Activation of muscarinic cholinergic receptors on 1321N1 human astrocytoma cells results in attenuation of cyclic AMP accumulation, apparently as a consequence of increases in phosphoinositide hydrolysis, Ca2+ mobilization, and the activation of a Ca2+ calmodulin-regulated phosphodiesterase. Preincubation of these cells with carbachol for 30-90 min resulted in a 15-30-fold shift to the right of the concentration effect curves for carbachol or methacholine for attenuation of cyclic AMP accumulation, with no change occurring in the maximal effect observed with either agonist. In contrast, preincubation with carbachol for 30-90 min resulted in an essentially complete loss of effects on cyclic AMP accumulation of the muscarinic receptor agonists oxotremorine, arecoline, and bethanechol. In contrast to carbachol and methacholine, oxotremorine, arecoline, and pilocarpine were much less effective inducers of desensitization. Inactivation of muscarinic receptors with propylbenzylcholine mustard, or preincubation of cells with 4 beta-phorbol 12 beta-myristate 13 alpha-acetate, which has been shown to markedly decrease the phosphoinositide response of 1321N1 cells to cholinergic agonists, produced differential effects on the cyclic AMP response to carbachol and oxotremorine that were similar to those observed after preincubation with carbachol. The data can be explained by the presence of "reserve" in the series of steps that result in muscarinic receptor-mediated activation of phosphodiesterase. That is, it is proposed that carbachol and methacholine mobilize much more Ca2+ than is necessary for maximal activation of phosphodiesterase, whereas oxotremorine and several other muscarinic receptor agonists do not.