|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
1 FLORAL DEVELOPMENT AND GE NETICS IN NYMPHAEALES By MI-JEONG YOO A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2008
2 2008 Mi-Jeong Yoo
3 To my parents
4 ACKNOWLEDGMENTS I thank my advisors, Douglas Soltis and Pamela Soltis, for their guidance and support throughout my PhD program. Their boundless enthus iasm for science and warm encouragement led me to complete this work and further help me continue scientific re search. I also would like to thank my committee members, Bernard Haus er, Edward Braun, and David Oppenheimer, for their help and support. I appreciate the use of Bernard Hausers la b facilities, such as gel-photo systems and microscopes with camera. My gratit ude for former and current Soltis lab members should be expressed here. In particular, I want to thank Jin Koh, Sangtae Kim, Matyas Buzgo, Andr Chanderbali, Chuck Bell, Matt Gitzen danner, Sam Brockington, and Monica Arakaki for their discussion and help with experimental pr ocedures and analytical methods. Also, I would like to thank Jin Koh and Sangtae Kim for field wo rk. Most of this work is funded and aided by the Floral Genome Project (FGP), and so I w ould like to give thanks to the FGP people, including Claude dePamphilis, Jim Leebens-Mack, Naomi Altman, and Lena Landherr. In addition, I express my gratitude to the ICBR staff at the University of Florida: Karen Kelly, who helped me use the SEM, and Gigi Ostrow, who provides the microarray facility. Finally, I highly thank my family, both immediate and extended, for their endless love and support. Particularly, I am thankful to my parents, who have always en couraged me to do what I want. Also, I thank my parents-in-law, who have been taking care of my babies, Yoo-Shin and Yoo-Min. Their devotion and support made it possible for me to do this work. I would like to express my thanks to my two kids, Yoo-Shin and Yoo-Min. You two provide me with invaluable experience and delight. Most importantly, I would like to thank my husband, Jin, for his love, trust, encouragement, and support. He is my best friend as well as an excellent discussion partner.
5 TABLE OF CONTENTS Page ACKNOWLEDGEMENTS ............................................................................................................ 4 LIST OF TABLES ................................................................................................................ .......... 8 LIST OF FIGURES ............................................................................................................... ......... 9 ABSTRACT ...................................................................................................................... ............ 11 CHAPTER 1 INTRODUCTION ................................................................................................................ 13 2 DIVERGENCE TIMES AND HIST ORICAL BIOGEOGRAPHY OF NYMPHAEALES .................................................................................................................. 18 Introduction ................................................................................................................. ........... 18 Materials and Methods ........................................................................................................ .. 20 Taxon Sampling and Topology ...................................................................................... 20 Tests of Rate Heterogeneity ........................................................................................... 21 Estimation of Ages ......................................................................................................... 2 1 Dispersal-Vicariance Analysis ....................................................................................... 24 Results .................................................................................................................................... 24 Tests of Rate Heterogeneity among Genes and Lineages .............................................. 24 Comparison of Estimates from Different DNA Data Sets ............................................. 25 Comparison of Estimates from Different Methods ........................................................ 25 Biogeography ................................................................................................................. 26 Discussion .................................................................................................................... .......... 27 Inferences from Estimates of Divergence Times ........................................................... 27 Reconciling Molecular-based Age Es timates with the Fossil Record ............................ 28 Biogeography ................................................................................................................. 33 Future Directions ............................................................................................................. ...... 35 3 COMPARATIVE STUDIES OF FLORAL DEVELOPMENT IN NYMPHAEALES ........ 43 Introduction ................................................................................................................. ........... 43 Materials and Methods ........................................................................................................ .. 46 Results .................................................................................................................................... 47 Developmental Series ..................................................................................................... 47 Cabomba caroliniana .............................................................................................. 47 Nuphar advena ........................................................................................................ 48 Nymphaea odorata .................................................................................................. 48 Developmental Features of the Third-Whorled organs of Nuphar advena .................... 49
6 Discussion .................................................................................................................... .......... 50 Comparison of Developmental Stages in Three Species ................................................ 50 Outer and Inner Tepals or Sepals and Petals ........................................................... 53 Identity of the Thrid-Whorled Organs of Nuphar advena : Petals or Staminodes .......... 55 4 EXPRESSION OF HOMOLOGUES OF MADS-BOX GENES IN FLOWERS OF TWO DIVERGENT WATER LILIES : THE BASAL ANGIOSPERM NYMPHAEALES AND THE BASAL EUDICOT NELUMBO ........................................... 66 Introduction ................................................................................................................. ........... 66 Materials and Methods ........................................................................................................ .. 72 Plants .............................................................................................................................. 72 RNA Extraction, RT-PCR, and Screening for Homologues of MADS-Box Genes ...... 72 Phylogenetic Analysis for Sequence Identification ........................................................ 73 Gene Expression Based on RQ-RT-PCR ....................................................................... 74 Results .................................................................................................................................... 76 Homologues of MADS-Box Genes in Nymphaeales ..................................................... 76 Analysis of Expression Profile by RQ-RT-PCR ............................................................ 77 Discussion .................................................................................................................... .......... 80 Comparison and Implication of Expre ssion Patterns of MADS-Box Genes in Nymphaeales .................................................................................................................. 80 Perianth Differentiation in Nymphaeales and Transition of Petaloid Staminodes to Stamens in Nymphaea .................................................................................................... 86 Floral Developmental Genetics in Nymphaea and Nelumbo .......................................... 91 Molecular Models for Floral Development in Nymphaeales ......................................... 94 Origin of the Floral Parts in Nymphaeales ..................................................................... 95 5 ANALYSIS OF THE FLORAL TRANSCRIPTOME OF A BASAL ANGIOSPERM, NUPHAR ADVENA (NYMPHAEACEAE) ........................................................................ 116 Introduction ................................................................................................................. ......... 116 Materials and Methods ........................................................................................................ 119 Sample Preparation, Probe Labeling and Microarray Hybridization ........................... 119 Microarray Design ........................................................................................................ 120 Data Acquisition and St atistical Analysis .................................................................... 120 Comparison of Microarray Data with Rela tive Quantitative Reverse Transcriptios PCR (RT-RQ-PCR) data .............................................................................................. 121 Comparative Floral Transcriptomics ............................................................................ 121 Results .................................................................................................................................. 122 Gene Expressed Differentially in Reproductive Organs .............................................. 122 Comparison of Microarray Data with RQ-RT-PCR Data ............................................ 124 Hierarchical Clustering ................................................................................................. 124 Comparative Floral Transcriptome ............................................................................... 125 Hierarchical clustering .......................................................................................... 125 Investigation of spatial ge ne expression patterns .................................................. 126
7 Discussion .................................................................................................................... ........ 127 Gene Expressed Differentially in Reproductive Organs .............................................. 127 Hierarchical Clustering ................................................................................................. 129 Comparative Floral Transcriptome ............................................................................... 135 Hierarchical clustering .......................................................................................... 135 Investigation of spatial ge ne expression patterns .................................................. 136 6 CONCLUSIONS ................................................................................................................. 159 LIST OF REFERENCES ............................................................................................................ 164 BIOGRAPHICAL SKETCH ...................................................................................................... 186
8 LIST OF TABLES Table page 2-1 List of taxa used in this study with GenBank accession numbers and references ............ 37 2-2 Divergence time estimation: Age estimates for nodes 2-9 for different age estimation methods ....................................................................................................................... ...... 38 2-3 Effect of fossil placement ................................................................................................ 40 3-1 Stage alignment using descriptions of flor al developmental stages for three taxa ........... 58 4-1 Primer information used in this study ............................................................................... 98 4-2 Comparison of floral gene expression patterns from diffe rent developmental stages of Nymphaea odorata ..................................................................................................... 100 4-3 Summary of expression patterns of floral genes in Nymphaeales and Nelumbo just prior to anthesis ............................................................................................................. .. 101 5-1 Number of genes up-regulated (t wo-fold higher expression level) in Nuphar floral tissue relative to leaves ................................................................................................... 140 5-2 Genes down-regulated in reproductive organs of Nuphar .............................................. 141 5-3 Genes up-regulated in reproductive organs of Nuphar ................................................... 144
9 LIST OF FIGURES Figure page 2-1 Phylogenetic tree of Nymphaeal es from Les et al. (1999), with Amborella Austrobaileyales, and gymnosperms outgroups added, based on many recent analyses ...................................................................................................................... ....... 41 2-2 Chronogram for Nymphaeales showing timing of inferred divergences and biogeographic events ........................................................................................................ 42 3-1 Developmental series of Cabomba caroliniana ............................................................... 60 3-2 Developmental series of Nuphar advena .......................................................................... 61 3-3 Developmental series of Nymphaea odorata .................................................................... 63 3-4 Developmental features of the third-whorled organs of Nuphar advena ......................... 64 4-1 Photographs of flowers of Nymphaeales and Nelumbo .................................................. 103 4-2 Alternatively spliced transcript of Nyod.AP3 ................................................................. 104 4-3 Phylogenetic analyses using 186 MADS-box gene homologues ................................... 105 4-4 Relative quantitative RT -PCR results of floral MADS-box gene homologues in Nymphaeales and Nelumbo............................................................................................. 109 4-5 Developmental features of Nuphar and Cabomba .......................................................... 115 5-1 Summary tree for angiosperms ....................................................................................... 146 5-2 Flower of Nuphar advena ............................................................................................... 147 5-3 Double-loop design for Nuphar microarray experiments ............................................... 148 5-4 Comparison of microarray da ta with RQ-RT-PCR data ................................................. 149 5-5 Hierarchical clustering di splays up-regulated or dow n-regulated gene clusters based on similarity of expression patterns ...................................................................... 151 5-6 Number of genes up-regulated in Nuphar floral tissue at iden tified four regulatory modules ....................................................................................................................... .... 152 5-7 Hierarchical clustering results of combined data sets of Arabidopsis Nuphar and Persea data set ................................................................................................................ 153 5-8 Scatter plots of spatial gene expression patterns in floral organs of Nuphar, Persea, and Arabidopsis .............................................................................................................. 154
10 5-9 Expanse of gene expression domain across floral organs of Nuphar Persea and Arabidopsis .................................................................................................................... 156 5-10 Number of genes with at least two-fold up-regulated in one organ and less than 1.5fold in all other floral organs of Nuphar Persea and Arabidopsis ............................... 158
11 Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy FLORAL DEVELOPMENT AND GE NETICS IN NYMPHAEALES By Mi-Jeong Yoo December 2008 Chair: Douglas E. Soltis Major: Botany Nymphaeales (water lilies) one of the basal lineages of angiosperms, and this clade is sister to all extant flowering plants except Amborella. Nymphaeales also exhibi t floral morphology that differs from other basal angiosperms, for exam ple, in having whorled phyllotaxis. Thus, studying floral evolution in Nymphaeales will help elucidate floral development in early-diverging angiosperms and contribute to understanding floral evolution and diversif ication of angiosperms as a whole. The goals of this study were to esti mate the divergence times of the major clades of Nymphaeales, to obtain floral developmen tal sequences for three study species ( Cabomba caroliniana Nuphar advena, and Nymphaea odorata ), to examine the expression patterns of putative floral organ identity genes in floral orga ns of these three species (and especially in the transition series from perianth to stamens in Nymphaea ), and to compare expression profiles of the floral transcriptome across floral organs of Nuphar using microarrays. Using a phylogeny based on the multiple genes, I found that extant Nymphaeales diversified into two major clades during the Eocene (44.6.9 mya), indicating that extant Nymphaeales diversified relatively recently, in contrast to their ancient fossil r ecord. The three study species follow the same 10 developmental stages in the very same order. Nuphar which is sister to other members of Nymphaeaceae, exhibits Nymphaea-like or Cabomba-like developmental features, and its
12 phylogenetic position suggests that sharing features between Nuphar and Cabomba might be ancestral characteristics to core Nymphaeales. E xpression patterns of floral genes in these three species of Nymphaeales partially follow the developmental model of well-studied eudicot systems, such as Arabidopsis However, analyses of th e floral transcriptome in Nuphar and comparisons with other plants show that the gene s involved downstream of floral organ identity genes are distinctive in their ex pression patterns in this species. Furthermore, there is greater overlap in floral transcriptional programs between floral parts in Nuphar than in the eudicot Arabidopsis paralleling the less specialized floral orga nization of basal angiosperms. The results for Nuphar coupled with similar data for the basal angiosperms, Persea indicate that the first flowers deployed developmental programs more broadly than the vast majority of modern species.
13 CHAPTER 1 INTRODUCTION Nymphaeales comprise nine genera and approxi mately 80 species of aquatic plants, with a worldwide distribution in tropical to temperate regions (Saarela et al. 2007; Schneider and Williamson 1993; Williamson and Schneider 1993) Nymphaeales are clearly monophyletic, appearing as a well-supported clad e in a series of molecular phylogenetic analyses (Borsch et al. 2007; Chase et al. 1993; Hilu et al. 2003; Lhne et al. 2007; Soltis et al. 2000). Phylogenetic relationships among genera of Ny mphaeales were well resolved and strongly supported in phylogenetic analyses of molecular data (Borsch et al. 2007; Lhne et al. 2007; Les et al. 1999). This clade is composed of th ree subclades: Cabombaceae, Nymphaeaceae, and Hydatellaceae. The former two subclades are traditionally recognized as two families, Cabombaceae and Nymphaeaceae (Richard 1828; Williamson and Schneider 1993), which share many features, including vegeta tive, reproductive, and molecular characters. The remaining subclade, Hydatellaceae, is sister to Cabo mbaceae + Nymphaeaceae and was recently placed in Nymphaeales based on molecular phylogenetics (Saarela et al. 2007). Furthermore, its placement in Nymphaeales is confirmed by the presence of a 4-celled embryo sac like that of other Nymphaeales (Friedman 2008). Howe ver, floral development diffe rs among the three families. For example, Cabomba and Brasenia (Cabombaceae) are characte rized by oligomerous flowers and simultaneous initiation of calyx and corolla, whereas Nymphaeaceae ( Barclaya Euryale, Nuphar Nymphaea, Ondinea and Victoria ) have polymerous flowers and exhibit a unidirectional order of perianth initiation (Endress 2001; Ito 1987; Les et al 1999; Schneider et al. 2003). Hydatellaceae (Hydatella and Trithuria ) show unique floral development: the presence of involucral bracts, no perianth, and separate staminate a nd carpellate flowers (Rudall et al.
14 2007). In this study, I only focused on three genera, Nuphar the phylogenetically derived Nymphaea (Nymphaeaceae), and Cabomba (Cabombaceae). Cabomba and Brasenia (Cabombaceae) (Cronquist 1981; Takhtajan 1980; Takhtajan 1997) are united by the presence of free-floating stem s, in addition to rhizomes. They also have distinct carpels, nutlike fruits, and six-tepaled flowers without petal-like staminodes (Judd et al. 2002). There are numerous non-DNA synapomorphi es for Nymphaeaceae, including the presence of star-shaped sclerids spirally inserted stamens with laminar filaments, and laminar placentation (Les et al. 1999; Moseley et al. 1993). Nuphar is sister to a ll other genera of Nymphaeaceae (Borsch et al. 2007, 2008; Les et al 1999; Lhne et al. 2007), and it has a superior ovary and monosulc ate pollen. In contrast to Nuphar Barclaya, Ondinea, Nymphaea, Euryale, and Victoria have perigynous/epigynous flowers and zonasulculate pollen (Ito 1987; Les et al. 1999). Barclaya is then sister to Euryale Victoria Nymphaea, and Ondinea; relationships among the latter ar e not clear. In recent phyl ogenetic studies, the monotypic Ondinea was embedded in the Nymphaea clade (Borsch et al. 2007, 2008; Lhne et al. 2007). A clade of Euryale and Victoria is sister to all species of Nymphaea (Borsch et al. 2007), or is included within the Nymphaea clade (Borsch et al. 2008; Lhne et al. 2007). In Nymphaea, Euryale, and Victoria the gynoecial vascular st rand originated from the receptacular plexus and is the source of the pe tal trace (Ito 1987; Les et al. 1999). The clade of Victoria and Euryale is united morphologically by prickles on the pe tiole and on the abax ial leaf surface. The flowers of Nymphaeales range from small, ca. 0.5-1.5 cm ( Cabomba) to very large, up to 50 cm ( Victoria ) (Schneider and Williamson 1993). Organ nu mber varies widely, with three or four sepals, zero to four ( Ondinea ) up to six to 51 ( Nymphaea) and 50-70 (Victoria ) petals, six ( Cabomba) up to 15-750 ( Nymphaea) stamens, and three to 14 ( Ondinea) up to five to 47
15 ( Nymphaea) carpels. The highest number of floral parts occurs in Victoria and Nymphaea, but Nuphar the sister to all remaining Nymphaeaceae, has also numerous parts. Thus, the proliferation of floral parts in Nymphaeales might be the result of several instances of secondary increase (Borsch et al 2008; Les et al. 1999). Several labs are trying to understand the gene tic architecture of floral development, including the origin and s ubsequent diversification of the flower (D. Soltis et al. 2002). The basal angiosperm lineages are of particul ar interest in studies of floral evolution due to their diversity in the arrangement and number of floral parts. Importantly, members of Nymphaeales exhibit floral morphology that differs from the other basalmost angiosperms (i.e., Amborella Austrobaileyales) in several features, including the presence of large flowers in some members (e.g., Nuphar Nymphaea, Euryale, and Victoria ) rather than small or moderate-sized flowers, whorled rather than spiral phyllotaxis, eudicot-like perianth differentiation (as in Cabomba and Nuphar ), and the occurrence of blue perianth organs containing anthocyani ns, which are absent in other basal angiosperms (Endress 2001; Schneider et al. 2003). Nymphaea les are sister to all extant angiosperms except Amborella (Borsch et al. 2005; Hilu et al. 2003; Jansen et al. 2007; Lhne and Borsch 2005; Leebens-Mack et al. 2005; Mathews and Donoghue 1999; Moore et al. 2007; Qiu et al. 1999; Qiu et al. 2005; Soltis et al. 1999; Soltis et al. 2000; Soltis et al. 2005; Zanis et al. 2002; Zanis et al. 2003). Thus, studying floral evolu tion in Nymphaeales will help elucidate floral evolution in early-diverging angios perms, and thus provide insight into the early diversification of the flower. Therefore, the main goals of this study were to: (1) estimate the divergence times of the major genera of Nymphaeales using a tree based on multiple genes, (2) obtain floral developmental sequences of Cabomba caroliniana Nuphar advena, and Nymphaea odorata (3)
16 examine the expression patterns of floral organ identity genes in the transitional series of stamens of Nymphaea, which has a complete series from petalo id staminodes to functional stamens, (4) examine the expression patterns of floral organ identity gene homologues in three genera ( Cabomba, Nuphar and Nymphaea), and (5) compare expression profiles of Nuphar floralderived genes in different floral and vegetative ti ssues using microarrays. These goals are addressed in the following fi ve chapters. In chapter 2, I estimated the divergence times of the major genera of Nymph aeales using a tree based on multiple genes and several methods (NPRS (Sanderson 1997), penali zed likelihood (Sanders on 2002), and Bayesian approaches (Thorne and Kishino 2002)). The es timated divergence times may help clarify the timing of evolutionary events suggest ed in my molecular studies below. In chapter 3, I present floral developmental sequences for Nuphar advena Nymphaea odorata and Cabomba caroliniana Although developmental studies of Nymphaeales have been reported previously (Endress 2001; Mose ley 1958, 1961, 1965, 1972; Moseley et al. 1993; Schneider et al. 2003), the complete developmental series from initiation of the floral meristem to anthesis provides the framework for understand ing the developmental characteristics across a diverse array of Nymphaeales. In chapter 4, I investigated the expression patterns of floral organ identity gene homologues in floral organs of Cabomba, Nuphar, and Nymphaea, and so called water lily Nelumbo, which is previously considered to be related to Nymphaeales, now recognized as eudicots, and compared their gene expression profiles. In particular, I focused on gene expression patterns associated with perianth differentiation among these three genera, and the transitional series of petaloid st aminodes to functional stamens of Nymphaea Also, I
17 investigated whether morphol ogical similarities between Nymphaea and Nelumbo are related to similar floral developmental genetics. In chapter 5, I compared of expression profiles of Nuphar floral-derived genes in different floral and vegetative tissues using microarrays. I used a cluster analysis and comparative approach to elucidate the floral transcriptome of Nuphar and compare it using data from other plants, such as Persea and Arabidopsis Chapter 6 is a general conclu sion in which I summarize what has been done and what I learned about floral developmental genetics and evolution in Nymphaeales. Furthermore, I suggest future directions that will help clarify flower developm ent in Nymphaeales as well as basal angiosperm lineages.
18 CHAPTER 2 DIVERGENCE TIMES AND HISTORICAL BIOGEOGRAPHY OF NYMPHAEALES Introduction 0*Nymphaeales comprise eight genera and approxi mately 70 species of aquatic plants, with a worldwide distribution in tropical to temperate regions (Schneider and Williamson 1993; Williamson and Schneider 1993). Nymphaeales ar e clearly monophyletic, appearing as a wellsupported clade in a series of molecular phylogene tic analyses (Chase et al. 1993; Hilu et al. 2003; Savolainen et al. 2000; Soltis et al. 2000; Zanis et al. 2002). Phylogenetic relationships among extant genera of Nymphaeales were well resolved and strongly supported in analyses of molecular and morphological data (Les et al. 1999). The order comprises two clades, traditionally recognized as two families (Richard 1828; Williamson and Schneider 1993), Cabombaceae and Nymphaeaceae, but sometimes recognized as a single family (Nymphaeaceae), generally with two subfamilies (Cabomboideae and Nymphaeoideae) (e.g., (APG II 2003; Caspary 1888; Henkel et al. 1907). We will recognize these two clades at the familial rank for consistency with most previous treatments. Cabomba and Brasenia (Cabombaceae) (Cronquist 1981; Takhtajan 1980; Takhtajan 1997) are united by the presence of free-floating stem s, in addition to rhizomes. They also have distinct carpels, nutlike fruits, and six-tepaled flowers without petal-like staminodes (Judd et al. 2002). Numerous morphological synapomorphies unite Nymphaeaceae, including the presence of star-shaped sclereids, spirally inserted stamens with laminar filaments, and laminar placentation (Les et al. 1999; Moseley et al. 1993). Within Nymphaeaceae, relationships can be summarized as follows: ( Nuphar ( Barclaya ( Ondinea ( Nymphaea ( Euryale, Victoria ))))) (Fig. 21). Species of Nuphar the second largest genus, share several unique morphological characters *Reprinted with permission from the American Society of Pl ant Taxonomists. Original publication: Yoo, M.-J., C. D. Bell, P. S. Soltis, and D. E. So ltis. 2005. Divergence Times and Historical Biogeography of Nymphaeales
19 that provide support for the monophyly of the genu s (Padgett et al. 1999), including pollen with ektexine spines (Takahashi 1992). Nuphar has hypogynous flowers and monosulcate pollen, while Barclaya, Ondinea, Nymphaea, Euryale and Victoria have perigynous/epigynous flowers and zonasulculate pollen (It o 1987; Les et al. 1999). Nymphaea is sister to the Victoria-Euryale clade, a relationship supported by anatomical characters (Ito 1987; Les et al. 1999); the monophyly of Nymphaea is weakly supported by sequence data (Borsch 2000). In these three genera, the gynoecial vascular stra nd originates from the receptacular plexus and is the source of the petal trace (Ito 1987; Les et al. 1999). Victoria and Euryale are united morphologically by prickles on the petiole and on the abaxial leaf surface. The robust phylogeny provided by Les et al. (1999) is also useful for estimating divergence times of major clades and genera within Nymph aeales. In this study we es timated the divergence times of genera of Nymphaeales using DNA sequ ence data from nuclear 18S rDNA and plastid rbcL and matK and the maximum parsimony tree of Les et al. (1999). We applied four methodsa strict molecular clock (Langley and Fitch 1974), nonparametric rate smoothing (Sanderson 1997, 1998), penalized likelihood (S anderson 2002), and a Bayesian method (Kishino et al. 2001; Thorne et al. 1998; Thorne and Kishin o 2002)to estimate divergence times. The use of molecular data to date divergen ces is laden with potenti al problems, and these, along with potential inadequacies of the fossil record, have be en reviewed (Magalln 2004; Sanderson and Doyle 2001; Sanderson et al. 2004; Soltis et al. 2002). Despite potential drawbacks, these approaches have provided esti mates that have converged and are in general agreement with the fossil record, especially regard ing the origin of the a ngiosperms (Bell et al. 2005; Magalln 2004; Sanderson et al. 2004). In add ition, as the estimated divergence times are
20 important for interpretation of historical biog eography, we inferred the historical biogeographic pattern of major clades in Nymphaeales in light of these divergence times. Materials and Methods Taxon Sampling and Topology Our molecular sampling employed the sequence da ta from Les et al. (1999) and includes eight species of Nymphaeales, representing the eight genera of the order. To estimate the divergence time of Nymphaeacea e and Cabombaceae, sequences of other basal angiosperm lineages, Amborella and three representatives of Austrobaileyales ( Austrobaileya Illicium and Schisandra ), were included. Four sequences of extant gymnosperms ( Larix, Taxus Gnetum and Ginkgo) were included to permit estimation of the age of the angiosperms. For each species, nuclear 18S DNA sequences and plastid rbcL and matK sequences were retrieved from GenBank (Table 2-1). The sequences were aligned us ing CLUSTAL X (Thompson et al. 1997) with the default options, and the alignment was then adjust ed manually. Data matrices used in this study are deposited in TreeBase (www.Treebase.org). The topology for Nymphaeales used in our study is the single maximum parsimony tree obtained by Les et al. (1999) inferred from an alysis of morphological characters, plus rbcL, matK and 18S rDNA sequences (Fig. 2-1). In gene ral, the individual and combined data sets (for all genes but 18S rDNA) gave very simila r topologies (reviewed in Les et al. 1999). The matK topology (see Hilu et al. 2003 ) is also highly similar to the Les et al. total evidence topology. Phylogenetic relati onships among Nymphaeales, Amborella trichopoda and Austrobaileyales were constrained to match those reported in a seri es of recent studies (Borsch et al. 2003; Graham and Olmstead 2000; Hilu et al. 2003; Mathews and Donoghue 1999; Parkinson
21 et al. 1999; Qiu et al. 1999; Soltis et al. 1999; Soltis et al. 200 0; Zanis et al. 2002) (Fig. 2-1). This topology was constrained durin g estimation of divergence times. Tests of Rate Heterogeneity For each gene taken separately and all genes combined, rate heterogeneity across lineages was tested using a likelihood rati o (LR) test (Felsenstein 1988) Significance was determined by comparing the difference between -ln likelihood (-lnL ) of the tree, with and without enforcing a molecular clock, based on a 2 distribution with n 2 degrees of freedom where n is the number of taxa. In addition, rate hetero geneity between pairs of genes wa s also tested using a LR test. Significance was assessed by comparing = -2logLR with a 2 distribution, in which LR=[ln L (ln L1 + ln L2)], where L1 is the likelihood of the tree with one data set, L2 is the likelihood of the tree with the second data set, and L is the likelihood of the tree with both data sets combined (Sanderson and Doyle 2001). Degrees of freedom were calculated according to Sanderson and Doyle (2001). Estimation of Ages Because all tests of rate heterogeneity among lineages were highly significant, we used three approaches that have been proposed for use with heterogeneous rates, NPRS (Sanderson 1997, 1998), PL (Sanderson 2002), and a Bayesian method (Kishino and Hasegawa 1989; Thorne et al. 1998; Thorne and Kishino 2002). A lthough a molecular clock was rejected for our data (which is typically the case in such analyses; reviewed in Sanderson et al. 2004), we calculated divergence times under a strict molecular clock, using the LF (Langley and Fitch 1974) method as implemented in the comput er program r8s v. 1.6 (Sanderson 2003), for comparison with methods that attempt to accomm odate rate heterogeneity. The LF method uses maximum likelihood to estimate divergence times under the assumption of rate constancy among
22 lineages (i.e., a molecular clock). All LF, NPRS and PL analyses were performed on a fixed topology (Fig. 2-1) with branch lengths (for si ngle genes and the combined gene data set) optimized under both maximum parsimony (MP) and maximum likelihood (ML) criteria using a GTR + model of nucleotide substitution in PAUP* 4.0 (Swofford 2002). We used a hierarchical likelihood ratio test (hLRT) to choose the best-fitting model for our data (Felsenstein 1981; Goldman 1993). Using NPRS, the trees with branch lengths we re transformed into ultrametric trees as implemented in 0r8s (Sanderson 2003). To transform relative time to absolute ages, we calibrated the trees by using the minimum age of the a ngiosperm crown group constrained to 131.8 mya which is based on the oldest unambiguous fossils (s ee Soltis et al. 2002b and references therein). To compute confidence intervals for the divergence times estimated, we used the bootstrap resampling method (Baldwin and Sanderson 1996). PL is a semi-parametric smoothing method that allows a range of modes for rate differences among lineages, from nearly clock-like behavior to a condition in which each branch is allowed its optimal substitution rate (Sande rson 2002). The PL method attempts to combine the statistical power of parametric methods w ith the robustness of non-parametric methods, by assigning penalties that limit rate changes between adjacent branches on a phylogeny. All PL analyses were conducted using th e computer program r8s. The program r8s implements a datadriven, cross-validation procedure that systema tically prunes terminals from the tree and then estimates parameters from the sub-matrix and a given smoothing value. It then predicts the data for pruned taxa using the estimated paramete rs. Finally, it calculate s a chi-squared error associated with the difference between the pr edicted and observed data. The optimal smoothing level is chosen as the one that minimi zes the chi-squared e rror (Sanderson 2002).
23 Bayesian methods (Kishino and Hasegawa 1989; Thorne et al. 1998; Thorne and Kishino 2002) that relax a strict molecular clock were also used to estimate divergence times using MULTIDIVTIME (available from J. Thorne, North Carolina State University). This parametric approach relaxes the assumption of a strict molecular clock with a continuous autocorrelation of substitution rates across the phylogeny and allows the use of several calibrations/time constraints. Divergence date estimation with MULTID IVTIME involved two steps. First, ESTBRANCHES was run to estimate branch lengths from the data and a fixed tree topology (Fig. 2-1) using the F84 (Felsenstein 1984; Kishino and Hasegawa 1989) model of sequence evolution, with rates allowed to vary among sites following a discrete gamma ( ) distribution with four rate categories (Yang 1994) along with their variance-covariance matrix. The F84 substitution model is the most complicated model currently available in this program, and it is therefore the most appropriate for comparison with the GTR model used with the ot her three methods. Parameter values for the F84 + model were estimated using the BA SEML program in PAML (Yang 2000). Next, the outgroups (in our case, the extant gymnosperm species) were pruned from the tree, and MULTIDIVTIME was used to estimate the prior and posterior ages of branching events, their standard deviations, and the 95% credib ility intervals via Ma rkov chain Monte Carlo (MCMC). The Markov chain was run for 1,000,000 generations and sampled every 100 generations after an initial bur n-in period of 10,000 cycles. To check for convergence of the MCMC, analyses were run from at l east two different starting points. Fossils assigned to Nymphaeales were repo rted recently by (Friis et al. 2001) and (Gandolfo et al. 2004). We attempted to reconc ile our estimated divergence times with the reported ages and putative phylogene tic positions of these fossils. For this part of our study, we used PL to estimate the age of the angiosperms using each fossil as a single calibration point.
24 These estimated ages were then compared with the fossil record and with other recent molecularbased estimates to assess the correspondence betw een the reported placements of these fossils and molecular-based dates. Only PL was used for these analyses because it is favored over NPRS in simulations, and it provides similar estima tes to those of the Bayesian analysis. Dispersal-Vicariance Analysis We analyzed the distributions of major clades and genera in Nymphaeales using DIVA (Ronquist 1996; Ronquist 1997). The program optimi zes distributions for each node of the tree by favoring vicariance events and minimizing the nu mber of assumed dispersals and extinctions (Ronquist 1997). DIVA assigns a cost to changes in distribution interpreted as extinctions or dispersals but no cost to changes interpreted as vicariance. Therefore, optimizing vicariance minimizes the cost, and those reconstructions are favored. The distributions retrieved by DIVA were compared with the estimated divergence times. Seven main ar eas of distribution were used: (1) North America, (2) Central America, (3) South America, (4) Europe, (5) Africa, (6) Asia, and (7) Australia. The dates estimated by the Baye sian method were used to construct a chronogram with biogeographic events i ndicated; although the estimated dates vary somewhat among methods, estimates obtained via both NPRS and PL fall within the 95% cred ibility intervals of the Bayesian estimates, indicating that al l estimates are essentially equivalent. Results Tests of Rate Heterogeneity among Genes and Lineages All genes, separate and combined, show significant rate heterogeneity among lineages based on likelihood ratio tests. In addition, all pair s of genes evolve at significantly different rates across this tree. Therefore, a molecular clock is rejected.
25 Comparison of Estimates from Different DNA Data Sets The estimated divergence times varied c onsiderably among genes (Table 2-2). For example, with NPRS, the estimates for node 3 (Nymphaeales) ranged from 43.9 mya (combined data set) to 61.1 mya (18S rDNA) (Table 2-2). Plastid genes generally yielded the oldest age estimates for node 2, 18S rDNA typically provided the oldest ages for all other nodes (Table 2-2). Divergence time estimates based on 18S rDNA seque nces have also been older than those obtained with other genes (e.g., rbcL, atpB ) in other studies (e.g., Soltis et al. 2002b). Comparison of Estimates from Different Methods Although the molecular clock was rejected for our data set, we calculated divergence times under a strict molecular clock, using the Langl ey-Fitch method, for comparison with the other three methods. The estimated divergence times us ing Langley-Fitch were generally younger than those inferred from NPRS, PL, and the Bayesian method (Table 2-2). We used NPRS to accommodate the significant rate heterogeneity among lineages, using both MP and ML branch lengths. The ML estima tes are older than the MP estimates inferred from matK and the combined data set (Table 2-2). In general, however, the divergence times obtained with the two branch-length optimizations were similar, mostly dating to the Eocene and Miocene, so the effects of the optimization method seem to be small. Although none of the standard de viations are large, for the more recent divergences (e.g., nodes 8 and 9; Nymphaea, Euryale+ Victoria ) the confidence intervals inferred using MP branch lengths were considerably narrower than those co mputed with ML branch lengths (Table 2-2). Furthermore, in many cases the confidence interv als using MP branch lengths did not cover the estimate from the original data (Table 2-2). Cons idering that the estimates based on the original data using both MP and ML branch lengths were very similar across gene s and that the bootstrap
26 confidence intervals using ML branch lengths co ntained the original es timates, the bootstrap analyses using MP branch lengths seem to be systematically underestimating node ages and associated variances. To our knowledge, this prob lem has not been reported before; in fact, previous estimates inferred from bootstrapping and the original data differ by only 2-4 million years across a time span of 50 to over 400 mya (So ltis et al. 2002). However, it may be best not to rely on standard deviations calcula ted with MP branch lengths and NPRS. The estimated ages obtained from PL are usua lly slightly younger than those obtained from NPRS (Table 2-2). Sanderson (2002) has shown th at PL outperforms NPRS in simulations. In our analyses, use of the Bayesian method provid es estimated ages that are similar to those derived from PL, except for the age of crow n-group Cabombaceae (node 4) as estimated using 18S rDNA data (34.6-32.6 mya (PL) vs. 17.2 mya (Bayesian); Table 2-2). The methods give similar values despite the use of different nucleotide substituti on models for the NPRS and PL analyses vs. the Bayesian analysis. Despite some differences among the three me thods that attempt to account for rate heterogeneity among lineages, the estimated diverg ence times are very similar for most of the nodes. For example, the 95% credibility interv al of the Bayesian method includes the ages estimated from both NPRS and PL. Further disc ussion is based on ages obtained with the Bayesian method. We also think it is reasonable that more data will give better branch length estimates, so we favor the estimated dates calculated from the combined data set. Biogeography The results of the DIVA analysis are shown in Figure 2-2. The ances tors of Nymphaeales were found to be distributed in the American and Eurasian continents, and the present distributional patterns require several dispersal and extin ction events (Fig. 2-2).
