o acceptance for Svpollen , and to thepresenc e ofth e^-linke d SCverfactor ; S3S4plant s can only be SCbeaus e of SCvtr.

From selfing theBCl-clon e 1053-27(= S3S4, SCveT—, acceptor),a clon ewit hth egenotyp e "S4S4,

ScveT*,acceptor "wa s identified and coded 1144-02.Selfin g of 1144-02 resulted, as expected for pollen-borne SC,i nS Cplant sonly ,al lbein g acceptor. Clone 1144-02wa sa smal ecompatibl ewit h clones 6221-17 (S1S4, NA) and 6223-40 (S3S4, NA),bu t not with 6221-32 (S3S4,NA ) 6221-37 (S3S4, NA)an d 6223-39 (S2S4, NA).Thi sindicate stha t adifferentia l reaction not only against Sv exists,bu t against theSC V„ factor aswell .

Self-incompatibility by UI-base d SC inhibition: expression ofth e pollen-SCver factor. 5 6 Thepopulation sTV an dTV (bot hSvSv an dexpecte dt ocontai nSC ver)ar ever wit htbr cytoplasm , yielding male and female fertile hybrid populationswhe n crossed with non-acceptor tbr asmal e parent. Twonon-accepto r based hybrid populations were investigated more closely,becaus e the Expression ofSC and SI inS.tbr-S.ve roffspring 65

Table3 .Reactio npattern s ofpopulatio n 1173(S3S3, SI,accepto r xS3Sv, SI,non-acceptor ) incrosses ,an d deduced genotype composition. The staminate parent of 1173 was known tob e S3Sv aal*. Elevenplant swer eselfe d (SC=self-compatible , SI= self-incompatible) , allwer eSC .Te nwer eteste d for acceptance based onberr y setand/o rpolle ntub epenetratio n (A=acceptor ,NA =non-acceptor) .Th e genotypesmentione d below (exceptclone s6 , 7 and 11)wer e additionally tested for pollen tubepenetratio n instyle so f clone 1024-02 (S3S10, NA). Pollen tube arrest isindicate d infractions compatibl e (C ) andincompatibl e (I ) pollen . S3 andS\: S-alleles;

SCy„= SC-factor from ver, - = no SC-factor; A =accepto r allele;I = inhibitor allele. Blank = not determined.

Plant nr. SI/SC Acceptor/non-acceptor Pollentub e arresti nS3S10, NA Deduced genotype

1173-01 SC NA 1/2 C 1/21 S3S3, SCm-, Aali

1173-03 SC NA 1/4 C 3/41 S3Sv, SCm-, Aali

1173-04 SC A I S3Sv , - -, Aaii

1173-05 SC A 1/2 C 1/21 S3S3, SCm-, Aaii 1173-06&11 SC NA S3S*, SC^.-, Aali

1173-08 SC A 1/2 C 1/2 I S3S3, SCm-, Aaii

1173-09 SC NA 1/4 C 3/4 I S3Sv, SCm-, Aaii

1173-10 SC A 1/4 C 3/41 S3Sv, SCm-, Aaii

expression of SCi nth ehybrid sdeviate d from expectation.Th efirst investigate dpopulation , coded 1173, was derived from a cross in which an acceptor tbr and anon-accpto r Fl clone (TV5 xtbr, S2S3, NA)wer einvolved .Th esecon dpopulation ,code d6484 ,wa sa nF lfrom th ecros sTV 6x non- acceptor S2S2 tbrclon e6233-12 . Thepopulatio n 1173 (=6234-0 8x 625 1-1 9= S3S3 (SI) ,accepto rx S3Sv (SI) ,non-acceptor ) showed the S3 band by IEF, for the pistillate parent was S3S3. Genotypes like S3Sv and S3S3 can be distinguished by test crossing but not by IEF.Therefore , the 1173 -plants were selfed, tested for acceptance of verpolle n and test crossed onth eS3S10 non-accepto r 1024-2 (table 3).Al l eleven obtained clones were self-compatible (Fig. 5a) and all of them, except 1173-02, were tested for acceptance. Five were non-acceptor, the other five were acceptor. The pollen tube penetration fractions (in)compatible pollen and site of pollen tube arrest) of three acceptors and three non- acceptors was investigated intes t crosses onclon e 1024-02 (S3S10, non-acceptor).55-polle nca n beinhibite db y SI,an dSv polle nb yUI .Th egenotypi cconstitutio no fth e 1173clone swa sdeduce d from combiningpolle ntub epenetratio ndat awit hacceptor -an dS Ibehaviour .Re-appearanc e ofS C innon-accepto r 1173-plantsindicate d thatparenta lclon e6251-1 9containe d thepolle n factor 5Cver and confirmed the expectationtha tit snon-accepto rbackgroun d wasaali. Clone 1173-04,whic h was self-compatible and acceptor, was fully incompatible on clone 1024-2. This proved, as expected, theheterozygosit y ofSC veri nclon e6251-1 9 aswel l asit sheterozygosit y for the5-locus . 66 Chapter 4

Fig.5 Pollen tubegrowt hmonitore d byfluorescence microscop y inanilin-blu estaine d styles ofnon-accepto r tbr-verhybrids .

Fig. 5a (left): Self-compatibility based on SCver pollen as monitored in styles from the non-acceptor 1173-08 (S3Sv SCver). Fig5 b(middle ) andFi g 5c(right) :S Ian dU Ii n styles from thenon-accepto r Fl clone 6484-06 asfunctio n ofth eploid y level of the pollen. Middle: incompatible reaction for S2 pollen after test crossing with an S2S2 pollinator. Upper arrow: approximate inhibition sitewhe n pollinated with vertyp e pollen. Lower arrow:approximat e inhibition site when pollinated with tbr S2 pollen. Selfing results in a mix ofthos e reactions. Right: Pollen tube growth after selfing of the tetraploidised clone 6484-06.S2Sv polle n isdeepl y penetrating the style. Other pollen tube types are earlier arrested.

Thesecon dpopulatio n investigated for aputativ esuppressio n of SC was theF l population 6484: TV6 x tbr 6233-11 (=SVS V x S2S2,aall). All fourteen 6484-clones (S2Sv,a*I* )wer e SI,an d as expected, incompatible for S2-an dnon-accepto r (NA)fo r verpolle n (Fig.6b) .A t least 150 flowers ofeac hplan twer esel fpollinated ,260 0pollination s intotal .Twelv eseed swer eobtaine db y end-of- season-compatibility, givingrise t o sixwea k andpoorl y flowering plantswit h areduce d fertility. All six seedlings showed the S2-glycoprotein band and were, based on a limited number of pollinations, self-incompatible andnon-acceptor . NoSvSv plant s were found, which has under the assumption that S2an d Svpolle ntube s are equally arrested, alikelihoo d of at least 17%. Pollinationswit hclon e 6484-06 (S2Sv, non-acceptor) onth eclone s6233-1 1(S2S2, NA) and 6223- 39 (S2S4,NA )wer e incompatible, but compatible on 6104-21 (S1S2, NA).Thi s observation ofa Expressionof SC and SI inS.tbr-S.ve roffspring 67

Fig. 6. Detail of fluorescence microscopy images of anilin blue stained pollinated styles. Reaction of pollen tubes in various crosses. Left panel (a-d):ruptur e oppolle n tubetip si nnorma lU Ireaction s and(e ) most common type of pollen tube tips of SI inhibited pollen tubes.Righ t panel:Deteail s ofmutua lweakenin g basedpseudo-compatibl e pollen tubes with the presumed genotype S2Sv in S2S2SvSv styles, as monitored in styles like shown in Fig 5c.Pane l fl: directly under the stigma; fl: halfway the style; f3:a t 3/5 of the style; f4: pollen tube arrest in the lower part of the style. differential reaction is highly similar to the observed crossing results with the clones 6541-06

(S3Sv, Aaii, SCV„ -) and 1144-02 {S4S4, A*ii, SCver* ) (see above). When clone 6104-21 was pollinated with 6484-06,a nestimate d 10-20 %o fth epolle n tubesmonitore d by UV-fluorescence microscopy was ofth e compatibletype .

Mutual weakening by combining S2 andSv inpolle n from atetraploi d NAtbr-ver hybri d Non-acceptor clone 6484-06 (S2Sv a*I*) was somatically doubled by tissue culture (data not shown). Twentytetraploidise d cloneswer eobtaine d andteste d for acceptance and incompatibility againstS2 polle n intes t crosses.Al ltetraploi dplant sbehave d asth eorigina l diploid clone6484-06 , they were incompatible for S2 pollen (from clones 6233-11, 1140-02 and 1140-05) and non- acceptor for ver (Fig. 5 b, c). One clone was male sterile, but the remaining nineteen of the tetraploidized cloneswer esufficientl y male fertile and allowed for selfing. These tetraploid clones showed alo wleve lo fSC ,whic hwa sstrictl y absenti nbot hth eorigina l 6484-06 clone andth enon - doubled tissuecultur ederive d controlplants .Fro m thesetetraploids ,berrie s with few seedswer e obtained in a much higher frequency than after selfing of the original diploids. Fluorescence microscopy onpollinate d styles showedtha t about 5-10 %o fth epolle ntube swer eo fa remarkabl y different type. In theuppe r 1/3 ofth e stylethe ywer e of acompatibl e type,wit h long,thi n tubes, withregularl yinterspace d smallcallos eplug s(Fig .6 :fl , f2),a susuall y found innorma lcompatibl e crosses. At about halfway the style the tubes became broader and irregular, sometimes even branched,wit hmor ebu tirregula rcallos edeposition , andfinally th epolle ntube swer earrested ,wit h much inflated pollen tubes (Fig.6 ß), differing from the normal incompatibility or unilateral 68 Chapter 4

incompatibility typei nthi smaterial . Theincompatibilit y reaction was completed at about theen d of the style, usually showing swollen pollen tube tips instead of ruptured ones (Fig. 6, f4 versus a,b,c,d). Doublingo fth egenom eallowe d for atyp eo fgen ecombinatio ntha ti sno t attainablei npolle n from adiploid .Heterozygosit y is,therefore , themos t likelyexplanation , andmutua lweakenin g between Svan dS2 ca nb ehel dresponsibl e forthis .Th eweakenin geffec t appearst ob ebarel y strong enough to bypass the UI reaction, though. So, the pollen part of the S-locus of ver appears to be still functional, not onlyi ntriggerin g anU I reaction,bu t also in causing the SIrelate d phenomenon of mutualweakenin g whentogethe rwit h atbr 5-allele .

Discussion

StylarS-allele suppression. Thebackcros s experimentswit hth eVT Van dV Tclone sindicate d that thesever accession sneithe r expressa cytoplasmic ,styl especifi c incompatibility suppressing factor norgeni esuppressin g factors,a spostulate db yKaufman n eta l(1992 )i nthei rmaterial .I ti spossibl e thatthei rmateria l contained afacto r sucha sreporte d forPetunia (Flaschenrie man dAscher , 1979; Danaan dAscher , 1986,b) ,whic hmigh t cause,fo r instance, (pseudo-)compatibility, although the expression ofthi s factor inPetunia appeare d to act only incis. Parthenogenesis , asrepore d tob e present in some ver lines (Abdalla, 1970), is another explanation for Kaufman's compatibility. Furthermore, the normal incompatibility reaction is not always reliable. Besides, modifiers are commonly found andearl ypollinatio nca nbypas sa notherwis efull y functional SIsyste m(Ei jlande r et al, 1997).Non eo fth e afore mentioned factors appearedt ooperat ei nou rmateria l and, therefore, theSI, S2 an dth eS3 allele swer e fully functional in our system.I nth ereciproca l type ofmaterial , TVT(e.g .th e 1053population) ,th epenetratio n ofS4 c.q . the arrest ofS3 wa s fully in accordance withthes e findings.

Stylarproteins andallelism. Althoug h Kl andK2 are accepted tob e alleles from the samegene , therei sstil la smal lchanc e(o fabou t0. 2 %)tha tthe yar enot ,becaus eth eanalysi so fth esegregatio n behaviourha dt ob ebase do nrelativel y smallpopulations .Th epresenc e ofKl orK2 togethe r with S2an dS3 (6251-22 )o rS3 an dS4 (e.g . 1053-27)i na diploi d plant proved that they areno t located on the same locus asth e5-alleles . SCca nb e found inplant s that lackbot hKl andK2 and plants withtw odifferen t S-allelesan dKl orK2 canb e SI.Thus ,mutua l weakening caused byKl orK2 is not an explanation for SC. Because of this and because of the relatedness of tbr and ver, translocation or duplication of S-alleles during the evolution of ver is therefore not a likely hypothesis to explain SC. Segregation ratioso fKl andSK2 i nF 2progenie s showed that allelism isver ylikely .I ti slikel ytha tKl andK2 arelocate da t(approximately ) thesam elocu sa sSK2 i ntbr Expression ofSC and SI inS.tbr-S.ve roffspring 69

andtha tthe y areprobabl y allelic.Allelis mo fKl andK2 wit hSK2 implie stha tthe y should belong toth e family ofendochitinase s (Wemmere t al, 1994).I t isunclea rwhethe r theypla y arol e inUI , because theKl-, K2-,SK2- and also theS£7-ban d couldno tb e linked toUI .

Acceptanceand differential reactions .Th e acceptance ofve rpolle nb y tbrha sbee n reconsidered severaltime s andha sbee npu t in abroade rperspectiv eb y comparing itwit hU I systems between other species aswel l (e.g . Abdalla, 1970;Abdall a andHermsen , 1972;Abdalla , 1974;Hermsen , 1977; Hermsen et al, 1977;Hermsen , 1979;Hermse n and Sawicka, 1979). Differential acceptor seriesan dpenetratio ncapacit ylevel spasse di nrevie wan dthi sseemed ,o nth efac eo fit ,t o conflict with apreviou s model ofHermse n eta l (1974),wher eonl yth eAalli system (dominant inhibitor epistatic over dominant acceptor gene) was discussed as an alternative for the postulated AtA2 system (only doublerécessives ,a,a ;a2a2, areacceptors) ;penetratio n capacity did not play arol ei n this analysis. Extending the Aali model by introducing more alleles and various dominance realtionships can bring all results in accordance with each other. Hermsen (1974) used pollen mixtures of S.ver PI 195172, and the other articles dealt with various accessions and separate genotypes of tbr and ver.I n our study differences inpenetratio n capacity between verP I 195172 based siblingswer edetecte dtoo . Wedi dno tus epolle n mixtures,bu t individual ver siblings were used fortestin gth e segregation ofacceptanc e andnon-acceptanc e inhybri dpopulations . When all individualteste rgenotype sfailed , eventhos ewit ha hig hpenetratio ncapacity ,i twa sconclude dtha t ateste d plant was non-acceptor. Basedo nthi s approach, fully expressednon-acceptanc e for allver teste rclone swa s detected inth e 1052an d 1054populations .Th e 1051an d 1053 populations wereentirel y acceptor for at leaston e ofth ever teste rclones .Th einbre dclon e6234-0 8wa sAA ( dat ano t shown)and ,consequently , ver must havebee nA'a {A' gives an differential acceptor reaction when compared with A; amus t be present because of the segregation of non-acceptors inpopulation s 1052 and 1054) and 6541-06 must havebee nAA '. Backcrosses with 6536-01 (S3S4AAii) resulte d in differentials, so 6536-01 must havecontaine dtw odifferen t co-dominant^-alleles , eg.AA'oxAA ".

Theself-compatibilizingfactors Svand SC ver. Therear evariou sway sb ywhic hplant sca nbecom e self-compatible. For instance, a gene can become silenced, but mutation of the coding region is another possibility which has been described for many crops. For Lycopersicon peruvianum, a diploid SI tomato species, it hasbee n reported that a singlepoin t mutation caused loss of the S- RNase activity, resulting in full SC (e.g. Kowyama et al, 1994), but a basic protein remained detectable. A frame shift, as reported for L.esculentum (Thompson et al, 1995), truncated the putative5-glycoprotein ,resultin g inanothe r IEFpoin t and loss ofit sRNa.se activity, thus causing SCo ftomato . Thepresenc eo fth enot-stylar-activ eSv allel ei na diploi d hybridtbr-ver plan tmean stha tther eca n 70 Chapter4 beonl y one activeS-allele .So ,al lS-heterozygou stbr clone swil lb ecompatibl e on such ahybrid . Ahybri d tbr-verplan twil l alsob eSC ,a slon ga sth eplan t isaccepto r (A*ii) for verpollen . So,th e effect ofth eSv allel ei nth estyl edepend so nth epresenc eo rabsenc eo fth eacceptanc e determining genesA andI.

Thepolle nactiv eSC-facto r SCyatha tha sbee nfound , appearedt ob eweakl ylinke dwit hth eS-locu s (chromosome 1). This is in agreement with the location of apolle n active SC factor inPetunia hybrida (calledpolle ninactivator) ,tha twa sreporte d tob ea ta distanc eo fabou t 20t o2 8map-unit s fromth eS-locu s(wherea sals oa styla rfacto rwa sreporte d atapproximatel yth esam edistance :Dan a et al, 1986 a,b; Flaschenriem et al, 1979). Therefore, this factor is clearly different from the SC factor reported for thediploi d tbrclone s G254 andB16 ,whic hwa s localised on chromosome 12 and not linked with the 5-locus on chromosome 1 (Hermsen et al, 1973, 1978). A remarkable characteristic of the 5Cver-factor is that its effectiveness of causing a mutual weakening effect depends on the genotype of the recipient. All results obtained so far can be explained by the assumption that SCVCI isno t effective in overcoming theU Ibarrier s caused by the genotypes aall and aali, but that other non-acceptor genotypes like aaii and A*I* allow for a 5Cver induced compatibilityi na combinatio ntha twoul db eincompatibl ewhe nonl y«S-allele swer eregarded .Thus , genotypes like 6484-6 (S2Sv, aa,li) and 6251-19(S3Sv, aa,Ii) can contain a self-compatibilizing factor (SCmT) but areS Ib y anUl-base dinhibitio n of SC.S Cca n show up inoffsprin g populations, but onlywhe n non-aal*genotype s segregate.

Recognition ofthe pollen part ofthe ver-S-locus. Thetacitl y acceptedassumptio ntha tpolle n ofth e vertyp ei sarreste db ynon-acceptors ,implie stha tth epolle npar to fth e5-locu sof ver i sstil lcapabl e ofcausin ga nU Ireactio nwhe npenetratin g anon-accepto rstyl e(se eals obelow) .Thi s is confirmed bysegregatio nratio si na populatio nbase do na non-acceptor(S2S4) x 625 1-1 9(S3Sv) cross,wher e only the S3 allele penetrated (data not shown). It is also in agreement with the pollen tube penetration types asreporte d for 1024-02x 1173,wher e alsoth eSC nr factor played arole . When theiSC^-facto r makesS3 polle no n6234-0 8 (S3S3, AA,ii )a scompatibl e asth eSv-polle ntype ,th e pollentha t ledt opopulatio n 1173 willhav eha d thecompositio n of 1 S3SC vctal:\ S3SC vcrai : 1

Sv ai : \ Sv al :1 Sv SCva ai : 1 Sv SCya ai. Not all possible genotypes for this cross have been found, forS3SvAaIi wa sno tdetecte d(se etabl e3 )an dthi si sth eonl ypossibl eS Igenotyp eresultin g from the cross S3S3AAii -- x S3Svaali 5Cver -.Thi s ispresumabl y due toth e limited population size,becaus e the segregation found was inagreemen t with the expected one.

