Potency and intrinsic activity of agonists depend on ligand structure, but are also regulated by receptor-G protein stoichiometry. A potential functional reserve in adenosine A(1) receptor-mediated G protein activation was investigated by stimulation of guanosine-5'-(gamma-[35S]thio)-triphosphate ([35S]GTPgammaS) binding by the full agonist 2-chloro-N(6)-cyclopentyladenosine (CCPA) and the partial agonist 5'-deoxy-5'-methylthioadenosine (MeSA). Pretreatment of rat brain membranes with the irreversible antagonist 1-propyl-3-[3-[[4-(fluorosulfonyl)benzoyl]oxy]-propyl]-8-cyclopentylxanthine revealed no classical receptor reserve for either agonist. The functional significance of the G protein coupling state of the receptor and occupancy of G proteins by guanine nucleotides was assessed after partial uncoupling of receptor-G protein complexes with N-ethylmaleimide and in the presence of increasing GDP concentrations. Agonist EC(50) values in G protein activation were increased after NEM pretreatment and at higher GDP concentrations, and a decrease in the relative intrinsic activity of MeSA was observed. The shift of agonist concentration-response curves to the right, the decrease in maximal effects and the decrease in relative intrinsic activity of the partial agonist point to a functional reserve which has to be attributed to GDP-free receptor-G protein complexes. The mechanisms of action of FSCPX, NEM and GDP were fully consistent with the two-state model of receptor activation. The apparent reserve revealed by GDP reflects a shift from spontaneously active GDP-free receptor-G protein complexes (RG)(*), which can bind [35S]GTPgammaS, to (RG) occupied by GDP. The abundance of (RG)(*) is favored by agonists and by the absence of GDP.