674
Chem. Res. Toxicol. 1993,6, 674-680
Cytochrome P45O-Mediated Oxidation of Pentafluorophenol to Tetrafluorobenzoquinone as the Primary Reaction Product Cathaline den Besten,? Peter J. van Bladeren,??$Erwin Duizer,tfs Jacques Vervoort,§ and Ivonne M. C. M. Rietjens*>s Department of Toxicology, Agricultural University, Wageningen, The Netherlands, Department of Biological Toxicology, TNO-Toxicology and Nutrition Institute, Zeist, The Netherlands, and Department of Biochemistry, Agricultural University, Wageningen, The Netherlands Received March 26, 1993
In the present study the oxidative dehalogenation of a para-halogenated phenol was studied using pentafluorophenol and its non-para-halogenated analogue 2,3,5,64etrafluorophenolas model compounds. 19F NMR was used to characterize the metabolite patterns. In order to study the primary oxidation products of the microsomal cytochrome P450-catalyzed conversion, the alternative oxygen donors cumene hydroperoxide (CumOOH) and iodosobenzene (IOB) were used in addition to the use of NADPH and molecular oxygen. In a NADPH/oxygen-driven reaction, but also in a CumOOH- or IOB-driven cytochrome P450 reaction, tetrafluorophenol was converted to tetrafluorohydroquinone. However, for pentafluorophenol, the formation of tetrafluorohydroquinone as a product of its cytochrome P450-mediated conversion was only observed in the NADPH-driven system. Addition of reducing equivalents such as NADH to the CumOOH or IOB incubations resulted in the formation of tetrafluorohydroquinone. From these data it was concluded that the primary reaction product of the cytochrome P45O-catalyzed conversion of pentafluorophenol is a reactive species that can be reduced to tetrafluorohydroquinone by NAD(P)H and, thus, must be tetrafluorobenzoquinone. Additional experiments with tetrafluorobenzoquinone, incubated in vitro with either microsomal protein or glutathione in the presence or absence of reducing equivalents, demonstrated that the tetrafluorobenzoquinone ends up bound to proteins, losing its fluorine atoms as fluoride anions. Thus, while cytochrome P450-mediated conversion of the 2,3,5,6-tetrafluorophenol results in the formation of tetrafluorohydroquinone as the primary reaction product, monooxygenation a t a fluorinated para position, such as in pentafluorophenol, results in the formation of the reactive tetrafluorobenzoquinone derivative as the primary reaction product. This direct formation of reactive benzoquinones upon cytochrome P450-catalyzed conversion of halogenated compounds may very well have toxicological implications.
Introduction In the toxicity of halogenated benzenes, bioactivation to chemically reactive species is thought to be involved (1, 2). Indeed, during microsomal metabolism of chlorinated benzenes a considerable amount of covalent binding to protein is observed. Secondary benzoquinonemetabolites formed upon further oxidation of the primary phenols have recently been implicated to be the reactive species involved in the metabolism of hexachlorobenzene (31, pentachlorobenzene (4), 1,2,4-trichlorohenzene(51, 1,2and 1,Cdichlorobenzene (6), and bromobenzene (7). In the past it has been suggestedthat oxidation of arenes at a halogenated position, as in the case of pentachlorophenol, proceeds by a mechanism analogous to arene oxidation at a nonhalogenated position, resulting in a hydroxylated product (€49).Tetrachlorohydroquinonehas indeed been detected as a metabolite of pentachlorophenol, both in vitro (9)and in vivo (10). Taking the P450 reaction * Address correspondence to this author at the Department of Biochemistry, Agricultural University, Dreijenlaan 3,6703HA Wageningen, The Netherlands; phone: 31-8370-82868;fax: 31-8370-84801. t Department of Toxicology, Agricultural University. Department of BiologicalToxicology,TNO-Toxicology and Nutrition Institute. 1 Department of Biochemistry, Agricultural University. * Abstract published in Advance ACS Abstracts, August 15,1993.
as a net two-electron transfer, the halogen is formally lost as a halogen cation (depicted in route a, Figure 1). However, the actual mechanism of (oxidative)dehalogenation resulting in a hydroxylated product is still a matter of debate. Direct formation of tetrahalobenzoquinone upon oxidation of a pentahalogenated phenol, accompanied by loss of a halogen anion (route c, Figure 11, has been suggested as one of the alternative pathways (1113). Theoretically, formation of the semiquinone anion radical accompanied by the formal loss of a halogen radical is also feasible upon monooxygenation of a pentahalophenol (route b, Figure 1). The current lack of experimental evidence favoring any of these mechanisms is due to several complicating factors that are present when performingmechanistic studies with these substrates. On the one hand, direct formation of benzoquinones may be masked by their swift covalent binding to protein due to their high reactivity (14). On the other hand, the presence of reducing equivalents such as NADPH makes it impossible to probe the cytochrome P450-dependent primary reaction product, since any benzoquinone formed may subsequently be reduced to its hydroquinone. In recent studies using '9F NMR (12,151 evidence was obtained for the formation of benzoquinone imines and
0893-228x/93/2706-0674$04.00/00 1993 American Chemical Society
Chem. Res. Toxicol., Vol. 6, No. 5, 1993 675
P450 Oxidation of Halophenols OH
X X
OH
OH
I
0
OH
+ X'
0
+ x
+ X'
Figure 1. Possible pathways for the cytochrome P450-mediated oxidation of a pentahalogenated phenol, taking into account the electron balance for the P450 reaction (X = halogen substituent). The pentahalogenated phenol might be oxidized to tetrahalohydroquinone via elimination of a halogen cation (route a); to a tetrahalosemiquinone anion radical through elimination of a halide radical (route b); or to tetrahalobenzoquinone via elimination of a halide anion (route c).
