Nuclear pseudogenes of mtDNA (NUMTS) suggest repeated distant inter-species hybridization among direct human ancestors

Introduction: Increasingly, the emergence and evolution of our species is being tied to genetic exchange between divergent lineages within ~1Ma (e.g., Neanderthals, Denisovans). However, little is known about genetic exchange during earlier (pre-1Ma) human evolution and between more divergent lineages. Results: We present evidence of hybridization within human lineage, show that it likely happened between highly divergent (~4.5My) lineages, more than once. We use analysis of nuclear pseudogenes of mtDNA (“NUMTs”). NUMTs are considered “mtDNA fossils”, as they preserve sequences of ancient mtDNA because mutational rate in the nucleus is much lower than in mtDNA. We demonstrate that a NUMT on human chromosome 5, which is shared by chimpanzee and gorilla, had descended from a mitochondrial genome that had been divergent from our ancestor’s mtDNA by ~4.5% at the time of pseudogene insertion. This implies that this pseudogene should have been inserted in a hominid that at that time had been diverged by about 4.5My of evolution from the hominid that at that time carried our mtDNA lineage. In order for this pseudogene and our mtDNA to end up in the same body, these two hominids should have mated with each other. The large divergence implies a distant interspecies (or even inter-generic) hybridization. Additionally, analysis of two other NUMTs (on Chr11 and Chr7) suggests that hybridization events occurred repeatedly. To exclude the large ancestral population size effect we show that mtDNA divergence in extant ape populations does not depend on population size. Discussion: It is thought that within mammals, it takes ~2-4My to establish reproductive isolation. However, fertile inter-generic hybrids have been documented among several primates, separated by ca. 4My. Very recently, hybridization between Colobine genera separated by ~5 My was reported to involve a NUMT scenario similar to what we had proposed human ancestors. Interestingly, phylogenic analysis consistently places the chr5 NUMT insertion around the time of the Homo/Pan split. Intriguingly, certain hominin fossils of that epoch have been interpreted alternately as more human-like or more ape-like. Such morphological mosaicisity could potentially be explained by hybridization. Fixation of NUMTs in question within population should have been rather efficient, since these pseudogenes appear to have been fixed in more than one population. Thus their spread across populations might have been driven by selection. Indeed, NUMTs on chr5 and chr11 are located in 3’ regions of functional genes. Most intriguingly, Ps11 is located 3’ to the RNF141/ZNF230 gene, essential for spermatogenesis. NUMT might have served as an expression modifier for RNF141, resulting in reproductive advantage. Indeed, RNF141 demonstrates selectively driven expression shift in testis of the ancestor of hominines.