Chem. Res. Toxicol. 1992, 5, 10-19
10
Articles A New Hypothesis for the Mechanism for Cytochrome P-450 Dependent Aerobic Conversion of Hexahalogenated Benzenes to Pentahalogenated Phenols Ivonne M. C. M. Rietjens* and Jacques Vervoort* Department of Biochemistry, Agricultural University, Wageningen, The Netherlands Received April 24, 1991 The mechanism for cytochrome P-450 dependent conversion of hexahalogenated benzenes was investigated. This was done mainly by studying the in vitro and in vivo biotransformation of pentafluorochlorobenzene using 19FNMR. Identification of the metabolites formed in vivo from pentafluorochlorobenzene demonstrated that the fluorine atom para with respect to the chlorine substituent was preferentially eliminated during biotransformation. 19F NMR data also demonstrated that this fluorine atom is eliminated as a fluoride anion. In vitro studies on the biotransformation of pentafluorochlorobenzenedemonstrated that defluorination was oxygen nor of an dependent and resulted in formation neither of 2,3,5,6-tetrafluoro-l-chlorobenzene amount of hydroxylated metabolites that could account for the fluoride anion formation. This formation of a high amount of fluoride anion, which was not accompanied by formation of a similar amount of other '?F NMR detectable products, demonstrates that the mechanism involved should also explain formation of a highly reactive intermediate that ends up bound to cellular macromolecules. On the basis of the results obtained, three of the four mechanisms proposed in the literature for the mechanism of the cytochrome P-450 dependent conversion of hexahalogenated benzenes to pentahalogenated phenols could be eliminated. Finally, results from molecular orbital computer calculations were compared to results from in vivo biotransformation studies for polyhalogenated benzenes. On the basis of the results obtained a new hypothesis for the mechanism of cytochrome P-450dependent conversion of hexahalogenated benzenes to pentahalogenated phenols was proposed. This mechanism suggests that the reaction might proceed by elimination of a halogen anion and formation of a benzohaloquinone with a positive charge on a halogen substituent para with respect to the dehalogenated position. In a subsequent chemical reduction this benzohaloquinone could be reduced by, for instance, NADPH to give the fiial pentahalogenated phenol, although due to its high reactivity it might also end up bound to cellular macromolecules.
Introduction The accumulation of halogen-containing aliphatic and aromatic compounds is a major factor adding to environmental pollution. The phenomenon originates from the widespread use of halogenated compounds in industry, commerce, and medicine and from the relatively large persistence of these compounds ( I , 2). For aromatic POlyhalogenated benzenes the relatively high persistence may be related to the metabolically difficult dehalogenation steps necessary to produce water-soluble metabolites. In mammalian liver, dehalogenation of the aromatic halogenated compounds is known to be cytochrome P-450 mediated ( I , 2 ) . Recently, we proposed a mechanism for the dehalogenation of fluorinated anilines (3, 4 ) . This mechanism appeared also to be valid for halogenated phenols.' The mechanism is summarized in Figure 1. As a result of cytochrome P-450dependent monooxygenation, the halogen atom para with respect to the amino or hydroxyl moiety is released from the molecule as a halogen *Address correspondence to these authors at the Department of Biochemistry, Agricultural University, Dreijenlaan 3, NL-6703 HA Wageningen, The Netherlands.
anion, leaving a benzoquinone (imine) derivative as the primary aniline- or phenol-derived reaction product. In the presence of reducing equivalents, like NAD(P)H,2the benzoquinone (imine) is chemically reduced to give the corresponding aminophenol or hydroquinone (Figure 1) (3,4). Although the above-mentioned mechanism could explain the oxidative dehalogenation of ortho- or para-halogenated phenols and anilines, the mechanism for dehalogenation at a meta position or of a fully halogenated benzene remains to be solved. Five mechanisms can be proposed for cytochrome P-450dependent dehalogenation of hexahalogenated benzenea, four of which were suggested before in the literature. These hypotheses are summarized in Figure 2. The first mechanism (mechanism 1) is based
' Den Besten et al., in preparation.
~~
Abbreviations: NMR, nuclear magnetic resonance; HXO, hypohalous acid (for example, HClO = hypochlorous acid); NADPH, reduced nicotinamide adenine dinucleotide phosphate; NADH, reduced nicotinamide adenine dinucleotide;NAD(P)H,NADH or NADPH; CYCLOPS, cyclically ordered phase sequence; GSH, reduced glutathione; HOMO, highest occupied molecular orbital, LUMO, lowest unoccupied molecular orbital; SOMO, single occupied molecular orbital; E(HOMO),energy of the HOMO; E(LUMO), energy of the LUMO HF, heat of formation.
0 1992 American Chemical Society 0893-228~/92/2705-0010$03.00/0
Chem. Res. Toxicol., Vol.5, No.1, 1992 11
Conversion of Hexahalogenated Benzenes by Cyt P-450
tp/ OH
other reductors) will produce the pentahalogenated phenol. This mechanism provides a hypothesis for the conversion of hexahalogenated benzenes to pentahalogenated phenols not suggested before in the literature for this reaction. In mechanism 4 (Figure 2) the cytochrome P-450 dependent monooxygenation proceeds by two subsequent cytochrome P-450 dependent reactions. First, a so-called reductive dehalogenation occurs, resulting in formation of a pentahalogenated benzene, followed by a cytochrome P-450 dependent monooxygenation at the dehalogenated position. Such a mechanism can be suggested on the basis of results published by Stewart et al. (61,who reported the formation of pentachlorobenzene upon microsomal conversion of hexachlorobenzene. In contrast to this,however, data have also been reported that excluded this mechanism, a t least for microsomal conversion of hexachlorobenzene to pentachlorophenol (7). Finally, a mechanism can be proposed that proceeds by formation of an epoxide that loses a halogen anion in a subsequent chemical reduction reaction with NAD(P)H (mechanism 5). Formation of an epoxide from (halogenated) aromatic compounds has been suggested before in the literature (5,8,9). Some authors even suggested the formation of a 1-keto-2,2,3,4,5,6-hexahalo-3,5-cyclodiene from the epoxide by an NIH shift before loas of the halogen occurs (5). However, for fluorine and iodine substituents it has been reported that they are lost from the molecule rather than giving rise to an NIH shift (9, lo), making the cyclodiene intermediate perhaps only valid for chlorinated or bromated compounds. In addition, loss of a halogen anion from the hexahalogenated epoxide without concomitant reduction would give rise to the benzohaloquinone cation intermediate
NH:! / OH I
Figure 1. Mechanism proposed for the cytochrome P-450dependent dehalogenation of ortho- and para-halogenated phenols and anilines ( 3 , 4 ) . X = F or C1.
on the mechanism proposed for the cytochrome P-450 dependent dehalogenation of hexachlorobenzene, resulting in formation of pentachlorophenol and elimination of a chlorine cation (5). Elimination of this halogen cation follows from the electron balance of the cytochrome P-450 monooxygenase reaction, providing two electrons, and not the four that would be necessary for elimination of a halogen anion in combination with formation of the phenol. Mechanism 2 (Figure 2) proceeds by oxygenation of the halogen substituent, resulting in an intermediate that produces the pentahalogenated phenol and eliminates the halogen as HXO (for example, HClO = hypochlorous acid) in a reaction with a water molecule (1, 5 ) . Mechanism 3 proceeds by oxidative dehalogenation of a fully halogenated benzene and is based on our model for oxidative dehalogenation of anilines and phenols (Figure 1)(3,4). It includes elimination of the halogen as a halogen anion and formation of an oxidized benzohaloquinone, with a positive charge on the halogen substituent para with respect to the dehalogenated position. This pmitive charge is stabilized through mesomeric structures. A chemical two-electron reduction of this compound by NAD(P)H (or
i
X'
xo
X 2/@XX
.