fluoride anions as the primary reaction products of the cytochrome P450-dependent monooxygenation of parafluorinated anilines. The present study was undertaken to test whether the cytochrome P450-dependent conversion of para-halogenated phenols proceeds by a similar reaction pathway. Pentafluorophenol and 2,3,5,6-tetrafluorophenol were used as model substrates to investigate the oxidation at a halogenated and a nonhalogenated position para with respect to the hydroxyl moiety.
Experimental Procedures Chemicals. Pentafluorophenol (PFP),' 2,3,5,6-tetrafluorophenol (TFP), and 2,3,5,6-tetrafluorohydroquinone(TFHQ) were purchased from Aldrich Chemie (Steinheim, FRG). 2,3,5,6Tetrafluorobenzoquinone (TFBQ) was from Fluorochem (Derbyshire, U.K.). NAD(P)H was from Boehringer (Mannheim, FRG). Cumene hydroperoxide (CumOOH)was from Aldrich (Steinheim, FRG). Iodosobenzene was synthesized from iodosobenzyldiacetate (Fluka, Switzerland) as described by Saltzman and Sharefkin (16). Preparation of Microsomes. Microsomes from untreated rata were prepared as previously described (3)and stored at -80 OC until used. Protein was determined according to Lowry using bovine serum albumin as a standard (17). Cytochrome P450 was determined as described by Omura and Sat0 (18). Microsomal Incubations. Microsomal incubations (final volume 2 mL) were carried out as follows. The fluorinated phenols (3 mM final concentration) in acetone (2.5% finalconcentration) were added to an incubation mixture containing 0.1 mM potassium phosphate, pH 7.4, and microsomal protein (2 nmol of P450/mL) in glass culture tubes with a Teflonized rubber cap. After 2 min of preincubation at 37 OC, the reaction was started by the addition of NADPH (1mM fiial concentration). Reactions were terminated after 5 min by the addition of HCl(O.6 N final 1 Abbreviations: PFP, pentafluorophenol; TFP, 2,3,5,6-tetrafiuoTFBQ, 2,3,5,6-tetrophenol; TFHQ, 2,3,5,6-tetrafluorohydroquinone; ratluorobenzoquinone;CumOOH, cumene hydroperoxide;IOB, iodoeobenzene; P450, cytochrome P450.
concentration). Ascorbic acid (1 mM final concentration) was also added to prevent oxidation of the hydroquinone formed. The reaction mixtures were frozen in liquid nitrogen and stored at -20 "C. After centrifugation upon thawing, the supernatant was made anaerobic by four cycles of evacuation and filling with argon and directly analyzed by l9F NMR. In some incubations, cumene hydroperoxide [CumOOH, 250 pM final concentration, added as a 10 mM suspension in water containing 10% (v/v) acetone] or iodosobenzene (IOB, 1 mM final concentration, added as a 10-min-sonified suspension of 20 mM in 10% dimethyl sulfoxide in water) was used as artificial oxygen donor for the cytochrome P450 reaction, and NADH (1 mM final concentration), as a reducing agent. For these incubations the preincubation mixture as well as the CumOOH or IOB solution and the NADH solution was made anaerobic by four cycles of evacuation and filling with argon. The reactions were started by the addition of CumOOH, IOB, and/or NADH and terminated after 5 (CumOOH) or 1 (IOB) min as described above. Chemical Incubations with TFBQ a n d TFHQ. To investigate the chemical reactivity of TFBQ and TFHQ under various incubation conditions, 20 p L of freshly dissolved TFBQ or TFHQ (50 mM in dimethyl sulfoxide, 0.5 mM final concentration) was added to 19F NMR tubes containing 2 mL of potassium phosphate (0.1 M, pH 7.4, final concentration) containing 100 pL of 2Hz0 and (final concentrations) 20 mM NADH, 2 nmol/mL microsomal P450,250 pM CumOOH, 1mM IOB, and/or 5 mM GSH as indicated. Upon addition of TFBQ or TFHQ the reaction mixtures were immediately analyzed by lgF NMR. Furthermore, in order to characterize metabolites covalently bound to protein, microsomes (2 nmol of P450/mL, final volume 10 mL) were incubated with TFBQ (1mM final concentration, 5 min, 37 "C). At the end of these incubations the protein was precipitated by the addition of 0.6 N HCl (final concentration) to the incubation mixture. The protein pellet was extensively washed with organic solvents of decreasing polarity as previously described (4). The final pellet of the TFBQ incubation was hydrolyzed in 3 mL of 6 N HC1 (24 h, 98 "C) and analyzed by lgF NMR, to detect possible amino acid-bound fluorinated material. l9F NMR Measurements. For l9F NMR measurements samples were routinely made oxygen-free by four cycles of evacuation and fiiing with argon gas. l9FNMR measurements were performed as previously described (12,15). Briefly, a Bruker CXP 300 or AMX 300 