27 Discussion Inferences from Estimates of Divergence Times The estimated divergence times indicate that crown-group Nymphaeales (node 3, Figs. 21, 2-2) date to the Eocene (44.6.9 mya); Nymp haeaceae and Cabombaceae split at that point. Extant genera of Nymphaeaceae began to divers ify in the late Eocene to early Oligocene (41.1.7 mya; node 5, Figs. 2-1, 2-2), and the two extant genera of Cabombaceae diverged during the Miocene (19.9.6 my a; node 4, Figs. 2-1, 2-2). These dates for crown-group Nymphaeales are slightly younger than fossil s eeds described from the Middle Eocene Allenby Formation of Princeton, British Columbia, and attributed to Nymphaeaceae (Cevallos-Ferriz and Stockey 1989). However, on the basis of th e fossil seed record, Cevallos-Ferriz and Stockey (1989) inferred that Cabombaceae a nd Nymphaeaceae probably diverged as early as the Middle Eocene, consistent with our older estimates for the age of this divergence. Our divergence time estimates indicate that extant Nymphaeales diversified relatively recently, whereas the stem lineage to Nymphaeales is old, based on a fossil attributed to Nymphaeales from the Early Cretaceous (125-115 my a; Friis et al. 2001) and a fossil attributed to Nymphaeaceae from the middle Cretaceous (~90 mya; Gandolfo et al. 2004). These results for Nymphaeales indicating recent dive rsification in an ancient lineag e agree with similar findings for the basal angiosperms Chloranthaceae (Zhang and Renner 2003) and Illicium (Illiciaceae; (Morris et al. 2007). The fossil record indicates clearly that Chloranthaceae represent one of the oldest angiosperm lineages, with unequivocal reproductive structures resembling those of Hedyosmum from the Barremian-Aptian boundary, approximately 125 mya (Doyle et al. 2003; Eklund et al. 2004; Friis et al. 1994; Friis 1997; Friis et al. 1999) for recent interpretations and lists of earlier references). However, divergence time estimates based on molecular data indicate
28 that the extant genera of Chloranth aceae are relatively young (i.e., 60-29 mya for Hedyosmum 22-11 mya for Chloranthus and 18-9 mya for Ascarina ; Zhang and Renner 2003). Reconciling Molecular-based Age Es timates with the Fossil Record Two putative, ancient water lily fossils are extremely important in discussions of the diversification of Nymphaeales : the water lily fossil reported by Friis et al. (2001) and Microvictoria described by Gandolfo et al. (2004). To reconcile mol ecular-based estimates of divergence times with the fossil dates, we placed the Friis et al. (2001) and the Gandolfo et al. (2004) fossils on several different nodes of the phylogenetic tree for extant Nymphaeales (Fig. 21) as calibration points, and then used PL to estimate the age of the angiosperms. Friis et al. (2001) reported a water lily flower fossil that they estimated to be 125 to 115 million years old, indicating the antiquity of the Nymphaeales lin eage. Although the arrangement of the carpels around a central protrusion of the floral ap ex is found in both some Nymphaeales and Illicium and despite abundant fossil seeds from the same local ity attributed to Illiciales, Friis et al. (2001) argued that the fossil belongs to Ny mphaeales because of its perigynous/epigynous perianth, syncarpous gynoecium, a nd the presence of numerous ovul es per carpel. In contrast, Illicium flowers are hypogynous with an apocarpous gynoecium and a single ovule per carpel. A phylogenetic analysis, based on th e morphological data set of Le s et al. (1999) for extant Nymphaeales, placed the fossil as sister to Nymphaeaceae; synapomorphies with extant Nymphaeaceae included a syncarpous gynoecium, a perigynous or epigynous perianth, and a central protrusion of the floral apex between the carpels (although neith er of the latter two characters is found consistently throughout Nymp haeaceae) (Friis et al. 2001). Based on this analysis, plesiomorphies for Nymphaeales (shared by the fossil and Cabombaceae) seem to include a trimerous perianth, monocolpate pollen, fl oral parts in apparent whorls, linear stamen
29 filaments, and separate radiating stigmatic areas ; some of these characters (monocolpate pollen, trimerous perianth, discontinuous stigmatic areas) are also shared by at least some species of Nuphar (Padgett et al. 1999; Schneider et al. 2003). Gandolfo et al. (2004) proposed that the Friis et al. water lily fossil is equally compatible with Illicium as well as other angiosperm families. The homoplasy in character of hypogyny vs. perigyny/epigyny and the large number of symplesiomorphies between the fossil and Cabomb aceae suggest that a reanalysis that includes taxa outside Nymphaeales may be informative regarding the placement of the fossil and the evolution of these characters. The placement of the water lily fossil as sister to all extant Nymphaeaceae (Friis et al. 2001) makes its point of attachment to the tree older than node 5, but younger than node 3 (Fig. 2-1). Therefore, the appropria te node for calibration using this fossil is node 3. Using this calibration point, the estimated age of the angios perms using PL is 414 mya (Table 2-3), an age that would place the origin of flowering plants sh ortly after the first appearance of land plants, approximately 450 mya based on the fossil recor d, and close to the in itial radiation of tracheophytes (Kenrick and Cran e 1997). However, given the hom oplasy in the character that place the fossil and given its similarity in some respects to Illicium (Friis et al. 2001; Gandolfo et al. 2004), it is possible that the fossil should be placed on the stem lineage leading to Nymphaeales, between nodes 2 and 3. Using the Friis et al. fossil as a calibration point at node 2, which comprises all angiosperms except Amborella results in a reasonable estimate for the age of the angiosperms (138.8 mya; Table 2-3). Henc e, moving the fossil calibration point just one node deeper in the tree reconciles the ages inferred from the fossil record and from molecular data. This placement, on the stem lineage to Nymphaeales, suggests that the fossil may have been part of an ancient assemblage that included Nymphaeales and Austrobaileyales.
30 A fossil flower from the Raritan Formation of New Jersey was recently placed within Nymphaeaceae (Gandolfo et al. 2004). Based on a phylogenetic analysis of a combined morphological and molecular data set, the new fossil taxon, Microvictoria was placed within the crown Nymphaeaceae, in a clade with the modern genera Victoria, Euryale and Nymphaea. The fossil Microvictoria dates to the Turonian (~90 mya), and the paleoclimate of the collection site is described as subtropical to tropical, wh ich corresponds to the climate of extant Victoria and Euryale. We used the age of Microvictoria as reported by Gandolfo et al. (2004) as a calibration point and repeated the divergen ce time exercise described above to evaluate the correspondence of molecularand fossil-derived ages Based on Gandolfo et al.s analysis, Euryale, Victoria Nymphaea, and Microvictoria form a polytomy with 68% bootstrap support. This placement most likely corresponds to node 8. However, calib ration using node 8 results in an angiosperm age estimate of 773 mya (Table 2-3). This placement of Microvictoria is not strongly supported. Alternative placements deeper in the tree, ba sed on higher bootstrap values, correspond to calibration at nodes 6 and 7, but calibration at these nodes results in age estimates for the angiosperms that are also too old, 410 and 687 mya, respectively. In fact, all estimates using this fossil are too old, until the fossil is placed outside Nymphaeales. For example, the age estimate for angiosperms is 311 mya when node 3, which subtends all crown-gro up Nymphaeales, is the calibration point; this age corresponds more closely to the age of extant seed plants (Mapes and Rothwell 1984; Mapes and Rothwell 1991). However, if node 2, which subtends all crown-group angiosperms except Amborella is the calibration point the age estimate for the angiosperms is 104 mya, which is too young, perhaps suggesting that Microvictoria might be best placed along the stem lineage leading to extant Nymphaeales (as found for the Friis et al. fossil), although
31 such a placement is at odds with the derived placement of Microvictoria within Nymphaeaceae favored by Gandolfo et al. (2004). However, Gandolfo et al. s topology disagrees with the morphological analysis of Les et al. (1999), who found strong support for Victoria + Euryale, consistent with many previous inferences of relationship in Nymphaeaceae; this sister-group relationship is not evident in Gandol fo et al.s tree, which places Victoria Euryale, Nymphaea and Microvictoria in a polytomy supported by number of pe tals greater than fi ve and number of stamens greater than 50. In the two s hortest trees of Gandolfo et al. (2004), Microvictoria is alternatively placed as either sister to Victoria or as sister to Victoria + Euryale. The collapse of the Victoria + Euryale sister group in the strict consensu s tree of Gandolfo et al. (2004) when Microvictoria is included must be due to the one-to-one structural and positional correspondence of floral organs of Victoria and Microvictoria reported by Gandolfo et al. (2004). Conflict between ch aracters that unite Victoria and Euryale and those that unite Victoria and Microvictoria is apparently responsible for the lack of resolution in this clade, although it is not possible to discern the pattern of character support from their study. There seem to be two possible explanations fo r the disparity between the fossil record for angiosperms and the age inferred here using PL and Microvictoria as placed within Nymphaeaceae by Gandolfo et al. (2004) as the calib ration point. Either methods of estimating divergence times from molecular data are highly dubious or the phylogenetic placement of Microvictoria in Gandolfo et al. (2004) may need to be reconsidered. Regarding the first possibility, several authors have reviewed the limitations of these methods (e.g., (Magalln 2004; Sanderson and Doyle 2001; Sanders on 2002; Sanderson et al. 2004; Solt is et al. 2002). However, despite these limitations, when combined data sets of multiple genes have been used, a series of recent studies has converged on similar, reasonable age estimates for the angiosperms (e.g.,
32 approximately130-190 mya; Magalln and Sanders on 2001; Sanderson and Doyle 2001; Soltis et al. 2002b; reviewed in Sanderson et al. 2004; Bell et al. 2005). Current evidence therefore indicates that, despite errors i nherent in the process of diverg ence time estimation, methods that account for rate heterogeneity among lineages typi cally provide estimates consistent with the fossil record, at least when all relevant lineages have been sampled and multiple genes have been included. Of course, lineage-specific rate deceleration, as observed in both angiosperms (Sanderson and Doyle 2001) and across tracheophytes (Soltis et al. 2002), may also account for the anomalously old ages inferred using both the Friis et al. (2001) and Gandolfo et al. (2004) fossils as calibration points. However, no such rate deceleration was observed for Nymphaeales in our study or in broader studies of angiospe rms (e.g., Zanis et al. 2002 ), although the apparent non-clocklike delay in di versification of most extant genera of Nymphaeales re lative to the age of the clade may contribute to the ol d ages estimated using these fossils. Alternatively, Microvictoria may be misplaced in the phylogenetic analysis of Gandolfo et al. (2004), perhaps due to homoplas y in the crucial morphological ch aracters scored and included in that study. That is, a now-extinct assemblage of early angiosperms may have possessed suites of traits not found in any extant groups. Friis et al (2000) stressed that many early angiosperm fossils exhibit character combinati ons unknown in extant angiosperms. Microvictoria may belong instead on the stem lineage to Nymphaeales a position that could not be evaluated by Gandolfo et al. (2004) because onl y Nymphaeales were included in their study. The placement of early angiosperm fossils in phylogenetic trees ma y be extremely challeng ing due to the mixture of characters present in early angiosperms. Howe ver, it would be worthwhile to reassess the features of both Microvictoria and the Friis et al. (2001) fossil and conduct additional phylogenetic analyses that include taxa from outside Nymphaeales.
33 Biogeography Nymphaeales have a worldwide distribution in tropical to temperate regions (Schneider and Williamson 1993; Williamson and Schneider 1993). The oldest putative fossil of Nymphaeales is from western Portugal (Fr iis et al. 2001), although caution regarding the phylogenetic placement of this fossil may be warra nted (see above). The fossil records from the Northern Hemisphere reveal that Brasenia and Nymphaea exhibited great species diversity in these areas during the Miocene (Cevallos-Ferr iz and Stockey 1989; Collinson 1980; Dorofeev 1973; Dorofeev 1974; Mai 1988). However, three genera ( Barclaya, Ondinea, and Victoria ) of Nymphaeales are not known from the fossil record, and their present distributions are restricted to small geographic areas (see below). The results of the DIVA analysis explain the present distributional pa ttern of Nymphaeales with seven inferred dispersal and two extinction events (Fi g. 2-2). Based on our estimated divergence times, Nymphaeales may have been wi dely distributed in the American continents and Eurasia during the Eocene (44.59 7.9 mya) (Fig. 2-2). This infere nce is consistent with the fossil record, with fossils of Nymphaeales found in the American continents and Eurasia (Anzotegui 2004; Cevallos-Ferriz and Stocke y 1989; Collinson 1980; Dorofeev 1974; Knobloch and Mai 1984). Cabombaceae were ancestrally distributed in th e American continents and Eurasia. After the two genera diverged, Cabomba either diversified only on the Am erican continents (Fig. 2-2) or, if it did occur in Eurasia, it subsequently became extinct there. Unfortunately, however, there is no fossil record of Cabomba to help infer the past distributional pattern of the genus. Two separate dispersals and an ex tinction event are needed to explain the present distributional pattern of Brasenia The distribution of Brasenia in Australia can be explained by dispersal from
34 southeastern Asia, but the presence of the genus in Africa is interpreted as ambiguous. Brasenia could have dispersed from either South Ameri ca or Europe to Africa. In the latter case, southward migration from Europe would have b een followed by the subsequent extinction of the genus in Europe, probably during glaciation. The fo ssil record lends support to this hypothesis; Brasenia was distributed in Europe during the Eo cene to Oligocene (C ollinson 1980). Based on our estimated divergence times, Cabombaceae diversified during the Miocene (19.91 5.6 mya), and at that time the connection between South America and Africa had already disappeared (Leys et al. 2002; Smith et al. 1994; Storey 1995), making this mi gration route unlikely (unless via long-distance dispersal). Da vis et al. (Davis et al. 2002a ; Davis et al. 2002b) provided evidence for a similar biogeographical scenario for members of Malpighiaceae with lineages occurring ancestrally in the Northern Hemisphere, but with subsequent migr ation to the tropics. Thus, the presence of Brasenia in Africa can be explained mo re plausibly by dispersal from Europe to Africa than by disper sal from South America to Africa. The present distribution of Nymphaeaceae provide s a more complicated pattern than that of Cabombaceae. The oldest Nuphar fossil is from the Paleocene of North America (Chen et al. 2004), and a fossil seed from the Early Eocene wa s recently reported from China (Chen et al. 2004). The fossil record th erefore indicates that Nuphar was already widespread in North America by the Early Eocene. The present distribu tion of Nuphar can be e xplained by retention of its ancestral distribution in the Northern Hemisphere. Nuphar is sister to the remaining extant Nymphaeaceae. Following this divergence, four of the remaining five genera ( Barclaya Ondinea, Euryale, and Victoria ) seem to have experienced range cont raction or separate dispersal events, based on their current distributional patterns. At present, these four gene ra are distributed in small geographic areas; Barclaya and Euryale are found only in southeastern Asia, Ondinea
35 occurs only in western Australia, and Victoria is native to South Americ a, especially the Amazon River basin and Paraguay. Fossils of Euryale are known from Eurasia (CevallosFerriz and Stockey 1989; Dorofeev 1974; Miki 1960). Thus, we can infer that the present distribution of Euryale might be the result of a more widespread distribution in the Northe rn Hemisphere followed by extinction in Europe, leaving the genus only in southeastern Asia. However, there are no fossils for Barclaya Ondinea, or Victoria so it is difficult to infer thei r past distributional patterns. Nymphaea has a cosmopolitan distribution, and it s present occurrence is wider than the inferred ancestral distribution of Ny mphaeales. The wide distribution of Nymphaea can be explained by three separate dispersal events (F ig. 2-2). The ancestral area of distribution for Nymphaea may have been North America, Europe, and Asia, because fossils of Nymphaea have been reported from North America and Eurasi a (Cevallos-Ferriz and Stockey 1989; Dorofeev 1974). Based on our estimated divergence times, Nymphaea may have experienced range expansion from North America to South Am erica via Central America. A fossil of Nymphaea from the Late Miocene of Argentina (Anzotegui 2004) supports our infere nce of range expansion from North America to South Ameri ca during the Miocen e. In addition, Nymphaea may have also dispersed from Eurasia into Africa a nd Australia during the Miocene. The modern worldwide distribution of Nymphaea may be the result of an ability of this lineage to adapt to a wider range of temperatures than other genera of Nymphaeales. Future Directions Estimates of divergence times and analyses of historical biogeogra phy rely heavily on the fossil record. In this study, we have relied on published interpretations of the placement of fossils of Nymphaeales. However, based on our a ttempts to reconcile fossil-based and molecular
36 dates, it appears that additional effort to plac e Nymphaealean fossils in the context of other lineages of early angiosperms is warranted. Like wise, the placement of fossils attributed to extant genera of Nymphaeales s hould be re-evaluated using exp licit analyses of characters and taxa. The morphological matrix of Les et al. (1999) provides an outst anding starting point for synthetic analyses of fossil and extant Nymphaeales.
37 Table 2-1. List of taxa used in this study with GenBank accession numbers and references Ingroup: Nymphaeales Cabombaceae. Brasenia schreberi J. F. Gmelin: 18S rDNA AF206874 (Soltis et al. 1999); rbcL M77031 (Les et al. 1991); matK AF092973 (Les et al. 1999). Cabomba caroliniana A. Gray: 18S rDNA AF206878 (Soltis et al. 1999); rbcL M77027 (Les et al. 1991); matK AF108719 (Les et al. 1999). Nymphaeaceae. Barclaya longifolia Wall.: 18S rDNA AF096692 (Les et al. 1999); rbcL M77028 (Les et al. 1991); matK AF092982 (Les et al. 1999). Euryale ferox Salisb.: 18S rDNA AF096694 (Les et al. 1999); rbcL M77035 (Les et al. 1991); matK AF092994 (Les et al. 1999). Nuphar variegata Durand: 18S rDNA AF096695 (Les et al. 1999); rbcL M77029 (Les et al. 1991); matK AF092979 (Les et al. 1999). Nymphaea odorata Aiton: 18S rDNA AF206973 (Soltis et al. 1999); rbcL M77034 (Les et al. 1991); matK AF092988 (Les et al. 1999). Ondinea purpurea den Hartog: 18S rDNA AF096697 (Les et al. 1999); rbcL AF102549 (Les et al. 1999); matK AF108722 (Les et al. 1999). Victoria cruziana Orb.: 18S rDNA AF096698 (Les et al. 1999); rbcL M77036 (Les et al. 1991). Victoria amazonica (Poepp.) Sowerby: matK AF092991 (Les et al. 1999). Amborellales. Amborellaceae. Amborella trichopoda Baill.: 18S rDNA U42497 (S oltis et al. 1997); rbcL L12628 (Qiu et al. 1993); matK AJ506156 (Goremykin et al. 2003). Austrobaileyales. Austrobaileyaceae. Austrobaileya scandens C. T. White: 18S rDNA U42503 (Soltis et al. 1997); rbcL L12632 (Qiu et al. 1993); matK AF543726 (Hilu et al. 2003). Schisandraceae. Illicium parviflorum Michx. ex Vent.: 18S rDNA L75832 (Soltis et al. 1997); rbcL L12652 (Qiu et al. 1993). Illicium floridanum J. Ellis: matK AF543738 (Hilu et al. 2003). Schisandra chinensis (Turcz.) Baill.: 18S rDNA L75842 (Soltis et al. 1997). Schisandra sphenanthera Rehder & E.H. Wilson: rbcL L12665 (Qiu et al. 1993). Schisandra rubriflora Rehder & E.H. Wilson: matK AF543750 (Hilu et al. 2003). Outgroup: Gymnosperms Taxaceae. Taxus mairei (Lemee et H.Lev.) S.Y. Hu ex T.S. Liu: 18S rDNA D16445 (Chaw et al. 1993); rbcL AB027316 (Chaw et al. 2000); matK AB024001(Cheng et al. 2000). Pinaceae. Larix leptolepis (Siebold & Zucc.) Gordon: 18S rDNA D85294 (Chaw et al. 1997). Larix decidua Mill.: rbcL AB019826 (Wang et al. 1999). Larix gmelini (Rupr.) Rupr.: matK AF143433 (Wang et al. 2000). Gnetaceae. Gnetum gnemon L.: 18S rDNA U42416 (Soltis et al. 1997); rbcL U72819 (Price 1996); matK AF542561 (Hilu et al. 2003). Ginkgoaceae. Ginkgo biloba L.: 18S rDNA D16448 (Chaw et al. 1993); rbcL AJ235804 (Chase et al. 1993); matK AF456370 (Quinn et al. 2002).
Table 2-2. Divergence time estimation: Age estimates for nodes 2-9 for different age estimati on methods (LF=Langley-Fitch; NPRS=nonparametric rate smoothing; PL=p enalized likelihood). Bootstrap= stan dard deviations from bootstrap resampling, =smoothing factor. CI=Bayesian credib ility intervals. All estimates base d on minimum age of the angiosperm clade (node 1, Fig. 2-1) constr ained to 131.8 mya (Soltis et al 2002). All dates estimated on fi xed tree (Les et al. 1999) with branch lengths for LF, NPRS, and PL optimized under maximum likelihood (ML) and maximum parsimony (MP) using a GTR+ model of sequence evolution; optimization of br anch lengths in Bayesian analysis under F84+ model Genes LFML LFMP N PRSMLBootstrap for NPRSML N PRSMP Bootstrap for NPRSMPPL ( ) PLML PL ( ) PLMP Bayesian 95% CI 18S rDNA 2 109.2 94.6 114.6 112.4.9 100.7 88.1.9 0.001 114.5 0.010 96.1 115.5.3 68.9-164.2 3 52.0 54.4 60.6 61.5.9 61.1 71.4.0 0.001 58.8 0.010 52.6 62.4.6 32.8-105.5 4 31.2 36.0 33.0 39.4.6 39.2 26.1.0 0.001 32.5 0.010 34.6 17.2.2 4.5-40.4 5 36.6 44.8 46.2 48.3.7 53.3 41.2.0 0.001 45.0 0.010 43.1 54.8.6 28.0-94.0 6 32.6 31.9 40.7 46.2.4 42.5 34.6.6 0.001 45.0 0.010 33.1 43.3.9 20.1-78.9 7 22.5 22.4 36.0 38.9.5 33.1 22.0.7 0.001 34.2 0.010 23.5 35.9.6 15.1-67.8 8 19.5 17.0 32.0 32.2.8 25.8 14.5.3 0.001 30.8 0.010 17.8 27.6.8 11.0-56.8 9 15.11 12.5 27.2 28.9.1 17.7 10.2.9 0.001 25.8 0.010 13.0 19.5.0 5.6-44.8 rbcL 2 111.3 104.9 115.4 115.2.5 107.7 83.7.5 0.001 112.4 0.0001 105.2 116.7.5 75.2-167.2 3 34.4 35.3 49.7 48.1.9 50.0 69.7.9 0.001 34.4 0.0001 35.6 42.3.5 20.5-80.9 4 20.7 20.7 28.7 27.5.0 27.7 14.4.3 0.001 20.6 0.0001 20.8 26.2.2 9.7-54.3 5 19.4 22.7 40.8 37.8.1 43.8 27.5.9 0.001 25.0 0.0001 23.1 27.1.2 12.1-59.4 6 15.2 17.8 33.8 34.8.9 36.6 21.2.7 0.001 19.3 0.0001 18.1 22.2.5 9.3-50.4 7 11.8 13.9 26.6 28.6.5 29.4 13.8.0 0.001 14.5 0.0001 14.1 17.4.8 6.7-41.1 8 8.9 9.0 19.9 24.5.1 19.5 10.2.7 0.001 10.6 0.0001 9.1 12.1.9 3.530.6 9 2.8 2.9 7.2 8.8.4 7.1 3.8.4 0.001 3.1 0.0001 2.9 4.7 .3 0.2-15.4 matK 2 109.6 95.5 118.3 122.4.0 102.4 71.8.0 0.0001 116.9 0.0001 99.9 116.4.3 75.9-167.8 3 28.9 29.1 56.2 49.6.1 44.3 53.0.9 0.0001 50.5 0.0001 37.9 34.4.8 19.1-57.4 4 13.4 9.8 26.5 24.2.4 16.2 6.2.2 0.0001 24.0 0.0001 12.4 16.5.4 3.7-20.1 5 19.5 22.3 48.3 32.2.2 39.1 18.0.2 0.0001 43.0 0.0001 32.5 33.2.8 18.2-55.7 6 15.7 17.5 34.3 42.4.4 31.8 11.9.8 0.0001 32.2 0.0001 26.2 27.4.9 14.9-48.3 7 13.4 7.7 26.5 23.6.1 15.1 4.7.2 0.0001 26.5 0.0001 12.1 14.8.5 6.1-31.1 38
Table 2-2. Continued. Genes LFML LFMP N PRSMLBootstrap for NPRSML N PRSMPBootstrap for NPRSMPPL ( ) PLML PL ( ) PLMP Bayesian 95% CI 8 7.5 7.2 25.7 23.6.1 13.6 4.1.0 0.0001 25.7 0.0001 11.0 11.7.6 4.3-25.6 9 6.2 5.9 20.5 19.1.0 10.4 3.1.8 0.0001 20.5 0.0001 8.7 8.5.7 2.2-20.2 Combined 2 110.9 97.4 116.3 115.1.9 100.8 73.4.7 0.0001 116.1 0.0001 97.9 118.4.2 107.5-127.9 3 33.7 34.7 46.9 45.3.6 43.9 55.8.7 0.0001 45.8 0.0001 37.6 44.6.9 29.5-60.7 4 18.6 16.3 24.8 23.9.1 19.9 9.5.2 0.0001 24.0 0.0001 16.4 19.9.6 11.0-32.9 5 27.8 26.5 44.6 37.2.7 37.9 20.4.8 0.0001 43.4 0.0001 31.3 41.1.7 26.8-56.9 6 21.3 20.3 37.3 36.1.3 30.1 14.5.3 0.0001 36.2 0.0001 24.6 34.7.3 21.5-50.0 7 11.8 11.8 23.0 23.0.4 18.4 6.8.6 0.0001 22.3 0.0001 14.7 23.3.2 13.0-37.1 8 10.5 9.4 20.2 20.7.6 14.0 5.1.1 0.0001 19.7 0.0001 11.5 20.4.8 11.0-33.2 9 7.6 6.7 13.8 14.8.3 8.9 3.2.8 0.0001 13.4 0.0001 7.8 14.5.8 7.2-25.5 39
40 Table 2-3. Effect of fossil placement. Inferred age (in millions of years) of the angiosperms if each of two described fossils is used as a calibration point at that node. All age estimates were computed using the penalized likelihood method (Sanderson 2002). Node numbers are given in Figure 2-1 Constrained node Inferred age of angiosperms (Gandolfo et al. ~90 mya) Inferred age of angiosperms (Friis et al. ~ 120 mya) 2 104.1 138.9 3 310.9 414.5 4 603.2 804.2 5 334.9 446.5 6 410.3 547.1 7 686.8 915.8 8 772.7 1030.3 9 1093.0 1457.3
Figure 2-1. Phylogenetic tree of Nymph aeales from Les et al. (1999), with Amborella Austrobaileyales, a nd gymnosperm outgroups added, based on many recent analyses (see text). Phylogram depi cts branch lengths based on the combined data set of 18S rDNA, rbcL and matK optimized via maximum parsimony. Numbers si gnify nodes for which divergence times were estimated. Solid star on branch leading to node 5 indicates the preferred placement of the Friis et al. (2001) fossil by those authors, and open star on branch leading to node 8 represents the preferred position of Microvictoria by Gandolfo et al. (2004). 41
42 Figure 2-2. Chronogram for Nymphaeales show ing timing of inferred divergences and biogeographic events. Closed bars repr esent dispersal events, and open bars indicate extinction events. Present distri butions are given as NA=North America, SA=South America, CA=Central Ameri ca, E=Europe, A=Asia, AF=Africa, AU=Australia. Divergence time estimates for each node are from the Bayesian analysis, based on the combined data set of 18S rDNA, rbcL and matK (Table 22). Scale at bottom indicates geological time scale: Pal = Paleocene, Eoc = Eocene, Oli = Oligocene, Mio = Miocene, Pl = Pliocene, KT = the CretaceousTertiary boundary.
43 CHAPTER 3 COMPARATIVE STUDIES OF FLORAL DEVELOPMENT IN NYMPHAEALES Introduction The molecular genetics of floral developm ent is well established for model eudicot plants such as Arabidopsis (Coen and Meyerowitz 1991; Ma and dePamphilis 2000; Smyth et al. 1990; Theissen et al. 2000; Weigel 1995; Zhao et al. 2001). The well-known ABCDE model, based on mutant studies in Arabidopsis explains floral organ formation by combinatorial action of spatially specific transc ription factors. According to this model, A and E class genes control sepal id entity; A, B, and E class genes control petal identity; B, C, and E class genes control stamen identity; C and E class genes control carpel identity; D and E class genes control ovule iden tity (Coen and Meyerowitz 1991; Colombo et al. 1995; Pelaz et al. 2000; Theissen 2001a). Many features of the ABCDE model are conserved across angiosperms (i.e., Petunia hybrida (Angenent et al. 1993); Silene latifolia (Hardenack et al. 1994); Gerbera hybrida (Yu et al. 1999); Oryza sativa (Fornara et al. 200 3; Kater et al. 2006; Kyozuka et al. 2000); Pisum sativum (Taylor et al. 2002); Zea mays (Whipple et al. 2004); Magnolia grandiflora (Kim et al. 2005); Persea americana (Chanderbali et al. 2006), Akebia trifoliata (Shan et al. 2006); Elaeis guineensis (Adam et al. 2007); Taihangia rupestris (L et al. 2007); Vitis vinifera (Poupin et al. 2007). Buzgo et al. (2004a) suggested that compar ative studies of developmental morphology and genetics will be required to build up a new set of model plants for floral genetics outside of those organisms studied to date, most of which are eudicots. However, so far these have been few studies involving basal angiosperms. Noteworthy studies includ e investigation of a basal eudicot, Eschscholzia californica (Becker et al. 2005), basal angiosperms, Amborella trichopoda (Buzgo et al. 2004), and a magnoliid, Persea americana (Buzgo et al. 2007). Additional comparative studies throughout angi osperms are needed for a comprehensive comparative floral developmental paradigm.
44 Nymphaeales are the sister to all extant angiosperms except Amborella (Borsch et al. 2005; Hilu et al. 2003; Jansen et al. 2007; Lhne and Borsch 2005; Leebens-Mack et al. 2005; Mathews and Donoghue 1999; Moore et al. 2007; Qiu et al. 1999; Qiu et al. 2005; Soltis et al. 1999; Soltis et al. 2000; Soltis et al. 2005; Zanis et al. 2002; Zanis et al. 2003). They exhibit distinct floral morphologies from those of the other basalmost angiosperms (i.e., Amborella Austrobaileyales) in severa l features, including the presence of large flowers in some members (i.e., Nuphar Nymphaea, Euryale, and Victoria ) rather than small or moderate-sized flowers, whorled rather than spiral phyllotaxis, eudicot-like perianth differentiation (as in Cabomba and Nuphar ), and the occurrence of blue perianth organs containing anthocyanins, whic h are absent in other basal angiosperms (Endress 2001; Schneider et al. 2003). Thus, studying floral de velopment in Nymphaeales will help elucidate floral evolution in early-dive rging angiosperms, and thus provide insight into the early diversification of the flower. There have been several previous floral developmental studies in Nymphaeales (Cutter 1957a, b, 1961; Endress 2001; Moseley 1958, 1961, 1965, 1972; Moseley et al. 1993; Schneider et al. 2003; Tucker and Douglas 1996). For example, Cutter (1957a, b, 1959, 1961) mainly focused on early floral development in Nuphar with an emphasis on the scaleappendage. She considered that organ to be a sepal because it is formed by the young primordium in the same position as a sepal of Nymphaea, although Moseley (1972) argued for its interpretation as a bract. However, most of these earlie r works focused only on specific genera, such as Nymphaea and Nuphar (Cutter 1957a, b, 1959, 1961; Moseley 1961, 1965, 1972), on specific developmental stages or or gans, or on morphological features (Ito 1983, 1984, 1986; Khanna 1964, 1967; Moseley 1958; Mosele y et al. 1984; Moseley et al. 1993; Osborn and Schneider 1988; Prance and Arias 1975; Schneider 1976; Schneider and Moore 1977; Schneider and Jeter 1982; Schneider 1983; Williamson and Moseley 1989; Williamson
45 and Schneider 1994). Recently, Endress (2001) exam ined floral development of extant basal angiosperms, and he included three species from Nymphaeales: Cabomba furcata Nuphar advena and Victoria cruziana However, the most extensive study of flower development in Nymphaeales is that of Schneider et al. (2003). They investigated developmental stages from flower initiation or sepal initiation to carpel form ation across all genera of water lilies except Barclaya and Hydatellaceae, the latter only re cently placed in Nymphaeales based on molecular phylogenetics (Saarel a et al. 2007). However, Schne ider et al. (2003) did not examine developmental stages after carpel fo rmation. Furthermore, they did not align the developmental sequences among genera; recommenda tions for standardizing descriptions of developmental sequences to allow for alignment among species were presented by Buzgo et al. (2004a), after Schneider et al.s (2003) study. In this study, I describe the floral developm ental series of three species of Nymphaeales, Cabomba caroliniana A. Gray, Nuphar advena Aiton, and Nymphaea odorata Aiton. First, I completed examination of the floral developm ental series of all three species. Early developmental stages are well-documented fo r these three species (Endress 2001; Schneider et al. 2003; Tucker and Douglas 1996), so I focused on later developmental stages from carpel formation to anthesis, and I then aligned the floral developmenta l series according to Buzgo et al. (2004a). I then compared the aligne d stages for these three species to each other and to other angiosperms. In addition, I examined the third whorl of Nuphar advena in detail. Organs in the third whorl were traditionally considered petals based on the pr esence of nectaries (e.g., Crow and Hellquist 2000; Endress 2001; Moseley 1965, 1972; Padgett et al. 1999; Schneider and Williamson 1993; Schneider et al. 2003; Wierse ma and Hellquist 1998 ; Zanis et al. 2003; Zomlefer 1994), while some authors interpreted these organs as staminodes based on their stamen-like appearance (Judd et al. 2002; Kim et al. 2002; Warner et al. 2008). To try to
46 elucidate whether these organs are petals or staminodes, I observed their development in detail and compared it with that of stamens. In th is chapter, I used the te rm petals to refer to the third-whorl organs. The results of this study were used in seve ral subsequent studies, for example, floral gene expression profiles at the level of specific genes (i.e., fl oral organ identity genes in chapter 4) and of the entire transcriptome (chapter 5). Materials and Methods Floral buds at various developmental stages ranging from flower in itiation to prior to anthesis were collected from the following taxa: Cabomba caroliniana, plants purchased from a local aquarium store, Ga inesville, FL, USA (Yoo 10020, FLAS); Nuphar advena, in Waccassasa River, near Waccassasa Preserve Area, Levy Co., FL, USA (Yoo & Koh 1000, FLAS); Nymphaea odorata, in Waccassasa Preserve Area, Levy Co., FL, USA (Yoo & Koh 1001, FLAS); and Nymphaea capensis Thunb., plants purchased from a local aquarium store, Gainesville, FL, USA (Yoo 10021, FLAS). Samples we re fixed in FAA (formalin, acetic acid, alcohol) and then transferred and stored in 70% ethanol. Studies were performed by scanning electr on microscopy (SEM) and serial microtome sectioning. For SEM, samples were dehydrate d, critical-point drie d, gold sputtered, and examined with a Hitachi S-4000 FE-SEM with an acceleration of 4.0 kV at the University of Floridas ICBR Electron Microscopy Laboratory. For the serial sections, samples were dehydrated in a series of alcohol, and then transferred to xylene and embedded in paraplast. Samples were sectioned to a thickness of 5 m to 20 m using a rotary microtome. Sectioned samples were stained with Sasss safranin-fast green or toluidine blue O (pH < 5) (Ruzin 1999). To facilitate comparison among the taxa surveyed, I employed the developmental landmarks for alignment of floral developmenta l stages proposed by Buzgo et al. (2004a).
