Comparative Mitochondrial Genome Analyses of Sesarmid and Other Brachyuran Crabs Reveal Gene Rearrangements and Phylogeny

Mitochondrial genomes (mitogenomes) are important for understanding molecular evolution and phylogenetic relationships. The complete mitogenome of Perisesarma bidens was determined, which is 15,641 bp in length. The A + T content of P. bidens mitogenome was 74.81%. The AT skew was slightly negative (−0.021). The 22 tRNAs ranged from 65 to 73 bp and were highly A + T biased. All tRNA genes had typical cloverleaf structures, except for the trnS1 gene, which lacked a dihydrouridine (DHU) arm. The gene order within the P. bidens mitogenome was identical to the pancrustacean ground pattern, except for the translocation of the trnH. Additionally, the gene order of trnI-trnQ-trnM in pancrustacean ground pattern became trnQ-trnI-trnM in P. bidens. Phylogenetic analyses supported the inclusion of P. bidens in Sesarmidae and the promotion of Sesarminae to Sesarmidae. The results will help us to better understand the status and evolutionary history of Grapsoidea crabs.


INTRODUCTION
Decapoda is the most diverse, species-rich group of crustaceans, containing many well-known animals, such as crayfish, lobsters, shrimps, hermit crabs, and "true" crabs (Shen et al., 2013;Basso et al., 2017). The true crabs belong to Brachyura, which is a diverse, economically important group, with about 7200 described species (De Grave et al., 2009;Ahyong et al., 2011). Brachyura is highly adaptable and can live on land and in both marine and fresh water. Therefore, crabs have become important groups for the study of evolution (Castro et al., 2015). Some Brachyura are edible and medicinal and have economic importance (Carpenter and Niem, 1998).
Most Brachyura are grouped into the Podotremata, Heterotremata, and Thoracotremata, with the latter two referred to as the Eubrachyura. However, the phylogenetic relationships within Eubrachyura remain controversial, particularly the relationships of the Sesarmidae and Varunidae, and between these two and Grapsoidea (Schubart et al., 2000(Schubart et al., , 2002Kitaura et al., 2002). The traditional classification of Grapsidae contains four subfamilies: Grapsinae, Plagusiinae, Sesarminae, and Varuninae (Schubart et al., 2000). Traditional methods place the following in the Sesarminae: Perisesarma bidens, Sesarmops sinensis, Clistocoeloma sinensis, Helice tientsinensis, Helice latimera, Helice wuana, and Metaplax longipes. Of these, P. bidens and S. sinensis should be Sesarmops crabs; C. sinensis should be a Clistocoeloma crab; H. tientsinensis, H. latimera, and H. wuana should be Helice crabs and M. longipes should be a Metaplax crab (Schubart et al., 2000). However, some scholars have suggested that Grapsidae should be promoted to Grapsoidea, promoting the four subfamilies to families, i.e., Grapsidae, Varunidae, Sesarmidae and Plagusiidae. Other studies have advised that Sesarmops and Clistocoeloma crabs, which originally belonged to Sesarminae, should belong to Sesarmidae, and that Helice and Metaplax crabs that originally belonged to Sesarminae should be transferred from the Sesarminae to the Varunidae (Kitaura et al., 2002;Schubart et al., 2002). The classification of these taxa remains unresolved. Sesarmid crabs are common in mangroves and can tolerate great variation in salinity along the environment (Theuerkauff et al., 2018). They are very good experimental research objects.
Many studies have investigated these relationships using nuclear DNA, mitochondrial DNA (mtDNA), and morphological character analyses. Some studies combined mtDNA and nuclear genes to reconstruct more reliable phylogenetic trees. However, the combination of these genes made alignment and model selection quite difficult (Foster, 2004;Cox et al., 2008). This has led to the conclusion that the taxon sampling is insufficient and unbalanced (Bergsten, 2005;Wägele and Mayer, 2007). It is evident that more species are necessary to improve the quality of the analyses and stability of phylogenetic trees (Brinkmann and Philippe, 2008).
The mitochondrial genome (mitogenome) has been widely used in phylogenetic analyses, due to its rich signals from sequence information and gene arrangement (Xin et al., 2017a,b). The mitogenome has a simple structure, haploid nature, maternal inheritance, and rapid evolutionary rate (Liu et al., 2015). The mitogenomes are closed circular double-stranded molecules in the range of 14-18 kb in most bilaterian animals, including 13 protein-coding genes (PCGs; cox1-3, cob, nad1-6 and nad4L, atp6, and atp8), 2 rRNA genes, 22 tRNA genes, and an ATrich region (control region) (Tang et al., 2003(Tang et al., , 2017(Tang et al., , 2018Xin et al., 2017a,b). The taxonomy of Sesarmid crabs has been studied extensively and benefited from recent refinements in species of Perisesarma and Sesarmops Shih et al., 2019;Ng et al., 2020). However, the mitogenome of P. bidens has not been analyzed. Here, we determined the mitogenome of P. bidens and used the mitogenomes of 65 species to construct phylogenetic trees to discuss the systemic status and genetic relationships of the controversial taxa, Sesarmidae and Grapsoidea.

