Structure−Activity Relationships for the Glutathione Conjugation of 2

The Hammett σp can be divided into an inductive (F) and a resonance (R) component. ...... (21) Department of Chemistry Columbia University (1994) Mac...
0 downloads 0 Views 210KB Size
Download PDF
Chem. Res. Toxicol. 1996, 9, 527-534

527

Structure-Activity Relationships for the Glutathione Conjugation of 2-Substituted 1-Chloro-4-nitrobenzenes by Rat Glutathione S-Transferase 4-4 Ellen M. van der Aar,† Marcel J. de Groot,†,‡ Greetje J. Bijloo,†,‡ Henk van der Goot,‡ and Nico P. E. Vermeulen*,† Leiden/Amsterdam Center for Drug Research, Department of Pharmacochemistry, Divisions of Molecular Toxicology and Medicinal Chemistry, Vrije Universiteit, De Boelelaan 1083, 1081 HV Amsterdam, The Netherlands Received August 8, 1995X

In the present study structure-activity relationships (SAR’s) are described for the experimentally determined kinetic parameters (Km, kcat, and kcat/Km) of the GST 4-4-catalyzed reaction between GSH and 10 2-substituted 1-chloro-4-nitrobenzenes. Steric, lipophilic, and electronic parameters were correlated with the kinetic parameters. Moreover, charge distributions and several energy values were calculated for the substrates and the corresponding Meisenheimer intermediates with MeS- as a model nucleophile for the thiolate anion of GSH and used in the regression analyses. The correlations obtained were compared with the corresponding SAR’s for the base-catalyzed GSH conjugation reaction at pH 9.2. A high correlation coefficient was found between the kinetic parameter ks for the base-catalyzed reaction and the Hammett substituent constant (σp). Much lower correlation coefficients were obtained with kcat and σp and with kcat/Km and σp. Moreover, the reaction constant F was significantly higher for the base-catalyzed than for the enzyme-catalyzed reaction. Also, high correlations were found between the kinetic parameters and the charges on the p-nitro substituent in the substrates. When ks was plotted against these charges, a linear relationship was found in which the slope was larger than the slope of a corresponding plot with kcat/Km. The Hammett σp can be divided into an inductive (F) and a resonance (R) component. With multiple regression between the kinetic parameters and F and R, higher correlation coefficients were obtained than with σp alone. Our observations suggest that the transition states for the base-catalyzed and the GST 4-4-catalyzed GSH conjugation reaction are different. Moreover, single classical physicochemical and computer-calculated molecular parameters and combinations of them can be an alternative approach for examining SAR’s for spontaneous and GSTcatalyzed GSH conjugation reactions.

Introduction The glutathione S-transferases (GSTs; EC 2.5.1.18)1 constitute a complex supergene family that metabolizes chemotherapeutic drugs, carcinogens, environmental pollutants, and a broad spectrum of other foreign and endogenous compounds (1). GSTs catalyze the nucleophilic attack of the tripeptide GSH to electrophilic substrates, which results in addition or substitution reactions, depending on the nature of the substrates (2). Mammalian cytosolic GST-isoenzymes have been categorized into four classes designated R, µ, π (3), and θ (4). Isoenzymes within a gene class typically exhibit a very high degree of amino acid sequence identity (60-80%), whereas intergene class similarities are considerably less (25-35%). * To whom correspondence should be addressed. Tel 020-4447590, Fax 020-4447610. † Division of Molecular Toxicology. ‡ Division of Medicinal Chemistry. X Abstract published in Advance ACS Abstracts, February 1, 1996. 1 Abbreviations: CSD, Cambridge Structural Database; DMA, distributed multipole analysis; Ehomo, energy of the highest occupied molecular orbital; Elumo, energy of the lowest unoccupied molecular orbital; GAMESS-UK, Generalised Atomic and Molecular Electronic Structure System UK-version; GS-conjugate, glutathione conjugate; GSTs, glutathione S-transferases; λmax, maximal absorbance wavelength; MeS-, methanethiolate anion or model nucleophile for the thiolate anion of GSH; RHF, restricted Hartree-Fock; SAR, structureactivity relationship; SDS, sodium dodecyl sulfate; STO, Slater type orbital; SV, split valence.

0893-228x/96/2709-0527$12.00/0

GSTs consist of two subunits, which can be identical (homodimers) or different (heterodimers). Each subunit contains a binding site for GSH (G-site) and a separate binding site for the electrophilic substrate (H-site). Recently, high-resolution crystal structures of class R (5), µ (6, 7), and π (8, 9) isoenzymes of different species in complex with GSH, various substrate and transition state analogues or products have been determined. It appeared that the GST isoenzymes share a number of general structural features but differ considerably in detail. The relationship between structure and the functional properties of GSTs is a central issue with respect to understanding their participation in xenobiotic and endogenous metabolism. Until now a limited amount of research has been done at structure-activity relationships (SAR’s) of GST isoenzymes. Keen et al. (10) showed that Hammett plots of the catalytic constants of rat liver GST 1-1 and 3-4 obtained with a series of 4-substituted 1-chloro-2-nitrobenzenes demonstrated a linear relationship between Hammett σ- substituent constants, reflecting the nucleophilic nature of the enzymatic reaction and its dependence on the electrophilicity of the substrates. Chen et al. (11) found the same linear correlation with the same substrates and rat GST 4-4. The F value of the © 1996 American Chemical Society

528 Chem. Res. Toxicol., Vol. 9, No. 2, 1996 Scheme 1. Reaction Coordinate Diagram for the Nucleophilic Addition of GS- to 2-Substituted 1-Chloro-4-nitrobenzene Derivativesa

van der Aar et al.

could be obtained between the rate constants of the GST 4-4-catalyzed reaction and of the chemical reaction with some striking differences, due to influences of the GST protein.

