Protein Tyrosine Nitration in Cytokine-activated Murine Macrophages INVOLVEMENT OF A PEROXIDASE/NITRITE PATHWAY RATHER THAN PEROXYNITRITE*

Peroxynitrite, formed in a rapid reaction of nitric oxide (NO) and superoxide anion radical (O 2 .), is thought to mediate protein tyrosine nitration in various inflammatory and infectious diseases. However, a recent in vitro study indicated that peroxynitrite exhibits poor nitrating efficiency at biologically relevant steady-state concentrations (Pfeiffer, S., Schmidt, K., and Mayer, B. (2000) J. Biol. Chem. 275, 6346–6352). To investigate the molecular mechanism of protein tyrosine nitration in intact cells, murine RAW 264.7 macrophages were activated with immunological stimuli, causing inducible NO synthase expression (interferon- (cid:1) in combination with either lipopolysaccharide or zymosan A), followed by the determination of protein-bound 3-nitrotyrosine levels and release of potential triggers of nitration (NO, O 2 ., H 2 O 2 , peroxynitrite, and nitrite). Levels of 3-nit- rotyrosine started to increase at 16–18 h and exhibited a maximum at 20–24 h post-stimulation. Formation of O 2 . was maximal at 1–5 h and decreased to base line 5 h after stimulation. Release of NO peaked at (cid:1) 6 and (cid:1) 9 h after stimulation with interferon- (cid:1) /lipopolysaccharide

after stimulation. Release of NO peaked at ϳ6 and ϳ9 h after stimulation with interferon-␥/lipopolysaccharide and interferon-␥/zymosan A, respectively, followed by a rapid decline to base line within the next 4 h. NO formation resulted in accumulation of nitrite, which leveled off at about 50 M 15 h post-stimulation. Significant release of peroxynitrite was detectable only upon treatment of cytokine-activated cells with phorbol 12myristate-13-acetate, which led to a 2.2-fold increase in dihydrorhodamine oxidation without significantly increasing the levels of 3-nitrotyrosine. Tyrosine nitration was inhibited by azide and catalase and mimicked by incubation of unstimulated cells with nitrite. Together with the striking discrepancy in the time course of NO/O 2 .
release versus 3-nitrotyrosine formation, these results suggest that protein tyrosine nitration in activated macrophages is caused by a nitrite-dependent peroxidase reaction rather than peroxynitrite.
The free radical nitric oxide (NO) is produced by constitutive and inducible nitric-oxide synthases and regulates numerous biological processes, including relaxation of blood vessels and neurotransmitter release in the brain. However, overproduc-tion of NO appears to contribute essentially to tissue injury in inflammatory and ischemic conditions (1). One of the mechanisms by which excess NO can injure tissues is by its nearly diffusion-controlled reaction with O 2 . to give peroxynitrite, a potent oxidant thought to be a key mediator of NO-mediated tissue injury in atherosclerosis, congestive heart failure, glutamate excitotoxicity, and other disease states involving inflammatory oxidative stress (2). There are several pieces of evidence implicating peroxynitrite as toxic agent in these pathologies as follows. (i) All of these diseases are associated with increased expression of inducible NO synthase, resulting in sustained formation of NO over relatively long periods of time, (ii) oxidative stress causes increased generation of O 2 . , (iii) authentic peroxynitrite triggers tyrosine nitration of a wide variety of proteins known to subserve important cellular functions that are lost upon nitration, and (iv) 3-nitrotyrosine levels have been observed in the injured tissues by both immunohistochemical techniques and quantitative analyses with HPLC 1 or gas chromatography-mass spectrometry (3). Despite this apparently conclusive link between oxidative tissue injury, peroxynitrite, and tyrosine nitration, direct evidence for peroxynitrite-mediated nitration in vivo is still lacking (4,5). This is of particular relevance because recent in vitro studies suggest that co-generation of NO and O 2 . , an obviously better approximation to the in vivo situation than bolus addition of concentrated peroxynitrite solutions, does not cause significant nitration of free tyrosine (6 -10). Although all of those studies, performed in four independent laboratories with a number of different NO/O 2 . -generating systems including pulse radiolysis, gave essentially identical results, Sawa et al. (11) recently reported on highly efficient tyrosine nitration by low fluxes of NO/O 2 . . The reason for this discrepancy is unclear.
