The Modulation of Oxygen Radical Production by Nitric Oxide in Mitochondria*

Biological systems that produce or are exposed to nitric oxide (NO⋅) exhibit changes in the rate of oxygen free radical production. Considering that mitochondria are the main intracellular source of oxygen radicals, and based on the recently documented production of NO⋅ by intact mitochondria, we investigated whether NO⋅, produced by the mitochondrial nitric-oxide synthase, could affect the generation of oxygen radicals. Toward this end, changes in H2O2production by rat liver mitochondria were monitored at different rates of endogenous NO⋅ production. The observed changes in H2O2 production indicated that NO⋅affected the rate of oxygen radical production by modulating the rate of O2 consumption at the cytochrome oxidase level. This mechanism was supported by these three experimental proofs: 1) the reciprocal correlation between H2O2production and respiratory rates under different conditions of NO⋅ production; 2) the pattern of oxidized/reduced carriers in the presence of NO⋅, which pointed to cytochrome oxidase as the crossover point; and 3) the reversibility of these effects, evidenced in the presence of oxymyoglobin, which excluded a significant role for other NO⋅-derived species such as peroxynitrite. Other sources of H2O2 investigated, such as the aerobic formation of nitrosoglutathione and the GSH-mediated decay of nitrosoglutathione, were found quantitatively negligible compared with the total rate of H2O2 production.

Biological systems that produce or are exposed to nitric oxide (NO ⅐ ) exhibit changes in the rate of oxygen free radical production. Considering that mitochondria are the main intracellular source of oxygen radicals, and based on the recently documented production of NO ⅐ by intact mitochondria, we investigated whether NO ⅐ , produced by the mitochondrial nitric-oxide synthase, could affect the generation of oxygen radicals. Toward this end, changes in H 2 O 2 production by rat liver mitochondria were monitored at different rates of endogenous NO ⅐ production. The observed changes in H 2 O 2 production indicated that NO ⅐ affected the rate of oxygen radical production by modulating the rate of O 2 consumption at the cytochrome oxidase level. This mechanism was supported by these three experimental proofs: 1) the reciprocal correlation between H 2 O 2 production and respiratory rates under different conditions of NO ⅐ production; 2) the pattern of oxidized/reduced carriers in the presence of NO ⅐ , which pointed to cytochrome oxidase as the crossover point; and 3) the reversibility of these effects, evidenced in the presence of oxymyoglobin, which excluded a significant role for other NO ⅐ -derived species such as peroxynitrite. Other sources of H 2 O 2 investigated, such as the aerobic formation of nitrosoglutathione and the GSH-mediated decay of nitrosoglutathione, were found quantitatively negligible compared with the total rate of H 2 O 2 production.
Biological systems of diverse complexity, when exposed to NO ⅐ 1 or stimulated to produce NO ⅐ , usually present changes in the rate of oxygen free radical production (Ref. 1 and references therein). Decreases in the rate of oxygen free radical production are often attributed to the diffusion-controlled reaction between O 2 . and NO ⅐ , which yields peroxynitrite. Increases in reactive oxygen species (ROS) production are usually associated with the damage and/or inactivation of mitochondrial components by peroxynitrite or peroxynitrite-like species.
Considering that under physiological conditions mitochondria constitute the main intracellular source of oxygen free radicals (2,3) and that these organelles can produce NO ⅐ (4 -8), it becomes of interest to examine whether the production of mitochondrial ROS is affected by endogenous NO ⅐ . The accurate documentation of the rates of production of nitrogen and oxygen radicals is an essential step to understanding the role of these interactions in different, more complex pathophysiological situations and underlines the relevance of studying the mechanisms behind these processes in relatively simpler models. Furthermore, transient changes in ROS production have become an area of intense research given the growing interest in the role of ROS as mediators of signal transduction pathways, organ preconditioning (9), and apoptotic processes (10).
In this study, we examined the rates of O 2 free radical production by intact mitochondria under various conditions of endogenous NO ⅐ production; different mechanisms underlying the NO ⅐ -O 2 . interactions were quantitatively investigated.
