474
Chem. Res. Toxicol. 1991,4,414-481
Reactive Intermediates Formed during the Peroxidative Oxidation of Anisidine Isomers David C. Thompson+and Thomas E. Eling* Laboratory of Molecular Biophysics, National Institute of Environmental Health Sciences, Research Triangle Park, North Carolina 27709
Received February 19, 1991 The ortho isomer of anisidine (2-methoxyaniline) causes urinary bladder tumors in both mice and rata while the para isomer (4methoxyaniline) is inactive. Since the urinary bladder contains substantial peroxidase activity, we investigated the peroxidative metabolism of both o- and p-anisidine using horseradish peroxidase as a model enzyme. Both isomers were excellent reducing cofactors for the oxidized state of horseradish peroxidase (HRF’),resulting in one-electron oxidation to free radicals. Using high-pressure liquid chromatography, we observed that HRP oxidized p-anisidine to a diimine metabolite which subsequently hydrolyzed to form a quinone imine. Also observed was a dimeric metabolite with an azo bond. Both the diimine and quinone imine metabolites were reactive toward nucleophiles. The quinone imine formed a conjugate with glutathione and was also reduced by glutathione or ascorbic acid. Higher concentrations of substrate (>1mM) led to the formation of polymeric produds (tetramer). Similar metabolites (diimine, quinone imine, azo dimer, polymers) were observed with o-anisidine. Using tritiumlabeled anisidine, we observed substantial metabolism-dependent covalent binding of both isomers to protein and DNA. These results demonstrate that horseradish peroxidase dependent metabolism of anisidine isomers yields similar metabolites, although some differences in reactivity of the respective intermediates with nucleophiles were observed.
Introductlon Aromatic amines represent a class of toxic and carcinogenic chemicals to which humans are widely exposed. Certain aromatic amines, such as benzidine, are known human carcinogens. The central hypothesis for the induction of cancer by these agents involves their oxidation to electrophilic intermediates which covalently link to DNA (1,2). Previous studies have focused on the hepatic metabolism of aromatic amines and the induction of liver cancer by these compounds (3-5). Yet, many aromatic amines also induce extrahepatic neoplasia, particularly bladder cancer. Our laboratory is interested in the involvement of peroxidase enzymes in the activation of chemical carcinogens which cause extrahepatic cancer. Bladder tissue is one of the tissues where peroxidase activity may be especially relevant in the development of neoplasia. This is because bladder epithelia contains low amounts of cytochrome P-450 activity but appreciable peroxidase activity (prostaglandin H synthase) (6, 7). Aromatic amines are generally good cosubstrates for peroxidases (81, and recent studies have investigated the hypothesis that peroxidases may be an important system for the activation of carcinogenic aromatic amines in extrahepatic tissue (9-1 1). In a carcinogenicity study of anisidine isomers in rats and mice conducted by the National Toxicology Program (12, 13)some striking differences were noted (Table I). The ortho isomer elicited a high incidence of urinary bladder carcinomas or papillomas in rata and mice. The effect was dose dependent although in rata even the low dose (5000 ppm) led to an almost 100% incidence of tumors. In direct contrast, the para isomer was ineffective in eliciting tumorigenesis in either rata or mice at any of the doses tested. One possible mechanism to explain these To whom correspondence should be addressed. Present address: Department of Medical Pharmacology and Toxicology, Texas ALM University, College Station, TX 77843.
Table I. Carcinogenicity Dataa o-anisidine urinary bladder carcinomas or papillomas (transitional cell) rata (dietary doses of 0, 5000,or loo00 ppm): males:* 0151,52154,52152 females: 0149,46/49,50/51 mice (dietary doses of 0, 2500,or 5000 ppm): males: 0/48,2/55,22/53 females: 0/50,1/51,22/50 p-anisidine no neoplasms associated with chemical exposure in mice or rats of either sex doses tested: rats (0, 3000, or 6000 ppm), mice (0,5OO0,or 10Ooo ppm) Data are from National Toxicology Program carcinogenicity study on aromatic amines. bNumbers refer to number of animals bearing tumors/number of animals receiving the given doae of test compound.
carcinogenicityobservations involves potential differences in metabolism or pharmacokinetics between the two isomers. Since these areas have not been thoroughly studied with the anisidine isomers, we initiated studies into their metabolism. In this report we describe the peroxidative oxidation of both anisidine isomers using horseradish peroxidase (HRP)’ as a model peroxidase.
Materlals and Methods Materials. Radiolabeled [G-3H]-o-anisidine(2.16Ci/mmol) and p-anisidine (1.40 Ci/mmol) were synthesized by Chemsyn Science Laboratories (Lenexa, KS). Radiochemical purity of both compounds was >98% (TU). &Phenyl-Cpentenyl hydroperoxide (PPHP) was obtained from Oxford Biomedical Research, Inc. (Oxford, MI). Nonlabeled 0- and panisidine (>99% purity based on HPLC) were purchased from Aldrich (Milwaukee, WI). Abbreviations: HRP, homeradieh peroddasq PHS, prw landin H synthaee, BHA, butylated hydroxyanisole; TLC, thin-layerxomatography; HPLC, high-pressure liquid chromatography; PPHP, 5phenyl-4-pentenylhydroperoxide.
