Nitration of Tyrosine 92 Mediates the Activation of Rat Microsomal Glutathione S-Transferase by Peroxynitrite*

There is increasing evidence that protein function can be modified by nitration of tyrosine residue(s), a reaction catalyzed by proteins with peroxidase activity, or that occurs by interaction with peroxynitrite, a highly reactive oxidant formed by the reaction of nitric oxide with superoxide. Although there are numerous reports describing loss of function after treatment of proteins with peroxynitrite, we recently demonstrated that the microsomal glutathione S-transferase 1 is activated rather than inactivated by peroxynitrite and suggested that this could be attributed to nitration of tyrosine residues rather than to other effects of peroxynitrite. In this report, the nitrated tyrosine residues of peroxynitrite-treated microsomal glutathione S-transferase 1 were characterized by mass spectrometry and their functional significance determined. Of the seven tyrosine residues present in the protein, only those at positions 92 and 153 were nitrated after treatment with peroxynitrite. Three mutants (Y92F, Y153F, and Y92F, Y153F) were created using site-directed mutagenesis and expressed in LLC-PK1 cells. Treatment of the microsomal fractions of these cells with peroxynitrite resulted in an ∼2-fold increase in enzyme activity in cells expressing the wild type microsomal glutathione S-transferase 1 or the Y153F mutant, whereas the enzyme activity of Y92F and double site mutant was unaffected. These results indicate that activation of microsomal glutathione S-transferase 1 by peroxynitrite is mediated by nitration of tyrosine residue 92 and represents one of the few examples in which a gain in function has been associated with nitration of a specific tyrosine residue.

MGST1 exists as a homotrimer of identical 17.3-kDa subunits and has a high degree of amino acid sequence similarity (83%) to human MGST1. In contrast to the cytosolic GSTs, an important characteristic of microsomal GST1 is that the enzyme can be activated by various treatments, including limited proteolysis, heating, radiation, and exposure to sulfhydryl modifying reagents and reactive oxygen species. In addition, since MGST1 possesses glutathione peroxidase activity, it has been suggested that this enzyme may play a protective role under conditions of oxidative stress (2,3).
Activation of MGST1 also occurs after exposure to reactive nitrogen species (RNS), including nitric oxide (NO) donors and peroxynitrite (ONOO Ϫ ) (3,4), in contrast to the cytosolic GSTs, which are inhibited by RNS (5). The ONOO Ϫ -induced activation of MGST1 was associated with tyrosine nitration, polymer formation, and protein fragmentation but not protein S-oxidation (3). This increase in enzyme activity by ONOO Ϫ is unusual, since tyrosine nitration of proteins is almost always associated with a loss of function. For example, a number of enzymes involved in protection against oxidative stress are inhibited by ONOO Ϫ , including superoxide dismutases (6), glutathione reductase (7), cytosolic GSTs (5), catalase (8,9), and glutathione peroxidase (8,10). The activation of MGST1 by RNS, together with its GSH peroxidase activity, suggest that this enzyme may play an important role in limiting the extent of oxidative tissue injury under pathophysiological conditions of excessive ONOO Ϫ formation, especially if other antioxidant defense mechanisms are compromised.
The tyrosine residues of MGST1 that are nitrated by ONOO Ϫ have not been identified. In the present study, we used two mass spectrometric techniques, MALDI-TOF MS and ESI-MS/MS, to identify the tyrosine residues of purified rat hepatic MGST1 that are nitrated after ONOO Ϫ treatment. Subsequently, we determined the functional consequences resulting from nitration of these tyrosine residues using sitedirected mutagenesis.
Enzyme Purification and Enzyme Activity Assay-Rat hepatic MGST1 was purified from male Sprague-Dawley rats (250 -300 g) by hydroxyapatite and CM-Sepharose chromatography as described (3). The purity of the enzyme preparation was assessed by SDS-PAGE on 15% gels under reducing conditions. The protein migrated as a single band at about 17 kDa (4). Prior to experiments, GSH was removed from the preparation by dialysis against buffer containing 10 mM potassium phosphate, pH 7.0, 0.1 mM EDTA, 1% Triton X-100, and 20% glycerol using a System 500 Microdialyzer (Pierce). Enzyme activity was determined by the spectrophotometric method of Habig et al. (11). Samples (1.0 ml) contained 100 mM potassium phosphate, pH 6.5, 0.5% Triton X-100, 1.0 mM GSH, and 1.0 mM 1-chloro-2,4-dinitrobenzene at 25°C.
