Inactivation of Glyceraldehyde-3-phosphate Dehydrogenase by a

Inactivation of Glyceraldehyde-3-phosphate. Dehydrogenase by a Reactive Metabolite of. Acetaminophen and Mass Spectral Characterization of an Arylated...
0 downloads 0 Views 201KB Size
Download PDF
Chem. Res. Toxicol. 1997, 10, 1097-1103

1097

Inactivation of Glyceraldehyde-3-phosphate Dehydrogenase by a Reactive Metabolite of Acetaminophen and Mass Spectral Characterization of an Arylated Active Site Peptide Eric C. Dietze, Andrea Scha¨fer,† James G. Omichinski, and Sidney D. Nelson* Department of Medicinal Chemistry, Box 357610, University of Washington, Seattle, Washington 98195 Received May 27, 1997X

Acetaminophen (4′-hydroxyacetanilide, APAP) is a widely used analgesic and antipyretic drug that can cause hepatic necrosis under some circumstances via cytochrome P450-mediated oxidation to a reactive metabolite, N-acetyl-p-benzoquinone imine (NAPQI). Although the mechanism of hepatocellular injury caused by APAP is not fully understood, it is known that NAPQI forms covalent adducts with several hepatocellular proteins. Reported here is the identification of one of these proteins as glyceraldehyde-3-phosphate dehydrogenase [GAPDH, D-glyceraldehyde-3-phosphate:NAD+ oxidoreductase (phosphorylating), EC 1.2.1.12]. Two hours after the administration of hepatotoxic doses of [14C]APAP to mice, at a time prior to overt cell damage, hepatocellular GAPDH activity was significantly decreased concurrent with the formation of a 14C-labeled GAPDH adduct. A nonhepatotoxic regioisomer of APAP, 3′hydroxyacetanilide (AMAP), was found to decrease GAPDH activity to a lesser extent than APAP, and radiolabel from [14C]AMAP bound to a lesser extent to GAPDH at a time when its overall binding to hepatocellular proteins was almost equivalent to that of APAP. In order to determine the nature of the covalent adduct between GAPDH and APAP, its major reactive and toxic metabolite, NAPQI, was incubated with purified porcine muscle GAPDH. Microsequencing analysis and fast atom bombardment mass spectrometry (FAB-MS) with collisioninduced dissociation (CID) were used to characterize one of the adducts as APAP bound to the cysteinyl sulfhydryl group of Cys-149 in the active site peptide of GAPDH.

Introduction Acetaminophen (4′-hydroxyacetanilide, APAP; Chart 1)1 is a widely used analgesic which can be hepatotoxic in humans and laboratory animals (for reviews, see refs 1-4). The hepatotoxicity appears to result from cytochrome P450-mediated formation of N-acetyl-p-benzoquinone imine (NAPQI; see Chart 1) from APAP in the liver (5-7). While the mechanism of hepatotoxicity caused by APAP has not been resolved, there is evidence that NAPQI is involved in both the formation of covalent adducts with proteins and the oxidation of proteins (14, 8-12). If arylation of proteins is important in the pathogenesis of liver cell damage caused by APAP, specific proteins must be important targets, inasmuch as a nonhepatotoxic regioisomer of APAP, 3′-hydroxyacetanilide (AMAP; Chart 1), binds as extensively to total protein in the liver of animals, but it does not cause hepatotoxicity (13, 14). Evidence that there is arylation of a select group of proteins by APAP, via its reactive metabolite NAPQI, has been obtained by immunodetection of protein-bound APAP in subcellular fractions of hepatocytes (15, 16) and by 2-dimensional electrophoretic detection of the dif* Address correspondence to this author. Tel: (206) 543-1419. Fax: (206) 685-9297. E-mail: [email protected]. † Present address: Hoechst AG, Algemeine Pharmaforschung, Instrumentelle Analytik G838, Frankfurt, Germany. X Abstract published in Advance ACS Abstracts, September 15, 1997. 1 Abbreviations: AMAP, 3′-hydroxyacetanilide; APAP, acetaminophen; CID, collision-induced dissociation; FAB, fast atom bombardment; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; G-3-P, glyceraldehyde-3-phosphate; NAPQI, N-acetyl-p-benzoquinone imine; PVDF, polyvinylidene difluoride.

S0893-228x(97)00090-8 CCC: $14.00

Chart 1. Structures of Acetaminophen (APAP), Its Nonhepatotoxic Positional Isomer AMAP, and the Toxic Metabolite of APAP, N-Acetyl-p-benzoquinone Imine (NAPQI)

ferential modulation and adduction of mouse hepatocellular proteins after doses of APAP and AMAP (17). One of the major liver cytosolic proteins that is arylated in mice administered hepatotoxic doses of APAP has been identified as a selenium binding protein (18, 19). However, its role in the pathogenesis of acetaminophen hepatotoxicity is unknown, as the function of the protein has not been characterized. Other proteins that form adducts with acetaminophen and have been identified include a microsomal subunit of glutamine synthetase (20), the mitochondrial proteins glutamate and aldehyde dehydrogenase (21, 22), and the cytosolic protein N-10formyltetrahydrofolate dehydrogenase (23). However, in none of the cases has the nature of the protein modification been determined. Because hepatocellular ATP concentrations are markedly decreased within 2 h of administration of hepatotoxic doses of APAP to mice (10, 11) but are not affected by treatment with its nonhepatotoxic regioisomer, AMAP, even though it binds as extensively to total hepatocellular protein at that time, we have been examining the effects © 1997 American Chemical Society

