Affinity Labeling of Rat Glutathione S -Transferase Isozyme 1-1 by 17 b -Iodoacetoxy-estradiol-3-sulfate*

Rat liver glutathione S -transferase, isozyme 1-1, catalyzes the glutathione-dependent isomerization of D 5 -an-drostene-3,17-dione and also binds steroid sulfates at a nonsubstrate inhibitory steroid site. 17 b -Iodoacetoxy-estradiol-3-sulfate, a reactive steroid analogue, pro-duces a time-dependent inactivation of this glutathione S -transferase to a limit of 60% residual activity. The rate constant for inactivation ( k obs ) exhibits a nonlinear de- pendence on reagent concentration with K I 5 71 m M and k max 5 0.0133 min 2 1 . Complete protection against inactivation is provided by 17 b -estradiol-3,17-disulfate, whereas D 5 -androstene-3,17-dione and S -methylgluta-thione have little effect on k obs . These results indicate that 17 b -iodoacetoxy-estradiol-3-sulfate reacts as an affinity label of the nonsubstrate steroid site rather than of the substrate sites occupied by D 5 -androstene-3,17-dione or glutathione. Loss of activity occurs concomi-tant with incorporation of about 1 mol 14 C-labeled rea-gent/mol enzyme dimer when the enzyme is maximally inactivated. Isolation of the labeled peptide from the chymotryptic digest shows that Cys 17 is the only enzy-mic amino acid modified. Covalent modification of Cys 17 neutralized m loss radioactivity to HPLC. fractionated by a Varian 5000 LC equipped with a Vydac C 18 reverse-phase column equilibrated with Solvent A (0.1% trifluoroacetic acid in At a flow rate of 1 ml/min, the peptides were separated by a linear gradient from 0% to 20% Solvent B (0.1% trifluoroacetic acid in acetonitrile) in 100 min followed by a linear gradient to 100% Solvent B in 30 min. The eluate was monitored by A 220 , and 1-ml fractions were collected. An aliquot (300 m each was added to 5 ml of Liquiscint to test for radioactivity. Determination of Separated se- quences purified were determined on an Applied Biosystems model 470A phase sequencer, equipped with a 120A phenylthiohydantoin analyzer. Molecular modeling conducted using Simulations, Inc. on

Glutathione S-transferases (GST) 1 (EC 2.5.1.18) constitute a family of detoxification enzymes that are involved in the metabolism of endogenous and xenobiotic compounds (1)(2)(3)(4). They catalyze the conjugation reaction of glutathione to a wide variety of electrophilic substrates. These conjugation products are more water-soluble than the xenobiotic substrates, and they can be further degraded or transported out of the cell. Glutathione S-transferases have been found in elevated levels within cancerous tumors and have been implicated in the development of resistance to anti-cancer drugs (5). The cytosolic enzymes are now grouped into seven classes and within a particular class they can exist as either homo-or heterodimers (1). There are crystal structures to represent most of the classes (6 -12). Each subunit of the dimer contains a glutathione-binding site and a xenobiotic site that can accommodate a wide variety of compounds.
