A Newly Recognised Beaked Whale ( Ziphiidae ) in the Tropical Indo-Pacific : Mesoplodon hotaula or M . ginkgodens hotaula

We present genetic and morphological data supporting the recognition of a previously described but unrecognised Mesoplodon beaked whale in the tropical Indo-Pacific. Currently known from at least seven specimens (Sri Lanka [1], Kiribati [1+], Hawai’i [3], Maldives [1], Seychelles [1]), this beaked whale is the sister-taxon to M. ginkgodens proper. The type specimen (Sri Lanka) was described as a new species, M. hotaula, in 1963, but the species was subsequently synonymised with M. ginkgodens by Moore and Gilmore (1965). Analyses of three mitochondrial genes and eight nuclear introns, together with distinct morphological features, suggest that this specimen and the others we have identified as belonging to this lineage represent a distinct species or subspecies of beaked whale. Introduction For the majority of beaked whales (family Ziphiidae), most of what we know has come from beachcast or stranded animals (Reeves et al., 2002). To assist with beaked whale identification and discovery, a comprehensive, validated DNA taxonomy for all known species in this group was established using sequences from mitochondrial DNA (mtDNA) control region (CR) and cytochrome b (CYB) genes (Dalebout et al. 2004, Dalebout et al. 2007). For the genus Mesoplodon, these markers showed consistently low intra-specific variation and comparatively high interspecific divergence. In phylogenetic analyses, DNA sequences from each of the known species clustered together to the exclusion of sequences from the other known species. All species also possessed diagnostic nucleotide substitutions that appear to distinguish them from all other Mesoplodon species (though it is recognised that the small sample sizes available may lead to underestimates of intra-specific variation). These results were in concordance with morphological diagnoses. Previous application of this DNA taxonomy has resulted in several significant findings: the description of a new species from the North Pacific (M. perrini; Dalebout et al. 2002); the resurrection of a long-forgotten species in the Southern Hemisphere (M. traversii; van Helden et al. 2002); and confirmation of the identity of the enigmatic “tropical bottlenose whale” (Indopacetus pacificus; Dalebout et al. 2003). This DNA taxonomy therefore offers a robust framework within which the discovery of a divergent lineage could indicate the existence of an unrecognised species or subspecies. Just such a lineage was reported by Dalebout et al. (2007), represented by several specimens which appeared to be closely related to Mesoplodon ginkgodens Nishiwaki and Kamiya 1958. Further specimens representing this divergent lineage have since been discovered. One of these is a specimen from Sri Lanka described as a new species, M. hotaula (Deraniyagala 1963a,b) but subsequently synonymised with M. ginkgodens by Moore and Gilmore (1965). M. ginkgodens is one of the rarest of beaked whale species. It is known from less than 30 strandings and there has yet to be a confirmed sighting in the wild (MacLeod et al. 2006). Given the few records to date and the fact that they are all strandings, little can be said about comparative distributions. However, the divergent “M. hotaula” lineage appears to be tropically distributed, while strandings of the other lineage, M. ginkgodens proper, have generally been in more temperate regions (Fig. 1). To assess the taxonomic status of this divergent lineage, DNA from three mitochondrial genes and eight nuclear introns, as well as morphological characters, were analysed. 1 There is no evidence that Deraniyagala was aware of the existence of M. ginkgodens when he described M. hotaula. SC/64/SM3 Dalebout et al. 2 Methods Material Examined Seven specimens of M. hotaula were examined (Table 1) and compared to all other known Mesoplodon species using phylogenetic analyses of mtDNA and nuclear gene sequences. (The initial conclusions of Dalebout et al. (2007) were based on mtDNA analyses of specimens 2 – 4). Museums and institutions holding specimens of M. hotaula are as follows: the National Museum, Colombo, Sri Lanka (n = 1), Smithsonian National Museum of Natural History, Washington DC, USA (USNM, n = 3), a private collection in the Republic of Maldives (n = 1), and the Island Conservation Society, Seychelles (n = 1). Specimen 2 from Kiribati is known only from a soft tissue sample held in the University of Auckland DNA and Tissue Archive, Auckland, New Zealand (see SC/64/SM4 Baker et al. for information on additional specimens represented by osteological material from this region). Genetic and morphological comparisons were made to six specimens of M. ginkgodens (Table 1), including the holotype (Nishiwaki and Kamiya 1958). For genetic comparisons to other Mesoplodon species, up to six specimens per species were sampled (see Dalebout et al. 2007 for details). Genetic Analyses The Polymerase Chain Reaction (PCR) was used to amplify