Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Molecular and genetic analyses of the Bel1 gene regulating ovule development in Arabidopsis thaliana Modrusan, Zora 1995

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_1995-983256.pdf [ 3.68MB ]
JSON: 831-1.0088876.json
JSON-LD: 831-1.0088876-ld.json
RDF/XML (Pretty): 831-1.0088876-rdf.xml
RDF/JSON: 831-1.0088876-rdf.json
Turtle: 831-1.0088876-turtle.txt
N-Triples: 831-1.0088876-rdf-ntriples.txt
Original Record: 831-1.0088876-source.json
Full Text

Full Text

MOLECULAR AND GENETIC ANALYSES OF THE BELl GENE REGULATINGOVULE DEVELOPMENT IN ARABIDOPSIS THALIANAbyZora ModrusanB.Sc., University of Zagreb, Croatia, 1986M.Sc., University of Zagreb, Croatia, 1988A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFDOCTOR OF PHILOSOPHYinTHE FACULTY OF GRADUATE STUDIESDEPARTMENT OF BOTANYWe accept this thesis as confirmingto the required standardTHE UNIVERSITY OF BRITISH COLUMBIAApril 1995© Zora Modrusan, 1995In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.(Signature)Department of ‘O TA/J L/The University of British ColumbiaVancouver, CanadaDate 2/. i. 5DE-6 (2/88)ABSTRACTPhenotypic characterization of the Arabidopsis thaliana mutant Bell revealednumerous morphological abnormalities of Bell ovules, including the absence of theinner integuments, homeotic transformation of the outer integuments into carpel-likestructures, and incomplete development of the female gametophytes. The homeotictransformation of outer integuments into carpelloid structures in Bell has beencorrelated with abnormal expression of class C gene expression in the Arabidopsisovule. Both direct and indirect evidence established such correlation: first, transcriptsof AQAMOUS, the only class C gene identified to date, were detected throughoutBell ovules at the time of formation of carpel-like structures and second, the highestfrequency of carpel-like structures was found in Bell plants grown underenvironmental conditions that also stimulate class C function. These data suggestedthat the BELl product functions as a negative regulator of AG during ovuledevelopment. Similar to Bell, occasional transformation of ovules into structures withcarpelloid features occurs in Ap2-6 mutant plants, indicating a role of the APETALA2gene in ovule development not previously described. AP2 function during earlyArabidopsis flower development is to suppress AG in the outermost whorls. My datasuggests that AP2 functions in a similar manner during ovule development.Two mutant alleles, beIl-2 and beIl-3, originate from Agrobacterium T-DNAmutagenized populations of Arabidopsis and both of them are due to T-DNAinsertions into the BELl locus. Genomic sequences adjacent to the T-DNA insert inthe beIl-2 were used to isolate the putative BELl gene (L. Reiser and R. Fischer,unpublished data). The identity of the cloned BELl sequence was confirmed bycloning the beIl-3 allele and demonstrating that the T-DNA lies within putative BELlgene and by complementation of the Bell mutant phenotype with the clonedsequence. Characterization of the BELl gene revealed its amino acid sequenceidentity to DNA binding domain of the homeobox family of transcription factors,suggesting that the BELl protein itself may function as a transcriptional regulator.The search for genes homologous to BELl resulted in identification of three newhomeobox genes, ABH1, ABH2, and ABH3. Together with BELl, these three genesdefine a novel homeobox gene subfamily in Arabidopsis.IIITABLE OF CONTENTSpageABSTRACT iiTABLE OF CONTENTS ivLIST OF FIGURES viiLIST OF TABLES ixABBREVIATIONS xACKNOWLEDGEMENTS xii1. LITERATURE REVIEW1.1 Introduction 11.2 Regulation of flower development in Arabidopsis thaliana 21.2.1 The Arabidopsis flower 31.2.2 A model for Arabidopsis flower development 51.2.3 Function and expression of organ-identity genes 61.2.4 The organ-identity gene products 111.2.5 The role of AGAMOUS in carpel morphogenesis 141.3. Plant sexual reproduction 161.3.1 Ovule morphology 161.3.2 Ovule and embryo sac development 191.3.3 Genetic studies of ovule and embryo sac development 211.4 Homeobox genes - regulators of development 231.4.1 The HD is a DNA-binding domain 251.4.2 Functional specificity of HD proteins 291.4.3 KNOTTED1 and KNOTTED1 related homeobox genes 311.4.4 Arabidopsis homeobox genes 351.5 Thesis objectives 382. MATERIALS AND METHODS2.1. Genetic and phenotypic plant analyses 402.1.1 Genetic nomenclature 402.1.2 Plant material and growth conditions 402.1.3 Plant resistance to kanamycin 412.1.4 Detection of the NPTII gene 422.1.5 Genetic mapping 422.1.6 Scanning electron microscopy 432.1.7 Construction of double mutants 43ivpage2.2 Molecular biology techniques 442.2.1 Standard methods of molecular biology 452.2.2 Strains, vectors, enzymes, media and reagents 462.2.3 Genomic DNA isolation and analyses 462.2.4 Plasmid rescue 472.2.5 Screening Arabidopsis cDNA libraries 472.2.6 Bacteriophage ) DNA isolation 482.2.7 DNA sequence determination 502.3 Generation of transgenic plants 512.3.1 Agrobacterium strains and vectors 512.3.2 Media for plant transformation 512.3.3 Root tissue culture 532.3.4 In planta vacuum infiltration 543. CHARACTERIZATION OF THE BELl MUTANT3.1 Introduction 563.2 Results 573.2.1 Isolation and genetic analyses of new bell alleles 573.2.2 Wild-type Arabidopsis ovules 603.2.3 Bell mutant ovule development 663.2.4 Bell ovules develop carpel-like structures 683.2.5 Mutations in BELl affect inflorescence termination 713.2.6 Mutations in AP2 result in carpelloid ovules 723.2.7 Bell-3/Ap2 and Bell-3/Ap3 double mutant analyses 733.3 Discussion 783.3.1 BELl gene is required for ovule development 783.3.2 Bell carpel-Iike structures 793.3.3 BELl influences inflorescence structure 813.3.4 AP2 gene is involved in ovule morphogenesis 823.3.5 Role of organ-identity genes in ovule development 834. CHARACTERIZATION OF THE BELl GENE4.1 Introduction 854.1.1 Isolation of the BELl gene 854.1.2 Expression of the BELl gene 874.2 Results 914.2.1 Cloning of the beIl-3 mutant allele 914.2.2 Complementation of the Bell mutant phenotype 944.2.3 Characterization of the BELl gene product 102Vpage4.3 Discussion 1064.3.1 Confirmation of the cloned sequence as the BELl gene 1064.3.2 T-DNA insertions in the BELl gene 1074.3.3 BELl is a novel homeobox gene 1084.3.4 Function of BELl in ovule/plant development 1105. ARABIDOPSIS GENES HOMOLOGOUS TO BELl5.1 Introduction 1135.2 Results 1145.2.1 Isolation of cDNA sequences 1145.2.2 Sequence analyses of cDNA sequences 1155.2.3 Identification of Arabidopsis BEL 1-like homeobox genes 1165.3 Discussion 1185.3.1 Arabidopsis BELl homeobox gene subfamily 118CONCLUDING REMARKS 121FURTHER INVESTIGATIONS 124BIBLIOGRAPHY 126viUST OF FIGURESpageFigure 1 .1. Morphology of wild-type Arabidopsis flower and the domains of 4expression of homeotic genes controlling floral organ-identity.Figure 1.2. Morphology and anatomy of angiosperm ovules. 17Figure 1.3. Comparison of amino acid sequence identity between theconsensus homeodomain and homodomains of genes from Drosophila,yeast, maize, and Arabidopsis. 26Figure 3.1. Position of BELl on chromosome 5 of Arabidopsis. 60Figure 3.2. Scanning electron micrographs of developing wild-type and Bellovules. 63Figure 3.3. Light micrographs of developing wild-type and Bell ovules. 65Figure 3.4. Carpel-like structures and carpel-sepal ovules in Bell and Ap2-6mutant siliques, respectively. 69Figure 3.5. Scanning electron micrographs of Ap3 mutant and Bell-3/Ap3double mutant flowers. 74Figure 3.6. Scanning electron micrographs of Ap2 mutant and Bell-3/Ap2double mutant flowers. 76Figure 4.1. Restriction fragment map of BELl genomic region withSouthern blot analysis showing T-DNA insertion within the 2.9 Kb Hindlllgenomic fragment of both beIl-2 and beIl-3 alleles. 87Figure 4.2. BELl cDNA nucleotide sequence and deduced amino acidsequence of the BELl protein. 88Figure 4.3. Diagrammatic representation of the structural organization ofthe BELl gene including the genomic DNA sequence, BELl mRNA, andBELl protein. 90Figure 4.4. Diagrammatic representation of the T-DNA in pGV3850:1003plasmid and T-DNA insertion within genomic DNA of beIl-3. 93viipageFigure 4.5. Diagrammatic representation of the BELl genomic sequenceswithin two constructs used for Agrobacterium-mediated transformation. 96Figure 4.6. Outline of genetic analyses used for identification ofcomplemented Bell plants. 99Figure 4.7. Southern blot analysis of transformed Arabidopsis plantsshowing distinct banding pattern characteristic of three forms of the BELllocus: BELl, beIl-2 and BEL1-T. 101Figure 4.8. Scanning electron micrographs of wild-type, Bell-3, andtransgenic ovules. 103Figure 4.9. Amino acid sequence identity within the homeodomain ofBELl and other plant homeobox genes highly homologous to BELl. 105Figure 5.1. Amino acid sequence identity between BELl and three BELlhomologs, ABH1, ABH2, and ABH3. 117VIIIUST OF TABLESpageTable 3.1. Wild-type and Bell ovule development. 67Table 3.2. Effects of growing conditions and genetic background ondevelopment of Bell carpel-like structures. 71ixABBREVIATIONSABH Arabidopsis BELl homologAG AGAMOUSAPi APETALA1AP2 APETALA2AP3 APETALA3BELl BELL1BEL1-T BELL 1 transgenebp basepair°C degree CelsiusCM callus inducing mediaCLS carpel-like structurecso carpel-sepal ovuleCol Columbia ecotypecm centimeter2,4-D 2,4-dichioro-phenoxyacetic acidDEFA DEFICIENSE einsteinEDTA ethylenediaminetetraacetic acid, disod ium salt d ihyd rateF1 first filial progenyF2 second filial progenyg gravitation forceGLO GLOBOSAGM germination mediahr hourHD homeodomainIAA indole-3-acetic acidIM infiltration media2ip N6-(2-isopentenyl)adenineKb kilobaseKanR kanamycin resistantKanS kanamycin sensitivexKy kilovoltI literLB Luria-BertaniLer Landsberg erecta ecotypeLM light microscopem meterM molarMADS-box sequence motif present in MCM1, AG, DEF, SRF lociMCM1 mini chromosome maintenance deficientMES 2 [N-Morpholino] ethanesulfonic acidmg milligrammm minuteml millilitermM millimolarNM naphthol-acetic acidng nanogramNPTII neomycin phosphotransferaseSIM shoot inducing mediaSDS sodium dodecyl sulfatePCR Polymerase chain reactionP1 PISTILLATARIM root inducing mediarpm revolutions per minuteRT room temperaturesec secondSEM scanning electron microscopeSRF serum response factorTE Tris-EDTA bufferWS Wasselevskija ecotypewt wild typepg microgramp1 microliterxiACKNOWLEDGEMENTSThere are many people to thank for the last several enjoyable years. First ofall, many thanks to my advisor, Dr. George Haughn. His guidance, enthusiasm,advice, and constructive criticism have been sincerely appreciated. I also thank therest of my supervisory committee for their interest and input in my research: Drs.Larry Fowke, John King and Gopalan Selvaraj at the University of Saskatchewan,Saskatoon, and here, at UBC, Drs. Carl Douglas and John Carison. I acknowledgethe financial support of both University of Saskatchewan and UBC.Special thanks to Dr. Alon Samach for generating enthusiastic workingatmosphere and generous help on the project, to Tamara Western for her patiencein reading this thesis, and all members, past and present, in Dr. G. Haughn’s lab atboth Universities. Both Dr. Robert Fischer and Leonore Reiser at the University ofBerkeley, California, were part of this project and here I would like to thank them forcollaborative efforts and communication of data.I would like to thank many of my friends for their constant encouragement andmore importantly, fun times outside the lab. Most especially, I would like to thank myfamily in Zagreb, Croatia, for their generous support and extreme understanding.xii1. LITERATURE REVIEW1.1 IntroductionThe development of each of the different floral organs, including sepals, petals,stamens, and carpels, begins with a group of cells representing a distinct, organ-specific primordium. Subsequent development of a primordium involves multiple celldivisions, cell expansion and cell-specific differentiation leading to a distinctmorphological structure. In the last few years, the determination of floral organ typehas been the central question in studies of Arabidopsis flower development; however,very little is known about how subsequent steps of organ-specific morphogenesis arespecified. Morphogenesis of the female reproductive organ, the carpel, includes bothdifferentiation of its typical parts (stigma, style and ovary) and development of distinctsubstructures, the ovules. .Ovules themselves function in plant sexual reproductionto house the female gametophyte, mediate fertilization and support the developmentof the embryo. To date, studies on ovules have focused primarily on the structuralaspects of their development and morphological and physiological changes occurringduring different phases of female gametophyte and embryo development. Mostrecently several research groups including Dr. Haughn’s laboratory have begun toinvestigate the genetic mechanisms controlling ovule morphogenesis usingArabidopsis thaliana mutants with defective ovule morphology. In this thesis Iinvestigate the role of one such gene, BELl. I provide evidence that BELl helps to1determine the fate of ovule primordia at least in part by negatively regulating the classC floral organ-identity function. Molecular characterization of BELl revealed that itbelongs to the homeobox gene family of transcription factors whose regulatoryfunction has been well studied in a range of eukaryotic organisms. Considering allthese aspects of the BELl gene, in the following literature review I have summarized:1) the molecular-genetic control of Arabidopsis flower development; 2) the structure,function and development of ovules; 3) structural and functional characteristics ofhomeobox genes themselves and in particular, homeobox genes identified in plants.1.2 Regulation of flower development in Arabidopsis thalianaArabidopsis flowers begin their development as a group of cells (floralmeristem) which arise on the flanks of the apical (inflorescence) meristem of theprimary shoot. Following the establishment of the floral meristem, organized groupsof cells known as floral organ primordia are formed in different positions within thefloral meristem. The cells of these primordia divide and then differentiate into the fourfloral organ types (sepals, petals, stamens, and carpels) that occupy preciselydefined positions within four concentric whorls, commonly referred to by number fromthe outermost to the innermost (whorls 1, 2, 3, and 4, respectively, Fig. 1.1A).In the last decade, morphological, genetic, and molecular studies of homeoticmutants in which at least one floral organ type is replaced by another, havesignificantly contributed to our understanding of flower development in Arabidopsis.2These studies identified numerous genes involved in the control of flowerdevelopment. Throughout these studies, the process of floral development has beendivided into two phases, the first one being determination of the floral meristem, andthe second being the specification of floral organ type. Thus, identified genes wereclassified in two groups known as meristem-identity (or FLIP, Eloral InitiationErogram) and organ-identity genes, respectively. Although convenient, this divisioninto two stages with corresponding group of genes is artificial because some of thegenes appear to function in both stages of flower development. However, despitean increasing amount of information about the genetic regulation of both stages offlower development, the mechanisms by which cells establish their position within theflower and subsequently differentiate into appropriate types are largely unknown.The following section summarizes the present knowledge on the specificationof floral organ identity (for reviews see Coen and Meyerowitz, 1991; Okamuro et al.,1993; Ma, 1994; Weigel and Meyerowitz, 1994; Haughn et al., 1995). A briefdescription of the wild-type flower is followed by a model which proposes howseveral regulatory genes can interact to direct floral organ identity. Genetic andmolecular evidence supporting the model is presented last.1.2.1 The Arabidopsis flowerWild-type Arabidopsis flowers possess the radial symmetry and organographytypical of the Brassicaceae (Polowick and Sawhney, 1986), with four concentric3A)1 2 3I API,AP2 IFigure 1.1. Morphology of wild-type Arabidopsis flower and the domains ofexpression of organ-identity genes controlling floral organ identity in these flowers.A) Wild-type Arabidopsis flower (magnification x28), and floral diagram of wild-typeArabidopsis flower showing the four concentric whorls of organs. 1St whorl sepals(Se), 2nd whorl petals (pe), 3d whorl = stamens (st), 4th whorl = carpels (c).B) Diagrammatic representation of the spatial expression of the organ-identity geneclasses A (APi, AP2), B (AP3, P1), and C (AG) in four whorls of wild-type Arabidopsisflowers.whorls of floral organs (Figure 1 .1A). The first (outermost) floral whorl contains foursepals, while four petals whose positions alternate with those of sepals are found inthe second whorl. Six stamens, four long medial and two short lateral occupy thethird whorl and a superior gynoecium composed of a two-chambered ovary, toppedwith a short style capped with stigmatic papillae, occupies the fourth (innermost)B)whorl whorl whorl whorlIsepalAGI A’3,PI IpetalIstamen carpel4whorl (Figure 1 .1A). The gynoecium contains approximately 30-50 ovules arrangedin four rows along the margins of fusion of the two carpels.The morphology and ontogeny of the Arabidopsis flower has been wellcharacterized (Muller, 1961; Bowman et al., 1989; Hill and Lord, 1989; Kunst et al.,1989), and divided into 13 defined stages from the first appearance of floralprimordium to anthesis (Smyth et al., 1990). New floral primordia arise in a spiralphyllotactic pattern at the top of the inflorescence. Thus, the inflorescence is araceme and it contains a complete developmental series of flowers, from theyoungest floral primordium at the apex to mature fruits toward the base.1.2.2 A model for Arabidopsis flower developmentA model for the control of floral organ identity has been proposed on the basisof genetic studies of Arabidopsis floral homeotic mutants (Haughn and Somerville,1988; Bowman et al., 1991a; Coen and Meyerowitz, 1991; Meyerowitz et al., 1991).Three classes of genes (organ-identity genes) have been defined (A, B, and C) eachof which functions in two adjacent whorls. The three classes act alone or incombination to specify a unique identity for the organs in each of the four whorls ofthe Arabidopsis flower (Figure 1.1B). Class A genes alone specify development ofsepals in the first whorl; class A and B genes in combination designate petals in thesecond whorl; class B together with class C genes control stamen fate in the thirdwhorl; and class C genes alone specify the development of the gynoecium in the5fourth whorl of the Arabidopsis flower. An important aspect of the model is that theA and C functions antagonize each other, such that A inhibits C in the two outermostwhorls, and C inhibits A in the two innermost whorls.At present, five Arabidopsis organ-identity genes have been identified andcharacterized: two class A genes, APETALA1 (APi, Irish and Sussex, 1990; Bowmanet al., 1993; Schultz and Haughn, 1993) and APETALA2 (AP2, Komaki et al., 1988;Bowman et al., 1989, 1991a; Kunst et al., 1989); two class B genes, APETALA3 (AP3,Bowman etal., 1989, 1991a) and PISTILLATA (P1, Hill and Lord, 1989; Bowman etal.,1989, 1991a); and one class C gene, AGAMOUS (AG, Bowman et al., 1989, 1991a,1991b). All five organ-identity genes, APi, AP2, AP3, P!, and AG, have been clonedand their gene products appear to function as transcription factors controlling theArabidopsis flower development (Mandel et al., 1992b; Jofuku et al., 1994; Jack etal., 1992; Goto and Meyerowitz, 1994; Yanofsky et al., 1990).1.2.3 Function and expression of organ-identity genesRecessive mutations in the AP2 gene cause homeotic transformation ofperianth organs into carpels and stamens and a reduction in organ number in thefirst three whorls (Komaki et al., 1988, Bowman et al., 1989, 1991a; Kunst et al.,1989). The transformation of sepals and petals into reproductive organs in Ap2mutant flowers indicates that wild type AP2 gene is required for normal developmentof perianth organs and thus represents a class A organ-identity gene (Figure 1.1B).6Moreover, the presence of carpels and stamens in the outermost floral whorls of Ap2mutant flowers suggests that gene activities responsible for development ofreproductive organs are normally suppressed by AP2 activity. Under specificenvironmental conditions, mutations in the AP2 result in the production ofinflorescence-like structures, the coflorescences, in the positions normally occupiedby flowers (Komaki et al., 1988; Okamuro et al., 1993; Schultz and Haughn, 1993;Shannon and Meeks-Wagner, 1993). The Ap2 mutant phenotype characterized bycoflorescence transformation revealed an additional role of AP2 in the determinationof floral meristem identity. Thus, beside being a class A organ-identity gene, the AP2can also be classed as one of the FLIP genes (see Schultz and Haughn, 1993). Inaddition to both described transformations, Ap2 mutant flowers show subtlephenotypic changes in the fourth floral whorl, including the incomplete fusion ofgynoecium, homeotic transformation of ovules, and defects in seed coat morphology(Kunst et al., 1989; Modrusan et al., 1994b; Jofuku et al., 1994). These phenotypicchanges suggest that AP2 has an additional role late in flower development.Expression studies have demonstrated the presence of AP2 transcripts in all fourfloral whorls and the whole inflorescence (Jofuku et aI., 1994), supporting its geneticroles in Arabidopsis flower development.Mutations in the APi gene disturb both stages of flower development, flowermeristem specification and floral organ identity. In the Api mutant flowers the floralmeristem is partially converted into an inflorescence shoot causing the production ofsecondary flowers in the axile of the first whorl organs. Thus, APi gene is required7for determination of floral meristem identity, and similar to AP2, can be considereda FLIP gene (Irish and Sussex, 1990; Bowman et al., 1993; Schultz and Haughn1993; Shannon and Meeks-Wagner, 1993). The perianth organ identity is alsoaffected by the absence of APi activity (Irish and Sussex, 1990; Mandel et al., 1992b;Bowman et al., 1993; Schultz and Haughn 1993). In the outermost whorl of certainapi mutant alleles (i.e. api-il), development of carpel-sepal organs similar to thoseseen in the first whorl of Ap2 mutant flowers suggest that APi is a class A organidentity gene (Schultz and Haughn, 1993; Figure 1.1 B). Moreover, the type of organtransformation suggests that, like AP2, APi has a role in suppressing thedevelopment of reproductive organs in the outermost floral whorls. The transcriptionpattern of APi revealed its initial transcription in young floral primordia and furtherlocalization to sepals and petals which corresponds to the affected parts of Apimutant flowers (Mandel et al., 1992b; Gustafson-Brown et al., 1994).Mutations in both the AP3 and P1 genes result in a similar mutant phenotypecharacterized by the conversion of petals and stamens into sepals and carpels,respectively (Hill and Lord, 1989; Bowman etal., 1989, 1991a; Jack etal., 1992). Themutant phenotypes suggest that both AP3 and P1 are required for development ofpetals and stamens in the second and third floral whorls and thus AP3 and Pt areclassed as B organ-identity genes (Figure i.1B). The initial patterns of P1 and AP3expression differ, Pt being transcribed in the three innermost and AP3 in the secondand third floral whorls (Goto and Meyerowitz, 1994). However, the later transcriptionpatterns of AP3 and P1 are coincident, both restricted to floral organ primordia that8subsequently give rise to stamens and petals (Jack et al., 1992; Goto andMeyerowitz, 1994) which is in agreement with their proposed roles in Arabidopsisflower development. The analyses of the expression pattern of AP3 and P1 genes infloral homeotic mutants Pi and Ap3, respectively, indicate that both genes requireeach other for proper gene expression (i.e. AP3 depends on F!, and vice versa; Jacket al., 1994; Goto and Meyerowitz, 1994). Moreover, studies in vitro have showedthat the AP3 and P1 gene products bind to each other in solution (Goto andMeyerowitz, 1994), suggesting that they function as a heterodimer in vivo. Theanalyses of beta-glucoronidase (GUS) gene expression in transgenic plantscontaining both the GUS reporter gene under the control of the AP3 promoter andthe AP3 gene under the control of the constitutive Cauliflower Mosaic Virus 35S(CAMV 35S) promoter demonstrated ectopic expression of GUS in the fourth whorl(Jack et al., 1994), indicating that AP3 positively autoregulates.At present, only one class C organ-identity gene has been recognized, AG(Figure 1.1 B). The Ag mutant flower is characterized by the homeotic conversion ofstamens into petals and the replacement of the pistil by an inner flower with the samestructure as the outer. Consequently, the flower structure is (sepals4, petals4,petals6), a pattern known as flower-within-a-flower (Bowman et al., 1989, 1991a).Thus theAG gene is required for the normal development of the reproductive organs,stamens and carpels, and for the determinacy of the floral meristem. Moreover,development of perianth organs throughout the Ag flower suggests that AG geneactivity normally suppresses such developmental events in the innermost floral whorls9(Bowman et al., 1991a). The pattern of AG transcription in the early development ofwild-type flowers is within the third and fourth whorl organ primordia which eventuallydevelop into stamens and carpels (Yanofsky et al., 1990; Drews et al., 1991). AGtranscripts are also detected in late floral development during morphogenesis of floralorgans. The late expression pattern demonstrated the presence of AG transcriptswithin specific cells of stamens and carpels (Bowman et al., 1991 b), suggesting theadditional late role of AG in the morphogenesis of the reproductive organs.The model for the control of Arabidopsis flower development suggestsantagonism between the A and C classes of organ-identity genes. The similarity ofmutant phenotypes between Api and Ap2 flowers indicate that both class A genes,APi and AP2, suppress development of reproductive organs in the two outermostwhorls, (Bowman et al., 1989, 1991a; Komaki et al., 1988; Kunst et al., 1989) andthus negatively regulate the class C gene(s). This observation is supported by thefact that the expression of the class C gene, AG, expands outward to all floral whorlsin the absence of AP2 gene activity (Drews et al., 1991). Moreover, ectopicexpression of AG or the AG homologs from Brass/ca, petunia, and tobacco in theouter whorls of transgenic flowers results in Ap2-like flowers (Mandel et al., 1992a;Mizukarni and Ma, 1992; Kempin et al., 1993; Tsuchimoto et al., 1993). In contrastto Ap2 mutants, the pattern of AG transcription is not significantly altered in Apimutant flowers (Gustafson-Brown et al., 1994). The proposed antagonism alsopredicts that class A genes are ectopically expressed in two innermost whorls of theclass C mutant Ag. In support of the predicted antagonism between A and C gene10classes, APi transcripts are found in the inner whorls of Ag mutant flowers(Gustafson-Brown et at., 1994). However, this is not the case with AP2, whosetranscripts are present in all floral whorls of wild type flowers (Jofuku et at., 1994).Recent modifications of the model for the control of organ identity in theArabidopsis flowers propose that APi and AP2 genes function in the two inner floralwhorls. This hypothesis is based on the phenotype of strong loss-of-function apimutant alleles in which third-whorl stamens are transformed into carpets (Schultz andHaughn, 1993). Moreover, conversion of fourth-whorl ovules into carpet-likestructures occur in strong ap2 alleles (Modrusan et at., 1994b). Both types oftransformation suggest down-regulation of AG in the third and