This article not subject to U S . Copyright. Published 1991 by the American Chemical Society
Anisidine Metabolism Horseradish peroxidase (HRP, type VI, 310 purpurogallin units/mg), calf thymus DNA, glutathione, phenol, sodium nitroprusside, and hydrogen peroxide (30% solution) were from Sigma (St. Louis, MO). Butylated hydroxyanisole (BHA) was obtained from Fluka Chemical Co.(Ronkonkoma, NY). All other chemicals were of the highest grade of purity available from local suppliers. High-Pressure Liquid Chromatography. Anisidine metabolism was followed by high-pressure liquid chromatography (HPLC). Incubations generally contained 1mM anisidine (0.2 pCi), 1 mM hydrogen peroxide, and 0.1 pg of HRP in a total volume of 1mL of 0.1 M phosphate buffer (pH 7.0). Reactions were stopped with potassium cyanide (10 mM f d concentration) and 2WpL aliquots injected directly onto the HPLC. The HPLC was a Waters system with both a UV detector (280 nm, SpectraPhysics Model 770) and a radioactivity detedor (Radiomatic Flo-One Beta). The column was a 5-pm Ultrasphere ODS (Beckman, 4.6 X 250 mm) preceded by a (2-18 guard column. Compounds were eluted with a water/methanol mobile phase which initially was 50% methanol for 15 min, followed by a linear increase to 100% methanol at 30 min, which remained in effect until 40 min when initial conditions were restored. Flow rate was 1 mL/min throughout the entire run. Metabolite Analysis. Metabolites were collected from thinlayer chromatographyor HPLC runs,concentrated, and analyzed by mass spectroscopy and/or proton nuclear magnetic resonance. For metabolites (quinone imines) collected by HPLC a scaled-up reaction volume of 50 mL was used. At an appropriate time, the reaction was stopped by extraction with 50 mL of ethyl acetate. The ethyl acetate extracts were concentrated on a rotary evaporator, and the residue was dissolved in 1mL of methanol. The sample was then chromatographed on a semipreparative HPLC column (10 X 250 mm, Hibar RP-18, Merck), and fractions containing peaks of interest were collected. An isocratic flow of 65% methanol, with a flow rate of 4 mL/min (UV detection at 280 nm), was used to elute the metabolites. Samples were concentrated by rotary evaporation and analyzed by mass spectroscopy. The quinone imines were reacted with glutathione for 1 h at 31 "C. The reaction products were concentrated by passage over a Prep-Sep C18column (Fisher), washed with acidic water, and eluted with 1mL of methanol. These samples were then separated and isolated by HPLC on the semipreparative column with an isocratic mobile phase of 50% methanol and 50% acidic water (0.25% perchloric acid and 0.25% acetic acid adjusted to pH 3.5). With p-anisidine three peaks were observed on the semipreparative column. One peak represented the reduced form of the quinone imine and was not a glutathione conjugate (8 min). The major peak (11min) was collected, analyzed by fast atom bombardment maas spectroscopy, and found to be a monoglutathione conjugate (MW 521) of the quinone imine. The minor peak (16 min) was not analyzed. Multiple peaks were also observed with o-anisidine (data not shown). Polymeric metabolites of anisidine (and BHA adducts) were isolated by thin-layer chromatography (TLC) using preparative silica gel TLC plates (1 mm, Merck). Extracts of 10-mL incubations were applied to the TLC plate and developed in either 80:20 or 6040 ratios of hexane/ethyl acetate (14). Bands of interest were scraped from the TLC plate, eluted with ethyl acetate, evaporated, and analyzed by mass spectroscopy. Mass spectroscopy was performed on a Concept I SQ hybrid mass spectrometer, Kratos Analytical (Manchester, England). Samples were analyzed in the electron impact (E1 at 70 eV), low-resolution mode and introduced into the mass spectrometer by using the direct probe. The data were then processed on a SUN 360 workstation utilizing MACH 3 software from Kratos Analytical. Covalent Binding. Reactions for measuring covalent binding to protein contained 1mM anisidine (0.5 pCi), 1 mM hydrogen peroxide, 0.1-1 pg of HRP, and 1mg of heat-denatured rat liver S-9 protein in 1mL of 0.1 M phosphate buffer (pH 7.0). Reactions were stopped by the addition of 50 p L of 100% trichloroacetic acid. After centrifugation, the protein pellets were repeatedly washed with methanol or methanol/ether (31) until background levels of radiation were found in the washes. The protein pellets were dissolved in 1 mL of 1 N sodium hydroxide with heating at 60 "C for 1 h and neutralized with 10 M HCl, and a 0.5-mL
Chem. Res. Toxicol., Vol. 4, No.4, 1991 475
