Complete mitochondrial genome of Clistocoeloma sinensis (Brachyura: Grapsoidea): Gene rearrangements and higher-level phylogeny of the Brachyura

Deciphering the animal mitochondrial genome (mitogenome) is very important to understand their molecular evolution and phylogenetic relationships. In this study, the complete mitogenome of Clistocoeloma sinensis was determined. The mitogenome of C. sinensis was 15,706 bp long, and its A+T content was 75.7%. The A+T skew of the mitogenome of C. sinensis was slightly negative (−0.020). All the transfer RNA genes had the typical cloverleaf structure, except for the trnS1 gene, which lacked a dihydroxyuridine arm. The two ribosomal RNA genes had 80.2% A+T content. The A+T-rich region spanned 684 bp. The gene order within the complete mitogenome of C. sinensis was identical to the pancrustacean ground pattern except for the translocation of trnH. Additionally, the gene order of trnI-trnQ-trnM in the pancrustacean ground pattern becomes trnQ-trnI-trnM in C. sinensis. Our phylogenetic analysis showed that C. sinensis and Sesarmops sinensis cluster together with high nodal support values, indicating that C. sinensis and S. sinensis have a sister group relationship. The results support that C. sinensis belongs to Grapsoidea, Sesarmidae. Our findings also indicate that Varunidae and Sesarmidae species share close relationships. Thus, mitogenomes are likely to be valuable tools for systematics in other groups of Crustacea.

its evolutionary status and rearrangement information by comparing it with complete Brachyuran mitogenomes available to date 20,21 . This information may provide insights into phylogenetic rearrangement and enable phylogenetic analysis.  Table 3. Summary of Clistocoeloma sinensis mitogenome.   Complete Mitogenome Analysis. The graphical map of the complete mitogenome was drawn using the online mitochondrial visualization tool mtviz 27 . The secondary cloverleaf structure and anticodon of transfer RNAs were identified using the tRNA-scan SE webserver 28 . Codon usage and the nucleotide composition of the mitogenome were determined using MEGA6. The sequences of 29 Brachyura species and Alpheus distinguendus were aligned using MAFFT 29 .
Phylogenetic Analysis. Twenty-eight complete Brachyura mitogenomes were downloaded from GenBank (https://www.ncbi.nlm.nih.gov/genbank/). In addition, the mitogenome of A. distinguendus was downloaded from GenBank and used as an outgroup taxon. GenBank sequence information is shown in Table 2.
The sequences were aligned with the mitochondrial sequences of closely related species. In order to remove the gaps in sequences, poorly aligned positions and divergent regions were removed using Gblocks 25 . Then, fasta sequences were converted to nex format sequences and phylip format sequences for Bayesian inference (BI) and Maximum likelihood (ML) analyses using online software (http://sequenceconversion.bugaco.com/converter/ biology/sequences/fasta_to_phylip.php). We used DAMBE to detect the saturation status of the sequences 30 .
We determined the taxonomic status of C. sinensis within Brachyura by reconstructing the phylogenetic tree. Nucleotide sequences from 30 mitogenome PCGs were combined. The dataset was run using two inference methods: BI and ML analyses. The former was performed using Mrbayes v3.2.1 31 , while ML analysis was performed using raxmlGUI 32 . The nucleotide substitution model was selected using Akaike information criterion implemented in Mrmodeltest v2.3 33,34 . The GTR+I+G model was the best model to examine nucleotide phylogenetic   analysis and molecular evolution. BI and ML analyses were performed under the GTRCAT model with nucleotide alignment (NT dataset) of the 13 mitochondrial PCGs. ML analyses were performed on 1000 bootstrapped datasets. The BI analysis ran as 4 simultaneous MCMC chains for 10,000,000 generations, sampled every 100 generations, and a burn-in of 5000 generations was used. The average standard deviation of split frequencies was less than 0.01, and the effective sample size determined using tracer v1.6 exceeded 200. These two findings indicate that our data was convergent. The resulting phylogenetic trees were visualized using FigTree v1.4.2.

