Design and synthesis of highly sensitive fluorogenic substrates for glutathione S-transferase and application for activity imaging in living cells

Article abstract of DOI:10.1021/ja802423n

Here we report the development of fluorogenic substrates for glutathione S-transferase (GST), a multigene-family enzyme mainly involved in detoxification of endogenous and exogenous compounds, including drug metabolism. GST is often overexpressed in a variety of malignancies and is involved in the development of resistance to various anticancer drugs. Despite the medical significance of this enzyme, no practical fluorogenic substrates for fluorescence imaging of GST activity or for high-throughput screening of GST inhibitors are yet available. So, we set out to develop new fluorogenic substrates for GST. In preliminary studies, we found that 3,4-dinitrobenzanilide (NNBA) is a specific substrate for GST and established the mechanisms of its glutathionylation and denitration. Using these results as a basis for off/on control of fluorescence, we designed and synthesized new fluorogenic substrates, DNAFs, and a cell membrane-permeable variant, DNAT-Me. These fluorogenic substrates provide a dramatic fluorescence increase upon GST-catalyzed glutathionylation and have excellent kinetic parameters for the present purpose. We were able to detect nuclear localization of GSH/GST activity in HuCCT1 cell lines with the use of DNAT-Me. These results indicate that the newly developed fluorogenic substrates should be useful not only for high-throughput GST-inhibitor screening but also for studies on the mechanisms of drug resistance in cancer cells.

Full text of DOI:10.1021/ja802423n

Published on Web 10/09/2008
Design and Synthesis of Highly Sensitive Fluorogenic
Substrates for Glutathione S-Transferase and Application for
Activity Imaging in Living Cells
Yuuta Fujikawa,†,‡ Yasuteru Urano,†,§ Toru Komatsu,†,‡ Kenjiro Hanaoka,†,‡
Hirotatsu Kojima,†,‡ Takuya Terai,†,‡ Hideshi Inoue,| and Tetsuo Nagano*,†,‡
Graduate School of Pharmaceutical Sciences, The UniVersity of Tokyo, 7-3-1 Hongo, Bunkyo-ku,
Tokyo 113-0033, Japan, CREST, Japan Science and Technology Agency, Kawaguchi, Saitama,
Japan, PRESTO, Japan Science and Technology Agency, Kawaguchi, Saitama, Japan, and
School of Life Sciences, Tokyo UniVersity of Pharmacy and Life Sciences, 1432-1 Horinouchi,
Hachioji, Tokyo 192-0392, Japan
Received April 7, 2008; E-mail: tlong@mol.f.u-tokyo.ac.jp
Abstract: Here we report the development of fluorogenic substrates for glutathione S-transferase (GST),
a multigene-family enzyme mainly involved in detoxification of endogenous and exogenous compounds,
including drug metabolism. GST is often overexpressed in a variety of malignancies and is involved in the
development of resistance to various anticancer drugs. Despite the medical significance of this enzyme,
no practical fluorogenic substrates for fluorescence imaging of GST activity or for high-throughput screening
of GST inhibitors are yet available. So, we set out to develop new fluorogenic substrates for GST. In
preliminary studies, we found that 3,4-dinitrobenzanilide (NNBA) is a specific substrate for GST and
established the mechanisms of its glutathionylation and denitration. Using these results as a basis for
off/on control of fluorescence, we designed and synthesized new fluorogenic substrates, DNAFs, and a
cell membrane-permeable variant, DNAT-Me. These fluorogenic substrates provide a dramatic fluorescence
increase upon GST-catalyzed glutathionylation and have excellent kinetic parameters for the present
purpose. We were able to detect nuclear localization of GSH/GST activity in HuCCT1 cell lines with the
use of DNAT-Me. These results indicate that the newly developed fluorogenic substrates should be useful
not only for high-throughput GST-inhibitor screening but also for studies on the mechanisms of drug
resistance in cancer cells.
Introduction
processes, including xenobiotic biotransformation, drug me-
tabolism, protection against endogenous ROS-induced toxic
Optical detection is one of the most powerful tools to visualize
intracellular and extracellular events with high sensitivity and
specificity. In the field of chemistry, many fluorescence probes
have been developed for metal ions and small organic com-
pounds, as well as for reactive oxygen species (ROS).1-4
However, there are few nonpeptidic reporter substrates that can
directly read out the activity of enzymes, except for hydrolases
and kinases. In this paper, we report new fluorogenic substrates
that can directly visualize the activity of glutathione S-transferase
(GST).
products, biosynthesis of prostaglandins and steroid hormones,
and degradation of aromatic amino acids.6 Generally, GST
catalyzes the conjugation of reduced tripeptide glutathione
(GSH) to target molecules, including a wide range of endog-
enous and exogenous electrophilic substrates. The Pi isoform
of GST, named GSTP, was first described as a placental isoform.
It seems to play a pivotal role in the carcinogenic process7-9
and in the development of drug resistance.10 GSTP is the
predominant GST isozyme overexpressed in a wide range of
human tumors (gastric, ovarian, nonsmall cell lung, breast,
colon, pancreas, lymphoma, etc.).11 It is usually localized in
cytosol, but some reports also describe its localization in the
GST is a family of dimeric enzymes involved in phase II
detoxification reactions.5 They participate in a wide range of
† The University of Tokyo.
(6) Hayes, J. D.; Flanagan, J. U.; Jowsey, I. R. Ann. ReV. Pharmacol.
Toxicol. 2005, 45, 51–88.
‡ CREST, JST Corporation.
§ PRESTO, JST Corporation.
(7) Dang, D. T.; Chen, F.; Kohli, M.; Rago, C.; Cummins, J. M.; Dang,
L. H. Cancer Res. 2005, 65, 9485–9494.
| Tokyo University of Pharmacy and Life Sciences.
(1) Charier, S.; Ruel, O.; Baudin, J. B.; Alcor, D.; Allemand, J. F.; Meglio,
A.; Jullien, L. Angew. Chem., Int. Ed. 2004, 43, 4785–4788.
(2) Maeda, H.; Katayama, K.; Matsuno, H.; Uno, T. Angew. Chem., Int.
Ed. 2006, 45, 1810–1813.
(8) Henderson, C. J.; Smith, A. G.; Ure, J.; Brown, K.; Bacon, E. J.; Wolf,
C. R. Pro. Natl. Acad. Sci. U.S.A. 1998, 95, 5275–5280.
(9) Elsby, R.; Kitteringham, N. R.; Goldring, C. E.; Lovatt, C. A.;
Chamberlain, M.; Henderson, C. J.; Wolf, C. R.; Park, B. K. J. Biol.
Chem. 2003, 278, 22243–22249.
(3) Miller, E. W.; Chang, C. J. Curr. Opin. Chem. Biol. 2007, 11, 620–
625.
(10) Tew, K. D. Cancer Res. 1994, 54, 4313–4320.
(11) Tew, K. D.; Monks, A.; Barone, L.; Rosser, D.; Akerman, G.; Montali,
J. A.; Wheatley, J. B.; Schmidt, D. E. Mol. Pharmacol. 1996, 50,
149–159.
(4) Tang, B.; Xing, Y. L.; Li, P.; Zhang, N.; Yu, F. B.; Yang, G. W.
J. Am. Chem. Soc. 2007, 129, 11666–11667.
(5) Armstrong, R. N. Chem. Res. Toxicol. 1997, 10, 2–18.
9
10.1021/ja802423n CCC: $40.75
2008 American Chemical Society
J. AM. CHEM. SOC. 2008, 130, 14533–14543 14533

A R T I C L E S
Fujikawa et al.
nucleus.12-17 Nuclear GSTP is strongly and inversely correlated
to survival of patients with certain types of cancers.14,18,19 So,
in addition to the activity of intracellular GST, its localization
is biologically important.