Table 11. Abilitv of Anisidine Isomers To Serve as HRP Reducing Cofactorsa % PPHP reduced group (mean f SE) control (no cofactor) 2 f l butylated hydroxyanisole 55 f 1 ascorbic acid 48 f 3 p-anisidin e 82 f 2 o-anisidine 0 " Incubations contained 100 pM PPHP, 200 pM cofactor, and 60 nM HRP in 0.1 M potassium citrate buffer (pH 5.5). The higher the percentage of PPHP reduced in this assay, the better reducing cofactor the test compound is. aliquot was counted for radioactivity and a 50-pL aliquot used to measure protein (15). Reactions for DNA covalent binding contained 1mM anisidine (1.0 pCi), 1mM hydrogen peroxide, 0.1 pg of HRP, and 1mg of calf thymus DNA in 1 mL of 0.1 M phosphate buffer (pH 7.0). Reactions were stopped by subsequent extraction with 2 volumes each of phenol, chloroform/isoamyl alcohol (241), and ethyl acetate. An aliquot (0.5 mL) of the final DNA solution was counted for radioactivity,and a 50-pL aliquot was used to measure DNA concentration by absorbance at 260 nm. Other Assays. Oxygen consumption was measured with a Clark-type electrode and a Yellow Springs Model 53 oxygen monitor. Reactions contained 1mM anisidine, 1mM hydrogen peroxide, 0.1 pg/mL HRP, and various amounts of glutathione in 1.5 mL of 0.1 M phosphate buffer (pH 7.0). Reactions were carried out at 37 OC. Ammonia formation was measured by a colorimetric procedure (Berthelot method) using phenol/hypochlorite (16). Reactions contained 1mM anisidme, 1mM hydrogen peroxide, and 1g/mL HRP in 1 mL of 0.1 M phosphate buffer (pH 7.0) and were stopped by extraction with 2 volumes of ethyl acetate. An aliquot (0.5 mL) of the remaining aqueous phase was used for the ammonia measurement. The ability of anisidine isomers to serve as reducing mubatrates for horseradish peroxidase was measured by using the hydroperoxide &phenyl-4pentenyl hydroperoxide (PPHP)as described by Weller et al. (17). Incubations contained 100 pM PPHP, 200 pM anisidine, and 60 nM horseradish peroxidase in 0.1 M potassium citrate buffer (pH 5.5). The amount of PPHP reduced in the presence and absence of anisidine was measured by HPLC. Spectral changes during anisidine oxidation were measured by using a Hewlett Packard 8452A diode array spectrophotometer. Reaction conditions consisted of 0.25 mM anisidine, 2 pg of HRP, and 0.25 mM hydrogen peroxide in a total of 2 mL of 0.1 M phosphate buffer, pH 7. Scans (350-830 nm) were taken at various time points after the addition of hydrogen peroxide.
Results S t a n d a r d incubations of 1 mM anisidine, 0.1 pg/mL HRP, and 1 mM hydrogen peroxide were carried out in 0.1 M phosphate buffer ( p H 7.0) with both anisidine isomers. Upon t h e addition of hydrogen peroxide to panisidine incubations, a reddish brown color gradually developed. At higher p-anisidine concentrations (>1mM) a reddish brown solid precipitated. With o-anisidine incubations, an orange color developed (lighter in color than p-anisidine) which also precipitated at high concentrations. T h e para isomer was oxidized at a much faster rate than the ortho isomer. Berry and Boyd (18)reported a 5.7-fold difference in the relative reaction rates of p-anisidine and o-anisidine with HRP (185 pM s-* for p-anisidine versus 32.7 pM s-l for o-anisidine). Many compounds, including aromatic amines (14), can function as reducing cofactors for peroxidases, being oxidized in the process to free radicals. We tested the ability of o-anisidine and p-anisidine t o function as reducing cofactors for HRP. We found t h a t both isomers were excellent electron donors t o the peroxidase (Table 11). T h e
Thompson and Eling
476 Chem. Res. Toxicol., Vol. 4, No. 4, 1991 2000
1004
1500
E,
1000
'0
500
0
1
2
3
4
5
6
7
8
9
10
mM GSH
Figure 1. Oxygen consumption by anisidine isomers in the presence of glutathione. Reactions contained 1 mM anisidine, 1 mM hydrogen peroxide, and 0.1 pg/mL HRP in 0.1 M phosphate buffer (pH 7.0). 100% = 240 pM oxygen. ortho isomer was slightly more efficient in donating electrons to HRP than was the para isomer. Ascorbate and butylated hydroxyanisole were used as controls in the m y (17). Detection of Anisidine Free Radicals. Aromatic amines generally form nitrogen-centered cation radicals when oxidized by peroxidases (19, 20). However, free radicals from aromatic amines are not readily detected by static electron spin resonance techniques. Therefore, in order to confirm the formation of a free radical, we measured oxygen consumption during the peroxidative oxidation of anisidine isomers in the presence of various amounta of glutathione. Aromatic amine free radicals readily react with glutathione, forming glutathionyl radicals which subsequently react with oxygen (21). We observed increased oxygen uptake in the presence of glutathione with both anisidine isomers, thus confirming freeradical formation (Figure 1). The reactive metabolite from the ortho isomer was more reactive toward glutathione than was the para isomer metabolite. In the presence of glutathione, the