Endothelial cells release several compounds, including prostacyclin, NO, and endothelium-derived hyperpolarizing factor (EDHF), that mediate the vascular effects of vasoactive hormones. The identity of EDHF remains unknown. Since arachidonic acid causes endothelium-dependent relaxations of coronary arteries through its metabolism to epoxyeicosatrienoic acids (EETs) by cytochrome P450, we wondered if the EETs represent EDHFs. Precontracted bovine coronary arteries relaxed in an endothelium-dependent manner to methacholine. The cytochrome P450 inhibitors, SKF 525A and miconazole, significantly attenuated these relaxations. They were also inhibited by tetraethylammonium (TEA),an inhibitor of Ca2+-activated K+ channels, and by high [K+]0 (20 mmol/L). Methacholine also caused hyperpolarization of coronary smooth muscle (-27 +/- 3.9 versus -40 +/- 5.1 mV), which was completely blocked by SKF 525A and miconazole. In vessels prelabeled with [3H] arachidonic acid, methacholine stimulated the release of 6-ketoprostaglandin F1alpha, 12-HETE, and the EETs. Arachidonic acid relaxed precontracted coronary arteries, which were also blocked by TEA, charybdotoxin, another Ca2+-activated K+ channel inhibitor, and high [K+]0. 14,15-EET, 11,12-EET, 8,9-EET, and 5,6-EET relaxed precontracted coronary vessels (EC50, 1 X 10(-6) mol/L). The four regioisomers were equally active. TEA, charybdotoxin, and high [K+]0 attenuated the EET relaxations. 11,12-EET hyperpolarized coronary smooth muscle cells from -37 +/- 0.2 to -59 +/- 0.3 mV. In the cell-attached mode of patch clamp, both 14,15-EET and 11,12-EET increased the open-state probability of a Ca2+-activated K+ channel in coronary smooth muscle cells. This effect was blocked by TEA and charybdotoxin. These data support the hypothesis that the EETs are EDHFs.