Evidence for a One-electron Mechanism of 2-Aminofluorene Oxidation by Prostaglandin H Synthase and Horseradish Peroxidase*

Transcription

1 THE JOURNAL OF BIOLOGICAL CHEMISTRY Val. 259, No. 22, Issue of November 25, pp ,1984 Printed in U.S.A. Evidence for a One-electron Mechanism of 2-Aminofluorene Oxidation by Prostaglandin H Synthase and Horseradish Peroxidase* (Received for publication, May 3, 1984) Jeff A. Boyd and Thomas E. Eling From the Laboratory of Molecular Biophysics, National Institute of Environmental Health Sciences, Research Triangle Park, North Carolina Previous studies have shown that the primary arylamine carcinogen 2-aminofluorene (2-AF) is oxidized by the prostaglandin H synthase peroxidase to mutagenic and electrophilic products capable of covalent binding to macromolecules. The present study was designed to identify the potential reactive intermediate(s) responsible for binding, and to characterize further the metabolic intermediates in 2-AF peroxidation. Both prostaglandin H synthase and horseradish peroxidase, with H202, oxidize 2-AF to azofluorene, 2- aminodlfluorenylamine (2-ADFA), 2-nitrofluorene, polymeric and nonorganic-extractable material. Both enzymes show greater activity at ph 5.0 than at ph 7.0. In the presence of either 2-t-butyl-4-methoxyphenol or 2,6-dimethylphenol, arylamine/phenol adducts were formed in high yield, with the nitrogen of either 2-AF or Z-ADFA coupled to the para position of the phenol (loss of -0CHs with 2-t-butyl-4-methoxy- phenol). These structures were confirmed by mass spectrometry and NMR spectroscopy. Acid hydrolysis of N-hydroxy-2-AF to yield the nitrenium ion, in the presence of a phenol, also results in adduct formation, but only at times >2 h and in very limited yield. The peroxidase-catalyzed adduct formation, however is rapid (e2 min) and extensive. These and other data support a one-electron pathway for 2-AF peroxidation, with a free radical or a free radical-derived product responsible for binding to protein and DNA.An N- hydroxy intermediate may therefore not be obligatory in the enzymatic activation of 2-AF to a mutagenic product. A broad range of xenobiotics may serve as reducing cofactors for the peroxidase activity of prostaglandin H synthase, undergoing oxidation in the process. Comprehensive reviews summarizing particular substrates, tissues and reaction mechanisms have appeared elsewhere (1,2). A previous report from this laboratory described the oxidation of2-af, a model arylamine carcinogen, by prostaglandin H synthase in vitro (3). The radiolabeled substrate was metabolized to material * A preliminary report of this work was presented at the American Association for Cancer Research Annual Meeting, May 9-12, 1984, in Toronto, Ontario, Canada. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked aduertkement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. The abbreviations used are: 2-AF, 2-aminofluorene; azofluorene, 2,2 -azobisfluorene; azoxyfluorene, 2,2 -azoxybisfluorene; BHA, 2-tbutyl-4-methoxyphenol; BHT, 2,6-di-t-butyl-p-cresol; DCQI, 2,6- dichloroquinone-4-chloroimide; DMP, 2,6-dimethylphenol; di-bha, 2,2 -dihydroxy-5,5 -dimethoxy-3,3 -di-t-butylbiphenyl; HPLC, highperformance liquid chromatography covalently bound to tissue macromolecules, water-soluble products, and two organic-extractable metabolites, azofluorene and 2-nitrofluorene. This reaction was compared to that catalyzed by two other well-characterized peroxidases, horseradish peroxidase and chloroperoxidase. The former yielded products qualitatively identical to those from the prostaglandin H synthase reaction, while the latter enzyme gave primarily 2-nitrosofluorene. Furthermore, no evidence for an N-hydroxy-2-AF intermediate was found in the prostaglandin H synthase (or horseradish peroxidase) reactions. The conclusion was made that a one-electron mechanism resulting in free radical metabolites probably accounted for a majority of the 2-AF oxidation products. These results proved interesting given that the widely accepted mechanism for the metabolic activation of primary arylamines centers around N-hydroxylation by the cytochrome P-450-dependent mixed function oxidase system of the liver (4, 5). Subsequent conjugation, transport, and hydrolysis result in binding of an electrophilic species to bladder epithelial cell DNA in this model (6, 7). Presumably, these DNA adducts then lead to the bladder tumors observed in susceptible species, including dog (8) and man (6). Recently, however, alternative mechanisms of metabolic activation of aromatic amines, including prostaglandin H synthase, have received increasing attention (9). A prominent feature of the prostaglandin H synthasedependent oxidation of xenobiotics is the possible target tissue relevance of this reaction. Numerous extrahepatic tissues possess equal or greater amounts of prostaglandin H synthase than mixed function oxidase activity. Local metabolic activation of the bladder-specific arylamine carcinogens by this peroxidase is therefore a distinct possibility. Other studies have demonstrated that prostaglandin H synthase converts a number of aromatic amines, including 2-AF, to products mutagenic in the Ames test (lo), and to products which bind to DNA (11) and trna (12). If the prostaglandin H synthasedependent oxidation of 2-AF to mutagenic electrophiles occurs via a one-electmn mechanism, the heretofore presumed scheme involving obligatory N-hydroxylation of 2-AF for its activation to a carcinogenic intermediate (5) would require modification. Here, we examine further whether 2-AF oxidation by prostaglandin H synthase occurs via a one-electron mechanism. Direct electron spin resonance measurements of the primary free radical have been unsuccessful, owing undoubtedly to the extreme instability/reactivity of this particular radical species. Also, we provide further evidence that N-hydroxylation is not occurring to any appreciable extent in this system. The study of phenol adduct formation proved extremely useful in supporting both of the above points, since one-electron arylamine oxidation products would be expected to attack phenols

2 13886 Peroxidation of 2-Aminofluorene to produce adducts, while the N-hydroxy product clearly does not. In addition, the one-electron chemical oxidants potassium ferricyanide and Fenton s reagent are utilized in a comparative manner to further support the proposed mechanism of 2-AF oxidation by peroxidase. The similarities between the prostaglandin H synthase and horseradish peroxidase-dependent oxidation of 2-AF are also documented, thus allowing the latter model peroxidase to be used as a tool in delineating the metabolic pathways of 2-AF peroxidation. EXPERIMENTAL PROCEDURES~ RESULTS Oxygen Uptake during 2-AF Metabolism-The peroxidasedependent oxidation of phenylbutazone, a reaction known to incorporate molecular oxygen (18), results in an easily quantifiable oxygen consumption tracing, as shown in Fig. 1. Using horseradish peroxidase, approximately 60% of the total dissolved oxygen was incorporated over a s time course at ph 5.0 (Fig. LA); at ph 7.0, approximately 30% oxygen utilization was observed (Fig. 1B). Under these same conditions, the tracing observed during 2-AFoxidation was no different from that observed when H202 was omitted from the reaction, indicating little or no incorporation of molecular oxygen. That 2-AF