UNIVERSIDADE ESTADUAL PAULISTA INSTITUTO DE BIOCIÊNCIAS BRUNO FRANCELINO DE MELO RELAÇÕES FILOGENÉTICAS NA SUPERFAMÍLIA ANOSTOMOIDEA (OSTARIOPHYSI: CHARACIFORMES) DOUTORADO BOTUCATU 2015! UNIVERSIDADE ESTADUAL PAULISTA INSTITUTO DE BIOCIÊNCIAS Bruno Francelino de Melo RELAÇÕES FILOGENÉTICAS NA SUPERFAMÍLIA ANOSTOMOIDEA (OSTARIOPHYSI: CHARACIFORMES) Tese apresentada ao Instituto de Biociências de Botucatu, Universidade Estadual Paulista – UNESP, como parte dos requisitos para obtenção do título de Doutor em Ciências Biológicas, Área de Concentração: Genética. Orientador: Dr. Claudio de Oliveira Co-orientador: Dr. Richard Peter Vari BOTUCATU 2015 Palavras-chave: Elementos ultraconservados; Filogenias; Filogenômica; Peixes; Sistemática. Melo, Bruno Francelino de. Relações filogenéticas na superfamília Anostomoidea (Ostariophysi: Characiformes) / Bruno Francelino de Melo. - Botucatu, 2015 Tese (doutorado) - Universidade Estadual Paulista "Júlio de Mesquita Filho", Instituto de Biociências de Botucatu Orientador: Claudio de Oliveira Coorientador: Richard Peter Vari Capes: 20200005 1. Peixe de água doce. 2. Peixe - Classificação. 3. Filogenia. 4. Characiformes. DIVISÃO TÉCNICA DE BIBLIOTECA E DOCUMENTAÇÃO - CÂMPUS DE BOTUCATU - UNESP BIBLIOTECÁRIA RESPONSÁVEL: ROSEMEIRE APARECIDA VICENTE-CRB 8/5651 FICHA CATALOGRÁFICA ELABORADA PELA SEÇÃO TÉC. AQUIS. TRATAMENTO DA INFORM. ! ii! “Aos meus pais, Osvaldo e Jaci, exemplos de força, humildade e sabedoria...” ! iii! AGRADECIMENTOS Expresso aqui meus sinceros agradecimentos às instituições e pessoas que contribuíram para a realização deste trabalho. Ao Prof. Claudio de Oliveira, pela orientação acadêmica em todas as fases do projeto, e às oportunidades oferecidas de se fazer pesquisa com qualidade, além da amizade, essenciais para minha formação. Ao Dr. Richard P. Vari, pela co-orientação de suma importância pra este projeto, pela acolhida no Smithsonian Institution, pela valiosa e sólida colaboração científica que possuímos hoje, e pela rica e forte amizade. Ao Dr. Brian L. Sidlauskas, por ter permitido realizar o doutorado-sanduíche na Oregon State University, por ser um dos responsáveis diretos pela minha inserção em projetos de alto nível científico e por ter aberto as portas para minha experiência internacional. Ao Dr. Michael E. Alfaro, por me acolher na University of California, Los Angeles, por ter possibilitado a viabilização das modernas técnicas de sequenciamento de elementos ultraconservados, e por ter permitido que esta importante etapa do projeto tenha sido realizada com sucesso. Aos colaboradores e amigos: Kendra Hoekzema, parceira de laboratório de biologia molecular; Ben Frable, parceiro de análises filogenéticas e discussões de resultados; Brant Faircloth, conhecedor dos elementos ultraconservados; Mike Burns, Casey Dillman, Mark Sabaj Perez, pelo conhecimento compartilhado referente a este projeto e pela receptividade nas instituições; José L. O. Birindelli, Raphael Covain e Ricardo M. C. Castro, pelo auxílio na identificação das espécies; ao Prof. Fausto Foresti pelos ensinamentos; ao Prof. Ricardo C. Benine, pelas grandiosas colaborações científicas de longa data. A várias coleções científicas, seus curadores e pesquisadores que possibilitaram a ampliação de nossa amostragem: DZSJRP, ANSP, AUM, DePaul University, FMNH, INPA, LIRP, MCP, MHNG, MNHN, MZUSP, OSU, ROM, STRI, UNIR, USNM. Aos amigos do Laboratório de Biologia e Genética de Peixes da UNESP e aos amigos de Botucatu pela contribuição e pelos inúmeros momentos de descontração. À Dra. Cristiane Shimabukuro-Dias e Amanda Alves pela preparação e envio de tecidos aos EUA. ! iv! A Universidade Estadual Paulista, Instituto de Biociências de Botucatu, Programa de Pós- Graduação em Ciências Biológicas (Genética) e Departamento de Morfologia, pela oportunidade e condições necessárias à realização deste estudo. Agradecimento especial à Ana Paula Taioqui, minha linda namorada, que esteve ao meu lado nesta importante etapa da minha vida profissional. Obrigado pela companhia, paciência e amor. Te amo! Agradecimento especial também aos meus pais Osvaldo e Maria Jaci, irmãos Ronaldo, Rosângela e Lucilene (e família: Ewerthon e meus lindos sobrinhos Bárbara e Bryan) e toda minha família Melo e Petrini, sempre oferecendo apoio incondicional. Este projeto foi financiado pela agências de fomento: FAPESP (proc. 2011/08374-1 e BEPE proc. 2013/16436-2), CNPq (# 143365/2011-8 e # 140193/2015-4) e NSF (DEB- 1257898). *** ! v! ! ! vi! AVISO Esta tese é parte dos requerimentos necessários à obtenção do título de Doutor em Ciências Biológicas, Genética, e como tal, não deve ser vista como uma publicação no senso do Código Internacional de Nomenclatura Zoológica (apesar de disponível publicamente sem restrições). Desta forma, quaisquer informações inéditas, opiniões e hipóteses, bem como nomes novos, mudanças taxonômicas não estão disponíveis na literatura zoológica. Pessoas interessadas devem estar cientes de que referências públicas ao conteúdo deste estudo, na sua presente forma, somente devem ser feitas com a aprovação prévia do autor. NOTICE This thesis is a partial requirement for the PhD degree in Biological Sciences with emphasis in Genetics and, as such, should not be considered as a publication in the sense of the International Code of Zoological Nomenclature (although it is available without restrictions). Therefore, any new information, opinions, and hypotheses, as well as new names, taxonomic changes are not available in the zoological literature. Interested people are advised that any public reference to this study, in its current form, should only be done after previous acceptance of the author. ! vii! RESUMO A superfamília Anostomoidea representa um grande grupo de peixes da ordem Characiformes, amplamente distribuído pelos ecossistemas de água-doce neotropicais e representa uma proporção significante da biomassa em algumas regiões. Anostomoidea é reconhecidamente monofilética e formada por quatro famílias: Anostomidae, Chilodontidae, Curimatidae e Prochilodontidae. As relações filogenéticas neste grupo têm sido estudadas por diversos autores, mas as hipóteses apresentadas são ainda discordantes. Internamente, as hipóteses de relação próxima entre Anostomidae e Chilodontidae e entre Curimatidae e Prochilodontidae são suportadas por diversas sinapomorfias baseadas em caracteres morfológicos, porém não completamente corroboradas por estudos moleculares. A incongruência dos estudos prévios associada a uma pequena representatividade das espécies utilizadas nesses estudos evidencia a importância de um estudo filogenético aprofundado usando uma ampla amostragem de Anostomoidea. Neste contexto, o principal objetivo do presente projeto foi estudar as relações filogenéticas entre as famílias, os gêneros e as espécies de Anostomoidea com base em sequências de três loci nucleares e três mitocondriais. Nós também utilizamos, pela primeira vez em peixes neotropicais, novas metodologias filogenômicas usando elementos ultraconservados do genoma. Os resultados das análises de cerca de 1500 elementos ultraconservados em 69 táxons e das análises com seis loci usando 422 táxons (~70% das espécies) revelaram similarmente uma nova hipótese de relacionamento interfamiliar (Anostomidae (Prochilodontidae (Chilodontidae + Curimatidae))). Apresentamos também as filogenias multilocus para cada uma das quatro famílias discutindo as relações intergenéricas e interespecíficas e comparando-as com estudos morfológicos prévios. Muitas hipóteses de relação foram corroboradas e as discordâncias são extensivamente discutidas, incluindo novas propostas de classificação. Também discutimos os processos evolutivos que podem ter sido responsáveis pela diversificação, numa escala temporal determinada por árvores datadas com base no registro fóssil disponível. Nossos resultados deverão contribuir para o entendimento das relações evolutivas que ocorreram na radiação do grande número de espécies que compõe a superfamília Anostomoidea, assim como em outros grandes subgrupos de Characiformes. Palavras-chave: Elementos ultraconservados, filogenias, filogenômica, peixes, sistemática. ! viii! ABSTRACT! The superfamily Anostomoidea represents one of the largest groups within the order Characiformes, widely distributed along Neotropical freshwater ecosystems, representing a significant biomass proportion in some regions. The Anostomoidea is recognizably monophyletic containing four families: Anostomidae, Chilodontidae, Curimatidae, and Prochilodontidae. Several authors have extensively studied the phylogenetic relationships within this group, albeit incongruences among the proposed hypotheses still exist. Several morphological synapomorphies support the close relationship between Anostomidae and Chilodontidae and this clade sister to Curimatidae and Prochilodontidae, whereas molecular phylogenies present distinct resolution. These incongruences associated to a low taxon sampling evidenced the necessity of an extensive phylogenetic study within the superfamily Anostomoidea using a broad taxon sampling. Herein, we aimed to suggest new hypotheses about the phylogenetic relationships among families, genera and species of Anostomoidea using sequences of three nuclear and three mitochondrial loci. We also used, for the first time in Neotropical fishes, modern phylogenomic methods that capture genomic-wide ultraconserved elements. Results from approximately 1500 ultraconserved elements spanning 69 taxa and from six genes with 422 taxa (~70% species) similarly agree with a novel interfamilial hypothesis (Anostomidae (Prochilodontidae (Chilodontidae + Curimatidae). We also presented multilocus phylogenies for each of the four family and we discussed intergeneric and interspecific relationships comparing with previous morphological studies. Many hypotheses of relationships are corroborated and the discordances are extensively discussed, including new proposals of classification. We also discussed evolutionary processes that might explain the species diversification in a temporal scale determined by time-scaled trees using the available fossil record. Our results will contribute to understand the evolutionary relationships driving the radiation of the large number of species of the superfamily Anostomoidea as well as in other major subgroups of Characiformes. Keywords: Phylogenies, phylogenomics, fishes, systematics, ultraconserved elements. ! Sumário(–(Doutorado(–(Bruno(F.(Melo( ! ! ! ! ! 9! SUMÁRIO PREFÁCIO ________________________________________________________________ 9 INTRODUÇÃO ____________________________________________________________ 12 Estudos filogenéticos em Characiformes _________________________________________ 12 Superfamília Anostomoidea ___________________________________________________ 16 Prochilodontidae + Curimatidae ________________________________________________ 19 Anostomidae + Chilodontidae _________________________________________________ 21 JUSTIFICATIVA __________________________________________________________ 24 OBJETIVOS ______________________________________________________________ 25 MATERIAL E MÉTODOS, RESULTADOS E DISCUSSÃO ______________________ 26 Phylogenomic analysis reveals multiple episodes of diversification and convergence within the radiation of South American anostomoid fishes (Ostariophysi: Characiformes) ___________ 28 Miocene radiations in the family Prochilodontidae (Ostariophysi: Characiformes) revealed by a time calibrated molecular phylogeny _____________________________________________ 52 The Neotropical fish family Curimatidae (Teleostei: Ostariophysi: Characiformes): classification and phylogenetic relationships based on molecular data ______________________________ 82 Molecular phylogenetics of the family Anostomidae (Teleostei: Characiformes) _________ 116 LITERATURA CITADA ___________________________________________________ 141 ANEXO 1: The first molecular phylogeny of Chilodontidae (Teleostei: Ostariophysi: Characiformes) reveals cryptic biodiversity and taxonomic uncertainty ________________ 157 ANEXO 2: New species of Cyphocharax (Characiformes: Curimatidae) from the upper rio Negro, Amazon basin ________________________________________________________ 166 ANEXO 3: Produção Acadêmica e Científica no Doutorado: 2011–2015 _______________ 173 Prefácio)–)Doutorado)–)Bruno)F.)Melo) ! 