Turning floral organs into leaves, leaves into floral organs Koji Goto*, Junko Kyozuka and John L Bowman

449 Turning floral organs into leaves, leaves into floral organs Koji Goto*, Junko Kyozuka and John L Bowman The development of the floral organs is specified by the combinations of three classes of gene for organ identity in the ABC model. Recently, molecular genetic studies have shown this model is applicable to grass plants as well as most eudicots. Transcription factor complexes of ABC and homologous proteins form the molecular basis of the ABC model. Addresses *Research Institute for Biological Sciences, Kayo-cho, Jobo, Okayama, 716-1241, Japan; e-mail: kgoto@v004.vaio.ne.jp Department of Biosciences, Nara Institute of Science and Technology, 8916-5, Takayama, Ikoma 630-0101, Japan; e-mail: junko@bs.aist-nara.ac.jp Section of Plant Biology, University of California Davis, Davis, California 95616, USA; e-mail: jlbowman@ucdavis.edu Correspondence: Koji Goto 2001, 11:449 456 0959-437X/01/$ see front matter 2001 Elsevier Science Ltd. All rights reserved. Abbreviations ag agamous agl agamous-like ap1 apetala1 ap2 apetala2 ap3 apetala3 cal cauliflower crc crabs claw MADS MCM1-AMOUS-DEFICIENS-SRF1 OsMADS Oryza sativa MADS box pi pistillata sep sepallata Si1 silky1 spt spatula Z Zea mays AMOUS-LIKE ZMM Zea mays MADS box Introduction The application of developmental genetics has led to progress in our understanding of how organ type is specified in the flower. The advances have been made through genetic analyses of mutants that specifically disrupt floral development, and through molecular cloning of the corresponding genes to reveal the nature of their biochemical function. A model (now known widely as the ABC model ; Figure 1) based on genetic data was proposed to explain how a limited set of genes acting alone and in combination could specify the identity of the floral organs [1 4]. Subsequent molecular analyses of the relevant genes and ectopic expression studies in Arabidopsis have largely supported this genetic model and have led to numerous refinements and further insights (reviewed in [5 8]). In this review we summarize recent progress made on three outstanding issues. First, how widely applicable is the ABC model, which was originally based on two evolutionarily divergent eudicot species Arabidopsis thaliana and Antirrhinum majus? Second, how is the activity of the floral homeotic genes confined to flower meristems? Third, are there additional genes that act in parallel with the floral homeotic ABC genes to promote floral organ identity? The floral homeotic mutants of Arabidopsis The Arabidopsis flower consists of four concentric whorls of organs: sepals, petals, stamens and carpels. Mutations in a set of floral homeotic genes result in the misinterpretation of positional information in the developing flower, and subsequent homeotic transformation of floral organ types. These floral homeotic mutants fall into three classes, designated A, B and C, and each of the mutations results in organ identity defects in two adjacent whorls. Class A mutants (apetala2 [ap2] and apetala1 [ap1]) have homeotic conversions in the first two whorls, with, in the case of ap2, the first whorl organs developing as carpels rather than sepals, and the second whorl organs developing as stamens rather than petals [1,4,9 15]. Class B mutants (pistillata [pi] and apetala3 [ap3]) have alterations in the middle two whorls, with sepals instead of petals developing in the second whorl and carpels instead of stamens in the third whorl [1,4,16 18]. The inner two whorls are affected in class C mutants (agamous [ag]), with petals developing in place of stamens and another flower replacing the carpels [1,4,19]. There are three basic tenets of the ABC model [4]: first, each of the classes of homeotic gene function acts in a field comprised of two adjacent whorls, the particular whorls being those that are altered when the corresponding genes are mutant; second, the combination of floral homeotic gene activities specifies the type of organ that develops; and third, the class A and class C activities are mutually antagonistic, such that loss of A results in C activity in all four whorls and vice versa. Turning floral organs into leaves The ABC model successfully predicts the phenotypes of the floral organs for most of the singly, doubly and triply mutant genotypes examined. But when all of the A, B and C classes of floral homeotic activities are removed, as in an ABC triple mutant, the model does not predict the type of organ that will develop. One might suppose that if all genes required for the specification of floral organ identity were removed, the identity of the resulting organs might represent a ground state, possibly a leaf-like organ. When the floral organs of the ABC triple mutant are examined, they exhibit features of both carpels (such as stigmatic tissue, fusion of organs along their margins and ovules) and leaves (stellate trichomes and stipules) [4,20 ]. Thus, there must exist genes that promote carpel development in the absence of gene function.

