On the phylogenetic position of a rare Iberian endemic mammal, the Pyrenean desman (Galemys pyrenaicus)

Abstract

The nucleotide sequences of the complete mitochondrial genome and nine partial nuclear genes of the Pyrenean desman (Galemys pyrenaicus) were determined in order to establish the relative phylogenetic position of this species at different taxonomic levels within the placental tree. Phylogenetic relationships of desman within the family Talpidae were inferred based on complete mitochondrial cytochrome b gene nucleotide sequence data. The Pyrenean desman was unambiguously recovered as sister group of the Russian desman (Desmana moschata) confirming the monophyly of the subfamily Desmaninae. However, phylogenetic relationships among major lineages within the Talpidae could not be confidently resolved. Phylogenetic analyses based on mitochondrial (at the amino acid level) and nuclear (at the nucleotide level) sequence data sets confidently placed desman within the Eulipotyphla (that also included moles, shrews, and hedgehogs), and partially recovered placental interordinal relationships. The monophyly of Laurasiatheria (including Eulipotyphla, Chiroptera, Carnivora, Pholidota, Perissodactyla, and Cetartiodactyla) was strongly supported. Mitochondrial amino acid sequences of Pholidota (pangolins) were found to bias phylogenetic inferences due to long-branch attraction effects. A Bayesian inference based on a combined mitochondrial and nuclear data set without Pholidota arrived at an almost fully resolved tree that supported the basal position of Eulipotyphla within Laurasiatheria.

Introduction

The Pyrenean desman Galemys pyrenaicus (E. Geoffroy Saint-Hilaire, 1811) is an endemic mammal species of the Iberian Peninsula that is presently considered as vulnerable in the 2004 IUCN Red List of Threatened Species (http://www.redlist.org). Until recently, the Pyrenean desman was found at both slopes of the Pyrenees, as well as at the Cantabrian, Iberian and Central mountain ranges. However, its distribution has significantly shrunk during the last decades due to habitat loss and fragmentation (Castien and Gosálbez, 1992, Gonzalez-Esteban et al., 2002). This rare species lives associated to aquatic habitats, and exhibits a highly specialized morphology with a characteristic bi-lobed snout adapted to capturing insects, as well as paddle-like hind feet, and an extremely long and laterally compressed tail adapted to swimming and diving (Palmeirim and Hoffman, 1983).

The Pyrenean desman is classified within the subfamily Desmaninae (Family Talpidae). The only other extant species included within this subfamily is the Russian desman, Desmana moschata that occurs in southwest Russia (Palmeirim and Hoffman, 1983, Hutterer, 1993). The striking distribution of the desmans into two widely separated geographical areas is likely linked to the reduction and isolation of cold areas after the last glaciation. In contrast to the specialized semiaquatic life style of the Desmaninae, the other two subfamilies of moles i.e., Uropsilinae and Talpinae are ambulatory, semi-fossorial or fossorial (Hutchinson, 1968, Yates and Moore, 1990, Hutterer, 1993, Whidden, 2000, Shinohara et al., 2003, Motowaka, 2004). The Uropsilinae includes only one genus (Uropsilus), and occurs in Asia (Hutterer, 1993, Motowaka, 2004). The Talpinae includes 14 genera that are classified into five tribes: Condylurini from North America, Scalopini from Eurasia and North America, Scaptonychini from Eurasia, Urotrichini from Eurasia and North America, and Talpini from Eurasia (Hutterer, 1993, Grenyer and Purvis, 2003, Motowaka, 2004). The phylogenetic relationships of desmans within Talpidae are elusive. The highly distinct morphology of desmans has complicated the finding of synapomorphies with other members of the Talpidae (Yates and Moore, 1990, Whidden, 2000; but see Grenyer and Purvis, 2003, Motowaka, 2004). On the other hand, only four recent molecular studies (Douady et al., 2002b, Douady and Douzery, 2003, Shinohara et al., 2003, Shinohara et al., 2004) using sequence data included desmans but none of them was specifically focused on elucidating their relative phylogenetic position.

