Kremer et al. describe a strain of the maternally inherited bacterium Wolbachia that induces parthenogenesis in its host, the wasp Asobara japonica; in certain populations females only produce female offspring when they are infected. A related Wolbachia strain in another Asobara wasp species (Asobara tabida) has quite a different effect—this other strain is essential for successful oogenesis but it does not cause parthenogenesis. Two related Wolbachia have, therefore, diverged and become essential for their wasp hosts, but without providing them any fitness benefit.

Wolbachia are well-known bacteria that live inside the reproductive tissue and other tissues of their insect hosts. The bacteria spread across generations by vertical transmission through females. The presence of these bacteria is associated with a range of host effects—from cytoplasmic incompatibility (where infected males prevent or reduce the production of viable embryos when they mate with uninfected females), through to feminization (genetically programmed males turning into females) and parthenogenesis (where Wolbachia cause females to produce only female offspring). Recently, it has become apparent that Wolbachia infections influence the fitness of their hosts in diverse ways, by altering patterns of reproduction, resistance to microbial infections and the provision of nutrients (Hedges et al., 2008; Teixeira et al., 2008; Ghedin et al., 2008). Changes in Wolbachia or their host genomes in the same species have now been documented (Hornett et al., 2006; Weeks et al., 2007) and suggest that evolutionary shifts in Wolbachia–host interactions (and their host effects) can be rapid. Wolbachia can therefore mediate rapid evolutionary shifts in host reproductive patterns and fitness effects.

Parasitoid wasps reproduce by producing offspring that then develop in the immature stages of moths, spiders, aphids and other arthropods. Both wasps, and their host arthropods, are often infected by Wolbachia and the bacterial infections are known to influence the fitness, cytoplasmic incompatibility and parthenogenesis of their wasp hosts. Parasitoid wasps may even facilitate the horizontal spread of Wolbachia across different arthropods as they attack a range of host species, although this is likely to be a rare event (Haine et al., 2005; Baldo et al., 2008).

Kremer et al. now provide evidence for multiple phenotypic effects that seems to have evolved in Wolbachia residing in parasitoids of the genus Asobara. In an earlier work on the species A. tabida, it was discovered that Wolbachia controls oogenesis by influencing programmed cell death (Pannebakker et al., 2007). When Wolbachia were removed from strains after an antibiotic treatment, there was no oogenesis and therefore no offspring production. Wolbachia have therefore become essential for one component of Asobara reproduction. In this paper, Kremer et al. describe a different strain of Wolbachia isolated from a Japanese population of the species A. japonica. This population was known to produce only female offspring through parthenogenesis. The authors show that parthenogenesis no longer occurs once the Wolbachia are removed because male offspring are then produced. These male offspring court and attempt to mate with females, but the females from this population are no longer receptive to the males. The Wolbachia have therefore become essential for the continued persistence of this A. japonica population, although other populations around Japan that have not been infected by Wolbachia continue to exist and exhibit sexual reproduction. Unlike the Wolbachia strain found in A. tabida, the new strain of Wolbachia is not essential for oogenesis, because offspring are produced when Wolbachia are removed through antibiotics.

To test the origin of the new Wolbachia strain from A. japonica, Kremer et al. sequence a series of Wolbachia genes used earlier in phylogenetic analyses. They find that the new strain is closely related to the strain from A. tabida that is essential for oogenesis, and not necessarily related closely to strains of Wolbachia from other hosts where the infection results in parthenogenesis. Therefore, two related strains of Wolbachia that occur in Asobara species have quite different effects, presumably because of recent evolutionary divergence in a few genes responsible for triggering different host phenotypes. However, it is also possible that the host genome dictates the expression of the phenotype induced by Wolbachia—this could be tested by transfering different Wolbachia strains across populations and species.

The spread of Wolbachia in insect populations depends on host effects as well as the fidelity of Wolbachia transmission across generations. For Wolbachia causing parthenogenesis to spread at the expense of other strains they need to provide an advantage to their hosts in terms of female offspring production, and exhibit a low rate of leakage of the infection when transmitted from parents to offspring (Stouthamer, 1997). However, in the case of some populations of A. japonica, it seems that the Wolbachia have basically become essential for the wasps, because females are no longer able to mate successfully with uninfected males. Presumably, there has been DNA sequence decay in the genes controlling female components of the mating interaction. Therefore, even if Wolbachia-infected strains no longer have a reproductive advantage, the infection is likely to persist in these populations of A. japonica, at least in the short term, and perhaps until they are invaded by sexually reproducing A. japonica or a close competitor. Wolbachia–host associations in parasitoids might then represent evolutionarily transient phenomena; they may develop rapidly, but then be prone to displacement, unless a fitness benefit for the host has meanwhile evolved.