<%BANNER%>

Evolution and Development of Petals in Aizoaceae (Caryophyllales)

Permanent Link: http://ufdc.ufl.edu/UFE0024370/00001

Material Information

Title: Evolution and Development of Petals in Aizoaceae (Caryophyllales)
Physical Description: 1 online resource (213 p.)
Language: english
Creator: Brockington, Samuel
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: aizoaceae, development, evolution, mads, perianth, petal, phylogeny, pigment
Botany -- Dissertations, Academic -- UF
Genre: Botany thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: In this dissertation, the Caryophyllales sensu lato are explored and developed as a system for studying the evolutionary development of the angiosperm perianth. The phylogenetic context of the Caryophyllales is defined with a molecular phylogenetic analysis employing eight plastid genes, two nuclear genes and the entire plastid Inverted Repeat for forty taxa representing the major families within the Caryophyllales sensu lato. Stochastic character mapping and parsimony reconstruction analyses reveal a minimum of nine independent origins of the differentiated perianth within the core Caryophyllales clade. Molluginaceae are revealed to be polyphyletic such that two genera, Hypertelis and Macarthuria, form disparate lineages with respect to Molluginaceae sensu stricto. The implications of this resolved polyphyly for our understanding of the evolution of pigmentation within the core Caryophyllales was explored by stochastic character mapping and parsimony reconstruction analyses. Multiple origins of betalains are suggested, with reversals to anthocyanins. The value of Multiple Displacement Amplification is demonstrated as a means to augment rare archival DNA stocks for molecular phylogenetic analysis. The evolution of various lineages of MADS-box transcription factors was explored in relation to the novel evolution of petaloid staminodes within the plant family Aizoaceae. Fifty-four new MADS-box gene loci were sequenced from six species of Caryophyllales. Phylogenetic analyses of these loci demonstrate an absence of gene duplication, therefore, increases in MADS-box gene complement do not play a role in the evolution of the novel petaloid staminodes in Aizoaceae. Finally the evidence for a shared genetic pathway for petal development was assessed by examining the morphological structure and genetic processes in two distinct petaloid organs in Aizoaceae, the petaloid staminodes and petaloid tepals. It is determined that at multiple levels the petaloid staminodes and petaloid tepals are not-homologous. There is no morphological or genetic evidence to recommend the concept of a conserved petal identity program in Aizoaceae. The petaloid organs in Aizoaceae should therefore be regarded as convergent evolutionary events that have occurred independently of a hypothesized petal identity program. Thus, the concept of a conserved petal identity program across all core eudicot angiosperms is refuted.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Samuel Brockington.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Soltis, Douglas E.
Local: Co-adviser: Soltis, Pamela S.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2011-08-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0024370:00001

Permanent Link: http://ufdc.ufl.edu/UFE0024370/00001

Material Information

Title: Evolution and Development of Petals in Aizoaceae (Caryophyllales)
Physical Description: 1 online resource (213 p.)
Language: english
Creator: Brockington, Samuel
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: aizoaceae, development, evolution, mads, perianth, petal, phylogeny, pigment
Botany -- Dissertations, Academic -- UF
Genre: Botany thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: In this dissertation, the Caryophyllales sensu lato are explored and developed as a system for studying the evolutionary development of the angiosperm perianth. The phylogenetic context of the Caryophyllales is defined with a molecular phylogenetic analysis employing eight plastid genes, two nuclear genes and the entire plastid Inverted Repeat for forty taxa representing the major families within the Caryophyllales sensu lato. Stochastic character mapping and parsimony reconstruction analyses reveal a minimum of nine independent origins of the differentiated perianth within the core Caryophyllales clade. Molluginaceae are revealed to be polyphyletic such that two genera, Hypertelis and Macarthuria, form disparate lineages with respect to Molluginaceae sensu stricto. The implications of this resolved polyphyly for our understanding of the evolution of pigmentation within the core Caryophyllales was explored by stochastic character mapping and parsimony reconstruction analyses. Multiple origins of betalains are suggested, with reversals to anthocyanins. The value of Multiple Displacement Amplification is demonstrated as a means to augment rare archival DNA stocks for molecular phylogenetic analysis. The evolution of various lineages of MADS-box transcription factors was explored in relation to the novel evolution of petaloid staminodes within the plant family Aizoaceae. Fifty-four new MADS-box gene loci were sequenced from six species of Caryophyllales. Phylogenetic analyses of these loci demonstrate an absence of gene duplication, therefore, increases in MADS-box gene complement do not play a role in the evolution of the novel petaloid staminodes in Aizoaceae. Finally the evidence for a shared genetic pathway for petal development was assessed by examining the morphological structure and genetic processes in two distinct petaloid organs in Aizoaceae, the petaloid staminodes and petaloid tepals. It is determined that at multiple levels the petaloid staminodes and petaloid tepals are not-homologous. There is no morphological or genetic evidence to recommend the concept of a conserved petal identity program in Aizoaceae. The petaloid organs in Aizoaceae should therefore be regarded as convergent evolutionary events that have occurred independently of a hypothesized petal identity program. Thus, the concept of a conserved petal identity program across all core eudicot angiosperms is refuted.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Samuel Brockington.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Soltis, Douglas E.
Local: Co-adviser: Soltis, Pamela S.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2011-08-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0024370:00001


This item has the following downloads:


Full Text

PAGE 1

1 EVOLUTION AND DEVELOPMENT OF PETALS IN AIZOACEAE (CARYOPHYLLALES) By SAMUEL FRASER BROCKINGTON A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2009

PAGE 2

2 2009 Samuel Fraser Brockington

PAGE 3

3 To my ma and pa

PAGE 4

4 ACKNOWLEDGMENTS I acknowledge the assistance of numerous people and organizations that have contributed to this doctoral work. First and foremost, I thank my advisors Douglas and Pamela Soltis for their patient support and example. I thank Walter S. Judd for inspiring my interest in the evolution of angiosperms and for his insight into various a spects of this project. I thank David Oppenheimer and Paris Gray for their expertise and use of their in -situ facilities. I thank Elena M. Kramer for her advice and expertise at various stages of this project. The morphological development work would not have been possible without the help of Paula J Rudall at the Royal Botanic Gardens, Kew and Michael Frohlich at the Natural History Museum, London. I thank my supervisory committee for taking the time to read this dissertation. I thank Mark Whitten for taking so many superb photos and sharing his photographic equipment. Kurt Neubig kindly shared equipment. I thank my two undergraduate helpers Alexandre Roolse and Jeremy Ramdial for putting in some long hours and collecting substantial amounts of data. The re search was funded by an Assembling the Tree of Life grant EF 0431266 (NSF) to D. E. Soltis and P. S. Soltis, Floral Genome Project NSF grant PGR 0115684 to D. E. Soltis and P. S. Soltis, and a DDIG grant DEB 0808342 (NSF) to D. E. Soltis and S. F. Brocking ton, and a Botanical Society of America Karling Award to S. F. Brockington. Finally I thank my friends in Florida and England, and my siblings Dan, Alice and Grace for their endless support and encouragement. Above all I thank my parents for their love an d dedication over so many years.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................................... 4 LIST OF TABLES ................................................................................................................................ 9 LIST OF FIGURES ............................................................................................................................ 11 LIST OF ABBREVIATIONS ............................................................................................................ 13 ABSTRACT ........................................................................................................................................ 14 CHAPTER 1 INTRODUCTION ....................................................................................................................... 16 The Angiosperm Perianth as a System for Evolutionary Developmental Biology (Evo Devo) ....................................................................................................................... 16 An Evo Devo Framework for the Study of Petal Evolution ............................................. 17 Caryophyllales: A Model Clade for the Study of Perianth Evolution .............................. 18 Dissertation Outline: Developing the Caryophyllales as a System for EvoDevo Research ............................................................................................................................ 18 2 PHYLOGENY OF THE CARYOPHYLLALES SENSU LATO : REVISITING HYPOTHESES ON POLLINATION BIOLOGY AND PERIANTH DIFFERENTIATION IN T HE CORE CARYOPHYLLALES ............................................... 20 Introduction ................................................................................................................................. 21 Materials and Methods ................................................................................................................ 26 Taxon Sampling ................................................................................................................... 26 DNA Isolation and Amplification ...................................................................................... 27 Alignment and Phylogenetic Analysis ............................................................................... 28 Character Reconstructions .................................................................................................. 30 Results .......................................................................................................................................... 34 Individual Plastid Single Copy Data Sets .......................................................................... 34 Inverted Repeat Data Set ..................................................................................................... 35 Combined Plastid Genes from the Single Copy Region ................................................... 35 Combined Plastid Single Copy and Nuclear Genes .......................................................... 36 Total Evidence Data Set ...................................................................................................... 36 Discussion .................................................................................................................................... 37 Phylogenetic Analyses ........................................................................................................ 38 Reconstruction of Pollination Mechanism ......................................................................... 42 Reconstruction of Perianth D ifferentiation ........................................................................ 44 Multiple Origins of Perianth Differentiation ..................................................................... 46 Recruitment of Preceding Bracts ........................................................................................ 47 Petaloid Modification of the Androecium .......................................................................... 50

PAGE 6

6 Differentiation of Homologous Perianth Parts .................................................................. 51 Caryophyllales as a System for Floral Evo Devo .............................................................. 52 3 LABILE EVOLUTION OF PIGMENTATION IN THE CORE CARYOPHYLLALES ..... 85 Introduction ................................................................................................................................. 86 The Unique Origin of Betalains .......................................................................................... 87 The Intercalation of Betalain and Anthocyanin Lineages ................................................. 88 Mutual Exclusion of Anthocyanins and Betalains ............................................................. 89 Biosynthetic Inhibition ........................................................................................................ 89 Materials And Method s .............................................................................................................. 91 Taxon Sampling and Data Assembly ................................................................................. 91 Phylogenetic Analyses ........................................................................................................ 91 Character Coding ................................................................................................................. 93 Character -state Reconstruction ........................................................................................... 93 Stochastic Mapping ............................................................................................................. 94 Prior Specification ............................................................................................................... 94 Rates of Transformation and Ancestral State Reconstruction .......................................... 95 Results .......................................................................................................................................... 96 Phylogenetic Relationships ................................................................................................. 96 Character Mapping with Unordered Coding ...................................................................... 97 Character Mapping with Ordered Coding .......................................................................... 98 Discussion .................................................................................................................................... 99 Polyphyletic Molluginaceae ................................................................................................ 99 Early Diverging Anthocyanic Lineages ........................................................................... 100 Derived Anthocyanic Lineages ......................................................................................... 101 Evidence for Reversals to Anthocyanin Pigm entation .................................................... 102 Effect of Ordered Character Coding Support for Ehrendorfer (1976. ......................... 103 The Occurrence of Unpigmented Lineages ...................................................................... 104 Multiple Origins of Betalain Pigmentation ...................................................................... 105 Conclusion ................................................................................................................................. 107 4 K EEP THE DNA ROLLING: M ULTIPLE DISPLACEMENT AMPLIFICATION OF ARCHIVAL PLANT DNA EXTRACTS ............................................................................... 122 Introduction ............................................................................................................................... 122 Materials and Metho ds .............................................................................................................. 124 DNA Extractions from Herbarium Specimens ................................................................ 124 Multiple Displacement Amplification .............................................................................. 124 PCR ..................................................................................................................................... 125 PCR Cleanup and Sequencing .......................................................................................... 126 Results ........................................................................................................................................ 126 Multiple Displacement Amplification .............................................................................. 126 PCR ..................................................................................................................................... 127 Sequencing ......................................................................................................................... 128 Discussion .................................................................................................................................. 128

PAGE 7

7 5 PHYLOGENY OF MADS -BOX GENE LINEAGES IN AIZOACEAE (CARYOPHYLLALES) ........................................................................................................... 133 Introduction ............................................................................................................................... 133 MIKC Genes and ABCDE Functions in Floral Organ Identity ...................................... 134 Phylogenetic Analyses Reveal the Evolutionary History of MIKC type Genes ........... 135 Phylogeny of A -Function MIKC type Genes .................................................................. 136 B-Function MIKC -type Genes .......................................................................................... 137 C and D -Function MIKC type Genes ............................................................................... 137 E -function MIKC type Genes ........................................................................................... 138 Duplications in Association with Morphological Var iation and Novelty ...................... 139 Methods ..................................................................................................................................... 141 Taxon Sampling ................................................................................................................. 141 RNA Isolat ion and Gene Amplification ........................................................................... 142 Alignments ......................................................................................................................... 142 Phylogenetic Analyses ...................................................................................................... 143 Results and Discussion ............................................................................................................. 144 Caryophyllid MIKC Complement .................................................................................... 144 Polygonales -specific Duplication in the Eu AP3 Lineage ............................................... 146 Absence of the eu PLE Lineage in Caryophyllales .......................................................... 147 Absence of Duplications Coinciding with the Inferred Origin of Petaloid Stamino des ..................................................................................................................... 148 6 PERIANTH EVOLUTION IN AIZOACEAE: NOT ALL CORE EUDICOT PETALS WERE CREATED EQUAL ..................................................................................................... 162 Introduction ............................................................................................................................... 163 Methods ..................................................................................................................................... 169 Plant Materials ................................................................................................................... 169 Scanning Electron Microscopy ......................................................................................... 169 RNA -RNA In -situ Hybridisation ...................................................................................... 169 Relative Quantitative RT PCR ......................................................................................... 170 Results ........................................................................................................................................ 171 Floral Ontogeny in Sesuvium portulacastrum ................................................................. 171 Floral Ontogeny in Delosperma napiforme ..................................................................... 172 Expression of AGAMOUS Homologs .............................................................................. 173 Expression of PISTILLATA Homologs ............................................................................ 174 Expression of APETALA3 Homologs ............................................................................... 175 Discussion .................................................................................................................................. 176 Perianth Ontogeny and Historical Homology of Petaloid Organs in Aizoaceae ........... 176 Petal Identity: A Morphological or Functional Concept? ............................................... 177 Implications of Correspondence between Leaf Sheath and Petaloid Tepal ................... 178 Assessing Morphological Homology between Different Petaloid Organs in Aizoaceae ....................................................................................................................... 180 Assessing Process Homology between Petaloid Staminodes and Petaloid Tepals ....... 181 Implications of AP3 and PI Homolog Expression Patterns ............................................ 182 Implications of AG Homolog Expression Patterns .......................................................... 184

PAGE 8

8 Evolution of the Aizoaeae Perianth in the Context of the Caryophyllales .................... 186 Conclusion.......................................................................................................................... 188 REFERENCES ................................................................................................................................. 197 BIOGRAPHICAL SKETCH ........................................................................................................... 213

PAGE 9

9 LIST OF TABLES Table page 2 1 Table showing for each data partition: aligned and analysed length of data partitions ..... 70 2 2 Voucher, Citation, and GenBank Accession No. for rbcL .................................................. 71 2 3 Voucher, Citation, and GenBank Accession No. for rpoC2 ............................................... 72 2 4 Voucher, Citation, and GenBank Accession No. for atpB .................................................. 73 2 5 Voucher, Citation, and GenBank Accession No. for psbBTN ............................................. 74 2 6 Voucher, Citation, and GenBank Accession No. for rps4 ................................................... 75 2 8 Voucher, Citation, and GenBank Accession No. for 18S rDNA ........................................ 77 2 9 Voucher, Citation, and GenBank Accession No. for 26S rDNA ........................................ 78 2 10 Voucher, Citation, and GenBank Accession No. for IR ...................................................... 79 2 11 Primers used for PCR and Sequencing ................................................................................. 80 2 1 2 Coding of Perianth for Parsimony Reconstruction and Stochastic Character Mapping, with references used to code taxa .......................................................................................... 81 2 13 Coding of Pollination for Parsimony Reconstruction and Stochastic Character Mapping, with references used to code taxa. ....................................................................... 82 2 14 Estimations of the posterior probability of ancestral perianth character states at each node ......................................................................................................................................... 83 2 15 Estimations of the posterior probability of ancestral pollination character states at each node. Coding Strategy for Pollination: Anemophilous = 0, Entomophilous = 1 ....... 84 3 1 Summary of hypotheses historically proposed to explain the evolution of pigmentation in the Caryophyllales, with limitations of each. .......................................... 115 3 2 Frequencies of each transformation for four alternative reconstructions employing unordered character coding. ................................................................................................ 116 3 3 Frequencies of each transformation for four alternative reconstructions employing ordered character coding ...................................................................................................... 117 3 4 Voucher, citation, and GenBank accession No. for rbcL (Voucher information not shown for previously published sequences). ...................................................................... 118

PAGE 10

10 3 5 Voucher, citation, and GenBank accession No. for matK (voucher information not shown for previously published sequences) ....................................................................... 119 3 6 The estimations of the posterior probability of ancest ral characters for the ordered and unordered evolution of pigments for all nodes presented graphically in Fig. 3. ....... 120 3 7 Pigment coding for each terminal and sources of information used to de signate characters states. Coding Strategy: Anthocyanin (0), Unpigmented (1), Betalain (2), Unknown (?). ........................................................................................................................ 121 4 1 Species of Caryophyllales used in this study: family, source, collection date, accession number and DNA extraction method. ................................................................ 132

PAGE 11

11 LIST OF FIGURES Figure page 2 1 Maximum likelihood (ML) tree resulting from GARLI analysis of total evidence data set ............................................................................................................................................ 55 2 2 Phylogram of single most parsimonious tree based on the total evidence data set ............ 56 2 3 Parsimony reconstruction (illustrated on an MP tree) and stochastic character mapping (illustrated on Bayesian consensus tree): A) Reconstruction of perianth evolution; B) Pollination syndromes. ................................................................................... 57 2 4 The d iverse forms of the perianth in the core Caryophyllales ............................................. 58 2 5 Trees derived from analyses of the atpB data set ................................................................. 59 2 6 Trees der ived from analyses of the matK data set ................................................................ 60 2 7 Trees derived from analyses of the ndhF data set ................................................................ 61 2 8 Trees derived from analys es of the psbBTN data set ........................................................... 62 2 9 Trees derived from analyses of the rbcL data set ................................................................. 63 2 10 Trees derived from analyses of the rpo C2 data set .............................................................. 64 2 11 Trees derived from analyses of the rps4 data set ................................................................. 65 2 12 Trees derived from analyses of the IR data set .................................................................... 66 2 13 Trees derived from analyses of the combined plastid single -copy gene data set ............... 67 2 14 Trees derived from analyses of the com bined single copy plastid gene and nuclear gene data sets. ......................................................................................................................... 68 2 15 Posterior probabilities of each rate category given each combination of E(T) and SD(T) and sampled from the prior with 10,000 realizations ............................................... 69 3 1 Phylogram of one of twelve most parsimonious trees based on the matK and rbcL data set for 11 genera of Molluginaceae ............................................................................. 109 3 2 Maximum likelihood (ML) tree resulting from GARLI analysis of matK and rbcL data set for 11 genera of Molluginaceae ............................................................................. 110 3 3 Reconstruction of pigment evolution. ................................................................................. 111 3 4 Prior simulations for ordered data set. ................................................................................ 112

PAGE 12

12 3 5 Prior simulations for unordered data set. ............................................................................ 113 3 6 Parsimony reconstruction of pigment evolution using unordered data sets and ordered data sets ................................................................................................................... 114 4 1 MDA treated reaction products (M1 8) and corresponding untre ated genomic equivalent (G1 8) ................................................................................................................. 131 4 2 Amplification of 1363 bp fragment of psbB fragment of nad7 gene, fragment of ITS for two of the eight species .................................................................................................. 131 5 1 ML tree derived from analysis of a DNA dataset comprising 121 MADS -box gene loci (including 54 loci isolated by this study) .................................................................... 149 5 2 Trees derived from analyses of the A -function DNA dataset comprising 82 MADS box gene loci (including 12 loci isolated by the study) ..................................................... 150 5 3 Trees derived from analyses of the B -function DNA dataset comprising 127 MADS -box gene loci (including 19 loci isolated by the study) ........................................ 152 5 4 Trees derived from analyses of the APETALA3 DNA dataset comprising 69 MADS box gene loci (including ten loci isolated by the st udy) .................................................... 154 5 5 Trees derived from analyses of the PISTILLATA DNA dataset comprising 47 MADS -box gene loci (including 6 loci isolated by the study) .......................................... 156 5 6 Trees derived from analyses of the C and D -function DNA dataset comprising 104 MADS -box gene loci (including nine loci isolated by the study) ..................................... 158 5 7 Trees derived f rom analyses of the E -function DNA dataset comprising 151 MADS -box gene loci (including eight loci isolated by the study) .................................... 160 6 1 Floral Development in Sesuvium portulacastrum .............................................................. 190 6 2 Floral development in Delosperma napiforme ................................................................... 191 6 3 Correspondence between perianth and leaf sheath in Sesuvium portulacastrum ............. 192 6 4 Expression of AGAMOUS ( AG ) homologs in Sesuvium portulacastrum ......................... 193 6 5 Expression of PISTILLATA ( PI ) homologs in Sesuvium portul acastrum ........................ 194 6 6 Expression of APETALA3 (AP3) homologs in Sesuvium portulacastrum ....................... 195 6 6 RQ -RT -PCR data for four species of Aiz oaceae and three gene homologs ( AG, AP3, and PI ) ................................................................................................................................... 196

PAGE 13

13 LIST OF ABBREVIATIONS bp Base Pair ML Maximum Likelihood MP Maximum Parsimony l micro litre

PAGE 14

14 Abstract of Dissertation Presented to the Graduate School of the Univer sity of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy EVOLUTION AND DEVELOPMENT OF PETALS IN AIZOACEAE (CARYOPHYLLALES) By SAMUEL FRASER BROCKINGTON August 2009 Chair: Douglas E. Soltis Cochair: Pamela S. Soltis Major: Botany In this dissertation, the Caryophyllales sensu lato are explored and developed as a system for studying the evolutionary development of the angiosperm perianth. The phylogenetic context of the Caryophyllales is defined with a molecular phylogenetic analysis employing eight plastid genes, two nuclear genes and the entire plastid Inverted Repeat for forty taxa representing the major families within the Caryophyllales sensu lato. Stochastic character mapping and parsimony reconstruction analyses reveal a minimum of nine independent origins of the differentiated perianth within the core Caryophyllales clade. Molluginaceae are revealed to be polyphyletic such that two genera, Hypertelis and Macarthuria, form disparate lineages wit h respect to Molluginaceae sensu stricto The implications of this resolved polyphyly for our understanding of the evolution of pigmentation within the core Caryophyllales was explored by stochastic character mapping and parsimony reconstruction analyses. Multiple origins of betalains are suggested, with reversals to anthocyanins. The value of Multiple Displacement Amplification is demonstrated as a means to augment rare archival DNA stocks for molecular phylogenetic analysis. The evolution of various linea ges of MADS -box transcription factors was explored in relation to the novel evolution of petaloid staminodes within the plant family Aizoaceae. Fifty -four new MADS -box gene loci were sequenced from six species of Caryophyllales. Phylogenetic analyses of these loci demonstrate an absence of gene duplication,

PAGE 15

15 therefore, increases in MADS -box gene complement do not play a role in the evolution of the novel petaloid staminodes in Aizoaceae. Finally the evidence for a shared genetic pathway for petal development was assessed by examining the morphological structure and genetic processes in two distinct petaloid organs in Aizoaceae, the petaloid staminodes and petaloid tepals. It is determined that at multiple levels the petaloid staminodes and petaloid tepals are not homologous. There is no morphological or genetic evidence to recommend the concept of a conserved petal identity program in Aizoaceae. The petaloid organs in Aizoaceae should therefore be regarded as convergent evolutionary events that have occurred independently of a hypothesized petal identity program. Thus, the concept of a conserved petal identity program across all core eudicot angiosperms is refuted.

PAGE 16

16 CHAPTER 1 INTRODUCTION The Angiosperm Perianth as a System for Evolutionary Developmental Biology (Evo -Devo) The perianth is a near -constant characteristic of angiosperms, giving many flowers their distinctive appearance and playing an important role in plant -pollinator interactions (Baum and Whitlock 1999). The importance of the perianth for angiosp erm biology has ensured a long and rich history of research. More recently, research into perianth evolution has been invigorated by the discipline of evolutionary developmental biology, which seeks to integrate new data from molecular phylogenetic analyse s, comparative developmental morphology, and comparative analyses of gene expression and function. In the context of the perianth, the evolution the petal in particular has become the focus of evolutionary developmental study. Several factors motivate e volutionary developmental research into the petal: 1) The genetic pathways underlying petal development in model organisms are relatively well understood (Coen and Meyerowitz, 1991). 2) Homologs of the MADS -box genes that are critical in petal developmenta l pathways have been isolated across the angiosperms (Kramer and Irish, 2000; Kim et al., 2005). 3) MADS -box gene homologs function in a spatially dependent fashion and so are amenable to comparative expression techniques (Kramer and Jaramillo, 2007). 4) D iverse perianth structures provide the necessary variation to explore how the evolution of developmental genetic pathways relates to the evolution of morphology and allow various principles of evolutionary developmental biology to be tested (Endress, 1996) In the context of petal evolution, these principles include: the role of heterotopy in morphological evolution; the developmental basis of homoplasy and homology; the relative frequency of parallel and convergent evolution; and the role of gene duplicati on in the origin of morphological novelty.

PAGE 17

17 An Evo -Devo Framework for the Study of Petal Evolution Classical theory of petal evolution suggests that petals have evolved multiple times independently (Takhtajan, 1991). This theory arose from the assumption th at the ancestor to the angiosperms was apetalous (Sun et al. 1998), and that petals in different lineages of angiosperms exhibit distinct suites of morphological traits reminiscent of either bracts or stamens. Bracteopetals describe petals that generally e xhibit leaf like characteristics: they initiate and develop earlier than the stamens; they are often spirally arranged on the same parastichies as the subtending bracts; their primordia are distinctly crescent -shaped; and they are supplied by three vascul ar traces. In contrast, andropetals exhibit stamen -like characteristics: they are developmentally delayed relative to the stamens; arranged on the same parastichies as the stamens; similar in appearance to stamen primordia at inception; and are supplied by a single vascular trace (Endress 1994). In an effort to synthesize these classical observations with contemporary developmental genetic data, Kramer and Irish (1999) proposed the concept of an ancestral petal identity program. Their hypothesis rests on the observation that petal development pathways are conserved between distant core eudicot lineages, and that the MADS -box gene components within these petal developmental programs are frequently associated with petals and petaloid organs across angiosperm s. Kramer and Irish hypothesized that de activation and subsequent repeated re activation of an ancestral genetic petal identity program, together with spatial redeployment in either bracts or stamens, generated petaloid organs of differing historical and positional homology in different lineages of angiosperms. Consequently, angiosperm petals, although appearing to have evolved independently many times, may actually share latent genetic homology in the form of a common petal identity pathway. This hypothes is is considered to be

PAGE 18

18 testable through the comparative analysis of the developmental genetic pathways that specify petal identity in bracteopetals and andropetals in distinct lineages of angiosperms. Caryophyllales: A Model Clade for the Study of Perianth Evolution The Caryophyllales is a group long recognized by its distinctive placentation and embryology, and is a major order of angiosperms representing about 5% of core eudicot diversity. The order exhibits diverse variation in perianth structure and mo rphology and is consequently a valuable clade in which to assess the concepts of a petal identity program. In essence, the patterns of perianth evolution in Caryophyllales imitate, on a smaller phylogenetic scale, the evolutionary trends traditionally thou ght to have taken place across the angiosperms as a whole ( Bessey, 1915; Takhtajan, 1991). The perianth varies from an undifferentiated to differentiated structure with the concomitant evolution of petals from either bracts or stamens, in varying positions in the flower. More recently, molecular phylogenetics has improved our understanding of intra -ordinal relationships within the Caryophyllales (Cuenoud et al., 2002), allowing more accurate interpretations of polarity in perianth evolution and enabling cle arer understanding of homology between different perianth structures. This improved phylogenetic resolution, combined with novel evolutionary developmental approaches has the potential to shed much light on the evolution of floral morphology in the Caryophyllales. In the following chapters, I take the first steps in developing the Caryophyllales as a system for the study of the evolutionary development of the angiosperm perianth, with special reference to the evolution of petals. Dissertation Outline: Deve loping the Caryophyllales as a System for Evo-Devo Research Chapter 2 defines the phylogenetic concept of Caryophyllales with a molecular phylogenetic analysis employing eight plastid genes, two nuclear genes and the entire plastid Inverted Repeat (~42,000bp) for 40 taxa representing the major families within the

PAGE 19

19 Caryophyllales sensu lato. Stochastic character mapping and parsimony reconstruction analyses reveal a minimum of nine independent origins of the differentiated perianth within the core Caryophylla les clade. Chapter 3 addresses the problem of the anthocyanic yet polyphyletic Molluginaceae, revealing that two genera, Hypertelis and Macarthuria, form disparate lineages with respect to Molluginaceae sensu stricto The implications of this resolved polyphyly for our understanding of the evolution of pigmentation within the core Caryophyllales is explored by stochastic character mapping and parsimony reconstruction analyses. Chapter 4 illustrates the value of Multiple Displacement Amplification as a means to augment rare archival DNA stocks for molecular phylogenetic analysis a strategy employed in Chapters 2 and 3. Chapter 5 describes the isolation of 54 MADS box gene loci representing 10 clades of MADS -box genes, from all four subfamilies within the fa mily Aizoaceae, as well representatives from Portulacaceae and Polygonaceae. Phylogenetic analyses using Parsimony, Likelihood and Bayesian criteria are employed to determine orthology and paralogy in these MADS box gene lineages and to identify putative g ene losses and duplication. Chapter 6 tests the concept of the petal identity program in the context of the family Aizoaceae (Caryophyllales), combining developmental morphological data and gene expression analyses of MADS -box gene homologs.

PAGE 20

20 CH APTER 2 PHYLOGENY OF THE CAR YOPHYLLALES SENSU LATO : REVISITING HYPOTHESES ON POLLINATION BIOLOGY AND PERIANTH DIFF ERENTIATION IN THE C ORE CARYOPHYLLALES Molecular phylogenetics has revolutionized our understanding of the Caryophyllales, and yet many relat ionships have remained uncertain, particularly at deeper levels. We have performed parsimony and maximum likelihood analyses on separate and combined data sets comprising eight plastid genes (~12,000bp), two nuclear genes (~5000bp) and the plastid inverted repeat (~24,000bp) giving a combined analyzed length of 42006bp for 36 species of Caryophyllales and four outgroups. We have recovered strong support for deep-level relationships across the order. Two major subclades are well -supported the non-core and core Caryophyllales; Rhabdodendron followed by Simmondsia are sisters to the core Caryophyllales; Limeum and Stegnosperma are successive sisters to the globular inclusion clade; Gisekia is a distinct lineage well separated from Rivina within the raphid e clade; and Rivina and Phytolaccaceae are disparate lineages with Rivina sister to Nyctaginaceae. The placement of Sarcobatus and relationships within the portulacaceous cohort remain problematic. Within the latter, Halophytum is sister to Basellaceae an d Didiereaceae, and the clade comprising Portulaca Talinum and Cactaceae is well supported. Classical hypotheses argued that the early Caryophyllales had evolved in open, dry, marginal environments at a time when pollinators were scarce and as such the an cestral caryophyllid flower was wind -pollinated with an undifferentiated perianth. We re -evaluated these hypotheses in light of our phylogeny and find little support for anemophily as the ancestral condition, however the early caryophyllid flower is sugges ted to have possessed an undifferentiated perianth. There has been a subsequent minimum of nine origins of differentiated perianth. We discuss the evidence for independent origins of

PAGE 21

21 differentiated perianth and highlight the research opportunities that thi s pattern offers to the field of evolutionary developmental genetics. Introduction Research interest in Caryophyllales has a long and rich history; core members of this lineage correspond to the old Centrospermae (central seeded), a group long recognized by its distinctive placentation and embryology (Braun, 1864; Eichler, 187578). Centrospermae became the focus of research and debate in the 1960s as one of the first groups whose circumscription was modified based on phytochemistry (Cronquist and Thorne, 1994). All but two of the ten families then recognized as belonging to the Centrospermae were discovered to possess betalain pigments instead of anthocyanins (Cronquist and Thorne, 1994). On the basis of this chemosystematic character, Cactaceae and Didie reaceae were reassigned to the Centrospermae, and several families of dubious affiliation were excluded (Cronquist and Thorne, 1994). Subsequent classifications recognized the Caryophyllales as a well -defined group on the basis of numerous morphological, ultrastructural and chemical characters (Cronquist, 1981, 1988; Dahlgren, 1975; Tahktajan, 1980; Thorne, 1976). Just prior to the emergence of DNA based molecular systematics, the Caryophyllales sensu stricto comprised 12 families (Cronquist, 1988; Tahktaja n, 1980; Thorne, 1992a): Phytolaccaceae, Achatocarpaceae, Nyctaginaceae, Aizoaceae, Didiereaceae, Cactaceae, Chenopodiaceae, Amaranthaceae, Portulacaceae, Basellaceae, Molluginaceae, and Caryophyllaceae. In addition, Polygonaceae and Plumbaginaceae have be en regarded as closely related to these 12 families by numerous systematists (Cronquist and Thorne, 1994). A series of molecular phylogenetic investigations has altered the concept of Caryophyllales provided in earlier classifications. Giannasi et al. (19 92) confirmed the close relationship of Polygonaceae and Plumbaginaceae with the Caryophyllales. Albert et al. (1992) and Williams et

PAGE 22

22 al. (1994) demonstrated the association of Droseraceae, Nepenthaceae and Drosophyllaceae with Polygonaceae and Plumbaginac eae. The carnivorous clade plus Polygonaceae/Plumbaginaceae was further expanded to include Ancistrocladaceae, Dioncophyllaceae, Frankeniaceae and Tamaricaceae (Fay et al., 1997). Numerous analyses have recognized this expanded clade, which is variously te rmed the non-core Caryophyllales (APGII, 2003; Cunoud et al., 2002), Caryophyllales II (Hilu et al., 2003), and Polygonales (Judd et al., 1999), as sister to Caryophyllales sensu stricto (Soltis et al., 2000; Soltis et al., 1999). Within Caryophyllale s s.s. or the core Caryophyllales, molecular data have resulted in several refinements in phylogeny and classification. Additional families recognized as belonging to the core Caryophyllales include Physenaceae and Asteropeiaceae (Morton et al., 1997), Rhabdodendraceae and Simmondsiaceae (Fay et al., 1997). Simmondsiaceae are supported as sister to the core Caryophyllales (Cunoud et al., 2002) with Physenaceae and Asteropeiaceae forming a strongly supported sister group (combined rbcL / matK analysis: Cun oud et al., 2002). Rhabdodendraceae have been associated with Simmondsiaceae as sister to the core Caryophyllales or as sister to both core and non -core Caryophyllales, but with little support for either position ( matK analysis: Cunoud et al., 2002). Mole cular studies have also identified and confirmed a number of polyphyletic groups within the core Caryophyllales. Recognition that Phytolaccaceae are polyphyletic (initially by Rettig et al.,1992) supports the delimitation of four families: Achatocarpaceae, Barbeuiaceae, Gisekiaceae and Stegnospermataceae (APG II, 2003; Cunoud et al., 2002). Achatocarpaceae may form a clade together with Caryophyllaceae and Amaranthaceae (combined matk/rbcL analysis: Cunoud et al., 2002). Stegnospermataceae are placed with out support as a successive sister lineage to the remainder of the core Caryophyllales (Cunoud et al., 2002), following the

PAGE 23

23 divergence of Asteropeiaceae, Physenaceae, Achatocarpaceae, Caryophyllaceae and Amaranthaceae. Barbeuia represents a distinct, isol ated lineage within the core Caryophyllales but of uncertain position (Cunoud et al., 2002). Rivina and Petiveria formerly both of Phytolaccaceae, are paraphyletic (combined matK/rbcL analysis: Cunoud et al., 2002) with respect to Phytolacca Rivina ha s been allied with the family Gisekiaceae, and Petiveria is placed sister to Rivina and Gisekiaceae (combined matK/rbcL analysis: Cunoud et al., 2002). Similarly in the matK analyses Hilleria is placed sister to Rivina and Ledenbergia is sister to these two taxa. Lophiocarpus originally placed within Phytolaccaceae, is now separated and placed sister to Corbichonia (Cunoud et al., 2002). Sarcobatus originally placed in the Chenopodiaceae, was recognized as the family Sarcobataceae (Behnke, 1997; APG II 2003) on the basis of distinct sieve -element plastids with respect to Chenopodiaceae (Behnke, 1997). This separation was supported by molecular analyses in which Sarcobataceae form a distinct lineage allied with the clade containing Aizoaceae, Phytolacca ceae and Nyctaginaceae, Gisekiaceae and Agdestis (Cunoud et al., 2002; Downie et al., 1997). The circumscription of Molluginaceae remains problematic; the family is likely to be polyphyletic. Previous authors have suggested that the inclusion of Macarthur ia and Polpoda is unlikely based on morphological observations, however, two genera have not been included in previous molecular analyses (Cunoud et al., 2002). Genera previously included within Molluginaceae ( Corbichonia, Limeum Gisekia ) form disparate lineages with respect to the type genus. However, the position of Limeum outside Molluginaceae is unsupported ( rbcL / matK : Cunoud et al., 2002). Mollugo, Adenogramma, Glischrothamnus, Glinus, Pharnaceum and Suessenguthiella constitute a monophyletic group that is sister to the portulacaceous cohort (Cunoud et al., 2002).