Isozyme 1-1, 2 a member of the ␣ class, efficiently catalyzes the isomerization reaction of ⌬ 5 -androstene-3,17-dione to ⌬ 4androstene-3,17-dione, which it binds at the substrate steroid site (13). In addition to this site, isozyme 1-1 also has a nonsubstrate steroid binding site that is located in the cleft between the two subunits (14,15). This site has been proposed to fulfill a transport function (5) or to act in controlling levels of steroids in target organs (16). The nonsubstrate site has a preference for steroid sulfates, which is illustrated by the more potent inhibitory effect of 17␤-estradiol-3,17-disulfate as compared with that of 17␤-estradiol. However, previous work in this laboratory (aimed at locating the nonsubstrate site) used the affinity label 3␤-(iodoacetoxy)dehydroisoandrosterone (3␤-IDA) (shown in Fig. 1), which is structurally related both to substrates of the enzyme, such as ⌬ 5 -androstene-3,17-dione, and to inhibitors of the enzyme, such as ⌬ 5 -androstene-3␤,17␤diol disulfate and 17␤-estradiol-3,17-disulfate. The 3␤-IDA modified Cys 17 and Cys 111 equally with an incorporation of 1 mol of reagent/mol enzyme subunit; analysis of molecular models suggested that the binding site of 3␤-IDA is located in the cleft between the subunits (15). Based on the previous data, we have now designed a more specific affinity label for the nonsubstrate steroid site: 17␤-iodoacetoxy-estradiol-3-sulfate (17␤-IES). This new compound features the negatively charged sulfate that should enhance and direct its binding and a reactive iodoacetoxy group at a position at the opposite end of the molecule from that of 3␤-IDA (Fig. 1). The iodide can be displaced from the iodoacetoxy group by nucleophilic attack by the side chains of several amino acids including Cys, Asp, Lys, Met, and His (17). In this paper, we demonstrate that this affinity label reacts specifically with Cys 17 at a single subunit of the enzyme dimer. Molecular modeling studies support the location of the nonsubstrate binding site within the cleft and the contribution of the sulfate moiety in orienting the ligand within the cleft. A preliminary version of this work has been presented (18).
Synthesis of 17␤-Iodoacetoxy-estradiol-3-sulfate-17␤-IES was synthesized from 17␤-estradiol-3-sulfate and iodoacetic acid by procedures based on the method of Pons et al. (21). One molar equivalent of 17␤-estradiol-3-sulfate, 1.1 molar equivalents of iodoacetic acid, and 2 molar equivalents of dicyclohexylcarbodiimide were combined in 15 ml of cellosolve. (For the radioactively labeled compound before addition to the reaction mixture, 125 Ci of radioactive iodoacetic acid was added to 0.83 mmol of unlabeled iodoacetic acid in a total of 5 ml.) The reaction was initiated by the addition of a catalytic amount of pyridine (250 l), and the reaction mixture was allowed to stir at room temperature for 1.5 h. The reaction was stopped by the addition of 3 ml of distilled water, and the mixture was centrifuged to remove the insoluble dicyclohexylurea. The organic layer, containing 17␤-IES, was lyophilized. The product was resuspended in 100 l of acetonitrile and was brought to a final volume of 1 ml by the addition of distilled water.
The 17␤-IES was purified by HPLC using a Varian 5000LC equipped with a Vydac C 18 column (1 ϫ 25 cm) and a UV-100 detector. The solvent system used was H 2 O (Solvent A) and acetonitrile (Solvent B). The column was equilibrated with solvent A containing 10% solvent B. After 10 min at 10% solvent B, a linear gradient was run to 100% B in 90 min at a flow rate of 1 ml/min. The effluent was monitored at 275 nm and 17␤-IES eluted at ϳ28 min. For comparison, the starting material, 17␤-estradiol-3-sulfate, elutes at ϳ23 min.
For the radioactively labeled compound, the specific radioactivity was 2.17 ϫ 10 11 cpm/mol. The product has a UV absorption spectrum with a maximum at 260 nm and a shoulder at 270 nm. The extinction coefficient at 260 nm was measured to be 1810 M Ϫ1 cm Ϫ1 , with the concentration determined from the specific radioactivity.
Reaction of 17␤-IES with Glutathione S-transferase, Isozyme 1-1-Glutathione S-transferase (0.2 mg/ml, 7.8 M enzyme subunits) was incubated in 0.1 M potassium phosphate buffer, pH 7.0, at 37°C with various concentrations of 17␤-IES. Control enzyme samples were incubated under the same conditions but without 17␤-IES. At various time points, an aliquot was removed from the incubation mixture, diluted, and assayed (30 l) for residual activity.
