Superoxide dismutase (SOD) rapidly scavenges superoxide (O°) and also prolongs the vasorelaxant effects of nitric oxide (NO), thought to be the endotheliumderived relaxing factor. This prolongation has been ascribed to prevention of the reaction between°2-with NO. We report that SOD supports a reversible reduction of NO to NO-. When cyanamide and catalase were used to generate NO-in the presence of SOD, NO was measured by the conversion of HbO2 to MetHb. When SOD[Cu(I)] was exposed to NO anaerobicafly, NO-was trapped by MetHb forming nitrosylmyoglobin. When NO was generated by 3-morpholinosydnonimine hydrochloride in the presence of SOD, NO-or a similar reductant was formed, which reduced catalase compound II and promoted the formation of the catalase[Fe(H)J-NO complex. It is, therefore, conceivable that SOD may protect NO and endothelium-derived relaxing factor by a mechanism in addition toO°scavenging and that NO-may be a physiologically important form of endothelium-derived relaxing factor.Nitric oxide is thought to be the endothelium-derived relaxing factor (EDRF), a vasodilator produced from arginine. The similar chemical properties of NO and EDRF, including their apparent reactivity with°2, support this proposal (1-4). The reaction of NO with O2 (reactions la and b) has been directly observed by using pulse-radiolysis (5), but evidence of the reaction of EDRF with°2 is indirect. Namely, superoxide dismutase (SOD) prolongs the effects of both EDRF and exogenous NO, whether or not there is a simultaneous addition of compounds thought to generate O°. This consistent effect of SOD has been attributed (1-4) to its known ability to catalyze 0°dismutation (6).One report (5) suggests the rate constant for reaction la may be much slower than diffusion limited (-56 x 106 M-1 s-1 at 370C). Also, the intracellular concentration of°2 is estimated to be quite low [e.g., in liver <66 pM (7) Analyses of the metabolites of NO in vivo also fail to prove that reaction la is significant. Although peroxynitrite (ONOO-) can decompose to NO2 and 0OH in the absence of SOD (9) or may convert to NO' in its presence (10), nitrate (NO-) is still a major product of reaction la (5, 9). Therefore, when reaction la is significant, SOD should not only protect NO but should also decrease NO production in favor of NO and NO-. However, product analyses have failed to meet the latter expectation (11, 12) and led us to examine reactions involving SOD and NO in more detail. METHODS 3-Morpholinosydnonimine hydrochloride (SIN-1) was from Cassella AG (Frankfurt), CuZn-SOD (from bovine erythrocytes), cyanamide, and hydroxylamine were from Sigma, and catalase (from bovine liver) was from Boehringer Mannheim. HbO2 was prepared by reduction of MetHb (Aldrich) with dithionite, oxygenation, and purification on a Sephadex G-25 column (in Krebs/Hepes buffer containing 140mM NaCl, 2.7 mM KCl, 0.42 mM NaH2PO4, 1 mM MgCl2, 1.8 mM CaCI2, 20 mM Hepes, adjusted to pH 7.4 at 37°C by adding NaOH) and measured at 415 nm (e = 131 mM-1-cm-1) (13). Metmyoglob...