PAGE 24

24 The portulacaceous cohort of Basellaceae, Cactaceae, Didiereaceae and Portulacaceae was initially proposed by Thorne (1976) and is supported by nonDNA characters such as presence of a f loral involucre, succulent tissue, mucilage and CAM photosynthesis (Cunoud et al., 2002; Nyffeler, 2007). The monophyly of the cohort was implied by early molecular analyses (Downie et al., 1997; Rettig et al., 1992); however, relationships within the gro up are unclear, and complicated by the gross paraphyly of Portulacaceae (suggested by Carolin, 1987 and Hershkovitz, 1993). The addition of molecular data has resulted in some clarification. In an analysis of ITS sequences, Hershkovitz and Zimmer (1997) su ggested that Cactaceae were embedded within Portulacaceae and sister to Portulaca, Anacampseros and relatives, and portions of Talinum (the ACPT clade; from Anacampseroteae, Cactaceae, Portulaca and Talinum : Nyffeler, 2007). These findings were confirmed a nd extended by Applequist and Wallace (2001) in an analysis of ndhF sequences; Talinum with Talinella Portulaca with Anacampseros and the Cactaceae form three distinct lineages within a well -supported clade. Although the monophyly of the ACPT clade seems clear, as summarized by Nyffeler (2007), the pattern of branching within the clade has varied among analyses. Outside of the ACPT clade, Hershkovitz and Zimmer (1997) found that Basellaceae and Didiereaceae form a distinct monophyletic group and are siste r to portulacaceous genera Portulacaria and Ceraria In addition, Applequist and Wallace (2001) described this same clade as consisting of three distinct lineages: Basellaceae; Didiereaceae with Calyptrotheca, Ceraria and Portulacaria; and a strongly suppo rted assemblage of genera including Claytonia Montia Calandrinia, Montiopsis Cistanthe Calyptridium and Phemeranthus Despite these advances in the understanding of subordinal relationships in Caryophyllales, a number of uncertainties remain. The posit ions of Limeum Stegnosperma and Barbeuia are all

PAGE 25

25 ambiguous. The relative position of Rhabdodendron and Simmondsia as possible sisters to either the core Caryophyllales or Caryophyllales s.l. is unclear. Relationships within the portulacaceous cohort are l argely unresolved, particularly within the clade containing Basellaceae, Didiereaceae and allied genera (Applequist and Wallace, 2001). The clade including Phytolaccaceae, Sarcobataceae, Nyctaginaceae, Gisekiaceae and Agdestidaceae also lacks internal reso lution. Additionally, some relationships have only been suggested on the basis of single -gene analyses (Cunoud et al., 2002) with many nodes scattered throughout the Caryophyllales lacking strong support. In an effort to resolve the remaining problematic deep level relationships within Caryophyllales, we constructed a much larger data set than employed previously. Our data set comprised eight plastid genes from single -copy (SC) regions, two nuclear genes and the entire plastid Inverted Repeat (IR; a combi ned analyzed length of 42,006 base pairs) for 40 taxa representing 31 families of the Caryophyllales and three families as outgroups. The unusual morphological and biochemical variation found within the Caryophyllales has fueled much speculation as to its evolutionary origins. Ehrendorfer (1976) outlined a plausible scenario to explain the coincident evolution of several unique characteristics found in the Caryophyllales, including floral variation and betalain pigmentation. He proposed that ancestral taxa in Caryophyllales occupied open, warm, dry and windy habitats with mineral soils. Reasons for assuming this ancestral habitat derive from the observation that many of the families in Caryophyllales currently inhabit xeric, marginal environments. Ehrendorfer (1976) argued that if the ancestral habitats were xeric, there would be strong selection for anemophily as pollinating insects would have been scarce in areas of little pioneer plant growth. He states that pollinators may have been scarce, as the time of the origin of the core Caryophyllales (104 111 MYBP:

PAGE 26

26 Wikstrom, 2001) predates the major diversification of insect pollinator lineages. He then argues that much of the floral variation and novel pigmentation in the core Caryophyllales could be interpret ed as the consequence of this anemophilous ancestry with reversals to zoophilly in extant lineages. However, as this study will demonstrate, the phylogenetic concept of the Caryophyllales has changed considerably since Ehrendorfer (1976). We evaluate Ehrendorfers hypotheses in light of a muchaltered phylogeny by examining patterns of pollination biology and perianth differentiation. We discuss the evolution of perianth differentiation in the context of the literature on perianth development within Caryop hyllales and use our phylogeny to identify broad trends in perianth evolution across the clade. Finally, we discuss the research opportunities that these patterns of morphological variation offer to the field of evolutionary developmental genetics. Mater ials and Methods Taxon Sampling In this analysis 31 families of Caryophyllales sensu APGII (2003; Cunoud et al., 2002) were represented. Som e families are monotypic (e.g. Drosophyllaceae, Halophytaceae, Stegnospermataceae), others comprise only one genu s (e.g. Asteropeiaceae, Nepenthaceae, Ancistrocladaceae, Frankeniaceae), or two or three genera (Achatocarpaceae, Dioncophyllaceae, Droseraceae, Limeaceae, Talinaceae). For larger potentially polyphyletic or paraphyletic families (e.g. Portulacaceae), mult iple genera were sampled to represent more of the phylogenetic diversity. The final data set included 36 taxa of Caryophyllales with an additional four taxa (Tetracera and Hibbertia representing Dilleniaceae, Berberidopsis and Vitis ) sampled as outgroups. Species, voucher information and GenBank accession numbers are given in tables 2 2 to 2 10). In some instances sequence data was combined from multiple species to represent a family this was judged not to significantly affect a family level analysis but the instances are

PAGE 27

27 listed here: Aizoaceae ( Delosperma napiforme Delosperma echinatum Delosperma cooperi ); Amaranthaceae ( Celosia argentea and Celosia cristata ); Cactaceae ( Opuntia microdasys and Opuntia dillenii); Didiereaceae ( Alluaudia ascendens and Alluaudia procera ); Dilleniaceae (Hibbertia volubilis and Hibbertia cuneiformis ); Gisekiaceae ( Gisekia africana and Gisekia pharnacioides ); Molluginaceae ( Limeum africanum and Limeum aetheopicum ); Plumbaginaceae (Limonium gibertii and Limonium arborescens, Plumbago zeylanica and Plumbago auriculata); Polygonaceae ( Polygonum sagittatum and Polygonum virginicum ); Portulacaceae ( Claytonia virginica and Claytonia perfoliata ). DNA Isolation and Amplification We isolated DNA following standard CTAB protocols (D oyle and Doyle, 1987) and Qiagen DNA extraction kits (Qiagen, Valencia, CA, USA). To augment depleted DNA stocks we carried out Multiple Displacement Amplification (MDA) using the Genomiphi kit (Amersham, Piscataway, NJ, USA) according to the manufacturer s instructions (Brockington et al., 2008). MDA -treated DNA was diluted 1:10 before further PCR amplification of targeted genes. We targeted 10 specific genes for sequencing eight plastid genes from the large and small SC regions and two nuclear genes; al l targeted genes and primers used for PCR and sequencing are provided in table 2 11. All PCR reactions contained Taq DNA polymerase (New England Biolabs MA, USA) and 10X Thermopol Reaction Buffer supplied by the manufacturer. The reaction volume was 25 l, and the final concentration of the components was Taq buffer (pH 8.8), MgCl2 (1.5 mM), 200 M dNTP, forward and reverse primers (1 M), 1U Taq polymerase, and 1 l of DNA. PCR cycling was carried out in an Eppendorf Mastercycler (Eppendorf, Westbury, NJ, USA (95oC for 3 min, followed by 30 35 cycles of 94oC for 30 sec, 50oC for 30 sec, and 72oC for 1 min with a final extension time of 7 min at 72oC). PCR products

PAGE 28

28 were purified using ExoSAP, and sequences were generated on an ABI 3730 XL DNA sequencer (Appl ied Biosystems, Inc., Fullerton, CA, USA) following the manufacturers protocol. Sequences were submitted to GenBank (numbers given in tables 2 2 to 2 10) The Amplification, Sequencing and Annotation of Plastomes (ASAP) method (Dhingra and Folta, 2005) wa s used to obtain the sequence of the plastid genome Inverted Repeat (IR) for 35 genera of Caryophyllales (the IR for Physena was not sequenced) and two members of Dilleniaceae. The published complete plastid sequencse of Spinacia (Schmitz Linneweber et al. 2001) and Plumbago (Moore et al., 2007 provided the IR sequence for these two taxa. The IR sequences were subsequently annotated using DOGMA (Wyman et al., 2004). IR sequences were submitted to GenBank (numbers given in tables 2 2 to 2 10). Alignment an d Phylogenetic Analysis Sequences were automatically aligned using Clustal X (Thompson et al., 1997) and then manually adjusted. Coding regions were aligned by predicted amino acid sequence. Regions at the beginning and ends of genes for which sequences we re incomplete, together with regions that were difficult to align, were excluded from the analysis. The total aligned lengths and the analyzed aligned lengths are given in table 1. Using the new sequences generated here together with those previously publi shed (cited in tables 2 2 to 2 10), we constructed six different data partitions: 1) individual plastid genes from the SC regions; 2) combined plastid genes from the SC regions; 3) two nuclear ribosomal RNA genes (18S rDNA and 26S rDNA); 4) plastid IR; 5) combined plastid SC and nuclear genes; 6) total evidence data set (all plastid and nuclear genes). All data partitions were subject to the following phylogenetic analyses. We used maximum parsimony (MP) and maximum likelihood (ML) to infer phylogeny. MP a nalyses were implemented in PAUP*4.0 (Swofford, 2000). Shortest trees were obtained using a heuristic search and 1,000 replicates of random taxon addition with tree -bisection -reconnection (TBR)

PAGE 29

29 branch swapping, saving all shortest trees per replicate. Boot strap support (BS) for relationships (Felsenstein, 1985) was estimated from 1 000 bootstrap replicates using 10 random taxon additions per replicate, with TBR branch swapping and saving all trees. For ML analyses we employed the program GARLI (Genetic Algorithm for Rapid Likelihood Inference; version 0.942) (Zwickl, 2000). GARLI conducts ML heuristic phylogenetic searches under the GTR model of nucleotide substitution, in addition to models that incorporate among-site rate variation, either assuming a gamma distribution ( ) or a proportion of invariable sites (I), or both. Analyses were run with default options, except that the significanttopochange parameter was reduced to 0.01 to make searches more stringent. ML bootstrap analyses were conducted with the default parameters and 100 replicates. We performed a strict consensus of five replicate GARLI analyses and topological differences resulting in collapsed nodes were annotated on the representative ML tree. Bayesian analyses were performed on the combined partition to generate trees for stochastic character mapping. Models of nucleotide substitution were determined using MrModeltest (Nylander, 2004). The Akaike information criterion (AIC) was used to select GTR+I+G as an appropriate model based on the rela tive informational distance between the ranked models. Analyses were implemented in MrBayes, version 3.1.2 (Huelsenbeck and Ronquist, 2001; Ronquist and Huelsenbeck, 2003). Two independent analyses each ran for 5 million generations, using four Markov chai ns, and with all other parameters at default values; trees were sampled every thousandth generation, with a burnin of 200,000 generations. Stationarity of the Markov Monte Carlo chain was determined by the average standard deviation of split frequencies b etween runs (after 5 million generations the average standard deviation was 0.004%) and by examination of the posterior in Tracer, version 1.3 (Rambaut and Drummond,

PAGE 30

30 2003). A majority rule consensus of post burnin trees were generated in PAUP*4.0 (Swoffor d, 2000), using the resulting posterior distribution of the trees. Character Reconstructions Parsimonybased reconstructions were achieved using standard unweighted parsimony character optimization and perfomed within Mesquite (Maddison and Maddison, 2008) Reconstructions focus on the core Caryophyllales. And were carried out using the MP topology derived from the total evidence data set. Reconstructions were further modeled by means of stochastic mapping techniques as described by Huelsenbeck et al. (2003) and implemented in SIMMAP (Bollback, 2006). This approach estimates the rates at which a discrete character undergoes state changes as it evolves through time. Bayesian estimation has several advantages over traditional parsimony based reconstruction. F irst, it allows one to average over equally likely topologies, which is valuable as the position of some taxa are poorly supported (e.g Limeum ) or poorly resolved (e.g taxa within the portulacaeous cohort and raphide clade). Second, it allows more than o ne character change per branch and is therefore a useful methodology for character reconstruction in the Caryophyllales a clade in which long branches are common. Posterior mapping requires the specification of prior values. The prior on the bias parame ter was fixed at 1/ k where k is the number of states (this being the recommended approach in SIMMAP for characters of more than two states; (Renner et al., 2007). We applied an empirical Bayesian approach in choosing appropriate priors for the substitutio n rate parameters following the method of (Couvreur et al., 2008a, 2008b). The gamma distribution of the substitution rate is governed by two hyperparameters defining the mean E(T) and the standard deviation SD(T). The values of these hyperparameters for t he prior gamma distribution were selected independently for each character using the number of realizations sampled from

PAGE 31

31 priors function in SIMMAP with 10,000 draws. A series of trials was performed (10,000 realizations in each) that systematically sampl ed for values of E(T) between 1 and 30, in combination with SD(T) values of either 1 and 5. The posterior distribution these combinations were visualized in Tracer v. 1.3 and further plotted as graphs of frequency against rate (see figure. 2 15). The post erior distribution curves derived from these trials allowed the selection of values of E(T) that gave highest sampling and allowed optimization of the E(T) value (Couvreur et al., 2008a, 2008b). A trial was also performed without specifying priors and allowing rates to be determined by branch lengths (as performed by Renner et al., 2007); however, the posterior distribution curves were generally highly skewed (see figure. 2 15), and thus this form of prior selection was not employed in subsequent analyses. Following exploration of different combinations of E(T) and SD(T), the prior E(T) values chosen for the characters were as follows: Perianth E(T) = 17, Pollination E(T) = 7 (marked with an asterix in figure. 2 15). For all of these values of E(T), an SD(T) value of 5 was applied in subsequent analyses, allowing a large standard deviation to accommodate uncertainty in mean rate of substitution. Following specification of priors, the rate and number of state transformations were estimated by 100 realizations on the 4800 post burnin trees (with branch lengths) from the Bayesian analyses. As recommended, branch lengths were rescaled so that the total tree length was 1 but the branch length proportions were maintained. The ancestral state at different nodes was assessed using a hierarchical Bayesian ancestral state reconstruction method implemented in the posterior ancestral states function of SIMMAP (Bollback, 2006). The nodes for which ancestral states were estimated are labeled in figure 2 3. The estimations of the posterior probability of ancestral character states at each node are listed in tables 2 14 and 2 15 and presented graphically on the nodes in figure 2 3.

PAGE 32

32 With parsimony reconstruction analyses, when more than one character state was present in th e family the representative taxon was coded as having more than one character state. In stochastic mapping analyses using SIMMAP terminals cannot be coded as having more than one state so in instances where more than one character state was present in the family, the representative taxon was coding as unknown (?). Information on pollination was derived primarily from entries in Kubitzki et al. (1993, 2003, 2007); pollination was coded as entomophilous or anemophilous. In the case where observations on polli nators have not been made, the character -state determination was unknown (?). All coding information is listed in tables 2 12 and 2 13. Our approach to coding perianth requires further clarification as there are many types of differentiated perianth wi thin Caryophyllales, and their homology is not always clear. Occurrences of differentiated perianth were given different character states where there are clear documented differences in development of the differentiated perianth. Data on perianth and devel opment were collated from the available literature (see figure 2 15). As reviewed by Ronse De Craene, 2008 criteria used to determine these differences in the literature include: meristic variation, sequence of organ initiation, difference in appearance at maturity, and presence of morphological intermediates. We were however interested in estimating the minimum number of origins of the differentiated perianth under parsimony and therefore applied a stringent approach to character coding, minimizing the num ber of character states to four: undifferentiated (0), differentiated with stamen -derived petaloid organs (1), differentiated with an involucre -derived outer whorl (2), and differentiated perianth of uncertain affinity (3). We employed a conservative approach to character coding, only assigning states 1 and 2 to taxa in which developmental morphological data was most conclusive. Where we were uncertain we assigned taxa to character

PAGE 33

33 state 3 differentiated perianth of uncertain affinity. We emphasize that t his coding does not reflect our belief that these instances of differentiated perianth are necessarily homologous, but in coding them as identical we ensure that estimation of the number of origins of differentiated perianth is conservative. The different iated perianths of Caryophyllaceae and Molluginaceae were assigned character state 2 due to the similar androecial nature of the petaloid perianth parts despite notable developmental differences (Ronse De Craene, 1998; 2007). In Aizoaceae, in the subfamili es Mesembryanthemoideae and Ruschoideae, the differentiated perianth is also the result of sterilization of outer members of a centrifugally initiating androecium and concomitant sepaloidy in the outer quincuncial whorl. Morphological intermediates clearly link the petaloid staminodes with fertile stamens (Brockington SF, unpublished observations; Ronse de Craene 2007). A similar form of perianth differentiation exists in Glinus in the Molluginaceae (Hofmann, 1994; Ronse de Craene 2007). However it is clear that Mesembryanthemoideae and Ruschoideae are derived within Aizoaceae and that the early diverging families do not have a differentiated perianth thus in contrast to Caryophyllaceae and Molluginaceae, the family is coded as undifferentiated for the pur pose of this reconstruction. Ronse de Craene (1998) describes the strong affinity of the petaloid members of the differentiated perianth to the androecium, in Caryophyllaceae. The involucral perianths of Nyctaginacaeae and the portulacaceous cohort were given the same character states in recognition of similar recruitment of bracteoles or bracts as a calyx, but again there are notable developmental differences (as described in Rohweder and Huber, 1974; Hofmann, 1994; Ronse De Craene, 2008). The quincuncia l differentiated perianth in Mirabilis jalapa is most likely the result of floral loss within an involucre (Vanvinckenroye et al. 1993) while the differentiated perianth in the portulacaceous cohort is probably the result of

PAGE 34

34 two addition phyllomes inserted between the bracteoles and inner petaloid quincuncial perianth members (Hofmann, 1994). The petals of Limeum and Stegnosperma are similar (and probably homologous to the petals in Caryophyllacaeae (Hofmann 1973, 1977; Ronse de Craene pers. comm) and are c oded as such. The nature of the perianth in both Asteropeia and Rhabdodendron is uncertain due to lack of detailed investigations and is coded as differentiated perianth of uncertain affinity. Cactaceae exhibit a great increase in perianth parts and thes e increases in floral merism and generally modified floral form make it challenging to determine correspondence between perianth in Cactaeae and its closest relatives in this study, Portulaca and Talinum Cactaceae was also therefore coded as of uncertai n homology. Non -core Caryophyllales and Outgroup taxa were coded as differentiated perianth of uncertain homology because the nature of the perianth relative to that within the core Caryophyllales is uncertain. Results Individual Plastid Single Copy Dat a Sets MP and ML trees from individual data sets are largely congruent with each other (figures 2 5 to 2 11; tree statistics shown in table 2 1). Consistent with the approximate nature of the GARLI approach to ML phylogeny estimation, replicate GARLI anal yses on the individual gene data sets do on occasion recover slightly different topologies. Taking into account nodes that are either unsupported or that collapse in the strict consensus however there are few instances of conflicting relationships between trees derived from different individual gene data sets. These examples of conflict include the following: in the matK MP tree, Delosperma and Gisekia were recovered as sister groups (51%); in the MP and ML ndhF tree, Spinacia and Stellaria were resolved as sister groups to the exclusion of Celosia (MP BS=100%); in the rbcL MP and ML tree Gisekia are sister to Rivina (MP BS=100%), Delosperma was sister to Phytolacca MP (BS=89%) and Stellaria was sister to the Amaranthaceae (MP BS = 53%); in the rpoC2 ML tre e

PAGE 35

35 alone, Phytolacca and Sarcobatus are recovered as sister groups (ML BS = 79%). Importantly, these anomalous relationships are not recovered or not supported in any other data sets, in either single gene or combined partitions. None of the trees derived from these individual plastid gene data sets gives good resolution across the tree, and deeper level relationships in particular are poorly supported. Inverted Repeat Data Set As with the individual plastid gene data sets, the IR partition generates MP and ML trees that are congruent (figure 2 12). Parsimony analyses recovered a single tree; replicate GARLI analyses recovered trees that differed only in the topology of the succulent clade. The IR tree differs from the previous analyses in the placement of Sarcobatus which is resolved as sister to Nyctaginaceae and Phytolaccaceae with strong support (ML BS = 100%). Furthermore, analyses of individual plastid genes resolve Talinum as sister to Portulaca and Cactaceae whereas the IR data set recovers Port ulaca and Talinum as sister to each other. Combined Plastid Genes from the Single Copy Region The combined plastid gene data set generated a single most parsimonious tree. Replicate GARLI analyses recovered the same ML topology (figure 2 13). Levels of bo otstrap support are higher in general in the ML tree than in the MP tree. Again, the MP and ML trees are largely congruent, although in the MP tree Sarcobatus and Rivina are sister to each other whereas in the ML tree Sarcobatus is placed without support w ith Phytolaccaceae and Nyctaginaceae. In the MP tree Limeum is placed without support as sister to Mollugo and the succulent clade and Stegnosperma is placed as sister to the globuloid inclusion clade; in the ML tree, Limeum and Stegnosperma are placed as successive sisters to the globular inclusion clade. As with the individual plastid gene trees, Talinum is resolved as sister to Portulaca and Cactaceae, but

PAGE 36

36 relationships among Didiereaceae, Basellaceae, Halophytum and Claytonia are either poorly su pported or unresolved. Combined Plastid Single Copy and Nuclear Genes The addition of the nuclear gene data set to the combined plastid genes has little effect on topology (figure 2 14). In contrast to analyses of the combined plastid gene data set, both the MP and ML analyses with the nuclear data recover Stegnosperma and Limeum as successive sisters to the globular inclusion clade. Again the MP and ML trees differ in their placement of Sarcobatus in the same way as in the combined plastid tree topologies, i.e. as sister to Rivina in the MP tree but sister to Rivina and Nyctaginaceae in the ML tree; both placements have low bootstrap support (~60%). As in the IR ML tree topology, the MP recovers Portulaca and Talinum as sister to Cactaceae but without s upport; in the ML tree, however, Talinum is sister to Portulaca plus Cactaceae. Total Evidence Data Set The total evidence data partition generated a single MP (figure 2 2) tree that agrees in topology with the ML tree (figure 2 1), except for the placem ent of Sarcobatus The MP and ML trees derived from the total evidence data set show more congruence with each other than the congruence found between MP and ML trees derived for any other data partition. As in previous combined analyses, in the MP tree, Sarcobatus and Rivina are placed sister to each other (BS=63%) while in the ML tree Sarcobatus is placed without support as sister to Nyctaginaceae plus Rivina The position of Sarcobatus therefore remains uncertain in these analyses. The ML topology was chosen as the basis of subsequent character reconstruction analyses because it is less prone to the problem of longbranch attraction (Felsenstein, 1978) and because the bootstrap values are higher than in the MP tree. The full topology of the tree is the refore described in detail here.

PAGE 37

37 Non -core Caryophyllales form a strongly supported (BS=100%) monophyletic group with two subclades. One clade comprises Plumbaginaceae with Polygonaceae resolved as sister to Frankeniaceae plus Tamaricaceae (all with BS of 100%). The second clade, containing the carnivorous taxa and relatives, comprises Drosophyllaceae with Ancistrocladaceae and Dioncophyllaceae (BS=100%), and Nepenthaceae with Droseraceae (BS=59%). Core Caryophyllales form a strongly supported group (BS= 100%) with Rhabdodendraceae as sister to the rest (BS=100%). Following the divergence of Rhabdodendraceae, the backbone of the tree is strongly supported and characterized by a grade of successively branching taxa, in the following order: Simmondsiaceae; A steropeiaceae with Physenaceae; a clade comprising Caryophyllaceae, Achatocarpaceae, and Amaranthaceae (BS=100%); and Stegnospermataceae (BS=100%). Subsequently, Limeum is placed as sister to the remaining members of Caryophyllales, which form two clades. In the first of these two clades, the topology is as follows: the earliest -diverging group is Aizoaceae, followed by Gisekia Phytolacca, Sarcobatus Rivina and Nyctaginaceae. In the second clade, Molluginaceae are sister to a group comprising Cactaceae, P ortulacaceae, Didiereaceae, Basellaceae, Halophytum and Claytonia. Within this group, Portulaca and Talinum are strongly supported as sister to Cactaceae; however, relationships among Didiereaceae, Basellaceae, Halophytum and Claytonia are poorly supported. For each of the character reconstructions, multiple state transitions are inferred within the core Caryophyllales. The patterns of character evolution derived from parsimony reconstruction and the inferred ancestral states derived from stochastic mapping analyses are illustrated in figure 2 3. Discussion Several broad molecular phylogenetic analyses have examined intra -ordinal relationships across the entire Caryophyllales sensu lato. Rettig et al. (1992) conducted an rbcL analysis of 12

PAGE 38

38 families; Downie and Palmer (1994) inferred phylogeny from chloroplast genome structural changes and Inverted Repeat restriction site variation in 11 families of Caryophyllales; Downie et al. (1997) compared sequences of ORF2280 ( ycf2 ) across 11 families. However, the most comprehensive study is that of Cunoud et al. (2002), who generated a partial matK sequence phylogeny (30 families, 121 genera). In Cunoud et al. (2002) a subset of the matK data was combined with previously published genes to generate a combined rbcL/ma tK phylogeny (19 families 53 genera) and a four -gene analysis that also incorporated atpB and 18S rDNA sequences (19 families, 25 genera). Although the taxonomic sampling of the matK phylogeny was extensive and dramatically improved our understanding of t he Caryophyllales phylogeny, the study suffered from restricted taxon sampling in the combined analyses, with just over half of the families in the core Caryophyllales represented in the rbcL/matK and rbcL/matK/atpB/18S data sets. Parsimony was the only optimality criterion used in these analyses and there were several soft incongruences among the matK, rbcL/matK and rbcL/matK/atpB/18S trees. Our analyses resolve many of these remaining uncertainties. Phylogenetic Analyses The earliest -diverging lineages in the core Caryophyllales are clarified and are well supported. Notably, Rhabdodendraceae followed by Simmondsiaceae are supported as sisters to the rest of the core Caryophyllales (both with 100% BS; figure 2 1). The position of Rhabdodendron had previously been ambiguous, recovered either as sister to both core and noncore Caryophyllales (in combined matK and rbcL analyses: Cunoud et al., 2002) or weakly supported as sister to Simmondsia, at the base of the core Caryophyllales (in matK analysis: Cuno ud et al., 2002). Following the divergence of Rhabdodendraceae and Simmondsiaceae, our analyses strongly support a clade of Physenaceae and Asteropeiaceae (100% BS) as sister to the remaining core Caryophyllales. Relatively little is known about these earl y -diverging lineages of

PAGE 39

39 core Caryophyllales from the perspective of morphology, and lack of data for these critical early lineages prevents a clear understanding of ancestral states within the core Caryophyllales. In the matK analysis of Cunoud et al. (2002), Caryophyllaceae and Achatocarpaceae plus Amaranthaceae s.l. branch successively as sister to the rest of the core Caryophyllales and do not form a clade with Achatocarpaceae and Amaranthaceae s.l. as suggested in the rbcL/matK analyses. In our analyse s, the clade comprising Caryophyllaceae, Achatocarpaceae and Amaranthaceae s.l. receives strong support (100% BS), agreeing with the combined analyses of Cunoud et al. (2002). Morphological synapomorphies for this clade remain elusive, probably in part be cause Achatocarpaceae are poorly studied. Molecular studies have consistently recovered a distinct clade within the core Caryophyllales (Cunoud et al., 2002; Downie et al., 1997; Giannasi et al., 1992; Rettig et al., 1992) termed the globular inclusion clade (Aizoaceae, Phytolaccaceae, Nyctaginaceae, Gisekiaceae, Molluginaceae, Portulacaceae, Didiereaceae, Basellaceae, Cactaceae; Cunoud et al., 2002) on account of distinctive P plastid characteristics (as found by Behnke, 1993). Consistent with previous analyses, Stegnosperma and Limeum are recovered as successive sisters to this globular inclusion clade but with greater support (BS 98%) than in earlier studies. Within the globular inclusion clade, two subclades are recovered that correspond to the raphide clade (Judd et al., 1999) and the succulent clade (Rettig et al., 1992). Molluginaceae are maximally supported as sister to this succulent clade (BS 100%). In Cunoud et al. (2002), Molluginaceae were placed as sister to the succulent clade but with no support in the matK and matK / rbcL analyses and moderate support (BS 70%) in the four gene analysis. Within the raphide clade, Gisekia is strongly supported as sister to Phytolaccaceae, Rivina Sarcobatus and Nyctaginaceae (BS 100%). This contradicts the findings of Cunoud et

PAGE 40

40 al. (2002) whose single -gene analyses variously place Gisekia as sister to Aizoaceae ( matK ) or Rivina (rbcL and rbcL / matK ) analyses. Out of the eight plastid genes that we analyzed, only rbcL supports a sister relationship between Gisekia and Rivina In addition, we provide further evidence and support for the separation of Rivina (Rivinioideae) from Phytolaccaceae and its placement as sister to Nyctaginaceae (BS=77%). The placement of Sarcobatus in our analyses is problemati c and its position varies in relation to Nyctaginaceae, Rivina and Phytolaccaceae. The raphide clade is under -sampled in this study and while we suggest alternative placements for Gisekia and Sarcobatus we recognize that increased taxon sampling (e.g. Agdestis which has been associated with Sarcobatus by Cunoud et al., 2002) could affect these findings. Taxa within the portulacaceous cohort have traditionally been treated at the rank of family; however, the degree of paraphyly in Portulacaceae sugge sts that phylogenetic resolution should be conceptually envisioned as an intrafamilial problem (Hershkovitz and Zimmer, 1997). For example, in analyses of ITS, the genetic divergence of Cactaceae from Portulacaceae is equal to or less than that between man y pairs of genera in Portulacaceae (Hershkovitz and Zimmer, 1997). Two methodological constraints limited our ability to address the question of phylogenetic relationships among Cactaceae and its portulacaceous relatives. First, the large amount of sequenc ing for each taxon limited the total number of taxa that could be sampled this under -sampling is particularly acute in the succulent clade, given the degree of paraphyly inherent in Portulacaceae. Second, the broad scope of the taxon sampling, i.e. the w hole of Caryophyllales s.l. meant that only slower -evolving coding genes rather than more rapidly evolving regions such as intergenic spacers could be sampled to permit alignment across the order. Consistent with the low levels of genetic divergence in this clade, very little informative variation was obtained for members of the portulacaceous cohort from these coding regions

PAGE 41

41 (despite sequencing over 40,000 base pairs). There are fewer than 20 substitutions on the branches leading to the clade containing H alophytum Alluaudia and Basella and fewer than 50 substitutions on the branch leading to Portulaca and Talinum Limitations aside, within the portulacaceous cohort, the monophyly of the ACPT clade comprising Cactaceae ( Pereskia plus Opuntia), Portulaca a nd Talinum is strongly supported (100% BS) in analyses of all partitions with the exception of the IR partition. Historically, however, different analyses have recovered different patterns within the ACPT clade depending on taxon sampling and phylogenetic markers used (reviewed in Nyffeler, 2007). In our analyses, different partitions and analytical methods also gave different branching patterns within the ACPT clade. Parsimony analyses of the IR partition recovered a branching pattern (with support less th an 50%) similar to the morphological cladistic analysis of Carolin (1987). Parsimony and GARLI analyses of the plastid gene partition found Talinum sister to Portulaca plus Cactaceae as proposed by the morphological cladistic analyses of Hershkovitz (1993) and the Bayesian molecular analysis of Nyffeler (2007). Parsimony and GARLI analyses of the total evidence and GARLI analyses of IR data sets recovered a well -supported branching pattern not found in previous analyses Cactaceae sister to Portulaca plus Talinum Total evidence, plastid gene, plastid plus nuclear and IR data sets all place Halophytum as sister to Basellaceae and Didiereaceae. This affiliation is consistent with Savolainen et al.s (2000) analysis of rbcL sequences (albeit with low taxon s ampling in the Caryophyllales), and was suggested by Bittrich (1993) on the basis of pollen morphology and by Hunziker et al. (2000) due to shared similarities in basic chromosome number ( x =12). The position of Claytonia however, is unstable in our analys es and is generally not in agreement with studies with better taxon sampling (Applequist and Wallace, 2001; Hershkovitz and Zimmer, 1997; Nyffeler, 2007).

PAGE 42

42 Claytonia together with associated portulacaceous genera, were placed in a clade with Basellaceae a nd Didiereaceae with reasonable support (80% BS; Applequist and Wallace, 2001). However, in our study only MP analyses of the matK and IR data sets were able to recover this relationship. Individual plastid genes placed Claytonia in a variety of positions while the combined data sets invariably placed Claytonia as sister to the rest of the portulacaceous cohort. Notably, our phylogeny derived from ndhF alone (the same gene employed by Applequist and Wallace, 2001) also recovered Claytonia as sister to the r est of the portulacaceous cohort. This suggests that the apparent instability in the placement of Claytonia may be the result of limited taxon sampling in our study; pruning the data set from Applequist and Wallace (2001) to match our taxon sampling genera ted a similarly anomalous placement of Claytoni a (data not shown). Reconstruction of Pollination Mechanism Our phylogeny differs considerably from the concept of the Caryophyllales that stimulated the speculations of Ehrendorfer (1976). The Caryophyllales sensu Ehrendorfer essentially correspond to the core Caryophyllales presented in this study; however, the composition and phylogeny of this clade has changed considerably. None of the four of the currently recognized early diverging lineages were recognize d as belonging to the Caryophyllales in the 1970s. Ehrendorfer was strongly influenced by the idea that the Chenopodiaceae (Amaranthaceae s.l.) with their reduced anemophilous flowers were representative of the ancestral Caryophyllid type. Consequently, he argued that anemophily was the ancestral condition because the early Caryophyllales had evolved in open, dry, marginal environments at a time when pollinators were scarce. These hypotheses to prove or disprove (Clement and Mabry, 1996); however, our phylo geny confirms that the Amaranthaceae constitute a relatively derived lineage. If pollinators were scarce at the time of origin of the Caryophyllales, this might also apply as a general limitation to other lineages of eudicots diverging at that time, but in any case, the relative timing

PAGE 43

43 and location of diversification in eudicot lineages and their respective pollinator lineages is unclear at present. Friedman and Barrett (2008) demonstrate a strong correlation between the occurrence of open habitat and anemo phily and provide support for the prevalence of anemophily in open habitats; however, this correlation may not necessarily be due to pollinator scarcity but rather to the selective advantage of wind pollination in an open environment. Moreover, parsimony r econstruction of the ancestral habitat would be ambiguous given that the extant members of early -diverging lineages of the core Caryophyllales occupy tropical under story (Asteropeiaceae and Rhabdodendraceae) or have a global holoarctic distribution (Caryo phyllaceae). Using the current phylogeny, parsimony-based character reconstruction and stochastic character mapping do not provide support for the hypothesis that the Caryophyllales were ancestrally wind -pollinated. R habdodendron, which is sister to all o ther core Caryophyllales, is described as visited by pollen -collecting bees (Prance, 2003) while extensive field observations suggest that Asteropeiaceae are also entomophilous (Birkinshaw et al., 2004). It is notable, however, that together with the windpollinated Simmondsia, two other early diverging lineages do at least exhibit morphological characteristics that are reminiscent of wind -pollinated flowers. Despite reports of bee visitation, Rhabdodendron exhibits very long anthers and sepaloid petals, l acks a nectary, possesses a gynoecium with only 1 or 2 ovules and a single seed in fruit, and has a relatively long stigma (P.K Endress, pers. comm.); perianth parts also fall off as the flower opens (Nelson and Prance, 1984). Physena exhibits very long an thers and no petaloid organs, lacks a nectary and has a large stigmatic surface. Coding Physena as wind -pollinated, however, does not alter the conclusion of the character mapping analyses, and thus there is little support for an anemophillous ancestry in the core Caryophyllales. Indeed, as noted by Clement and

PAGE 44

44 Mabry (1996), even if one accepts the highly reduced inconspicuous flowers of Amaranthaceae s.l. as archetypal, it is not necessary to invoke wind pollination at is has already been noted that many o f the diminutive flowers in Amaranthaceae are probably entomophilous (Blackwell and Powell, 1981; Kuhn, 1993). Reconstruction of Perianth Differentiation Despite recovering entomophily as ancestral, our character reconstruction analyses suggest that an undifferentiated perianth arose early within the core Caryophyllales (in agreement with Ehrendorfer, 1976; Ronse De Craene, 2008). This perianth type has been strongly correlated with anemophily (Friedman and Barret 2008). Parsimony reconstruction infers that the evolution of this undifferentiated, uniseriate condition evolved after the divergence of Rhabdodendron while the stochastic mapping analyses recovers the basalmost node in core Caryophyllales as either undifferentiated or differentiated, with equal pr obability. Subsequent nodes along the backbone of the tree until the divergence of Molluginaceae are recovered as undifferentiated (with greater than 0.99 posterior probability). A discussion of perianth evolution within the core Caryophyllales is compl icated by the great diversity of floral structure within the order and uncertainty in defining the correspondence of these structures both within Caryophyllales and with respect to floral organs in other eudicots. Observations by Ronse De Craene (2007, 200 8) suggest that although petal organs in eudicots may appear homologous with respect to position and superficial appearance, the variable expression of features reminiscent of either stamens or bracts means that petals in different lineages of core eudicot s are of uncertain homology and may have been differently derived. This viewpoint argues against the widespread notion that the petals within eudicots are invariably derived from stamens and makes it difficult to homologize between perianth parts even with in the core eudicots (Ronse De Craene, 2007; 2008). In the Caryophyllales, these difficulties are

PAGE 45

45 compounded by different floral structures and the limitations of established terminology. The specific terms petal, sepal, corolla and calyx are not usefully applied to Caryophyllid taxa as they imply not only the characteristics and function of an organ but also the position of the organ (Endress, 1994; Jaramillo and Kramer, 2007). In core eudicots, for example, the term petal implies both the showiness of the perianth part and the position of the organ in the second whorl of the flower (Jaramillo and Kramer, 2007). However, in comparing the differentiated perianth found in many Caryophyllales with the bipartite perianth of most core eudicots, the positional fe ature alone is not necessarily a sufficient criterion of homology, and terms such as petal that imply positional correspondence are misleading. Similarly, the term bipartite perianth should also be avoided, as this implies the presence of distinct peri anth whorls. While distinct perianth whorls may be found in some families in the Caryophyllales e.g Limeaceae (Hofmann, 1973) and Caryophyllaceae (Rohweder, 1967), differentiation in a spiral phyllotaxis occurs in Cactaceae. For the purposes of this discu ssion, therefore, we use the term differentiated to describe a perianth that comprises at least two distinct types of organ that perform the functions commonly ascribed to the calyx and corolla. We refer to members of a differentiated perianth as either petaloid or sepaloid (i.e., resembling the petals or sepals of other core eudicots and putatively performing similar functions without necessarily being homologous by positional criterion alone). Finally, because the terms petaloid and sepaloid refer only to a superficial resemblance and putatively similar function, within Caryophyllales we apply these terms to structures that are clearly non -homologous in other respects. The terms sepaloid tepal and petaloid tepal are applied to the quincuncial periant h parts that are present in core Caryophyllales while petaloid staminodes refer to perianth parts that are clearly androecium derived.

PAGE 46

46 Multiple Origins of Perianth Differentiation Our analyses suggest that there have subsequently been a minimum of nin e independent origins of a differentiated perianth within the Caryophyllales. This is more than the minimum suggested by Ronse De Craene (2008) who, in a broad survey of eudicots, cites five origins of petals in the core Caryophyllales, occurring in Ste gnospermataceae, Aizoaceae, Portulacaceae clade, Caryophyllaceae, Molluginaceae. Considering the reconstructions provided by Ronse De Craene (2008), the difference in our respective estimations of perianth differentiation can be attributed to several fact ors. Most significantly, coding and definition of the perianth differs in our studies; e.g. Nyctaginaceae and the portulacaceous cohort are coded as petals a bsent (Ronse De Craene, 2008, F igure 2 3) but by our definition both Mirabilis jalapa (Nyctaginac eae) and the portulaceous cohort have a differentiated perianth and are listed as polymorphic and differentiated, respectively Similarly, Glinus is a member of the Molluginacaeae that possesses putatively staminodial petals (Hofmannn, 1994; pp. 137, 141) but is coded as petals absent by Ronse De Craene (2008). In our analysis, Molluginaceae are coded as polymorphic, exhibiting both taxa with a uniseriate, undifferentiated perianth and taxa with differentiated perianth. A different tree topology may als o be a factor contributing to the different results, e.g. Rhabdodendron is sister to the core Caryophyllales (this study), and the placement of Corbichonia and Glinus as successive sisters to the globuloid clade by Ronse De Craene (2008) is erroneous bas ed on current understanding; Corbichonia is most likely sister to the raphide clade (Cunoud et al. 2002), and Glinus (Molluginaceae) is sister to the portulacaceous cohort (70% BS according to Cunoud et al., 2002 and 100% BS in this study). Finally, du e to the broader scope of the study by Ronse De Craene (2008) (i.e. all eudicots), lineages of core Caryophyllales with differentiated perianth are also under -sampled, with both Asteropeia and Limeum excluded from his study. Consequently, differences in em phasis, sampling, coding, and

PAGE 47

47 tree topology may all have contributed to the differences between Ronse De Craenes 2008 results and those we report here. Despite these differences, the five independent origins of petals described by Ronse De Craene (2008 ) are included in the nine independent derivations of differentiated perianth inferred in our study. These nine origins occur in: Asteropeiaceae; Caryophyllaceae (although several genera do not have differentiated perianth); Stegnospermataceae; some specie s of Limeum ; Corbichonia (not sequenced in this study); the subfamilies Mesembryanthemoideae and Ruschioideae within Aizoaceae; Mirabilis in Nyctaginaceae; Glinus in Molluginaceae; the portulacaceous cohort (Portulacaceae, Didiereaceae, Basellaceae) includ ing Cactaceae. The number of origins of differentiated perianth could well increase depending on the final placement of the enigmatic Macarthuria and increased resolution of phylogenetic relationships within Caryophyllaceae. Developmental evidence (where a vailable) is consistent with these independent origins of differentiated perianth indicated by character reconstruction analyses. Below, we discuss the developmental evidence for perianth differentiation by different mechanisms in these nine lineages: thro ugh differentiation of putatively homologous organs and through the recruitment of floral structures derived either from the androecium or preceeding bract. Recruitment of Preceding Bracts The secondary recruitment of preceding bracts to from perianth pa rts has occurred twice within the globular inclusion clade, once in Nyctaginaceae and again in the portulacaceous cohort (F igure 2 3). Developmental studies suggest that the mechanism underlying the recruitment of preceding bracts is different in these distinct lineages. Within Nyctaginaceae, an involucre may have evolved more than once, occurring also in Abronia, Allionia Boerhavia, Bougainvillea Mirabilis, Nyctaginia and Tripterocalyx (Douglas and Manos, 2007). In Boerhavia Sharma (1963) describes a n involucre surrounding five lateral flowers and one central

PAGE 48

48 flower. In Mirabilis however there appears to a tendency to reduction in floral number. In Mirabilis nyctagineus only the first three leaves of the involucre produce axillary flowers (Hofmannn, 1994). In Mirabilis jalapa each flower has a differentiated perianth with a calyx of five fused parts that has been secondarily derived from an invol ucre of bracts (F igure 2 4; L). This variation within Mirabilis suggests that the apparent bipartite perian th in M. jalapa may have been derived through reduction in floral number. Subsequent loss of five lateral flowers is inferred, leaving a single central flower with the involucre appearing as a pseudocalyx (Vanvinckenroye et al., 1993). In Nyctaginaceae, fl oral loss within an involucre of bracts would appear to result in an apparently differentiated perianth, although the association of the involucre (calyx) with the rest of the flower is weak (Rohweder and Huber, 1974). Portulacaceae s.l. Didiereaceae and Basellaceae share a distinct floral morphology that emerges following th e divergence of Molluginaceae (F igure 2 3). These lineages possess flowers with an involucre which in contrast to Mirabilis (Nyctaginaceae) comprises two leafy phyllomes that are inser ted below the petaloid members of the perianth. Importantly, their developmental origin is probably different from the involucre found in Nyctaginaceae as Hofmann (1994) comments that axillary products are never formed in the axes of these phyllomes. They are termed involucral phyllomes by Hofmann (1994), reflecting the belief that these organs are additional phyllomes inserted between the bracteoles and the sepals; however, there have been other interpretations as to the nature of these phyllomes. Sharma (1954) (reviewed in Milby, 1980), who examined vascular anatomy in Portulaca and Talinum concluded that the flowers are essentially dimerous with the pentamerous petaloid perianth inferred as a derived condition. These alternatives will merit further deve lopmental study as phylogenetic understanding within this group is clarified but it is valuable to consider these different interpretations in light of the

PAGE 49

49 current phylogeny and per ianth reconstruction analysis (F igure 2 3). The ancestral floral condition of the portulacaceous cohort is uniseriate pentamery; therefore, Sharmas interpretation (1954) suggests reduction to a dimerous state followed by a reversal to a pentamerous condition. Irrespective of the developmental origin of these two phyllomes, in ma ny species they cover the developing floral meristem very early in development and thus perform the function of a calyx in a differentiated perianth (Hofmann, 1994). Subfunctionalization of perianth roles may have facilitated the high degree of petaloidy i n the inner quincuncial peria nth members of these families (F igure 2 4; B, I, L): the involucral phyllomes cover the inner bud very early and take over the function of the calyx. Therefore, the sepals [uniseriate pentamers] behave like petals (Hofmann, 1994). Within the portulacaceous cohort, a very different floral structure is found in Cactaceae. In contrast to the perianth of Portulacaceae s.l., Didiereaceae and Basellaceae, Cactaceae exhibit a great increase in perianth parts. Increases in floral me rism and generally modified floral form make it challenging to determine correspondence between perianth in Cactaeae and its closest relatives in this study, Portulaca and Talinum The perianth parts of Cactaceae are suggested to be bracteal rather than staminodial in their homology (Buxbaum, 195055 pp 122123; Ronse De Craene, 2007; Ronse De Craene, 2008) and are arranged in a spiral phyllotaxy. The perianth may have arisen by inclusion and differentiation of supernumary bracts (Ronse De Craene, 2008) o r simply by formation of additional bracts. Differentiation of the perianth occurs, with outer sepaloid parts an d highly petaloid inner parts (F igure 2 4; D). This high degree of differentiation together with a spiral phyllotaxy is unusual within the Caryophyllales however Ronse de Crane (2008) highlights that this combination of floral characters (large, spirally arising petals with a multi -staminate androecium) occurs in several derived lineages in the core eudicots. Endress

PAGE 50

50 (2002) suggested that increas es in numbers of stamens and/or carpels may result in an increase in size of the flower, greater plasticity, and an irregular petal development This developmental interpretation is consistent with our reconstruction analyses, which do not argue for an ind ependent origin of differentiated perianth in Cactacaeae; rather, an increase in meristem size and merosity of reproductive organs may be in part responsible for the unusual perianth in Cactaceae. Petaloid Modification of the Androecium Reconstruction anal yses suggest that perianth differentiation through sterilization and petaloid modification of the outer members of a centrifugally initiating androecium has arisen a minimum of three times in Caryophyllales (F igure 2 3): clea r examples occur in Aizoaceae (F igure 2 4; C), Molluginaceae and Corbichonia (not sampled in this study but shown to be a distinct lineage within the raphide clade: Cunoud et al., 2002). In Glinu s, Corbichonia and Aizoaceae, the petaloid structures can be readily interpreted as differentiated staminodial structures (Ronse De Craene, 2008). For example, within Aizoaceae subfamilies, Ruschioideae and Mesembryanthemoideae, androecial development proceeds centrifugally, and the basipetal members become progressively more sterile and petal oid with intermediates conceptually linking the outermost petals to the inner fertile stamens (F igure 2 4; C). A similar situation has been described in Glinus in the Molluginaceae (Hofmann, 1994) and in Corbichonia (Ronse De Craene, 2007). Petaloid membe rs of the differentiated perianth in Caryophyllaceae, Limeum Stegnosperma and Macarthuria have also been attributed to the androecium (Hofmann, 1973; Ronse De Craene, 2007; 2008). The reconstruction analyses suggest that these differentiated perianths have occurred independently and thus merit further developmental study. The assessment of homology between the petaloid members of the differentiated perianth and the

PAGE 51

51 androecium is complicated by a high degree of variability in androecium organization, proces ses of reduction and differences in phyllotaxy. However, several lines of evidence suggest an androecial origin of the petals in Caryophyllaceae (Rohweder, 1967) (reviewed in Ronse De Craene 1998, 2007, 2008). Ronse De Craene et al. (1998) review the pre sence and absence of petals in 52 genera of Caryophyllaceae: nine genera lack petals, 11 genera have both species with petals and species without, while the remaining 32 genera in the survey possess petals. It remains unclear whether the absence of petals is ancestral in Caryophyllaceae or whether instances of petal loss have occurred. The most comprehensive molecular phylogeny of Caryophyllaceae to date (Fior et al., 2006) sampled only two genera with apetalous members ( Paronychia and Sagina), but none of the entirely apetalous genera were sampled. Differentiation of Homologous Perianth Parts Despite the high degree of variation in floral structure found in different lineages of Caryophyllales, there are key common elements. Almost all lineages within the order possess five perianth members that are organized in a uniseriate quincuncial arrangement (with the exception of Cactaceae, which is multiseriate). Occasionally, one of the members in this series has been lost to give a tetramerous perianth e.g. in Te tragonia and Mesembryanthemum in Aizoaceae, Rivinoideae and Didiereaceae, but these cases of tetramery are clearly derived from pentamerous ancestors. These uniseriate quincuncial perianth members are probably homologous, given their constancy, position i n the flower and common phyllotaxy. These putatively homologous organs have, however, undergone considerable differentiation in certain lineages, which often correlates with the emergence of a differentiated perianth. For example, in Aizoaceae, members of the early diverging subfamilies Sesuvioideae and Aizooideae possess a quincuncial uniseriate perianth whose members are petaloid on the adaxial surface and sepaloid on the abaxial side. In the derived subfamilies Mesembryanthemoideae and Ruschioideae, the

PAGE 52

52 androecium is polyandrous, and the outer members of centrifugally initiating stamens are sterile, resulting in a differentiated perianth with androecial -derived petaloid organs. Concomitantly, the outer quincuncial uniseriate perianth loses all petaloid ch aracteristics and resembles only a calyx. In instances where differentiation of the perianth has been achieved through recruitment of involucral bracts and/or bracteoles (Portulacaceae and Mirabilis ), the involucral organs act as a calyx and the now -inner uniseriate quincuncial perianth members are considerably more showy and petaloid (compare the showy petaloid perianth of the portulacaceous cohort ( Figure 2 4; B, D and I) with the diminutive simple perianth of some genera in Molluginaceae. Seemingly homol ogous perianth parts within Caryophyllales can be either petaloid, e.g. Nyctaginaceae (Figure 2 4; A, L) and Portulacaceae ( Figure 2 4; I), or sepaloid e.g. Limeum Stegnosperma (Figure 2 4; E), Molluginaceae, Ruschioideae and Mesembryanthemoideae ( Figure 2 4; C), Caryophyllaceae ( Figure 2 4; J) and Simmmondsia or chimeric, e.g. Sesuvioideae/Aizooideae (Figure 2 4; F), Hypertelis ( Figure 2 4; G). Caryophyllales as a System for Floral Evo -D evo Nine independent origins of a differentiated perianth, the conc omitant independent evolution of petaloid from either androecial or bracteal organs and varying degrees of petaloid differentiation in homologous structures across the order make the Caryophyllales a valuable system for exploring the evolutionary developme ntal genetics of petaloidy in core eudicots. In the majority of core eudicots whose petal developmental genetics have been examined (e.g. in Arabidopsis thaliana, Antirrhinum majus, Solanum lycopersicon, Nicotiana tabacum, Petunia hybrida), differentiation of the petals is strongly influenced by MADS box transcription factors: APETALA3 (AP3 ) and PISTILLATA ( PI ) (in A. thaliana) and their orthologues (Irish and Kramer, 1998; Kramer and Irish, 1999). In these core eudicot species, AP3 and PI orthologues are expressed throughout the development of the petal and their ubiquitous expression in the

PAGE 53

53 petal has been shown to be necessary for normal petal development in A. thaliana and A. majus (Bowman et al., 1989; Sommer et al., 1991; Zachgo et al., 1995). It seems apparent that these genes play a conserved role in petal identity in the core eudicots examined so far, yet core eudicot petals have also traditionally been considered to be homologous, stamen -derived organs: this homology has been invoked to explain such developmental genetic similarities (Irish and Kramer, 1998). More recently, however, the assertion that petals in core eudicots are largely homologous and predominantly stamen derived has been questioned (Ronse De Craene, 2007). Although in Caryophyllales the homology of the perianth parts to petals in other core eudicots is uncertain, it is clear that many lineages (Sesuvioideae, Nyctaginaceae, Portulacaceae, Cactaceae) possess petaloid organs that are bracteal rather than staminal in origin. Furthermore the occurrences of stamen -derived petals within Caryophyllales (Caryophyllaceae, Aizoaceae, Glinus and Corbichonia) are phylogenetically, derived, independent events. These independent occurrences are valuable for further study as there are very few exampl es of petals within core eudicots that are unquestionably stamen -derived (Ronse De Craene, 2007). The pattern of perianth evolution in the Caryophyllales therefore presents a unique opportunity to address long standing questions regarding differences and/or similarities in the developmental genetics of bracteopetals and andropetals (Ronse De Craene, 2008) articulated by Ronse De Craene (2008) this question remains highly pertinent in studies of floral diversification. The improved phylogenetic understand ing reported here provides opportunities for comparing bracteopetalous and andropetalous lineages that have arisen more recently than both basal angiosperms (traditionally considered to bear bracteopetals) and the early-diverging eudicot lineages (with the ir presumed andropetals). The Caryophyllales are a well defined clade within core eudicots but, in a sense, the patterns of perianth evolution discussed here recapitulate

PAGE 54

54 (on a smaller phylogenetic scale) the evolutionary trends traditionally thought to ha ve taken place across the angiosperms as a whole (Bessey, 1915; Takhtajan, 1991). Therefore, despite uncertainty surrounding the precise correspondence of the caryophyllid perianth with the perianth of other eudicots, evo-devo investigations in the Caryophyllales may have far reaching implications for our understanding of petal evolution and perianth differentiation. Is there latent developmental genetic homology underlying these derived and oft -seemingly dissimilar occurrences of perianth differentiation i n Caryophyllales? What is the involvement of AP3 and PI orthologues in these bracteopetals in Caryophyllales? How do expression and function of AP3 and PI orthologues in caryophyllid bracteopetals compare with expression and function in the derived instanc es of andropetals? How expression and function of AP3 and PI compare in the different occurrences of andropetals? Evolutionary developmental approaches to these questions are currently underway (Brockington et al., 2007) and may shed light on the evolutionary origins and homology of these diverse perianth forms in relation to other petals in core euicots.

