NUMEROUS TRANSPOSED SEQUENCES OF MITOCHONDRIAL CYTOCHROME-OXIDASE I-II IN APHIDS OF THE GENUS SITOBION (HEMIPTERA, APHIDIDAE)

Polymerase chain reaction (PCR) products corresponding to 803 bp of th e cytochrome oxidase subunits I and II region of mitochondrial DNA (mt DNA COI-II) were deduced to consist of multiple haplotypes in three Si tobion species. We investigated the molecular basis of these observati ons. PCR products were cloned, and six clones from one individual per species were sequenced. In each individual, one sequence was found com monly, but also two or three divergent sequences were seen. The diverg ent sequences were shown to be nonmitochondrial by sequencing from pur ified mtDNA and Southern blotting experiments. All seven nonmitochondr ial clones sequenced to completion were unique. Nonmitochondrial seque nces have a high proportion of unique sites, and very few characters a re shared between nonmitochondrial clones to the exclusion of mtDNA. F rom these data, we infer that fragments of mtDNA have been transposed separately (probably into aphid chromosomes), at a frequency only know n to be equalled in humans. The transposition phenomenon appears to oc cur infrequently or not at all in closely related genera and other aph ids investigated. Patterns of nucleotide substitution in mtDNA inferre d over a parsimony tree are very different from those in transposed se quences. Compared with mtDNA, nonmitochondrial sequences have less cod on position bias, more even exchanges between A, G, C and T, and a hig her proportion of nonsynonymous replacements. Although these data are consistent with the transposed sequences being under less constraint t han mtDNA, changes in the nonmitochondrial sequences are not random: t here remains significant position bias, and probable excesses of synon ymous replacements and of conservative inferred amino acid replacement s. We conclude that a proportion of the inferred change in the nonmito chondrial sequences occurred before transposition. We believe that Sit obion aphids (and other species exhibiting mtDNA transposition) may be important for studying the molecular evolution of mtDNA and pseudogen es. However, our data highlight the need to establish the true evoluti onary relationships between sequences in comparative investigations.