Table 1. Initial Velocity in the Reaction of 10 µM Test Compound
with 1.0 mM GSH in the Absence/Presence of hGSTP1
initial velocity (µM/sec)
without hGSTP with hGSTP activationa
So far, the measurement of GST activity has often been done
with the aid of a chromogenic substrate, 2,4-dinitrochloroben-
zene (CDNB). But this method suffers from low sensitivity of
the chromogenic compound, as well as the fast spontaneous
reaction with GSH. Several fluorescent probes or detection
methods have been developed for highly sensitive detection of
GST activity20-23 (Figure S1, Supporting Information). But,
there is no practical fluorogenic substrate that permits the
measurement of GSH/GST activity in living cells. For example,
Haugland and colleagues developed PFB-F, which contains a
pentafluorobenzoyl moiety (Figure S1a, Supporting Informa-
tion).20 The pentafluorobenzoyl moiety of PFB-F is glutathio-
nylated by GST, but the fluorescence properties are not changed
by glutathionylation. Thus, the measurement of intracellular total
GST activity using CDNB and PFB-F requires careful cell
homogenization to avoid degradation by proteases. Not only is
there a possibility of misestimating intracellular enzyme activity
but also information about the distribution of enzymatic activity
is completely lost. Bimanes, such as monobromobimane (mBBr)
and monochlorobimane (mBCl)23,24 are used for detecting
intracellular glutathionylation (Figure S1b, Supporting Informa-
tion). Before reaction, bimanes are fluorescently silent, but once
glutathionylated, they become strongly fluorescent. This activa-
tion of fluorescence is a useful property for GSH/GST activity
imaging, but bimanes are excited in the UV range (<400 nm),
which damages cells and tissues of interest, and high background
fluorescence is also a problem. Furthermore, as bimanes have
high reactivity to GSH in the absence of GST, it is difficult to
evaluate GST activity, particularly in living cells. Practical
fluorogenic substrates that would make it possible to visualize
GSH/GST activity in living cells by means of fluorescence
microscopy are still needed, for example, for examining the
characteristics of GST-overexpressing cancer cells and cells with
the multidrug-resistance phenotype.
2,4-Chlorodinitrobenzene (CDNB)
Pentafluorobenzanilide (PFBA)b
4-Chlor-3-nitrobenzanilide (CNBA)
4-Fluoro-3-nitrobenzanilide (FNBA)
3,4-Dinitrobenzanilide (NNBA)
0.2
1.6
10.5
-
N.D.c
N.D.
N.D.
0.02
N.D.
N.D.
N.D.
N.D.
0.2
9.0
0.2
-
N.C.d
528
N.C.
N.C.
4-Chloro-3-nitroacetophenone (CNAP)
4-Chloro-3-nitrobenzophenone (CNBP)
0.2
a Activation ) fold change of initial velocity in with/without GSTP.
b HPLC confirmed. c N.D.; not detectable. d N.C.; not calculable.
of all, we explored and found NNBA as a specific substrate for
recombinant human GSTP1-1 (hGSTP1). Based on this
molecule, we developed new fluorogenic substrates for GST,
DNAFs, and DNAT-Me. DNAFs are highly sensitive and useful
to evaluate recombinant and lysate GST activity in vitro and
DNAT-Me is suitable to visualize the distribution of intracellular
GST activity in fluorescence microscopic observation in living
cells.
Results and Discussion
Evaluation of Specific Substrates for GST. Generally, devel-
opment of enzymatically activated fluorogenic substrates faces
two main difficulties. One is to achieve fluorescence activation
by specific reaction, and the other is to ensure that the target
enzyme recognizes the compound as a specific substrate. In
particular, the latter is the first requisite for the development of
enzymatically activated fluorogenic substrates. Both nonenzy-
matic and enzyme-catalyzed nucleophilic conjugations of GSH
to CDNB are believed to follow a two-step SNAr addition/
elimination mechanism. Substrates glutathionylated by GST are
usually electron-deficient compounds such as CDNB, DCNB
(1,2-dichloro-4-nitrobenzene), and other nitro-group-containing
aromatic and/or aliphatic compounds. Although these com-
pounds are glutathionylated enzymatically, the noncatalyzed
reaction with GSH is relatively fast. So we needed to find a
substrate that would specifically undergo GST-catalyzed glu-
tathionylation. For this purpose, we spectrophotometrically
examined the reactivity of 7 compounds, including nitroben-
zanilide derivatives and nitrobenzene derivatives, toward GSH/
hGSTP1 (Figure S2, Supporting Information). The results are
summarized in Table. 1. The highest value of activation was
seen with NNBA (activation ≈ 530). Previously reported
compounds, CNAP, CNBP, and a new compound, FNBA,
reacted with GSH only in the presence of hGSTP1. This feature
is ideal in terms of specific enzyme-catalyzed reaction, but the
reaction rates are extremely slow compared with that of NNBA
(about 45-fold difference). Threfore, these compounds were not
suitable for our purpose. Further, CNBA did not react under
these conditions, though it was previously reported to be a
substrate for rat GST. Interestingly, the absorbance spectrum
of PFBA, which is the reactive moiety of PFB-F, did not change
upon reaction with GSH/hGSTP1 under these conditions.
Glutathionylation of PFBA was not detectable with HPLC even
at 100 µM. As shown in Figure 1A and B, NNBA was
recognized by hGSTP1 and had excellent kinetic parameters
Here, we present the first practical fluorogenic substrates for
GST, developed on the basis of a rational design strategy. First
(12) Maguire, N. C.; Drozd, K.; Luff, R. D.; Kantor, R. R. S. Acta Cytol.
1991, 35, 94–99.
(13) Goto, S.; Ihara, Y.; Urata, Y.; Izumi, S.; Abe, K.; Koji, T.; Kondo, T.
FASEB J. 2001, 15, 2702–2714.
(14) Soh, Y.; Goto, S.; Kitajima, M.; Moriyama, S.; Kotera, K.; Nakayama,
T.; Nakajima, H.; Kondo, T.; Ishimaru, T. Clin. Oncol. 2005, 17, 264–
270.
(15) Goto, S.; Kamada, K.; Soh, Y.; Ihara, Y.; Kondo, T. Jpn. J. Cancer
Res. 2002, 93, 1047–1056.
(16) Hall, A. G.; McGuckin, A. G.; Pearson, A. D. J.; Cattan, A. R.;
Malcolm, A. J.; Reid, M. M. J. Clin. Pathol. 1994, 47, 468–469.
(17) Campbell, J. A. H.; Corrigall, A. V.; Guy, A.; Kirsch, R. E. Cancer
1991, 67, 1608–1613.
(18) Ali-Osman, F.; Brunner, J. M.; Kutluk, T. M.; Hess, K. Clin. Cancer
Res. 1997, 3, 2253–2261.
(19) Allen, T. C.; Granville, L. A.; Cagle, P. T.; Haque, A.; Zander, D. S.;
Barrios, R. Human Pathol. 2007, 38, 220–227.
(20) Arttamangkul, S.; Bhalgat, M. K.; Haugland, R. P.; Diwu, Z. J.; Liu,
J. X.; Klaubert, D. H.; Haugland, R. P. Anal. Biochem. 1999, 269,
410–417.
(21) Svensson, R.; Greno, C.; Johansson, A. S.; Mannervik, B.; Morgen-
stern, R. Anal. Biochem. 2002, 311, 171–178.
(22) Zhou, W. H.; Shultz, J. W.; Murphy, N.; Hawkins, E. M.; Bernad, L.;
Good, T.; Moothart, L.; Frackman, S.; Klaubert, D. H.; Bulleit, R. F.;
Wood, K. V. Chem. Commun. 2006, 4620–4622.
(23) Nauen, R.; Stumpf, N. Anal. Biochem. 2002, 303, 194–198.
(24) Hulbert, P. B.; Yakubu, S. I. J. Pharm. Pharmacol. 1983, 35, 384–
386.
(KM ) 23.2 µM, kcat ) 367 s-1, kcat/KM ) 1.6 × 107 M-1 sec-1
)
for the hGSTP1-catalyzed reaction (Figure 1C). After comple-
tion of the reaction, nitrite was present in quantitative yield in
the reaction mixture, suggesting that a nitro group was released
9
14534 J. AM. CHEM. SOC. VOL. 130, NO. 44, 2008

Highly Sensitive Fluorogenic Substrates for GST
A R T I C L E S
Figure 1. NNBA is good a substrate for GST. (A) Absorbance spectra of 10 µM NNBA and its reaction product in the presence of 1.0 mM GSH and 0.0241
µg/mL hGSTP1 in phosphate-buffered saline (PBS, pH 7.4, 0.2% DMSO as a cosolvent) at 37 °C. The arrow shows the direction of the absorbance changes.