PAGE 55

55 Figure 2 1. Maximum likelihood (ML) tree resulting from GARLI analysis of total evidence data set (2 nuclear genes, 8 plastid genes from the single c opy region, and the Inverted Repeat) for 36 members of the Caryophyllales and 4 outgroups. Numbers above branches are bootstrap values. [ lnL score 210420.96]

PAGE 56

56 Figure 2 2. Phylogram of single most parsimonious tree based on the total evidence data set (2 nuclear genes, 8 plastid genes from the single -copy region, and the Inverted Repeat) for 36 members of the Caryophyllales and 4 outgroups. Numbers above branches are bootstrap values. [Length 27604, Consistency Index = 0.605, Retention Index = 0.538]

PAGE 57

57 F igure 2 3. Parsimony reco nstruction (illustrated on an MP tree) and stochastic character mapping (illustrated on Bayesian consensus tree): A) Reconstruction of perianth evolution ; B) P ollination syndromes

PAGE 58

58 Figure 2 4 The diverse forms of the perianth i n the core Caryophyllales. A Bougainvillea sp. (Nyctaginaceae); B Claytonia sp. ( Ron Wolf) (Portulacaceae s.l .); C Mesembryanthemum cordifolia (Aizoaceae); D Opuntia humifusa (Cactaceae); E Stegnosperma sp. ( Debra Valov) (Stegnospermataceae); F Sesuvium portulacastrum (Aizoaceae); G Hypertelis salsoloides (Molluginaceae); H Chenopodium sp. ( Brian Johnston) (Amaranthaceae); I Portulaca oleracea ( Kurt Neubig) (Portulacaceae); J Stellaria media ( Kurt Neubig) Caryophyllaceae; K P hytolacca americana ( Kurt Neubig) (Phytolaccaceae) ; L Mirabilis jalapa ( Walter Judd ) (Nyctaginaceae).

PAGE 59

59 Figure 2 5 Trees derived from analyses of the atpB data set (missing taxa: Gisekia and Alluaudia ).a) 1 of 23 MP trees; Numbers above branches are bootstrap support values; starred nodes collapse in strict consensus. [Length 1296, Consistency Index = 0.525, Retention Index = 0.566] b) ML tree resulting from GARLI analysis; Numbers above branches are bootstrap support values; starred nodes collapse in strict consensus. [ -lnL score 8854.1973]

PAGE 60

60 Figure 2 6. Trees derived from analyses of the matK data set (missing taxa: Physena ). a) 1 of 2 MP trees; numbers above branches are bootstrap support values. [Length 3360, Consistency Index = 0.519, Retent ion Index = 0.522] b) ML tree; numbers above branches are bootstrap support values. [ -lnL score 18173.934]

PAGE 61

61 Figure 2 7. Trees derived from analyses of the ndhF data set (missing taxa: Drosera, Drosophyllum, Simmondsia ). a)1 of 20 MP trees; numbers above branches are bootstrap support values; starred nodes collapse in strict consensus. [Length 3159, Consistency Index = 0.514, Retention Index = 0.493]. b) ML tree; numbers above branches are bootstrap support values. [ lnL score 17917.229]

PAGE 62

62 Figure 2 8. Tre es derived from analyses of the psbBTN data set. a) 1 of 15 MP trees; numbers above branches are bootstrap support values; starred nodes collapse in strict consensus. [Length 1639, Consistency Index = 0.494, Retention Index = 0.541]. b) ML tree; numbers ab ove branches are bootstrap support values. [ -lnL score 10692.37]

PAGE 63

63 Figure 2 9 Trees derived from analyses of the rbcL data set. a) 1 of 73 MP trees; numbers above branches are bootstrap support values; starred nodes collapse in strict consensus. [Length 1451, Consistency Index = 0.522, Retention Index = 0.601]. b) ML tree; numbers above branches are bootstrap support values; starred nodes collapse in strict consensus. [ lnL score 9770.5926]

PAGE 64

64 Figure 2 10. Trees derived from analyses of the rpoC2 data set ( missing taxa: Physena ). a) 1 of 3 MP trees; numbers above branches are bootstrap support values. [Length 4529, Consistency Index = 0.576, Retention Index = 0.567]. b) ML tree; numbers above branches are bootstrap support values. [ -lnL score 28139.564]

PAGE 65

65 Figure 2 11. Trees derived from analyses of the rps4 data set ( missing taxa: Drosera ). a) 1 of 122,155 MP trees; numbers above branches are bootstrap support values; starred nodes collapse in strict consensus [Length 545, Consistency Index = 0.646, Retent ion Index = 0.623]. b) ML tree; numbers above branches are bootstrap support values; starred nodes collapse in strict consensus [ lnL score 3826.098]

PAGE 66

66 Figure 2 12. Trees derived from analyses o f the IR data set (missing taxa: Limeum, Physena, Plumbago). a) Single MP tree; numbers above branches are bootstrap support values; starred nodes collapse in strict consensus. [Length 8649, Consistency Index = 0.781, Retention Index = 0.600]. b) ML tree; numbers above branches are bootstrap support values; starred nodes collapse in strict consensus. [ lnL score 86224.798]

PAGE 67

67 Figure 2 13. Trees derived from analyses of the combined plastid single -copy gene data set. a) Single MP tree; numbers above branches are bootstrap support values; starred nodes collapse in stri ct consensus. [Length 16158, Consistency Index = 0.530, Retention Index = 0.535] b) ML tree; numbers above branches are bootstrap support values; starred nodes collapse in strict consensus. [ lnL score 99154.419]

PAGE 68

68 Figure 2 14. Trees derived from analyses of the combined single copy plastid gene and nuclear gene data sets. a) Single MP; numbers above branches are bootstrap support values; starred nodes collapse in strict consensus. [Length 18939, Consistency Index = 0.524, Retention Index = 0.523]. b) ML T ree; numbers above branches are bootstrap support values; starred nodes collapse in strict consensus. [ lnL score 121955.45]

PAGE 69

69 Figure 2 15. Posterior probabilities of each rate category given each combination of E(T) and SD(T) and sampled from the prior with 10,000 reali zations.

PAGE 70

70 Table 2 1. Table showing for each data partition: aligned and analy sed length of data partitions; N o. of MP trees; Length of MP tree; Retention and Consistency Indices; ML score; length of ML tree under parsimony Data Partition T otal Aligned Length (bp) Analysed Aligned Length (bp) No. of MP Trees Length of MP Tree Consistency Index of MP Tree Retention Index of MP Tree ML Tree Score ML Tree Length under Parsimony atpB 1497 1497 23 1296 0.525 0.566 8854.1973 1302 matk 165 0 1650 2 3360 0.519 0.522 18173.934 3370 ndhF 2319 2182 20 3159 0.514 0.493 17917.229 3169 psbBTN 1780 1780 15 1639 0.494 0.541 10692.37 1644 rbcL 1449 1449 73 1451 0.522 0.601 9770.5926 1459 rpoC2 3903 3652 3 4529 0.576 0.567 28139.564 4533 rps4 609 609 122,155 545 0.646 0.623 3826.098 564 IR 29410 23966 1 8649 0.781 0.600 86224.798 8654 18S + 26S 5221 5221 14 2713 0.502 0.469 21309.042 2724 SC Plastid 13207 12819 1 16158 0.530 0.535 99154.419 16171 SC Plastid + Nuclear 18428 18040 1 18939 0.524 0 .523 121955.45 18962 SC Plastid + Nuclear +IR 47838 42,006 1 27604 0.605 0.538 210420.96 27613

PAGE 71

71 Table 2 2. Voucher, Citation, and GenBank Accession No. for rbcL Family Species Voucher Citation Embl Achatocarpaceae Phaulothamnus spinescens Manhart and Re ttig s.n. Rettig et al.1992 M97887 Aizoaceae Delosperma echinatum Chase 2539 K Savolainen et al. 2000 AJ235778 Amaranthaceae Celosia argentea Bot. Gard. Mainz (no voucher) Kadereit et al. 2003 AY270072 Amaranthaceae Spinacia oleracea not published Schmi tz Linneweber, 2001 AJ400848 Ancistrocladaceae Ancistrocladus korupensis Fay s.n Fay et al. 1997 Z97636 Asteropeiaceae Asteropeia micraster Civeyrel s.n K Soltis et al. 2000 AF206737 Basellaceae Basella alba Wilson s.n Rettig et al.1992 M62564 Berberi dopsidaceae Berberidopsis sp. Moore 326 this study Cactaceae Opuntia dillenii Greuter s.n. Edwards et al. 2005 AY875233 Cactaceae Pereskia aculeata Manhart and Rettig s.n. Rettig et al.1992 M97888 Caryophyllaceae Stellaria media Mort s.n WS Soltis et a l. 2000 AF206823 Didiereaceae Alluaudia procera Manhart and Rettig s.n. Rettig et al.1992 M62563 Dilleniaceae Hibbertia volubilis Hoot 9222 UWM Hoot et al. 1999 AF093721 Dilleniaceae Tetracera asiatica Chase 1238 K Savolainen et al. 2000 AJ235796 Dionc ophyllaceae Triphyophyllum peltatum Fay s.n Fay et al. 1997 Z97637 Droseraceae Drosera capensis Williams D1 LVC Albert et al. 1994 L01909 Drosophyllaceae Drosophyllum lusitanicum Williams D100 LVC Albert et al. 1994 L01907 Frankeniaceae Frankenia pul verulenta Collenette 6/93 K Savolainen et al. 2000 Z97638 Gisekiaceae Gisekia pharnacioides Manhart and Rettig s.n. Rettig et al.1992 M97890 Halophytaceae Halophytum ameghinoi Tortosa, Bartoli, and Chubut s.n. Savolainen et al. 2000 AJ403024 Molluginac eae Limeum aethiopicum not published unpublished AF132095 Molluginaceae Mollugo verticellata Hershkovitz 37 WS Soltis et al. 2000 M62566 Nepenthaceae Nepenthes alata M.W Chase 145 (NCU) Albert et al. 1994 L01936 Nyctaginaceae Bougainvillea glabra Chase 2485 K Savolainen et al. 2000 M88340 Nyctaginaceae Mirabilis jalapa Hershkovitz 60 WS Soltis et al. 2000 M62565 Physenaceae Physena madagascariensis Morton n.s Morton et al., 1997 Y13116 Phytolaccaceae Phytolacca americana Wells 4366 (1997) Kress et al. 2005 DQ006112 Phytolaccaceae Rivinia humulis Manhart and Rettig s.n. Rettig et al.1992 M62569 Plumbaginaceae Limonium gibertii not published Galmes et al. 2005 AJ786659 Plumbaginaceae Plumbago auriculata Voucher UCR Giannasi et al. 1992 M77701 Pol ygonaceae Polygonum sagittatum not published Kress et al. 2005 DQ006118 Portulacaceae Claytonia perfoliata not published unpublished AF132093 Portulacaceae Portulaca oleracea Applequist 7, unk, ISC Edwards et al. 2005 AY875249 Portulacaceae Talinum pa niculatum Edwards 6, Garden, YU Edwards et al. 2005 AY875214 Rhabdodendraceae Rhabdodendron amazonicum Fay s.n Fay et al. 1997 Z97649 Sarcobataceae Sarcobatus vermiculatus not published unpublished AF132088 Simmondsiaceae Simmondsia chinensis S. Boyd e t al. 3555 (F) Hoot et al. 1999 AF093732 Stegnospermataceae Stegnosperma halmifolium Miss.Bot. Gar Rettig et al.1992 M62571 Tamaricaceae Tamarix pentandra Fay s.n Fay et al. 1997 Z97650 Vitaceae Vitis vinifera not published Jansen et al., 2006 DQ424856

PAGE 72

72 Table 2 3. Voucher, Citation, and GenBank Accession No. for rpoC2 Family Species Vouncher Citation Embl Achatocarpaceae Phaulothamnus spinescens J. Panhart s.n this study Aizoaceae Delosperma napiforme Brockington S700 this study Amaranthaceae Celosia cristata Qiu 94153 IND this study Amaranthaceae Spinacia oleracea unpublished Schmitz Linneweber, 2001 AJ400848 Ancistrocladaceae Ancistrocladus korupensis Gereau et al. 5203 MO this study Asteropeiaceae Asteropeia micraster Civeyrel s.n. K this study Basellaceae Basella alba Qiu 02055 Univ this study Bereberidopsidaceae Berberidopsis corallina Moore 326 this study Cactaceae Opuntia microdasys D. Soltis s.n this study Cactaceae Pereskia aculeata D. Soltis 2645 this study C aryophyllaceae Stellaria media D. Soltis s.n this study Didiereaceae Alluaudia ascendens Qiu 97030 IND this study Dilleniaceae Hibbertia cuneiformis Qiu 97020 IND this study Dilleniaceae Tetracera asiatica Prance 30760 K this study Dioncophy llaceae Triphyophyllum peltatum Chase 663 K this study Droseraceae Drosera capensis D. Soltis s.n this study Drosophyllaceae Drosophyllum lusitanicum J. Cortez s.n, GBG this study Frankeniaceae Frankenia pulverulenta Collenette 6/93 K this stud y Gisekiaceae Gisekia africana this study Halophytaceae Halophytum ameghinoi Tortosa + Bartoli, Chubut, s.n this study Molluginaceae Limeum africanum Goldblatt et al. 11512 this study Molluginaceae Mollugo verticellata Moore 321 this study Nepenthaceae Nepenthes alata Botany Greenhouse, UF this study Nyctaginaceae Bougainvillea glabra Moore 323 this study Nyctaginaceae Mirabilis jalapa D.Soltis 2638 this study Physenaceae Physena madagascariensis Miller et al. 8817 MO this stud y Phytolaccaceae Phytolacca americana Brockington s.n this study Phytolaccaceae Rivinia humulis D.Soltis 2643 this study Plumbaginaceae Limonium arborescens Moore 318 this study Plumbaginaceae Plumbago auriculata Moore 306 this study Pol ygonaceae Polygonum virginicum D.Soltis 2656 this study Portulacaceae Claytonia virginica this study Portulacaceae Portulaca oleracea Moore 322 this study Portulacaceae Talinum paniculatum D.Soltis 2646 this study Rhabdodendraceae Rhabdodendron amazonicum E. Ribiro 1187 K this study Sarcobataceae Sarcobatus vermiculatus King and Garvey, 13892 MO this study Simmondsiaceae Simmondsia chinensis Qiu 96120 IND this study Stegnospermataceae Stegnosperma halmifolium Martin et al. s.n MO this study Tamaricaceae Tamarix pentandra Chase 252 NCU this study Vitaceae Vitis vinifera unpublished Jansen et al., 2006 DQ424856

PAGE 73

73 Table 2 4. Voucher, Citation, and GenBank Accession No. for atpB Family Species Voucher Citation Embl Achatoc arpaceae Phaulothamnus spinescens J. Panhart s.n This study Aizoaceae Delosperma echinatum Chase 2539 K Savolainen et al. 2000 AJ235452 Amaranthaceae Celosia argentea no voucher Soltis et al. 2000 AF209559 Amaranthaceae Spinacia oleracea not published Schmitz Linneweber, 2001 AF528861 Ancistrocladaceae Ancistrocladus korupensis not published not published AF209526 Asteropeiaceae Asteropeia micraster Civeyrel s.n K Soltis et al. 2000 AF209533 Basellaceae Basella alba Qiu 02055 this study Berber idopsidaceae Berberidopsis sp. Moore 326 this study Cactaceae Opuntia microdasys D. Soltis s.n this study Cactaceae Pereskia aculeata Soltis and Soltis, s.n Soltis et al. 2000 AF209648 Caryophyllaceae Stellaria media Mort s.n WS Soltis et al. 2000 AF209680 Didiereaceae Alluaudia ascendens Qiu 97030 IND this study Dilleniaceae Hibbertia volubilis Hoot 922 UWM Hoot et al. 1999 AF092120 Dilleniaceae Tetracera asiatica Chase 1238 K Savolainen et al. 2000 AJ235622 Dioncophyllaceae Triphyophyllum peltatum Chase 663K Soltis et al. 2000 AF209693 Droseraceae Drosera capensis K.Cameron 2134 NY Cameron et al. 2002 AY096110 Drosophyllaceae Drosophyllum lusitanicum J. Horn s.n Cameron et al. 2002 AY096113 Frankeniaceae Frankenia pulverulenta Collenette 6/93 K Savolainen et al. 2000 AJ235476 Gisekiaceae Gisekia africana this study Halophytaceae Halophytum ameghinoi Tortosa + Bartoli, Chubut, s.n this study Molluginaceae Limeum africanum Goldblatt et al. 11512 Klak et al. 2002 AJ532610 Mollu ginaceae Mollugo verticellata Hershkovitz 37 WS Soltis et al. 2000 AF209631 Nepenthaceae Nepenthes alata M.W Chase 145 NCU Albert et al. 1994 AJ235542 Nyctaginaceae Bougainvillea glabra Chase 2485 K Savolainen et al. 2000 AJ235415 Nyctaginaceae Mirabi lis jalapa Hershkovitz 60 WS Soltis et al. 2000 AJ532611 Physenaceae Physena madagascariensis Miller et al. 8817 MO this study Phytolaccaceae Phytolacca americana unpublished Kai et al., 2002 AF528855 Phytolaccaceae Rivinia humulis D.Soltis 2643 t his study Plumbaginaceae Limonium dendroides Chase 706 K Soltis et al. 2000 AF209620 Plumbaginaceae Plumbago zeylanica Chase 994 K Soltis et al. 2000 AJ235565 Polygonaceae Polygonum sachalinense Chase 896 K Soltis et al. 2000 AJ235569 Portulacaceae Claytonia virginica this study Portulacaceae Portulaca grandiflora Soltis s.n WS Soltis et al. 2000 AF209659 Portulacaceae Talinum paniculatum D.Soltis 2646 this study Rhabdodendraceae Rhabdodendron amazonicum Ribeiro 1187 K Savolainen et al. 20 00 AJ235578 Sarcobataceae Sarcobatus vermiculatus King and Garvey, 13892 MO this study Simmondsiaceae Simmondsia chinensis S. Boyd et al. 3555 (F) Hoot et al. 1999 AF093401 Stegnospermataceae Stegnosperma halmifolium Martin et al. s.n MO this study Tamaricaceae Tamarix pentandra Chase 262 NCU Soltis et al. 2000 AF209684 Vitaceae Vitis vinifera unpublished Jansen et al., 2006 DQ424856

PAGE 74

74 Table 2 5. Voucher, Citation, and GenBank Accession No. for psbBTN Family Species Voucher Citation Embl Ach atocarpaceae Phaulothamnus spinescens J. Panhart s.n this study Aizoaceae Delosperma napiforme Brockington S700 this study Amaranthaceae Celosia cristata Qiu 94153 IND this study Amaranthaceae Spinacia oleracea unpublished Schmitz Linneweber, 2001 AJ400848 Ancistrocladaceae Ancistrocladus korupensis Gereau et al. 5203 MO this study Asteropeiaceae Asteropeia micraster Civeyrel s.n., K this study Basellaceae Basella alba Qiu 02055 this study Bereberidopsidaceae Berberidopsis corallina Moo re 326 this study Cactaceae Opuntia microdasys D. Soltis s.n this study Cactaceae Pereskia aculeata D. Soltis 2645 this study Caryophyllaceae Stellaria media D. Soltis s.n this study Didiereaceae Alluaudia ascendens Qiu 97030 IND this study Dilleniaceae Hibbertia cuneiformis Qiu 97020 IND this study Dilleniaceae Tetracera asiatica Prance 30760 K this study Dioncophyllaceae Triphyophyllum peltatum Chase 663 K this study Droseraceae Drosera capensis D. Soltis s.n this study Dr osophyllaceae Drosophyllum lusitanicum J. Cortez s.n, GBG this study Frankeniaceae Frankenia pulverulenta Collenette 6/93 K this study Gisekiaceae Gisekia africana this study Halophytaceae Halophytum ameghinoi Tortosa + Bartoli, Chubut, s.n th is study Molluginaceae Limeum africanum Goldblatt et al. 11512 this study Molluginaceae Mollugo verticellata Moore 321 this study Nepenthaceae Nepenthes alata Botany Greenhouse, UF this study Nyctaginaceae Bougainvillea glabra Moore 323 this s tudy Nyctaginaceae Mirabilis jalapa D.Soltis 2638 this study Physenaceae Physena madagascariensis Miller et al. 8817 MO this study Phytolaccaceae Phytolacca americana Brockington s.n this study Phytolaccaceae Rivinia humulis D.Soltis 2643 t his study Plumbaginaceae Limonium arborescens Moore 318 this study Plumbaginaceae Plumbago auriculata Moore 306 this study Polygonaceae Polygonum virginicum D.Soltis 2656 this study Portulacaceae Claytonia virginica this study Portulac aceae Portulaca oleracea Moore 322 this study Portulacaceae Talinum paniculatum D.Soltis 2646 this study Rhabdodendraceae Rhabdodendron amazonicum E. Ribiro 1187 K this study Sarcobataceae Sarcobatus vermiculatus King and Garvey, 13892 MO this study Simmondsiaceae Simmondsia chinensis Qiu 96120 IND this study Stegnospermataceae Stegnosperma halmifolium Martin et al. s.n MO this study Tamaricaceae Tamarix pentandra Chase 252 NCU this study Vitaceae Vitis vinifera unpublished Jansen et al., 2006 DQ424856

PAGE 75

75 Table 2 6. Voucher, Citation, and GenBank Accession No. for rps4 Family Species Voucher Citation Embl Achatocarpaceae Phaulothamnus spinescens J. Panhart s.n this study Aizoaceae Delosperma napiforme Brockington S700 this stu dy Amaranthaceae Celosia cristata Qiu 94153 IND this study Amaranthaceae Spinacia oleracea unpublished Schmitz Linneweber, 2001 AJ400848 Ancistrocladaceae Ancistrocladus korupensis Gereau et al. 5203 MO this study Asteropeiaceae Asteropeia micraster Civeyrel s.n., K this study Basellaceae Basella alba Qiu 02055 this study Bereberidopsidaceae Berberidopsis corallina Moore 326 this study Cactaceae Opuntia microdasys D. Soltis s.n this study Cactaceae Pereskia aculeata D. Soltis 2645 t his study Caryophyllaceae Stellaria media D. Soltis s.n this study Didiereaceae Alluaudia ascendens Qiu 97030 IND this study Dilleniaceae Hibbertia cuneiformis Qiu 97020 IND this study Dilleniaceae Tetracera asiatica Prance 30760 K this stud y Dioncophyllaceae Triphyophyllum peltatum Chase 663 K this study Droseraceae Drosera capensis D. Soltis s.n this study Drosophyllaceae Drosophyllum lusitanicum J. Cortez s.n, GBG this study Frankeniaceae Frankenia pulverulenta Collenette 6/ 93 K this study Gisekiaceae Gisekia africana this study Halophytaceae Halophytum ameghinoi Tortosa + Bartoli, Chubut, s.n this study Molluginaceae Limeum africanum Goldblatt et al. 11512 this study Molluginaceae Mollugo verticellata Moore 321 this study Nepenthaceae Nepenthes alata Botany Greenhouse, UF this study Nyctaginaceae Bougainvillea glabra Moore 323 this study Nyctaginaceae Mirabilis jalapa D.Soltis 2638 this study Physenaceae Physena madagascariensis Miller et al. 881 7 MO this study Phytolaccaceae Phytolacca americana Brockington s.n this study Phytolaccaceae Rivinia humulis D.Soltis 2643 this study Plumbaginaceae Limonium arborescens Moore 318 this study Plumbaginaceae Plumbago auriculata Moore 306 thi s study Polygonaceae Polygonum virginicum D.Soltis 2656 this study Portulacaceae Claytonia virginica this study Portulacaceae Portulaca oleracea Moore 322 this study Portulacaceae Talinum paniculatum D.Soltis 2646 this study Rhabdodendra ceae Rhabdodendron amazonicum E. Ribiro 1187 K this study Sarcobataceae Sarcobatus vermiculatus King and Garvey, 13892 MO this study Simmondsiaceae Simmondsia chinensis Qiu 96120 IND this study Stegnospermataceae Stegnosperma halmifolium Martin et al. s.n MO this study Tamaricaceae Tamarix pentandra Chase 252 NCU this study Vitaceae Vitis vinifera unpublished Jansen et al., 2006 DQ424856

PAGE 76

76 Table 2 7. Voucher, Citation, and GenBank Accession No. for ndhF Family Species Voucher Citation Embl Achatocarpaceae Phaulothamnus spinescens J. Panhart s.n this study Aizoaceae Delosperma cooperi DQ855864. Amaranthaceae Celosia argentea AY959890 Amaranthaceae Spinacia oleracea not published Schmitz Linneweber, 2001 AY090621 Ancistro cladaceae Ancistrocladus korupensis Gereau et al. 5203 MO this study Asteropeiaceae Asteropeia micraster Civeyrel s.n., K this study Basellaceae Basella alba Qiu 02055 this study Berberidopsidaceae Berberidopsis corallina Moore 326 this study Cactaceae Opuntia microdasys D. Soltis s.n this study Cactaceae Pereskia aculeata D. Soltis 2645 this study Caryophyllaceae Stellaria media D. Soltis s.n this study Didiereaceae Alluaudia ascendens Qiu 97030 IND this study Dilleniaceae Hibbe rtia cuneiformis Qiu 97020 IND this study Dilleniaceae Tetracera asiatica AJ236277 Dioncophyllaceae Triphyophyllum peltatum Chase 663 K this study Droseraceae Drosera capensis D. Soltis s.n this study Drosophyllaceae Drosophyllum lusitani cum J. Cortez s.n, GBG this study Frankeniaceae Frankenia pulverulenta Collenette 6/93 K this study Gisekiaceae Gisekia africana this study Halophytaceae Halophytum ameghinoi Tortosa + Bartoli, Chubut, s.n this study Molluginaceae Limeum afr icanum Goldblatt et al. 11512 this study Molluginaceae Mollugo verticellata Moore 321 this study Nepenthaceae Nepenthes alata Botany Greenhouse, UF this study Nyctaginaceae Bougainvillea alba AF194825 Nyctaginaceae Mirabilis jalapa D.Soltis 2638 this study Physenaceae Physena madagascariensis Miller et al. 8817 MO this study Phytolaccaceae Phytolacca americana Brockington s.n this study Phytolaccaceae Rivinia humulis D.Soltis 2643 this study Plumbaginaceae Limonium arborescens Moore 318 this study Plumbaginaceae Plumbago zeylanica AJ236279 Polygonaceae Polygonum virginicum D.Soltis 2656 this study Portulacaceae Claytonia virginica AF194856 Portulacaceae Portulaca oleracea Moore 322 this study Portulacaceae Talinum paniculatum D.Soltis 2646 this study Rhabdodendraceae Rhabdodendron amazonicum E. Ribiro 1187 K this study Sarcobataceae Sarcobatus vermiculatus King and Garvey, 13892 MO this study Simmondsiaceae Simmondsia chinensis Qiu 96120 IND thi s study Stegnospermataceae Stegnosperma halmifolium Martin et al. s.n MO this study Tamaricaceae Tamarix pentandra Chase 252 NCU this study Vitaceae Vitis vinifera not published Jansen et al., 2006 DQ424856

PAGE 77

77 Table 2 8. Voucher, Citation, and GenBank Accession No. for 18S rDNA Family Species Voucher Citation Embl Achatocarpaceae Phaulothamnus spinescens Manhart and Rettig s.n. this study Aizoaceae Delosperma echinatum Chase 2539 K this study Amaranthaceae Celosia argentea Bot. Gard. Ma inz (no voucher) Kadereit et al. 2003 AF206883 Amaranthaceae Spinacia oleracea not published Schmitz Linneweber, 2001 L24420 Ancistrocladaceae Ancistrocladus korupensis Fay s.n Fay et al. 1997 AF206846 Asteropeiaceae Asteropeia micraster Civeyrel s.n K Soltis et al. 2000 AF206857 Basellaceae Basella alba Wilson s.n this study Berberidopsidaceae Berberidopsis corallina Moore 326 this study AF206866 Cactaceae Opuntia microdasys D. Soltis s.n this study Cactaceae Pereskia aculeata Manhart and Rett ig s.n. Rettig et al.1992 AF206986 Caryophyllaceae Stellaria media Mort s.n WS Soltis et al. 2000 AF207027 Didiereaceae Alluaudia procera Manhart and Rettig s.n. this study Dilleniaceae Hibbertia volubilis Hoot 9222 UWM Hoot et al. 1999 AF094542 Dilleniaceae Tetracera asiatica Chase 1238 K Savolainen et al. 2000 AJ235982 Dioncophyllaceae Triphyophyllum peltatum Fay s.n Fay et al. 1997 AF207049 Droseraceae Drosera capensis Williams D1 LVC Albert et al. 1994 U42532 Drosophyllaceae Drosophyllum lusitanicum Williams D100 LVC Albert et al. 1994 AB072556 Frankeniaceae Frankenia pulverulenta Collenette 6/93 K Savolainen et al. 2000 AF206914 Gisekiaceae Gisekia pharnacioides Manhart and Rettig s.n. this study Halophytaceae Halophytum ameghinoi Tortosa, Bartoli, and Chubut s.n. this study Molluginaceae Limeum aethiopicum not published unpublished AF094554 Molluginaceae Mollugo verticellata Hershkovitz 37 WS this study Nepenthaceae Nepenthes alata M.W Chase 145 (NCU) Albert et al. 1994 U4 2787 Nyctaginaceae Bougainvillea glabra Chase 2485 K Savolainen et al. 2000 AF206873 Nyctaginaceae Mirabilis jalapa Hershkovitz 60 WS this study Physenaceae Physena madagascariensis Morton n.s this study Phytolaccaceae Phytolacca americana Wel ls 4366 (1997) this study Phytolaccaceae Rivinia humulis Manhart and Rettig s.n. this study Plumbaginaceae Limonium gibertii not published Galmes et al. 2005 AF206953 Plumbaginaceae Plumbago auriculata Voucher UCR Giannasi et al. 1992 U42795 P olygonaceae Polygonum sagittatum not published this study Portulacaceae Claytonia perfoliata not published this study Portulacaceae Portulaca oleracea Applequist 7, unk, ISC this study Portulacaceae Talinum paniculatum Edwards 6, Garden, YU this study Rhabdodendraceae Rhabdodendron amazonicum Fay s.n Fay et al. 1997 AF207007 Sarcobataceae Sarcobatus vermiculatus not published this study Simmondsiaceae Simmondsia chinensis S. Boyd et al. 3555 (F) Hoot et al. 1999 AF094562 Stegnospermataceae Stegnosperma halmifolium Miss.Bot. Gar this study Tamaricaceae Tamarix pentandra Fay s.n Fay et al. 1997 AF207033 Vitaceae Vitis vinifera AF321271

PAGE 78

78 Table 2 9. Voucher, Citation, and GenBank Accession No. for 26S rDNA Family Species Vouche r Citation Embl Achatocarpaceae Phaulothamnus spinescens J. Panhart s.n this study Aizoaceae Delosperma napiforme Brockington S700 this study Amaranthaceae Celosia cristata Qiu 94153 IND this study Amaranthaceae Spinacia oleracea unpublished thi s study Ancistrocladaceae Ancistrocladus korupensis Gereau et al. 5203 MO this study Asteropeiaceae Asteropeia micraster Civeyrel s.n K Soltis et al. 2000 AF479090 Basellaceae Basella alba Qiu 02055 Berberidopsidaceae Berberidopsis sp. Moore 326 this study AF389242 Cactaceae Opuntia microdasys D. Soltis s.n this study Cactaceae Pereskia aculeata Manhart and Rettig s.n. Rettig et al.1992 AF479092 Caryophyllaceae Stellaria media Mort s.n WS Soltis et al. 2000 AF479084 Didiereaceae Alluaudi a ascendens Qiu 97030 IND this study Dilleniaceae Hibbertia cuneiformis Qiu 97020 IND this study Dilleniaceae Tetracera asiatica Chase 1238 K Savolainen et al. 2000 AF479097 Dioncophyllaceae Triphyophyllum peltatum Fay s.n Fay et al. 1997 AF479091 Droseraceae Drosera capensis Williams D1 LVC Albert et al. 1994 AF389248 Drosophyllaceae Drosophyllum lusitanicum J. Cortez s.n, GBG this study Frankeniaceae Frankenia pulverulenta Collenette 6/93 K this study Gisekiaceae Gisekia africana th is study Halophytaceae Halophytum ameghinoi Tortosa + Bartoli, Chubut, s.n this study Molluginaceae Limeum africanum Goldblatt et al. 11512 this study Molluginaceae Mollugo verticellata Moore 321 this study Nepenthaceae Nepenthes alata M.W Ch ase 145 (NCU) Albert et al. 1994 AF389260 Nyctaginaceae Bougainvillea glabra Moore 323 this study Nyctaginaceae Mirabilis jalapa D.Soltis 2638 this study Physenaceae Physena madagascariensis Miller et al. 8817 MO this study Phytolaccaceae Phyto lacca americana Brockington s.n this study Phytolaccaceae Rivinia humulis D.Soltis 2643 this study Plumbaginaceae Limonium arborescens Moore 318 this study Plumbaginaceae Plumbago auriculata Voucher UCR Giannasi et al. 1992 AF036492 Polygona ceae Polygonum sagittatum not published Kress et al. 2005 AF479085 Portulacaceae Claytonia virginica this study Portulacaceae Portulaca oleracea Applequist 7, unk, ISC Edwards et al. 2005 AF479093 Portulacaceae Talinum paniculatum D.Soltis 2646 th is study Rhabdodendraceae Rhabdodendron amazonicum E. Ribiro 1187 K this study Sarcobataceae Sarcobatus vermiculatus King and Garvey, 13892 MO this study Simmondsiaceae Simmondsia chinensis Qiu 96120 IND this study Stegnospermataceae Stegnosperma halmifolium Martin et al. s.n MO this study Tamaricaceae Tamarix pentandra Chase 252 NCU this study AF479083 Vitaceae Vitis vinifera unpublished this study AF479207

PAGE 79

79 Table 2 10. Voucher, Citation, and GenBank Accession No. for IR Family Species Voucher Citation Embl Achatocarpaceae Phaulothamnus spinescens J. Panhart s.n this study Aizoaceae Delosperma napiforme Brockington S700 this study Amaranthaceae Celosia cristata Qiu 94153 IND this study Amaranthaceae Spinacia oleracea unpublished Schmitz Linneweber, 2001 AJ400848 Ancistrocladaceae Ancistrocladus korupensis Gereau et al. 5203 MO this study Asteropeiaceae Asteropeia micraster Civeyrel s.n., K this study Basellaceae Basella alba Qiu 02055 this study Bereberidopsi daceae Berberidopsis corallina Moore 326 this study Cactaceae Opuntia microdasys D. Soltis s.n this study Cactaceae Pereskia aculeata D. Soltis 2645 this study Caryophyllaceae Stellaria media D. Soltis s.n this study Didiereaceae Alluaudia as cendens Qiu 97030 IND this study Dilleniaceae Hibbertia cuneiformis Qiu 97020 IND this study Dilleniaceae Tetracera asiatica Prance 30760 K this study Dioncophyllaceae Triphyophyllum peltatum Chase 663 K this study Droseraceae Drosera capensis D. Soltis s.n this study Drosophyllaceae Drosophyllum lusitanicum J. Cortez s.n, GBG this study Frankeniaceae Frankenia pulverulenta Collenette 6/93 K this study Gisekiaceae Gisekia africana this study Halophytaceae Halophytum ameghinoi Tortosa + Bartoli, Chubut, s.n this study Molluginaceae Limeum africanum Goldblatt et al. 11512 this study Molluginaceae Mollugo verticellata Moore 321 this study Nepenthaceae Nepenthes alata Botany Greenhouse, UF this study Nyctaginaceae Bou gainvillea glabra Moore 323 this study Nyctaginaceae Mirabilis jalapa D.Soltis 2638 this study Physenaceae Physena madagascariensis Miller et al. 8817 MO this study Phytolaccaceae Phytolacca americana Brockington s.n this study Phytolaccacea e Rivinia humulis D.Soltis 2643 this study Plumbaginaceae Limonium arborescens Moore 318 this study Plumbaginaceae Plumbago auriculata Moore 306 this study Polygonaceae Polygonum virginicum D.Soltis 2656 this study Portulacaceae Claytonia virginica this study Portulacaceae Portulaca oleracea Moore 322 this study Portulacaceae Talinum paniculatum D.Soltis 2646 this study Rhabdodendraceae Rhabdodendron amazonicum E. Ribiro 1187 K this study Sarcobataceae Sarcobatus vermiculatus King and Garvey, 13892 MO this study Simmondsiaceae Simmondsia chinensis Qiu 96120 IND this study Stegnospermataceae Stegnosperma halmifolium Martin et al. s.n MO this study Tamaricaceae Tamarix pentandra Chase 252 NCU this study Vitaceae Vi tis vinifera unpublished Jansen et al., 2006 DQ424856