The striking difference between peroxynitrite generated in situ at relatively low fluxes and bolus addition of authentic peroxynitrite appears to be a consequence of the different steady-state concentrations that are achieved with the two experimental protocols; at low (submicromolar) steady-state concentrations, the reaction of peroxynitrite with tyrosine was found to give almost exclusively dityrosine, i.e. the product of tyrosyl radical dimerization, whereas 3-nitrotyrosine is the major product at the fairly high concentrations of peroxynitrite * This work was supported by the Fonds zur Förderung der Wissenschaftlichen Forschung in Austria. 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.
‡ Recipient of an Austrian Academy of Sciences APART fellowship (APART 7/98).
§ To whom correspondence should be addressed. that occur upon bolus addition of the authentic compound (8).
Recent studies have revealed an alternative mechanism of tyrosine nitration with potential in vivo relevance (4). Heme peroxidases such as myeloperoxidase or eosinophil peroxidase have been shown to utilize H 2 O 2 to oxidize nitrite to a reactive nitrogen oxide species that triggers nitration of protein tyrosine residues and other phenolic compounds (12)(13)(14). Since inflammatory processes are typically associated with an infiltration of phagocytes, which contain high levels of heme peroxidases, this pathway has to be considered as a possible alternative to peroxynitrite in mediating protein tyrosine nitration in vivo. Intriguingly, the dependence of peroxidase-catalyzed nitration on the local levels of NO 2 Ϫ and H 2 O 2 suggests that the peroxidase pathway operates under exactly the conditions that favor formation of peroxynitrite, i.e. increased formation of both NO and O 2 . , the reactive precursors of NO 2 Ϫ and H 2 O 2 , respectively.
Thus, the experimental evidence currently available does not allow a decision as to which of the two pathways is responsible for tyrosine nitration in vivo. Although several in vitro studies with NO/O 2 . -generating systems (see above) argue against peroxynitrite as a mediator of nitration, it should be emphasized that those studies were performed with highly artificial in vitro systems not necessarily reflecting the in vivo situation. As a first approach in addressing this issue, we attempted to clarify the cellular pathways mediating protein tyrosine nitration in cultured macrophages activated with established immunological stimuli. As a model system we used the murine macrophage RAW 264.7 cell line. These cells are known to express high levels of inducible NO synthase and 3-nitrotyrosine-like immunoreactivity in response to immunological challenge with IFN-␥ in combination with LPS or zymosan (15). Upon cell activation, we measured several key parameters of the NO/O 2 . /peroxynitrite pathway as a function of time and compared the data with protein tyrosine nitration. These experiments revealed a striking discrepancy in the time course of NO formation and nitration and showed that peroxidase inhibitors as well as catalase attenuated the formation of 3-nitrotyrosine, whereas peroxynitrite scavengers had no significant effects. Based on these results we suggest that tyrosine nitration in cytokine-activated macrophages is mediated by a peroxidase/nitrite pathway rather than NO/O 2 . -derived peroxynitrite.

EXPERIMENTAL PROCEDURES
Materials-DHR and 3-nitrotyrosine were from Fluka (Vienna, Austria). Recombinant mouse IFN-␥ and pronase were from Roche Molecular Biochemicals (Vienna, Austria). MnTBAP was from Alexis (Vienna, Austria). Rabbit anti-human myeloperoxidase antibody was from DAKO (Vienna, Austria). Human myeloperoxidase was from Planta Naturstoffe (Vienna, Austria). The 3-nitrotyrosine antibody (clone 1A6, mouse monoclonal IgG, 100 g/100 l) was from Upstate Biotechnology (Lake Placid, NY). Penicillin, amphotericin, and fetal calf serum were from PAA Laboratories GmbH (Linz, Austria). The ECL Western blotting detection system was obtained from Amersham Pharmacia Biotech. Centrifuge tube filters (0.22 m cellulose acetate) were from Szabo (Vienna, Austria). Lipopolysaccharide was from Salmonella typhosa; bovine erythrocytes SOD, horse heart cytochrome c (type VI), and all other chemicals were from Sigma. PEG-Cat and PEG-SOD were prepared according to Beckman et al. (16).
