Aldehyde oxidase functions as a superoxide generating NADH oxidase: an important redox regulated pathway of cellular oxygen radical formation

The enzyme aldehyde oxidase (AO) is a member of the molybdenum hydroxylase family that includes xanthine oxidoreductase (XOR); however, its physiological substrates and functions remain unclear. Moreover, little is known about its role in cellular redox stress. Utilizing electron paramagnetic resonance spin trapping, we measured the role of AO in the generation of reactive oxygen species (ROS) through the oxidation of NADH and the effects of inhibitors of AO on NADH-mediated superoxide (O(2)(??)) generation. NADH was found to be a good substrate for AO with apparent K(m) and V(max) values of 29 μM and 12 nmol min(-1) mg(-1), respectively. From O(2)(??) generation measurements by cytochrome c reduction the apparent K(m) and V(max) values of NADH for AO were 11 μM and 15 nmol min(-1) mg(-1), respectively. With NADH oxidation by AO, ≥65% of the total electron flux led to O(2)(??) generation. Diphenyleneiodonium completely inhibited AO-mediated O(2)(??) production, confirming that this occurs at the FAD site. Inhibitors of this NADH-derived O(2)(??) generation were studied with amidone the most potent exerting complete inhibition at 100 μM concentration, while 150 μM menadione, raloxifene, or β-estradiol led to 81%, 46%, or 26% inhibition, respectively. From the kinetic data, and the levels of AO and NADH, O(2)(??) production was estimated to be ~89 and ~4 nM/s in liver and heart, respectively, much higher than that estimated for XOR under similar conditions. Owing to the ubiquitous distribution of NADH, aldehydes, and other endogenous AO substrates, AO is predicted to have an important role in cellular redox stress and related disease pathogenesis.     

Role of cellular superoxide dismutase against reactive oxygen metabolite injury in cultured bovine aortic endothelial cells.

We examined the protective effect of cellular superoxide dismutase against extracellular hydrogen peroxide in cultured bovine aortic endothelial cells. 51Cr-labeled cells were exposed to hydrogen peroxide generated by glucose oxidase/glucose. Glucose oxidase caused a dose-dependent increase of 51Cr release. Pretreatment with diethyldithiocarbamate enhanced injury induced by glucose oxidase, corresponding with the degree of inhibition of endogenous superoxide dismutase activity. Inhibition of cellular superoxide dismutase by diethyldithiocarbamate was not associated either with alteration of other antioxidant defenses or with potentiation of nonoxidant injury. Enhanced glucose oxidase damage by diethyldithiocarbamate was prevented by chelating cellular iron. Inhibition of cellular xanthine oxidase neither prevented lysis by hydrogen peroxide nor diminished enhanced susceptibility by diethyldithiocarbamate. These results suggest that, in cultured endothelial cells: 1) cellular superoxide is involved in mediating hydrogen peroxide-induced damage; 2) superoxide, which would be generated upon exposure to excess hydrogen peroxide independently of cellular xanthine oxidase, promotes the Haber-Weiss reaction by initiating reduction of stored iron (Fe3+) to Fe2+; 3) cellular iron catalyzes the production of a more toxic species from these two oxygen metabolites; 4) cellular superoxide dismutase plays a critical role in preventing hydrogen peroxide damage by scavenging superoxide and consequently by inhibiting the generation of the toxic species.     

