Panulirus ornatus (Fabricius, 1798)

Qian, Guanghui, Zhao, Qiang, Wang, An, Zhu, Lin, Zhou, Kaiya & Sun, Hongying, 2011, Two new decapod (Crustacea: Malacostraca) complete mitochondrial genomes: bearings on the phylogenetic relationships within the Decapoda, Zoological Journal of the Linnean Society 162 (3), pp. 471-481 : 474-478

P. ORNATUS MITOCHONDRIAL GENOMES

The complete mt sequences of the snapping shrimp A. distinguendus and ornate rock lobster P. ornatus comprise 15 700 and 16 105 bp, with overall AT content 60.2 and 66.7%, respectively. The genetic content and organization of the two genomes are summarized in Figure 1 . They contain 13 protein-coding genes, two rRNAs, 22 tRNAs, and a major noncoding region that is typical for many other arthropods, but no tRNA gene for trnS (AGN) [trn for Serine (S 1, AGN)] has been located in A. distinguendus . Although a pseudogene-like trnS (AGN) is identified at the same location of the genome as in P. ornatus , its anticodon is UCG instead of the usual UCU ( Fig. 2 ). The coding region of tRNAs is also absent in the genomes of the two crustacean taxa, the hydrothermal vent galatheid crab Shinkaia crosnieri ( Decapoda : Anomura) in which trnS (UCN), trnW, trnC, and trnY are absent ( Yang & Yang, 2008), and the common sea slater Ligia oceanica (Isopoda) in which trnR is absent ( Kilpert & Podsiadlowski, 2006). It is proposed in this study that the transformation of the pseudogene-like trnS (AGN) to functional trnS (AGN) in A. distinguendus could be realized by changing the guanine to uridine using post-transcriptional RNA editing, and this mechanism of RNA modification was proved by Janke & Paabo (1993). A major noncoding region containing runs of AT dinucleotides was identified in rrnS and trnI genes of both the A. distinguendus (890 bp, AT% = 76.6) and P. ornatus (1176 bp, AT% = 69.4) mtDNAs. A duplicate of 89 bp sequences was found in the noncoding region of P. ornatus ( Fig. 3A, C ), but not in its congeneric species Panulirus japonicus ( Yamauchi et al., 2002) . However, the noncoding regions of both Panulirus species have the potential to form stem-loop structures ( Fig. 3A, B ), a common characteristic of control regions that is probably associated with replication and transcription ( Zhang, Szymura & Hewitt, 1995).

Genome arrangement is identical in A. distinguendus and P. ornatus , except that in A. distinguendus trnE (transfer, trn for Glutamate) is located between the cob and trnS (UCN) genes with opposite orientation. The rearranged gene order found in A. distinguendus represents a novel rearrangement in the Crustacea ( Fig. 1 ). Intramitochondrial recombination ( Dowton & Austin, 1999) and tandem duplication followed by random deletion may account for the inversion and translocation of the trnE gene, respectively.

PHYLOGENETIC RELATIONSHIPS WITHIN THE DECAPODA

Figure 4 shows the topologies derived from ML and Bayesian inferences based on 31 mt protein-encoding gene sets, i.e. nucleotide and amino acid sequences. The monophyly of the Decapoda is consistently supported by high ML bootstrap percentages (BPs) in both nucleotide (90%) and amino acid datasets (89%), and by high Bayesian posterior probabilities (BPPs) in both datasets (1.00 and 1.00). Within the monophyletic decapods, two well-supported clades were recovered, which correspond to the two suborders Dendrobranchiata (clade a, BP = 100 and 100, BPP = 1.00 and 1.00), and Pleocyemata (clade b, BP = 98 and 95, BPP = 1.0 and 1.0). The Pleocyemata was further subdivided into two subclades, clade c and clade d with strong support from two datasets, with the former corresponding to the Caridea clade (BP = 100 and 100, BPP = 1.00 and 1.00). The assemblage of Astacidea, Achelata, Anomura, and Brachyura (clade d) is strongly supported by high MLBP (100 and 100) and BPP (1.00 and 1.00) in both datasets, corresponding to the Reptantia clade. Within clade d, a sister-group relationship between the Anomura and Brachyura is supported by ML analyses (BP = 75 and 93) and strongly by Bayesian inference (BPP = 1.00 and = 1.00).

