Chem. Res. Toxicol. 1996, 9, 667-675
667
Conformational Aspects of Glutathione Conjugates of Chlorinated Alkenes: A Computational Study Joong-Youn Shim† and Ann M. Richard* National Health and Environmental Effects Research Laboratory, U.S. Environmental Protection Agency, Research Triangle Park, North Carolina 27711 Received November 27, 1995X
The nephrotoxicity of halogenated alkenes is due to the β-lyase mediated bioactivation of the hepatic glutathione (GS) conjugate to mutagenic or cytotoxic reactive species in kidney. Experimental evidence obtained for regioisomers and geometric isomers of haloalkene GS conjugates indicates that different isomers may be metabolized and excreted at different rates, follow different metabolic pathways, and exhibit different toxicities. Computational methods were applied in the present work to a conformational study of GS-haloalkene conjugates to determine the relative stabilities of possible regioisomers and geometric isomers of the conjugates. The halogenated alkenes studied were 1,1,2-trichloroethylene (TCE), hexachloro1,3-butadiene (HCBD), and 1,1,2-trichloro-3,3,3-trifluoro-1-propene (TCTFP). Calculated energies of GS conjugate products were used to approximately infer relative product abundance under synthetic and in vivo conditions. This approach neglects differential solvent effects and enzyme selectivity and assumes a late transition state for GS conjugation and/or some thermodynamic control of the conjugation process. Relative population predictions of GS conjugate isomers, based on computed energies, were in agreement with experimental synthetic and in vivo isomer determinations in the case of TCE, where careful analytical characterization of the isomers was definitive. In the case of HCBD, where analytical determinations were not performed and isomer assignments were based on general reactivity concepts, calculations from the present study supported one GS conjugate isomer assignment and disagreed with the other. Finally, in the case of TCTFP, the calculations predicted that three isomers would have similar populations, whereas only two were detected in the experimental study.
Introduction Halogenated alkenes, such as 1,1,2-trichloroethylene (TCE),1 tetrachloroethylene (PERC), and hexachloro-1,3butadiene (HCBD), are high volume industrial solvents and ubiquitous environmental pollutants. Concern for adverse health effects due to widespread exposure to these chemicals has been fueled by evidence that they induce carcinomas in the proximal tubules of rodent kidney (1-3). Several halogenated alkenes, including TCE, HCBD, 1,1,2-trichloro-3,3,3-trifluoro-1-propene (TCTFP), 1,1,2,2-tetrafluoroethene, hexafluoro-1-propene, 1-chloro-1,2,2-trifluoroethene, and 1,1-dichloro-2,2-difluoroethene, are also selectively nephrotoxic and induce proximal tubular damage in kidneys (4), suggesting a relationship between the mechanisms of nephrotoxicity and nephrocarcinogenicity. Although microsomal cytochrome P-450 oxidative metabolism is considered a primary metabolic pathway for many of these chemicals * Corresponding author, at MD-68, US-EPA, Research Triangle Park, NC 27711. Phone: (919) 541-3934; Fax: (919) 541-0694; Email:
[email protected]. † Postdoctoral fellow in the Curriculum of Toxicology, UNC-Chapel Hill, Chapel Hill, NC. X Abstract published in Advance ACS Abstracts, March 15, 1996. 1 Abbreviations: TCE, 1,1,2-trichloroethylene; PERC, tetrachloroethylene; HCBD, hexachloro-1,3-butadiene; TCTFP, 1,1,2-trichloro3,3,3-trifluoro-1-propene; GS, thiolate or bound form of glutathione; GSH, glutathione; GST, glutathione S-transferase; 1,2-DCVG, S-(1,2dichlorovinyl)glutathione; 2,2-DCVG, S-(2,2-dichlorovinyl)glutathione; 1,2-DCVC, S-(1,2-dichlorovinyl)-L-cysteine; 2,2-DCVC, S-(2,2-dichlorovinyl)-L-cysteine; 1,2-NAc-DCVC, N-acetyl-S-(1,2-dichlorovinyl)-Lcysteine; 2,2-NAc-DCVC, N-acetyl-S-(2,2-dichlorovinyl)-L-cysteine; LUMO, lowest unoccupied molecular orbital; PCBG, S-(pentachloro1,3-butadienyl)glutathione; PCMTB, pentachloro-1-(methylthio)-1,3butadiene.
