Synthesis and characterization of a series of highly fluorogenic substrates for glutathione transferases, a general strategy

Article abstract of DOI:10.1021/ja205500y

Glutathione transferases (GSTs) are used in biotechnology applications as fusion partners for facile purification and are also overexpressed in certain tumors. Consequently, there is a need for sensitive detection of the enzymes. Here we describe a genera

Full text of DOI:10.1021/ja205500y

ARTICLE
pubs.acs.org/JACS
Synthesis and Characterization of a Series of Highly Fluorogenic
Substrates for Glutathione Transferases, a General Strategy
Jie Zhang,† Aya Shibata,‡ Mika Ito,‡,§ Satoshi Shuto,§ Yoshihiro Ito,‡ Bengt Mannervik,||,^ Hiroshi Abe,*,‡ and
Ralf Morgenstern*,†
†Institute of Environmental Medicine, Division of Biochemical Toxicology, Karolinska Institutet, SE 17177 Stockholm, Sweden
‡Nano Medical Engineering Laboratory, RIKEN Advanced Science Institute, 2-1 Hirosawa, Wako-Shi, Saitama 351-0198, Japan
§Faculty of Pharmaceutical Sciences, Hokkaido University, Kita-12, Nishi-6, Kita-ku, Sapporo 060-0812, Japan
Department of Biochemistry, Uppsala University, Biomedical Center, Box 576, SE-751 23 Uppsala, Sweden
^Department of Neurochemistry, Stockholm University, SE-10691 Stockholm, Sweden
S Supporting Information
b
ABSTRACT:
Glutathione transferases (GSTs) are used in biotechnology applications as fusion partners for facile purification and are also
overexpressed in certain tumors. Consequently, there is a need for sensitive detection of the enzymes. Here we describe a general
strategy for the synthesis and characterization of novel fluorogenic substrates for GSTs. The substrates were synthesized by introducing
an electrophilic sulfonamide linkage to fluorescent molecules containing an amino group [e.g., 2,4-dinitrobenzenesulfonamide (DNs)
derivatives of coumarin, cresyl violet, and rhodamine]. The derivatives were essentially nonfluorescent, and upon GST catalyzed
cleavage of the dinitrobenzenesulfonamide, free fluorophore is released (and 1-glutathionyl-2,4-dinitrobenzene + SO2). All the
coumarin-, cresyl violet- and rhodamine-based fluorogenic probes turned out to be good substrates for most GSTs, especially for
GSTA1À1, in terms of strong fluorescence increases (71À1200-fold), high kcat/Km values (104À107 MÀ1 sÀ1) and significant rate
enhancements (106À109-fold). The substrates weresuccessfully applied to quantitate verylow levelsofGST activity in cell extracts and
DNs-cresyl violet was also successfully applied to the imaging of microsomal MGST1 activity in living cells. The cresyl violet stained
cells retained their fluorescence after fixation, which is a very useful property. In summary, we describe a general and versatile strategy to
generate fluorogenic GST substrates, some of them providing the most sensitive assays so far described for GSTs.
’ INTRODUCTION
glutathione (GSH) is added to an electrophilic center of a
hydrophobic compound that can be of both endogenous and
exogenous origin.1 In addition to the GSH conjugation activity,
these enzymes also carry out a wide range of other functions,
including glutathione peroxidase (GPx) and isomerase activity,
biosynthesis of steroid hormones and prostaglandins, and mod-
ulation of signaling pathways.2 GSTs play an important role in
cellular protection from environmental and oxidative stress yet
are also implicated in cellular resistance to drugs.3À5 Mammalian
GSTs are divided into three main families: cytosolic, mitochon-
drial, and membrane-bound microsomal. The cytosolic family is
further divided into seven classes: alpha, mu, omega, pi, sigma,
Fluorogenic substrates are crucial to develop sensitive biotech-
nology applications (e.g., screening/reporter assays), single mole-
cule enzymology, and specific cell and tissue staining. A method
that allows the conversion of existing ultrabright fluorophores to
highly fluorogenic, essentially nonfluorescent, enzyme substrates is
therefore of strong interest. Here we describe a general strategy
that converts free amino group containing fluorophores to glu-
tathione transferase (GST) substrates. The GSTs are remarkably
indiscriminate, allowing a variety of fluorophore structures. As
glutathione transferases are overexpressed in many tumors and
used as biotechnological tools (easily purified fusion partners), this
strategy should be of general use.
Glutathione transferases (GSTs, EC 2.5.1.18) are phase II
detoxification enzymes that catalyze the basic reaction in which
Received: June 22, 2011
Published: July 25, 2011
r
2011 American Chemical Society
14109
dx.doi.org/10.1021/ja205500y J. Am. Chem. Soc. 2011, 133, 14109–14119
|

Journal of the American Chemical Society
ARTICLE
Table 1. Reported Fluorogenic Substrates as Well as the Universal Chromogenic Substrate CDNB for GSTs
kcat/Km (MÀ1 sÀ1
)
rate enhancement
substrate
product QE
GSTA1À1
GSTM1À1
GSTP1À1
MGST1
GSTA1À1
GSTM1À1
GSTP1À1
MGST1
ref
NBD-Cl
SBD-Cl
mBCl
2.7 Â 106
3.2 Â 104
7.8 Â 104
1.4 Â 105
6.8 Â 105
1.5 Â 106
1.4 Â 104
7.6 Â 104
3.9 Â 106
3.1 Â 104
4.0 Â 104
2.1 Â 106
3.6 Â 106
1.4 Â 104
3.1 Â 105
2.6 Â 105
1.3 Â 107
3.8 Â 105
1.6 Â 106
1.3 Â 106
6
7
0.191
5.8 Â 103
5.5 Â 105
1.1 Â 107
1.1 Â 107
2.2 Â 104
2.4 Â 105
7.8 Â 104
3.4 Â 106
1.7 Â 107
2.0 Â 106
9.4 Â 104
2.2 Â 104
1.1 Â 107
4.0 Â 107
2.0 Â 106
1.8 Â 104
3.1 Â 105
5.5 Â 107
1.2 Â 108
3.1 Â 106
8, 9
10
11
11
6
Cadan
8 Â 103
4.4 Â 103
4.8 Â 107
DNAF1
DNAF2
CDNB
0.101
0.257
3.3 Â 105
theta and zeta. The microsomal GSTs are designated as “membrane-
associated proteins in eicosanoid and glutathione metabolism”
(MAPEG) and include microsomal GST1 (MGST1), MGST2,
MGST3, leukotriene C4 synthase, 5-lipoxygenase activating protein,
and microsomal prostaglandin E synthase 1. MAPEG are structurally
distinct from cytosolic and mitochondrial GSTs but are functionally
similar in their ability to exhibit GSH conjugating and GPx ac-
tivity3,4,12 (one member, 5-lipoxygenase activating protein, is not
catalytically active). Among GSTs, only MGST1 is activated by
sulfhydryl reagents such as N-ethylmaleimide (NEM) [e.g., GST
activity toward 1-chloro-2,4-dinitrochlorobenzene (CDNB)13À18
and GPx activity toward cumene hydroperoxide19], and the site
of modification is a single cysteine residue, Cys49, within the
polypeptide.20
Figure 1. (A) Potential fluorogenic substrates GST-catalyzed reaction
with GSH and (B) their structures.
So far, the measurement of GST activity has often been done
with the aid of the chromogenic substrate, CDNB, which is a
good substrate for many GSTs.1 Despite its significance for
detection of GST activity, its use is often limited by low sen-
sitivity and selectivity (absorbance at 340 nm), and it should be
are very advantageous and can be applied for the synthesis of
sulfonamides, labile toward GSH/GST, which may serve as
fluorogenic substrates to release amino-containing fluorophores.
