Ancestral polymorphism at the major histocompatibility complex (MHCIIß) in the Nesospiza bunting species complex and its sister species (Rowettia goughensis)

Abstract

Background: The major histocompatibility complex (MHC) is an important component of the vertebrate immune system and is frequently used to characterise adaptive variation in wild populations due to its co-evolution with pathogens. Passerine birds have an exceptionally diverse MHC with multiple gene copies and large numbers of alleles compared to other avian taxa. The Nesospiza bunting species complex (two species on Nightingale Island; one species with three sub-species on Inaccessible Island) represents a rapid adaptive radiation at a small, isolated archipelago, and is thus an excellent model for the study of adaptation and speciation. In this first study of MHC in Nesospiza buntings, we aim to characterize MHCIIß variation, determine the strength of selection acting at this gene region and assess the level of shared polymorphism between the Nesospiza species complex and its putative sister taxon, Rowettia goughensis, from Gough Island.

Results: In total, 23 unique alleles were found in 14 Nesospiza and 2 R. goughensis individuals encoding at least four presumably functional loci and two pseudogenes. There was no evidence of ongoing selection on the peptide binding region (PBR). Of the 23 alleles, 15 were found on both the islands inhabited by Nesospiza species, and seven in both Nesospiza and Rowettia; indications of shared, ancestral polymorphism. A gene tree of Nesospiza MHCIIß alleles with several other passerine birds shows three highly supported Nesospiza-specific groups. All R. goughensis alleles were shared with Nesospiza, and these alleles were found in all three Nesospiza sequence groups in the gene tree, suggesting that most of the observed variation predates their phylogenetic split.

Conclusions: Lack of evidence of selection on the PBR, together with shared polymorphism across the gene tree, suggests that population variation of MHCIIß among Nesospiza and Rowettia is due to ancestral polymorphism rather than local selective forces. Weak or no selection pressure could be attributed to low parasite load at these isolated Atlantic islands. The deep divergence between the highly supported Nesospiza-specific sequence Groups 2 and 3, and the clustering of Group 3 close to the distantly related passerines, provide strong support for preserved ancestral polymorphism, and present evidence of one of the rare cases of extensive ancestral polymorphism in birds.

Figure 1

Geographic location, vegetational composition, and Nesospiza populations occurring at the Tristan da Cunha archipelago. Location of the Tristan da Cunha archipelago in the South Atlantic Ocean with the three main islands: Tristan, Inaccessible, and Nightingale. The vegetational composition, and occurring species and morpho-types of Nesospiza buntings are shown for Inaccessible and Nightingale islands (adapted from reference 37 and Google Maps).

Figure 2

Nesospiza MHCIIβ gene tree. Neighbour-Joining tree showing the three well supported Nesospiza MHCIIβ exon 2 allele clusters. Of the 23 alleles, 21 were found in the N. acunhae individuals on Inaccessible Island (Neso1-8, 10–13, 15–23), 14 in the N. wilkinsi and N. questi on Nightingale (Neso1, 3, 4, 7, 9, 11, 13–15, 17, 18, 20, 22, 23), and 7 in R. goughensis (Neso5, 9, 13–15, 17, 23). Bootstrap support <70% are not shown.

Figure 3

Passerine MHCIIβ gene tree. Gene tree of the MHCIIβ exon 2 sequences of Nesospiza and sequences of several other passerine species obtained from GenBank. A Bayesian analysis of 159 bp of sequences, with independent mutational models applied to each codon position (Position 1: TIM3ef + I + G; Position 2: TVM + G; Position 3: TPM2uf + G). Bayesian posterior probabilities are indicated at the nodes; values < 0.95 are not shown.

Figure 4

Assignment of Nesospiza MHCIIβ peptide binding sites. Alignment of MHCIIβ exon 2 amino acid sequences of Neso01 – 23 indicating amino acid differences between Groups 1–3 sequences. Amino acid identity is shown by “.” and an alignment gap by “-”. Alignment with human HLA-DRB*04 (GenBank accession: NM_021983) was used to assign peptide binding sites (*) according to Brown et al.[44] and Tong et al.[45]