Preparation of Peroxynitrite (ONOO Ϫ ) and Treatment of MGST1 -ONOO Ϫ was synthesized from acidified nitrite and H 2 O 2 as described (12). Prior to each experiment, excess H 2 O 2 in ONOO Ϫ solutions was removed by MnO 2 chromatography (3). Purified enzyme (20 g/ml) or microsomal fractions (200 -400 g/ml protein) in 100 mM potassium phosphate, pH 7.0, containing 100 M diethylenetriaminepentaacetic acid (DTPA), were exposed to the indicated concentrations of ONOO Ϫ at room temperature for 10 s. To keep the pH value of the reaction unchanged, ONOO Ϫ was added to the protein or microsomal solution as a small volume during vigorous mixing, and the reaction was terminated by dilution of the sample into the assay buffer for the determination of enzyme activity or into the sample buffer for SDS-PAGE. For immunoblot analysis, samples treated with ONOO Ϫ were resolved on a 15% SDS-PAGE gel under non-reducing conditions. Proteins were transferred to polyvinylidene difluoride membranes and incubated with a 3-nitrotyrosine-specific antibody. The immunoreactive protein bands were visualized by enhanced chemiluminescence.
Identification of Nitration Sites by MALDI-TOF MS-Protein bands corresponding to the MGST1 monomer (17.3 kDa) with and without ONOO Ϫ treatment were excised from Coomassie Blue-stained, 15% reducing SDS-PAGE gels (approximately 2 g of purified protein/ band). Gel slices were washed three times with 50% ACN, 25 mM ammonium bicarbonate, pH 8.0 (15 min each time under gentle agitation) to remove the Coomassie Blue and were then incubated with 100% ACN for 5 min and vacuum-dried. To reduce and alkylate proteins, dried gel slices were incubated with 10 mM DTT in 100 mM ammonium bicarbonate at 55°C for 45 min followed by incubation with 55 mM iodoacetamide in 100 mM ammonium bicarbonate for 30 min at room temperature in the dark. Gel slices were then washed with 25 mM ammonium bicarbonate and dried with 100% ACN as described above. In-gel digestion was performed by swelling the dried gel slices with 10 g/ml trypsin in 25 mM ammonium bicarbonate, pH 8.0, for 30 min on ice. Excess trypsin solution was removed and digestion continued at 37°C for 16 h. Tryptic peptides were extracted with 50% ACN/5% trifluoroacetic acid, dried under vacuum, and reconstituted in 3 l of 70% ACN, 0.1% trifluoroacetic acid. Reconstituted extract (0.5 l) was mixed with 0.5 l of matrix (10 mg/ml ␣-cyano-4-hydroxy-trans-cinnamic acid in 70% ACN, 0.1% trifluoroacetic acid), spotted onto a stainless steel 100-well MS plate, and air dried. Samples were analyzed using a Voyager DE-Pro MALDI-TOF mass spectrometer (Applied Biosystems, Foster City, CA) operated in the delayed extraction/reflector mode with an accelerating voltage of 20 kV, grid voltage setting of 72%, and a 50-ns delay. Three spectra (100 laser shots/spectrum) were obtained for each sample. External calibration was performed using a Sequazyme Peptide Mass Standard kit (Perseptive Biosystems, Framingham, MA) containing the following standards: des-Arg-bradykinin, angiotensin-1, and Glu-fibrinopeptide B. Peptide mass fingerprinting was conducted with the data base search tool MS-Fit using Protein Prospector (version 3.2.1), available at prospector.ucsf.edu/. The modifications due to nitration were included in the search for both control and ONOO Ϫ -treated samples. The criteria for identification of the protein by MALDI-TOF MS were: a match of at least four peptides from the protein compared with the data base of rat proteins with similar to observed molecular mass and with an error of less than 50 ppm for all peptides, including modified ones.