Rhodamine 110 (Rh) is highly fluorescent and resistant to
photobleaching. By modifying Rh with the strongly electron
withdrawing 2,4-dinitrobenzenesulfonyl (DNs) group, the mol-
ecule becomes essentially nonfluorescent, and the resulting fluoro-
genic bis-2,4-dinitrobenzenesulfonyl rhodamine (bis-DNs-Rh)
has been used for the detection of biological thiols as well as thiols
in HeLa cells.30 Previous studies in our laboratory showed that the
electrophilic center R to the DNs group is attacked by GSH and
that this reaction is catalyzed by both cytosolic and membrane-
bound GSTs releasing the free Rh dye.31 These results were quite
promising and suggested that any amino-containing fluorophore
could be derivatized in a similar way and serve as a potential GST
substrate. This design strategy is straightforward and versatile and
can provide significantly greater sensitivity, compared to methods
where the fluorophore produces a GS conjugate. To demonstrate
the working hypothesis, synthesis of different fluorogenic probes
for GSTs was carried out, including that of 7-dinitrobenzenesulfo-
namino-4-methyl-3-coumarinylacetic acid (DNs-Coum), DNs-N-
acetylrhodamine (DNs-AcRh), and DNs-cresyl violet (DNs-CV).
Compared to bis-DNs-Rh, DNs-AcRh contains only a single
hydrolysis-sensitive sulfonamide moiety, thus avoiding the com-
plicated biphasic kinetics of disubstituted Rh-based substrates. In
order to examine the influence of chemical reactivity on enzyme
activity and background rates, para-substituted 4-mononitroben-
zenesofonyl (MNs), 4-toluenesulfonyl (Ts), and pentafluoroben-
zenesulfonyl (PBs) derivatives were synthesized as well. The
potential substrates and their reaction with GSH are depicted in
Figure 1. Fluorescence emissions of the products shift gradually to
6
stressed that certain forms of the enzyme (e.g., MGST1 and
GST theta21) display low activity with CDNB and consequently
may be overlooked in samples analyzed exclusively by use of
CDNB. Fluorometric detection utilizing several fluorogenic
molecules has been developed for highly sensitive detection of
GSH/GST activity. Representative fluorogenic reagents are 4-chloro-
7-nitrobenzofurazan (NBD-Cl),6 ammonium 4-chloro-7-sulfo-
benzofurazan (SBD-Cl),7 monochlorobimane (mBCl),8,9,22,23
6-chloroacetyl-2-dimethylaminonaphthalene (Cadan),10 and di-
nitrobenzoylaminofluorescein (DNAF)11 (Table 1). These re-
agents are essentially nonfluorescent themselves and can under-
go GST-catalyzed reactions to produce fluorescent GSH adducts,
thus enabling highly sensitive GSH/GST activity detection.
However, these probes display either high nonenzymatic reac-
tion rates, low efficiency, narrow enzyme selectivity, low fluor-
escence quantum efficiency (QE), and poor stability of the GS
conjugate products or have been tested only with cytosolic GSTs.
Although the fluorescence off/on mechanism of the above-
mentioned fluorogenic substrates differ,11,24À27 all of them show
substantial fluorescence enhancements upon GSH conjugation.
In addition to GSH conjugation activity, GSTs have been
reported to exhibit sulfonamidase activity and catalyze the GSH-
mediated cleavage of sulfonamide bonds to form a GS conjugate,
the corresponding amine and sulfur dioxide.28,29 StructureÀ
activity studies with a variety of sulfonamides indicate that an
electrophilic center R to the sulfonyl group is required for
cleavage. On the other hand, the stability of sulfonamides is less
dependent on the nature of the amine moiety.29 These features
14110
dx.doi.org/10.1021/ja205500y |J. Am. Chem. Soc. 2011, 133, 14109–14119

Journal of the American Chemical Society
ARTICLE
Table 2. Quantum Efficiencies of the Compoundsa
a specific reaction, and the other is to ensure that the target
enzyme recognizes the compound as a specific substrate.11 To
fulfill these two prerequisites, we first measured the fluorescence
quantum efficiencies of the fluorogenic probes and their fluor-
ophore metabolites. As shown in Table 2, the fluorogenic probes,
DNs-Coum, DNs-AcRh, MNs-AcRh, PBs-AcRh, DNs-CV, and
MNs-CV, show very low fluorescence quantum efficiencies, in
contrast to the high efficiencies of Coum, AcRh, and CV. This
result is ideal in terms of fluorescence activation. Next, we needed
to find out whether these sulfonamides can be readily cleaved by
the GSÀ thiolate and whether these reactions are specifically cat-
alyzed by GSTs. The probes were incubated in solution with or
without GSH/GST, and the resulting absorption and fluores-
cence spectra are shown in Figures 2 and 3, respectively. The
DNs-Coum and Coum showed a maximum absorption at 342
and 341 nm, respectively, in 10 mM PBS (pH 7.4) (Figure 2A). A
decrease in the absorption of DNs-AcRh at 490 nm was observed
compared with AcRh due to the formation of lactone ring
(Figure 2B). The DNs-CV and CV showed a maximum absorp-
tion at 607 and 594 nm, where lower absorption was observed in
DNs-CV (Figure 2C). Without GSH/GST, no significant fluor-
escence was observed from the fluorogenic probes (Figure 3).
After the addition of GSH to the solution of fluorogenic probes, a
low emission appeared around 450, 520, and 620 nm, respec-
tively. On the other hand, with GSH/GST, a strong emission
appeared and the emission was enhanced 62À290-fold. The
results demonstrated that all probes are potentially good fluoro-
genic substrates for GSTs. Therefore, in the next round of
experiments, specific activities of both cytosolic and membrane-
bound GSTs toward these probes, as well as steady-state kinetic
characteristics, were carefully examined.
fluorogenic
fluorescence
substrate
QE
product
QE
activation fold
DNs-Coum
DNs-AcRh
MNs-AcRh
PBs-AcRh
DNs-CV
0.009
Coum
AcRh
0.641
0.387
71
110
66
0.0036
0.0059
0.0256
0.0004
0.0002
15
CV
0.23
580
1200
MNs-CV
a DNs-Coum and Coum measurements were done in sodium phosphate
buffer (100 mM, pH 10). Compounds were excited at 375 nm. Quantum
yields are determined by using 4-methylumbelliferone (0.63, 0.1 M
phosphate buffer (pH 10) as a standard. DNs-AcRh, MNs-AcRh, PBs-
AcRh, and AcRh measurements were done in sodium phosphate buffer
(100 mM, pH 7.4). Compounds were excited at 490 nm. Quantum
yields are determined by using fluorescein (0.85, 0.1 M NaOH) as
standard. DNs-CV, MNs-CV, and CV measurements were done in
sodium phosphate buffer (100 mM, pH 7.4). Compounds were excited
at 540 nm. Quantum yields are determined by using CV (0.54, MeOH)
as standard.
longer wavelengths, from 450 to 620 nm. Therefore, this series of
probes can offer multiple color detection in living cells and other
biological matrices, giving blue, green and red color under
fluorescence microscopy. An important application of our fluoro-
genic GST substrates could be in imaging of cancer cells that
overexpress GSTs. However, this implies broad isoenzyme sub-
strate selectivity.
Here we present, for the first time, a general strategy to produce
fluorogenic substrates for both cytosolic and membrane-bound
GSTs. First of all, specific activities were determined for a panel of
recombinant cytosolic GSTs and purified MGST1. High activities
were obtained with DNs-Coum, DNs-AcRh and DNs-CV. Steady-
state kinetic analysis of these probes was then performed with four
representative enzymes, GSTA1À1, GSTM1À1, GSTP1À1, and
MGST1. Furthermore, DNs-Coum, DNs-AcRh, and DNs-CV
were highly sensitive and useful to quantitate GST activity in cell
extracts, and DNs-CV was successfully applied to the imaging of
MGST1 activity in living cells.
GST Activity toward Fluorogenic Probes. Specific activities
were determined for cytosolic GSTs, GSTA1À1,GSTA2À2,GSTA3À3
,
GSTA4À4, GSTM1À1, GSTM2À2, GSTP1À1, and GSTT1À1, as
well as membrane-bound MGST1, using DNs-Coum, DNs-
AcRh, MNs-AcRh, PBs-AcRh, DNs-CV, MNs-CV, and Ts-CV
(Table 3). In general, all of the probes, except MNs-CV and Ts-
CV, are good substrates for all GSTs, except for GSTT1À1. The
highest activities were found with DNs-AcRh, DNs-Coum, and
DNs-CV, and the activities for GSTA1À1 with these probes were
more than 2-fold higher than for other GSTs. With DNs-Coum
and DNs-AcRh, MGST1 showed much lower activity compared
with cytosolic GSTs (except GSTT1À1). In contrast, with DNs-
CV, the activities of MGST1 were higher or comparable with
those of cytosolic GSTs (except GSTA1À1).