fourth whorls by APiand AP2 genes. The patterns of transcription of APi or AP2 genes in Arabidopsiswild-type flower oppose or support, respectively, such observations (Mandel et at.,1992b; Jofuku et at., 1994)1.2.4 The organ-identity gene productsThe five Arabidopsis organ-identity genes previously mentioned have beencloned (APi, Mandel et al., 1992b; AP2, Jofuku et at., 1994; AP3, Jack et al., 1992;P1, Goto and Meyerowitz, 1994; AG, Yanofsky et at., 1990). Genes homologous tothe Arabidopsis organ-identity genes APi, AP3, P1 and AG have been alsocharacterized in the distantly related dicotyledonous species Antirrhinum, asSQUAMOSA (SQA), DEFICIENS (DEFA), GLOBOSA (GLO) and PLENA (PLE),11respectively (Schwarz-Sommer et al., 1990; Sommer et al., 1990; Huijser et al., 1992;Trobner et al., 1992; Bradley et at., 1993). Nucleotide and amino acid sequenceanalyses have revealed that APi, AP3, P1, and AG genes, as well as their homologsfrom AntIrrhinum, encode members of the same family of putative transcriptionfactors. Each gene product is characterized by a N-terminal region of 60 amino acidresidues which shows a high degree of sequence similarity to the conserved DNA-binding and dimerization domains of two known transcription factors, mammalianserum esponse Eactor (SRF, Norman et al., 1988), and yeast Mini-hromosomeMaintenance (MCM1, Passmore et al., 1988). The conserved domain shared by allof these genes has been designated the MADS-box for the initials of the first fouridentified family members (Iy1CM1, AG, EFA, RF, Schwarz-Sommer et al., 1990).To date, numerous MADS-box genes involved in floral morphogenesis have beenisolated from variety of dicotyledonous plant species, including Arabidopsis itself (Maet al., 1991), Petunia hybrida (pMADS1, pMADS2, pMADS3, fbp 1, fbp2, Angenent etal., 1992, 1993, 1994; van der Krol et al., 1993; Tsuchimoto et al., 1993), Brassicanapus (BAG1, Mandel et al., 1992a), Nicotiana tabacum (NAQ1, Kempin et al., 1993),Lycopersicon esculentum (TM3, TM4, TM5, TM6, TM8, TAG1, Pnuelli et at., 1991,1994), and the monocot Zea mays (ZAG1, ZAG2, Schmidt et al., 1993). These datasuggest that a large family of MADS-box transcription factors function predominantlyas regulators of flower development (for review see Davies and Schwarz-Sommer,1994), and that the regulatory mechanisms involved in the control of floral organidentity are similar to those in Arabidopsis and other distantly related angiosperms.12Much of what is known about the regulatory function of MADS-box genescomes from in vitro studies of MCM1 and SRF. These studies demonstrate that thebasic N-terminal portion of the SRF MADS-box region makes contacts with specificDNA sequences and thus corresponds to a sequence-specific DNA binding domain(Pollock and Treisman 1991). The rest of the MADS-box region of both SAF andMCM1, together with a C-terminal extension of the box, mediates dimerization as wellas interactions with other regulatory proteins (Norman et al., 1988; Ammerer, 1990;Christ and Tye, 1991; Mueller and Nordheim, 1991; Primig et al., 1991). Studies invitro demonstrated that the MADS-box domain of DEFA and GLO from Antirrhinum,as well as that from the Arabidopsis gene AG is sufficient for recognition and bindingof specific DNA sequences (Schwarz-Sommer et al., 1992; Trobner et al., 1992;Shiraishi et al. 1993; Huang et al., 1993). In contrast to MCM1 and SRF, the plantMADS-box genes share an additional domain of homology, termed the K-box, dueto its structural similarity to the coiled-coil domain of keratin (Ma et al., 1991). TheK-box, which is composed of 65 amino acid residues, is presumably capable offorming two or three amphipathic ct. helices where regularly spaced hydrophobicresidues are brought together on one face of the helix (termed a coiled-coil).Moreover, coiled-coil motifs of identical or two different polypeptides can interact topromote homo- or heterodimerization. Thus, the presence of a coiled-coil suggeststhat either dimerization or interaction with accessory transcription factors is involvedin the regulatory function of plant MADS-box genes (Ma et al., 1991; Pnuelli et al.,1991). Evidence for heterodimerization of MADS-box proteins came from studies in13vitro of AP3 and P1 (Goto and Meyerowitz, 1994) as well as their Antirrhinumhomologs DEFA and GLO (Schwarz-Sommer et al., 1992; Trobner et al., 1992). Bothsets of studies suggest that homologous pairs, AP3/PI and DEFA/GLO function as aheterodimer in Arabidopsis and Antirrhinum, respectively.The organ-identity gene AP2 differs from the others such that its gene productdoes not belong to MADS-box family of transcription factors (Jofuku et al., 1994).Nevertheless, AP2 encodes a protein that contains highly an acidic serine-rich domainand a putative nuclear localization signal, both strongly suggestive that AP2 functionsas a transcription factor. Moreover, the so-called AP2-domain seems capable offorming amphipathic c helices, indicating that the AP2 protein may interact with otherproteins (Jofuku et al., 1994).1.2.5 The role of AGAMQUS in carpel morphogenesisThe transcripts of organ-identity genes APi, AP3 and AG have been detectedin discrete domains of maturing organs, late in floral development (Bowman et al.,1991b; Jack et al., 1992; Mandel et al., 1992b). It is possible therefore that inaddition to specifying the identity of floral organ primordia, these organ-identity geneshave roles in late floral organ development (Okada et aL, 1989). For example, AG isthe only class C gene known to be required for development of the femalereproductive organ, the carpel, in the Arabidopsis flower (Bowman et al., 1989,1991a). Initially, AG transcripts are evenly distributed throughout fourth-whorl carpel14primordia. Later in floral development, the expression of AG becomes restricted tospecific cell types, for example, to stigmatic papillae and developing ovules. Similarearly and late expression patterns have been observed for the AG homologs PLEfrom Antirrhinum and ZAG1 from maize (Bradley et al., 1993; Schmidt et al., 1993).Together, these data suggest that AG or its homologs are required for carpel-specificmorphogenesis, including ovule differentiation late in Arabidopsis flower development.However, genetic evidence challenges this hypothesis. Although Ag mutant flowersdo not produce any organ with carpel-like characteristics, Ag/Ap2 double mutantflowers have morphologically normal ovules and stigmatic papillae, suggesting thatthe AG is not necessary for these aspects of carpel development (Bowman et al.,1991a).Although direct data supporting AG function late in carpel morphogenesis aremissing, several genes identified through molecular or genetic approaches appearto be involved in various aspects of such developmental events, and thus representpotential candidates for regulators of carpel morphogenesis. Six genes designatedAOL 1-AGL6 encoding putative MADS-box transcription factors have been isolatedbased on their sequence homology to the MADS-box domain of the AG gene (Ma etal., 1991). One of them, the AGL1 gene is transcribed throughout carpels, inparticular within ovules, possibly influencing ovule morphogenesis. In addition, twogenes regulating certain aspects of carpel morphogenesis, BELL1 (BELl) and SHORTLNTEGUMENTS (SIN1), have been identified through Arabidopsis mutants defectivein ovule morphogenesis (Robinson-Beers et al., 1992; Modrusan et al., 1994b). One15of these genes, BELl, has been shown to negatively regulate AG, suppressing theformation of carpel-like structures during development of ovules (Modrusan et al.,1994b; Ray et aL, 1994). The regulatory function of the BELl gene in Arabidopsisflower and ovule development is discussed later in detail (see Results).1.3. Plant sexual reproductionThe life cycle of flowering plants is characterized by the alternation ofgenerations between a diploid sporophyte and a haploid gametophyte. In contrastto lower plant species characterized by free-living, nutritionally independentgametophytes, the development of male and female gametophytes of angiospermstakes place on the parent plant, within the specialized structures of reproductive floralorgans. The male gametophytes (pollen grains) develop within the anthers ofstamens, whereas the female gametophyte (embryo sac) is a product of the ovulelocated within the gynoecium. Formation of the male and female gametes (spermand egg cells) within the pollen grain and embryo sac, respectively, marks the initialstep of plant reproduction. In flowering plants, sexual reproduction designates thefusion of haploid sperm and egg cells, resulting in the formation of a new diploidsporophyte or embryo.1.3.1 Ovule morphology16A) B)• •.--irma: iilegumeri(- (wJ1.44 sac- •3’-.<c’ •-4ScdsaFigure 1.2. The morphology and anatomy of angiosperm ovule.A) Structure of the angiosperm ovule consisting of nucellus, inner and outerinteguments, and funiculus (taken from Bouman, 1984).B) Diagram of Arabidopsis ovule with Polygonum-type of embryo sac as seen in amedial longitudinal section (taken from Mansfield, 1991). Funiculus (F), nucellus (N),inner integument (I), micropyle (Mi), egg cell (EC), synergids (SC), antipodal cells(AC), central cell (CC), vascular bundle (VB).Ovules are specialized reproductive structures that are the site of femalegametophyte formation, fertilization, and development of the embryo. Differentaspects of ovular morphology have been studied in various plant species (for reviewsee Bouman, 1984). To date, the morphology, histology, and ultrastructuralcharacteristics of Arabidopsis ovules have been well characterized (Misra, 1962;Webb and Gunning 1990, 1991; Guignard et al., 1991; Mansfield and Briarty, 1991;Mansfield et al., 1991; Robinson-Beers et aL, 1992; Modrusan et al., 1994a, 1994b).17The ovule commonly consists of three parts: the nucellus, one or twointeguments, and the funiculus (Figure 1.2A, Esau, 1977). The nucellus ormegasporangium consists of both vegetative and sporogenous cells. The femalegametophyte develops within the nucellus from a sporogenous archesporial cell(s)which undergo(es) meiosis (megasporogenesis), followed by few mitotic divisions ofthe haploid megaspore (megagametogenesis). The nucellusis enveloped by one or,more often, two integuments which overgrow the nucellus, leaving an opening, themicropyle, for pollen tube entry. The role of the integuments is to protect andnourish the nucellus, and later, in mature seed, they form the seed-coat. Thefuniculus, a stalk-like structure, connects the ovule with the ovary wall and usuallycontains a single vascular strand through which water and nutrients are supplied.The region where integuments are inserted is known as the chalaza, and it ispositioned opposite to the micropylar end of the ovule.Numerous classifications of angiosperm ovules are based on the position,number, and origin of the different ovular parts (Davies, 1966). For example, ovulesmay have two, one, or no integuments, and are consequently termed bitegmic,unitegmic, and ategmic. Furthermore, a minimum of three principal morphologicaltypes of ovules are recognized depending on the degree of ovule curvature (Giffordand Foster, 1989). The orthotropous types have erect ovules in which funiculus andnucellus lie in a straight line, anatropous ovules have the micropyle adjacent to thefuniculus as result of a 1800 curvature of the funiculus, and campylotropous ovuleshave the funiculus attached midway between the micropyle and chalaza, as a result18of the curvature of both the nucellus and the funiculus. Campylotropus ovules areunique in having curved embryo sacs and when curvature is extreme, they areclassified as a fourth, amphitropous type (Bocquet, 1959). Another classification ofovules is based on the pattern of development of the archesporial cell(s) within thenucellus. In crassinucellate ovules the archesporial cell(s) is always separated fromthe nucellar epidermis by one or several cell layers. In contrast, tenuinucellate ovuleshave the archesporial cell(s) directly below the nucellar epidermis.1.3.2 Ovule and embryo sac developmentThe ovule primordia derive from the placental tissue of the ovary wall (Esau,1977; Bouman, 1984). Initiation of the ovule primordium is a result of periclinaldivisions of two subdermal cell layers (L2 and L3) and anticlinal cell divisions ofdermal layer (Li). Soon after, the subdermal cell(s) within the primordium enlarg(es)and display(s) a prominent nucleus. This archesporial cell(s) may function directlyas the megasporocyte, or it may undergo one mitotic division to produce themegasporocyte. At this stage of development the apex and the base of ovuleprimordium differentiate into the nucellus and funiculus, respectively. Following thedifferentiation of ovule primordium, the integuments are initiated at the base of thenucellus; the inner integument is most often dermal in origin, while the outerintegument is usually derived from both the Li and L2 layers. At the same time,megasporogenesis is initiated within the nucellus. As the development of the female19gametophyte proceeds, the inner and outer integuments continue to enlarge andovergrow the nucellus. In most angiosperm ovules, the nucellus is absorbed beforethe maturity of the female gametophyte and thus, the embryo sac comes in directcontact with the inner integument. The innermost cell layer of the inner integumentthen differentiates into a unique cell layer, the endothelium. In ovules where thenucellus does not degenerate an endothelial layer does not form.Megasporogenesis and megagametogenesis lead to the formation of thefemale gametophyte. During megasporogenesis, the megasporocyte undergoesmeiosis to produce four megaspores. During megagametogenesis, themegaspore(s) undergoes several mitotic divisions, and together with subsequentnuclear migration and fusion, cytokinesis, and cell differentiation, result in formationof a mature embryo sac. The pattern of embryo sac development varies considerablyamong plant species, but the majority of them, including that of Arabidopsis, are ofthe Polygonum-type (Russel, 1978; Mansfield etal., 1990; Webb and Gunning, 1990).A Polygonum-type embryo sac derives from a single functional megaspore locatedat the chalazal end which undergoes three successive mitotic divisions to producethe seven-celled eight-nucleate embryo sac.In most angiosperm ovules, the Polygonum-type embryo sac consists of thefollowing seven cells: the egg cell, two synergids, the central cell, and three antipodalcells (Figure 1 .2B, Esau, 1977). The egg cell lies at the micropylar end of the embryosac and eventually fuses with the sperm cell to form a diploid embryo. The egg cellis highly polarized due to a large vacuole which restrains the nucleus and most of the20cytoplasm to the chalazal end of the cell (Cass et al., 1985; Sumner and VanCaeseele, 1989). As well, there are usually few ribosomes and organelles in the eggcell, consistent with its quiescent metabolic state (Mansfield et al., 1990). Adjacentto the egg cell, there are two synergids that are the primary site of pollen tubedischarge, and thus they have an important role in fertilization. The central cellcontaining two partially fused nuclei (diploid) is located in the centre of the embryo•sac. During fertilization, the central cell fuses with another sperm cell to form theendosperm (triploid) which functions as a source of nutrients for the developingembryo. Finally, at the chalazal end ‘of the embryo sac, there are three antipodalcells whose function in plant sexual reproduction has not been specified. However,the antipodal cells like the synergids and the central cell appear to have highmetabolic activity.1.3.3 Genetic studies of ovule and embryo sac developmentTo date, very little is known about the genetic mechanisms regulating theinitiation and development of ovules, or formation of the female gametophyte inflowering plants. Several mutations affecting such developmental events have beendescribed in different plant species, the majority of which disrupt meiosis within bothmale and female gametophytes (Kenell and Homer, 1985; Benavante et al., 1989;Golubovskaya et al., 1992). In the maize mutant mel and in the dyad mutant ofDatura, altered development of the embryo sac is the result of defects in21megasporogenesis (Golubovskaya, 1989; Satina and Blakeslee, 1935).Genetic studies undertaken in Arabidopsis have identified several othermutations affecting the development of ovule (for review see Reiser and Fischer,1993). The Female ametophyte Eactor (Gt) mutant has ovules that contain multiplearchesporial cells, an increased number of megaspores, and twin embryo sacs, butotherwise are normal in morphology (Redei, 1964). Other Arabidopsis mutants thatare defective in ovule morphogenesis include Bell, Sinl and Ovule Mutant-2 (Ovm2)(Robinson-Beers et al., 1992; Modrusan et aL, 1994b; Reiser and Fischer, 1993). Thecommon feature âf all these mutants is the absence of a functional embryo sacaccompanied by various defects in the development of integuments. For example,Sini mutant ovules possess both integuments but their development is altered dueto the abnormal cell elongation. Moreover, the development of Sini ovules differssignificantly in various genetic backgrounds and is dependent on an interaction of theSINlgene and its modifier MOD1 (Lang et al., 1994). In Bell ovules, the innerintegument is never initiated whereas the outer one develops abnormally (RobinsonBeers et al., 1992; Modrusan et al., 1994b). Finally, Ovm2 mutant ovules arecharacterized by absence of the inner and outer integuments (Reiser and Fischer,1993). Together, these Arabidopsis mutants affected in distinct steps of ovulemorphogenesis represent useful tools for the genetic dissection of pathwaysregulating ovule development.The phenomenon of ovules developing into the floral organ type from whichthey derive, the carpel, has been recognized for a long time. Since 1841,22development of carpelloid ovules was recognized in Primula and Cheirantus (deCandolle and de Candolle, 1841) and in carnations (Masters, 1869). About onehundred years later, such aberrant ovule development was examined in more detail.Placental tissue of tobacco ovaries cultured in vitro result in transformation of ovulesinto stigmatoid and carpelloid outgrowths (Hicks and McHughen, 1974, 1977) as wellas rudimentary carpets (Haccius et al., 1974). These studies also revealedrequirements for various nutrients and plant hormones for normal development ofovules (Beasley and Ting, 1974; Peterson, 1973). Similar studies in vitro demonstratethat ovules of two tobacco mutant plants, Mrg3 and Mrg9, have the capacity todevelop into carpelloid structures, however, only at the time when their developmentas ovules has not yet been committed (Evans and Malmberg, 1989). Together, theseobservations demonstrate the capacity of ovules to undertake alternativedevelopmental pathways and moreover, suggest that complex but precisemechanisms control their morphogenesis.1.4. Homeobox genes - regulators of developmentBased on a domain of conserved amino acid sequence, the BELl generepresents a member of a group of regulatory genes, the homeobox genes.Homeobox genes are characterized by 180 bp DNA segment, the homebox, whichencodes a precisely defined part of the protein, the homeodomain (HD, McGinnis etat., 1984; Scott and Weiner, 1984). The HD has been identified as a DNA-binding23domain that confers sequence-specific binding (for review see Gehring et al., 1994a).The function of homeobox genes is to regulate gene expression, in particular, thetranscription of subordinate target genes (for reviews see Levine and Hoey, 1988;Affolter et at., 1990a).Homeobox genes have been extensively studied as regulators of developmentin the fruit fly, Drosophila melanogaster (for review see Gehring, 1987). They wereoriginally identified through molecular-genetic studies of homeotic mutants in whichone body part is transformed into the homologous structure found on another bodysegment. These studies revealed that homeobox genes are involved in embryopattern formation and specification of segmental identity along the anterior-posteriorbody axis (for reviews see lngham, 1988; McGinnis and Krumlauf, 1992; Lawrenceand Morata, 1994). In the past few years homeobox genes have been found invariety of eukaryotic organisms ranging from yeast to vertebrates, and their generalrole in development, with respect to the morphological specification of the body plan,appears to be conserved throughout the evolution (for review see Kenyon, 1994).The recent discovery of homeobox genes in higher plants has suggested thateven though plant and animal development is very different, the basic regulatorymechanisms may be conserved among higher eukaryotes. In contrast to Drosophilahomeobox genes, the role of their plant counterparts is still largely unknown.However, some homeobox genes have been shown to affect primarily variousaspects of plant development including determination of cell fate (Hake, 1992;Schena et al., 1993; Rerie et al., 1994). At present, the majority of plant homeobox24genes have been identified in two species that have been extensively used formolecular-genetic studies, maize (Vollbrecht et al., 1991; Bellman and Werr, 1992),and Arabidopsis (Ruberti et al., 1991; Mattsson et al., 1992; Schena and Davies,1992; Schindler et al., 1993; Lincoln et al., 1994; Korfhage et al., 1994; Rerie et al.,1994). These studies identified at least 25 different genes that have been recentlyclassified into several homeobox gene subfamilies (Kerstetter et al., 1994; Schenaand Davies, 1994). Since Arabidopsis homeobox gene BELl is discussed in detailin this thesis, the following section includes description of main structural andfunctional features of the HD, followed by a summary on known homeobox genes inflowering plants.1.4.1 The HD is a DNA-binding domainThe HD includes a sequence of 60 amino acid residues that are highlyconserved among all 350 identified homeobox genes, thus defined as consensus HDsequence (Figure 1.3, for reviews see Scott et al., 1989; Gehring et al., 1994b). Inall HDs analyzed four amino acid residues are invariant (W48, F49, N51, R53) and eightresidues are highly conserved (R5,Q12, L16, F20, L40, I/V451 K55, K57, Figure 1.3), implyingthat they are required for some specific functions common to all homeobox genes.The three dimensional structure of the HD determined by nuclear magneticresonance (NMR) spectroscopy has demonstrated that the core of the HD consistsof four helical regions folded into a tight globular structure (Qian et al., 1989; Billeter251 10 20 30 40 50 60N-term HelixI Loop Helixil Turn HelixillV V V V V V ****V VPRRKRTAYT RYQILELEKEFH FNRLT RRRRIELZHSL NLT ERQVKIWFQNRRMKWKKEN cons.-K-G-QT T I--A- C AntpDEKP---FS SE--AR-KR--N E E---QQ-SSE- G-N -A K-A-I--ST en-GHRF-KE VRI-ESWF1’KNI E-P--D TKGLEN-MKNT S-S RI---N-VS---R-E-TIT MAT2KSP-GKSSI SP-?R2EEEQVF RRKQSL NSKEK-EVPKK GI- PL- -RV--I-K--RS- NATalKKGKLIPKEA -Q---SWWDQHY KWP-PS ETQKVA--E-T G-D LK--N--I-Q-IHW-PS KN1SThPKGHFG PVINQK-HEH-K P-PS -SVKES--EEL G-T F- --NK--ET--HS?RV7-S ‘4FIJJ2ASSSSACKQ- DPKTQR-YIS-Q E-Q-PD KATKES- -KE- QM- VK- -NN- -KH- -WSINSKP HAT3. 1KKG-LPKE -QK--’IWWELHY KWP-PS ESEKX7A--EST G-D QK--NN--I-Q-K-HW-PS I4T1Figure 1.3. Comparison of amino acid sequences between the consensus HDand HDs of homeobox genes from different organisms, including Drosophila (Ant,en), yeast (MATa2, MATa 1), maize (KN1, ZMHOX1A) and Arabidopsis (HAT3. 1, andKNAT1). The numbers designate amino acid residue positions within the HD; thepositions of helices (I, II and Ill) are indicated. In the consensus HD sequence, fourinvariant amino acid residues (*) and eight residues that are highly conserved amongmost HD (‘r) are indicated in bold. Within other homeobox genes, residues that arein bold represent similar amino acid al., 1990). Helix I, preceded by a flexible N-terminal arm, is connected by a loopto helix II, which is separated by a tight turn (tripeptide) of helix Ill and its extension,the flexible helix IV at the C-terminus (Figure 1.3). The structural motif formed byhelix II, the connecting turn, and helices III and IV, is identical to the helix-turn-helix(HTH) motif found in many prokaryotic regulatory proteins. Moreover, this HTH motifof several prokaryotic proteins, for example, A repressor, Lac repressor, and Croprotein, and the HD of the Drosophila protein Antennapedia (Ant) may be preciselysuperimposed when their backbone structures are compared (Gehring et al., 1990),indicating a high conservation of the HTH motif among distantly related species.Another structural feature of the HD is its overall globular conformation held together26by a core of 10 amino acid residues and their side chains (Qian et al., 1989; Bilteteret al., 1990). Two of these residues are invariant (Trp, Phe49), whereas the othersare either highly conserved (i.e., Leu16, Leu38, Leu, lle/Val45) or exclusivelyhydrophobic in all known HD proteins. The high conservation of core amino acidresidues among different HDs implies that all HD proteins may adopt similar threedimensional structures, and thus perform similar regulatory functions.The properties of HD proteins required to bind specific DNA sequences havebeen analyzed by both studies in vitro (i.e. in vitro DNA binding, DNA footprinting, gelmobility shift assays) and in vivo (i.e., transactivation assays; for review see Laughon,1991). The first evidence for the requirement of the HD for site-specific recognitioncame from studies on the yeast MATh2 (Johnson and Herskowitz, 1985; HaIl andJohnson, 1987) and Drosophila engrailed (en) and Ant proteins (Desplan et at., 1985;Muller et al., 1988). The target DNA sequences identified by several distantly relatedHD proteins in vitro were found to be very similar to each other (Desptan et al., 1988;Hoey and Levine, 1988), and moreover, the HD proteins were shown to recognizedegenerate DNA sequences (Baumruker et aL, 1988). Soon after, it was found thatthe majority of the HD proteins, including ones from invertebrates and vertebrates,recognize DNA sequences containing a 4 bp TAAT core motif (Hayashi and Scott,1990). Mutational analysis demonstrated that each conserved bp in the TAATsequence contributes to the high affinity DNA binding by HD proteins (Laughon,1991).The three dimensional structures of HD-DNA complexes were obtained both27by NMR spectroscopy for the Ant protein (Otting al., 1990; Billeter et al., 1993) andby X-ray crystallography for the en (Kissinger et al., 1990) and MATc2 (Wolberger etal., 1991) polypeptides. The mode of docking of the HD to DNA has been wellconserved among all complexes analyzed even though three HDs belong to distantlyrelated homeobox genes. The third helix, so-called recognition helix, of the HD ispositioned in the major groove of the DNA where it makes specific contacts with 4bp core motif TAAT. In addition, the flexible N-terminal arm of the HD contacts 2 bpin the minor groove of the DNA. Besides interactions in both major and minorgrooves of DNA, the HD establishes numerous contacts with the DNA backbone. Anindependent line of evidence supporting this model of DNA binding by the HD wasprovided at the same time by ethylation and methylation interference studies (Affolteret al., 1990b; Percival-Smith et al., 1990). Moreover, some of the essential featuresof HD-DNA complex have been verified by studies in vivo where the mutations inDrosophila genes Ant, fushi tarazu (ftz), Paired (Prd), and bicoid (bcd) altered eithertheir binding to DNA or functional specificity of the HDs (Hanes and Brent, 1989;Treisman et al., 1989; Schier and Gehring, 1992). These experiments pinpointed theimportance of the amino acid residue in position 9 of the recognition helix (position50 of the HD) for the DNA-binding specificity of the HD. When eukaryotic HDs werecompared to the HTH motif of prokaryotic regulatory proteins, a number ofdifferences were found, including the positions of amino acid residues important forDNA binding specificity (for review see Treisman et at., 1992).281.4.2 Functional specificity of HD proteinsAs mentioned above, a