0893-228x/96/2709-0667$12.00/0
(5-7), experimental evidence supports an alternative mechanism for the selective nephrotoxicity and nephrocarcinogenicity involving a glutathione (GS) activation pathway (8-11). GS conjugation, which normally acts as a metabolic detoxification pathway in mammals, is known to occasionally lead to toxicity through production of reactive thiols in the liver or kidneys. The proposed mechanism for GS activation of haloalkenes involves GS conjugation in the liver, enzymatic processing by renal peptidases to yield the corresponding cysteine conjugates, and subsequent cysteine conjugate β-lyase mediated bioactivation to reactive species in the proximal tubules of the kidneys (see Figure 1). The cysteine conjugates can be direct acting or can undergo further metabolic transformation by β-lyase to produce the ultimate mutagenic and cytotoxic reactive species in kidney (12, 13). Evidence to support this mechanism for haloalkenes has been provided by experiments that have detected cysteine conjugates and mercapturic acid metabolites produced by in vivo exposure to haloalkenes, and by direct toxicity and mutagenicity testing of these metabolites, in the presence and absence of β-lyase activation (14). The formation of the GS-haloalkene conjugates in biological systems is catalyzed by glutathione S-transferase (GST) and proceeds via an addition/elimination mechanism (15). As shown in Figure 2, GS conjugates of halogenated alkenes can exist as regioisomers, e.g., the cis- and trans-S-(1,2-dichlorovinyl)glutathione (1,2-DCVG) conjugates, and as geometric (or structural) isomers, e.g., the S-(1,2-dichlorovinyl)glutathione and the S-(2,2-dichlorovinyl)glutathione (2,2-DCVG) conjugates. Whether or not a mixture of isomers is produced by a synthetic pathway or under in vivo biological conditions © 1996 American Chemical Society
668 Chem. Res. Toxicol., Vol. 9, No. 3, 1996
Shim and Richard
Figure 1. Glutathione conjugation pathway for halogenated alkenes; an example of 1,2-dichlorovinyl conjugation isomers, where R ) -H [TCE], -C(Cl)dC(Cl)2 [HCBD], and -CF3 [TCTFP].
Figure 2. Possible conjugation products from GS and various halogenated alkenes catalyzed by GST, where Gly ) glycine, Glu ) glutamate, and R ) -H [TCE], -C(Cl)dC(Cl)2 [HCBD], and -CF3 [TCTFP].
depends upon various factors, including the nature of the synthetic pathway employed or the regioselectivity of the GST catalyzed conjugation reaction. Whether or not such isomers are experimentally detected and properly identified depends on factors such as the availability and correct identification of GS conjugate chemical standards for use in analytical determinations of in vivo metabolites. Interest in the accurate assignment of GS conjugate isomers of haloalkenes is fueled by experimental evidence indicating that different GS conjugate isomers are metabolized and excreted at different rates, follow different metabolic pathways, and exhibit different toxicities (16, 17). In investigations to ascertain mechanisms or identify metabolic products, incorrect assumptions concerning predominant or minor isomer forms could lead to erroneous interpretations of the data (18). A general characteristic of the GSTs is their ability to bind to and form GS conjugates with an extremely broad range of electrophilic chemical substrates (19). There is some evidence in the literature, however, for GST regioselectivity in GS conjugation of haloalkenes. In a study of chlorotrifluroethene (17, 20), GST selectivity resulting in preferential formation of one conjugate diastereomer was found using the microsomal GST fraction. In contrast, an equimolar mixture of conjugate diastereomers was found using cytosolic GST, which had a slightly lower activity than microsomal GST. Although the authors enumerated a variety of possible factors that could produce stereoselectivity in GS conjugation, including specific binding in the GST active site or directional protonation of the carbanionic intermediate while bound to GST, the mechanistic and structural requirements for such an outcome remain unknown.
