Reaction of Peroxynitrite with Mn-Superoxide Dismutase

Manganese superoxide dismutase (Mn-SOD), a critical mitochondrial antioxidant enzyme, becomes inactivated and nitrated in vitro and potentially in vivo by peroxynitrite. Since peroxynitrite readily reacts with transition metal centers, we assessed the role of the manganese ion in the reaction between peroxynitrite and Mn-SOD. Peroxynitrite reacts with human recombinant and Escherichia coli Mn-SOD with a second order rate constant of 1.0 ± 0.2 × 105 and 1.4 ± 0.2 × 105 m − 1s− 1 at pH 7.47 and 37 °C, respectively. TheE. coli apoenzyme, obtained by removing the manganese ion from the active site, presents a rate constant <104 m − 1 s− 1for the reaction with peroxynitrite, whereas that of the manganese-reconstituted apoenzyme (apo/Mn) was comparable to that of the holoenzyme. Peroxynitrite-dependent nitration of 4-hydroxyphenylacetic acid was increased 21% by Mn-SOD. The apo/Mn also promoted nitration, but the apo and the zinc-substituted apoenzyme (apo/Zn) enzymes did not. The extent of tyrosine nitration in the enzyme was also affected by the presence and nature (i.e.manganese or zinc) of the metal center in the active site. For comparative purposes, we also studied the reaction of peroxynitrite with low molecular weight complexes of manganese and zinc with tetrakis-(4-benzoic acid) porphyrin (tbap). Mn(tbap) reacts with peroxynitrite with a rate constant of 6.8 ± 0.1 × 104 m − 1s− 1 and maximally increases nitration yields by 350%. Zn(tbap), on the other hand, affords protection against nitration. Our results indicate that the manganese ion in Mn-SOD plays an important role in the decomposition kinetics of peroxynitrite and in peroxynitrite-dependent nitration of self and remote tyrosine residues.

SOD plays an active role detoxifying the cell from this species. In addition, pharmacological agents and cytokines that promote intracellular reactive oxygen species production, like paraquat (7) and tumor necrosis factor-␣ (8), induce Mn-SOD expression. Experiments with knock-out mice shed further light on the relevance of this enzyme, with Mn-SOD-deficient mice surviving only up to 3 weeks of age (9,10) and presenting many features of mitochondrial disease associated with reactive oxygen species toxicity (11).
Nitric oxide (NO ⅐ ) is a relatively unreactive free radical formed by nitric oxide synthase (12). However, fast reaction of nitric oxide with superoxide gives rise to peroxynitrite anion (ONOO Ϫ ), a potent oxidant (13)(14)(15). Peroxynitrite is formed during sepsis, inflammation, excitotoxicity, and ischemiareperfusion of tissues, conditions under which the cellular production of nitric oxide and superoxide increase (12, 16 -18), and participates in reactions related with the pathological expression of these processes. Recent reports regarding the presence of nitric oxide synthase in the mitochondria (19 -21), along with the easy diffusion of nitric oxide through membranes (22), make the intramitochondrial formation of peroxynitrite possible and highlight the relevance of its interactions with intramitochondrial targets.
Mn-SOD inhibits peroxynitrite formation in mitochondria, but it may be oxidatively inactivated by excess peroxynitrite. Indeed, peroxynitrite-mediated nitration of tyrosine 34 of human Mn-SOD results in enzyme inactivation in vitro (23,24). The presence of a nitrated and dysfunctional enzyme in rejected human renal allografts (25) strongly supports the relevance of this process in vivo. Mn-SOD inactivation, due to tyrosine 34 nitration, would lead to an increase in peroxynitrite formation that would in turn impair mitochondrial energy metabolism (26) and signal apoptotic cell death (27).
Peroxynitrite decomposition is catalyzed by a variety of Lewis acids (28). These are electron-accepting compounds like H ϩ (29,30), carbon dioxide (31), and transition metals (29) that favor the cleavage of the O-O bond and lead to the formation of nitrating species. In the case of H ϩ , peroxynitrous acid (ONOOH) undergoes homolysis to hydroxyl radical and nitrogen dioxide with yields up to 30% (28,32), resulting in nitration yields in the range of 6 -10% (29,30). In the case of transition metal-containing compounds, such as metalloproteins and Mnporphyrin SOD mimetics, higher nitration yields are obtained (29,(33)(34)(35). These facts led us to consider that the manganeseion should play an important role in the reaction of peroxynitrite with Mn-SOD. In addition, manganese may facilitate the formation of nitrating species at the active site, which could react with the critical tyrosine 34.
