The orthology of HLA-E and H2-Qa1 is hidden by their concerted evolution with other MHC class I molecules

Abstract

Background: Whether MHC molecules undergo concerted evolution or not has been the subject of a long-standing debate.

Results: By comparing sequences of eight functional homologues of HLA-E from primates and rodents with those of MHC class Ia molecules from the same eight species, we find that different portions of MHC class I molecules undergo different patterns of evolution. By focusing our analyses sequentially on these various portions, we have obtained clear evidence for concerted evolution of MHC class I molecules, suggesting the occurrence of extensive interallelic and intergenic exchanges. Intra-species homogenisation of sequences is particularly noticeable at the level of exon 4, which codes for the alpha3 domain, but our results suggest that homogenisation also concerns certain residues of the alpha1-alpha2 codomain that lie outside the antigen recognition site.

Conclusion: A model is presented in which Darwinian selective pressures due to pathogens could, at the same time, favour diversification of MHC class Ia molecules and promote concerted evolution of separate loci by spreading advantageous motifs arising by mutations in individual MHC molecules to other alleles and to other loci of the MHC region. This would also allow MHC molecules to co-evolve with the proteins with which they interact to fulfil their functions of antigen presentation and regulation of NK cell activity. One of the raisons d'être of the MHC may therefore be to favour at the same time both diversification of MHC class Ia molecules and homogenisation of the whole pool of MHC class I molecules (Ia and Ib) involved in antigen presentation.

Reviewers: This article was reviewed by Stephan Beck, Lutz Walter and Pierre Pontarotti.

Figure 1

Schematic representation of the 'birth an death' and the 'concerted evolution' models. In phylogenetic analyses, class Ia (solid boxes) and class Ib (hatched) molecules within a species are more similar to each other than to class Ia or class Ib molecules in another species. According to the 'birth and death' model, this reflects the propensity of MHC molecules to derive from a single ancestral sequence by successive gene duplications. Another explanation, the 'concerted evolution' model, is that class Ia and class Ib molecules tend to become like each other due to frequent events of gene conversion, without necessarily deriving from a single sequence.

Figure 3

Topologic distribution of the CD94L-specific and species-specific residues. A) The locations of CD94L-specific (blue) and species-specific (red) residues on the schematised structure of an α1/α2 co-domain. B) The residues of the α3 domain that have undergone intra-species homogenisation in at least one of the eight species studied (red numbers in Fig. 2) are represented in red 'spacefill' mode. Residues 222 and 227, which influence interactions of class I molecules respectively with tapasin and calreticulin, appear in black (and not red). The structure used is that of the rat MHC molecule RT1-A^a, bound to a 13-mer peptide of mitochondrial origin (acc. 1ED3) [82]. Representation of the 3D structure was generated with the Deep View Swiss-PdbViewer. The α1 domain is in orange, α2 in yellow, α3 in grey, beta 2 microglobulin in green and the peptide in light blue. The supine orientation of the molecule used here corresponds to that which is naturally adopted by MHC molecules on the plasma membrane, according to Mitra et al. [34]. The area indicated by blue dots corresponds to the footprint of the CD8 molecule [33].

Figure 2

Identification of species-specific and CD94L-specific residues. Comparison of the CD94L protein sequences (blue) with those of the class Ia molecules of their respective species (black). The sequences of the murine class Ib molecules M3, which are orthologous between mouse and rat were also included (green). Sequences from rodent are on the top 15 lines, and from primates on the lower 14 the alignment (see Materials and Methods for key to the abbreviations used). The 'pretty' output of the GCG software was used to generate this alignment of the MHC molecules extra-cellular portions (α1 domain: 1–90, α2: 91–180, α3: 181–270). Positions in agreement with the consensus are indicated by -. The positions of CD94L-specific residues, i.e. those that were found in at least 3 of the 4 CD94L rodent or primate sequences and not in the corresponding class Ia sequences are indicated by a blue number above the alignment. Conversely, certain residues were identified that differed from the general consensus, but were common to all the sequences of at least one species. These positions, where intra-species homogenisation has taken place, are indicated by a red number above the alignment.

Figure 4

When whole length sequences are compared, class Ib molecules cluster by function within taxa. A) A schematic representation of the divergence of the eight species studied here. The divergence times, indicated in millions of years, are those estimated from the fossil record. B) A comparison of the whole length of the aligned protein sequences shown in Fig. 2 (270 aa). Percentage bootstrapping support values are indicated of the branches (values above 60% are generally considered as highly significant). Class Ia molecules appear in red. SLA is a pig class Ia molecule chosen as an outlier, since the split with the ancestor of pigs occurred much earlier than that between primates and rodents [83]. CD94L molecules are shown in blue, and M3 molecules in green.

Figure 5

Intra-species concerted evolution is particularly prominent in the α3 domain. A) The tree was generated using the protein sequences of the α3 domain (last 90 aa of the aligned protein sequences shown in Fig. 2). B) The tree was generated using the protein sequences for the α1/α2 co-domain (first 180 aa of the aligned protein sequences shown in Fig. 2). Percentage bootstrapping support values above 60% are generally considered as highly significant. Class Ia molecules appear in red. SLA is a pig class Ia molecule chosen as an outlier, since the split with the ancestor of pigs occurred much earlier than that between primates and rodents [83]. CD94L molecules are shown in blue, and M3 molecules in green.

Figure 6

In the α1/α2 co-domain, residues inside and outside the PBR evolve differently. A) The tree was generated by comparing the 113 residues of the α1/α2 co-domain that fall outside of the antigen recognition site (ROARS). B) The tree was generated by comparing the 67 residues that take part in the PBR. Percentage bootstrapping support values above 60% are generally considered as highly significant. Class Ia molecules appear in red. SLA is a pig class Ia molecule chosen as an outlier, since the split with the ancestor of pigs occurred much earlier than that between primates and rodents [83]. CD94L molecules are shown in blue, and M3 molecules in green.