Buffers and Solutions-All solutions were prepared freshly each day. Water was from a Milli-Q reagent water system from Millipore (Vienna, Austria; resistance Ն 18 megaohms ϫ cm Ϫ1 ). DHR was dissolved in acetonitrile to 10  Culture and Activation of Macrophages-RAW 264.7 macrophages were cultured in Petri dishes (diameter, 90 mm) at 37°C and 5% CO 2 in Dulbecco's modified Eagle's medium supplemented with 10% (v/v) heat-inactivated fetal calf serum, penicillin (100 units/ml), amphotericin (1.25 g/ml), and NaHCO 3 (3.7 g/l) as described (17). Cells were grown to confluence (ϳ5 ϫ 10 7 cells/dish) and incubated for up to 48 h in the presence of IFN-␥ (50 units/ml) and either LPS (0.5 g/ml) or zymosan A (0.5 mg/ml) in fresh phenol red-free Dulbecco's modified Eagle's medium. At the time points indicated in the text and graphs, the activated cells were assayed for the following parameters: nitrite accumulation in the culture medium, release of NO, O 2 . , and H 2 O 2 , DHR oxidation, and intracellular levels of protein-bound 3-nitrotyrosine.
Determination of Nitrite Accumulation-The concentration of nitrite in the cell culture supernatants was determined photometrically with the Griess assay as described previously (18).
Determination of NO Release-The culture medium was removed, and the cells (one Petri dish for each measurement) were washed with PBS, harvested, centrifuged, and resuspended in 0.5 ml of PBS. NO release was continuously monitored with a Clark-type NO-sensitive electrode (Iso-NO, World Precision Instruments, Berlin, Germany) at 37°C in disposable tubes (19). After 1 min, 5 l of a 0.1 M solution of L-arginine (final concentration, 1 mM) was injected. NO formation was quantified from the initial release rates obtained after injection of L-arginine using the Macintosh CHART software.
Determination of O 2 . Release-Rates of O 2 . release were measured as PEG-SOD-inhibitable reduction of acetylated cytochrome c as described (20). At the indicated time points, the cells were washed three times and equilibrated for 30 min in PBS followed by incubation with 10 M acetylated cytochrome c for 45 min with and without 150 units of PEG-SOD/ml. The cell supernatants were centrifuged at 1,300 ϫ g for 3 min followed by the determination of the absorbance at 550 nm against PEG-SOD-containing blanks. PEG-SOD inhibited total cytochrome c reduction by ϳ80%. O 2 . release was calculated using an extinction coefficient of 27,700 M Ϫ1 ϫ cm Ϫ1 at 550 nm (21).
Determination of H 2 O 2 Formation-Formation of H 2 O 2 was measured as horseradish peroxidase-catalyzed oxidation of fluorescent scopoletin as described (22). At the indicated time points, macrophages were washed three times with PBS and incubated for 45 min with an assay mixture containing 30 M scopoletin, 1 mM NaN 3 , and 10 units/ml horseradish peroxidase in Krebs-Ringer phosphate buffer. Supernatants were centrifuged at 1,300 ϫ g for 3 min followed by the determination of the fluorescence at excitation and emission wavelengths of 305 and 470 nm, respectively. The fluorescence of the assay mixture without cells was subtracted as the blank. The method was calibrated with standard solutions of H 2 O 2 adjusted photometrically using an extinction coefficient of 40 M Ϫ1 ϫ cm Ϫ1 at 240 nm.
Determination of DHR Oxidation-Oxidation of DHR was determined as a measure of peroxynitrite formation (23) . At the indicated time points, the cells were washed three times with PBS and incubated for 45 min in PBS containing 0.1 mM DHR and 0.1 mM of the metal chelator diethylenetriaminepentaacetic acid. The cell supernatants were centrifuged at 1,300 ϫ g for 3 min followed by determination of the absorbance at 500 nm against blank samples obtained by incubation of the assay mixture without cells. DHR oxidation was calculated using an extinction coefficient of 78,800 M Ϫ1 ϫ cm Ϫ1 at 500 nm (23).
