Chem. Res. Toxicol. 2000, 13, 237-244
237
Disulfiram Generates a Stable N,N-Diethylcarbamoyl Adduct on Cys-125 of Rat Hemoglobin β-Chains in Vivo John C. L. Erve,† Ole N. Jensen,‡ Holly S. Valentine,† Venkataraman Amarnath,† and William M. Valentine*,† Department of Pathology and Center in Molecular Toxicology, Vanderbilt University Medical Center, Nashville, Tennessee 37232-2561, and Department of Molecular Biology, Odense University/ University of Southern Denmark, Odense 5230 M, Denmark Received November 11, 1999
Disulfiram (DSF) is a drug used in aversion therapy to treat alcoholics and acts by inhibiting mitochondrial low-Km aldehyde dehydrogenase. Investigations into the mechanisms for in vivo inactivation suggest that the DSF metabolite S-methyl-N,N-diethylthiocarbamate sulfoxide reacts irreversibly with an active site Cys. This work aimed to determine if DSF generates monothiocarbamate adducts on cysteine residues in vivo by examining hemoglobin. SpragueDawley rats were treated with DSF po for 2, 4, and 6 weeks. Rats have four different globin β-chains, of which three (β-1-3) contain two cysteine residues each. MALDI-TOF MS analysis of two new globin species from DSF-treated rats collected by HPLC revealed increments of 99 Da above the mass of the unmodified chains (β-2 and β-3). In a separate experiment, the globin mixture was digested for 2 h with Glu-C and reanalyzed by MALDI-TOF MS. Results showed a peptide at m/z 2716.3 having a mass 99 Da higher than a known Cys-containing peptide. Subsequently, the Glu-C digest was analyzed using Q-TOF tandem MS, enabling observation of the +4 charge state of the peptide with m/z 2716.3. This peptide was fragmented to produce y-sequence ions that located the modification to Cys-125 (present on both β-2 and β-3). Cys125 is the most reactive of two cysteine residues on these β-chains. To confirm the structure of the modification, globin was hydrolyzed with 6 N HCl at 110 °C for 18 h. The adduct survived these conditions so that S-(N,N-diethylcarbamoyl)cysteine was detected in the hydrolysates of treated rats on the basis of comparison with the tandem MS spectrum of a standard. These results extend the findings of others obtained using glutathione conjugates and demonstrate the ability of DSF to covalently modify Cys residues of proteins in a manner consistent with the production of S-methyl-N,N-diethylthiocarbamate sulfoxide, or sulfone, intermediates.
Introduction Disulfiram[bis(diethyldithiocarbamoyl)disulfide](DSF),1 or Antabuse, has been used in alcohol aversion therapy to treat alcoholism for almost 50 years since its implementation by Larsen et al. in 1948 (1). The basis of this therapy relies on the series of unpleasant reactions, known as the DSF ethanol reaction (2), which include intense flushing of the face and/or neck, tachycardia, nausea, and vomiting, that occur after ingestion of ethanol (1, 3). Blood levels of acetaldehyde, the toxic proximate metabolite of ethanol metabolism, is increased in patients on Antabuse who drink ethanol (3) and is thought responsible for many of these effects. It is believed that acetaldehyde accumulation arises due to inhibition of low-Km liver mitochondrial aldehyde dehydrogenase (EC 1.2.1.3, ALDH) brought about by DSF (1, * To whom correspondence should be addressed. † Vanderbilt University Medical Center. ‡ Odense University/University of Southern Denmark. 1 Abbreviations: ACN, acetonitrile; CID, collision-induced dissociation; DSF, disulfiram; DTT, dithiothreitol; DTNB, dithiobis(2,2′nitrobenzoic) acid; DETC-MeSO, S-methyl diethylthiocarbamate sulfoxide; DETC-MeSO2, S-methyl diethylthiocarbamate sulfone; DETCCys, diethylcarbamoylcysteine; DEDC, diethyldithiocarbamate; DEDCMe, S-methyl diethyldithiocarbamate; ESI, electrospray ionization; GSH, glutathione; MALDI-TOF MS, matrix-assisted laser desorption ionization time-of-flight mass spectrometry; MeOH, methanol; Q-TOF, quadrupole orthogonal time-of-flight; SA, sinapic acid; TFA, trifluoroacetic acid.
