We present entire sequences of two hymenopteran mitochondrial genomes and the major portion of three others. We combined these data with nine previously sequenced hymenopteran mitochondrial genomes. This allowed us to infer and analyze the evolution of the 67 mitochondrial gene rearrangements so far found in this order. All of these involve tRNA genes, whereas four also involve larger (protein-coding or ribosomal RNA) genes. We find that the vast majority of mitochondrial gene rearrangements are independently derived. A maximum of four of these rearrangements represent shared, derived organizations, whereas three are convergently derived. The remaining mitochondrial gene rearrangements represent new mitochondrial genome organizations. These data are consistent with the proposal that there are an enormous number of alternative mitochondrial genome organizations possible and that mitochondrial genome organization is, for the most part, selectively neutral. Nevertheless, some mitochondrial genes appear less mobile than others. Genes close to the noncoding region are generally more mobile but only marginally so. Some mitochondrial genes rearrange in a pattern consistent with the duplication/random loss model, but more mitochondrial genes move in a pattern inconsistent with this model. An increased rate of mitochondrial gene rearrangement is not tightly associated with the evolution of parasitism. Although parasitic lineages tend to have more mitochondrial gene rearrangements than nonparasitic lineages, there are exceptions (e.g., Orussus and Schlettererius). It is likely that only a small proportion of the total number of mitochondrial gene rearrangements that have occurred during the evolution of the Hymenoptera have been sampled in the present study.

The arrangement of tRNA genes at the junction of the cytochrome oxidase II and ATPase 8 genes was examined across a broad range of Hymenoptera. Seven distinct arrangements of tRNA genes were identified among a group of wasps that have diverged over the last 180 Myr (suborder Apocrita); many of the rearrangements represent evolutionarily independent events. Approximately equal proportions of local rearrangements, inversions, and translocations were observed, in contrast to vertebrate mitochondria, in which local rearrangements predominate. Surprisingly, homoplasy was evident among certain types of rearrangement; a reversal of the plesiomorphic gene order has arisen on three separate occasions in the Insecta, while the tRNA(H) gene has been translocated to this locus on two separate occasions. Phylogenetic analysis indicates that this gene translocation is real and is not an artifactual translocation resulting from the duplication of a resident tRNA gene followed by mutation of the anticodon. The nature of the intergenic sequences surrounding this region does not indicate that it should be especially prone to rearrangement; it does not generally have the tandem or inverted repeats that might facilitate this plasticity. Intriguingly, these findings are consistent with the view that during the evolution of the Hymenoptera, rearrangements increased at the same time that the rate of point mutations and compositional bias also increased. This association may direct investigations into mitochondrial genome plasticity in other invertebrate lineages.

No abstract