Glutathiolation of Proteins by Glutathione Disulfide S -Oxide Derived from S -Nitrosoglutathione MODIFICATIONS OF RAT BRAIN NEUROGRANIN/RC3 AND

S -Nitrosoglutathione (GSNO) undergoes spontaneous degradation that generates several nitrogen-containing compounds and oxidized glutathione derivatives. We identified glutathione sulfonic acid, glutathione disulfide S -oxide (GS(O)SG), glutathione disulfide S -dioxide, and GSSG as the major decomposition products of GSNO. Each of these compounds and GSNO were tested for their efficacies to modify rat brain neurogranin/RC3 (Ng) and neuromodulin/GAP-43 (Nm). Among them, GS(O)SG was found to be the most potent in causing glutathiolation of both proteins; four glutathiones were incorporated into the four Cys residues of Ng, and two were incorporated into the two Cys residues of Nm. Ng and Nm are two in vivo substrates of protein kinase C; their phosphorylations by protein kinase C attenuate the binding affinities of both proteins for calmodulin. When compared with their respective unmodified forms, the glutathiolated Ng was a poorer substrate and glutathiolated Nm a better substrate for protein kinase C. Glutathiolation of these two proteins caused no change in their binding affinities for calmodulin. Treatment of [ 35 S]cysteine-labeled rat brain slices with xan-thine/xanthine oxidase or a combination of xanthine/

Protein S-glutathiolation can be induced in cells by mild oxidative stress (1). GSSG has been shown to oxidatively regulate the activity of several purified enzymes including carbonic anhydrase III (2,3), protein kinase C (PKC) 1 (4), human aldose reductase (5), and human immunodeficiency virus, type I protease (6), and in each case the effects of glutathiolation can be reversed by reducing agents. As the concentration of reduced GSH in the mammalian cells is in the millimolar range and that of GSSG is less than 5% of GSH, glutathiolation of proteins by GSSG in vivo is not likely an efficient mechanism. More recently, the superoxide-induced glutathiolation of protein (7) and that induced by peroxynitrite, nitric oxide (NO), and nitrosothiol, in particular, S-nitrosoglutathione (GSNO), are thought to be the main avenues leading to protein Sthiolation (8 -12). In mammalian cells, a relatively high concentration of GSH (0.5-10 mM) serves as an NO sink to form GSNO (13)(14)(15)(16), which can undergo transnitrosylation with protein sulfhydryl group to form S-nitrosoprotein and GSH or to form protein-GSH mixed disulfide and nitroxyl (17)(18)(19)(20). GSNO can also release NO in the presence of cuprous ion (21), ascorbate (22), or thiols (20,23) and serves as a possible source of nitrsonium or nitroxyl ions (24). In addition, GSNO is unstable in aqueous solution and undergoes decomposition, which is believed to be homolytic cleavage of the S-N bond to give NO and a thiyl radical (25,26). Indeed, the reactions involving GSNO are fairly complex and generate many potential products including ammonia, NO, nitrous oxide, nitrite, sulfinamide, hydroxylamine, and several oxidized forms of glutathione (20,23). Recently, it was found that freshly prepared GSNO was effective in S-nitrosylation of proteins through transnitrosylation, whereas the decomposed GSNO was more effective in S-glutathiolation of proteins (10). It was suggested that glutathione sulfenic acid was the active component for glutathiolation of proteins.