Materials and Methods

a Catalysis by the enzyme (- - -) compared to the base-catalyzed reaction (s) involves, beyond simple deprotonation of the thiol, stabilization of the transition state for formation of the Meisenheimer complex.

base-catalyzed reaction was 3.4, while the F value for the turnover number of the GST 4-4-catalyzed reaction with GSH was only 1.2. This suggests that there are differences in the transition states for the base- and enzymecatalyzed reactions. The kinetic parameters and relative stereoselectivity of the rat GST 4-4-catalyzed addition of GSH to five parasubstituted 4-phenyl-3-buten-2-ones were determined by Kubo and Armstrong (12). The log kcat/Km values for the formation of the diastereomeric products A and B correlated highly with the substituent constant σp and the Hansch hydrophobic substituent constant π. Also the log of the ratio of the two diastereomeric products correlated with these parameters. The different F values for the formation of the products A and B in these correlations indicated that the two diastereomeric transition states differ in electronic character. It was proposed that one orientation of the substrate in the active site of the enzyme, which results in attack of GS- on one prochiral face of the enone, provides more effective dispersion of charge in the transition state than the other orientation (12). Nucleophilic substitutions on electron-deficient positions in aromatic molecules are generally thought to occur via Meisenheimer or σ-complex intermediates (13) as illustrated in Scheme 1 for a 2-substituted 1-chloro-4nitrobenzene. Using different nucleophiles and leaving groups, strong evidence was obtained that formation of the Meisenheimer intermediate is the rate-limiting step in such reactions. Kinetic evidence suggests that the GST-catalyzed addition of GSH to electron-deficient arenes also proceeds with rate-limiting formation of the σ-complex intermediate (11). In a previous study we examined the ability of four different rat GST isoenzymes from three different classes to catalyze the conjugation of GSH to 11 2-substituted 1-chloro-4-nitrobenzenes (14). For some substrates a pronounced GST isoenzyme selectivity was found. In the present study an attempt was made to correlate the kinetic parameters of GST 4-4 for the GSH conjugation of 10 2-substituted 1-chloro-4-nitrobenzenes (Km, kcat, and kcat/Km) with classical physicochemical parameters and computer-calculated molecular parameters of the substrates and the Meisenheimer intermediates. These SAR’s were compared with the corresponding SAR’s for the chemical reactivity of GSH with the chloronitrobenzene derivatives. It appeared that similar correlations

Materials. The origin, syntheses, and purities of all substrates used in this study (Table 1) have been described previously (14). Buffers and chemical reagents were of the highest quality commercially available. GST isoenzymes were purified from rat liver using affinity chromatography (Shexylglutathione-Sepharose 6B), and isolation of GST 4-4 was achieved by chromatofocusing as described previously (15). Purity was confirmed by sodium dodecyl sulfate (SDS) gel electrophoresis, isoelectric focusing, and HPLC analysis as described by Vos et al. (16) and Bogaards et al. (17). Protein was determined by the method of Lowry, using bovine serum albumin as standard (18). Enzyme Kinetics. The determination of the GST 4-4 kinetic parameters Km, kcat, and kcat/Km of the GSH conjugation of 2-substituted 1-chloro-4-nitrobenzenes has been described in a previous study (14). In short, the formation of GS-conjugates was followed spectrophotometrically with time at 37 °C at the maximal absorbance wavelength (λmax) of the GS-conjugates. The final assay medium contained 0.1 M potassium phosphate buffer (pH 6.5, 0.1 mM EDTA), GST 4-4 (different amounts depending on substrate tested), 1 mM GSH, and varying concentrations of substrate. Kinetics of the Uncatalyzed Reaction. The rate constants of the uncatalyzed reaction between GSH and the 2-substituted 1-chloro-4-nitrobenzenes were determined in a 0.1 M potassium phosphate buffer (pH 9.2) at a temperature of 50 °C. The final concentration of GSH in the assay was 10 mM, while the substrate concentrations were varied (0.05-4 mM). The increase in absorption at the λmax of the GS-conjugates was recorded spectrophotometrically. The second order rate constants of the chemical reaction (ks) were calculated by dividing the absorption increase per minute by the GSH concentration, the substrate concentration, and the extinction coefficient of the corresponding GS-conjugate. All ks values are means ( SD of at least three independent experiments. Determination of Computer-Calculated Molecular Parameters. The initial conformations of substrates 1 and 9 were retrieved from the Cambridge Structural Database (CSD (19)). The initial conformations of the other substrates and the corresponding Meisenheimer complexes with MeS- as a model nucleophile for the thiolate anion of GSH were generated using the molecular modeling package ChemX (20). Where necessary, a systematic conformational analysis was carried out around all rotable axes, to obtain the global minimum conformation using the molecular modeling program Macromodel (21, 22) with the Amber force field (23, 24). The geometries of all substrates and Meisenheimer complexes were subsequently ab initio optimized at the RHF (restricted Hartree-Fock) level using the STO-3G (Slater type orbital comprized of 3 Gaussians) (25) minimal basis set. On the resulting STO-3G geometries, a single point energy calculation and DMA (distributed multipole analysis) calculation (26) was performed using the RHF method in a SV (split valence) 6-31G (27, 28) basis set.2 For the substrates 2-6, several conformations were ab initio investigated. The quantum chemical program package GAMESS-UK (29, 30), implemented on IBM/ RS6000 workstations and on a CRAY-YMP supercomputer, was used for the ab initio calculations described. Structure-Activity Relationships. Simple and multiple regression analyses were applied for determination of correlations between the kinetic parameters of the GST 4-4-catalyzed and the base-catalyzed GSH conjugation reaction with physicochemical parameters (electronic, steric, and lipophilic) of the 2 Due to unavailability of a SV 6-31G basis set for bromine (substrate 10) a SV 3-21G basis set was used.