X
p450
7-
Xx@:
X
+HXo
YO
X
X
H
x+:
X ' +
P450
*;*x
X
X X
5)
Xx@x X
X
p450 XxQx
X
X
X
7-7 NAD(P)H
X'
t
XlQX X X
NAD(P)'
Figure 2. Hypotheses for the cytochrome P-450dependent dehalogenation of hexahalogenated benzenes. For explanation see text. X = F or C1.
12 Chem. Res. Toxicol., Vol. 5, No. 1, 1992 suggested to occur in mechanism 3. The present investigation was undertaken to study t h e likeliness of these possible mechanisms for cytochrome P-450 dependent hydroxylation at a halogenated position in hexahalogenated benzenes. This was done by studying the in vitro and in vivo biotransformation of hexahalogenated benzenes using 19F NMR. T h e results obtained provide evidence which unequivocally eliminates three of the five mechanisms and points at a new hypothesis for t h e cytochrome P-450 dependent conversion of hexahalogenated benzenes to pentahalogenated phenols.
Experimental Procedures Chemicals. 2,3,5,6-Tetratluorophenol, hexafluorobenzene, and 2,3,5,6-tetrafluorohydroquinonewere purchased from Aldrich (Steinheim, FRG). Pentafluorophenol and 2,3,5,6-tetrafluoro1-chlorobenzene were from Sigma (St. Louis, MO). Pentafluorochlorobenzene, 1,3,5-trifluoro-2,4,6-trichlorobenzene, pentafluorobenzene, and 2,3,5,6-tetrafluoro-l-chlorobenzene were purchased from Fluorochem (Derbyshire, U.K.). 2,3,5,6-Tetrafluoro-4-chlorophenolwas synthesized from pentafluorochlorobenzene essentially following the procedure described for synthesis of 2,4-dichlorophenol from l-fluoro-2,4-dichlorobenzene (11,12). In short, 1g of pentafluorochlorobenzene was added to 20 mL of 3 M sodium methanolate in methanol. The reaction mixture was refluxed a t 120 OC, for 3.5 h. During overnight cooling the residual benzene was evaporated from the mixture. The 2,3,5,6-tetrafluoro-4-chlorophenol formed in this reaction mixture was 7% of all products detected by 19FNMR and was separated from d l other fluorine-containing reaction products on a silica gel column (1 cm diameter, 20 cm height) (230-400 mesh ASTM) (Merck, Darmstadt, FRG) eluted with petroleum ether with slowly increasing ( 0 4 % ) percentage of ethanol. The fractions that were pure on the basis of 19FNMR analysis were pooled. The 2,3,5,6-tetrafluoro-4chlorophenolthus prepared was pure as judged from 19FNMR spectra. The compound was identified on the basis of its proton-decoupled and proton-coupled '9 NMR spectra and on the basis of results from an arylsulfatase/& glucuronidase-treated urine sample from a rat exposed to 2,3,5,6-tetrafluoro-l-chlorobenzene. Proton-decoupled '9 NMR spectra of the synthesized compound in 0.1 M potassium phosphate, pH 7.6, showed two resonances of equal intensity at -152.8 and -169.5 ppm and with a JF-Fof 18.3 Hz. Proton-coupled spectra were similar, demonstrating that the compound does not contain an aromatic proton substituent. The 19FNMR spectra demonstrated that the four fluor substituents are present in two equivalent pairs and that therefore the fluorine atom at C4 has been eliminated from 2,3,4,5,6-pentafluoro-l-chlorobenzene. Further identification of the compound was based on the analysis of the urine from a rat exposed to 2,3,5,6-tetrafluoro-lchlorobenzene. For this pentahalogenated benzene hydroxylation at the nonsubstituted position in the aromatic ring will be a major biotransformation route. 19FNMR analysis of the ethyl acetate extract of the arylsdfatase/&glucuronidase-treated urine (0-24 h) from a 2,3,5,6-tetrafluoro-l-chlorobenzene-exposed rat (diluted in 0.1 M potassium phosphate, pH 7.6, prior to extraction) demonstrated the presence of only one main (>SO%) ethyl acetate extractable product having l9FNMR characteristics which are completely identical to those of the synthesized 2,3,5,6-tetrafluoro-4-chlorophenol. This further supports the identification of the 2,3,5,6-tetrafluoro-4-chlorophenol resonances at -152.8 and -169.5 ppm. In Vivo Exposure of Rats to Halogenated Benzenes or Phenols. Male Wistar rats (300-350 g) were exposed to 300 pmol of the desired halogenated compound/kg body weight administered in olive oil by oral injection. After dosing, 24-h urine samples were collected for 7 days. Blanks received olive oil alone. Analysis of Urine Samples. Urine samples were analyzed by lgF NMR, generally after 1:l dilution in 0.2 M potassium phosphate, pH 7.6. If necessary, urine samples were concentrated 4 times by freeze-drying. For quantification, 4-fluorobenzoicacid was added to the samples at a final concentration of 58.5 pM. Concentrations of the various metabolites and thus total recoveries
Rietjens and Veruoort could be calculated by comparison of the integrals of the 'BF NMR resonances of the metabolites to the integral of the 19F NMR resonance of 4-fluorobenzoic acid. Enzyme hydrolysis of urine samples was carried out essentially as described before (13). In short, for &glucuronidase treatment, 1.2 mL of 0.2 M KH2PO4/Na2HPO4,pH 6.2, containing 8 units of &glucuronidase (from Escherichia coli K12) (Boehringer, Mannheim, FRG) was added to 1.2 mL of urine sample. The mixture was made oxygen free by four cycles of evacuating and flushing with argon and incubated at 37 "C for 16 h. For arylaulfatase/@-glucuronidase treatment 40 pL of the enzyme mixture from Helix pomatia (Boehringer, Mannheim, FRG) was added to 1.2 mL of urine sample diluted with 1.2 mL of 0.1 h4 potassium acetate, pH 5.2. Samples were made oxygen free and incubated at 37 OC for 16 h. 19FNMR Measurements. 19F NMR measurements were performed on a Bruker CXF' 300 NMR spectrometer as described before (3,4,13).Norell (Landisville,NJ) 1 0 " NMR tubes were used. The sample volume was 1.71 mL, containing 100 pL of %I20 for locking the magnetic field. Proton decoupling was achieved with the Waltz-16 pulse sequence ( 1 4 ) at -16 dB from 20 W. Nuclear Overhauser effects were eliminated using the inverse gated decoupling technique. Spectra were obtained with 25-30° pulses (6 pa), a 50-Wz spectral width, repetition time of 1s, quadrature phase detection, and quadrature phase cycling (CYCLOPS). Between 5000 and 50 OOO scans were recorded, depending on the concentration of fluorine metabolites present and the signal to noise ratio required. Chemical shifts are reported relative to CFC13. Preparation of Microsomes. Microsomes were prepared as described before (10)from the perfused livers of male Wistar rats (300-350 g) pretreated with dexamethasone (Sigma, St. Louis, MO) (300 mg/kg body weight, using a stock solution of 90 mg/mL in water containing 2% Tween-80, administered by oral injection, daily for 4 days). Microsomal cytochrome P-450 content was measured as described by Omura and Sat0 (15). Cytosol was obtained as the supematant from the f i t 105OOOg ultracentrifuge step performed during preparation of microsomes from untreated, control rata. Protein content was determined by the method of Lowry et al. (16) using bovine serum albumin (Sigma, St. Louis, MO) as the standard. Microsomal