We have previously shown that platelet ADP receptors are coupled to G-proteins by measuring the binding of [35S]guanosine-5'-[gamma-thio]triphosphate ([35S]GTP gamma S) to human platelet membranes stimulated with ADP. In order to identify the activated G-proteins, we used an approach which combines photolabelling of receptor-activated G-proteins with 4-azidoanilido-[alpha-32P]GTP and immunoprecipitation of the G-protein alpha-subunits with subtype-specific antibodies. Stimulation of human platelet membranes with ADP resulted in an increase in 4-azidoanilido-[alpha-32P]GTP incorporation into the immunoprecipitates of G alpha i but not of G alpha q proteins, whereas stimulation with the thromboxane analogue U46619 resulted in an increase in 4-azidoanilido-[alpha-32P]GTP incorporation into the immunoprecipitates of G alpha q but not of G alpha i proteins, and thrombin activated both G-proteins. This effect of ADP was concentration dependent and inhibited by the class P2 purinoceptor (P2T) antagonist ATP. Using specific antisera against subtypes of Gi proteins, we found that ADP stimulated labelling of the G alpha 12 immunoprecipitate, but not of the G alpha 13 precipitate. G alpha i1 was not detectable by immunoblotting of platelet membrane proteins. These data suggest that ADP inhibits cAMP formation by activation of G alpha 12 proteins and add evidence in support of the hypothesis that human platelet ADP receptors do not activate PLC through Gq activation.