Modulation of superoxide-dependent oxidation and hydroxylation reactions by nitric oxide.

The rapid and spontaneous interaction between superoxide (O2-.) and nitric oxide (NO) to yield the potent oxidants peroxynitrite (ONOO-) and peroxynitrous acid (ONOOH), has been suggested to represent an important pathway by which tissue may be injured during inflammation. Although several groups of investigators have demonstrated substantial oxidizing and cytotoxic activities of chemically synthesized ONOO-, there has been little information available quantifying the interaction between O2-. and NO in the absence or the presence of redox-active iron. Using the hypoxanthine (HX)/xanthine oxidase system to generate various fluxes of O2-. and H2O2 and the spontaneous decomposition of the spermine/NO adduct to produce various fluxes of NO, we found that in the absence of redox-active iron, the simultaneous production of equimolar fluxes of O2-. and NO increased the oxidation of dihydrorhodamine (DHR) from normally undetectable levels to approximately 15 microM, suggesting the formation of a potent oxidant. Superoxide dismutase, but not catalase, inhibited this oxidative reaction, suggesting that O2-. and not hydrogen peroxide (H2O2) interacts with NO to generate a potent oxidizing agent. Excess production of either radical virtually eliminated the oxidation of DHR. In the presence of 5 microM Fe+3-EDTA to insure optimum O2-.-driven Fenton chemistry, NO enhanced modestly HX/xanthine oxidase-induced oxidation of DHR. As expected, both superoxide dismutase and catalase inhibited this Fe-catalyzed oxidation reaction. Excess NO production with respect to O2-. flux produced only modest inhibition (33%) of DHR oxidation. In a separate series of studies, we found that equimolar fluxes of O2-. and NO in the absence of iron only modestly enhanced hydroxylation of benzoic acid from undetectable levels to 0.6 microM 2-hydroxybenzoate. In the presence of 5 microM Fe+3-EDTA, HX/xanthine oxidase-mediated hydroxylation of benzoic acid increased dramatically from undetectable levels to 4.5 microM of the hydroxylated product. Superoxide dismutase and catalase were both effective at inhibiting this classic O2-.-driven Fenton reaction. Interestingly, NO inhibited this iron-catalyzed hydroxylation reaction in a concentration-dependent manner such that fluxes of NO approximating those of O2-. and H2O2 virtually abolished the hydroxylation of benzoic acid. We conclude that in the absence of iron, equimolar fluxes of NO and O2-. interact to yield potent oxidants such as ONOO-/ONOOH, which oxidize organic compounds. Excess production of either radical remarkably inhibits these oxidative reactions. In the presence of low molecular weight redox-active iron complexes, NO may enhance or inhibit O2-.-dependent oxidation and hydroxylation reactions depending upon their relative fluxes.

It is becoming increasingly apparent that certain types of inflammatory tissue injury are mediated by reactive metabolites of oxygen and nitrogen. For example, it has been demonstrated that administration of superoxide dismutase is effective at attenuating the tissue injury observed in experimental models of arthritis, chronic gut inflammation, and immune complex-induced pulmonary injury (1)(2)(3). Furthermore, models of joint, bowel, and lung inflammation have been shown to be associated with enhanced production of nitrogen oxides derived from the free radical nitric oxide (NO) (4 -6). Indeed, recent studies have demonstrated that inhibition of NO synthase also provides substantial protection against the inflammatory tissue injury observed in these models of acute and chronic inflammation (4 -6). These data suggest that both superoxide (O 2 . ) and NO are important mediators of inflammation-induced tissue injury and dysfunction.  (24). Therefore, the objectives of the present study were to: (a) systematically quantify the oxidizing and hydroxylating activity of NO and/or O 2 . in the absence or the presence of redox-active iron and (b) characterize these reactions using different fluxes of each radical. The physiological significance of our findings is discussed.

