An electron-sharing network involved in the catalytic mechanism is functionally conserved in different glutathione transferase classes.

In Anopheles dirus glutathione transferase D3-3, there are electrostatic interactions between the negatively charged glutamyl alpha-carboxylate group of glutathione, the positively charged Arg-66, and the negatively charged Asp-100. This ionic interaction is stabilized by a network of hydrogen bonds from Ser-65, Thr-158, Thr-162, and a conserved water-mediated contact. This alternating ionic bridge interaction between negatively and positively charged residues stabilized by a network of hydrogen bonding we have named an electron-sharing network. We show that the electron-sharing network assists the glutamyl alpha-carboxylate of glutathione to function as a catalytic base accepting the proton from the thiol group forming an anionic glutathione, which is a crucial step in the glutathione transferase (GST) catalysis. Kinetic studies demonstrate that the mutation of electron-sharing network residues results in a decreased ability to lower the pKa of the thiol group of glutathione. Although the residues that contribute to the electron-sharing network are not conserved in the primary sequence, structural characterizations indicate that the presence of the network can be mapped to the same region in all GST classes. A structural diversification but functional conservation suggests a significant role for the electron-sharing network in catalysis as the purpose was maintained during the divergent evolution of GSTs. This network appears to be a functionally conserved motif that contributes to the "base-assisted deprotonation" model suggested to be essential for the glutathione ionization step of the catalytic mechanism.

Glutathione transferases (GSTs, 1 EC 2.5.1.18) are a superfamily of multifunctional enzymes involved in the cellular detoxification of various physiological and xenobiotic substances (1)(2)(3)(4). The enzyme catalyzes the nucleophilic addition of the glutathione (GSH) sulfhydryl group to electrophilic centers of such organic compounds. The glutathione conjugates formed are rendered more water soluble, thereby facilitating their eventual elimination (3)(4)(5)(6)(7). The catalytic strategy can be di-vided into steps involving binding of substrates to the enzyme active site, activation of GSH by thiol deprotonation, nucleophilic attack by the thiolate at the electrophilic center, product formation, and product release (8 -11).
The cytosolic GSTs have been subdivided into at least 13 distinct evolutionary classes designated Alpha, Mu, Pi, Theta, Sigma, Zeta, Kappa, Omega, Phi, Tau, Delta, Epsilon, and Beta on the basis of their primary structure, immunological properties, and substrate specificities (3,(12)(13)(14)(15)(16)(17)(18)(19). GSTs possess a high specificity for glutathione as the nucleophile but exhibit broad specificity with regard to structurally diverse electrophilic second substrates. Individual classes of GSTs exhibit overlapping but distinct hydrophobic substrate and ligand binding specificities while retaining a high specificity toward the thiol substrate glutathione. Hence, a step of the activation of GSH by deprotonation of the thiol group to enhance the nucleophilicity for reaction with diverse types of electrophilic centers is crucial for the enzymatic catalysis. This thiol deprotonation results from the lowering of the pK a of the thiol group of the enzymebound GSH from 9.0 to between 6.0 and 6.9, which causes a 200 -300-fold rate acceleration at physiological pH (6,20).
Many of the mechanistic investigations of GSTs have emphasized the importance of a conserved tyrosine or serine residue at the glutathione binding site of certain GSTs (Tyr-8 for Pi, Tyr-9 for Alpha, Tyr-6 for Mu, and Ser-9 for Delta class) in facilitating the thiol deprotonation (10,(21)(22)(23). The hydroxyl group of this tyrosine or serine residue that is within hydrogenbonding distance of the thiol group of enzyme-bound glutathione is considered to be required for the correct orientation and stabilization of the deprotonated thiolate anion in the active site (10, 22, 24 -26).
Apart from the foregoing, several experimental observations have implicated an alternative mechanism named "base-assisted deprotonation" (25,27). This model has been proposed on the basis of the function of the glutamyl ␣-carboxylate of glutathione as a catalytic base that accepts the thiol proton from the thiol group. In this regard, it has become apparent that apart from the thiol group, the ␣-carboxylate of the Glu residue is also a crucial functional group of the glutathione contributing to the catalytic mechanism. It has been reported that an alternative thiol substrate in which the glutamyl part of the tripeptide was decarboxylated (GABA-Cys-Gly) displayed a large decrease in the catalytic efficiency of GSTA1-1 using 1-chloro-2,4-dinitrobenzene (CDNB) as the electrophilic substrate (28). The decrease was postulated to be partially due to a raised pK a value of the active site-bound thiol group of glutathione resulting in an inability of the enzyme to promote ionization of decarboxylated GSH. Introduction of a carboxylate in the glutathione binding site of the enzyme in a location generally occupied by the glutamyl ␣-carboxylate partially res-cued the activity lost by the deletion of the ␣-carboxylate from glutathione, as the ability to ionize the thiol was improved.