spectrometer operating at 282.3 MHz with a 10-mm l9F Bruker probe head and Norell (Landisville, NJ) 10-mmNMR tubes were used. The sample volume was 1.71mL, containing 100 pL of 2H20, and the temperature was kept at 7 "C. Proton decoupling was achieved with the Waltz-16 pulse sequence at -16 dB from 20 W. Nuclear Overhauser effects were eliminated using the inverse gated decoupling technique. Spectra were obtained with 30° pulses (6 ps), a 50-kHz spectral width, repetition time of 1s, quadrature phase detection, and quadrature phase cycling (CYCLOPS). Between 500 and 50 OOO scans were recorded depending on the concentration of the compounds and the signal to noise ratio required. To improve the signal to noise ratio, the free induction decays were multiplied by an exponential decay resulting in an increase of the line width by 5 Hz. The concentrations of the substrates and products were determined from the integrals of their l9FNMR resonances. Concentrations were calculated by comparison of these integrals to the integral from the l9F NMR resonance of p-fluorobenzoic acid added to each sample as an internal standard. Chemical shifta are reported relative to CFCls. Identification of 19F NMR Resonances. The "3F NMR chemical shifts for the standard compounds relative to CFCls acquired under the standard experimental conditions [incubations performed in 0.1 M potassium phosphate, pH 7.4, stopped by acidifying the reaction by the addition of 0.6 N HC1 (final concentration), after which 1 mM ascorbic acid was added; measured at 7 "C]were as follows: fluoride anion, -164.1 ppm; pentafluorophenol, -167.2 ppm (F2, F6), -169.5 ppm (F3, F5),
676
Chem. Res. Toxicol., Vol. 6,No. 5, 1993
-174.7 ppm (F4);2,3,5,6-tetrafluorophenol, -146.0 ppm (F3,F5), -166.9 ppm (F2,F6);2,3,5,6-tetrafluorohydroquinone, -168.4ppm (F2,F3, F5, F6). Resonances could be ascribed to specificfluorine substituents on the basis of (i) relative signal intensities, (ii)
splittingpatterns in proton-coupled and -decoupledspectra,and (iii)the knowledge that meta substituentsonly slightly influence 19FNMR chemical shifts, whereas ortho and para substituents have a much larger effect. IBF NMR chemical shifts of the compounds in 0.1 M potassium phosphate, pH 7.4 (for the chemical reactivityexperiments),could also be identified on the basis of added reference compounds.
1
I
,
IS
$)I
I
den Besten et al.
H
Results and Discussion NADPH-Dependent Microsomal Oxidation of Fluorophenols. Incubation of TFP or its para-fluorinated analog, PFP, with rat liver microsomes in the absence of an electron donor did not result in formation of metabolites (Figures 2a and 3a). In the presence of NADPH and molecular oxygen, formation of TFHQ is observed for both TFP and PFP (singlet at -168.4 ppm, Figures 2b and 3b). In addition to TFHQ, formation of fluoride anion was observed especially in incubations with PFP (singlet a t -164.1 ppm, Figure 3b). Quantitative results derived from the spectra of Figures 2 and 3 are presented in Table I. From these data it can be derived that TFP and PFP are metabolized to a comparable extent in a NADPH-driven P450-dependent conversion. However,the ratio of fluoride anion to TFHQ formed was 0.4 and 4.1 in incubations with TFP and PFP, respectively, and thus higher than the values expected on the basis of a stoichiometric conversion of TFP to TFHQ and of PFP to TFHQ and F(respectively 0 and 1.0). This phenomenon might be ascribed to autoxidation of TFHQ to TFBQ and fluoride anion production due to the high reactivity of TFBQ (see below). CumOOH-Dependent Microsomal Conversion of TFP and PFP. To investigate the nature of the primary reaction product upon cytochrome P450-catalyzed conversion of TFP and PFP, standard microsomal incubation conditions are unsuitable, since any benzoquinone formed can be reduced by NADPH to its hydroquinone form (see below). To circumvent this problem, incubations were performed under anaerobic conditions using CumOOH as an artificial oxygen donor to create the reactive (FeO)3+ species. The CumOOH-dependent P450 monooxygenase activity has been suggested to proceed either via heterolytic oxygen-oxygen bond cleavage, resulting in (FeO)3+formation, or via homolytic oxygen-oxygen bond cleavage, resulting in (FeO-H)3+and a peroxide radical that initiates the reaction by hydrogen abstraction (19-23). However, for hydroxylation of aromatic carbon atoms, the peroxidedriven cytochrome P450 reaction may very well proceed to a significant extent by the heterolytic cleavage, leading to the (FeO)3+intermediate, which is also formed in the NADPH/oxygen-driven reaction (22, 23). Figures 2c and 3c present the results of microsomal incubations with TFP and PFP with CumOOH as the artificial oxygen donor. In incubations of TFP under anaerobic conditions with CumOOH the formation of TFHQ was readily observed (Figure 2c). Since in the absence of reducing equivalents TFHQ cannot be formed from TFBQ (see below), TFHQ must be the primary product resulting from the cytochrome P450 conversion of TFP. However, when PFP was used as a substrate, the formation of TFHQ was not observed in the CumOOHdriven reaction mixture (Figure 3c). Nevertheless, Cum-
