Chem. Res. Toxicol. 1996,8, 780-791
780
Investigations of Glutathione Conjugation in Vitro by lH NlMR Spectroscopy. Uncatalyzed and Glutathione Transferase-Catalyzed Reactions Gina Kubal,' David J. Meyer,t Richard E. Norman,§and Peter J. Sadler*?+ Department of Chemistry, Birkbeck College, University of London, London WClH OPP, U.K., CRC Molecular Toxicology Group, Department of Biochemistry and Molecular Biology, University College London, Windeyer Building, Cleveland Street, London W1P 6DB, U.K., and Department of Chemistry, Duquesne University, Pittsburgh, Pennsylvania 15282 Received February 13, 1995@ Conjugation reactions of glutathione (GSH) and related thiols with diethyl maleate (DEM) and other a,p-unsaturated carbonyl compounds have been investigated by 'H NMR spectroscopy. The products from the reaction with DEM and diethyl fumarate (DEF) are shown to be the diastereomers of 5'-(a$-diethoxycarbonylethy1)glutathione. During the course of the reaction, DEM isomerized to DEF, and the rate of isomerization was dependent upon whether the solvent was lH2O or 2H20. The observed rate data exhibit apparent second order kinetic behavior. The reaction of maleate with GSH was considerably slower, and solvent-dependent isomerization was observed, while little reaction of fumarate with GSH was observed a t pH 6.5. Reaction of DEM with N-acetyl-L-cysteine followed a similar course to that of GSH, and although L-cysteine reacted rapidly with DEM, it did not promote the isomerization of DEM. Reactions involving penicillamine and N-acetylpenicillamine were considerably slower. Conjugation reactions catalyzed by commercial GSH transferases and selected rat and human purified isoenzymes were also investigated. Of those isoenzymes studied, rat GSH transferase 4-4 was found to exert the greatest degree of stereo control in conjugation reactions with DEF.
Introduction
ecules. Therefore. dedetion of heDatic GSH through GSH-transferase-catalyzedconjugation can be both rapid A large number of compounds with electrophilic centers and effective. conjugate with GSH under catalysis by GSH-transferase The conjugation of GSH with diethyl maleate (DEM), (GST, EC 2.5.1.18)192(1-8). All cytosolic GSH transdiethyl fumarate (DEF), and other a,P-unsaturated carferases have a common dimeric structure with subunits bonyl compounds was first reported in 1967 by Boyland of molecular mass between 23 and 26 kDa (5). To date, and Chasseaud (11). The effective conjugation of GSH rat, pig, and human GSTs have been characterized in by DEM has led to the use of DEM as a GSH-depleting greatest detail, including determinations of the X-ray agent (12-14). However, despite this, there appears to structures of rat GST 3-3 (mu class), pig and human GST be no detailed study to date of this conjugation reaction pi, and human GST Al-1 (alpha class), as reviewed in in vitro, either catalyzed or uncatalyzed. ref 9. Although GSH-transferases have been shown to In recent years, high-field nuclear magnetic resonance be present in many animal tissues, they are usually most spectroscopy has been increasingly used for the identiabundant in the liver where they often represent 2-5% fication of GSH conjugates. 'H NMR (500 MHz) has been of soluble protein (10). The major GST subunits of rat used to study the products from the conjugation of GSH liver are 1and 2 (alpha class) and 3 and 4 (mu class). In with 4-nitrosophenetole (151, and I3C-IH 2D heterohuman liver, subunits A1 and A2 predominate, and lower nuclear correlated spectroscopy has assisted in the amounts of subunit M1 are present in ca. 50% of structural determination of the various cyclized adducts individuals (5). Under the right conditions, the presence formed during the conjugation of GSH with formaldehyde of GSH-transferase minimizes the possible nonenzymatic (16). Products arising from the conjugation of GSH with conjugation of hydrophobic electrophiles with macromolthe carcinogenic vinyl halide 1,l-dichloroethylene have been analyzed by IH NMR (17), and IH and 19Fevidence has been reported for stereochemical control of the * To whom correspondence should be addressed. + Birkbeck College, University of London. conjugation reaction of GSH with chlorotrifluoroethene i University College London. (18). Various NMR studies of GSH conjugation have Duquesne University. been extensively reviewed (19-21). Abstract published in Advance ACS Abstracts, June 15, 1995. GST nomenclature is based on a numerical system for the We chose to study the conjugation reaction of GSH with chronological order of the characterization of GSH transferase subDEM as a possible means of trapping intracellular GSH units. The human GSH transferases until recently were denoted by in a form suitable for direct 'H NMR analysis; although Greek letters according to their acidic, basic, or near-neutral isoelectric points (6, 7). Now a gene classification is used for human enzymes, IH NMR signals from GSH are readily detected for some and that used in this work is designated Alphal-2 ( d y ) or GSTA1-2 cells, e.g., erythrocytes (22), for others, e.g. hepatocytes, (8). they are not (23). In this work, the reactions of GSH and Abbreviations: GSH, glutathione; DEM, diethyl maleate; DEF, dimethyl fumarate; TSP, sodium 3-(trimethyl~ilyl)[2,2,3,3-~H]propi- other selected thiols with DEM and related compounds onate; DCEG, S-(a$-bis(ethoxycarbony1)ethyl)glutathione; DTNB, 5,5were studied using high resolution IH NMR spectroscopy. dithiobis(2-nitrobenzoic acid); TNB,5-thio-2-nitrobenzoic acid pH*, pH As this technique provides an overall view of events meter reading in 2Hz0: p2H = pH* + 0.4 (24). I
&
@
0893-228x/95/2708-0780$09.00/0
0 1995 American Chemical Society
'HNMR Studies of Glutathione Conjugation occurring in the sample at the time of NMR measurement, new information has been gained on the course of the conjugation reactions. Complete lH and 13C spectral assignments of the conjugation products were assisted by computer simulation and the use of 2D 13C-'H heteronuclear correlated spectroscopy. Enzyme-catalyzed conjugation reactions were also investigated and their stereoselectivity was examined.
Experimental Procedures Instrumentation. lH NMR data were collected at 500 MHz using a Bruker AM-500 spectrometer (Medical Research Council Biomedical NMR Centre, Mill Hill, London) and 5 mm NMR tubes. Typically, data were collected using 16 k data points with a spectral width of 5000 Hz and 45-60" pulses separated by a relaxation delay of 1.5-2.5 s. Where appropriate, the intense 'HzO signal was suppressed by application of either presaturation or continuous secondary irradiation a t the relevant frequency. 13C spectra were acquired at 100.6 MHz on a Bruker AM-400 spectrometer (Medical Research Council Biomedical NMR Centre) and 10 mm NMR tubes. Typically d a t a were collected using 64 k data points with a spectral width of 30 kHz and broad band 'H decoupling. Two-dimensional COSY (correlated spectroscopy) spectra were accumulated using similar conditions. Spectral simulation was achieved using Parameter Adjustment in NMR by Iteration Calculation (PANIC) provided by Bruker Spectrospin. The root mean square difference in frequencies between transitions for the theoretical and experimental spectra were within the 0.6 Hz error limit. The spectrometers were operated in quadrature detection mode, and d a t a were collected at 25 "C unless otherwise stated. Sodium 3-(trimethylsilyl)[2,2,3,3-2Hlpropionate (TSP) was used as a chemical shift reference. Nonenzymatic Conjugation. (A) Materials. Sources of chemicals were as follows: glutathione (reduced), fumarate, and N-acetyl-L-cysteine from Sigma Chemical Co. (Poole, Dorset, U.K.); diethyl maleate and maleate from Aldrich Chemical Co. Ltd. (Gillingham, Dorset, U.K.); diethyl fumarate from Lancaster Synthesis Ltd. (Lancaster, U.K.);N-acetyl-D-penicillamine and DL-penicillamine from Fluka Chemicals Ltd. (Gillingham, Dorset, U.K.). Solutions of each of the reactants were made up to a concentration of 35 mM in phosphate buffer (0.2 M) prepared i n 2H20 (pH* = 6.6 (24)h3 For samples prepared in 'HzO, 10% 2Hz0 was added to provide a field-frequency lock. (B)Method. In most cases, GSH and other thiol conjugates were synthesized by reaction of equimolar quantities of the reactants at 25 "C. For 5 mM reactions, DEM and DEF were sulfoxide. added as a 10% v/v solution in [1,1,1,3,3,3-*H]dimethyl For the majority of the reactions, *H20 buffer was used. However, some of the reactions carried out in lHzO were first allowed to go to completion before lyophilization and redissolving in 2H20. This procedure did not alter the spectrum of the reaction product when compared to that obtained in 'HzO. The loss of reactants and formation of products during conjugation were monitored by collecting a series of spectra at regular time intervals and measuring the changes in area of selected resonances. For GSH loss and DCEG formation, the cysteine a C H signals at 4.57 and 4.65 ppm were monitored, respectively. To follow changes in either the DEM or DEF concentrations, the singlets a t 6.46 and 6.91 ppm were measured corresponding to the vinylic protons of unreacted DEM and DEF, respectively. Absolute concentrations were calculated by comparing peak areas at specified time intervals with the same peak of the control GSH sample before addition of the reactant. As well as a chemical shift reference, TSP also served as a n internal concentration standard for this procedure. Enzymatic Conjugation. (A) Materials. Various batches of r a t hepatic GSH-transferases were purchased from Sigma a s lyophilized samples. The individual isoenzyme distribution within the batches was unknown and was unavailable from the G. Kubal, M. Pue, and P. J. Sadler, unpublished.
