Quasi-Mendelian Paternal Inheritance of mitochondrial DNA: A notorious artifact, or anticipated mtDNA behavior?

A recent report by Luo et al (2018) in PNAS (DOI:10.1073/pnas.1810946115) presented evidence of biparental inheritance of mitochondrial DNA. The pattern of inheritance, however, resembled that of a nuclear gene. The authors explained this peculiarity with Mendelian segregation of a faulty gatekeeper gene that permits survival of paternal mtDNA in the oocyte. Three other groups (Vissing, 2019; Lutz-Bonengel and Parson, 2019; Salas et al, 2019), however, posited the observation was an artifact of inheritance of mtDNA nuclear pseudogenes (NUMTs), present in the father’s nuclear genome. We present justification that both interpretations are incorrect, but that the original authors did, in fact, observe biparental inheritance of mtDNA. Our alternative model assumes that because of initially low paternal mtDNA copy number these copies are randomly partitioned into nascent cell lineages. The paternal mtDNA haplotype must have a selective advantage, so ‘seeded’ cells will tend to proceed to fixation of the paternal haplotype in the course of development. We use modeling to emulate the dynamics of paternal genomes and predict their mode of inheritance and distribution in somatic tissue. The resulting offspring is a mosaic of cells that are purely maternal or purely paternal – including in the germline. This mosaicism explains the quasi-Mendelian segregation of the paternal mDNA. Our model is based on known aspects of mtDNA biology and explains all of the experimental observations outlined in Luo et. al., including maternal inheritance of the grand-paternal mtDNA.