For the purposes of this study, it is assumed that the initial GS conjugation product isomer ratios are maintained in the corresponding cysteine conjugates or mercapturic acids detected in experimental studies of in vivo metabolites. This first level of approximation assumes that the GS conjugation step is rate limiting and neglects isomer selectivity known to occur to some degree in β-lyase activation and N-acetylation (16, 18, 21). In addition, in the absence of specific knowledge of the enzyme interaction, a basic working assumption is that relative free energies of products, i.e., the GS conjugate isomers of the haloalkenes, are relevant and predictive of the outcome of the conjugation process in vivo or under synthetic conditions. A major, recognized limitation of this approach is that it neglects possible stereoselectivity of the GS conjugation process and differential solvent effects on isomer formation in the in vivo situation. However, the working assumption is supported by data that suggest a late transition state for GS conjugation (19), which, according to the Hammond postulate, implies that reaction kinetics are influenced by product state stability (22). In addition, product energies are relevant to the thermodynamics of the GS conjugation process, particularly under synthetic conditions. Hence, calculated geometries and relative energies of the haloalkeneGS conjugates provide fundamental data that can potentially complement analytical methods and heuristic arguments used in the identification of isomers produced under a variety of experimental conditions. To assess the relative stabilities of various GS conjugate isomers in the present study, computational methods were applied to a search for stable conformations of
Conformations of Haloalkene Glutathione Conjugates
Chem. Res. Toxicol., Vol. 9, No. 3, 1996 669
Figure 3. Structures of TCE, HCBD, and TCTFP.
the GS conjugates of the halogenated alkenes, TCE, HCBD, and TCTFP (see Figure 3), based on energy considerations. Upon inspection, each identified conformation was grouped into one of the three major isomer classes, i.e., a 1,2-DCVG isomer (cis or trans) or a 2,2DCVG isomer. Energies of individual conformers, and Boltzmann averages over groups of conformers for particular isomers, were then used to infer relative probabilities of forming regioisomers or geometric isomers under various experimental conditions. These findings are discussed in the context of some published experimental investigations.
Methods Systematic conformational searches, with incremental sampling of angles and dihedrals, were performed using the SYBYL molecular mechanics force field (23) for each of the three studied haloalkenes, TCE, HCBD, and TCTFP, and the semiempirical AM1 (24) method only for the smallest haloalkene, TCE, for comparison purposes. The conformational searches were carried out within the SPARTAN molecular modeling program (Wavefunction, Inc., version 3.1) residing on a Silicon Graphics Indigo R4400XZ workstation. Geometries of all studied molecules were fully optimized using the semiempirical AM1 method within SPARTAN. In addition, fully optimized geometries and corresponding energies were calculated at the ab initio RHF/6-31G* level for the unconjugated haloalkenes, and at the MP2/6-31G* level for selected isomers of TCTFP to verify the lowest energy conformation. Single point energies of the haloalkene conjugates were calculated for AM1 optimized geometries at the RHF/ STO-3G and 6-31G* basis set levels within SPARTAN or GAUSSIAN 92 (25). All computations involving the use of GAUSSIAN 92 were carried out on a Cray Y-MP located at the EPA-National Environmental Supercomputer Center in Bay City, MI. The initial geometry of the free, protonated form of glutathione (GSH) was extracted from the X-ray structure of GS complexed with the mu class of GSTs (1GST.PDB) from the Brookhaven Protein Data Base (PDB) and was subjected to semiempirical AM1 geometry optimization and subsequent ab initio geometry optimization. The required geometries of the GS-haloalkene conjugates were constructed starting from the AM1 optimized geometry of the free GSH. Both AM1 and SYBYL conformational searches were carried out within SPARTAN for each positional isomer of the TCE conjugates, with all 12 torsional angles of the backbone of GS constrained to the conformation obtained from AM1 optimization of the X-ray structure. A SYBYL conformational search within SPARTAN was carried out for each positional isomer of the HCBD and TCTFP conjugates, with all 12 torsional angles of the backbone of GS constrained in the same manner as for TCE. All conformations found from the conformational search having greater than 1% of the Boltzmann population (at 25 °C), according to the SPARTAN analysis, were subjected to AM1 geometry optimization without backbone constraints. The AM1 optimized structures were examined to identify duplicate conformers, followed by low level ab initio RHF/STO-3G single point calculations of all structurally unique conformations to confirm relative energy information and to identify potential low energy conformers whose AM1 energy might have been overestimated. Both AM1 geometry-optimized and RHF/STO-3G single point
energies, obtained at reasonable computational cost, were used to identify low energy isomers for single point RHF energy evaluation using the more computationally intensive 6-31G* basis set, which includes polarization functions for more accurate treatment of the electron density of the chlorine atoms. A Boltzmann distribution of the RHF/6-31G* energy conformers at physiological temperature, 37 °C, was used to compute a Boltzmann average energy of each major isomer form of the GS conjugates of TCE, HCBD, and TCTFP by weighting each conformer energy according to relative population.
