( PDF ) Filogenia molecular, taxas evolutivas, tempo de divergência e herança organelar em Passiflora L. (Passifloraceae)

Download PDF

ads:

Universidade Federal do Rio Grande do Sul

Instituto de Biociências

Programa de Pós-Graduação em Genética e Biologia Molecular

FILOGENIA MOLECULAR, TAXAS EVOLUTIVAS, TEMPO DE

DIVERGÊNCIA E HERANÇA ORGANELAR EM Passiflora L.

(PASSIFLORACEAE)

Valéria Cunha Muschner

Orientador: Francisco Mauro Salzano

Co-Orientadora: Loreta Brandão de Freitas

Tese submetida ao Programa de Pós-Graduação em

Genética e Biologia Molecular da UFRGS como

requisito parcial para a obtenção do grau de Doutor

em Ciências.

Porto Alegre

Agosto de 2005

ads:

Livros Grátis

http://www.livrosgratis.com.br

Milhares de livros grátis para download.

INSTITUIÇÕES E FONTES FINANCIADORAS:

- Laboratório de Evolução Molecular, Departamento de Genética, Instituto de Biociências,

UFRGS

- Centro de Biologia Genômica e Molecular, Faculdade de Biociências, PUCRS

- Programa de Apoio a Núcleos de Excelência (PRONEX)

- Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)

- Fundação de Amparo à Pesquisa do Estado do Rio Grande do Sul (FAPERGS)

- Pró-Reitoria de Pesquisa da Universidade Federal do Rio Grande do Sul (PROPESQ-

UFRGS)

ads:

Dedico esta tese aos meus pais, que sempre me

incentivaram a fazer o que eu gosto! Amo vocês!

AGRADECIMENTOS

Agradeço em primeiro lugar a Deus.

Aos meus orientadores Francisco M. Salzano, Loreta B. Freitas e Sandro L. Bonatto.

Ao Prof. Salzano pelo exemplo, carinho e amizade!

À Loreta também pelo exemplo, carinho e amizade, mas acima de tudo pela

confiança. Por ter me apresentado o fantástico mundo das “Passifloras”! Resumindo,

obrigada por TUDO!

Ao Sandro pela ajuda nas análises!

Aos meus maravilhosos pais Willi e Iracema! Pela vida, pelos ensinamentos, pela

força... Por TUDO!

Às minhas queridas irmãs Fernanda e Adriana pelo carinho, preocupação e

companheirismo! Amo vocês!

À Taia por ser mais do que uma amiga, ser uma verdadeira irmã, me escutar, me

ajudar e sempre ter uma palavra de alento! Te adoro!

Ao Alexandre pela amizade e também por me fazer rir muito!

Aos meus queridos amigos do Laboratório de Evolução Molecular: Aline, Carlos

André, Claudia, Clênio, Dânae, Franceli, Geraldo, Jaqueline, Jéferson, Joana, Laci, Lúcia,

Pakisa, Patrícia e Priscilla pelos momentos de descontração! Todos vocês são especiais e

ocupam um lugar importante no meu coração!

À Aline pela amizade e conselhos sempre bem vindos! E também por toda força

quando estava terminando de escrever a tese! Obrigada!

À Laci, pelas caronas, pelas longas conversas sérias e também por aquelas cheias de

besteiras, pela amizade maravilhosa... Te adoro! E também por ter me ajudado a formatar e

imprimir essa tese!

Ao “Charles Andrew” pelas “dicas”, pela amizade, por me escutar e por me fazer rir!

À Pakisa pela ajuda na amplificação das amostras e amizade!

Aos meus mais novos amigos Pri e Jéferson que foram muito importantes na etapa

final da redação desta tese! Obrigada pela força!

À Jaque que também me deu uma força na etapa final! Por me escutar!

À grande amiga Verônica! Obrigada pelas jantinhas, pela amizade, pelo carinho! És

uma amiga maravilhosa! Te adoro Vê!

À Franceli pela amizade e por ter sido a minha primeira co-orientada!

Aos amigos do Departamento de Genética! Não citarei nomes para não cometer

nenhuma injustiça.

Ao Elmo e à Ellen pela competência e amizade!

À Cladinara por seqüenciar as minhas amostras!

Ao Hugo, sem ele acho que não teria conseguido finalizar a tese no prazo...

Ao Duda por ter me dado várias dicas de análise!

Aos avós do Tiago, Vô Olavo e Vó Zola, agora meus avós também, pelo carinho!

Aos meus queridos tios, Luciano e Iolanda, que mesmo longe, sempre torceram

muito por mim!

Ao CNPq por ter financiado meus estudos desde a iniciação científica.

A todos que, direta ou indiretamente, ajudaram para que esse trabalho se tornasse

possível!

SUMÁRIO

RESUMO..................................................................................................................... 7

ABSTRACT................................................................................................................. 10

CAPÍTULO I INTRODUÇÃO.................................................................................... 13

I.1 A Família Passifloraceae........................................................................................ 14

I.2 O Gênero Passiflora L............................................................................................ 15

I.3 A Classificação Taxonômica................................................................................... 16

I.4 A Origem das Angiospermas.................................................................................. 19

I.5 Taxas de Substituição Nucleotídica em Plantas...................................................... 21

I.6 O Relógio Molecular............................................................................................... 27

I.7 O Relaxamento do Relógio Molecular.................................................................... 30

I.8 Padrões de Herança Organelar em Plantas.............................................................. 32

CAPÍTULO II OBJETIVOS........................................................................................ 35

CAPÍTULO III 1º ARTIGO: PHYLOGENETIC RELATIONSHIPS AMONG

Passiflora (PASSIFLORACEAE) SPECIES: MOLECULAR DATA

STRENGTHEN A NEW TAXONOMIC PROPOSAL FOR

SUBGENERA..............................................................................................................

CAPÍTULO IV 2º ARTIGO: DIVERGENCE TIME AND EVOLUTIONARY

RATES IN Passiflora...................................................................................................

CAPÍTULO V 3º ARTIGO: ORGANELLAR INHERITANCE IN Passiflora

(PASSIFLORACEAE)................................................................................................

119

CAPÍTULO VI DISCUSSÃO..................................................................................... 136

REFERÊNCIAS BIBLIOGRÁFICAS........................................................................ 145

RESUMO

Passiflora é um gênero neotropical que apresenta uma grande variabilidade floral e

foliar, o que dificulta enormemente sua classificação taxonômica. A característica mais

marcante do gênero é a corona de filamentos em suas flores. Para melhor entender a

taxonomia de Passiflora, foram analisadas sete regiões de seu DNA, englobando os

genomas plastidial, mitocondrial e nuclear de 104 espécies. Essas espécies incluem 19 dos

23 subgêneros da classificação morfológica antiga e todos os quatro subgêneros da

classificação mais recente, também baseada em caracteres externos. Os resultados

corroboraram a proposta mais recente, que divide o gênero em quatro subgêneros

(Astrophea, Decaloba, Deidamioides e Passiflora), com a adição de mais um,

Tryphostemmatoides. Os três subgêneros com o número de espécies mais representativo

(Astrophea, Decaloba e Passiflora) formam grupos monofiléticos estatisticamente bem

fundamentados. No entanto, Deidamioides não é monofilético em todas as análises e

marcadores, enquanto que Tryphostemmatoides é representado por apenas uma espécie.

Foram também estudadas as prováveis datas de surgimento de Passiflora e sua

diversificação nos três principais subgêneros, através do estudo de quatro regiões do DNA,

englobando também os três genomas, em 70 espécies. Verificou-se que o gênero Passiflora

deve ter aparecido há cerca de 42 milhões de anos atrás (Ma); Decaloba parece ter sido o

primeiro subgênero a se estabelecer (35 Ma), enquanto que os subgêneros Astrophea e

Passiflora devem ter se diversificado há 24 Ma. Estes dois subgêneros parecem ter tido

uma radiação rápida em comparação a Decaloba, que possui árvores filogenéticas com

comprimentos dos ramos significativamente maiores que os outros dois. As adaptações

morfológicas a diferentes polinizadores devem ter sido as principais responsáveis pela

radiação rápida. A herança organelar de dois subgêneros, Decaloba e Passiflora, foi

investigada através de um e quatro híbridos interespecíficos, respectivamente. A herança

RESUMO

foi estritamente materna para a mitocôndria e o cloroplasto em Decaloba, enquanto que no

subgênero Passiflora ocorreu herança materna para o DNA mitocondrial e paterna para o

DNA plastidial. Esses resultados reforçam os dados obtidos pela filogenia, no que se refere

à diferenciação e diversificação dos subgêneros, pois há evidências de que a herança

materna dos cloroplastos seja ancestral em plantas.

ABSTRACT

Passiflora is a neotropical genus which presents a high floral and foliar variability,

making its taxonomic classification very difficult. The most marked characteristic of the

genus is the corona of filaments in its flowers. To better understand Passiflora’s taxonomy

seven regions of its DNA, including the plastid, mitochondrial, and nuclear genomes, were

studied in 104 species. These species include 19 of the 23 subgenera of the old

morphological classification and all the four subgenera of the most recent classification.

The results confirmed the most recent proposal, which divides the genus in four subgenera

(Astrophea, Decaloba, Deidamioides and Passiflora) with one more addition,

Tryphostemmatoides. The three subgenera with the most representative number of species

(Astrophea, Decaloba and Passiflora) form statistically well supported monophyletic

groups. But Deidamioides is not monophyletic in all analyses and markers, while

Tryphostemmatoides is represented by just one species. Estimations were also made of the

date of Passiflora emergence and its diversification in the three main subgenera by the

study of four DNA regions, also including the three genomes, in 70 species. The Passiflora

genus should have appeared at about 42 million years ago (Ma); Decaloba was the first

established subgenus (35 Ma) while the Astrophea and Passiflora subgenera should have

diversified at 24 Ma. These two latter subgenera seem to have had a fast radiation in

comparison to Decaloba, which presents significantly higher branch lengths in the

phylogenetic trees as compared to the two others. Morphological adaptations to different

pollinator agents could be the main responsible for this fast radiation. The organelle

inheritance of two subgenera, Decaloba and Passiflora, was investigated through the study

of respectively one and four interspecific hybrids. The inheritance was strictly maternal for

the mitochondria and chloroplast in Decaloba, while in the Passiflora subgenus maternal

inheritance for the mitochondrial DNA but paternal for chloroplast DNA occurred. These

ABSTRACT

results strengthen the phylogenetic data which differentiated the subgenera, as well as the

subgenera divergence, since there is evidence that chloroplast maternal inheritance is

ancestral in plants.

CAPÍTULO I

INTRODUÇÃO

CAPÍTULO I

I.1. A Família Passifloraceae

A família Passifloraceae ocorre, predominantemente, em áreas tropicais e

subtropicais, possuindo cerca de 17 gêneros e 650 espécies, que se distribuem nas Tribos

Paropsieae e Passifloreae. Essas espécies são trepadeiras ou lianas com gavinhas axilares

(inflorescências modificadas), ocasionalmente podem ser arbustos ou até árvores e, nestes

casos, as gavinhas estão ausentes (Judd et al. 1999). Essas plantas possuem uma grande

variabilidade foliar e floral. A característica mais marcante da família é a presença de uma

corona filamentosa em suas flores que, segundo Judd et al. (1999) apoiaria a monofilia de

Passifloraceae. Para esses autores, a tribo Paropsieae, que contém árvores e arbustos que

perderam as gavinhas, provavelmente representa um complexo basal parafilético dentro da

família, enquanto que Passiflorieae é monofilética, como evidenciado pelo hábito

escandente, gavinhas axilares e flores especializadas. No Novo Mundo a família é

representada por cinco gêneros (Passiflora, Mitostemma, Dilkea, Ancistrothyrsus e

Tetrastylis, embora esse último já tenha sido incluído ao gênero Passiflora por Feuillet &

MacDougal, 2003 e Muschner et al., 2003), todos da tribo Passifloreae. As famílias

Malesherbiaceae e Turneraceae foram incluídas em Passifloraceae pela APG II (2003), por

possuírem glicosídios cianogênicos. Além disso, Turneraceae e Passifloraceae têm

glândulas foliares e transmissão biparental ou paterna dos cloroplastos (Shore et al. 1994;

Ulmer & MacDougal 2004; V. C. Muschner et al. em preparação) e Malesherbiaceae e

Passifloraceae apresentam a corona de filamentos.

CAPÍTULO I

I.2. O Gênero Passiflora L.

O nome do gênero, passiflora, origina-se do termo “flor da paixão ou

passionflower”, atribuído por Cieza de León em 1553, numa alusão à crucificação de

Cristo. O nome do gênero foi adotado por Carl von Linné, em 1753, no Species Plantarum

(Ulmer & MacDougal 2004).

O gênero Passiflora L. é o mais representativo da família Passifloraceae possuindo

cerca de 525 espécies. Encontra-se distribuído em regiões tropicais do Novo Mundo,

raramente na Ásia e Austrália. As características do gênero são: presença de corona de

filamentos e cinco estames e órgão sexuais elevados em uma coluna conspícua, o

androginóforo. Outra característica peculiar é a grande variabilidade foliar encontrada

nesse gênero, que segundo MacDougal (1994) é a maior encontrada entre todas as

angiospermas. Suas flores também são muito variáveis em tamanho e cor, com a corona e

o perianto diversamente orientados e desenvolvidos, sendo que todas essas características

devem ter surgido de um processo co-evolutivo com os agentes polinizadores (MacDougal

1994). A reprodução funciona predominantemente por fecundação cruzada. A

autofecundação é rara e, quando ocorre, forma frutos menores e com poucas sementes

(Semir & Brown 1975). Várias classes de animais atuam como polinizadores das diferentes

espécies do gênero, mas em geral a melitofilia (abelhas) é a síndrome floral predominante;

também ocorrem, em um número menor de espécies, as síndromes da ornitofilia (beija-

flores), quiropterofilia (morcegos) e esfingofilia (mariposas) (Semir & Brown 1975;

Koschnitzke 1993; Koschnitzke & Sazima 1997; Varassin & Silva 1999; Varassin et al.

2001). A dispersão das sementes é freqüentemente feita por aves e morcegos, que são

atraídos pela coloração e pelo cheiro dos frutos maduros (Semir & Brown 1975), embora

CAPÍTULO I

pequenos mamíferos já tenham sido observados alimentando-se dos frutos de algumas

passifloras (Williams et al. 2000; Koehler-Santos et al. in press).

Devido à grande variabilidade floral e foliar encontrada em Passiflora, a

classificação taxonômica é bastante complexa.

I.3. A Classificação Taxonômica

A taxonomia de Passiflora está baseada em diversos caracteres florais e

vegetativos, levando a uma complexa subdivisão taxonômica em subgêneros, seções e

séries. De acordo com Killip (1938), o gênero poderia ser subdividido em 22 subgêneros e

dez seções, mas Escobar (1989) sugeriu mudanças nessa classificação, descrevendo um

novo subgênero. Ambas as classificações foram baseadas somente em características

morfológicas, especialmente de estrutura floral.

Mais recentemente, Feuillet & MacDougal (2003) agruparam todas as espécies de

Passiflora em quatro categorias principais ou subgêneros: Astrophea, Decaloba,

Deidamioides e Passiflora. A classificação desses autores também é baseada,

exclusivamente, em características morfológicas.

Nosso grupo já vem estudando este gênero há alguns anos quanto a aspectos

moleculares, tanto no que se refere à delimitação taxonômica (Muschner et al. 2003)

quanto aos processos de especiação (Lorenz-Lemke et al. 2005). Muschner et al. (2003),

baseados na análise de duas seqüências de DNA não codificadoras (ITS do rDNA nuclear e

espaçador intergênico trnL-trnF do cloroplasto) e uma codificadora (rps4 do cloroplasto),

sugeriram a redistribuição dos 12 subgêneros estudados (segundo a classificação de Killip

1938) para apenas três, devido à formação de três clados com altos valores de suporte

CAPÍTULO I

estatístico (bootstraps > 99). Interessantemente, foram evidenciados diferentes tamanhos

de ramos entre os três clados. O tempo de geração de cada espécie pode estar envolvido na

diferença encontrada entre dois dos clados (Passiflora e Decaloba). Apesar da pouca

informação sobre fatores ecológicos que possam estar envolvidos nesse processo, Benson

et al. (1975) foram os primeiros a afirmar que espécies do subgênero Decaloba de Killip

(pertencente ao clado Decaloba de Muschner et al. 2003) possuem tempo de geração mais

curto que espécies do subgênero Passiflora (componente do clado Passiflora de Muschner

et al. 2003). É possível que este fator possa estar acelerando a taxa evolutiva nas espécies

do subgênero Decaloba. Os resultados de Muschner et al.. (2003) corroboram

perfeitamente a nova classificação proposta por Feuillet & MacDougal (2003), podendo os

clados observados serem nomeados como subgêneros.

O subgênero Astrophea é, indubitavelmente, o mais diferenciado dentro do gênero

Passiflora, pois algumas dessas espécies não se parecem muito com as espécies do gênero

à primeira vista. São 57 espécies de árvores, arbustos ou lianas arbustivas, com folhas não-

lobadas, sendo a maioria nativa do norte da América do Sul. Os arbustos e, principalmente,

as árvores, nos quais foram perdidas as gavinhas, diferem consideravelmente das outras

espécies trepadeiras devido ao hábito e presença de crescimento secundário do lenho. A

grande árvore P. macrophylla, por exemplo, produz as maiores folhas do gênero, podendo

atingir até 95 cm de comprimento. Esse subgênero é dividido por Feuillet & MacDougal

(2003) em duas superseções e cinco seções. As espécies da seção Astrophea são arbóreas

ou arbustivas e possuem flores com tubo floral curto e coloração branca, o que forma um

contraste com a corona amarelada. Presume-se que essas espécies sejam polinizadas por

grandes abelhas, mas a literatura carece de estudos sobre este tema. Já a seção

Boryastrophea tem espécies que são predominantemente lianas e apresentam flores

CAPÍTULO I

alaranjadas ou purpúreas, com um tubo floral conspícuo, mais longo que as sépalas, e uma

corona de filamentos reduzida. A ecologia floral destas espécies sugere que elas sejam

polinizadas por beija-flores.

O subgênero Decaloba inclui mais de 200 espécies, geralmente pequenas

trepadeiras, com a maioria das espécies possuindo flores pequenas e folhas variegadas ou

bi-lobadas. A grande maioria das espécies é polinizada por abelhas ou vespas, mas existe

um número delas adaptadas à polinização por beija-flores (especialmente aquelas com

flores avermelhadas e androginóforo longo como, por exemplo, P. murucuja), além de uma

comprovadamente polinizada por morcegos noturnos (P. penduliflora). A maioria das

espécies possui flores brancas ou esverdeadas; no entanto, a palheta de cores florais é

quase completa neste subgênero. Atualmente as espécies estão distribuídas em oito

superseções, sendo a superseção Decaloba a maior delas com cerca de 120 espécies. Nesta

superseção foram incluídos os subgêneros Murucuja, Pseudomurucuja, Psilanthus e

Astephia de Killip (1938), além das espécies de Decaloba. As espécies do subgênero

Adopogyne de Killip (1938) foram incluídas na superseção Multiflora.

O subgênero Passiflora compreende cerca de 240 espécies caracterizadas por flores

grandes, que geralmente têm uma corona com faixas de diversas cores. A corona é também

a plataforma para as abelhas e outros insetos que são atraídos pelo odor destas flores. Esse

subgênero contém também muitas espécies com a corona de filamentos reduzida e que são

polinizadas por beija-flores. A maioria delas é classificada na superseção Tacsonia. As

superseções Distephana e Coccinea também apresentam espécies polinizadas por beija-

flores. É neste subgênero que podem ser encontradas plantas de importância econômica,

tais como P. edulis, P. ligularis e P. tarminiana. Com exceção de P. edulis f. edulis, P.

tarminiana e P. foetida, quase todas as espécies do subgênero Passiflora são auto-

CAPÍTULO I

incompatíveis e requerem polinização cruzada. Os subgêneros Adenosepala,

Calopathanthus, Distephana, Dysosmia, Dysosmioides, Granadilla, Granadillastrum,

Manicata, Rathea, Tacsonia, Tacsonioides e Tacsoniopsis de Killip (1938) e Escobar

(1989) foram agregados ao subgênero Passiflora, que se encontra dividido em seis

superseções.

O subgênero Deidamioides é o menor de todos, contendo apenas 13 espécies

relativamente primitivas. Esse estado primitivo é confirmado pelo surgimento das flores

diretamente a partir das gavinhas, um fenômeno raro no gênero Passiflora. Deidamioides é

subdividido em cinco seções, duas das quais são monoespecíficas. O gênero Tetrastylis foi

incluído neste subgênero dentro da seção Tetrastylis e as duas espécies que o compõem

passaram a ser denominadas de P. ovalis e P. contracta. Neste subgênero foram incluídos

os subgêneros Deidamioides, Tryphostemmatoides e Polyanthea de Killip (1938).

