Analysis of expression, structure and evolution of non-classical class I major histocompatibility complex genes

Abstract

Class I major histocompatibility molecules (MHC) are 45 kilodalton (kD) glycoproteins that associate with a smaller 12 kD polypeptide, [beta]2-microglobulin. In the BALB/c mouse, there are three classical class I molecules, H-2K [superscript d], D [superscript d], and L [superscript d], which are expressed throughout the body and present viral antigens to cytotoxic T lymphocytes (CTLs). In addition to the genes that encode the three classical class I antigens, the BALB/c genome contains 32 genes that structurally resemble the classical class I genes, and therefore possibly encode class I molecules. A few of the non-classical class I genes have been shown to encode molecules, TL, Qa-1, Qa-2, Q10, Qb-1, and Hmt, which are expressed in a generally tissue-specific manner, and probably do not act as restriction elements. However, it is unclear what function these molecules play, or why such a large gene family is maintained if only three viral antigen-presenting restriction elements are required by the murine immune system. DNA sequences were obtained from each of the 35 class I genes of the BALB/c mouse of the transmembrane domain-encoding fifth exon. Based on nucleotide sequence similarity, the fifth exons could be divided into seven groups that share little similarity with each other. In addition, the majority of the fifth exons are able to encode a transmembrane domain that can be separated into a proline-rich connecting peptide, a hydrophobic transmembrane segment, and a ctyoplasmic portion that includes basic anchoring residues. Since this conservation occurs in spite of extensive variation of nucleotide sequence in these exons, it is likely that selective pressure exists to maintain a functional structure in the majority of class I genes. A cDNA library was constructed from a thymus from a five-week-old BALB/c mouse. From this library, 69 class I cDNA transcripts from 15 different class I genes were isolated and analyzed. Included were three novel transcripts from Tla subregion genes, the T9 [superscript c], T17 [superscript c], and T18 [superscript c] genes. Sequence analysis of these clones reveals that the T9 [superscript c] gene is probably a pseudogene, while the T18 [superscript c] gene has an open reading frame in at least exons 2, 3, 4, and 5. A fourth cDNA clone was a transcript from the Thy19.4 gene, a gene that had not been previously isolated on a recombinant DNA clone. The isolation of transcripts from such a relatively large number of genes suggests that the number of expressed and perhaps functionally important class I genes may be larger than previously believed, and that expression of class I recognition structures may be important for cell-cell interactions within the thymus. To further pursue the characterization of the Thy19.4 gene, a genomic clone containing this gene was isolated from a size-selected insert library, and the DNA sequence of the Thy19.4 gene was obtained. The Thy19.4 gene contains an open reading frame, and in several aspects resembles the genes that encode the transplantation antigens. These similarities include a shared exon/intron structure and shared amino acid sequence motifs. In addition, PCR amplification experiments using tissue cDNA demonstrates that the Thy19.4 gene is expressed in a variety of tissues. However, unlike the classical transplantation antigens, the Thy19.4 gene maps distal to the H-2 region, in the Hmt region. These studies have demonstrated that class I gene transcription is more extensive than previously believed. Some of the expressed genes, like the T18 [superscript c] and Thy19.4 genes, appear to be able to encode class I molecules which may share structural characteristics with the classical transplantation antigens and may possibly serve as recognition structures in cell-cell interaction events. In addition, examination of the transmembrane domain exon of each of the 35 class I genes suggests that some selective constraint is acting on the majority of members of this family of genes, thus raising the possibility that many of the nonclassical class I genes encode functionally important products.