Selection of antibody epitopes in an immunopathogenic neural autoantigen

Results with two well-characterized self-antigens, cytochrome c and myelin basic protein, have led to differing opinions regarding the predominant specificities of autoantibodies, whether regions of sequence diversity or 'structurally inherent features' of a protein determine favored antigenic sites. To further examine this question, 16 antibody epitopes have been mapped on a highly immunopathogenic autoantigen, retinal S-antigen (S-Ag). The epitopes were characterized for: (1) sequence diversity and cross-reactivity on S-antigens from several species; (2) conformational dependency; and (3) probability of their occurrence on the surface of S-antigen. A single C-terminal region containing sequence diversity was most frequently recognized, but no evidence for recognition of any other regions of sequence diversity was found. Thirteen of 16 monoclonal antibodies raised to native S-Ag bound epitopes strongly predicted to be on the surface of S-antigen. Conversely, only one of six antibody preparations raised to peptides or affinity-purified on peptides was found to recognize an epitope predicted to be on the surface, suggesting a good correlation between specificity for conformation-dependent sites and surface probability based on the surface prediction algorithm. Three of these six antibodies which preferred denatured epitopes bound sites which overlapped or coincided with T cell sites; two of these T cell sites are immunopathogenic. The epitopes recognized on denatured antigen and peptides were similar whether the antibodies were elicited with intact human or bovine S-antigen or with cyanogen bromide-cleaved peptides. Our data suggests that in the case of S-antigen, structural features are more significant factors in epitope selection than sequence diversity.