7I I
-120
I
.140
I
I
- 1 60
I
I
- 1 80
PPM Figure 2. 19FNMR spectra from microsomal incubationswith 2,3,5,6-tetrafluorophenol.Different electron or oxygen donors
were used. (a) Blank under aerobic conditions but without electron donor; (b) NADPH under aerobic conditions; (c) CumOOH under anaerobic conditions; (d) CumOOH + NADH under anaerobic conditions. Microsomal incubationswith substrate and NADH gave similar results as blank incubations (19F NMR spectrum not shown). Controlincubationswith substrate and either CumOOH or NADH did not result in any product formation(l9FNMR spectranot shown). The resonancemarked "IS" is from the internal standard p-fluorobenzoic acid. OOH-dependent microsomal conversion of PFP resulted in the formation of a significant amount of fluoride anion (Figure 3c, Table I), indicating that PFP is converted into a defluorinated product other than TFHQ. Addition of reducing equivalents in the form of NADH, to incubations containing PFP as a substrate and CumOOH as the artificial oxygen donor for the cytochrome P450 reaction, resulted again in the formation of TFHQ (Figure 3d, Table I). In addition, a concomitant decline in the amount of fluoride anion production was observed upon NADH addition to the PFP/CumOOH incubation (Figure 3d, Table I). Apparently, conversion of PFP results in the formation of a metabolite that can be reduced to TFHQ
Chem. Res. Toxicol., Vol. 6, No. 5, 1993 677
P450 Oxidation of Halophenols
Table I. Conversion of 2,3,5,6-TFP and PFP to TFHQ and Fluoride Anion by Rat Liver Microsomes in NAD(P)H and/or Artificial Oxygen Donor-Driven Reactions, As Determined by 19F NMR. nmol (total incubation time)-* (nmol of P450)-1* reaction total ratio substrate supported by F- TFHQ conversionc ETFHQ TFP C0.5 CO.1 0 NADH C0.5 CO.1 0 NADPH 3.1 7.2 8.0 0.4 CumOOH 107 3.8 31 28 CumOOH+ 4.8 16 17 0.3 NADH IOB 74 1.1 20 67 3.1 2.6 3.4 IOB + NADH 1.2 PFP C0.5 CO.1 0 NADH C0.5 CO.1 0 NADPH 14 3.4 6.2 4.1 CumOOH 211 CO.1 42 >2100 CumOOH+ 31 14 20 2.2 NADH IOB 78 CO.1 16 >780 IOB+ NADH 4.4 2.4 3.3 1.8 a Iodosobenzene (1OB)- and cumene hydroperoxide (Cum0OH)driven reactions were carried out under anaerobic conditions. Incubations were carried out as described in Experimental Procedures. * Total incubation time was 5 min for all incubations except for the IOB incubations, for which the total incubation time was 1 min. c Calculated as the amount of fluoride anions divided by respectively 4 (TFP) or 5 (PFP), plus the amount of TFHQ formed. This calculation assumes that TFBQ loses all four fluoro atoms as fluoride anions upon ita reaction with microsomal macromolecules.
I - 120
I
.140
I
r
- 1 60
quinone derivative. Another part may end up bound to macromolecules. The absence of a l9FNMR resonance representing TFBQ in the 19F NMR spectra can be explained by the fact that this compound is reactive and avidly binds to microsomal macromolecules, in a Michaeltype addition-elimination reaction, losing its fluoro substituents as fluoride anions. Additional results supporting this chemical reactivity of the TFBQ are reported below. In addition, previous studies on the microsomal conversion of the chlorinated analog [14Clpentachlorophenolclearly demonstrated the formation of protein-bound adducts
--- 1 80
PPM Figure 3. 1BF NMR spectra from microsomal incubations with pentafluorophenol. Different electron or oxygen donors were used. (a) Blank under aerobic conditions but without electron donor; (b) NADPHunder aerobic conditions; (c)CumOOH under anaerobic conditions; (d) CumOOH + NADH under anaerobic conditions. Microsomal incubations with substrate and NADH gave similar results as blank incubations (l9F NMR spectrum not shown). Control incubations with substrate and either CumOOH or NADH did not result in any product formation (l9F NMR spectra not shown). The resonance marked "IS" is from the internal standard p-fluorobenzoic acid.