In this work, we have studied the role of the manganese metal center in the decomposition kinetics of peroxynitrite and in peroxynitrite-dependent nitration of the enzyme and nonprotein aromatic residues. These studies shed light on the nature of peroxynitrite reaction with Mn-SOD and provide further rationale to account for the toxic actions that peroxynitrite may promote in vivo.
A rabbit polyclonal antibody against nitrotyrosine was raised with nitrated keyhole limpet hemocyanin and purified in our laboratory by affinity chromatography as described elsewhere (36). The rabbit polyclonal antibody against human Mn-SOD was a kind gift from Dr. Ling-Yi Chang (University of Colorado). The donkey monoclonal antibody against rabbit IgG, linked to horseradish peroxidase, nitrocellulose (0.45 m pore size, Hybond C extra), and luminol-enhanced chemiluminescence detection kit (ECL) were obtained from Amersham Pharmacia Biotech.
The reagents for Mn-SOD expression and purification were the following: tryptone and yeast extract for LB medium were obtained from Difco, and methyl viologen (paraquat), ampicillin, Tris-HCl, cytochrome c, CaCl 2 , DNase, RNase, Sephadex G-25 and CM-Sepharose were purchased from Sigma. KCl and potassium phosphate were from Mallinckrodt, and MnCl 2 was from Matheson, Coleman, and Bell. (NH 4 ) 2 SO 4 (enzyme-grade) was the highest grade commercially available.
Expression and Purification of Human Recombinant Mn-SOD-E. coli sodAsodB strain QC774 lacking Mn-SOD and Fe-SOD was transformed with the pGB1 expression vector (39) containing the coding sequence of wild type human recombinant Mn-SOD (hrMn-SOD), minus that encoding residues 2-24 (i.e. minus the mitochondrial targeting sequence). The cells were grown in LB medium supplemented with 0.2 mM MnCl 2 , 50 g/ml ampicillin, and 30 M paraquat to induce the overexpression of pGB1. The wild type pGB1-hrMn-SOD-transformed cell cultures were incubated overnight at 37°C. The cultures were incubated with orbital shaking at 200 rpm. The cells were harvested and broken by ultrasonication on ice. The crude extracts were treated with 100 units/ml DNase, 10 mg/ml RNase, 10 mM MgCl 2 , 10 mM CaCl 2 , 0.15 M NaCl in 10 mM K 2 PO 4 , pH 7.0, at room temperature for 1 h. Mn-SOD was precipitated by (NH 4 ) 2 SO 4 fractionation between 65 and 80% saturation. The precipitate obtained at 80% saturation was desalted by gel filtration chromatography on Sephadex G-25 in 20 mM potassium acetate buffer, pH 6.6. The proteins were then applied to a CM-Sepharose column and eluted in a gradient of 0 -100 mM KCl. The yield of purified hrMn-SOD was 72%.
SOD Activity-SOD activity was determined measuring the inhibition of the reduction of cytochrome c by the xanthine-xanthine oxidase system (40). The concentrations of E. coli Mn-SOD and hrMn-SOD were measured by absorbance at 282 nm (⑀ 282 ϭ 8.67 ϫ 10 4 M Ϫ1 cm Ϫ1 (3)) and 280 nm (⑀ 280 ϭ 1.81 ϫ 10 5 M Ϫ1 cm Ϫ1 (41,42)) respectively, and by the bicinchoninic acid method obtaining concordant results. Enzyme preparations of hrMn-SOD and E. coli Mn-SOD, used in the different assays, typically had specific activities of 2500 and 3200 units/mg, respectively.