oxidation was extensive, however,was evidenced by a marked inhibition of oxygen uptake by phenylbutazone when both substrates were included in the reaction mixture. Similar results were obtained using prostaglandin H synthase. As with horseradish peroxidase, greater metabolism 8OL A,,,,,,, 801,,,,,,, TIME {rc) TIME [roc) of phenylbutazone was observed at ph 5.0 (Fig. IC) than at ph 7.0 (Fig. ID). Traces for 2-AF incubation mixtures with and without HzOz were identical. Spectrophotometric Analysis of 2-AF Oxidation-The oxidation of 2-AF by horseradish peroxidase/hz02 at either ph 5.0 or 7.0 results in the rapid formation of a blue color. Fig. 2 shows the uv visible spectra of the disappearance of 2-AF and concurrent appearance of this blue chromophore, centered at approximately 600 nm. An isosbestic point is observed at 310 nm. Enzyme concentrations were varied from 25 ng/ml to 10 pg/ml, Hz02 concentrations varied from 10 to 400 p ~ 2-AF, concentration varied from 50 to 200 p ~ and, ph varied from 3.7 to 7.0. Both the rate of formation and the final intensity of the blue chromophore varied directly with enzyme, H202, and 2-AF concentration. As the ph of the reaction was raised, the rate and extent of blue chromophore formation decreased markedly. The titration of Hz02 in this reaction was carried out at ph 5.0. At an enzyme concentration which results in virtually instantaneous metabolism of 2-AF, the amount of blue color formed (Aeoo) was measured as a function of H2OZ added, using a fixed concentration of 50 p~ 2-AF. The blue chromophore formation saturated at approximately 50 p~ HzOz; the addition of greater concentrations of Hz02 did not increase the absorbance seen at 50 p~ Hz02. The oxidation of 2-AI? by the prostaglandin H synthase/ H202 system also resulted in the rapid formation of a dark color, although greater turbidity and a lesser rate and extent of metabolism made it difficult to distinguish a precise color. Analysis of this reaction by uv visible spectrophotometry was difficult for the same reasons; however, the disappearance of 2-AF and an isosbestic point at 310nmcould clearly be distinguished at ph 5.0. The concurrent formation of a chromophore in the visible region was shifted to approximately 650 nm. If the horseradish peroxidase/h2o2 reaction, ph 5.0, was run in the presence of 0.2 mg/ml heat-denatured solubilized prostaglandin H synthase, or if this protein was added to a maximal concentration of the bluecolor, the visible chromophore was shifted to 650 nm as well, and was clearly light green in color. When BHA was included in 2-AF reaction mixtures, the addition of HzOz resulted in the rapid formation of a bright pink color. The formation of this pink chromophore (centered around 520 nm) as generated by the prostaglandin H synthase system at ph 5.0 isshownin Fig., 3. The visible spectra.81c TIME (aoc) TIME (aoc] FIG. 1. Oxygraph tracings of 2-AF, phenylbutazone (PB), or 2-AF and phenylbutazone oxidation by peroxidase. A, horseradish peroxidase, ph 5.0; B, horseradish peroxidase, ph 7.0; C, prostaglandin H synthase, ph 5.0; D, prostaglandin H synthase, ph 7.0. Reaction conditions are described under Experimental Procedures. Arrows indicates addition of Hz02. e z 3.60 $.40 e a.20 Portions of this paper (including Experimental Procedures, Tables IV and V, and Figs. 5 and 6) are presented in miniprint at the end of this paper. Miniprint is easily read with the aid of a standard magnifying glass. Full size photocopies are available from the Journal of Biological Chemistry, 9650 Rockville Pike, Bethesda, MD Reauest Document No. 84M cite the authors, and include a check or money order for $3.20 per set of photocopies. Full size photocopies are also included in the microfilm edition of the Journal that is available from Waverly Press I I I I I i T WAVELENGTH inml FIG. 2. Optical spectra of 2-AF oxidation by horseradish peroxidase at ph 5.0. The reaction was initiated by the addition of HzOz, and spectra recorded every 2 s. Reaction conditions are described under Experimental Procedures. The arrows refer to growth or decay of the appropriate chromophore.

3 Peroxidation of 2-Aminofluorene I\\\\ \ I I 1 1 I WAVELENGTH (nm) FIG. 3. Optical spectra of 2-AF oxidation, in the presence of BHA, by prostaglandin H synthase at ph 5.0. The reaction was initiated by the addition of H202, and spectra recorded every 4 s. Reaction conditions are described under Experimental Procedures. The arrow refers to growth of the chromophore. generated by all four systems (either enzyme at either ph value) were very similar (+lo nm). The rate of formation and intensity of the pink color varied directly with the concentration of enzyme, substrates, and ph as described above for the oxidation of 2-AF alone. The inclusion of DMP in 2-AF reaction mixtures resulted in the similar formation of a violet color. In the presence of either BHT or 2-t-buty1-4-methylpheno1, the final visible spectrum was identical to that obtained from the oxidation of 2-AF alone, although it developed at a slightly slower rate. Product Analysis and Quantitation of r3h]2-af Metabolism-As quantitated by HPLC analysis, the prostaglandin H synthase/h202 and horseradish peroxidase/h202 systems both oxidize [3H]2-AF to organic-extractable and nonorganicextractable metabolites, at ph 5.0 and 7.0 (Table I). In the prostaglandin H synthase system, a slightly greater extent of total metabolism, as measured by recovered 2-AF, was observed at ph 5.0 (45%) than at ph 7.0 (34%), while in the horseradish peroxidase system, >90% metabolism was seen at both ph values. The prostaglandin H synthase-dependent reaction resulted in approximately 9 and 7% (at ph 5.0 and 7.0, respectively) of the total [3H]2-AF becoming covalently bound to tissue macromolecules. Nonorganic-extractable polymeric material was also formed in both enzyme systems. In addition to the 2-nitrofluorene and azofluorene products 1 identified in the previous study (3), organic-extractable products from both enzyme systems included a large number (10-12) of polymeric/oligomeric products migrating as dark, sometimes colored bands at the origin or with a low RF on preparative TLC plates. These products eluted together as a single, rather broad peak on HPLC analysis, and were quantitated together as organic-extractable polymer in Table I. Only one of these products was small enough to volatilize from the heated mass spectrometer probe, and was identified from its mass spectrum as a nitrogen to fluorene ring dimer, 2-aminodifluorenylamine. The blue color formed in the horseradish peroxidase system at ph 5.0 was rapidly and completely bleached upon the addition of 100 PM ascorbic acid. At a 10 pg/ml concentration of horseradish peroxidase, the complete development of the blue color was instantaneous. The subsequent addition of ascorbic acid resulted in the conversion of approximately 70% of the total radiolabeled material to 2-aminodifluorenylamine (RF = 0.35), as determined by TLC in diethyl etherln-hexane, 7:3. The remainder of the products isolated under these conditions consisted of polymeric material (RF = origin ), azofluorene (Rp = 0.90), and 2-AF (RF = 0.50). A striking ph effect was also observed in the horseradish peroxidase system, in regard to the qualitative nature of the products formed. At ph 7.0, approximately 50% of the total starting material was oxidized