A Re-examination of Wolbachia-induced Cytoplasmic Incompatibility in California Drosophila Simulans

Background: In California Drosophila simulans, the maternally inherited Riverside strain Wolbachia infection (wRi) provides a paradigm for rapid spread of Wolbachia in nature and rapid evolutionary change. wRi induces cytoplasmic incompatibility (CI), where crosses between infected males and uninfected females produce reduced egg-hatch. The three parameters governing wRi infection-frequency dynamics quantify: the fidelity of maternal transmission, the level of cytoplasmic incompatibility, and the relative fecundity of infected females. We last estimated these parameters in nature in 1993. Here we provide new estimates, under both field and laboratory conditions. Five years ago, we found that wRi had apparently evolved over 15 years to enhance the fecundity of infected females; here we examine whether CI intensity has also evolved.

Methodology/principal Findings: New estimates using wild-caught flies indicate that the three key parameters have remained relatively stable since the early 1990s. As predicted by our three-parameter model using field-estimated parameter values, population infection frequencies remain about 93%. Despite this relative stability, laboratory data based on reciprocal crosses and introgression suggest that wRi may have evolved to produce less intense CI (i.e., higher egg hatch from incompatible crosses). In contrast, we find no evidence that D. simulans has evolved to lower the susceptibility of uninfected females to CI.

Conclusions/significance: Evolution of wRi that reduces CI is consistent with counterintuitive theoretical predictions that within-population selection on CI-causing Wolbachia does not act to increase CI. Within taxa, CI is likely to evolve mainly via pleiotropic effects associated with the primary targets of selection on Wolbachia, i.e., host fecundity and transmission fidelity. Despite continuous, strong selection, D. simulans has not evolved appreciably to suppress CI. Our data demonstrate a lack of standing genetic variation for CI resistance in the host.