47 Stages 1 to 5 were well describe d from previous studies for all th ree species investigated here (Cutter 1957b; Endress 2001; Moseley 1961, 1972; Schneider et al. 2003; Tucker and Douglas 1996), so I mainly focused on later developmental stages although I included some early developmental stages to ensure overlap. Results Developmental Stages Flower development in the three species inve stigated here were aligned according to the developmental landmarks proposed by Buz go et al. (2004a) and generally proceeds through 10 developmental stages in the same order (Table 3-1). Description of floral developmental stages of each species is followed below. Cabomba caroliniana: The flower of Cabomba caroliniana consists of four trimerous whorls: three sepals, three petals, six stam ens, and three carpels. Early developmental features of Cabomba were well studied by Moseley et al (1984), Tucker and Douglas (1996), Endress (2001), and Schneider et al. (2003), so I only focuse d on developmental features from carpel initiation to anthesis. At stage 5 ( carpel initiation), carpel primordia arise from a flat apical meristem (Fig. 3-1 A ) and become cylinder-shaped (Fig. 3-1 B ). Later, carpels become ascidiate at stage 6, at which time microsporangia initiate at the abaxial side of the stamen (Fig. 3-1C ). At this stage, the anther locule is filled with sporogenous tissue (Fig. 31D ). Just after this, ovules ini tiate along the dorsal wall of th e locule (stage 7: Fig. 3-1 E ), and their development continues until megaspores have formed (Fig. 3-1 H J ). After ovule initiation, male meiosis occurs within the anthers (stage 8): the microspore mother cell and tetrads were observed in Fig. 3-1 F and G respectively. At stage 9 (female meiosis), megaspores are formed (Fig. 3-1 H I ), and nuclei were observed within the ovule (Fig. 3-1 J ). Anthesis occurs over two consecutive days: duri ng the first day the flower is pistillate, and
48 the stigma is receptive (Fig. 3-1 L ), and in the second day the fl ower is staminate, and the stamens are dehiscent (Fig. 3-1 M ). Nuphar advena: The flower of Nuphar advena has a single bract, six sepals in two whorls, 16 to 18 petals in two whorls, numer ous stamens, and many carpels. Although many researchers had studied floral development in Nuphar most focused on various features, such as bracts and flower initiati on (Cutter 1957a, b), and early fl oral development (Endress 2001; Moseley 1965, 1972; Schneider et al. 2003). I in cluded several early developmental stages. At stage 2 (initiation of sepals), six sepal prim ordia are initiated in two whorls; three of them occurred in the first whorl, and the other thr ee sepals in the second whorl. There is a long plastochron between initiation of the third and fourth sepals, so at this stage a big size difference between sepals was observed (Fig. 3-2 A ). At stage 3, petal primordia are initiated in the third whorl (Fig. 3-2B ), and two trimerous whorls of pe tals appear at stage 5 (carpel initiation; Fig. 3-2 H and Fig. 3-4D ). Stamen primordia are initiat ed in the fourth whorls, and later they showed orthostichies (Fig. 3-2 B-E ). The carpel primordia br oaden at stage 5 (Fig. 3-2F), and later carpellary ridges are formed and deepen (Fig. 3-2 G H ). Next, microsporangia are formed on the adaxial surface of stamens (Fig. 3-2 I J ), and ovules are initiated along the lateral and dorsal walls of carpel locules (Fig. 3-2 K ). At stage 8, male meiosis initiates; the pollen sac is fi lled with sporogenous tissues (Fig. 3-2 L ). Female meiosis occurs at stage 9, at which point nucle i are observed within ovules (Fig. 3-2 N O ). As in Cabomba caroliniana, anthesis occurs over two consecutive days: pistillate flower in the first day, and staminate flower in the second day. Nymphaea odorata : The flower of Nymphaea odorata consists of several tetramerous whorls; four sepals, 28 petals, numerous stamens, and many carpels. The general developmental features of Nymphaea were studies by Moseley (1961) and Schneider et al. (2003). After sepal initiation, peta l primordia are initiated alternating with the first four
49 perianth members, which are sepals (Fig. 3-3 A ). At stage 4, stamen primorida are initiated along the dome-shaped apical meristem, but the central region is flat (Fig. 3-3 B ). Later, the apical meristem becomes depressed, and along th e peripheral region of this central region more stamens are formed (Fig. 3-3 C ). At a later stage, the cent ral area bulges, and it becomes a floral apex (Fig. 3-3 D ). When a visible floral apex has formed, carpel primorida are initiated at the periph eral region (Fig. 3-3 E ). At stage 6, microsporangia are formed at the adaxial surface of the stamens, and a floral ap ex becomes a ball-shaped structure (Fig. 3-3 F H ). Also, at this time, carpellar y locules are formed (Fig. 3-3 G ), and later ovules initiate along the lateral and dorsal wall s of those locules (Fig. 3-3 I ). After ovule initiation, male meiosis occurs; the pollen sac is fill ed with sporogenous tissue (Fig. 3-3 J ). At stage 9, female meiosis initiates; ovules are an atropous, and at this time the megaspore is formed in each ovule (Fig. 3-3 K L ). Anthesis also occurs over two c onsecutive days, and it shows the same patterns as those observed in Cabomba and Nuphar Developmental Features of the Third-Whorled Organs of Nuphar advena At anthesis, there are 16 to 18 stamen-like organs or petals in th e third whorl between two perianth whorls and stamens of Nuphar advena. In fact, they are developed from the third, fourth, and/or fifth whor ls of a floral bud in early de velopmental stages (see below). However, later in development, they are densely packed in a li mited space, so they appear to be present in the third whorl at anthsis. In this chapter, I refer thses third-whorl organs to petals. Petal primordia arise at the third whorl (Fig. 3-4 A ), and they are much larger than stamen primordia at the stage of stamen initia tion. At stage 4, the petal primordia broaden and become flat (Fig. 3-4 B ), and at the stage of ca rpel initiations (stage 5) a tip of the stamen primordia is projected and petal primordia become perianth-like structures (Fig. 3-4 C ). There seem to be two whorls of petal primordia (Fig. 3-4 D ); however, at maturity, petals appear to
50 occurr in the third whorl (Fig. 3-4 E ). 16 to 18 petals are observed at anthesis (Fig. 3-4 F ). At stage 9 (female meiosis), stamens have four mi crosporangia, and petals have nectaries on the abaxial side (Fig. 3-4 E ). Stamens have a ridge in the middle of the abaxial side (Fig. 3-4 G H ), but petals do not have su ch a structure (Fig. 3-4 I-O ). There are also intermediate forms of petal-like structures which have characterist ics of petals as well as stamens (Fig. 3-4I-N ). They have sterile or fertile microsporangi a on the adaxial surface, but no ridges on the abaxial side, and generally they do not have nectaries on the abax ial surface like normal petals (Fig. 3-4I-L ). However, there are some petals with sterile mi crosporangia on the adaxial surface and nectaries on the abaxial side (Fig. 3-4 M ). At stage 9, stamens have four fertile microsporangia, and the st amen tip is projected (Fig. 3-4 N ). However, a ridge on the abaxial side of stamens is ve ry weakly developed (Fig. 3-4 O ). At anthesis, mature petals have two weak ridges on the ad axial side (Fig. 3-4 P ), and the stamens are now dehiscent (Fig. 34Q ). Discussion Comparison of Developmental Stages in Three Species of Nymphaeales Flower development in the thre e species investigated here, Cabomba caroliniana, Nuphar advena, and Nymphaea odorata proceed through 10 developmental stages in the same order (Table 3-1). However, variation o ccurs among species in se veral stages. At stage 1 (flower initiation) in Nuphar and Nymphaea, floral buds arise as lateral shoots on the rhizome, so they replace a leaf in the ontogenetic spiral (Cutter 1957a, b; Schneider et al. 2003). However, in Cabomba flowers are axillary to floatin g leaves, which have decussate phyllotaxy if two floating leaves are present. Ge nerally, the next flower occurs proximally and distantly against the previous flower. All three species exhibit acrope tal organogeny. Two or more whor ls of perianth initiate in a similar way in all three species. In Cabomba, the trimerous whorl of sepals emerges
51 simultaneously, and subsequently a trimerous whorl of petals initiates in alternate position to the sepals. Nuphar displays similar developmental features to Cabomba; a trimerous whorl of sepals initiates, although there is a time lapse among the firs t three sepals (they are not simultaneous), and then after a long plastochron, another trimerous whorl of sepals emerges in alternate position to the previous sepal primordia. In contrast, Nymphaea exhibits a different sequence of sepal in itiation although it also has whor led phyllotaxy. The four sepals are initiated in a unidirectional sequence; first, the abaxial sepal, next the two lateral sepals, and last, the adaxial sepal. Next, the first four petal primordia occur alternating with the first four sepals, and two additional petal whorls fo rm. Finally, in the fifth and sixth whorl of N. odorata eight petal primordia initiate in alterna ting position to the previous petal primordia, resulting in many petals in Nymphaea. The difference among the three species studied is clear at stage 4 (initiation of stamens). In Cabomba, six stamen primordia form in alte rnate position with petal primordia. Nuphar and Nymphaea have numerous stamens, so they have a bigger apical meristem. In Nuphar the central region becomes dome-shaped, so many stamen primordia are formed along the peripheral region. Nymphaea also has a bulging apical meri stem at stage 4, but it becomes depressed as stamen primordia are formed. Afte r stamen primordia formation, the central area gets swollen again and becomes the floral ap ex. Around this floral apex, carpel primordia initiate. The main developmental difference at this point between Nuphar and Nymphaea is that the floral apex (central area) becomes the carpels in Nuphar while only the peripheral region gives rise to carpel primordia in Nymphaea. The central area of Nymphaea becomes the apical residuum, and this is considered a derived character for members of Nymphaeaceae, such as Ondinea, Victoria and Euryale (Schneider et al. 2003). In Cabomba, a trimerous whorl of carpel primordia initiates in alternat e position to th ree of the paired stamens.
52 At stage 6, microsporangia initiate on the abaxial (Cabomba) or adaxial ( Nuphar and Nymphaea) side of the stamens. All three species s how a similarity in ovule initiation; ovules initiate along the dorsal (Cabomba) or dorsal/lateral ( Nuphar and Nymphaea) walls of the locules. After ovule init iation, male meiosis initiates. At stage 9, ovules are anatropous in all three species, and the megaspore is observed. At anthesis, they all have protogynous flowers, so the flower is pistillate on the firs t day and is staminate on the second day. From comparison of developmental stag es in all three species, sometimes Nuphar exhibits Nymphaea -like or Cabomba-like features. For example, Nuphar exhibits Nymphaealike characteristics; in flower initiation of both species, the floral primordia replace a leaf in a phyllotactic spiral. Also, the stamen primordia of Nuphar are initiated along the peripheral region of the central area as those of Nymphaea are formed around the periphery of the central area. Nuphar and Nymphaea also exhibit similar microsporangia initiation; four microsporangia initiate on the adaxial surface of the stamens. Finally, ovules of Nuphar and Nymphaea initiate along the lateral and dorsal walls of locules, which is due to their syncarpous carpels. However, Nuphar also exhibits Cabomba-like developmental features. First of all, initiation of sepals a nd petals occurs in the same way in Nuphar and Cabomba; a trimerous whorl of sepals emerges simultaneously ( Cabomba) or subsequently ( Nuphar ), and another trimerous whorl of petals (or sepals in Nuphar ) initiates in an alternate position with the first three sepals. Next, six stamen or petal primordia arise in the third whorl of Cabomba and Nuphar respectively, in a double position against the previous petal or sepal primodia. These developmental features of Nuphar compared to Cabomba and Nymphaea may be explainable by the phylogenetic position of Nuphar Nymphaeaceae and Cabombaceae are sister families that share many features, incl uding vegetative, reproductive, and molecular characters (Borsch et al. 2007, 2008; Lhne et al. 2007; Les et al. 1999). Within Nymphaeaceae, Nuphar is sister to all remaining genera within Nymphaeaceae (Borsch et al.
53 2007, 2008; Lhne et al. 2007; Les et al. 1999). Although Nuphar shares several morphological features with other members of Nymphaeaceae, Nuphar also shares some morphological characteristics with Cabomba, such as, anasulcate pollen, absence of staminodes (although its absence is doubtable), separate or discontinuous stigmatic surface, and flower maturation above water (Ito 1987; Le s et al. 1999). Based on these characteristics, several authors have suggested that Nupharaceae be recognized as a separate family (Kerner von Marilaun 1891; Nakai 1943; Takhtajan 1997). However, despite of these morphological traits, its phylogenetic position as as sister to other Nymphaeaceae is strongly supported by three genome analyses (Borsch et al. 2008). Therefore, sharing developmental features of Nuphar with Cabomba may reflect ancestral state for core Nymphaeales. Outer and Inner Tepals or Sepals and Petals Although Endress (2008) reviewed the perianth structure of Cabomba and Nuphar here I focused on whether perianth members of the three study species are differentiated into sepals and petals. In fact, there has long been a question as to whether the outer sepaloid and inner petaloid tepals of these water lilies are homologous to the sepals and petals of eudicots, respectively. However, most researchers use the term outer and inner tepals instead of sepals and petals for basal angiosperms because of only slight morphological differentiation between them and their spiral arrangement. For Nymphaeales, many authors have used the terms sepals and petals for the outer and inner tepals respectively, based on their position and color difference. In this study, I used term se pals and petals because the two whorls are somewhat differentiated from each other (see Warner et al. 2008). In Cabomba, following their simultaneous appearan ce, sepal primordia continue their development, so at stage 8 (male meiosis), th ey enclose a bud. In contrast, petal primordia remain very small and do not expand until just prior to anthesis. In addition to this developmental retardation in pe tals, petals are easily disti nguished from sepals by having
54 auricular-shaped nectaries at the base However, both sepals and petals in Cabomba have a single vascular trace (Ito 1986; Moseley et al. 1984), similar sh eath-like texture, and similar color. Therefore, the two organs are not eas ily distinguishable except for developmental features and the presence of nectaries. Su ch undifferentiated perianth members are common in monocots. In Nuphar two whorls of sepals initiate, and they are much bigger than the petal primordia. Endress (2001) poi nted out that petals of Nuphar are retarded in development relative to the sepals and stamens. However, I could not observe any retardation in petal development in Nuphar in this study. Petal primordia are initiated after se pal formation, and they continue their development. I can see br oad and spatulate petals at stage 4 (stamen initiation), and some petals become thicker and larger than stamens at stage 5 (carpel initiation). As a result, in Nuphar petals are different from sepals with regard to development. Also, the six sepals have three vascular traces, while petals and stamens each have a single trace (Moseley 1958). Therefore, considering that there are no members in Nymphaeales having two whorls of sepals, and the similarity of petals to stamens (see below), the petals of Nuphar may not be homologous to petals of other Nymphaeales. Alternatively, the first three perianth parts are sepals, and the next three perianth members may be petals. In Nymphaea, four sepal primordia are initiated in the first whorl and develop fast, so at stage 4 (initiation of stamens), they are much bigger than petal primordia. Petal primordia also develop continuously, so at the later stage of stamen initiation they consist of several whorls (or layers) of petals. Th erefore, there is no distinctiv e difference between sepals and petals in Nymphaea except for their position and color. Also, these organs all have three vascular traces (Moseley 1961). Sepals are generally differentiated from petals in many wa ys; their position (1st whorl for sepals vs. 2nd whorl for petals), functi on (protection in sepals vs attraction of pollinators
55 in petals), development (spiral initiation of se pals vs. simultaneous initiation of petals), vascular system (three traces in sepals vs. one trace in petals), color (greenish sepals vs. colorful petals), epidermal cell type (flat in sepals vs. conical in petals), and general similarity (foliage leaves vs. stamens) (reviewed in Albert et al. 1998 and Warner et al. in press). However, these characteristics are found in eudi cots, which have bipartite perianth members. Based on the features above, perianth memb ers of Nymphaeales are distinguished by a limited number of traits. Thus, sepa ls and petals may not be appr opriate terms for the perianth in Nymphaeales even though perianth memb ers are differentiated to some degree. Identity of the Third-Whorled Organs of Nuphar advena : Petals or Staminodes Petals of Nuphar can be easily distinguished fr om stamens by their position (occurrence at the third whorl) and size (muc h wider than stamens). However, the outer appearance of petals of Nuphar advena is very similar to those of stamens, although there are several differences. Petals are oblong to spatul ate with truncate tip, a nd they have nectaries on the abaxial surface, while stamens are obl ong with projected tip with two pairs of microsporangia on the adaxial surface. They also share developmental features. Petal primordia arise alternately w ith the first three sepal primordia in double position (Endress 2001). Thus, the first six petals are present in th e third whorl at stages 2-5. Next, nine petal primordia formed in the fourth whorl, but thes e two whorls are relatively indistinctive after stage 5 as their size increases. S ubsequently, in the very next w horl, stamen primordia initiate. I observed 16 to 18 petals in flowers from the Waccassasa River population. However, from those two whorls of petal primordia only 15 peta ls can be formed. Our results from serial sectioning showed that one to three stamens from the outermo st stamen primordia convert into petals, perhaps due to thei r position. These stamens are located next to the petal whorls, and sometimes extra space is present between petal primordia. Therefore, some stamen primordia can be placed at that position, and they can become petals. In this case, they have
56 sterile or fertile microsporangia on the adaxia l surface, but do not have nectaries on the adaxial surface, although there ar e exceptional cases (see below). Petals and stamens initiate in different whorls; petals from the third and fourth whorls, and stamens from the fifth whorl to the more inner whorls. Also, the primordia of stamens and petals are different in shape as well as position. Petal primordia become flat and scalelike structures as stages proceed (Fig. 3-4), while stamen primordia are dome-shaped and the apices of primordia become projected (sterile appendages) (Padgett 200 7). At stage 8, petals have nectaries on the abaxial surf ace and two ridges on the adaxial surface. At the same stage, stamens have well-developed microsporangia on the adaxial surface and a conspicuous ridge on the abaxial surface. I observed intermediate forms of petals, which have microsporangia on the adaxial surface, ranging from sterile (or a borted) to fertile (fully developed). Those petals are relatively small and present in the bo rders between petals and stamens. Also, they do not have a ridge on the abaxial surface nor nect aries. At anthesis, the number of petals ranges from 16 to 18, and sometimes there ar e intermediate forms between stamens and petals in petal whorls. In Nymphaeales, perianth members from the first and second whorls (sepals and petals, respectively) are similar in shape and size. For example, Cabomba has two petaloid whorls (similar to monocots), and Nymphaea has four greenish sepals and many petals. However, only Nuphar has been regarded as having perianth me mbers of different size; petals are very small compared to sepals. In addition, sepals occur in the first tw o whorls, resulting in presence of petals in the third whorls. In contrast to the presence of staminodes in Nymphaeaceae, but not Cabombaceae, Moseley (1958) noted that Nuphar does not have staminodes because there is no morphological gr adual transition from stamens to petals. To define the identity of the organs at the third whorl, we have to think of three categories of homology; historical homology, positional homology, and process homology
57 (Albert et al. 1998). First of all, these organs of Nuphar are positionally homologous to stamens of other members of Nymphaeales, particularly in having a small number of perianth organs, as do Cabomba, Brasenia and Ondinea. Considering the important phylogenetic position of Barclaya as subsequent sister to all other Nymphaeaceae after Nuphar developmental data for Barclaya is required to elucidate this problem (Borsch et al. 2007; Borsch et al. 2008; Lhne et al. 2007; Les et al. 1999). Also, when we examine the developmental features of the organs in the third whorls, they are much more similar to stamens, not to the organs of the first tw o whorls. In addition, developmental evidence presented here also suggests that intermediate forms between stamens and petals at the third whorls might have originated from stamens, although there is no gr adual transition from stamens to these organs. There is controversy in the definition of a staminode, but aborted stamens, whatever their structures are, should obviously be referred to as st aminodes (Ronse De Craene and Smets 2001). According to this view, the organs in the third whorl of Nuphar should be considered staminodes because some of them sh ow reminiscent characte ristics of stamens, for example, sterile or fertile microsporangia on the abaxial surface. In addition, their position between two whorls of perianth s and stamens, and similar developmental features to stamens further support this idea. Howe ver, detailed investigation of staminodes from other members of Nymphaeales is needed to assess whether they are all developmentally and historically homologous. Also, comparison of gene expression patterns of these organs with those of stamens will be helpful to clarify this problem.
58Table 3-1. Stage alignment using descriptions of floral developmental stages for three taxa Stage, developmental landmark Cabomba caroliniana Nuphar advena Nymphaea odorata 1. Flower initiation 1Flowers are axillary to floating leaves, which have decussate phyllotaxy (0.17 mm) 4, 5Flowers appear as lateral shoots on the rhizome (0.2-0.3 mm) 3Flowers appear as lateral shoots on the rhizome (0.09-0.18 mm) 2. Initiation of bract & sepals 2A trimerous whorl of sepals emerges simultaneously (0.17 mm in C. furcata ) 2, 4, 5A long plastochron between the initiation of the third and fourth organs (0.28-0.7 mm) Four calyx members are initiated in unidirectional sequence: first, the abaxial sepal; next the two later sepals; and last the adaxial sepal 3. Initiation of petals 1, 3A trimerous whorl of petals initiates in alternate positions to sepals (0.14-0.21 mm) 2, 5Petal primordia emerge in double positions (0.7-0.9 mm) 3The first four petals are initiated alternating with sepals (0.3-0.65 mm) 4. Initiation of stamens 1, 3Six stamen primordia initiate (0.18-0.36 mm) 2, 5The floral apex becomes domeshaped (0.88-1.10 mm) 2, 6Stamen primordia are initiated in whorls around peri phery of inactive central region, central region becomes depressed, and later floral apex bulges (0.8-1.6 mm) 5. Carpel initiation 1, 3A trimerous whorl of carpel primordia initiates in alternate position to three of the stamens (0.32-0.46 mm) 2, 5Carpellary primordia broaden laterally slightly (1.5-3.0 mm) 6Carpels rise around periphery of floral apex (2.0 mm) 6. Microsporangia initiation Four microsporangia initiate on abaxial side of stamens (0.6 mm) Four microsporangia initiate on adaxial side of stamens (3.2-4.0 mm) Four microsporangia initiate on adaxial side of stamens (2.8-4.0 mm) 7. Ovule initiation Ovules in itiate along the dorsal walls of locules (0.8 mm) Ovules initiate along the lateral and dorsal walls of locules (4.6-6.4 mm) Ovules initiate along the lateral and dorsal walls of locules (4.0-4.8 mm) 8. Male meiosis Formation of sporogenous tissue in anthers, microspore formation (1.2 mm) Formation of sporogenous tissue in anthers (6.7-9.2 mm) Formation of sporogenous tissue in anthers, microspore formation (6.4-7.4 mm)
59Table 3-1. Continued Stage, developmental landmark Cabomba caroliniana Nuphar advena Nymphaea odorata 9. Female meiosis Ovule anatropous, with megaspore (1.6 mm) Ovule anatropous, with megaspore (11.2 mm) Ovule anatropous, with megaspore (6.0-9.6 mm) 10. Anthesis 1 Stigma receptive Stigma receptive Stigma receptive 10. Anthesis 2 Stamens dehiscent St amens dehiscent Stamens dehiscent NoteI have observed developmental stages from stamen initiation to anthesis. Some developmental stages were taken from previo us studies: 1Tucker and Douglas (1996); 2Endress (2001); 3Schneider et al. (2003); 4Cutter (1957); 5Moseley (1972); 6Moseley (1961).
60 Figure 3-1. Developmental series of Cabomba caroliniana. All longitudinal sections. A B Stage 5: carpel initiation. A Carpel primorida emerge. B Carpel formed. C D Stage 6, microsporangia initiation. C Four microsporangia in itiate on abaxial side of stamens. D Enlarged image of microsporangia in C E Stage 7, ovule initiation: ovules initiates al ong the dorsal walls of locules. F G Stage 8, male meiosis. F microspore mother cells formed. G tetrads formed. H J Stage 9, female meiosis. H female meiosis occurred. I Enlarged image of H microspore formed. J nuclei (arrow) formed. K endosperm formed. L M Stage 10, anthesis. L The first day of flowering, stigma receptive. M The second day of flowering, stamens dehiscent. Cp: carpel, P: petal, S: sepal, St: stamen. Scale bars are as follows: A-G, I-J = 0.1 mm, H, K = 1 mm
61 Figure 3-2. Developmental series of Nuphar advena. A Stage 2, initiation sepal: first three sepals formed, and other three se pals initiated (SEM, top view). B E Stage 4, initiation of stamens. B Petal whorl formed, and st amen primordia initiated, longitudinal section. C Slightly older stage, cent ral region becomes dome-shaped (SEM, top view). D Older stage of stamen initiation: many stamen primordia formed (SEM, top view). E same stage as D, longitudinal section. F-H Stage 5, carpel initiation. F Flat carpel primordia fo rmed, longitudinal section. G Carpellary ridge formed, longitudinal section. H Carpellary ridge deepen (SEM, top view). I J Stage 6, microsporangia initiation. I Four microsporangia initiate on adaxial side of stamens, longitudinal section. J Enlarged image of I K Stage 7, ovule initiation: ovules initiates along the lateral and dorsal walls of locules, longitudinal section. L Stage 8, male meiosis: sporogenous tissue, longitudinal section. M-O Stage 9, female meiosis. M female meiosis initiated, longitudinal section. N O nuclei observed, longitudi nal section. Cp=carpel, ms=microsporangia, P=petal, S=sepal, St=s tamen. Scale bars are as follows: A-D J, N, O = 0.1 mm, E-I, K-M = 1 mm.
62 Figure 3-2. Continued.
63 Figure 3-3. Developmental series of Nymphaea odorata. All longitudinal sections except A (SEM, top view). A Stage 3, initiation of petals: the first four petals are initiated alternating with sepals. B D Stage 4, stamen initiation. B Stamen primordia initiated, central region is flat. C initiation of later stamens, central region is depressed. D floral apex bulges. E Stage 5, carpel initiation: carp el primodria (arrow) formed. F G Stage 6, microsporangia initiation. F four microsporangia form ed on adaxial sides of stamens. G Slightly older stage, carpel locule formation. H cross-sectioned image of microsporangia from stage 6. I Stage 7, ovule initiation: ovules initia tes along the lateral and dorsal walls of locules. J Stage 8, male meiosis: pollen sac is filled with sporogenous tissues. K, Stage 9, female meiosis: ovule anatropous, with megaspore. A=floral apex, Cp=carpel, ms=microsporangia, P=petal, S=sepal, St=stamen. Scale bars are as follows: A-C, H, L = 0.1 mm, D-G, I-K = 1 mm.
64 Figure 3-4. Developmental features of the third-whorled organs of Nuphar advena. A B Stage 4, stamen initiation. A stamen primordia are alternative with petal primordia (SEM, abaxial view). B Slightly older stage (SEM, abaxial view). C D Stage 5, carpel initiation. C a tip of stamens is projected (SEM, abaxial view). D two whorls of sepals (green dots) and two whorls of petals (red and yellow dots) were formed (SEM, abaxial view). E-O Stage 9, female meiosis. E The most outer whorl is composed of 16 petals and nectaries are on the abaxial side of petals. F, At antheis, view from the bottom, six sepals are removed. 18 petals formed at the third whorl, and they have nectaries on the abaxial surface. G Adaxial (right) and abaxial (left) view of stamens: four microsporagia formed in adaxial side. H Abaxial view of stamens: there is a ridge in the middle of stamen (SEM). I Adaxial (right) and ab axial (left) view of petals: there is a nectary on the top area in abaxial side. J Adaxial (right) and abaxial (left) view of petals with two microsporangia. K Adaxial (right) and abaxial (left) view of petals with four microsporangia. L Adaxial (right) and ab axial (left) view of petals with four fertile microsporangia. The tip is projected. M Abnormal petal at anthesis with adaxial (right) and abaxial (l eft) view. It has aborted microsporangia on the adaxial surface as well as nectaries on the abaxial surface. N Abnormal petal with aborted microsporangia (SEM, adaxial view). O Adaxial (right) and abaxial (left) view of petals at anthesis : two ridges are present. P Adaxial (two right) and abaxail (two left) views of stamens at anthesis: stamens are dehiscent and one ridge is present in the middle area. Scale bars are as follows: A-C, H = 0.1 mm D-G, I-P = 1 mm.
65 Figure 3-4. Continued.
66 CHAPTER 4 EXPRESSION OF HOMOLOGUES OF MADS-BOX GENES IN FLOWERS OF TWO DIVERGENT WATER LILI ES: THE BASAL ANGIOSPE RM NYMPHAEALES AND THE BASAL EUDICOT NELUMBO Introduction The basal angiosperm lineages are of particular in terest in studies of fl oral evolution due to their diversity in the arrangement and number of floral parts (e.g., Endress 1994, 2001; Soltis et al. 2002; Zanis et al. 2003). Ny mphaeales are the sister to all extant angiosperms except Amborella (Borsch et al. 2005; Hilu et al. 2003; Jansen et al. 2007; Lhne and Borsch 2005; Leebens-Mack et al. 2005; Mathews and Donoghue 1999; Moore et al. 20 07; Qiu et al. 1999; Qiu et al. 2005; Soltis et al. 1999; Soltis et al. 2000; Soltis et al. 2005; Zanis et al. 2002; Zanis et al. 2003). Members of Nymphaeales exhibit floral morphology that differs from that of the other basal lineages of angiosperms (i.e., Amborella Austrobaileyales) in se veral features, including the presence of large flowers in some members (i.e., Nuphar Nymphaea, Euryale, and Victoria ) rather than small or moderate-sized flowers, whor led rather than spiral phyllotaxis, eudicot-like perianth differentiation (i.e., Cabomba and Nuphar ), and the occurrence of blue perianth organs containing anthocyanins, which ar e absent in other basal angiosperms (Endress 2001; Schneider et al. 2003). Thus, studying floral development in Nymphaeales will help elucidate floral evolution in early-diverging angios perms and thus provide insight into the early diversification of the flower. Nymphaeales comprise nine genera and 80 aquatic species distributed in tropical to temperate regions around the world (Saarel a et al. 2007; Schneider and Williamson 1993; Sokoloff et al. 2008; Williamson and Schneider 1993). This clade is composed of three subclades: Cabombaceae, Nymphaeaceae, and Hydatellaceae. The former two subclades have long been recognized as two closely related families (Richard 1828; Williamson and Schneider
67 1993) that share many features, including vegetativ e, reproductive, and molecular characters. The remaining subclade, Hydatellaceae, is sister to Cabombaceae + Nymphaeaceae and was only recently placed in Nymphaeales based on molecular phylogenetics (Saarela et al. 2007). The placement of Hydatellaceae in Nymphaeales is further supported by the presence of a 4-celled embryo sac like that of other Nymphaeales (F riedman 2008). However, floral development differs among the three families. For example, Cabomba and Brasenia (Cabombaceae) are characterized by oligomerous flowers and simultane ous initiation of calyx and corolla, whereas Nymphaeaceae (Barclaya Euryale, Nuphar Nymphaea, Ondinea and Victoria ) have polymerous flowers and exhibit a unidirectional or der of perianth initiation (Endress 2001; Ito 1987; Les et al. 1999; Schneider et al. 2003). H ydatellaceae show unique floral developmental features: the presence of involucral bracts, no pe rianth, and separate staminate and carpellate flowers (Rudall et al. 2007). For this study, we focused on three genera of Nymphaeales, Cabomba, Nuphar which is sister to all other genera of Nymphaeaceae, and the phylogenetically derived Nymphaea (Fig. 1; Borsch et al. 2008). Cabomba has two petaloid whorls (similar to monocots) that have been variously described as undifferentiated tepa ls (Crow and Hellquist 2000; Judd et al. 2002; Zomlefer 1994) or a perianth differentiated into sepals (1st whorl) and petals (2nd whorl) (Endress 2001; Fassett 1953; Ito 1986, 1987; Moseley et al 1993; rgaard 1991; Schneider and Jeter 1982; Soltis et al. 2005; Wiersema and Hellquist 199 8; Zanis et al. 2003). The inne r perianth (petals) of Cabomba is distinguished by the auricular-shaped n ectaries at the side of each petal (Fig. 1 A ; Endress 2001; Schneider et al. 2003). In a ddition, the inner perianth (petals) of Cabomba shows retarded development (Endress 2001), as in eudicots. The petals remain small until late development in the floral bud when the outer perianth and stamens develop fully, and they
68 expand just prior to anthesis (Endress 2001; Schneider et al. 2003). This developmental retardation in petals is also observed in Nuphar (Endress 2001). The perianth of Nuphar is well differentiated into sepals (1st and 2nd whorls) and petals (the remaining perianth members) (e.g., Crow a nd Hellquist 2000; Endress 2001; Moseley 1965, 1972; Padgett et al. 1999; Schneider and Williams on 1993; Schneider et al. 2003; Wiersema and Hellquist 1998; Zanis et al. 2003; Zomlefer 1994). In fact, the two whorls of sepals show a difference in color: the outer sepals are green, and the inner sepals are yellow (Fig. 1 B ; Padgett et al. 1999; Warner et al. 2008). Therefore, several authors ha ve regarded the 1st whorl of Nuphar flowers as sepals or outer tepals and the 2nd whorl as petals or inner tepals (Judd et al. 2002; Kim et al. 2005; Warner et al. 2008). The remaini ng perianth members are then regarded as staminodes (Judd et al. 2002; Kim et al. 2005; Warner et al. 2008). Nymphaea has attracted the attention of many botan ists because of its distinctive floral morphology. It is a classic example of a gradua l transition among adjacent organ types with a complete range from petaloid staminodes to functional stamens (Crow and Hellquist 2000; Judd et al. 2002; Schneider and Williamson 1993; Wier sema and Hellquist 1998; Zomlefer 1994). Typically, the innermost stamens are functional, but the androecium shows a gradual transition from inner functional stamens toward outer petalo id perianth members; the apical portion of the anther becomes smaller, and the laminar filament area broadens toward the perianth members (Fig.1C D ). The outermost stamens, also referred to as petaloid staminodes, could be derived from either petals or stamens, but their origins have not yet been studied in detail. Furthermore, the perianth of Nymphaea has been variously interpreted as undifferentiated (Doyle and Endress 2000; Llamas 2003; Soltis et al. 2005) or slightly differentia ted into sepals (1st whorl) and petals (the remaining perianth members) (Crow a nd Hellquist 2000; Ito 1987; Judd et al. 2002;
69 Schneider and Williamson 1993; Schneid er et al. 2003; Wiersema and Hellquist 1998; Zanis et al. 2003; Zomlefer 1994). Recently, Warner et al. (2008) suggested that the perianth organs of Nymphaeales are differentiated, but these plants l ack the typical sepals and peta ls of eudicots based on SEM examination. They showed that both sepaloid and petaloid areas are present on individual perianth members, and sepals and petals only di ffer in the occurrence of trichomes on the abaxial side of the sepals. Thus, Warner et al. (2008) suggest the term tepal is more appropriate for the perianth organs of Nymphaeales because the peri anth members of other basal angiosperms, such as Amborella and Austrobaileyales, are considered to represent tepals (i.e., Endress 2001; Buzgo et al. 2004b). However, in this study we use the terms sepals and petals for Cabomba Nymphaea, and Nuphar because the perianth organs are, in fact, differentiated (Cro w and Hellquist 2000; Endress 2001; Fassett 1953; Ito 1986, 1987; Mo seley 1965, 1972; rgaard 1991; Schneider and Jeter 1982; Schneider and Williamson 1993; Schneid er et al. 2003; Wiersema 1988; Wiersema and Hellquist 1998; Zanis et al. 2003; Zomlefer 1994). Nelumbo (Nelumbonaceae) was historically thought to have a close relationship with Nymphaeales due to a superficia l similarity in overall floral morphology and habitat (Cronquist 1988; Heywood 1993). However, molecular data have shown that Nelumbo is an early-diverging eudicot related to Platanaceae and Proteaceae (APG 1998; APG II 2003; Chase et al. 1993; Soltis et al. 2000; Worberg et al. 2007), and some morphological features also support this placement (Drinnan et al. 1994; Hoot et al 1999). Although the floral similarity between the distantly related Nelumbo and Nymphaeales is an example of convergence rather than homology, this similarity is of interest with resp ect to floral developmental genetics. Nelumbo has a spirally arranged, undifferentiated peri anth, and the innermost perian th members have a staminal
70 appendage (Hayes et al. 2000; Vogel and Hadacek 2004). In Nelumbo, the first two perianth parts (sepals) are greenish and smaller than the other perianth parts (pe tals). The abaxial side of the outermost petals is greenish white, while the inner petals are pinkish white (Hayes et al. 2000; Vogel and Hadacek 2004). Unlike Nymphaea, the sepals of Nelumbo are smaller than the petals, and have different developmental features For example, two sepal primordia enclose the floral apex and the sepals become hood-shaped with two basal auricles at the sides at early developmental stages (Hayes et al. 2000). However, at anthesis the difference between sepals and petals in Nelumbo is unclear except for organ size. The innermost petals of Nelumbo have a staminal appendage, but no anthers (M.-J. Y oo, personal observation), a feature somewhat similar to the petaloid staminodes of Nymphaea. In flowering plants, many MADS-box gene s are involved in the well-known ABCDE model of floral organ identity. In the model organism Arabidopsis thaliana A and E class genes control sepal identity; A, B, and E class genes control petal identity; B, C, and E class genes control stamen identity; C and E class genes control carpel identi ty; D and E class genes control ovule identity (Coen and Meyerowitz 1991; Colo mbo et al. 1995; Pelaz et al. 2000; Theissen 2001a). At least some aspects of the ABCDE mo del are conserved across angiosperms (i.e., Petunia hybrida (Angenent et al. 1993); Silene latifolia (Hardenack et al. 1994); Gerbera hybrida (Yu et al. 1999); Oryza sativa (Fornara et al. 2003; Kater et al. 2006; Kyozuka et al. 2000); Pisum sativum (Taylor et al. 2002); Zea mays (Whipple et al. 2004); Magnolia grandiflora (Kim et al. 2005), Persea americana (Chanderbali et al. 2006), Akebia trifoliata (Shan et al. 2006); Elaeis guineensis (Adam et al. 2007); Taihangia rupestris (L et al. 2007); Vitis vinifera (Poupin et al. 2007)). However, other features are not c onserved throughout flowering plants, with important variations observed, especially in basal eudicots and basal
71 angiosperms (i.e., Kim et al. 2005; Kramer et al. 2003; van Tunen et al. 1993; Kanno et al. 2003). For whorled sepaloid or petaloid perianths, the shifting boundaries model or the sliding boundaries model was introduced to explain sepaloidy or petaloidy in some taxa by shifting the B class gene expression area to the outer whorl (Bowman 1997; Kramer et al. 2003). Recently, Buzgo et al. (2004, 2005) proposed the fading borders model to explain morphological intergradation of floral organs in the basal angi osperms. According to this model, each floral organ identity gene is broadly expressed across adjacent floral organs, but only weakly expressed at the edges (borders) of its zone of activity (Buzgo et al. 2 004; Buzgo et al. 2005). The fading borders model is supported by gene expression studies in several basal angiosperms (Kim et al. 2005). We investigated the molecular determinati on of floral development in Nymphaeales. Specifically, we identified homologues of we ll-known floral organ identity genes and investigated their expression patterns across an array of Nymphaeales with varying floral morphologies ( Cabomba, Nuphar and Nymphaea) and in the basal eudicot Nelumbo. We focused on the following three issues: (1) patterns of floral gene expression and the transitions from outer perianth to inner perianth to petaloid staminodes to stamens in Nymphaea; (2) gene expression and perianth differentiation in Cabomba, Nuphar, and Nymphaea; and (3) gene expression patterns and morphological convergence in Nelumbo and Nymphaea. We surveyed expression profiles of several MADS-box gene homologues of the A, B, C, D, and E classes and the GGM13 ( Bsister) and AGL6 clades (Becker and Theissen 2003; MartinezCastilla and Alvarez-Buylla 2003; Nam et al. 2004; Theissen et al. 1996). As for a homologue of APETALA3 ( AP3), Stellari et al. (2004) repor ted alternative splicing of the Nymphaea AP3 transcript. They found fo ur different kinds of Nymphaea AP3 transcript and suggested that there
72 might be diverse regulatory mechanisms in AP3 across divergent taxa (Stellari et al. 2004). In our preliminary experiments, we also observed several bands corresponding to AP3, and therefore designed a specific primer set to invest igate the expression pattern of specific kinds of AP3 transcripts. We also compared our results here with expression patterns from other basal angiosperms including Amborella and Persea (Chanderbali et al. 2006; Kim et al. 2005; Zahn et al. 2005). This study will shed additional light on floral developmental genetics in early angiosperms. Material and Methods Plants We collected samples from the following sources: Cabomba caroliniana plants purchased from a local aquarium store, Ga inesville, FL, USA (Yoo 10020, FLAS); Nuphar advena, Waccassasa River, near Waccassasa Preserve Area, Levy Co., FL, USA (Yoo & Koh 1000, FLAS); Nymphaea odorata, Waccassasa Preserve Area, Levy Co., FL, USA (Yoo & Koh 1001, FLAS); Nymphaea capensis plants purchased from a local aquarium store, Gainesville, FL, USA (Yoo 10021, FLAS); and Nelumbo nucifera cultivated at the Kanapaha Botanical Garden, Gainesville, FL, USA (Yoo & Koh 1002, FLAS). Floral buds or flowers from early developmental stages to anthesis were collected and preserved in liquid n itrogen and stored at -80 RNA Extraction, RT-PCR, and Screening for Homologues of MADS-Box Genes We extracted RNA from whole floral buds using the RNeasy Plant Mini Kit (Qiagen, Stanford, CA, USA). The modified method of Kim et al. (2004) was used for RNA isolation from Nuphar advena. This method, which consists of two parts, a CTAB DNA extraction protocol (Doyle and Doyle 1987) and subsequent use of the RNeasy Plant Mini Kit, increased
73 the amount of isolated RNA. Reverse tr anscription was perf ormed by following the manufacturers directions usi ng Super-Script II RNase H-reve rse transcriptase (Invitrogen, Carlsbad, CA, USA) and polyT primer (5-C CG GAT CCT CTA GAG CGG CCG C(T)17-3). PCR reactions were performed using MADS gene-specific degenerate primers (5-GGG GTA CCA AYM GIC ARG TIA CIT AY T CIA AGM GIM G-3) and the polyT primer used in reverse transcription (Kramer et al. 1998). PCR conditions were those employed by Kramer et al. (1998). PCR bands over 800 bp in size were excise d from agarose gels and purified using the Geneclean Turbo Kit (QBio-Gene, Carlsbad, CA USA). Purified DNAs were cloned using the TOPO TA Cloning Kit (I nvitrogen), and plasmid DNAs were purified from cloned cells through the FastPlasmid Mini Kit (Eppendorf, Westbur y, NY, USA). Sequences were determined by cycle sequencing reactions using the CEQ DTCS -Quick Start Kit (Beck man Coulter, Fullerton, CA, USA). Phylogenetic Analysis for Sequence Identification To determine the putative identity of newly obtained gene sequences, we applied a BLAST search followed by phylogenetic analysis. To assign our putative MADS-box genes to the appropriate subfamily, we added our sequences to a large data set of sequences representing major subfamilies of MIKCc-type MADS genes (Becker and Theissen 2003) and recently obtained sequences from basal angiosperms (K im et al. 2005). A total of 186 amino acid sequences was aligned using CLUSTAL X (ver. 1.83) (Thompson et al. 1997) with manual adjustments by eye. Maximum parsimony an alysis was performed using PAUP* 4.0b10 (Swofford 2002). The search strategy involved 10 random addition replicates with TBR branch swapping, saving all optimal trees. To assess support for each node, a bootstrap analysis
74 (Felsenstein 1985) was conducted using 100 bootstrap replicates, each with 10 random addition replicates and TBR branch swapping, saving all optimal trees. Putative homologues of Arabidopsis ABCDE genes were denoted using the first two letters of the genus name and the first two letters of th e species epithet of the species from which the sequence was isolated, followed by the Arabidopsis gene abbreviation and a number to indicate the gene copy detected, if more than one; for example, the APETALA1 ( AP1) homologue of Cabomba caroliniana is Caca.AP1. Gene Expression Based on RQ-RT-PCR To investigate the expression pa tterns of homologues of A-, B-, C-, D-, and E-class genes and GGM13 ( Bsister) and AGL6 lineages in floral organs, we employed relative quantitative reverse transcriptasepolymerase chain reacti on (RQ-RT-PCR) because of its reliability in studies of floral genes (i.e., Adam et al. 2007; Chanderbali et al. 2006; Kim et al. 2005; Kramer et al. 1998; Kramer et al. 2003). For RQ-RT-PCR, we dissected each floral orga n from buds collected just before anthesis for all taxa except Nymphaea odorata For N. odorata, we used two different developmental stages, an early developmental stage (female-pr emeiotic; prior to stage 9; Buzgo et al. 2004a) and flowers at anthesis (stage 10; Buzgo et al 2004a), to trace cha nges in gene expression patterns. Young leaves were used for comparison with perianth parts and reproductive organs. For Cabomba two leaf tissues (submerged and floati ng leaves) were included. Floating leaves are produced by flowering shoots an d are axillary to the inflorescen ce. In contrast to petiolate and diand trichotomously dissected submerged l eaves, floating leaves are peltate and the lamina is entire (Fig. 4-1 A ) (Fassett 1953; rgaard 1991)
75 Total RNAs were extracted from each organ sample using the RNeasy Plant Mini Kit (Qiagen). Extracted RNA was treated with DNase to remove potential contamination by genomic DNA (DNase-free kit from Ambion, Austin, TX, USA). Reverse transcription using RNA from each floral part was performed following the ma nufacturers directions using SuperScriptTM II RNase H-reverse transcriptase (Invitrogen). Fo r reverse transcripti on, we used a randomhexamer instead of the polyT primer because the 18S ribosomal RNA gene was used as an internal control. RQ-RT-PCR was performed using a gene-specific primer pair (Table 4-1), the 18S rRNA gene primer pair (internal control), and a competitive primer pair to the 18S rRNA gene primers (competimers) following the protocol of QuantumRNA (Ambion). For AP3 homologues from Nymphaea we designed a specific primer set, AP3-4 and AP3-5, to reduce the number of bands. The reverse primer, AP 3-4, is from the sequence of exon 7 of Nyod.AP3, and the forward primer, AP3-5, is designed from seque nces of exons 5 and 6 (Fig. 4-2). Using this primer set, two classes of AP3 transcript, class I and class VI, were detectable (Fig. 4-2). The optimal ratio of the 18S primer pair to competim ers was tested for each gene to obtain a similar level of PCR signal between the 18S rRNA and that of each gene. The optimal ratio ranged from 3:7 to 4:6 for the genes that we surveyed. PCR r eactions for all genes were performed with 27 or 28 cycles at 95 (30 sec), 56 (30 sec), and 72 (30 sec) using an Eppendorf Mastercycler (Brinkmann, Westbury, NY, USA) with 50 ng of total cDNA template. Twenty from each PCR reaction were run in a 2% (w/v) agarose gel containing 10-4 % (w/v) ethidium bromide in TAE buffer. Gel images were analyzed us ing KODAK 1D Image Analysis Software (Kodak, Rochester, NY, USA). At least three replicates of RQ-RT-PCR were performed for each gene, and the relative PCR intensity of the specific gene to that of the 18S rRNA gene was calculated for each gene. The identity of each PCR product was confirmed by sequencing. To compare
76 relative gene expression levels among the fl oral organs of each taxon, we applied the quantification method of Kim et al. (2005). Expression levels were de signated as follows: -, not expressed; +, <0.1; ++, 0.1-0.4; +++, 0.4-1 (Kim et al. 2005). The genes with +++ were highly expressed relative to the internal control in every PCR reaction. The genes with ++ were consistently expressed in all PCR reactions, bu t their expression level was relatively low. The genes with + were variable in their gene expression levels, rangi ng from very weak expression to no expression in some PCR reactions. Results Homologues of MADS-Box Genes in Nymphaeales We report here the following homologue s of MADS-box genes in Nymphaeales: homologues of APETALA1 ( AP1), Caca.AP1 from Cabomba caroliniana Nenu.AP1 from Nelumbo nucifera ; homologues of APETALA3 ( AP3), Caca.AP3, Nyod.AP3 from Nymphaea odorata Nyca.AP3 from Nymphaea capensis Nenu.AP3-1 Nenu.AP3-2 from Nelumbo nucifera ; homologues of PISTILLATA ( PI ), Caca.PI Nyod.PI Nyca.PI Nenu.PI ; homologues of AGAMOUS ( AG ), Caca.AG Nyod.AG1-1 Nyod.AG1-2 Nyod.AG2 Nyca.AG1 Nyca.AG2 Nenu.AG ; homologues of AGAMOUS-LIKE11 ( AGL11 ), Nyod.AG3 Nyca.AG3 ; homologues of AGAMOUS-LIKE2/4/9 ( AGL2/4/9 ; SEP1/2/3 ), Caca.AGL2-1( Caca.SEP1-1 ), Caca.AGL22( Caca.SEP1-2 ), Caca.AGL2-3 ( Caca.SEP1-3 ), Nuad.AGL4 ( Nuad.SEP2), Nuad.AGL 9 ( Nuad.SEP3) from Nuphar advena, Nyod.AGL2 ( Nyod.SEP1 ), Nenu.AGL2 ( Nenu.SEP1 ), Nenu.AGL9 ( Nenu.SEP3 ); homologues of AGAMOUS-LIKE6 ( AGL6 ), Caca.AGL6-1, Caca.AGL6-2, Caca.AGL6-3, Nuad.AGL6 Nyod.AGL6, Nenu.AGL6; and homologues of Bsister ( Bs), Caca.Bs1 and Caca.Bs2.
