domain variability is clustered into three discrete regions, two of which divide the Ap chains into subgroups, suggesting an evolutionary history for the separation of alleles in inbred strains of mice. The amino acid sequences of these five chains are compared to each other and to previously published I-Ap chains. Correlations are made between the primary structural differences and the serologic and immune response characteristics mapping to the I-A subregion.The intimate relationship between I-region associated (Ia) antigens and immune responsiveness to a multitude of natural and synthetic antigens has been known for some time (1, 2).Recent studies using either anti-Ia antisera or monoclonal antibodies to block immune responses strongly suggest that Ia antigens are the products of immune response (Ir) genes (3-5), but how a limited number of Ia molecules differentially regulate responses to a myriad of antigens is yet to be resolved. It seems likely that their role in the presentation of antigen to imtnunocompetent cells is of fundamental importance and that primary structural alterations in the a and (3 chains of the class II heterodimer result in a failure to present antigen in a configuration appropriate for recognition. These alterations must also account for the multiple epitopes recognized by both anti-Ia antibodies and alloreactive T cells.In an effort to address the nature of primary structural differences among class II molecules and how these differences affect immune responses, we have undertaken the cloning and sequencing of cDNAs encoding I-Ap chains from several mouse haplotypes. This should facilitate the eventual assignment of particular serologic epitopes and restriction elements to specific amino acids in these chains. Such analysis is a first step in designing experiments in which immune responsiveness can be manipulated via structural alterations in Ia molecules and then, perhaps, understood. Clones hybridizing to the human DQp or the murine Al cDNA probes were tested for the presence of restriction enzyme sites consistent with AP but not Ed coding sequences (9). The longest such clones were subcloned into the EcoRI site ofpBR322 and from there into the EcoRI site of M13 mp8 for sequencing. Sequencing was performed by the dideoxysequencing method of Biggin et al. (10). All clones were sequenced using the M13 universal primer (UP) from the EcoRI ends, as shown in Fig. 1. In addition, five 18-bp synthetic oligonucleotide primers homologous to conserved regions of the known AP and DQO sequences (8,(11)(12)(13)(14) were constructed and used to obtain sequence information on portions of the clones inaccessible from the EcoRI ends (Fig. 1, B-H productive (8, 9, 11-17).
MATERIALS AND METHODS