Recent studies in the Anopheles dirus GSTD3-3 (adGSTD3-3) have provided information bearing on the base-assisted deprotonation model (29). It has been demonstrated that the pK a value of the active site-bound GSH of a mutant, where the Arg-66 that directly interacts with the glutamyl ␣-carboxylate was changed into alanine, was shifted ϳ1 pH unit toward higher pH values. The positively charged Arg-66 would provide a counter ion that helps stabilize the ␣-carboxylate, thereby facilitating ionization of the thiol.
Observations on the active site structure have provided a further insight into the enzyme catalysis (30). The configuration of the glutamyl ␣-carboxylate group of glutathione, together with the G-site residues Ser-65, Arg-66, Asp-100, Thr-158, and Thr-162 makes possible an electron-sharing network for the distribution of a charge that could be in the form of either a proton or an electron. The aim of the present study is to ascertain the validity of this proposed electron-sharing network. Our study shows that the electron-sharing network helps promote the ionization of the thiol group of GSH through the base-assisted deprotonation model.

MATERIALS AND METHODS
Site-directed Mutagenesis-The plasmid pET3a-adgstD3, previously described (31), was used to generate the mutants via PCR-based sitedirected mutagenesis. The proposed residues Ser-65, Arg-66, Asp-100, Thr-158, Thr-162, and Thr-158/Thr-162 were substituted with Ala by using the mutagenic primers that have been designed according to the 5Ј-and 3Ј-sequence of the adgstD3 wild type gene (GenBank TM accession number AF273039). Each mutant was randomly screened by restriction digestion analysis. Mutant plasmids could be distinguished from the template by digestion with the restriction enzyme corresponding to the restriction recognition site introduced by the mutagenic primers. The full-length GST coding sequences of the plasmids carrying S65A, R66A, D100A, T158A, T162A, and T158A/T162A mutations were verified by the dideoxy chain termination method.
Heterologous Expression and Purification-The proteins were expressed from the pET3a-adgstD3 vector in Escherichia coli BL21 (DE3)pLysS. The cells were grown to A 600 ϭ 0.5, and expression was induced by addition of 0.1 mM isopropyl 1-thio-␤-galactopyranoside. Following induction for 3 h, cells were collected by centrifugation and lysed using sonication. The soluble recombinant GST proteins were purified by GSTrap affinity chromatography (Amersham Biosciences) or S-hexylglutathione-agarose (Sigma) affinity chromatography in the case of low affinity toward the glutathione ligand. The protein concentration was determined by the Bradford method using bovine serum albumin as a standard (32). The results showed that all mutant enzymes, S65A, R66A, D100A, Thr158Ala, T162A, and T158A/T162A were successfully expressed in E. coli and purified by affinity chromatography.
Steady-state Kinetics-Steady-state kinetics were studied for wild type and mutant enzymes at varying concentrations of CDNB and GSH in 0.1 M phosphate buffer, pH 6.5. The reaction was monitored at 340 nm, ⑀ ϭ 9600 M Ϫ1 cm Ϫ1 . Apparent kinetic constants, k cat , K m , and k cat /K m were determined by fitting the collected data to a Michaelis-Menten equation by non-linear regression analysis using GraphPad Prism (GraphPad software, San Diego, CA).
pH Dependence of Kinetic Constants-The pH dependence of k cat / K m CDNB was obtained as stated above by recording the enzymatic reaction in the following buffers: 0.1 M sodium acetate buffers (from pH 5.0 -5.5) and 0.1 M potassium phosphate buffer (from pH 6.0 -8.5). The control studies showed that the affinity of the enzyme toward GSH does not change in the pH range utilized. pK a values were obtained by fitting the data to equation y ϭ y lim /(1 ϩ 10 pKa ϪpH) (10).
