Evolution of petaloid sepals independent of shifts in B-class MADS box gene expression
http://www.springerlink.com ( Publisher's URL )
Full Citation
Permanent Link: http://ufdc.ufl.edu/IR00001229/00001
 Material Information
Title: Evolution of petaloid sepals independent of shifts in B-class MADS box gene expression
Series Title: Development Genes and Evolution 2012 222:19-28 DOI 10.1007/s00427-011-0385-1
Physical Description: Journal Article
Creator: Landis, Jacob B.
Publisher: SpringerLink
Place of Publication: Development Genes and Evolution
Publication Date: December 24, 2011
Abstract: Attractive petals are an integral component of animal-pollinated flowers, and in many flowering plant species are restricted to the second floral whorl. Interestingly, multiple times during angiosperm evolution, petaloid characteristics have expanded to adjacent floral whorls or to extra-floral organs. Here we investigate developmental characteristics of petaloid sepals in Rhodochiton atrosanguineum, a close relative of the model species Antirrhinum majus (snapdragon). We undertook this in two ways, first using scanning electron microscopy (SEM) we investigate micromorphology of petals and sepals, followed by expression studies of genes usually responsible for the formation of petaloid structures. From our data, we conclude that R. atrosanguineum petaloid sepals lack micromorphological characteristics of petals, and that petaloid sepals did not evolve through regulatory evolution of B-class MADS box genes, which have been shown to specify second whorl petal identity in a number of model flowering plant species including snapdragon. These data, in conjunction with other studies, suggests multiple convergent pathways for the evolution of showy sepals.
Acquisition: Collected for University of Florida's Institutional Repository by the UFIR Self-Submittal tool. Submitted by Jacob Landis.
Publication Status: Published
 Record Information
Source Institution: University of Florida Institutional Repository
Holding Location: University of Florida
Rights Management: Applicable rights reserved.
System ID: IR00001229:00001


This item has the following downloads:

manuscript ( PDF )

manuscript ( DOC )

Full Text


Landis et al 1 Evolution of petaloid sepals independent of shifts in B class MADS box gene expression 1 2 Jacob B. Landis 1,2 Laryssa L. Barnett 3 and Lena C. Hileman 3 4 Department of Ecology and Evolutionary Biology, University of Kansas, 1200 Sunnyside Avenue, Lawrence, KS 5 6 6045, USA 6 7 1 Author for correspondence email: jblandis@ufl.edu, phone: 352 273 2977 fax : 352 392 3704 8 9 2 Present Address: Department of Biology, University of Florida, Gainesville, Florida 32608 10 3 Present Address: Department of Biology, Duke University, Durh am, North Carolina 27708 11 12 13 14 15 16 17 18 19 20 21 22 23


Landis et al 2 ABSTRACT 24 Attractive petals are an integral component of animal pollinated flowers, and in many flowering plant spec ies are 25 restricted to the second floral whorl. Interestingly, multiple times during angiosperm evoluti on, petaloid 26 characteristics have expanded to adjacent floral whorls or to extra floral organs. Here we investigate developmental 27 characteristics of petaloid sepals in Rhodochiton atrosanguineum a close relative of the model species Antirrhinum 28 majus (sna pdragon). We undertook this in two ways, first using scanning electron microscopy (SEM) we investigate 29 micromorphology of petals and sepals, followed by expression studies of genes usually responsible for the formation 30 of petaloid structures. From our dat a, we conclude that R. atrosanguineum petaloid sepals lack micromorphological 31 characteristics of petals, and that petaloid sepals did not evolve through regulatory evolution of B class MADS box 32 genes, which have been shown to specify second whorl petal ide ntity in a number of model flowering plant species 33 including snapdragon. These data, in conjunction with other studies, suggests multiple convergent pathways for the 34 evolution of showy sepals. 35 36 Key words: ABC model; Antirrhinum majus (snapdragon); B class genes ; petal identity; Rhodochiton 37 atrosanguineum 38 39 INTRODUCTION 40 Flowers show a wide range of diversity, much of which can be attributed to evolutionary changes in the 41 size, shape, number and color of petals or petaloid organs, which function to attract po llinators. It has been 42 suggested that petals have been gained ( De Craene 2007, Rasmussen et al. 2009, Soltis et al. 2009, Zanis et al. 2003) 43 and lost ( Jaramillo and Kramer 2007, Wu et al. 2007) multiple times during angiosperm evolution Indeed, in some 44 cases, such as, but not limited to, the genus Cornus additional floral and extra floral organs, including bracts, have 45 evolved petaloid features ( Brockington et al. 2009, Geuten et al. 2006, Maturen 2008, Rasmussen et al. 2009, Zhang 46 et al. 2008) 47 The ty pical core eudicot flower consists of four concentric whorls of floral organs. The two outer most 48 organ whorls develop into a differentiated perianth surrounding the reproductive organs. In this typical flower, leaf 49 like sepals develop in the outer perian th whorl known as the calyx, and petals occupy the inner perianth whorl 50 referred to as the corolla. Petals are often highly complex and are morphologically distinct from leaves and sepals 51


