Superoxide reduction by <Emphasis Type="Italic">Archaeoglobus fulgidus</Emphasis> desulfoferrodoxin: comparison with neelaredoxin

Superoxide reductases (SORs) are non-heme iron-containing enzymes that remove superoxide by reducing it to hydrogen peroxide. The active center of SORs consists of a ferrous ion coordinated by four histidines and one cysteine in a square-pyramidal geometry. In the 2Fe-SOR, a distinct family of SORs, there is an additional desulforedoxin-like site that does not appear to be involved in SOR activity. Our previous studies on recombinant Archaeoglobus fulgidus neelaredoxin (1Fe-SOR) have shown that the reaction with superoxide involves the formation of a transient ferric form that, upon protonation, decays to yield an Fe³⁺–OH species, followed by binding of glutamate to the ferric ion via replacement of hydroxide (Rodrigues et al. in Biochemistry 45:9266–9278, 2006). Here, we report the characterization of recombinant desulfoferrodoxin from the same organism, which is a member of the 2Fe-SOR family, and show that the steps involved in the superoxide reduction are similar in both families of SOR. The electron donation to the SOR from its redox partner, rubredoxin, is also presented here. Electronic supplementary material Supplementary material is available in the online version of this article at and is accessible for authorized users.     

The mechanism(s) of superoxide reduction by superoxide reductases in vitro and in vivo

Exposure of obligately anaerobic bacteria and archaea to transiently aerobic or micro-aerobic growth habitats requires that these microorganisms protect against oxidative stress resulting from adventitious dioxygen reduction. Superoxide reductases (SORs), which catalyze reduction of superoxide to hydrogen peroxide, have been identified as one component of a novel oxidative stress protection system in anaerobic bacteria and archaea. SORs contain a unique non-heme [Fe(His)(4)(Cys)] active site. This Commentary addresses the mechanism of superoxide reduction catalyzed by this unique active site in SORs both in vitro and in vivo.     

Superoxide reduction mechanism of Archaeoglobus fulgidus one-iron superoxide reductase

Superoxide reductases (SORs), iron-centered enzymes responsible for reducing superoxide (O2(-)) to hydrogen peroxide, are found in many anaerobic and microaerophilic prokaryotes. The rapid reaction with an exogenous electron donor renders the reductase activity catalytic. Here, we demonstrate using pulse radiolysis that the initial reaction between O2(-) and Archaeoglobus fulgidus neelaredoxin, a one-iron SOR, leads to a short-lived transient that immediately disappears to yield a solvent-bound ferric species in acid-base equilibrium. Through comparison of wild-type neelaredoxin with mutants lacking the ferric ion coordinating glutamate, we demonstrate that the remaining step is related to the final coordination of this ligand to the oxidized metal center and kinetically characterize it for the first time, by pulse radiolysis and stopped-flow kinetics. The way exogenous phosphate perturbs the kinetics of superoxide reduction by neelaredoxin and mutant proteins was also investigated.     

Intermolecular electron transfer in two-iron superoxide reductase: a putative role for the desulforedoxin center as an electron donor to the iron active site

Superoxide reductase (SOR) is a superoxide detoxification system present in some microorganisms. Its active site consists of an unusual mononuclear iron center with an FeN4S1 coordination which catalyzes the one-electron reduction of superoxide to form hydrogen peroxide. Different classes of SORs have been described depending on the presence of an additional rubredoxin-like, desulforedoxin iron center, whose function has remained unknown until now. In this work, we investigated the mechanism of the reduction of the SOR iron active site using the NADPH:flavodoxin oxidoreductase from Escherichia coli, which was previously shown to efficiently transfer electrons to the Desulfoarculus baarsii SOR. When present, the additional rubredoxin-like iron center could function as an electronic relay between cellular reductases and the iron active site for superoxide reduction. This electron transfer was mainly intermolecular, between the rubredoxin-like iron center of one SOR and the iron active site of another SOR. These data provide the first experimental evidence for a possible role of the rubredoxin-like iron center in the superoxide detoxifying activity of SOR.     

