Search

Orthoptera is the most diverse order of polyneopterans, and the forewing and hindwing of its members exhibit extremely variability from full length to complete loss in many groups; thus, this order provides a good model for studying the effects of insect flight ability on the evolutionary constraints on and evolutionary rate of the mitochondrial genome. Based on a data set of mitochondrial genomes from 171 species, including 43 newly determined, we reconstructed Orthoptera phylogenetic relationships and estimated the divergence times of this group. The results supported Caelifera and Ensifera as two monophyletic groups, and revealed that Orthoptera originated in the Carboniferous (298.997 Mya). The date of divergence between the suborders Caelifera and Ensifera was 255.705 Mya, in the late Permian. The major lineages of Acrididae seemed to have radiated in the Cenozoic, and the six patterns of rearrangement of 171 Orthoptera mitogenomes mostly occurred in the Cretaceous and Cenozoic. Based on phylogenetic relationships and ancestral state reconstruction, we analysed the evolutionary selection pressure on and evolutionary rate of mitochondrial protein-coding genes (mPCGs). The results indicated that during approximately 300 Mya of evolution, these genes experienced purifying selection to maintain their function. Flightless orthopteran insects accumulated more non-synonymous mutations than flying species and experienced more relaxed evolutionary constraints. The different wing types had different evolutionary rates, and the mean evolutionary rate of Orthoptera mitochondrial mPCGs was 13.554 × 10−9 subs/s/y. The differences in selection pressures and evolutionary rates observed between the mitochondrial genomes suggested that functional constraints due to locomotion play an important role in the evolution of mitochondrial DNA in orthopteran insects with different wing types.

Although mitogenomes are useful tools for inferring evolutionary history, only a few representative ones can be used for most Ensifera lineages. Thirty-two ensiferan mitogenomes were determined using ABI Sanger sequencing and standard primer walking of 2–3 overlapping Long-PCR fragments, or Illumina® HiSeq2000 for “shotgun” sequenced long-PCR-amplified mitochondrial or total genomic DNA. Six patterns of gene arrangements, including the novel trnR-trnSAGN-trnA-trnN-trnG-nad3 in Lipotactes tripyrga (Lipotactinae), were identified from 59 ensiferan mitogenomes. The results suggest that trnM-trnI-trnQ and trnA-trnR-trnE-trnSAGN-trnN-trnF rearrangements might be a shared derived character in Pseudophyllinae and Gryllidae, respectively. We found base composition biases in our dataset, which potentially complicate the inference of higher-level ensiferan phylogeny. Site-heterogeneous Bayesian inference (BI) and site-homogeneous maximum likelihood (ML) analyses recovered all ensiferan superfamilies as monophyletic. The site-homogeneous BI analysis failed to recover the monophyly of Stenopelmatoidea. As Schizodactyloidea was only represented by Comicus campestris, its monophyly could not be tested. In the Triassic/Jurassic boundary, Ensifera diverged into grylloid and non-grylloid clades. All analyses confirmed Grylloidea and Gryllotalpoidea as sister groups. Site-heterogeneous BI analysis found Schizodactyloidea as the most basal lineage and sister to the clade formed by Grylloidea and Gryllotalpoidea, but the site-homogeneous analyses placed it basally to the non-grylloid clade and recovered a sister relationship between Tettigonioidea and (Hagloidea, Rhaphidophoroidea, Stenopelmatoidea), although this clade had a low support. The site-heterogeneous BI analysis found Tettigonioidea and Hagloidea were sister groups (posterior probability (PP) = 0.99), Stenopelmatoidea was sister to (Tettigonioidea, Hagloidea) (PP > 0.91), and Rhaphidophoroidea was basal to the non-grylloid clade. At a lower level, all analyses divided Tettigonioidea into Phaneropteridae and Tettigoniidae.