Measurement of Incorporation of 17␤-IES into Glutathione S-Trans-ferase-Glutathione S-transferase (0.2 mg/ml) was incubated with 500 M [ 14 C]17␤-IES at pH 7.0 under standard reaction conditions. Aliquots were withdrawn at various times, and excess reagent was removed by the gel centrifugation method using two successive Sephadex G-50 columns (5 ml) equilibrated with 0.1 M potassium phosphate buffer, pH 7.5 (23). The protein concentration in the filtrate was determined using the Bio-Rad protein assay, based on the Bradford method, using a Bio-Rad 2550 RIA plate reader with a 600-nm filter (24). Unmodified GST 1-1 was used to generate the standard concentration curve. The amount of reagent present was determined by radioactivity using a Packard 1500 Liquid scintillation counter. Incorporation was expressed as mol 17␤-IES/mol of enzyme subunit. Preparation and Separation of Proteolytic Digest of Modified Glutathione S-Transferase-Glutathione S-transferase (0.2 mg/ml) was incubated with 500 M 14 C-labeled 17␤-IES at pH 7.0 under standard reaction conditions for 3 h, at which time the enzyme was maximally inactivated. Excess reagent was removed as described above. Solid guanidine HCl was added to make a 5 M guanidine-HCl solution and was incubated for 1 h at 37°C to denature the protein, followed by treatment with 10 mM N-ethylmaleimide at 25°C for 30 min to block free cysteine residues. The solution was then dialyzed against 6 liters of 10 mM ammonium bicarbonate, pH 8, at 4°C with one change for a total of 18 h, after which the sample was lyophilized.
The enzyme was solubilized by adding 250 l of 8 M urea in 10 mM ammonium bicarbonate, pH 8.0, and incubating at 37°C for 1 h. The solution was then diluted with 10 mM ammonium bicarbonate to bring the final concentration of urea to 2 M. Chymotrypsin was added (10% w/w) at 2 h intervals while incubating at 37°C. The ester bond between the iodoacetic acid and estradiol-3-sulfate was subsequently hydrolyzed by adding 2 N NaOH to yield 0.2 N NaOH and then incubating the enzyme digest at 25°C for 2 h. The solution was then neutralized by adding HCl to yield 0.2 N. The solution was filtered through a 0.45 M filter, with no loss of radioactivity and was subjected to HPLC.
The chymotryptic peptides were fractionated by a Varian 5000 LC equipped with a Vydac C 18 reverse-phase column equilibrated with Solvent A (0.1% trifluoroacetic acid in water). At a flow rate of 1 ml/min, the peptides were separated by a linear gradient from 0% to 20% Solvent B (0.1% trifluoroacetic acid in acetonitrile) in 100 min followed by a linear gradient to 100% Solvent B in 30 min. The eluate was monitored by A 220 , and 1-ml fractions were collected. An aliquot (300 l) from each fraction was added to 5 ml of Liquiscint to test for radioactivity.
Sequence Determination of Separated Peptides-The amino acid sequences of purified peptides were determined on an Applied Biosystems model 470A gas phase protein/peptide sequencer, equipped with a model 120A phenylthiohydantoin analyzer.
was constructed as described previously (19) based on the known crystal structure of human liver isozyme 1-1 (1GUH). The structure of 17␤-IES was constructed using the Builder module. Docking of 17␤-IES was done manually based on the energy minimized structure of 17␤estradiol-3,17-disulfate docked into isozyme 1-1 (15).

RESULTS
Inactivation of Rat Liver Glutathione S-transferase 1-1 by 17␤-IES-Incubation of rat GST 1-1 (0.2 mg/ml, 7.8 M enzyme subunits), with 300 M 17␤-IES, when assayed with 1-chloro-2,4 dinitrobenzene, results in a time-dependent loss of enzyme activity that reaches a limit of 60% of the original activity, as is illustrated in Fig. 2A. After 180 min, excess reagent was removed, and a second addition of 300 M 17␤-IES was added; no further decrease in activity occurred. Because the activity levels off at 60% at long incubation times and over a range of 17␤-IES concentrations, the data were calculated using 60% as the end point (Fig. 2B). Control enzyme incubated under the same conditions but with no reagent present shows no loss of activity. The k obs for inactivation was calculated from the slope of ln([E t Ϫ E ϱ ]/[E 0 Ϫ E ϱ ]) versus time where E t is the enzyme activity at time t, E 0 is the original enzyme activity, and E ϱ is the enzyme activity at long times, which is equal to 0.6 (E 0 ). The reaction obeys pseudo-first order kinetics with a rate constant of 0.0125 min Ϫ1 (Fig. 2B).