large number of HD proteins recognize extremelysimilar DNA sequences in vitro which is in sharp contrast to the distinctly differentbiological effects exerted by these proteins in vivo (for reviews see Gehring 1987;Krumlauf, 1994). The low selectivity of the HD proteins in vitro suggests that someadditional molecular strategies are involved in determining their functional specificityin vivo. The specific spatial and temporal pattern of gene expression, varying levelsof transcripts, differential regulation (i.e., postranscriptional, postranslational), and theavailability of different transacting factors in specific cells/tissues contributesignificantly to the specificity of action with respect to target gene regulation.Moreover, the presence of multiple DNA-binding sites of varying affinity may influencethe functional specificity of the different HD proteins. For example, the Drosophila ftzprotein sufficiently binds DNA and activates transcription only when multiple medium-affinity binding sites are present (Schier and Gehring, 1992). However, besidesmultiple medium-affinity sites, conserved DNA-binding sites for other regulatoryproteins are required for proper ftz gene function (Schier and Gehring, 1993). Thedifferential regulatory activity may also involve selective protein-protein interactions.The cofactors may increase activity, affinity or specificity of HD proteins for their targetsites either by binding the DNA or interacting with other proteins. At present, thereare several examples of such combinatorial interaction between HD proteins andspecific cofactors. VP1 6, encoded by the herpes simplex virus, is a non-DNA-binding29protein which in association with HD protein Oct-i activates transcription. The VP16differentially influences the ability of mammalian octamer proteins Oct-i and Oct-2,which bind indistinguishably to the same DNA sequence, to activate transcription(Stern et al., 1989). Furthermore, mammalian HD protein Phoxi enhances the DNA-binding activity of another mammalian transcription factor, the MADS-box protein SRF(Grueneberg et al., 1992). The specificity of the yeast MATh2 HD protein isdetermined in two ways: first, the MATh2 increases its own weak DNA-binding affinityby forming homodimers or heterodimers (MATc2/MATh2, MATh2/MATa1); second,the association of MATc2 with the specifià cofactors MATa1 or MCM1 determine itsregulatory activity (MATh2/MATa1 represses haploid-specific genes, MATx2/MCM1represses a-specific genes; for review see Herskowitz, 1989). Another possibility toaccount for the specificity of proteins is that the sequence of a binding site may alterthe conformation of a protein. Such conformational change is then transmitted to apart of the protein outside the HD, thus, affecting the ability of the protein to activatetranscription.There are several examples where HD proteins contain additional conserveddomains, most often another DNA-binding domain, that adds further specificity totheir DNA recognition. The presence of two DNA-binding domains in the sameprotein strongly increases its specificity by adding a second binding site, and bydetermining the precise spacing between these two sites. For example, the Pairedbox (Prd) is a second DNA binding domain that can act independently orcooperatively with the HD to enhance the specificity of the Prd-HD proteins (Treisman30et al., 1991). The two most common DNA-binding domains found in eukaryoticregulatory proteins, the HD and the zinc finger, may be in the same polypeptide.Indeed, the presence of multiple HDs and zinc finger domains have been reported(Fortini et al., 1991). The POU domain (for PJt-1, Qct-1, Qct-2, nc-86 proteins) isalways found together with a POU type of HD, characterized by a cysteine residueat position 9 of the recognition helix (Herr et al., 1988). The POU domain contactsthe DNA next to the HD binding site but it is unable to bind in the absence of the HD(Kristie and Sharp, 1990). Analyses of mutant Pit-i proteins demonstrated that thePOU domain is essential for sequence specific, high-affinity DNA binding as well asPit-i dimerization (Ingraham et al., 1990). Finally, the HD has been found associatedwith a cysteine-rich domain called LIM (for lin-li, jsl-1, miec-3; Freyd et al., 1990).The function of the LIM domain itself is not known, however, in contrast to the otherconserved domains, it appears to inhibit DNA binding by the HD, thereby, negativelyregulating the regulatory activity of LIM-HD proteins (Sanchez-Garcia et al., 1993).The association of various domains with the HD served as a base for classificationof HD proteins (Scott et al., 1989; Gehring et al., 1994b).1.4.3 KNOTTED 1 and KNOTTED 1-related homeobox genesHomeobox gene KNO7TED1 (KN1) from maize was the first gene from higherplants that joined the large family of homeobox genes. The KN1 gene was isolatedby transposon tagging (Hake et al., 1989) and was shown to encode a HD protein31(Vollbrecht et al., 1991), containing the four invariant (W48, F49, N51, and R53) and fourout of eight of the highly conserved amino acid residues in the HD (Q12, L4c, Q, R55;Figure 1.3). All known Kni mutant phenotypes are the result of dominant gain-of-function mutations caused by a Ds2 transposable element insertion within the KN1locus (Hake et al., 1989; Veit et al., 1990). Kni mutations primarily affect thedevelopment of the leaf blade where uncontrolled cell divisions along the lateral veinsresult in formation of outgrowths or knots. The knots are present in all cell layers ofthe leaf blade although it has been shown through clonal analysis that only thegenotype of the innermost cell layer is important for the mutant phenotype (Sinha andHake, 1990). KN1 expression studies revealed that the KN1 gene is transcribed onlyat very low levels in wild-type leaves, and strongly in vegetative and floral meristems(Smith et al., 1992). The ectopic expression of KN1 gene in mutant leaves wasconsistent with the presence of Kni mutant phenotype. Constitutive expression ofKN1 gene caused dramatic changes in the phenotype of transgenic tobacco plants(Sinha et al., 1993). These phenotypes were dependent on the level of KN1 proteinand were characterized by a lack of apical dominance, thus the transgenic plantswere dwarfed in overall height and leaf size. Based on the KN1 expression studiesand phenotypes of Kni mutant plants, it has been suggested that the KN1 genefunctions to maintain the indeterminate state of shoot apical meristem, and thereforeits down-regulation results in the initiation of determinate organs such as leaves andfloral organs.To date, genes homologous to KN1 have been identified in variety of plant32species including Arabidopsis (Lincoln et al., 1994), rice (Matsuoka et al., 1993), andsoybean (Ma et al., 1994). Arabidopsis genes KNAT1 and KNAT2 (for iJiOTTED-likein Arabidopsis thaliana) share 53% and 40% amino acid identity, respectively, with theKN1 protein. The KNAT1 and KN1 genes have similar patterns of expression; bothtranscripts are detected primarily in the shoot apical meristem and they decreaseduring the leaf primordia initiation. Moreover, transgenic Arabidopsis plants whereeither KNAT1 or KN1 has been constitutively expressed had similar phenotypescharacterized by highly abnormal leaf morphology (Lincoln et al., 1994). Together,these observations suggest that the KNAT1 is an Arabidopsis gene homologous tothe maize KN1 gene. KN1 homolog from rice, the OSH1 gene (Qriza s_ativaijomeobox gene 1) has been also overexpressed in Arabidopsis, rice (Matsuoka etal., 1993), and tobacco (Kano-Murakami et al., 1993). All transformants showedabnormal phenotypes, typically forming clumps of shoot apices that never elongated.Such phenotypes suggested that OSH1 functions as a morphological regulator in thesame manner as K!S.I1. In contrast to KNAT1 and OSH1, the SBH1 gene (SoybeanHomeobox gene 1) was transcribed strongly in early-stage somatic embryos, weaklyin soybean stems, and was completely silent in other plant tissues (Ma et al., 1994).The proposed role of SBH1 in plant embryo development has been based on itsenhanced expression during embryogenesis and high homology to the maize KN1gene.When cloning KN1, two KN1 related homeobox genes, ZMH1 and ZMH2 wereidentified (Vollbrecht et al., 1991). Recently, the homeobox sequence of KN1 has33been used to isolate ten additional sequences designated as KNOX genes (K_Nirelated homeob; Kerstetter et al., 1994). Although the HD sequences of all twelvegenes are very similar, they have been grouped into two classes based on thedifferences outside the conserved recognition helix. The HDs in class I (eight genes)share 73%-89% identical residues with the KN1 peptide whereas a lower percentageof homology (55%-58%) is found for class II genes (four genes). Moreover, theexpression of the two KNOX gene classes is different; class I genes are highlytranscribed in meristem-rich tissues, similar to the KN1 itself; class II genes aregenerally expressed in all tissues but their expression patterns show noticeablediversity (Kerstetter et al., 1994). The expression pattern of three members of classI, KNOX, KNOX and RS1, has been examined in detail (Jackson etal., 1994). All threegenes are transcribed in shoot meristems (i.e., vegetative and reproductive) but notin determinate lateral organs (i.e., leaves, flowers). Moreover, the genes have distincttranscription patterns in vegetative shoot meristems that together indicate the site ofsubsequent leaf initiation (Jackson et al., 1994). Thus, the KNOX, KNOX and RS1genes may function to establish the pattern of morphogenesis in the vegetative shoot.The screening of a maize expression library with the cis-acting regulatoryelement from the Shrunken promoter, has revealed two homeobox genes, ZMHOX1Aand ZMHOX1B, which are distantly related to the maize KNOX gene family (Bellmanand Werr, 1992). The ZMHOX1A protein contains different conserved domains suchas a HD, a leucine zipper, a POU-B subdomain, and a cysteine-rich motif. Thespecific function of these motifs is not yet known, although it seems likely that they34contribute to the functional specificity of the ZMHOX1A gene product. Based on itsamino acid sequence, the maize protein ZMHOX1A shows the highest homology tothe Arabidopsis protein HAT3. 1 which contains a typical HD and a cysteine-rich motif(Schindler et al., 1993).1.4.4 Arabidopsis homeobox genesIn the last few years, Arabidopsis has been extensively used to search for newhomeobox genes, through sequence homology with the consensus HD sequence.However, the biological roles of most are still unknown. Based on the structuralcharacteristics of all identified HD proteins, Arabidopsis HD proteins and theircorresponding genes can be divided in three distinct classes. One class contains HDproteins with two conserved domains, the HD and the leucine zipper motif and arethus designated the HD-Zip class (Ruberti et al., 1991; Mattsson et al., 1992; Schenaand Davies 1992, 1994). Plant homeodomain finger proteins (PHD-finger) is anotherclass of Arabidopsis HD proteins characterized by a conserved cysteine-rich motif(Schindler et al., 1993; Korthage et al., 1994). Finally, HD proteins with atypical, butdistinct, HD represent a third class (Rerie et al., 1994).Most of the Arabidopsis homeobox genes are members of the HD-Zip genesuperfamily (Schena and Davies, 1992). The HD-Zip proteins have an unusualstructure in that they have a leucine zipper motif adjacent to the HD. The leucinezipper motif is a coiled-coil with a series of leucine residues spaced by exactly seven35amino acids. The distribution of residues permits the formation of an amphipathic chelix which dimerizes, forming a protein-protein complex (Landschulz et al., 1988).The close proximity of the leucine zipper motif to the HD suggests that the HD-Zipproteins recognize dyad-symmetrical DNA sequences. This was confirmed in vitrofor HD-Zip protein ATHD1 which dimerizes via its leucine zipper and binds a 9-bpdyad-symmetric sequence (CAAT(NT)ATTG, Sessa et at., 1993). The correct spatialrelationship between the HD and leucine zipper motif is crucial for successful DNAbinding. In HD-Zip proteins ATHD1 and ATHD2 the spatial arrangement between theDNA binding and the dimerization domains is similar to that found in other regulatoryproteins, the basic-leucine zipper (b-Zip) transcription factors (Vinson et al., 1989).Besides ATHD1 and ATHD2 (Ruberti et al., 1991), other members of Arabidopsis HDZip gene superfamily are HAT4, 5, 22, 24 (Schena and Davies, 1992), ATHD3(Mattsson et at., 1992), ATHD4 (Carabelli et al., 1993), ATHD5, 6, 7, 8 (Morelli et al.,1993), and HAT1, 2, 3, 4, 5, 7, 8, 9, 14, 22 (Schena and Davies, 1994). The HD-Zipsuperfamily has been divided into two subfamilies (Carabelli et al., 1993; Schena andDavies, 1994), and a model for the evolution of these subfamilies has been proposed(Schena and Davies, 1994).Because HD-Zip genes have been found only in higher plants, we mayspeculate that they mediate processes unique to plants, one of them beingdevelopmental responses to environmental stimuli. Some evidence for such afunction of Arabidopsis HD-Zip genes come from studies of ATHD2 and ATHD4whose expression in vegetative and reproductive tissues is induced by far-red-rich36light treatments (Carabelli et al., 1993). Moreover, plants without HAT4 or withectopic HAT4 expression exhibit changes in morphology and developmental ratesimilar to those caused by the deficient or excess light stimulation, respectively(Schena et al., 1993). Together, these observations suggest that some HD-Zip geneproducts might be part of a signal transduction pathway by which light affects plantgrowth and development.The HAT3. 1 gene is anArabidopsis homeobox gene that encodes a typical HDand an N-terminal cysteine-rich motif designated the PHD-finger (Schindler et al.,1993). The cysteine-rich domain contains eight regularly spaced cysteine/histidineresidues which resemble other metal-binding domains, for example, the zinc finger(Berg, 1990). The similarity between the cysteine-rich domain of HAT3.1 and knownmetal-binding domains implies a functional similarity. Although the regulatoryfunction of the cysteine-rich domain or HAT3.1 protein has not yet beendemonstrated, its expression pattern demonstrated that the HAT3. 1 gene primarilyfunctions in root tissues (Schindler et al., 1993). The Arabidopsis HD protein PRHAand its parsley homolog PRHP were identified on the basis of their binding to thespecific DNA regulatory element of parsley pathogenesis-related gene PR2 (Korfhageet al., 1994). Both PRHA and PRHP proteins contain a typical HD and a cysteine-richmotif and thereby have homology to the PHD-finger protein HAT3.1. Based onstudies in vitro, for example DNA-binding studies, immunoprecipitation, and transientexpression, PRHA and PRHP have been proposed to be transcriptional regulators ofthe PR2 gene (Korthage et al., 1994).37The homeobox gene GL2 encodes a protein characterized by a typical HDthus defining the third class of Arabidopsis HD proteins (Rerie et al., 1994). GL2gene has been cloned using a T-DNA tagged mutant allele g12-2 affected in trichomedevelopment. The phenotype of G12 mutants as well as the transcription pattern ofthe GL2 in the trichome progenitor cells suggest that Arabidopsis HD protein GL2may regulate developmental events which result in cell elongation during trichomeformation (Rerie et al., 1994). In the same class with GL2, there are several KNATgenes that were identified by sequence homology to the maize KN1 gene, containingone conserved domain, the HD (Kerstetter et al., 1994). Based on the expressionpattern studies, two of the KNAT genes (KNAT1 and KNAT2) have been characterizedas regulators of plant morphogenesis (Lincoln et al., 1994). The BELl gene whichis the focus of this thesis shows the highest sequence homology to the KNAT genes,and therefore, it belongs to the same class of Arabidopsis homeobox genes.1.5. THESIS OBJECTIVESIn this thesis, I have used a molecular-genetic approach to investigate thecontrol of ovule morphogenesis in Arabidopsis thaliana. First, I characterized thedevelopment of the Bell mutant plants that have altered ovule morphology. Second,I determined the interaction of the BELl gene with the Arabidopsis organ-identitygenes. Following the isolation of a putative BELl clone (L. Reiser and R. Fischer,unpublished data), I verified its identity through characterization of the beIl-3 mutant38allele and complementation of the Bell mutant phenotype. Finally, in collaborationwith Dr. Alon Samach, I cloned three homeobox regulatory genes closely related toBELl.392. MATERIALS AND METHODS2.1. Genetic and phenotypic plant analyses2.1.1 Genetic nomenclatureI will use standard Arabidopsis genetic nomenclature throughout this thesis.Wild type genes are indicated by uppercase, italicised letters (e.g., AGAMOUS),however, a two or three letter abbreviation is normally used (e.g., AG). In caseswhere several genes result in a similar phenotype they are differentiated using anumber (e.g., APi, AP2, AP3). Mutant alleles are written in lowercase, italics (e.g.,ag). Multiple alleles of a single gene are distinguished by hyphenated numbers (e.g.,ag-i allele was isolated before ag-2). Mutant phenotypes are denoted by the sameabbreviation used for the corresponding gene, having the first letter uppercase, noitalics (e.g., Ag-i). Proteins are indicated by uppercase, non italicised letters (e.g.,AG).2.1.2 Plant material and growth conditionsA line segregating for the Bell mutant phenotype was isolated from acollection of Arabidopsis thaliana plants, ecotype Wasselewskija (WS), transformedwith T-DNA of Agrobacterium tumefaciens (Feldmann, 1991). Before being used for40the analyses, homozygous mutant plants were back-crossed at least three times tothe wild type. All Bell mutant plants were maintained as heterozygotes and analyzedwithin F2 segregating populations.The marker lines used for genetic mapping were Wi 00 (Landsberg erecta, an,api, er, py, hy2, gil, bp, cer2, msi, ff3, Koornneef et al., 1987), and TC 35 (Columbiax Landsberg erecta, yl, pgm, ttg; gift from T. Caspar, Central Experimental Station,Dupont Co., Wilmington, DE). Mutant lines used for construction of double mutantswere Ap2-5, Ap2-6 (Kunst et al., 1989), Ap3-1 (gift from M. Koornneef, Departmentof Genetics, Wageningen Agricultural University, The Netherlands), and Ap3-3 (giftfrom E. Meyerowitz, California Institute of Technology, Pasadena, CA).Plants were normally grown at 22°C under continuous fluorescent illumination(Gro-Lux, Sylvania) of 80-120 pE m2 sec1 photosynthetically active radiation (PAR)on Terra-lite Redi-earth prepared by W.R. Grace & Co. Canada Ltd., Ajax, Ontario.Photoperiodic regimes used other than 24 hrs light (continuous light, CL) were 16 hrs(100-150 pE m2 sec1 PAR) and 10 hrs (150-180 pE m2 sec1 PAR), and othertemperatures (16°C) were used as indicated.2.1.3 Plant resistance to kanamycinResistance of Bell mutant plants to the antibiotic kanamycin was determinedby growing seedlings on minimal media (At medium, Haughn and Somerville, 1986)containing 50 mg/I kanamycin sulfate (Sigma). Seeds were first sterilized with 20%41bleach containing 0.001% Triton-X100 for 15 mm, washed three times with steriledistilled water (dH2O), and then distributed on sterile plates. After incubation at 4°Cfor 3 days, plates were maintained at room temperature (RT) beneath fluorescentlights on 16 hrs Iight/8 hrs dark cycle. After 10 to 14 days, seedlings with greenleaves and well developed root system were scored as kanamycin resistant (KanR)whereas kanamycin sensitive (KanS) seedlings had poorly developed roots and whitecotyledons.2.1.4 Detection of the NPTII geneThe neomycin phosphotransferase Il gene (NPTI!) which confers KanR wasalso identified in the genome using the Polymerase chain reaction (PCR). Plantgenomic DNA (Edwards et al., 1991) was combined with right and left 18 bpoligonucletide NPTI! primers (gift from R. Datla, Plant Biotechnology Institute,Saskatoon, SK.), dNTPs (Pharmacia), Taq polymerase (Gibco-BRL), and PCR buffer(Gibco-BRL). A DNA thermal cycler (model 480, Perkin-Elmer) was used to amplifyNPTII sequence using the following program: jst = 96°C for 2 mi 2 = 94°C for 30see, 3d = 55°C for 30 see, 72°C for 3 mm, 5th = 35 cycles of steps 2,3, and 4,= 72°C for 5 mm, 7th = 4°C for unlimited time.2.1.5 Genetic mapping42The BELl gene was mapped genetically by crossing the Bell-3 mutant plantto appropriate marker lines, and determining frequencies of recombinant progeny inthe F2 population. Map distances were calculated using the method of Suiter et al.,(1983).2.1.6 Scanning electron microscopyFor scanning electron microscopy (SEM), flowers at different stages ofdevelopment were vacuum infiltrated with 3% glutaraldehyde in 0.02 M sodiumphosphate buffer (pH 7.2) and fixed overnight at 4°C. The samples were rinsedseveral times in the same buffer. Flowers at the youngest stages were post-fixed inosmium-tetroxide for 2 hrs to make them more visible. All samples were dehydratedin a graded acetone series (10%, 30%, 50%, 70%, 80%, 96%, 100%) before criticalpoint drying in liquid carbon dioxide and mounting on stubs. In most cases perianthorgans were removed from flowers, and pistils were then dissected with pulled glassneedles. Stubs with prepared material were coated with gold in a sputter coater(Edwards S15OB) and examined with a scanning electron microscope (Philips 505)at an accelerating voltage of 30 kilovolts.2.1.7 Construction of double mutantsBell-3 plants were crossed to plants homozygous for either ap2-5, ap2-6, ap3-431, or ap3-3. The double mutant plants were identified and verified for each cross asfollows:beIl-3/ap2-5 and beIl-3/ap2-6. Among the F2 progeny of Bell -3 x Ap2-5 and Bell -3x Ap2-6 crosses, four phenotypes were found in a ratio of 388 wild type to 142 Ap2-5to 135 BeIl-3 to 42 with the novel phenotype (9:3:3:1, X2=1.06, P>0.5) and 104 wildtype to 35 Ap2-6 to 24 Bell-3 to 8 with the novel phenotype (9:3:3:1, X2=3.56P>0.25) respectively. To confirm that the novel phenotype represented the doublemutant, we allowed F2 progeny with the Ap2 phenotype to self. In the case of 10Ap2-5 plants, five segregated for the novel phenotype in a ratio of 328 Ap2-5 to 104novel (3:1,X2O.20, P>0.5). In the case of 10 Ap2-6 plants, four segregated for thenovel phenotype in a ratio 82 Ap2-6 to 21 novel (3:1, X2=0.012, P>0.9). Fourputative Bell -3/Ap2-5 F2 plants and three putative Bell -3/Ap2-6 F2 plants were shownto be homozygous for ap2-5 (ap2-6) and beIl-3 by testcrossing to homozygous Ap2-5 (Ap2-6) and heterozygous Bell -3 plants. When a homozygous Ap2-5 (Ap2-6) plantwas used for the testcross, all F1 progenies showed Ap2-5 (Ap2-6) mutant phenotype.Among the F1 progenies of crosses Bell-3/Ap2-5 and BeIl-3/Ap2-6 x heterozygousBell-3, two phenotypes were found in a ratio of 42 wild type to 39 BeIl-3 (1:1,X2=0. 11, P>0.5) and 17 wild type to 15 Bell -3 (1:1, X2=0. 125, P>0.75), respectively.Several Ap2-5 and Ap2-6 plants among the F1 progeny from the crosses Bell -3/Ap2-5x Ap2-5 or Bell-3/Ap2-6 x Ap2-6, respectively, were allowed to self, and doublemutants ap2-5/beIl-3 or ap2-6/beIl-3 among their F1 progeny were used forphenotypic analysis.44beIl-3/ap3-1 and beIl-3/ap3-3. Among the F2 progeny of crosses Ap3-1 x BeIl-3and Ap3-3 x Bell-3, a novel phenotype with characteristics of both BeIl-3 and Ap3was observed. Phenotypes in the F2 generation were in a ratio of 252 wild type to73 Ap3-1 to 65 Bell-3 to 21 Ap3-1/Bell-3 (9:3:3:1,X2=4.78, P>O.1) and 233 wild typeto 80 Ap3-3 to 72 Bell-3 to 22 Ap3-3/Bell-3 (9:3:3:1, X2=0.93, P>0.75). Becauseboth BeIl-3/Ap3-1 and Bell-3/Ap3-3 double mutants were male and female sterile,I was unable to perform further genetic analysis on these plants, and phenotypicanalysis was carried out on the putative double mutants from the F2 generation.2.2 Molecular biology techniques2.2.1 Standard methods of molecular biologyRestriction enzyme digestion, agarose gel electrophoresis, DNA modificationand ligation, and bacterial transformation were performed as described in Sambrooket al., (1989). Isolation of small amounts of plasmid DNA was performed accordingto the alkaline denaturation method (Birnboim and Doly, 1979). The Geneclean(Gibco-BRL) and Random priming kits (Gibco-BRL) with their enclosed protocolswere used for recovery of DNA from agarose and radioactive labelling(32P-dCTP) ofDNA probes, respectively. Amplification of certain DNA genomic fragments was doneusing PCR (Innis et al., 1990) and a DNA thermal cycler (model 480, Perkin-Elmer).452.2.2 Strains, vectors, enzymes, media, and reagentsThe bacterial strains used for transformation and growth of bacteriophage Awere DH5c, and LE392, respectively (Sambrook et aL, 1989); the electro-competentbacterial strain ElectroMax DH1OBTM (Gibco-BRL) was used for electro-transformationduring plasmid rescue. The cloning vector pT7T318U (Pharmacia) was commonlyused for subcloning DNA fragments. All restriction and modifying enzymes and theiraccompanying buffers (Gibco-BRL) were used according to the manufacturer’sinstructions. DNasel, RNaseA, and proteinaseK (Sigma) were normally made asstock solutions and kept at -20°C (Sambrook et aL, 1989). Lucia-Bertani (LB) or Abroth were used for standard bacterial or bacteriophage A liquid cultures and platesrespectively; S.O.C. media was used for the recovery of transformed bacterial cells;SM buffer was used to make bacteriophage A suspension and its serial dilutions; allthese as well as other frequently used reagents and buffers, Tris, EDTA, TE, TBE,SSC, and SSPE, were made and used as described in Sambrook et al., (1989).2.2.3 Genomic DNA isolation and analysisPlant genomic DNA was isolated using a CsCI/EtBr density gradient(Sambrook et al., 1989) or quick CTAB preparation (Dean et al., 1992). For Southernblot analysis DNA was transferred to an Hybond N+ nylon membrane (Amersham)according to the method of the membrane manufacturer.462.2.4 Plasmid rescueThe genomic DNA of Bell-3 plants adjacent to the T-DNA insert was recoveredby plasmid rescue (Behringer and Medford, 1992). Genomic DNA (3 pg) wasdigested with restriction enzyme BgI II at 37°C for at least 4 hrs. The digested DNAwas ligated with T4 DNA ligase at 16°C for 16 hrs. Following the precipitation of theligation mixture, electra-competent cells (20 p1) were transformed with ligated DNA(50-1 00 ng). Electroporation was done using the gene pulser and pulse controller(BioRad) with an electric field strength of 25 kV/cm and voltage decay half time of 8-10 milliseconds. Transformed cells were immediately incubated in S.O.C. medium(1 ml) at 37°C for 1 hr with rapid shaking (225 rpm). Aliquots of cells (50 p1 and 200p1) were plated on plates containing LB media and either 100 mg/I of ampicillin(Sigma) or 50 mg/I kanamycin sulfate (Sigma) and incubated overnight at 37°C toselect transformants. Plasmid DNA from transformants was isolated (Birnboim andDaly, 1979) and the T-DNA and its flanking plant genomic sequences subcloned intopT7T318U vector (Pharmacia).2.2.5 Screening Arabidopsis cDNA librariesTwo Arabidopsis cDNA libraries, a A-YES library prepared from adult plants(001-0; Schena and Davies, 1992) and a floral specific library in ANM 1149 (gift fromHans Sommer, Max-Planck-lnstitut fur Zuchtungsforschung, Cologne, Germany), were47plated on ten 8-cm plates per library at a density of 5,000 plaques per plate. Theplaques were transferred to Hybond N+ nylon filters (Amersham) as recommendedby the filter manufacturer. Filters were prehybridized for approximately 24 hrs at 50°Cin 25 ml of 5X SSPE/5X Denhardt’s solution/0.5% SDS and 0.5 ml of salmon spermDNA (1 mg/mI, sodium salt type Ill, Sigma). Hybridization was carried out forapproximately 24 hrs at 50°C in 25 ml of prehybridization solution supplemented with32PdCTP labelled DNA probe (50 ng, specific activity lxlO9dpm/pg). The BELlhomeodomain sequence (PCR amplified fragment from 1490 to 1846 bp of BELlcDNA) was used as a probe to screen both cDNA libraries. Following hybridization,membrane was washed at 50°C for 3 hrs with 2X SSC/0.1% SDS. Positive plaqueswere picked into SM buffer (1 ml) containing a drop of chloroform and used forsecondary and tertiary screens to purify the initial positives of both cDNA libraries.2.2.6 Bacteriophage A DNA isolationA modified version of the small scale liquid culture procedure from Sambrooket al., (1989) was used to isolate bacteriophage A DNA. A single plaque was pickedinto SM buffer (1 ml) containing a drop of chloroform and stored at 4°C for 4-6 hrs.The bacteriophage A suspension (0.5 ml, approximately 3x1 06 bacteriophages) wasmixed with bacterial cells (1 ml O.D.