Removal of Manganese from the Active Site, Reconstitution and Substitution by Zinc-E. coli Mn-SOD was dialyzed for 16 -20 h against 20 mM 8-hydroxyquinoline, 2.5 M guanidinium hydrochloride, 5 mM Tris chloride, and 0.1 mM EDTA at pH 3.8 and 4°C (43). This caused a loss of manganese-from the active site and of enzyme activity. This enzyme was then dialyzed for 16 h at 4°C and pH 7. The apoenzyme and the reconstituted enzymes were assayed for SOD activity and protein content as described above. Recovered specific activities as percentages of original activity were-apoenzyme 0%, apo/Mn enzyme 21%, and apo/Zn enzyme 1%.
Metal Analyses-Zinc and manganese content of the enzymes were determined with a graphite furnace atomic absorption spectrometer (Spectra 20, Varian Instruments, Victoria, Australia). Calibration curves of each element were made from dilutions in deionized water of atomic absorption standards. The enzyme samples were diluted with water and metal ion content calibrated against standards.
Kinetic Studies-The kinetics of peroxynitrite decomposition in absence and presence of enzyme were studied in a stopped-flow spectrophotometer (Applied Photophysics, SF.17MV) with a mixing time of less than 2 ms, at 302 nm. An initial rate approach was used to analyze the data; the first 0.1 s were fit to a linear plot, and the rate constant was determined as the ratio between the slope and the difference between the initial and final absorbance (A o Ϫ A f ). To ensure the accuracy of the rate constant determinations, 200 absorbance measurements were acquired during the initial part of the reaction (first 0.2 s) and 200 further points were acquired until more than 99.9% peroxynitrite had decomposed (0.2-10 s) (45).
Kinetics of Mn(tbap) and Zn(tbap) reaction with peroxynitrite were studied under pseudo-first order conditions with peroxynitrite in excess over the porphyrin, following absorbance changes on the porphyrin, at 468 and 421 nm, respectively, as described previously (34). Data obtained in the first 0.03-0.2 s were fit to single exponential and pseudofirst order rate constants determined. Reactions were performed at 37.0 Ϯ 0.1°C, and the final pH of the mixture was measured at the outlet.
Kinetics of HPA nitration by peroxynitrite, in the absence and in presence of Zn(tbap), were studied using an initial rate approach, at 430 nm.
Nitration of HPA-Nitration of HPA by peroxynitrite was assessed spectrophotometrically. After the reaction had taken place, the pH of the solution was adjusted to 10 -11 with 6 N NaOH, and absorbance was recorded at 430 nm. Absorbance of a control containing everything except peroxynitrite was subtracted before determining-the nitro-HPA concentration (⑀ 430 ϭ 4400 M Ϫ1 cm Ϫ1 ) (29). Percent yield was calculated with respect to initial peroxynitrite concentration.
Nitration of HPA was also studied by high performance liquid chromatography (HPLC)-based techniques. Standards and samples of HPA and nitro-HPA were separated using a Gilson 306 pump (Wilson Medical Electronics, Inc.) and a C 18 -derivatized silica column. Samples were eluted with 50 mM potassium phosphate, pH 3, and HPLC-grade meth-anol. Elution conditions consisted of 5 min at 10% (v/v) methanol, followed by a 10-min linear gradient from 5 to 45% (v/v) methanol and finally 15 min at 45% (v/v) methanol. Absorbance was monitored at 280 and 365 nm. Nitro-HPA concentration was calculated from the peak area using a calibration curve performed with a standard. Samples that contained protein were filtered using a cellulose membrane with a 10,000 Da cut-off, previous to injection.
Western Blot Analyses-SDS-polyacrylamide gel electrophoresis of protein samples was performed on 13% polyacrylamide gels, and proteins were transferred electrophoretically (20 mA, 16 h) to nitrocellulose membranes. Membranes were blocked with 5% bovine serum albumin (BSA) in 50 mM Tris chloride, pH 7.4, 150 mM NaCl (TBS), 0.6% Tween 20 (blocking buffer). For detection of nitrotyrosine nitrocellulose membranes were incubated (1 h at 25°C) with 0.2 mg/ml anti-nitrotyrosine antibody (1/1000 dilution) in blocking buffer. After extensive washing in TBS, 0.6% Tween 20, the immunocomplexed membranes were probed (1 h at 25°C) with horseradish peroxidase-linked secondary antibody (1/6000 dilution) in TBS, 0.1% BSA, 0.3% Tween 20. Probed membranes were washed with TBS, 0.3% Tween 20, and immunoreactive proteins were visualized with a luminol-enhanced chemiluminescence detection kit (ECL). Photographs of the Western blots were then scanned, and relative nitration was determined by densitometric techniques using the data analyzer program Scion Image (Scion Corp.). Images were set to gray scale, and average pixel intensity (sum of all the gray values of all the pixels divided by the number of pixels) of each band was determined. Band area was also determined. The product of the band area and the average pixel intensity is proportional with the amount of nitrotyrosine in the blot.
