The sex-determining systems of arthropods are surprisingly diverse. Some species have male or female heterogametic sex chromosomes while other species do not have sex chromosomes. Most species are diploids but some species, including wasps, ants, thrips and mites, are haplodiploids (n in males; 2n in females). Many of the sexual aberrations, such as sexual mosaics, sex-specific lethality and conversion of sexuality, can be explained by developmental defects including double fertilization of a binucleate egg, loss of a sex chromosome or perturbation of sex-determining gene expression, which occur accidentally or are induced by certain environmental conditions. However, recent studies have revealed that such sexual aberrations can be caused by various groups of vertically-transmitted endosymbiotic microbes such as bacteria of the genera Wolbachia, Rickettsia, Arsenophonus, Spiroplasma and Cardinium, as well as microsporidian protists. In this review, we first summarize the accumulated data on endosymbiont-induced sexual aberrations, and then discuss how such endosymbionts affect the developmental system of their hosts and what kinds of ecological and evolutionary effects these endosymbionts have on their host populations.

When the butterfly Eurema hecabe is infected with two different strains (wHecCI2 and wHecFem2) of the bacterial endosymbiont Wolbachia, genetic males are transformed into functional females, resulting in production of all-female broods. In an attempt to understand how and when the Wolbachia endosymbiont feminizes genetically male insects, larval insects were fed an antibiotic-containing diet beginning at different developmental stages until pupation. When the adult insects emerged, strikingly, many of them exhibited sexually intermediate traits in their wings, reproductive organs, and genitalia. The expression of intersexual phenotypes was strong in the insects treated from first instar, moderate in the insects treated from third instar, and weak in the insects treated from fourth instar. The insects treated from early larval instar grew and pupated normally but frequently failed to emerge and died in the pupal case. The dead insects in the pupal case contained lower densities of the feminizing Wolbachia endosymbiont than the successfully emerged insects, although none of them were completely cured of the symbiont infection. These results suggest the following: (i) the antibiotic treatment suppressed the population of feminizing Wolbachia endosymbionts; (ii) the suppression probably resulted in attenuated feminizing activity of the symbiont, leading to expression of intersexual host traits; (iii) many of the insects suffered pupal mortality, possibly due to either intersexual defects or Wolbachia-mediated addiction; and hence (iv) the feminizing Wolbachia endosymbiont continuously acts on the host insects during larval development for expression of female phenotypes under a male genotype. Our finding may prompt reconsideration of the notion that Wolbachia-induced reproductive manipulations are already complete before the early embryonic stage and provide insights into the mechanism underlying the symbiontinduced reversal of insect sex.

It was recently recognized that in Japan, the common yellow butterfly, Eurema hecabe, consists of two sibling species, which have been unnamed yet and tentatively called yellow (Y) type and brown (B) type. We investigated the diversity of nuclear and mitochondrial genes in Japanese populations of Y type and B type of E. hecabe. The phylogeny based on nuclear genes agreed with the distinction between Y type and B type, which had been also supported by a wide array of biological data. However, the phylogeny based on mitochondrial genes did not reflect the distinction. PCR survey of Wolbachia revealed that B-type populations were all infected while Y-type populations contained both infected and uninfected individuals. A single genotype of Wolbachia, which was inferred to be a CI-inducing strain from their wsp gene sequence, was prevalent in these populations. Notably, the mitochondrial phylogeny was in perfect agreement with the pattern of Wolbachia infection, suggesting that the Wolbachia infection had affected the mitochondrial genetic structure of the host insects. Probably, the Wolbachia strain and the associated mitochondrial genomes have been occasionally introduced from B-type populations to Y-type populations through migration and subsequent interspecific hybridization, and CI-driven population sweep has been spreading the Wolbachia strain and the particular mitochondrial haplotypes, which originated from B-type populations, into Y-type populations. On the basis of these results together with the geological and biogeographical knowledge of the Japanese Archipelago, we proposed an evolutionary hypothesis on the invasion and spread of Wolbachia infection in B-type and Y-type of E. hecabe.

