The effects of mercury on vascular smooth muscle results in vasoconstriction, but the mechanism of this action is not elucidated yet. To investigate this issue we examined the effects of HgCl(2) in the isolated rat tail vascular bed. The tail artery was dissected, cannulated, and perfused at a constant flow (2.5 ml/min) with Krebs solution plus EDTA 0.03 mM at 36 degrees C. After equilibration for 30 min the effects of increasing concentrations of HgCl(2) (0.5, 1, 2, 5, and 10 microM) on the perfusion pressure were investigated. Concentrations of HgCl(2), 2 microM and above, significantly increased perfusion pressure. Blockade of alpha receptors (prazosin 84 ng/ml) did not alter the responses to HgCl(2), suggesting that the metal does not induce the release of neurotransmitters from sympathetic nerve terminals. To investigate the possible role of endothelium on the vasoconstriction produced by HgCl(2), preparations were precontracted with 10(-7) M phenylepherine or perfused with 5 microM HgCl(2) for 20 min. Acetylcholine-vasodilated preparations precontracted with phenylepherine demonstrating the integrity of the endothelial nitric oxide-releasing mechanism. In contrast, after perfusion with 5 microM HgCl(2), the vasodilation produced by acetylcholine was abolished. In the presence of either phenylephrine or HgCl(2) the effects of sodium nitroprusside remained unchanged. Pretreatment with 30 microM indomethacin fully prevented the HgCl(2)-induced vasoconstriction. However, the endothelium-dependent vasodilation in response to acetylcholine was significantly reduced after indomethacin plus HgCl(2) treatment, meanwhile the vasodilation produced by nitroprusside remained unchanged. Pretreatment with L-arginine (1 mM) did not prevent the vasoconstriction induced by HgCl(2), nor did it restore the ability of acetylcholine to produce vasodilation, and it did not alter the response to sodium nitroprusside. The possibility of HgCl(2)'s actions mediated by the formation of free radicals was also investigated. The administration of 10 mM histidine significantly reduced the vasoconstrictor response if used before HgCl(2) treatment without improving the reduced vasodilation produced by acetylcholine. These results are consistent with the hypothesis that the vasoconstriction produced by HgCl(2) may be mediated by the formation of superoxide anions, stimulating the production of a COX-derived vasoconstrictor agent and by reducing the endothelial vasodilator activity.