Dualfunction ofthe pollen part ofthe S-locus: polyploidy-effects. Amos t striking result on theSv allele was obtained after doubling of genotype 6484-06 (S2Sv; NA).Althoug h it was known that thestyla rpar t ofth ever 5-allel ewa sno t active,a dua l function ofth epolle npar t (contributing to both SIan dUI )coul d stillb epresent . Somatic doubling of an^-heterozygou sclon ewil l normally lead to SC, because of heterozygosity of 50 % of the pollen for the ^-alleles involved. Mutual Expression ofSC and SI inS.tbr-S.ve roffspring 71 weakening, an SI related phenomenon, can then be active. It was known from the diploid that normal recombination does not result in self-compatibility ofS2 or Svpollen , irrespective of the presenceo r absenceo fSC waan dwil lno tpla y arol ei na somaticall y doubledplant .SvSv polle n will be inhibited by UI. Pollen being homozygous S2S2 will be blocked by SI and competitive interaction does not play any role here (Eijlander et al, 1997). The more compatible polle tubes respond inth euppe rpar t ofth e style likeS2S3 polle ntube s in anS2S3 o rS2S2S3S3 style.Mutua l weakening between Svan d S2i nS2Sv polle n offers thebes t explanation for this. Theobservatio n thatSv i scapabl eo fraisin g aU Ireaction , theafore n mentioned polyploidy based SC and the dual function hypothesis ofLewi s and Crowe (1958)justif y the assumption that Svi s alsocapabl eo fcausin gmutua lweakening .Thus ,th eS-locu so fve rcontain sa pollen-par ttha ti sstil l active and co-dominant in SI.Furthermore ,th e fact that the combination ofSv + S2 in the pollen (thusbein g^-heterozygous )i s alsobreakin g downth eU Ireactio n in the upper part of thestyle ,i s a strong indication, if not proof, that the S-locus is involved in the UI reaction. However, in the lower part of the style a second reaction typebecam e visible, that has remained unnoticed in the diploid situation. Thisha spresumabl ybee nmaske do rprevente db yth emuc h stronger UI reaction inth euppe rpar t ofth estyle .

Generalconsiderations onSI, SCand UI factors. Chetelat andDeVern a (1991)mad e it likelytha t "expressiono funilatera lincompatibilit y inpolle nof Lycopersicon pennelliii sdetermine db y major loci onchromosome s 1,6 and 10",wit hth eremar k that the locus onchromosom e 1 mapped near or on the S-locus. When the tomato linkage maps of Chetelat and deVern a (1991) are integrated with the potato map of Van Eck et al (1994), the flower colour locus maps on or close to the UI related locus. When these results for linkage can be extrapolated to the solanaceous family of Nicotiana,mor e can be said about this locus.Pande y (1981) described for Nicotianaglauca (SC species) the phenomenon that SI was very strongly linked with flower colour. Although the explanation was rather speculative andpartiall y flawed byto o many reported S-allelespresent , it canb eregarde d asS Ib yU Ibase dinhibitio n of SC,a sreporte d here.I nthi s study itwa s showntha t withinth eU I system,tha t operatesbetwee n tbran dver, theU Ireactio n inth e style istwo-fol d and directed against the pollen-active part of theS-locu si n verpolle n and that this pollen part of the locusi nver has still SIrelate dpropertie s(capabl eo fcausin g mutual weakening).Th e UI-reaction isals odirecte d againstth epolle nactiv eSC veTfactor , but differs inth emod e of expression from the reaction against the S-locus. All this supports, at least for the pollen part, the dual function hypothesis for the S-locus(contributio n to SI and UI) of Lewis and Crowe (1958) and is also in agreementwit hth eobservation s of Chetelat andDeVern a (1991).I t also illustrates that excluding thepossibility o fth e SIsyste mfrom contributin g toUI ,a spropose db yHogenboo m (1973),i sno t correct andtha tprevalenc e ofth eexpressio n "unilateral incongruity"ove r "unilateral incompatibi­ lity"i sno t always justified. Chapter 5 Contribution ofth e5-locu st o Unilateral Incompatibility when crossing S.verrucosum(SC )wit hS.tuberosum (SI )

Abstract.

Diploidpotato ,Solanum tuberosum (tbr),i scharacterize db y aone-locu s (S)gametophyti c self-in­ compatibility (SI)system .Th ediploi dwil d speciesS.verrucosum (ver) isself-compatibl e (SC),an d forms anexceptio nt oth erul etha tdiploi dtuber-bearin gSolanum speciesar eSI .Th ecros sver x tbr issuccessful , butgive srise t ocytoplasmi cmal e sterileF l hybrids.Th ereciproca l cross,tbr x ver, usually fails. Thisphenomeno n is called unilateral incompatibility orunilatera l incongruity (UI). Plantsshowin gth eU Ireactio nar ecalle dnon-acceptor s(NA )fo rth ever pollen .However ,exceptio ­ nallytbr plant swer efoun d toaccep tver pollen ;th eF l hybridsthu sobtaine dwer efull y male fertile. Now tbrx veroffsprin g couldb eteste d for thecontributio n of functional 5-allelest oUI . An antisense S2 construct was introduced into an52-homozygou snon-accepto r by crossing with atransgeni cS2 antisens eexpressor ,an db ytransformatio n ofthi sconstruc ti na S2Sv tbrx verhybrid , that was incompatible for S2-an d .SV-pollen,thu s showing SI andUI . Crossing the transformants withS2S2 an dSvSv teste rclone sshowe dtha tth esuppressio no fth eS Ireactio nagains tS2 coincide d with abreak-dow n ofth eU Ireactio n against verpollen . Theanalysi so fth esegregatio nratio sfo r SI/SCan dA/N Ai ntbr x verhybri dpopulation s revealed that ver does contain non-acceptor factors against own pollen, not expressed in ver, but only in species-hybridsituation swher e^-glycoprotein sar eexpressed .Thes efindings ar ei naccordanc ewit h someearlie rreport stha tth eS-locu si sinvolve d inbot hS Ian dUI .Her eth ewhol e SIan dU I system canb e explained by adua l function ofth e5-locu s(polle n and style genes contributing tobot h SI and UI), the acceptor geneA and its epistatic inhibitor geneI, apollen-expresse d 5Cver factor. A modeli spresente d explainingobserve d results aswel l as allowingprediction s based onth e afore­ mentioned intergenic interactions.

This chapter is submitted for publication in a slightly modified version as:Ronald Eijlander, Munikote S. Ramanna, Michael Ficker and Evert Jacobsen. Occurrence of Self-Compatibility, Self-Incompatibility and Unilateral Incompatibility after crossing S.tuberosum (SI) with S.verrucosum (SC): II Contribution of theS-locu st o Unilateral Incompatibility. 74 Chapter5

Introduction.

Self-incompatibility (SI)i sa mechanis m bywhic h manyplan t speciesprotec t themselves from in­ breeding by selfing. There are many mechanisms, some based on floral morphology, others on difference ofmaturatio ntim eo fmal ean dfemal ereproductiv eorgan swithi na flowe ro rplant .Ther e areals omechanism stha tar ebase do nth einteractio nbetwee npolle n(tubes )an dth epistil .Thi sma y happena tal lstage sbetwee nlandin go fa polle ngrai no nth estigmati csurfac ean dfertilisation . There aretw omajo r incompatibility systems:th esporophyti c one,wher eincompatibilit y iscontrolle d by the interaction of the genotypes of the pollen parent and the style parent (sporophytes), and the gametophytic one,wher eth eincompatibilit y isdetermine d byth einteractio n ofth egenotyp eo fth e pistillateparen t andth egenotyp eo fth epolle n grain (gametophyte). Irrespective ofth e incompati­ bilitysystem ,i ti sfrequently foun d thathybridisatio n ofrelate dspecie si s possiblei non e direction only.I ti susuall yfoun d whenself-compatibl e (SC)specie sar ecrosse dwit h related SIspecie s(e.g . Anderson and De Winter, 1931; Mather, 1943;Lewi s and Crowe, 1958; De Nettancourt, 1977, etc),that the SIspecie sar esuccessfu l aspollinators ,bu tno t as pistillateparent s (SC x SI- >F l; S I x SC- >-) . Thisphenomeno n iscalle dunilatera l incompatibility orunilatera l incongruity (Hogen- boom, 1973). Hogenboom tried to distinguish between SI and UI and introduced the concept "incongruity" for inhibitory reactions that are not based on self-incompatibilty, arguing that incompatibility andincongruit y areseparat ephenomena . Fora lon gtim e adebat ei songoin g about apossibl einvolvemen t ofth eS-locu si nUI ,and ,directl yrelate dt othis ,whethe r inthi s connection theter m incongruity orincompatibilit y should beused . For the Brassicaceae (sporophytic system) it was found likely that the S-locus is involved in UI (Hiscock and Dickinson, 1993). The genetic analysis of an interspecific hybrid system inLyco- persicon,(Solanaceae) ,wit h aon elocu sgametophyti c system,showe d the likelihood of (apar t of) theS-locu sbein ginvolve d inU I(Chetela t andDeVerna , 1991). Forth e Solanaceous species,ver i nparticular , ageneti cmode l for the evolution of speciesfrom S I to SCwa spostulated , inwhic h the dual function ofth e»S-locus ,a spropose d by Lewis and Crowe (1958),i scrucial .Her eth eS-locu scontribute st obot hS Ian dUI .Th etwo-power scompetitio nhypo ­ thesis is based on this and on the co-evolution of sympatric SC and SI species (Abdalla, 1970; Abdalla and Hermsen, 1972;Abdalla , 1974).Th edevelopmen t of CMS (Abdalla, 1970; Abdalla andHermsen , 1971)i ncas eo finterspecifi c hybridisation with vera spistillat eparen t ison eo f the necessary results inthi s hypothesis,turnin g most hybrids into a"dea d end". The reciprocal cross wouldb epossibl eonl ywhe nth eS Ispecie sha sn o UIgenes ,tha tcoul dinhibi tth epostulate dS callel e ofth e SC species.Hybri d progeny created with such t/Z-lackingplant s could open the possibility to investigate theU I model for the contribution ofbot h stylar and pollen determined factors. Ina diploi dpotat opopulatio n originatingfrom tw oS Cparents ,no t accepting verpollen , Hermsen eta l(1974 )detecte dclone stha twer eS Io rS Can dwer eaccepto rfo rver. Theyanalyse dth egenetic s Contribution ofthe S-locusto UI 75 ofacceptanc eb ypollinatin g severaltbr population swit hver polle nmixture san dfoun d segregation of acceptors (A) and non-acceptors (NA).Tw o modelswer eteste d for fitting the ratios.Th e first model was based on two independent loci,A, andA 2, where onlyth e doublerecessiv e genotypes

(axaxa2a^) areacceptors .Thi smode lwa ssimila rt oth ehypothesi sb yGru n andAuberti n (1966),bu t didno tfit tw oou t offourtee n observedratios .Mode l 2wa sbase d on the independent genesA and 1.1i sa ninhibitor , epistatic overth edominan t acceptorgen eA. Here,onl y UAA and iiAagenotype s couldb eacceptor . Thismode l fitted theobserve d segregationsver ywell . The segregation ofth eS - allelesan dth eSC-facto r tSl, thatwa spresen ti nth eorigina lparents ,segregate d independenlyfrom theacceptance .A thoroug hanalysi so fthi s andrelate d material by Hermsene t al(1977) , inwhic h verpolle nwa sno tpooled ,reveale d akin d ofgene-fo r generelationshi p inpenetratio n and barrier capacityo fpolle n and style,resultin g in adifferentia l reaction pattern. Their explanations and ex­ pressionscorrespon dwit hthos euse dfo rth eincongruit yhypothesi s(Hogenboom , 1973).Th eresult s were,however , notconclusiv e aboutth eexac tmechanism , andth e authors left open the possibility of other explanations. Iti sclea rfro m the study on SCi nth ehybri d system oftbr x vertha tbot h theS-locu sfro m vero r tbran d(non-)acceptanc epla y arol ei nth eexpressio n of SCan d SI(Eijlande r et al.,b ,submitted) . Forth epolle npar t ofth eS-locus , adua l function wasmad e likely,thu s introducing atleas t apar t ofth eS-locu si nbot hth e SIan dth eU Ihypothesis ,whic hi s inaccordanc ewit hth econclusion s of Chetelat and DeVerna (1991) andFoola d (1996).Murfet t et al(1996 ) showed by amolecula r ap­ proach intobacc otha t introduction ofa nactiv eS-allel eca ncontribut e to unilateral incompatibility in those solanaceous species, but they did not link this to a genetic model. We used here both a molecular andgeneti c approach for thehybri d system oftbr x vert oprov e that, like inNicotiana, the stylar S-locusproduc t (S-glycoproteins, also called S-RNases)ca n contribute to UI. We also integrated the results into an already existing genetic model for UI, explaining why interspecific hybridisation canresul t inunexpecte d appearance ofS Co rS I/ U Ibase d crossingbarrier s between Fl hybrids and theparenta l species.

Materials and methods.

Themateria l that was used, wasbase d on the expectations that factors likeS-alleles , (responsible forth eS Ireaction) ,th eaccepto rgen eA (dominan tove rnon-accepto rallel ea) tha tcause sacceptanc e of ver pollen, and the acceptor-inhibitor gene I (epistatic overA, thus in dominant form always causing UI) couldb e identified by electrophoresis (S-alleles),b ytest-crossin g with tbr,ver o rb y selling. The S2 antisense construct that was introduced here, was earlier proven to be effective against the S2 incompatibility allele, suppressing the synthesis of the S2 glycoprotein. With this material (see laterfo r details)materia lcoul db ecreate d andselecte dt oanswe rth equestio nwhethe r the S-locuscontribute s to theunilatera l incompatibility reaction. 76 Chapter5

Non-transgenicclones. SixF 2populations ,code d 1061t o 1066(u pt ote nplant s each) wereobtaine db y selfing ofth e Fl- acceptorclone s6541-01 ,-02 ,-03 ,-06 , -11an d-2 5respectively .Thes e6541 -plant sar eal ltbr x ver

Fl hybrids(6234-0 8x ver P I 195172-27)o fth egenotyp eS3SvAAii SCva- orS3SvAaii SCver-, and thusaccepto r ofve rpollen .SC vai sa pollen-expresse d self-compatibilizing factor, derivedfrom ver (Eijlander etal. ,b ,submitted) .Th eF 2population s 1061t o 1066wer eexpecte d to segregate S3S3, S3Svan dSvSv genotypes ,and ,wheneve r a6541-paren twa saccepto r ofth etyp eAa, alsoi nAA an d Aa acceptor andi naa non-accepto r genotypes.I ntbr, theaa genotype sbehav ea snon-acceptor s of verpollen , thus showingU Iwit hver. Population 6484(1 4plants )originate dfrom th ecros so fth enon-accepto r tbrclon e 6233-11 (S2S2, aall) with the fertile near-ver cloneTV 6-14 (14thplan t ofth e 5thbackcros s generation of the Fl acceptor-fftr (T)x ver (V)) . TV6-14wa so fth egenotyp eSvSv (nostyle-activit y ofth e S-allele)an d containedth epollen-expresse d SC-factorSC ya.Sv an dSC verwer etransmitte dt oth e6484-population . Theplant so fthi s6484-populatio nhav epreviousl ybee nteste dfo racceptanc ean dfo r SC.Al lplant s were non-acceptor for verpolle n and self-incompatible (SI). SIo fth e 6484-plantswa s explained bya combinatio no fgametophyti cself-incompatibilit ywit hS2, Ul-base dre jectio no fSv an da specia l , interactionbetwee n genes expressed inpolle nan dstyle ,directe d agains S2+SCverpolle n (SIb yUI - based inhibition of SC;Eijlande r et al.,b ,submitted) .

Transgenic clones. Clone 6484-06 was a randomly chosen SI genotype out of the afore-mentioned non-transgenic population,wit hth egenotyp eS2Sv an dhad ,presumably ,th eaali non-accepto r genotype.A sstate d above,i tcontaine d apolle nexpresse d "self-compatibilizing" factor (SCver).Thi s factor appeared to beexpresse di ncombinatio nwit ha functiona l tbrS-allel e(causin ga mutual-weakning-lik e effect), butonl ywhe nth etbr o rhybri drecipien twa sno to fth eaall or aalinon-accepto rgenotyp e (Eijlander et al., b, submitted). The clone 6484-06 was transformed with the 52-antisense construct (P35- S2AS)2(Eijlande re tal. ,a ,submitted )i nAgrobacterium tumefaciens LBA440 4(Hoekem ae tal. , 1983). Transformation wascarrie dou ta sdescribe db yFlips ee ta l(1994 )an dEijlande r eta l(a ,submitted) . Population 1184wa s obtained by pollinating tbr clone 1140-02 (S2S2,SI ,NA ) (Eijlander et al., 1997;Eijlande r eta l b,submitted ) with transgenic tbrclon e (P35-S2AS)2- V -6 .Thi s transgenic clone V is of the genotype S2S10, NA, expressing the afore-mentioned antisense S2 construct

(Eijlander et al., a, submitted) andi tbear s apollen-expresse d SC factor (no t SCver)derive d from clone 1024-02 (Kuipers et al., 1994)),allowin g for ^-penetration inS2 expressin g styles.Du et o thepollen-S Cfactor , S2S2genotype sca nb eobtained ,tha t are consequently all expressing this SC factor again. The antisense S2 construct was transmitted by pollen. So, it was tested whether offspring plants showed an antisense S2induce d reduction of^-incompatibility . Thepopulatio n was screened by IEF and test crossed with S2 pollen for presence of S2S2 homozygotes that expressed the52-antisens e construct. Subsequently, theplant swer eteste d for UIb ytes t crossing with ver. Contribution ofthe S-locusto UI 77

Monitoring incompatibility(UI & SI) Testcrosse sfo rmonitorin g biologicalexpressio no fantisens eeffect s onS Ian dU Irespectivel ywer e performed bypollinatio n with pollen from S2S2homozygote s (Eijlander et al., 1997) and pollen mixtures of the ver (PI 195172) offspring clones 6555-31, 6555-33 and 1076-06. Styles were harvested 48hour s after pollination and fixed and stained accordingt o themodifie d technique of Martin(1959) .Polle ntub epenetratio n was observed under a fluorescent microscope (Eijlander et al., 1997).

Proteingel electrophoresis Up to 50 mg of plant tissue was ground in an Eppendorf tube with 20-100 ml 5m M potassium phosphatep H6.0 , 2.5% (w/v )sucrose , 0.1% (v/v )b-mercaptoethanol ,usin ga ground-glas spestle . Singlestyl eextract swer emad ei na volum e of2 5m lextractio n buffer. After centrifugation ofth e homogenate at 14000g for 15 min, the supernatant was fractionated on horizontal thin-layer isoelectric focussing (IEF)Polyacrylamid e gels(pH3.5-10 ) (Schmidt-Stohn, 1979) or agarose gels (Hypurege lV G 1020,Isola b inc).Silve rstainin gwa sperforme d asdescribe db yKirc he ta l(1989 ) for thePolyacrylamid e gels and according to company specifications for the agarosegels .

