Major histocompatibility complex class I evolution in songbirds: universal primers, rapid evolution and base compositional shifts in exon 3

Abstract

Genes of the Major Histocompatibility Complex (MHC) have become an important marker for the investigation of adaptive genetic variation in vertebrates because of their critical role in pathogen resistance. However, despite significant advances in the last few years the characterization of MHC variation in non-model species still remains a challenging task due to the redundancy and high variation of this gene complex. Here we report the utility of a single pair of primers for the cross-amplification of the third exon of MHC class I genes, which encodes the more polymorphic half of the peptide-binding region (PBR), in oscine passerines (songbirds; Aves: Passeriformes), a group especially challenging for MHC characterization due to the presence of large and complex MHC multigene families. In our survey, although the primers failed to amplify exon 3 from two suboscine passerine birds, they amplified exon 3 of multiple MHC class I genes in all 16 species of oscine songbirds tested, yielding a total of 120 sequences. The 16 songbird species belong to 14 different families, primarily within the Passerida, but also in the Corvida. Using a conservative approach based on the analysis of cloned amplicons (n = 16) from each species, we found between 3 and 10 MHC sequences per individual. Each allele repertoire was highly divergent, with the overall number of polymorphic sites per species ranging from 33 to 108 (out of 264 sites) and the average number of nucleotide differences between alleles ranging from 14.67 to 43.67. Our survey in songbirds allowed us to compare macroevolutionary dynamics of exon 3 between songbirds and non-passerine birds. We found compelling evidence of positive selection acting specifically upon peptide-binding codons across birds, and we estimate the strength of diversifying selection in songbirds to be about twice that in non-passerines. Analysis using comparative methods suggest weaker evidence for a higher GC content in the 3rd codon position of exon 3 in non-passerine birds, a pattern that contrasts with among-clade GC patterns found in other avian studies and may suggests different mutational mechanisms. Our primers represent a useful tool for the characterization of functional and evolutionarily relevant MHC variation across the hyperdiverse songbirds.

Keywords: 454 pyrosequencing; Adaptive variation; Comparative methods; Diversifying selection; GC content; Immune response; Major histocompatibility complex; Pathogen-mediated selection.

Figure 1. Schematic representation of part of an MHC class I gene.

Arrows indicate the location of the primers used in this study. Both the coding sequences of exon 2 and exon 3 comprise the antigen-binding region of MHC class I molecules. Exons are represented by boxes and the lines connecting boxes represent introns.

Figure 2. Neighbor-joining tree of the passerine MHC class I sequences (exon 3) here isolated plus additional exon 3 sequences isolated in other avian species.

Bootstrap support for the main branches of the tree are indicated. A more detailed depiction of this tree is provided in Fig. S1.

Figure 3. Neighbor-net network of the MHC class I sequences (exon 3, N = 120) isolated from the 16 songbird species investigated in this study.

Only the main clusters of sequences are labeled for simplicity. Those species whose sequences fall into a single cluster are indicated by asterisks.

Figure 4. Distribution of positively selected sites in exon 3 of songbird class I genes as estimated by PAML (model 2).

Red columns indicate the class of sites with a high probability of ω > 1. In this model ω₁ = 0.25 and applies to ∼ 59.3% of the codons (blue); ω₂ = 1 at ∼ 28.2% of the sites (green); and ω₃ = 3.53 at ∼ 12.6% of the sites (red). Asterisks indicate codons assumed to comprise the avian PBR in the MEGA analysis and crosses indicate PBR residues in the human HLA-A2 molecule (Björkman et al., 1987; Saper, Bjorkman & Wiley, 1991).

Figure 5. Base compositional variation in the 3rd codon position of avian class I MHC genes (exon 3).

Clades are indicated according to the key provided. The topology depicted here is a neighbor-joining tree as described in Methods, however, the branch lengths have been ultrametricized as described in Methods to conduct the comparative tests. The branch leading to songbirds is indicated. The topology of this tree differs slightly from that in Fig. 2 because this is a simple neighbor-joining tree, rather than a bootstrap consensus of trees as in Fig. 2.