1098 Chem. Res. Toxicol., Vol. 10, No. 10, 1997

of APAP and AMAP treatment on the structure and function of hepatic enzymes involved in ATP synthesis. One of the enzymes, glyceraldehyde-3-phosphate dehydrogenase (EC 1.2.1.12, GAPDH), is a cytosolic enzyme involved in the glycolytic pathway of ATP synthesis and contains an active site cysteine residue required for catalysis (24). Since it is known that NAPQI, the toxic metabolite of APAP, binds primarily to cysteine residues on proteins (25, 26) GAPDH activity may be decreased by reaction with NAPQI. As described in this article, we have found that within 2 h of treatment of mice with hepatotoxic doses of APAP, GAPDH activity is significantly decreased and radiolabel from APAP becomes bound to the enzyme. This is in contrast to the nonhepatotoxic AMAP, which inhibits and is bound to GAPDH to a much lesser extent. Furthermore, we have determined that the major toxic metabolite of APAP, NAPQI, is a potent inhibitor of purified porcine muscle GAPDH and binds to Cys-149, the active site cysteine residue (27).

Experimental Procedures Materials. Silver(I) oxide, 99.99%, was purchased from Aldrich Chemical Co. (Milwaukee, WI). Reagents for gel electrophoresis and Western blotting were obtained from BioRad (Hercules, CA). B6C3/Fl mice (male, 20-25 g, SPF) were obtained from Charles River Laboratories (Wilmington, MA) and had free access to Purina rodent chow and water. Male, New Zealand white rabbits were purchased from R & R Rabbitry (Bellevue, WA) and also had free access to food and water. Trifluoroacetic acid and BCA reagent were obtained from Pierce Chemical Co. (Rockford, IL). HPLC solvents were obtained from J. T. Baker Chemical Co. (Phillipsburg, NJ). Porcine muscle GAPDH and other biochemical reagents were purchased from Sigma Chemical Co. (St. Louis, MO). Sigma Radiochemicals (St. Louis, MO) supplied [ring-U-14C]APAP (10.2 mCi/mmol) and [ring-U-14C]AMAP (16.2 mCi/mmol). The radiochemical purity of both compounds was >97% as determined by HPLC on a reverse phase 5 µm Ultrasphere ODS column (25 cm × 10 mm; Rainin Instruments, Berkeley, CA). Elution was isocratic with water/acetonitrile/glacial acetic acid (80:180:20, by volume) at a flow rate of 2 mL/min. NAPQI was synthesized as described previously (28) as was [14C]NAPQI (25). Preparation of Anti-GAPDH Antibody. Porcine muscle GAPDH was mixed with Freund’s complete adjuvant for the initial immunization and with Freund’s incomplete adjuvant for all subsequent immunizations. All immunizations were given subcutaneously in multiple locations. The initial immunization was 500 µg/rabbit and was followed, at intervals of 1 week, with four injections of 250 µg GAPDH each. One week after the last injection the rabbits were sacrificed and the blood was harvested by cardiac puncture. The serum was prepared, shown to be specific for GAPDH, and stored at -80 °C under argon. Inhibition, Reactivation, and Assay of GAPDH. GAPDH (3 mg/mL) was solubilized at room temperature in pH 7.4 potassium phosphate buffer (100 mM; 0.1 mM EDTA) and treated with NAPQI dissolved in acetonitrile at 100× the final concentration used. All controls contained equal amounts of acetonitrile. Both the ratio of NAPQI to GAPDH and the preincubation time of NAPQI with GAPDH were varied depending on the experiment. GAPDH was assayed using (()glyceraldehyde-3-phosphate (G-3-P) according to published methods (29). All samples were diluted to be in the linear range of the assay. Inactivation of GAPDH in Vivo. Mice were fasted for 20 h prior to receiving APAP or AMAP. APAP and AMAP were dissolved in sterile, isotonic saline at a concentration of 30 mg/ mL and injected intraperitonially at 400-500 mg/kg of body weight (400 mg/kg when 14C-labeled APAP or AMAP was administered). Control mice were injected with sterile, isotonic

Dietze et al. Table 1. Inhibition of GAPDH Activity in Mouse Liver by APAP and AMAPa substance administered

GAPDH activity [µmol/(min‚mg of protein)]

initial activity (%)

normal saline APAP AMAP

1.30 ( 0.05 0.22 ( 0.02 0.96 ( 0.04

100 17 74

a GAPDH activity was determined in liver cytosol of mice 2 h after the administration by ip injection of normal saline vehicle or 500 mg/kg doses of APAP and AMAP as described in Experimental Procedures. Results are expressed as means ( SE of 5 mice/group.