Landis et al 3 due to several characteristics, including the presence of colored p igmentation, conical or elongated epidermal cell 52 shape, lack of stomata, and lack of palisade mesophyll (reviewed in De Craene 2008, Glover 2007, Irish 2009) 53 Another characteristic that distinguishes petals from sepals and leaves is the necessary gene pr oducts for petal 54 specification (reviewed in Irish 2009) Interestingly, many angiosperms develop flowers which do not conform to 55 the typical form described above most notably, those in which all perianth organs exhibit a petaloid appearance 56 ( e.g., orchi ds, lilies, columbines, magnolias; reviewed in Kramer 2007, Litt and Kramer 2010) and the extent to 57 which outer whorl perianth organs adopt petaloid characteristics has been studied in multiple taxa ( Geuten et al. 58 2006, Kanno et al. 2003, Maturen 2008, Na kamura et al. 2005, Park et al. 2004) 59 For the typical flower with a differentiated perianth, the ABC model posits that the combination of three 60 classes of genes, termed A, B, and C, function in overlapping domains to specify the formation of the differe nt floral 61 organ identities ( Bowman et al. 1991, Coen and Meyerowitz 1991, Coen et al. 1991, Schwarz Sommer et al. 1990, 62 Trobner et al. 1992) Presence of A class function determines sepals in the outer whorl, with co occurrence of A 63 and B class function determining petal identity in the second whorl. Stamens are determined in the third whorl by 64 co occurrence of B and C class function, and in the fourth whorl, occurrence of C class function determines the 65 identity of carpels. Therefore, based on the ABC model, B class proteins play a critical developmental role in 66 establishing petal identity, and in the differentiation of petal from sepal identity within the perianth. 67 B class genes comprise two lineages: the APETALA3 / DEFICIENS lineage ( Jack et al. 1992, Sommer et al. 68 1990) and the PISTILATA/GLOBOSA lineage ( Goto and Meyerowitz 1994, Trobner et al. 1992) Since their original 69 characterization in Arabidopsis thaliana L. (Brassicaceae) and Antirrhinum majus L. (Plantaginaceae) ( Bowman et 70 al. 1989, Carpente r and Coen 1990, Krizek and Meyerowitz 1996, Schwarz Sommer et al. 1990, Sommer et al. 71 1990) many additional studies have demonstrated that B class function in establishing second whorl petal identity is 72 conserved across angiosperms including flowering tob acco ( Nicotiana benthamiana Domin, Solanaceae; Liu et al. 73 2004) tomato ( Solanum lycopersicum L., Solanaceae; de Martino et al. 2006) petunia ( Petunia hybrida Juss., 74 Solanaceae; Rijpkema et al. 2006, Vandenbussche et al. 2004) poppy ( Papaver somniferum L ., Papaveraceae; Drea 75 et al. 2007 ) and represents a conserved developmental pathway for lodicule specification in grasses (Kang et al. 76 1998, Ambrose et al. 2000, Prasad and Vijayraghavan 2003, Xiao et al. 2003, Whipple et al. 2004). 77 As mentioned above m any core eudicot species develop perianths that are undifferentiated or exhibit 78 reduced differentiation and therefore may not poses floral developmental programs that conform fully to the 79


Landis et al 4 canonical ABC model. To explain the developmental genetic program underlying undifferentiated perianths, the 80 sliding boundary model was developed ( Bowman 1997, Kramer et al. 2003) According to this model, A and B 81 class function in both the inner and outer whorl perianth organs leads to an expansion of petal identity a cross the 82 entire perianth. Data from multiple species, including tulip ( Tulipa gesneriana L., Liliaceae; Kanno et al. 2003) lily 83 ( Lilium longiflorum Thunb., Liliaceae; Tzeng and Yang 2001 ) lily of the nile ( Agapanthus praecox Willd., 84 Alliaceae; Nakamura et al. 2005) and water lilies ( Cabomba caroliniana Gray, Cabombaceae; Yoo et al. 2010) 85 support this hypothesized genetic model as an explanation for the presence of an entirely petaloid perianth. 86 However, many other species possess perianths that exhibit differentiation but with sepals that appear petaloid. For 87 example, sepals may appear petaloid in shape, coloration or micromorphology, but not develop identically to second 88 whorl petals Examples of these flowers include columbine ( Aquilegia vulgaris L., Ranunculaceae; Kramer et al. 89 2007) common heather ( Calluna vulgaris (L.) Hull, Ericaceae; Borchert et al. 2009) gerbera ( Gerbera hybrida L., 90 Asteraceae; Broholm et al. 2010), orchids ( Habenaria radiata (Thunb.) Spreng., Orchidaceae; Kim et al. 2007) 91 im patiens ( Impatiens hawkeri W. Bull, Balsaminaceae; Geuten et al. 2006) garden asparagus ( Asparagus officinalis 92 L., Asparagaceae; Park et al. 2003, Park et al. 2004) and members of the Aizoaceae ( Brockington et al. 2011; 93 Brockington et al. 2009) Within this diverse group, there does not appear to be a clear developmental genetic 94 program explaining the formation of first whorl petaloid organs. 95 The focus of this study is Rhodochiton atrosanguineum L. (Plantaginaceae), a close relative of the model 96 species snapdragon (Fig 1a). Rhodochiton atrosanguineum flowers do not appear to phenotypically adhere to the 97 ABC model in the same fashion as snapdragon flowers with their distinct sepals and petals (Fig. 1 c ) Rhodochiton 98 atrosanguineum flowers have outer and i nner whorl perianth organs that are morphologically distinct, yet both 99 whorls of organs exhibit a petaloid appearance (Fig 1 b ). In this study we aim to determine the extent to which outer 100 whorl perianth organs of R. atrosanguineum exhibit petal identity b eyond pigmentation of sepals. Therefore, we 101 investigate perianth micromorphology with the specific hypothesis that R. atrosanguineum outer whorl perianth 102 organs will exhibit characteristics of epidermal cell shape resembling inner whorl petals. Additiona lly, we test the 103 applicability of the ABC genetic model, with the specific prediction that expression of B class orthologs DEF and 104 GLO will be detected in R. atrosanguineum outer whorl perianth organs if they have adopted petal identity. 105 METHODS 106


Landis et al 5 Plant mat erial Seeds of R. atrosanguineum were obtained from B and T World Seeds (http://www.b and 107 t world seeds.com) and A. majus seeds, accession number ANTI 11 (D2836), were obtained from the Gatersleben 108 collection (Leibniz Institute of Plant Genetics and Crop Research, http://www.ipk gaterlseben.de). Voucher 109 specimens of R. atrosanguineum (JL001 and JL002) and A. majus (JL004) have been placed in the R. L. McGregor 110 Herbarium (KANU), University of Kansas. Flower material for both species was collected from pla nts grown in the 111 greenhouse at the University of Kansas. 112 Scanning electron microscopy Mature flowers of A. majus and R. atrosanguineum were fixed in 113 glutaraldehyde (5% glutaraldehyde solution in 0.1 M phosphate buffer) overnight and then dehydrated throu gh an 114 ethanol series. Dehydrated flowers were critical point dried using a Tousimis critical point dryer and then dissected 115 into sepals, base of petal tube, and petal lobe. Tissue was collected for imaging both adaxial and abaxial surfaces. 116 Specimens we re mounted on stubs, sputter coated with gold, and viewed with a D. Leo field emissi on scanning 117 electron microscope. 118 Isolation of RaDEF and RaGLO orthologs Total RNA was isolated from immature R. atrosanguineum 119 flowers using Tri Reagent following the man ufacturer instructions (Ambion, Austin, Texas, USA) and DNAse 120 treated using TurboDNA (Ambion, Austin, Texas, USA). cDNA was generated using 1 g of total RNA in a 15 l 121 cDNA synthesis reaction using iScript Synthesis Kit following the manufacturer instru ctions ( BioRad Hercules 122 California, USA). Orthologs of DEF and GLO were isolated from floral cDNA by reverse transcriptase PCR (RT 123 PCR) using the degenerative forward primer (5' AACAGGCARCTIACITAYTC verse PolyT QT 124 primer (5' GACTCGAGTCGACATGGA(T) 18 ( Hileman et al. 2006) This primer combination yields near full 125 length gene sequences, lacking only the 21 amino acids at the 5' end of the gene. RT PCR reactions contained 2 ul 126 of 1:10 diluted cDNA, 1.25 units Taq (Sigma Aldrich, St. Louis, Missouri, USA), 10X PCR buffer, 0.5 M of each 127 primer, and 0.8 mM dNTPs. PCR reactions were run for 40 cycles with an annealing temperature of 47 C. PCR 128 products were subjected to gel electrophoresis using a 1.5% agarose gel and gel purified using the Wizard SV Gel 129 and PCR Clean Up System Kit (Promega, Madison, Wisconsin, USA) before being cloned. Gel purified PCR 130 products were cloned into the pGEM T vector system (Promega, Madison, Wisconsin, USA) following the 131 manufacturer instructions. Twenty clones were sequenced using M13 forward and M13 reverse primers in order to 132 identify multiple gene copies amplified by our RT PCR approach ( Howarth and Baum 2005) 133


