Crystal structures and kinetic studies of human Kappa class glutathione transferase provide insights into the catalytic mechanism

Article abstract of DOI:10.1042/BJ20110753

GSTs (glutathione transferases) are a family of enzymes that primarily catalyse nucleophilic addition of the thiol of GSH (reduced glutathione) to a variety of hydrophobic electrophiles in the cellular detoxification of cytotoxic and genotoxic compounds.

Full text of DOI:10.1042/BJ20110753

Biochem. J. (2011) 439, 215–225 (Printed in Great Britain) doi:10.1042/BJ20110753
215
Crystal structures and kinetic studies of human Kappa class glutathione
transferase provide insights into the catalytic mechanism
1
2
Bing WANG*†, Yingjie PENG* , Tianlong ZHANG* and Jianping DING*
*
State Key Laboratory of Molecular Biology and Research Center for Structural Biology, Institute of Biochemistry and Cell Biology, Shanghai Institutes for Biological Sciences, Chinese
Academy of Sciences, Shanghai 200031, China, and †Graduate School of Chinese Academy of Sciences, 320 Yue-Yang Road, Shanghai 200031, China
GSTs (glutathione transferases) are a family of enzymes that
primarily catalyse nucleophilic addition of the thiol of GSH
(GSH-binding site) and the H site (hydrophobic substrate-binding
16
site). The conserved Ser at the G site functions as the catalytic
residue in the deprotonation of the thiol group and the conserved
(
reduced glutathione) to a variety of hydrophobic electrophiles in
69
200
201
202
the cellular detoxification of cytotoxic and genotoxic compounds.
GSTks (Kappa class GSTs) are a distinct class because of their
unique cellular localization, function and structure. In the present
paper we report the crystal structures of hGSTk (human GSTk)
in apo-form and in complex with GTX (S-hexylglutathione) and
steady-state kinetic studies, revealing insights into the catalytic
mechanism of hGSTk and other GSTks. Substrate binding induces
a conformational change of the active site from an ‘open’
conformation in the apo-form to a ‘closed’ conformation in
the GTX-bound complex, facilitating formations of the G site
Asp , Ser , Asp and Arg form a network of interactions with
γ -glutamyl carboxylate to stabilize the thiolate anion. The H site
is a large hydrophobic pocket with conformational flexibility to
allow the binding of different hydrophobic substrates. The kinetic
mechanism of hGSTk conforms to a rapid equilibrium random
sequential Bi Bi model.
Key words: catalytic mechanism, conformational change, crystal
structure, Kappa class glutathione transferase (GSTk), kinetic
measurement, substrate-binding pocket.
INTRODUCTION
anion, (iii) nucleophilic attack of the thiolate on the electrophilic
centre of the substrate to form a thioether bond, and (iv) release
of the resulting product [1,17–20]. The activation of GSH to form
the thiolate anion is the critical step in the catalytic reaction. It is
proposed that a conserved residue at the active site (tyrosine in
Alpha, Mu and Pi classes, serine in Theta and Delta classes, and
cysteine in the Omega class) facilitates the thiol deprotonation by
forming a hydrogen bond with the thiol group of GSH [19,21–23].
It is also proposed that the γ -glutamyl carboxylate group of GSH
functions as the catalytic base to accept the proton from the thiol
group of GSH, and an electron-sharing network helps to stabilize
the thiolate anion [20,24,25].
GSTks (Kappa class GSTs) are suggested to form a distinct
family of the Trx (thioredoxin)-fold superfamily, different from
the soluble GSTs [5,26–29]. They localize in the matrix of
mitochondria and peroxisomes of the cells, and exhibit a GSH-
dependent conjugation activity with model substrates [26,28,30],
and possibly other activities including peroxidase and isomerase
[27,29]. The structure of rGSTk (rat mitochondrial GSTk) in
complex with GSH demonstrated that the enzyme is structurally
more similar to DsbA than to the soluble GSTs, suggesting that
it was evolved in a parallel pathway from that of the soluble
GSTs [23,29,31]. DsbA is a protein disulfide oxidoreductase that
catalyses the formation of disulfide bonds in Escherichia coli.
It was reported that overexpression of GSTk in mouse cells up-
regulates the multimerization of adiponectin molecules, probably
through formation of intracellular disulfide bonds [32]. However,
GSTks do not contain a conserved CXXC motif at the active
site which is required by DsbA to catalyse the oxidation and
GSTs (glutathione transferases) are a family of multifunctional
enzymes that play important roles primarily in the cellular
detoxification of cytotoxic and genotoxic compounds. The main
chemistry of GSTs is to catalyse nucleophilic addition of the thiol
group of GSH (reduced glutathione), the tripeptide γ -Glu-Cys-
Gly, to a wide variety of hydrophobic electrophiles [1–6]. The
products of the conjugation are more soluble non-toxic peptide
derivatives. GSTs exist in all eukaryotes as well as in many
bacteria, which have been divided into a number of classes based
on their amino acid sequence homology in combination with other
criteria, including tertiary structure similarity, substrate specificity
and immunological properties [1,6–8]. Although GSTs share less
than 30% identity among different classes, almost all canonical
GSTs (also called soluble or cytosolic GSTs) function as dimers.
The residues forming the G site (GSH-binding site) are highly
conserved, whereas the residues forming the H site (hydrophobic
substrate-binding site) are less conserved to allow a wide range
of substrate selectivity. In addition to their function in cellular
detoxification, GSTs are implicated to play important roles in cell
signalling and other cellular processes [9–11], and are associated
with many human diseases [12–16].
On the basis of structural and biochemical studies of several
classes of soluble GSTs, in particular the Alpha, Mu and
Pi classes of GSTs, the catalytic mechanism of nucleophilic
aromatic substitution reaction by GSTs has been postulated to
consist of four steps: (i) substrate binding to the active site, (ii)
activation of GSH by thiol deprotonation to form the thiolate
Abbreviations used: CDNB, 1-chloro-2,4-dinitrobenzene; DTT, dithiothreitol; GSDNB, S-(2,4-dinitrophenyl)glutathione; GSF, glutathione sulfinate; GSH,
reduced glutathione; G site, GSH-binding site; GST, glutathione transferase; GSTk, Kappa class GST; GTX, S-hexylglutathione; HCCA, 2-hydroxychromene-
2
-carboxylic acid; hGSTk, human GSTk; H site, hydrophobic substrate-binding site; PEG, poly(ethylene glycol); rGSTk, rat GSTk; RMSD, root mean square
deviation; Trx, thioredoxin.
1
Present address: Department of Molecular Biology, The Scripps Research Institute, 10550 North Torrey Pines Road, La Jolla, CA 92037, U.S.A.
2
To whom correspondence should be addressed (email jpding@sibs.ac.cn).
The structures of hGSTk in apo-form and in complex with GTX have been deposited with the RCSB PDB under accession codes 3RPP and 3RPN
respectively.
ꢀc The Authors Journal compilation ꢀc 2011 Biochemical Society

216
B. Wang and others
isomerization of disulfide bonds in substrate proteins [33,34], and
thus are unlikely to catalyse a DsbA-type reaction. The structure
of hGSTk (human GSTk) in complex with GSF (glutathione
sulfinate) showed that hGSTk has a similar overall structure to
that of rGSTk and its Trx-fold domain is structurally more similar
to that of human Theta class GST than the other soluble GSTs,
although the two enzymes share a very low sequence identity
Table 1 Summary of diffraction data and structure refinement statistics
Numbers in parentheses indicate the highest resolution shell.
Measurement
hGSTk
hGSTk-GTX
Diffraction data
Space group
C2
P2
1
(
19%) [35]. In addition, the active site of both rGSTk and hGSTk
Cell parameters
a, b, c ( A˚ )
146.1, 84.3, 87.3
90.0, 122.1, 90.0
50.0–1.80 (1.86–1.80)
511931
66.4, 199.8, 66.7
90.0, 116.1, 90.0
50.0–1.90 (1.97–1.90)
413733
is similar to that of the soluble GSTs, and GSTks appear to
function in a manner similar to that of the soluble GSTs [23].
To understand the catalytic mechanism of GSTks, we carried
out structural and biochemical studies of hGSTk. In the present
paper we report the crystal structures of hGSTk in apo-form and
in complex with an inhibitor, namely GTX (S-hexylglutathione),
and the mutagenesis and steady-state kinetic studies of the
enzyme. The structural and biochemical data taken together
reveal that hGSTk utilizes an induced-fit mechanism for substrate
binding and a rapid equilibrium random sequential Bi Bi kinetic
mechanism for catalysis, and suggest that other GSTks might
function in a similar manner.
