Many species of insect are host to symbiotic microorganisms. These symbionts have traditionally been classified as mutual-ists (beneficial), parasitic (harmful), or commensal (neutral).
A variety of microorganisms among these symbionts are vertically transmitted, that is, they are inherited. The survival of these heritable symbi-onts is almost totally dependent on the success of their host. It follows that these symbionts should evolve characteristics that increase host survival. However, some so-called “ultraselfish” symbionts manipulate their hosts to the symbiont’s benefit, even when this is to the detriment of the host. The strategies used by such microorganisms include sterilization of noncarriers, feminization of genetic males, induction of asexual reproduction, and biasing host sex ratios in favor of females by killing male hosts. The most common of these reproductive parasites are the ultra-selfish bacteria.
ULTRASELFISH, MATERNALLY INHERITED BACTERIA
Inherited bacteria that live in the cytoplasm of host cells are transmitted to subsequent generations through the female line. This is because female gametes are heavily resourced with cytoplasm, whereas sperm contain negligible amounts. A maternally inherited bacterium can thus increase its prevalence, either by increasing the fitness of infected compared with uninfected hosts or by biasing the sex ratio in favor of females. The phenomenon of cytoplasmic incompatibility (CI), which involves the sterilization of noncarriers, follows the first of these options. Three other strategies, feminization, parthenogenesis induction (PI), and male-killing (M-k), skew populations in favor of females.
CYTOPLASMIC INCOMPATIBILITY
CI was first noted when some crosses between Culex pipiens mosquitoes from European, Asian, and American populations failed to produce offspring or produced offspring only when the crosses were carried out in one direction. Observed patterns of reproductive successes and failures demonstrated maternal inheritance (usually 95-100% efficiency), indicating the possible involvement of some cytoplasmic factor, hence the name given to the phenomenon. A bacterium of the genus Wolbachia was found to be responsible. Subsequently, CI has been reported from many insect orders.
CI occurs when a male carrying a CI agent mates with a female that does not bear the same agent. The bacteria allegedly secrete a chemical into the sperm of their host that kills zygotes formed within the female parent if they do not bear the same bacteria. In essence then, an uninfected female that mates with an infected male is rendered sterile thereafter. Other mating combinations are not affected.
CI bacteria can give rise to two patterns of reproductive failure. If one host population has the symbiont, whereas the other lacks it, matings fail in only one direction; that is, the incompatibility is unidirectional. If, however, each population harbors a different strain of CI bacterium, matings fail in both directions and the incompatibility is said to be bidirectional. For example, in the fruit fly Drosophila simulans, both unidirectional and bidirectional incompatibility have been recorded. Indeed, some CI host species carry more than one strain of CI Wolbachia, with individuals in some populations hosting two strains concurrently.
Host Range and Identity of the Agents Causing CI
CI is commonly reported when crosses of insects from geographically distinct populations are made. It has been reported most often from species of Diptera. However, Coleoptera, Heteroptera, Lepidoptera, Hymenoptera, and Orthoptera have also been found to exhibit the phenomenon and other orders will certainly be added to the list.
The CI phenotype can be cured by treatment with antibiotics or by exposing infected strains to high temperatures (—37°C) for several days. This suggests that the agents responsible for CI are bacteria. Most CI agents that have been identified belong to the Wolbachia pipientis complex, the exception being Cardinium species from the phylum Bacteroidetes that cause CI in some Encarsia wasps. The Wolbachia complex represents a widespread group of a-Proteobacteria, whose members are known to infect hosts from all the major orders of insects as well as some other arthropods (mites, spiders, isopods) and phyla (nematodes). Not all Wolbachia infections result in CI. Wolbachia is known to manipulate hosts in a variety of other ways and some appear to have no overt effect on their hosts, whereas some are beneficial. Estimates from molecular surveys using Wolbachia-specific probes suggest that between 20% and 25% of insect species harbor this bacterium.
Phylogenetic analysis has shown that Wolbachia in taxonomically very different hosts are more similar to one another than Wolbachia in very closely related hosts, indicating that Wolbachia can transfer between host species.