Ethics Statement
We have taken a close look at the website 1 . We found that the species Perisesarma bidens is not considered endangered or protected species, the IUCN status for this species is "Not evaluated." Similarly, the species Perisesarma bidens is also Not endangered or protected species in China. No special permit Is required to collect crabs at selected sites in China. the sampling locations are Not privately-owned or natural protected areas, the collection of this species is legal in China. So we can use this species for experiments and subsequent analysis.

Sample Collection
Specimens of P. bidens were collected from the seaside of Zhangzhou City, Fujian Province, China, identified using the morphological methods of Dai (1999) and molecular identification with COI marker, and preserved in 95% ethanol at -20 • C until DNA extraction. Voucher specimens of P. bidens were deposited in the Jiangsu Provincial Key Laboratory of Coastal Wetland Bioresources and Environmental Protection, School of Ocean and Biological Engineering, Yancheng Teachers University, Yancheng, China.

DNA Extraction, PCR, and Genome Sequencing
Total genomic DNA was extracted from muscle using a genomic DNA extraction kit (Sangon, China), following the manufacturer's instructions, and was visualized on 1.0% agarose gels. The complete mitogenome was obtained using a combination of conventional PCR and long PCR to amplify overlapping fragments spanning the entire mitogenome. Initially, conserved sequences, such as cox1, cox3, nad5, nad4, and rrnS, were amplified by conventional PCR using universal primers synthesized by Beijing Sunbiotech (Tang et al., 2003(Tang et al., , 2017(Tang et al., , 2018Liu et al., 2015;Xin et al., 2017a,b).
The PCR conditions for conserved sequences followed a standard three-step protocol, with an initial denaturing at 96 • C

Annotation and Alignment
The sequence was annotated using DNASTAR (DNASTAR, Madison, WI, United States). The locations of the PCGs, rRNA genes, tRNA genes, and CR were initially identified using the MITOS Web Server 2 . The PCG coding regions were further identified using the NCBI ORF Finder 3 . Two rRNA genes were identified by alignment with published brachyuran sequences. Codon usage and the nucleotide composition of the mitogenomes were determined using MEGA6 (Tamura et al., 2013). The nucleotide sequence of the complete P. bidens mitogenome was deposited in the NCBI database under accession no. KY808394. Gene orders in the complete mitogenome were also inferred through the MITOS Web Server.

Phylogenetic Analyses
We used nucleotide (NT) sequences for phylogenetic analyses. The sequences were aligned with MAFFT using the default settings (Katoh et al., 2002). Gaps in the sequences were removed using Gblocks (Castresana, 2000), and the saturation of the sequences was examined using DAMBE (Xia and Xie, 2001), which indicated that the sequences were not saturated and were suitable for phylogenetic analyses. Complete mitogenomes of 65 decapods (60 crabs plus 5 outgroups) were downloaded from NCBI (

