We have recently reported that exogenous thrombin produced a dose- and endothelium-dependent coronary vasodilation in both intact open-chested dogs and in isolated dog coronary artery preparations. To determine whether the observed vasodilatory effect may be related to thrombin proteolytic enzymatic activity, effects of other proteases, such as trypsin, chymotrypsin, and pepsin, on the mechanical responses of isolated dog coronary arteries were studied. Among the four proteases evaluated, only thrombin (0.01-0.1 U/ml) and trypsin (0.03-0.67 U/ml) consistently produced a potent dose- and endothelium-dependent relaxation, that was reproducible with repeated testings. Addition of chymotrypsin (0.01-1.0 U/ml) produced only a minimal effect and was not reproducible, while addition of pepsin, as much as 10 U/ml, did not produce any effect. The specific soybean trypsin inhibitor and aprotinin, but not heparin and hirudin, competitively shifted the trypsin dose-response to the right, whereas heparin, hirudin, and antithrombin III proved to be more effective than trypsin inhibitors in inhibiting the thrombin-induced vasodilation. In all cases, the thrombin- and trypsin-induced vasodilation were equally sensitive to inhibition by the specific synthetic thrombin inhibitor, PPACK (D-phenylalanyl-L-propyl-L-arginine chloromethyl ketone, 1-30 nM). PPACK, however, had no effect on the other endothelium-dependent coronary vasodilators, such as acetylcholine and adenosine triphosphate, in our isolated dog coronary artery preparations. Biochemical determinations of the amidolytic activity of thrombin, using Tosylglycyl-L-prolyl-L-arginine-p-nitroanilide as a chromogen, also indicated a similar PPACK and heparin-antithrombin III dose-dependent inhibition of the thrombin enzymatic activity.(ABSTRACT TRUNCATED AT 250 WORDS)