’ RESULTS
Synthesis of the Fluorogenic Probes. The synthesis of DNs-
Coum was described in our previous paper.32 DNs-AcRh, MNs-
AcRh, and PBs-AcRh were synthesized according to Scheme
S1 (Supporting Information). The amino group of mono-
Boc-rhodamine 11033 was protected by an acetyl group. And
then, the Boc group was removed by treatment with 4 M HCl/
dioxane solution to give the AcRh. The desired products,
benzene sulfonamide-protected rhodamine derivatives (DNs-
AcRh, MNs-AcRh and PBs-AcRh), were obtained by treatment
of AcRh with DNs-Cl, MNs-Cl, or PBs-Cl in pyridine/CH2Cl2
for 5À16 h in 37%, 42%, or 47% yield, respectively. The synthesis
of DNs-CV, MNs-CV, and Ts-CV was carried out according to
Scheme S2 (Supporting Information). The treatment of CV with
DNs-Cl, MNs-Cl, or Ts-Cl under various basic conditions
successfully gave the desired products, DNs-CV (30%), MNs-
CV (69%), or Ts-CV (47%).
When the DNs group was replaced by a MNs group, a dra-
matic decrease (more than 1000-fold) was observed in activity for all
GSTs, presumably due to the low reactivity. In this case, the activities
for GSTM1À1 and GSTM2À2 with MNs-AcRh and MNs-CV were
higher than for other GSTs. The replacement of DNs group with a
Ts group resulted in loss of activity for all GSTs. When the DNs
group was replaced by PBs group, the activities were more than 100-
fold lower for all GSTs. The activity for GSTP1À1 with PBs-AcRh
was higher than for other GSTs. Thus, manipulating substrate
reactivity alters enzyme selectivity.
Steady-State Kinetics with Selected Fluorogenic Probes.
In order to gain deeper insight into the rate behavior toward
different fluorogenic probes, a steady-state kinetic analysis of
DNs-Coum, DNs-AcRh, MNs-AcRh, PBs-AcRh, and DNs-CV
was performed with three representative cytosolic GSTs,
GSTA1À1, GSTM1À1, and GSTP1À1, and membrane-bound
MGST1. To determine whether a highly efficient fluorogenic
Characterization of the Fluorogenic Probes. Generally,
development of fluorogenic substrates for enzyme assays faces
two main difficulties. One is to achieve fluorescence activation by
14111
dx.doi.org/10.1021/ja205500y |J. Am. Chem. Soc. 2011, 133, 14109–14119

Journal of the American Chemical Society
ARTICLE
Figure 2. Absorption spectra of 20 μM fluorogenic probes and 20 μM products in 10 mM PBS (pH7.4). (A) DNs-Coum and Coum, (B) DNs-AcRh
and AcRh, and (C) DNs-CV and CV. Absorption spectra were observed by UV/vis spectrophotometer (V-550, JASCO).
Figure 3. Fluorescence measurement of the fluorogenic probes in the absence or presence of GSH/GST. Reactions were performed in 10 mM PBS
(pH7.4) with 1 μM probe, 10 mM GSH, and 2 μg/mL GST for 8 min at 37 ꢀC. Fluorescence spectra were observed by fluorescence spectrophotometry
(FP-6500, JASCO). The fluorescence with excitation of (A, DNs-Coum) 345 nm was obtained a scan range of 400À550 nm, (B, DNs-AcRh) 490 nm
was obtained a scan range of 500À600 nm, and (C, DNs-CV) 540 nm was obtained a scan range of 500À750 nm.
Table 3. Specific Activity of GSTs with Different Fluorogenic Substratesa
specific activity (nmol minÀ1 mgÀ1
)
substrate
GSTA1À1
GSTA2À2
GSTA3À3
GSTA4À4
GSTM1À1
GSTM2À2
GSTP1À1
GSTT1À1
ND
MGST1
130 ( 2
DNs-Coum 1700 ( 160 230 ( 30
660 ( 210 370 ( 60
930 ( 100 520 ( 100
810 ( 110
330 ( 40
990 ( 90
0.31 ( 0.06
0.45 ( 0.06
410 ( 50
690 ( 60
DNs-AcRh
6700 ( 200 490 ( 30
2700 ( 200
2300 ( 560 0.28 ( 0.04
0.22 ( 0.01 ND
420 ( 60
MNs-AcRh 0.23 ( 0.4
0.18 ( 0.01 0.24 ( 0.1 0.13 ( 0.04 0.31 ( 0.02
0.15 ( 0.01 3.1 ( 0.4 0.47 ( 0.04 0.94 ( 0.09
0.0011 ( 0.0001
PBs-AcRh
DNs-CV
MNs-CV
Ts-CV
6.7 ( 0.3
1500 ( 100 150 ( 20
9.0 ( 0.6
460 ( 50
0.0016 ( 0.0001 0.50 ( 0.06
480 ( 90
ND
770 ( 210
ND
550 ( 100
0.062 ( 0.007
560 ( 80
ND
ND
ND
0.085 ( 0.013 0.038 ( 0.021 ND
ND
ND
ND
ND
ND
ND ND ND
ND
ND
a Specific activities were obtained in 0.1 M potassium phosphate buffer pH 6.5 (with 0.1% Triton for MGST1), 5 mM GSH, and DNs-Coum, DNs-AcRh,
MNs-AcRh, PBs-AcRh, DNs-CV, MNs-CV, or Ts-CV as second substrate. The concentration of DNs-Coum used was 40 μM, the concentration of
DNs-AcRh, MNs-AcRh, and PBs-AcRh used was 25 μM, and the concentration of DNs-CV, MNs-CV, and Ts-CV used was 2.5 μM. The final DMSO
concentration in all assays was 1.0% (v/v). ND, not detectable. Values are means ( standard deviations (n = 3).
substrate has been found, the most useful catalytic constant is
kcat/Km, sometimes called the specificity constant, representing
the apparent second-order rate constant for the association of
substrate and free enzyme to form product and free enzyme.
There is an upper limit to kcat/Km, imposed by diffusion, i.e., the
rate at which enzyme and substrate collide in aqueous solution.
The diffusion-controlled limit is 108À109 MÀ1
s
À1. Enzymes
having a kca3t4/,3K5m near this range are said to have achieved catalytic
perfection.
k /Km values of GSTs with different fluorogenic
cat
probes are shown in Figure 4. GSTA1À1 is the most efficient
enzyme for all the DNs-derivatized fluorogenic substrates
(DNs-Coum, DNs-CV, and DNs-AcRh), with kcat/Km values
14112
dx.doi.org/10.1021/ja205500y |J. Am. Chem. Soc. 2011, 133, 14109–14119

Journal of the American Chemical Society
ARTICLE
Figure 4. Catalytic efficiencies of GSTs with different fluorogenic substrates. The catalytic efficiencies (kcat/Km) of GSTA1À1, GSTM1À1, GSTP1À1
,
and MGST1 were determined with (A) DNs-Coum, (B) DNs-CV, (C) DNs-AcRh, (D) MNs-AcRh, and (E) PBs-AcRh. Values are means ( standard
errors of the mean (n g 6).
