Structure−Activity Relationship for the Biotransformation of

1036

Chem. Res. Toxicol. 2002, 15, 1036-1041

Structure-Activity Relationship for the Biotransformation of Haloalkenes by Rat Liver Microsomal Glutathione Transferase 1 Larry J. Jolivette† and M. W. Anders* Department of Pharmacology and Physiology, University of Rochester, 601 Elmwood Avenue, Box 711, Rochester, New York 14642 Received February 21, 2002

Many haloalkenes are nephrotoxic in rodents, and experimental evidence supports a glutathione-dependent bioactivation pathway that leads to nephrotoxicity or nephrocarcinogenicity, or both. The reaction of glutathione with haloalkenes is catalyzed by cytosolic glutathione transferases (cGST) and microsomal glutathione transferase 1 (MGST1). The aim of this study was to develop a computational approach to predict the competency of cGST and MGST1 to catalyze the reaction of glutathione with a range of haloalkenes. The hypothesis tested was that the semiempirically computed energy of the lowest unoccupied molecular orbital (ELUMO) of a haloalkene may be used to predict the competency of cGST and MGST1 to catalyze its reaction with glutathione. The MGST1- and cGST-catalyzed reaction of glutathione with nine haloalkenes with ELUMO values ranging from -1.14 to 0.38 eV was determined experimentally. The data indicated that the ELUMO values for haloalkenes were inversely related to the specific activity of the MGST1-catalyzed reaction but not the cGST-catalyzed reaction. These data also demonstrated that MGST1 catalyzed the reaction of glutathione with haloalkenes with ELUMO values equal to or more negative than -0.73 eV and that cGST catalyzed the reaction of glutathione with haloalkenes with ELUMO values more negative than -0.06 eV.

Introduction Haloalkenes are high-volume chemicals used in industrial, synthetic, and pharmaceutical applications and are common environmental pollutants. All of the haloalkenes investigated in this study, i.e., trichloroethene, tetrachloroethene, 1,1,2-trichloro-3,3,3-trifluoro-1-propene, hexafluoropropene, and 2-(fluoromethoxy)-1,1,3,3,3-pentafluoro1-propene (Compound A), are nephrotoxic in rodents (19; for a review, see 10). The bioactivation of haloalkenes involves hepatic glutathione S-conjugate formation, enzymatic hydrolysis of the glutathione S-conjugates by peptidases to the corresponding cysteine S-conjugates, and bioactivation of the cysteine S-conjugates to reactive intermediates by renal cysteine conjugate β-lyase. The reaction of glutathione with haloalkenes is catalyzed by both microsomal (MGST1)1 and cytosolic glutathione transferases (cGST) (5, 11-13) or, in some cases, exclusively by cGST (14, 15). No methodology is available, however, to predict whether microsomal or cytosolic glutathione transferases are competent to catalyze the reaction of glutathione with a given haloalkene. Previous studies reported a quantitative structureactivity relationship (QSAR) for the enzyme-catalyzed * Address correspondence to this author at the Department of Pharmacology and Physiology, University of Rochester Medical Center, 601 Elmwood Ave., Box 711, Rochester, NY 14642. Telephone: 585275-1678; Fax: 585-273-2652; E-mail: [email protected]. † Present address: Preclinical Drug Discovery, Cardiovascular and Urogenitary Center for Excellence in Drug Discovery, GlaxoSmithKline, UW2720, 709 Swedeland Rd., King of Prussia, PA 19406. 1 Abbreviations: cGST, cytosolic glutathione transferase; MGST1, microsomal glutathione transferase1; ELUMO, energy of the lowest unoccupied molecular orbital; QSAR, quantitative structure-activity relationship; HOMO, highest occupied molecular orbital; LUMO, lowest unoccupied molecular orbital.

reaction of glutathione with a series of fluoronitrobenzenes (16, 17). In these studies, the natural logarithm of the rate of glutathione conjugation catalyzed by rat liver cytosolic proteins, purified rat liver glutathione transferases 1-1 and 3-3, and purified human liver glutathione transferases A1-1 and M1a-1a correlates with the calculated energy of the lowest unoccupied molecular orbital (ELUMO) and with the calculated relative heat of formation of the Meisenheimer complex for the series of fluoronitrobenzenes with methanethiolate as the model nucleophile. The aim of the present study was to test the hypothesis that a structure-activity relationship exists that predicts the competency of rat liver microsomal glutathione transferase 1 (MGST1) to catalyze the reaction of glutathione with haloalkenes. A computational chemistry approach, analogous to the studies of Rietjens et al. and Soffers et al. (16, 17) in which ELUMO values of fluoronitrobenzenes were correlated with the cGST-catalyzed glutathione S-conjugation activity of fluoronitrobenzenes, was utilized in the present study. The results of this study demonstrated that ELUMO values might be used as a tool to predict the competency of MGST1 to catalyze the reaction of glutathione with haloalkenes and to predict rank-order reactivity for cGST- and MGST1catalyzed glutathione S-conjugation activity among haloalkenes.