450 Pattern formation and developmental mechanisms Figure 1 AP2 + A AP3 + PI B AP3 + PI B AP2 + A C se pe st ca se pe st ca st 1 2 3 4 3 2 1 pe se The ABC model and the specification of floral organ identity. A wild-type flower and its floral diagram. The ABC model of the specification of floral organ identity depicts how three classes of floral homeotic genes can specify the identity of each of the four whorls of floral organs. Class A alone specifies sepals, classes A + B specify petals, classes B + C specify stamens, and class C alone specifies carpels. Abbreviations: ca, carpel; pe, petal; se, sepal; st, stamen. A section through a floral primordium, depicted as a set of boxes with the regions representing each whorl indicated beneath each column. The top boxes each represent a single field (A, red; B, yellow; C, blue). The distribution of floral homeotic gene products present in each of the whorls is indicated inside the boxes, with the phenotype of the organs in each whorl indicated beneath the lower set of boxes: sepal (red), petal (orange), stamen (green), carpel (blue). In a series of elegant genetic experiments, Alvarez and Smyth [20 ] have demonstrated that CRABS CLAW (CRC) and SPATULA (SPT) act in an -independent manner to promote several aspects of carpel differentiation. Both spt and crc mutations have a phenotypic effect in an ABC triple mutant background, reducing the amount and type of carpel tissues that develop, with mutations in spt having the more marked effect ([20 ]; Figure 2). In the ap2 pi ag spt quadruple mutant, residual occasional ovule-like structures develop at the margins of the phylloid floral organs. Some of these marginal outgrowths also develop in ap2 pi ag spt crc pentuple mutants [20 ], which suggests that additional genes also promote -independent develop-ment of marginal carpel tissues, candidates being LEUNIG and AINTEGUMENTA [21 ]. Consistent with the presence of additional redundant genes is the observation that spt single mutants do not lack most marginal tissues, but instead primarily affect the differentiation of septal tissues [20 ]. Because ag mutations are epistatic to spt mutations, SPT, like, may also be negatively regulated by A class genes [20 ]. SPT encodes a basic helix loop helix transcription factor and its expression in the developing carpel is consistent with the proposed role of promoting the growth of marginal tissues [22 ]. SPT is also expressed in many different tissues outside the carpel, suggesting that it may act redundantly in diverse developmental processes. The negative regulation of SPT by the A class genes is at the transcriptional level and, furthermore, the restriction of SPT expression to the marginal tissues of the carpel is mediated by ETTIN activity [22 ]. In ettin mutants, marginal tissue (e.g. septum) develops ectopically in non-marginal positions [23]. The ectopic development of septum in ettin mutants is mediated by ectopic SPT activity, as confirmed by the epistasis of SPT mutations over ettin mutations with respect to this phenotype [22,24]. Is ABC model applicable to the grass flower? Grass species have flowers with highly derived structures, and the identity of the lodicules, palea and lemma, which surround the reproductive organs has been unclear (Figure 3; [25 27]). Issues regarding the evolution of floral organs in grass species are now being addressed using the ABC model as a baseline. Grass flower homologs of ABC genes The isolation of putative class A, B and C genes from rice and maize suggests that the basic mechanisms of flower development are probably conserved between grasses and eudicots [28 33,34,35 ]. In both maize and rice, class C genes are expressed in the inner two whorls, class B gene transcripts accumulate in stamens and lodicules, and a putative rice class A gene, RA, is expressed in lodicules, the palea and lemma [34 ]. Although expression patterns can be correlative, a functional assay is needed to further assess any potential homology. Recently, Ambrose et al. [35 ] have identified silky1 (si1) as the AP3 ortholog in maize. Remarkably, in si1 mutants, stamens and lodicules are homeotically transformed into carpels and palea/lemma-like structures, respectively [35 ]. Likewise, reduced function of the rice PI ortholog (OsMADS4) by antisense methodology results in a similar phenotype [36]. These data imply that there is a homologous relationship between lodicules and petals, and that B function is conserved in grass flowers.