Table 4. Steady-State Kinetic Parameters of GST with Alternative Fluorogenic Substratesa
substrate
parameters
GSTA1À1
GSTM1À1
GSTP1À1
MGST1
47 ( 12
DNs-Coum
Km (μM)
kcat (sÀ1
14 ( 2
12 ( 3
17 ( 5
)
1.0 ( 0.1
0.47 ( 0.05
0.43 ( 0.06
0.24 ( 0.03
(5.1 ( 1.5) Â 103
4.5 ( 0.7
1.5 ( 0.1
(3.3 ( 0.6) Â 105
38 ( 8
kcat/Km (MÀ1 sÀ1
)
)
)
)
)
(7.1 ( 1.2) Â 104
0.18 ( 0.03
(3.9 ( 1.1) Â 104
0.45 ( 0.15
(2.5 ( 0.8) Â 104
0.49 ( 0.13
DNs-CV
Km (μM)
kcat (sÀ1
)
0.60 ( 0.03
0.23 ( 0.02
0.25 ( 0.02
kcat/Km (MÀ1 sÀ1
(3.3 ( 0.8) Â 106
0.29 ( 0.02
(5.1 ( 2.0) Â 105
0.72 ( 0.12
(5.1 ( 1.4) Â 105
0.87 ( 0.18
DNs-AcRh
MNs-AcRh
PBs-AcRh
Km (μM)
kcat (sÀ1
)
3.1 ( 0.1
1.2 ( 0.1
1.1 ( 0.1
1.0 ( 0.1
(2.6 ( 0.6) Â 104
>25b
NDc
0.056 ( 0.009d
>200f
kcat/Km (MÀ1 sÀ1
(1.1 ( 0.1) Â 107
1.5 ( 0.2
(1.7 ( 0.3) Â 106
0.87 ( 0.11
(1.3 ( 0.3) Â 106
0.99 ( 0.19
Km (μM)
kcat (sÀ1
)
(1.0 ( 0.4) Â 10À4
67 ( 28
(1.4 ( 0.4) Â 10À4
160 ( 50
(1.1 ( 0.1) Â 10À4
110 ( 20
kcat/Km (MÀ1 sÀ1
Km (μM)
17 ( 5
6.8 ( 1.5
41 ( 8e
kcat (sÀ1
kcat/Km (MÀ1 sÀ1
)
(4.4 ( 0.6) Â 10À3
260 ( 80
(4.7 ( 0.4) Â 10À4
69 ( 16
(9.1 ( 1.0) Â 10À3
220 ( 50
NDc
16 ( 1d
a The cytosolic GST (GSTA1À1, GSTM1À1, or GSTP1À1) was assayed in 0.1 M potassium phosphate buffer pH 6.5, at a constant GSH concentration of
5 mM and varying concentrations of DNs-Coum (1.25À40 μM), DNs-CV (0.1À2.5 μM), DNs-AcRh (0.049À6.25 μM), MNs-AcRh (0.39À25 μM),
or PBs-AcRh (0.78À50 μM), whereas MGST1 was assayed in 0.1 M potassium phosphate buffer pH 6.5 with 0.1% Triton, at a constant GSH
concentration of 5 mM and varying concentrations of DNs-Coum (2.5À80 μM), DNs-CV (0.63À20 μM), DNs-AcRh (1.56À100 μM), MNs-AcRh
(12.5 and 25 μM), or PBs-AcRh (25À200 μM). The final concentration of the solvent DMSO was <1% unless otherwise specified. Error limits are given
by Graphpad Prism 5 after fitting by nonlinear regression; in general, at least 18 observations were included in the data sets. b The hydrophobic nature of
MNs-AcRh makes it difficult to obtain saturation kinetics. The highest concentration used was 25 μM. c ND, not determined. d The apparent kcat/Km
value was obtained using the MichaelisÀMenten relationship at low substrate concentration: v/[E][S] = kcat/Km, when [S] < Km. Values are means (
standard errors of the mean (n g 6). e The enhanced inner-filter effect at high concentrations of PBs-AcRh makes it difficult to obtain saturation kinetics.
The highest concentration used was 50 μM. f The enhanced DMSO concentration (the concentration of the PBs-AcRh stock solution was 10 mM in
DMSO) at high concentrations of PBs-AcRh makes it difficult to obtain saturation kinetics. The highest concentration used was 200 μM (2% DMSO as
cosolvent).
approaching 105, 107, and 108, respectively. For DNs-Coum and
DNs-AcRh, MGST1 showed much lower efficiencies compared
with all cytosolic GSTs. Whereas for DNs-CV, the efficiencies of
MGST1 were comparable with those of GSTM1À1 and
GSTP1À1, with kcat/Km values approaching 106.
Table 5. Second-Order Rate Constant for the Nonenzymatic
Reactiona
knoncat (MÀ1 sÀ1
)
substrate
without 0.1% Triton X-100
with 0.1% Triton X-100
The effect of the substituent in the aromatic ring, R to the
sulfonyl group, on the catalytic efficiency was investigated with
the Rh-based fluorogenic substrates (DNs-AcRh, MNs-AcRh,
and PBs-AcRh). The catalytic efficiencies with DNs-AcRh as a
substrate were generally 4 orders of magnitude greater than those
with MNs-AcRh and 3 orders of magnitude greater than those
with PBs-AcRh. For DNs-AcRh and PBs-AcRh, GSTA1À1
showed higher efficiencies compared with other GSTs, whereas
for MNs-AcRh, the kcat/Km value for GSTM1À1 was higher than
for other GSTs (Figure 4CÀE).
DNs-Coum
DNs-CV
0.017 ( 0.001
0.016 ( 0.001
(7.7 ( 0.8) Â 10À4
(1.4 ( 0.1) Â 10À3
NDb
(1.4 ( 0.3) Â 10À3
0.16 ( 0.01
DNs-AcRh
MNs-AcRh
PBs-AcRh
(4.9 ( 0.3) Â 10À6
(1.5 ( 0.1) Â 10À5
(1.4 ( 0.1) Â 10À5
a The second-order rate constant for the nonenzymatic reaction (knoncat
)
was obtained in 0.1 M potassium phosphate buffer pH 6.5 (without or
with 0.1% Triton), at a constant GSH concentration of 5 mM and
varying concentrations of DNs-Coum, DNs-CV, DNs-AcRh, MNs-
AcRh, or PBs-AcRh, same as mentioned in Table 4. b ND, not detectable.
Values are means ( standard errors of the mean (n g 6).
kcat and Km values are summarized in Table 4. kcat values are in
the 1 sÀ1 range, and the high catalytic efficiencies are depending
14113
dx.doi.org/10.1021/ja205500y |J. Am. Chem. Soc. 2011, 133, 14109–14119

Journal of the American Chemical Society
ARTICLE
Table 6. Rate Enhancements Produced by GSTsa
GSTA1À1
GSTM1À1
GSTP1À1
MGST1
DNs-Coum
DNs-CV
(4.1 ( 0.7) Â 106
(2.4 ( 1.1) Â 109
(6.8 ( 0.6) Â 107
(1.4 ( 0.6) Â 107
(1.7 ( 0.6) Â 107
(2.3 ( 0.6) Â 106
(3.6 ( 2.1) Â 108
(1.1 ( 0.2) Â 107
(3.3 ( 1.1) Â 107
(4.5 ( 1.2) Â 106
(1.5 ( 0.5) Â 106
(3.6 ( 2.1) Â 108
(8.0 ( 1.8) Â 106
(2.3 ( 0.5) Â 107
(1.4 ( 0.4) Â 107
(3.2 ( 0.9) Â 105
(4.3 ( 1.2) Â 108
(1.9 ( 0.4) Â 107
ND
DNs-AcRh
MNs-AcRh
PBs-AcRh
(1.1 ( 0.2) Â 106
a For assay conditions, see Tables 4 and 5. Values are means ( standard errors of the mean (n g 6).
on low Km values. The low Km values for the DNs derivatives are
most likely due to the high reactivity of the compounds, as the
less reactive analogues MNs and PBs display considerably higher
Km values. This rate behavior precludes the simplest MichaelisÀ
Menten mechanism, where Km is a combination of substrate
release/binding and chemical conversion (kÀ1 + k2)/k1, consis-
tent with earlier kinetic characterization of these enzymes.36
In conclusion, so far, all the DNs-derivatized fluorogenic
substrates (DNs-AcRh, DNs-Coum, and DNs-CV) are good
substrates for most GSTs, except GSTT1À1. The broad enzyme
specificity, high catalytic efficiencies, and the highly sensitive
fluorescence-based assay enable these fluorogenic substrates to
be attractive alternatives to the universal chromogenic substrate
CDNB for GST activity measurement. The nature of the
electrophilic group in the sulfonyl moiety markedly affects the
enzyme reactivity and selectivity. A strong electrophilic center
R to the sulfonyl group is required for high GSH/GST activity.