Determination of Protein-bound 3-Nitrotyrosine-Protein-bound 3-nitrotyrosine was determined by HPLC with electrochemical detection after derivatization to N-AcATyr following a protocol described recently (24). The cells were homogenized in 0.1 M phosphate buffer, pH 7.4, and adjusted to a protein concentration of 16 -30 mg of protein/ml. Protein was determined with the Bradford method using bovine serum albumin as a standard (25) . Homogenates (0. 5 ml) were precipitated with 0.6 ml of HPLC grade acetonitrile, thoroughly vortexed and centrifuged (1000 ϫ g), followed by resuspension of the precipitates in 0.1 M phosphate buffer, pH 7.4, and sonication for ϳ10 s at 50 watts. This procedure was repeated three times to efficiently wash out non-protein material. The final suspensions were incubated overnight (16 -20 h) at 50°C with 1-2 mg of pronase and 0.5 mM CaCl 2 . Subsequently, 350-l aliquots of the samples were centrifuged (20,000 ϫ g), and an equal volume of 3 M phosphate buffer, pH 9.6, was added followed by the addition of 25 l of acetic anhydride. After 10 min of incubation at ambient temperature, ethyl acetate (1 ml) and formic acid (0.2 ml) were added. The samples were thoroughly vortexed for 30 s and then centrifuged at 20,000 ϫ g for 1 min. The ethyl acetate phase was concentrated to dryness under a gentle stream of N 2 at 50°C. For deacetylation of the phenolic acetate group, the samples were resuspended in 1 N NaOH (60 l). After 30 min of incubation at 37°C, 60 l of 1 M phosphate buffer, pH 6.5, was added followed by the addition of 0.1 M sodium dithionite (10 l) to reduce the nitro substituent to the corresponding amine. The samples were incubated for 10 min at ambient temperature, acidified by addition of concentrated hydrochloric acid (20 l), and centrifuged at 20,000 ϫ g for 10 min in centrifuge tube filters. Aliquots (100 l) were injected onto a 250 ϫ 4 mm C 18 reversed phase HPLC column (LiChrospher 100 RP-18, 5-m particle size, Merck) and eluted with 10 mM H 3 PO 4 at 0.7 ml/min. The performance of the column decreased gradually over time. This loss in resolution was overcome by supplementing the solvent with up to 2% (v/v) methanol. N-AcATyr was detected electrochemically with an ESA Coulochem II detector. The potentials of the two electrodes were set to Ϫ70 mV and ϩ70 mV (versus palladium), respectively. The method was calibrated daily with authentic N-AcATyr (5-500 nM) prepared as described (24). The recovery of authentic 3-nitrotyrosine added to homogenates of resting RAW 264.7 macrophages was 69.3 Ϯ 12.8%.
Immunostaining-To visualize 3-nitrotyrosine formation, RAW 264.7 macrophages were subjected to immunostaining with a monoclonal antibody. The cells were grown to confluence on L-polylysinetreated cover slides followed by activation with IFN-␥/Zy in phenol red-free Dulbecco's modified Eagle's medium for 24 h. After 14 h of activation the test compounds (methionine, 0.25 mM; MnTBAP, 50 M; KCN, 0, 25 mM; NaN 3 , 0, 25 mM, and PEG-Cat, 2000 units/ml) were added followed by incubation for a further 10 h. As a negative control, non-activated macrophages were incubated under identical conditions for 24 h. For positive control, the cells were treated with authentic peroxynitrite (1 mM) for 1 h. After incubation, the cover slides were gently rinsed three times with PBS and fixed for 1 h with a solution, pH 6.5, containing Na 2 HPO 4 (6.5 g/liter), NaH 2 PO 4 (4 g/liter), (v/v) 15 ml/liter methanol (15 ml/liter; v/v), and formaldehyde (100 ml of 37%/ liter, v/v). Thereafter, cover slides were gently rinsed three times with PBS. The 3-nitrotyrosine antibody was diluted to 10 g/ml in PBS containing 1% bovine serum albumin (w/v). 50 l of this solution were carefully applied to each cover slide to cover the entire surface and incubated for 1 h at 37°C under humidified atmosphere. After gentle rinsing of the cover slides three times with PBS, 50 l of biotinylated goat anti-mouse IgG (part of the mouse ExtrAvidin peroxidase staining kit obtained from Sigma diluted 1/20 in PBS containing 1% bovine serum albumin) were applied on the cover slides and incubated for 30 min at 37°C under a humidified atmosphere. Then cover slides were again gently rinsed three times with PBS followed by the application of 100 l of ExtrAvidin peroxidase (10 g/ml in PBS) on each cover slide and incubation for 30 min at 37°C under humidified atmosphere and rinsing of the slides with PBS. The staining solution was prepared by mixing 0.2 ml of 20 mg of 3-amino-9-ethylcarbazole in 2.5 ml of dimethylformamide with 3.8 ml of 0.05 M acetate buffer, pH 5.0. Before use, 20 l of 3% (v/v) H 2 O 2 were added to the staining solution, and 100 l were applied on the cover slides until the appropriate color development (3-4 min). Reactions were terminated by rinsing the slides gently with distilled water.