Turning floral organs into leaves, leaves into floral organs Goto, Kyozuka and Bowman 451 Figure 2 Figure 3 le st pa ap2 pi ag ap2 pi ag spt lo ca le ca lo st pa ap2 pi ag (ABC) triple mutants develop placentae bearing ovules, septum containing transmitting tract tissue, style tissue, and stigmatic papillae along the margins of their floral organs [4,20 ]. The only major carpel tissue not found in ap2 pi ag triple mutants is the specialized histology of the ovary wall [20 ]. In contrast, floral organs of ap2 pi ag spt quadruple mutants lack most of the carpelloid marginal tissues found in the ABC triple mutant, indicating that SPT promotes the differentiation of all these tissue types a role not obvious from the spt single mutant phenotype in which only the septum is lacking [20 ]. Identity of lemma and palea In contrast to the progress in understanding the identity of lodicules, the nature of lemma and palea is still uncertain. Palea/lemma-like structures develop in place of the lodicules in si1 mutants, which suggests that there is a possible homology between palea/lemma and sepals [35 ]. In addition, the expression pattern of the RA is consistent with the palea/lemma representing the calyx [34 ]. In contrast, an absence of any morphological changes in the palea or lemma in Act1::R plants (J Kyozuka, unpublished data) argues against this view and raises the possibility that the palea and lemma may not be floral organs whose identities are controlled by class A, B and C genes. However, it is also possible that ectopic expression of the class C gene is not sufficient to specify carpel identity in grasses. Recently, Jeon et al. [37] demonstrated that a mutation in OsMADS1, an L9 subfamily rice MADS-box gene, resulted in the elongation of palea and lemma and a reduction in floral meristem determinacy. Analysis of loss-of-function phenotypes of the grass class A genes, as well as genes with related sequences, such as OsMADS1, will be critical in elucidating the nature of the palea/lemma. Homology between lodicules and eudicot petals In eudicots, the ABC model has enabled the prediction of modified flower structures through manipulating the expression of the floral organ identity genes. For example, ectopic expression of by a constitutive promoter results in homeotic transformations of sepals to carpels and petals to stamens in eudicot flowers [38,39]. In rice, the ectopic expression of R/OsMADS3, a rice ortholog driven by the strong Actin1 promoter, caused homeotic conversions of lodicules to stamens (J Kyokuka, unpublished data). A rice spikelet and diagram of a transverse section through a flower. A rice flower, which is enclosed by a pair of bract-like structures called the lemma (le) and palea (pa), is composed of a pistil comprising three carpels (ca), six stamens (st) and two lodicules (lo). The lemma is detached from original position (arrow) to show lodicules. This increases the evidence in favour of lodicules being genetically homologous to petals. Together, homology between lodicules and eudicot petals is supported by three different lines of evidence: first, analyses of RNA expression patterns of class A and B genes; second, loss-offunction mutant phenotypes of class B genes; and third, gain-of-function alleles of class C genes. Together, these results suggest that the ABC genes control floral organ development in a similar manner in the grasses and the eudicots, which span the majority of angiosperm phylogenetic diversity [40]. Carpel development in grass flowers Grass species may have acquired unique mechanisms for carpel development. It has been proposed that in maize the C function is shared by two orthologs, Z1 and ZMM2 [29,30]. The expression patterns of Z1 and ZMM2 suggest that these genes may be involved in determining the carpel and stamen identities, respectively. In the zag1 mutant, however, in spite of a biased expression of Z1 in the carpel, there were no clear morphological alterations in this organ and only floral meristem determinacy was affected [29]. This raises the possibility that carpel identity may not depend solely on the class C homeotic genes in grass species, and that the function of another unknown gene(s) may also be required for the specification of carpels. Turning leaves into floral organs Molecular cloning of the ABC genes in Arabidopsis has revealed that all except AP2 share a conserved DNAbinding domain, called the MADS domain or MADS box [3,14,15,17 19]. By manipulating the expression of the ABC genes, one can construct flowers with of any floral organs in any whorl [4,39,41]. Ectopic expression of the ABC genes in leaves fails, however, to convert them