MATERIALS AND METHODS
Chemicals-Hypoxanthine (HX), 1 benzoic acid (BA), 2-hydroxybenzoic acid (HB), dimethyl formamide, spermine, diethylenetriaminepentaacetic acid, and cytochrome c (horse heart) were purchased from Sigma. Potassium superoxide (KO 2 ) was obtained from Pfaltz and Bauer, Inc. (Waterbury, CT). Xanthine oxidase derived from bovine milk was supplied by Calbiochem, and rhodamine 123 (RH) and dihydrorhodamine 123 (DHR) were purchased from Molecular Probes, Inc. (Eugene, OR). The spermine/NO adduct (Sp/NO) was a generous gift from Dr. Larry Keefer (National Cancer Institute, Frederick, MD). Human recombinant copper-zinc superoxide dismutase was obtained from Kabi-Pharmacia AB (Uppsala, Sweden), and catalase was purchased from Boehringer Mannheim. The tetramethylammonium salt of ONOO Ϫ was synthesized according to the method of Bohle et al. (25) ) were generated at 37°C in a total reaction volume of 500 l containing 20 mM potassium phosphate buffer (pH 7.4), 0.15 M NaCl, 0.5 mM HX, and various concentrations of xanthine oxidase (0 -10 milliunits/ml). The initial rates of O 2 . generation were determined spectrophotometrically by measuring the superoxide dismutase-inhibitable reduction of cytochrome c at 550 nm and ranged from 0.25 to approximately 10 nmol/min at 37°C (26). We used relatively large concentrations of HX and short incubation times in order to minimize the formation of urate and thus its potential interference as a free radical scavenger as well as to produce fluxes of O 2 . and H 2 O 2 that were approximately equal (27). Hydrogen peroxide production by HX/ xanthine oxidase at 37°C was determined using the peroxidase-catalyzed, H 2 O 2 -dependent oxidation of p-hydroxyphenylacetic acid to yield its fluorescent product 2,2Ј-dihydroxybiphenyl-5,5Ј-diacetate (28). We found that under our conditions, 1 milliunit/ml xanthine oxidase catalyzed the formation of approximately 1.0 nmol/min each of O 2 . and H 2 O 2 .
Nitric oxide was generated using the spontaneous decomposition of the Sp/NO. Sp/NO spontaneously decomposes at 37°C and pH 7.4 at a known and constant rate (t1 ⁄2 ϭ 39 min; Ref. 29) to yield 2 mol of NO/mol of adduct. Sp/NO solutions were prepared fresh each day as a 10 mM stock solution in ice-cold 10 mM NaOH and stored on ice until used. Various concentrations of Sp/NO (0 -200 M) were incubated at 37°C in a total reaction volume of 500 l containing 20 mM potassium phosphate buffer (pH 7.4) and 0.15 M NaCl. The initial rate of NO generation via Sp/NO decomposition at 37°C for 50 and 100 M Sp/NO was determined electrochemically using a NO-specific electrode (World Precision Instruments, Sarasota, FL) and was found to be 0.87 Ϯ 0.13 and 2.36 Ϯ 0.12 nmol NO/min, respectively. These values agreed well with those calculated for 50 and 100 M Sp/NO (1.1 and 2.2 nmol NO/min, respectively) based upon the published t1 ⁄2 ϭ 39 min at 37°C and pH 7.4 (29) Detection of Oxidizing and Hydroxylating Species-Oxidant formation was detected using the oxidant-sensitive nonfluorescent probe DHR, which when oxidized by two electrons is converted to the highly fluorescent RH product (30). Preliminary data from our laboratory confirmed previously published reports (31)  .
-driven Fenton chemistry and OH ⅐ production. Following the 30-min incubation period, reactions were terminated by dilution with 1.0 ml of cold phosphate-buffered saline (pH 7.4). Rhodamine formation was quantified using fluorescence spectroscopy in which an excitation wavelength of 500 nm and an emission wavelength of 536 nm were used. The concentration of RH was then calculated using regression values (i.e. slope and intercept values) obtained from plots using RH standards. Hydroxylating activity of the various systems described above was quantified by measuring the hydroxylation of BA to HB. At pH 7.4, the principal products of BA hydroxylation are the monohydroxylated derivatives, 2-, 3-, and 4-hydroxybenzoate, with 2-hydroxybenzoic acid representing approximately 60% of total product (32). Fluorescence emission of the HB derivative represents greater than 95% of total emission detected at 410 nm (i.e. for the three purified derivatives at equimolar concentrations and physiological pH (32)). 500-l reaction volumes containing 20 mM potassium phosphate buffer (pH 7.4) and 0.15 M NaCl. 1.0 mM benzoic acid, 0.5 mM HX, and various concentrations of xanthine oxidase or Sp/NO were incubated for 60 min at 37°C. For some experiments, catalase (15 g/ml) or superoxide dismutase (100 g/ml) were included in the reaction volumes, whereas in other experiments catalase and superoxide dismutase were omitted and 5 M Fe ϩ3 -EDTA was included. Following the 60-min incubation period, reactions were terminated by dilution with 0.5 ml of cold phosphatebuffered saline (pH 7.4). Production of HB was quantified by measuring the fluorescence obtained with excitation and emission wavelengths of 290 and 410 nm, respectively. The concentration of HB was determined using HB standards. All fluorescence emission measurements and spec-tra were obtained using an Aminco/Bowman Series 2 luminescence spectrometer (SLM Instruments, Inc., Rochester, NY).
