Biochemical Journal

Research article

Structural contributions of Delta class glutathione transferase active-site residues to catalysis

Jantana Wongsantichon , Robert C. Robinson ,  Albert J. Ketterman


GST (glutathione transferase) is a dimeric enzyme recognized for biotransformation of xenobiotics and endogenous toxic compounds. In the present study, residues forming the hydrophobic substrate-binding site (H-site) of a Delta class enzyme were investigated in detail for the first time by site-directed mutagenesis and crystallographic studies. Enzyme kinetics reveal that Tyr111 indirectly stabilizes GSH binding, Tyr119 modulates hydrophobic substrate binding and Phe123 indirectly modulates catalysis. Mutations at Tyr111 and Phe123 also showed evidence for positive co-operativity for GSH and 1-chloro-2,4-dinitrobenzene respectively, strongly suggesting a role for these residues in manipulating subunit–subunit communication. In the present paper we report crystal structures of the wild-type enzyme, and two mutants, in complex with S-hexylglutathione. This study has identified an aromatic ‘zipper’ in the H-site contributing a network of aromatic π–π interactions. Several residues of the cluster directly interact with the hydrophobic substrate, whereas others indirectly maintain conformational stability of the dimeric structure through the C-terminal domain (domain II). The Y119E mutant structure shows major main-chain rearrangement of domain II. This reorganization is moderated through the ‘zipper’ that contributes to the H-site remodelling, thus illustrating a role in co-substrate binding modulation. The F123A structure shows molecular rearrangement of the H-site in one subunit, but not the other, explaining weakened hydrophobic substrate binding and kinetic co-operativity effects of Phe123 mutations. The three crystal structures provide comprehensive evidence of the aromatic ‘zipper’ residues having an impact upon protein stability, catalysis and specificity. Consequently, ‘zipper’ residues appear to modulate and co-ordinate substrate processing through permissive flexing.

  • crystal structure
  • glutathione transferase (GST)
  • structural motif
  • structure–function relationship
  • substrate specificity


GST (glutathione transferase; EC is eminently recognized as a detoxifying enzyme that is widely distributed in all organisms [1]. The enzymatic biotransformation within the active site is generally initiated by conjugation between glutathione and hydrophobic electrophilic toxic substances including drugs, carcinogens, herbicides and insecticides [2]. Amino acid residues forming the G-site (glutathione-binding site) appear to be highly conserved between GST classes and therefore have been studied for their roles in enzyme catalysis [36]. Amino acid residues forming the H-site (hydrophobic substrate-binding site) are more diverse allowing wide-ranging substrate selectivity. The situation is complicated further in that the active site appears to undergo rearrangement to have different residues interacting with different substrates [7,8]. The roles of specific H-site residues remain unexplored.

To date, multiple forms of GSTs of various organisms have been discovered and assigned into at least 14 distinct classes. The nomenclature and classification for this enzyme superfamily is principally based on amino acid sequence identity [9,10]. In mammalian species, there are seven identified cytosolic GST classes consisting of Alpha [11], Mu [12], Pi [12], Sigma [13], Theta [14], Omega [15] and Zeta [16]. Other classes are identified in non-mammalian species including Beta [17], Delta [18], Epsilon [19], Lambda [20], Phi [21], Tau [22] and the recently designated Rho class [23]. Many GST classes have counterparts across different phyla, but some are kingdom-specific, such as Alpha, Mu and Pi are mammal-specific [12,24]; Beta is bacterial-specific [24]; Delta and Epsilon are insect-specific [24]; and Phi and Tau are plant-specific classes [25].

GSTs catalyse CDNB (1-chloro-2,4-dinitrobenzene) as a common substrate [14,26]. Particular GST classes may display substrate selectivity towards other hydrophobic electrophilic compounds. Nevertheless, many GSTs appear to exhibit a degree of cross-specificity for a number of substrates [27]. Amino acids forming the H-site represent the first sphere of interaction with the co-substrate, therefore studying the roles of these residues would extend our understanding of enzyme catalysis and substrate specificity.

Previously, a crystal structure of adGSTD4-4 (Anopheles dirus GST Delta class homodimer of class 4) in an apo-form was obtained and five potential H-site residues next to the G-site were proposed [28]. The first three residues (Tyr111, Tyr119 and Phe123) are located on helix α4 and the other two residues (Phe212 and Tyr215) are on helix α8 at the C-terminus. Amino-acid-sequence alignment of Delta class GSTs demonstrates that Tyr111 and Tyr119 are highly conserved across insect species; Phe123 is conserved within alternatively spliced isoforms, among orthologous transcripts from other insect species and in some other Delta class members; whereas the Phe212 and Tyr215 positions have greater variation (Supplementary Figure S1 at These latter two residues on helix α8 have been investigated previously [29].

Using recombinant enzyme from the malaria vector A. dirus we have determined three new crystal structures of enzyme–inhibitor complexes, including wild-type and two mutants, in order to shed light on structure–function relationships in enzyme catalysis. These data implicate a structural contribution of highly conserved H-site residues, Tyr111, Tyr119 and Phe123, in enzyme catalysis and stability.



CDNB, DCNB (1,2-dichloro-4-nitrobenzene), EA (ethacrynic acid), PNPB (p-nitrophenethyl bromide), PNBC (p-nitrobenzyl chloride) and shGSH (S-hexylglutathione) were purchased from Sigma–Aldrich. GSH was from Bio Basic.

Recombinant enzymes

PCR-based site-directed mutagenesis using the Stratagene QuikChange® site-directed mutagenesis kit was used to introduce point mutations to the previously constructed plasmid of adGSTD4 wild-type enzyme [30]. All oligonucleotide primers (Sigma Proligo®) were designed independently using Vector NTI® Suite 8 software. Oligonucleotides containing the desired change with the addition or removal of a restriction endonuclease recognition site for positive-clone screening were synthesized. Recombinant plasmids were transformed into Escherichia coli DH5α strain for screening and E. coli BL21(DE3)pLysS for protein expression. The entire coding regions were verified by DNA sequencing at least twice. All GST enzymes were highly expressed as soluble proteins of approx. 25 kDa under 0.1 mM IPTG (isopropyl β-D-thiogalactoside) induction at 25 °C. Cell harvest and lysate preparation was performed as previously described [31]. Glutathione-affinity chromatography (GSTrap™ FF column; Amersham Biosciences) was initially used in an attempt to purify the recombinant enzymes. If this was not successful, a two-step procedure involving a cation exchanger (HiTrap™ SP-XL; Amersham Biosciences) and hydrophobic interaction chromatography (HiTrap™ Phenyl-Sepharose HP, Amersham Biosciences) was used [32]. All purified recombinant enzymes showed a single band on SDS/PAGE, indicating purity.

Enzyme characterization

Enzyme activity conjugating GSH to the CDNB substrate was determined by monitoring the increase in absorbance at 340 nm over time using a SpectraMax® 250 spectrophotometer. The standard GST assay was performed in 0.1 M potassium phosphate buffer (pH 6.5) in the presence of 3 mM CDNB and 10 mM GSH. The molar absorption coefficient of 9.6 mM−1·cm−1 [33] was used to convert the absorbance into moles. Kinetic parameter determination has been described previously [31]. One-way ANOVA and a post-hoc test by Dunnett's analysis were used to determine significant differences between the properties of engineered recombinant enzymes compared with the wild-type.

Substrate specificities of enzymes were determined by conjugation of GSH to the following hydrophobic electrophilic substrates: CDNB, DCNB, EA, PNPB and PNBC. Specific activities were calculated according to the molar absorption coefficient for each substrate [33].

The DCNB assay was performed in 0.1 M potassium phosphate buffer (pH 7.5) in the presence of 10 mM GSH and DCNB varied from 0.5 to 10 mM. The molar absorption coefficient of 8.5 mM−1·cm−1 was used to convert the absorbance into moles.