(Blue, NNBA; magenta, after reaction for 600 s.) (B) Time courses of reaction of 10 µM NNBA under diffrent conditions (magenta; 1.0 mM GSH and
0.0241 µg/mL hGSTP1, black; 1.0 mM GSH). (C) Michaelis-Menten plot for NNBA. Assay conditions were as follows: 3-100 µM NNBA in PBS (pH
7.4, 1.0% DMSO as a cosolvent) containing 1.0 mM GSH, and 0.241 µg/mL hGSTP1. (D) Linear relation between concentrations of NNBA and released
nitrite. After completion of the reaction at various NNBA concentrations (6-100 µM) in (C), nitrite released from NNBA in the reaction mixture was
measured using a NO2/NO3 Assay Kit-C II (Colorimetric) Griess Reagent Kit (Dojindo, Japan).
Table 2. Electrochemical Properties of Chromogens for GSTs
as a leaving group with substitution of the sulfhydryl group of
GSH (Figure 1D). Which nitro group was eliminated? To
address this question, glutathionylated NNBA, named GS-NBA,
was synthesized from GSH and NNBA under alkaline condi-
tions. In the 1H NMR spectrum, the chemical shift of the
5-position proton in NNBA was different from in GS-NBA
(Figure S3, Supporting Information), suggesting that the 4-nitro
group in NNBA is the leaving group. Thus, NNBA is a GST
substrate offering a high ratio of enzymatic/nonenzymatic
reaction rates, with excellent enzymatic kinetics for the intended
4-substitutent
Ered (V vs SCE)a
purpose. After finding this specific substrate, we next had to
consider how the fluorescence can be quenched/activated as a
result of the reaction. In general, nitro groups greatly lower the
LUMO energy level of aromatic compounds due to their strong
electron-withdrawing effect. In a previous report, we showed
that an electron-deficient benzene moiety can quench the fluo-
rescence of a fluorophore via an intramolecular photoinduced
electron transfer (PeT) process from the excited fluorophore to
the electron-deficient benzene moiety (donor-excited PeT;
d-PeT).25-27 In this mechanism, change of the LUMO level on
the reaction moiety leads to a change of the fluorescence
quantum yield upon specific reaction. Because denitration of
FNBA
CNBA
NNBA
TNBA
F
Cl
NO2
SEt
-1.08
-1.08
-0.72
-1.15
a Measured in acetonitrile.
NNBA occurs upon glutathionylation, we thought the LUMO
level of NNBA would rise. To address this question, we
measured the redox potential of NNBA by cyclic voltammetry
to estimate the extent of the shift upon reaction with GSH/hGST.
As shown in Table 2, NNBA has an exceptionally low reduction
potential. Compared with the glutathione-adduct model com-
pound, 4-ethylthio-3-nitrobenzanilide (TNBA) (-1.15 V vs
SCE), NNBA has a low reduction potential (-0.72 V vs SCE).
This large difference in reduction potential suggests that the
fluorescence intensity may be modulated via a d-PeT quenching
mechanim upon glutathionylation, when a fluorophore is at-
tached at an appropriate distance from the reactive moiety
(Figure 2).
(25) Ueno, T.; Urano, Y.; Setsukinai, K.; Takakusa, H.; Kojima, H.;
Kikuchi, K.; Ohkubo, K.; Fukuzumi, S.; Nagano, T. J. Am. Chem.
Soc. 2004, 126, 14079–14085.
(26) Ueno, T.; Urano, Y.; Kojima, H.; Nagano, T. J. Am. Chem. Soc. 2006,
128, 10640–10641.
(27) Mineno, T.; Ueno, T.; Urano, Y.; Kojima, H.; Nagano, T. Org. Lett.
2006, 8, 5963–5966.
9
J. AM. CHEM. SOC. VOL. 130, NO. 44, 2008 14535

A R T I C L E S
Fujikawa et al.
Figure 2. Strategy for the development of GST-catalyzed GSH substitution-activated quenched specific fluorescence probe. The xanthene moiety seems
likely to be orthogonal to the enzyme substrate moiety in PFB-F. Therefore, we hypothesized that the xanthene moiety might not interfere with the enzymatic
recognition. Thus, pentafluorobenzanilide is expected to be recognized by GST as a substrate. Optimization of this “lead” specific substrate would provide
a more reactive and electrochemically specific substrate. The optimized substrate can be recombined with the xanthene moiety to provide a quenched
fluorogenic substrate for GST.
Scheme 1. Synthetic Scheme for DNAFs
Synthesis of DNAFs. We designed fluorescence probes based
on the results of the above reactive moiety search, as illustrated
in Figure 1. Synthesis of fluorogenic substrates followed Scheme
1. Because of the availability of aminofluorescein isomers (5
and 6-isomer), we could synthesize dinitrobenzamide fluores-
ceins, named DNAFs, by means of very simple procedures.
Briefly, 3,4-dinitrobenzoic acid was converted to the acid
chloride by refluxing in thionyl chloride, followed by reaction
with the dipivalic ester-protected aminofluorescein. After depro-
tection with 2 N NaOH or sodium methoxide in MeOH, the
product was purified by means of semipreparative reverse-phase
HPLC. We named the 5- and 6-isomer derivatives DNAF1 and
DNAF2, respectively.
of hGSTP1. That is, a very fast increase of fluorescence intensity
was seen upon reaction with GSH/hGSTP1 (Figure 3A).
Fluorescence of DNAF1 was highly quenched (the quantum
efficacy (Q.E.) was 0.003), but the apparent Q.E. increased to
0.101 after completion of the reaction (Table 3). A 34-fold
increase of Q.E. was achieved. The linear correlation of enzyme
concentration to initial velocity calculated from the rate of
fluorescence increase suggests that the fluorescence increase of
DNAF1 is hGSTP1-activity dependent (Figure 3B). HPLC and
UPLC-MS analysis revealed that the main fluorescent product
is the glutathionylated DNAF1 (Figure 4). Thus, the glutathio-
nylation and subsequent denitration result in strong fluorescence
activation (Scheme 2). The specificity of DNAF1 for GSH/
hGSTP1 is confirmed in Figure S5 (Supporting Information).
The apparent Q.E. values of DNAF2 before and after reaction
were 0.008 and 0.257, respectively (Table 3). DNAF2 has a
Photochemical Properties of DNAFs. Photochemical proper-
ties of the synthesized compounds are shown in Table 3. As
expected, the newly developed DNAFs were good substrates
9
14536 J. AM. CHEM. SOC. VOL. 130, NO. 44, 2008

Highly Sensitive Fluorogenic Substrates for GST
A R T I C L E S
Table 3. Fluorescence Properties and Kinetic Parameters of
mBBr, DNAF1, and DNAF2a
App kcat/KM
(×104 M-1 sec-1 b
)
Absmax Emmax apparent k2 (GSH)
substrate glutathionylation (nm) (nm) Q.E. (M-1sec-1) hGSTA1 hGSTM1 hGSTP1
mBBr
before
after
before
after
before
after
394 483 0.009
388 487 0.191
493 512 0.003
492 516 0.101
492 511 0.008
492 513 0.257
1.5
258
68
734
213
DNAF1
DNAF2
0.2
210 1107
359 1066
0.09
149
a Fluorescence quantum yield was calculated using fluorescein (in 0.1
N NaOH aq.) and quinine sulfate (in 0.1 N H2SO4) as a standard. Assay
was performed in PBS (pH 7.4) at 37 °C. The GSH concentration was
1.0 mM for all substrates. The same protein concentrations as shown in
Table 2 were used for DNAFs. In the mBBr assay, protein con-
centrations were 157 ng/mL, 320 ng/mL, and 24.1 ng/mL for GSTA1,
GSTM1, GSTP1, respectively. b Apparent kcat/KM parameter when 1.0
mM GSH used.
somewhat higher Q.E. than DNAF1 and the change of Q.E.
upon reaction was about 32-fold. The observation that the
absorbance spectra of DNAFs did not change upon reaction
suggests that DNAFs are quenched via the PeT mechanism, as
expected (Figure S4, Supporting Information). In conclusion,
we believe this is the first report of a fluorescence probe utilizing
replacement of a nitro group as an off/on mechanism for
fluorescence.