Genome Structure, Organization, and Composition
The complete mitogenome of P. bidens was a circular of 15,641 bp (GenBank accession no. KY808394). Its size was within the range observed in completely sequenced brachyuran species. The mitogenome composition (A: 36.61%, T: 38.20%, C: 15.13%, G: 10.06%) was strongly A + T biased which accounts for 74.81%, and exhibited with negative AT-skew (-0.021). The AT-skew of the mitogenomes of most crabs were negative, for example, H. wuana (Tang et al., 2018), S. sinensis (Tang et al., 2017), H. tientsinensis (Xin et al., 2017b), C. sinensis (Xin et al., 2017a), the AT-skew value of mitogenomes in other crabs had also been calculated and counted in related studies (Xin et al., 2017a,b). The genes were typical of animal mitogenomes, with 22 tRNA genes, 13 PCGs, 2 rRNA genes, and a CR ( Table 2). Overall, 4 of the 13 PCGs (nad5, nad4, nad4L, and nad1), 8 tRNAs [trnQ, trnC, trnY, trnF, trnH, trnP, trnL (CUN), and trnV], and 2 rRNAs (rrnL and rrnS) were encoded by the minority strand, while the other 23 genes were encoded by the majority strand ( Table 2 and Figure 1). The 13 PCGs ranged from 159 to 1731 bp. Of 22 tRNA genes, 8 were encoded by the L-strand and the remaining 14 by the H-strand. All tRNAs had the typical clover-leaf secondary structures observed in mitochondrial tRNA genes, except for trnS1 (AGN), which lacked a stable dihydrouridine (DHU) arm; this has been observed in several animals, including insect and brachyuran mitogenomes (Liu et al., 2015;Xin et al., 2017a,b). Figure 2 shows the relative synonymous codon usage (RSCU) of P. bidens. The codon usage was biased with a high frequency of AT compared to GC in the third codon position. The codon usage analysis revealed that the leucine 2 (Leu2), isoleucine (Ile), phenylalanine (Phe) codon families were most frequently utilized, while cysteine (Cys) family was the least used (Figure 3).

Gene Order in Sesarmidae
The gene order of P. bidens was identical to other Sesarmidae species in our study. In contrast to the inferred ancestral gene sequences of Pancrustaceans, where trnH was located between nad5 and nad4, here it was found between trnE and trnF. In Pancrustaceans, the tRNA gene sequences between CR and trnM was trnI-trnQ, but here was trnQ-trnI ( Figure 4A). The duplication/random loss model was used to explain the rearrangements seen in Sesarmidae (Moritz and Brown, 1987;Macey et al., 1997;Boore and Brown, 1998). The movement of trnH can be explained as follows. First, gene duplication occurred in trnF, nad5, and trnH, changing the arrangement of trnF-nad5-trnH to trnF-nad5-trnH-trnF-nad5-trnH. Then, the redundant trnF, nad5, and trnH genes were lost at random. Finally, the new gene order of trnH-trnF-nad5 was formed ( Figure 4B). The order principles of trnQ moving from the junction between trnI and trnM to between the CR and trnI could also be explained similarly ( Figure 4C).

Gene Order of Crabs From Other Families
The gene orders of all species are shown in Figure 5. The gene sequences within 13 families were the same. The gene order pattern of Macrophthalmus japonicus (Ocypodoidea, Macrophthalmidae) was identical to that of other Varunidae. The gene orders of Damithrax spinosissimus (Majoidea, Mithracidae) and Dynomene pilumnoides (Dromioidea, Dynomenidae) were different, as were those of two Xenograpsidae crabs (Xenograpsus testudinatus and X. ngatama). However, two Majidae crabs (Maja  FIGURE 4 | Generation of the P. bidens mitochondrial gene arrangement. The duplication/random loss, recombination, and duplication/non-random loss models were used to explain the principle of gene rearrangement. (A) Comparison of gene order in mitogenome of Perisesarma bidens and Pancrustacean ground pattern. tRNA genes are indicated by the singer letter IUPAC-IUB abbreviation with S1 = AGN, S2 = UCN, L1 = CUN, and L2 = UUR, where as protein and rRNA genes are labled with three letter codes. (B) Gene duplication occurred in trnF, nad5, and trnH, changing the arrangement of trnF-nad5-trnH to trnF-nad5-trnH -trnF-nad5-trnH. Then, the redundant trnF, nad5, and trnH genes were lost at random. Finally, the new gene order of trnH-trnF-nad5 was formed. (C) Gene duplication occurred in trnI, trnQ, and trnM, changing the arrangement of trnI-trnQ-trnM to trnI-trnQ-trnM-trnI-trnQ-trnM. Then, the redundant trnI, trnM, and trnQ genes were lost at random. Finally, the new gene order of trnQ-trnI-trnM was formed.
squinado and M. crispata) had the same gene order. Interestingly, although there were only four species of Potamoide, each showed a different gene order.