=0.2, approximately 1.6x10cells) and incubatedfor 20 mm at 37°C. Lysis was obtained by adding 10 ml of LB media containing 5mMCaCl2 and incubating cells and bacteriophages for approximately 9 hrs with constant48shaking (225 rpm) at 37°C. Chloroform (250 p1) was added when the lysate wasclear and incubated for another 15 mm in the shaker. Following centrifugation of thelysate (8,000 g for 10 mm at 4°C), the recovered supernatant was treated withRNaseA and DNasel (final concentration of 10 pg/mI each) for 1 hr at 37°C.Precipitation of bacteriophage was done with an equal volume of aqueous solutioncontaining 20% polyethylene glycol (PEG 8,000, Sigma) and 2M NaCI for 1 hr at 0°C.The bacteriophages were then recovered by centrifugation (10,000 g for 20 mm at4°C) and resuspended in SM buffer (0.7 ml) by vortexing. Bacteriophage DNA wasisolated by treating the bacteriophage with 10% SDS and 0.5M EDTA, pH 8.0 (7 p1each), and proteinaseK (final concentration of 50 pg/mI, Sigma) at 65°C for 20 mm.Finally, A DNA was extracted twice with buffered phenol, once with chloroformisoamyl alcohol (24:1), and precipitated with an equal volume of isopropanol. TheDNA pellet was washed with 75% ethanol, dried, and resuspended in TE buffer (pH8.0, 50 p1).The cDNA inserts of the positive plaques scored after the tertiary screen of twocDNA libraries were released with the EcoR I enzyme (A-YES, Elledge et al., 1991;ANM 1149, Murray, 1983). These cDNA fragments were then subcloned intopT7T318U vector (Pharmacia) suitable for further analyses including restrictiondigestion, Southern analyses and DNA sequencing. Southern blot analyses wereperformed using the same conditions as described for the initial screen ofArabidopsis cDNA libraries.492.2.7 DNA sequence determinationA modified alkaline-lysis/polyethylene glycol precipitation procedure was usedto prepare DNA for sequencing (Terminator cycle sequencing protocol, appendix A,Applied Biosystems Inc.,). For determination of DNA sequence, the DyedeoxyTerminator Cycle Sequencing method was used (Terminator Cycle Sequencingprotocol, Applied Biosystems Inc.). DNA template (1.0 pg of double stranded DNA),primers (3.2 pmol of each), terminator premix containing four fluorescent dye-labelleddideoxy nucleotides (9.5 p1, Applied Biosystems Inc.), and dH2O (to final volume of20 p1) were mixed together and cycled (DNA thermal cycler model 480, Perkin-Elmer)using the following program: = rapid ramp to 96°C, 2 96°C for 30 sec, 3 =rapid ramp to 50°C, 4th = 50°C for 15 sec, 5th = rapid 60°C, 6th = 60°C for4 mm, 7th 25 cycles of steps 1,2,3,4,5, and 6, 8t = rapid ramp to 4°C and hold forunlimited time. The removal of dye-labelled dideoxy nucleotides (dyedeoxyterminators) prior to sequencing was done using the centri-sep columns and themethod recommended by the column manufacturer (Princeton Separations, Inc.).Purified DNA sequencing products were further handled by the Nucleic Acid andProtein Sequencing (NAPS) laboratory, UBC, Vancouver. The fluorescence-baseddetermination of DNA sequences was done with an automated DNA sequencer(model 373, Applied Biosystems Inc.). Basic sequence manipulations and sequencehomology searches were done with Alignment (Scientific and educational software)and PC Gene (IntelliGenetics, Inc.).502.3 Generation of transgenic plantsTwo techniques, root tissue culture (Valvekens et al.,1988) and in plantavacuum infiltration (Bechtold et at., 1993; Chang et at., 1994; Katavic et at., 1994)were used forAgrobacterium-mediated transformation ofArabidopsis plants. For roottissue culture I used tissue from both F1 and F2 progenies of the cross wt (WS) xBell-2, whereas for in planta vacuum infiltration only the F2 population segregatingfor Bet 1-2 plants was used.2.3.1 Agrobacterium strains and vectorsThe Agrobacterium strain LBA4404 bearing a helper, nopaline plasmid, MP9O,and derivatives of binary plasmid vectors pCGN 1548 (McBride and Summerfelt,1990) or pRD400 (Datla et al., 1992), were used for both transformation techniques.Agrobacterium cultures were usually grow at 28°C for 24 hrs in LB mediasupplemented with appropriate antibiotics (for pCGN 1548 100 mg/I streptomycin and25 mg/I of gentomycin; for pRD400 100 mg/I of streptomycin and 50 mg/I ofkanamycin).2.3.2 Media for plant transformationThe media used for Agrobacterium-mediated transformation by root tissue51culture and in planta vacuum infiltration are listed below. Plant growth substances2,4-dichloro-phenoxyacetic acid (2,4-D), kinetin, N6-(2isopentenyl) adenine (21p),indole-3-acetic acid (IAA), and benzylaminopurine were made as stock solutions indimethyl sulfoxide (DMSO), filtered and kept at -20°C. Stock solutions of thiamineHCL, nicotinic acid, pyridxine, and naphthol-acetic acid (NAA), as well as antibioticscarbenicillin and kanamycin were made with sterile dH2O, filtered and stored at -20°Cfor several weeks. Plant growth substances and antibiotics were mostly purchasedfrom Sigma or other manufacturers as indicated.At minimal media: 5 mI/i 1M KNO3, 2.5 mI/I 1M KH2Po4,2.5 mI/I 20 mM FeEDTA, 2mi/I 1 M MgSO4,2 mI/I 1 M CaNO3x 4 H20, 1 mi/i micronutrients (70mM H3B0, 14mMMnCI2 x 4 H20, 0.5mM CuSO4, 1mM ZnSO4 x 7 H20, 1.2mM NaMoO4,lOmM NaCI,0.01mM CoC12 x 6 H20).Germination media (GM): At minimal media, 0.5% sucrose, 0.8% phytagar (pH 5.7with 1M KOH).Callus inducing media (CIM): lx Gamborg B-5 (Gibco-BRL), 0.5 g/i 2[N-Morpholino]ethanesulfonic acid (MES, pH 5.7 with 1M KOH), 0.8% phytagar, 0.5 mg/I2,4-D, 0.05 mg/I kinetin.Shoot inducing media (S1M): lx Gamborg B-S (Gibco-BRL), 0.5 g/I MES (pH 5.7 with1M KOH), 0.8% phytagar, 5.0 mg/i 2ip, 0.15 mg/I IAA.Root inducing media (RIM): 1X Murashige-Skoog basal salt mixture, 2% sucrose, 100mg/I inositol, 0.5 g/I MES (pH 5.7 with 1M KOH), 0.8% phytagar, 1 mg/I thiamineHCL, 0.5 mg/I nicotinic acid, 0.5 mg/I pyridxine, 1 mg/I NAA.52Infiltration media (IM): 0.5X Murashige Skoog basal salt mixture, 1X B5 vitamins(Gibco-BRL), 5.0% sucrose, 10 pg/I benzylaminopurine.Selection plates: At minimal media (pH 5.7 with 1M KOH), 0.7% agar, kanamycin (50mg/I).2.3.2 Root tissue cultureThe protocol of Valvekens et al., (1988) for transformation of Arabidopsis rootexplants was followed with the changes noted. Seeds (20-50) were first sterilized aspreviously described (section 2.1.3), transferred into a flask containing GM (50 ml),and placed on a shaker (50 rpm, 22°C, CL). After two to three weeks seedlings wereremoved from the flask and roots were excised. Roots were spread on platescontaining CIM, the plates were taped with porous surgical tape (Micropore), andincubated in a growth chamber for 3 days (22°C, CL). These were then infected withAgrobacterium by transferring to a tube containing liquid CIM (10 ml) and bacterialculture (0.5 ml), and incubating for 5 mm with gentle mixing, washed with liquid CIM,blotted, transferred onto fresh CIM plates, and placed back in the growth chamber.After 48-72 hrs roots were washed with liquid CIM and carbenicillin (500 mg/I) toremove excess Agrobacterium, cut into 0.5-1 cm segments, transferred in a clumpon plates containing SIM, carbenicillin (500 mg/I), and kanamycin (50 mg/I), andplaced back in the growth chamber for two to three weeks. Once green calli startedto appear (approximately 10 days later), they were cut away from the dead explant53material and transferred weekly to fresh SIM plates which facilitated shootdevelopment. When rosette leaves developed on the regenerated shoot, the rosetteswere excised and transferred to plates containing RIM and kanamycin (50 mg/I).Soon after bolting, transformants were placed into glass tubes capped with a poroussponges where they grew further, rooted, and eventually set seeds.2.3.3 In planta vacuum infiltrationThe in planta transformation technique I used is a àombination of in planta(Chang et al., 1994; Katavic et al., 1994) and its vacuum infiltration modification(Bechtold et al., 1993). Seeds (10-20) were planted in pots (4.5”) that were firstplaced at +4°C for 48 hrs and then transferred to a growth chamber (22°C, CL).When primary bolts achieved 1.5-2 cm in height, the inflorescences were cut off andthe wounds were inoculated with Agrobacterium culture. Plants were grown furtheruntil multiple secondary bolts appeared. These inflorescences were cut off again andplants were then exposed to vacuum infiltration. The Agrobacterium culture used forvacuum infiltration was prepared by adding saturated bacterial culture (grown for 3days in 25 ml of LB media) to 400 ml of LB containing appropriate antibiotics. Theculture was grown for additional 24 hrs (until O.D.= 1.5-2.0). Cells were harvestedby centrifugation (5,000 g for 10 mm at RT) and resuspended in lM (1-1.5 I). Eachpot with wounded plants was submerged into a petri dish (15-cm diameter)containing bacterial suspension (200 ml), and placed within a bell jar. A vacuum was54applied for 5 mm, or until bubbles started to form, and then rapidly released.Following the vacuum infiltration, treated plants were covered with saran-wrap for first24 hrs and placed back in growth chamber. The vacuum infiltrated plants (T0transformants) were grown until their seed was ready to harvest. Those T1 seedswere analyzed for plants that were KanR (T1 KanR transformants) as follows. T1seeds were sterilized as usual (section 2.1.3) and then evenly spread on sterilecheesecloths overlying the selection plates. Approximately 1500 T1 seeds wereplated on a single plate, plates were taped with porous surgical tape (Micropore),incubated at 4°C for 48 hrs, and transferred to a growth chamber (22°C, CL). T1KanR transformed plants were recognized as seedlings with green leaves anddeveloped root system after 10-14 days on selection plates. They were transferredinto soil, grown to maturity, and single-plant harvested.553. CHARACTERIZATION OF THE BELl MUTANT3.1 IntroductionAs discussed in Chapter 1, morphological, genetic and molecular studies offloral mutants significantly contributed to our understanding of mechanisms regulatingflower development in Arabidopsis. These studies identified genes involved in thecontrol of flower development, their regulatory functions and interactions. In contrast,very little is known about the genetic mechanisms regulating development of highlyspecialized reproductive structures, the ovules, and formation of the femalegametophyte. Thus, a similar approach using Arabidopsis female sterile mutantsdefective in ovule development, has been undertaken to identify genetic mechanismsinvolved in this developmental event.The ontogeny, morphology and anatomy of wild-type Arabidopsis ovules hasbeen examined first to serve as a basis for comparison to mutant ovules. Theanalyses of ovule development in two novel Arabidopsis mutants, Female sterilemutant 1 (Fmsl) and Fruitless 1 (Ftsl), as well as the previously characterized floralmutant Ap2, are presented. The specific roles of these genes in Arabidopsis ovuledevelopment, their interactions, and correlation with the organ-identity gene AG havebeen investigated, and are discussed in detail.563.2 Results3.2.1 Isolation and genetic analysis of new bell allelesTo isolate mutations that affect ovule development, approximately 8000Arabidopsis (Wasselewskija ecotype) lines transformed with the T-DNA ofAgrobacterium (Feldmann, 1991) were screened for mutations that reduce fertility onthe basis of reduced silique size. Female-sterile mutants were identified usingreciprocal crosses with the wild type. Siliques from these female sterile mutants werethen dissected to identify possible abnormalities in ovule development. Two lines,designated female sterile (Fmsl) and fruitless (Ftsl), had similar phenotypes, asdescribed below.To determine the genetic basis of the mutant phenotypes, both Fmsl and Fts2mutant plants were used as a male parent in a cross to a wild-type plant. All of theF1 progeny examined had the fertile wild-type phenotype. F2 progeny from the crosswith Fms yielded 1569 fertile and 570 sterile plants (3:1, X2=3.0981, P>0.05). F2progeny from the cross with Fts yielded 2324 fertile and 748 sterile plants (3:1,X2=0.6944, P>0.25). These results indicated that both sterile Fmsl and Ftslphenotypes are due to single recessive nuclear mutations.A trans complementation tests were performed to determine if the Fmsl andFtsl phenotypes are the result of mutations in the same gene. Reciprocal crossesinvolving a male parent homozygous for one mutation and a female parent57heterozygous for the other mutation were performed. The F1 progeny from FTS1/ftslx fmsl/fmsl were 42 fertile to 38 sterile (1:1, X2=O.2, P>O.5), whereas the F1 progenyfrom FMS1/fmsl x ftsl/ftsl were 88 fertile to 87 sterile (1:1,X2=O.005, P>O.9). Theseresults indicated that fmsl and ftsl are alleles. While my research was in progress,a female sterile mutant, generated by ethyl methanesulfonate mutagenesis, with aphenotype similar to the Fmsl and Ftsl was described (designated Bell -1, Robinson-Beers et aL, 1992). To determine the genetic relationship between Bell and Fmsl,we crossed female BELl/bell x tmsl/fmsl male. Among 35 F1 progeny tested, 19were fertile and 16 were sterile (1:1, X2=O.26, P>O.5), indicating that bell and fmslrepresent different alleles of the same genetic locus. For this reason fmsl and ftslhave been designated as bell-2 and bell-3, respectively.As mentioned above, Bell-2 and Bell-3 mutant plants were isolated from acollection of Arabidopsis plants transformed with the T-DNA of Agrobacterium(Feldmann, 1991). The T-DNA carries a marker gene NPTII which confers KanR(dominant trait). To determine if the Bell-3 mutant phenotype cosegregated with aT-DNA insert, I analyzed F2 progeny of the cross wild type x BeIl-3. The F2 progenysegregated in the ratio 1488 KanR : 461 KanS (3:1,X2=1.8854, P>O.1), suggestingpresence of a single T-DNA insert. Furthermore, KanR plants segregated for theBell -3 phenotype in a ratio of 494 wt: 214 Bell-3 (2:1,X2=3.076, P>O.05), indicatingthat the Bell-3 mutant phenotype is linked to the T-DNA insert. Two different wayswere used to test that all Bell-3 mutant plants (beIl-3/bell-3) are also homozygousfor the T-DNA insert (NPTII/NPTII). Six different Bell-3 mutant plants from the F258progeny (cross wt x Bell-3) were crossed to wild-type plants, and all F1 progenyexamined were KanR (157 plants). Since the probability that all F1 progeny (from 6Bell-3 plants) are KanR if the mutant phenotype and the T-DNA are not linked is1/1 000 (probability of Bell-3 mutant plant being homozygous for KanR = 1/3 and 6plants were used, thus (1/3)6=0.001), this data supported linkage of the Bell-3mutant phenotype and the NPT1I gene. The linkage of the Bell-3 mutant phenotypeand the T-DNA was also determined by following the segregation of the NPTII gene(PCR) in F2 progeny of the cross wt x Bell-3. The F2 progenies segregated for NPTIIin a ratio of 44 NPTII: 16 NPTII (3:1,X2=0.086, P>0.75). Furthermore, all 25 Bell -3mutant plants from the F2 progeny showed presence of NPTII. Since the probabilitythat all Bell-3 mutant plants will show the presence of the NPTII gene if these twotraits are not linked is 7/1 0000 (probability of an F2 plant being NPTII = 3/4 and 25plants were used, thus (3/4)25=0.0007), these data confirmed the proposed linkageof the BeIl-3 mutant phenotype and the T-DNA.The map position of the BELl was determined by genetic crosses to markerlines W100 (Koornneef et al., 1987) and TC35. Data from 189 F2 progeny obtainedfrom the cross W100 x Bell-3 indicated that the BELl was genetically linked to themarker transparent testa (TT3) located on the chromosome five. A more accuratemap position for the BELl was obtained by analyzing 749 F2 progeny from the crossTC35 x Bell-3 which indicated linkage of the BELl to TTG (23.24 ± 3.4 map units)and to POM (19.52 ± 3.4 map units). Figure 3.1 illustrates the chromosomal positionof BELl with respect to the current linkage map of Arabidopsis.59ttg bell pgm yiI I I(35.5) (71.1) (93.3)23.3+/-3.4 19.5+/-3.435.3+/-3.1 24.0+/-3.4Figure 3.1. Position of BELl on chromosome five of Arabidopsis. Arrows spandistances between genetic markers and the numbers beneath the arrows representdistance and standard error in map units. Numbers in parentheses represent thechromosomal location of the markers from the current linkage map of Arabidopsis.3.2.2 Wild-type Arabidopsis ovuleI examined first the ontogeny and morphology of wild-type ovules usingscanning electron microscopy (SEM, Figure 3.2) to serve as a basis for comparisonto mutant ovules. Leonore Reiser analyzed the anatomy of ovules using lightmicrocopy (LM, Figure 3.3). The stages within ovule development were correlatedwith other events in floral development (Smyth et al., 1990).Stage 8 of Arabidopsis flower development is characterized by the formationof an open gynoecial cylinder (Figure 3.2A). At this time, four files of ovule primordiaare initiated, two from each of the placental tissues located at the carpel margins(Figure 3.2B). Ovule primordia initiate asynchronously; initiation starts in the lowerpart and progresses toward the opening of:the gynoecial cylinder. Later events inovule development appear to be more synchronous. The gynoecial cylindercontinues to grow as an open tube, although its apex becomes tapered (stage 9,Figure 3.2D). The central septum forms and divides the cylinder into two locules,60and the two files of enlarging ovule primordia in each locule display an interlockingpattern (Figure 3.2E). Within each primordium, the megasporocyte is recognized byits size and can be found directly below the apex of the nucellus (Figure 3.3A).Complete closure of the gynoecial cylinder at the top and the appearance ofimmature papillae cells indicates the beginning of stage 10 (Figure 3.2G). Initiationof the inner integument is marked by the cell outgrowth at the base of the nucellarregion (Figure 3.2H). As the stigmatic papillae develop at the tip of the gynoecium(stage 11, Figure 3.2J), the nucellus and funiculus can easily be distinguished on thebasis of both morphology and cell surface features. They are separated by two ringsof cells comprising the inner and outer integument primordia (Figure 3.2K). Withinthe nucellus, the megasporocyte enters meiosis (Figure 3.3C).By stage 12, megasporogenesis has resulted in the formation of a linear tetradin which only the chalazal megaspore survives (Figure 3.3E). During the same periodboth integuments exhibit intensive growth that can be divided into substages. Inearly stage 12, the outer integument grows asymmetrically through increased celldivisions on the side facing the central septum (Figure 3.2N). Growth of the innerand outer integuments continues upward, enclosing the nucellus (Figure 3.20). Thenucellus changes its orientation with respect to the funiculus from being parallel tobeing perpendicular (anatropous ovule). Growth of the integuments contributes toovule enlargement. Concurrently, the ovary expands, the stylar region elongates, andtypical style cells differentiate below the stigmatic papillae (stage 12, Figure 3.2M).At ovule maturity (stage 13), the gynoecium possesses well differentiated61stigmatic papillae, style, and ovaries (Figure 3.2R). The outer integument completelyovergrows the inner integument, forming the micropylar opening (Figure 3.2S). Thenucellus degenerates and the embryo sac is appressed by a newly differentiated celllayer of the inner integument, the endothelium. Within the embryo sac,megagametogenesis is completed with the formation of seven-celled, eight-nucleatefemale gametophyte (Figure 3.3G). No obvious morphological changes ofgynoecium and unfertilized ovules occurred after stage 13.Figure 3.2. SEM of developing wild-type and Bell ovules and correspondinggynoecial morphology from stage 8 to stage 13 of Arabidopsis flower development.Structures have been given the following abbreviations: nucellus (flu), funiculus (fu),inner integument primordium (up), outer integument primordium (oip), innerintegument (ii), outer integument (oi), integument-like structure (ils).A) Open gynoecial cylinder of a stage 8 wild-type flower. Magnification x257.B) Single carpel of gynoecial cylinder of a stage 8 wild-type flower. Two files of ovuleprimordia are initiated along carpel margins. Magnification x1179.C) Ovule primordia within dissected gynoecial cylinder of a stage 8 Bell flower.Magnification x760.D) Constricted gynoecial cylinder of a stage 9 wild-type flower. Magnification xl 51.E) Side view into gynoecial cylinder of a stage 9 wild-type flower. Note staggeredarrangement of enlarged ovule primordia. Magnification x720.F) Enlarged ovule primordia within gynoecial cylinder of a stage 9 Bell flower.Magnification x827.G) Closed gynoecial cylinder of a stage 10 wild-type flower. Magnification xl 20.H) Elongated ovule primordia within gynoecial cylinder of a stage 10 wild-type flower.Primordia differentiate into funiculus and nucellus. Arrow indicates initiation of innerintegument primordium. Magnification x859.I) Ovule primordia within gynoecial cylinder of a stage 10 Bell flower. The nucellusand funiculus have differentiated. Arrow indicates initiation of outer integument.Magnification x556.J) Gynoecium with stigmatic papillae of a stage 11 wild-type flower. Magnificationx108.K) Developing ovules within gynoecium of a stage 11 wild-type flower. Arrowsindicate the inner and outer integument primordia at the base of nucellus.Magnification x647.L) Developing ovules within gynoecium of a stage 11 Bell flower. Arrow indicatescollar tissue at the base of the nucellus. Magnification x394.62M) Gynoecium with well-developed stigmatic papillae and style of a stage 12 wild-type flower. Magnification x76.N and 0) Ovules within gynoecium of a stage 12 wild-type flower. Arrows indicatethe inner and outer integuments enclosing the nucellus. Magnification x603 for (N)and x508 for (0).P and Q) Ovules within gynoecium of a stage 12 Bell flower. Arrows indicateintegument-like structure (ils) around the nucellus. Magnification x299 for (F) andx352 for (Q).• - • -- 0 -•163R) Fully developed gynoecium with stigmatic papillae, style, and ovaries of a stage13 wild-type flower. Magnification x56.S) Mature ovules within gynoecium of stage 13 wild-type flower. Arrow indicatesmicropylar opening. Magnification x243.T) Ovules within gynoecium of a stage 13 Bell flower. Arrow indicates the nucellus,which is not completely enclosed by integument-like structures. Magnification x240.64Figure 3.3. LM of developing wild-type and Bell ovules. Structures have been giventhe following abbreviations: central septum (cs), funiculus (fu), inner integumentprimordium (lip), outer integument primordium (oip), inner integument (ii), outerintegument (oi), integument-like structure (ils).1-I65A) Ovule primordium of a stage 9 wild-type flower. Arrow indicates megasporocyte.Magnification xl 400.B) Ovule primordium of a stage 9 Bell flower. Arrow indicates megasporocyte.Magnification x21 20.C) Developing ovule of a stage 11 wild-type flower. The inner and outer integumentprimordia have been initiated. The premeiotic megasporocyte is apparent within thenucellus. Magnification xl 720.D) Developing ovule of a stage 11 Bell flower. Arrows indicate the position wherethe inner integument primordium is expected to form; arrowhead indicates the collartissue at the position of the outer integument primordium. Magnification xl 840.E) Median section of ovule of a stage 12 wild-type flower. Arrowhead indicatesdensely staining cells of the linear tetrad that are degenerating; the arrow indicatesthe surviving chalazal megaspore. Magnification xl 500.F) Ovule with integument-like structure of a stage 12 Bell flower. Arrowheads andarrows indicate degenerating cells of linear tetrad and surviving chalazal megasporerespectively. Magnification xl 880.G) Oblique section through mature ovule of stage 13 wild-type flower. Arrowindicates the egg cell in contact with synergids; arrowhead indicates the polar nuclei.Antipodal cells are not visible in this section. Magnification xl 760.H) Ovule of a stage 13 Bell flower. Arrows indicate two postmitotic cellsrepresenting the two-nucleate embryo sac. Magnification x21 Bell mutant ovule developmentThe results on ontogeny, morphology, and anatomy of Bell ovules are shownin Figures 3.2 and 3.3, and summarized in Table 3.1, Relative to the wild type, Bellovules do not show any deviation during early development (stages 8 and 9). Bellovule primordia are initiated from the placental tissue and enlarge at the same timeand in the same place as the wild type (Figure 3.2C and 3.2F). Concomitantly, themegasporocyte below the apex of the nucellus differentiates (Figure 3.3B).Differences between Bell and wild-type ovule morphogenesis are firstdiscernable at inner integument initiation (stage 10). No structural changes wereobserved at the expected position of the inner integument primordium, suggesting66Ovule8Opencylindar(60-80urn)Ovuleprimordiainitiation9CylinderconstrictedatapexOvuleprirnordiaenlarge(300urn)10Closedcylindar(350urn)Integumentprirnordiaformed11StigmaticpapillaedevelopIntegumentprimordia(425-550urn)growth12Styledevelops(650-850urn)Integumentsovergrownucellus1 3FullyformedgynoeciumwithIntegumentdevelopmentstigma,style,ovaries(950urn)complete,micropyleformed17Greensilique(>7mm)SeedcQatformationSinglecollarofcellsformsbelowthenucellusProliferationofcollarcellsIntegument-likestructuredevelopsContinuedgrowthofintegument-likestructureCarpel-likestructuresformoccasionallyTableStage3.1.Wild-typeandBellOvuleDevelopment.MorphologyofgynoeciumWildtypeBellGametophyteOvuleGametophyteNCNCNCMegasporocytevisibleMeiosisbeginsMeiosiscompletedMegagametogenesisMatureembryosacEmbryogenesisNC NC NC EmbryosacdevelopmentarrestsAbsenceoffunctionalembryosacNC=denotesnochangefromthewildtypethat the inner integument is not initiated. A single outgrowth, represented by adisorganized cluster of cells, arises at a time and position corresponding to that ofthe wild-type outer integument primordium (Figure 3.21). By stage 11, the putativeouter integument primordium enlarges and generates an irregular collar of tissue atthe base of the nucellus (Figure 3.2L). Asymmetric growth of the collar of tissueforms individual protuberances (stage 12, Figure 3.2P). These cell outgrowthssurround the nucellus, generating an integument-like structure (Figure 3.2Q). Atstage 13, Bell ovules have a distinct morphology because the integument-likestructure surrounds but does not enclose the nucellus (Figure 3.2T).Embryo sac development is also affected in Bell mutant ovules. Developmentof the female gametophyte of Bell ovules proceeds as in wild-type ovules, with theformation of the linear tetrad of megaspores (Figure 3.3F). Shortly afterwards,however, embryo sac development is arrested and results in a reduced number ofpostmitotic cells and remnants of degenerated megaspores (Figure 3.3H).3.2.4 Bell ovules develop carpel-like structuresDuring stage 17 of wild-type flower development, floral organs senesce andabscise, and unfertilized ovules degenerate. Figure 3.4 shows that Bell ovules canhave one of two different fates. Most Bell ovules, like the unfertilized wild-typeovules, degenerate (Figure 3.4A). However, in some Bell ovules, cells from theregion of the integument-like structure continue to grow and differentiate into a68a) (0Figure 3.4. Carpel-like structures and carpel-sepal ovules in Bell and Ap2-6 mutantsiliques, respectively. Structures have been given the following abbreviations:funiculus (fu), ovule (ov), filament (f), carpet-like structure (cls), carpel-sepal ovule(cso), ovary (ovr), style (sty), stigmatic papillae (sp).A) Dissected stage 17 mutant silique containing degenerating Bell ovules.Magnification xl 51.B) Stage 17 Bell ovule developing into carpel-like structure while two adjacent ovulesare degenerating. Magnification xl 50.C) Developed carpel-like structure within a stage 17 Bell mutant silique. Cell typescharacteristic of the carpet, including stigmatic papillar, stylar, and ovarian cells, havedifferentiated. Magnification x118.D) Carpet-like structure attached to placenta by a funiculus within a stage 17 Bellmutant silique. Arrow indicates newly initiated ovule on carpel-like structure.Magnification xl 24.E) Dissected stage 17 Bell silique from a population segregating for Wasselewskijaand Landsberg erecta genetic backgrounds. Numerous Bell carpet-like structurescan be observed. Magnification x60.F) Dissected gynoecium of a stage 13 Ap2-6 mutant flower containingmorphologically normal ovules, carpet-sepal ovules, and filaments. Magnification x51.G) Termination of the apical meristem of the Bell-3 primary shoot in a mass ofcarpetloid structures. Magnification x35.H) Terminal differentiation of the Bell-3 inflorescence in a pistil-like structure.Magnification x8l.structure with carpet-like morphology (Figure 3.4B) and cell types (Figure 3.40). Insome cases, second-order mutant ovutes are produced along the margins of thecarpet-like structure (CLS, Figure 3.4D). In all cases, the carpet-like structureseventually degenerated.As shown in Table 3.2, growth conditions and genetic background influencedthe frequency of the carpet-like structures. In the WS ecotype, under standardgrowth conditions (22°C, CL), 63% of the beIl-3 homozygous plants displayed carpetlike structures, with one to two structures per silique (294 carpet-like structures per228 sitiques). In contrast, carpet-like structures were not observed in the WS ecotypewhen plants were grown at 16°C or when the daytength was reduced from 24 to 1070or 16 hrs. The frequency of carpel-like structure formation was found to differ in F2populations generated by crossing Bell-3 to wild type plants of ecotypes Landsbergerecta (Ler) and Columbia (Cot). The most dramatic increase in frequency wasobserved in the outcross to Ler, where 90% of the F2 plants homozygous for beIl-3displayed carpet-like structures under standard growth conditions, and as many as10 structures were observed per silique (Figure 3.4E). In all ecotypes tested,reducing the temperature or daylength also reduced the frequency of carpet-likestructure formation (Table 3.2).Temperature Daylength Genetic Background(light/dark) WS F2 WS x Col F2 WS x Ler22 C 24 hours 63% (164) 50% (32) 90% (97)16 C 24 hours 0% (38)•.’ 