Figure 3. (A) Fluorescence spectra of DNAF1 (0.5 µM in PBS containing
0.1% DMSO as a cosolvent, pH 7.4) before (blue) and after (magenta)
addition of 1.0 mM GSH and 0.241 µg/mL hGSTP. (B) Linear relationship
of initial velocity of DNAF1 fluorescence increase to concentration of
hGSTP1.
Kinetic Parameters of DNAFs. We next evaluated the kinetics
of the enzymatic reaction of DNAFs in detail. Table S1
(Supporting Information) summarizes the specific activity of
glutathionylation of DNAFs catalyzed by various GST isozymes.
Because of quenching at high concentrations of DNAFs (ex.
50 µM), we measured the specific activity at 0.5 µM concentra-
tion of DNAFs. DNAF1 appeared to be more susceptible to
hGSTP1 compared with DNAF2. The kinetic parameters of
mBBr and DNAFs are listed in Table 3. mBBr reacted much
faster than DNAFs with GSH in the absence of GST. Thus,
DNAFs are superior to known substrates in terms of sensitivity
and spontaneous reaction rate with GSH. We anticipate that
DNAFs would be well suited for in vitro assays of GST, for
example, for high throughput screening (HTS) of GST inhibitors.
To test the feasibility of this, we confirmed the inhibitory effect
of three known GST inhibitors with DNAF1 as a substrate
(Figure S6, Supporting Information). The apparent kcat/KM values
were derived using the Michaelis-Menten relationship at low
concentrations (0.5-1.5 µM) of DNAFs: v/([E] ·[S]) ) kcat/KM,
when KM > [S]. All enzymes displayed a linear rate vs substrate
concentration behavior in this concentration region. The kcat/
KM values of DNAFs were greater than those of mBBr,
particularly in the hGSTP1-catalyzed reaction. Both DNAFs
were best with hGSTP1, but the preference for hGSTP1 was
greater in the case of DNAF1, the 5-isomer. For DNAF1, kcat/
KM was calculated as 1107 × 104 (sec-1 M-1) for hGSTP1,
which is slightly lower (about 0.7-fold) than that of NNBA,
1582 × 104 (sec-1 M-1). This may be due to the presence of
the fluorophore.
cal vein endothelial cell line). Specific activities of these cell
lysates with DNAF1 are listed in Table 4. The highest specific
activity in was found in the A549 cell line. Higher GST specific
activity in A549 than in the normal counterpart, OUS11, may
reflect overexpression of the Pi subtype. Thus, DNAF1 could
evaluate GST activities from seven different cell lines. Highly
sensitive and specific assay with submicromolar concentrations
of DNAF1 and reduced numbers of cells might be suitable to
replace the conventional method of GST activity measurement
in cell lysates with chromogenic CDNB.
Design, Synthesis, and Evaluation of Membrane-Permeable
Derivative of DNAF1. DNAF1 is plasma membrane-imperme-
able due to the moderate hydrophilicity of the fluorescein
moiety. So, for application to living cells, we require a method
for introduction of DNAF1 into the cells. One approach is to
use a protecting group that is cleaved within the cell. In general,
hydrophilic fluorescence probes are introduced into cells using
protecting groups, such as acetoxymethyl group for carboxylic
acid,28 or acetate ester for phenol, which are hydrolyzed by
intracellular esterases. But we felt that this approach is not
suitable in the present case because of the possible complex
effect of the ester hydrolysis process on the fluorescence increase
arising from the reaction with GSH/GST. Another approach is
to use the more hydrophobic TokyoGreen (TG) scaffold, which
might be better suited for the imaging of GST/GSH activity in
living cells. So, we designed and synthesized the TokyoGreen
derivative of DNAF1. According to the general synthetic
procedure of TGs, we attempted halogen-lithium exchange
reaction with the iodonitrobenzene derivative, but this was
unsuccessful. We thought the failure of the synthesis might have
been due to the presence of the nitro group, and tried the
procedure reported by Peterson and co-workers.29 DNAT-Me
was synthesized in 6 steps (Scheme 3). Following reduction to
obtain 2-methyl-4-nitrobenzyl alcohol (1) from the correspond-
ing benzoic acid as a starting material, oxidation by PCC gave
Cell Lysate Experiment. Prior to biological application within
cells, we tested whether DNAF1 could detect differences of GST
activity in cell lysates prepared from seven cell lines: A549
(human lung adenocarcinoma cell line), HuCCT1 (human
cholangiocarcinoma cell line), HL-60 (human leukemia cell
line), HT1080 (fibrosarcoma cell line), OUS11 (cell line derived
from normal tissue of lung cancer patient), HeLa (human
cervical adenocarcinoma cell line), and HUVEC (human umbili-
9
J. AM. CHEM. SOC. VOL. 130, NO. 44, 2008 14537

A R T I C L E S
Fujikawa et al.
Figure 4. Confirmation of product formation by means of HPLC and LC-MS. HPLC chromatograms with absorbance (A) and fluorescence detection (B).
((1) Before reaction, (2) at 50 s after addition of hGSTP, (3) at 600 s (after completion of the reaction)). Reaction was performed with hGSTP (0.0723
µg/mL), reduced GSH (1.0 mM), and DNAF (100 µM) in PBS (pH 7.4) (1.0% DMSO as a cosolvent) at 37 °C. Absorbance was detected at 490 nm, and
fluorescence was detected at 514 nm with excitation at 490 nm. The HPLC conditions: linear gradient (eluent, 0 min 16% CH3CN/0.1 M TEAA aq. ≈ 20
min, 80% CH3CN/0.1 M TEAA aq.; flow rate ) 1.0 mL/min). (C) UPLC-MS chromatogram with absorbance detection at 490 nm of sample (3) in HPLC
analysis. (D) Mass spectrum of the peak at 5.3 min and corresponding m/z values analyzed by mass spectrometry in the negative ion mode. UPLC conditions:
eluent, 0 min 5% CH3CN/5 mM TEAA aq. ≈ 1.00 min, 95% CH3CN/5 mM TEAA aq. ≈ 1.50 min, 5% CH3CN/5 mM TEAA aq. ≈ 1.51 min, 5% CH3CN/5
mM TEAA aq. MS Scan: scan time 0.10 s, cone voltage; 30.0 V.
Scheme 2. GST-Catalyzed Glutathionylation of DNAF1
Table 4. Specific Activity of GSTa
due to the absence/presence of a 4-substituent in resorcinol.
Therefore, we used an excess of resorcinol to avoid side reaction.
cell line
A549 HuCCT1 HL60 HT1080 OUS11 HUVEC HeLa
This condensation afforded triarylmethane 3 in 40% yield. Then,
3 was refluxed with excess TsOH to afford crude 2-Me-4-NO2-
TG (4), which was reduced to 2-Me-4-NH2-TG (5) in 3% yield
from 3 (two steps). The desired product was thus synthesized
by almost the same procedure as DNAFs in 8% yield, and was
designated DNAT-Me. The fluorescence increase of DNAT-
Me in PBS (pH 7.4) containing 1.0 mM GSH and hGSTM1 at
37 °C is shown in Figure 5. The Q.E. of DNAT-Me is 0.009,
somewhat higher than that of DNAF1, and increased to 0.204
after completion of the reaction in the presence of GSH and
hGSTP1. Although the activation ratio of about 23 is less
than that of DNAF1 (about 34-fold), we think it is sufficient
GST activity
37.5
23.5 10.0
9.5
6.2
3.8
1.7
(pmol/sec/mg protein) (5.34 (2.2 (0.4 (1.8 (1.2 (0.3 (0.6
a GST activity was measured using DNAF1 in total cell lysates as
described in Materials and Methods. The enzymatic activity was
normalized for protein concentration and is shown as specific activity
(pmol/sec/mg protein). Data are means of three independent analysis.
2-methyl-4-nitrobenzaldehyde (2) in relatively high yield.