3). On the basis of early in vitro work with ALDH, in vivo DSF inhibition was thought to result from an intramolecular disulfide bond involving an active site Cys (4). However, evidence now suggests that metabolism of DSF to a reactive metabolite(s) is required for the irreversible inactivation observed in vivo (5-7). The disulfide bond of DSF (1) can be reduced in blood (Scheme 1) presumably by glutathione (GSH) (8), or by albumin (9). The resulting diethyldithiocarbamate (DEDC) (2) can be methylated by thiol methyl transferase or thiopurine methyltransferase (10, 11) to give rise to methyl diethyldithiocarbamate (DEDC-Me) (3). Oxidation of the sulfonyl group of DEDC-Me by a number of different P450s such as P450 3A4/5 (7) produces methyldiethythiocarbamate (DETC-Me) (4), which can undergo additional oxidation to generate the sulfoxide (DETCMeSO) (5). Further oxidation to a sulfone (DETC-MeSO2) (6) may also occur, and the oxygenated metabolites have been shown to be capable of inactivating ALDH in vitro with the reactivity of DETC-MeSO2 being higher than that of DETC-MeSO (6). GSH conjugates (7) that were consistent with reaction between the Cys of GSH and DETC-MeSO and/or DETC-MeSO2 were detected in bile of rats treated with DSF or diethyldithiocarbamate (12). The reactivity of DSF metabolites with proteins in vivo has not been investigated in detail. For example, al-
10.1021/tx990191n CCC: $19.00 © 2000 American Chemical Society Published on Web 03/16/2000
238
Chem. Res. Toxicol., Vol. 13, No. 4, 2000 Scheme 1
though irreversible inhibition of human ALDH by DETCMeSO has been demonstrated in vitro (13, 14), detection of the adduct on ALDH in vivo has not been reported. DSF hepatotoxicity is a rare consequence of DSF therapy but is associated with high mortality rates (15). Recently, a case of DSF hepatitis in which autoantibodies were formed against P450 was reported (16). Such findings suggest that proteins are modified following DSF therapy, at least in some individuals, although the nature of the hapten has not been elucidated. The primary purpose of this study was to determine if administration of DSF resulted in covalent modification of hemoglobin in vivo. To achieve this goal, we used a combination of mass spectrometric techniques to characterize an adduct on Cys-125 of the β-chain of hemoglobin following oral administration of DSF to rats.
Materials and Methods Chemicals. Endoproteinase Glu-C was purchased from Boehringer Mannheim (Mannheim, Germany). Sinapic acid (SA) was purchased from Fluka Chemie (Buchs, Switzerland). DSF in feed was prepared by Bio-Serve (Frenchtown, NJ). Acetonitrile (ACN), methanol (MeOH), and water were all HPLC grade. Constantly boiling (6 N) HCl was from Pierce (Rockford, IL). Nano-electrospray (ESI) needles were obtained from Protana A/S (Odense, Denmark). Myoglobin, human insulin, n-octyl β-Dglucopyranoside, and dithiothreitol (DTT) were from Sigma (St. Louis, MO). SelfPack 20-R2 and 50-R1 POROS reverse phase packing material was purchased from PerSeptive Biosystems Inc. (Framingham, MA). Carbonyl sulfide was obtained from Aldrich Chemical Co. (Milwaukee, WI). Chemical Synthesis. (1) N,N-Diethylthiocarbamate Methyl Ester (DETC-Me). Diethylamine (1.1 mL, 10 mmol), ethanol (10 mL), and 10 N NaOH (1 mL) were taken in a 25 mL two-neck flask, and the flask was fitted with a Dewar condenser containing dry ice and 2-propanol. The flask was stirred and cooled in a dry ice/2-propanol bath (-70 °C), and carbonyl sulfide (∼1 mL) from a gas cylinder was condensed into the flask. The second neck was stoppered, and the temperature of the bath was allowed to rise to 10 °C and maintained at that temperature while the Dewar condenser was kept cold.