Results

The effect of antisenseS2 on SI andU I inS2Sv andS2S2 non-acceptor genotypes

The molecular approach of gain and loss of function was very successful in proving that the S- glycoproteinspla ya ke yrol ei nth eS Ireactio no fth esolanaceou sspecies .Her ew edescrib eth elos s of function approach as used in potato (Eijlander et al., a, submitted), but now applied to two different S2containin ggenotype so fnon-acceptor s for verpollen :th etbr-ver hybri d 6484-6(S2Sv) and apur e tbrpopulation : 1184,containin gth e^-genotype sS2S10 an d S2S2. The SI,N Aclon e6484- 6wa stransforme d withth eantisens eS2 construc t (P35-S2AS)2.A tota l of 40transgeni chybri dplant swa s analysed for theexpressio n ofth eS2 antisense transgene. Testing by pollination with only 52-pollenreveale d that in one genotype, (P35-£2AS)2-6484-6-4, the SI reactionagains t S2wa s effectively suppressed.Thi swa sconfirme d byIE Fo fth e stylar extract that showeda reductio no f^-glycoprotei ncontent .S2 polle nwa scompatibl ewit hthi stransgeni chybri d clone,clearl y contrasting withth e strong SIreactio n inth euntransforme d clone6484-6 . The same 40 clones were also tested for acceptance of verpolle n by monitoring the pollen tube ingrowth and seed set. Thirty-nineplant s showed aver y strongU Ireaction , comparablewit h that of the untransformed genotype. Only one clone was altered in this respect. This was the same transgenicclon ea sth eon etha tha dbecom ecompatibl efo rS2 pollen :clon e(P35-S2AS)2-6484-6-4 . 78 Chapter5

acceptance non-acceptance antisense S2 antisense S2

(P35-S2AS)2-6484-6-4 6536-01 x ver 6484-06 x ver x ver 1184-01 x ver Fig. 1 Reactions of different stylar genotypes on ver pollen. Effects of àntisense-S2-RNase based inhibition of S2 RNaseproductio n onU I inS2-RNase expressin g non-acceptors (c,middl eright an dd , far right) compared with normal acceptance (a, far left) andnon-acceptance (b, middle left) of ver pollen. Insets in panel d: callose plugs in the most compatible pollen tubes. See also text.

TheU Ireactio n against ver, though stillintac t (incompletepolle ntub epenetratio n andthu sn osee d set),wa sstrongl yreduce dan ddiffere d clearlyfro m thosei nth eothe rclones .Here ,man yver polle n tubesreache d to 2/3o fth estyle ,wherea s inth enon-transgeni c plant S2polle n did not surpass 1/4 tol/3 ofth e stylean dver polle nwa seve ninhibite d at 1/5 to 1/4 ofth estyl e(Fig . 1 b,c) .I n thever pollentube s therewa s still alo t ofcallos e deposition andman y far-reaching pollen tubes showed inhibitionphenomen a likethickenin go fth etubes ,irregula rshapes ,spong ycallos edepositio n along large stretches ofth epolle n tubes and inflation ofth etips . Thecros s 1140-02x (P35 -£?AS)2- V- 6 ( =S2S2 x [S2S10, pollen-expressed SC,+ 2 copie s ofS2 antisense])resulted , asexpected ,i nS2S2 andS2S10 offsprin g plants.Twent yo fthe m flowered well and were tested for SI and UI reaction by selfing and test crossing with S2 and Svpollen . Seven cloneswer e SIan dwer eo fth eS2S10 genotype .Thirtee nwer epollen-base d SCan d seven of them hadth edesire dS2S2 genotype.Tes tcrossin gdetecte d antisense induced S2 suppression at various Contribution ofthe S-locus to UI 79 levelsi nsevera lS2S2 an dS2S10 genotypes .Clon e 1184-01wa sS2S2, an d expressed the strongest antisenseinduce d S2suppressio n amongal lS2S2 plant stested . However, this suppression wasno t as strong as that in the original transgenic parent. Variation in SI for S2 ranged from pseudo- compatible to compatible with S2pollen , which ranparalle l to theS2 glycoprotei n content ofth e styles. Clone 1184-04,als o S2S2,showe d variable reaction patterns from strictly incompatible to compatible with S2pollen . Theclone s that showedn o antisense effect onth e SIreactio n against S2polle n were, as expected, allnon-accepto r ofve rpollen . Reduction ofth eU Ireactio n upt o acceptancewa s only detected in S2S2plant stha t showed the^-antisens e effect on SI.Th eU Ireactio n against verpolle n in clone 1184-01range d from normal UI to highlypseudo-compatibl e (Fig. Id), with limited seed set and clone 1184-04wa s slightly less compatible with verpollen . TheU Ireactio n was simultaneously suppressed withth e SIreactio n forS2 polle n andra nparalle l to thedecreas e ofth e^-glycoprotei n content ofth estyle .Whe nth ebreakdow n ofU Iwa s strong,th epolle ntube s appearedt ob enormal , with small,regula r interspaced calloseplug s(se einset si nfigure Id) . Inth e lowerpar t ofth e style, however, many tubes, but not all, showed reaction patterns as observed after pollination of the antisense-.S2transgeni c 6484-6 clonewit h verpollen(se e above). Thedifferenc e inreactio npatter ni nparticula r inclon e 1184-01(see dse twit hver) comparedt otha t inclon e(P35-S2AS)2-6484-6- 4(inhibitio n at 1/3 ofth e style),probabl y indicates thatbot h clones havedifferen t non-acceptor genotypes.Nevertheless ,bot htype s ofmateria l confirm that reduction ofth estyla rS2 glycoprotei n contentcoincide swit hth esimultaneou sbreak-dow n ofth e SIan dth e UI reaction, thus implying that at least the^-glycoprotei n contributes toth eU I reaction.

Testing for (non)acceptance ofSvSv aaiigenotype s with verpolle n

It was shown that theS-locu si s involved in the UIreaction . The observations suggested that this contribution depended also on the type of non-acceptance, including the possibility that non- acceptanceca nb eexpresse deve ni nth eabsenc eo fa nactiv e5-glycoprotein . Genotypeswit hinactiv e or defective S-allelesma y shed more light overthi s question. Those genotypes must have anon - acceptorbackgroun d (non-^4*ii) inorde rt ose ean yeffec t ofsuc ha defec t onth eU Iexpression .Th e ^-alleles of ver show no stylar activity, so the introgression of these alleles in a non-acceptor background may serve as an example ofnon-functiona l tbr^-alleles .

The hybrid population 6541 (S3Sv, SCver—) is known to contain solely Aaii and AAii acceptor genotypes,base do ntes t crossingwit hver polle n(Eijlande r et al, b,submitted) .Afte r selling,an y Aaiiaccepto rparen t isexpecte d to segregateint oacceptor s andnon-acceptor s in a3: 1rati o ofA *ii andaaii genotypes . Selfpollen , capable ofpenetratin g the style,i so fth e genotype Sv,S3 SCver or

SvSC verThus ,whe nn oothe rselectio nmechanism spla y arole ,one-thir d ofth e genotypes will be of thedesire d SvSvgenotype , allth e others S3Svo rS3S3. Five ofth e sixpopulation s of selfed 6541 clones suffered from inbreeding depression and noneo f thepopulation s gave over seven testable flowering plants. 80 Chapter5

Four self populations showed the expected segregation for acceptance (AAii or Aaii) and non- acceptance(aaii) whe nS3S3 an dS3Sv genotype swer eteste dfo racceptanc eo fve rpollen ,indicatin g descendence from Aaii genotypes. SvSv plants were separately selected to be tested for acceptance.The hypothesis Ho is,tha tan ySvSv aaii genotyp ewil lb enon-acceptor , indicating that this type ofU I is independent of style-expressed S-alleles,an d Ha istha t SvSvaaii genotype s are acceptor because they lack style-active S-alleles. Those four segregating populations gave 27 - flowering plants, eight ofthe mbein gSvSv (n oS3 glycoprotein detected).A sexpecte d underHa , allSvSv genotype s were acceptor. Under Ho the likelyhood ofinadvertentl y not detecting anSvSv aaii genotype is P(k =0,N = 8|p = 0.25) = 0.758 = 0.10. These observations show that there isa strong indication for 5-alleledependenc e ofnon-acceptanc e expression inaaii genotypes .

Analysis ofver o nN Ab y analysis oflikelihood s ofexpressio n patterns in population 6484.

Asmentione dbefore , oneo fth eparent s ofth eF l population 6541(tbr 6234-0 8 x ver)mus t have beenheterozygou s(Aa) forth eaccepto rgen eA, becausenon-acceptanc e(NA )segregate d inth eF2 - populations(code d 1061 to 1066),a swel la si nsom eo fth eBC 1 populationsmentione db yEi jlande r et al (b, submitted).Th e tbrparen t 6234-08 was S3S3AAU, thus implying that verwa s the most likely source for a, becauseo fit sorigi n denoteda v. Thiswa si naccordanc e with old data onsegre ­ gationo faccepto r linesi nver x tbr an d (verx tbr)x ver crosses ,base do nhybridisation s of verwit h thediploi d tbrclon e G254 (SIS3, Aaii) (unpublished results).Jus t like theresult s from the afore­ mentioned experiment, these unpublished results point to the possible existence of aaii ver genotypes,bu t areno t conclusive.

Research on the expression of SC in tbr-verhybrid s (Eijlander et al., b, submitted) resulted in a complex hypothesis concerning the suppression ofpolle n mediated SC in specific non-acceptor genotypes. This hypothesis, explained hereafter, suggested that the near-ver line TV6 could have containedth enon-accepto r allelea v.Th ecros swit hth e self-incompatible non-acceptor clone 6233- 6 11 (S2S2, aaii)woul dthe nresul ti na tleas tsom eaaji v genotypes.Theoreticall y onea allelei nTV couldhav ebee nderive d from theorigina l tbraccepto r clone,bu t thepresenc e of asecon d a allele 6 (thus TV = avaji oraaji) should imply thepresenc e ofthi srecessiv e allele in the verbackcros s parents.The inhibitor alleleI seemst o obstruct further analysis ofinteractio nbetwee n S-locusan d aa, because of its epistatic behaviour, but the presence of SCva can bypass this problem. It was hypothesised that thepolle n expressed 5Cver-factor, as found in thehybri d clones 6484-06 and in 6541-06(Eijlande r etal. ,b ,submitted) ,i sno t functioning inaal* styles.Clon e6484-0 6woul dthe n be SIb yUI-base d SC-inhibition. Based onthis ,a nanalysi sha sbee n made for thepossibl e genetic constitutions ofth eparenta lclone sTV 6-1 4an dtbr 6233-1 1o fpopulatio n 6484(tabl e 1).Fro m this 6 analysisi tca nb ededuce d thatTV -14bein gA vaji hasa maximu m likelyhood as lowa s2% . With 6 arelativ elikelyhoo do f96 %th emos tlikel ygenotyp eo fTV -14i saaji ora vaji andtha to f6233-1 1 aaii. Thedominan t geneƒ coul d notb eteste d here onS-allel edependance . Contribution ofthe S-locus to UI

Table 1.Likelihoo d (y) for erratically missing a SCplan t inpopulatio n 6484 (14plants ) under the assumptions 6 thatSC ya isinhibite di na nS2Sv aaii stylebu texpresse di na nS2SvAaIi style an dtha tTV -14ca nb eeithe rSC ytr -

orSC verSC y„. Thisi scalculate d for alltheoreticall ypossibl e(non )accepto rbackground s forbot hparenta lclones . TV4saccepto rfo rve rpolle nan dself-compatible , 6233-11i s52-homozygous ,self-incompatibl e andnon-accepto r for ver pollen.

S 6 Genotype of TV -14 Genotype of 6233-11 YifTV"-14 = SCv„- YifTV -14 = SC„r5Cv„ SvSv, AAH S2S2, AAII 0.00006 0 S2S2, Aaii 0.00006 0

S2S2, aaii 0.00006 0

SvSv, Aaii S2S2, All 0.00006 0

S2S2, Aaii 0.0014 0.00000

S2S2, aaii 0.0178 0.00006

SvSv, aaii S2S2, AAII 0.00006 0

(non-acceptor?) S2S2, Aaii 0.0178 0.00006

S2S2, aaii 1 1

Discussion

Contribution ofthe S-locus to UI The approach in testing the contribution ofth e stylarpar t ofth eS-locu st o non-acceptance of ver pollen tbr styleswa sbase d onthre edifferen t typeso fmaterial : - hybrids that did not expess atbr 5-allel ebu t with aputativ e non-acceptor genetic background (SvSvaaii)

- hybridstha twer enon-acceptors ,containe da polle nSC yerfacto rbu tdi dno tsho wS Cwher ei twa s expected, unlesscertai n interactionswer epostulate d (SIb yUl-base dS Cinhibition , directed by aal* styles against SC„„ containing pollen). This hypothesis allowed for the analysis of the acceptance of self-pollen in ver, - non-acceptorclone swit honl yon etbr 5-allel e(S2S2 o rS2Sv) tha twer eantisens eS2 transformant s andshowin g antisenseinhibitio n ofth eS2 alllele ,thu s disrupting aputativ econtributio n ofth e stylarproduc t to theU I reaction. Hybridstha twer ebot hmal ean dfemal e fertile andtha tcontaine donl ySv wer eobtaine db y selling Fl hybrids.Th efive sel fpopulations , 1060t o 1065,gav e strong indications aboutth e relationship between (non)acceptance andth eS-locus ,becaus e noSvSv non-accepto r genotypeswer e detected 82 Chapter 5

(H0: SvSv aaii =non-acceptor) . The alternative hypothesis Ha (SvSvaaii = acceptor), that in this material any non-acceptor ofth egenotyp e aaiineed s astyle-expresse d 5-allelet o exhibit itsnon - acceptornature ,canno tb eprove n by direct identification ofaaii accepto r genotypes,bu t can only beestablishe db yth erejectio n ofth ehypothesi s thatSvSv aaii genotype s arenon-acceptor s at any time. Thus,H aha s alikelihoo d of90% .

Thepostulate d suppressiono fth eSC m factor inaaii and aaii styles ,a soccur safte r selfing ofclone s ofth e648 4populatio n (TV6x S2S2 aaii), enabled atheoretica l genetic analysis ofTV 6. When the commonly applied confidence level of 5%i sapplied , it canb e concluded that TV6-14i sSvSv aaii 6 SCve*.Thi simplie s that the genotype ofve r isexpecte d tob eaaii. TV -14i sself-compatibl e and hassuccessfull y beenbackcrosse dwit hver, resultin gi n TV7. Previousdat a(Hermsen , unpublished) onsegregatio n ofve rx tbrhybrid s into acceptors andnon-acceptor s support our finding thatAaii and evenaaii vergenotype s do exist,tha t are acceptor for self-pollen.

Another approach to test for the dual function of the S-locus was to introduce the sense S2 glycoprotein constructsb ytransformatio n (Eijlander etal. ,submitted) .An ySvSv aaii genotyp etha t expresses theS2 transgene , should then change from acceptor to non-acceptor for verpollen.Thi s transformation hasbee nperforme d on several verclone s with the genotypeAaii or aaii.Th e only clone that has been analysed thoroughly for being aaii, was TV6-14. Despite the fact that the transformations were successful and resulted in nearly 100% (transgenic) callus formation, no régénérantswer e obtained (datano t shown),thu s disabling this option. Murfett et al (1996)wer e successful withthi sapproac hb ycausin ga nU Ireactio nagains tNicotiana tabacum andN. glutinosa (both SC-species) pollen in N. tabacumwhe n the introduced S2 ^-glycoprotein ofN. alata (SI- species)wa sexpresse d athig hlevels ,bu tthi sapproac hfaile d forth eS Cspecie sN. plumbaginifolia, indicating different UIbackgrounds . Although inou rexperiment s theapproac ho fsense-transformation-induce d UI failed, the approach ofknockin gdow nU Ib ya nantisens eS-allel ewa ssuccessful.Th e numbero fantisens eS2 transgeni c 6484-6 plants showing antisense effect was much lower than previously reported for clone V (Eijlander etal. ,a ,submitted) ,th epolle nparen to fpopulatio n 1184.Th eS2 incompatibilit y reaction inth enon-transforme d plants isquit e strong. Anexplanatio n for this low frequency might be that because ofth e absence of anadditiona l style-expressed S-allele, anup-regulatio n ofth e52-allel e isobtained .Thi sexplanatio ni sno tunlikel ybecaus edown-regulatio no f^-allele sb yover-expressio n oftransgeni cS-allele sha sbee n shown (Eijlander et al., a, submitted).Natura l weakening ofth e SI reactionb ymodifie r genesi smentione db yman yauthor sfo rman ycrop s(Mather , 1943;Takahashi , 1973; Henny and Ascher, 1976;Litzo w and Ascher, 1983) and it istherefor e tempting to assume thatther eare ,a sa nalternativ epossibility ,modifie r genesi nthi sclon etha tenhanc eth eS2 expressio n rathertha nweake n it,thu slimitin gth eantisens eeffect . That this stronger SIreactio n iscause d by anon-acceptanc ebackgroun dca nneithe rb econfirme d norrejected ,becaus emor eloc itha nA and ƒar e likely to play a role.The extremely strong SI and UI reaction in the original 6484 hybrid Contribution ofthe S-locus to UI 83 population showed that those plants, including genotype 6484-06, were highly reliable in those reactions. Thiswa sconfirme d for allth etransgenic s derivedfrom plan t 6484-06,excep t transgenic number four, thatprovide d theproo f for S-glycoprotein involvement inUI . Inadditio nt othis ,th etransgeni c clones(i n648 4an d 1184populations ) that showed abreak-dow n ofU I formed anadditiona l controlpopulatio n bythemselves ,becaus eth e fluctuation in strength of theU Ireactio n coincided withth e fluctuation in strength of SI andi nS2 glycoprotein content. There is a small discrepancy inth ebreak-dow n ofS Ifo r S2polle n andU I for verpolle n between the transformed hybrid 6484-06 and the anti-sense S2 expressing transgenic tbr clone 1184-01. Althoughbot hclone s showed abreak-dow n ofU Itha t coincidedwit hth ebreak-dow n of SIfo r the S2 allele,th e latter transgenic showed astronge r break-down ofUI , that resulted in seed set after pollinationwit hver. Thismigh tb edu et o astronge r expression ofantisens eS2, but 1184-01bein g anon-accepto r ofth etyp eA *I* oraaii insteadofaal*, isprobabl y abette rexplanation . Basedbot h onth eanalyse spresente dher ean do nth emode lfo rS Cexpression /inhibitio npropose db y Eijlander eta l(b ,submitted) ,6484-0 6 should haveth e genotype aaii,thu s also blocking the SC-factor from ver.Thi simplie s anadditiona l UI factor to react upon and astronge rU Ireactio n inthi sbreaking - down situation. Thatth e 1184populatio n showed aweake r antisenseinduce dbreak-dow n ofth eS2 incompatibility reaction, might be due to variation in transmission of the T-DNA, or to the homozygosity ofth eS2 allel e inth e offspring (Heeres et al., 1998). Itwa s proven (Murfet t et al., 1994;Le e et al., 1994;Eijlande r et al., submitted) that the stylar SI reactioni sdetermine db y^-glycoproteins .Chetela tan dDeVern a(1991 )showe dtha ti nL. pennellii- pollenU Iwa sdetermine db ya tleas tthre eloci ,on eo fthe m mapping ono rnea rth eS-locus . Foolad (1996) found even more UI related loci, accounting for skewed segregations.That ver is still expressing a pollen factor, was proven in our experiments on SC in the interspecific hybrids (Eijlander etal , submitted).Theoretically , aU Ideterminin g factor might existtha ti sclosel ylinke d to the S-locus. Breaking ofthi s linkage might be as difficult asbreakin g the linkage between the pollen- and style-factor ofth eS-complex .Becaus eth epolle n factor andth e style factor areclosel y linkedan dtogethe r constitute theS-locus , the involvement ofth eS-glycoprotei n inU I shouldno t be surprising anymore. Murfett et al (1996) showed that in Nicotiana the introduction of an S- glycoprotein caninduc eU I characteristics, that wereno tpresen t before. By antisense suppression theycoul deliminat eS Ian dU Ireactio npattern sfo rrespectivel yN. alata(SI )an dN. plumbaginifolia (SC)fro mTV .plumbaginifolia x N. alata(S Ispecies ,NA )hybrids .Thes eresult sar ecomparabl ewit h thosew e found in the antisense S2expressin g clones 6484-6-4 and 1184-1.