Condensation with resorcinol was employed the reported
conditions, except for the use of excess resorcinol. Polymeri-
zation occurred at the ratio of aldehyde 2: resorcinol ) 1:2.
The difference between our result and that of Peterson may be
9
14538 J. AM. CHEM. SOC. VOL. 130, NO. 44, 2008

Highly Sensitive Fluorogenic Substrates for GST
A R T I C L E S
Scheme 3. Synthetic Scheme of DNAT-Me
to detect intracellular GSH/GST activity. Next, we evaluated
the membrane permeability of DNAT-Me with DNAF1 using
HL60 cells. DNAT-Me had a very low Q.E. before glutathio-
nylation, but its entry into cells could be confirmed by highly
sensitive flow cytometric analysis. As shown Figure 5B, even
after 30 s, higher fluorescence intensity was observed in HL60
cells incubated with DNAT-Me was seen with DNAF1, which
showed almost the same intensity as nontreated cells. This
suggests that DNAT-Me has high membrane permeability,
and that extra-/intracellular equilibrium is rapidly attained.
A high intracellular concentration of DNAT-Me should
increase the probability of collision with GSTs in the
cytoplasm, leading to rapid formation of the fluorescent GS-
adduct.
Visualization of GSH/GST Activity in Living Cells by Fluo-
rescence Microscopy. In the lysate experiment, we found that
the activity of the GSH/GST system is high in cholangiocar-
cinoma cell line HuCCT1. Cholangiocarcinoma is a cancer that
is highly resistant to various anticancer drugs, and has a poor
prognosis.30-32 Nakajima and colleagues reported that GSTP1-1
is a resistance factor for anticancer drugs in cholangiocarci-
noma.33 In particular, high nuclear GSTP activity is a poor
prognostic factor. So we investigated whether nuclear localiza-
tion of GST activity in HuCCT1 could be imaged by fluores-
cence microscopy. Cultured cells were washed once with PBS
and incubated with 2 µM DNAT-Me for 15 min. As shown in
Figure 6, a strong fluorescence signal was seen in the nucleus.
HPLC analysis confirmed that the fluorescent product of DNAT-
Me was the same GS-adduct as the reaction with GSH/hGSTP1
(Figure S7, Supporting Information). The nuclear localization
of high fluorescence was also confirmed by confocal micro-
scopic observation after coincubation with a nuclear-specific
dye, SYTO-red (Figure 7). GSH-depleting agent, N-ethylma-
leimide (NEM), abolished the fluorescence, which confirmed
that the nuclear fluorescence signal can be attributed to the GS-
adduct (Figure S8, Supporting Information). It has been reported
that nuclear localization of GSTP in uterine cancer34 and glioma
cells18 is negatively correlated with patients’ survival. Goto and
colleagues demonstrated that this negative correlation was at
least partly due to the effect of nuclear GSTP in protecting DNA
against damage by anticancer drugs and lipid peroxides.13,35
Because cholangiocarcinoma is a highly resistant to various
Figure 5. Membrane-permeable fluorogenic substrate, DNAT-Me. (A)
Fluorescence increase of DNAT-Me in PBS (pH 7.4, 0.1% DMSO as a
cosolvent), containing 1.0 mM GSH, and hGSTM1 (1.569 µg/mL) at 37
°C. Excitation and emission wavelengths were 490 and 514 nm, respectively.
(B) High permeability of DNAT-Me. HL60 cells (5.0 × 105 cells/mL) were
mixed with 0.5 µM DNAF1 or DNAT-Me in PBS after 30 s and evaluated
with a BD LSR II flow cytometer. The gray peak represents untreated
samples. Blue and green represent 0.5 µM of DNAT-Me and DNAF1,
respectively.
Figure 6. Fluorescence microscopic images of living HuCCT1 cells. (A)
Brightfield and (B) fluorescence image of cells incubated with 2 µM DNAT-
Me for 15 min at 37 °C. Exposure time was 300 ms, and ND was 25%.
9
J. AM. CHEM. SOC. VOL. 130, NO. 44, 2008 14539

A R T I C L E S
Fujikawa et al.
Figure 7. Confocal microscopic imaging of DNAT-Me and SYTO-loaded HuCCT1 cells. HuCCT1 cells were pretreated with 5 µM CCCP in HBSS for 15
min at 37 °C. Then, the medium was changed to fresh HBSS containing DNAT-Me (2 µM), SYTO (1 µM), and CCCP (5 µM), and incubation was
continued 15 min. The cells were washed with PBS twice and fluorescence images were aquired through FITC and Cy5 filters for DNAT-Me and SYTO,
respectively.
anticancer drugs, the visualization of the nuclear GSTP activity
in HuCCT1 with DNAT-Me is of considerable interest and
confirms the utility of this membrane-permeable fluorogenic
substrate. This is first report of fluorescence imaging of the
intracellular distribution of GST activity using an activatable
fluorogenic substrate.
the lack of a high-through put assay format has restricted the
development of GST subtype-selective inhibitors. DNAFs are
markedly superior to known substrates (such as CDNB and
mBBr), especially in terms of sensitivity, and should be suitable
for high-throughput screening. The membrane-permeable variant
DNAT-Me is available for imaging GST activity in living cells,
and was used to visualize high GST activity, possibly due to
the Pi subtype, in the nucleus of HuCCT1 cells under a
fluorescence microscope. Fluorescence imaging of GSH/GST
activity is expected to be useful in many fields, such as obtaining
biological insights into drug-resistance phenotypes with high
spatio-temporal resolution, and in diagnosis for the detection
of GST-overexpressing preneoplastic foci and drug resistance.
Further studies on the biological applications of DNAFs and
DNAT-Me are in progress.
Conclusion
We have developed the first off/on-type fluorogenic substrates
for GSTs, DNAFs and DNAT-Me, and applied the latter for
imaging GST activity in living cells. GST is of clinical interest,
because it is overexpressed in various type of cancer cells and
is involved in drug resistance. Nevertheless, no practical
fluorescence probe for GST activity has been available. We
designed and synthesized fluorogenic substrates with an off/on
fluorescence switching mechanism based on the replacement
of a nitro group with GSH. The DNAFs, have excellent kinetic
parameters for multiwell plate experiments. GSTP is an
important inhibitor of JNK activity and may function as a
regulator of cell proliferation and apoptosis.9,36,37 The ability
of GSTP1 to inhibit JNK appears to be dependent upon the
presence of GSTP1 monomer in the cytosol of cells.38 So GSTP
inhibitors are candidate anticancer agents,39 and indeed many
GSTP-selective inhibitors have been developed.40-42 However,
Experimental Section
Materials and General Instrumentation. General chemicals
were purchased from Tokyo Chemical Industries, Wako Pure
Chemical, or Aldrich Chemical Co., and were used without further
purification. Special chemicals were DMSO (fluorometric grade,
Dojindo) and TBAP (electrochemical grade, Fluka). All solvents
were used after appropriate distillation or purification. NMR spectra
were recorded on a JNM-LA300 (JEOL) instrument at 300 MHz
1
for H NMR and 75 MHz for 13C NMR and on a JNM-LA400
(JEOL) instrument at 400 MHz for 1H NMR and 100 MHz for
13C. ESI mass spectra were measured with a JMS-T100LC
AccuToF (JEOL). UV-visible spectra were obtained on a Shi-
madzu UV-1600. Fluorescence photometric studies were performed
on a Hitachi F4500.
(28) Tsien, R. Y. Nature 1981, 290, 527–528.
(29) Mottram, L. F.; Maddox, E.; Schwab, M.; Beaufils, F.; Peterson, B. R.
Org. Lett. 2007, 9, 3741–3744.
(30) Patel, T. Hepatology 2001, 33, 1353–1357.
(31) Isa, T.; Kusano, T.; Shimoji, H.; Takeshima, Y.; Muto, Y.; Furukawa,
M. Am. J. Surg. 2001, 181, 507–511.
Cyclic Voltammetry. Cyclic voltammetry was performed as
reported.26
(32) Okuda, K.; Nakanuma, Y.; Miyazaki, M. J. Gastroenterol. Hepatol.
2002, 17, 1049–1055.