saline. The mice were sacrificed by cervical dislocation 2 h after the injection and the livers harvested. The livers were rinsed twice in cold pH 7.4 Tris buffer (10 mM; 250 mM sucrose, 1.0 mM EDTA) and homogenized in 5 volumes of the same buffer with 10 strokes of a glass-Teflon homogenizer. Cytosol was prepared by centrifuging the homogenate at 4 °C and 900g for 10 min, 7000g for 10 min, 20,000g for 30 min, and 105000g for 90 min. Thiol content of the cytosol was measured with Ellman’s reagent and protein content using BCA reagent. Unless GAPDH was assayed immediately, the cytosol was stored at -80 °C. Gel Electrophoresis, Western Blotting, and Imaging. Cytosolic proteins were separated using a 10% SDS-PAGE system, transferred to polyvinylidene difluoride (PVDF) membrane, and either stained with naphthol blue or treated with anti-GAPDH antibody and goat anti-rabbit IgG conjugated to horseradish peroxidase to locate GAPDH. [14C]Adducts were visualized and quantitated with a Molecular Devices phosphorimaging system and 14C-labeled protein standards. Protein levels were quantitated from Coomassie blue-stained SDSPAGE gels using a Millipore BioImage system. The BioImage system was also used to align the Coomassie blue-stained gels with the phosphorimager output. Mass Spectrometry (MS). Control or NAPQI-inactivated samples of GAPDH were dialyzed against two changes × 1000 volumes of pH 7.8 ammonium bicarbonate buffer (50 mM; 0.1 mM EDTA) digested with DPCC-treated trypsin (37 °C for 360 min with additions of 1%, w/w, trypsin at 0, 60, and 120 min) in the same buffer. The digests were frozen and lyophilized. The lyophilized digests were stored at -80 °C until used. The digests were dissolved in 70% formic acid, and 100-150 µg of digested GAPDH was injected onto a 250 mm × 4.6 mm i.d., C18 reverse phase HPLC column. The peptides were separated using a linear gradient from 5% aqueous acetonitrile: 0.1% trifluoroacetic acid to 50% aqueous acetonitrile:0.075% trifluoroacetic acid over 60 min at a flow of 1 mL/min. Peaks were collected in polypropylene tubes, frozen, lyophilized, and stored at -80 °C for MS analysis. Individual peptide fractions were characterized by fast atom bombardment-MS (FAB-MS) to identify the active site peptide. The NAPQI-labeled active site peptide was then analyzed by tandem MS. FAB/MS spectra were recorded on a VG 70-SEQ hybrid tandem mass spectrometer of EBqQ geometry (VG Analytical Ltd., Manchester, U.K.), equipped with a saddle-field fast atom gun (Ion Tech Ltd., Teddington, Middlesex, U.K.) and a VG 11/250 data system. The samples (1-5 µg) were dissolved in methanol:1 N HCl (1:1, 50 µL), and aliquots (1 µL) were added to a glycerol:thioglycerol:methanol (1:1:1) matrix on a FAB target. Sample introduction was performed with a conventional FAB probe, and ionization was achieved following bombardment with a primary beam of xenon (8 keV). Conventional FAB-MS spectra were recorded via the data system at an accelerating voltage of 8 kV and a nominal mass resolution of M/∆M 2000 (10% valley). Daughter ion spectra (MS-MS) of the MH+ ion of the active site tryptic peptide were obtained by CID in the first [radio frequency (rf) only] quadrupole. The collision energies were controlled by floating the rf-only quadrupole at potentials from 5 to 100 V below the accelerating voltage (8 kV). The collision energies, therefore, varied between 5 and 100 eV in the laboratory frame of reference. The parent MH+ ion of the

Inactivation of GAPDH by Acetaminophen

Chem. Res. Toxicol., Vol. 10, No. 10, 1997 1099 summed, centroids were assigned, and the peaks were subsequently mass measured by the data system.

Results

Figure 1. Comparison of 14C-incorporation from APAP and AMAP and staining by anti-GAPDH antibody. SDS-PAGE gels of 150 µg of cytosol from livers of B6C3/F1 mice which had been dosed with 400 mg/kg of either 14C-labeled APAP or AMAP. The gels were run and the protein was transferred to PVDF which was cut in separate lanes and either exposed to a Phosphorimager or developed with anti-GAPDH antibody as described in Experimental Procedures. analyte was selected by the adjustment of the magnetic field strength of the magnet (B), and daughter ion spectra were recorded by scanning the quadrupole mass analyzer (Q) from m/z 50 to 1200 over a period of 10 s. The spectra were recorded in the MCA (multichannel analyzer) mode; 5-10 scans were