10! PREFÁCIO Estudos abordando a história evolutiva através de reconstrução filogenética das famílias de Characiformes foram, até o início do novo século, predominantemente baseados em caracteres morfológicos. Buscando contribuir para o conhecimento da evolução de grandes subgrupos de Characiformes, esta tese objetivou analisar as relações filogenéticas de quatro famílias (Anostomidae, Chilodontidae, Curimatidae e Prochilodontidae) que formam a superfamília Anostomoidea, com base em caracteres moleculares utilizando diferentes metodologias. Os resultados apresentados aqui são frutos de uma importante rede de colaboração científica que foi implementada para compreender os processos evolutivos responsáveis pela enorme radiação de espécies da superfamília Anostomoidea. Reitero que estes achados são cooperativamente divididos entre os membros deste grupo de pesquisa. Da mesma forma, esta tese é parte do trabalho intelectual de todos os membros desta equipe e de pesquisas constantes que, indubitavelmente, irão se manter ao longo dos próximos anos. Inicialmente, apresento uma introdução geral com detalhes de todo o histórico de estudos realizados e cada hipótese levantada, utilizando diferentes metodologias, tais como uso de caracteres morfológicos e moleculares. Muitas destas hipóteses foram primariamente propostas em estudos de longa data por Richard Vari (USNM), que se dedicou imensamente em compreender as relações entre esses táxons. Apresento seguidamente a justificativa e os objetivos deste trabalho. Em seguida, apresento quatro capítulos inéditos contendo suas metodologias específicas, os resultados encontrados com detalhadas discussões. No primeiro, apresento os resultados das primeiras análises utilizando filogenômica com elementos ultraconservados do genoma. Neste, que é o primeiro estudo filogenômico com peixes Neotropicais, utilizamos cerca de 1500 loci ortólogos, distribuídos ao longo do genoma de 69 táxons, e reconstruímos as relações entre as famílias e entre a maioria dos gêneros de Anostomoidea. A árvore final obtida foi datada, com base no registro fóssil, para identificar, temporalmente, os eventos que promoveram a diversificação na superfamília. Comparo também com a topologia final da filogenia multilocus com 422 táxons e 70% de cobertura das espécies. A filogenia de elementos ultraconservados é a principal hipótese molecular fortemente sustentada das relações interfamiliares que nosso grupo de pesquisa já obteve. Prefácio)–)Doutorado)–)Bruno)F.)Melo) ! 11! No segundo capítulo, apresento a filogenia molecular da família Prochilodontidae utilizando dados multilocus, com 90% de representatividade das espécies. Mostro também os resultados das análises de relógio molecular, datado com o registro fóssil, e os eventos que podem ter proporcionado a radiação das espécies da família. No terceiro capítulo, apresento a primeira filogenia molecular dos peixes detritívoros da grande família Curimatidae, com 75% de cobertura das espécies. Neste capítulo, apresentamos novas hipóteses de relacionamento entre gêneros e espécies. Também sugerimos mudanças na classificação dos táxons, com propostas de sinonimização de um dos gêneros e de realocação de algumas espécies, com o objetivo de se reconhecer agrupamentos monofiléticos. No quarto e último capítulo, apresentamos as relações entre representantes da família Anostomidae, a maior família entre as quatro de Anostomoidea, com um total de 62% de cobertura das espécies. Novos relacionamentos são propostos, assim como o reconhecimento de vários clados monofiléticos dentro da linhagem que contém o gênero Leporinus. Em síntese, apresentamos também as relações entre todos os táxons de Anostomoidea utilizando dados multilocus e discutimos brevemente as principais discrepâncias em relação aos resultados com elementos ultraconservados. Ao final, incluí dois artigos já publicados. O Anexo 1 contém o trabalho das relações filogenéticas em Chilodontidae. Neste artigo, sustentamos vários agrupamentos previamente propostos mas descobrimos várias espécies crípticas dentro dos dois gêneros que compõem a família, através de análises multilocus. O Anexo 2 contém um artigo de descrição de uma nova espécie de Cyphocharax da bacia amazônica, fruto também das nossas pesquisas com taxonomia em Curimatidae. O Anexo 3 mostra a minha produção acadêmica e científica durante este doutorado. Assim, finalizamos essa contribuição para o entendimento das relações filogenéticas na superfamília Anostomoidea. *** Introdução*–*Doutorado*–*Bruno*F.*Melo* ! 12! INTRODUÇÃO Estudos filogenéticos em Characiformes A fauna de peixes neotropicais de água-doce continentais é composta por mais de 7.000 espécies, sendo de longe a mais rica fauna de vertebrados continentais no planeta (Albert and Reis 2011a, 2011b, e referências citadas). A ictiofauna é dominada, tanto em termos de diversidade taxonômica quanto em biomassa, por peixes da superordem Ostariophysi, série Otophysi, que alcançam aproximadamente 73% das espécies descritas, divididas primariamente entre Siluriformes (~37% das espécies) e Characiformes (~33% das espécies) e, secundariamente, por Gymnotiformes (~3% das espécies); as espécies restantes dividem-se entre Cichlidae (~9%), Cyprinodontiformes (~12%), ambas da superordem Acanthopterygii, e ainda todo um conjunto (~7% do total de espécies) formado por vários grupos taxonômicos (Potamotrygonidae, Lepidosirenidae, Osteoglossidae, Engraulidae, Clupeidae, Synbranchidae, Sciaenidae, Gobiidae, Nandidae, Belonidae, Tetraodontidae, etc.). A ordem Characiformes constitui um dos maiores e mais diversificados grupos de peixes de água-doce contendo atualmente cerca de 2.100 espécies válidas, alocadas em cerca de 270 gêneros e 23 famílias (Eschmeyer e Fong 2015). Possui distribuição ampla nas maiores drenagens da África subsaariana, América do Sul (exceto o sul da Argentina e Chile), América Central, México e o sul dos Estados Unidos. É representada por espécimes de diversos tamanhos, desde miniaturas (sensu Weitzman e Vari 1988) como algumas espécies de Crenuchidae, Cheirodontinae e Lebiasinidae até alguns de grande porte, atuando como importantes recursos para pesca (e.g. Brycon, Hydrocinus, Leporinus, Piaractus, Prochilodus, Salminus). A idade mínima das principais linhagens de peixes neotropicais (ordens Characiformes, Siluriformes e Gymnotiformes) data do período Cretáceo, entre 90 e 112 milhões de anos (Ma) (Lundberg et al. 1998). Estudos recentes, utilizando o registro fóssil com datação de relógios moleculares, corroboraram a origem de diversificação de Characiformes no Cretáceo há aproximadamente 108 Ma (Near et al. 2012; Arroyave et al. 2013). Espécies pertencentes à ordem Characiformes formam um agrupamento natural, como proposto por Müller (1842) e reafirmado por Müller e Troschel (1845) e, um século depois, por Greenwood et al. (1966). Após o advento da metodologia cladística ou filogenética (Hennig Introdução*–*Doutorado*–*Bruno*F.*Melo* ! 13! 1950, 1966), a ordem Characiformes foi analisada por Fink e Fink (1981), que propuseram sua monofilia com base em sete sinapomorfias. No entanto, as relações entre ordens de Ostariophysi, assim como o estabelecimento da monofilia de cada uma delas, não estão completamente esclarecidas. Nakatani et al. (2011; Fig. 2) resumem as incongruências no relacionamento entre as ordens de Ostariophysi apresentadas por diferentes hipóteses (Fink e Fink 1981; Dimmick e Larson 1996; Ortí e Meyer 1996; Ortí e Meyer 1997; Saitoh et al. 2003; Peng et al. 2006; Lavoué et al. 2008; Li et al. 2008; Poulsen et al. 2009). Utilizando mitogenomas de mais de 100 espécimes de Otophysi, os autores também não encontraram suporte para a monofilia de Characiformes, sugerindo que a subordem Characoidei (Characiformes americanos) é mais relacionada à Siluriformes do que Citharinoidei (Characiformes africanos), tornando os Characiformes parafiléticos. Estudos futuros são necessários para testar a monofilia de Characiformes utilizando um maior número de caracteres em escala genômica. Somente uma publicação até agora analisou as relações de alto nível entre os principais componentes de Characiformes com base em dados morfológicos (Buckup 1998). Este estudo analisou 80 caracteres de 27 grupos internos com representantes de todos os Characiformes conhecidos menos Cynodontidae, Gasteropelecidae e Serrasalmidae (previamente reconhecidas como subfamílias de Characidae) e apenas sete gêneros de Characidae. Outros estudos morfológicos focaram em questões variando da relação entre algumas famílias a relações dentro de famílias ou entre seus componentes. Estes incluem os Alestidae (Zanata e Vari 2005), Anostomidae (Vari 1983; Sidlauskas e Vari 2008), Characidae (Reis 1989; Lucena e Menezes 1998; Malabarba 1998; Toledo-Piza 2000; Vari e Harold 2001; Malabarba e Weitzman 2003; Mirande 2009, 2010; Mattox e Toledo-Piza 2012), Chilodontidae (Vari 1983; Vari et al. 1995), Citharinidae (Vari 1979), Crenuchidae (Buckup 1998), Curimatidae (Vari 1983, 1989a), Distichodontidae (Vari 1979), Erythrinidae (Oyakawa e Mattox 2009; Mattox et al. 2014) Hemiodontidae (Langeani 1998), Prochilodontidae (Vari 1983; Castro e Vari 2004), dentre outros. Métodos moleculares têm sido utilizados há vários anos para estudar as relações internas em Characiformes. Os primeiros estudos foram publicados por Ortí e Meyer (1997). Baseados em sequências parciais dos genes 12S e 16S rRNA (cerca de 880 pb), os autores procuraram investigar as relações dentro das ordens de Ostariophysi, em famílias em Characiformes e entre gêneros em Serrasalminae (=Serrasalmidae). Ainda que uma hipótese filogenética de Introdução*–*Doutorado*–*Bruno*F.*Melo* ! 14! relacionamento entre os Characiformes seja apresentada, Ortí e Meyer (1997) notaram que as relações entre as famílias não puderam ser reconstruídas com segurança. Entre as hipóteses interessantes propostas nestes estudos estão a posição basal das famílias africanas Citharinidae e Distichodontidae, a hipótese de que a união de grupos africanos e neotropicais não constituem grupos naturais e a hipótese de que Serrasalmidae não é intimamente relacionada com Characidae. Calcagnotto et al. (2005) apresentaram um estudo filogenético amplo com base na análise de dois genes mitocondriais e quatro genes nucleares em 124 táxons de Characiformes, incluindo 59 representantes de famílias africanas, embora não tenham incluído representantes das famílias Curimatidae e Gasteropelecidae. O grande número de táxons africanos analisados neste estudo permitiu corroborar a natureza monofilética das famílias Citharinidae, Distichodontidae, Alestidae e Hepsetidae (sendo a última a única monotípica) e novamente rejeitando o conceito de que grupos africanos e neotropicais formam unidades monofiléticas. Entretanto, o pequeno número de táxons neotropicais analisado, principalmente Characidae, não permitiu uma melhor resolução da hipótese apresentada. Characidae, a maior família de peixes de água-doce Neotropical, ainda é a que apresenta os maiores problemas taxonômicos e sistemáticos. Utilizando dados morfológicos, Mirande (2009, 2010) suporta a monofilia da família com base em oito sinapomorfias, embora sete delas possuam homoplasias. Com dados moleculares, Javonillo et al. (2010) propuseram três unidades monofiléticas definidas então como clados A, B e C dentro de Characidae. Subsequentemente, Oliveira et al. (2011) apresentaram a primeira filogenia de Characiformes incluindo representantes de todas as famílias e uma grande amostragem de Characidae. Os autores utilizaram dois loci mitocondriais e três nucleares de 213 espécies representando todas as famílias então reconhecidas, e apresentaram uma nova hipótese de classificação dentro da ordem Characiformes (Fig. 1). Nesta hipótese, diversos grupos antes considerados como membros de Characidae são alocados em outros grupos monofiléticos e Characidae é redefinido como um grande grupo monofilético com quatro subclados monofiléticos. !! ! Introdução*–*Doutorado*–*Bruno*F.