Mitochondrial pseudogenes plague locust nuclear genomes

Mitochondrial DNA (mtDNA) sequences are commonly used in phylogenetic analyses and as DNA barcodes. However, recent studies have shown that the presence of nuclear copies of mtDNA sequences (numts) can mislead phylogenetic inferences and complicate species identification. Numts have been described in a variety of taxa including grasshoppers. We studied a region of the mitochondrial Cytochrome Oxidase 1 (COI) gene in two grasshopper species, Locusta migratoria (Lm) and Chortoicetes terminifera (Ct). We examined intraindividual variation in COI by analysing sequences from multiple clones of PCR products (114 and 110 clones in Lm and Ct, respectively). Neither the use of different DNA extraction protocols (Qiagen, Chelex and Alkaline lysis procedure) nor optimized PCR conditions successfully controlled for the presence of numts. In both locust species, we found (i) a prevalent haplotype matching with true mitocondrial COI sequences referenced in GenBank (36.8% and 28.2% of the clones in Lm and Ct, respectively), and (ii) numerous unique COIlike sequences (61.4% and 70% of the clones in Lm and Ct, respectively), hereafter referred as numts. Numts could be categorized as either proximal-numts (= 1% divergent) or distantnumts (= 10% divergent) based on their level of similarity to the true mitochondrial COI sequence. Intra-specific haplotype network analyses generally revealed that only proximalnumts connect to the primary mtCOI haplotype from the same individual (yielding a star-like network topology). Phylogenetic analysis of COI sequences from the Oedipodinae subfamily showed that numts were recently-derived and clustered within species¿ lineages. However, numts can be very similar across populations within species, and may pose a problem for phylogeographic analyses of populations and closely-related species. (Texte intégral