formation of other metabolites from anisidine isomers was inhibited (data not shown). Characterization of Anisidine Metabolites. The anisidine metabolites formed during peroxidase incubations were analyzed by HPLC. Figure 2 illustrated typical chromatograms from both isomers obtained under standard reaction conditions and using a radioactivity detector. With p-anisidine five peaks were observed (upper panel). Peak 1 represented the solvent front while peak 2 was an artifact which coeluted with the reduced form of peak 4. This peak was present in small, relatively constant amounts in our reactions. Peak 3 was a diimine metabolite and peak 4 a quinone imine metabolite of p-anisidine. Peak 5 was an azo dimer (4,4'-dimethoxyazobenzene,MW 242). At higher concentrations (>1 mM) a tetramer (MW 454)was formed which precipitated from solution. This tetramer [2-amino-5-(p-anisidino)benzoquinonedi-pmethoxyphenylimine] has been reported in previous studies (13,22,23). The parent compound, p-anisidine, eluted at 4 min and was completely metabolized when this incubation was analyzed. The quinone imine metabolite (peak 4)was quite stable and could be isolated and analyzed by MS and NMR. The mass spectrum of the quinone imine from p-anisidine (reduced form, MW 215)is shown in the upper panel of Figure 3. (The quinone imine was reduced with ascorbic acid prior to analysis by mass spectroscopy because this was necessary for the less stable quinone imine of o-anisidine.) Proton NMR of the quinone imine confirmed the structure of the quinone imine, revealing 1 methoxyl group (3 H, 3.8 ppm), 8 ring protons
0 0
5
IO
15
20
25
30
35
40
Time (min)
o-anisidine 750
4
2
0
5
10
15
20
25
30
35
40
Time (min)
Figure 2. HPLC chromatogram of anisidine metabolites. Standard incubations (1mM anisidine, 1mM hydrogen peroxide, and 1 pg/mL HRP in 0.1 M phosphate buffer, pH 7) were stopped after 60 min and analyzed by HPLC. For p-anisidine (upper panel) the peak numbers refer to the following metabolites: (1) solvent front; (2) reduced fom of quinone imine; (3) diimine; (4) quinone imine; and (5) azo dimer. For o-anisidine (lower panel) the peak numbers refer to the following metabolites: (1)solvent front (and polar metabolite?);(2) quinone imine; (3) diimine; (4) quinone imine (?); (5) diimine (?); (6) tetramer; (7) azo dimer. Parent compounds of both isomers elute at 4 min under these conditione.
(6.5-7.4ppm), and no NH protons (data not shown). The diimine metabolite (peak 3) was not sufficiently stable to isolate, however, and rapidly hydrolyzed to the quinone imine. The diimine could be trapped with butylated hydroxyanisoleas described by Josephy and van Damme (24), and this resulted in a blue-colored adduct (MW 360). The mass spectrum of this adduct is shown in the upper panel of Figure 4. The lower panel of Figure 2 shows a corresponding chromatogram of the HRP metabolites of o-anisidine. With the ortho isomer seven peaks were obtained. Peak 1, which represented the solvent front, was a much more significant percentage of the total metabolites with 0anisidine than peak 1 in the p-anisidine chromatogram. This indicates that perhaps there are polar metabolites being formed as well as the other more lipophilic metabolites. Peak 2 represents a diimine and peak 3 a quinone imine metabolite, similar to the p-anisidine metabolites. Peaks 4 and 5 are unidentified but probably also are diimine and quinone imine metabolites which have lost a
Chem. Res. Toxicol., Vol. 4, No. 4, 1991 477
Anisidine Metabolism
Metabolism of p-anisidine by HRP 0
0
120
140
160
-;i-
,
180
V
0
11115.
10
I
20
-
.
-
.
30
- p-oniridine - dilmine - quinoneimine
-
I
I
.
40
I
50
,
-
,
60
Time (min) Metabolism of o-anisidine by HRP
100 90 80
A o
70
P t
-
1000
60
-
50
0
E
40
30
o-anisidine diimine quinoneimine
600 500 400
20
Figure 3. Mass spectra of reduced quinone imine metabolites of p- and o-anisidine. Upper panel: p-Anisidine. Lower panel: o-Anisidine.
90.
ea.
n.
TI
4 100 90
d0
70
1.9
w II ., 100
150
,
,
dl
I
I
200
250
300
350
Mlll
Figure 4. Mass spectra of BHA adducts of diimine metabolites of p- and o-anisidine. Upper panel: p-Anisidine. Lower panel: 0-
Anisidine.
methoxyl group. We feel this is likely because peaks 2-5 all disappear when glutathione is added to the reaction mixture (not shown) and also because the time course of the formation of these two peaks is similar to those of peaks 2 and 3. Peaks 4 and 5 were not analyzed because they were unstable, and the quantity formed was low
Time (min)
Figure 5. Time course for HRP-dependentmetabolism of anisidine isomers. Standard incubations [lmM anisidine (0.2 pCi), 1mM hydrogen peroxide, and 0.1 ~lg/mLHRP in 0.1 M phosphate buffer, pH 71 were stopped at the times indicated and analyzed by HPLC using a radioactivity detector.