Landis et al 6 Phylogenetic Analysis Putative orthologs of RaDEF an d RaGLO were aligned to additional B class 134 ( DEF/GLO like) genes down loaded from Genbank using MUSCLE ( Edgar 2004) followed by manual adjustment 135 in MacClade v4.08 ( Maddison and Maddison 2005) Nucleotide sequence alignments were used to generate 136 estimates of the gene phylogeny under Maximum Parsimony (MP), Maximum Likelihood (ML) and Bayesian 137 criteria. Maximum likelihood was implemented in Garli ( Zwickl 2006) using the GTR + I + model of molecular 138 evolution. Support values using ML were generated with 1 000 bootstrap replicates in Garli as described above. 139 Support values were also generated using MP in Paup* 4.0 ( Swofford 2002) with 1000 heuristic bootstrap replicates 140 and the TBR branch swapping algorithm. Bayesian criterion was implemented using MrBayes ( Huelsenbeck and 141 Ronquist 2001, Ronquist and Huelsenbeck 2003) and the GTR + I + model of molecular evolution with two 142 Markov chains running for 1,000,000 generations sampling every 100th generation. At completion of runs, the two 143 chains were checked f or convergence and the first 25% of saved trees were discarded as initial burn in. Remaining 144 trees were used to calculate posterior probabilities of node support in a 50% majority rule consensus tree. 145 Expression of RaDEF and RaGLO by Reverse Transcriptas e (RT) PCR Rhodochiton atrosanguineum 146 RNA was extracted and cDNA generated from the four floral organ types of multiple flowers: outer whorl petaloid 147 sepals, petals, stamens and carpels, in three distinct floral size classes. The small size class include d the earliest 148 stage flower buds t hat could be hand dissected, with corolla length in this size class ranging from 4.0 to 7.0 mm. 149 Corolla length in the medium size class had a range of 15.0 to 18.0 mm. The large size class included flowers just 150 pre anthe sis, with corolla length in this size class ranging from 39.5 to 40.5 mm. Expression patterns were 151 characterized using gene specific primers designed to amplify fragments of RaDEF and RaGLO of ca. 150 200 bp 152 of the open reading frame. Gene specific prime rs for amplifying RaDEF were RaDEF F ( 5' 153 AGCTTGAACGATCTGGGCTA 3') and RaDEF R (5' GTGCGGATCCTCTCTTCTTG 3') and primers for 154 amplifying RaGLO were RaGLO F (5' GGGACGTCAGCTCTCAAAA 3') and RaGLO R (5' 155 ATCGTATACCCCCTGGCTTT 3'). ACTIN was used a loading contro l as described by Prasad et al. (2001) RT 156 PCR reactions included 2 l of 1:10 diluted cDNA, 1.25 units Taq (Sigma Aldrich, St. Louis, Missouri, USA), 10X 157 PCR buffer, 0.5 M of each primer, and 0.8 mM dNTPs. PCR conditions consisted of 26 cycles with an annealing 158 temperature of 55 C for all genes tested. The number of cycles was determined as that representing the linear range 159 of amplification from a PCR product curve i ncluding reactions run for 22 40 cycles. Triplicate RT PCR reactions 160 for each cDNA, i ncluding RT PCR negative control cDNAs ( RT), were conducted to ensure consistency. 161


Landis et al 7 Expression of RaDEF and RaGLO by in situ mRNA hybridization Flower buds of R. atrosanguineum 162 were fixed in FAA (47.5% ethanol, 5% acetic acid, 3.7% formaldehyde) for 8 ho urs, stained with eosin Y, 163 dehydrated and wax embedded as described by Jackson (1991) and Preston and Kellogg (2007) Gene specific 164 probe templates of RaDEF and RaGLO were generated using primers RaDEF F (5' 165 AATACATCAGTCCCACCACAGC 3'), RaDEF R (5' GCAAAGC AAATGTGGTAAGGTC 3'), RaGLO F (5' 166 TCATCATCTTTGCTAGTTCTG 3'), and RaGLO R (5' TCCTGAAGATTAGGCTGCATTG 3'). Probes were 520 167 bp and 488 bp long for RaDEF and RaGLO respectively, with forward primers in the I domain and reverse primers 168 in the C terminal coding region of each gene ( Yang et al. 2003) to exclude amplification of the highly conserved 169 MADS domain. All PCR products for probe generation were cloned into the pGEM T vector (Promega, Madison, 170 Wisconsin, USA) and confirmed by sequencing. Sense and antis ense riboprobes for RaDEF and RaGLO were 171 generated using T7 and SP6 RNA polymerase (Roche, Indianapolis, Indiana, USA) incorporating digoxygenin 172 dNTPs (Roche, Indianapolis, Indiana, USA) according to the manufacturers instructions. Probe hydrolysis follow ed 173 Jackson (1991) to yield fragments c. 150 bp long. In situ hybridization was performed on longitudinal sections of 174 multiple inflorescences as in Jackson (1991 ). Multiple floral sections hybridized with antisense or sense probe were 175 visualized to ensure consistent assessment of gene expression. Images were documented using a Leica DM5000B 176 microscope attached to a Leica DFC300FX camera. Images were imported into Adobe Photoshop and adjusted for 177 contrast, brightness and color balance. 178 RESULTS 179 Micromorph ological analysis of petaloid sepals To determine whether R. atrosanguineum petaloid 180 sepals exhibit micromorphological cell shape characteristics found in R. atrosanguineum petals, SEM analyses were 181 undertaken. Figure 2 shows image comparisons of epiderm al cell shape from both the abaxial and adaxial surface of 182 snapdragon (Fig. 2a i ) and R. atrosanguineum (Fig. 2 j r ) leaves, sepals or petaloid sepals, and petals. Leaves (Fig. 183 2a,e, j n ), sepals and petaloid sepals (Fig. 2b,f, k o ) of both species show a co nsistent jigsaw shaped cellular pattern 184 on both the abaxial and adaxial surfaces, with stomata found predominantly on the abaxial surface of these organs 185 (Fig. 2a,b, j k ). Jigsaw shaped cells are also observed on the abaxial surface of both snapdragon and R. 186 atrosanguineum petal lobes (Fig. 2d, m ); interestingly, stomata are also found on the abaxial surface of R. 187 atrosanguineum petal lobes (Fig. 2 m ), but not the abaxial surface of snapdragon petal lobes (Fig. 2d). Rhodochiton 188 atrosanguineum abaxial petal lobe epidermal cells not only develop stomata but also appear more domed, or 189