HPLC Analysis. HPLC analysis was performed on an Inertsil
ODS-3 (4.6 × 250 mm) column using an HPLC system composed
of a pump (G1312A, Agilent) and a detector (G1315B or G1321A,
Agilent). In the LC-MS experiment to analyze the reaction mixture
of DNAF1 and GSH in the presence of recombinant human
GSTP1-1, LC analysis was done with an Inertsil ODS-3 (2.1 ×
150 mm) column using a system composed of a pump (G1312A,
Agilent) and a detector (G1314A, Agilent).
Synthesis of Compounds. 4-Fluoro-3-nitrobenzanilide (FNBA).
To a solution of oxalyl chloride (745 mg, 5.87 mmol) in 15 mL of
CH2Cl2, DMF (457 µL, 5.87 mmol) was added. This solution was
stirred for 10 min at -20 °C, then 4-fluoro-3-nitrobenzoic acid (249
mg, 1.35 mmol) was added. Stirring was continued for 30 min,
then 10 mL of a CH2Cl2 solution of aniline (124 µL, 1.35 mmol)
and DIEA (1.2 mL, 7.22 mmol) was added at -20 °C. The mixture
was stirred for 2.5 h at room temperature, then, after confirmation
of disappearance of the starting materials, evaporated to dryness.
The residue was poured into AcOEt (100 mL). The organic solution
was washed with H2O (100 mL × 3) and brine, dried over
anhydrous Na2SO4, and filtered. The filtrate was evaporated to
dryness. The residue was purified by a preparative TLC (AcOEt/
(33) Nakajima, T.; Takayama, T.; Miyanishi, K.; Nobuoka, A.; Hayashi,
T.; Abe, T.; Kato, J.; Sakon, K.; Naniwa, Y.; Tanabe, H.; Niitsu, Y.
J. Pharmacol. Exp. Ther. 2003, 306, 861–869.
(34) Shiratori, Y.; Soma, Y.; Maruyama, H.; Sato, S.; Takano, A.; Sato,
K. Cancer Res. 1987, 47, 6806–6809.
(35) Kamada, K.; Goto, S.; Okunaga, T.; Ihara, Y.; Tsuji, K.; Kawai, Y.;
Uchida, K.; Osawa, T.; Matsuo, T.; Nagata, I.; Kondo, T. Free Radical
Biol. Med. 2004, 37, 1875–1884.
(36) Bernardini, S.; Bernassola, F.; Cortese, C.; Ballerini, S.; Melino, G.;
Motti, C.; Bellincampi, L.; Iori, R.; Federici, G. J. Cell. Biochem.
2000, 77, 645–653.
(37) Gilot, D.; Loyer, P.; Corlu, A.; Glaise, D.; Lagadic-Gossmann, D.;
Atfi, A.; Morel, F.; Ichijo, H.; Guguen-Guillouzo, C. J. Biol. Chem.
2002, 277, 49220–49229.
(38) Adler, V.; Yin, Z. M.; Fuchs, S. Y.; Benezra, M.; Rosario, L.; Tew,
K. D.; Pincus, M. R.; Sardana, M.; Henderson, C. J.; Wolf, C. R.;
Davis, R. J.; Ronai, Z. EMBO J. 1999, 18, 1321–1334.
(39) Tew, K. D. Biochem. Pharmacol. 2007, 73, 1257–1269.
(40) Lyttle, M. H.; Hocker, M. D.; Hui, H. C.; Caldwell, C. G.; Aaron,
D. T.; Engqvistgoldstein, A.; Flatgaard, J. E.; Bauer, K. E. J. Med.
Chem. 1994, 37, 189–194.
(41) Burg, D.; Mulder, G. J. Drug Metab. ReV. 2002, 34, 821–863.
(42) Ang, W. H.; De Luca, A.; Chapuis-Bernasconi, C.; Juillerat-Jeanneret,
L.; Lo Bello, M.; Dyson, P. J. ChemMedchem 2007, 2, 1799–1806.
9
14540 J. AM. CHEM. SOC. VOL. 130, NO. 44, 2008

Highly Sensitive Fluorogenic Substrates for GST
A R T I C L E S
n-hexane ) 1/1) to afford the title compound 29.6 mg (8% yield,
slight yellowish powder). 1H NMR (300 MHz, DMSO-d6) δ: 10.57
(1H, s), 8.73 (1H, dd, J ) 7.2, 2.3 Hz), 8.38 (1H, ddd, J ) 8.7,
4.2, 2.20 Hz), 7.73-7.80 (3H, m), 7.37 (2H, t, J ) 8.0 Hz), 7.13
(1H, t, J ) 7.4 Hz) 13C NMR (75 MHz, DMSO-d6) δ: 162.1, 138.3,
136.4, 135.4, 131.4, 128.4, 125.5, 123.9, 120.3, 118.5, 118.3 HRMS
(ESI-TOF) m/z Found 259.0506 [M - H]-, calculated 259.0519
(-1.26 mmu).
The residue was dissolved in 10 mL of CH2Cl2, and cooled to -5
°C. A CH2Cl2 (5 mL) solution of 5-aminofluorescein dipivaloyl
(336 mg, 0.652 mmol) and DIEA (260 µL, 1.56 mmol) was added
dropwise, and the reaction mixture was stirred for 4 h at room
temperature. The solvent was removed and the residue was
redissolved in 10 mL of MeOH, then 10 mL of aqueous 2 N NaOH
was added dropwise with stirring overnight at ambient temper-
ature. The reaction was quenched by the addition of 8 mL of
aqueous 2 N HCl. The residue was purified by means of
preparative reverse-phase HPLC to give the title compound, 27
4-Chloro-3-nitrobenzanilide (CNBA). To a solution of 4-chloro-
3-nitrobenzoyl chloride (327 mg, 1.49 mmol) in 8 mL of dioxane,
aniline (272 µL, 2.98 mmol) was added dropwise with stirring at
room temperature. After 10 min, a white residue appeared, and the
mixture was poured into AcOEt (100 mL). The organic solution
was washed with H2O (50 mL × 3) and brine, dried over anhydrous
Na2SO4, and filtered. The filtrate was evaporated to dryness. The
residue was purified on a silica gel column (AcOEt/n-hexane )
5/1) to afford the title compound, 377 mg (91% yield, slight
yellowish solid). 1H NMR (300 MHz, DMSO-d6) δ: 10.53 (1H, s,
br), 8.62 (1H, d, J ) 2.2 Hz), 8.25 (1H, dd, J ) 8.4, 2.2 Hz), 7.96
1
mg (7.2% yield, orange solid). H NMR (300 MHz, DMSO-d6)
δ: 11.10 (1H, s), 10.11 (1H, br), 8.78 (1H, d, J ) 1.7 Hz), 8.53
(1H, dd, J ) 8.4, 1.7 Hz), 8.43-8.47 (m, 2H), 8.07 (1H, dd, J
) 8.4, 1.8 Hz), 7.31 (1H, d, J ) 8.4 Hz), 6.67 (2H, d, J ) 2.2
Hz), 6.62 (2H, d, J ) 8.6 Hz), 6.55 (2H, dd, J ) 8.7, 2.3 Hz)
13C NMR (100 MHz, DMSO-d6) δ: 168.5, 162.4, 159.5, 151.9,
148.0, 143.5, 141.7, 140.0, 139.4, 134.0, 129.1, 127.6, 126.9,
126.2, 125.1, 124.7, 114.9, 112.6, 109.5, 102.2, 83.2 HRMS
(ESI-TOF) m/z Found 540.0790 [M - H]-, calculated 540.0679,
(+2.98 mmu).
