Identification, origin and evolution of a new family of mammalian MHC class I genes

All jawed vertebrate species examined so far have major histocompatibility complex (MHC) class I molecules that trigger cell-medieted immunity by displaying endogenously generated peptides to CD8+ cytotoxic T lymphocytes through the interaction with αβ T cell receptors (TCR). These molecules are encoded in the MHC region and known to be highly polymorphic. They are called classical class I . There are several families that show structural homology to classical class I molecules but have different functions, which are called non-classical class I genes. Some of them are not involved in immune responses. They are encoded not only inside the MHC but also outside the MHC. Recently, a novel non-classical class I family was found near the leukocyte receptor complex (LRC) on mouse chromosome 7 in our laboratory. This newly discovered family was named Mill (MHC class I-like located near the LRC). The Mill family consists of two functional genes, Mill1 and Mill2, and two pseudogenes, Mill-ps1 and Mill-ps2 showed sequence similarities to human MICA and MICB genes (about 40% amino acid sequence identity). This close relationship between the Mill and MIC families was also supported by phylogenetic analyses. The overall amino acid sequence identity of the α1-α3 domains between MILL1 and MILL2 was about 70%. However, each domain showed different sequence homologies. Sequence identities of the α1, α2 and α3 domains were 43%, 87%, 78%, respectively. Another notable feature of Mill1 and Mill2 was that they had an extra exon (exon2) between the signal peptide-encording exon and thea 1 domain-encording exon, which wasn\u27t observed in any class I families. While exon2 of Mill1 was transcribed constantly, exon2 of Mill2 alternaitively spliced out. This extra exon of Mill2 turned out to be part of a repetitive element, SINE-B4, but exon2 of Mill1 was not part of such repetitive elements, which indicate that Mill1 and Mill2 acquired their extra exons independently. Mill1 was expressed at rather restricted tissues such as skin and eyes neonatal mice, whereas Mill2 was ubiquitously expressed at low levels. To understand the evolutionary dynamics of this family and gain insights into its function, I carried out comparative analysis between mice and rats. The Mill family of rat consists of only Mill1 and Mill2. There was no orthologous copy of mouse Mill-ps1 and Mill-ps2. Sequence identities of MILL1 and MlLL2 between mice and rats were 71% and 74% (α1-α3 domains), respectively. Apparently, this degree of sequence conservation was the least among the class I genes that share orthologous copies between mouse and rat. However, in no domain, non-synonymous substitution rates (dN) signicantly exceeded synonymous substitution rates (ds); hence there was no obvious evidence for the positive selection. Interestingly, the α2 domain of mouse rat Mill1 showed most similarity to that of mouse Mill2 rather than rat Mill1. The α2 domain of rat Mill1 showed most similarity to that of rat Mill2. Similarity, the α3 domains of mouse Mill1 and rat Mill1 showed most similarities to those of mouse Mill2 and rat Mill1, respectively. The other domains and introns did not show such relationship. These observations suggested the occurrence of gene-conversion like events at the α2 and α3 domains. Polymorphism analysis revealed that most of inbred rat strains (12 of 14 strains) had Mill2 pseudogene owing to the same single premature stop codon mutation, which indicates that the functional constraint on Mill2 was low. Deduced exon/intron structures of mouse and rat Mill genes were similar but not identical The most remarkable difference was the absence of the extra exon in Mill2. Although the corresponding sequence to the extra exon was found at intron 1 of rat Mill2, no mature mRNA containing this sequence was observed in any tissues. Nevertheless the amino acid sequence of rat Mill2 was longer than that of mouse Mill2, because, in the rat, the region corresponding to the mouse exon 5 (α3 domain), intron 5, and exon 6 (CP/TM/CYT) became a single exon without any frameshift. Computational homology searches revealed that the emergence of the Mill family could be traced back to the common ancestor of extant eutherian mammals and marsupials. A computational search followed by the isolation of cDNA from horse revealed that Mill is not a rodent specific class I family, although the horse Mill gene was a transcribed pseudogene. Further computational analysis revealed the existing of the Mill family was not found in any other species examined so far. So, it was suggested that the mill family was lost from several lineages independently, possibly owing to low functional constraints. Interestingly, Mill and MIC families tended to show reciprocal distribution in mammalian species. In other words, no species have both Mill and MIC families simultaneously, although it is not yet known about functional correlation between them. Thus, the Mill family is supposed to have undergone dynamic changes through mammalian evolution, which may be due to the species specific functions or overlapping function with other genes