Host-associated population substructure in generalist parasitoids: cyniipid oak gall communities as a model system.

Oak gallwasps (Hymenoptera, Cynipidae) induce a huge diversity of species-specific galls on oaks and related Fagaceae, each with a characteristic host plant association, location on the host, phenology and morphology. In addition to the gallwasp, these galls harbour closed and species-rich communities of cynipid inquilines, and parasitoids that attack the gall-inducer, inquilines, or both. A central feature of these communities is that many of the parasitoids are apparent generalists, attacking a wide diversity of hosts. Such apparent generalism has long raised important theoretical issues concerning the coexistence of multiple hosts, since early population dynamic models predicted that where two or more host species share a common natural enemy only the species with the highest intrinsic growth rate should survive . Several mechanisms have been suggested that could allow the co-existence of numbers of host species that share natural enemies, including physical or temporal host refuges from parasitoid attack, and parasitoid behaviours resulting in apparent preferences for certain host species. Studies of a range of parasitoid-host sytems have shown that parasitoids show considerable population variation in host location behaviour. Variation in host location preferences may be the result of learned or conditioned responses, but can also indicate underlying genetic variance in associated parasitoid traits. A range of plausible theoretical models predict that, given such genetic variation, an initially generalist parasitoid population should evolve towards specialised host races, each attacking a subset of the total host range. This prediction is supported by a range of studies showing apparently generalist parasitoids to consist of cryptic, sympatric parasitoid host races or species. The potential for variation in parasitoid host location behaviour is central to the ongoing debate on the possible adaptive significance of gall traits. All oak gall parasitoids attack their concealed hosts by locating the gall and drilling through the surrounding gall tissue. Selective modification of gall traits (such as location on the host plant, morphology, colour, and scent) and shifts between alternative oak hosts have long been seen as possible routes by which gallwasps and other gall-inducers might escape natural enemies, and achieve ‘enemy-free space’. However, taken at face value the apparent generalism of oak gall parasitoids suggests that diversity in gall traits has little significance for their ability to attack a wide diversity of hosts. Demonstration of the existence of ecotypes within oak gall parasitoids, however, would provide strong support for a causative link between gall traits and community composition, and hence support for cynipid escape from natural enemies through divergence in gall traits. Preliminary support for geographic variation in host preference comes from the observation that at a given location, oak gall parasitoids commonly only attack approximately a third of the hosts that they are known to attack over their entire distribution . However, the mechanisms by which oak gall parasitoids identify host galls and the population structure of apparently generalist species remain almost completely unstudied.

This proposal involves two complementary approaches to issues raised by the presence of generalist parasitoids in oak cynipid galls:

1. Genetic analyses of the population structure of selected generalist parasitoids.
This component is divided into two parts: (i) An analysis of genetic subdivision between parasitoid populations associated with galls on different oak hosts in long-established gallwasp communities and (ii) discrimination among alternative hypotheses for the origin of parasitoids recruiting to invading gallwasps in Britain.
2. Experimental analysis of the cues used by parasitoids to discriminate among alternative host galls. Analyses will concentrate on the roles of volatile plant compounds for parasitoid host searching and for the assessment of galls for suitable hosts.

We examined host-associated population substructure in 5 parasitoids associated with oak gallwasp hosts, using DNA sequence data (all 5) and microsatellites (marked *): Eurytoma brunniventris*, Sycophila biguttata*, Ormyrus nitidulus, Megastigmus dorsalis* and Megastigmus stigmatizans*. Individuals were sampled to allow us to examine population structure
(i) across host galls in ancient, glacial refuge communities (Hungary, Iran), and
(ii) associated with introduction of invading host gallwasps in NW Europe, including the UK.
Main Findings: Host-associated population structure
1. There is no evidence for gall host-specific population structure in any of these generalist parasitoids. All species showed genetic structure compatible with ability of individual lineages to switch frequently between structurally diverse galls on different parts of the oak.
2. Megastigmus dorsalis showed evidence of population genetic structure at the level of gallwasp hosts on different groups of host oaks (white oaks, section Quercus, versus black oaks, section Cerris).
3. Three parasitoid species comprise divergent cryptic lineages for mitochondrial and nuclear genes, suggesting cryptic species. Both Ormyrus nitidulus and Megastigmus dorsalis contain 2 distinct lineages that diverged at least 2 million years ago (mya), while Eurytoma brunniventris comprises at least 5 lineages that diverged at least 10 mya. All of these cryptic lineages attack a wide range of gallwasp hosts.
These results suggest that apparent generalism in gallwasp community parasitoids is genuine, contrary to the predictions of coevolution theory. We hypothesise that persistence of specialised parasitoid lineages is prevented by the unpredictability of host gallwasp populations in time and space. Above some critical threshold, such variation will impose strong selection on parasitoids to maintain the ability to attack whichever host is locally abundant.
Community phylogeography
4. Multiple parasitoid species show intraspecific genetic structure suggesting an eastern origin (from Turkey or Iran), followed by westwards range expansion over the period 2-5 mya. This matches signatures of an eastern origin in their gallwasp hosts, suggesting that the community may have spread as an intact set of interacting species over distances of at least 5,000 km.
5. Sycophila biguttata was shown to be a wholly parthenogenetic parasitoid, which shows almost no genetic variation across the Western Palaearctic. This adds to a small but growing group of asexually reproducing parasitoids.
DNA barcodes
6. DNA barcodes allow identification of taxa no the basis of DNA sequence data. This approach revealed the cryptic parasitoid species described in 3. Above. We find that cryptic lineages are even more abundant in another component of the gallwasp community – the inquiline gallwasps of the tribe Synergini, which inhabit galls, made by other species. Further, many morphological inquiline species have identical DNA sequences, suggesting that morphology is not a good guide to species limits. Overall, DNA barcodes have triggered a full-scale reanalysis of relationships in this important group.