◦
α, β, γ ( )
Resolution ( A˚ )
Observed reflections
Unique reflections [I/σ(I)>0]
Average I/σ(I)
Completeness (%)
82817
21.6 (4.9)
100.0 (100.0)
10.0 (23.5)
116840
19.7 (3.6)
95.8 (98.7)
9.8 (32.7)
R
merge (%)*
Refinement and structure model
Reflections [Foꢀ0σ(F
o
)]
Working set
78674
4142
15.2
19.5
5914
649
110664
5836
18.3
Test set
R factor (%)†
Free R factor (%)†
21.8
Number of non-hydrogen atoms
Number of amino acid residues
11418
1320
804
MATERIALS AND METHODS
Number of water molecules
752
Average B factor ( A˚ )
2
Protein expression and purification
All atoms
Protein
GTX
27.6
26.3
–
38.9
38.1
45.2
47.1
The hGSTk protein was expressed in E. coli and purified
by affinity chromatography using a nickel-chelating resin, as
described previously [35]. The purified protein was dialysed
Water molecules
36.7
RMSDs
against buffer containing 20 mM NaH
2
PO (pH 7.4), 20 mM
4
Bond lengths ( A˚ )
0.006
0.954
0.007
0.989
NaCl, 1 mM EDTA and 1 mM DTT (dithiothreitol) for
biochemical studies. The protein used for crystallization of the
apo-form of hGSTk was dialysed against buffer containing 20 mM
◦
Bond angles ( )
Ramachandran plot (%)
Most favoured regions
Allowed regions
92.4
7.6
91.5
8.5
NaH
2
PO (pH 7.4), 20 mM NaCl, 1 mM EDTA and 7.2 mM 2-
4
ꢀ
ꢀ
ꢀ
ꢀ
mercaptoethanol. The protein used for crystallization of hGSTk in
complex with GTX was dialysed against buffer containing 20 mM
Hepes (pH 7.0), 50 mM NaCl, 1 mM EDTA and 1 mM DTT.
Constructs of the hGSTk mutants containing point mutations were
*R
=
=
|I (hkl) − ꢁI(hkl)ꢂ|/
Ii(hkl).
i
merge
factor
hkl
hkl ||F
i
i
hkl
ꢀ
ꢀ
†
R
o
| − |F
c
||/ hkl |F
o
|.
®
generated using the QuikChange Site-Directed Mutagenesis
with the program CNS [38] following the standard protocols, and
the final structure refinement was performed using the program
PHENIX [39]. No non-crystallographic symmetry constraint was
applied in the refinement. The manual model building was per-
formed with the program COOT [40]. In the initial difference
Fourier maps, there was strong electron density corresponding
to a GTX molecule in each hGSTk subunit in the hGSTk–GTX
complex. The statistics of the structure refinements and the final
structure models are summarized in Table 1. The quality of
the stereochemistry of the structure models was validated with the
program PROCHECK [41]. The molecular graphic images were
prepared using the program PyMOL (http://www.pymol.org).
kit (Stratagene) and were verified by DNA sequencing. The
procedures for expression and purification of the mutants were
the same as those for the wild-type protein.
Crystallization and diffraction data collection
Crystals of the apo-form of hGSTk were grown by the hanging
◦
drop vapour diffusion method at 4 C in drops containing
equal volumes (0.5 μl) of the protein solution (5 mg/ml) and
the crystallization solution {0.2 M Mg(NO
3
) and 15% PEG
2
[
poly(ethylene glycol)] 3350}. Crystals of hGSTk in complex
with GTX were grown by the hanging drop vapour diffusion
◦
method at 20 C in drops containing equal volumes (0.5 μl) of the
protein solution (10 mg/ml) supplemented with GTX (Sigma) at a
molar ratio of 1:2 and the crystallization solution (0.2 M NaSCN
and 20% PEG3350). Diffraction data of the apo-form and GTX-
bound hGSTk were collected to 1.80 Å (1 Å = 0.1 nm) and 1.90
Å resolution respectively, at 100 K from flash-cooled crystals at
the Shanghai Synchrotron Radiation Facility, Shanghai, China,
and processed using the HKL2000 suite [36]. The diffraction data
statistics are summarized in Table 1.
Enzymatic activity assay and kinetic analysis
The enzymatic activity assay of both wild-type and mutant
hGSTk catalysing the addition of GSH to CDNB (1-chloro-2,4-
dinitrobenzene) was carried out using a modified protocol, as
described previously [31]. All steady-state kinetic experiments
were performed in buffer containing 20 mM NaH
2
PO
4
(pH 7.4),
◦
20 mM NaCl, 1 mM EDTA and 1 mM DTT at 25 C, and the
amount of the product was measured at 338 nm using a Beckman
Coulter DU800 spectrophotometer. The enzymatic reaction rate
was linear up to 60 s after initiation and the initial rate was
measured at 2 s intervals for a total period of 60 s. The non-
enzymatic reaction rate served as a control and was subtracted
from the enzymatic rate. Owing to the low solubility of CDNB
in water, ethanol was used as the solvent for CDNB. To avoid a
Structure determination and refinement
Crystal structures of both the apo-form and GTX-bound
hGSTk were solved using the molecular replacement method
implemented in the program MOLREP [37] with the structure
of hGSTk in complex with GSF (PDB code 1YZX) [35] as the
search model. The initial structure refinement was carried out
ꢀc The Authors Journal compilation ꢀc 2011 Biochemical Society

Catalytic mechanism of human GSTk
217
Table 2 Steady-state kinetic parameters of wild-type and mutant hGSTk
containing point mutation at the G site towards GSH
Table 3 Steady-state kinetic parameters of wild-type and mutant hGSTk
containing point mutation at the H site towards CDNB
The kinetic parameters are the means + S.D. of triplicate determinations. ND means that the
The kinetic parameters are the means + S.D. of triplicate determinations. ND means that the
−
−
enzymatic activity is too low to be detected.
enzymatic activity is too low to be detected.
GSH
− 1
(M− 1·s− 1
CDNB
− 1
(M− 1·s− 1
hGSTk
K
m
(mM)
k
cat (s
)
k
cat/K
m
)
hGSTk
K
m
(mM)
k
cat (s
)
k
cat/K
m
)
3
3
Wild-type
S16A
Y18F
Y18H
Y18L
Y18A
S19A
N53A
P56A
K62A
K62E
0.19 + 0.02
1.18 + 0.02
(6.21 + 0.52)×10
Wild-type
L15A
P17A
I44A
P55A
P56A
L58A
M66A
F83A
F83W
F87A
F87W
L88A
M91A
M91K
L92A
1.10 + 0.12
3.94 + 0.23
(3.58 + 0.45)×10
−
17.34 + 4.79
−
−
0.17 + 0.03
−
−
−
−
−
2
3
3
3
4
3
9.80 + 3.20
2.59 + 0.43
1.19 + 0.18
(7.37 + 1.41)×10
−
−
−
−
20.45 + 3.53
−
1.56 + 0.10
−
(7.63 + 1.40)×10
0.34 + 0.04
2.38 + 0.11
(7.00 + 0.85)×10
−
−
−
−
ND
ND
ND
ND
ND
ND
2.01 + 0.44
3.28 + 0.46
(1.63 + 0.42)×10
−
−
−
1.08 + 0.14
4.50 + 0.28
(4.17 + 0.60)×10
−
−
−
0.74 + 0.07
7.83 + 0.40
(1.06 + 0.12)×10
−
−
−
3
2
3
2
1.61 + 0.41
3.74 + 0.51
(2.32 + 0.67)×10
1.45 + 0.16
2.40 + 0.14
(1.66 + 0.21)×10
−
−
−
−
−
−
4.53 + 0.38
1.75 + 0.05
(3.86 + 0.34)×10
>5
ND
−
−
−
1.68 + 0.31
2.29 + 0.16
(1.36 + 0.27)×10
1.30 + 0.16
6.24 + 0.44
(4.80 + 0.68)×10
3
3
3
3
3
3
3
−
−
−
−
−
−
1.25 + 0.07
0.89 + 0.01
(7.12 + 0.39)×10
0.64 + 0.07
3.18 + 0.15
(4.97 + 0.61)×10
−
−
0.75 + 0.04
−
−
−
−
−
25.26 + 3.89
(2.97 + 0.48)×10
0.91 + 0.10
4.18 + 0.22
(4.59 + 0.55)×10
−
−
−
−
−
D69T
60.89 + 13.66
1.33 + 0.11
−
(2.18 + 0.52)×10
1.38 + 0.17
5.78 + 0.41
(4.19 + 0.59)×10
−
105.39 + 21.83
−
−
3.04 + 0.74
−
−
−
−
D69A
L183A
S200A
D201A
R202A
0.32 + 0.04
4.44 + 0.74
5.00 + 0.58
(1.13 + 0.23)×10
−
−
−
−
9.81 + 1.75
0.47 + 0.03
(4.79 + 0.90)×10
0.62 0.05
+
5.50 0.20
+
(8.87 0.80)×10
+
−
−
−
−
−
−
2
1.03 + 0.19
0.99 + 0.06
(9.61 + 1.88)×10
0.63 0.07
>5
+
5.11 0.24
ND
+
(8.11 0.96)×10
+
−
−
−
−
−
−
4.74 + 1.08
0.47 + 0.01
(9.92 + 2.26)×10
−
−
−
123.60 + 74.89
0.67 + 0.20
5.42 + 3.63
4
W126A
0.17 0.02
+
2.94 0.09
+
(1.73 0.16)×10
+
−
−
−
−
−
−
potential effect of ethanol on the enzymatic activity, the ethanol
concentration was adjusted to 5% (v/v) in both the activity assay
system and the control system. All experiments were repeated
three times under the same conditions.