Kinetics and mechanisms of catalase in peroxisomes of the mitochondrial fraction

1. The primary intermediate of catalase and hydrogen peroxide was identified and investigated in peroxisome-rich mitochondrial fractions of rat liver. On the basis of kinetic constants determined in vitro, it is possible to calculate with reasonable precision the molecular statistics of catalase action in the peroxisomes. 2. The endogenous hydrogen peroxide generation is adequate to sustain a concentration of the catalase intermediate (p(m)/e) of 60-70% of the hydrogen peroxide saturation value. Total amount of catalase corresponds to 0.12-0.15nmol of haem iron/mg of protein. In State 1 the rate of hydrogen peroxide generation corresponds to 0.9nmol/min per mg of protein or 5% of the mitochondrial respiratory rate in State 4. 3. Partial saturation of the catalase intermediate with hydrogen peroxide (p(m)/e) in the mitochondrial fraction suggests its significant peroxidatic activity towards its endogenous hydrogen donor. A variation of this value (p(m)/e) from 0.3 in State 4 to 0 under anaerobic conditions is observed. 4. For a particular preparation the hydrogen peroxide generation rate in the substrate-supplemented State 4 corresponds to 0.17s(-1) (eqn. 6), the hydrogen peroxide concentration to 2.5nm and the hydrogen-donor concentration (in terms of ethanol) to 0.12mm. The reaction is 70% peroxidatic and 30% catalatic. 5. A co-ordinated production of both oxidizing and reducing substrates for catalase in the mitochondrial fraction is suggested by a 2.2-fold increase of hydrogen peroxide generation and a threefold increase in hydrogen-donor generation in the State 1 to State 4 transition. 6. Additional hydrogen peroxide generation provided by the urate oxidase system of peroxisomes (8-12nmol of uric acid oxidized/min per mg of protein) permits saturation of the catalase with hydrogen peroxide to haem occupancy of 40% compared with values of 36% for a purified rat liver catalase ofk(1)=1.7x10(7)m(-1).s(-1) and k'(4)=2.6x10(7)m(-1). s(-1)(Chance, Greenstein & Roughton, 1952). 7. The turnover of the catalase ethyl hydrogen peroxide intermediate (k'(3)) in the peroxisomes is initially very rapid since endogenous hydrogen peroxide acts as a hydrogen donor. k'(3) decreases fivefold in the uncoupled state of the mitochondria.     

Identification of superoxide production by Arabidopsis thaliana aldehyde oxidases AAO1 and AAO3

Plant aldehyde oxidases (AOs) have gained great attention during the last years as they catalyze the last step in the biosynthesis of the phytohormone abscisic acid by oxidation of abscisic aldehyde. Furthermore, oxidation of indole-3-acetaldehyde by AOs is likely to represent one route to produce another phytohormone, indole-3-acetic acid, and thus, AOs play important roles in many aspects of plant growth and development. In the present work we demonstrate that heterologously expressed AAO1 and AAO3, two prominent members of the AO family from Arabidopsis thaliana, do not only generate hydrogen peroxide but also superoxide anions by transferring aldehyde-derived electrons to molecular oxygen. In support of this, superoxide production has also been found for native AO proteins in Arabidopsis leaf extracts. In addition to their aldehyde oxidation activity, AAO1 and AAO3 were found to exhibit NADH oxidase activity, which likewise is associated with the production of superoxide anions. According to these results and due to the fact that molecular oxygen is the only known physiological electron acceptor of AOs, the production of hydrogen peroxide and/or superoxide has to be considered in any physiological condition in which aldehydes or NADH serve as substrate for AOs. In this respect, conditions such as natural senescence and stress-induced stomatal movement, which both require simultaneously elevated levels of abscisic acid and hydrogen peroxide/superoxide, are likely to benefit from AOs in two ways, namely by formation of abscisic acid and by concomitant formation of reactive oxygen species.     

Hyperoxia enhances lung and liver nuclear superoxide generation

Porcine lung and liver nuclei generated superoxide (O-2) at a rate which increased with increasing oxygen concentration. NADH-dependent O-2 generation increased from 0 to 2.21 +/- 0.11 nmol/min per mg protein for lung nuclei and from 0.16 +/- 0.09 to 1.34 +/- 0.14 nmol/min per mg protein for liver nuclei, when oxygen concentration increased from 0 to 100%. NADPH-dependent O-2 generation increased similarly in liver nuclei (from 0.20 +/- 0.09 to 1.20 +/- 0.12 nmol/min per mg protein), while lung nuclei produced only 0.45 +/- 0.09 nmol/min per mg protein at 100% oxygen. NADH and NADPH had an additive effect on O-2 generation by liver nuclei, yielding 2.58 +/- 0.21 nmol/min per mg protein at 100% oxygen. Very little or no superoxide dismutase activity was present in washed nuclear preparations. The oxygen-dependence of nuclear O-2 generation shows that nuclear-derived partially reduced species of oxygen may affect nuclear function during hyperoxia or other metabolic situations where overproduction of oxygen radicals is problematic.     