77 Gene annotation is based on our phylogenetic analysis, and the subf amilial placement of each gene was determined based on >70% bootst rap support (Fig. 4-3). From this study, we report a newly identif ied orthologue of AP1 (A function in A. thaliana ) from Cabomba and Nelumbo, orthologues of AP3 and PI (B function in A. thaliana) from Cabomba Nymphaea and Nelumbo, orthologues of AG (C function in A. thaliana ) from Cabomba, Nymphaea, and Nelumbo, an orthologue of AGL11 (D function in A. thaliana) from Nymphaea orthologues of AGL2/4/9 (E function in A. thaliana) from Cabomba Nuphar Nymphaea, and Nelumbo orthologues of AGL6 (tepals, carpels, and ovule development in some angiosperms) from Cabomba, Nuphar Nymphaea, and Nelumbo, and orthologues of Bsister (ovule and female gametophyte development in A. thaliana Petunia and Gnetum ) from Cabomba. Analysis of Expression Profile by RQ-RT-PCR In Cabomba, Caca.AP1 was expressed in all floral organs and also leaves Expression of Caca.PI was detected in sepals, petals, and stamens, and Caca.AG was expressed in stamens and carpels. Caca.AP3, however, was expressed in all floral organs and in all leaves, with the strongest expression in sepals, petals and stamens. The two homologues of Bsister genes from Cabomba were expressed in carpels only. In addition, the three AGL2 homologues, Caca.AGL21, Caca.AGL2-2, and Caca.AGL2-3 (Fig. 4-3) exhibited almost identical gene expression patterns, with strong signal in all floral organs, although Caca.AGL2-3 was slightly expressed in floating leaves (Fig. 4-4 A ). The three AGL6 homologues were expresse d in different tissues. Caca.AGL6-1 and Caca.AGL6-2 were detected in all floral organs, although the expression levels of Caca.AGL6-1 were very low in stamens and carpels. Caca.AGL6-3 was expressed in carpels only (Fig. 4-4A ).
78 For Nuphar Nuad.AP1 was strongly expressed in sepa ls, carpels, and leaves, and transcripts of Nuad.AG were detected in petals, stam ens, and carpels. Signals of both Nuad.AGL4 and Nuad.AGL9 were detected in al l floral organs, and Nuad.AGL6 was expressed in sepals and carpels (Fig. 4-4 B ). The gene expression patterns of sepals (1st whorl) and outermost petals (2nd whorl) of Nymphaea were very similar for most of th e genes we investigated (Fig. 4-4 C ; Tables 4-2, 3) at both developmental stages studie d. However, two paralogues of Nyod.AG1 and Nyod.AG3 showed different gene expression patterns betw een sepals and outermo st petals (Fig. 4-4C ). Nyod.AG1-1 and Nyod.AG1-2 were expressed in both of these organs at an early developmental stage although Nyod.AG1-1 showed relatively low expression level in petals relative to that of sepals. However, at anthesis, neither Nyod.AG1-1 nor Nyod.AG1-2 was detected in sepals. Nyod.AG3 was weakly expressed in sepals, but not in the outermost petals at the early developmental stage (Fig. 4-4 C ). At both developmental stages investigated in N. odorata, petals and petaloid staminodes had identical expression patte rns for all genes (Fig. 4-4 C ; Table 4-2), although there are differences in expression levels. For example, the expression level of the AP3 transcript class I was higher in petals than in staminodes. Th ere is no difference in expression for genes investigated here between outermost and innermost stamens (Fig. 4-4 C ; Table 4-2). Three AG homologues were identified from Nymphaea odorata (Fig. 4-3). Both Nyod.AG1-1 and Nyod.AG1-2 are expressed in all floral organs at the early developmental stage, but their expression patterns are narrower at anth esis (i.e., no expression in sepals, carpels, and ovules). Nyod.AG2 was expressed in stamens, carpels, a nd ovules at the early developmental stage investigated, and the expr ession level was relatively high in carpels. However, its
79 expression area changed at anthesis. The expression of Nyod.AG2 was only detected in innermost petals and petaloid staminodes, but the expression level was relatively low in innermost petals (Fig. 4-4 C ; Table 4-2). Based on this expression pattern, Nyod.AG2 may be a functional equivalent of AG of Arabidopsis In contrast to the differe nt gene expression patterns of the AG homologues of N. odorata, two AG homologues of N. capensis Nyca.AG1 and Nyca.AG2 show similar gene expression patterns to each other. They were expressed in all floral organs except sepals although Nyca.AG1 was very weakly expressed in sepals (Fig. 4-4 D ). Expression of the AGL11 homologues, Nyod.AG3 and Nyca.AG3 was observed in ovules while Nyod.AG3 was also expressed in sepals at the early developmental stage investigated (Fig. 4-4 C D ). In addition to the four kinds of AP3 transcripts previously reported from Nymphaea (Stellari et al. 2004), we found two more classes, classes V and VI (Fig. 4-2). We examined the expression patterns of two classes of Nymphaea AP3 transcript, class I an d class VI, which were detected with the AP3-4 and AP3-5 primer set (F ig. 4-2). These two tran scripts show different gene expression patterns. The class I Nyod.AP3 transcript was expressed in all floral organs and leaves at the early developmental stage studied although the expression level in staminodes and ovules was relatively low. However, at anthesis, expression was not observe d in carpels (ovules) or leaves. The class VI Nyod.AP3 transcript was detected in stam ens and carpels (ovules) only at the early developmental stage. However, at anthesis this transcript was found in all floral organs except carpels (ovules). In Nelumbo, gene expression patterns differed between the two sepals and the outermost petals, especially for the two AP3 homologues, which belong to euAP3 lineage (Fig. 4-4 E ). The outermost and inner petals exhibited the same expression profiles (Fig. 4-4 E ). Nenu.AP1 was
80 expressed in all floral organs and leaves, and Nenu.AGL2 and Nenu.AGL9 were detected in all floral organs. Nenu.AGL6 was also expressed in all floral organs, but the expression level was higher in sepals, petals, and car pels than in stamens and the staminal appendage (Fig. 4-4 E ). The AG homologue, Nenu.AG was strongly expressed in stam ens and carpels only. Expression of Nenu.PI was observed in all floral organs except carpels. Two AP3 homologues showed different expression patterns: Nenu.AP3-1 expression was detected in pe tals, the staminal appendage, stamens, and carpels, whereas Nenu.AP3-2 was expressed in petals, the staminal appendage, and stamens (Fig. 4-4E ). Discussion Comparison and Implication of Expression Patterns of MADS-Box Genes in Nymphaeales To elucidate the molecular determination of floral development in Nymphaeales, we investigated the expression patte rns of homologues of floral ogan identity (MADS-box) genes from across a diverse array of Nymphaeales: Cabomba, Nuphar and Nymphaea. Most of the floral organ identity gene homologues detected in Nymphaeales are expressed in all floral organs (Fig. 4-4, Tables 4-1, 2). The expression pattern in each genus of Nymphaeales examined here differs, however, perhaps in correlation with their morphological differences. For example, AP3/ PI homologues are similarly expresse d in all floral organs in Nuphar and Nymphaea. However, in Cabomba, the PI homologue is not expressed in car pels but in sepals, petals, and stamens, while the AP3 homologue is additionally expressed in leaves and carpels at relatively low levels. This expression patt ern of B-class homologues from Cabomba can be explained by the shifting boundary model (Bowman 1997) or th e sliding boundaries model (Kramer et al. 2003). In this model, when B-class gene expressi on is expanded to incl ude the entir e perianth, only a petaloid perianth results. In Cabomba, petals differ from sepals in having retarded
81 development, and in the presence of nectar ies and the absence of trichomes (Endress 2001; Schneider et al. 2003; Warner et al. 2008). However, their overal l morphologies are quite similar to each other, much like the petaloid tepals of many monocots. Thus, the shared expression profile observed for PI and AP3 homologues in both perianth whorls of Cabomba is reasonable based on their similar petaloid appearance (i.e., Tulipa Kanno et al. 2003; va n Tunen et al. 1993; Persea Chanderbali et al. 2006). Transcripts of C-cl ass homologues from Cabomba are observed in stamens and carpels only (Fig. 4-4A Table 4-2), in agreement with data fo r eudicot models and monocots (Davies et al. 1999; Kang et al. 1995; Kater et al. 1998; Kempin et al. 1993; Schmidt et al. 1993; Yanofsky et al. 1990). A similar pattern of AG homologue expression was observed in Nymphaea (see below) and other basal angiosperms, including Amborella and magnoliids (Chanderbali et al. 2006; Kim et al. 2005). These results imply that AG homologues have a conserved function throughout the angiosperms. However, the AG homologue of Nuphar is also expressed in petals (Fig. 4-4B ); one AG homologue is also e xpressed in tepals of Persea (Chanderbali et al. 2006) and in inner tepals of Illicium (Kim et al. 2005). Th e expression patterns of Nuad.AG will be discussed further below. We identified three AG homologues in Nymphaea, and Nyod.AG2 is thought to be a functional equivalent of AG based on the expression profile we observed in early development (Fig. 4-4C ; Table 4-2). Two other AG homologues, Nyod.AG1-1 and Nyod.AG1-2 are expressed in all floral organs earl y in development (Fig. 4-4 C ; Table 4-2). In addition to their different gene expression patterns, these three AG homologues are included in different subclades: Nyod.AG1-1 and Nyod.AG1-2 form a clade with AG homologues from other Nymphaea species, while Nyod.AG2 is clustered separately in the C lineage (F ig.4-3). As a result, th ere are two subclades
82 of AG homologues in Nymphaea and two AG homologues identified from N. capensis Nyca.AG1 and Nyca.AG2 are also included in each subclade However, we could not identify a second AG homologue from Nuphar or Cabomba indicating a duplication event might have occurred only in Nymphaea, although we examined only two species of Nymphaea Therefore, to determine whether a duplication event in the C lineage of the AG subfamily is specific to Nymphaea (or to Nymphaea and other Nymphaeaceae not sampled here) or occurred earlier in the history of Nymphaeales, a more comprehensiv e search of representative species of each genus is required. Because Nyod.AG1 and Nyod.AG2 are differentially expressed, there may be functional differentiation between these AG homologues in Nymphaea. To investigate this possibility, we examined the expression patterns of Nyca.AG1 and Nyca.AG2 : these two AG homologues are expressed in all floral orga ns except sepals (Fig. 4-4 D ), implying there is no differentiation between the two copies at the st age of development studied in N. capensis Thus, to investigate expression differentiation among duplicate AG homologues in N. odorata, we need to exploit other methods, for example, in situ hybridization throughout vari ous developmental stages. AGL11 homologues (D lineage) from two Nymphaea species ( Nyod.AG3 from N. odorata and Nyca.AG3 from N. capensis ) and two Bsister homologues from Cabomba ( Caca.Bs1 and Caca.Bs2) were restricted in their expr ession to ovules only (Fig. 4-4A C D ). Both D lineage and Bsister genes are known to specify ovule or female gametophyte identities (Becker et al. 2002; de Folter et al. 2006; Nesi et al 2002; Rounsley et al. 1995; Tzeng et al. 2002), so the expression profiles of these three genes from Nymphaeales agree with these functions and further suggest that their function may be conser ved throughout the flowering plants.
83 Three remaining genes, SQUA ( AP1 ), AGL2 ( SEP1), and AGL6, are phylogenetically close (Becker and Theissen 2003; Kim et al. unpublished; Martinez-Castilla an d Alvarez-Buylla 2003), but their expression profil es are quite diverse. AGL2 homologues from Cabomba, Nuphar and Nymphaea are expressed in all fl oral organs, similar to AGL2 genes in other basal angiosperms and eudicot model plants (Cha nderbali et al. 2006; Honma and Goto 2001; Kim et al. 2005; Pelaz et al. 2000; Zahn et al. 2005), suggesting that these genes may be conserved in their expression patterns, and possibly their f unction throughout the flowering plants. For AP1 homologues, functional studies ar e very limited. To date, only AP1 of Arabidopsis and MtPIM ( AP1 orthologue) of Medicago truncatula exhibit a true A function (Benlloch et al. 2006; Bowman et al. 1993; Irish and Susse x 1990; Litt and Irish 2003). In Antirrhinum although the expression pattern of the AP1 homologue SQUA is the same as that of AP1, mutant analysis failed to exhibit the A function of SQUA (Davies et al. 2006; Huijser et al. 1992; Taylor et al. 2002). In fact, recent study of the ap1 mutant of Arabidopsis suggests that AP1 function is not essential for sepal and petal development: over-expression of AGL24 seems to sufficient to mask the defects of sepals and petals in the ap1-1 mutant, thus, in the absence of AGL24 mutants partially recover their wild-type phenotypes (Yu et al. 2004). Therefore, together with the expression patterns of AP1 homologues from other eudicots such as Petunia (Rijpkema et al. 2006), Antirrhinum (Davies et al. 2006), and Gerbera (Teeri et al. 2006), true A function of AP1 is questionable. AP1 homologues from Cabomba and Nuphar also displayed differe nt expression patterns. In Cabomba, the AP1 homologue is expressed in all floral organs and leaf tissues (Fig. 4-4 A ), while the transcrip t of the AP1 homologue of Nuphar is detected in sepals, carpels, and leaves (Fig. 4-4B ). Kim et al. (2005) earlier suggested that AP1 homologues may have different
84 functions in basal angiosperm taxa than in eudicots based on the broader patterns of gene expression in these basal lineages; our da ta further support this idea. In fact, AP1 is a product of gene duplication (Litt and Irish 2003): basal angi osperms and early-diverging eudicots have only FUL-like genes while core eudicots contain euAP1 and euFUL genes due to a duplication event near the base of the core eudicots. Interestingly, this duplication event in the AP1 lineage is apparently coupled with the orig in of true sepals, although seve ral studies have questioned the role of AP1 homologues in conferring true A functi on gene (see above). Al bert et al. (1998) estimated phylogenetic timing of the sepal/petal distinction using the phy logeny of Chase et al. (1993). Through optimization analyses, Albert et al. (1998) found that the characteristic sepaloid/petaloid bipartite state of Arabidopsis may have become fixed in the ancestral core eudicots. Thus, euAP1 might have gained a new function in sepal formation after duplication of the FUL-like gene. This implies that basal angiospe rms may not have a true A-function gene because this function did not evolve until after the duplication that yielded euAP1 and euFUL genes. However, a FUL-like gene or its relatives may be i nvolved in perianth development. FULlike genes are expressed in all fl oral organs and leaf tissue in basal angiosperms, such as Cabomba and Magnolia (Kim et al. 2005). AGL6 homologues are known to be expressed in inflorescence buds, tepals, carpels, and ovules of eudicots (Fan et al. 2007; Hsu et al. 2003; Ma et al 1991; Rounsley et al. 1995). For Nymphaeales, similar patterns of expre ssion are observed. For example, in Nuphar the AGL6 homologue is expressed in sepals (1st and 2nd whorls) and carpels onl y. Interestingly, this expression pattern is exactly the same as that obtained for the AP1 homologue of Nuphar (Fig. 44B ). In contrast, the Nymphaea AGL6 homologue is detected in sepals, petals, petaloid staminodes, and ovules in early development. In Cabomba, three AGL6 homologues were
85 identified (Fig. 4-3), and their gene expression patterns differ: Caca.AGL6-1 and Caca.AGL6-2 are expressed in all floral organs, although Caca.AGL6-1 is very weakly expressed in stamens and carpels, and Caca.AGL6-3 is detected in carpels only (Fig. 4-4 A ). These expression patterns indicate the possibility of subfunctionalizati on after duplication of AGL6 homologues in Cabomba, as in Arabidopsis (Duarte et al. 2006). In Arabidopsis there are two paralogues in the AGL6 lineage, AGL6 and AGL13 (Martinez-Castilla and Alvarez-Buylla 2003; Rounsley et al. 1995): AGL6 is expressed in inflorescence buds (Ma et al. 1991; Rounsley et al. 1995), whereas AGL13 is expressed in ovules only (Rounsley et al. 1995). Before the duplication event, one copy of AGL6 might have played a role in both pe rianth and ovule development. After the duplication event, however the two copies of AGL6 in Cabomba may have partitioned the function of the ancestral gene; one copy functions in perianth development, and the other functions in ovule development. Therefore, the complement of both copies can represent the functional capability of the ancestr al gene (Lynch and Conery 2000). These gene expression patterns suggest a possible role of AGL6 homologues in perianth and ovule development in Nymphaeales; these role s may be present in all flowering plants (Chanderbali et al. 2006; Fan et al. 2007; Hsu et al. 2003; Kim et al. unpublished; Rounsley et al. 1995). Also, some AGL6 homologues from basal angiosperms are expressed in perianth organs only (e. g., Magnolia Kim et al. 2005, Liriodendron, Kim et al. unpublished, and Persea Chanderbali et al. 2006 ), indicating that AGL6 homologues may be candidates for genes involved in perianth developm ent in basal angiosperms. The expression patterns of Nymphaea AGL2 AGL6 PI and AG homologues early in development are more similar to those observed for Nuphar floral gene homologues than those from Cabomba This is reasonable considering the closer phylogeneti c relationship of Nymphaea
86 and Nuphar and their floral similarities. For ex ample, A and B homologues of all three Nymphaeales examined here show broader gene expression patterns compared to those reported from eudicot model plants (Angenent et al. 1993; Davies et al. 1999; Goto and Meyerowitz 1994; Jack et al. 1992; Jack et al. 1994; SchwarzSommer et al. 1992; Trbne r et al. 1992; Yu et al. 1999). However, the expression patterns of C, D, E, and Bsister homologues are almost the same as those reported from eudicot models (Col ombo et al. 1995; Davies et al. 1999; de Folter et al. 2006; Honma and Goto 2001; Nesi et al. 2002; Pelaz et al. 2000 ; Tzeng et al. 2002; Yanofsky et al. 1990). In particular, the expressi on patterns of homologues involved in female gametophyte development are conserved in those few angiosperms and gymnosperms investigated to date (Becker et al. 2002; de Folter et al 2006; Zhang et al. 2004). Gene expression patterns observed in Nymphaeales furthe r support the conservation of function of C, D, and GGM13 homologues in all seed plants. Perianth Differentiation in Nymphaeales and Tr ansition of Petaloid Staminodes to Stamens in Nymphaea In Nymphaea the sepals and outermost petals have similar gene expression patterns at both developmental stages examined, excep t for genes that belong to the AG subfamily (Fig. 4-4C ; Table 4-1). Two AG homologues, Nyod.AG1-1 and Nyod.AG1-2 are expressed in all floral organs early in development, but their expression is not detected in sepals at anthesis (Fig. 4-4 C ; Table 4-1). Also, Nyod.AG2 a putative functional equivalent of AG of Arabidopsis is expressed in the innermost petals but not the rest of the perianth of Nymphaea at anthesis. Nyod.AG3 an AGL11 homologue in Nymphaea is expressed in sepals but not outer perianth members early in development. As a result, sepals, outermost petals, and innermost petals of Nymphaea differ slightly in their floral gene expression patterns at both develo pmental stages (Table 4-1). In
87 accordance with previous research (Crow and Hellquist 2000; Ito 1987; Judd et al. 2002; Schneider and Williamson 1993; Schneid er et al. 2003; Wiersema and Hellquist 1998; Zanis et al. 2003; Zomlefer 1994), Warner et al. ( 2008) noted that perianth parts of Nymphaea are morphologically and anatomically slightly diffe rentiated (i.e., by pres ence or absence of trichomes and papillate cells on the surf ace). However, sepals and petals of Nymphaea lack the typical sepal and petal charac teristics of eudicots (Warner et al. 2008). In eudicots, A and B class genes are responsible for pe rianth differentiation, for example, A class gene is required for sepal identity and B class gene for petal id entity (Coen and Meyerowitz 1991; Goto and Meyerowitz 1994; Jack et al. 1992; Schwarz-Sommer et al. 1990). However, in Nymphaea A and B class genes show the same expression patterns across all perianth parts. In addition, AGL6 homologues, potentially involved in perianth development, are al so expressed in all perianth parts. Hence, the absence of a well-differentiate d perianth in Nymphaeal es coupled with broad expression of floral organ identity genes agrees w ith data for other basal angiosperms (Kim et al. 2005). The slight perianth differentiation seen in Nymphaea might be due to the differences detected in gene expression patterns of AG homologues. Typically, the innermost stamens of Nymphaea are functional, but stamens show a gradual transition toward the perianth members; the apical portion of the anther is smaller, and the laminar filament area gets broader toward th e perianth members (Fi g. 4-1; Schneider and Williamson 1993; Zomlefer 1994; Wiersema and Hellquist 1998; Crow and Hellquist 2000; Judd et al. 2002). In our study, petaloid staminodes share their floral gene expression patterns with the innermost petals, suggesting that innermost pe tals might have originated from petaloid staminodes (or outer stamens). Pr eviously, based on the morphological similarity between petals and petaloid staminodes (or outermost stamens), petals of some taxa in Nymphaeales were
88 thought to have originated from stamens (see Albert et al. 1998), and our re sults also support this idea. However, the expression patterns of floral genes in staminodes, such as AG2 AGL6 and AP3 class VI transcript, differ fr om those observed in stamens even though staminodes also have anthers. Although their similar morphology indicat es staminodes may have originated from stamens, their genetic similarities should be investigated using stamen-specific genes for clarification of their homology (see below). In Nymphaea, most of the MADS-box genes ( Nyod.PI Nyod.AP3-class VI, Nyod.AG1 Nyod.AGL11 and Nyod.AGL6) are expressed in a smaller and narrower area later in development (anthesis) than earlier. In particular B-class genes are not ex pressed in carpels and ovules at anthesis. Also, the expression patterns of AG homologues differ between developmental stages, and those differences migh t contribute to the development of slightly differentiated sepals and petals. We also compared the floral developmental genetics of Cabomba and Nuphar with that of Nymphaea. The morphological features of Cabomba and Nuphar differ from Nymphaea In Cabomba, there are two petaloid perianth whorls; the ou ter whorl of sepals is very similar to the petals that have nectariferous parts, which ab sorb UV strongly (Endress 2001). However, there is no difference in floral gene expression patterns between sepa ls and petals in Cabomba (Fig. 44A ). When we consider that sepals and petals in Cabomba share characteristic s such as the color (purplish white), texture, development (simultaneous initiation), and vasculat ure (one trace) (Ito 1986; Les et al. 1999; Schneider et al. 2003), possessing the same gene expression patterns for the genes examined is not surprising. Unlike other members of Nymphaeales, Nuphar has two whorls of sepals, which differ slightly in coloration (Crow and Hellquist 2000; Endress 2001; Mo seley 1965, 1972; Schneider
89 and Williamson 1993; Schneider et al. 2003; Warner et al. 2008; Wiersema and Hellquist 1998; Zanis et al. 2003; Zomlefer 1994). Nuphar also has a whorl of nect ar-producing petals between sepals and stamens (Fig. 4-1 B ), which are sometimes considered staminodes (Judd et al. 2002; Kim et al. 2005; Warner et al. 2008). Based on fl oral gene expression prof iles, no differences are detected between the two whorls of sepals (Fig. 4-4B; Fig. 5b in Kim et al. 2005). However, petals differ from sepals in the expression patterns of AG AGL6 and AP1 homologues. Transcripts of AP1 and AGL6 homologues are detected in sepals, but not petals of Nuphar. In particular, transcripts of the AG homologue are found in peta ls, but not sepals of Nuphar (Fig. 44B ), while C-function homologues of Nymphaea and Cabomba are expressed in stamens and carpels only (Fig. 4-3A D) Also, petals of Nuphar exhibit the same gene expression patterns as those observed for stamens (Fig. 4-4 B ). Considering these gene expres sion patterns, the petals of Nuphar might have originated from stamens (andropetals, as in Persea Chanderbali et al. 2006). In fact, the development of petals of Nuphar is initiated in double pos itions against each sepal primordium, just like the origins of stamens in Cabomba (Fig. 4-5; Endress 2001), suggesting a close developmental relate dness between petals of Nuphar and stamens of Cabomba The two leading hypotheses for the origin of petals are derivation from bracts (b racteopetals) and from stamens (andropetals), respectively; basal angios perms are considered to have a bract-derived perianth while eudicots have andropetals (Hiepko 1965; Takhtajan 1991). However, recently, Chanderbali et al. (2006) hypothe sized that the tepals of Persea (Lauraceae) are derived from stamens, not from bracts. Also, bracteopetals are considered to be more common than previously recognized in the core eudicots (Ronse De Craene 2008). Therefore, together with the expression patterns of the floral organ identity genes, ba sed on their stamen-like appearance, developmental similarities to stamens (Yoo et al. in prep.), and their position between sepals and stamens, petals
90 might be of staminal origin despite the presen ce of an abaxial nectar y. If the petals of Nuphar are staminodes, the outer and inner sepals can be regarded as sepals and petals, respectively, based on their positions, as in other members of Nymphaeales. In fact, distinguishing staminodes from petals in Nymphaeales is not clear because if petals are derived from stamens, petals, stamens, and staminodes would all be homologous. Although most Nymphaeaceae but not Cabombaceae are reporte d to have staminodes (Heinsbroek and Van Heel 1969; Judd et al. 2002; Les et al. 1999; Moseley 1958), a detailed study of staminodes has not been done. Moseley (1958) found that staminode s have the same number of vascular bundles as are found in stamens, but only Barclaya and Victoria follow this rule. Although he doubted the presence of staminodes in Nymphaea and Nuphar he showed that there were only three vascular bundles in sepals, petals, and stamens in Nymphaea. Stamens of Nuphar have one trace while petals have three vascular bundles (Moseley 1958); however, traces of stamens and petals are from the same common system which supplie s the stamens and carpels (Moseley 1958). In general, therefore, staminodes of Nymphaeaceae are similar to stamens in the number of vascular bundles, but no information is available for staminodes of Ondinea and Euryale. In this study, we examined two species, Nuphar advena and Nymphaea odorata which have staminodes between petals and stamens. They share their gene expression patterns with stamens ( Nuphar ) and innermost petals ( Nymphaea), indicating that staminodes (petals) of Nuphar and innermost petals of Nymphaea might have originated from stamen s and staminodes (outermost stamens), respectively. However, we investigated only tw o species of Nymphaeaceae with a few genes, so the entities and features of staminodes in Nymphaeaceae should be reevaluated both morphologically and developmentally. Also, using stamen-specific genes, their genetic
91 homology can be investigated further. Those comparative studies ac ross diverse taxa of Nymphaeaceae will be helpful for elucid ating the identity of staminodes. Based on the gene expression data obtained for Nymphaea, Cabomba, and Nuphar the morphological differences among them may be expl ained by different flor al genetic programs. As a result, sepals and petals are not appropriate terms for the perianth members of Nymphaeales because they differ from typical sepals and pe tals of eudicot model organisms in morphology, development, and genetics. Endress (2006) noted that it is appropriate to use sepaloid or petaloid tepals to describe organs in Nymphaea due to the ambiguous flor al organ identities in these plants. Therefore, considering phylogene tic, morphological, developmental, and genetic data, it might be better to use sepaloid or petaloid tepals, fo llowing Endress (2006). According to these terms, Nymphaea species have four sepaloid tepals and numerous petaloid tepals, and Cabomba species have only six petaloid tepals. Nuphar has sepaloid tepals in the first whorl and three petaloid tepals in the second whorl. The or gans between the petaloid tepals and stamens in Nuphar might be additional whorls of petalo id tepals or petaloid staminodes. Floral Developmental Genetics in Nymphaea and Nelumbo To determine whether the floral morphological similarity of Nelumbo and Nymphaea is based on similarities in developmental genetics, we investigated the expression patterns of floral organ identity genes in Nelumbo and compared them to those obtained from Nymphaea. In Nelumbo, the expression pattern of Nenu.AP1 is similar to that of other FUL-like genes that are AP1 homologues in early-diverging e udicots and basal angiosperms: FUL-like genes are not true A-function genes, and their expression area is not re stricted to sepals and petals, but extends to all floral organs and leaf tissue (reviewed in Litt 2007), and our results also follow those expression patterns.