Two hours after treatment of mice with a hepatotoxic dose of APAP (500 mg/kg ip), GAPDH activity in liver cytosol was inhibited by 83% (Table 1). After an equivalent dose of the nonhepatotoxic regioisomer, AMAP, GAPDH activity was only inhibited by 26%. The 2 h time was chosen because hepatic ATP concentrations are maximally decreased to approximately 60% of controls by that time in mice treated with hepatotoxic doses of APAP (10, 11). Separation of liver cytosolic proteins from mice treated with [14C]APAP and [14C]AMAP by SDS-PAGE revealed the presence of several protein bands that contained radiolabel as detected by phosphorimage analysis (Figure 1). One of these bands had an Mr ∼ 38 kDa and reacted with GAPDH antibodies on Western blots (Figure 1). Moreover, N-terminal sequence analysis of this band after electroelution from PVDF membranes revealed a sequence identical with the first seven amino acids of porcine muscle GAPDH and a murine lymphocyte GAPDH (H2N-V-K-V-G-V-N-G) (27). Gene bank sequences for GAPDH from other sources gave matches for five of the seven N-terminal amino acids. Because of limitations on the amounts of 14C-radiolabeled analogs of APAP and AMAP available to us and the specific activity of radiolabel required to detect the protein adducts, doses of 400 mg/kg [14C]APAP and [14C]AMAP were administered to mice in this study. This dose of APAP was only marginally hepatotoxic and caused 15-20% decreases in liver GAPDH activity in the mice 2 h after the administration of drug. The GAPDH from the mouse that received [14C]APAP contained 0.16

Figure 2. Inhibition of porcine muscle GAPDH by NAPQI. GAPDH (3 mg/mL) was inhibited by the addition of varying amounts of NAPQI to achieve different molar concentration ratios of NAPQI([I]) to [GAPDH]. The preincubation time of NAPQI and GAPDH was for 5 min prior to the assay of GAPDH activity as described in Experimental Procedures. The molar titration-inhibition data were analyzed as a biphasic plot, and linear regression analysis of the first linear portion of the curve provided an estimate of the efficiency of inactivation as shown. The activity of uninhibited GAPDH controls was 36 ( 4 µmol/(min‚mg of protein).

1100 Chem. Res. Toxicol., Vol. 10, No. 10, 1997

Dietze et al.

Table 2. Amino Acid Sequence Analysis of the Active Site Peptide of Pig Muscle GAPDH Labeled by [14C]NAPQI cycle

amino acid released

pig muscle GAPDH

1 2 3 4 5 6 7 8 9 10

Ile Val Ser Asn Ala Ser X Thr Thr Asn

Ile-143 Val-144 Ser-145 Asn-146 Ala-147 Ser-148 Cys-149 Thr-150 Thr-151 Asn-152

mol of [14C]APAP radioequiv bound/mol of GAPDH monomer. Previous studies using stopped-flow techniques have shown that the cysteinyl-containing tripeptide, glutathione, reacts with NAPQI with a second-order rate constant of 104-105 M-1 s-1 (30), and NAPQI was found to inactivate GAPDH at a rate too rapid to accurately measure. However, estimation of the efficiency of inactivation of GAPDH could be determined by incubating a fixed concentration of enzyme with increasing amounts of inhibitor (Figure 2). The percent activity remaining after 30 min of incubation was plotted vs the ratio of inhibitor to enzyme. Extrapolation of the first linear portion of the curve gave an intercept on the horizontal plot (x-intercept) of 3.7, indicating that approximately 4 mol of NAPQI were required to inhibit 1 mol of GAPDH. Because each mole of GAPDH contains four equivalent active subunits, one molecule of NAPQI is apparently required for the inactivation of each subunit. Reasons for the lack of complete inactivation of 1 mol of GAPDH by 4 mol of NAPQI were not investigated. Previously we have found that NAPQI undergoes comproportionation and polymerization in aqueous media (28), which would decrease the amount of NAPQI available to react with enzyme, thereby decreasing the efficiency of inactivation of GAPDH. In addition, there are 3 cysteinyl thiols/subunit of porcine muscle GAPDH (31) that could potentially be arylated by NAPQI. Although the active site cysteinyl thiol is the most reactive (31), it is possible that NAPQI arylates one or more of the other three cysteinyl thiol groups. If arylation of one or more of these thiol groups does not markedly affect enzyme activity, the efficiency of inactivation of GAPDH by NAPQI would be decreased as well. Indirect evidence for reaction of NAPQI with the active site cysteine (Cys-149) of porcine muscle GAPDH was provided by results of coincubation experiments with the normal enzyme substrate G-3-P and the cofactor NAD+. G-3-P forms a thiohemiacetal with the active site cysteine residue during catalysis and was found to protect against inactivation by NAPQI, whereas coincubation with the cofactor NAD+ had essentially no effect on the inactivation (data not shown). Direct evidence for the reaction of NAPQI with Cys149 was obtained by analysis of an active site peptide obtained by trypsin digestion of GAPDH modified with either 4-vinylpyridine or [14C]NAPQI. The 4-vinylpyridine-modified peptide was isolated by HPLC and found to have a sequence identical to 10 amino acids of the porcine muscle GAPDH active site residue (Table 2). Note: 4-Vinylpyridine is used to identify cysteine residues in the microsequencing facility at the University of

Figure 3. Incorporation of [14C]APAP into GAPDH. (A) HPLC of GAPDH inactivated with [14C]NAPQI and digested with trypsin. The peptides were separated on a C18 reverse phase column with a gradient from 5% aqueous acetonitrile (0.1% TFA) to 50% aqueous acetonitrile (0.075% TFA) over 60 min. Approximately 100-150 µg of the GAPDH digest was injected. (B) Positive ion FAB-MS of the [14C]NAPQI-labeled peptide (30.25 min). The 14C-containing peptide which eluted at 30.25 min was collected, lyophilized, and analyzed by FAB-MS as described in Experimental Procedures. The proposed structure of the (M + H+) peak containing APAP is shown.