to azofluorene, while an equal amount of polymeric/oligomeric material (including 2-aminodifluorenylamine), and a little azofluorene, was formed at ph 5.0 (Table I). When BHA was included in [3H]2-AF reaction mixtures, the pink color formed was completely organic extractable and found to consist of two distinct 2-AF/BHA adducts. The structures of these adducts are shown in Fig. 4; the procedures for structural identification are described below. In the prostaglandin H synthase/h202 system (Table 11), the formation of adduct 2 predominates over that of adduct 1 at both ph values. As before, total metabolism and adduct formation were greater at ph 5.0 than 7.0. At ph 5.0, chromatographic analysis of the organic extract shows little or no evidence of the numerous 2-AF polymeric products observed when BHA was left out of the reaction mixtures. At ph 7.0, the only nonadduct product seen is azofluorene. A significant amount of primarily adduct 2 was still formed when BHA was added to reaction mixtures after 2-AF oxidation was essentially complete (as determined by 2-AF disappearance). The dark color TABLE I Prostaglandin H synthase and horseradish peroxidase-dependent 2-AF oxidation products All values are mean f S.D., n = 3, based on HPLC analysis. Organic-extractable products Non-organic extractable Incubation mixture 2-AF 2-Nitrofluorene Polymer Azofluorene Covalently Water bound soluble nmollincubation PHs f * 2.0 f f f f 2.1 PHsb & f f f & f 0.4 HRP 15.9 f f f f ? 1.1 HRPd 21.0 f f * f f 2.5 Reaction mixture contains: 100 mm sodium acetate buffer, ph 5.0, 0.2 mg/ml solubilized prostaglandin H synthase (PHs), 50 p~ [3H]2-AF(250 nmol total), and 100 pm H@z. * Reaction mixture contains: 100 mm potassium phosphate buffer, ph 7.0, 0.2 mg/ml solubilized prostaglandin H synthase, 50 PM [3H]2-AF(250 nmol total), and 100 p~ HzOZ. Reaction mixture contains: 100 mm sodium acetate buffer, ph 5.0, 1 pg/ml horseradish peroxidase (HRP), 50 p~ [3H]2-AF(250 nmol total), and 100 pm HzOZ. Reaction mixture contains: 100 mm potassium phosphate buffer, ph 7.0, 1 pg/ml horseradish peroxidase, 50 ~ L M [3H]2-AF(250 nmol total), and 100 PM Hz%

4 13888 Peroxidation of 2-Amirtofluorene ADDUCT 1 ADDUCT 2 ADDUCT 3 ADDUCT A CH3 FIG. 4. Proposed structures of 2-AFfphenolic adducts. Evidence for structural identification of these adducts is presented under Results. (green) rapidly changed to pink/violet when the BHA was added. Two adducts were also formed when DMP was included in [3H]2-AF reaction mixtures (Fig. 4). A slightly greater amount of the monomer adduct (adduct 3) than the dimer adduct (adduct 4) is observed with DMP, however. No adducts were formed when either BHT or 2-t-butyl-4- methylphenol was incubated with [3H]2-AF; the radioactive product chromatography profiles were identical to those obtained from reaction mixtures containing no phenolic compound. In the horseradish peroxidase/h202 system, roughly equal amounts of adducts 1 and 2 were formed when BHA was included in 2-AF reaction mixtures at ph 5.0 (Table 111). Again, the inclusion of BHA in reaction mixtures eliminated entirely the many polymeric products observed upon chromatographic analysis of organic extracts from 2-AF reaction mixtures not containing BHA. When BHA was added after complete development of the blue color seen at ph 5.0 (approximately 30 s), virtually all of the product isolated was adduct 2. At ph 7.0, adduct 2 formation was favored, and significant amounts of both adducts were seen when BHA was added after 30 s. Azofluorene was also formed at ph 7.0. The inclusion of DMP also resulted in the formation of both adducts; adduct 4 predominates at ph 5.0, while adduct 3 predominates at ph 7.0. If DMP is added to the reaction mixture at ph 5.0 after 30 s, the blue color rapidly changes to the violet color characteristic of adduct 4. As in the prostaglandin H synthase system, no 2-AF/ phenolic adducts were formed in the presence of BHT or 2-tbutyl-4-methylphenol. HPLC Analysis of 2-AF Oxidation Products-An HPLC system was developed to separate all 2-AF oxidation products and phenolic adducts described in this study. Their retention times are shown in Table IV. Oxidation of [ TIBHA and DMP by Peroxidase-Under conditions identical to those employed above for the generation of 2-AFlphenolic adducts (1 pg/ml horseradish peroxi- dase, 100 PM BHA, 100 p~ HZO2, 37 C), [14C]BHA was oxidized to a metabolite which co-chromatographed with di- BHA. At both ph 5.0 and 7.0, approximately 10% of the [ C] BHA was converted to di-bha. A small, yet measurable amount of metabolite(s) was nonorganic-extractable. Under those conditions previously reported (15) to oxidize [ *C]BHA to di-bha (200 pg/ml horseradish peroxidase, 500 p~ BHA, 500 pm H202, ph 7.0, 37 C), approximately 50% of the starting material was isolated as &-BHA. The oxidation of DMP to a yellow/green product, completely organic extractable, was also observed at both ph 5.0 and 7.0. This product has been previously identified (20) as a DMP dimer, 3,5,3,5 -tetramethyldipheno-4,4 -quinone. The reaction conditions for: the present study were identical to those described for 2-AF oxidation and adduct formation. Chemical Oxidation of 2-AF-The oxidation of 2-AF by potassium ferricyanide resulted in organic-extractable material consisting primarily of blackbrown polymeric products which remained at the origin or migrated with low RF on TLC plates, employing the chloroform/methanol (191) solvent system. These products eluted from the HPLC column in the region of organic-extractable polymer. A significant amount (5-1096) of azofluorene was also isolated and identified from this system. From reaction mixtures containing 2-AF and BHA, the products obtained were polymeric material similar to that just described, di-bha and a 2-AFIBHA adduct consisting of one 2-AF molecule and two BHA moieties. This structure (below) was confirmed by mass spectrometry (Mt = 507) and NMR spectroscopy. The oxidation of 2-AF by Fenton s reagent yielded four distinct products. Azofluorene, 2-aminodifluorenylamine, and 2-nitrosofluorene were isolated and identified by uv visible Reaction mixture TABLE I1 Prostaglandin H synthase-dependent Z-AF/phenolic adduct formation All values are mean f S.D., n = 3, based on HPLC analysis. products Organic-extractable Adduct Water Covalently 2-AF Adduct 1 Adduct 2 Adduct 3 soluble hound Non-organic extractable nml/incubutian BHA 64.0 f f f rt rt 0.1 BHA after 1 min 70.0 f f f f f 0.1 BHA _ f C f k 0.1 BHA after 1 mind C 2.8 f rt k f 0.2 DMP f f f f f 0.9 DMPf f f k f f 0.1 a Reaction mixture contains: 100 mm sodium acetate buffer, ph 5.0, 0.2 mg/ml solubilized prostaglandin H synthase, 50 p~ [3H]2-AF(250 nmol total), 100 p~ BHA, and 100 pm HZ% Same as a, except BHA added after 1 min of reaction. Reaction mixture contains: 100 mm potassium phosphate buffer, ph 7.0, 0.2 mg/ml solubilized prostaglandin H synthase, 50 p~ [3H]2-AF(250 nmol total), 100 p~ BHA, and 100 p~ H~OZ. Same as c, except BHA added after 1 min of reaction. e Reaction mixture contains: 100 mm sodium acetate buffer, ph 5.0, 0.2 mg/ml solubilized prostaglandin H synthase, 50 p~ [3H]2-AF(250 nmol total), 100 p~ DMP, and 100 pm H&Z. Reaction mixture contains: 100 mm potassium phosphate buffer, ph 7.0, 0.2 mg/ml solubilized prostaglandin H synthase, 50 p~ [3H]2-AF(250 nmol total), 100 p~ DMP, and 100 pm H202.