Chem. Res. Toxicol., Vol. 8, No. 5, 1995 781 supplier. Individual purified isoenzymes: 1-1,2-2,3-3, 4-4 (rat hepatic), and GSTA1-2 (human hepatic) were prepared as described previously (25, 26). Their purities were '97% as assessed by reverse phase HPLC analysis (27).
(B)Method. (i) Synthesis of GSH Conjugates Using Commercial GSH Transferases. GSH conjugates were synthesized by GSH transferase-catalyzed reactions of equimolar quantities ( 5 mM) of GSH and reactant. Reactions were carried out at 25 "C in 0.2 M phosphate buffer in 2H20, pHWca. 6.1, since previous workers (11) have shown that although GSHtransferases show optimum activity at pH 7.6, by lowering the pH to 6.5 (or pH* 6.1 in 2H20)2 the rates of the nonenzymatic reaction and aerobic oxidation of GSH are reduced. The quantity of lyophilized enzyme used was chosen to be comparable with the GSH transferase/GSH ratio reported for rat liver (ca. 50 unitd5.5 pmol of GSH) (28, 291, using the GSH transferase activities supplied by Sigma Chemical Co., which were measured using the standard assay (30, 31). Conjugation reactions were initiated by addition of substrate to a solution of GSH and GSH-transferases.
(ii) Synthesis of GSH Conjugates Using Selected Individual GSH Transferase Isoenzymes. GSTs 1-1,3-3, and 4-4 and GSTA1-2 were prepared in aqueous phosphate buffer (0.15 M); GSH transferase 2-2 was in a solution of aqueous NaCl (0.1 M). All isoenzymes were stored in solution a t -20 "C. For conjugation reactions, the isoenzymes were removed from storage and thawed for 15 min prior to use. GSH conjugates were synthesized by reaction of equimolar quantities (3 mM) of GSH and either DEM or DEF in the presence of one of the isoenzymes. All conjugation reactions were carried out at 37 "C in the suspending medium, supplemented with further phosphate buffer in 2H20 (0.15 M) where necessary. The quantity of enzyme used was, as far as possible, chosen to be comparable with the individual isoenzyme/GSH transferase ratio for rat liver and h u m a n liver: 1-1,1.0; 2-2, 0.3; 3-3, 0.6; 4-4, 0.7; A1-2, 0.3 mg mL-l. In most cases, once conjugation was complete, samples were lyophilized, redissolved in 2H20, and reexamined by single-pulse NMR to improve the resolution of the spectra. For each catalyzed conjugation reaction examined, a standard uncatalyzed reaction was also prepared under identical conditions as a control.
Results
N M R Studies on the Coqjugation of GSH with Diethyl Maleate and Diethyl Fumarate. (A) 'H NMR Studies. The 'H NMR spectrum of the product from the conjugation reaction of GSH with DEM in 'HzO phosphate buffer (pH = 6.5) is shown in Figure 1, and the data are collected in Table 1. The spectrum reveals pairs of resonances in approximately 1:l ratios arising from the alkene moiety (Ha, Hb, and Hc) and from the cysteinyl CHZ moiety (H1 and H2) of the products (for labels see Figure 2). These pairs of signals arise from the formation of diastereomers of S-(a,P-bis(ethoxycarbony1)ethyl)glutathione (DCEG). Conjugation of GSH with DEM or with DEF produces the same final product as judged by lH and 13C NMR spectroscopy. DCEG exists as a pair of diastereomers since the addition of sulfur to the a carbon4of the alkene creates a chiral center; see Figure 2. The diastereomers are labeled I and I1 and are assumed to possess the absolute configurations S and R , respectively, at the a carbon of the alkene moiety. The corresponding NMR signals arising from these two diastereomers are distinguished by either undashed (isomer I) or dashed (isomer 11) labels. These NMR assignments are arbitrary since ~~~~~~~~~~~~~~~
* The terms a and B carbon, when referring to the alkene-derived
moiety of the conjugation product, refer to the carbon atom attached to sulfur and the other alkenyl-derived carbon, respectively.
782 Chem. Res. Toxicol., Vol. 8, No. 5, 1995
Kubal et al.
Hg
H
H2 H2' w-
I
'
1
-
l
~
I
I
3.2
HH
1
H'
I
Hd He
"
4.0
l
'
l
'
l
'
~
~
"
3.0
~
3.0
'
l
'
l
~
"
~
~
"
l
'
l
'
~
,
p
2.0
Figure 1. 500 MHz 'H NMR spectrum of DCEG resulting from the conjugation of GSH with DEM in 1HzO phosphate buffer (pH 6.5).
no attempt was made to assign peaks to absolute configurations. With the exception of the cysteinyl -CH2-, the GSH moiety of DCEG gave extensively overlapped signals for the two isomers (Figure 1,Table 1). Assignment of the ester -CH3 and -CH2- protons was achieved through selective proton decoupling experiments and comparison with the spectra obtained from the enzymecatalyzed system (vide infra).Pairing of cysteinyl -CH2signals (i.e., H1 with H2, and H1' with H2')was achieved through proton decoupling and 2D-correlated spectroscopy (COSY). The 2D spectra also assisted in resolving the coupling pattern between protons Ha, Hb, and H', since each of these protons gives rise to a doublet of doublets at 500 MHz. The assignments were confirmed by simulating the spectra of the relevant spin systems (Ha,Hb, and Hc)(Figure 3a,b). The chemical shifts and coupling constants for each isomer are listed in Table 1. The proton adding to the p carbon of the alkene during conjugation appears to be derived from the solvent. This is supported by changes in the Ha, Hb, and H' regions of the spectrum when the reaction is conducted in 2 H ~ 0 phosphate buffer (pH* = 6.1),as shown in Figure 3c,d. The coupling pattern between Ha, Hb, and Hc was resolved by 2D experiments. Due to the acquisition of deuterium at the p carbon atom of the alkene moiety, the Ha and Hb signals in each isomer are doublets since they are coupled to only one other proton, H'. Since deuterium can add at either of two magnetically inequivalent positions (labeled Hama'and Hb/Hb')fo* each isomer, there are four possible conjugation products with closely related chemical shifts. This assignment was confirmed by simulating the spectrum for the relevant spin systems (Ha, Hb, and Hc) for each of the products and comparing the sum to the experimental spectrum. The data are collected in Table 1.