To measure the effect of mutations on the binding of GSH,
the reaction mixture consisted of a constant concentration of
CDNB (0.5 mM, Sigma) and various concentrations of GSH
Scheme 1 The conjugation reaction of GSH and CDNB catalysed by hGSTk
(
0.05–200 mM, Sigma). To measure the effect of mutations on the
binding of CDNB, the reaction mixture consisted of a constant
concentration of GSH (1 mM) and various concentrations of
factor of 0.195 (Table 1). There are three hGSTk molecules in
an asymmetric unit with two molecules forming a dimer and
the third forming a dimer with a two-fold crystallographic axis-
related molecule. The crystal structure of hGSTk in complex with
GTX was refined to 1.90 Å resolution, yielding an R factor of
0.183 and a free R factor of 0.218 (Table 1). There are six
hGSTk molecules in an asymmetric unit forming three dimers
and each hGSTk is bound with a GTX molecule at the active site.
No significant conformational difference exists among the three
hGSTk molecules in the apo-form and the six hGSTk molecules in
the GTX-bound complex [the RMSD (root mean square deviation)
values are approximately 0.2 Å and 0.3 Å respectively]. GTX is an
inhibitor of GSTs with a hexyl group that conjugates to the thiol
group of GSH. In the structure of the hGSTk–GTX complex,
there was strong electron density without ambiguity for the GSH
moiety of GTX with an identical conformation at the active site of
each subunit. However, the hexyl group of GTX has two distinct
conformations in the two subunits of the dimer with well-defined
electron density and occupies part of the H site. Consistently, the
CDNB (0.05–5 mM). The apparent kinetic parameters K
m
and kcat
(
Tables 2 and 3) were determined by fitting the kinetic data to the
Michaelis–Menten equation using non-linear regression analysis
implemented in the program Prism 4.0 (GraphPad Software).
To calculate the initial velocity of the enzymatic reaction, the
kinetic data were fitted to eqn (1) for the rapid equilibrium random
sequential Bi Bi model [42,43] according to Scheme 1.
V
max [GSH] [CDNB]
v = αK
GSH CDNB
GSH
CDNB
Kd
+ αKd [CDNB] + αKd
[GSH] + [GSH] [CDNB]
d
(
1)
where v is the initial velocity of the enzymatic reaction, α is
GSH
the coupling factor, K
d
is the dissociation constant of GSH,
CDNB
and K
d
is the dissociation constant of CDNB (dissociation
GSH
constants of hGSTk for GSH and CDNB: K
d
0.16 + 0.04 mM,
−
CDNB
K
d
1.23 + 0.59 mM, α 0.51 + 0.04). Inhibition studies were
−
−
carried out in a similar manner as for the initial velocity studies
using GSDNB [S-(2,4-dinitrophenyl)glutathione] (provided by
the Xu laboratory, Shanghai University, Shanghai, China) as the
product inhibitor. The inhibitor was fixed at several concentrations
ranging from 0 to 66.7 μM. The parameters of the rate equation
were obtained from the Lineweaver–Burk plots and the secondary
plots that were constructed as described previously [44].
2
average B factor of the hexyl group of GTX (68.4 Å ) is much
2
higher than that of the GSH moiety of GTX (38.3 Å ) and that
2
of the protein (38.1 Å ), indicating a high flexibility of the hexyl
group.
The overall structure of hGSTk in the apo-form and the GTX-
bound complex is very similar to that in the GSF-bound complex
[
35] with an RMSD of <0.48 Å for 204 Cα atoms between the
RESULTS
apo-form and the GSF-bound complex, and <0.70 Å for 217
Cα atoms between the GTX-bound and GSF-bound complexes.
hGSTk consists of a Trx-fold domain and an α-helical domain
(Figures 1A and 1B). The Trx-fold domain (domain I) is composed
of a βαβα (β1α1β2α2) motif and a ββα (β3β4α10) motif
Overall structures of the apo-form and GTX-bound hGSTk
The crystal structure of the apo-form of hGSTk was refined to
1
.80 Å resolution, yielding an R factor of 0.152 and a free R
ꢀc The Authors Journal compilation ꢀc 2011 Biochemical Society

218
B. Wang and others
Figure 1 Overall structures of hGSTk
A) Overall structure of the apo-form of hGSTk. The disordered regions of residues Asp –Gly are indicated with dotted lines. The active sites are circled with dashed lines. (B) Overall structure
5
0
52
(
of the hGSTk–GTX complex. The bound GTX molecules are shown as ball-and-stick models. The active sites are circled with dashed lines. (C) Superimposition of the monomers of three hGSTk
5
0
52
structures. The apo-form of hGSTk is shown in pink, the GTX-bound hGSTk is in green, and the GSF-bound hGSTk is in yellow (PDB code 1YZX). The disordered region of residues Asp –Gly
in the apo-form hGSTk structure is indicated with a dotted line. The bound GTX molecule is shown as a ball-and-stick model. The regions with substantial conformational differences are circled
with dashed lines. (D) Sequence alignment of GSTks from different species. The alignment was generated by ESPript [51] with the secondary structures of hGSTk at the top of the alignment. The
conserved residues forming the G site are marked with red stars. HsGSTk, Homo sapiens GSTk1; MmGSTk, Mus musculus GSTk1; RnGSTk, Rattus norvegicus GSTk1; XtGSTk, Xenopus tropicalis
GSTk1; OmGSTk, Oncorhynchus mykiss GSTk1; CeGSTk, Caenorhabditis elegans GSTk1; and PpHCCA-iso, Pseudomonas putida HCCA isomerase.
forming a four-stranded β-sheet surrounded by three α-helices.
The α-helical domain (domain II) is composed of seven α-helices
denoted with a prime) are involved in direct interactions with the
GSH moiety of GTX (see the results and discussion below). In
50
52
(α3–α9), and is inserted between the N-terminal βαβα motif and
the apo-form of hGSTk, residues Asp –Gly are disordered and
48 49 53 54
the C-terminal ββα motif of the Trx-fold domain. This domain
topology is similar to that of rGSTk, but different from that of
the soluble GSTs in which the Trx-fold domain and the α-helical
domain are linked together contiguously. The dimeric hGSTk
has an open-wings butterfly shape which is also different from the
canonical globular shape of the soluble GSTs. The dimer interface
is dominated by hydrophobic interactions between residues of
domain I of one subunit and domain II of the other. As in
the flanking regions (Met –Lys and Asn –Lys ) exhibit higher
55 59
B factors. Furthermore, residues Pro –Leu form a loop and are
56
positioned away from the G site (the Cα atom of Pro in the apo-
form of hGSTk is displaced by approximately 14 Å away from that
in the GSF/GTX-bound hGSTk). In the hGSTk–GTX complex,
the α3–α4 loop assumes different conformations in the two
subunits to accommodate the hexyl group of GTX with different
conformations, both of which are slightly different from those
in the apo-form of hGSTk and the hGSTk–GSF complex (see
the results and discussion below). These results suggest that the
G and H sites have conformational flexibilities and the substrate
binding induces conformational changes at the active site.
184
most of the soluble GSTs, the conserved Pro at the N-terminal
β3 adopts a cis-conformation. Previous studies have shown that
the conserved cis-proline residue in the soluble GSTs plays an
important role in maintaining the conformational stability of the
active site [45,46].