Results Selected dihedral angles for the glutathione portion of GS conjugates of diverse substrates are defined and listed in Table 1. Comparison is made between the X-ray crystallographic GS conjugate geometries corresponding to a variety of substrates, and the calculated GS conjugate geometries corresponding to the halogenated alkenes. From the 1GST.PDB file of the Brookhaven Protein Data Bank, the dihedral angle ω1(N-CR-Cβ-S) of the free form of GS bound in the active site of mu class GST is measured as -90.3°, where one of the Cβ hydrogens of the cysteinyl residue is oriented toward the oxygen of the carbonyl group of the cysteinyl residue to maximize H-bonding. This dihedral angle, obtained from X-ray structures, ranges from -79.6° to -31.9° for the various electrophilic, non-haloalkene substrates listed in Table 1, which indicates possible constraints on the orientation of the substrate undergoing GS conjugation in the presence of the GST. Selected GS conjugates of the haloalkenes, having the lowest energies according to RHF/6-31G* single point energy evaluation of AM1 optimized geometries, show similar ranges of ω1 dihedral angles (-74.7° to -53.4°), with the exception of a structure whose GS backbone differs significantly from the X-ray determined GS backbone structures. Conservation of the GS backbone structure for the GS-haloalkene conjugates is illustrated in Figure 4. Geometry optimized energies of GSH and the studied haloalkenes at the AM1 and RHF/6-31G* calculation levels are presented in Table 2. Calculations indicate that the most stable conformation of HCBD is with a torsional angle ω(CdCsC)C) of about 90° (95.6° and 88.5° from the AM1 and RHF/6-31G* calculations, respectively), where the steric repulsion between chlorines is minimized. AM1 also found planar cis and trans conformers (not shown) whose energies were 10 kcal/mol higher than the lowest energy nonplanar conformer. These planar conformers were considered suspect, however, since they were not identified by the ab initio RHF/ 6-31G* method, and since AM1 has known difficulty optimizing from planar starting geometries. As shown in Table 3, the geometry of the nonplanar HCBD conformer obtained from the RHF/6-31G* calculation is in good agreement with the geometry obtained by electron diffraction experiment (26). While AM1 identified only a single conformer of TCTFP, two stable conformational isomers were identi-
670 Chem. Res. Toxicol., Vol. 9, No. 3, 1996
Shim and Richard
Table 1. Selected Dihedral Angles for Glutathione Conjugates: X-Ray Crystal Structures and Calculated Geometries
GS Conjugate 10-hydroxy-9,10-dihydrophenanthrenea 2,4,6-trinitrocyclohexadienec 2,4-dinitrobenzened p-nitrobenzenee hexanef haloalkenesg 1-GS TCE CFZ 3 1-GS TCE CFE 3 2-GS TCE CF 2h 1-GS HCBD CFZ 3 1-GS HCBD CFE 3 2-GS HCBD CF 10 1-GS TFTCP CFZ 7 1-GS TFTCP CFE 8 2-GS TFTCP CF 5
ω1(N-CR-Cβ-S)
ω2(CR-Cβ-S-C)
-55.7 -65.4b -56.2 -56.5 -59.6 -79.6 -60.9 -66.0 -61.7 -31.9
-82.3 -78.6 -94.2 -97.9 -55.3 -44.3 -90.0 -88.1 -115.4 -129.1
-71.9 -74.7 29.1 -55.4 -53.4 -62.6 -58.4 -54.2 -57.5
164.0 165.3 86.7 172.7 -68.2 151.8 -77.4 177.1 137.4
ω3(Cβ-S-CdC)
162.0 -79.4 178.2 -154.7 158.8 72.5 173.8 -171.1 107.4
a From 3GST. PDB of the Brookhaven Protein Data Bank (PDB). b GS conjugate bound to second subunit of GST in slightly different conformation. c From 4GST. PDB of the Brookhaven PDB. d From 5GST. PDB of the Brookhaven PDB. e From 1GLQ. PDB of the Brookhaven PDB. f From 2GLR. PDB of the Brookhaven PDB. g From the present study, AM1 GO, RHF/6-31G* SP minimum energy geometries (see text): 1-GS-...-CFZ ) 1,2-cis, 1-GS-...-CFE ) 1,2-trans, and 2-GS-... ) 2,2 isomer. h The backbone structure was significantly different from the X-ray determined backbone structures of the GS conjugates.
