The gamma-aminobutyrate-synthesizing enzyme glutamate decarboxylase (GAD; L-glutamate 1-carboxy-lyase, EC 4.1.1.15) is a major target of autoantibodies associated with both early and late stages of pancreatic beta-cell destruction and development of type 1 diabetes. We have used five monoclonal anti-islet-cell antibodies (MICAs 1,2,3,4, and 6) derived from a newly diagnosed diabetic patient to probe the autoimmune epitopes in the enzyme. All the MICAs specifically recognized the smaller GAD protein, GAD65, and did not recognize the nonallelic GAD67 protein. A series of N-terminal, C-terminal, and internal deletion mutants, as well as protein footprinting, were used to identify the target regions in GAD65. Immunoprecipitation revealed two major native epitope areas in the GAD65 molecule. The first, defined by MICAs 1 and 3, is destroyed by deleting 41 amino acids at the C terminus but is also dependent on intact amino acids 244-295. This epitope (or epitopes) may span both middle and C-terminal domains of the protein. The second conformational epitope region, defined by MICAs 4 and 6, is dependent on intact amino acids 245-295 but is not affected by deletion of 110 amino acids at the C terminus and is therefore confined to domain(s) in the middle of the molecule. MICA 2 recognizes a linear epitope close to the C terminus. Thus, the N-terminal domain of GAD65, which differs most significantly from GAD67, does not harbor the MICA epitopes. Rather subtle amino acid differences in the middle and C-terminal domains define the GAD65-specific autoimmune epitopes. Analysis of sera from 10 type 1 diabetic patients suggests that MICAs 1, 3, 4, and 6 represent a common epitope recognition in this disease, whereas the MICA 2 epitope is rare. Furthermore, autoantibodies in some sera are restricted to the MICA 1/3 epitope, suggesting that this epitope may represent a single dominant epitope in the early phases of beta-cell autoimmunity.