Inactivation of gamma-glutamylcysteine synthetase, but not of glutamine synthetase, by S-sulfocysteine and S-sulfohomocysteine.

The reactions catalyzed by gamma-glutamylcysteine synthetase and glutamine synthetase are thought to proceed via enzyme-bound gamma-glutamyl phosphate intermediates. We investigated the possibility that S-sulfocysteine and S-sulfohomocysteine might act as analogs of gamma-glutamyl phosphate or of the associated putative tetrahedral intermediates. The D- and L-enantiomers of S-sulfocysteine and S-sulfohomocysteine were found to rapidly inactivate rat kidney gamma-glutamylcysteine synthetase but to be reversible inhibitors of sheep brain glutamine synthetase. Inactivation of gamma-glutamylcysteine synthetase does not require ATP and is associated with noncovalent binding of close to 1 mol of inactivator/mol of enzyme. The findings indicate that the S-sulfo amino acids are transition-state analogs, and that binding of S-sulfo amino acid to the enzyme induces formation of a very stable enzyme-inactivator complex. The data suggest that stabilization of the enzyme-inactivator complex results from interactions involving the sulfenyl sulfur atom of the S-sulfo amino acid and the active site thiol group of the enzyme.

preliminary account of these findings has appeared (Moore, W., Wiener, H., and Meister, A. (1987) Fed. Proc. 46, 1809) and was presented at the Meeting of the American Society of Biological Chemistry, June 8-12, 1987, Philadelphia, PA. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. NHZOH).
y-Glutamyl phosphate formation is consistent with early "0 tracer studies (11,12), and is supported by kinetic data from That the reaction catalyzed by y-glutamylcysteine synthetase follows a similar pathway is consistent with "0 studies (14, 15) and in accord with the finding that this enzyme also catalyzes reaction 3 and is inactivated by L-methionine-Ssulfoximine (16, 17) (reaction 6).
It has been concluded that both reactions involve formation of a tetrahedral intermediate by interaction of NH, (glutamine synthetase) or the a-amino group of cysteine (y-glutamylcysteine synthetase) with enzyme-bound y-glutamyl phosphate and that L-methionine-S-sulfoximine phosphate formation thus reflects an aspect of the normal catalytic mechanisms (10). L-Methionine-S-sulfoximine phosphate may be viewed as an analog of the tetrahedral transition-state intermediate; it is bound noncovalently to the enzymes and dissociates so slowly that inhibition is essentially irreversible.
Despite these similarities, there are some notable differences between the two enzymes; these include differences in the rates and extents of the cyclization reactions (reaction 3), and different effects of higher homologs of methionine sulfoximine on the two synthetases (10,18,19,41). Several observations indicate that rat kidney y-glutamylcysteine synthetase (in contrast to glutamine synthetase) has a thiol group at or close to its binding site for glutamate (20)(21)(22)(23)(24). This is considered further below.
In the present studies we have examined the effects of a series of Bunte salts, i.e. the S-sulfo derivatives of the L-and D-isomers of cysteine and homocysteine (111) on the two synthetase reactions. These compounds were studied because it was thought that they might act as analogs of y-glutamyl phosphate (I) or of the putative tetrahedral intermediates (IV) involved in these reactions. We found that the S-sulfo  amino acids, which are reversible inhibitors of glutamine synthetase, markedly inactivate rat kidney y-glutamylcysteine synthetase. Inactivation, which does not require ATP, is associated with stoichiometric binding of the inactivators to the enzyme. The inactivators are bound noncovalently and may be released intact from the enzyme.

RESULTS
Inuctiuatwn of y-Glutamykysteine Synthetase by S-Sulfo Derivatives of Cysteine and Homocysteine-The S-sulfo compounds were found not to be substrates (in place of L-glutamate or L-cysteine), as determined by amino acid analysis. When added to the standard reaction mixtures, the S-sulfo compounds were inhibitory; the corresponding apparent Ki values are given in Table I (Miniprint). Inhibition by S-sulfo-L-cysteine was essentially competitive with respect to L-glutamate ( Fig. 2, Miniprint), whereas inhibition by S-sulfo-Lhomocysteine ( Fig. 3, Miniprint) and the corresponding Disomers was mixed. The apparent K,,, value for L-glutamate was found to be 1.6 mM, in agreement with earlier findings (24, 33). The data (Table I) suggest that the affinity of the enzyme for the S-sulfo-L-amino acids is somewhat less than that for L-glutamate. In agreement with earlier studies (20), DL-2-amino-4-phosphonobutyrate was found to inhibit, as also did L-homocysteate and L-homocysteine sulfinate ( Table   1).