Incubations. Microsomal incubations were carried out at 37 OC in 0.1 M potassium phosphate, pH 7.6, containing 2.0 nmol of microsomal cytochrome P-450/mL and 0.3 mM halogenated benzene, added to the incubation mixture as 1% v/v of a 30 mM stock solution in ethanol. To some incubations, vitamin C was added at a final concentration of 1.0 mM to make sure that products from monooxygenation would be reduced as efficiently as possible. The reactions were started by addition of NADPH (2 mM final concentration) and carried out in a closed reaction vessel to prevent evaporation of the substrate. Reactions were terminated after 20 min by freezing the mixtures into liquid nitrogen. Blanks were carried out by adding distilled water instead of NADPH. Cytosolic Incubations. Cytosolic incubations with pentawere fluorochlorobenzene and 1,3,5trifluoro-2,4,6-trichlorobe~ne carried out in 0.1 M sodium phosphate, pH 6.5, in the presence of 3.5 mg/mL cytosolic protein, 1mM glutathione in its reduced form (GSH) (Sigma, St. Louis, MO), and 0.3 mM of the hexahalogenated benzene added as 1% v/v of a 30 mM stock solution in ethanol. The incubation was started by the addition of the benzene and carried out at 37 "C for 20 min in a closed reaction vessel to prevent evaporation of the substrate. The incubations with hexafluorobenzene were carried out in the presence of twice the amount of cytosolic protein (7.0 mg/mL) for 180 min. This was necessary because conversion of hexafluorobenzene appeared to occur at a much lower rate. Molecular Orbital Computer Calculations. Computer calculations were performed on a Silicon Graphics Iris 4D/% using Quanta/Charmm (Polygen Inc., Reading, U.K.). The semiempirical molecular orbital method was used, applying the AM1 Hamiltonian from the AMPAC program (Quantum Chemistry Program Exchange No. 506, Indiana University, Bloomington, IN). All calculations were carried out with PRECISE criteria. For all calculations the self-consistent field was achieved. Geometries were optimized for all bond lengths, bond angles, and torsion
Chem. Res. Toxicol., Vol. 5, No. 1, 1992 13
Conversion of Hexahalogenated Benzenes by Cyt P-450 oso, F
a
a
F
F
F
F
F
F
F
F
F
F r f l
F
is
F
F
F
F
F
F F
r-+l
L
Lil
F
F
is
-120
-140
-160
-180
PPM Figure 3. lgF NMR spectrum of the urine from a rat exposed to (a) pentafluorophenol or (b) 2,3,5,6-tetrafluorophenol.Resonances were identified as described in the text. The resonance marked "is"is from the internal standard 4-fluorobenzoicacid.
angles using the Fletcher-Powell criteria. Calculation of log Poetanol Values. log P-,,] values were calculated by the method of Rekker and de Kort (17) and represent the log of the partition ratio of the compound over octanol and water.
Results In Vivo Biotransformation of Pentafluorophenol and 2,3,5,6-Tetrafluorophenol.To identify 19FNMR resonances of theoretically possible pentafluorochlorobenzene metabolites, urine metabolites formed upon exposure of the rata to pentafluorophenol or 2,3,5,6-tetrafluorophenol were identified. Figure 3a presents the 19F NMR spectrum of the 24-h urine of a rat exposed to pentafluorophenol. Metabolites could be identified on the basis of the 19FNMR spectra of the same urine treated with either 8-glucuronidase or 8-glucuronidaselarylsulfatase preparations. Upon 8-glucuronidase treatment the resonances at -159.5, -165.4, and -167.7 ppm disappeared, giving rise to three new resonances at -171.8, -173.2, and -185.9 ppm, which-on the basis of the spectrum of the reference compound-represent pentafluorophenol. This identifies the resonances of the pentafluorophenyl glucuronide. Treatment of the urine sample with @-glucuronidase/ arylsulfatase caused not only the disappearance of the pentafluorophenyl glucuronide resonances but also the disappearance of the resonances at -158.1, -162.7, and -167.1 ppm, with a concomitant proportional additional increase of the pentafluorophenol resonances. This identifies the 19FNMR resonances of pentafluorophenyl sul-
a
1
-120
-140
-160
-180
PPM Figure 4. 'BF NMR spectrum of the concentrated urine of a rat exposed to (a) heduorobenzene or (b) pentafluomhlorobenzene. Resonance were identified as described in the text. The resonance marked Ys" is from the internal standard 4-fluorobenzoicacid.
fate. In the same way the resonances of the glucuronide and sulfate conjugate of 2,3,5,6-tetrafluorophenoland of the sulfate conjugate of 2,3,5,6-tetrafluorohydroquinone could be identified using the urine of a 2,3,5,6-tetrafluorophenol-exposed rat (Figure 3b)The 19F NMR resonances of fluoride anion, hexafluorobenzene, pentafluorochlorobenzene, 2,3,5,6-tetrafluoro-1-chlorobenzene,pentafluorophenol, 2,3,5,6-tetrafluorophenol, and 2,3,5,6-tetrafluorohydroquinonewere identified using the commercially available reference compounds. The lBFNMR resonances of 2,3,5,6-tetrafluoro-4chlorophenol were identified after synthesis of the compound from pentafluorochlorobenzene. Table I summarizes the 19F NMR resonances thus identified. In Vivo Metabolism of Hexafluorobenzene. Figure 4a shows the 19FNMR spectrum of the 4 times concentrated 24-h urine from a hexafluorobenzene-exposedrat. Some of the identified metabolites summarized in Table I were readily observed, demonstrating that in vivo hexafluorobenzene is indeed defluorinated in a monooxygenation reaction to give pentafluorophenol. The pentafluorophenol formed is preferentially sulfated and excreted into urine. Metabolites from secondary monooxygenation reactions, i.e., 2,3,5,6-tetrafluorohydroquinone derivatives, were not observed. This is not surprising taking into account the results obtained in in vivo experiments with pentafluorophenol where less than 2% of
Rietjens and Vervoort
14 Chem. Res. Toxicol., Vol. 5, No. 1, 1992 Table I. Chemical Shifts of the I9F NMR Resonances of the Fluorinated Benzenes, Phenols, and Their Metabolic Derivatives. Identified in the Present Study" compound chemical shift, ppm fluoride anion -123.0 hexafluorobenzene -167.0 -145.5 (F2,F6); -160.6 (F4); pentafluorochlorobenzene -165.8 (F3,F5) 2,3,5,6-tetrafluoro-l-chloro- -142.7 (F3,F5);-146.0 (F2,F6) benzene -171.8 (F2,F6);-173.2 (F3,F5); pentafluorophenol -185.9 (F4) -149.3 (F3,F5);-170.6 (F2,F6) 2,3,5,6-tetrafluorophenol 2,3,5,6-tetrafluoro-4-chloro- -152.8 (F3,F5);-169.5 (F2,F6) phenol -171.2 (F2,F3,F5,F6) 2,3,5,6-tetrafluorohydroquinone -143.7 (F3,F5);-158.2 (F2,F6) 2,3,5,6-tetrafluorophenyl sulfate -144.2 (F3,F5);-159.8 (F2,F6) 2,3,5,6-tetrafluorophenyl glucuronide 2,3,5,6-tetrafluoro-4-hydroxy- -164.5 (F2,F6);-171.5 (F3,F5) phenyl sulfate -159.5 (F2,F6);-165.4 (F4); 2,3,4,5,6-pentafluorophenyl glucuronide -167.7 (F3,F5) -158.1 (F2,F6);-162.7 (F4); 2,3,4,5,6-pentafluorophenyl sulfate -167.1 (F3,F6)
8
-120
-140
-160
1
-180
PPM Figure 5. 19FNMR spectrum from a cytosolic GSH containing incubation with pentafluorochlorobenzene, identifying the 19F NMR resonances of its glutathione adduct.