Determination of hydrogen peroxide generated by reduced nicotinamide adenine dinucleotide oxidase

Hydrogen peroxide can be conveniently determined using horseradish peroxidase (HRP) and 2,2'-azino-bis(3-ethylbenzthiazoline-6-sulfonic acid). However, interference occurs among assay components in the presence of reduced nicotinamide adenine dinucleotide (NADH) that is also a substrate of NADH oxidase. So, depletion of NADH is required before using the HRP method. Here, we report simple and rapid procedures to accurately determine hydrogen peroxide generated by NADH oxidase. All procedures developed were based on the extreme acid lability of NADH and the stability of hydrogen peroxide, because NADH was decomposed at pH 2.0 or 3.0 for 10 min, while hydrogen peroxide was stable at pH 2.0 or 3.0 for at least 60 min. Acidification and neutralization were carried out by adjusting sample containing NADH up to 30 microM to pH 2.0 for 10 min before neutralizing it back to pH 7.0. Then, hydrogen peroxide in the sample was measured using the HRP method and its determination limit was found to be about 0.3 microM. Alternatively, hydrogen peroxide in samples containing NADH up to 100 microM could be quantitated using a modified HRP method that required an acidification step only, which was found to have a determination limit of about 3 microM hydrogen peroxide in original samples.     

Organ distribution of aldehyde dehydrogenase activity in the rainbow trout (Salmo gairdneri Richardson)

1. Aldehyde dehydrogenase activity was measured in gills, muscle, brain, intestine, kidney, heart and liver of rainbow trout, using 3,4-dihydroxyphenylacetaldehyde (the biogenic aldehyde derived from dopamine) as the substrate. 2. Aldehyde dehydrogenase activity was found to be present in all of the organs studied. 3. The highest activity was found in the liver (276 nmol/min.g wet wt of tissue). 4. A remarkably high activity was found in the heart (117 nmol/min.g). 5. The gills showed the lowest activity (1.9 nmol/min.g).     

Aldehyde dehydrogenase activity in blood from various vertebrates

1. Aldehyde dehydrogenase activity was determined in whole blood samples from 17 selected vertebrates of 5 classes, using 3,4-dihydroxyphenylacetaldehyde (the aldehyde derived from dopamine) as substrate. 2. Aldehyde dehydrogenase activity in blood was widely but unevenly distributed among the species studied. 3. Mean aldehyde dehydrogenase activities in the range of 40-140 nmol/min.ml blood (measured at 37 degrees C, pH 8.8) were found in blood from man, monkey, rabbit, guinea pig and mouse (C57BL and NMRI strains), with the highest activity in rabbit blood. 4. Much lower aldehyde dehydrogenase activities (0.5-7.5 nmol/min.ml blood) were found in blood from Sprague-Dawley and Wistar rat, dog, cat, horse, pig, chicken, caiman, frog and rainbow trout, whereas the activities in blood from DBA mouse, cow, sheep and crucian carp were close to the detection limit.     

Cellularly Generated Active Oxygen Species and Hela Cell Proliferation

In HeLa cells evidence is provided that active oxygen species such as hydrogen peroxide and superoxide at low levels are important growth regulatory signals. They may constitute a novel regulatory redox system of control superimposed upon the established cell growth signal transduction pathways. Whilst for example hydrogen peroxide can be added exogenously to elicit growth responses in these cells, it is clear that cellularly generated superoxide and hydrogen peroxide are important. Experiments with superoxide dismutase, superoxide dismutase mimics and inhibitors of both superoxide dismutase and xanthine oxidase suggest that superoxide generated intracellularly and superoxide released extracellularly are both relevant to growth control in HeLa cells.     

Cellularly Generated Active Oxygen Species and Hela Cell Proliferation

In HeLa cells evidence is provided that active oxygen species such as hydrogen peroxide and superoxide at low levels are important growth regulatory signals. They may constitute a novel regulatory redox system of control superimposed upon the established cell growth signal transduction pathways. Whilst for example hydrogen peroxide can be added exogenously to elicit growth responses in these cells, it is clear that cellularly generated superoxide and hydrogen peroxide are important. Experiments with superoxide dismutase, superoxide dismutase mimics and inhibitors of both superoxide dismutase and xanthine oxidase suggest that superoxide generated intracellularly and superoxide released extracellularly are both relevant to growth control in HeLa cells.     