Kinetics of the superoxide reductase catalytic cycle

The steady state kinetics of a Desulfovibrio (D.) vulgaris superoxide reductase (SOR) turnover cycle, in which superoxide is catalytically reduced to hydrogen peroxide at a [Fe(His)4(Cys)] active site, are reported. A proximal electron donor, rubredoxin, was used to supply reducing equivalents from NADPH via ferredoxin: NADP+ oxidoreductase, and xanthine/xanthine oxidase was used to provide a calibrated flux of superoxide. SOR turnover in this system was well coupled, i.e. approximately 2O*2 reduced:NADPH oxidized over a 10-fold range of superoxide flux. The reduction of the ferric SOR active site by reduced rubredoxin was independently measured to have a second-order rate constant of approximately 1 x 10(6) m-1 s-1. Analysis of the kinetics showed that: (i) 1 microM SOR can convert a 10 microM/min superoxide flux to a steady state superoxide concentration of 10(-10) m, during which SOR turns over about once every 6 s, (ii) the diffusion-controlled reaction of reduced SOR with superoxide is the slowest process during turnover, and (iii) neither ligation nor deligation of the active site carboxylate of SOR limits the turnover rate. An intracellular SOR concentration on the order of 10 microM is estimated to be the minimum required for lowering superoxide to sublethal levels in aerobically growing SOD knockout mutants of Escherichia coli. SORs from Desulfovibrio gigas and Treponema pallidum showed similar turnover rates when substituted for the D. vulgaris SOR, whereas superoxide dismutases showed no SOR activity in our assay. These results provide quantitative support for previous suggestions that, in times of oxidative stress, SORs efficiently divert intracellular reducing equivalents to superoxide.     

Fe–O versus O–O bond cleavage in reactive iron peroxide intermediates of superoxide reductase

It is generally accepted that the catalytic cycles of superoxide reductases (SORs) and cytochromes P450 involve a ferric hydroperoxo intermediate at a mononuclear iron center with a coordination sphere consisting of four equatorial nitrogen ligands and one axial cysteine thiolate trans to the hydroperoxide. However, although SORs and P450s have similar intermediates, SORs selectively cleave the Fe–O bond and liberate peroxide, whereas P450s cleave the O–O bond to yield a high-valent iron center. This difference has attracted the interest of researchers, and is further explored here. Meta hybrid DFT (M06-2X) results for the reactivity of the putative peroxo/hydroperoxo reaction intermediates in the catalytic cycle of SORs were found to indicate a high-spin preference in all cases. An exploration of the energy profiles for Fe–O and O–O bond cleavage in all spin states in both ferric and ferrous models revealed that Fe–O bond cleavage always occurs more easily than O–O bond cleavage. While O–O bond cleavage appears to be thermodynamically and kinetically unfeasible in ferric hydrogen peroxide complexes, it could occur as a minor (significantly disfavored) side reaction in the interaction of ferrous SOR with hydrogen peroxide.     

Photochemical processes observed during the reaction of superoxide reductase from Desulfoarculus baarsii with superoxide: Re-evaluation of the reaction mechanism

Superoxide reductase SOR is an enzyme involved in superoxide detoxification in some microorganisms. Its active site consists of a non-heme ferrous center in an unusual [Fe(NHis)₄ (SCys)₁] square pyramidal pentacoordination that efficiently reduces superoxide into hydrogen peroxide. In previous works, the reaction mechanism of the SOR from Desulfoarculus baarsii enzyme, studied by pulse radiolysis, was shown to involve the formation of two reaction intermediates T1 and T2. However, the absorption spectrum of T2 was reported with an unusual sharp band at 625 nm, very different from that reported for other SORs. In this work, we show that the sharp band at 625 nm observed by pulse radiolysis reflects the presence of photochemical processes that occurs at the level of the transient species formed during the reaction of SOR with superoxide. These processes do not change the stoichiometry of the global reaction. These data highlight remarkable photochemical properties for these reaction intermediates, not previously suspected for iron-peroxide species formed in the SOR active site. We have reinvestigated the reaction mechanism of the SOR from D. baarsii by pulse radiolysis in the absence of these photochemical processes. The T1 and T2 intermediates now appear to have absorption spectra similar to those reported for the Archaeoglobus fulgidus SOR enzymes. Although for some enzymes of the family only one transient was reported, on the whole, the reaction mechanisms of the different SORs studied so far seem very similar, which is in agreement with the strong sequence and structure homologies of their active sites.     