15% (21) 70% (45)22 C 10/14 hours 0% (25) 0% (05) 0% (22)22 C 16/8 hours 0% (33) ND NDTable 3.2. Effects of growing conditions and genetic background on the percentageof Bell plants having carpet-like structures in stage 17 siliques. Number in bracketsindicates the total number of Bell mutant plants examined. Four siliques per plantwere analyzed. WS = Wasselewskija, Cot = Columbia, Ler = Landsberg erecta, ND= not determined.3.2.5 Mutations in BELl affect inflorescence terminationIn addition to the ovule defects, both Betl-2 and BeIl-3 inflorescences oftenterminate in carpeltoid organs (Figure 3.4G) that are commonly fused in a pistil-likestructure (Figure 3.4H). These pistil-like structures do not seem to result in earlytermination of the inflorescence since the number of nodes produced by BeIl-3primary shoots is not significantly different from that of wild type.713.2.6 Mutations in AP2 result in carpelloid ovulesRecessive mutations in the AP2 gene cause homeotic transformation ofperianth organs to reproductive organs and stamens to carpels, a reduction in organnumber in the first three whorls and abnormalities in gynoecial morphology (Komakiet al., 1988; Kunst et al., 1989; Bowman et al., 1991a). As Figure 3.4F shows, thegynoecium of the Ap2-6 flower contains morphologically normal ovules, thinfilaments, and structures with both carpel and sepal characteristics. The latterstructures, so called carpel-sepal ovules (cso), are analogous to the sepal-carpelorgans in the outermost whorls of the Ap2-6 mutant flower (Kunst et al., 1989).Carpel-sepal ovules and thin filaments may also form in the place of ovules on thesepal-carpel organs of Ap2 mutant flowers (see below).The abnormal Ap2-6 ovules bear no resemblance to wild-type ovules in eithermorphology or cell surface features (Figure 3.4F). In addition, the Ap2-6 carpel-sepalovules are recognizable by stage 13. Therefore, in contrast to Bell carpel-likestructures, the Ap2-6 carpel-sepal ovules develop early and do not appear to bederived specifically from the integument region.The Ap2-6 ovule abnormalities segregate together with the other characteristicsof the Ap2-6 mutant phenotype, suggesting that they are due to the ap2-6 allele. Inaddition, abnormal ovules are not specific to the ap2-6 allele since we have observedsimilar defects in another strong allele of AP2 mutant, ap2-7.723.2.7 Bell/Ap2 and Bell/Ap3 double mutant analysisThe bell mutation was introduced into plants homozygous for ap2 and ap3,and the resulting double mutant plant phenotypes (Beli/Ap2, BeIl-Ap3) wereanalyzed for two reasons. First, ap2 and ap3 mutations result in ovule-bearingcarpels in the first and third floral whorls, respectively (Bowman et al., 1989; Kunstet al., 1989). Thus, the double mutant phenotypes would allow us to determine if theBELl is required for development of ovules regardless of their position within aflower. Second, because both BELl and AP2 genes affect development of ovules,phenotypic analysis of the Bell/Ap2 double mutants can be useful in determining theextent of gene interaction. Putative double mutant plants were identified on the basisof a novel phenotype and were genetically confirmed (see Materials and Methods).Bell/Ap3. Flowers of plants homozygous for recessive mutations in AP3 have sepal-like organs in the second whorl and carpelloid organs in the third whorl. Dependingon the ap3 allele and growth conditions, the mutant phenotype varies in degree ofthird whorl carpelloidy (Bowman et al., 1989, 1991a). Ap3-1 flowers grown at 22°Chave six stamen-carpel or carpel-like organs in the third whorl. In contrast, thestamen-carpel intermediate organs in the third whorl of Ap3-3 flowers are typicallyfused to the fourth whorl gynoecium (Bowman et al., 1991a). Figures 3.5A and 3.5Bshow that morphologically normal ovules can develop in the third whorl of Ap3-1 andAp3-3 flowers, indicating that ap3-1 and ap3-3 have no obvious effect on ovulemorphogenesis.73Figure 3.5. SEM of stage 13 flowers from Ap3 mutants and Bell-3/Ap3 doublemutants.A) Single carpel-like organ with normal ovules from the third whorl of an Ap3-1 flower.Magnification x94.B) Ap3-3 flower with perianth partially removed. Third whorl carpels fuse to the fourthwhorl gynoecium, and morphologically normal ovules are apparent. Magnificationx50.C) Single carpel-like organ from the third whorl of Bell-3/Ap3-1 double mutant flower.Bell ovules develop at the carpel margin. Magnification x89.D) Dissected gynoecium of Bell-3/Ap3-3 double mutant flower. Bell ovules developwithin the gynoecium, and filaments arise from the third floral whorl (perianthremoved). Magnification x65.In Bell-3/Ap3-1 and Bell-3/Ap3-3 double mutantflowers only Bell-type ovulesdevelop on the third whorl organs and within the gynoecium (Figure 3.5C and 3.5D).In the later stages of flower development (Le., stage 17) the carpel-Iike structurescharacteristic for Bell mutants were observed in siliques of both double mutants.Thus, the phenotype of both Bell-3/Ap3-1 and Bell -3/Ap3-3 double mutants indicatedthat the BELl gene is required for normal ovule development in the third whorl74carpels.Bell/Ap2. A weak AP2 allele, ap2-5, causes sepal-to-carpel and petal-to-stamentransformations in the first and second whorls, respectively. As shown in Figure 3.6A,morphologically normal ovules develop on the first whorl sepal-carpel organs of Ap2-5 flowers. However, some of the structures developing in the position of ovules arecarpel-like, judging from the cell surface features (Figure 3.6A). In addition, they haveone of two general morphologies that were described previously for the gynoeciumof Ap2-6 flowers: a carpel-sepal ovule or a thin filament often ending with stigmaticpapillae. The first whorl organs of Bell-3/Ap2-5 double mutant flower bear Bell-typeovules and carpel-sepal ovules typical for Ap2 (Figure 3.6B). At stage 13 of flowerdevelopment, the fourth whorl gynoecium of Bell-3/Ap2-5 double mutant flowerscontains Bell ovules and carpel-sepal ovules (Figure 3.6C), and the former one bystage 17 occasionally develop into carpel-like structures typical for Bell.A strong AP2 allele, ap2-6, causes sepal-to-carpel, petal-to-carpel, andoccasionally stamen-to-carpel transformations and a reduction in organ number inthe first three whorls (Kunst et al., 1989). The first and second whorls of an Ap2-6flower are occupied by sepal-carpel organs bearing morphologically normal ovules,deformed ovule-like structures, carpel-sepal ovules and thin filaments (Figure 3.6D).As shown on Figure 3.6E, the first whorl organs of Bell -3/Ap2-6 double mutant flowerbear abnormal ovule types characteristic of Bell and Ap2 and morphologicallynormal ovules were never observed (Figure 3.6E). The fourth whorl gynoecium ofBell-3/Ap2-6 double mutant flowers typically contains both Bell-type ovules and a75Figure 3.6. SEM of stage 13 flowers of Ap2 mutants and Bell-3/Ap2 double mutants.Structures have been given the following abbreviations: ovule (ov) and carpel-sepalovules (cso).A) Carpel-sepal organ from the first whorl of Ap2-5 flower. At the margins, normalovules and carpel-sepal ovule develop. Magnification x137:B) Portion of a carpel-sepal organ from the first whorl of Bell-3/Ap2-5 double mutantflower. Arrow indicates a degenerating Bell ovule. Magnification x85.76C) Upper part of dissected Bell-3/Ap2-5 double mutant gynoecium. Arrow indicatesBell ovule, arrowhead indicates developing carpel-like structure. Magnification xl 52.D) Portion of a carpel-sepal organ from the first whorl of an Ap2-6 flower.Morphologically normal ovules and several developing carpel-sepal ovules areapparent. Magnification xl 07.E) Carpel-sepal organ from the first whorl of a Bell-3/Ap2-6 double mutant flower.Arrow indicates developing Bell ovule. Magnification x120.F) Dissected gynoecium of Bell-3/Ap2-6 double mutant flower exhibiting a few Bellovules (arrow) and a novel ovule-type distinct in morphology and cell structure fromeither Ap2-6 carpel-sepal ovules and the Bell carpel-like structures (arrowhead).Magnification x80.novel class of carpelloid ovules (Figure 3.6F). These novel carpelloid ovules aredistinct from either Ap2-6 carpel-sepal ovules or Bell carpel-like structures. Theyusually have a flat, leaflike shape, and are comprised of cells whose shape andsurface features highly resemble ovary cells. Differentiation of other cell types ofcarpel that are typical for Bell carpel-like structures and Ap2-6 carpel-sepal ovules,such as stigmatic papillar and sepaloid cells, respectively, have never been observedon Bell-3/Ap2-6 carpelloid ovules. Since double mutant analysis did not revealepistasis between BELl and AP2, both genes likely function independently withinovules, suppressing carpel development. To determine whether the novel phenotypewas specific to the Bell-3/Ap2-6 double mutant and not due to other differences ingenetic background, we crossed wild-type Ccl x beIl-3 (WS) and wild-type WS x ap2-6 (Ccl). The novel ovule morphology of Bell-3/Ap2-6 double mutants was notobserved among the progeny of either cross.Taken together, double mutant analyses of the beil-3 and alleles of AP2, ap2-5and ap2-6, suggested that BELl gene is required for development of morphologicallynormal ovules in the outer floral whorls of the Arabidopsis flower. Further, the novel77carpelloid ovules observed in the Bell-3/Ap2-6 double mutants indicated that BELlfunction is not dependent on AP2 expression, and vice versa, in the ovule.3.3. Discussion3.3.1 BELl is required for ovule developmentI have characterized the phenotype of female-sterile mutant of Arabidopsis,Bell, which is defective in ovule development and is the result of recessive mutationof the BELl gene. Phenotypic comparison of two alleles, bell-2 and bell-3, did notreveal any significant differences between them, thus, most of the mutant analyseshave been done with single allele beli-3. Based on the recessive nature of the bell-2and bell-3 alleles, and the fact that another recessive allele bell-i with a similarphenotype has been identified (Robinson-Beers et al., 1992), I will assume that thesealleles represent loss-of-function of the BELl gene.The phenotypic analyses of Bell and the double mutants support the followinggeneral conclusions concerning the function of the BELl in wild-type Arabidopsisflowers:1) BELl is necessary for development of morphologically normal ovules, regardlessof the floral whorl in which they appear. This suggests that the action of BELl issubordinate to carpel formation.2) BELl does not appear to be required for the initiation of ovule primordia and early78development of ovules because the ontogeny of Bell ovules is normal until the timeof integument initiation. Normal integuments are not found in Bell ovules, indicatingthat BELl is required for integument development. One possibility is that the BELlmay play a specific, direct role in development of integuments or alternatively, theBELl gene may be needed for ovule development to progress beyond stage 10,thus, having an indirect influence on the development of integuments.3) BELl gene product plays a role in development of the female gametophyte.Arrested megagametogenesis was observed in all sectioned Bell ovules. This defectcould be a pleiotropic effect resulting from the absence of integuments, or the BELlgene may have a more direct role in the embryo sac development.3.3.2 Bell carpel-like structuresThe most intriguing aspect of the Bell mutant phenotype is homeotictransformation of the outer integument into a carpel-like structure (from stage 13 to17). This Bell mutant phenotype suggests possible uncontrolled expression of classC gene AG within mutant ovules. To date, the hypothesis that BELl gene is involvedin negative regulation of AG during ovule development has been supported by twoindependent pieces of evidence. First, in situ hybridization analyses havedemonstrated spatial distribution of AG expression throughout Bell ovules at stage13 of flower development (provided by G. Haughn; Modrusan et al., l994b),suggesting suppression of AG transcription by BELl late in wild-type ovules.79Second, constitutive expression of an AG transgene in Arabidopsis plants resulted inan ovule morphology similar to that of Bell mutant plants (Ray et al., 1994),indicating further that BELl itself is involved in negative regulation of AG expressionin Arabidopsis ovules.The development of carpel-like structures is variable such that only someovules on some plants possess these structures. Moreover, the frequency of carpellike structures per silique and per plant is influenced by environmental conditions andgenetic background. Interestingly, the environmental conditions that favourdevelopment of carpel-like structures (long photoperiod and high temperatures) alsopromote flowering in general (Martinez-Zapater et al., 1995). Furthermore, of thethree ecotypes examined, the one that flowers the earliest (Martinez-Zapater et al.,1995), Ler, is also the one that has the highest frequency of carpel-like structures.Together, these data suggested that development of carpel-like structures isdependent on the strength of the floral inductive signal. A similar correlation betweenfloral inductive conditions and the degree of carpelloidy has been observed in severalother Arabidopsis mutants with altered floral morphology, and it was attributed to theinfluence of inductive conditions on the activity of class C organ-identity gene(s)(Schultz and Haughn, 1993). In the case of Bell mutants, the formation of carpel-likestructures also depends on the level of class C gene, AG, which can vary from ovuleto ovule. In general, floral inductive conditions which tend to increase AG level alsopromote formation of carpel-like structures.803.3.3 BELl influences intlorescence structureMutations in BELl result in the termination of the apical meristem of theprimary shoot in a carpelloid structure, a phenomenon recently described as terminaldifferentiation (Hensel et al., 1994). Analogous aberrant termination of inflorescencemeristems have been observed in male sterile (Chaudhury, 1993; Hensel et al., 1994),reduced-fertility (Hensel et al., 1994), and flower-development mutant plants (Hualaand Sussex 1992; Schultz and Haughn, 1993). The intlorescence defects in floralmutants (i.e., Lfy; Huala and Sussex, 1992; Schultz and Haughn, 1993) andtransgenic Arabidopsis plants in which AG is constitutively expressed (Mizukami andMa, 1992) have been primarily attributed to ectopic expression of AG in the primarymeristem. Similarly, BELl may negatively regulate carpel development ininflorescence meristems directly, as a regulator of AG, or indirectly, as a sterile plant.The sterility has been correlated with the terminal differentiation in the male sterilemutant msl-l as well as de-truited wild-type plants (Hensel et al., 1994). Based onthese studies, the alternative fate of inflorescence differentiation appears as a resultof absence of a “communication system” between developing siliques andinflorescence tissue, typical of male and female sterile plants. Although both sterilityand altered expression of AG are associated with disruption of inflorescencedifferentiation, it seems likely that the former one influences the strength of the floralinduction which in turn affects development of the primary shoot (Schultz andHaughn, 1993; Hensel et al., 1994).813.3.4 AP2 gene is involved in ovule morphogenesisThe role of the AP2 gene in perianth organ development (Komaki et at., 1988;Kunst et at., 1989; Bowman et al., 1989, 1991a) and in the initiation of flowerdevelopment (Irish and Sussex, 1990; Shannon and Meeks-Wagner 1993; Schultzand Haughn, 1993) has been well studied. TheAP2 gene is also required for normaldevelopment of the reproductive organs (Komaki et al., 1988; Kunst et al., 1989;Bowman et at., 1991a; Schultz et at., 1991), ovules (Modrusan et al., 1994b), andseeds (Jofuku et al., 1994).Abnormal Ap2-6 ovules do not resemble wild-type ovules either in cell surfacefeatures or morphology, although both of them appear to develop during the sameperiod of flower development (stages 8 to 13). Thus, as in floral organ development,AP2 seems to function relatively early in ovule development. The carpel-tike featuresof the abnormal Ap2-6 ovutes suggested that the AP2 gene is involved in regulationof AG expression early in ovule development. This hypothesis is consistent with theknown role of AP2 in inhibiting expression of AG in the outer two whorls during flowerdevelopment (Kunst et al., 1989; Bowman et al., 1991a; Drews et at., 1991).Morphologically normal ovules are also found in the Ap2-6 gynoecium. Theirdevelopment may be due to remaining AP2 activity in the Ap2-6 mutant.Alternatively, the role of the AP2 gene may be to limit the choices of the ovuleprimordium to a single developmental program in the wild-type plant. If this weretrue, the absence of AP2 activity in the Ap2-6 mutant would allow the ovule primordia82to commit to one of several fates, only one of them being development ofmorphologically normal ovules.The interaction between AP2 and BELl genes during ovule development hasbeen tested through double mutant analyses. The fact that a novel type of carpelloidovules was observed in the Bell-3/Ap2-6 gynoecium indicated that BELl activity ispresent in the Ap2-6 ovules and thatAP2 activity is present in Bell ovules. Therefore,the AP2 function in ovules does not appear to depend on BELl, and vice versa.Based on Bell-3/Ap2-6 double mutant phenotype where novel carpelloid ovules cannot be described as more carpel-like than those of Bell or Ap2-6, it has been difficultto assess the relative contribution of BELl and AP2 to the negative regulation of thecarpel program during ovule development.3.3.5 Role of organ-identity genes in ovule developmentCombination of three classes of organ-identity genes (A,B, and C) withoverlapping expression domains specify the floral organ fates in the Arabidopsisflower (Coen and Meyerowitz, 1991; Okamuro et al., 1993; Ma, 1994; Weigel andMeyerowitz, 1994). As discussed above, the class A gene AP2 appears to play a rolein ovule development. In addition, ectopic expression of class C genes clearly haseffects on ovule development (Mandel et al., 1992a; Ray et al., 1994). By contrast,the role of class B genes during wild-type ovule development is not known.To date, in situ hybridization experiments showed that the AP3 and AG genes83are transcribed in very specific subsets of cells of mature wild-type ovules (Jack etal., 1992; Bowman et al., 1991b, Modrusan et al., 1994b). Moreover, in developedovules (stage 13), the relative position of the AP3 transcripts with respect to AGtranscripts, is analogous to what is seen in the early floral meristem. This similarityin expression pattern of organ-identity genes in the flower and ovule suggest that theovule, like the flower, uses the cues from organ-identity genes to activate appropriatedevelopmental programs. Despite these observations, the function of organ-identitygenes in ovule development is not fully understood. Ovules do not develop on Agmutants (Bowman et al., 1989), they are abnormal on strong Ap2 mutants (Modrusanet al., 1994b), yet morphologically normal ovules are found on Ag/Ap2 doublemutants (Bowman et al, 1991b) and on strong Ap3 mutants. Thus, if indeed theorgan-identity genes play a role in ovule development, then either their role is asubtle one or there is enough gene redundancy to allow normal ovulemorphogenesis in their absence.844. CHARACTERIZATION OF THE BELl GENE4.1 IntroductionAs described in the previous Chapter, I have demonstrated that the beIl-3mutant allele is T-DNA tagged with the NPTII gene. Our collaborator, Leonore Reiserin Dr. Fischer’s group (University of California, Berkeley) used Southern blot analysesto show that the beIl-2 mutant allele is also T-DNA tagged. Two T-DNA taggedalleles, beIl-2 and beIl-3, facilitated cloning of BELl which was undertaken toinvestigate the function of the gene at the molecular level. A putative BELl clone wasisolated by Leonore Reiser and further molecular analyses have been donecoordinately by the laboratories of Dr. Fischer and Dr. Haughn. The main focus ofmy work on BELl was to confirm the identity of the cloned sequence bycharacterizing the beIl-3 allele and complementing the Bell mutant phenotype.Throughout this work, I was assisted by Dr. Alon Samach, whose invaluablecomments have been greatly appreciated. Other aspects of the molecularcharacterization of BELl, obtained by colleagues in Dr. Fischer’s group, provide anecessary background for results described here. Thus, they are briefly summarizedbelow.4.1.1 Isolation of the BELl gene85BELl was cloned by screening a Bell-2 mutant library with the right bordersequence of the T-DNA (L. Reiser and R. Fischer, unpublished data). Positive cloneswere analyzed and one, containing the genomic DNA sequence flanking the T-DNAinsert, was used to screen a wild-type Arabidopsis genomic library. Based on theidentification of overlapping genomic clones, a restriction fragment map of theputative BELl genomic region was generated. Furthermore, Southern blot analysesdemonstrated that the T-DNA is inserted within a single, 2.9 Kb Hindlll genomicfragment of both beIl-2 and beIl-3 mutant alleles (Figure 4.1). The putative BELlgenomic sequence was used to screen a floral specific cDNA library, resulting inisolation of a 1.8 Kb cDNA sequence. Soon after, Northern blot analysis indicated thepresence of 2.4 Kb BELl mRNA in Arabidopsis flowers, suggesting that the full lengthcDNA had not been identified. Different cDNA libraries, including leaf, silique andfloral specific one, were screened for 5’ end sequences of the BELl cDNA. Fulllength 2.4 Kb BELl cDNA sequences were isolated from both leaf and floral specificcDNA libraries.Both BELl genomic and cDNA nucleotide sequences have been determined(L. Reiser, N. Chad, L. Margossian and R. Fischer, unpublished data). The sequenceof the 2.4 Kb BELl cDNA and the deduced amino acid sequence of a putative BELlprotein are shown in Figure 4.2. By comparing the cDNA and genomic nucleotidesequences, the structural organization of the BELl gene has been determined: BELlcontains four exons and three introns (Figure 4.3).86BELl genomic region1Kb5’ end 3’ endHindlIl HindlIHindill Hindill6.7 KbHindlil1.4 Kb 2.9 Kba)— 0.