Generation of superoxide anion and hydrogen peroxide during redox cycling of 5-(4-nitrophenyl)-penta-2,4-dienal by mammalian microsomes and enzymes

5-(4-Nitrophenyl)penta-2,4-dienal (NPPD) stimulated NADPH-supported oxygen consumption by rat liver microsomes in a concentration-dependent manner. The NPPD stimulation of O2 uptake was not inhibited by metyrapone and was decreased in the presence of NADP+ and p-hydroxymercuribenzoate. These observations suggest that the NPPD initial reduction step is mediated by NADPH-cytochrome P-450 reductase and not by cytochrome P-450. Spin-trapping studies using 5,5-dimethyl-1-pyrroline N-oxide (DMPO) revealed the formation of superoxide anion upon incubation of NPPD, NADPH, DMPO and rat liver microsomes. Hydrogen peroxide generation was also detected in these incubations, thus confirming redox cycling of NPPD under aerobic conditions. NPPD stimulated oxygen consumption, superoxide anion formation and hydrogen peroxide generation by rat kidney, testes and brain microsomes. Other enzymes capable of nitroreduction (NADH dehydrogenase, xanthine oxidase, glutathione reductase, and NADP+ ferredoxin oxidoreductase) were also found to stimulate redox cycling of NPPD. The ability of NPPD to induce superoxide anion and hydrogen peroxide formation might play a role in its reported mutagenicity.     

External alternative NADH dehydrogenase of Saccharomyces cerevisiae: a potential source of superoxide

Three rotenone-insensitive NADH dehydrogenases are present in the mitochondria of yeast Saccharomyces cerevisiae, which lack complex I. To elucidate the functions of these enzymes, superoxide production was determined in yeast mitochondria. The low levels of hydrogen peroxide (0.10 to 0.18 nmol/min/mg) produced in mitochondria incubated with succinate, malate, or NADH were stimulated 9-fold by antimycin A. Myxothiazol and stigmatellin blocked completely hydrogen peroxide formation with succinate or malate, indicating that the cytochrome bc(1) complex is the source of superoxide; however, these inhibitors only inhibited 46% hydrogen peroxide formation with NADH as substrate. Diphenyliodonium inhibited hydrogen peroxide formation (with NADH as substrate) by 64%. Superoxide formation, determined by EPR and acetylated cytochrome c reduction in mitochondria was stimulated by antimycin A, and partially inhibited by myxothiazol and stigmatellin. Proteinase K digestion of mitoplasts reduced 95% NADH dehydrogenase activity with a similar inhibition of superoxide production. Mild detergent treatment of the proteinase-treated mitoplasts resulted in an increase in NADH dehydrogenase activity due to the oxidation of exogenous NADH by the internal NADH dehydrogenase; however, little increase in superoxide production was observed. These results suggest that the external NADH dehydrogenase is a potential source of superoxide in S. cerevisiae mitochondria.     

Individual variation in hepatic aldehyde oxidase activity

Aldehyde oxidase (AO) is a molybdo-flavo enzyme expressed predominantly in the liver, lung, and kidney. AO plays a major role in oxidation of aldehydes, as well as oxidation of various N-heterocyclic compounds of pharmacological and toxicological importance including antiviral (famciclovir), antimalarial (quinine), antitumour (methotrexate), and nicotine. The aim of this study was to investigate cytosolic aldehyde oxidase activity in human liver. Cytosolic AO was characterised using both the metabolism of N-[(2-dimethylamino)ethyl] acridine-4-carboxamide (DACA) and benzaldehyde to form DACA-9(10H)-acridone (quantified by HPLC with fluorescence detection) and benzoic acid (quantified spectrophotometrically). Thirteen livers (10 female, 3 male) were examined. The intrinsic clearance (Vmax/Km) of DACA varied 18-fold (0.03-0.50 m/min/mg). Vmax ranged from 0.20-3.10 nmol/ min/mg, and Km ranged from 3.5-14.2 microM. In the same specimens, the intrinsic clearance for benzaldehyde varied 5-fold (0.40-1.8 ml/min/mg). Vmax ranged from 3.60-12.6 nmol/min/mg and Km ranged from 3.6-14.6 microM. Furthermore, there were no differences in AO activity between male and female human livers, nor was there any relationship to age of donor (range 29-73 years), smoking status, or disease status. In conclusion, our results showed that there are variations in AO activity in human liver. These variations in aldehyde oxidase activity might reflect individual variations or they might be due to AO stability during processing and storage.     

