Using a chemiluminescence method in the present study, we measured nitric oxide and one-electron oxidation products of nitric oxide (NOX) released from porcine coronary artery segments in response to bradykinin, ADP, and the calcium ionophore A23187. Total NOX was compared with the bioactivity of endothelium-derived relaxing factors (EDRF) by a biodetector ring preparation before and after inhibition of L-arginine-dependent nitric oxide synthesis and in the presence of indomethacin. Under basal conditions, arterial segments released NOX and relaxed biodetector rings. Bradykinin, ADP, and A23187 elicited vasorelaxation greater than that observed basally; A23187, but not bradykinin or ADP, caused additional release of NOX greater than that measured basally. Hemoglobin completely reversed vasorelaxation elicited by all three agonists. We compared the amount of nitric oxide released under basal conditions and after stimulation with bradykinin, ADP, and A23187 with the amount of authentic nitric oxide necessary to elicit a bioequivalent response. Authentic nitric oxide did not account for the observed bioactivity as compared with the amount of nitric oxide actually measured in arterial segment effluent. To investigate whether a second non-nitric oxide-containing compound was responsible for the increased bioactivity and the discrepancy between the bioactivity and quantity of nitric oxide measured, we exposed arterial segments to omega-nitro-L-arginine methyl ester to inhibit L-arginine-dependent synthesis of nitroso compounds. The drug completely abolished the nitric oxide signal derived from both basally released and A23187-stimulated relaxing factor and completely reversed vasorelaxation. In contrast, omega-nitro-L-arginine methyl ester only partially reversed bradykinin-stimulated vasorelaxation.(ABSTRACT TRUNCATED AT 250 WORDS)