Genetic structure and contrasting selection pattern at two major histocompatibility complex genes in wild house mouse populations

Abstract

The mammalian major histocompatibility complex (MHC) is a tightly linked cluster of immune genes, and is often thought of as inherited as a unit. This has led to the hope that studying a single MHC gene will reveal patterns of evolution representative of the MHC as a whole. In this study we analyse a 1000-km transect of MHC variation traversing the European house mouse hybrid zone to compare signals of selection and patterns of diversification at two closely linked MHC class II genes, H-2Aa and H-2Eb. We show that although they are 0.01 cM apart (that is, recombination is expected only once in 10 000 meioses), disparate evolutionary patterns were detected. H-2Aa shows higher allelic polymorphism, faster allelic turnover due to higher mutation rates, stronger positive selection at antigen-binding sites and higher population structuring than H-2Eb. H-2Eb alleles are maintained in the gene pool for longer, including over separation of the subspecies, some H-2Eb alleles are positively and others negatively selected and some of the alleles are not expressed. We conclude that studies on MHC genes in wild-living vertebrates can give substantially different results depending on the MHC gene examined and that the level of polymorphism in a related species is a poor criterion for gene choice.

Figure 1

Position of sampled localities on the map of Europe: Arzdorf (ARZ), Schweben (SCH), Straas (STR), Buskovice (BUS), Studenec (STU) and Cejkov (CEJ), with the sample size in brackets. Dashed line schematically represents the house mouse hybrid zone. (Below) The hybrid localities between Straas and Buskovice: Tyniste (TYN), Horni Slavkov (HOSL), Rudolec (RUD), Mostov (MOS), Obilna (OBIL), Milhostov (MIL), Mostov (MOS), Doubi (DOU), Luzna (LUZ), Hurka (HUR), Thierstein (THI) and Lehsen (LEH).

Figure 2

Amplification of H-2Aa and pH-2Eb exon 2: position and sequences of PCR primers, reaction setup and cycling conditions. Shaded boxes represent exon 2, and the lines represent adjacent introns. Arrows indicate position of primers. PCR setup and conditions are identical for the two pairs of primers.

Figure 3

Amino acid alignment of H-2Aa and H-2Eb exon 2. Alignment of loci and codon numbering follows the IMGT (international ImMunoGeneTics information system) (Lefranc et al., 2005), based on structural and functional comparisons. Codons present at one locus but absent in the other are in grey. Parts of H-2Aa sequence that were not genotyped are dotted. Sequences are aligned against GenBank reference sequences for mouse MHC. The alleles that have been sequenced previously have the GenBank accession numbers attached. Antigen-binding codons (ABS) and codons under selection are shown below the alignments. Selection was analysed for three groups of sequences: musculus, domesticus and pooled musculus, domesticus and hybrid-specific sequences (m+d+H). Positively selected codons identified by random effects likelihood (REL), fixed effects likelihood (FEL) and single likelihood ancestral counting (SLAC) are shown as black squares, by OmegaMap as grey squares, negatively selected codons as shaded squares and antigen-binding codons as asterisks.

Figure 4

Neighbour-joining trees of (a) H-2Aa and (b) H-2Eb exon 2 sequences of mice, including sequences discovered in this study (in bold, labelled as, for example, Mumu-Aa*1). Sequences stored in GenBank have the accession numbers attached. The following taxa were included: Mus musculus castaneus (Muca), Mus musculus domesticus (Mudo), Mus musculus molossinus (Mumo), Mus musculus musculus (Mumu), Mus spicilegus (Musi) and Mus spretus (Musp). Arvicola terrestris (Arte) sequences were used to root the trees. Bootstrap values >60 are shown. The H-2Eb sequences highlighted in grey are under negative selection according to GA-branch analysis of sequences found in this study.