PAGE 80

80 Table 2 11. Primers used for PCR and Sequencing rpoC2 Primers 26S Primers rpoC2.80F TAGATCACTTCGGAATGGC 26S.N.nc.S1 CGACCCCAGGTCAGGCG rpoC2.750F GGTCGTGTATTAGCRGACG 26S.N.nc.S3 AGGGAAGCGGATGGGGG C rpoC2.1000F GGTATTATTGCGGGTCAATC 26S.N.nc.S5 CGTGCAAATCGTTCGTCT rpoC2.1000R GATTGACCCGCAATAATACC 26S.N.nc.S7 GATGAGTAGGAGGGCGCG rpoC2.1400F AGGGRGARATGGCAYTGGAG 26S.N.nc.S9 AATGTAGGCAAGGGAAGT rpoC2.1400R CTCCARTGCCATYTCYCCCT 26S.N.nc.S11 AATCAGCGGGGAAAGAAG rpoC2.1950F GATCCTCGATACAGAAGAAAGAGTTC 26S.N.nc.S13 CCTATCATTGTGAAGCAG rpoC2.1950R GAACTCTTTCTTCTGTATCGAGGATC 26S.N.nc.S14 TTATGACTGAACGCCTCT rpoC2.2100F TCATTCCCGAGGAAGTRCAT 26S.268rev GCATTCCCAAACAACCCGAC rpoC2.2100R ATGYACTTCCTCGGGAATGA 26S.5 41rev GCATTCCCAAACAAGCCGAC rpoC2.2300F GTGGCATCTTGATACCRCCAG 26S.950rev GCTATCCTGAGGGAAACTTC rpoC2.2300R CTGGRGGTATCAAGATGCCAC 26S.1499rev ACCCATGTGCAAGTGCCGTT rpoC2.2600F CAAGTATTCAATTAGTTCGGACTTG 26S.1839rev TTCACCTTGGAGACCTGATG rpoC2.2600R CAAGTCCGA ACTAATTGAATACTTG 26S.2426rev MCTACACCTCTCAAGTCAT rpoC2.3000F CATCTAATTGTTTTCGAATGGGTCC 26S.2782rev GGTAACTTTTCTGACACCTC rpoC2.3000R GGACCCATTCGAAAACAATTAGATG 26S.3058rev TTCGCGCCACTGGCTTTTCA rpoC2.3400F CATTYAATTTGAATTGGTATTTTCTSC 26S.950rev GCTATCCTGAG GGAAACTTC rpoC2.3400R GGAGAAAATACCAATTCAAATTYAATG 26S.1499rev ACCCATGTGCAAGTGCCGTT rpoC2 3700R ATGYACTTCCTCGGGAATGA 26S.1839rev TTCACCTTGGAGACCTGATG rpoC2.4200R TTTCAGGCCTTTYARCCARTC 26S.2426rev MCTACACCTCTCAAGTCAT rps4 Primers 26S.2782rev GGTAACTTTTCT GACACCTC rps4.1F ATG TCR CGT TAC CGA GGG 26S.3058rev TTCGCGCCACTGGCTTTTCA rps4.8F GTT ACC GAG GRC CTC GTT TC 26S.3331rev ATCTCAGTGGATCGTGGCAG rps4.48F GGC TTT ACC RGG ACT AAC 18S Primers rps4.124F TCTCAATATCGTATTCGTYTAGAAG 18S.C18L CGACTTCTCCTTCCTCTC trnsR TACCGAGGGTTCGAATC 18S.NS1 GTAGTCATATGCTTGTCTC ndhF Primers 18S.42F GATTAAGCCATGCATGTGTAAGTATGAAC ndhF.1F ATGGAACAACATATAATATGC 18S.1706R CGGTGTGTACAAAGGGCAGGGACGTAGTC ndhF.73F CCCTTCATTCCRCTTCCAGTTCCT 18S.1740R CGCACCATTCAATCGGTAGGAGCGAC ndhF.95 F CTTTRTTAATAGGARYVGGACTTC atpB Primers ndhF.142F AATYTTCGDCGYATMTGGGCTT atpB.S2 TATGAGAATCAATCCTACTACTTCT ndhF.274F CTTACTTCTATTATGTCAATATTAAT atpB.S1494R CAGTACACAAAGATTTAAGGTCAT ndhF.536F TTG TAA CTA ATC GTG TAG GGG A rbcL Primers ndhF.536R TCCCCTAC ACGATTAGTTACAA rbcL.Z1 ATGTCACCACAAACAGAAACTAAAGCAAGT ndhF.803F CTATGGTAGCGGCGGGAATTTTTC rbcL.3' CTCGGAGCTCCTTTTAGTAAAAGATTGGGCCGA ndhF.803R GAAAAATTCCCGCCGCTACCATAG psbBTNH Primers ndhF.972F GTCTCAATTGGGTTATATGATG psbBNTH.60F ATGGGTTTGCCTTGGTATCGTGTTCA TAC ndhF.972R CATCATATAACCCAATTGAGAC psbBNTH.B60F CATACAGCTCTAGTTKCTGGTTGG ndhF.1006F CCCARTTGAGACATTGTMGAATARGC psbBNTH.66R CCAAAAGTRAACCAACCCCTTGGAC ndhF.1318F GGATTAACGCATTTTATATGTTTCG psbBNTH.B66R CCCCTTGGACTRCTACGAAAAAACC ndhF.1318R CGAAACATATAAAATGCGTTAATCC psbBNTH.65F TGCCTACTTTTTTTGAAACATTTCC ndhF.1603R GCATAGTATTGTCCGATTCATRAGG psbBNTH.71R CCCATMAAAGGAGTAGTYCCCC ndhF.2110R CCCCCTAYATATTTGATACCTTCTCC psbBNTH.B71R CCAGGAGCTACTTTACCATATTC ndhF.2153R GGAATTCCATCAATKATCGTYTATC

PAGE 81

81 Table 2 1 2. Coding of Perianth for Parsimony Reconstruction and Stochastic Character Mapping, with references used to code taxa Taxa Parsimony Character Mapping Stochastic Character Mapping Literature Cited Alluaudia 1 1 (Hofmann 1994) Ancistrocladus 3 3 (Porembski S 2003) Asteropeia 3 3 (Kubitzki K 2003) Basella 1 1 (Hofmann 1994) Berberidopsis 3 3 (Kubitzki K 2007) Bougainvillea 0 0 (Vanvinckenroye et al. 1993) Celosia 0 0 (Townsend 1993) Claytonia 1 1 (Hofmann 1994) Delosperma 0 0 (Hartmann 1993) Drosera 3 3 (Kubitzki 2003a) Drosophyllum 3 3 (Kubitzki 2003b) Frankenia 3 3 (Kubitzki 2003c) Gisekia 0 0 (Stevens 2001 onwards) Halophytum 0 0 (Pozner and Cocucci 2006) Hibbertia 3 3 (Horn 2007) Limeum 0/2 ? (Hofmann 1994) Limonium 3 3 (Kubitzki 1993) Mirabilis 1 1 (Bittrich and Kuhn 1993) Mollugo 0 0 (Endress and Bittrich 1993) Nepenthes 0 0 (Kubitzki 2003d) Opuntia 3 3 (Buxbaum 195055) Pereskia 3 3 (Buxbaum 195055) Phaulothamnus 0 0 (Bittrich 1993a) Physena 0 0 (Dickison 2003) Phytolacca 0 0 (Rohwer 1993) Plumbago 3 3 (Kubitzki 1993) Polygonum 0 0 (Brandbyge 1993) Portulaca 1 1 (Hofmann 1994) Rhabdodendron 3 3 (Prance 2003) Rivina 0 0 (Rohwer 1993) Sarcobatus 0 0 (Stevens 2001 onwards) Simmondsia 0 0 (Schmid 1978) Spinacea 0 0 (Townsend 1993) Stegnosperma 2 2 (Hofmann 1994) Stellaria 0/2 ? (Ronse De Craene et al. 1998) Talinum 1 1 (Vanvinckenroye and Smets 1996) Tamarix 3 3 (Gaskin 2003) Tetracera 3 3 (H orn 2007) Triphyophyllum 3 3 (Porembski and Barthlott 2003) Vitis 3 3 (Wen 2007)

PAGE 82

82 Table 2 13. Coding of Pollination for Parsimony Reconstruction and Stochastic Character Mapping, with references used to code taxa. Taxa Parsimony Character Map ping Stochastic Character Mapping Source of coding Information Alluaudia 1 1 (Kubitzki K 1993) Ancistrocladus ? ? (Porembski S 2003) Asteropeia 1 1 (Birkinshaw et al. 2004) Basella 1 1 (Sperling and Bittrich 1993) Berbe ridopsis ? ? (Kubitzki K 2007) Bougainvillea 1 1 (Bittrich and Kuhn 1993) Celosia 0/1 ? (Townsend 1993) Claytonia 1 1 (Carolin 1993) Delosperma 1 1 (Hartmann 1993) Drosera 1 1 (Kubitzki 2003a) Drosophy llum 1 1 (Kubitzki 2003b) Frankenia 1 1 (Kubitzki 2003c) Gisekia ? ? Halophytum 0 0 (Pozner and Cocucci 2006) Hibbertia 1 1 (Horn 2007) Limeum 1 1 (Endress and Bittrich 1993) Limonium 1 1 (Kubitzki 1993) Mirabilis 1 1 (Bittrich and Kuhn 1993) Mollugo 0/1 ? (Endress and Bittrich 1993) Nepenthes 1 1 (Kubitzki 2003d) Opuntia 1 1 (Barthlott and Hunt 1993) Pereskia 1 1 (Barthlott and Hu nt 1993) Phaulothamnus 0 0 (Bittrich 1993a) Physena ? ? (Dickison 2003) Phytolacca 1 1 (Rohwer 1993) Plumbago 1 1 (Kubitzki 1993) Polygonum 0/1 ? (Brandbyge 1993) Portulaca 1 1 (Carolin 1993) Rhabdodendron 1 1 (Prance 2003) Rivina 1 1 (Rohwer 1993) Sarcobatus ? ? Simmondsia 0 0 (Kohler 2003) Spinacea 0/1 ? (Kuhn 1993) Stegnosperma ? ? Stellaria 1 1 (Bittrich 1993b) Talinum 1 1 (Carolin 1993) Tamarix 0/1 ? (Gaskin 2003) Tetracera 1 1 (Horn 2007) Triphyophyllum ? ? (Porembski and Barthlott 2003) Vitis 1 1 (Wen 2007)

PAGE 83

83 Table 2 14. E stimations of the posterior probabili ty of ancestral peria nth character states at each node. Coding Strategy for Perianth: 0= Uniseriate Undifferentiated Perianth, 1 = Differentiated Perianth with involucral derived outer whorl 2 = Differentiated Perianth with stamen -derived petaloid organs 3 = Differentiated per ianth of uncertain affinity Name Freq Character 0 Character 1 Character 2 Character 3 Node A 1 0.5059 0.0006 0.0005 0.493 Node B 1 0.9956 0.0003 0.0003 0.0039 Node C 1 0.9936 0.0009 0.0009 0.0046 Node D 1 0.6279 0.0211 0.0209 0.3301 Node E 1 0.9981 0. 0005 0.0001 0.0012 Node F 1 0.9989 0.0004 0.0003 0.0004 Node G 1 0.9995 0.0002 0.0002 0.0002 Node H 1 0.9968 0.0012 0.0001 0.0019 Node I 1 0.9761 0.0217 0.0004 0.0018 Node J 0.622 0.9788 0.0209 0.0002 0.0001 Node K 1 0.9965 0.0015 0.0017 0.0003 Node L 1 0.9998 0 0.0001 0 Node M 1 1 0 0 0 Node N 0.9973 0.9996 0.0003 0 0 Node O 1 0.3924 0.6039 0.002 0.0017 Node P 1 0.0001 0.9999 0 0 Node Q 0.9884 0 1 0 0 Node R 1 0.0001 0.9982 0 0.0017 Node S 0.9994 0.0002 0.9998 0 0

PAGE 84

84 Table 2 15. E stim ations of the posterior probabili ty of ancestral pollination character states at each node. Coding Strategy for Pollination: Anemophilous = 0, Entomophilous = 1 Name Freq Character 0 Character 1 Node A 1 0.0007 0.9993 Node B 1 0.017 0.983 Node C 1 0.007 0.993 Node D 1 0.0148 0.9852 Node E 1 0.0023 0.9977 Node F 1 0.1052 0.8948 Node G 1 0.9781 0.0219 Node H 1 0.0002 0.9998 Node I 1 0 1 Node J 0.622 0 1 Node K 1 0.0002 0.9998 Node L 1 0.0004 0.9996 Node M 1 0.0001 0.9999 Node N 0.9973 0.0001 0.9999 Node O 1 0.0007 0.9993 Node P 1 0 1 Node Q 0.9884 0 1 Node R 1 0 1 Node S 0.9994 0.0002 0.9998

PAGE 85

85 CHAPTER 3 LABILE EVOLUTION OF PIGMENTATION IN THE CORE CARYOPHYLLALES Flavonoids, anthocyanins and carotenoids are the predominant pigments in flowe ring plants and play critical roles in floral and fruit coloration. However, in certain families within the core Caryophyllales (a major clade within the core eudicots) pigments known as betalains, which are structurally and biosynthetically distinct from flavonoids and anthocyanins, replace the latter in providing red to yellow pigmentation. The isolated occurrence of betalains in the core Caryophyllales has stimulated over half a century of speculation, debate and experimentation. Anthocyanins and betalai ns are mutually exclusive with two families of core Caryophyllales (Caryophyllaceae and Molluginaceae) producing anthocyanains and the remaining families producing betalains. The extensive phylogenetic analysis reported here demonstrates that the anthocya nic Molluginaceae are polyphyletic such that Hypertelis and Macarthuria should be excluded from Molluginaceae s.s (comprising nine genera, Adenogramma, Polpoda, Psammotropha, Coelanthum Pharnaceum Suessenguthiella, Glinus Glischrothamnus and Mollugo). As a result, reconstructions indicate that four (as opposed to two) anthocyanic lineages occur within the core Caryophyllales. Hypertelis arises toward the base of the raphide clade while Macarthuria constitutes a distinct lineage following divergence of Asteropeiaceae and Physenaceae. The ancestral pigment status of Molluginaceae s.s. becomes uncertain due to lack of published pigment data in genera of this now narrowly defined family. We explore the effect of this altered topology on inferences of pigme nt evolution in the core Caryophyllales. Data for Caryophyllaceae and Macarthuria imply that the early -diverging lineages of core Caryophyllales may have been anthocyanic although their pigment status is unknown, and analyses suggest a minimum of two reversals to anthocyanin pigmentation with the core Caryophyllales. Although the biochemical switch to betalains has traditionally been considered to have had only one origin,

PAGE 86

86 seven out of eight character -state reconstruction analyses suggest multiple origins of betalain pigmentation within the clade. Consequently, we discuss molecular and biochemical evidence that might support this finding. Introductio n Pigments perform critical roles in the biology of angiosperms, acting as visible signals to attract insec ts, birds and mammals for pollination and seed dispersal, and to protect plants from damage caused by UV or visible light. Flavonoids, especially anthocyanins, are widely distributed across angiosperms where they are responsible for the yellow to blue col oration found in flowers, leaves, fruit and seed (Tanaka et al. 2008). Betalains are yellow, orange to red pigments that are structurally and biosynthetically distinct from flavonoids and anthocyanins and in flowering plants are found only in certain famil ies within the core Caryophyllales (Bischoff 1876; reviewed in Mabry 1964), a clade of ~ 29 families and ~ 9000 species (APG II, 2003). In these betalain -producing families, a diversity of flavonoids, including the immediate anthocyanin precursors, l eucoa nthocyanidins have been detected and yet anthocyanins are not present (Bate Smith 1962; Bittrich and Amaral 1991); thus in the Caryophyllales, the production of betalains apparently substitutes for the otherwise ubiquitous anthocyanins. However two fami lies in the core Caryophyllales, Molluginaceae and Caryophyllaceae, produce only anthocyanins and not betalains, suggesting that the presence of anthocyanins or betalains is mutually exclusive (Mabry 1964). The novel nature of this chemosystematic characte r has enge ndered considerable debate (e.g. Cronquist and Thorne 1994). Some early taxonomic tr eatments, influenc ed by this mutual exclusion, placed the Molluginaceae and Caryophyllaceae together (Caryophyllinae) separated from the betalain producing linea ges (Chenopodiinae) (Mabry 1976; Cronquist and Thorne 1994)

PAGE 87

87 S ubsequent molecular phylogenetic analyses however, determined that Molluginaceae and Caryophyllaceae form disparate lineages within Caryophyllales (Rettig et al. 1992; Manhart and Rettig 1994) each allied to derived clades containing betalain producing families (Cunoud et al. 2002). Molluginaceae are sister to a large clade of betalain -producing succulent plant families including Cac taceae, while Caryophyllaceae are sister to both Ach atocarpa ceae, which apparently lack pigments (Clement et al. 1994), and Amaranthaceae, which produce betalains. Mo lluginaceae occupy a derived position within the core Caryophyllales follo wing the divergence of betalain -producing lineages (Cunoud et al. 2002) s uggesting the possibility of a t least one reversal from betalain pigmentation to anthocyanin pigmentation. Many evolutionary scenarios have been suggested to explain this complex distribution of pigment types, and yet hypotheses concerning the evolution o f pigmentation in the core Caryophyllales must be able to account for the following: 1) the unique presence of betalain pigmentation in the core Caryophyllales; 2) the intercalation of anthocyanin -pigmented taxa with betalain -pigmented lineages; and 3) the mutual exclusivity of the two pigment types. Although many of the hypotheses that have been advanced to date are not mutually exclusive, none of them alone easily explains these phenomena (as summarized in table 3 1 and discussed below). The U niq ue Origin of Betalains Ehrendorfer (1976) argued that the unique presence of betalains in the Caryophyllales was the consequence of an unusual evolutionary history, in which the ancestor to the Caryophyllales evolved in arid to semi arid conditions prior to the ra diation of major pollinator lineages. In this open, pollinator -deprived environment, wind pollination may have prevailed, and anthocyanic pigmentation was lost due to lack of need. Subsequently, following the radiation of pollinator lineages and colonizati on of less marginal habitats, reversion to zoophily engendered a return to pigmentation in the form of betalains rather than anthocyanins (Ehrendorfer 1976). This

PAGE 88

88 hypothesis is difficult to evaluate (Clement and Mabry 1996), and the concept of the Caryophy llales has changed considerably since Ehrendorfer (1976). Many morphological features of extant Caryophyllales are consistent with a wind -pollinated ancestor, including the predominant uniovular condition (Friedman and Barrett 2008) and an ancestrally unis eriate undifferentiated flower (i.e., a perianth of one organ type rather than differentiated sepals and petals). However several recently identified early -diverging lineages are probably entomophilous (Rhabdodendraceae and Asteropeiaceae) (Brockington et al. in press). Furthermore, anthocyanins and betalains accumulate in both vegetative and reproductive tissues, and their presence in vegetative organs is inconsistent with their evolutionary loss or gain due to absence or presence of pollinators alone. Th e Intercalation of Betalain and Anthocyanin L ineages In an attempt to explain the intercalation of betalain and anthocyanin lineages, Clement and Mabry (1996) suggested that the two classes of compounds might have co-occurred in an ancestor to the core Caryophyllales. The two pigments might have been selectively maintained in ancestors of extant Caryophyllales, with the subsequent loss of one or the other of the pigments in extant lineages. Clement and Mabry (1996) did not identify the nature of the select ion maintaining this co -occurrence, or its loss, but Cronquist (1977) proposed that betalains evolved due to their repellant (and fungicidal) properties (Mabry 1980; Piatelli 1981) rather than pigmentation and thus the two pigments could have co existed du e to complementary functions. The most significant objection to this hypothesis is that to date no extant taxa have been identified that possess both anthocyanins and betalains a surprising observation if they were complementary in function. It is import ant to note, however, that no pigment data exist for the four earliest -diverging lineages of core Caryophyllales ( Rhabdodendron, Simmondsia, Asteropeia and Physena ) as in general these are hard -to -obtain taxa from remote geographical

PAGE 89

89 areas. It remains poss ible that the two pigment types coexist in these unexplored taxa however the maintenance of both pigments in an early ancestor does not provide mechanisms for their ultimate mutual exclusion. Clement and Mabry (1996) referred to the stochastic loss of one or the other pigment in extant lineages, and yet the use of the term stochastic circumvents specific mechanisms to explain loss of a pigment. Mutual E x clusion of Anthocyanins and B etalains Additional hypotheses are needed to explain the mutual exclus ion of these pigment types and/or the replacement of one pigment type with another. Clement and Mabry (1996) explored the relative metabolic cost of betalain and anthocyanin synthesis but reached no clear conclusions in part because the full physiological roles of anthocyanins and betalains are not thoroughly understood and in part because true metabolic costs are not easily assessed. The relative effectiveness of the pigments have also been discussed in terms of molar absorptivity, with the suggestion tha t much smaller amounts of betacyanin are needed to absorb equivalent amounts of visible light relative to their anthocyanic counterparts, and betacyanins are thus more cost -effective(Clement and Mabry 1996). This precludes the significant absorption of U V wavelengths by anthocyanins. The recent discovery that yellow betaxanthins are fluorescent (Gandia -Herrero et al. 2005) and the increasingly recognized importance of fluorescence for biological signaling (Arnold et al. 2002; Mazel et al. 2004) highlight a further asset that betalains may possess over anthocyanins and complicates analyses of the cost effectiveness of pigments. Biosynthetic Inhibition Stafford (1994), taking a different line of reasoning, proposed that the remarkably similar patterns in accumulation of betalains and anthocyanins in both vegetative and floral tissues are indicative of a common regulatory system. In this situation, should the ultimate step in anthocyanin synthesis be in some way inhibited by an endproduct of betalain synt hesis

PAGE 90

90 (Stafford, 1994), then anthocyanin accumulation might be repressed and replaced entirely by betalain pigmentation. To invoke repression of anthocyanin pigmentation by betalain synthesis is problematic, however, when explaining the occurrence of anthocyanins in lineages diverging after the inferred origin of betalain pigmentation. Confidence in the phylogeny of the core Caryophyllales is fundamental to consideration of this complex, and sometimes contradictory, set of hypotheses. The anthocyanic Mollug inaceae have been thought to be polyphyletic with the doubtful inclusion of Polpoda (due to lack of sympodial growth; Endress and Bittrich 1993) and Macarthuria (due to the angular inclusion in P plastids; Cunoud et al. 2002) Resolution of this putative polyphyly might be expected to clarify patterns of pigment evolution. To el ucidate the full extent of the anthocyaninproducing lineages and their phylogenetic pattern in relation to betalain -producing lineages, we perf ormed a phylogenetic analysis on gene ra within the Molluginaceae using multiple genetic markers. No attempt has been made to rigorously evaluate hypotheses of pigment evolution since the advent of DNA -based phylogenies for the core Caryophyllales (Clement and Mabry 1996) other than simple rec onstructions in Soltis et al. (2005) and Cunoud (2006). Furthermore, the significance of apparently pigment -less lineages (lacking anthocyanins and betalains) for understanding pigment evolution has been largely ignored (with the exception of Cunoud et al. 2002), despite their frequent occurrence (e.g., Achatocarpaceae, Limeum and members of Molluginaceae, according to Clement and Mabry, 1996). In this study we: (1) conduct a phylogenetic analysis of all genera within Molluginaceae in the context of th e core Caryophyllaes, (2) explore concepts of pigment evolution in light of a revised molecular phylogeny using a variety of character -state reconstruction methods and character -coding strategies and (3) assess the implications of a polyphyletic Molluginac eae for inferring pigment evolution.

PAGE 91

91 Materials And Methods Taxon Sampling and Data A ssembly All 11 genera currently assigned to Molluginaceae were sampled (presented tables 3 4 and 3 5). The p lastid genes matK and rbcL were targeted for amplification. Am plification and sequencing protocols are previously described (Brockington et al. 2008). S equences generated in this study were combined with previously published sequences f or 35 additional genera representing all major lineages within the core Caryophyll ales and four outgroups ( Hibbertia Tetracera Vitis and Berberidopsis ). In some instances sequence data were combined from multiple species to represent a family; in 10 of 11 cases the species that were combined are congeneric. This approach was judged not to significantly affect the primarily family -level analysis performed in this study, but the instances are listed here: Aizoaceae ( Delosperma napiforme and Delosperma echinatum ); Amaranthaceae ( Celosia argentea and Celosia trigyna ); Cactaceae ( Opuntia quimilo and Opuntia stricta ); Didiereaeae ( Alluaudia ascendens and Alluaudia procera); Dilleniaceae ( Hibbertia volubilis and Dillenia indica ); Gisekiaceae ( Gisekia africana and Gisekia pharnacioides ); Molluginaceae ( Limeum africanum and Limeum aetheopicum ); Plumbaginaceae ( Limonium gibertii and Limonium latifolium ); Polygonaceae (Polygonum sagittatum and Polygonum cespitosum ); Portulacaceae ( Claytonia magarhiza and Claytonia perfoliata ); Tamaricaceae ( Tamarix pentandra and Tamarix canariensis ). Phylogenetic Analyses Sequences were automatically aligned using Clustal X (Thompson et al., 1997) and then manually adjusted. Coding regions were aligned by predicted amino acid sequence. Regions at the beginning s and ends of genes for which sequences were incomple te, together with regions that were difficult to align, were excluded from the analysis. The matK and rbcL data sets were combined into a single data partition. This partition was subject to phylogenetic analyses using

PAGE 92

92 three optimality criteria: m aximum pa rsimony (MP), maximum likelihood (ML) and Bayesian posterior probability MP analyses were implemented in PAUP*4.0 (Swofford 2000) Shortest trees were obtai ned using a heuristic search with 1 000 replicates of random taxon addition with tree bisection -re connection (TBR) branch swapping, saving all sh ortest trees per replicate. Bootstrap support (BS) for relationships was estimated from 1 000 bootstrap replicates using 10 random taxon additions per replicate, with TBR branch swapping saving all trees Fo r ML analyses we employed GARLI (Genetic Algorithm for Rapid Likeli hood Inference; version 0.942, Zwickl, 2006), which performs maximum likelihood heuristic phyloge netic searches under the GTR model of nucleotide substitution, in addition to models that in corporate among -site rate variation, either assuming a gamma distribution ( ) or a proportion of invariable sites (I), or both. Models of nucleotide substitution were determined using MrModeltest (Nylander 2004). The Akaike information criterion (AIC) was used to select GTR+I+G as an appropriate model based on the relative informational distance between the ranked models. Analyses were run with default options, except that the significanttopochange parameter was reduced to 0.01 to make searches more stri ngent. Five replicate analyses were performed in GARLI ML bootstrap analyses were conducted in GARLI with the default parameters and 100 replicates. Bayesian a nalyses were implem ented in MrBayes, version 3.1.2 (Huelsenbeck and Ronquist 2001; Ronquist a nd Huelsenbeck 2003). Two independent analyses each ran for 5 million generations, using four Markov chains, and with all other parameters at default values; trees were sampled every thousandth generation, with a burnin of 20,000 generations. Stationarity of the Markov chain Monte Carlo was determined by the average standard deviation

PAGE 93

93 of split frequencies between runs (after 5 million generations the average standard deviation was 0.004%) and by examination of the posterior in Tracer, version 1.3 (Rambaut and Drummond 2003) All trees from the two runs were combined in a majority rule consensus tree generated in PAU P*4.0 (Swofford 2000), using the posterior distribution of the trees. Character Coding Pigment type was coded as a multistate character: betal ain pigmented, anthocyanin pigmented, or pigment less. Due to the apparent mutual exclusion of these two types of pigments no terminals were coded as equivocal; where data w ere unavailable or uncertain, taxa were coded as unknown. Coding matrices and sources of information on coding are given in table 3 6. To allow for different hypotheses of pigment evolution, two alternative coding strategies were employed: unordered and ordered. The unordered strategy permitted direct transitions between all chara cter states allowing direct transitions to occur between anthocyanin and betalain pigmentation. The ordered strategy allowed transitions between betalain and anthocyanin pigmentation to occur only thro ugh an intervening unpigmented state. These two codi ng strategies represent alternative traditional hypotheses concerning the evolutionary history and biology of pigmentation within the core Caryophyllales. An ordered coding strategy in which betalains and anthocyanins can arise only via a pigment less cond ition is consistent with the hypothesis of Ehrendorfer (1976). The unordered coding strategy, in which transitions directly occur between betalains and anthocyanins, is more consistent with hypotheses suggesting selective advantage of betalains (Clement an d Mabry 1996) or regulatory inhibition of the anthocyanin pathway by betalains (Stafford, 1994). Character -state Reconstruction Parsimonybased character -state reconstructions on both ordered and unordered characters were performed in Mesquite (Maddison a nd Maddison 2008) using the ML tree topology with

PAGE 94

94 the highest likelihood score (Figure 3 2). The ML topology was chosen to best represent the phylogeny of the Caryophyllales because of the high degree of rate heterogeneity detected among lineages (see belo w). Characters were mapped under ACCTRAN and DELTRAN resolving options, and the all most parsimonious states option. The number of state -to state transformations is listed in table 3 2. Stochastic Mapping Changes in pigmentation were further modeled by me ans of stochastic mapping techniques a s described by Huelsenbeck et al. (2003) and implemented in SIMMAP (Bollback 2006) This approach estimates the rates at which a discrete character undergoes state changes as it evolves through time. Bayesian estimatio n has some advantages over tra ditional parsimony -based reconstruction (Huelsenbeck et al. 2003). First it allows one to average over equally likely topologies, which is valuable as the position of Hypertelis within the raphide clade is unresolved or poo rly supported and the position of Macarthuria is weakly supported in some analyses. Second, it allows more than one character change per branch and is therefore a useful methodology for character -state reconstruction in the Caryophyllales a clade in whic h long branches are common. Prior Specification Posterior mapping requires the specification of prior values. The prior on the bias parameter was fixed at 1/ k where k is the number of states (this being the recommended approach in SIMMAP (Bollback 2006) for characters of more than two states ; Renner et al. 2007); the bias parameter is therefore 0.333 (3 states: betalain pigmented, anthocyanin pigmented and unpigmented). We applied an empirical Bayesian approach in choosing appropriate priors for the subs titution rate parameters (Couvreur et al. 2008a, 2008b) The gamma distribution of the substitution rate is governed by two hyp erparameters defining the

PAGE 95

95 mean E(T) and the standard deviat ion SD(T). The values of these hyperparameters for the prior gamma dis tribution were selected i ndependently for each character -coding strategy (ordered vs unordered) by using the number of realizations sampled from priors function in SIMMAP with 10,000 draws. A series of trials was performe d (10,000 realizations in each) that systematically sampled for values of E(T) between 1 and 30, in combination w ith SD(T) values of either 1 or 5. The posterior dis tribution of these combinations was visualized in Tracer v 1.3 and further plotted as graphs of frequency against r ate (Fi gures 3 4 and 3 5 ). The posterior distribution curves derived from these trials allowed the select ion of values of E(T) that gave highest sampling and allowed optimization of the E(T) value (Couvreur et al., 2008a, 2008b) A trial was also performed without specifying priors and allowing rates to be determined by branch lengths (as performed by Renner et al. 2007); however, the posterior distribution curves were generally highly skewed, and this form of prior selection was not employed in subsequent analys es. For the unordered coding strategy, the chosen prior mean rat e E(T) was 14 (Figure 3 4 ), and for the ordered coding strategy, the chosen prior mea n rate E(T) was 25 (Figure 3 5 ). For both of these values, an SD(T) value of 5 was ap plied in subsequent an alyses, allowing a large standard deviation to accommodate uncertainty in the mean rate of character substitution e (Couvreur et al., 2008a, 2008b) Rates of Transformation and Ancestral State R econstruction Following specification of priors, the rate an d number of state transformations were estimated by 100 realizations on the 4800 post burnin trees (with branch lengths) from the Bayesian analyses. As recommended, branch lengths were rescaled so that the total tree length was 1 but the branch length proportions were maintained. The total number of transformations and average number of state to -state transformations were recorded for each cod ing strategy (recorded in tables 3 2, and 3 3 ). The ancestral state at different nodes was assessed using a

PAGE 96

96 hierarc hical Bayesian ancestral state reconstruction method implemented in the posterior ancestral states function of SIMMAP. The nodes for which ancestral states were estimated are labeled in Figure 3 3. The estimations of the posterior pro bability of ancestra l character states for the ordered and unordered evolution of pigment s are listed in table 3 7 and present ed graphically on the nodes in Figure 3 3 The effect of missing data (i.e., unknown pigmentation character states at terminals) on the stochastic map ping analyses was examined by pruning unknown taxa from the Bayesian-derived trees. Analyses were then re -run on these pruned trees to assess the effect of pruning unknown taxa on estimations of state -to -state transformations. Results Phylogenetic Relatio nships Maximum parsimony analysis of the combined matK and rbcL data set yielded 12 trees. Nine of the Molluginacaeae genera ( Glinus Glischrothamnus Mollugo, Adenogramma, Polpoda Psammotropha, Suessenguthiella, Coelanthum and Pharnaceum ) constitute a s trongly supported monophyletic clade (BS = 98%), hereafter, Molluginaceae sensu stricto s.s. sister to the portulacaceous cohort (Figure 3 1). These nine genera form three subclades: the clade containing Glinus Glischrothamnus and Mollugo is sister to a clade comprising (1) Adenogramma, Polpoda and Psammotropha, and (2) a clade comprising Suessenguthiella, Coelanthum and Pharnaceum The strict consensus tree derived from the replicates of the ML analysis and the consensus tree derived from the Bayesian analyses are almost identical to the MP strict consensus tree with respect to the placement of the genera within Molluginaceae s.s The sole exc eption is the relative position of Adenogramma, Polpoda and Psammotropha ; however internal relationships within this clade are unresolved i n the strict MP consensus tree and unsupported in the GARLI bootstrap analysis (Figure 31) Two of the sampled taxa, Hypertelis and Macarthuria, are not placed in monophyly with the rest of the Molluginaceae.

PAGE 97

97 Macarthuria is res olved as sister to the rest of the core Caryophyllales following the divergence of Rhabdodendron, Simmondsia, Asteropeia and Physena Hypertelis is placed as sister to members of the raphide clade following the divergence of Corbichonia (Figure 3 1 and 3 2). The placements of Hypertelis and Macarthuria are identical in both the MP and ML trees. Character Mapping with Unordered Coding Using parsimony, a nthocyanin pigmentation is recovered for the most basal nodes in the core Caryophyllales ( Figure 3 3 an d Figure 6 3 ). H owever following the divergence of Macarthuria, reconstruction of pigment evolution varies depending on whether ACCTRAN or DELTRAN optimization is used. With the ACCTRAN option, following the divergence of Macarthuria, node G (Figure 3 3 a nd Figure 6 3) is reconstructed as betalain pigmented; thus a single origin of betalain pigmentation is inferred, with three subsequent reversals to anthocyanin pigmentation in Caryophyllaceae, Hypertelis and Molluginaceae s.s. Furthermore, three instanc es of pigmentation loss are inferred: in Achatocarpaceae, Limeum and, via anthocyanin pigmentation, the subclade containing Adenogramma, Psammotropha and Polpoda within Molluginaceae s.s. With the DELTRAN option, nod e G is recovered as anthocyaninpigmente d (Figure 3 3 and Figure 6 3), with two origins of betalain pigmentation, one following the divergence of the clade comprising Caryophyllaceae, Achatocarpaceae and Amaranthaceae (the CAA clade) and one within this same clade, in the Amaranthaceae. Follow ing the origin of betalain pigmentation, there are two subsequent reversals to anthocyanin pigmentation (rather than the three under ACCTRAN) : in Hypertelis and Molluginaceae s.s. The same instances of pigmentation loss are inferred under DELTRAN as in ACC TRAN. Using the stochastic mapping approach, the total mean number of pigment transformations using unordered coding was 15.60 (see table 32). The mean number of transitions to betalain pigmentation was 3.43 (anthocyanin to betalain = 2.1147, non-pigmente d to betalain = 1.3162).

PAGE 98

98 Pruning of earlydiverging taxa of unknown pigment status reduced the mean number of transitions to betalain pigmentation to 3.23. Pruning all taxa of unknown pigment status reduced the mean number of transitions to betalain pigmen tation to 2.39. Transitions between other pigment states are shown in table 3 2. The reconstruction of ancestral states derived from the stochastic mapping approach is shown in Figure 3 3b). Character Mapping with Ordered Coding Using parsimony reconstru ction with ordered coding, anthocyanin pigmentation is again recovered for the most basal nodes in the core Caryophyllales (nodes C, D, E, F) but again, following the divergence of Macarthuria, reconstruction of pigment evolution varies dramatically depen ding on whether ACCTRAN or DELTRAN optimization is used. With ACCTRAN optimization, node G (Figure 3 3 and Figure 3 6) is inferred to be ancestrally lacking in pigment with one reversal to betalain pigmentation in Amaranthaceae and a reversal to anthocyanin pigmentation in Caryophyllaceae. Following the divergence of the CAA clade, an origin of betalain pigmentation is inferred on the branch leading to the rest of the core Caryophyllales with two reversals to anthocyanin pigmentation in Hypertelis and M olluginaceae s.s Complete loss of pigmentation is inferred to have occurred three times, once following the divergence of Macarthuria, once in Limeum and again in a subclade of Molluginaceae s.s. (Adenogramma). The DELTRAN optimization differs considerabl y from ACCTRAN in its reconstruction of ancestral states (Figure 3 6) T he basalmost nodes and the node giving rise to the CAA clade are reconstructed as anthocyanic. Within the CAA clade, pigment is inferred as lost on the branch leading to Achatoca r pa ceae and Amaranthaceae but subsequently regained in Amaranthaceae. Following the divergence of the CAA clade pigmentation is inferred as lost on the branch leading to the rest of the core Caryophyllales. From this unpigmented condition four separate or igins of betalain pigmentation and three o rigins of anthocyanin pigmentation are

PAGE 99

99 inferred The instances of derived betalain pigmentation are inferred in Stegnospermataceae, Corbichonia, the raphide clade following divergence of Hypertelis and on the br anch leading to the portulacaceous clade. The three instances of anthocyanin pigmentation are inferred to have occur red in Mollugo, Pharnaceum and Hypertelis Using the stochastic mapping approach the total mean number of pigment transformations using ordered coding was 24.74 (table 3 3). The mean number of transitions to betalain pigmentation was 4.76. Pruning of early-diverging taxa of unknown pigment status reduced the mean number of transitions to betalain pigmentation to 3.7. Pruning of all taxa of unknown pigment status reduced the mean number of transitions to betalain pigmentation to 3.2. Transitions between other pigment states are shown in table 3 3. The reconstruction of ancestral states derived from the stochastic mapping approach is shown i n Figure 3 3d). Discussion Polyphyletic Molluginaceae Our analyses demonstrate that the anthocyanic lineage Molluginaceae is polyphyletic such that two genera originally assigned to the Molluginaceae ( Hypertelis and Macarthuria) now form distinct lineages (Figure 3 1 and 32). Limeum and Corbichonia also previously belonging to the Molluginaceae have previously been shown to form distinct lineages. Molluginaceae s.s are re defined as comprising nine genera: Adenogramma, Polpoda, Psammotropha Coelanthum Pharnaceum Suessenguthiella, Glinus Glischrothamnus and Mollugo. The anthocyanic genus Hypertelis is placed as sister to all of the raphide clade except Corbichonia and the anthocyanic Macarthuria is assigned to a grade of early-diverging core Caryophyllales. Given this polyphyly, the number of anthocyanic lineages within the core Caryophyllales has increased from two to four, with each major betalain pigmented clade (the portulacaceous clade, raphide clade and CAA clade) subtended completely or i n part by an anthocyanin -pigmented lineage (Figure 3 1,

PAGE 100

100 3 2 and 3 3). Furthermore, the anthocyanic genus Macarthuria is placed prior to the divergence of any known betalain-pigmented taxa. The presence of an angular protein inclusion in the P plastid of Ma carthuria (Behnke 1994) is consistent with its position outside of the globular inclusion clade (Brockington et al. in press; Cunoud et al. 2002). We conducted several types of character -state reconstruction analyses to evaluate the effect of this new topology on concepts of pigment evolution within the Caryophyllales. These analyses do not provide a consistent impression of the evolution of pigmentation. While we recognize that these character -state reconstructions cannot solve the complexity of pigmen t evolution in the core Caryophyllales, the phylogenetic topology provided here, together with these analyses, highlights some key questions and uncertainties. Here we discuss the salient similarities and differences in our reconstruction analyses with the aim of defining themes and questions for future research. Early-Diverging Anthocyanic Lineages Caryophyllaceae, represented in our analyses by Stellaria are the most species rich and best -studied of the anthocyanic lineages within the core Caryophyllales ; with 10 genera analyzed, only anthocyanins have been detected (Clement et al. 1994). Due to the early divergence of Caryophyllaceae, within the core Caryophyllales (Figure 3 1 and 3 2), it is uncertain whether the presence of anthocyanins is a derived st ate or the plesiomorphic condition. At the level of ancestral character -state reconstruction, distinguishing between these two alternatives depends on the pigment status of the early -diverging grade, which, with the exception of Macarthuria, is unknown. In parsimony reconstruction, the early -diverging lineages are always recovered as anthocyanic (Figure 3 4). However, this result is influenced by the position of Caryophyllaceae and the position of the anthocyanic Macarthuria, identified by this study as a member of the early -diverging grade. Thus, although the pigment status of the other

PAGE 101

101 early diverging lineages is unknown, the position of anthocyanic Macarthuria, together with the parsimony reconstruction analyses, suggests they could also be anthocyanic. The stochastic mapping analyses, however are less supportive for the anthocyanic condition of the unknown early diverging lineages (Figure 3 3). If the occurrence of anthocyanins is plesiomorphic within Caryophyllales, then the presence of anthocyanins in Caryophyllaceae could represent a retention of this plesiomorphic condition (parsimony; DELTRAN; Figure 3 6) rather than a derivation from a pigment less or betalain condition (ACCTRAN; Figure 3 6: stochastic mapping analyses; Figure 3 3). Clearly, determining the pigment status of recently identified earlydiverging lineages ( Rhabdodendron, Simmondsia, Physena and Asteropeia ) is crucial for clarifying the early evolution of pigmentation, yet the geographical isolation of these often rare endemics makes th is challenging. Derived Anthocyanic Lineages Molluginaceae s.s occupies a more derived position than the Caryophyllaceae, and consequently in all reconstruction analyses, the origin of anthocyanins within this clade is invariably inferred as derived, eit her from a lack of pigments or from betalains. Although Molluginaceae are widely cited as an anthocyanic family, this claim is based on limited data, partly as a result of the restricted circumscription presented in this study. Of the nine genera that now constitute Molluginaceae as defined in this study only three have been examined for the presence of pigments: Adenogramma, Pharnaceum and Mollugo. Of these, presence of anthocyanin in Mollugo was reported with no reference to experimental data (Mears 1976 ), and anthocyanins in Pharnaceum have been reported only as unpublished data (Clement et al. 1994). Due to the resolved polyphyly presented in this study, there are currently no genera within Molluginaceae that have published data to support an anthocyani c status for the family. The absence of pigments reported in Adenogramma might be due to the fact that only dried stems