Replacement of the ortho- and para-dinitro (DNs) group with
para-mononitro (MNs) or PBs group results in a dramatic
decrease in both activity and catalytic efficiency for all GSTs,
and the replacement with a para-methyl (Ts) group resulted in
loss of activity. The DNs-derivatized fluorogenic substrates dis-
play selectivity for GSTA1À1, whereas the MNs-derivatized
fluorogenic substrates display selectivity for GSTM1À1 and
GSTM2À2, and the PBs-derivatized fluorogenic substrates dis-
play selectivity for GSTA1À1 and GSTP1À1. The nature of the
amine moiety (fluorophore) influences the enzyme reactivity and
selectivity as well, though the differences are not remarkable. For
cytosolic GSTs, the activities and catalytic efficiencies with DNs-
Rh were mostly higher than those with DNs-Coum and DNs-
CV, whereas membrane-bound MGST1 showed higher activity
and catalytic efficiency with DNs-CV.
In order to assess the rate enhancement, the second-order rate
constant of the nonenzymatic reaction, knoncat, was determined
to compare with the apparent second-order rate constant for the
enzyme-catalyzed reaction, kcat/Km. Since the enzymatic reaction
was performed both with and without 0.1% Triton X-100, the
same assay conditions were applied to determine knoncat, and the
results are summarized in Table 5. The chemical properties of the
different derivatives are clearly reflected in knoncat in a logical
fashion, as the reactivity of the aromatic substituent is decreased
(DNs-AcRh to PBs-AcRh and MNs-AcRh). Interestingly, the
fluorogenic moiety has a 2 orders of magnitude influence on the
background rate, with DNs-CV showing a very low value. In the
presence of detergent, the more hydrophobic fluorogenic mol-
ecules display much lower rates, as they are probably shielded from
reaction with the hydrophilic GSH by partitioning into the
detergent micelles. The observed trend in nonenzymatic reactivity
is reflected also in the kcat/Km values (Figure 4), resulting in
quite similar rate enhancement for DNs-AcRh, MNs-AcRh, and
Figure 5. Specific activity of GST in cell lysate with fluorogenic substrate:
(A) 40μMDNs-Coum, (B) 25μMDNs-AcRh, and(C) 2.5μMDNs-CV.
Values are means ( standard errors of the mean (n g 3).
PBs-AcRh (Table 6). Overall the cresyl violet derivative displays
superior rate enhancement with all enzymes tested.
GST Activity in MCF7 Cell Lysates. It is often desirable to be
able to detect very low levels of enzymatic activities in cell
extracts, tissues, or clinical specimens. Therefore, we tested
whether DNs-Coum, DNs-AcRh, and DNs-CV could measure
the GST activity in MCF7 cell lysates and whether the differences
of GST activity in MGST1-expressing and vector control MCF7
cells could be detected by these fluorogenic substrates. Human
MCF7 breast cancer cells have very low GST activity and have
often been used for transfection experiments with GSTs.37À40
For all fluorogenic substrates tested, extracts from MGST1-
expressing MCF7 cells were more efficient at catalyzing the
reaction compared with the vector control cells (under normal
conditions and after NEM preincubation to activate MGST1).
The most significant differences in specific activity between
MGST1-expressing cells and vector control cells were found
with DNs-CV, consistent with DNs-CV being the best substrate
for MGST1 (Figure 5). The activity could be increased in
MGST1-expressing cell extracts by NEM preincubation as ex-
pected, whereas the activity in the vector control cells was
decreased by NEM preincubation. The low (and NEM sensitive)
activity in vector control cells most likely stems from the low
14114
dx.doi.org/10.1021/ja205500y |J. Am. Chem. Soc. 2011, 133, 14109–14119

Journal of the American Chemical Society
ARTICLE
Figure 6. Fluorescence microscopic imaging of live MCF7 cells. The MGST1-expressing cells and the control cells were incubated with 2.5 μM DNs-CV
for 30 min at 37 ꢀC without or with NEM pretreatment.
Figure 7. Fluorescence microscopic imaging of MCF7 cells. The MGST1-expressing cells and the control cells were incubated with 2.5 μM DNs-CV for
30 min at 37 ꢀC without or with NEM pretreatment. Cells were then fixed for 15 min at room temperature with 2% formaldehyde and DNA was stained
with 1 μg/mL Hoechst 33342 for 10 min at room temperature to localize nuclei.
endogenous levels of cytosolic Mu class GST, since GSTA and
GSTP are essentially not expressed in MCF7 cells.37,39,40
area adjacent to the nucleus (Figure 7). The localization of the
released dye is similar to when CV is added directly to cells and
thus a property of the dye itself. After fixing the cells, the dye was
still detected, which should be a quite useful property for
conveniently detecting GST-expressing cells.
Fluorescence Microscopic Imaging of MCF7 Cells. Since
DNs-Coum, DNs-AcRh, and DNs-CV could detect differences in
GST activity in cell lysates prepared from MGST1-expressing cells
and vector control cells, we further investigated whether the
differences could be imaged by fluorescence microscopy in living
cells. DNs-Coum and DNs-AcRh did not produce a marked
increase of fluorescence in MGST1 overexpressing or control cells.
This is probably due to the higher water solubility or that their
metabolites (Coum and AcRh) are easily pumped out of the cells.
Incubation of cells with DNs-CV, however, resulted in a time-
dependent increase in fluorescence within 30 min, as shown in
Figure 6. A stronger fluorescence signal was seen in MGST1-
expressing cells compared with the vector control cells. Preincuba-
tion with the GSH-depleting agent, NEM, abolished the fluores-
cence, which confirmed that the cellular fluorescence signal can be
attributed to the GSH-mediated cleavage of sulfonamide to form
the corresponding fluorescent amine, cresyl violet. Cells were then
stained with Hoechst 33342 to localize nuclei, and it appeared that
the fluorescence signal from the released CV was localized to an
’ DISCUSSION
A General Strategy to Produce a Novel Series of Fluoro-
genic Substrates for GSTs. Assays wherein a fluorogenic
reagent is reacted with a thiol group to form a fluorescent adduct
have been known for some time. There are now many types
of thiol-reactive dyes reported for GSH/GST activity mea-
surement, including benzofurazan derivatives (NBD-Cl and
SBD-Cl), bimane derivatives (mBCl), naphthalene derivatives
(Cadan), DNAFs, and DNAT-Me. These compounds are essen-
tially nonfluorescent, through either intramolecular charge trans-
fer (ICT) blocking (NBD-Cl, SBD-Cl, mBCl, and Cadan) or
photon-induced electron transfer (PET) (DNAFs and DNAT-
Me). The substitution of the electron-withdrawing Cl atom
(NBD-Cl, SBD-Cl, mBCl, and Cadan) or nitro group (DNAFs)
by the GSÀ thiolate introduces an electron-donating group that
14115
dx.doi.org/10.1021/ja205500y |J. Am. Chem. Soc. 2011, 133, 14109–14119

Journal of the American Chemical Society
ARTICLE
increase upon GST-catalyzed reaction with GSH (Table 2).