Interaction between Peroxynitrite and Excess Superoxide or Nitric Oxide-The ability of excess NO to interact with chemically synthesized ONOO Ϫ was assessed using NO generated from Sp/NO and the tetramethylammonium salt of ONOO Ϫ . Briefly, NO formation (1.0 nmol/ min) was continuously monitored using electrochemical detection (WPI NO electrode) in the absence or the presence of a bolus addition of a small aliquot of ONOO Ϫ (11 nmol) in a 2-ml reaction volume containing 50 mM phosphate buffer. The rapid disappearance of the NO signal signified the interaction between NO and ONOO Ϫ . The interaction between excess O 2 . and ONOO Ϫ was determined by mixing a 10-fold molar excess of potassium superoxide (KO 2 ) prepared in ice-cold 0.1 N NaOH containing 0.1 mM diethylenetriaminepentaacetic acid with an alkaline solution of the ONOO Ϫ salt prepared as described above. A small aliquot alkaline O 2 . /ONOO Ϫ was immediately (Ͻ1 min) added to a 1-ml reaction volume containing 50 mM potassium phosphate buffer (pH 7.4) and 50 M DHR. Oxidation of DHR was compared with that when the same amount of alkaline ONOO Ϫ was added. A decrease in the oxidation of DHR was used as evidence to suggest that O 2 .
decomposes ONOO Ϫ . resulted in the production of RH in a pattern similar to that produced above such that DHR oxidation was maximal when fluxes of NO and O 2 . were approximately equal at 1.0 nmol/min for each (Fig. 3). Moreover, the yield of RH decreased substan-tially with further increases in O 2 . flux such that DHR oxidation was inhibited by 75% when fluxes of O 2 . reached a rate of 5.0 nmol/min (5-fold that of NO formation) (Fig. 3). We found that excess O 2 . may interact with ONOO Ϫ and partially decompose this oxidant as measured by the ability of O 2 . to decrease the ability of ONOO Ϫ to oxidize DHR (Fig. 4).  The addition of catalase or superoxide dismutase to solutions containing Fe ϩ3 -EDTA and HX/xanthine oxidase attenuated RH production by Ͼ90% (Fig. 7)     NO to yield the oxidant or oxidants (Fig. 5). These data also confirm a previous report (31) that found that neither O 2 . , H 2 O 2 , nor NO per se is capable of oxidizing substantial amounts of DHR in the absence of redox active metals such as iron or hemoproteins. Only oxidants such as those derived from Fenton-type reactions, ferryl hemoproteins, or ONOO Ϫ /ONOOH are potent enough oxidizing agents to oxidize DHR. Indeed, decomposition of peroxynitrous acid to nitrate has been suggested to proceed via a rate-limiting isomerization reaction that yields a potent oxidizing agent capable of hydroxylating organic substrates (8). Thus, we also assessed the ability of O 2 .

Oxidation of Dihydrorhodamine in the Absence of Iron-
and NO to interact (in the absence of iron and H 2 O 2 ) to hydroxylate BA. We found a similar pattern of hydroxylation of BA as observed for DHR oxidation in that equimolar fluxes (1.0 nmol/ min) of O 2 . and NO appeared to synergize to hydroxylate BA to HB (Fig. 8, A and B), although the magnitude of this hydroxylation reaction was rather small (Ͻ15% that with iron and H 2 O 2 present). Although we have not definitely identified ONOO Ϫ /ONOOH as the oxidants produced in this system, we expect that this would be the likely reaction pathway because of the rapid interaction between O 2 . and NO and because of the lack of alternative explanations for the production of equally potent oxidants. We speculate that the decreased production of RH or HB in the absence of iron and H 2 O 2 but in the presence of either excess NO or O 2 . may be accounted for on the basis of either secondary chemical interactions occurring directly between NO or O 2 . and ONOOH. It may also be due to the interaction between NO or O 2 . with free radical intermediates of DHR or BA to yield adducts with diminished fluorescence. The latter possibility does not appear to be a major pathway because we did not observe dramatic inhibition of DHR oxidation by excess NO in the iron-containing system (Fig. 6A) nor did nitrosation of HB by NO-derived nitrosating agents attenuate its fluorescence (data not shown). The former hypothesis appears to be the more viable explanation.