Erve et al. Carbonyl sulfide evaporated and condensed into the reaction mixture. After 4 h, the condenser was removed and iodomethane (0.7 mL) was added. The stirring was continued for 2 h before ethanol was removed, and the residue was distributed between water (10 mL) and chloroform (20 mL) layers. The latter was separated, dried, and evaporated. (2) N,N-Diethylthiocarbamate Methyl Sulfone (DETCMeSO2). A solution of DETC-Me (1.1 g, 7.5 mmol) in dichloromethane (15 mL) was cooled in ice, and 3-chloroperoxybenzoic acid (3.75 g, 15 mmol) in the same solvent (20 mL) was added with stirring. The reaction mixture was stirred in ice for 30 min and at room temperature for 4 h. The reaction was checked by TLC (a 9:1 hexane/ethyl acetate mixture). The solid was filtered, and the filtrate was washed with saturated NaHCO3 (2 × 25 mL), dried, and concentrated. Purification was accomplished with flash chromatography (hexane, followed by a 5:1 hexane/ ethyl acetate mixture). The presence of the sulfone in the fractions was checked by GC: 1H NMR δ 1.23 and 1.30 (t, 6H, J ) 7.1 Hz, CH3), 3.14 (s, 3H, CH3S), 3.42 and 3.75 (q, 4H, J ) 7.1 Hz, CH2); 13C NMR δ 12.1 and 13.7 (CH3), 39.5 (CH3S), 41.4 and 42.5 (CH2), 160.5 (CdO). (3) S-(N,N-Diethylcarbamoyl)cysteine (DETC-Cys). The pH of a solution of Cys (0.52 g, 4 mmol) in water (20 mL) was adjusted to 8 with 1 N NaOH and the mixture stirred with DETC-MeSO2 in methanol (20 mL) under Ar for 2 h. The precipitated solid was filtered and washed with 1:1 MeOH/H2O (20 mL). The filtrate was concentrated before purification over a column of silica (40-200 mesh) with 20% H2O in ACN. The fractions exhibiting a UV maximum around 220 nm were combined and concentrated to obtain the cysteinyl adduct (TLC with 20% H2O in ACN; Rf ) 0.4): 13C NMR δ 13.2 and 13.6 (CH3CH2), 31.4 (SCH2), 43.5 and 44.0 (CH3CH2), 55.8 (CH), 168.8 (NCdO), 173.0 (CO2H); ESI-MS m/z 221 (M + H)+. In Vivo Exposures. This study was performed in accordance with the NIH Guide for Care and Use of Laboratory Animals and was approved by the Vanderbilt animal care committee. Male Sprague-Dawley rats (Harlan Sprague Dawley, Inc.) (n ) 6; 2 controls), 6-7 weeks old, were used after a 2 week acclimation period. The rats were housed under a diurnal light cycle with water (ad libitum) and fed 1% DSF in a nutritionally complete pelleted diet for 14, 28, or 40 days ad libitum. Body weights were checked weekly to assess food intake and weight gain. Globin Purification. Rats were bled under general anesthesia (15 mg of Xylazine/kg of body weight, combined with 100 mg of Ketamine/kg of body weight) by exsanguination from the abdominal aorta. Blood was collected in a heparinized syringe and centrifuged to separate the plasma from the red cells. A 5 mM phosphate-buffered saline solution (pH 7.4) containing 150 mM NaCl was added in an equal volume to the red cells, and the mixture resuspended and centrifuged at 4000g. The supernatant and buffy coat were discarded, and the washing procedure was repeated twice. The washed red cells were lysed with twice their volume of 5 mM phosphate buffer (pH 7.4) and centrifuged at 14000g for 25 min. The hemolysate was mixed with 1 mL of 1 M ascorbic acid and added dropwise to 50 mL of cold 2.5% oxalic acid in acetone, and globin was allowed to precipitate for 15 min. Samples were then centrifuged at 12000g for 10 min; the supernatant was aspirated, and the globin was washed by adding 5-6 mL of acetone, resuspended with a metal spatula, and centrifuged at 12000g for 10 min. The supernatant was removed and the globin dried under a stream of nitrogen and stored at -78 °C. Chromatography. Separation of globin chains was accomplished by dissolving globin from control or DSF-treated rats in HPLC solvent A to a final concentration of 1 mg/mL and injecting 20 µL onto a C4 RP column (4.6 mm × 150 mm, 300 Å, 5 µm, Vydac, Hesperia, CA). Proteins were eluted at a rate of 1 mL/min with a linear gradient from 44 to 57% B over the course of 40 min [solvent A is ACN/H2O/TFA (20:60:0.01 v/v/v); solvent B is ACN/H2O/TFA (60:40:0.008 v/v/v)]. The eluant was monitored at 214 nm. Protein fractions were collected and