Detection of Mn-SOD dimers was carried out in the same way, probing with a 1/6000 dilution of the antibody against human Mn-SOD.
Electrospray-Mass Spectrometry Studies-All electrospray ionization-mass spectrometry experiments were performed on a PE-Sciex (Concord, Ontario, Canada) API-III triple-quadrupole mass spectrometer equipped with an atmospheric pressure ion source. Positive ion mass spectra were acquired for capillary LC-MS. For the capillary LC-MS analysis, the effluent from a 300-m inner diameter ϫ 15 cm capillary Vydac C18 column (LC-Packings, San Francisco, California) was introduced directly into the ionization needle of the mass spectrometer. The column was equilibrated in 0.1% formic acid, and the flow rate was maintained at 5.7 l/min by splitting the 0.4 ml/min flow from a Hewlett-Packard model 1050 HPLC system with an Accurate brand (1:70) stream splitter and the capillary column. The protein was desalted by dialysis against deionized water, concentrated using spin filters (Fisher, 5000 M r cut off), and loaded in 0.1% formic acid that was maintained for 3-5 min of the run, followed by a linear gradient of 0 -80% acetonitrile in 0.1% formic acid over the next 10 min.
Mn-SOD Structure Analysis-E. coli Mn-SOD structure analysis was performed using the Swiss PDB Viewer program (Glaxo Wellcome Experimental Research).
General Conditions-All experiments involving peroxynitrite reactions were carried out in 100 mM potassium phosphate buffer, 0.1 mM DTPA at 37°C and pH 7.4 Ϯ 0.1, unless otherwise specified. These buffers were prepared daily to minimize carbon dioxide contamination. Experiments reported herein were performed a minimum of three times with similar results obtained. Results are expressed as means Ϯ S.D. or by a representative example. Graphics and curve fitting were generated in Slide-Write 2.1 for Windows (Advanced Graphic Software Inc.).

Kinetics of Peroxynitrite Reaction with Mn-SOD-
The decay of peroxynitrite (0.1 mM) was followed in absence and in presence of hrMn-SOD (10 M) (Fig. 1). At these concentrations pseudo-first order conditions are not achieved, and the kinetic traces of peroxynitrite decay in the presence of the enzyme did not follow a single exponential function (Fig. 1, inset). Indeed, the rate of peroxynitrite decomposition in the presence of hrMn-SOD was initially faster, reflecting the reaction of peroxynitrite with the enzyme (Fig. 1, inset).
To obtain the rate constant of peroxynitrite reaction with hrMn-SOD, the decay of peroxynitrite (0.2 mM) was followed in the absence and in presence of different concentrations (2.5-15 M) of hrMn-SOD tetramer, obtaining plots such as presented in Fig. 1. The apparent rate constant of peroxynitrite decomposition (k obs ) was determined by measuring the initial rate of peroxynitrite decay (i.e. during the first 100 ms), as reported recently (45). The plot of the apparent rate constants of peroxynitrite decomposition as a function of Mn-SOD concentrations was linear (Fig. 2). The slope of such plot rendered a second order rate constant for the reaction of hrMn-SOD with peroxynitrite of 1. The fast reaction of peroxynitrite with the enzyme suggested that the manganese ion in the active site should be the primary target of the oxidant. Considering this and the fact that other compounds that contain manganese ions (33)(34)(35) have been reported to catalyze the nitration of phenols, it seemed reasonable that the reaction of peroxynitrite with the manganese ion would lead to the formation of a nitrating species that could be responsible for the nitration of tyrosine 34. These considerations led us to assess the role of the metal center in peroxynitrite decomposition kinetics and peroxynitrite-mediated nitration of the enzyme.
