Evidence for a Free Radical Mechanism of Styrene-Glutathione Conjugate Formation Catalyzed by Prostaglandin H Synthase and Horseradish Peroxidase*

Transcription

1 THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 261, No. 34, Issue of December 5, pp ,1986 Printed in U. S. A. Evidence for a Free Radical Mechanism of Styrene-Glutathione Conjugate Formation Catalyzed by Prostaglandin H Synthase and Horseradish Peroxidase* (Received for publication, May 12, 1986) Beresford H. Stock$, Jorg Schreiber, Christian Guenat, Ronald P. Mason, John R. Bend, and Thomas E. Elinge From the Laboratories of Molecular Biophysics and Pharmacology, National Institute of Environmental Health Sciences, National Institutes of Health, Research Triangle Park, North Carolina We have proposed, using styrene as a model, a new mechanism for the formation of glutathione conjugates that is independent of epoxide formation but dependent on the oxidation of glutathione to a thiyl radical by peroxidases such as prostaglandin H synthase or horseradish peroxidase. The thiyl radical reacts with styrene to yield a carbon-centered radical which subsequently reacts with molecular oxygen to give the styrene-glutathione conjugate. We have used electron spin resonance spin trapping techniques to detect the pro- posed free radical intermediates. A styrene carboncentered radical was trapped using the spin traps 5,5- dimethyl-l-pyrroline N-oxide (DMPO) and t-nitrosobutane. The position of the carbon-centered radical was confirmed to be at carbon 7 by the use of specific H-labeled styrenes. The addition of the spin trap DMPO inhibited both the utilization of molecular oxygen and the formation of styrene-glutathione conjugates. Under anaerobic conditions additional styreneglutathione conjugates were formed, one of which was identified by fast atom bombardment mass spectrometry as S-(2-phenyl)ethylglutathione. The glutathione thiyl radical intermediate was observed by spin trapping with DMPO. These results support the proposed free radical-mediated formation of styrene-glutathione conjugates by peroxidase enzymes. The reaction of glutathione with epoxides catalyzed by the glutathione S-transferases (EC ) is an important step in both the formation of biologically active compounds and in the detoxication of chemical carcinogens (1). For example, glutathione S-transferases catalyze the formation of the potent inflammatory agent leukotriene C, from its epoxide precursor leukotriene & (2). In addition, glutathione S-transferases convert epoxides of chemical carcinogens to glutathione conjugates, an important mechanism for the inactivation of these reactive carcinogenic metabolites. Recently, using the model compound styrene, we have shown the formation of glutathione adducts catalyzed by peroxidases (3). The formation of styrene-glutathione conjugates was independent of epoxide formation and catalysis by glutathione S-transferase but required the initial oxidation of * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked aduertiement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. $ Visiting Scientist from the Pharmacy School, South Australian Institute of Technology, Adelaide, South Australia. To whom correspondence should be addressed glutathione to its thiyl free radical by the peroxidase and the subsequent reactions of this radical with styrene. As shown in Fig. 1, a free radical mechanism was proposed for glutathione conjugate formation based upon product analysis, incubation conditions, and the known chemistry of peroxidases. The peroxidase oxidized glutathione to its thiyl free radical which then adds to styrene. The styrene carbon-centered radical thus formed would then react with molecular oxygen to eventually yield the styrene-glutathione conjugate. The inclusion in the incubation mixture of reducing cofactors for the peroxidase (phenol or aminopyrine) greatly enhanced thiyl radical formation (4) and hence styrene-glutathione conjugate formation (3). However, direct evidence for the free radical intermediates involved in the styrene-glutathione conjugate formation was lacking. In this paper, we have used ESR spin trapping techniques to detect and characterize the proposed free radical intermediates in an effort to obtain additional evidence for the proposed mechanism. Also, we have isolated and characterized an additional styrene-glutathione adduct which is formed by the peroxidases under anaerobic conditions. EXPERIMENTAL PROCEDURES Materials Arachidonic acid (99% pure) was obtained from NuChek Preparations Inc., Elysian, MN. Aminopyrine, diethylenetriaminepentaacetic acid, glutathione, horseradish peroxidase (Type VI), hydrogen peroxide (30%), indomethacin, and t-nitrosobutane (t-nb ) were obtained from Sigma. Ring-labeled [ Clstyrene, 1.83 Ci/mol and 97% radiochemically pure, was obtained from Amersham Corp. [2-3H]Glycine-labeled glutathione, 5 Ci/mmol and 98% pure, was purchased from New England Nuclear. Unlabeled styrene (gold labeled), >99% pure, and [U- H] styrene (uniformly deuterated), 98% pure, were obtained from Aldrich. [8- H2]Styrene, 98.9% pure, and [7- H]styrene, 99.2% pure, were obtained from Merck. 5,5-Dimethyl-l-pyrroline N-oxide (DMPO), obtained from Aldrich, was used after purification on charcoal (5). Glacial acetic acid, ammonium hydroxide, and phenol were products of Mallinckrodt Chemical Works, and HPLC grade methanol, HPLC grade water, and n-hexane (certified) were purchased from Fisher. Anhydrous glycerol, analytical grade, and potassium phosphate, monobasic, were obtained from J. T. Baker Chemical Co. Hydrofluor was purchased from National Diagnostics, Somerville, NJ. The abbreviations used are: t-nb, t-nitrosobutane; DMPO, 5,5- dimethyl-l-pyrrcline N-oxide; FAB, fast atom bombardment; HEPES, 4-(2-hydroxyethyl)-l-piperazineethanesulfonic acid; HPLC, high performance liquid chromatography; MS, mass spectrometry; RSV, ram seminal vesicles; mt, millitesla; HPETE, hydroperoxyeicosatetraenoic acid.

