To characterize vasodilator capacity of small coronary arteries (200-350 mu m diameter) in the setting of congestive heart failure, we examined relaxation responses to acetylcholine (10(-9)-10(-4) M) and nitroglycerin (10(-9)-10(-4) M), in the absence and presence of the nitric oxide precursor, L-arginine (10(-4) M). Congestive heart failure was reliably induced in dogs by rapid ventricular pacing (250 beats.min(-1) for 4 weeks). Maximum relaxations (means +/- S.E.) to each vasodilator are expressed as a percentage of the relaxation response to papaverine (10(-4) M). Relaxation responses to the endothelium-dependent relaxing agent, acetylcholine, were not altered at heart failure, or in the presence of L-arginine. Contrary to acetylcholine, relaxations to nitroglycerin were significantly enhanced in heart failure compared to control (83 +/- 25% vs. 25 +/- 6%, respectively, P < 0.05). Although L-arginine, alone, did not cause any vasodilator response in coronary microvessels, it was able to potentiate nitroglycerin relaxations at control (no L-arginine: 25 +/- 6% vs. L-arginine: 135 +/- 66%). In contrast, at heart failure, L-arginine diminished nitroglycerin relaxations (no L-arginine: 83 +/- 25%, vs. L-arginine: 48 +/- 15%). These data indicate a unique vasodilator profile in small coronary arteries at heart failure: endothelium-dependent relaxations are unaltered, whereas responses to nitroglycerin are augmented. Addition of the nitric oxide precursor, L-arginine, did not affect acetylcholine relaxation, yet surprisingly had a differential effect in response to nitroglycerin. Moreover, inhibition of nitric oxide synthase with N-omega-nitro-L-arginine elicited concentration-dependent constriction in heart failure but not control coronary microvessels. In summary, our study suggests an important role for nitric oxide in vasodilator control of coronary microvessels, which may modify nitrovasodilator therapy in congestive heart failure.