Table 2. Calculated Energies of Geometry Optimized Free Form of Glutathione (GSH) and Halogenated Alkenes E(AM1) GOa
E(RHF/6-31G*) GOa
E(MP2/6-31G*) GOa
GSH -268.398 -877671.234 TCE -8.406 -912850.088 HCBD -7.528 -1824968.213 TCTFP CONF 1 -155.610b -1123443.715 (69.2)c -1124252.416 (91.9) -1123443.259 (30.8) -1124250.905 (8.1) CONF 2 a Computed energies (in kcal/mol) of fully geometry optimized (GO) structures at indicated level of approximation. b AM1 was able to identify only a single conformer of TCTFP, CONF 1, even when the RHF/6-31G* optimized geometry for CONF 2 was used as the starting geometry for the AM1 optimization. c Numbers in parentheses (in %) represent Boltzmann averaged population of isomer as a fraction of total population at 37 °C.
Figure 4. Superposition of the of most stable conformations of glutathione conjugates of TCE, HCBD, and TCTFP, with the left side of the figure showing the conserved glutathione region of the conjugates.
fied with ab initio methods: one form (CONF 1) with the chlorine at the central carbon atom eclipsed with one of the terminal fluorides of the CF3 group, and the other form (CONF 2) with the same chlorine trans to one of the terminal fluorines of the CF3 group. RHF/6-31G* calculations found CONF 1 favored by less than 0.5 kcal/ mol over the CONF 2, whereas geometry optimization
at the higher MP2/6-31G* level predicted the CONF 1 isomer to be 1.5 kcal/mol more stable than the CONF 2 isomer. Due to its lower energy, only CONF_1 GS conjugates were considered in the latter part of this study. Interestingly, an electron diffraction study of hexafluoropropene (27) reported predominant existence of a conformer similar to CONF 2 of TCTFP, i.e., with the fluorine at the central carbon atom trans to a fluorine atom in the CF3 group. Consistent with these experimental findings, and in contrast to the results for TCTFP, computed energies of the two conformers of hexafluoropropene fully optimized at the RHF/6-31G* level indicate that the CONF 2 isomer is more stable than CONF 1 by 1.627 kcal/mol. Hence, due to the larger bulk of the central halogen in TCTFP versus hexafluoropropene, a chlorine rather than a fluorine, calculations indicate a shift in favor of a staggered conformation for TCTFP,
Conformations of Haloalkene Glutathione Conjugates Table 3. Molecular Structure of HCBD
electron diffractiona c
r(C1-Cl)avg r(C2-Cl) r(CdC)avg r(CsC) ∠(CdCsC)avge ∠(CsCsCl)avg ∠(CdCsCl)avg ω(CdC-CdC)
1.716(3)d 1.716(3) 1.341(5) 1.485(9) 122.6(4) 115.8(4) 122.50(10) 89(3)
RHF/6-31G* GOb 1.7178 1.7298 1.3208 1.4836 123.406 114.643 122.830 88.5
a Results obtained from Gundersen et al. (26) electron diffraction study at 373 K; bond lengths obtained directly from P(r)/r radial distribution function. b Results obtained from GAUSSIAN 92 fully geometry optimized (GO) structure computed at the RHF/6-31G* basis set level. c Bond lengths in angstroms, Å. d Experimental error in last significant figure indicated in parentheses. e Angles in degrees.