Mechanisms of Incompatibility
The precise mechanics of incompatibility are not well understood, partly because the mechanism of incompatibility appears to vary between host species. Various lines of evidence suggest that CI Wolbachia modify host paternal chromosomes during spermatogen-esis. In C. pipiens, incompatibility results from failure of sperm bearing CI Wolbachia to fuse correctly with female gametes that lack the same Wolbachia. In fruit flies (Drosophila) and moths (Ephestia kuehniella), embryo development appears to be suppressed at an early stage, suggesting loss of mitotic synchrony. In the jewel wasp, Nasonia vitripennis, Wolbachia interferes with condensation of the paternal chromosome set during the first mitotic cell division of the embryo so that the paternal chromosome set is lost. All progeny that get the Wolbachia thus carry just the maternal chromosome set, are thus hap-loid, and develop as males. Here, CI results in a male-biased sex ratio.
The Population Dynamics of CI
Bacteria that cause CI spread through host populations because infected females have an advantage over uninfected females, which are rendered sterile once they have mated with an infected male. The precise dynamics of CI are influenced by several factors, including (1) the vertical transmission of the bacterium, (2) the fitness costs or benefits it imposes on its host, (3) any horizontal transmission, and (4) host suppression systems.
In general, the prevalence levels of CI bacteria in infected populations are high and often approach 100%. Few studies of the population dynamics of CI have been conducted. The most intensive studies involve drosophilid hosts. For example, in D. simulans in California, rapid spread of a CI strain of Wolbachia was monitored during the 1980s and the 1990s. A theoretical model, based upon the levels of incompatibility, the vertical transmission efficiency of the bacterium, and the assumption that bearing the symbiont had no direct fitness effect on its bearer, produced theoretical predictions close to the observed rates of increase.
FEMINIZATION
Microbes and Sex Determination
Microorganisms of three types are known to cause feminization of their hosts. One type, microsporidians and other simple protists,as yet are known to infect only crustaceans. The other types are both bacteria, with Wolbachia, causing feminization in crustaceans and insects, and Cardinium causing feminization in mites. In essence, what these microorganisms do is turn genetic males into females.
The best-studied cases of feminization involve crustaceans, such as the woodlouse, Armadillium vulgare. However, many aspects of the feminization system in crustaceans are probably similar to those in insects, of which two species of moth, Ostrinia furnacalis and Ostrinia scapulalis, harbor feminizing Wolbachia.
Sex determination in the Lepidoptera is normally based upon sex chromosomes; females are the heterogametic sex, carrying two different sex chromosomes (WZ), and males are homogametic (ZZ). In O. furnacalis, all-female strains were first reported in 1998. Antibiotic treatment showed the female biases to be tetracycline sensitive, and molecular analysis revealed a Wolbachia that is taxo-nomically similar to that causing feminization in A. vulgare.
In parallel to A. vulgare, cure of feminizer lines with antibiotics led to a strong male bias in progeny of cured females. These moths are ancestrally female heterogametic, with the result that feminized ZZ males, when cured of the bacterium and mated by normal (ZZ) males, can produce only male offspring. It is perhaps no coincidence that in almost all examples of symbiont-induced feminization, the host is ancestrally female heterogametic, and further similar instances in the Lepidoptera and other taxa with heterogametic females are likely to be discovered in the future.
Feminization in the mite Brevipalpus phoenicis is extraordinary because this mite can live and reproduce entirely in the haploid state. This species reproduces by thelytokous parthenogenesis. When infected with a feminizing Cardinium, haploid individuals develop as females, rather than males, and these infected females are able to produce haploid female offspring.
Mechanism of Feminization
The mechanism of feminization in the two lepidopteran examples has not been determined. However, in A. vulgare, the difference between developing into a male and developing into a female appears to depend on the activity of a single gene that blocks the expression of one or more genes that cause the differentiation of the androgenic gland.
PARTHENOGENESIS INDUCERS
Microbe-Induced Parthenogenesis in the Hymenoptera
Much of the biotic world indulges in sexual reproduction. However, some organisms reproduce asexually, either ancestrally or secondarily when asexuality has evolved from sexuality. Secondary asexual reproduction, or parthenogenesis, occurs with new individuals arising from unfertilized eggs.