Assessing the relationships amongst decapod taxa is an intensely debated issue in crustacean phylogeny, with various inferences and testing of different characteristics from morphological, embryological, and spermatological to molecular data. The classical classification proposed the departure of two clades within the order Decapoda , the Natantia for the swimming forms and the Reptantia for the walking forms ( Boas, 1880). As an alternative, the splitting of the suborder Dendrobranchiata and Pleocyemata within the order Decapoda has been widely accepted recently, although Holthuis (1993) felt that treating the penaeoids as a separate group ( Dendrobranchiata ) in the same rank with the combination of Natantia + Macrura Reptantia + Anomura + Brachyura ( Pleocyemata ) was unsatisfactory (see Martin & Davis, 2001).

Recent phylogenetic studies based on mt genomic data have exhibited ambiguity in recovering the monophyly of the Decapoda , the Pleocyemata , and even the status of Meiura of the Reptantia (but see Ki et al., 2009; Liu & Cui, 2009). Several factors can introduce phylogenetic errors or technical artefacts in tree reconstruction when using mt proteins. An important factor is the appropriate choice of out- group. This is the case for the Caridea , when using stomatopod taxa as outgroups, the Caridea is grouped either with Penaeoidea forming a natantian clade ( Yang & Yang, 2008), or with nonpenaeoid decapods forming a pleocyematan clade ( Shen et al., 2007; Ki et al., 2009; Liu & Cui, 2009), or constitutes a separate branch as a basal lineage to all other decapods ( Yang & Yang, 2008; Shen et al., 2009). Distantly related outgroups, as with two branchiopods in Peregrino-Uriarte et al. (2009), and with both branchiopod and isopod taxa in Ivey & Santos (2007), might also have resulted in the separated status of the Caridea . The Isopoda , Stomatopoda , and Euphausiacea are three related groups with the Decapoda . The possibility of using these taxa as outgroups in phylogenetic inference of the Decapoda was tested here. The result indicated an artefactual position of the Caridean clade as sistergroup of the Penaeoidea with Isopoda used as an outgroup (see Supporting Information Figure 1 ). Recent phylogenetic studies revealed a strongly supported clade, the combination of Anaspidacea + Stomatopoda + Euphausiacea + Decapoda ( Jenner et al., 2009) , which also implies a close relationship to the Decapoda , the Euphausiacea , and the Stomatopoda , instead of the Isopoda . The choice of the two best-fit outgroups ( Euphausiacea and Lysiosquillina ) according to our phylogenetic tests and the latest evidence, strongly supports the relationship ‘ Decapoda = Pleocyemata + Dendrobranchiata’ at both the nucleotide and amino acid levels ( Fig. 4 ). This two branch classification was also verified by Tsang et al. (2008) and Chu et al. (2009), as well as Bracken et al. (2009a) with a combination of mt and nuclear sequences, but recently the monophyly of the Pleocyemata was rejected by Bracken et al. (2010) using almost the same molecular markers. The resolving power of multiple nuclear and mt genes in the decapod phylogeny was questionable because of the unstable position of the caridean lineage ( Toon et al., 2009; Bracken et al., 2010).

Taxonomic sampling is another important factor affecting the accuracy of the decapod phylogeny when using mt genomes. This is the case for the Meiura. As an achelatan taxon and anomuran, it was grouped together with the brachyurans based on Bayesian inference of nucleotide data in Miller & Austin (2006), and as an achelatan plus astaciean it was grouped together with the brachyurans from ML analyses of amino acid data in Shen et al. (2007). These ambiguous statuses of the Meiura were not modified until the taxon sampling from the Anomura increased ( Yang & Yang, 2008; Ki et al., 2009; Liu & Cui, 2009; Shen et al., 2009). The modified status is also verified by our datasets if we exclude one anomuran species, Shinkaia crosnieri . In our analyses, the topology ‘Meirura = Brachyura + Anomura’ was strongly recovered at both the nucleotide and amino acid levels. The Meiura clade was also strongly supported by previous research based on morphological and molecular evidence ( Scholtz & Richter, 1995; Dixon et al., 2003; Ahyong & O’Meally, 2004; Schram & Dixon, 2004) and moderately supported recently by Tsang et al. (2008) and Chu et al. (2009) using two nuclear protein-coding genes. However, there is no support for the monophyly of the Meiura clade from recent phylogenetic inferences using multiple nuclear and mt genes ( Toon et al., 2009; Bracken et al., 2009a, 2010). The relationships within the Meiura remain to be corroborated by further studies, albeit the monophyletic Meiura clade is strongly supported from our results and most other evidence.