In addition to the 2,3,5,6-tetrafluoro-4-chlorophenol derivatives, I9FNMR resonances from two additional main metabolites are observed in the urine 19FNMR spectrum from a pentafluorochlorobenzene-exposedrat (Figure 4b). "Chemical shifts were determined in 0.1 M potassium phosFirst, the 19FNMR resonance at -123.0 ppm can be asphate, pH 7.6 at 7 "C, and presented relative to CFC13. Resocribed to fluoride anion, as this resonance increases upon nances could be ascribed to specific fluorine substituents on the addition of potassium fluoride. From the results obtained basis of (i) relative signal intensities, (ii) proton-coupled and proit was calculated that for pentafluorochlorobenzene the ton-decoupled spectra, and (iii) the knowledge that meta substitufluoride anion elimination was 2.0 f 0.2 times the amount ents only slightly influence I9F NMR chemical shifts, ortho and para substituents having a much larger effect on the "F NMR of tetrafluorinated urine metabolites. In addition, two resonances of fluorine substituents in the aromatic ring. resonances of equal intensity, at -136.9 and -145.0 ppm, are observed, which demonstrate a splitting pattern characteristic for an AA'XX' spin system. This indicates the phenol is finally converted to 2,3,5,6-tetrafluorothat it must be a 2,3,5,6-tetrafluorobenzenederivative. hydroquinone metabolites (Figure 3a). The resonances Because the proton-coupled spectrum shows a similar ascribed to a pentafluoro sulfur substituted metabolite are splitting pattern, it can be concluded that neither the 1identifed on the basis of the cytosolic incubation with nor the 4-position is occupied by a proton. Further results, hexafluorobenzene, described below. described below, demonstrated that the resonances repFrom the spectrum presented it can be seen that the resent a C1 sulfur substituted 2,3,5,6-tetrafluoro-4-chloro amount of fluoride anion excreted is higher than the metabolite from pentafluorochlorobenzene. amount of fluorinated metabolites present in urine. After Conversion of Pentafluorochlorobenzene and correction for the low amount of fluoride anion excretion Hexafluorobenzene by Cytosolic Glutathione S detected in the urine from untreated control rats it can Transferases. Figure 5 presents the 19FNMR spectrum be calculated that the amount of fluoride anion is 6.7 f from a GSH-containing cytosolic incubation with penta0.1 times the amount of pentafluorinated metabolites exfluorochlorobenzene. This incubation was carried out to creted. identify the 19FNMR resonances of sulfur-substituted In Vivo Metabolite Pattern of Pentafluorochloropentafluorochlorobenzene. In the presence of GSH, cybenzene. Figure 4b presents the 19FNMR spectrum of tosolic glutathione S-transferases might catalyze the forthe 4 times concentrated urine of a pentafluorochloromation of glutathione adducts from the hexahalogenated benzene-exposed rat. The spectrum does not show any of benzene (5). From Figure 5 it can be derived that in these the 19FNMR resonances observed in the urine metabolite incubations indeed a new metabolite is formed. Three '9F pattern of a hexafluorobenzene-exposedrat (Figure 4a). NMR resonances were observed that are not present in Nevertheless,fluorine-containingmetabolites and fluoride control incubations carried out in the absence of GSH. anion are clearly present in the urine (Figure 4b). The resonance at -123.0 ppm again represents fluoride Treatment of the urine with @-glucuronidaseeliminated anion, which is apparently eliminated from pentafluorothe two resonances at -146.6 and -158.7 ppm, giving rise chlorobenzene upon glutathione addition. The resonances to a proportional increase of the resonances at -152.8 and -169.5 ppm identified as 2,3,5,6-tetrafluoro-4-chlorophenol at -136.9 and -145.0 ppm show an AA'XX' splitting pattern, which does not change in proton-coupled spectra, (Table I). Treatment with arylsulfatase/@-glucuronidase demonstrating that they represent a 2,3,5,6-tetrafluoro additionallyeliminated the resonances at -145.9 and -157.0 derivative, which must be 2,3,5,6-tetrafluoro-4-chloroppm with a proportional increase of the 2,3,5,6-tetrabenzene conjugated with glutathione at the C1 position. fluoro-4-chlorophenol 19FNMR resonances. Because these resonances are at similar positions and show The results presented demonstrate that upon cytoa similar splitting pattern as the ones for the "unknown" chrome P-450 dependent monooxygenation of pentametabolite in the urine metabolite pattern of Figure 4b, fluorochlorobenzene, the fluorine atom para with respect the resonances in the urine spectrum of Figure 4b are to the chlorine substituent is preferentially eliminated from the molecule, resulting in 2,3,5,6-tetrafluoro-4-chloro- ascribed to a sulfur-substituted urine metabolite derived from a glutathione adduct of pentafluorochlorobenzene. phenol, which is subsequently sulfated or glucuronidated Modifications in the glutathione side chain, expected to and excreted into urine.
Chem. Res. Toxicol., Vol. 5, No. 1, 1992 15
Conversion of Hexahalogenated Benzenes by Cyt P-450
I
I
F
is
I
I
I
-120 I
I
-120
!
-140
I
-160
-180
PPM Figure 6. I9F NMR spectrum of an aerobic microsomal incubation with pentafluorochlorobenzene. Liver microsomes from a dexamethasone-pretreatedrat were used. The arrows mark the the resonance positions for 2,3,5,6-tetrafluoro-4-chlorobenzene, product that would result from reductive dehalogenation.