Various types of parthenogenesis exist. Normal haplodiploid species, in which females result from fertilized eggs and males from unfertilized eggs, are called arrhenotokous. In deuterotoky, both males and females arise from unfertilized eggs, whereas thelytoky involves the production of only female offspring without fertilization. The production of only female offspring, without the need for males, raises the possible involvement of cytoplasmic symbionts whose interests are favored by female hosts, which can vertically transmit them. This involvement is realized in a number of haplodiploid Hymenoptera and Thysanoptera, as well as some mites.
The first report of bacterially induced thelytoky was in the para-sitoid hymenopteran Trichogramma pretiosum. Here, administration of various antibiotics or of high temperatures to thelytokous females led to the production of males among the offspring. This led to the deduction that some sort of cytoplasmic microorganism, probably a bacterium, caused thelytoky in this species. Microscopic examination revealed a bacterium present in eggs of some thelytokous lines, but not in antibiotic-treated, temperature-treated, or arrhenotok-ous lines. Molecular genetic analysis revealed the presence of a Wolbachia.
Evidence of microbe-induced parthenogenesis has been revealed subsequently in over 50 species of parasitoid wasp, in the predatory thrip Franklinothrips vespiformis, and in the collembolan Folsomia candida, as well as in some species of mite (Bryobia and Brevipalpus spp.). The microbe involved is usually Wolbachia) although bacteria of the genera Cardinium and Rickettsia cause PI in a few cases. The bacteria involved do not form single taxonomic groups within their genera, but are intermixed with bacteria that affect hosts in other ways (e.g., CI and male-killing), suggesting that PI has evolved independently several times. Alternatively, the genes that cause PI may be horizontally transferred between bacterial taxa, through DNA exchange involving virus-like particles or plasmids. A third possibility is that the same bacterium causes different effects in different hosts, for example, parthenogenesis in haplodiploids and CI or male-killing in diploids. Phylogenetic analysis of a number of PI Wolbachia, particularly in the genus Trichogramma, has provided evidence for interspecific horizontal transmission of the symbiont.
Mechanism of Parthenogenesis Induction
Investigations of parthenogenetic organisms have shown variation in the way in which diploid offspring can result from unfertilized eggs. This diversity can be divided into two basic groups: either meiosis is modified with the diploid number of chromosomes being retained or, following meiosis, the diploid chromosome number is restored after the formation of a single pronucleus, usually by the fusion of two haploid mitotic products. In the examples of Wolbachia-induced parthenogenesis that have been studied microscopically, the latter of these two routes applies. In infected females, either both sets of chromosomes migrate to the same pole during the first mitotic division of the meiotic product (e.g., T. pretiosum) or, following the first mitotic division, the two mitotic nuclei fuse (e.g., Diplolepis rosae). The ways in which Wolbachia affect the behavior of host chromosomes during the first mitotic division are currently not known.
Dynamics of Wolbachia Infection
In most Hymenoptera with PI Wolbachia, thelytoky has been fixed. However, in many Trichogramma species, both thelytokous and arrhenotokous individuals occur together. The reasons for polymorphism in some species and monomorphism in others are not clear, but incomplete vertical transmission, levels of horizontal transmission, negative-fitness effects of the symbionts on their hosts, and suppresser genes may all have an influence.
MALE-KILLERS
Types of Male-Killers
Male-killing is perhaps the least sophisticated of the mechanisms of sex ratio distortion practiced by inherited symbionts. The basic mechanism involved is for the symbiont to kill male but not female
hosts. Among arthropods, M-k has been reported from five orders of insects (Hemiptera, Coleoptera, Lepidoptera, Hymenoptera, and Diptera) and two species of mite. In addition, perhaps because M-k is easily evolved, M-k’s themselves are taxonomically diverse, with a- and Y-Proteobacteria, Mycoplasma, Flavobacteria, and Microsporidia being recorded. Two M-k strategies have been reported. The first strategy, which involves killing male hosts late in their development (typically in the final larval instar), allows symbionts to replicate within their host, so that large numbers of individuals are released from the host’s corpse. Here, the population dynamics of the M-k depend on both horizontal and vertical transmission. Reported late, M-k’s are confined to Microsporidia infecting mosquitoes.