5 Reaction mixture Peroxidation of 2-Aminofluorene TABLE 111 Horseradish peroxidase-dependent 2-AF/phenolic formation adduct All values are mean k S.D., n = 3, based on HPLC analysis. Organic-extractable products 2-AF Adduct 1 Adduct 2 Adduct 3 Adduct 4 Non-organic extractable nmoijincubutian BHA" 2.5 f f * * 0.1 BHA after 30 sb 2.7 f k f f 2.7 BHA' 56.4 f & f 0.1 BHA after 30 sd 36.2 * * f * 1.6 DMP" 5.2 f C * * 0.2 DMPf 83.6 f f f f 0.6 Reaction mixture contains: 100 mm sodium acetate buffer, ph 5.0,l pg/ml horseradish peroxidase, 50 pm ['HI 2-AF (250 nmol total), 100 p~ BHA, and 100 PM HZ&. Same as a, except BHA added after 30 s of reaction. Reaction mixture contains: 100 mm potassium phosphate buffer, ph 7.0, 1 pg/ml horseradish peroxidase, 50 FM [3H]2-AF(250 nmol total), 100 p~ BHA, and 100 PM H202. Same as c, except BHA added after 30 s of reaction. e Reaction mixture contains: 100 mm sodium acetate buffer, ph 5.0,1 pg/ml horseradish peroxidaee, 50 JJM ['HI 2-AF (250 nmol total), 100 p~ DMP, and 100 p~ HzOZ. ' Reaction mixture contains: 100 mm potassium phosphate buffer, ph 7.0, 1 pg/ml horseradish peroxidase, 50 p~ [3H]2-AF (250 nmol total), 100 PM DMP, and 100 ~ b4 H&. "04 > OCH, 'C (CH3)3 spectrophotometry and mass spectrometry, and accounted for 10-20% each of the total products, depending on the conditions used. The remainder of the organic-extractable material consisted of polymeric/oligomeric products with the characteristic chromatographic properties. Approximately 10% of the starting material was recovered as unchanged 2-AF. Reaction of N-Hydroxy-2-AF with Phenols-Under conditions which generate the arylnitrenium ion from N-hydroxy- 2-AF (ph 5.0,37 "C), no 2-AFJBHA adducts were detected at time points of 2 min or 1 h. HPLC analysis of organic extracts from these reaction mixtures indicated the presence of only 2-AF, 2-nitrosofluorene, and azoxyfluorene, the breakdown products of the hydroxylamine. At the 4-h time point, however, incubation extracts were noticeably pink in color. HPLC analysis yielded a pink product which co-chromatographed with the 2-AF/BHA adduct 1. This product accounted for approximately 10-20% of the total 2-AF at ph 5.0 and ~ 5 % at ph 7.0. Similar results were obtained with DMP. No adducts were detected at time points of 2 min or 1 h, but small amounts of a violet product co-chromatographing with the 2-AFIDMP adduct 3 were detected at the 4-h time point, ph 5.0. No DMP adducts were observed at any time point at ph 7.0. When [3H]N-hydroxy-2-AF was included in enzymatic reaction mixtures containing a phenol, peroxidase, and H202, no color formation was observed with either BHA or DMP at either ph 5.0 or 7.0. HPLC and uv visible spectrophotometric analysis of organic-extractable material indicated only the presence of [3H]N-hydroxy-2-AF breakdown products, i.e. 2- AF, 2-nitrosofluorene, and azoxyfluorene. Structural Identification of Products-The mass spectral and uv visible spectral characteristics of 2-nitrofluorene and azofluorene were previously described in detail (3). The mass spectrum of 2-aminodifluorenylamine is shown in Fig. 5A. A molecular ion, also the base peak, was observed at m/e 360. The peak at m/e 180 is from either fragment resulting from cleavage of a nitrogen to fluorene ring bond. Cleavage of the other fluorene to nitrogen bond results in a fluorene moiety (m/e 165) and a diaminofluorene fragment (m/e 195). Although the coupling of nitrogen to fluorene is most likely at carbon-7 as drawn, this spectrum only supports a nitrogen to fluorene ring bond. A bond at another carbon atom, especially C-1 or C-3, cannot be ruled out solely upon interpretation of the mass spectrum. The uv visible spectrum of Z-aminodifluorenylamine has an absorption maximum in ethanol at 359 nm, with a shoulder at 290 nm. The mass spectra of the Z-AF/phenolic adducts are shown in Fig. 5, B-E. For the 2-AFIBHA adduct (adduct l), a molecular ion, also the base peak, was observed at m/e 327. This was confirmed as corresponding to C,HzlON by the exact mass measurement. C~~HZON Observed mass: Calculated The fragments at mje 312, 298,285, and 272 (M - 15, M - 29, M - 42, and M - 55, respectively) appear to arise from the BHA moiety, since identical fragments of the same relative intensity appear in the spectrum of the DCQIJBHA standard, described under "Experimental Procedures." The peak at m/e 165 results from cleavage of the nitrogen to fluorene bond with charge retention on the fluorene ring, while the peak at m/e 179 results from cleavage of the nitrogen to phenolic bond, with charge retention on the nitrogencontaining fragment. The mass spectrum of the 2-iminodifluorenylamine/BHA adduct (adduct 2) is shown in Fig. 5C. A molecular ion, also the base peak, is observed at m/e 506; the peak at m/e 508 most likely results from reduction of the quinonoid ring by the heated mass spectrometer probe; temperatures approaching 400 "C were required to volatilize the 2-iminodifluorenylamine/phenolic adducts. Such M + 2 peaks are commonly seen with other quinonoid compounds (21), including the benzidine/bha adduct (22). Additional evidence for the quinonoid as opposed to phenolic adduct structure (or a mixture of the two) was obtained by chromatographic analysis. The purified adduct gives a single, sharp, and intensely colored

6 13890 Peroxidation of 2-Aminofluorene peak (or band) upon analysis by two TLC systems and HPLC. Also, the absence of any spectral change at high PH, as with indophenols, was further evidence for the absence of any phenolic function in the molecule (23). The peak at m/e 358 resuits from cieavage of the nitrogen to quinonoid bond, with charge retention on the nitrogen-containing fragment. Peaks at d e 180 and m/e 165 result from cleavage of the nitrogen to either fluorene bond, with charge retention on the nitrogencontaining fragment, and fluorenyl fragment, respectidy. The Z-AF/DMP adduct (adduct 3) gives the mass spectrum shown in Fig. 50. A molecular ion and base peak was observed at rnle 299 (with M + 2, as explained above, at nle 301). The fragments at m/e 284, 256, and 227 (M - 15, 43, and 72, respectively) appear to arise from the DMP portion of adduct, since identical fragments of the same relative intensity are observed in the spectrum of the DCQI/DMP adduct standard. The peak at m/e 165 results from cleavage of the nitrogen to fluorene bond with charge retention on the fluorene fragment, while the peak at m/e 179 results from cleavage of the nitrogen to phenolic bond with charge retention on the nitrogencontaining fragment. The mass spectrum of the 2-iminodifluorenylamine/DMP adduct (adduct 4) is shown in Fig. 5E. The molecular ion and base peak at m/e 478 is again accompanied by the M -k 2 reduction product peak. The peak at m/e 344 results from cleavage of the nitrogen to nuorenyl bond with charge retention on the difluorenylamine fragment. Finally, the peak at m/e 165 results from cleavage of the nitrogen to fluorene bond with charge retention on the fluorenyl fragment. Some additional evidence for the proposed structures of the 2-irninodifluorenylaminejphenolic adducts, as well as for the 2-aminodifluorenylamine molecule itself, was obtained by proton NMR spectral analysis of adducts 2 and 4. AIthough the resulting spectra were extremely complex, several conclusions could be drawn. The data are summarized in Table V, while the spectrum of adduct 2 is shown in Fig. 6. For the 2- iminodifluorenylamine/bha adduct (adduct 2), the presence of one or more 2-AF moieties as well as a BHA molecule was indicated by the fluorenyl and t-butyl protons. The presence of two t-butyl. groups and C-6' quinoid protons is consistent with the presence of cis-trans isomers of the iminelphenol adduct, due to the nonlinear central C-N-C bond. The similar structure of the DCQIJBHA adduct, and its NMR spectrum, have been studied and described in detail (24). The presence of the secondary amine group between the two fluorene rings was also evident. Support for these assignments was gained by their good correiation with standard spectra from 2-aminofluorene and the DCQI/BHA adduct standard, the protons of which were easily assigned. The 2-imindifluorenylamine/DMP adduct (adduct 4) yielded a spectrum that was-interpreted similarly, however, the protons from only one isomer have been assigned. The presence of aromatic protons from the fluorene moiety(s) and methyl protons from the quinoid moiety were both evident, correlating well with their corresponding standard spectra. The broad peakfrom the secondary amine proton was also present. either as an intermediate or final product. The horseradish peroxidase/h202 system gave qualitatively identical results. Primary aromatic amines and diamines are very efficient cofactors for the reduction of Hz02 by horseradish peroxidase. The extensive metabolism studies of Saunders (for review, see Ref. 251, together with the mechanistic studies of Chance (261, George (27L and Yamazaki (28, 29) have clearly established the now well-known scheme: Peroxidase + HzOZ compound I Compound I + AH2 - compound II + AH' Compound I1 + AH2 peroxidase + AH' The resulting free radicals may react with themselves or