Wolbachia belonging to Alphaproteobacteria are transovarially transmitted bacteria responsible for reproductive alterations in a wide range of arthropods. In natural populations of the butterfly Eurema hecabe, there are two different types of Wolbachia-infected individuals. Individuals singly infected with Wolbachia strain wHecCI exhibit strong cytoplasmic incompatibility, whereas those doubly infected with wHecCI and wHecFem exhibit feminization. Here, we examined the infection frequencies and population densities of each Wolbachia strain in different host tissues (ovary, testis, fat body, midgut, Malpighian tubule and leg), and the cost of infection in offspring produced by single-infected and double-infected mothers of E. hecabe. The vertical transmission rate of wHecCI was nearly 100%, and that of wHecFem was c. 80%. The wHecCI densities were 10(3)-10(4)-fold higher than the wHecFem densities. In most tissues, the wHecCI densities were significantly higher in offspring of single-infected mothers than in offspring of double-infected mothers. In offspring of double-infected mothers, however, the wHecCI densities were not affected by the presence of wHecFem, suggesting a lack of interaction between the wHecCI and wHecFem densities. The offspring development time was dependent on the infection status of the mothers. These results imply that the maternal infection status affects the Wolbachia densities and fitness of offspring.

No abstract

1. Complete feminisation of genetic males into functional females, a unique case among insects, is known in Eurema mandarina (former Eurema hecabe Y type) that are infected with two strains of Wolbachia, w CI Em and w Fem Em .2. Here, we newly found that a proportion of wild-caught E. hecabe (former E. hecabe B type) produced only female offspring. Cytogenetic observations indicated that individuals of E. hecabe displaying the all-female trait were genetically male (i.e. feminisation).3. Multilocus sequence typing analyses demonstrated that the feminised individuals of E. hecabe were infected with two Wolbachia strains, w CI Eh and w Fem Eh , that were indistinguishable from w CI Em and w Fem Em , respectively. 4. Even identical strains of Wolbachia can be regulated differently depending on the host genetic background. Therefore, we compared the infection densities and vertical transmission efficiencies of Wolbachia between feminised E. mandarina and E. hecabe, but detected no significant differences in these traits.5. The possible routes by which the two Wolbachia strains have transferred between E. mandarina and E. hecabe are discussed.

Wolbachia are maternally inherited endosymbiotic bacteria of invertebrates that can manipulate the reproductive systems of their arthropod hosts in a variety of ways. To establish a useful model system for investigating the mechanism of Wolbachia-induced host feminization, we conducted the following series of experiments: (1) feminizing Wolbachia of the butterfly, Eurema hecabe, were transferred into cell cultures of the silkmoth, Bombyx mori, and (2) the transfected Wolbachia in cell cultures were inoculated into B. mori at four immature stages. Wolbachia were successfully transfected into the cell cultures and stably maintained for more than 1 year (>30 passages). However, none of the inoculated insects produced mature oocytes that were Wolbachia-positive. This finding was consistent with the fact that Wolbachia was not detected in individuals in subsequent generations. In contrast, Wolbachia were detected at relatively high frequencies (60-80% of individuals) in the somatic tissues of inoculated insects. Real-time quantitative polymerase chain reaction revealed that the Wolbachia densities in the cultured cells were approximately tenfold higher than those in the native host E. hecabe. Among B. mori individuals inoculated at various developmental stages, those inoculated at early stages exhibited higher Wolbachia densities at the adult stage. The Wolbachia densities in individuals inoculated at the second-instar stage were comparable to those in intact E. hecabe. These results suggest that infection and/or proliferation of Wolbachia in germline cells are actively hindered by regulation in B. mori but feasible in somatic cells and that the Wolbachia densities in somatic tissues are regulated by the living host insects.

Wolbachia, belonging to Alphaproteobacteria, is ubiquitously found in arthropods and filarial nematodes, and is known to manipulate the reproduction of its hosts in various ways, such as feminization, male killing, induction of parthenogenesis or induction of cytoplasmic incompatibility. We found that the Wolbachia infection frequencies of the butterfly Colias erate poliographus were high (85.7-100%) in seven Japanese populations. Crossing experiments and rearing revealed that the Wolbachia strain exhibited strong cytoplasmic incompatibility and perfect vertical transmission in C. erate poliographus. Moreover, a comparison of the survival rates between infected and cured broods suggested that Wolbachia infection had beneficial effects on host fitness. Our findings suggested that the high infection frequencies in Japanese populations have been accomplished by these advantageous traits of the Wolbachia strain. Furthermore, the multilocus sequence typing (MLST) scheme revealed that the Wolbachia in C. erate poliographus is a novel strain (ST141), belonging to supergroup B.