>o9.6 Kb6.0 Kb5.1 Kb4.6 Kb2.9 KbFigure 4.1. Restriction fragment map of BELl genomic region and Southern blotanalysis showing that the T-DNA is inserted within the 2.9 Kb HindlIl BELl fragment.Arrows span the size (Kb) of different Hindlil fragments. The 2.9 Kb Hindlil fragmentwas used as a probe for Southern blot analysis with genornic DNA from wild type,Bell-2 and Bell-3 plants, and the 2.9 Kb Hindlil fragment itself (control). The size(Kb) of hybridizing fragments is indicated.4.1.2 Expression of the BELl geneExpression of the BELl gene at the transcriptional level has been determined2.9 Kb Hindill87I bell-3189177265352MARDQFYGHNNHHHQEQQHQM441I NQ IQGFDETNQNPTDHHHYNHQ I FGSNS529NMGMMI D FSKQQQIRMTS GS DHHHHHHQT617 GTGGTGGTAcJCAGAcTTCTGGAAGATTcTTCCTGCCATGAGAcTATGCAATGTTAATAATGM’TTCCCAAGTGAAGTSGGTDQNQLLEDS SSAMRLCNVNND FPS EV705NDERPPQRPSQGLSLSLSSSNPTS ISLQS793FELRPQQQQQGYSGNKSTQHQNLQHTQMM881MMMMNSHHQNNNNNNHQHHNHHQFQ I GS S K969YLS PAQELLS E FCS LGVKE SD EEVMMMKH1057 AAGAAAGCAGGTAAACACAATGGGACPCMGTCACCACA3CAACAGTCAACPTGACCM1’C1’GCGACTACTTKKKQKGKQQEEWDTSHHS NND QHD Q SATTV1145SS KKHVPPLHS LEFMELQKRKAKLLSMLE E* * * _* *___1233LKRRYGHYREQMRVAAAAFEAAVGLGGAE*__ * * * *1321I YTALASRAMSRHFRCLKD GLVGQ I QATS1409QALGEREEDNRAVS IAARGET PRLRLLD QA1497 mGcGGcAAcAGAArCGThCGCCAAJGACTCrTGTrGACGCrCATCCrTGGCGTCCACAACGCGGc1’TGcCTGAACGCGcAGTCLRQQKS YRQMTLVDAHPWR P QRGLPERAVV1585TTLRAWLFEHFLHPYPS DVD KH I LARQTGV1673LS RS QVSNWF I NARVRLWK PM I E EMYC EE_______________________________* * * *1 761 CAAGAAGTGAACAAATGG3ATTACAAACCCGGATGTCGTACTAAACCGGACCCGGACCAGTTGTCCGTGTCGAACCGGAATCTRS EQME I TNPMM I DT KPD PD Q LI RVE P ES* * 88 * * * * *1849LS S IVTNPTSKSGHNS THGTMS LGS T FD F1937S LYGNQAVT YAGEGGPRGDVSLTLGLQRN2025DGNGGVS LALSPVTAQGGQLFYGRDH I EEG2113PVQYSASMLDDDQVQNLPYRNLMGAQLLH2201DIV2289 AAAAACCAAATATTTATIAAAAAAAMAAAAAAAAAAFigure 4.2. BELl cDNA sequence and deduced amino acid sequence of BELlprotein. Putative functional and/or structural domains within BELl are indicated asfollows: HD (underlined), acidic region (E and D amino acid residues with *), coiledcoil (amino acid residues with *), amino acid homopolymers (= = =), nucleartargeting motif (- - - -). The positions of three introns are marked by arrowhead (v)within the BELl cDNA sequence while the arrow (I) indicates the site of the T-DNAinsertion in Leonore Reiser using Northern blot and in situ hybridization analyses. Northernblot analyses demonstrated the presence of BELl transcripts indifferent plant tissues,including seedlings, leaves, flowers, and siliques. The pattern of BELl transcriptionin variety of plant tissues is surprising with respect to the Bell mutant phenotypewhich does not show any significant changes in vegetative development. Theexpression of BELl was also examined in three mutant alleles, bell-i (pointmutation), beIl-2, and bell-3 (T-DNA insertions) by Northern blot analyses.Preliminary data showed that BELl transcripts of expected 2.4 Kb size are presentin the bell-i allele. In the bell-2 allele, transcripts of smaller size than that of the wildtype are found whereas BELl transcripts were not detected in the beIi-3 allele(L.Reiser and R. Fischer, unpublished data). In situ hybridization analysis has89BELl genomic DNABELl mRNA5, 3’BELl proteinNH2L______Coiled coil610COOHAcidic domainHomeodomainFigure 4.3. Structure of the BELl gene and its product based on the comparison ofthe BELl genomic and BELl cDNA nucleotide sequences. The four exons areindicated by open boxes while filled boxes represent untranslated regions. The 5’and 3’ untranslated sequences located before and after the translated region,respectively, are indicated on the BELl mRNA. Within the BELl protein, arrowsindicate the relative position of the three different functional/structural domains.showed that BELl transcripts are located within ovules of wild-type Arabidopsisflowers.I5,1KbTAA904.2 Results4.2.1 Cloning of the beIl-3 mutant allelePlant genomic sequences flanking the T-DNA insert in the beIl-3 mutant allelehave been isolated for two reasons; first, to confirm the identity of the cloned BELlsequence by showing that the T-DNA of the Bell-3 mutant is inserted into the putativeBELl gene, and second, to characterize the T-DNA insertion in the BELl locus at themolecular level. Since the Agrobacterium T-DNA includes a pBR322 vector sequencecontaining an origin of replication and a gene conferring resistance to ampicillin inbacteria, plasmid rescue was used to clone the beIl-3 allele. To rescue the entire TDNA with flanking genomic sequences, genomic DNA isolated from Bell-3 plants wasdigested with the restriction enzyme BgI II. Bgl II was chosen because there are noBgI II restriction sites within the T-DNA. Digested DNA was ligated, electroporatedinto bacterial cells, and ampicillin resistant colonies were selected. Approximately tenampicillin resistant transformants were screened by colony hybridization using BELlgenomic sequence as a probe. Seven transformants designated G1-G7 hybridizedto BELl. By Southern blot analyses, plasmid DNA (pGl) isolated from the Gitransformant was shown to contain the most sequences homologous to BELl.Restriction fragment mapping was used to determine a relative map of the rescuedplasmid pGl. The plasmid DNA was digested with different enzymes, including thosewhose restriction sites were known within the T-DNA insert (i.e., EcoR I, Hind III, Pst91I, Sal I; Figure 4.4A, Velten and Schell, 1985). A map of pGl was constructed bycomparing the obtained restriction fragment patterns with those of the T-DNA. Basedon the generated map (Figure 4.4B), the rescued 24.1 Kb pGl plasmid containedapproximately 20.5 Kb of T-DNA sequences flanked by plant genomic sequenceswhich hybridized to the BELl clone. Figure 4.4B shows the putative arrangement ofT-DNA sequences interrupting the beIl-3 allele. Comparison of the T-DNA sequencesin pGV 3850:1003 plasmid (Velten and Schell, 1985) used for plant transformation andsequences present in the pGl revealed that the Bell-3 T-DNA was an inverted repeat(compare Figure 4.4A and B).The presumptive site of T-DNA insertion within the BELl genomic region wasproposed previously to be within the 2.9 Kb Hind Ill genomic fragment in both beIl-2and beIl-3 (L. Reiser and R. Fischer, unpublished data, Figure 4.1). Indeed, the plantgenomic sequences that were rescued when cloning the bell-3 allele coincide withthe 2.9 Kb Hind Ill fragment. Moreover, the plasmid rescue technique demonstratedthat the site of T-DNA insertion in bell-3 lies within the 1.3 Kb 5’ Hind 111-3’ BgI II(Figure 4.4B). Two pGl DNA fragments, 1.3 Kb Hind lII-EcoR I and 2.4 Kb Bgl ll-PstI, containing the putative junctions of T-DNA and Bell-3 genomic sequences weresubcloned and sequenced. The 1.3 Kb Hind lII-EcoR I fragment located at the 5’ endof the insertion was sequenced from the EcoR I site (Figure 4.4B). Approximately 300bp sequence showed homology to the 1 region of the T-DNA (Figure 4.4B).Similarly, at the 3’ end of the insertion, the 2.4 Kb BgI ll-Pst I fragment was sequencedfrom the BgI II site (Figure 4.4B). The first 245 bp of the sequence was completely92A) T-DNA (pGV385O:1003)EiEo o ——o .—opBR322 Tn903 Tn5 1B) T-DNA in beIl-3 allele1 KbFigure 4.4. Diagrammatic representation of the cloning of the beIl-3 mutant allele.A) Structure of the T-DNA in pGV385O:1003 plasmid. Different T-DNA sequenceshave been designated as follows: left border (LB), right border (RB), pBR322 vector(pBR322), sequences encoding KanR in bacteria (Tn903), sequences encoding KanRin plants (l’plant promoter and bacterial kanamycin resistance gene Tn5), nopalinesynthetase gene (NOS). The main restriction enzyme sites are indicated.B) Structure and position of the T-DNA in beIl-3. A relative map of rescued plasm IdpGl containing the T-DNA and flanking genomic sequences is drawn above thecorresponding BELl genomic region. The insertion of T-DNA into 1.3 Hindlll-Bglllfragment is indicated by straight lines. Arrows above the EcoRl and Bglll sitesindicate starting points and direction of sequencing.LB00uJ CuC),pBR322-4L=) .9 0z wlJCI)ZV pBR322RBINOSRB•jNOS____ ___________I RBNOS_.0__ — =_.9_. )____111111___________pGlr9O3PBR38pBR3225’Z5.9 .9 .9 8I I I Cl) LLiUJXI I II I I III =2(31.3 1.61.6 1.4 + 2.9* 9.63’93homologous to the previously cloned BEL 1 sequence and thus confirmed the identityof the isolated clone as BELl sequence. The remaining 55 bp following the BELlgenomic sequence did not show sequence identity with pBR322 as expected fromthe restriction mapping results (Figure 4.4B). Thus, the site where BELl sequenceis interrupted by T-DNA was determined at the 3’ end of the insertion. Compared tothe sequence of BELl eDNA, the site of T-DNA insertion in the beIl-3 lies within the5’ untranslated sequence (46 bp from beginning, Figure 4.2). Similar analyses of thesite and arrangement of T-DNA sequences were determined for the beIl-2 allele (L.Reiser and R. Fischer, unpublished data).4.2.2 Complementation of the Bell mutant phenotypeIn addition to molecular characterization of the 1-DNA insertion in the beIl-3allele, complementation of the Bell mutant phenotype was used to confirm theidentity of the cloned BELl sequence. Briefly, genomic DNA sequences containingthe putative BELl gene were inserted into the Agrobacterium binary vectorspCGN 1548 or pRD400 that were later used for plant transformation. Both pCGN1548and pRD400 carried the NPTI! gene conferring KanR as a selectable marker in plants.Since the exact position of the BELl gene was not known when these experimentswere initiated, two genomic fragments differing in the length of 3’ end sequences wereused for transformations. A 10.6 Kb Xba I BELl fragment was cloned into the Xba Isite of the plant transformation binary vector pCGN 1548, resulting in plasmid pAS394(Figure 4.5). An 8.6 Kb EcoR I BELl fragment was cloned into the EcoR I site of thepRD400 binary vector, generating plasmid pZM4 (Figure 4.5). Compared to thepresent structure of BELl gene (Figure 4.3), both constructs included approximately4.6 Kb upstream and either 0.6 Kb (pZM4) or 2.6 Kb (pAS3) sequences downstreamfrom the gene. The constructs were first transformed into Escherichia coil where theidentity and orientation of genomic fragments in pAS3 and pZM4 was confirmed byrestriction endonuclease digestions. In both pAS3 and pZM4 óonstructs the 5’ endgenomic sequences were adjacent to the T-DNA left border (Figure 4.5). Subsequenttransformation of Agrobacterium with the pAS3 and pZM4 generated bacterial strainsAS3 and ZM4, respectively.Because of sterility, I was not able to transform Bell-2 mutant plants directly.Instead, F1 and F2 progenies of the cross wt x Bell-2 were used for transformationexperiments. Mutant allele beil-2 was used for two reasons: it contained a T-DNAinsertion which was used later in segregation analyses and, unlike the beil-3 allele,due to rearrangements the T-DNA, did not include the NPTII gene conferring KanRwhich could interfere with selection of plant transformants. To increase the certaintyof obtaining transformants, two different techniques, root tissue culture and in piantavacuum infiltration, were used to generate transgenic Arabidopsis plants.Root tissue culture transformation. The technique described by Valvekens et al.,(1988) has been followed (see Materials and Methods). A total of 178 (76 obtainedwith ZM4 and 102 with AS3) independent KanR calli were used to generate 247 KanRplants (T1 transformants). Among 247 T1, 65 produced seeds and thus were valuable95BELl genomic clone1Kb=0 cu D C) 00 —Cu DC) .D Cu .E . Cu 0 C) CuUJXU) Z a Z C!) WW >< CoI5’ 3’4 Xbal BELl fragmentEcoRlBELlfragmentXbaJ BELl fragment in pCGN 1548 = pAS3= = =—c) D == — —0 00 (13. Cu . .E C3)C3) . Cu 00XC!) Z I CX]CI3 I C/) WW ><4N__________________________ ______________LB IPrd NPTI! ITerI 54 10.6 Kb 3I RBEcoRl BELl fragment in pRD400 = pZM40 D 00C.) (13 ..(130UJCI) I I (/)UJN jLB [Ten NPT(! P1d RB54 8.6Kb 3’Figure 4.5. BELl genomic clone and two constructs, pAS3 and pZM4, used forAgrobacterium-m ed iáted transformation. Main restriction sites are indicated. pAS3and pZM4 were generated by using unique Xbal and EcoRl sites (in bold) of pCGN1548 and pRD400, respectively. T-DNA sequences have been designated as follows:right border (RB), left border (LB), neomycin phosphotransferase gene (NPTII) withits promoter (Pro) and terminator (Ter).96for further analyses. These 65 T1 plants were derived from 49 calli representingindependent transformation events. Therefore, the regeneration efficiency was 27.5%(49 transformants/1 78 calli). Since transformants were recovered and analyzed morequickly using the in planta technique, genetic and molecular analyses of these 49 T1have not been carried out further. However, their analyses would be similar to thoseof transformants generated by in planta vacuum transformation described below. Thetransformation efficiency obtained in all transformations was similar, and thus, it doesnot seem to depend either on the bacterial strain (LBA4404 AS3 or ZM4) or genotypeof seeds used for transformation.In planta transformation. Both, the original in planta transformation technique(Chang et aL, 1994, Katavic et al., 1994) and its modification, vacuum infiltration(Bechtold et al., 1993), was used in combination for plant transformation (seeMaterials and Methods). Seeds from 15-20 treated plants (T0) were bulk harvested(T1 seeds), and screened for transformants (T1 plants) by growing on plates withkanamycin. T1 seeds harvested from 15-20 treated plants usually gave rise to variousnumber of KanR T1 plants, from one to thirty four. A total of 140 KanR T1 plants wereobtained; they probably include both unique transformants and siblings that originatefrom the same transformation event.During the transformation, the putative BELl genomic sequence, termed theBELl transgene (BEL1-T), and the NPTI! gene were presumably transferred into thegenome of T0 plants. My goal was to identify and analyze the phenotype of Bell-2mutant plants that were homozygous for the beIl-2 allele and contained BEL l-T. To97reach that goal, progenies of T1 plants were analyzed for segregation of the beIl-2allele, NPTII, and BEL1-T as follows.Figure 4.6 illustrates the genetic cross scheme of T1 KanR plants. The firststep was to identify T1 KanR plants that would segregate for the beIl-2 allele, thus allT1 plants (genotype = BEL 1-2/BEL 1-2, BEL 1-2fbeIl-2, or beIl-2/beIl-2) were crossedto mutant plants homozygous for the beIl-2. The F1 progeny of one T1 plant (no.AS31, Figure 4.6) contained both wild type and Bell-2 mutant plants, indicating thatthe initial T1 parent was heterozygous for beIl-2 (BEL1-2/beIl-2). Next, the F1 plants(designated by numbers from 1 to 6, Figure 4.6) were analyzed for the presence ofthe NPTII gene using PCR. Three F1 plants (nos. 3, 5, and 6) that contained NPTII(genotype BEL 1-2/beIl-2, NPTII/nptll or beIl-2/beIl-2, NPTII/npti!) were selfed andtheir progenies were analyzed for segregation of the beIl-2 allele and KanR marker.The segregation of these two traits was followed to determine how many T-DNAcopies were present, and to determine the frequency of the Bell-2 mutant phenotypesegregating in the population. F2 progeny of plant no. 5 segregated in the ratio 3KanR: 1 KanS (43:11, X2=O.62, P>O.1, Figure 4.6), typical for a single T-DNA insert.The ratio observed in two other F2 progenies was 15 KanR: 1 KanS (no. 3=114:13,X2=3.43, P>O.05; no. 6=68:6, sample too small for X2 analysis; Figure 4.6),suggesting presence of two T-DNA inserts, If BEL1-T complemented the beIl-2mutation, then the F2 progeny no. 5 should segregate for the Bell-2 mutantphenotype in ratio 15 wt: 1 Bell-2 whereas F2 progenies no. 3 and 6 in ratio 63 wt1 Bell-2. Indeed, Bell-2 mutant phenotypes were found with similar frequencies in98ANALYSES of KanR T1 PLANTSCROSS: KanR T1 x KanS BeIl-2GENOTYPE BEL 1-2/BEL 1-2, NPT!I/nptll beIl-2/beIl-2, nptl!/nptllBEL 1-2/beIl-2, NPTII/nptllbe! 1-2/be! 1-2, NPTII/nptllKanR T1 no. AS31 = BEL1-2/beIl-2, NPT!I/npt!! x beIl-2/be!1-2, npt!!/nptllF1 PLANTS No.: 1 2 3 4 5 6PHENOTYPE: wt Bell-2 wt wt wt wtNPTI!: - - + + +F1 GENOTYPES = BEL1-2/beIl-2, NPTII/nptll No. 3, 5, and 6beIl-2/beIl-2, NPTI!/nptllBEL1-2/beIl-2, npt!!/npt!I No. 1 and 4beIl-2/beIl-2, npt!!/npt!! No. 2self-fertilizedSEGREGATION ANALYSES for KanR MARKER and PHENOTYPE in F2:F1 plant F2 progeniesNo. 3: 114 KanR: 13 KanS, 83 wt: 1 Bell-2No.5: 43 KanR: 11 KanS, 57 wt:1 BeIl-2No. 6: 68 KanR: 6 KanS, 65 wt: 1 BeIl-2Figure 4.6. Genetic cross scheme of KanR T1 plants. The F1 and F2 progenies of thecross KanR T1 x KanS Bell-2 were analyzed for segregation of NPTII and mutantphenotype. The possible genotypes of Kan T1 and F1 progeny is indicated.99all three F2 populations (no. 5=57:1; no. 3=83:1; no. 6=65:1; samples too small forX2 analyses, Figure 4.6). The low frequency of Bell-2 mutant plants (beIl-2/beIl-2,nptI!/nptll) suggested that BEL1-Tcomplementated Bell phenotype. Within the sameprogeny, plants homozygous for the beIl-2 allele and homo- or heterozygous for theintroduced T-DNA (boIl-2/beIl-2, NPTII/NPTII or bell-2/beIl-2, NPT!I/nptll,respectively) should be present in frequency of 3/16. Such plants would representcomplemented Bell-2 mutant plants, thus ten plants from the F2 progeny of plant no.5 were analyzed by Southern blot analyses for the presence of the BEL 1-T transgene.Based on distinct hybridization patterns, three different forms of the BELllocus, including the wild type allele (BELl), mutant allele (beIl-2), and introducedBELl transgene (BEL1-T) could be distinguished by Southern blot analyses (Figure4.7). Therefore, the BELl genotype for any given plant could be easily determinedas follows. Briefly, the genomic DNA from ten KanR plants (all from F2 progeny of no.5 plant, Figure 4.6) was digested with BgI Il enzyme and hybridized with the 3’endBELl cDNA sequence (0.9 Kb fragment). The fact that there are no BgI II sites withinthe T-DNA sequences greatly facilitated differentiation of three forms of BELl locus.The wild type BELl allele appeared as a fragment of 9.4 Kb, in Bell-2 the sizeincreased to 22.2 Kb due to the insertion of T-DNA, and the BEL l-T was recognizedas a fragment of 15 Kb in size. The size of the BEL1-T fragment (15 Kb) was notknown in advance since it depends on the site of insertion into a plant genome. Asexpected, the Southern blot revealed the presence of the BEL1-T in all ten KanRplants that were derived from the same insertional event (15 Kb band in all lanes,1002 4 • 6 7 8 9 1.0 c.IC2WTFigure 4.7. Southern blot showing distinct banding pattern characteristic of threeforms of the BELl locus: BELl, beIl-2 and BEL1-T. Genomic DNA digested with BgIII was hybridized with BELl cDNA sequence as a probe. The lanes representfollowing samples: ten different transformants (lanes 1 to 10), Bell-2 planthomozygous for mutant allele (beIl-2/beIl-2, lane Cl), heterozygous Bell-2 plant(BeIl-2/beIl-2, lane C2) and wild type plant (lane WT). The sizes (Kb) of hybridizedfragments is indicated.Figure 4.7). Besides having the BEL1-T, three plants were homozygous for the wildtype BELl allele (9.4 Kb band, lanes 4, 6, and 7), three plants were BEL1-2/beIl-2heterozygous (9.4 Kb and 22.2 Kb bands, lanes 1, 9, and 10), and four plants werehomozygous for the mutant bell-2 allele (22.2 Kb band, lanes 2, 3, 5, and 8, Figure4.7). The phenotypes of the four plants that were homozygous for the bell-2 alleleand contained the BEL1-T transgene were analyzed.To determine the function of the BEL1-T in four plants identified as—22.2 Kb—15kb101complemented Bell-2, the morphology of ovules was examined in detail. In general,the phenotype of four complemented Bell-2 plants was similar to the wild type. Thetransgenic plants produced seeds and thus differ from the Bell-2 mutant plants.Analyses of their siliques under the dissecting microscope revealed presence of oval-shaped ovules, similar to those of the wild type. Transgenic ovules were alsoexamined by SEM which revealed that both inner and outer integuments enclose thenucellus (Figure 4.8). Minor morphological abnormalities were detected in someovules, for example, unlike wild type ovules, the outer integument did not completelyovergrow the nucellus. Conclusively, the alteration in ovule morphology observed inthe four transgenic plants suggested that BEL 1-Tfunctionally complemented the Bell -2 mutant plants, suggesting that the 10.6 Kb Xbal fragment used for transformationcontains the BELl gene.4.2.3 Characterization of the BELl gene productThe BELl cDNA sequence includes an 1830 bp open reading frame encodinga polypeptide of 610 amino acid residues (Figure 4.3). The predicted amino acidsequence of BELl protein shares amino acid identity with consensus homeodomainsequence (HD, Figure 1.3), classifying the BELl as a homeobox gene. As discussedpreviously, the highly conserved HD functions as sequence specific DNA-bindingdomain (for review see Affolter et al., 1990a; Gehring et al., 1994b). Within the HD,there are four invariable and eight highly conserved amino acid residues which serve102Figure 4.8. SEM of three types of ovules: wild type (A, magnification x573), Bell-3mutant (B, magnification x427), and transgenic ovule (C, magnification x625).103as a trademark of all HD proteins (Figure 1.3). The BELl HD contains the fourinvariable (W, F49, N1, and R53) and three out of the eight highly conserved aminoacid residues (L, R, R57, Figure 4.9).The BELl HD shares 26 out of 60 amino acid residues with the HD of theArabidopsis homeobox gene KNAT1, thus having 43.3% sequence similarity (Figure4.9). BELl and KNAT1 are not homologous in other protein regions outside the HD.Furthermore, the BELl HD has homology to the HDs of the maize gene KN1 (38%sequence identity) and KN1 related genes from maize or other plant species, forexample, ZMH1, ZMH2, OSH1, and SBH1 (Figure 4.9). Twenty percent amino acidsequence identity was also found between the HDs of the BELl and the yeast celltype specific factor MATa1 (Figure 4.9). A notable feature shared by all abovementioned homeobox genes (Figure 4.9) is the amino acid residue at position 9 ofthe recognition helix (lw), which partly determines the DNA binding specificity of HDproteins (Treisman et al., 1989). The other highly conserved motif located within theloop between helix 1 and helix 2 is the tripeptideP24Y56. Besides BELl, the PYPmotif has been found in all identified maize homeobox genes as well as in theArabidopsis genes KNAT1 and KNAT2 (Lincoln et al., 1994), suggesting a possiblestructural and/or functional relationship between BELl and the KN1-related genes.Although this motif appears common in plant homeobox genes, the PYP occurs rarelyin homeobox genes of Drosophila (cut, Blochinger et al., 1988) and yeast (MAT2-P,Kelly et al., 1988).In addition to the conserved HD, the BELl protein possesses several other1041 10 20 30 40 50 60N-term HelixI Loop Helixil Turn HelixillV V V V ** * * V VPQRGLPERA V]1RAWLFEHF LKPYPS DVDKHILARQT GLS RSQVSNWFflThRVRLWKPM BELlKKGK--KE- RQK-LT-WEL-Y KW---- ESE-VA--ES- --D QK-IN Q-K-H---S Ic\TAT1IG(GK--RE- RQ-LD-WNV-N KW---T EG--IS--EE- --D QK-IN Q-K-H---S KNT2KGKLPK-A- RQQ-LS-WDQ-Y KW---- E-VA--ES- --D LK-IN Q-K-H---S K1\T1KGKLPK-ZR QQL-EW-NRHYK KW---- ESQ-LA--ES- --D QK-fl Q-K-H---S SBH1KKGK--I- RQQ-LN-WEL-Y KW---- ESQ-VA--ES- --0 LK-IN Q-K-H---S OSH1AGKLPGDTI’ TSI-KQ-WQ--S KW---T ED--1K-VEE- --Q LK-NN Q-K-N-HNN Zt41-I1ALPGD1T JS--K--WQ-S KW---T EE--JR-VQE- --Q LK-IN Q-K-N-HNN ZIVJH2KSPKGKSSI SPQA.- -F-EQV- RRKQSL NSKEKEEVAKK - -T PL- -RV- - - -K-M-SK MATa1Figure 4.9. Comparison of the HD amino acid sequence between BELl and otherhomeobox genes, including Arabidopsis KNAT1 and KNAT2, maize KN1, ZMH1, andZMH2, soybean SBH1, rice OSH1, and yeast MATa1. The numbers designate aminoacid residue positions within the HD, the position of helices (I, II and Ill) is indicated.In the consensus HD sequence, four invariant amino acid residues (*) and eightresidues that are highly conserved among most HD (v) are in bold. Within otherhomeobox genes, residues that are in bold represent similar amino acid substitutions.interesting features. A cluster of negatively charged, acidic amino acid residues islocated immediately after the HD (from residues 454 to 489, Figure 4.3). Similaracidic domains rich in aspartic and glutamic acids have been found in otherArabidopsis HD proteins (Carabelli et al., 1993; Rerie et al., 1994), and have beenpostulated to function as a transcription activation domains (Ptashne and Gann,1990). The possible function of the BELl gene product as a transcription factor hasbeen supported by the presence of a putative bipartite nuclear-targeting sequence(Figure 42), suggesting the nuclear localization of the BELl protein. Four short105motifs rich in either histidine (H), glutamine (Q), methionine (M), or aspargine (N) arefound near the N-terminus of the BELl protein (Figure 4.2). Although the function ofsuch structural motifs is not known, similar amino acid homopolymers have beenfound in various types of transcription factors including HD proteins (Smoller et al.,1990). The polyhistidine array also occurs in KN1 (Vollbrecht et al., 1991) and KNAT1(Lincoln et al., 1994). The predicted secondary structure of BELl suggests itscapability to form a coiled-coil (from residues 269 to 307, Figure 4.3). The coiled-coilformation has been shown to mediate protein-protein interactions (Ho et al., 1994),thus suggesting possible interactions between BELl and other proteins. Formationof homo- or heterodimers commonly results in enhanced affinity for the targetsequences as well as functional activation of transcription (Laughon, 1991).4.3. Discussion4.3.1 Confirmation of the cloned sequence as the BELl geneAs discussed previously, both recessive alleles beIl-2 and beIl-3 result fromT-DNA insertions in the BELl locus. A putative BELl gene was cloned usingsequences adjacent to the T-DNA of Bell-2 mutant plants (L. Reiser and R. Fischer,unpublished data). To prove that the isolated sequences are indeed sequences ofthe BELl gene, two approaches were undertaken. First, the molecularcharacterization of the beIl-3 mutant allele demonstrated that a T-DNA lies within the106first exon of the putative BELl gene. Second, the putative BELl gene complementedthe beIl-2 mutation resulting in plants with functional, morphologically normal ovules.Both lines of evidence strongly suggest that the cloned sequences represent theBELl gene.4.3.2 T-DNA insertions in the BELl geneThe first step in my characterization of the T-DNA insertional mutant Bell-3 wasto examine the linkage between the T-DNA insert and mutant phenotype (Marks andFeldmann, 1989; Errampalli et al., 1991). The genetic segregation data obtained wereconsistent with the hypothesis that a T-DNA insert containing the NPTII gene hasinterrupted a gene required for normal ovule morphogenesis. It is interesting that inthe beIl-2 allele the inserted T-DNA sequences did not include a functional NPTIIgene (L. Reiser and R. Fischer, unpublished data). Presence of T-DNA inserts withfull, partial, or negligible NPTII activity has been previously observed (Errampalli et al.,1991), and both deletion and altered expression appear to be responsible for suchpatterns.The Bell-2 and BeIl-3 mutant lines were generated at different times andthrough separate T-DNA transformation experiments, thus they must represent uniquemutational events at the same locus. The uniqueness of each mutation at themolecular level was determined through characterization of both beIl-2 and beIl-3mutant alleles. With respect to the pattern of insertion of T-DNA sequences, these107two mutant alleles differ significantly. A T-DNA insert presents an inverted repeat inthe beIl-3 allele whereas tandem repeats of mainly right border sequences areinserted within beIl-2 (L. Reiser and R. Fischer, unpublished data). The wide rangeof different T-DNA structures observed within plant genomes (Feldmann et al., 1989;Herman and Marks, 1989; Jones et al., 1987; Jorgensen et a!., 1987) indicates thecomplexity of the T-DNA insertional events.Similar to the pattern of inserted T-DNA sequences, the site of the T-DNAinsertion also varies in the two mutant alleles. In beIl-2 the T-DNA is inserted in thefirst intron of the BELl gene (L. Reiser and R. Fischer, unpublished data), while in thebeIl-3 allele, the T-DNA interrupts the first exon. The observed difference in the siteof T-DNA insertion in beIl-2 and beIl-3 alleles may influence the pattern of BELltranscription. Northern analyses detected BELl transcripts of smaller size than thatof the wild type in Bell-2 mutant plants while in BeIl-3 the transcripts are absent (L.Reiser and R. Fischer, unpublished data). It is conceivable that in the beIl-2 allelethe T-DNA insert affects RNA processing (splicing), producing the BELl transcriptsof smaller size. Since BELl transcripts are absent in the beIl-3 allele, the insertionof T-DNA may interfere with transcription or generate unstable transcripts. Thus, itseems likely that beIl-3 is a null mutant allele. Since the phenotype of ovules is verysimilar in both BeIl-2 and Bell-3 mutant plants, the truncated BELl gene product, ifpresent, appears to be nonfunctional in Bell-2 mutant plants.4.3.3 BELl is a novel homeobox gene108The BELl gene shares amino acid sequence identity with a DNA-bindingdomain, the homeodomain, and thus belongs to the homeobox gene family oftranscription factors. The function of the BELl gene as a transcription factor hasbeen supported by structural features of the putative BELl protein, including apossible acidic transcription activation domain and nuclear-targeting motif.Furthermore, the BELl region corresponding to the transcriptional activation domainactivated expression of a reporter gene (A. Samach and G. Haughn, unpublisheddata) in the yeast 2-hybrid system (Fields and Sternglanz, 1994). The possibility offormation of a coiled-coil within the BELl protein suggests that it may interact withother gene products. Preliminary data obtained in the 2-hybrid yeast system supportsuch protein-protein interaction since several gene products that possibly interact withBELl have been identified (A. Samach and G. Haughn, unpublished data). Takentogether, the HD and other structural/functional characteristics found in BELl stronglysuggest that the BELl gene product functions as a transcriptional regulator.The majority of cloned Arabidopsis homeobox genes encode proteins thatinclude a leucine zipper motif downstream from the HD, and thus are designated asHD-Zip proteins (Schena and Davies, 1992). A few other genes, such as HAT3. 1 andGL2, are structurally different homeobox genes encoding distinct HD-finger or atypical HD protein, respectively (Schindler et al., 1993; Rerie et al., 1994). Withrespect to size, the BELl protein is substantially larger than the HD-Zip proteins andabout the same size as the HAT3.l or GL2 proteins (Schindler et al., 1993, Rerie etal., 1994). However, no significant homology, except for the amino acid residues109within the HD, has been found between BELl, HAT3. 