Landis et al 8 lenticular, than corresponding snapdragon epidermal cells (Fig. 2d, m ). In both species elongated cells were found 190 on both adaxial and abaxial surfaces at the base of the corolla tube (Fig. 2c,g, l p ). The major difference between the 191 two species is found on the surface of adaxial petal lobes. The adaxial surface of A. majus petal lobes (Fig. 2h ,i ) 192 exhibit papillose conical cells as previously documented ( Noda et al. 1994, Perez Rodriguez et al. 2005) while the 193 adaxial petal lobes of R. atrosanguineum (Fig 2 q,r ) lack conical cells. Rhodochiton atrosanguineum adaxial petal 194 lobe epidermal cells are more domed than the papillose conical epidermal cells found in snapdragon, and are 195 referred to as lenticular ( Kay et al. 1981) SEM images show no distinct micromorphological differences between 196 petaloid sepals and leaf like sepals of R. atrosanguineum and snapdragon, respectively, and R. atrosanguineum 197 petaloid sepals do not resemble R atrosanguineum petals at the micromorphological scale. 198 Isolation and phylogenetic assessment of RaDEF and RaGLO Two B class genes were isolated from 199 floral cDNA of R. atrosanguineum and deposited in GenBank ( RaDEF JQ173625 ; RaGLO JQ173626 ). ML, MP 200 and Bayesian phylogenetic estimates place one of these two genes ( RaDEF ) in a well supported clade with A. majus 201 DEF and the other gene ( RaGLO ) in a clade with A. majus GLO (Fig. 3). RaDEF is nested within a clade of 202 orthologs from snapdragon and Misopates orontium (L.) Raf. (Plantaginaceae), all of which are members of the tribe 203 Antirrhineae. This clade has bootstrap support values of 97% (MP) and 98% (ML), and a posterior probability of 204 1.0 (Fig. 3). The placement of RaGLO is also in a well supported cla de with GLO like genes from snapdragon and 205 M. orontium with bootstrap values of 100% (MP) and 90% (ML) and a posterior probability of 1.0 (Fig. 3). The 206 sister relationship between snapdragon and M. orontium is highly supported and reflects species relati onships based 207 on other molecular markers ( Vargas et al. 2004) 208 Expression of RaDEF and RaGLO Scoring of RT PCR gene expression was dichotomous presence or 209 absence of RaDEF or RaGLO transcripts in sampled tissues. RaDEF and RaGLO transcripts were dete cted by RT 210 PCR in petals and stamens, but not in petaloid sepals or carpels, for the three size classes of flowers (Fig. 4 ). In 211 addition RaDEF (Fig. 5a) and RaGLO (Fig. 5c) expression was detected in developing petals and stamens, but not 212 developing sepa ls of early stage R. atrosanguineum flowers by in situ mRNA hybridization. These results 213 demonstrate that from early through late stages of flower development, the B class genes, RaDEF and RaGLO are 214 not expressed in R. atrosanguineum petaloid sepals. 215 DISC USSION 216 Comment [JL1]: Ove rly repetitive information regarding the results of gene expression was removed.


Landis et al 9 Rhodochiton atrosanguineum sepals are brightly pigmented giving them a superfic ial petaloid appearance 217 (Fig. 1b ). The objective of this study was to determine to what extent R. atrosanguineum petaloid sepals have 218 adopted petal characteristics at t he micromorphological and molecular level. Specifically, we used SEM and gene 219 expression studies to test the following hypotheses: 1) R. atrosanguineum petaloid sepals exhibit 220 micromorphological characteristics found in adjacent petals, and 2) petaloid se pal development in R. 221 atrosanguineum is associated with B class MADS box gene regulatory evolution. We found that petaloid sepals of 222 R. atrosanguineum do not morphologically resemble the petals of R. atrosanguineum similar to studies in Impatiens 223 ( Geuten et al. 2006) Additionally, we found that the ABC genetic model, specifically expression of B class genes 224 restricted to the second and third flower whorls, is conserved in R. atrosanguineum despite development of petaloid 225 sepals. 226 If, in R. atrosanguine um the sepals have evolved a petaloid appearance due to expansion of petal identity 227 to the first floral whorl, then we expect epidermal cell shape in these outer whorl organs to resemble epidermal cell 228 shapes found in the second whorl petals Rhodochiton atrosanguineum petaloid sepals develop jigsaw shaped cells, 229 very similar to the cells found in sepals of snapdragon (Figs. 2b and 2 k ). These jigsaw shaped cells are distinct from 230 the adaxial epidermal cells of R. atrosanguineum petals (Figs. 2h ,i and 2 q, r ), which are dome shaped and distinguish 231 second whorl petals at the micromorphological level from R. atrosanguineum petaloid sepals and leaves It is 232 noteworthy that conical or papillose cell shape is found on the adaxial petal surface of many flowering p lant species 233 ( Christensen and Hansen 1998, De Craene 2008, Kay et al. 1981, Whitney and Glover 2007) and these predicted 234 papillose cells were found on the adaxial epidermis of snapdragon petals as previously described ( Noda et al. 1994, 235 Perez Rodriguez et al. 2005) (Fig. 2h and i ). Cell micromorphology of R. atrosanguineum petal lobes differs from 236 snapdragon lacking true papillose cells shape (Figs. 2h and 2p). The ultimate cause of micromorphological 237 differences between R. atrosanguineum and snapdragon petals is unknown but may reflect evolutionary shifts in 238 biotic pollination mechanisms ( Cronk and Ojeda 2008, Di Stilio et al. 2009); s napdragon is bee pollinated ( Glover 239 and Martin 1998, Whitney et al. 2009) while R. atrosanguineum is pollinated by humm ingbirds ( Sutton 1988) 240 A single copy of DEF and GLO were isolated from R. atrosanguineum and, based on our phylogenetic 241 estimates (Fig. 3), are orthologous to DEF and GLO from snapdragon, respectively. The single copy of RaDEF and 242 RaGLO are likely the only B class homologs in R. atrosanguineum The degenerative forward primer that was used 243 to isolate these genes was situated in the MADS domain which is highly conserved ( Yang et al. 2003) and the 244 Comment [JL2]: Information about cell shape was condensed to make the discussion more concise.