by NAD(P)H, suggesting it to be TFBQ. As NADH and NADPH are well-known two-electron reductants, the observation eliminates formation of the tetrafluorosemiquinone as a primary metabolite that can be reduced to TFHQ by NAD(P)H as this requires a single-electron reductant. Together, these results imply that the cytochrome P450-dependent conversion of PFP results in formation of the reactive TFBQ as the primary reaction product. On the basis of the electron balance of the cytochrome P450 reaction this implies that the fluorine is lost from the molecule as a fluoride anion. In the presence of reducing equivalents [e.g., NAD(P)Hl the benzoquinone metabolite is (partly) reduced to its hydro-
(3, 9).
Comparison of the 19FNMR spectra in Figures 2 and 3 also demonstrates that replacing NADPH and molecular oxygen as cofactors for the cytochrome P450-mediated monooxygenation of fluorophenols by CumOOH results in a more efficient oxidation of both TFP and PFP (Table I). This phenomenon has also been described for the conversion of fluorinated anilines in a peroxide-driven P450-catalyzed conversion (12). The reason underlying this effect may be that in the NADPH-driven reaction the donation of the first or second electron to, respectively, the iron-substrate-porphyrin or the iron-substrateoxygen-porphyrin complex is rate-limiting. These reaction steps are circumvented in the CumOOH-driven reaction. Another factor then becomes rate-limiting in the CumOOH-driven reaction. Furthermore, in incubations with TFP the ratio of fluoride anion to TFHQ is increased to 28 in the CumOOHdriven reaction (Figure 212, Table I) as compared to 0.4 in the NADPH-driven reaction (Figure 2b, Table I). In the CumOOH-driven reaction in the presence of NADH, however, the ratio is low again, Le., 0.3 (Figure 2d, Table I). In the incubations with PFP (Figure 3) the ratio of fluoride anion to TFHQ formed is always significantly higher than observed for the analogous TFP incubation.
678 Chem. Res. Toxicol., Vol. 6, No. 5, 1993
This can be explained by the fact that in the PFP incubations the reactive TFBQ, giving rise to fluoride anion production in subsequent reactions, is the primary reaction product, whereas in incubations with TFP, formation of this fluoride anion producing TFBQ is dependent on secondary oxidation of the primary metabolite TFHQ. Iodosobenzene (1OB)-Dependent Microsomal Conversion of Fluorophenols. Iodosobenzene (IOB) was also used as an artificial oxygen donor for the cytochrome P450 reaction. IOB-supported cytochrome P450 reactions are generally accepted to proceed by the same cytochrome P450(Fe0)3+ intermediate as formed in the NADPH/ oxygen-driven reaction (21,22). The results obtained for the conversion of TFP and PFP in IOB-driven anaerobic microsomal incubations in the absence or presence of NADH are also presented in Table I. In the presence of NADH the total conversion was much lower than observed in the IOB-driven reaction in the absence of reducing equivalents (Table I). This phenomenoncould be ascribed to reduction and, thus, inactivation of the iodosobenzene by NADH, demonstrated by a decrease of the NADH absorbance at 340 nm upon addition of iodosobenzene (data not shown). In addition, in the absence of NADH, the IOB-driven microsomal conversion of T F P resulted in a ratio of fluoride anion to TFHQ formed that was higher than the ratio observed in the CumOOH system, suggesting an even more efficient chemical oxidation of the TFHQ formed by IOB than by CumOOH. Nevertheless, the results of the IOB experiments showed the same tendencies as the CumOOH-driven reactions; Le., (i) significant formation of TFHQ was observed in incubations with TFP as a substrate, but not with PFP as a substrate, and (ii) addition of NADH to the IOB-driven PFP system resulted in significant formation of TFHQ accompanied by a reduction in the relative amount of fluoride anions (Table I). These results further support that the primary metabolites formed in the PFP and TFP systems are different and that the primary reaction product formed in the PFP system must be a compound that can be chemically reduced to TFHQ by NADH, demonstrating it to be TFBQ. Chemical Reactivity of T F B Q and TFHQ under Varying Reaction Conditions, Additional experiments were performed to demonstrate that NADH is capable of (chemically) reducing TFBQ into TFHQ, and to demonstrate possible fluoride anion production from TFBQ under various conditions (Figure 4). When TFBQ is added to a potassium phosphate buffer, the compound appeared to be highly unstable, resulting in formation of fluoride anions and unidentified reaction products represented by various new, unidentified '9F NMR resonances (Figure 4a). Since addition of ascorbic acid (1 mM final concentration) to this incubation did not result in the formation of TFHQ, none of the peaks observed in the spectrum in Figure 4a represents TFBQ. Addition of TFBQ to a potassium phosphate buffer containing a 40-fold excess of NADH (compared to TFBQ) results in its reduction to TFHQ (Figure 4b). In an incubation of TFBQ in the presence of microsomes the main I9F NMR signal observed is that of fluoride anion (Figure 4c). In the presence of microsomes and a 40-fold excess of NADH, TFBQ is again converted to TFHQ, but still significant formation of fluoride anion is observed (Figure 4d),most likelyresulting from covalent binding of TFBQ to microsomal protein in a Michael-type addition-elimination reaction with an
den Besten et al.