Results and Discussion
Genome Structure and Organization. The mitogenome of C. sinensis is 15,706 bp long, and its gene content is same as that most known Brachyura: 13 PCGs, 2 rRNA genes, and 22 tRNA genes plus CR (Table 3 and  (Table 4). However, different regions have different A+T contents. The CR had the highest A+T content (82.9%), whereas the PCG region had the lowest A+T content (74.2%) ( Table 5).
Protein-Coding Genes. Among the 13 PCGs, 9 (nad2, cox1, cox2, atp8, atp6, cox3, nad3, nad6, and cob) were coded on the J strand, while the rest (nad5, nad4, nad4L, and nad1) were on the N strand. The 13 PCGs ranged in size from 159 to 1731 bp (Table 3). Their A+T content was 74.2% and AT skew was −0.026 (Table 5). The relative synonymous codon usage for C. sinensis at the third position is shown in Fig. 2. The usage of both two-and four-fold degenerate codons was biased toward the use of codons abundant in A or T (Table 6), which is consistent with other Brachyura species 35-37 . Transfer RNAs, Ribosomal RNAs, and A+T-Rich Region. Like most Brachyura mtDNA, the C. sinensis mitogenome contains a set of 22 tRNAs genes (Fig. 3), although this feature is not very well conserved in animal mtDNA. The tRNAs ranged in size from 64 to 73 bp and showed a strong A+T bias, as these bases accounted for 76.2% of the DNA. Further, they exhibited a negative AT skew (−0.010) ( Table 5). Fourteen tRNA genes were present on the J strand and eight were on the N strand. All the tRNA genes had the typical cloverleaf structure, except for the trnS1 gene, whose dihydroxyuridine arm was instead just a simple loop (Fig. 3). These features are common in most Brachyura mitogenomes [35][36][37] . The secondary cloverleaf structure of 18 tRNAs was examined using tRNA-scan SE; 4 tRNAs not detected by tRNAscan-SE were found in the unannotated regions by sequence similarity to the tRNAs of other crabs. The 2 rRNA genes with 80.2% total A+T content and positive AT skew (0.007) ( Table 5) were located between trnL1 and trnV and between trnV and CR. rrnL is 1336 bp while rrnS is 832 bp. The CR located between rrnS and trnQ, spans 684 bp. This region contains 82.9% AT nucleotides, with a positive AT skew (0.047) and negative GC skew (−0.228) ( Table 5).
Gene Arrangement. Gene order within the complete mitogenome of C. sinensis is similar to the pancrustacean ground pattern [38][39][40] (Fig. 4A), except for the translocation of trnH. Typically, the trnH gene is located between the nad4 and nad5 genes in the pancrustacean ground pattern, but in C. sinensis, it is between the trnE and trnF genes (Fig. 4B). This translocation was also observed in the mitogenomes of Brachyura crabs available in GenBank that were compared with the C. sinensis mitogenome. In addition, in the pancrustacean ground pattern, the tRNA gene order between the CR and nad2 is trnI-trnQ-trnM. However, in C. sinensis, it is trnQ-trnI-trnM (Fig. 4B). The tRNA rearrangements are generally considered to be a consequence of tandem duplication of part of the mitogenome 41 . Similar non-coding sequences are present at the position of trnI originally occupied by the transposed trnQ in C. sinensis. Because these intergenic sequences have similar lengths to those of typical tRNA genes, they were presumed to be remnants of the trnQ gene and its boundary sequences 42 . The gene order  of C. sinensis is identical to that of S. sinensis (Fig. 4B), which indicates that C. sinensis may belong to the group Sesarmidae of the superfamily Grapsoidea and that C. sinensis and S. sinensis probably belong to sister groups. The gene sequences of Varunidae species (Eriocheir japonica sinensis, E. j. hepuensis, E. j. japonica, and Helice latimera) are identical (Fig. 4C). As shown in Fig. 4D, the order and orientation of genes in 7 families are uniform. The order of genes in C. sinensis sequences is different from that in the sequences of the mitogenomes of these 7 families because of the rearrangement of two tRNA genes between CR and trnM: the placement of genes between CR and trnM in C. sinensis is CR-trnQ-trnI-trnM, while that in the 7 families is CR-trnI-trnQ-trnM. In this case, tandem duplication of gene regions may be the most likely mechanism for mitochondrial gene rearrangement, which includes trnI and trnQ, followed by loss of supernumerary genes 43,44 . Slipped-strand mispairing occurred first, followed by gene deletion 45 . Partial PCGs, tRNAs, and rRNAs of Damithrax spinosissimus, G. dehaani, and Xenograpsus testudinatus appear to be rearranged compared to C. sinensis (Fig. 4E-G).
Phylogenetic analysis. Our analyses were based on the NT dataset in mitogenomes derived from 29 Brachyura species belonging to 12 families (Varunidae, Xenograpsidae, Homolidae, Menippidae, Mithracidae, Potamidae, Portunidae, Raninidae, Bythograeidae, Sesarmidae, Grapsidae, and Dotillidae). The data matrix (15,706 bp in all) was analysed using the model-based evolutionary methods of BI and ML analyses (Fig. 5). The ML and BI analyses of the dataset gave the same tree topology. It is obvious that C. sinensis and S. sinensis clustered in one branch in the phylogenetic tree with high nodal support values (Fig. 5), indicating that C. sinensis  and S. sinensis have a sister group relationship. This result supported that C. sinensis belongs to Grapsoidea, Sesarmidae. From the phylogenetic tree, we found that X. testudinatus and two Sesarmidae species formed a group and showed close relationships. X. testudinatus, which was originally placed in Varunidae, has been transferred to its own family (Xenograpsidae) 21,46 . Analysis of the nucleotide sequences of the 13 mitochondrial PCGs using BI and ML showed that E. j. sinensis, E. j. hepuensis, E. j. japonica, and H. latimera clustered together with high statistical support, showing that these species have a sister group relationship and belong to Grapsoidea, Varunidae. Our phylogenetic analysis indicated that Sesarmidae species, Xenograpsidae species and Varunidae species have close relationships 47 . In addition, P. crassipes belongs to Grapsoidea, Grapsidae 48 .
The phylogenetic position of Ilyoplax deschampsi is always within Grapsoidea 21,47,49,50 . I. deschampsi belongs to the family Dotillidae, Ocypodoidea. The real phylogenetic position of I. deschampsi should be closer to the Grapsoidea species that shown in Fig. 5. Recent studies on the genus Ucides have also shown similar classification 51,52 . G. dehaani belongs to Potamidae, Potamoidea 53 . However, the phylogenetic tree showed that Potamidae are associated closely with Varunidae, Grapsidae, Sesarmidae, Dotillidae, and Xenograpsidae. This result is in agreement to that inferred from 23 Brachyuran crabs, in which the author use the two mitogenomes 21 . Phylogenetic relationships between I. deschampsi, G. dehaani and Grapsoidea species need to be reconsidered by integrating more mitogenomic data. More mitogenomic data will also lead to a better overall understanding the phylogenetic relationships among Brachyuran crabs.
Availability of data and materials. The data set supporting the results of this article is available at NCBI (KU589292).