Landis et al 10 primer was designed to encompass variation in DEF and GL O across multiple Antirrhineae sequences. A 245 combination of RT PCR and in situ mRNA hybridization analyses demonstrate that both RaDEF and RaGLO 246 expression is restricted to the petals and stamens from early to late stage s of R. atrosanguineum flower develo pment 247 (Fig s 4 and 5 ). Therefore, expansion of B class gene expression to outer whorl perianth organs is not responsible for 248 the petaloid appearance of R. atrosanguineum sepals. Interestingly, other studies have implicated another MADS 249 box protein, SEP3 in conjunction with either DEF or GLO in the evolution of a petaloid first whorl floral organs 250 ( Geuten et al. 2006) Because neither RaDEF nor RaGLO are expressed in the petaloid sepals a model invoking the 251 combined action of RaSEP with either RaDEF or RaGLO protein in first whorl organs can be rejected Although R. 252 atrosanguinem outer whorl organs are petaloid in appearance primarily due to coloration they are morphologically 253 quite distinct from the inner whorl petals (Fig. 1b and 4 ) suggesting that changes in the anthocyanin pathway alone 254 ( Weiss 2000, Whittall et al. 2006) may underlie the evolution of petal like sepals 255 This study joins other studies that together indicate there is not a single developmental genetic model that 256 explains the evolutio n of petaloid sepals ( Borchert et al. 2009, Brockington 2009, Broholm et al. 2010, Geuten et al. 257 2006, Kim et al. 2007, Kramer et al., 2007, Park et al. 2004). Clearly convergent mechanisms lead to the 258 development of petaloid sepals, and may involve evolu tionary changes at the level of B class gene expression, 259 upstream or downstream of B class genes, or parallel pathways ( Jaramillo and Kramer 2004, Kramer et al. 2007, Litt 260 and Kramer 2010) Interestingly R. atrosanguineum develops a differentiated periant h (Fig. 1b) common to core 261 eudicots, lacks duplicates of DEF and GLO and gene expression data from this species does not support the 262 expansion of B class genes to the outer perianth whorl. Because B class genes play a critical role in establishing a 263 bipa rtite perianth, constraints are likely present for the evolution of petaloid sepals by mechanisms involving B class 264 gene regulation ( Hileman and Irish 2009, Kramer et al. 2003) Strikingly, when expansion of B class gene 265 expression is associated with the evolution of petaloid outer whorl perianth organs, it is restricted to species that 266 have a history of gene duplication in DEF and/or GLO lineages, or species lacking a differentiated perianth ( Kanno 267 et al. 2003, Litt and Kramer 2010, Nakamura et al. 2005, Sharma et al. 2011, Tzeng and Yang 2001, Yoo et al. 268 2010 ) This study provides support for the emerging pattern that there are multiple, convergent developmental 269 genetic mechanisms underlying independent transitions to petaloidy in the outer whorl perian th, and that constraint 270 lies, at least in part, in whether there is morphological differentiation within the perianth, and whether there is a 271 history of duplication in the B class gene lineage s 272 Comment [JL3]: Information was removed to shorten the discussion, as well as address comments from reviewers about SEP3 and the inclusion of anthocyanin pathway. Comment [JL4]: Kramer references removed from this section in response to reviewer's comments


Landis et al 11 273 FIGURE CAPTIONS 274 275 Fig. 1 Phylogenetic context of study spec ies. a) ITS phylogeny adapted from Vargas et al. (2004) with new 276 sequences added, showing relationships within the tribe Antirrhineae (Plantaginaceae). Focal genera Rhodochiton 277 and Antirrhinum are bold faced to exemplify their relationship to each other. b) Rhodochiton atrosanguineum with 278 petaloid sepals, c) Antirrhinum majus (snapdragon) with showy p ink petals and leaf like sepals 279 280 Fig. 2 Scanning electron microscope (SEM) images from Antirrhinum majus (snapdragon; a i) and Rhodochiton 281 atrosanguineum ( j r ). Individual SEM images of A. majus are a) abaxial leaf, b) abaxial sepal, c) abaxial base of 282 petal tube, d) abaxial petal lobe, e) adaxial leaf, f) adaxial sepal, g) adaxial base of petal tube h) adaxial petal lobe 283 top down, i) adaxial petal lobe 60 angled view. Individual SEM images of R. atrosanguineum are j) abaxial leaf, k) 284 abaxial sepal, l) abaxial base of petal tube, m) abaxial petal lobe n) adaxial leaf, o) adaxial sepal, p) adaxial base of 285 petal tube, q) adaxial petal lobe, r) adaxial peta l lobe 60 angled view Images show conserved jig saw shaped 286 patterns of cell shape in leaves and sepals of the two species, and conserved elongated tubular cells at the base of the 287 petal tube between both species. Rhodochiton atrosanguineum lacks conica l cells on the adaxial surface of the petal 288 lobe as seen in snapdragon where as the abaxial petal lobe of R. atrosanguineum has a more defined cell shape than 289 those seen in snapdragon. Scale bars in c, g p are 30 m, scale bars in h, i, q are 10 m, all others are 20 m 290 291 Fig. 3 Phylogeny of DEFICIENS ( DEF ) and GLOBOSA ( GLO ) orthologs showing placement of newly sequenced 292 Rhodochiton atrosanguineum DEF and GLO. Shown is the maximum likelihood (ML) tree with support values 293 from 1000 replicate Maximum Pars imony (MP) bootstrap analysis/1000 replicate ML bootstrap analysis/Bayesian 294 posterior probabilities. Only support values of >75% for bootstrap and >0.95 Bayesian posterior probabilities are 295 shown on the phylogeny. The following are GenBank accession numb ers for taxa other than R. atrosanguineum : 296 Antirrhinum majus L. ; AmDEF X52023, AmGLO AB516403. Arabidopsis thaliana L.; AtAP3 NM_115294, AtPI 297 NM_122031. Brassica napus L.; BnAP3 DQ372719. Camellia japonica L.; GQ141126. Chelone glabra L.; 298 CgDEF AY52400 8. Diospyros digyna Jacq.; DdGLO GQ141136. Lycopersicon esculentum L.; LeTAP3 299 DQ674532, LeTPI DQ674531. Mimulus guttatus DC.; MgDEFA AY524012, MgDEFB AY524020. Mimulus 300