Disulfiram Metabolite Adduct Hemoglobin in Vivo lyophilized prior to subsequent mass analysis. The HPLC system consisted of dual LKB 2150 pumps with a diode array detector (Shimadzu, SpectraChrom SPD-M10A). Mass Spectrometry. MALDI-TOF MS analyses of proteolytic digests (both on probe and solution) were performed on a Voyager Elite (PerSeptive Biosystems Inc.) time-of-flight (TOF) mass spectrometer equipped with delayed extraction. The ion acceleration was 20.5 kV, the grid voltage 97.5%, and the guide wire voltage 0.3%. The instrument was operated with a delay time of 350 ns in the linear mode. Mass spectra were acquired as the sum of ions generated by irradiation of the target with 100-300 laser pulses (337 nm N2 laser) and processed using GRAMS/386 software (Galactic Industries Corp., Salem, NH) and GPMAW (LightHouse Data, Odense, Denmark). Internal calibration was accomplished using masses of the two known peptide fragments derived from β-3, [27-43] and [2790] appearing at m/z 2098.5 and 7047.1, respectively. MALDITOF MS of intact proteins was performed on a Bruker Reflex II mass spectrometer (Bruker-Franzen Analytik GmbH, Bremen, Germany). The matrix solution was prepared by saturating ACN/H2O/TFA (50:49:0.1 v/v/v) with SA. For analysis of peptide mixtures generated by enzymatic digestion in solution, 0.6 µL of sample was mixed with 0.6 µL of SA matrix, and the mixture was applied to the sample plate by the dried-droplet technique and air-dried. For intact protein analysis, 1 µL of micropurified DSF globin (1 mg/mL in 20% ACN) was mixed with the internal standards insulin and myoglobin and the mixture applied by the dried droplet technique with the SA matrix. LaserOne software (European Molecular Biology Laboratory, Heidelberg, Germany) was used to analyze data acquired on the Bruker mass spectrometer. Tandem mass spectrometry was performed on a quadrupoleorthogonal time-of-flight instrument (Q-TOF) (Micromass, Manchester, U.K.) equipped with a nano-ESI ion source. Approximately 3 µL of peptide mixture, or 1 µL of acid hydrolysate, was loaded into the sample needle. The needle voltage was 900 V, the cone voltage at 30-40 V, and the focus at 0 V. The aperture was adjusted between 2 and 20 V, and resolutions (LM and HM) were between 0 and 8 V. Collision energy varied between 15 and 20 eV (12 eV for acid hydrolysate), and Ar was used as a collision gas (analyzer reading of 1 × 10-4 Torr). Data were collected over 2-5 min and peptide tandem MS data deconvoluted using MaxEnt sequencing software provided with the instrument. Enzymatic Digestion. Enzymatic digestion of protein on the MALDI probe sample plate was performed by redisolving the protein-SA matrix spot with 1 µL of NH4HCO3 (100 mM, pH 7.8) and adding 0.3 µg of Glu-C. The sample plate was then placed in a humidity chamber and the chamber maintained at 37 °C for 1.5 h. After digestion had occurred, 0.5 µL of SA matrix and 0.5 µL of 1% formic acid were added to the digest and the mixture was allowed to recrystallize before reanalysis by MALDI-TOF MS. In solution, digestion was performed in 0.6 mL Eppendorf tubes by adding 5 µL of a globin mixture (1 mg/ mL), 10 µL of a NH4HCO3 buffer (pH 6.5), and 1.5 µg of Glu-C. Digestion was performed for 2 h at 37 °C. For some samples, more extensive digestion of the β-chains was desired. This was accomplished by adding, after the initial 2 h, 5 µL of n-octyl β-D-glucopyranoside (20 mM) to aid in denaturing the globin and an additional 1 µg of Glu-C, followed by incubation for 45 min at 37 °C. Acid Hydrolysis. Approximately 1 mg of globin was added to an Eppendorf tube and placed in a hydrolysis vial. A portion of 6 N HCl (180 µL) was placed in the bottom of the vial which was flushed with Ar gas before being placed in an oven at 110 °C for 18 h. Upon completion of hydrolysis, samples were reconstituted with 100 µL of a MeOH/H2O/formic acid mixture (50:49:1) and filtered using a spin filter with a pore size of 0.2 µm. Samples were stored at -20 °C before being analyzed by Q-TOF MS as described above. Reaction with DTT. Globin (4 µL, 1 mg/mL) isolated from rats exposed to DSF was mixed with 8 µL of a NH4HCO3 buffer