SDS-Polyacrylamide Gel Electrophoresis and Immunoblotting-Cell homogenates were subjected to SDS-polyacrylamide gel electrophoresis on 12% slab gels (26) and transferred onto nitrocellulose membranes in 25 mM Tris/HCl, pH 8.3, containing 192 mM glycine, 0.02% (w/v) SDS, and 20% (v/v) methanol at 250 mA for 90 min. Unspecific binding sites were saturated by overnight incubation of the membranes at 4°C in TBST containing 3% (w/v) ovalbumin. Subsequently, the membranes were washed twice for 5 min followed by incubation for 2 h with the anti-myeloperoxidase antibody diluted 1:500 in TBST containing 0.3% (w/v) ovalbumin. Subsequently, the membranes were washed twice for 15 min with TBST and incubated for 1 h with horseradish peroxidaselabeled anti-rabbit-IgG antibody that had been diluted 1:5000 in TBST buffer containing 0.3% (w/v) ovalbumin. Finally, the membranes were washed three times for 20 min with TBST buffer and processed with the ECL Western-blotting detection system according to the recommendations of Amersham Pharmacia Biotech.
Data Analysis and Statistics-Unless otherwise indicated, release rates are expressed as amounts of product (pmol or pg)/min/mg of total cell protein. Results are the mean values Ϯ S.E. of n experiments as indicated in the figure legends. The statistical significance of the data shown in Fig. 6 was evaluated by analysis of variance using Fisher's protected least significant difference test.

Time Course of NO and NO 2
Ϫ Formation by Activated Macrophages-Activation of RAW 264.7 macrophages with either IFN-␥/LPS or IFN-␥/Zy led to a pronounced release of NO accompanied by an accumulation of nitrite in the cell culture media. As shown in Fig. 1A, the maximal rates of NO release were 116.2 Ϯ 15.0 and 90.9 Ϯ 11.5 pmol ϫ min Ϫ1 ϫ mg Ϫ1 at 7 and 9 h after stimulation with IFN-␥/LPS and IFN-␥/Zy, re-spectively. Note that with both stimuli, NO release was virtually back to base line after 14 h of incubation. The inset in Fig.  1A shows that NO release was markedly increased upon the addition of L-arginine, an observation that agrees well with previous studies reporting on a pronounced dependence of macrophage NO synthesis on extracellular substrate supply (27,28). NO release was not significantly affected by the addition of SOD (1000 units/ml). The signal rapidly declined to zero upon the addition of the NO scavenger hemoglobin, demonstrating the specificity of the Clark-type NO electrode.
The time course of nitrite accumulation in the cell culture supernatant was virtually identical with both combinations of stimuli (Fig. 1B). Nitrite levels progressively increased from 4 to 15 h of incubation followed by a plateau corresponding to nitrite concentrations of about 50 M. Conversion of the rates of NO release from macrophages activated with either cytokine combination (Fig. 1A) to accumulating concentrations revealed that the decrease in the rates of NO release is in good accordance with the observed reduction in the rate of nitrite accumulation; based on the nitrite data, the apparent recovery of NO detected with the Clark electrode was ϳ50% (not shown). These results indicate that macrophage NO synthesis ceased after about 15 h of cell activation, presumably due to inducible NO synthase inactivation and/or limiting cofactor supply. Interestingly, the small but significant rightward shift of the time course of NO release from macrophages activated with IFN-␥/Zyas compared with that from IFN-␥/LPS-stimulated cells was not paralleled by a significant difference in the time course of nitrite accumulation, suggesting that specific intracellular pathways may affect net NO release from activated macrophages. protocols, but the overall fluxes were ϳ1000-fold higher (note the different scales in the two y axes of Fig. 2B).
We considered the possibility that the apparent decrease in O 2 . formation was a consequence of a rapid reaction of O 2 . with NO to yield peroxynitrite and carried out two sets of experiments to test this hypothesis. First, we repeated the experiments shown in Fig. 2A using cells treated with a high concentration of a non-selective NO synthase inhibitor (L-NNA; 1 mM). L-NNA almost completely inhibited nitrite accumulation in the cell culture supernatant (data not shown) but had no effect on the rates of O 2 . release measured 7 h after cell stimulation (inset to Fig. 2A). Secondly, we determined the time course of DHR oxidation as a measure for peroxynitrite formation. Neither of the two protocols of macrophage activation (IFN-␥/LPS and IFN-␥/Zy) resulted in a considerable increase in the rates of DHR oxidation (1-3 pmol ϫ min Ϫ1 ϫ mg Ϫ1 ), which was insensitive to L-NNA (data not shown). Together, these results argue against peroxynitrite as a major reactive nitrogen species formed by activated macrophages.
