MHC polymorphism and host-pathogen interactions: The case of Borrelia in its reservoir host, the bank vole Myodes glareolus

The major histocompatibility complex (MHC) class IIB genes exhibit extensive allelic polymorphism, most likely maintained by pathogen-mediated balancing selection (PMBS). PMBS may operate in the form of heterozygote advantage (HA), and/or through the interaction of pathogens and specific MHC alleles via fluctuating selection (FS) or negative frequency-dependent selection (NFDS). In particular, NFDS is one of the primary models used to explain the extraordinary number of alleles found in the MHC genes. However, there is as yet limited empirical evidence for this hypothesis. Here, I use the tick-borne bacterium Borrelia afzelii in bank voles as a model system to study the maintenance of MHC class IIB variation for resistance to infectious diseases. I analyse both infection prevalence and infection intensity of Borrelia in a large-scale mark-recapture study (>500 bank voles). The prevalence and incidence of infection was higher in adults than immatures and peaked during the summer. Infection intensities decreased over time within individuals, and the loss of infection was significantly higher than the diagnostic error rate, in particular during winter, suggesting that bank voles can clear B. afzelii infections. I examined the rate ratio of replacement to silent amino acid substitutions in the peptide-binding region of the polymorphic MHC class II DRB and DQB genes in bank voles. I found signs of positive selection in both genes, but maximum likelihood models detected at least twice as many positively selected sites in DQB as in DRB, suggesting that DQB has been under stronger positive selection than DRB over evolutionary time. For this reason, I focused on DQB in my analyses of bank vole resistance to Borrelia. I analysed the effect of DQB-genotypes on both B. afzelii prevalence and strain-specific infection intensity. I tested 12 different DQB haplotypes for associations with infection prevalence. I found no evidence for HA in regards to B. afzelii prevalence. However, I found two significant associations between DQB haplotypes and infection prevalence: One haplotype was associated with resistance, while another was associated with susceptibility. The DQB intra-genotypic divergence was higher for voles with the resistance haplotype and lower in voles with the susceptible haplotype, compared to other voles. According to the divergent allele advantage hypothesis (DAA), individuals with a minimal overlap in antigen recognition of MHC-molecules should be selected for. I suggest that higher divergence in voles with resistance haplotypes (alleles), is consistent with selection where NFDS works in concert with DAA. The key assumption of NFDS is that there is a host-pathogen G×G for resistance to infection. To investigate if there is G×G between DQB and B. afzelii in bank voles, I analysed DQB haplotypes and infection intensity per B. afzelii strain. I found that one DQB haplotype had strain specific effects on quantitative resistance: Voles with haplotype H06 had lower infection intensities than voles without H06, with regards to some strains of B. afzelii. However, bank voles carrying the H06 haplotype had higher infection intensities of some other B.afzelii strains. This significant interaction between host MHC class II DQB haplotype and B. afzelii strains shows that there is potential for NFDS to maintain MHC allelic richness in a natural population of bank voles