Chem. Res. Toxicol., Vol. 13, No. 4, 2000 239
Figure 1. MALDI/TOF MS spectra of new globin peaks isolated from DSF-treated rats (6 week DSF administration in feed at 1% w/w). The peaks at m/z 15 948 (A) and 15 962 (B) correspond to diethylcarbamoyl adducts (+99 Da) on β-3 and β-2, respectively. Samples were dissolved in SA matrix and spectra acquired as described in Materials and Methods. (pH 8.5) to which 4 µL of 50 mM DTT was added. Samples were incubated for 2 h at 37 °C prior to microcolumn desalting (see below) and MALDI-TOF MS analysis. The identical procedure was repeated at pH 10.5 with incubation at either 37 or 56 °C. Micropurification of Globin and Globin Digests. Prior to ESI analysis of Glu-C digests, or MALDI-TOF MS analysis of intact globin, samples were desalted using a micropurification column containing 20-R2, or 50-R1, POROS reverse phase material, respectively. The micropurification column consisted of a Gel-Loader pipet tip prepared according to the method of Gobom et al. (17). The micropurification column was washed once with 70% ACN and 0.1% TFA and then equilibrated with 5% formic acid. Then 5-10 µL of digest, or 4 µL globin, was loaded onto the column. Peptide, or globin, was washed with 15 µL of 5% formic acid, while proteins had an additional wash of 15 µL of a mixture of 5% MeOH and 5% formic acid. Peptides, or globin, were eluted with 10 µL of a 50:45:5 (v/v/v) MeOH/ H2O/formic acid mixture into a 0.6 mL Eppendorf tube and stored at -20 °C prior to analysis. For some globin analyses, proteins were eluted with 3 µL of SA matrix and applied directly to the MALDI-TOF MS target in small spots.
Results DSF Globin Analysis by MALDI-TOF MS. Two new peaks in the chromatograph of globin from DSF-treated rats were collected and lyophilized prior to MALDI-TOF MS. The peak with a retention time of 44 min exhibited a prominent molecular ion with m/z 15 948.2 ( 0.3 (Figure 1A), while the peak with a retention time of 42 min exhibited a molecular ion with m/z 15 962.2 ( 0.6 (Figure 1Β). Both these species were 99.1 ( 0.6 Da higher in mass than β-2 and β-3, respectively. SA matrix adducts were observed about 200 Da higher in mass than the protein signals. DSF Globin and DTT. Incubations (2 h) of DSF globin in the presence of DTT (pH 8.5 or 10.5) at either 37 or 56 °C produced no apparent loss of the adduct from either β-2 or β-3, as judged by MALDI-TOF MS. The DSF globin spectra before and after incubation with DTT appeared to be identical (data not shown). Glu-C Digestion of Globin and Analysis by MALDITOF MS. After on-probe digestion for 1 h followed by mass analysis using MALDI-TOF MS, 20 peptide fragments could be assigned to peptides derived from β-3, for which the amino acid sequence was available in the data banks (Swiss Prot accession number P02091) (Table
240
Chem. Res. Toxicol., Vol. 13, No. 4, 2000
Erve et al.
Table 1. Observed Glu-C Peptides in DSF-Treated Rats and Control Globin
c
fragment
observed
predicted
∆Da
∆ppm
91-101a 80-96 27-43b 102-121 115-138 27-47 or 80-101 99-121 122-146 Mod[122-146]a 27-56c 48-79 91-121 57-90c 44-79 8-43 8-47 48-90 48-90d 102-146 Mod[102-146]a 44-90 27-73 27-79 or 48-101 27-90b 8-90
1306.5 1927.4 2098.5 2261.8 2467.9 2505.0 2603.2 2617.4 2716.4 3322.8 3363.1 3549.1 3743.1 3769.8 4007.1 4413.2 4561.1 4577.2 4859.8 4959.1 4967.7 5184.8 5848.9 7047.1 8955.1
1306.5 1927.1 2098.5 2261.7 2467.9 2504.9 2603.2 2617.1 2716.2 3322.8 3362.9 3549.2 3743.1 3769.3 4006.6 4413.2 4561.1 4577.3 4859.8 4958.9 4967.0 5185.0 5848.7 7047.1 8955.2
-0.01 -0.3 -0.02 -0.10 0.00 -0.18 -0.13 -0.25 -0.25 -0.02 -0.13 0.07 0.19 -0.49 -0.55 -0.25 0.13 0.10 -0.05 -0.22 0.64 0.13 -0.15 -0.01 -0.10
-9 -128 -10 -44 0 -70 -50 -96 -89 -6 -38 20 51 130 -137 -57 28 22 -10 -44 128 26 -26 -2 7
a DSF globin digest only. b Used as an internal standard. Control digest only. d Contains oxidized Met.