Washington. The NAPQI-modified residue also was isolated by HPLC (Figure 3A, peak eluting at 30 min) and found to have an identical sequence with the exception of a blank reading at Cys-149 (Table 2). Although trypsin digestion of GAPDH yields an active site peptide of 17 amino acids (31), bulky adducts at Cys-149 apparently lead to additional cleavage as has been noted by others (32). A FAB-MS of the NAPQI-modified peptide gave a protonated molecular ion at m/z 1159 (Figure 3B) consistent with the addition of NAPQI to the active site peptide containing amino acids 143-152 of porcine muscle GAPDH. A CID mass spectrum for the protonated molecular ion (M + H+) of m/z 1159 is shown in Figure 4. Significant ions are depicted above the spectrum. The ions are consistent with both the peptide backbone sequence and cysteine as the site of covalent modification. The latter is based on shifts in masses of several sequence ions by 149 amu (b7, b9, y4, y5, y6, and y8) and by the relatively intense daughter ion at m/z 975 indicative of neutral loss of APAP plus the cysteinyl thiol group (MH+ - 184). This reaction yields the protonated dehydroalanyl peptide. A low-abundance ion at m/z 183 also suggests that the cysteinyl thiol is attached to the APAP structure, most likely at the C-3′ position of the aromatic ring, based on previous structural studies of reactions of cysteinyl thiol-containing compounds with NAPQI (33, 34).

Inactivation of GAPDH by Acetaminophen

Chem. Res. Toxicol., Vol. 10, No. 10, 1997 1101

Figure 4. CID mass spectrum for the molecular ion (M + H+) of m/z 1159 of the modified active site peptide. Significant ions indicating the sequence of this peptide are depicted above the spectrum. Fragments that contain Cys-149 (Cys*) show a shift of 149 amu for APAP indicating that this is the site of covalent binding. A relatively intense ion at m/z 975 (MH+ - 184) indicates the loss of mercaptoacetaminophen from the peptide, and a low-intensity ion at m/z 183 is consistent with this structure as well.

The molar binding ratio, under the conditions used in this experiment in vitro, was between 1.2 and 1.4 mol of NAPQI/mol of GAPDH monomer. The HPLC chromatogram (Figure 3A) indicates several additional peptides that contain radiolabel from APAP, some of which may contain adducts to other cysteinyl residues. Some of these adducts may result from reactions of NAPQI polymers with GAPDH, as it is known that NAPQI undergoes rapid polymerization under the conditions of the reaction (28). A more complete analysis of several other adducts is ongoing.

Scheme 1. Possible Reactions of NAPQI with an Active Site Cysteine Residue (Cys-149) in GAPDH To Form a Hypothetical 3′-Adduct (top) or a Hypothetical ipso-Adduct (bottom)

Discussion Results of these studies indicate that hepatocellular GAPDH is one of the early protein targets of arylation by NAPQI, the major reactive and toxic metabolite of the widely used drug APAP. Administration of a hepatotoxic dose of APAP to mice leads to loss of activity of GAPDH (Table 1) concurrent with its arylation within 2 h of dosing (Figure 1). In contrast, reactive metabolites of AMAP, a nonhepatotoxic regioisomer of APAP, bind less extensively (Figure 1) and cause less inactivation of GAPDH (Table 1). Using antibodies to protein-bound APAP, other investigators have detected an APAPprotein adduct of apparent Mr ∼ 38 kDa in the cytosolic fraction obtained from both mice (12) and human (35) livers after hepatotoxic doses of APAP. However, the protein was not identified. Our results suggest that this protein is GAPDH. Studies in vitro with porcine muscle GAPDH clearly implicate the active site cysteine residue (Cys-149) as a site of arylation by NAPQI that leads to loss of enzyme

activity (Figures 2-4). Normally, this cysteine residue forms a thiohemiacetal with the substrate G-3-P (3638), but it can react with other soft electrophiles (39, 40). As previously noted, NAPQI is a soft electrophile that reacts with cysteinyl thiols of peptides and proteins to form a 3′-S-cysteinyl adduct (vida infra). Though we believe it is this adduct that is formed with Cys-149 of GAPDH, another thiol adduct of NAPQI that has been suggested is an ipso-adduct at the imine carbon (30, 41), and we cannot rule out such a possibility (Scheme 1). Whatever the case, this is the first report to our knowledge of the structural characterization of a protein adduct of acetaminophen. It should be noted that mechanisms other than arylation of the active site thiol of GAPDH may be involved in its inactivation or modification by APAP reactive metabolites. For example, it has been found that GAP-