with the central chlorine trans to a terminal fluorine, i.e., CONF 2. The conformational search results were tabulated for all identified unique conformations of TCE, HCBD, and TCTFP conjugates, where uniqueness was determined by visual inspection of structures. A listing of selected dihedrals for the individual conformers, energies of the AM1 geometry optimized structures, ab initio RHF/STO3G single point energies for all conformers, and RHF/631G* single point energies for selected conformers is available (see Supporting Information). A conformational search was carried out for the smallest of the three haloalkene conjugates, GS TCE, using both the semiempirical AM1 and the SYBYL molecular mechanics force field options within SPARTAN, with very similar results. Although SYBYL identified a greater number of conformers initially, upon closer structural examination and AM1 geometry optimization, many were found to be structurally non-unique. Due to the significantly greater computational effort associated with the AM1 conformational search, only the SYBYL conformational search was performed for HCBD and TCTFP, with subsequent AM1 geometry optimization of all SYBYL-identified conformers. AM1 geometry optimizations were relied upon to provide reasonably accurate geometries, while ab initio methods were assumed to provide more accurate indications of the relative energies of the isomers (28). For this purpose, RHF/STO-3G level calculations were used as a rough screen for identifying the lowest energy conformers from among the full conformational analysis results, while higher level RHF/6-31G* calculations were only performed on this reduced set of conformers. A number of general observations concerning the conformational results were made. For 7 out of the 9 isomers considered, the lowest energy was assigned to the same conformer by both STO-3G and 6-31G* basis set level calculations. In the remaining 2 cases, the 6-31G* energy of the lowest energy conformer identified by STO-3G was within 0.11 and 1.5 kcal/mol of the lowest energy conformer identified by 6-31G*. Also, when measured by AM1 energies, in all cases the AM1 minimum energy conformer was within 1 kcal/mol of the minimum energy conformer identified by either STO-3G
Chem. Res. Toxicol., Vol. 9, No. 3, 1996 671
or 6-31G* basis set calculations. This indicated consistency between the various methods for identifying low energy conformers, providing support for our use of the RHF/6-31G* results to infer relative isomer energies. Boltzmann average energies of each isomeric form of the GS conjugated halogenated alkenes at physiological temperature, 37 °C, are presented in Table 4. These values represent estimates of the average energy of the major isomeric forms when individual conformer energies are weighted by their Boltzmann populations. In addition, scaled Boltzmann averaged energies, with respect to a zero point energy, are compared in the last column of this table to the scaled individual minimum conformer energies obtained in earlier results (see Supporting Information). The Boltzmann weighting is seen to have little effect on the relative energies of the various isomers.
Discussion TCE has been the focus of intensive toxicological investigation, with a large body of work centered on study of the cytotoxic and mutagenic GS conjugate, DCVG, and its metabolite, S-(dichlorovinyl)-L-cysteine (DCVC) (29, 30). Investigators have used synthesized chemical standards of DCVG, DCVC, and related derivatives for the analytical detection, monitoring, and characterization of these metabolites in in vivo studies. For the generation of DCVG and DCVC standards, most have used the synthetic pathway of McKinney et al. (31). This procedure involves base-catalyzed transdehydrohalogenation through a symmetrical 1,2-dichloroacetylene intermediate, followed by nucleophilic addition of the GS thiolate anion, and is expected to result in selective formation of the trans-1,2-DCVG isomer as shown in paths “a” and “b” in Figure 5 (18). However, if the reaction proceeds under mildly basic conditions closer to the in vivo state, where TCE remains in the neutral form, a thiolate anion directly attacks TCE rather than an acetylene intermediate. Under these conditions, mixtures of trans/cis-1,2DCVG and 2,2-DCVG isomers are formed in proportions that depend upon product energies, differences in electrophilicity of the two carbon centers of TCE, and/or differences in steric and electronic interactions between the approaching thiolate anion and TCE, as shown in paths “c” and “d” in Figure 5 (15). Commandeur and Vermeulen (18) reported synthesis of