Critical experimental test of the possibility of "paternal leakage" of mitochondrial DNA

Most previous data suggesting maternal inheritance of mtDNA have come from single-generation mating experiments, and most of the analytical techniques utilized would not have detected paternal mtDNA molecules in progeny at levels less than about 5%. Long-term mating experiments, in which a fertile female lineage derived from hybridization of two species with distinguishable mtDNAs is backcrossed recurrently to the male parental species, provide an ideal opportunity to assess possible low-level paternal leakage. We have analyzed the 45- and 91-generation backcross progeny of such matings between two species of lepidopteran insects [Heliothis (Noctuidae)], using autoradiographic techniques that can detect rare mtDNA molecules in less than 1 part per 500. The analysis failed to detect any paternal mtDNA and sets an upper limit of paternal leakage at about 1 molecule per 25,000 per generation in this system.