Acetylcholine relaxes rat aorta and increases aortic cyclic GMP levels by a mechanism (or mechanisms) dependent on the endothelium and on extracellular calcium. Therefore, the effects of representatives of different subclasses of calcium entry blockers, verapamil, nifedipine, diltiazem, and bepridil, on maximal acetylcholine (1 microM)-induced increases in cyclic GMP levels were investigated in rat isolated aorta. None of these compounds, at a concentration (3 microM) sufficient to maximally inhibit agonist-stimulated Ca2+ influx into vascular smooth muscle cells, significantly affected either the basal or the acetylcholine-stimulated tissue cyclic GMP levels. On replacing all but 20 mM Na+ by choline, a condition that might be expected to limit or even abolish Na+-Ca2+ exchange, or in the presence of amiloride (1 mM), an inhibitor of Na+-Ca2+ exchange, acetylcholine-stimulated increases in tissue cyclic GMP levels were abolished or inhibited by about 80%, respectively. In choline containing solution acetylcholine relaxant responses were abolished. The presence of amiloride, or the replacement of Na+ by choline, had no effect on increases in cyclic GMP levels evoked by sodium nitroprusside (0.3 microM), an agent that stimulates cyclic GMP formation in smooth muscle without intervention of the endothelium. Replacement of Na+ by Li+ but not the other treatments depressed basal tissue cyclic GMP levels by about 45% but did not abolish either acetylcholine- or sodium nitroprusside-induced relaxant responses. However, the time course of relaxant responses elicited by both these relaxant agonists in precontracted rat aortic rings with endothelium was altered by Li+ replacement; the half-time to relaxation to acetylcholine was increased by about 70-fold. It is concluded that calcium channels, as characterized in smooth muscle and cardiac tissue, are not involved in the stimulated liberation of an endothelial-derived relaxant factor by acetylcholine, but that an Na+-Ca2+ exchange process may be of importance.