The major conformational differences between the apo-form
and the GTX-bound hGSTk occur at the active site, in particular
Structure of the G site
46
61
the α2–α3 loop (residues Ile –Arg ) which forms part of the G
The G site of hGSTk resides in a hydrophilic cleft at the dimer
interface and is formed by structural elements of both subunits,
including helices α1 and α3 and the β4–α10 loop of domain I,
and the two connecting loops between domain I and domain II
(the α2–α3 and α9–β3 loops) of one subunit, and helix α3 and
the β4–α10 loop of the other subunit (Figures 1A and 1B). This
is slightly different from the soluble GSTs in which the G site is
located in a cleft between domain I of one subunit and domain
81
90
and H sites, and the α3–α4 loop (residues Ile –Val ) which forms
part of the H site (Figure 1C). In the hGSTk–GTX complex which
53
60
is similar to the hGSTk–GSF complex, residues Asn –Pro
form a surface-exposed loop with the region of residues
55
59
Pro –Leu forming a short 310 α-helix, and several residues of
53
56
this region including Asn6 and Pro , and a residue from the
ꢃ
2
other subunit, namely Lys (residues of the adjacent subunit are
ꢀc The Authors Journal compilation ꢀc 2011 Biochemical Society

Catalytic mechanism of human GSTk
219
Figure 2 Structure of the G site
A) Schematic diagram showing the interactions of the GSH moiety of GTX with the surrounding residues in the hGSTk–GTX complex. The salt bridges and hydrogen bonds are shown as dotted
(
lines, and the hydrophobic interactions are shown as dashed lines. The distances (in A˚ ) of the interactions are indicated. (B) A detailed view of the interactions between the two γ -glutamyl moieties
2
00
18
69
of GTX at the adjacent G sites in the hGSTk–GTX complex. The γ -glutamyl moiety of GTX is shown as a ball-and-stick model, and the main chain of Ser and the side chains of Tyr , Asp ,
2
00
201
202
Ser , Asp and Arg are shown as stick models. Three water molecules are shown as red spheres. The network of hydrogen bonds and salt bridges are indicated with dotted lines. (C) Structural
comparison of the G site of hGSTk in the apo-form (pink), the GTX-bound complex (green) and the GSF-bound complex (yellow). The disordered region of residues Asp –Gly in the apo-form of
hGSTk is indicated with a pink dotted line. GTX is shown as a ball-and-stick model, and the side chains of Ser , Ser , Asn , Pro , Lys and Leu , and the main chain of Leu are shown as
stick models. The hydrogen-bonding and salt bridge interactions are indicated with dotted lines.
5
0
52
ꢃ
1
6
19
53
56
62
183
183
1
8
18
II of the other, and residues of domain I provide the interactions
for the specific recognition of GSH [1,17–19]. In the hGSTk–
GTX complex, the GSH moiety of GTX maintains interactions
with several conserved residues of hGSTk (Figure 2A), similar
to those in the hGSTk–GSF complex [35], and most of the
interactions are also conserved in the rGSTk–GSH complex
of Tyr (2.7 Å). Meanwhile, the phenolic ring of Tyr forms
hydrophobic contacts with the γ -glutamyl moiety and the thiol
group of GSH. Additionally, there are three water molecules at
the dimer interface to mediate the interactions between the two
γ -glutamyl moieties of the adjacent subunits (Figure 2B). To
investigate the functional roles of these residues in catalysis, we
performed mutagenesis and steady-state kinetics studies (Table 2).
[
31]. Specifically, the γ -glutamyl moiety of GSH is mainly
recognized by the conserved strand-turn-helix motif (the β4–α10
Mutation R202A resulted in a significantly increased K of GSH
m
ꢃ
202
200
201
loop) consisting of Ser , Asp and Arg
(Figures 2A and
(661-fold) and a slightly decreased kcat (1.8-fold), and mutation
2
B). The γ -glutamyl carboxylate group makes a salt bridge with
D201A resulted in a moderately increased K of GSH (25-fold)
m
ꢃ
02
2
the side-chain amino of Arg (3.1 Å) and two hydrogen bonds
with the main-chain amide (2.7 Å) and the side-chain hydroxy
group (3.0 Å) of Ser respectively, and the γ -glutamyl amino
and a slightly decreased kcat (2.5-fold). Mutation S200A had
relatively less profound effects on the Km of GSH (increased
5.5-fold) and the kcat (decreased 1.2-fold) than mutations D201A
200
69
group makes a salt bridge with the side-chain carboxyl group of
and R202A. Replacement of Asp by alanine resulted in a
ꢃ
201
202
Asp (3.0 Å). The positively charged side chain of Arg also
forms two salt bridges with the negatively charged side chains
564-fold increase in the Km of GSH and a 3.7-fold decrease
in the k , and that by threonine resulted in a 326-fold increase in
cat
201
69
69
of Asp and Asp (2.8 Å for both). The side chain of Asp
makes another hydrogen bond with the side-chain hydroxy group
the K of GSH, but had no appreciable effect on the k . These
m
cat
200
69
201
results indicate that the conserved residues Asp , Ser , Asp
ꢀc The Authors Journal compilation ꢀc 2011 Biochemical Society

220
B. Wang and others
2
02
and Arg play very important roles in the binding of GSH,
but less profound roles in catalysis through stabilization of the
Structure of the H site
Owing to the lack of a hydrophobic substrate-bound GSTk
structure, the exact binding site for the hydrophobic substrate
in GSTk is not well defined. In the hGSTk–GTX complex, there
is a large hydrophobic pocket near the G site, and the hexyl group
of GTX occupies part of the pocket with distinct conformations
in the two subunits of the dimer and the two conformers differ
18
γ -glutamyl moiety of GSH. Surprisingly, mutation of Tyr to
alanine, leucine or histidine led to undetected binding of GSH
and enzymatic activity, and mutation of Tyr to phenylalanine
yielded a markedly increased K of GSH (109-fold), but had a
slightly increased kcat (1.3-fold), indicating that Tyr plays critical
roles in both GSH binding and catalysis. It is possible that Tyr is
not only involved in stabilization of the γ -glutamyl carboxylate
of GSH via the hydrogen-bonding interaction with Asp , but also
in precise positioning and/or correct orientation of the thiol group
and the γ -glutamyl moiety of GSH via hydrophobic interactions.
As in the soluble GSTs and rGSTk, the sulfur atom of the
cysteinyl moiety of GSH is recognized via a hydrogen bond (3.0
Å) by the side-chain hydroxy group of Ser at the N-terminus
of α1, which is further stabilized via a hydrogen bond by the
side-chain hydroxy group of Ser (2.8 Å) (Figures 2A and 2C).
Additionally, the main-chain amino and carbonyl groups of the
cysteinyl moiety form two hydrogen bonds with the main-chain
carbonyl and amino groups of Leu at the α9–β3 loop (2.8 Å for
both). The equivalent of Ser in rGSTk and the soluble GSTs was
proposed to act as a catalytic residue in the ionization of the thiol
group of GSH [19,21–23,31]. The functional roles of Ser and
the other residues in catalysis are confirmed by mutagenesis and
steady-state kinetic studies (Table 2). The biochemical data show
that mutation S16A caused a greatly increased K
fold) and a markedly decreased kcat (6.9-fold), mutation S19A
resulted in a slightly increased K of GSH (8.6-fold) but a
moderately increased kcat (3.2-fold), and mutation L183A caused
a 53-fold increase in the K of GSH and a 2.5-fold decrease in
the kcat. These results support the notion that the conserved Ser
plays a critical role in catalysis. Structural comparison shows
that all of these residues and the structural elements interacting
with the cysteinyl moiety of GSH maintain similar conformations
in the apo-form and GSF/GTX-bound hGSTk, indicating that
the proposed catalytic residue maintains a stable conformation in
the substrate binding.
The carboxylate of the glycyl moiety of GSH makes a hydrogen
bond with the side-chain amide of Asn (2.7 Å) and the main
chain of the glycyl moiety makes hydrophobic contacts with the
side chain of Pro (Figures 2A and 2C). The biochemical data
show that mutation N53A resulted in a 24-fold increase in the
of GSH and a 1.5-fold increase in the kcat, and mutation P56A
yielded a 9-fold increase in the K of GSH and a 1.9-fold increase
in the kcat (Table 2). Since mutations of Asn and Pro , both of
which are located in the α2–α3 connecting loop that undergoes
conformational change upon the GTX/GSF binding, have much
18
m
18
18
◦
by 70 : in subunit A, the hexyl group has hydrophobic contacts
15
17
66
88
91
92
with the side chains of Leu , Pro , Met , Leu , Met and Leu ;
in subunit B, the hexyl group has hydrophobic contacts with the
69
17
18
66
83
88
126
side chains of Pro , Tyr , Met , Phe , Leu and Trp , and
67
the main chain of Ala (Figure 3A). Concurrently, the α3–α4
loop assumes different conformations in the two subunits which
are also different from those in the apo-form and GSF-bound
hGSTk structures (Figure 1C). Compared with that in subunit A,
the α3–α4 loop in subunit B shifts towards helix α3 with the
16
19
85
Cα atom of Lys in the middle of the loop displaced by 4.5 Å.