Analysis of Trypsin Cleavage Products by ESI-MS/MS-Samples
were analyzed with a ESI/MS/MS system composed of Waters Capillary LC (Waters Corp., Milford, MA,) coupled to a Finnigan LCQ Deca iontrap mass spectrometer (ThermoFinnigan, San Jose, CA) equipped with a nanospray ionization source. Separation was performed using a picofrit column (PFC7515-PP18 -5-3PK), New Objective, Woburn, MA) with a 75 m inner diameter and a 15 m tip. The stationary phase was 5 m ProteoPep C18 matrix packed in a column of 5 cm length. The mobile phase was ACN, 0.1% trifluoroacetic acid. Separation of peptides was achieved using a gradient of ACN from 5 to 65% in 45 min, followed by a wash with 95% ACN. MS/MS spectra were analyzed using SEQUEST (ThermoFinnigan) software. For the ESI-LC MS/MS spectra, a correlation score above 2 was considered significant for peptides with a charge of 2 to accept the deduced sequence, whereas the sequence of peptides with a lower score was deduced manually.
Site-directed Mutagenesis-Plasmid (pCMV5) containing rat liver MGST1 cDNA was a generous gift from Dr. Ralf Morgenstern (Institute of Environmental Medicine, Karolinska Institute, Stockholm, Sweden). The plasmid was digested with EcoRI (Promega), and the MGST1 cDNA was inserted into pcDNA3.1(Ϫ). All mutations were generated using the QuikChange TM site-directed mutagenesis kit (Stratagene, La Jolla, CA). pcDNA3.1-MGST1 was used as template for generating mutants Y92F and Y153F, and subsequently pcDNA3.1-MGST (Y153F) was used as the template for generation of double mutants (Y92F and Y153F). Oligonucleotides bearing mismatched bases at the residues to be mutated (underlined) were synthesized by ACGT Corp. (Toronto, Ontario, Canada) as follows: Y92F (5Ј-C GGC CTC CTG TAC TCC CTG AGC GG-3Ј) and Y153F (5Ј-G CTC AGG AGC AGA CTG TAC TTG TAA AGA GAT TTG TGA CC-3Ј). All mutations were confirmed by DNA sequencing.
Cell Culture and Gene Transfection-LLC-PK1 cells were obtained from the American Type Culture Collection (Manassas, VA) and grown in Dulbecco's modified Eagle's medium/F-12 medium (1:1) supplemented with 10% fetal bovine serum, 5 g/ml insulin, 2 mM glutamine, 10 mM HEPES, pH 6.9, 50 units/ml penicillin, and 50 g/ml streptomycin. For transient transfections, LLC-PK1 cells were seeded in 150-mm cell culture plates, and the expression plasmids were introduced using FuGENE 6 (Roche Applied Science), according to the manufacturer's instructions. After 24-h incubation with the transfection reagents, cells were harvested, lysed by sonication in 100 mM potassium phosphate, pH 7.4, and microsomal fractions (100,000 ϫ g pellet) prepared by ultracentrifugation. MGST1 activity in both ONOO Ϫ -treated and non-treated microsomal fractions was assessed as described above. MGST1 activity determinations were performed in triplicate, and data are presented as the mean Ϯ S.D. Data were analyzed by the appropriate statistical test, as indicated. p values Ͻ0.05 were considered statistically significant.

RESULTS
Tyrosine Nitration and Activation of MGST1 by ONOO Ϫ -Exposure of purified MGST1 to 2.0 mM ONOO Ϫ resulted in a 4.3-fold increase in enzyme activity (p Ͻ 0.001) (Fig. 1). However, decayed ONOO Ϫ or 2.0 mM H 2 O 2 did not affect enzyme activity (Fig. 1). The presence of 3-nitrotyrosine in the activated protein was demonstrated using SDS-PAGE (under non-reducing conditions) and immunoblot analysis employing a monoclonal anti-nitrotyrosine antibody. Immunoreactive staining of the 17.3-kDa MGST1 monomer increased in intensity after exposure to increasing concentration of ONOO Ϫ (Fig. 2). In addition, there was staining evident in proteins bands of about 34 and 50 kDa, as observed previously (3), indicating the formation of tyrosine-nitrated dimers and trimers.