92 The PI and AP3 homologues of Nelumbo, Nenu.PI Nenu.AP3-1 and Nenu.AP3-2 are commonly expressed in all petals (outer, inner, and with a stam inal appendage) and stamens like the B-class genes of eudicot model plants. However, Nenu.PI is additionally expressed in sepals while Nenu.AP3-1 is also detected in carpels. Considering that PI and AP3 homologues function properly as heterodimers in Arabidopsis (Goto and Meyerowitz 1994; Honma and Goto 2001; Riechmann et al. 1996), Antirrhinum (Schwarz-Sommer et al. 1992; Zachgo et al. 1995), Petunia (Vandenbussche et al. 2004), and maize (Whipple et al 2004), B-class homologue expression in sepals and carpels in Nelumbo is surprising; Nenu.PI and Nenu.AP3-1 may not actually function in sepals or carpels respectively, because both are requi red for proper function (Goto and Meyerowitz 1994; Honma and Goto 2001; Riech mann et al. 1996; Vandenbussche et al. 2004; Whipple et al. 2004; Zachgo et al 1995). Therefore, together w ith developmental differences between sepals and petals (Hayes et al. 2000), this result suggests that differentiation of sepals and petals in Nelumbo might be due to differe nt expression of B-class genes; expression of both AP3 and PI homologues in petals, bu t expression of only the PI homologue in sepals. In fact, similar phenomenon was observed in expression patterns of PI and AP3 homologues of Ranunculaceae (Kramer et al. 2003): due to many duplications in B-class gene lineage in Ranunculaceae, some members of this family have several copies of PI and AP3 homologues which expression patterns differ. Howe ver, considering their proper functioning as heterodimers, they are only functional in petaloid sepals, petals, and stamens like Nelumbo (Kramer et al. 2003). These variations observed in expression patterns of B-class homologues in Nelumbo and Ranunculaceae might be due to the uns ettled ABC program in early-diverging eudicots. Also, as Kramer et al. (2003) suggested, lineage-sp ecific duplication events and
93 subsequent diversification in their function co uld be plausible explanation, which should be further addressed with more detailed study. The AG homologue from Nelumbo, Nenu.AG shows the same expression pattern as that of the AG homologue from Nymphaea, further supporting the id ea that the function of AG homologues may be conserved throughout the flowering plants (see above). Homologues of the remaining three genes, AGL6 AGL2 and AGL9 show the same expression patterns in Nelumbo and Nymphaea. The expression patterns of AGL2 and AGL9 homologues in Nelumbo are similar to AGL2 genes in other flowering plants (Chanderbali et al. 2006; Honma and Goto 2001; Kim et al. 2005; Pelaz et al. 2000; Zahn et al. 2005). The AGL6 homologue from Nelumbo also shows similar expression patterns to those obtained from Nymphaea, indicating its involvement in perianth and carpel development. However, both Nenu.AP1 and Nenu.AGL6 show similar expre ssion patterns, although Nenu.AP1 is also expressed in leaf tissue. In summary, in Nelumbo sepals are differentiated from pe tals based on their developmental sequence and gene expression (i.e., the expression pattern of a PI homologue, but not an AP3 homologue). Unlike the undifferentiated tepals (sepaloid and peta loid tepals) of Nymphaea Nelumbo has differentiated sepals and petals. Also, pe tals with a staminal appendage share their expression patterns with outer most and inner petals in Nelumbo indicating that petals with the staminal appendage are not homologous to stamens. Therefore, even though there is morphologica l similarity between the perianth of Nymphaea and Nelumbo, the floral gene expression patterns of the two genera differ from each other: broad floral gene expression patterns are obse rved in B-class homologues from Nymphaea, but the expression of B-class homologues in Nelumbo follows the classic ABCDE eudicot model.
94 In general, however, the expression of other floral organ identity gene homologues in both Nelumbo and Nymphaea generally follows the eudicot model. Thus, based on different organ identity gene profiles, simila r floral morphologies between Nelumbo and Nymphaea might be just morphological convergence. Molecular Models of Floral Development in Nymphaeales In this study, we focus on MADS-box genes b ecause they play key roles in regulating floral organ identities. Expression of Ca nd E-class homologues in Nymphaeales generally follows the classic ABCDE eudicot model. Howe ver, B-class homologues of Nymphaeales show generally broad expression patterns compared to those obtained from eudicot model organisms. These results are consistent with the fading borders model (Buzgo et al. 2004; Buzgo et al. 2005) and may explain the intergradation of perianth organs, and of perianth to stamens in Nymphaeales. Similar results were found in other basal angiosperms with gradual differentiation of floral organs (Chanderbali et al. 2006; Kim et al. 2005). However, Cabomba, which has discrete whorls of floral organs with a pe rianth differentiated by position, fits the shifting boundary model (Bowman 1997) or the slidi ng boundaries model (Kramer et al. 2003). Therefore, within Nymphaeales, both the fading borders model, which appears to be ancestral for the angiosperms (Kim et al. 2005), and the sliding boundaries model are present in divergent clades. However, furthe r work is needed to evaluate the details of the fading borders model in Nymphaeales. No difference in signal strength is apparent between the margin and center of a genes zone of activity, for examples between sepaloid tepals and carpels. Similar results were obtained previously for PI / AP3 in Nuphar also with RT-RQ-PCR (Kim et al. 2005). However, using RNA in situ hybridization, both PI / AP3 homologues were strongly expressed in stamens and staminodes (petals in this study), but also very weakly expressed in sepaloid and
95 petaloid tepals rather than highl y expressed in all floral organs (Kim et al. 2005). This result shows that the fading borders model indeed might be applicable for the PI / AP3 homologues of Nuphar but the sensitivity of RNA in situ hybridization is needed to evaluate the details of the models predictions. Therefore, to determine the details of the fading borders model in Nuphar and Nymphaea, further research is needed using additional developmental stages and additional genes, and greater sampling of species of Nymphaeales. Origin of the Floral Parts in Nymphaeales Since the first appearance of angiosperms in earth about 130 mya, origin of flower and their subsequent strong diversification have been a mystery fo r a long time, which is often called Darwins abominable mystery (Crepet 1998, 2000; Frohlich 1999; Frohlich and Parker 2000; Frohlich 2003; Ma and dePamphilis 2000; Theiss en et al. 2002; Theissen and Becker 2004; Theissen and Melzer 2007; Winter et al. 2002). Recently, based on the molecular and developmental genetics, several th eories and hypothesis have been proposed for origin of flower; among them are the out-of-male hypothesis, th e out-of-female hypothesis, and the mostly male theory (Frohlich and Parker 2000; Fr ohlich 2003, 2006; Theissen et al. 2002; Theissen and Becker 2004; Theissen and Melzer 2007). The first two theories were proposed based on the fact that gymnosperms, the closest relatives of angios perms, have orthologues of B-class genes and their expression patterns are distinct in reproductive organs (Theissen et al. 2002; Theissen and Becker 2004). For example, B-cl ass homologues were expressed only in male reproductive units (Becker et al. 2000; Becker et al 2003; Fukui et al. 2001; Moura dov et al. 1999; Sundstrm et al. 1999; Winter et al. 1999), while C/D-class homologue s were expressed in both male and female reproductive units (Jager et al. 2003; Rutledge et al. 1998; Tandre et al. 1998; Winter et al. 1999). These expression patterns facilitate the out-o f-male or out-of-female hypothesis, explaining
96 that hermaphroditic flower originated from either a male cone or a female cone, respectively, and changes in B gene expression played a critical role in evolution of the other reproductive organ (Theissen et al. 2002; Theissen and Becker 2004; Theissen and Me lzer 2007). For example, in the out-of-male hypothesis, B genes were expressed in male cones, but down-regulation of B genes in the upper region of male cones resulted in development of the female structures there. In the out-of-female hypothesis, ectopic expressi on of B gene in lower region of female cones led to development of male reproductive organs (Theissen et al. 2002). The mostly male theory was developed based on the LEAFY duplications prior to the split between angiosperms and gymnosperms and the loss of one copy in angiosperm lineages (Frohlich and Parker 2000). In this theory, two copies of LEAFY LEAFY and NEEDLY were required for specifying male and female cones in gymnosperms, respectively. The loss of NEEDLY in angiosperms caused to retain more genes relatively active in male c ones, and the minimum set of female genes were restricted in ovule identity (Frohlich and Parker 2000; Fr ohlich 2003, 2006). Therefore, in general this theory is similar to the out-of-ma le hypothesis in respect to derivation of flower organization more from the male structure than from the female structure (Frohlich and Parker 2000; Frohlich 2003, 2006). However, in the mostly ma le theory, the female structure developed ectopically in male cones (F rohlich and Parker 2000; Frohlich 2003, 2006). Although these theories do not propose the mechanism for peri anth development, later Baum and Hileman (2006) suggested that sterilizati on of the outer stamens led to evolution of perianth organs. Our data, especially the expres sion patterns of B-class homologues in Nymphaeales, fit the outof-male hypothesis. First, PI and AP3 homologues of Nuphar are expressed all floral organs, but the expression of PI homologue is not detected in carpels later in develo pment (Kim et al. 2005). This is also observed in B-class homologues of Nymphaea; Nyod.PI and Nyod.AP3 are expressed
97 in all floral organs early in development while th eir expression are not detect ed in carpels later in development (Fig. 4-4 C ). Therefore, broad gene expressi on patterns of B-class homologues with restriction of their expression in carpels later in development support the idea that downregulation of B-class homologues might have resulted in female reproductive units in the upper region of male structures. Surprisingly, sim ilar expression pattern was also observed in PI of Arabidopsis : unlike AP3, PI transcripts of Arabidopsis were detected in the second, third and fourth whorls of flowers at stage 3 (petal initiation), but PI was not detected in the fourth whorl after stage 4 (stamen initiation), indicating that PI expression early in development may not be related to carpel specification (Goto and Meyerowitz 1994). However, this feature observed in Arabidopsis might be vestige of ancestral feature wh ich might have been a rule for flowering plants. Furthermore, perianth organs of Nymph aeales share their B gene expression patterns with stamens (Fig. 4-4A C ; Kim et al. 2005). In addition, Yoo et al. (in prep.) also suggested that perianth organs might have originated from stam ens based on analysis of global gene expression profiles using microarray. Thus, together with microarray data, our data indicate that the out-ofmale hypothesis, which assumes that all floral pa rts might have derived from male reproductive units, might be true for ancestors of Nymphaeaceae.
98Table 4-1. Primer information used in this study Taxon Gene name Forward sequence Reverse sequence Name: sequence Name: sequence GLO subfamily Cabomba Caca.PI CcPI-1: GAACCAAGAGCT GGAGAGAA CcPI-2: ACGGTTGCATCGCCGCCA Nymphaea Nyod.PI Nyod.PI-3: ATCAGTGTCCTCTGT GACGC Nyod.PI-4: CTGCTGCAAGTTAGGCTGTA Nelumbo Nenu.PI Nenu.PI: CTCAGCAACGAAATAG ACAG Nenu.PI-2: GCTGGATTGGCTGCACACG DEF subfamily Cabomba Caca.AP3 CcAP3-3: GTTGAGTTACACC GAGCTG CcAP3-4: TCATGCTAGCCTCAAGTCA Nymphaea Nyod.AP3 Nyod.AP3-5: GAAAAAGATAAGGTTA GCCG Nyod.AP3-4: CACG CAGATTAGGATGGCTG Nelumbo Nenu.AP3-1 Nenu.AP3-1: CGATCTGAGTGTGGACGAGC Nenu.AP3-2: CATCCTGAAGATTGGGCTG Nenu.AP3-2 Nenu.AP3-3: CGATATGAGCATCGAAG AAC Nenu.AP3-4: CAACCTGAAGATTAGGATG AG subfamily Cabomba Caca.AG-1 CcAG-7: ACCTTGAAGTG AAGCTGGAG CcAG-8: GAGGCTCTCCAGTCATCATA Caca.AG-2 CcAG-9: ATGCTCAATATTATCAACAGG CcAG-8: GAGGCTCTCCAGTCATCATA Nuphar Nuad.AG Nuad.AG-5: ATGCTGGGTGAAGGAAT CAG Nuad.AG-6: TCATCCAAGTTGTAGTGCCG Nymphaea Nyod.AG1-1 Nyod.AG4-1: AGAATAAATTGGA GAGAAGC Nyod.AG4-2: TCCATTAGATTGATGTGCTG Nyod.AG1-2 Nyod.AG2-3: GTGAGAATGAAAGAGCGC AGC Nyod.AG2-4: TACGACAATGAGACAACGGC Nyod.AG2 Nyod.AG-1: CTCCAAC TCGGGAACCGTTAC Nyod.AG-4: ATGCTGCTGCTGCTGCTGATG Nyod.AG3 Nyod.AG3-1: ATCCTTCAGAATGCGAAC AG Nyod.AG3-2: TCAGCAACCTTGGCTCGAA Nyca.AG1 Nyca.AG1-1: GAGACCTCAGATCTTTAGAG Nyca.AG1-2: TCATCCAAGTTGCAAGGCAG Nyca.AG2 Nyca.AG2-3: ACTTAGAAGGCAAACT GGAG Nyca.AG2-4: TCATCCAAGTTGAAGGGC Nyca.AG3 Nyca.AG3-3: GACCAAGTTGAGGCAG CAG Nyca.AG3-4: GGTTGGAGTAGTGGATCATG Nelumbo Nenu.AG Nenu.AG-1: GCTCTTAGCACTA TGACTG Nenu.AG-2: CTTGGCGAGAGTAATGGTG
99Table 4-1. Continued Taxon Gene name Forward sequence Reverse sequence Name: sequence Name: sequence SQUA subfamily Cabomba Caca.AP1 CcAP1-1: CATCAGACTCCAGCATAATG CcAP1-2: ATAGACTCCTGCATCAGCTG Nuphar Nuad.AP1 NuadAP1-3: GTGAAGATCTCGAGCCA TTGAG NuadAP1-4: GCTAGTCTGAGTTTGCAGGTG Nelumbo Nenu.AP1 Nene.AP1-1: TAGAGCAACAGCTTGACAC Nenu.AP1-2: GTGAGTTGGTCCGAGCATG AGL2 subfamily Cabomba Caca.AGL2-1 CcSEP-1: TCTCTGCGATGCTGAGGTC CcSEP-2a: CA TATACTGTGTCTTGGTAG Caca.AGL2-2 CcSEP-1a: GTACCAGAAGTGTAGC TATG CcSEP-2a: GAACGTACTGGATGTGCTG Caca.AGL2-3 CcSEP-3: TTGCGATGCGGAGGT CG CcSEP-4: TGAGGTCAGCAAGTTGATC Nuphar Nuad.AGL4 Nuad.AGL4-1: CATCTTGAGCAGCAGC TAGAG Nuad.AGL4-2: CA GTTATCTGCTCTTGCGCA Nuad.AGL9 Nuad.AGL9-1: ACTGGA GCGGTATCAGAAGT Nuad.AGL9-2: CTTGTTAGCTTCGATCAGCA Nymphaea Nyod.AGL2 Nyod.SEP-1: AGCTACTGTGCCAT CTCGAGA Nyod.SEP-2: ATGTGCTGCGCACTGTTATC Nelumbo Nenu.AGL2 Nenu.AGL2-1: GCTTGAACAGCTTG AGCATC Nenu.AGL2-2: GTTATCTGATCTGGACCAAC Nenu.AGL9 Nenu.AGL9-1: CTACAGCGGTCACAGA GGA Nenu.AGL9-2: GCTCACACTCTACAGGATG AGL6 subfamily Cabomba Caca.AGL6-1 CcAGL6-1: TGATGGAGCA GATGGATGA CcAGL6-2: ATGAAGTTGCTCTCAGGAG Caca.AGL6-2 CcAGL6-3: CATTAGACAACAGCA TAGC CcAGL6-4: GCTCTGAGAACATGACCA Caca.AGL6-3 CcAGL6-5: TACGATGCATTTGACAACAG CcAGL6-6: CATGCGATTCTAGCTGCA Nuphar Nuad.AGL6 Nuad.AGL6-1: GCAAAGCTGAAGGCAAGATA Nuad.AGL6-2: CATGTCCTTGTGATTCTAGCT Nymphaea Nyod.AGL6 Nyod.AGL6-1: CAACACTGCTGCTG CTGCGA Nyod.AGL6-2: GTTCACGTCTCCAAGGTGAC Nelumbo Nenu.AGL6 Nenu.AGL6-1: ACCAGTGCGTCC GACCGTGA Nenu.AGL6-2: TGATGTCTCCAAGATGACGT GGM13 subfamily Cabomba Caca.Bs1 CcBs-1: GAGGATCTTGCCA CTCTCAC CcBS-2: TGGTGGATGGCTGCATGATC Caca.Bs2 CcBs-1: GAGGATCTTGCCACT CTCAC CcBS-4: ATGTACGGCCGCAGGATC
100Table 4-2. Comparison of floral gene expression patterns from different developmental stages of Nymphaea odorata Gene Leaves Sepals Outer petals Innermost petals Staminodes Stamens Innermost stamens Carpels Ovules Nyod.PIa Nyod.PIb +++ +++ +++ +++ +++ +++ +++ +++ +++ +++ +++ +++ Nyod.AP3 Class Ia Class Ib +++ +++ +++ ++ +++ +++ + +++ ++ +++ +++ +++ ++ Class IVa Class IVb ++ ++ ++ +++ +++ ++ ++ +++ +++ Nyod.AG1-1a Nyod.AG1-1b +++ ++ ++ +++ ++ +++ +++ +++ +++ ++ ++ Nyod.AG1-2a Nyod.AG1-2b ++ ++ +++ +++ + +++ ++ +++ +++ +++ + Nyod.AG2a Nyod.AG2b ++ +++ ++ +++ +++ +++ ++ Nyod.AG3a Nyod.AG3b +++ +++ +++ Nyod.AGL2a Nyod.AGL2b +++ +++ ++ +++ +++ ++ +++ +++ +++ +++ +++ ++ +++ +++ Nyod.AGL6a Nyod.AGL6b +++ +++ ++ +++ +++ ++ +++ +++ aGene expression patterns from an early developmental stage bGene expression patterns at anthesis
101Table 4-3. Summary of expression patterns of floral genes in Nymphaeales and Nelumbo just prior to anthesis Taxa Gene LeavesSepals 1st Petals 2n d Inner stamens 3rdCarpels 4th GLO subfamily Cabomba Nuphar Nymphaea Nelumbo DEF subfamily Cabomba Nuphar Nymphaea Nelumbo AG subfamily C lineage Cabomba Nuphar Nymphaea Nelumbo D lineage Nymphaea SQUA subfamily Cabomba Nuphar Nelumbo Caca.PI Nuad.PI a Nyod.PI Nene.PI Caca.AP3 Nuad.AP3-1 a Nuad.AP3-2 a Nyod.AP3(class I) Nyod.AP3(class VI) Nene.AP3-1 Nene.AP3-2 Caca.AG Nuad.AG a Nyod.AG1-1 Nyca.AG1 Nyod.AG1-2 Nyod.AG2 Nyca.AG2 Nene.AG Nyod.AG3 Nyca.AG3 Caca.AP1 Nuad.AP1 Nene.AP1 + +++ +++ +++ ++ +++ +++ +++ +++ +++ +++ +++ +++ +++ ++ +++ +++ +++ +++ +++ +++ +++ +++ +++ +++ +++ +++ ++ +++ +++ ++ +++ ++ ++ +++ +++ +++ +++ +++ +++ +++ +++ +++ ++ +++ +++ +++ +++ +++ +++ +++ ++ +++ +++ +++ +++ ++ ++ +++ + +++ +++ +++ +++ +++ +++ +++ ++ +++ +++ +++ +++ +++ ovules ovules +++ +++ ++
102Table 4-3. Continued Taxa Gene Leaves Sepals 1st Petals 2n d Inner stamens 3rdCarpels 4th AGL2 subfamily AGL2 lineage Cabomba Nuphar Nymphaea Nelumbo AGL9 lineage Nuphar Nelumbo AGL6 subfamily Cabomba Nuphar Nymphaea Nelumbo GGM13 subfamily Cabomba Caca.AGL2-1 Caca.AGL2-2 Caca.AGL2-3 Nuad.AGL2b Nuad.AGL4 Nyod.AGL2 Nene.AGL2 Nuad.AGL9 Nene.AGL9 Caca.AGL6-1 Caca.AGL6-2 Caca.AGL6-3 Nuad.AGL6 Nyod.AGL6 Nene.AGL6 Caca.Bs1 Caca.Bs2 ++ +++ +++ +++ +++ +++ +++ +++ +++ ++ +++ +++ +++ +++ ++ +++ +++ +++ +++ +++ +++ +++ +++ +++ +++ +++ ++ +++ +++ +++ +++ +++ +++ +++ +++ +++ +++ + +++ + +++ +++ +++ +++ +++ +++ +++ +++ +++ + +++ +++ +++ ovules ++ +++ +++ NOTEFor comparison, the expression patterns of Nymphaea odorata were obtained from an early developmental stage. aData from Kim et al. (2005) bData from Zahn et al. (2005)
103 Figure. 4-1. Photographs of flowers of Nymphaeales and Nelumbo A Cabomba Caroliniana; peltate floating le aves and diand trichotomously dissected subm erged leaves, and flower. B Nuphar advena ; flower and petals. C Nymphaea odorata flower. D Dissected floral parts of N. odorata; sequence showing from four green sepals to 28 white petals (the 1st and 2nd rows) and from petaloid staminodes (the 3rd row) to functional stamens (the 4th and 5th rows), and two views of carpels and ovules are presented (photo credit: S. Kim). E Nelumbo nucifera flower.
104 Figure. 4-2. Alternatively spliced transcript of Nyod.AP3. A Intron/exon structure of the Nyod.AP3 genomic sequences. Designations of the exons and introns are based on comparison to genomic structure of NymAP3 (Stellari et al. 2004) and Arabidopsis AP3 (Jack et al. 1992). B Inferred structure of cDNA splicing variants (Ste llari et al. 2004). The number of clones for each class is shown in parentheses.
105 Figure. 4-3. Phylogenetic analyses using 186 MADS-box gene homologues. A One of the most parsimonious trees. Triangle symbol indica tes the collapsed branch in the strict consensus tree of 1355 shortest trees from a maximum parsimony analysis of MADS genes (6815 steps, consistency inde x=0.3859 and retention index=0.7568). MADSbox gene homologues newly identified and us ed in this study (underlined) are placed as a member of a well-supported major cl ade of MADS-box subfamily (thickened nodes). B Enlarged tree of AGL2 AGL6 and SQUA lineages. C Enlarged tree of AG lineage. D Enlarged tree of DEF, GLO and Bsister lineages.
106 Figure. 4-3. Continued.
107 Figure. 4-3. Continued
108 Figure. 4-3. Continued.
109 Figure. 4-4. Relative quantitative RT -PCR results of floral MADS-box ge ne homologues in Nymphaeales and Nelumbo. Standard deviations are indicated for each value. sLF, submerged leaves ; fLF, floating leaves; SE, sepa ls; PE, petals; ST, stamens; CA, carpels; OSE, outer sepals; ISE, inne r sepals; LF, leaves; OPE, outer petals; IPE, inner petals; SN, staminodes; ST1, outermost stamens; ST2, innermost stamens; OV, ovules; FA; floral apex; SA, staminal appendages. In Nymphaea odorata E shows expression patterns from early developmental stage, and A indicates expression patterns from at anthesis.
110 Figure 4-4. Continued.
111 Figure 4-4. Continued.
112 Figure 4-4. Continued.
113 Figure 4-4. Continued.
114 Figure 4-4. Continued.
115 Figure. 4-5. Developm ental features of Nuphar and Cabomba A Petal initiation (3) in Nuphar advena. B Stamen initiation (3) in Cabomba caroliniana. Images from Endress (2001).
116 CHAPTER 5 ANALYSIS OF THE FLORAL TRANSCRIPTOME OF A BASAL ANGIOSPERM, NUPHAR ADVENA (NYMPHAEACEAE) Introduction Extensive genetic analyses, particularly of Arabidopsis and Antirrhinum floral mutants, have led to the development of the ABCDE mo del for the molecular mechanism controlling floral organ identity: A and E class genes control sepal identity; A, B, and E class genes control petal identity; B, C, and E class genes control stamen identity; C and E class genes control carpel identity; D and E class genes c ontrol ovule identity (Coen and Me yerowitz 1991; Colombo et al. 1995; Ditta et al. 2004; Pe laz et al. 2000). In Arabidopsis A function is provided by APETALA1 ( AP1) and APETALA2 ( AP2), B function by APETALA3 ( AP3 ) and PISTILLATA ( PI ), C function by AGAMOUS ( AG ), and E function by multiple SEPALLATA gene products ( SEP14 ). Some aspects of the ABCDE model ar e conserved across eudicot and monocot angiosperms (i.e., Petunia hybrida (Angenent et al. 1993); Silene latifolia (Hardenack et al. 1994); Gerbera hybrida (Yu et al. 1999); Oryza sativa (Fornara et al. 200 3; Kater et al. 2006; Kyozuka et al. 2000); Pisum sativum (Taylor et al. 2002); Zea mays (Whipple et al. 2004); Magnolia grandiflora (Kim et al. 2005), Persea americana (Chanderbali et al. 2006), Akebia trifoliata (Shan et al. 2006); Elaeis guineensis (Adam et al. 2007); Taihangia rupestris (L et al. 2007); Vitis vinifera (Poupin et al. 2007)). All of the ABCD E genes are transcription factors, and are thought to interact with genes involved in the establishment of floral meristem identity or floral organ formation (Jack 2004; Theissen 2001b; Wellmer et al. 2006). However, relatively few downstream targets have been identified, possibly due in part to limitations of genetic approaches based on mutant analyses (Wellmer et al. 2006). Also, many duplication events leading to functional redundancy may hinder el ucidation of functional roles in flower
117 development (Arora et al. 2007; Cho et al. 2002; Moore et al. 2005; Moore and Purugganan 2005; Wellmer et al. 2006). Recent studies of gene expression profiling usin g microarray techniques provide valuable information for understanding the transcriptiona l programs of plants on a genome-wide scale (Alves-Ferreira et al. 2007; Arora et al. 2007; Birnbaum et al. 2003; Cho et al. 2002; Wellmer et al. 2004; Wellmer and Riechmann 2005; Wellmer et al. 2006). In particular, several studies identified genes with potentially important role s in flower development (Gomez-Mena et al. 2005; Hennig et al. 2004; Schmid et al. 2003; Schmid et al. 2005; Wellmer et al. 2006; Zhang et al. 2005). For example, many potential signaling co mponents, such as protein kinases and other putative signaling proteins, and transcription factors, were up-regulated during reproductive development (Hennig et al. 2004). Among the latter are members of the YABBY, MADS box, and MYB gene families, consistent with geneti c studies of floral organ development (Bowman 2000; Golz and Hudson 1999; Ng and Yanofsky 2001; Theissen et al. 2000). Other studies have targeted the expression profiles of specific intra-floral organs and/or tissues, for example, stamens (Alves-Ferreira et al. 2007; Wang et al. 2005; Wellmer et al. 2004), stigma (Li et al. 2007), female gametophyte (Yu et al. 2005), and embryo sac (Jones-Rhoades et al. 2007). Microarray-based analyses are also useful for comparing the expression profiles of duplicated genes or members of gene families within organisms (Arora et al. 2007; Schmid et al. 2005; Wellmer et al. 2006). For example, Schmid et al. (2005) showed that large gene families, such as WRKY and MADS transcription factors, which are important developmental regulators, exhibited correlated expression patterns; many WRKY genes are pr eferentially expressed in leaves while many MADS-box genes are mainly expr essed in flowers and fruits. Wellmer et al. (2006) also showed that in Arabidopsis closely related genes in gene families are highly
118 correlated in their temporal gene expression patterns, and the majority of those genes is upregulated during certain developmental stages. As illustrated by the studies above, nearly all transcriptome analyses of floral development have so far focused on Arabidopsis (Gomez-Mena et al. 2005; Henni g et al. 2004; Schmid et al. 2003; Schmid et al. 2005; Wellmer et al. 2004; Wellm er et al. 2006; Zhang et al. 2005). Here, we report the floral transcriptome profiling of Nuphar advena, a member of the Nymphaeales (water lilies) which, together with Hydatellaceae (Saarela et al. 2007), lie sister to all other extant angiosperms but Amborella (Fig. 5-1) (Borsch et al. 2005; Hilu et al. 2003; Jansen et al. 2007; Lhne and Borsch 2005; Leebens-Mack et al. 2005; Mathews and Donoghue 1999; Moore et al. 2007; Qiu et al. 1999; Qiu et al. 2005; Soltis et al. 1999; Soltis et al. 2000; Soltis et al. 2005; Zanis et al. 2002; Zanis et al 2003). Nymphaeales consist of two families, Cabombaceae and Nymphaeaceae, and reconstructions of floral evol ution suggest that a small number of floral parts and whorled phyllotaxy (trimery ) are ancestral characteristics (Borsch et al. 2008; Les et al. 1999; Zanis et al. 2003), wher eas flowers with large number of floral parts, e.g. Nymphaea, Victoria and other water lilies, are derived in Nymphaeaceae (Borsch et al. 2008; Les et al. 1999; Zanis et al. 2003). Nuphar is sister to all other genera of Nymphaeaceae (Borsch et al. 2007; Borsch et al. 2008; Lhne et al. 2007; Les et al. 1999), and has trimerous flowers of moderate size; the ancestral character states of Nymphaeales (Les et al. 1999). The flower of Nuphar advena consists of two perianth whorls (t hree sepals and three petals), numerous staminodes, numerous stamens, and numerous ca rpels in a syncarpous gynoecium (Fig. 5-2). The two perianth whorls differ in color: the sepals in the outer whorl are green, and the petals in inner whorl are yellow (Fig. 5-2; Padgett et al. 1999; Warner et al. 2008). The staminodes are smaller
119 than the sepals and petals, but broader than the stamens, with a nectary on their adaxial surface (Fig. 5-2). We have used floral ESTs collected by the Floral Genome Pr oject (Albert et al. 2005; 1http://fgp.bio.psu.edu/fgp/index.html) to conduct the first investigation of the floral transcriptional profile in Nuphar one of the basalmost angiosperm lineages. Specifically, we assessed gene expression levels in both young an d medium-aged floral buds, as well as sepals, petals, stamens, carpels, fruits, relative to leaves, to provide an assessment of the genes involved in floral development in one of the basal-mo st angiosperms. Also, we conducted comparative analyses of similar data sets for Arabidopsis (Schmid et al. 2005) and Persea (Chanderbali et al. submitted), the latter a member of the large magnoliid clade of basal angiosperms, to infer evolutionary trends in floral transcriptomes. Materials and Methods Sample Preparation, Probe Labeling and Microarray Hybridization Tissues were collected from four i ndividuals (biologica l replicates) of Nuphar advena in Pennsylvania, USA. The tissues include young leaf tissue, young floral buds (Ybd) at the premicrosporangia initiation stage, medium-aged fl oral buds (Mbd) at the pre-meiotic stage, and sepals, petals, stamens, and carpels, dissected fr om flowers at anthesis. Total RNA was extracted from all tissue samples using the RNeasy Plant Mi ni Kit (Qiagen, Stanford, CA, USA). Both the quantity and quality of RNA were assessed us ing a 2100 Agilent Bioanalyzer. RNA transcripts were labeled using the Low-input RNA labeling kit (Agilent Inc.) and hybridized to the arrays according to the manufacturers protocol.
120 Microarray Design We used custom microarrays produced by Agilent Technologies (Palo Alto, CA), containing 10,187 60-mer oligonucleotid e probes, targeting 6,220 unique Nuphar floral transcripts collected from a pre-meiotic flor al cDNA library by the Floral Genome Project (Albert et al. 2005, Soltis et al. 2007; 2http://fgp.bio.psu.edu/fgp/index.h tml). The oligonucleotide probes were in situ synthesized and randomly arranged on th e arrays. For quality control checks of hybridization, 544 Agilent controls were included in the arrays. We measured expression level in different developmental stages of the flower (pre-microsporangia initiation stage vs. the premeiotic stage in floral buds), sepals, petals, st amens, carpels, fruits, and young leaves, using an interwoven double loop for eight tiss ues with 16 arrays (Fig. 5-3), which minimizes the variance of pair-wise comparisons between any two tissues and efficiently detects differentially expressed genes. Data Acquisition and Statistical Analysis Microarrays were scanned with an Agile nt DNA microarray scanner using Agilents Feature Extraction Software 9.1.3 (Agilent Inc.). Raw data were imported into the Bioconductor package Limma and processed as previously de scribed (Chanderbali et al. submitted). After quality control checks for hybridiz ation, arrays were background corrected and loess normalized within arrays and Aq normalized between arrays (Smyth et al. 2006; Yang and Thorne 2003). To identify significantly differentially expre ssed genes among the eight tissue samples, we employed a one-way empirical Bayes ANOVA using single channel analysis while considering correlation between channels at each spot (Smyth 2004; Smyth et al. 2005). Data for all genes showing differential expr ession were assembled by hierarchical clustering (Eisen et al. 1998), as implemented by de Hoon et al. (2002), into gr oups with similar gene expression patterns.
