Key result
Overwhelming experimental evidence supports uncoupled superoxide formation as an intrinsic catalytic function of all three nitric oxide synthase isoforms, contrary to recent claims of autoxidation.
Key points are not available for this paper at this time.
Design
Editorial
Answers are grounded in this paper and cite their sources.
This commentary defends the established biochemical concept that nitric oxide synthases enzymatically generate superoxide, refuting a recent paper that attributed this to non-enzymatic autoxidation.
During the last decade, dozens of studies from at least 10 independent laboratories have demonstrated enzymatic superoxide formation by nitric oxide synthases (NOS). However, in a recent article published in FEBS Letters Dr. Xu suggested that the observed superoxide generation resulted from autoxidation of redox-active cofactors and not from an enzymatic reaction [1]. In the following I wish to summarize the overwhelming experimental evidence in favor of uncoupled superoxide formation as an intrinsic catalytic function of all three NOS isoforms. In addition, I will provide a possible explanation for Dr. Xu's negative findings. In 1991, neuronal NOS was identified as a complex oxidoreductase with multiple enzymatic functions [2]. One of the reactions catalyzed by the enzyme purified from pig brain was the oxidation of NADPH in the absence of the substrate L-arginine. This reaction, resembling uncoupled substrate oxidation by cytochrome P450s, was accompanied by consumption of molecular oxygen and strictly dependent on the presence of Ca2+/calmodulin. In subsequent studies it was shown that subsaturating concentrations of either the substrate L-arginine or the cofactor tetrahydrobiopterin (BH4) provoke the uncoupling of NADPH-dependent reductive oxygen activation from L-arginine oxidation, resulting in the generation of superoxide and hydrogen peroxide instead of NO [3, 4]. The transfer of electrons from NADPH to the heme iron, where the reductive oxygen activation takes place, requires bound calmodulin (for review see [5]). Thus, based on current knowledge it is no surprise that peroxide formation by neuronal NOS was found to be strictly dependent on the presence of Ca2+/calmodulin and to be blocked by the heme site inhibitor N G-nitro-L-arginine [3, 4]. More recently, similar data were obtained with the inducible [6] and endothelial [7] isoforms, demonstrating that enzymatic formation of superoxide/hydrogen peroxide is a general feature of NOS and that these reduced oxygen species are certainly not ‘non-specific by-products’ of the reaction, as suggested by Dr. Xu [1]. Dr. Xu is correct in stating that some of the reducing cofactors essential to NOS catalysis, especially BH4 and flavin mononucleotide, produce superoxide in the course of autoxidation reactions. Albeit this non-enzymatic superoxide production results in the inactivation of enzymatically produced NO and interferes with the measurement of free NO in vitro [8], it must not be confused with the highly specific and tightly regulated NOS-catalyzed reaction. Thus, Ca2+/calmodulin-dependent superoxide formation by BH4-deficient neuronal NOS (saturated with L-arginine) is completely inhibited by excess BH4, demonstrating the essential role of the pterin for the coupling of NADPH-dependent oxygen reduction to L-arginine oxidation [3]. In other words, the non-enzymatic autoxidation reactions stressed by Dr. Xu are not relevant when peroxide formation is measured against Ca2+- or calmodulin-deficient blanks. The EPR data Dr. Xu shows in Fig. 1 of the report [1] are in striking disagreement with all previous EPR spin trapping studies performed with NOS (e.g. [4, 6, 7]). It is a difficult task to explain this discrepancy with the limited amount of information given in the paper. Dr. Xu observed that addition of purified neuronal NOS to a mixture of cofactors did not increase the EPR signal obtained with the cofactors alone. However, it remains unclear how much NOS-derived superoxide was expected and whether the experimental design would have made it possible to detect the difference. Unfortunately, neither the source nor the specific activity of the enzyme preparation is given. In fact, Dr. Xu has not even tested whether the added enzyme was active at all under the conditions of superoxide detection. In this context it is puzzling that, according to the figure legend, the EPR measurements were performed in the absence of added Ca2+ ions. Although sufficient (micromolar) Ca2+ is usually present in buffer solutions, enzyme preparations often contain chelators to inhibit tissue proteases. Considering the fairly high final NOS concentration of 12 μg/ml, chelators possibly present in the enzyme stock solutions could have reduced the free Ca2+ concentration in the final assay mixture below the threshold level of ∼0.5 μM required for calmodulin binding to neuronal NOS. I wish to emphasize that this is only one out of many possible reasons for the apparent lack of activity of neuronal NOS in Dr. Xu's experiments. In any case, a careful experimenter is expected to perform appropriate positive controls before questioning a well established scientific concept with preliminary negative data.
No takes yet. Share an insight, caveat, or question.
Bernd Mayer (2000) reported a letter. Overwhelming experimental evidence supports uncoupled superoxide formation as an intrinsic catalytic function of all three nitric oxide synthase isoforms, contrary to recent claims of autoxidation.
Synapse has enriched 10 related papers. Consider them for comparative context: