Metabolism of 1, 2-Dichloro-1-fluoroethane and 1-Fluoro-1, 2, 2

1-Fluoro-1,2,2-trichloroethane: Electronic Factors. Govern the Regioselectivity of Cytochrome. P450-Dependent Oxidation. Hequn Yin, M. W. Anders, and ...
0 downloads 0 Views 405KB Size
50

Chem. Res. Toxicol. 1996, 9, 50-57

Metabolism of 1,2-Dichloro-1-fluoroethane and 1-Fluoro-1,2,2-trichloroethane: Electronic Factors Govern the Regioselectivity of Cytochrome P450-Dependent Oxidation Hequn Yin, M. W. Anders, and Jeffrey P. Jones* Department of Pharmacology, Box 711, University of Rochester, School of Medicine and Dentistry, Rochester, New York 14642 Received May 18, 1995X

1,2-Dichloro-1-fluoroethane (HCFC-141) and 1,1,2-trichloro-2-fluoroethane (HCFC-131) were chosen as models to study the regioselectivity of halogenated alkane metabolism. Metabolites in the urine of rats given HCFC-131 ip were the following: inorganic fluoride, chlorofluoroacetic acid, dichloroacetic acid, N-(2-hydroxyethyl)chlorofluoroacetamide, and three unidentified minor metabolites. In vitro incubation of HCFC-131 with either rat liver microsomes from pyridinetreated rats or expressed human cytochrome P450 2E1 isozyme in the presence of NADPH gave fluoride, chlorofluoroacetic acid, and dichloroacetic acid as metabolites. HCFC-141 was biotransformed in rats to inorganic fluoride, chlorofluoroacetic acid, 2-chloro-2-fluoroethanol, and 2-chloro-2-fluoroethyl glucuronide, which were detected in urine. Incubation of HCFC141 with NADPH-fortified liver microsomes from pyridine-induced rats or expressed human cytochrome P450 2E1 afforded fluoride, chlorofluoroacetaldehyde hydrate, and chloroacetic acid as products. The metabolites identified were consistent with a cytochrome P450-dependent oxidation mechanism. The data also indicated that phosphatidylethanolamine may be a cellular target for chlorofluoroacetyl chloride, a reactive intermediate generated from HCFC-131 by cytochrome P450-dependent oxidation. Chlorofluoroacetic acid given to rats ip was largely recovered in the rat urine, although the formation of inorganic fluoride as a metabolite was observed. The mechanism of defluorination of chlorofluoroacetic acid is not clear. Regioselective oxidation by cytochrome P450 was observed between the two potential oxidizable sites in HCFC141 and in HCFC-131. Comparison of the observed ratio of oxidation at different sites in in vitro experiments with the calculated activation energies for hydrogen-atom abstraction from these sites indicated that electronic factors are the primary determinant of regioselectivity. In vivo regioselectivity could not be compared with theory since this ratio does not reflect the true regioselectivity due to differences in excretion, reabsorption, secondary metabolism (e.g., fluoride generation from chlorofluoroacetic acid), other routes of fluoride formation, and limitation of the method of detection.

Introduction (HCFCs)1

Hydrochlorofluorocarbons and hydrofluorocarbons (HFCs) are partially halogenated alkanes that are being considered as replacements for chlorofluorocarbons (CFCs) (1). CFCs are perhalogenated alkanes that enjoy wide use as refrigerants, cleaning agents, blowing agents, and propellants (1), but are being phased out because of their ozone-depleting potential (2). The presence of C-H bonds in hydro(chloro)fluorocarbons (H(C)FCs) makes them labile to degradation in the troposphere and therefore reduces their ozone-depleting potentials in the stratosphere (1). These C-H bonds also render H(C)FCs labile to oxidative metabolism, which may be associated with the generation of toxic metabolites. The large-scale production and use of H(C)FCs may be accompanied by the potential for human exposure. Hence, a clear understanding of their metabolism and toxicity is essential. The metabolism of some H(C)FCs has been studied (38). Cytochromes P450 were identified as the sole enAbstract published in Advance ACS Abstracts, November 1, 1995. Abbreviations: HCFCs, hydrochlorofluorocarbons; HFCs, hydrofluorocarbons; H(C)FCs, hydro(chloro)fluorocarbons; CFCs, chlorofluorocarbons; HCFC-141, 1,2-dichloro-1-fluoroethane; HCFC-131, 1,1,2trichloro-2-fluoroethane. X 1

0893-228x/96/2709-0050$12.00/0

zymes that catalyze the initial oxidation of H(C)FCs (4, 6). Most of the H(C)FCs studied thus far, such as 2-dichloro-1,1,1-trifluoroethane (HCFC-133a) (9) and 2,2-dichloro-1,1,1-trifluoroethane (HCFC-123) (10), possess only one oxidizable site, i.e., identical C-H bonds in each molecule. In the present study we chose two compounds, each bearing two oxidizable sites, namely, 1,2-dichloro-1-fluoroethane (HCFC-141) and 1,1,2-trichloro2-fluoroethane (HCFC-131) (see Figure 1 for structures), as model compounds to address the following questions: first, what are the metabolic fates of these H(C)FCs both in vivo and in vitro; second, is there regioselectivity between the two potential oxidizable sites in each molecule and what are the factors that govern the regioselectivity. The hypothesis being tested is that regioselectivity in cytochrome P450-dependent biotransformations of H(C)FCs is primarily governed by electronic factors.

Experimental Procedures Instrumental Analyses. 19F NMR spectra were acquired with a Bruker WP-270 spectrometer operating at 254.18 MHz with D2O as the lock signal. The acquisition parameters were as follows: excitation pulse width, 7 µs; relaxation delay, 0.5-