I.4. A Origem das Angiospermas

Segundo Charles Darwin a origem das angiospermas seria “um abominável

mistério”, e ainda hoje esta questão permanece como um problema altamente controverso.

Acredita-se que a radiação das angiospermas tenha ocorrido há cerca de 115 milhões de

anos (Ma) atrás, aproximadamente na metade do Cretáceo, tendo o grupo dominado a flora

terrestre há cerca de 90 Ma, situação que continua até os dias de hoje (Lidgard & Crane,

1988).

As relações entre as linhagens de angiospermas têm sido uma questão de difícil

resolução, pois o enraizamento do clado, usando dados morfológicos, é problemático e o

registro fóssil é insuficiente (Crane et al. 1995). Durante as duas últimas décadas, têm sido

CAPÍTULO I

alcançados progressos significativos no sentido da resolução deste problema. Análises

filogenéticas combinando caracteres morfológicos e seqüências de DNA têm procurado

resolver as relações entre as linhagens (Donoghue & Doyle 1989; Chase et al. 1993, 2000;

Doyle et al. 1994; Soltis 1997; Nandi et al. 1998; Qiu et al. 1999; Soltis et al. 1999a, 2000;

Savolainen et al. 2000a, b; Wikström et al. 2001; APG II 2003; Hilu et al. 2003), e têm

sido observados padrões altamente congruentes com relação ao enraizamento do clado

(Mathews & Donoghue 1999; Qiu et al. 1999; Soltis et al. 1999a; Sanderson 2003a; Davis

et al. 2004; Bell et al. 2005). Estes achados têm renovado o interesse e o foco para o

registro fóssil, particularmente aqueles dos depósitos do Cretáceo.

Davies et al. (2004) usaram a abordagem metodológica de supertrees para

reconstruir uma árvore datada das angiospermas a partir de dados moleculares. Eles

concluíram que suas análises indicam um padrão ligeiramente lábil da taxa de

diversificação, e corroboram a suspeita de Darwin de que não existe uma explicação

simples para o mistério da diversificação desse grupo de plantas. A calibragem da

diversificação utilizada para as maiores linhagens não mostrou uma radiação rápida recente

das basais. Houve, porém, numerosas mudanças nas taxas de diversificação com aumentos

grandes nas taxas evolutivas em períodos recentes. A diversificação não seria dirigida por

poucas grandes inovações-chave, mas por um processo mais complexo no qual haveria

instabilidade com “vencedores” e “perdedores” em todos os níveis e mudanças repetidas.

Os registros fósseis mais antigos das angiospermas são os de pólen do período

Valanginiano-Hauterivano, 141-132 Ma antes do presente, AP (Brenner & Bickoff 1992;

Huge 1994; Brenner 1996). O depósito do Aptiano-Albiano (125-97 Ma AP) de Portugal,

que contém rápida expansão de diversidade morfológica em flores, sementes e pólen,

também foi bastante estudado (Friis et al. 1999). Outros estudos, referentes a depósitos do

CAPÍTULO I

Cenomaniano-Campaniano, têm indicado idades mais recentes que variam de 97-74 Ma

AP (Basinger & Dilcher 1984; Herendeen et al. 1999). A discussão sobre a idade das

angiospermas é ampla e longa, e até que sejam encontrados registros fósseis mais

completos, não será esgotada com facilidade. Uma alternativa que pode ajudar na

estimativa do tempo de divergência deste grupo é a utilização de seqüências de DNA.

Wikström et al. (2001), baseando-se no seqüenciamento do DNA de três regiões

codificadoras (genes rbcL e atpB do DNA plastidial e 18S do DNA ribossomal nuclear),

estimaram que as angiospermas teriam surgido entre o final e a metade do Jurássico (179-

158 Ma AP) e que as eudicotiledôneas teriam surgido do início do Jurássico até a metade

do Cretáceo (147-131 Ma AP). Estas estimativas são mais antigas que as obtidas com os

registros fósseis, mas estão sendo bem aceitas por diversos pesquisadores em tentativas de

datar a origem de diversos grupos de plantas (Sanderson 2003a; Davies et al. 2004; Bell et

al. 2005). No presente estudo estaremos aceitando a proposição de que o ramo que deu

origem à família Passifloraceae teria surgido há 36-32 Ma AP (Wikström et al. 2001).

I.5. Taxas de Substituição Nucleotídica em Plantas

I.5.1. Cloroplasto

O genoma plastidial (cpDNA) é uma molécula com aproximadamente 150 mil

pares de bases (kbp), que codifica cerca de 100 funções genéticas (Clegg & Zurawski

1991) e sua organização molecular e evolução têm sido amplamente estudadas. Estudos de

mapeamento gênico dessa organela em algas e plantas terrestres confirmaram a impressão

de forte conservação de seu conteúdo gênico (Clegg et al. 1994; Martin et al. 2005). Entre

CAPÍTULO I

as plantas terrestres o conteúdo gênico é quase totalmente conservado, embora tenham sido

demonstradas transferências de função do genoma plastidial para o nuclear (Downie &

Palmer 1991; Huang et al. 2005).

O conteúdo gênico conservado e a taxa de substituição nucleotídica relativamente

baixa em genes codificadores de proteínas, têm tornado o genoma plastidial ideal para

estudos evolutivos em plantas (Clegg 1993; Martin et al. 2005). Durante os últimos dez a

quinze anos, houve uma explosão de publicações sobre filogenias moleculares construídas

a partir do cpDNA (p. ex. Chase et al. 1993; Soltis et al. 2000; Stefanović & Olmstead

2004; Salamin et al. 2005; Young & dePamphilis 2005).

O interesse por filogenias precisas supera o âmbito puramente descritivo. A origem

das adaptações morfológicas pode ser colocada em um contexto filogenético para

reconstruir a seqüência precisa das mudanças genéticas (e moleculares) que originaram

novas estruturas (p. ex., Clegg et al. 1994) ou os genes de desenvolvimento são usados

como marcadores filogenéticos para explicar a diversidade na estrutura floral, que ainda é

o principal elemento de identificação taxonômica (Becker & Theiβen 2003; Kaufmann et

al. 2005).

Estudos têm demonstrado, por outro lado, que alguns genes do cloroplasto violam a

constância do relógio molecular, porque as taxas evolutivas variam entre as linhagens mais

abrangentes de plantas (Bousquet et al. 1992; Gaut et al. 1992). No entanto, tem sido

sugerido que a maioria dessas variações seria devida a diferenças no tempo de geração

(Gaut et al. 1992; Doyle & Gaut 2000). Segundo esses últimos autores, tanto o relógio

molecular calibrado pelo tempo, quanto o calibrado pelo tempo de geração são consistentes

com a Teoria da Neutralidade.

CAPÍTULO I

A grande maioria das análises filogenéticas envolvendo grandes grupos

taxonômicos em níveis acima de famílias envolvem o gene rbcL, codificador da

subunidade maior da rubisco (Soltis et al. 2000; APG II 2003; Salamin et al. 2005). O gene

rps4, codificador da proteína 4 da subunidade menor do ribossomo plastidial, tem sido

pouco utilizado em estudos sobre a filogenia de plantas (Nadot et al. 1994; Souza-Chies et

al. 1997; Soltis et al. 2002; Rydin et al. 2004). No entanto, o estudo realizado por

Muschner et al. (2003), com 35 espécies do gênero Passiflora, demonstrou que essa região

tem grande potencial para desvendar as relações filogenéticas dentro deste gênero.

Regiões não-codificadoras do cpDNA têm sido amplamente utilizadas nas análises

filogenéticas em níveis taxonômicos intra e interespecíficos (Mes et al. 2000; Muschner et

al. 2003), devido às taxas de substituição nucleotídica elevadas. Dentre as regiões

plastidiais não-codificadoras melhor estudadas nas análises filogenéticas de angiospermas

encontram-se o espaçador intergênico trnL-trnF e o intron do gene trnL (Mes et al. 2000;

Holt et al. 2004; Lledó et al. 2005). Em 1991, Taberlet et al. publicaram primers universais

para essas regiões, os quais vêm sendo utilizados numa extensa gama de espécies com

excelentes resultados (Chen et al. 2005; Wang et al. 2005).

Por outro lado, as seqüências não-codificadoras, algumas vezes, contêm mais

inserções/deleções (indels) que substituições (Golenberg et al. 1993), as quais podem

dificultar o alinhamento das seqüências e a determinação das homologias (Kelchner 2000).

Enquanto alguns autores argumentam que os indels não devem ser tratados como

caracteres informativos (Golenberg et al. 1993), outros alegam que eles contêm

informação filogenética importante e que, por esse motivo, devem ser incluídos nas

análises (p.ex. Kelchner 2000).

CAPÍTULO I

I.5.2. Genes Nucleares

Um gene nuclear amplamente utilizado para a filogenia das angiospermas é o 18S

rDNA (p.ex. Chaw et al. 2000; Soltis et al. 2000). O DNA ribossomal (rDNA) é a região

do genoma que codifica os componentes do RNA dos ribossomos. O rDNA eucariótico

está organizado em tandem, com milhares de cópias no genoma. Cada unidade de repetição

consiste de genes que codificam a subunidade menor (18S), a subunidade maior (26S) e o

rDNA nuclear 5,8S, sendo que essas regiões codificadoras são separadas por espaçadores

(Schlötterer 1998). Essas cópias são homogeneizadas por evolução em concerto, que pode

ocorrer devido à permuta desigual e/ou conversão gênica (Hamby & Zimmer 1992; Koch

et al.. 2003).

Nickrent & Soltis (1995) estudaram a taxa evolutiva e a resolução filogenética do

gene 18S rDNA inteiro bem como as do gene rbcL de 59 angiospermas. A comparação

mostrou que o rbcL é cerca de três vezes mais variável que o 18S rDNA. No entanto, por

causa do maior comprimento deste último, a razão do número de sítios filogeneticamente

informativos por molécula é somente 1,4 vezes maior para o rbcL que para o 18S rDNA.

As análises de parcimônia mostraram que diversos clados foram fortemente apoiados pelos

dois genes, levando os autores a concluir que as seqüências do 18S rDNA inteiras são

suficientemente variáveis para o desenvolvimento de estudos filogenéticos no grupo das

angiospermas.

Diversos autores têm publicado filogenias robustas das angiospermas a partir do

gene 18S rDNA, por exemplo, Soltis et al. (1997), que analisaram 223 espécies, e Soltis et

al. (2000), que estudaram 560 espécies. Apesar disso, o 18S rDNA sozinho forneceu

poucos caracteres filogeneticamente informativos para resolver adequadamente as relações

CAPÍTULO I

entre e dentro dessas plantas (Soltis et al. 1999b; Kim et al. 2004). Kuzoff et al. (1998)

demonstraram que a seqüência do gene 26S rDNA tem um grande potencial para a

reconstrução filogenética, em níveis taxonômicos comparáveis aos investigados com o 18S

rDNA. Eles concluíram ainda que o 26S rDNA evolui de 1,6 a 2,2 vezes mais rápido e tem

cerca de 3,3 vezes mais sítios informativos que o 18S rDNA. Além disso, propuseram que

os segmentos de expansão do 26S rDNA evoluem de 1,2 a 3 vezes mais rápido que o gene

rbcL, com 1,5 vezes mais sítios informativos. A utilidade filogenética dessa região em

angiospermas tem sido demonstrada em diversos estudos (p. ex. Fishbein et al. 2001; Zanis

et al. 2002; Korall & Kenrick 2004; Schönenberger et al. 2005).

I.5.3. A Mitocôndria

Wolfe et al. (1987) mostraram que as taxas de substituição sinônimas em genes

mitocondriais de angiospermas são anormalmente baixas (cerca de 50 a 100 vezes menor)

quando comparadas às dos genes mitocondriais de mamíferos. Além disso, demonstraram

que as taxas de substituição do mtDNA são algumas vezes menores que às do cloroplasto e

10 a 20 vezes menores que as taxas de substituição dos genomas nucleares. Palmer &

Herbon (1988) confirmaram que as taxas de substituição nucleotídica do genoma

mitocondrial inteiro são baixas (inclusive em regiões não codificadoras) e mostraram

separação entre as taxas de evolução de seqüências e estrutura. Estudos subseqüentes têm

confirmado essas taxas de substituição muito baixas (Gaut 1998; Muse 2000). No entanto,

variações moderadas nas taxas de substituição sinônimas (maiores que 7 vezes) têm sido

encontradas na comparação entre diversos grupos de plantas (Eyre-Walker & Gaut 1997;

Laroche & Bousquet 1999; Whittle & Johnston 2002). Na maioria dos casos, a correlação é

CAPÍTULO I

feita com genes plastidiais e / ou nucleares. Forças evolutivas que operam nos dois

genomas organelares ou nos três genomas, como transmissão dos genomas organelares ou

efeito do tempo de geração, respectivamente, são explicações para esses padrões. Estudos

filogenéticos, embora desenvolvidos sem um enfoque quantitativo, sugerem que exista

uma certa variação entre os diversos grupos de plantas (Beckert et al. 1999; Bowe et al.

2000; Chaw et al. 2000; Barkman et al. 2004; Davis et al. 2004). Cho et al. (2004)

encontraram taxas de substituições sinônimas do mtDNA extraordinariamente elevadas e

variáveis no gênero Plantago (Plantaginaceae). Estes autores tentaram explicar essa alta

taxa de variação no gênero e argumentaram que a transferência de genes mitocondriais

para o núcleo, os níveis elevados de edição de RNA, o tempo de geração, e mecanismos de

mutação e reparo poderiam ser os responsáveis.

Mas o interesse em incluir um terceiro genoma nas análises filogenéticas tem

aumentado. O mtDNA tem sido relativamente pouco utilizado na filogenia de plantas

devido, também, ao alto grau de recombinação intra-molecular que apresenta na maioria

das espécies vegetais estudadas, o que torna difícil o estudo de grandes regiões de seu

DNA (Palmer & Herbon 1988). A análise de pequenas regiões do mtDNA pode vir a

minimizar este problema. Por esse motivo, regiões intrônicas e espaçadoras têm-se

mostrado uma promessa na sistemática molecular de níveis taxonômicos moderadamente

mais baixos (Demesure et al. 1995; Freudenstein & Chase 2001; Duminil et al. 2002;

Dombrovska & Qiu 2004). Um dos primeiros destes fragmentos do mtDNA a ser

investigado é um intron do gene da NADH desidrogenase. A subunidade I desse gene

(nad1) tem cinco exons (a-e) em Oenothera, alguns dos quais estão altamente dispersos e

envolvidos no mecanismo de processamento em trans (Wissinger et al. 1991). O intron

entre os exons b e c é relativamente pequeno (de 1422-1464 pb em Triticum, Citrullus e

CAPÍTULO I

Oenothera). Esse é um intron do grupo II, que está envolvido no mecanismo de

processamento em cis, cuja estrutura secundária característica facilita seu auto-

processamento. Na primeira vez que essa região foi usada em análises filogenéticas, foi

identificada pouca variação inter e intra-específica (Demesure et al. 1995). No entanto,

Chen & Sun (1998) encontraram variação comparável à de outros marcadores plastidiais e

nucleares em espécies de Spiranthes, e Freudenstein & Chase (2001) observaram ampla

variação, principalmente relacionada a eventos de inserção/deleção, dentro da família

Orchidaceae. Embora essa não seja uma região codificadora, acredita-se que esteja sujeita

a pressão seletiva devido à estrutura de sua alça principal, necessária para o auto-

processamento. Outras regiões do mtDNA parecem conter informações filogenéticas muito

úteis (Freudenstein & Chase 2001), principalmente relacionadas com eventos de

inserção/deleção. A subunidade cinco do gene da NADH desidrogenase também possui

cinco exons e introns evolvidos com o processamento em trans (Souza et al. 1991).

Laroche et al. (1997) encontraram uma boa variação entre espécies de angiospermas para o

intron entre os exons d e e, e embora até o momento essa região não tenha sido estudada

em análises filogenéticas de angiospermas, ela parece promissora, pelo menos para

algumas espécies.

I.6. O Relógio Molecular

Até pouco tempo o estudo da evolução dos organismos era baseado somente em

registros fósseis e evidências genéticas a nível protéico. Com o advento da biologia

molecular e o surgimento de ferramentas poderosas de bioinformática, a facilidade de

acessar e analisar seqüências de DNA tem aumentado, tornando possível, portanto, a

CAPÍTULO I

comparação entre espécies através de seus genomas. Com essas novas tecnologias foram

desenvolvidas filogenias moleculares detalhadas para diversos grupos de plantas, nos mais

diferentes níveis taxonômicos.

Uma das pressuposições usadas para a produção de filogenias moleculares é a

hipótese do relógio molecular. Essa hipótese propõe que a taxa de mudanças evolutivas em

nível de seqüências de DNA é constante em diferentes linhagens. A hipótese ainda é

bastante controversa, gerando discussões entre os biólogos evolutivos, especificamente o

debate neutralista x selecionista. Os neutralistas tomam o relógio molecular como base

para inferências sobre a evolução dos organismos supondo a constância das taxas de

mutação neutras (Kimura 1968, 1969, 1983). Já os selecionistas consideram as variações

encontradas no relógio molecular como evidência contra essa constância. As taxas de

substituições nucleotídicas, ou fixação, em um sítio nucleotídico por ano (k) em uma

população diplóide de tamanho 2N, são iguais ao número de novas mutações (neutras,

deletérias ou vantajosas) que surgirem por ano (u) multiplicado por sua probabilidade de

fixação (f). A expressão matemática para isso é k = 2N.u.f (Li, 1997). Para uma mutação

neutra, a probabilidade de fixação é simplesmente o inverso do tamanho populacional

(1/2N); então a taxa de substituição de uma mutação deste tipo é k = (2N).(1/2N).u = u.

Portanto, a sua taxa de substituição depende somente da taxa de mutação e não de outros

fatores, como o tamanho populacional.

Kimura (1968) postulou que a maioria das mutações, em nível molecular, seria

devida à fixação aleatória das mutações neutras ou quase neutras. Essa hipótese é hoje

conhecida como a teoria neutra da evolução molecular (Kimura, 1983). No entanto,

durante os anos 70 muitos dados comparativos, principalmente para seqüências protéicas

de mamíferos, forneceram exemplos de heterogeneidade no ritmo evolutivo e o debate

CAPÍTULO I

sobre a constância do relógio molecular começou. Wu & Li (1985) demonstraram que as

taxas evolutivas de roedores são maiores que as de humanos. As variações encontradas

entre as taxas de substituições podem ser causadas pelas diferenças no tempo de geração

entre as espécies, ou seja, espécies que têm um tempo de geração mais curto teriam taxas

de substituição nucleotídica maiores que aquelas que têm um tempo longo de geração. Por

exemplo, Gaut et al. (1992), através de análises realizadas com o gene rbcL do cloroplasto,

verificaram que em gramíneas as taxas de substituições nucleotídicas eram cinco vezes

maiores que em palmeiras. No entanto, Whittle & Johnston (2003), após a análise de 24

pares de espécies filogeneticamente independentes (cada par contendo uma espécie com

história de vida anual e outra perene, ou espécies arbóreas com tempos de geração

respectivamente curto e longo), não encontraram qualquer evidência de que o tempo de

geração estivesse correlacionado com as taxas evolutivas das regiões nucleares estudadas.

A avaliação precisa do relógio molecular exige que o tamanho dos ramos de uma

árvore filogenética sejam proporcionais; para satisfazer esta exigência há, muitas vezes, a

remoção de algumas espécies da árvore, o que, em muito casos, impossibilita a datação das

divergências, além de fazer uso ineficiente dos dados (Yang & Yoder 2003).

A definição do tamanho do ramo é um dos principais motivos da deficiência dos

métodos que utilizam o pressuposto do relógio molecular. Este é interpretado como o

produto da taxa de evolução (µ) pelo tempo (T), que representa a quantidade de evolução

ocorrida desde o evento da cladogênese. Ao contrário do que é comumente pensado, o

tamanho do ramo não representa a taxa evolutiva µ (Yang & Yoder 2003). Para calcular o

tamanho de um ramo é necessário que se suponha a constância de um dos fatores acima,

para que se calcule o outro. Em uma árvore de máxima verossimilhança ou de distância em

que os ramos variam livremente, supõe-se que T é igual a 1, sendo as diferenças

CAPÍTULO I

observadas oriundas exclusivamente da variação na taxa µ. O grande problema dos

métodos que testam o relógio molecular é que eles não calculam µ e T independentemente,

ou seja, sem supor a constância das taxas ao longo dos ramos. A decomposição das duas

quantidades que definem o tamanho de um determinado ramo não é simples, pois a

ausência de informações a priori impossibilita tal cálculo. No entanto, é possível obter-se

informações em relação aos valores de T devido ao registro fóssil, permitindo que se

conheça algo com relação aos tempos de divergência entre as espécies.