Neurogranin/RC3 (Ng) and neuromodulin/GAP-43 (Nm) are two prominent PKC substrates in the brain. Phoshorylations of both Ng and Nm reduce their binding affinities for calmodulin (CaM) (27,28). Rat brain Ng contains four, and Nm contains two Cys residues; these Cys residues in Ng form two pairs of intramolecular disulfides upon oxidation by NO and other oxidants (29,30), and those in Nm undergo palmitoylation (31). Intramolecular disulfide formation renders Ng a poorer substrate of PKC and also reduces its binding affinity for CaM (29). The effect of oxidation of the two Cys residues of Nm has not been elucidated. However, it was shown that treatment of cultured dorsal root ganglion and PC-12 cells with 3-morpholino-sydononimine, which generates both NO and superoxide, inhibited palmitoylation of Nm (32). Recently, we found that treatment of Ng with GSNO caused oxidation to form intramolecular disulfides, as well as glutathiolation; the extent of glutathiolation was greatly increased upon incubation of Ng with the decomposed GSNO (33). In this study, by using mass spec-trometry, we have identified several glutathione derivatives as the degradation products of GSNO, including glutathione sulfonic acid (GSO 3 H), glutathione disulfide S-oxide (GS(O)SG), glutathione disulfide S-dioxide (GS(O) 2 SG), and GSSG. Among them, GS(O)SG was the most potent in causing glutathiolation of Ng and Nm. The level of this compound was found to increase upon treatment of rat brain slices with oxidants. Ng had been positively identified as a target of thiolation under oxidative stress.

EXPERIMENTAL PROCEDURES
Materials-The following materials were obtained from the indicated sources: GSNO was from Calbiochem; sodium nitroprusside (SNP), diamide, bovine serum albumin, GSH, iodoacetamide (IAM), GSSG, xanthine (X), and xanthine oxidase (XO) were from Sigma; H 2 O 2 was from Fisher Scientific; phosphatidylserine and dioleoylglycerol were from Avanti Polar Lipids; BCA protein determination reagent was from Pierce; [ 35 S]cysteine was from PerkinElmer Life Sciences; protein A-Sepharose was from Amersham Pharmacia Biotech; and horseradish peroxidase-conjugated goat anti-rabbit IgG was from Bio-Rad.
Purification of the Various Glutathione Derivatives of GSNO-GSNO (100 mM) dissolved in water was kept at room temperature for 24 -72 h, and the solution was injected into a C 18 reverse phase HPLC column (connecting two Vydac 218TP54, 0.46 ϫ 25 cm) eluted with 0.1% trifluoroacetic acid for 2 h at a flow rate of 0.5 ml/min. Fractions of 0.25 ml were collected, lyophilized, and subjected to electrospray mass spectrometry analysis. The concentrations of the various glutathione derivatives were determined by amino acid composition analysis.
Preparation of Rat Brain Slices and Treatment with Oxidants-Brain from adult Harlan Sprague Dawley rat (200 -250 g) was removed immediately after decapitation and placed in ice-cold artificial cerebrospinal fluid (ACSF; 125 mM NaCl, 2.5 mM KCl, 2 mM CaCl 2 , 26 mM NaHCO 3 , 1.25 mM NaH 2 PO 4 , 1 mM MgCl 2 , 25 mM glucose, bubbled with 95% O 2 /5% CO 2 ) for 3-5 min. Cortical slices (400-m thickness) were prepared using a Vibratome 1000 sectioning device with the brain submerged in ice-cold ACSF bubbled with 95% O 2 /5% CO 2 . The slices were incubated for 1 h in ACSF containing 200 M cycloheximide to block protein synthesis and then [ 35 S]cysteine (22 Ci/ml) was added to label the intracellular GSH for 1-2 h. The slices were treated with X (0.5 mM)-XO (80 milliunits) and X-XO plus SNP (0.5 mM) for 15 min. After incubation, the slices were washed 3 times with ice-cold ACSF containing 1 mM cysteine, followed by the addition of 100 mM IAM, and homogenized. The homogenate was treated with 5% HClO 4 , the acidsoluble fraction was analyzed by HPLC, and radioactivity was determined by scintillation counting. Identification of the various glutathione derivatives was based on their retention times compared with those of the authentic compounds.