Fluoride/Chloride Leaving Group Substitution-The second order kinetic constants at pH 6.5 for the spontaneous reaction of GSH with CDNB and FDNB (1-fluoro-2,4-dinitrobenzene) and the catalytic center activities at pH 6.5 for adGSTD3-3 with CDNB and FDNB as cosubstrates were obtained as described previously (33).
Substrate Specificity-The specific activities of the enzymes were determined by spectrophotometer with five different substrates: CDNB, 1,2-dichloro-4-nitrobenzene, ethacrynic acid, PNPBr (p-nitrophenethyl bromide), and p-nitrobenzyl chloride as previously described (34). However, the results showed that there is no detectable activity for all recombinant enzymes toward PNPBr as a cosubstrate.
Viscosity Effect on the Kinetic Parameters-The effect of viscosity on kinetic parameters was obtained by using 0.1 M potassium phosphate buffer, pH 6.5, with varying glycerol concentrations. Viscosity values () at 25°C were calculated as previously described (35).
Half-life Determination-The thermal stability assay was performed to determine half-life of the GST proteins at 45°C. The wild type and mutant enzymes were incubated at 45°C at the protein concentration of 1 mg/ml. The inactivation time courses were determined by withdrawing suitable aliquots at the different time points for assay of remaining activity to calculate half-life of the enzyme.

RESULTS
Steady-state kinetic constants were obtained with varying concentrations of glutathione and CDNB substrate. Michaelis-Menten kinetic analysis was performed using non-linear regression (Table I). All the mutants showed increased K m values for glutathione, except T158A and T162A, which are third sphere residues. In particular, the mutants S65A, R66A, D100A, and T158A/T162A showed K m increases in the range of 5-21-fold. Conversely, no significant differences were found in K m values for CDNB substrate, except for D100A, which was ϳ4-fold higher, when compared with the wild type enzyme.
The mutation effects on k cat values in the nucleophilic aromatic substitution reaction with CDNB were significantly decreased for R66A and D100A, ϳ10and 60-fold, respectively. The remaining enzymes showed the catalytic center activity to be decreased to ϳ20 -50% lower than wild type.
The pH dependence of k cat /K m CDNB should reflect a kinetically relevant ionization of the GST⅐GSH complex. Therefore, an apparent pK a value of 6.36 was determined for the wild type GST enzyme (Fig. 1). Then, to characterize the influence of the proposed electron-sharing network on GSH thiol ionization, the pK a values for these mutant enzymes were measured by this kinetic approach (Table I). Each of the alanine mutations at the proposed network residues exhibited an increase in pK a for the bound GSH ranging from ϳ0.5-1.1 pH unit higher than that found for the wild type. values for the thiol group of GSH of wild type and mutants of adGSTD3-3 for the CDNB conjugation reaction at pH 6.5 and 25°C The enzyme activities were measured at varying concentrations of CDNB and GSH in 0.1 M phosphate buffer, pH 6.5. The pK a was obtained by using 0.1 M sodium acetate buffers (from pH 5.0 to 5.5) and 0.1 M potassium phosphate buffer (from pH 6.0 to 8.5). The reaction was monitored at 340 nm, ⑀ ϭ 9600 M Ϫ1 cm Ϫ1 . It is well established that the bimolecular nucleophilic substitution reactions precede through a -complex intermediate (36). Thus, the rate-limiting formation of a spontaneous -complex intermediate can be increased by substitution of chlorine in the CDNB molecule with fluorine, which is more electronegative. With regard to the results, the ratio of the catalytic rate of GSH with FDNB and CDNB was comparable with the ratio of the second-order rate constants for the spontaneous uncatalyzed reaction. That is, k cat FDNB /k cat CDNB ϭ 40 is similar to k c FDNB /k c CDNB ϭ 47, which indicates that the -complex formation is the rate-limiting step. Though the k cat of the GST variants reflected different sensitivity to the nature of the leaving group, the catalytic efficiency (k cat /K m ) was nearly unchanged (Table II). Therefore it appears that an alteration of the relative turnover number is a consequence of changes in binding affinity toward different substrate leaving groups rather than a reflection of the rate of -complex formation.