Materials and Methods Materials. Hydroxyapatite and glutathione were obtained from Sigma Chemical Co. (St. Louis, MO). Carboxymethyl Sepharose was obtained from BioRad Inc. (Hercules, CA). Male, Sprague-Dawley rats, 300-350 g, were obtained from Charles

10.1021/tx0255222 CCC: $22.00 © 2002 American Chemical Society Published on Web 07/26/2002

Haloalkene Glutathione Conjugation SAR River (Wilmington, MA). 1-Chloro-2,4-dinitrobenzene, trichloroethene, and tetrachloroethene were obtained form Aldrich Chemical Co. (Milwaukee, WI). Hexafluoropropene, trans-1,2dichloroethene, and 1,1-dichloroethene were obtained from Lancaster Synthesis (Windham, NH). 1,1,2-Trichloro-3,3,3trifluoro-1-propene, 1-bromo-2,2-difluoroethene, and 1,1-dibromo2,2-difluoroethene were obtained from PCR Incorporated (Gainesville, FL). 2-(Fluoromethoxy)-1,1,3,3,3-pentafluoro-1-propene was provided by Abbott Laboratories (Abbott Park, IL). Microsomal glutathione transferase 1 (MGST1) was purified as previously described (18). S-(1,2,2-Trichlorovinyl)glutathione was synthesized as previously described by Dekant et al. (4). S-[2-(Fluoromethoxy)-1,1,3,3,3-pentafluoropropyl]glutathione was synthesized as described by Iyer et al. (19). Both S-(1,2,2trichlorovinyl)glutathione and S-[2-(fluoromethoxy)-1,1,3,3,3pentafluoropropyl]glutathione were purified by recrystallization from 85% ethanol as previously described by Dohn et al. (11). The purity (>95%) and identification of synthetic glutathione S-conjugates were determined with atmospheric pressure chemical ionization (APCI) LC/MS. Molecular Orbital Calculations. The ELUMO and the highest occupied molecular orbital (HOMO) energies were computed for the haloalkenes in Figure 1 and for ethanethiolate, respectively. Structures of the haloalkenes were constructed with the Chem3D Ultra software application (version 5.0, CambridgeSoft Inc., Cambridge, MA). Semiempirical molecular orbital calculations were performed with the AM1 Hamiltonian within the MOPAC 98 module contained within the Chem3D Ultra software package. Geometries were optimized in the gas phase for all bond lengths, bond angles, and torsional angles with PRECISE criteria. Assay of Enzyme-Catalyzed Haloalkene-Derived Glutathione S-Conjugate Formation. The formation of glutathione S-conjugates of the haloalkenes shown in Figure 1 was studied with purified rat liver MGST1 and rat liver cytosolic proteins. Reaction mixtures were incubated at 37 °C for 10 and 20 min in 0.1 M potassium phosphate (pH 7.4) containing 0.1% Triton X-100, 0.1 mM EDTA, 1 mM glutathione, 0.1-15 µg of purified rat liver MGST1 or 0.1-4 mg of rat liver cytosolic protein, and 4 mM haloalkene in a final volume of 1 mL. The reactions were terminated with the addition of 40 µL of 70% trichloroacetic acid. The precipitated proteins were removed by centrifugation, and the supernatant was collected for LC/MS analysis. Separation and Quantification of Glutathione S-Conjugate Formation. The reaction products were fractionated with an Agilent 1100 LC/MS system coupled to a Waters NovaPak C18 column (2 × 150 mm). Forty-microliter samples of the supernatants were injected onto the column. The reaction products were eluted with 15% acetonitrile in water containing 0.1% trifluoroacetic acid. The glutathione S-conjugates were detected by monitoring UV absorbance at 202 nm and APCI LC/ MS operating in the positive scan mode (150-550 m/z). The mass spectrophotometer operating conditions were as follows: fragmentor voltage, 120 V; nebulizer pressure, 60 psi; corona voltage, 4 µA; capillary voltage, 3500 V; and drying gas temperature and flow rate, 350 °C and 5 mL/min, respectively. UV absorbance (202 nm) chromatogram peak areas were integrated with the software provided with the LC/MS system. The concentration of the glutathione S-conjugates in the supernatants was determined by comparison of the UV absorbance (202 nm) chromatogram peak area with that of standard curves generated from glutathione, S-(1,2,2-trichlorovinyl)glutathione, and S-[2-(fluoromethoxy)-1,1,3,3,3-pentafluoropropyl]glutathione, all of which yielded nearly identical calibration curves. The lower limit for quantification of glutathione conjugate formation by UV absorbance was 0.5 nmol of glutathione conjugate/mL.

Results Molecular Orbital Calculations. The lowest unoccupied molecular orbitals (LUMO) and the energies of

Chem. Res. Toxicol., Vol. 15, No. 8, 2002 1037

Figure 1. Calculated energies of the lowest unoccupied molecular orbitals (ELUMO) for (A) hexafluoropropene, (B) 2-(fluoromethoxy)-1,1,3,3,3-pentafluoro-1-propene, (C) 1,1,2-trichloro3,3,3-trifluoro-1-propene, (D) 1,1-dibromo-2,2-difluoroethene, (E) tetrachloroethene, (F) 1-bromo-2,2-difluoroethene, (G) trichloroethene, (H) trans-1,2-dichloroethene, and (I) 1,1-dichloroethene. The ELUMO values of the haloalkenes were computed with the MOPAC 98 module in Chem3D v. 5.0 (CambridgeSoft). Semiempirical molecular orbital calculations were performed with the AM1 Hamiltonian within the MOPAC 98 program. Geometries were optimized in the gas phase for all bond lengths, bond angles, and torsional angles with PRECISE criteria.

Figure 2. Lowest unoccupied molecular orbitals (LUMO) for 1,1,2-trichloro-3,3,3-trifluoropropene (panel A) and 1,1-dichloroethene (panel B). The shape and size of the LUMO of these haloalkenes are representative of the LUMO for each of the haloalkenes tested in this study. LUMO of the haloalkenes were computed with the MOPAC 98 module in Chem3D v. 5.0 (CambridgeSoft). Semiempirical molecular orbital calculations were performed with the AM1 Hamiltonian within the MOPAC 98 program. Geometries were optimized in the gas phase for all bond lengths, bond angles, and torsional angles with PRECISE criteria.

the LUMO (ELUMO) were computed for the haloalkenes shown in Figure 1. The LUMO of each the haloalkenes was similar in shape and size, as shown in Figure 2 for the LUMO of 1,1,2-trichloro-3,3,3-trifluoropropene and 1,1-dichloroethene. The ELUMO values of the haloalkenes studied ranged from -1.14 to 0.38 eV (Figure 1). Enzyme-Catalyzed Haloalkene-Derived Glutathione S-Conjugate Formation. Haloalkenes were incubated with glutathione in the presence and absence of rat liver cytosolic proteins or purified rat liver MGST1, as described above. The rates of glutathione conjugate formation were determined after incubation times that resulted in less than 10% loss of the limiting reactant, i.e., glutathione, and the rates of product formation were the same in reaction mixtures incubated for 10 and 20

1038

Chem. Res. Toxicol., Vol. 15, No. 8, 2002

Jolivette and Anders

Table 1. Specific Activity of the Enzyme-Catalyzed Glutathione Conjugation Reaction by Purified Rat Liver MGST1 and Rat Liver Cytosolic Protein for Various Haloalkenes glutathione S-conjugates

substrate hexafluoropropene 2-(fluoromethoxy)-1,1,3,3,3-pentafluoro-1-propene 1,1,2-trichloro-3,3,3-trifluoropropene 1,1-dibromodifluoroethene tetrachloroethene 1-bromo-2,2-difluoroethene trichloroethene trans-1,2-dichloroethene 1,1-dichloroethene a

retention time (min)

parent (MH+) ion (m/z)

5.9 6.3 6.9 8.4 14.4 15.9 6.0 6.7 3.1 2.9 3.8 nd nd

438 453 488 468 470 470 530 438 450 402 402 nd nd

glutathione transferase activity sp act.a [µmol min-1 (mg of MGST1 protein)-1]

sp act.a [nmol min-1 (mg of cytosolic protein)-1]

6.22 ( 0. 38

122 ( 1.4

5.00 ( 0.51

174 ( 11

4.39 ( 0.23

139 ( 2.7

1.66 ( 0.27 ndb nd nd

62.0 ( 1.2 0.069 ( 0.005 low activityc low activity

nd nd

nd nd

Data are shown as means ( SD, n ) 6. b nd, no conjugate formation detected. c Conjugate formation detected but not quantified.

min, indicating linear rates of product formation. For all haloalkenes studied, no conjugate formation was detected when glutathione or substrate was not included in the incubation mixture. All rates of conjugate formation were corrected for nonenzymatic product formation. The results are summarized in Table 1 and are described below for each haloalkene. Hexafluoropropene. Incubation of hexafluoropropene with glutathione in the presence of purified MGST1 or rat liver cytosolic proteins resulted in the time- and protein-concentration-dependent formation of two conjugates that eluted at 5.9 and 6.3 min and were detected by both UV and APCI mass spectral analyses. The mass spectrum of the product that eluted at 5.9 min showed a parent (MH+) ion at m/z 438, indicating that the conjugate was a substitution product that resulted from the reaction of glutathione with hexafluoropropene. The mass spectrum of the conjugate that eluted at 6.3 min showed a MH+ ion at m/z 458, indicating that the conjugate formed was an addition product that resulted from the reaction of glutathione with hexafluoropropene. These conjugates were assigned as S-(1,2,3,3,3-pentafluoro-1propenyl)glutathione (m/z 438) and S-(1,1,2,3,3,3-hexafluoropropyl)glutathione (m/z 458) based on the structural assignments of the glutathione S-conjugates that resulted from the reaction of glutathione with hexafluoropropene reported by Koob and DeKant (13). The enzyme-catalyzed rates of conjugate formation are shown in Table 1. 2-(Fluoromethoxy)-1,1,3,3,3-pentafluoro-1-propene. Incubation of 2-(fluoromethoxy)-1,1,3,3,3-pentafluoro-1-propene with glutathione in the presence of purified rat liver MGST1 or rat liver cytosolic proteins resulted in the timeand protein-concentration-dependent formation of two conjugates that eluted at 6.9 and 8.4 min and were detected by both UV and APCI mass spectral analyses. The mass spectrum of the conjugate that eluted at 6.9 min showed a MH+ ion at m/z 488, indicating that the conjugate was an addition product that resulted from the reaction of glutathione with 2-(fluoromethoxy)-1,1,3,3,3pentafluoro-1-propene. The mass spectrum of the conjugate that eluted at 8.4 min showed a MH+ ion at m/z 468, indicating that the conjugate was a substitution product that resulted from the reaction of glutathione with 2-(fluoromethoxy)-1,1,3,3,3-pentafluoro-1-propene. These conjugates were assigned as S-[2-(fluoromethoxy)-1,1,3,3,3pentafluoropropyl]glutathione (m/z 488) and S-[2-(fluo-

romethoxy)-1,3,3,3-tetrafluoro-1-propenyl]glutathione (m/z 468) based on the structural assignments of the glutathione S-conjugates that resulted from the reaction of glutathione with 2-(fluoromethoxy)-1,1,3,3,3-pentafluoro-1propene reported by Jin et al. (12). The enzyme-catalyzed rates of conjugate formation are shown in Table 1. 1,1,2-Trichloro-3,3,3-trifluoro-1-propene. Incubation of 1,1,2-trichloro-3,3,3-trifluoro-1-propene with glutathione in the presence of purified rat liver MGST1 or rat liver cytosolic proteins resulted in the time- and proteinconcentration-dependent formation of two conjugates that eluted at 14.4 and 15.9 min and were detected with UV and APCI mass spectral analyses. The products that eluted at 14.4 and 15.9 min had identical mass spectra with MH+ ions of m/z 470, indicating that both products were substitution products that resulted from the reaction of glutathione with 1,1,2-trichloro-3,3,3-trifluoro-1propene. The mass spectra showed isotope clusters with an abundance ratio 100:66.5:14.1 for the MH+, MH+2, and MH+4 ions, indicating that both conjugates contained two chlorine atoms. The mass and number of chlorine atoms in these products were consistent with the report of Vamvakas et al., who identified the glutathione Sconjugates of 1,1,2-trichloro-3,3,3-trifluoro-1-propene as (E)- and (Z)-S-(1,2-dichloro-3,3,3-trifluoro-1-propenyl)glutathione (5). The enzyme-catalyzed rates of conjugate formation are shown in Table 1. 1,1-Dibromo-2,2-difluoroethene. Incubation of 1,1-dibromo-2,2-difluoroethene with glutathione in the presence of purified rat liver MGST1 or rat liver cytosolic proteins resulted in the time- and protein-concentrationdependent formation of a single conjugate that eluted at 6.0 min and was detected by UV and APCI mass spectral analyses. The mass spectrum of the conjugate showed a MH+ ion of m/z 528 and an isotope cluster with an abundance ratio of 50.7:100.0:55.7 for the MH+, MH+2, and MH+4, indicating that the conjugate contained two bromine atoms. The mass and number of bromine atoms indicated that this product was an addition product that resulted from the reaction of glutathione with 1,1dibromo-2,2-difluoroethene, likely S-(2,2-dibromo-1,1-difluoroethyl)glutathione. The enzyme-catalyzed rate of conjugate formation is shown in Table 1. Tetrachloroethene. Incubation of tetrachloroethene with glutathione in the presence of rat liver cytosolic protein resulted in the time- and protein-dependent formation

Haloalkene Glutathione Conjugation SAR

Chem. Res. Toxicol., Vol. 15, No. 8, 2002 1039

Figure 3. Relationship between the natural logarithm of the specific activity of the MGST1-catalyzed and the rat liver cytosolic protein-catalyzed glutathione conjugation reactions and the ELUMO values of haloalkenes: (A) hexafluoropropene; (B) 2-(fluoromethoxy)1,1,3,3,3-pentafluoro-1-propene; (C) 1,1,2-trichloro-3,3,3-trifluoro-1-propene; (D) 1,1-dibromo-2,2-difluoroethene; and (E) tetrachloroethene. The ELUMO values of the haloalkenes were computed with the MOPAC 98 module in Chem3D v. 5.0 (CambridgeSoft), as described under Materials and Methods. The haloalkenes were incubated with glutathione in the presence or absence of purified MGST1 or rat liver cytosolic proteins in 0.1 M phosphate buffer (pH 7.4) containing 0.1% Triton X-100 and 0.1 mM EDTA, for 10-20 min at 37 °C. Glutathione S-conjugate formation was determined with HPLC analysis, as described under Materials and Methods. Data are shown as means ( SD, n ) 6. The correlation coefficient (r2) of the linear relationship of the specific activities for the MGST1-catalyzed reactions with ELUMO was 0.943.

of a conjugate that eluted at 6.7 min and was detected with UV absorbance and APCI mass spectral analyses. The mass spectrum of the conjugate showed a MH+ ion of m/z 436 and an isotope cluster with an abundance ratio of 99.8:100.0:37.7:6.1 for the MH+, MH+2, MH+4, and MH+6 ions, indicating that the conjugate contained three chlorine atoms. The mass and number of chlorine atoms in the product are consistent with the glutathione Sconjugate identified previously as S-(1,2,2-trichlorovinyl)glutathione formed by the reaction of glutathione with tetrachloroethene (4). The enzyme-catalyzed rate of conjugate formation is shown in Table 1. Trichloroethene. Incubation of trichloroethene with glutathione in the presence of rat liver cytosolic protein resulted in the time- and protein-concentration-dependent formation of two conjugates that eluted at 2.9 and 3.8 min and were detected by single-ion monitoring of m/z 402, which is the mass of the reported glutathione S-conjugates that is formed by the reaction of glutathione with trichloroethene [S-(1,2-dichlorovinyl)glutathione and S-(2,2-dichlorovinyl)glutathione] (20, 21). The conjugates were not detected by UV absorbance analysis. The enzyme-catalyzed rate of conjugate was not calculated because no conjugate formation was detected by UV absorbance analysis. 1-Bromo-2,2-difluoroethene. Incubation of 1-bromo-2,2difluoroethene with glutathione in the presence of rat liver cytosolic protein resulted in the time- and proteinconcentration-dependent formation of a glutathione Sconjugate that eluted at 3.1 min and was detected by APCI LC/MS, but not by UV absorbance analysis. The mass spectrum of the conjugate showed a MH+ ion of m/z 450; the mass spectrum showed nearly equal abundances of the MH+ and MH+2 ions, indicating that the product contained one bromine atom. These data indicated that the product was an addition product, likely S-(2-bromo1,1-difluoroethyl)glutathione formed by the reaction of glutathione with 1-bromo-2,2-difluoroethene. The enzymecatalyzed rate of conjugate formation was not calculated because no conjugate formation was detected by UV absorbance analysis.

Dichloroethenes. Incubation of trans-1,2-dichloroethene or 1,1-dichloroethene with glutathione in the presence of purified rat liver MGST1 or rat liver cytosolic protein did not yield glutathione S-conjugates that were detected by UV and APCI LC/MS analyses. Structure-Activity Relationship. The natural logarithms of the specific activities for the conversion of the haloalkenes studied (Figure 1) to glutathione S-conjugates by purified rat liver MGST1 and rat liver cytosolic proteins were compared with the ELUMO values of haloalkenes (Figure 3). The natural logarithms of the specific activities for the purified rat liver MGST1catalyzed reactions were inversely proportional to the calculated ELUMO values, whereas the natural logarithms of the specific activities for the rat liver cytosolic proteincatalyzed reactions were not inversely proportional to the calculated ELUMO values.

Discussion The formation of glutathione S-conjugates of haloalkenes is the first step in the bioactivation of nephrotoxic haloalkenes via the cysteine conjugate β-lyase pathway. Cytosolic glutathione transferases and microsomal glutathione transferase 1 are competent to catalyze the reaction of glutathione with some haloalkenes (5, 1113). Previous studies of the glutathione-dependent metabolism of several of the haloalkenes in this study utilized rat liver microsomal and cytosolic fractions as the catalysts. DePierre and Morgenstern reported that some cytosolic glutathione transferase is associated with rat liver microsomal membranes (22). This may interfere in the interpretation of results in which microsomal fractions, rather than purified MGST1, were used to catalyze the reaction of glutathione with haloalkenes. For example, there are conflicting reports of the glutathione conjugation of trichloroethene and tetrachloroethene: early reports indicated that rates of conjugate formation were higher with microsomal fractions than with cytosolic fractions (4, 23). Subsequent investigations showed, however, that microsomal fractions fail to catalyze the reaction of glutathione with trichloroethene or tetrachloroethene (14, 15). This discrepancy may be due to the