The ratios of the diastereomers of DCEG which are products of reaction with DEM or DEF were the same, irrespective of the relative concentrations of the reactants and the pH. (B)Deuterium Isotope Shifts. The chemical shifts for DCEG are independent of the nature of the solvent with the exception of Ha, Hb, and Hc. The isotopic shifts of protons Hama'or Hb/Hb'and H'/H'' when either Hama' or Hb/Hb'is replaced by 2H were determined; see Table 2. As expected, the two-bond deuterium isotope shifts for Hama' and Hb/Hb'are greater than the three-bond shifts for Hc/Hc'. As well as altering the shifts of neighboring protons, deuterium addition caused a marked broadening of the geminal proton resonances, an effect due to the coupling of the quadrupolar 2H with the geminal 'H. N M R spectra of deuterated (C)13CNMR Studies. and nondeuterated DCEG confirmed the assignments made from the 'H NMR studies. Because of the greater range of the 13C spectrum relative to that of the 'H spectrum (approximately 200 ppm vs 10 ppm), individual 13C resonances are well spaced, making it easier to observe chemical shift differences between isomers, particularly for carbon atoms close to the point of conjugation. Chemical shift assignments for DCEG were made by 13C-lH heteronuclear 2D-correlated spectroscopy and comparison with reported 13C NMR studies of GSH (32,331. A comparison of the spectra from samples in 'HzO and 2 H ~ clearly 0 identified the carbon atom which became deuterated; the sharp singlets which were seen for Ca in lHzO were observed as broad triplets when the reaction was carried out in 2H20,which indicates the site of deuterium attachment. The data are collected in Table 3.
p
m
lH NMR Studies
Chem. Res. Toxicol., Vol. 8, No. 5, 1995 783
of Glutathione Conjugation
Table 1. lH NMR Data assignment Cys H1 Cys H2 Cys H3 Gly H4 Gly HS Glu H6 Glu H7 Glu Ha Glu H9 Glu HIO alkyl-Ha alkyl-Hb alkyl-HC ester Hd ester He ester Hf ester Hg J(1,a J(1,3) J12a J14.5) J16.7) J16,a) J(6,9) J(7,a) JI~,~J J(s,g)
Jwo) J(9,io) Jla,bJ J(a,c~ Jlb,c) Jidn JIe,€!1
DCEG
(IP
2.97 (d of d) 3.29 (d of d) 4.65 (d of d) 3.75 (d) 3.81 (d) 2.52 (m) 2.57 (m) 2.17 (m) 2.17 (m) 3.78 (d of d) 3.00 (d of d) (2.99) (d) 2.86 (d of d) (2.85) (d) 3.86 (d of d) (3.86) (d) 4.27 (9) 4.18 (9) 1.30 (t) 1.25 (t) 14.2 9.3 4.8 17.2 15.3 7.4 7.6 7.7 7.8 15.3 6.4 6.3 17.1 9.6 5.6 7.2 7.2
DCEG (IIP 3.07 3.24 4.64 3.75 3.81 2.52 2.56 2.16 2.16 3.77 2.99 (2.98) 2.90 (2.89) 3.88 (3.87) 4.25 4.18 1.29 1.25 14.1 8.6 5.1 17.1 15.5 7.4 7.6 7.7 7.8 15.4 6.4 6.3 17.0 9.4 5.9 7.2 7.2
GSH
Chemical Shih, 6 (ppm) 2.93 (m) 2.98 (m) 4.57 (d of d) 3.78 (8) 2.55 (m) 2.58 (m) 2.17 (m) 2.17 (m) 3.79 (d of d)
Coupling Constants (Hz) 14.1 7.1 5.2 15.3 7.4 7.8 7.5 7.7 15.5 6.4 6.2
GSSG
DEMb
DEFb
4.28 (9)
4.30 (9)
1.31 (t)
1.32 (t)
7.2
7.2
2.97 (d of d) 3.32 (d of d) 4.57 (d of d) 3.76 (d) 3.81 (d) 2.53 (m) 2.57 (m) 2.16 (m) 2.16 (m) 3.78 (d of d)
14.2 9.7 4.5 17.2 15.5 7.5 7.8 7.5 7.7 15.0 6.3 6.1
a !l"he values in parentheses are for spectra in 2H20 buffer, see text. The multiplicities of peaks for isomer 11 are as given for isomer I. The labels for protons are defined in Figure 1. DEM is diethyl maleate, and DEF is diethyl fumarate.
(D)Kinetics of Conjugation. The conjugation reactions of GSH with DEM and DEF were monitored as a function of time and of concentration (5,15, and 30 mM in both alkene and GSH). Reaction profiles for the uncatalyzed conjugation of GSH (5 mM) with DEM and GSH (5 mM) with DEF, in 2H20, are shown in Figure 4a,b. During the course of the DEM conjugation reaction a significant proportion of the substrate isomerized to DEF. This was apparent from the appearance in the 'H NMR spectrum of a singlet at 6.91 ppm, the chemical shift of the vinylic protons of DEF, and was confirmed by the addition of DEF. Thus, GSH catalyzes the isomerization of DEM to the thermodynamically more stable trans isomer, DEF. In contrast, no isomerization was observed during the conjugation of GSH with DEF. Analogous studies in 'H2O buffer were attempted, but the dynamic range problem caused by the intense 'HzO signal (ca. 110 M in protons) prohibited studies at 5 and 15 m M concentrations. The water signal could not be suppressed sufficiently with secondary irradiation to enable detection of peaks within 0.1 ppm of the water peak. No other method of water suppression was attempted. At GSH and substrate concentrations of 30 mM, all peaks were readily observed with secondary irradiation of the 'H2O signal. No allowance was made for possible effects on the intensity of the cysteine aCH peak at 4.55 ppm due to the nearby secondary irradiation of lHzO. Although isomerization of DEM to DEF was s t i l l observed in 'H20, it occurred to a lesser extent. To rule
out the possibility that the small amount of isomerization observed during conjugation in 'HzO was the result of the 10% 2H20 added prior to spectral acquisition, a conjugation reaction of GSH with DEM was carried out in the absence of any % 2 0 . There was very little difference in the extent of isomerization. The rate data were found to obey apparent second order kinetics (Figure 4c,d), and the observed rate constants are collected in Table 4. When the pH* of the reaction between GSH and DEM was increased to 9, conjugation was complete within minutes. The increase in pH* did not affect the relative quantities of the diastereomers formed or affect isomerization. When the pH* of the reaction was decreased to 3.5, no reaction was observed even after 20 h although a small proportion (ca. 5%) of DEM had isomerized to DEF. To assess the degree of spontaneous isomerization of DEM to DEF, a 30 mM solution of DEM in 2Hz0 phosphate buffer (pH* = 6.1) was prepared. No isomerization was observed after the sample had stood for 10 h at 22 "C, although a small amount of hydrolysis occurred. Since DEM isomerizes during conjugation with GSH, the reversibility of the process was investigated via lH or 2Hincorporation from the solvent at various stages of the reaction. When aliquots of 'H2O buffer (up to 50% of sample volume) were titrated into a sample of DCEG (formed after complete reaction between GSH and DEM in 2H20buffer), and the sample was left to stand at room temperature for ca. 90 min at 22 "C, no spectral changes
784 Chem. Res. Toxicol., Vol. 8, No. 5, 1995
GS’ I
HI
i ‘ H
CH~CHZOOC~COOCHZCH~
bS’ (lS,ZS,3S-Configuration)
(lS,2S.3R.Confguration)
Figure 2. Diastereomers (I and 11) of DCEG formed in a n approximately 1:l ratio on conjugation of GSH with diethyl maleate in phosphate buffer.