Structure-Activity Relationships of GST 4-4

Chem. Res. Toxicol., Vol. 9, No. 2, 1996 529

Table 1. R Groups in 2-Substituted 1-Chloro-4-nitrobenzenes and the Corresponding Reference Numbers Used in This Studya substrate

R

ks (×10-4 µM-1‚min-1)

Km (µM)

kcat (min-1)

kcat/Km (×10-2 µM-1‚min-1)

1 2 3 4 5 6 7 8 9 10

NO2 CHO COC6H5 CO2CH3 CO2(CH2)3CH3 CO2C(CH3)3 CN CF3 Cl Br

420 ( 20 5.1 ( 0.1 2.8 ( 0.1 4.3 ( 0.1 3.9 ( 0.5 2.9 ( 0.2 48 ( 1 1.1 ( 0.1 1.0 ( 0.03 1.0 ( 0.2

156 ( 36 63 ( 12 154 ( 18 554 ( 97 341 ( 26 b 292 ( 48 435 ( 53 855 ( 107 381 ( 70

137 ( 14 1.0 ( 0.15 2.2 ( 0.5 6.5 ( 1.4 12.1 ( 0.8 b 23.5 ( 4.3 8.9 ( 1.5 5.4 ( 1.5 6.1 ( 0.8

93 ( 19 1.5 ( 0.4 1.5 ( 0.3 0.9 ( 0.1 3.5 ( 0.1 b 8 ( 1.2 2.0 ( 0.03 0.6 ( 0.2 1.6 ( 0.2

a The second order rate constant of the spontaneous reaction between GSH and these substrates (k ) is shown (assay is described in s Materials and Methods). The kinetic parameters of the GST 4-4-catalyzed reaction between GSH and these substrates (Km, kcat, and b kcat/Km) are shown and are taken from Van der Aar et al. (14). No Lineweaver-Burk plots were determined due to the limited solubility of substrate 6. substituents and the computer-calculated molecular parameters of the substrates and the Meisenheimer complexes with a MeSmodel nucleophile. Only correlations with a Student’s t test’s t value of >|2| were considered to be significant. In case of multiple regression the intercorrelation between the two independent parameters was checked.

Results and Discussion In the present study, the experimentally determined kinetic parameters of GST 4-4 for the GSH conjugation of 2-substituted 1-chloro-4-nitrobenzenes (1-10, Table 1), reported previously (14), and of the kinetic parameters of the corresponding base-catalyzed reaction (Table 1) were correlated with physicochemical parameters and computer-calculated molecular parameters of the substrates and the Meisenheimer complexes with MeS- as a model nucleophile. On the basis of the comparison of the correlations of the base- and GST 4-4-catalyzed GSH conjugation reaction with the physicochemical and calculated molecular parameters, conclusions could be drawn about the rate-determining step in both reactions and about the influence of the GST 4-4 protein. Comparison of Base- and GST 4-4-Catalyzed GSH Conjugation. Comparison of Km, kcat, and kcat/Km of GST 4-4 with 10 different substrates has been done in a previous study (14). Briefly, substrate 9, with a 2-Cl substituent, showed the highest Km (855 µM), suggestive of a low apparent affinity. Substrates with more bulky 2-substituents, like CO2(CH2)3CH3 and Br (substrates 5 and 10), had lower Km values, suggesting that there are no important steric restrictions in this area of the active site of GST 4-4. Substrates 1 and 7 showed the highest and second highest kcat and kcat/Km, indicating that these two substituents stabilize the corresponding Meisenheimer intermediates most efficiently. The second order rate constant of the base-catalyzed GSH conjugation reaction (ks) was approximately 100fold lower than the kcat/Km values of GST 4-4. Substrates 1 and 7 showed respectively the highest and second highest GSH conjugation rate in both the base-catalyzed and the GST 4-4-catalyzed reaction. When the enzymeand base-catalyzed GSH conjugation rates (log k) were plotted against the examined substrates (Figure 1), a different pattern was observed. In an attempt to explain these differences, we applied regression analyses to determine possible correlations between the kinetic parameters of both reactions and classical physicochemical parameters of the substituents and computercalculated molecular parameters of the substrates and the corresponding Meisenheimer complex model molecules.

Figure 1. Comparison of the kinetic parameter ks of the basecatalyzed (--O--) and kcat/Km of the GST 4-4-catalyzed (-b-) GSH conjugation reaction of 2-substituted 1-chloro-4-nitrobenzenes 1-10.

Classical Physicochemical Parameters. To probe the steric effects of the substituents, the Taft steric parameter Es (taken from ref 31) and the multidimensional Sterimol parameters (L, B1, and B5) (32) were considered. The hydrophobic fragment constant f of Rekker (33) was used to probe lipophilicity effects of the substituents. As electronic parameter, the Hammett σ constants were used (taken from ref 31). Some of the substituents investigated in the present study are directly conjugated with the reaction center, and this phenomenon causes “through resonance”. Such interactions may either facilitate or hinder attainment of the transition state. A different Hammett constant, named σ-, has therefore been defined in order to correct for “through resonance” (34). Moreover, the σ values are position-dependent: σ for a given substituent in the meta position (σm) is different from that in the para position (σp). In the present study, ortho substituents were varied and no experimentally determined Hammett substituent constants are known for ortho substituents. In a study of Hess et al., 13C-H coupling constants of orthosubstituted toluenes, anisoles, and benzaldehydes were used to calculate σo (35). Depending on the nature of the remainder of the molecules, the calculated σo values appeared to correlate well with σp values for a series of substituents. Apparently, para parameters can sometimes be used for ortho substituents, although they can lead to discrepancies as well. Also, attempts were undertaken to describe σo by including extra inductive and steric components to σp (36). Swain and Lupton (37) defined two electronic parameters to separate the induc-

530 Chem. Res. Toxicol., Vol. 9, No. 2, 1996

van der Aar et al.