other molecules in any number of ways, including: 2AH' -.) AH - AH ZAH -+ A + AH2 The focus of the present work was to test the hypothesis that prostaglandin H synthase and horseradish peroxidase oxidize 2-AF by a similar one-electron mechanism, and to further examine the possible involvement of N-hydroxy-2-AF as an intermediate in the activation of 2-AF to a mutagenic electrophile by prostaglandin H synthase. Since H202 was found in the previous study (3) to initiate the same reaction yielding the same products as arachidonic acid (which is metabolized to the lipid hydroperoxide, prostaglandin G2, by the prostaglandin H synthase cyclo-oxygenase activity), HzOz was used throughout this study to facilitate a more direct comparison between enzymes and simplify the reactions involved. Electron spin resonance spectroscopy of 2-AF oxidation by either enzyme gave only a broad, single line spectrum characteristic of the synthetic melanins (data not shown). This signal was most likely the result of a polymer radical(s) produced by the rapid and extensive coupling of primary oxidation products, and lacked the hyperfine structure necessary for resolution. Other experimental approaches were clearly necessary to support the hypothesis of a free radical mechanism of 2-AF peroxidation. The transient blue color observed upon 2-AF oxidation by horseradish peroxidase/hzoa was suggested in our previous study to be a charge transfer complex of some form. This hypothesis was based upon the possible analogy with a blue benzidine charge transfer complex, existing in equilibrium with the cation radical, produced in peroxidase systems (30, 31). The current study suggests that this is not, however, the case with 2-AF. First, no spectral shift upon dilution of the blue color, a distinguishing characteristic of charge transfer complexes, was observed. Second, a maximal concentration of the blue chromophore could not be altered or decreased by increasing concentrations of H202, suggesting that further oxidation was not possible. Several lines of evidence, as follows, suggest that the blue color is due to absorbance by a diimine intermediate with the following structure: DISCUSSION In our previous study of 2-AF oxidation by prostaglandin H synthase, the peroxidase activity of this enzyme complex was found to rapidly oxidize 2". The majority of the products observed, including azofluorene, water-soluble polymeric material, and material covalently bound to protein, were consistent with a one-electron oxidation mechanism. Furthermore, no evidence could be found for N-hydroxy-2-AF, & N A N " First, if the blue compound(s) is reduced with ascorbic acid before extensive polymerization occurs, the major product isolated is 2-aminodifluorenylamine, the expected reduction product of the above structure. Second, if either BHA or DM '

7 is added in a similar manner to a maximal concentration Of the blue color, 2-iminodifluorenylamine/phenolic adducts are the major products isolated. An electrophilic diimine compound such as that shown above would react readily with phenols to produce adducts, in a manner similar to that of the Gibbs reagent (DCQI). Indeed, the reaction of the benzidine diimine with phenols produces similar adducts which have been thoroughly characterized (22, 32). Third, titration experiments show that the blue color formation saturates at a concentration of H202 in the range of 1 molar equivalent for each mole of 2-AF present in the reaction mixture. This is the stoichiometry expected if all of the 2-AF in the reaction mixture were oxidized to the proposed 2- iminodifluorenylimine. Finally, an analogous structure accounts for a transient blue color observed during aniline oxidation by horseradish peroxidase (33). A quantitative analysis of the stable [3H]2-AF oxidation products also provides evidence for a one-electron mechanism, similar with both enzymes. In addition to the azofluorene identified in the previous study, a second 2-AF dimer, 2- aminodifluorenylamine, was isolated and identified. Both of these products are characteristic of radical dimerization; nitrogen to nitrogen dimerization would produce hydrazofluorene, which may rapidly autoxidize to azofluorene, while nitrogen to carbon dimerization would produce 2-aminodifluorenylamine directly. It should be re-emphasized that the exact location of the N-C bond cannot be determined from this data; C-7, C-1, and C-3 are the most likely sites for coupling to N. It is conceivable that azofluorene could be formed directly by the coupling of two nitrenium ions, but the spontaneous coupling of two positively charged nitrogen atoms is extremely unlikely. Control experiments in the previous study (3) also ruled out the formation of azofluorene by condensation of N-hydroxy-2-AF or 2-nitrosofluorene with the parent amine. Dimerization of aromatic amine radicals in the N-C manner has been documented in the peroxidative oxidation of aniline (33), 4-phenyl-2,6-dimethylaniline (34), and substituted anilines (with elimination of the para substituents) (35,36), as well as in th electrochemical oxidation of aniline (37). The electrochemical oxidation of 2-AF in acetonitrile has also been studied (38). The hypothesis was made that the resulting amine radicals couple to produce 2-aminodifluorenylamine (with the N-C bond at either C-1 or C-3). This intermediate is then further oxidized to a dark green diimine compound analogous to the product shown above. Preliminary evidence also indicates that p-phenetidine is oxidized by peroxidase to a nitrogen to ring dimer, with elimination of the para-ethoxy group (39). N-N coupling of arylamine radicals to produce azo-compounds has been documented in the peroxidative oxidation of 4-chloroaniline (35), o-dianisidine (40), 3,5,3',5'-tetramethylbenzidine (30), and benzidine (31). It is probable that the formation of azofluorene was greatly favored in the horseradish peroxidase system at ph 7.0 because of the deprotonated amine radicals at this ph. The product profile shifted in favor of 2-aminodifluorenylamine/polymer at ph 5.0, most likely reflecting the unlikely N-N coupling of two charged cation radicals. This ph difference was not as dramatic in the prostaglandin H synthase system. One final observation from the quantitative metabolism studies involves the extent of metabolism and its relationship to ph. Although a greater rate and total extent of metabolism of amine substrates is neither novel nor unexpected in the horseradish peroxidase system at acidic ph, relative to neutral ph, similar results with the prostaglandin H synthase peroxidase activity in this study proved interesting. Studies measuring oxygen incorporation into phenylbu- Peroxidation of 2-Aminofluorene tazone, 2-AF disappearance, and 2-AF product formation (including covalently bound material) all indicated that both the rate and total extent of metabolism were greater at ph 5.0 than at ph 7.0. To our knowledge, this is the first report of greater prostaglandin H synthase peroxidase activity, in regard to a particular substrate, at ph 5.0 than at neutral ph. The prostaglandin H synthase-dependent metabolism of benzidine at ph 5.0 was demonstrated (41), but no direct comparison was made with the same reaction at neutral ph. A scheme which accounts for a one-electron oxidation pathway leading to the formation of 2-AF metabolites, and consistent with the above data, is shown in Fig. 7. First, a one- electron abstraction by the peroxidase produces the 2-AF cation radical; deprotonation would then yield the uncharged radical species. This radical may then couple, as discussed above, to produce the stable products azofluorene or %aminodifluorenylamine, or initiate radical polymerization to produce the product(s) evident in static electron spin resonance studies. This radical may also attack macromolecules, even- tually resulting in covalently bound adducts. Since the steadystate concentration of the primary free radical would not be expected to reach the level necessary in order to serve as a further substrate for the peroxidase, and a passive loss of the second electron is unlikely, formation of the 2-AF nitrenium ion appears unfavorable. Small amounts of this reactive molecule could be formed, however, by oxygen abstraction of the second electron (to yield O;), or by disproportionation of two radicals to produce the parent amine and a nitrenium ion. It is probable that 2-aminodifluorenylamine may also serve as a substrate for peroxidase. Indeed, the analogous aniline dimer, 4-aminodiphenylamine, is readily oxidized by horseradish peroxidase to the same products as aniline (31). Furthermore, the hypothesized 2-aminodifluorenylamine intermediate observed during the electrochemical oxidation of 2- AF was shown to be more easily oxidized than 2-AF (38). An unambiguous synthesis of 2-aminodifluorenylamine, in order to test this hypothesis directly, has to this point proven unsuccessful. Based on the evidence presented above for the formation of 2-aminodifluorenylamine and its corresponding imine, however, it is difficult to envision a more likely pathway. The one-electron oxidation of 2-aminodifluorenylamine would produce the free radical, which