1040

Chem. Res. Toxicol., Vol. 15, No. 8, 2002

contamination of the microsomal fraction with cytosolic enzymes in the earlier reports that were not present in the latter studies. Hence, purified MGST1 was used in this study to avoid possible errors resulting from cytosolic protein contamination. The aim of this investigation was to develop a methodology to predict the competency of glutathione transferases to catalyze the reaction of glutathione with haloalkenes. The hypothesis tested was that the ELUMO of potential haloalkene substrates could be used as a computational parameter to predict the competency of rat liver cytosolic protein and purified rat liver MGST1 to catalyze the reaction of glutathione with haloalkenes. The structure-activity relationship described in this work is based on the frontier molecular orbital theory that states that, apart from electrostatic interactions, the overlap between orbitals favors the reaction between an electron donor and an electron acceptor. Energetically, a high-lying occupied orbital in the donor may overlap with a low-lying empty orbital in the acceptor, leading to a net stabilization (24). The strength of the interaction between the reactants is dependent on the amount of overlap between the HOMO and LUMO, where greater overlap of the orbitals results in stronger interactions. Second, the strength of the interaction between the reactants is dependent on the energy difference between the HOMO and LUMO, where smaller differences in energy between the orbitals result in stronger interactions. ELUMO values were computed for several haloalkenes with semiempirical methodology and are summarized in Figure 1. The size and shape of the orbitals were similar for each of the haloalkenes (Figure 2), and, therefore, these parameters were not useful for developing a structure-activity relationship. The ELUMO for the haloalkenes proved to be a differentiating parameter. Generally, the ELUMO values became more positive as halogens were replaced with protons or when fluorine atoms were substituted with less electronegative halogens. The HOMO energy value for ethanethiolate, a surrogate for glutathione thiolate anion, was -2.5 eV (data not shown). Thus, according to frontier molecular orbital theory, the haloalkenes with ELUMO closer in value to -2.5 eV should react at a greater rate with glutathione than haloalkenes with ELUMO further in value from -2.5 eV. This basis for a structure-activity relationship was tested with hexafluoropropene, 2-(fluoromethoxy)-1,1,3,3,3pentafluoro-1-propene, 1,1,2-trichloro-3,3,3-trifluoro-1propene, 1,1-dibromo-2,2-difloroethene, tetrachloroethene, trichloroethene, 1-bromo-2,2-difluoroethene, trans1,2-dichloroethene, and 1,1-dichloroethene. These haloalkenes range in ELUMO from -1.14 to 0.379 eV. The competency of rat liver cytosolic protein, as a source of cGSTs, and purified MGST1 to catalyze the reaction of glutathione with these haloalkenes is summarized in Table 1. The glutathione-dependent metabolism of hexafluoropropene, 2-(fluoromethoxy)-1,1,3,3,3-pentafluoro-1propene, and 1,1,2-trichloro-3,3,3-trifluoro-1-propene has been studied previously (5, 12, 13). In the present study, the rate of the cytosolic protein-catalyzed reaction of glutathione with each of these haloalkenes was consistent with the published rates. The rate of the purified MGST1catalyzed reaction of glutathione with these haloalkenes has not been previously reported. The rate of formation of the products of the reaction of glutathione with hexafluoropropene was 6.22 ( 0.38 µmol min-1 (mg of MGST1 protein)-1 and 122 ( 1.4 nmol min-1 (mg of

Jolivette and Anders

cytosolic protein)-1. The rate of formation of the products of the reaction of glutathione with 2-(fluoromethoxy)1,1,3,3,3-pentafluoro-1-propene was 5.00 ( 0.51 µmol min-1 (mg of MGST1 protein)-1 and 174 ( 11 nmol min-1 (mg of cytosolic protein)-1. The rate of formation of the products of the reaction of glutathione with 1,1,2-trichloro3,3,3-trifluoro-1-propene was 4.39 ( 0.23 µmol min-1 (mg of MGST1 protein)-1 and 139 ( 2.7 nmol min-1 (mg of cytosolic protein)-1. The glutathione-dependent metabolism of 1,1-dibromo2,2-difluoroethene has not been reported previously. The data generated in this study demonstrated that both rat liver cytosolic proteins and MGST1 were competent to catalyze the reaction of glutathione with 1,1-dibromo-2,2difluoroethene. A single glutathione conjugate was observed for both the cytosolic protein-catalyzed and the MGST1-catalyzed reactions of glutathione with 1,1dibromo-2,2-difluoroethene. APCI LC/MS analysis was consistent with the formation of S-(2,2-dibromo-1,1difluoroethyl)glutathione from 1,1-dibromo-2,2-difluoroethene. The rate of conjugate formation was 1.66 ( 0.27 µmol min-1 (mg of MGST1 protein)-1 and 62.0 ( 1.2 nmol min-1 (mg of cytosolic protein)-1. Glutathione S-conjugates were not detected when MGST1 was incubated with glutathione and trichloroethene, tetrachloroethene, and 1-bromo-2,2-difluoroethene, whereas rat liver cytosolic protein catalyzed the reaction of glutathione with these haloalkenes. Neither MGST1 nor rat liver cytosolic protein catalyzed the reaction of glutathione with trans-1,2-dichloroethene or 1,1-dichloroethene. A linear relationship was observed when the ELUMO values for hexafluoropropene, 2-(fluoromethoxy)-1,1,3,3,3pentafluoro-1-propene, 1,1,2-trichloro-3,3,3-trifluoro-1propene, and 1,1-dibromo-2,2-difluoroethene were compared with the natural logarithms of the specific activities for the glutathione conjugation reaction catalyzed by purified rat liver MGST1 (Figure 3). However, this linear relationship consists of only four data points that do not span a wide range in terms of values of specific activities and thus cannot be extrapolated for values outside of the ELUMO range of -1.14 to -0.73 eV. Based on the data contained herein and published reports of the MGST1catalyzed reaction of glutathione with tetrafluoroethene, chlorotrifluoroethene, and hexachlorobutadiene (25-27), all of which have ELUMO values more negative than -0.59 eV, it appears that haloalkenes with ELUMO values more negative that -0.59 eV are substrates for purified rat liver MGST1 and haloalkenes with ELUMO values more positive than -0.44 eV are not substrates for purified rat liver MGST1. A linear relationship was not observed, however, when the ELUMO values for hexafluoropropene, 2-(fluoromethoxy)1,1,3,3,3-pentafluoro-1-propene, 1,1,2-trichloro-3,3,3-trifluoro-1-propene, 1,1-dibromo-2,2-difluoroethene, and tetrachloroethene were compared with the natural logarithms of the specific activities for the glutathione conjugation reaction catalyzed by rat liver cytosolic protein (Figure 3). The lack of a linear relationship between ELUMO values of these haloalkenes and the specific activities of the cytosolic protein-catalyzed reactions may be attributed to the participation of multiple classes of cytosolic glutathione transferases in the enzyme-catalyzed reactions of glutathione with the different haloalkenes. Additionally, the rank-order of the cytosolic protein specific activities for hexafluoropropene, 2-(fluoromethoxy)-1,1,

Haloalkene Glutathione Conjugation SAR

3,3,3-pentafluoro-1-propene, 1,1,2-trichloro-3,3,3-trifluoro1-propene, and 1,1-dibromo-2,2-difluoroethene was different than that of the MGST1-catalyzed reactions; therefore, the same computed parameter cannot be expected to correlate with both sets of activities. In summary, a linear structure-activity relationship existed for the MGST1-catalyzed reaction of glutathione with hexafluoropropene, 2-(fluoromethoxy)-1,1,3,3,3-pentafluoro-1-propene, 1,1,2-trichloro-3,3,3-trifluoro-1-propene, and 1,1-dibromo-2,2-difluoroethene. Rank-order reactivity may be predicted for MGST1-catalyzed reactions of glutathione with haloalkenes such that haloalkenes with more negative ELUMO values have a greater specific activity for the enzyme-catalyzed glutathione conjugation reaction than haloalkenes with less negative ELUMO values. These results indicate that the chemical reactivity of the substrate plays a vital role in the rate at which the glutathione-dependent biotransformation of the haloalkenes occurs. This finding may be of value in understanding and predicting the route of metabolism of haloalkenes and their associated toxicities. Further experimental and computational (particularly at higher levels of theory) studies to extend the range of haloalkenes are needed to determine whether a predictive model can be developed and validated.

Acknowledgment. This research was supported by National Institutes of Environmental Health Sciences Grant ES03127 (M.W.A.) and National Institutes of Environmental Health Sciences Training Grant ES07026 (L.J.J.). We thank Dr. Ann M. Richard for critical evaluation of this work, and Agilent Technologies (Palo Alto, CA) for the use of an LC/MS during the conduct of this research.

References (1) Bruckner, J. V., Davis, B. D., and Blancata, J. N. (1989) Metabolism, toxicity and carcinogenicity of trichloroethylene. Crit. Rev. Toxicol. 20, 31-50. (2) Davidson, I., and Beliles, W. F. (1991) Consideration of the target organ toxicity of trichloroethylene in terms of metabolite toxicity and pharmacokinetics. Drug Metab. Rev. 23, 493-599. (3) DeKant, W., Metzler, M., and Henschler, D. (1986) Identification of S-1,2,2-trichlorovinyl-N-acetylcysteine as a urinary metabolite of tetrachoroethylene: bioactivation through glutathione conjugation as a possible explanation of its nephrocarcinogenity. J. Biochem. Toxicol. 1, 57-72. (4) DeKant, W., Martens, G., Vamvakas, S., Metzler, M., and Henschler, D. (1987) Bioactivation of tetrachloroethylene. Role of glutathione S-transferase catalyzed conjugation versus cytochrome P-450 dependent phospholipid alkylation. Drug Metab. Dispos. 15, 702-709. (5) Vamvakas, S., Kremling, E., and DeKant, W. (1989) Metabolic activation of the nephrotoxic haloalkene 1,1,2-trichloro-3,3,3trifluoro-1-propene by glutathione conjugation. Biochem. Pharmacol. 38, 2297-2304. (6) Potter, C. L., Gandolfi, A. J., Nagle, R., and Clayton, J. W. (1981) Effects of inhaled chlorotrifluoroethylene and hexafluoropropene on the rat kidney. Toxicol. Appl. Pharmacol. 59, 431-440. (7) Kandel, L., Laster, M. J., Eger, E. I., II, Kerschmann, R. L., and Martin, J. (1995) Nephrotoxicity in rats undergoing a one-hour exposure to Compound A. Anesth. Analg. 81, 559-563. (8) Gonsowski, C. T., Laster, M. J., Eger, E. I., II, Ferrell, L. D., and Kerschmann, R. L. (1994) Toxicity of Compound A in rats. Effects of increasing duration of administration. Anesthesiology 81, 566573.

