Gallwasp Lifecycles

Some gallwasps have truly weird and wonderful lifecycles, and the following description and explanation gets complex. It's just complicated, and I apologise in advance for this! With the considerable help of Rachel Atkinson and discussion with Roger Folliot, I have tried to keep everything as simple as possible.

Oak and Sycamore gallwasps are rare examples of a lifecycle called cyclical parthenogenesis. This involves alternation between sexually and asexually reproducing generations, and outside the gallwasps it is known only from six other animal groups - monogonont rotifers, digenean flukes, cladoceran crustaceans, and 3 insect lineages: the cecidomyiid gall midges (Diptera), adelgids and aphids (Homopteran bugs) and one species of beetle (Coleoptera). In most cyclically parthenogenetic animals, including aphids and cladoceran waterfleas , reproduction is predominantly asexual, with a single sexual generation each year triggered by a change in environmental conditions. Cyclical parthenogenesis in oak cynipids is unusual in that the two reproductive modes are strictly alternating and there is (at most) just a single generation of each type per year. No-one knows what triggers the switch between generations.

Alternation of generations in gallwasps has only been revealed by detailed rearing experiments involving the tracking of the offspring of individual females for several generations. The lifecycle I now describe has been shown for several European members of the genera Andricus, Cynips and Neuroterus and seems to represent the most complex state. The lifecycle involves two different types of asexually reproducing females - called androphores and gynephores. Androphores produce only sons (from haploid eggs by meiosis), while gynephores produce only sexual females (from unfertilised diploid eggs). The sexual females can also be divided into two types: one produces only androphores, while the other produces only gynephores. Sexual gallwasp females are weird compared to most Hymenoptera in that they cannot produce sons from unfertilised eggs and must mate to produce the next generation of asexual females. This is summarised in the diagram below.

The sycamore gallwasp Pediaspis aceris and several oak gallwasps (the best studied is the oak apple gallwasp Biorhiza pallida) deviate from this lifecycle. So few oak gallwasps have been studied in detail that all sorts of other deviations may exist.

  1. Biorhiza pallida and some Andricus species depart from the androphore/gynephore dichotomy, and individual asexual generation females are able to produce both males and females from unfertilised eggs (an ability called deuterotoky). In most cases, the departure from producing purely offspring of a single sex is slight, but some asexual females of Biorhiza pallida (called gynandrophores) produce approximately equal numbers of males and females.
  2. In contrast to the lifecycle shown above, the sexual generation females of Pediaspis aceris and Biorhiza pallida are able to produce small numbers of viable offspring without mating. However, just to confuse matters, these offspring are not haploid males (as they would be in typical hymenopteran arrhenotoky), but diploid asexual females! (c) Finally, a proportion of the sexual generation females of Biorhiza pallida and Andricus quercusradicis mated to a single male give rise to both androphores and gynephores.

How does this happen?

The genetic mechanism underlying alternation of generations in gallwasps remains completely unknown, and hypotheses are unavoidably complex. Explanatory hypotheses fall into two types - those involving the genesis of two types of sexual female by gynephores (the lefthand diagram below) and those involving the genesis of two types of male by androphores (the righthand diagram below). For two genotypes of sexual female to be generated by a single gynephore, the cytology of egg production must include a process generating genetic diversity in the daughters. In an asexual female, this could only be achieved by automixis from a heterozygous mother. Recent population genetic work in my lab by Rachel Atkinson on Biorhiza pallida and Andricus curvator shows, however, that in these species the offspring of a single gynephore are genetically identical, and are produced not by automixis, but by clonal apomixis, arguing against this proposal (Atkinson et al 2003).

The second hypothesis (the righthand diagram below) assumes that all sexual generation females are able to produce either gynephores or androphores, and that which type of asexual offspring a female generates depends on which of two possible male types she mates with. This hypothesis predicts that matings between a range of virgin females and the same male should all give rise to the same type of asexual female - a conclusion supported by experiments conducted by Roger Folliot with Andricus kollari. The original mechanism proposed to explain these data envisaged a gene for which gynephores and sexual females were obligately homozygous, but for which androphores were obligately heterozygous. This proposal was based on the view that gynephores reproduce by a form of automictic gamete duplication that can give rise only to homozygous offspring. In contrast, androphores must produce haploid sons by meiosis, and if heterozygous would inevitably give rise to the two types of male required in the righthand diagram above. Though an elegant explanation, available evidence suggests that sexual females are not produced by automixis. In Biorhiza pallida and Andricus curvator, daughters of a single gynephore are always genetically identical, but can be heterozygous. This is incompatible with gamete duplication, and instead implies clonal apomixis. Revealing the underlying mechanism remains one of the most interesting and least studied aspects gallwasp biology.

