1. In the rat hepatic artery, the acetylcholine-induced relaxation mediated by endothelium-derived hyperpolarizing factor (EDHF) is abolished by a combination of apamin and charybdotoxin, inhibitors of small (SKCa) and large (BKCa) conductance calcium-sensitive potassium (K)-channels, respectively, but not by each toxin alone. The selective BKCa inhibitor iberiotoxin cannot replace charybdotoxin in this combination. Since delayed rectifier K-channels (KV) represent another target for charybdotoxin, we explored the possible involvement of KV in EDHF-mediated relaxation in this artery. 2. The KV inhibitors, agitoxin-2 (0.3 microM), kaliotoxin (0.3 microM), beta-dendrotoxin (0.3 microM), dofetilide (1 microM) and terikalant (10 microM), each in combination with apamin (0.3 microM) had no effect on the EDHF-mediated relaxation induced by acetylcholine in the presence of N omega-nitro-L-arginine (0.3 mM) and indomethacin (10 microM), inhibitors of nitric oxide (NO) synthase and cyclo-oxygenase, respectively (n = 2-3). Although the KV inhibitor margatoxin (0.3 microM) was also without effect (n = 5), the combination of margatoxin and apamin produced a small inhibition of the response (pEC50 and Emax values were 7.5 +/- 0.0 and 95 +/- 1% in the absence and 7.0 +/- 0.1 and 81 +/- 6% in the presence of margatoxin plus apamin, respectively; n = 6; P < 0.05). 3. Ciclazindol (10 microM) partially inhibited the EDHF-mediated relaxation by shifting the acetylcholine-concentration-response curve 12 fold to the right (n = 6; P < 0.05) and abolished the response when combined with apamin (0.3 microM; n = 6). This combination did not inhibit acetylcholine-induced relaxations mediated by endothelium-derived NO (n = 5). 4. A 4-aminopyridine-sensitive delayed rectifier current (IK(V)) was identified in freshly-isolated single smooth muscle cells from rat hepatic artery. None of the cells displayed a rapidly-activating and -inactivating A-type current. Neither charybdotoxin (0.3 microM; n = 3) nor ciclazindol (10 microM; n = 5), alone or in combination with apamin (0.3 microM; n = 4-5), had an effect on IK(V). A tenfold higher concentration of ciclazindol (0.1 mM, n = 4) markedly inhibited IK(V), but this effect was not increased in the additional presence of apamin (0.3 microM; n = 2). 5. By use of membranes prepared from rat brain cortex. [125I]-charybdotoxin binding was consistent with an interaction at a single site with a KD of approximately 25 pM. [125I]-charybdotoxin binding was unaffected by iberiotoxin (0.1 microM, n = 6), but was increased by apamin in a concentration-dependent manner (Emax 43 +/- 10%, P < 0.05 and pEC50 7.1 +/- 0.2; n = 7-8). Agitoxin-2 (10 nM) displaced [125I]-charybdotoxin binding by 91 +/- 3% (n = 6) and prevented the effect of apamin (1 microM; n = 6). 6. It is concluded that the EDHF-mediated relaxation in the rat hepatic artery is not mediated by the opening of either KV or BKCa. Instead, the target K-channels for EDHF seem to be structurally related to both KV and BKCa. The possibility that a subtype of SKCa may be the target for EDHF is discussed.