*Melo* ! 15! Figura 1. Relações filogenéticas em Characiformes com base em dados moleculares (Oliveira et al. 2011). Filogenias moleculares recentes vêm sendo publicadas para diversos subgrupos de Characiformes. Dentro de Characidae, estudos com Aphyocharacinae (Tagliacollo et al. 2012) e Cheirodontinae (Mariguela et al. 2013b) permitiram uma melhor definição dessas subfamílias com propostas de organização de diversos gêneros, mas também apontaram para diversos possíveis problemas taxonômicos não suspeitados até o momento. A nível de gênero, Mariguela et al. (2013a) analisaram 29 espécies de Moenkhausia e identificaram cinco unidades monofiléticas que mostram o gênero como polifilético. Por outro lado, o emprego de dados moleculares tem-se mostrado muito útil, permitindo uma maior segurança na descrição de espécies (Benine et al. 2009; Melo et al. 2011; Silva et al. 2013) e até mesmo gêneros (Netto- Ferreira et al. 2013) em Characidae. Assim, estudos com maior número de gêneros e outros marcadores moleculares podem ajudar a elucidar as relações existentes dentro da ordem Characiformes e ainda permitir o reconhecimento de novos grupos taxonômicos. Estudos extensivos em outras famílias de Characiformes também têm mostrado resultados relevantes. Arroyave e Stiassny (2011) estudaram a família africana Alestidae e não recuperaram sua monofilia pois o gênero Chalceus foi alocado dentro do grupo de Characiformes sul-americanos. Além disso, os autores não recuperaram a monofilia de vários dos seus gêneros. Subsequentemente, Arroyave et al. (2013) publicaram um trabalho contendo muitos representantes de Distichodontidae no qual corroboram a monofilia da família e de vários Introdução*–*Doutorado*–*Bruno*F.*Melo* ! 16! gêneros, mas também suportam o parafiletismo de alguns gêneros como Phago, Nannocharax, e Hemigrammocharax, este recentemente sinonimizado em Nannocharax (Jerep e Vari 2014). A filogenia molecular de Gasteropelecidae (Abe et al. 2013) mostrou que vários grupos (gêneros e espécies) não são monofiléticos, havendo portanto a necessidade de uma revisão da família. Em Bryconidae (Abe et al. 2014), os resultados mostraram que o principal gênero, Brycon, não é monofilético, mas quatro grandes grupos foram identificados na família, sendo Salminus e Brycon pesu os únicos grupos monofiléticos. A filogenia molecular de Chilodontidae (Melo et al. 2014) corroborou a monofilia de cada um dos gêneros mas revelou também a presença de várias espécies crípticas e a necessidade de revisão taxonômica. Um estudo recentemente publicado em Serrasalmidae (Thompson et al. 2014) sugere que essa família tenha se originado no período Cretáceo, mas a diversificação das piranhas (o grupo mais numeroso) seria mais recente, tendo se originado durante o Eoceno. Superfamília Anostomoidea A relação próxima entre as famílias Anostomidae, Chilodontidae, Curimatidae e Prochilodontidae dentro da ordem Characiformes foi primeiramente estudada por autores pré- cladistas (Günther 1864; Boulenger 1904; Regan 1911; Eigenmann 1917; Gregory e Conrad 1938; Roberts 1973; Géry 1977) que, por muitas vezes incluíram outras famílias nas análises, como Parodontidae, Lebiasinidae e Hemiodontidae. Entretanto, foi num extenso estudo baseado principalmente em osteologia que Vari (1983) propôs o monofiletismo do grupo das quatro famílias, denominado superfamília Anostomoidea, sustentado por quatro sinapomorfias. O autor também propõe a relação ((Curimatidae + Prochilodontidae)(Anostomidae + Chilodontidae)) sustentadas por 11 e 15 sinapomorfias para o primeiro e o segundo clado, respectivamente (Fig. 2). Essas relações interfamiliares foram subsequentemente corroboradas por vários estudos morfológicos (Buckup 1998; Castro e Vari 2004; Moreira 2007; Sidlauskas 2008). Introdução*–*Doutorado*–*Bruno*F.*Melo* ! 17! ! Figura 2. Relações filogenéticas entre as quatro famílias de Anostomoidea com dados morfológicos, modificada de Vari (1983). Números representam as respectivas sinapomorfias que sustentam cada clado. A monofilia de Anostomoidea, proposta pelos trabalhos anteriores e sustentada por caracteres morfológicos foi, entretanto, rejeitada por Ortí e Meyer (1997) usando dados moleculares. O estudo corrobora, no entanto, o agrupamento formado por Prochilodontidae e Curimatidae proposto por Vari (1983), e este aparecendo como grupo-irmão de Hemiodontidae. A família Chilodontidae surpreendentemente agrupa-se com Crenuchidae, ambas relacionadas com Anostomidae. Assim, as relações em Anostomoidea aparecem de forma distinta das propostas anteriores. Embora os autores tenham incluído representantes das 18 famílias então reconhecidas de Characiformes (sensu Nelson 2006) e vários grupos de Characidae, os genes mitocondriais escolhidos (12S e 16S) não resolveram as relações devido, reconhecidamente, à saturação das mudanças nucleotídicas e/ou ao modo de diversificação evolutiva dos táxons, como discutido pelos próprios autores (Ortí e Meyer 1997: 91). Diferentemente do estudo anterior, Calcagnotto et al. (2005) corroboraram Anostomoidea (sensu Vari 1983), porém os autores não usaram nenhum representante de Curimatidae, impossibilitando se testar a hipótese de relações próximas entre as quatro famílias. A topologia apontou discordância com os estudos anteriores, evidenciando Chilodontidae irmão de Prochilodontidae, e estes relacionados com Anostomidae. Outro resultado interessante refere-se ao grupo-irmão de Anostomoidea, composto por Parodontidae e Hemiodontidae, previamente Introdução*–*Doutorado*–*Bruno*F.*Melo* ! 18! indicados como aparentados em estudos pré-cladistas (e.g. Boulenger 1904) e em estudos recentes (e.g. Moreira 2007). Utilizando uma ampla amostragem dentro de Characiformes e Characidae, Oliveira et al. (2011) corroboram a monofilia de Anostomoidea (Fig. 1) com altos índices de suporte estatístico, sendo este agrupado num grande clado composto por Cynodontidae, Serrasalmidae, Hemiodontidae e Parodontidae. O agrupamento formado por Prochilodontidae e Curimatidae dentro de Anostomoidea é corroborado. Entretanto, este trabalho não resolve completamente a posição de Chilodontidae, sendo esta alocada em diferentes clados nas diferentes análises (Bayesiana, máxima verossimilhança e máxima parcimônia). Os autores sugerem que a hipótese de relacionamento das famílias proposta por Vari (1983) seja testada com a inclusão de mais táxons. Melo et al. (2014; Anexo 1) testaram as hipóteses de relacionamento na família Chilodontidae propostas por Vari et al. (1995) corroborando o monofiletismo da família e dos dois gêneros, mas também evidenciou a existência de diversas espécies crípticas além de uma necessidade de revisão taxonômica, principalmente em Caenotropus labyrinthicus e Chilodus punctatus sensu lato. Mesmo que os autores tenham utilizado um número relativamente baixo de representantes de outras famílias de Anostomoidea, as relações entre as quatro famílias novamente não foram reconstruídas como a proposta de Vari (1983). Pela primeira vez, a relação se apresentou como (Prochilodontidae (Chilodontidae + Curimatidae)) e Anostomidae como um clado não resolvido na base da filogenia. Os autores destacam, considerando esta hipótese molecular, duas hipóteses para explicar tal arranjo: todos os caracteres comuns a Anostomidae e Chilodontidae são convergentes, ou eles estiveram presentes no ancestral comum de Anostomoidea e foram subsequentemente perdidos durante a evolução de Curimatidae e Prochilodontidae. Embora o trabalho de Ortí e Meyer (1997) não recupere a monofilia de Anostomoidea, todos os demais estudos moleculares corroboram este agrupamento. Assim, nos quatro trabalhos utilizando caracteres moleculares, as posições de cada uma das famílias de Anostomoidea ainda permanecem incertas e a hipótese de Vari (1983) ainda não foi satisfatoriamente refutada. Em suma, cinco diferentes padrões de relacionamento (Fig. 2 do Capítulo 1) entre as famílias de Anostomoidea foram propostas até o presente, através de estudos morfológicos realizados por Vari (1983) e corroborados por outros estudos morfológicos (e.g. Buckup 1998) e moleculares Introdução*–*Doutorado*–*Bruno*F.*Melo* ! 19! realizados por Ortí e Meyer (1997), Calcagnotto et al. (2005), Oliveira et al. (2011) e Melo et al. (2014). Prochilodontidae + Curimatidae Estudos anteriores já haviam mostrado relação próxima entre essas duas famílias, mesmo no período pré-cladístico. Günther (1864) foi um dos primeiros a relacionar Curimatus (atuais Curimatidae) com Prochilodus num clado chamado “grupo Curimatina”, além de diversos outros rearranjos propostos por outros autores, envolvendo Parodontidae, Citharinidae e Hemiodontidae (ver revisão em Vari 1983). No entanto, análises baseadas em caracteres morfológicos realizadas por Vari (1983) apontaram Prochilodontidae como mais relacionada à Curimatidae sustentados por 11 sinapomorfias (Fig. 2). Este relacionamento é corroborado por Vari (1989a) que incluiu mais quatro novas sinapomorfias, e também por Ortí e Meyer (1997), Buckup (1998) e Oliveira et al. (2011). Alternativamente, Calcagnotto et al. (2005) diferem dos anteriores, indicando uma relação de Prochilodontidae com Chilodontidae por uma nítida razão: a ausência de Curimatidae. Melo et al. (2014, Anexo 1) encontraram uma nova reconstrução, com Curimatidae sendo irmão de Chilodontidae, e este clado irmão de Prochilodontidae com moderado suporte estatístico, embora os autores tenham apenas usado uma grande amostragem de Chilodontidae, foco principal do estudo. Vari (1983) sustentou a monofilia de Prochilodontidae baseando-se em 18 sinapomorfias associadas principalmente aos arcos branquiais e hióide, suspensório e maxilas. Num extenso trabalho, Castro e Vari (2004) utilizaram 125 caracteres e corroboraram esta hipótese adicionando 40 outras sinapomorfias para a família. Estes autores delimitam a família em três gêneros monofiléticos com um total de 21 espécies, em que Prochilodus é considerado grupo- irmão do clado Semaprochilodus + Ichthyoelephas (Fig. 3). No entanto, as relações interespecíficas não foram completamente resolvidas em Prochilodus e Semaprochilodus. Introdução*–*Doutorado*–*Bruno*F.*Melo* ! 20! ! Figura 3. Relações intergenéricas em Prochilodontidae com dados morfológicos (Castro e Vari 2004). Curimatidae é considerada monofilética e sustentada por dezenove sinapomorfias (Vari 1983, 1989a). As relações intra e intergenéricas são compreendidas devido a uma série de trabalhos realizados por Vari (1982a, 1983, 1984, 1989a, 1989b, 1989c, 1989d, 1991, 1992a, 1992b). Embora as relações mais basais tenham sido mais resolutivas, os gêneros Curimatella, Cyphocharax, Pseudocurimata e Steindachnerina formam um clado não resolvido e sem monofiletismo estabelecido (Vari 1989a) (Fig. 4). A família possui, atualmente, 105 espécies válidas distribuídas em oito gêneros, e diversas espécies novas têm sido constantemente descritas (e.g. Vari et al. 2010, 2012; Netto-Ferreira e Vari 2011; Melo e Vari 2014, Anexo 2). Fig. 4. Relações intergenéricas em Curimatidae com dados morfológicos (Vari 1989a). Introdução*–*Doutorado*–*Bruno*F.*Melo* ! 