compared to peaks 2 and 3. Peak 6 is a polymer (tetramer, MW 488), and peak 7 is an azo dimer (2,2'-dimethoxyazobenzene, MW 242). Both cis and trans isomers of the azo dimer were isolated. Parent o-anisidine eluted at 4 min under these conditions. The quinone imine metabolite of o-anisidine was not as stable as the para isomer. The mass spectrum of the reduced form of the quinone imine (MW 245) is shown in the lower panel of Figure 3. The BHA adduct of the corresponding diimine (MW 390) is shown in the lower panel of Figure 4. We did not observe a BHA adduct corresponding to a diimine in which the methoxyl group had been lost (MW 360). This indicated that coupling reactions with o-anisidine occur predominately at the para position, similar to p-anisidine. Time Course of Oxidation. Figure 5 illustrates the time course of formation of the diimine and quinone imine metabolites of both anisidine isomers as well as the disappearance of the parent compounds. As can be seen in the upper panel, the diimine metabolite of p-anisidine is formed first and reaches a peak concentration at approximately 5 min. The concentration of diimine begins to steadily decline after this time point due to ita hydrolysis into the quinone imine. For p-anisidine these two metabolites account for about 80% of all products formed, as determined by the percent of radioactivity recovered in these peaks. In Figure 6, the visible spectra of the p-anisidinelHRP reaction are shown over a time course of 60 min. At the earliest scan (5 min) a peak is observed at 450 nm. This peak gradually shifts over the 60-min time course to about 490 nm. This clearly reflects the initial appearance of the diimine (Am= = 450 nm) and ita subsequent conversion to the quinone imine (A, = 500 nm). A similar shift was observed with o-anisidine.
Thompson and Eling
478 Chem. Res. Toxicol., Vol. 4, No. 4, 1991
0.9
o.8 0.7
1
1
1
11
p-anisidine
4
0.9
0.8
o-anisidine
0.7 0.6 c
0.5
s"
0.4 0.3
0.3
0.2
O.'
-.-
t
0.1
1 I
0.0
350 400 450 500 550 600 650 700 750 800
350 400 450 500 550 600 650 700 750 BOO
Wavelength (nm)
Wovelength (nm)
Figure 6. Time course of spectral changes ocmrriy during HRp-dependent anisidine oxidation. Fbactions contained 0.25 mM anisidine, 0.25 mM hydrogen peroxide, and 1pg/mL HRP in 0.1 M phosphate buffer (pH 7). Spectra were recorded at 5,10,20,30,40,50, and 60 min after the addition of hydrogen peroxide. 500
Table 111. Covalent Binding of ['HIAnisidine Isomers to
T
p-anisidine
Protein
A
reaction p-anisidine complete reaction" -hydrogen peroxide
-HRP
7 o-anisidine 0
30
60
90
+5 mM glutathione +1 mM ascorbic acid o-anisidine complete reaction -hydrogen peroxide 120
Time (min) Figure 7. Time course of ammonia formation during HRP-dependent oxidation of anisidine isomers. Reactions contained 1 mM anisidine, 1 mM hydrogen peroxide, and 1pg/mL HRP in 0.1 M phosphate buffer (pH 7.0). In the lower panel of Figure 5, the formation of diimine and that of quinone imine follow a similar time course for o-anisidine but the products accounted for only 40-50 % of the metabolites formed. There is a larger proportion of nonpolar polymeric products (peaks 7 and 8) formed as well as polar products (peak 11, in addition to the possible secondary diimine and quinone imine metabolites. The conversion of diimine to quinone imine involves the loss of ammonia. The ammonia formed during the metabolism of both o-anisidine and p-anisidine is shown in Figure 7. Ammonia is released by both isomers, further indicating that diimine to quinone imine conversion is taking place. The formation of ammonia follows the same time course as the formation of quinone imine (compare with Figure 5). The amount of ammonia formed (250 pM at 60 min for p-anisidine) corresponds well with the amount of quinone imine formed (500pM at 60 min), since 1mol of ammonia is released for every 2 mol of anisidine metabolized to a quinone imine. Reaction of Anisidine Metabolites with Nucleophiles. We were specifically interested in the reactivity of the metabolites formed from both anisidine isomers. We studied the covalent binding of each isomer to both protein and DNA as well as the formation of covalent adducts with glutathione. As shown in Table 111,HRP-derived reactive metabolites from both isomers bound extensively to pro-
-HRP
+5 mM glutathione +1 mM ascorbic acid +1 mM p-anisidine
nmol bound/mn of Drotein ~
~~
42.2 f 1.0 3.5 f 0 3.1 f 0.1 2.5 f 0.1 3.1 f 0.1 74.9 f 1.4 3.2 f 0.2 3.2 f 0.3 1.6 f 0.1 23.7 f 0.1 16.4 f 3.0
a Complete reactions contained 1 mM anisidine (0.5pCi), 1 mM hydrogen peroxide, 0.1 pg of HRP, and 1 mg of heat-denatwed rat liver S9 protein in 1 mL of 0.1 M phosphate buffer (pH 7.0). Inhibitors were added at the concentrations indicated. Incubations were carried out at room temperature for 10 min and were stopped by the addition of 50 pL of 100% trichloroacetic acid.