Table 2. Classical Physicochemical Parameters of the Substituents Used in This Study substrate

σp

σp-

F

R

B1

f

1 2 3 4 5 6 7 8 9 10

0.78 0.42 0.43 0.45 0.45 0.45 0.66 0.54 0.23 0.23

1.27 1.02 0.87 0.76 0.76 a 1.01 0.71 0.21 0.23

0.67 0.31 0.30 0.33 0.33 a 0.55 0.38 0.41 0.44

0.16 0.13 0.16 0.15 0.15 a 0.19 0.19 -0.15 -0.17

1.70 1.60 1.92 1.64 1.64 a 1.60 1.99 1.80 1.95

-0.039 -0.333 0.926 0.181 1.738 1.738 -0.155 1.223 0.933 1.134

a This physicochemical parameter was not available for the tertbutyl ester derivative.

Table 3. Computer-Calculated Molecular Parameters of Substrates Examined in This Study (See Materials and Methods)

Table 4. Computer-Calculated Molecular Parameters of Meisenheimer Intermediates with MeS- Used as a Model Nucleophile for GSH (See Materials and Methods) chargea chargea ortho p-nitro chargea subsubElumob sub- attacked Ehomob ∆Ef,mcb strate C atom stituent stituent (kJ/mol) (kJ/mol) (kJ/mol) 1 2 3 4 5 6 7 8 9 10

0.269 0.330 0.327 0.298 0.302 0.302 0.310 0.323 0.339 0.348

-0.328 0.018 0.036 -0.052 -0.051 -0.058 -0.237 -0.003 -0.208 -0.068

-0.380 -0.417 -0.404 -0.425 -0.420 -0.426 -0.426 -0.430 -0.465 -0.456

379.4 445.2 439.5 460.4 455.4 459.1 438.3 442.7 458.5 459.6

-436.7 -388.8 -394.5 -381.6 -387.0 -382.7 -395.8 -401.2 -366.1 -356.2

-209.5 -164.3 -154.7 -146.7 -148.8 -143.7 -170.4 -167.7 -138.3 -129.6

a Charges were obtained from DMA calculations. b Energy values were obtained from SV 6-31G calculations.

chargea chargea chargea attacked ortho p-nitro Elumob Ehomob substrate C atom substituent substituent (kJ/mol) (kJ/mol) 1 2 3 4 5 6 7 8 9 10

0.265 0.234 0.216 0.222 0.219 0.216 0.264 0.242 0.150 0.139

-0.148 0.167 0.233 0.118 0.123 -0.102 -0.135 0.106 -0.130 0.015

-0.169 -0.196 -0.201 -0.193 -0.193 -0.196 -0.184 -0.191 -0.204 -0.205

44.3 3.1 20.9 15.0 16.1 20.3 19.2 15.5 1.9 12.7

-1094.4 -1055.0 -930.7 -1029.9 -1029.6 -1023.9 -1060.6 -1068.4 -1021.4 -979.4

a Charges were obtained from DMA calculations. b Energy values were obtained from SV 6-31G calculations.

tive component of the electronic effects of substituents (field effect, F) from the resonance component (R) (taken from ref 31). The classical physicochemical parameters of the substituents actually used in this study are listed in Table 2. Computer-Calculated Molecular Parameters. Charge and energy calculations of the substrates and the Meisenheimer intermediates with MeS- as a model nucleophile afforded the charge distributions of the molecules; among others, charges at the C1 atom attacked, charges at the ortho substituents and p-nitro substituents, and the energies of the highest occupied molecular orbital (Ehomo) and of the lowest unoccupied molecular orbital (Elumo) were obtained. For the Meisenheimer complexes, the energies of formation (∆Ef,mc) were calculated by subtracting the energies of the substrates and MeS- from the energies of the Meisenheimer complexes. Furthermore, the energy differences between Elumo and Ehomo were calculated and used in the regression studies. Also, the differences in charges at the attacked C1 atom, the ortho substituent and the p-nitro substituent in the substrates and the Meisenheimer intermediates were calculated and used in the correlation study. The computer-calculated molecular parameters of the substrates are shown in Table 3 and those of the Meisenheimer intermediates in Table 4. The charge and energy differences between substrates and intermediates are not shown, but can be calculated from the given data. Correlations with Km. Firstly, the log Km values for the GSH conjugation of the 2-substituted 1-chloro-4nitrobenzenes by GST 4-4 were correlated with all steric, lipophilic, and electronic parameters mentioned above. No statistically significant correlations were observed (t < |2|), indicating that the affinity of these substrates toward GST 4-4 (i.e. formation of ES complexes, Scheme

Figure 2. Hammett plots of the log of the GST 4-4 kinetic constants kcat (b) and kcat/Km (O) and of the base-catalyzed reaction ks (2) for the GSH conjugation of 2-substituted 1-chloro4-nitrobenzenes vs σp. Electrophilic substrates used and the corresponding σp values are listed in Tables 1 and 2.

1) is not dependent on any of these parameters. Also the computer-calculated charge distributions and energies and their differences in the substrates and Meisenheimer complexes were analyzed for correlations with Km. None of the parameters shown in Tables 3 and 4 led to statistically significant correlations (t < |2|). Correlations of ks, kcat, and kcat/Km with Electronic Parameters. The log of the second order rate constant of the base-catalyzed GSH conjugation of 2-substituted 1-chloro-4-nitrobenzenes (log ks) correlated well with the Hammett σp constant:

log ks ) 4.58((0.47)σp - 5.26((0.23) r ) 0.970

s ) 0.233

(1)

n)8

The CF3-substituted substrate (8) appeared to be an outlier, and furthermore, substrate 6 has been excluded from the regression analysis because no data were available for the GST 4-4-catalyzed conjugation of this compound. To make a well-founded comparison between the base- and enzyme-catalyzed GSH conjugation, it is necessary to include an identical data set into the SAR’s. Equation 1 indicates that the more electron-withdrawing the ortho substituents are, the more efficiently the Meisenheimer intermediates are stabilized and the higher ks becomes. The 2-substituents of substrates 1-7 all possess the ability for “through-resonance”, and the ks values were