Evidence against the Involvement of Peroxynitrite in Tyrosine Nitration-Protein-bound 3-nitrotyrosine was measured in the cell extracts as the N-acetyl-amino derivative (N-AcATyr). As expected, treatment of macrophages with authentic peroxynitrite (1 mM final) resulted in a pronounced increase in tyrosine nitration from 19.4 Ϯ 17.3 to 855.9 Ϯ 270.2 pg of N-AcATyr/mg of cellular protein (n ϭ 3 each). Fig. 3 shows that a significant increase in nitration was also observed upon activation of the macrophages with either IFN-␥/Zy or IFN-␥/LPS. The time course of N-AcATyr formation was similar with both combinations of stimuli, although IFN-␥/Zy led to about a 3-fold higher product formation than IFN-␥/LPS (385.3 Ϯ 77.8 and 127.9 Ϯ 8.7 pg ϫ mg Ϫ1 , respectively). Nitration occurred with a pronounced lag phase of 6 (IFN-␥/Zy) to 18 h (IFN-␥/LPS), was maximal 24 h post-stimulation, and slowly declined during the next 24 h.
Thus, we observed a pronounced difference in the time course of protein tyrosine nitration and NO/O 2 . formation such that nitration started to increase at a time when the rates of NO/O 2 . had already declined close to basal levels. These results argue against peroxynitrite as a mediator of tyrosine nitration in activated macrophages. However, because of the apparent lack of peroxynitrite formation, as evident from the lack of significant DHR oxidation by activated cells, the data do not exclude that peroxynitrite, if produced, is capable of nitrating cellular proteins. To clarify this issue, we treated IFN-␥/LPSactivated macrophages with PMA, which is known to trigger O 2 . formation through activation of protein kinase C (29). This protocol was expected to result in an intracellular co-generation of both NO (through induction of NO synthase) and O 2 .
(through activation of protein kinase C). Fig. 4A shows that PMA (0.5 M) led to ϳ2-fold increase in the rates of DHR oxidation from 2.6 Ϯ 0.1 to 5.4 Ϯ 0.2 pmol ϫ min Ϫ1 ϫ mg Ϫ1 after 30 min. This effect of PMA was sensitive to inhibition of NO synthase and protein kinase C with L-NNA (1 mM) and H-7 (30 M), respectively (Fig. 4B). These data indicated that coactivation of macrophages with IFN-␥/LPS and PMA resulted in formation of peroxynitrite, detectable as increased rates of DHR oxidation. However, as shown in Fig. 4C, this apparent peroxynitrite formation was not accompanied by an increase in the yields of protein nitration, measured 3.5 h after the addition of PMA, i.e. 9 h after IFN-␥/LPS. Note that the lack of effect of IFN-␥/LPS on nitration after 9 h of incubation is in accordance with the results shown in Fig. 3.
Evidence for the Involvement of a Nitrite/Peroxidase Path- way in Tyrosine Nitration-So far, the results indicated that macrophages activated with typical NO synthase inducers did not produce significant amounts of peroxynitrite unless the NADPH oxidase pathway was additionally stimulated to generate O 2 . . The finding that peroxynitrite even when produced did not trigger significant nitration extended our observations made previously with cell-free systems and free tyrosine (7,8) to protein nitration in intact cells. Based on these data, we speculated that a nitrite/peroxidase pathway (5) rather than peroxynitrite may be responsible for protein tyrosine nitration in activated macrophages. The known peroxidase pathways of tyrosine nitration do all utilize nitrite as a substrate (5). Thus, assuming the involvement of such a pathway in nitration would imply that induction of macrophage NO synthase with IFN-␥/Zy or IFN-␥/LPS is mimicked by treatment of the cells with nitrite. To address this issue, we measured N-AcATyr levels in non-activated macrophages treated for 5 h with increasing concentrations of nitrite. As shown in Fig. 5A, incubation with 20 -100 M nitrite led to a 3-4-fold increase in nitration. The amount of N-AcATyr in the nitrite-treated cells (100 -150 pg ϫ mg Ϫ1 ) was clearly less than in IFN-␥/Zy-activated cells but identical to that measured 24 h after activation of the macrophages with IFN-␥/LPS (see Fig. 3). A time course of nitration in response to 20 M nitrite is shown in Fig. 5B.