Table IV. Effect of HRP Concentration on the Covalent Binding of Anisidine Isomers to Protein nmol bound/ma of protein HRP concn, pg/mL p-anisidine o-anisidine 0.1 0.2 0.5 1.0
7.3 f 1.1 10.7 f 0.9 15.2 f 2.1 18.0 f 0.7
8.5 f 0.5 15.9 f 0.4 21.4 f 0.6 23.2 f 0.9
" Reaction conditions are the same as described in the footnote to Table I11 except for the varying concentrations of HRP. The reactions were stopped after 0.5 min incubation time. tein. The ortho isomer bound to protein in higher amounts than the para isomer. Glutathione and ascorbic acid inhibited the covalent binding of each isomer. Ascorbic acid was not as effective, however, at inhibiting covalent binding from o-anisidine metabolites as with p-anisidine metabolites. Since the higher degree of protein binding with oanisidine compared to p-anisidine may simply be a consequence of the differences in rates of metabolism between the two isomers, we measured protein binding at various enzyme concentrations (Table IV). At comparable rates
Chem. Res. Toxicol., Vol. 4, No. 4, 1991 479
Anisidine Metabolism
Scheme I. Proposed Metabolic Pathways for the HRP-Dependent Oxidation of Anisidine Isomers
p-anisidine
BHA conjugate mw 360
o-anisidine
C(CH,r, &H&Y& azodimer mw242
BHA conjugate mw 390
H3C0
f
&* 6peroxidase,
GSH, Ascorbate
___)
0
o-anirldlne dilmlne
quinonelmlne
tetramer mw 488
of metabolism (e.g., 0.5 pg/mL HRP for o-anisidine and 0.1 pg/mL HRP for p-anisidine), or a t the same enzyme concentration, o-anisidine always gave more covalently bound product than did p-anisidine. This indicated that the reactive intermediates from o-anisidine are more reactive toward protein than the intermediates from p anisidine. When quinone imine from p-anisidine was isolated from these reactions and reacted with glutathione, the solution immediately lost ita reddish brown color, and glutathione conjugates were formed. With p-anisidine, two conjugates were detected by HPLC using [3H]glutathione as well as the reduced form of the quinone imine. The largest peak was isolated, and its mass spectra (FABMS) revealed it was a monoglutathione conjugate (MW 521, data not shown). The other conjugate may represent a diglutathione conjugate (25). The diimine metabolites were also reactive toward glutathione. When a typical reaction mixture containing both diimine and quinone imine was reacted with glutathione, both diimine and quinone imine
Table V. Covalent Binding of Anisidine Isomers to DNA reaction pmol bound/ma of DNA o-anisidine complete reactiona 293 f 112 +5 mM glutathione 467 f 128 +1 mM ascorbic acid nd p-anisidine complete reaction 318 f 89 +5 mM glutathione 1412 f 106 +I mM ascorbic acid 76 f 64 OComplete reaction contained 1 mM anisidine (1.0 pCi), 1 mM hydrogen peroxide, 0.1 pg/mL HRP, and 1 mg of calf thymus DNA in 1 mL of 0.1 M phosphate buffer (pH 7.0). Reactions were terminated after 30-min incubation at room temperature. Values represent mean fSE of triplicate determinations.
peaks on the HPLC chromatogram disappeared (data not shown). With o-anisidine, all four peaks representing quinone imine and diimine metabolites disappeared and glutathione conjugates were detected by using [3H]glutathione.
Thompson and Eling
480 Chem. Res. Tonicol., Vol. 4, No. 4, 1991
Covalent binding of reactive metabolites of anisidine isomers to DNA was also measured. Table V shows that reactive metabolites from both isomers readily bound to DNA. In these incubations, p-anisidine usually bound in slightly higher amounts to DNA than did o-anisidine. The diimine metabolite appeared to be responsible for most of the binding, since the isolated quinone imine did not undergo a significant color change when DNA was added to it (as occurred when glutathione was added to it). In the presence of ascorbic acid, the covalent binding was inhibited. However, when glutathione was present, the covalent binding was greatly increased, especially with p-anisidine.