1, and GL2. Moreover, outsidethe HD, the amino acid sequence of BELl is not homologous to any other proteinavailable in the databank. Therefore, BELl is distinct from all other identifiedhomeobox genes in Arabidopsis.4.3.4 Function of BELl in ovule/plant developmentBesides GL2 (Rerie et al., 1994), BELl is the only homeobox gene known inArabidopsis and other plant species for which a loss of function phenotype has beenwell defined (Robinson-Beers et al., 1992; Modrusan et al., 1994b; Ray et al., 1994).The analyses of Bell mutant plants suggested that BELl prevents the developmentof carpels within the ovules by repressing the AG gene. As a member of homeoboxgene family of transcription factors, BELl could be considered part of the regulatorycascade directing the ovule development.The molecular mechanisms through which BELl negatively regulates AGduring ovule development are not yet known. Since a potential transcriptionalactivation domain has been identified in the putative BELl protein, BELl may activatetranscription of another gene(s) whose activity results in suppression of AG.Alternatively, BELl gene may function in combination with other gene product(s) toregulate AG expression directly. Based on the presence of a structural motif whichallows protein-protein interactions and on the pattern of transcription within theovules, potential candidates for the interaction with BELl are ovule-specific MADS-box110(i.e., AGL1, Ma et al., 1991) and homeobox genes (A. Samach and G. Haughn,unpublished data). In an analogous manner, the yeast homeobox gene MATa2interacts with MADS-box gene MCM1 or homeobox gene MATa1 and as suchnegatively regulate a-specific or haploid-specific genes, respectively (for review seeHerskowitz, 1989). The interaction of two homeobox genes, MATa2 and MATa1, ismediated by a coiled-coil (Ho et at., 1994), a structural motif recognized in theputative BELl protein. The interaction between HD and MADS-box proteins havebeen characterized previously in mammalian systems (Grueneberg, 1992). Therefore,interaction of the BELl protein with other ovule-specific regulatory gene products isa potential mechanism of AG regulation in ovules.To date, analyses of the BELl transcription pattern have not yet beencompleted. Preliminary data from Northern analyses demonstrated that BELl istranscribed in developing flowers, siliques, seedlings, and leaves (L. Reiser and R.Fischer, unpublished data). This is a surprising finding because Bell mutant plantsdo not show any morphological changes in vegetative development. Possibleexplanations for such differences between the mutant phenotype and the distributionof the BELl transcripts include post-transcriptional regulation of BELl or lack of atissue specific factor with which BELl interacts which limits its function to ovules.Alternatively, other gene activities may compensate for the loss of BELl in vegetativeplant tissues (gene redundancy). Analyses of the BELl transcripts and protein inflowers and vegetative tissues is necessary to resolve these possibilities. Aninteresting observation is that AP2 is also transcribed in vegetative tissues and thus,111both genes may negatively regulate AG throughout shoot development. If true, theovule-specific effects of ap2 and bell mutations could be only explained bypostulating that other gene(s) playing similar roles in vegetative tissues compensatefor loss of AP2 and BELl activity. Functional gene redundancy involving the APi andCAULIFLOWER MADS-box genes in Arabidopsis flower development has beenidentified previously (Kempin et al., 1995).1125. ARABIDOPSIS THALIANA GENES HOMOLOGOUS TO BELl5.1 IntroductionAlthough in the last several years plant homeobox genes have been extensivelystudied, for example, the KN1 and numerous HAT genes (Hake, 1992; Schena andDavies, 1992, 1994), the understanding of their regulatory roles in plant developmentis still incomplete. Since homologous genes often have related functions within thesame processes, the identification and characterization of closely related homologousgenes can contribute to an understanding of the biological process underinvestigation. This approach has been used in maize and Arabidopsis and resultedin identification of two large KN1 and HAT homeobox gene subfamilies, respectively(Kerstetter et al., 1994; Schena and Davies, 1994). In an analogous manner, mostof the information on the control of Arabidopsis flower development has beenobtained by investigating a class of regulatory genes encoding MADS-boxtranscription factors (for review see Davies and Schwarz-Sommer, 1994). Moreover,MADS-box sequence from the AG gene has been used to identify another sixmembers of the same subfamily (Ma et al., 1991) that share structural similarity aswell as related functions within flower development.As described previously, the BELl gene contains conserved sequencesencoding the HD, typical of the homeobox class of transcription factors. PreliminarySouthern blot data have demonstrated that the BELl gene belongs to a small gene113subfamily (L. Reiser and R. Fischer, unpublished data). Therefore, in collaborationwith Dr. Alon Samach, I used the BELl HD sequence to search for novel regulatorygenes, possibly with related functions. Identification and characterization of geneshomologous to BELl would allow us to study the function of individual members aswell as the whole gene subfamily, and thus contribute to a more completeunderstanding of the regulation of processes involved in ovule and/or plantdevelopment.5.2 Results5.2.1 Isolation of cDNA sequencesTwo Arabidopsis cDNA libraries, floral specific and whole plant, were screenedfor genes homologous to the BELl. A PCR amplified fragment of the BELl HD (from1490 to 1846 bp of BELl cDNA, Figure 4.2) was used as a probe to screen bothlibraries. The primary screen and rescreening, as well as subsequent Southernanalyses of positive clones were done under low stringency conditions (see Materialsand Methods). After screening 50,000 plaques of each library, eleven and ninepositive clones were selected from floral specific and the whole plant cDNA library,respectively. Clones that originate from the floral specific eDNA library weredesignated as 1 G, 2G, 3G, ..., 11 G, whereas ones from the whole plant cDNA librarywere named 1Y, 2Y, 3Y, ..., 9Y.114To analyze putative eDNA sequences that hybridized to the BELl HD, ?. DNAwas isolated from each of twenty positive bacteriophage isolates and digested withEcoR I enzyme to release the cDNA sequences. Most cDNA sequences wererepresented by a single fragment whose size ranged from 0.4 to 1.6 Kb from the floralspecific library or 0.5 to 2.5 Kb from the whole plant cDNA library. eDNA sequenceswere verified for their homology with the BELl HD sequence by Southern blotanalysis. All 20 cDNA EcoR I fragments were subcloned into pT7T318U vectorsuitable for sequence determination. However, before DNA sequencing, severalenzymes including BgI II, EcoR I, Hind Ill, Sac I, and Xba I were used to generaterestriction fragment maps of positive clones. The restriction fragment map of a eDNAfragment from the clone 2Y was the same as that of the BELl cDNA and moreover,the 2Y clone hybridized more strongly to the BELl HD than the others. Thus, thecDNA fragment from one out of the twenty clones possibly represented the BELlcDNA sequence. The restriction fragment mapping also suggested similarity betweenother cDNA sequences which were grouped together, possibly representingoverlapping cDNA clones. The relationship between putative overlapping cDNAsequences within each group was verified by Southern blot analysis. However, thefinal proof that these fragments represent overlapping cDNA sequences came fromdetermination of DNA sequences.5.2.2 Sequence analyses of cDNA sequences115Sequencing of all twenty cDNA sequences identified four distinct genes withhomology to BELl. As expected, the clone 2Y contained the BELl cDNA sequence.Three additional groups of genes with homology to BELl were recognized: the firstgroup contained cDNA sequences from clones 4Y and 6Y; the second group includedcDNA sequences from clones 1Y, 8Y and 9Y; and the third group consisted of DNAsequences from five clones, 3Y, 5Y, 7Y, 7G and 13G. The longest cDNA sequencewithin each group, the 6Y, 9Y and 13G, were characterized further. Sequencing ofthe remaining cDNA sequences (clones 1G, 3G, 4G, 5G, 6G, 8G, 9G, hG, and 12G),which were significantly less homologous to the BELl HD based on Southern blotdata, did not reveal homology to BELl or any other gene sequence available in thedatabank, and thus they were not considered further.5.2.3 Identification of Arabidopsis BELl-like homeobox genesAs mentioned above, three groups of overlapping cDNA sequences whoserepresentatives are clones 6Y, 9Y and 1 3G, showed high homology to the BELl gene,and thus they were designated ABH1, ABH2, and ABH3 genes, respectively(Ajabidopsis BEL1 Homolog). The ABH1, ABH2, and ABH3 genes share nucleotidesequence identity with BELl of 78.8%, 70.0%, and 70.5%, respectively. ABH1 andABH2 sequences originate from the whole plant cDNA library while the ABH3sequence has been found in both floral specific and whole plant cDNA libraries. Thesize of ABH1, ABH2, and ABH3 cDNA sequences are 1.32 Kb, 1.1 Kb, and 1.58 Kb,116BELl MARDQFYGHNNHHHQEQQHQMflQIQQFDETNQNPTDHHHYNHQIFGSNSNMGIDFsK 60BELl QQQIRMTSGSDHHHHHHQTSGGTDQNQLLEDSSSAMRLCNVNDFPSEVNDERPPQRPSQ 120A.BH3 -TLE-P-V-SSN- 13BELl GLSLSLSSSNPTSISLQSFELRPQQQQQQYSGNKSTQHQNLQHTQSHHQN1NN 180ABH3 -K---VHQHH -DQIL-SSVYNNN-- NGVGFY- 38BELl NNHQHHNHHQFQIGSSKYLSPAQELLSEFCSLGKESDEEVKHKKKQKGKQQEEWDT 240ABH3 -YRYETSGFVSSVLR-R- -K-T-Q- -D-VV-V_R-DLKLGN__-KM-ND- --DFTMGL 73BELl SHHSNNDQHDQSATTS SKKHVPPLHSLEFNELQKRKAKILSMLEELK RRYGHYREQM 297ABH3 VITLQKMINLNRRS -SPS-RQ---SK-S--FT-VG--RWVK--N--HH-- 122ABH2 --RQQEQ---RQ-HQ-- 16BELl RVAAFEAAVGLGGAEIYTALASRSRHFRCLKDGLVGQIQATSQALGEREEDNRAVS 357ABH3 EAL-SS--MVT---A-KP--SV-LNRI R-AIKE---VIRGK TS_____ 352ABH2 QMVISS--Q-X--- S--LKTI--Q-X---EAIA---K-ANKS---EDSV - 172ABH1 NSS-YV-D AVA-L-SCEL--DKDAAGISS- 52BELl IAARGET_PRLRLLDQAIIRQQKSYRQ_MTLVDAHP_________________________ 390ABH3 DEQ--RI----Y---R----RDLH-QLGI-RPA______________________ 385ABH2 GVG-F-GS--KFV-HH----RALQ-LGMIQ--SN____________________ 208ABHI GLTK------C--E-S----P.AFH-_-GMMEQEA______________________ 85N-terminus Helix I Loop Helix II Turn Helix IIIV V V V V V ****VVBELl WRPQRGLPERA VTTLRAWLFEHF LHPYPS DVDKHILARQT GLS RSQVSNWFINARVRLWKPM 452ABH3 NS -SI K ESE-IM-SIC KN--A 447A.BH2 SV X K -S---M--K-- --T -x 270ABHZ --S S -NI----I L VP- 147Figure 5.1. Comparison of amino acid sequences of BELl (in bold) and three BELlhomologs, ABH1, ABH2, and ABH3. Identical amino acid residues are indicated withdashed line (-) while the gaps, represented by straight line, are introduced to obtainthe highest homology. Position of helices (I, II, Ill), invariant (*) and highly conserved(v) amino acid residues are indicated within the HD. Numbers indicate relativeposition of amino acid residues within each protein.respectively, and they all encode HD proteins (Figure 5.1). Within the HD, the aminoacid sequence identity between ABH1, ABH2, and ABH3 and BELl is 83%, 85%, and75%, respectively. Amino acid sequence homology between the ABH and BELlproteins was also found in regions outside the HD, for example from amino acid117residues 272-310 (50% for ABH3, 44% for ABH2), from 325-351 (61% for ABH3, 50%ior ABH2, and 48% for ABH1), and from 362-376 (75% for ABH3, 53% for ABH2, and80% for ABH1 (numbers as in BELl, Figure 5.1). Based on the amino acid sequencehomology with BELl (from 272-310 residues), it is possible that ABH2 and ABH3proteins may contain coiled-coil motif. Although the function of the two other highlyhomologous protein regions is not known in BELl, the high conservation of suchmotifs suggests their importance either for proper protein structure and/or function.Taken together, based on high homology to the BELl, the three genes ABH1, ABH2,and ABH3 represent new members of the BELl homeobox gene subfamily.5.3 Discussion5.3.1 Arabidopsis BELl homeobox gene subfamilyUsing a known gene sequence as a tool in a search for other genes has beena common approach to identify genes within the same or different plant species. Forexample, the maize KN1 HD sequence was extensively used to isolate many KN1related maize genes (ZMH1, ZMH2, Vollbrecht et al., 1991; KNOX genes, Kerstetteret al., 1994), or genes homologous to KN1 in other plant species (SBH1, Ma et al.,1994; OSH1, Matsuoka et al., 1993; KNAT1 and KNAT2, Lincoln et al., 1994). Here,we used the BELl HD sequence in a similar manner and identified three newhomeobox genes, ABH1, ABH2, and ABH3.118Homeobox genes are grouped into distinct subfamilies based on the degreeof conservation of their HD (Scott et al., 1989; Carabelli et al., 1993). At present, twosuperfamilies of homeobox genes, HD-Zip and KN1-like, have been recognized inArabidopsis and maize, respectively. Since both superfamilies are comprised ofmany homologous genes, each one has been divided into two classes based onsequence homology among the individual members (Schena and Davies, 1994;Kerstetter et al., 1994). As described previously, the BELl HD is significantly differentfrom all other identified plant homeobox genes, implying that it represents a distinctArabidopsis homeobox gene. Identification of three genes, ABH1, ABH2, and ABH3,which are highly homologous to BELl within the HD (83%, 85%, and 75%,respectively) suggests that they, together with the BELl, represent a novel BELlhomeobox gene subfamily. The BELl and homologous ABH1, ABH2, and ABH3genes share essentially identical amino acid residues in the recognition helix and highhomology in other protein regions. Although the regions of homology do notcoincide with BELl transcriptional activation domain, isolation of full length ABH cDNAsequences will reveal if such domains exist in ABH gene products.The homology within the HDs of the BELl and three ABH genes suggest first,that all members of the BELl gene subfamily possess comparable DNA-bindingspecificities, and second, that their gene products may function in a similar manner.Based on the structural similarity, ABH genes may represent candidates which cancompensate for the loss of BELl function in vegetative plant tissues. Further analysesof the three new homeobox genes, ABH1, ABH2, and ABH3, may provide information119concerning not just their regulatory roles but also that of BELl, and thus we intendto proceed with their characterization. Identification of full length ABH cDNAsequences will contribute to characterization of structural/functional features ofputative ABH proteins. To determine if the three ABH genes represent novel loci,their chromosomal positions on the Arabidopsis linkage map will be determined byrestriction fragment length polymorphism mapping. Northern blot analyses will beused to determine the expression pattern of the three genes in different tissues ofArabidopsis plants. More precise localization of gene transcripts within certain typesof tissue and/or cell types during development will be obtained by in situ hybridizationanalysis. To gain an insight into the regulatory roles of the ABH1, ABH2, and ABH3genes, transgenic experiments, including ectopic gene expression (overexpression)and antisense suppression will be employed. Similar to BELl, experiments designedto determine the molecular mechanisms through which these genes function,including DNA-protein and protein-protein interactions, will contribute to theunderstanding of their regulatory functions. These analyses will hopefully reveal thebiological importance of the BELl homeobox gene subfamily in Arabidopsis.120CONCLUDING REMARKSMy analyses of the Arabidopsis thaliana mutants Bell and Ap2, defective inovule development, have shown first that the wild type BELl gene regulatesdevelopment of ovules in the Arabidopsis flower, and second that the organ-identitygene AP2 also has a role in this developmental event. Moreover, in Bell and Ap2mutant plants similar homeotic transformation of ovules into structures with carpel-likefeatures occurs occasionally, suggesting that in wild-type ovules both BELl and AP2genes prevent the activation of alternative developmental programs, such as the onefor carpel formation. The development of the carpel in the Arabidopsis flower isdetermined by the organ-identity gene AG. Therefore, by repressing thedevelopmental program of carpel within ovules, BELl appears to function as anegative regulator of AG during ovule development. Like BELl, AP2 may function ina similar manner, resembling one of its roles, suppression of AG, during earlydevelopment of the Arabidopsis flower.Bell-2 and Bell-3 mutant plants were generated by Agrobacteriumtransformation. In both beIl-2 and beIl-3 mutant alleles, distinct T-DNA sequenceswere inserted into the BELl locus. Plant genomic sequence adjacent to the T-DNAin beIl-2 was cloned and used to isolate the wild type BELl gene (L. Reiser and R.Fischer, unpublished data). I have confirmed the identity of the cloned sequence asthe BELl gene by first verifying that the same genomic sequences are presentadjacent to the 1-DNA in the beIl-3 allele, and second by complementing the Bell121mutant phenotype with the cloned sequence. Subsequent analyses of the BELlsequence revealed that it has sequence similarity to the homeobox gene family oftranscription factors, and thus likely functions as a transcription factor itself. Theputative function of BELl as a transcriptional regulator coincides with its geneticallydefined role as a repressor of the AG gene during ovule development.Since BELl functions as a regulatory gene in plant development, its sequencewas used to identify three additional Arabidopsis homeobox genes, ABH1, ABH2, andABH3. Based on the high sequence homology between BELl and the three ABHgenes, they may be involved in regulation of similar processes and together with theBELl, comprise a new BELl homeobox gene subfamily in Arabidopsis. The functionsof individual members of the BELl gene subfamily should be further elucidated aswell as the role of the whole family within plant development.Since this research project on the molecular and genetic analyses of the BELlgene has been done in collaboration with Dr. R. Fischer’s group, in conclusion Iwould like to list my unique contributions to this research project:1) Characterization of ovule development in wild-type Arabidopsis which served asa basis for comparison to the Bell mutant ovules.2) Genetic analysis of the bell mutation including mapping of the BELl gene.3) Phenotypic characterization of Bell mutant plants, in particular SEM analysis ofovule development.4) Construction and analysis of Bell/Ap3 and Bell/Ap2 double mutant plants.5) Determination of linkage between Bell-3 mutant phenotype and a T-DNA insertion.1226) Molecular characterization of the T-DNA insertion in the beIl-3 allele.7) Complementation of the ovular defects in Bell mutants with the cloned BELl gene.8) Identification of three novel homeobox genes ABH1, ABH2, and ABH3 that arehighly homologous to the BELl gene.123FURTHER INVESTIGATIONSThe cloned BELl gene represents a valuable tool for numerous investigationswhich eventually will contribute to a more precise picture of BELl function inArabidopsis plant/ovule development. Although preliminary data on the BELlexpression pattern exists, more detailed studies of the BELl transcription pattern inthe wild type plant/flower are required to refine the BELl function to specific tissuesand/or cell types. Expression studies of BELl in different mutant backgrounds, forexample Ag, Ap2, and Bell, may further indicate possible relationships between BELland Arabidopsis organ-identity genes. BELl gene expression should also beexamined at the protein level by immuno-localization. This analysis would reveal thesimilarities and/or differences between BELl transcript and BELl protein localizationand thus identify the sites of actual gene function.The role of BELl in ovule development may be further defined through geneticanalyses of various Arabidopsis mutants. For example, characterization of a triplemutant Ag/Ap2/Bell may reveal whether negative regulation of the AG is the sole, oralternatively, one among other roles of BELl in ovule morphogenesis. Although therole of AP2 in ovule development was recognized, the function of AP2 during ovuledevelopment has not been characterized in detail. Based on genetic studies, AP2and BELl genes function independently to prevent development of carpels withinovules. It would be interesting to compare the molecular mechanisms by which AP2and BELl accomplish such regulatory functions.Transgenic studies may also provide useful information regarding BELl124function in plant/ovule development. Based on the phenotype of transgenicArabidopsis plants containing the BELl gene under the control of a constitutivepromoter, we could verify the proposed role of BELl as negative regulator of AG, forexample if transgenic phenotype is Ag-like. Moreover, such experiments mayuncover additional aspects of the role of BELl in Arabidopsis plant/ovuledevelopment that were not obvious from Bell mutant analyses. The value of doingsuch experiments rely on the results of the BELl expression studies, for example, ifBELl is constitutively expressed in Arabidopsis flower/plant, the overexpressionexperiments would not be initiated.Besides experiments designed to investigate the role of BELl in Arabidopsisplant/ovule development, the function of BELl as a transcriptional regulator shouldbe examined. The specific DNA sequences that BELl protein binds to could bedetermined by in vitro (gel mobility shift assay, DNA footprinting) as well as in vivoDNA-binding studies (i.e, in yeast system). Identification of specific DNA bindingsites (i.e., within the AG promoter) would also distinguish between direct or indirectregulation of AG by BELl. Since protein-protein interactions commonly mediatenumerous regulatory functions, including transcriptional regulation, the interactionsbetween BELl and other proteins will be examined. It is possible that the BELlprotein participates in such interactions via its coiled-coil and thus function either ashomo- or heterodimer. Both in vitro (i.e., in solution) and in vivo (i.e., the yeast 2-hybrid system) studies could be used to examine BELl for protein-proteininteractions. Finally, if present, the biological significance of such BELl-protein Xinteractions should be examined in plants.125BIBUOGRAPHYAffolter, M., Schier, A., and Gehring, W.J. (1990a). Homeodomain proteins andregulation of gene expression. Curr. Opin. Cell Bid. 2, 485-495.Affolter, M., Percival-Smith, A., Muller, M., Leupin, W., aqd Gehring, W.J. (1990b).DNA binding properties of the purified Antennapedia homeodomain. Proc. Natl. Acad.Sci. USA 87, 4093-4097.Ammerer, G. (1990). Identification, purification, and cloning of a polypeptide(PRTF/GRM) that binds to mating-specific promoter elements in yeast. Genes Dev.4, 299-312.Angenent, G.C., Buscher, M., Franken, J., Mol, J.N.M., and van Tunen, A.J. (1992).Differential expression of two MADS box genes in wild-type and mutant petuniaflowers. Plant Cell 4, 983-993.Angenent, G.C., Franken, J., Buscher, M., Colombo, L, and van Tunen, A (1993).Petal and stamen formation in petunia is regulated by the homeotic gene fbpl. PlantCell 4, 101-112.Angenent, G.C., Franken, J., Buscher, M., Weiss, D., and van Tunen, A (1994). Cosupression of the petunia homeotic gene fbp2 affects the identity of the generativemeristem. Plant J. 5, 33-44.Beasley, C.A., and Ting, l.P. (1974). Effects of plant growth substances on in vitrofiber development from unfertilized cotton ovules. Amer. J. Bat. 6, 188-194.Baumruker, T., Sturm, R. and Herr, W. (1988). OBP100 binds remarkably degenerateoctamer motifs through specific interactions with flanking sequences. Genes Dev 2,1400-141 3.Bechtold, N., Ellis, J., and Pelletier, G. (1993). InplantaAgrobacterium mediated genetransfer by infiltration of adult Arabidopsis thaliana plants. C. R. Acad. Sci. paris,Sciences de Ia vie/Life sciences 316, 1194-1199.Berg, J.M. (1990). Zinc fingers and other metal-binding domains. J. Biol. Chem. 265,6513-6516.Behringer, F.J., and Medford, J.l. (1992). A plasmid rescue technique for the recoveryof plant DNA disrupted by T-DNA insertion. Plant Mol. Biol. Rep. 10, 190-198.Bellmann, R., and Werr, W. (1992). Zmhox 1A, the product of a novel maizehomeobox gene, interacts with the Shrunken 26 bp feedback control element. EMBO126J. 11, 3367-3374.Benevante, R.S., Skorupska, H., Palmer, R.G., and Shoemaker, R. (1989). Embryosac development in the cv.ks male sterile, female sterile line of soybean ‘(Glycinemax.). Am. J. Bot. 76, 1756-1768.Billeter, M., Qian, Y.Q., Oiling, G., Muller, M., Gehring, W.J. and Wuthrich, K. (1990).Determination of the three-dimensional structure of the Antennapedia homeodomainfrom Drosophila in solution by 1H nuclear magnetic resonance spectroscopy. J. McI.Biol. 214, 183-197.Billeter, M., Qian, Y.Q., Oiling, G., Muller, M., Gehring, W.J. and Wuthrich, K. (1993).Determination of the NMR solution structure of an Antennapedia homeodomain-DNAcomplex. J. Mol. Biol. 234, 1084-1093.Birnboim, H.C., and Doly, J. (1979). A rapid alkaline extraction procedure forscreening recombinant plasmid DNA. Nucl. Acid Res. 7, 1513-1523.Blochlinger, K., Bodmer, R., Jack, J., Jan, LY., and Jan, Y.N. (1988). Primarystructure and expression of a product from cut, a locus involved in specifying sensoryorgan identity in Drosophila. Nature 333, 629-635.Bocquet, G. (1959). The campylotropous ovule. Phytomorphology 9, 222-227.Bouman, F. (1984). The ovule. In: Embryology of Angiosperms, ed. B.M. John,Springer-Verlag, Berlin, pp. 123-158.Bowman, J.I., Smyth, D.R., and Meyerowitz, E.M. (1989). Genes directing flowerdevelopment in Arabidopsis. Plant Cell 1, 37-52.Bowman, J.L, Smyth, D.R., and Meyerowilz, E.M. (1991a). Genetic interactionsamong floral homeotic genes of Arabidopsis. Development 112, 1-20.Bowman, J.L, Drews, G.N., and Meyerowitz, E.M. (1991b). Expression of theArabidopsis floral homeotic gene AGAMOUS is restricted to specific cell types late inflower development. Plant Cell 3, 749-758.Bowman, J.L, Alvarez, J., Weigel, D., Meyerowitz, E.M., and Smyth, DR (1993).Control of flower development in Arabidopsis thaliana by APETALA1 and interactinggenes. Development 119, 721-743.Bradley, D., Carpenter, R., Sommer, H., Hartley, N., and Coen, E. (1993).Complementary floral homeotic phenotypes result from opposite orientations of atransposon at the plena locus of Antirrhinum. Cell 72, 85-95.127de Candolle, A.P., and de Candolle, A. (1841). Monstruosites vegetales. Imp.Petitpierre, Neuchatel, Switzerland.Cass, D., Petaya, D.J., and Robertson, B.L (1985). Megagametophyte developmentin Hordeum vulgare. 1. Early megagametogenesis and the nature of cell wallformation. Can. J. Bot. 63, 2164-2171.Carabelli, M., Sessa, G., Baima, S.., Morelli, G. and Ruberti, I. (1993). The ArabidopsisAthd-2 and -4 genes are strongly induced by far-red-rich light. Plant J. 4, 469-479.Chang, S.S., Park, S.