Thegene s directlyresponsibl e for (non-)acceptanceA andI (ora , anda 2, Hermsen et al., 1974;o r

IM,an d UI2, Abdalla, 1974) segregate independently ofth eS-alleles .However , acceptance isno t completely independent of the S-locus. That this has not been detected previously in thetbr-ver system, canb e attributedt oth etechnica l difficulties encountered inth epas t in detectingS-alleles , but also to therar e occurrence ofth e relevant genotypes in self- andbackcros s populations. 84 Chapter5

Descriptivemodel for interactions ofSC, SI and UIdetermininggenes. Allth eresult sobtaine d on SC,S I andU I anddiscusse d inthi s and in our afore mentioned article, canb e explained byth e lociA andI, the5-allele s from thetw o species (tested here:S2,S3,S4 an d

Sv),th epollen-activ eself-compatibilizin g factorSC vaan dth eearlie rdiscusse d interactionsbetwee n them.Th eoccurrenc eo fdifferential s (tbr-genotype stha tar eacceptor sfo ron ever genotyp ebu tnon - acceptors for another ver)a sdescribe d byHermse n and Sawicka (1979) and also observed byth e present authors may be explained by the existence ofmor etha ntw o alleles ofA and/or a dosage effect of the genes involved. The observations made and itspredicte d (incompatibility reactions are shown in table 2.Thi stabl e dealsonl ywit hth e allelesA and a.Her e individual reactions of pollengenotype s in styles ofvariou s tbran dver base d recipients are indicated. Pollen can be ofa puretbr type , ahybri dtyp eo ra ver type .Ploid y effects areals oindicated : the left block deals with haploidpolle nfrom diploid san dth eothe rbloc k (right) shows somepolle n typesproduce d by, for

Table 2. Predicted and/or observed interactions between pollen and style in various acceptor and non-acceptor back­ grounds, based on tbr and ver. S2 and S3: 5-alleles from tbr, Sv = S-allele from ver. S- = silenced S-allele (by antisense) from tbr. Recipients with the genotype SvSv react like S-Sv. Pollen with S2S3 and S2Sv: diploid pollen as produced by tetraploids or by arestitutio n mechanism active in diploids. A and a: acceptor alleles. A= acceptor,

/and i: inhibitor alleles, ƒi s epistatic overA. The genotype aaii needs an active S-allele to cause UI. SCy„ = pollen-

expressed self-compatibilizing factor, not effective in a Slbr * aal* style. PC= pseudo-compatible= incomplete incompatible reaction, potentially giving some seed set. The question mark ? is placed where the effect of counter­ acting powers is unknown and depends on dominance relationships. See also text.

Pollen genotypes

Style genotypes produced by diploids produced by tetraploids

S3 S3SC„, S2 S2SCm Sv SvSC„, S2S3 S2Sv S2SvSC,„

S2S2,A*ii + + - + + + + + +

tbr S2S2,A*Ii + + - + - - + + + S2S2, aali + 7 - - - - + + 7

S2S2, aaii + + - + - - + PC/+ + S2Sv,A*ii + + - + + + + + + tbr S2Sv,A*I* + + - + - - + + +

X S2Sv, aal* + 7 - - - - + PC PC? ver S2Sv, aaii + + - + - - + + +

tbr S-Sv,A*ii + + + + + + + + +

X S-Sv,A*I* + + + + _? —? + + +

ver, S-Sv, aal* + 7 + ? _? —? + + + asS2 S-Sv, aaii + + + + + + + + +

ver SvSv, aaii + + + + + + + + + Contribution ofthe S-locusto UI 85 instance, tetraploids. The inactivation of a tbrS-allel e (stylar expression) by means of antisense gives,whe n 100%effective , functionally anSv allele .SvSv genotype s aretherefor e not mentioned. TheS Ireactio n againstS2 polle ni na nS2S2 stylei spresumabl y strongertha n ina nS2S3 (M cCub - bin etal. , 1997)o r anS2Sv styl e(dosag eeffect) . Iti spostulate d that pollen containing both Svan d

SCvercause sa norma l orenhance dU Ireactio nan dtha ther eth einteractio n liketha tbetwee n S2an d

SCm isno t effective ina nS2Sx aal* styl e(Eijlande r etal. , submitted).Questio nmark s indicatetha t the final effect depends on the balance of different effects, like in S2S2aaIix S3SCyer, where inhibition of5C vermightb eepistati cove rth ecompatibilit y ofS3. Co-dominance of S3will leadt o amoderatel y compatible reaction and epistasy to full compatibility. Svpolle nwil lpresumabl y be inhibited in anyI -containing style,unles sI toodepend s on^-allel e activity. Alleles for UI other thanA versus a andI versus iar eno t consideredhere .

Most ofth e^-glycoprotein so fth e solanaceaous specieshav eRNase propertie s and areessentia l for activity of SI (e.g.McClur e et al., 1989) and aretherefor e also addressed asS-RNases. Loss ofS- RNaseactivit y hasbee n shownt o result in self-compatibility (Royo et al., 1994;Kowyam a et al., 1994). Thatabsenc eo fRNase activit ywoul db eenoug ht obypas sa U Ireaction ,i sno ttrue .Ther ear emor e factors involved in this, as is exemplified by the UI reaction between N.plumbaginifolia (SC- species)an d aS Clin e(S Cdu et oabsenc eo f S-RNase activity) ofth e SI speciesN.alata (Murfett et al., 1996). UI was still intact here.L. esculentum and L.peruvianum, tw o species that are less relatedt oeac hothe rtha nver an dtbr are ,sho wa ver ystron gU Ireaction . Heretoo ,th ecros si sonl y successful whenperforme d onth e SCspecie sL. esculentum. There is,however , aS C line from L. peruvianum, LA2157, that is SC due to a mutation in the coding region for the S-glycoprotein, causing loss ofRNase activity (Kowyama et al., 1994;Roy o et al., 1994).Ric k (1986) reported, however,tha tal lth eline sinvestigated ,retaine dthei rU Ireactio nagains tL. esculentum. Ourresult s confirmed hisobservatio n thatplant shomozygou s forthi smutation ,ar ehighl yU Iwhe n pollinated by L. esculentum.Althoug h thepolle n tubes from thispollinato r appeared to penetrate perhaps a tentho fa styl elengt hdeepe rint oth e styles ofth e SChomozygou s plants than in styles from SISI or SISC plants (R.Eijlander, unpublished data), the differences were minute: all plants were definitely UI forL .esculentum.The simplest explanation for such astron g UI reaction, even when the ^-glycoproteins possess no RNase activity, might be due to a gene being different from the genes dealt with here. Such a gene might be strongly expressed when species are more distantly related and not as close as tbr and ver.Othe r explanations migthb e apleiotropi c effect of/, ex­ pressed in wider crosses, or stronger alleles of this gene in Lycopersicon. Another attractive explanation might be that an S-glycoprotein may not not need RNase properties to trigger a UI reaction(se eals olater) .I twoul dtherefor e beessentia lt otes twhethe raccepto rplant stha tar eo fth e aaiigenotyp eremai naccepto rwhe ntransforme d witha construc t codingfo r an^-glycoprotei ntha t lacks RNase activity, like the modified Petunia inflataS3 glycoprotein that was described by McCubbin et al(1997) . 86 Chapter 5

Examples of S-glycoprotein s not being necessary for a functional UI reaction can be found within Solanum.The species S. brevidens (brd) is a SC species and may lack S-RNase activities in the style, for IEF of the stylar proteins of a brd clone used here showed a single basic protein at the position ofKl. Backcross progeny of atbr+brd fusion showed thisban d togetherwit h ^-glycoproteinbands , suggesting similar behaviour of brd and ver and thus Kl homology (unpublished results) and thus absence of^-glycoproteins . This species shows, apart from crossing barriers due to problems during seed development, also UI reactions when crossed with SI Solanum species (Pandey, 1962). Additionally, brd and its relative S. etuberosum (etb) show unilateral incompatibility with ver. Despite the barriers at the seed formation level, they could be crossed with ver as pistillate parent (Hermsen,1983), supporting the belief that the latter species is a more recent SC one.

Discrepancies The phenomenon that pollen tubes of the ver-type can be arrested by the ^-glycoproteins of tbr cannot be explained when a specific inhibition or transport of S-glycoproteins over the membrane occurs, as proposed for the SI reaction (McClure et al., 1989,1990; review by Kao and McCubbin (1997)), because ver pollen is not expected to possess those S-allele specificities, and should subsequently be indifferent to tbr S-RNases. Additionally, the SI reaction is dependent on the RNase activity of the ^-glycoprotein (Royo et al., 1994), which supports the hypothesis of transport over the membrane. Irrespective of which model (selective uptake over the membrane versus random uptake plus selective activation/inactivation of the S-glycoproteins: McClure et al., 1989; Clark et al., 1990; Kirch, 1993) is applied, the selective procedure causes a problem. So, when one of these models is basically correct, it must be modified or extended by introducing an additional function for ^-glycoproteins in UI systems, but outside the pollen tube. As already proposed, the expression of modified ^-glycoproteins (transgenes) or antisense suppression of the production of RNase- activity lacking S-glycoproteins as in the SC lines of L. peruvianum can give more insight in other mechanisms of causing pollen tube inhibition. If those /Wase-activity lacking ^-glycoproteins are capable indeed of causing an UI reaction, the possibility is opened that 5-glycoproteins cause a signal transduction as presumably takes place in the SI system of Papaver rhoeas (Franklin-Tong and Franklin, 1993) or in the sporophytic system of the Brassicaceae (e.g. Stein et al., 1991;Nasralla h et al., 1994), with one important difference: here S-allele non-matching products are triggering a reaction instead of the 5-allele matching products. This implies that within a species the compatible crosses, likeS1S2 x S2S3, the compatible non-matching of the style and the S3- pollen should not trigger a UI reaction. Here the products of the acceptor gene Ala and the inhibitor gene Himus t either block this signal transduction in case of species-own pollen or, when the membrane-bound ^-glycoprotein needs an additional factor, to enable the signal transduction in case of non-self-species pollen. From this it may be clear, that the contribution of the 5-locus to UI complicates some hypotheses and that both SI and UI might be more complicated than was expected. Chapter 6 General discussion

Diploidpotat o expresses aone-locu s gametophytic incompatibility (GSI) system. The diploid potato population derived from the clones G254, G609 and B16 (Olsder and Hermsen, 1976; Hermsen, 1978a,b,c )prove dt ob ea valuabl esourc efo rresearch ,an dfo r GSIresearc hi nparticular . The first two dihaploids werereporte d tob ederive d from the cultivar Gineke,th e dihaploid B16 was derived from a complex interspecific hybrid created by Black. Diallel crosses allowed for a classification ofS-alleles ,randoml y assignedSI toS4. Plantmateria l derived thereofwa s used for biochemical and molecular classification of factors involved in SI (e.g., Kirch et al. 1989; Thompson etal , 1991;Kirch , 1993;Wemmer, 1994;Peil , 1995). This basic material was used for thecreatio no fwel lperformin g 5-homo-an dhetero2ygotes ,an dth eplan tmateria l wasteste d onSI . Test crosses, iso-electric focusing of stylar extracts, Southern blotting and investigation of SC sourcesmad e clear however, that Ginekecoul dno tb eth e direct sourceo fth eprimar y dihaploids G254 and G609, but more likely an indirect one. This uncertainty about the ancestors of this material has, though, no consequences for thevalidit y ofth e SIresult s obtained on thismaterial . Thisi sno tonl ycorroborate db yth econsistenc y ofthes eresults ,bu tals ob y thereport s concerning SIi n other solanaceous species.

Obtainingreliable S- heterozygous andhomozygous self-incompatible tester clones. Itwa s argued in chapter 2tha tpollen-bom eP C canb euse d tocreat eS-homozygote stha tare , nevertheless,reliabl ei nthei rstyla rS Ireaction ,an dvic eversa .Th eS-homozygou steste rclone stha t were required for the SIresearc h described anduse d here (chapters 2-5),neede d tob ereliabl e in theirpollen-born e SIresponse .Th e counterfeit pollination method proved tob e avaluabl e tool in obtaining material with a strong SI response in pollen and style. Utilisation of strong pseudo- compatibility (PC)alread ypresen t in5-heterozygotes ,showe dt ob eunattractiv e forth e production of Si-reliable clones, because of the apparent heritability of PC, causing PC even in the S- homozygotes. The 5-homozygotes that were used as tester pollinators, showed neither pollen- expressed nor style-expressed pseudo-compatibility, andwer ethu swel l selected. Mosto fth eselecte d cloneswer eteste d forthei rtransformatio n efficiency. Although there was some variation in both regeneration ability (giving régénérants from a stem expiant) and trans­ formation ability( =givin gtransforme d cells,se eals oKharbotly , 1995;Kharbotl ye tal , 1995),littl e progressb ybreedin g fortransformatio n efficiency (=transgenic sobtaine dpe rtransforme d expiant) was anticipated. Therefore it was decided that this trait could be introduced by crossing with materialtha twa sgoo di nthi srespec t(se eAppendi x 1)an dvaluabl emateria lcoul db eselected .Th e selectedmateria lwa sreliabl ei nit sstyla rexpressio n ofSI ,bu tal lF l plantswer eS Cdu et o apolle n expressed SCfactor , liketh eon efoun d in G254 andB16 .Transgenic s obtained from these clones couldb eteste d for their stylar SI expression. Bybackcrossin g cloneswithou t this SC factor have 88 Chapter 6 beenobtained . Only alimite d number ofwel l performing cloneswit h areliabl e pollen- and style expressed SIreactio nhav ebee nselected , likeclon e6618-0 2(VI) . Thecontinuatio no fth ebreedin g andselectio nprocedures ,wit hth eai mo f combining allth e good factors ofth e SIpopulatio n with the transformation efficiency of the second population was the reason why a wide range of transformable cloneswa sused , instead of sticking to oneclone .

Expressionof thesense andanti-sense S-RNase basedconstructs. It isclea r from both the anti-sense experiments withSI andS2 tha t acomplet e suppression of thecorrespondin gS-glycoprotei nproductio n isno trequire dt oobtai na nefficien t break-down ofth e incompatibility reaction.A sligh treductio n ofth eS-RNase result si nsom epseudo-compatibilit y for pollen carryingtha tS-allele ,an da stron greductio nresult si n completecompatibility . This finding isno tsurprising ,knowin gtha tearl ypollinatio no fflower s from clonestha tar eslo wi nth ebuilding - upo fth e stylarglycoprotei n content frequently leadt osee d set.Thi swa s also found and described inchapte r2 .Th egain-of-functio n approach (introduction ofsense-S2 constructs) inpotat o showed that the strong S2 expressing clones displayed aver y strong SIreaction , in extreme cases barely allowingth epolle nt openetrat eth estigm amor etha nhal fa millimetre .Th ewea k expressors hardly inhibited thepolle n tubepenetration , thus confirming theresult s obtained by the loss-of function approach,bu tthe nfrom th eothe rsid eo fth eS Ireactio n spectrum.Analo gresult swer eobtaine db y gainan dlos so ffunctio n experimentswit h5-allel ebase d constructs inNicotiana alata (Murfet t et al., 1994) andPetunia inflata(Le e et al., 1994).Th e sense-inhibition approach inpotat o was not successful inthi srespec t thatth eendogenou s andhomologou s allelewa sno t silenced or otherwise rendered ineffective. It was in some cases, however, successful in causing a break-down of the incompatibilityreactio no fth enot-targete dnon-homologou s S-allele.I nthi scas eth eapparen t over- expression ofth etransgen ecause d somekin d ofdown-regulatio n ofth enon-homologou s allele.I t is also possible that the endogenous homologous allele was down regulated too, but this is impossiblet odemonstrat ebecaus ebot hth etransgen e construct pSK2QS2 an dth e endogenousS2 genecod efo r thesam eproduct .Thi sdow nregulatio n was also observed in sometransgenic s with thegenotyp eS3SJ0 +pSK2 QS2, knockin g downbot hS3 an dS10. I tmight , therefore, bepossibl e tofind i nnatur egenotype swit ha combinatio no fS-allele stha t showdominanc eo fon e^-allel eove r anotherb y somekin d ofdown-regulation , butther e aret o daten oreport s ofthi s yet. Thus,bot hth esens ean dth eanti-sens e experimentsconfir m thatth e5-glycoprotein sar eresponsibl e forth estyla r sideo fth e SIreaction , aswa s also earlier found inPetunia an dNicotiana (Le e et al, 1994;Murfet t et al, 1994).

Contribution ofthe S-locusto unilateral incompatibility/incongruity. Asalread y statedi nth echapter s4 and 5,i t isdebatabl ewhethe r theS-locu si sinvolve d inUI . Untilth erepor t ofMurfet t eta l(1996 ) showingproo f for 5-involvement,onl y strong indications, Generaldiscussion 89 favouring or opposing this involvement, were reported. The apparent absence of S-locus contributiont oUI ,reporte di nsom ecases ,le dt oth ebelie ftha tunilatera lincongruit ywa sth eprope r expression, whereas acontributio n supported theexpressio n unilateral incompatibility. Thewor ko fHermse nan dco-worker so nU Ibetwee ndiploi dpotat ospecie smad ealread y clear that the expression ofU Iwa s independent from thetbr S-alleles ,becaus e alltype s of5-heterozy - gotes andeve n someS-homozygote scoul db efoun d thatshowe deithe rU Io rth eopposit ereaction , acceptance. Also,th e absence orpresenc e oftSI -cause d SCcoul d not be linked to acceptance or non-acceptance for ver-, etb-o rftrcf-pollen. I nthes eexperiment s tbrgenotype s inwhic hS-RNase activitywa s lacking,wer eno tused , soth einfluenc e of absenceo f activetbr S-allele so nU I could notb etested . Ver, however, lacksS-RNase activity ,bu tnevertheles s crossing barriers not based on differences in EBN,wer e encountered (Abdalla,1970).Thi s supports the incongruity hypothesis, andals oth e SCmutan t found inLycopersicon peruvuianum, thatlacke dS-RNase activity, retained its UI reaction against L. esculentum.I n the latter case it can be argued that 5-glycoprotein was present, but it had lost its RNasepropert y due to apoin t mutation only,whic h still allowed for a non-RNase dependent UImechanism . In the case of ver this is unlikely, however, because no S- glycoproteins havebee nreporte d tob epresen t inthi sspecies . Theresearc ho fMurfet t eta l(1996) ,wh oworke d with gainan d lossof S-RNase expression in Nicotiana,showe d evidence for thecontributio n ofth e^-glycoprotein st oUI . Because UI related reactions were not present in all the interspecific crossings where they were anticipated, which showedtha tals oi nNicotiana mor egene so rallele so fU Ideterminin ggene sar einvolved . Thiswa s in agreement with our detection of different levels of UI breakdown after the loss ofS-RNase activityan dwit hth e existenceo fth epreviousl ypostulate d genesI andA. Wecoul d confirm here thatth eS-RNases cancontribut et oUI .I swa s alsoclea rtha t theU Ireactio n iscomplex , it knows several sites of activity in the style andha s different mechanisms. Itbecam e also cleartha t theS- locuso f veri srecognise d inth eU Ireactio nwhe nnon-accepto r tbrgenotype s arepollinate d with verpolle n orwit h pollen from tbr-verhybrids ,whic h isi n accordance with localisation of pollen determinedU Ilinke dfactor son -o ri nclos elinkag ewit hth eS-locu s(Chetela t andD eVerna , 1991). The pollen-part of the ver5-locu s is still active and this might be aprerequisit e for raising aU I reaction. Thusi twa s showntha tth eS-locus ,bot h styleand polle npart ,reall y cancontribut e to the UI reaction.