1102 Chem. Res. Toxicol., Vol. 10, No. 10, 1997

DH is sensitive to inactivation and S-thiolation during oxidative stress (42-44). NAPQI is a strong oxidizing quinone-like compound and is known to cause oxidant stress under some conditions (3, 8-12). Under our reducing assay conditions, we would not have detected such modifications. Unfortunately, after the administration of hepatotoxic doses of APAP to mice at a time when >80% of GAPDH activity was inhibited, we were not able to isolate enough modified GAPDH to obtain a mass spectrum of the arylated active site peptide. Therefore, it is possible that a large fraction of the GAPDH inactivated in vivo may have been S-thiolated or may have formed an ipso-adduct unstable to our workup conditions. Further work is required in this regard. Whatever the mechanism of GAPDH modification by APAP reactive metabolites, inhibition and/or inactivation of this enzyme may have several effects. Not only is GAPDH a key enzyme in glycolysis that catalyzes the oxidative phosphorylation of glyceraldehyde-3-phosphate to form 1,3-diphosphoglycerate in the presence of NAD+ and inorganic phosphate (24), this enzyme also appears to be involved in the regulation of activities of other cytosolic and membrane-bound proteins (45-47) and of nuclear transfer RNA (48, 49). Furthermore, nitrosylation of the active site thiol group of GAPDH apparently initiates its own modification by NADH and possibly ADP-ribosylation (50-52). Thus, new roles for GAPDH in DNA repair, gene expression, endocytosis, and cellular toxicity have been suggested (53, 54). Although we do not know yet if inhibition of GAPDH by APAP reactive metabolites plays a role in APAP hepatotoxicity, it will be of interest to investigate the effects of modification of GAPDH by APAP reactive metabolites on the varied functions of this enzyme.

Acknowledgment. The authors thank Santosh Kumar (Department of Biochemistry, University of Washington) for performing the microsequencing analyses. This study was supported by Grant GM-25418 (S.D.N.) from the National Institutes of Health.

References (1) Hinson, J. A. (1980) Biochemical toxicology of acetaminophen. Rev. Biochem. Toxicol. 2, 103-130. (2) Black, M. (1984) Acetaminophen hepatotoxicity. Annu. Rev. Med. 35, 577-593. (3) Nelson, S. D. (1990) Molecular mechanisms of the hepatotoxicity caused by acetaminophen. Semin. Liver Dis. 10, 267-278. (4) Vermeulen, N. P. E., Bessems, J. G. M., and van de Straat, R. (1992) Molecular aspects of paracetamol-induced hepatotoxicity and its mechanism-based prevention. Drug Metab. Rev. 24, 367407. (5) Mitchell, J. R., Jollow, D. J., Potter, W. Z., Davis, D. C., Gillette, J. R. and Brodie, B. B. (1973) Acetaminophen-induced hepatic necrosis. I. Role of drug metabolism. J. Pharmacol. Exp. Ther. 187, 185-194. (6) Dahlin, D. C., Miwa, G. T., Lu, A. Y. H., and Nelson, S. D. (1984) N-Acetyl-p-benzoquinone imine: A cytochrome P-450-mediated oxidation product of acetaminophen. Proc. Natl. Acad. Sci. U.S.A. 81, 1327-1331. (7) Patten, C. J., Thomas, P. E., Guy, R. L., Lee, M., Gonzalez, F. J., Guengerich, F. P., and Yang, C. S. (1993) Cytochrome P450 enzymes involved in acetaminophen activation by rat and human liver microsomes and their kinetics. Chem. Res. Toxicol. 6, 511518. (8) Tee, L. B. G., Boobis, A. R., Hugett, A. C., and Davies, D. S. (1986) Reversal of acetaminophen toxicity in isolated hamster hepatocytes by dithiothreitol. Toxicol. Appl. Pharmacol. 83, 294-314. (9) Kyle, M. E., Nakae, D., Serroni, A., and Farber, J. L. (1988) Superoxide dismutase and catalase protect cultured hepatocytes from the cytotoxicity of acetaminophen. Mol. Pharmacol. 84, 584589.