Structure-Activity Relationships of GST 4-4

Chem. Res. Toxicol., Vol. 9, No. 2, 1996 531

therefore expected to correlate with the Hammett electronic parameter corrected for “through-resonance”. Chen et al. (11) found an excellent correlation between ks and σp- for 4-substituted 1-chloro-2-nitrobenzenes. The fact that in the present study ks did not correlate with σpindicates that “through-resonance” on the ortho position is not taking place, possibly because of steric hindrance in the Meisenheimer intermediates between the ortho substituent and the Cl atom at C1, which is acting as leaving group. In the Meisenheimer intermediate C1 is sp3-hybridized, taking much more space, compared to the sp2-hybridized C1 in the ground state (Scheme 1). The kinetic parameters kcat and kcat/Km of the GST 4-4catalyzed reaction correlated considerably less with σp:

log kcat ) 2.42((0.99)σp - 0.20((0.49) r ) 0.705

s ) 0.499

n)8

log kcat/Km ) 3.23((0.72)σp - 3.01((0.35) r ) 0.877

s ) 0.362

(2)

(3)

n)8

Substrate 8 was excluded from this regression analysis because it was an outlier in eq 1. Based on the high correlation coefficient observed between ks and σp and the low correlation coefficients between kcat and σp and kcat/Km and σp, it may be concluded that the GST 4-4catalyzed reaction clearly depends less on the electronwithdrawing capacity of the ortho substituent than the corresponding base-catalyzed reaction. This is also clear from the slopes of the Hammett plots (reaction constant F) in Figure 2, which give an indication about the sensitivity of the reaction to electronic effects of the substituents. It can be seen that the slope in the Hammett plot of the base-catalyzed reaction is clearly higher (F ) 4.58) than that of the GST 4-4-catalyzed reaction (F ) 2.42 for kcat and F ) 3.23 for kcat/Km). This suggests that there are differences in the transition states for the two reactions. It is possible that steps other than the formation of a Meisenheimer complex are partially rate-determining in the GST 4-4-catalyzed reaction. It is not very likely, however, that release of the respective GS-conjugates from the active site of GST 4-4, a phenomenon found for the release of 1-(S-glutathionyl)-2,4dinitrobenzene from the active site of GST 3-3 (38), is rate-limiting. In the latter isoenzyme a hydrogen bond between Ser209 and Tyr115 is proposed to be formed, by which elimination of this GS-conjugate from the active site is restricted (38). In GST 4-4 residue 209 is an Ala residue (39), not capable of forming this specific hydrogen bond (40). Also, release of a Meisenheimer intermediate from the enzyme surface as the initial product can become rate-limiting. This process might be expected to be related to the Hammett σp since the stability of the enzyme-intermediate complex and the transition state for release could be influenced by the electron density and distribution in the Meisenheimer complexes. The correlation between kcat and kcat/Km and σp was not very high (eqs 2 and 3), so probably the release of Meisenheimer complex out of the active site is not rate-limiting. It is also possible that the smaller F value of the enzymatic reaction when compared to the spontaneous reaction reflects an earlier, more reactant-like transition state for the nucleophilic attack of the thiolate on the enzyme surface. GSH is efficiently deprotonated by GSTs via formation of a stabilizing hydrogen bond with a Tyr residue (41). Moreover, GSTs position GSH and the

Figure 3. Correlations between the kinetic constants of the base-catalyzed (ks, 2) and the GST 4-4-catalyzed (kcat, b, and kcat/Km, O) GSH conjugation reaction of 2-substituted 1-chloro4-nitrobenzenes and the calculated charges on the p-nitro substituents of the substrates (charge p-NO2 (S)). The correlation coefficients of the correlation between ks, kcat, and kcat/Km and the charges are 0.984, 0.927, and 0.927, respectively, and the slopes of the equations are 74, 42, and 54, respectively.

electrophilic substrate in such an orientation that attack of GS- proceeds in the most efficient way (proximity effect). Because only low correlations were found between the kinetic parameters kcat and kcat/Km of GST 4-4 and classical electronic parameters, an attempt was made to correlate those kinetic parameters with computercalculated charge distributions in the substrates and the corresponding Meisenheimer complexes with MeS- as a model nucleophile. No statistically significant correlations were found between the kinetic parameters and charge distributions (charge on attacked C1 atom, ortho and para substituents, Table 4) in the Meisenheimer model intermediates. In the Meisenheimer complexes a net charge of -1 is divided over the whole molecule and the corresponding charge distributions depend on the electron-withdrawing capacity of the aromatic ring substituents. Apparently, none of the kinetic parameters are determined by these charge distributions: neither ks, kcat, nor kcat/Km correlated significantly with the charge on the ortho substituents in the substrates (Table 3). It was expected that the charge on the ortho substituents would reflect the electron-withdrawing capacity of the ortho substituent corrected for steric influences. However, this appeared not to be the case. A high correlation coefficient was found between ks, kcat, and kcat/Km and the calculated charge on the p-nitro group in the substrates (Figure 3). The latter parameter indirectly reflects the electronwithdrawing capacity of the ortho substituents. From the differences in slope in the plots of Figure 3, it can be seen that the ks of the spontaneous GSH conjugation reaction depends more on the electron-withdrawing capacity of the substituents (slope ) 74) than the GST 4-4-catalyzed reaction (slope kcat ) 42 and slope kcat/Km ) 54). This again suggests a difference in transition state between the base- and GST 4-4-catalyzed conjugation reaction. Multiple Regression Analyses. The Hammett σp constant can be separated into two parts: the inductive or field effect F and a resonance component R (37). When the kinetic parameters (ks and kcat/Km) were correlated simultaneously with F and R, higher correlation coefficients were found than with σp alone (eqs 1 and 3):