Landis et al 12 kelloggi (Curran ex Greene) Curran ex A. Gray; MkDEF AY530545. Misopates orontium ( L.) Raf.; MoDEF 301 AM162207, MoGLO Am162211. Napoleona vogelii Hook. & Planch; NvGLO GQ141117. Papaver somniferum 302 L.; PsAP3 1 EF071993, PsAP3 2 EF071992, PsPI 1 EF071994, PsPI 2 EF071995. Paulownia tomentosa (Thunb.) 303 Steud.; PtDEF AY524018. Petunia hybrid a Juss.; PhDEF DQ539416. Phlox paniculata L.; PpDEF GQ141172, 304 GQ141129. Saxifraga caryana L.; ScAP3 DQ479367. Syringa vulgaris L.; SvAP3 DQ479367, SvPI 1 AF052861. 305 Torenia fournieri L.; TfGLO AB359952 306 307 Fig. 4 RT PCR conducted on cDNA generated from th ree size classes of R. atrosanguineum flowers separated into 308 their four floral organs. Small class flowers had a corolla length ranging from 4 7 mm, medium class flowers had a 309 corolla length ranging from 15 18 mm, with large class flowers having corolla l engths of 39.5 40.5 mm. For all size 310 classes, DEFICIENS and GLOBOSA were only expressed in the petals and stamens as seen by presence of bands in 311 these organs. ACTIN was used as a loading control during RT PCR analysis to insure the integrity cDNA. RT 312 samples served as negative controls, and no bands were visible in these reactions. ACTIN was also tested to show 313 integrity of cDNA samples. ACTIN was expressed in all cDNA samples but in none of the RT samples Petaloid 314 sepals = sep, petals = pet, stame ns = sta, and carpels = car 315 316 Fig. 5 In situ hybridization conducted on early stage flower buds of R. atrosanguineum. a) Expression of Ra DEF 317 using antisense probe. b) Sense probe control for RaDEF c) Expression of RaGLO using antisense probe. d) Sense 318 probe control for RaGLO. Dark blue staining using antisense probes for RaDEF and RaGLO show that expression 319 of these genes is limited to the developing petals and stamens of R. atrosanguineum Petaloid sepals = sep, petals = 320 pet, stamens = sta, and carpe ls = car 321 322 Acknowledgments 323 The authors thank Dr. Jill Preston for assistance with lab work and discussions on early versions of this manuscript, 324 Dr. David Moore and the imaging facility at the University of Kansas for help with SEMs, and Dr. Mark Mort and 325 l ab for help with p h ylogenetic analyses. This work was supported by the National Science Foundation (grant IOS 326 0616025 to L.C.H.) 327 328


Landis et al 13 References 329 Ambrose B, Lerner D, Ciceri P, Padilla C, Yanofsky M, Schmidt R (2000) Molecular and genetic analyses of the 330 silky 1 gene reveal conservation in floral organ specification between eudicots and monocots. Mol Cell 331 5:569 579 332 Borchert T, Eckardt K, Fuchs J, Krueger K, Hohe A (2009) 'Who's who' in two different flower types of Calluna 333 vulgaris (Ericaceae): morphological and molecular analyses of flower organ identity. BMC Plant Biology 334 9:148 335 Bowman JL (1997) Evolutionary conservation of angiosperm flower development at the molecular and genetic 336 levels. J Biosci 22:515 527 337 Bowman JL, Smyth DR, Meyerowitz EM (1991) Genetic int eractions among floral homeotic genes of Arabidopsis 338 Development 112:1 20 339 Bowman JL, Smyth DR, Meyerowitz EM (1989) Genes directing flower development in Arabidopsis Plant Cell 340 1:37 52 341 Brockington SF, Rudall PJ, Frohlich MW, Oppenheimer DG, Soltis PS, So ltis DE (2011) "Living stones" reveal 342 alternative petal identity programs within the core eudicots. Plant Journal 0.1111/j.1365 343 313X.2011.04797.x 344 Brockington SF, Alexandre R, Ramdial J, Moore MJ, Crawley S, Dhingra A, Hilu K, Soltis DE, Soltis PS (2009) 345 Ph ylogeny of the Caryophyllales sensu lato : Revisiting hypotheses on pollination biology and perianth 346 differentiation in the core Caryophyllales. Int J Plant Sci 170:627 643 347 Brockington SF (2009) Evolution and development of petals in Aizoaceae (Caryophyllal es). Dissertation, University 348 of Florida 349 Broholm SK, Pollanen E, Ruokolainen S, Tahtiharju S, Kotilainen M, Albert VA, Elomaa P, Teeri TH (2010) 350 Functional characterization of B class MADS box transcription factors in Gerbera hybrida J Exp Bot 351 61:75 85 352 C arpenter R, Coen E (1990) Floral homeotic mutations produced by transposon mutagenesis in Antirrhinum majus 353 Genes Dev 4:1483 1493 354 Christensen K, Hansen H (1998) SEM studies of epidermal patterns in the angiosperms. Opera Botanica 135:1 91 355