Although the direct reaction of ONOOH with either NO or O 2 . has not been definitively demonstrated, it has been suggested to be thermodynamically possible (33,34 Fig. 6A). These data suggest that as the ratio of NO/O 2 . increased from 0 to 2, oxidation of DHR increased by approximately 30% (Fig. 6A). As the ratio was increased further to 4.5, RH production was reduced by 40%. These data are reminiscent of those reported by Rubbo et al. (10) using xanthine oxidase-dependent iron-catalyzed lipid peroxidation. As expected, generation of O 2 . and H 2 O 2 (1.0 nmol/ min each) in the presence of iron increased production of HB tremendously (i.e. ϳ4000 nM) (Fig. 10A). Remarkably, the addition of NO to this iron-catalyzed hydroxylation system dramatically inhibited hydroxylation of BA such that equimolar fluxes of NO inhibited hydroxylation by 80% (Fig. 10A). Kanner et al. (38) recently suggested that NO may modulate ironmediated oxidative reactions by forming nitrosyl complexes with ferrous iron or by the direct interaction of NO with H 2 O 2 .
The sequence of reactions involving NO and iron may proceed as follows: The efficiency of such interactions could explain the results in Fig. 10A. Indeed it is well known that NO binds under physiological conditions with ferrous heme containing compounds (e.g. hemoglobin and myoglobin), and moreover these reactions are the chemical basis of current methodology used for NO detection (39). . and NO are produced (i.e. inflammatory foci) is dependent upon the relative fluxes of NO and O 2 . in the extracellular space. Our data suggest that excess production of one radical over the other may act as an endogenous modulator of ONOO Ϫ formation such that the steady state levels of this potent cytotoxic oxidant never accumulates above a certain amount. Indeed, the spontaneous acid-catalyzed decomposition of another potent oxidant, hypothiocyanous acid, is an example of autocatalytic regulation of oxidant formation (40). (b) NO may enhance or inhibit oxidation and hydroxylation reactions depending upon the absence or the presence of low molecular weight, redoxactive metals such as iron. Normally, there is little low molecular weight iron (e.g. amino acid, carbohydrate, or nucleotide chelates of iron) present in most cells and tissue, with the vast majority of this metal sequestered in its redox-inactive, protein-bound forms such as ferritin-or transferrin-bound iron. However, it is known that certain reductants (e.g. ascorbate and O 2 . ) are very effective in releasing iron from ferritin by reducing ferritin-bound Fe ϩ3 to Fe ϩ2 , which is no longer capable of being bound by the protein (41,42). In addition to ferritin, there is also a small but significant pool of low molecular weight iron chelate (e.g. non-protein-bound iron) located within cells. Studies by Deighton and Hider (43) have identified this low molecular weight iron chelate as a glutamate-iron complex (molecular weight of 1000 -1500) that can easily exchange its iron with other more potent chelators. Important consequences of the oxidant/hydroxylation-dependent flux of NO and O 2 . are temporal and spatial considerations. To achieve the maximum oxidant resulting from NO/ O 2 . , the site orientation and timing of the formation of these two species is crucial. The timing of the superoxide production relative to the NO can be distinctly different in vivo and have a limited overlap under some immunological and pathophysiological conditions. For instance, superoxide formation of neutrophils reaches a flux 10 times higher than that of NO within the first few minutes after treatment with phorbol ester (36). However, the flux of superoxide formed quickly subsides within an hour, whereas the NO production continues for several hours. The time overlap in which the flux of these two radicals is one to one is for a very limited time; therefore the amount of peroxynitrite formed is small. Conversely, cytokine-stimulated alveolar macrophage are thought to generate both NO and O 2 .
at the same sustained rate for long period of time, implying that the oxidant formed may be intentionally held high in this specific cell line (15). Yet, RAW macrophages appear to generate solely nitrosating agents via the NO/O 2 reaction without the presence of superoxide (35). This switching between oxidation, hydroxylation, and nitrosation appears to be well orchestrated in the immune response to pathogens and appears to be critical in host defense. Although NO and superoxide can be generated from the same cell type and cytokine influence, kinetic considerations must be carefully examined to determine the reactive intermediates involved.