When the enzyme was preincubated with the S-sulfo amino compounds it was rapidly inactivated. Inactivation was timedependent ( Fig. 4). Surprisingly, the initial rates of inactiva-Portions of this paper (including "Experimental Procedures," Figs. 2, 3, 5, 6 and 7, and Tables I, 11, and IV) are presented in miniprint at the end of this paper. Miniprint is easily read with the aid of a standard magnifying glass. Full size photocopies are available from the Journal of Biological Chemistry, 9650 Rockville Pike, Bethesda, MD 20814. Request Document No. 87M-2688, cite the authors, and include a check or money order for $6.00 per set of photocopies. Full size photocopies are also included in the microfilm edition of the Journal that is available from Waverly Press. tion were not markedly different for the four S-sulfo compounds; as shown in Fig. 4, the initial rates of inactivation were in the decreasing order S-sulfo-L-cysteine 2 s-SUlfO-Dcysteine > S-sulfo-L-homocysteine 2 S-sulfo-D-homocysteine. The extent of inactivation, after prolonged incubation, was about the same with the four S-sulfo compounds. Notably, preincubation under the same conditions with L-homocysteate, L-homocysteine sulfinate, and DL-2-amino-4-phosphonobutyrate did not produce inactivation.
Inactivation by the S-sulfo compounds was found to be dependent upon the concentration of the inactivator. For example, after preincubation for 1 min with S-sulfo-D-homocysteine at concentrations of 0.1,0.5, and 5 mM, the respective activities were 87, 72, and 11% of the untreated controls. The effects of preincubation of the enzyme with s-SUlfO-Dhomocysteine are given in Fig. 5 (Miniprint). The rates of inactivation were much less (<lo%) in the absence of added divalent metal ion, and substantially greater in the presence of added Mn2+ than with M$+, although the extent of inactivation (>go%) was about the same. In the presence of 10 mM S-sulfo-L-homocysteine and 0.25 mM divalent metal ion, complete inactivation required 1 and 60 min, respectively, with Mn2+ and M e .
Data on protection of the enzyme against inactivation by S-sulfo-L-cysteine by various substrates and products are given in Table I1 (Miniprint). Complete protection was observed with L-y-glutamyl-L-a-aminobutyrate, L-glutamate + ATP, and ADP; similar results were obtained in studies on S-sulfo-D-cysteine and the L-and D-isomers of S-sulfo-homocysteine.
Previous studies showed that treatment of the enzyme with cystamine leads to formation of a mixed disulfide between cysteamine and a thiol at or close to the binding site for glutamate (21). This leads to inactivation of the enzyme, which is readily reversible by treatment with dithiothreitol. Prior treatment of the enzyme with cystamine was found to prevent interaction of the enzyme with chloroketone inacti-vators (~-2-amino-4-oxo-5-chloropentanoate (ZO), the L-and D-isomers of 3-amino-1-chloro-2-pentanone (21), methionine sulfoximine (21), and 4-methylene-D-glutamate (24). In the present studies we found that prior treatment of the enzyme with cystamine prevents its interaction with the S-sulfo amino acids. Thus, treatment with cystamine followed by treatment with S-sulfo amino acid led to an enzyme preparation that was completely reactivated by dithiothreitol (Table 111, experiments 4 and 6 ) , whereas enzyme inactivated by treatment with S-sulfo amino acids (experiments 3 and 5) was not reactivated by treatment with dithiothreitol. These studies indicate that the S-sulfo amino acids inactivate yglutamylcysteine synthetase in a manner which is not reversed by treatment with dithiothreitol and suggest that the inhibitors bind to the active center of the enzyme in the regions of the glutamate and ATP-binding sites.
Stoichiometry and Nature of Binding-When y-glutamylcysteine synthetase was incubated with [35S]sulfenyl-labeled sulfo-L-cysteine and then subjected to gel filtration (Fig. 6, Miniprint), the enzyme fraction (completely separated from the low molecular weight fraction) was found to be enzymatically inactive and to contain radioactivity equivalent to 1.1 mol of S-sulfo-L-cysteine/mol of enzyme. In a duplicate experiment, a value of 0.93 mol/mol of enzyme was found. In an analogous experiment in which the sulfonyl sulfur atom of S-sulfo-L-cysteine was labeled with 35S, the binding of inactivator to the enzyme was found to be equivalent to that found in the experiments with the sulfenyl 35S-labeled compound. The findings therefore suggest that the entire S-sulfo-L-cysteine molecule binds to the enzyme. The findings argue strongly against mixed disulfide formation between inhibitor and an enzyme thiol; this is also in accord with the finding that dithiothreitol does not reactivate the enzyme.
