The present study was intended to examine the relaxant effects of berberine in rat isolated mesenteric arteries. Berberine produced a rightward shift of the concentration-response curve to phenylephrine and significantly reduced the maximal contractile response to phenylephrine. Berberine (10(-7)-3x10(-5) M) also relaxed the phenylephrine- and 9,11-dideoxy-11alpha, 9alpha-epoxy-methanoprostaglandin F(2alpha)-precontracted arteries with respective IC(50) values of 1.48+/-0.16x10(-6) and 2.23+/-0. 22x10(-6) M. Removal of a functional endothelium significantly attenuated the berberine-induced relaxation (IC(50): 4.73+/-0. 32x10(-6) M) without affecting the maximum relaxant response. Pretreatment with N(G)-nitro-L-arginine methyl ester (L-NAME) or methylene blue reduced the relaxant effect of berberine, and L-arginine (10(-3) M) partially antagonized the effect of L-NAME. In contrast, pretreatment with 10(-6) M glibenclamide or 10(-5) M indomethacin had no effect. Berberine (10(-5) M) reduced over by 50% the transient contraction induced by caffeine or phenylephrine in endothelium-denuded rings bathed in Ca(2+)-free Krebs solution. Pretreatment with putative K(+) channel blockers, such as tetrapentylammonium ions (1-3x10(-6) M), 4-aminopyridine (10(-3) M), or Ba(2+) (3x10(-4) M), significantly attenuated the berberine-induced relaxation in endothelium-denuded arteries. In contrast, tetraethylammonium ions (3x10(-3) M), charybdotoxin (10(-7) M) or glibenclamide (10(-6) M) were without effect. Berberine reduced the high-K(+)-induced sustained contraction and the relaxant response to berberine was greater in rings with endothelium (IC(50): 4.41+/-0.47x10(-6) M) than in those without endothelium (IC(50): 8.73+/-0.74x10(-6) M). However, berberine (10(-6)-10(-4) M) did not affect the high-K(+)-induced increase of intracellular [Ca(2+)] in cultured aortic smooth muscle cells. Berberine did not affect active phorbol ester-induced contraction in Ca(2+)-free Krebs solution. In addition, berberine inhibited proliferation of cultured rat aortic smooth muscle cells with an IC(50) of 2.3+/-0.43x10(-5) M. These findings suggest that berberine could act at both endothelium and the underlying vascular smooth muscle to induce relaxation. Nitric oxide from endothelium may account primarily for the berberine-induced endothelium-dependent relaxation, while activation of tetrapentylammonium-, 4-aminopyridine- and Ba(2+)-sensitive K(+) channels, inhibition of intracellular Ca(2+) release from caffeine-sensitive pools, or a direct relaxant effect, is likely responsible for the berberine-induced endothelium-independent relaxation. Mechanisms related to either Ca(2+) influx or protein kinase C activation may not be involved. Both vasorelaxant and antiproliferative effects may contribute to a long-term benefit of berberine in the vascular system.