Characterization of superoxide-producing sites in isolated brain mitochondria

Mitochondrial respiratory chain complexes I and III have been shown to produce superoxide but the exact contribution and localization of individual sites have remained unclear. We approached this question investigating the effects of oxygen, substrates, inhibitors, and of the NAD+/NADH redox couple on H2O2 and superoxide production of isolated mitochondria from rat and human brain. Although rat brain mitochondria in the presence of glutamate+malate alone do generate only small amounts of H2O2 (0.04 +/- 0.02 nmol H2O2/min/mg), a substantial production is observed after the addition of the complex I inhibitor rotenone (0.68 +/- 0.25 nmol H2O2/min/mg) or in the presence of the respiratory substrate succinate alone (0.80 +/- 0.27 nmol H2O2/min/mg). The maximal rate of H2O2 generation by respiratory chain complex III observed in the presence of antimycin A was considerably lower (0.14 +/- 0.07 nmol H2O2/min/mg). Similar observations were made for mitochondria isolated from human parahippocampal gyrus. This is an indication that most of the superoxide radicals are produced at complex I and that high rates of production of reactive oxygen species are features of respiratory chain-inhibited mitochondria and of reversed electron flow, respectively. We determined the redox potential of the superoxide production site at complex I to be equal to -295 mV. This and the sensitivity to inhibitors suggest that the site of superoxide generation at complex I is most likely the flavine mononucleotide moiety. Because short-term incubation of rat brain mitochondria with H2O2 induced increased H2O2 production at this site we propose that reactive oxygen species can activate a self-accelerating vicious cycle causing mitochondrial damage and neuronal cell death.     

Enzymatic production of hydrogen peroxide and acetaldehyde in a pressure reactor

Alcohol oxidase, an enzyme which exhibits relatively weak substrate specificity among short chain alcohols, forms the corresponding aldehyde and hydrogen peroxide as coproduct. The ability of alcohol oxidase from Pichia pastoris yeast to convert ethanol to acetaldehyde and hydrogen peroxide was examined in an oxygen pressure reactor under conditions, such that oxygen availability was sufficient to permit rapid catalysis. Hydrogen peroxide levels of approximately 1.8/M (6% w/w) were attained in 2-3 h with 2.8 muM enzyme, corresponding to a productivity of approximately 30 g peroxide/g enzyme. Optimal conditions (within equipment limitations) were 900 psi oxygen, 2.6M ethanol, at 4 degrees C. Similar levels of products were reached in the reactor using enzyme immobilized covalently on controlled pore glass and noncovalently on an anion exchange support. Recycle of covalently immobilized enzyme was not possible as a result of enzyme inactivation after a single run. Limited recycle of noncovalently immobilized enzyme was accomplished with substantial decreases in levels of product attainable on each cycle.     

Adriamycin stimulated superoxide formation in submitochondrial particles.

Adriamycin (doxorubicin), an anticancer agent, stimulated the formation of superoxide in submitochondrial particles isolated from bovine heart. Superoxide formation was detected by oxygen uptake, by the cooxidation of epinephrine to adrenochrome and by the reduction of acetylated cytochrome c. These processes were sensitive to superoxide dismutase (SOD). Rotenone-insensitive oxidation of NADH by the mitochondrial respiratory chain in the presence of oxygen caused the formation of approx 4 nmol of superoxide per min/mg of protein. Adriamycin at a concentration of 400 micron stimulated the rate of superoxide formation 6-fold to 25 nmol.min-1.mg-1, but this was not a maximum rate. Approximately 50 micron adriamycin was estimated to be sufficient for obtaining one-half maximal stimulation. Hydrogen peroxide accumulated as a final reaction product. Measurements of the relative catalase activity of blood-free tissues of rabbits and rats indicated that heart contained 2 to 4% of the catalase activity of liver or kidney. An enhanced production of superoxide and hydrogen peroxide and the relatively low catalase content of heart tissue may be factors in the cardiotoxicity induced by adriamycin chemotherapy if a similar reaction occurs in vivo.     