The length of cynipid lifecycles.

In most oak gallwasps the sexual generation gall (the gall containing the sexual generation) develops in the spring or early summer, while the asexual generation gall develops through the summer and autumn of the same year. Asexual generation females emerge from their galls in autumn and lay eggs that remain dormant until the following spring, or overwinter in the gall. There are many deviations from this general pattern and most oak gallwasps show considerable plasticity in response to environmental fluctuation. Many have lifecycles in which the asexual generation obligately or facultatively requires more than one year to develop (e.g. Biorhiza pallida), while in others the sexual and asexual generations each take a year (e.g. Andricus albopunctatus). In A. albopunctatus and similar species the galls of both generations are often found together, representing two cohorts a year out of phase. Unless the lifecycle is occasionally completed in a single year, these cohorts are effectively discrete sets of genotypes. In Andricus kollari the lifecycle is annual in the south of its range (Asia Minor and southern Europe), but takes two years in northern Scotland. Within Britain, there is a gradual south to north increase in the proportion of individuals taking two years to develop. As for A. albopunctatus, sexual and asexual generation galls present in the same year belong to cohorts a year out of phase.

Evidence for loss of sex from gallwasp lifecycles

Many oak gallwasps are known only from a single generation, and in the majority of cases the known generation is asexual. This raises the question of whether some or many of them have lost the sexual generation from their lifecycle and become obligately parthenogenetic. Secondary loss of sex has occurred in five of the six taxa of cyclically parthenogenetic animals, repeatedly in some groups, and is thus perhaps to be expected in oak gallwasps. Loss of sex is widespread in other gallwasps (particularly herb and rose gallwasps), and here it is due to infection by a bacterium, Wolbachia. Work in our group by Antonis Rokas suggests strongly that Wolbachia infection does not have the same impact on oak gallwasp reproduction (Rokas et al. 2002). To date, ability to bypass the sexual generation is seriously suggested for only three oak gallwasps. Plagiotrochus suberi is a European oak cynipid that that is cyclically parthenogenetic in Europe, but seems to have lost its sexual generation and become obligately parthenogenetic as an invader in North America. Andricus targionii is a wholly parthenogenetic species from eastern Asia, and is the only purely parthenogenetic species that overlaps in its distribution with its probable cyclically parthenogenetic ancestors. Andricus quadrilineatus is a European species that may represent an intermediate step in the loss of the sexual generation; a small proportion of asexual females are able to produce both sexual and asexual offspring.

Demonstration that oak gallwasps can sustain purely asexual lifecycles raises the question of how many of the species known only from parthenogenetic generations are genuinely obligately parthenogenetic. Until recently, rearing experiments have been the only technique available to resolve this issue. Population genetic analysis and DNA sequencing provide recently-developed alternatives.

  1. Analysis of gene frequences in population data can detect signals of sexual reproduction in apparently purely parthenogenetic taxa. Wholly parthenogenetic populations lack genetic recombination among alleles, and also inherit their chromosomes as entire sets. These processes are predicted to lead, respectively, to departures from Hardy-Weinberg and linkage equilibrium. Taxa that show no significant departure from these genetic equilibria are thus unlikely to be purely parthenogenetic, and more likely to possess a cryptic sexual generation. All 7 putatively asexual European Andricus species studied to date have been shown to have cryptic sexual generations in this way, confirmed in one case by collection of the sexual generation female.
  2. DNA sequencing can be used to match previously unpaired sexual and asexual generations sharing identical sequence. Sequencing has also revealed that samples of sexual generation adults thought on the basis of morphology to represent a single taxon in fact contain the previously unidentified sexual generations of several species. It is clear that gallwasps whose asexual generations may be clearly distinguishable as galls or insects may have sexual generations for which both galls and insects are currently indistinguishable. The consensus from these results is that while purely asexual oak gallwasps do exist, lifecycles in this group should be assumed to be cyclically parthenogenetic until proven otherwise.