1. The role of endothelium-derived hyperpolarizing factor and voltage-operated Ca2+ channels in mediating endothelium-dependent, NG-nitro-L-arginine (L-NOARG; 100 microM) -resistant relaxations to bradykinin (BK), was examined in isolated rings of endothelium-intact bovine left anterior descending coronary artery. 2. Rings of artery were contracted isometrically to approximately 40% or their respective maximum contraction to 125 mM KCl Krebs solution (KPSSmax) with the thromboxane A2-mimetic, U46619. Relaxations to BK and the endothelium-independent NO donor, S-nitroso-N-acetylpenicillamine (SNAP), were normalized as percentages of reversal of the initial contraction to U46619. All experiments were carried out in the presence of indomethacin (3 microM). 3. BK caused concentration-dependent relaxations [sensitivity (pEC50) 9.88 +/- 0.05; maximum relaxation (Rmax), 103.3 +/- 0.5%] in U46619-contracted rings of bovine coronary artery. L-NOARG (100 microM) caused a significant (P < 0.01) 3 fold reduction in the sensitivity to BK (pEC50, 9.27 +/- 0.11) without affecting the Rmax (101.8 +/- 2.3%). A similar, significant 3 fold reduction in sensitivity to BK with no change in Rmax was observed after treatment with oxyhaemoglobin (20 microM; pEC50, 9.18 +/- 0.13, P < 0.001) or a combination of oxyhaemoglobin (20 microM) and L-NOARG (100 microM; pEC50, 9.08 +/- 0.10, P < 0.001). Oxyhaemoglobin (20 microM) either alone or in combination with L-NOARG (100 microM) caused an approximate 600 fold decrease in the sensitivity to SNAP. 4. The L-type voltage-operated Ca2+ channel inhibitor, nifedipine (0.3 microM-3 microM), reduced the maximum contraction (Fmax) to isotonic 68 mM KCl Krebs solution (103.5 +/- 2.0% KPSSmax) by 85-90% (P < 0.001); yet, the highest concentration of nifedipine (3 microM) caused only a small but significant reduction in both the sensitivity and Fmax to U46619. By contrast, nifedipine (3 microM) had no effect on the relaxation response to BK. Furthermore, a combination of nifedipine (3 microM) and L-NOARG (100 microM) had no further inhibitory effects on relaxations to BK (pEC50, 8.79 +/- 0.10; Rmax, 101.7 +/- 2.4%) than did L-NOARG (100 microM) alone (pEC50, 9.05 +/- 0.12; Rmax, 99.62 +/- 1.19). Also, nifedipine (0.3 microM and 3 microM) had no effect on the maximum relaxation to the K+ channel opener, levcromakalim (0.3 microM). 5. In the presence of nifedipine (0.3 microM to control contractions induced by high KCl) and isotonic 68 mM KCl Krebs solution (to inhibit K+ channel activity), relaxations to BK (pEC50, 9.42 +/- 0.10; Rmax, 93.9 +/- 1.8%) were similar to those observed in normal Krebs solution (pEC50, 9.58 +/- 0.09; Rmax, 98.4 +/- 0.8%). However, in the presence of 68 mM KCl Krebs solution the inhibitory effect of L-NOARG (100 microM) on relaxations to BK (pEC50, 8.53 +/- 0.20; Rmax, 31.0 +/- 11.3%) was markedly greater than that in normal KCl Krebs solution (pEC50, 9.12 +/- 0.08; Rmax, 91.5 +/- 2.0%). Similar treatment with 68 mM KCl Krebs had no effect on relaxations to the NO donor, SNAP, yet abolished the response to the K+ channel opener, levcromakalim (0.3 microM). 6. In summary, this study has shown that (1) NO synthesis in response to BK in bovine coronary artery endothelial cells in situ is likely to be abolished by L-NOARG, (2) NO-independent relaxations to BK are markedly attenuated by 68 mM KCl-containing Krebs, which, in the absence of L-NOARG, had no effect, (3) nifedipine blocked contractions to a maximum-depolarizing stimulus (KCl) yet had no effect on NO-independent relaxations to BK, and (4) maximum relaxations to levcromakalim were abolished by 68 mM KCl Krebs but were not affected by nifedipine. Therefore, we hypothesize that if smooth muscle hyperpolarization is involved in non-NO-, endothelium-dependent relaxation in bovine coronary arteries contracted with U46619, then it can accomplish this via a mechanism which does not i