I.7. O Relaxamento do Relógio Molecular

Sanderson (1997) foi o primeiro a explorar a questão da decomposição do tamanho

dos ramos de uma árvore filogenética. Ele propôs um método que estima o tempo de

divergência quando as taxas evolutivas variam entre as linhagens, supondo que as

mudanças são autocorrelacionadas e que a taxa de mudança é herdada, a partir de uma

linhagem ancestral, por seus descendentes imediatos. Através de técnicas de otimização, o

método procura pela solução que minimiza as taxas inferidas de mudança.

O programa “r8s”, criado a partir do método proposto e descrito acima (Sanderson

2003b), estima as taxas absolutas de evolução molecular e os tempos de divergência. A

estimativa dos parâmetros é feita de diversas formas, a partir dos padrões de máxima

verossimilhança, em um contexto global, ou postulando um relógio molecular local, para

métodos experimentais semiparamétricos ou não paramétricos que relaxam a estringência

de sua admissão usando procedimentos homogêneos. O ponto de partida é uma árvore

filogenética e a estimativa do comprimento dos ramos (número de substituições ao longo

dos ramos). Além disso, podem ser adicionados um ou mais pontos de calibragem para

CAPÍTULO I

permitir uma escala das taxas e tempos para unidades reais. Essas calibragens podem ser

realizadas de duas maneiras: a escolha de uma idade fixa para um nó da árvore, ou a

imposição de uma idade mínima ou máxima em um determinado nó, o que geralmente

reflete o conteúdo de informações obtidas a partir das evidências fósseis. Os nós terminais

podem ocorrer em qualquer ponto de calibragem, permitindo a investigação da taxa de

variação nas filogenias. Finalmente, é possível determinar-se todos os tempos de

divergência e avaliar a taxa de variação molecular a partir de diversos modelos de

ajustamento (smoothing).

O algoritmo de Sanderson (2002) não utiliza distribuições probabilísticas para

modelar a evolução das taxas e dos tempos e, portanto, é fundamentalmente não-

paramétrico.

Thorne et al. (1998), Kishino et al. (2001) e Thorne & Kishino (2002) propuseram

uma outra solução para o problema da dissociação de tempos e taxas. Esses autores

admitiram explicitamente distribuições probabilísticas. O método usa uma abordagem

baiesiana para inferir os tempos de divergência e as taxas evolutivas. A idéia básica é

adotar distribuições a priori dos parâmetros µ e T e, posteriormente, verificar o impacto

dos dados (alinhamento das seqüências e pontos de calibragem) nessas distribuições.

Este enfoque tem sido amplamente utilizado no cálculo de tempos de divergência,

pois possibilita o uso de múltiplos genes e de modelos evolutivos complexos, ao contrário

do método de Sanderson (1997, 2002). Além disso, incorpora a adoção de intervalos

temporais para os pontos de calibragem. Simulações feitas pelos autores indicam que o

método é robusto e valores a priori diferentes dos parâmetros tendem a convergir para uma

mesma distribuição posterior (Kishino et al. 2001, Thorne & Kishino 2002).

CAPÍTULO I

I.8. Padrões de Herança Organelar em Plantas

As plantas possuem três genomas: nuclear, plastidial e mitocondrial. Cloroplastos e

mitocôndrias são organelas eucarióticas de origem endossimbiôntica. A maioria das células

eucarióticas contém dezenas a centenas de mitocôndrias que geram energia, enquanto as

células das plantas contêm dezenas a centenas de cloroplastos que desempenham a

fotossíntese como principal função. Mitocôndrias e cloroplastos são herdados de uma

maneira não-Mendeliana em todos os organismos estudados (revisões em Birky 1995,

2001). A herança dos genomas citoplasmáticos é freqüentemente materna, mas existem

numerosas exceções que resultam em diferentes graus de herança paterna ou biparental do

mtDNA ou cpDNA (Koperlainen 2004). Essa variedade de padrões de herança sugere que

têm sido adotadas diferentes estratégias entre os diferentes organismos considerados. A

perda do padrão universal da herança materna também indica que o sistema não-

Mendeliano não é provavelmente uma mera conseqüência da assimetria no tamanho dos

gametas. Por causa dos variáveis graus de herança uniparental, segregação durante as

divisões mitóticas e meióticas e múltiplas cópias desses genomas em cada célula,

processos evolutivos que agem nos genomas do cloroplasto e da mitocôndria diferem

daqueles que governam os de genomas nucleares.

A maioria das angiospermas exibe herança materna do cpDNA, mas cerca de um

terço dos gêneros investigados mostra algum grau de herança biparental dos cloroplastos

(Corriveau & Coleman 1988, Mogensen 1996, Zhang et al. 2003). Por outro lado, pouco se

sabe sobre a herança do mtDNA em plantas, mas esse parece ser maternalmente herdado

(Sodmergen et al. 2002; Mohanty et al. 2003). Dentre as angiospermas melhor

investigadas estão o gênero Actinidia, que possui herança estritamente paterna do cpDNA

CAPÍTULO I

(Testolin & Cipriani 1997, Chat et al. 1999); as espécies Medicago sativa e Turnera

ulmifolia, que exibem herança paterna, materna e biparental (Shore & Triassi 1998); e o

gênero Pelargonium, que possui um padrão de herança biparental para ambos cpDNA e

mtDNA (Guo & Hu 1995). A herança paterna do mtDNA já foi detectada em

Cyclobalanopsis glauca (Lin et al. 2003). Casos de padrões de herança organelar não-

usuais têm sido registrados na análise de híbridos interespecíficos ou intragenéricos

produzidos artificialmente, como é o caso do gênero Larrea, onde a herança do genoma

plastidial é paterna (Yang et al. 2000). Em cruzamentos entre Citrus-Poncirus (Moreira et

al. 2002), a herança do mtDNA é parcialmente biparental.

Dentre as gimnospermas, as coníferas herdam o cpDNA, exclusiva ou

predominantemente do genitor masculino, enquanto que outros grupos, como Ephedra,

Ginkgo e Zamia, parecem ter herança materna dos cloroplastos (Mogensen 1996;

Koperlainen 2004). Dependendo do grupo de gimnospermas, o mtDNA pode ter herança

materna, paterna ou biparental.

Existe uma variedade de mecanismos pelos quais as organelas podem ser ou não

transmitidas para a prole. Os mecanismos que resultam na supressão da herança

citoplasmática paterna em angiospermas incluem a exclusão ou perda de organelas

citoplasmáticas das células germinativas ou espermáticas, a exclusão do citoplasma

masculino na fusão gamética e a degradação do DNA organelar dentro da célula

germinativa e/ou células espermáticas (Morgensen 1996, Nagata et al. 1997). A eliminação

materna pode ser ocasionada pela transformação das organelas citoplasmáticas ao longo do

ovo/zigoto ou pela sua degeneração antes da fusão gamética (Morgensen 1996, Brums &

Owens 2000). No entanto, nenhum desses mecanismos de eliminação é perfeitamente

efetivo em todos os casos.

CAPÍTULO I

Liu et al. (2004) acreditam que o desenvolvimento do controle da herança do

genoma dos cloroplastos ocorreu independentemente nas angiospermas, sendo possível que

a herança materna tenha se tornado dominante antes do seu surgimento Estes autores

pressupõem, ainda, que os mecanismos para a herança dos cloroplastos nessas plantas

devam ter se desenvolvido independentemente e mais tardiamente que os de herança da

mitocôndria, devido ao fato de que o modo de herança dos plastídios varia

consideravelmente, estando associados aos processos de especiação.

O entendimento da diversidade genética é necessário, tanto para o desenvolvimento

de estratégias eficientes de conservação, quanto para o delineamento de programas de

exploração sustentável da biodiversidade. Análises moleculares são úteis para avaliar a

diversidade e filogenia das espécies e para a compreensão dos processos de especiação. A

determinação do modo de herança dos plastídios e mitocôndrias em plantas é um passo

importante para o uso desses genomas no traçado da história evolutiva das espécies e na

compreensão de suas implicações.

CAPÍTULO II

OBJETIVOS

CAPÍTULO II

II.1. Geral

O objetivo geral desse trabalho é contribuir para o esclarecimento da história

evolutiva do gênero Passiflora a partir da utilização de seqüências de DNA e diversos

métodos de análise.

II.2. Específicos

1. Analisar as relações filogenéticas dentro do gênero Passiflora, principalmente no

que diz respeito ao seu agrupamento em subgêneros, e contribuir para o melhor

entendimento da taxonomia clássica.

2. Correlacionar os resultados obtidos com características ecológicas e bioquímicas.

3. Datar o surgimento de Passiflora e a diversificação de seus principais subgêneros,

associando essas datas com eventos biogeográficos.

4. Determinar os modos de herança organelar no gênero e tentar correlacioná-los com

a filogenia e a datação de sua diversificação evolutiva.

CAPÍTULO III

1º ARTIGO

A ser submetido para a revista Systematic Biology

PHYLOGENETIC RELATIONSHIPS AMONG Passiflora

(PASSIFLORACEAE) SPECIES: MOLECULAR DATA

STRENGTHEN A NEW TAXONOMIC PROPOSAL FOR

SUBGENERA

Jul– xx –05

Running head: PASSIFLORA’S PHYLOGENY

Phylogenetic relationships among Passiflora (Passifloraceae)

species: Molecular data strengthen a new taxonomic proposal

for subgenera

VALÉRIA C. MUSCHNER

, ALINE P. LORENZ–LEMKE

, PAKISA D. TOGNI

, ARMANDO C.

CERVI

, SANDRO L. BONATTO

, FRANCISCO M. SALZANO

, AND LORETA B. FREITAS

Departamento de Genética, Instituto de Biociências, Universidade Federal do Rio Grande do Sul,

Caixa Postal 15053, 91501–970 Porto Alegre, RS, Brazil; E–mail: [email protected];

[email protected]r

Departamento de Botânica, Universidade Federal do Paraná, Caixa Postal, 19031, 81531–970,

Curitiba, PR, Brazil

Centro de Biologia Genômica e Molecular, Faculdade de Biociências, Pontifícia Universidade

Católica do Rio Grande do Sul, Ipiranga 6681, 90610–001 Porto Alegre, RS, Brazil

Correspondence: Loreta B. Freitas, Departamento de Genética, Instituto de

Biociências, Universidade Federal do Rio Grande do Sul, Caixa Postal 15053, 91501–

970 Porto Alegre, RS, Brazil. Phone: 55 51 33166715. Fax: 55 51 33166727. E–mail:

loreta.freitas@ufrgs.br

CAPÍTULO III

Abstract – Passiflora is a genus with more than 500 species, showing large flower

complexity and diversity; its habit range from climbing herbaceous shrubs to trees. These

characteristics condition taxonomic problems. The genus’ first classification divided it into

22 subgenera, and afterwards an additional subgenus was added to them. However,

recently a new classification system grouped the 23 subgenera in just four. To further

understand and complement a first phylogeny of the genus published by us, 104 species,

representing 19 of 23 subgenera of the first classifications, and all four of the most recent

proposal were investigated. Seven molecular markers were used: (a) the rbcL and rps4

genes, trnL intron, and trnL-trnF intergenic spacers of plastid DNA; (b) nad1 b/c and nad5

d/e of mitochondrial DNA; and (c) a partial sequence of the 26S nuclear DNA, totaling

about 6,300 base pairs. The monophyly of Astrophea, Decaloba, and Passiflora subgenera

was highly supported independently of the phylogenetic analysis employed, but the fourth

subgenus (Deidamioides), as originally proposed, proved to be polyphyletic. We hereby

classified a restriction in the delimitation of Deidamioides, and the addition of E. P.

Killip’s Tryphostemmatoides to the subgenera classification, that therefore would be

composed by five taxonomic units.

Key Words: Passiflora, phylogeny, taxonomic classification, cpDNA, mtDNA, nuclear

DNA, combined genetic analysis.

CAPÍTULO III

The Passifloraceae family of the Malpighiales order is composed by 19 genera

(APG II, 2003), one of its characteristics being the presence of cyclopentenoid cyanogenic

glycosides and of a hypanthium-like structure that does not bear the stamens. Passiflora is

the largest genus of the family with 525 species distributed especially in the tropical region

(Vanderplank, 1996; Cervi, 1997; Ulmer and MacDougal, 2004). The majority of these

species is herbaceous, but there also shrubs and trees among them. Killip (1938) and

MacDougal (1994) asserted that among the Angiosperms no other group presented such a

large foliar diversity. In addition, its flowers displayed ample variation in size and color,

with the corona and perianth showing diverse orientation and development. Coevolution

with insect pollinators has been suggested as an explanation for these features

(MacDougal, 1994). Based on morphology only (especially flower structures) Killip

(1938) and Escobar (1989) concluded that the genus could be divided into 23 subgenera,

with diverse series and sections.

Feuillet and MacDougal (2003), on the other hand, proposed a drastic taxonomic

reevaluation of the genus that according to them would consist of only four subgenera

(Astrophea, Decaloba, Deidamioides and Passiflora). The Astrophea subgenus remained

unchanged in relation to the previous classification, with 57 species divided in six sections;

214 species were attributed to Decaloba, the majority with x = 6 chromosomes, distributed

by eight supersections and five sections; 13 species, grouped into five sections, were

classified as Deidamioides; while Passiflora included more than 236 species mainly with x

= 9 chromosomes, separated in six supersections, 13 sections, and 11 series. This

classification, however, did not include Old World species. Krosnick and Freudenstein

(2005) studied species of this region (subgenus Decaloba, supersection Disemma)

confirming the supersection monophyly.

CAPÍTULO III

The first Passiflora molecular phylogeny, published by Muschner et al. (2003),

included 11 of the previously suggested 23 subgenera, studied for two non-coding regions

[the nuclear ribosomal internal transcribed spacers (nrITS) and the plastid trnL-trnF spacer

regions], while the rps4 plastid gene was also investigated for a more restrict, but

representative sample. They found three clearly defined clades (involving the Passiflora,

Decaloba, and Astrophea subgenera) while the Deidamioides subgenus remained

undefined due to the small number of studied species classified in it. Other attempts to

elucidate the genus’ phylogeny did not yield good association with morphologic and/or

ecological data, and did not agree with Feuillet and MacDougal (2003) proposition in

relation to the number of subgenera and their composition (Yockteng and Nadot, 2004;

Plotze et al., 2005).

Sequences of plastid, mitochondrial, and nuclear (especially ribosomal) DNA

have been extensively utilized to study plant (especially Angiosperm) phylogenies (Soltis

et al., 1998; Qiu et al, 1999; Kuzoff and Gasser, 2000). This strategy of combining

multiple genes with different functions from the three plant genomes should reduce the

homoplasies generated by gene-function and/or genomic specific phenomena, such as

heterogeneity of rates of change, GC-content bias, RNA editing, and protein structural

constraints (Qiu et al., 1999). Rokas et al. (2003) showed that as the number of genes

increases in a phylogenetic analysis, the better the tree reflects the species’ phylogeny. The

same type of relationship was examined by Rokas and Carrol (2005), who concluded that

for phylogenetic precision the number of the genes considered is a more important

determinant than the number of taxa examined. However, branch representativiness should

also be taken into consideration, and when a large number of taxa is being studied the ideal

number of markers should be decided in cost-benefit terms.

CAPÍTULO III

The objectives of the present work were: (a) to examine Passiflora’s three

genomes, assessing their phylogenetic utility; (b) test the genus monophyly; (c) compare

the results obtained with Feuillet and MacDougal’s (2003) infrageneric classification; (d)

evaluate different methods of phylogenetic reconstruction; and (e)generally verify the

relationships among species in a large number of such taxa in this genus.

MATERIALS AND METHODS

Taxon sampling

The 104 species representing nineteen of Killip’s (1938) and Escobar’s (1989)

subgenera, as well as all four Feuillet and MacDougal’s (2003) subgenera investigated are

listed in Table 1, together with representatives from seven other genera of Passifloraceae

(Adenia Forssk., Ancistrothyrsus Harms, Barteria Hook.f., Deidamia Noronha ex Thouars,

Dilkea Mast., Mitostemma Mast., Paropsia Noronha ex Thouars) utilized as outgroups.

Tetrastylis Barb.Rodr. (re-classified as Passiflora by Feuillet and MacDougal [2003]), was

also included in our analyses. This sampling involved all species of Passifloraceae from

which we could obtain suitable material to extract DNA, containing taxa from a wide

distributional range in South and Central America. Among the outgroups, representatives

of the two tribes of Passifloraceae were considered, as well as members of the Turneraceae

and Malesherbiaceae, included in the Passifloraceae by APG II (2003).

CAPÍTULO III

DNA Extraction, Amplification, and Sequencing

Total DNA was extracted from fresh leaves dried in silica gel or obtained from

herbarium material, using Roy et al.’s (1992) method with a few adaptations.

Seven DNA regions were sampled: the rbcL and rps4 genes, trnL intron and

trnL–trnF intergenic spacer from the plastid genome; nad1 b/c and nad5 d/e introns from

the mitochondrial genome; and a partial portion of the 26S gene from the nuclear

ribosomal genome. These regions were amplified with primers 1F and 1460R (Savolainen

et al., 2000), rps45’ and rps43’ (Souza-Chies et al., 1997), c, d, e and f (Taberlet et al.,

1991), nad1/2 and nad1/3 (Duminil et al., 2002), mt3 and mt6 (Souza et al., 1991), N–

nc26S1 and 1229r (Kuzoff et al., 1998). Sequencing primers were used as listed by these

authors except for the nad1 b/c intron, for which we constructed internal primers specific

for Passiflora. PCR products were purified using the polyethylene glycol / NaCl

precipitation method of Dunn and Blattner (1987). Sequencing was performed on a

MegaBace 1000 (Amersham Biosciences) machine using the DYEnamicTM ET

termination cycle sequencing premix kit (Amersham Biosciences) and following the

manufacturer’s protocol. The sequences were deposited with Genbank (Accession nos.

given in Table 1). Sequence alignments were conducted using the ClustalX 1.81 program

(Thompson et al., 1994, 2001) and manually refined. Regions of ambiguous alignment

were excluded from the analyses.

Model Selection and Phylogenetic Analyses

Models for maximum likelihood (ML), neighbor-joining (NJ), and Bayesian (BA)

analyses were selected based in two approaches, the Akaike Information Criterion (AIC) of

the 56 models implemented in Modeltest (Posada and Crandall, 1998) and the evaluation

CAPÍTULO III

of the same 56 models in DT–ModSel (Minin et al., 2003). The latter uses a Bayesian

Information Criterion (BIC) that combines branch–length estimates, model fit, and a

penalty for overfitting in a statistically rigorous way. Since the models chosen by BIC were

practically identical to those selected by AIC, we decided to use the AIC approach.

All analyses were performed for (1) each region separately; (2) the cpDNA data;

(3) the mtDNA data; and (4) the combined seven regions data. The combined set for all

loci includes 6,382 nucleotides with 75 Passiflora species representing all subgenera

investigated and eight outgroups (Adenia, Barteria, Deidamia, Dilkea, Mitostemma,

Paropsia, Malesherbia, and Turnera). In the combined results, 2.87% of the cells were

coded as missing information.

Maximum parsimony – Equally weighted parsimony (maximum parsimony or

MP) analyses were performed by heuristic search with TBR branch swapping, the

MULPARS option, and 10 random–addition replicates. Gaps were treated as missing data.

Since an excessive number of most parsimonious trees did not allow searches to be

completed for the each region separately under the described search parameters, a heuristic

search of 1,000 random addition replicates was conducted, with 100 trees saved per

replicate. Bootstrap statistical support (Felsenstein, 1985) was carried out with 1,000

replications of heuristic search and simple taxon addition, with the ALL TREES SAVED

option. The g1 statistic (Hillis, 1991) of skewed tree-length distribution was calculated

from 10,000 random trees, to measure the phylogenetic information content of the seven

DNA regions independently and for the combined data.

Maximum likelihood – In each ML analysis we used the model of sequence

evolution as suggested by Modeltest. The ML tree estimation used heuristic searches with

neighbor-joining starting trees, and TBR branch swapping was conducted with PAUP*

CAPÍTULO III

(Swofford, 1998). We performed 100 replicates of nonparametric bootstrap (Felsenstein,

1985) using the FASTSTEP option, to obtain the confidence of the ML topology. The ML

analysis was also performed in Treefinder (Jobb et al., 2004) and PHYML (Guindon and

Gascuel, 2003), with appropriate models of sequence evolution. The trees obtained by

these two last programs had their likelihood scores recalculated in PAUP* and were then

employed to evaluate the performance in each of the three programs. For this procedure

Shimodaira and Hasegawa’s (1999; SH) test, as well as consensus trees among

phylogenies obtained from different programs were used.

Distances – For the distance analyses, trees were constructed in PAUP* using the

neighbor-joining method (NJ; Saitou and Nei, 1987) with models selected by Modeltest,

proportional (p), and logDet (Steel, 1994; Lockhart et al., 1994) distances. LogDet or

paralinear distances were calculated to test the possible influence of nucleotide

composition differences in the phylogeny (Nei and Kumar, 2000). Reliability of the trees

was tested using 10,000 bootstrap replications (Hedges, 1992).