were observed. Similarly, when aliquots of 2Hz0 buffer were also titrated into a sample of DCEG formed in ‘HzO, no spectral changes were observed. ‘HNlMR Studies on the Coqjugation of GSH with Maleate and Fumarate. It has been reported that, on administration of DEM to cells, tissues, and whole animals, DEM is rapidly hydrolyzed to maleate by tissue esterases (34). Consequently, the reactions of GSH with maleate and fumarate were also investigated. The reaction of 5 mM GSH with maleate at pH* = 6.1 was relatively slow at 25 “C. The GSH concentration decreased to 3.9 mM after 7 h. Some of this loss was through oxidation of GSH to GSSG (ca. 0.2 mM), despite purging the solvents and solutions with nitrogen. Additionally, extensive isomerization of maleate to fumarate was observed; e.g., after 3 h, the ratio of unreacted maleate to fumarate was ca. 1:9. However, in lHzO (pH 6.51, the extent of isomerization was greatly reduced, as seen previously for the reaction with DEM. In a separate experiment, the reaction of GSH with fumarate was found to be slower still (no reaction was observed after 12 h). When the pH*’s of the conjugation reactions in 2Hz0 were lowered to 3.5, reactions of both maleate and fumarate were faster with GSH (1:l);after 7 h, ca. 50% of the substrate (maleate or fumarate) had reacted with GSH. Also at this pH* (3.5),isomerization of maleate to fumarate was significantly inhibited compared to pH* 6.1; after 3 h at pH* 3.5, the ratio of unreacted maleate to fumarate was ca. 8 5 1 , compared to 1:9 at pH* 6.1. No isomerization of fumarate to maleate was observed under any of the conditions employed. Reactions of Diethyl Maleate with Other Thiols. Reactions of DEM with N-acetyl-L-cysteine, L-cysteine, N-acetyl-D-penicillamine, and DL-penicillaminewere stud-
Kubal et al. ied a t DEM:thiol (1:l) in both 2 H ~ and 0 lHzO buffers (pH* 6.1 and pH 6.5, respectively). The results for the reaction of DEM with N-acetyl-L-cysteine were very similar to those found for GSH; i.e., the production of diastereomers, the solvent-derived 2H or lH on the p carbon of the alkene, and the isomerization of DEM to DEF. In contrast, while two diastereomers of approximately equal proportions were produced from the reaction of DEM with L-cysteine, there was no evidence for isomerization. For the reaction of DEM with N-acetylD-penicillamine, no evidence for either isomerization or conjugate formation was observed. Only very weak resonances were observed for the reaction of DEM with DL-penicillamine even after 20 h, together with a small amount (ca. 2%)of DEF. Catalysis by GSH Transferases. (A)Commercial Enzymes. IH NMR studies of the enzyme-catalyzed conjugation of GSH with selected a$-unsaturated carbonyl compounds were carried out under essentially the same conditions as described for uncatalyzed conjugation reactions. The GSH-transferase/GSH ratios used were chosen to be comparable with those reported for rat liver (28, 29). Under these conditions, GSH-transferases increased the rate of conjugation of GSH with DEM (and DEF) by approximately 20-fold, see Table 4. Careful comparison of the ‘H NMR spectrum of the catalyzed reaction with that of the uncatalyzed reaction revealed that the enzyme mixture had exerted some stereo control during conjugation (Figure 5). Resonances for isomer I predominated over resonances for isomer 11. GSH-transferases were also found to influence the position at which the hydrogen atom became attached to the alkenyl ,8 carbon. For isomer I (the major product) deuterium addition occurred predominantly at the Hb site. For isomer I1 (the minor product) it was difficult to assess the selectivity of deuterium incorporation since the resonances are reduced in intensity and the Ha‘resonance overlaps with one from the Cys -CH2- protons (Figure 5). Selective deuteration was also reflected in the resonances for the Hc/Hc’protons. Thus, although resonances for all four products arising from conjugation in 2Hz0 phosphate buffer were observed, one product predominated. The relative proportions of the various products are collected in Table 5. GSH-transferases also increased the rate of conjugation of GSH with DEF (>20-fold, data not shown). Interestingly, the ‘H NMR spectrum of the products indicated that while GSH transferase catalysis again increased the relative yield of isomer I, the predominant site of deuterium addition was at Ha, opposite to that for DEM (Figure 5C). The distribution of products is collected in Table 5. In addition to influencing the stereochemistry of conjugation, the presence of GSH transferases also reduced the amount of isomerization of DEM to DEF. When the quantity of GSH-transferases used to catalyze the conjugation reaction was reduced by 30%,the stereospecificity of the reaction diminished, and the amount of isomerization of DEM to DEF increased. It seems likely that, at these low GSH-transferase concentrations, the uncatalyzed reaction contributes significantly to the ratio of the final products. Since the uncatalyzed reaction produces equal amounts of isomers I and 11,the apparent stereospecificity is reduced. In contrast to the catalyzed conjugation of the diesters, the presence of GSH-transferases had no apparent effect on the conjugation of GSH with either maleate or
‘HNMR Studies of Glutathione Conjugation Glu-OCHN
Chem. Res. Toxicol., Vol. 8, No. 5, 1995 785
OCNH-GlY
G I u - O Y ~ F M * G ~
H2XF3 HC-C-COOCH~~CH,’ s
s Hc-~-COOCH~dCH3‘ D/Hb-~-CO0CH~Y!H~g
Hb-$-COOCH,eCH# Ha HC’ HC
Ha Ha’
Ha Ha,
D/Hn
Hb’
HC’ HC
Hb
HC’ HC
3.9
3.0
2.9
3.0
2.9
Figure 3. IH NMR spectra showing (a) simulated and (b) experimental resonances for spin systems labeled Ha, Hb, and HC(isomer I) and Ha’, Hb, and HC‘(isomer 11) of DCEG after the conjugation of GSH with DEM i n lH2O buffer (pH 6.5), and 1H NMR spectra of DCEG formed by conjugation of GSH with DEM in (c) *HzO buffer and (d) ‘HzO buffer. Resonances for protons Ha, Hb, and HC (diastereomer I) and Ha’, Hb,and Hc‘(diastereomer 11) are shown. In 2Hz0, deuterium is incorporated into positions Ha or Hb (or Ha’ or Hb), effectively causing decoupling of these protons.
Table 2. Deuterium Isotope Shifts for DCEG deuterium isotope shift (ppmp isomer II
isomer I Ha Hb HC
0.015 (7.9) 0.016 (7.7) 0.005 (2.5)
Ha’ Hb’ HC’
0.015 (7.1) 0.014 (7.3) 0.004 (2.2)
a Values in Hz are shown in parentheses. On replacing lH with 2H, the resonance for the neighboring proton shifts to low field.
Table 3.
I3C NMR
Chemical Shifts (ppm)
DCEGa assignment C2 c4 c5 c7
C8 c9
Ca
Cb Cc,d Ce,f
IC
11‘
46.28 56.02 35.82 34.25 29.12 57.09 38.85 45.29 65.71 16.10
46.28 55.54 35.36 34.25 29.12 57.09 39.14 44.44 65.01 16.10
GSH 46.16 58.43 28.34 34.16 28.97 56.91
DEMb
129.23 (CH) 164.40 (C-0) 60.24 13.20
DEFb
132.75 163.71 60.28 13.05
‘HzO buffer, see text. DEM is diethyl maleate, and DEF is diethyl fumarate. e I and I1 refer to the different isomers. Carbon atoms are numbered as shown below.
DCEG
s
H-~~-COOC~H,CCH, €I-$‘- COO@H,C%, H
fumarate at pH* 6.1, nor did GSH-transferases significantly affect the isomerization of maleate to fumarate. The experiments were repeated using different batches of GSH-transferases from the same commercial supplier, and it was found that the stereospecificity, in terms of isomer distribution, of the enzyme-catalyzed conjugation reactions of GSH with DEM or DEF depended on the specific isoenzyme distribution in each batch, Table 5.
(B)Purified Isoenzymes. In order to assess the contributions of different isoenzymes to stereocontrol, conjugation reactions of GSH with DEM were catalyzed by several purified homodimeric rat hepatic GSH-transferase isoenzymes (1-1,2-2,3-3,and 4-4) and one human hepatic transferase (GSTA1-2) and monitored by ‘H NMR. Since the conjugation reactions catalyzed by the individual isoenzymes were conducted in ‘HzO buffer in which the isoenzymes were prepared, information about the site of deuterium addition was not obtained, except for the 4-4 case. The DCEG isomer distributions of the isoenzyme-catalyzed conjugation reactions are collected in Table 5. No attempts were made to determine the kinetics of any of these catalyzed reactions. With DEM conjugation to GSH, the presence of GSHtransferase 1-1favored the formation of a small excess of isomer I1 and inhibited isomerization of DEM to DEF. The data obtained for the GSH-transferase 1-1catalyzed conjugation reaction were very similar to those for the catalyzed conjugation reactions using the second batch of commercial GSH-transferases. In contrast, GSHtransferase 1-1catalysis of GSH conjugation with DEF resulted in the production of approximately equal proportions of isomers I and 11. For GSH-transferase 2-2, only the conjugation of GSH with DEM was investigated. No evidence was found for any stereocontrol of conjugation, and isomerization of DEM to DEF was apparently unaffected. Diethyl maleate conjugation catalyzed by GSH-transferase 3-3 favored an excess of isomer I. A small amount of isomerization of DEM had occurred by the end of conjugation (ca. 10-15%), similar to that of the uncatalyzed reaction in ‘HzO. For DEF conjugation, the presence of GSH-transferase 3-3 favored the formation of a small excess of isomer 11. Since it was possible to prepare a concentrated solution of GST 4-4 in 2 H ~ buffer, 0 it was possible to determine the site of deuterium attachment in the DCEG product. With DEM, isomer I was favored, and the isomerization of DEM to DEF was significantly reduced. This confirmed that the isoenzyme was directly involved in the reaction, since the extent of uncatalyzed isomerization
786 Chem. Res. Toxicol., Vol. 8, No. 5, 1995
Kubal et al.