In another set of experiments we studied the effects of peroxynitrite/O 2 . scavengers, peroxidase inhibitors, and catalase. To minimize nonspecific and/or cytotoxic effects such as the described inhibition of NO formation by manganese porphyrins and catalase (30,31), the compounds were added 14 -15 h after IFN-␥/Zy followed by 9 -10 h of incubation (24 h total) and subsequent determination of protein-bound 3-nitrotyrosine. To account for this protocol, the results shown in Fig.  6 are expressed as the rates of N-AcATyr formation (pg ϫ h Ϫ1 ϫ mg Ϫ1 ) during the 9 -10 h of treatment (i.e. from 14 -15 to 24 h after the addition of IFN-␥/Zy). Under control conditions, this rate was 15.1 Ϯ 2.38 pg ϫ h Ϫ1 ϫ mg Ϫ1 . NaN 3 and KCN (0. 25 mM each), the most commonly used inhibitors of all types of heme peroxidases (32), reduced the rate of N-AcATyr formation to 2.90 Ϯ 2.50 and 7.03 Ϯ 4.68 pg ϫ h Ϫ1 ϫ mg Ϫ1 , respectively. The effect of NaN 3 was statistically significant (p Ͻ 0.05). Peroxidase-catalyzed nitration was reported to be strictly dependent on H 2 O 2 (12). Indeed, tyrosine nitration by activated macrophages was completely blocked by incubation of the activated cells with PEG-Cat (Ϫ1.23 Ϯ 1.99 pg ϫ h Ϫ1 ϫ mg Ϫ1 ). In contrast, tyrosine nitration was not significantly affected by the peroxynitrite/O 2 . scavengers methionine (0.25 mM) and MnTBAP ( were corroborated by immunostaining of the cells with a monoclonal 3-nitrotyrosine antibody. As shown in Fig. 7 . or peroxynitrite with MnTBAP or methionine, respectively (panels C and D), was much less effective. The results confirm the quantitative data shown in Fig. 6, suggesting that a peroxidase/nitrite pathway rather than peroxynitrite was responsible for tyrosine nitration in our experiments. Among several heme peroxidases, myeloperoxidase appeared to be a likely candidate catalyzing nitrite/H 2 O 2 -dependent tyrosine nitration, but it is unclear whether this enzyme occurs in RAW 264.7 macrophages (35,36). We attempted to clarify the possible involvement of myeloperoxidase in protein nitration using the fairly selective myeloperoxidase inhibitor 4-aminobenzoic acid hydrazide (32). However, this drug reduced the apparent N-AcATyr levels even below basal levels (data not shown), presumably because of an interference with 3-nitrotyrosine derivatization and/or electrochemical detection of the N-AcATyr derivative. Moreover, several published myeloperoxidase activity assays yielded negative results. To allow a more sensitive detection of this enzyme, we used a selective myeloperoxidase antibody (37,38) for immunoblotting of macrophage homogenates. As illustrated by a representative blot shown in Fig. 8, the antiserum recognized a protein with an apparent molecular mass of ϳ57 kDa, which comigrated with the 57-kDa band of purified human myeloperoxidase (lane 4). Similar amounts of the protein were found in non-activated macrophages (lane 1) and in cells activated with IFN-␥/LPS (lane 2) or IFN-␥/Zy (lane 3). These results suggest that RAW 264.7 macrophages contain small amounts of myeloperoxidase that might contribute to tyrosine nitration upon cytokine activation of the L-arginine/NO pathway in these cells. DISCUSSION In the present study we investigated the mechanisms of protein tyrosine nitration in activated RAW 264.7 murine macrophages with a special focus on the potential involvement of peroxynitrite and heme peroxidases. In contrast to an earlier report (39), our data do not support the view that peroxynitrite is a major reactive nitrogen species formed by macrophages activated with either IFN-␥/Zy or IFN-␥/LPS. With both combinations of stimuli, we observed a pronounced release of NO and accumulation of nitrite in the cell culture supernatant. Although maximal rates of NO release were observed 6 -8 h after stimulation, release of O 2 . was maximal at much earlier time points, i.e. 1-3 h after stimulation, and then rapidly declined. It was conceivable that the decline in the rates of O 2 .  (1-3 versus 10 -25 pmol ϫ min Ϫ1 ϫ mg Ϫ1 ), indicating that peroxynitrite was not a major reactive nitrogen species released from activated macrophages. These results do not exclude that peroxynitrite was formed at low steady-state concentrations inside the cells and rapidly consumed by scavengers, e.g. GSH, ascorbate, or urate, but they argue against the common view that large amounts of peroxynitrite are released from activated macrophages to kill adjacent target cells. In fact, there are a few previous studies suggesting that the killing of pathogens and tumor cells by activated macrophages is mediated by NO rather than peroxynitrite (40 -42).