452 Pattern formation and developmental mechanisms Figure 4 AP3::GUS reporter gene expression in plants constitutively expressing PI AP3 and PI:VP16 AP3. AP3::GUS is activated by PI AP3 in the floral organs, but not in the vegetative organs (i.e. leaves). When the viral transactivation domain VP16 is fused to PI (PI:VP16), AP3::GUS becomes activated in the leaves. This implies that a flower-specific cofactor with a transactivation domain is required for complete B gene activity. into floral organs [39,41]. Thus, the ABC genes are necessary for the formation of floral organs, but they are not sufficient for the conversion from vegetative leaves into floral organs. Another factor is needed More than 200 years ago, Goethe proposed that floral organs are the result of a transformation ( metamorphosis ) of the basic leaves. Thus, there should exist as yet unidentified factors required for this transformation or, alternatively, it might be that vegetative leaves are not the basic organ from which the floral organs were derived. Recently, Honma and Goto [42 ] found the missing factor by searching for proteins that interact with the ABC MADS-box proteins. The B gene proteins, PI and AP3, act as a heterodimer and autoregulate their own transcription ([41,43 45]; reviewed in [8]). The transcription of an AP3::GUS reporter gene is activated by PI AP3 without de novo protein synthesis in the flower [43], but is not activated outside the flower even when PI and AP3 are expressed constitutively ([41,43]; Figure 4). However, AP3::GUS expression is observed in the leaves when a viral transcriptional activator domain, VP16, is fused with PI (35S::PI:VP16; Figure 4). This suggested that a flower-specific cofactor might supply a transcriptional activator domain to the PI AP3 complex. Thus, the putative missing factor was isolated by fishing for proteins that interact with the PI AP3 complex, and was identified as another MADS-box gene, L9 (agamous-like 9) [42,46]. Other kinds of proteins were not isolated in the yeast two-hybrid screening, suggesting that the entire complex may be composed solely of MADS-box proteins, and that L9 may supply a transcriptional activator domain. Notably, Davies et al. [47] and Egea- Cortines et al. [48 ] have demonstrated that Antirrhinum ABC MADS-box proteins exhibit a higher affinity and specificity with target DNA sequences when part of a ternary complex than when part of a dimer. L9 () is the missing factor Recently, L9 was renamed SEPLALLATA3 () based on its mutant phenotype [49 ]. The sep3 mutant itself shows only a subtle phenotype, but in combination with sep1 and sep2 mutants (knockouts of the closely related genes L2 and L4, respectively), all floral organs develop as sepal-like organs. Thus, the triply mutant sep123 flowers resemble those of bc double mutants [49 ]. The most striking evidence demonstrating that is the missing factor is the phenotypes of doubly, triply or quadruply transgenic plants combining ectopic expression of with ectopic expression of the ABC genes. Remarkably, leaves are converted into petals by ectopic expression of PI AP3, PI AP3 [42 ] or PI AP3 [50 ], into stamens by PI AP3 [42 ], and into carpels by (K Goto, unpublished data; Figure 5). These observations indicate that when gene function is added in combination with ABCgenes, vegetative leaves can be converted into floral organs, supporting Goethe s theory. Complexes of the MADS-box proteins are the molecular basis of the ABC model Further analyses of protein protein interactions suggest that not only interacts with PI and AP3, but also serves a scaffold between PI AP3 and [42 ]. Although PI AP3 and (the B and C proteins) cannot interact directly, if is added then the quartet complex, PI AP3 (third whorl complex), can form. The interaction suggests that the fourth whorl complex should be. Likewise, PI AP3 can interact with both and, and and interact with each other [42 ]; therefore, the second whorl complex should be PI AP3. What about the first whorl? The SEP proteins may not be required for a first whorl complex as the sep123 triple mutant develops normal sepals. Because MADS-box proteins form a dimer to bind their target DNA sequence (called the CArG-box ; reviewed in [8]), might form

Turning floral organs into leaves, leaves into floral organs Goto, Kyozuka and Bowman 453 Figure 5 Phenotypes of (a,b) wild-type and (c f) transgenic plants. A similar developmental stage (shortly after bolting) is shown in (a,c,e). The inflorescence stem (is), cauline leaf (cl) and lateral shoot (ls) are depicted in (b,d,f). Rosette leaves (rl) are transformed (c) into petals by the complex PI AP3 and (e) into carpels by the complex. Cotyledons (ct) are not transformed. Cauline leaves (cl) are converted (d) into petals by the complex PI AP3 and (f) into stamens by the complex PI AP3. Is (c) (e) ct rl rl ct ct rl (d) Is (f) Is cl cl cl is is is homodimers or heterodimers with CAULIFLOWER (CAL). and CAL can interact with each other (K Goto, unpublished results), but cal single mutants do not exhibit any aberrant phenotype although they enhance ap1 phenotypes [13,51]. Ectopic expression of or CAL does not result in a transformation of leaves into sepals [51,52], however, which suggests the existence of an as yet unidentified factor required for the promotion of sepal identity in conjunction with. On these molecular bases, the classical ABC model should be modified by adding SEP as a limiting factor that restricts the activity of the ABC genes to the floral organs (Figure 6). The functional redundancy of the SEP genes has been shown recently by both loss-of-function alleles [49 ] and gain-of-function alleles [42,50 ]. Ectopic expression of PI AP3, as well as of PI AP3 SEP1 SEP2, can convert leaves into petals [50 ]. The three SEP genes form a monophyletic clade within the