Authentic nitric oxide (NO; 0.1 - 10 micromoles) caused transient, dose-dependent relaxation of phenylephrine-induced tone without changing membrane potential in mesenteric arteries. Larger doses, above 10 micromoles, did not evoke more relaxation (maximal relaxation to 150 micromoles NO in denuded arteries, 69+/-7%, n=8) but stimulated muscle hyperpolarization (maximum 19+/-3 mV, n=5). The soluble guanylyl cyclase inhibitor, 1H-[1,2,4]oxadiazolo[4,3-a]quinoxalin-1-one (ODQ; 10 microM), abolished relaxation to low doses of NO (n=4), but did not modify hyperpolarization with higher doses of NO (n=4). The potassium channel blocker charybdotoxin (ChTX; 50 nM) abolished hyperpolarization to high doses of NO and significantly reduced the maximal relaxation (to 43+/-6%, n=4; P<0.01). ODQ and ChTX together abolished tension and membrane potential change to all doses of NO (n=4). All relaxations to 3-morpholino-sydnonimine (SIN-1; 0.01 - 10 microM) were associated with hyperpolarization. When the endothelium was intact, ChTX inhibited hyperpolarization and relaxation to SIN-1 (n=5), while iberiotoxin (IbTX; 50 nM) or 4-aminopyridine (4-AP; 500 microM) reduced relaxation by 40% and 20%, respectively and by 80% in combination (n=6 in each case). In denuded arteries, relaxation to SIN-1 was unaffected by either ChTX or ODQ alone, but abolished by the inhibitors together (n=6). Alone, 4-AP did not alter relaxation, but in the presence of ODQ it reduced the maximal response by around 45% (n=6; P<0.01). 4-AP, ODQ and IbTX together inhibited relaxation to SIN-1 by 75% (n=6; P<0.01). Therefore, cyclic guanosine 3',5'-monophosphate (cyclic GMP)-independent smooth muscle hyperpolarization, possibly involving direct activation of calcium-activated and voltage-sensitive potassium channels, contributes to relaxation evoked by authentic NO and SIN-1. However, the importance of each pathway depends on the source of NO and with SIN-1 the relative contribution from each pathway is modified by the endothelium.