Bayesian analyses – Bayesian phylogeny estimation (BA) was performed in

MrBayes v3.0b4 (Ronquist and Huelsenbeck, 2003). Each gene was assigned to its own

model of sequence evolution. One cold and three heated Markov chain Monte Carlo

(mcmc) chains run for 2x10

generations were used, with trees sampled every 100

generation, using a random tree as a starting point and a temperature parameter value of

0.2. The mcmc runs were repeated three times as a safeguard against spurious results.

Burn–in, or the time for each parameter to reach a stationary state, was determined by

visual inspection when the log-likelihood values achieved an asymptote over a large

number of generations (the first 2,000 trees were discarded; 200,000 generations for each

CAPÍTULO III

analysis). To calculate the posterior probability of each bipartition, a 50% majority–rule

consensus tree was constructed from the remaining trees using PAUP*.

Evolutionary rates – The molecular–clock hypothesis was tested with the LR test

by calculating the log likelihood score of the chosen model with the molecular clock

enforced and comparing it with the log likelihood score without the molecular clock

enforced (Muse and Weir 1992; Baldwin and Sanderson 1998). The number of degrees of

freedom is equivalent to the number of terminals minus two (Sorhannus and Van Bell,

1999). The two-cluster test of Takezaki et al. (1995) was performed using PHYLTEST

(Kumar, 1996) to evaluate the relative rates between clades. Additionally, average

nucleotide diversities and their standard errors within each subgenus were calculated by the

Mega3 program (Kumar et al., 2004) with the p-distance option and 2,000 bootstrap

replications.

RESULTS

Data Set Characteristics

The general characteristics for each data set are summarized in Table 2;

alignments are available at the Systematic Biology website. The number of taxa analyzed

for each region differs because in specific cases we were unable to adequately amplify the

DNA of a given segment. Some points needing consideration are present below.

rbcL – A nine base pair (bp), or three amino acid (aa) deletion was observed in

Paropsia braunii between positions 120 and 130. It is not know if this deletion may affect

the protein’s function, but the possibility that in this species a pseudogene occurs at this

region should be contemplated. Passiflora sanguinolenta’s sequence presented several stop

CAPÍTULO III

codons, also suggesting the presence of a pseudogene, and therefore was excluded from the

analyses. The DNA of P. clathrata, P. palmeri e Paropsia guinensii, were only partially

sequenced, since it was not possible to amplify the whole region. This suggests a large

deletion or mutation in the annealing portion of the reverse primer. The nucleotide

diversity observed in this region is quite high (0.039 ± 0.004; Table 3), and if data from 32

Malpighiales genera (extracted from GenBank) are compared (data not shown) the

diversity obtained is 0.051 ± 0.004.

rps4 – Several insertion/deletion events (indels) were observed in this gene.

Passiflora suberosa, P. coriacea e P. xiikzodz share a 21 bp (or 7 aa) insertion between

positions 27 and 49, while P. coriacea and P. xiikzodz have another 33 bp insertion

between 441 and 475.

trnL intron – Here also indels were found, and some characterize monophyletic

groups. For instance a six bp deletion occurs between nucleotides (nt) 157 and 154 in

Decaloba.

trnL–trnF intergenic spacer – P. coriacea, P. suberosa e P. xiikzodz presented a

15 bp insertion between nt 105 and 121. Almost all species (exception P. clathrata) show a

11 bp insertion between positions 66 and 78 as compared to outgroups. P. actinia, P.

elegans e P. sidaefolia have a five nt duplication between positions 316 and 322.

nad1 b/c intron – Several indels were observed here, but our previous study

(Muschner et al., 2003), showed that the majority is not phylogenetically informative.

Therefore, they were not considered in the final alignment. Large deletions occur between

nt 686 and 1645. A deletion of 768 bp, starting at position 687, is found in P. umbilicata,

while another of 684 bp, beginning at nt 747, is observed in P. mixta, P. trifoliata and P.

tripartita var. mollissima. Additionally, starting at nt 810, there is a 250 bp deletion in P.

CAPÍTULO III

multiflora and other of 805 bp in Barteria fistulosa. A deletion of 450 bp, beginning at nt

1,125, is observed in P. actinia, P. caerulea, P. edmundoi, P.racemosa, P. reflexiflora, P.

sidaefolia, P. sprucei e P. tenuifila. The evolutionary meaning of these large deletions will

be discussed by V. C. Muschner et al. in a forthcoming paper. Finally, an 11 bp deletion

between nt 161 and 173 is found in all species of the Decaloba subgenus, but is also

present in Deidamia sp. and Turnera subulata.

nad5 d/e intron – Indels can occur, but are less frequent than those found in the

other noncoding regions. Paropsia braunii has a large 173 bp insertion starting at nt 441

and another of 22 bp beginning at nt 610; species of the subgenus Passiflora share a five

bp insertion between positions 316 e 322.

26S – This region practically does not show indels, that when present involve one

nucleotide only.

Phylogenetic Analyses

The topologies of all trees, independently of the types of analysis employed, were

very similar; therefore only selected examples of those obtained by Bayesian or maximum

likelihood methods will be presented. The models selected by Modeltest are listed in Table

The g1 statistics of the distribution of 10,000 random trees indicates that all seven

combinations of data sets are significantly structured, suggesting that the DNA sequence

variation is not random with respect to phylogeny.

Generally, all trees (Figures 1 – 5) showed three monophyletic groups,

corresponding to the Astrophea, Decaloba, and Passiflora subgenera, plus a small set with

P. cirrhiflora and P. ovalis. P. tryphostemmatoides is consistently positioned at the base of

CAPÍTULO III

the Astrophea subgenus. In what follows we will consider first the cpDNA, mtDNA and

nrDNA separately, and afterwards the combined analyses.

Chloroplast DNA – Decaloba, Astrophea, and Passiflora clusters are highly

supported by all methods used (MP, NJ, ML and BA; for the latter see Figure 1), bootstrap

(BS) and posterior probabilities (PP) being higher than 80 in almost all isolated analyses.

The exceptions were rbcL and the trnL–trnF intergenic spacer, in which Astrophea is not

supported. However, when the four plastid regions are considered together the statistical

support for Decaloba are equal to 100 in all analyses, vary between 99 and 100 for

Passiflora, and between 88 (ML and BA), 74 (MP) and 60 (NJ) for Astrophea. The coding

regions (rbcL and rps4) show a better resolution within each of the three groups when

compared to the noncoding segments (trnL intron and trnL–trnF intergenic spacer). In the

combined analysis (Figure 1) species having the same main pollinator generally group

together.

Mitochondrial DNA – In the isolated analyses only the Decaloba monophyly is

supported. But when the two regions are considered jointly Astrophea is also supported in

the ML and BA (Figure 2) approaches. The interspecific relationships are also better

supported when the two regions are considered together.

Nuclear DNA – If the 26S nrDNA gene is considered in isolation, again only the

Decaloba and Astrophea clusters are supported (Figure 3).

Combined data – Tests using combined data from different systems showed

limitations in the detection of incongruences (Reeves et al., 2001; Yoder et al., 2001);

therefore we decided, as Wiens (1998), Reeves et al. (2001) and van den Berg et al. (2005),

to closely examine especially groups supported in the individual analyses. But note that in

CAPÍTULO III

this case (Figures 4 and 5) Decaloba’s monophyly is supported by PP and BS values of

100, of 98 to 100 for Passiflora, and of 94 to 100 for Astrophea.

Evolutionary Rates

A difference that can be visually observed especially in Figures 4 and 5 is the

larger branches of the Decaloba clade when compared to the others. A Z statistic (p <

0.005) calculated with the PHYLTEST package rejects a constant rate of evolution of the

subgenus Decaloba in relation to the others for all DNA regions examined (except the

trnL–trnF intergenic spacer). This test was made with an equal number of species for each

subgenus, since it showed to be very conservative when all species were compared. The

result of the LR test of the molecular–clock hypothesis for

the combined data set is –2 Ln =

508.96, df = 81, P < 0.001. The nucleotide diversity was always higher in Decaloba, as

compared to the other main subgenera (Table 3).

Comparison Among Phylogenetic Methods

The methods employed (Maximum Parsimony, Neighbor–Joining, Maximum

Likelihood, Bayesian), in a general way generated very similar topologies, although a few

within-cluster relationships changed (data not shown). The SH test among the ML trees

generated by PAUP*, PHYML, and Treefinder showed that they do not differ

significantly.

Taxonomic Implications

Decaloba relationships – This subgenus according to Feuillet and MacDougal

(2003), groups Killips’ (1938) Adopogyne, Astephia, Decaloba, Murucuja,

CAPÍTULO III

Pseudomurucuja and Psilanthus subgenera. Our results highly support this view (Figures 1

and 5). For instance, P. penduliflora and P. tacsonioides, as well P. tulae and P. murucuja,

placed in different subgenera by Killip (1938), (but not by Feuillet and MacDougal, 2003),

generally cluster together in the majority of our trees. P. coriacea, P. suberosa, and P.

xiikzodz groups together in all trees. Morphologically they are peculiar by the absence of

petals and Feuillet and MacDougal (2003) placed them in supersection Cieca . P.

sanguinolenta, inserted in the Decaloba supersection by Feuillet and MacDougal (2003),

was first classified in the Psilanthus subgenus by Killip (1938). This last author mentioned

that due to its leaf shape, integument, and bract absence this species should show affinity

with P. rubra. Since we could only study the trnL-trnF intergenic spacer of P. rubra, its

relationship of P. sanguinolenta and P. capsularis is supported by a PP = 71 (data not

shown). The relationships of P. sanguinolenta and P. capsularis is supported in all trees

with PP or BS = 100, then our results support their position in section Xerogona. P.

misera, P. pohlii, P. organensis and P. tricuspis generally group together also. According

to Killip (1938) P. misera and P. tricuspis are related due to their relatively wide inner

corona. P. organensis leaves can also be easily confounded with those of P. pohlii and P.

tricuspis. Feuillet and MacDougal (2003) cluster these species together in section

Decaloba. Our results confirm this classification this and relate them with other Decaloba

species. Another interesting cluster is that observed in the cpDNA tree (Figure 1) between

P. helleri and P. talamancensis which, according to Ulmer and MacDougal (2004) show

almost identical leaves. P. lancetillensis and P. microstipula occur in a basal position in the

Decaloba clade, with high support in all analyses except that of the 26S gene (Figure 3).

According to MacDougal and Hansen (2003) they are intimately related; Feuillet and

MacDougal (2003) classified them in the Pterosperma section of the Decaloba subgenus,

CAPÍTULO III

but Killip (1938) and Feuillet and MacDougal (1999) placed them in the Deidamioides

subgenus. Our results confirm their relationships with Decaloba, but in a divergent

position from the remaining of this subgenus species (Figures 4 and 5).

Astrophea relationships – This subgenus includes species identified with the

same label by Killip (1938) and Feuillet and MacDougal (2003) and the clade is highly

supported. According to Feuillet and MacDougal (2003) the Astrophea subgenus can be

divided into two supersections: Astrophea and Pseudoastrophea. Our analyses, in a general

way, confirm the existence of these groups, but with some exceptions. A larger number of

species of this subgenus should be studied to verify this question.

Passiflora relationships – This group is composed by Killip’s (1938)

Calopathanthus, Distephana, Dysosmia, Dysosmioides, Granadillastrum, Passiflora,

Tacsonia and Tacsonioides subgenera, as well by Escobar’s (1989) Manicata. This

composition is the same proposed by Feuillet and MacDougal (2003). There are clusters

with high statistical support within this subgenus, such as P. actinia, P. elegans and P.

sidaefolia in the combined analyses and cpDNA tree (Figures 1, 4 and 5). Actually,

Muschner et al. (2003) and Lorenz–Lemke et al. (2005) verified that they are evolutionary

closely related. Another well supported cluster refers to P. clathrata, P. foetida and P.

palmeri of Killip’s (1938) Dysosmia subgenus, placed in the subgenus Passiflora,

supersection Stipulata and section Dysosmia by Feuillet and MacDougal (2003). P.

campanulata, P. setulosa and P. villosa classified in the Dysosmioides subgenus by Killip

(1938) cluster with high values of statistical support. P. speciosa and P. vitifolia were

classified in subgenus Distephana by Killip (1938). Feuillet and MacDougal (2003) placed

them in supersection Distephana of Passiflora and we confirmed this classification. They

are both pollinated by hummingbirds. A seven-species cluster with a high bootstrap value

CAPÍTULO III

formed by P. antioquiensis, P. trisecta, P. manicata, P. mathewsii, P. tripartita var.

mollissima, P. mixta and P. trifoliata was observed. These species were variously

classified in different groups by morphological analysis (Killip, 1938; Escobar, 1989), but

were included in the supersection Tacsonioides by Feuillet and MacDougal (2003).

Undetermined relationships – P. cirrhiflora, classified in the subgenus

Polyanthea by Killip (1938), and Tetrastylis (Passiflora) ovalis, included by Feuillet and

MacDougal (2003) in the Deidamioides subgenus, generally clustered together. This

grouping however shows different positions in the trees of the present study, with

undetermined relationships to Decaloba and/or Passiflora. Besides, the statistical support

of this relationship is high only in the Bayesian analyses of the combined 7 regions (Figure

4) or of the cpDNA tree (Figure 1).

P. tryphostemmatoides, classified in Killip’s (1938) Tryphostemmatoides,

subgenus, but included in the Deidamioides subgenus by Feuillet and MacDougal (2003),

occurs closer to the Astrophea base in all but one of our trees, differing markedly from

species of this subgenus in its mtDNA (Figure 2). We suggest that this species could be

considered a separate subgenus.

Outgroups – In the phylogenies involving the seven DNA regions (Figures 4 and

5) Passiflora can be considered as a monophyletic group. Dilkea johannesii and

Mitostemma brevifilis generally occur together at the base of the Passiflora genus cluster,

with the exception of the 26S nrDNA tree (Figure 3). According to Killip (1938), species

of these genera are morphologically very similar, especially in their woody habit.

Deidamia and Adenia always cluster with high statistical support. Curiously, Adenia has

dioic flowers, while Deidamia (and the majority of the Passifloraceae) is hermaphroditic.

Rooting of the trees was made with Malesherbia and Turnera, since despite being included

CAPÍTULO III

in the Passifloraceae they are phylogenetically more distant (Wikström et al., 2001). But in

all trees the outgroup genera that cluster together are Barteria and Malesherbia. The first is

classified in Old World’s (Africa and Madagascar) Paropsieae, while Malesherbia occurs

in the Andes and coastal deserts of Chile, Peru and Argentina (Gengler–Nowak, 2003).

Turnera, distributed in subtropical and tropical America (Neffa and Fernández, 2000) and

Paropsia, which occurs in the Old World together with Barteria and Malesherbia, are in a

more basal position as compared to Adenia and Deidamia.

DISCUSSION

Taxonomic Implications

The three main clades found in this study had already been detected by Muschner

et al. (2003) in a first analysis with a smaller number of taxa and molecular markers.

Yockteng and Nadot (2004) also identified the existence of these clades, but proposed five

other subgenera.

The Decaloba subgenus was highly supported in the present study, despite the

fact that only a fraction (14%) of the 214 species classified in it had been studied. Six of

the eight supersections proposed by Feuillet and MacDougal (2003) were studied and the

majority was confirmed by us. For example, P. coriacea, P. suberosa e P. xiikzodz,

classified by them in the Cieca supersection, clustered together in our trees. However, P.

capsularis and P. sanguinolenta of the Xerogona section were inside the Decaloba section

in the combined analyses. Species classified by Feuillet and MacDougal (2003) as the

Pterosperma supersection were always observed in a basal position in relation to Decaloba

in the present study. According to Ulmer and MacDougal (2004) there are two basal

CAPÍTULO III

supersections in this subgenus, Pterosperma and Hanhniopathanthus, but the latter was not

represented in our studies. These authors also maintained that Pterosperma would be most

primitive, since it possesses flowers borne off the tendrils.

Another well-supported subgenus is Astrophea, with 57 species according to

Feuillet and MacDougal (2003), and represented here by 12 (21%) species. In a general

way our results confirmed the two supersections proposed by these authors, but a cluster

well supported in almost all of our trees, P. candida plus P. citrifolia, were classified in

different supersections (Pseudoastrophea and Boryastrophea, respectively) by them. In the

two phylogenies including all data (Figures 4 and 5) Astrophea is basal in relation to

Passiflora, with high statistical support. These species, in contrast to other Passiflora that

are herbaceous, lianas or small shrubs, consist of trees, shrubs or woody climbing plants.

Besides, the genus characteristic tendrils are reduced to spines or aculei and the leaves can

reach 95 cm of length, as in P. macrophylla (Ulmer and MacDougal, 2004). As asserted by

Benson et al. (1975), this position suggests that the arborescent habit may be ancestral in

the genus. This is confirmed by the Dilkea – Mitostemma group’s position, basal to

Astrophea, since according to Killip (1938) they could be wood vines, subscandent shrubs,

or small trees, as the Old World genera. Kim et al (2004) maintained that the woody habit

is the ancestral state in several major eudicotyledon clades. Besides Dilkea also does not

posses tentrils.

The subgenus Passiflora is composed by 236 species, of which 24% are here

represented. This subgenus is also highly supported in our analyses, confirming Feuillet

and MacDougal’s (2003) classification. However, although in some cases we could discern

agreement with supersections, sections and/or series of these authors, discrepancies also

occur. For instance, P. mendoncaei, P. reflexiflora and P. luetzelburgii, placed in the

CAPÍTULO III

Tacsonioides subgenus by Killip (1938) and classified in the Stipulata supersection,

Tacsonioides section by Feuillet and MacDougal (2003), do not cluster in our phylogenies.

In our analyses the Deidamioides subgenus proposed by Feuillet and MacDougal

(2003) is polyphyletic. Similar results were obtained by Krosnick and Freudenstein (2005).

We suggest that of the three species studied here (P. cirrhiflora, P. ovalis e P.

tryphostemmatoides), only two should be classified as Deidamioides, since P.

tryphostemmatoides does not cluster with them and is placed in the trees near Astrophea,

although with clear differences. A further suggestion, therefore, is that P.

tryphostemmatoides be mantained in the Tryphostemmatoides subgenus of Killip (1938).

Our proposal is that the genus Passiflora should be classified in five subgenera:

Astrophea, Decaloba, Passiflora, Deidamioides and Tryphostemmatoides.

Comparison Among the Three Main Clades

There are significant differences among the three main clades. To begin with, the

species of subgenus Decaloba present significantly smaller flowers than the Passiflora or

Astrophea subgenera (P < 0.05), as initially reported by Muschner et al. (2003). In

addition, Souza et al. (2004) investigated the variation in genome size in eight Passiflora

taxa, seven of the Passiflora and one of the Decaloba subgenera. P. suberosa, the only

Decaloba representative, was the species with the lowest 2C DNA content.

V. C. Muschner et al. (unpublished) analyzed four Passiflora interspecific

hybrids, three from the Passiflora and one from the Decaloba subgenera, in relation to the

sequences of four DNA regions (rps4 gene, trnL intron, trnL–trnF and psbA–trnH

intergenic spacers) and observed that the first group presents exclusively paternal while the

CAPÍTULO III

second shows maternal cpDNA inheritance. On the other hand, the mtDNA is inherited

exclusively from the maternal parent in all these hybrids.

Studies on the reproductive systems have been generally performed in P. edulis,

of the Passiflora subgenus (Rêgo et al., 1999, 2000; Suassuna et al., 2003), no reports

existing in Decaloba or Astrophea. The subgenus Passiflora presents autocompatible and

autoincompatible species (Vasconcelos, 1991; Lindberg and Olesen, 2001; Rêgo et al.,

2000; Suassuna et al., 2003). According to Lewis (1979), Passiflora species with large

flowers are autoincompatible, while those with small flowers are autocompatible. The

same is true in Amsinckia (Boraginaceae; Barret, 2002).

As reported by Snow and MacDougal (1993) and De Melo et al. (2001), the

Decaloba subgenus presents x = 6, while the Passiflora subgenus shows x = 9

chromosomes [but P. foetida, of the latter subgenus and studied by De Melo et al. (2001)

has x = 10]. The Astrophea subgenus would have x = 12 like Adenia, strengthening its

basal position within the genus. But De Melo and Guerra (2003) suggested that x = 6

would be the ancestral genome for the genus, while groups with x = 9, x = 10 and x = 12

would have a tetraploid origin with descending disploidy and inactivation of some

redundant gene sites, especially those of 5S rDNA.

Plotze et al. (2005) developed a morphometric method for the analysis of

Passiflora leaves. They studied six species of the Passiflora and four of the Decaloba

subgenera and detected two clusters, one composed by four of six Passiflora species and

the other with species of the two subgenera. Therefore, no clear association between this

characteristic and subgenus classification was found, leaf form not being a good indicator

for this type of comparison.

