1. Clozapine has recently been claimed to behave as a selective and full agonist at the cloned m4 muscarinic receptor artificially expressed in Chinese hamster ovary (CHO) cells. In the present study we have investigated whether clozapine could activate the rat striatal muscarinic receptors coupled to the inhibition of adenylyl cyclase activity, considered as pharmacologically equivalent to the m4 gene product. In addition, we have examined the effect of the drug on various functional responses following the activation of the cloned m4 receptor expressed in CHO cells. 2. In rat striatum, clozapine (1 nM-10 microM) caused a slight inhibition of forskolin-stimulated adenylyl cyclase activity, which was not counteracted by 10 microM atropine. On the other hand, clozapine antagonized the inhibitory effect of acetylcholine with a pA2 value of 7.51. Moreover, clozapine (1 microM) failed to inhibit dopamine D1 receptor stimulation of adenylyl cyclase activity, but counteracted the inhibitory effect of carbachol (CCh). Clozapine displaced [3H]-N-methylscopolamine ([3H]-NMS) bound to striatal M4 receptors with a monophasic inhibitory curve and a pKi value of 7.69. The clozapine inhibition was not affected by the addition of guanosine-5'-O-(thio)triphosphate (GTPgammaS). 3. In intact CHO cells, clozapine inhibited forskolin-stimulated cyclic AMP accumulation with an EC50 of 31 nM. This effect was antagonized by atropine. CCh produced a biphasic effect on cyclic AMP levels, inhibiting at concentrations up to 1 microM (EC50=50 nM) and stimulating at higher concentrations (EC50 = 7 microM). Clozapine (0.3-5 microM) antagonized the CCh stimulation of cyclic AMP with a pKi value of 7.47. Similar results were obtained when the adenylyl cyclase activity was assayed in CHO cell membranes. 4. In CHO cells pretreated with the receptor alkylating agent 1-ethoxycarbonyl-2-ethoxy-1,2-dihydroquinoline (10 microM), the maximal inhibitory effect of clozapine on cyclic AMP formation was markedly reduced, whereas the CCh inhibitory curve was shifted to the right with no change in the maximum. 5. As in rat striatum, in CHO cell membranes the displacement of [3H]-NMS binding by clozapine yielded a monophasic curve which was not affected by GTPgammaS. 6. Clozapine (10 nM-10 microM) had a small stimulant effect (approximately 20%) on the binding of [35S]-GTPgammaS to CHO cell membranes, whereas CCh caused a 250% increase of radioligand binding. Moreover, clozapine (50 nM-5 microM) antagonized the CCh-stimulated [35S]-GTPgammaS binding with a pA2 value of 7.48. 7. These results show that at the striatal M4 receptors clozapine is a potent and competitive antagonist, whereas at the cloned m4 receptor it elicits both agonist and antagonist effects. Thus, clozapine behaves as a partial agonist, rather than as a full agonist, at the m4 receptor subtype, with intrinsic activity changing as a function of the coupling efficiency of the receptor to effector molecules.