Our data on intracellular protein tyrosine nitration suggest that peroxynitrite, even if produced, does not contribute to nitration in activated macrophages. When cells were treated with either IFN-␥/Zy or IFN-␥/LPS, formation of protein-bound 3-nitrotyrosine was considerably delayed and became significant only when the production of NO/nitrite had virtually ceased, indicating that nitration was not dependent on the presence of active NO synthase and, thus, not mediated by peroxynitrite. This conclusion is further supported by our results obtained with the peroxynitrite scavengers methionine and MnTBAP, both of which had no considerable effects on nitration (cf. Figs. 6 and 7 (8). The present results suggest that this conclusion is not confined to test tube chemistry but also holds for the chemical reactivity of the NO/O 2 . /tyrosine system in intact cells. The time course of 3-nitrotyrosine formation indicated that the nitration reaction was dependent on the intracellular levels of nitrite, the stable oxidation product of NO. Indeed, treatment of non-stimulated macrophages with nitrite led to formation of similar amounts of 3-nitrotyrosine as activation of the cells with IFN-␥/LPS. The ϳ3-fold greater nitrating efficiency of IFN-␥/Zy, despite similar rates of NO formation, suggests that additional factors contribute to tyrosine nitration. Several analytical methods applied previously for the determination of 3-nitrotyrosine in cell and tissue extracts were shown to yield false positive results in the presence of nitrite (4,5). However, the fairly sophisticated method described by Shigenaga et al. (24) that we used in this study involves rigorous removal of nitrite and showed no increase in the N-AcATyr levels when cell homogenates were treated with 0.1 mM nitrite before sam-ple preparation (data not shown). The nitrite dependence of tyrosine nitration pointed to the involvement of a heme peroxidase as reported previously for nitration in neutrophils (13) and eosinophils (14). This assumption was confirmed by the complete inhibition of 3-nitrotyrosine formation upon removal of H 2 O 2 with PEG-CAT and the pronounced inhibitory effect of the heme-site peroxidase inhibitor NaN 3 . Together, these data strongly suggest that an as yet unidentified peroxidase catalyzes tyrosine nitration in cytokine-activated RAW 264.7 macrophages. Similar evidence against peroxynitrite as a mediator of tyrosine nitration was recently obtained with primary murine peritoneal macrophages activated in vitro with IFN-␥/ LPS as well as macrophages isolated from mice subjected to systemic inflammation by treatment with heat-inactivated Corynebacterium parvum. 2 The occurrence of heme peroxidases in macrophages is still a controversial issue. It has been shown that myeloperoxidase activity declines rapidly during differentiation of monocytes into macrophages (43), but peroxidase activity was detected in several types of resident tissue macrophages, such as rodent and human alveolar as well as peritoneal macrophages (44 -47). Unlike monocytes, which express peroxidase in the primary lysosomes (48), resident macrophages contain peroxidase activity in the rough endoplasmic reticulum and the perinuclear cisterna (48 -50). It has been suggested that macrophage peroxidase activity results from acquisition of exogenous peroxidases by vesicular transport or phagocytosis of peroxidasepositive cells (51). However, myeloperoxidase mRNA has been isolated from thioglycollate-elicited mouse peritoneal macrophages, indicating that myeloperoxidase gene expression is inducible in macrophages by selected immunological stimuli (52). This is further indicated by a recent study showing that granulocyte macrophage colony-stimulating factor up-regulates expression of active myeloperoxidase in macrophages residing in atherosclerotic lesions (53). The macrophage cell line RAW 264.7 that we used in this study was established from murine peritoneal macrophages elicited by the intraperitoneal injection of Abelson leukemia virus. Thus, it is conceivable that myeloperoxidase is indeed expressed in this cell line, as suggested by our immunoblot analyses. So far, macrophage heme peroxidases have been only poorly characterized (54), and further work is required to clarify the role of these enzymes in macrophage-dependent nitration and cytotoxicity.