CAPÍTULO III

As previously mentioned Passifloraceae species, including those previously

classified as Turneraceae or Malesherbiaceae, recently referred as the Passifloraceous

group (APG II, 2003) are known to produce, with some few exceptions, cyclopentanoid

cyanohydrin glycosides 1-10. Flacourtiaceae and Achariaceae species also produce these

compounds, but no other plants have this property (Clausen et al., 2002). Jaroszewski et al.

(2002) verified the presence of these substances in P. apetala, P. cuneata, P. indecora, P.

kalbreyeri, P. murucuja, P. perfoliata, P. biflora, P. discophora e P. herbetiana, of the

Decaloba subgenus and in P. x violacea, a hybrid between P. caerulea e P. racemosa, of

the Passiflora subgenus. No such compounds were found in P. aurantia, P. gibertii, P.

ligularis, P. manicata, P. platyloba and P. tripartita, species of the Passiflora subgenus,

nor in P. lindeniana of the Astrophea subgenus. It is therefore possible that the presence of

these substances is restricted to Decaloba, their occurrence in the Passiflora hybrid being

due to the exceptional opening of the respective metabolic route due to the hybridization

process.

The association Heliconius-Passiflora is one of the better studied insect-plant

associations. Heliconiine butterflies only feed, as larvae, Passifloraceae leaves. Benson et

al. (1975) and Brown (1981) reviewed the Heliconiine-Passifloraceae interaction and

concluded that there is specificity between determined groups of Heliconius and different

Passiflora subgenera. Ehrlich and Raven (1964) suggested that these correlations can be

explained by coevolutive processes responsible for the high present diversification of

modern plants and herbivorous insects. Species of the Heliconius numata-melpomene

group feed especially Granadilla (of the Passiflora subgenus) plants. According to these

authors both taxa are considered primitive in their genera, possessing non-specialized

morphology and behavior. On the other hand, members of the Heliconius erato-charitonia

CAPÍTULO III

species group feed plants of the Plectostemma (Decaloba) subgenus, while larvae of the

Heliconius sara-sapho group feed plants both of the Plectostemma (Decaloba) and

Astrophea subgenera. These other associations could be derived phenomena. Smiley

(1985) confirmed these relationships and concluded that the Decaloba subgenus evolved

without a chemical barrier against herbivory, while many species of the Passiflora

subgenus are protected from a large number of Heliconius larvae. Besides, species of this

subgenus developed morphologic structures which prevent or disfavor butterfly

oviposition, such as those which resemble egg layers, or present high pilosity (Ulmer and

MacDougal, 2004).

Maternal, Paternal, and Biparental Phylogenies

The mode of inheritance of a characteristic has obvious influence in its evolution

(Harris and Ingram, 1991). In addition, traits inherited by parents of only one side can

furnish precious details about sex-specific conditions important for the evolution of the

organism as a whole, such as interpopulation gene flow (Collevatti et al., 2001, 2003;

Liepelt et al., 2002). Ideally, therefore, evolutionary investigations should include, as was

done in the present study, characteristics inherited just by the maternal or paternal line, as

well as others with biparental inheritance (Shore and Triassi, 1998).

In plants and algae mitochondria and chloroplasts can be inherited through distinct

sexual lineages. Generally, in hermaphrodite species both organelles are transmitted

together by the gamete of the same sex, female in the majority of the Angiosperms

(Dumolin-Lapègue et al., 1998, Moreira et al., 2002) and male in some Gymnosperms

(Chesnoy, 1987; Hagemann, 1992). Exceptions, however, were observed; in Pinaceae

(Neale and Sederoff, 1989) and Actinidia (Actinidiaceae) chloroplasts are paternally and

CAPÍTULO III

mitochondria maternally inherited (Chat et al., 1999; Burban and Petit, 2003); while the

opposite was observed in Musa acuminata (Musaceae) and Cucumis (Cucurbitaceae)

(Fauré et al., 1994; Havey et al., 1998).

Similarities and dissimilarities were obtained in our phylogenetic cpDNA,

mtDNA, and nrDNA trees. All of them, however, point for Decaloba subgenus

monophyly. On the other hand, when the placement of the Passiflora and Astrophea

subgenera is considered, intergenomic differences occur. Mention was already made that

the pattern of organelle inheritance differs among these subgenera; if some of the species

considered had a hybrid origin, discrepancies could occur in relation to the others. Another

question refers to different rates of nucleotide substitution; the differences observed

between Decaloba and the other subgenera in this regard were already mentioned. But the

intergenomic differences in such rates should also consider (Riesenberg, et al., 1996; Chat

et al., 2004). In our case it should be pointed out that the four cpDNA regions presented

five times more parsimoniously informative sites than the mtDNA and/or nrDNA markers.

Pollinization Agents

The high floral diversity found in the Passiflora genus is certainly intimately

related to the different pollinization forms that exist in this genus. The corona of filaments,

between the perianth and stamens, is highly variable, showing different colors, forms,

smells, and filament disposition. According to Endress (1994), the series of most external

filaments are involved in pollinator attraction, while the two inner ones furnish mechanical

protection to the nectar chamber.

Passiflora ancestors were probably especially pollinated by bees. Hummingbird

pollinization probably occurred independently more than once in the genus and in all

CAPÍTULO III

subgenera that were also pollinated by wasps, butterflies, and bats (MacDougal, 1994).

Examples are seen in the phylogeny of Figure 1.

Pollination by small and large bees are found in Passiflora taxa with flattish to cup-

shaped hypanthia, and with a range of erect, lateral to pendulous flowers. This flower form

is generally found in the Decaloba subgenus, but also in several species of the Passiflora

and Astrophea subgenera (MacDougal, 1994).

According to MacDougal (1994), hummingbird pollination is common in several

Passiflora groups, and was documented in the Granadillastrum, Tacsonia, Dysosmia

(presently Passiflora), Decaloba, Murucuja e Pseudomurucuja (presently Decaloba)

subgenera. In our material widely separated species like P. speciosa and P. vitifolia, in the

Passiflora, P. sanguinolenta, in the Decaloba, and P. amoena, in the Astrophea subgenera

are all pollinated by hummingbirds, suggesting independent origins. The latter observation

agrees with Ulmer and MacDougal’s (2004) assertion that hummingbirds pollinate species

of the Astrophea Boryastrophea section (to which P. amoena belongs).

Passiflora species typically pollinated by bats (P. mucronata and P. galbana, for

instance) seem to attract the animals by their scent (Sazima and Sazima, 1978). This type

of pollination seems also to have independently developed many times in Passiflora, as

suggested by its presence in evolutionary distinct species (P. penduliflora in the Decaloba,

P. galbana, P. trisecta and P. trifoliata in the Passiflora, and P. ovalis in the Deidamioides

subgenera).

CONCLUSIONS

The present study illustrates the utility of molecular analyses for a series of

comparisons of both evolutionary and taxonomic importance. The observed pattern of

CAPÍTULO III

genetic variation shows clear intergenomic differences, one especially curious finding

being the large deletions found in the nad1 b/c intron. Also of note is the higher rate of

change observed in the Decaloba clade, which is reflected in higher nucleotide diversity

indices. The three main subgenera identified through the molecular approach show

distinctiveness in flower and genome size, cpDNA inheritance, chromosome numbers,

secondary compounds content, and diverse degree of relationships with pollinators and

predators. In addition to these three subgenera we are proposing the existence of two

others. Within these entities clear associations between patterns of molecular phylogenetic

relationships and morphological traits could be found. We maintain that global evaluations

such as this one, involving different genomes, good phylogenetic markers, and careful

taxonomic sampling, are essential for the unraveling of the complex evolutionary histories

of the organic world.

ACKNOWLEDGMENTS

We thank Mark Chase, Maurizio Vecchia, Marcelo S. Guerra-Filho, Natoniel

Franklin de Melo, Cláudio Mondin, Teonildes S. Nunes, Marcelo C. Dornelas, Cássio van

den Berg, Roxana Yockteng, Sophie Nadot, Karla Gengler, Fernando Campos Neto, Luis

Carlos Bernacci, Alessandra Selbach, Alba Lins and Shawn Krosnick for specimen

donations. This research was financially supported by Programa de Apoio a Núcleos de

Excelência (PRONEX), Conselho Nacional de Desenvolvimento Científico e Tecnológico

(CNPq), Fundação de Amparo à Pesquisa do Estado do Rio Grande do Sul (FAPERGS),

and Pró-Reitoria de Pesquisa da Universidade Federal do Rio Grande do Sul (PROPESQ-

UFRGS).

CAPÍTULO III

REFERENCES

APG II. 2003. An update of the Angiosperm Phylogeny Group classification for the orders

and families of flowering plants: APG II. Bot. J. Linn. Soc. 141: 399–436.

Bakker, F. T., A. Culham, C. E. Pankhurst, and M. Gibby. 2000. Mitochondrial and

chloroplast DNA–based phylogeny of Pelargonium Geraniaceae). Am. J. Bot. 87: 727–

734.

Baldwin, B. G., and M. J. Sanderson. 1998. Age and rate of diversification of the Hawaiian

silversword alliance (Compositae).Proc. Natl. Acad. Sci. USA 95:9402–9406.

Barret, S. C. H. 2002. The evolution of plant sexual diversity. Nat. Rev. Genet. 3:274–284.

Benson, W. W., K. S. Brown, and L. E. Gilbert. 1975. Coevolution of plants and

herbivores: Passion flower butterflies. Evolution 29: 659–680.

Bowe, L. M., G. Coat, and C. W. Depamphilis. 2000. Phylogeny of seed plants based on

all three genomic compartments: Extant Gymnosperms are monophyletic and Gnetales’

closest relatives are Conifers. Proc. Natl. Acad. Sci. USA 97: 4092–4097.

Brown, K. S. 1981. The biology of Heliconius and related genera. Ann. Rev. Entomol. 26:

427–456.

Burban, C., and R. J. Petit. 2003. Phylogeography of maritime pine inferred with organelle

markers having contrasted inheritance. Mol. Ecol. 12: 1487–1495.

Cervi, A. C. 1997. Passifloraceae do Brasil. Estudo do gênero Passiflora L. subgênero

Passiflora. Fontqueria 45: 1–92.

Chat J., L. Chalak, and R. J. Petit. 1999. Strict paternal inheritance of chloroplast DNA and

maternal inheritance of mitochondrial DNA in intraspecific crosses of kiwifruit. Theor.

Appl. Genet. 99: 314–322.

CAPÍTULO III

Chat, J., B, Jáuregui, R. J. Petit, and S. Nadot. 2004. Reticulate evolution in kiwifruit

(Actinidia, Actinidaceae) identified by comparing their maternal and paternal

phylogenies. Am. J. Bot. 91: 736–747.

Chaw, S.–M., C. L. Parkinson, Y. Cheng, T. M. Vincent, and J. D. Palmer.2000. Seed

plant phylogeny inferred from all three plant genomes: Monophyly of extant

gymnosperms and origin of Gnetales from conifers. Proc. Natl. Acad. Sci. USA 97:

4086–4091.

Chesnoy L. 1987. La reproduction sexuée des Gymnosperms. Bull. Soc. Bot. France 134:

63–85.

Cho, Y., J. P. Mower, Y.–L. Qiu, and J. D. Palmer. 2004. Mitochondrial rates are

extraordinarily elevated and variable in a genus of flowering plants. Proc. Natl. Acad.

Sci. USA 101: 17741–17746.

Clausen, V., K. Frydenvang, R. Koopman, L. B. Jørgensen, D. K. Abbiw, P. Ekpe, and J.

W. Jaroszewski. 2002. Plant analysis by butterflies: Occurrence of

cyclopentenylglycines in Passifloraceae, Flacourtiaceae, and Turneraceae and

discovery of the novel nonproteinogenic amino acid 2–(3’–cyclopentenyl)glycine in

Rinorea. J. Nat. Prod. 65: 542–547.

Collevatti R. G., D. Grattapaglia, and J. D. Hay. 2001. High resolution microsatellite based

analysis of the mating system allows the detection of significant biparental inbreeding

in Caryocar brasiliense, and endangered tropical tree species. Heredity 86: 60–67.

Collevatti R. G., D. Grattapaglia, and J. D. Hay. 2003. Evidence for multiple maternal

lineages of Caryocar brasiliense populations in the Brazilian Cerrado based on the

analysis of chloroplast DNA sequences and microsatellite haplotype variation. Mol.

Ecol. 12: 105–115.

CAPÍTULO III

De Melo, N. F., A. C. Cervi, and M. Guerra. 2001. Karyology and cytotaxonomy of the

genus Passiflora L. (Passifloraceae). Plant Syst. Evol. 226: 69–84.

De Melo, N. F., and M. Guerra. 2003. Variability of the 5S and 45S rDNA sites in

Passiflora L. species with distinct base chromosome numbers. Ann. Bot. 92: 309–316.

Duminil J., M. H. Pemonge, and R. J. Petit. 2002. A set of 35 consensus primer pairs

amplifying genes and introns of plant mitochondrial DNA. Mol. Ecol. Notes 2: 428–

430.

Dumolin–Lapègue S, Pemonge MH, and R. J. Petit. 1998. Association between chloroplast

mitochondrial lineages in oaks. Mol. Biol. Evol. 15: 1321–1331.

Dunn I. S., F. R. Blattner. 1987. Charons 36 to 40: multi enzyme, high capacity,

recombination deficient replacement vectors with polylinkers and polystuffers. Nucl.

Ac. Res. 15: 2677–2698.

Ehrlich, P. R., and P. H. Raven. 1964. Butterflies and plants: A study in coevolution.

Evolution 18: 586–608.

Endress, P. K. 1994. Diversity and evolutionary biology of tropical flowers. Cambridge

University Press, Cambridge.

Escobar, L. K. 1989. A new subgenus and five new species in Passiflora (Passifloraceae)

from South America. Ann. Miss. Bot. Gard. 76: 877–855.

Farris, J. S., M. Källersjö, A. G. Kluge, and C. Bult. 1994. Testing significance of

congruence. Cladistics 10: 315–319.

Fauré S., J.–L. Noyer, F. Carreel, J.–P. Horry, F. Bakry, and C. Lanaud. 1994. Maternal

inheritance of chloroplast genome and paternal inheritance of mitochondrial genome in

bananas (Musa acuminata). Curr. Genet. 25: 265–269.

CAPÍTULO III

Felsenstein, J. 1985. Confidence limits on phylogenies: An approach using the bootstrap.

Evolution 39: 783–791.

Feuillet, C. P. and J. M. MacDougal. 1999. Abstract 4295. XVI International Botanical

Congress, Saint Louis, Missouri, USA. http://www.biologie.uni–hamburg.de/b–

online/ibc99/ibc/abstracts/listen/abstracts/4295.html.

Feuillet, C. P., and J. M. MacDougal. 2003. A new infrageneric classification of

Passiflora. Passiflora 13: 34–38.

Freudenstein, J. V., and M. W. Chase. 2001. Analysis of mitochondrial nad1 b–c intron

sequences in Orchidaceae: utility and coding of length change characters. Syst. Bot. 26:

643–657.

Gengler–Nowak, K. M. 2003. Molecular phylogeny and taxonomy of Malesherbiaceae.

Syst. Bot. 28:333–344.

Goffinet B., C. J. Cox, A. J. Shaw, and T. A. J. Hedderson. 2001. The bryophyta (mosses):

Systematic and evolutionary inferences from an rps4 gene (cpDNA) phylogeny. Ann.

Bot. 87: 191–208.

Guindon S., and O. Gascuel. 2003. A simple, fast, and accurate algorithm to estimate large

phylogenies by maximum likelihood. Syst. Biol. 52: 696–704.

Hagemann R. 1992. Plastid genetics in higher plants. Pages 65–96. in Cell organelles (R.G.

Herrmann, ed.). Springer–Verlag, Berlin.

Harris S. A., and R. Ingram. 1991. Chloroplast DNA and biosystematics: The effects of

intraspecific diversity and plastid transmission. Taxon 40: 393–412.

Havey M. J., J. D. McCreight, B. Rhodes, and G. Taurick. 1998. Differential transmission

of the Cucumis organellar genomes. Theor. Appl. Genet. 97: 122–128.

CAPÍTULO III

Hedges, S. B. 1992. The number of replications needed for accurate estimation of the

bootstrap P value in phylogenetic studies. Mol. Biol. Evol. 9: 366–369.

Hillis, D. M. 1991. Discriminating between phylogenetic signal and random noise in DNA

sequences. In (M. M. Myamoto, and J. Craft, eds.) Phylogenetic analysis of DNA

sequences, 278 – 294, Oxford University Press, New York.

Jaroszewski J. W., E. S. Olafsdottir, P. Wellendorph, J. Christensen, H. Franzyk, B.

Somanadhan, B. A. Budnik, L. B. Jørgensen, and V. Clausen. 2002. Cyanohydrin

glycosides of Passiflora: Distribution pattern, a saturated cyclopentane derivative from

P. guatemalensis, and formation of pseudocyanogenic alpha–hydroxyamides as

isolation artifacts. Phytochemistry 59: 501–511.

Jobb, G., A. von Haeseler, and K. Strimmer. 2004. TREEFINDER: a powerful graphical

analysis environment for molecular phylogenetics. BMC Evol. Biol. 4: 18.

Killip, E. P. 1938. The American species of Passifloraceae. Field Museum of Natural

History, Botanical Series 19: 1–613.

Kim, S., D. E. Soltis, P. S. Soltis, M. J. Zanis, and Y. Suh. 2004. Phylogenetic

relationships among early-diverging eudicots based on four genes: were the eudicots

ancestrally woody? Mol. Phylogenet. Evol. 31: 16-30.

Krosnick, S. E., and J. V. Freudenstein. 2005. Monophyly and floral character homology

of Old World Passiflora (Subgenus Decaloba: Supersection Disemma) Syst. Bot. 30:

139–152.

Kumar, S. 1996. PHYLTEST: phylogeny hypothesis testing software. Pennsylvania State

University, University Park, Pennsylvania, USA.

Kumar, S., K. Tamura, and M. Nei. 2004. MEGA3: Integrated software for Molecular

Evolutionary Genetics Analysis and sequence alignment. Brief. Bioinform. 5:150–163.

CAPÍTULO III

Kuzoff R. K., and C. S. Gasser. 2000. Recent progress in reconstructing Angiosperm

phylogeny. Trends Plant Sci. 5: 330–336.

Kuzoff R. K., J. A. Sweere, D. E. Soltis, P. S. Soltis, and E. A. Zimmer. 1998. The

phylogenetic potential of entire 26S rDNA sequences in plants. Mol. Biol. Evol. 15:

251–263.

Lewis, D. 1979. Evolution of incompatibility: The price of a gene bank and long–term

insurance. Pages 53–56. in Sexual incompatibility in plants (D. Lewis, ed.). Edward

Arnold, London.

Liepelt S., R. Bialozyt, and B. Ziegenhagen 2002. Wind–dispersed pollen mediates

postglacial gene flow among refugia. Proc. Natl. Acad. Sci. USA 99: 14590–14594.

Lindberg, A. B., and J. M. Olesen. 2001. The fragility of extreme specialization: Passiflora

mixta and its pollinating hummingbird Ensifera ensifera. J. Trop. Ecol. 17: 323–329.

Lockhart, P. J., M. A. Steel, M. D. Hendy, and D. Penny. 1994. Recovering evolutionary

trees under a more realistic model of sequence evolution. Mol. Bio. Evol. 11: 605–612.

Lorenz–Lemke A. P., V. C. Muschner, S. L. Bonatto, A. C. Cervi, F. M. Salzano, and L. B.

Freitas 2005. Phylogeographic inferences concerning evolution of Brazilian Passiflora

actinia and P. elegans (Passifloraceae) based on ITS (nrDNA) variation. Ann. Bot. 95:

799–806.

MacDougal, J. M. 1994. Revision of Passiflora subgenus Decaloba section

Pseudodysosmia (Passifloraceae). Syst. Bot. Monogr. 41: 1–46.

MacDougal, J. M., and A. K. Hansen. 2003. A new nection of Passiflora, subgenus

Decaloba (Passifloraceae), from Central America, with two new species. Novon 13:

459–466.

CAPÍTULO III

Minin, V. , Z. Abdo , P. Joyce , and J. Sullivan. 2003. Performance–based selection of

likelihood models for phylogeny estimation. Syst. Biol. 52: 674–682.

Moreira C. D., F. G. Gmitter, J. W. Grosser, S. Huang, V. M. Ortega, and C. D. Chase

2002. Inheritance of organelle DNA sequences in Citrus-Poncirus intergeneric cross. J.

Hered. 93: 174–178.

Muschner, V. C., A. P. Lorenz, A. C. Cervi, S. L. Bonatto, T. T. Souza–Chies, F. M.

Salzano, and L. B. Freitas. 2003. A first molecular phylogenetic analysis of Passiflora

(Passifloracae). Am. J. Bot. 90: 1229–1238.