Characterization of Protein S-Thiolation-[ 35 S]Cysteine-labeled rat brain slices previously treated with H 2 O 2 (0.5 mM), diamide (0.25 mM), SNP (0.5 mM), X (0.5 mM)-XO (80 milliunits), and a combination of X-XO and SNP for 15 min were kept frozen on dry ice and homogenized in 50 mM Tris-Cl buffer, pH 7.5, containing 0.5% Nonidet P-40, and 100 mM IAM. The homogenates were centrifuged at 20,000 ϫ g for 20 min, and the protein concentration in the supernatant was determined with the BCA reagent. For the identification of thiolated proteins, the homogenates (100 g) treated with or without 50 mM DTT were heated at 95°C for 5 min and resolved by 10 -20% gradient SDS polyacrylamide gel electrophoresis, and proteins were transferred to nitrocellulose membrane for autoradiography. The differences in the radioactivities among the various protein bands from those samples treated with or without 50 mM DTT reflected the extent of protein S-thiolation. The same membrane was used for the identification of Nm and Ng by immunoblot with a preparation of antibody (number 270) recognizing both proteins (34).
Immunoprecipitation of Thiolated Ng and Nm-The [ 35 S]cysteinelabeled rat brain slices were homogenized in Buffer A (20 mM Tris-Cl, pH 7.8, 2 mM EDTA, 2 mM EGTA, 50 mM KF, 5 mM sodium pyrophosphate, 5 g each of leupeptin, aprotinin, pepstatin A, and chymostatin, 50 M okadaic acid, 0.5% Nonidet P-40, 0.1% SDS, and 100 mM IAM). To 600 g of the homogenate (in 1 ml of Buffer A containing 150 mM NaCl) 25 l of antibody number 270 was added and incubated at 4°C overnight. The immune complexes were collected by centrifugation following incubation with 100 l of 50% (w/v) protein A-Sepharose (in Buffer A plus 150 mM NaCl) at room temperature for 2 h and washed twice with Buffer A plus 150 mM NaCl and twice with Buffer A alone. Ng and Nm were dissociated from the immune complexes by adding 100 l of 100 mM glycine, pH 3.0, and the supernatant was neutralized (ϳpH 7) by 1.5 M Tris-Cl, pH 8.8. 35 S-labeled proteins were detected by autoradiography after SDS-PAGE and transfer to nitrocellulose membrane. The same membrane was also used for immunoblot to quantify Ng and Nm.
Preparation of Rat Brain Ng and Nm and Recombinant Ng-Rat brain Ng (34) and Nm (35) and the recombinant Ng (36) were purified to homogeneity as described previously. For the preparations of reduced Ng (red-Ng) and Nm (red-Nm), purified proteins were incubated in 20 mM Tris-Cl, pH 7.5, 1 mM EDTA, and 1 mM DTT for 10 min. The reaction mixtures were applied to a C 4 reverse phase HPLC column (Vydac 214TP54, 0.46 ϫ 25 cm) and eluted with 0.1% trifluoroacetic acid (Solvent A) and 0.1% trifluoroacetic acid plus acetonitrile (Solvent B) gradient at 0.5 ml/min as follows: 0 -10 min, Solvent B kept at 0%; 10 -15 min, Solvent B increased to 15%; 15-55 min, Solvent B increased to 60%; and 55-60 min, Solvent B increased to 100%. Fractions containing Ng and Nm were lyophilized and dissolved in N 2 -saturated water immediately before use. GS-Ng and GS-Nm prepared by incubation of red-Ng and red-Nm with GS(O)SG were purified by HPLC under the same conditions.
Chromatography of GS-Ng and GS-Nm on CaM-affinity Column-GS-Ng or GS-Nm (10 -30 g) was applied to a CaM-Sepharose column (1-ml bed volume) equilibrated with 50 mM Tris-Cl buffer, pH 7.5, containing 1 mM EDTA, 1 mM EGTA, 0.1 M NaCl, and 2 mM ascorbic acid. The column was washed with 2.5 ml of the same buffer and eluted with 2.5 ml of 50 mM Tris-Cl buffer, pH 7.5, containing 1 mM EDTA, 1 mM EGTA, 0.5 M NaCl, 6 mM CaCl 2 , and 2 mM ascorbic acid. The DTT-reduced Ng or Nm (incubation in 20 mM Tris-Cl, pH 7.5, 1 mM EDTA, and 1 mM DTT at 30°C for 30 min followed by the addition of 20 mM ascorbic acid) was analyzed under the same conditions except that both the washing and elution buffers also contained 2 mM DTT. Fractions (0.16 ml) were collected for assay of Ng or Nm by exhaustive phosphorylation with PKC or analyzed by immunoblot with an antibody recognizing both proteins.