the mercapturic acid, N-acetyl-S-(2,2-dichlorovinyl)-L-cysteine (2,2-NAcDCVC), under the latter experimental conditions. When the McKinney method is employed in the synthesis of glutathione conjugates of HCBD or TCTFP, where no hydrogens are available for base-catalyzed dehydrohalogenation to an acetylenic intermediate, the situation is analogous to the weak base condition described above. Hence, for both HCBD and TCTFP, the synthesis would be expected to produce a mixture of regioisomers (cis and trans) and geometric isomers (1,1 and 1,2). Calculated geometries and relative energies of the haloalkene-GS conjugates complement analytical methods and heuristic arguments used in the identification of such isomers. GS-TCE Conjugation. Dekant et al. (30) reported identification of DCVG and its corresponding mercapturic acid, NAc-DCVC, as urinary metabolites of TCE, and as evidence of a GS detoxification pathway. Identification of NAc-DCVC was made by chromatographic and mass spectral comparisons to a NAc-DCVC reference standard, synthesized according to the McKinney method described
672 Chem. Res. Toxicol., Vol. 9, No. 3, 1996
Shim and Richard
Table 4. Calculated Boltzmann Averaged Energies (in kcal/mol) and Relative Energies of GS Conjugate Isomers for Halogenated Alkenes at 37 °C
TCE
1-conjugate
trans cis
2-conjugate HCBD
1-conjugate
trans cis
2-conjugate TCTFP
1-conjugate 2-conjugate
trans cis
AM1 GOa EBoltzmannb
RHF/6-31G* SPa EBoltzmannb
RHF/6-31G* SP scaled EBoltzmannc
RHF/6-31G* SP scaled Emind
-266.356 -266.034 -268.715
-1501819.451 -1501817.294 -1501819.910
0.459 2.616 0.000
0.330 2.700 0.000
-266.131 -267.291 -265.869
-2413933.465 -2413937.114 -2413934.386
3.649 0.000 2.728
3.872 0.000 2.702
-417.369 -413.576 -414.916
-1712408.179 -1712407.983 -1712408.161
0.000 0.196 0.018
0.000 0.132 0.089
a
Level of approximation for calculation of energy: AM1 energy evaluation of AM1 geometry optimized (GO) structures, RHF/6-31G* single point (SP) energy evaluation of AM1 GO structures. b EBoltzmann is the average energy computed by weighting each conformer energy by the Boltzmann population fraction at 37 °C. c Energies scaled according to minimum Boltzmann averaged energy conformer in each haloalkene group. d Scaled energies of the un-Boltzmann averaged individual minimum energy conformers from the RHF/6-31G* SP results (see Supporting Information).
Figure 5. Formation of various isomers of DCVG from TCE and the thiolate form of glutathione.
above. Using HPLC separation procedures, these investigators were able to resolve the NAc-DCVC peak into two peaks, with relative intensities of 2:1. Arguments based on observed fragmentation pattern and reaction mechanisms were provided to support assignment of the minor peak to the trans-1,2 (E) isomer of NAc-DCVC, while the major peak was assumed to correspond to the cis-1,2 (Z) isomer of NAc-DCVC. These peak assignments were called into question by later published work of Commandeur and Vermeulen (18), who observed that the synthesis of NAc-DCVC in DMSO and under physiological weak base conditions (through a thiolate rather than an acetylene intermediate) yielded 2,2-NAc-DCVC as a major product (90%). Using chemical standards for trans-1,2-NAc-DCVC, synthesized by the McKinney method, and 2,2-NAc-DCVC, synthesized under weak base conditions, Commandeur and Vermeulen (18) assigned the two in vivo metabolites of TCE, that had been previously identified as the trans- and cis-1,2-isomers by Dekant et al. (30), as the trans-1,2- and 2,2-isomers, respectively. In a later publication, Dekant and coworkers (32) revised their original assignments to agree with those of Commandeur and Vermeulen (18). The present calculations indicate that the 2,2-DCVG isomer is the most stable of the three TCE conjugate
isomers, by comparison of individual conformer energies (see Table S1 in Supporting Information) and by comparison of Boltzmann average energies in Table 4. This result is in agreement with experiment, where calculated relative abundances of DCVG are assumed to apply to the subsequently produced mercapturic acids. AM1 energies in Table 4 indicate that the 2,2-isomer is nearly 2.5 kcal/mol more stable than either the cis or trans 1,2isomer and that the latter two isomers are very close in energy, with the trans isomer slightly more stable than the cis isomer. The RHF/6-31G* ab initio energies also find the 2,2-DCVG isomer to be the most stable of the three isomers, slightly more stable than the trans-1,2DCVG isomer, and >2.5 kcal/mol more stable than the cis-1,2-isomer. Hence, the predicted product ratio of 2,2DCVG and trans-1,2-DCVG formation at the higher level of theory is consistent with the observations of Commandeur and Vermeulen (18), who reported 2,2-DCVG as the major product and trans-1,2-DCVG as the minor product of in vivo metabolism. Given the large barrier to interconversion of the cis- and trans-1,2-forms (estimated at >100 kcal/mol), negligible thermal rearrangement between these two isomers is expected. Hence, relative populations of the three isomers will be determined, not according to Boltzmann statistics, but by differences in