Herein, we demonstrated that mitochondria could generate O 2 . and NO ⅐ ; however, unlike other biological systems, no irreversible damage to mitochondrial components was observed. The enhanced ROS production detected under NO ⅐ -producing conditions was attributed to an increase in the steady-state level of reduced components of the respiratory chain, whereas a small contribution was attributed to the aerobic decay of S-nitrosoglutathione (GSNO) in the presence of GSH. Finally, the implications of these results were discussed in terms of the ability of NO ⅐ to modulate the rates of oxygen free radical production and/or consumption and the pathophysiological consequences associated with this process.

Biological Materials
Liver mitochondria were isolated from adult Wistar rats by differential centrifugation and purified through Percoll centrifugation (5). This procedure yielded intact mitochondria minimally contaminated with other subcellular compartments (5). Essentially, male rats (body weight of 180 -220 g) were anesthetized using a CO 2 chamber. The livers were quickly removed; washed with 0.22 M mannitol, 70 mM sucrose, 0.5 mM EGTA, 2 mM Hepes, and 0.1% defatted bovine serum albumin, pH 7.4 (Buffer A); and homogenized in a 10:1 buffer to liver w/w ratio. Large cell debris and nuclei were pelleted by centrifuging at 600 ϫ g for 5 min in a Sorvall SS34 rotor. This supernatant was filtered through two layers of cheesecloth to remove fat. Mitochondria were pelleted by centrifuging the supernatant for 10 min at 10,300 ϫ g in the same rotor. After suspending the pellet in 5 ml of the previous buffer, the mitochon-* This work was supported by Grant MCB 9724060 from the National Science Foundation and by a grant from the United Mitochondrial Disease Foundation. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Biochemical Analyses
Oxygen Consumption-The oxygen uptake of mitochondria was measured using a Clark-type O 2 electrode from Hansatech (King's Lynn, UK) at 30°C (5, 7). Intact, purified mitochondria (0.5-2 mg of protein/ml) were maintained in 1 ml of reaction medium in the presence of 10 mM succinate as substrate and, where indicated, different amounts of L-Arg or N G -monomethyl-L-arginine (NMMA).
NO ⅐ Detection-The production of NO ⅐ by mitochondria (1-2 mg/ml) was evaluated as the change in the absorbance at 581-592 nm at 22°C using a dual wavelength, double beam SLM-Aminco DW-2C UV-visible spectrophotometer. The sample cuvette contained 2 ml of reaction medium, 0.1-0.3 mM oxymyoglobin, 10 mM succinate, and 0.1 mM Larginine under constant stirring. When NMMA was used in the incubation, the protein concentration was 3-5 mg/ml. Ten M catalase and superoxide dismutase were also present in the reaction medium to avoid unspecific side reactions (namely, the reaction of NO ⅐ with the superoxide anion or of oxymyoglobin with peroxynitrite/hydrogen peroxide); however, no significant changes were found with or without these enzymes. The rates of NO ⅐ production were taken during the first 4 -5 min when the ratios of myoglobin:catalase (in heme):NO ⅐ were 28:6:1, thereby avoiding a significant, if any, inhibition of catalase by NO ⅐ . The L-citrulline production was monitored by a modified colorimetric technique from Prescott and Jones (11).
Hydrogen Peroxide Production-The rate of H 2 O 2 production by intact mitochondria (0.12-0.15 mg of protein/ml) was measured in reaction medium with 5 M horseradish peroxidase and 1 mM scopoletin (12). This high concentration of scopoletin was required to compete with GSH for the horseradish peroxidase-H 2 O 2 complex (data not shown). The recovery of H 2 O 2 was 80 -90% assayed with glucose/glucose oxidase.