The next step was to observe the effect of viscosity on kinetic parameters to study the rate-determining step of the catalytic reaction. A decrease of the rate constant by increasing the medium viscosity should reflect the weight of diffusion events on catalysis (37). It would indicate that the rate-limiting step is related to diffusion-controlled motions of the protein or the dissociation of the product. A plot of the reciprocal of the relative catalytic constant (k cat o /k cat ) against the relative viscosity (/ o ) should be linear. The slope should be equal to unity when the product release or structural transition is limited by a strictly diffusional barrier. If the slope approaches zero, either the chemistry or another non-diffusion barrier is ratelimiting. For adGSTD3-3 wild type, a plot of the inverse relative rate constant (k cat o /k cat ) versus the relative viscosity (/ o ) gives a linear dependence with a slope (1.14 Ϯ 0.01) very close to unity (Fig. 2). In contrast, the R66A mutant yields plots fully viscosity independent with a slope approaching zero. The other mutants, S65A, D100A, T158A, T162A, and T158A/T162A, exhibited k cat values with different degrees of viscosity dependence compared with the wild type enzyme.
The activity of the enzymes toward various hydrophobic substrates revealed that the specificity or the interactions with these substrates differed (Fig. 3). This result suggests that changes in the electron-sharing network residues cause a rearrangement of the active site, resulting in changes in the topology of the active site pocket and/or the ability of the responsible residues in the active site pocket of the enzyme to interact with hydrophobic substrates by an induced-fit mechanism (26,38).
The heat inactivation assay for adGSTD3-3 wild type was performed at different temperatures and demonstrated that the GST activity began to decrease at 45°C (39). This temperature was used to determine half-life stabilities for the mutant enzymes. The half-life corresponds to the time of preincubation when the mutant enzymes still have 50% remaining activity. All mutant enzymes, except R66A and T162A, showed no significant difference in half-life compared with the wild type (Table III). The R66A mutant enzyme was more stable than wild type by ϳ60-fold. In contrast, the alanine replacement at the Thr-162 position decreased the stability of the enzyme to ϳ7 times lower than the wild type. DISCUSSION Comparing the six subunits of adGSTD3-3 in the crystal structure, we observed an apparent electron-sharing network consisting of an ionic bridge interaction between the negatively charged glutamyl ␣-carboxylate group of glutathione, positively charged Arg-66, and negatively charged Asp-100 (Fig. 4). These three functional groups appear to form a resonance motif that is stabilized by a network of hydrogen bonds between Ser-65, Thr-158, Thr-162, and a conserved water-mediated contact (Fig. 5). To test the hypothesis of this electron-sharing network involvement in enzyme catalysis, the five G-site residues forming the network, Ser-65, Arg-66, Asp-100, Thr-158, and Thr-162, were replaced with alanine.
The kinetic studies demonstrated that the mutations reduced k cat , which resulted from a disruption of the electron distribution network. A dramatic decrease in enzyme activity was observed for the alanine replacements at Arg-66 and Asp-   100. We propose that Arg-66 and Asp-100 are key residues in this network because they generate an ionic bridge that is critical to the function of this motif. Lesser effects were observed for the remaining positions, suggesting these results may be due to the loss from the network of several stabilizing hydrogen bonds, as the key residues are still present. The question arose whether the effect on enzyme catalysis was due to a change in the rate-limiting step of the catalytic mechanism. The pH dependence of the kinetic parameters in the binary complex was determined to observe the ionization process. A crucial property of GSTs is their ability to lower the pK a of the thiol group of the bound GSH. The data show that the mutations caused an increase in the pK a values for the key residues, Arg-66 and Asp-100, of ϳ1 pH unit. The other network position mutants, Ser-65, Thr-158, and Thr-162, gave increases in pK a of ϳ0.5 pH unit when compared with the wild type. This suggests that the ionic bridge interaction and the stabilizing H-bonds of Ser-65, Thr-158, and Thr-162, as well as the water-bridge contact, are important for the ionization process. To test this concept, we engineered the wild type enzyme to introduce a more positively charged histidine into the Thr-158 position to expand the ionic bridge interaction. The k cat of this mutant enzyme, T158H, is nearly two times greater and the pK a of the bound GSH ϳ 0.2 pH unit lower than the wild type ( Table I).