Identification of Sites of Tyrosine Nitration in MGST1-Peroxynitrite-treated and non-treated MGST1 were separated by 15% SDS-PAGE, and the monomeric MGST1 (17.3 kDa) underwent in-gel digestion with trypsin. Peptide mass fingerprinting by MALDI-TOF MS confirmed the identification of the 17.3-kDa band as rat MGST1 (SwissProt accession number P08011, M r 17,472, pI 9.6) (ϳ57% protein coverage for both control (non-treated) and ONOO Ϫ -treated samples). There are 7 tyrosine residues in MGST1 (Table 1). Tryptic peptides containing these tyrosine residues were observed by either MALDI-TOF MS and/or ESI-MS/MS with the exception of the tryptic peptide containing Tyr 18 . Most likely this was not observed due to poor ionization of the peptide fragment containing Tyr 18 . Based on the sequence of MGST1, digestion using either Asp-N or Glu-C would not have been predicted to produce shorter peptides containing Tyr 18 . Digestion with thermolysine produced a large number of very short peptides not suitable for identification by MALDI-TOF MS.
The tryptic peptide containing Tyr 92 (a.a. 74 -107) was observed in both untreated and ONOO Ϫ -treated MGST1 by ESI-MS/MS, whereas the nitrated form of the Tyr 92 peptide was detected only in the ONOO Ϫ -treated sample. Tyrosines 115 and 120 present within the same peptide (1897.03 m/z (aa 114 -129) were observed using both types of MS, but there was no evidence for nitration of either of these tyrosine residues in ONOO Ϫ -treated samples. Similarly, Tyr 137 and Tyr 145 lie within the tryptic peptide comprising residues 131-147 (1851.95 m/z). Nitration of these two tyrosine residues also was not detected. However, in all cases, the methionine residue at position 143 was always found in the oxidized form, and this modification could have prevented the identification of other potential modifications within the same peptide. The tyrosine residue located near the C terminus, Tyr 153 , was found in a short 3-amino acid peptide (aa 152-154) by ESI-MS/MS or as a peptide that resulted from a single missed cleavage comprising the penultimate 5 amino acids (aa 150 -154) by MALDI-TOF MS. Nitration at Tyr 153 in ONOO Ϫ -treated samples was observed in both peptides. The MS data are summarized in Table 1. The Sequest summary for parental and nitrated peptides observed in ONOO Ϫ -treated samples and a schematic representation of the MS/MS sequence for peptides containing nitrated Tyr 153 are shown in Fig. 3.
Cellular Functional Analysis-The functional consequences of nitration of tyrosine residues Tyr 92 and Tyr 153 of MGST1 were determined by comparing the wild type protein to several MGST1 mutants, in which these tyrosine residues were substituted with phenylalanine, alone or in combination (Y92F, Y153F, and a Y92F,Y153F double mutant). These mutations would prevent any potential nitration at the mutated amino acid site while maintaining similar overall chemistry. Expression vectors containing the wild type or mutated MGST1 (both single and double sites) were transiently transfected in LLC-PK1 cells and the MGST1 activity in the microsomal fractions assessed. Immunoreactive MGST1 protein was readily detected in transfected cells but not in non-transfected cells or cells transfected with empty vector (Fig.  4A), consistent with a 2-3-fold increase in basal MGST1 activity in MGST1-transfected cells. (Fig. 4B). As expected, the basal MGST1 activity was similar in microsomes containing either the wild type or mutated proteins. However, after treatment with 2.0 mM ONOO Ϫ for 10 s, there was a significant increase in MGST1 activity only in microsomal fractions from cells transfected with wild type MGST1 or with the Y153F mutant (Fig. 4B). In contrast, MGST1 activity in the Y92F mutant or the double mutant was unaltered by ONOO Ϫ , suggesting that nitration of Tyr 92 is required for the peroxynitrite-induced activation of MGST1.