Figure 2. MALDI/TOF MS spectra of Glu-C digests of rat globin from DSF-treated rats (6 week DSF administration in feed at 1% w/w) (A) and control rat (B). Peptides are labeled with the corresponding sequence of the β-3 globin chain. Peptides [27-43] (m/z 2098.5) and [27-90] (m/z 7047.1, not in spectrum) were used for internal calibration. The peptides Mod[122-146] and Mod[102-146] have a 99 Da mass increment above those of their corresponding unmodified peptides. NI denotes not identified. Spectra were acquired on a PerSeptive Voyager Elite mass spectrometer as described in Materials and Methods.
1). These fragments had mass errors ranging from -0.55 to 0.13 Da and covered 100% of the amino acid sequence. Two peptide signals with high intensities were observed at m/z 2716.4 and 4959.1 in globin from a rat exposed to DSF (Figure 2A) that were not present in digests from control rat globin (Figure 2B). Both new peptides had molecular masses 99 Da higher than the peptides [122146] and [102-146] observed at m/z 2617.4 and 4859.8, respectively. Both of these peptides contain Cys-125 and were designated as the modified peptides, Mod[122-146] and Mod[102-146]. Peptides at m/z 1306.5, 1927.4, and 3549.1 were identified as fragments [91-101], [80-96], and [91-121], respectively, all of which contain Cys-93. None of these Cys-93-containing peptides could be paired
with a 99 Da heavier peptide. A small peptide at m/z 2467.9 could be assigned to Glu-C fragment [115-138], which also contains Cys-125. However, no peptide with a mass 99 Da higher than this could be observed. The control globin digest had many of the same Glu-C fragments as the DSF-treated globin, and the overall appearance was similar with a few exceptions. For example, the high-intensity peptide at m/z 4967.0 corresponding to [44-90] in the control spectrum was much smaller in the DSF-treated globin spectrum, possibly due to suppression. Peptide fragments seen in the control but not in the DSF-treated sample include [57-90] and [2756], while fragment [91-101] was seen in the DSFtreated, but not in the control, digest. Solution phase enzymatic digestion was performed to effect more complete digestion with the objective to generate more of the peptide Mod[122-146]. For the solution digests, fewer large peptides were observed by MALDI-TOF MS and the incompletely digested Mod[102-146] disappeared over time to generate Mod[122-146] whose abundance increased, as judged by MALDI-TOF MS (data not shown). Glu-C Digestion of HPLC-Purified Protein Fractions. Protein-containing fractions collected following separation by RP-HPLC were assessed by MALDI-TOF MS. The fractions corresponding to β-2 or β-3 isolated from control animals were then separately digested enzymatically on-probe with Glu-C. In both of these digests, no sign of peptide signals corresponding to Mod[122-146] or Mod[102-146] was observed. Similarly, fractions corresponding to adducted β-2 or β-3 from DSFtreated rats were also separately digested. In both digests, a signal at m/z 2716.4, corresponding to Mod[122-146], was observed, while no m/z 2617.4 was identified. However, although a signal at m/z 4959.1 corresponding to Mod[102-146] was observed in both digests from exposed animals, a small signal at m/z 4859.8 representing the corresponding unmodified peptide was also observed. Tandem MS of Glu-C Digests. The DSF-treated globin Glu-C digest was analyzed by nano-ESI. The peptide corresponding to Mod[122-146] ionized to produce a +4 ion at m/z 679.3. This peptide was selected by Q1 and subjected to collision-induced dissociation (CID) in the collision cell, resulting in extensive fragmentation. The resolution was sufficient to determine the charge states of the individual fragment ions. The resulting spectrum was deconvoluted using MaxEnt sequence software that converts multiply charged ion signals to singly charged ion signals (Figure 3). Following deconvolution, a nearly complete sequence of yn fragment ions could be identified, consistent with the sequence Thr-ProCys-Ala-Gln-(Ala)2-Phe-Gln-Lys-(Val)2-Ala-Gly-Val-AlaSer-Ala-Leu-Ala-His-Lys-Tyr-His (Table 2). The fragment ions at m/z 2369, 2466, and 2567 were shifted 99 Da higher than the predicted y22-y24 ions, respectively. These latter yn ions all contain Cys-125 near the Nterminus. The fragment at m/z 584.3 was consistent with either y4 or b4 shifted by 99 Da. Other shifted bn ions that were observed were b7-b9, b12-b14, b20, b23, and b24. The low-mass region of the spectrum did not contain any ion signals at m/z 72, 100, or 175 as observed previously in the tandem MS spectrum for GSH conjugates of DSF reactive metabolites (12), or the modified octapeptide previously reported for DETC-MeSO2-modified ALDH (14).