Removal of Manganese from the Active Site and Reconstitution with Manganese or Zinc Ions-To study the role of the metal center in peroxynitrite reaction with Mn-SOD, the metal was removed from the active site of the E. coli enzyme, and the apoenzyme (apo) was obtained. The apoenzyme was then reconstituted with manganese (apo/Mn) or substituted with zinc (apo/Zn). The activity and metal content of these enzymes is shown in Table I. The native enzyme presented 0.7 manganeseatoms per monomer, in agreement with that reported for the E. coli enzyme (3). The manganese-content of the holoenzyme, apo, apo/Mn, and apo/Zn enzymes correlated with the activity, in agreement with the literature (51). The apo/Zn enzyme presented slightly more than one atom of zinc per monomer, implying the existence of unspecific binding of the metal to the enzyme.
Zinc was considered to be a good candidate for the substitution of manganese and evaluation of peroxynitrite reactions with Mn-SOD. On one hand, the Zn(II) and Mn(II) ions have similar sizes (ionic radius of 0.74 and 0.80 Å, respectively (52)), and zinc is reported to bind to Mn-SOD in a stoichiometric amount displacing manganese-from the active site (51), so the apo/Zn conformation would probably be similar to the native one. On the other hand, zinc is not capable of rendering high oxidation states (53), such as those proposed to participate in tyrosine nitration in the case of manganese and iron (i.e. oxomanganese (OϭMn(IV)) and oxo-iron (OϭFe(IV))) (34, 35), so the apo/Zn enzyme would not be an efficient promoter of nitration reactions.
The apo, apo/Mn, and apo/Zn forms of hrMn-SOD were obtained in 0.5 M sucrose. When the disaccharide was extracted by dialysis, these enzymes largely precipitated (60 -70%). The different stability of the human and E. coli preparations may be due to their different quaternary structures, tetrameric and dimeric, respectively.

Role of the Metal Center in Peroxynitrite Decomposition
Kinetics-Peroxynitrite decomposition kinetics in the presence of the apo, apo/Mn, and apo/Zn forms of the enzyme was assessed as described in Fig. 1. Table II reports the second order rate constants determined for the decomposition of peroxynitrite in the presence and absence of 2.5 M enzyme. The apoenzyme did not affect peroxynitrite decomposition kinetics in a detectable way, so the product of its second order rate constant and the concentration of enzyme is less than 10% of that determined for peroxynitrite alone (k apo ϫ [apo]Ͻ 0.1 ϫ k ONOO Ϫ). These considerations indicate that the second order rate constant of the apoenzyme must be smaller than 4 ϫ 10 4 M Ϫ1 s Ϫ1 and probably reflects the reaction rate of peroxynitrite with the enzyme amino acids (45). The apo/Mn rate constant was 37% that of the holoenzyme, in agreement with the data obtained for the SODspecific activity, which is in turn proportional to the manganese content (51). Thus it is reasonable to assume that apo/Mn preparations with the same manganese content than the holoenzyme will recover 100% of the rate constant value. The apo/Zn presented an unexpectedly high rate constant, approximately twice that obtained for the holoenzyme. These results support the idea that metal center plays a central role in peroxynitrite decomposition kinetics.
Role of the Metal Center in Enzyme Nitration-E. coli Mn-SOD, apo, apo/Mn, and apo/Zn (5 M) were incubated with peroxynitrite (0.1 mM). Tyrosine nitration was assessed by immunoblot techniques using a highly specific anti-nitrotyrosine antibody (Fig. 3). The apo/Mn enzyme was more nitrated than the apo/Zn enzyme, suggesting that the metal center could be involved in protein-tyrosine nitration. Although the apo/Mn was more nitrated than the native enzyme, the latter presented larger inactivation with respect to the untreated enzyme, consistent with subtle changes in protein conformation on apo/Mn. Surprisingly, the apoenzyme was more nitrated than the na- The blot was scanned and analyzed by densitometric techniques. Nitration is expressed as relative to the native enzyme exposed to peroxynitrite. Activity is expressed relative to the native enzyme incubated in absence of peroxynitrite. tive, apo/Mn, and apo/Zn enzymes, which must be related to a larger surface exposure and accessibility to tyrosine residues in the apoenzyme, due to the denaturalization and metal extraction process. Mass spectrometry studies of the native enzyme and the apoenzyme showed that incubation with peroxynitrite resulted in the formation of a species with a molecular mass 47 Da higher than the control enzyme (22,944 to 22,991 Da), consistent with the addition of a single nitro group to both enzymes (Fig. 4). Integration of the area below the peaks showed that the nitration yields with respect to the protein were higher in the apoenzyme than in the holoenzyme, in agreement with the results obtained by immunoblot techniques.