Detection of GSNO-Sample preparation proceeded as follows. Aliquots of the reactions were withdrawn at different time points. The aliquots were treated to avoid thiol oxidation or thiol-disulfide exchange, to terminate any metabolic process, and to prevent any artifactual GSNO production. To satisfy these criteria, the samples were kept on ice protected from light during the duration of the procedure and promptly precipitated with perchloric acid in the presence of a metal chelator, desferrioxamine mesylate. The samples (0.8 ml) were treated with 20 l of 10 mM desferrioxamine mesylate and 36 l of 70% perchloric acid, kept on ice for 10 min, and centrifuged at 14,000 rpm for 10 min at 4°C. The supernatants were treated with 1 volume of 0.15 M K 2 HPO 4 , 10 mM sodium citrate, and 5 mM EDTA (phosphate-citrate-EDTA). The samples were kept on ice for 30 min and centrifuged at 14,000 rpm for 15 min at 4°C. The supernatants were filtered through a 2-m filter and kept at Ϫ80°C until HPLC analysis, usually performed during the same day. The samples were analyzed by HPLC using a modified method from Stamler and Feelisch (13). The separation and detection of GSNO were performed by using a C18 Aqua HPLC column (column specifications: length and inner diameter, 250 mm ϫ 4.6; particle size, 5 m) (Phenomenex™, Torrance, CA) connected with a SecurityGuard™ column from the same manufacturer (4 ϫ 3 mm C18 cartridge). The different components were eluted with an isocratic gradient of 99% 0.1 M monochloroacetic acid, 0.125 mM EDTA, 1.5 mM sodium octyl sulfate, pH 2.8, and 1% acetonitrile. The mobile phase was purged with helium, and the flow rate was set at 0.8 ml/min. The chromatography equipment consisted of a Class-VP Shimadzu HPLC system equipped with two high pressure LC10ADvp pumps and an automatic injector Sil-10ADvp. The chromatograms were analyzed using the software provided with the system (version 5.03). The peaks were monitored by a photodiode array detector (215 and 336 nm) and an electrochemical detector (L-ECD-6A from Shimadzu set at Ϫ300 mV) connected in series. The identification of the derivatives was performed by comparison with synthetic GSNO. This latter compound was prepared according to Hart (14) (15) with the modifications introduced by Yonetani (16), using bovine serum albumin as the standard.
Statistics-Data were expressed as mean Ϯ S.E. and were evaluated by one-way analysis of variance (ANOVA).

Endogenous Nitric Oxide Modulates Hydrogen Peroxide Production by
Mitochondria-The addition of increasing concentrations of L-Arg to intact, coupled mitochondria in State 4 (using either glutamate-malate or succinate) stimulated the production of NO ⅐ (Fig. 1A). The initial velocity plotted versus L-Arg concentration represented a right rectangular hyperbola; the V max and K m values were 1.5 nmol of NO ⅐ ϫ (min ⅐ mg protein) Ϫ1 and 5 M, respectively. Comparable results were obtained by measuring L-citrulline, cosynthesized with NO ⅐ by NOS (V max ϭ 1.2 Ϯ 0.2 nmol of L-citrulline/min ϫ mg of protein and K m ϭ 8 M). The rate of NO ⅐ production was negligible at L-Arg concentrations of 50 -100 ϫ K m . This effect was accompanied by an increase in the rate of oxygen consumption (3-5 times), found to be resistant to ADP stimulation, and comparable with the increase obtained in the presence of carbonyl cyanide p-trifluoromethoxyphenylhydrazone. These experimental observations indicated that the lack of NO ⅐ production at high concentrations of L-Arg was attributed to the uncoupling of mitochondria by this amino acid.
