A partial sequence of the mitochondrial cytochrome oxidase subunit I (COI) gene is widely used as a molecular marker for species identification in animals, also termed a DNA barcode. However, the presence of more than one sequence type in a single individual, also known as heteroplasmy, is one of the shortcomings of barcode identification. In this study, we examined the extent and divergence of COI heteroplasmy, including nuclear-encoded mitochondrial pseudogenes (NUMTs), at the genomic-DNA level from 13 insect species including orthopteran Anapodisma miramae, and a long fragment of mitochondrial DNA and cDNA from A. miramae as templates. When multiple numbers of clones originated from genomic DNA were sequenced, heteroplasmy was prevalent in all species and NUMTs were observed in five species. Long fragment DNA (∼13.5 kb) also is a source of heteroplasmic amplification, but the divergent haplotypes and NUMTs obtained from genomic DNA were not detected in A. miramae. On the other hand, cDNA was relatively heteroplasmy-free. Consistently, one dominant haplotype was always obtained from the genomic DNA-origin clones in all species and also from the long fragment- and cDNA-origin clones in the two tested individuals of A. miramae. Furthermore, the dominant haplotype was identical in sequence, regardless of the DNA source in A. miramae. Thus, one possible solution to avoid the barcoding problem in relationship to heteroplasmy could be the acquisition of multiple numbers of barcoding sequences to determine a dominant haplotype that can be assigned as barcoding sequence for a given species.
Keywords: Anapodisma miramae; COI; DNA barcoding; NUMTs; Orthoptera; heteroplasmy; mitochondrial DNA; pseudo-sequence.