532 Chem. Res. Toxicol., Vol. 9, No. 2, 1996

log ks ) 3.61((0.46)R + 5.19((0.52)F - 5.61((0.23) (4) r ) 0.985

s ) 0.181

n)8

log kcat/Km ) 2.02((0.74)R + 4.33((0.83)F - 3.50((0.36) (5) r ) 0.936

s ) 0.290

n)8

Substrate 8 appeared to be an outlier in the correlation with ks, and consequently was excluded from the correlation with kcat/Km. When eqs 4 and 5 were transformed to log k - c1(R) ) c2(F) + c3, the plots indicated in Figure 4 were obtained. It can be seen that the slopes in the plots of the base- and enzyme-catalyzed conjugation are similar, and so these two reactions depend in the same way on the field effect F. The correlation between kcat and F and R simultaneously was not significant (t values < |2|). Clearly, when σp is divided into inductive and resonance components, the correlation coefficients increase. This is caused by the possibility to give more emphasis to one of the two components, in this case the inductive component. The correlation coefficients between ks and kcat/Km with F alone were 0.784 and 0.833, respectively, while with R alone no statistically significant correlations were found. However, combination of both parameters apparently gave the highest correlations (eqs 4 and 5). When a multiple regression was performed between the charges on the attacked C1 atom in the substrates and the F values, high correlations for ks and kcat/Km were found as well (substrate 8 as outlier in ks correlation):

log ks ) 3.82((0.47)F + 12.26((1.33)(charge C atom) - 7.38((0.29) (6) r ) 0.989

s ) 0.155

n)8

log kcat/Km ) 3.61((0.94)F + 6.46((2.67)(charge C atom) - 4.42((0.58) (7) r ) 0.927

s ) 0.310

n)8

The correlations with the single parameters were significantly lower: log ks and log kcat/Km with F, 0.784 and 0.833, respectively, and with the calculated charges on the attacked C1 atom, 0.834 and 0.666, respectively. From eqs 6 and 7 it can be seen that the contribution of the field effect (F) in both correlations is the same ((3.7), indicating that both the base- and the enzyme-catalyzed reaction depend in a similar way on the field effects of the substituents. However, the contribution of the charge on the attacked C1 atom in the substrates (Table 3) is twice as high in the spontaneous reaction (eq 6) compared to the GST 4-4-catalyzed reaction (eq 7). The fact that the GST 4-4-catalyzed reaction apparently depends less on the charge on C1 can be explained by the more efficient deprotonation of GSH by the enzyme, thus leading to a stronger nucleophilic GS- species, compared to the basecatalyzed reaction. Also, via formation of hydrogen bonds between atoms in the substrates and amino acids in the active site of the GST protein, charge distributions at the substrates, and in particular at C1, could change, and possibly diminish the differences in charge. In all correlations between ks and electronic parameters, the CF3-substituted compound (8) appeared to be an outlier. The GSH conjugation rate was lower than

van der Aar et al.

calculated from the equations on account of the electronwithdrawing capacity of CF3. This can possibly be explained by the high negative charge on the F atoms (-0.30), leading to a negatively charged electrostatic field around the CF3 substituent, subsequently leading to repulsion of GS- when attacking C1 for the formation of the Meisenheimer complex. In the case of enzymecatalyzed conjugation of substrate 8, this electrostatic field could be neutralized by the effects of the enzyme, for example, an interaction with a positively charged amino acid. Spontaneous vs Enzyme-Catalyzed Reaction. When the enzyme- and base-catalyzed GSH conjugation reactions are compared, a relatively low correlation coefficient was obtained between kcat/Km and ks:

log kcat/Km ) 0.68((0.11) log ks + 0.67((0.38) r ) 0.915

s ) 0.283

(8)

n)9

A correlation coefficient of 0.915 indicates that additional factors influence the GST 4-4-catalyzed reaction when compared to the base-catalyzed. A higher correlation coefficient was found by including the steric Sterimol B1 factor:

log kcat/Km ) 0.82((0.11) log ks + 1.35((0.60)B1 1.27((0.92) (9) r ) 0.954

s ) 0.226

n)9

In the Sterimol system, L is defined as the length of the substituent along the axis of the bond between the first atom of the substituent and the parent molecule and B1 is the minimum width (32). The influence of B1 in the correlation indicates that, among others, steric factors are causing the difference between the enzyme- and basecatalyzed conjugation of 2-substituted 1-chloro-4-nitrobenzenes to GSH. Another parameter which appeared to increase the correlation coefficient between kcat/Km and ks was the lipophilicity fragment constant f of Rekker (33):

log kcat/Km ) 0.86((0.11) log ks + 0.35((0.13)f - 1.04((0.31) (10) r ) 0.963

s ) 0.204

n)9

Clearly, the active site of GST 4-4 provides such an environment that lipophilic interactions increase the rate of GSH conjugation, compared to the base-catalyzed reaction.

Conclusions In this SAR study, kinetic parameters of the basecatalyzed (ks) and GST 4-4-catalyzed (Km, kcat, kcat/Km) GSH conjugation reaction of 10 different 2-substituted 1-chloro-4-nitrobenzenes (Table 1) were correlated with steric, electronic, and lipophilic parameters (Table 2). Moreover, charge distributions and energies of the substrates (Table 3) and corresponding Meisenheimer intermediates with MeS- as model nucleophile for the thiolate anion of GSH were calculated (Table 4) and included in the SAR study. It appeared that ks of the base-catalyzed reaction correlated well with the electronic Hammett σp parameter (r ) 0.970), while kcat and kcat/Km of the enzyme-catalyzed

Structure-Activity Relationships of GST 4-4

Chem. Res. Toxicol., Vol. 9, No. 2, 1996 533

(4)

(5)

(6)

(7)

Figure 4. Effect of F and R on the base-catalyzed (ks, 2) and GST 4-4-catalyzed (kcat/Km, O) GSH conjugation of 2-substituted 1-chloro-4-nitrobenzenes. Equations 4 and 5 are transformed to log k - c1(R) ) c2(F) + c3. For ks, c1 ) 3.61, c2 ) 5.19, c3 ) -5.61. For kcat/Km, c1 ) 2.02, c2 ) 4.33, c3 ) -3.50.