Recent phylogenetic studies (Murphy et al., 2001a, Murphy et al., 2001b, Douady et al., 2002b, Douady et al., 2004, Roca et al., 2004) based mainly on nuclear gene sequence data clustered moles into a new order termed Eulipotyphla (Waddell et al., 1999), which also included shrews (Soricidae), hedgehogs (Erinaceidae), and solenodons (Solenodontidae) (Murphy et al., 2001b, Douady et al., 2002b, Roca et al., 2004). Eulipotyphla was placed together with carnivores, pangolins, bats, cetartiodactyls and perissodactyls into a new superordinal clade called Laurasiatheria (Waddell et al., 1999, Madsen et al., 2001, Murphy et al., 2001b, Douady et al., 2002b). In contrast, phylogenetic analyses based on complete mitochondrial genomes initially rejected the monophyly of Eulipotyphla by placing the hedgehog in a basal position of the placental tree, and the remaining Eulipotyphla taxa in a more derived position (Cao et al., 2000, Mouchaty et al., 2000a, Mouchaty et al., 2000b, Nikaido et al., 2001, Arnason et al., 2002). However, the hedgehog mitochondrial sequence data (Krettek et al., 1995) exhibited extremely fast rates of evolution, and it was later demonstrated that the basal position of hedgehog was an artifact due to long-branch attraction by outgroup taxa (Waddell et al., 2001), and could be partly corrected using a more dense taxon sampling (Lin et al., 2002). Furthermore, a maximum likelihood analysis using an appropriate substitution model that corrected for site-heterogeneity was able to recover hedgehogs as members of the Eulipotyphla based on mitochondrial protein sequences (Nikaido et al., 2003). These results reconciled mitochondrial evidence with nuclear (Murphy et al., 2001b, Douady et al., 2002b, Roca et al., 2004) and morphological (MacPhee and Novacek, 1993, McKenna and Bell, 1997, Grenyer and Purvis, 2003) evidences.

Although there is recent wide agreement on the monophyly of both Eulipotyphla (Douady et al., 2002b, Douady et al., 2004, Lin et al., 2002, Nikaido et al., 2003, Roca et al., 2004) and Laurasiatheria (Waddell et al., 1999, Madsen et al., 2001, Murphy et al., 2001b, Lin et al., 2002, Nikaido et al., 2003), the exact phylogenetic position of Eulipotyphla is not fully resolved (Narita et al., 2001). Although this order has been mostly placed as sister group of all other Laurasiatheria (Murphy et al., 2001b, Lin et al., 2002, Nikaido et al., 2003), in some instances it is recovered as sister group of Chiroptera (e.g. Cao et al., 2000, Madsen et al., 2001, Narita et al., 2001, Nikaido et al., 2001). Furthermore, another related open question is the relative phylogenetic position of the order Perissodactyla within Laurasiatheria either as sister group of Cetartiodactyla (Euungulata hypothesis) or of Carnivora + Pholidota (Zooamata hypothesis) (Waddell et al., 1999). Many authors (Douady et al., 2002b, Lin et al., 2002, Nikaido et al., 2003) agree that further resolution among competing phylogenetic hypotheses on laurasiatherian intrarelationships partly relies on a more thorough taxon sampling of Eulipotyphla, a group that was underrepresented in early phylogenetic studies based on molecular data.

Here, we present the complete mitochondrial genome sequence as well as nucleotide sequence data from nine nuclear genes of G. pyrenaicus. We performed a multigene approach because it has been shown that phylogenies based on different partitions can be most confident in having resolved branches (Springer et al., 1999). We conducted different phylogenetic analyses based on the new sequence data with the main aim of clarifying the relative phylogenetic positions of desmans within Talpidae and Eulipotyphla, and of Eulipotyphla within Laurasiatheria. In particular, we wanted to test whether adding a novel non-fast evolving species of Eulipotyphla into the phylogenetic analyses would improve mitochondrial support of the monophyly of this order.