The viscosity experiment elucidated that the rate-limiting step in the enzyme-catalyzed reaction by adGSTD3-3 is a nonphysical step (Fig. 2). The mutation at Arg-66 position changed the rate-determining step from a physical to a non-physical step, which is possibly the ionization step. However, alteration of the other network residues decreased the viscosity effects to intermediate values (0 Ͻ slope Ͻ 1), indicating that the ratelimiting step is not strictly dependent on a diffusional barrier and that other viscosity-dependent motions or conformational changes of the mutated proteins contribute to the rate-limiting step of the catalytic reaction (37). This suggests that the structural integrity or flexibility of functionally important regions, such as this electron-sharing network, of the mutated enzymes has been altered. This idea is supported by the differences in substrate specificity, which confirm active site changes, and the differences in half-life, which confirm structural movement changes. Previously, we have observed that changing residues that form the packing of the active site wall can influence the topology of the active site, which can affect both the binding mechanism as well as the structural maintenance of the enzyme (29).
However, the stability changes also might be due to the loss/change of structure of the electron-sharing network. The Thr-158 of the electron-sharing network can form two hydrogen bonds with Ile-154 and Ala-155, a conserved loop-helix substructure of the conserved folding module (GXXh(S/ T)XXDh) (X is any residue and h is a hydrophobic residue) (40). The local sequence of the conserved folding module in adGSTD3-3 is G 149 DSLTIADL 157 . This conserved folding module is present in all GST classes and is composed of two structural motifs at the N-terminal region of the ␣6 helix. These two structural motifs are the N-capping box ((S/T)XXD) and the hydrophobic staple motif in which two hydrophobic residues flank the N-capping box (41)(42)(43). Previous investigations demonstrated that single point mutations of residues that formed the conserved folding module had a dramatic effect on protein stability (40 -43). The ␣6 helix loop substructure (GST motif II)  is stabilized by a network of hydrogen bonds, and crystallographic studies indicate that these amino acid substitutions destabilized GST motif II through a partial or complete loss of the hydrogen bond network (44). Therefore the mutations in the electron-sharing network residues also appear to have an influence on the packing and/or stability of the protein because of the change to the hydrogen bond network in the motif II region. The amino acid alignments of known Delta class GSTs show the electron-sharing network residues are identical or functionally conserved within the class. A structural observation of several available crystal structures, adGSTD3-3, adGSTD4 -4, adGSTD5-5, adGSTD6 -6, agGSTD1-6, and LcGST, illustrate that an identical electron-sharing network is located in the same position. These equivalent residues show an acceptable range of distances between them to form the necessary ionic or hydrogen bonding. The water-mediated contact also is observed in several of the tertiary structures. Therefore, the electronsharing network appears to be conserved in the Delta class. The question arose of whether the electron-sharing network is conserved among all GST classes or is specific to the Delta class.
A primary sequence alignment of all GST classes suggests that there is no conserved equivalent electron-sharing network. However, similar features of the electron-sharing network, that is the ionic bridge interaction between negatively and positively charged residues stabilized by a network of hydrogen bonds, can still be observed in the same region but with slightly different residue positions (Fig. 6). For example, the putative electron-sharing network in hGSTP1 consists of Arg-13, Gln-64, Ser-65, Glu-97, Asp-98, and Cys-101. Alignment of Pi class GST shows that 4 of 6 putative electron-sharing network residues are perfectly conserved in position. Previous investigations demonstrated that mutations of residues forming the putative Pi GST electron-sharing network yielded dramatic effects on the enzyme catalysis (45)(46)(47)(48). A substitution of the charged residue at the Arg-13, Gln-64, or Asp-98 position decreased specific activity more than 95%. The decreased activities of these mutants gave a larger effect than the removal of the conserved Tyr-7 hydroxyl group, which is ϳ90% lower than wild type. Moreover, the k cat /K m versus pH profile for the D98N mutant was shifted by 0.5 pH unit in the alkaline direction. Hence, it was proposed that Asp-98 participated in proton transfer in the catalytic mechanism (48). In hGSTA1-1, the putative electron-sharing network is composed of Arg-15, Thr-68, and Glu-104. The Arg-15 and Thr-68 are strictly conserved in GST Alpha class. The mutation of Arg-15 to alanine or histidine caused a substantial reduction in the specific activity (200-or 400-fold, respectively), one order of magnitude more pronounced than the effect of the Y9F mutation (24,49). In