Tests of population genetic models of the segregation distorter system in wild populations of Drosophila melanogaster

Abstract

The Segregation Distorter system of meiotic drive in Drosophila melanogaster consists of a haplotype of second chromosome loci (SD) which together mediate the dysfunction of sperm carrying a variable number of 240 base pair satellite DNA repeats located in the centromere region of SD+ homologs. These alleles are found worldwide in D. melanogaster populations in an apparently stable polymorphism that is unexpected under the simplest model of meiotic drive dynamics. The present study investigates whether a high level of repeat number mutation from insensitive (Rspi) to sensitive (RspS) repeat number is the force that balances the drive effect of SD in removing Rspi alleles. Iterative computer simulations of six hypothetical models of mutational processes produced equilibria suggesting that mutation-drive balance can produce stable polymorphism, but at frequencies different from those found in nature. Rates of mutation required to maintain polymorphism are on the order of 10.3 to 10- 1 changes per generation. Southern blot analysis of native forms of SD and SD+ chromosomes from a natural Hawaii population indicated >100 and 18 copies of the Rsp repeat in canonical Rsps and Rspi forms respectively. This agrees with previous studies showing the correlation between copy number and sensitivity. No changes 10 repeat copy number in chromosomal isolines were observed through approximately 25 generations. The three native isolines were used to establish population cages, which were sampled at intervals for Sd, Rspi and Rsps frequencies. After 300 days, Rsps was lost from all cages, which therefore did not reproduce the polymorphism observed in nature. Minimum X2 analysis shows that none of the six mutational models was a good predictor of the observed frequencies. Therefore, the hypothesized models are not supported by these data. Cages containing only native Rspi and Rsps were established and changes in Rsp frequencies measured. Estimates of the relative fitness of the alleles the three cages was s = 0.087, 0.025 and 0.068 respectively. These selective coefficients between Rsp alleles from the same natural population are much less than those found in previous studies using lab chromosomes and are too low to explain the observed SD polymorphism in this population.