TABLE 1 Result of mass spectrometric characterization of ONOO ؊ -treated versus non-treated MGST1
Peptides containing tyrosines where either not detected (ND) or detected in control only (m/z/Ϫ) or in both control and ONOO Ϫ -treated samples (m/z/m/z). 137, 145, and 153) in addition to a single cysteine residue (Cys 49 ). Six of the seven tyrosine residues are located in the C-terminal half of the protein. Electron crystallography of two-dimensional crystals of MGST1 has been used to construct a three-dimensional model of the enzyme. In this model, the trimeric enzyme exhibits three repeats of a left-handed four-helix bundle motif (13). The catalytic center of the enzyme is thought to be contained in the first cytosolic loop of the protein between peripheral transmembrane helices designated D and A. Only two of the tyrosine residues (Tyr 120 and Tyr 153 ) are located in putative non-membrane domains, the former in the second cytosolic loop and the latter at the intra-lumen C-terminal; the other five are located in one of the four putative transmembrane domains (13). The first cytosolic loop does not contain any tyrosine residues (13) but does contain Cys 49 , the alkylation or oxidative modification of which results in increased catalytic activity and is thought to act as a sensor of chemical or oxidative stress (14). However, oxidation of Cys 49 by reactive oxygen species does not appear to cause enzyme activation, since we (Ref. 3 and Fig. 1) and others (15) (16,17), incubations with H 2 O 2 were performed in the presence of GSH, and the increase in enzyme activity was attributed to S-glutathiolation or polymer formation, rather than oxidation of the Cys 49 per se. In our studies, incubations with H 2 O 2 were performed in the absence of GSH, and therefore S-glutathiolation of Cys 49 and subsequent activation of the enzyme would not be expected.

MALDI-TOF MS fragment mass
To identify sites of tyrosine nitration of MGST1, we utilized mass spectrometric techniques using both MALDI and ESI to maximize cov- Cells were harvested, lysed, and 5 g of microsomal protein resolved on 15% SDS-PAGE gels. MGST1 was detected using a polyclonal antibody raised against purified rat MGST1. WT, non-transfected cells; Vec, cells transfected with empty vector. B, effect of ONOO Ϫ on microsomal GST activity in MGST1 wild type and MGST1 mutants expressed in LLC-PK1 cells. Microsomal fractions (200 -400 g/ml protein) isolated from cells transiently transfected with the indicated MGST1-contaning plasmids were exposed to 2.0 mM ONOO Ϫ in 100 mM potassium phosphate, pH 7.0, containing 100 M DTPA, at room temperature for 10 s. GST activity assay of non-treated and ONOO Ϫ -treated microsomes was carried out as described under "Experimental Procedures." *, p Ͻ 0.05 compared with matched, non-treated microsomes (Student's t test for paired data). erage of the protein and allow observation of most of the tyrosine residues. Spectra of the tryptic digests of the MGST1 obtained using both techniques showed the presence of nitration of Tyr 153 in ONOO Ϫtreated but not in the untreated samples. A second nitration site at Tyr 92 was detected by ESI-MS/MS (sequence confirmation shown in Fig. 3,  upper panel). Modification of the other tyrosine residues, Tyr 115 , Tyr 120 , Tyr 137 , and Tyr 145 , was not detected. However, Tyr 18 was not detected by either MS instrument, and hence we could not assess its nitration status. It is noteworthy that the methionine residue present in the peptide that contains Tyr 137 and Tyr 145 (aa 130 -146) was always found in the oxidized form. Methionine was not present in peptides containing Tyr 92 , Tyr 115 , Tyr 120 , or Tyr 153 . Oxidation of methionine is a common modification that occurs prior to protein preparation as well as during protein purification and analysis. Methionine oxidation can cause suppression of ionization and can prevent or mask identification of any other modifications within the same peptide. Therefore, confirmation of the presence or absence of nitration at Tyr 137 and/or Tyr 145 would have to be investigated by other means, possibly by either chemical derivatization of the nitro group to enhance its ionization or by other non-MS techniques such as immunoassay with site specific antibodies.
Peroxynitrite is a strong biological oxidant that has been implicated in diverse forms of free radical-induced tissue injury. It can react with a variety of biological molecules to modify cellular structure and function, and there are now many reports in the literature describing the inhibitory effect of ONOO Ϫ on the enzymatic activity of a variety of proteins, including several enzymes involved in defense against oxidative stress. In most cases this inhibitory activity has been attributed to nitration of critical tyrosine residues, although in some, oxidation of critical sulfhydryl groups (18 -20) or disruption of active site zinc-thiolate clusters (21,22) has been proposed. In marked contrast, however, treatment with ONOO Ϫ increases MGST1 activity in both hepatic microsomes or in purified enzyme preparations (Ref. 3 and Fig. 1) and also increases the GSH peroxidase activity of the enzyme (3).
