Structure and function of three novel MHC class I antigens derived from a C3H ultraviolet-induced fibrosarcoma.

Abstract

The UV-induced, C3H fibrosarcoma, 1591, expresses at least three unique MHC class I antigens not found on normal C3H tissue. Here we report the complete DNA sequence of the three novel class I genes encoding these molecules, and describe in detail the recognition of the individual products by tumor-reactive and allospecific CTL. Remarkably, although C3H does not appear to express H-2L locus information, this C3H tumor expresses two distinct antigens, termed A149 and A166, which are extremely homologous to each other and to the H-2Ld antigen from BALB/c. The gene encoding the third novel class I antigen from 1591, A216, is quite homologous to H-2Kk) throughout its 3' end. Since all three of these genes account for polymorphic restriction fragments not found in C3H, it is likely that they were derived by recombination from the endogenous class I genes of C3H. The DNA sequence homology of A149, A166, and H-2Ld is especially significant given the functional conservation observed between the products of these genes. Limited sequence substitutions appear to correlate with some of the discrete serological differences observed between these molecules. In addition, both A149 and A166 crossreact, but to differing extents, with H-2Ld at the level of T cell recognition. Our results are consistent with the view that CTL recognize complex conformational determinants on class I molecules, but extend previous observations by comparing a set of antigens with discrete and overlapping structural and functional differences.

The murine MHC class I genes, H-2Dq and H-2Lq, are strikingly homologous to each other, H-2Ld, and two genes reported to encode tumor-specific antigens.

“…In the coding sequences compared, the DQ gene is identical to the A166 gene, and the Lq gene is identical to the A149 gene. Furthermore, all of the previously published serological and functional characterizations of the D9 AND Lv CLASS I GENES A166 and A149 gene products are consistent with their identity to the D9 and Lq molecules (8,9,17). However, the Dk region was previously shown to contain a single gene, D't (13,14).…”

“…The considerable homology (>97%) of Lq to Dq supports the notion that these genes resulted from a duplication of an Ld-like progenitor. The near identity of Dq and Lq with the previously reported tumor-specific genes, A166 and A149 (17), from the UVinduced H-2k fibrosarcoma 1591, was unexpected. This observation raises the distinct possibility that the A166 and A149 genes represent genetic contamination rather than novel class I gene products expressed by tumor cells .…”

Role of the Major Histocompatibility Complex Class I Antigens in Tumor Growth and Metastasis

“…Two of these genes are highly homologous to the L d gene from BALB/c mice, and a third is quite homologous to the Kk gene throughout its 3' region (99). Furthermore, analysis of fi broblastic lines transfected with each of the three novel class I genes indicates that these sequences encode products recognized by the same monoclonal antibodies and CTL clones that are reactive with the original 1591 tumor (98,99).…”

A novel H-2K splice form: predictions for other alternative H-2 splicing events

“…All cDNAs were sequenced by the supercoil method of dideoxynucleotide sequencing (Chen and Seeberg 1983) using [e~-3sS]dATP (Amersham) and either Escherichia coli DNA polymerase I Klenow fragment (Boehringer Mannheim) or Sequenas e (U. S. Biochemical Co., Cleveland, Ohio). Oligonucleotide primers homologous to the T3 and T7 promoters flanking the cDNA inserts, as well as a panel of internal H-2K-and H-2D-specific oligonucleotide primers corresponding to evenly spaced regions on both strands of the respective gene, as previously described (Linsk et al 1986), were used to determine the complete DNA sequence of each clone. DNA sequence data were analyzed using the programs on the BIONET computer network (Intelligenetics, Palo Alto, California).…”

“…In an effort to predict such possible differential splicing patterns, several H-2 genes, including H-2U (Watts et al 1987, Arnold et aI. 1984, H-2K b (Weiss et al 1983), H-2K ~ ), H-2L d (Linsk et al 1986), H-2D b (Watts et al 1987), H-2D ~ (Watts et al 1987), H-2D p (Schepart et al 1986), and H-2D ~ (Sher et al 1986), were examined for the presence of cryptic splice sites.…”