occur when the glutathione adduct is metabolized in the mercapturic acid pathway, will not affect the electronic density in the benzene ring and thus will not affect the chemical shift of the aromatic fluorine substituents. In a similar way, the 19FNMR resonances at -137.0, -156.2,and -165.4 ppm in Figure 4a can be ascribed to a sulfur-substituted pentafluorobenzene derivative. Microsomal Metabolism of Pentafluorochlorobenzene. The results described above clearly point at a mechanism for dehalogenation of pentafluorochlorobenzene in which the fluorine and not the chlorine substituent is preferentially eliminated from the molecule. This excludes mechanisms 1 and 2,as for these reactions preferential elimination of the chlorine substituent would occur (see Discussion). Further experiments were undertaken to investigate the contribution of mechanism 3, 4,or 5 (Figure 2) to this reaction. Figure 6 shows the 19F NMR spectrum of an aerobic microsomal incubation with pentafluorochlorobenzene. Formation of fluoride anion is readily observed. No resonances were observed at -142.7 and -146.0 ppm representing the product from reductive dehalogenation, i.e., 2,3,5,6-tetrafluoro-l-chlorobenzene. A small amount of 2,3,5,6-tetrafluorochlorophenol can be observed in the spectrum (Figure 6). When the microsomal incubation was carried out in the absence of oxygen, conditions favoring reductive dehalogenation, fluoride anion formation was less than 3% of the fluoride anion formation under aerobic conditions. Furthermore, again no 2,3,5,6tetrafluoro-1-chlorobenzeneformation could be observed. These results eliminate mechanism 4 for cytochrome P-450 dependent pentafluorochlorobenzene conversion. From the spectrum presented in Figure 6 it also follows that the amount of fluoride anion formed is about 50 times the amount of metabolite(s) detected. This points at formation of a reactive metabolite in the microsomal system that readily binds to microsomal protein. As a result, the molecular mass of the metabolite increases substantially, resulting in such a line broadening of ita 19F NMR resonance(s) (decrease in T2) that the signal(s)can
I
1
-140
PPM Figure 7. 19F NMR spectrum of the urine from a rat exposed to 1,3,5-trifluoro-2,4,6-trichlorobenzene. Resonances were identified as described in the text. The resonance marked “is”is from the internal standard 4-fluorobenzoicacid. no longer be observed in the ’%’ NMR spectrum. Further experimenta demonstrated that the intermediate formed must be highly reactive as addition of vitamin C to the microsomal incubation mixture did not change the metabolite pattern observed. The protein binding observed can be the result of two possible bioactivation pathways. First, the benzohaloquinone cation intermediate of mechanism 3 would be a likely candidate for protein binding, losing an additional fluorine substituent as fluoride anion upon protein binding. The mechanism would be comparable to the protein binding of halogenated benzoquinone (imines) reported before in the literature (3,5).However, protein binding might also be ascribed to the epoxide intermediate of mechanism 5. Comparison of the in Vivo Biotransformation of Hexafluorobenzene, Pentafluorochlorobenzene, and 1,3,5-Trifluoro-2,4,6-trichlorobenzene. Figure 7 presents the 19FNMR spectrum of the 24-h urine of a 1,3,5trifluoro-2,4,6-trichlorobenzeneex~ed rat. The spectrum shows four main metabolite peaks, which were not present in a 2:l ratio. From this observation it follows that no 1,3,5-trifluoro-containingmetabolites are formed. Thus these main metabolites must result from preferential elimination of a fluorine, not a chlorine, substituent upon biotransformation. Fluorine elimination will result in metabolites.with two equivalent fluorine substituents, that give rise to one 19FNMR peak, as observed (Figure 7). Treatment of the urine with @-glucuronidaseresulted in disappearance of the resonance at -113.2 ppm and was accompanied by an increase of the resonance at -113.5 ppm. On the basis of all this, the resonance at -113.2 ppm was ascribed to 3,5-difluoro-2,4,6-trichlorophenyl glucuronide. Treatment of the urine with arylsulfatase/@glucuronidase resulted only in the disappearance of the 3,5-difluoro-2,4,6-trichlorophenol glucuronide resonance while the resonance at -113.5 ppm remained present. Thus the resonance a t -113.5 ppm most likely represents both 3,5-difluoro-2,4,6-trichlorophenol and -phenyl sulfate. Furthermore, incubation of 1,3,5-trifluoro-2,4,6-trichlorobenzene with cytosol and glutathione did not result in formation of a product with its resonance at -113.5 ppm,
16 Chem. Res. Toxicol., Vol. 5, No. 1, 1992 Table 11. Urinary Recovery Profiles of Rats Exposed to Hexahalogenated or Pentahalogenated Benzenesa recovery, % of the total amount of fluorine administered 24-48 h compound 0-24 h 48-72 h hexafluorobenzene 2.8 f 0.9 0.2 f 0.1 0 (0.8f 0.2) (0) (0) pentafluorochlorobenzene 10.0 f 2.8 2.2 f 0.7 0.2 f 0.1 (1.8f 0.5) (0.3f 0.2) (0.1f 0.1) 1,3,5-trifluoro-2,4,6-trichlor19.0 f 0.8 12.4 f 0.3 6.9 f 0.1 obenzene (9.4f 0.6) (5.8f 0.3) (2.9f 0.1) pentafluorobenzene 81.5 f 15.3 10.9 f 4.3 0.4 f 0.1 2,3,5,6-tetrafluoro-l-chloro-70.9 f 3.2 17.7 f 7.1 0.8 f 0.2 benzene
Rietjens and Vervoort
:@:
a
F
‘The dose administered was 300 pmollkg body weight. The values presented are the mean f SEM for 3-5 rats. Values in parentheses represent urinary recoveries calculated on the basis of only the I9F NMR resonances ascribed to monodefluorinated hydroxylated metabolites (phenol, phenyl glucuronide, and phenyl sulfate derivatives).
indicating that this resonance at -113.5 ppm in the urine spectra does not represent a 3,5-difluoro-2,4,6-trichloro1-glutathione derivative. Table I1 summarizes the in vivo urine recovery profile for hexafluorobenzene, pentafluorochlorobenzene, and 1,3,5-trifluoro-2,4,6-trichlorobenzene. Assuming a similar rate of uptake from the intestines and transport to the liver, it can be derived from these results that the rate of in vivo conversion of these hexahalogenated benzenes into urine-excretable metabolites and also into defluorinated hydroxylated urine metabolites increases with the number of chlorine substituents. On the basis of the high log Poctanol coefficients for all three compounds, calculated to be 3.3, 3.8, and 4.9 for hexafluorobenzene, pentafluorochlorobenzene,and 1,3,5trifluoro-2,4,6-trichlorobenzene, respectively, it can be concluded that uptake characteristics for all three compounds will be regulated by passive diffusion processes. From these log Podanolvalues it also follows that differential liver uptake of the three hexahalogenated benzenes, due to differential fat accumulation, cannot be the reason underlying the differences in in vivo metabolism. When fat accumulation would be the factor causing the differential conversion of the three hexahalogenated benzenes to urine metabolites, recoveries would decrease in the opposite order than observed, 1,3,5-trifluoro-2,4,6-trichlorobenzene being the most hydrophobic compound. Furthermore, urinary recovery data for the pentahalogenated pentafluorobenzene and 2,3,5,6-tetrafluoro1-chlorobenzene, presented in Figure 8 and Table 11, demonstrate that the differences observed for the hexahalogenated benzenes are also not a result of a difference in clearance of the volatile benzenes in expired air. This follows from the facts that (i) pentafluorobenzene and 2,3,5,6-tetrafluoro-l-chlorobenzene, having almost similar boiling points as hexafluorobenzene and pentafluorochlorobenzene, respectively, show 0-72-h recoveries of over 90%, suggesting that only very limited clearance in the expired air occurs, and that (ii) 0-24-h urinary recovery of the pentafluorobenzene is not lower than that of the less volatile 2,3,5,6-tetrafluoro-l-chlorobenzene. Molecular Orbital Computer Calculations. From the results obtained so far, mechanisms 1, 2, and 4 could be excluded as possible routes for the formation of hydroxylated metabolites from the hexahalogenated benzenes. Thus, mechanism 3 and mechanism 5 provided the best explanation for the cytochrome P-450 dependent oxidative dehalogenation of the hexahalogenated benzenes.
Ir-IT-
CI
, -120
Figure 8.