MADS-box genes of Arabidopsis (Figure 7, SEP clade). The observation that both PI AP3 and PI AP3 can convert leaves into petals when constitutively expressed [42 ] argues that and are also redundant in some of their functions. This is supported by the fact that the SEP genes and belong to the SEP/ superclade (Figure 7). One possible redundant function may be to act as a transactivation component in the complexes. B and C genes do not show any transactivation activity, but genes in the SEP/ superclade shows such activity, although their relative activities vary (Figure 7). Conclusions and perspectives The ABC model has been surprisingly successful in explaining how a small number of regulatory genes, acting alone and in combination, specify the identity of the floral organs. If all the ABC gene activities were compromised, one might expect a ground state of the floral organ to be revealed what Goethe referred to as the ideal basic organ. abc triple mutant flowers comprise not leaves, but organs that have characteristics of both leaves and carpels. It is tempting to speculate that the pathways mediated by SPT and other marginal-tissue promoting genes might represent evolutionarily ancient genetic networks directing the development of sporophylls that predate the evolution Figure 6 1 se A:, CAL CAL PI 2 pe AP3 B: PI, AP3 A revised ABC model. In addition to the original ABC genes, the SEP genes have a crucial role in providing transactivation activity and serving as a scaffold between B and C proteins. Boxes indicate the region of gene expression detected by in situ RNA hybridization. Ovals show proteins, and contact between ovals indicates a protein interaction. For simplicity, the homodimer is not indicated. ca, carpel; pe, petal; se, sepal; st, stamen. PI 3 st AP3 C: 4 ca

454 Pattern formation and developmental mechanisms Figure 7 FRUITFULL /SEP super clade clade PI/AP3 clade clade SEP clade CAL L13 L6 L3 SEP2 SEP1 L11 SHATTERPROOF2 SHATTERPROOF1 PI AP3 Relative LUC/R-LUC activity 150 100 50 0 Control PI PI-VP16 AP3 PI+AP3 PI-VP16+AP3 CAL SEP1 SEP2 Phylogenetic tree of the Arabidopsis MADS-box proteins and transactivation activity of ABC and SEP genes. Protein sequences were aligned by ClustalW (http://www.ddbj.nig.ac.jp/ E-mail/clustalw-e.html), and the results are depicted using TreeView (http://taxonomy.zoology.gla.ac.uk/rod/treeview.html). The ABC MADS-box proteins are located in three distinct clades: the /SEP superclade, the clade and the PI/AP3 superclade. Although not all members have been tested, proteins belonging to the /SEP superclade show transactivation activity. Transactivation assays were performed by particle bombardment into onion epidermal cells using a CArG::Luciferase reporter gene [42 ]. of flowers. Subsequently, the ABC genetic network would have been recruited and integrated with these pre-existing genetic programs, leading to the evolution of the flower in angiosperms. In this scheme, the marginal tissues and nonmarginal tissues of the Arabidopsis carpel would have separate evolutionary origins, and a ground state of the angiosperm floral organ might be a sporophyll. Studies of orthologous genes in other angiosperms and in non-flowering plants are needed to address this issue. The ABC model is widely applicable to the angiosperms, including grasses. Grass flowers are morphologically quite different from those of other angiosperms, especially in the peripheral two whorls. Molecular genetic studies of maize and rice suggest that the palea and lemma may be similar to the sepals of eudicots, and that lodicules are probably homologous to petals of eudicots. Protein complexes are the molecular basis of the genetic ABC model: A + B + specifies petals; B + C + specifies stamens; C+ specifies carpels. Because leaves can be converted into floral organs, these complexes must be sufficient for floral organ identity. It is noteworthy that as well as the ABC genes are homologous MADS-box transcription factors, and that different complexes of homologous proteins result in specific floral organ activity. Biochemical studies are now required to reveal how DNA-binding affinities are modulated by these complexes in vivo, because all MADS-box proteins bind to CArG boxes in vitro. It is also intriguing that some MADS-box proteins belonging to the /SEP superclade have transactivation activity, whereas others belonging to the PI/AP3 or clades do not. When and how they obtained or lost such activity during their evolution might have influenced the evolution of floral organs. Acknowledgements Work on floral organ identity in K Goto s laboratory is supported by the Japan Society for the Promotion of Science (JSPS-RFTF96L00403). J Kyozuka s project is partly supported by CREST, JST. J Bowman is supported by a grant from National Science Foundation. References and recommended reading Papers of particular interest, published within the annual period of review, have been highlighted as: of special interest of outstanding interest 1. Bowman JL, Smyth DR, Meyerowitz EM: Genes directing flower development in Arabidopsis. Plant Cell 1989, 1:37-52. 2. Carpenter R, Coen ES: Floral homeotic mutations produced by transposon-mutagenesis in Antirrhinum majus. Genes Dev 1990, 4:1483-1493.