A thermal stability assay was performed as described previously [29].

Enzyme crystallization

Purified enzymes at more than 10 mg/ml were desalted into 50 mM Tris/HCl (pH 7.5) using Amicon® Ultra-15 centrifugal filter devices (Millipore) and then filtered through an Ultrafree-MC 0.22 μm pore-size centrifugal filter unit (Millipore). The proteins were freshly prepared for crystallization using the hanging-drop vapour diffusion method at 16 °C and 23 °C. Grid-screens or 2D (two-dimensional) optimizations were performed to obtain favourable crystal growth. Crystallizing conditions were carried out by mixing 2 μl of enzyme (in the presence of 5 mM shGSH) with 2 μl of precipitant solution comprising 0.1 M cacodylate (pH 6.6), 29–34% of PEG4000 [poly(ethylene glycol) 4000] and 0.13–0.16 M sodium acetate. The reservoir volume was 0.5 ml. In most cases, the first crystals appeared within 1–2 days.

Data collection and structural determination

Protein crystals were exposed to X-rays at a synchrotron light source (National Synchrotron Radiation Research Center, NSRRC, Hsinchu Science Park, Taiwan) on beamline BL13B. Diffraction images were processed and scaled using the HKL2000 program package [34]. The CCP4 package [35] and program O (Uppsala University software) were employed for structural refinement and model building. The previous model of an apo-form wild-type enzyme (PDB code: 1JLW) was applied as a template for molecular replacement of the WT:shGSH (WT is wild-type) dataset. Then, the resulting complete model was used as a template for other datasets. Datasets used for refinement of the final models for the WT:shGSH, Y119E and F123A were collected at 1.8, 2.2 and 1.7 Å (1 Å=0.1 nm) resolution respectively. Models were validated with the structure validation tools PROCHECK [36] and Molprobity [37].

Structural figures

Molecular graphic images were produced using the UCSF Chimera package from the Resource for Biocomputing, Visualization and Informatics at the University of California, San Francisco (supported by the National Institutes of Health grant number P41 RR-01081) [38] and PyMOL (DeLano Scientific;


Substrate specificity

A PCR-based site-directed mutagenesis technique was used to individually replace the amino acids of Tyr111, Tyr119 and Phe123 with several substitutes. All recombinant enzymes were individually screened for enzymatic and physical properties.

Five different hydrophobic substrates were used to monitor activity changes for H-site recombinant enzymes compared with wild-type activities (Supplementary Table S1 at All of the engineered enzymes displayed lower activity towards PNPB and EA, but possessed diverse activities towards CDNB, DCNB and PNBC (Supplementary Table S1). The data suggest that the orientation of the substrates or packing of the active site is altered by the replaced residues which has an impact on catalysis. Since the activity of the Y119F mutant towards DCNB was increased to approx. 18-fold that of the wild-type enzyme (Supplementary Table S1), the steady-state kinetic constants for DCNB conjugation were determined. Results showed that the catalytic rate (kcat) of the Y119F mutant was indeed greater than that of the wild-type enzyme; however, binding affinity was diminished, as indicated by the larger KmDCNB (Table 1). These data suggest that Tyr119 plays a role not only in hydrophobic substrate binding, but also in packing of the active-site pocket which influences chemical interactions involved in enzyme catalysis.

View this table:
Table 1 Kinetic constants of GSH conjugation to DCNB substrate

The data are the means±S.D. for at least three independent experiments.

Catalytic properties

Steady-state kinetics were studied by measuring GSH-conjugating CDNB activity.