2 15916 Free Radical Mechanism Glutathione for Conjugate Formation OH 4 00' FIG. 1. Proposed free radical mechanism for peroxidasecatalyzed formation of styrene-glutathione conjugates. Methods RSV Microsome Preparation-Ram seminal vesicles were obtained from a slaughterhouse and stored at -70"C until use. Tissue was thawed, trimmed of excess fat and connective tissue, and microsomes were prepared as previously described (6) with the exception that homogenization was carried out in 1.15% KC1 containing 10 mm HEPES/NaOH, ph 7.5, and 1% bovine serum albumin. Microsomes were resuspended in 1.15% KC1 containing 10 mm HEPES, ph 7.5. Protein determinations were made using the method of Lowry et al. (7). Microsomal prostaglandin H synthase activity was determined prior to each experiment by measuring oxygen incorporation into arachidonic acid with a Clark-type oxygen electrode (4). Incubation Conditions-Incubation mixtures consisted of 100 mm potassium phosphate, ph 7.6, with 1 mm diethylenetriaminepentaacetic acid, 100 pm [14C]styrene(0.5 pci), 1 mm glutathione, 100 pm arachidonic acid, and 1 mg of RSV microsomal protein (and, where appropriate, either 500 p~ phenol or aminopyrine) in a final volume of 1 ml. In some experiments [3H]glutathione(5 pci) was used. When H,O, replaced arachidonic acid it was present at 100 PM, and when horseradish peroxidase was used as the peroxidase it was present at a concentration of 20 pg/ml. The reaction mixtures were preincubated at 37 "C for 2 min in the absence of either arachidonic acid or H202. One of these substrates was then added to initiate the reaction, followed by incubation for 5 min. The reaction was stopped by immersion in an ethanol/dry ice bath. After hexane extraction to remove excess styrene the mixture was chromatographed on a ClS pbondapak column as described in HPLC analysis. Aerobic Versus Anaerobic Incubation Conditions-100 mm potassium phosphate buffer, ph 7.6, containing 1 mm diethylenetriaminepentaacetic acid was boiled to remove dissolved air. Aliquots were then allowed to cool, first to room temperature, then on ice, while being continuously bubbled either with argon or oxygen gas to saturate the buffers with the respective gas. Incubation mixtures consisted of the appropriate 100 mm potassium phosphate buffer (ph 7.6), 1 mm styrene, 5 mm glutathione, 1 mm H202, 500 p~ phenol, and 80 pg of horseradish peroxidase in a final volume of 4 ml. The phenol and horseradish peroxidase were added to individual 8-ml screw-cap tubes equipped with rubber septa and bubbled again with argon or oxygen. The tubes were then sealed under the appropriate gas, and [14C]styrene in methanol and glutathione dissolved in argon-gassed buffer were injected into the reaction mixture. After 2 min of preincubation at 37 "C the reaction was initiated by the injection of H202 diluted with degassed water through the rubber septum and then maintained at 37 "C for 5 min. The reaction was terminated by freezing in a dry ice/ethanol bath. The reaction mixture was left in sealed tubes until HPLC analysis was carried out by direct injection onto the column without a hexane extraction. Oxygen Uptake Assay-The utilization of molecular oxygen during peroxidase-dependent formation of styrene-glutathione conjugates was studied using a Clark-type oxygen electrode (Yellow Springs Instruments) attached to a Fisher Recordall Series Reactions were carried out in a 2-ml water-jacketed cell maintained at 37 "C and consisted of 10 mm potassium phosphate, ph 7.6, ["Clstyrene concentrations ranging from 25 to 200 pm, 1 mm glutathione, 100 pm H,O,, and 40 pg of horseradish peroxidase in the presence of either 500 p~ aminopyrine or 250 p~ phenol. The reaction was initiated by the addition of H202 after 2-min preincubation at 37 "C. The formation of styrene-glutathione conjugates under these conditions was confirmed by HPLC analysis of the incubation mixtures. HPLC Analysis-A Waters HPLC system consisting of a model U6K injector, two model 6000A solvent delivery pumps, a model 720 system controller, and a preparative C,, pbondapak steel column (7.8 mm X 30 cm) was used for HPLC analysis. A Flo-one p radioactive flow detector (Radiomatic Instruments and Chemical Co., Inc., Tampa, FL) linked with a Qume visual display and a C-Itoh 8510 printer was used to measure snd record the radioactivity in the eluant at 6-5 intervals. Hydrofluor was used as the scintillation fluid. Radioactivity was corrected for quenching and converted to dpm/fraction. The elution system for total styrene-glutathione conjugates consisted of 10 min of an isocratic phase of water followed by a 0-100% linear methanol/ water gradient over the next 30 min. Over the next 10 min the gradient was reversed from 100% methanol to 100% water. Where resolution of the products of anaerobic and aerobic incubations was required, the water phase was replaced by 0.01 M ammonium acetate, ph In both cases the flow rate was 2 ml/min. Electron Spin Resonance (ESR) Measurements-ESR spectra were recorded with a Varian E-109 spectrometer in a flat cell for aqueous solutions mounted in a TM,,, cavity. A modulation frequency of 100 khz was used, and all spectra were