The secret in their MHC: variation and selection in a free living population of great tits

Understanding the genetic basis of fitness differences has been a major goal for evolutionary biologists over the last two decades. Although there are many studies investigating how natural selection can promote local adaptation, few have succeeded to find the link between genotype and fitness of the phenotype. Polymorphic genes of the major histocompatibility complex (Mhc) are excellent candidates for such associations as they are a central component of the vertebrate immune system, playing an important role in parasite resistance, and hence can have direct effects on survival of their bearers. Although associations between Mhc and disease resistance are frequently documented, the epidemiological basis of the host-parasite interaction is often lacking and few studies have investigated the role that Mhc genes play in individual variation in fitness; thus comparatively little is known about the fitness consequences of Mhc in wild populations. Furthermore, the majority of work to date has involved testing associations between Mhc genotypes and disease. However, the mechanism by which any direct selection on the Mhc acts, depends on how genotypes map to the functional properties of Mhc molecules. The aim of this thesis was to characterize Mhc alleles in terms of their predicted functional properties and to investigate whether and how selection operates on Mhc class I functional variation using the great tit (Parus major) population at Wytham Woods as a model host species. Through a comprehensive characterization effort and the use of 454 pyrosequencing platform, I performed a detailed analysis of genetic variation at Mhc class I exon 3 and grouped alleles with similar antigen-binding affinities into supertypes to classify functionally distinct Mhc types. There was extreme complexity at the Mhc class I of the great tit both in terms of allelic diversity and gene number. A total of 862 alleles were detected from 857 individuals; the highest number yet characterized in a wild bird species. The functional alleles were clustered into 17 supertypes; there was clear evidence that functional alleles were under strong balancing selection. To understand the role of Mhc in disease resistance, I examined the linkage between Mhc supertypes, Plasmodium infection and great tit survival, and showed that certain functional variants of Mhc confer resistance to two divergent Plasmodium parasite species that are common in the environment. I further investigated the fitness consequences of functional variation at Mhc, using mark-recapture methods and long-term breeding data; and tested the hypotheses that selection: (i) maximizes Mhc diversity; (ii) optimizes Mhc diversity, or (iii) favours specific functional variants. I found that the presence of three different supertypes was associated with three different components of individual fitness: adult survival, annual recruitment probabilities and lifetime reproductive success. In contrast, there was no evidence for a selective advantage of Mhc functional diversity, either in terms of maximal or optimal supertype diversity. Finally, I explored the role that Mhc plays in female mate choice decisions and examined the reproductive fitness consequences of Mhc-dependent mating patterns. There was little evidence to suggest that functional dissimilarity at Mhc has any influence on female mate choice decisions or that dissimilarity at Mhc affects the reproductive output of the social pair. Overall, this thesis provides strong support for the suggestion that selection favours specific functional variants of Mhc, possibly as a result of supertype-specific resistance or susceptibility to parasites that exert strong selective pressures on their hosts; whereas there is no support for selection favouring maximal or optimal Mhc diversity. More importantly it demonstrates that functional variants of Mhc class I loci are an important determinant of individual fitness in natural populations.This thesis is not currently available in ORA