The enzyme catalytic rate (kcat) of the nucleophilic aromatic substitution reaction with the classical GST substrate CDNB is shown in Table 2. Although Tyr111 is an H-site residue, steady-state kinetics illustrate a catalytic influence on GSH binding (the G-site). The negatively charged replacement of Y111E was the most unfavourable, with GSH-binding affinity decreasing approx. 14-fold relative to the wild-type enzyme. This replacement also decreased the kcat to approx. 15-fold less than the wild-type enzyme activity. As a result, the catalytic efficiency (kcat/KmGSH) of the Y111E mutant is reduced to approx. 225-fold less than the wild-type enzyme. In addition, the mutant enzyme demonstrated positive co-operativity upon GSH binding, indicating that the residue replacement affects the communication between subunits of the dimeric enzyme. The CDNB-binding affinity was also decreased approx. 2.6-fold relative to the wild-type enzyme, probably due to conformational rearrangement of the H-site. The positively charged replacement of the Y111H mutant showed a slightly different catalytic rate compared with wild-type. Nevertheless, the enzyme displayed a 4-fold lower binding affinity towards GSH, as well as positive co-operativity. However, no significant difference was observed in KmCDNB for this mutant enzyme when compared with the wild-type. The Y111A, Y111-S and T111F mutants revealed no significant differences in either KmGSH or KmCDNB values, yet the enzymes displayed different kcat values in a range from 1.5- to approx. 4-fold lower than the wild-type enzyme activity (P<0.01). These results suggest that Tyr111 has a contribution to CDNB substrate binding and is additionally involved in stabilization of glutathione substrate binding, probably through the molecular packing of the active-site pocket.

View this table:
Table 2 Steady-state kinetic constants of H-site mutants, using GSH and CDNB as substrates

The data are the means±S.D. for at least three independent experiments. Bold italic numbers represent positive co-operativity of substrate binding. ANOVA revealed significant differences compared with the wild-type, where indicated (†P<0.01 and §P<0.05). The absence of a symbol indicates no significant difference compared with the wild-type. Calculations of the Hill coefficient (h) and Km for co-operativity kinetics were performed as described previously [31].


Amino acid substitutions of Tyr119 had no significant effect on GSH binding as indicated by KmGSH (Table 2). This suggests that this residue position has no influence on the G-site topology. The recombinant proteins possessed similar catalytic rates to that of the wild-type. It is of interest that replacements in this position show either a weakening or strengthening effect on the CDNB-binding affinity. Surprisingly, the Y119F mutant showed a 9-fold greater CDNB-binding affinity compared with that of the wild-type enzyme. Tyr119 is the equivalent residue to the important active-site residue Tyr108 in human Pi GST. However, this observation is inconsistent with the reports for Pi GST which showed no effects for Y108F using CDNB as a substrate [40]. In contrast with Y119F, Y119S and Y119E mutants showed diminished CDNB-binding affinity of approx. 2- to 3-fold compared with the wild-type enzyme, whereas Y119A and Y119H mutants showed no effects. The results demonstrate that this residue position in the structure of adGSTD4-4 is significant for modulating hydrophobic substrate binding. An equivalent residue of Tyr119 that was investigated in GST class Pi (Tyr108), Mu (Tyr115) and Sigma (Phe106) revealed that the contribution of this residue position varied among the GST classes [13,4042]. Overall findings show that the presence or removal of a hydroxy group in this position has a different influence on catalysis depending on the types of reactions/substrates, as well as whether the rate-limiting step in catalysis is a physical or chemical step. Therefore the residue acts as a modulator for both enzyme catalysis and co-substrate recognition. From our data, the equivalent residue Tyr119 in Delta class appears to perform a similar role, which suggests that this residue position has been functionally conserved during the evolutionary development of the GST classes.


All Phe123-mutated enzymes exhibited significantly lower maximum velocities in the range from 1.7- to 4-fold less than the wild-type (P<0.01). It was striking that all four amino acid replacements resulted in decreased CDNB-binding affinity concomitant with positive co-operativity effects (Table 2). The cooperativity for CDNB substrate was previously reported in a study of subunit interface residues of the same enzyme, adGSTD4-4 [43]. Positive co-operativity indicates that the binding of CDNB to one subunit facilitates the binding of another CDNB molecule in the neighbouring subunit. Hence, it appears that the position 123 residue contributes to stabilizing subunit–subunit interactions in addition to binding of co-substrate at the H-site.