Substituents at the C-7 position are well-known to affect the
fluorescence properties of coumarins. DNs-protection quenches
the emission of 7-aminocoumarins, which are highly fluorescent
when deprotonated. DNs-AcRh decreased absorption at 490 nm,
indicating a closed lactone form. On the other hand, AcRh in its
open lactone form emits a strong fluorescence signal. In DNs-CV,
the reasons for the loss of fluorescence can be attributed to the
conversion of an electron-donating amine group to an electroneg-
ative sulfonamide group.42 The most important feature of our
substrates is that they are the most efficient fluorogenic GST
substrates so far produced and that they generate a free
(unconjugated) fluorophore. In order to facilitate the comparison
of different fluorogenic substrate/enzyme combinations, we here
suggest a bar graph where the most important features can be
viewed in a simple fashion (Figure 8). By depicting the combina-
tion of QE (being 0À1), log kcat/Km divided by 10 (resulting in a
scale that at best approaches 1), and rate enhancement (RE) (in
this caseusinglog RE divided by 10, againresulting in a scale thatat
best approaches 1 for the substrates/enzymes used here) in a
single bar with three sections it is easy to compare quantum
efficiency, catalytic efficiency, and rate enhancement (over the
noncatalyzed reaction). The combined bar gives an indication of
the best substrate and how far from perfection the substrate is (i.e.,
physical limits in terms of QE and kcat/Km). For example, DNAFs
were the most recently reported fluorogenic substrates for GSTs
with catalytic efficiencies reaching 107 for GSTP1À1 and rate
enhancements reaching 107À108.11 By comparing DNAFs to our
substrates using the bar graph approach, we can easily see that our
substrates are somewhat closer to the goal of perfection, especially
for GSTA1À1 (in theory approaching a combined score of 3).
Applications. Extremely bright fluorescent molecules can be
used to study single enzyme molecules and thus offer insight into
enzyme behavior. Furthermore, GSTs play key roles in inflam-
mation, pathophysiology, and tumor drug resistance. Therefore,
practical fluorogenic substrates for GSTs should be extremely
useful and with potential applications in many areas. First of all, it
has become evident that the collective behavior of enzymes that
are studied in traditional setups for steady-state and pre-steady-
state kinetics cannot be taken for granted as reflecting enzyme
behavior at the single molecule level.43 No information on single
enzyme molecular behavior has been gathered for GSTs; in fact,
only very few enzymes have been amenable to this type of
analysis. In order to achieve this, extremely stable and efficient
fluorogenic substrates are needed. Our discovery that 7-amino-4-
methyl-3-coumarinylacetic acid, cresyl violet, and rhodamine can
be efficiently released by GSTs opens the door to study GSTs at
single molecule level. Second, successful quantitation of very low
levels of GST activity in cell extracts and imaging of GST activity
in living cells could be very useful for identifying tumor cells that
overexpress GSTs. Here the release of a fluorescent molecule
(not a glutathione conjugate that is usually pumped out of cells)
makes it possible to achieve fluorophore cell retention (as
demonstrated for CV). Additional applications of this novel
series of fluorogenic substrates include sensitive detection of
hybrid proteins generated by GST fusion and their binding
partners for use as a signal amplification system in biodetection.
Figure 8. Comparison of different fluorogenic substrate/enzyme com-
binations. Quantum efficiency (QE), kcat/Km, and rate enhancements
(RE) are from Tables 1, 2, 4, and 6. A perfect substrate (see the text)
could in principle reach a combined score g2.5.
results in a dramatic increase in fluorescence. In general, for the
compounds so far described, the fluorescence off/on switching
happens upon glutathionylation and the fluorophore ends-up in a
GS-conjugate. Many of these compounds still have drawbacks for
use as a fluorogenic reagent in terms of fluorescence activation,
catalytic efficiency, and rate enhancement, and therefore, there
exists a need for improved fluorogenic substrates. Desirable
properties are high and selective reactivity to GSH/GST, low
background reaction, and highly stable fluorescent products,
thereby enabling sensitive detection of very low levels of GST
activity in biological samples. Here we report a general strategy to
derivatize any amino group containing fluorophore and convert
it to a GST substrate. This affords great versatility in terms of
choosing fluorophores. So far, the different fluorescent substit-
uents tested did not result in strong enzyme selectivity. Although
it is a useful property to have broad spectrum substrates, highly
selective substrates are also desirable. By altering the reactivity of
the sulfonamide, very efficient substrates were created (kcat/Km
up to 107 MÀ1 À1, Figure 4). In addition, reactivity was also
s
coupled to changes in enzyme selectivity. Therefore, to fine-tune
reactivity36 and alter hydrophobic properties41 opens possibili-
ties to engineer substrate isoenzyme selectivity utilizing the
electrophilic portion of the fluorogenic substrate. The rate
enhancements obtained (105À109, Table 6) are equivalent to,
or 2 orders of magnitude greater than, those obtained for the
universal substrate CDNB (See Table 1; kcat/Km, 104À105
MÀ1
s
À1; rate enhancement, 106À107 6) . If one compares
specific activities toward our best fluorogenic substrates and
CDNB, the latter is still around 10-fold higher with several
GSTs,1 indicating that there is room for improvement; however,
as the Km for CDNB is generally high,6 considerably higher
substrate concentrations have to be used.
’ CONCLUSIONS
Comparison to Other Fluorogenic Substrates, a General
Approach. The DNs group quenches the fluorescence efficiently
through different mechanisms and provides a significant fluorescent
A general strategy to synthesize strongly fluorogenic and
efficient substrates applicable for most glutathione transferases
14116
dx.doi.org/10.1021/ja205500y |J. Am. Chem. Soc. 2011, 133, 14109–14119

Journal of the American Chemical Society
ARTICLE
has been developed. Amino group containing fluorophores are
derivatized with activated benzene sulfonamides, effectively
preventing fluorescence. The enzyme reaction releases the ori-
ginal fluorophore, a property that allows great versatility for
enzymology, biotechnology applications, and cellular imaging.
established for quantification. Coum and CV were obtained from Sigma-
Aldrich (St. Louis, MO).
Determination of Steady-State Kinetic Constants. Kinetic
parameters Km, kcat, and kcat/Km for GSTA1À1, GSTM1À1, GSTP1À1
,
and MGST1 were determined with DNs-Coum, DNs-AcRh, MNs-
AcRh, PBs-AcRh, and DNs-CV as substrates. The same assay conditions
as those described above were used. Cytosolic GSTs (GSTA1À1
,
GSTM1À1, or GSTP1À1) were assayed in 0.1 M potassium phosphate
buffer pH 6.5, at a constant GSH concentration of 5 mM and varying
concentrations of DNs-Coum (1.25À40 μM), DNs-AcRh (0.049À
6.25 μM), MNs-AcRh (0.39À25 μM), PBs-AcRh (0.78À50 μM), or
DNs-CV (0.1À2.5 μM), whereas MGST1 was assayed in 0.1 M po-
tassium phosphate buffer pH 6.5 containing 0.1% Triton, at a constant
GSH concentration of 5 mM and varying concentrations of DNs-Coum
(2.5À80 μM), DNs-AcRh (1.56À100 μM), MNs-AcRh (12.5À
25 μM), PBs-AcRh (25À200 μM), or DNs-CV (0.63À20 μM). The
values of the steady-state kinetic constants were determined by fitting the
MichaelisÀMenten equation to the data by nonlinear regression analysis
using GraphPad Prism 5 (GraphPad Software, Inc., San Diego, CA). In
cases where saturation was not reached, kcat/Km was determined by fitting
the equation v = (kcat/Km)[E][S] to the low substrate concentration data,
where Km > [S]. The kcat value was calculated per subunit for cytosolic
GSTs (25.5 kDa) and per trimer for MGST1 (51 kDa for MGST1), since
the enzyme displays one-third of the sites reactivity.15
’ EXPERIMENTAL METHODS
Enzyme Preparation. Human GSTA1À1, GSTA2À2, GSTA3À3
,
and GSTA4À4 were heterologously expressed from the pET-21a (+)
vector in Escherichia coli BL-21 DE3 cells (Novagen, Madison, WI) and
purified from bacterial lysate using a HiTrap SP cation-exchange column
(Amersham Biosciences) as described previously.44 Human GSTM1À1
and GSTM2À2 were heterologously expressed from the pKK-D vector45
in E. coli XL1-Blue cells (Strategene, La Jolla, CA) and purified by affinity
chromatography as described previously.46,47 Human GSTP1À1 was ex-
pressed and purified as described previously.48 Human GSTT1À1 was
expressed and purified as described previously.49 The high purity of
the enzymes was confirmed by SDS/PAGE stained with Commassie
Brilliant Blue R-250. MGST1 was purified from male SpragueÀDawley
rat livers as described previously,18 with the exception that 0.2% Triton
X-100 was used in the last purification step. Protein concentration was
determined using the method of Peterson with bovine serum albumin as
standard50 or by spectroscopic quantification (cytosolic GSTs) based on
known extinction coefficients at 280 nm.