, -140
-160
-180
PPM
19F NMR spectra of the urine from a rat exposed to (a) pentafluorobenzene or (b) 2,3,5,6-tetrafluoro-l-chlorobenzene. Resonances were identitied as described in the text. The resonance marked “is” is from the internal standard 4-fluorobenzoic acid.
Molecular orbital calculations were carried out to see whether theoretical data would be in accordance with (i) the observation of increased urinary recovery with increasing number of chlorine substituents, with (ii)the high reactivity of the possible intermediate, and with (iii) the decreased ratio fluoride anion:organic urine metabolites excreted with increasing number of chlorine substituents. When radical attack is known to be the rate-limiting factor for a reaction, the reaction rate might become dependent on the energy level of either the HOMO (highest occupied molecular orbital) [E(HOMO)] or the LUMO (lowest unoccupied molecular orbital) [E(LUMO)] of the species attacked (18). This possible influence of E(HOM0) or E(LUM0) on the reaction rate originates from the fact that molecular orbital overlap, and thus reactivity, will increase with a smaller energy difference between the SOMO (single occupied molecular orbital) of the radical and either the HOMO or the LUMO of the other reactant (18).
Table III presents results from the calculation of HOMO and LUMO energies for the three hexahalogenated benzenes studied. These data provide some reason for increased rate of attack of the P-450 iron-oxene in the sequence 1,3,5-trifluoro-2,4,6-trichlorobenzene> pentafluorochlorobenzene > hexafluorobenzene, as a less negative HOMO energy might result in a more efficient interaction between the benzene HOMO and the iron-oxene SOMO (18). The differences in E(HOM0) are, however, small and might therefore not explain the recovery differences actually observed in vivo (Table II). Furthermore,
Chem. Res. Toricol., Vol. 5, No. 1, 1992 17
Conversion of Hexahalogenated Benzenes by Cyt P-450
hexahalogenated benzene hexafluorobenzene pentafluorochlorobenzene 1,3,5-trifluoro-2,4,6trichlorobenzene
Table 111. Results from Molecular Orbital Computer Calculations" E(LUM0). E(LUM0) relative AHF for hexahalo:' relative AHF for for formation of the E(HOMO), genated formation of the benzohalohexahalogenated hexahalogenated benzene, benzhaloquinone quinone epoxide, kcal/mol benzene, eV eV cation, kcal/mol cation, eV -10.37 -1.48 0 -9.02 0 +15.6 -10.24 -1.42 -0.9 -8.88
-10.10
-1.23
-4.6
-8.60
+22.0
E(LUMO), epoxide, eV -1.98 -1.86 -1.72
" Differences in heat of formation (AHF) are presented relative to the value for the conversion of hexafluorobenzene to its pentafluorobenzoquinone cation (or epoxide). AHF values for the reactions were calculated as the difference between the heat of formation for the benzohaloquinone cation (or the hexahalogenated epoxide) minus the heat of formation for the hexahalogenated benzene. it seems more reasonable that for mechanism 3 the halogen elimination step resulting in fluoride anion elimination and formation of the benzohaloquinone cation will be the rate-limiting step. Molecular orbital computer calculations provide a possibility to calculate the relative change in heat of formation for this reaction for the three different hexahalogenated benzenes. Taking the reaction for hexafluorobenzene as the reference, conversion of pentafluorochlorobenzene to the 2,3,5,6-tetrafluoro-4-chlorobenzoquinone is 0.9 kcal/mol easier, and conversion of 1,3,5-trifluoro-2,4,6-trichlorobenzene to 3,5-difluoro-2,4,6trichlorobenzoquinone is 4.6 kcal/mol easier. This demonstrates that additional chlorine substituents seem to stabilize the relatively unstable benzohaloquinone cation. A reduced instability of the benzohaloquinone cation intermediate going from hexafluorobenzene > pentafluorochlorobenzene > 1,3,5-trifluoro-2,4,6-trichlorobenzene would explain the increased urine 0-72-h recoveries, as well as the decreased relative amount of fluoride anion elimination observed going from hexafluorobenzene > pentafluorochlorobenzene > 1,3,5-trifluoro-2,4,6-trichlorobenzene. It should be stressed that the energy values presented in Table I11 are not absolute energy changes, as neither the energy of the iron-oxene in the P-450 active site taking part in the reaction nor that of the fluoride anion eliminated are included in the calculations. However, as these changes are the same when the reactions of hexafluorobenzene, pentafluorochlorobenzene, or 1,3,5-trifluoro2,4,6-trichlorobenzenein mechanism 3 are compared to one another, this will only affect the absolute values obtained, not the relative differences between the three reactions, which are the differences presented in Table 111. Table I11 also presents E(LUM0) values for the benzohaloquinone cations. From the relatively low values obtained it can be derived that these intermediates will be readily reduced or-in other words-that benzohaloquinones are strong oxidizing agents. Besides, extremely low E(LUM0) values, as observed for the benzohaloquinone cations, could make them easy substrates for nucleophilic attack by electrons in HOMO orbitals of, for instance, cellular macromolecules. Similar values presented in Table I11 for the hexahalogenated epoxide intermediates formed in mechanism 5 demonstrate that E(LUM0) values for the hexahalogenated epoxide and the relative AHF for formation of the hexahalogenated epoxide intermediates do not explain the in vivo biotransformation characteristicsobserved as well. The relative AHF for formation of the hexahalogenated epoxide even shows a trend opposite to the one that would explain the urinary recovery results.
