Mitochondrial mutation rate revisited: hot spots and polymorphism

• Published in: Nature genetics Vol. 18; no. 2; pp. 109 - 110
• Main Authors: Jazin, Elena, Soodyall, Himla, Jalonen, Paula, Lindholm, Eva, Stoneking, Mark, Gyllensten, Ulf
• Format: Journal Article
• Language: English
• Published: United States 01.02.1998
• Subjects: DNA, Mitochondrial - genetics; Humans; Meiosis; Mitochondrial Myopathies - genetics; Point Mutation; Polymorphism, Genetic
• ISSN: 1061-4036; 1546-1718
• DOI: 10.1038/ng0298-109

Parsons et al. report a mutation rate for the mtDNA control region (CR) of 1.5-2.5/base pair/Myr, roughly 20 times higher than the substitution rate estimated from phylogenetic analysis. Our results from a similar size study are not consistent with that of Parsons et al. We analysed positions 1-370 of the mtDNA CR in 33 large maternal lineages, corresponding to a total of 288 meiotic (or generational) events. No homoplasmic point mutations were found, yielding an estimate of the mutation rate (<0.46 (99% CI=0.0-1.52) /basepair/Myr) that is significantly lower (Fisher's exact Test P=0.032) than that reported by Parsons et al. A number of additional studies of the mutation rate of mtDNA in families are also available. For example, Soodyall et al. did not find any homoplasmic mutations among 108 transmissions from mother to offspring, while Bendall et al. reported 2 substitutions in 170 transmissions in the second hypervariable region (HVII) and Howell et al., 1 in 81 transmissions. One possibility for the difference in mutation rate between the studies is that fewer transmissions were actually examined in the other family studies compared to Parsons et al., increasing the probability that reversions may lower the estimate of mutation rate. However, the studies do not appear to differ substantially with respect to the percentage of transmissions examined (60-70%). A second possibility is heterogeneity in mutation rate among families, due either to nuclear factors influencing the replication/repair machinery or sequence polymorphisms that render the DNA structure more liable to mutations. For example, we note that the majority of point mutations reported by Parsons et al. were found in only one (AFDIL) of the four sets of families. We consider heterogeneity of mutation rate among families unlikely, because the number of maternal lineages examined is not substantially smaller in our study than in Parsons et al., and the population origin is at least partly the same. A final possibility is that the disease state of the individuals may affect the mutation rate. Several of the materials examined were collected from families with various diseases. In particular, the pedigree examined by Howell et al. segregates with the mitochondrial disease Leber's hereditary optic neuroretinopathy (LHON). Overall, there were no obvious relationships between the observed mutation rate and disease state in these studies. Because we are unable to identify any factor that can explain the higher rate in the data of Parsons et al., it appears appropriate to consider the estimate of the pooled data as more accurate than that of any individual set of data. The pooled data from studies in which the same portion of the CR was analysed yield a mutation rate of 7/804 events, corresponding to 1.17 (99% CI=0.15-2.2)/per base pair/Myr. This interval is clearly not significantly different from the substitution rate estimated from phylogenetic studies of 0.15 (99% CI=0-0.30).