Phylogenetic relationships among nematodes have been revised in the past two decades using sequences of nuclear small subunit (SSU) rRNA gene [12–14]. The current working hypothesis of Nematoda phylogeny incorporated evidence from both morphology, ecology and nuclear SSU rRNA gene sequence analyses [14, 15]. Recently, Kang et al.  tested the hypothesis of the Nematoda phylogeny with the mt genome sequences of 25 species. Kang et al.  showed strong support to the monophyly of the class Chromadorea, represented by 16 species from the order Rhabditida; this study, however, could not establish the monophyly of the class Enoplea, represented by nine species from the subclass Dorylaimia. Park et al.  and Sultana et al.  expanded the taxon sampling and inferred phylogenies with mt genome sequences of 36 and 41 species of nematodes respectively. Park et al.  showed strong support for the monophylies of both Chromadorea and Enoplea. Sultana et al. , however, could not establish the monophyly of the class Enoplea in most of their analyses. A major difference between the mt genome phylogenies and the current working hypothesis of the Nematode phylogeny is on the monophyly of the suborder Spirurina, which includes Clade III nematodes proposed in Blaxter et al. . In the current working hypothesis, the suborder Spirurina is monophyletic, which includes the infraorders Ascaridomorpha, Spiruromorpha and Oxyuridomorpha, whereas the infraorder Rhabditomorpha is in the suborder Rhabditina [13, 14]. Further, Ascaridomorpha is most closely related to Spiruromorpha, whereas Oxyuridomorpha is sister to the group that includes Ascaridomorpha + Spiruromorpha. In the phylogenies inferred from mt genome sequences, however, Ascaridomorpha is most closely related to Rhabditomorpha; in most analyses, Oxyuridomorpha is sister to Ascaridomorpha + Rhabditomorpha, whereas Spiruromorpha is sister to the group that contains Ascaridomorpha, Rhabditomorpha and Oxyuridomorpha. Kang et al. , Park et al.  and Sultana et al.  contained much less taxa compared to previous phylogenetic studies on nematodes with nuclear SSU rRNA gene sequences (e.g. >200 taxa in Meldal et al. ). The novel phylogenetic relationship inferred from mt genome sequences, thus, needs to be tested further with more taxa from a wide range of nematode lineages.
In the present study, we sequenced the complete mt genomes of three Ascaridia species from the family Ascaridiidae of the superfamily Heterakoidea, which was not represented in previous studies (Kang et al. ; Park et al. ; Sultana et al. ). Furthermore, we inferred the phylogenetic relationships among 65 nematode species, for which complete mt genomes have been sequenced to date. Our analyses support the division of nematodes into two classes, Chromadorea and Enoplea, represented by the order Rhabditida and the subclass Dorylaimia, respectively, in the present study. Both the Rhabditida and the Dorylaimia were monophyletic with strong support, regardless the methods of phylogenetic analysis used. Within the order Rhabditida, the suborder Rhabditina was monophyletic; the other two suborders, Spirurina and Tylenchina, however, were both paraphyletic in all of the three phylogenetic analyses in this study. Monophyly of the suborder Spirurina was well supported in nuclear SSU rRNA gene sequence analyses [12, 14], but was rejected in mt genome sequence analyses in Kang et al. , Park et al. , Sultana et al.  and the present study. Phylogenetic analyses of mt genome sequences in all of these four studies support strongly a close relationship between Ascaridomorpha and Rhabditomorpha to the exclusion of Oxyuridomorpha and Spiruromorpha. Apparently, additional markers for phylogenetic inference are required to resolve the controversy on the Spirurina between mt genome sequence phylogeny and nuclear SSU rRNA phylogeny.
Gene arrangement in mt genomes provides a source of information for phylogenetic inference, independent from mt genome sequences . It is noteworthy that 10 species of Ascaridomorpha and 20 species of Rhabditomorpha included in our phylogenetic analyses share a common pattern of gene arrangement, GA3, in their mt genomes (Figure 2). GA3 is also present in Pristionchus pacificus, the only species from the infraorder Diplogasteromorpha, which is closely related to the species from the superfamily Rhabitoidea of the Rhabditomorpha (Figures 4, 5, 6). Indeed, GA3 is the most common pattern of mt gene arrangement observed in nematodes to date; nevertheless, this pattern is present only in species of the Ascaridomorpha, the Rhabditomorpha and the Diplogasteromorpha. The possibility that GA3 is ancestral to the phylum Nematoda can be rejected with confidence thanks to the discovery of mt gene arrangement pattern GA23 in three Trichinellida species [29, 30]. GA23 shares obvious similarity with that of many arthropods and would resemble the ancestral pattern of mt gene arrangement of the phylum Nematoda more than any other patterns found in the nematodes. There is no evidence either that GA3 is ancestral to the class Chromadorea. Given the strong support to the close relationship between Ascaridomorpha and Rhabditomorpha + Diplogasteromorpha indicated by mt genome sequence analyses in the present study, the most parsimonious explanation is that GA3 is a shared derived character (i.e. synapomorphy) of the clade Ascaridomorpha + Rhabditomorpha + Diplogasteromorpha. The other six patterns, GA1, GA2, GA4, GA5, GA6 and GA25, observed in either Ascaridomorpha and Rhabditomorpha or in species closely related to these two infraorders, can be derived from GA3 by only a few rearrangement events (Figure 2), e.g. GA1, observed in the three Ascaridia species, can be converted from GA3 by three rearrangement events (see above).
Monophyly of the suborder Tylenchina could not be established previously in nuclear SSU RNA gene sequence analyses . No species from this suborder were included in two previous analyses of mt genome sequences [23, 24]. The present study included four species from the Tylenchina and indicated the paraphyly of this suborder. The positions of the four species of the Tylenchina change depending on the methods of phylogenetic inference used in this study. For instance, Steinernema carpocapsae is sister to the suborder Rhabditina in Bayesian analysis (Figure 4) and MP analysis (Figure 6); it is, however, sister to Ascaridomorpha + Rhabditomorpha + Diplogasteromorpha in ML analysis (Figure 5). Further test with more species from this suborder apparently is needed in order to clarify the phylogenetic position of the suborder Tylenchina. Indeed, a very recent study by Sultana et al.  included six species from the Tylenchina and showed the paraphyly of this suborder with strong support. The paraphyly of the infraorder Rhabditomorpha with respect to the Diplogasteromorpha will also need further test with expanded taxon sampling as only one species, Pristionchus pacificus, from the Diplogasteromorpha was included in the present study. No species from the subclass Enoplia, which has two orders and eight suborders , was included in our analyses. For the subclass Chromadoria, all of the species included in this study were from the order Rhabditida whereas the other five orders, Plectida, Araeolaimida, Monhysterida, Desmodorida, and Chromadorida, were not represented. Expanding taxon sampling from these lineages of nematodes is clearly the next step for phylogenetic studies of nematodes with mt genome sequences.