Discussion The present study investigates the mechanism for cy-
tochrome P-450 dependent conversion of fully halogenated benzenes. In theory, on the basis of literature data (3-9), five mechanisms could be provided for the cytochrome P-450 dependent dehalogenation coupled to hydroxylation at the dehalogenated position (Figure 2). Mechanism 2, which proceeds by oxygenation of a halogen substituent, followed by HXO elimination and phenol formation as a result of H 2 0 addition (Figure 2), was already eliminated in the literature (5,7). On the basis of microsomal conversion experiments with 180-labeled H20 and hexachlorobenzene in which no incorporation of the l80label in the pentachlorophenol was observed, mechanism 2 was excluded (5, 7). However, migration of the C10 moiety, resulting in binding of the oxygen instead of the chlorine atom to the aromatic ring, followed by H20 attack on the chlorine, would also explain that no l80 incorporation into the phenol was observed. The present article, however, provides another line of evidence which eliminates mechanism 2. In vivo biotransformation of pentafluorochlorobenzeneclearly resulted in preferential elimination of the fluorine at the position para with respect to the chlorine substituent and not of the chlorine substituent from pentafluorochlorobenzene. This eliminates mechanism 2 because oxygenation of a halogen will be less likely for a fluorine than for a chlorine substituent. This assumption can be made on the basis of data published by Guengerich (19),demonstrating that bromine substituents already showed a much lower rate for such a reaction than iodine substituents. These considerations imply that mechanism 2 would preferentially eliminate the chlorine not the fluorine from pentafluorochlorobenzene,which is not observed. The preferential elimination of fluorine substituent and not the chlorine substituent from pentafluorochlorobenzene also excludes mechanism 1. This conclusion follows from the fact that a mechanism proceeding by halogen cation elimination would favor elimination of the less electronegative chlorine substituent over elimination of the highly electronegative fluorine substituent, which is not what is actually observed. Results from in vitro microsomal incubations further demonstrated that reductive defluorination was not observed as a result of cytochrome P-450 dependent conversion of pentafluorochlorobenzene (mechanism 4). Under aerobic conditions fluoride anion formation was readily observed (Figure 61, but no formation of 2,3,5,6-tetrafluorochlorobenzene was seen. In addition, under anaerobic conditions, which should favor reductive dehalogenation, fluoride anion formation was less than 3% of the amount observed in aerobic incubations. These data exclude mechanism 4. In addition, the results from Figure 6 demonstrated that the aerobic microsomal fluoride anion formation was also not accompanied by formation of an equal amount of other
18 Chem. Res. Toxicol., Vol. 5, No. 1, 1992
Rietjens and Vervoort 0 II
c
OH
1
I CI
binding to cellular macromolecules
+
F
Figure 9. Hypothesis for the mechanism of the cytochrome P-450 dependent dehalogenation of hexahalogenated benzenes. (FeO)3+ represents the iron-oxene species present in the active site of the cytochrome P-450 catalyzing the reaction. dehalogenated metabolites. This observation implies that the metabolite formed upon the cytochrome P-450 monooxygenative defluorination must be highly reactive, binding to microsomal proteins, which makes it NMR invisible. Following mechanism 5,defluorination would be a result of epoxide reduction and would be accompanied by formation of a similar amount of hydroxylated metabolites. This is, however, not observed upon in vitro microsomal conversion of pentafluorochlorobenzene (Figure 6). An alternative explanation for the fluoride anion formation that would still permit mechanism 5 would be a process in which the hexahalogenated epoxide intermediate of mechanism 5 loses a fluoride anion upon binding to cellular protein. Mechanism 3 could also explain the results from the in vitro incubations (Figure 6), demonstrating fluoride anion formation not accompanied by 19FNMR visible product formation. Formation of the benzohaloquinone cation would be accompanied by proportional fluoride anion formation, and binding of the reactive benzohaloquinone cation to cellular macromolecules would result in formation of additional fluoride anions. Such a process would be comparable to the well-known protein binding of benzoquinone (imines) (3, 5). Additional data from molecular orbital computer calculations and results from the in vivo recoveries of three hexahalogenated benzenes provided results that seem to favor mechanism 3 over mechanism 5. In vivo urinary recoveries appear to increase in the sequence hexafluorobenzene C pentafluorochlorobenzene C 1,3,5-trifluoro2,4,6-trichlorobenzene. Furthermore, a decreased ratio of urine fluoride anion:other metabolites was observed in the sequence hexafluorobenzene > pentafluorochlorobenzene > 1,3,5-trifluoro-2,4,6-trichiorobenzene, being 6.7 > 2.0 > 1.3, respectively. Molecular orbital computer calculations demonstrated formation of the benzohaloquinone cation from pentafluorochlorobenzene and 1,3,5-trifluoro-2,4,6trichlorobenzene to be respectively 0.9 and 4.6 kcal/mol thermodynamically favored over formation of the benzohaloquinone cation from hexafluorobenzene. A more stabilized benzohaloquinone cation with increasing number of chlorine substituents would explain both the order of the urinary recoveries as well as the changes in the ratio fluoride anion:metabolite excretion observed. Besides, the relatively low E(LUM0) of the benzohaloquinone cations would explain their high reactivity in, for instance, a reaction involving a nucleophilic attack of HOMO electrons from a cellular macromolecule. Similar theoretical data calculated for the hexahalogenated epoxide intermediates did not explain the in vivo and in vitro biotransformation characteristics as well. However, definite exclusion of mechanism 5 might have to await synthesis of the epoxide intermediates and their incubation with NADPH. Furthermore, one could also suggest elimination of a fluoride anion from the hexahalogenated epoxide before
its chemical reduction to the phenol (mechanism 5). This would result in formation of the benzohaloquinone cation intermediate of mechanism 3. Such a reaction pathway would combine mechanism 3 and 5 to a certain extent but would still imply the reactive benzohaloquinone cation intermediate and its reduction to give the pentahalogenated phenol (=mechanism 3). On the basis of all results presented, we conclude that a mechanism with a benzohaloquinone cation intermediate provides the best hypothesis for the cytochrome P-450 dependent conversion of hexahalogenated benzenes to pentahalogenated phenols. Figure 9 presents this hypothesis for the mechanism of the cytochrome P-450 dependent dehalogenation of pentafluorochlorobenzene, taking into account the iron-oxene intermediate supposed to be the enzyme form actually catalyzing the substrate conversion (20, 21). The mechanism might also be of importance for the cytochrome P-450 dependent aerobic dehalogenation of other halogenated benzenes. Finally, it should be stressed that the results of the present study clearly demonstrate that, upon the first cytochrome P-450 dependent conversion of a hexahalogenated benzene, reactive intermediates are formed that might lose a halogen anion upon reaction with cellular macromolecules. This implies that secondary metabolism converting the phenol formed in the first reaction step to a reactive quinone, or to a hydroquinone that might be reoxidized to give the reactive quinone, is not a prerequisite for bioactivation of hexahalogenated benzenes, as suggested before in the literature (5).
Acknowledgment. Thanks are extended to G. van Tintelen and J. Haas for help with animal handling and values. to S. Boeren for calculation of the log Poctanol Registry No. F,16984-48-8; pentafluorophenol, 771-61-9; 2,3,5,6-tetrafluorophenol, 769-39-1; pentatluorophenylglucuronide,
137570-48-0;pentafluorophenyl sulfate, 137570-49-1; 2,3,5,6tetrafluorophenyl glucuronide, 137570-50-4;2,3,5,6-tetrafluorophenyl sulfate, 137570-51-5; 2,3,5,6-tetrafluorohydroquinone, 527-21-9;hexafluorobenzene, 392-56-3; pentafluorochlorobenzene, 344-07-0; 2,3,5,6-tetrafluoro-4-chlorophenylglucuronide, 137570-52-6; 2,3,5,6-tetrafluoro-4-chlorophenyl sulfate, 13757053-7; 2,3,5,6-tetrafluoro-4-chlorophenol,4232-66-0; S-(2,3,5,6tetrafluoro-4-chlorophenyl)glutathione,137570-54-8; 1,3,5-trifluoro-2,4,6-trichlorobenzene, 319-88-0; 3,5-difluoro-2,4,6-trichlorophenyl glucuronide, 137570-55-9;3,5-difluoro-2,4,6-trichlorophenol, 2992-91-8; 3,5difluoro-2,4,6-trichlorophenyl sulfate, 137570-56-0;cytochrome P-450, 9035-51-2.