Protein stability

Heat inactivation at 45 °C was employed to determine the thermal stability of H-site mutant enzymes. The results show that kinetic properties and physical properties of the enzymes are quite independent. For instance, the negatively charged amino acid replacements Y111E, Y119E and F123E showed a negative effect on enzyme catalysis through either catalytic rate or binding affinity influence (Table 2), but all three recombinant proteins exhibited improved stability of approx. 34-, 22- and 7-fold greater than the wild-type respectively (Supplementary Figure S2 at A difference in half-life is evident for rearrangements that influenced conformational ensembles attainable by the proteins.

Crystal structures

Although the wild-type apo-structure is available, obtaining a ligand-bound complex was necessary to gain information on the roles of particular amino acids in the H-site. Several recombinants with remarkable features were subjected to crystallization experiments, but only crystal structures of the wild-type enzyme, Y119E and F123A mutant proteins were successfully obtained by co-crystallizing with shGSH. These three new structures were elucidated by molecular replacement. The previous apoenzyme structure (PDB code: 1JLW) [28] was used as template for WT:shGSH and the resulting model was subsequently used for the other two structures. Data collection, refinement statistics and model content are shown in Supplementary Table S2 (at

The dimeric structure displays a GST canonical fold with two identical subunits that are well aligned with an RMSD (root mean square deviation) of 0.303 Å. The overlaid subunit configurations (Figure 1) demonstrate alternative arrangement for the S-hexyl moiety between the subunits. This suggests that active-site formation in the native form is quite dynamic to accommodate hydrophobic substrates. Structural investigation shows that only the glutathionyl moiety of the ligand forms hydrogen bonds to the enzyme at His50, Ile52, Glu65 and Ser66 with distances between 2.6 and 3.0 Å. Therefore these G-site residues appear to be significant for initial binding of GSH and enzyme catalysis as shown in previous studies [6,44]. In contrast, there is no highly conserved interaction between the S-hexyl moiety and the H-site. All three highly conserved H-site residues in the present study appear to simply form a hydrophobic wall of the active-site cavity to locate the hexyl moiety of the ligand. However, an observation of note was the presence of an aromatic zipper at the interface between helix α4 and α6-7-8 (Figure 2). This motif is formed by an extended cluster of aromatic residues within domain II, the C-terminal domain. This elongated aromatic motif has extensive internal π–π and H-bond (hydrogen bond) interactions that distinguishes it from the usual hydrophobic packing. The ‘zipper’ consists of the following 12 residues: Tyr111, Phe114, Tyr117, Tyr118, Tyr119, Phe123, Tyr175, Tyr180, Tyr185, Tyr212, Tyr215 and Phe216. This motif appears to maintain structural integrity of the domain with aromatic hydrophobic interactions and therefore is involved in conformational stability of the H-site.

Figure 1 Stereo view of the WT:shGSH dimeric structure showing alternative arrangements of the S-hexyl ligand

The ribbon representation is the superposition of the bottom subunit on to the top subunit with a RMSD of 0.303 Å. The loop connecting helix α4–5 was clipped for viewing clarification.

Figure 2 An aromatic zipper in domain II of WT:shGSH

Aromatic zipper residues are in stick representation and shGSH in ball-and-stick representation. Residues in yellow were mutated to assess the impact of the ‘zipper’ motif.

