The mitochondrial genome of Indonesian coelacanth Latimeria menadoensis (Sarcopterygii: Coelacanthiformes) and divergence time estimation between the two coelacanths

Abstract

We determined the whole mitochondrial genome sequence for Indonesian coelacanth Latimeria menadoensis. The genome content and organization were identical to that of typical vertebrates including Comoran coelacanth, Latimeria chalumnae. The overall nucleotide differences between the two species (excluding the control region) was 4.28%. The divergence time between the two species was estimated using whole mitochondrial genome data from the two coelacanths and 26 actinopterygians that represent major actinopterygian lineages plus an outgroup. Partitioned Bayesian analyses were conducted with the two data sets that comprised concatenated amino acid sequences from 12 protein-coding genes (excluding ND6 gene) and concatenated nucleotide sequences from 12 protein-coding genes (without 3rd codon positions), 22 transfer RNA genes, and two ribosomal RNA genes. The molecular clock analysis was also conducted with the concatenated amino acid sequences from the 12 protein-coding genes after removing faster or more slowly evolving sequences. Using the sarcopterygian–actinopterygian split as a calibration point (450 Mya), divergence time estimation between L. menadoensis and L. chalumnae fell in the range of 40–30 Mya, which is much older than those of the previous studies (<6.3 Mya). Assuming that the most recent ancestor of Latimeria was distributed continuously along the deep coasts of Africa through Eurasia, our estimate is in agreement with the hypothesis that the collision of India with Eurasia (50 Mya) and the subsequent siltation caused by the formation of major rivers resulted in a coelacanth habitat disjunction that allowed populations on either side of India to diverge.

Introduction

Coelacanths were believed to have gone extinct more than 80 Mya until the sensational rediscovery of one surviving member of this lineage, Latimeria chalumnae, in 1938 (Thomson, 1991). Since then, more than 200 coelacanths have been caught off the Comoro archipelago near the eastern coast of Africa in the Indian Ocean (Forey, 1998). During the period of September 1997 through July 1998, two coelacanths were captured off the coast of Manado, Sulawesi, Indonesia, some 10,000 km east of the southwestern Indian Ocean (Erdmann et al., 1998). These coelacanths are the first individuals recorded from a location outside the western Indian Ocean. The extensive interviews with Indonesian fishermen, combined with the vast distance from the Comoro archipelago, supported the idea that the Indonesian coelacanths are part of an established north Sulawesi population, and not simply waifs from the Comoran population (Forey, 1998).

Taxonomic confusion raised by overlapping morphological variations has posed problems in relation to coelacanth dispersal and biogeography. Pouyaud et al. (1999) described the Indonesian coelacanth as a new species, Latimeria menadoensis, based on nine morphological and meristic differences. However, Erdmann et al. (1999) suggested that the Indonesian coelacanth is morphologically similar to the Comoran coelacanth based on a preliminary comparison of external morphological measurements. Based on a survey of the literature, Holder et al. (1999) also stated that the case for morphological differentiation of L. menadoensis is much more tenuous than originally reported (Pouyaud et al., 1999) and concluded that one important morphological character that should be considered in future examination is that of scale ornamentation.

Two independent research groups have published results from divergence time estimation of the two coelacanths using partial mitochondrial gene sequences (6.3–4.7 Mya: Holder et al., 1999; 1.3 Mya: Pouyaud et al., 1999). Nevertheless, divergence time between the two coelacanths has remained ambiguous. In general, lineage-specific variation in rate of molecular evolution complicates molecular dating because a calibration rate estimated from one lineage may not be an accurate representation of the rate in other lineages (Bromham and Penny, 2003). These two studies, however, used rates of evolution from amphibians or teleosts for estimation without conducting the test of rate variation among lineages. Moreover, estimation of divergence time is generally more difficult than reconstruction of a phylogenetic tree, because no gene would evolve at a constant rate (Glazko and Nei, 2003). Considering the effect of rate variation among small number of genes, it is no wonder that the analyses based on partial mitochondrial gene sequences may be significantly biased.

In this study, we determined whole mitochondrial genome for Indonesian coelacanth L. menadoensis and compared the new sequence to that already reported for Comoran coelacanth L. chalumnae. Recent authors have used many genes to estimate divergence times in the hope of reducing the effect of rate variation (Nei and Glazko, 2002). We used whole mitochondrial genome sequences to estimate the divergence time between the two coelacanths because they have been demonstrated in recent studies as being useful for estimating the divergence times among basal lineages within tetrapods (Kumazawa et al., 2004) and within primates (Schrago and Russo, 2003). Two distinct methodologies were employed to estimate divergence time: the partitioned Bayesian approach and the molecular clock approach. Based on the molecular evidence, we evaluated several alternative hypotheses about the speciation of the two coelacanths.