The rates of NO ⅐ production in mitochondria, maintained in State 4, were decreased significantly when NMMA was included in the reaction mixture (Fig. 1A). The addition of this compound did not affect the V max but increased the apparent K m values (15 and 30 M, respectively). These results indicated that this compound is acting as a competitive inhibitor of mitochondrial NOS, as has been shown by other reports (17)(18)(19). The K I was calculated from the slope of a plot of K m(app) versus [NMMA] (Fig. 1B), resulting in a value (15 M) similar to that reported for mouseinducible NOS (13 M) (18). Even at high concentrations of NMMA (0.5 mM), a slow rate of NO ⅐ was detectable (0.02 nmol of NO ⅐ ϫ (min ⅐ mg protein) Ϫ1 ) not accompanied by L-citrulline production. This slow rate has been attributed to the nonenzymatic release of NO ⅐ from GSNO in the presence of reduced glutathione and superoxide dismutase. 2 To examine whether endogenous NO ⅐ affected the mitochondrial production of oxygen free radicals, the production of H 2 O 2 The addition of L-Arg to intact, coupled substrate-supplemented mitochondria increased the rate of H 2 O 2 production in a dose-response manner ( Fig. 2A). The concentrations of L-Arg required for the half-maximal effect were 29.6 M (glutamatemalate) and 34 M (succinate). At saturating concentrations of L-Arg, the rate of H 2 O 2 production increased by 42% (glutamatemalate) and 58% (succinate), whereas the respiratory rates decreased 33% (glutamate-malate) and 27% (succinate) (Fig. 2B). The concentrations of L-Arg required for the 50% inhibition of the respiratory rates were 24.7 M (glutamate-malate) and 31.5 M (succinate). The effects observed on the respiratory rates and H 2 O 2 production were completely reversed by adding 0.2-1 mM oxymyoglobin to the reaction mixture (not shown). These latter experiments might wrongly indicate that the rate of H 2 O 2 production was decreased by the reaction of H 2 O 2 with oxymyoglobin via its pseudocatalatic/peroxidatic activity, without the participation of NO ⅐ . This possibility seemed unlikely based on the following evidence. First, the rate constants for the reaction of oxymyoglobin or oxyhemoglobin with H 2 O 2 in their peroxidatic and catalatic activities are 4 -5 orders of magnitude smaller than those corresponding to the reaction of oxymyoglobin with NO ⅐ (21-23), indicating that this molecule will preferentially react with oxymyoglobin rather than with H 2 O 2 . Second, the addition of NMMA, which blocks the production of NO ⅐ by mitochondrial NOS (independently of the presence of oxymyoglobin) decreased the rate of H 2 O 2 production, excluding the role of oxymyoglobin in the modulation of H 2 O 2 production by NO ⅐ -producing mitochondria (see below).
The addition of NMMA to L-Arg-and succinate-supplemented mitochondria decreased the rate of H 2 O 2 production and increased the respiratory rates in a dose-response manner (Fig. 3, A and B, respectively). The concentrations of NMMA required for the half-maximal rate of H 2 O 2 production were 13.9 (glutamate-malate) and 13.8 M (succinate); at saturating concentrations of NMMA, the rate of H 2 O 2 production was decreased 42% (malate-glutamate) and 37% (succinate). The concentrations of NMMA required for the half-maximal respiratory rates were 15.7 (glutamate-malate) and 10.2 M (succinate); at saturating concentrations of NMMA, the respiratory rates were increased by 85% (malate-glutamate) and 39% (succinate). The effects observed on the respiratory rates and H 2 O 2 production mediated by NMMA were completely reversed by adding increasing amounts of L-Arg (not shown).
The reciprocal association between ROS and NO ⅐ production with the O 2 uptake (Figs. 1-3, Table I) and the similar concentrations of NMMA and L-Arg required for the half-maximal effects for the NO ⅐ production, respiratory rates, and H 2 O 2 production (Figs. 1-3) indicated that these events were biochemically linked, acting at a common site. The involvement of NO ⅐ in these effects was supported by the reversibility of the effects observed with NMMA (upon the addition of L-Arg) and L-Arg (upon the addition of oxymyoglobin). These results also precluded a significant role for peroxynitrite because this species causes irreversible damage of respiratory chain components (24). Furthermore, the fact that oxymyoglobin reversed the modulation of the O 2 uptake and ROS production not only excluded ONOO Ϫ but also the involvement of other NO ⅐ -derived species, such as nitrate, also produced in the presence of oxymyoglobin (23).
Although the association between respiratory rates, hydrogen peroxide production, and NO ⅐ generation was apparent from these results, no quantitative correlation was demonstrated. Moreover, given the fact that mitochondria are endowed with a pool of GSNO 2 and that its synthesis (25) and decay may produce oxygen radicals (26) at rates that may be significant, we sought to explore these processes using quantitative strategies.
Role of S-Nitrosoglutathione Metabolism in the Hydrogen Peroxide Generation by Mitochondria-The major mitochondrial thiol is constituted by glutathione (about 10 mM) (27). Glutathione is synthesized in the cytosol, imported to the mitochondria, and involved in a protective role against oxidative stress (28). Thiols react with NO ⅐ and/or its metabolites, resulting in the formation of stable NO ⅐ donors or nitrosothiols (29). Recent reports suggest that the formation of GSNO in vivo may include the production of O 2 . via the oxidation of the intermediate GSNO . (25) (Equations 1 and 2).