Y119E:shGSH comprises an inhibitor residing in both subunits in similar configuration. The final model shows intact peptide chains of 217 and 216 amino acids, although electron density for B123 to B129 of subunit B was of poor quality. The segment could only be traced from a 2FoFc map, at a contour level of 0.5σ, and the whole region was built into the peptide chain. Low electron density of the region reflects a high degree of flexibility at the loop connecting helix α4 and α5. Owing to glutamate disruption of the aromatic ‘zipper’, major structural rearrangement was found, especially in the first subunit or subunit A (Figure 3c). Accordingly, the residue appears to regulate affinity for the co-substrate binding as indicated by Km values in Table 2. The Y119E mutant possesses the lowest CDNB-binding affinity, whereas Y119F strengthens the binding 28-fold relative to the Glu119 mutant. The side chain of Tyr119 is located between the side chains of Tyr111 and Phe123 which together face towards the conjugated substrate. As a result, the flanking aromatic rings appear to contribute to structural integrity of the domain. The difference between the two subunits of Y119E in the α4-loop-α5 region shows molecular flexing of the active-site pocket. This structural flexibility of the mutant also effects the ionic interactions at the dimeric interface. In the wild-type structure, Glu116 interacts with Arg134 in the same subunit; but in the mutant structures, Glu116 instead interacts with the Gln112, as well as the Glu116 from the other subunit (Figure 3c). However, the residue substitution does not have a detrimental effect on enzyme function (Table 2). In fact, the Y119E mutant can catalyse DCNB approx. 12-fold better than the wild-type enzyme (Table 1). It is of interest that the enzyme has a high structural adjustability to a modification at this position.

Figure 3 Structural differences between WT:shGSH, Y119E:shGSH and F123A:shGSH

(a) shGSH conformation with H-site residues of subunit A; (b) shGSH conformation with H-site residues of subunit B; (c) charged interactions at the dimer interface, the mutated residues are in yellow.

F123A:shGSH displays alternative orientations for the S-hexyl moiety. Without the aromatic ring in position 123, the active-site pocket supplies a unique arrangement for substrate binding. The first peptide chain in this structure was defined to 218 amino acids, whereas the second chain was 213 amino acids with an undefined region in the H-site. In subunit A, alanine residue substitution for Phe123 apparently provides a sub-pocket for the hexyl moiety, as well as contributing to movement of neighbouring residues that affects the H-site topology (Figure 3a and 3b). In the second subunit, the lack of electron density for residues 120–129 again demonstrates H-site fluctuation to accommodate different binding modes, which is similar to the weak electron density region in Y119E:shGSH. It is also notable that Phe212, one of the H-site residues and part of the aromatic zipper, was relocated due to an unwound helix-8 in this subunit. The residue was previously investigated by site-directed mutagenesis and was reported to be important for the rate-limiting step of the nucleophilic aromatic substitution reaction [29]. Therefore Phe123 plays an indirect role in modulation of enzyme catalysis. Given that all affected regions in the three structures are essential parts that form the H-site, it suggests that, after binding GSH, flexing occurs to accommodate the hydrophobic substrate binding. The extent of this flexing is determined by the aromatic cluster interactions. With the loss of Phe123 interactions, the flexing became larger in one subunit as shown by the loss of electron density. The movement of this region appears to be analogous to the α2 helix movement for binding GSH [45] as the enzyme undergoes ‘induced fit’ relocation for recognition and binding of the hydrophobic substrate. Accordingly, the flexibility of this region is controlled and modulated through the interactions of the aromatic ‘zipper’ to accommodate different target molecules.

Active-site rearrangements

Although no protein crystal of the Tyr111 mutant was obtained, deductions can be made from related crystal structures in the present study. WT:shGSH shows that the hydroxy groups of Tyr111 and of Tyr119 are in H-bonding distance of 2.7 Å. In fact, H-bonds are found extensively throughout the H-site region. Additionally, substitutions of Tyr111 with glutamate or histidine residues would disrupt hydrophobicity of the binding site and therefore lower the affinity towards CDNB. Rearrangement of charge distribution in the region includes fine-tuning of side-chain interactions to residues in proximity, as well as alteration of substrate orientations. First, it appears that charged replacements of Tyr111 may directly disrupt the positioning of GSH in the active-site pocket, therefore lowering the substrate-binding affinity (Table 2). Secondly, investigation of the first sphere contacts of Tyr111 in the WT:shGSH structure, suggests that there is a high possibility that the charged amino acid replacement at this position interacts with Arg67 in the active site. Structural observations suggest that the substitution of glutamate at the Tyr111 position probably removes the H-bond with Tyr119, but favours formation of a salt bridge with Arg67 and additional H-bonds with Asp106 and Thr171. Accordingly, the Y111E mutant showed an increased half-life and was the most stable recombinant enzyme in the present study, with a 34-fold greater stability than wild-type.