Dietze et al. (10) Tirmenstein, M. A., and Nelson, S. D. (1990) Acetaminopheninduced oxidation of protein thios: Contribution of impaired thiol metabolizing enzymes to the breakdown of adenine nucleotides. J. Biol. Chem. 265, 3059-3065. (11) Jaeschke, H. (1990) Glutathione disulfide formation and oxidant stress during acetaminophen-induced hepatotoxicity in mice in vivo. J. Pharmacol. Exp. Ther. 255, 935-941. (12) Birge, R. B., Bartalone, J. B., Cohen, S. D., Khairallah, E. A., and Smolin, L. (1991) A comparison of proteins S-thiolated by glutathione to those arylated by acetaminophen. Biochem. Pharmacol. 42, 5197-5207. (13) Tirmenstein, M. A., and Nelson, S. D. (1989) Subcellular binding effects on calcium homeostasis produced by acetaminophen and a non-hepatotoxic regioisomer, 3′-hydroxyacetanilide. J. Biol. Chem. 264, 9814-9819. (14) Roberts, S. A., Price, V. F., and Jollow, D. J. (1990) Acetaminophen structure-toxicity studies: In vivio covalent binding of a nonhepatotoxic analog, 3-hydroxyacetanilide. Toxicol. Appl. Pharmacol. 105, 195-208. (15) Bartalone, J. B., Sparks, K., Cohen, S. D., and Khairallah, E. A. (1987) Immunochemical detection of acetaminophen-bound liver proteins. Biochem. Pharmacol. 36, 1193-1196. (16) Roberts, D. W., Pumford, N. R., Potter, D. W., Benson, R. W., and Hinson, J. A. (1987) A sensitive immunochemical assay for acetaminophen-protein adducts. J. Pharmacol. Exp. Ther. 241, 527-533. (17) Myers, T. G., Dietz, E. C., Anderson, N. L., Khairallah, E. A., Cohen, S. D., and Nelson, S. D. (1995) A comparative study of mouse liver proteins arylated by reactive metabolites of acetaminophen and its nonhepatotoxic regioisomer, 3′-hydroxyacetanilide. Chem. Res. Toxicol. 8, 403-413. (18) Pumford, N. R., Martin, B. M., and Hinson, J. A. (1992) A metabolite of acetaminophen covalently binds to the 56 kDa selenium binding protein. Biochem. Biophys. Res. Commun. 182, 1348-1355. (19) Bartalone, J. B., Birge, R. B., Bulera, S. J., Bruno, M. K., Nishanian, E. V., Cohen, S. D., and Khairallah, E. A. (1992) Purification, antibody production, and partial amino acid sequence of the 58-kDa acetaminophen binding liver proteins. Toxicol. Appl. Pharmacol. 113, 19-29. (20) Bulera, S. J., Birge, R. B., Cohen, S. D., and Khairallah, E. A. (1995) Identification of the mouse liver 44-kDa acetaminophen binding protein as a subunit of glutamine synthetase. Toxicol. Appl. Pharmacol. 134, 313-320. (21) Halmes, C. N., Hinson, J. A., Martin, B. M., and Pumford, N. (1996) Glutamate dehydrogenase covalently binds to a reactive metabolite of acetaminophen. Chem. Res. Toxicol. 9, 541-546. (22) Landin, J. S., Cohen, S. D., and Khairallah, E. A. (1996) Identification of a 54-kDa mitochondrial acetaminophen-binding protein as aldehyde dehydrogenase. Toxicol. Appl. Pharmacol. 141, 299-307. (23) Pumford, N. R., Halmes, N. C., Martin, B. M., Cook, R. J., Wagner, C., and Hinson, J. A. (1997) Covalent binding of acetaminophen to N-10-formyl-tetrahydrofolate dehydrogenase in mice. J. Pharmacol. Exp. Ther. 280, 501-505. (24) Harris, J. I., and Waters, M. (1976) Glyceraldehyde-3-phosphate dehydrogenase. In The Enzymes, 3rd ed. (Boyer, P. D., Ed.) Vol. 13, pp 1-49, Academic Press, New York. (25) Streeter, A. J., Dahlin, D. C., Nelson, S. D., and Baillie, T. A. (1984) The covalent binding of acetaminophen to protein: Evidence for cysteine residues as major sites of arylation in vitro. Chem.-Biol. Interact. 48, 349-366. (26) Hoffman, K.-J., Streeter, A. J., Axworthy, D. B., and Baillie, T. A. (1985) Identification of the major covalent adduct formed in vitro and in vivo between acetaminophen and mouse liver proteins. Mol. Pharmacol. 27, 566-573. (27) Sabath, D. E., Broome, H. E., and Prystowsky, M. B. (1990) Glyceraldehyde-3-phosphate dehydrogenase mRNA is a major interleukin 2-inducible transcript in a cloned T-helper lymphocyte. Gene 91, 185-191. (28) Dahlin, D. C., and Nelson, S. D. (1982) Synthesis, decomposition kinetics, and preliminary toxicological studies on pure N-acetylp-benzoquinone imine, a proposed toxic metabolite of acetaminophen. J. Med. Chem. 25, 885-886. (29) Dagher, S. M., and Deal, W. C. (1982) Glyceralde-3-phosphate dehydrogenase from pig liver. Methods Enzymol. 89, 310-316. (30) Coles, B., Wilson, I., Wardman, P., Hinson, J. A., Nelson, S. D., and Ketterer, B. (1988) The spontaneous and enzymic reaction of N-acetyl-p-benzoquinone imine with glutathione. Arch. Biochem. Biophys. 264, 253-260.