The effect of exogenous spermidine on superoxide dismutase activity,H<Subscript>2</Subscript>O<Subscript>2</Subscript> and superoxide radical level in barley leaves under water deficit conditions

Barley (Hordeum vulgare) seedlings were treated with spermidine prior to water deficit to determine whether this polyamine is able to affect the activity of superoxide dismutase -SOD (EC 1.15.1.1) responsible for hydrogen peroxide and superoxide radical level. Short-term dehydration (24h) resulted in decrease of the SOD specific activity and a distinct increase in the superoxide anion and hydrogen peroxide contents. Polyamine treatment caused a substantial reduction in the contents of these two stress-raised reactive oxygen species and thereby lowered the oxidative stress in plant cells. Antioxidant system as an important component of the water-stress-protective mechanism can be changed by polyamines, able to moderate the radical scavenging system and to lessen in this way the oxidative stress.     

Modulation of catalase peroxidatic and catalatic activity by nitric oxide

Previously, we found that catalase enhanced the protection afforded by superoxide dismutase to Escherichia coli against the simultaneous generation of superoxide and nitric oxide (Brunelli et al., Arch. Biochem. Biophys. 316:327-334, 1995). Hydrogen peroxide itself was not toxic in this system in the presence or absence of superoxide dismutase. We therefore investigated whether catalase might consume nitric oxide in addition to hydrogen peroxide. Catalase rapidly formed a reversible complex stoichiometrically with nitric oxide with the Soret band shifting from 406 to 426 nm and two new peaks appeared at 540 and at 575 nm, consistent with the formation of a ferrous-nitrosyl complex. Catalase consumed more nitric oxide upon the addition of hydrogen peroxide. Conversely, micromolar concentrations of nitric oxide slowed the catalase-mediated decomposition of hydrogen peroxide. Catalase pretreated with nitric oxide and hydrogen peroxide regained full activity after dialysis. Our results suggest that catalase can slowly consume nitric oxide while nitric oxide modestly inhibits catalase-dependent scavenging of hydrogen peroxide. The protective effects of catalase in combination with superoxide dismutase may result from two actions; reducing peroxynitrite formation by scavenging nitric oxide and by scavenging hydrogen peroxide before it reacts with superoxide dismutase to form additional superoxide.     

Oxycytochrome P-450: its breakdown to superoxide for the formation of hydrogen peroxide.

Four different experimental studies are described which were designed to evaluate the role of oxycytochrome P-450 in the formation of superoxide anions and hydrogen peroxide. The use of lipophilic copper chelates with superoxide dismutase like activity revealed that the primary site of interaction of these agents is related to the inhibition of the flavoprotein. NADPH-cytochrome P-450 reductase. Measurements of the proton assisted nucleophilic displacement of superoxide from oxycytochrome P-450 by high concentrations of sodium azide indicated an increase in the rate of hydrogen peroxide formation concomitant with the inhibition of the N-demethylation of ethylmorphine. Studies on the effect of NADH on the rate of hydrogen peroxide formation during NADPH oxidation by liver microsomes failed to reveal a stimulatory or synergistic effect in a manner analogous to results obtained during the cytochrome P-450 dependent oxidation of substrates such as ethylmorphine. These results suggest that hydrogen peroxide formation may not require the reduction of oxycytochrome P-450 to peroxycytochrome P-450. Measurements of the reduction of succinylated cytochrome c using purified cytochrome P-450 and the flavoprotein, NADPH-cytochrome P-450 reductase, directly demonstrate the formation of superoxide anions. It is concluded that oxycytochrome P-450 may decompose to generate hydrogen peroxide.     

