Molecular Phylogenetics and Evolution 56 (2010) 49–63 Contents lists available at ScienceDirect Molecular Phylogenetics and Evolution journal homepage: www.elsevier.com/locate/ympev Pattern and timing of diversification of the mammalian order Carnivora inferred from multiple nuclear gene sequences Eduardo Eizirik a,b,c,*, William J. Murphy d, Klaus-Peter Koepfli e, Warren E. Johnson b, Jerry W. Dragoo f, Robert K. Wayne e, Stephen J. O’Brien b a Faculdade de Biociências, PUCRS, Av. Ipiranga 6681, Porto Alegre, RS 90619-900, Brazil b Laboratory of Genomic Diversity, NCI-Frederick, NIH, Frederick, MD 21702-1201, USA c Instituto Pró-Carnívoros, Atibaia, SP, Brazil d Department of Veterinary Integrative Biosciences, Texas A&M University, College Station, TX 77843-4458, USA e Department of Ecology and Evolutionary Biology, University of California, Los Angeles, CA 90095-1606, USA f Department of Biology, University of New Mexico, Albuquerque, NM 87131-1091, USA article info abstract Article history: The mammalian order Carnivora has attracted the attention of scientists of various disciplines for dec- Received 30 June 2009 ades, leading to intense interest in defining its supra-familial relationships. In the last few years, major Revised 21 January 2010 changes to the topological structure of the carnivoran tree have been proposed and supported by various Accepted 29 January 2010 molecular data sets, radically changing the traditional view of family composition in this order. Although Available online 4 February 2010 a sequence of molecular studies have established a growing consensus with respect to most inter-familial relationships, no analysis so far has included all carnivoran lineages (both feliform and caniform) in an Keywords: integrated data set, so as to determine comparative patterns of diversification. Moreover, no study con- Carnivores ducted thus far has estimated divergence dates among all carnivoran families, which is an important Mammals Nuclear markers requirement in the attempt to understand the patterns and tempo of diversification in this group. In this Supermatrix study, we have investigated the phylogenetic relationships among carnivoran families, and performed Molecular dating molecular dating analyses of the inferred nodes. We assembled a molecular supermatrix containing 14 Divtime genes (7765 bp), most of which have not been previously used in supra-familial carnivoran phylogenet- BEAST ics, for 50 different genera representing all carnivoran families. Analysis of this data set led to consistent and robust resolution of all supra-familial nodes in the carnivoran tree, and allowed the construction of a molecular timescale for the evolution of this mammalian order. Ó 2010 Elsevier Inc. All rights reserved. 1. Introduction There are currently 286 recognized living carnivoran species, classified into 125 different genera (Wilson and Mittermeier, The mammalian order Carnivora exhibits a remarkable diversity 2009) traditionally placed in 11 families (Nowak, 1999; Wozen- of form and function, evolved as adaptations to widely different craft, 1993). Due to their diversity, public and scientific appeal, habitats, ranging from equatorial deserts and forests to temperate and rich fossil record, carnivorans have been historically the sub- mountains and polar marine environments. Carnivorans are very ject of extensive evolutionary studies, including numerous at- widespread geographically, and demonstrate one of the most ex- tempts to resolve phylogenetic relationships among some or all treme cases of size variation among all mammalian orders (from of their lineages (e.g. Bininda-Emonds et al., 1999; Flynn et al., a 45 g weasel to a 3700 kg elephant seal). Members of this group 2000, 2005; Wozencraft, 1989). Phylogenies of the order Carnivora range from charismatic species well known to the general public have been used to make inferences on processes involved in taxon (e.g. cats, dogs, bears) to mysterious organisms about which almost diversification patterns, tempo and mode of character evolution, nothing is known beyond museum materials used in original taxon and conservation-related issues. As many carnivoran species have descriptions (MacDonald, 2001; Nowak, 1999). suffered tremendous anthropogenic impact on their populations and habitats, many of them are endangered or likely to become so in the future (Gittleman et al., 2001). A phylogenetic framework (Bininda-Emonds et al., 1999) has been used to make assessments * Corresponding author. Address: Faculdade de Biociências, PUCRS, Av. Ipiranga of biological and geographic features related to extinction vulner- 6681, Prédio 12, Porto Alegre, RS 90619-900, Brazil. Fax: +55 51 3320 3568. E-mail addresses: [email protected], [email