The exposure of the holoenzyme to peroxynitrite under these conditions resulted not only in nitration but also in a small degree of dimerization of the enzyme. Densitometric analysis of Western blots, using an antibody against human Mn-SOD, revealed that ϳ3-5% of the native enzyme was present as dimer after the exposure to peroxynitrite (not shown).
Mn-SOD and Peroxynitrite-dependent Nitration of Phenols-4-Hydroxyphenylacetic acid (5 mM) was exposed to peroxynitrite (1 mM) in the absence and in the presence of increasing concentrations of E. coli Mn-SOD (Fig. 5A). While in absence of enzyme 10.3% of the added peroxynitrite was recovered as nitro-HPA, and in presence of the E. coli Mn-SOD, nitration yields increased with the concentration of enzyme, fitting a hyperbolic profile. A maximum of 12.5% nitration yield could be predicted from these data. Considering the maximum nitration yield (%R max ) and that obtained in absence of enzyme (%R 0 ) (see Equation 1) the maximum increase in nitration yield was calculated to be 21%.
The Mn-SOD-promoted increase in nitration yields was also assessed by HPLC techniques. HPA (5 mM) was incubated with peroxynitrite (1 mM) in the absence and presence of E. coli Mn-SOD (5 M). Mn-SOD was extracted from the samples by filtration, and HPA and nitro-HPA were separated using a reverse phase chromatography column, presenting elution times of 19.7 and 24.1 min, respectively. Nitro-HPA was quantified measuring the peak area obtained at 360 nm. In the absence of the enzyme a nitration yield of 10.9 Ϯ 0.7% was obtained, whereas in the presence of Mn-SOD this increased to 12.6 Ϯ 0.4% (not shown), in complete agreement with the data obtained by spectrophotometric techniques. Then the role of the metal center, in promoting peroxynitritedependent nitration of HPA, was assessed. HPA (5 mM) was exposed to peroxynitrite (1 mM) in the presence of holo, apo, apo/Mn, and apo/Zn enzymes (5 M) (Fig. 5B), and percentage increases in nitration yields were determined to be 14.3, 0, 10.6, and 0.6%, respectively. These results unambiguously show that the manganese-ion is responsible for the increase in Kinetics of Peroxynitrite Decomposition in the Presence of Mn(tbap) and Zn(tbap)-Due to the surprisingly high rate constant obtained for the reaction of peroxynitrite with the apo/Zn, compared with that obtained for the native and apo/Mn, and to obtain a better comprehension of peroxynitrite reactivity with these metals, we studied the reaction of peroxynitrite with low molecular weight complexes of manganese and zinc with a substituted porphyrin (tbap).
The reaction rate of the Mn(tbap) and Zn(tbap) (8 M) with peroxynitrite was followed at 468 and 421 nm, the respective Soret peak wavelengths of these porphyrins. Peroxynitrite was present in 5-50-fold excess (0.04 -4 mM) over the porphyrin achieving pseudo-first order conditions. The kinetic traces of the Mn(tbap) reaction with excess peroxynitrite displayed a biphasic pattern as follows: a first order descent (Fig. 6A), followed by a slow recovery of the absorbance values. A total absorbance recovery was observed for Mn(tbap), similar to that described for the reaction of other manganese-porphyrins (34). In the case of Zn(tbap), the kinetics were more complex. An initial rapid descent (Fig 6A, inset) was followed by a slower one. This second descent was independent of peroxynitrite concentration and had a k obs value similar to that of the protoncatalyzed decomposition of peroxynitrite, suggesting reactions between peroxynitrite-derived radicals and the porphyrin moiety. This latter idea is consistent with the fact that Zn(tbap) recovered its initial absorbance only partially. The observed rate constants determined from the exponential fit in the first 30 -200 ms (Fig. 6A, inset) were plotted in function of peroxynitrite concentration (Fig. 6B). From the slope of these plots, second order rate constants for the reactions of Mn(tbap) and Zn(tbap) with peroxynitrite of 6.8 Ϯ 0.1 ϫ 10 4 and 4.9 Ϯ 0.1 ϫ 10 5 M Ϫ1 s Ϫ1 at pH 7.2 and 37°C were obtained, respectively. The ratio between the rate constants obtained from the Mnand Zn-porphyrins is in good agreement with that obtained with the native enzyme and zinc-substituted apoenzyme.