© 1996 American Chemical Society

Regioselective Metabolism of HCFC-141 and HCFC-131

Figure 1. Structures of HCFC-141 and HCFC-131. 1.0 s for metabolite detection and 1.7 s for quantitative measurements; sweep width (SW), 5000-20 000 Hz; and number of scans, 1000-30 000. Chemical shifts were referenced to a 5.0 mM solution of trifluoroacetamide in D2O (δtrifluoroacetamide ) 0 ppm) that was contained in a capillary coaxial tube. Spectra were recorded at room temperature in a 5-mm tube with the sample spinning at ∼25 rpm. The ratio of various fluorinecontaining metabolites was quantified by integration of resonances; the variations in longitudinal relaxation time (T1) of the metabolites were corrected by choosing relaxation times that were long enough to allow full relaxation of all fluorine nuclei. Mass spectra were recorded with a Hewlett-Packard 5880A GC equipped with a HP-1 capillary column (dimethyl silicone gum, 25 m × 0.2 mm × 0.5 µm film thickness) and coupled to a HP-5970 mass selective detector (70 eV, electron impact). Conditions for GC analyses were as follows: splitless injection; injector temperature, 250 °C; initial column temperature, 30 °C for 0.5 min; program rate, 10 °C/min to 200 °C; final column temperature of 200 °C for 2.5 min; interface temperature, 250 °C; carrier gas, He. Fluoride ion concentrations were measured with a fluorideselective electrode (Orion Model 90-01, Orion Research Inc., Boston, MA) connected to a potentiometer (Corning Model 125 pH meter, Corning, NY). Samples were prepared by mixing 1.5 mL of sample and 1.5 mL of total-ionic-strength adjusting buffer (1 M acetic acid, 1 M sodium chloride, and 0.012 M 1,2cyclohexanediaminetetraacetic acid) in deionized water (6-8). The electrode was calibrated prior to use with solutions of known fluoride concentrations. Zero potential was established with a solution containing a 1.5-mL sample of control rat urine or a 1.5-mL sample of an incubation mixture that lacked NADPH and 1.5 mL of total-ionic-strength adjusting buffer. Chemicals. HCFC-141 and HCFC-131 were purchased from PCR Inc. (Gainsville, FL) and were purified by distillation. Only fractions that were pure by 19F and 1H NMR were used. β-Glucuronidase (EC 3.2.1.31, type B3 from bovine liver) and acylase I (EC 3.5.1.14, from porcine kidney) were purchased from Sigma Chemical Co. (St. Louis, MO). Hep G2 cells were obtained from American Type Culture Collection (Rockville, MD). TK- cells and cytochrome P450 2E1 recombinant vaccinia virus were gifts from Dr. Kenneth Korzekwa, National Cancer Institute (Bethesda, MD). All other chemicals were of the highest grade commercially available. Synthesis of Chlorofluoroacetyl Chloride and Fluoride. The synthetic intermediate 2-chloro-1,1,2-trifluoroethyl ethyl ether was synthesized by the reaction of 2-chloro-1,1,2-trifluoroethene with sodium ethoxide in ethanol according to the method of Englund (11). (Caution: Chlorofluoroacetyl halides are reactive and may be harmful if inhaled or absorbed by skin. All procedures should be conducted in an efficient fume hood.) A mixture of chlorofluoroacetyl chloride and fluoride was synthesized according to Nguyen (12) from 2-chloro-1,1,2trifluoroethyl ethyl ether. The final product was a mixture of chlorofluoroacetyl chloride and fluoride. Analytical data: 1H NMR (CDCl3, δTMS ) 0 ppm): 6.37 ppm (d, J ) 50 Hz); 19F NMR (CDCl3, δtrifluoroacetamide ) 0 ppm): -70.18 ppm (d, J ) 50 Hz). This mixture was used in the subsequent synthesis without further purification. Synthesis of N-(2-Hydroxyethyl)chlorofluoroacetamide. Ethanolamine (0.3 mL, 4.97mmol) and 2 mL of dry ether along with a magnetic stirring bar were placed in a 20-mL glass vial. The vial was then sealed with a Teflon-lined septum crimp-top cap. The septum was punctured with a needle and tubing connected to a CaSO4 drying tube. A mixture of chlorofluoroacetyl chloride and fluoride (0.2 mL) was then added dropwise over a period of 5 min with a syringe. The mixture was stirred

Chem. Res. Toxicol., Vol. 9, No. 1, 1996 51 for 5 min at room temperature. The reaction mixture was added to 4% NaHCO3 solution and extracted with ether. The ether fraction was dried over sodium sulfate and evaporated to dryness. GC/MS analysis showed N-(2-hydroxyethyl)chlorofluoroacetamide was formed in a purity of >95%. 1H NMR (acetonitrile-d6, δTMS ) 0 ppm): 7.17 (s, NH), 6.42 (d, CH, 2JHF ) 50.6 Hz), 3.57 (t, CH2O, 3J ) 5.5 Hz), 3.33 (t, CH2N, 3J ) 5.7 Hz), 3.03 (s, OH); 19F NMR (D2O, δtrifluoroacetamide ) 0 ppm): -70.26 ppm (2JHF ) 50.6 Hz); mass spectrometry (EI): m/z 154 (M - 1), 137 (M - H2O), 124 (M - CH2OH), 88 (M - CHFCl), 67 (CHFCl), 60 (M - COCHFCl). Synthesis of Chlorofluoroacetic Acid. Chlororofluoroacetic acid was prepared by hydrolysis of chlorofluoroacetyl chloride and fluoride (13) and was purified by fractional distillation. 1H NMR (DMSO-d6, δTMS ) 0 ppm): 11.6 ppm (s), 6.37 (d, J ) 50 Hz); 19F NMR (D2O, δtrifluoroacetamide ) 0 ppm): -60.58 ppm (J ) 50.0 Hz). Synthesis of Chlorofluoroacetaldehyde Hydrate. Chlorofluoroacetaldehyde hydrate was synthesized by the reduction of methyl chlorofluoroacetate with diisobutylaluminium hydride. Methyl chlorofluoroacetate was prepared by methylation of chlorofluoroacetic acid with diazomethane. A magnetic stirring bar, 10 mL of ethyl ether dried over sodium metal, and 2.0 g (15.8 mmol) of methyl chlorofluoroacetate were contained in a three-neck, round-bottom flask fitted with a nitrogen inlet, an addition funnel, and a condensing tube connected to a drying tube. The mixture was cooled to -78 °C in a dry ice-ethanol bath while maintaining a gentle flow of dry nitrogen. Diisobutylaluminium hydride (17.5 mL of 1 M solution in THF, 17.5 mmol) was added dropwise through the addition funnel over a period of 60 min. The temperature was maintained at -78 °C. The mixture was stirred for 15 min and then poured into 50 mL of ice-water containing 4% (v/v) sulfuric acid. After stirring for 60 min at room temperature, the slurry was extracted with ether (6 × 15 mL). The ether was evaporated, and the residue was distilled in vacuo. A viscous colorless liquid was collected (1.4 g), bp 55.0-57.5 °C/100 mmHg; yield 77%. 1H NMR (acetonitrile-d6, δTMS ) 0 ppm): 5.97 (dd, CHFCl, 2JHF ) 47.8 Hz, 3JHH ) 3.9 Hz), 4.72 (dd, CH(OH)2, 3JHF ) 6.8 Hz, 3JHH ) 3.9 Hz); 19F NMR (D2O, δtrifluoroacetamide ) 0 ppm): -72.70 (dd, 2J 3 HF ) 47.0 Hz, JHH ) 6.6 Hz). Synthesis of 2-Chloro-2-fluoroethanol. Methyl chlorofluoroacetate (7 g, 0.0553 mol) and 100 mL of dry ether were contained in a three-neck, round-bottom flask equipped with a stirrer, an addition funnel, a condensing tube connected to a CaSO4 drying tube, and a nitrogen inlet. Lithium aluminium hydride (600 mL of 0.104 M solution in ether, 0.0624 mol) was added through the addition funnel. The mixture was heated at reflux for 90 min. A few drops of water were added to the mixture, 50 mL of 2 N H2SO4 was added, and the mixture was stirred overnight. The mixture was extracted with ether, and the crude product was distilled. 2-Chloro-2-fluoroethanol was collected as a colorless liquid, bp 75-80 °C/165 mm Hg; yield 65%. 1H NMR (acetonitrile-d6, δTMS ) 0 ppm): 6.37 (dt, CHFCl, 2J 3 3 HF ) 48.1 Hz, JHH ) 4.4 Hz), 3.64 (dd, CH2, JHF ) 19.4 Hz, 3J 19 HH ) 3.9 Hz); F NMR (D2O, δtrifluoroacetamide ) 0 ppm): -72.70 (dd, 2JHF ) 47.0 Hz, 3JHH ) 6.6 Hz). Animal Treatment. Male, Fischer 344 rats (Charles River Breeding Laboratories, Kingston, NY; 250-275 g) were given 10 mmol/kg of HCFC-131 or HCFC-141 ip dissolved in 0.2 mL of corn oil or 10 mmol/kg of chlorofluoroacetic acid in saline (14). The animals were placed in metabolic cages immediately after treatment and were kept on a 12-h light-and-dark cycle. Urine was collected for up to 24 h in tubes maintained at -78 °C with dry ice and was stored frozen until analyzed. Urinary Metabolite Identification. 19F NMR spectroscopy was used as the primary tool in metabolite identification. Typically, 0.5 mL of rat urine and 0.05 mL of D2O were placed in a 5-mm NMR tube along with a coaxial tube containing 5 mM trifluoroacetamide in D2O. 19F NMR spectra were recorded at room temperature. The presence of glucuronide conjugates of metabolites of HCFC-141 was investigated by incubating a mixture of 0.4 mL

52

Chem. Res. Toxicol., Vol. 9, No. 1, 1996

Yin et al.

Figure 2. Flow chart of batch separation of metabolites from urine of rats given HCFC-131. of urine, 0.1 mL of 500 mM potassium phosphate buffer (pH 5.0), 0.1 mL of D2O, and 5 mg of β-glucuronidase in a microcentrifuge tube at 37 °C for 3 h. After centrifugation, the supernatant was examined by 19F NMR spectroscopy. The formation of 2-chloro-2-fluoroethanol from HCFC-141 was examined by comparing the 19F NMR chemical shift, multiplicity, and coupling constants with that of a synthetic standard as well as by 19F NMR coresonance spectra. The observation that incubation of the urine of HCFC-141-treated rats with β-glucuronidase resulted in the intensity increase of this signal supported the assignment. The formation of chlorofluoroacetic acid as a metabolite of HCFC-131 and HCFC-141 was investigated by comparing the observed 19F NMR chemical shifts and spectral patterns with those of synthetic chlorofluoroacetic acid. Furthermore, chlorofluoroacetic acid in urine from HCFC-131- and HCFC-141treated rats and dichloroacetic acid from HCFC-131-treated rats were converted to their benzyl esters (15) and analyzed by GC/ MS analysis, which showed mass spectra identical with those esters formed from authentic acids. The formation of inorganic fluoride as a metabolite of both HCFC-131 and HCFC-141 was determined by 19F NMR spectroscopy and with an ion-selective electrode (16). Urine (0.5 mL) from HCFC-131-treated rats was incubated with 5 mg of acylase I and 0.1 mL 100 mM potassium phosphate buffer at a final pH 7.4 for 16.5 h at 37 °C. Control experiments were conducted in the absence of enzyme. Urine (1.0 mL) from HCFC-131-treated rats was incubated with 3 mL of freshly prepared whole kidney homogenate (37 mg of protein/mL) and 1 mL of 0.132 M potassium phosphate buffer (pH 7.2) at 37 °C for 6 h. The buffer concentration and pH used were optimal for acylase I activity (17). Whole kidney homogenates were prepared from kidneys from Sprague-Dawley rats by homogenization in 0.066 M potassium phosphate buffer (pH 7.2) with a Dounce homogenizer (17). Purification and Characterization of N-(2-Hydroxyethyl)chlorofluoroacetamide in the Urine of Rats Given HCFC-131. Urine (80 mL) from HCFC-131-treated rats was brought to pH 1.4 with hydrochloric acid, and 400 mL of acetone was added to precipitate proteins. After filtration, the acetone was removed by evaporation, and the aqueous solution was subjected to batch separation (see flow chart in Figure 2). Examination of 19F NMR spectra of each fraction showed that the organic acid fraction H and the organic base fraction D contained the same major organic metabolite with a 19F NMR chemical shift of -69.3 ppm, whereas fraction C (amphoteric compounds) contained two minor metabolites, P1 and P2, with chemical shifts of -69.6 and -69.7 ppm, respectively. After drying over sodium sulfate, organic acid fraction H was con-

centrated and derivatized with diazomethane; the product was applied to a silica gel column (20 cm × 1 cm) eluted with a gradient of petroleum ether/ether from 2:1 to pure ether. Fractions were collected and analyzed by 19F NMR spectroscopy. The fractions that contained the major fluorine-containing organic metabolite with a chemical shift of -69.21 ppm were pooled; GC/MS analysis showed the presence of one peak. Analytical data: 1H NMR (acetonitrile-d6, δTMS ) 0 ppm): 7.17 (s, NH), 6.42 (d, CH, 2JHF ) 50.6 Hz), 3.57 (t, CH2O, 3J ) 5.5 Hz), 3.33 (t, CH2N, 3J ) 5.7 Hz), 3.03 (s, OH); 19F NMR (D2O, pH not adjusted): -70.15 ppm (2JHF ) 51.1 Hz); mass spectrometry: m/z 137 (M - H2O), 124 (M - CH2OH), 88 (M CHFCl), 67 (CHFCl), 60 (M - COCHFCl). Silica gel column chromatography of organic base fraction D resulted in the identification of same metabolite. Cation exchange (AG50WX8, 50-100 mesh, pH 4.5) column separation of amphoteric compound C eluted with potassium chloride failed to generate fractions with sufficient fluorine-containing metabolites for detection by 19F NMR spectroscopy. Expression of Cytochrome P450 2E1 Isozymes. Cytochrome P450 2E1 isozyme was expressed in Hep G2 cells, which contain P450 reductase but lack cytochrome P450 activity, according to the methods described by Gonzalez et al. (18). The frozen Hep G2 cells that expressed cytochrome P450 2E1 isozyme were quickly thawed, sonicated with a probe sonicator (50% setting) for 15 pulses, mixed, and sonicated again for 18 pulses. The cell lysate was centrifuged at 100000g for 15 min, and the centrifugate was resuspended in 100 mM potassium phosphate buffer (pH 7.0). Protein concentrations were measured by the method of Bradford (19). Cytochrome P450 contents were measured by CO difference spectroscopy (20). Preparation of Human and Rat Liver Microsomes. Frozen human liver was a gift from Dr. James Harris, Food and Drug Administration. Rat liver tissue was obtained from male, Fischer 344 rats (250-275 g) after anesthetization with ether. In some cases, rats were given 100 mg of pyridine/kg/day ip for 4 days to induce cytochrome P450 2E1 (14). Microsomes were prepared from human and rat liver tissues by homogenization at 4 °C in 100 mM potassium phosphate buffer (pH 7.0) with a Dounce homogenizer. The homogenate was centrifuged at 10000g for 20 min; the supernatant thus obtained was centrifuged at 100000g for 70 min. The microsomal pellets were resuspended at 4 °C by gentle stirring in 100 mM potassium phosphate buffer (pH 7.0). Protein concentrations were measured by the method of Bradford (19). The cytochrome P450 contents were measured by CO difference spectroscopy (20). In Vitro Microsomal Incubations with HCFC-141 and HCFC-131. Incubation mixtures contained pyridine-induced rat liver microsomes (2 mg of protein/mL) or expressed human

Regioselective Metabolism of HCFC-141 and HCFC-131

Chem. Res. Toxicol., Vol. 9, No. 1, 1996 53 Table 1.

19F

NMR Spectral Properties of Metabolites from HCFC-131 in Rats

metabolite fluoride chlorofluoroacetic acid N-(2-hydroxyethyl)chlorofluoroacetamide unknown P1 unknown P2 unknown P3 HCFC-131

Figure 3. 19F NMR spectra of metabolites derived from HCFC131. (A) Urine of rats given HCFC-131 and (B) microsomal incubation mixture of HCFC-131 in the presence of NADPH. See Experimental Procedures for details. cytochrome P450 2E1 (366 pmol of cytochrome P450/mL), 2 mM NADPH, 10 mM HCFC-131 or HCFC-141, and 100 mM potassium phosphate buffer (pH 7.0) and were contained in 20-mL crimp-top seal vials and incubated at 37 °C. Reactions were initiated by addition of NADPH and terminated after 30 min by heating at 70 °C for 5 min and chilling in an ice-water bath for 10 min. The reaction mixtures were stored frozen until analyzed. After thawing, the mixtures were centrifuged for 5 min, and 0.5 mL of the supernatant together with 0.05 mL of D2O was analyzed by 19F NMR spectroscopy and GC/MS, as described above. Determination of Regioselectivity. Two methods were used to determine the relative amounts of metabolites that resulted from oxidation at the two oxidizable sites in HCFC131 and in HCFC-141. First, acid metabolites were quantified by single-ion monitoring GC/MS after conversion to their benzyl esters (15). Pentafluoropropanoic acid was added as internal standard, and standard curves were generated with known amounts of synthetic acids. A plot of the ratio (integration counts) of trihaloacetate:pentafluoropropionate was linear with mole ratio of trihaloacetate:pentafluoropropionate in the range tested (0-100 nmol). Second, the fluorine-containing metabolites were quantified by integration of resonances in 19F NMR spectra. The results obtained by the latter method should, however, be considered as semiquantitative, particularly for inorganic fluoride because of its long relaxation time. Regioselectivity was measured in the in vitro experiments by comparing the amount of products resulting from different sites of oxidation. Regioselectivity for HCFC-141 was expressed as the ratio of the two products, fluoride and chlorofluoroacetaldehyde hydrate, determined by 19F NMR. Regioselectivity of HCFC131 was expressed either as the ratio dichloroacetic acid: chlorofluoroacetic acid, determined by GC/MS, or as the ratio fluoride:chloroactic acid, determined by 19F NMR. Calculation of Activation Energies. The activation energies for the cytochrome P450-dependent hydrogen-atom abstraction from each C-H bond in HCFC-131 and in HCFC-141 were calculated according to the method of Korzekwa et al. (21). Gasphase heats of formation for substrates, intermediate carboncentered radicals, p-nitrosophenol, and p-nitrosophenoxy radical were calculated by the MOPAC program (version 6.0) with AM1 Hamiltonian and precise criteria on a Silicon Graphics workstation. Molecular structures were constructed and input decks were generated by Sybyl program (Tripos Associates, Inc., St. Louis, MO).

Results In Vivo Metabolism of HCFC-131. The 19F NMR spectrum of the urine from rats given HCFC-131 is shown in Figure 3A. The spectral values for each metabolite are listed in Table 1.

chemical shift (ppm)a multiplicityb -42.74 -60.55 -69.21

s d d

-69.32 -69.57 -69.69 -63.69

d d d dd

coupling constants (Hz) 2J HF 2J HF

) 52.5 ) 51.2

2J HF ) 50.0 2J HF ) 51.2 2J HF ) 47.6 2J HF ) 50.3, 3J HF ) 10.8

a Chemical shifts are reported as ppm referenced to trifluoroacetamide (δ ) 0 ppm). 19F NMR spectra were recorded as described under Experimental Procedures. b s ) singlet, d ) doublet, dd ) doublet of doublets.

The broad singlet centered at -42.74 ppm was assigned to inorganic fluoride, as determined by coresonance with authentic sodium fluoride. Inorganic fluoride (1.1 mM) in the urine of rats given HCFC-131 was also detected with a fluoride-selective electrode. The doublet centered at -60.55 ppm (2JHF ) 47.5 Hz) was assigned to chlorofluoroacetic acid. The chemical shift, splitting pattern, and proton-fluorine coupling constants were identical to those of a synthetic standard. The assignment was supported by addition of synthetic chlorofluoroacetic acid to the urine sample and by GC/ MS analysis after conversion to benzyl ester. Benzyl chlorofluoroacetate (retention time 17.7 min) derived from urine had a MS fragmentation pattern identical with that prepared from standard chlorofluoroacetic acid: m/z 202 (M+), 91 (base peak, CH2C6H5). Similarly, dichloroacetic acid was identified as a metabolite: benzyl dichloroacetate (retention time 18.3 min) derived from urine had an MS fragmentation pattern identical with that prepared from standard dichloroacetic acid: m/z 218 (M+), 91 (base peak, CH2C6H5). The metabolite corresponding to the major doublet centered at -69.21 ppm (J ) 51.2 Hz) was purified from urine of HCFC-131-treated rats by batch separation (Figure 2) and chromatography as described under Experimental Procedures and was identified as N-(2hydroxyethyl)chlorofluoroacetamide. GC/MS analysis showed a retention time (12.6 min) and an MS fragmentation pattern identical with that of a synthetic standard (data not shown). 1H NMR spectra of the metabolite matched that of synthetic standard (see Experimental Procedures). Three other metabolites were detected, but not identified; their 19F NMR chemical shifts were as follows: unknown P1, -69.32 ppm (J ) 50.0 Hz); unknown P2, -69.57 ppm (J ) 51.2Hz); and unknown P3, -69.69 ppm (J ) 47.6 Hz). In an effort to characterize these metabolites, urine from rats given HCFC-131 was subjected to various enzymatic incubations. β-Glucuronidase incubation at pH 5.0 resulted in the disappearance of the doublet corresponding to metabolite P3 in the 19F NMR spectrum (Figure 4B), but no detectable change in the intensity of other resonances was observed and a new resonance was not detected. Since N-(2-hydroxyethyl)chlorofluoroacetamide has a free hydroxyl group and was present in relatively large amounts, unknown P3 may be N-(2hydroxyethyl)chlofluoroacetamide glucuronide. Incubation of synthetic N-(2-hydroxyethyl)chlofluoroacetamide

54

Chem. Res. Toxicol., Vol. 9, No. 1, 1996

Yin et al.

Figure 5. 19F NMR spectra of metabolites of HCFC-141. (A) Urine of rats given HCFC-141 and (B) microsomal incubation mixture of HCFC-141 in the presence of NADPH. See Experimental Procedures for details.

Figure 4. 19F NMR spectra (-60 to -70 ppm) of urine of rats given HCFC-131: (A) without treatment; (B) after incubation with β-glucuronidase at pH 5.0 for 3 h; (C) after incubation with acylase I for 16.5 h at 37 °C (pH 7.4); and (D) after incubation with whole kidney homogenates for 6 h at 37 °C (pH 7.2). See Experimental Procedures for details.

with freshly isolated rat liver microsomes at 30 °C for 22.5 h in the presence of 20 mg of uridine 5′-diphosphoglucuronic acid/mL and 100 mM potassium phosphate buffer (final pH 7.0) failed to generate a glucuronide conjugate detectable by 19F NMR spectroscopy. The chemical shifts of esters of chlorofluoroacetic acid are generally more upfield (e.g., methyl chlorofluoroacetate has a 19F NMR chemical shift of -70.5 ppm (J ) 48.3 Hz)) than the observed resonances. Therefore, the possibility that unknown metabolites P1 and P2 were amides was considered. Incubation of urine with acylase I for 16.5 h resulted in the diminution of resonances corresponding to unknown metabolites P1 and P2 and a concurrent increase in the resonance corresponding to chlorofluoroacetic acid (Figure 4C), whereas control experiments in the absence of enzyme failed to result in such a change. Rat kidney is rich in aminoacylases (17); therefore, urine samples were incubated with freshly prepared whole kidney homogenates, which resulted in the loss of the resonances corresponding to unknown metabolites P1 and P2 and a concurrent increase in the resonances assigned to chlorofluoroacetic acid (Figure 4D). Batch separation experiments showed that metabolites P1 and P2 were not extractable with ether under either acidic or basic conditions, indicating that they contain at least one acid group (e.g., carboxylic group) and one basic group (e.g., amino group). Hence, it is possible that metabolites P1 and P2 are chlorofluoroacetylated peptides or amino acids. Unchanged HCFC131 was not detected in urine by 19F NMR spectroscopy. In Vitro Biotransformation of HCFC-131. Inorganic fluoride, -42.74 ppm, chlorofluoroacetic acid, -60.55 ppm (J ) 47.5), and the unchanged HCFC-131, -63.9 ppm (J ) 50.3, 10.8), were detected by 19F NMR spec-

troscopy after incubation of HCFC-131 with liver microsomes from pyridine-induced rats in the presence of NADPH (Figure 3B). The same products were detected in the incubation with expressed cytochrome P450 2E1 enzyme. In Vivo Metabolism of HCFC-141. The 19F NMR spectrum of the urine of rats given HCFC-141 is shown in Figure 5A. The spectral data of each fluorine-containing metabolite are listed in Table 2. The broad singlet centered at -44.04 ppm was identified as inorganic fluoride, as described above. The small doublet centered at -60.59 ppm (J ) 52.5 Hz) was characterized as chlorofluoroacetic acid. The two sets of doublets of doublets of doublets centered at -66.26 ppm (2JHF ) 49.8 Hz, 3JHFa ) 20.3 Hz, 3JHFb ) 17.0 Hz) and -66.65 ppm (2JHF ) 51.3 Hz, 3JHFa ) 24.7 Hz, 3JHFb ) 17.8 Hz) were assigned to diastereomeric mixtures of 2-chloro-2-fluoroethyl glucuronides. The expanded spectrum is shown in Figure 6A. Free 2-chloro-2-fluoroethanol was also observed as a doublet of triplets centered at -67.06 ppm (2JHF ) 50.1 Hz, 3JHF ) 20.1 Hz). Incubation of the urine with β-glucuronidase resulted in a complete loss of the resonances assigned to 2-chloro-2-fluoroethyl glucuronide and a concomitant increase in intensity of the resonances assigned to 2-chloro-2-fluoroethanol (Figure 6B), which had chemical shifts, multiplicities, and proton-fluorine coupling constants identical with those of synthetic 2-chloro-2-fluoroethanol. Addition of synthetic 2-chloro2-fluoroethanol to the biological sample confirmed the assignment. The ratio of the diastereomeric isomers of 2-chloro-2-fluoroethyl glucuronide was determined as 1:1.6 by integration of peaks in the 19F NMR spectrum. Chloroacetic acid was not detected by GC/MS in the urine. Unchanged HCFC-141 was not detected by 19F NMR spectroscopy in urine. In Vitro Biotransformation of HCFC-141. Figure 5B shows the 19F NMR spectrum recorded after a 30min incubation of HCFC-141 with liver microsomes from pyridine-induced rats in the presence of 2 mM NADPH at 37 °C (pH 7.0). The broad singlet centered at -43.6 ppm was identified as inorganic fluoride. The small doublet centered at -72.69 ppm (2JHF ) 53.9, 3JHF ) 6.6 Hz) was assigned to chlorofluoroacetaldehyde hydrate. Synthetic chlorofluoroacetaldehyde hydrate coresonated with the peaks assigned to the metabolite. Unchanged HCFC-141 was also detected: -61.80 ppm (double

Regioselective Metabolism of HCFC-141 and HCFC-131 Table 2.