In order to assess the RE, the second-order rate constant of the
nonenzymatic reaction knoncat was determined and compared with the
apparent second-order rate constant for the enzyme-catalyzed reaction
kcat/Km (i.e., kcat/Km divided by knoncat).
Measurement of GST Activity with Standard Substrates.
)
The specific activity of the purified cytosolic GSTs (except GSTT1À1
Cell Culture. Cell lines derived from human breast carcinoma
(MCF7 cells) were cultured as described in ref 53. In short, the
MGST1-expressing and vector control MCF7 cells, which had been
transfected with a vector for overexpressing rat-MGST1 or the same
vector containing the antisense DNA, were cultured in Dulbecco’s
modified Eagle’s medium (DMEM, Gibco 41965), supplemented with
10% fetal bovine serum, 100 U/mL penicillin, 100 μg/mL streptomycin,
1 mM sodium pyruvate (Gibco, Paisley, UK), and 1 mg/mL Geneticin
(G418, GE Healthcare, Piscataway, NJ) at 37 ꢀC and 5% CO2 in a
humidified environment.
Activity Measurements with Cell Lysate. Cells were cultured
in a T162 flask. When confluent, cells were washed with 1Â phosphate-
buffered saline (PBS), trypsinized, resuspended in culture medium, and
centrifuged at 2000 rpm for 5 min. The medium was removed and the
pellet was resuspended in 0.1 M potassium phosphate buffer pH 6.5
containing 1% Triton X-100, transferred to an Eppendorf tube, and kept
on ice. Sonication was performed for 3 Â 10 s at 6 A on ice with an MSE
Soniprep 150 (Sanyo Gallenkamp PLC, Leicester, UK). The GST
activity was then determined with the fluorogenic probes in 0.1 M
potassium phosphate buffer pH 6.5 containing 0.1% Triton, using the
same assay protocol as that described under “Novel GST Activity Assay
Based on Monitoring Fluorophore Release”. MGST1 was activated by
incubation of the cell suspension with 5 mM NEM (Sigma-Aldrich, St.
Louis, MO) on ice for 15 min, and the reaction was terminated by the
addition of 5 mM GSH. Protein concentration was determined by the
Micro BCA protein assay kit (Pierce, Rockford, IL; No. 23235) in a 96-
well plate with bovine serum albumin as standard.
and MGST1 was measured using 5 mM GSH (Sigma-Aldrich, St. Louis,
MO) and 0.5 mM CDNB (Merck, Darmstadt, Germany) as second
substrate in a 100 μL cuvette with a single-beam Philips PU8700 UV/
visible spectrophotometer (Philips Scientific & Analytical Equipment,
Cambridge, UK) by following the change in absorbance at 340 nm. The
specific activity of GSTT1À1 was determined with 0.5 mM 1,2-epoxy-
3-(p-nitrophenoxy)propane (EPNP, Sigma-Aldrich, St. Louis, MO)
instead of CDNB as second substrate by monitoring the change in
absorbance at 360 nm. The molar extinction coefficients used for
CDNB conjugation and EPNP conjugation were 9.6 and 0.5 mMÀ1cmÀ1
,
respectively.51 All enzyme activity measurements were performed at
room temperature. The cytosolic GSTs were assayed in 0.1 M potassium
phosphate buffer pH 6.5, whereas MGST1 was assayed in 0.1 M pota-
ssium phosphate buffer pH 6.5 containing 0.1% Triton X-100 (required
for enzyme solubility, Sigma-Aldrich, St. Louis, MO). Enzymatic activ-
ities were calculated after correction for the nonenzymatic reaction.
These measurements were performed in order to validate the activity
and stability of the enzymes. The specific activities were in general
agreement with the values reported previously.48,52
Novel GST Activity Assay Based on Monitoring Fluoro-
phore Release. The conjugation of GSH with the fluorogenic probes
catalyzed by cytosolic and microsomal GSTs was measured with a
Shimadzu RF-510LC fluorescence spectrophotometer (Analytical In-
struments Division, Kyoto, Japan) by monitoring the release of the
relevant fluorophores. Essentially the same assay protocol as that
described above was used with DNs-Coum, DNs-AcRh, MNs-AcRh,
PBs-AcRh, DNs-CV, MNs-CV, or Ts-CV as second substrate. The
concentration of DNs-Coum used was 40 μM, the concentration of
DNs-AcRh, MNs-AcRh, and PBs-AcRh used was 25 μM, and the
concentration of DNs-CV, MNs-CV, and Ts-CV used was 2.5 μM.
Initial rates were calculated from the relative increase in fluorescence
intensity when 7-amino-4-methyl-3-coumarinylacetic acid (Coum, Ex =
Fluorescence Microscopic Imaging. The MGST1-expressing
and vector control MCF7 cells were seeded at a density of 1 Â 105 cells/
well in a 6-well plate and cultured until 80À90% confluent (∼48 h).
Imaging experiments were then carried out. Cells were washed with 1 Â
PBS and incubated with one of the fluorogenic probes (DNs-Coum,
DNs-AcRh, or DNs-CV) in serum-free DMEM without phenol red
(Gibco 21063) (0.2% DMSO as a cosolvent) at 37 ꢀC for 30 min. The
medium was replaced with new serum-free DMEM without phenol red,
and fluorescence images were captured using an inverted Nikon
375 nm, Em = 450 nm), N-acetylrhodamine (AcRh, Ex = 490 nm, Em
=
522 nm) or cresyl violet (CV, Ex = 540 nm, Em = 620 nm) was formed.
Calibration curves with Coum, AcRh, and CV in 0.1 M potassium
phosphate buffer pH 6.5 with and without 0.1% Triton X-100 were
14117
dx.doi.org/10.1021/ja205500y |J. Am. Chem. Soc. 2011, 133, 14109–14119

Journal of the American Chemical Society
ARTICLE
ECLIPSE TE2000-S fluorescence microscope at 400Â magnification.
Cells were then fixed for 15 min at room temperature with 2%
formaldehyde and DNA was stained with 1 μg/mL Hoechst 33342
(Invitrogen, Paisley, UK) for 10 min on shaker at room temperature to
localize nuclei. Fluorescence images were captured again at 400Â
magnification. As a negative control, cells were pretreated with 500 μM
NEM for 15 min before adding either of the fluorogenic probes.
(16) Meyer, D. J.; Coles, B.; Pemble, S. E.; Gilmore, K. S.; Fraser,
G. M.; Ketterer, B. Biochem. J. 1991, 274 (Pt 2), 409.
(17) Bolton, R. M.; Haritos, V. S.; Whitehouse, M. W.; Ahokas, J. T.
Anal. Biochem. 1994, 216, 418.
(18) Nauen, R.; Stumpf, N. Anal. Biochem. 2002, 303, 194.
(19) Cook, J. A.; Iype, S. N.; Mitchell, J. B. Cancer Res. 1991,
51, 1606.
(20) Shrieve, D. C.; Bump, E. A.; Rice, G. C. J. Biol. Chem. 1988,
263, 14107.
(21) Rice, G. C.; Bump, E. A.; Shrieve, D. C.; Lee, W.; Kovacs, M.
Cancer Res. 1986, 46, 6105.
’ ASSOCIATED CONTENT
S
(22) Svensson, R.; Greno, C.; Johansson, A. S.; Mannervik, B.;
Morgenstern, R. Anal. Biochem. 2002, 311, 171.
(23) Fujikawa, Y.; Urano, Y.; Komatsu, T.; Hanaoka, K.; Kojima, H.;
Terai, T.; Inoue, H.; Nagano, T. J. Am. Chem. Soc. 2008, 130, 14533.
(24) de Silva, A. P.; Gunaratne, H. Q.; Gunnlaugsson, T.; Huxley,
A. J.; McCoy, C. P.; Rademacher, J. T.; Rice, T. E. Chem Rev 1997,
97, 1515.