If this mechanism had been active in mitochondria, assuming oxygen is the only suitable electron acceptor for Equation 2, increases in NO ⅐ production would have been followed by increases in the rates of O 2 . and GSNO production. Studies were undertaken to explore the occurrence of this mechanism in mitochondria and to explore whether the increase in oxygen radical production observed in NO ⅐ -producing mitochondria could be associated with the formation of GSNO. In addition, the decay of GSNO, which includes a complex chemistry of homolytic and heterolytic mechanisms (26,30), could also result in the production of hydrogen peroxide. Toward this end, the rates of the production and decay of GSNO were evaluated in terms of their potential to contribute to the rate of hydrogen peroxide produced by mitochondria.The concentration of GSNO from intact mitochondria (analyzed and quantified by HPLC as The O 2 uptake was measured using 1 mg of protein/ml suspended in reaction medium as described under "Experimental Procedures." The rate of H 2 O 2 production was measured using 0.12 mg of protein/ml in reaction medium using the horseradish peroxidase-scopoletin method. Other experimental conditions were described under "Experimental Procedures." described under "Experimental Procedures") was 0.37 Ϯ 0.03 nmol/mg of protein, and its rate of production (under optimal conditions for NO ⅐ production) was 10 -20 pmol of GSNO (min mg ⅐ protein) Ϫ1 . According to the mechanism outlined before, if this rate were equivalent to that of O 2 . production (see Equation   2), it would have represented a small fraction (3-5%) of the rate of hydrogen peroxide obtained with NO ⅐ -producing mitochondria (Table I), constituting a minor contribution to the total rate of H 2 O 2 production. The decay of GSNO may result in H 2 O 2 production if the pathway indicated in Fig.4 is operative. To evaluate the contribution to the rate of H 2 O 2 by GSNO decay, the rate of H 2 O 2 generation was evaluated in a control (mitochondria treated with NMMA for 30 min) and in GSNO-depleted mitochondria (mitochondria incubated with NMMA at 25°C for 3-4 h until the concentration of GSNO was not detectable by HPLC). No significant difference was found in the rates of H 2 O 2 production between GSNO-depleted and control mitochondria (98 Ϯ 30 and 137 Ϯ 30 M/min, respectively; mean Ϯ S.E.; nϭ3), indicating that the decay of GSNO did not contribute significantly to the H 2 O 2 generation by mitochondria; however, it did not preclude in vivo situations in which higher concentrations of GSNO may be achieved, and its decay may have a substantial contribution to the rate of H 2 O 2 production.
Interaction of Nitric Oxide and Respiratory Chain Carriers-We investigated whether the modulation of ROS by NO ⅐ -producing mitochondria was the result of the interaction between NO ⅐ and certain electron carriers of the respiratory chain. This hypothesis is based on the high affinity of NO ⅐ for metallo-containing proteins (31,32) and on the fact that certain inhibitors of the respiratory chain (those that increase the steady-state reduction level of electron carriers at the NADH dehydrogenase and ubiquinol-cytochrome c reductase segments) enhance the production of ROS.
Toward this end, the steady-state reduction of the carriers of the respiratory chain was evaluated in NO ⅐ -producing mitochondria by visible spectroscopy (Fig. 5). Under these conditions, 77% of cytochromes b and c was reduced, whereas 50% of cytochrome oxidase was reduced. The steady-state reduction of cytochrome oxidase (in percentage) was found within the experimental range of the inhibition of the oxygen consumption under identical conditions. This result agreed with those pre-viously obtained in which changes in the respiratory rates correlated with those in cytochrome oxidase activity under identical conditions of NO ⅐ production (7).