Muse,S. V., and B. S. Weir. 1992. Testing for equality of evolutionary rates. Genetics 132,

269–276.

Neale D. B., and R. R. Sederoff. 1989. Paternal inheritance of chloroplast DNA and

maternal inheritance of mitochondrial DNA in loblolly pine. Theor. Appl.Genet.

77:212–216.

Neffa V. G. S., and A. Fernández. 2000. Chromosome studies in Turnera (Turneraceae).

Genet. Mol. Biol. 23: 925–930.

Nei, M., and S. Kumar. 2000. Molecular evolution and phylogenetics. Oxford University

Press, Oxford, UK.

Pérez–Losada M, Bond–Buckup G, Jara CG, and Crandall KA. 2004. Molecular

systematics and biogeography of the southern South American freshwater "crabs"

Aegla (Decapoda: Anomura: Aeglidae) using multiple heuristic tree search approaches.

Syst. Biol. 53: 767–780.

CAPÍTULO III

Plotze R. D., M. Falvo, J. G. Padua, L. C. Bernacci, M. L. C. Vieira, G. C. X. Oliveira, and

O. M. Bruno. 2005. Leaf shape analysis using the multiscale Minkowski fractal

dimension, a new morphometric method: A study with Passiflora (Passifloraceae).

Can. J. Bot. 83: 287–301.

Posada, D., and K. A. Crandall. 1998. MODELTEST: Testing the model of DNA

substitution. Bioinformatics 14: 817–818.

Qiu, Y.–L., J. Lee, F. Bernasconi–Quadroni, D. E. Soltis, P. S. Soltis, M. Zanis, E. A.

Zimmer, Z. Chen, V. Savolainen, and M. W. Chase. 1999. The earliest Angiosperms:

Evidence from mitochondrial, plastid and nuclear genomes. Nature 402:404–7.

Reeves, G., M. W. Chase, P. Goldblatt, P. Rudall, M. F. Fay, A. V. Cox, B. Lejeune, and

T. Souza–Chies. 2001. Molecular systematics of Iridaceae: evidence from four plastid

regions. Am. J. Bot. 88: 2074–2087.

Rêgo, M. M., C. H. Bruckner, E. A. M. da Silva, F. L. Finger, D. L. de Siqueira, and A. A.

Fernandes. 1999. Self–incompatibility in passion fruit: Evidence of two locus genetic

control. Theor. Appl. Genet. 98: 564–568.

Rêgo, M. M., E. R. Rêgo, C. H. Bruckner, E. A. M. da Silva, F. L. Finger, and K. J. C.

Pereira. 2000. Pollen tube behavior in yellow passion fruit following compatible and

incompatible crosses. Theor. Appl. Genet. 101: 685–689.

Riesenberg, L. H., J. Whitton, and C. R. Linder. 1996. Molecular marker incongruence in

plant hybrid zones and phylogenetic trees. Acta Bot. Neerl. 45: 243–262.

Rokas A., and S. B. Carroll. 2005. More genes or more taxa? The relative contribution of

gene number and taxon number to phylogenetic accuracy. Mol. Biol. Evol. 22: 1337–

1344.

CAPÍTULO III

Rokas A., B. L. Williams, N. King, and S. B. Carrol. 2003. Genome–scale approaches to

resolving incongruence in molecular phylogenies. Nature 425: 798–804.

Ronquist F., and J. P. Huelsenbeck. 2003. MrBayes 3: Bayesian phylogenetic inference

under mixed models. Bioinformatics 19: 1572–1574.

Roy A., N. Frascaria, J. MacKay, and J. Bousquet. 1992. Segregating random amplified

polymorphic DNAs (RAPDs) in Betula alleghaniensis. Theor. Appl. Genet. 85: 173–

180.

Saitou, N., and M. Nei. 1987. The neighbor–joining method: A new method for

reconstructing phylogenetic trees. Mol. Biol. Evol. 9: 945–967.

Savolainen V., M. W. Chase, S. B. Hoot, C.M. Morton, D. E. Soltis, C. Bayer, M. F. Fay,

A. Y. de Bruijn, S. Sullivan, and Y. L. Qiu. 2000. Phylogenetics of flowering plants

based on combined analysis of plastid atpB and rbcL gene sequences. Syst Biol. 49:

306–362.

Sazima, M., and I. Sazima. 1978. Bat pollination of the passion flower, Passiflora

mucronata, in southeastern Brazil. Biotropica 10: 100–109.

Shimodaira, H., and M. Hasegawa. 1999. Multiple comparisons of loglikelihoods with

applications to phylogenetic inference. Mol. Biol. Evol. 16: 1114–1116.

Shore J. S., and M. Triassi. 1998. Paternally biased cpDNA inheritance in Turnera

ulmifolia (Turneraceae). Am. J. Bot. 85: 328–332.

Smiley, J. T. 1985. Are chemical barriers necessary for evolution of the butterfly–plant

associations? Oecologia 65: 580–583.

Snow, N., and J. M. MacDougal. 1993. New chromosome reports in Passiflora

(Passifloraceae). Syst. Bot. 18: 261–273.

CAPÍTULO III

Soltis, D. E., P. S. Soltis, M. E. Mort, M. W. Chase, V. Savolainen, S. B. Hoot, and C. M.

Morton. 1998. Inferring complex phylogenies using parsimony: An empirical approach

using three large DNA data sets for Angiosperms. Syst. Biol. 47: 32–42.

Sorhannus, U., and C. Van Bell. 1999. Testing for equality of molecular evolutionary rates:

A comparison between a relative–rate test and a likelihood ratio test. Mol. Biol. Evol.

16:848–855.

Souza, A. P. M.–F. Jubier, E. Delcher, D. Lancelin, and B. Lejeune. 1991. A trans–splicing

model for the expression of the tripartite nad5 gene in wheat and maize mitochondria.

Plant Cell 3: 1363–1378.

Souza, M. M., G. Palomino, T. N. S. Pereira, M. G. Pereira, and A. P. Viana. 2004. Flow

cytometric analysis of genome size variation in some Passiflora species. Hereditas 141:

31–38.

Souza–Chies, T. T., G. Bittar, S. Nadot, L. Carter, E. Besin, and B. Lejeune. 1997.

Phylogenetic analysis of Iridaceae with parsimony and distance methods using the

plastid gene rps4. Plant Syst. Evol. 204: 109–123.

Steel, M. 1994. Recovering a tree from the Markov leaf colouration it generates under a

Markov model. Appl. Math. Lett. 7: 19–23.

Suassuna, T. M. F., C. H. Bruckner, C. R. de Carvalho, and A. Borém. 2003. Self–

incompatibility in passionfruit: evidence of gametophytic–sporophytic control. Theor.

Appl. Genet. 106: 298–302.

Swofford, D. L. 1998. PAUP*: phylogenetic analysis using parsimony (*and other

methods). Version 4. Sinauer Associates, Sunderland, Massachusetts, USA.

Taberlet, P., L. Gielly, G. Patou, and J. Bouvet. 1991. Universal primers for amplification

of three non–coding regions of chloroplast DNA. Plant Mol. Biol. 17: 1105–1109.

CAPÍTULO III

Takezaki, N., A. Rzhetsky, and M. Nei. 1995. Phylogenetic test of the molecular clock and

linearized tree. Mol. Phylogenet. Evol. 12: 823–833.

Thompson, J. D., D. G. Higgins, and T. J. Gibson. 1994. ClustalW: Improving the

sensivity of progressive multiple sequence weighting, positions–specific gap penalties

and weight matrix choice. Nucl. Ac. Res. 22: 4673–4680.

Thompson, J. D., D. G. Higgins, and T. J. Gibson. 2001. ClustalX. Program available at

fttp://ftp–igbmc.u–strasbg.fr/pub/clustalx/.

Ulmer, T., and J. M. MacDougal. 2004. Passiflora hybrids and cultivars.

http://www.passionflow.co.uk/downloads1.htm.

Ulmer, T., and MacDougal. 2004. Passiflora: Passionflowers of the world. Timber Press,

Portland.

Van den Berg, C., D. H. Goldman, J. V. Freudenstein, A. M. Pridgeon, K. M. Cameron,

and M. W. Chase. 2005. An overview of the phylogenetic relationships within

Epidendroideae inferred from multiple DNA regions and recircumscription of

Epidendreae and Arethuseae (Orchidaceae). Am. J. Bot. 92:613–624.

Vanderplank J. 1996. Passion flowers. 2nd edn. Massachusetts Institute of Technology,

Cambridge.

Varassin, I. G., J. R. Trigo, and M. Sazima. 2001. The role of nectar production, flower

pigments and odour in pollination of four species of Passiflora (Passifloraceae). Bot. J.

Linn. Soc. 136: 139–152.

Vasconcellos, M. A. D. S. 1991. Observaciones sobre incompatibilidad floral y de botones

florales em fase de pré–antesis em el maracuya dulce, Passiflora alata Dryand. Pages

95–97 in Primer simpósio internacional de Passifloras, memórias (F. A. Vallejo, ed.).

Palmira, Colombia.

CAPÍTULO III

Werner, O., and J. Guerra. 2004. Molecular phylogeography of the moss Tortula muralis

Hedw. (Pottiaceae) based on chloroplast rps4 gene sequence data. Plant Biol. 6: 147–

157.

Wiens J. J., 1998. Combining data sets with different phylogenetic histories. Syst. Biol. 47:

568–581.

Wikström, N., V. Savolainen, and M. W. Chase. 2001. Evolution of the Angiosperms:

Calibrating the family tree. Proc. Royal Soc. London 268: 2211–2220.

Yockteng, R., and S. Nadot. 2004. Phylogenetic relationships among Passiflora species

based on the glutamine synthetase nuclear gene expressed in chloroplast (ncpGS). Mol.

Phylogenet. Evol. 31: 379–396.

Yoder, A. D., J. A. Irwin, and B. A. Payseur. 2001. Failure of the ILD to determine data

combinability for slow loris phylogeny. Syst. Biol. 50: 408–424.

Table 1: List of the species studied, their taxonomic classification, source of collection, and GenBank accession numbers for the DNA sequences.

GenBank numbers

Subgenera or tribe

(Killip, 1938)

This study's

proposal

Species

Data of collection

rbcL rps4 trnL trnL-trnF

nad1 b/c nad5 d/e

26S

Astrophea Astrophea P. amoena L. K. Escobar

Italy, Ripalta Cremasca,

Colection (MV)

DQ123300 DQ123407 DQ123017 DQ123486 DQ123214 DQ123128 DQ122935

P. arborea Spreng.

Panama (RY)

DQ123301 DQ123408 DQ123018 DQ123487 DQ123215 DQ123129 DQ122936

P. candida (P. & E.) Mast.

Italy, Ripalta Cremasca,

Colection (MV)

DQ123302 DQ123409 DQ123019

N/A

DQ123216 DQ123130 N/A

P. ceratocarpa Silveira

Brazil, PA (LCB)

DQ123303 DQ123410 DQ123020 DQ123488 DQ123217 DQ123131 DQ122937

P. citrifolia (Juss.) Mast.

French Guiana (MV)

DQ123304 AY212311 DQ123021 AY210958 DQ123218 DQ123132 DQ122938

P. haematostigma Mart. ex Mast.

Guaratuba, PR (ACC)

DQ123305 AY212292 DQ123022 AY032773 DQ123219 DQ123133 DQ122939

P. kawensis Feuillet

French Guiana (RY)

DQ123306 DQ123411 DQ123023 DQ123489 DQ123220 DQ123134 DQ122940

P. lindeniana Tr. & Pl.

Italy, Ripalta Cremasca,

Colection (MV)

DQ123307 DQ123412 DQ123024 DQ123490 DQ123221 DQ123135 DQ122941

P. macrophylla Spruce ex Mast.

Brazil (MV)

DQ123308 AY212313 DQ123025 AY210965 DQ123222 DQ123136 DQ122942

P. mansoi (Mart.) Mast.

Chapadão do Sul, MS

(ACC)

DQ123309 AY212307 DQ123026 AY102401 DQ123223 DQ123137 DQ122943

P. pittieri Mast.

Italy, Ripalta Cremasca,

Colection (MV)

DQ123310 DQ123413 DQ123027 DQ123491 DQ123224 DQ123138 DQ122944

P. rhamnifolia Mast.

Cabo Frio, RJ (TSN)

DQ123299 DQ123406 DQ123016 DQ123485 DQ123213 DQ123127 N/A

Adopogyne Decaloba P. multiflora L.

Dominica (MV)

DQ123297 DQ123404 DQ123014 AY210967 DQ123211 DQ123125 DQ122933

Astephia P. penduliflora Bertero ex DC.

Blois-France-Greenhouse

(RY)

DQ123298 DQ123405 DQ123015 DQ123484 DQ123212 DQ123126 DQ122934

Decaloba P. capsularis L.

Quatro Barras, PR

(ACC)

DQ123312 DQ123415 DQ123029 AY032775 DQ123226 DQ123140 DQ122946

P. coriacea Juss.

Colombia (MV)

DQ123313 DQ123416 DQ123030 AY210959 DQ123227 DQ123141 DQ122947

P. helleri Peyer

Mexico (MV)

DQ123314 DQ123417 DQ123031 AY210962 DQ123228 DQ123142 DQ122948

P. lobbi subsp. ayaucuchoensis

Skrabal & Weigend

Peru (MC)

DQ123315 DQ123419 DQ123032 DQ123493 N/A N/A N/A

P. lobbi subsp. obtusiloba (Mast.)

Skrabal & Weigend

Peru (MC)

DQ123316 DQ123418 DQ123033 DQ123494 DQ123229 DQ123143 N/A

P. misera HBK.

Santa Maria, RS (PASS)

DQ123317 DQ123420 DQ123034 AY032777 DQ123230 DQ123144 DQ122949

Table 1 (Cont.)

GenBank numbers

Subgenera or tribe

(Killip, 1938)

This study's

proposal

Species

Data of collection

rbcL rps4 trnL trnL-trnF

nad1 b/c nad5 d/e

26S

P. morifolia Mast. in Mart.

Brazil, RS (PASS)

DQ123318 AY212314 DQ123035 AY032780 DQ123231 DQ123145 DQ122950

P. organensis Gardn.

Brazil, PR (ACC)

DQ123319 DQ123421 DQ123036 AY032779 DQ123232 DQ123146 DQ122951

P. ornithoura Mast.

Guatemala (MV)

DQ123320 DQ123422 DQ123037 AY210968 DQ123233 DQ123147 DQ122952

P. pohlii Mast. in Mart.

Pirapora, MG (ACC)

DQ123321 DQ123423 DQ123038 AY032778 DQ123234 DQ123148 DQ122953

P. podlechii Skrabal & Weigend

Peru (MC)

N/A DQ123403 DQ123013 DQ123483 DQ123210 N/A N/A

P. punctata L.

Peru (MV)

DQ123322 DQ123424 DQ123039 AY210969 DQ123235 N/A DQ122954

P. rovirosae Killip

Italy, Ripalta Cremasca,

Colection (MV)

N/A DQ123425 DQ123040 N/A N/A N/A N/A

P. rubra L.

Brazil, PE (MG)

N/A N/A N/A AY032776 N/A N/A N/A

P. rufa Feuillet

French Guiana (MV)

DQ123323 AY212315 DQ123041 AY210971 DQ123236 DQ123149 DQ122955

P. sexflora Juss.

Dominican Republic

(MV)

DQ123324 DQ123426 DQ123042 AY210974 DQ123237 DQ123150 DQ122956

P. suberosa L.

Brazil, RS (PASS)

DQ123325 DQ123427 DQ123043 AY032774 DQ123238 DQ123151 DQ122957

P. talamancensis Killip

Costa Rica (MV)

DQ123326 DQ123428 DQ123044 AY210976 DQ123239 DQ123152 DQ122958

P. tricuspis Mast. in Mart.

Brazil, SP (MCD)

DQ123327 DQ123429 DQ123045 AY102396 DQ123240 DQ123153 DQ122959

P. trifasciata Lemaire

Pitangui, MG (NFM)

DQ123328 DQ123430 DQ123046 AY210980 N/A N/A N/A

P. truncata Regel

Brazil, SC (ACC)

N/A DQ123431 DQ123047 AY102390 N/A N/A N/A

P. vespertilio L.

Brazil, PA (LCB)

DQ123329 DQ123432 DQ123048 DQ123495 N/A N/A N/A

P. xiikzodz MacDougal

Italy, Ripalta Cremasca,

Colection (MV)

DQ123330 DQ123433 DQ123049 AY210975 DQ123241 DQ123154 DQ122960

Deidamioides P. lancetillensis MacDougal &

Meerman

French Guiana (MV)

DQ123331 AY212312 DQ123050 AY210963 DQ123242 DQ123155 DQ122961

P. microstipula Gilbert &

MacDougal

Mexico (MV)

DQ123332 DQ123434 DQ123051 AY210966 DQ123243 DQ123156 DQ122962

Murucuja P. murucuja L.

Blois-France-Greenhouse

(RY)

DQ123345 DQ123442 DQ123064 DQ123501 DQ123255 DQ123168 DQ122974

P. tulae Urban

Puerto Rico (MV)

DQ123346 DQ123443 DQ123065 AY102392 DQ123256 DQ123169 DQ122975

Pseudomurucuja P. cupraea L.

Bahamas (MV)

DQ123378 DQ123459 DQ123102 DQ123513 DQ123274 DQ123186 DQ122993

Table 1 (Cont.)

GenBank numbers

Subgenera or tribe

(Killip, 1938)

This study's

proposal

Species

Data of collection

rbcL rps4 trnL trnL-trnF

nad1 b/c nad5 d/e

26S

P. tacsonioides Griseb.

Blois-France-Greenhouse

(RY)

DQ123379 DQ123461 DQ123103 DQ123514 DQ123275 DQ123187 DQ122995

Psilanthus P. sanguinolenta Mast.

Ecology & Evolutionary

Biology Conservatory,

Univ. Connecticut (RY)

N/A DQ123462 DQ123104 DQ123515 DQ123276 DQ123188 DQ122996

Calopathanthus Passiflora P. racemosa Brot.

Brazil, RJ (FCN)

DQ123311 DQ123414 DQ123028 DQ123492 DQ123225 DQ123139 DQ122945

Distephana P. coccinea Aubl.

Peru (MC)

DQ123333 DQ123435 N/A N/A N/A N/A N/A

P. speciosa Gardn.

Brazil, MS (ACC)

DQ123334 AY212293 DQ123052 AY102402 DQ123244 DQ123157 DQ122963

P. vitifolia HBK.

Colombia (MV)

DQ123335 DQ123436 DQ123053 AY210977 DQ123245 DQ123158 DQ122964

Dysosmia P. clathrata Mast.

Brazil, MG (FCN)

DQ123336 DQ123437 DQ123054 DQ123496 DQ123246 DQ123159 DQ122965

P. foetida L.

Brazil, PE (NFM)

DQ123337 AY212291 DQ123055 AY032763 DQ123247 DQ123160 DQ122966

P. palmeri var. sublanceolata

Killip

Italy, Ripalta Cremasca,

Colection (MV)

DQ123338 DQ123438 DQ123056 DQ123497 DQ123248 DQ123161 DQ122967

Dysosmioides P. campanulata Mast.

Brazil, PR (ACC)

DQ123339 AY212317 DQ123057 AY032760 DQ123249 DQ123162 DQ122968

P. setulosa Killip

Brazil, PR (ACC)

DQ123340 AY212297 DQ123058 AY032761 DQ123250 DQ123163 DQ122969

P. villosa Vell.

Brazil, MG (ACC)

DQ123341 AY212308 DQ123059 AY102403 DQ123251 DQ123164 DQ122970

Granalillastrum P. antioquiensis Karst.

Italy, Ripalta Cremasca,

Colection (MV)

DQ123342 DQ123439 DQ123060 DQ123498 DQ123252 DQ123165 DQ122971

P. trisectaMast.

Blois-France-Greenhouse

(RY)

DQ123343 DQ123440 DQ123061 DQ123499 DQ123253 DQ123166 DQ122972

Manicata P. manicata (Juss.) Pers.

Blois-France-Greenhouse

(RY)

DQ123344 DQ123441 DQ123062 DQ123500 DQ123254 DQ123167 DQ122973

Passiflora P. actinia Hook

Brazil, RS (PASS)

DQ123347 AY212301 DQ123065 AY032767 DQ123257 DQ123170 DQ122976

P. acuminata DC.

Brazil, PA (AL)

N/A AY212301 DQ123066 DQ123502 N/A N/A N/A

P. alata Curtis

Brazil, RS (PASS)

DQ123348 AY212323 DQ123067 AY032765 DQ123258 DQ123171 DQ122977

P. ambigua Hemsl.

Brazil, MT (LCB)

DQ123349 DQ123444 DQ123068 DQ123503 DQ123259 DQ123172 DQ122978

P. amethystina Mikan

Brazil, MG (MCD)

N/A AY212323 DQ123069 AY102397 N/A N/A N/A

P. caerulea L.

Brazil, RS (PASS)

DQ123350 AY212316 DQ123070 AY032772 DQ123260 DQ123173 DQ122979

P. cincinnata Mast.

Brazil, MS (ACC)

DQ123351 AY212294 DQ123071 AY102400 DQ123261 DQ123174 DQ122980

Table 1 (Cont.)