19F

metabolite fluoride chlorofluoroacetic acid 2-chloro-2-fluoroethyl glucuronide 2-chloro-2-fluoroethanol chlorofluoroacetaldehyde hydrate HCFC-141

Chem. Res. Toxicol., Vol. 9, No. 1, 1996 55

NMR Spectral Properties of Metabolites from HCFC-141 in Rats chemical shift (ppm)a

multiplicityb

-44.04 -60.59 isomer 1: -66.26 isomer 2: -66.65 -67.06 -72.69 -61.69

s d ddd ddd dt dd dt

coupling constants (Hz) 2J

HF

2J

HF

) 52.5 ) 49.8, 3JHaF ) 20.3, 3JHbF ) 17.0 2J 3 3 HF ) 51.3, JHaF ) 24.7, JHbF ) 17.8 2J 3 HF ) 50.1, JHF ) 20.1 2J 3 HF ) 53.9, JHF ) 6.6 2J 2 HF ) 50.3, JHF ) 18.0

a Chemical shifts are reported as ppm referenced to trifluoroacetamide (δ ) 0 ppm). 19F NMR spectra were recorded as described under Experimental Procedures. b s ) singlet, d ) doublet, dd ) doublet of doublets, ddd ) doublet of doublets of doublets, dt ) doublet of triplets.

Figure 6. 19F NMR spectra (-65.4 to -67.6 ppm) of urine of rats given HCFC-141: (A) without treatment; (B) after incubation with β-glucuronidase at 37 °C (pH 5.0) for 3 h. See Experimental Procedures for details.

triplet, 2JHF ) 50.1 Hz, 3JHF ) 20.1 Hz). GC/MS analysis showed the presence of chloroacetic acid after conversion to benzyl chloroacetate: rentention time 17.7 min; m/z 184 (M+), 91 (base peak, CH2C6H5), 77 (ClCH2CO). The retention time and MS fragmentation pattern were identical with those of synthetic benzyl chloroacetate. In Vivo Metabolism of Chlorofluoroacetic Acid. In order to determine whether dihaloacetic acids are subject to metabolism, rats were given synthetic chlorofluoroacetic acid in saline, and the urine was analyzed by 19F NMR spectroscopy (data not shown). The parent compound (-60.59 ppm, J ) 52.5 Hz) was recovered as the major fluorine-containing product in the urine. A substantial amount of inorganic fluoride (-43.0 ppm, singlet) was detected as a metabolite. Regioselectivity of the Biotransformation of HCFC-131 and HCFC-141. Both HCFC-141 and HCFC131 have two dissimilar C-H bonds that may be oxidized by cytochromes P450. Regioselectivity in the cytochrome P450-dependent biotransformation of HCFC-131, expressed as the ratio of oxidation at carbons C1 and C2, was measured by GC/ MS (see Experimental Procedures). The ratio of dichloroacetic acid:chlorofluoroacetic acid was 0.25:1 in rat liver microsomal incubations and 0.49:1 in expressed human cytochrome P450 2E1 system (Table 3); when measured by 19F NMR spectroscopy (see Experimental Procedures), the ratio of fluoride:chlorofluoroacetic acid was 0.48:1 in rat liver microsomal incubations (Table 3). The ratio of fluoride:all organofluorine compounds in in vivo experiments measured by 19F NMR spectrosopy was 2.1:1. This ratio does not reflect the true regioselectivity because of differences in excretion, reabsorption, secondary metabolism (e.g., fluoride generation from chlorofluoroacetic acid), other routes of fluoride formation, and limitations of the method of detection.

The in vitro ratio of oxidation at carbon C1 versus carbon C2 in HCFC-141 was 11.8:1, expressed as the ratio fluoride:chlorofluoroacetaldehyde hydrate, in rat liver microsomal incubation (Table 3). The in vivo ratio of fluoride:all organofluorine metabolites in the urine of rats given HCFC-141 was 35.3:1 (Table 3). 19F NMR spectroscopy was used for ratio determination in both experiments. Again, the in vivo ratio does not reflect the true regioselectivity for the reasons mentioned above. These data demonstrated the existence of regioselectivity in the biotransformation of HCFC-131 and HCFC141 by cytochrome P450. The ratio measured in vitro reflects the true regioselectivity, whereas the in vivo data are composites of several metabolic steps and do not represent the regioselective oxidation. Calculations of Activation Energies of the Cytochrome P450-Mediated Hydrogen-Atom Abstraction from Substrates. The activation energy for the hydrogen-atom abstraction step from C1-H and C2-H bonds in each substrate was calculated by the method of Korzekwa et al. (21). The electronically based theoretical regioselectivity, expressed as the ratio of oxidation at carbons C1 and C2, was calculated from the activation energies difference for hydrogen abstraction from carbons C1 and C2:

k1/k2 ) exp[-(∆Hact.1 - ∆Hact.2)/RT]

(1)

where k1 and k2 represent the rate constants for hydrogenatom abstraction at carbons C1 and C2, respectively; ∆Hact.1 and ∆Hact.2 represent activation energies; and R is the gas constant, T the absolute temperature. The results are shown in Table 3. The predicted ratio of oxidation by cytochrome P450 at carbons C1 and C2 was 0.78:1 for HCFC-131 and 31.6:1 for HCFC-141.

Discussion HCFC-131 and HCFC-141 were chosen as models to investigate the metabolism of H(C)FCs and to investigate the regioselectivity of the biotransformation of substrates bearing multiple oxidizable sites. The metabolites identified (Tables 1 and 2) were consistent with previously identified pathways for the cytochrome P450-dependent biotransformation mechanism of small halogenated alkanes. The proposed mechanism of metabolite formation in vivo from HCFC-131, HCFC-141, and chlorofluoroacetic acid is depicted in Figure 7. The first step in the cytochrome P450-dependent oxidation is hydrogen-atom abstraction at either carbon C1 or C2 in HCFC-141 or HCFC-131 by a heme-ironoxy radical, to form a carbon-centered radical, followed by oxygen rebound to give geminal halohydrins 2a, 2b, 3a, or 3b. Elimination of HX from 2a or 2b gives acyl

56

Chem. Res. Toxicol., Vol. 9, No. 1, 1996

Yin et al.