protected] (E. ability, which were proposed to serve as guides in the design of Eizirik). conservation strategies. For such purposes, as well as for other 1055-7903/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.ympev.2010.01.033 50 E. Eizirik et al. / Molecular Phylogenetics and Evolution 56 (2010) 49–63 biological applications, it is critical to assess whether the underly- generating a large, multi-gene data set composed exclusively of ing phylogenies are accurate, and to obtain a stable evolutionary segments from the nuclear genome. Nuclear sequences have been framework for this group. Likewise, insights from the rich carnivo- found to be more informative than mtDNA at different phyloge- ran fossil record can be greatly augmented by synergistic interac- netic levels (e.g. Koepfli and Wayne, 2003; Springer et al., 2001), tion with a well-established evolutionary timescale derived from and have been successfully used to resolve various portions of molecular data. the mammalian phylogeny (e.g. Amrine-Madsen et al., 2003; Eizir- The order Carnivora is divided into two main evolutionary lin- ik et al., 2001, 2004; Johnson et al., 2006; Koepfli et al., 2006, 2007, eages: the suborders Feliformia and Caniformia. The suborder Fel- 2008; Koepfli and Wayne, 2003; Murphy et al., 2001a,b; Sato et al., iformia has traditionally comprised four families (Felidae, 2006; Yu et al., 2004; Janecka et al., 2007). Concatenation of multi- Herpestidae, Hyaenidae and Viverridae), while Caniformia would ple independent segments has been shown to produce an amplifi- contain the terrestrial families Canidae, Mustelidae, Procyonidae cation of the phylogenetic signal, usually leading to well-resolved and Ursidae, along with the marine carnivores (Pinnipedia, which and supported trees (e.g. Rokas et al., 2003; de Queiroz and Gatesy, include the families Otariidae, Odobenidae and Phocidae). Canifor- 2007). In particular, we selected a novel set of genes, most of which mia is further subdivided into the Cynoidea (containing the family have not been used previously in higher-level carnivoran phyloge- Canidae, thought to be the deepest divergence in this group) and netics (e.g. Flynn et al., 2005; Gaubert and Veron, 2003; Yoder Arctoidea (with the six remaining families in this suborder). This et al., 2003; Sato et al., 2004, 2006; Yu et al., 2004), thus providing taxonomic arrangement was first proposed by Flower (1869) on an independent test for many recently proposed supra-familial the basis of the form and structure of the auditory bulla, and has hypotheses. Using this data set and multiple inferential ap- since been consistently supported by numerous phylogenetic stud- proaches, we arrived at a well-resolved phylogeny presenting con- ies employing other anatomical/morphological and molecular gruence among methods and high support for all higher-level characters (e.g. Flynn and Wesley-Hunt, 2005). Despite this consis- nodes. Divergence dating analyses based on this data set produced tency with respect to higher level relationships, extensive contro- an evolutionary timescale of living carnivoran lineages, and led to versy has dominated the evolutionary literature regarding the inferences on historical processes involved in the diversification of phylogenetic relationships among families in each suborder, the this mammalian order. placement of enigmatic taxa (e.g. giant and red pandas, walrus, and the Malagasy fossa [Cryptoprocta]) and even the monophyly 2. Materials and methods of several families. The extensive literature covering these contro- versies will not be described here in detail (see Flynn and Wesley- 2.1. Taxon sampling Hunt (2005) and Eizirik and Murphy (2009) for reviews), and we will focus only on the most recent developments regarding these The major goal of our taxon-sampling scheme was to represent issues. all extant carnivoran families, as well as the most basal divergence Recent challenges to the monophyly of traditional families have within each family (i.e. the base of each crown-group). For that included the following propositions: (i) the African Palm civet purpose, we included divergent genera (one species for each) from (Nandinia binotata) is a basal feliform, and not included in the Viv- all traditionally recognized carnivore families, as well as all addi- erridae (Flynn, 1996; Hunt, 1987); (ii) Asian linsangs (genus Prion- tional lineages whose membership in traditional families had been odon) are also removed from the Viverridae, and constitute the questioned by previous studies (e.g. Dragoo and Honeycutt, 1997; sister-group to felids (Gaubert and Veron, 2003); (iii) Malagasy car- Flynn and Nedbal, 1998; Gaubert and Veron, 2003; Yoder et al., nivores usually placed in the Herpestidae and Viverridae