121 Specifically, log2 ratios relative to leaf for 3,333 genes with a differential expression probability greater than 95% (p-value <0.05) were read into CLUSTER 3.0 (de Hoon et al. 2002) and hierarchical clustering performed via centroid linkage of Pearson co rrelations. The cluster results were visualized using Java TreeView (Saldanha 2004). We also searched log2 transformed expression data of differentially expressed gene s for transcripts with at least two-fold upregulation in floral organs re lative to leaves to identify ge nes significantly up-regulated in particular floral organs and/or floral stages. Comparison of Microarray Data with Relative Quantitative Reverse Transcription PCR (RT-RQ-PCR) Data We compared our microarray da ta with RT-RQ-PCR data for Nuphar from Kim et al. (2005), Zahn et al. (2005), and Yoo et al. (in prep.) for Nuphar homologs of several floral organ identity genes; AGAMOUS ( AG ), PISTILLATA ( PI ), APETALA3 ( AP3 ), AGAMOUS-LIKE 2 ( AGL2; SEP1), AGAMOUS-LIKE 6 ( AGL6 ), and APETALA1 ( AP1). From our microarray data for Nuphar we used the absolute amount of gene expre ssion in each tissue and then compared those expression patterns relative to leaves. In this way we assessed whether ge ne expression patterns from microarray and RT-RQ-PCR are consistent with each other. Comparative Floral Transcriptomics To investigate evolutionary pa tterns in floral transcriptom es we compared the expression profiles of differentially expressed Nuphar genes with those of their putative homologs in the model angiosperm Arabidopsis representing the derived eudi cot lineage Brassicaceae, and Persea of the magnoliid clade of basal angiosperms. In the absence of reliable gene family phylogenies, except for MADS box genes (Kim et al. 2005), we based gene homology assignments on amino acid sequence similarity (best reciprocal BLAST E score < 10-5). We
122 combined the AtGenExpress (Development) expression data for Arabidopsis (Schmid et al. 2005) and the Persea data set (Chanderbali et al. submitted) with our Nuphar data set based on putative homology to construct two separate data sets: (1) Nuphar and Arabidopsis homologs ( Nuphar subset 1), and (2) Nuphar Persea and Arabidopsis homologs ( Nuphar subset 2). Each data set was subjected to hierar chical clustering analyses, pair -wise plots of gene expression levels in floral organs within species to compar e the extent of correlation in their transcriptional profiles, and sorted by organ of primary positive expression ratios relative to leaves to compare the relative extent of expressi on domains. Finally, we searched for genes with organ-specific expression in accordance with the criteria of Cha nderbali et al. (submitted), where organ-specific expression is defined as an at l east two-fold up-regulation in the ta rget floral organ and less than 1.5-fold increase in all others. Results Genes Expressed Differential ly in Reproductive Organs We identified 3,333 differentially expressed genes among the eight tissue samples (P < 0.05; FDR = 0.38%), of which 2,165 were up-regulate d in floral tissues, including 831 at twofold or higher expression levels relative to leaves (Table 5-1). Thus, 53.6% (3,333 out of 6,220) and 13.4% (831 out of 6,220) of the genes surveyed are differentially expressed and up-regulated in floral tissues relative to leaves with a mi nimum of two-fold higher expression levels. These values found in Nuphar are low relative to those obtaine d from gene expression profiles of Persea which was studied using a similar research strategy: 77.9% of ge nes (4,744 out of 6,086) showed differential expressi on, and 17.8% of them (1,084 out of 6,086) were up-regulated in floral tissues and/or stages rela tive to leaves. We identified only 22 and 23 genes at least 2-fold up-regulated in carpels and stamens of Nuphar whereas 68 and 105 such genes were up-
123 regulated in carpels and stamens of Persea respectively (Table 5-1). Most of the genes upregulated in Nuphar floral organs are primarily expre ssed in perianth organs and stamens (35.8%), followed by perianth organs alone (20. 3%). Therefore, the most common pattern was observed in both perianth members and stamen s with 88 genes, followed by both sepals and petals, and all floral organs (Table 5-1). Among the genes down-regulated in Nuphar flowers are transcription factors such as members of the bHLH, homeobox, SBP (S QUAMOSA-pROMOTER B INDING P ROTEIN), and WRKY families (Table 5-2). Several are related to leaf morphogenesis in Arabidopsis for example, homologs of CURLY LEAF (Kim et al. 1998), BRASSINOSTEROID-INSENSITIVE 2 (Perez-Perez et al. 2002), and FASCIATA 2 (Kaya et al. 2001) (Table 5-2). However, several genes related to flower development are also down-regulated in some reproductive organs. For example, a homolog of ABNORMAL FLORAL ORGANS required for flower organ formation and development in Arabidopsis (Chen et al. 1999), is down-expr essed in all reproductive organs except Ybd, indicating an ex clusively early role in Nuphar As observed for non-floral gene s, many genes up-regulated in Nuphar flowers are transcription factors of the bHLH, bZIP, MADS, and Myb-re lated gene families (Table 5-3). All the Nuphar MADS-box genes examined, except the AP1 homolog, are up-regulated in reproductive organs. bZIP transcription factors are diverse in their functions, including light and stress signaling, floral transition, and seed development (Ja koby et al. 2002; Nijhawan et al. 2008). Eight members of the bZIP family are differentially expressed in Nuphar all mainly expressed in Ybd. NAC-LIKE ( NAP, ACTIVATED BY AP3/PI ) plays a role in cell morphogenesis and leaf senescence (Guo and Gan 2006; Li et al 2004) and floral development in Arabidopsis particularly in petals and stamens (S ablowski and Meyerowitz 1998), but the Nuphar homolog is
124 expressed in early floral st ages (Ybd and Mbd) and not in mature stamens and petals. KANADI1 ( KAN1 ) is known to regulate l eaf and carpel polarity in Arabidopsis and is expressed temporarily on the abaxial side of initiating floral-organ primordia (Kerste tter et al. 2001). The Nuphar KAN1 homolog is expressed in pe tals, stamens, and carpels. Comparison of Microarray Data with RT-RQ-PCR Data Previously reported RT-RQ-PCR data for Nuphar homologs AG PI AP3 AP1, and SEP1 (Kim et al. 2005, Zahn et al. 2005) generally agre e with the expression patterns obtained in our microarray experiments (Fig. 5-4). This is especially true for homologs of AP3 and PI Transcripts of a Nuphar AP1 homolog were barely detectable in floral organs with RQ-RT-PCR, and the same holds true in our microarray data (Fig. 5-4). The Nuphar AG homolog was only detected in stamens and carpels by RQ-RT-PCR techniques, and is strongly up-regulated in stamens and carpels relative to leaves accordi ng to the microarray datas, but it is also upregulated, albeit at lower expres sion levels, in sepals and peta ls (Fig. 5-4). Similarly, the Nuphar AGL6 homolog was detected in sepal, petals, and carpels with RQ-RT-PCR (Yoo et al. in prep) whereas some stamen expression was detected with microarrays, again at lower levels compared to the other floral tissues (Fig. 5-4). Hierarchical Clustering Hierarchical clustering sorted differentially ex pressed genes into flor ally up-regulated and down-regulated clusters (Fig. 5-5). Within the up-regulated cluster, two major subsets accommodate genes expressed primarily in carpels, including the Nuphar homolog of AG and genes expressed broadly across the flower but primarily in stamens, sepals and petals including homologs of AP3, PI SEP1, and AGL6 The AP1 homolog was placed in the down-regulated cluster together with most genes primarily e xpressed during the earliest developmental stages
125 sampled. A second probe targeting the PI homolog detected a similar expression pattern and was also placed in the down-regulated cluster (Fi g. 5-5). Clustering of ti ssues separated the two floral buds from mature organs suggesting ex tensive turnover in transcriptional programs through developmental stages, and grouped sepals with petals, followed by stamens, then carpels in the hierarchy of floral organ relationships (Fig. 5-5). Comparative Floral Transcriptomics Hierarchical Clustering: Among the differentially expressed Nuphar genes, 2,449 have homologs in the AtGenExpress data set for Arabidopsis of which 957 are also present in the Persea data set. Hierarchical cluste ring of the two matrices generate d gene clusters characterized by up regulation in Nuphar but down-regulation in Arabidopsis and vice versa (Fig. 5-7 A ,) and up-regulation in either Nuphar Persea or both, but down-regulation in Arabidopsis, and vice versa (Fig. 5-7 B ,). We observed that a very small proportion of the genes show similar expression patterns between Nuphar and Arabidopsis (Fig. 5-7A ). Those genes are involved in metabolism and energy production, and this is also found in Nuphar subset 2 ( Nuphar Arabidopsis and Persea ). Floral organs clustered into groups in accordance with species, with Nuphar and Persea flowers transcriptionally closer to each other than either is to Arabidopsis flowers. Also, in both Nuphar and Persea, perianth organs cluster together, follo wed by stamens, as in analyses of the total Nuphar (Fig. 5-7A ) and Persea (Chanderbali et al. submi tted) data sets. However, Arabidopsis petals always cluster with stamens, and sepals are either placed distant from the other floral organs (Fig. 5-7 A ; Nuphar-Arabidopsis data set) or next to the petal/stamen cluster (Fig. 5-7B ; Nuphar-Persea-Arabidopsis data set).
126 Investigation of Spatial Gene Expression Patterns: Three analyses were conducted to examine spatial gene expression patterns across fl oral organs. First, in scatter plots comparing the transcriptional profil es of all pairs of flor al organs (Fig. 5-8), Arabidopsis floral organs are more divergent from each other than those of Nuphar and Persea Petals and stamens share the highest similarity among Arabidopsis organs, but with rather low correlations (r2=0.44), compared to almost linear correspondence between petals (inner tepa ls) and sepals (outer tepals) for Nuphar and Persea perianth organs, correlations of r2=0.93 and 0.96, respectively (Fig. 5-8). Perianth organs and stamens are also highly co rrelated in the two basal angiosperms, with r2=0.79 and 0.73 for Nuphar and r2=0.63 and 0.67 for Persea comparing stamens and outer and inner perianth whorls, respectiv ely. However, low correlation between carpels and other floral organs of Persea are observed (i.e., r2<0.4). Next, to examine the relative expression dom ains of each floral organ, we sorted expression level of genes with positive log2 value according to organ of primary expression. In all three data sets ( Nuphar Nuphar subsets 1 and 2), Nuphar genes showed extended expression domain from one organ to adjacent organs (Fig. 5-9 A-C ). Persea genes also exhibited similar trends to those of Nuphar, while in Arabidopsis there is little spill over into adjacent organs (Fig. 5-9 B C ) (see also Chanderbali et al. submitted). Organ-specific expression in Nuphar in accordance with the criter ia of Chanderbali et al. (submitted), identified 22 carpel-specific genes, 23 specific to stamens, and 11 and eight specific to petals and sepals, respectively. In our search for genes specific to combinations of adjacent floral organs, we identified two genes specific to carpels and stamens, one specific to stamens and petals, 50 specific to petals and sepals, a nd 88 specific to perianth organs and stamens. Similar analyses for subsets 1 and 2 of the Nuphar data set are presented in Fig. 10. In general,
127 we find stronger representation of genes with organ-combination-specific than with organspecific expression in Nuphar with perianth organs plus st amens accounting for ~43% of the observed instances, followed by se pals plus petals (Fig. 5-10 A-B ). In Persea approximately 33% of the genes exhibiting up-regulatio n are co-expressed in outer + inne r tepals and perianth organs + stamens (Fig. 5-10 A-B ). However, in Arabidopsis a relatively low number of genes is coexpressed in carpels /stamens and stamens/petals (Fig. 5-10 A-B ). Discussion Genes Expressed Differential ly in Reproductive Organs From the microarray experiments, we found that in Nuphar the number of genes differentially expressed and up-regu lated in floral tissues relative to leaves (with a minimum of two-fold higher expression), is low relative to the number obtained from gene expression profiles of Persea (Table 5-1). The difference betw een the expression profiles of Nuphar and Persea is mainly due to the low number of gene s expressed in carpels and stamens of Nuphar In fact, most of the genes up-regulated in floral tissues a nd/or stages in Nuphar are found in perianth members, but only a very small number of th e genes are found in ca rpels and stamens in Nuphar In contrast, in Persea a majority of the genes up-regulated in reproductive organs is found in stamens and carpels (Table 5-1). This finding can be explained by floral developmental features of Nuphar Although a cDNA library was constructed us ing the floral buds from premeiotic stages, perianth members of Nuphar at premeotic stages are relati vely larger than stamens and carpels. Therefore, there is a gr eater possibility of including mo re genes related to perianth development in the library. This inference is further supported when we examined the genes upregulated in specific floral organs in the combined data set from Nuphar Arabidopsis and Persea ( Nuphar subset 2). Even though we have the same gene homologs in Nuphar subset 2 for
128 these three species, in Nuphar most of the genes are up-regul ated in perianth members and stamens, while the majority of the genes are up-regulated in stamens and carpels in Persea and carpels in Arabidopsis (Fig. 5-10 B ). Therefore, the difference in the proportions of the genes differentially expressed between Nuphar and Persea can be explained by their different floral morphologies. When we checked for downor up-regulated genes in reproductive organs, we found that some of them may have a different function in Nuphar compared to their function in Arabidopsis For example, AFO is required for floral organ forma tion and development (Chen et al. 1999); thus, it is mainly expressed in the very early stage of floral development. In our experiment, this gene homolog is weakly expressed in Ybd, the stag e at which all initial floral organ formation takes place, so down-regulation in reproductive orga ns is consistent with expression patterns in a later stage. Also, a homolog of BARELY ANY MERISTEM 2 ( BAM2, At3g49670), a regulator of early anther development in Arabidopsis (Hord et al. 2006), is also down-regulated in all reproductive tissues in Nuphar However, since we only incl uded floral tissues before microsporangia initiation, we would not expect to see expression in our samples. Importantly, other gene homologs in Nuphar such as NAP, KAN1 VRN5 and TUF are expressed in different floral organs compared to their function in Arabidopsis Thus, the homologs of these genes may have different functions in Nuphar than in Arabidopsis ; the roles of these genes in flower development in Nuphar should be addressed using other approach. In general, microarray data for homologs of floral organ identity genes in Nuphar confirm expression patterns obtained in ea rlier RQ-RT-PCR experiments (Yoo et al. in prep.). In those instances of inconsiste ncy, microarray data suggested broader expression patterns than RQ-RTPCR data for homologs of AP1, AG and AGL6 Whether these reflect greater sensitivity, un-
129 specific binding, or experimental variation in cDNA populations, rema ins unclear. RQ-RT-PCR uses sequence-specific primer sets to amplify target cDNA, whereas microarray experiments are based on hybridization between probes and target cRNA. Therefor e, if duplicate gene copies with sufficient sequence similar ity exist in the cDNA population, the probe may bind to both and wider gene expression patterns coul d be detected if duplicates co llectively exhibi t broader gene expression domains. Similar resu lts would be obtained if microa rray probes target conserved motifs. Also, expression levels measured by RQ-RT-PCR are relative to that of an internal control, therefore, direct co mparison between these two experiments may only be partially reliable. When we tried to examine the expression patterns of Nuphar AP1 and AGL6 homologs without the internal control, thos e transcripts were detected in a ll floral tissues even though there is variation in signal intensities as shown in microarray data (data not shown). However, as for an AG homolog from Nuphar we still could not detect any AG transcripts in sepals and petals. Closer examination of microarray ex pression levels in flor al tissues relative to leaf tissue, shows that the expression levels are relatively high in stamens and carpels (Fi g. 5-4), suggesting that expression levels in sepal and petals may lie below the thresh old sensitivity of RQ-RT-PCR. However, the broader expression pattern obser ved in microarrays might be due to the nonspecificity of the Nuphar AG probe, since it is designed be tween the MADS-box and I region. However, BLAST analyses of the probes against NCBI nr database only hit other members of the AG gene lineage; therefore, there is a hi gher possibility th at it binds other AG homologs in the transcriptome of Nuphar sepals and petals than more distantly re lated MADS-box genes. Hierarchical Clustering The four main gene clusters correspond to distinct spatial expression patterns. The AG cluster contains the Nuphar homolog of AG and most of the genes with two-fold or higher
130 expression levels in carpels rela tive to leaves (Fig. 5-6). The AG homolog from Nuphar is strongly up-regulated in all floral buds and floral tissues, although its expression is restricted to stamens and carpels in RQ-RT-PCR data (Yoo et al. in prep.). Other genes highly expressed in carpels are related to diverse function s, such as carbon metabolism (At1g23760, At1g48100, At5g66460), stress response (At1g56340, At3g58450), and systemic interaction with the environment (At2g46370, At3g25230). The AP3/ PI / AGL2 / AGL6 module contains the largest number of genes among the four regulatory modules, and AGL6 homolog of Nuphar was separated from the main cluster with other two genes, serine carboxypeptidase-like 45 precursor ( SCPL45, At1g28110) and expressed protein (At2g35880) (Fig. 5-5). Thes e three genes exhibited a twofold higher expression in all floral tissues and carpels rela tive to leaves (Fig. 5-6). In Arabidopsis SCPL45 is expressed everywhere, including seedlings, young leaves, flowers, and fruits (Fraser et al. 2005). However, its expression is very weak in old leaves and mature organs, such as stem and root. A SCPL45 homolog from Nuphar is also up-regulated in all floral organs and medium-sized floral buds, but not in leaves. AGL6 (At2g45650) is particularly highly expres sed in sepals, petals and carpels of Arabidopsis (Schmid et al. 2005), and this expr ession pattern is also shown by its Nuphar homolog (Fig. 5-4). Most of the genes in the AP3/ PI / AGL2 / AGL6 module were expressed primarily in petals, sepals and petals, or both perianth members and stamens (Fig. 5-6). Together with homologs of floral organ identity genes, such as AP3 PI and AGL2 MYB or MYB-related transcription factors are up-regulated in perianth members and/ or stamens relative to leaves (two-fold higher expression level).
131 The AP1 module includes an AP1 homolog of Nuphar and most of the genes in this module were up-regulated in floral buds (Fig. 5-6). Some of th em are involved in meristem activity or flowering transition in Arabidopsis First, AP1 is an A-function gene and helps to regulate sepal identity in Arabidopsis (Bowman et al. 1993), but its inferred function (based on its expression pattern) differs in Nuphar It is up-regulated during early flower development relative to leaves, especially at the pre-microsporangia initiation st age, but not in floral organs. Also, ATSKP1 (At1g75950), which is mainly involved in meristem activity is included in this module. Porat et al. (1998) showed that mRNA of ATSKP1 accumulated in all of the plant meristems, including inflorescence and floral meri stems, and our data is consistent with their observation. The last module contains a PI homolog of Nuphar, and most of the genes showed similar expression patterns to those of genes in the AP3/ PI / AGL2 module. However, genes included in the PI module were also highl y expressed in floral buds. In contrast to a PI homolog in the AP3/ PI / AGL2 module, this PI shows relatively high expression in both floral buds rather than floral tissues of Nuphar Considering we include only one PI homolog in the array, this observation might be due to the non-speci ficity of probes. The probe in the PI module is designed from the MADS-box region, so it can po tentially hybridize to other MADS-box genes that are highly expressed in early floral de velopmental stages. However, the probe in the AP3/PI/AGL2 module is designed from sequences between the K-box and the C-terminal region, so it is more PI -specific and its expression pattern is more accurate. For example, most of the genes from the AP3/ PI / AGL2 / AGL6 module are mainly expressed in perianth members and stamens (Fig. 5-6). However, when we examine gene expression in detail, there can be differences in the gene expression level between the floral organs. For example, AGL6
132 (At2g45650) is highly expressed in all floral organs of Arabidopsis (Schmid et al. 2005), and this expression pattern is also observed in Nuphar However, the expression level of an AGL6 homolog in stamens is half the expression leve l of other floral organs (Fig. 5-4). Similar expression patterns are shown by the AGL6 homologs of Persea and Hordeum from microarray data (Chanderbali et al. submitted; Druka et al. 2006) In Persea two AGL6 homologs are highly expressed in outer and inner tepa ls and carpels, but s how relatively low expression in stamens (Chanderbali et al. submitted). Also, an AGL6 homolog of Hordeum is up-regulated in inflorescence, bracts, and carpels, but the expression level is three times less in anthers (Druka et al. 2006). These data indicate that AGL6 homologs in basal angiosperms might differ in their function from that of the model eudicot Arabidopsis In Arabidopsis there are two paralogs in the AGL6 lineage, AGL6 and AGL13 (Martinez-Castilla and AlvarezBuylla 2003; Rounsley et al. 1995): AGL6 is expressed in inflorescence buds (Dua rte et al. 2006; Ma et al. 1991), whereas AGL13 is expressed in ovules only (Rounsley et al. 1995). Thus, before the duplication event, one copy of AGL6 might have played a role in both perianth and ovule development. For example, Mena et al. (1995) showed that two AGL6 homologs, ZAG3 and ZAG5 are present in maize, and ZAG3 functions in inflorescence and carpe l development although this gene is expressed in sterile floral organs that correspond to sepa ls and petals of eudicots. ZAG5 is homologous to ZAG3 so it is suggested that these two genes were duplicated via polyploidy and subsequently diverged (Mena et al. 1995). However, the function of ZAG5 is not well known. Thus, we cannot tell whether divergence in function of AGL6 homologs occurs in maize. In other monocots, Oncidium and Hyacinth AGL6 homologs are detected in floral buds, perianth members (lip or tepals), and carpels, but not in stamens (Fan et al. 2007 ; Hsu et al. 2003). In addition, through transgenenic experime nts, these tw o genes have been shown to play a role in
133 the regulation of flower transi tion and organ formation, as does AP1 in Arabidopsis (Fan et al. 2007; Hsu et al. 2003). In Nymphaeales, AGL6 homologs from Nymphaea and Cabomba are also expressed in perianth members a nd carpels (Yoo et al. in prep.). In other basal angiosperms, AGL6 transcripts were detected only in perianth members ( Magnolia and Liriodendron Kim et al. 2005, Persea, Chanderbali et al. 2006). Together with these reported expression patterns, AGL6 homologs might contribute to perianth de velopment in basal angiosperms (and monocots) (as a putative A-function gene) because they are ma inly expressed in perianth members (Kim et al. unpubl. data; Yoo et al. in prep). However, AGL6 homologs are also involved in carpel development in most basal angiosperms and mono cots; thus, more extensive study of expression throughout the angiosperms is needed to evaluate the idea of AGL6 homologs as A-function genes which promote perianth organ identity in basal angiosperms. In the AP3/ PI / AGL2 / AGL6 module, together with homologs of floral organ identity genes, such as AP3, PI and AGL2 MYB or MYB-related transcript ion factors are up-regulated in perianth members and/or stamens relative to le aves (two-fold higher ex pression level). This result is consistent with th e study of Henning et al. (2004), which showed that many potential signaling components, such as protein kinases, tr anscription factors, and other putative signaling proteins, were up-regulated during reproductive development. However, most genes in this module do not have Arabidopsis homologs, so it is hard to infer their function. Another MADS-box gene homolog exhibiting a different expression pattern from its function in Arabidopsis is an AP1 homolog of Nuphar Originally, AP1 was identified as an Afunction gene to regulate sepal identity in Arabidopsis (Bowman et al. 1993). However, a recent study of ap1 mutants of Arabidopsis suggests that AP1 function is not essential for sepal and petal development: over-expression of AGL24 seems to be responsible for defects of sepals and
134 petals in the ap1-1 mutants. Thus, in the absence of AGL24 ap1 mutants partially recover their wild type phenotypes (Yu et al. 2004). Therefore, together with the expression patterns of AP1 homologs from other eudicots such as Petunia (Rijpkema et al. 2006), Antirrhinum (Davies et al. 2006), and Gerbera (Teeri et al. 2006), the identity of th e true A function gene is questionable. The AP1 homolog from Nuphar also displayed different expression patterns from AP1 in Arabidopsis The AP1 homolog from Nuphar is up-regulated during early flower development relative to leaves, especially at the pre-microsporangia initiation st age, but not in floral organs. This expression pattern is c onsistent with several recen t studies, suggesting that AP1 homologs may not be floral organ identity genes (Chen et al. 2008; Huijser et al. 1992; Shan et al. 2007; Taylor et al. 2002; Zhang et al. 2008). Rather than be ing expressed only in sepals and petals, most AP1 homologs are expressed in both vegetative and reproductive organs, showing the highest expression in inflorescen ce and floral meristem (Shan et al. 2007). Our microarray data and RQ-RT-PCR data (Yoo et al. in prep.) also show a similar pattern; the AP1 homolog is expressed in both vegetative and reproductive organs, exhibiting the highest expression in floral buds. Thus, considering that AP1 is involved in both vegetative and reproductive development in other species, our result for Nuphar is consistent with other studi es (Chen et al. 2008; Huijser et al. 1992; Shan et al. 2007; Ta ylor et al. 2002; Zhang et al. 2008) (Shan et al. 2007). The four modules obtained in the Nuphar floral transcriptome a ppear to exploit similar gene contents, for example, MADS-box genes an d other transcription f actors, to those of Arabidopsis and suggest their functional conservati on in major regulatory elements between these two species (Fig. 5-6). However, we have only focused this discussion on genes related to floral meristem and organ forma tion, so further comparison between Nuphar and Arabidopsis is required (see below).
135 Comparative Floral Transcriptomics Hierarchical Clustering: From Nuphar subsets 1 and 2, we found that the majority of the genes showing similar expression patterns is involved in metabolism and energy production. Also, MADS-box gene homologs from Nuphar, Persea and Arabidopsis exhibit similar expression patterns (Fig. 5-7 B ). However, other genes regulating flower development, such as NAP, AFO are expressed in different fl oral organs in the three sp ecies, indicating that those genes function differently in these three species. This results suggest that even though the major elements in floral organ forma tion work in similar ways among these species, other genes acting downstream of these key floral regulators act di fferently in these three species. Therefore, modification and refinement in downstream ge ne function may be responsible for floral diversification acro ss the angiosperms. Arrays also clustered into three groups that correspond to species. In contrast to Nuphar and Persea which perianth members are cluste red together with stamens, in Arabidopsis petals and stamens are clustered together with carpels in Nuphar subset 1 and with sepals in Nuphar subset 2 (Fig. 5-7 A, B ). Considering that the branch lengths between the cluster of petals and stamens and other organs are relativ ely short, this difference may be meaningless. Petals in basal angiosperms are thought to have been derived fr om bracts (bracteopetals ), while petals of eudicots are considered to be derived from of stamens (andropetals) (Endress 2001; Hiepko 1965; Takhtajan 1991). However, th e petals (inner tepals) of Nuphar and Persea are less differentiated from the sepals (outer tepals) than observed in most eudicots, so the petals of these basal taxa may be not homologous to petals of Arabidopsis Furthermore, the two perianth whorls of Nuphar and Persea clusterd with stamens, indicating that the perianth organs of these two species may have originated from stamens (i .e., the petals would therefore be andropetals).
136 Although this result contradicts the classic view of bracteopetals in basal angiosperms, other studies also supported the idea of andropetals in Nuphar (Yoo et al. in prep.) and Persea (Chanderbali et al. 2006). Investigation of Spatial Gene Expression Patterns: To examine whether there is any relationship between gene expressi on patterns of floral organs, we drew scatter plots using the transcriptional profiles of floral organs in Nuphar, Persea, and Arabidopsis We observed a strong linearity between perianth members of Nuphar and Persea, and this similarity in transcriptional profile might reflect their largely undiffere ntiated morphology, although their position and some subtle morphologi cal characters at least in Nuphar can distinguish these two whorls. However, Persea differs from Nuphar in the low correlation between carpels and other floral organs in the former (i.e., r2<0.4), suggesting different flor al transcriptomic programs are exploited in carpels. In contrast to Nuphar and Persea, Arabidopsis showed a very weak correlation among the floral organs (0.07
137 The trends observed above were shown in a numerical way. There are large differences in the genes strongly expressed in one floral organ, but with lower expression in all other floral organs among three species. For Nuphar a large number of genes is expressed in petals/sepals and petals/sepals/stamens, indicating that in Nuphar these three organs share very similar patterns of expression. Significan tly, 88 genes are shared by these three organs, indicating a strong similarity among three organs. A similar pattern is observed in Persea ; approximately 30% of the genes with up-regulation are co-expresse d in outer and inner tepals and both perianth members and stamens (63 and 67 genes, respectively; Fig. 5-10 A ). In Persea however, 105 genes are up-regulated only in stam ens, indicating that stamens of Persea have a more unique transcriptome than do stamens of Nuphar (see below). In Arabidopsis a relatively low number of genes is co-expressed in carpels/stamens and stamens/petals (Fig. 5-10 A B ). In particular, more than 2.5 times many genes are co-expressed in petals/stamens compared to the number of gene s co-expressed in petals /stamens, supporting the idea of andropetal orgin in eudicot. The three species compared here ( Nuphar Persea, and Arabidopsis ) show differences in spatial expression patterns of flor al genes, and those patterns are correlated with different floral morphologies. For example, Arabidopsis a derived core eudicot, has well-differentiated floral organs, and the number of genes expressed in more than one floral organ is very small, indicating that each floral organ has a uni que floral transcriptional program. This likely typifies most eudicots. In Persea a magnoliid, similar floral transc riptional programs characterize the morphologically undifferentiated inner and outer te pals and these programs differ from those for carpels and stamens. Although stamens share a nu mber of genes with perianth members, many genes are exclusively up-regulated in stamens of Persea. This pattern in Persea may be
138 explained by morphology; Persea stamens possess well-differentiated anthers and filaments like Arabidopsis so more genes are possibly associated with microsporogenesis. In Nuphar which exhibits only modest differentiation between sepals and petals, petals appear to have a slightly different floral transc riptional program from sepals. Howe ver, the high proportion of genes expressed in both petals and se pals also implies that differe ntiation between them is not extensive; similarly, the morphological distinction between sepals a nd petals is also slight in Nuphar (see Warner et al. 2008). In addition, the higher proportion of genes commonly expressed in perianth members and stamens co mpared to the number of genes expressed exclusively in stamens suggests that the fl oral developmental program in stamens of Nuphar is not unique, as it is in Persea and Arabidopsis We infer that differences among these species in spatial gene expression patterns am ong floral organs are consistent with their different floral morphologies. In summary, the similarities of transcripti on profiles among floral organs indicate that Nuphar and Persea exploit the floral transcriptome in a similar way while Arabidopsis employs a much more divergent and refined program, which a ppears to have originated well after the origin of the flower itself. However, Nuphar and Persea also show differences in some spatial gene expression patterns, although they are similar in some aspects of floral morphology. For example, both share the same merosity (trimery in both species) and largely undif ferentiated perianth members. In Nuphar most of the floral genes are invol ved in the development of perianth members and stamens, and the floral transc riptional programs of these organs overlap substantially. However, Persea also exhibits differe nt patterns from Nuphar ; although inner and outer tepals share similar ge ne expression patterns in Persea most of the genes are also
139 expressed in stamens and carpels. Thus, each flor al organ (tepals, stamens, and carpels) of Persea is under a relatively unique transcriptional program compared to Nuphar Floral organ identity genes, such as PI AP3, AG and AGL6 showed similar expression patterns in Nuphar Persea and Arabidopsis but most of the genes inve stigated exhibited different expression patterns (Fig. 5-7). This result suggests that even though the major elements in floral organ formation work in similar ways among sp ecies, other genes acting downstream of these key floral regulators ac t differently in these three speci es. Therefore, modification and refinement in downstream gene function may be responsible for floral di versification across the angiosperms.
140Table 5-1. Number of genes up-regulated (two-fold higher expression level) in Nuphar floral tissue relative to leaves Organ No. of genes Organ No. of genes Sepals 43 Petals 113 Sepals + Carpels 2 Petals + Fruits 1 Sepals + Fruits 4 Petals + Stamens 32 Sepals + Petals 101 Petals + Stamens + Carpels + Fruits 1 Sepals + Petals + Carpels 5 Petals + Stamens + Fruits 3 Sepals + Petals + Fruits 13 Stamens 73 Sepals + Petals + Stamens 124 Stamens + Carpels 2 Sepals + Petals + Stamens + Carpels 16 Stamens + Carpels + Fruits 1 Sepals + Petals + Stamens + Carpels +Fruits 25 Stamens + Fruits 13 Sepals + Petals + Stamens + Fruits 48 Carpels 23 Sepals + Stamens 30 Carpels + Fruits 7 Sepals + Stamens + Fruits 10 Fruits 65 Buds 76
141Table 5-2. Genes down-regulated in reproductive organs of Nuphar. Genes in red are involve d in leaf morphogenesis. TFs=transcription factors, Mbd: medium -aged floral buds at the pre-meiotic st age, Ybd: young floral buds at the premicrosporangia initiation stage. Ath Homolog Descript ion Expression area At5g42700 ABI3-VP1 family (TFs ) All reproductive organs At5g58280 ABI3-VP1 family Sepal, petal, stamen At2g41710 AP2-EREBP Family All reproductive oragns except Mbd At4g37750 AINTEGUMENTA (ANT), AP2-EREBP fa mily (TFs) All reproductive organs At1g30330 ARF6, ARF family (TFs) All re productive oragns except carpel At5g62000 Auxin Responsive Factor 2 (ARF2), ARF family (TFs) All floral tissues At1g05805 AtbHLH128, bHLH family (TFs) All re productive oragns except sepal, petal At1g22490 AtbHLH94, bHLH family (TFs) A ll reproductive oragns except Ybd At1g63650 ENHANCER OF GLABRA 3 (EGL3, AtbHLH3), bHLH family (TFs) All reproductive oragns except Ybd At3g26744 INDUCER OF CBF EXPRE SSION 1 (ICE1, AtbHLH116), bHLH family (TFs) All reproductive organs At4g01460 AtbHLH57, bHLH family (T Fs) All reproductive organs At4g09820 TRANSPARENT TESTA 8 (TT8, AtbHLH42), bHLH family (TFs) Ybd, sepal, petal, stamen At1g06070 AtbZIP69, bZIP family (TFs) Sepal, petal, stamen At2g26580 YABBY5, C2C2-YABBY family (T Fs) Ybd, stamen, carpel, fruit At2g45190 ABNORMAL FLORAL ORG ANS (AFO, YAB1, FIL), C2C2-YABBY family (TFs) All reproductive oragns except Ybd At3g44750 C2H2 family (TFs) Ybd, sepal, petal, stamen At3g48430 C2H2 family (TFs) Sepal, petal, stamen At3g12680 ENHANCER OF AG-4 1 (HUA1), C3H family (TFs) All floral tissues At1g52150 ATHB15, Homeobox family (TFs ) All reproductive organs At1g62990 KNOTTED-LIKE HOMEOBOX OF ARABIDOPSIS THALIANA 7 (KNAT7), Homeobox family (TFs) All reproductive organs At1g73360 HOMEODOMAIN GLABROUS11 (HDG11), Homeobox family (TFs) All reproductive organs At2g34710 PHABULOSA (PHB), Homeobox family (TFs) All reproductive oragns except Ybd
142Table 5-2. Continued Ath Homolog Descript ion Expression area At3g61150 HOMEODOMAIN GLABROUS1 (HDG1), Home obox family (TFs) All floral tissues At5g11510 AtMYB3R4, MYB family (TFs ) All reproductive organs At5g53200 TRIPTYCHON (TRY), WRKY family (TFs ) All floral tissu es except fruit At1g08560 SYNTAXIN OF PLANTS 111 (SYP111, KNOLLE) All reproductive organs At1g09570 PHYTOCHROME A (PHYA) All floral tissues At1g48410 ARGONAUTE 1 (AGO1) All re productive organs except Mbd At1g70940 PIN-FORMED 3 (PIN3) All re productive oragns except Ybd At2g19520 FVE All reproductive organs At2g23380 CURLY LEAF (CLF) All repr oductive oragns except fruit At2g42260 UV-B-INSENSITIVE 4 (UVI4) All reproductive oragns except Ybd AT3G15670 late embryogenesis abundant protein (LEA protein) all floral tissues At3g19820 DWARF 1 (DWF1) All reproductive oragns except Ybd At3g49670 BARELY ANY MERISTEM 2 (BAM2) All reproductive oragns except Ybd At4g08980 FBW2 all floral tissues At4g18710 BRASSINOSTEROID-INSENSITIVE 2 (BIN2) All floral tissu es except fruit At4g32551 LEUNIG (LUG) A ll floral tissues At4g39400 BRASSINOSTEROID INSENSITIVE 1 (BRI1) All reproductive oragns except Ybd At5g08370 Arabidopsis thaliana ALPHA-GALACTOSID ASE 2 (AtAGAL2) All reproductive organs At5g58230 MULTICOPY SUPRESSOR OF IRA1 (MSI1) All floral tissues At5g64630 FASCIATA 2 (FAS2) All reproductive oragns except Ybd At2g42200 SQUAMOSA PROMOTER BINDI NG PROTEIN-LIKE 9 (SPL9), SBP family (TFs) All reproductive oragns except Ybd At3g60030 SPL12, SBP family (TFs) All reproductive oragns except Ybd At5g50670 SPL13, SBP family (TFs) All reproductive oragns except Ybd At5g08330 TCP family (TFs) Stamen, carpel, fruit
143Table 5-2. Continued Ath Homolog Descript ion Expression area At2g37260 TRANSPARENT TESTA GLAB RA 2 (TTG2, AtWRKY44), WRKY family (TFs) All reproductive organs At4g26640 AtWRKY20, WRKY family (TFs) All floral tissues At5g26170 AtWRKY50, WRKY family (TFs) A ll reproductive oragns except Mbd At5g28650 AtWRKY74, WRKY family (TFs) All floral tissues At4g24660 ATHB22, ZF-HD family (TFs ) All reproductive organs
144Table 5-3. Genes up-regulated in reproductive organs of Nuphar. Genes in red are involved in flower development. TFs=transcription factors, Mbd: medium-aged floral buds at the pre-meiotic stage, Ybd: young floral buds at the pre-micr osporangia initiation stage. Ath Homolog Descript ion Expression area At1g22490 AtbHLH94, bHLH family (TFs) Ybd At2g20180 PHY-INTERACTING FACTOR 1 (PIF1, AtbHLH15), bHLH family (TFs) All reproductive organs except fruit At3g23210 AtbHLH34, bHLH family (TFs) A ll reproductive organs except Ybd At1g58110 bZIP family (TFs) Ybd At2g40950 AtbZIP17, bZIP family (TFs) Ybd At3g23210 PHYTOCHROME INTERACTING FACTOR 3-LIKE 5 (PIL5), bZIP family (TFs) Mbd, Fruits At3g58120 AtbZIP61, bZIP family (TFs) Ybd At5g06950 AtbZIP20 (AHBP-1B), bZIP family (TFs) Ybd At5g39660 CYCLING DOF FACTOR 2 (CDF2), C2C2-Dof family (TFs) Mbd, Ybd At1g02040 C2H2 family (TFs) Stamens At5g14140 C2H2 family (TFs) Petal, stamen, fruits At1g30500 CCAAT-HAP2 family (TFs) Sepal, petal, stamen At2g01060 G2-like family (TFs) Mbd At5g16560 KANADI1 (KAN1), G2-like family (TFs) Petal, stamen, carpel At1g27050 ARABIDOPSIS THALIANA HOMEO BOX PROTEIN 54 (ATHB54), Homeobox family (TFs) Sepal, petal, stamen At1g69120 APETALA1, MADS family (TFs) Mbd, Ybd At2g45650 AGAMOUS-LIKE6, MADS family (TFs) All reproductive organs At3g54340 APETALA3, MADS family (T Fs) All reproductive organs At4g18960 AGAMOUS, MADS family (TFs ) All reproductive organs At5g15800 AGAMOUS-LIKE2, MADS family (TFs) All reproductive organs At5g20240 PISTILLATA, MADS family (TFs) All reproductive organs At3g47600 AtMYB94, MYB family (TFs) Sepal, petal, stamen, fruit
145Table 5-3. Continued Ath Homolog Descript ion Expression area At1g01060 LATE ELONGATED HYPOCOTYL 1 (LHY1), MYB-re lated family (TFs) Sepal, petal, stamen At1g74840 MYB-related family (TFs) Stamen, carpel AT3G09600 MYB-related family (TFs) Sepal, petal At5g52660 MYB-related family (TFs) All reproductive orgasn except fruit At1g69490 NAC-LIKE, ACTIVATED BY AP3/PI (NAP), NAC family (TFs) Mbd, Ybd At1g10940 ARABIDOPSIS SERINE/THREONINE KINASE 1 (A SK1) All reproductive organs except Ybd At1g19270 DA1, LIM domain-containing pr otein Sepal, petal, stamen At3g15354 SPA1-RELATED 3 (SPA3) All floral tissues except fruit At1g10670 ACLA-1 Sepal, petal, stamen At3g24440 VERNALIZATION 5 (VRN5), a PHD fi nger protein Sepal, petal, stamen At4g11150 VACUOLAR ATP SYNTHASE SUBUNIT E1 (T UF) All floral tissues except fruit
146 Figure 5-1. Summary tree for angiosperms. This tree is modified from the plastid genome trees of Jansen et al. (2007) an d Moore et al. (2007), and onl y a few representatives of asterids and rosids are included. Nuphar advena belongs to Nymphaeales (red) which are the sister to all extant angiosperms except Amborella (Amborellales).
