The effects of four ATP-dependent K+ channel blockers (hypoglycemic sulfonylureas) against morphine- and U50488H-induced antinociception were evaluated using the tail flick test in mice. None of the sulfonylureas tested significantly modified tail flick latency in control animals. However, i.c.v. pretreatment with gliquidone (0.4-1.6 micrograms/mouse), glipizide (2.5-10 micrograms/mouse), glibenclamide (10-40 micrograms/mouse) or tolbutamide (20-80 micrograms/mouse) dose dependently antagonized morphine-induced antinociception approximately equieffectively, the only difference being in potency: gliquidone > glipizide > glibenclamide > tolbutamide. This effect of sulfonylureas was very specific, since none antagonized the antinociception elicited by U50488H even at doses twice as great as the dose that induced maximum antagonism of morphine antinociception. Because morphine, but not U50488H, opens K+ channels in neurons and because the order of potency of the different sulfonylureas for blocking ATP-dependent K+ channels in neurons and for antagonizing morphine antinociception is the same, we suggest that morphine antinociception is mediated by the opening of ATP-dependent K+ channels.