The second strategy involves killing males early in their development, typically during embryogenesis. In killing its host, the M-k dies, but in doing so, it increases the fitness of copies of itself in female hosts. The death of the M-k in male hosts means that early M-k does not involve significant levels of horizontal transmission. The increase in the fitness of infected females resulting from M-k is called fitness compensation, because for a M-k to invade a host population, infected females must have higher fitness than unin-fected females to compensate for the death of the infected males. The nature of cytoplasmic bacterial reproduction means that many or all progeny of an infected mother will carry copies of the sym-biont that are virtually identical by descent. If by killing male hosts the symbiont thereby increases the fitness of the dead males’ sisters, a fitness benefit will also accrue to “exact” copies of the male-killer within these females.
Two advantages of M-k have been identified. First, the probability of mating with close relatives, specifically brothers, will be reduced, thus avoiding the harmful effects of inbreeding. An infected female whose brothers have been killed has no siblings to mate with. The second advantage involves resource reallocation. If the resources that would have been consumed by males of an infected female become available specifically to female siblings, these females will benefit.
Early M-k has been recorded from many insect hosts, and to date, all confirmed early M-k’s are bacteria. A diverse array of bacteria is involved, suggesting that M-k has evolved independently many times.
The distribution of M-k’s among insects is not random. Certain groups have behavioral traits and ecologies that provide a context in which high resource benefits are obtained by daughters of infected females specifically as a result of the death of males. In these instances, the resources made available to females from the death of males become preferentially available to infected compared with uninfected females.
Hot spots for male-killing are known in the milkweed bugs (Hemiptera), nymphalid butterflies (particularly of the genus Acraea), and ladybug beetles (Coccinellidae).
Early Male-Killing in Ladybugs
Female-biased sex ratios were first reported in some families of the ladybug Adalia bipunctata in Russia in 1947. The female-biased sex ratio trait was shown to be maternally inherited. More recent work has demonstrated that the trait is curable with both antibiotics and heat treatment, suggesting that a bacterial agent is responsible. Male-killers have been discovered in 13 other species of ladybug. In each case, the female sex ratio bias is associated with the death of approximately 50% of eggs before they hatch (Fig. 1). This leads to the interpretation that the eggs of infected females have a normal sex ratio at fertilization, but that the secondary sex ratio is biased toward females because male embryos die before hatching.
FIGURE 1 An egg clutch laid by a male-killer-infected female Harmonia axyridis. Approximately half of the larvae (the females) have hatched and the neonates have begun consuming the soma of the dead male eggs.
Experimental work has shown that female larvae from mothers bearing a M-k gain a considerable nutritional advantage from consuming the contents of their dead brothers’ eggs. This means that they are larger when they disperse from their egg clutch than larvae from uninfected mothers, and they have longer to find and subdue their first prey before dying of starvation.
Detailed examination of resource reallocation in coccinellids has provided a set of criteria that are predicted to allow the invasion of M-k’s. The ladybug must feed on an ephemeral and often limited food, such as aphids. Neonate larval mortality from starvation must be significant. The ladybug must also lay eggs in clutches, and neonate larvae must consume unhatched eggs in their clutch. Lack of one or more of these criteria should make a ladybug reflective to M-k invasion. All ladybugs known to harbor M-k’s possess all the relevant criteria.
The Mechanism of Male-Killing
The mechanism by which bacteria discriminate between male and female hosts and kill only the former has only been elucidated in Drosophila melanogaster artificially infected with the M-k Spiroplasma poulsonii. Here, the M-k failed to kill males with mutations in any of the five protein-coding genes in the dosage compensation complex. However, as sex determination systems of species that host M-k’s vary (Drosophila and ladybugs have XX females and XY males, butterflies have ZZ males and ZW females, and the hymenop-teran parasitoid N. vitripennis has typically arrhenotokous haplodip-loid sex determination), it is unlikely that the various M-k’s all detect that they are in males in the same manner.