Concentration Dependence of the Rate of Inactivation-GST 1-1 (0.2 mg/ml, 7.8 M enzyme subunits) was incubated with 20 -300 M of 17␤-IES as described above, to determine the rate of inactivation at various reagent concentrations (Fig. 3). The apparent rate constant k obs exhibits a nonlinear dependence on reagent concentration. This type of curve is typical of an affinity label, suggesting that a reversible enzyme-reagent complex is formed prior to the irreversible modifcation of the enzyme (26). The curve can be described by the equation k obs ϭ k max /(1 ϩ K I /[17␤-IES]), where K I is the apparent dissociation constant of the enzyme-reagent complex, and k max is the maximum rate of inactivation at saturating concentrations of the reagent. A least squares fit of the observed data yields K I ϭ 71.4 M and k max ϭ 0.0133 min Ϫ1 .
Effect of Ligands on the Inactivation Rate of GST 1-1 by 17␤-IES-Various ligand analogues were added to the reaction mixture to determine whether they could protect against the inactivation of the enzyme by 100 M 17␤-IES. The results, given in Table I, are expressed as k ϩL /k ϪL , where k ϩL is the rate constant for inactivation in the presence of a particular ligand, and k ϪL is the rate constant for inactivation in the absence of a particular ligand. Glutathione derivatives (Table I,   17␤-Estradiol-17-sulfate (500 M) 0.14 9.
17␤-Estradiol-3,17-disulfate (500 M) 0.00 a k ϩL /k ϪL was determined by the ratio of initial inactivation rate with ligand present to that observed in the absence of ligand. site but distinct from it. Electrophilic substrates, such as ⌬ 5androstene-3,17-dione (Table I,  Isolation and Characterization of Chymotryptic Peptides from 17␤-IES Modified GST 1-1-Maximally inactivated GST 1-1 was prepared and digested with chymotrypsin. The digest was fractionated by HPLC using a reverse-phase column (C 18 ) equilibrated with 0.1% trifluoroacetic acid and an acetonitrile gradient (Fig. 5). One radioactive peptide peak was observed on HPLC. Because the ester linkage of 17␤-IES (Fig. 1) was hydrolyzed before the digest was applied to HPLC, the steroid moiety was removed, and the peptide is expected to be labeled with the radioactive carboxymethyl group. The fractions corresponding to this peak were pooled, lyophilized, and subjected to gas phase amino acid sequencing. The results are shown in Table II. The sequence Glu-Xaa-Ile-Arg-Trp corresponds to residues 16 -20 in the known amino acid sequence. None of the common phenylthiohydantoin derivatives was detected in cycle 2; instead, there was a peak with a retention time between that of phenylthiohydantoin-Ser and phenylthiohydantoin-Asn. This peak corresponds to that of a phenylthiohydantoin-carboxymethylcysteine standard, indicating that a Cys in this  Trp (14) a Retention time is 7.8 min between PTH-Asn and PTH-Ser.
Labeling of Glutathione S-Transferase by Reactive Steroid position had been modified. Thus, Cys 17 of GST1-1 is the amino acid target of 17␤-IES.