147 Figure 5-2. Flower of Nuphar advena. A Side view of whole flower. B Three sepals in the first whorl and three petals in the second whorl. C staminodes with nectary, adaxial (left) and abaxial (right) view. D stamens with anther, adaxial (two left) and abaxial (two right) view.
148 Figure 5-3. Double loop design for Nuphar microarray experiments. Each line indicates a single array, and different color means different bi ological replicates. The tissue at the head of the arrow is labeled red, and the ti ssue at the tail is labeled green.
149 Figure 5-4. Comparison of microarr ay data with RQ-RT-PCR-data. A Microarray data of AP1, AP3-a AP3-b PI-a and PI-b For AP3 and PI two different probes were designed from one sequence. B RQ-RT-PCR data of AP1 (Yoo et al., in prep.), AP3-1, AP3-2 and PI (Kim et al., 1995). C Microarray data of AG AGL2-a AGL2-b and AGL6 For AGL2 two different probes were designed from one sequence. D RQ-RT-PCR data of AG (Yoo et al., in prep.), AGL2 (Zahn et al., 1995), AGL6 (Yoo et al., in prep). For all RQ-RT-PCR data, the expression level was relative to 18S rRNA internal control, so the maximum expression level is 1.
150 Figure 5-4. Continued.
151 Figure 5-5. Hierarchical clusteri ng displays up-regulated or down-regulated gene clusters based on similarity of expression patterns. Identified five regulatory modules are sequentially enlarged in the next of dendrogram. The color scale is presented below dendrogram; red indicates up-regulation, and green shows down-regulation in each floral organ relative to leaf. Top, array tree; left, gene tree. Mbd: medium-aged floral buds at the pre-meiotic stage, Ybd: young buds at the pre-microsporangia initiation stage.
152 Figure 5-6. Number of genes up-regulated in Nuphar floral tissue at identified four regulatory modules.
153 Figure 5-7. Hierarchical clustering results of combined data sets of Arabidopsis Nuphar and Persea data set. A clustering result from comparison of Nuphar with Arabidopsis only. I: up-regulation of Arabidopsis genes, II: down-regulation of Arabidopsis genes. B clustering result from comparison of Nuphar with Persea and Arabidopsis I: down-regulation of Arabidopsis genes, II: up-regulation of Arabidopsis genes. The color scale is presented below dendrogram ; red indicates up-regulation, and green shows down-regulation in each floral organ re lative to leaf. Top, array tree; left, gene tree.
154 Figure 5-8. Scatter plots of spatial gene expression patterns in floral organs of Nuphar, Persea, and Arabidopsis For this graphs we used the whole data set from Nuphar Persea (Chanderbali et al. submitted), and Arabidopsis (Schmid et al. 2005). The log2 transformed expression data were used, so the positive va lue indicates up-regulation and the negative value indicates down-regulation in each floral organ relative to leaf. The r2-value of the regression lin e (blue) is presented.
155 Figure 5-8. Continued.
156 Figure 5-9. Expanse of gene expressi on domain across floral organs of Nuphar Persea and Arabidopsis Genes overexpressed in each floral or gan compared to other are sorted by their positive value accordi ng to primary expression domain. A from the whole data set of Nuphar, Persea (Chanderbali et al. submitted), and Arabidopsis (Schimid et al. 2005). B from Nuphar Persea and Arabidopsis homologues. SE: sepals, PE: petals, ST: stamens, CA: carpels, OT: outer te pals, IT: inner tepals. The color scale is presented at the bottom; red indicat es up-regulation, and green shows downregulation in each floral organ relative to leaf.
157 Figure 5-9. Continued.
158 Figure 5-10. Number of genes with at least two-fold up-regulated in one organ a nd less than 1.5-fold in all other floral organs of Nuphar Persea and Arabidopsis A genes up-regulated in organspecific and organ-combinati on from the whole data set of Nuphar, Persea (Chanderbali et al. submitted), and Arabidopsis (Schimid et al. 2005). B genes up-regulated in organspecific and organ-combination from from Nuphar, Persea, and Arabidopsis homologues.
159 CHAPTER 6 CONCLUSIONS The goals of this study were to understand th e phylogeny and timing of diversification of the waterlily clade (Nymphaeales) as well as the floral development and developmental genetics in this clade. Three species from Cabombaceae and Nymphaeaceae were used to construct floral developmental series for these families and to examine floral gene expression patterns. One species, Nuphar advena (Nymphaeaceae), was chosen for detailed floral transcriptome analysis using microarrays. First, to elucidate the ages of Nymphaeales lineages, we es timated their divergence times. Using published nuclear 18S rDNA and plastid rbcL and matK DNA sequences and a published topology for Nymphaeales, we estimated the diverg ence times of genera in this clade. We applied four different methods, a strict molecular clock, nonparametric rate smoothing (NPRS), penalized likelihood (PL), and a Bayesian method, to estimate divergence times. We calibrated the trees by using the minimum age of the angi osperm crown group constrained to 131.8 mya. Our results indicate that extant Nymphaeales dive rsified into two major clades corresponding to Cabombaceae and Nymphaeaceae during the Eocene (44.6.9 mya); extant genera of Nymphaeaceae date to 41.1.7 mya, and extant Cabombaceae diversifie d during the Miocene (19.9.6 mya). Whereas the stem lineage of Nymphaeales is old based on fossil evidence (125115 mya), our results indicate that extant Nymphaeales diversified relatively recently. In another set of analyses we used PL to estimate th e age of the angiosperms using two prominent Nymphaeales fossils as calibration points. These analyses suggest that these Nymphaeales fossils may be attached at deeper nodes than pro posed in earlier studies, that is, these fossils may not represent Nymphaeales. Using dispersal-vicaria nce analysis, we infer that the ancestor of Nymphaeales occupied the American and Eurasian continents during the Eocene and that the
160 present distributional patterns re quire several subsequent dispersal and extinction events. This biogeographic inference is s upported by the fossil record. In chapter 3, we examined the floral developmental series of three species ( Cabomba caroliniana Nuphar advena, and Nymphaea odorata ) and aligned them. Although there are previous studies of flower development of Ny mphaeales, comparative analyses have not been conducted. We found that the three study species follow the same 10 developmental stages in the very same order. Nuphar which is sister to other members of Nymphaeaceae, exhibits Nymphaea-like or Cabomba -like developmental features, and its phylogenetic position suggests that sharing features between Nuphar and Cabomba might be ancestral ch aracteristics to core Nymphaeales. The problematic origin of the third whorl of floral organs in Nuphar is investigated in detail developmentally and morphologically. Comparison of these organs with stamens of Nuphar indicates that third-whor led organs originated from stamens. However, when compared with perianth organs from Cabomba and Nymphaea, it is found that those organs may be not homologous with perianth organs. For eluc idating its organ ident ity, more comprehensive genetic research is required fo r other members of Nymphaeales. In chapter 4, to provide insights into the fl oral developmental genetics of Nymphaeales, I investigated the expressi on patterns of floral organ identity genes representing major lineages of MADS-box genes using relativ e quantitative RT-PCR in Cabomba, Nuphar and Nymphaea. Because of the similarity in floral struct ure between Nymphaeaceae and the basal eudicot Nelumbo, we conducted the same experiments in the latter as well. I focu sed on (1) perianth differentiation in all species exam ined and (2) the transition of pe taloid staminodes to stamens in Nymphaea. For B-class gene homologues, in Cabomba, expression patterns fit the sliding boundaries model, whereas Nuphar and Nymphaea have generally broad patterns of gene
161 expression. Other gene homologues from these species follow the classic ABCDE eudicot model except AP1 homologues. In contrast, the expression of floral ge ne homologues in the basal eudicot Nelumbo follows the classic AB CDE eudicot model except AP1 homologue. Considering morphology together with developmen tal and genetic data, sepaloid or petaloid tepals are more appropriate terms than sepals or petals for pe rianth organs of Nymphaeales. Also, the gene expression patterns in the transi tional sequence of stamens of Nymphaea suggest that petals of Nymphaea originated from petaloid staminodes or stamens. Despite the superficial floral similarity of Nelumbo to Nymphaea, expression of floral genes in Nelumbo differs from that of Nuphar Nymphaea, and other basal angiosperms. In addi tion, based on expression data obtained from this study, I infer that th e out-of-male hypothesis, deriva tion of floral parts from male structures, might be true fo r ancestors of Nymphaeaceae. In chapter 5, I investigated the floral transcriptomes of Nuphar advena in detail. Most of our understanding of the floral transcriptome comes from anal yses of model eudicots and monocots. I present the first floral transcriptom e profiling of one of the basalmost angiosperms, Nuphar advena (Nymphaeaceae). I used custom microarrays from Agilent Technologies containing 60-mer oligonucleotide probes, and representing approximately 6,200 unique Nuphar floral transcripts obtained from the Floral Ge nome Project (http://fgp.bi o.psu.edu/fgp/index.html). We investigated gene expression in two floral buds from differen t floral developmental stages, sepals, petals, stamens, carpels, fruits, and l eaves using a double loop design. I id entified 3,333 floral transcripts that were significantly differentially expr essed among eight tissues (with a significance of P<0.05; FDR=0.0381), and 1,624 ge nes showed a minimum of two-fold differential expression in at leas t one of the reproductive tissue rela tive to leaves. In particular, homologues of AGAMOUS-LIKE 6 ( AGL6 ), AGAMOUS ( AG ), APETALA3 ( AP3), PISTILLATA
162 ( PI ) and AGAMOUS-LIKE 2 ( AGL2 ) were up-regulated in floral tissues compared to leaves. These results are consistent with the expression patterns previously reported for these tissues in Nuphar using RQ-RT-PCR. Hierarch ical clustering was performe d to group tissues and genes based on similarities in gene expression patterns, and perianth members clustered together with stamens in hierarchical clusteri ng, indicating a strong similarity in their gene expression profiles. The analysis identified four regulatory m odules, and each module exhibited a distinctive expression pattern. I compared the floral transcriptome of Nuphar with those of the core eudicot Arabidopsis and the basal anigiosperm Persea (Lauraceae) with regard to spatial gene expression patterns and correlation of genes expresse d in each floral tissue. I found that Nuphar has a less well defined floral transcript ional program compared to Arabidopsis Floral transcriptional programs of sepals, petals, and st amens are not well-defined in Nuphar There is overlap in gene expression among these organs. Similar results have been reported for Persea other basal angiosperm. In Arabidopsis in contrast, each floral organ has a well defined transcriptional program. These findings are in agreement with the floral morphologies and phylogenetic positions of these three species. Furthermore, most floral organ identity genes exhibited similar expression patterns among the thr ee species, but genes involved downstream of floral organ formation and other pathways show distinct expression patterns. Theref ore, modification and refinement in spatial expression patterns of thos e genes may, in part, be responsible for floral diversification across angiosperms. During the period of this researh, Hydate llaceae was added to Nymphaeales based on molecular phylogenetics. Therefore, future deve lopmental genetics work should include this new member of this clade. First, estimation of divergence times in Nymphaeales should be done including this new member. Hydatellaceae is sister to Cabombaceae + Nymphaeaceae, and they
163 are found in India, Australila, and New Zeala nd. Thus, inclusion of th is family may alter divergence time estimates of the major lineages and also impact inferences on historical biogeography in Nymphaeales. Also, Hydatell aceae exhibit unique floral developmental features: the presence of involucral bracts, no pe rianth, and separate staminate and carpellate flowers. Therefore, application of the same approaches performed in this study will shed additional light on floral development geneti cs in Nymphaeales and, further, in early angiosperms.
164 LIST OF REFERENCES Adam, H, S Jouannic, Y Orieux, F Morcillo, F Richaud, Y Duval, JW Tregear 2007 Functional characterization of MADS box genes involved in the determination of oil palm flower structure. J Exp Bot 58: 1245-1259. Albert, VA, MHG Gustafsson, LD Laurenzio 1998 Ontogenetic systematic, molecular developmental genetics, and the angiosperm petal. in : DE Soltis, PS Soltis, JJ Doyle eds. Molecular systematic of plants II DNA sequencing. Kluwer Academic Publishers, Dordrecht, The Netherlands. Albert, VA, DE Soltis, JE Carlson, WG Farmerie, PK Wall, DC Ilut, TM Solow, LA Mueller, LL Landherr, Y Hu, M Buzgo, S Kim, MJ Yoo, MW Frohlich, R Perl-Treves, SE Schlarbaum, BJ Bliss, X Zhang, SD Ta nksley, DG Oppenheimer, PS Soltis, H Ma, CW DePamphilis, JH Leebens-Mack 2005 Floral gene resources from basal angiosperms for comparative genomics research. BMC Plant Biol 5: 5. Alves-Ferreira, M, F Wellmer, A Banhara, V Kumar, JL Riechmann, EM Meyerowitz 2007 Global expression profiling applied to the analysis of Arabidopsis stamen development. Plant Physiol 145: 747-762. Angenent, GC, J Franken, M Busscher, L Co lombo, AJ van Tunen 1993 Petal and stamen formation in petunia is regulated by the homeotic gene fbp1. Plant J 4: 101-112. Anzotegui, M 2004 Megaflora of the chiquimil form ation (Late Miocene) in the Santa Maria and Villavil valleys, Catamarca and Tucuman provinces, Argentina. Ameghiniana 41: 303-314. APG 1998 An ordial classification for the families of flowering plants. Ann Mo Bot Gard 85: 531-553. APG II 2003 An update of the Angiosperm Phylogeny Group classification for the orders and families of flowering plants. Bot J Linn Soc 141: 339-436. Arora, R, P Agarwal, S Ray, AK Singh, VP Singh, AK Tyagi, S Kapoor 2007 MADS-box gene family in rice: genome-wide identificati on, organization and expression profiling during reproductive development and stress. BMC Genomics 8: 242. Baldwin, BG, MJ Sanderson 1996 Age and rate of di versification of the Hawaiian silversword alliance (Compositae). Proc Na tl Acad Sci U S A 95: 9402-9406. Baum, DA, LC Hileman 2006 A developmental gene tic model for the origin of the flower. in : C Ainsworth ed. Flowering and its manipul ation. Blackwell Publishing, Sheffield, UK. Becker, A, K Kaufmann, A Fr eialdenhoven, C Vincent, MA Li, H Saedler, G Theissen 2002 A novel MADS-box gene subfamily with a siste r-group relationship to class B floral homeotic genes. Mol Genet Genomics 266: 942-950.
165 Becker, A, H Saedler, G Theissen 2003 Distinct MADS-box gene expression patterns in the reproductive cones of the gymnosperm Gnetum gnemon Dev Genes Evol 213: 567-572. Becker, A, G Theissen 2003 The major clades of MADS-box genes and their role in the development and evolution of flowering plants. Mol Phylogen et Evol 29: 464-489. Becker, A, KU Winter, B Meyer, H Saedler, G Theissen 2000 MADS-Box gene diversity in seed plants 300 million years ago. Mol Biol Evol 17: 1425-1434. Becker, S, S Gleissbergy, DR Smyth 2005 Floral and vegetative morphogenesis in California poppy ( Eschscholzia californica Cham.). Int J Plant Sci 166: 537-555. Bell, CD, DE Soltis, PS Soltis 2005 The age of the angiosperms: a molecular timescale without a clock. Evolution 59: 1245-1258. Benlloch, R, I d'Erfurth, C Ferrandiz, V Co sson, JP Beltran, LA Canas, A Kondorosi, F Madueno, P Ratet 2006 Isolation of mtpim proves Tnt1 a useful reverse genetics tool in Medicago truncatula and uncovers new aspects of AP1-like functions in legumes. Plant Physiol 142: 972-983. Birnbaum, K, DE Shasha, JY Wang, JW Jung, GM Lambert, DW Galbraith, PN Benfey 2003 A gene expression map of the Arabidopsis root. Science 302: 1956-1960. Borsch, T 2000 Phylogeny and evolution of the genus Nymphaea (Nymphaeaceae). Dissertation. Universitt Bonn, Bonn, Germany. Borsch, T, KW Hilu, D Quandt, V Wilde, C Neinhuis, W Barthlott 2003 Noncoding plastid trnTtrnF sequences reveal a well reso lved phylogeny of basal angiosperms. J Evol Biol 16: 558-576. Borsch, T, KW Hilu, JH Wiersema, C L hne, W Barthlott, V Wilde 2007 Phylogeny of Nymphaea (Nymphaeaceae): Evidence from substitutions and microstructural changes in the chloroplast trnT-trnF regi on. Int J Plant Sci 168: 639-671. Borsch, T, C Lhne, K Mller, KW Hilu, S Wa nke, A Worberg, W Barthl ott, C Neinhuis, D Quandt 2005 Towards understanding basal angios perm diversification: recent insights using rapidly evolving genomic regions. Nova Acta Leopold 92: 85-110. Borsch, T, C Lhne, JH Wiersema 2008 Phylogeny and evolutionary patterns in Nymphaeales: integrating genes, genomes and morphology. Taxon 57: 1052-1081. Bowman, JL, J Alvarez, D Weigel, EM Meyerowitz, DR Smyth 1993 Control of flower development in Arabidopsis thaliana by APETALA1 and interacting genes. Development 119: 721-743. Bowman, JL 1997 Evolutionary conservation of angiosperm flower development at the molecular and genetic leve ls. J Biosci 22: 515-527.
166 Bowman, JL 2000 The YABBY gene family and abax ial cell fate. Curr Opin Plant Biol 3: 17-22. Buzgo, M, AS Chanderbali, S Kim, Z Zheng, D Oppenheimer, PS Soltis, DE Soltis 2007 Floral developmental morphology of Persea americana (avocado, Lauraceae): the oddities of male organ identity. Int J Plant Sci 168: 261-284. Buzgo, M, PS Soltis, DE Soltis 2004 Floral developmental morphology of Amborella trichopoda (Amborellaceae). Int J Plant Sci 165: 925-947. Buzgo, M, PS Soltis, S Kim, DE Soltis 2005 The making of the flower. Biologist 52: 149-154. Caspary, R 1888 Nymphaeaceae. in : A Engler, K Prantl eds. Die Natrlichen Pflanzenfamilien. Vol. 3. Wilhelm Engelmann, Leipzig, Germany. Cevallos-Ferriz, SRS, RA Stockey 1989 Permineralized fruits and seeds from the Princeton Chert (Middle Eocene) of British Columbia: Nymphacaeceae. Bot Gaz 150: 207-217. Chanderbali, AS, S Kim, M Buzgo, Z Zheng, DG Oppenheimer, DE Soltis, PS Soltis 2006 Genetic footprints of stamen ances tors guide perianth evolution in Persea (Lauraceae). Int J Plant Sci 167: 1075-1089. Chase, MW, DE Soltis, RG Olmstead, D Morgan, DH Les, BD Mishler, MR Dunvall, RA Price, HG Hills, Y-L Qiu, KA Kron, JH Rettig, E Conti, JD Palmer, JH Manhart, KJ Sytsma, HJ Michaels, WJ Kress, KG Karol, WD Clark, M Hedren, BS Gaut, RK Jansen, KJ Kim, CF Wimpee, JF Smith, GR Furnier, SH Strauss, QY Xiang, GM Plunkett, PS Soltis, SM Swensen, SE Williams, PA Gadek, CJ Qui nn, LE Eguiarte, E Golenberg, GH Learn, SW Graham, SCH Barrett, S Dayanadan, VA Albert 1993 Phylogenies of seed plants: an analysis of nucleotide sequences from the plastid gene rbcL Ann Mo Bot Gard 80: 528580. Chaw, SM, H Long, BS Wang, A Zharkikh, WH Li 1993 The phylogenetic position of Taxaceae based on 18S rRNA sequences. J Mol Evol 37: 624-630. Chaw, SM, CL Parkinson, Y Cheng, TM Vin cent, JD Palmer 2000 Seed plant phylogeny inferred from all three plant genomes: monophyl y of extant gymnosperms and origin of Gnetales from conifers. Proc Natl Acad Sci U S A 97: 4086-4091. Chaw, SM, A Zharkikh, HM Sung, TC Lau, WH Li 1997 Molecular phylogeny of extant gymnosperms and seed plant evolution: analys is of nuclear 18S rRNA sequences. Mol Biol Evol 14: 56-68. Chen, I, SR Manchester, Z Chen 2004 Anatomically preserved seeds of Nuphar (Nymphaeaceae) from the early Eocene of Wutu, Shandong Province, China. Am J Bot 91: 1265-1272. Chen, MK, IC Lin, CH Yang 2008 Functi onal analysis of three lily ( Lilium longiflorum ) APETALA1-like MADS box genes in regulating flor al transition and formation. Plant Cell Physiol 49: 704-717.
167 Chen, Q, A Atkinson, D Otsuga, T Ch ristensen, L Reynolds, GN Drews 1999 The Arabidopsis FILAMENTOUS FLOWER gene is required for flower formation. Development 126: 27152726. Cheng, Y, RG Nicolson, K Tripp, SM Chaw 2000 Phylogeny of Taxaceae and Cephalotaxaceae genera inferred from chloroplast matK gene and nuclear rDNA ITS region. Mol Phylogenet Evol 14: 353-365. Cho, Y, J Fernandes, SH Kim, V Walbot 2002 Ge ne-expression profile co mparisons distinguish seven organs of maize. Ge nome Biol 3: research0045. Coen, ES, EM Meyerowitz 1991 The war of the whor ls: genetic interactions controlling flower development. Nature 353: 31-37. Collinson, ME 1980 Recent and Tertia ry seeds of the Nymphaeaceae sensu lato with a revision of Brasenia ovula (Brong.) Reid and Chandler. Ann Bot 46: 603-632. Colombo, L, J Franken, E Koetje J van Went, HJ Dons, GC A ngenent, AJ van Tunen 1995 The Petunia MADS box gene FBP11 determines ovule identity. Plant Cell 7: 1859-1868. Crepet, WL 1998 The abominable mystery. Science 282: 1653-1654. Crepet, WL 2000 Progress in understanding angios perm history, success, and relationships: Darwins abominably 'perpl exing phenomenon'. Proc Natl Acad Sci U S A 97: 1293912941. Cronquist, A 1981 An integrated system of classification of flowering plants. Columbia University Press, New York, NY. Cronquist, A 1988 The evolution and classification of flowering plants. New York Botanical Garden, Bronx, New York. Crow, GE, CB Hellquist 2000 Cabombaceae and Ny mphaeaceae Aquatic and wetland plants of northeastern North America: a revised and en larged edition of Norman C. Fassett's A manual of aquatic plants. University of Wisconsin Press, Madison, WI, USA. Cutter, EG 1957a Studies of morphogenesis in the Nymphaeaceae. I. Introduction: some aspects of the morphology of Nuphar lutea (L.) Sm. and Nymphaea alba L. Phytomorphology 7: 45-56. Cutter, EG 1957b Studies of morphogenesis in the Nymphaeaceae. II. Floral development in Nuphar and Nymphaea: bracts and calyx. P hytomorphology 7: 57-73. Cutter, EG 1959 Studies of morphogenesis in the Nymphaeaceae. IV. Early floral development in species of Nuphar Phytomorphology 9: 263-275. Cutter, EG 1961 The inception and distribution of flowers in the Nymphaeaceae. Proc Linn Soc Bot 172: 93-100.
168 Davies, B, M Cartolano, Z Schwarz-Sommer 2006 Flower development: The Antirrhinum perspective. Adv Bot Res 44: 280-321. Davies, B, P Motte, E Keck, H Saedle r, H Sommer, Z Schwarz-Sommer 1999 PLENA and FARINELLI : redundancy and regulatory in teractions between two Antirrhinum MADS-box factors controlling flower development. EMBO J 18: 4023-4034. Davis, CC, CD Bell, PW Fritsch, S Mathews 2002a Phylogeny of Acridocarpus-Brachylophon (Malpighiaceae): implications fo r tertiary tropical floras and Afroasian biogeography. Evolution 56: 2395-2405. Davis, CC, CD Bell, S Mathews, MJ Donoghue 2002b Laurasian migration explains Gondwanan disjunctions: evidence from Malpighiaceae. Proc Natl Acad Sci U S A 99: 6833-6837. de Folter, S, AV Shchennikova, J Franken, M Busscher, R Baskar, U Grossniklaus, GC Angenent, RG Immink 2006 A Bsister MADS-box gene involved in ovule and seed development in Petunia and Arabidopsis Plant J 47: 934-946. de Hoon, MJ, S Imoto, S Miyano 2002 Statistical an alysis of a small set of time-ordered gene expression data using linear spli nes. Bioinformatics 18: 1477-1485. Ditta, G, A Pinyopich, P Robles, S Pelaz, MF Yanofsky 2004 The SEP4 gene of Arabidopsis thaliana functions in floral or gan and meristem identi ty. Curr Biol 14: 1935-1940. Dorofeev, PI 1973 Systematics of ancestral forms of Brasenia Paleontological Journal 2: 212227. Dorofeev, PI 1974 Nymphaeales. in : A Takhtajan ed. Magnoliophyta Fossilia URSS. Vol. 1. Nauka, Leningrad, Russia. Doyle, JA, H Eklund, PS Herendeen 2003 Floral e volution in Chloranthaceae: Implication of a morphological phylogenetic analysis. Int J Plant Sci 164 Suppl: S365-S382. Doyle, JA, PK Endress 2000 Morphological phylogenetic analysis of basal angiosperms: Comparison and combination with molecu lar data. Int J Plant Sci 161: 5121-5153. Doyle, JJ, JL Doyle 1987 Rapid DNA isolation from small amount of fresh leaf tissue. Phytochem Bull 19: 11-15. Drinnan, AN, PR Crane, SB Hoot 1994 Patterns of fl oral evolution in the ear ly diversification of non-magnoliid dicotyledons (eudicots). Pl Syst Evol Suppl 8: 93-122. Druka, A, G Muehlbauer, I Druka, R Caldo, U Baumann, N Rostoks, A Schreiber, R Wise, T Close, A Kleinhofs, A Graner, A Schulman, P Langridge, K Sato, P Hayes, J McNicol, D Marshall, R Waugh 2006 An atlas of gene expr ession from seed to seed through barley development. Funct Integr Genomics 6: 202-211.
169 Duarte, JM, L Cui, PK Wall, Q Zhang, X Zhang, J Leebens-Mack, H Ma, N Altman, CW dePamphilis 2006 Expression pattern shifts following duplication indicative of subfunctionalization and neofunctionalization in regulatory genes of Arabidopsis Mol Biol Evol 23: 469-478. Eisen, MB, PT Spellman, PO Brown, D Botstein 1998 Cluster analysis a nd display of genomewide expression patterns. Proc Natl Acad Sci U S A 95: 14863-14868. Eklund, H, JA Doyle, PS Herendeen 2004 Morpho logical phylogenetic an alysis of living and fossil Chloranthaceae. Int J Plant Sci 165: 107-151. Endress, PK 2001 The flowers in extant basal angiosperms and inferences on ancestral flowers. Int J Plant Sci 162: 1111-1140. Endress, PK 2006 Angiosperm floral evoluti on: Morphological developmental framework. Adv Bot Res 44: 1-61. Endress, PK 2008 Perianth biology in the basal grade of extant angiosperms. Int J Plant Sci 169: 844-862. Fan, J, W Li, X Dong, W Guo, H S hu 2007 Ectopic expression of a Hyacinth AGL6 homolog caused earlier flowering and homeotic conversion in Arabidopsis Sci China C Life Sci 50: 676-689. Fassett, N 1953 A monograph of Cabomba. Castanea 13: 116-128. Felsenstein, J 1981 Evolutionary trees from DNA seque nces: a maximum likelihood approach. J Mol Evol 17: 368-376. Felsenstein, J 1984 Distance methods for inferring phylogenies: a justifica tion. Evolution 38: 1624. Felsenstein, J 1985 Confidence limits on phylogeni es: an approach using the bootstrap. Evolution 39: 783-791. Felsenstein, J 1988 Phylogenies from molecular sequences: inference and reliability. Annu Rev Genet 22: 521-565. Fornara, F, G Marziani, L Mizzi, MM Kater, L Colombo 2003 MADS-box genes controlling flower development in rice. Plant Biology 1: 16-22. Fraser, CM, LW Rider, C Chapple 2005 An expr ession and bioinformatics analysis of the Arabidopsis serine carboxypeptidase -like gene family. Plant Physiol 138: 1136-1148. Friedman, WE 2008 Hydatellaceae are water lilies with gymnospermous tende ncies. Nature 453: 94-97.
170 Friis, EM 1997 Fossil history of magnoliid angiosperms. in : K Iwatsuki, PR Raven eds. Evolution and diversification of la nd plants. Springer, Tokyo, Japan. Friis, EM, KR Pedersen, PR Crane 1994 Angiosperm floral structures from the Early Cretaceous of Portugal. Plant Syst Evol 8: 31-49. Friis, EM, KR Pedersen, PR Crane 1999 Early angios perm diversification: the diversity of pollen associated with angiosperm reproductive stru ctures in Early Cretaceous floras from Portugal. Ann Mo Bot Gard 86: 259-296. Friis, EM, KR Pedersen, PR Crane 2001 Fossil ev idence of water lilies (Nymphaeales) in the Early Cretaceous. Nature 410: 357-360. Frohlich, MW 1999 MADS about Gnetales. Proc Natl Acad Sci U S A 96: 8811-8813. Frohlich, MW, DS Parker 2000 The mostly male theo ry of flower evolutionary origins: from genes to fossils. Syst Bot 25: 155-170. Frohlich, MW 2003 An evolutionary scenario for the origin of flowers. Nat Rev Genet 4: 559566. Frohlich, MW 2006 Recent developments regarding th e evolutionary origin of flowers. Adv Bot Res 44: 63-127. Fukui, M, N Futamura, Y Mukai, Y Wang, A Nagao, K Shinohara 2001 Ancestral MADS box genes in Sugi, Cryptomeria japonica D. Don (Taxodiaceae), homologous to the B function genes in angiosperms. Plant Cell Physiol 42: 566-575. Gandolfo, MA, KC Nixon, WL Crepet 2004 Cret aceous flowers of Nymphaeaceae and implications for complex insect entrapment pollination mechanisms in early angiosperms. Proc Natl Acad Sci U S A 101: 8056-8060. Goldman, N 1993 Statistical tests of models of DNA substitution. J Mol Evol 36: 182-198. Golz, JF, A Hudson 1999 Plant development: YABBYs claw to the fore. Curr Biol 9: R861-863. Gomez-Mena, C, S de Folter, MM Costa, GC Angenent, R Sablowski 2005 Transcriptional program controlled by the floral homeotic gene AGAMOUS during early organogenesis. Development 132: 429-438. Goremykin, VV, KI Hirsch-Ernst, S Wolf l, FH Hellwig 2003 Analysis of the Amborella trichopoda chloroplast genome sequence suggests that Amborella is not a basal angiosperm. Mol Biol Evol 20: 1499-1505. Goto, K, EM Meyerowitz 1994 Function and regulation of the Arabidopsis floral homeotic gene PISTILLATA Genes Dev 8: 1548-1560.
171 Graham, SW, RG Olmstead 2000 Utility of 17 chloroplast genes for inferring the phylogeny of the basal angiosperms. Am J Bot 87: 1712-1730. Guo, Y, S Gan 2006 AtNAP, a NAC family transcrip tion factor, has an impor tant role in leaf senescence. Plant J 46: 601-612. Hardenack, S, D Ye, H Saedler, S Grant 1994 Comparison of MADS box gene expression in developing male and female flowers of the dioecious plant white campion. Plant Cell 6: 1775-1787. Hayes, V, EL Schneider, S Carl quist 2000 Floral development of Nelumbo nucifera (Nelumbonaceae). Int J Plant Sci 161: 183-191. Heinsbroek, PG, WA Van Heel 1969 Note on the b earing of the pattern of vascular bundles on the morphology of the stamens of Victoria amazonica (Poepp.) Sowerby. Proc K ned A kad Wet, Ser C 72: 431-444. Henkel, F, F Rehnelt, L Dittmann 1907 Das Buch der Nymphaeaceen oder Seerosengewchse. Verlag F. Henkel, Darmstadt, Germany. Hennig, L, W Gruissem, U Grossniklaus, C K ohler 2004 Transcriptional programs of early reproductive stages in Arabidopsis Plant Physiol 135: 1765-1775. Heywood, VH 1993 Flowering plants of the worl d. Oxford University Press, New York. Hiepko, P 1965 Vergleichend-morphologische und entwicklungsgeschichtliche Untersuchungen uber das Perianth bei den Polycarpicae. Bot Jahrb Syst 84: 359-508. Hilu, KW, T Borsch, K Mller, DE Soltis, PS Soltis, V Savolainen, MW Chase, MP Powell, LA Alice, R Evans, H Sauquet, C Neinhuis, TAB Slotta, JG Rohwer, CS Campbell, LW Chatrou 2003 Angiosperm phylogeny based on matK sequence information. Am J Bot 90: 1758-1776. Honma, T, K Goto 2001 Complexes of MADS-box prot eins are sufficient to convert leaves into floral organs. Nature 409: 525-529. Hoot, SB, S Magallon, PR Crane 1999 Phylogeny of basal eudicots based on three molecular data sets: atp B, rbc L, and 18S nuclear ribosomal DNA sequences. Ann Mo Bot Gard 86: 1-32. Hord, CL, C Chen, BJ Deyoung, SE Clark, H Ma 2006 The BAM1/ BAM2 receptor-like kinases are important regulators of Arabidopsis early anther development. Plant Cell 18: 16671680. Hsu, HF, CH Huang, LT Chou, CH Yang 2003 Ectopic expression of an orchid ( Oncidium Gower Ramsey) AGL6 -like gene promotes flowering by activating flowering time genes in Arabidopsis thaliana Plant Cell Physiol 44: 783-794.
172 Huijser, P, J Klein, WE Lonnig, H Meijer, H Saedler, H Sommer 1992 Bracteomania, an inflorescence anomaly, is caused by the loss of function of the MADS-box gene squamosa in Antirrhinum majus. EMBO J 11: 1239-1249. Irish, VF, IM Sussex 1990 Function of the apetala-1 gene during Arabidopsis floral development. Plant Cell 2: 741-753. Ito, M 1983 Studies in the floral morphology an d anatomy of Nymphaeales. I. The morphology of vascular bundles in the flower of Nymphaea tetragona George. Acta Phytotax Geobot 34: 18-26. Ito, M 1984 Studies in the floral morphology an d anatomy of the Nymphaeales. II. Floral anatomy of Nymphaea tetragona George. Acta Phytotax Geobot 35: 18-26. Ito, M 1986 Studies in the floral morphology and anatomy of Nymphaeales. III. Floral anatomy of Brasenia schreberi Gmel. and Cabomba caroliniana A. Gray. Bot Mag (Tokyo) 99: 169-184. Ito, M 1987 Phylogenetic systematics of th e Nymphaeales. Bot Mag (Tokyo) 100: 17-36. Jack, T 2004 Molecular and genetic mechanisms of floral control. Plant Cell 16 Suppl: S1-17. Jack, T, LL Brockman, EM Meyerowitz 1992 The homeotic gene APETALA3 of Arabidopsis thaliana encodes a MADS box and is expressed in petals and stamens. Cell 68: 683-697. Jack, T, GL Fox, EM Meyerowitz 1994 Arabidopsis homeotic gene APETALA3 ectopic expression: transcriptional and posttranscriptional regulation determine floral organ identity. Cell 76: 703-716. Jager, M, A Hassanin, M Manuel, H Le G uyader, J Deutsch 2003 MADS-box genes in Ginkgo biloba and the evolution of the AGAMOUS family. Mol Biol Evol 20: 842-854. Jakoby, M, B Weisshaar, W Droge-L aser, J Vicente-Carbajosa, J Tiedemann, T Kroj, F Parcy 2002 bZIP transcription factors in Arabidopsis Trends Plant Sci 7: 106-111. Jansen, RK, Z Cai, LA Raubeson, H Daniell, CW Depamphilis, J Leebens-Mack, KF Muller, M Guisinger-Bellian, RC Haberle, AK Hansen, TW Chumley, SB Lee, R Peery, JR McNeal, JV Kuehl, JL Boore 2007 Analysis of 81 genes from 64 plastid genomes resolves relationships in angiosperms and identifies ge nome-scale evolutionary patterns. Proc Natl Acad Sci U S A 104: 19369-19374. Jones-Rhoades, MW, JO Borevitz, D Preuss 2007 Genome-wide expressi on profiling of the Arabidopsis female gametophyte identifies families of small, secreted proteins. PLoS Genet 3: 1848-1861. Judd, WS, CS Campbell, EA Kellogg, PF St evens, MJ Donoghue 2002 Nymphaeales Plant systematics: a phylogenetic approach. Sinauer Associates, Sunderland, MA.