Phosphorylation of Ng and Nm by PKC-Ng or Nm was phosphorylated by PKC in 30 mM Tris-Cl buffer, pH 7.5, containing 6 mM MgCl 2 , 0.12 mM [␥-32 P]ATP, 50 g/ml phosphatidylserine, 10 g/ml dioleoylglycerol, 0.4 mM CaCl 2 , Ϯ 5 mM DTT, and Ng or Nm. For analysis of the kinetic parameters of GS-Ng and GS-Nm phosphorylations by PKC, DTT was omitted. 32 P incorporations into Ng and Nm were measured by the Dowex AG1 ϫ8/DEAE-cellulose mini column method (34). Kinetic constants were determined using the ALLFIT program (37).

Electrospray Ionization Mass Spectrometry (ES-MS)-ES-MS
was performed using an M-Scan Quattro II upgraded Bio-Q instrument with quadrupole analyzer. Myoglobin was used to calibrate the instrument. Sample aliquots of 10 l were injected into the instrument, and elution was carried out using a mixture of acetonitrile, 0.1% trifluoroacetic acid, methoxyethanol, and isopropanol (1:1:1:1) at a flow rate of 5 l/min. The spectra were deconvoluted, and the masses were expressed in Da.

Oxidation of Ng by GSNO and Its Decomposition Products-
Rat brain Ng is susceptible to oxidation by several oxidants resulting in the formation of intramolecular disulfides. The oxidized Ng exhibits an increased electrophoretic mobility in nonreducing SDS-PAGE (29). Incubation of red-Ng with a low concentration (Ͻ1 mM) of freshly prepared GSNO resulted in the formation of a fast-migrating oxidized Ng containing intramolecular disulfide bridges. At a higher GSNO concentration (Ͼ3 mM) this fast-migrating oxidized Ng was slightly reduced (Fig. 1A). This phenomenon became more pronounced when the red-Ng was treated with a partially decomposed GSNO preparation (Fig. 1B). A decrease in the intramolecular disulfide-bridged form of Ng at a higher concentration of GSNO, especially with the decomposed GSNO, suggests that a different type of modification of Ng sulfhydryl group occurs. To characterize the mechanism of modification of Ng by GSNO, the reactions were terminated by addition of 100 mM IAM, purified by HPLC, and analyzed by ES-MS. For controls, treatment of red-Ng (7494.8 Da) with IAM resulted in an incorporation of 4 acetamides (AM)/Ng (7723.6 Da) (Fig. 1C), whereas treatment of the air-oxidized Ng (7491.9 Da) with IAM did not result in any incorporation of AM (Fig. 1D). For brevity, we refer to the Ng containing intramolecular disulfide(s) as ox-Ng. By ES-MS analysis, the ox-Ng containing one pair of intramolecular disulfides will incorporate 2 AM, and that containing two pairs of intramolecular disulfides will incorporate none. Incubation of red-Ng with 3 mM of the partially decomposed GSNO for 5 s resulted in an extensive glutathiolation (1-3 glutathiolated residues (GS-)-/Ng) (Fig. 1E), and after 30 min of incubation a majority of Ng was glutathiolated to 4 GS-/Ng (8718.1 Da), and a minor fraction contained 2 GS-/Ng with one pair of intramolecular disulfides (8105.4 Da) (Fig. 1F). It should be noted that in these experiments the reaction mixture also contained 1 mM DTT to reduce the purified Ng before treatment so that the concentration of GSNO required to oxidize Ng was relatively high. Nevertheless, these results indicate that the partially decomposed GSNO contains a potent glutathiolating component.