Fe(3+)-eta(2)-peroxo species in superoxide reductase from Treponema pallidum. Comparison with Desulfoarculus baarsii

Superoxide reductases (SORs) are superoxide (O2-)-detoxifying enzymes that catalyse the reduction of O2- into hydrogen peroxide. Three different classes of SOR have been reported on the basis of the presence or not of an additional N-terminal domain. They all share a similar active site, with an unusual non-heme Fe atom coordinated by four equatorial histidines and one axial cysteine residues. Crucial catalytic reaction intermediates of SOR are purported to be Fe(3+)-(hydro)peroxo species. Using resonance Raman spectroscopy, we compared the vibrational properties of the Fe3+ active site of two different classes of SOR, from Desulfoarculus baarsii and Treponema pallidum, along with their ferrocyanide and their peroxo complexes. In both species, rapid treatment with H2O2 results in the stabilization of a side-on high spin Fe(3+)-(eta(2)-OO) peroxo species. Comparison of these two peroxo species reveals significant differences in vibrational frequencies and bond strengths of the Fe-O2 (weaker) and O-O (stronger) bonds for the T. pallidum enzyme. Thus, the two peroxo adducts in these two SORs have different stabilities which are also seen to be correlated with differences in the Fe-S coordination strengths as gauged by the Fe-S vibrational frequencies. This was interpreted from structural variations in the two active sites, resulting in differences in the electron donating properties of the trans cysteine ligand. Our results suggest that the structural differences observed in the active site of different classes of SORs should be a determining factor for the rate of release of the iron-peroxo intermediate during enzymatic turnover.     

Product of extracellular-superoxide dismutase catalysis

Extracellular-superoxide dismutase is a tetrameric enzyme containing four copper atoms. It has previously been shown to catalyse the decay of the superoxide radical, but the resulting product was not determined. In a xanthine oxidase-xanthine system in which about 30% of the electron flux resulted in superoxide radical formation, accumulation of hydrogen peroxide was determined. Catalysis of superoxide radical decay by extracellular-superoxide dismutase was found to result in hydrogen peroxide formation. The catalysed reaction is thus identical to those of previously investigated superoxide dismutases. Human manganese superoxide dismutase was also found to dismute the superoxide radical to hydrogen peroxide and water.     

Photo-irradiated titanium dioxide catalyzes site specific DNA damage via generation of hydrogen peroxide

Titanium dioxide (TiO₂) is a potential photosensitizer for photodynamic therapy. In this study, the mechanism of DNA damage catalyzed by photo-irradiated TiO₂ was examined using [₃₂P]-5'-end-labeled DNA fragments obtained from human genes. Photo-irradiated TiO₂ (anatase and rutile) caused DNA cleavage frequently at the guanine residue in the presence of Cu(II) after E. coli formamidopyrimidine-DNA glycosylase treatment, and the thymine residue was also cleaved after piperidine treatment. Catalase, SOD and bathocuproine, a chelator of Cu(I), inhibited the DNA damage, suggesting the involvement of hydrogen peroxide, superoxide and Cu(I). The photocatalytic generation of Cu(I) from Cu(II) was decreased by the addition of SOD. These findings suggest that the inhibitory effect of SOD on DNA damage is due to the inhibition of the reduction of Cu(II) by superoxide. We also measured the formation of 8-oxo-7,8-dihydro-2' -deoxyguanosine, an indicator of oxidative DNA damage, and showed that anatase is more active than rutile. On the other hand, high concentration of anatase caused DNA damage in the absence of Cu(II). Typical free hydroxyl radical scavengers, such as ethanol, mannnitol, sodium formate and DMSO, inhibited the copper-independent DNA photodamage by anatase. In conclusion, photo-irradiated TiO₂ particles catalyze the copper-mediated site-specific DNA damage via the formation of hydrogen peroxide rather than that of a free hydroxyl radical. This DNA-damaging mechanism may participate in the phototoxicity of TiO₂.     

