Generation of superoxide from nitric oxide synthase

The detection and quantification of superoxide anion in biological systems and from nitric oxide synthase (NOS) is one of the most challenging goals in the field of nitric oxide free radical research. Most of the available methodologies for detecting superoxide are indirect and unsuitable for detecting superoxide from NOS. In a recent contribution to this journal, Dr. Xu [1] questioned the formation of superoxide by endothelial NOS (eNOS) from electron paramagnetic resonance spin trapping experiments using 5,5-dimethyl-1-pyrroline N-oxide (DMPO). The author concluded that superoxide is not formed by eNOS [1]. Electron paramagnetic resonance (EPR or ESR) is the only direct method to quantify superoxide and other free radicals. However, under physiological conditions the steady-state concentration of superoxide is low, which hinders its direct EPR detection. The EPR spin trapping technique enables detection of transient free radicals such as superoxide and hydroxyl radical [2, 3]. This technique uses a nitrone compound that reacts with a free radical species to form a persistent radical adduct. For decades investigators used DMPO to detect superoxide. The first EPR spin trapping evidence for the calcium/calmodulin-dependent generation of superoxide from neuronal NOS was obtained using DMPO [4]. Although DMPO also detects superoxide from eNOS upon activation by calcium/calmodulin (Fig. 1, trace A; cf. trace C), the DMPO-superoxide adduct (DMPO-OOH) rapidly decays to form the DMPO-hydroxyl adduct (DMPO-OH) thus making superoxide quantification nearly impossible. Recently new and improved spin traps have been developed [5, 6]. These are the carboxylated analogs of DMPO, 5-ethoxycarbonyl-5-methyl-1-pyrroline N-oxide (EMPO) and the phosphorylated analog 5-diethoxyphosphoryl-5-methyl-1-pyrroline N-oxide (DEPMPO). Both EMPO and DEPMPO superoxide adducts do not decay to the corresponding hydroxyl adducts. Quantification of superoxide is, therefore, possible with these new spin traps [5]. As shown in Fig. 1, both EMPO (trace D) and DEPMPO (trace E) react with superoxide generated from eNOS to yield more persistent superoxide adducts than DMPO (trace F) as demonstrated by the higher signal-to-noise ratio of the EPR spectra (traces D and E) at the same rate of superoxide formation. It is important to note that increasing the concentration of DMPO in the incubation mixtures did not significantly improve superoxide detection, as shown in Fig. 1 (trace F). Superoxide was only marginally detected with resting enzyme and abolished by the addition of SOD (not shown). These data clearly demonstrate that DMPO is not an appropriate spin trap for detecting superoxide from NOS. In view of the recent advancements in EPR spin trapping detection of superoxide, we believe that the use of a more suitable spin trap would be desirable before drawing conclusions about NOS biochemistry based on spin trapping experiments with DMPO. Previously, we and others have presented definitive evidence for the generation of superoxide from the oxygenase domain of endothelial and neuronal NOS by a calcium/calmodulin-dependent mechanism [7, 8]. Thus, Dr. Xu's findings are dramatically opposite to those reported by us and others [4, 7, 8]. Using DEPMPO we estimated that 7 pmol eNOS presenting an enzyme activity of 133.3 nmol [14C]citrulline/min/mg protein generated 6.7±0.3 μM of superoxide adduct [7]. Current NOS research indicates that BH4 critically controls one- and two-electron mechanism of oxygen activation by NOS [8]. It is likely that a comprehensive study on the mechanism of oxygen activation by NOS will reveal the role of BH4 in superoxide, hydrogen peroxide and ultimately NOS turnover. This knowledge is critical to understanding the role of NOS in free radical pathology. All NOS isoforms (endothelial, neuronal and inducible) generate superoxide in the absence of BH4 at different rates. Extrapolation of these findings to a cellular milieu is undoubtedly a challenging task.