could initiate radical polymerization or bind to macromolecules. Unlike the monomer amino radical, however, the abstraction of a second electron from the dimer radical would lead to a more stable intermediate, due to the additional electronegative nitrogen atom present on the fluorene ring. The diimine compound could thereby be formed, subsequently reacting with macromolecules to give covalent adducts or polymerizing to form the observed dark precipitate. Note that the diimine of 2- aminodifluorenylamine may exist in equilibrium with the nitrenium ion. It is still unclear how 2-nitrofluorene (2-13% of the total products) is formed in these two peroxidase systems; a oneelectron mechanism for its formation, although complicated, is conceivable. Small amounts of a 2-AF nitrenium ion could be formed, adding water to produce the hydroxamic acid. This compound could then proceed through a series of oxidations (4 electrons total) to 2-nitrofluorene. Results in our previous study (3) suggested that N-hydroxy-2-AF and 2-nitrosofluorene could be rapidly oxidized by peroxidase to 2-nitrofluorene. The oxygen uptake studies are consistent with a oneelectron pathway for 2-AF oxidation, since no molecular oxygen is incorporated, even though 2-AF metabolism is extensive. The formation of 2-nitrofluorene via the above hypothesized one-electron pathway, utilizing water as the

8 13892 Peroxidatwn of 2-Arninofluorene &NSO NITROFLUORENE / 2-AMINODIFLUORENYLAMINE I FIG. 7. Proposed pathway for the oxidation of 2-AF by prostaglandin 3 synthase or horseradish peroxidase. A detailed description of this scheme is presented under "Discussion." Ar-NH2, either 2-AF or 2- aminodifluorenylamine. source of at least one oxygen atom, is also consistent with the oxygen uptake data. In our previous study (3), we reported that the inclusion of BHA in 2-AFIperoxidase reaction mixtures resulted in the rapid formation of a pink color, and hypothesized that this was a 2-AF/BHA adduct. The results of the present study indicate that this color actually consists of at least two adducts, 2-AFIBHA and 2-iminodifluorenylamine/BHA. Analogous adducts are formed when DMP is substituted for BHA. The proposed structures of these adducts (Fig. 4) are supported by mass spectrometry, NMR spectroscopy, and comparison with similar adducts formed by DCQI (see "Experimental Procedures") and benzidine (22, 32). The location of the nitrogen to carbon bond at the para position of the phenolic compound is particularly interesting with the BHA adducts, since spontaneous loss of the para-methoxy group is required. Similar reactions have been documented with the so-called ipso attack of hydroxyl radicals on methoxylated phenols (42), and in the coupling of an anilo radical with a substituted phenol radical to produce adducts analogous to those described in the present study (43). Consistent with these observations are the results we obtained when BHT or 2-t-butyl-4-methylphenol was substituted for BHA or DMP in 2-AF reaction mixtures. No adducts are formed with either phenolic compound, which is the expected result since the paru-methyl function is a poor leaving group, and the meta position is unactivated in the phenolic ring system. A scheme accounting for the formation of these phenolic adducts, and consistent with the one-electron chemistry of 2- AF outlined above, is detailed in Fig. 8. If either BHA or DMP is present in 2-AF reaction mixtures, the 2-AF radical may attack a phenol molecule, producing an adduct radical. Further one-electron oxidation of this radical would result in the stable adduct structures shown (adducts 1 and 3). In the case of BHA, simultaneous loss of the methoxy anion during 2-AF radical attack would lead to the quinonoid adduct structure directly. In the case of DMP, since loss of a hydride ion is extremely unlikely, further oxidation of the phenolic form of the adduct is required. Although further oxidation by the peroxidase is possible, oxidation by 2-iminodifluorenylimine, yielding 2-aminodifluorenylamine and the quinonoid form of the adduct is likely. The 2-aminodifluorenylamine radical may also attack the phenols, resulting in adducts 2 and 4 through a mechanism exactly as described for the 2-AF radical. The diimine form of 2-aminodifluorenylamine may also react with the phenols, producing adducts 2 and 4 directly, as shown in Fig. 8. Like aromatic amines, phenols are also known to undergo one-electron oxidation by peroxidase (44). For BHA in particular, oxidation by horseradish peroxidase led to the formulation of &-BHA, via the BHA free radical (45). The same BHA product was obtained from oxidation by prostaglandin H synthase (15). Since the oxidation conditions employed in previous studies were much stronger than those used in the

9 Peroxidation of 2-Aminofluorene Adducts 2 and 4 Adduct 1 I Adduct 3 Adduct 4 FIG. 8. Proposed pathway for the formation of phenolic adducts by 2-AF radical products. A detailed description of this scheme is presented under "Discussion." R, fluorene. current study, experiments were carried out with [I4C]BHA to determine whether di-bha was produced under the same conditions used for adduct formation in this study. Since these experiments indicated that BHA (and DMP) were readily oxidized, the possible effect of the competitive oxidation of the phenol with 2-AF in this study must be considered. In the absence of a phenol, greater than 90% 2-AF metabolism is observed at both ph values in the horseradish peroxidase system (Table I). When either BHA or DMP is included, however, significantly greater amounts of unchanged 2-AF are recovered at ph 7.0 than at ph 5.0 (Table 111). This is the result expected if the phenols were competing with the amine for oxidation at the higher ph (where more of the phenol would exist in an ionized, therefore more easily oxidized, state). This hypothesis is supported by the data obtained from the potassium ferricyanide-catalyzed oxidation of 2-AF in the presence of BHA (at ph 10.0). The formation of only the 2- AF/di-BHA adduct most likely reflects the attack of a 2-AF radical on a molecule of di-bha, formed first by the preferential oxidation of BHA at this ph. This phenomenon is also apparent in the prostaglandin H synthase system (Table 11). Although a one-electron mechanism of 2-AF oxidation was consistent with the results reported above, it remained to be established whether or not an N-hydroxy-2-AF intermediate was the reactive species involved. The previously established theory of arylamine oxidation to reactive electrophiles requires N-hydroxylation, followed by acid-catalyzed hydrolysis to a nitreniurn ion (6). This ultimate electrophile then reacts with DNA to produce adducts. Our results indicated that when N-hydroxy-2-AF was incubated with BHA or DMP at ph 5.0 for 4 h, a small amount of the corresponding adduct was formed. At time points of 1 h or less, or at ph 7.0 at any time point, no adducts were formed. These results are consistent with the time course and ph dependence of DNA adduct formation by acid-catalyzed hydrolysis of N-hydroxy- 2-naphthylamine (6). The rapid (<2 min) and extensive for- mation of 2-AFlphenolic adducts at ph 5.0 or 7.0 in the prostaglandin H synthase or horseradish peroxidase systems clearly rules out a nitrenium ion derived from N-hydroxy-2- AF as the reactive intermediate involved in adduct formation. Furthermore, the lack of 2-AF/phenol adduct formation when N-hydroxy-2-AF was substituted for 2-AF in peroxidase/h202 reaction mixtures demonstrates that a metabolite of N-hydroxy-2-AF, free radical or otherwise, is not involved in adduct formation. The chemical oxidation of 2-AF by potassium ferricyanide, a known one-electron oxidant, gave a similar product profile as the peroxidase system at non-acidic ph, namely organicextractable polymer and azofluorene. In addition, a 2-AF/ phenolic adduct was formed when BHA was included in the reaction mixture. Oxidation of 2-AF by Fenton's reagent at ph 5.0 produced 2-aminodifluorenylamine in addition to azofluorene and polymer, a result consistent with the enzymatic reactions. The formation of 2-nitrosofluorene in the Fenton system is most likely the result of a hydroxyl radical being the oxidant in this system. In summary, the data presented in this paper provide evidence that the prostaglandin H synthase peroxidase and horseradish peroxidase oxidize 2-AF similarly via a one-elec- tron mechanism. The possibility of a direct two-electron oxidation of 2-AF to yield a nitrenium ion cannot be rigorously excluded, based on the data in this paper. This appears very unlikely, however, based on known peroxidase/amine chemistry, the product profile (especially azofluorene), and ESR data (polymer radical). In either case, data from this and the previous study further suggest that the reactive oxidation products in both enzyme systems are not derived from N- hydroxy-2-af. We propose that the electrophilic species responsible for the prostaglandin H synthase-catalyzed binding of 2-AF to tissue macromolecules, DNA (ll), and trna (12) is either a free radical or a free radical-derived product (the diimine derived from oxidation of 2-aminodifluorenylamine).