Concurrenceof hypotheses on UI Asstate di nchapte r 5,th emode l fornon-acceptanc e ofGra nan dAuberti n (1966)di dno t differ significantly from one of the models proposed by Hermsen et al (1974). However, the best explaining hypothesis ofHermse n andco-workers , introduced adominan t gene for acceptance in stead of arecessiv e one. Abdalla and Hermsen (1972) developed the two-powers-hypothesis, an evolutionary model as also discussed by several other authors(e.g . Lewis and Crowe, 1958),bu t 90 Chapter 6

in the two-powers-hypothesis the defence reaction of the SC species against most forms of introgression of genesfrom othe r species plays an important role.Th ebasi c assumption in these hypotheseswas ,tha t an SI speciesevolve s stepwise into an SCspecie s following therul e SI- > Sc - Sc' -> SC. Accordingt oAbdall a (1970),"th e UIgenes havedevelope d through the challenge of hybridisation between SI and SCpopulations "an d for ver"th e sensitive plasmons of SC species havedevelope d inth ecounteractin g competition subsequent tocrossin g of SCan d SIpopulations , particularly after the development of UIgenes" .Thus , finally an SCspecie swil lhav e a genotype like ScSc uiui and the SI species for instance S1S2 UM. Little was known then how SI can be converted into SC.Sc ca ntherefor e alsostan d fora nactiv e5-allel elinke d with a pollen-expressed SC-factor, as has been described in chapter 4. Style-expressed SC factors as described by Flaschenriem andAsche r(1979 )an dDan aan dAsche r(1986b )ca nthu sals ob ecounteracte db yth e developmento fUI genes .Al lthes ehypotheses ,includin gth eincongruit yhypothesi so fHogenboo m (1973) canb ebrough t in accordance with each other, when theS-locu si s subdivided into at least a pollen part and a stylepart , against which separate UIgene s canb e developed, which can have various alleles and dominance relationships. Thus,non-acceptanc e genes can also have dominant acceptance alleles or show intermediate reactions when heterozygous. The appearance of SI affecting genes, like minor genes, causing pseudo-compatibility, the pollen and style SC factors reported for Petuniab y Dana and Ascher (1986 a,b), tSl in tbrclon e G254,5Cverin verca n thus trigger the development of corresponding UIgenes . When these genes areno t located within the S-locus complex, the UI genes are then 5-locus dependent evolved, but potentially S-locus independent inexpression . These UIgene s arethe n truly incongruity genes.Thu s amixe d system withunilatera l incompatibility andunilatera l incongruity genesca nevolve .Th einhibito r geneI, as described in chapters 4 and 5,migh t be such anincongruit y gene.I t isno t known against which factor it is directed, soi tmigh t stillbelon g to theunilatera l incompatibility genesgroup .

Pollenexpressed SC-factors. Iti salread yargue dtha tther eexist sa variet yi npollen-expresse d SC(chapte r2) .Ther ei st odat e noproo ffo r SCdu et omutatio n ofth epolle n factor ofth e5-locus ,wherea sther ei sampl eevidenc e for SCcausin gmutation s inth e styleexpresse dS-gene .Lik ei nB1 6an dG254 ,whic hwer e SCdu e to the pollen-expressed factor called tSl, the well-transformable clone 1024-02 proved to homozygous for pollen-expressed SC. Iti stemptin g to assume allelism with tSl, especially after it was shown that tSl also causes SCwit hSI pollen. The offspring of 1024-02di d not show any strongpreferentia l penetrationo fSI, S2, S3,o rS10 pollen .Al lfou r^-homozygote swer efoun d after selfing of SC Fl offspring clones in more or less equal amounts. Hosaka and Hanneman (1998) detected skewed segregations of SCan d SIi n offspring from S.acaule an dS. phureja. Theirbes t fitting hypothesiswa stha tth epollen-expresse d SCfacto r Sli(S-locu sinhibitor ) acted sporophytic, and deviates in this respect from the expression pattern of tSl, which was a quite remarkable finding. It is,o n the other hand, like tSl, localised on chromosome 12(Hanneman , pers.comm) , Generaldiscussion 91 which still keeps open the possibility of allelism of SU and tSl. The appearance of two pollen- expressed SCallele s in 1024-2,unrelate d with G254 and B16, showstha t SC factors canb e more common than believed. In Petunia hybrida Dana and Ascher (1986a) detected a pollen-expressed SC factor at approximately the same distance from S asSC ver in the tbr-verhybrids . If thesegene s are allelic, this Petunia SC factor might show acceptance dependent penetration too, but nothing has been reported on this yet. The complex expression pattern of SCvetan d its dependence on the proper geneticbackgroun d ofth e seedparen t isexplaine d inth e chapters4 and 5.Th e stylar suppression ofpolle n expressed SCver, asdescribe d in the chapters 4 and 5,explain s why intercrossing oftw o self-incompatible potato speciesca nbrin g about self-compatibility. For example, species 1 canb e describedb y abasi cgenotype ,S1S2 AA ii scsc, whic hrepresent s apopulatio n of self-incompatible genotypes,accepto r for species2 .Specie s2 canthe nb eS3S4 aa IISCsc (and in somecase s scsc), whereaall inhibitsth epenetratio n of SC,bu t allows for thepenetratio n of scan dth e5-alleles ,a s well asfo r thepolle n of species 1.A nF l ofthes etw o species canthe nb e S1S3/S1S4/S2S3/S2S4, Aa, Ii, scsc/Scsc.Onl y theAa Ii Scscplant s will be SC.Thi sphenomeno n of sudden appearance of SCafte r intercrossing SIgenotype si sknow n amongpotat obreeder s working with interspecific hybrids (e.g.,i n complexphu-stn hybrids,Hermsen ,per s comm.).

Interactionbetween pollen andstyle. The S-RNase is transported overth e membrane of thepolle n tube and isbelieve d to cause an incompatibilityreactio nb ydegradin gRNA ,whic hfinall y resultsi npolle ntub earrest .A sexplaine d inchapte r 1,eithe r thetranspor t overth emembran e isallel e specific, orth euptak e is nonspecific followed by an inhibition of the non-self S-RNases. Analysis of DNA sequences and protein structures of S-RNasesgav e alreadymor e insight in conserved regions (C1-C5), (hyper)variable regions(V1-V5 )an dpresume d identity(V1-V5 ,excep tV3 )an dactivit y(C2 ,C3 )determinin gpart s of theS-RNases (e.g. . Ioerger et al., 1991; Tsai et al., 1992;Newbigi n et al, 1993;Simms , 1993; McCubbin et al., 1997).Mutatio n studieshav ebee nperforme d onth eregion s that were expected tob eresponsibl e forth e activity (e.g.Huan ge tal. , 1994)o rth eidentity .Th e activity was easily be disrupted by replacing a histidine residue in the activity parts (e.g. McCubbin et al., 1997) and interchanging hyper variable regions could disrupt the identity (Matton et al., 1997;Zure k etal. , 1997),bu t specific identity determiningregion swithi nth ehype rvariabl eregion s could to dateno t be identified (except, for at least, 4 aminoacid coding triplets). The mechanism of recognition between pollen and styleremaine d still unclear. The^SO-driven S2 sense constructs,reporte d inchapte r 3, showed in somecase s detectable expression in complete anther extracts. Nevertheless, the pollen appeared to stay completely functional. It was not clear whether the pollen produced this S-RNase,o r the surrounding anther tissue. Expression by thepolle n itself could theoretically have resulted in immediate pollen tube growth arrest orpolle n death.Kirc h eta l (1995) expressed already apotat o S-RNase inpolle n of Nicotiana tabacum. The pollen remained fully functional and capable of giving seed set. This 92 Chapter 6 showed that there is no cytotoxic effect of the S-RNaseo n the pollen tube, although the authors themselves camewit hpossibl e explanations whythi sconclusio n might beincorrect . An incorrect genetic background (wrong species) might havebee n such areason , or the lack of the necessary protein.Eve n absenceo fa prope rphosphorylation , directly after S-RNase entering thepolle n tube (as in a normal situation) might be a reason for this. A Ca2+ dependent phosphorylation is a possibility and putative mediating proteins have been detected inNicotiana alata b y Kunz et al (1997).Tha tphosphorylatio n andCa 2+pla y animportan t rolei nth epolle n tube growth and/or the incompatibility reaction appearst ob e likely. Itwa s shownb yL i et al(1994 )tha t Ca2+play s ake y role inth e growth and development ofth epolle ntub etip .I n chapter 1,i t ispointe d out that Ca2+ and phosphorylation play key roles in the SI systems of poppy and brassica, so this comes all togetherno t asa surprise . That acytotoxi c effect ina norma lincompatibl ecombinatio n canb erejected , isshow nb y the style grafting experiments of Lush and Clarke (1997). Incompatibility reactions were, at least partly, reversible and supports the ideatha t thepolle n tube actively synthesises RNA, which is degraded byth e self-type S-RNasewhe ntransporte d overth emembrane .

Discrepancybetween pollen recognition inSI and UI. As stated earlier in chapter 5, there is aproble m when explaining the contribution of the S- glycoproteins (S-RNases)t o both SI and UI. In the incompatibility reaction, there is a selective mechanism. Self-type pollen transports selectively theS-RNase overth emembran e or selectively inactivatesnon-sel fS-RNases. Inth eU Ireaction , however, non-selfS-RNases caus e an inhibition reaction. This justifies to consider an alternative, or additional mechanism for the interaction betweenpolle nan dstyle .Whe nth eS-RNases areno ttransporte d overth emembrane ,i tmigh tcaus e a signaltransductio n overth emembrane . Somehow theremus t be amechanis m that discriminates betweenspecies-sel f andnon-specie s selfcombination so fpolle ntub ean dS-RNases. Probablyher e the products of the acceptance genes and different acceptance alleles play arole . Species-self is always recognised by this product (from for instance, bothA and a) and should then disable or interrupt thesigna ltransductio n cascadecause dby , for instance,th ecombinatio n ofS2-RNase an d anSI pollentub eo fth e samespecies .Polle nfro m veri sthe n allowed topenetrat e any styleo f tbr, as long as it is of the genotypeA*ii. Absence ofth eproduc t from A allows now for the S-RNase induced signal transduction, finally resulting inpolle n tube arrest. There are several ways to test properties of this model. Style grafting asperforme d by Lush and Clarke (1997) is in the case of signaltransductio n unlikelyt ob eabl et ocaus ea reversio n ofth eU Ireactio n when analogous with theS Irespons e inpoppy . When theintroductio n ofa nS-glycoprotein , lacking theRNas e activity (such as described by McCubbin et al, 1997) in plants that express UI when transformed with correct S-RNases,i t is clear that identity is the determinant and not activity, thus allowing for a hypothesised additional interaction mechanism. Ake y factor inelucidatio n moreaspect s ofS Ian dU I stillremain sth eidentificatio n and cloning of pollen-S'-locus factors. Appendix 1: Selection ofwel l performing and welltransformabl e clones, useful for SI research. Somecharacteristic s and pedigree.

There are from various tobacco species accessions available that are highly efficient for transformation. Additionally, they areeasil y grown inth e greenhouse,wit h more generations per yeartha npotato .Becaus eS-allele sca nvar ywithi na specie seve nmor etha tsom eS-allele sbetwee n species do,i ti stemptin gt ous etobacc o for transformation experiments withpotato-S-allel ebase d constructs.Thi s approachwa s followed indeed (e.g.,Kirch , 1992;Kirc h et al., 1995;Ficke r et al., (1998),bu ti sno t applicable for allaspect so fS Iresearch .Interspecifi c crossingmigh tb enecessar y for testing the biological effect of a construct. Then not only factors like interspecific crossing barries (e.g.,unilatera l incompatibility) canpla y arole ,bu t also thedifferen t genetic background ofth ehos tca ncaus esubtl einteraction stha twer eno t anticipated (Kirche t al, 1995;Murfet t etal , 1995), changing or inactivating the transgene's effect. Thus, for some basic experiments the constructs shouldb eteste d within the species ofth eS-allel eorigin . Therefore, well transformable potato cloneswit h areliabl e SIreactio n were aprerequisite . Screening of the selected SI clones for ahig h efficiency of transformation ability byAgro- bacterium tumefaciens showedtha tther ewa s somevariatio ni nregeneratio n capacitys o that some transformants couldb eobtained ,bu tnon eo fth eclone sshowe dbot h agoo d transformation ability anda goo dregeneratio ncapacit y(unpublishe dresults) .Breedin g forthes ecombine dabilitie swoul d most likely, when using thebes t performing genotypes, have taken several crossing generations. Some well-transformable diploid potato clones were available that were not related to the SI material. In a combined effort (see also Kharbotly, 1995) crosses were made and progenies were screened for both transformation ability aswel l asothe r criteria asmentione d inchapte r 2 for the basic clones.A goo dperformanc e forvigour flowering,, mal e and female fertility and areliabl e SI reactionwer eprerequisites .Th eclone sha dt ocontai n atleas t an5 7o ra nS2 allel ei norde r tob eo f use for sense and anti-sense transformation experiments. Clone 1024-2,on e ofth ebes t transformable clones available,flowered lat ebu t abundantly. This clone was self-compatible, but it was expected to segregate for SC and SI clones in the Fl progenies. However, none ofth e 60teste d offspring clones was SI.Th e cloneswer e SC due toa pollen-expressed factor like the one found in G254 and B16, indicating that 1024-2 was homozygous for SC. Segregation patterns of the 5-alleles showed that 1024-2 contained S3 and anotherS-allele ,tha tprobabl y belongs to the SI-S3 family. The SIreactio n against S3polle n was reliable: none of the offspring genotypes originated from fertilisation by Si-pollen (Fig. 1).Th e stylar SIreactio n ofth eselecte d cloneswa s alsover yreliabl e (for SI andS2 i nth e648 6 and 6487 population respectively, for S3onl y onSI S3 o rS2S3), thu s eligeble asgoo d clones for testing of the stylarcontributio n ofSI .Th eselectio n ofwel l transformable cloneswa smor e labourious.Th e screeningmetho ddevelope db yKharbotl y(1995 )prove d tob ea powerfu l tool,bu tth e performance in SIresearc h formed atim ebottl enec ki nthis .Onl y few cloneswer e found that met all criteria to a satisfying extent. 94 Appendix1

Plant A16, an offspring of clone 1024-2,wa s superior in transformation efficiency and was veryvigorous ,bu tha dpoo r characteristics like,lat e flowering, functional male sterility, very poor intube rset ,easil ywilting ,wil dbranching ,an di tdi dno t expressth edesire dS-alleles .Mos t ofthes e characteristics weretransmitte d to itsprogeny . Thus,onl y few transgenic clonesou t of hundreds screened,wer e suitable for ourresearc h on SI.Wel lperforming , reliable self-incompatible clones wereunde rrepresented ,bu tfoun d ina lat estag eo fth eproject , whichexplain s alsowh yth epollen - mediated SC clones predominate in this thesis. It also explains why a less easily transformable clonelik e 195/5wa sused .Al l selected clones were,however , reliable inthei r stylar expression of SI. Only the transformed SI clones that were diploid (thus not spontaneous somatically doubled) couldb euse d directly to test the influence ofth e constructs onth emal e expressed SI reaction. The pedigree of the transformable clones is presented in table 1, as well as the S-allele composition of the clones or populations, as far as investigated. Most of the clones in the 6618- population suffered from alo wdegre eo fflowerin g andfertilit y problems. Population 1122ha d an excess of SC clones and a tendency to flower malformation. Many clones, though vigorous, suffered from alo wmal efertility . Population 1120segregate d forsom eusefu l clones,bu tth evigou r was somewhat reduced.Man yclone sha d atendenc y forpseudo-compatibilit y under unfavourable climatological conditions. Clone V is excellent in its transformation efficiency (an average of 80% in 5 weeks). It flowers middle late and fairly well,polle n fertility isexcellent . Itstuberisatio n is late and only acceptable for vegetativepropagation , whenraise d in early springo rgrow nunde r short day conditions. Clones R2 and R5 are slow in transformation: nine weeks are required for 50% transformation efficiency. Theplant s arevigorous ,flowe r early andprofusel y with excellent fertility. Tuberisation in small pots is early and good.