1.o
v)
-
0.8
E
2
0.6
m W
cz
-
g 0.4
.-
-m
GSH DCEG DEF
0.2 I
0.0 4 0
80
-
1.o
$ 0.8
240
160 time (min)
0
50
100
50
100
150 200 time (min)
250
300
250
300
(c) DCEG DEM
e!
m
%
0.6
W Q W
.z-m 0.4 m
a:
0.2 0.0
0
80
160 time (min)
240
0
150 time (min)
200
Figure 4. Reaction profiles showing the conjugation of GSH with (a) DEM and (b) DEF (all reactants 5 mM in 2Hz0buffer), and corresponding second order plots for (c) DEM and (d) DEF, where [AI0 and [AIt are the concentrations of GSH at time zero and t , respectively, measured from NMR peak areas.
Table 4. Pseudo Second Order Rate Constants for Cosiugation of GSH with DEM and DEF at 25 “C ~~
conjugation conditions
1O6k,dDEM)”~ 106k,t,,(DEF)a (M-l min-l) (M-I min-I)
uncatalyzed in *HzO uncatalyzed in IH20 catalyzed batch 1 uncatalyzed in ‘HzO ( 3 4 ) a
2.3 C! 0.2 2.5 i 0.4 37.6 0.79
4.3 i 0.3 4.8 f 0.6 5.3
Observed rate constants for DEM and DEF.
increased in 2Hz0 solution. There appeared to be little preference for the site of deuterium addition since the signal intensities of Ha and Hb were comparable. GSH-transferase 4-4 again favored the formation of isomer I from DEF, and the site of deuterium attachment was mainly at Ha. The difference in the relative intensities of the resonances for each isomer produced by isoenzyme catalysis was by far the most pronounced for GSH-transferase 4-4. Also, the results were similar to the commercial GSH transferase-catalyzed reactions with DEF. For GSTA1-2, only the conjugation of GSH with DEM was investigated. There was no significant difference in the DCEG isomer ratio compared to the uncatalyzed reaction. However, the isomerization of DEM to DEF was greatly enhanced.
Discussion Conjugation reactions of GSH with DEM and DEF were originally studied by Boyland and Chasseaud (11, 35). They determined the kinetics of the reaction by following the disappearance of unconjugated GSH from
the reaction mixture via addition of DTNB. The latter rapidly reacts with unconjugated GSH to form a GS-TNB conjugate as well as reduced TNB and GSSG, and the TNB absorbance was monitored at 412 nm. The use of ’H NMR spectroscopy has enabled us to identify several other features of conjugation reactions of GSH with DEM and other substrates: the production of diastereomers, the isomerization of DEM to DEF, and the observation that DCEG, once formed, is not in equilibrium with the reactants or intermediates. The formation of diastereomers upon GSH conjugation has been observed previously (18) using a different substrate (l,l,l-trichloroethene). The latter workers also examined the effect of crude preparations of cytosolic and mitochondrial GSHtransferases on the stereoselectivity of diastereomer formation. In the present work, we have been able to study the stereoselectivity of purified isoenzymes. Direct Coqjugation. The role of GSH as a cofactor in catalyzed cis-truns isomerization reactions is well documented (36-381, particularly for maleate and its derivatives. We have shown here that, even in the absence of enzymes, isomerization of DEM still proceeds at a finite rate under the inflence of GSH. Isomerization of the substrate leads to the production of the thermodynamically more stable trans isomer (reverse isomerization is not observed). This isomerization has important consequences for kinetic studies (vide infra). The experiments reported here allow us to comment on the mechanism for the conjugation reaction which can be considered to be a Michael type addition reaction, Figure 6. The pH* dependence of the DEM conjugation reaction (rapid reaction at pH* 9, slow reaction at pH*
Chem. Res. Toxicol., Vol.8, No. 5, 1995 787
‘H NMR Studies of Glutathione Conjugation HC
Ha’
Hb
ua’
HC’HC
H2 H2’ nn
,
,
,
I
’ 314
,
I
I
/
’ 3:O Figure 5. IH NMR spectra of DCEG. (a) Formed from the uncatalyzed reaction of GSH with DEM in deuterated phosphate buffer (pH* 6.1),(b) as (a) but catalyzed by commercial GSH transferase, and (c) a s (b) but with DEF. Predominance of isomer I over isomer I1 can be seen in (b) and (c) (e.g., compare the intensities of H2 and HT). Note that Hb/Hb’is preferentially deuterated in (b) whereas in (c) Ha/Ha’is preferentially deuterated.
3.8
3.5) is consistent with GS- being involved in the rate limiting step (pKa(SH) 9.65 (39)). The observed rate constants for the conjugation of GSH with DEM (kobs (DEM))determined by NMR (Table 4) were about 3 times greater than that obtained by Boyland and Chasseaud (0.79 x mM-’ min-l). This reaction was found to be much more complex than the conjugation of GSH with DEF since DEM isomerizes during the course of the reaction. Thus, k,b,(DEM) includes contributions from the following: (i) the formation of the intermediates from DEM (kADEM)); (ii) the reverse rate (k,(DEM)); (iii) the isomerization of DEM to DEF (hi); (iv) KADEF); and (VI MDEF), where kADEM) is the rate constant of interest. Since Boyland and Chasseaud (I11 determined the rate constants from data collected during the first 5 min of reaction, this would have substantially decreased the influence of isomerization on the measured rate, and their k,b,(DEM) FZ KADEM). In the NMR work reported here, it was found that nearly 2% of DEM isomerized to DEF within 5 min after initiating conjugation. In our studies k,b,(DEM) was determined from data obtained over several hours, and isomerization of DEM occurred both in 2Hz0and, to a lesser extent, in ‘HzO. Thus, DEF is always present, and our pseudo second order rate constant (K,b,(DEM)) is likely to include contributions from at least five separate contributing rate constants: KADEM), K,(DEM), ki,KADEF), and KLDEF). The determination of these latter rate constants is not trivial and is beyond the scope of the present work. The rate constants determined for the conjugation of GSH with DEF compared well (within experimental error) with the value of 5.3 x mM-l min-’ obtained by Boyland and Chasseaud. It must be noted, however, that the con-
stants determined by NMR are observed rate constants and exhibit apparent second order kinetic behavior. This observed rate constant for DEF conjugation (k,b,(DEF)) most likely contains both a rate constant for the formation of the intermediates of conjugation and the reverse reaction (k,(DEF)). Geometric isomerizations of DEM and maleate were dependent on whether the solvent was IHzO or 2 H ~ 0The . observed difference between isomerization rates in these solvents may result from the different lifetimes of the intermediateb) in the two solvents. A bond to protium can be broken as much as 18 times faster than a bond to deuterium (40). Thus, we suggest that in 2 H ~ the 0 intermediates are sufficiently long-lived (before 2Haddition) to allow rotation about the Ca-Cp bond followed by re-formation of GS- and alkene. Overall, the reaction rate for formation of DCEG is independent of solvent since the RDS does not involve cleavage of X-H/D from the intermediateb). The overall rate of conjugation is still determined by the rate determining step, attack by GS-. Isomerization of DEM to DEF in the presence of GSH does not involve 2H incorporation. This was demonstrated by the observation that reaction of DEM with GSH at a 1:2 mol ratio in 2Hz0 buffer resulted in no significant loss of GSH after the disappearance of both the DEM and DEF signals from the spectrum; i.e., the molar loss of DEM and DEF is equal to the loss of GSH. If deuterium had been incorporated into either DEM or DEF during the isomerization process, then the signal intensity observed for the DEM and DEF would not reflect the total quantity of DEM and DEF, and the decrease in signal intensity of DEM and DEF would be greater than the decrease in signal intensity of the GSH. This observation suggests that acquisition of hydrogen at the p carbon of the alkenyl moiety takes place during the final step of the conjugation reaction (Figure 6). The final step (formation of DCEG) is effectively irreversible: no C-WH exchange was observed from lHzO to 2 H ~or 0 vice versa. In their study of the conjugation of chlorotrifluoroethene with GSH, Dohn et al. (18) proposed that the hydrogen atom acquired on addition of GSH across the double bond arose from the solvent or from a pool of hydrogen atoms in exchange with the solvent. The results reported here appear to be in agreement with this proposal, as well as with the observations reported by Seltzer in 1961 (41) that deuterium (from 2Hz0)is not incorporated into fumarate during the isomerization of maleate and that deuterium is not lost during the potassium thiocyanate-catalyzed isomerization of [2,32Hlmaleic acid in IHzO. Although GSH is known to conjugate or take part in addition reactions with carbonyl groups (as in its reaction with formaldehyde (1611, there is no NMR evidence to support a similar reaction with any of the a,p-unsaturated carbonyls studied here; i.e., CH~CHZOCOCH=CHC(OH)(SG)OCHZCH~ was not formed. In the case of maleate and fumarate, the opposite pH* dependence was observed to that of DEM and DEF (i.e., reactions were faster at pH* 3.5 compared to pH* 6.1). This suggests that the dominant factor here is the low electrophilicity of Ca (or Cp) since both carbonyl groups are deprotonated (2nd pK:s 5.83 and 4.10 for maleic and fumaric acids, respectively (42)). At low pH, attack presumably involves GSH rather than GS-. At high pH the high negative charge presumably destabilizes the
788 Chem. Res. Toxicol., Vol.8,No. 5, 1995
Kubal et al.