The phylogenetic relationships of 39 species of Eneopterinae crickets are reconstructed using four molecular markers (16S rRNA, 12S rRNA, cytochrome b, 18S rRNA) and a large morphological data set. Phylogenetic analysis via direct optimisation of DNA sequence data using parsimony as optimality criterion is done for six combinations of weighting parameter sets in a sensitivity analysis. The results are discussed in a twofold purpose: first, in term of significance of the molecular markers for phylogeny reconstruction in Ensifera, as our study represents the first molecular phylogeny performed for this insect suborder at this level of diversity; second, in term of corroboration of a previous phylogeny of Eneopterinae, built on morphological data alone. The four molecular markers all convey phylogenetic signal, although variously distributed on the tree. The monophyly of the subfamily, that of three over five tribes, and of 10 over 13 genera, are recovered. Finally, previous hypotheses on the evolution of acoustic devices and signals in the Eneopterinae clade are briefly tested, and supported, by our new data set.

Representatives of the Orthopteran suborder Ensifera (crickets, katydids, and related insects) are well known for acoustic signals produced in the contexts of courtship and mate recognition. We present a phylogenetic estimate of Ensifera for a sample of 51 taxonomically diverse exemplars, using sequences from 18S, 28S, and 16S rRNA. The results support a monophyletic Ensifera, monophyly of most ensiferan families, and the superfamily Gryllacridoidea which would include Stenopelmatidae, Anostostomatidae, Gryllacrididae, and Lezina. Schizodactylidae was recovered as the sister lineage to Grylloidea, and both Rhaphidophoridae and Tettigoniidae were found to be more closely related to Grylloidea than has been suggested by prior studies. The ambidextrously stridulating haglid Cyphoderris was found to be basal (or sister) to a clade that contains both Grylloidea and Tettigoniidae. Tree comparison tests with the concatenated molecular data found our phylogeny to be significantly better at explaining our data than three recent phylogenetic hypotheses based on morphological characters. A high degree of conflict exists between the molecular and morphological data, possibly indicating that much homoplasy is present in Ensifera, particularly in acoustic structures. In contrast to prior evolutionary hypotheses based on most parsimonious ancestral state reconstructions, we propose that tegminal stridulation and tibial tympana are ancestral to Ensifera and were lost multiple times, especially within the Gryllidae.

The phylogenetic utility of mitochondrial genomes (mtgenomes) is examined using the framework of a preliminary phylogeny of Orthoptera. This study presents five newly sequenced genomes from four orthopteran families. While all ensiferan and polyneopteran taxa retain the ancestral gene order, all caeliferan lineages including the newly sequenced caeliferan species contain a tRNA rearrangement from the insect ground plan tRNA(Lys)(K)-tRNA(Asp)(D) swapping to tRNA(Asp) (D)-tRNA(Lys) (K) confirming that this rearrangement is a possible molecular synapomorphy for this suborder. The phylogenetic signal in mtgenomes is rigorously examined under the analytical regimens of parsimony, maximum likelihood and Bayesian inference, along with how gene inclusion/exclusion, data recoding, gap coding, and different partitioning schemes influence the phylogenetic reconstruction. When all available data are analyzed simultaneously, the monophyly of Orthoptera and its two suborders, Caelifera and Ensifera, are consistently recovered in the context of our taxon sampling, regardless of the optimality criteria. When protein-coding genes are analyzed as a single partition, nearly identical topology to the combined analyses is recovered, suggesting that much of the signals of the mtgenome come from the protein-coding genes. Transfer and ribosomal RNAs perform poorly when analyzed individually, but contribute signal when analyzed in combination with the protein-coding genes. Inclusion of third codon position of the protein-coding genes does not negatively affect the phylogenetic reconstruction when all genes are analyzed together, whereas recoding of the protein-coding genes into amino acid sequences introduces artificial resolution. Over-partitioning in a Bayesian framework appears to have a negative effect in achieving convergence. Our findings suggest that the best phylogenetic inferences are made when all available nucleotide data from the mtgenome are analyzed simultaneously, and that the mtgenome data can resolve over a wide time scale from the Permian (approximately 260 MYA) to the Tertiary (approximately 50 MYA).