Turning floral organs into leaves, leaves into floral organs Goto, Kyozuka and Bowman 455 3. Schwarz-Sommer Z, Huijser P, Nacken W, Saedler H, Sommer H: Genetic control of flower development: homeotic genes of Antirrhinum majus. Science 1990, 250:931-936. 4. Bowman JL, Smyth DR, Meyerowitz EM: Genetic interactions among floral homeotic genes of Arabidopsis. Development 1991, 112:1-20. 5. Weigel D, Meyerowitz EM: The ABCs of floral homeotic genes. Cell 1994, 78:203-209. 6. Coen ES: Floral symmetry. EMBO J 1996, 15:6777-6788. 7. Yanofsky MF: Floral meristems to floral organs genes controlling early events in Arabidopsis flower development. Annu Rev Plant Physiol 1995, 46:167-188. 8. Riechmann JL, Meyerowitz EM: MADS domain proteins in plant development. Biol Chem 1997, 378:1079-1101. 9. Komaki MK, Okada K, Nishino E, Shimura Y: Isolation and characterization of novel mutants of Arabidopsis thaliana defective in flower development. Development 1988, 104:195-203. 10. Kunst L, Klenz JE, Martinez-Zapater J, Haughn GW: AP2 gene determines the identity of perianth organs in flowers of Arabidopsis thaliana. Plant Cell 1989, 1:1195-1208. 11. Irish VF, Sussex IM: Function of the apetala1-1 gene during Arabidopsis floral development. Plant Cell 1990, 2:741-751. 12. Mandel MA, Gustafson-Brown C, Savidge B, Yanofsky MF: Molecular characterization of the Arabidopsis floral homeotic gene apetala1. Nature 1992, 360:273-277. 13. Bowman JL, Alvarez J, Weigel D, Meyerowitz EM, Smyth DR: Control of flower development in Arabidopsis thaliana by apetala1 and interacting genes. Development 1993, 119:721-743. 14. Gustafson-Brown C, Savidge B, Yanofsky MF: Regulation of the Arabidopsis floral homeotic gene apetala1. Cell 1994, 76:131-143. 15. Jofuku KD, den Boer BGW, Van Montagu M, Okamuro JK: Control of Arabidopsis flower and seed development by the homeotic gene apetala2. Plant Cell 1994, 6:1211-1225. 16. Hill JP, Lord EM: Floral development in Arabidopsis thaliana: a comparison of the wild type and the homeotic pistillata mutant. Can J Bot 1989, 67:2922-2936. 17. Jack T, Brockman LL, Meyerowitz EM: The homeotic gene apetala3 of Arabidopsis thaliana encodes a MADS-box and is expressed in petals and stamens. Cell 1992, 68:683-697. 18. Goto K, Meyerowitz EM: Function and regulation of the Arabidopsis floral homeotic gene pistillata. Genes Dev 1994, 8:1548-1560. 19. Yanofsky MF, Ma H, Bowman JL, Drews GN, Feldmann KA, Meyerowitz EM: The protein encoded by the Arabidopsis homeotic gene agamous resembles transcription factors. Nature 1990, 346:35-40. 20. Alvarez J, Smyth DR: crabs claw and spatula, two Arabidopsis genes that control carpel development in parallel with agamous. Development 1999, 126:2377-2386. The genetic interactions between crc, spt and the ABC floral homeotic genes are described and it is demonstrated that CRC and SPT control several aspects of carpel development independently of the C class gene AMOUS. Specifically, was shown only to be absolutely required for the differentiation of the ovary wall, whereas SPT can promote the growth of all of the marginal tissues of the carpel. 21. Liu ZC, Franks RG, Klink VP: Regulation of gynoecium marginal tissue formation by leunig and aintegumenta. Plant Cell 2000, 12:1879-1891. The leunig aintegumenta double mutant lacks most marginal tissues in the carpel, implicating both genes in the development of these tissues. 