Conformations of Haloalkene Glutathione Conjugates
rates of product formation for each isomer. Since the reactant, TCE, is the same for each of the three GS conjugate reaction pathways leading to the three isomer products, product energy is assumed to be a factor in determining relative product formation rate in the present study. Preference for formation of the 2,2-DCVG conjugate isomer is also supported by consideration of frontier molecular orbital densities for the reactant, TCE. An approximate measure of the differential electrophilic reactivity potential of C1 and C2 positions in TCE is provided by the virtual relative density of the lowest unoccupied molecular orbital (LUMO) in the regions of these two carbons. Inspection of the 3D graphical representation of the LUMO density plotted on the charge density surface in SPARTAN (data not shown), computed for the RHF/6-31G* fully optimized structure of TCE, indicated significantly greater LUMO density at the C2 position (carbon with attached hydrogen), consistent with GS attack at that position to form the 2,2DCVG product. GS-HCBD Conjugation. Nash et al. (33) observed three bands from column chromatography of rat biliary metabolites after HCBD administration. From the major band (55% of the total radioactivity), they identified HCBD-GS conjugation products by comparison to the 1-S(pentachloro-1,3-butadienyl)glutathione (1-PCBG) standard, synthesized by the method of McKinney mentioned above. Wolf et al. (34) observed two radiolabeled isotopic mass peaks from the HCBD-GS metabolites, with relative intensities of 9:1, whose mixture gave an identical mass spectrum to the synthetic PCBG conjugate(s) obtained by the McKinney method. The peaks were assigned to the cis and trans 1-positional isomers of PCBG (1-PCBG), where the glutathionyl moiety is attached to the terminal carbon, although the authors were unable to specifically assign the major or minor peak. Reichert et al. (35) synthesized the methanethiol conjugate and mercaptoacetate conjugate of HCBD, pentachloro-1-(methylthio)-1,3butadiene (PCMTB) and pentachloro[(carboxymethyl)thio]-1,3-butadiene, as standards for use in identification of HCBD metabolites in rat urine. The major peak in the GC analysis of PCMTB was assigned as a 1-positional isomer (whether cis or trans was not determined) and had a mass spectrum very similar to that of the major HCBD metabolite. Two additional minor peaks from GC were identified and assigned as the other 1-positional isomer and the 2-positional isomer. Dekant et al. (36) observed three major metabolites after incubation of HCBD with rat liver cytosolic fraction and GS. Two of the three metabolites were inseparable by GC/MS and yielded identical mass spectra to the synthetic PCBG standard. These latter two isomers, having a formation ratio of 20:1, were assigned as the trans and cis forms of 1-PCBG, respectively. RHF/6-31G* energies in Table 4 indicate that the cis1-GS conjugate of HCBD is the most stable isomer, with the trans-1-GS conjugate computed to be less stable than the 2-GS conjugate isomer by a relatively large energy margin. In agreement with these calculated energies, visual inspection of graphical representations of LUMO densities for HCBD within SPARTAN (data not shown) indicated clear preference for an incoming nucleophile at the terminal carbon, C1 position, rather than at the C2 position. The inference that the cis-1-GS conjugate would be the energetically preferred product, however, is in conflict with Dekant et al. (36) who assigned the
Chem. Res. Toxicol., Vol. 9, No. 3, 1996 673
major product of HCBD conjugation as the trans-1-GS isomer. The Dekant et al. (36) assignment was based on a general reactivity argument for nucleophilic vinylic addition/elimination reactions (15), rather than analytical determination. A key element of this argument was the implicit assumption of a 120° rotation about the C1-C2 bond to produce the trans product. However, it is known from experiment (26) and the present calculation that, unlike butadiene, HCBD has a nonplanar, sharply bent structure (C-C-C-C torsional angle of