recorded with a microwave power of20 milliwatts. The scan range was set to 5.0 mt in incubations containing t-nb and 8.0 mt in incubations with DMPO. The scan time was 4 min, and the time constant was 0.5 or 1 s. Receiver gain and modulation amplitude are given in the legend to the figures. All investigations were carried out at room temperature in 10 mm potassium phosphate buffer, ph 7.6. The final concentrations of the reagents are given in the figure legends. Three ml of the incubation mixture were loaded into the flat cell by aspiration using a rapid sampling device (8). Mass Spectral (MS) Analysis by Fast Atom Bombardment (FAB)- The sample dissolved in 50% methanol/water was loaded onto a FAB target covered with a glycerol oxalic acid matrix. The oxalic acid was included to enhance the formation of the protonated molecular ion. Analyses were performed on a VGZAB 4F (VG Analytical, Manchester, England) tandem instrument of BE-EB geometry (magnetic:electrostatic-e1ectrostatic:magnetic). The xenon beam generated under FAB conditions gave a collision energy of 8 kev. The secondary ions were accelerated at 8 kev and the spectrum recorded on a DATA system PDP (VG Analytical). An MS-MS spectrum was measured by selecting the m/z 412 ion with MS1 (BE). RESULTS Horseradish Peroxidase-catalyzed Free Radical Formation- Styrene, glutathione, horseradish peroxidase, and HzOz were incubated at room temperature in 10 mm phosphate buffer (ph 7.6) containing 5.8 mm spin trap t-nb. Some incubation mixtures also contained aminopyrine, which acts as reducing a cofactor for the peroxidase and significantly increases the formation of styrene-glutathione conjugates (3). The reaction was initiated by the addition of HzOz, the incubation mixture was aspirated into a quartz flat cell, and the ESR spectrum recorded on a Varian E-109 spectrometer as described under "Methods." The ESR spectrum obtained from this incubation (Fig. 2A) shows the existence of a carbon-centered radical adduct which is characterized by its couplings an = 1.6 mt and uh = 0.37 mt and also the t-nb-hydrogen atom adduct with hyperfine coupling constants an = 1.46 mt and ah = 1.44 mt (9). However, the demonstration by ESR of a carbon-centered radical signal does not itself establish that this free radical arose from the styrene molecule. Therefore, the above incubation was repeated in the presence of uniformly deuterated [U-'H]styrene. Substitution of [U-'Hlstyrene for styrene gave an ESR spectrum (Fig. 2B) with the characteristic collapse of the hydrogen doublets into the unresolved deuterium triplets. This establishes that the carbon-centered radical is located on the styrene rather than on other molecules in the incubation mixture. To further localize the position of the carbon-centered free radical this experiment was repeated with styrene labeled in the C-7 position ([7-'H]styrene) or in the C-8 position ([8- 'H2]styrene). Incubation with the [8-'Hz]styrene gave a spectrum identical with nondeuterated styrene (Fig. 2A) while the spectrum with [7-'H]styrene was identical with that of uni-

3 Free Radical Mechanism for Glutathione Conjugate Formation mt B - AP - H2OZ - HRP showed a strong transient ESR signal due to the aminopyrine cation free radical which masks the presence of spin trapped species. This is seen in Fig. 2C which shows the signal obtained immediately after initiating the reaction in the absence of styrene. There is no evidence of a styrene carbon-centered radical, but there were a few large lines at the low field end of the spectrum which rapidly decayed and which we have assigned to the aminopyrine cation free radical (10). A much stronger aminopyrine cation radical signal was observed in the presence of styrene but in the absence of glutathione (Fig. 20). Moreover, the characteristic blue color of the radical was observed on visual inspection. The formation of the aminopyrine cation radical by peroxidases and its subsequent reduction by glutathione was reported previously (4). This is shown by the larger aminopyrine radical ESR signal in the absence of glutathione (Fig. 20). The presence of a reducing cofactor such as aminopyrine in the incubation increases the formation of the carbon-centered styrene free radical. In the absence of aminopyrine little styrene free radical was observed (Fig. 2E). No radical was detected in the absence of H202 or horseradish peroxidase (Fig. 2, F and G). The substitution of phenol for aminopyrine also results in an increase in formation of the styrene carbon-centered radical, confirming the importance of a good reducing substrate for the efficient formation of the styrene free radical. These results are in agreement with our previous finding which showed increased peroxidase-mediated styrene-glutathione conjugate formation in the presence of phenol or aminopyrine (3). Substitution of the t-nb by DMPO in the incubation mixtures containing phenol gave ESR spectra showing the presence of 3 radical species (Fig. 3). The rapidly decaying signal with un = 1.54 mt, and uh = 1.62 mt was assigned to the DMPO-glutathione thiyl radical adduct and indicates that the glutathione thiyl radical is an intermediate in the reaction