Enzymatic and Non-Enzymatic Oxygenation of Tyrosine

Tyrosinase isolated from cultured human melanoma cells was studied for tyrosine oxygenation activity. l -Tyrosine and d -tyrosine were used as substrates and dopa was measured with HPLC and electrochemical detection as the product of oxygenation. Incubations were performed in the presence or absence of dopamine as co-substrate. Oxygenation of l -tyrosine occurred only in the presence of dopamine as co-substrate. No oxygenation of d -tyrosine was found, and we conclude that human tyrosinase is characterised by exclusive specificity for the l -isomer of tyrosine in its oxygenase function. It has recently been suggested that superoxide anion is a preferential oxygen substrate for human tyrosinase. Incubations were therefore performed with l - and d -tyrosine, human tyrosinase, and xanthine/xanthine oxidase in the system, generating superoxide anion and hydrogen peroxide. Considerable formation of dopa was observed, but the quantity was the same irrespective of whether d -tyrosine or l -tyrosine was used as the substrate. Furthermore, formation of dopa occurred in a xanthine/xanthine oxidase system when bovine serum albumin (BSA) was substituted for tyrosinase. Our results provide no evidence that superoxide anion is an oxygen substrate for human tyrosinase. In the incubate containing xanthine/xanthine oxidase, catalase completely inhibited dopa formation, and superoxide dismutase and mannitol each strongly inhibited dopa formation. The results are compatible with hydroxyl radicals being responsible for the formation of dopa, since such radicals may be secondarily formed in the presence of superoxide anion and hydrogen peroxide.     

Heat shock augments angiotensin II-induced vascular contraction through increased production of reactive oxygen species

A temporal increase in temperature triggers a series of stress responses and alters vascular smooth muscle (VSM) contraction induced by agonist stimulation. Here we examined the role of reactive oxygen species (ROS) in heat shock-dependent augmentation of angiotensin II (AngII)-induced VSM contraction. Endothelium-denuded rat aortic rings were treated with heat shock for 45 min at 42 °C and then subjected to assays for the production of force, ROS, and the expression of ROS-related enzymes. AngII-induced contraction was enhanced in heat shock-treated aorta. AngII-induced production of hydrogen peroxide and superoxide were elevated in response to the heat shock treatment. Pre-treatment with superoxide dismutases (SOD) mimetic and inhibitors for glutathione peroxidase and NADPH oxidase but not for xanthine oxidase eliminated an increase in the AngII-induced contraction in the heat shock-treated aorta. Heat shock increased the expression of p47phox, a cytosolic subunit of NADPH oxidase, but not Cu-Zn-SOD and Mn-SOD. In addition, heat shock increased contraction that was evoked by hydrogen peroxide and pyrogallol. These results suggest that heat shock causes an elevation of ROS as well as a sensitization of ROS signal resulting in an augmentation of VSM contraction in response to agonist.     

Gastropod mollusc aliphatic alcohol oxidase: subcellular localisation and properties

The digestive gland and other tissues of several species of terrestrial gastropod mollusc contain an aliphatic alcohol oxidase activity (EC1.1.3.13). The enzyme is FAD dependent, consumes oxygen and generates hydrogen peroxide and the corresponding aldehyde. Saturated primary alcohols are favoured as substrates with octanol preferred with an apparent K_m of 3–4 μM. The activity is clearly distinguishable from previously reported molluscan aromatic alcohol oxidase (EC1.1.3.7) on the basis of FAD dependence, sensitivity to heat treatment and high salt concentration and with regard to substrate preferences. The aliphatic alcohol oxidase is membrane associated and most likely localised to the endoplasmic reticulum. Extraction of membranes with 1% Igipal solubilises the enzyme in active form. This enzyme is a further example of an oxidase apparently restricted to molluscs.     

Hydrogen peroxide release from human blood platelets

The release of hydrogen peroxide from human blood platelets after stimulation with particulate membrane-perturbing agents has been determined by fluorescence using scopoletin as the detecting agent. Platelet suspensions containing less than 1 polymorphonuclear leukocyte/10⁸ platelets showed a significant release of hydrogen peroxide (6.11 nmol/10⁹ platelets per 20 min, S.D., 0.26, n=9) after addition of zymosan or latex particles, compared to unstimulated platelets. The release of hydrogen peroxide was only observed when the scopoletin was added to the platelet suspensions during the stimulation. Any attempt to determine hydrogen peroxide release in the supernatant at the end of the incubation with zymosan or latex failed. A NADH-dependent production of hydrogen peroxide was observed by measuring the difference of oxygen uptake in the presence and absence of catalase (500 units), which was not inhibited by potassium cyanide (1 mM). By this method the NADH-dependent cyanide-insensitive peroxide production and release was 6.0 nmol/10⁹ platelets per 20 min from resting platelets (S.D., 2, n=6) vs. 15 nmol/10⁹ platelets per 20 min from stimulated platelets (S.D., 2, n=6).