Factors affecting the relative abundance of nuclear copies of mitochondrial DNA (numts) in hominoids

Although nuclear copies of mitochondrial DNA (numts) can originate from any portion of the mitochondrial genome, evidence from humans suggests that more variable parts of the mitochondrial genome, such as the mitochondrial control region (MCR), are under-represented in the nucleus. This apparent deficit might arise from the erosion of sequence identity in numts originating from rapidly evolving mitochondrial sequences. However, the extent to which mitochondrial sequence properties impacts the number of numts detected in genomic surveys has not been evaluated. In order to address this question, we: (1) conducted exhaustive BLAST searches of MCR numts in three hominoid genomes; (2) assessed numt prevalence across the four MCR sub-domains (HV1, CCD, HV2, and MCR(F)); (3) estimated their insertion rates in great apes (Hominoidea); and (4) examined the relationship between mitochondrial DNA variability and numt prevalence in sequences originating from MCR and coding regions of the mitochondrial genome. Results indicate a marked deficit of numts from HV2 and MCR(F) MCR sub-domains in all three species. These MCR sub-domains exhibited the highest proportion of variable sites and the lowest number of detected numts per mitochondrial site. Variation in MCR insertion rate between lineages was also observed with a pronounced burst in recent integrations within chimpanzees and orangutans. A deficit of numts from HV2/MCR(F) was observed regardless of age, whereas HV1 is under-represented only in older numts (>25 million years). Finally, more variable mitochondrial genes also exhibit a lower identity with nuclear copies and because of this, appear to be under-represented in human numt databases.

A comparative approach shows differences in patterns of numt insertion during hominoid evolution

Nuclear integrations of mitochondrial DNA (numts) are widespread among eukaryotes, although their prevalence differs greatly among taxa. Most knowledge of numt evolution comes from analyses of whole-genome sequences of single species or, more recently, from genomic comparisons across vast phylogenetic distances. Here we employ a comparative approach using human and chimpanzee genome sequence data to infer differences in the patterns and processes underlying numt integrations. We identified 66 numts that have integrated into the chimpanzee nuclear genome since the human-chimp divergence, which is significantly greater than the 37 numts observed in humans. By comparing these closely related species, we accurately reconstructed the preintegration target site sequence and deduced nucleotide changes associated with numt integration. From >100 species-specific numts, we quantified the frequency of small insertions, deletions, duplications, and instances of microhomology. Most human and chimpanzee numt integrations were accompanied by microhomology and short indels of the kind typically observed in the nonhomologous end-joining pathway of DNA double-strand break repair. Human-specific numts have integrated into regions with a significant deficit of transposable elements; however, the same was not seen in chimpanzees. From a separate data set, we also found evidence for an apparent increase in the rate of numt insertions in the last common ancestor of humans and the great apes using a polymerase chain reaction-based screen. Last, phylogenetic analyses indicate that mitochondrial-numt alignments must be at least 500 bp, and preferably >1 kb in length, to accurately reconstruct hominoid phylogeny and recover the correct point of numt insertion.