Complete mitochondrial genomes of Carcinoscorpius rotundicauda and Tachypleus tridentatus (Xiphosura, Arthropoda) and implications for chelicerate phylogenetic studies

Abstract

Horseshoe crabs (order Xiphosura) are often referred to as an ancient order of marine chelicerates and have been considered as keystone taxa for the understanding of chelicerate evolution. However, the mitochondrial genome of this order is only available from a single species, Limulus polyphemus. In the present study, we analyzed the complete mitochondrial genomes from two Asian horseshoe crabs, Carcinoscorpius rotundicauda and Tachypleus tridentatus to offer novel data for the evolutionary relationship within Xiphosura and their position in the chelicerate phylogeny. The mitochondrial genomes of C. rotundicauda (15,033 bp) and T. tridentatus (15,006 bp) encode 13 protein-coding genes, two ribosomal RNA (rRNA) genes, and 22 transfer RNA (tRNA) genes. Overall sequences and genome structure of two Asian species were highly similar to that of Limulus polyphemus, though clear differences among three were found in the stem-loop structure of the putative control region. In the phylogenetic analysis with complete mitochondrial genomes of 43 chelicerate species, C. rotundicauda and T. tridentatus were recovered as a monophyly, while L. polyphemus solely formed an independent clade. Xiphosuran species were placed at the basal root of the tree, and major other chelicerate taxa were clustered in a single monophyly, clearly confirming that horseshoe crabs composed an ancestral taxon among chelicerates. By contrast, the phylogenetic tree without the information of Asian horseshoe crabs did not support monophyletic clustering of other chelicerates. In conclusion, our analyses may provide more robust and reliable perspective on the study of evolutionary history for chelicerates than earlier analyses with a single Atlantic species.

Keywords: Carcinoscorpius rotundicauda; Horseshoe crabs; Mitochondrial genome; Phylogenetics; Tachypleus tridentatus; Xiphosura.

Fig 1

Gene map of the mitochondrial genome of Asian horseshoe crabs, Carcinoscorpius rotundicauda and Tachypleus tridentatus. Thirteen protein-coding genes and 2 ribosomal RNA genes are abbreviated as follows: nad1-6 and nad4L, NADH dehydrogenase subunits 1-6 and 4L; cox 1-3, cytochrome C oxidase subunits I-III; cob, cytochrome b apoenzyme; atp 6 and 8, ATPase subunits 6 and 8. The 22 transfer RNA genes are identified by the IUPAC amino acid single-letter codes.

Fig 2

Putative tRNA secondary structures predicted from the 22 tRNA gene sequences found in the (A) Carcinoscorpius rotundicauda and (B) Tachypleus tridentatus mitochondrial genome.

Fig 2

Putative tRNA secondary structures predicted from the 22 tRNA gene sequences found in the (A) Carcinoscorpius rotundicauda and (B) Tachypleus tridentatus mitochondrial genome.

Fig 3

Comparison in the primary and secondary structures of non-coding regions found in the mitochondrial genomes of Carcinoscorpius rotundicauda, Tachypleus tridentatus and Limulus polyphemus. The CR pair-wise alignment of the three xiphosuran species (A): the dots indicate nts identical to those of the first line in the alignment. Hairpin-like structure (60 bp and 57 bp) from the control region observed in C. rotundicauda (B), T. tridentatus (C), and L. polyphemus (D). An alignment of the gap region between trnS1 (UCN) and nad1 in three xiphosuran species (E): the box indicates a conserved region across three horseshoe crabs. L; L. polyphemus, C; C. rotundicauda, T; T. tridentatus.

Fig 4

Maximum likelihood phylogenetic trees of chelicerates based on the amino acid sequences of concatenated 12 mitochondrial protein coding genes. Tree was reconstructed either (A) with three horseshoe species (C. rotundicauda, T. tridentatus and L. polyphemus) or (B) with only L. polyphemus. Numbers at the branch indicate the percentages from ML bootstrapping (left) and Bayesian posterior probabilities (right).

Fig 4

Maximum likelihood phylogenetic trees of chelicerates based on the amino acid sequences of concatenated 12 mitochondrial protein coding genes. Tree was reconstructed either (A) with three horseshoe species (C. rotundicauda, T. tridentatus and L. polyphemus) or (B) with only L. polyphemus. Numbers at the branch indicate the percentages from ML bootstrapping (left) and Bayesian posterior probabilities (right).