Mitochondrial heteroplasmy and paternal leakage in natural populations of Silene vulgaris, a gynodioecious plant.

It is currently thought that most angiosperms transmit their mitochondrial genomes maternally. Maternal transmission limits opportunities for genetic heterogeneity (heteroplasmy) of the mitochondrial genome within individuals. Recent studies of the gynodioecious species Silene vulgaris and Silene acaulis, however, document both direct and indirect evidence of mitochondrial heteroplasmy, suggesting that the mitochondrial genome is at times transmitted via paternal leakage. This heteroplasmy allows the generation of multi-locus recombinants, as documented in recent studies of both species. A prior study that employed quantitative PCR (q-PCR) on a limited sample provided direct evidence of heteroplasmy in the mitochondrial gene atp1 in S. vulgaris. Here, we apply the q-PCR methods to a much larger sample and extend them to incorporate the study of an additional atp1 haplotype along with two other haplotypes of the mitochondrial gene cox1 to evaluate the origin, extent, and transmission of mitochondrial genome heteroplasmy in S. vulgaris. We first calibrate our q-PCR methods experimentally and then use them to quantify heteroplasmy in 408 S. vulgaris individuals sampled from 22 natural populations located in Virginia, New York, and Tennessee. Sixty-one individuals exhibit heteroplasmy, including five that exhibited the joint heteroplasmy at both loci that is a prerequisite for effective recombination. The heteroplasmic individuals were distributed among 18 of the populations studied, demonstrating that heteroplasmy is a widespread phenomenon in this species. Further, we compare mother and offspring from 71 families to determine the rate of heteroplasmy gained and lost via paternal leakage and vegetative sorting across generations. Of 17 sibships exhibiting cox1 heteroplasmy and 14 sibships exhibiting atp1 heteroplasmy, more than half of the observations of heteroplasmy are generated via paternal leakage at the time of fertilization, with the rest being inherited from a heteroplasmic mother. Moreover, we show that the average paternal contribution during paternal leakage is about 12%. These findings are surprising, given that the current understanding of gynodioecy assumes that mitochondrial cytoplasmic male sterility elements are strictly maternally inherited. Knowledge of the dynamics of mitochondrial populations within individuals plays an important role in understanding the evolution of gynodioecy, and we discuss our findings within this context.