C Population 6487

Si

S10 S2 S3/ faint second S10 SKI SK2 Fig 1. Silver stained Iso Electric Focusing patterns of offspring of clone 1024-2. Most left: clone 1024-02. Left: population 92-6487 (535/0 xS1S2->S2S3 + 525/0); Right: 92-6486 (53570 x SlS3-> S1S3 +5/5/0). The 5/0 allele shows sometimes a faint second band, here at the position of 53. The penetration of only 5/ in population 6486 and only 52 inpopulatio n 6487 confirms that the band on the5 3 position has the S3 identity. Selection of well transformable clones 95

Table 1.Pedigre e of well transformable clones and some characteristics. See for clone numbers and references also chapters 2 and 3

Clone Mother Father S-alleles SC/SI Remarks

86-04-176 doubled monoploid, amf amf

87-10175-5 86-04-176 87.0008

87.0007 = SH82-62-247

87.0008 = SH82-70-297

87.1024-1 86-040-231 87.0007

87.1024-2 86-040-231 87.0007 S3S10 SC well transformable, Amf amf

87.1029-31 87.1017-5 87.1024-1

91-6222-40 G254 G609 S2S3 SI see chapter 2

91-6104-19 S2S3 S1S1, sc S1S3 SI see chapter 2

91-6167-2 Her-64 87.1029-31 S9S11

A16 91-6167-2 1024-2 S9S10 SC functionally male sterile

93-4002-3 91-6222-24 A16Ü! S2SJ0 SI

VI (93-6618-02) A16 91-6222-40 S3S10 SI

R2 (92-6486-4) 1024-2 91-6104-19 S1S3 SC R3 (92-6486-..) 1024-2 91-6104-19 S1S10 sc R5 (92-6486-19) 1024-2 91-6104-19 S1S3 sc V(92-6487-09) 1024-2 91-6222-40 S2S10 sc 94-1120-... R5 91-6105-06 SI/S3+S2/S4 SC/SI

94-1122-... R5 93-4002-03 S1/S3+S2/S10 SC/SI References

Abdalla MMF (1970) Inbreeding,heterosis , fertility, plasmon differentiation and Phytophthora resistance in Solanum verrucosum Schlechtd., and some interspecific crosses in Solanum. Thesis Wageningen Agricultural University. Agr. Res. Rep. 748.p p 213 Abdalla MMF (1974)Unilatera l incompatibility inplan t species:analysi san d implications. Egypt. J. Genet. Cytol. 3: 133-154 Abdalla MMF, Hermsen JGTh (1971a) A Two-loci system of gametophytic incompatibility in Solanum phureja and S.stenotomum. Euphytica 20: 345-350 Abdalla MMF,Hermse n JGTh (1971b)Th eplasmon-geni cbasi so fpolle n lobedness andtetra d in Solanum verrucosum hybrids and duplicate linkage groups. Genetica 42, 261-270 Abdalla MMF,Hermse n JGTh (1972)Unilatera l incompatibility: hypotheses, debate andit s implications for plant breeding. Euphytica 21, 32-47 Anderson E,d eWinte r D (1931)Th egeneti c analysis ofa nunusua lrelationshi p between self-sterility and self-fertility inNicotiana. Arm. Mo. Botan. Gdn. 18,97-11 6 Anonymus. Brochure: De Victoria bloeit weer in de hortus. Stichting Uitgeverij De Oude Stad, i.s.m. Vereniging Vrienden van deHortus.Amsterdam . Bateman AJ (1955) Self-incompatibility systemsi n angiosperms. III. Cruciferae.Heredity 9: 52-68 Bevan M (1984)Binar yAgrobacterium vector sfo rplan ttransformation . NucleicAcid sRes . 12:8711-872 1 Breukelen EWM van, Ramanna MS,Hermse n JGTh (1977) Parthenogenetic monohaploids (2n=x=12)from Solanum tuberosum L. and S. verrucosumSchlechtd . and the production of homozygous potato diploids. Euphytica 26:263-27 1 Brown CR, Adiwilaga KD (1991) Use of rescue pollination to make a complex interspecific cross in potato. AmericanPotat oJourna l 68:(12) 813-820 Charlesworth D(1988 )Evolutio n ofhomomorphi c sporophytic self-incompatibility. Heredity 60:446-45 3 Chetelat,RT ,DeVern aJ W(1991 )Expressio no funilatera l incompatibility inpolle no fLycopersicon pennellii isdetermine db ymajo r loci onchromosome s 1,6an d 10. TheorApp lGene t 82(6) :704-712 . Clark KR, Okuley JJ, CollinsPD , SimsT L (1990) Sequencevariabilit y anddevelopmenta l expression of S- allelesi nself-incompatibl e andpseudocompatibl ePetunia. ThePlan tCel l2 : 815-826 ClarkeA ,Newbigi n E(1993 )Molecula raspect so fself-incompatibilit y inflowerin g plants.Ann uRev .Genet . 27: 257-279 Clulow-S-A;Wilkinson-M-J ;Waugh-R ;Baird-E ;Demaine-M-J ;Powell-W ; 1991. Cytological andmolecula r observations on Solanumphurej a induced dihaploidpotatoes .Theo rApp lGene t 82(5):545-55 1 ConnerJA ,Tantikanjan a T, SteinJC ,Kandasam y MK,Nasralla hJB ,Nasralla hM E(1997 ) Transgene-induced silencing of5-locu sgene san drelate d genesi nBrassica. ThePlan tJourna l 11(4) :809-82 3 ComishEC ,Pettit tJM ,Boni gI ,Clark eA R(1987 ) Developmentally controlled expression ofa gen e associated with self-incompatibility inMcofen a alata. Nature 326:99-10 2 CraneMB ,Lawrenc eWJ C(1929 )Genetica lan dcytologica l aspectso fincompatibilit y andsterilit yi n cultivated fruits. J.Pomol. Hort. Sei.7 ,276-301 . DanaMN ,Asche rP D(1985 )Discriminatin g styles(DS )a pollen-mediate d pseudo-self compatibility (PMPSC) inPertunia hybrida Hort.Euphytic a 35:237-244 . Dana MN, Ascher PD (1986a) Sexually localized expression ofpseudo-sel f compatibility (PSC) in Petunia x hybrida Hort. 1. Pollen inactivation.Theo rApp lGene t 71: 573-577. 98 References

Dana MN, Ascher PD(1986b ) Sexually localized expression ofpseudo-sel f compatibility (PSC) inPetunia x hybrida Hort. 1.Styla rinactivation .Theo rApp lGene t 71: 577-584. DeNettancour t D. Seeals oNettancour tD de .

DeNettancourt D (1997)Incompatibilit y inAngiosperms . SexPlan tRepro d 10: 185-199 EckH Jvan ,Jacob sJME ,Ber gPM Mva nden ,Stiekem aWJ ,Jacobse nE (1994 )Th einheritanc e of anthocyanin pigmentationi npotat o(Solanum tuberosumL )an dmappin go ftube rski ncolou rloc iusin gRFLPs .Heredit y73 : 410-421 EldikG Jva n(1996 )Gene sexpresse di npollinate dpistil sof Solanum tuberosum.Thesis ,Katholiek e Universiteit Nijmegen,p p 155 El-KharbotlyA (1995 )Transposo n Tagging:Toward sth eisolatio n ofth eresistanc e R-Genesi npotat o against Phytophthora infestans (Mont.)d eBary .Thesi sWageninge nAgricultura lUniversit yNL .p p 99.ISB N90-5485 - 454-5 El-KharbotlyA ,Jacobse nE ,Stiekem aWJ ,Pereir aA (1995 ) Genetic localisation oftransformatio n competence indiploi dpotato .Theor .Appl .Genet . 91:557-562 . El-Kharbotly A,Jacob sJME ,Linte l HekkertB te ,Jacobse n E,Ramann a MS, StiekemaWJ ,Pereir a A(1996 ) Localization of ßs-transposon containing T-DNA inserts in dihaploid transgenic potato: linkage to theRJ resistance geneagains tPhytophthora infestans (Mont.) deBary .Genom e 39:249-257 . Eijlander R,Ramann aMS ,Jacobse nE (1997 )Selectio no fvigorou san dfertil e S-homo-an dheterozygou s tester clones from self-incompatible diploidpotato ,Solanum tuberosum L.Euph .97 :97-11 1 Ficker M,Wemme rT , ThompsonR D(1998 )A promote rdirectin ghig hleve lexpressio n inpistil so f transgenic plants(i.p) . Ficker M, Kirch H-H, Eijlander R, Jacobsen E, Thompson RD (1998) Multiple elements of theS2-RNase promoterfro m potato(Solanum tuberosumL. )ar erequire dfo r celltype-specifi c expression intransgeni cpotat o and tobacco.Mo l GenGene t257 : 132-142 Flaschenriem DR, Ascher PD (1979) S Allele Discrimination in Styles of Petuniahybrida Bearin g Stylar- conditioned Pseudo-self-compatibility. Theor.Appl .Genet . 55:23-28 FlipseE ,Huisma nJG ,Vrie sB Jde ,Bergervoe tJEM ,Jacobse nE ,Visse r RGF(1994 )Expressio n ofa wild-typ e GBSS-geneintroduce d intoa namylose-free potat omutan tb yAgrobacterium tumefaciens andth e inheritance of theinsert s onmicrospori c level.Theor .Appl .Genet . 88:369-37 5 Foolad-MR(1996 )Unilatera lincompatibilit ya sa majo r causeo fskewe dsegregatio ni nth ecros sbetwee nLyco- persiconesculentum andL. pennellii. Plant-Cell-Reports., 15(8):627-63 3. FooteH ,Rid eJ , franklin-Tong VE,Walke r E,Lawrenc eMJ ,Frankli nFC H(1994 )Clonin g andexpressio no f a distinctive classo f self-incompatibility (S)gen e from Papaverrhoeas L.Pro cNat l Acad SeiUS A 91:2265 - 2269 Franklin FCH,Lawrenc e MJ,Franklin-Ton g VE(1995 ) Cell andmolecula rbiolog y ofself-incompatibilit y in flowering plants.IntRevCyto l 158: 1-64 Franklin-TongN ,Frankli nC (1993 ) Gametophytic self-incompatibility: Contrasting mechanisms forNicotiana ma Papaver. Trends inCel lBiolog y 3(10):340-34 5 Franklin-Tong VE,Ruut h E,Marmer yP ,Lawrenc e MJ,Frankli n FCH (1989) Characterization of astigmati c component ofPapaver rhoeas L.whic h exhibitsth e specific activity ofa selfincompatibilit y (S)gen eproduct . NewPhytol 112:307-31 5 References 99

Franklin-TongVE ,Rid e JP,Rea dN ,Trewava s AJ,Frankli nFC H (1993)Th e self-incompatibility responsei n Papaverrhoeas ismediate db y cytosolicfree calcium . ThePlan tJourna l 4: 163-177 Franklin-TongVE ,Rid eJP , FranklinFC H(1995 )Recombina t stigmatic self-incompatibility (S-)protei n elicits aCa 2+transien t inpolle no f Papaverrhoeas. ThePlan tJourna l 8:299-30 7 Franklin-Tong VE, Drebak BK, Alan AC, Watkins PAC, Trewavas JA (1996) Growth of Pollen Tubes of Papaverrhoeas I sRegulatee db ya Slow-Movin gCalciu mWav ePropagate db yInosito l 1,4,5-Trophosphate. The Plant Cell,8 : 1305-1321 Franklin-TongVE ,Hacket tG ,Heple rP K(1997 )Ratio-imagin go fCa 2+,inth eself-incompatibilit y responsei n pollentube sof Papaver rhoeas. ThePlan tJourna l 12(6) : 1375-1386

Gastel AJG van (1976) Mutability of the self-incompatibility locus and identification of the S-bearing chromosome inNicotiana alata. Thesis.Agricultural-Research-Reports. , No. 852,89pp . Gastel AJG van, Nettancourt D de (1975) The generation of new incompatibility alleles. Incompatibility Newsletter 6:66-6 9 Gastel AJG van; Carluccio F (1975) Identification of the S-bearing chromosome in Nicotiana alata. Incompatibility Newsletter. 6: 87-90 Grun P,Auberti n M (1966)Th e inheritance andexpressio n ofunilatera l incompatibility inSolanum. Heredity 212: 131-138 Heeres P, Jacobsen E,Visse r RGF (1998) Behaviour of genetically modified amylose free potato clones as progenitors ina breedin gprogra m Euphytica 98(3 ) :169-17 5 HennyRJ , AscherP D(1976 )Th einheritanc e ofpseudo-self-compatibilit y (PSC) inNemesia strumosa Benth. TheorAppl Genet48 : 185-195. Hermsen,JGTh . (1977)Genera l Considerations OnInterspecifi c Hybridization. Interspecific Hybridization In PlantBreeding .Proc . 8thEUCARPI A Congress,Madrid , Spain.P p299-304 . Hermsen JGTh (1978a) Genetics ofself-incompatibilit y indihaploid s oîSolanum tuberosumL .2 . Detection and identification of allpossibl e incompatibility and compatibility genotypes in sixFl' s from interdihaploid crosses. Euphytica 27: 1-11. Hermsen JGTh (1978b) Genetics of self-incompatibility indihaploid s oîSolanum tuberosum L.3 .Lethalit y ofS-bearin gtranslocatio n homozygotes. Euphytica 27: 13-17 HermsenJGT h (1978c). Genetics of self-incompatibility in dihaploids oîSolanum tuberosum L.4 . Linkage between anS-bearin gtranslocatio n and alocu s for virescens.Euphytica 27: 381-384. Hermsen JGTh (1979) Factors controlling interspecific crossability and their bearing on the strategy for breakingbarrier st ointercrossin go ftuber-bearin gSolanum species.Proc .Conf . BroadeningGenet .Bas eCrops , Wageningen (1978) Pudoc,WageningenPp311-31 8 Hermsen JGTh (1983)Utilizatio n ofWid eCrosse s inPotat oBreeding .In :Presen t andFutur e Strategies for PotatoBreedin g and Improvement. Perport ofth eXXV Iplannin g conference. CEP,Lima ,Peru ,p p 115-132. Hermsen JGTh, Sawicka E (1979) Incompatibility and incongruity in tuber-bearing Solanum species.In: Hawkes JG,Leste rNR , Skelding AD(eds )Th ebiolog y andtaxonom y ofth eSolanaceae. Linnaean Society ofLondo n Academic 445-453 HermsenJGTh ,Verdeniu sJ (1973 )Selectio nfrom Solanum tuberosum groupphureja o fgenotype scombinin g high-frequency haploid induction withhomozygosit y for embryo-spot. Euphytica 22:244-25 9 Hermsen JGTh,Ramann a MS,Voge lJ (1973 )Th elocalisatio n ofa recessiv e gene for chlorophyll deficiency indiploi d Solanumtuberosum bymean s oftrisomi e analysis.Can .J . Genet. Cytol. 15: 807-813 100 References

Hermsen JGTh, Olsder J,Hovin g E,Janse n P (1974) Acceptance ofself-compatibl e pollen from Solanum verrucosum indihaploid s from S.tuberosum. In:H.F.Linsken s (Ed.),Fertilizatio n inhighe rplants .P p 37-40. Hermsen, JGTh, Govaert I, Hoekstra S,va nLoo n C,Neefje s C (1977) Analysis of the effect of parental genotypes oncrossabilit y ofdiploi d Solanumtuberosum with S.verrucosum. Agene-for-gen e relationship? Proc. 8thEucarpi a Congress,Madrid , Spain.P p 305-312 Hermsen JGTh, Sonneveld J,Vreugdenhi l D,Ballegooije n Dvan , Staas-Ebregt E (1978a) Genetics ofself - compatibility indihaploid s ofSolanum tuberosum L. 5.th eS-genotyp eo ftetraploi d potato cultivar Gineke. Euphytica27 : 385-388 HermsenJGTh ,Taylo rLM ,Breukele n EWMvan ,Lipsk iA (1978b )Inheritanc e ofgeneti cmarker s from two potato dihaploids andthei rrespectiv eparen t cultivars.Euphytic a 27: 681-688 Heslop-Harrison Y,Heslop-Harriso n J, Shivanna KR(1981 ) Heterostyly inPrimula. 1.Fine-structura lan d cytochemical features ofth e stigma and style inPrimula vulgarisHuds .Protoplasm a 107: 171-187 Hiscock SJ,Dickinso nH G(1993 )Unilatera lincompatibilit ywithi nth eBrassicaceae : Further evidence forth e involvement ofth e self-incompatibility (S)-locus.Theor .Appl .Genet . 86(6):744-75 3 Hoekema A, Hirsch PR,Hooykaa s PJJ, Schilperoort RA (1983) Abinar y plant vector strategy based on separation of Vir-an dT-regio no fth eAgrobacterium tumefaciens Ti-plasmid.Nature 303 ,179-180 .

HogenboomN G(1973 )A mode l forincongruit y inintimat epartne rrelationships .Euphytic a 22: 219-233 Höfgen R,Willmitze rL (1988).Storag e ofcompeten t cellsfo rAgrobacterium transformation . NuclAcid s Res 16:987 7 Huang S,Le eH-S ,Karunananda a B,Ka oT- H(1994 ) Ribonuclease activity ofPetunia inflata S protein si s essential forrejectio n of self-pollen. Plant Cell 6(7): 1021-1028 IkedaS ,Nasralla hJB ,Dixi tR ,Preis sS ,Nasralla hM E(1997 )A naquaporin-lik egen erequire d forth eBrassica self-incompatibility response. Science 276: 1564-1566 IoergerT R ClarkAG ,Ka oT- H (1990)Polymorphis m atth eself-incompatibilit y locusi nSolanacea epredate s speciation. Proc.Natl .Acad . Sei.US AGenetics . 87:9732-973 5 IoergerTR ,Gohlk eJR ,X uB ,Ka oT- H(1991 )Primar y structural features ofth e self-incompatibility protein in Solanaceae. SexPlan trepro d 4: 81-8 7 Jacobsen E, Hovenkamp-Hermelink JHM,Krijgshe d HT,Nijda m H,Pijnacke r LP,Withol t B,Feenstr a WJ (1989).Phenotypi can dgenotypi ccharacterizatio no famylose-fre e starchmutan to fpotato .Euphytic a44:43-4 8 Iwanaga M, Freyre R Watanabe K (1991) Breaking the crossability barriers between disomic tetraploid Solanumacaule an dtetrasomi c tetraploidS. tuberosum. Euphytica 52:(3 ) 183-191 Kao T-H, McCubbin AG(1997 ) Molecular and biochemical bases of gametophytic self-incompatibility in Solanaceae. PlantPhysiol .Bioche m 35 (3): 171-176 Kaufmann H, Salamini F, Thompson RD (1991) Sequence variability and gene structure at the self- incompatibility locuso f Solanumtuberosum. Mol.Gen .Genet .226,457-46 6 Kaufmann H, Kirch H, Wemmer T, Peil A, Lottspeich F, Uhrig H, Salamini F, Thompson R (1992) Sporophytic andgametophyti c self-incompatibility. In:Crest iM ,Tiezz iA (eds) .Sex .PI . Repr.Springer-Verlag Berlin. 115-126 Kharbotly AEl : seeals o El-Kharbotly Kirch H-H (1993) Strukturelle undfunctionell e Analyse derRegulatio n zweier S-RNaseGen e aus Solanum tuberosumL .i ntransgene n Pflantzen. Dissertation, Universität Köln (Thesis,Universit y of Cologne) References 101

KirchH-H ,Uhri gH ,Lottspeic hF ,Salamin iF ,Thompso nR D(1989 ) Characterization ofprotein s associated with self-incompatibility in Solanumtuberosum. Theor. Appl. Genet. 78:581-58 8 KirchH-H ,L iYQ ,Seu lU ,Thompso n RD(1995) Theexpressio n ofa potat o(Solanum tuberosum)S-RNas e genei nNicotiana tabacumpollen . Sex.Plan tReprod . 8(2):77-84 . Kowyama Y, Kunz C, Lewis I, Newbigin E, Clarke AE, Anderson MA (1994) Self-compatibility in a Lycopersiconperuvianum variant (LA2157)i sassociate d with alac ko fstyl e S-RNaseactivity .Theo r Appl Genet 88(6-7): 859-864 Kubicki,B (1969) .Investigation s on sexdeterminatio n incucumbe r (Cucumissativus L. ) Multiple alleles of locus Acr. Genetica Polonica 10: 23-68. Kuipers AGJ, Soppe WJJ, Jacobsen E, Visser RGF (1994) Field evaluation of transgenic potato plants expressing an antisense granule-bound starch synthase gene: Increase of the antisense effect during tuber growth.Plan tMolecula r Biology 26(6): 1759-1773 KunzC ,Chan gA ,Faur eJD ,Clark eAE ,Poly aGM ,Anderso nM A(1996) .Phosphorylatio n ofstyl eS-RNase s byCa 2+-dependen tprotei n kinases from pollen tubes.Se xPlan t Reprod.9:25-3 4 Kyhse-Anderson J (1984) Electroblottingo fmultipl e gels:A simple apparatuswithou tbuffe r tank for rapid transfer ofprotein s from Polyacrylamide tonitrocellulose .J .Biochem .Biophys .Method s 10,203-20 9 LawrenceM J(1975 )Th egenetic so fself-incompatibilit y inPapaver rhoeas. Proc Roy SocLon d [Biol] 188: 275-285 Lawrence MJ,Afaza l M,Kendric kJ (1978)Th egenetica l control of self-incompatibility inPapaver rhoeas L.Heredit y 40: 239-285 Lee H-S, Singh A, Kao T-H (1992) RNase X2 A pistil-specific ribonuclease from Petunia inflata, shares sequence similarity withsolanaceaous Sproteins.Plan tMo lBio l 20: 1131-1141 Lee H-S,Huan g S, Kao T-H (1994) S-proteinscontro l rejection of incompatible pollen inPetunia inflata. Nature 367: 560-563. LewisD (1943) Physiology ofself-incompatibility . Journal ofGenetic s 45: 171-185 LewisD (1961) Chromosomefragments an d mutations ofth e incompatibility gene.Natur e (190:990-991. LewisD ,Crow e LK (1958)Unilatera l incompatibility in flowering Plants. Heredity 12:233-25 6 LiX ,Nied lJ ,Hayma nD ,Langridg eP (1994 )Clonin ga putativ e self-incompatibility genefrom th epolle n of the grassPhalaris coerulescens. PlantCel l6:1932-193 2 LiYQ ,Faler i C,Thompso n RD,Tiezz i A,Eijlande r R, Cresti M (1994)Cytochemica l immunolocalization ofth eabundan tpisti lprotei n SK2i npotat o (Solanum tuberosum). Sex.Plan t Reprod. 7(3): 164-168 Liedl, BE, Anderson NO (1994) Statistical differentiation between self incompatibility and pseudo-self compatibility inPertunia hybrida Hort .Usin g female andmal e coefficient of crossability. SexPlan t Reprod 7:229-238 Litzow ME, Ascher PD (1983)Th e inheritance ofpseudo-sel f compatibility (PSC) inRaphanaus sativus L. Euphytica32:9-15 Lundqvist A(1956 ) Self-incompatibility inry eI .Geneti c control ofth ediploid .Heredita s 42: 239-348 Lundqvist A (1990) Variability within and among populations in the 4-gene system for control of self- incompatibility inRanunculus polyanthemos. Heredita s 74: 161-168 Lundqvist A (1991) Four-locus 5-gene control of self-incompatibility made probable in Lilium martagon. Hereditas 114:57-6 3 102 References