21! Anostomidae + Chilodontidae Vari (1983) sustentou a monofilia deste clado com 15 sinapomorfias associadas ao aparato branquial e aos sistemas internos do corpo (Fig. 2). Posteriormente, esta proposta foi corroborada por outros estudos morfológicos (Buckup 1998; Moreira 2007; Sidlauskas 2008), mas não por estudos moleculares, os quais indicam grupos diferentes como irmãos de Chilodontidae e de Anostomidae. Ortí e Meyer (1997), por exemplo, recuperaram a relação (Anostomidae (Chilodontidae + Crenuchidae)). Calcagnotto et al. (2005) encontraram a relação (Anostomidae (Chilodontidae + Prochilodontidae)) com baixo suporte estatístico. Oliveira et al. (2011) também não recuperaram a hipótese de Vari (1983) sendo que Anostomidae e Chilodontidae aparecem com posição indefinida na base do clado de Anostomoidea. Melo et al. (2014) novamente não recuperam a relação de Anostomidae com Chilodontidae, sendo que esta última família é resolvida como irmã de Curimatidae. Embora esses estudos mostrem avanços no nosso conhecimento sistemático em Characiformes, muitos deles possuem similaridades nos caracteres moleculares, uma vez que pertencem aos mesmos loci, e também devido ao foco dos estudos não serem exatamente as famílias de Anostomoidea. Assim, a hipótese de Vari (1983) ainda não foi corroborada com caracteres moleculares. Diversos estudos desde Günther (1864) com o ‘grupo Anostomina’ até Géry (1977) têm sido realizados com os representantes da atual família Anostomidae. No contexto cladista, Vari (1983) sustenta sua monofilia com base em 15 sinapomorfias, tendo Chilodontidae como grupo- irmão (Fig. 2). Internamente, Winterbottom (1980) estabelece a monofilia da subfamília ‘Anostominae’ que engloba Anostomus, Gnathodolus, Pseudanos, Sartor e Synaptolaemus, externamente distinta dos demais gêneros pela posição superior da boca, porém com poucas espécies da atual composição de toda a família. Revisões completas foram realizadas nos gêneros Abramites (Vari e Williams 1987) e Laemolyta (Mautari e Menezes 2006), enquanto que parciais vem sendo feitas em Leporinus (e.g. Garavello 1979; Britski et al. 2012; Birindelli e Britski 2013), Pseudanos (Sidlauskas e Santos 2010; Birindelli et al. 2012) e Schizodon (e.g. Bergmann 1988; Garavello e Britski 1990). Recentemente, Sidlauskas e Vari (2008) realizaram uma extensiva análise filogenética na família Anostomidae baseada em caracteres morfológicos, sustentando a monofilia de todos os gêneros da família, exceto Leporinus (Fig. 5). O estudo é bastante resolutivo nas relações entre os gêneros Anostomoides, Rhytiodus, Schizodon, Laemolyta e na posição de Leporellus como grupo-irmão de todos os demais membros da Introdução*–*Doutorado*–*Bruno*F.*Melo* ! 22! família. Os autores descrevem Petulanos como um novo gênero alocado na subfamília Anostominae e ressuscitam o gênero Hypomasticus. No entanto, os resultados mostram Abramites e os demais gêneros (Anostomoides, Laemolyta, Rhythiodus, Schizodon e os membros de Anostominae) alocados dentro do clado contendo as espécies de Leporinus, evidenciando seu parafiletismo. Além disso, os autores amostraram poucas espécies de Leporinus (14 de ~90 válidas), evidenciando que mais estudos filogenéticos são necessários para entender as relações intergenéricas em Anostomidae. A família possui atualmente 14 gêneros (sensu Sidlauskas e Vari 2008) contendo cerca de 150 espécies válidas. ! Fig. 5. Relações filogenéticas em Anostomidae baseada em caracteres morfológicos (Sidlauskas e Vari 2008). Introdução*–*Doutorado*–*Bruno*F.*Melo* ! 23! Com dois gêneros (Caenotropus e Chilodus) e oito espécies, algumas de interesse para a aquariofilia, a família Chilodontidae é distribuída pelas bacias do Orinoco, Amazonas, rios independentes das Guianas e no rio Parnaíba. Com base em 26 sinapomorfias, Vari (1983) sustenta a monofilia de Chilodontidae (Fig. 2), sendo corroborada por Vari et al. (1995) que adicionam dez sinapomorfias para o grupo. Os autores ainda propõem Chilodus e Caenotropus como grupos monofiléticos sustentados por oito sinapomorfias cada (Fig. 6), esclarecem ainda as relações intragenéricas em Caenotropus e discutem a biogeografia de Chilodontidae com base nas suas relações filogenéticas. Isbrücker e Nijssen (1988) revisaram o gênero Chilodus, Vari e Ortega (1997) descreveram Chilodus fritillus e Scharcansky e Lucena (2007) descrevem Caronotropus schizodon e comentaram sobre as relações interespecíficas em Caenotropus. Melo et al. (2014, Anexo 1) apresentaram a primeira filogenia molecular de Chilodontidae e corroboraram a hipótese de monofiletismo de cada um dos gêneros (Vari et al. 1995), mas encontraram evidências que suportam a hipótese da presença de espécies crípticas dentro do clado Chilodus punctatus sensu lato e ainda a necessidade de revisões taxonômicas dentro da família. Figura 6. Relações filogenéticas em Chilodontidae com dados morfológicos (Vari et al. 1995). Introdução*–*Doutorado*–*Bruno*F.*Melo* ! 24! JUSTIFICATIVA A utilização de caracteres moleculares nas análises filogenéticas em peixes tem crescido consideravelmente nas últimas décadas, desde os primeiros trabalhos, como o de Bermingham e Avise (1986), até aqueles envolvendo peixes neotropicais (Chiachio et al. 2008; López- Fernández et al. 2010; Lundberg et al. 2011; Arce H et al. 2013; Roxo et al. 2014). Mais especificamente, diversos estudos vêm sendo publicados com diferentes famílias de Characiformes (Calcagnotto et al. 2005; Javonillo et al. 2010; Arroyave e Stiassny 2011; Oliveira et al. 2011; Tagliacollo et al. 2012; Abe et al. 2013, 2014; Arroyave et al. 2013; Mariguela et al. 2013a, 2013b; Thompson et al. 2014; Melo et al. 2014). Esses trabalhos têm contribuído significativamente para uma melhor compreensão das relações entre grupos complexos ainda não esclarecidos. Contudo, não existem estudos em sistemática molecular abordando as relações internas de nenhuma das famílias contidas em Anostomoidea, exceto o trabalho recentemente publicado da família Chilodontidae (Melo et al. 2014; Anexo 1). Adicionalmente, as relações filogenéticas entre as famílias de Anostomoidea propostas por Vari (1983) ainda não foram testadas utilizando uma ampla amostragem, como já abordado nesta introdução. Os únicos estudos que estudaram as relações nos Anostomoidea foram aqueles realizados com grupos abrangentes, suprafamiliares (Ortí e Meyer 1997; Buckup 1998; Calcagnotto et al. 2005; Moreira 2007; Oliveira et al. 2011) ou intrafamiliares (Vari 1989a; Vari et al. 1995; Castro e Vari 2004; Sidlauskas e Vari 2008; Melo et al. 2014). Após o estudo de Vari (1983), nenhum outro objetivou testar as relações filogenéticas entre as famílias de Anostomoidea e de seus grupos relacionados. A reconstrução filogenética é fundamental para o entendimento das relações evolutivas entre as famílias que compõe Anostomoidea, e consequentemente entre os gêneros de cada uma delas. A incongruência entre os resultados propostos, associada a uma pequena representatividade da diversidade de espécies da superfamília nestes estudos, evidencia a importância de um estudo aprofundado utilizando uma grande amostragem. Nesse ponto, análises filogenéticas utilizando dados moleculares se tornam muito importantes para um melhor entendimento da diversidade e relações na superfamília Anostomoidea. Objetivos*–*Doutorado*–*Bruno*F.*Melo* ! 25! OBJETIVOS Baseando-se nas hipóteses prévias de relacionamento entre membros da superfamília Anostomoidea, e desta com outras famílias, o presente estudo teve como objetivos: 1) Testar as hipóteses de relações filogenéticas entre as famílias de Anostomoidea com base em caracteres moleculares; 2) Testar as hipóteses de relações filogenéticas entre os gêneros e espécies de cada uma das famílias de Anostomoidea com base em caracteres moleculares; 3) Testar novas metodologias filogenômicas e novos métodos de análise filogenética e de calibração de relógio molecular na superfamília Anostomoidea; 4) Desenvolver novas filogenias intrafamiliares mais bem resolvidas para servir de base para futuros estudos taxonômicos e identificar clados problemáticos que requerem análises mais aprofundadas; 5) Gerar filogenias consistentes que poderão ser usadas em futuros estudos evolutivos. Doutorado(–(Bruno(F.(Melo( ! 26! MATERIAL E MÉTODOS, RESULTADOS E DISCUSSÃO As sessões de Material e Métodos, Resultados e Discussão da presente tese estão contidos em quatro manuscritos, que seguem os padrões atuais de publicação científica. Os manuscritos estão organizados da seguinte forma: 1) Phylogenomic analysis reveals multiple episodes of diversification and convergence within the radiation of South American anostomoid fishes (Ostariophysi: Characiformes) 2) Miocene radiations in the family Prochilodontidae (Ostariophysi: Characiformes) revealed by a time-calibrated molecular phylogeny 3) The Neotropical fish family Curimatidae (Teleostei: Ostariophysi: Characiformes): classification and phylogenetic relationships based on molecular data ! 4) Molecular phylogenetics of the family Anostomidae (Teleostei: Characiformes) Além deles, apresentamos em dois Anexos, o artigo publicado sobre a filogenia molecular de Chilodontidae, e o artigo de descrição de uma nova espécie de Cyphocharax, sendo parte dos resultados desta tese. Doutorado(–(Bruno(F.(Melo( ! 27! Capítulo)1)–)Doutorado)–)Bruno)F.)Melo) ! ! 28! Phylogenomic analysis reveals multiple episodes of diversification and convergence within the radiation of South American anostomoid fishes (Ostariophysi: Characiformes) Kendra Hoekzema1, Bruno F. Melo2, Michael E. Alfaro3, Michael D. Burns1, Brant C. Faircloth3,4, Benjamin W. Frable1, Claudio Oliveira2, Richard P. Vari5, and Brian L. Sidlauskas1,5* 1Department of Fisheries and Wildlife, Oregon State University, Corvallis, OR, USA. 2Departamento de Morfologia, Instituto de Biociências, Universidade Estadual Paulista, Botucatu, SP, Brazil. 3Department of Ecology and Evolutionary Biology, University of California Los Angeles, Los Angeles, CA, USA. 4Department of Biological Sciences, Louisiana State University, Baton Rouge, LA, USA. 5Department of Vertebrate Zoology, National Museum of Natural History, Smithsonian Institution, Washington DC, USA. ! *Corresponding author: B. L. Sidlauskas (brian.sidlauskas@oregonstate.edu) Target journal: Current Biology Running title: Phylogenomics of Anostomoid Fishes Highlights: • First phylogenomic study for Anostomoidea, a major clade of Neotropical fishes • Novel ostariophysan probe set capable of capturing 1500 ultraconserved loci • Remarkable morphological convergences in gill arches and oral jaws • The ecomorphologically diverse Anostomidae may represent a rare riverine adaptive radiation. Keywords: Headstanders, Anostomoidea, Neotropics, Convergent Evolution, Next-generation Sequencing, UCEs, Osteology, Tempo and Mode, Adaptive Radiation Capítulo)1)–)Doutorado)–)Bruno)F.)Melo) ! ! 29! Abstract Background. Using a phylogenomic approach spanning approximately 1000 ultraconserved elements in a new ostariophysan-specific array, we present the first genus-level molecular phylogeny for the Neotropical fishes in superfamily Anostomoidea. This clade of approximately 300 species includes three primarily or exclusively detrivorous families, plus Anostomidae, an ecomorphologically diverse and species-rich family of omnivores, insectivores and herbivores. Together these four families are among the most important components of South American fish diversity. Results. Phylogenomics confirm the monophyly of all four anostomoid families, and return a topology that is largely congruent with an earlier osteological dataset spanning nearly 500 characters. However, novel placements of family Chilodontidae and subfamily Anostominae imply two instances of remarkable morphological convergence, one related to gill arch morphology and the other involving repeated evolution of upturned oral jaws. Anostomidae also has a much younger crown group than does the detritivorous clade, implying unexpectedly swift evolution of species and morphologies within the family of greatest ecomorphological diversity. Conclusions. Phylogenomics reveal several surprising convergences in this diverse clade of fishes, and contradict previous macroevolutionary hypotheses that assumed roughly equal ages of the two main subclades. Given its young age and quick genesis of morphological and species diversity, family Anostomidae may represent a rare example of a riverine adaptive radiation. Fully testing this hypothesis will require linking morphological and ecological data to a species- level phylogeny. Given the high support for relationships at deep and shallow nodes obtained in this study, UCE-based phylogenomics will provide an excellent method for generating such a densely-sampled phylogeny in the future. Capítulo)1)–)Doutorado)–)Bruno)F.)Melo) ! ! 