Currently, whereas a consensus has widely been reached on the monophyly of the reptantian clade, the phylogenetic affinities of the basal groups within the Reptantia remain controversial ( Scholtz & Richter, 1995; Ahyong & O’Meally, 2004; Schram & Dixon, 2004; Porter et al., 2005; Tsang et al., 2008; Toon et al., 2009; Bracken et al., 2009a, 2010). The present phylogenetic analyses and other related studies on mt genomes indicate that the Meiura clade is in a sistergroup position with other reptantians, Palinura and Astacidea ( Yang & Yang, 2008; Ki et al., 2009; Liu & Cui, 2009; Shen et al., 2009). Mitochondrial genome sampling for the key reptantian taxa, such as the Thalassinidea and Polychelida, is needed to clarify the relationships within the Reptantia.

Apart from outgroup selection and taxonomic sampling, phylogenetic signals in mt protein data and model specifications are important factors that can add noise as a result of multiple substitution processes and erroneous homology hypotheses caused by ambiguous sequence alignment. For example, the third codon positions of nucleotide sequences were omitted before alignment, thereby causing potential problems for the selection of homologous sequences ( Yang & Yang, 2008). The bias in nucleotide composition ( Peregrino-Uriarte et al., 2009; Shen et al., 2009), can also influence phylogenetic reconstruction because of substitution saturation, as indicated by our analyses. Additionally, although the choice of an appropriate evolutionary model could have resulted in a much higher likelihood and reflected true phylogenetic relationships ( Rota-Stabelli, Yang & Telford, 2009), the substitution models were not selected for the amino acid sequences that was conducted in Peregrino-Uriarte et al. (2009). Overall, our study demonstrates that phylogenetic signals can be missing because of the systematic errors listed above and thus lead to incorrect phylogenetic inferences for the Decapoda when using mt genomes. However, our results strongly suggested that by decreasing the systematic errors to the utmost degree, it is possible to extract useful historical signals from mt protein sets. This indicates that mt protein data can be promising in the supply of independent evidence for the resolution of relationships amongst different groups within the Decapoda . Considering the possible discrepancy between gene trees and species trees ( Maddison, 1997; Lavrov, Wang & Kelly, 2008), a combined analysis of additional mt genomes and/or other independent molecular sampling accompanied by morphological evidence is likely to resolve further the decapod crustacean tree.

Figures

Figure 1. Mitochondrial genome maps of Alpheus distinguendus and Panulirus ornatus. Protein-coding genes in the majority strand (J strand) are transcribed in a clockwise direction, except for those in the minority strand (N strand) indicated by underlining. The two ribosomal RNA genes are encoded by the N strand. All the darkly coloured transfer RNA (tRNA) genes encoded by the J and N strand are presented outside and inside the circular gene map, respectively. The tRNA genes are given as single-letter amino acid codes, except those encoding leucine and serine, which are labelled L1 (trnL(CUN)) and L2 (trnL (UUR)), and as S1 (trnS(AGN)) and S2 (trnS(UCN)), respectively. Numbers inside the circles represent the size of the fragments separating two adjacent genes (positive values) or the amount of shared nucleotides between two overlapping genes (negative values).

Figure 2. Comparisons amongst the putative secondary structures for functional trnR (CGA) (A) and pseudogene-like trnS (AGN) (B) from the mitochondrial (mt) genome of Alpheus distinguendus, and the functional trnS (AGN) (C) from the mt genome of Panulirus ornatus. AU/GC pairs are indicated by dashes and GU bonds are indicated by dots. The nucleotide base located in the third codon positions of pseudogene-like trnS (AGN) is G instead of the usual base U, which is indicated by an arrow.

Figure 3. The secondary structures of the noncoding regions in the mt genome of Panulirus ornatus and Panulirus japonicus are labelled by A and B, respectively. The duplicate 89 bp sequences in the noncoding region of P. ornatus are compared in their secondary structures by arrows pointing to four numbers i.e. 1, 89, 428, and 516 sites in A, which correspond to Repeat sequence 1 and 2 (RS1 & RS2), in C.

Figure 4. Phylogenetic analyses of the nucleotide sequence (A) and amino acid datasets (B) derived from the Decapoda using the maximum likelihood (ML) analyses and Bayesian inferences (BI), respectively. Bayesian posterior probability (BPP) and bootstrap support (BP) of each node are denoted by the asterisk indicating the well-supported values (BPP = 1.00 and simultaneously BP> 90), or otherwise are presented directly. The phylogram shown is from BI except that the a’ clade was based on amino acid datasets from ML analyses. Scale bars represent 0.1 mutations per site.