It has previously been shown that covalent incorporation of the photoreactive adenosine derivative (R)-2-azido-N6-p-hydroxy-phenylisopropyladenosine [(R)-AHPIA] into the A1 adenosine receptor of intact fat cells leads to a persistent activation of this receptor, resulting in a reduction of cellular cAMP levels [Mol. Pharmacol. 30:403-409 (1986)]. In contrast, covalent incorporation of (R)-AHPIA into human platelet membranes, which contain only stimulatory A2 adenosine receptors, reduces adenylate cyclase stimulation via these receptors. This effect of (R)-AHPIA is specific for the A2 receptor and can be prevented by the adenosine receptor antagonist theophylline. Binding studies indicate that up to 90% of A2 receptors can be blocked by photoincorporation of (R)-AHPIA. However, the remaining 10-20% of A2 receptors are sufficient to mediate an adenylate cyclase stimulation of up to 50% of the control value. Similarly, the activation via these 10-20% of receptors occurs with a half-life that is only 2 times longer than that in control membranes. This indicates the presence of a receptor reserve, with respect to both the extent and the rate of adenylate cyclase stimulation. These observations require a modification of the models of receptor-adenylate cyclase coupling, which is described in the accompanying paper [Mol. Pharmacol. 39:524-530 (1991)].