Characterization of the Decomposition Products of GSNO-GSNO in aqueous solution was kept at room temperature for 24 -72 h and separated by C 18 reverse phase HPLC to characterize the decomposition products. The freshly prepared GSNO was eluted from HPLC as a major component having m/z ϭ 337 (data not shown). A partially decomposed GSNO preparation contained several components labeled as P-1, -2 a , -2 b , -3, -4, and -5 with m/z ϭ 356, 629, 629, 645, 337, and 613, respectively (Fig. 2). Based on the estimated masses of these products, we predict that P-1 is GSO 3 H (mass, 356 Da), P-2 a and P-2 b are stereoisomers of GS(O)SG (mass, 629 Da), P-3 is GS(O) 2 SG (mass, 645 Da), P-4 is GSNO (mass, 337 Da), and P-5 is GSSG (mass, 613 Da). The multiple ionic species seen in P-4, GSNO, are because of fragmentation of the parent compound during ES-MS, i.e. the m/z ϭ 307 species is the one without NO.
The effects of these compounds on the modification of Ng were analyzed by SDS-PAGE and ES-MS (Fig. 3). GSO 3 H, GS(O) 2 SG, GSNO, and GSSG were effective in causing Ng oxidation-forming intramolecular disulfide, whereas both GS(O)SG stereoisomers (P-2 a and P-2 b ) were unique in causing modification of Ng characteristic for glutathiolation. The Ng modified by these compounds (3 mM for 5 min) was analyzed by ES-MS. P-1, GSO 3 H, caused partial oxidation of Ng generating one intramolecular disulfide bond without glutathiolation (7624.3 Da) (Fig. 3, P-1). This oxidized form of Ng contained one extra oxygen over the expected species, which should have one intramolecular disulfide and two free sulfhydryl groups being modified by IAM with a resulting mass of 7608 Da. We speculate that this is because of modification of the methionine residue to form methionine sulfoxide. P-2, GS(O)SG, caused extensive glutathiolation of Ng with the majority of the proteins containing 4 GS-/Ng (8717.8 Da) (Fig. 3, P-2). The identity of another modified Ng of 8063.4 Da is unknown. P-3, GS(O) 2 SG, caused partial oxidation of Ng forming one intramolecular disulfide without glutathiolation, such as ox-Ng plus 2 AM (7609.3 Da) (Fig. 3, P-3). It should be noted that the major Ng species after treatment with GS(O) 2 SG contained 3 AM (7666.8 Da), suggesting that one of the sulfhydryl groups is readily modified by this compound so that it is no longer susceptible to modification by IAM. It appears that this particular modification is very susceptible to fragmentation during ES-MS, i.e. Cys-OH, so that it is not detectable in the ES-MS spectra. In addition, we were unable to assign the structure of a 7561.9-Da species after treatment with GS(O) 2 SG. P-4, GSNO, caused Ng oxidation to form intramolecular disulfides and partial glutathiolation (1 GS-/Ng); however, the reaction products, which may also include nitrosylated Ng, were rather complex, and many of them had not been identified (Fig. 3,  P-4). The Ng-Cys-NO cannot be modified by IAM but is susceptible to fragmentation during ES-MS so that certain Ng species shown in the ES-MS spectra contained an odd number of modified sulfhydryl groups. P-5, GSSG, caused mainly oxidation of Ng to form intramolecular disulfides with relatively little glutathiolation ( Fig. 3, P-5). The oxidized Ng also contained one extra oxygen presumably because of formation of methionine sulfoxide. These results indicate that P-2, GS(O)SG, is the most potent glutathiolating agent among the various products derived from GSNO. The formation of GS(O)SG can also be detected by bubbling either GSH or GSSG solution with NO gas (data not shown).