GenBank numbers

Subgenera or tribe

(Killip, 1938)

This study's

proposal

Species

Data of collection

rbcL rps4 trnL trnL-trnF

nad1 b/c nad5 d/e

26S

P. edmundoi Sacco

Brazil, BA (NFM)

DQ123352 AY212302 DQ123072 AY102399 DQ123262 DQ123175 DQ122981

P. edulis Sims

Brazil, RS (PASS)

DQ123353 AY212303 DQ123073 AY032769 DQ123263 DQ123176 DQ122982

P. eichleriana Mast.

Brazil, RS (PASS)

DQ123354 AY212304 DQ123074 AY102388 N/A N/A N/A

P. elegans Mast.

Brazil, RS (PASS)

DQ123355 AY212295 DQ123075 AY032766 DQ123264 DQ123177 DQ122983

P.exura

Italy, Ripalta Cremasca,

Colection (MV)

DQ123356 DQ123445 DQ123076 DQ123504

N/A

N/A N/A

P. gabrielliana sp. new

French Guiana (MV)

DQ123357 AY212319 DQ123077 AY210960 N/A N/A N/A

. P. galbana Mast.

Camocin S. Felix, PE

(NFM)

DQ123358 DQ123446 DQ123078 AY032770 DQ123265 DQ123178 DQ122984

P. garkey Mast.

French Guiana (MV)

DQ123359 AY212320 DQ123079 AY210961 N/A N/A N/A

P. incarnata L.

Brazil, SP (BGJ)

DQ123360 AY212306 DQ123080 AY032768 DQ123266 DQ123179 DQ122985

P. ischnoclada

Brazil, PA (LCB)

N/A DQ123447 DQ123081 DQ123505 N/A N/A N/A

P. jilekii Wawra

Brazil, SC (ACC)

DQ123361 AY212318 DQ123082 AY102387 DQ123267 DQ123180 DQ122986

P. kermesina Link & Otto

Brazil, SP (BGJ)

N/A DQ123448 DQ123083 AY032762 N/A N/A N/A

P. maliformis L.

Dominica (MV)

DQ123362 AY212321 DQ123084 AY210964 DQ123268 DQ123181 DQ122987

P. miersii Mast. in Mart.

Brazil, SP (PASS)

DQ123363 DQ123449 DQ123085 AY102395 DQ123269 DQ123182 DQ122988

P. mucronata Lam.

Brazil, PE (MG)

N/A DQ123450 DQ123086 DQ123506 N/A N/A N/A

P. nitida Kunth

Brazil, MT (LCB)

DQ123364 DQ123451 DQ123087 N/A N/A N/A N/A

P. odontophylla Harms ex Glaz.

Brazil, MG (FCN)

DQ123365 DQ123452 DQ123088 DQ123507 N/A N/A N/A

P. quadrangularis L.

Brazil, SP (BGJ)

DQ123366 AY212322 DQ123089 AY032764 N/A N/A N/A

P. recurva Mast in Mart.

Brazil, MG (ACC)

DQ123367 AY212310 DQ123090 AY102391 N/A N/A N/A

P. ripariaMart.

Brazil, PA (LCB)

DQ123368 DQ123453 DQ123091 DQ123508 N/A N/A N/A

P. serratifolia L.

Surinam (MV)

DQ123369 DQ123454 DQ123092 AY210973 N/A N/A N/A

P. serratodigitata L.

Martinique (MV)

DQ123370 DQ123455 DQ123093 AY210972 N/A N/A N/A

P. setacea DC.

Brazil, SP (BGJ)

DQ123371 AY212296 DQ123094 AY102398 N/A N/A N/A

P. sidaefolia M. Roemer

Brazil, MG (MCD)

DQ123372 AY212298 DQ123095 AY102394 DQ123270 DQ123183 DQ122989

P. sprucei Mast.

Italy, Ripalta Cremasca,

Colection (MV)

DQ123373 DQ123456 DQ123096 DQ123509 DQ123271 DQ123184 DQ122990

P. tenuifila Killip

Brazil, RS (PASS)

DQ123374 AY212299 DQ123097 AY032771 DQ123272 N/A DQ122991

Table 1 (Cont.)

GenBank numbers

Subgenera or tribe

(Killip, 1938)

This study's

proposal

Species

Data of collection

rbcL rps4 trnL trnL-trnF

nad1 b/c nad5 d/e

26S

P. trintae Sacco

Brazil, BA (TSN)

DQ123375 DQ123457 DQ123098 DQ123510 N/A N/A N/A

P. urubicensis Cervi

Brazil, SC (ACC)

N/A AY212300 DQ123099 AY102393 N/A N/A N/A

P. watsoniana Mast.

Brazil, BA (AS)

DQ123376 DQ123458 DQ123100 DQ123511 N/A N/A N/A

Tacsonia P. mathewsii (Mast.) Killip

Blois-France-Greenhouse

(RY)

DQ123380 DQ123463 DQ123105 DQ123516 DQ123277 DQ123190 DQ122994

P. mixta L. f.

(RY)

DQ123381 DQ123464 DQ123106 DQ123517 DQ123278 DQ123191 DQ122997

P. tripartita var. mollissima

(Juss.) Poir.

(RY)

DQ123382 DQ123465 DQ123107 DQ123518 DQ123279 DQ123192 DQ122998

P. trifoliata Cav.

Peru (MC)

DQ123383 DQ123466 DQ123108 DQ123519 DQ123280 DQ123193 N/A

Tacsonioides P. luetzelburgii Harms

Brazil, BA (TSN)

DQ123384 DQ123467 DQ123109 DQ123520 DQ123281 DQ123194 DQ122999

P. mendoncaei Harms

Brazil, PR (ACC)

DQ123385 DQ123468 DQ123110 AY102389 DQ123282 N/A DQ123000

P. reflexiflora Cav.

Ecuador (MV)

DQ123386 DQ123469 DQ123111 AY210970 DQ123283 DQ123195 DQ123001

P. umbilicata (Griseb.) Harms

Blois-France-Greenhouse

(RY)

DQ123387 DQ123470 DQ123112 DQ123521 DQ123284 N/A DQ123002

Polyanthea Deidamioides P. cirrhiflora Juss.

Italy, Ripalta Cremasca,

Colection (MV)

DQ123377 DQ123459 DQ123101 DQ123512 DQ123273 DQ123185 DQ122992

Genus P. ovalis

Brazil, BA (TSN)

DQ123401 AY216662 DQ123122 AY210978 DQ123295 DQ123207 DQ123010

Tryphostemmatoides

P. tryphostemmatoides Harms

Blois-France-Greenhouse

(RY)

DQ123388 DQ123471 DQ123113 DQ123522 DQ123285 DQ123196 DQ123003

Passifloreae tribe Adenia isoalensis

(MC)

DQ123389 DQ123472 DQ123115 N/A DQ123286 DQ123198 DQ123004

Adenia keramanthus

(MC)

DQ123390 DQ123473 DQ123114 AY102405 DQ123287 DQ123197 DQ123005

Ancistrothyrsus sp.

Peru (MC)

DQ123391 DQ123474 N/A DQ123523 N/A N/A N/A

Deidamia sp.

(MC)

DQ123394 DQ123477 DQ123117 DQ123526 DQ123289 DQ123201 DQ123007

Dilkea cf johannesii Barb. Rodr.

Peru (MC)

DQ123399 DQ123478 DQ123118 DQ123527 DQ123290 DQ123202 DQ123008

Mitostemma brevifilis

Brazil, MS (ACC)

DQ123400 AY212309 DQ123119 AY102386 DQ123291 DQ123203 DQ123009

Paropsieae tribe Barteria fitulosa

Nigeria (MC)

DQ123392 DQ123475 N/A DQ123524 DQ123288 DQ123199 DQ123006

Barteria sp.

(MC)

DQ123393 DQ123476 DQ123116 DQ123525 N/A DQ123200 N/A

Paropsia braunii

Tanzania (MC)

DQ123395 DQ123479 N/A N/A N/A DQ123204 N/A

Paropsia brazzeana

Africa (MC)

DQ123396 DQ123480 DQ123120 DQ123528 DQ123292 DQ123205 N/A

Table 1 (Cont.)

GenBank numbers

Subgenera or tribe

(Killip, 1938)

This study's

proposal

Species

Data of collection

rbcL rps4 trnL trnL-trnF

nad1 b/c nad5 d/e

26S

Paropsia guneensis

Gold Coast (MC)

DQ123397 DQ123481 N/A N/A N/A N/A N/A

Paropsia madagascariensis

(MC)

AF206802 AY216663 DQ123121 AY102404 DQ123293 DQ123206 N/A

Malesherbiaceae Malesherbia linearifolia

Chile (KG, SK)

DQ123402 DQ123482 DQ123123 DQ123529 DQ123294 DQ123208 DQ123011

Turneraceae Turnera subulata

Brazil, BA (CB)

DQ123398 N/A DQ123124 DQ123530 DQ123296 DQ123209 DQ123012

Collectors: ACC = A. C. Cervi; AL = A. Lins; AS = A. Selbach; BGJ = Banco de Germoplasma de Jaboticabal; CB = C. van den Berg; FCN = F. Campos

Neto; KG = K. Gengler; LCB = L. C. Bernacci; MC=Mark Chase; MCD = M. C. Dornelas; MG = M. Guerra; MV = M. Vecchia; NFM = N. F. Melo; PASS =

our group; RY = R. Yockteng; SK = S. Krosnick; TSN = T. S. Nunes.

Brazilian states: BA = Bahia; MG = Minas Gerais; MS = Mato Grosso do Sul; MT = Mato Grosso; PA = Pará; PR = Paraná; PE = Pernambuco; RJ = Rio de

Janeiro; RS = Rio Grande do Sul; SC = Santa Catarina; SP = São Paulo

TABLE 2: Number of taxa and outgroups considered, sequence and phylogenetic tree characteristics of the data.

DNA regions

Combined data

Characteristics

rbcL rps4

intron

trnL trnL–trnF

nad1 b/c nad5 d/e 26S

cpDNA mtDNA All

Number of taxa 93 100 99 100 77 72 71 99 77 74

Number of outgroups 17 14 12 12 11 13 9 15 9 9

Sequence characteristics

Length of sequenced region

(range)

683 – 1345 479 – 600 314 – 546 239 – 324

898 – 1610 1103 – 1357 722 – 1125

1111 – 2773 771 – 2063

4 430 – 5 913

Aligned length

1345 666 664 393 1787 1545 1127 3068 2187 6382

Analyzed aligned length

1345 666 664 393 843 1344 1127 3068 2187 6382

Variable sites

376 247 178 155 180 228 227 965 342 1480

Parsimony informative sites

237 178 109 71 90 83 123 592 121 705

Pairwise uncorrected

distances (range, %)

0 – 9.1 0 – 10.0 0 – 6.0 0 – 13.3 0 – 4.5 0 – 3.3 0 – 4.2 0 – 16.2 0 – 4.5 0 – 5.4

Average AT content

55.3 63.4 69.6 64.5 46.0 52.1 39.4 60.8 49.6 53.0

Tree characteristics

Parsimony

Number of trees

57 419 2 069 620 1 205 5 382 132 4 2 427 576

Length

951 519 294 271 208 320 549 2 233 438 2 862

CI/RI

0.51/0.81 0.65/0.89 0.74/0.92 0.72/0.89

0.92/0.96 0.80/0.90 0.52/0.78 0.56/0.82 0.84/0.93 0.62/0.82

Likelihood

PAUP score

7 813.471 3 956.488 2 858.708 2 212.631

2 575.551 3 846.605 4 598.699 18 268.774 6 040.234 27 418.611

PhyML score

7 812.643 3 998.949 2 834.748 2 160.828

2 544.575 3 895.092 4 635.927 18 294.505 6 022.834 27 430.730

Treefinder score

7 992.700 3 932.725 2 751.871 2 168.797

2 577.286 3 802.093 4 722.206 18 494.149 6 004.187 27 429.117

Without outgroups.

Consistency Index / Retention Index.

TABLE 3: Average values for the nucleotide diversity found in the seven DNA regions. Values within parentheses refer to standard errors.

rbcL rps4 trnL intron trnL–trnF nad1 b/c nad5 d/e 26S

Total 0.039 (0.004) 0.053 (0.005) 0.017 (0.004) 0.043 (0.009) 0.015 (0.003) 0.013 (0.002) 0.006 (0.002)

Passiflora genus 0.033 (0.004) 0.041 (0.004) 0.025 (0.004) 0.041 (0.009) 0.013 (0.003) 0.011 (0.002) 0.012 (0.002)

Passiflora subgenus 0.013 (0.002) 0.006 (0.001) 0.008 (0.002) 0.014 (0.004) 0.005 (0.001) 0.001 (0.000) 0.007 (0.001)

Decaloba subgenus 0.029 (0.003) 0.033 (0.004) 0.021 (0.003) 0.033 (0.005) 0.009 (0.001) 0.010 (0.002) 0.016 (0.002)

Astrophea subgenus 0.007 (0.001) 0.018 (0.003) 0.013 (0.003) 0.014 (0.004) 0.001 (0.000) 0.001 (0.000) 0.008 (0.001)

Outgroups 0.053 (0.005) 0.095 (0,008) 0.020 (0.005) 0.051 (0.007) 0.042 (0.003) 0.027 (0.002) 0.029 (0.001)

TABLE 4: Models selected by Modeltest and used in the Bayesian (BA) and Maximum

Likelihood (ML) inferences.

DNA region Model Selected by AIC

rbcL GTR+I+G

Shape = 0.5651 Pinvar = 0.8313

rps4 TVM+I+G

Shape = 1.1134 Pinvar = 0.2475

intron trnL GTR+I+G

Shape = 1.1285 Pivar = 0.1483

trnL–trnF GTR+G

Shape = 0.7787 Pivar = 0

nad1 b/c TVM+G

Shape = 0.7621 Pivar = 0

nad5 d/e GTR+I+G

Shape = 0.9423 Pivar = 0.3599

26S GTR+I+G

Shape = 0.5061 Pivar = 0.6080

cp TVM+I+G

Shape = 0.8897 Pinvar = 0.4484

mt TVM+I+G

Shape = 0.9362 Pinvar = 0.4690

Combined GTR+I+G

Shape = 0.8313 Pinvar = 0.5651

Figure 1: Phylogenetic tree based on the four cpDNA regions and obtained by the Bayesian

approach. Numbers above the branches refer to posterior probabilities (PP)

values. Pollinator agents are represented as follows: = bat; =

hummingbird; = wasp; = bee.

Figure 2: Phylogenetic tree considering the two mtDNA regions, obtained by the

Bayesian method. Numbers above the branches refer to posterior

probabilities (PP) values.

Figure 3: Phylogenetic tree for the 26S nrDNA gene obtained by the Bayesian

approach. Numbers above the branches refer to posterior probabilities (PP)

values.

Figure 4: Phylogenetic tree considering all seven DNA regions studied by the

Bayesian analysis. Numbers above the branches refer to posterior

probabilities (PP) values.

Figure 5: Maximum likelihood phylogenetic tree considering all the seven DNA

regions. Numbers above the branches refer to bootstrap (BS) values.

Figure 1

P. podlechii

P. lobii var. ayaucuchoensis

P. rufa

P. truncata

P. morifolia

P. multiflora

P. penduliflora

P. tacsonioides

P. helleri

P. talamancensis

P. murucuja

P. tulae

P. misera

P. pohlii

P. trifasciata

P. vespertilio

P. organensis

P. tricuspis

P. punctata

P. ornithoura

P. capsularis

P. sanguinolenta

P. sexf lora

P. coriacea

P. suberosa

P. xiikzodz

P. cupraea

P. lancetillensis

P. microstipula

P. cirrhiflora

Tetrastylis ovalis

P. racemosa

P. galbana

P. mucronata

P. campanulata

P. set ulosa

P. villosa

P. amethystina

P. exura

P. garkey

P. kermesina

P. miersii

P. watsoniana

P. edmundoi

P. urubiscencis

P. caerulea

P. eichleriana

P. tenui fila

P. sprucei

P. reflexif lora

P. ischnoclada

P. jilekii

P. mendoncae

P. umbi licata

P. coccine a

P. speciosa

P. vitifolia

P. recurva

P. set acea

P. trintae

P. actinia

P. elegans

P. sidaefolia

P. alat a

P. quadrangularis

P. ambi gua

P. riparia

P. odontophylla

P. gabrielliana

P. nitida

P. incarnata

P. serratifolia

P. serratodigitata

P. anti oquiensis

P. mani cata

P. mathewsii

P. tripartita

P. mixta

P. trisecta

P. trifol iata

P. clathrata

P. foetida

P. palm eri

P. cincinnata

P. eduli s

P. maliformis

P. luetzelburgii

P. rhamnifolia

P. haematostigma

P. ceratocarpa

P. kawensis

P. mansoi

P. amoena

P. candida

P. citrifolia

P. lindeniana

P. macrophylla

P. pittieri

P. tryphostemmatoides

P. arborea

Dilkea johannesii

Mitostemma brevifilis

Adenia isoalensis

Deidamia sp.

Adenia keramanthus

Ancistrothyrsus sp.

Barteria fistul osa

Paropsia guineensis

Paropsia braunii

Barteria sp.

Paropsia brazzeana

Paropsia madagascariensis

Malesherbia linearifolia

Turnera subulata

100

Subgenus

Decaloba

Subgenus Deidamioides

Subgenus

Passiflora

Subgenus

Astrophea

Subgenus Tryphostemmatoides

Figure 2

P. podlechii

P. multiflora

P. penduliflora

P. talamancensis

P. murucuja

P. tulae

P. tacsonioides

P. helleri

P. misera

P. pohlii

P. organensis

P. tricu spis

P. ornitho ura

P. punctata

P. cupraea

P. capsularis

P. sanguinolenta

P. sexflora

P. coriacea

P. suberosa

P. xiikzodz

P. morifolia

P. rufa

P. lobii var. ayaucuchoensis

P. lancetillensis

P. microstipula

P. amoena

P. candida

P. citrifolia

P. rhamnifolia

P. arborea

P. ceratocarpa

P. haematostigma

P. kawensis

P. lindeniana

P. macrophylla

P. mansoi

P. pittieri

Parop sia brazzeana

P. racemosa

P. edmundoi

P. tenuifila

P. miersii

P. sprucei

P. reflexiflora

P. coccinea

P. speciosa

P. vitifolia

P. campanulata

P. setulosa

P. villosa

P. incarnata

P. actinia

P. alata

P. umbilicata

P. ambigua

P. caerulea

P. cincinn ata

P. edulis

P. elegans

P. galbana

P. jilekii

P. maliformis

P. sidaefolia

P. luetzelburg ii

P. mendoncae

P. clathrata

P. foetida

P. palmeri

P. antioquiensis

P. manica ta

P. mathew sii

P. trisecta

P. trifoliata

P. mixta

P. tripartita

Adenia isoalensis

Deidamia sp.

P. cirrhiflora

Tetrastylis ovalis

P. tryphostemmatoides

Dilkea johannesii

Mitostemma brevifilis

Turnera subulata

Barteria fistulosa

Malesherbia linearifolia

100

Subgenus

Decaloba

Subgenus

Passiflora

Subgenus

Astrophea

Subgenus Deidamioides

Subgenus Tryphostemmatoides

Figure 3

P. multiflora

P. coriacea

P. suberosa

P. xiikzodz

P. penduliflora

P. capsularis

P. sanguinolenta

P. helleri

P. misera

P. organensis

P. pohlii

P. tricuspis

P. cupraea

P. ornithoura

P. tacsonioides

P. punctata

P. sexflora

P. talamancensis

P. murucuja

P. tulae

P. rufa

P. morifolia

Adenia isoalensis

Deidamia sp.