(1H, d, J ) 8.4 Hz), 7.75 (2H, m), 7.37 (2H, m), 7.13 (2H, m) 13
C
DNAF2. 3,4-Dinitrobenzoic acid (23.5 mg, 0.11 mmol) was
suspended in 6 mL of thionyl chloride and catalytic amount (a few
drops, via a Pasteur pipet) of DMF was added. The mixture was
refluxed for 3 h. Toluene (100 mL) was added to the reaction
mixture, which was then evaporated in vacuo to dryness. The
residue was dissolved in 5 mL of CH2Cl2 and cooled to -5 °C. A
CH2Cl2 (2.5 mL) solution of 6-aminofluorescein dipivaloyl (53.5
mg, 0.10 mmol) and DIEA (50 µL, 0.30 mmol) was added
dropwise, and the reaction mixture was stirred for 4 h at room
temperature. The solvent was removed and the residue was
redissolved in 10 mL of MeOH, then 10 mL of a solution of NaOMe
(26 mg, 0.48 mmol) in MeOH was added dropwise with stirring
overnight at room temperature. The reaction was quenched with
8 mL of aqueous 2 N HCl. The residue obtained by evaporation
was purified by means of preparative reverse-phase HPLC to
NMR (75 MHz, DMSO-d6) δ: 162.4, 147.3, 138.5, 134.9, 132.8,
131.9, 128.7, 128.0, 124.9, 124.9, 124.2, 120.5 HRMS (ESI-TOF)
m/z Found 464.2350 [M - H]-, calculated 275.0220 for
C13H8ClN2O3 (-0.38 mmu). Anal. Calcd for C13H9ClN2O3: N,
10.13; C, 56.43; H, 3.28, found N, 10.14; C, 56.53; H, 3.42.
4-Ethylthio-3-nitrobenzanilide (TNBA). To a solution of
sodium ethanethiolate (54.7 mg, 0.65 mmol) in 5 mL of HMPA,
CNBA (183.6 mg, 0.66 mmol) in 5 mL of HMPA was added
dropwise. The mixture was stirred overnight at room temperature,
then poured into AcOEt (100 mL). The organic solution was
washed with H2O (100 mL × 3) and brine, dried over anhydrous
Na2SO4, and filtered. The filtrate was evaporated to dryness. The
residue was purified by a preparative TLC (AcOEt/n-hexane )
1/1) to afford the title compound 52.3 mg (26% yield, yellow
powder). 1H NMR (300 MHz, DMSO-d6) δ: 10.51 (1H, s), 8.80
(1H, d, J ) 2.0 Hz), 8.26 (1H, dd, J ) 8.5, 1.9 Hz), 7.75-7.79
(3H, m), 7.37 (2H, t, J ) 7.98 Hz), 7.12 (1H, t, J ) 7.4 Hz),
3.15 (2H, q, J ) 7.3 Hz), 1.31 (3H, t, J ) 7.3 Hz) 13C NMR
(75 MHz, DMSO-d6) δ: 162.9, 145.0, 140.8, 138.7, 132.7, 131.0,
128.7, 127.2, 125.1, 124.0, 120.5, 25.5, 12.8 HRMS (ESI-TOF)
m/z Found 301.0660 [M - H]-, calculated 301.0647 for
C15H13N2O3S (+1.27 mmu).
3,4-Dinitrobenzanilide (NNBA). To a solution of oxalyl chloride
(1.17 g, 9.2 mmol) in 15 mL of CH2Cl2, DMF (717 µL, 9.2 mmol)
was added. The mixture was stirred for 10 min at -20 °C, then a
CH2Cl2/THF (1:1) solution (20 mL) of 3,4-dinitrobenzoic acid (777
mg, 3.7 mmol) was added. Stirring was continued for 30 min, then
a solution of aniline (334 µL, 3.7 mmol) and DIEA (1.2 mL, 7.3
mmol) in 10 mL of THF was added at -20 °C. The mixture was
stirred for 1 h at room temperature, then after confirmation of
disappearance of the starting materials, evaporated to dryness. The
residue was poured into CH2Cl2 (100 mL). The organic solution
was washed with H2O (100 mL x 3) and brine, dried over
anhydrous Na2SO4, and filtered. The filtrate was evaporated to
dryness. The residue was purified on a silica gel column (CH2Cl2)
to afford the title compound, 299.6 mg (29% yield, slight
yellowish powder). 1H NMR (300 MHz, DMSO-d6) δ: 10.78
(1H, s, br), 8.80 (1H, d, J ) 1.7 Hz), 8.55 (1H, dd, J ) 8.3, 1.7
Hz), 8.47 (1H, d, J ) 8.3 Hz), 7.81 (2H, d, J ) 8.2 Hz),
7.39-7.49 (2H, m), 7.19-7.24 (1H, m) 13C NMR (75 MHz,
DMSO-d6) δ: 161.8, 143.3, 141.6, 139.8, 138.3, 133.8, 128.8,
126.0, 124.9, 124.5, 120.5 HRMS (ESI-TOF) m/z Found
286.0462 [M - H]-, calculated 286.0464 for C13H8N3O5 (-0.18
mmu).
1
give the title compound, 7.4 mg (13% yield, orange solid). H
NMR (300 MHz, DMSO-d6) δ: 11.03 (1H, s), 10.15 (2H, s),
8.68 (1H, d, J ) 1.6 Hz), 8.41 (1H, dd, J ) 8.4, 1. Six Hz),
8.37 (1H, d, J ) 8.4 Hz), 8.06 (1H, dd, J ) 8.4, 1.5 Hz), 8.02
(1H, d, J ) 8.4 Hz), 7.64 (1H, d, J ) 1.5 Hz), 6.68 (2H, d, J )
2.3 Hz), 6.64 (2H, d, J ) 8.7 Hz), 6.57 (2H, dd, J ) 8.7, 2.3
Hz). 13C NMR (100 MHz, DMSO-d6) δ: 168.2, 162.5, 159.5,
154.4, 151.7, 144.8, 143.4, 141.5, 139.1, 134.0, 129.1, 126.0,
125.8, 125.0, 121.7, 121.3, 113.9, 112.7, 109.4, 102.2, 82.3
HRMS (ESI-TOF) m/z Found 540.0656 [M - H]-, calculated
540.0679, (-2.27 mmu).
NBDHEX. NBDHEX was synthesized from NBD-Cl and 6-
mercapto-1-hexanol according to the literature43 with minor
1
modifications. H NMR (300 MHz, CDCl3) δ: 8.41 (1H, d, J )
7.9 Hz, a), 7.15 (1H, d, J ) 8.1 Hz, b), 3.68 (2H, t, J ) 6.3 Hz, h),
3.28 (2H, t, J ) 7.3 Hz, c), 1.44-1.93 (8H, m, d-h) 13C NMR (75
MHz, CDCl3) δ 149.2, 142.5, 142.0, 132.5, 130.7, 120.2, 62.7, 32.4,
31.7, 28.6, 27.9, 25.3 HRMS (ESI+) Calculated for C12H15N3NaO4S
(M + Na+); 320.0681, found; 320.06557 (-2.52 mmu).
Synthesis of DNAT-Me. 2-Methyl-4-nitrobenzyl Alcohol (1).
1 was synthesized from 2-methyl-4-nitrobenzoic acid according to
the literature.44 1H NMR (300 MHz, DMSO-d6) δ: 8.07 (1H, dd,
J ) 8.4, 2.3 Hz), 8.02 (1H, d, J ) 2.3 Hz), 7.62 (1H, d, J ) 8.4
Hz), 4.80 (2H, d, J ) 5.4 Hz), 2.39 (3H, s,), 1.94 (1H, t, J ) 5.4
Hz) 13C NMR (100 MHz, CDCl3) δ 147.1, 146.0, 136.9, 127.1,
121.2, 62.4, 18.6. LRMS (EI+); 167.
2-Methyl-4-nitrobenzaldehyde (2). To a stirred CH2Cl2 solution
(200 mL) containing pyridinium chlorochromate 3.4 g (15.9 mmol)
and anhydrous magnesium sulfate 3.5 g, 1.41 g of 1 (8.4 mmol)
DNAF1. 3,4-Dinitrobenzoic acid (148 mg, 0.70 mmol) was
suspended in 10 mL of thionyl chloride, and a catalytic amount (a
few drops, via a Pasteur pipet) of DMF was added. The mixture
was refluxed for 2.5 h. An excess of toluene was added to the
reaction mixture, which was then evaporated in vacuo to dryness.
(43) Ricci, G.; De Maria, F.; Antonini, G.; Turella, P.; Bullo, A.; Stella,
L.; Filomeni, G.; Federici, G.; Caccuri, A. M. J. Biol. Chem. 2005,
280, 26397–26405.
(44) Hay, M. P.; Sykes, B. M.; Denny, W. A.; O’Connor, C. J. J. Chem.