Role of iron in adriamycin biochemistry

Adriamycin forms a chelate with Fe(III) that exhibits complex redox chemistry. The drug ligand is able to directly reduce the bound Fe(III) with the concomitant production of a one-electron oxidized drug radical. This Fe(II) can reduce oxygen to hydrogen peroxide and cleave the peroxide to yield the hydroxyl radical. In addition, the drug X Fe complex can catalyze the transfer of electrons from reduced glutathione to molecular oxygen to yield superoxide, hydrogen peroxide, and hydroxyl radicals. The adriamycin X Fe complex binds to DNA to form a ternary drug X Fe X DNA complex, which is also able to catalyze the thiol-dependent reduction of oxygen and the formation of hydroxyl radical from hydrogen peroxide. As a consequence of this chemistry, the adriamycin X Fe complex can cleave DNA on the addition of glutathione or hydrogen peroxide. Although less well defined, the adriamycin X Fe complex can bind to cell membranes and cause oxidative destruction of these membranes in the presence of thiols or hydrogen peroxide.     

Flavin-linked Erv-family sulfhydryl oxidases release superoxide anion during catalytic turnover

Typically, simple flavoprotein oxidases couple the oxidation of their substrates with the formation of hydrogen peroxide without release of significant levels of the superoxide ion. However, two evolutionarily related single-domain sulfhydryl oxidases (Erv2p; a yeast endoplasmic reticulum resident protein and augmenter of liver regeneration, ALR, an enzyme predominantly found in the mitochondrial intermembrane) release up to ~30% of the oxygen they reduce as the superoxide ion. Both enzymes oxidize dithiol substrates via a redox-active disulfide adjacent to the flavin cofactor within the helix-rich Erv domain. Subsequent reduction of the flavin is followed by transfer of reducing equivalents to molecular oxygen. Superoxide release was initially detected using tris(3-hydroxypropyl)phosphine (THP) as an alternative reducing substrate to dithiothreitol (DTT). THP, and other phosphines, showed anomalously high turnover numbers with Erv2p and ALR in the oxygen electrode, but oxygen consumption was drastically suppressed upon the addition of superoxide dismutase. The superoxide ion initiates a radical chain reaction promoting the aerobic oxidation of phosphines with the formation of hydrogen peroxide. Use of a known flux of superoxide generated by the xanthine/xanthine oxidase system showed that one superoxide ion stimulates the reduction of 27 and 4.5 molecules of oxygen using THP and tris(2-carboxyethyl)phosphine (TCEP), respectively. This superoxide-dependent amplification of oxygen consumption by phosphines provides a new kinetic method for the detection of superoxide. Superoxide release was also observed by a standard chemiluminescence method using a luciferin analogue (MCLA) when 2 mM DTT was employed as a substrate of Erv2p and ALR. The percentage of superoxide released from Erv2p increased to ~65% when monomeric mutants of the normally homodimeric enzyme were used. In contrast, monomeric multidomain quiescin sulfhydryl oxidase enzymes that also contain an Erv FAD-binding fold release only 1-5% of their total reduced oxygen species as the superoxide ion. Aspects of the mechanism and possible physiological significance of superoxide release from these Erv-domain flavoproteins are discussed.     

Reductase and oxidase activity of rat liver cytochrome P450 with 2,3,5,6-tetramethylbenzoquinone as substrate.

The main objective of the present study was to investigate the proposed role of cytochrome P450 in the reductive metabolism of quinones as well as in the formation of reduced oxygen species in liver microsomes from phenobarbital (PB-microsomes) and beta-naphthoflavone (beta NF-microsomes) pretreated rats. In the present study, 2,3,5,6-tetramethylbenzoquinone (TMQ) was chosen as a model quinone. Anaerobic one-electron reduction of TMQ by PB-microsomes showed relatively strong electron spin resonance (ESR) signals of the oxygen-centered semiquinone free radical (TMSQ), whereas these signals were hardly detectable with beta NF-microsomes. Under aerobic conditions TMSQ formation was diminished and concomitant reduction of molecular oxygen occurred in PB-microsomes. Interestingly, TMQ-induced superoxide anion radicals, measured by ESR (using the spin trap 5,5'-dimethyl-1-pyrroline-N-oxide), and hydrogen peroxide generation was found to occur with beta NF-microsomes as well. Furthermore, SK&F 525-A (a type I ligand inhibitor of cytochrome P450) inhibited TMQ-induced hydrogen peroxide formation in both PB- and beta NF-microsomes. However, metyrapone and imidazole (type II ligand inhibitors of cytochrome P450) inhibited molecular oxygen reduction in beta NF-microsomes and not in PB-microsomes. The present study indicates that cytochrome P450-mediated one-electron reduction of TMQ to TMSQ and subsequent redox cycling of TMSQ with molecular oxygen constitutes the major source for superoxide anion radical and hydrogen peroxide generation in PB-microsomes (i.e. from the reductase activity of cytochrome P450). However, most of the superoxide anion radical formed upon aerobic incubation of TMQ with beta NF-microsomes originates directly from the dioxyanion-ferri-cytochrome P450 complex (i.e. from the oxidase activity of cytochrome P450). In conclusion, both the one-electron reduction of TMQ and molecular oxygen were found to be cytochrome P450 dependent. Apparently, both the reductase and oxidase activities of cytochrome P450 may be involved in the reductive cytotoxicity of chemotherapeutic agents containing the quinoid moiety.     

