As described above, the endosymbiont hypothesis proposes that mitochondrial genomes arose from an α-proteobacterium in an ancient cell-fusion event that occurred around 1.5 billion years ago. To explain the very small size of current mitochondrial DNAs, DNA sequences were envisaged to have been shed by the original genome, but with many of them subsequently integrating into what would become the nuclear genome. That prob ably happened comparatively quickly after the ancient cell-fusion event (at least when referenced against evolutionary timescales) and had the advantage of removing most of the genes from close exposure to harmful free radicals generated by the oxidative phosphorylation system.
In addition to this ancient sequence transfer between the two precursor genomes, there has also been evolutionarily recent and ongoing transfer of mitochondrial DNA sequences to the nuclear genome. Analysis of the human genome reference sequence using standard BLAST programs reveals over 750 nuclear sequences that are imperfect copies of mtDNA sequences, with sizes ranging from tens of nucleotides up to 14,654 nucleotides in length. The transferred mtDNA sequences have acquired inactivating mutations over time and so are sometimes described as nuclear mitochondrial pseudogenes, but they are more generally known as nuclear mitochondrial DNA sequences (NUMTs). In total they account for over 627 kb of the nuclear genome; see Table 1 for five prominent examples. Many human NUMT sequences are present in some haplotypes but not in others and so constitute insertion/deletion polymorphisms.
Table1. EXAMPLES OF NUCLEAR MITOCHONDRIAL SEQUENCES (NUMTs) THAT ARE CLOSELY RELATED IN SEQUENCE TO mtDNA
NUMTs are found in the nuclear genomes of a wide range of eukaryotes, and their existence means that even in complex eukaryotic cells, mitochondrial DNA sequences can somehow move to the nucleus, and then integrate into the nuclear genome. Integration appears to occur using the nonhomologous end-joining DNA repair mechanism that specializes in repairing double-strand DNA breaks. There is less certainty about how mitochondrial DNA sequences get to the nucleus, but it most likely begins with natural degradation of damaged or abnormal mitochondria, and release of mitochondrial DNA fragments into the cytoplasm.
The great majority of human NUMTs are not of recent origin (the oldest may have entered the nuclear genome about 60 million years ago). However, the insertion process is ongoing: a small proportion of NUMT loci are polymorphic, being present in some individuals but absent in others (note that the human genome reference sequence lacks some NUMTs present in the human population). And occasional de novo insertion of mtDNA sequences into the nuclear genome is known to disrupt gene expression, causing disease.