reaction correlated considerably less (r ) 0.705 and 0.877, respectively). Also, a larger slope (F value) in the Hammett plot (Figure 2) was found for the spontaneous reaction. Both observations indicate that the basecatalyzed GSH conjugation reaction depends more on the electronic effects of the substituents than the corresponding enzyme-catalyzed reaction and that there are differences between the transition states of the base- and GST 4-4-catalyzed reaction. The Hammett σp constant can be divided into an inductive effect (F) and a resonance component (R). When the ks and kcat/Km values were correlated with F and R simultaneously (Figure 4), higher correlation coefficients were found than with σp alone. A multiple regression between the charges on the attacked C atom in the substrates and the F values for the different substituents led to high correlation coefficients for ks and kcat/Km. The contribution of the field effect (F) in both correlations was the same, the one of the charge on the attacked C1 atom in the substrates was twice as high in the spontaneous reaction when compared to the enzymecatalyzed reaction. The fact that all correlations with kcat/Km were considerably better than those with kcat is explainable since these two steady-state kinetic constants are composed of different rate constants and reflect free energy differences between different ground states and possibly different transition states. When kcat/Km was correlated with ks, a relatively low correlation coefficient was found (r ) 0.915). Inclusion of the steric Sterimol B1 parameter or the lipophilicity fragment constant f led to a significant increase of the correlation coefficients (r ) 0.954 and 0.963, respectively). This suggests that steric and lipophilic factors cause the observed differences between the base- and GST 4-4catalyzed conjugation of 2-substituted 1-chloro-4-nitrobenzenes to GSH.

References (1) Commandeur, J. N. M., Stijntjes, G. J., and Vermeulen, N. P. E. (1995) Enzymes and transport systems involved in the formation and disposition of glutathione S-conjugates. Pharmacol. Rev. 47, 271-330. (2) Armstrong, R. N. (1991) Glutathione S-transferases: reaction mechanism, structure, and function. Chem. Res. Toxicol. 4, 131140. (3) Mannervik, B., A° lin, P., Guthenberg, C., Jenssen, H., Tahir, M. K., Warholm, M., and Jornvall, H. (1985) Identification of three

(8)

(9)

(10) (11)

(12)

(13) (14)

(15)

(16)

(17)

(18) (19) (20) (21) (22)

(23)

classes of glutathione transferases common to several mammalian species. Correlation between structural data and enzymatic properties. Proc. Natl. Acad. Sci. U.S.A. 82, 7202-7206. Meyer, D. J., Coles, B., Pemble, S. E., Gilmore, K. S., Fraser, G. M., and Ketterer, B. (1991) Theta, a new class of glutathione transferases purified from rat and man liver. Biochem. J. 274, 409-414. Sinning, I., Kleywegt, G. J., Cowan, S. W., Reinemer, P., Dirr, H. W., Huber, R., Gilliland, G. L., Armstrong, R. N., Ji, X., Board, P. G., Olin, B., Mannervik, B., and Jones, T. A. (1993) Structure determination and refinement of human-alpha class glutathione transferase A1-1, and a comparison with the mu and pi class enzymes. J. Mol. Biol. 232, 192-212. Ji, X., Zhang, P., Armstrong, R. N., and Gilliland, G. L. (1992) The three-dimensional structure of glutathione S-transferase from the Mu gene class. Structural analysis of the binary complex of isoenzyme 3-3 and glutathione at 2.2 Å resolution. Biochemistry 31, 10169-10184. Ji, X., Johnson, W. W., Sesay, M. A., Dickert, L., Prasad, S. M., Ammon, H. L., Armstrong, R. N., and Gilliland, G. L. (1994) Structure and function of the xenobiotic substrate binding site of a glutathione S-transferase as revealed by X-ray crystallographic analysis of product complexes with the diastereomers of 9-(Sglutathionyl)-10-hydroxy-9,10-dihydrophenanthrene. Biochemistry 33, 1043-1052. Reinemer, P., Dirr, H. W., Ladenstein, R., Schaffer, J., Gallay, O., and Huber, R. (1991) The three-dimensional structure of class pi glutathione S-transferase in complex with glutathione sulfonate at 2.3 Å resolution. EMBO J. 10, 1997-2005. Reinemer, P., Dirr, H. W., Ladenstein, R., Huber, R., Lo Bello, M., Federici, G., and Parker, M. W. (1992) Three-dimensional structure of class π glutathione S-transferase from human placenta in complex with S-hexylglutathione at 2.8 Å resolution. J. Mol. Biol. 227, 214-226. Keen, J. H., Habig, W. H., and Jakoby, W. B. (1976) Mechanism for the several activities of the glutathione S-transferases. J. Biol. Chem. 251, 6183-6188. Chen, W.-J., Graminski, G. F., and Armstrong, R. N. (1988) Dissection of the catalytic mechanism of isoenzyme 4-4 of glutathione S-transferase with alternative substrates. Biochemistry 27, 647-654. Kubo, Y., and Armstrong, R. N. (1989) Observation of a substituent effect on the stereoselectivity of glutathione S-transferase toward para-substituted 4-phenyl-3-buten-2-ones. Chem. Res. Toxicol. 2, 144-145. Miller, J. (1968) in Reaction Mechanisms in Organic Chemistry (Eaborn, C., and Chapman, N. B., Eds.) Vol. 8, pp 137-179, Elsevier, New York. Van der Aar, E. M., Commandeur, J. N. M., Buikema, D., Te Koppele, J. M., Van Ommen, B., Van Bladeren, P. J., and Vermeulen, N. P. E. (1996) Enzyme kinetics and substrate selectivities of rat glutathione S-transferase isoenzymes towards a series of new 2-substituted 1-chloro-4-nitrobenzenes. Xenobiotica, in press. Ploemen, J. H. T. M., Bogaards, J. J. P., Veldink, G. A., Van Ommen, B., Jansen, D. H. M., and Van Bladeren, P. J. (1993) Isoenzyme selective irreversible inhibition of rat and human glutathione S-transferases by ethacrynic acid and two brominated derivatives. Biochem. Pharmacol. 45, 633-639. Vos, R. M. E., Snoek, M. C., Van Berkel, W. J. H., Mu¨ller, F., and Van Bladeren, P. J. (1988) Differential induction of rat hepatic glutathione S-transferase isoenzymes by hexachlorobenzene and benzyl isothiocyanate. Biochem. Pharmacol. 37, 1077-1082. Bogaards, J. J. P., Van Ommen, B., and Van Bladeren, P. J. (1989) An improved method for the separation and quantification of glutathione S-transferase subunits in rat tissue using highperformance liquid chromatography. J. Chromatogr. 474, 435440. Lowry, O. H., Rosebrough, N. J., Farr, A. L., and Randall, R. J. (1951) Protein measurement with the Folin phenol reagent. J. Biol. Chem. 193, 265-275. Allan, F. H., and Kennard, O. (1993) 3D search and research using the Cambridge Structural Database. Chemical Design Automation News 8, 31-37. Chemical Design Ltd. (1993) ChemX, version January 1993. Department of Chemistry Columbia University (1994) Macromodel, version 4.5. Mohamadi, F., Richards, N. G. J., Guida, W. C., Liskamp, R., Lipton, M., Caufield, C., Chang, G., Hendrickson, T., and Still, W. C. (1990) MacroModelsAn integrated software system for modeling organic and bioorganic molecules using molecular mechanics. J. Comput. Chem. 11, 440-467. Weiner, S. J., Kollman, P. A., Case, D. A., Singh, U. C., Ghio, C., Alagona, G., Profeta, S., and Weiner, P. (1984) A new force field for molecular mechanical simulation of nucleic acids and proteins.