Efficacy of GS(O)SG in the Glutathiolation of Ng and
Nm-To test the efficacy of GS(O)SG in the glutathiolation of protein, freshly reduced Ng (80 M) was separated from DTT by HPLC and was incubated with increasing concentrations of GS(O)SG for 5 min. After stopping the reactions with IAM the reaction products were analyzed by ES-MS (Fig. 4). The reduced Ng contained four free sulfhydryl groups, which were completely modified by IAM resulting in Ng plus 4 AM (7723.7 Da) (Fig. 4A). At 50 M of GS(O)SG, Ng was oxidized to form two intramolecular disulfides (7491.2 Da) or one intramolecular disulfide plus 2 GS-/Ng (8104.1 Da) (Fig. 4B). At 100 M, the majority of Ng contained one intramolecular disulfide and 2 GS-/Ng (Fig. 4C). As the concentrations of GS(O)SG were increased to 200 (Fig. 4D), 500 (Fig. 4E), and 1000 M (Fig. 4F We also tested the efficacy of GS(O)SG on the glutathiolation of Nm, which contains two Cys residues. In the control, treatment of red-Nm with IAM resulted in the incorporation of 2 AM into rat brain Nm, which contained multiple phosphorylated species (Fig. 5A). Incubation of red-Nm (25 M) with an equimolar concentration of GS(O)SG resulted in ϳ50% of Nm being glutathiolated to 2 GS-/Nm (Fig. 5B). The nonglutathiolated Nm, however, did not contain any AM, suggesting that both -SH groups are modified by this compound to form Nm-Cys-OH. As the concentrations of GS(O)SG were increased to 50 (Fig. 5C), 100 (Fig. 5D), 200 (Fig. 5E) ation by PKC and Binding to CaM-The single PKC phosphorylation site of Ng (Ser-36) is located near the center of the molecule, and the four glutathiolation sites are at Cys-3, -4, -9, and -51. The GS-Ng exhibited a higher K m value (35.9 Ϯ 4.6 versus 24.1 Ϯ 3.9 M) and lower V max value (1010 Ϯ 70 versus 2110 Ϯ 120 units/mg) than that of the red-Ng; thus GS-Ng is a poorer substrate for PKC compared with red-Ng (V max /K m was 28 versus 88) (Table I). Nm also contains a single PKC phosphorylation site (Ser-41) within the Ng/Nm homologous region. Glutathiolation of Nm at Cys-3 and -4 resulted in a reduction of its K m value for PKC to nearly one-half that of the reduced Nm (11.2 Ϯ 1.6 versus 20.3 Ϯ 2.5 M), however, without an effect on V max values (1390 Ϯ 120 versus 1330 Ϯ 87 units/mg). Thus, GS-Nm became a better substrate for PKC compared with red-Nm (V max /K m was 120 versus 70). Glutathiolation of Ng and Nm had no effect on their bindings to the CaM-affinity column in the presence of EGTA (data not shown).

Generation of GS(O)SG and other Oxidized Glutathione Species in Rat Brain Slices Treated with
Oxidants-Rat brain slices were incubated with [ 35 S]cysteine to label glutathione in the presence of cycloheximide followed by treatment with X-XO and a combination of X-XO and SNP. These experiments were designed to test whether generation of GS(O)SG was responsive to oxidative and nitrosative stresses. Previously, we have shown that SNP was one of the most potent NO donors tested, including 2-(N,N-diethylamino)-diazenolate-2-oxide and S-nitroso-N-acetylpenicillamine (29,30,36), in causing Ng oxidation. The perchloric acid-soluble fractions were analyzed by C 18 reverse phase HPLC (Fig. 6). Identification of the various glutathione derivatives was based on their retention times as compared with the authentic compounds. GSO 3 H and [ 35 S]cysteine had the same retention time; they were not separable by this chromatography. Treatment of the slices with X-XO, which generates superoxide, caused an increase in both GSSG and GS(O)SG (Fig. 6B) as compared with the control (Fig. 6A), and a greater increase in these two compounds was seen with a combination of X-XO and SNP (Fig. 6C). The latter treatment also resulted in a slight increase in GSNO; however, because of its instability quantification of GSNO was not vigorously pursued.