fragments from three mitochondrial genes (control region – CR, cytochrome b – CYB, cytochrome oxidase I – COI), seven nuclear autosomal introns (biglycan – BGN, catalase – CAT, rhodopsin – RHO, cytotoxic T-lymphocyte-associated serine esterase 3 – CTLA3, cholinergic receptor-nicotinic alpha polypeptide 1 – CHRNA1, muscle actin ACT, major histocompatibility complex class II – DQA) and one nuclear Y-chromosome intron (DBY7). For all details laboratory procedures and genetic analyses, see Dalebout et al. (2004) and Dalebout et al. (2008), or contact the lead author. Results Genetics Mitochondrial DNA – CR fragments (658 bp) were successfully sequenced from all seven specimens of M. hotaula. CYB fragments (384 – 706 bp) were successfully sequenced from only four specimens due to degraded nature of much of the available material. A COI fragment (987 bp) was successfully sequenced from one specimen (Table 2). These mtDNA fragments were also successfully sequenced from up to six specimens of M. ginkgodens. For mtDNA CR, comparisons between M. hotaula and M. ginkgodens revealed 35 variable sites, of which 18 appear to represent fixed differences between the species (Dα, net divergence 3.6% ± 0.91%). In comparisons including all Mesoplodon species, Dα ranged from 3.1 % to 8.3%. For mtDNA CYB, comparisons between M. hotaula and M. ginkgodens revealed 31 variable sites, of which 26 appear to represent fixed differences between the species, including 4 non-synonymous substitutions (Dα, net divergence 8.2% ± 1.79%). In comparisons including all other known Mesoplodon species, Dα ranged from 5.5% to 16.6%. For mtDNA COI, comparisons between M. hotaula and M. ginkgodens revealed 64 variable sites, of which 49 appear to represent fixed differences between the species (Dα, net divergence 5.5% ± 0.76%). In comparisons including a subset of Mesoplodon species (M. mirus, M. europaeus, M. densirostris), Dα ranged from 5.5% to 10.0%. Intra-specific diversity for M. hotaula at the mtDNA CR and CYB was low, in line with trends observed in other Mesoplodon species. For CR, the Pacific Ocean specimens (Kiribati and Hawai’i) shared the same haplotype, while the Indian Ocean specimens (Sri Lanka, Maldives, Seychelles) differed from this by 3 – 7 bp. For CYB, the Pacific Ocean specimens differed from one another by 1 bp, while the only Indian Ocean specimen sequenced for this locus to date (Seychelles) differed from this by 4 – 5 bp. In phylogenetic analyses of the combined mtDNA CR and CYB sequences (819 bp) including all known Mesoplodon species, the M. hotaula and M. ginkgodens specimens clustered together in two strongly-supported, species-specific clades (bootstrap scores 100%, posterior probabilities 1.00) which were reciprocally monophyletic to one another (Fig. 2). The sister-species relationship of these taxa was also strongly supported (bootstrap 91%, posterior probability 1.00). All other Mesoplodon species formed similar strongly-supported, species-specific clades, with branch lengths reflecting the relatively low genetic diversity observed within species and the comparatively large genetic divergence observed between them (see also Dalebout et al. 2002, Dalebout et al. 2004, Dalebout et al. 2007). Individual analyses of the CR and CYB datasets revealed the same pattern (Dalebout et al. 2007). Nuclear introns – Intron fragments were successfully amplified from seven autosomal genes: BGN, 706 bp; CAT, 559 bp; RHO, 166 bp; CTLA3, 305 bp; CHRNA1, 366 bp; ACT, 925 bp; and DQA, 456 bp. Due to degraded nature of much of the material available, each species was represented by only a single specimen for these analyses. Over all these introns combined (3348 bp), all Mesoplodon species analysed, including the M. ginkgodens-M. hotaula complex, possessed nucleotide substitutions that distinguished them from all other Mesoplodon species (Table 3; see also Dalebout et al. 2008). Further, M. ginkgodens also possessed one nucleotide substitution that distinguished it from M. hotaula and all other Mesoplodon species. SC/64/SM3 Dalebout et al. 3 Intron fragments were also successfully amplified from one Y-chromosome gene (DBY7, 241 bp). Due to degraded nature of much of the material available and the male-only nature of this marker, each species was again represented by only a single specimen. Our inclusion of data from this marker has several advantages. First, under random mating, the effective population size of this non-recombining chromosome is 1⁄4 that of single copy autosomal markers such that the accumulation of mutations through genetic drift occurs far more rapidly. Second, the Y-chromosome is exposed to mutations that have arisen only in the male germline, giving us a male-specific marker to compare to the femalespecific mtDNA markers. For DBY7, the majority of Mesoplodon species sampled, including M. hotaula and M. ginkgodens, possessed at least one nucleotide substitution that distinguished them from all other species in this genus (Table 4). Only one other sister-species pair (M. bowdoini an