Table 5. Isomer Distributions for Catalyzed GSH Conjugation Reactions S
"$: H'
Reaction
I(i)
2H
:.if:
G
G
G
I(ii)
*:$:
H$r
II(i)
Hb
G S
2H
II(ii)
Hb
(A) Using commercial enzymes in 2H20 buffer 9%
%
%
%
54. 33,43
13, I , 12
20, 41. 34
13, 19, 1 1
Batch 1,2,3
8,22, 16
5 1, 44,48
10, 18, 12
31, 16,24
Uncataly sedb
28+1
24f4
2552
23+4
QEMa Batch 1,2,3
REP
(B) Using purified isoenzymes with the dimeric structure indicatedc.
DEMa
Isomer ratio Uncatalysedb
I - I1 ( 1:1)
I - I1 (1:l)
I1 > I (1.3:l) I > II (1.S:I) I > 11 (1.5:l) I > II (1.S:l) I-II(1:l)
I I1 (1:l)
IsoenzvmeC 1- 1 2-2 3-3 4-4 A1-2
DEP Isomer ratio
-
-d
II > I (1.3:l) I >> 11 (2:l) -d
Effect on DEM
isomerization
none
decrease none none decrease increase
a DEM refers to conjugation with diethyl maleate and DEF with diethyl fumarate. Refer to Figure 1 for structures of isomers I and 11. Average of three experiments. Values are quoted for one experiment only. Not studied.
intermediate and isomerization is still observed, but little product is formed. Although both N-acetyl-L-cysteine and L-cysteine were found to conjugate readily with DEM and DEF, the rate of conjugation of DEM with N-acetyl-L-cysteine was relatively slow compared to that with L-cysteine, and only N-acetyl-L-cysteine was found to significantly catalyze the isomerization of DEM to DEF. Thus, N-acetyl-L-cysteine behaves more like GSH than L-cysteine, and it would appear that the amide linkage between L-cysteine and L-glutamine in GSH exerts some influence upon the isomerization reaction. This influence may be the result of stabilization of the intermediate product(s j of conjugation, providing more time for isomerization before reformation of the reactants. Donation of a proton from the amine group of L-cysteine in turn may enhance product formation and inhibit isomerization. N-AcetylL-cysteine has been shown to be an efficient substrate for rat liver microsomal GST (43), but not for cytosolic GST (44). Surprisingly, uncatalyzed reactions of Nacetylpenicillamine and DL-penicillaminewith GSH were very slow or did not occur at all. Also, no isomerization of DEM to DEF was observed after 24 h. It is not easy to rationalize the differences in the behavior of these thiol substrates in terms of differences in their pK,(SH) (9.65,
10.36,9.52,10.66, and 10.33 for GSH (391, L-cysteine (451, N-acetyl-L-cysteine (461, m-penicillamine (451, and Nacetyl-D-penicillamine (42j, respectively, and the effects may be due to unfavorable steric interactions involving the methyl groups of the penicillamine derivatives. Enzymatically Catalyzed Studies. The results from the commercial mixtures of GSH transferases suggest that commercial samples vary considerably with respect to the proportions of the individual isoenzymes contained within each sample. This is reflected by the different distribution of product isomers for a particular batch of enzymes. This variation is not surprising since isoenzyme levels in rat livers and their distribution are tissue characteristic and vary with developmental stage and exposure to hormones and other inducers. It is interesting to note that GST 1-1exhibits some stereospecificity toward the cis isomer DEM and GST 4-4 toward both DEM and DEF. In contrast, GSH-transferase 1-1did not lead to any stereospecificity in the reaction with DEF. Detailed studies by Boehlert and Armstrong (47) and Cobb et al. (48) have shown that GST 4-4exerts pronounced stereoselectivity in conjugation reactions of arene and azaarene oxides. In contrast, the 3-3 isoenzyme was found to be less stereoselective in these
‘HNMR Studies of Glutathione Conjugation I I
H, Q t - O E t E‘Y
1
Chem. Res. Toxicol., Vol. 8, No. 5, 1995 789
H-C-?
H
I
GS
H
E
H
E
I I H-C--C-C-OEW I II DEM
L
GS
0
GS
GS-
E I H-C-c
GA
H
I I H-C-C=C-OEt I I
0-
-
W2H
/
-W~H
C ‘
-OEt
II
0
Figure 6. Possible mechanism for the conjugation of GSH with DEM, where E is COOCH2CH3 and X is probably OH.