sequence. Also spin trapped was a carbon-centered radical (an = 1.6 mt and uh = 2.24 mt) which was assigned to the styrene a-carbon radical. If styrene was omitted from the incubation no carbon-centered radical was detected (results not shown). A third species was also detected (the unmarked singlecenter line) which could not be identified. This radical had a FIG. 2. The t-nb spin trap system contained 10 mm potassium phosphate buffer, ph 7.6,5.8 mm t-nb, 1.3 mm styrene, 1.7 mm glutathione, 1.7 mm Hz~z, 1 mm aminopyrine, and 8 pg of horseradish peroxidase in a final volume of 3 ml. The reactions, at room temperature, were initiated by addition of the enzyme. The ESR spectrometer conditions were: modulation amplitude, mt; scan rate, 5.0 mt/4 min; time constant, 1 s (A) and 0.5 s (other spectra). Relative gain settings are given at B, D, and H. A, complete system (t-nb-styrene carbon-centered radical adduct with an = 1.6 mt, ah = 0.37 mt; t-nb one-electron reduction product with an = 1.46 mt, a" = 1.44 mt); B, as A with [U-'HI- or [7-'H] styrene instead of [U-'HI- or [8-'H2]styrene; (t-nb-['hlstyrene carbon-centered radical adduct with an = 1.6 mt, azh = 0.06 mt); C, as A without styrene; D, as A without GSH, (aminopyrine cation free radical (unassigned in aqueous solution)); E, as A without aminopyrine; F, as A without HzOz; G, as A without enzyme; H, as A with 200 PM phenol instead of aminopyrine. formly deuterated styrene (Fig. 2B). These results show unambiguously that the unpaired electron is localized on the a- carbon or C-7 position of styrene. The ESR spectra shown in Fig. 2, A and B, were recorded 5 min after initiation of the reaction because the initial spectrum obtained with incubations containing aminopyrine FIG. 3. The DMPO spin trap system contained 10 mm potassium phosphate buffer, ph 7.6, 50 mm DMPO, 2.5 mm styrene, 1.7 mm HzO,, 250 p~ phenol, and 4 fig of horseradish peroxidase in a final volume of 3 ml. ESR spectrometer conditions were: modulation amplitude, mt; scan rate, 8.0 mt/4 min; and time constant, 0.5 s. The spectrum consists of three radical species: a DMPO adduct with a styrene carbon-centered radical (an = 1.6 mt, ah = 2.24 mt (X)), the DMPO-glutathione thiyl radical adduct (an = 1.54 mt, ah = 1.62 mt (O)), and an unassigned free radical (unmarked center line) without hyperfine structure and with g =

4 15918 Free Radical Mechanism for Glutathione Conjugate Formation g factor of as determined with Fremy's salt and di-tbutyl nitroxide standards. It was not spin trapped by DMPO as it was clearly visible in the absence of the spin trap and apparently did not represent a by-product of the formation of the carbon-centered radical as its formation was dependent on the presence of glutathione, phenol, horseradish peroxidase, and H202, but not styrene. Prostaglandin H Synthase-catalyzed Free Radical Formation-Our earlier investigation indicated that the styreneglutathione conjugates formed by horseradish peroxidase were identical to those produced by the peroxidase activity of prostaglandin H synthase (3). As this latter enzyme represents a more important physiological system we have investigated the ESR signals obtained from incubation mixtures containing prostaglandin H synthase, arachidonic acid as cyclooxygenase substrate, either aminopyrine or phenol as reducing cofactor to facilitate the peroxidase activity, and styrene. The ESR spectrum obtained with the prostaglandin H synthase system and the spin trap t-nb is shown in Fig. 4A. This spectrum shows the carbon-centered styrene and hydrogen atom radical adducts identical to those found with the horseradish peroxidase system. An additional signal, indicated by the arrow, was also observed. This signal (also shown in Fig. 40 by arrows) with an = 1.85 mt is assigned to the t- NB-glutathione thiyl radical adduct. This assignment is in agreement with the results of Felix et al. (11) who found the t-nb-glutathione thiyl radical adduct to be formed during the photoreduction of hematoporphyrin by glutathione. This radical adduct is very unstable (see decay in Fig. 40, arrows). Fig. 4B represents the spectrum obtained with [7-'H]styrene, and as observed using horseradish peroxidase, a hydrogen doublet collapsed into a triplet. This confirmed the location of the carbon-centered radical at the a-carbon or C-7 position of styrene. The addition of indomethacin, an inhibitor of the cyclooxygenase activity of prostaglandin H synthase, abolishes all signals because of the lack of formation of the peroxide substrate for the peroxidase activity of the prostaglandin H synthase system (Fig. 4C). The omission of styrene (Fig. 40) gave the t-nb-hydrogen atom adduct and the t-nb-glutathione thiyl radical adduct described above but no carbon-centered radical. Also another fast decaying triplet (marked X) with an = mt was observed. This triplet is the ESR signal of the spin trapped product that gave the one line radical signal in Fig. 3 and was obtained from an incubation containing phenol, glutathione, horseradish peroxidase, and H202. Without a spin trap this system gave a signal with g = , but if t-nb was added a triplet with an = mt, identical with that in