To determine the extent of thiolation of proteins in the [ 35 S]cysteine-labeled brain slices, the tissue extracts were resolved by nonreducing (Fig. 7A) and reducing SDS-PAGE (Fig.  7B), and proteins were transferred to nitrocellulose membrane and analyzed by autoradiography. Following treatments of the slices with H 2 O 2 , SNP, diamide, X-XO, and X-XO/SNP several prominent 35 S-labeled proteins of 43, 39,36,32,28,20,17, and 14 kDa were observed to be thiolated as evidenced by their release of 35 S in the reducing gel. The overall patterns of protein thiolation by these oxidants were qualitatively similar. However, thiolation of certain proteins was more distinct by treatment with different oxidants; e.g. thiolation of the 43-kDa protein was more prominent by diamide and X-XO, and that of the 36-kDa protein was more prominent by H 2 O 2 , X-XO, and X-XO/SNP. Characterization of Ng Thiolation by Immunoprecipitation-The various oxidant-treated 35 S-labeled brain slices were analyzed for the thiolation of Ng and Nm by immunoprecipitation with antibody number 270, which recognizes both Ng and Nm. Following resolution by nonreducing SDS-PAGE, transfer of proteins to nitrocellulose membrane, and autoradiography, it was evident that thiolation of Ng was enhanced by treatments of the slices with H 2 O 2 , SNP, diamide, X-XO, and X-XO/SNP (Fig. 8A). No detectable thiolation of Nm was seen in these experiments. Immunoblot of the same nitrocellulose membrane showed that both Ng and Nm were immunoprecipitated by the antibody (Fig. 8B), and the oxidation of Ng to form intramolecular disulfides induced by these oxidants appeared to exhibit similar responses as those of thiolation. The extents of thiolation and oxidation to form intramolecular disulfides were more  prominent by treatments with SNP, X-XO, and X-XO/SNP than those with H 2 O 2 and diamide. The increase in thiolation of Ng induced by the former three treatments ranged from 3-4-fold, and those by the latter two treatments ranged from 0.6 -1.6fold more than the untreated control from two separate experiments. Thiolation of Ng detected in the immunoprecipitate was further confirmed by running the SDS-PAGE under reducing conditions, in which all the 35 S bound to the protein was removed (Fig. 8C), and all the intramolecular disulfide forms of Ng were reduced (Fig. 8D).

DISCUSSION
While testing the effect of GSNO on the oxidation of Ng to form intramolecular disulfide, we found that the freshly prepared GSNO was effective in this modification, but a partially decomposed GSNO was more effective in causing glutathiolation (33). Analysis of the decomposition products of GSNO led us to identify GS(O)SG as one of the most potent glutathiolating agents among the various glutathione derivatives tested, including GSO 3 H, GS(O) 2 SG, GSNO, and GSSG (Fig. 3). Although these latter compounds can oxidize Ng to form intramo-lecular disulfide bonds, they are not very effective for glutathiolation of this protein. Glutathiolation of protein by GS(O)SG likely proceeds by the following reactions, where R 1 and R 2 are either protein or any sulfhydryl-containing compound. Modification of proteins containing multiple sulfhydryl groups, such as Ng, by GS(O)SG is complicated by two competing reactions, namely, formation of intramolecular disulfide and glutathiolation. Partially glutathiolated Ng can be driven to form intramolecular disulfide, but the intramolecular disulfide form of Ng cannot be glutathiolated. Thus, at a low GS(O)SG concentration Ng forms intramolecular disulfide, and at a high concentration Ng is glutathiolated. When the ratio of GS(O)SG/-SH is equal or greater than one, both Ng and Nm are stoichiometrically glutathiolated.