22. Heisler MGB, Atkinson A, Bylstra YH, Walsh R, Smyth DR: Spatula, a gene that controls development of carpel margin tissues in Arabidopsis encodes a bhlh protein. Development 2001, 128:1089-1098. The cloning and expression pattern of SPT is described. The expression pattern of SPT is consistent with its proposed role in promoting the growth of carpel marginal tissues, and several regulators of SPT mrna accumulation are identified. The authors propose that restricting SPT activity to the marginal regions of the carpel is mediated in part by the activity of ETTIN. 23. Sessions RA, Zambryski PC: Arabidopsis gynoecium structure in the wild type and in ettin mutants. Development 1995, 121:1519-1532. 24. Alvarez J, Smyth DR: Genetic pathways controlling carpel development in Arabidopsis thaliana. J Plant Res 1998, 111:295-298. 25. Schmidt RJ, Ambrose BA: The blooming of glass flower development. Curr Opin Plant Biol 1998, 1:60-67. 26. Ma H, depamphilis C: The ABCs of floral evolution. Cell 2000, 101:5-8. 27. McSteen P, Laudencia-Chingcuanco D, Colasanti J: A floret by any other name: control of meristem identity in maize. Trends Plant Sci 2000, 5:61-66. 28. Schmidt RJ, Veit B, Mandel MA, Mena M, Hake S, Yanofsky M: Identification and molecular characterization of Z1, the maize homolog of the Arabidopsis floral homeotic gene Agamous. Plant Cell 1993, 5:729-737. 29. Mena M, Ambrose BA, Meeley RB, Briggs SP, Yanofsky M, Schmidt RJ: Diversification of C-function activity in maize flower development. Science 1996, 274:1537-1540. 30. Theissen G, Strater T, Fischer A, Saedler H: Structual characterization, chromosomal localization and phylogenetic evaluation of two pairs of Agamous-like MADS-box genes from maize. Gene 1995, 156:155-166. 31. Kang HG, Noh YS, Chung YY, Costa M, An K, An G: Phenotypic alterations of petals and sepals by ectopic expression of a rice MADS box gene in tobacco. Plant Mol Biol 1995, 29:1-10. 32. Chung YY, Kim SR, Kang HG, Noh YS, Park MC, Finkel D, An G: Characterization of two rice MADS box genes homologous to GLOBOSA. Plant Science 1995, 109:45-56. 33. Moon YH, Jung JY, Kang HG, An G: Identification of a rice APETALA 3 homologue by yeast two-hybrid screening. Plant Mol Biol 1999. 40:167-177. 34. Kyozuka J, Kobayashi T, Morita M, Shimamoto K: Spatially and temporally regulated expression of rice MADS box genes with similarity to Arabidopsis class A, B and C genes. Plant Cell Physiol 2000, 41:710-718. The expression patterns of class A, B and C genes of rice are determined by in situ hybridization analysis. The significant finding of this paper is that the expression of grass class A gene is observed in lodicules, palea and lemma. 35. Ambrose BA, Lerner DR, Ciceri P, Padilla CM, Yanofsky M, Schmidt RJ: Molecular and genetic analysis of the Silky1 gene reveal conservation in floral organ specification between eudicots and monocots. Mol Cell 2000, 5:569-579. The authors isolate the causal gene of a maize si1 as an ortholog of AP3. si1 is shown to be expressed in primordia of stamen and lodicules. This paper clearly shows that lodicules correspond to dicot petals, and that class B gene function is conserved between dicot and grass species. 36. Kang HG, Jeon JS, Lee S, An G: Identification of class B and class C organ identity genes from rice plants. Plant Mol Biol 1998, 38:1021-1029. 37. Jeon JS, Jang S, Lee S, Nam J, Kim C, Lee SH, Chung YY, KimSR, Lee YH, Cho YG, An G: Leafy hull sterile1 is a homeotic mutation in a rice MADS box gene affecting rice flower development. Plant Cell 2000, 12:871-884. 