Fig. 40, developed. Attempts to further identify this radical were unsuccessful. It was established that the signal was not due to SO; (g = ) (12) or $0; (g = ) (13) and, as previously noted, was independent of styrene. Thus, it probably did not reflect part of the major pathway of styrene-glutathione conjugate formation but rather a secondary reaction involving phenol and glutathione. The omission of glutathione abolished all radical formation indicating that the formation of the thiyl radical precedes, and is essential for, the formation of the styrene carboncentered radical (Fig. 4E). The omission of phenol (Fig. 4F) and the use of boiled RSV microsomes (Fig. 41) both resulted in the production of only the hydrogen atom radical adduct formed by reduction of t-nb (3) which is unrelated to styreneglutathione conjugate formation. The omission of arachidonic acid resulted in a small carbon-centered styrene radical adduct (Fig. 4G) which is thought to be due to the presence of trace amounts of H202 formed from glutathione oxidation (14). C. - -r D. E. 0.5 mt Complete: AA RSV Phenol Styrene GSH tnb + lndo -Styrene - GSH -Phenol - AA +CAT Boiled RSV J. - tnb FIG. 4. The RSV microsomal t-nb spin trap system contained 10 mm potassium phosphate buffer, ph 7.6, 5.8 mm t- NB, 1.3 mm styrene, 1.7 mm glutathione, 220 b~ arachidonic acid, 250 p~ phenol, and 1.5 mg of RSV microsomal protein in a final volume of 3 ml. ESR spectrometer conditions were: modulation amplitude, mt; scan rate, 5.0 mt/4 min; and time constant, 0.5 s. A, complete system (showing the t-nb-styrene car- bon-centered radical adduct (an = 1.6 mt, ah = 0.37 mt), the t-nbhydrogen atom adduct as t-nb one-electron reduction product (an = 1.46 mt, ah = 1.44 mt), and the t-nb-glutathione thiyl radical adduct (first line marked with an arrow, an = 1.85 mt). B, as A with [U-'H]- or [7-'H]styrene instead of [U-'HI- or [8-'H2]styrene (an = 1.6 mt, azh = 0.06 mt). C, as A with 0.4 mm indomethacin; D, as A without styrene showing the t-nb-hydrogen atom adduct (an = 1.46 mt, ah = 1.44 mt), the t-nb-glutathione thiyl radical adduct (an = 1.85 mt, marked with arrows), and a t-nb adduct with an = mt (marked with crosses) which corresponds to the g = free radical in Fig. 3. E, as A without glutathione; F, as A without phenol; G, as A without arachidonic acid ( AA); H, as A without arachidonic acid but with 7500 units/ml catalase (CAT); I, as A without microsomes or with boiled microsomes; J, as A without t-nb; K, a8 A with 1 mm aminopyrine (AP) without phenol. This explanation was supported by the use of catalase (Fig. 4H), where catalase inhibited the formation of the small carbon-centered styrene radical adduct (Fig. 4G). No signal was observed in the absence of t-nb (Fig. 4J). The spectrum obtained by substituting DMPO for t-nb with the prostaglandin H synthase system is shown in Fig. 5. The existence of the decaying glutathione thiyl free radical and the development of the styrene carbon-centered radical as shoulders beside the main lines of the DMPO-glutathione thiyl radical adduct are seen in Fig. 5A. These small shoulders, which are identical to the carbon-centered radical species (X)

5 Free Radical Mechanism Glutathione for Conjugate Formation Complete: DMPO GSH Styrene AA RSV AP 6. + lndo u. - - GSH E. - AP F. - - Boiled RSV +Phenol - AP Indo+ 15- HPETE FIG. 5. The RSV microsomal DMPO spin trap system contained 10 mm potassium phosphate buffer, ph 7.6, 50 mm DMPO, 1.3 mm styrene, 1.7 mm glutathione, 220 p~ arachidonic acid, 1 m~ aminopyrine, and 1.5 mg of RSV microsomal protein in a final volume of 3 ml. ESR spectrometer conditions were: modulation amplitude, mt scan rate, 8.0 mt/4 min; and time constant, 0.5 s. Relative gain settings are given for C and G. A, complete system with the DMPO-glutathione thiyl radical adduct (an = 1.54 mt, ah = 1.62 mt) and the DMPO-styrene carbon-centered radical adduct visible as shoulders beside the four lines of the DMPOglutathione thiyl radical adducts. B, as A with 0.4 mm indomethacin; C, as A without styrene; D, as A without glutathione; E, as A without aminopyrine (AP); F, as A without arachidonic acid (AA), DMPO, RSV microsomes, or with boiled microsomes; G, as A with 250 p~ phenol instead of aminopyrine; H, as A with 0.4 mm indomethacin and 16 pm 15-HPETE instead of arachidonic acid. the presence of indomethacin. This demonstrates that the peroxidase activity of prostaglandin H synthase is responsible for the formation of these free radicals. The Involvement of Molecular Oxygen in the Formation of Styrene-Glutathione Conjugates-The mechanism proposed for conjugate formation (Fig. 1) requires the addition of OXYgen to the styrene carbon-centered radical to yield the styreneglutathione conjugates. To demonstrate a requirement for molecular oxygen by the peroxidase during synthesis of styrene-glutathione conjugates, incubations were saturated with either oxygen or argon. After incubation under oxygen or argon the reaction mixture was applied directly to a CIS pbondapak column and eluted with an 0.01 M ammonium acetate (ph 5.35)/methanol gradient as previously described (3). The results (Fig. 6) show that in the presence of oxygen