Multilevel selection analysis reveals diversity in selfish mitochondrial genome dynamics

Abstract

AbstractMutant copies of mitochondrial DNA (mtDNA) can behave as selfish entities, propagating within organisms at the expense of host fitness. We previously developed experiments to separately measure the within-host proliferation, and host fitness cost, of selfish mtDNA. By focusing on a single mutant genome, the uaDf5 variant, our previous work raises the question of generalizability: do selfish genomes uniformly overcome a heavy fitness cost by proliferating to high frequency, or do their population dynamics vary? By applying a standardized multilevel selection analysis across a collection of mitochondrial mutants, we uncover variation in cheating strategies. Mutants that impose a heavy fitness cost nevertheless persist by outcompeting wildtype mtDNA within hosts. Conversely, some mutants fail to outcompete wildtype mtDNA but nevertheless persist, maintaining a frequency range at which host fitness cost is negligible. These findings suggest that the population dynamics of mutant mtDNA depend on the loci affected.