References MacDonald, T. L. (1983) Chemical mechanisms of halocarbon metabolism. CRC Crit. Rev. Tonicol. 11, 85120. (2) Wiersma, D. A., Schnellmann, R. G., and Sipes, I. G. (1984) Pathways of halogenated hydrocarbon metabolism. In Foreign compound metabolism (Caldwell, J., and Paulson, G. D., Eds.) pp
(1)
Chem. Res. Toxicol. 1992,5, 19-25 53-64, Taylor & Francis, New York. (3) Rietjens, I. M. C. M., Tyrakowska, B., Veeger, C., and Vervoort, J. (1990) Reaction pathways for biodehalogenation of fluorinated anilines. Eur. J. Biochem. 194, 945-954. (4) Rietjens, I. M. C. M., and Vervoort, J. (1991) Bioactivation of 4-fluorinated anilines to benzoquinoneiminea as primary reaction products. Chem.-Biol. Interact. 22, 263-281. (5) van Ommen, B., and van Bladeren, P. J. (1989) Possible reactive
intermediates in the oxidative biotransformation of hexachlorobenzene. Drug Metab. Drug Interact. 7 , 213-243. (6) Stewart, F. P., and Smith, A. G. (1986) Metabolism of the “mixed”cytochrome P-450 inducer hexachlorobenzene by rat liver microsomes. Biochem. Pharmacol. 35, 2163-2170. (7) van Ommen, B., Adang, A. E. P., Brader, L., Posthumus, M. A,, MUer, F., and van Bladeren, P. J. (1986) The microsomal metabolism of hexachlorobenzene. Origin of the covalent binding to protein. Biochem. Pharmacol. 35,3233-3238. (8) Jerina, D. M., and Daly, J. W. (1974) Arene oxides: A new aspect of drug metabolism. Science 185,573-582. (9) Daly, J. W., Jerina, D. M., and Witkop, B. (1972) Arene oxides and the NIH shift: The metabolism, toxicity and carcinogenicity of aromatic compounds. Erperientia 28, 1129-1164. (10) Rietjens, I. M. C. M., and Vervoort, J. (1989) Microsomal metabolism of fluoroanilines. Xenobiotica 19,1297-1305. (11) Hudlicky, M. (1972) Reactions of organic fluorine compounds. In Chemistry of Organic Fluorine Compounds (Hudlicky, M., Ed.) Chapter 5, pp 170-518, Ellis Horwood Ltd. Halsted Press, John Wiley & Sons, New York. (12) van de Lande, L. M. F. (1932) L’action du mgthylate de sodium
19
sur quelques d6rivgs du m6tadichlorobenzene. Rec. Trav. Chim. PUYS-BUS 51,9&113. (13) Vervoort, J., De Jager, P. A., Steenbergen, J., and Rietjens, I. M. C. M. (1990) Development of a l9F n.m.r. method for studies on the in vivo and in vitro metabolism of 2-fluoroaniline. Xenobiotica 20, 657-670. (14) Shaka, A. J., Keeler, J., and Freeman, R. (1983) Evaluation of a new broad band decoupling sequence: Waltz 16. J. Magn. Reson. 55,313-340. (15) Omura, T., and Sato, R. (1964) The carbon monoxide-binding pigment of liver microsomes. 1. Evidence for its hemoprotein nature. J. Biol. Chem. 239,2370-2318. (16) Lowry, 0. H., Rosebrough, N. L., Farr, A. L., and Randall, R. J. (1951) Protein measurement with the folin phenol reagent. J. Biol. Chem. 193,265-275. (17) Rekker, R. F., and de Kort, H. M. (1979) The hydrophobic fragmental constant; an extension to a 1000 data point set. Eur. J. Med. Chem. Chim. Ther. 14,479-488. (18) Fleming, I. (1976) in Frontier Orbitakr and Organic Chemical Reactions, John Wiley & Sons, New York. (19) Guengerich, F. P. (1989) Oxidation of halogenated compounds by cytochrome P-450, peroxidases and model metalloporphyrins. J. Biol. Chem. 264,17198-17205. (20) Ortiz de Montellano, P. R. (1986) Oxygen activation and transfer. In Cytochrome P-450. Structure, Mechanism and Biochemistry (Ortiz de Montellano, P. R., Ed.) pp 217-271, Plenum Press, New York. (21) Guengerich, F. P., and MacDonald, T. L. (1990) Mechanisms of cytochrome P-450 catalysis. FASEB J. 4, 2453-2459.
Separation of (+)-sun- and (-)-anti-Benzo[ a Ipyrene Dihydrodiol Epoxide-DNA Adducts in 32P-PostlabelingAnalysis Ashok P. Reddy, Donna Pruess-Schwartz, and Lawrence J. Marnett* Department of Chemistry, Wayne State University, Detroit, Michigan 48202 Received June 13, 1991 The (+)-enantiomer of 7,8-dihydroxy-7,&dihydrobenzo[a]pyrene(BP-7,8-diol) is a diagnostic probe for cytochrome P-450 and non-cytochrome P-450 pathways of dihydrodiol epoxidation. The principle products of epoxidation are the (+I-syn-dihydrodiol epoxide [(+)-syn-BPDE] and the (-)-anti-dihydrodiol epoxide [ (-)-anti-BPDE] . Chromatographic conditions are described that separate the major deoxynucleoside 3’,5’-bisphosphate adducts derived from these dihydrodiol epoxides on commercial poly(ethy1enimine)thin-layer plates. Inclusion of boric acid and magnesium chloride in the D4 solvent is a key feature of the separation. Reasonable separation of these bisphosphate adducts from the major deoxynucleoside 3’,5’-bisphosphate adduct derived from (+)-anti-BPDE is also observed. 3?P-Postlabeling analysis of DNA adducts produced following topical administration of benzo[a] pyrene to mouse skin suggests that cytochrome P-450 plays a major role in its metabolism to DNA binding derivatives.
Introduction Benzo[a]pyrene (BP)’ is a ubiquitous environmental pollutant and a potent animal carcinogen (1,2).Extensive experimental evidence implicates dihydrodiol epoxides (BPDE) as major contributors to the genotoxic activity of BP (3). BPDEs are formed from BP by sequential oxidation, hydration, and oxidation as indicated in Figure 1 (4-10). Four possible BPDEs exist that differ significantly in DNA binding capacity, mutagenicity, and tumor-initiating activity. The extent to which individual BPDE’s form in a particular cell is determined by the balance of *To whom correspondence should be addressed at The A. B. Hancock, Jr., Memorial Laboratory for Cancer Research, Department of Biochemistry, Center in Molecular Toxicology, Vanderbilt University School of Medicine, Nashville, TN 37232-0146.
enzymes that catalyze each transformation (7,8,11-13). Epoxidation of BP to BP-7,8-oxide is catalyzed by cytochromes P-450and hydration of BP-7,8-oxide to BP-7,8diol by epoxide hydrolase (5,9). Epoxidation of BP-7,8diol is catalyzed by cytochromes P-450 and by peroxidases, or it is effected by peroxyl radicals generated by lipid peroxidation, metal-catalyzed hydroperoxide decomposiAbbreviations: BP, benzo[a]pyrene;BHA,butylated hydroxyanieole;
8-NF, 8-naphthoflavone; (+)-BP-7,8-diol,7(S),B(S)-dihydroxy-7,8-dihydrobenzo[alpyrene; BPDE, 7,8-dihydroxy-9,10-epoxy-7,8,9,lO-tetrahydrobenzo[a]pyrene; (+)-anti-BPDE, 7(R),8(S)-dihydroxy-9(S),lO(R)epoxy-7,8,9,10-tetrahydrobenzo[a]pyrene; (-)-anti-BPDE, 7(S),8(R)-dihydroxy-9(R),10(S)-epoxy-7,8,9,1O-tetrahydrobenzo[a]pyrene; (+)-synBPDE, 7(S),8(R)-dihydroxy-9(S),10(R)-epoxy-7,8,9,10-tetrahydrobenzo[alpyrene; (-)-syn-BPDE, 7(R),8(S)-dihydroxy-9(R),lO(S)-epoxy7,8,9,lO-tetrahydrobenzo[aIpyrene;PEI, poly(ethy1enimine);TBA, tet-
rabutylammonium chloride; TLC, thin-layer chromatography; RAL, relative adduct labeling.
oa93-22a~/92/2705-ooi9$03.oo/o 0 1992 American Chemical Society