534 Chem. Res. Toxicol., Vol. 9, No. 2, 1996 J. Am. Chem. Soc. 106, 765-784. (24) Weiner, S. J., Kollman, P. A., Nguyen, D. T., and Case, D. A. (1986) An all atom force field for simulations of proteins and nucleic acids. J. Comput. Chem. 7, 230-252. (25) Hehre, W. J., Stewart, R. F., and Pople, J. A. (1969) Self consistent molecular orbital methods. I. Use of Gaussian expansions of Slater-type atomic orbitals. J. Chem. Phys. 51, 2657-2664. (26) Stone, A. J. (1981) Distributed multipole analysis, or how to describe a molecular charge distribution. Chem. Phys. Lett. 83, 233-239. (27) Binkley, J. S., Pople, J. A., and Hehre, W. J. (1980) Self consistent molecular orbital methods. 21. Small split-valence basis sets for first-row elements. J. Am. Chem. Soc. 102, 939-947. (28) Gordon, M. S., Binkley, J. S., Pople, J. A., Pietro, W. J., and Hehre, W. J. (1982) Self consistent molecular orbital methods. 22. Small split-valence basis sets for second-row elements. J. Am. Chem. Soc. 104, 2797-2803. (29) Dupuis, M., Spangler, D., and Wendoloski, J. (1980) NRCC Program No. QG01 (GAMESS). (30) Guest, M. F., Lenthe, J. H. v., Kendrick, J., Schoffel, K., Sherwood, P., Harrison, R. J., with contributions from Amos, R. D., Buenker, R. J., Dupuis, M., Handy, N. C., Hillier, I. H., Knowles, P. J., Bonacic-Koutecky, V., Niessen, W. v., Saunders, V. R., and Stone, A. J. (1993) GAMESS-UK, IBM RS6000 v2.1. (31) Hansch, H., and Leo, A. (1979) Substituent constants for correlation analysis in chemistry and biology, Wiley, New York. (32) Silipo, C., and Vittoria, A. (1990) Three-dimensional structure of drugs. In Comprehensive medicinal chemistry (Hansch, C., Sammes, P. G., and Taylor, J. B., Eds.) Vol. 4, pp 153-204, Pergamon Press, Oxford. (33) Rekker, R. F., and Mannhold, R. (1992) Calculation of drug

van der Aar et al.

(34) (35) (36) (37) (38)

(39)

(40)

(41)

lipophilicity. The hydrophobic fragmental constant approach, VCH, Weinheim, Germany. Hammett, L. P. (1940) Physical Organic Chemistry, McGraw-Hill, New York. Hess, R. E., Schaeffer, C. D. and Yoder, C. H. (1971) 13C-H coupling constants as a probe of ortho-substituent effects. J. Org. Chem. 36, 2201-2202. Fujita, T., and Nishioka, T. (1976) The analysis of the ortho effect. In Progress in Physical Organic Chemistry (Taft, R. W., Ed.) Vol. 12, pp 49-89, Wiley, New York. Swain, C. G., and Lupton, E. C. (1968) Field and resonance components of substituent effects. J. Am. Chem. Soc. 90, 43284337. Johnson, W. W., Liu, S. X., Ji, X. H., Gilliland, G. L., and Armstrong, R. N. (1993) Tyrosine-115 participates both in chemical and physical steps of the catalytic mechanism of a glutathione S-transferase. J. Biol. Chem. 268, 11508-11511. Ding, G. J. F., Ding, V. D. H., Rodkey, J. A., Bennett, C. D., Lu, A. Y. H., and Pickett, C. B. (1986) Rat liver glutathione Stransferases. DNA sequence analysis of a Yb2 cDNA clone and regulation of the Yb1 and Yb2 mRNAs by phenobarbital. J. Biol. Chem. 261, 7952-7957. De Groot, M. J., Vermeulen, N. P. E., Mullenders, D. L. J., and Donne´-Op den Kelder, G. M. (1996) A homology model for rat mu class glutathione S-transferase 4-4. Chem. Res. Toxicol. 9, 2840. Liu, S., Zhang, P., Ji, X., Johnson, W. W., Gilliland, G. L., and Armstrong, R. N. (1992) Contribution of tyrosine 6 to the catalytic mechanism of isoenzyme 3-3 of glutathione S-transferase. J. Biol. Chem. 267, 4296-4299.

TX9501391