I F
1 IF
C
IF
d
IF
e
I
I
.120
.160
.140
.loo
PPM
Figure 4. 19FNMR spectraobtained upon the addition of freshly dissolved TFBQ (tetrafluorobenzoquinone) to anaerobic potassium phosphate solutions containing various additional components. (a) No extra addition; (b)40-fold excess NADH; (c) 2 HM microsomal P450; (d) 40-fold excess NADH plus 2 fiM microsomal P450; and (e) 10-fold excess reduced glutathione. Resonances were identified on the basis of added reference compounds.
electrophilic protein residue, analogous to reactions of tetrachlorobenzoquinone (3). To investigate this possible binding of TFBQ to microsomal protein in more detail, acid hydrolysis of the thoroughly washed precipitated protein material from an incubation of TFBQ with microsomes was performed to detect possible amino acidbound TFBQ derivatives. Without protein hydrolysis, the substantial increase in molecular weight of proteinbound TFBQ caused by its binding to the protein gives rise to an increase in its rotational correlation time (decrease in T2) and consequently such an intensive line broadening of ita l9FNMR resonance that the signal cannot be observed in the 19F NMR spectrum. Hydrolysis of the protein into the amino acid residues, however, decreases the molecular weight of the fluorine-containing adduct, making its 19F NMR signals visible again. However, IgF NMR analysis of the sample derived from a 10-mL incubation containing 1 mM TFBQ and 2 nmol/mL microsomal P450, concentrated into 3 mL of acid hydrolysate, demonstrated no fluorine resonances in spite of the 1 pM sensitivity limit of the measurement, Le., 0.03 74 of the TFBQ added. This result indicates the elimination of all four fluoro substituents from TFBQ upon covalent binding to nucleophilic moieties of cellular macromole-
Chem. Res. Toxicol., Vol. 6,No.5, 1993 679
P450 Oxidation of Halophenols
radical o-complex
cationic o-comolex H
0
+ xFe
Fe3+
0
Figure 5. Proposed reaction pathway for the cytochrome P450-catalyzed conversion of phenols at a halogenated para position and a nonhalogenated para position (X = halogen). The intermediates presented between brackets are hypothetical and are based on a reactionpathway proceeding by the a-additionmechanism. Formation of the cationic a-adduct from the radical a-adduct results from electron abstraction from the aromatic ring by the Fe of the cytochrome P450.
cules. This was further confirmed by results from an experiment in which TFBQ was added to a potassium phosphate buffer, pH 7.4, containing a 10-fold excess of GSH (Figure 4e). The complete elimination of all fluorine atoms from TFBQ as fluoride anions was observed: no other 19FNMR resonance was visible in these incubations. Additional incubations were performed with TFHQ, to explain the fluoride anion formation in microsomal incubations where TFHQ instead of TFBQ was formed as the primary reaction product. Incubation of TFHQ in phosphate buffer under aerobic conditions, or under anaerobic conditions in the presence of CumOOH or IOB, resulted in significant breakdown accompanied by formation of fluoride anions (‘9F NMR spectra not shown). Thus, this fluoride anion formation most likely results from the oxidation of TFHQ to the reactive TFBQ, which loses fluoride anions upon subsequent reactions.
