Molecular Signatures of Adaptation: MHC Evolution and Natural Selection in European Rabbits

The Major Histocompatibility Complex (MHC) is a multi-gene family, covering a genomically complex region that plays a pivotal role in immune surveillance and pathogen recognition. Within this region, only a subset of MHC class I (MHC-I) genes encodes classical molecules that present peptides to CD8+ T cells. These molecules exhibit extraordinary genetic polymorphism, particularly in their peptide- binding domains, enabling them to bind a diverse array of pathogen-derived peptides and initiate targeted immune responses. While MHC-I genes evolve rapidly under pathogen pressure, their complex genomic architecture — characterised by gene duplication, gene conversion, and strong linkage disequilibrium — poses significant challenges for tracking evolutionary changes in natural populations. The myxomatosis pandemic in European rabbits (Oryctolagus cuniculus) provides an unprecedented opportunity to examine MHC-I evolution under intense viral selection. Following the release of myxoma virus as a biological control in 1950s in Australia, France and the United Kingdom, host and virus underwent rapid coevolution, with viral virulence declining and host resistance increasing dramatically within years. Although genome-wide analyses identified the MHC-I region as a target of selection during the pandemic, the specific genetic changes underlying resistance remained unknown. This thesis presents a comprehensive characterisation of MHC class I genes in Eu- ropean rabbits. Using PacBio long-read transcriptome sequencing from 68 modern rabbits, we identified and manually annotated 9 MHC-I paralogs on chromosome 12, named from Orcu-U1 to Orcu-U9. From these paralogs, 79 distinct alleles were identified, substantially improving our understanding of the MHC-I region in rabbits. Among these genes, Orcu-U1 and Orcu-U2 showed characteristics of classical MHC-I molecules — conserved residues for peptide presentation, high level of polymorphism, and wide and high tissue expression. To track MHC-I variation through time, we first performed a targeted amplicon se- quencing of Orcu-U1 and Orcu-U2 in 166 rabbits from modern and pre-RHDV populations, revealing 25 previously unidentified alleles. We then adapted the HLA- typing program OptiType to genotype the historical specimens collected before the release of myxoma virus, based on the knowledge of identified MHC-I alleles. Through simulations mimicking the fragmentation and deamination patterns of mu- seum DNA, we established reliable coverage thresholds for reliable genotyping of historical samples using the adapted program. Finally, we integrated these datasets to reconstruct changes in MHC-I frequencies across three populations in Australia, France, and Britain. Comparative analysis of modern and historical datasets suggested pronounced shifts in MHC-I allele frequencies before and after the release of myxoma virus. Using a time-aware modelling approach that accounts for historical admixture patterns, we identified six SNPs in the coding regions of Orcu-U1 and Orcu-U2 that showed significant changes at the genome-wide level. The most striking pattern was the parallel increase in frequency of two alleles, one from U1 and the other from U2, across all three populations. Through statistical phasing, we demonstrated that these two alleles were predominantly found on the same haplotype, exhibiting over 90% linkage disequilibrium patterns between them. To distinguish selective pressures from myxoma virus and rabbit haemorrhagic disease virus (RHDV), which emerged in rabbit populations in 1984, we used a Bayesian approach to estimate temporal changes in selection coefficients. We found evidence of positive selection on the rising MHC-I haplotype in Australian and British populations prior to RHDV emergence, with no significant changes in selection pressure after RHDV appeared. These results confirm that MHC-I adaptation was primarily driven by myxomatosis rather than subsequent viral challenges. Beyond myxoma virus, rabbits have experienced multiple selective pressures in their recent evolutionary history as they colonised new habitats and encountered other novel pathogens. A striking example is their adaptation to desert in Australia, where strong selection acts on coat colour polymorphism. Field phenotypic obser- vations indicated that 25% of rabbits in desert population exhibit a ginger-coated phenotype, which is encoded by an allele containing a 30-bp deletion in the MC1R gene that was introgressed from domestic rabbits, while alleles with no deletion encode grey phenotype. Analysis of rabbit remains from eagle nests showed a significant underrepresentation of ginger phenotypes compared to the field observations, suggesting that eagles preferentially prey upon grey rabbits. Visual modelling experi- ments demonstrated that grey rabbits have higher chromatic contrast against desert substrates than ginger rabbits, providing a potential explanation for this selective predation. This finding illustrates how colonisation of novel environments can expose invasive species to unexpected selective pressures that drive rapid evolutionary changes. We also explored potential genomic signatures of selection by rabbit haemorrhagic disease virus 2 (RHDV2), which emerged in Australia in 2015. We detected no sig- nificant signals of selection comparing samples collected before and after RHDV2’s release, possibly due to the limited time since its emergence.

Name dc:type.qualificationname

Doctor of Philosophy (PhD)

Level dc:type.qualificationlevel

Doctoral

Grantor dc:publisher.institution

University of Cambridge

Year dc:date.issued

2025

Author dc:creator

• Zhang, Yexin

Advisor dc:contributor.advisor

• Jiggins, Francis

• European Rabbits
• Evolutionary Genetics
• MHC-I
• Natural Selection

Licence

• https://apollo8-f-pro.lib.cam.ac.uk/bitstreams/88e3c1f5-bc34-499d-a567-31e386253170/download
• http://purl.org/NET/rdflicense/allrightsreserved

DOI dc:identifier.doi

https://doi.org/10.17863/CAM.120035

OAI identifier oai:identifier

oai:www.repository.cam.ac.uk:1810/387107