83
87
91
In addition, Phe , Phe and Met assume different side-chain
conformations to accommodate the hexyl group of GTX with
different conformations.
183
16
To better understand the binding site of the hydrophobic
substrate in hGSTk, we docked a GSDNB molecule, the
conjugation product of GSH and CDNB, into the active site of
hGSTk based on the position of the GSH moiety of GTX or GSF.
GSDNB can be docked into the active site of hGSTk in both the
apo-form and the GTX/GSF-bound complex, and particularly
the CDNB moiety of GSDNB can be accommodated into the
large hydrophobic pocket without obvious steric conflict with
the protein (Figures 3B and 3C). On the basis of the modelling
studies, the H site of hGSTk can be defined as a large hydrophobic
16
m
of GSH (93-
m
m
16
15
18
44
pocket composed of Leu –Tyr of the β1–α1 loop and Ile
of helix α2 of domain I, Met of helix α3, Leu –Leu of the
66
88
92
126
55
α3-α4 loop, and Trp of helix α6 of domain II, and Pro –
58
Leu of the α2–α3 loop which links domain I and domain
II (Figure 3B). This is also slightly different from the soluble
GSTs in which the H site is mainly formed by the structural
elements of domain II [1,17–19]. Owing to the conformational
differences of the α3–α4 loop and the hexyl group of GTX
in the two subunits of the hGSTk–GTX complex, the CDNB
moiety of GSDNB can be docked into the H site in two slightly
differed conformations (Figure 3B). This implies that the shape
and size of the H site in hGSTk may vary to some extent to
bind different hydrophobic substrates. The kinetic data show
that mutations of the residues forming the H site had various
effects on the hydrophobic substrate binding but insignificant
53
56
K
m
m
53
56
15
44
effects on the kcat (Table 3). Specifically, mutations of Leu , Ile ,
66
88
92
Met , Leu and Leu which make close hydrophobic interactions
with the CDNB moiety of GSDNB and undergo conformational
changes upon the binding of GTX or GSF, had relatively large
less effect on the K of GSH than the residues that are involved in
m
interactions with the γ -glutamyl and cysteinyl moieties of GSH,
these results suggest that the α2–α3 loop is involved in the binding
of GSH, but plays a less critical role than the residues interacting
with the γ -glutamyl and cysteinyl moieties. Previously it was
reported that the soluble GSTs contain a functionally conserved
basic residue (histidine, lysine or arginine) at the active site that
interacts directly with the glycyl carboxylate of GSH and plays an
effects on the K
m
of CDNB (increased approximately 2–5-fold),
probably because of the loss of favourable hydrophobic contacts.
17
126
Mutations of Pro and Trp , which also make close hydrophobic
interactions with the CDNB moiety but have no conformational
change upon the substrate binding, yielded a markedly decreased
K of CDNB (approximately 3–6-fold), probably because these
m
important role in GSH binding and catalysis [47]. In the hGSTk–
GTX complex, the side-chain amino group of Lys makes a salt
bridge with the carboxylate of the glycyl moiety of GSH (2.8 Å)
ꢃ
mutations might render the H site more flexible and/or more
spacious to accommodate the hydrophobic substrate. The other
62
55
56
58
83
87
91
residues, including Pro , Pro , Leu , Phe , Phe and Met ,
have less or indirect interactions with the CDNB moiety, and
their mutations had minor effects on the K of CDNB (<2 fold).
53
as well as a hydrogen bond with the side-chain carbonyl of Asn
62
(
3.5 Å) (Figures 2A and 2C). Mutation of Lys to glutamate or
alanine increased the K of GSH 135-fold or 7-fold, but had a
minor effect on the kcat (reduced 1.6- or 1.3-fold) (Table 2). These
results indicate that Lys of hGSTk plays a similar functional
role in the binding of GSH to that of the conserved basic residue
of the soluble GSTs.
m
m
ꢃ
2
Kinetic analysis of the catalytic reaction
6
Previously it was proposed that the catalytic reaction of rGSTk
utilizes a rapid equilibrium ordered mechanism with GSH binding
ꢀc The Authors Journal compilation ꢀc 2011 Biochemical Society

Catalytic mechanism of human GSTk
221
Figure 3 Structure of the H site
A) A detailed view showing that the hexyl group of GTX assumes different conformations in the two subunits of the hGSTk–GTX complex. The GTX molecules are covered with a representative
(
SIGMAA-weighted 2F map (1σ contour level). The residues that have hydrophobic interactions with the hexyl group of GTX are shown as stick models and covered with the van der Waals
surfaces. (B) Structure of the H site showing the interactions of the docked GSDNB with the surrounding residues in the hGSTk–GTX complex. Owing to the conformational differences of the hexyl
group of GTX in the two subunits, the CDNB moiety of GSDNB can be docked into the H site in two slightly different conformations. (C) Electrostatic surfaces of hGSTk in the apo-form (left-hand
panel), the GSF-bound complex (middle panel) and the GTX-bound complex (right-hand panel) showing the structure of the H site. The disordered regions of residues Asp –Gly in the apo-form
are indicated with dotted circles. The docked GSDNB molecule at the H site is shown in yellow. Some residues of the active site are shown as stick models for reference.
o
− F
c
5
0
52
first which induces conformational changes at the active site to
facilitate CDNB binding [31]. To gain insights into the kinetic
mechanism of hGSTk, we carried out steady-state kinetic studies
using GSH and CDNB as the substrates. When GSH was used
as the variable substrate and CDNB was used at fixed concent-
rations, the intersecting point of the Lineweaver–Burk plots is
(Figures 4C and 4D). The K for GSH is 65.9 + 3.2 μM
i
−
and 111.9 + 28.9 μM at the CDNB concentration of 0.5 mM and
−
1 mM respectively. These data suggest that the kinetic mechanism
of hGSTk conforms to the rapid equilibrium random sequential
Bi Bi model [44]. The random binding of GSH and CDNB has
a coupling factor α of 0.51 + 0.04, indicating that the affinity
−
above the horizontal axis, indicating that the apparent K
m
of
for CDNB increases in the presence of GSH approximately 2-
GSH
CDNB
CDNB decreases as the GSH concentration increases (Figure 4A).
A similar pattern was obtained when CDNB was used as the
variable substrate and GSH at fixed concentrations (Figure 4B).
These intersecting initial velocity patterns indicate that the
kinetics of hGSTk are consistent with a sequential mechanism
with both substrates binding to the enzyme before the product
is released [44]. However, as the curvatures of the reciprocal
plots of the Lineweaver–Burk plots are too small, it is difficult
to fit the kinetic data to a steady-state random sequential Bi Bi
model or a steady-state ordered sequential Bi Bi model. To resolve
the issue on the binding order of the two substrates, we further
performed the inhibition studies with the product GSDNB. The
results show that GSDNB exhibits a competitive inhibition and
a mixed-type inhibition towards GSH and CDNB respectively
fold and vice versa (Kd
is 0.16 + 0.04 mM, and K
is
d
−
1.23 + 0.59 mM, α 0.51 + 0.04). Nevertheless, hGSTkbinds GSH
−
−
CDNB
much more tightly than CDNB as the dissociation constant K
d
GSH
is much higher than Kd . The coupling of the random binding
of GSH and CDNB also implies the existence of an intra-subunit
communication between the G and H sites. These kinetic data are
supported, in part, by the structural data that in the apo-form of
hGSTk both G and H sites are accessible to the solvent to allow the
random binding of the substrates, and the binding of GSH at the
G site induces conformational changes of the α2–α3 and α3–α4
loops to facilitate the formation of the H site and presumably the
binding of CDNB. We expect that the binding of CDNB at the H
site would also induce conformational changes of the α2–α3 and
α3–α4 loops to facilitate the formation of the G site and the
ꢀc The Authors Journal compilation ꢀc 2011 Biochemical Society

222
B. Wang and others
Figure 4 Steady-state kinetics and inhibition analysis of hGSTk
(
A) Initial velocity analysis with GSH at various concentrations and CDNB at fixed concentrations: 0.2 (ꢀ), 0.5 (᭺), 0.75 (∇) and 1.0 (ᮀ) mM. (B) Initial velocity analysis with CDNB at various
concentrations and GSH at fixed concentrations: 0.5 (ꢀ), 0.75 (᭺), 1.5 (∇) and 2.0 (ᮀ) mM. (C) Lineweaver–Burk plots for the inhibition of hGSTk by product GSDNB at different concentrations of
GSH. (D) Lineweaver–Burk plots for the inhibition of hGSTk by product GSDNB at different concentrations of CDNB. The assays were performed in the absence or presence of different concentrations
of GSDNB: 0 (᭹), 13.3 (᭺), 33.4 (᭜) and 66.7 (᭛) μM.