Luu D-T, Heizmann P, Dumas C, Trick M, Cappadocia M (1997). Involvement of SLR1 genes in pollen adhesion toth e stigmatic surface inBrassicaceae . SexPlan t Reprod 10:227-23 5 MartinF W(1959 )Stainin gan dobservatin gpolle ntube si nstyl eb ymean so ffluorescence . StainTechnol .34 : 125-128. Mather, K. (1943)Polygeni c inheritance andnatura l selection.Biologica l Reviews 18:32-64. MattonDP ,Mae sO ,Laubli nG ,Xik eQ ,Bertran d C,Mors eD ,Cappadoci a M(1997) . Hypervariable Domains of Self-Incompatibility RNasea Mediate Allele-Specific Pollen Recognition. ThePlan t Cell 9: 1757-1766 Matzke MA,Primi gM , Trnovsky J,Matzk e AJM (1989) Revesible methylation andinactivatio n ofmarke r genes in sequentially transformed tobaccoplants .EmboJourna l 8: 643-649 McClure BA (1996) S RNase and interspecific pollen rejection in the genus Nicotiana: multiple pollen-rejection pathways contribute to unilateral incompatibility between self-incompatible and self- compatible species.Plant-Cell. , 8(6):943-95 8 McClure BA, Haring V, Ebert PR, Anderson MA, Simpson RJ, Sakiyama F, Clarke AE (1989) Style self- incompatibility products ofNicotiana alata ar eribonucleases. Natur e 323: 955-957 McClure BA, Gray JE,Anderso n MA, Clarke AE (1990) Self-incompatibility inNicotian a alata involves degradation ofpolle nrRNA .Nature ,347 :757-76 0 McCubbinAG ,Chun gY-Y ,Ka oT (1997)A Mutan tS3 Rnas e ofPetunia inflata lacking Rnase Activity has anallele-specifi c dominantnegativ e effect on self-incompatibility interactions. MurashigeT , SkoogF (1962 )A revise dmediu m forrapi d growth andbioassay s withtobacc o tissue cultures. Physiol. Plant 15,473-497. Murfett J,Atherto nTL ,Mo uB ,Gasse rCS , McGur eB A(1994 )S-RNase expresse d intransgeni cNicotiona causes^-allele-specifi c pollen rejection. Nature 367:563-56 6 Murfett J,Bourqu eJE ,M cClur eB A(1995 )Antisens e suppression of S-RNase expression inNicotiona usin g RNApolymeras e II-an d Ill-transcribed gene constructs.Plan tMol .Biol .29 :201-21 2 Murfett J, StrabalaTJ ,Zure kDM ,Mo u B,Beeche r B, McClure BA (1996) SRNas e and interspecific pollen rejection inth e genusNicotiana: multiple pollen-rejection pathways contribute tounilatera l incompatibility between self-incompatible and self-compatible species.Plant-Cell. , 8(6):943-95 8 Mutschler,MA ,Lied lB A(1994 )Interspecifi c crossingbarrier si nLycopersicon andthei r relationship toself - incompatibility. In:E.G.Williams ,A.E .Clarke ,R.B .Kno x (Eds),Geneti c control of self-incompatibility and reproductive development in flowering plants.P p 164-188. Nasrallah JB (1997) Evolution of the Brassica self-incompatibility locus: A look into S-locus gene polymorphisms. Proc.Nat l Acad. SeiUS A 94: 9516-9519 Nasrallah JB, Nasrallah ME (1984) Electrophoretic heterogeneity exhibited by the 5-allele specific glycoproteins ofBrassica oleracea. Experientia 40: 279-281 Nasrallah JB, Kao T-H, Goldberg ML, Nasrallah ME (1985) A cDNA clone encoding an S-specific glycoprotein from Brassicaoleracea. Nature 318:263-267 NasrallahJB ,Ka oT-H,Che nC-H ,Goldber gML ,Nasralla hM E(1987 )Amin oaci dsequence so fglycoprotein s encoded bythre e alleleso fth eS-locu sof Brassica oleracea. Nature 326:617-61 9 Nasrallah JB, Stein JC, Kandasamy MK, Nasrallah ME (1994) Signaling the Arrest of Pollen Tube Development in Self-Incompatible Plants. Science 266: 1505-1508 Nettancourt, D.d e(1977 )Incompatibilit y inAngiosperms . In: Frankel R, Gall GAE, Grossman M, Linskens References 103

HF,d eZeeu wD(Eds) . Monographs onTheoretica l andApplie dGenetics ,Vol .3 Springer-Verlag ,Berlin ,p p 230 Newbigin E,Anderso nMA ,Clark eA E(1993) .Gametophyti c self-incompatibility systems.Plan t Cell 5(10): 1315-1324 Nicholls MS (1987) Pollen flow, self-pollination and gender specialisation: factors affecting seed-set inth e tristylous speciesLythrum salicaria (Lythraceae) .Plan t Systematicsan d Evolution 156: 151-157 Olsder J, Hermsen JGTh (1976) Genetics of self-compatibility in dihaploids of Solanumtuberosum L . I. Breedingbehaviou r oftw o self-compatible dihaploids. Euphytica 25: 597-607 Pandey KK (1962)Interspecifi c incompatibility inSolatium species.Am.J.Bot .49:874-882 . Pandey KK (1963) Stigmatic secretion and bud-pollinations in self- and cross-incompatible plants. Naturwissenschaften 50: 408-409. PandeyK K(1970 )Tim ean dsit eo fo fth eS-gen eaction ,breedin gsystem san drelationship si nincompatibility . Euphytica 19:364-37 2 Pandey KK (1981) Evolution ofunilatera l incompatibility inflowerin g plants: further evidence in favour of twin specificities controlling intra-an dinterspecifi c incompatibility. TheNe wPhytologis t 89:705-728 . PeilA (1995 )Identifizierun g undKlassifizierun g von S-Alleleni nSolanum undvergleichend e Analysen von Solanaceae-S-Mlelen. Inaugural-Dissertation zurErlangun gde sGrade sDokto rde rAgrarwissenschafte n (Dr. agr.)de rhohe nLandwirtschaftliche n Fakultät derRheinische n Friedrich-Wilhelms-Universitätz uBonn .pp . 147. RamuluK Sree(1982 )Geneti cinstabilit y atth eS-locu so fLycopersicon peruvianum plantsregenerate d from invitroculture ofanthers :generatio no fne wS-specificitie san d5-allel ereversions .Heredit y4 9 (3):319-330 . Pochards,A J (1986)Plan t Breeding Systems.Georg eAlle nan dUnwin , London. Rick CM (1986) Reproductive Isolation in theLycopersicon peruvianum Complex. In: WG D'Arcy (Ed.), Solanaceae: Biology and Systematics.P p477-495 . Riley HP (1932) Self-sterility in Shepherds' purse.Genetic s 17:231-29 5 Rowlands DG (1964) Self-incompatibility in sexuallypropagate d cultivated plants. Euphytica 13,157-162 . RoyoJ ,Kun z C,Kowyam aY ,Anderso nM ,Clark eAE ,Newbigi nE (1994) .Los so f ahistidin eresidu e atth e activesit eo f S-locusribonuclease i sassociate dwit hself-compatibilit y inLycopersicon peruvianum.Troc. Nat Acad. SeiUSA , 91:6511-6514 Sambrook J, Fritsch EF,Maniati s T (1989) Molecular Cloning. A laboratory manual. 2nd ed.. Cold Spring Harbour Laboratory Press.Ne w York. Schmidt-Stohn G (1979) Eine Methode zur isoelektrischen Fokussierung von Proteinen in Flachgelen für Serienuntersuchungen inde r Genetik.Z .Pflanzenzüchtun g 83, 81-90. SharmaKD ,Boye s JW (1961) Modified incompatibility ofBuckwhea t following irradiation. Can. J. Botan. 39: 1241 Shivanna KR, Rangaswamy N (1969) Overcoming self-incompatibility in Petunia axillaris. 1. Delayed pollination, pollination with storedpollen , andbu dpollination . Phytomorphology (19,372-38 0 Shivanna KR, Heslop-Harrison J, Heslop-Harrison Y (1981) Heterostyly in Primula. 2. Sites of pollen inhibition,an deffect s ofpisti lconstituent s oncompatibl e andincompatibl e pollen-tube growth. Protoplasma 107: 319-337 SimsT L(1993 )Geneti cRegulatio n ofSelf-Incompatibility . CriticalReview s inPlan t Sciences 12(1/2) :129 - 167 104 References

SingA ,Ka oT- H (1992) Gametophyticself-incompatibility . Int.Rev .Cyt . 140:449-48 3 SingsitC , Hanneman REJr . (1991)Rescuin gabortiv e inter-EBN potatohybrid s through double pollination andembry o culture.Plan t CellReport s 9:(9 )475-47 8 StamM ,Mo lJNM ,Koote rJ M(1997) . TheSilenc eo fGene si nTransgeni cPlants .Annal s ofBotan y 79: 3-12 SteinJC ,Howlet tB ,Boye sDC ,Nasralla hME ,Nasralla h JB (199 1) Molecula r cloningo f aputativ e receptor proteinkinas egen eencode d atth eself-incompatibilit y locusof Brassica oleracea. ProcNat l Acad SeiUS A 88: 8816-8820 Stein JC,Dixi t R, Nasrallah ME, Nasrallah JB (1996) SRK, the stigma-specific S locus receptor kinase of Brassica,i stargete dt oth eplasm a membrane intransgeni c tobacco.Plan t Cell 8:429-44 5 StephensonAG ,Dought yJ ,Dixo n S,Ellema n C,Hiscoc k S,Dickinso n HG(1997 )Th emal e determinant of self-incompatibility inBrassica oleracea islocate d inth epolle n coating.Th eplan t Journal 12(6) :1351-135 9 Suzuki G, Watanaba M, Kai N, Matsuda N, Toriyama K, Takayama S, Isogai A, Hinata K (1997). Three members ofth eS mutagen e family arelinke dt oth eS locuso fBrassica. Mo l gen Genet 256: 257-264 TakahasiH (1973 )Genetica lan dphysiologica l analysiso fpseudo-self-compatibilit y inPetunia hybrida.Jap n J Genet 48: 27-33 TanLW ,Jackso nJ F(1988 )Stigm aprotein so fth etw oloc iself-incompatibl e grassPhalaris coerulescens. Sex PlantRepro d 1:25-27 Tantikanjana T,Nasralla hME ,Stei nJC ,Che nC-H,Nasralla hJ B(1993 )A nalternativ etranscrip to fth e5-locu s glycoprotein geni na clas sI Ipollen-recessiv e self-incompatibility haplotype ofBrassica oleracea encode sa membrane anchoredprotein .Plan t Cell 5:657-66 6 ThompsonRD ,Uhri gH ,Hermse nJGTh , Salamini F,Kaufman n H(1991 )Investigatio n ofa self-compatible mutationi nSolanum tuberosumclone sinhibitin g S-alleleactivit yi npolle ndifferentially . MolGe nGene t226 : 283-288 Thompson RD, Kirch H-H, Ficjer M, Kaufmann H, Liu J-Q, Peil A, Seul U, Salamini F, Uhrig H. (1995) Molecular genetic analysis of self-incompatibility in potato (Solanum tuberosum L). Advances in Plant Breeding 18: 139-148 TöpferR ,Maa sC ,Hörricke-Grandpierr e C,Schel lJ ,Steinbis sH-H .(1993) .Expressio nvector s for high-level geneexpressio n in dicotyledonous undmonocotyledonou s plants.Method s inEnzymology , 217,109pp. TrognitzBR ,Schmiedich eP E(1993 )A ne wloo ka tincompatibilit yrelationship s inhighe rplants . Sex.Plant . Reprod. 6: 183-190. Uijtewaal BA, HuigenD-J, Hermsen JGTh (1987) Production of potato monohaploids (2n=x=12) through pricklepollination . TAG 73: 751-758 VisserRG F(1991 )Regeneratio n andtransformatio n ofpotat ob yAgrobacterium tumefaciens. InPlan tTissu e Culture Manual B5: 1-9. Kluwer Academic Publishers. Wallace DH (1979) Interactions of S-allelesi n sporophytically controlled self-incompatibility of Brassica. Theor.Appl .Genet . 54(5): 193-201. Waugh R, Baird E, Powell W (1992) The use of RAPD markers for the detection of gene introgression in potato. Plant Cell Reports 11(9):466-46 9 Webber, JM (1948)Nucella r embryony andheterozygosi s in Citrus evolution. Citrus Industry, 1. WehlingP ,Hackau fB ,Wrick eG (1994 )Phosphorilatio n ofpolle nprotein s inrelatio n to self-incompatibility inry e (Secalecéréale L) SexPlan tRepro d 7:67-7 5 References 105

Wemmer T (1991)Characterisierun g desProtein sSK2 u . SeinesGen si nSolanum Spec .Diplomarbeit , Uni­ versität Köln (Thesis,Universit y ofCologne) . WemmerT ,Kaufman n H,Kirc hH-H , Schneider K,Lottspeic h F,Thompso n RD(1994 )Th emos t abundant soluble basic protein of the stylar transmitting tract in potato (Solanumtuberosum L. ) is an endochitinase. Planta 194: 264-273. Wen-jun S,Ford eB G(1989 )Efficien t transformation ofAgrobacterium spp.b y high voltage electroporation. NucleicAcids Research Vol . 17,No .2 0 Wilkinson MJ, Bennett ST, Clulow SA, Allainguillaume J, Harding K, Bennett MD (1995) Evidence for somatic translocation duringpotat o dihaploid induction.Heredit y 74(2): 146-151 Wilson TMA, Saunders K,Dowson-Day , MJ, SleatDE ,Trachse l H, Mundry KW (1993) Effects ofth e 5'- leader sequence ofTobacc o Mosaic VirusRNA , or derivatives thereof, on foreign mRNA andnativ e viral gene expression. Nato ASI Series,Vo l H49.Post-Transcriptiona l Control of Gene Expression. Wing D, Koncz C, Schell J (1989) Conserved function inNicoliana tabacum of a single Drosophila hsp 70promote r heat shock elementwhe n fused toa minima l T-DNApromoter . Mol. Gen. Genet. 219:9-1 6 Summary