Consequences of Population Sex Ratio Distortion
Rather more is known about the evolutionary effects of M-k’s on their hosts. The prevalence of M-k’s varies from less than 5% to over 95% in different insect groups. In the ladybug Harmonia axyridis, the prevalence of a male-killing Spiroplasma is closely correlated to the population sex ratio. At low prevalence, the population sex ratio deviates little from 50% of each sex. However, at high prevalence (50% of females infected), over two-thirds of the population are female. Similarly, female-biased populations of A. bipunctata have been reported. More spectacularly, in the butterflies Acraea ence-don, Acraea encedana and Hypolimnas bolina, some populations
are 95% females. Such highly biased sex ratios have the potential to impact on the evolution of reproductive strategies. This potential may be the cause of a number of unusual features of the courtship and copulatory behavior of these species.
First, both A. bipunctata and H. axyridis exhibit a variety of genetically determined mate choice patterns, including, unusually, weak male choice of females. Furthermore, in A. bipunctata, the investment that a male makes in an individual copulation decreases as the ratio of females to males increases. Male A. bipunctata may transfer one, two, or three spermatophores (sperm packages) during a single copulation. In populations with a 1:1 sex ratio, males transfer, on average, just under two spermatophores. Conversely, in populations approaching two females for each male, males rarely transfer more than a single spermatophore (average 1.2).
In highly female-biased populations of the butterflies A. ence-don and A. encedana, many females die without mating. This suggests that males are limiting. The result has been the evolution of sex-role-reversed behaviors. The most striking of these is of female lekking behavior. Male leks, which are resource-lacking arenas where males congregate and compete for mates, are known in many taxa. In A. encedon and A. encedana, it is virgin females that congregate at resource-lacking sites. Within these arenas, females compete for matings.
IMPLICATIONS AND APPLICATIONS
Ultraselfish inherited bacteria adopt a strategy of harming their host to aid their own persistence and spread. Early reports of such symbionts, which caused abnormalities in host sex ratios or reproductive patterns, were long thought of as evolutionary oddities. Over the past two decades, that view has been changed as advances in molecular genetics have revealed that many insects are host to these reproductive manipulators. Surveys have shown that 20-25% of insect species harbor bacteria of a single genus, Wolbachia. It is likely that similar surveys directed toward other clades of inherited bacteria will lead to the conclusion that species of insects that host antagonist bacteria are in the majority.
Invasion by ultraselfish inherited bacteria can have far-reaching effects on the evolution of hosts. It has already been shown that the selection pressures imposed by these cytoplasmic symbionts have led to the evolution of nuclear suppresser genes in some insects (e.g., H. bolina and Menochilus sexmaculatus) . Such genes provide evidence of the intragenomic conflict that results from the maternal inheritance of these microbes. Invasion and spread of any maternally inherited symbiont through a host population will also cause an immediate decrease in the diversity of host cytoplasmic organelles, such as mitochondria, for they are inherited together. The spread of a symbiont, therefore, pulls the organelles of its original female host through the host population with it. Further effects of CI, PI, feminizing, and M-k bacteria undoubtedly await discovery. Given the antagonistic interactions between deleterious symbionts and their hosts, it is difficult to believe that a complete understanding of the biology of any insect species can be achieved without consideration of the symbionts it carries.
Deleterious inherited bacteria, such as Wolbachia, currently have a high scientific profile. This is at least partly because they may have a future role in pest control. Several strategies of use can be envisaged. Endosymbionts that cause female biases in host species that are already used as biological control agents, such as parasitoid wasps or aphidophagous coccinellids, may be used to increase the impact of the control agent or reduce costs of a control program. Alternatively, mass release of CI agents could be used to cause sterility in recipient populations, much as mass releases of males sterilized by radiation have been used to control Cochliomyia hominivorax. A third strategy would be to use these bacteria to transfer genes. If a symbiont bearing “useful” genes were likely to spread through a target population to fixation, as is likely of a CI agent, the useful genes need to be introduced by only a relatively small initial release. Finally, Wolbachia could be used to affect gene expression in arthropod or nematode hosts if a gene transformation system could be developed for this bacterium.
The study of ultraselfish inherited bacteria in insects has blossomed over the past decade. The focus on this group of extraordinary parasites is likely to increase, as the breadth of their influence on host evolution becomes more clearly understood.
