and subsequent male-specific repressive mutations are of no relevance, or Wolbachia persists at reduced prevalence levels so that conditions for subsequent male-specific repressive mutations are eased due to lowered CI-levels.
Interestingly, these thresholds are up to 100% higher than for a male-specific repressive mutation (compare to figure 2).
In the absence of survival costs, male-specific repressive mutations always spread as long as transmission is imperfect (t<1; as shown analytically in eqs. 31&32 of Appendix S1).
However, if the male-specific repressive mutation's penetrance is lower, then Wolbachia will persist, albeit at lower prevalence as a result of the mutation's spread.
If fecundity costs are present (i.e. f>0), then effective transmission rates remain below 1 so that subsequent spread of male-specific repressive mutations is still possible.
In contrast, conditions for the subsequent spread of additional male-specific repressive modifiers are eased (i.e. threshold levels of e fall with lower lCI).
Generally, the consequences are reminiscent of a male-specific repressive mutation: The mutation induces an autonomous rescue-function in the host (irrespective of whether mutant females are infected with Wolbachia) that leads to a decrease in population CI-levels.
When we introduce a survival cost c that is inflicted on mutant males, the repressive mutation's penetrance e needs to be above a certain threshold ethr for the mutation to spread.
However, our study shows that male-specific repressive mutations lowering CI will spread for a wider range of conditions than will female-specific mutations lowering CI (i.e. a 'nuclear rescue construct', sensu Sinkins and Godfray[24]).
We find that fixation of sub-threshold male-specific repressive mutations (that decrease CI-levels and lower Wolbachia-prevalence) eases the conditions for the spread of subsequent mutations of similar kind, so that Wolbachia may be lost by way of sequential host adaptation.
