Evolution in eggs and phases: experimental evolution of fecundity and reproductive timing in Caenorhabditis elegans

Abstract

To examine the role of natural selection in fecundity in a variety of Caenorhabditis elegans genetic backgrounds, we used an experimental evolution protocol to evolve 14 distinct genetic strains over 15-20 generations. We were able to generate 790 distinct genealogies, which provided information on both the effects of natural selection and the evolvability of each strain. Among these genotypes are a wild-type (N2) and a collection of mutants with targeted mutations in the daf-c, daf-d and AMPK pathways. Differences are observed in reproductive fitness along with related changes in reproductive timing. The majority of selective effects on fecundity occur during the first few generations of evolution, while the negative selection for reproductive timing occurs on longer time scales. In addition, positive selection on fecundity results in positive and negative strain-dependent selection on reproductive timing. A derivative of population size per generation called reproductive carry-over (RCO) may be informative in terms of developmental selection. While these findings transcend mutations in a specific gene, changes in the RCO measure may nevertheless be products of selection. In conclusion, the broader implications of these findings are discussed, particularly in the context of genotype-fitness maps and the role of uncharacterized mutations in individual variation and evolvability.

Keywords: evolution of development; experimental evolution; reproductive dynamics.

Figure 1.

Evolutionary trajectory (measured using mean population size per generation) for 11 strains over 20 generations and 3 strains over 15 generations, normalized using exponential smoothing (α = 0.1).

Figure 2.

Twenty generations of experimental evolution on the wild-type (N2) strain (62 genealogies). Normalized population size is a non-evolved control subtracted from the measured population size.

Figure 3.

Twenty generations of experimental evolution on the AMPK mutant strains (62 genealogies for aak-1, 50 genealogies for aak-2 and 58 genealogies for aak-1; aak-2). Normalized population size is a non-evolved control subtracted from the measured population size.

Figure 4.

Twenty generations of experimental evolution on the daf-7(e1372) strain (62 genealogies). Normalized population size is a non-evolved control subtracted from the measured population size.

Figure 5.

Fifteen generations of experimental evolution on the AMPK/daf-c mutant strains (47 genealogies for aak-1; daf-7, 51 genealogies for aak-2; daf-7 and 53 genealogies for aak-1; aak-2; daf-7). Normalized population size is a non-evolved control subtracted from the measured population size.

Figure 6.

A heat map of reproductive carry-over (RCO) measurement values for 14 strains (13 mutant strains and 1 wild-type) over 15–20 generations. All bins (vertical axis) are of size 110 except for the zero bin. The coloured legend indicates the relative frequency of individual population/generation instances result in a range of RCO values, while the numbers in each cell reveal the exact number of generational replicates that exhibit a value with that range.

Figure 7.

Cumulative distribution functions of the RCO measure for three AMPK mutant genotypes, a daf-c mutant (daf-7) and the wild-type (N2).

Figure 8.

Cumulative distribution functions of the RCO measure for three AMPK/daf-c mutants (aak-1; daf-7, aak-2; daf-7 and aak-1; aak-2; daf-7), a daf-c mutant (daf-7) and the wild-type (N2).

Figure 9.

Normalized population size versus five discrete generations of the wild-type (N2) genotype and three AMPK mutant genotypes: N2 (a), aak-1 (b), aak-2 (c), aak-1; aak-2 (d). Starred pairs are statistically significant (two-tailed t-test).

Figure 10.

Normalized population size versus five discrete generations of the daf-7 mutant genotype and three AMPK/daf-c mutant genotypes: daf-7 (a), aak-1; daf-7 (b), aak-2; daf-7 (c), aak-1; aak-2; daf-7 (d).