Mn(tbap) and Zn(tbap) Catalysis of Peroxynitrite-dependent Nitration of HPA-Mn(tbap)
(1-20 M) increased peroxynitrite (1 mM)-dependent yield of nitration of HPA (5 mM) in a dosedependent fashion. These results fit a hyperbolic plot, and a maximum increase in nitration yields of 350% was determined, in agreement with previous reports (33). On the other hand, Zn(tbap) did not promote peroxynitrite-dependent nitration of HPA (Fig. 7), and in fact at high concentrations of Zn(tbap) (20 M) a small decrease in nitration yields was observed, in agreement with recent reports (35). These results are in agreement with those obtained with the native enzyme and zinc-substituted apoenzyme, supporting the role for the active site manganese ion in promoting the peroxynitrite-dependent nitration of low molecular weight phenols. At the same time it is clear that either the environment of the manganese ion in the enzyme or the poor accessibility of HPA to the active site makes the enzyme a less efficient promoter of peroxynitrite-dependent  nitration of phenols, compared with the manganese-porphyrin.
The fast reaction of Zn(tbap) with peroxynitrite in conjunction with a marginal effect on nitration yields supports the idea that zinc is behaving like a Lewis acid, resulting in nitration yields similar to those of H ϩ -catalyzed nitration. This hypothesis was further evaluated by experiments showing that Zn(tbap) (10 -50 M) accelerated peroxynitrite (1 mM)-dependent HPA (5 mM) nitration and diminished peroxynitrite nitration yields in presence of carbon dioxide (0.2 mM) (not shown).
Effect of Scavengers on Mn-SOD Nitration and Inactivation-Different compounds known to interact either with peroxynitrite or with hydroxyl radical or nitrogen dioxide were assessed for their ability to inhibit or increase the nitration and inactivation of E. coli Mn-SOD (5 M) by peroxynitrite (0.5 mM) (Fig. 8A). Coincubation of Mn-SOD with 1 mM glutathione (GSH) and HPA largely prevented both nitration and inactivation. Dimethyl sulfoxide (Me 2 SO) (10 mM), a well known hydroxyl radical scavenger, was a weak inhibitor of inactivation and nitration, implying a modest role for the radical pathway in the nitration of the holoenzyme. Most interestingly, bicarbonate (HCO 3 Ϫ ) (1 mM carbon dioxide) also rendered partial protection at this enzyme concentration.
At higher concentrations of Mn-SOD, HCO 3 Ϫ enhanced the nitration and inactivation of the enzyme by peroxynitrite, whereas HPA and GSH (5 mM) achieved total protection, the latter both in absence and in presence of HCO 3 Ϫ .

DISCUSSION
Peroxynitrite reacts with human recombinant and E. coli Mn-SOD in a direct reaction with second order rate constants of 1.0 Ϯ 0.2 ϫ 10 5 and 1.4 Ϯ 0.2 ϫ 10 5 M Ϫ1 s Ϫ1 at pH 7.47 and 37°C, respectively. The rate constant for the reaction with the E. coli apoenzyme was at least 1 order of magnitude smaller, whereas that of the apo/Mn was comparable with that of the holoenzyme, confirming that the reaction between Mn-SOD and peroxynitrite largely depends on the presence of the metal in the active site. The reactions of peroxynitrite with metalloproteins are typically fast; therefore, these are likely to be major targets in vivo. In this context, the reactivity of peroxynitrite with Mn-SOD is similar to that previously reported for mitochondrial aconitase (47), cytochrome c 2ϩ (48), alcohol dehydrogenase (46), and peroxidases (49).