PAGE 102

102 were analyzed ( Cunoud et al. 2002). Presence of pigments in vegetative tissue is more likely to be dependent on environmental conditions, unlike floral tissue, which tends to be constitutively pigmented (Stafford 1994). Clement et al. (1994) reported that most genera in Molluginaceae lack detectable pigments (again without reference to experimental data); if this is accurate, pigments m ight have been secondarily lost in an ancestor to Molluginaceae with the presence of anthocyanins in Pharnaceum and Mollugo a further derivation (DELTRAN; Figure 3 6 and stochastic mapping; ordered coding; Figure 3 6). The isolated status of anthocyanic H ypertelis towards the base of the raphide clade is surprising from the point of view of pigment evolution. Although Hypertelis was determined to be anthocyanic (Beck et al. 1962), Clement et al. (1994) cast doubt on this finding (but later cite Hypertelis as anthocyanic in Clement and Mabry, 1996). In light of its current phylogenetic position, the presence of anthocyanins in Hypertelis should be re -evaluated. Evidence for Reversals to Anthocyanin P igmentation Due to the derived position of Molluginaceae a nd Hypertelis all reconstruction analyses suggest reversals to anthocyanins within the core Caryophyllales (tables 3 1 and 3 2, Figure 3 3 and Figure 6 3), however, historically, reversals to anthocyanins have not been discussed. Elements of the anthocyan in synthesis pathway must be conserved in betalainpigmented plants due to the presence and accumulation of various precursors such as flavonols, leucoanthocyanidins and proanthocyanidins (Clement and Mabry 1996). This inference has been confirmed by recen t studies demonstrating the presence of genes encoding enzymes of the anthocyanin synthetic pathway within betalain-pigmented taxa (Shimada et al. 2005; 2007). Interestingly, the betalain pigmented Phytolacca americana and Spinacia oleracea possess fully f unctioning enzymes for the penultimate stages of the anthocyanin pathway d ihydrof l avonol reductase (DFR) and anthocyanidin synthase (ANS) (Shimada et al. 2005). The genes encoding

PAGE 103

103 DFR and ANS are transcribed in the seeds in correlation with production of proanthocyanidins and yet are not transcribed in other parts of the plant. It is probable, therefore, that the absence of anthocyanic pigments is due to altered transcriptional regulation rather than loss of enzymatic genes. This finding is important beca use if the enzymatic and regulatory machinery are maintained in the betalain -pigmented or unpigmented taxa, then reactivated transcriptional regulation provides a simple mechanism for the reappearence of anthocyanic lineages. Identifying the primary caus e for a switch from betalain to anthocyanin production, however, remains problematic. Effect of Ordered Character C oding S upport for Ehrendo rfer (1976. In this study, different forms of character coding were employed to reflect alternative interpretatio ns of evolutionary, developmental and biosynthetic evidence: and the different coding strategies have a significant effect on the reconstruction of pigment evolution (table 32 and 3 3, Figure 3 3). Use of ordered coding, which only allows pigments to ari se from an unpigmented state, increased the total number of transformations from seven (unordered analyses) to as many as ten, in the parsimony reconstruction analyses, and from 15 transformations (unordered analyses) to 24 in the stochastic mapping analys es (tables 3 1 and 3 2). These increases were anticipated, as a single transition between pigments in the ordered character coding analyses would automatically be doubled due to the intervening unpigmented stage. In stochastic mapping analyses, as expected ordered coding also resulted in an increase in unpigmented dwell time which is a measure of the amount of branch length that is occupied by the unpigmented state (tables 3 1 and 3 2). There are, however, more subtle consequences of ordered coding. The intercalation of betalain and anthocyanic lineages is such that ordered character -state parsimony reconstruction with DELTRAN optimality reconstructs the backbone of the tree as lacking pigment, with seven separate origins of betalains and anthocyanins ari sing

PAGE 104

104 from this unpigmented condition (Figure 6 3). The ordered stochastic mapping analysis also reconstructs the early backbone of the tree as unpigmented (Figure 32). Therefore, the pattern of distribution of pigment types across the core Caryophyllales can be somewhat consistent with the Ehrendorfer hypothesis (1976) of ancestral lack of pigments, if the assumption is made that no direct transitions can occur between betalains and anthocyanins. The Occurrence of U n pigmented L ineages The presence of deri ved unpigmented lineages is, in general, difficult to explain. Above we highlighted the possibility that, in concordance with Ehrendorfer (1976), these lineages might have retained the plesiomorphic, unpigmented state. Only one analysis (DELTRAN; Figure 3 6) was able to reconstruct this scenario with the consequence that eight origins of pigment are also inferred. However, in all other analyses, at least three losses of pigment are always inferred. Biological explanations for these instances of pigment loss are lacking, with the exception of Achatocarpaceae, which exhibits a wind pollination syndrome. Limeum is suggested to be unpigmented and yet some species of Limeum possess a differentiated perianth with five small whitish petaloid structures (Endress an d Bittrich 1993), clearly adapted for entomophily. In Molluginaceae s.s although the flowers are often diminutive, many genera ( Psammotropha, Glinus Pharnaceum and Coelanthum ) have nectariferous secretory tissue (Endress and Bittrich 1993), and flowers of Glinus possess staminodial petals, all indicative of entomophily. Complete loss of pigment in these entomophilous taxa would be surprising. Again, the appearance and possible function of pigments in vegetative tissue ensures that the above explanations based solely on floral biology remain unsatisfactory. As emphasized by our character -state reconstruction analyses, the presence and coding of these derived unpigmented taxa have significant effects on reconstructions of pigment evolution, and therefore it is important to resolve these uncertainties. Are these taxa truly unpigmented or

PAGE 105

105 are the pigment levels just low? To what extent are the structural genes for anthocyanin and betalain synthesis present in these unpigmented taxa? Due to the lack of publishe d experimental data, it is unclear if analyses of these unpigmented taxa were conducted using fresh material, or on dried material in which pigments can degrade. These uncertainties may be resolved by the use of fresh floral material, application of sensitive methods for pigment detection such as HPLC, and possibly use of semi quantitative RT -PCR to assess the transcription of key pigment synthesis genes such as anthocyanin synthase. Multiple Origins of B etalain Pigmentation The distribution of anthocyani c taxa at or towards the base of three predominantly betalain -pigmented clades (Molluginaceae sister to the portulacaceous cohort; Hypertelis within the raphide clade; Caryophyllaceae sister to the Achatocarpaceae and Amaranthaceae) is provocative. Toge ther with the early divergence of two of the anthocyanic taxa ( Macarthuria and Caryophyllaceae), it suggests the possibility that the core Caryophyllales were ancestrally anthocyanic, with subsequent multiple origins of betalains (Figure 3 3 and Figure 3 6). Multiple origins of betalains is supported by seven out of eight of the reconstruction analyses, which recover at least two transformations to betalains from either anthocyanins or from the unpigmented state (tables 3 1 and 3 2). In stochastic mapping analyses, we observed higher rates of character transformation along branches leading to terminals of unknown pigment status because transformations along these branches are not forced to culminate at a pigment type in unknown terminals. Inclusion of ter minals of unknown pigment status could consequently inflate estimations of character transformation. We therefore pruned out all branches to unknown terminals in SIMMAP and demonstrated that multiple origins of betalains were recovered even when taxa of un known pigment status were excluded from the analysis. The possibility of multiple origins of betalain pigmentation has been raised by Soltis et al. (2005) and Cunoud

PAGE 106

106 (2006) but has not been discussed in the literature. The betalain pathway is unique to t his clade, and with a complex pathway, would seem unlikely to have evolved more than once. Therefore, here we consider additional evidence that may be consistent with this multiple origins scenario. Betalain biosynthesis has only been partly characteri zed in a limited number of species and certain key enzymes have not yet been isolated. All of the enzymes that have been characterized (i.e DOPA 4, 5 -dioxygenase, betanidin 50 -glucosyl transferase and betanidin 6 0 glucosyl transferase) have homologues in nonbetalain -synthesizing plants (Tanaka et al. 2008). DOPA 4, 5 -dioxygenase converts dihydroxyphenylalanine (DOPA) into betalamic acid a precursor for both betaxanthins and betacyanins. The DOPA 4, 5 dioxygenase orthologues from anthocyanic plants have been shown to have DOPA 4, 5 dioxygenase catalytic activity in vitro (Tanaka et al. 2008) This suggests that key enzymes in betalain synthesis may be derived from orthologues in anthocyanic plants (Sasaki et al. 2005). Betanidin 5 0 -glucosyl transfe rase and betanidin 6 0 -glucosyl transferase convert the precursor betanidin to betanin. These enzymes indiscriminately catalyze in vitro the transfer of glucose from UDP -glucose to hydroxyl groups of not just betanidin but also flavonols, anthocyanidins and flavones (Vogt et al. 1997); this indiscriminate catalysis highlights the possibility that certain enzymatic steps in the betalain biosynthetic pathway could be performed by enzymes with broad substrate specificity. This broad substrate specificity may have existed prior to the origin of betalain synthesis and might have facilitated the co -option of these enzymes to the betalain biosynthetic pathway. The two glucosyl transferases (betanidin 5 0 -glucosyl -transferase and betanidin 6 0 glucosyl transfe rase) only share 15% amino acid identity and belong to different lineages of glucosyl transferases; in each lineage the betanidin glucosyl transferases are sister to glucosyl transferases involved in flavonoid or anthocyanin pathways (Vogt 2002). The exist ence of

PAGE 107

107 betanidin glucosyl transferases derived from distinct lineages emphasizes that recruitment of genes to the betalain biosynthetic pathway may have occurred independently from distinct gene lineages. Evidence of such independent gene recruitment enco urages speculation that betalain pigmentation in distinct organismal lineages may also be the consequence of independent cooption of ancestral homologues. Biochemical and molecular investigations in a rang e of caryophyllid taxa have suggested several se mi indepe ndent pathways in the synthesis of betalains, as discussed by Tanaka et al. (2008) Betaxanthins can be derived from betalamic acid and amino acids conjugates Some betalains may be synthesized by tyrosinase or polyphenol oxidase after the condensation of betalamic acid and amino acid (Gandia Herrero et al. 2005). Additionally, betalains can be synthesized by two semi independent pathways from the intermediate dihydroxyphenylalanine via the betalamic biosynthetic pathway or the via cyclo -dihydroxy phenylalanine glycoside pathway ( Tanaka et al. 2008). Our study asks whether this variability in betalain synthesis pathways in different taxa could be in part a reflection of independent origins of betalain pigmentation: a hypothesis that deserve s further consideration. Conclusion The reds, blues and violet colors of the plant world are typically due to anthocyanins with the exception of one clade of angiosperms (the core Caryophyllales) in which anthocyanins are inexplicably replaced by betalains in mos t members of the clade. The betalain pathway has traditionally been considered to have had a single origin but our phylogenetic analyses of the core Caryophyllales (including the polyphyletic Molluginaceae), coupled with multiple methods of character recon struction indicate a likely minimum of two origins of betalain pigmentation in this clade, as well as possible reversals to anthocyanin pigmentation. Additionally w e highlight the unknown pigment status of recently identified early -diverging lineages and the now -

PAGE 108

108 uncertain pigment status of Molluginaceae s.s We provide a robust phylogenetic framework with which to further explore the biology of betalain synthesis and to address the complex and labile evolution of betalains and anthocyanins within the core Caryophyllales.

PAGE 109

109 Figure 3 1 Phylogram of one of twelve most parsimonious trees based on the matK and rbcL data set for 11 genera of Molluginaceae, 35 mem bers of the Caryophyllales and four outgroups. All taxa newly sequenced in this study and previously assigned to Molluginaceae are marked in bold. Numbers above branches are bootstrap values. Starred nodes collapse in strict consensus.

PAGE 110

110 Figure 3 2. Maximum likelihood (ML) tree resulting from GARLI analysis of matK and rbcL data set for 11 genera of Molluginaceae, 35 mem bers of the Caryophyllales and four outgroups. All taxa newly sequenced in this study and previously assigned to Molluginaceae are marked in bold. Numbers above branches are bootstrap values. Starred nodes collapse in strict consensus of 5 replicate GARLI trees.

PAGE 111

111 Figure 3 3. Reconstruction of pigment evolution: a) parsimony reconstruction over MP tree with unordered data set; b) Stochastic mapping reconstruction of pigment evolution on a Bayesian consensus tree using unorder ed dataset; c) parsimony reconstruction over MP tree with ordered data set d) stochastic mapping reconstruction of pigment evolution on a Bayesian consensus tree using ordered data set.

PAGE 112

112 Figure 3 4. Prior simulations for ordered data set. Posterior proba bilities of each rate category given each combination of E(T) and SD(T) and sampled from the prior with 10,000 realizations. X axis: rate of substitution. Y axis: sampling frequency of each discrete category.

PAGE 113

113 Figure 3 5. Prior simulations for unordered data set. Posterior probabilities of each rate category given each combination of E(T) and SD(T) and sampled from the prior with 10,000 realizations. X axis: rate of substitution. Y axis: sampling frequency of each discrete category.

PAGE 114

114 Figure 3 6. Parsim ony reconstruction of pigment evolution using unordered data sets and ordered data sets: a) unordered: DELTRAN; b) unordered: ACCTRAN; c) ordered: DELTRAN; d) unordered: ACCTRAN.

PAGE 115

11 5 Table 3 1. Summary of hypotheses historically proposed to explain the evoluti on of pigmentation in the Caryophyllales, with limitations of each. Hypotheses Proposes to Explain Limitations Loss of pigmentation through anemophily (Ehrendorfer, 1976) Origin of betalains. Mutual exclusivity of anthocyanins and betalains. In current m olecular phylogeny anthocyanic and betalain lineages are intercalated. Pigments occur in vegetative parts. Little evidence for anemophily in early diverging lineages. Selective advantage of betalains over anthocyanins (Clement and Mabry, 1994) Origin of betalains. Mutual exclusivity of anthocyanins and betalains. Little evidence for selective advantage of either betalains or anthocyanins. Functional equivalency between pigments followed by stochastic loss (Clement and Mabry, 1994) Intercalation of anth ocyanic and betalain pigmented lineages. Mutual exclusion argues against functional equivalency. Reason for stochastic loss unclear. Inhibition of anthocyanin pathway by betalains (Stafford, 1994) The mutual exclusivity of anthocyanins and betalains. Li ttle evidence for inhibition. Hard to explain possible reversals to anthocyanin following origin of betalains.

PAGE 116

11 6 Table 3 2. Frequencies of each transformation for four alternative reconstructions employing unordered character coding (parsimony with AC CTRAN and DELTRAN optimality, stochastic mapping with all taxa included, and stochastic mapping with taxa of unknown pigment status pruned). Transformations to betalains marked in bold. Transformation Type Parsimony: ACCTRAN Parsimony: DELTRAN Stochastic M apping (total) Stochastic Mapping (unknown terminals pruned) Anthocyanin to Pigment less 1 2 1.86 1.38 Pigment less to Anthocyanin 0 0 1.89 1.42 Betalain to Pigment less 2 1 3.4 2.4 Pigment less to Betalain 0 0 1.3 0.94 Anthocyanin to Betalain 1 2 2.1 1.45 Betalain to Anthocyanin 3 2 5 2.4 Total Transformations 7 7 15.603 11.33 Total Transformations to Anthocyanin 3 2 6.89 3.82 Total Transformations to Pigment less 3 3 5.26 3.78 Total Transformations to Betalain 1 2 3.4 2.39 Anthocyanin Dwell Tim e n/a n/a 0.548 0.5753 Pigment less Dwell Time n/a n/a 0.082 0.0647 Betalain Dwell Time n/a n/a 0.3722 0.36

PAGE 117

11 7 Table 3 3. Frequencies of each transformation for four alternative reconstructions employing ordered character coding (parsimony with ACCTRAN and DELTRAN optimality, stochastic mapping with all taxa included, and stochastic mapping with taxa of unknown pigment status pruned). Transformations to betalains marked in bold. Transformation Type Parsimony: ACCTRAN Parsimony: DELTRAN Stochastic Mapp ing (total) Stochastic Mapping ( unknown terminals pruned) Anthocyanin to Pigment less 2 2 4.7985 3.86 Pigment less to Anthocyanin 1 3 8.5503 6.95 Betalain to Pigment less 1 0 6.6034 5.14 Pigment less to Betalain 2 5 4.7597 3.16 Anthocyanin to Betalain 0 0 0 0 Betalain to Anthocyanin 2 0 0 0 Total Transformations 8 10 24.81 19.13 Total Transformations to Anthocyanin 3 3 8.56 6.95 Total Transformations to Pigment less 3 2 11.40 9.0 Total Transformations to Betalain 2 5 4.7597 3.16 Anthocyanin Dwell Time n/a n/a 0.5165 0.54 Pigment less Dwell Time n/a n/a 0.1736 1.339 Betalain Dwell Time n/a n/a 0.3099 0.3229

PAGE 118

118 Table 3 4. Voucher, citation, and GenBank accession No. for rbcL (Voucher information not shown for previously published sequences). F a mily S pecies V oucher C itation Embl Molluginaceae Adenogramma sp. Unvouchered, 11501 (K) This Study Didieraceae Alluaudia ascendens Cunoud et al, 2002 AY0425 Ancistrocladac Ancistrocladus korupensis Meimberg et al. 2001 AF31593 Asteropeiac eae Asteropeia micraster Cunoud et al, 2002 AY0425 Basellaceae Basella alba Crawley et al, unpublished Berberidopsida Berberidopsis corralina Cunoud et al, 2002 A042554 Nyctaginaceae Bougainvillea glabra Crawley et al, unpublished Ama ranthaceae Celosia trigyna Muller and Borsch, 2005 AY5148 Portulacaeae Claytonia megarhiza OQuinn and Hufford 2005 AY7641 Molluginaceae Coelanthum grandiflorum Thomspson and Le Roux, This Study Lophiocarpacea Corbichonia decumbens Cunoud et al, 2002 AY0425 Aizoaceae Delosperma cooperi Nyffeler 2007 DQ5584 Droseraceae Dillenia indica Crawley et al, unpublished Drosophyllacea Drosera capensis Cameron et al, 2002 Frankeniaceae Drosophyllum lusitanicum Muller and Borsch, 2005 AY5148 Gisekiaceae Frankenia pulverulenta Muller and Borsch, 2005 AY5148 Molluginaceae Gisekia africana JR Clarkson, 5499 (K) Cunoud et al, 2002 AY0425 Molluginaceae Glinus lotoides RM Harley, 19007 (K) Cunoud et al, 2002 Halophytaceae Glischrothamnus ulei Cunoud et al, 2002 AY0425 Dilleniaceae Halophytum ameghinoi Muller and Borsch, 2005 AY5148 Molluginaceae Hypertelis salsoloides Brockington SF, Moll 01 This study Limeaceae Limeum africanum Cr awley et al, unpublished Plumbaginacea Limonium latifolium Muller and Borsch 2005 AY5148 Nyctaginaceae Mirabilis jalapa Cunoud et al, 2002 AY0426 Molluginaceae Macarthuria sp. N Byrnes, 2701 (K) This study Molluginaceae Mollugo verticellat a Crawley et al, unpublished Nepenthaceae Nepenthes alata Meimberg et al. 2001 AF31589 Cactaceae Opuntia quimilo Nyffeler, 2002 AY0152 Cactaceae Pereskia aculeata Edwards, 2005 AY8753 Molluginaceae Pharnaceum sp. Unvouchered, 11503 (K) Cu noud et al, 2002 AY0426 Achatocarpacea Phaulothamnus spinescens Muller and Borsch, 2005 AY5148 Phytolaccaceae Phytolacca americana Crawley et al, unpublished Plumbaginacea Plumbago auriculata Crawley et al, unpublished Molluginaceae Polp oda capesis This study Polygonaceae Polygonum cespitosum J P H Acocks, 17405 (K) Crawley et al, unpublished Portulacaceae Portulaca oleracea Edwards et al, 2005 AY8753 Molluginaceae Psammotropha myriantha This study Rhabdodendrac Rhabdod endron amazonicum Bidgood et al, 2142 (K) Crawley et al, unpublished Phytolaccaceae Rivinia humulis Muller and Borsch 2005 AY5148 Sarcobataceae Sarcobatus vermiculatus Cunoud et al. 2002 AY0426 Simmondsiacea Simmondsia chinensis Muller and Borsch 2005 AY5148 Amaranthaceae Spinacia oleracea Schmitz Linneweber et al. AJ40084 Stegnospermata Stegnosperma halmifolium Crawley et al, unpublished Caryophyllacea Stellaria media Fior et al. 2006 AY9326 Mollug inaceae Suessenguthiella scleranthoides Cunoud et al 2002 AY0426 Portulaceae Talinum paniculatum Merxmueller and Giess, Nyffeler 2002 AY0152 Tamaricaceae Tamarix canariensis Cunoud et al 2002 AY0426 Dilleniaceae Tetracera asiatica Cunoud et al. 2002 AY0426 Dioncophyllace Triphyophyllum peltatum Meimberg et al. 2001 AF31594 Vitaceae Vitis aestivalis Fishbein et al. 2001 AF27463

PAGE 119

119 Table 3 5. Voucher, citation, and GenBank accession No. for matK (voucher information not shown for previously published sequences) F amily S pecies V oucher C itation Embl Molluginaceae Adenogramma sp. Unvouchered, 11501 (K) This Study Didieraceae Alluaudia ascendens Cuenoud et al, 2002 AY0425 Ancistrocladac Ancistrocladus korupensis Meimberg et a l. 2001 AF31593 Asteropeiaceae Asteropeia micraster Cuenoud et al, 2002 AY0425 Basellaceae Basella alba Crawley et al, unpublished Berberidopsida Berberidopsis corralina Cuenoud et al, 2002 A042554 Nyctaginaceae Bougainvil lea glabra Crawley et al, unpublished Amaranthaceae Celosia trigyna Muller and Borsch, 2005 AY5148 Portulacaeae Claytonia megarhiza OQuinn and Hufford 2005 AY7641 Molluginaceae Coelanthum grandiflorum Thomspson and Le Roux, This S tudy Lophiocarpacea Corbichonia decumbens Cuenoud et al, 2002 AY0425 Aizoaceae Delosperma cooperi Nyffeler 2007 DQ5584 Droseraceae Dillenia indica Crawley et al, unpublished Drosophyllacea Drosera capensis Cameron et al, 2002 Franken iaceae Drosophyllum lusitanicum Muller and Borsch, 2005 AY5148 Gisekiaceae Frankenia pulverulenta Muller and Borsch, 2005 AY5148 Molluginaceae Gisekia africana JR Clarkson 5499 (K) Cuenoud et al, 2002 AY0425 Molluginaceae Glinus lotoides RM Harley 19007 (K) Cuenoud et al, 2002 Halophytaceae Glischrothamnus ulei Cuenoud et al, 2002 AY0425 Dilleniaceae Halophytum ameghinoi Muller and Borsch, 2005 AY5148 Molluginaceae Hypertelis salsoloides Brockington SF Moll 01 This study Limeaceae Limeum africanum Crawley et al, unpublished Plumbaginacea Limonium latifolium Muller and Borsch 2005 AY5148 Nyctaginaceae Mirabilis jalapa Cuenoud et al, 2002 AY0426 Molluginaceae Macarthuria sp. N Byrnes 2701 (K) T his study Molluginaceae Mollugo verticellata Crawley et al, unpublished Nepenthaceae Nepenthes alata Meimberg et al. 2001 AF31589 Cactaceae Opuntia quimilo Nyffeler, 2002 AY0152 Cactaceae Pereskia aculeata Edwards 2005 AY8753 Mollugina ceae Pharnaceum sp. Unvouchered, 11503 (K) Cuenoud et al, 2002 AY0426 Achatocarpacea Phaulothamnus spinescens Muller and Borsch, 2005 AY5148 Phytolaccaceae Phytolacca americana Crawley et al, unpublished Plumbaginacea Plumbago auriculata Cr awley et al, unpublished Molluginaceae Polpoda capesis This study Polygonaceae Polygonum cespitosum J P H Acocks, 17 405 (K) Crawley et al, unpublished Portulacaceae Portulaca oleracea Edwards et al, 2005 AY8753 Molluginaceae Psammotropha myri antha This study Rhabdodendrac Rhabdodendron amazonicum Bidgood et al, 2142 (K) Crawley et al, unpublished Phytolaccaceae Rivinia humulis Muller and Borsch 2005 AY5148 Sarcobataceae Sarcobatus vermiculatus Cuenoud et al. 2002 AY0 426 Simmondsiacea Simmondsia chinensis Muller and Borsch 2005 AY5148 Amaranthaceae Spinacia oleracea Schmitz Linneweber et al. AJ40084 Stegnospermata Stegnosperma halmifolium Crawley et al, unpublished Caryophyllacea Stellar ia media Fior et al. 2006 AY9326 Molluginaceae Suessenguthiella scleranthoides Cuenoud et al 2002 AY0426 Portulaceae Talinum paniculatum Merxmueller and Giess Nyffeler 2002 AY0152 Tamaricaceae Tamarix canariensis Cuenoud et al 20 02 AY0426 Dilleniaceae Tetracera asiatica Cuenoud et al. 2002 AY0426 Dioncophyllace Triphyophyllum peltatum Meimberg et al. 2001 AF31594 Vitaceae Vitis aestivalis Fishbein et al. 2001 AF27463

PAGE 120

120 Table 3 6. The estimations of the posterior probability of ancestral characters for the ordered and unordered evolution of pigment s for all nodes presented graphically in Fig 3. Node Clade Description Frequency of Clade Un ordered Ordered Anthocyanin Pigment less Betalain Anthocyanin Pigment l ess Betalain A Caryophyllales s.l 1.0000 0.9918 0.0012 0.007 0.9953 0.0047 0 B non core Caryophylales 1.0000 0.9999 0 0.0001 0.9999 0.0001 0 C core Caryophyllales 0.7029 0.7143 0.0085 0.2772 0.6176 0.3819 0.0005 D Simmondsia + core Caryophyllales 1.000 0.2162 0.0177 0.766 0.1781 0.782 0.0438 E Asteropeiaceae/Physenaceae + core Caryophyllales 1.0000 0.0279 0.0153 0.9567 0.0125 0.9289 0.0586 F Macarthuria + core Caryophyllales 1.0000 0.0163 0.0056 0.978 0.074 0.9454 0.0473 G CAA clade + core Caryophy llales 0.6575 0.0006 0.017 0.9977 0.002 0.3165 0.6834 H CAA clade 0.9609 0.046 0.1632 0.7908 0.0115 0.611 0.3776 I Amaranthaceae/Achatocarpaceae 0.9998 0.0069 0.6231 0.37 0.0014 0.7947 0.2038 J Stegnosperma + core Caryophyllales 1.0000 0.0001 0.0004 0 .9995 0 0.0313 0.9687 K Limeum + globuloid inclusion clade 0.9992 0.0003 0.0014 0.9983 0 0.0222 0.9778 L globuloid inclusion clade 0.9255 0.0004 0.0002 0.9994 0 0.0207 0.9793 M raphide clade 1.0000 0.0006 0.0001 0.9993 0 0.0093 0.9907 N Mollugina ceae + succulent clade 1.0000 0.013 0.0025 0.9845 0.0008 0.0805 0.9188 O succulent clade 1.0000 0.0001 0 0.9999 0 0.0012 0.9988 P Molluginaceae s.s. 1.0000 0.7974 0.0348 0.1678 0.4472 0.4459 0.0819

PAGE 121

121 Table 3 7 Pigment coding for each termina l and sources of information used to designate characters states. Coding Strategy : Anthocyan in (0), Unpigmented (1), Betalain (2), U nknown (?). Taxa Character Mapping Literature Cited Adenogramma 1 Cuenound et al., 2002 Alluaudia 2 Cle ment et al., 1994 Ancistrocladus ? Asteropeia ? Basella 2 Clement et al., 1994 Berberidopsis ? Bougainvillea 2 Clement et al., 1994 Celosia 2 Clement et al., 1994 Claytonia 2 Clement et al., 1994 Coelant hum ? Corbichonia 2 Cuenound et al., 2002 Delosperma 2 Clement et al., 1994 Drosera 0 (Di Gregorio and Dipalma 1966) Drosophyllum ? Frankenia 0 (Beale et al. 1941) Gisekia 2 Clement et al., 1994 Glin us ? Glischrothamnus ? Halophytum 2 Clement et al., 1994 Hibbertia 0 (Price and Sturgess 1938) Hypertelis 0 Clement et al., 1994 Limeum 1 Clement et al., 1994 Limonium 0 (Price and Sturgess 1938) M acarthuria 0 Clement et al., 1994 Mirabilis 2 Clement et al., 1994 Mollugo 0 Clement et al., 1994 Nepenthes 0 (Schaefer and Ruxton 2008) Opuntia 2 Clement et al., 1994 Pereskia 2 Clement et al., 1994 Pharn aceum 0 Clement et al., 1994 Phaulothamnus 1 Clement et al., 1994 Physena ? Phytolacca 2 Clement et al., 1994 Plumbago 0 (Price and Sturgess 1938) Polpoda ? Clement et al., 1994 Polygonum 0 (Price and Sturgess 1938) Portulaca 2 Clement et al., 1994 Psammotropha ? Rhabdodendron ? Rivina 2 Clement et al., 1994 Sarcobatus 2 Cuenound et al., 2002 Simmondsia ? Spinacea 2 Clement et al., 1994 Stegnosp erma 2 Clement et al., 1994 Stellaria 0 Clement et al., 1994 Talinum 2 Clement et al., 1994 Tamarix 2 (Beale et al. 1941) Tetracera 0 (Price and Sturgess 1938) Triphyophyllum ? Vitis 0 (Robinson and Robinson 1933)

PAGE 122

122 CHAPTER 4 KEEP THE DNA ROLLING: M ULTIPLE DISPLACEMENT AMPLIFICATION OF ARCHIVAL PLANT DNA EXTRACTS Insufficient reserves of genomic DNA can hamper molecular phylogenetic analysis. Highthroughput genetic techniques that require relatively large amounts of DNA, the difficulty in obtaining samples of taxa from remote regions, and re -sampling of limited archival DNA by repeated phylogenetic surveys can often limit the DNA available for study. To provide a possible solution to this problem, we applied Multiple Displacement Amplification (MDA) to eight archival genomic DNA extracts. The performance of MDA -treated DNA versus untreated genomic extract was evaluated by PCR amplification of three common phylogenetic markers (psbB nad7, and ITS) acro ss a dilution series. Generally, amplification of all three genetic markers from the MDA treated DNA dilutions was greater than from equivalent dilutions of untreated genomic template. These results indicate that genes from all three plant genomes were amp lified and that copies of the target genes psbB nad7, and ITS were substantially increased during the MDA procedure. Sequencing of the psbB nad7, and ITS PCR products from both the MDA treated DNA and the untreated template was used to assess the fidelity of the MDA procedure. Sequences from the MDA treated DNA and the untreated genomic template differed by 1.2 x 104 %, which is within the margin of Taq error. These findings emphasize the significance of Multiple Displacement Amplification for optimization of weak PCR, maintenance of depleted genetic stocks, increasing density of taxon sampling, and improving consistency between different phylogenetic analyses. Introduction Many plant systematists work with taxa that are rare or difficult to collect, maki ng it difficult or impossible to obtain samples for DNA isolation. Small fragments of silica dried material may be the only source of material for hard to -obtain taxa. Herbarium collections and

PAGE 123

123 DNA banks are therefore an invaluable source of material for p hylogenetic studies and are increasingly important as a source of DNA for molecular analyses (Drabkova et al., 2002) The success of th ese molecular studies depends on DNA of adequate quantity and quality and yet extraction of sufficient amounts of high-quality DNA from herbarium specimens or silica -dried material can be challenging. Herbaria may limit destructive sampling of specimens in order to preserve collections relatively intact, specimens may be rare and unavailable for sampling, and the DNA of herbarium specimens may have degraded because of age or poor preservation, yielding low quantities of DNA. Furthermore, the DNA of hardto -obtain samples is often used repeatedly in phylogenetic surveys (Chase et al., 1993; Soltis et al., 2000) and is ultimately depleted. In recent years a number of methods have been developed for whole genome amplific ation that generate considerable amounts of DNA from minute quantities of starting material (Pinard et al., 2006) Rolling Circle Amplification (RCA), also known as Multiple Displacement Amplification (MDA), is a widely used method for whole genome ampl ification that employs a bacteriophage 29 DNA polymerase together with exonuclease resistant degenerate primers to amplify genomic DNA isothermally at 30oC (Dean et al., 2002; Lage et al., 2003) MDA has a number of advantages over PCR -based methods of genome amplification. These include a less biased amplification of the genome, greater yields of DNA, and higher fidelity to the template DNA (Pinard et al., 2006) The importance of genomic data in biological scie nces often places considerable demand on supplies of DNA; therefore, there is interest in the use of MDA to augment stocks of low abundance DNA (Lasken and Egholm, 2003) Several studies from a range of biological disciplines have already demonstrated the utility of MDA in the molecular analysis of rare

PAGE 124

124 biological material. Examples of low abundance biological samples that have been successfully amplified using MDA include: limited medical samples such as buccal swabs and single cell samples generated by laser capture microscopy (Lasken and Egholm, 2003) single microorganisms (Raghunathan et al., 2005) and single spores of abuscular mycorrhizal fungi (Gadkar and Rillig, 2005) Despite these successful applications of MDA, the technique does not seem to be have been widely adop ted by plant systematists, although supplies of DNA may often be limited. The aim of this study was to assess the efficacy of MDA for amplification of plant DNA extractions by amplifying and sequencing common phylogenetic markers from MDA treated DNA. We a mplified genes from each of the three plant genomes: psbB (plastid genome), nad7 (mitochondrial genome), and ITS (nuclear genome) with the goal of investigating the PCR performance of MDA treated DNA over its untreated template. Materials and Methods DNA E xtractions from Herbarium S pecimen s DNA from eight different species (table 4 1) belonging to the angiosperm order Caryophyllales was used as template for MDA. These DNA samples were extracted from materials of varying age and condition, which was reflecte d in the variable quality and quantity of the DNA extracted ( Figure 4 1) Multiple Displacement Amplification DNA extracts from the eight species were amplified using the Genomiphi kit (Amersham, Piscataway, NJ, USA) according to the manufacturers instructions. Reactions were performed using crude DNA extract. In each case, 1 l of crude DNA extract was combined with 9 l of sample buffer, heated to 95oC for 3 min, and then chilled to 4oC. To each tube, 9 l of reaction buffer and 1 l of Genomiphi enzyme mix were added, and the reactions incubated at 30oC for

PAGE 125

125 16 hrs. A negative control was performed in which no input DNA was added. On completion of the MDA amplification, a 1 l aliquot of the reaction was fractionated by electrophoresis on a 1.5% agarose g el, stained with ethidium bromide, and visualized under UV light. For comparison a 1 l aliquot of the untreated DNA extraction was run along with its respective MDA treated DNA. PCR Untreated DNA extract and MDA -treated DNA were both serially diluted: 1:5, 1:10, 1:100, and 1:200. To assess the value of MDA -treated DNA versus untreated DNA extract, PCR was performed across this dilution series. Gene regions from the three plant genomes were amplified: psbB nad7, and ITS Amplification and sequencing was performed in accordance with previously published protocols (Soltis et al., 2000) psbB was amplified according to previously published protocols (Soltis et al., 2000) (95oC for 3min, followed by 3035 cycles of 94oC for 30 sec, 50oC for 30 sec, and 72oC for 1 min with a final extension time of 7 min at 72oC) using primers 60F (ATG GGT TTG CCT TGG TAT CGT GTT CAT AC] and 66R (CCA AAA GTR AAC CAA CCC CTT GGA C) (Graham and Olmstead, 2000) which amplify 1362 bp between positions 1 and 1362 (inclusive) of psbB nad7 was amplified ( 95oC for 2 min, foll owed by 30 cycles of 94oC for 30 sec, 50oC for 30 sec, and 72oC for 1 min with a final extension time of 7 min at 72oC) using primers nad7/2 (GCT TTA CCT TAT TCT GAT CG) and nad7/3 (TGT TCT TGG GCC ATC ATA GA) (Duminil et al., 2002) which amplify ~ 1000 bp product ITS was amplified (95oC for 2 min, followed by 5 cycles of 94oC for 30sec, 53oC for 30 sec, and 72oC for 1 min with a step down to 28 cycles of 94oC for 30 sec, 48oC for 30 sec, and 72oC for 1min) using primers ITSA (GGA AGG AGA AGT CGT AAC AAG G) and ITSB (CTT TTC CTC CGC TTA TTG ATA TG) (Blattner, 1999) which amplify a 650700 bp

PAGE 126

126 product The PCR products were fractionated by electrophoresis on a 1.5% agarose gel, stained with ethidium bromide, and visualized under UV light. PCR Cleanup and Sequencing To remove unused primers and nucleotides, PCR product was treated with ExoSAP -IT (USB Corporation, Cleveland, OH). All PCR products were sequenced using the same primers as for the PCR with the exception of the psbB products, which were sequenced using nested primers B60F (CAT ACA GCT CTA GTT KCT GGT TGG) and B66R (CCC CTT GGA CTR CTA CGA AAA ACA CC) (Grah am and Olmstead, 2000) Sequencing was performed on the Applied Biosystems 3730xl capillary DNA sequencer. Raw data signals were automatically base-called and subsequently manually checked using Sequencher 4.2 (Gene Codes Corporation, Ann Arbor, MI, USA). Sequences were aligned using Se-Al (Department of Zoology, Oxford, UK). The sequences generated by this study have been submitted to Genbank: EU410352EU410375 Results Multiple Displacement Amplification MDA on all eight DNA extracts resulted in an increa se in high-molecular -weight DNA (10 11Kb) relative to the template DNA ( Figure 4 1). This increase in high molecular -weight DNA is especially apparent for Mollugo verticillata and Mirabilis jalapa from which either very little or highly degraded DNA was initially extracted. An amplification product also occurs in the negative control. This is expected when using the Genomiphi kit and is most likely due to the non -specific amplification of degenerate hexamer primers rather than to DNA contamination (Amersham Biosciences, Protocol 74004229). The presence of this amplification product in the no -template control, however, prevents immediate evaluation of the success of MDA in specifically amplifying herbarium DNA. PCR was therefore performed to evaluate the util ity of the MDA treated DNA.

PAGE 127

127 PCR Amplification of psbB gene was initially performed on a 1:25 dilution of the MDA treated DNA. All 8 MDA-treated DNAs generated a strong amplification product of the expected size (results not shown). PCR performed on the notemplate MDA control product did not generate the expected amplification product, demonstrating that the amplification products from PCR of MDA treated DNA were not due to spurious amplification during the MDA procedure. To assess the effectiveness of the MDA -treated DNA versus the original DNA template, amplification of psbB nad7, and ITS was performed across a dilution series. Reactions contained 0.2, 0.1, 0.01, or 0.005 l of either the untreated genomic template or its MDA treated equivalent. As a gen eral trend, for all species and at all dilutions, the MDA -treated product generated a similar or stronger PCR amplification than the corresponding genomic template (Figure 4 2). This is particularly apparent for the amplification of psbB, nad7 and ITS from M. verticillata and M. jalapa (Figure 4 2). Differences between dilutions for the amplification of ITS are less striking but still discernable. For all DNA regions however, i t is most informative to compare the 1:5 and 1:10 MDA treated dilution with the respective 1:100 and 1:200 dilutions of the untreated genomic template. This is an appropriate comparison because the MDA procedure effectively dilutes the initial genomic template by 1:20 (1 l of template in a 20 l MDA reaction) and thus following furt her dilution, the 1:5 (M1) and 1:100 (G3), and 1:10 (M2) and 1:200 (G4) represent equivalent dilutions of the initial genomic stock. Amplifications from the 1:5 and 1:10 dilutions of MDA treated DNA are without exception considerably stronger than their re spective 1:100 and 1:200 dilutions of genomic DNA across all 8 species for all three genes (for example see Figure 4 2)).

PAGE 128

128 Sequencing PCR products from the untreated DNA and corresponding MDA treated DNA were sequenced to ensure that the amplifications wer e not due to spurious DNA amplified during the MDA procedure. In all cases, the sequences from untreated and MDA treated DNA were nearly identical with a difference of only 1.2 x 104 %. These differences are within the margin of Taq error, demonstrating that no contamination had occurred during the MDA procedure and that replication of DNA in the MDA treatment had been of high fidelity. Discussion Phylogenetic studies of plants are based on ever increasing portions of the nuclear, mitochondrial, and plas tid genomes and utilize high-throughput sequencing technology that can often exhaust reserves of DNA. Archived collections of DNA can gradually degrade with use and age and may be difficult to replenish or restore. Collaborative projects rely on the same s amples for genetic analysis, yet limited stocks of archival DNA force participating laboratories to exchange minute quantities of DNA. Increased taxon sampling in molecular phylogenetics has created a demand for DNA from difficult to obtain species and DNA from these hard to -obtain species is often available in only limited amounts. For these reasons, DNA is often a limiting resource in molecular phylogenetic analyses. Strong initial amplification of the psbB gene from MDA treated DNA suggested that templat e DNA was successfully replicated during the MDA procedure. Comparisons of psbB nad7, and ITS amplification across dilutions of untreated and MDA -treated DNA confirmed that amplification was generally greater from MDA -treated DNA than from the original te mplate (Figure 4 2). This was particularly apparent in the amplification of the psbB and nad7 genes; however, differences in the amplification of ITS from MDA -treated DNA versus genomic DNA though present, were less distinct. This might be due to the effi ciency of the ITS amplification

PAGE 129

129 protocol and to the high abundance of the 18S rDNA template in the genomic extract. Nonetheless, the generally increased amplification from MDA treated DNA indicates that copies of the target genes psbB nad7, and ITS were s ubstantially increased during the MDA reaction as all three genomes were amplified during the MDA procedure. This enhanced amplification suggests that MDA could be a rapid and efficient procedure for the optimization of PCR for genes that weakly amplify fr om the original template. The comparisons of 1:5 (MDA treated) with 1:100 (untreated genomic), and 1:10 (MDA treated) with 1:200 (untreated genomic) dilutions revealed dramatic increases in amplification from MDA treated DNA. A relatively small investment of 1 l of genomic DNA generates 20 l of MDA -treated DNA, and improved subsequent PCR of the genes of interest. This illustrates the value of MDA for the augmentation of depleted DNA stocks. Sequencing of the PCR products confirmed that MDA replicated the genomic template with a high degree of fidelity as expected from previous studies that have more rigorously assessed the fidelity of the 29 DNA polymerase (Pinard et al., 2006) These findings demonstrate that Multiple Displacement Amplification is a useful procedure for pre amplification of precious archival DNA prior to subsequent genetic analysis Pretreatment by MDA enhanced amplification of all three molecular markers from the three plant genomes, while maintaining fidelity to the original templat e sequence. MDA therefore offers a way to alleviate problems of depleted DNA stocks and help to maintain DNA standards to ensure sampling consistency among phylogenetic studies. Multiple rounds of MDA may even be feasible ensuring almost infinite suppl ies of DNA (Sato et al., 2005) It is important to note that all the DNA extracts utilized in this study studies were of sufficient quality to amplify the phylogenetic markers from the genomic extract prior to MDA treatment. Successful MDA requires the input of good quality DNA and thus it is not clear that

PAGE 130

130 this procedure would permit the augmentation of severely degraded DNA stocks. In our experien ce, poor quality DNA derived from very old or poorly preserved herbarium material may not be an appropriate substrate for MDA (data not shown). Furthermore, MDA enzymes, as in PCR, may be susceptible to inhibitory substances that can be present in extracte d DNA. These limitations mean that the usefulness of MDA must be determined empirically, taking into account the quality of the extracted DNA and its performance in PCR prior to MDA treatment. We have observed that if successful PCR has been performed on a DNA extract, MDA is also likely to perform well. Despite these final cautionary comments, this approach still provides a cheap and efficient method for enhancing and maintaining minute DNA stocks from numerous invaluable museum collections worldwide.