Discussion We investigated the peroxidase-mediated oxidation of anisidine isomers using HRP as a model peroxidase. We identified reactive intermediates (diimine and quinone imine) as well as stable metabolites (dimers and polymers) from both isomers. The proposed major pathways of anisidine oxidation by HRP are depicted in Scheme I, as well as the structures of the metabolites. The major product of HRP oxidation of both isomers is a diimine metabolite, which further reacts with water to yield a quinone imine. With both isomers, coupling to form the diimine occurs at the para position. With p-anisidine this results in the elimination of methanol (22). With o-anisidine, methanol elimination results only if coupling occurs at the ortho position. If this occurs, it is a minor pathway, since we did not detect BHA-diimine adducts of the appropriate molecular weight with o-anisidine. The peroxidative oxidation of p-anisidine has been previously studied by Daniels and Saunders (22).They reported the HRP-catalyzed demethoxylation of p anisidine and isolated a tetrameric polymer as the major metabolite (80%). In addition, they isolated a pentameric polymer (tetramethoxyazophenine) as a minor metabolite ( 5 % ) and also identified the azo dimer in very small quantities. The incubation conditions contained a high amount of p-anisidine (2%) compared with our reaction conditions (1 mM or 0.012%). We observed both the tetramer and the am dimer metabolites in our incubations, and the amount formed (percent of total metabolites) increased as the concentration of anisidine increased. The tetramer was also identified as a metabolite of p-anisidine by using fungal phenol oxidase and laccase (13, 23). From a toxicologic and carcinogenic perspective, the most relevant metabolites formed via peroxidase-mediated activation of anisidine are the diimine and quinone imine. The formation of these intermediates suggest that the anisidine isomers would be highly toxic, since quinone imines from other aromatic amines have been implicated as possible toxic (26-30) and perhaps carcinogenic metabolites (25,31). Both of these anisidine metabolites were reactive with cellular nucleophiles. Metabolites from oanisidine appeared to be consistently more reactive with protein and glutathione than metabolites from p-anisidine. This was true whether the isomers were metabolized at the same rate or at equal enzyme concentrations (Table IV). In addition, the initial free radical metabolite of o-anisidine was more reactive toward glutathione than was the p anisidine free radical (Figure 1). With DNA as the target nucleophile, however, the covalent binding of anisidine isomers was quantitatively very similar. In the presence of glutathione, the covalent binding to DNA of both isomers was increased, particularly for p anisidine. This phenomena has been previously reported for p-phenetidine, which is an aromatic amine with a structure similar to that of anisidine (25). While gluta-
thione increased DNA binding of p-phenetidine, these authors reported that, in contrast, other thiols such as cysteine and N-acetylcysteine inhibited DNA binding under similar conditions. In addition, the glutathionedependent increase in p-phenetidine-DNA conjugates could be dialyzed out of the DNA down to levels obtained without glutathione, indicating that the conjugates were noncovalent. Therefore, we did not further investigate the anisidine-glutathione-DNA conjugates at this time. Our results point out some interesting similarities as well as differences in the HRP-dependent oxidation of anisidine isomers. The differences observed with HRP suggest that differences may also exist using mammalian peroxidases such as prostaglandin H synthase or renal or bladder tissue enzymes. Care must be exercised in extrapolating results from HRP studies to mammalian peroxidases, however, since recent studies have shown that some substrates, including aromatic amines, show very different affinities for PHS versus HRP (32). Other studies have shown that unique oxidation mechanisms with PHS (peroxyl radical mediated reactions) may lead to the formation of different metabolites that are not observed with HRP under similar reaction conditions.2 Further studies with tissue enzymes must be carried out, therefore, before the possible involvement of peroxidases in the carcinogenicityof anisidine can be adequately evaluated.
Acknowledgment. We thank Steve McGown for analyzing our samples on the mass spectrometer.
References (1) Miller, J. A. (1970)Carcinogenesis by chemicals-An overview. (G. H. A. Clowes Memorial Lecture). Cancer Res. 30,559-576.
(2) Miller, E. C. (1978)Some current perspectives on chemical carcinogenesis in humans and experimental animals: Presidential address. Cancer Res. 38,1479-1496. (3) Spitz, S.,Maguigan, W. H., and Dobriner, K. (1954)The carcinogenic action of benzidine. Cancer 3, 789-804. (4) Zenser, T. V.,and Davis, B.B. (1984)Enzyme systems involved in the formation of reactive metabolites in the renal medulla: Cooxidation via prostaglandin H synthase. Fundam. Appl. Toxicol. 4,922-929. (5) Robertson, I. G.C., Sivarajah, K., Eling, T. E., and Zeiger, E. (1983)Activation of some aromatic amines to mutagenic products by prostaglandin endoperoxide synthetase. Cancer Res. 43, 476-480. (6) Flammang, T. J., Yamazoe, Y., Benson, R. W., Roberte, D. W., Potter, D. W., Chu, D. 2. J., Lang, N. P., and Kadlubar, F. F. (1989)Arachidonic acid-dependent peroxidative activation of carcinogenic arylamines by extrahepatic human tissue microsomes. Cancer Res. 49, 1977-1982. (7) Mohandas, J., Duggin, G. G., Horvath, J. S., and Tiller, D. J. (1981)Metabolic activation of acetaminophen (paracetamol) mediated by cytochrome P-450mixed-function oxidase and prostaglandin endoperoxidesynthetase by rabbit kidney. Toxicol. Appl. Pharmacol. 61, 252-259. (8) Bull, A. W. (1987)Reducing substrate activity of some aromatic amines for prostglandin H synthase. Carcinogenesis 8,387-390. (9) Kadlubar, F. F.,Frederick, C. B., Weis, C. C., and Zenser, T. V. (1982)Prostaglandin endoperoxide synthetase-mediated metabolism of carcinogenic aromatic amines and their binding to DNA and protein. Biochem. Biophys. Res. Commun. 108, 253-258. (10) Petry, T. W., Josephy, P. D., Pagano, D. A., Zeiger, E., Knecht, K. T., and Eling, T. E. (1989)Prostaglandin hydroperoxidasedependent activation of heterocyclic aromatic amines. Carcinogenesis 10,2201-2207. (11) Wise, R.W., Zenser, T. V., Rice, J. R., and Davis, B. B. (1986) Peroxidatic metabolism of benzidine by intact tissue: A prostaglandin H synthase-mediated process. Carcinogenesis 7,111-115. (12) National Cancer Institute (1978)Bioassay of o-anisidine hydrochloride for possible carcinogenicitjj. Technical Report No. 89, National Cancer Institute, Bethesda, MD. ~~
* J. Curtis and T. Eling, unpublished observations.