Adenia keramanthus

P. lancetillensis

P. microstipula

P. amoena

P. ceratocarpa

P. kawensis

P. mansoi

P. haematostigma

P. pittieri

P. citrifolia

P. arborea

P. lindeniana

P. macrophylla

P. tryphostemmatoides

P. cirrhiflora

Tetrastylis ovalis

Dilkea johannesii

Mitostemma brevifilis

P. racemosa

P. galbana

P. clathrata

P. foetida

P. palmeri

P. antioquiensis

P. mathewsii

P. mixta

P. tripartita

P. trisecta

P. luetzelburgii

P. campanulata

P. setulosa

P. villosa

P. actinia

P. alata

P. ambigua

P. caerulea

P. incarnata

P. edmundoi

P. miersii

P. tenuifila

P. jilekii

P. maliformis

P. sidaefolia

P. sprucei

P. mendoncae

P. reflexiflora

P. umbilicata

P. speciosa

P. vitifolia

P. cincinnata

P. edulis

P. elegans

P. manicata

Barteria fistulosa

Malesherbia linearifolia

Turnera subulata

100

Subgenus

Decaloba

Subgenus

Passiflora

Subgenus

Astrophea

Subgenus Tryphostemmatoides

Subgenus Deidamioides

Subgenus Decaloba

Figure 4

P. multiflora

P. penduliflora

P. tacsonioides

P. helleri

P. talamancensis

P. murucuja

P. tulae

P. misera

P. pohlii

P. organensis

P. tricuspis

P. punctata

P. ornithoura

P. cupraea

P. capsularis

P. sanguinolenta

P. sexflora

P. coriacea

P. suberosa

P. xiikzodz

P. lobii var. ayaucuchoensis

P. rufa

P. morifolia

P. lancetillensis

P. microstipula

P. cirrhiflora

Tetrastylis ovalis

P. racemosa

P. galbana

P. campanulata

P. setulosa

P. villosa

P. caerulea

P. edmundoi

P. miersii

P. jilekii

P. mendoncae

P. sprucei

P. reflexiflora

P. tenuifila

P. umbi licata

P. speciosa

P. vitifolia

P. actinia

P. elegans

P. sidaefolia

P. alata

P. ambigua

P. incarnata

P. cincinnata

P. edulis

P. maliformis

P. luetzelburgii

P. clathrata

P. foetida

P. palmeri

P. antioquiensis

P. mani cata

P. mathewsii

P. tripartita

P. mixta

P. trisecta

P. trifoliata

P. rahmnifolia

P. haematostigma

P. ceratocarpa

P. kawensis

P. mansoi

P. amoena

P. candida

P. citrifolia

P. lindeniana

P. macrophylla

P. pittieri

P. tryphostematoides

Dilkea johannesii

Mitostemma brevifilis

Adenia sp.

Deidamia sp.

Paropsia sp.

Turnera subulata

Barteria sp.

Malesherbi a linearifolia

100

Subgenus

Decaloba

Subgenus

Passiflora

Subgenus

Astrophea

Subgenus Tryph ostemmatoide s

Subgenus Deidamioides

Figure 5

P. multiflora

P. lobii var. ayaucuchoensis

P. rufa

P. morifolia

P. penduliflora

P. tacsonioides

P. helleri

P. talamancensis

P. murucuja

P. tulae

P. misera

P. pohlii

P. organensis

P. tricuspis

P. punctata

P. ornithoura

P. capsularis

P. sanguinolenta

P. sexflora

P. coriacea

P. suberosa

P. xiikzodz

P. cupraea

P. lancetillensis

P. microstipula

P. cirrhiflora

Tetrastylis ovalis

P. racemosa

P. galbana

P. tenuifila

P. edmundoi

P. miersii

P. caerulea

P. campanulata

P. setulosa

P. villosa

P. jilekii

P. mendoncae

P. sprucei

P. reflexiflora

P. umbilicata

P. speciosa

P. vitifolia

P. actinia

P. elegans

P. sidaefolia

P. alata

P. ambigua

P. incarnata

P. cincinnata

P. edulis

P. maliformis

P. luetzelburgii

P. clathrata

P. foetida

P. palmeri

P. antioquiensis

P. mathewsii

P. tripartita

P. mixta

P. manicata

P. trisecta

P. trifoliata

P. rhamnifolia

P. haematostigma

P. ceratocarpa

P. mansoi

P. kawensis

P. amoena

P. candida

P. citrifolia

P. lindeniana

P. macrophylla

P. pittieri

P. tryphostemmatoides

Dilkea johannesii

Mitostemma brevifilis

Adenia sp.

Deidamia sp.

Paropsia sp.

Turnera subulata

Barteria sp.

Malesherbia linearifolia

10 changes

100

Subgenus

Decaloba

Subgnenus

Passiflora

Subgenus

Astrophea

Subgenus Deidamioides

Subgenus Tryphostemmatoides

CAPÍTULO IV

2º ARTIGO

A ser submetido para a revista Systematic Botany

DIVERGENCE TIME AND EVOLUTIONARY RATES IN Passiflora

Running head: PASSIFLORA EVOLUTIONARY HISTORY

Divergence Time and Evolutionary Rates in Passiflora

VALÉRIA C. MUSCHNER

, SANDRO L. BONATTO

, FRANCISCO M. SALZANO

and LORETA B.

FREITAS

1,3

Departamento de Genética, Instituto de Biociências, Universidade Federal do Rio Grande do Sul,

Caixa Postal 15053, 91501–970 Porto Alegre, RS, Brazil

Centro de Biologia Genômica e Molecular, Faculdade de Biociências, Pontifícia Universidade

Católica do Rio Grande do Sul, Ipiranga 6681, 90610–001 Porto Alegre, RS, Brazil

Author for Correspondence (loreta.freit[email protected])

Address to which proofs should be sent

Loreta B. Freitas

Departamento de Genética, UFRGS

Caixa Postal 15053

91501–970 Porto Alegre, RS

Brazil

Phone: 55 51 33166715. Fax: 55 51 33166727.

E–mail: loreta.freit[email protected]

CAPÍTULO IV

ABSTRACT A total of 70 species of Passiflora distributed over five subgenera,

plus nine outgroups from related genera and representatives from eight other angiosperm

families were examined in relation to two chloroplast (rbcL, rps4), one mitochondrial (b/c

intron of the nad1 gene), and one nuclear ribosomal (26S) DNA regions to establish

evolutionary rates and divergence times. In a separate analysis the monophyly of Southeast

Asian and Australian species of the Disemma section was confirmed. Passiflora shows

heterogeneous substitution rates among subgenera, probably related to coevolution with

pollinator agents. Its disjunct distribution (Americas / Southeast Asia–Australia) could

probably be explained by Trans–Pacific dispersion which occurred about 42 million years

ago (Ma). The first of the three main subgenera to separate was Decaloba, at about 35 Ma,

the diversification of Passiflora and Astrophea occurring much later (24 Ma).

CAPÍTULO IV

The advantages of DNA sequences for the inference of phylogenetic relationships

among different organisms has been amply demonstrated (Savolainen et al. 2000; APG II,

2003; Muschner et al. 2003) and several studies had identified variation in the rates of

evolution of genes and lineages (Avise 1994; Li 1997; Page and Holmes 1998; Hebert et

al. 2002; Soltis et al. 2002). Those, in turn, led to the development of methods for the

estimation of divergence time (Rambaut and Bromham 1998; Sanderson 2002; Thorne and

Kishino 2002). The use of local strict molecular clock (Zuckerkandal and Pauling 1962,

1965), which follow the principles of the neutral theory of molecular evolution (Kimura

1983), calibrated by the fossils or tectonic moviments, is being increasingly used as a

biogeographical tool for dating the divergence between clades (Renner 2004; Givinish and

Renner 2004).

Renner et al. (2001) asserted that the difficulties encountered in the performance

of biogeographic works dealing with tropical angiosperms, and specifically the

identification of sister groups and divergence times, are due to their high variability and the

scarcity of study material. Biogeographical analyses of the tropical flora attribute

Transtropical disjunctions at high taxonomic levels to the Gondwana break–up (Raven and

Axelrod 1974; Gentry 1982, 1993; Barlow 1990; Burnham and Graham 1999). This

interpretation, however, imply divergence times of 100–90 millions of year ago (Ma)

between the African neotropical clades, and higher values for taxa also found in Indochina

and Southeast Asia. In addiction, fossils supporting the Gondwana break–up are available

for only a few pantropical eudicot families. In the absence of an adequate fossil record for

key areas like South America (Burnham and Graham 1999), the controversy between

break–up Gondwana explanations and those which rely in more recent long–distance

dispersion events for the interpretation of present distribution patterns remains unsettled.

CAPÍTULO IV

Sanmartín e Ronquist (2004) found that in plants the hierarchical patterns observed by

them were incongruent with the sequence of commonly accepted geological events, and

that they seem to have been molded by more recent dispersion and extinction events that

occurred in the southern hemisphere.

Passiflora L. is a large neotropical genus which has about 520 species. Its

taxonomy was until recently rather complicated due to the high variability of its flowers

and vegetative structures. Feuillet and MacDougal (2003), based in morphological

characteristics, divided the genus in four subgenera: Astrophea, Decaloba, Deidamioides,

and Passiflora. Muschner et al. (in mss.) studied seven DNA regions distributed over the

chloroplast (cp), mitochondrial (mt) and nuclear (n) genomes, suggesting that the genus

should be subdivided in the four above–indicated subgenera, plus Tryphostemmatoides.

According to Ulmer and MacDougal (2004) Astrophea and Deidamioides occur in Central

and South America, Passiflora in North and South America, while Decaloba has a disjunct

distribution, with a group in North and South America and another in Southeast Asia and

Australia (Fig. 1). Tryphostemmatoides, with only two species, was up to now only found

in Colombia (South America).

Astrophea, Decaloba, and Passiflora exhibit monophyletic groups with high

statistical support in the molecular analyses (Muschner et al. 2003; Krosnick and

Freudenstein 2005). The latter authors based their analyses in six Southeast Asian and

three Australian species of the Disemma section of the Decaloba subgenus.

The present study investigated four regions of the three genomes (cpDNA,

mtDNA, nDNA) of 70 widely distributed species of Passiflora, to estimate the dates of the

genus probable origin, of subgenera diversification, and their relationship with

biogeographical and/or historical events.

CAPÍTULO IV

MATERIALS AND METHODS

Taxon Sampling and Laboratory Methods. The Passiflora species studied, their

subgenera classification, nine species from eight other genera used as outgroups, the source

of the material, and the GenBank accession numbers of the sequences determined are

displayed in Table 1. Since there is no fossil record for the Passifloraceae (maybe because

these plants are pollinated by animals; see Proctor et al. 1996) sequences from

representatives of seven other angiosperm genera for which fossil data information are

available were obtained from GenBank for purposes of comparison and tree calibration.

Information about them is given in Table 2.

Total DNA was extracted by Roy et al.’s (1992) method, and the regions studied

were as follows: 1. cpDNA rbcL and rps4 genes, amplified with primers 1F and 1460R

(Savolainen et al. 2000) and rps45’ and rps43’ (Souza–Chies et al. 1997) respectively; 2.

mtDNA intron b/c of the gene nad1 gene, amplified by nad1/2 and nad1/3 (Duminil et al.

2002); and 3. partial nuclear ribosomal DNA 26S gene, amplified by N–nc26S1 and 1229r

(Kuzoff et al. 1998). Sequencing primers were the same indicated by the above–indicated

authors, except for the nad1 b/c intron, for which an internal primer was devised, specific

for Passiflora (5’ – ATTCACATAGAGACAGACT). PCR products were cleaned using

the polyethylene glycol/NaCl precipitation method of Dunn and Blattner (1987).

Sequencing was performed in a MegaBace 1000 machine (Amersham Biosciences) using a

DYEnamicTM ET termination cycle sequencing premix kit (Amersham Biosciences), with

the protocol provided by the manufacturer. Sequence alignments were manually

conducted, with ambiguous regions being excluded from the analyses. The sequences were

deposited in the Genbank (accession numbers provided in Table 1).

CAPÍTULO IV

Molecular dating. Bayesian methods (Thorne et al. 1998. Kishino et al. 2001; Thorne and

Kishino 2002) were employed using the MULTIDISTRIBUTE package (available at

http://statgen.nscu.edu/thorne/multidivtime). This parametric approach relaxes the

hypothesis of a strict molecular clock through a continuous autocorrelation in the

substitution rates across the phylogeny and allows the use of several calibration time

constraints.

MULTIDIVTIME divergence date estimates were performed in two stages: first

the parameters of nucleotide substitution were estimated using the BASEML program of

the PAML package (Yang 1997), using the F84 + G model (with five rate categories). Tree

topology was estimated by maximum likelihood (ML) with the PAUP*4.0 program

(Swofford, 1998) using the model selected by the MODELTEST (GTR + I + G) (Posada

and Crandall 1998). The PAML2MODELINF program from the MULTIDISTRIBUTE

package was employed to convert the BASEML output in an archive acceptable for

ESTBRANCHES, also from the MULTIDISTRIBUTE package. The latter was used to

calculate the variance–covariance matrix and the respective trees’ branch lengths.

ESTBRANCHES outputs served as MULTIDIVTIME inputs and the divergence

times were estimated using the appropriate calibration points. A priori and a posteriori

branch ages, their standard errors, and the 95% confidence intervals were inferred via

Markov chain Monte Carlo (MCMC) calculations. MCMC was run for 1,000,000

generations and sampled every 100 generations after an initial bur–in period of 10,000

cycles. To check for convergence analyses were run from two different starting points.

The following prior distributions were used in these analyses: 108 Ma (standard

deviation, SD: 54 Ma) for the expected time between tip and root if there were no

constraints; 0.001 (SD: 0.0005) substitutions per site per million year for the rate of the

CAPÍTULO IV

root node; 0.1 (SD: 0.1) for the parameters which determine the magnitude of

autocorrelation per million years; and 300 Ma for the larger value of the time unit between

the root and the tips. The parameters were chosen following the MULTIDIVTIME

program manual.

Old World Passiflora. Sequences of the cpDNA trnL–trnF and of the nrDNA ITS1 and

ITS2 spacers previously published by Muschner et al. (2003) and Krosnick and

Freudenstein (2005), in a total of 32 Passiflora species of the Decaloba subgenus, were

used to confirm the Disemma section monophyly. For this purpose we utilized the

maximum likelihood (using PAUP* 4.0) and Bayesian (MrBayes v3.0b4; Ronquist and

Huelsenbeck 2003) programs.

RESULTS

Sequence and Tree Characteristics. The size of alignment of the four combined DNA

regions was of 3,040 nucleotides, since due to distant relationships of the Passifloraceae

with the other families used for tree calibration large regions had to be excluded from the

analysis. The average rate of nucleotide substitution obtained by the MULTIDIVTIME

program was of 0.0002 substitutions / site / million years. This rate is however higher in

the Decaloba subgenus, as can be observed by the branch lenghts of the ML tree displayed

in Fig. 2.

Since Krosnick and Freudenstein (2005) included a limited number of New

World species in their analysis we decided to reanalyze their data including a larger

number of Decaloba subgenus taxa. The results of the Bayesian analysis are presented in

Fig. 3. Disemma’s section monophyly was confirmed with high statistical support (PP:

CAPÍTULO IV

100; BS: 97), the Australian and Southeast Asian species separating as distinct clusters.

Basically the same results were obtained using a maximum likelihood approach.

Divergence Date Estimates. Fig. 2 shows the divergence dates obtained. The separation

between Old World (Barteria and Paropsia genera) and New World taxa appears to have

occurred about 56 ± 18 million years ago (Ma). The next event (diversification of the

Passifloraceae family in New World) should be have happened 44 ± 17 Ma, and shortly

afterwards (42 ± 16 Ma) the branching of the Passiflora genus occurred. The first of the

subgenera which separated from the others was Decaloba (35 ± 15 Ma), while the split

between the two others should have occurred 24 ± 14 Ma.

DISCUSSION

This is the first study which considers Passiflora’s biogeographical history and its

subgenera diversification. Wikström et al. (2001) in a work involving 560 plant families

and three molecular markers (the rbcL e atpB cpDNA genes and 18S nrDNA) estimated a

diversification time for the Passifloraceae of 32–36 Ma. We are proposing an older (56

Ma) date, and note that Bremer et al. (2004), after studying six cpDNA markers in a larger

number of Asterid species, also obtained older dates from those of the above–indicated

authors for this group of plants. The application of molecular dating techniques to diverse

sets of data are helping to identify changes in speciation rates and to address questions

about the ecological, geographical, and temporal factors which may influence these rates

(Penington and Dick 2004).

We tried to minimize the problems which may arise by use of calibration points

just in the terminal nodes of a phylogeny (Sanderson 1997), by the inclusion of several

CAPÍTULO IV

external groups, and the highest number of dates possible. Despite this fact, the ages

derived from the present study should be considered with caution, as a first approximation

of the main events which occurred in Passiflora’s evolution.

Even though the sampling at the molecular level of this study is lower than those

of similar analyses (Renner 2004; Richardson et al. 2004; Bell and Donoghue 2005; Yuan

et al. 2005) there are evidences for a Gondwanic origin of the Passifloraceae family, and

according to Raven and Axelrod (1974) the migration between South America and Africa

could have occurred even after the Gondwana break–up, 90–105 Ma. Morley (2003)

reviewed the potential world migration routes for the megathermal angiosperms,

suggesting, for example, that the South American and African connections should have

existed up to the Oligocene (around 35 Ma). This should have happened via stepping stone

dispersal across islands of the Rio Grande Rise and the Walvis Ridge (which according to

Parrish 1993 was above water southwest of the coast of Africa up until that time), as well

as through the Sierra Leone Rise.

Dispersion through Laurasia during the Eocene climatic optimum, when the

conditions supported a tropical vegetation, could be the best explanation for many

organisms that now have a disjunct distribution among the tropics of South America,

Africa, and Southeast Asia (Richardson et al. 2004). Other studies (Renner et al. 2001;

Davis et al. 2002) suggested a boreotropical migration into southern areas during the

Oligocene and Miocene (35–23 Ma). Molecular phylogenetic studies have also

demonstrated that the role of long–distance dispersals to explain modern distribution

patterns may have been subestimated (Renner et al. 2001; Renner, 2004; Yuan et al. 2005).

The Passiflora disjunct distribution could be explained by a Trans–Pacific

(Sanmartín and Ronquist 2004). The [(South America, New Zealand) Australia]

CAPÍTULO IV

100

relationship is the most frequent found in the flora and fauna of the Southern Hemisphere,

and is in conflict with the geologically predicted vicariance patterns (Renner et al. 2000;

Winkworth et al. 2002). Sanmartín and Ronquist (2004) documented a highly

asymmetrical plant directional dispersion, a westward long–distance dispersal from South

América to New Zealand against the prevailing wind and oceanic currents (Winkworth et

al. 2002). Instead of direct jumps, the dispersal could have occurred in a stepping–stone

way along the Antarctic coastline (Renner et al. 2000). This hypothesis is supported by the

presence of temperate florests in these areas until at least the Pliocene (Swenson and

Bremer 1997; Sanmartín and Ronquist, 2004). This dispersal could have been mediated by

the west–flowing East Wind Drift, which runs close to the Antarctic coast, or could have

followed the West Wind Drift around Antarctica, involving dispersal first to the

subantarctic islands (and/or Australia) and from there to New Zealand (Swenson and

Bremer, 1997). This pattern would explain the clade monophyly that we observed in the

Southeast Asian and Australian species of Passiflora.

Differences in evolutionary rates are widespread in plants (Muse 2000), and can

be ascribed to intrinsic factors like genome type (cp, mt, nuclear) or specific regions, as

well as to extrinsic ones like the speed of group speciaton or population size (Bousquet et

al. 1992; Muse 2000; Andreasen and Baldwin 2001; Barraclough and Savolainen 2001).

The heterogeneous and high rates of nucleotide substitution found in Passiflora, that do not

seem to be gene- or genome-specific (data not shown), can be used as a comparison for

those found in other taxa.

As was already mentioned, the Decaloba subgenus has larger branches than those

of the Passiflora and Astrophea subgenera, which presented a pattern of accelerated

radiation. The mechanisms that lead to high lineage diversification constitute a central

CAPÍTULO IV

101

question in evolutionary biology (Malcomber 2002). The classical examples of rapid

species radiation are all geographically restricted, with high sympatry levels. In these

cases, low levels of interspecific competition in poor habitats have been suggested as

important (Jensen 1990; Liem 1990). But key innovations could also increase the radiation

of a given lineage (Hodges and Arnold 1994; Malcomber 2002). Although interspecific

competition could be involved in the rapid diversification of the Passiflora and Astrophea

subgenera, it is more likely that the key factors are related to plant morphology and

pollination agents, as detailed below.

Species of the Passiflora subgenus are characterized by handsome flowers that

are usually dominated by a corona that is usually zoned or banded in different colors. The

corona is the seat of scent production, and this subgenus contains de majority of intensely

fragrant species. The corona is also the landing platform for bees and other insects which

are attracted by odor (Ulmer and MacDougal 2004). In the species pollinated by

hummingbirds, however, the corona is usually reduced.

On the other hand, Astrophea flowers have a short floral tube and a white

coloring that contrast with the often bright yellow corona. They are probably pollinated by

large bees. In other section of this subgenus orange or reddish to purple flowers occur with

a conspicuous floral tube that is longer than the sepals, and a reduced corona. In these

cases the pollinators are probably hummingbirds (Ulmer and MacDougal 2004).

We are, therefore, observing a wonderful example of coevolution between flower

morphology and pollination agents that are now being corroborated by molecular data.

CAPÍTULO IV

102