PAGE 131

131 Figure 4 1. MDA -treated reaction products (M1 8) and corresponding untreated genomic equivalent (G1 8). M1 & G1: H. ameghinoi ; M2 & G2: M. verticillata ; M3 & G3 : M. jalapa; M4 & G4: D. capensis ; M5 & G5: L. arborescens ; M6 & G6: T.paniculatum ; M7 & G7: P.virginianum ; M8 & G8 : R. amazonicum C: the control MDA reaction to which no genomic template was added. L: l DNA/ Pst1 (MBI Fermentas Inc, Hanover, NH, USA) Figure 4 2. Amplification of 1363 bp fragment of psbB fragment of nad7 ge ne fragment of ITS for two of the eight species, Mollugo verticillata and Mirabilis jalapa. Amplifications across dilution gradient of MDA treated template (1:5; M1, 1:10; M2, 1:100; M3, 1:200; M4) and untreated genomic template (1:5; G1, 1:10; G2, 1:100; G3, 1:200; G4). 100bp ladder (Bioneer, Alameda, CA, USA).

PAGE 132

132 Table 4 1. Species of Caryophyllales used in this study: family, source, collection date, accession number and DNA extraction method. Species Family Voucher Date of Collection Extraction Method Halophytum ameghinoi Halophytaceae Tortosa et al. s.n K 1997 CTAB Mollugo verticillata Molluginaceae Moore 321 FLAS 2006 DNAeasy Mirablis jalapa Nyctaginaceae Soltis 2638 FLAS 2004 CTAB Drosera capensis Droseraceae Moore 267 FLAS 2006 DNAeasy Limmonium a rborescens Plumbaginaceae Chase 1649 K Uncertain CTAB Talinum paniculatum Portulaccaceae Soltis 2646 FLAS 2004 CTAB Polygonum virginicum Polygonaceae Soltis 2656 FLAS 2005 DNAeasy Rhabdodendron amazonicum Rhabdodendraceae E. Ribiro 1187 K 1993 CTAB

PAGE 133

133 CHAPTER 5 PHYLOGENY OF MADS -BOX GENE LINEAGES IN AIZO A CEAE (CARYOPHYLLALES ) Introduction The family of MADS -box genes (for MCM1 AGAMOUS DEFICIENS and SRF; Sommer et al., 1990) encodes transcription factors that play importa nt roles in developmental processes across e ukaryotes (Theissen et al., 2000). MADS -box genes have been the focus of part icular interest in plants be notable gene diversification together with the involvement of MADS box genes in the development of evoluti onar il y significant traits, particularly in the context of the flower. Describing the diversification of MADS -box genes (in particular the MIKC -type lineage ; Alvarez -Buylla et al., 2000) in relation to the radiation and morphological evolution of land plan ts is a major theme in the field of plant evolutionary development (Becker and Theissen, 2003; Theissen et al., 2000). The MIKC lineage of MADS -box genes is characterized by the presence of four distinct protein dom ains: M MADS; I Intervening ; K Kera tin like; and C C -terminus (Alvarez Buylla et al., 2000) The highly conserved MADS domain at the N terminus contains the CArG elements that are essential for DNA binding and plays a role in protein dimeriz ation (Norman et al., 1988; Pollock and Treisma n, 1991). The Intervening and Keratin-like domains ar e important in mediating dimeriz ation and influence the specificity of interaction between MADS -box proteins (Krizek and Meyerowitz, 1996; Pnueli et al., 1991; Riechmann et al., 1996). The C terminus is highly variable at the sequence level but short, conserved and lineage -specific motifs have been identified in the C termini of many clades of MADS -box genes. Recent evidence suggests that the C terminus may contri bute to the formation of higher -order prot ein complexes between dimers of MIKC type proteins (Egea Cortines et al., 1999).

PAGE 134

134 MIKC Genes and ABCDE Functions in Floral Organ Identity The MIKC -type genes have become a focus for studies in evolution of development largely due to their role in specifying floral organ identity, underpinning the conceptual framework of the ABC model. On the basis of floral homeotic mutants (Bowman et al., 1989; Carpenter and Coen, 1990; Komaki et al., 1988), the ABC model proposed that the identity of the four concentric w horls of floral organs in Arabidopsis thaliana is determined by the combinatorial activity of three functional classes of gene activity (Bowman et al., 1991; Coen and Meyerowitz, 1991; Meyerowitz et al., 1991). A -function alone specifies sepal identity; A function in combi nation with B -function specifies petal identity; B -function together with C function specifies stamen identity; and C -function alone specifies Carpel identity (Coen and Meyerowitz, 1991). MIKC genes were identified as the key genetic compo nents underlying homeotic mutant phenotypes and defining the A, B, and C functional classes of gene activity. For example in Arabidopsis thaliana, the APETALA1 (AP1 ) locus is involved i n determining sepal identity (A -function) (Irish and Sussex, 1990), the paralogous loci PISTILLATA ( PI ) and APETALA3 (AP3 ) are involved in petal and stamen identity (B -function) (Goto and Meyerowitz, 1994; Jack et al., 1992; Jack et al., 1994) and the AGAMOUS (AG ) locus is involved in determining stamen and carpel identity (C -function) (Mizukami and Ma, 1992; Yanofsky et al., 1990). Parallel studies in Antirrhinum majus demonstrated that DEFICIENS (DEF ) and GLOBOSA (GLO ) (orthologs of PI and AP3 ) specify B -function (Sommer et al., 1990; Trobner et al., 1992); additionally a pa ralog of AGAMOUS the PLENA ( PLE ) locus specifies C -function in A. majus (Bradley et al., 1993). Subsequent studies in A. thaliana identified MIKC type genes such as SEEDSTICK (STK ) together with homologs in other organi s ms (e.g FLORAL BINDING PROTEIN 7 a nd 11 in Petunia hybrida) that are involved in specifying ovule identity (D function) (Angenent et al., 1995; Favaro et al., 2003; Pinyopich et al., 2003). A fifth functional

PAGE 135

135 class (E -function) is represented in A. thaliana by the AGL2 3 4 9 genes (als o referred to as SEPALLATA genes), which are required for specification of all floral organ identities (Pelaz et al., 2000); a similar function for SEPALLATA homologs has been demonstrated in P. hybrida (Vandenbussche et al., 2003). Functional analyses of MIKC -type genes in a broad range angiosperm taxa (e.g. Arabidopsis Antirrhinum Petunia Gerbera Lycopersicon Papaver and Oryza ), together with expressionbased studies in basal eudicots (Ranunculales), magnoliids, and earlydiverging angiosperms (the ANITA grade) have established that homologs of the above ABCDE function MIKC type genes play an important and conserved role in floral devel opment across all angiosperms (reviewed in Irish, 2003; Ng and Yanofs ky, 2001; Theissen et al., 2000; Irish and Kram er, 1998) Phylogenetic Analyses Reveal the Evolutionary History of MIKC -type G enes The evolutionary history of the MIKC type genes is characterized by repeated duplications and divergence that have occurred throughout the evolution of angiosperms. The per vasive nature of these events ensures that comparative studies of MADS -box gene expression and function in relation to the evolution of morphological structures in different organsims require a comprehensive phylogenetic understanding of the MIKC type gene family (Becker and Theissen, 2003; Jaramillo and Kramer, 2007; Litt and Irish, 2003). Phylogenetic analysis is used to identify clades of MIKC type genes and identify orthology (homology through speciation) vs paralogy (homology through duplication) of MI KC type genes isolated fro m different organismal lineages. To date numerous phylogenetic analyses have been conducted on those MIKC -type gene lineages containing homologs of genes that have been assigned to the ABCDE functions of floral organ identity in A rabidopsis thaliana and A ntirrhinum majus Here the main findings of these

PAGE 136

136 studies are summarized largely to define expectations of the MIKC gene compleme nt in the core eudicot lineage Caryophyllales, but also to highlight the potential role of gene dupli cation in the evolution of morphological variation. In summarizing these phylogenetic studies the MIKC type gene complement and nomenclature of model organism A. thaliana is employed as a reference. Phylogeny of A -Function MIKC -type Genes The AP1 (A -funct ion) gene has three closely related genes in A. thaliana : AGAMOUS LIKE 79 (AGL79 ), FRUITFUL ( FUL ) and CAULIFLOWER (CAL ). CAL and API have been shown to be the product of a recent Brassicaceae -specific duplication (Litt and Irish, 2003) ; however AGL79 FU L and the inferred ancestor of AP1 and CAL are representatives of ancient duplications that occurred prior to the origin of core eudicots (Litt and Irish, 2003; Shan et al., 2007). Phylogenetic analyses of homologs isolated across angiosperms i ndicate that as a whole, early diverging lineages of angiosperms possessed a sing le AP1 homolog (excluding lineage specific duplications in e.g Magnolia and Peperomia ), however these genes have been termed FUL like due to their sequence similarity with the FUL lo cus in A. thaliana (Litt and Irish, 2003). The ancestral FUL -like lineage has undergone several duplication events associated with the radiation of major angiosperm lineages: a single duplication within the Ranunculales (Litt and Irish, 2003); one duplicat ion within the monocots and one duplication at the base of the monocots (Preston and Kellogg, 2006); and two duplications at the base of the core eudicots (Litt and Irish, 2003; Shan et al., 2007). As a result of these last eudicot -specific two duplication s core eudicots tend to exhibit three paralogs of the original FUL -like lineage: the euAP1 lineage (containing the AP1 locus, the euFUL lineage (containing the FUL locus) and the eu FU L -like or AGL79 lineage (containing the AGL79 locus) (Litt and Irish, 2003; Shan et al., 2007).

PAGE 137

137 B-Function MIKC -type Genes Kramer et al. (1998) was the first phylogenetic analysis to clearly demonstrate the complexity of MIKC gene evolution in angiosperms. In A. thaliana two MIKC genes, AP3 and PI have been ascribed to B -fun ction H owever isolation and phylogenetic isolation of homologs from a range of eudicot, magnoliid, and basal angiosperm taxa has revealed a complicated history of gene duplication, together with gene loss (Kim et al., 2004; Kramer et al., 1998) The AP3 a nd PI paralogs are probably present in all extant angiosperms and are representatives of an ancient duplication that preceded the emergence of the earliest extant angiosperm Amborella (Kim et al., 2004) The PI lineage has also undergone additional duplica tion events for example the many recent duplications in Ranunculaceae (Kramer et al., 2003) The AP3 lineage has undergone numerous further duplications; again in the Ranunculales (giving ris e to three paralogous lineages (Kramer et al., 2003) but also at t he base of the core eudicots giving rise to two paralogous lineages, termed eu AP3 (containing the AP3 locus) and TM6 (named after the TM6 locus in Lycopersicon Kramer et al., 1998) The TM6 paralog appear s to have been lost in the model organims A. thalia na and A. majus but is present in many core eudicots and thus constitutes a major eudicot gene lineage. The eudicot -specific duplications in the AP3 lineage ensure that three major paralogous B -function gene lineages are present in c ore eudicots: the eu AP3 lineage the PI lineage and the TM6 lineage (Kramer et al., 1998). C and D -Function MIKC -type Genes In Arabidopsis the AGAMOUS subfamily of MIKC type genes is represented by four paralogs AGAMOUS (AG ), SHATTERPROOF 1 and 2 (SHP 1 and 2 ), and SEEDSTICK (STK ). AG and STK are representatives of an ancient duplication that preceded the last common ancestor of the angiosperms (Kramer et al., 2004; Zahn et al., 2006). This duplication led to two major lineages of AGAMOUS genes in angiosperms termed the AG line age and the AGL11 lineage (a

PAGE 138

138 synonym of STK ) (Zahn et al., 2006). The AG lineage is largely involved in specifying C function while the AG11 lineage tends to be res tricted to ovule development (D -function) (Kramer et al., 2004; Zahn et al., 2006). Within the AG lineage further duplications have occurred within the monocots, the Ranunculaceae and within the eudicots. The AG lineage duplication within or at the base of the eudicots has generated two subsequent lineages, the eu AG lineage and eu PLENA ( PLE ) li neage (named after the PLENA locus in A. majus ). Due to uncertainty over deep level relationships within the core eudicots and poor taxon sampling of the AGAMOUS lineage s it is currently not clear where this duplication took place. This duplication may ha ve arisen at the base of the rosid and asterid clades as only members of the rosid and aste rid lineages possesses both homologs of the AG and PLE lineages (Kramer et al., 2004; Zahn et al., 2006). However Kr amer EM reports a possible PLE homolog in Gunner a suggesting that this duplication may have arisen earlier at the base of the core eudicots (pers.comm). SHP 1 and SHP2 from A. thalinana belong to the PLE lineage but the presence of two loci in A. thaliana suggests a recent dupl ication perhaps specific to Arabidopsis (Kramer et al., 2004; Zahn et al., 2006). Core eudicots may be expected to have a minimum of two AG paralogs belonging to the eu AG lineage (containing the AG lineage) and the AGL11 (containing the STK locus). Depending on the precise timing of the eu AG / eu PLE duplication, core eudicots may be expected to possess a third paralog from the PLE lineage (containing the SH P1 and 2 loci). E-function MIKC -type Genes In A. thaliana the SEPALLATA (SEP ) subfamily of MIKC type genes is represented by four l oci, AGAMOUS like ( AGL ) 2, 3, 4, and 9 (also named SEPALLATA 1 4 2 and 3 respectively) (Ditta et al., 2004; Pelaz et al., 2000). These four paralogs are representatives of several duplication events (Zahn et al., 2005). The first duplication prior to the emergence of extant angiosperms gave rise to two distinct lineages the AGL9 and AGL2 / 3 / 4 lineages (Zahn et

PAGE 139

139 al., 2005). These lineages have subsequently undergone additional duplication in association with the radiation of major angiosperm lineages. The AGL2 / 3 / 4 lineage has undergone two major duplications in monocot lineages (Malcomber and Kellogg, 2005) with a further two duplications within the core eudicots (Zahn et al., 2005). The core eudicot -specific duplications of the AGL2 / 3 / 4 lineage have given rise to three lineages: the AGL2 lineage (containing the AGL2 and AGL4 loci); the FBP9 lineage (named after the representative locus isolated from Petunia hybrida the FBP9 paralog appears to have been lost in A. thaliana ); and the AGL3 lineage (containin g the AGL3 locus). The AGL9 lineage has undergone duplication within the monocots (Malcomber and Kellogg, 2005) and a further possible duplication within the core eudicots, although both paralogs from this putative core eudicot duplication have been isolat ed only from asterid lineages. Core eudicots can be expected to possess representatives of up to five paralogous SEP gene lineages including three homologs of the AGL2 FBP9 AGL3 lineage, and up to two homologs from the AGL9 lineage (Zahn et al., 2006) Duplications in Association with Morphological Variation and Novelty Gene duplication is an important phenomenon for studies in evolutionary development due to the potential of paralogs to acquire altered gene functions that might contribute to altered or novel morphology. Three basic models have been proposed to account for the potential consequences of a gene duplication event. The pseudogenization model argues that if paralogs remain completely redundant in function they will not be selectively mai ntained over evolutionary time and thus one paralog will rapidly accumulate deleterious mutations and subsequently be lost. The second model sub -functionalization argues that given the often modular nature of protein structure and/or function, paralog s could be selectively maintained if the various functions of the ancestral homolog were parsed out between its descendent para logs. In the third model, neo functionalization, one of the two paralogs derived from a duplication

PAGE 140

140 event would retain the func tion of the ancestral homolog while the second would acquire a novel function again resulting in the selective maintenance of both copies. Although there is evidence for all these potentialities with respect to the evolution of MIKC type genes here, a ca se study of potential neo -functi onalization in the Ranunculales is briefly reviewed. As previously highlighted, there have been several duplications in the AP3 and PI lineages within the Ranunculales (Kramer et al., 2003). Kramer et al. (2007) investigated the evolution of three AP3 (AqvAP3 1, 2 and -3 ) paralogs and the Aqv PI homolog in connection to the evolution of novel staminodial organs in Aquilegia The staminodia in Aquilegia are considered derived from stamens on the basis of their early ontogeny and their positional homology to stamens in other genera of Ranunculales. The stamens and staminodia in Aquilegia share similar expression patterns of AqvAP3 -1 Aqv AP3 -2 and Aqv PI early in development, perhaps reflecting the derivation of the staminodia from stamens. Subsequently, however, different AP3 paralogs are expressed in the stamens and staminodia, with Aqv AP3 -2 maintained in the stamens and AqvAP3 -1 in staminodia (Kramer et al., 2007). This partitioning of expression of AP3 paralogs in association with the evolution of a novel floral organ may signify a neofunctionali zation event in which the AqvAP3 -1 paralog has evolved to specify an alternative floral morphology. This case study from the Ranunculales sets a precedent for investigating instances of perianth differentiation in the c ore Caryophyllales. In Chapter 2 several instances of perianth differentiation derived from a uniseriate undifferentiated condition were identified in the core Caryophyllales. Several of these differentiated perianths are inferred to have evolved through the recruitment and modification of an droecial organs (Caryophyllaceae, Glinus Corbichonia, and Aizoaceae ) to generate novel floral organs in similar manner to that envisaged in Ranunculales Given the similar pattern of modification of androecial -derived organs it is reasonable to ask

PAGE 141

141 whethe r the emergence of these novel structures is associated with similar gene duplication events in the MIKC type homologs that are implicated in the ABCDE functions of floral development. Here, the ev olution of MIKC type homologs is investigated in the l inea ge Aizoaceae, a lineage of Caryophyllales, in which petaloid organs are thought to have arisen through sterilization and modification of stamens. The two early -diverging subfamilies within Aizoaceae, Sesuvioideae and Tetragonoideae, lack petaloid staminode s and form a grade leading to a monophyletic clade containing the subfamilies Mesembryanthemoideae and Ruschoideae, both of which possess petaloid staminodes. Our hypothesis is that, if duplication in ABCDE MIKC type genes is involved in the emergence of petaloid staminodes, these duplication events should coincide with the origin of the Ruschoideae and Mesembryanthoideae lineages. These duplication events would be detectable by an increase in the MIKC type gene complement in these derived sub-families, wit h the presence of additional paralogs. The focus here is primar i ly the evolution of those MIKC type genes that contribute to the ABC functions of floral organ identity although the methodology used ensures that homologs underlying D and E functions are als o recovered as are MIKC -type genes from a wide range of additional gene families Methods Taxon Sampling The species sampled included five species of Aizoaceae (core Caryophyllaes) representing the four sub families in Aizoaceae: Sesuvium portulacastrum (Sesuvioideae), Tetragonia teragonoides (Aizooideae), Mesembryanthemum crystallinum and Mesembryanthemum cordifolia (Mesembryanthemoideae) and Delosperma napiforme (Ruschoideae). Two additonal species of Car yophyllales were also included: Antigonon leptopus (Polygonaceae: noncore Caryophyllales) and Portulaca grandiflora (Portulacaceae: core Caryophyllales).

PAGE 142

142 RNA Isolation and Gene Amplification Total RNA was extracted from RN Aeasy extraction kit (Qiagen, Carlsbad, California, USA ). cDNA was synthesized using Superscript II (Invitrogen Carlsbad, USA ) according to the manufacturers instruction s. Ampli publi shed studies (A -function MIKC: Litt and Irish, 2003), B -function MIKC: Kr amer et al., 1998), C -function MIKC: Stellari et al., 2004), using degenera te PCR thermoc ycling conditions. PCR bands over 500 bp in size were excised from the agarose gel and puri Geneclean II Kit (QBio Gene, Carlsbad, California, USA). Puri TOPO TA Cloning Kit (Invitrogen). Plasmid DNAs were amplif ied using the Templiphi Cycle (Amersham, Piscataway, NJ, USA) and sequences were generated on an ABI 3730 XL DNA sequencer (Applied Biosystems, Inc., Fullerton, CA, USA) following the manufacturers protocol. Alignments Newly isolated genes were identifie d by distinctive C terminal motifs or subject to a BLAST search to preliminarily assign them to a given clade within the MADS box gene family New sequences were then automatically aligned by predicted amino acid sequence using Clustal X (Thompson et al., 1997) with previously published sequences representing all major clades of MIKC genes. RevTrans 1.4 ( http://www.cbs.dtu.dk/services/RevTrans/ ) was employed to generate a corresponding codonaligned D NA alignment. This codonaligned DNA dataset was analysed under the maximum likelihood (ML) criterion (see method section phylogenetic analyses) On the basis of this preliminary phylogenetic analysis distinct clades or monophyletic groups of clades that contained newly isolated sequences were separated into different

PAGE 143

143 alignments to be further analyzed separately. Additional previously published sequences that shared af finity to these clades (based on BLAST results and previous ly published phylogenetic a nalyses) were added to increase taxon sampling in each separate alignment. Within each group, sequences were automatically aligned by predicted amino acid sequence using Clustal X (Thompson et al., 1997). RevTrans 1.4 was employed to generate a correspondi ng codonaligned DNA alignment. These separate alignments with their included gene lineages were as follows: A -function (APETALA1 FRUITFUL and FRUITFUL like homologs ); B function ( APETALA3 TM6 PISTILLATA and B -sister lineages), C and D -function ( AGAMO US and AGAMOUS like 11 lineages); E -function ( SEPALLATA gene lineages). For each group the DNA partition was subject ed to phylogenetic reconstruction by maximum parsimony (MP), maximum likelihood (ML) and Bayesian analyses. Phylogenetic Analyses MP analy ses were implemented in PAUP*4.0 (Swofford, 2000). Shortest trees were obtained using a heuristic search and 1,000 replicates of random taxon addition with tree bisection -reconnection (TBR) branch swapping, saving all shortest trees per replicate. Bootstra p support (BS) for relationships (Felsenstein, 1985) was estimated from 1 000 bootstrap replicates using 10 random taxon additions per replicate, with TBR branch swapping and saving all trees. For ML analyses we employed the program GARLI (Genetic Algorith m for Rapid Likelihood Inference; version 0.942) (Zwickl, 2000). GARLI conducts ML heuristic phylogenetic searches under the GTR model of nucleotide substitution, in addition to models that incorporate among-site rate variation, either assuming a gamma dis tribution ( ) or a proportion of invariable sites (I), or both. Models of nucleotide substitution were determined using

PAGE 144

144 MrModeltest (Nylander, 2004). The Akaike information criterion (AIC) was used to select GTR+I+G as an appropriate model based on the rel ative informational distance between the ranked models. Analyses were run with default options, except that the significanttopochange parameter was reduced to 0.01 to make searches more stringent. ML bootstrap analyses were conducted with the default par ameters and 100 replicates. We performed a strict consensus of five replicate GARLI analyses and topological differences resulting in collapsed nodes were annotated on the representative ML tree. Models of nucleotide substitution were determined using MrM odeltest (Nylander, 2004). The Akaike information criterion (AIC) was used to select GTR+I+G as an appropriate model based on the relative informational distance between the ranked models. Bayesian a nalyses were implemented in MrBayes, version 3.1.2 (Huels enbeck and Ronquist, 2001; Ronquist and Huelsenbeck, 2003). Two independent analyses each ran for 5 million generations, using four Markov chains, and with all other parameters at default values; trees were sampled every thousandth generation, with a burnin of 200,000 generations. Stationarity of the Markov Monte Carlo chain was determined by the average standard deviation of split frequencies between runs (after 5 million generations the average standard deviation was 0.004%) and by examination of the pos terior in Tracer, version 1.3 (Rambaut and Drummond, 2003). A majorityrule consensus of post burnin trees was generated in PAUP*4.0 (Swofford, 2000), using the resulting posterior distribution of the trees. Results and Discussion Caryophyllid MIKC Comple ment ML analysis of the all MIKC data set revealed that 50 distinc t MIKC -type loci representing ten major clades of MADS -box genes had been isolated from seven previously unrepresented t axa within the Caryophyllales (Figure 5 1) Almost all major lineage s of the

PAGE 145

145 ABCDE MIKC -type gene lineages are represented by these caryophyllid loci (with the exception of the eu PLE lineage, (see below). A -Function : Representatives of the eu AP1 and eu FUL lineages were isolated from all 5 species of Aizoaceae and additiona lly two eu FUL -like loci were isolated from D.napiforme and M. cordifolia B -function : Representatives of the eu AP3 PI and TM6 lineages were isolated from all five species of Aizoaceae and two eu AP3 paralogs were isolated from A. leptopus Two additional loci belonging to the B -sister lineage were identified from P. oleracea and T tetragonoides Cand D -function : R epresentatives of the eu AG lineage were recovered from all five species of Aizoaceae and two loci belonging to the AGL11 lineage were isolated from D. napiforme and T. tetragonoides E -function : Three SEP homologs were obtained from two species of Aizoaceae ( D. napiforme, M. cordifolia and T. tetragonoides and in addition four SEP homologs were obtained from A. leptopus Consequently four gene lineages within the SEP sub -family ( AGL2 FBP9 AGL3 and AGL9 ) are represented by caryophyllid loci isolated in this study. The diversity of caryophyllid MADS -box genes isolated here is representative of the expected complement predicted by previous phylog enetic analyses, although this study provides some of the first caryophyllid placeholders in lineages that previously were not represented by loci from the Caryophyllales. These include homologs from non-ABCDE function lineages STMADS11, AGL15 and TM3 P. oleracea MADS1 and T. tetragonoides MADS6 are the first B sister homologs to be isolated from the Caryophyllales while A. leptopus MADS6 is the first caryophyllid placeholder for the FBP9 lineage of the SEP subfamily. On the whole, the inclusion of these 54 new caryophyllid loci does not alter our understanding of the broad relationships between the different gene lineages in which they are included. The relative order of branching between some paralogous lineages remains uncertain

PAGE 146

146 either because deeper level nodes lack support or because there is topological conflict between phylogenetic analyses using different optimality criterion (e.g. contrast the topology of the eu AP1 eu FUL and FUL like lineages in the MP, ML and Bayesian analyses in Figure 5 2; contrast with Litt et al., 2003). While broad taxon sampling, including caryophyllid taxa, is important to resolve these deeper level uncertainties, the most critical taxa to sample are those that are phylogenetically close to point of duplication i.e. mo stly non -core and early diverging eudicot taxa. It is not surprising, therefore, that the inclusion of additional relatively derived caryophyllid loci does not clarify these deeper -level uncertainties. Polygonales -specific D uplication in the Eu AP3 Lineage The isolation of two AP3 paralogs in Antigonon leptopus (Polygonaceae) is notable given that two AP3 paralogs have also been detected in another genus in the Polygonaceae, Rumex acetosa This suggested a possible Polygonales -specific duplication in the e u AP3 lineage. Phylogenetic analysis of this putative duplication is complicated by the much higher molecular substitution rates in the AP3 homologs in R. acetosa versus A leptopus as inferred by branch lengths. The MP analyses suggest two recent separate duplications as the paralogs in each species are recovered as sister to each other i.e. Rumex RAD1 is sister to Rumex RAD2 and Antigonon MADS2 is sister to Antigonon MADS4 (Figure 5 3a). In contrast the ML and Bayesian approaches suggest an older Polygonac eae -specific duplication with Rumex RAD1 sister to Antigonon MADS2 and Rumex RAD2 sister to Antigonon MADS4 (Figure 5 -3b and 5 3c). The ML and Bayesian topologies would seem more credible given the inherent difficulties that long branches pose for MP phylogenetic analyses and the better performance of model -based ML and Bayesian approaches in situations of substitution rate heterogeneity (Gadagkar and Kumar, 2005). On the other hand, analysis of the RAD1 and RAD2 paralogs finds little evidence for neo or su bfunctional ization at the mRNA expression level as b oth paralogs are apparently id entically

PAGE 147

147 expressed in both staminate and carpellate flowers of Rumex acetosa (Ainsworth et al., 1995). It is therefore currently unclear what selective forces are maintainin g the co -existence of these paralogs if they are derived from an older duplication event. Absence of the eu PLE Lineage in Caryophyllales Although representatives of most of the expected eudicot -specific lineages of ABCDE MIKC type genes were isolated in t his study, one notable exception is the absence of the eu PLE lineage homologs in Caryophyllales (Figure 5 6). Additionally, eu PLE homologs have not been isolated by five previous studies that report the isolation of eu AG homologs which constitute the siste r lineage to eu PLE A total of 11 caryophyllid taxa are therefore represented by eu AG homologs, and yet no eu PLE homologs have been detected. There are three possible explanations for this: 1) the eu PLE homologs in Caryophyllales may have diverged in seque nce such that they were un amplifiable by the primers used in this and previous studies. This would seem unlikely as the primers used were able to amplify more distant related AGL11 homologs within the AGAMOUS subfamily ( Delosperma MADS8 and Tetragonia MAD S17 in this study; Phytolacca AG2, (Kramer et al., 2004). 2) The eu PLE and eu AG lineages derive from a duplication that occurred after the divergence of the Caryophyllales within the core eudicots. This scenario remains possible as it currently unclear whe re the duplication giving rise to eu AG and eu PLE lineages occurred. Only rosid and asterid lineages possess both paralogs suggesting that the duplication may have arisen just prior to the divergence of the rosid and asterid lineages; uncertainty as to the branching order of major eudicot lineages further complicates evaluation of this scenario. Additionally EM Kramer reports a possible PLE homolog in the early-diverging eudicot lineage Gunnera (pers comm.), which would make the following third explanation m ore probable. 3) The duplication giving rise to the euAG and eu PLE lineages occurred at the base of the core eudicots as suggested by Kramer et al., (2004) but the eu PLE lineage was subsequently

PAGE 148

148 lost in the Caryophyllales. As highlighted previously there i s widespread evidence for instances of paralog loss in lineages of MIKC -type gene lineages; in the AGAMOUS subfamily, a notable example is the apparent loss of AGL11 lineage homologs in the Ranunculales (Kramer et al 2004). This hypothesis would need to be explored at the genomic level to demonstrate that the PLE loci have indeed been lost. Absence of Duplications Coinciding with the Inferred Origin of Petaloid Staminodes It is evident from these phylogenetic analyses that there have been no duplications in the MADS -box gene lineages associated with the inferred point of origin of petaloid staminodes. In particular, there is no increase in gene complement in the AGAMOUS APETALA3 and PISTILLATA lineages, which, due to their roles in stamen and petal identi ty, are of particular interest with respect to the evolution of petaloid staminodes of Ruschoideae and Mesembryanthoideae. We therefore reject the hypothesis that the evolution of petaloid staminodes could be linked to duplication and subsequent neofunctionalization in key MADS box gene lineages.

PAGE 149

149 Figure 5 1. ML tree derived from analysis of a DNA dataset comprising 121 MADS box gene loci (including 54 loci isolated by this study) Nodes with circles c ollapse in strict consensus of five trees; numbers above branches represent bootstrap support values, red branches leading to terminals in bold text signify loci isolated by this study. Ten major clades are revealed and labeled based on the nomenclature in (Becker and Theissen, 2003).

PAGE 150

150 Figure 5 2. Trees deri ved from analyses of the A -function DNA dataset comprising 82 MADS box gene loci (including 12 loci isolated by the study). Lineages coded by organismal affinity: gymnosperms (black), magnoliids and basal angiosperms (red), monocots (yellow), noncore eu dicots (pink), core eudicots (blue), caryophyllids (green). Loci isolated in this study are highlighted in bold text. a) Single MP tree, numbers above branch es are bootstrap support values. b) ML tree; numbers above branches are bootstrap support values; N odes with circles collapse in strict consensus c) 50% majority rule consensus tree derived from post burnin posterior distribution of the trees derived from bayesian analysis. Clades are labeled based on the nomenclature in (Litt and Irish, 2003)

PAGE 151

151 Fig ure 5 2. Continued

PAGE 152

152 Figure 5 3. Trees derived from analyses of the B -function DNA dataset comprising 127 MADS -box gene loci (including 19 loci isolated by the study). Lineages coded by organismal affinity: gymnosperms (black), magnoliids and basal angiosperms (red), monocots (yellow), non-core eudicots (pink), core eudicots (blue), caryophyllids (green). Loci isolated in this study a re highlighted in bold text. a) one of five MP trees, numbers above branches are bootstrap support values b) ML tree; nu mbers above branches are bootstrap support values; n odes with circles collapse in strict consensus c) 50% majority rule consensus tree derived from post burnin posterior distribution of the trees derived from bayesian analysis. Clades are labeled based on the nomenclature in (Stellari et al., 2004)

PAGE 153

153 Figure 5 3. Continued

PAGE 154

154 Figure 5 4. Trees derived from analyses of the APETALA3 DNA dataset comprising 69 MADS box gene loci (including ten loci isolated by the study). Lineages coded by organismal affin ity: magnoliids and basal angiosperms (red), monocots (yellow), non-core eudicots (pink), core eudicots (blue), caryophyllids (green). Loci isolated in this study ar e highlighted in bold text. a) one of two MP trees, numbers above branches are bootstrap su pport values b) ML tree; numbers above branches are bootstrap support values; n odes with circles collapse in strict consensus c) 50% majority rule consensus tree derived from post burnin posterior distribution of the trees derived from bayesian analys is. Clades are labeled based on the nomenclature in (Stellari et al., 2004)

PAGE 155

155 Figure 5 4. Continued

PAGE 156

156 Figure 5 5. Trees derived from analyses of the PISTILLATA DNA dataset comprising 47 MADS box gene loci (including 6 loci isolated by the study). Lineage s coded by organismal affinity: magnoliids and basal angiosperms (red), monocots (yellow), non-core eudicots (pink), core eudicots (blue), caryophyllids (green). Loci isolated in this study ar e highlighted in bold text. a) one of 13 MP trees, numbers above branch es are bootstrap support values b) ML tree; numbers above branches are bootstrap support values; n odes with circles collapse in strict consensus c) 50% majority rule consensus tree derived from post burn-in posterior distribution of the trees de rived from bayesian analysis. Clades are labeled based on the nomenclature in (Stellari et al., 2004).

PAGE 157

157 Figure 5 5 Continued

PAGE 158

158 Figure 5 6. Trees derive d from analyses of the C and D -function DNA dataset comprising 104 MADS -box gene loci (including nine loci isolated by the study). Lineages coded by organismal affinity: gymnosperms (black), magnoliids and basal angiosperms (red), monocots (yellow), non-core eudicots (pink), core eudicots (blue), caryophyllids (green). Loci isolated in this study ar e hig hlighted in bold text. a) one of 16 MP trees, numbers above branch es are bootstrap support values b) ML tree; numbers above branches are bootstrap support values; n odes with circles collapse in strict consensus c) 50% majority rule consensus tree derived from post burn-in posterior distribution of the trees derived from bayesian analysis. Clades are labeled based on the nomenclature in (Kramer et al., 2004; Zahn et al., 2006)

PAGE 159

159 Figure 5 6. Continued

PAGE 160

160 Figure 5 7. Trees derived from analyses of the E -function DNA dataset comprising 151 MADS -box gene loci (including eight loci isolated by the study). Lineages coded by organismal affinity: gymnosperms (black), magnoliids and basal angiosperms (red), monocots (yellow), non-core eudicots (pink), core eudi cots (blue), caryophyllids (green). Loci isolated in this study ar e highlighted in bold text. a) one of four MP trees, numbers above branch es are bootstrap support values b) ML tree; numbers above branches are bootstrap support values; n odes with circles collapse in strict consensus c) 50% majority rule consensus tree derived from post burnin posterior distribution of the trees derived from bayesian analysis. Clades are labeled based on the nomenclature in (Zahn et al., 2005)

PAGE 161

161 Figure 5 7. Continued

PAGE 162

162 CHAPTER 6 PERIANTH EVOLUTION I N AIZOACEAE: NOT ALL CORE EUDICOT PETALS WERE CREATED EQUAL The existence of similar developmental genetic pathways in the petals of distantly related eudicot model organisms led to the proposal of a common petal identity progr am (Bowman 1997; Ambrose et al. 2000). According to this hypothesis, a petal identity program, regulated by APETALA3 and PISTILLATA MADS -box genes, evolved to specify petals or petaloid morphology in an early angiosperm ancestor. This hypothesis is attract ive from the evolutionary standpoint as a petal identity program could engender organ identity independently of position in the flower or historical homology of the organs. The observation that petaloid organs can occupy both first and second whorls of the flower and can be derived from different organs (bracts or stamens) could therefore be explained by simple heterotopic changes in the activity of an ancestral petal identity program. We test the concept of a common petal identity program in the Caryophyll id family Aizoaceae, which possesses two distinct petaloid organs, petaloid tepals and petaloid staminodes derived from bracts and stamens respectively. We assess to what extent the morphology of these organs in Aizoaceae can be attributed to their differi ng historical derivation and whether they share similar morphological features indicative of a common petal identity pathway. Furthermore we analyze the gene expression patterns of three MADS box gene homologs APETALA3 PISTILLATA and AGAMOUS to assess the genetic evidence for a common petal identity program. We find little evidence at the morphological or genetic level for a shared petal identity program between petaloid tepals and staminodes. Moreover the petaloid organs in Aizoaceae exhibit unusual expre ssion patterns of MADS box genes, unlike that described in other core eudicot petals. We discuss these findings in comparison to developmental genetic data from other angiosperm taxa, and in the context of perianth evolution in the Caryophyllales. We sugge st that both petaloid organs have unique genetic and morphological

PAGE 163

163 attributes, implying that petaloid morphology has evolved independently of an ancestral petal identity program. Introduction Petals are the showy attractive organs that comprise the second whorl in the perianth of most core eudicot flowers. Several similarities in the genetic regulation of the second whorl petal (Coen and Meyerowitz, 1991) were revealed by comparison of two distantly related eudicot taxa Antirrhinum majus and Arabidopsis thaliana First, in the absence of the C -function MADS box gene AGAMOUS (AG ) ( PLENA in A. majus ) co activity of B -function MADS box genes APETALA3 (AP3 ) and PISTILLATA ( PI ) ( DEFICIENS and GLOBOSA respectively in A. majus ) is necessary for development of the petal in the second whorl of the flower (Bradley et al., 1993; Goto and Meyerowitz, 1994; Jack et al., 1992; Mizukami and Ma, 1992); Second, persistent coactivity of B class MADS -box genes through late stages of petal development is necessary to maintain t he expression of characteristic petal features (Bowman et al., 1991; Sommer et al., 1991; Zachgo et al., 1995). Third, heterotopic expression of B -function MADS box genes in the first whorl of the flower is sufficient to induce ectopic petal morphology in the sepals (Coen and Meyerowitz, 1991; Krizek and Meyerowitz, 1996). From these observations it has been inferred that the common ancestor of A. majus and A. thaliana possessed a petal development pathway regulated by orthologs of AP3 / PI and DEF / GLO Furth ermore, this has led to the suggestion of an ancestral petal identity program conserved at least in core eudicots and possibly across angiosperms (Kramer and Jaramillo, 2005): a concept that has been strengthened by additional taxon sampling. AP3 and/or PI are implicated in the development of petals in additional angiosperm taxa; e.g., Aquilegia (Kramer et al., 2007), Aristolochia (Jaramillo and Kramer, 2004) Gerbera (Yu et al., 1999) Lycopersicon (de Martino et al., 2006) Magnolia (Kim et al.,

PAGE 164

164 2005) Ni cotiana (Liu et al., 2004) Papaver (Drea et al., 2007), Petunia (Vanderkrol et al., 1993) and Silene (Hardenack et al., 1994) Central to the concept of a conserved petal identity program is an assumption that the role of B -class genes in petal development is conserved because, in the absence of C -class gene activity, B -class gene function is required for appearance of petal characteristics such as pigmentation, texture, ultra -structure, and epidermal cell shape. In other words, activity of B class gene s in petals is conserved in distantly related taxa because B -class gene orthologs evolved to regulate, in a common ancestor, a genetic program that engenders the expression of petal features (Baum and Whitlock, 1999; Kramer and Jaramillo, 2005). Conserved AP3 and PI homolog co activity in second -whorl petals could also be described as process homology sensu Gilbert and Bolker (2001). The concept of the petal identity program rests on the idea that the process homology of petals is best explained in correlat ion with the morphological homology of petals i.e. the appearance of homologous petal traits. However morphological homology, corresponding to special characteristics sensu Remane (1952), is only one criterion by which petals may be considered homologous Additional criteria include: 1) positional homology sensu Remane (1952), i.e., petals in core eudicots and other angiosperms may be homologous by the criterion of positional homology as they commonly occupy the same position in the second whorl of the fl ower and; 2) historical homology sensu Mayr (1982) i.e. core eudicot petals may be considered homologous if they are derived from the same structure that is present in the common ancestor (e.g. andropetals derived from stamens). A key question then is th e extent to which the process homology of petals can be explained by morphological homology as opposed to positional or historical homology.