Protein S-thiolation, especially S-glutathiolation, has been recognized as one of the physiological responses to nitrosative and oxidative stresses. The mechanism by which these stresses induce protein S-thiolation is poorly understood. Several mechanisms have been proposed for protein glutathiolation including the following: 1) thiol-disulfide exchange between protein thiols and GSSG (38); 2) oxidation of protein thiols by oxyradicals or H 2 O 2 to form thiyl radicals or sulfenic acids and then to interact with GSH to produce mixed disulfide (39); 3) nucleophilic attack of protein thiolate on GSNO to produce mixed disulfide (9 -12); 4) oxidation of GSH to form sulfenic acid and then interact with protein thiols to form mixed disulfides (10); and 5) nitrosation of protein thiols followed by interaction with GSH to form mixed disulfides (9, 10). The present study suggests another mechanism that utilizes GS(O)SG as a potential GS-donor for protein glutathiolation. GS(O)SG is more potent than GSNO in glutathiolation of protein, and this degradation product of GSNO may account for some of the effects of GSNO. GS(O)SG is present at a low level in the control rat brain slices labeled with [ 35 S]cysteine and is increased under oxidative stress. GS(O)SG and GS(O) 2 SG have been shown to be the major products generated by oxidation of GSH with H 2 O 2 (40). We also showed that GS(O)SG could be formed by oxidation of both GSH and GSSG with NO (data not shown). Although GSSG can be reduced by glutathione reductase to form GSH, it is unknown whether GS(O)SG and GS(O 2 )SG are also substrates of the reductase. Accumulation of GS(O)SG under basal physiological conditions could provide tonic glutathiolation of certain proteins that employ this modification mechanism for regulation of their activities or subcellular localization.
Thiolation of Ng has been positively identified by immunoprecipitation of the [ 35 S]cysteine-labeled rat brain cortical slices, whereas thiolation of Nm was not detected under the same conditions. The two Cys residues of Nm are the potential sites of palmitoylation, which is involved in the association of this protein with the membrane (31). It is apparent that the acylated Nm cannot be thiolated. Under the basal conditions, in which the slices were incubated with ACSF without added oxidants, Ng was thiolated to a lower level as compared with those treated with oxidants. In addition, Ng also subjected to oxidation to form intramolecular disulfides in the presence of the oxidants. These findings indicate that Ng can undergo multiple oxidative modifications at its Cys residues. The nature of thiolation of Ng in the brain slices has not been determined; we do not know whether it is because of glutathiolation or mixed disulfides between Ng and Cys. It is interesting to note that both oxidation of Ng to form intramolecular disulfides and glutathiolation convert this protein to become a poorer substrate of PKC by reducing the V max value (see Table I and Ref. 29); however, the former modification, but not the latter, also causes a reduction in the binding affinity of Ng toward CaM. Because the partially thiolated Ng can be converted into the intramolecular disulfide form of Ng, one may consider thiolation as an intermediate step toward intramolecular disulfide formation. Thus, oxidative stress may cause these two modifications of Ng and result in an attenuation of Ng phosphorylation by PKC.
In summary, the current study provides evidence that GS(O)SG derived from GSNO is one of the most potent glutathiolating agents known so far. This compound can be easily prepared from the decomposed GSNO, by bubbling GSH or GSSG solution with NO gas, or by treatment of GSH with H 2 O 2 . GS(O)SG is relatively stable upon storage but is very reactive toward sulfhydryl group. GS(O)SG apparently is generated in situ following oxidative stress and may be responsible for the glutathiolation of proteins in vivo. Here, we have also demonstrated that Ng is thiolated in rat brain slices and that the extent of thiolation is enhanced under oxidative stress. FIG. 8. Identification of Ng thiolation by immunoprecipitation. Rat cortical slices were labeled with [ 35 S]cysteine and treated with H 2 O 2 , SNP, diamide, X-XO, and X-XO/SNP for 15 min (see text for details). Tissue extracts (600 g/ml) were incubated with antiserum number 270 (25 l), and the immune complexes were recovered by adding protein A-Sepharose. Ng was eluted from protein A-Sepharose with glycine, pH 3, and neutralized with Tris-Cl. Proteins were analyzed by nonreducing (A and B) and reducing (C and D) 10 -20% SDS-PAGE and transferred to nitrocellulose membrane. The same membranes were analyzed by autoradiography (A and C) and immunoblot with antibody number 270 (C and D). Note that the levels of Ng among the various treated samples determined by immunoblot were comparable (B), but the extents of thiolation (A), as well as the formation of ox-Ng (B), were increased in the oxidant-treated samples. Both the thiolated Ng (A) and intramolecular disulfide-bridged form of Ng (ox.Ng) (B) disappeared following separation on reducing gel (C and D).