Table 3. Regioselectivity of the Cytochrome P450-Dependent Oxidation at C1 and C2 in HCFC-131 and in HCFC-141 and the Theoretical Ratio of Hydrogen-Atom Abstraction from Carbons C1 and C2 substratea

sites of oxidation

pyridine-induced rat liver microsomes

expressed cytochrome P450 2E1

rat urineb

calcd ∆Hact. (kcal/mol)

theoretical ratio

HCFC-131

C1/C2

0.25:1 (GC/MS) 0.48:1 (NMR)

0.49:1 (GC/MS)

2.1:1 (NMR)

C1: 25.383 C2: 25.232

0.78:1

HCFC-141

C1/C2

NDc

35.3:1 (NMR)

C1: 24.305 C2: 26.350

11.8:1 (NMR)

31.6:1

a Racemic mixtures were used for study. b Numbers are expressed as integration of fluoride over integration of all organic fluorinecontaining metabolites combined in 19F NMR spectra. c Not done.

Figure 7. Proposed mechanism for the biotransformation of HCFC-141 and HCFC-131 in rats. UDPGT ) uridine diphosphoglucuronosyl transferase, UDPGA ) uridine diphosphoglucuronic acid, GSH ) glutathione, GST ) glutathione transferase, Nu: ) nucleophiles.

halides 4a or 4b, which may be hydrolyzed to give chloroacetic acid 5a or dichloroacetic acid 5b. Chloroacetic acid is metabolized by glutathione transferasedependent pathways (22), but was not detected in rat urine in the present experiments. Alternatively, 4a or 4b may react with cellular nucleophiles to form covalent adducts; these are not detectable by 19F NMR spectroscopy because no fluorine nuclei are present. Elimination of HX from 3a or 3b may generate chlorofluoroacetaldehyde (7a) or chlorofluoroacetyl chloride (7b), respectively. Chlorofluoroacetaldehyde can be further oxidized to chlorofluoroacetic acid (9a), which in turn may be further metabolized to inorganic fluoride. Alternatively, 7a can be reduced to 2-chloro-2-fluoroethanol (10a), which was converted to a mixture of diastereomeric glucuronides 11a through the action of UDP-glucuronosyl transferase. Chlorofluoroacetyl chloride (7b) may react with water or with cellular nucleophiles to form chlorofluoroacetic acid (8b) (R ) OH), N-(2-hydroxyethyl)chlorofluoroacetamide (8b) (R ) NHCH2CH2OH), or other products. The 2-hydroxyethylamine moiety in urinary adduct 8b (R ) NHCH2CH2OH) may be derived from phosphatidylethanolamine. A similar metabolite, N-(2-hydroxyethyl)trifluoroacetamide, was identified in the urine of patients anesthetized with halothane (23). Subsequent multinuclear NMR spectroscopic studies identified phosphotidylethanolamine as the molecular target for trifluoroacetyl halides (24). Harris et al. (3) showed that

trifluoroacetyl halides modify the -amino group of Llysine residues in proteins, an event associated with neoantigen formation and toxicity. The modification of the amino group in phosphotidylethanolamine, a lipid component, may cause membrane dysfunction. It appears from the present study and others (3) that amino groups are the primary targets for stable adduct formation by acyl halides. However, it is not clear whether the formation of amides is the kinetically favored product, or whether these products are thermodynamically more stable than other products. Chlorofluoroacetic acid given ip was largely recovered unchanged in the urine of rats. Fluoride was detected as a metabolite of chlorofluoroacetic acid. The mechanism of defluorination has not been established. Chloroacetic acid is extensively metabolized in vivo to give S-(carboxymethyl)cysteine and thiodiacetic acid by the action of the glutathione transferases (22). Dichloroacetic acid is present as a contaminant in drinking water (25, 26) and is metabolized to glyoxylic acid and oxalic acid (27). Cytochrome P450-dependent regioselective oxidation of HCFC-131 (C1/C2 ) 0.25-0.49:1) and HCFC-141 (C1/ C2 ) 11.8:1) was observed, based on the ratio of products from different sites of oxidation (Table 3). The observed regioselectivity is in agreement with the theoretical values based on an electronic model (C1/C2 ) 0.78:1 for HCFC-131 and C1/C2 ) 35.3:1 for HCFC-141), indicating that electronic effects are the primary factors that govern

Regioselective Metabolism of HCFC-141 and HCFC-131

regioselectivity. It should be noted that the theoretical calculations overestimate the observed ratios, perhaps for the following two reasons: (1) Semiempirical calculations (e.g., AM1) have errors associated with them (28), and the p-nitrosophenoxy radical (PNR model) (21) may not be a perfect mimic of the heme-iron-oxy radical involved in cytochrome P450 reactions. Indeed, the calculated activation energies (24.3-26.4 kcal/mol, see Table 3) overestimate the true values. The turnover of HCFCs by P4502E1 is in the range of 5.1-0.067 nmol/ (min‚nmol of P450) (29), which corresponds to an activation energy of 14.3-17.0 kcal/mol assuming an Arrehnius constant of 109 s-1, or 18.5-21.2 kcal/mol assuming an Arrehnius constant of 1012 s-1. (2) Presumably, there are two spatial orientations of a substrate in the enzyme active site, each orientation leading to oxidation at carbons C1 or C2. If apoprotein-substrate interaction is significant and one orientation is preferred over another, or if switching between the two orientations is slow, the theoretically calculated intrinsic ratio of rate constants for hydrogen-atom abstraction may be partially masked, i.e., the overall observed regioselectivity may not reflect fully the theoretical electronic factors. In conclusion, the identified metabolites of HCFC-141 and HCFC-131 in rats and with human cyto-chrome P450 2E1 point to a cytochrome P450-depen-dent phase I biotransformation mechanism. Regioselectivity between two oxidizable sites in each molecule was observed. Comparison of electronic factor-based theoretical calculations and experimental data suggested that electronic effects are the primary factors that govern regioselectivity.

Acknowledgment. We thank Sandy Morgan for her assistance in preparing the manuscript and Grace A. Bennett for the synthesis of chlorofluoroacetic acid and chlorofluoroethanol. The authors are grateful to Dr. Kenneth Kozekwa for providing TK- cells and P-4502E1 recombinant vaccinia virus. This research was supported by National Institute of Enviromental Health Science Grant ES05407 to M.W.A. H.Y. was supported by a fellowship from Pharmaceutical Research and Manufacturers of America Foundation, Inc.