In a previous study, we found that MGST1 was much more susceptible to sulfhydryl oxidation than to tyrosine nitration (3). Exposure of MGST1 to 0.1 mM ONOO Ϫ resulted in 80% oxidation of sulfhydryl groups, whereas tyrosine nitration was barely detectable. However, the extent of Cys 49 oxidation by ONOO Ϫ did not correlate well with the increase in enzyme activity, suggesting that S-oxidation of MGST1 by ONOO Ϫ was not sufficient to cause the observed enzyme activation. In contrast, the dose-dependent increase in tyrosine nitration by ONOO Ϫ paralleled quite closely the increase in enzyme activity, suggesting that tyrosine nitration was the basis for the gain of function (3). The results of the site-directed mutagenesis studies reported herein provide direct evidence that this is the case, since in cells expressing the Y92F mutant, ONOO Ϫ was completely without effect.
The active site of MGST1 is localized on the cytosolic side of endoplasm reticulum (23). The MGST1 trimer binds one molecule of GSH, and subunit communication is thought to play a role in the activation of the enzyme (24,25). Thiolate formation follows a biphasic kinetic pattern, with rapid binding of GSH followed by slower formation of thiolate (26). The presumed conformational changes brought about by modification of Cys 49 by N-ethylmaleimide markedly increases the rate of thiolate formation and increases k cat . Although nitration of Tyr 153 was shown to occur, this residue is the penultimate C-terminal residue, far removed from the catalytic site of the enzyme, and thus might not be expected to influence enzyme conformation or activity. This is supported by the site-directed mutagenesis data (Fig. 4), in which microsomes from cells expressing the Y153F mutant exhibited significant increases in MGST1 activity after exposure to ONOO Ϫ . On the other hand, Tyr 92 is located in the transmembrane domain on the C-terminal side of the cytosolic loop containing the catalytic domain. Considering the location of Tyr 92 , it seems reasonable to suppose that in a manner analogous to that occurring after structural modification of Cys 49 , a conformational change brought about by nitration of this tyrosine residue might increase the rate of thiolate formation and increase catalytic activity. Therefore, in simile with the role of Cys 49 as a sensor of chemical and oxidative stress, the nitration of Tyr 92 could function as a sensor of nitrosative stress.
Data from this and other studies have shown that activation of MGST1 by RNS occurs by two distinct mechanisms: S-nitrosylation of Cys 49 and nitration of Tyr 92 , and the question arises as to the relative importance of these two mechanisms in vivo. Using the purified enzyme, the degree of activation by ONOO Ϫ is about 2.5-fold greater than that by GSNO (3,4). Using hepatic microsomes, enzyme activity is increased about 3-fold by ONOO Ϫ and is not increased by GSNO. Thus in purified enzyme or microsomal preparations, ONOO Ϫ is a better activator of the enzyme. The lack of effect of GSNO could be due to contamination of the microsomal fraction by cytosolic GSTs, whose activity is inhibited by GSNO, with a net effect of no change in activity. However, in a recent report in which S-nitrosylation of MGST1 was assessed using biotin switch methodology, no evidence was found for S-nitrosylation of the enzyme after incubation of hepatic microsomes with GSNO or in microsomes from rats treated with lipopolysaccharide to induce nitrosative stress (27). We would therefore speculate that in vivo, under pathological situations associated with increased ONOO Ϫ formation, that tyrosine nitration would be the more relevant modification of MGST1.
In a previous study we found that addition of ONOO Ϫ -treated MGST1 to microsomal membrane preparations caused a marked reduction in iron-induced lipid peroxidation, consistent with the finding that the GSH peroxidase activity of MGST1 is increased after treatment with ONOO Ϫ and suggesting a protective function for the nitrated enzyme (3). Under conditions of oxidative/nitrosative stress, most antioxidant enzymes are inactivated by S-oxidation or/and tyrosine nitration by ONOO Ϫ (5-10). The data obtained in the present study suggest that nitration of Tyr 92 and subsequent activation of MGST1 after exposure to ONOO Ϫ may be an important mechanism for cellular protection against nitrosative stress under circumstances in which other antioxidant defense mechanisms are compromised.