173 Kang, HG, YS Noh, YY Chung, MA Costa, K An, G An 1995 Phenotypic alterations of petal and sepal by ectopic expression of a rice MADS box gene in tobacco. Plant Mol Biol 29: 1-10. Kater, MM, L Colombo, J Franken, M Busscher, S Masiero, MM Van Lookeren Campagne, GC Angenent 1998 Multiple AGAMOUS homologs from cucumber and Petunia differ in their ability to induce reproductive or gan fate. Plant Cell 10: 171-182. Kater, MM, L Dreni, L Colombo 2006 Func tional conservation of MADS-box factors controlling floral organ identity in rice and Arabidopsis J Exp Bot 57: 3433-3444. Kaya, H, KI Shibahara, KI Taoka, M Iwabuchi, B Stillman, T Araki 2001 FASCIATA genes for chromatin assembly factor-1 in Arabidopsis maintain the cellular organization of apical meristems. Cell 104: 131-142. Kempin, SA, MA Mandel, MF Yanofsky 1993 Conversion of perianth into reproductive organs by ectopic expression of the t obacco floral homeotic gene NAG1 Plant Physiol 103: 10411046. Kenrick, P, PR Crane 1997 The origin and early di versification of land pl ants: a cladistic study. Smithsonian Institution Press, Washington, D.C. Kerner von Marilaun, A 1891 Pflanzenleben, Bd. 2: Geschichte der Pflanzen. Verlag des Bibliographischen Instit uts, Leipzig, Germany. Kerstetter, RA, K Bollman, RA Ta ylor, K Bomblies, RS Poethig 2001 KANADI regulates organ polarity in Arabidopsis Nature 411: 706-709. Khanna, P 1964 Morphological and embryologi cal studies in the Nymphaeaceae. I. Euryale ferox Salisb. Proc Indian Acad Sci 59: 237-243. Khanna, P 1967 Morphological and embryologi cal studies in the Nymphaeaceae. III. Victoria cruziana d'Orb. and Nymphaea stellata Willd. Bot Mag (Tokyo) 80: 305-312. Kim, GT, K Shoda, T Tsuge, KH Cho, H Uchimi ya, R Yokoyama, K Nishitani, H Tsukaya 2002 The ANGUSTIFOLIA gene of Arabidopsis a plant CtBP gene, regulates leaf-cell expansion, the arrangement of cortical microtubu les in leaf cells and expression of a gene involved in cell-wall form ation. EMBO J 21: 1267-1279. Kim, GT, H Tsukaya, H Uchimiya 1998 The CURLY LEAF gene controls both division and elongation of cells during the expa nsion of the leaf blade in Arabidopsis thaliana Planta 206: 175-183. Kim, S, J Koh, MJ Yoo, H Kong, Y Hu, H Ma, PS Soltis, DE Soltis 2005 Expression of floral MADS-box genes in basal angiosperms: implications for the evolution of floral regulators. Plant J 43: 724-744.
174 Kim, S, LM Zahn, Z Zheng, AS Chanderbali, DG Oppenheimer, H Ma, PS Soltis, DE Soltis unpublished The evolution of the AGL6 -like MADS-box genes. Kishino, H, M Hasegawa 1989 Evaluation of the maximum likelihood estimate of the evolutionary tree topologies from DNA sequence data, and the branching order in hominoidea. J Mol Evol 29: 170-179. Kishino, H, JL Thorne, WJ Bruno 2001 Performa nce of a divergence time estimation method under a probabilistic model of rate evolution. Mol Biol Evol 18: 352-361. Knobloch, E, DH Mai 1984 Neue Gattungen nach Frchten und Samen aus dem Cenoman bis Maastricht (Kreide) von Mitteleuro pa. Feddes Repertorium 95: 3-41. Kramer, EM, VS Di Stilio, PM Schluter 2003 Co mplex patterns of gene duplication in the APETALA3 and PISTILLATA lineages of the Ranunculaceae Int J Plant Sci 164: 1-11. Kramer, EM, RL Dorit, VF Irish 1998 Molecular evolution of genes cont rolling petal and stamen development: duplication a nd divergence within the APETALA3 and PISTILLATA MADSbox gene lineages. Genetics 149: 765-783. Kyozuka, J, T Kobayashi, M Morita, K Shimam oto 2000 Spatially and temporally regulated expression of rice MADS box genes with similarity to Arabidopsis class A, B and C genes. Plant Cell Physiol 41: 710-718. Langley, CH, WM Fitch 1974 An examination of the constancy of the rate of molecular evolution. J Mol Evol 3: 161-177. Leebens-Mack, J, LA Raubeson, L Cui, JV Kueh l, MH Fourcade, TW Chumley, JL Boore, RK Jansen, CW depamphilis 2005 Identifying the basal angiosperm node in chloroplast genome phylogenies: sampling one's way out of the Felsenstein zone. Mol Biol Evol 22: 1948-1963. Les, DH, DK Garvin, CF Wimpee 1991 Molecula r evolutionary history of ancient aquatic angiosperms. Proc Natl Acad Sci U S A 88: 10119-10123. Les, DH, EL Schneider, DJ Padgett, PS Soltis, DE Soltis, M Zanis 1999 Phylogeny, classification and floral evolution of water lilies (Nymphaeaceae; Nymphaeales): A synthesis of non-molecular, rbcL, matK, and 18S rDNA data. Syst Bot 24: 28-46. Leys, RS, JB Cooper, MP Schwarz 2002 Molecu lar phylogeny and historical biogeography of the large carpenter bees, genus Xylocopa (Hymenoptera: Apidae). Biol J Linn Soc 77: 249266. Li, M, W Xu, W Yang, Z Kong, Y Xue 2007 Genomewide gene expression profiling reveals conserved and novel molecular functions of th e stigma in rice. Plant Physiol 144: 17971812.
175 Li, Y, K Sorefan, G Hemmann, MW Bevan 2004 Arabidopsis NAP and PIR regulate actin-based cell morphogenesis and multiple developmenta l processes. Plant Physiol 136: 3616-3627. Litt, A 2007 An evaluation of A-function: evidence from the APETALA1 and APETALA2 gene lineages. Int J Plant Sci 168: 73-91 Litt, A, VF Irish 2003 Duplicati on and Diversification in the APETALA1/FRUITFULL Floral Homeotic Gene Lineage: Implications for th e Evolution of Floral Development. Genetics 165: 821-833. Llamas, KA 2003 Tropical flowering plants: A guide to identification a nd cultivation. Timber Press, Portland, OR. Lhne, C, T Borsch 2005 Molecular evolution and phylogenetic utility of the petD group II intron: a case study in basal angi osperms. Mol Biol Evol 22: 317-332. Lhne, C, T Borsch, JH Wiersema 2007 Phyloge netic analysis of Ny mphaeales using fastevolving and noncoding chloroplast ma rkers. Bot J Linn Soc 154: 141-163. L, S, X Du, W Lu, K Chong, Z Meng 2007 Two AGAMOUS-like MADS-box genes from Taihangia rupestris (Rosaceae) reveal independent trajecto ries in the evolution of class C and class D floral homeotic f unctions. Evol Dev 9: 92-104. Lynch, M, JS Conery 2000 The evolutionary fate and consequences of duplicate genes. Science 290: 1151-1155. Ma, H, C dePamphilis 2000 The ABCs of floral evolution. Cell 101: 5-8. Ma, H, MF Yanofsky, EM Meyerowitz 1991 AGL1-AGL6 an Arabidopsis gene family with similarity to floral homeotic and transc ription factor genes. Genes Dev 5: 484-495. Magalln, S 2004 Dating lineages: Mo lecular and paleontological approaches to the temporal framework of clades. Int J Plant Sci 165 Suppl: S7-S21. Mai, DH 1988 New nymphaealean fossils from the Ter tiary of central Europe. Tertiary Research 9: 87-96. Mapes, G, GW Rothwell 1984 Permineralized ov ulate cones of Lebachia from the late Palaeozoic limestones of Kansas. Paleontology 27: 69-94. Mapes, G, GW Rothwell 1991 Structure and relati onships of primitive conifers. Neues Jahrbuch fr Geologie und Palontologie Abhandlungen 183. Martinez-Castilla, LP, ER Alvarez-B uylla 2003 Adaptive evolution in the Arabidopsis MADSbox gene family inferred from its complete resolved phylogeny. Proc Natl Acad Sci U S A 100: 13407-13412.
176 Mathews, S, MJ Donoghue 1999 The root of a ngiosperm phylogeny inferred from duplicate phytochrome genes. Science 286: 947-950. Mena, M, MA Mandel, DR Lerner, MF Yanofsk y, RJ Schmidt 1995 A characterization of the MADS-box gene family in maize. Plant J 8: 845-854. Miki, S 1960 Nymphaeaceae remains in Japan, with new fossil genus Eoeuryale. Journal of the Institute of Polytechnics, Osaka City University, Japan D-11: 63-78. Moore, MJ, CD Bell, PS Soltis, DE Soltis 2007 Using plastid genome-scale data to resolve enigmatic relationships among basal angios perms. Proc Natl Acad Sci U S A 104: 1936319368. Moore, RC, SR Grant, MD Purugganan 2005 Molecular population genetics of redundant floralregulatory genes in Arabidopsis th aliana. Mol Biol Evol 22: 91-103. Moore, RC, MD Purugganan 2005 The evolutionary dynamics of plant duplicate genes. Curr Opin Plant Biol 8: 122-128. Morris, AB, CD Bell, JW Clayton, WS Judd, DE Soltis, PS Soltis 2007 Phylogeny and divergence time estimation in Illicium with implications for New World biogeography. Syst Bot 32: 236-249. Moseley, MF 1958 Morphological studies of the Ny mphaeaceae. I. The nature of the stamens. Phytomorphology 8: 1-29. Moseley, MF 1961 Morphological studies of the Nymphaeaceae. II. THe flower of Nymphaea Bot Gaz 122: 233-259. Moseley, MF 1965 Morphological studies of the Nymphaeaceae. III. The floral anatomy of Nuphar Phytomorphology 15: 54-84. Moseley, MF 1972 Morphological studies of Nymphaeaceae. VI. Development of flower of Nuphar Phytomorphology 21: 253-283. Moseley, MF, IJ Mehta, PS Williamson, H Kosakai 1984 Morphological studies of the Nymphaeaceae (sensu lato). XIII. Contributions to the vegetative and floral structure of Cabomba. Am J Bot 71: 902-924. Moseley, MF, EL Schneider, PS Williamson 1993 Phyl ogenetic interpretati on from selected floral vasculature character s in the Nymphaeaceae sensu lato. Aquat Bot 44: 325-342. Mouradov, A, B Hamdorf, RD Teasdale, JT Kim, KU Winter, G Theissen 1999 A DEF/GLO like MADS-box gene from a gymnosperm: Pinus radiata contains an ortholog of angiosperm B class floral homeotic genes. Dev Genet 25: 245-252. Nakai, T 1943 Ordines, familae, tribi, genera, sectiones, species, varietates, formae et combinationes novae a Prof. Nakai-Takenosin adhuc tu novis edita. Appendix:
177 Quaestiones characterium naturalium planta rum vel Extractus ex praelectionibus pro alumnis botanicus Universitatis Imperia lis Tokyoensis per annos 1926. Imperial University of Tokyo, Tokyo, Japan. Nam, J, J Kim, S Lee, G An, H Ma, M Nei 2004 Type I MADS-box genes have experienced faster birth-and-death evolu tion than type II MADS-box gene s in angiosperms. Proc Natl Acad Sci U S A 101: 1910-1915. Nesi, N, I Debeaujon, C Jond, AJ Stewart, GI Jenkins, M Caboche, L Lepiniec 2002 The TRANSPARENT TESTA16 locus encodes the ARABIDOPSIS BSISTER MADS domain protein and is required for proper development and pigmentation of the seed coat. Plant Cell 14: 2463-2479. Ng, M, MF Yanofsky 2001 Function and evoluti on of the plant MADS-box gene family. Nat Rev Genet 2: 186-195. Nijhawan, A, M Jain, AK Tyagi, JP Khurana 2008 Genomic survey and gene expression analysis of the basic leucine zipper tr anscription factor family in rice. Plant Physiol 146: 333-350. rgaard, M 1991 The genus Cabomba (Cabombaceae)-a taxonomic study. Nord J Bot 11: 179203. Osborn, JM, EL Schneider 1988 Morphological st udies of the Nymphaeaceae sensu lato. XVI. The floral biology Of Brasenia schreberi Ann Mo Bot Gard 75: 778-794. Padgett, DJ, DH Les, GE Crow 1999 Phylogenetic relationships in Nuphar (Nymphaeaceae): evidence from morphology, chloroplast DNA, a nd nuclear ribosomal DNA. Am J Bot 86: 1316-1324. Padgett, DJ 2007 A monograph of Nuphar (Nymphaeaceae). Rhodora 109: 1-95. Parkinson, CL, KL Adams, JD Palmer 1999 Mul tigene analyses identif y the three earliest lineages of extant flowering plants. Curr Biol 9: 1485-1488. Pelaz, S, GS Ditta, E Baumann, E Wisman, MF Yanofsky 2000 B and C floral organ identity functions require SEPALLATA MADS-box genes. Nature 405: 200-203. Perez-Perez, JM, MR Ponce, JL Micol 2002 The UCU1 Arabidopsis gene encodes a SHAGGY/GSK3-like kinase required for cell e xpansion along the proximodistal axis. Dev Biol 242: 161-173. Poupin, MJ, F Federici, C Medina JT Matus, T Timmermann, P Arce-Johnson 2007 Isolation of the three grape sub-lineages of B-class MADS-box TM6, PISTILLATA and APETALA3 genes which are differentially expressed during flower and fruit development. Gene 404: 10-24. Prance, GT, JR Arias 1975 A st udy of the floral biology of Victoria amazonica (Poepp.) Sowerby (Nymphaeaceae). Acta Amazonica 5: 109-139.
178 Qiu, YL, MW Chase, DH Les, CR Parks 1993 Molecular phylogenetics of the Magnoliidae: cladistic analysis of nucleotide sequences of the plastid gene rbcL Ann Mo Bot Gard 80: 587-606. Qiu, YL, O Dombrovska, J Lee, LB Li, BA Whitlock, F Bernasconi-Quadroni, JS Rest, CC Davis, T Borsch, KW Hilu, SS Renner, DE Soltis, PS Soltis, MJ Zanis, JJ Cannone, RR Gutell, M Powell, V Savolainen, LW Chatrou, MW Chase 2005 Phylogenetic analyses of basal angiosperms based on nine plastid, mitochondrial, and nuclear genes. Int J Plant Sci 166: 815-842. Qiu, YL, J Lee, F Bernasconi-Quadroni, DE Soltis, PS Soltis, M Zanis, EA Zimmer, Z Chen, V Savolainen, MW Chase 1999 The earliest angiosperms: evidence from mitochondrial, plastid and nuclear geno mes. Nature 402: 404-407. Quinn, CJ, RA Price, PA Gadek 2002 Familial concep ts and relationships in the conifers based on rbcL and matK sequence comparisons. Kew Bull 57: 513-531. Richard, A 1828 Nouveaux lments de botanique et de physiologie vgtale. Bechet, Paris, France. Riechmann, JL, BA Krizek, EM Meyerowitz 1996 Dimerization specificity of Arabidopsis MADS domain homeotic proteins AP ETALA1, APETALA3, PISTILLATA, and AGAMOUS. Proc Natl Acad Sci U S A 93: 4793-4798. Rijpkema, A, T Gerats, M Vandenbussche 2006 Genetics of floral development in Petunia Adv Bot Res 44: 237-278. Ronquist, F 1996 DIVA: dispersal-vi cariance analysis, vers. 1.1. U ppsala University, Uppsala, Sweden. Ronquist, F 1997 Dispersal-vicariance analysis: a new approach to th e quantification of historical biogeography. Syst Biol 46: 195-203. Ronse De Craene, LP 2008 Homology and evolution of petals in the core eudicots. Syst Bot 33: 301-325. Ronse De Craene, LP, EF Smets 2001 Staminode s: Their morphological and evolutionary significance. Bot Rev 67: 351-402. Rounsley, SD, GS Ditta, MF Yanofsky 1995 Divers e roles for MADS box genes in Arabidopsis development. Plant Cell 7: 1259-1269. Rudall, PJ, DD Sokoloff, MV Remizowa, JG Conr an, JI Davis, TD Macfarlane, DW Stevenson 2007 Morphology of Hydatellaceae, an anomalous aquatic family recently recognized as an early-divergent angiosperm lineage. Am J Bot 94: 1073-1092. Rutledge, R, S Regan, O Nicolas, P Fobert, C Cote, W Bosnich, C Kauffeldt, G Sunohara, A Seguin, D Stewart 1998 Char acterization of an AGAMOUS homologue from the conifer
179 black spruce (Picea mariana) that produces floral homeotic conversions when expressed in Arabidopsis Plant J 15: 625-634. Ruzin, SE 1999 Plant microtechnique and microsc opy. Oxford University Press, Oxford, UK. Saarela, JM, HS Rai, JA Doyle, PK Endress, S Mathews, AD Marchant, BG Briggs, SW Graham 2007 Hydatellaceae identified as a new branch near the base of the angiosperm phylogenetic tree. Nature 446: 312-315. Sablowski, RW, EM Meyero witz 1998 A homolog of NO APICAL MERISTEM is an immediate target of the floral homeotic genes APETALA3/PISTILLATA. Cell 92: 93-103. Saldanha, AJ 2004 Java Treeview--e xtensible visualization of mi croarray data. Bioinformatics 20: 3246-3248. Sanderson, MJ 1997 A nonparametric approach to esti mating divergence times in the absence of rate constancy. Mol Biol Evol 14: 1218-1231. Sanderson, MJ 1998 Estimating rate and time in molecular phylogenies: beyond the molecular clock? in : DE Soltis, PS Soltis, JJ Doyle eds. Mo lecular systematics of plants II, DNA sequencing. Kluwer, Boston, MA. Sanderson, MJ 2002 Estimating absolute rates of mo lecular evolution and divergence times: a penalized likelihood approach. Mol Biol Evol 19: 101-109. Sanderson, MJ 2003 r8s: inferring absolute rates of molecular evolution a nd divergence times in the absence of a molecular cl ock. Bioinformatics 19: 301-302. Sanderson, MJ, JA Doyle 2001 Sources of error and confidence intervals in estimating the age of angiosperms from rbcL and 18S rDNA data. Am J Bot 88: 1499-1516. Sanderson, MJ, JL Thorne, N Wikstrm, K Bremer 2004 Molecular evidence on plant divergence times. Am J Bot 91: 1656-1665. Schmid, M, TS Davison, SR Henz, UJ Pape, M Demar, M Vingron, B Scholkopf, D Weigel, JU Lohmann 2005 A gene expression map of Arabidopsis thaliana development. Nat Genet 37: 501-506. Schmid, M, NH Uhlenhaut, F Godard, M Dema r, R Bressan, D Weigel, JU Lohmann 2003 Dissection of floral induction pathways using global expression analysis. Development 130: 6001-6012. Schmidt, RJ, B Veit, MA Mandel, M Mena, S Hake, MF Yanofsky 1993 Identification and molecular characterization of ZAG1 the maize homolog of the Arabidopsis floral homeotic gene AGAMOUS Plant Cell 5: 729-737. Schneider, EL 1976 The floral anatomy of Victoria Schomb. (Nymphaeaceae). Bot J Linn Soc 72: 115-148.
180 Schneider, EL 1983 Gross morphol ogy and floral biology of Ondinea purpurea den Hartog. Aust J Bot 31: 371-382. Schneider, EL, JM Jeter 1982 Morphological studie s of the nymphaeaceae. Xii. The floral biology of Cabomba caroliniana. Am J Bot 69: 1410-1419. Schneider, EL, LA Moore 1977 Morphological studies of the Nymphaeaceae. VII. The floral biology of Nuphar lutea subsp. macrophylla Brittonia 29: 88-99. Schneider, EL, SC Tucker, PS Williamson 2003 Flor al development in the Nymphaeales. Int J Plant Sci 164 Suppl 5: S279-S292. Schneider, EL, PS Williamson 1993 Nymphaeaceae. in : K Kubitzki, JG Rohwer, V Bittrich eds. The families and genera of vascular plan ts. Vol. 2. Springer, Berlin, Germany. Schwarz-Sommer, Z, I Hue, P Huijser, PJ Flor R Hansen, F Tetens, WE Lonnig, H Saedler, H Sommer 1992 Characterization of the Antirrhinum floral homeotic MADS-box gene DEFICIENS : evidence for DNA binding and autoregula tion of its persistent expression throughout flower development. EMBO J 11: 251-263. Schwarz-Sommer, Z, P Huijser, W Nacken, H Saedler, H Sommer 1990 Genetic Control of Flower Development by Homeotic Genes in Antirrhinum majus. Science 250: 931-936. Shan, H, K Su, W Lu, H Kong, Z Chen, Z Meng 2006 Conservation and divergence of candidate class B genes in Akebia trifoliata (Lardizabalaceae). De v Genes Evol 216: 785-795. Shan, H, N Zhang, C Liu, G Xu, J Zhang, Z Che n, H Kong 2007 Patterns of gene duplication and functional diversification dur ing the evolution of the AP1/SQUA subfamily of plant MADS-box genes. Mol Phylogenet Evol 44: 26-41. Smith, AG, DG Smith, BM Funnel 1994 Atlas of Me sozoic and Cenozoic Coastlines. Cambridge University Press, Cambridge, UK. Smyth, DR, JL Bowman, EM Meyerowitz 1990 Early flower development in Arabidopsis Plant Cell 2: 755-767. Smyth, GK 2004 Linear models and empirical bayes methods for assessing differential expression in microarray experiments. St at Appl Genet Mol Biol 3: Article3. Smyth, GK, J Michaud, HS Scott 2005 Use of w ithin-array replicate spots for assessing differential expression in microarray experiments. Bioinformatics 21: 2067-2075. Smyth, GK, M Ritchie, N Thorne, J Wettenhall 2006 LIMMA: Linear Models for Microarray Data User's Guide. Sokoloff, DD, MV Retnizowa, TD Macfarlane PJ Rudall 2008 Classi fication of the earlydivergent angiosperm family Hydatellaceae: one genus instead of two, four new species and sexual dimorphism in di oecious taxa. Taxon 57: 179-200.
181 Soltis, DE, H Ma, MW Frohlich, PS Soltis, VA Albert, DG Oppenheimer, NS Altman, C dePamphilis, J Leebens-Mack 2007 The floral ge nome: an evolutionary history of gene duplication and shifting patte rns of gene expression. Trends Plant Sci 12: 358-367. Soltis, DE, PS Soltis, DL Nickrent, LA Johnson, WJ Hahn, SB Hoot, JA Sa kamoto, RK Kuzoff, KA Kron, MW Chase, SM Swensen, EA Zimmer, SM Chaw, LJ Gillespie, KJ Systma 1997 Angiosperm phylogeny inferred from 18S ribosomal DNA sequences. Ann Mo Bot Gard 84: 1-49. Soltis, DE, PS Soltis, MW Chase, ME Mort, DC Albach, M Zanis, V Savolainen, WH Hahn, SB Hoot, MF Fay, M Axtell, SM Swensen, LM Pr ince, WJ Kress, KC Nixon, JS Farris 2000 Angiosperm phylogeny inferred from 18S rDNA, rbcL and atpB sequences. Bot J Linn Soc 133: 381-461. Soltis, DE, PS Soltis, PK Endress, MW Chase 2005 Phylogeny and evolution of angiosperms. Sinauer Associates, Sunderland, MA. Soltis, PS, DE Soltis, MW Chase 1999 Angiospe rm phylogeny inferred from multiple genes as a tool for comparative biology. Nature 402: 402-404. Soltis, PS, DE Soltis, V Savolainen, PR Crane, TG Barraclough 2002 Rate heterogeneity among lineages of tracheophytes: in tegration of molecular and fossil data and evidence for molecular living fossils. Proc Natl Acad Sci U S A 99: 4430-4435. Stellari, GM, MA Jaramillo, EM Kramer 2004 Evolution of the APETALA3 and PISTILLATA lineages of MADS-box-containing genes in the basal angiosperms. Mol Biol Evol 21: 506519. Storey, BC 1995 The role of mantle plumes in continental breakup: case histories from Gondwanaland. Nature 377: 301-308. Sundstrm, J, A Carlsbecker, ME Svensson, M Svenson, U Johanson, G Theissen, P Engstrm 1999 MADS-box genes active in developing pollen cones of Norway spruce ( Picea abies) are homologous to the B-class floral homeotic genes in angiosperms. Dev Genet 25: 253266. Swofford, DL 2002 PAUP* 4.0b10: phylogenetic analysis using parsimony (*and other methods). Sinauer Associat es, Sunderland, MA, USA. Takahashi, M 1992 Development of spinous exine in Nuphar japonicum De Candolle (Nymphaeaceae). Review of Palae obotany and Palynology 75: 317-322. Takhtajan, A 1980 Outline of the classification of flowering plants (Magnoliophyta). Bot Rev 46: 225-359. Takhtajan, A 1991 Evolutionary trends in flowering plants. Columbia University Press, New York, NY, USA.
182 Takhtajan, A 1997 Diversity and cla ssification of flowering plants. Columbia University Press, New York, NY. Tandre, K, M Svenson, ME Svensson, P Engstr om 1998 Conservation of gene structure and activity in the regulati on of reproductive organ developmen t of conifers and angiosperms. Plant J 15: 615-623. Taylor, SA, JM Hofer, IC Murfet, JD Solli nger, SR Singer, MR Knox, TH Ellis 2002 PROLIFERATING INFLORESCENCE MERISTEM a MADS-box gene that regulates floral meristem identity in pea. Plant Physiol 129: 1150-1159. Teeri, TH, M Kotilainen, A Uimari, S Ruokolai nen, YP Ng, U Malm, E Pollanen, S Broholm, R Laitinen, P Elomaa, VA Albert 2006 Floral developmental genetics of Gerbera (Asteraceae). Adv Bot Res 44: 323-351. Theissen, G 2001a Development of floral organ identity: stories from the MADS house. Curr Opin Plant Biol 4: 75-85. Theissen, G 2001b Flower development, genetics of. in : S Brenner, JH Miller eds. Encyclopedia of Genetics. Academic Press. Theissen, G, A Becker 2004 Gymnosperm orthologues of class B floral homeotic genes and their impact on understanding flower orig in. CRC Crit Rev Plant Sci 23: 129-148. Theissen, G, A Becker, A Di Rosa, A Kanno, JT Kim, T Munster, KU Winter, H Saedler 2000 A short history of MADS-box genes in plants. Plant Mol Biol 42: 115-149. Theissen, G, A Becker, K-U Winter, T Mnste r, C Kirchner, H Saedler 2002 How the land plants learned their floral ABCs : the role of MADS-box genes in the evolutionary origin of flowers. in : Q Cronk, R Bateman, J Hawkins eds. Developmental Genetics and Plant Evolution. Taylor & Francis, London, UK. Theissen, G, JT Kim, H Saedler 1996 Classification and phylogeny of the MADS-box multigene family suggest defined roles of MADSbox gene subfamilies in the morphological evolution of eukaryotes J Mol Evol 43: 484-516. Theissen, G, R Melzer 2007 Molecular mechanisms underlying origin and di versification of the angiosperm flower. Ann Bot 100: 603-619. Thompson, JD, TJ Gibson, F Plewniak, F Jean mougin, DG Higgins 1997 The ClustalX windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Res 25: 4876-4882. Thorne, JL, H Kishino 2002 Divergence time and e volutionary rate estimation with multilocus data. Syst Biol 51: 689-702. Thorne, JL, H Kishino, IS Painter 1998 Estimating th e rate of evolution of the rate of molecular evolution. Mol Biol Evol 15: 1647-1657.
183 Trbner, W, L Ramirez, P Motte, I Hue, P Huijser, WE Lonnig, H Saedler, H Sommer, Z Schwarz-Sommer 1992 GLOBOSA : a homeotic gene which interacts with DEFICIENS in the control of Antirrhinum floral organogenesis. EMBO J 11: 4693-4704. Tucker, SC, AW Douglas 1996 Floral structure, development, and relationships of paleoherbs: Saruma Cabomba, Lactoris and selected Piperales. in : DW Taylor, LJ Hickey eds. Flowering plant origin, evol ution, and phylogeny. Chapman & Hall, New York, NY, USA. Tzeng, TY, HY Chen, CH Yang 2002 Ectopic expr ession of carpel-specific MADS box genes from lily and lisianthus causes similar hom eotic conversion of sepal and petal in Arabidopsis Plant Physiol 130: 1827-1836. Vandenbussche, M, J Zethof, S Royaert, K We terings, T Gerats 2004 The duplicated B-class heterodimer model: whorl-specific effect s and complex geneti c interactions in Petunia hybrida flower development. Plant Cell 16: 741-754. Vogel, S, F Hadacek 2004 Contributions to the functional anatomy and biology of Nelumbo nucifera (Nelumbonaceae) III. An ecological reappraisa l of floral organs. Pl Syst Evol 249: 173-189. Wang, XQ, DC Tank, T Sang 2000 Phylogeny and dive rgence times in Pinaceae: evidence from three genomes. Mol Biol Evol 17: 773-781. Wang, XR, Y Tsumura, H Yoshimaru, K Nagasaka AE Szmidt 1999 Phylogenetic relationships of Eurasian pines ( Pinus Pinaceae) based on chloroplast rbcL matK RPL20-RPS18 spacer, and trnV intron sequences. Am J Bot 86: 1742-1753. Wang, Z, Y Liang, C Li, Y Xu, L Lan, D Zhao, C Chen, Z Xu, Y Xue, K Chong 2005 Microarray analysis of gene expression involved in anther development in rice ( Oryza sativa L.). Plant Mol Biol 58: 721-737. Warner, KA, PJ Rudall, WF Michael 2008 Differen tiation of perianth orga ns in Nymphaeales. Taxon 57: 1096-1109. Weigel, D 1995 The genetics of flower deve lopment: from floral induction to ovule morphogenesis. Annu Rev Genet 29: 19-39. Wellmer, F, M Alves-Ferreira, A Dubois, JL Riechmann, EM Meyerowitz 2006 Genome-wide analysis of gene expression during early Arabidopsis flower development. PLoS Genet 2: e117. Wellmer, F, JL Riechmann 2005 Gene network analysis in plant development by genomic technologies. Int J Dev Biol 49: 745-759. Wellmer, F, JL Riechmann, M Alves-Ferreira, EM Meyerowitz 2004 Genome-wide analysis of spatial gene expression in Arabidopsis flowers. Plant Cell 16: 1314-1326.
184 Whipple, CJ, P Ciceri, CM Padilla, BA Am brose, SL Bandong, RJ Schmidt 2004 Conservation of B-class floral homeotic ge ne function between maize and Arabidopsis Development 131: 6083-6091. Wiersema, JH 1988 A monograph of Nymphaea subgenus Hydrocallis (Nymphaeaceae). American Society of Plant Taxonomists, Ann Arbor, MI, USA. Wiersema, JH, CB Hellquist 1998 Nymphaeaceae. in : FoNAE Committee ed. Flora of North America. Vol. 3. Oxford University Press, New York, NY, USA. Williamson, PS, MF Moseley 1989 Morphological Studies of the Nymphaeaceae sensu lato. XVII. Floral anatomy of Ondinea purpurea subspecies purpurea (Nymphaeaceae). Am J Bot 76: 1779-1794. Williamson, PS, EL Schneider 1993 Cabombaceae. in : K Kubitzki, JG Rohwer, V Bittrich eds. The families and genera of vascular plan ts. Vol. 2. Springer, Berlin, Germany. Williamson, PS, EL Schneide r 1994 Floral aspects of Barclaya (Nymphaeaceae): pollination, ontogeny and structure. Pl Syst Evol 8 suppl: 159-173. Winter, KU, A Becker, T Munster, JT Kim, H Saedler, G Theissen 1999 MADS-box genes reveal that gnetophytes are more closely related to conifers than to flowering plants. Proc Natl Acad Sci U S A 96: 7342-7347. Winter, KU, C Weiser, K Kauf mann, A Bohne, C Kirchner, A Kanno, H Saedler, G Theissen 2002 Evolution of class B floral homeotic proteins: obligate heterodimerization originated from homodimerization. Mo l Biol Evol 19: 587-596. Worberg, A, D Quandt, AM Barniske, C Lohne KW Hilu, T Borsch 2007 Phylogeny of basal eudicots: Insights from non-coding and ra pidly evolving DNA. Organisms diversity & Evolution 7: 55-77. Yang, YH, NP Thorne 2003 Normalization for two-color cDNA microarray data. in : DR Goldstein ed. Science and Statistics: A Fest schrift for Terry Speed, IMS Lecture Notes Monograph Series. Vol. 40. Institute of Mathematical Statistics, Beachwood, OH. Yang, Z 1994 Maximum likelihood phylogenetic estim ation from DNA sequences with variable rates over sites: approximate methods. J Mol Evol 39: 306-314. Yang, Z 2000 Phylogenetic analysis by maximum likelihood (PAML). University College London, London, UK. Yanofsky, MF, H Ma, JL Bowman, GN Drews, KA Feldmann, EM Meyerowitz 1990 The protein encoded by the Arabidopsis homeotic gene agamous resembles transcription factors. Nature 346: 35-39.
185 Yu, D, M Kotilainen, E Pollanen, M Mehto, P El omaa, Y Helariutta, VA Albert, TH Teeri 1999 Organ identity genes and modified patterns of flower development in Gerbera hybrida (Asteraceae). Plant J 17: 51-62. Yu, H, T Ito, F Wellmer, EM Meyerowitz 2004 Repression of AGAMOUS-LIKE 24 is a crucial step in promoting flower development. Nat Genet 36: 157-161. Yu, HJ, P Hogan, V Sundaresan 2005 Analysis of the female gametophyte transcriptome of Arabidopsis by comparative expression profiling. Plant Physiol 139: 1853-1869. Zachgo, S, A Silva Ede, P Motte, W Trobner, H Saedler, Z Schwarz-Sommer 1995 Functional analysis of the Antirrhinum floral homeotic DEFICIENS gene in vivo and in vitro by using a temperature-sensitive muta nt. Development 121: 2861-2875. Zahn, LM, H Kong, JH Leebens-Mack, S Kim, PS Soltis, LL Landherr, DE Soltis, CW Depamphilis, H Ma 2005 The evolution of the SEPALLATA subfamily of MADS-box genes: a pre-angiosperm origin with multip le duplications throughout angiosperm history. Genetics 169: 2209-2223. Zanis, MJ, DE Soltis, PS Soltis, S Mathews, MJ Donoghue 2002 The root of the angiosperms revisited. Proc Natl Acad Sci U S A 99: 6848-6853. Zanis, MJ, PS Soltis, YL Qiu, E Zimmer, DE Soltis 2003 Phylogenetic analyses and perianth evolution in basal angiosperms. Ann Mo Bot Gard 90: 129-150. Zhang, L-B, S Renner 2003 The deepest split in Chloranthaceae as resolved by chloroplast sequences. Int J Plant Sci 164 Suppl: S383-S392. Zhang, L, Y Xu, R Ma 2008 Molecular cloning, iden tification, and chromosomal localization of two MADS box genes in peach ( Prunus persica ). J Genet Genomics 35: 365-372. Zhang, P, HT Tan, KH Pwee, PP Kumar 2004 Conserva tion of class C function of floral organ development during 300 million years of evolution from gymnosperms to angiosperms. Plant J 37: 566-577. Zhang, X, B Feng, Q Zhang, D Zhang, N Altman, H Ma 2005 Genome-wide expression profiling and identification of gene activities during ea rly flower development in Arabidopsis. Plant Mol Biol 58: 401-419. Zhao, D, Q Yu, C Chen, H Ma 2001 Genetic control of reproductive meristems. in : MT McManus, B Veit eds. Meristematic tissues in plant growth and development. Sheffield Academic Press, Sheffield, England. Zomlefer, WB 1994 Nymphaeaceae. in : WB Zomlefer ed. Guide to flowering plant families. University of North Carolina Press, Chapel Hill, NC.
186 BIOGRAPHICAL SKETCH Mi-Jeong Yoo was born in Chungju, Korea, a nd spent her childhood in the country, where she developed her interest in na ture. She received her Bachelor of Science in biology from Seoul national university, Seoul, Korea, in March 1998. She began her grad uate career in the Department of Biological Sciences at Seoul National University, and received her Master of Science in March 2000. After gr aduation, she decided to study abroad. While preparing her application, she worked at the herbarium of Seoul National University (SNU) as a specimen manager. Mi-Jeong was admitted to the University of Florida, so she started her doctoral work with Drs. Douglas and Pamela Soltis in Fall 2002. She studied the floral developmental genetics and evolution in water lilies. Mi-Jeong gra duated with a PhD in bot any in December 2008.