30! Introduction Systematic biology aims to reconstruct the evolutionary relationships among organisms in order to understand the genesis of biodiversity. Recent technological improvements have catalyzed remarkable progress in this field by permitting researchers to sequence genome-level data, harness increasingly powerful computers, and communicate effortless across the globe (Chen and Mayden 2010; Faircloth et al. 2012; Lemmon et al. 2012). With these new phylogenomic tools, systematics have forged new collaborations between biologists and informaticians to assemble unprecedentedly large and well-supported phylogenies for groups of organisms ranging from plants (Timme et al. 2012) to mammals (Meredith et al. 2011; McCormack et al. 2012). However, phylogenomic methods have not yet been applied to many clades, including some of the most prevalent evolutionary radiations in the hyperdiverse tropics. By targeting such understudied areas and organisms, we move towards a more complete understanding of the distribution of diversity across the global tree-of-life. The approximately 300 fish species of superfamily Anostomoidea represent one of the most diverse, ubiquitous and important evolutionary radiations within the Neotropical ichthyofauna, which is the richest fish fauna in the world (Albert and Reis 2011). Their ranks include large migratory prochilodontid species that support commercial fisheries, the popular headstanders of the aquarium trade, the brightly colored and ecomorphologically diverse members of family Anostomidae, and numerous cryptic species of detritivorous curimatids (Fig. 1). Perhaps because of their exceptional ecological and morphological diversity, anostomoids have been frequently treated in systematic (Winterbottom 1980; Vari 1983, 1989a; Vari et al. 1995; Castro and Vari 2004; Sidlauskas and Vari 2008; Melo et al. 2014), taxonomic (Britski and Garavello 1993; Garavello 2000; Mautari and Menezes 2006; Birindelli et al. 2013), evolutionary (Sidlauskas 2007, 2008) and biogeographic (Vari 1988; Sidlauskas and Vari 2012) studies, and they rank among the most intensively-studied South American fishes. Each of the four anostomoid families has a highly distinct head anatomy (Vari 1983), and the approximately 150 known species of Anostomidae have diversified into a wide range of mouth positions (Winterbottom 1980; Sidlauskas 2007), tooth configurations (Géry 1961, 1977; Sidlauskas and Vari 2008) and diets (Santos and Rosa 1988; Balassa et al. 2004), while species within the other three families vary much less in head shape and trophic position (Sidlauskas 2007, 2008). Capítulo)1)–)Doutorado)–)Bruno)F.)Melo) ! ! 31! Figure 1. Diversity of anostomoid fishes. 1. Leporellus vittatus; 2. Anostomoides laticeps; 3. Schizodon fasciatus; 4. Schizodon vittatus; 5. Laemolyta proxima; 6. Ichthyoelephas longirostris; 7. Prochilodus hartii; 8. Prochilodus lineatus; 9. Prochilodus nigricans; 10. Anostomus brevior; 11. Anostomus anostomus; 12. Anostomus ternetzi; 13. Pseudanos irinae; 14. Synaptolaemus latofasciatus; 15. Petulanos plicatus; 16. Prochilodus rubrotaeniatus; 17. Prochilodus vimboides; 18. Semaprochilodus laticeps; 19. Semaprochilodus varii; 20. Caenotropus maculosus; 21. Caenotropus mestomorgmatos; 22. Caenotropus labyrinthicus; 23. Caenotropus schizodon; 24. Curimatopsis crypticus; 25. Curimata vittata; 26. Curimata cyprinoides; 27. Curimatella alburna; 28. Psectrogaster ciliata; 29. Curimata roseni; 30. Steindachnerina planiventris; 31. Chilodus gracilis; 32. Chilodus punctatus; 33. Chilodus fritillus; 34. Chilodus zunevei; 35. Curimatopsis crypticus; 36. Cyphocharax spilurus; 37. Cyphocharax helleri; 38. Curimatella immaculata; 39. Psectrogaster essequibensis; 40. Steindachnerina guentheri; 41. Steindachnerina varii. Scale bars indicate one centimeter. Photographs of Anostomidae (1-5; 10-15) and Curimatidae (24-30; 35-41) are from Sidlauskas and Vari (2012), and photographs of Chilodontidae (20-23; 31-34) are from Melo et al. (2014); photographs of Prochilodontidae (6-9; 16-19) will be published by Melo et al. (in prep., Cap. 2). Capítulo)1)–)Doutorado)–)Bruno)F.)Melo) ! ! 32! The skeletal and muscular diversity that characterize all four families has been described and analyzed in exceptional detail, and a family-level phylogeny originally inferred from 104 morphological characters (Vari 1983) and expanded to nearly 500 characters (Vari 1982, 1984, 1989a, 1989b, 1989c, 1989d, 1991; Vari et al. 1995; Castro and Vari 2004; Sidlauskas and Vari 2008) has stood as canon for nearly forty years. In turn, that phylogeny provided the conceptual framework for studies suggesting that the remarkable morphological diversity in a clade containing Anostomidae and Chilodontidae resulted from a lack of constraint and a greater tendency towards innovation, rather than a faster rate of change that in the two exclusively detritivorous families, Curimatidae and Prochilodontidae (Sidlauskas 2007, 2008). Given the taxonomic richness, ecological diversity and evolutionary interest of these fishes, it is remarkable that no one has ever proposed a comprehensive molecular phylogeny for them. A few multilocus molecular studies at higher (Calcagnotto et al. 2005; Oliveira et al. 2011) or lower (Melo et al. 2014) taxonomic levels have included a handful of anostomoid taxa, and several of these have obtained novel relationships among the four anostomoid families by placing Chilodontidae, the smallest of the four, as sister to Prochilodontidae (Calcagnotto et al. 2005) or Curimatidae (Melo et al. 2014) or in a polytomy with those families (Oliveira et al. 2011), rather than as sister to Anostomidae, which is the preferred placement based on morphological (Roberts 1969; Vari 1983; Buckup 1998; Sidlauskas 2008; Sidlauskas and Vari 2008) and mitochondrial ribosomal data (Ortí and Meyer 1997) (Fig. 2). Such a novel placement would challenge our understanding of morphological evolution by implying numerous convergences or reversals in body systems ranging from the bones of the upper gill arches and the pharyngeal dentition to the intercostal ligaments connecting the ribs (Vari et al. 1995). However, because previous molecular studies (Ortí and Meyer 1997; Calcagnotto et al. 2005; Oliveira et al. 2011; Melo et al. 2014) have sampled extremely few taxa and employed no more than six loci, they cannot support a definitive revision of relationships. Formal tests of convergence and congruence between molecules and morphology demand the same rigor in genetic analysis that has already been achieved for morphology. Capítulo)1)–)Doutorado)–)Bruno)F.)Melo) ! ! 33! Next-generation sequencing and phylogenomics allow for unprecedented resolution of relationships within many groups across the Tree of Life (Faircloth et al. 2012; Lemmon et al. 2012; Lemmon and Lemmon 2013). One such method uses ultraconserved elements (UCEs) (Bejerano et al. 2004; Stephen et al. 2008; Faircloth et al. 2012), which can be found across the genome and are, as the name implies, highly conserved across a broad taxonomic range (Stephen et al. 2008; Faircloth et al. 2012). These elements are flanked by variable nuclear regions, which can be enriched, sequenced, and analyzed using phylogenetic methods (Faircloth et al. 2012). Ultraconserved elements datasets dwarf traditional Sanger matrices, and have strongly supported phylogenetic reconstructions for insects (Faircloth et al. 2015), fishes (Faircloth et al. 2013), reptiles/birds (Crawford et al. 2012; McCormack et al. 2013; Smith et al. 2014), and mammals (McCormack et al. 2012), even for frequently debated or difficult questions such as the phylogenetic placement of turtles (Crawford et al. 2012), the relationship between parrots and passerines (McCormack et al. 2013) or the monophyly of holostean fishes (Faircloth et al. 2013). Here, we apply the UCE approach to the superfamily Anostomoidea with a novel ostariophysan-specific set of sequence capture probes targeting more than 1000 UCE regions. Figure 2. Prior phylogenies involving Anostomoidea based on morphological A (Vari 1983; Sidlauskas 2008) and molecular data B (Ortí and Meyer 1997), C (Calcagnotto et al. 2005), D (Oliveira et al. 2011), E (Melo et al. 2014). Capítulo)1)–)Doutorado)–)Bruno)F.)Melo) ! ! 34! The phylogenomic perspective provided by these data permits the first rigorous molecular test of the morphological hypothesis of relationships of these fishes, and reveals the temporal context and progression of this fascinating evolutionary radiation. While confirming substantial congruence between the two datasets, the UCE data also provide strong support for several novel groupings of taxa, illustrate at least two remarkable instances of convergence, reveal an unexpectedly young crown radiation of the ecomorphologically diverse family Anostomidae, and confirm the power of this new approach to phylogenetics. Methods Taxon sampling We sampled 69 ingroup species (43 specimens from Anostomidae, five from Chilodontidae, 27 from Curimatidae, and six from Prochilodontidae) and six representants of five characiform families (see Table S1 for museum catalog numbers). Taxa were chosen to achieve the widest possible coverage at the level of genera given the available tissues. Taxonomic identifications are based on physical or photographic examination of voucher specimens (Fig. S2). Target enrichment and assembly We extracted DNA using a DNeasy Tissue kit (Qiagen Inc.) following the instructions of the manufacturers or a modified NaCl extraction protocol (Lopera-Barrero et al. 2008). We prepared sequencing libraries from purified and sheared DNA using Illumina library preparation kits (Kapa Biosystems), following Faircloth et al. (2013). Libraries were PCR amplified (14 cycles) and brought to 147 ng/µL using a Speed-Vac before target enrichment. We pooled libraries to a total of 500ng of DNA per group and enriched the pooled libraries using a probe set designed for ostariophysian fishes (‘MYBaits’; MYcroarray, Inc.). Identification of UCEs and design of the silico probes is detailed in Faircloth and Alfaro (in prep.). We quantified enriched, indexed libraries using qPCR (Kapa Biosystems), and we sequenced equimolar pools of enriched, indexed libraries using one lane of paired-end, 100 bp sequencing on an Illumina HiSeq 2000 in the Center for Genome Research and Biocomputing at Oregon State University. After sequencing, we trimmed adapter contamination, low quality bases, and sequences containing ambiguous base calls using the Illumiprocessor workflow (available at Capítulo)1)–)Doutorado)–)Bruno)F.)Melo) ! ! 