binding of GSH. The crystal structure of hGSTk in complex with
CDNB is needed to address this issue.
of hydrogen bonds mediated by surrounding residues (serine and
threonine) and water molecules. The residues contributing to
the electron-sharing network are not conserved in the primary
sequence, but are structurally and functionally conserved in most
of the soluble GSTs including the Alpha, Pi, Theta and Delta
classes [25]. Previously, a molecular dynamics simulation study
further suggested that a water molecule at the active site acts as a
bridge to assist the transfer of the proton from the thiol group to
the γ -glutamyl carboxylate of GSH after an initial conformational
rearrangement of GSH [17].
hGSTk contains a strictly conserved Ser at the active site
which forms a hydrogen bond with the sulfur atom of GTX in
the GTX-bound or GSF in the GSF-bound complex (Figures 1D
and 2A). In addition, hGSTk also contains a strictly conserved
SDR motif which forms a network of salt bridges and hydrogen
bonds with the γ -glutamyl carboxylate of GSH (Figures 1D
and 2A). In the hGSTk–GTX and hGSTk–GSF complexes,
DISCUSSION
GSTks constitute a distinct family of the Trx-fold superfamily,
different from the soluble GSTs in both function and catalytic
mechanism [5,26–29]. The previous structural studies of rGSTk
and hGSTk have revealed some unique structural features
of GSTks. We determined the crystal structures of hGSTk
in the apo-form and in complex with GTX and carried out
mutagenesis and steady-state kinetic studies of hGSTk. The
structural and biochemical data taken together provide new
insights into the catalytic mechanism of hGSTk and the other
GSTks.
In the conjugation reaction catalysed by the soluble GSTs,
the ionization of GSH to form the thiolate anion is an essential
step. Previous structural and biochemical studies of the soluble
GSTs have led to the proposal of two GSH activation models.
The active-site catalytic residue model suggests that a conserved
residue (tyrosine in Alpha, Mu and Pi classes, serine in Theta
and Delta classes, and cysteine in the Omega class) at the active
site functions as a base to receive the proton from the thiol
group of GSH and stabilizes the thiolate anion via a hydrogen
bond [19,21–23]. Alternatively, the base-assisted deprotonation
model suggests that the γ -glutamyl carboxylate of GSH acts
as a base to accept the proton from the thiol group of GSH,
which is assisted by a network of hydrophilic interactions to
share the negative charge of the γ -glutamyl carboxylate and to
stabilize the thiolate anion [20,24,25]. The proposed electron-
sharing network is characterized by electrostatic interactions
between the γ -glutamyl carboxylate, a positively charged residue
16
the γ -glutamyl carboxylate of GSH forms an electrostatic
ꢃ
202
interaction with the positively charged Arg which makes two
201
electrostatic interactions with the negatively charged Asp and
69
Asp (Figure 2B). These interactions are further stabilized by
18
200
a network of hydrogen bonds involving Tyr , Ser and three
water molecules. Our mutagenesis data show that mutations of
16
Ser and the residues involved in the formation of the interaction
network with the γ -glutamyl carboxylate have significant effects
on the binding of GSH and the kcat, underscoring their functional
16
importance in catalysis (Table 2). However, mutation of Ser
that makes a direct hydrogen bond with the thiol group of
GSH has the most significant effect on the kcat than the other
residues, and mutations of the residues that are involved in salt-
bridge interactions with the γ -glutamyl carboxylate of GSH have
the most significant effects on the binding of GSH than the
other residues. Furthermore, the structural data show that the γ -
glutamyl carboxylate of GSH is positioned >8 Å away from the
(
primarily arginine), and a negatively charged residue (primarily
glutamate or aspartate), which are further stabilized by an array
ꢀc The Authors Journal compilation ꢀc 2011 Biochemical Society

Catalytic mechanism of human GSTk
223
thiol group, suggesting that it is unlikely to be involved directly
or indirectly via a water molecule in the deprotonation of the thiol
group. Thus our structural and biochemical data together support
the active-site catalytic residue model for the GSH activation in
it is conceivable that the binding of a hydrophobic substrate at
the H site may induce conformational changes of the H and G
sites, leading to the formation of a ‘closed’ active site which,
however, has a great flexibility and can undergo conformational
changes at the G and/or H sites to allow the binding of GSH. The
crystal structure of hGSTk in complex with CDNB will eventually
resolve this issue.
16
which Ser functions as the catalytic base in thiol deprotonation,
and the γ -glutamyl carboxylate of GSH plays an important role in
stabilizing the deprotonated thiolate anion rather than functioning
69
as a catalytic base. The network of interactions involving Asp ,
It has been shown previously that several classes of the dimeric
soluble GSTs (including the Alpha, Mu, Pi and Delta classes)
exhibit a positive co-operativity between the two active sites upon
GSH binding [47–50]. These soluble GSTs contain an aromatic
residue in the connecting loop between α2 and β3 at the active site.
This aromatic residue forms interactions with several hydrophobic
residues of the adjacent subunit and acts as a hydrophobic lock in
the dimerization, and thus plays a pivotal role in the co-operative
work of the two active sites. For hGSTk, the kinetic data of
both wild-type and mutant enzymes fit well with the Michealis–
Menten equation and the regression coefficients of the plots are
all greater than 0.95 at a variety of substrate concentrations
(Tables 2, 3 and 4), which are consistent with the previous studies
of rGSTk [31]. These results indicate that the two active sites of
the dimeric hGSTk do not display co-operativity for both GSH
and CDNB. On the other hand, the structures of hGSTk show
that the G site of hGSTk is formed by structural elements of both
subunits and several conserved residues of both subunits provide
the interactions for the specific recognition of GSH (Figures 1A
and 1B). In particular, the interactions responsible for stabilization
of the γ -glutamyl and glycyl carboxylates of GSH in one subunit
200
201
202
Ser , Asp , Arg and water molecules help to stabilize the
negatively charged γ -glutamyl carboxylate of GSH. In addition,
18
the strictly conserved Tyr also plays an important role in catalysis
through stabilization and precise positioning of the γ -glutamyl
carboxylate and the thiol group of GSH.
Structural comparison of the apo-form and the GTX- and GSF-
bound hGSTk reveals that substrate binding induces substantial
conformational changes of the active site from an ‘open’
conformation to a ‘closed’ conformation. In particular, the α2–
α3 loop which constitutes part of the G and H sites assumes a
partially disordered loop structure and is positioned away from
the active site in the apo-form of hGSTk, but adopts an ordered
loop structure and interacts directly with the GSH moiety of
GSF or GTX in the substrate-bound complexes. The kinetic
data confirm that the key residues of this region, including
53
56
62
Asn , Pro and Lys , play important roles in the binding
of GSH and in catalysis. In addition, the α3–α4 loop which
constitutes part of the H site also assumes different conformations
among different hGSTk structures, and is involved in the binding
of the hexyl group of GTX. The kinetic data also confirm that
ꢃ
ꢃ
202
88
92
62
the key residues of this region, particularly Leu and Leu , are
involved in the binding of the hydrophobic substrate. In addition,
the structural data show that the G and H sites are coupled
together via a shared structural element (the α2–α3 loop), and
the conformational change at one site upon the binding of one
substrate could induce conformational change of the other site to
facilitate the binding of the other substrate. This is consistent with
the kinetic data showing that hGSTk utilizes a rapid equilibrium
random sequential Bi Bi mechanism, and the random binding of
GSH and CDNB are coupled together with a coupling factor
of 0.51. Although conformational changes of the active site upon
substrate binding have been observed in some classes of the
soluble GSTs [43,48], the conformational change of the α2–α3
loop at the active site of hGSTk upon substrate binding is unique
as this loop connects domains I and II and forms part of the G and
H sites. The large conformational changes of both G and H sites
in hGSTk might be beneficial to the enzyme to be able to bind
a wide range of substrates, including the canonical hydrophobic
substrates and potential protein substrates.
In the apo-form of hGSTk, the active site assumes an ‘open’
conformation, and both the G and H sites are accessible to the
solvent allowing the random binding of the substrates (Figure 3C).