-K., Kim, B.C., Kang, B.J., Kim, D.U., and Nam, H.G. (1994).Stable genetic transformation of Arabidopsis thaliana by Agrobacterium inoculationin planta. Plant J. 5, 551 -558.Chaudhury, AM. (1993). Nuclear genes controlling male fertility. Plant Cell 5, 1277-1283.Christ, C., and Tye, B.R. (1991). Functional domains of the yeasttranscription/replication factor MCM1. Genes Dev. 5, 751-763.Coen, E.S., and Meyerowitz, E.M. (1991). The war of the whorls: Genetic interactionscontrolling flower development. Nature 353, 31-37.Datla, R.S.S., Hammerlindl, J.K., Panchuk, B., Pelcher, LE, and Keller, W. (1992).Modified binary plant transformation vectors with the wild-type gene encoding NPTII.Gene 211, 383-384.Davies, B., and Schwarz-Sommer, Z. (1994). Control of floral organ identity byhomeotic MADS-box transcription factors. In Plant promoters and transcriptionfactors: results and problems in cell differentiation 20, ed. Nover, L., Springer-Verlag,235-258.Davies, G.L (1966). Systematic embryology of angiosperms. ed. John Wiley andSons, New York, pp. 123-157.Dean, C., Sjodin, C., Page, T., Jones, J., and Lister, C. (1992). Behaviour of the maizetransposable element Ac in Arabidopsis thaliana. Plant J. 2, 69-81.Desplan, C., Theis, J. and OFarrell, P.H. (1985). The Drosophila developmental gene,engrailed, encodes a sequence-specific DNA binding activity. Nature 318, 630-635.Desplan, C., Theis, J. and O’Farrell, P.H. (1988). The sequence specificity ofhomeodomain-DNA interaction. Cell 54, 1081-1090.128Drews, G.N., Bowman, J.L, and Meyerowitz, E.M. (1991). Negative regulation of theArabidopsis gene AGAMOUS by the APETALA2 product. Cell 65, 991-1002.Edwards, K., Johnstone, C., and Thompson, C. (1991). A simple and rapid methodfor the preparation of plant genomic DNA for PCR analysis. Nuci. Acid Res. 19, 1349.Elledge, S.J., Mulligan, J.T., Ramer, S.W., Spotlswood, M., and Davies, R.W. (1991).AYES: a multifunctional eDNA expression vector for the isolation of genes bycomplementation of yeast and Escherichia coil mutations. Proc. Natl. Acad. Sci. USA88, 1731-1735.Errampalli, D., Patton, D., Castle, L, Mickelson, L, Hensen, K., Schnall, J., Feldmann,K., and Meinke, D. (1991). Embryonic lethals and T-DNA insertional mutagenesis inArabidopsis. Plant Cell 3, 149-157.Esau, K. (1977). Anatomy of seed plants. ed. John Wiley and Sons, New York, pp.403-427.Evans, Pt, and Malmberg, R..L (1989). Alternative pathways of tobacco placentaldevelopment: Time of commitment and analysis of a mutant. Dev. Biol. 136, 273-283.Feldmann, K.A. (1991). T-DNA insertion mutagenesis of Arabidopsis: Mutationalspectrum. Plant J. 1, 71-82.Feldmann, K.A., Marks, M.D., Christianson, M.L, and Quatrano, R.S. (1989). A dwarfmutant of Arabidopsis generated by T-DNA insertional mutagenesis. Science 243,1351-1 354.Fields, S., and Stemglanz, R. (1994). The two-hybrid system: an assay for protein-protein interactions. TIG 10, 286-292.Fortini, M.E., Zhichun, L, and Rubin, G.M. (1991). The Drosophila zfh-1 and zfh-2genes encode novel proteins containing both zinc-finger and homeodomain motifs.Mech. Develop. 34, 113-122.Freyd, G., Kim, S.K. and Horvitz, H.R. (1990). Novel cysteine-rich motif andhomeodomain in the product of the Caenorhabditis elegans cell lineage gene un-il.Nature 344, 876-879.Gehring, W.J. (1987). Homeo boxes in the study of development. Science 236, 1245-1252.Gehring, W.J., Affolter, M. and Burglin, K. (1994b). Homeodomain proteins. Annu.Rev. Biochem. 63, 437-526.129Gehring, W.J., Muller, M., Affolter, M., Percival-Smith, A., Billeter, M., Qian, Y.Q.,Otting, G., and Wuthrich, K. (1990). The structure of the homeodomain and itsfunctional implications. Trends. Genet. 6, 323-329.Gehring, W.J., Qian, Y., Billeter, M., Furokubo-Tokunaga, K., Schier, A.F., ResendezPerez, D., Affolter, M., Otting, G. and Wutrich, K. (1994a). Homeodomain-DNArecognition. Cell 78, 211-223.Gifford, E.M., and Foster, A.S. (1989). Morphology and Evolution of Vascular Plant.ed. W.H. Freeman, New York, pp. 634-71 3.Golubovskaya, l.N. (1989). Meiosis in maize: mel genes and conception of geneticcontrolof meiosis. Adv. Gemet. 26, 149-193.Golubovskaya, I.N., Avalkina, N.A., and Sheridan, W.F. (1992). Effects of severalmelotic mutants on female meiosis in maize. Dev. Genet. 13, 411-424.Goto, K., and Meyerowitz, E.M. (1994). Function and regulation of the Arabidopsisfloral homeotic gene PISTILLATA. Genes Dev. 8, 1548-1560.Grueneberg, D.A., Natesan, S., Alexandre, C. and Gilman, M.Z. (1992). Human andDrosophila homeodomain proteins that enhance the DNA-binding activity of serumresponse factro. Science 357, 1089-1094.Guignard, J.L, Agier, C., and Bury, M. (1991). Developpement de l’embryonzygotique chez l’Arabidopsis thaliana Schur. Bull. Soc. Bot. Fr., 138, 149-154.Gustafson-Brown, C., Savidge, B., and Yanofsky, M.F. (1994). Regulation of theArabidopsis floral homeotic gene APETALA1. Cell 76, 131-143.Haccius, B., Bhandari, N.N., and Hausner, G. (1974). In vitro transformation of ovulesinto rudimentary pistils in Nicotiana tabacum L. J. Exp. Bot. 25, 695-704.Hake, S., Vollbrecht, E. and Freeling, M. (1989). Cloning Knotted, the dominantmorphological mutant in maize using Ds2 as a transposon tag. EMBO J. 8, 15-22.Hake, S. (1992) Unraveling the knots of plant development. Trends. Genet. 8, 109-114.Hall, M.N. and Johnson, A.D. (1987). Homeodomain of the yeast repressor c2 is asequence-specific DNA-binding domain but is not sufficient for repression. Science237, 1007-1012.Hanes, S.D. and Brent, R. (1989). DNA specificity of the bicoid activator protein is130determined by homeodomain recognition helix residue 9. Cell 57, 1275-1283.Haughn, G.W., and Sommerville, C.R. (1986). Sulfonylurea-resistant mutants ofArabidopsis thaliana. Mol. Gen. Genet. 204, 430-434.Haughn, G.W., and Sommerville, C.R. (1988). Genetic control of morphogenesis inArabidopsis. Dev. Genet. 9, 73-89.Haughn, G.W., Schullz, E.A., and Martinez-Zapater. (1995). The regulation offlowering in Arabidopsis thaliana: meristems, morphogenesis and mutants. Can. J.Bot. (in press).Hayashi, S. and Scott, M.P. (1990). What determines the specificity of action ofDrosophila homeodomains protiens? Cell 63, 883-893.Hensel, L, Nelson, M.A., Richmond, TA., and Bleecker, A.B. (1994). The fate ofinflorescence meristems is controlled by developing fruits in Arabidopsis. PlantPhysiol. 106, 863-876.Herr, W., Sturm, R.A., Clerc, R.G., Corcoran, LM., Baltimore, D., Sharp, P.A.,lngrham, HA., Rosenfeld, M.G., Finney, M., Ruvkun, G., and Horvitz, H.R. (1988). ThePOU domain: a large conserved region in the mammalian pit-i, oct-i, oct-2, andCaenorhabditis elegans unc-86 gene products. Genes. Dev. 2, 1513-1516.Herrman, P.L, and Marks, M.D. (1989). Trichome development in Arabidopsisthaliana. Il. Isolation and complementation of the GLABROUS1 gene. Plant Cell 1,1051 -1 055.Herskowitz, I. (1989). A regulatory hierarchy for cell specialization in yeast. Nature342, 749-757.Hicks, G.S., and McHughen A. (1974). Altered morphogensis of placental tissues oftobacco in vitro: Stigmatoid and carpelloid outgrowths. Planta 121, 193-1 96.Hicks, G.S., and McHughen A. (1977). Ovule development in vitro from isolatedtobacco placental tissue. Plant Sci. Leff. 8, 141-145.Hill, J.P., and Lord, E.M. (1989). Floral development in Arabidopsis thaliana,comparison of the wild-type and the homeotic pistillata mutant. Can. J. Bat. 67, 2922-2936.Ho, Chi-Yip, Adamson, G., Hodges, R.S., and Smith, M. (1994). Heterodimerizationof the yeast MATa1 and MATcc2 proteins is mediated by two leucine zipper-like coiledcoil motifs. EMBO J. 13, 1403-1413.131Hoey, T. and Levine, M. (1988). Divergent homeo box proteins recognize similarsequences in Drosophila. Nature 322, 858-861.Huala, E. and Sussex, LM. (1992). LEAFY interacts with floral homeotic genes toregulate Arabidopsis floral development. Plant Cell 4, 901-913.Huang, H., Mizukami, Y., Hu, Y., and Ma, H. (1993). Isolation and characterization ofthe binding sequences for the product of the Arabidopsis floral homeotic geneAGAMOUS. Nucl. Acid Res. 21, 4769-4776.Huijser, P., Klein, J., Lonnig, W.-E., Meijer, H., Seadler, H., and Sommer, H. (1992).Bracteomania, an inflorescence anomaly, is caused by the loss of function of theMADS-box gene squamosa in Antirrhinum majus. EMBO J. 11, 1239-1249.lngham, P.W. (1988). The molecular genetics of embryonic pattern formation inDrosophila. Nature 335, 25-34.Ingraham, H.A., Flynn, S.E., Voss, J.W., Albert, V.R., Kapilof, M.S., Wilson, L, andRosenfeld, M.G. (1990). The POU-specific domain of Pit-i is essential for sequence-specific high affinity DNA binding and DNA dependent Pit-i -Pit-i interactions. Cell 61,1021-1033.Innis, M.A., Gelfand, D.H., Sninsky, J.J., and White, T.J. (1990). In: PCR protocols,a guide to methods and applications. Acadamic Press, Inc., Harcourt BraceJovanovich, New York, pp. 13-21.Irish, V.F. and Sussex, l.M. (1990). Function of the apetaia-1 gene duringArabidopsisfloral development. Plant Cell 2, 741 -753.Jack, T., Brockman, LL, and Meyerowitz, E.M. (1992). The homeotic geneAPETALA3of Arabidopsis thaliana encodes a MADS box and is expressed in petals andstamens. Cell 68, 683-697.Jack, T., Fox, G.L, and Meyerowilz, E.M. (1994). Arabidopsis homeotic geneAPETALA3 ectopic expression: transcriptional and posttranscriptional regulationdetermine floral organ identity. Cell 76, 703-716.Jackson, D., Veit, B. and Hake, S. (1994). Expression of maize KNQTTED1 relatedhomeobox genes in the shoot apical meristem predicts patterns of morphogenesisin the vegetative shoot. Development 120, 405-41 3.Jofuku, K.D., den Boer, B.G.W., Van Montagu, M., and Okamuro, J. (1994). Controlof Arabidopsis flower and seed development by the homeotic gene APETALA2. PlantCell 6, 1211-i 225.132Johnson, A.D. and Herskowitz, I. (1985). A repressor (MATt2product) and its operatorcontrol expression of a set of cell type specific genes in yeast. Cell 42, 237-247.Jones, J.D.G., Gilbert, D.E., Grady, K.L, and Jorgensen, R.A. (1987). T-DNA structureand gene expression in petunia plants transformed by Agrobacterium C58 derivatives.McI. Gen. Genet. 207, 478-486.Jorgensen, R., Snyder, C., and Jones, J.D.G. (1987). T-DNA is organizedpredominantly in inverted repeat structures in plants transformed with Agrobacteriumtumefaciens C58 derivatives. Mol. Gen. Genet. 207, 471-477.Kano-Murakami, Y., Yanai, T., Tagiri, A. and Matsuoka, M. (1993). A rice homeoticgene, OSH1, causes unusual phenotypes in transgenic tobacco. FEBS 334, 365-368.Katavic, V., Haughn, G.W., Reed, D., Martin, M., and Kunst, L (1994). In plantatransformation of Arabidopsis thaliana. Mo!. Gen. Genet. 245, 363-370.Kelly, M., Burke, J., Smith, M., Kiar, A., and Beach, D. (1988). Four mating type genescontrol sexual differentiation in the fission yeast. EMBO J. 7, 1537-1547.Kempin, S.A., Mandel, M.A., and Yanoisky, M.F. (1993). Conversion of perianth intoreproductive organs by ectopic expression of the tobacco floral homeotic geneNAG1. Plant Physic!. 103, 1041-1046.Kempin, S.A., Savidge, B., and Yanofsky, M.F. (1995). Molecular basis of thecauliflower phenotype in Arabidopsis. Science 267, 522-525.Kenell, J.K., and Homer, H.T. (1985). Influence of the soybean male-sterile gene(msl) on the development of the female gametophyte. Can. J. Genet. Cytol. 27, 200-209.Kenyon, C. (1994). If birds can fly, why can’t we? Homeotic genes and evolution. Cell78, 175-180.Kerstetter, R., Vollbrecht, E., Lowe, B., Veit, B., Yamaguchi, J., and Hake, 5. (1994).Sequence analysis and expression patterns divide the maize knottedl-like homeoboxgenes into two classes. Plant Cell 6, 1877-1887.Kissinger, C.R., Liu, B., Martin-Blanco, E., Komberg, T.B. and Pabo, C.O. (1990).Crystal structure of an engralled homeodomain-DNA complex at 2.8 A resolution: aframework for understanding homeodomain-DNA interaction. Cell 63, 579-590.Komaki, M.K., Okada, K., Nishino, E., and Shimura, Y. (1988). Isolation andcharacterization of novel mutants of Arabidopsis thaliana defective in flower133development. Development 104, 195-203.Koornneef, M., Hanhart, C.J., van Loenen-Martinet, E.P., and van der Veen, J.H.(1987). A marker line, that allows the detection of linkage on all Arabidopsischromosomes. Arabidopsis lnf. Serv. 23, 46-50.Korfhage, U., Treini, G.F., Hahlbrock, K. and Somssich, I.E. (1994). Planthomeodomain protein involved in transcriptional regulation of a pathogen defense-related gene. Plant Cell 6, 695-708.Kristie, T.M., and Sharp, PA. (1990). Interactions of the Oct-i POU subdomains withspecific DNA sequences and with the HSV alpha-trans-activatir protein. Genes. Dev.4, 2383-2396.Krumlauf, R. (1994). Hox genes in vertebrate development. Cell 78, 191-201.Kunst, L, Klenz, J.E., Martinez-Zapater, J., and Haughn, G.W. (1989). AP2 genedetermines the identity of perianth organs in flowers of Arabidopsis thaliana. Plant Cell1, 1195-1208.Landschulz, W.H., Johnson, P.F. and McKnight, S.L (1988). The leucine-zipper: Ahypothetical structure common to a new class of DNA binding proteins. Science 240,1759-1748.Lang, J.D., Ray, S., and Ray, A. (1994). sini, a mutation affecting female fertility inArabidopsis, interacts with modi, its recessive modifier. Genetics 137, 1101-1110.Laughon, A. (1991). DNA binding specificity of homeodomains. Biochemistry 30,11357-11367.Lawrence, P.A. and Morata, G. (1994). Homeobox genes: their function in Drosophilasegmentation and pattern formation. Cell 78, 181-189.Levine, M. and Hoey, T. (1988). Homeobox proteins as sequence specifictranscription factors. Cell 55, 537-540.Lincoln, C., Long, J., Yamaguchi, J., Serikawa, K., and Hake, S. (1994). A knottedilike homeobox gene in Arabidopsis is expressed in the vegetative meristem anddramatically alters leaf morphology when overexpressed in transgenic plants. PlantCell 6, 1859-1976.Ma, H. (1994). The unfolding drama of flower development: recent results fromgenetic and molecular analyses. Genes Dev. 8, 745-756.134Ma, H., McMullen, M.D. and Finer, J.J. (1994). Identification of a homeoboxconataining gene with enhanced expression during soybean (Glycine max L.) somaticembryo development. Plant Mol. Biol. 24, 465-473.Ma, H., Yanofsky, M.F., and Meyerowilz, E.M. (1991). AGL1-AGL6, an Arabidopsisgene family with similarity to floral homeotic and transcription factor genes. GenesDev. 5, 484-495.Mandel, M.A., Bowman, J.L, Kempin, S.A., Ma, H., Meyerowilz, E.M., and Yanofsky,M.F. (1992a). Manipulation of flowerstructure in transgenictobacco. Cell 71, 133-143.Mandel, M.A., Gustafson.Brown, C., Savidge, B., and Yanofsky, M.F. (1992b).Molecular characterization of the Arabidopsis floral homeotic gene APETALA 1. Nature360, 273-277.Mansfield, S.G., and Briarty, LG. (1991). Early embryogenesis in Arabidopsis thaliana.II. The developing embryo. Can. J. Bot. 69, 461-476.Mansfield, S.G., Briarty, LG., and Emi, S. (1991). Early embryogenesis inArabidopsisthaliana. I. The mature embryo sac. Can. J. Bot. 69, 447-460.Marks, M.D., and Feldmann, K.A. (1989). Trichome development in Arabidopsisthaliana. I. T-DNA tagging of the GLABROUS1 gene. Plant Cell 1, 1043-1050.Martinez-Zapater, J.M., Coupland, G., Dean, C., and Koomneef, M.K.. (1995). Thetransition to flowering in Arabidopsis. In Arabidopsis, eds. E.M. Meyerowitz and C.R.Somerville, Cold Spring Harbor Press (in press).Masters, M.T. (1869). Vegetable teratology: an account of the principle deviationsfrom the usual construction of plants. Ray Society, London.Matsuoka, M., lchikawa, H., Saito, A., Tada, Y., Fujimura, T. and Kano-Murakami, Y.(1993). Expression of a rice homeobox gene causes altered morphology oftransgenic plants. Plant Cell 5, 1039-1048.Mattsson, J., Soderman, E., Svenson, M., Borkird, C. and Engstrom, P. (1992). A newhomeobox-leucine zipper gene from Arabidopsis thaliana. Plant Mol. Biol. 18, 1019-1022.McBride, KE., and Summerfelt, KR. (1990). Improved binary vectors forAgrobacterium-mediated plant transformation. Plant Mol. Biol. Rep. 14, 269-276.McGinnis, W., Levine, M.S., Hafen, E., Kuroiwa, A. and Gehring, W.J. (1984). Aconserved DNA sequence in homeotic genes of the Drosophila antennapedia135complexes. Nature 308, 428-433.McGinnis, W. and Krumlauf, R. (1992). Homeobox genes and axial patterning. Cell68, 283-302.Meyerowflz, E.M., Bowman, J.L, Brockman, L, Drews, G.N., Jack, T., Sieburth, LE.,and Weigel, D. (1991). A genetic and molecular model for flower development inArabidopsis thaliana. Development Suppl. 1, 157-167.Misra, R.C. (1962). Contribution to the embryology of Arabidopsis thalianum (Gay &Monn.). Univ. J. Res. Agra. 11, 191-198.Mizukami, Y. and Ma, H. (1992). Ectopic expression of the floral homeotic geneAGAMQUS in transgenic Arabidopsis plants alters floral organ identity. Cell 71, 119-131.Modrusan, Z., Reiser, L, Fischer, L, and Haughn, G. (1994a). Ontogeny of theArabidopsis wild type ovule. In: Arabidopsis, an atlas of morphology anddevelopment, ed. J. Bowman, Springer-Verlag, New York, pp. 302.Modrusan, Z., Reiser, L, Feldmann, K.A., Fischer, R.L, and Haughn, G.W. (1994b).Homeotic transformation of ovules into carpel-Iike structures in Arabidopsis. Plant Cell6, 333-349.Morelli, G., Baima, S., Carabelli, M., Lucchetti, S., Nobili, F.,Sessa, G. and Ruberti,I. (1993). Identification of distinct families of HD-Zip proteins in Arabidopsis thaliana.J. Cell. Biochem. Suppl. 17B, pp. 43.Mueller, C.G., and Nordheim, A. (1991). A protein domain conserved between yeastMCM1 and human SRF directs ternary complex formation. EMBO J. 10, 989-1003.Muller, A. (1961). Zur Charakterisierung der Blueten und lnfloreszen von Arabidopsisthaliana (L.) Heynh. Kulturpflanze 9, 364-393.Muller, M., Affolter, M., Leupin, W., Otting, G., Wuthrich, K. and Gehring, W.J. (1988).Isolation and sequence-specific DNA binding of the Antennapedia homeodomain.EMBO J. 7, 4299-4304.Murray, N.E. (1983). Phage lambda and molecular cloning. In: Lambda II, eds. R.W.Hendrix, J.W. Roberts, F.W. Stahl, and R.A. Weisberg, Cold Spring Harbor LaboratryPress, NY. pp. 395-432.Norman, C., Runswick, M., Pollock, R., and Triesman, R. (1988). Isolation of cDNAclones encoding SRF, a transcription factor that binds to the c-fos serum response136element. Cell 55, 989-1003.Okada, K., Komaki, M.K., and Shimura, Y. (1989). Mutational analysis of pistilstructure and development of Arabidopsis thaliana. Cell Duff. Dev. 28, 27-38.Okamuro, J.K., den Boer, B.G.W., and Jofuku, lCD. (1993). Regulation of Arabidopsisflower development. Plant Cell 5, 1183-1193.Oiling, G., Qian, Y.Q.., Billeter, M., Aflolter, M., Gehring, W.J., and Wuthrich, K. (1990).Protein-DNA contacts in the structure of a homeo domain-DNA complex determinedby nuclear magnetic resonance-spectroscopy in solution. EMBO J. 9, 3085-3092.Passmore, S., Maine, G.T., Elible, C., Christ, C., and Tye, B.K. (1988). ASaccharomyces cerevisiae protein involved in plasmid mainteneance is necessary formating of MATa cells. J. Mol. Bid. 204, 593-606.Percival-Smith, A., Muller, M., Affolter, M. and Gehring, W.J. (1990). The interactionwith DNA of wild-type and mutant fushi tarazu homeodomains. EMBO J. 9, 3967-3974.Peterson, CM. (1973). Nutritional requirements for ovule formation in excised pistilsof Nigella. Amer. J. Bot. 60, 381 -386.Pnuelli, L, Abu-Abeid, M., Zamir, D., Nacken, W., Schwarz-Sommer, Z., and Lifschilz,E. (1991). The MADS box gene family in tomato: temporal expression during floraldevelopment, conserved secondary structures and homology with homeotic genesfrom Antirrhinum and Arabidopsis. Plant J. 1, 255-266.Pnuelli, L, Hareven, D., Rounsley, S.D., Yanofsky, M.F., and Lifschitz, E. (1994).Isolation of the tomato AGAMQUS gene TAG1 and analysis of its homeotic role intransgenic plants. Plant Cell 6, 163-173.Pollock, R., and Treisman, R. (1991). Human SRF-related proteins: DNA-bindingproperties and potential regulatory targets. Genes. Dev. 5, 2327-2341.Polowick, P.L, and Sawhney, V.K. (1986). A scanning electron microscopic study onthe initiation and development of floral organs of Brassica napus (cv. westar). Am. J.Bot. 73, 254-263.Primig, M., Winkler, H., and Ammerer, G. (1991). The DNA binding andoligomerization domain of MCM1 is sufficient for its interaction with other regulatoryproteins. EMBO J. 10, 4209-4218.Ptashne. M., and Gann, A.F. (1990). Activators and targets. Nature 346, 329-331.137Qian, Y.Q., Billeter, M., Oiling, G., Muller, M., Gehring, W.J. and Wuthrich, K. (1989).The structure of Antennapedia homeo domain determined by NMR spectroscopy insolution: comparison with prokaryotic repressors. Cell 59, 573-580.Ray,A., Robinson-Beers, K, Ray, S., Baker, S.C., Lang, J.D., Preuss, D., Milligan,S.B., and Gasser, C.S. (1994). The Arabidopsis floral homeotic gene BELL controlsovule development through negative regulation of AGAMOUS. Proc. Nati. Acad. Sd.USA 91, 5761-5765.Redel, G. (1964). Non-mendelian megagametogenesis in Arabidopsis. Genetics 51,857-872.Reiser, L, and Fischer, R. L (1993). The ovule and the embryo sac. Plant Cell 5,1291-1301.Rerie, W.G., Feldmann, K.A. and Marks, M.D. (1994). The GLABRA2 gene encodesa homeo domain protein required for normal trichome development in Arabidopsis.Genes Dev. 8, 1388-1399.Robinson-Beers, K., Pruitt, R.E., and Gasser, CS. (1992). Ovule development in wild-type Arabidopsis and two female-sterile mutants. Plant Cell 4, 1237-1249.Ruberti, I., Sessa, G., Lucchetti, S., and Morelli, G. (1991). A novel class of plantproteins containing a homeodomain with a closely linked leucine zipper motif. EMBOJ. 10, 1787-1791.Russel, S.D. (1978). Fine structure of megagametophyte development in Zea mays.Can. J. Bot. 57, 1093-1110.Sambrook, J., Fritsch, E.F., and Maniatis, T. (1989). Molecular cloning, a laboratorymanual. Cold Spring Harbor Laboratry Press, NY.Sanchez-Garcia, I., Osada, H., Forster, A., and Rabbitts, T.H. (1993). The Cysteinerich LIM domains inhibit DNA binding by the associated homeodomains in lsl-1.EMBO J. 12, 4243-4250.Satina, S., and Blakeslee, A.F. (1935). Cytological effects of a gene in Datura whichcauses dyad formation in sporogenesis. Bot. Gaz. 96, 521 -532.Schena, M. and Davies, R.W. (1992). HD-Zip proteins: members of and Arabidopsishomeodomain protein superfamily. Proc. Nati. Acad. Sci. USA 89, 3894-3898.Schena, M. and Davies, R.W. (1994). Structure of homebox-leucine zipper genessuggests a model for the evolution of gene families. Proc. NatI. Acad. Sci. USA 91,1388393-8397.Schena, M., Lloyd, A.M. and Davies, R.W. (1993). The HAT4 gene of Arabidopsisencodes a developmental regulator. Genes Dev. 7, 367-379.Schier, &F. and Gehring, W.J. (1992). Direct homeodomain-DNA interaction in theregulation of the fushi tarazu gene. Nature 356, 804-807.Schier, A.F. and Gehring, W.J. (1993). Analysis of a fushi tara.zu autoregulatoryelement: multiple sequence elements contribute to enhancer activity. EMBO J. 12,1111-1119.Schindler, U., Beckmann, H. and Cashmore, A.R. (1993). HAT3.1, a novel Arabidopsishomeodomian protein containing a conserved cysteine-rich region. Plant J. 4, 137-150.Schmidt, R.J., Veit, B., Mandel, M.A., Mena, M., Hake, S., and Yanofsky, M.F. (1993).Identification and molecular characterization of ZAG1, the maize homolog of theArabidopsis floral homeotic gene AGAMQUS. Plant Cell 5, 729-737.Schultz, E.A., and Haughn, G.W. (1993). Genetic analysis of floral initiation process(FLIP) in Arabidopsis. Development 119, 745-765.Schultz, E.A., Pickett, F.B. and Haughn, G.W. (1991). The FLOW gene productregulates the expression domain of homeotic genes AP3 and P1 in Arabidopsisflowers. Plant Cell 3, 1221-1237.Schwa,z-Sommer, Z., Huijser, P., Nacken, W., Saedler, H., and Sommer, H. (1990).Genetic control of flower development by homeotic genes in Antirrhinum majus.Science 250, 931-936.Schwarz-Sommer, Z., Hue, I., Huijser, P., Flor, P.J., Hansen, R., Tetens, F., Lonnig,W., Saedler, H., and Sommer, H. (1992). Characterization of the Antirrhinum floralhomeotic MADS-box gene deflciens: evidence for DNA binding and autoregulationof its persistent expression throughout flower development. EMBO J. 11, 251-1 63.Scott, M.P., and Weiner, A.J. (1984). Structural relationship among genes that controldevelopment: sequence homology between the Antennapedia, Ultrabithorax and fushitarazu loci of Drosophila. Proc. NatI. Acad. Sd. USA 81, 4115-4119.Scott, M.P., Tamkun, J.W. and Harizell, G.W., Ill (1989). The structure and functionof the homeodomain. BBA Rev. Cancer 989, 25-48.Sessa, G., Morelli, G. and Ruberti, I. (1993). The Athd-1 and -2 HD-Zip domains139homodimerize forming complexes of different DNA binding specificities. EMBO J. 12,3507-3517.Shannon, S., Meeks-Wagner, D.R. (1993). Genetic interactions that regulateinflorescence development in Arabidopsis. Plant Cell 5, 639-655.Shiraishi, H., Okada, K., and Shimura, Y. (1993). Nucleotide sequences recognizedby the AGAMOUS MADS domain of Arabidopsis thaliana in vitro. Plant J. 4, 385-398.Sinha, N.R. and Hake, S. (1990). Mutant characters of Knotted maize leaves aredetermined in the innermost tissue layers. Dev. Biol. 141, 203-21 0.Sinha, N.R., Williams, R.E. and Hake, S. (1993). Overexpression of the maize homeobox gene KNOTTED-i, causes a switch from determinate to indeterminate cell fates.Genes Dev. 7, 787-795.Smith, LG., Greene, B., Veit, B. and Hake, S. (1992). A dominant mutation in themaize homeobox gene, knotted-i, causes its ectopic expression in leaf cells withaltered fates. Development 116, 21-30.Smoller, D., Friedal, C., Schmid, A., Bettler, D., Lam, L, and Yedvobnick, B. (1990).The Drosophila neurogenic locus mastermind encodes a nuclear protein unusuallyrich in amino acid homopolymers. Genes Dev. 4, 1688-1700.Smyth, D.R., Bowman, J.L, and Meyerowitz, E.M. (1990). Early flower developmentin Arabidopsis. Plant Cell 2, 755-767.Sommer, H., Beltran, J, Huijser, P., Pape, H., Lonnig, W.-E., Seadler, H., andSchwarz-Sommer, Z. (1990). Deficiens, a homeotic gene involved in the control offlower morphogenesis in Antirrhinum majus: the protein shows homology totranscription factors. EMBO J. 9, 605-613.Suiter, K.A., Wendel, J.F., and Case, J.S. (1983). LINKAGE-i: A PASCAL computerprogram for the detection and analysis of genetic linkage. J. Hered. 74, 203-204.Sumner, M.J., and Van Caeseele, L (1989). Ovule development in Brassicacampestris. A light microscope study. Can. J. Bot. 66, 2459-2469.Stern, S.A., Tanaka, M. and Herr, W. (1989). The Oct-i homeodomain directsformation of a multiprotein-DNA complex with the HSV transactivatior VP16. Nature341, 624-630.Treisman, J., Gonczy, P., Vashishtha, M., Harris, E. and Desplan, C. (1989). A singleamino acid can determine the DNA binding specificity of homeodomain proteins. Cell14059, 553-562.Treisman, J., Harris, E., and Desplan, C. (1991). The Paired box encodes a secondDNA-binding domain in the Paired homeodomain protein. Genes. Dev. 5, 594-604.Treisman, J., Harris, E., Wilson, D. and Desplan, C. (1992). The homeodomain: a newface for the helix-turn-helix? Bioessays 14, 145-150.Trobner, W., Ramirez, L, Motte, P., Hue, I., Huijser, P., Lonnig, W.-E., Seadler, H.,Sommer, H., and Schwarz-Sommer, Z. (1992). GLOBOSA: a homeotic gene whichinteracts with DEFICIENS in the control of Antirrhinum floral organogenesis. EMBO J.11, 4693-4704.Tsuchimoto, S., van der Krol, A.R., and Chua, N.-H. (1993). Ectopic expression ofpMADS3 in transgenic petunia phenocopies the petunia blind mutant. Plant Cell 5,843-853.Valvekens, D., Van Montagu, M., and Van Lijsebettens, M. (1988). Agrobacteriumtumefaciens-mediated transformation of Arabidopsis thaliana root explants by usingkanamycin selection. Proc. NatI. Acad. Sci. USA 85, 5536-5540.van der Krol, A.R., Brunelle, A., Tsuchimoto, S., and Chua, N.-H. (1993). Functionalanalysis of petunai floral homeotic MADS box gene pMADS1. Genes Dev. 7, 1214-1228.Vinson, C.R., Sigler, P.B. and McKnoght, S.L (1989). Scissors-grip model for DNArecognition by a family of leucine zipper proteins. Science 246, 911-916.Veit, B., Vollbrecht, E., Mathem, J. and Hake, S. (1990). A tandem duplication causesthe Kni-O alleles of knotted, a dominant morphological mutant of maize. Genetics125, 623-631.Velten, J., and Schell, J. (1985). Selection-expression plasmid vectors for use ingenetic transformation of higher plants. Nuci. Acids Res. 13, 6981-6998.Vollbrecht, E., Veit, B., Sinha, N. and Hake, S. (1991). The developmental geneKnotted-i is a member of a maize homeobox gene family. Nature 350, 241 -243.Webb, M.C., and Gunning, B.E.S. (1990). Embryo sac development in Arabidopsisthaliana. 1. Megasporogenesis, including the microtubular cytoskeleton. Sex PlantReprod. 3, 244-256.Webb, M.C., and Gunning, B.E.S. (1991). The microtubular cytoskeleton duringdevelopment of the zygote, proembryo and free-nuclear endosperm in Arabidopsis141thaliana (L.) Heynh. Planta 184, 187-195.Weigel, D., and Meyerowitz, E.M. (1994). The ABCs of floral homeotic genes. Cell 78,203-209.Wolberger, C., Vershon, AK, Uu, B, Johnson, A.D. and Pabo, C.O. (1991). Crystalstructure of a MATa2 homeo domain-operator complex siggests a general model forhomeo domain-DNA interactions. Cell 67, 517-528.Yanofsky, M.F.., Ma, H., Bowman, J.L, Drews, G.N., Feldmann, K.A., and Meyerowilz,EM. (1990). The protein encoded by the Arabidopsis homeotic gene AQAMOUSresembles transcription factors. Nature 346, 35-40.142


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items