Disulfiram Metabolite Adduct Hemoglobin in Vivo
Chem. Res. Toxicol., Vol. 13, No. 4, 2000 241
Figure 3. Charge-deconvoluted tandem MS spectrum of the peptide designated Mod[122-146] in which its +4 charge state (m/z 709) was fragmented. The spectrum was acquired on a Micromass Q-TOF instrument equipped with a nano-electrospray source. The charge-deconvoluted spectrum exhibits singly charged sequence ions. The gap of 202 Da between y21 and y22 is caused by the adduct on Cys-125. Table 2. Observed and Predicted yn Fragement Ions Following CID of Mod[122-146] yn
amino acid
observed
predicted
∆Da
∆ppm
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
His Tyr Lys His Ala Leu Ala Ser Ala Val Gly Ala Val Val Lys Gln Phe Ala Ala Gln Ala Cys Pro Thr
156.09 319.10 447.23 584.30 655.34 768.44 na 925.48 997.52 1096.57 1153.61 1224.69 1323.72 1422.81 1550.91 naa 1826.02 1897.07 1968.11 2096.21 2167.23 2369.30b 2466.36b 2567.39b
156.08 319.14 447.24 584.30 655.33 768.42 839.45 926.49 997.52 1096.59 1153.61 1224.65 1323.72 1422.79 1550.88 1678.94 1826.01 1897.05 1968.08 2096.14 2167.18 2369.26b 2466.31b 2567.36b
0.01 -0.14 0.01 0.005 0.01 0.02 -0.01 -0.002 -0.02 -0.005 0.04 0 0.02 0.03 0.01 0.02 0.03 0.07 0.05 0.04 0.05 0.03
96.1 -128 -6.7 8.6 16.8 28.6 -10.8 -2.0 -21.9 -4.3 36.7 0 17.6 20.0 4.9 15.3 15.2 32.4 21.7 18.2 20.3 14.0
a
Not observed. b Contains adduct.
Tandem MS of DETC-Cys and Globin Acid Hydrolysates. The parent ion of authentic DETC-Cys at m/z 221.0 was subjected to CID to produce fragment ions at m/z 204.0, 175.0, 134.0, 100.0, 76.0, 74.0, and 72.0 (Figure 4A). The control hydrolysate produced many fragment ions with m/z values close to those seen in the standard, but ions at m/z 204.0, 100.0, and 76.0 were absent (Figure 4B). In addition, ions were observed at m/z 132.0, 118.0, 116.0, 90.0, and 86.0. Acid hydrolysates from globin obtained from treated rats (2 and 4 weeks) displayed all the ions observed in the standard (Figure 4C,D) and control globin hydrolysates.