In the GTX- or GSF-bound hGSTk, the binding of GTX or
GSF renders the enzyme a ‘closed’ active site owing to the
conformational changes of the α2–α3 and α3–α4 loops, and
the H site is inaccessible to the solvent even though it is only
partially occupied or unoccupied (Figure 3C). The ‘closed’ H
site in the GTX- or GSF-bound hGSTk structure appears to be
in disagreement with the proposed kinetic mechanism of random
sequential binding of GSH and CDNB. However, our structural
data also show that the α3–α4 loop in both the GSF- and GTX-
bound complexes exhibits relatively high B factors, indicating a
high flexibility of this region even in the presence of GSH. Thus
we propose that the binding of GSH induces the formation of a
involve two positively charged residues (Lys and Arg ) of the
adjacent subunit. This is different from the soluble GSTs in which
the G site is located in a cleft between domain I of one subunit
and domain II of the other subunit, but only residues of domain
I provide the interactions for the specific recognition of GSH.
Compared with the soluble GSTs, the two G sites of the dimeric
hGSTk are much closer to each other owing to the unique domain
topology and the tight dimer interface (the amino groups of the
γ -glutamyls of the two GTX molecules at the adjacent G sites are
only 5.5 Å apart) (Figure 2B). A similar case is observed in rGSTk
[31]. Taken together, these results demonstrate that the two active
sites of hGSTk are connected, but do not exhibit co-operativity
for the substrate binding.
GSTks exist widely in the mitochondria of a variety of species
ranging from bacteria to mammals. Sequence alignment of GSTks
from different species indicates that the residues forming the G
site and the residues interacting with GSH are highly conserved
16
in all GSTks, including Ser that interacts with the sulfur of
the thiol group of GSH and functions as the catalytic residue;
6
9
200
201
202
Asp , Ser , Asp and Arg that are involved in interactions
53 56
with the γ -glutamyl moiety of GSH; and Asn , Pro and
62
Lys that are involved in interactions with the glycyl moiety
of GSH (Figure 1D). In addition, the α2–α3 connecting loop
that undergoes conformational change upon substrate binding is
also highly conserved in GSTks. On the other hand, the residues
involved in hydrophobic substrate binding are less conserved
among different GSTks, even between human and rodent GSTks
(Figure 1D). Nevertheless, the structural elements forming the
H site and the shape of the H site appear to be comparable in
human and rat GSTks. In addition, HCCA (2-hydroxychromene-
2-carboxylic acid) isomerase, a key enzyme in the naphthalene
catabolic pathway of Pseudomonas putida which is suggested
to function as a GSTk [27], has a similar overall structure and
catalytic active site as hGSTk, even though the two enzymes
share a low sequence identity of approximately 20% (results
not shown). The G and H sites of HCCA isomerase are also
comparable with those of hGSTk and undergo conformational
‘
closed’ active site, but the α3–α4 loop that forms part of the H
site has a great flexibility and can undergo conformational change
to allow the binding of the hydrophobic substrate. Conversely,
ꢀc The Authors Journal compilation ꢀc 2011 Biochemical Society

224
B. Wang and others
changes upon substrate binding [27]. From an evolutionary
perspective, variation of the residues forming the H site in
GSTks from different species might confer the varied selectivity
of the enzymes for a wide range of substrates and/or different
enzymatic activities towards the same substrate depending on the
environment to reach a better balance between the detrimental and
beneficial effects of the enzymes. Taken together, it is very likely
that the other GSTks might utilize a similar catalytic mechanism
as that of hGSTk.
14 Singh, N., Sinha, N., Kumar, S., Pandey, C. M. and Agrawal, S. (2011) Glutathione
S-transferase gene polymorphism as a susceptibility factor for acute myocardial infarction
and smoking in the north Indian population. Cardiology 118, 16–21
1
5
Hamed, R. R., Maharem, T. M., Abdel-Meguid, N., Sabry, G. M., Abdalla, A. M. and
Guneidy, R. A. (2011) Purification and biochemical characterization of glutathione
S-transferase from Down syndrome and normal children erythrocytes: a comparative
study. Res. Dev. Disabil. 32, 1470–1482
1
1
1
1
6
7
8
9
Vasieva, O. (2011) The many faces of glutathione transferase pi. Curr. Mol. Med. 11,
1
29–139
Dourado, D. F., Fernandes, P. A., Mannervik, B. and Ramos, M. J. (2008) Glutathione
transferase: new model for glutathione activation. Chemistry 14, 9591–9598
Armstrong, R. N., Rife, C. and Wang, Z. (2001) Structure, mechanism and evolution of
thiol transferases. Chem.-Biol. Interact. 133, 167–169
Caccuri, A. M., Antonini, G., Nicotra, M., Battistoni, A., Lo Bello, M., Board, P. G., Parker,
M. W. and Ricci, G. (1997) Catalytic mechanism and role of hydroxyl residues in the
active site of theta class glutathione S-transferases: investigation of Ser-9 and Tyr-113 in
a glutathione S-transferase from the Australian sheep blowfly, Lucilia cuprina. J. Biol.
Chem. 272, 29681–29686
AUTHOR CONTRIBUTION
Bing Wang carried out the structural and biochemical studies and drafted the paper. Yingjie
Peng and Tianlong Zhang participated in the structural studies. Jianping Ding conceived
the study, participated in the experimental design, and wrote the paper.
2
2
2
0
1
2
Caccuri, A. M., Antonini, G., Board, P. G., Parker, M. W., Nicotra, M., Lo Bello, M.,
Federici, G. and Ricci, G. (1999) Proton release on binding of glutathione to Alpha, Mu
and Delta class glutathione transferases. Biochem. J. 344, 419–425
Kolm, R. H., Sroga, G. E. and Mannervik, B. (1992) Participation of the phenolic hydroxyl
group of Tyr-8 in the catalytic mechanism of human glutathione transferase P1-1.
Biochem. J. 285, 537–540
Allardyce, C. S., McDonagh, P. D., Lian, L. Y., Wolf, C. R. and Roberts, G. C. (1999) The
role of tyrosine-9 and the C-terminal helix in the catalytic mechanism of Alpha-class
glutathione S-transferases. Biochem. J. 343, 525–531
ACKNOWLEDGEMENTS
We thank the staff members at the Shanghai Synchrotron Radiation Facility, Shanghai,
China, for support in diffraction data collection, Professor Bin Xu at Shanghai University
for providing GSDNB, and other members of our group for helpful discussion.
FUNDING
This work was supported by the Ministry of Science and Technology of China [grant
numbers 2011CB911102, 2011CB966301]; the National Natural Science Foundation
of China [grant number 30730028]; the Chinese Academy of Sciences [grant number
SIBS2008002]; and the Science and Technology Commission of Shanghai Municipality
23 Atkinson, H. J. and Babbitt, P. C. (2009) Glutathione transferases are structural and
functional outliers in the thioredoxin fold. Biochemistry 48, 11108–11116
2
2
4
5
Gustafsson, A., Pettersson, P. L., Grehn, L., Jemth, P. and Mannervik, B. (2001) Role of
the glutamyl alpha-carboxylate of the substrate glutathione in the catalytic mechanism of
human glutathione transferase A1-1. Biochemistry 40, 15835–15845
Winayanuwattikun, P. and Ketterman, A. J. (2005) An electron-sharing network involved
in the catalytic mechanism is functionally conserved in different glutathione transferase
classes. J. Biol. Chem. 280, 31776–31782
[grant number 10JC1416500].
REFERENCES
1
2
3
4
5
6
Armstrong, R. N. (1997) Structure, catalytic mechanism, and evolution of the glutathione
transferases. Chem. Res. Toxicol. 10, 2–18
Oakley, A. J. (2005) Glutathione transferases: new functions. Curr. Opin. Struct. Biol. 15,
26 Jowsey, I. R., Thomson, R. E., Orton, T. C., Elcombe, C. R. and Hayes, J. D. (2003)
Biochemical and genetic characterization of a murine class Kappa glutathione
S-transferase. Biochem. J. 373, 559–569
7
16–723
Frova, C. (2006) Glutathione transferases in the genomics era: new insights and
perspectives. Biomol. Eng. 23, 149–169
Hayes, J. D., Flanagan, J. U. and Jowsey, I. R. (2005) Glutathione transferases. Annu. Rev.
