The roles of kinin and protease-activated receptors (PAR) in endothelium-dependent relaxations to the serine protease, trypsin, were examined in rings of bovine left anterior descending coronary artery (LAD). Trypsin (0.01-30 U/ml) caused biphasic, endothelium-dependent relaxations-a high potency (0.01-0.3 U/ml), low efficacy relaxation [maximum relaxation (R (max)), 9.0 +/- 5.1%] followed by a lower potency (1-30 U/ml) but high efficacy (R (max), 90.4 +/- 5.5%) relaxation, which was abolished by aprotinin. Captopril (10 microM) caused an approximately 10-fold leftward shift of the second phase response such that the first phase was masked. The second phase relaxation to trypsin was inhibited in a concentration-dependent, non-surmountable manner by the B2 antagonist, HOE-140. At 3 nM HOE-140, the second phase response to trypsin was abolished unmasking the first phase. Kallikrein (0.0003-0.3 U/ml) caused monophasic, endothelium-dependent relaxations (R (max), 33.7 +/- 14.6%), which were potentiated by captopril (R (max), 94.2 +/- 1.0%) and abolished by HOE-140. In the presence of captopril, the second phase relaxation to trypsin was only minimally inhibited by either N(G)-nitro-L: -arginine (100 microM) or 67 mM [K(+)](o) alone but markedly reduced when these two treatments were combined (R (max), 26.1 +/- 11.6% versus 98.6 +/- 2.9% in controls). The PAR1-activating peptide, SFLLRN (0.1-30 microM), but not the PAR2-activating peptide, SLIGRL, caused concentration-dependent relaxations (pEC(50), 5.9 +/- 0.0%; R (max), 43.3 +/- 8.3%). In conclusion, trypsin causes endothelium-dependent relaxations in the bovine LAD predominantly via release of endogenous BK, which in turn activates endothelial B2 receptors. Only a minor role for PAR1-like receptors was evident in this tissue.