10 13894 Peroxidation of 2-Arninofluorene Although phenol adduct formation is not necessarily a model for DNA binding, several Iines of evidence support this model. Other work from this laboratory (46) has shown that both prostaglandin H synthase and horseradish peroxidase catalyze the extensive binding of 2-AF to DNA; the major adduct being unique and distinct from the C8-dG-AF adduct obtained from N-hydroxy-2-AF (nitrenium ion). Furthermore, the species responsible for DNA binding is extremely short-lived, as evidenced by greatly reduced binding when DNA is added 1 min after initiation of 2-AF oxidation. It therefore appears likely that the same non-nitrenium ion species responsible for phenol adduct formation and DNA binding is a free radical. That this reaction may be catalyzed by prostaglandin H synthase at either acidic or neutral ph, utilizing arachidonic acid, lipid hydroperoxides, or H202, and at a high turnover number characteristic of peroxidases, make it a highly versatile reaction. The possible in vivo relevance of this reaction, especially in regard to the activation of aryl- amine bladder carcinogens, deserves further investigation. Further studies of the peroxidation of other aromatic amines are also underway, since we expect neither that all of these compounds will be oxidized by the same mechanism, nor that all peroxidases will act on a particular substrate in the same manner. Acknowledgments-we gratefully acknowledge the helpful discussions with and contributions from Drs. Ron Mason and Paul West. Dean Marbury provided expert assistance in performing the mass spectrometry; NMR spectra were provided by Dr. Tony Ribeiro. We also thank Robin Sorrel1 for preparing the manuscript. REFERENCES 1. Eling, T. E., Boyd, J. A., Reed, G. A. Mason, R. P., and Sivarajah, K. (1983) Drug Metab. Reu. 14, Marnett, L. J., and Eling, T. E. (1983) in Reuierus in Biochemical Toxicology (Hodgson, E., Bend, J. R., and Philpot, R. M., eds) Vol. 5. DD Elsevier/North-Holland Biomedical Press, New Yo& 3. Boyd, J. A., Harvan, D. J., and Eling, T. E. (1983) J. Bid. Chem 258, Frederick, C. B., Mays, J. B., Ziegler, D. M., Cuengerich, F. P., and Kadlubar, F. F. (1982) Cancer Res. 42, Miller, E. C., and Miller, J. A. (1981) Cancer (Phila.) 47, Kadlubar, F. F., Miller, J. A., and Miller, E. C. (1977) Cancer Res. 37, Poupko, J. M., Hearn, W. L., and Radomski, J. L. (1979) Toxicol. Appl. Pharmacol. 50, Radomski, J. L. (1979) Annu. Reu. Pharmacol. Toxcol. 19, O Brien, P. J. (1984) in Free Radicals in Biology (Pryor, W. A., ed) Vol. 6, pp , Academic Press, New York 10. Robertson, I. G. C., Sivarajah, K., Eling, T. E., and Zeiger, E. (1983) Cancer Res. 43, Kadlubar, F. F., Frederick, C. B., Weis, C. C., and Zenser, T. V. (1982) Biochern. Biophys. Res. Comm. 108, Morton, K. C., King, C. M., Vaught, J. B., Wang, C. Y., Lee, M.- S., and Marnett, L. J. (1983) Biachern. Biophys. Res. Comm Westra, J. G. (1981) Carcinogenesis (Lond.) 2, Josephy, P. D., and Van Damme, A. (1984) Anal. Ckm. 56, Rahimtula, A. (1983) Chem.-Bwl. Interactions 45, Hewgill, F. R., and Hewitt, D. G. (1967) J. Chem. SOC. (Land.) (C) Miyarnoto, T., Ogino, N., Yamarnoto, S., and Hayaishi, 0. (1976) J. Biol. Chem. 252, Siedlik, P. H., and Marnett, L. J. (1984) Methods Enzymol. 105, Hildebrandt, A. G., and Roots, 1. (1975) Arch. Biochem. Biophys. 171, Saunders, B. C., and Stark, B. P. (1967) Tetrahedron 23, Brown, K. C., and Corbett, J. F. (1979) J. SOC. Cosmet. Chem. 30, Josephy, P. D., and Van Damrne, A. (1984) Biochem. Pharmacol. 33, Corbett, J. F. (1970) J. Chem. Soc. (Lond.)(Bf Josephy, P. D., and Lenkinski, R. E. (1984) J. Chromatogr. 294, Saunders, B. C. (1973) in Inorganic Biochemistry (Eichhorn, G. L., ed) Vols. 1 and 2, pp , Elsevier/North-Holland Biomedical Press, New York 26. Chance, B. (1952) Arch. Biochem. Biophys. 41, George, P. (1952) Nature (Lond.) 169, Yamszaki, I., Mason, H. S., and Piette, L. (1960) J. Biol. Chern. 235, Yamazaki, I. (1977) in Free Radicals in Biology (Pryor, W. A., ed) Vol. 3, , Academic Press, New York 30. Josephy, P. D., Eling, T., and Mason, R. P. (1982) J. Biol. Chem. 257, Josephy, P. D., Eling, T. E., and Mason, R. P. (1983) J. Biol. Chem. 258, Josephy, P. D., Mason, R. P., and Eling, T. E. (1982) Carcinogenesis (Lond.) 3, Mann, P. J. G., and Saunders, B. C. (1935) Proc. R. SOC. Lo&. B Biol. Sci. 119, Baker, P. B., and Saunders, B. C. (1974) Tetrahedron 30, Holland, V. R., and Saunders, B. C. (1968) Tetrahedron 24, Holland, V. R., Roberts, B. M., and Saunders, 3. C. (1969) Tetrahedron 25, Mohilner, D. M., Adams, R. N., and Argersinger, W. J. (1962) J. Am. Chem. SOC. 84, Yasukouchi, K., Taniguchi, X., Yamaguchi, H., Miyaguchi, K., and Horie, K. (1979) Bull. Chen. SOC. Jpn. 52, Moldeus, P., Larsson, R., Ross, D., Andersson, B., Nordenskjoid, M., Rahimtula, A., and Lindeke, B. (1983) in Extrahepatic Drug Metabolism and Chemical Carcinogenesis (Rydstrom, J., Montelius, J., and Bengtsson, M., eds) pp , Elsevier/North Holland Biomedical Press, New York 40. Claiborne, A., and Fridovich, I. (1979) Biochemistry 18, Wise, R. W., Zenser, T. V., and Davis, B. B. (1983) Carcinogenesis (Lord.) 4, Steenken, S., and O Neill, P. (1977) J. Phys. Chern Rieker. A.. and Kessler, H. (1966) 2. Naturforsch. 21b, Job, D., and Dunford, H. B. (1976) Eur. J. Biochem. 66, Sgaragli, G., Corte, L. D., Puliti, R., De Sarlo, F., Francalanci, R., Guarna, A., Dolara, P., and Komarynsky, M. (1980) Biochem. Pharmacol. 29, Krauss, R. S., Reed, G. A., and Eling, T. E. (1984) Proceedings Am. Assoc. Cancer Res. 25,84