Pharmacol. Toxicol. 45, 51–88
Nebert, D. W. and Vasiliou, V. (2004) Analysis of the glutathione S-transferase (GST) gene
family. Hum. Genomics 1, 460–464
Sheehan, D., Meade, G., Foley, V. M. and Dowd, C. A. (2001) Structure, function and
evolution of glutathione transferases: implications for classification of non-mammalian
members of an ancient enzyme superfamily. Biochem. J. 360, 1–16
Mannervik, B., Board, P. G., Hayes, J. D., Listowsky, I. and Pearson, W. R. (2005)
Nomenclature for mammalian soluble glutathione transferases. Methods Enzymol. 401,
2
7
Thompson, L. C., Ladner, J. E., Codreanu, S. G., Harp, J., Gilliland, G. L. and Armstrong,
R. N. (2007) 2-Hydroxychromene-2-carboxylic acid isomerase: a kappa class glutathione
transferase from Pseudomonas putida. Biochemistry 46, 6710–6722
Pemble, S. E., Wardle, A. F. and Taylor, J. B. (1996) Glutathione S-transferase class
Kappa: characterization by the cloning of rat mitochondrial GST and identification of a
human homologue. Biochem. J. 319, 749–754
2
8
29 Robinson, A., Huttley, G. A., Booth, H. S. and Board, P. G. (2004) Modelling and
bioinformatics studies of the human Kappa-class glutathione transferase predict a novel
third glutathione transferase family with similarity to prokaryotic
7
2
-hydroxychromene-2-carboxylate isomerases. Biochem. J. 379, 541–552
1
–8
3
3
0
1
Morel, F., Rauch, C., Petit, E., Piton, A., Theret, N., Coles, B. and Guillouzo, A. (2004)
Gene and protein characterization of the human glutathione S-transferase kappa and
evidence for a peroxisomal localization. J. Biol. Chem. 279, 16246–16253
Ladner, J. E., Parsons, J. F., Rife, C. L., Gilliland, G. L. and Armstrong, R. N. (2004)
Parallel evolutionary pathways for glutathione transferases: structure and mechanism of
the mitochondrial class kappa enzyme rGSTK1-1. Biochemistry 43, 352–361
8
9
Allocati, N., Federici, L., Masulli, M. and Di Ilio, C. (2009) Glutathione transferases in
bacteria. FEBS J. 276, 58–75
Adler, V., Yin, Z., Fuchs, S. Y., Benezra, M., Rosario, L., Tew, K. D., Pincus, M. R.,
Sardana, M., Henderson, C. J., Wolf, C. R. et al. (1999) Regulation of JNK signaling by
GSTp. EMBO J. 18, 1321–1334
1
0
Cho, S. G., Lee, Y.H, Park, H. S., Ryoo, K., Kang, K. W., Park, J., Eom, S. J., Kim, M. J.,
Chang, T. S., Choi, S. Y. et al. (2001) Glutathione S-transferase mu modulates the
stress-activated signals by suppressing apoptosis signal-regulating kinase 1. J. Biol.
Chem. 276, 12749–12755
32 Liu, M., Zhou, L., Xu, A., Lam, K. S., Wetzel, M. D., Xiang, R., Zhang, J., Xin, X., Dong, L.
Q. and Liu, F. (2008) A disulfide-bond A oxidoreductase-like protein (DsbA-L) regulates
adiponectin multimerization. Proc. Natl. Acad. Sci. U.S.A. 105, 18302–18307
1
1
1
2
Johansson, A. S. and Mannervik, B. (2001) Human glutathione transferase A3-3, a highly
efficient catalyst of double-bond isomerization in the biosynthetic pathway of steroid
hormones. J. Biol. Chem. 276, 33061–33065
Li, Y. J., Oliveira, S. A., Xu, P., Martin, E. R., Stenger, J. E., Scherzer, C. R., Hauser, M. A.,
Scott, W. K., Small, G. W., Nance, M. A. et al. (2003) Glutathione S-transferase omega-1
modifies age-at-onset of Alzheimer disease and Parkinson disease. Hum. Mol. Genet. 12,
3
3
Inaba, K., Murakami, S., Suzuki, M., Nakagawa, A., Yamashita, E., Okada, K. and Ito, K.
2006) Crystal structure of the DsbB-DsbA complex reveals a mechanism of disulfide
bond generation. Cell 127, 789–801
Ito, K. and Inaba, K. (2008) The disulfide bond formation (Dsb) system. Curr. Opin. Struct.
Biol. 18, 450–458
Li, J., Xia, Z. and Ding, J. (2005) Thioredoxin-like domain of human kappa class
glutathione transferase reveals sequence homology and structure similarity to the theta
class enzyme. Protein Sci. 14, 2361–2369
(
3
3
4
5
3
259–3267
1
3
Pedersen, J. Z., De Maria, F., Turella, P., Federici, G., Mattei, M., Fabrini, R., Dawood, K.
F., Massimi, M., Caccuri, A. M. and Ricci, G. (2007) Glutathione transferases sequester
toxic dinitrosyl-iron complexes in cells: a protection mechanism against excess nitric
oxide. J. Biol. Chem. 282, 6364–6371
36 Otwinowski, Z. and Minor, W. (1997) Processing of X-ray diffraction data collected in
oscillation mode. Methods Enzymol. 276, 307–326
ꢀc The Authors Journal compilation ꢀc 2011 Biochemical Society

Catalytic mechanism of human GSTk
225
3
3
3
7
8
9
Vagin, A. and Teplyakov, A. (2010) Molecular replacement with MOLREP. Acta Crystallogr.
Sect. D Biol. Crystallogr. 66, 22–25
Brunger, A. T. (2007) Version 1.2 of the Crystallography and NMR system. Nat. Protoc. 2,
45 Allocati, N., Casalone, E., Masulli, M., Ceccarelli, I., Carletti, E., Parker, M. W. and Di Ilio,
C. (1999) Functional analysis of the evolutionarily conserved proline 53 residue in
Proteus mirabilis glutathione transferase B1-1. FEBS Lett. 445, 347–350
46 Nathaniel, C., Wallace, L. A., Burke, J. and Dirr, H. W. (2003) The role of an evolutionarily
conserved cis-proline in the thioredoxin-like domain of human class Alpha glutathione
transferase A1-1. Biochem. J. 372, 241–246
47 Vararattanavech, A. and Ketterman, A. J. (2007) A functionally conserved basic residue in
glutathione transferases interacts with the glycine moiety of glutathione and is pivotal for
enzyme catalysis. Biochem. J. 406, 247–256
2
728–2733
Adams, P. D., Grosse-Kunstleve, R. W., Hung, L. W., Ioerger, T. R., McCoy, A. J., Moriarty,
N. W., Read, R. J., Sacchettini, J. C., Sauter, N. K. and Terwilliger, T. C. (2002) PHENIX:
building new software for automated crystallographic structure determination. Acta
Crystallogr. Sect. D Biol. Crystallogr. 58, 1948–1954
Emsley, P. and Cowtan, K. (2004) Coot: model-building tools for molecular graphics. Acta
Crystallogr. Sect. D Biol. Crystallogr. 60, 2126–2132
4
4
0
1
48 Oakley, A. J., Lo Bello, M., Ricci, G., Federici, G. and Parker, M. W. (1998) Evidence for an
induced-fit mechanism operating in pi class glutathione transferases. Biochemistry 37,
9912–9917
Laskowski, R. A., MacArthur, M. W., Moss, D. S. and Thornton, J. M. (1993) PROCHECK:
a program to check the stereochemical quality of protein structures. J. Appl. Crystallogr.
2
6, 283–291
49 Ricci, G., Lo Bello, M., Caccurri, A. M., Pastore, A., Nuccetelli, M., Parker, M. W. and
Federici, G. (1995) Site-directed mutagenesis of human glutathione transferase P1-1:
mutation of Cys-47 induces a positive cooperativity in glutathione transferase P1-1. J.
Biol. Chem. 270, 1243–1248
4
4
2
3
Tang, S. S. and Chang, G. G. (1995) Steady-state kinetics and chemical mechanism of
octopus hepatopancreatic glutathione transferase. Biochem. J. 309,
3
47–353
Axarli, I., Dhavala, P., Papageorgiou, A. C. and Labrou, N. E. (2009) Crystal structure of
Glycine max glutathione transferase in complex with glutathione: investigation of the
mechanism operating by the Tau class glutathione transferases. Biochem. J. 422,
50 Ricci, G., Caccuri, A. M., Lo Bello, M., Rosato, N., Mei, G., Nicotra, M., Chiessi, E.,
Mazzetti, A. P. and Federici, G. (1996) Structural flexibility modulates the activity of
human glutathione transferase P1-1: role of helix 2 flexibility in the catalytic mechanism.
J. Biol. Chem. 271, 16187–16192
2
47–256
4
4
Segel, I. H. (1975) Enzyme Kinetics: Behavior and Analysis of Rapid Equilibrium
and Steady-State Enzyme Systems, pp. 560–664, John Wiley and Sons, New York
51 Gouet, P., Courcelle, E., Stuart, D. I. and Metoz, F. (1999) ESPript: analysis of multiple
sequence alignments in PostScript. Bioinformatics 15, 305–308
Received 28 April 2011/28 June 2011; accepted 28 June 2011
Published as BJ Immediate Publication 28 June 2011, doi:10.1042/BJ20110753
ꢀc The Authors Journal compilation ꢀc 2011 Biochemical Society