A mitochondria-specific mutational signature of aging: increased rate of A > G substitutions on the heavy strand

Abstract

The mutational spectrum of the mitochondrial DNA (mtDNA) does not resemble any of the known mutational signatures of the nuclear genome and variation in mtDNA mutational spectra between different organisms is still incomprehensible. Since mitochondria are responsible for aerobic respiration, it is expected that mtDNA mutational spectrum is affected by oxidative damage. Assuming that oxidative damage increases with age, we analyse mtDNA mutagenesis of different species in regards to their generation length. Analysing, (i) dozens of thousands of somatic mtDNA mutations in samples of different ages (ii) 70053 polymorphic synonymous mtDNA substitutions reconstructed in 424 mammalian species with different generation lengths and (iii) synonymous nucleotide content of 650 complete mitochondrial genomes of mammalian species we observed that the frequency of AH > GH substitutions (H: heavy strand notation) is twice bigger in species with high versus low generation length making their mtDNA more AH poor and GH rich. Considering that AH > GH substitutions are also sensitive to the time spent single-stranded (TSSS) during asynchronous mtDNA replication we demonstrated that AH > GH substitution rate is a function of both species-specific generation length and position-specific TSSS. We propose that AH > GH is a mitochondria-specific signature of oxidative damage associated with both aging and TSSS.

Figure 1.

A_H> G_H mtDNA mutational gradient is increasing with the sample age. (A) Asynchronous replication of mtDNA is associated with a long time spent single stranded (TSSS) by the parental heavy strand. TSSS in turn is associated with the high frequency of two the most common mtDNA transitions: C_H> T_H and A_H> G_H (daughter heavy strand: dashed black line; parental heavy strand: bold thickening black line reflecting the TSSS; daughter light strand: dashed gray line; parental light strand: solid gray line; OH: origin of replication of daughter heavy strand, OL: origin of replication of daughter light strand). (B) Gradients of C_H> T_H and A_H> G_H mutations along the major arc of mtDNA are more pronounced in humans versus old mice and in old mice versus young mice. Both intercepts and slopes are increasing with sample age. (C) A_H> G_H substitution rate is increasing faster in aged samples. Upper panel: barplots visualize the slopes of the linear regressions between the mutation frequency and TSSS. Middle panel: A_H> G_H slopes increase faster with age as compared to C_H> T_H slopes. Boxplots are based on the ratio of slopes derived from 1000 bootstrapped samples. Bottom panel: frequency of A_H> G_H in the total mutational spectrum is increasing with age (P-values from all three pairwise comparisons are less than 1.583e−08, Mann–Whitney U test). ‘***’ marks P values < 0.001.

Figure 2.

Variation in neutral mtDNA mutational spectrum of mammals is driven by the generation length. (A) An average mtDNA mutational spectrum of mammalian species (N = 611). Mutational spectrum is a probability of each nucleotide to mutate to each other based on the observed and normalized frequencies of twelve types of nucleotide substitutions in four-fold degenerate synonymous sites of all available within-species polymorphisms of mtDNA protein-coding genes. (B) Mutational spectra vary with species-specific generation length (N = 424). A_H> G_H is the type of substitutions, frequency of which stronger correlated with the generation length. It shows approximately two-fold difference between the mammalian with very short and very long generation length. (C) The principal component analysis (PCA) of mtDNA mutational spectra of mammalian species (N = 424). Left panel: the biplot of the principal component analyses (first and the second components explains 16% and 12% of variation correspondingly). C_H> T_H has the highest loading on the first principal component while A_H> G_H has the highest loading on the second principal component. Note that we plotted negative PC2 to make it positively correlated with generation length. Right panel: The second principal component correlates with the generation length in mammals. Generation length is color-coded from dark green (the shortest generation length) to light green (the longest generation length).

Figure 3.

The long-term effect of the mutational bias: neutral nucleotide content in mammalian species. (A) A correlation between the expected (obtained in simulations) and observed neutral nucleotide content of mammals with very short and very long generation length. Due to an excess of A_H> G_H substitutions in long-lived mammals (marked by red circles) they are more A_H poor and G_H rich for both expected and observed values. A location of all data points (red and blue circles) near the diagonal shows that mtDNA of mammals is close enough to a neutral equilibrium. However, short-lived species (marked by blue circles) are even closer to the diagonal (horizontal dotted blue line towards the diagonal are shorter than the red dotted lines), suggesting that they are evolving faster towards the neutral equilibrium. (B) Nucleotide frequencies in neutral sites of all 13 protein-coding genes as a function of generation length—fraction of A_H is decreasing while fraction of G_H is increasing (N = 650). (C) Structure of mtDNA of two mammalian species with extreme generation lengths: a honey possum and a whale. Upper panel: frequencies of A_H (red) and G_H (grey) nucleotides along the major arc of mtDNA of the most short-lived (honey possum) and the most long-lived (whale) mammalian species from our dataset. Each bar represents the nucleotide frequency in a 20-nucleotide window. In both mammals, A_H is decreasing and G_H is increasing along the major arc of mtDNA: from the bottom left (origin of replication of light strand) to the top right (origin of replication of heavy strand). However, additionally to the gradient, mtDNA of a whale has an integral, genome-wide, deficit of A_H and excess of G_H - a signature of an increased generation length. Bottom panel: heatmaps visualize asymmetry of the codon usage of 12 protein-coding genes (all except ND6). Whale is more contrast than honey possum in terms of an asymmetry driven by the age-related T_L> C_L (A_H> G_H) substitutions. Heatmaps of both species are equally contrasted in terms of an asymmetry driven by G_L> A_L (C_H> T_H) substitutions, which have high and similar (not age related) substitution rate in both species.

Figure 4.

(A) Changes in nucleotide content along mtDNA of short- and long-lived mammals (N = 650). All genes (except for ND6) located in the major arc are ranked according to the time spent single stranded: from COX1 to CYTB. Pairs of boxplots for each gene represent G_HA_H skew for short- and long-lived mammals splitted by the median generation length. G_HA_H is increasing with both gene-specific TSSS and the species-specific generation length. (B) A visual summary of the main finding: A_H> G_H substitution rate (marked as red gradient) is increasing with both gene-specific TSSS and the species-specific generation length. The effect size of the G_L is comparable with the effect size of TSSS. C_H> T_H substitution rate (marked as grey gradient) is sensitive to TSSS only.