To further explore the nature of the bound inhibitor, the enzyme was inactivated by incubation with a preparation of S-sulfo-L-cysteine in which the sulfenyl and sulfonyl sulfur atoms were equivalently labeled with 35S. The enzyme-inhibitor complex was isolated by gel filtration as described above. Attempts to remove the labeled inhibitor from the enzyme by incubation in 50 mM Tris (pH 8.2) and MgC12 (5 mM) at 37 "C were unsuccessful. However, when the labeled enzyme was incubated at 37 "C in Tris (50 mM In experiment 4, cystamine was added 2 min prior to addition of Ssulfocysteine. In experiment 6, cystamine was added 2 min prior to addition of S-sulfohomocysteine. After incubation of the complete reaction mixtures for 120 min at 37 'C, a portion (10 pl) was removed and assayed for enzyme activity in the presence and absence of 1 mM dithiothreitol. enzyme activity was observed. After incubation for 20 h about 13% of the original enzyme activity returned. The enzyme could not be further reactivated by prolonging the incubation time because of substantial denaturation.

Experiment
The 35S label could readily be removed from the enzyme under denaturing conditions; for example, by placing the enzyme solution at 100 "C for 5 min or by treating it with 18% trichloroacetic acid. After denaturation, all of the radioactivity was readily separated from the protein by gel filtration on Sephadex G-50 or by use of a Centricon-10 microconcentrator. The radioactive material was shown to be S-sulfocysteine by several procedures. (a) The radioactive material comigrated with authentic S-sulfocysteine on thin layer chromatography. (b) The radioactive material did not bind to a column of AG-50 (H+). (c) After treatment of the radioactive material with NaBH,, followed by its application to a column of AG-50 (H+), half of the radioactivity applied to the column washed through the column on elution with water. The other half of the radioactivity eluted with 2 M NH,OH. The findings are in accord with the predicted (34) reduction of S-sulfocysteine to sulfite and cysteine; this was confirmed in the present studies in experiments in which authentic S-sulfocysteine was treated with NaBH,. Inactivation of y-Glutamylcysteine Synthetase by S-Sulfo-Homocysteine-y-Glutamylcysteine synthetase was incubated with 10 mM S-sulfo-~-~'~C]homocysteine, and the enzyme was then isolated by gel filtration as described above. The enzyme, which was 90% inactivated, was found to bind 0.84 mol of inhibitor/mol of inactive enzyme; a value of 0.81 was obtained in a duplicate experiment. When the enzyme was inactivated by incubating it in a solution that contained 10 mM S-sulfo-DL-homocysteine containing an excess of 5.1% of S-sulfo-~-['~C]homocysteine, the isolated, labeled, inactivated enzyme contained 0.48 mol of S-sulfo-~-['~C]homocysteine per mol of enzyme. A value of 0.43 would be expected if the D-and L-isomers combine with and dissociate from the enzyme at the same rates. Studies cited above (Fig. 4) indicate that the L-and D-iSOInerS inactivate the enzyme at very similar rates. It may therefore be concluded that the enantiomers of S-sulfohomocysteine bind at the same enzyme site.
Inhibition of Glutamine Synthetase by the S-Sulfo Amino Acids-In contrast to the results obtained with y-glutamylcysteine synthetase, preincubation of glutamine synthetase with the S-sulfo derivatives of L-and D-cysteine and L-and D-homocysteine did not lead to inactivation, even after 3 h of preincubation. When added to standard assay mixtures the S-sulfo amino acids of the L-configuration produced inhibition that is competitive with respect to L-glutamate (Table  IV, Miniprint). The inhibition produced by the D-isomers is mixed with respect to L-glutamate suggesting that the Disomers bind in a manner which is different from the Lisomers; cf. Ref. 24.

DISCUSSION
We initially considered the possibility that inactivation of y-glutamylcysteine synthetase by S-sulfo amino acids was the result of covalent modification of the enzyme. However, failure of dithiothreitol to reactivate, and the data obtained with the sulfonyl-and ~ulfenyl-~~S-labeled compounds show clearly that inactivation is associated with tight noncovalent binding of the inactivator to the enzyme. Inactivation by S-sulfo amino acids thus resembles that produced by methionine sulfoximine phosphate (and its higher homologs), except that ATP is not required. However, S-sulfo amino acids do not inactivate glutamine synthetase whereas methionine sulfoximine (in the presence of MgATP) inactivates both glutamine synthetase and y-glutamylcysteine synthetase. The potential in uiuo usefulness of S-sulfo amino acids as selective inactivators of y-glutamylcysteine synthetase seems worthy of note; studies on this possibility are in progress. Other in vivo interactions are suggested by the reports that S-sulfocysteine is found in the urine of patients with sulfite oxidase deficiency (35) and that this compound may serve as an excitatory neurotransmitter (36, 37).