PAGE 165

165 Morphological homology, inherent in the concept of an angiosperm -wide petal identity program, has dominated the in terpretation of comparative MADS box gene expression data (e.g. Kim et al., 2005). However, both positional and historical homology criteria have also been invoked to explain the particular pattern expressions of MADS box genes For example, in an analysis of the evolution of the lodicules in Poaceae, Whipple et al. (2007) invoked positional homology and rejected morphological homology when they argued, B class genes specify second whorl identit y in many angiosperms [despite the fact] that the details of s econd whorl morpholog y are variable In a similar rejection, Drea et al. (2007) contended that the common role for AP3 and PI is in specifying a spatially limited regional domain [the second and third whorls] within the flower. In contrast, Kramer and Irish (2000) appeal to historical homology when they suggested that the expression of AP3 and PI homologs in stamens makes it likely that andropetals [stamen -derived petals] would also utilize these genes in their development. In the context of the gynos temium in Aristolochia Jaramillo and Kramer (2004) also invoke historical homology when they suggest that B -class genes may serve as molecular markers for staminally derived tissue. Finally, Chanderbali et al. (2006) suggested that the expression of AG in the second whorl of the perianth in Persea is a genetic footprint resulting from the historical derivation of the second whorl tepals from stamens. That these in terpretations are competing or co -existing in the literature reflects a number of issues: 1) p etal identity is difficult to define from a comparative morph ological standpoint (Endress, 1996; Jaramillo and Kramer, 2004; Kramer and Jaramillo, 2005); 2) v ariable floral structure merosity and phyllotaxy obscure s positional correspondence between f loral organs (Endress, 1996); 3) historical homology of the perianth can be difficul t to determine and the criteria for doing so have been questioned (Ronse De Craene, 2007, 2008); 4) l imited taxon

PAGE 166

166 sampling forces comparative a nalysis across vast phylogene tic distance over which our ability to define the nature of the homology between the perianth of distant taxa is diminished (Jaramillo and Kramer, 2007); 5) arguments that seek to explain process homology through the positional, historical, and morphologic al homology of the second perianth whorl are not exclusive as each may have explanatory power in different morphological structures, taxonomic groups, phylogenetic levels and episodes of angiosperm evolution. As subsequently discussed, however, explanation s of process homology based on different homology criteria have distinct implications for our understanding of morphological evolution. If process homology in the petals of angiosperms is explained through association with appearance of petal characteristi cs, then an ancestral petal identity program becomes a viable hypothesis. A petal identity program is appealing in the context of petal evolution because, as suggested by genetic manipulation in model organisms, it could operate in a homeotic fashion and engender organ identity independently of position in the flower or historical homology of the organ (Kramer and Jaramillo, 2005). The presence of petals in both first and second whorls of the flower and their derivation from different organs (e.g. bracteo petals from bracts and andropetals from stamens Tahktajan, 1991) could consequently be explained through heterotopic changes in the activity of a petal identity program (Kramer and Jaramillo, 2005). An an cestral petal identity program therefore advocat es parallelism ( the independent evolution of the same derived trait via the same developmental changes sensu Patterson, 1982) and makes evolutionary transitions between different organs a more simple matter than could be envisioned by models of gradual modi In contrast, explanations of process homology in petals based on positional or historical homology allow for convergence (the evolution of morphologically similar traits with a distinct

PAGE 167

167 developmental basis, sensu Patte rson 1982). I f process homology in the second perianth whorl is better explai ned through positional homology then the repeated role of AP3 and PI homologs in d irecting petal development could be due to repeated independent recruitment of the gen es through position -dependent developmental constraint (Drea et al., 2007; Kramer et al., 1998) In this scenario, second whorl organs may share process homology because of position, yet similar morphological appearance would not necessarily be due to this process h omology, but could be the result of convergent morphological differentiation. Similarly if historical homology of organs in the second whorl best explains observed process homology then the conserved function of AP3 and PI homologs could be attributed to a common derivation of s econd -whorl organs from stamens ( whose development is also dependent in part on the activity of AP3 and PI homologs, Kramer et al., 1998). This interpretation would predict conserved AP3 and PI homolog activity in all stamen -derived petals (but not necessarily bract -derived petals) yet allow morphological similarity to be the result of convergent differentiation of developmental processes downstream of AP3 and PI In the face of these competing interpretations, evaluating the concept of a petal identity program requires perianth variation with a pattern of character change that can be accurately reconstructed across a well resolved organismal phylogeny. In addition, it requires an experimental system in which it is possible to assess perianth variation clearly from the perspective of several homology criteria. The angiosperm family Aizoaceae (Caryophyllales) is one such system, exhibiting two distinct types of perianth. The subfamilies Sesuvioideae and Aizooideae both possess a periant h that is simple, comprising organs (termed petaloid tepals) that are petaloid on the adaxial side and sepaloid on the abaxial side. In contrast, the perianth of both the subfamilies Ruschoideae and Mesembryanthemoideae is differentiated, with an outer

PAGE 168

168 whorl of sepals and an inner whorl of petals (termed petaloid staminodes). The term petaloid is applied here as a homologyneutral term that simply describes the superficially similar appearance of these organs to each other and to petals in other core eudi cots, without reference to specific characters or homology criteria. Molecular phylogenetic studies indicate that Sesuvioideae and Aizooideae form an earlydiverging grade within Aizoaceae (Klak et al., 2003) such that the differentiated perianth of Rusch oideae and Mesembryanthemoideae is derived (Brockington et al., 2009). Importantly, the petaloid perianth in the early -diverging subfamilies is thought to derive from bracteal organs and thus is not homologous in a historical sense to what are considered t o be stamen -derived petaloid staminodes present in the derived subfamilies. Furthermore the sepals in derived sub-families are derived from the petaloid tepals of early -diverging subfamilies although the petaloid appearance of the tepals has been lost in the evolution of the differentiated perianth in Ruschoideae and Mesembryathoideae, in which petaloid staminodes have assumed the attractive role. The above pattern of perianth evolution makes Aizoaceae a valuable system in which to assess the concept of th e petal identity program against an underlying conflict in historical homology. In this study we: (1) confirm the different historical derivation of the two types of petaloid organ (2) assess the extent to which the differing derivation of the petaloid organs influences their morphological appearance; (3) examine whether historically non-homologous petaloid organs share common characteristics consistent with a morphological concept of shared petal identity; (4) examine the expression patterns of homologs of the organ identity genes, AGAMOUS (AG ) APETALA3 (AP3 ) and PISTILLATA ( PI ) to assess whether the differing petaloid organs share a common developmental genetic program and; (5) compare these findings

PAGE 169

169 with data from core eudicot model organisms and other angiosperm taxa to assess the applicability of a conserved petal identity program to Aizoaceae. Methods Plant Materials Mesembryanthemum cordifolia Delosperma napiforme Tetragonia tetragonoides were grown in a greenhouse at the University of Florida. Flo wering material of Sesuvium portulacastrum was collected from Cedar Key, Florida. Vouchered specimens are deposited at the University of Florida herbarium (FLAS). Scanning Electron Micro scopy Ontogenetic analysis was performed on S. portulacastrum and D. napiforme to investigate the respective homology of the perianth organs and to provide a morphological framework to interpret RNA RNA in -situ hybridizations. Developing inf lorescences were dissected and fixed in f resh ly prepared FAA (3.7% formalde hyde, 50% ethanol, 5% acetic acid). Samples were dehydrated through absolute ethanol and critical -point dried using an Autosamdri 815B CPD (Tousimis Research, Rockville, Maryland, USA), then coated with platinum using an Emitech (Kent, UK) K550 sputter coater and e xamined using a Hitachi (Wokingham, UK) coldemission SEM S4700 II. Some images were colored using Adobe Photoshop (San Jose, California, USA) RNA -RNA In -s itu Hybridisation RNA -RNA in -situ hybridizations were performed on two species in Aizoaceae re presenting two perianth types; S. portulacastrum and D. napiforme. Developing inf lorescences were dissected and fixed in f resh ly prepared FAA (3.7% formalde hyde, 50% ethanol, 5% acetic acid) overnight at 4oC. Anti -sense and sense probes were designed to t he C terminus and 3UTR region of the homologs of AP3 (SpAP3 and DnAP3 ), PI (SpPI and DnPI ) and AG (SpAG and

PAGE 170

170 DnAG ) and synthesized using a DIG RNA labeling kit with a T3 polymerase (Roche, Indianapolis, USA). Probes were hydrolyzed to approximately 150bp. Probe hybridization, post hybridization treatment and immuno localization was performed as previously described (Malcomber and Kellogg, 2004). Colorimetric staining was performed with NBT/BCIP for 1 2 and slides were photographed using a Zeiss Axiocam MRc5 digital camera mounted on a Zeiss Axioskop 2 Plus microscope (Carl Zeiss MicroImaging, Thornwood, NY) with bright illumination. Relative Quantitative RT -PCR Relative Quantitative RT PCR (RQ RT -PCR) was performed in four species representing all four subfamilies in Aizoaceae: S. portulacastrum (Sesuvioideae), T. tetragonia (Aizooideae), Mesembryanthemum cordifolia (Mesembryanthoideae) and D. napiforme (Ruschoideae). Flowers were dissected into separate pools of organs prior to RNA extraction ( S. portulacastrum / T. tetragonia; Vegetative Leaves, Petaloid Tepals, Stamens, and Carpels: Mesembryanthemum cordifolia / D napiforme ; Sepals, Petaloid Staminodes, Stamens, and Carpels). For S. portulacastrum and D. napiforme prior to organ dissection, flowers were separated into different size pools (<4mm, 5 8mm, >10mm) in order to assess gene expression levels at different stages of development. Only mature flowers of T. tetragonia were employed in RQ -RT -PCR analyses due to the small size of the flowers so expression data from T. tetragonia only represents late -stage gene expression. In the case of M cordifolia flowers of different sizes were admixed prior to cDNA manufacture so expression data also predominantly represents late stage gene expression. RNA extraction and cDNA manufacture was performed as in Chapter 5. RNA was normalized between different pools using a Nanodr op (Thermoscientific, Delaware, USA) prior to cDNA manufacture. We performed RQ RT PCR using a gene -speci (designed to AG, AP3 and PI homologs) and using a specific primer pair to the 18S rRNA gene

PAGE 171

171 following the protocol of QuantumRNA (Ambio n California, USA ). The 18S rRNA gene was used as a control. All RT -PCR products were confir med initially by size and subsequently by direct sequencing Products were run on a 1.5% agarose gel and digitally photographed. Results Floral Ontogeny in Sesuvi um portulacastrum Tepal primordia arise in a quincuncial 2/5 arrangement (Figure 6 1, A F). The primordia are crescent shaped with a broad base of insertion (Figure 6 1, A, B). The gynoecial primordium becomes visible prior to the emergence of distinct and roecial primordia. Androecial primordia arise in a somewhat chaotic fasciculated pattern with a tendency to centrifugal initiation (Figure 6 1, I, J, K). The gynoecium develops with three carpels. Outer stamens develop and hide the inner stamens and develo ping carpel later in development (Figure 6 1, K, L). The tepal primordia differentiate into an upper and lower zone early in development (Figure 6 3, A). In S. portulacastrum an adaxial cross -zone gives rise to a ligule that marks the boundary between the upper and lower zones (colored green and pink, respectively, in Figure 6 3), allowing the two zones to be tracked through the ontogeny of the tepal. Early in development, the unifacial upper zone comprises the bulk of the tepal (Figure 6 3, B). Subsequent ly, however, the bifacial lower zone expands to form the bulk of the mature tepal (Figure 63, C H). The petaloid lamina of the tepal is therefore derived from the lower zone while a distal unifacial tip represents the upper leaf zone (Figure 6 3, G). The derivation of the petaloid portion of the tepal from the lower leaf zone suggests homology between the lamina of the tepal and the leaf sheath of the vegetative leaf (Figure 6 3, L), which is also derived from a lower leaf zone. In turn, the lamina of the vegetative leaf (Figure 6 3, I), derived from the upper zone, is represented in the tepal by the reduced unifacial distal tip.

PAGE 172

172 The correspondence between leaf sheath and perianth reveals that several morphological features in the perianth of S. portulacast rum are also present in the vegetative l eaf sheath. 1) The perianth in S. portulacastrum is singular in its strong abaxial/adaxial differentiation, a condition that is also found in t he leaf sheath (cf. Figure 6 1, M, N a nd Figure 6 3, G, H, J, K). 2) The epidermis of the showy adaxial tissue in the perianth has a simple convex cell type that initially seems unique within the context of the flower but is actually identical to the epidermal cell type found on the adaxia l surface of the leaf sheath (c f Figu re 6 1, M and Figure 63, M). 3) The distinctive flanges of delicate petal tissue, seen easily from the abaxial s ide of the perianth (Figure 6 3, H), correspond to the hyaline margins of tissue i n the leaf sheath (Figure 6 3, K). Floral Ontogeny in Delosp erma napiforme Sepals arise in a quincuncial 2/5 arrangement in D. napiforme (data not shown). Removal of the sepals reveals that five carpel primordia comprising the gynoecium are visible prior to the emergence of the initial androecial primordia, which arise in alternation with the carpel primordial (Figure 6 2, A, B). Further primordia arise outside of these inner androecial primordia in a centrifugal direction (Figure 6 2, C). Only the innermost primordia develop into fertile stamens (Figure 6 2, F, G); in a mature flower the fertile stamens form a single ring around the gynoecium Figure 6 2, I, J, K). Subsequent centrifugally initiating primordia develop into sterile staminodes or petaloid staminodes (Figure 6 2, G, H, I); early in development these sta minodes resemble stamens (Figure 6 2, L). Centrifugally initiated primordia are increasingly petaloid the further they are initiated from the inner fertile stamens (Figure 6 2, I). The petaloid staminodes are broad relative to the filament of the stamens a nd cover all fertile organs in development (Figure 6 2, I). The epidermal cells on the petals are undifferentiated (Figure 6 2, M) and resemble the epidermal cells on the stamen filament (Figure 6 2, N).

PAGE 173

173 Expression of AGAMOUS Homologs In general, expressi on of AG homologs was strong and easily detected by RNARNA in situ hybridisation. In S. portulacastrum SpAG is expressed early in the development of the androecium and gynoecium and continues through to the late stages of stamen and carpel differentiatio n (Figure 6 4, A -F). SpAG is also expressed in developing microand mega sporangia (Figure 6 4, F). In D. napiforme DnAG is expressed in the meristem prior to the emergence of androecial and gynoecial primordia (Figure 6 4, G). Expression is strong in th e developing carpel and in all members of the androecium: it is strongly expressed in the primordia that eventually form sterile staminodes as well as in fertile stamen primordia (Figure 6 4, H, I). Sections that simultaneously capture staminodes at varying stages of maturation reveal the dynamic change in expression of DnAG (Figure 6 4, I, J, K). In sterile staminodes Dn AG is expressed throughout the organs until after they have reached approximately 100 M in length (Figure 6 4: J, staminode 3). After this stage in staminode maturation, expression of DnAG appears to become restricted in a distal direction until expression is restricted to the tip of the maturing sterile staminode (Figure 6 4, J, staminode 2; K, staminode 3). Later in development this rest ricted expression is lost (Figure 6 4, J, staminode 1, K, staminode 2 and 3). DnAG is strongly expressed in the ovary wall and the developing pollen and ovules (Figure 6 4, J, K, L). RQ -RT -PCR mostly supports these data. Expression of DnAG is strong throughout stamen and carpel development, but although initially expressed in petaloid staminodes, expression weakens and is lost by the time flowers are greater than >10mm in diameter (Figure 6 7). ApAG is strongly expressed in stamens and carpels but not dete cted in the petaloid staminodes possibly due to the late stage flowers that were used to manufacture cDNA pools from M. cordifolia

PAGE 174

174 SpAG is expressed strongly throughout development of stamens and carpels and TtAG is strongly expressed in late -stage stamen and carpels of T tetragonoides (Figure 6 7) Expression of PISTILLATA Homologs In general, expression of PI homologs was strong and easily detected. In S. portulacastrum SpPI expression is restricted to the androecium and absent from the developing carpe ls and tepals (Figure 6 5, A F). Expression in the androecium persists through the late stages of stamen differentiation and is present in both filament and differentiated anther (Figure 6 5: E, F). SpPI expression is absent from developing ovules (Figure 6 5; F). In D. napiforme DnPI expression is restricted to androecial primordia giving rise to fertile stamens and staminodes (Figure 6 5; G). Expression is initially strong in all androecial primordia. In fertile stamens, expression is maintained through differentiation of the filament and anther (Figure 6 5; I, K). Expression remains strong in the developing filament; within the anthers strong expression is restricted to the anther locule wall (Figure 6 5; I, K). In contrast, in primordia giving rise to outer staminodes, DnPI expression is initially strong but rapidly weakens in maturation (Figure 6 5; J, c.f. staminodes 1, 2 and 3) as expression becomes restricted in a distal direction (Figure 6 5; K c.f. staminodes 2, and 3) (in a similar manner to DnAG ) Expression is lost by the time the staminodes reach ~300 M in length (Figure 6 5, K staminode 1). DnPI expression is absent from the ovules (Figure 6 5; I and K). RQ -RT -PCR supports these data (Figure 6 7). DnPI is expressed strongly in stamens and p etaloid stamens; expression weakens throughout petaloid staminodes but is maintained to late stages in stamen development. AcPI is expressed in stamen and petaloid staminodes. SpPI is expressed throughout stamen development and TtPI is expressed in stamens of late stage flowers.

PAGE 175

175 Expression of APETALA3 Homologs In general, the expression of AP3 homologs was weaker and harder to detect than the expression of AG and PI homologs. In S. portulacastrum SpAP3 expression is restricted to the androecium and absent from the developing carpels and tepals (Figure 6 6, A -F). Expression in the androecium persists through the late stages of stamen differentiation and is present in both filament and differentiated anther (Figure 66, D, E). SpAP3 expression is strong in de veloping ovules (Figure 6 6, F). In D. napiforme Dn AP3 expression is restricted to androecial primordia giving rise to fertile stamens and staminodes (Figure 6 6, G) and to the ovules. Expression is initially strong in all androecial primordia. In fertile stamens, expression is maintained through differentiation of the filament and anthers (Figure 6 6, I, J). In contrast, in primordia giving rise to staminodes, Dn AP3 expression is initially strong but weakens in maturation of staminodes (Figure 6 6, J, cf. staminodes 1, 2 and 3), and expression becomes restricted in a distal direction (Figure 6 5, K cf. staminodes 1, 2, and 3) (in a similar manner to DnAG ). Expression is lost by the time the staminodes reach ~300 M in length (Figure 6 5, K staminode 1). D nAP3 is strongly expressed in the ovules (Figure 6 5, I and K). RQ -RT -PCR supports this data (Figure 6 7). DnAP3 is expressed throughout stamen development but although initially strongly expressed in petaloid staminodes, subsequently weakens through deve lopment. Weak DnAP3 expression is also detected in carpels, consistent with expression in ovules detected by RNA -RNA in situ hybridization. AcAP3 is detected in petaloid staminodes and stamens, with weak expression in carpels. SpAP3 and TtAP3 are both stro ngly detected in stamens with weak expression in carpels.

PAGE 176

176 Discussion Perianth Ontogeny and Historical Homology of Petaloid Organs in Aizoaceae The distinct historical homology of the petaloid organs in Aizoaceae is an essential experimental feature of Aizo aceae as a system for testing the concept of the petal identity program. Here we summarize the evidence supporting the different derivation of petaloid tepals (in Sesuvioideae and Aizooideae) and petaloid staminodes (in Mesembryanthoideae and Ruschoideae). The petaloid tepals of Sesuvioideae and Aizooideae meet traditional criteria of bract derived petals, as the tepal primordia are crescent -shaped, have a broad base at insertion (Figure 6 1, B D), and are supplied by three vascular traces (Hofmann, 1994) Ontogenetic analysis of the perianth in S. portulacastrum further supports a leaf -like development of petaloid tepals (Figure 6 3, A L). Tepal primordia differentiate into an upper and lower zone; such differentiation is a well -established first stage in the morphogenesis of the leaves of many species of angiosperms (Kaplan, 1973). The relative expansion of these two zones reveals that the petaloid lamina of the tepal is derived from the lower zone and thus corresponds to the leaf sheath in the vegetative leaves of S. portulacastrum The sepals of Mesembryanthemoideae and Ruschoideae are clearly homologous to the tepals in Sesuvioideae and Aizooideae. Sepal primordia in Mesembryanthemoideae and Ruschoideae exhibit similar differentiation into upper and low er zones (Payer, 1857) (although the lower zone does not undergo much expansion in development), occupy a similar position relative to the floral bracts, and have a similar quincuncial aestivation (Payer, 1857). In contrast to the tepals of Sesuvioideae and Aizooideae, however, the adaxial surface of the sepals of Mesembryanthemoideae and Ruschoideae is not petaloid.

PAGE 177

177 The petaloid staminodes of Mesembryanthemoideae and Ruschoideae can be considered stamen -derived, according to traditional criteria, as they p ossess a singular vascular trace (Hofmann, 1994) and a narrow point of insertion (Figure 6 2, L). In addition, several lines of evidence peculiar to floral development in Aizoaceae support the interpretation of stamenderived petaloid structures in Mesembr yanthemoideae and Ruschoideae. 1) both petals and stamens develop from primordia initiating in a centrifugal direction (Figure 6 2, A H) (Hofmann, 1994). 2) Centrifugal initiation of the androecium also occurs in the early-diverging subfamily Aizooideae (e.g. in the genus Gunniopsis ), however in Aizooideae all centrifugally initiating primordia develop into fertile stamens (Hofmann, 1994). 3) Within the genus Delosperma positionally homologous primordia can develop into petaloid staminodes or stamens, depe ndent on the species. For example, in D. napiforme all but the innermost primordia develop into petals (Figure 6 2, F, I, J, K), but in D. apetala only the outermost primordia develop into petals. 4) Stamens and petals are conceptually linked by intermedia tes, as floral organs developing closest to the fertile stamens are increasingly filamentous while outermost organs are increasingly petaloid (Hofmann, 1994). 5) The petaloid staminodes resemble stamens early in development (Figure 6 2, L). Petal Identit y: A Morphological or Functional Concept? The concept of petal identity, suggested by the ectopic expression of petal characteristics in gain -of -function mutants, argues for a floral -specific genetic program that turns on a suite of petal characteristics. As a comparative concept, however, petal identity is problematic. There are no diagnostic morphological features that can define a comparative concept of petal identity across angiosperms (Kramer et al., 2007; Whipple et al., 2007), as the characteristic s commonly attributed to petals are not always present, do not always co -occur, and are not always restricted to petals (Ronse De Craene, 2007, 2008). Petal identity is therefore applied to a diverse range of

PAGE 178

17 8 variable perianth organs in absence of definabl e or consistent morphological similarity. We argue that in the absence of a strict demonstration of morphological homology, the comparative concept of petal identity is in general based on functional analogy. This perceived functional analogy is then maske d by the use of intangible pseudo-morphological terms such as petaloidy and showiness. That similar function does not necessarily reflect underlying homology is well documented by numerous examples of non -homologous yet functionally similar structures (e.g. wings of bats and flies, Bolker and Raff, 1996). Consequently, functional similarity can obscure underlying differences and has received rigorous criticism (Bolker and Raff, 1996). Such criticism is echoed in our analysis of the morphology of the pe rianth in Aizoaceae, which finds little morphological evidence to support homology between functionally equivalent and superficially similar petaloid organs found in different sub -families. Implications of Correspondence between Leaf Sheath and Petaloid Tepal In assessing the morphological concept of petal identity in Aizoaceae, we need to distinguish between morphological characters that have bearing on a comparative concept of petal identity, versus those that are related to the historical homology of th e petaloid organ. Morphology that can be attributed to historical homology of the organ is difficult to incorporate within the conceptual framework of petal identity, because petal identity has been hypothesized to operate independently of the derivation of the floral organ (Kramer and Jaramillo, 2005). In practice it is easy to conflate historical and morphological homology, but the need for distinction is illustrated by the correspondence of the petaloid tepal and leaf sheath in S. portulacastrum Viewed from the perspective of the organ identity concept, the petaloid tepal is chimeric with adaxial petal identity and abaxial sepal identity. However, with the exception of pigmentation, we find little tepal -specific morphological differentiation relative to the leaf sheath; both have the same strong adaxial/abaxial differentiation, epidermal cell morphology, and delicate margins (Figure 6 3, cf.

PAGE 179

179 G, H, J, K). We suggest therefore that the chimeric morphology of the tepal should not be interpreted within the fr amework of organ identity, but is better interpreted as a largely derivative morphology (in essence, a pigmented leaf sheath) that is perceived as petaloid in the context of the flower. Additionally, the correspondence of the petaloid tepal and leaf shea th is insightful due to the correspondance of the delicate petal -like tissue and hyaline margins of the leaf sheath (Figure 6 3, H and K). This observation suggests that petal like tissue can have unexpected developmental origins. We speculate that hyaline margins may be pre adapted for co -option as pigmented petal tissue as they are both delicate and, perhaps more significantly, achlorophyllous. The role of hyaline margins in the perianth is largely unstudied (see Rohweder, 1967, 1970; cited in Hoffman, 1994); however additional taxa within the Caryophyllales also exhibit a sepal like perianth with delicate petaloid margins (e.g. Polycarpon (Figure 6 3, N) and Hypertelis (Figure 6 3, O). The correspondence between hyaline margins and petal tissue implies th at a petal -like appearance may be achieved differently depending on the historical derivation of the petaloid organ; this complicates our morphological comparison of petaloid tepals and staminodes. Although it is tempting to argue that a hyaline origin of delicate petal like tissue in S. portulacastrum demonstrates non -homology with respect to the delicate tissue in petaloid staminodes of Aizoaceae, this risks conflating concepts of historical and morphological homology. Delicate tissue is commonly regarded as a distinctive feature of petals (Whipple et al., 2007); however delicacy is yet another intangible term without a defined histological basis. As such, there may still be morphological characteristics that are shared by the delicate tissue of the tepals and the petaloid staminodes in Aizoaceae.

PAGE 180

180 Assessing Morphological Homology between Different Petaloid Organs in Aizoaceae Given the uncertainty introduced by the different derivations of petaloid organs within Aizoaceae, what morphological characteristi cs should one evaluate in the framework of comparative petal identity? Despite their distinctive appearance, petals can have a very simple structure (Endress and Matthews, 2006) which makes it difficult to define discrete morphological characters with which to test homology. Perhaps the most useful characters then are those that have been shown to have adaptive significance to pollinator attraction and which may be present irrespective of the position or derivation of the petal. We focused on three such ada ptive characteristics that have been demonstrated to enhance petal -pollinator interactions: conical cell morphology on the epidermis (Glover and Martin, 1998; Kevan and Lane, 1985; Noda et al., 1994), epidermal cuticular striation (Whitney et al., 2009), a nd pigmentation (Waser and Price, 1981). We surveyed the epidermis of S. portulacastrum and D. napiforme (Figure 6 1, 2, 3) and several additional genera in Aizoaceae: Aizoon with petaloid tepals has a similar cell type to Sesuvium ; in Tetragonia all plan t surfaces including petaloid tepals are covered in bladder cells (data not shown); Mesembryanthemum and Mestoklema have petaloid staminodes with similar cell type to Delosperma (data not shown); Lampranthus has petaloid staminodes with flattened tabular e pidermal cells similar to Delosperma (Christensen and Hansen, 1998). With respect to epidermal morphology and cuticular striation, we find no similarity between the petaloid tepals and petaloid staminodes. In both petaloid staminodes and petaloid tepals the epidermal morphology is remarkably undifferentiated with an absence of conical cells and cuticular striations. This lack of differentiation is not necessarily a point of a similarity however, as the epidermal cell types in D. napiforme and S. portulacast rum most resemble those found on the organs to which they are historically homologous (stamen filaments in the case of the staminodes, and the adaxial surface of the leaf sheath in the case of petaloid tepals). It is only

PAGE 181

181 with respect to pigmentation that the petaloid tepals and petaloid staminodes can be considered morphologically homologous. In conclusion, we find that the morphological appearance of the petaloid organs in Aizoaceae can be largely explained through appeal to historical derivation. There are inherent limitations of assessing homology in simple and largely undifferentiated structures; however there is little evidence for morphological homology between petaloid staminodes and petaloid tepals to support a concept of shared petal identity in A izoaceae. One exception is pigmentation, which is not a trivial similarity but the manifestation of several complex developmental pathways regulating processes such as pigment synthesis, manufacture of leuco or chromoplasts, and the turning off of chlorop last synthesis (Irish, 2009). Therefore we next assess the evidence for process homology between petaloid staminodes and petaloid tepals Assessing Process Homology between Petaloid Staminodes and Petaloid Tepals The term process homology as coined by Gil bert and Bolker (2001) refers to similarity in developmental pathways comprised of homologous proteins and related by common ancestry. Unfortunately, the developmental pathway that leads to the expression of petal characteristics in model organisms is stil l poorly characterized and here we can only assess process homology at the level of gene expression of the AP3 PI and AG homologs. We realize that the expression pattern of three gene homologs does not necessarily constitute process homology per se howeve r these three genes perform critical roles in genetic pathways of organ identity in model organisms. Differences in the expression patterns of these genes would at least be suggestive of differences in process homology although we accept that genetic simil arities may still exist at different hierarchical levels than that represented by MADS box gene expression. From the perspective of genetic programs of petal identity we expected the B -class gene homologs SpAP3 and SpPI to be expressed in the petaloid adaxial lamina of the tepal and the

PAGE 182

182 lower zone of the primordia from which the petaloid lamina is derived. Contrary to this expectation, SpAP3 and SpPI are not expressed at any point in the development of the petaloid tepal but that expression of SpAP3 and SpP I is restricted to the stamens, where they are expressed throughout the development of the filament and anthers. In contrast, in D. napiforme DnAP3 and DnPI are expressed not only in the filaments and anthers of fertile stamens, but also in the petaloid staminodes. Petaloid tepals and petaloid staminodes can also be distinguished at the level of AG homolog expression as AG homologs are expressed in petaloid staminodes but are not expressed in petaloid tepals. Therefore at the level of AP3 PI and AG homol og expression, we find no evidence for process homology between the petaloid tepals and petaloid staminodes. This supports our morphological interpretation that there is little to unite petaloid tepals and petaloid staminodes within the framework of a shar ed petal identity program. Implications of AP3 and PI Homolog Expression Patterns The petaloid appearance of the tepal is not dependent on SpAP3 or SpPI activity. To our knowledge this is the first evidence that demonstrates the complete absence of AP3 a nd PI homologs in a petaloid organ within the core eudicots, although previous studies have also conceptually decoupled petal identity and B -class gene expression in other angiosperm taxa. On the basis of late stage RT -PCR Geuten et al. (2006) suggest that AP3 homologs but not PI homologs are expressed in the petaloid sepals of Impatiens hawkeri (Balsaminaceae). However, late stage RT PCR alone is a potentially misleading method of assaying gene expression with respect to organ identity (Jaramillo and Krame r, 2007). Jaramillo and Kramer (2004) found expression of AP3 and PI homologs in the showy uniseriate perianth of Aristolochia but noted that, in contrast to model organisms, the onset of expression occurs late in the differentiation of the perianth and i s not correlated with the showiest parts of the perianth. Park et al. (2004, 2003) demonstrated that AP3 and PI homologs are not expressed in the outermost petaloid perianth

PAGE 183

183 whorl in the homochlamydeous flowers of Asparagus Silencing of the PI homolog in Aquilegia suggests that Aq v PI is required for petal identity in the second whorl but not the petaloid first whorl (Kramer et al., 2007). Finally, Whipple et al., (2007) found expression of AP3 and PI homologs irrespective of a petaloid appearance in the se cond whorl of members of Poales. The absence of SpAP3 and SpPI activity in the petaloid tepals of Sesuvium portulacastrum extends these observations to the core eudicots, and contradicts earlier hypotheses (Kramer and Irish, 1999, 2000) that all petaloid o rgans in core eudicots might share a common AP3 and PI dependent petal identity pathway. The spatial and temporal expression patterns of DnAP 3 and DnPI within the petaloid staminodes also bear on the concept of a core eudicot petal identity program. Expres sion of DnAP3 and DnPI is initially strong in the primordia but is subsequently restricted to the distal tip of the developing staminode. Signal through in -situ hybridisation is undetectable by the time petaloid staminodes reach ~500 M in length. This expr ession pattern differs from that of stamens as in both S. portulac a strum and D. napiforme the AP3 and PI homologs are expressed throughout filament and anther development. RNA -RNA in -situ hybridization can become unreliable at later stages of organ develo pment due to diffuse signal caused by vacuolarization and cell expansion (Kramer and Jaramillo, 2005). However we have several reason to reject this artifactual explanation: 1) We were able to detect strong AP3 and PI homolog expression in expanded stamen filaments of both S. portulacastrum and D. napiforme (Figure 6 5 E, K and Figure 6 6, E) We detected spatially distinct patterns in temporally separated staminodes despite very slight differences in overall length of the staminode (Figure 6 5, K) 3) Strength of signal does not correlate with cell size (data not shown) 4) The in -situ hybridization data is supported by temporal changes in gene expression detected by RQ RT PCR (Figure 6 -7).

PAGE 184

184 This pattern of expression is of interest because it is similar to t hat reported in the petaloid organs of Ranunculales (Kramer and Irish, 1999). Constant expression until late stages of petal development is necessary for the maintenance of petal identity in A. thaliana and A. majus Given inconstant expression of B -class homologs in Ranunculus and Papaver Kramer and Irish (1999) hypothesized that the developmental genetic processes operating in the petals in Ranunculales were different from petals in core eudicot species. They suggested that these developmental differences might be the consequence of the independent evolution of petals in the Ranunculales and/or the phylogenetic position of the order outside of core eudicots, prior to the fixation of a petal identity program. However, the theory of Developmental Systems Dr ift offers an alternative explanation (Kramer and Jaramillo, 2005; True and Haag, 2001). Our description of a similar pattern within Aizoaceae suggests that even within core eudicots there may be considerable variation in AP3 and PI expression in stamen -de rived petaloid organs. Following the argumentation of Kramer and Irish (1999), the detection of this pattern within the core eudicots argues against the notion of a core eudicot petal identity program that has simply been turned on in the petaloid staminodes of Aizoaceae. Implications of AG Homolog Expression Patterns The specification of petal identity by AP3 and PI homologs in A. thaliana and A. majus is contingent on an absence of AG homolog activity; however, the expression of AG homologs is rarely ex amined when assessing evidence for a conserved petal identity program. The expression pattern of DnAG in petaloid staminodes further limits our expectation of a typical eudicot petal identity program operating in the petaloid staminodes. In D. napiforme DnAG is expressed not only in the gynoecium and androecium, but also in the petaloid staminodes. This finding is not surprising from the perspective of historical homology as the petals in D napiforme are derived from stamens, and thus one might expect Dn AG to be retained in the development of the

PAGE 185

185 staminodes (Chanderbali et al, 2006). The spatial expression pattern is consistent with this interpretation because although DnAG is expressed early on it is subsequently lost perhaps allowing staminodes to becom e sterile rather than developing into fertile stamens. However, DnAG expression in the staminodes is unexpected in comparison with the petals of A. thaliana and A. majus AG homologs are not expressed during the development of the petals of these model spe cies and yet traditionally the petals of these species are also considered to be stamen derived. The expression of DnAG in staminodes is therefore a further point of non-homology with respect to the petals of other core eudicots. Notably, in the petaloid staminodes of D. napiforme the strong expression of DnAP3 and DnPI seems entirely coincident with strong DnAG expression. All three genes exhibit the same patterns of strong early expression and subsequent loss. At no point in the development of the petal oid staminodes are DnAP3 and DnPI strongly co -expressed in absence of DnAG expression. This co incident expression of DnAG DnAP3 and DnPI conceptually limits the opportunity for a DnAP3 and DnPI driven petal identity pathway unless the co activity of Dn AG DnAP3 and DnPI is decoupled. The quartet model of MADS -box protein interactions (Saedler and Theissen, 2001), in which the coactivity of AP3 and PI and AG is mediated by additional SEPELLATA (SEP ) MADS box proteins, provides a possible mechanism for a decoupling of DnAG DnAP3 and DnPI activity. For example, in Gerbera it has been discovered that a specific SEP locus is required for mediating C function. Antisense knockdown of this locus, GRCD1 results in the transformation of stamens into petaloid staminodes. However, no C functionspecific SEP has been described in Caryophilids, or other core eudicots for that matter. This raises another possibility that transient, early expression of the B+C code in staminodia commits them to an alternative orga n identity. Notably, the down regulation of B and C genes

PAGE 186

186 does not result in the organs reversion to tepal identity but, rather, their development into a completely distinct fourth identity. It will be important, therefore, to examine other floral MADS box genes to determine what loci might be involved in distinguishing these organs from stamens on the o ne hand and tepals on the other. Alternatively, the similar coincident expression and loss of DnAG DnAP3 and DnPI might simply signify the wholesale turni ng off of a stamen developmental program. Evolution of the Aizoaeae Perianth in the Context of the Caryophyllales We have assessed the different perianth stuctures in Aizoaceae from the perspective of historical, morphological and process homology and fin d numerous anomalies that are difficult to explain from the perspective of a common petal identity program: 1) lack of shared morphological features between petaloid staminodes and petaloid tepals, 2) a potentially different mechanism for the origin of pet aloid tissue in tepals, 3) absence of AP3 and PI homologs from the petaloid tepals of Sesuvioideae and Ruschoideae, 4) coincident expression of AP3 PI and AG homologs in petaloid staminodes of Ruschoideae and Mesembyanthoideae ,and 5) loss of AP3 PI and AG homologs through the development of the petaloid staminodes. So far we have discussed our data in the context of Aizoaceae with some reference to other core eudicot model organisms, however the unique evolutionary history of the Caryophyllales provides additional insight. From the character reconstruction analyses it is evident that early in the evolution of the Caryophyllales, one whorl of perianth was lost such that the simple uniseriate perianth of the Sesuvioideae/Aizooideae type predominates throu ghout the Caryophyllales (Brockington et al., 2009). Traditionally it has been assumed that the inner petal whorl of the perianth was lost such that the uniseriate tepals in Caryophyllales correspond to the sepals of other core eudicots (Hofmann, 1994). Th e absence of AP3 and PI homolog expression in the petaloid tepals of

PAGE 187

187 Aizoaceae is consistent with this interpretation as core eudicot sepals also lack B -class gene function. This appeal to historical homology suggests that tepals (corresponding to sepals) have reacquired petaloid characteristics (as in Sesuvioideae and Aizooideae), independently of an AP3 and PI dependent petal identity program. If the petaloid appearance in the tepals of Sesuvioideae and Aizooideae have been achieved independently of AP3 a nd PI homologs this may well be the case in the numerous other lineages of the Caryophyllales which possess showy organs that are positionally or historically homologous to the tepals of Aizoaceae e.g the tepals of Molluginaceae, Nyctaginaceae and Hypertel is and the inner perianth whorls of Portulacaceae, Montiaceae, and Didieraceae. A sepal -derived interpretation of the petaloid tepals in Sesuvioideae and Aizooideae therefore predicts widespread convergent evolution of petaloid organs in additional lineage s of Caryophyllales, independent of the AP3 and PI dependent genetic programs operating in other core eudicots. Character reconstruction analyses suggest that following a loss of a perianth whorl, multiple separate origins of a differentiated perianth resu lt from the recruitment of staminodes to function as petals e.g the petaloid staminodes in Aizoaceae and the petals of Caryophyllaceae, Corbichonia and Glinus (Molluginaceae) (Brockington et al., 2009). Arguably, this is precisely the pattern of perianth e volution that the concept of an ancestral petal identity program is intended to explain: o nce such a homeotic petal identity program evolved, deactivation and reactivation of the genetic pathway could produce independent loss es and gains of petaloid organ s (Kramer and Jaramillo, 2005). A loss of petals, early in the evolution of the core Caryophyllales, would seem to necessitate loss of an AP3 and PI driven petal identity pathway. However the absence of AP3 and PI homologs from the tepals and sepals in Ai zoaceae implies that an AP3 and PI dependent petal identity pathway was not maintained by spatial redeployment

PAGE 188

188 to the tepals at least at the point of divergence of Aizoaceae. Therefore it is not clear how, following the complete loss of the second whorl or gans, an AP3 and PI dependent pathway would then be maintained. This is perhaps reflected by genetic differences in the development of stamen -derived petals in Aizoaceae compared with Caryophyllaceae. Hardenack et al. (1994) describes the expression of AP3 and PI homologs early in the development of the petals in Silene but the expression of the AG homolog SLM1 is clearly absent from all stages of petal development, similar to other core eudicots. This signifies that in lineages within the Caryophyllales, i ndependent recruitment of stamens to function as petals has entailed dissimilar developmental genetic processes; it is difficult to reconcile these observations with a petal identity program that has been maintained across the Caryophyllales despite losses and gains of petals. Thus far the petaloid staminodes of Aizoaceae appear to be novel floral organs, uniquely derived within the family, and exhibiting a unique developmental genetic program with respect to the stamen derived petals of Caryophyllaceae and other core eudicots. Conclusion In this study we have analyzed two different perianth structures each possessing functionally similar petaloid organs of differing historical homology. We found little morphological evidence for a concept of petal identi ty that unites these different petaloid structures. This finding was supported by expression data suggesting dissimilar developmental genetic processes operating in petaloid staminodes versus petaloid tepals. We therefore reject the hypothesis that hetereo topic expression of an AP3 and PI driven petal identity pathway underlies superficially similar appearance of petaloid staminodes and petaloid tepals. Furthermore the gene expression patterns in these petaloid organs do not match those of the petals in any other core eudicot organisms studied to date. Consequently we reject the hypothesis

PAGE 189

189 that a core eudicot petal identity program operates in Aizoaceae and suggest that, even within core eudicots, not all petals were created equal.

PAGE 190

190 Figure 6 1 Fl oral Development in Sesuvium portulacastrum ( A -F ). Early development of the pentamerous uniseriate perainth. ( A ) Meristem encloded by floral bracts. ( B ) Emergence of first two members of perianth. (C and D) Emergence of third and fouth tepal primordia. (E ) Five tepal primordia in 2/5 quincuncial arrangement. (F) Five tepals enclosing floral meristem. (G -L) Development of the androecium and gynoecium. (G) Tepals parted to reveal underlying development of androecium and gynoecium. (H and I) Tepals removed to reveal developing carpel primordia in the center of the flower surrounded by early stamen primordial. (J) Tricarpellate gynoecium is revealed together with fasciculated androecial development. (K) Developing stamens beginning to enclose gynoecium, (L) Sta mens with clear anther and filament differentiation enclose the gynoecium. (M) Cell type on the adaxial petaloid epidermis of mature tepal. (N) Stomata on the abaxial sepaloid epidermis of the mature tepal. (O) Mature floral form of Sesuvium portulacastrum

PAGE 191

191 Figure 6 2. Floral development in Delosperma napiforme (outer sepals removed). ( A ) floral meristem prior to emergence of androecial and gynoecial primordia. ( B and C) Five carpel primordia emerge at the centre of the meristem. ( D and E ) Further androe cial primordia initiate to the outside of existing stamen primordia in a centrifugal fashion. (F ) Inner androecial primordia differentiate into stamens with anthers clearly visible. (G ) Outer androecial primordia differentiate into sterile lamina petal str uctures. ( H ) Petals clearly visible enclosing underlying fertile stamens and carpels. ( I) Mature flower opened to reveal multiple petal whorls surrounding single whorl of fertile stamens and five -carpellate gynoecium. ( J ) Outer petals removed to reveal sin gle whorl of developing differentiated stamnes surrounding gynoecium. ( K) Outer petals removed to reveal stamens and carpels at anthesis. ( L ) Stamens flanked on the left by a developing sterile staminode. ( M ) Cell type on the epidermis of the mature stamen filament. ( N) Cell type on the epidermis of the mature petals. ( O ) Mature floral form of Delosperma napiforme

PAGE 192

192 Figure 6 3. Correspondence between perianth and leaf sheath in Sesuvium portulacastrum ( A -F ) Expansion of the lower leaf zone through devel opment of the tepals. (A) Single primordium with distinct upper (green) and lower (pink) zones. ( B and C) Early in development the upper leaf zone forms the bulk of the tepal. ( D, E and F ) Expansion of lower leaf zone. ( F ) Lower leaf zone forms the bulk of the tepal and upper leaf zone forms distal tip of the tepal. ( G and H ) Adaxial and abaxial surface of a mature tepal with distal tip (arrow in H ). ( I) Vegetative leaf illustrating the leaf lamina and leaf sheath. ( J ) Adaxial surface of the leaf sheath. ( K ) Abaxial surface of the leaf sheath. Comparison of ( K) with ( H ) demonstrates relationship between petaloid flanges of tepal and marginal tissue of the leaf sheath (arrows). ( L ) Leaf sheath attached to the shoot node. ( M ) Cell type on the adaxial surface of the leaf sheath. ( N) Petaloid margins in Polycarpon tetraphyllum (arrows). (O ) Petaloid margins in Hypertelis salsoloides (arrows).

PAGE 193

193 Figure 6 4. Expression of AGAMOUS (AG ) homologs in Sesuvium portulacastrum (A F ) and Delosperma napiforme (G -L ). Tepal (Te), Androecium (A), Gynoecium (G), Stamen (Sta), Filament (Fil), Anther (Ant), Staminode (St), Ovary (Ov). ( A ) Weak SpAG expression in the centre of the meristem, absent from tepals. ( B an d C ) SpAG expression in gynoecium and androecium. ( D ) SpAG expres sion in androecium and absence in tepals. ( E ) SpAG expression in carpels and stamens. ( F ) SpAG expression in the developing ovules; weak expression in filaments and anthers of differentiated stamens. (G) Dn AG expression in the carpel and stamen primordia. (H ) Dn AG expression in carpel and stamen primordia and in outer first and second primordial that form petaloid staminodes. ( I) Dn AG expression in gynoecium and all fertile and sterile members of the androecium. ( J ) Dn AG expression strong in anthers and fi laments of inner stamens; lost from developing staminode 1; expression restricted to the tip of staminode 2 (black arrow), strong expression in emerging staminode 3 (~100 M in length) (black arrow). (K) Expression of Dn AG in anthers; absent now in staminod es 1 and 2; expression restricted to distal tip of staminode 3 (~200uM in length). ( L ) Dn AG expression in developing ovules.