35! https://github.com/faircloth-lab/illumiprocessor), and assembled reads into contigs using Trinity (https://github.com/faircloth-lab/phyluce/blob/master/bin/assembly/assemblo_trinity.py). Following assembly, we used PHYLUCE (https://github.com/faircloth-lab/phyluce) and a custom Python program (match_contigs_to_probes.py) integrating LASTZ (Harris 2007) to align contigs to the set of probes/UCEs we used for enrichment while removing reciprocal and non- reciprocal duplicate hits from the data set. We used additional components of PHYLUCE to query the database, generate fasta files for the UCE loci, align contigs with MAFFT (Katoh et al. 2005) and trim contigs representing UCEs, in parallel, across the selected taxa prior to phylogenetic analysis following the procedure described in Faircloth et al. (2012). Phylogenetic analysis For Bayesian analysis, we used a custom script (run_mraic.py) wrapping a modified MrAIC 1.4.4 (Nylander 2004) to find the best-fitting, finite-sites substitution model for each UCE locus. We grouped loci having similar substitution models (selected by AICc) into the same partition, and we assigned the partition specific substitution model to all loci concatenated within each partition. For maximum likelihood analyses, we maintained the partitions identified in the Bayesian analysis and we modeled each partition using GTR+GAMMA. We performed Bayesian analysis of the concatenated data set using MrBayes 3.1 (Ronquist and Huelsenbeck 2003) and two independent runs (4 chains each) of 5,000,000 iterations each, sampling trees every 500 iterations, to yield a total of 10,000 trees. We sampled the last 5,000 trees after checking results for convergence by visualizing the log of posterior probability within and between the independent runs for each analysis, ensuring the average standard deviation of split frequencies was 0.001, and ensuring the potential scale reduction factor for estimated parameters was approximately 1.0. We performed maximum likelihood analysis of the concatenated data in RAxML (Stamatakis 2006) using the standard bootstrapping algorithm and 500 bootstrap replicates. We time-calibrated the maximum likelihood phylogeny using a penalized likelihood approach (Sanderson 2002; Paradis 2013) as implemented in Chronos (Paradis 2013) in the R (R Development Core Team 2013) package ‘ape’ (Paradis et al. 2004). Four fossils provided calibrations (Supplementary material 3) and we chose a moderate value of lambda (R Development Core Team 2013). Capítulo)1)–)Doutorado)–)Bruno)F.)Melo) ! ! 36! Results Following enrichment and sequencing, we obtained an average of 5,347,980 reads per species, which we assembled into an average of contigs having an average length of 622,376 bp (Table 1). After removing contigs that matched no UCEs and UCE loci that matched multiple contigs, we enriched unique contigs matching UCE loci from each species. This dataset was filtered to 75% completeness, which resulted in 1,077 UCE loci, totaling 469,960 bp with an average of 436 bp per contig; the complete dataset had only 68 UCE loci and 40,730 bp. Table 1. Summary values describing the analysis of the UCE data. Maximum likelihood analysis produced a single, completely resolved topology (Fig. 3) with slight support differences for Bayesian analysis (Fig. S4). 75% complete trees are shown as Figs. S4–S5. Before focusing on the novel arrangements implied by the phylogenomic results, it is important to acknowledge that much of this molecular topology agrees with earlier hypotheses Incomplete dataset 75% complete dataset Alignment summary [Alignment] number of loci 2302 loci 1077 loci [Alignment] total length 774,792 bp 469,960 bp [Alignment] mean bp per locus 336.57 bp 436.36 bp [Alignment] min bp per locus 109 bp 147 bp [Alignment] max bp per locus 1996 bp 1996 bp Taxon summary [Taxa] mean 53.23 taxa 80.17 taxa [Taxa] min 3 taxa 65 taxa [Taxa] max 87 taxa 87 taxa Missing data from trim summary [Missing data] mean 6.19% 5.31% [Missing data] min 0% 0.09% [Missing data] max 26.57% 15.38% Character count summary [Characters] total number 47,467,378 bp 38,241,583 bp [Nucleotides] total number 41,006,322 bp 32,739,348 bp Data matrix completeness summary Matrix 50% 1437 alignments 1077 alignments Matrix 55% 1370 alignments 1077 alignments Matrix 60% 1298 alignments 1077 alignments Matrix 65% 1235 alignments 1077 alignments Matrix 70% 1164 alignments 1077 alignments Matrix 75% 1077 alignments 1077 alignments Matrix 80% 1013 alignments 1013 alignments Matrix 85% 930 alignments 930 alignments Matrix 90% 789 alignments 789 alignments Matrix 95% 591 alignments 591 alignments Capítulo)1)–)Doutorado)–)Bruno)F.)Melo) ! ! 37! based upon morphology (Fig. 4). These results confirm the monophyly of all four families and of all genera not previously suspected to be paraphyletic on the basis of morphological analysis (e.g. Leporinus and Cyphocharax), and provide unprecedented resolution within the latter two genera, both of which presented insoluble challenges in earlier morphological studies. Affinities of many genera are exactly as expected, including the close relationships between Anostomoides, Laemolyta, Rhytiodus and Schizodon, the presence of a clade containing Curimatella, Cyphocharax and Steindachnerina, the recovery of the subfamily Anostominae sensu Winterbottom (1980) (Anostomus, Pseudanos and Synaptolaemus), and the close affinity of Leporinus striatus and Abramites hypselonotus, which in turn allows a very precise time calibration using the Miocene fossil †Leporinus scalabrinii, which shares their morphological synapomorphies. Within the two largest families, the position of Leporellus and Curimatopsis as sister to the majority of other genera is upheld. Overall, far more congruence than discord exists between the molecular and morphological datasets (Fig. 4). Of the several disagreements, the most remarkable is the novel placement of the small family Chilodontidae within a clade also containing Curimatidae and Prochilodontidae. It is most likely sister to Curimatidae, but that grouping is subtended by an exceptionally small internode, and these three families essentially form a polytomy. Nevertheless, the eight chilodontid species have long been considered as sister to Anostomidae, and their placement herein implies convergence in either the substantial set of characters that have traditionally united those two families, or those that have united Curimatidae and Prochilodontidae. The grouping of the anostomine genera (Anostomus, Pseudanos and Synaptolaemus) with Leporellus, and not Laemolyta, is similarly unexpected. With the exception of Leporellus, all of these genera possess strongly upturned jaws and highly multicuspidate premaxillary teeth, and were reconstructed in a single clade in prior morphological analysis (Sidlauskas and Vari 2008). Their wide separation herein implies that anostomids have evolved apparently equivalent morphologies at least twice, and that morphological divergence within the clade containing Leporellus is particularly noteworthy, given that Leporellus possesses a downturned mouth entirely unlikely that of the upturned anostomines. Other conflicting relationships include the placement of the trans-Andean clade Pseudocurimata with Psectrogaster instead of as sister to the clade containing Cyphocharax, Curimatella and Steindachnerina, and the placement of Capítulo)1)–)Doutorado)–)Bruno)F.)Melo) ! ! 38! Ichthyoelephas, a genus with immensely hypertrophied lips, as sister to all other members of Prochilodontidae, rather than as sister to Semaprochilodus alone. Time calibration revealed unexpectedly major differences in the age of the crown clades of the two halves of the anostomoid radiation (Fig. 5). The three detritivorous families split from each other around 90 million years ago, while diversification within the living Anostomidae dates back only half that long. Given the roughly equal species richnesses of the two major subclades, Anostomidae must have experienced a much faster net rate of cladogenesis. The extremely short internodes reconstructed within the paraphyletic genus Leporinus further support the inference of accelerated speciation within that portion of the tree. Discussion The novel scenarios proposed by the molecular results involve an exceptional amount of osteological convergence and call earlier macroevolutionary conclusions into question. Yet, they appear with high statistical support for the 1,077 UCE loci analysis, and must be taken seriously. While provocative, these data are based upon only 23% species diversity within Anostomoidea. We propose to evaluate those hypotheses using much greater taxon sampling reaching 80% taxon coverage. Is Anostomidae an Adaptive Radiation? Possibly. The extremely young age and apparently rapid speciation within the most ecomorphologically diverse clade suggests that evolutionary dynamics in this portion of the tree differ remarkably from those in the detritivorous families. A match between rate of speciation and ecomorphological diversification is one of several predictions of adaptive radiation. Examples of adaptive radiation are well know for cichlid fishes in the African Great Lakes (Kocher 2004; Seehausen 2006; Turner 2007; Wagner et al. 2012) and were predicted by both ecological opportunity and sexual selection (Wagner et al. 2012). But, it is not clear whether the long apparent delay before the onset of speciation in Anostomidae is real, or reflects the extinction of early branching species. Rates of morphological and ecological change have not yet been formally measured, and functional implications of the morphological shifts have not been quantified. Do UCEs provide an effective phylogenomic approach? Yes. The phylogenetic analyses showed extremely high support for almost every node in the phylogeny at both deep and shallow levels. Ultraconserved loci have clear ability to reconstruct a tree that makes good biological Capítulo)1)–)Doutorado)–)Bruno)F.)Melo) ! ! 39! sense and largely agrees with a wealth of prior knowledge about morphology. As its prevalence and application grows, phylogenomics promises to systematists a new age of biodiversity discovery (Pennisi 2013). Comparisons between UCE and multilocus topologies. We incorporated preliminary results of the multilocus analysis and present for the first time the multilocus phylogeny of 422 with 13 outgroup taxa of the superfamily Anostomoidea using six loci (Fig. 6). As noted, the resolution among families is exactly the same than the UCE topology (Figs. 3–5) with (Anostomidae (Prochilodontidae (Chilodontidae + Curimatidae). Major discrepancies are the support values in which the UCE methodology shows very well supported clades in all instances of the tree (basal and derived clades) whereas multilocus analyses present some poorly-supported clades mainly at the familial resolution (weak clades of Fig. 6). Overall, the UCE topology obviously reconstructs the most likely molecular phylogeny by having the most robust molecular dataset. Capítulo)1)–)Doutorado)–)Bruno)F.)Melo) ! ! 40! Figure 3. Phylogenetic relationships of Anostomoidea and outgroups based on a maximum likelihood analysis of 1,077 UCE loci. Prochilodontidae Chilodontidae Curimatidae Anostomidae curimata_sp psectrogaster_sp Capítulo)1)–)Doutorado)–)Bruno)F.)Melo) ! ! 41! Figure 4. Bayesian phylogeny generated through ExaBayes evidencing clades in agreement or in conflict with prior morphological hypotheses (Vari 1983; Castro and Vari 2004; Sidlauskas and Vari 2008) and clades that first resolved the unresolved relationships. Prochilodontidae Chilodontidae Curimatidae Anostomidae curimata_sp psectrogaster_sp ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! Agreement Resolves the Unresolved Conflict Capítulo)1)–)Doutorado)–)Bruno)F.)Melo) ! ! 42! Figure 5. Time-calibrated phylogeny using four fossil calibrations (grey circles). Branch widths set proportional to inferred rate of molecular evolution. Branches of the four anostomoid families are colored as in previous figures. Numbers after names designate duplicated specimens. Capítulo)1)–)Doutorado)–)Bruno)F.)Melo) ! ! 43! Figure 6. Best maximum likelihood tree including all analyzed taxa of Anostomoidea using a multilocus dataset showing the resolution (Anostomidae (Prochilodontidae (Chilodontidae + Curimatidae). Branches of the four anostomoid families are colored as in previous figures. Branch weight is determined by bootstrap support. Capítulo)1)–)Doutorado)–)Bruno)F.)Melo) ! ! 