Concluding Remarks Based on results of the present study, the best hypothesis for the reaction pathway describing the cytochrome P450dependent oxidation of halophenols at a para-halogenated and a non-para-halogenated position is the scheme put forward in Figure 5. Cytochrome P450-dependent conversion at a para-unsubstituted position proceeds via the formation of the hydroquinone as the primary product, which may subsequently be oxidized to its quinone derivative. However, the cytochrome P450-mediated conversion of a para-halogenated phenol is proposed to proceed by the direct formation of a benzoquinone and the loss of the halogen from the molecule as a halogen anion. These halogenated benzoquinones are highly reactive compounds and may covalently interact with cellular macromolecules. However, in the presence of NAD(P)H or other reducing equivalents part of the benzoquinone may be reduced to its (nonreactive) hydroquinone form. In this study CumOOH and IOB were used as the oxygen donor for the cytochrome P450-mediated oxidation of fluorinated phenols. For the CumOOH-driven reaction, independent of the mechanism for cleavage of the peroxide bond, the present data provide evidence for a reaction
resulting in formation of TFBQ as the primary reaction product for the conversion of PFP. However, evidence exists (22,23) that a peroxide-supported cytochrome P450catalyzed aromatic hydroxylation proceeds a t least in part by heterolytic cleavage of the peroxide bond and, thus, a P450(Fe0)3+ intermediate similar to the one formed in the NADPH/oxygen-driven reaction. An IOB-driven cytochrome P450 reaction is generally accepted to proceed by the same cytochrome P450(Fe0I3+ intermediate as formed in the NADPH/oxygen-supported reaction (21, 22). Therefore, the formation of TFBQ as the primary reaction product in the P450-catalyzed oxidative dehalogenation of PFP not only holds for the CumOOH- and IOB-driven reactions, but will also be relevant for the NADPH/oxygen-driven reaction. Such a cytochrome P450-catalyzed mechanism for oxidative dehalogenation of para-fluorinated phenols would be similar to the mechanism demonstrated by Husain et al. (24)for the p-hydroxybenzoate hydroxylase-catalyzed oxidative defluorination of fluorinated 4-hydroxybenzoates. Recent studies on the P450-catalyzedconversionof parafluorinated anilines also indicated the direct formation of benzoquinone imines with the concomitant loss of fluoride anions (12,151.It becomes tempting to suggest that the proposed pathway may represent a general mechanism for the cytochrome P450-dependent oxidation of substrates, which meet the following requirements: (i) the presence of an acidic proton in one of the substituents in the aromatic ring; and (ii) substituents with a “goodleaving group” character (e.g., halogens) in the ortho or para position. In fact, evidence has been gathered that the proposed mechanism may even be extended to compounds with electron-donating substituents in general, such as fully halogenated aromatics, e.g., hexahalobenzenes, resulting in the formation of a benzohaloquinone cation intermediate which may subsequently be reduced to its phenol derivative (25). The proposed pathway for oxidative dehalogenation may also be a relevant pathway in the P450-dependent oxidation of bromobenzene to protein alkylating species. In a recent in vivo study it was shown that the majority of the protein S-adducts in rat liver after bromobenzene
680 Chem. Res. Toxicol., Vol. 6, No. 5, 1993
exposure consisted of debrominated benzoquinone adducts (13,261.This phenomenon can readily be explained by the mechanism proposed in the present study, in which the cytochrome P450-dependent oxidation of p-bromophenol, a primary metabolite of bromobenzene, results in the formation of a debrominated benzoquinone with the loss of the bromine from the molecule as a bromide anion. The reactive benzoquinone may subsequently bind to cellular macromolecules. Quinone metabolites may play a role in the toxicity of several xenobiotics, through arylation of cellular macromolecules and/or through their capacity to form a redox shuttle with their reduced counterparts, thereby generating reactive oxygen species and imposing conditions of cellular oxidative stress. The direct formation of benzoquinones may have toxicological implications, especially because in vivo it can be expected that hydroquinones will not always be oxidized to their reactive quinone analogs, due to (i) competing pathways of conjugation and (ii) the presence of reducing equivalents such as NAD(P)H and ascorbic acid. However, for aromatic compounds from which benzoquinones are formed as a direct result of the cytochrome P450-mediated conversion, formation of the reactive products cannot be circumvented.
den Besten et al. bromobenzene and derivatives. Evidence for quinones as reactive metabolites. Xenobiotica 18, 501-510. (8)Ahlborg, U. A., Larsaon, K., and Thumber, T. (1978) Metabolism of pentachlorophenol in uiuo and in vitro. Arch. Toxicol. 40,45-53. (9) Van Ommen, B., Adang, A. E. P., MUer, F., and Van Bladeren, P. J. (1986)The microsomal metabolism of pentachlorophenol and ita covalent binding to protein and DNA. Chem.-Biol. Interact. 60, 1-11.
