Only rare classical MHC-I alleles are highly expressed in the European house sparrow

The exceptional polymorphism observed within genes of the major histocompatibility complex (MHC), a core component of the vertebrate immune system, has long fascinated biologists. The highly polymorphic classical MHC class-I (MHC-I) genes are maintained by pathogen-mediated balancing selection (PMBS), as shown by many sites subject to positive selection, while the more monomorphic MHC-I genes show signatures of purifying selection. In line with PMBS, at any point in time, rare classical MHC alleles are more likely than common classical MHC alleles to confer a selective advantage in host-pathogen interactions. Combining genomic and expression data from the blood of wild house sparrows Passer domesticus, we found that only rare classical MHC-I alleles were highly expressed, while common classical MHC-I alleles were lowly expressed or not expressed. Moreover, highly expressed rare classical MHC-I alleles had more positively selected sites, indicating exposure to stronger PMBS, compared with lowly expressed classical alleles. As predicted, the level of expression was unrelated to allele frequency in the monomorphic non-classical MHC-I alleles. Going beyond previous studies, we offer a fine-scale view of selection on classical MHC-I genes in a wild population by revealing differences in the strength of PMBS according to allele frequency and expression level.

Methods
Exon 3 of MHC class I was amplified and sequenced on the Illumina MiSeq platform from gDNA and cDNA originating from blood samples collected from 28 house sparrows (Passer domesticus) at four sites across Europe. Reads were trimmed and filtered to identify unique allele sequences in gDNA and cDNA. Alleles were classified as classical or non-classical, with non-classical alleles identified as having a 6 bp deletion. Filtered read counts were converted into allele presence/absence (1/0) for gDNA and cDNA, and cDNA read counts were designated as the expression levels for (i) nucleotide alleles (MHCI_cDNA_gDNA_nt_geno_pheno.txt) and (ii) amino acid alleles (MHCI_cDNA_gDNA_aa_geno_pheno.txt). Allele frequencies (i.e. number of individuals carrying each allele) were calculated. The numbers of unique genomic and expressed alleles were summed for each bird (MHCI_cDNA_gDNA_total_alleles_pheno.txt). Accompanying phenotypic data include country of origin (Bulgaria, Poland, Spain or Sweden) and season of sample collection (spring/autumn). For full methods, refer to the published article.