44! Author Contributions BLS, RPV, CO and MEA obtained funding and designed the project. BLS, RPV and CO facilitated the international exchange of samples. KH, BFM and MWB collected data. BCF and MEA wrote custom scripts and advised the phylogenomic analysis. KH, BFM, MEA, BCF, MWB and BLS performed analyses. BLS and RPV interpreted the results in the context of morphology. BLS and KH wrote most of the manuscript. BFM and RPV contributed sections of writing. BCF, BLS, BFM and KH prepared the figures. All authors edited and revised the manuscript. Acknowledgements We thank M. Sabaj-Perez (ANSP), N. Lujan and J. Armbruster (AUM), W. Aguirre (DePaul Universtiy), S. Mochel and K. Swagel (FMNH), M. de Pinna, O. Oyakawa, J. Birindelli and M. Gianeti (MZUSP) and R. Reina (STRI) for curatorial assistance and/or loan of tissues and vouchers. We especially thank C. Shimabukuro-Dias and A. Alves for helping with the preparation and shipment of tissues at LBP. Mark Dasenko and the CGRB CoreLabs at Oregon State provided assistance with the Illumina sequencing platform. NSF (DEB-1257898) and FAPESP (proc. 2011/08374-1 and proc. BEPE 2013/16436-2) funded this project. Capítulo)1)–)Doutorado)–)Bruno)F.)Melo) ! ! 45! Table S1. Voucher information for fish species used in this study. Collection acronyms follow Sabaj Pérez (2013). Family Species Voucher Tissue number Anostomidae Abramites hypselonotus ANSP 178126 1712 Abramites hypselonotus AUM 53775 T08985 Anostomoides laticeps AUM 53813 T09941 Anostomoides laticeps FMNH 123875 T45 Anostomus brevior FMNH ex ANSP 189141 6879 Hypomasticus julii MZUSP 97206 7275 Laemolyta garmani OS 18777 PE10-089 Laemolyta proxima OS 18778 PE10-094 Leporellus vittatus ANSP 182609 P6322 Leporinus agassizi OS 18369 PE10-073 Leporinus agassizi OS 18775 PE10-090 Leporinus cf. arimaspi ANSP 191183 BO6133 Leporinus brunneus AUM 54394 T09305 Leporinus cf. trifasciatus LBP 12722 43631 Leporinus copelandii LBP 6449 29075 Leporinus fasciatus ANSP 180321 NS61 Leporinus fasciatus AUM 43300 P4315 Leporinus maculatus FMNH Ex ANSP 189041 7035 Leporinus maculatus ANSP 189041 6982 Leporinus muyscorum ANSP 192864 6606 Leporinus octofasciatus LBP 3939 19827 Leporinus octomaculatus LBP 13249 69389 Leporinus ortomaculatus ANSP 191182 BO6132 Leporinus piau LBP 340 4309 Leporinus piavussu LBP 3304 19848 Leporinus reinhardti LBP 259 4156 Leporinus sp. “Nanay” ANSP 180317 T32 Leporinus steindachneri LBP 8098 37533 Leporinus striatus LBP 3180 16871 Leporinus subniger OS 18726 PE10-022 Leporinus taeniatus LBP 329 4267 Leporinus trifasciatus OS 18311 PE10-108 Pseudanos gracilis FMNH113530 T57 Pseudanos gracilis ANSP 180320 T73 Pseudanos trimaculatus OS 18779 PE10-021 Pseudanos winterbottomi AUM 39855 V120 Pseudanos winterbottomi AUM 39292 V170 Rhytiodus argenteofuscus ANSP 180242 T44 Rhytiodus microlepis ANSP 182603 P6326 Schizodon fasciatus OS 18310 PE10-104 Schizodon intermedius LBP 3305 19845 Schizodon scotorhabdotus AUM 53654 T09707 Synaptolaemus latofasciatus MZUSP 97460 mz7230 Chilodontidae Caenotropus maculosus ANSP 189159 6999 Caenotropus mestomorgmatos OS 18772 PE10-093 Chilodus gracilis LBP 6962 33397 Chilodus gracilis LBP 6962 33398 Chilodus punctatus OS 18781 PE10-100 Curimatidae Curimata acutirostris LBP 5761 28139 Curimata incompta ANSP 191354 BO6125 Curimata roseni ANSP 180389 T38 Curimata roseni LBP 12864 53467 Capítulo)1)–)Doutorado)–)Bruno)F.)Melo) ! ! 46! Curimata vittata LBP 13846 57302 Curimatella alburna LBP 5431 57171 Curimatella meyeri LBP 4253 22748 Curimatopsis crypticus ANSP 189091 7057 Curimatopsis macrolepis OS 18337 PE10-086 Curimatopsis macrolepis OS 18337 PE10-088 Cyphocharax magdalenae STRI-1001 Cyphocharax magdalenae BFD01631 STRI-1145 Cyphocharax plumbeus OS 18373 PE10-071 Cyphocharax spilurus ANSP 189157 6877 Psectrogaster amazonica OS 18313 PE10-111 Psectrogaster ciliata ANSP 191145 BO6120 Psectrogaster rutiloides ANSP178284 1576 Pseudocurimata lineopunctata BFD01628 STRI-1389 Pseudocurimata lineopunctata BFD01630 STRI-881 Pseudocurimata troschelli Univ Guayaquil 1542 PSB-201 Pseudocurimata troschelli Univ Guayaquil 1542 PSB-202 Steindachnerina argentea BFD01625 STRI-4270 Steindachnerina bimaculata LBP 173 4023 Steindachnerina guentheri AUM 54441 T09661 Steindachnerina dobula ANSP 180798 4123 Steindachnerina hypostoma ANSP 178124 1742 Steindachnerina hypostoma ANSP 178124 1736 Steindachnerina leucisca ANSP178125 1744 Steindachnerina planiventris OS 18324 PE10-150 Prochilodontidae Ichthyoelephas longirostris ANSP 192865 6609 Prochilodus mariae ANSP 40692 P4313 Prochilodus nigricans OS 18792 PE10-045 Semaprochilodus insignis ANSP 180205 T43 Semaprochilodus kneri ANSP 187277 P4298 Semaprochilodus varii ANSP 187435 6929 Serrasalmidae Myleus schomburgkii OS 18990 PE10-044 Characidae Moenkhausia oligolepis OS 18747 PE10-162 Hemiodontidae Anodus elongatus LBP 4244 22720 Anodus orinocensis LBP 2210 15614 Parodontidae Apareiodon ibitiensis LBP 2890 18635 Crenuchidae Crenuchus spilurus LBP 10622 24012 Capítulo)1)–)Doutorado)–)Bruno)F.)M elo)! ! 47! Supp. m aterial 2. R epresentatives of the voucher specim ens analyzed in this study: 1. M oenkhausia oligolepis, O S 18747, Peru, R ío N anay, A m azon basin; 2. M yleus schom burgkii, O S uncatalogued, Peru, R ío N anay, A m azon basin; 3. Anodus elongatus, LB P 4244, B razil, R io Juruá, A m azon basin; 4. Apareiodon ibitiensis, LB P 2890, B razil, R io G rande, upper Paraná basin; 5. Laem olyta garm ani, O S 18777, Peru, R ío N anay, A m azon basin; 6. Laem olyta proxim a, O S 18778, Peru, R ío N anay, A m azon basin; 7. Leporinus agassizi, O S 18369, Peru, R ío N anay, A m azon basin; 8. Leporinus agassizi, O S 18775, Peru, R ío N anay, A m azon basin; 9. Leporinus copelandii, LB P 6449, B razil, R io Paraíba do Sul; 10. Leporinus octom aculatus, LB P 13249, B razil, R io Tapajós, A m azon basin; 11. Leporinus piau, LB P 340, B razil, R io São Francisco; 12. Leporinus reinhardti, LB P 259, B razil, R io São Francisco; 13. Leporinus subniger, O S 18726, Peru, R ío N anay, A m azon basin; 14. Leporinus steindachneri, LB P 8098, B razil, R io M ucuri; 15. Leporinus striatus, LB P 3180, B razil, R io Paranapanem a, upper Paraná basin; 16. Pseudanos trim aculatus, O S 18779, Peru, R ío N anay, A m azon basin; 17. Leporinus octofasciatus, LB P 3939, B razil, R io Paranapanem a, upper Paraná basin; 18. Leporinus cf. trifasciatus, LB P 12722, B razil, R io A raguaia, A m azon basin; 19. Leporinus trifasciatus, O S 18311, Peru, Iquitos m arket, A m azon basin; 20. Schizodon interm edius, LB P 3305, B razil, R io Tietê, upper Paraná basin; 21. Schizodon fasciatus, O S 18310, Peru, Iquitos m arket, A m azon basin; 22. C aenotropus m estom orgm atos, O S 18772, Peru, R ío N anay, A m azon basin; 23. C hilodus gracilis, LB P 6962, B razil, R io N egro, A m azon basin; 24. C hilodus gracilis, LB P 6962, B razil, R io N egro, A m azon basin; 25. C hilodus punctatus, O S 18781, Peru, R ío N anay, A m azon basin; 26. C urim ata acutirostris, LB P 5761, B razil, R io A raguaia, A m azon basin; 27. C urim ata knerii, LB P 12864, B razil, R io Tapajós, A m azon basin; 28. C urim ata vittata, LB P 13846, B razil, R io Tapajós, A m azon basin; 29. C urim atopsis m acrolepis, O S 18337, Peru, R ío N anay, A m azon basin; 30. C urim atopsis m acrolepis, O S 18767, Peru, R ío N anay, A m azon basin; 31. C urim atella m eyeri, LB P 4253, B razil, R io Juruá, A m azon basin; 32. Steindachnerina bim aculata, LB P 173, B razil, R io Purus, A m azon basin; 33. C yphocharax plum beus, O S 18373, Peru, R ío N anay, A m azon basin; 34. Psectrogaster am azonica, O S 18313, Peru, Iquitos m arket, A m azon basin; 35. Steindachnerina planiventris, O S18324, Peru, Iquitos m arket, A m azon basin; 36. Prochilodus nigricans, O S 18792, Peru, R ío Itaya, A m azon basin. Capítulo)1)–)Doutorado)–)Bruno)F.)Melo) ! 48! Supplementary material 3 Fossil calibrations used in chronogram construction †Cyphocharax mosesi, UNG-TA 854. Articulated complete skeleton. Late Oligocene or very early Miocene of the Taubaté basin, Tremembé formation, São Paulo, Brazil. Minimally 22.5 mya (Travassos and Santos 1955; Malabarba 1996; Reis 1998) used to calibrate minimum age of node containing all living Cyphocharax species. †Plesiocurimata alvarengai (Pz.DBAV.UERJ.405-A) from the same formation (Figueiredo and Costa-Carvalho 1999) is a member of Curimatidae, but its possession of teeth at adult stages of ontogeny (which no living member of the family does) suggests that it falls outside of the crown group. Thus, †Plesiocurimata alvarengai provides no additional information for the purposes of time- calibration. Leporinus sp. Jaw teeth and pharyngeal teeth (Roberts 1975) remains from Cuenca basin of Ecuador at 19 million years ago. Most likely a Leporinus but possibly representative of Leporellus (Roberts 1975). Leporinus sp. IGM 253295. Symphyseal premaxillary tooth from La Venta formation of Colombia (Lundberg 1997), between 13.5 and 11.5 mya (Guerrero 1997). Used to set the minimum age of the clade containing all living species of Leporinus. †Leporinus scalabrinii. MACN A-9880. Incomplete articulated skull. Late Miocene deposits of Entre Ríos, Argentina between 6 and 9 mya (Bogan et al. 2012). Several osteological synapomorphies support the placement of this fossil within a clade containing Leporinus striatus and Abramites hypselonotus (Bogan et al. 2012). These two species appear in close proximity in the UCE phylogeny, and the fossil was used to set a minimum age for the node representing their common ancestor. Capítulo)1)–)Doutorado)–)Bruno)F.)Melo) ! 49! Fig. S4. Phylogenetic relationships of Anostomoidea and outgroups based on a Bayesian analysis of the 75% complete matrix. Numbers on branches represent posterior probabilities. Capítulo)1)–)Doutorado)–)Bruno)F.)Melo) ! 50! Fig. S5. Phylogenetic relationships of Anostomoidea and outgroups based on a maximum likelihood analysis of the 75% complete matrix. Numbers on branches represent percentages of bootstrap support. Capítulo)2)–)Doutorado)–)Bruno)F.)Melo) ! 51! Capítulo)2)–)Doutorado)–)Bruno)F.)Melo) ! 52! Miocene radiations in the family Prochilodontidae (Ostariophysi: Characiformes) revealed by a time calibrated molecular phylogeny Bruno F. Melo1*, Kendra Hoekzema2, Benjamin W. Frable2, Richard P. Vari3, Brian L. Sidlauskas2,3, Claudio Oliveira1 1Departamento de Morfologia, Instituto de Biociências, Universidade Estadual Paulista, Botucatu SP, Brazil 2Department of Fisheries and Wildlife, Oregon State University, Corvallis OR, USA. 3Department of Vertebrate Zoology, National Museum of Natural History, Smithsonian Institution, Washington DC, USA. * Corresponding author: B. F. Melo (melo@ibb.unesp.br). Target journal: BMC Evolutionary Biology Running title: Molecular phylogeny of the characiform family Prochilodontidae Highlights: • First molecular phylogeny for Prochilodontidae. • Ichthyoelephas sister to Prochilodus and Semaprochilodus. • Most relationships corroborate previous morphological hypothesis and incongruences are discussed. • Divergence between Ichthyoelephas and other genera in the Late Cretaceous and internal radiations during the Miocene. Keywords: Diversification, evolution, multilocus, phylogeny, Neotropics. Capítulo)2)–)Doutorado)–)Bruno)F.)Melo) ! 53! Abstract The large migratory detritivores in the characiform family Prochilodontidae represent one of the most commercially important fish groups in South America, but have never been the subject of a comprehensive molecular phylogenetic analysis. With three mitochondrial and three nuclear loci from 90% of the recognized prochilodontid taxa, we generated a new phylogenetic hypothesis, time-calibrated it using three fossils, and reconstructed the evolutionary history of the family. Our molecular phylogeny corroborated most previous morphological hypotheses of relationships and resolved some longstanding polytomies. Several morphologically similar species are embedded within the current concept of Prochilodus nigricans, and this group clearly merits taxonomic revision. The timetree shows that the split of the trans-Andean genus Ichthyoelephas from other prochilodontids occurring in the Late Cretaceous at approximately 80 million years ago (mya). Most internal radiations was recovered to be occurred during the Miocene since 30 mya and that the diversification within Prochilodus is much more recent that those within Semaprochilodus. Capítulo)2)–)Doutorado)–)Bruno)F.)Melo) ! 