Kinetic evidence of horseradish peroxidase oxidation by compound I.

The kinetics of compound II formation, obtained upon mixing a highly purified horseradish peroxidase and hydrogen peroxide, was spectrophotometrically studied at three wavelengths in the absence of an added reducing agent. Our experiments confirm George's finding that more than one mole of compound II is formed per mole of hydrogen peroxide added. The new mechanism that we propose, contrary to the mechanism of George, is only valid when compound II is obtained in the absence of an added donor. Moreover, it is not inconsistent with the classical Chance mechanism of oxidation of an added donor by the system peroxidase -- hydrogen peroxide. According to this new mechanism, in the absence of an added donor, compound II formation involved two pathways. The first pathway is the monomolecular reduction of compound I by the endogenous donor, and the second pathway is the formation of two moles of compound II through the oxidoreduction reaction between one mole of peroxidase and one mole of compound I.     

Involvement of hydrogen peroxide in histamine-induced modulation of WM35 human malignant melanoma cell proliferation

Histamine is a recognized growth factor in melanoma, and exogenous histamine produces a dual effect on proliferation. We have previously reported that histamine at micromolar concentrations reduces the proliferation of melanoma cell lines. To investigate the mechanism by which histamine inhibits proliferation of WM35 human melanoma cells, we have studied the involvement of histamine in reactive oxygen species production and antioxidant enzyme regulation in these cells. Results indicate that histamine treatment (10 μM) significantly increased hydrogen peroxide levels, whereas it slightly decreased superoxide levels associated with an enhancement of superoxide dismutase and a reduction in catalase activity. Additionally, catalase treatment reversed the inhibitory effect of histamine on proliferation, and various treatments that reduce hydrogen peroxide formation increased proliferation of these cells. Furthermore, we demonstrate that the inhibition of proliferation produced by histamine was mediated at least in part by an induction of cell senescence. We conclude that hydrogen peroxide is involved in histamine-mediated modulation of proliferation in malignant melanoma cells.     

Characterization of superoxide production from aldehyde oxidase: an important source of oxidants in biological tissues

Aldehyde oxidase, a molybdoflavoenzyme that plays an important role in aldehyde biotransformation, requires oxygen as substrate and produces reduced oxygen species. However, little information is available regarding its importance in cellular redox stress. Therefore, studies were undertaken to characterize its superoxide and hydrogen peroxide production. Aldehyde oxidase was purified to >98% purity and exhibited a single band at approximately 290 kDa on native polyacrylamide gradient gel electrophoresis. Superoxide generation was measured and quantitated by cytochrome c reduction and EPR spin trapping with p-dimethyl aminocinnamaldehyde as reducing substrate. Prominent superoxide generation was observed with an initial rate of 295 nmol min(-1) mg(-1). Electrochemical measurements of oxygen consumption and hydrogen peroxide formation yielded values of 650 and 355 nmol min(-1) mg(-1). In view of the ubiquitous distribution of aldehydes in tissues, aldehyde oxidase can be an important basal source of superoxide that would be enhanced in disease settings where cellular aldehyde levels are increased.     

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.