reactions, and the substrate binding cavity has recently been defined (49). Our results offer some insight to the active site of GSHtransferases. It is well-known that GSH-transferases interact specifically with GSH at a discrete site called the G-site (5, 50, 38, 51, 521, involving a network of specific polar interactions with amino acid side chains of the protein (9). The G-site has a relatively high specificity, demonstrated by the small number of compounds which can substitute for GSH as a substrate or act as inhibitor by binding at this site (52). A second substrate binding site (the H-site) is adjacent to the G-site; it is hydrophobic and is responsible for the distinctly different electrophilic substrate specificities of the various GSH-transferase subunits (9,50,51). Thus, although the enzymes are highly specific in their requirement for GSH, they can use a wide variety of dissimilar second substrates. Since a limited number of isoenzymes utilize a vast range of substrates, then the requirements of the substrate binding site may not be too stringent. Those substrates which possess a sufficiently reactive electrophilic center and are capable of binding to the H-site are subject to nucleophilic attack by bound GSH or GS-. During their studies on the isomerization of maleate to fumarate, Keen and Jakoby (51) found that GSH-transferases were inactive toward this cis-trans isomerization. They suggested that this was due to the inability of maleate to bind to the hydrophobic site of the enzyme due to its highly charged character. This explanation may also account for the apparent inability of the enzyme to catalyze the conjugation of GSH with maleate (or fumarate) (vide supra). Alternatively, maleate may bind competitively at the G-site. Reported mechanistic studies of GSH transferasecatalyzed reactions have been inconclusive. It has been suggested that the enzymic reaction of the GSH-transferases may only be the result of bringing the two substrates into close proximity (53). In contrast, it has also been suggested that the GSH-transferases serve to increase the nucleophilicity of the -SH group (51,54). There has been some spectroscopicevidence that suggests
that the thiol pKa of GSH bound to rat GSH-transferase 4-4 is at least two pKa units more acidic than that of unbound GSH. Thus, GSH is bound to the enzyme as a thiolate anion (55),which is now known to be stabilized by H-bonding to an active-site Tyr residue (9). The inhibition of isomerization of DEM to DEF by the enzyme, coupled with the introduction of some stereospecificity of deuterium attachment, suggests that the enzyme not only lowers the pKa of GSH but also binds the substrate in a specific orientation. These results would appear to support the presence of a substrate binding site which holds the substrate in a specific orientation. This could restrict rotation about the CaCP bond of the alkenyl moiety and lead to stereospecificity of deuterium attachment. Alternatively, the H atom donor may be positioned at the H-site with a particular orientation relative to the substrate, and the stereospecificity may be the result of an enhanced rate of H attachment, rather than the result of “nonrotation” of the substrate. Differences in substrate binding at the H-site andor orientation of the H-site toward the G-site may explain the observed stereoselectivities for conjugation with DEM and DEF, as well as the differences in stereospecificity between isoenzymes. For example, in one isoenzyme, DEM may be bound in the H-site in one orientation and DEF may be bound in the opposite orientation, while the orientation of one or both substrates may be reversed in another isoenzyme. Boyland and Chasseaud (35)carried out heat inactivation studies which suggested that the reactions of GSH with DEM and DEF were catalyzed by different enzymes. Our results would appear to support their conclusions. Our results are consistent with the model for enzyme action presented by Keen and Jakoby (51). The isoenzymes appear to interact specifically with GSH at a discrete site and bind the second substrate at an adjacent hydrophobic locus. Thus, while DEM and DEF, which possess sufficiently reactive electrophilic centers and are sufficiently hydrophobic to bind to the H-site, are effective substrates for GSH-transferase, maleate and fumarate, which are less hydrophobic, are not. It would appear
790 Chem. Res. Toxicol., Vol. 8, No. 5, 1995
from our results that both GSH and substrate may be bound to the enzyme in a fixed orientation during conjugation. In addition to the stereo control, this would explain the inhibition of isomerization since, in the bound state, the substrate may not be able to rotate before attachment of the hydrogen. Alternatively, these results can be explained if the site of hydrogen atom addition has a fxed orientation relative to the substrate, and hydrogen addition is fast compared to rotation about the Ca-CP bond. Finally, the ability of GSH-transferases to catalyze isomerization reactions is well-known (51,56, 57). The positional isomerization of double bonds in the conversion of A5-3-ketosteroids to A4-3-ketosteroids and the conversion of maleate analogues to their trans isomers are reported to occur at a faster rate in the presence of GSHtransferases (57). However, as described here, commercial GSH-transferases did not significantly catalyze the geometric isomerization of maleate to fumarate or of DEM to DEF. Indeed, our studies on the catalyzed . conjugation of GSH with DEM and maleate showed that, at the highest GSWGSH-transferase ratios used (1.5 mg/5 mmol of GSH), isomerization of DEM was almost completely inhibited. In contrast, using human GSTA12, the isomerization of DEM to DEF was significantly enhanced when compared to the control reaction. The reason for this difference is not understood but may be due to the high isomerase activity with A5-3-ketosteroids, which is characteristic of the basic human GSH-transferases (57).
Conclusions The direct conjugation of GSH with DEM and DEF results in a racemic mixture of conjugation products, while the presence of certain GSH-transferase isoenzymes exerts some stereo control over conjugation. The hydrogen atom which is added during conjugation apparently arises from the solvent (or is solvent exchangeable). Significant amounts of DEM isomerize to DEF during nonenzymatic conjugation, and this isomerization is inhibited by the presence of certain GSH-transferase isoenzymes. Maleate and fumarate conjugate with GSH relatively slowly, and GSH-transferases have no apparent effect on this reaction, and thus it would appear that if DEM is rapidly hydrolyzed to maleate in vivo, it is no longer available to conjugate with GSH. We have used NMR spectroscopy to investigate the conjugation of maleate and fumarate derivatives in intact liver cells (~581,~ and these results will be reported elsewhere.
Acknowledgment. We thank the SERC, MRC, CRC, and SmithKline Beecham for their support for this work. We also thank Professor B. Ketterer for helpful discussion and Drs. M. Pue and D. Ross (SKB)for their encouragement with these studies. References (1) Boyland, E., and Chasseaud, L. F. (1969) The role of glutathione
and glutathione S-transferases in mercapturic acid biosynthesis. Adv. Enzymol. 32, 173-219. (2) Jakoby, W. B., Ed. (1981) Detoxification and drug metabolism: conjugation and related systems. Methods Enzymol. 77. (3) Douglas, K. T. (1988) Reactivity of glutathione in model systems for glutathione S-transferases and related enzymes. In Glutathione Conjugation: Mechanisms and Biological Significances (Sies, H., and Ketterer, B., Eds.) pp 1-41, Academic Press, London.
Kubal et al. (4) Douglas, K. T. (1989) Chemical synthesis of glutathione and analogues. In Glutathione: Chemical, Biochemical and Medical Aspects (Dolphin, D., Poulson, R., and Avramovic, O., Eds.) Part A, pp 243-279, John Wiley and Sons, New York. (5) Sies, H., and Ketterer, B., Eds. (1988) Glutathione Conjugation: Mechanisms and Biological Significance, Academic Press, London. (6) Jakoby, W. B., Ketterer, B., and Mannervik, B. (1984)Glutathione transferases: nomenclature. Biochem. Pharmacol. 33, 25392540. (7) Mannervik, B. (1985) The isozymes of glutathione transferase. Adv. Enzymol. Relat. Areas Mol. Biol. 57, 357-417. ( 8 ) Mannervik, B., Awasthi, Y. C., Board, P. G., Hayes, J. D., Di Ilio, C., Ketterer, B., Listowsky, I., Morgenstern, R., Muramatsu, M., Pearson, W. R., Pickett, C. B., Sato, K., Widersten, M., and Wolf, C. R. (1992) Nomenclature for human glutathione transferases. Biochem. J . 282, 305-308. (9) Dirr, H., Reinemer, P., and Huber, R. (1994) X-ray crystal structures of cytosolic glutathione S-transferases. Implications for protein architecture, substrate recognition and catalytic function. Eur. J . Biochem. 220, 645-661. (10) Hayes, J. D., McLellan, L. I., Stockman, P. K., Chalmers, J., and Beckett, G. J. (1987) Glutathione S-transferases in man: the relationship between rat and human enzymes. Biochem. SOC. Trans. 15, 721-725. (11) Boyland, E., and Chasseaud, L. F. (1967) Enzyme-catalysed conjugations of glutathione with unsaturated compounds. Biochem. J . 104, 95-102. (12) Boyland, E., and Chasseaud, L. F. (1970) The effect of some carbonyl compounds on rat liver glutathione levels. Biochem. Pharmacol. 19, 1526-1528. (13) Plummer, J. L., Smith, B. R., Sies, H., and Bend, J. R. (1981) Chemical depletion of glutathione in vivo. Methods Enzymol. 77, 50-59. (14) Costa, L. G., and Murphy, S. D. (1986) Effect of diethylmaleate and other glutathione depletors on protein synthesis. Biochem. Pharmacol. 35, 3383-3388. (15) Mulder, G. J., Kadlubar, F. F., Mays, J . B., and Hinson, J. A. (1984) Reaction of mutagenic phenacetin' metabolites with glutathione and DNA. Possible implications for toxicity. Mol. Pharmacol. 26, 342-347. (16) Naylor, S., Mason, R. P., Sanders, J . K. M., and Williams, D. H. (1988) Formaldehyde adducts of glutathione. Biochem. J . 249, 573-579. (17) Liebler, D. C., Latwesen, D. G., and Reeder, T. C. (1988) S42Chloroacetyl)glutathione, a reactive glutathione thiol ester and a putative metabolite of 1,l-dichloroethylene. Biochemistry 27, 3652-3657. (18) Dohn, D. R., Quebbemann, A. J., Borch, R. F., and Anders, M. W. (1985) Enzymatic reaction of chlorotrifluoroethene with glutathione: 19FNMR evidence for stereochemical control of the reaction. Biochemistry 24, 5137-5143. (19) Rabenstein, D. L. (1989)Metal complexes of glutathione and their biological significance. In Glutathione: Chemical, Biochemical, and Medical Aspects (Dolphin, D., Poulson, R., and Avramovic, O., Eds.) Part A, pp 147-186, John Wiley and Sons, New York. (20) Rabenstein, D. L., and Keire, D. A. (1989) Nuclear magnetic resonance spectroscopy of glutathione. In Glutathione: Chemical, Biochemical, and Medical Aspects (Dolphin, D., Poulson, R., and Avramovic, O., Eds.) Part A, pp 67-102, John Wiley and Sons, New York. (21) Kadima, W., and Rabenstein, D. L. (1990) Nuclear magnetic resonance studies of the solution chemistry of metal complexes. 26. Mixed ligand complexes of cadmium, nitrilotriacetic acid, glutathione and related ligands. J . Inorg. Biochem. 38,277-288. (22) Brown, F. F., Campbell, I. D., Kuchel, P. W., and Rabenstein, D. C. (1977) Human erythrocyte metabolism studies by 'H spin-echo NMR. FEBS Lett. 82, 12-16. (23) Nicholson, J. K., Timbrell, J . A,, Bales, J . R., and Sadler, P. J. (1985) A high resolution proton nuclear magnetic resonance approach to the study of hepatocyte and drug metabolism. Mol. Pharmacol. 27, 634-643. (24) Glasoe, P. K., and Long, F. A. (1960) Use of glass electrodes to measure acidities in deuterium oxide. J . Phys. Chem. 64, 188190. (25) Beale, D., Meyer, D. J., Taylor, J. B., and Ketterer, B. (1983) Evidence that the Yb subunits of hepatic glutathione transferase represent two different but related families of polypeptides. Eur. J . Biochem. 137, 125-129. (26) Cmarik, J. L., Inskeep, P. B., Meredith, M. J., Meyer, D. J., Ketterer, B., and Guengerich, F. P. (1990) Selectivity of rat and human glutathione S-transferases in activation of ethylene dibromide by glutathione conjugation and DNA binding and induction of unscheduled DNA synthesis in human hepatocytes. Cancer Res. 50, 2747-2752.