Chem. Res. Toxicol., Vol. 15, No. 8, 2002 1041 (9) Gonsowski, C. T., Laster, M. J., Eger, E. I., II, Ferrell, L. D., and Kerschmann, R. L. (1994) Toxicity of Compound A in rats. Effects of a 3-hour administration. Anesthesiology 80, 556-565. (10) Anders, M. W., and DeKant, W. (1998) Glutathione-dependent bioactivation of haloalkenes. Annu. Rev. Pharmacol. Toxicol. 38, 501-537. (11) Dohn, D. R., Quebbemann, A. J., Borch, R. F., and Anders, M. W. (1985) Enzymatic reaction of chlorotrifluoroethene with glutathione: 19F NMR evidence for stereochemical control of the reaction. Biochemistry 24, 5137-5143. (12) Jin, L., Davis, M. R., Kharasch, E. D., Doss, G. A., and Baillie, T. A. (1996) Identification in rat bile of glutathione conjugates of fluoromethyl 2,2-difluoro-1-(trifluoromethyl)vinyl ether, a nephrotoxic degradate of the anesthetic agent sevoflurane. Chem. Res. Toxicol. 9, 555-561. (13) Koob, M., and DeKant, W. (1990) Metabolism of hexafluoropropene. Evidence for bioactivation by glutathione conjugate formation in the kidney. Drug Metab. Dispos. 18, 911-916. (14) Lash, L. H., Qian, W., Putt, D. A., Jacobs, K., Elfarra, A. A., Krause, R. J., and Parker, J. C. (1998) Glutathione conjugation of trichloroethylene in rats and mice: sex-, species-, and tissuedependent differences. Drug Metab. Dispos. 26, 12-19. (15) DeKant, W., Birner, G., Werner, M., and Parker, J. (1998) Glutathione conjugation of perchloroethene in subcellular fractions from rodent and human liver and kidney. Chem.-Biol. Interact. 116, 31-43. (16) Rietjens, I. M., Soffers, A. E., Hooiveld, G. J., Veeger, C., and Vervoort, J. (1995) Quantitative structure-activity relationships based on computer calculated parameters for the overall rate of glutathione S-transferase catalyzed conjugation of a series of fluoronitrobenzenes. Chem. Res. Toxicol. 8, 481-488. (17) Soffers, A. E., Ploemen, J. H., Moonen, M. J., Wobbes, T., van Ommen, B., Vervoort, J., van Bladeren, P. J., and Rietjens, I. M. (1996) Regioselectivity and quantitative structure-activity relationships for the conjugation of a series of fluoronitrobenzenes by purified glutathione S-transferase enzymes from rat and man. Chem. Res. Toxicol. 9, 638-646. (18) Weinander, R., Mosialou, E., DeJong, J., Tu, C.-P. D., Dypbukt, J., Bergman, T., Barnes, H. J., Hoog, J.-O., and Morgenstern, R. (1995) Heterologous expression of rat liver microsomal glutathione transferase in simian COS cells and Escherichia coli. Biochem. J. 311, 861-866. (19) Iyer, R. A., Baggs, R. B., and Anders, M. W. (1997) Nephrotoxicity of the glutathione and cysteine S-conjugates of the sevoflurane degradation product 2-(fluoromethoxy)-1,1,3,3,3-pentafluoro-1propene (Compound A) in male Fischer 344 rats. J. Pharmacol. Exp. Ther. 283, 1544-1551. (20) Commandeur, J. N., and Vermeulen, N. P. (1990) Identification of N-acetyl(2,2-dichlorovinyl)- and N-acetyl(1,2-dichlorovinyl)-Lcysteine as two regioisomeric mercapturic acids of trichloroethylene in the rat. Chem. Res. Toxicol. 3, 212-218. (21) Lash, L. H., Xu, Y., Elfarra, A. A., Duescher, R. J., and Parker, J. C. (1995) Glutathione-dependent metabolism of trichloroethylene in isolated liver and kidney cells of rats and its role in mitochondrial and cellular toxicity. Drug Metab. Dispos. 23, 846853. (22) DePierre, J. W., and Morgenstern, R. (1983) Comparison of the distribution of microsomal and cytosolic glutathione S-transferase activities in different organs of the rat. Biochem. Pharmacol. 32, 721-723. (23) DeKant, W., Koob, M., and Henschler, D. (1989) Metabolism of trichloroethenesin vivo and in vitro evidence for activation by glutathione conjugation. Chem.-Biol. Interact. 73, 89-101. (24) Fukui, K. (1971) Recognition of stereochemical paths by orbital interaction. Acc. Chem. Res. 4, 57-64. (25) Odum, J., and Green, T. (1984) The metabolism and nephrotoxicity of tetrafluoroethylene in the rat. Toxicol. Appl. Pharmacol. 76, 306-318. (26) Hargus, S. J., Fitzsimmons, M. E., Aniya, Y., and Anders, M. W. (1991) Stereochemistry of the microsomal glutathione S-transferase catalyzed addition of glutathione to chlorotrifluoroethene. Biochemistry 30, 717-721. (27) Wallin, A., Gerdes, R. G., Morgenstern, R., Jones, T. W., and Ormstad, K. (1988) Features of microsomal and cytosolic glutathione conjugation of hexachlorobutadiene in rat liver. Chem.Biol. Interact. 68, 1-11.

TX0255222