DISCUSSION
17␤-Iodoacetoxy-estradiol-3-sulfate acts as an affinity label of rat liver glutathione S-transferase isozyme 1-1. Upon incubation of the enzyme with 17␤-IES, a time-dependent loss of activity is observed, yielding a maximum loss of 40% of the original activity. The rate of inactivation exhibits nonlinear dependence on reagent concentration, as is typical of an affinity label, for which an enzyme-reagent complex forms prior to irreversible modification. Partial protection against inactivation is provided by glutathione derivatives; long chain derivatives, such as S-hexylglutathione, provide more protection than do shorter chain derivatives, like S-methylglutathione, indicating that 17␤-IES is binding in a site close to the glutathione site but not within the site. Electrophilic substrate analogues, such as ⌬ 5 -androstene-3,17-dione, do not offer any protection, demonstrating that 17␤-IES does not bind within the electrophilic substrate site. Steroid sulfates are most effective in protecting against inactivation of GST, 1-1, with 17␤-estradiol-3,17-disulfate providing complete protection. These results indicate that 17␤-IES is binding and reacting within the nonsubstrate steroid binding site.
In the case of glutathionyl S- [4-(succinimidyl)benzophenone), only one subunit is modified, yet the enzyme is completely inactivated. The modification of one subunit thus can abolish the enzyme activity of both subunits and, because this label does not occupy the nonsubstrate site, the inhibition is probably the result of a subtle conformational change rather than a physical barrier to the binding of the substrate (28). There is also complete inactivation by the aflatoxin conjugate, although in this case, the bound conjugate extends into the cleft and therefore may be inhibiting completely either because it is blocking access to the active site of the unmodified subunit or because it induces a conformational change (25).
In the present case, maximum reaction with 17␤-IES results in the loss of only 40% of activity; it is likely that the unmodified subunit retains full activity, whereas the other subunit with modified Cys 17 is 80% inactive. Incorporation of 17␤-IES on one subunit apparently prevents a second molecule from binding to and reacting with the other subunit, but, in contrast to the previous examples, this does not cause complete inactivation of both subunits. The 17␤-IES reacts at the steroid site, which is distinct from the active site, and thus there is still some residual activity in the modified subunit, whereas the catalytic site on the other subunit functions independently and is completely active. These results indicate that the observation of apparent cooperativity between the subunits of gluta-thione S-transferase depends on the particular binding site that is being examined.
A homology model for the rat 1-1 isozyme was generated from the crystal structure of the human glutathione S-transferase 1-1. The reagent was manually docked into the model based on an energy-minimized structure of 17␤-estradiol-3,17disulfate bound to GST 1-1 and the assumptions that the iodoacetoxy group of the 17␤-IES must be close to the sulfhydryl group of Cys 17 as well as in an orientation to modify only one subunit. The structure shown in Fig. 6 meets these requirements.
In the proposed model, there is 1 mol of 17␤-IES bound in the cleft between the subunits of the enzyme. The reactive iodoacetoxy group is about 3.4 Å from the sulfhydryl group of Cys A17 . This orientation prevents a second molecule from reacting at Cys 17 on subunit B. 17␤-IES appears to bind more specifically than does 3␤-(iodoacetoxy)dehydroisoandrosterone, which modified both Cys 17 and Cys 111 (15). In the case of 3␤-IDA, reaction at the two sites were mutually exclusive, i.e. reaction with Cys 17 on one subunit excludes binding and reaction with Cys 17 on the other subunit. We now propose that the more specific reaction of 17␤-IES with only Cys 17 is due to an interaction between the sulfate group of 17␤-IES and the guanidino group of Arg 14 ; this interaction would orient the reagent within the binding cleft. Based on the model, the charged sulfate group of 17␤-IES is about 3.1 Å from the guanidino group of Arg 14 .
In summary, 17␤-IES functions as an affinity label of the nonsubstrate steroid site of rat liver glutathione S-transferase, isozyme 1-1. Upon incubation with 17␤-IES, the enzyme loses 40% of its activity, incorporates about 0.5 mol of reagent/enzyme subunit, and is modified only at Cys 17 . Protection against inactivation by 17␤-IES is best provided by steroid sulfates, such as 17␤-estradiol-3,17-disulfate, indicating that Cys 17 is within the nonsubstrate steroid binding site of the enzyme and that its binding is more specific than that of 3␤-IDA because of the interaction of the sulfate group with the side chain of Arg 14 . Based on analysis of molecular models, this nonsubstrate site is located within the cleft between the two subunits of the enzyme.