‘HNMR Studies of Glutathione Conjugation (27) Meyer, D. J., Harris, J. M., Gilmore, K. S., Coles, B., Kensler, T. W., and Ketterer, B. (1993)Quantitation of tissue- and sex-specific induction of rat GSH transferase subunits by dietary 1,a-dithiole3-thiones. Carcinogenesis 14, 567-572. (28) Habig, W. H., and Jakoby, W. B. (1981) Glutathione S-transferases (rat and human). Methods Enzymol. 77, 218-231. (29) Akerboom, T. P. M., and Sies, H. (1981) Assay of glutathione, glutathione disulfide, and glutathione mixed disulfides in biological samples. Methods Enzymol. 77, 373-382. (30) Habig, W. H., Pabst, M. J., Fleischner, G., Gatmaitan, Z., Arias, I. M., and Jakoby, W. B. (1974) The identity of glutathione S-transferase B with ligandin, a major binding protein of liver. Proc. Natl. Acad. Sei. U.SA. 71, 3879-3882. (31) Habig, W. H., and Jakoby, W. B. (1981)Assays for differentiation of glutathione S-transferases (Dissociation equilibrium of glutathione. Fourier transform-carbon-13 NMR spectroscopic investigation of the pH-dependence of charge distribution). Methods Enzymol. 77, 398-405. (32) Jung, G., Breitmaier, E., and Voelter, W. (1972) Dissoziantionsgleichgewichte von glutathion. Eur. J . Biochem. 24, 438-445. (33) Fuhr, B. J., and Rabenstein, D. L. (1973) Nuclear magnetic resonance studies of the solution chemistry of metal complexes. E.The binding of cadmium, zinc, lead, and mercury by glutathione. J . Am. Chem. Soc. 95,6944-6950. (34) Meister, A,, and Anderson, M. E. (1983)Glutathione. Annu. Rev. Biochem. 52, 711-760. (35) Boyland, E., and Chasseaud, L. F. (1968) Enzymes catalysing conjugations of glutathione with a,P-unsaturated carbonyl compounds. Biochem. J . 109, 651-661. (36) Sher, W., and Jakoby, W. B. (1969) Maleate isomerase. J . Biol. Chem. 244, 1878-1882. (37) Seltzer, S. (1972) Cis-trans isomerization. In The Enzymes (Boyer, P. D., Ed.) Vol. 6, pp 381-406, Academic, San Diego. (38) Douglas, K. T. (1986) Mechanism of action of glutathionedependent enzymes. Adu. Enzymol. 59, 103-167. (39) Martell, A. E., and Smith, R. M. (1982) Critical Stability Constants, Vol. 5, Plenum Press, New York. (40) Atkins, P. W. (1982) Physical Chemistry, 2nd ed., pp 984-985, Oxford University Press, Oxford. (41) Seltzer, S. (1961)Secondary a-D isotope effect-mechanism of cistrans catalyzed isomerization of maleic acid. J . Am. Chem. SOC. 83, 1861-1865. (42) Martell, A. E., and Smith, R. M. (1979) Critical Stability Constants, Vol. 3, 2nd ed., Plenum Press, New York. (43) Weinander, R., Anderson, C., and Morgenstern, R. (1994) Identification of N-acetylcysteine as a new substrate for rat liver microsomal glutathione transferase. A study of thiol ligands. J . Biol. Chem. 269, 71-76. (44) Habig, W. H., Pabst, M. J., and Jakoby, W. B. (1974) Glutathione S-transferases. The first enzymatic step in mercapturic acid formation. J . Biol. Chem. 249, 7130-7139.
Chem. Res. Toxicol., Vol. 8, No. 5, 1995 791 (45) Penin, D. D. (1979) Stability of Metal-Ion Complexes: Part B, Organic Ligands, Pergamon Press, Oxford. (46) Sergeant, E. P., and Dempsey, B. (1979)Ionisation Constants of Organic Acids in Aqueous Solution, Pergamon Press, Oxford. (47) Boehlert, C. C., and Armstrong, R. N. (1984) Investigation of the kinetic and stereochemical recognition of arene and azaarene oxides by isozymes A2 and C2 of glutathione S-transferases. Biochem. Biophys. Res. Commun. 121, 980-986. (48) Cobb, D., Boehlert, C., Lewis, D., and Armstrong, R. N. (1983) Stereoselectivity of isozyme C of glutathione S-transferase toward arene and azaarene oxides. Biochemistry 22, 805-812. (49) 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 94sglutathiony1)-10-hydroxydihydrophenanthrene. Biochemistry 33, 1043-1052. (50) Mannervik, B. (1985) The isozymes of glutathione transferase. Adv. Enzymol. Relat. Areas Mol. Biol. 57, 357-417. (51) Keen, J. H., and Jakoby, W. B. (1978) Glutathione transferases. Catalysis of nucleophilic reactions of glutathione. J . Biol. Chem. 253,5654-5657. (52) Adang, A. E. P., Brussee, J., Meyer, D. J., Coles, B., Ketterer, B., Van der Gen, A., and Mulder, G. J. (1988) Substrate specificity of rat liver glutathione S-transferase isoenzymes for a series of glutathione analogues, modified at the y-glutamyl moiety. Biochem. J . 255, 721-724. (53) Jakoby, W. B. (1978) The glutathione S-transferases: a group of multifunctional detoxification proteins. Adu. Enzymol. Relat. Areas Mol. Biol. 46, 383-414. (54) 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. (55) Graminski, G. F., Chen, W.-J., and Armstrong, R. N. (1987) Spectroscopic and kinetic detection of the thiolate anion of glutathione in the active site of glutathione transferase. Fed. Proc. 46, 1934. (56) Benson, A. M., Talalay, P., Keen, J. H., and Jakoby, W. B. (1977) Relationship between the soluble glutathione-dependent A5-3ketosteroid isomerase and the glutathione S-transferases of the liver. Proc. Natl. Acad. Sei. U S A . 74, 158-162. (57) Warholm, M., Guthenberg, C., and Mannervik, B. (1983) Molecular and catalytic properties of glutathione transferase p from human liver: an enzyme efficiently conjugating epoxides. Biochemistry 22, 3610-3617. (58) Kubal, G. (1990) NMR studies of hepatocytes and glutathione conjugation, Ph.D. Thesis, University of London.
TX950024X