Discussion The metabolism of DSF to the electrophilic metabolites that have been suggested to inhibit ALDH in vivo has been described in detail (2, 5, 18). The first step is the
Figure 4. Spectrum of DETC-Cys (m/z 221) acquired on a Micromass Q-TOF instrument equipped with a nano-ESI source as described in Materials and Methods. Spectra are from a standard (A), a control rat (B), a rat after the 2 week DSF administration (C), and a rat after the 4 week DSF administration (D). The fragment ion at m/z 100 corresponds to cleavage of the diethyl carbamoyl adduct and is observed in the standard and hydrolysates prepared from exposed rats (C and D), but is not in the control (B).
reduction of DSF (1) (Scheme 1) by GSH, or albumin, to DEDC (2) which can be methylated by several enzymes to DEDC-Me (3). Oxidation of the thiocarbonyl carbon by P450s of DEDC-Me generates DETC-Me (4). Further thioxidation of the sulfur can generate the sulfoxide, DETC-MeSO (5), or the sulfone, DETC-MeSO2 (6). Either 5 or 6 may react with Cys on GSH (7), or sulfhydrylcontaining proteins (8), to generate an N,N-diethylcarbamoyl Cys residue. In work reported here, following acid hydrolysis of intact globin, DETC-Cys (9) could be released and quantified. Following administration of DSF to rats, MALDI-TOF MS analysis of HPLC-purified proteins demonstrated the presence of two new protein species with molecular masses 99 Da higher than those of β-2 and β-3 that were not present in samples from control rats. The 99 Da molecular mass increase suggested the presence of a diethylcarbamoyl adduct (calculated as 99 Da) on Cys. Both of these globin chains contain a Cys-125 (19), which is a highly reactive nucleophile (20). Although β-1 also contains a Cys-125, no evidence for adduct formation was observed on this chain. This apparent lack of reactivity could be due to slight differences in structure that prevent accessibility of Cys-125 by the reactive metabolite. It is also possible, however, that low levels of adduction took place, but could not be detected. The
242
Chem. Res. Toxicol., Vol. 13, No. 4, 2000
major R-chain, and a slightly smaller protein that could represent the minor R-chain, were observed. However, no evidence of adduction on either of these chains was detected by MALDI-TOF MS. Although two Cys residues are present on the R-chain, they are both masked and hence not reactive (20). Others have reported the reversibility of the diethylcarbamoyl Cys adduct following incubation of adducted yeast ALDH with the reducing agent DTT at 37 °C (21). In our work, the adduct could not be removed with DTT (pH 8.5 or 10.5) after 2 h at either 37 or 56 °C as judged by MALDI-TOF MS analysis following incubation. Such an observation may be explained by some unique feature of ALDH, such as the three adjacent active site Cys residues (14), which make the adduct labile compared to the adduct on globin. Another factor may relate to differences in how the protein was handled prior to incubation with DTT. In our work, the globin had been precipitated and stored lyophilized for several months prior to analysis. To locate the adduct observed on β-2 and β-3, the proteins were enzymatically digested into smaller peptides. The endoproteinase Glu-C has been used successfully to digest a variety of β-chains (22) which at pH 7.8 in ammonium bicarbonate buffer cleave at the carboxy side of Glu (23). Adding Glu-C directly to the unfractionated globin on the MALDI-TOF MS probe (see Materials and Methods for details) resulted in extensive digestion of the β-chains with minimal digestion of the R-chains. This finding allowed mass analysis of the β-chains without the prior need to separate the individual subunits using chromatography. Following Glu-C digestion, two peptides were observed whose molecular masses were 99 Da higher than those of two peptide fragments known to contain Cys-125. The larger of these peptides was an incomplete digestion product in which one Glu-C cleavage site had not been cleaved. Although several β-chains had been digested simultaneously, because of sequence homology, many of the predicted Glu-C peptide fragments were identical, including the peptide containing Cys-125. It was apparent from the MALDI-TOF MS spectrum that these two modified peptides were of particularly high intensity even though they only represented a small fraction of the total peptides present. This fortuitous occurrence may be explained by the increased ionizing potential of the peptide following addition of the less polar diethylcarbamoyl moiety to the polar sulfhydryl group of Cys. Individual β-2 and β-3 chains were HPLC purified and digested with Glu-C to show that neither these chains gave rise to any modified peptides. Likewise, individual adducted β-2 and β-3 chains were purified prior to Glu-C digestion, and both modified chains gave rise exclusively to Mod[122-146] at m/z 2617.4 with no sign of unmodified [122-146]. Regarding the larger Mod[102-146], resulting from one missed cleavage, both of the modified chains gave rise to a peptide at m/z 4859.8, but a small amount of unmodified [102-146] was also observed (