(25) Uchiyama, S.; Takehira, K.; Kohtani, S.; Imai, K.; Nakagaki, R.;
Tobita, S.; Santa, T. Org. Biomol. Chem. 2003, 1, 1067.
(26) Jiang, W.; Fu, Q.; Fan, H.; Ho, J.; Wang, W. Angew. Chem., Int.
Ed. 2007, 46, 8445.
Supporting Information. Details of probe synthesis and
b
1H NMR and 13C NMR data. This material is available free of
charge via the Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION
Corresponding Author
ralf.morgenstern@ki.se; h-abe@riken.jp
’ ACKNOWLEDGMENT
(27) Pu, L. Chem. Rev. 2004, 104, 1687.
(28) Koeplinger, K. A.; Zhao, Z.; Peterson, T.; Leone, J. W.;
Schwende, F. S.; Heinrikson, R. L.; Tomasselli, A. G. Drug Metab.
Dispos. 1999, 27, 986.
(29) Zhao, Z.; Koeplinger, K. A.; Peterson, T.; Conradi, R. A.;
Burton, P. S.; Suarato, A.; Heinrikson, R. L.; Tomasselli, A. G. Drug
Metab. Dispos. 1999, 27, 992.
(30) Shibata, A.; Furukawa, K.; Abe, H.; Tsuneda, S.; Ito, Y. Bioorg.
Med. Chem. Lett. 2008, 18, 2246.
(31) Alander, J.; Johansson, K.; Heuser, V. D.; Farebo, H.; Jarvliden,
J.; Abe, H.; Shibata, A.; Ito, M.; Ito, Y.; Morgenstern, R. Anal. Biochem.
2009, 390, 52.
These studies were supported by the Swedish Research Council,
Swedish Foundation for Strategic Research, VINNOVA,
and funds from Karolinska Institutet. H.A. was financially supported
by MEXT (Ministry of Education, Culture, Sports, Science and
Technology of Japan). A.S. was financially supported by Grant-in-
Aid for Young Scientists (B) (22790124) and the Special Post-
doctoral Researcher Program of RIKEN. We are grateful to Dr.
Koshino for NMR analysis (Molecular Characterization Team,
RIKEN).
(32) Shibata, A.; Abe, H.; Ito, M.; Kondo, Y.; Shimizu, S.; Aikawa, K.;
Ito, Y. Chem. Commun. (Cambridge) 2009, 6586.
(33) Lavis, L. D.; Chao, T. Y.; Raines, R. T. ACS Chem. Biol. 2006,
1, 252.
’ REFERENCES
(1) Mannervik, B.; Danielson, U. H. CRC Crit. Rev. Biochem. 1988,
23, 283.
(34) In Lehninger Principles of Biochemistry, 4th ed.; Nelson, D. L.,
Cox, M. M., Eds.; W. H. Freeman & Co.: New York, 2004; p 190.
(35) Albery, W. J.; Knowles, J. R. Biochemistry 1976, 15, 5631.
(36) Morgenstern, R.; Lundqvist, G.; Hancock, V.; DePierre, J. W.
J. Biol. Chem. 1988, 263, 6671.
(37) Paumi, C. M.; Ledford, B. G.; Smitherman, P. K.; Townsend,
A. J.; Morrow, C. S. J. Biol. Chem. 2001, 276, 7952.
(38) Weinander, R.; Mosialou, E.; DeJong, J.; Tu, C. P.; Dypbukt, J.;
Bergman, T.; Barnes, H. J.; Hoog, J. O.; Morgenstern, R. Biochem.
J. 1995, 311 (Pt 3), 861.
(39) Townsend, A. J.; Tu, C. P.; Cowan, K. H. Mol. Pharmacol. 1992,
41, 230.
(40) Morrow, C. S.; Chiu, J.; Cowan, K. H. J. Biol. Chem. 1992,
267, 10544.
(41) Morgenstern, R. Chem. Res. Toxicol. 1998, 11, 703.
(42) Rahavendran, S. V.; Karnes, H. T. J. Pharm. Biomed. Anal. 1996,
15, 83.
(2) Josephy, P. D.; Mannervik, B. Molecular Toxicology; 2nd ed.;
Oxford University Press: New York, 2006.
(3) Hayes, J. D.; Flanagan, J. U.; Jowsey, I. R. Annu. Rev. Pharmacol.
Toxicol. 2005, 45, 51.
(4) McIlwain, C. C.; Townsend, D. M.; Tew, K. D. Oncogene 2006,
25, 1639.
(5) Sheehan, D.; Meade, G.; Foley, V. M.; Dowd, C. A. Biochem. J.
2001, 360, 1.
(6) Zimniak, P.; Singh, S. P. In Toxicology of Glutathione Transferases;
Awasthi, Y. C., Ed.; CRC Press, Taylor & Francis Group, LLC: Boca
Raton, FL, 2007; p 11.
(7) Morgenstern, R.; DePierre, J. W.; Ernster, L. Biochem. Biophys.
Res. Commun. 1979, 87, 657.
(8) Morgenstern, R.; Meijer, J.; Depierre, J. W.; Ernster, L. Eur.
J. Biochem. 1980, 104, 167.
(9) Morgenstern, R.; Svensson, R.; Bernat, B. A.; Armstrong, R. N.
Biochemistry 2001, 40, 3378.
(43) Yang, H.; Luo, G.; Karnchanaphanurach, P.; Louie, T. M.; Rech,
I.; Cova, S.; Xun, L.; Xie, X. S. Science 2003, 302, 262.
(44) Gustafsson, A.; Mannervik, B. J. Mol. Biol. 1999, 288, 787.
(45) Bjornestedt, R.; Widersten, M.; Board, P. G.; Mannervik, B.
Biochem. J. 1992, 282 (Pt 2), 505.
(46) Widersten, M.; Huang, M.; Mannervik, B. Protein Expression
Purif. 1996, 7, 367.
(47) Johansson, A. S.; Bolton-Grob, R.; Mannervik, B. Protein
Expression Purif. 1999, 17, 105.
(10) Shimoji, M.; Aniya, Y.; Morgenstern, R. In Toxicology of
Glutathione Transferases; Awasthi, Y. C., Ed.; CRC Press, Taylor &
Francis Group, LLC: Boca Raton, FL, 2007; p 293.
(11) Svensson, R.; Alander, J.; Armstrong, R. N.; Morgenstern, R.
Biochemistry 2004, 43, 8869.
(12) Morgenstern, R.; Guthenberg, C.; Depierre, J. W. Eur. J. Biochem.
1982, 128, 243.
(13) Morgenstern, R.; DePierre, J. W. Eur. J. Biochem. 1983,
134, 591.
(48) Johansson, A. S.; Stenberg, G.; Widersten, M.; Mannervik, B.
J. Mol. Biol. 1998, 278, 687.
(49) Shokeer, A.; Larsson, A. K.; Mannervik, B. Biochem. J. 2005,
388, 387.
(14) Weinander, R.; Ekstrom, L.; Andersson, C.; Raza, H.; Bergman,
T.; Morgenstern, R. J. Biol. Chem. 1997, 272, 8871.
(15) Ricci, G.; Caccuri, A. M.; Lo Bello, M.; Pastore, A.; Piemonte,
F.; Federici, G. Anal. Biochem. 1994, 218, 463.
14118
dx.doi.org/10.1021/ja205500y |J. Am. Chem. Soc. 2011, 133, 14109–14119

Journal of the American Chemical Society
ARTICLE
(50) Peterson, G. L. Anal. Biochem. 1977, 83, 346.
(51) Habig, W. H.; Pabst, M. J.; Jakoby, W. B. J. Biol. Chem. 1974,
249, 7130.
(52) Mannervik, B.; Widersten, M. In Advances in Drug Metabolism
in Man; Pacifici, G. M., Fracchia, G. N., Eds.; European Commission:
Luxembourg, 1995, p 407.
(53) Johansson, K.; Ahlen, K.; Rinaldi, R.; Sahlander, K.; Siritantikorn,
A.; Morgenstern, R. Carcinogenesis 2007, 28, 465.
14119
dx.doi.org/10.1021/ja205500y |J. Am. Chem. Soc. 2011, 133, 14109–14119