Landis et al 14 Coen E, Doyle S, Romero J, Elliott R, Magrath R, Carpenter R (1991) Homeotic genes controlling flower 356 development in Antirrhinum Development 113:149 155 357 Coen E, Meyerowitz E (1991) The war of the whorls genetic interactions controlling flower development. Nature 358 353:31 37 359 Cronk Q, Ojeda I (2008) Bird pollinated flowers in an evolutionary and molecular context. J Exp Bot 59:715 727 360 De Craene LPR (2008) Homology and evolution of petals in the core eudicots. Syst Bot 33:301 325 361 De Craene LPR (2007) Are petals sterile stame ns or bracts? The origin and evolution of petals in the core eudicots. 362 Annals of Botany 100:621 630 363 de Martino G, Pan I, Emmanuel E, Levy A, Irish VF (2006) Functional analyses of two tomato APETALA3 genes 364 demonstrate diversification in their roles in regu lating floral development. Plant Cell 18:1833 1845 365 Di Stilio VS, Martin C, Schulfer AF, Connelly CF (2009) An ortholog of MIXTA like2 controls epidermal cell shape 366 in flowers of Thalictrum New Phytol 183:718 728 367 Drea S, Hileman LC, De Martino G, Irish VF (2007) Functional analyses of genetic pathways controlling petal 368 specification in poppy. Development 134:4157 4166 369 Edgar R (2004) MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res 370 32:1792 1797 371 Geuten K, Becker A, Kaufmann K, Caris P, Janssens S, Viaene T, Theissen G, Smets E (2006) Petaloidy and petal 372 identity MADS box genes in the balsaminoid genera Impatiens and Marcgravia Plant Journal 47:501 518 373 Glover BJ (2007) Understanding flowers and flowering: an integra ted approach. Oxford University Press, Oxford; 374 New York 375 Glover B, Martin C (1998) The role of petal cell shape and pigmentation in pollination success in Antirrhinum 376 majus Heredity 80:778 784 377 Goto K, Meyerowitz E (1994) Function and regulation of the Arab idopsis floral homeotic gene PISTILLATA Genes 378 Dev 8:1548 1560 379 Hileman LC, Irish VF (2009) More is better: the uses of developmental genetic data to reconstruct perianth 380 evolution. Am J Bot 96:83 95 381 Hileman LC, Sundstrom JF, Litt A, Chen M, Shumba T, Irish VF (2006) Molecular and phylogenetic analyses of the 382 MADS box gene family in tomato. Mol Biol Evol 23:2245 2258 383


Landis et al 15 Howarth D, Baum D (2005) Genealogical evidence of homoploid hybrid speciation in an adaptive radiation of 384 Scaevola ( Goodeniaceae ) in the Hawaii an Islands. Evolution 59:948 961 385 Huelsenbeck J, Ronquist F (2001) MRBAYES: Bayesian inference of phylogenetic trees. Bioinformatics 17:754 386 755 387 Irish VF (2009) Evolution of petal identity. J Exp Bot 60:2517 2527 388 Jack T, Brockman L, Meyerowitz E (1992) The h omeotic gene APETALA 3 of Arabidopsis thaliana encodes a 389 Mads box and is expressed in petals and stamens. Cell 68:683 697 390 Jackson D (1991) in situ hybridization in plants. Molecular Plant Pathology: A practical approach 1:163 174 391 Jaramillo MA, Kramer EM (20 07) Molecular evolution of the petal and stamen identity genes, APETALA3 and 392 PISTILLATA after petal loss in the Piperales. Mol Phylogenet Evol 44:598 609 393 Jaramillo M, Kramer E (2004) APETALA3 and PISTILLATA homologs exhibit novel expression patterns in th e 394 unique perianth of Aristolochia (Aristolochiaceae). Evol Dev 6:449 458 395 Kang H, Jeon J, Lee S, An G (1998) Identification of class B and class C floral organ identity genes from rice plants. 396 Plant Mol Biol 38:1021 1029 397 Kanno A, Saeki H, Kameya T, Saedler H, Theissen G (2003) Heterotopic expression of class B floral homeotic 398 genes supports a modified ABC model for tulip ( Tulipa gesneriana ). Plant Mol Biol 52:831 841 399 Kay Q, Daoud H, Stirton C (1981) Pigment distribution, light reflection and cell structure i n petals. Bot J Linn Soc 400 83:57 83 401 Kim S, Yun P, Fukuda T, Ochiai T, Yokoyama J, Kameya T, Kanno A (2007) Expression of a DEFICIENS like 402 gene correlates with the differentiation between sepal and petal in the orchid, Habenaria radiata 403 (Orchidaceae). Plant S cience 172:319 326 404 Kramer EM (2007) Understanding the genetic basis of floral diversity. Bioscience 57:479 487 405 Kramer EM, Holappa L, Gould B, Jaramillo MA, Setnikov D, Santiago PM (2007) Elaboration of B gene function 406 to include the identity of novel flora l organs in the lower eudicot Aquilegia Plant Cell 19:750 766 407 Kramer E, Di Stilio V, Schluter P (2003) Complex patterns of gene duplication in the APETALA3 and PISTILLATA 408 lineages of the Ranunculaceae. Int J Plant Sci 164:1 11 409 Krizek B, Meyerowitz E (1996 ) The Arabidopsis homeotic genes APETALA3 and PISTILLATA are sufficient to 410 provide the B class organ identity function. Development 122:11 22 411


Landis et al 16 Litt A, Kramer EM (2010) The ABC model and the diversification of floral organ identity. Semin Cell Dev Biol 412 21:12 9 137 413 Liu Y, Nakayama N, Schiff M, Litt A, Irish V, Dinesh Kumar S (2004) Virus induced gene silencing of a 414 DEFICIENS ortholog in Nicotiana benthamiana Plant Mol Biol 54:701 711 415 Maddison D, Maddison W (2005) MacClade 4: analysis of phylogeny and character evolution. 4.08 416 Maturen N (2008) Genetic analysis of the evolution of petaloid bracts in dogwoods. Dissertation, University of 417 Michigan 418 Nakamura T, Fukuda T, Nakano M, Hasebe M, Kameya T, Kanno A (2005) The modified ABC model explains the 419 development of t he petaloid perianth of Agapanthus praecox ssp orientalis (Agapanthaceae) flowers. Plant 420 Mol Biol 58:435 445 421 Noda K, Glover B, Linstead P, Martin C (1994) Flower color intensity depends on specialized cell shape controlled 422 by a Myb related transcription fa ctor. Nature 369:661 664 423 Park J, Ishikawa Y, Ochiai T, Kanno A, Kameya T (2004) Two GLOBOSA like genes are expressed in second and 424 third whorls of homochlamydeous flowers in Asparagus officinalis L. Plant and Cell Physiology 45:325 425 332 426 Park J, Ishikawa Y, Yoshida R, Kanno A, Kameya T (2003) Expression of AoDEF a B functional MADS box gene, 427 in stamens and inner tepals of the dioecious species Asparagus officinalis L. Plant Mol Biol 51:867 875 428 Perez Rodriguez M, Jaffe F, Butelli E, Glover B, Martin C (2005) Development of three different cell types is 429 associated with the activity of a specific MYB transcription factor in the ventral petal of Antirrhinum majus 430 flowers. Development 132:359 370 431 Prasad K, Vijayraghavan U (2003) Double stranded RNA interference of a rice PI/GLO paralog, OsMADS2 432 uncovers its second whorl specific function in floral organ patterning. Genetics 165:2301 2305 433 Prasad K, Sriram P, Kumar C, Kushalappa K, Vijayraghavan U (2001) Ectopic expression of rice OsMADS 1 reveals 434 a role in specifyin g the lemma and palea, grass floral organs analogous to sepals. Dev Genes Evol 211:281 435 290 436 Preston JC, Kellogg EA (2007) Conservation and divergence of APETALA1/FRUITFULL like gene function in 437 grasses: evidence from gene expression analyses. Plant Journal 52:69 81 438