Experimental evidence for the association between the increased reduced state of the respiratory chain carriers and H 2    Rat liver mitochondria were isolated as described under "Experimental Procedures." The respiratory control ratio and phosphate to oxygen ratio number of these preparations were 4.7 and 2.1, respectively, assayed in the presence of 10 mM succinate and 0.45 mM ADP in reaction buffer (0.225 M sucrose, 10 mM potassium phosphate, 5 mM MgCl 2 , 20 mM Hepes, pH 7.4). The absorption spectra were recorded using an SLM-Aminco DW-2C UV-visible spectrophotometer as described previously (33). The difference spectra of samples, which contained 1 mg of mitochondrial protein in reaction buffer, were recorded under the following conditions: 20 mM succinate (State 4; striped bar), succinate plus 0.5 mM ADP (State 3; white bar), succinate and 50 nM antimycin (State 4 plus antimycin; gray bar); and succinate and 0.15 mM L-arginine (State 4 plus L-arginine; black bar). The concentrations of each carrier were calculated using the following extinction coefficients: cytochromes c-c 1 , ⑀ 550 -540 ϭ 19 mM Ϫ1 cm Ϫ1 ; cytochromes b K-T , ⑀ 564 -575 ϭ 20 mM Ϫ1 cm Ϫ1 ; cytochromes aa 3 , ⑀ 605-630 ϭ 16 mM Ϫ1 cm Ϫ1 . production in NO ⅐ -producing mitochondria was provided by evaluating the steady-state reduction of the respiratory chain carriers and the rate of ROS production under a variety of conditions, including those that entailed changes in NO ⅐ production (Fig. 6). A significant correlation was found between the reduction level of cytochrome b (Fig. 6) or other components of the respiratory chain (not shown) and the production of H 2 O 2 (Fig. 6) (r ϭ 0.958), indicating that an increased reduction of these carriers under a variety of conditions (including those in which NO ⅐ is produced) leads to an increased rate of ROS production.
Based on our results, it seemed that the modulation of ROS production was accomplished through the reversible inhibition of the mitochondrial respiratory rate by NO ⅐ at the cytochrome oxidase level. The transient decrease (or inhibition, depending on the concentration of NO ⅐ and oxygen) in the respiratory rate leads to an increased steady-state reduction of the respiratory chain carriers, which in turn by reacting with oxygen resulted in the production of ROS. It is likely that cytochrome oxidase represents the crossover point in NO ⅐ -producing mitochondria (Fig. 5), excluding a significant interaction of NO ⅐ with other available components of the respiratory chain (cytochromes c or b) under these experimental conditions. However, if cytochrome oxidase were the crossover point in this process, it would have been fully reduced. This apparent discrepancy can be bridged by considering that the interaction of NO ⅐ with cytochrome oxidase is complex, constituted by the reversible binding of NO ⅐ to the heme a 3 of cytochrome oxidase and by the reduction of NO ⅐ to N 2 O by this oxidase (7).
Considering the results presented in this study, it could be wrongly surmised that the effect of NO ⅐ on the respiratory chain is not different from that of other xenobiotics used to enhance ROS production in mitochondria (e.g. NADH-supplemented mitochondria plus rotenone or FADH 2 -supplemented mitochondria plus antimycin). Nitric oxide has two distinctive properties from the aforementioned inhibitors; it is endogenously produced (4 -8), and it has a transient effect (5,7). This latter property, based on its cytochrome c oxidase-mediated catabolism, allows a slow but continuous flow of electrons through the chain, even when a complete suppression of the O 2 consumption has been accomplished. This type of modulation of the respiratory chain has the advantage of avoiding a complete reduction of the components of the respiratory chain and the subsequent burst of ROS when lower ratios of [NO ⅐ ]/[O 2 ] are encountered.
Conclusions-The mitochondrial production of ROS has been considered as a side process of the normal oxidative metabolism, its rate ranging between two levels determined by the mitochondrial metabolic states, namely States 4 (maximum) and 3 (minimum) (2). This study demonstrated that the mitochondrial production of ROS is not limited to these two values and may exhibit a degree of values modulated by endogenous NO ⅐ . Cellular conditions that would affect the availability of either L-Arg or other cofactors required for NOS activity would generate variable amounts of NO ⅐ and O 2 . . Although further studies need to be performed to fully understand the implications of these observations, two different groups of reports may support the importance of our observations. First, some studies support the important role of manganese superoxide dismutase in a system that produces both O 2 . and NO ⅐ in preventing the formation of the powerful oxidant peroxynitrite and the deleterious effect that the inactivation/lack of this enzyme may cause (20). Second, some studies examine the role of mitochondrial ROS, and that of endogenous agents that modulate their concentrations, as activators of intracellular signaling cascades involved in a variety of responses.