there was a nearly quantitative conversion to a product that cochromatographed with authentic samples of the styreneglutathione conjugates previously described (3) (i.e. a mixture of (2s)- and (2R)-S-(2-phenyl-2-hydroxyethyl)glutathione). However, under an argon atmosphere several additional metabolites were observed with only minimal formation of these styrene-glutathione conjugates. The involvement of molecular oxygen in the reaction was further confirmed by measuring oxygen utilization with an oxygen electrode in the presence of various concentrations of styrene and either aminopyrine or phenol (Table I). Oxygen utilization was clearly dependent on the styrene concentration. Oxygen utilization in the absence of styrene reflects the oxidation of reduced glutathione (14, 15). The Influence of DMPO on the Reaction Sequence-According to our proposed mechanism, the presence of DMPO in the incubation mixture should inhibit oxygen uptake and styrene-glutathione conjugate formation. The addition of increasing concentrations of DMPO produced a concentrationdependent inhibition of oxygen uptake (Table 11). These results indicate that DMPO traps the carbon-centered and/ or thiyl radical, thereby competing with molecular oxygen and thus reducing its utilization. The addition of DMPO also inhibited styrene-glutathione conjugate formation as measured by HPLC analysis (Table 111). Inhibition was observed in the presence or absence of phenol and was dependent on the concentration of DMPO. Anaerobic Metabolite Formation-As shown in Fig. 6 the n in Fig. 3, are not observed in the absence of styrene (Fig. 5C). In the absence of aminopyrine a small DMPO-glutathione 6 - thiyl radical adduct signal was present (Fig. 5E). Replacing aminopyrine by phenol resulted in an increase in ESR signal n, I I intensity consistent with phenol producing styrene-glutathione conjugates more efficiently than aminopyrine (Fig. 5G). Elution Time (min) These results are in agreement with our previous studies FIG. 6. The HPLC profile of radioactive products generated indicating that glutathione is a substrate, although only a from incubation of [ Clstyrene with glutathione and horseweak one, for the peroxidase activity of prostaglandin H radish peroxidase under an atmosphere of oxygen (-) or synthase (15) but that thiyl radical formation and styrene- under an atmosphere of argon (---) as described under glutathione conjugate formation are enhanced by aminopyrine Methods. The products were injected directly onto a CIS pbondaor phenol. No detectable ESR signal was observed in the pak steel column (7.8 mm X 30 cm) and eluted with the 0.01 M absence of glutathione (Fig. 50), arachidonic acid, RSV mi- ammonium acetate (ph 5.3.9, methanol system as described under Methods. The elution profile under oxygen showed a major peak crosomes, DMPO, or in the presence of boiled microsomes (I) at 27.3 min with some residual styrene at 42.9 min. The elution (Fig. 5F). The identical ESR spectrum was observed upon profile under argon showed 4 major peaks at 26.2 (O), 27.3 (I), 31.6 replacing the arachidonic acid with 15-HPETE (Fig. 5H) in (31, and 33.8 (4) min. t B l5 i lot D ~ ~ D I S ~ ~ J O ~ U

6 With Free Radical Mechanism for Glutathione Conjugate Formation TABLE I The influence of styrene concentration on oxygen uptake in the presence of aminopyrine or phenol Values expressed as nmol of oxygen uptake/s. All values reported are mean rt S.D., n = 3. amino- Styrene con- With centration PM 0 25 pyrineb Incubation mixture phenol 3.3 rt rt rt rt f rt rt rt rt 2.9 Incubation mixture contained 10 mm potassium phosphate, ph 7.6, 1 mm glutathione, 100 p~ Hz02, 40 pg of horseradish peroxidase, and the concentration of [14C]styrene indicated in the table in a final volume of 2 ml. Aminopyrine concentration = 500 pm. Phenol concentration = 250 pm. TABLE I1 The influence of DMPO on oxygen uptake All values reported are mean -t S.D., n = 3. DMPO concentration Oxygen uptake Percentage inhibition of oxygen uptake from control system mm nmol/s rt rt rt rt f Control incubation system contained 10 mm potassium phosphate, ph 7.6, 200 pm styrene, 1 mm glutathione, 100 pm H202, 250 p~ phenol, and 40 pg of horseradish peroxidase in a final volume of 2 ml. mixture was chromatographed on a Cls HBondapak column with the 0.01 M ammonium acetate (ph 5.35)/methanol gradient. The results in Fig. 7 show the distribution of both 14C and 3H labels and indicate the formation of five distinguishable peaks containing both styrene and glutathione. The major 3H peak between 5 and 12 min represented the large excess of reduced and oxidized glutathione present. There was insufficient material present under the peaks at 26.2 (peak 0) and 30.2 min beak 2) to permit further investigation; how- ever, it was noted that the ratio of [3H]glutathione/[ 4C] styrene in the 26.2 min peak (peak 0) was double that in all other metabolites containing both labels. The peak eluting at 27.3 min (peak I) co-eluted with (2R)- and (2S)-S-(2-phenyl- 2-hydroxyethyl)glutathione, the major metabolites formed under aerobic conditions and identified previously (3). The material present in peak 3 at 31.7 min and peak 4 at 33.8 