Phylogenetic Analyses
The phylogenetic trees were constructed based on 13 PCGs under ML and BI methods, which resulted in congruent tree topologies, except for minor differences within "Grapsoidea + Ocypodoidea" (Figure 6). As shown in Figure 6, P. bidens formed a wellsupported clade with Parasesarmops tripectinis (BP = 100; BPP = 1). (P. bidens + P. tripectinis) clade, (S. sinensis + S. neglectum) clade, (C. sinensis + M. depressus) clade were well supported with each other; these results were in accordance with the information provided by the same genes orders of P. bidens, P. tripectinis, S. sinensis, S. neglectum, C. sinensis, and M. depressus. Moreover, S. neglectum, M. depressus, and P. tripectinis all belonged to Sesarmidae (Park et al., 2018). Therefore, P. bidens, S. sinensis, and C. sinensis should belong to Sesarmidae rather than to Sesarminae. The species originally belonging to the Sesarminae should belong to the Sesarmidae. These results agree with previous analyses using the mitogenome of one species (Tang et al., 2017;Xin et al., 2017a,b).
In our study, two families (Potamidae and Parathelphusidae) were primarily freshwater crabs and were recognized as true heterotremes (Guinot et al., 2013). The systemic status of primary freshwater crabs had stimulated interest because of their high value and diversity (Cumberlidge et al., 2009;Klaus et al., 2010). The monophyly of Potamidae and Parathelphusidae was confirmed based on morphological and molecular analyses.
However, there still were uncertainties regarding the phylogenetic placement of Potamidae and Parathelphusidae (Xing et al., 2017). Von Sternberg and Cumberlidge (2001) suggested that these two families Potamidae and Parathelphusidae should be placed in Thoracotremata. Here, the Thoracotremata contained Grapsoidea and Ocypodoidea crabs. Our results showed that four heterotreme crabs (Geothelphusa dehaani, Longpotamon xiushuiense, Huananpotamon lichuanense, and Somanniathelphusa boyangensis) were actually more closely associated with thoracotreme crabs, showing that Heterotremata was not monophyletic; this result was in accordance with that inferred from 23 brachyuran crabs, in which the author use the two mitogenomes (Ji et al., 2014). Within Podotremata, the clade was monophyletic. The six crabs formed a robust clade [(Homolidae + Dynomenidae) + Raninidae]. Within Heterotremata, the phylogenetic relationships were clear, with FIGURE 6 | Phylogenetic trees were constructed using BI and ML methods based on NT dataset. Bootstrap values (BP) (IQ-Tree) and Bayesian posterior probability (BPP) of each node are shown as BP based on NT dataset/BPP based on NT dataset. C. destructor, C. similis, N. maculatus, P. camtschaticus, and P. longicarpus were used as outgroups. The supermatrix underlying this figure is as a Supplementary File. the exception of the four potamid crabs, which were outside of the heterotreme crabs.

DATA AVAILABILITY STATEMENT
The datasets generated for this study can be found in the GenBank accession no. KY808394.

AUTHOR CONTRIBUTIONS
Q-NL, B-PT, and X-MY conceived and designed the study. Z-ZX, Q-NL, Y-YT, and T-TY conducted the molecular work and data analysis. Z-ZX and Y-TL drafted the manuscript. Z-ZX, Q-NL, Y-YT, YS, D-ZZ, C-LZ, and T-TY prepared all figures and tables. Z-ZX and Q-NL performed the phylogenetic analyses. Z-ZX, Q-NL, B-PT, and X-MY contributed to drafting the manuscript. All authors contributed to the article and approved the submitted version.