Inactivation of GAPDH by Acetaminophen (31) Perham, R. N. (1969) The comparative structure of mammalian glyceraldehyde-3-phosphate dehydrogenases. Biochem. J. 111, 17-21. (32) Chen, S., Lee, T. D., Legesse, K., and Shively, J. E. (1986) Identification of the arylazido-β-alanyl-NAD+-modified site in rabbit muscle glyceraldehyde-3-phosphate dehydrogenase by microsequencing and fast atom bombardment mass spectrometry. Biochemistry 25, 5391-5395. (33) Nelson, S. D., Vaishnav, Y., Kambara, H., and Baillie, T. A. (1981) Comparative EI, CI, and FD mass spectra of some thioether metabolites of acetaminophen. Biomed. Mass Spectrom. 8, 244251. (34) Hinson, J. A., Monks, T. J., Hong, M., Highet, R. J., and Pohl, L. R. (1982) 3-(Glutathion-S-yl)acetaminophen: A biliary metabolite of acetaminophen. Drug Metab. Dispos. 10, 47-50. (35) Birge, R. B., Bartalone, J. B., Emeigh-Hart, S. G., Nishanian, E. V., Tyson, C. A., Khairallah, E. A., and Cohen, S. D. (1990) Acetaminophen hepatotoxicity: Correspondence of selective protein arylation in human and mouse liver in vitro, in culture and in vivo. Toxicol. Appl. Pharmacol. 105, 472-482. (36) Segal, H. L., and Boyer, P. D. (1953) The role of sulfhydryl groups in the activity of D-glyceraldehyde-3-phosphate dehydrogenase. J. Biol. Chem. 204, 265-281. (37) Moras, D., Olsen, K. W., Sabesan, M. N., Buehner, M., Ford, G. C., and Rossman, M. G. (1975) Studies of asymmetry in the threedimensional structure of lobster D-glyceraldehyde-3-phosphate dehydrogenase. J. Biol. Chem. 250, 9137-9162. (38) Skarzyniski, T., Moody, P. C. E., and Wonacott, A. J. (1987) Structure of holo-glyceraldehyde-3-phosphate dehydrogenase from Bacillus stearothermophilus at 1.8Å resolution. J. Mol. Biol. 193, 171-187. (39) Uchida, K., and Stadtman, E. R. (1993) Covalent attachment of 4-hydroxynonenal to glyceraldehyde-3-phosphate dehydrogenase: A possible involvement of intra- and intermolecular crosslinking reaction. J. Biol. Chem. 268, 6388-6393. (40) Cane, D. E., and Sohng, J.-K. (1994) Inhibition of glyceraldehyde3-phosphate dehydrogenase by pentalenolactone. 2. Identification of the site of alkylation by tetrahydropentalenolactone. Biochemistry 33, 6524-6530. (41) Smith, C. V., and Mitchell, J. R. (1985) Acetaminophen hepatotoxicity in vivo is not accompanied by oxidant stress. Biochem. Biophys. Res. Commun. 133, 329-336. (42) Brodie, A. E., and Reed, D. J. (1990) Cellular recovery of glyceraldehyde-3-phosphate dehydrogenase activity and thiol status after exposure to hydroperoxides. Arch. Biochem. Biophys. 276, 212-218.

Chem. Res. Toxicol., Vol. 10, No. 10, 1997 1103 (43) Graven, K. K., Troxler, R. F., Kornfeld, H., Panchenko, M. V., and Farber, H. W. (1994) Regulation of endothelial cell glyceraldehyde-3-phosphate dehydrogenase expression by hypoxia. J. Biol. Chem. 269, 24446-24453. (44) Ravichandran, V., Seres, T., Moriguchi, T., Thomas, J. A., and Johnston, R. B., Jr. (1994) S-Thiolation of glyceraldehyde-3phosphate dehydrogenase induced by phagocytosis-associated respiratory burst in blood monocytes. J. Biol. Chem. 269, 2501025015. (45) Knull, H. R., and Walsh, J. L. (1992) Association of glycolytic enzymes with the cytoskeleton. Curr. Top. Cell Regul. 33, 1530. (46) Rogalski, A. A., Steck, T. L., and Waseem, A. (1989) Association of glyceraldehyde-3-phosphate dehydrogenase with the plasma membrane of the intact human red blood cell. J. Biol. Chem. 264, 6438-6446. (47) Hsu, S.-C., and Molday, R. S. (1990) Glyceraldehyde-3-phosphate dehydrogenase is a major protein associated with the plasma membrane of retinal photoreceptor outer segments. J. Biol. Chem. 265, 13308-13313. (48) Singh, R., and Green, M. R. (1993) Sequence-specific binding of transfer RNA by glyceraldehyde-3-phosphate dehydrogenase. Science 259, 365-368. (49) Nagy, E., and Rigby, W. F. C. (1995) Glyceraldehyde-3-phosphate dehydrogenase selectively binds AU-rich RNA in the NAD+binding region (Rossmann Fold). J. Biol. Chem. 270, 2755-2763. (50) Molina y Vedia, L., McDonald, B. M., Reep, B., Bru¨ne, B., Di Silvio, M., Billiar, T. R., and Lapetina, E. G. (1992) Nitric oxideinduced S-nitrosylation of glyceraldehyde-3-phosphate dehydrogenase inhibits enzymic activity and increases endogenous ADPribosylation. J. Biol. Chem. 267, 24929-24932. (51) McDonald, L. J., and Moss, J. (1993) Stimulation by nitric oxide of an NAD linkage to glyceraldehyde-3-phosphate dehydrogenase. Proc. Natl. Acad. Sci. U.S.A. 90, 6238-6241. (52) Mohr, S., Stamler, J. S., and Bru¨ne, B. (1996) Posttranslational modification of glyceraldehyde-3-phosphate dehydrogenase by S-nitrosylation and subsequent NADH attachment. J. Biol. Chem. 271, 4209-4214. (53) Sirover, M. A. (1996) Emerging new functions of the glycolytic protein, glyceraldehyde-3-phosphate dehydrogenase, in mammalian cells. Life Sci. 58, 2271-2277. (54) Roses, A. D., Burke, J. R., Vance, J. M., and Strittmatter, W. J. (1996) A role for GAPDH in apoptosis and neurodegeneration. Nature Med. 2, 609-610.

TX970090U