Zipper motif roles

To investigate the H-site, the residues Tyr111, Tyr119 and Phe123 were initially chosen. Upon characterization, these residues appeared to exhibit only moderate effects on catalysis. Generally, these effects would most probably originate through indirect structural modifications and not direct interactions with the studied substrates. With a closer inspection of the structures, it was found that these residues were part of an aromatic motif that appeared to contribute structurally to domain II with interactions between α-helices 4, 5, 6, 7 and 8. Structure-based sequence alignment was employed to analyse six Drosophila melanogaster, six human and two mouse GST classes {Expresso (3D-Coffee) at; [46]}. The analysis showed that although the ‘zipper’ motif is of a similar extent in the Delta and Epsilon classes several, but not all, residue positions of this motif appear to be conserved in most of the classes (Supplementary Figure S3 at In several classes, the aromatic residues appear to be replaced with hydrophobic residues, thereby changing the nature of the motif. However, although the aromatic ‘zipper’ motif is not conserved in all GST classes, several of the aromatic positions appear to be conserved across classes, suggesting a functional importance.

The aromatic ‘zipper’ motif interacts at several points to stabilize α helix 5. It was noted that the outer part of helix 5, specifically Phe144 and Gln140, interacted with His50 in the active site of the other subunit. Previously it has been shown that His50 is a key basic residue for GSH ionization through the glycine moiety [44]. Therefore we proposed that other residues in the aromatic ‘zipper’ motif could exert similar effects on catalysis through structural influences of the ‘zipper’ changes. To test this concept three new constructs were made with alanine mutations at Phe114, Tyr118 and Tyr175. These residues are all interior positions and not part of the active site (Figure 2). Characterization of these engineered enzymes showed similar effects on catalysis (Table 2 and Supplementary Table S1). It was of interest to note that two of these engineered enzymes also displayed co-operativity in GSH kinetics with Hill coefficients greater than 1.5. We also observed that these three positions had influence on the overall stability of the protein as two decreased the half-life, whereas Y118A increased the half-life 6.8-fold (Supplementary Figure S2). In conclusion, the aromatic ‘zipper’ motif contributes to modulating the protein structure, which has an impact on protein stability and the topological arrangement of the active site. The active-site arrangement in turn affects substrate specificity and catalysis.


Jantana Wongsantichon, Robert Robinson and Albert Ketterman designed the research. Jantana Wongsantichon performed the research. Jantana Wongsantichon, Robert Robinson and Albert Ketterman analysed the data. Jantana Wongsantichon and Albert Ketterman wrote the paper.


This work was supported by the Thailand Research Fund; and the National Synchrotron Research Center, Thailand [NSRC grant number 2550/09]. J. W. was supported by a postdoctoral research fellowship from the Commission of Higher Education (CHE), Thailand. R. C. R. is supported by A*STAR, Singapore.


We thank the beamline staff at the National Synchrotron Radiation Research Center (NSRRC), Hsinchu, Taiwan, for data collection and processing assistance.


  • Co-ordinates have been deposited in the RSCB Protein Data Bank with accession codes of 3F63 for WT:shGSH, 3G7J for Y119E:shGSH and 3F6D for F123A:shGSH.

Abbreviations: adGSTD4-4, Anopheles dirus GST Delta class homodimer of class 4; CDNB, 1-chloro-2,4-dinitrobenzene; DCNB, 1,2-dichloro-4-nitrobenzene; EA, ethacrynic acid; G-site, glutathione-binding site; GST, glutathione transferase; H-bond, hydrogen bond; H-site, hydrophobic substrate-binding site; PNBC, p-nitrobenzyl chloride; PNPB, p-nitrophenethyl bromide; RMSD, root mean square deviation; shGSH, S-hexylglutathione


View Abstract