min, which corresponded to the major products formed under essentially anaerobic conditions (Fig. 6) was collected, evaporated to dryness at room temperature under reduced pressure, and further analyzed. The material obtained from peak 3 at 31.7 min was analyzed byfab using MS-MS techniques. Fig. 8 shows the mass spectrum of peak 3 and indicates that this metabolite, which is formed during anaerobic incubation, lacks the oxygen atom on the C-7 or a-carbon position of styrene. The structure is consistent with S-(2-phenyl)ethylglutathione. The major ion at 412 corresponds to the MH+ ion. The other ions are formed as indicated in the structure shown in Fig. 8. The peak at 33.8 min (peak 4) was clearly separated from the others, but isolation under reduced pressure proved un- TABLE 111 The influence of the radical spin trap DMPO on the formation of styrene-&tathione conjugates DMPO concentration mm Percentage inhibition of conju. gate formation Absence of phenolb Presence of phenolcd Control incubation system contained 10 mm potassium phosphate, ph 7.6, 100 pm [ Clstyrene, 1 mm glutathione, 100 pm H202, and 20 pg of horseradish peroxidase in a total of 1 ml. Conjugate formation in the absence of DMPO and phenol was 12.2 nmol/ml incubation. e Phenol concentration = 500 pm. Conjugate formation in the absence of DMPO but in the presence of 500 pm phenol was 81.3 nmol/ml incubation. Elution Time lminl FIG. 7. The HPLC profile of radioactive products generated from the incubation of [ Clstyrene (-) and [SH]glutathione (---) with horseradish peroxidase under the conditions described under Methods. absence of oxygen from incubation mixtures resulted in the formation of additional styrene metabolites. To investigate mle the chemical nature of these additional products, incubations containing both [14C]styrene and [3H]glutathione were per formed under conditions of high substrate concentrations and, FIG. 8. Mass spectrum of peroxidase-produced S-(2- therefore, potentially limiting oxygen concentrations. After pheny1)ethylglutathione. Spectrum was obtained using FAB and hexane extraction to remove excess styrene the resultant MS-MS techniques as described under Methods i

7 Free Radical Mechanism for Glutathione Conjugate Formation successful. Decomposition occurred with a variety of work-up procedures, and we were unable to elucidate the structure of this second product (peak 4) formed during anaerobic incubation. DISCUSSION The formation of styrene-glutathione conjugates is catalyzed by prostaglandin hydroperoxidase or horseradish peroxidase, and their formation is dependent on the presence of peroxide substrate, styrene, and glutathione (3). Two thioethers which are formed by glutathione addition to the C-8 or terminal carbon of styrene are present in equimolar concentrations and are identified as the diastereomers, (2s)- and (2R)-S-(2-phenyl-2-hydroxyethyl)glutathione. Their formation occurs in the absence of styrene epoxidation or transferase-catalyzed glutathione conjugation and is significantly enhanced by the presence of effective substrates for peroxidase enzyme activity such as phenol or aminopyrine. Evidence suggests (3) that thioether formation catalyzed by peroxidase proceeds by a free radical mechanism as outlined in Fig. 9. Peroxidases are known to oxidize glutathione to its thiyl radical (15, 16), and the ESR data presented in this paper indicate the presence of a thiyl radical in the incubation mixture containing peroxidase, styrene, and glutathione. Aminopyrine, which is a more efficient peroxidase substrate than glutathione, was preferentially oxidized to the aminopyrine cation radical (4). This radical subsequently oxidized the glutathione to its thiyl radical. The aminopyrine cation free radical was observed in the styrene incubation system, particularly in incubations without glutathione. An enhanced formation of thiyl radical in the presence of aminopyrine, and the other peroxidase substrate phenol, is consistent with enhanced styrene-glutathione conjugate formation in their presence. Although the addition of phenol to the incubation also significantly enhanced thiyl radical and styrene-glutathione conjugate formation, we did not detect the phenoxy radical due to its inherent instability. Nevertheless, the data obtained by ESR trapping experiments are consistent with the mechanism shown in Fig. 9 and indicate the intermediacy of a thiyl radical in styrene-glutathione conjugate formation by peroxidase. The addition of thiyl radicals including glutathione thiyl 14 ROOH \ radical to olefinic double bonds such as in styrene is well established (17-21). Reaction of a thiyl radical with styrene resulted in a carbon-centered radical which was detected with the spin traps DMPO and t-nb. The position of the free radical at C-7 of styrene was confirmed by the use of specifically *H-labeled styrene. The detection of the radical at the a-carbon (C-7) is in agreement with previous studies (3) and with the position of the resultant hydroxyl and glutathione groups on the isolated styrene-glutathione conjugates. Styrene carbon-centered free radical formation was dependent not only on peroxide, peroxidase, and styrene but