Landis et al 17 Rasmussen DA, Kramer EM, Zimmer EA (2009) One size fits all? Molecular evidence for a commonly inherited 439 petal identity program in Ranunculales. Am J Bot 96:96 109 440 Rijpkema AS, Royaert S, Zethof J, van der Weerden G, Gerats T, Vandenbussche M (200 6) Analysis of the Petunia 441 TM6 MADS box gene reveals functional divergence within the DEF/AP3 lineage. Plant Cell 18:1819 1832 442 Ronquist F, Huelsenbeck J (2003) MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 443 19:1572 1574 444 Schwa rz Sommer Z, Huijser P, Nacken W, Saedler H, Sommer H (1990) Genetic control of flower development by 445 homeotic genes in Antirrhinum majus Science 250:931 936 446 Sharma B, Guo C, Kong H, Kramer EM (2011) Petal specific subfunctionalization of an APETALA3 para log in the 447 Ranunculales and its implications for petal evolution. New Phytol 191:870 883 448 Soltis PS, Brockington SF, Yoo M, Piedrahita A, Latvis M, Moore MJ, Chanderbali AS, Soltis DE (2009) Floral 449 variation and floral genetics in basal Angiosperms. Am J Bo t 96:110 128 450 Sommer H, Beltran JP, Huijser P, Pape H, Lonnig WE, Saedler H, Schwarz Sommer Z (1990) DEFICIENS a 451 homeotic gene involved in the control of flower morphogenesis in Antirrhinum majus the protein shows 452 homology to transcription factors. EMBO J 9:605 613 453 Sutton DA (1988) A revision of the tribe Antirrhineae. British Museum Natural History; Oxford University Press, 454 London; New York 455 Swofford D (2002) PAUP*. Phylogenetic analysis using parsimony (*and other methods) 4 456 Trobner W, Ramirez L, Motte P Hue I, Huijser P, Lonnig WE, Saedler H, Sommer H, Schwarz Sommer Z (1992) 457 GLOBOSA a homeotic gene which interacts with DEFICIENS in the control of Antirrhinum floral 458 organogenesis. EMBO J 11:4693 4704 459 Tzeng TY, Yang CH (2001) A MADS box gene from lily ( Lilium longiflorum ) is sufficient to generate dominant 460 negative mutation by interacting with PISTILLATA (PI) in Arabidopsis thaliana Plant and Cell Physiology 461 42:1156 1168 462 Vandenbussche M, Zethof J, Royaert S, Weterings K, Gerats T (2004) The duplicated B class heterodimer model: 463 Whorl specific effects and complex genetic interactions in Petunia hybrida flower development. Plant Cell 464 16:741 754 465


Landis et al 18 Vargas P, Rossello JA, Oyama R, Guemes J (2004) Molecular evidence for naturalness of genera in the tribe 466 Antirr hineae (Scrophulariaceae) and three independent evolutionary lineages from the New World and the 467 Old. Plant Syst Evol 249:151 172 468 Weiss D (2000) Regulation of flower pigmentation and growth: Multiple signaling pathways control anthocyanin 469 synthesis in expa nding petals. Physiol Plantarum 110:152 157 470 Whipple CJ, Ciceri P, Padilla CM, Ambrose BA, Bandong SL, Schmidt RJ (2004) Conservation of B class floral 471 homeotic gene function between maize and Arabidopsis Development 131:6083 6091 472 Whitney HM, Chittka L, Br uce TJA, Glover BJ (2009) Conical epidermal cells allow bees to grip flowers and 473 increase foraging efficiency. Current Biology 19:948 953 474 Whitney HM, Glover BJ (2007) Morphology and development of floral features recognised by pollinators. 475 Arthropod Plant Interactions 1:147 158 476 Whittall JB, Voelckel C, Kliebenstein DJ, Hodges SA (2006) Convergence, constraint and the role of gene 477 expression during adaptive radiation: floral anthocyanins in Aquilegia Mol Ecol 15:4645 4657 478 Wu H, Su H, Hu J (2007) The identif ication of A B C and E class MADS box genes and implications for 479 perianth evolution in the basal eudicot Trochodendron aralioides (Trochodendraceae). Int J Plant Sci 480 168:775 799 481 Xiao H, Wang Y, Liu DF, Wang WM, Li XB, Zhao XF, Xu JC, Zhai WX, Zhu LH (2003) Functional analysis of the 482 rice AP3 homologue OsMADS16 by RNA interference. Plant Mol Biol 52:957 966 483 Yang YZ, Fanning L, Jack T (2003) The K domain mediates heterodimerization of the Arabidopsis floral organ 484 identity proteins, APETALA3 and PISTILLA TA Plant Journal 33:47 59 485 Yoo M, Soltis PS, Soltis DE (2010) Expression of floral MADS box genes in two divergent water lilies: 486 Nymphaeales and Nelumbo Int J Plant Sci 171:121 146 487 Zanis MJ, Soltis PS, Qiu YL, Zimmer E, Soltis DE (2003) Phylogenetic analy ses and perianth evolution in basal 488 angiosperms. Ann Mo Bot Gard 90:129 150 489 Zhang W, Xiang Q, Thomas DT, Wiegmann BM, Frohlich MW, Soltis DE (2008) Molecular evolution of 490 PISTILLATA like genes in the dogwood genus Cornus (Cornaceae). Mol Phylogenet Evol 47 :175 195 491 Zwickl D (2006) Genetic algorithm approaches for the phylogenetic analysis of large biological sequence datasets 492 under the maximum likelihood criterion. Dissertation, University of Texas Austin 493