Anisidine Metabolism (13) Department of Health, Education, and Welfare (1978)Bioassay of p-anisidine hydrochloride for possible carcinogenicity. DHEW Publication No. 78-1371,DHEW, Rockville, MD. (14) Hoff, T. Lui, S.-Y., and Bollag, J.-M. (1985)Transformation of halogen, alkyl-, and alkoxy-substituted anilines by a laccase of Trametes versicolor. Appl. Environ. Microbiol. 49, 1040-1045. (15) Lowry, 0.H., Rosebrough, N. J., Farr, A. L., and Randall, R. J. (1951)Protein measurement with the Folin phenol reagent. J. Biol. Chem. 193,265-275. (16) Ngo, T. T., Phan, A. P. H., Yam, C. F., and Lenhoff, H. M. (1982)Interference in determination of ammonia with the hypochlorite-alkaline phenol method of Berthelot. Anal. Chem. 54, 46-49. (17) Weller, P. E., Markey, C. M., and Marnett, L. J. (1985)Enzymatic reduction of 5-phenyl-4-pentenyl-hydroperoxide:Detection of peroxidases and identification of peroxidase reducing substrates. Arch. Biochem. Biophys. 243,633-643. (18) Berry, D. F., and Boyd, S. A. (1984) Oxidative coupling of phenols and anilines by peroxidase: structure-activity relationships. Soil Sci. SOC.Am. J. 48,565-569. (19) Boyd, J. A., and Eling, T. E. (1985)Metabolism of aromatic amines by prostaglandin H synthase. Environ. Health Perspect. 64,45-51. (20) Josephy, P. D., Mason, R. P., and Eling, T. E. (1982)Cooxidation of the clinical reagent 3,5,3’,5’-tetramethylbenzidineby prostaglandin synthase. Cancer Res. 42,2567-2570. (21) Ross,D., Norbeck, K., and Moldbus, P. (1985)The generation and subsequent fate of glutathionyl radicals in biological systems. J. Biol. Chem. 260, 15028-15032. (22) Daniels, D. G.H., and Saunders, B. C. (1951)Studies in peroxidase action. Part VI. (a) The oxidation of p-anisidine: A further example of ready demethoxylation. (b) Some theoretical considerations. J. Chem. SOC.,2112-2118. (23) Sjoblad, R. D., and Bollag, J.-M. (1977)Oxidative coupling of aromatic pesticide intermediates by a fungal phenol oxidase. Appl. Environ. Microbiol. 33,906-910.
Chem. Res. Toxicol., Vol. 4, No. 4, 1991 481 (24) Josephy, P.D., and van Damme, A. (1984)Reaction of 4-substituted phenols with benzidine in a peroxidase sytem. Biochem. Pharmacol. 33, 1155-1156. (25) Andersson, B., Larsson, R., Rahimtula, A., and Moldbus, P. (1984)Prostaglandin synthase and horseradish peroxidase catalyzed DNA-binding of p-phenetidine. Carcinogenesis 5,161-165. (26) Rundgren, M., Porobek, D. J., Harvison, P. J., Cotgreave, I. A., Moldbus, P., and Nelson, S. D. (1988)Comparative cytotoxic effects of N-acetyl-p-benzoquinone imine and two dimethylated analogues. Mol. Pharmacol. 34,566-572. (27) Larsson, R.,Ross,D., Nordenskjijld, Lindeke, B., Oleeon, L.-I., and Moldbus, P. (1984)Reactive products formed by peroxidase catalyzed oxidation of p-phenetidine. Chem.-Biol. Interact. 52, 1-14. (28) Ross, D., Larsson, R., Andersson, B., Nilason, U.,Lindquist, T., Lindeke, B., and Moldbus, P. (1985) The oxidation of p phenetidine by horseradish peroxidase and prostaglandin synthase and the fate of glutathione during such oxidations. Biochem. Pharmacol. 34,343-351. (29) Larsson, R.,Lindqvist, T., Lindeke, B., and Moldbus, P. (1986) Cellular effects of N-(4-ethoxyphenyl)-p-benzoquinoneimine. A p-phenetidine metabolite formed during peroxidase reactions. (!hem.-Biol. Interact. 60,317-330. (30) Dugue, B., Paoletti, C., and Meunier, B. (1984)Covalent binding of the antitumor agent P-methyl-9-hydroxy-ellipticinium acetate (NSC 264137) on RNA and poly A in vitro. Biochem. Biophys. Res. Commun. 124,416-422. (31) Zenser, T. V.,Mattammal, M., Armbrecht, H., and Davis, B. (1980)Benzidine binding to nucleic acids mediated by the peroxidative activity of prostaglandin endoperoxide synthetase. Cancer Res. 40,2839-2845. (32) Markey, C. M., Alward, A,, Weller, P. E., and Marnett, L. J. (1987)Quantitative studies of hydroperoxide reduction by prostaglandin H synthase. Reducing substrate specificity and the relationship of peroxidase to cyclooxygenase activities. J. Biol. Chem. 262,6266-6279.