Inchapte r 1 anovervie wi sgive no fth emajo r mechanismsoperatin gi nAngiosperm stha tpreven t or limit thedegre e ofinbreeding . Thetw o major systems that function onth ebasi so finteractio n between pollen and stigma/style, areth e sporophytic andth egametophyti c self-incompatibility systems (SSI and GSI). Theplan t is called the sporophyte andpolle n and egg cells are called gametophtytes.I nth esporophyti c system,th epolle n grainscarr yth e information aboutth epolle n donori nthei rcoating .Thus ,th epolle ncoatin gdoe sno treflec t thepolle n genotypebu t depositsi n it reflects the genotype of the pollen donor andth e dominance relationships between the self- incompatibility alleles(5-alleles) .Whe nth erecipien t hasincompatibilit y characteristicsi ncommo n with thepolle n coating,th ecombinatio n willb eincompatibl e andpolle n germination and pollen tube growthwil lb e arrested ono ri nth estigma .I nth eBrassicaceae ,a grou p displaying SSI, signal transduction seemst ob ea nimportan t mechanism fortriggerin g anS Iresponse . Inth egametophyti c self-incompatibility system,th epolle nreflect s thegenotyp eo fth epolle ngrai n itself. When the incompatibility allele(s) of the pollen grain areme t by a similar allele inth e recipient, thepolle n tube growth will be arrested. Thus, selfing provokes a gametophytic self- incompatibility (SI)response .Non-matchin g of^-allele s betweenplant so fth esam especie sresult s in a compatible combination. Most diploid Solanaceous species display GSI. Thestyle s contain extracellularly theproduct s ofth e style-expressed ^-alleles, the^-glycoproteins . About thepolle n components, contributing to SI, little is known, but ^-heterozygosity in thepolle n causes self- compatibility. The cultivated potato,Solanum tuberosum L.(tbr), isa tetraploi d andbehaves , due to ^-heterozygosity in the pollen, as self-compatible species, whereas diploid potato generally possesses anactiv e operating GSI system. Thereexist ,however , alsodiploi d speciestha t areself-compatibl e (SC).Frequently , regardlesso f SSI andGSI ,th eS Ispecie sca nb ecrosse dwit hrelate d SCspecie sonl y when the latter areuse da s female parents.Thi smeans ,th eS Cspecie s canb euse d asth epolle n acceptor (acceptance),bu t the SI species rejects the pollen of the SC species (non-acceptance). This phenomenon, in which interspecific hybridisation can occur inonl y one direction, iscalle d Unilateral Incompatibility or Unilateral Incongruity (UI). Inchapte r 2i ti sdescribe d howth ebasi cplan tmaterial , used forS Iresearch , was developed and selected. Vigour, abundant flowering anda goo d fertility wereprerequisite s forthi s material, but themos t important characteristic wasa reliabl e SIreactio ni npolle n andstyle .Th e combinationo f these characteristics israrel y found indiploi d tbr.Fro m adiploi d tbrpopulation , expressing four different ^-alleles,plant scoul db eselecte d foral lsi xS-heterozygosit y classes,tha tme t allth e afore mentioned criteria.S-allel ecompositio n couldb eteste db yperformin g test crosses,bu ti nadditio n to this, stylar extracts were analysed byiso-electri c focusing, followed bysilve r staining. The S- glycoproteins, also called S-RNases because ofthei r RNase properties, focus inth ebasi c parto f the gels. The selected material was used for the creation and selection of SI plants that were homozygous forth eS-alleles .Normally ,th eS I systemwil lpreven tS-homozygotisation ,unles sth e SI system is weakened by pollen- or style expressed minor or major Si-suppressor genes. A weakening ofth e SIrespons e can cause seed setafte r selfing. Thisi scalle d pseudo-compatibility (PC). Occasionally, however, somepolle n tubesmanag et openetrat eth e style ,eve nwhe n theS I 108 Summary system is fully functional andP C canb e excluded. The seed set will then,howeve r be too lowt o establish asink-sourc erelationshi ptha ti sstron genoug ht ocaus eberr y formation: the flowers will drop.Th eS.phureja (phu)clone sIvP3 5an dIvP4 8ar enormall ycompatibl ewit hdiploi d tbr, butth e hybridsee dha sth eremarkabl ean dusefu l characteristic,tha tth eembryo' shav ea noda lband ,whic h is visible through the seed coat as a seed spot at the first node between hypocotyl and the cotyledons.Pollinatio nwit hthos ephu clones after making crossestha t were incompatible, caused berry formation. This additional pollination is called "counterfeit pollination". Spotless seed, harvestedfrom thos eberries ,yielde dbot h5-heterozygote san dS-homozygotes .Analysi so nth esee d set andth e strength ofth e SIreactio ni nthi soffsprin g showed that, evenwhe nth eorigina l parents wereselecte d forthei rgoo d SIreaction ,weakene d SIwa spresent ,tha t couldb eexpresse d in either the pollen or the style. It was shown that this had a heritable character. From this material, S- homozygotescoul db eselecte dtha twer ereliabl ei nthei rS Ireactio nan dtha t served asteste rclones , as described inth echapter s 3,4 and5 . The selected material, described in chapter 2,wa spoo r in itstransformatio n efficiency. For the functional analysisof , for instance,S-allel ebase d constructs, an efficient transformation system is essential.I twa s decided,therefore , toselec t forthi strait .Transformatio n efficiency wasintroduce d from otherunrelate d sources.Wel ltransformabl e cloneswit h areliabl e stylar SIexpressio n could beselecte dfrom thi smateria l (Appendix 2),tha twer euse dfo ra gain-an dlos so ffunctio n approach. Sense (chapter 3)an dantisens e S-RNaseconstruct s (chapters 3an d 5)wer e introduced by genetic transformation. Indeed, senseS2 transgen e constructs,drive nb y thepromote r ofth e style-specific endochitinase SK2,wer e ablet o cause anincompatibilit y reaction against S2polle n inplant s that did not contain theS2 allel ewhe n not transformed. Some ofthos econstruct s showed such ahig h levelo fexpression ,tha tdu et osom emechanism ,th eendogenou s5-allele swer edown-regulate d and becamecompatibl e for the endogenous S-alleles,whils t remaining incompatible for the transgene S-allele. Theantisens eSl-RNase andS2-RNase construct swer eabl et oreduc eth eexpressio n ofth e corresponding 57 and 52-alleles, which resulted in abreak-dow n ofth e incompatibility reaction againstth e corresponding SI and52-pollen .Thus ,th e gain and loss of function approach showed theke yrol eo fth eS-RNases inth e stylar side ofth e self-incompatibility reaction. In chapter 4 it is described why ver is self-compatible and how this is expressed in hybrid offspring, when crossed with self-incompatible tbr.Whe nth e former species is used as recipient, thehybrid s suffer from cytoplasmic male sterility, thus disabling afurthe r analysis of inheritance andexpressio n of SI, SCan d UI.Th ereciproca l cross fails normally, asalread y stated, duet o UI. However, someo fth epotat o clones,describe d in chapter 2, accepted ver-pollen and yielded male and female fertile hybrid offspring. Those particular potato clones are called "acceptors"for ver pollen,a sa nexceptio nt oth erul eo fUI .Plant stha t showU I arethu scalle d "non-acceptors".I twa s shown that the species verca nb e SCdu et o at least two different reasons: 1)ther e isn o stylar S- glycoprotein anda styla r SIrespons e istherefor e disabled, and 2)ther e isa pollen-expresse d self- compatibilizing factor, SCva. This SCva-factorwa s linked with theS-locu so f ver,a t an estimated distanceo f 18cM .SC m isals ocapabl eo f suppressing the SIreactio n against pollen-expressed tbr ^-alleles.Thi s suppression depends,however , onth e genotype ofth epolle n recipient. Acceptors Summary 109

allow for thepenetratio n ofSC vercarryin gpollen ,bu t specific non-acceptors can inhibit this type ofpollen . It was shown that there exist differential reactions against verpollen , and in particular,

also againstth eSC ve!factor . Experimentswit h somatically doubled hybrids showed that where the stylarpart ofth eS-locu so fve ri sinactive ,tha tth epolle npar to fth e^-comple xi sno t only capable of triggering a UI reaction, but also in causing the Si-based phenomenon of mutual weakening. Mutual weakening isth ephenomeno n that whentw o different S-alleles are expressed in apolle n grain, the GSI reaction in the style will not take place anymore, even when those ^-alleles are expressed inth estyl ea swell .Thus ,a dua lfunctio n ofth epolle npar to fth eS-locu si smad e likely. In chapter 5th e gain and loss of function approach, asdescribe d in chapter 3, was used totes t whetherth e stylarpar to fth eS-locu si sinvolve d inth eU Irespons etoo .Th esens e approach failed, due to the absence of transgenic verrégénérants ,bu t the loss-of-function was successful. Both a transgenicnon-accepto r tbrx verhybri d and atransgeni c non-acceptor tbrclone ,bot h expressing onlyth eS2 allel ei nth estyle ,showe d acollaps eo fth eU Ireactio ntha t coincided withth e antisense S2cause dbreak-dow no fth eS Irespons eagains tS2 pollen .Th eS-locu scomple x showsthu sa dua l function for bothth epolle npar t andth e stylarpart ,bot h contributing toth e SIan d theU Iresponse . Itwa smad e likelytha tver ca nhav ea putativ enon-accepto rbackgroun d for selfpollen , but that its expression requires S-glycoproteinst o be expressed. In this chapter it is discussed why the most importanthypothese s about UI areno t necessarily conflicting with eachother . An explaining and predictivemode lwit hinteraction so fa rang eo fgene san dallele si spresented . Themos t important genes andthei rpropertie s are: - theacceptanc egen e A, whichcause sacceptanc e (aa genotypesbein gnon-acceptors) ,bu t knows different alleles that cause differential reactions against verpollen , - the inhibitor gene/ , which causesnon-acceptanc e and is epistatic over A,

- thepollen-expresse d SCfacto r SCva,whic hi si nwea k linkagewit h the S-locus,cause spolle n to becompatibl ei n anystyle ,excep tthos ewit hth e genotypes aall andaali, inwhic h it is inactive or even causes aU I reaction, - theS-locu s complex withbot h apolle ncomponen t and astyl e component, inwhic hth e pollen verfacto r triggersa U Iresponse ,an dth eactiv estyla rpar ti sneede d for aU Ireactio n inaaii non - acceptor genotypes.Th elatte rexplain swh yth eintroductio no fa nactiv eS-allel ei n a SC species (sucha sreporte db yMurfet t etal. , 1996)ca nbrin gabou t asudde nS Io rU Irespons ean dwh yver canb e aputativ e non-acceptor for selfpollen , without becoming self-incompatible. As a consequence of this, the expression "Unilateral Incompatibility" cannot completely be replaced by the expression "Unilateral Incongruity". The latter expression isvali d in cases where the5-locu sdoe sno t contribute to theU I response atall . Inth efina l chapter someo f theresult s are discussed in abroade r context. The last part stresses thatth edua l function ofth eS-locu simplicate s that,withi nth eexistin g model ofS-RNase activit y inth eS Isystem ,a secon dfunctio n ofth eS-glycoprotein s mustb epostulated .Thi sca nb etriggerin g a signal transduction, resulting in a SI likeresponse ,resultin g inth e arrest ofth epolle n tube,bu t whichma yb e independent ofth eRNas epropertie s ofth e^-glycoproteins . Samenvatting.

In hoofdstuk 1word t een overzicht gegeven van de belangrijkste mechanismen in bloeiende planten om inteelt te beperken of te voorkomen. Sommige mechanismen zijn direct herkenbaar zoals het voorkomen van uitsluitend mannelijke of vrouwlijke bloemen op een plant. Andere mechanismen zijn gebaseerdo phe twe lo fnie tdoorlate nva npollenbuize n door destij l enzij n niet direct zichtbaar. Verschillende van deze mechanismen worden nader toegelicht. De twee hoofdgroepen zijn "sporofytische zelf-incompatibiliteit"e n"gametofytisch e zelf-incompatibiliteit". De plant zelf wordt de sporofyt genoemd, en het stuifmeel de gametofyt. Bij de sporofytische incompatibiliteit wordt, vereenvoudigd gesteld, de informatie van debestuiverplan t meegegeven met de stuifmeelkorrel. Deze informatie zit aan de buitenkant. Als de bestoven plant een "incompatibiliteitskenmerk"gemeenschappelij k heeft met debuitenkan t vanhe t stuifmeel, wordt debevruchtin g onmogelijk gemaakt. Wat eraa ngenetisch e informatie inhe t stuifmeel zit,i sda n niet meer van belang. Dit heet daarom sporofytische incompatibiliteit. Bij gametofytische incompatibiliteit isnie t zozeer debuitenkan t van het stuifmeel van belang, alswe l de genetische inhoud van het stuifmeel. Als de genetische inhoud van de stuifmeelkorrels (pollenkorrels) voor "incompatibiliteit" wordt weerspiegeld in de te bevruchten bloem, wordt de doorgroei van de pollenbuis geremd en gestopt. Zelfbestuiving leidt dus tot een "gametofytische zelfincompatibiliteits-reactie". Hetmechanism e dat in demeest e aardappelsoorten actief is, isd e zogenaamde "gametofytische zelfincompatibiliteit". Inhoofdstu k 2word tvertel dho ehe tbasis-onderzoeksmateriaa l isgeselecteer d enwa t daarui t is gekomen. Hetbasismateriaa l vloeide voort uit kruisingen die ooit gemaakt warentusse n diploïde klonen (24chromosomen) ,di eui td enormale ,tetraploïd e aardappel (Solanum tuberosum (tbr),me t 48 chromosomen) verkregen waren De te gebruiken planten moesten uiteraard goed bloeien en mannelijk en vrouwlijk vruchtbaar (fertiel) zijn, maar bovenal heel betrouwbaar in hun zelf- incompatibiliteitsreactie (SI). Opdi t soort eigenschappen is eri nd e eersterond e geselecteerd. Om duidelijk onderzoek tekunne n doen aanhee l specifieke varianten van de incompatibiliteitsgenen, de "S-allelen", is geprobeerd planten te maken die slechts één type van zo'n S-allel bevatten, de zogenaamde S-homozygoten.Di tdruis t eigenlijk tegenhe t "incompatibiliteitsmechanisme" in,wi l mentoc hno g eenbetrouwbar e SIreacti ebehouden . Soms lukt het eenpollenkorre l toch om door tegroeie ne nd eeice lt ebevruchten ,waa rdi to pgron dva nd e SIreacti enie tverwach twas . Omdat eenlag ezaadzettin g meestal totvroegtijdig e vruchtval leidt,word tdi t soortzaa d zelden verkregen. Daarom is er gebruik gemaakt van een aanvullende bestuiving, waarbij er wel voldoende zaad gevormdwordt . Hetzeldzam ezaa dzi tda nverstop ttusse nhe t"reddende "zaad . Het reddende zaad iseenvoudi gt eherkenne n aanklein evlekje s ophe t embryo,wa t door dezaadhui d zichtbaar is,e n hetzeldzam ezaa dhee fzo' nvlekj e dusniet .Ui tdi tzeldzam ezaa dzij nS-homozygote ngeselecteer d dievitaa lware ne nbetrouwbaa ri nhu nSi-reactie .D eidentificati e vand e5-genotype ngebeurd enie t alleen middels toetskraisingen, maar ook door stijlextracten via gel-electroforese (IEF) en zilverkleuring te analyseren. DeS-allele nproducere n ind e stijl dezogenaamd e S-glycoproteïnen, vanwege hun RNA-afbrekende eigenschappen ook wel S-RNasengenoemd , en deze zijn na IEF goed teherkennen . Dezetechnie k is door alle experimentele hoofdstukken gebruikt. Al dit materiaal heeft aan de basis gestaan van de aardappellijnen die gebruikt zijn in het Samenvatting 111 vervolgonderzoek. Omdat er ook getransformeerd moest worden, wat ook wel "genetische modificatie" wordt genoemd, is erbovendie n gezocht naar lijnen dieredelij k efficient getransfor­ meerdkonde nworden .Hiervoo rmoes te rverde r gekruist engescreen d worden.He t materiaal wat hieruit isvoortgekome n(weergegeve ni nd eappendi xe n kortbesproke n inhe t discussiehoofdstuk), isgebruik t voor de transformatie-experimenten zoalsbeschreve n in dehoofstukke n 3 en5 . In hoofdstuk 3worde n deresultate n van zogenaamde "gain and loss of function" experimenten beschreven.Doo rhe tvi atransformati e toevoegen vanee nextr age nda t codeert voor eenS-RNase, is het mogelijk aan de stijlkant een extra incompatibiliteitsgroep tot expressie te brengen. Dit is gebeurtvoo rhe t£2-allel .Al sdi t overmatigto t expressiekomt ,ka n dit erto e leiden dat deplan t de productieva nd eandere ,niet-transgene ,5-alle lproducte n terugschroeft. Daardoorkunne n de ander incompatibiliteitsreacties koment evervallen .Z oi saangetoon d datd eS3 e n S10allelen , dieweini g overeenkomsvertone nme tS2, z ogoe dal suitgeschakel d kondenworden . Hetblee k ook mogelijk omhe tSI enhe tS2 alle lui t te schakelen doorhe t introduceren van "antisense"constructen . Deze antisense-constructen bevatten eendee lva nhe t coderende stukDN Ava n eenge ncoderen d voor eenS-RNase, maarda nomgedraaid .Dez enonsens-informati e ontregelt opd eéé no fander e manier de expressie van het correcte gen. Het SI-ge nproduceerd e nogredelij k watSl-RNase, maar toch bleken sommige planten compatibel geworden te zijn voor SI pollen. Het 52-allel kon vrijwel volledig uitgeschakeld worden; er was in sommige planten zo goed als geen S2-RNasemee r te detecteren. Ook deze planten waren nu hun zelf-incompatibiliteit voor 52-pollen kwijt. Hiermee werd aangetoond dat de S-RNases, die co-segregeerden met de incompatibiliteitsgroepen, rechtstreeksbetrokke n zijn bij de incompatibiliteitsreactie. In hoofdstuk 4 wordt een eerste link gelegd tussen zelf-incompatibiliteit en een interspecifieke kruisingsbarrière, de zogenaamde unilaterale incompatibiliteit. Solanumverrucosum (ver) i s een zelfcompatibele wilde soort dienau w verwant isaa nd ecultuuraardappel . Dediploïd e variant van onze cultuuraardappel kan normaal gesproken deze soortwe l bevruchten, maar omgekeerd niet. Enkele diploïde tbrklonen , diereed si nhoofdstu k 2besproke n waren,vormde n een uitzondering opdez eregel .Me tdez eklone nblee khe tmogelij k fertielehybride nt emake ndi ei n daaropvolgende kruisingsgeneratieskonde nworde ngeanalyseerd .He tblee kda tver o mtwe eredene n zelfcompatibel konzijn : 1)e rwa s geenS-RNase productie ,du se rko naa nd estijlkan t geen SIreacti e veroorzaakt worden en 2) er was een gen dat aan de pollenkant in bepaalde gevallen een incompatible pollenkorreltoc h doorko n laten groeien.He tmechanism e achterda tlaatst eblee k een ingewikkeld in elkaar te zitten. Voortsblee k dat depollenkan t vand eS-locu sva nS. ver no g steeds actiefwas , diti ntegenstellin g tothe tstijl-expressiedeel . Bovendienblee kd epollenkan t eentweeledig e functie tehebben :he tdroe gbi j aand eS Ireacti emaa roo kaa nd eU Ireactie .Verde rblee k ookno g datd e UIreacti etusse n tbre nver o ptenminst e twee enmogelij k drieverschillend e reactiemechanismen berust. In hoofdstuk vijf werd er verder geanalyseerd aan zowel ver zelf als aan transgene tbr-ver 112Samenvatting

hybridene ntransgen enakomelinge nva nantisens eS2 plante nui thoofdstu k 3.He tuitschakele n van hetenig eactiev eS-alle li ndez eplante nme tbehul pva nantisens everoorzaakt e eenineenstorte nva n deunilateral e incompatibiliteitsreactie. Verderblee k dat verzel f eenzogenaamd e "nonacceptor" achtergrond voor eigenpolle nko nhebben ,hetgee ndoo rhe t ontbreken vanee n actief5-alle lechte r geen enkele consequentie bleek te hebben. Al deze resultaten kunnen met al reeds bestaande modellen goed in overeenstemming worden gebracht, mits er enige flexibiliteit wordt betracht aangaandedominatieverhoudinge n van genen enallelen . Dezelf-incompatibiliteitslocu s blijkt wel degelijk bij te kunnen dragen aan deunilateral e incompatibiliteitsreactie. Alsee nconsequenti e van dezebijdrag e van deS-locu saa nzowe l dezelf-incompatibilitei t als aan de unilaterale incompatibiliteit moet het model voor de interactie tussen pollenbuis en stijl uitgebreid worden. Alleen een specifieke opname of activatie van het S-RNase in de pollenbuis blijkt niet meer te volstaan om e.e.a. met elkaar in overeenstemming te brengen, aangezien de unilaterale incompatibiliteitsreactie niet-allel specifiek is. Bij handhaving van de bestaande modellenmoe te ree ntweed epa dgepostuleer dworden ,da tmogelij k eensignaaltransducti e inhoudt. Curriculumvitae

Curriculum vitae

Ronald Eijlander werd op 4 juni 1958 geboren te Amsterdam. In 1977 behaalde hij het VWO-diploma aan het Berlingh College te Beverwijk. Aansluitend begon hij met de studie Plantenveredeling aan de Landbouwuniversiteit Wageningen (LUW). De doctoraalstudie omvatte twee stages en drie afstudeervakken. De eerste stage was in de aardappelveredling bij de ZPC in Friesland, de andere in de rubberveredeling aan het IRCA te Ivoorkust. Het eerste afstudeervak betrof cytogenetisch onderzoek aan rogge bij de vakgroep Erfelijkheids­ leer, het tweede betrof de epidemiologische ontwikkeling van de schimmelziekte "wolf in spinazie bij de vakgroep Fytopathologie en het derde vak betrof hybride-groeikracht in aardappel na kruising van cultuurmateriaal en een aangepaste wilde verwant, aan de vak­ groep Plantenveredeling. Hij studeerde af injanuar i 1987. Hierna volgen een aantal onder­ zoeksbanen die vooral betrekking hadden op aardappel, bij het voormalige SVP en ITAL. Deze instituten zijn later opgegaan in het huidige CPRO-DLO. In april 1991 is de overstap gemaakt naar de LUW, omdaa r aan een AIO-schapt e beginnen bij de voormalige vakgroep Plantenveredeling. Het betrof een EEG-gefinancierd project dat ten doel had mechanismen van gametofytische zelfincompatibiliteit te onderzoeken. Aardappel was daar een geschikt modelgewas voor, en aldus werd er door de promovendus wederom aan aardappel gewerkt. Deresultate n van hetpromotie-onderzoe k staan beschreven indi t proefschrift.