PAGE 194

194 Figure 6 5. Expression of PISTILLATA ( PI ) homologs in Sesuvium portulacastrum (A F ) and Delosperma napiforme (G -L ). Tepal (Te), Androecium (A), Gynoecium (G), Stamen (Sta), Filament (Fil), Anther (Ant), Staminode (St), Ovary (Ov). ( A ) Strong SpPI expression marks the site of origin of stamens; no expression developing tepals. ( B an d C) Sp PI expression in developing androecium. ( D ) Section across the androecium show SpPI expression maintained in developing androecium and absence in tepals. (E ) Sp PI expression in differentiating stamens; absent in the carpel ( F ) Sp PI in expression mature stamens; absent from carpels and developing o vules. ( G H and I) Dn PI expression in emerging and developing androecial primordia. ( J ) Expression maintained in anthers and filament; expression in innermost stamimodes (1 and 2) weaker than outermost staminode (3). ( K ) Expression maintained in anther and filament (white arrow), expression weak in staminode 1; restricted to tip in staminode 2; strong in staminode 3 (~150 M in length). (L) Expression of Dn PI absent now in staminodes 1; expression almost entirely lost in staminode 2 and restricted to the di stal tip of staminode 3 (~250uM in length); Dn PI expression absent in developing ovules.

PAGE 195

195 Figure 6 6. Expression of APETALA3 (AP3) homologs in Sesuvium portulacastrum (A F ) and Delosperma napiforme (G -L ). Tepal (Te), Androecium (A), Gynoecium (G), Stamen (Sta), Filament (Fil), Anther (Ant), Staminode (St), Ovary (Ov), Ovules (Ovl). (A ) Strong SpAP3 expression marks the site of origin of stamens; no expression in developing tepals. ( B an d C ) SpAP3 expression in developing androecium. ( D and E ) SpAP3 expres sion in differentiating stamens; absent in the carpel. ( F ) SpAP3 expression in developing ovules. ( G H and I) Dn AP3 expression in emerging and developing androecial primordia. ( I) Expression in developing placenta, expression maintained in anthers and fil ament; expression in stamimode 1 weaker than in staminode 2. ( J ) Expression strong in locule walls, absent in staminode 1, expression restricted to distal tip of staminode 2, stronger in staminode 3. ( K ) Expression absent in all staminodes, maintained in a nther locule wall and in ovules. (L) Dn AP3 expression strong in developing ovules.

PAGE 196

196 Figure 6 6. RQ RT PCR data for four species of Aizoaceae and three gene homologs ( AG, AP3, and PI ). S. portulacastrum and D. napiforme are examined at three different st ages of floral size (<4mm, 5 8mm, >10mm in diameter). T. tetragonoides was assessed at the mature flower only and M cordifolia was assessed using admixed flowers of different sizes. (Vlf = vegetative leaf, Sep = sepals, Pte = petaloid tepals, Pst = petalo id staminode, Car = carpel). 18S rRNA was employed as a control.

PAGE 197

197 REFERENCES Ainsworth, C, S Crossley, V Buchananwollaston, M Thangavelu, J Parker 1995 Male and female flowers of the dioecious plant sorrel show different patterns of MADS box gene -expr ession. Plant Cell 7: 15831598. Albert, VA, SE Williams, MW Chase 1992 Carnivorous Plants: Phylogeny and Structural Evolution. Science 257: 14911495 Alvarez -Buylla, ER, S Pelaz, SJ Liljegren, SE Gold, C Burgeff, GS Ditta, LR de Pouplana, L Martinez -Castilla, MF Yanofsky 2000 An ancestral MADS -box gene duplication occurred before the divergence of plants and animals. Proceedings of the National Academy of Sciences of the United States of America 97: 53285333. Angenent, GC, J Franken, M Busscher, A Vandijk en, JL Vanwent, HJM Dons, AJ Vantunen 1995 A novel class of MADS box genes is invo lved in ovule development in Petunia Plant Cell 7: 15691582. APG II 2003 An update of the angiosperm phylogeny group classification for the orders and families of flowering plants: APG II. Botanical Journal of the Linnean Society 141: 399436. Applequist, WL, RS Wallace 2001 Phylogeny of the portulacaceous cohort based on ndhF sequence data. Systematic Botany 26: 406 419 Arnold, KE, IPF Owens, NJ Marshall 2002 Fluorescent si gnaling in parrots. Science 295: 9292. Bate -Smith, E 1962 The phenolic constituents of plants and their taxonomic significance. Botanical Journal of the Linnean Society 58: 95 173. Baum, DA, BA Whitlock 1999 Plant development: Genetic clues to petal evolution. Current Biology 9: R525-R527. Beale, G, J Price, V Sturgess 19 41 A survey of anthocyanins. VII The natural selection of flower colour. Proceedings of the Royal Society of London. Series B, Biological Sciences 130: 113126. Beck, E, H Merxmuller, H W agner 1962 Uber die Art der anthocyane bei Plumbaginaceen, Alsinoideen und M olluginaceen. Planta 58: 220 224. Becker, A, G Theissen 2003 The major clades of MADS -box genes and their role in the development and evolution of flowering plants. Molecular Phylo genetics and Evolution 29: 464489. Behnke, HD 1994 Sieve Element Plastids: Their Significance for the Evolution and Systematics of the Order. Pages 87121 in HD Behnke and TJ Mabry, eds. Caryophyllales: Evolution and Systematics. Springer, Berlin. Behnke, HD 1997 Sarcobataceae a new family of Caryophyllales. Taxon 46: 495507.

PAGE 198

198 Bessey, CE 1915 The phylogenetic taxonomy of flowering plants. Annals of the Missouri Botanical Garden 2: 108 164. Birkinshaw, C, R Edmond, C Hong-Wa, C Rajeriarison, F Randriantaf ika, G Schatz 2004. Red lists for Malagasy plants II: Asteropeiaceae. Missouri Botanical Garden, Antananarivo University. Bischoff, H. 1876. Das caryophyllinenroth Inaugural Dissertation. Universitat Tubingen. Bittrich V 1993 Halophytaceae. Pages 320321 i n K Kubitzki, JG Rohwer, V Bittrich, eds. Flowering plants, Dicotyledons, Magnoliid, Hamamelid, and Caryophyllid families. Springer Verlag, Berlin; New York. Bittrich, B, M Amaral 1991 Proanthocyanidins in the testa of centrospermous seeds. Biochemical Sys tematics and Ecology 19: 319321. Blackwell, WH and Powell MJ 1981 A preliminary note on pollination in the Chenopodiaceae. Annals of the Missouri Botanical Garden 68: 524526 Blattner, FR 1999 Direct amplification of the entire its region from poorly pres erved plant material using recombinant PCR. Biotechniques 27: 1180 Bolker, JA, RA Raff 1996 Developmental genetics and traditional homology. Bioessays 18: 489494. Bollback, JP 2006 SIMMAP: Stochastic character mapping of discrete traits on phylogenies. BM C Bioinformatics 7: Bowman, JL, DR Smyth, EM Meyerowitz 1989 Genes directing flower development in Arabidopsis Plant Cell 1: 37 52. Bowman, JL, DR Smyth, EM Meyerowitz 1991 Genetic interactions among floral homeotic genes of Arabidopsis Development 112: 1 20. Bradley, D, R Carpenter, H Sommer, N Hartley, E Coen 1993 Complementary floral homeotic phenotypes result from opposite orientatio ns of a transposon at the PLENA locus of Antirrhinum Cell 72: 8595. Braun, A 1864 ebersicht der Natrlichen Systems nach der Anordnung Derselben. Pages 2267 in P Ascherson, ed. Flora der Provinz Brandenburg, der Altmark und des Herzogthums Magdeburg. Hirschwald, Berlin. Brockington, SF, E Mavrodiev, J Ramdial, A Dhingra, PS Soltis, DE Soltis 2008 Keep the DNA Rolling: Multiple Displacement Amplification of archival plant DNA extracts. Taxon 53: 944 948

PAGE 199

199 Brockington, SF, Frohlich M, PJ Rudall, DE Soltis, PS Soltis 2007 Differing patterns of MADS box gene expression associated with shifts in petaloidy within Aizoaceae (Car yophyllales). 2007. Page 191 in Program and Abstract Book, Plant Biology and Botany Congress, Chicago: pp 191. Brockington, SF, R Alexandre, J Ramdial, MJ Moore, S Crawley, A Dhingra, K Hilu, DE Soltis, PS Soltis 2009 Phylogeny of the caryophyllales sensu l ato: Revisiting hypotheses on pollination biology and perianth differentiation in the core caryophyllales. International Journal of Plant Sciences 170: 627643. Buxbaum, F 195055 Morphology of Cacti. Pasadena, Abbey Garden Press. Carolin R 1987 A review of the family Portulacaceae. Australian Journal of Botany 35: 383412. Carpenter, R, ES Coen 1990 Floral homeotic mutations produced by transposon -mutagenesis in Antirrhinum majus Genes & Development 4: 14831493. Chanderbali, AS, S Kim, M Buzgo, Z Zheng, DG Oppenheimer, DE Soltis, PS Soltist 2006 Genetic footprints of stamen ancestors guide perianth evolution in Persea (Lauraceae). International Journal of Plant Sciences 167: 10751089. Chase, MW, DE Soltis, RG Olmstead, D Morgan, DH Les, BD Mishler, MR Du vall, RA Price, HG Hills, YL Qiu, KA Kron, JH Rettig, E Conti, JD Palmer, JR 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 Quinn, LE Eguiarte, E Golenberg, GH Learn, SW Graham, SCH Barrett, S Dayanandan, VA Albert 1993 Phylogenetics of seed plants an analysis of nucleotide sequ ences from the plastid gene rbcL Annals of the Missouri Bot anical Garden 80: 528 580. Christensen, K, H Hansen 1998 Sem -studes of epidermal patterns of petals in angiosperms. Opera Botanica 135: 191. Clement, J, T Mabry 1996 Pigment evolution in the C aryophyllales: A systematic overview. Botanica Acta 109: 360367. Clement, J, T Mabry, H Wyler, A Dreiding 1994 Chemical review and evolutionary significance of the betalains. in: H D Behnke, T Mabry eds. Caryophyllales: Evolution and systematics. Springer, Berlin. Coen, ES, EM Meyerowitz 1991 The war of the whorls genetic interactions controlling flower development. Nature 353: 3137. Couvreur, TLP, JE Richardson, MSM Sos ef, RHJ Erkens, LW Chatrou 2008 Evolution of Syncarpy and other morphological characters in African Annonaceae: A posterior mapping approach. Molec ular Phylogenetics and Evolution 42: 302318 Couvreur, TLP. 2008. Revealing the secrets of African Annonaceae: Systematics, evolution and biogeography of the syncarpous genera Isolona and Monodora. Pages 15 33 D. Phil. Thesis, Wageningen University.

PAGE 200

200 Cronqu ist, A 1981 An integrated system of classification of flowering plants. Columbia University Press, New York. Cronquist, A 1988 The evolution and classification of flowering plants. New York Botanical Garden, Bronx. Cronquist, A, RF Thorne 1994 Nomenclatural and taxonomic history. Pages 87 12 in HD Behnke and TJ Mabry, eds. Caryophyllales: Evolution and Systematics. Springer, Berlin. Cuenoud, P, V Savolainen, LW Chatrou, M Powell, RJ Grayer, MW Chase 2002 Molecular phylogenetics of Caryophyllales based on nuclear 18S rDNA and plastid rbcL, atpB and matK DNA sequences. American Journal of Botany 89: 132144. Dahlgren, R 1975 A system of classification of the angiosperms to be used to demonstrate the distribution of characters. Botaniska Notiser 128: 119147. de Martino, G, I Pan, E Emmanuel, A Levy, VF Irish 2006 Functional analyses of two tomato apetala3 genes demonstrate diversification in their roles in regulating floral development. Plant Cell 18: 18331845. Dean, FB, S Hosono, L Fang, X Wu, AF Faruqi, P Bray -Ward, Z Sun, Q Zong, Y Du, J Du, M Driscoll, W Song, SF Kingsmore, M Egholm, RS Lasken 2002 Comprehensive human genome amplification using multiple displacement amplification. Proceedings of the National Academy of Sciences of the United States of Am ericaU S A 99: 52615266. Dhingra, A, KM Folta 2005 ASAP: Amplification, Sequencing and Annotation of Plastomes. BMC Genomics 6: 176. Di Gregorio, J, J Dipalma 1966 Anthocyanin in Dionaea muscipula (Ellis) (Venus Flytrap). Nature 212: 12641265. Ditta, G, A Pinyopich, P Robles, S Pelaz, MF Yanofsky 2004 The SEP4 gene of Arabidopsis thaliana functions in floral organ and meristem identity. Current Biology 14: 19351940. Douglas, NA, PS Manos. 2007. Molecular phylogeny of Nyctaginaceae: Taxonomy, biogeography and characters associated with a radiation of xerophytic genera in North America. American Journal of Botany. Vol. 94. Downie, SR, DS Katz -Downie, KJ Cho 1997 Relationships in the Caryophyllales as suggested by phylogenetic analyses of partial chloroplas t DNA ORF2280 homolog sequences. American Journal of Botany 84: 253273. Downie, SR, JD Palmer 1994 A chloroplast DNA phylogeny of the Caryophyllales based on structural and inverted repeat restriction site variation. Systematic Botany 19: 487 487. Doyle, J, J Doyle 1987 Genomic plant DNA preparation from fresh tissue CTAB method. Phytochemical Bulletin 19: 11.

PAGE 201

201 Drabkova, L, J Kirschner, C Vlcek 2002 Comparison of seven DNA extraction and amplification protocols in historical herbarium specimens of Juncaceae Plant Molecular Biology Reporter 20: 161175. Drea, S, LC Hileman, G De Martino, VF Irish 2007 Functional analyses of genetic pathways controlling petal specification in poppy. Development 134: 41574166. Duminil, J, MH Pemonge, RJ Petit 2002 A set of 35 consensus primer pairs amplifying genes and introns of plant mitochondrial DNA. Molecular Ecology Notes 2: 428 430. Egea Cortines, M, H Saedler, H Sommer 1999 Ternary complex formation between the MADS box proteins SQUAMOSA DEFICIENS and GLOBOSA is invol ved in the control of floral architecture in Antirrhinum majus EMBO Journal 18: 53705379. Ehrendorfer E 1976 Closing remarks: Systematics and evolution of centrospermous families. Plant Systematics and Evolution 126: 99106. Eichler, A 18751878 Blthen diagramme I/II. Engelmann, Leipzig, Germany Endress, M, V Bittrich 1993 Molluginaceae i n K Kubitski ed T he families and genera of vascular plants. Vol. 2. Springer Berlin. Endress, P K 2002. Morphology and angiosperm systematics in the molecular era Botani cal Review 68: 545 570. Endress, PK 1994 Diversity and evolutionary biology of tropical flowers. Cambridge University Press, Cambridge. Endress, PK, ML Matthews 2006 Elaborate petals and staminodes in eudicots: Diversity, function, and evolution. Organism s Diversity and Evolution 6: 257293. Favaro, R, A Pinyopich, R Battaglia, M Kooiker, L Borghi, G Ditta, MF Yanofsky, MM Kater, L Colombo 2003 MADS box protein complexes control carpel and ovule development in Arabidopsis Plant Cell 15: 26032611. Fay, MF KM Cameron, GT Prance, MD Lledo, MW Chase 1997 Familial relationships of Rhabdodendron (Rhabdodendraceae) plastid rbcL sequences indicate a caryophyllid placement. Kew Bulletin 52: 923932. Felsenstein, J 1978 Cases in which parsimony or compatability me thods will be positively misleading. Systematic Zoology 27: 401410. Felsenstein, J 1985 Confidence limits on phylogenies an approach using the bootstrap. Evolution 39: 783791. Fior, S, PO Karis, G Casazza, L Minuto, F Sala 2006 Molecular phylogeny of t he Caryophyllaceae (Caryophyllales) inferred from chloroplast matK and nuclear rDNA ITS sequences. American Journal of Botany 93: 399 411.

PAGE 202

202 Friedman, J, SCH Barrett 2008 A phylogenetic analysis of the evolution of wind pollination in the angiosperms. Intern ational Journal of Plant Sciences 169: 4958. Gadagkar, SR, S Kumar 2005 Maximum likelihood outperforms maximum parsimony even when evolutionary rates are heterotachous. Molecular Biology and Evolution 22: 21392141. Gadkar, V, MC Rillig 2005 Application of phi29 DNA polymerase mediated whole genome amplification on single spores of arbuscular mycorrhizal (am) fungi. FEMS Microbiol Lett 242: 6571. Gandia Herrero, F, J Escribano, F Garcia Carmona. 2005. Betaxanthins as pigments responsible for visible fluor escence in flowers Planta. Vol. 222. 1 12 Giannasi, DE, G Zurawski, G Learn, MT Clegg 1992 Evolutionary relationships of the Caryophyllidae based on comparative rbcL sequences. Systematic Botany 17: 1 15. Gilbert, SF, JA Bolker 2001 Homologies of process a nd modular elements of embryonic construction. Journal of Experimental Zoology 291: 112. Glover, BJ, C Martin 1998 The role of petal cell shape and pigmentation in pollination success in Antirrhinum majus Heredity 80: 778784. Goto, K, EM Meyerowitz 1994 Function and regulation of the Arabidopsis floral homeotic gene PISTILLATA Genes and Development 8: 15481560. Graham, SW, RG Olmstead 2000 Utility of 17 chloroplast genes for inferring the phylogeny of the basal angiosperms. American Journal of Botany 8 7: 17121730. 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: 17751787. Hershkovitz, MA 1993 Revised circumscriptions and subgener ic taxonomies of Calandrinia and Montiopsis (Portulacaceae) with Notes on Phylogeny of the portulacaceous Alliance. Annals of the Missouri Botanical Garden 80: 333365 Hershkovitz, MA, EA Zimmer 1997 On the evolutionary origins of the cacti. Taxon 46: 217232. Hilu, KW, T Borsch, K Muller, 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. American Jour nal of Botany 90: 17581776. Hofmann, U 1973 Morphologische Untersushungen zur Umgrenzung und Gliederung der Aizoaceen. Botanische Jahrbcher fr Systematik 93: 247 324. Hofmann, U 1994 Flower morphology and ontogeny. in: H. D. Behnke and T. J. Mabry [ed s.], Caryophyllales: Evolution and Systematics 123166. Springer, Berlin.

PAGE 203

203 Hofmann, U 1994 Flower morphology and ontogeny. Pages 123166 in HD Behnke and TJ Mabry, eds. Caryophyllales: Evolution and Systematics. Springer, Berlin. Huelsenbeck, JP, F Ronquist 2001 MrBayes: Bayesian inference of phylogenetic trees. Bioinformatics 17: 754755. Hunziker, JH, R Pozner, A Escobar 2000 Chromosome number in Halophytum ameghinoi (Halophytaceae). Plant Systematics and Evolution 221: 125127. Irish, VF 2003 The evolution of floral homeotic gene function. Bioessays 25: 637646. Irish, VF, EM Kramer 1998 Genetic and molecular analysis of angiosperm flower development. Advances in Botanical Research, Vol 28 : 197230. Irish, VF, IM Sussex 1990 Function of the APETALA1 gene d uring Arabidopsis floral development. Plant Cell 2: 741753. Jack, T, GL Fox, EM Meyerowitz 1994 Arabidopsis homeotic gene APETALA3 ectopic expression transcriptional and posttranscriptional regulation determine floral organ identity. Cell 76: 703716. J ack, 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: 683697. Jansen, RK, C Kaittanis, C Saski, SB Lee, J Tomkins, AJ Alverson, H Daniell 2006 Phyloge netic analyses of Vitis (Vitaceae) based on complete chloroplast genome sequences: Effects of taxon sampling and phylogenetic methods on resolving relationships among rosids. BMC Evolutionary Biology 6 Jaramillo, MA, EM Kramer 2004 APETAL3 and PISTILLATA h omologs exhibit novel expression patterns in the unique perianth of Aristolochia (Aristolochiaceae). Evolution and Development 6: 449458. Jaramillo, MA, EM Kramer 2007 The role of developmental genetics in understanding homology and morphological evolutio n in plants. International Journal of Plant Sciences 168: 6172. Jaramillo, MA, EM Kramer 2007 The role of developmental genetics in understanding homology and morphological evolution in plants. International Journal of Plant Sciences 168: 6172. Judd, WS, CS Campb ell, EA Kellogg, PF Stevens 2007 Plant systematics: A phylogenetic approach. Sinauer Associates, Sunderland. Kaplan, DR. 1973. The problem of leaf morphology and evolution in the monocotyledons Quarterly Review of Biology. Kevan, P, M Lane 1985 F lower petal microtexture is a tactile cue for bees. Proceedings of the National Academy of Sciences of the United States of America 82: 47504752.

PAGE 204

204 Kim, S, J Koh, MJ Yoo, HZ 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 Journal 43: 724744. Kim, ST, MJ Yoo, VA Albert, JS Farris, PS Soltis, DE Soltis 2004 Phylogeny and diversification of B -function MADS -box genes in angiosperms: Evolutionary and functional implications of a 260million -year -old duplication. American Journal of Botany 91: 2102 2118. Klak, C, A Khunou, G Reeves, T Hedderson 2003 A phylogenetic hypothesis for the Aizoaceae (Caryophyllales) based on four plastid DNA regions. American Journal of Botany 90: 14331445. Komaki, MK, K Okada, E Nishino, Y Shimura 1988 Isolation and characterization of novel mutants of Arabidopsis thaliana defective in flower development. Development 104: 195-&. Kramer, EM, L Holappa, B Gould, MA Jaramillo, D Setnikov, PM Santiago 2007 Elaboration of b gene function to include the identity of novel floral organs in the lower eudicot Aquilegia Plant Cell 19: 750766. Kramer, EM, MA Jaramillo 2005 Genetic basis for innovations in floral organ identity. Journal of Experimental Zoology Part B -Molecular and Developmental Evolution 304B: 526535. Kramer, EM, MA Jaramillo, VS Di Stilio 2004 Patterns of gene duplication and functional evolution during the diversification of the AGAMOUS subfamily of MADS -box genes in a ngiosperms. Genetics 166: 10111023. Kramer, EM, RL Dorit, VF Irish 1998 Molecular evolution of genes controlling petal and stamen development: Duplication and divergence within the APETALA3 and PISTILLATA MADS -box gene lineages. Genetics 149: 765783. Kra mer, EM, VF Irish 1999 Evolution of genetic mechanisms controlling petal development. Nature 399: 144148. Kramer, EM, VF Irish 2000 Evolution of the petal and stamen developmental programs: Evidence from comparative studies of the lower eudicots and basal angiosperms. International Journal of Plant Sciences 161: S29 S40. Kramer, EM, VS Di Stilio, PM Schluter 2003 Complex patterns of gene duplication in the APETALA3 and PISTILLATA lineages of the Ranunculaceae. International Journal of Plant Sciences 164: 1 11. Krizek, BA, EM Meyerowitz 1996 Mapping the protein regions responsible for the functional specificities of the Arabidopsis MADS domain organ identity proteins. Proceedings of the National Academy of Sciences of the United States of America 93: 40634070. Krizek, BA, EM Meyerowitz 1996 The Arabidopsis homeotic genes APETELA3 and PISTILLATA are sufficient to provide the B class organ identity function. Development 122: 1122.

PAGE 205

205 Kubitzki, K, C Bayer 2003 Flowering plants, Dicotyledons: Malvales, Capparales, and nonbetalain Caryophyllales. Springer, Berlin; New York. Kubitzki, K, C Bayer 2007 Flowering Plants. Eudicots: Berberidopsidales, Buxales, Crossosomatales, Fabales p.p., Geraniales, Gunnerales, Myrtales p.p., Proteales, Saxifragales, Vitales, Zygophyl lales, Clusiaceae Alliance, Passifloraceae Alliance, Dilleniaceae, Huaceae, Picramniaceae, Sabiaceae. Springer, Berlin; New York. Kubitzki, K, JG Rohwer, V Bittrich 1993 Flowering plants, Dicotyledons: Magnoliid, Hamamelid, and Caryophyllid families. Sprin ger -Verlag, Berlin; New York. Kuhn, U 1993 Chenopodiaceae. in: K Kubitzki, J Rohwer, V Bittrich eds. The families and genera of vascular plants. Vol. II. Springer -Verlag. Lage, JM, JH Leamon, T Pejovic, S Hamann, M Lacey, D Dillon, R Segraves, B Vossbrinck, A Gonzalez, D Pinkel, DG Albertson, J Costa, PM Lizardi 2003 Whole genome analysis of genetic alterations in small DNA samples using hyperbranched strand displacement amplification and array cgh. Genome Res earch 13: 294307. Lasken, RS, M Egholm 2003 Who le genome amplification: Abundant supplies of DNA from precious samples or clinical specimens. Trends Biotechnol ogy 21: 531535. Litt, A, VF Irish 2003 Duplication and diversification in the APETALA1/FRUITFULL floral homeotic gene lineage: Implications for the evolution of floral development. Genetics 165: 821833. Liu, Y, N Nakayama, M Schiff, A Litt, VF Irish, SP Dinesh -Kumar 2004 Virus induced gene silencing of a DEFICIENS ortholog in Nicotiana benthamiana. Plant Molecular Biology 54: 701711. Mabry, T 1 964 The betacyanins, a new class of red violet pigments, and their phylogenetic significance. Roland Press, New York. Mabry, T 1976 Pigment dichotomy and DNA -RNA hybridization data for centrospermous families. Plant Systematics and Evolution 126: 7994. Ma bry, T 1980 Betalains. in: E Bell, B Charwood eds. Encyclopedia of plant physiology, secondary plant products. Vol. 8. New series. Springer -Verlag, Berlin. Maddison, W, D Maddison 2008 Mesquite: A modular system for evolutionary analysis. Version 2.5 http: //mesquiteproject.Org. Maddison, W., and D. Maddison. 2008. Mesquite: a modular system for evolutionary analysis. Version 2.5 http://mesquiteproject.org. Malcomber, ST, EA Kellogg 2004 Heterogeneous expression patt erns and separate roles of the SEPALLATA gene leafy hull sterile1 in grasses. Plant Cell 16: 1692 1706.

PAGE 206

206 Malcomber, ST, EA Kellogg 2005 SEPALLATA gene diversification: Brave new whorls. Trends in Plant Science 10: 427 435. Manhart, J, JH Rettig 1994 Gene se quence data. in: H D Behnke, T Mabry eds. Caryophyllales: Evolution and Systematics. Springer, Berlin. Mayr, E 1982 Growth of biological thought. The Belknap Press of of Harvard University Press, Cambridge. Mazel, CH, TW Cronin, RL Caldwell, NJ Marshall 20 04 Fluorescent enhancement of signaling in a mantis shrimp. Science 303: 5151. Mears, J 1976 Guide to research in plant taxonomy. Chemical Plant Taxonomy Newsletter 25: 1114. Meyerowitz, EM, JL Bowman, LL Brockman, GN Drews, T Jack, LE Sieburth, D Weigel 1991 A genetic and molecular -model for flower development in Arabidopsis thaliana. Development: 157. Milby TH 1980 Studies in the Floral Anatomy of Claytonia (Portulacaceae). American Journal of Botany 67: 10461050 Mizukami, Y, H Ma 1992 Ectopic expressi on of the floral homeotic gene AGAMOUS in transgenic Arabidopsis plants alters floral organ identity. Cell 71: 119 131. Moore, MJ, CD Bell, PS Soltis, DE Soltis 2007 Using plastid genome -scale data to resolve enigmatic relationships among basal angiosperms Proceedings of the National Academy of Sciences of the United States of America 104: 1936319368. Morton, CM, KG Karol, MW Chase 1997 Taxonomic affinities of Physena (Physenaceae) and Asteropeia (Theaceae). Botanical Review 63: 231 239. Nelson, B, GT Pra nce 1984 Observations on the pollination of Rhabdodendron macrophyllum (Spruce ex Benth.) Huber. Acta Amazonica 14: 411426 Ng, M, MF Yanofsky 2001 Function and evolution of the plant MADS -box gene family. Nature Reviews Genetics 2: 186 195. Noda, K, BJ G lover, P Linstead, C Martin 1994 Flower color intensity depends on specialized cell -shape controlled by a myb related transcription factor. Nature 369: 661-664. Norman, C, M Runswick, R Pollock, R Treisman 1988 Isolation and properties of cDNA clones encod ing SRF a transcription factor that binds to the c -FOS serum response element. Cell 55: 9891003. Nyffeler, R. 2007. The closest relatives of cacti: Insights from phylogenetic analyses of chloroplast and mitochondrial sequences with special emphasis on re lationships in the tribe Anacampseroteae. American Journal of Botany: 94 89 101

PAGE 207

207 Nylander, J 2004 Mrmodeltest v2. Program distributed by the author. Evolutionary Biology Centre Uppsala Univ. Park, JH, Y Ishikawa, R Yoshida, A Kanno, T Kameya 2003 Expression of AODEF, a B Functional MADS -box gene, in stamens and inner tepals of the dioecious species Asparagus officina lis l. Plant Molecular Biology 51: 867875. Park, JH, Y Ishikawa, T Ochiai, A Kanno, T Kameya 2004 Two GLOBOSA like genes are expressed in secon d and third whorls of homochlamydeous flowers in a Asparagus officinalis l. Plant and Cell Physiology 45: 325332. Payer, J B 1857 Traite d'organogenie comparee de la fleur. Masson, Paris. Pelaz, S, GS Ditta, E Baumann, E Wisman, MF Yanofsky 2000 B and C f loral organ identity functions require SEPALLATA MADS -box genes. Nature 405: 200203. Piatelli, M 1981 The betalains: Structure, biosynthesis, and chemical taxonomy. in: C EE ed. The biochemistry of plants. Vol. 7. Acasemic Press. Pinard, R, A de Winter, G J Sarkis, MB Gerstein, KR Tartaro, RN Plant, M Egholm, JM Rothberg, JH Leamon 2006 Assessment of whole genome amplificationinduced bias through high -throughput, massively parallel whole ge nome sequencing. BMC Genomics 7 Pinyopich, A, GS Ditta, B Savidge SJ Liljegren, E Baumann, E Wisman, MF Yanofsky 2003 Assessing the redundancy of MADS -box genes during carpel and ovule development. Nature 424: 8588. Pnueli, L, M Abuabeid, D Zamir, W Nacken, Z Schwarzsommer, E Lifschitz 1991 The MADS box gene family in tomato temporal expression during floral development, conserved secondary structures and homology with homeotic genes from Antirrhinum and Arabidopsis Plant Journal 1: 255266. Pollock, R, R Treisman 1991 Human SRF related proteins DNA-binding proper ties and potential regulatory targets. Genes & Development 5: 23272341. Prance, GT 2003 Rhabdodendraceae. In K Kubitzki, C Bayer eds. The families and genera of vascular plants 5: 339341. Springer, Berlin. Preston, JC, EA Kellogg 2006 Reconstructing the evolutionary history of paralogous APETALA1 / FRUITFULL like genes in grasses (poaceae). Genetics 174: 421437. Price, J, V Sturgess 1 938 A survey of anthocyanins. VI Biochemical Journal 32: 16581660. Raghunathan, A, HR Ferguson, Jr., CJ Bornarth, W Song, M Driscoll, RS Lasken 2005 Genomic DNA amplification from a single bacterium. Appl ied Environ mental Microbiol ogy 71: 33423347.

PAGE 208

208 Rambaut, A, A Drummond. 2003 Tracer. Ver 1.http://tree.Bio.Ed.Ac.Uk/software/tracer/ March 2008 Rasmussen, DA, EM Kramer, EA Z immer 2009 One size fits all? Molecular evidence for a commonly inherited petal identity program in ranunculales. American Journal of Botany 96: 96109. Remane, A 1952 Die grundlagen des naturlichen Systems, der vergleichenden Anatomie und der Phylogenetik Akademische Verlagsgesellschaft, Leipzig. Rettig, JH, HD Wilson, JR Manhart 1992 Phylogeny of the Caryophyllales gene sequence data. Taxon 41: 201209. Riechmann, JL, BA Krizek, EM Meyerowitz 1996 Dimerization specificity of Arabidopsis MADS domain ho meotic proteins APETALA1 APETALA3, PISTILLATA and AGAMOUS Proceedings of the National Academy of Sciences of the United States of America 93: 47934798. Robinson, G, R Robinson 1933 A survey of anthocyanins: Notes on the distribution of leucoanthocyani ns. Biochemical Journal 27: 206212. Rohweder, O 1967 Centrospermenstudien 3. Bltenntwicklng und Bltenblau bei Silenoideen (Caryophyllaceae). Botanische Jahrbcher fr Systematik 86: 130185. Rohweder, O, K Huber (1974) Centropsermenstudien. 7. Beobacht ungen und Anmerkungen zur Morpholgie und Entwicklungsgeschichte einiger Nyctaginaceae. Botanische Jahrbcher fr Systematik 90: 447468 Ronquist, F, JP Huelsenbeck 2003 MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19: 15721574. Ronse De Craene, L 2008 Homology and evolution of petals in the core eudicots. Systematic Botany 33: 301325. Ronse De Craene, LP 2007 Are petals sterile stamens or bracts? The origin and evolution of petals in the core eudicots. Annals of Botany 10 0: 621630. Ronse De Craene, LP 2007 Are petals sterile stamens or bracts? The origin and evolution of petals in the core eudicots. Annals of Botany 100: 621 630. Ronse De Craene, LP 2008 Homology and evolution of petals in the core eudicots. Systematic Bo tany 33: 301325. Ronse De Craene, LP, EF Smets, P Vanvinckenroye 1998 Pseudodiplostemony and its implications for the evolution of the androecium in the Caryophyllaceae. Journal of Plant Research 111: 2543. Saedler, H, G Theissen 2001 Plant biology. Flor al quartets. Nature 409 525529.

PAGE 209

209 Sasaki, N, K Wada, T Koda, K Kasahara, T Adachi. 2005. Isolation and characterization of cdnas encoding an enzyme with glu cosyltransferase activity Plant and Cell Physiology. Sato, M, M Ohtsuka, Y Ohmi 2005 Usefulness of re peated genomiphi, a phi29 DNA polymerase based rolling circle amplification kit, for generation of large amounts of plasmid DNA. Biomolecular Engineering 22: 129132. Schaefer, HM, GD Ruxton 2008 Fatal attraction: Carnivorous plants roll out the red carpet to lure insects. Biology Letters 4: 153155. Schmitz -Linneweber, C, RM Maier, JP Alcaraz, A Cottet, RG Herrmann, R Mache 2001 The plastid chromosome of spinach ( Spinacia oleracea): Complete nucleotide sequence and gene organization. Plant Molecular Biolog y 45: 307315. Shan, HY, N Zhan, CJ Liu, GX Xu, J Zhang, ZD Chen, HZ Kong 2007 Patterns of gene duplication and functional diversification during the evolution of the AP1/SQUA subfamily of plant MADS -box genes. Molecular Phylogenetics and Evolution 44: 2641. Sharma, HP 1954 Studies in the order Centrospermales I Vascular anatomy of certain species of Portulacaceae. Journal of the Indian Botanical Society 33: 98111 Sharma, HP 1963 Contributions to the morphology of the Nyctaginaceae. 2: Floral anatomy of some species. Proceedings of the Indian Academy of Sciences 57: 149 163. Shimada, S, H Otsuki, M Sakuta 2007 Transcriptional control of anthocyanin biosynthetic genes in the caryophyllales. Journal of Experimental Botany 58: 957967. Shimada, S, YT Inoue, M Sakuta. 2005. Anthocyanidin synthase in non anthocyanin -producing caryophyllales species Plant J ournal Vol. 44. Soltis DE, PS Soltis, PK Endress, MW Chase 2005 Phylogeny and Evolution of Angiosperms. Sinauer, Sunderland. 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 Prince, WJ Kress, KC Nixon, JS Farris 2000 Angiosperm phylogeny inferred from 18S rDNA, rbcL and atpB sequences. Botanical Journal of the Linnean Society 133: 381461. Soltis, PS, DE Soltis, MW Chase 1999 Angiosperm phylogeny inferred from multiple genes as a tool for comparative biology. Nature 402: 402404. Sommer, H, JP Beltran, P Huijser, H Pape, WE Lonnig, H Saedler, Z Schwarzsommer 1990 DEFICIENS a homeoti c gene involved in the control of flower morphogenesis in Antirrhinum majus the protein shows homology to transcription factors. EMBO Journal 9: 605 613. Sommer, H, W Nacken, P Beltran, P Huijser, H Pape, R Hansen, P Flor, H Saedler, Z Schwarzsommer 1991 Properties of DEFICENS a homeotic gene involved in the control of flower morphogenesis in Antirrhinum majus Development: 169175.

PAGE 210

210 Stafford, HA. 1994. Anthocyanins and betalains: Evolution of the mutually e xclusive pathways Plant Science Stellari, GM, MA Jaramillo, EM Kramer 2004 Evolution of the APETALA3 and PISTILLATA lineages of MADS -box -containing genes in the basal angiosperms. Molecular Biology and Evolution 21: 506519. Strack, D, T Vogt, W Schliemann. 2003. Recent advances in betalain research Phy tochemistry. Vol. 62. Swofford, D 2000 PAUP: Phylogenetic analysis using parsimony and other methods, version 4. Tahktajan, A 1980 Outline of the classification of flowering plants (M agnoliophyta). Botanical Review 46: 225359. Takhtajan, AL 1991 Evolution ary trends in flowering plants. Columbia University Press, New York. Tanaka, Y, N Sasaki, A Ohmiya 2008 Biosynthesis of plant pigments: Anthocyanins, betalains and carotenoids. Plant Journal 54: 733 749. 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 Molecular Biology 42: 115 149. Thompson, JD, TJ Gibson, F Plewniak, F Jeanmougin, DG Higgins 1997 The Clustal X windows interface: Flexible strategies for multiple sequ ence alignment aided by quality analysis tools. Nucleic Acids Research 25: 48764882. Thorne, R 1976 A phylogenetic classification of the Angiospermae. Evolutionary biology 9: 35 106. Thorne, R 1992 An updated classification of the flowering plants. Aliso 13: 365389. Trobner, W, L Ramirez, P Motte, I Hue, P Huijser, WE Lonnig, H Saedler, H Sommer, Z Schwarzsommer 1992 GLOBOSA a homeotic gene which interacts with DEFICIENS in the control of Antirrhinum floral organogenesis. EMBO Journal 11: 46934704. Tru e, J, H Haag 2001 Developmental systems drift and flexibility in evolutionary trajectories. Evolution of Development 3: 109119. Vandenbussche, M, J Zethof, E Souer, R Koes, GB Tornielli, M Pezzotti, S Ferrario, GC Angenent, T Gerats 2003 Toward the analys is of the Petunia MADS -box gene family by reverse and forward transposon insert ion mutagenesis approaches: B, C, and D floral organ identity functions require SEPALLATA like mads box genes in petunia. Plant Cell 15: 26802693. Vanderkrol, AR, A Brunelle, S Tsuchimoto, NH Chua 1993 Functional analysis of Petunia floral homeotic mads box gene PMADS1 Genes & Development 7: 12141228.

PAGE 211

211 Vanvinckenroye P, E Cresens, LPR Decraene, E Smets (1993) A Comparative Floral Developmental Study in Pisonia Bougainvillea a nd Mirabilis (Nyctaginaceae) with Special Emphasis on the Gynoecium and Floral Nectaries. Bulletin du Jardin Botanique National de Belgique 62: 6996. Vanvinckenroye, P, E Smets 1996 Floral ontogeny of five species of Talinum and of related taxa (Portulaca ceae). Journal of Plant Research 109: 387402. Vogt, T 2002 Substrate specificity and sequence analysis define a polyphyletic origin of betanidin 5 and 6 o -glucos yltransferase from Dorotheanthus bellidiformis Planta 214: 492495. Vogt, T, E Zimmermann, R Grimm, M Meyer, D Strack 1997 Are the characteristics of betanidin glucosyltransferases fr om cell -suspension cultures of Dorotheanthus bellidiformis indicative of their phylogenetic relationship with flavonoid glucosyltransferases? Planta 203: 349361. Wa ser, NM, MV Price 1981 Pollinator choice and stabilizing selection for flower color in delphinium -nelsonii. Evolution 35: 376390. Whipple, CJ, MJ Zanis, EA Kellogg, RJ Schmidt 2007 Conservation of B class gene expression in the second whorl of a basal gra ss and outgroups links the origin of lodicules and petals. Proceedings of the National Academy of Sciences of the United States of America 104: 10811086. Whitney, HM, M Kolle, P Andrew, L Chittka, U Steiner, BJ Glover 2009 Floral iridescence, produced by diffractive optics, acts as a cue for animal pollinators. Science 323: 130133. Wikstrom N, V Savolainen, MW Chase 2001 Evolution of the angiosperms: calibrating the family tree. Proceedings of the Royal Society London B Biological Series 268: 22112220. W illiams, SE, VA Albert, MW Chase 1994 Relationships of Droseraceae: A Cladistic Analysis of rbcL Sequence and Morphological Data. American Journal of Botany 81: 10271037 Wyman, SK, RK Jansen, JL Boore 2004 Automatic annotation of organellar genomes with d ogma. Bioinformatics 20: 32523255. Yanofsky, MF, H Ma, JL Bowman, GN Drews, KA Feldmann, EM Meyerowitz 1990 The protein encoded by the Arabidopsis homeotic gene AGAMOU S resembles transcription factors. Nature 346: 3539. Yu, DY, M Kotilainen, E Pollanen, M Mehto, P Elomaa, Y Helariutta, VA Albert, TH Teeri 1999 Organ identity genes and modified patterns of flower development in gerbera hybrida (asteraceae). Plant Journal 17: 51 62. Zachgo, S, ED Silva, P Motte, W Trobner, H Saedler, Z Schwarzsommer 1995 Functional analysis of the Antirrhinum floral homeotic DEFICIENS gene in vivo and invitro by using a temperature-sensitive mutant. Development 121: 28612875.

PAGE 212

212 Zahn, LM, HZ King, 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 preangiosperm origin with multiple duplications throughout angiosperm history. Genetics 169: 22092223. Zahn, LM, JH Leebens -Mack, JM Arrington, Y Hu, LL Landherr, CW dePamphilis, A Becker G Theissen, H Ma 2006 Conservation and divergence in the AGAMOUS subfamily of MADS -box genes: Evidence of independent suband neofunct ionalization events. Evolution and Development 8: 30 45. Zwickl, D 2000 Genetic algorithm approaches for the phylogenet ic analysis of large biological sequence datasets under the maximum likelihood criterion. PhD Thesis University of Texas at Austin

PAGE 213

213 BIOGRAPHICAL SKETCH Samuel Fraser Brockington was born 27 July 1979 in Manchester, England, son of Ian Fraser Brockington an d Diana Hilary Pink. He attended King Edwards School, Edgbaston, Birmingham. Subsequently he taught at St Marks School, Jane Furse, South Africa before attending the University of Edinburgh where he was awarded a First Class degree in Biology with honors i n Plant Science. After attending a Tropical Botany Field Course run by Walter. S. Judd of the Univeristy of Florida, he subsequently entered the graduate program at the Depeartment of Botany and Florida Museum of Natural History at the University of Florid a in 2003. He complete his PhD in 2009. He goes on to a Marie Curie Fellowship at the Department of Plant Science at the University of Cambridge, England.