38. Mandel MA, Bowman JL, Kempin SA, Ma H, Meyerowitz EM, Yanofsky MF: Manipulation of flower structure in transgenic tobacco. Cell 1992, 71:133-143. 39. Mizukami Y, Ma H: Ectopic expression of the floral homeotic gene Agamous in transgenic Arabidopsis plants alters floral organ identity. Cell 1992, 71:119-131. 40. Kuzoff RK, Gasser CS: Recent progress in reconstructing angiosperm phylogeny. Trends Plant Sci 2000, 5:330-336. 41. Krizek BA, Meyerowitz EM: The Arabidopsis homeotic genes apetala3 and pistillata are sufficient to provide the B class organ identity function. Development 1996, 122:11-22. 42. Honma T, Goto K: Complexes of MADS-box proteins are sufficient to convert leaves into floral organs. Nature 2001, 409:525-529. is cloned as an interacting factor of the PI AP3 (B protein) complex. The authors show by protein protein interactions that PI AP3 and

456 Pattern formation and developmental mechanisms PI AP3 are possible candidates for the second and third whorl complexes, respectively. They also prove by transgenic analyses that these protein complexes are sufficient for the formation of petals (second whorl organ) and satamens (third whorl organ). Thus, these protein complexes are the molecular basis of the ABC model. 43. Honma T, Goto K: The Arabidopsis floral homeotic gene pistillata is regulated by discrete cis-elements responsive to induction and maintenance signals. Development 2000, 127:2021-2030. 44. Tilly JJ, Allen DW, Jack T: The CArG boxes in the promoter of the Arabidopsis floral organ identity gene apetala3 mediate diverse regulatory effects. Development 1998, 125:1647-1657. 45. Hill TA, Day CD, Zondlo SC, Thackeray, Irish VF: Discrete spatial and temporal cis-acting elements regulate transcription of the Arabidopsis floral homeotic gene apetala3. Development 1998, 125:1711-1721. 46. Mandel MA, Yanofsky MF: The Arabidopsis L9 MADS box gene is expressed in young flower primordia. Sex Plant Reprod 1998, 11:22-28. 47. Davies B, Egea-Cortines M, de Andrade Silva E, Saedler H, Sommer H: Multiple interactions amongst floral homeotic MADS box proteins. EMBO J 1996, 15:4330-4343. 48. Egea-Cortines M, Saedler H, Sommer H: Ternary complex formation between the MADS-box proteins SQUAMOSA, DEFICIENS and GLOBOSA is involved in the control of floral architecture in Antirrhinum majus. EMBO J 1999, 18:5370-5379. Antirrhinum A (SQUAMOSA) and B (DEFICIENS and GLOBOSA) proteins make complexes in vitro and in yeast. The DNA-binding affinity and sequence specificity of AB complex were altered from that of A-homodimer or B-heterodimer. 49. Pelaz S, Ditta GS, Baumann E, Wisman E, Yanofsky MF: B and C floral organ identity functions require sepallata MADS-box genes. Nature 2000, 405:200-203. Each single mutant of Sep123 shows a subtle phenotype, but the triple mutant of sep123 produces only sepals, thus resembling the phenotype of bc double mutants. As B and C gene expression is not affected by the sep123 triple mutation, BC genes are not regulated by SEP genes but require the activity of SEP genes for floral organ identity. 50. Pelaz S, Tapia-Lopez R, Alvarez-Buylla ER, Yanofsky MF: Conversion of leaves into petals in Arabidopsis. Curr Biol 2001, 11:182-184. Ectopic expression of AP3 PI or AP3 PI SEP2 in Arabidopsis results in the conversion of leaves into petals. 51. Kempin SA, Savidge B, Yanofsky MF: Molecular basis of the cauliflower phenotype in Arabidopsis. Science 1995, 267:522-525. 52. Mandel MA, Yanofsky MF: A gene triggering flower formation in Arabidopsis. Nature 1995, 377:522-524.