11 Peroxidation of 2-Aminofluorene supp1enentary materia1 TO EVIDENCE FOR A ONE-ELECTRONMECHANISM OF 2-PMINOFLUORENEOXIDATION BY PROSTAGLANOIN SYNTHASEANDHORSERADISHPEROXIDASE H Jeff A. Boyd and Thanas E. Cling EXPERIMENTAL PROCEDURES Matemall - (rl;s-3h12-af(120 mcilnmo1) dnd (&-3HlH-hydmxy-2-AF 1147 mci/ml) rere m f m n M duest Research Inrtitute. Kansas City, F10. The 13H12-AF was conristent- >9BX pure as detemined by periodic HPLC analyrir. Unlabeled 2-AF and 2-nitmfluorene :re obtained frm Aldrich Chemical Co. BHA 12-t-autyl-4-nuthoxy-phenoll. BHT (2.6-di-Lbutyl- crerol). and 2-t-butyl-4-methyl-phcnol &'e pmducts Of FlUkb. A.G. Horseradish peraxi%re (Type Vi), MP. 8mYbic acid. OCQI. and deutemted chlomfom, (99.96 ata 4 Dl *re Obtained frm Si- Chmical Co. POtaslim carbonate. potassium ferricyanide, fermus sulfate. dimdim EOTA, I-butyl-hydr uinone. idanthane. hydrazine hydrate, and palladim on carbon (5x1 *we fm Aldrich. lpac1methyl iodide I10 mcilml was purchased frrm NR England Nuclear. HPLC grade methanol. YPLC grade *Iter, and Hz02 (30%) were obtained frm Fisher Scientific Co. Al other materials -PC reagent grade and obtained fm cmercial rourcer. S "thesis Of standards and substrates - N-hydmry-2-AF YIP synthesized by reduction Of 2-nitrifluorenc with palladim On carbm and hydrazine hydrate as described by Yestra (13). 2-Nitmrofluorene dmxyflyorene. and aloflyo?ene IKC Syntherized fm N-hydrwy-2-AF. then analyzed by uv-vliible rpectrophetmtry and mass IpeCtmtry as preyiously described (31. KQl/phenol adduct standards em generated by adding 15-1 Of DCQl to an equimolar amunt Of elther BHA 01 MV in 30 ml of 25 nll sodim bonte buffer. ph 9.0, After 30 min. the mlxtwes were acidified with several dmps of HCl luntil the blue color turned redorangel and the adducts extracted with 2 Yolms of water-saturated ethyl acetatelcther. 1:l. The add& were purified by preparative TLC on 1wO un silica gel GF plater (Analtech) in a solvent systm Of etherlhexane. 1:l. Mars spectra and WllR spectra were Cnpletely Consirtent with adducts of the follaing Structure. This leaction and structures have been described elreuhere in detail (141. "o@nao CI allaed to pmcecd for 20 min; products -?e extracted and Identified dr described above for the en2mtic reactions. The oxidation sf 2-AF by Fenton'r reagent was studled by adding H 02 to a flnal concentration of 1 m to I 1w ml solution cansirtin9 Of 100 nw wdium acetate buffer. ph 5.0. or LOO nll ptdpsim phosphate buffer, ph nh fefroui Sulfate. 1 n(l dirod?m EOTA, dnd 100 yn 2-AF. After 20 nin. the pmdwtr were extracted and identified dl described for the enzpatic reactions. Reaction Of N-hydmxy-2-AF with phenols - TO study the porrlble reaction between N- hydmxy-2-af and phenols. a time course of incvbationr was carried out at both ph 5.0 and 7.0. Reaction mixtulex contained 1W nll IOdiUm acetate Puffer. ph 5.0. or 100 n(l ~otdlll~rn phosphate buffer ph UM N-hydroxy-2-AF and 500 YM BHA or MP Reactions were run ~n the ddvk. undh argon: at 3pC for 2 min. 1 k. or 4 hr. with constant 19ltdtian Products were extracted wth *ate?-saturated ethyl acetatelether, 1'1. at the approppiate time pornt. then separated by HPLC dl described above. PmduCt6 were identified by cochromatography with authentic standards and by Uv-visible rpectraphatometry. To determine whethel N-hydroxy-2-AF or 1 metabolite thermf could react ith a phenol to pmduce adducts under active Oxidizing conditions. experiments mployrng lyhin-h~dro~~- 2-Ai a? a pssiblc perexidare substrate in the presence of Phenols were also carried wt Reaction mxtwes contained 100 n*l sodium acetate buffer. ph 5.0. or LOO nl( FOtdlliUm phosphate buffer, ph vglml horseradish peroxidase. 50 d4 13H)N-hydroxy-2-AF. 100 ym BHA or DIP. and 100 wn H 02 The reaction was run and the products extracted dl dercrlbed &bow for the 2-AF reactfoni. products were analyzed by HPLC. StwctwaI identificatran of pmducts - Al products were initrally charartenzed by HPLC. nc. and uv-visible absorbance in ethanol. Mass spectra we~e Obtained by direct probe. 70 ev electron impact ionization using a Vt 7070E mall Spectmmeter equlpped,mth d VG 2050 data wstm. S-ler ere rcanned at 1000 rerolvinq power frm mle 600 to 50 at a scan speed (lac)blia was synthesized fm t-butyl-hydmquinone and l'4c)methyl iodide, then purified 1% described in detail 115). The purified pmduct MI r95% purt as detemined by TLC analynr on 250 silica gel G plates IAmltechl in a solvent systm of benzenelethyl acetate, 19:l. The (%lbha was quantitated by its colorimetric rolction with DWI IS described abvei an extinction coefficient of 6076 (109 I ) at 576 m (in acetone] was determined for the adduct generated using knom munts of unlabeled BHA and MPI. The specific activity Of the synthetic (I4C)BHA MS 250 ucilml. The dl-bha standard lds synthesized by oxidation Of BHA with p+tassim ferricyanide as described (16). Pmstaqlmdin H rvnthare prewration - IticVOSml fm ram IWinll VCIiCICI *le prepared as described previously 01. with the exception that the final ~c~~~pension MI in 100 nll wtassim phosphate buffer. ph 7.0. Prostaglandin H synthase was solubilized fm. rm seminal vesicle micmsans wing 1% Tern-20 as described (17). then stored at -7oOC until used. Micmra~l and soiubili2m enzyne preparations could be used interchangeably for a11 experimntr with qualitatively identical results. Generally. solubilized pmrtaglandin H synthase MI used fol a11 rpectmplmtmtric wrk while either e n w pmparation MS used for other cxperinmtr as indicated. Enrylc activity IUS assured prior to each experiwnt by lnnltorin.2 the incomeration of oxmen into arachidonic acid using I Clark-twe oxv~en adequate 2-AF 2-NltrOfl"oTe"e 2-~~nadifluorenylaminelpolper Adduct bdduct ~dduct a Adduct n Azofluorene TABLE V NMR spectra of 2-AOFAlphenOllc adducts 2-ilninodifluarenylaminelBHA t-butyl ladduct butyl (broad) -NH :::;; (d. J= C6 H, quinoid : : E Id, J=lO.BCI -C6 H. 4UlnDid H and C5 H of qulnoid. ammatic f1"orenyl proton5 P-im~nod~fluorenylaa~neiDnP (adduct 4)