54! Introduction Over 2,000 species from Africa, Central America and South America, including piranhas, tetras, and the impressive tigerfish, form the megadiverse ostariophysian order Characiformes. In the last decade, several molecular phylogenetic analyses (Ortí et al. 2008; Arroyave and Stiassny 2011; Oliveira et al. 2011; Tagliacollo et al. 2012; Abe et al. 2013, 2014; Arroyave et al. 2013; Mariguela et al. 2013; Melo et al. 2014; Thompson et al. 2014) have revisited inter- and intrafamilial relationships within this order that had previously received only morphological attention. Molecular and morphological datasets have agreed on numerous phylogenetic questions in Characiformes, for instance the monophyly of superfamily Anostomoidea, which includes the families Anostomidae, Chilodontidae, Curimatidae, and Prochilodontidae (Vari 1983; Sidlauskas 2008; Oliveira et al. 2011). In other instances, molecular and morphological hypotheses have diverged, as in the case of the African family Alestidae. Morphological studies (Zanata and Vari 2005; Mirande 2009, 2010) placed Alestidae sister to the South American genus Chalceus whereas molecular studies (Calcagnotto et al. 2005; Arroyave and Stiassny 2011; Oliveira et al. 2011) recovered Alestidae sister to the African Hepsetidae with Chalceus phylogenetically distant, as a newly proposed family Chalceidae (Oliveira et al. 2011). These later authors (Oliveira et al. 2011) also rejected the monophyly of Cynodontidae as proposed by Lucena and Menezes (1998), but corroborated Cynodontidae as proposed by Toledo-Piza (2000). Notwithstanding these advances, the position of many characiform families (e.g. Crenuchidae, Ctenoluciidae, Lebiasinidae, Serrasalmidae) remains uncertain. These discordances and uncertainties demonstrate the need to reexamine established characiform families via robust molecular datasets in order to understand the evolution of this major teleost lineage. The characiform superfamily Anostomoidea, containing four trophically diverse families and about 300 species, has received extensive systematic and revisionary treatment (Vari 1983; Sidlauskas 2008) and its monophyly seems certain. However, its internal familial relationships have been explored in detail only with morphology (Vari 1992b, 1982, 1984, 1989a, 1989b, 1989c, 1989d, 1991, 1992a; Vari et al. 1995; Castro and Vari 2004; Sidlauskas and Vari 2008), and merit reexamination with molecular data. A recent study of the eight described species in Chilodontidae (Melo et al. 2014), the smallest of the four anostomoid families, reinforced the monophyly of the two genera but also revealed several cryptic species, suggested potential necessary synonymies, and differed from the established morphological hypothesis in the Capítulo)2)–)Doutorado)–)Bruno)F.)Melo) ! 55! placement of one species (Vari et al. 1995). That analysis further demonstrates the utility of molecular data in taxonomic and evolutionary studies within characiforms, and suggests that comparable studies of the three more species-rich families in the Anostomoidea (Anostomidae, Curimatidae and Prochilodontidae) will yield similar insights and surprises. Herein, we generate the first comprehensive molecular phylogeny of the detritivorous family Prochilodontidae, commonly known as curimbatás and jaraquis in Portuguese, bocachicos in Spanish, and flannel-mouth characiforms in English (Castro and Vari 2003). The family includes three genera: the remarkable, trans-Andean Ichthyoelephas with a unique, highly developed upper lip (2 species, Fig. 1a), and two abundant, migratory, and commercially important genera, Prochilodus (13 species, Fig. 1b–f) and Semaprochilodus (6 species, 1g–h). Prochilodus is present in all major river systems of South America on both sides of the Andes while Semaprochilodus is broadly distributed east of the Andes, throughout the Amazon and Orinoco basins and some coastal rivers of the Guiana Shield. Figure 1. Live representative specimens of Prochilodontidae. (A) Ichthyoelephas longirostris, ANSP 192865, Colombia, Río Magdalena basin; (B) Prochilodus hartii, Brazil, Rio Jequitinhonha; (C) P. lineatus, Brazil, Rio Paraguai basin; (D) P. nigricans, Brazil, Rio Xingu, Amazon basin; (E) P. rubrotaeniatus, Brazil, Rio Branco, Amazon basin (F) P. vimboides, Brazil, Rio Mucuri; (G) Semaprochilodus laticeps (right after dead), MCNG 51963, Venezuela, Río Orinoco basin; (H) S. varii, ANSP 187435, Suriname, Marowijne river basin. Scale bars represent one centimeter. Photos by M. Sabaj Perez (A, D, G and H), T. C. Pessali (B), J. Sabino (C), B. F. Melo (E) and L. M. Sarmento- Soares (F). Capítulo)2)–)Doutorado)–)Bruno)F.)Melo) ! 56! Prochilodontidae is readily distinguished externally from all other characiforms by having numerous rows of relatively small teeth attached to the fleshy lips and structurally unique upper and lower jaws that support the expanded lips. These lips form an evertable (extendable outwards) oral disk (Castro and Vari 2004) which, in combination with the multiple tooth rows, serves to scrape periphyton and detritus off submerged surfaces and to suck detritus off the bottom of water bodies (Fig. 1c) (Bowen 1983; Bowen et al. 1984). Prochilodontids feed towards the base of the food chain and form a major link in the carbon flow within aquatic systems in the Neotropics (Taylor et al. 2006). Their specialization on a hyperabundant energy resource may account for the large bodies and enormous population sizes (Carolsfeld et al. 2003) of many species, which in turn allows these species to sustain the most important commercial and subsistence fisheries in many regions of South America (Bayley and Petrere Jr. 1989; Ribeiro and Petrere Jr. 1990; Camargo and Petrere Jr. 2001; Garcia et al. 2009). In many settings, prochilodontids additionally act as ecosystem engineers (Flecker 1996), and their removal can significantly impact broader riverine aquatic faunas (Ribeiro and Petrere Jr. 1990; Taylor et al. 2006). They are among the most important food resources for people across South America, albeit with anthropogenically induced declines in some fisheries (e.g. Prochilodus lineatus in the upper Río Paraguai (Mateus et al. 2004)). Aside from their unusual feeding apparatus and incredible biomass, prochilodontids are famous for their large seasonal migrations which can span many hundreds of river kilometers (Paiva and Bastos 1981; Godinho and Kynard 2006). Because of those long migrations, natural populations of some prochilodontid species span enormous geographic ranges. For example, a study of population structure in Prochilodus (Sivasundar et al. 2001) suggested a remarkable intraspecific genetic homogeneity throughout each of the Magdalena, Orinoco, and Amazon basins, with major watershed boundaries structuring their diversity. Recent proposals to dam major portions of the Amazon and its tributaries (e.g. those under construction in the Madeira, Tapajós, and Xingu river systems) lead to serious concerns that these barriers will likely dramatically diminish, if not eliminate, flooded downstream areas and their marginal lagoons considered crucial for genetic diversity in Prochilodus (Melo et al. 2013). These major impoundments will also impede the long distance migrations prevalent in these ecologically important fishes with consequent fragmentation of populations. Such fragmentation may hold Capítulo)2)–)Doutorado)–)Bruno)F.)Melo) ! 57! unknown consequences for the reproduction and life history of one of the most important components of the artisanal and commercial fisheries in South America. Despite the economic importance of Prochilodontidae, the systematics of this group remains incompletely resolved. Vari (1983) proposed monophyly of Prochilodontidae based on 18 synapomorphies related to dentition and cranial osteology, but examined only six species of the family. Later, in an extensive morphological phylogeny and taxonomic revision, Castro and Vari (2004) proposed 40 additional synapomorphies for the family, identified dozens of characters separating the three recognized genera, generated both generic and interspecific phylogenies, and hypothesized Prochilodus as sister to a clade containing Ichthyoelephas and Semaprochilodus. Despite their broad osteological sampling, Castro and Vari (2004) were able to identify only a few informative characters as to the relationships within Prochilodus and Semaprochilodus and only partially resolved the species-level phylogenies within those genera. This situation derives from the remarkable morphological similarity of species within each of these genera, with species differing primarily in scale counts, body morphometrics, nuanced aspects of coloration, and most significantly the watersheds to which they are endemic. Morphological stasis among species in the Prochilodontidae might be expected given their drainage-wide ranges and their low levels of sympatry, which would imply a primarily allopatric speciation mechanism without requisite major morphological evolution during speciation. Although biologically predictable, the lack of informative morphological characters at the species-level presents a significant obstacle in reconstructing evolutionary history based on such data sources. Molecular approaches, thereby, offer the surest path forward. In this study, we combine mitochondrial and nuclear loci to revisit the phylogenetics of Prochilodontidae and advance our understanding of their evolution. We develop a time-scaled phylogeny for the family and outgroups using molecular-clock Bayesian estimation methods. With nearly complete taxon sampling, we test Vari’s (1983) hypothesis of the monophyly of the family, Castro and Vari’s (2004) hypotheses of relationships among genera and species, and present the first fully resolved species-level phylogeny for the family. Material and Methods Taxon sampling Capítulo)2)–)Doutorado)–)Bruno)F.)Melo) ! 58! We sampled 51 terminal taxa covering 19 of 21 known species (90%; only Prochilodus britskii and Ichthyoelephas humeralis missing) and all three genera. The outgroup taxa included species from each of the other three anostomoid families (Anostomidae, Chilodontidae, Curimatidae), representatives of three families (Hemiodontidae, Parodontidae and Serrasalmidae) previously hypothesized to be closely related to Anostomoidea, and one specimen of Brycon pesu, which was used to root the trees. We used tissues preserved in 95% ethanol or a saturated DMSO/NaCl solution, primarily from specimens deposited in museum and university collections (Table 1; abbreviations follow Sabaj Pérez (2013)). The tissue sample of Ichthyoelephas longirostris (ANSP 192865) has only a photographic voucher (Fig. 1a), but the highly distinctive morphology of this species leaves no doubt as to its proper identification. The Fig. 2 represents sampling localities for each analyzed prochilodontid species and was produced with QuantumGIS 10.07.28. Multilocus sequencing We extracted DNA from tissue muscles or fins with a DNeasy Tissue kit (Qiagen Inc.) following manufacturer’s instructions or a modified NaCl extraction protocol adapted from Lopera-Barrero et al. (2008). Following recent phylogenetic studies with characiforms (Oliveira et al. 2011; Abe et al. 2013; Melo et al. 2014), we amplified partial sequences of the mitochondrial genes 16S rRNA (16S, 510 bp), cytochrome oxidase C subunit 1 (COI, 656 bp) and cytochrome B (Cytb, 990 bp) using one round of polymerase chain reaction (PCR). Additionally, we obtained sequences of the nuclear myosin heavy chain 6 gene (Myh6, 710 bp), recombination activating gene 1 (Rag1, 1378 bp), and recombination activating gene 2 (Rag2, 1029 bp) through nested- PCR following Oliveira et al. (2011). Primers used in this study are listed on Table 2. We did 12.5 µl reactions containing 9.075 µl of double-distilled water, 1.25 µl 5x reaction buffer, 0.375 MgCl2, 0.25 µl dNTP mix at 8 mM, 0.25 µl of each primer at 10 µM (list of primers in Table 2), 0.05 µl P