(10)Renner, G., and Hopfer, C. (1990)Metabolic studies on pentachlorophenol (PCP) in rata. Xenobiotica 10, 573-582. (11) Van Ommen, B., and Van Bladeren, P. J. (1989) Possible reactive intermediates in the oxidative biotransformation of hexachlorobenzene. Drug Metab. Drug Interoct. 7,213-243. (12) Rietjens, I. M. C. M., and Vervoort, J. (1991) Bioactivation of 4-fluorinated anilines to benzoquinoneimines as primary reaction products. Chem.-Biol. Interact. 77,263-281. (13) Zheng, J., and Hanzlik, R. P. (1992) Dihydroxylated mercapturic acid metabolites of bromobenzene. Chem. Res. Toxicol. 5,561-567. (14) Van Ommen, B., Voncken, J. W., MOller, F., and Van Bladeren, P. J. (1988) The oxidation of tetrachloro-1,4-hydroquinoneby microsomes and purified cytochrome P450b. Implications for covalent binding to protein and involvement of reactive oxygen species. &em.-Biol. Interact. 65, 247-259. (15) Rietjens, I. M. C. M., Tyrakoweka, B., Veeger, C.,and Vervoort, J. (1990) Reaction pathways for biodehalogenation of fluorinated anilines. Eur. J.Biochem. 194,945-964. (16) Saltzman,H., andShiuefkin, J. G. (1973)Iodmbenzene. In Organic Syntheses, Collect. Vol. V, pp 658-659, John Wiley & Sons, New York. (17) Lowry, 0. H., Roeabrough, N. J. Fan, A. L.,and Randall, R. J. (1951)Protein measurement with the Folin phenol reagent. J.Biol. Acknowledgment. Part of this study was financially Chem. 193,265-275. supported by Grant 427.531 from the Dutch Organisation (18) Omura, T., and Sato, R. (1964)The carbon monoxidebinding pigment for the Advancement of Pure Research (NWO). The of liver microsomes. 11. Solubilisation, purification and properties. authors acknowledge ir. N. H. P. Cnubben for providing J.Biol. Chem. 239, 2379-2385. the iodosobenzene. (19) Blake, R. C., and Coon, M. J. (1981) On the mechanism of action of cytochrome P450. Evaluation of homolytic and heterolytic mechanisms of oxygen-oxygen bond cleavage during substrate References hydroxylation by peroxides. J. Biol. Chem. 266, 12127-12133. (20) Coon, M. J., Ding, X., Perecky, S. J., and Vaz, A. D. N. (1992) (1) Brodie, B. B., Reid, W. D., Cho, A. K., Sipes, G., and Gillete, J. R. Cytochrome P450: progress and predictions. FASEB J.6,669-673. (1971) Poseible mechanism of liver necrosis caused by aromatic (21) Ortiz de Montellano, P. R. (1986) Oxygen activation and transfer. organic compounds. Roc. Natl. Acad. Sci. U.S.A. 68,160-164. In Cytochrome P-450. Structure, Mechanism and Biochemistry (2) Narasimhan,N.,Weller,P.E.,Buben,J.A.,Wiley,R.A.,andHanzlik, (Ortiz de Montellano, P. R., Ed.) pp 217-271, Plenum Press, New R. P. (1988) Microsomal metabolism and covalent binding of [sH/ York. W]-bromobenzene. Evidence for quinones aa reactive metabolites. Xenobiotica 18, 491-499. (22) McMurry, T. J., and Groves, J. T. (1986) Metalloporphyrin models (3) Van Ommen, B., Adang, A. E. P., Brader, L.,Posthumus, M. A., for Cytochrome P-450. In CytochromeP-450. Structure,Mechunism Miiller, F., and Van Bladeren, P. J. (1986) The microsomal and Biochemistry (Ortiz de Montellano, P. R., Ed.) pp 1-28, Plenum metabolism of hexachlorobenzene. Origin of the covalent binding Press, New York. to protein. Biochem. Pharmacol. 359, 3228-3238. (23) Yumibe, N. S., and Thompson, J. A. (1988) Fate of free radicals (4) Den Beaten, C., Peters, M. M. C. G., and Van Bladeren, P. J. (1989) generated during one-electron reductions of 4-alkyl-1,4-peroxyThe metabolism of pentachlorobenzene by rat liver microsomes: quinola by cytochrome P-450. Chem. Res. Toxicol. 1, 385-390. The nature of the reactive intermediates formed. Biochem.Biodws. _ (24) Husain, M., Entach, B., Ballou, D. P., Massey, V., and Chapman, P. Res. Commun. 163, 1275-1281. J. (1980) Fluoride elimination from substrates in hydroxylation (5) Den Besten. C., Smink, M. C. C., de Vries. E. J., and Van Bladeren, reactions catalyzed by p-hydroxybenzoate hydroxylase. J. Biol. P. J. (1991) Metabolic activation of 1,2,4-trichlorobenzene and Chem. 255, 4189-4197. pentachlorobenzeneby rat liver microsomes: a major role for quinone (25) Rietjens, I. M. C. M., and Vervoort, J. (1992) A new hypothesis for metabolites. Toxicol. Appl. Pharmacol. 108, 223-233. the mechanism for cytochrome P-450 dependent aerobic conversion (6) Den Besten, C., Ellenbroek, M., Van der Ree, M. A. E., Rietjene, I. of hexahalogenated benzenes to pentahalogenated phenols. Chem. M. C. M., and Van Bladeren, P. J. (1992)The involvementof primary Res. Toxicol. 5, 10-19. and secondary metabolism in the covalent binding of 1,2- and 1,4(26) Slaughter,D. E.,and Hanzlik, R. P. (1991)Identification of epoxidedichlorobenzenes. Chem.-Biol. Interact. 84,259-275. and quinone-derived bromobenzene adducta to protein sulfur (7) Buben, J. A,, Narasimhan, N., and Hanzlik, R. P. (1988) Effecta of nucleophiles. Chem. Res. Toxicol. 4, 349-359. chemical and enzymic probes on microsomal covalent binding of