addition, the corresponding pK a values of the Arg-15 mutants increase at least 0.5 pH unit when compared with the wild type (24). Mutation of the hydroxyl group that is hydrogen bonded to the ␣-carboxylate of the glutamate residue of glutathione (T68V) caused a shift of the pH dependence of the enzymecatalyzed reaction ϳ 1.5 pH units to more basic values as compared with the wild type (27). They also successfully mimicked the ionic bridge interaction by introducing the carboxylate group into a location generally occupied by the glutamyl ␣-carboxylate of hGSTA1-1. A T68E mutation increased catalytic efficiency with the decarboxylated analogue of GSH 10fold and reduced the pK a value of the active site-bound decarboxylated analogue of GSH by ϳ1 pH unit (28). For hGSTT2-2, Ser-67, Asp-104, Cys-105, and Arg-107 are proposed to be the putative electron-sharing network residues. Interestingly, Arg-107 is in hydrogen bonding distance of the main chain carbonyl of the ␥-glutamyl moiety of GSH and forms an interaction with the thiol sulfur of GSH either directly or through a water molecule (38). The replacement of Arg-107 by alanine remarkably increases the apparent pK a of the bound GSH from 6.1 to 7.8 (50,51). Arg-107 is a crucial residue in the electron-sharing network involved in the activation of the GSH, and it is strictly conserved in the Theta class. Comparisons between GST class Pi, Alpha, Theta, and Delta indicate that the electron-sharing network residues are identical or functionally conserved even through the interclass variations of GSTs. Although the electron-sharing network in other GSTs has not been characterized as such and non-conservation of the primary sequences of the residues forming the network is observed, nevertheless the structural comparisons of available crystal structures suggest that the presence of the network can be mapped to the same region in all GST classes based upon the possible distances between electron-sharing atoms and mutational studies in the other classes. Therefore, it appears that the proposed electron-sharing network in all GSTs derives from two critical residues that form ionic bridge interactions between the negatively charged glutamyl ␣-carboxylate group of glutathione, a positively charged residue (primarily Arg) and a negatively charged residue (Glu or Asp) stabilized by hydrogenbonding networks with surrounding residues (Ser, Thr, and/or water-mediated contact). The extent of the electron-sharing network seems to vary between GST classes. However, with regard to mutational studies in several GSTs, the electronsharing residues share similar characteristics: changes in either of the 2 key residues, the positively and negatively charged residues forming the salt-bridge interactions, nearly abolish enzymatic activity whereas replacement of residues involved in the hydrogen-bonding network decreases activity only ϳ20 -50%. In addition to an ionization step, which is modulated by the base-assisted deprotonation and the electronsharing network, the catalytic rate of a GST also depends upon other processes in the catalytic mechanism, e.g. nucleophilic substitution, product formation, and product dissociation. However, it is reasonable to suggest that a more extensive network might correlate with an increasing catalytic rate as it has a superior ability to lower the pK a value of bound GSH. As shown in Fig. 6, Delta and Pi class GSTs appear to have more cooperative residues in the network than Alpha class; therefore, the ability to lower the pK a of the bound GSH for Pi (pK a ϭ 6.2 Ϯ 0.1) (25) and Delta (pK a ϭ 6.3 Ϯ 0.1) (29) is greater than Alpha class (pK a ϭ 6.7 Ϯ 0.1) (25). However, the three enzymes exhibit similarities in catalytic activities toward CDNB substrate, Pi (k cat ϭ 34 Ϯ 2 s Ϫ1 ) (40), Alpha (k cat ϭ 48 Ϯ 4 s Ϫ1 ) (52), and Delta (k cat ϭ 35 Ϯ 1 s Ϫ1 ) (29). This demonstrates that additional factors in the catalytic mechanism have significant contributions to the catalytic rate and must be taken into account. One known example for such a factor is the C-terminal region of Alpha class GST, which is thought to help guide the reactants through multiple mechanisms into the transition state, resulting in an enhancement of enzymatic activity (52).
In conclusion, the present work, together with previous studies, supports the hypothesis of an electron-sharing network involved in the ionization of GSH. This network is characterized by an electrostatic interaction between negatively and positively charged amino acids stabilized by an array of hydrogen bonds. Therefore, an advantage of this type of extended network is that the electron-sharing burden is distributed among multiple residues that may be unable to fully support the function individually. This network appears to be a functionally conserved motif that contributes to the base-assisted deprotonation model suggested to be essential for the GSH ionization step of the catalytic mechanism. A structural diversification but functional conservation suggests a significant role for the electron-sharing network in catalysis as the purpose was maintained during the divergent evolution of GSTs.
