1. The aims of the present study were to investigate the effects of fresh garlic and one of its active metabolites, allicin, on rat isolated pulmonary arteries (RPA). 2. In endothelium-intact and phenylephrine-precontracted RPA, the addition of a water or a 5% ethanol extract of fresh garlic (1-500 microg/mL) resulted in a dose-dependent relaxation reaching a maximum (mean +/- SEM) of -91 +/- 3 and -93 +/- 2%, respectively, with an ED(50) of 113 +/- 12 and 106 +/- 10 microg/mL, respectively. The vasorelaxation was readily reversible upon washing and no tachyphylaxis was noted. 3. An extract of the external garlic storage leaf produced a significantly greater relaxation than the inner stem. Microfiltration of extracts with a 10,000 molecular sieve did not attenuate relaxation. Inactivation of alliinase and allicin formation, with either boiling of the garlic clove for 30 min or 100% ethanol treatment, completely abolished relaxation. In contrast, similar treatment of crushed garlic with formed allicin retained the relaxation response. 4. Pure allicin produced a similar relaxation as garlic extract, with an EC(50) of approximately 0.8 microg/mL. Disruption of endothelium or N(G)-nitro-L-arginine methyl ester pretreatment attenuated the relaxation, whereas indomethacin had no effect. 5. Prior garlic (500 microg/mL) treatment enhanced acetylcholine relaxation by shifting the response curve to the left, but had no effect on nitric oxide (NO) donor-induced responses. 6. These results demonstrate that garlic and the active metabolite allicin are capable of eliciting a NO-dependent relaxation in RPA and that this response is likely to be mediated via garlic activation of NO formation rather than its stabilization.