References (1) Anders, M. W. (1991) Metabolism and toxicity of hydrochlorofluorocarbons: Current knowledge and needs for the future. Environ. Health Perspect. 96, 185-191. (2) Molina, M. J., and Rowland, F. S. (1974) Stratospheric sink for chlorofluoromethanes: Chlorine atom-catalyzed destruction of ozone. Nature 249, 810-812. (3) Harris, J. W., Pohl, L. R., Martin, J. L., and Anders, M. W. (1991) Tissue acylation by the chlorofluorocarbon substitute 2,2-dichloro1,1,1-trifluoroethane (HCFC-123). Proc. Natl. Acad. Sci. U.S.A. 88, 1407-1410. (4) Harris, J. W., and Anders, M. W. (1991) Metabolism of the hydrochlorofluorocarbon 1,2-dichloro-1,1-difluoroethane. Chem. Res. Toxicol. 4, 180-186. (5) Harris, J. W., and Anders, M. W. (1991) In vivo metabolism of the hydrochlorofluorocarbon 1,1-dichloro-1-fluoroethane (HCFC141b). Biochem. Pharmacol. 41, R13-R16. (6) Olson, M. J., Reidy, C. A., Johnson, J. T., and Pederson, T. C. (1990) Oxidative defluorination of 1,1,1,2-tetrafluoroethane (R134a) by rat liver microsomes. Drug Metab. Dispos. 18, 992-998.

Chem. Res. Toxicol., Vol. 9, No. 1, 1996 57 (7) Olson, M. J., Reidy, C. A., and Johnson, J. T. (1990) Defluorination of 1,1,1,2-tetrafluoroethane (R-134a) by rat hepatocytes. Biochem. Biophys. Res. Commun. 166, 1390-1397. (8) Olson, M. J., Kim, S. G., Reidy, C. A., Johnson, J. T., and Novak, R. F. (1991) Oxidation of 1,1,1,2-tetrafluoroethane (R-134a) in rat liver microsomes is catalyzed primarily by cytochrome P450IIE1. Drug Metab. Dispos. 19, 298-303. (9) Yin, H., Jones, J. P., and Anders, M. W. (1995) Metabolism of 1-fluoro-1,1,2-trichloroethane, 1,2-dichloro-1,1-difluoroethane, and 2-chloro-1,1,1-trifluoroethane. Chem. Res. Toxicol. 8, 262-268. (10) Harris, J. W., Jones, J. P., Martin, J. L., LaRosa, A. C., Olson, M. J., Pohl, L. R., and Anders, M. W. (1992) Pentahaloethanebased chlorofluorocarbon substitutes and halothane: Correlation of in vivo hepatic protein-trifluoroacetylation and urinary trifluoroacetic acid excretion with calculated enthalpies of activation. Chem. Res. Toxicol. 5, 720-725. (11) Eugland, B. (1963) 2-Chloro-1,1,2-trifluoroethyl ethyl ether. In Organic Syntheses, Collect. Vol. 4, pp 184-188, John Wiley & Sons, New York. (12) Pearson, D. L. (1990) Cornell Univeristy Doctoral Thesis, pp 74-78. (13) Middleton, W. J. (1979) One-step method for converting esters to acyl chlorides. J. Org. Chem. 44, 2291-2292. (14) Carlson, G. P., and Day, B. J. (1992) Induction by pyridine of cytochrome P4502E1 and xenobiotic metabolism in rat lung and liver. Pharmacology 44, 117-123. (15) Gyllenhaal, O., and Ehrsson, H. (1975) Determination of sulphonamides by electron-capture gas chromatography. Preparation and properties of perfluoroacyl and pentafluorobenzyl derivatives. J. Chromatogr. 107, 327-333. (16) Olson, M. J., and Surbrook, S. E., Jr. (1991) Defluorination of the CFC-substitute 1,1,1,2-tetrafluoroethane: Comparison in human, rat and rabbit hepatic microsomes. Toxicol. Lett. 59, 89-99. (17) Paik, W. K., Bloch-Frankenthal, L., Birnbaum, S. M., Winitz, M., and Greenstein, J. P. (1957) -Lysine acylase. Arch. Biochem. Biophys. 69, 56-66. (18) Gonzalez, F. J., Aoyama, T., and Gelboin, H. V. (1991) Expression of mammalian cytochrome P450 using vaccinia virus. Methods Enzymol. 206, 85-92. (19) Bradford, M. M. (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72, 248-254. (20) Bomont, H. L., Tarbit, M. H., Humphrey, M. J., and Houston, J. B. (1994) Disposition of azole antifungal agents. II. Hepatic binding and clearance of dichlorophenyl-bis-triazolylpropanol (DTP) in the rat. Pharm. Res. 11, 951-60. (21) Korzekwa, K. R., Jones, J. P., and Gillette, J. R. (1990) Theoretical studies on cytochrome P-450 mediated hydroxylation: A predictive model for hydrogen atom abstractions. J. Am. Chem. Soc. 112, 7042-7046. (22) Yllner, S. (1971) Metabolism of chloroacetate-1-14C in the mouse. Acta Pharmacol. Toxicol. 30, 69-80. (23) Cohen, E. N., Trudell, J. R., Edmunds, H. N., and Watson, E. (1975) Urinary metabolites of halothane in man. Anesthesiology 43, 392-401. (24) Muller, R., and Stier, A. (1982) Modification of liver microsomal lipids by halothane metabolites: a multinuclear NMR spectroscopic study. Naunyn-Schmiedebergs Archiv. Pharmacol. 321, 234-237. (25) Kanhai, W., Koob, M., Dekant, W., and Henschler, D. (1991) Metabolism of 14C-dichloroethyne in rats. Xenobiotica 21, 905916. (26) McCall, S. N., Jurgens, P., and Ivanetich, K. M. (1983) Hepatic microsomal metabolism of the dichloroethanes. Biochem. Pharmacol. 32, 207-213. (27) Lin, E. L. C., Mattox, J. K., and Daniel, F. B. (1993) Tissue distribution, excretion, and urinary metabolites of dichloroacetic acid in the male fischer 344 rat. J. Toxicol. Environ. Health 38, 19-32. (28) Anders, M. W., Yin, H., and Jones, J. P. (1995) Application of computational chemistry in biological reactive intermediates. Biological Reactive Intermediates (in press). (29) Yin, H. (1995) University of Rochester Doctoral Thesis, pp 101-141.

TX950086N