The surprisingly high rate constant determined for the reaction between peroxynitrite and the zinc-substituted apoenzyme suggested that the decomposition kinetics of peroxynitrite could also be affected by the nonredox metal zinc (53). Indeed, kinetic rate determinations for the reaction of peroxynitrite with Mn(tbap) and Zn(tbap), revealed that the zinc-substituted porphyrin reacts with peroxynitrite faster than its manganese counterpart. The decomposition of peroxynitrite in the presence of Mn-SOD and Mn(tbap) is attributed to a redox reaction which involves the oxidation of the metal ion (34,54). However, in the presence of the zinc-substituted apoenzyme or Zn(tbap), peroxynitrite decomposition must proceed through a different pathway that may involve the utilization of the zinc ion as a Lewis acid.
Peroxynitrite reaction with Mn-SOD leads to the formation of nitrating species, capable of modifying low molecular weight aromatic compounds in a manganese-dependent process (Fig.  5). In addition, and as reported for the human recombinant Mn-SOD (23,24), peroxynitrite promoted the inactivation of the E. coli enzyme mainly by the nitration of one tyrosine residue. While in the apoenzyme peroxynitrite-derived hydroxyl radical and nitrogen dioxide-mediated tyrosine nitration, the manganese ion played an important role in tyrosine nitration in the holoenzyme (Fig. 3). The nitrating species could be either a nitronium ion (NO 2 ϩ ) bound to the metal (29,55) or an oxo-manganese complex plus nitrogen dioxide (34,54). Nevertheless, the kinetic traces of peroxynitrite decomposition in the presence of Mn-SOD do not fit single exponential kinetics indicating that the enzyme is being consumed in the reaction with peroxynitrite, so even though it enhances nitration yields, it is not a true catalyst. In all, metal cofactors prone to undergo redox transitions, such as manganese, iron, or copper, are likely to promote site-specific nitration of protein aromatic residues; this idea is consistent with previous data on prosta- Ϫ . Nitration is expressed relative to the native enzyme exposed to peroxynitrite. Activity is expressed relative to the native enzyme incubated in absence of peroxynitrite (Ctrl). cyclin synthase and Cu,Zn-SOD as well (55)(56)(57).
E. coli Mn-SOD tyrosine residues have different degrees of solvent accessibility, as revealed by the analysis of the native structure using the Swiss PDB Viewer program. Tyrosine 34 is less accessible to the solvent than Tyr-2, Tyr-9, and Tyr-11 and equally accessible as Tyr-173, Tyr-174, and Tyr-184, but tyrosine 34 is the residue located closest to the active site, only at 5 Å from the manganese ion. The attraction of peroxynitrite to the active site, probably by the basic residues in the channel entrance, and its reaction with the manganese ion leading to the formation of nitrating species provide a reasonable explanation to the fact that tyrosine 34 is the tyrosine residue most susceptible to nitration by peroxynitrite (23,24).
Considering the specific activity of purified Mn-SOD (4000 Ϯ 1000 units/mg (1,4,58,59)) and that observed in mitochondria (10 Ϯ 2 units/mg, in heart and liver mitochondria), 2 as well as the enzyme molecular mass (88 kDa) and mitochondrial volume (1.2 l/mg), a concentration of Mn-SOD inside the mitochondria of 20 Ϯ 10 M (80 M subunits) was estimated. In concentrations similar to those found in the mitochondria, GSH protected Mn-SOD from nitration and inactivation both in the absence and presence of carbon dioxide (Fig. 8). These data underscore the role of GSH as a mitochondrial antioxidant and raise the question on the mechanism of nitration and inactivation of the enzyme in vivo. In this sense, Mn-SOD reported to copurify with QH 2 :cytocrome c reductase (complex III) (60) may be partially associated to the sites of mitochondrial superoxide production, hence of peroxynitrite formation, constituting a preferential target for the oxidant.
Since Mn-SOD is a critical mitochondrial antioxidant, its nitration represents a severe hazard that will promote oxidative damage and may ultimately signal cell death. Strategies directed to attenuate nitration of Mn-SOD tyrosine 34 should result in a better mitochondrial and cellular outcome under conditions of excess peroxynitrite formation.