also on the presence of glutathione in the reaction mixture. This, together with the pattern of formation and decay of the individual radicals, suggests that thiyl radical formation precedes styrene carbon-centered radical formation. The styrene carbon-centered radical would be expected to react with molecular oxygen to yield a styrene-conjugate peroxyl radical whichwould eventually be reduced to the hydroxyl group present in the thioether conjugate. The styrene peroxyl radical probably abstracts hydrogen from GSH to form the styrene-conjugate hydroperoxide and the glutathione thiyl radical (Fig. 9). This glutathione thiyl radical could add to styrene creating the possibility of a chain reaction, which would yield additional styrene-glutathione conjugate. The styrene-conjugate hydroperoxide could be reduced to the alcohol either by peroxidases or nonenzymatically by GSH (15). The spin trap DMPO inhibited the incorporation of molecular oxygen into styrene and reduced the formation of the styrene-glutathione conjugates as measured by HPLC. In the absence of molecular oxygen, the styrene carbon-centered radical undergoes a different reaction sequence than that observed in the presence of oxygen to yield additional styreneglutathione metabolites. We isolated and characterized by FAB MS-MS an additional metabolite as S-(2-phenyl)- ethylglutathione which could be formed via hydrogen abstraction from GSH by the carbon-centered radical. Glutathione thiyl radical might also react with the styrene carbon-centered radical to form a diglutathione styrene conjugate. The ratio of [3H]glutathione to [14C]styrene in metabolite peak 0 is twice that of the other styrene-glutathione conjugates which suggests that peak 0 is a diglutathione-styrene derivative, but oc:-ch /2 ROH t u ach=ch2- H H FIG. 9. A model for the formation of styrene-glutathione conjugates catalyzed by peroxidase and their increased formation in the presence of a peroxidase substrate. 4 ROO GSH PEROXIDASE Gs + vp H PEROXIDASE H H I;/*RoH or PhO@ GS H

8 15922 Free Radical Mechanism for Glutathione Conjugate Formation we were not able to characterize the metabolite further. The major styrene-glutathione metabolite formed under anaerobic conditions was unstable and also was not characterized. The formation of additional metabolites under conditions of oxygen depletion was noted in other studies (17, 18). In summary, we propose that two enzymatic mechanisms exist for the formation of styrene-glutathione conjugates. The first is the classical reaction sequence in which styrene is first converted to styrene 7,8-oxide which reacts enzymatically or nonenzymatically with glutathione. A second mechanism, shown in Fig. 9, is free radical in nature. Initially a peroxidasedependent thiyl radical reacts with styrene to yield a styrene carbon-centered free radical at position C-7. Under aerobic conditions, the carbon-centered radical reacts with molecular oxygen to form a styrene peroxyl radical which eventually produces the styrene-glutathione conjugate. In the absence of oxygen the carbon-centered radical undergoes other reactions to form additional metabolites, some of which are uncharacterized. Aminopyrine or phenol is also oxidized by the peroxidase to a free radical metabolite which is reduced by glutathione with the concomitant formation of the thiyl radical. By this mechanism enhanced thiyl radical and styreneglutathione conjugate formation are observed. Both the ESR experiments described here and identification of the stable styrene-glutathione conjugates formed strongly support the mechanism described in Fig. 9. REFERENCES 1. Oesch, F. (1973) Xenobwtica 3, Soderstrom, M., Mannervik, B., Orning, L., and Hammarstrom, S. (1985) Bwchem. Bwphys. Res. Commun. 128, Stock, B. H., Bend, J. R., and Eling, T. E. (1986) J. Biol. Chem. 261, Eling, T. E., Mason, R. P., and Sivarajah, K. (1985) J. Biol. Chem. 260, Buettner, G. R., and Oberley, L. W. (1978) Biochem. Biophys. Res. Commun Parkes, D. G., and Eling, T. E. (1974) Biochemistry 13, Lowry, 0. H., Rosebrough, N. J., Farr, A. L., and Randall, R. J. (1951) J. Biol. Chem. 193, Mason, R. P. (1984) Methods Enzymol. 105, Chignell, C. F., Kalyanaraman, B., Mason, R. P., and Sik, R. H. (1980) Photochem. Photobwl. 32, Lasker, J. M., Sivarajah, K., Mason, R. P., Kalyanaraman, B., Abou-Donia, M. B., andeling, T. E. (1981) J. Bwl. Chem. 256, Felix, C.C., Reszka, K., and Sealy, R.C. (1983) Photochem. Phtobiol. 37, Mottley, C., Harman, L. S., and Mason, R. P. (1985) Biochem. Phurmal. 34, Mottley, C., Trice, T. B., andmason, R. P. (1982) Mol. Phurmacol. 22, Harman, L. S., Mottley, C., and Mason, R. P. (1984) J. Biol. Chem. 269, Eling, T. E., Curtis, J. F., Harman, L. S., and Mason, R. P. (1986) J. Biol. Chem. 261, Harman, L. S., Carver, D. K., Schreiber, J., and Mason, R. P. (1986) J. Biol. Chem. 261, Oswald, A. A., Hudson, B. E., Rogers, G., and Noel, F. (1962) J. Org. Chem. 27, Oswald, A. A., Griesbaum, K., Thaler, W. A., and Hudson, B. E. (1962) J. Am. Chem. SOC. 84, Ito, O., and Matsuda, M. (1983) J. Org. Chem. 48, Ito, O., and Matsuda, M. (1984) J. Phys. Chem. 88, Ford, J. F., Pitkethley, R. C., and Young, V. 0. (1958) Tetrahedron 4,