BACKGROUND: Although hypoxic pulmonary vasoconstriction (HPC) and hypoxic coronary vasodilatation (HCD) have been recognized by many researchers, the precise mechanism remains unknown. As isolated arteries will constrict or relax in vitro in response to hypoxia, the oxygen sensor/transduction mechanism must reside in the arterial smooth muscle, the endothelium, or both. Unfortunately, much of the current evidence is conflicting, especially concerning to the dependency of HPC and HCD on the endothelium and the role of the K+ channel. Therefore, this experiment was attempted to clarify the dependency of HPC and HCD on the endothelium and the role of the K+ channel on HPC and HCD. METHODS: HPC was investigated in isolated main pulmonary arteries precontracted with norepinephrine (NE). HCD was investigated in isolated left circumflex coronary artery precontracted with prostaglandin F2 alpha. Vascular rings were suspended for isometric tension recording in an organ chamber filled with Krebs-Henseleit solution. Hypoxia was induced by gassing the chamber with 95% N2 +5% CO2, which was maintained for 15 - 25 min. RESULTS: 1)Hypoxia elicited a vasoconstriction in NE-precontracted pulmonary arteries with endothelium, but a vasodilatation in PGF 2 alpha-precontracted coronary arteries with and without endothelium. There was no difference between the amplitude of the HPC and HCD induced by two consecutive hypoxic challenges and the effect of normoxic and hyperoxic control Krebs-Henseleit solution on subsequent response to hypoxia. 2)Inhibition of NO synthesis by the treatment with Nw-nitro-L-arginine reduced HPC in pulmonary arteries, but inhibition of the cyclooxygenase pathway by treatment with indomethacin had no effect on HPC and HCD, respectively. 3)Blockades of the TEA-sensitive K+ channel abolished HPC and HCD. 4)Apamin, a small conductance Ca2+/-activated K+ (KCa) channel blocker, and iberiotoxin, a large conductance KCa channel blocker, had no effect on the HCD. 5)Glibenclamide, an ATP-sensitive K+ (KATP) channel blocker, reduced HCD. 6)Cromakalim, an K(ATP) channel opener, relaxed the coronary artery precontracted with prostaglandin F2 alpha. The degree of relaxation by cromakalim was similar to that by hypoxia and glibenclamide reduced both hypoxia- and cromakalim-induced vasodilations. 7)Verapamil, a Ca2+ entry blocker, caffeine, a Ca2+ emptying drug; and ryanodine, an inhibitor of Ca2+ release from SR, reduced HPC, respectively. CONCLUSION: HPC is dependent on the endothelium and is considered to be induced by inhibition of the mechanisms of NO-dependent vasodilation while HCD is independent of the endothelium and is considered to be induced by activation of the K(ATP) channel.