FIG. 6. Putative electron distribution network in the other GST classes. The electron-sharing network, which is an ionic interaction between negatively charged and positively charged residues stabilized by a network of hydrogen bonds, can be observed in the same area involving different residues in the different GST enzymes. A, human P i hGSTP1-1 (PDB accession number 3GSS); B, human Alpha hGSTA1-1 (PDB accession number 1PKW); C, rat Mu rGSTM1-1 (PDB accession number 5FWG); D, human Omega hGSTO1-1 (PDB accession number 1EEM); E, human Theta hGSTT2-2 (PDB accession number 3LIR); F, squid Sigma GST (PDB accession number 2GSQ); and G, wheat Tau GSTI (PDB accession number 1GWC).The green line shows the putative electron movement pathway for distances between 2.5 and 3.0 Å.

VOLUME 280 (2005) PAGES 31776 -31782
An electron-sharing network involved in the catalytic mechanism is functionally conserved in different glutathione transferase classes.

PAGES 31777-31779:
We inadvertently did not cite previously published data properly in the legends of several tables and figures. The corrected Table 1 is shown in its entirety below. The legends for Tables 2 and 3 and Figs. 2 and 3 should read as stated below. TABLE 2. Effect of fluoride/chloride leaving group substitution on the rate of catalysis. The ratio of kinetic constants for the conjugation reaction catalyzing by adGSTD3-3 enzymes of GSH with CDNB and FDNB as co-substrates was calculated at pH 6.5. The data for wild type, S65A, and R66A have been reported previously in another format (29) and are shown here for purposes of comparison. TABLE 3. Thermal stability of wild type and mutants of adGSTD3-3 at 45°C. The wild type and mutant enzymes were incubated at 45°C at the protein concentration of 1 mg/ml. The inactivation time courses were determined by withdrawing suitable aliquots at the different time points for assay of the remaining activity to calculate the half-life of the enzyme. The data for wild type, S65A, and R66A have been reported previously (29) and are shown here for purposes of comparison. Figure 2. Viscosity effect on kinetic constants of wild-type and mutant enzymes. The effect of viscosity on kinetic constants was assayed by using 0.1 M potassium phosphate buffer, pH 6.5, with various glycerol concentrations. Dependence of the reciprocal of the relative turnover number (k cat 0 /k cat ) on the relative viscosity (/ 0 ) for CDNB as cosubstrate with WT (f), S65A (OE), R66A (), D100A (ࡗ), T158A (F), T162A (Ⅺ), and T158A/T162A (E). The experiment was performed in triplicate. and the lines were calculated by linear regression analysis. The slopes of the linear regression lines are 1.14 Ϯ 0.01 for wild type, 0.12 Ϯ 0.09 for S65A, Ϫ0.12 Ϯ 0.01 for R66A, 0.37 Ϯ 0.03 for D100A, 0.72 Ϯ 0.02 for T158A, 0.47 Ϯ 0.01 for T162A, and 0.63 Ϯ 0.01 for T158A/ T162A. The data for wild type, S65A, and R66A have been reported previously (29) and are shown here for purposes of comparison. Figure 3. Substrate-specific activity as a percent change compared with the adGSTD3-3 (wild type). The four substrates used for enzyme activity assays were: CDNB (1-chloro-2,4-dinitrobenzene), DCNB (1,2dichloro-4-nitrobenzene), EA (ethacrynic acid), and PNBC (p-nitrobenzyl chloride). The experiment was performed in triplicate, and the data are mean Ϯ S.D. The data for S65A and R66A have been reported previously (29) and are shown here for purposes of comparison.

TABLE 1
Steady-state kinetic parameters and pK a values for the thiol group of GSH of wild type and mutants of adGSTD3-3 for the CDNB conjugation reaction at pH 6.5 and 25°C The enzyme activities were measured at varying concentrations of CDNB and GSH in 0.1 M phosphate buffer, pH 6.5. The pK a value was obtained by using 0.1 M sodium acetate buffers (from pH 5.0 to 5.5) and 0.1 M potassium phosphate buffer (from pH 6.0 to 8.5). The reaction was monitored at 340 nm, ⑀ ϭ 9600 M Ϫ1 cm Ϫ1 .