Selfish Mitochondrial DNA Proliferates and Diversifies in Small, but not Large, Experimental Populations of Caenorhabditis briggsae

Abstract

Evolutionary interactions across levels of biological organization contribute to a variety of fundamental processes including genome evolution, reproductive mode transitions, species diversification, and extinction. Evolutionary theory predicts that so-called "selfish" genetic elements will proliferate when the host effective population size (Ne) is small, but direct tests of this prediction remain few. We analyzed the evolutionary dynamics of deletion-containing mitochondrial DNA (ΔmtDNA) molecules, previously characterized as selfish elements, in six different natural strains of the nematode Caenorhabditis briggsae allowed to undergo experimental evolution in a range of population sizes (N = 1, 10, 100, and 1,000) for a maximum of 50 generations. Mitochondrial DNA (mtDNA) was analyzed for replicate lineages at each five-generation time point. Ten different ΔmtDNA molecule types were observed and characterized across generations in the experimental populations. Consistent with predictions from evolutionary theory, lab lines evolved in small-population sizes (e.g., nematode N = 1) were more susceptible to accumulation of high levels of preexisting ΔmtDNA compared with those evolved in larger populations. New ΔmtDNA elements were observed to increase in frequency and persist across time points, but almost exclusively at small population sizes. In some cases, ΔmtDNA levels decreased across generations when population size was large (nematode N = 1,000). Different natural strains of C. briggsae varied in their susceptibilities to ΔmtDNA accumulation, owing in part to preexisting compensatory mtDNA alleles in some strains that prevent deletion formation. This analysis directly demonstrates that the evolutionary trajectories of ΔmtDNA elements depend upon the population-genetic environments and molecular-genetic features of their hosts.

Keywords: experimental evolution; genetic conflict; mitochondrial DNA; nematode; population size; selfish genetic elements.

Fig. 1.

— Phylogenetic relationships based on mtDNA of the six C. briggsae natural strains used for experimental evolution progenitors (A) and schematic of the ΔmtDNA-C locus (B). In (A), Roman numerals indicate the intraspecific mtDNA clade designations defined in Raboin et al. (2010). Clade I is also commonly referred to as the “tropical” clade, Clade II as the temperate clade, and Clade III as the “equatorial” Clade (Cutter et al. 2006). The blue branches lead to strains that encode the putative compensatory mutations associated with the ΔmtDNA-C locus (Howe and Denver 2008). In (B), schematics of the nad5 deletion region are shown; Roman numerals on the left indicate the intraspecific clades (sensu Raboin et al. 2010) in which different states are observed. The arrows indicate positions of the 21-bp direct repeats associated with deletion formation. The dashed line indicates DNA sequences missing in ΔmtDNA-C molecules. The blue arrow indicates the position of the direct repeat bearing putative compensatory mutations (observed in some Clade II strains). Clade III strains lack Ψnad5-2 elements and associated deletions, as indicated in the bottom mtDNA gene model. All Clade I strains and some Clade II strains have perfect direct repeats.

Fig. 2.

— Levels of evolution schematic. The diagram illustrates hierarchical biological levels of organization in the study system, from laboratory C. briggsae populations to mtDNA molecules.

Fig. 3.

— Locations of newly detected ΔmtDNA molecules. The numbers in the rectangles correspond with the numbers assigned to the deletions in table 1. Genes are named in the diagram and ψ indicates the ψnad5-2 element. One variant, ΔmtDNA-9, is not shown because its deletion boundaries are on the opposite side of the mtDNA chromosome (table 2). The arrowheads show the positions of primers used; black arrowheads indicate original primer set for analyzing ΔmtDNA-C, CbMt_01F-58 R (see Materials and Methods). The gray arrowheads indicate the new primers (102 F, 103 F, 105 R, and 106 R) were designed to discriminate between ΔmtDNA-C and new ΔmtDNA types.

Fig. 4.

— Detection and characterization of ΔmtDNA. Electrophoresis analysis of mtDNA length variation in generations 10–50 (lanes 1–9, white numbers across the lanes label the generation) for a N = 1 sample from C. briggsae strain HK104. CM+ indicates a strain with a putative compensatory mutation in the Ψnad5-2 element (see fig. 1); CM− indicates a strain that lacks a compensatory mutation in Ψnad5-2. The 1 kb+ molecular marker (Invitrogen) is shown in the far right lane of (A), the 100 bp molecular marker (Invitrogen) is shown in the far right lane of (B). Long-PCR amplification results are shown in (A), displaying the heteroplasmic co-occurrence of a wild-type mtDNA molecules (larger amplicon) along with ΔmtDNA-C (smaller amplicon), in early generations (lanes 1–3). The detection of a PCR product associated with a new smaller mtDNA deletion (ΔmtDNA-7) can be seen at generation 25. A more subtle size shift toward a new even smaller ΔmtDNA (ΔmtDNA-8) occurs between generations 40 and 45. Follow-up standard PCR amplification of the same samples is shown in (B) using a set of primers closer to the focal deletion region. The canonical deletion produces a 540 bp band, ΔmtDNA-7 a 197 bp band, and ΔmtDNA-8 a 163 bp band.

Fig. 5.

— Composition of mtDNA at five-generation intervals within experimental lines across 6 strains, 50 generations, and 4 population sizes. Each square shows mtDNA composition within a single experimental line measured at a particular five-generation interval. Light blue squares indicate that only wild-type mtDNA genomes were detected, dark blue squares indicate that a mixture of wild-type/ΔmtDNA-C was detected, and dark red squares indicate that only ΔmtDNA-C PCR banding types were detected. A transition from dark blue to light blue indicates a decrease in ΔmtDNA-C between time points; a transition from dark blue to dark red indicates an increase in ΔmtDNA-C. Appearance of new or previously undetectable deletions is represented by bright red squares; the number within these squares corresponds to those in figure 3 and table 2, describing each of these deletions. Orange squares denote the detection of a second new or previously undetected deletion within a line. Black squares indicate line extinctions. Gray squares indicate missing data.

Fig. 6.

— Statistical predicted probabilities for ΔmtDNA categories. The probabilities of receiving ΔmtDNA categorical scores of 1–5 for strains EG4181 (top panels) or HK104 (bottom panels) at each population size studied. A score of 1 indicated the intact banding pattern (no ΔmtDNA of any type detectable, light blue squares in fig. 5), a 2 showed the intermediate banding pattern (both intact and ΔmtDNA-C, dark blue squares in fig. 5), and a 3 indicated the deletion pattern (ΔmtDNA-C band only visible on gel, dark red squares in fig. 5). A score of 4 showed the presence of a new ΔmtDNA type (bright red squares in fig. 5) and a 5 indicated the occurrence of second new ΔmtDNA type (arising after an initial new type scored as 4 in the same line, orange squares in fig. 5). Probabilities were calculated based on the ordered logistic regression model presented in table 3.