Inactivation of y-glutamylcysteine synthetase is time-dependent and is also dependent upon concentration of inactivator. The possibility that inactivation is due to slow formation of an enzyme-inhibitor complex (i.e. E1 formation is slower than ES formation) is unlikely because in such a mechanism the initial velocity of the enzyme-catalyzed reaction would be expected to be independent of concentration of inactivator. The experimental findings seem to support a mechanism of inactivation in which E T is formed rapidly and is then converted to a different form (EI*). According to this mechanism (Scheme I), which has been termed "slow-binding inhibition" (38-40), an initial complex (EI) is rapidly formed, and undergoes slow (relative to the rate of E1 formation) isomerization to a more stable enzyme-inhibitor complex (EI*), from which the inhibitor may be released, but at a very slow rate. In this scheme k3 and k4 are much smaller than kl, kz, or the rate constants involved in the catalytic reaction; for inactivation of the enzyme to occur, k4 must be much smaller than k3. The K, value determined from initial velocity studies in the presence of inhibitor is k2/kl. K;*, the overall dissociation constant of inhibitor from EI*, is Kik4/(k3 + k4). The rate constant (k3) for the isomerization of E1 to EI* can be determined by measuring the effect of varying inhibitor concentration on the rate of inactivation. From such data it is possible to estimate k, by extrapolation (Fig. 7, Miniprint); a value of 2.5 min" was obtained. An estimate of k4 may be obtained by following the regain of enzyme activity from EI*. Thus, enzyme which had been inactivated by S-sulfo-L-cysteine was reactivated by incubation with substrates; about 13% of the initial enzyme activity returned after 20 h from which k, may be estimated to be about 1 X min-'. This corresponds to a half-life of about 4 days for the S-sulfo-L-cysteine-enzyme complex. Ki* for this complex is thus estimated to be about 1 x 10-7 M.
Such slow binding inhibition of enzymes has been reported with adenosine deaminase (42), ribulosebisphosphate carboxylate-oxygenase (43), and alanine racemase (44). It has been postulated that in each instance, the initially formed E1 complex in some manner perturbs the enzyme sufficiently to produce an EI* complex that binds to the enzyme more tightly than EI. The chemical nature of these interactions is not yet clear.
The present studies offer an approach to a mechanistic understanding of the way in which S-sulfo compounds inhibit y-glutamylcysteine synthetase. Thus, the S-sulfo amino acids appear to act as transition-state analogs which resemble the high-energy metastable y-glutamyl phosphate intermediate. S-sulfo-L-cysteine is about lo4 smaller than the K, value for L-glutamate).
It is significant that homocysteate (-0,s-CH,-CH,-CH(NH3+)COO-) does not inactivate y-glutamylcysteine synthetase. Thus, the sulfenyl sulfur atom of S-sulfocysteine seems to be essential for inactivation since its replacement by a methylene moiety abolishes ability to inactivate. Since the methylene group and the sulfur atom are of similar size and hydrophobicity, it would appear that stabilization of the inactivator-enzyme complex involves the sulfenyl sulfur atom of the inactivator. We suggest that the active site enzyme thiol group plays a role in this interaction.' The present findings lead to the tentative conclusion that the sulfenyl sulfur atom of the thiosulfate moiety of the inactivator, by virtue of its high polarizability, promotes rearrangement of the enzyme-bound thiosulfate group; the resulting resonance structure appears to be stabilized by the enzyme (Fig. 8). Such a structure would be expected to have greater net negative charge on the thiosulfate oxygen atoms and a greater positive charge on the sulfenyl sulfur atom than found in the ground state. The enzyme active site thiol might interact through the thiosulfate oxygen atoms or the sulfenyl sulfur atom. It is relevant to note that the difference in free energy between E1 and EI* is about 5.7 kcal/mol (from AG = -RT In k3/k4) and thus similar to that for a hydrogen bond. These possibilities require further study.
Biol. Chem. 237,2932Chem. 237, -2940 'It is relevant to note that a y-glutamylcysteine synthetase of bacterial origin has recently been isolated which does not have an active site thiol moiety (47); interestingly, the purified bacterial enzyme is not inactivated by S-sulfo amino acids. That the bacterial enzyme is, like the mammalian enzyme, inhibited by buthionine sulfoximine in the presence of ATP, suggests that the binding of Ssulfo amino acids and that of buthionine sulfoximine are not identical phenomena, and emphasizes the role of the enzyme active site thiol in the binding of S-sulfo amino acids.