Soc., Perkin Trans. 1 1999, 2759–2770.
9
J. AM. CHEM. SOC. VOL. 130, NO. 44, 2008 14541

A R T I C L E S
Fujikawa et al.
was added. The mixture was stirred for 6 h at ambient temperature.
The organic solution was washed with water (100 mL × 3) and
brine, dried over anhydrous sodium sulfate, and filtered. The filtrate
was evaporated to dryness. The residue was purified on a silica gel
column (Silicagel 60N, CH2Cl2/n-hexane ) 4/1) to afford the title
compound 1.15 g (83% yield, slight yellowish solid) 1H NMR (300
MHz, CDCl3) δ: 10.40 (1H, s), 8.19 (1H, dd, J ) 8.3, 2.2 Hz),
8.15 (1H, d, J ) 2.2 Hz), 7.99 (1H, d, J ) 8.3 Hz), 2.80 (3H, s)
13C NMR (75 MHz, CDCl3) δ: 190.9, 150.2, 142.2, 137.9, 132.4,
126.6, 121.3, 19.5 LRMS (EI+); 165.
added. The reaction mixture was stirred at room temperature for
2 h, then acidified with 2 N HCl to pH 3-4. MeOH was evaporated,
and the target compound was extracted with AcOEt (20 mL × 3).
The organic layer was washed with brine, dried over anhydrous
Na2SO4, and filtered. The mixture was lyophilized and purified by
HPLC (buffer A; 100 mM triethylamine acetate (TEAA) buffer
(pH 7.0), B; 80% CH3CN/20% TEAA; eluted by running a gradient
from 20-100% buffer B at an elution rate of 5 mL/min) to give a
solution containing the title compound. This solution was acidified
with AcOH and extracted with AcOEt (50 mL). The AcOEt solution
was washed with brine, dried over anhydrous Na2SO4, and filtered.
The filtrate was evaporated to give the title compound 2.4 mg (9%
yield, red orange solid). 1H NMR (400 MHz, DMSO-d6) δ: 11.03
(1H, s), 8.86 (1H, dd, J ) 1.5 Hz), 8.61 (1H, dd, J ) 8.3, 1.5 Hz),
8.52 (1H, d, J ) 8.3 Hz), 7.97 (1H, d, J ) 2.0 Hz), 7.93 (2H, dd,
J ) 8.3, 2.0 Hz), 7.38 (1H, d, J ) 8.3 Hz), 6.97 (2H, d, J ) 9.3
Hz), 6.80-6.53 (4H, m), 2.11 (3H, s) 13C NMR (100 MHz, DMSO-
d6) δ 171.6, 163.1, 158.6, 152.9, 151.2, 149.2, 148.9, 145.9, 143.4,
139.6, 139.3, 137.7, 135.6, 134.5, 131.3, 127.5, 109.6, 29.0 HRMS
(ESI-) Calculated for C27H16N3O8 [M - H]-; 510.0938, found;
510.0940 (+0.29 mmu).
4,4′-[(2-Methyl-4-nitrophenyl)methylene]bis(benzene-1,3-di-
ol) (3). Resorcinol (1.4 g, 12.6 mmol) and 2-methoxy-4-nitroben-
zaldehyde (2) (214.8 mg, 1.3 mmol) were dissolved in CH2Cl2/
ether (1:1, 20 mL) and stirred at 4 °C. Methanesulfonic acid (2.0
mL) was added and the reaction mixture was stirred at ambient
temperature for 15 min. The solution was evaporated and redis-
solved in ethyl acetate (100 mL). The organic solution was washed
with water (100 mL × 3) and brine, dried over anhydrous Na2SO4,
and filtered. The filtrate was evaporated to dryness. The residue
was purified by a silica gel chromatography (Silicagel 60N, CH2Cl2/
MeOH, 9:1) to give the title compound, 189 mg (40% yield,
1
Expression and Purification of GSTs. cDNA clones of human
GSTA1 and GSTM1 were purchased from ATCC. pET-21a(+) and
pET-21b(+) plasmid DNAs were purchased from Novagen. Cloning
of pET-GST was conducted according to general cloning proce-
dures. Overexpression of human GSTs was achieved by transform-
ing Escherichia coli BL21 (DE3) cells with pET-21b(+)-hGSTA1,
pET-21a(+)-T7-hGSTM1, and pET-21b(+)-hGSTP1, respectively.
A single colony was inoculate into 5 mL of 2 × YT supplemented
with ampicillin (100 mg/mL) and grown overnight at 37 °C with
shaking. This overnight seed culture was then used to inoculate 1
L of fresh growth medium and grown at 37 °C until induction with
isopropyl thio-ꢀ-D-galactoside (IPTG) (1.0 mM final concentration).
After agitation for 3 h at 37 °C, the cells were collected by
centrifugation. The cells were suspended in working solution (50
mM phosphate, 300 mM NaCl) and sonicated. (10 pulses of 20s,
on ice). Cell debris was removed by centrifugation at 10 000 rpm
for 60 min at 4 °C, after which the cell lysate supernatant was
transferred to swollen glutathione agarose, which was stirred slowly
at 4 °C. The mixture was washed with buffer A (3 mM EDTA, 20
mM potassium phosphate (pH 7.0), and 3 mM mercaptoethanol)
and centrifuged (1000 rpm, 3 min), and these processes were
repeated ten times. GSTs were eluted with GSH solution (10 mM
GSH, 50 mM Tris (hydroxymethyl) aminomethane, 0.4 M NaCl).
The eluate was dialysed against buffer A overnight, then buffer B
(3 mM EDTA, 20 mM potassium phosphate (pH 7.0), 3 mM
2-mercaptoethanol, 20% glycerol) for 6 h. Purified GSTs were
checked by SDS-PAGE (acrylamide 12.5%).
GST Specific Activity Measurement. GST assays with CDNB
as a substrate were performed on a C-550 UV/vis spectrophotometer
equipped with a stirrer controller and water bath (EYELA NCB-
120) maintained at 25 °C. The standard assay mixture contained
1.0 mM GSH, 1.0 mM CDNB (2% EtOH as a cosolvent), and 0.1
mM EDTA in 3 mL of 100 mM potassium phosphate buffer, pH
6.5. The activity was determined spectrophotometrically at 340 nm
(ꢀ ) 9.6 mM-1 cm-1) within 20 s after substrate addition. Reported
enzymatic specific activities are the averages of three independent
determinations at three different concentrations of enzyme in the
range where concentration-initial rate proportionality was main-
tained. Assays with DNAFs as a substrate were performed at 37
°C. The standard assay mixture contained 1.0 mM GSH, 0.5 µM
DNAF1 or 2 (0.1% DMSO as a cosolvent), and phosphate-buffered
saline (PBS, pH 7.4). The activity was determined spectrophoto-
metrically at 514 nm with excitation at 490 nm. The rate of product
formation was estimated by use of the equation, Initial rate ) P ×
(Ft - F0)/{t × (FMax - F0)}, where Ft and F0 represent the
fluorescence at times t and 0 min, P is the amount of product of a
known concentration in solution, and FMax is the fluorescence due
to P of product. Reported enzymatic specific activities are the
yellowish solid). H NMR (300 MHz, CD3OD) δ: 7.93 (1H, d, J
) 2.3 Hz), 7.84 (1H, dd, J ) 8.5, 2.3 Hz), 6.95 (1H, d, J ) 8.5
Hz), 6.39 (2H, d, J ) 8.3 Hz), 6.26 (2H, d, J ) 2.4 Hz), 6.13 (2H,
dd, J ) 8.3, 2.4 Hz), 5.99 (1H, s), 2.23 (3H, s) 13C NMR (100
MHz, CD3OD) δ: 158.0, 157.1, 154.0, 147.2, 140.0, 131.4, 130.0,
125.4, 121.4, 121.2, 107.0, 103.5, 40.8, 19.6 HRMS (ESI-TOF)
Calculated for C20H16NO6 [M-H]-; 366.0978, found; 366.0962
(-1.55 mmu).
6-Hydroxy-9-(2-methyl-4-aminophenyl)-3H-xanthen-3-one (2-
Me-4-NH2-TokyoGreen) (5). Compound 3 (106 mg, 0.29 mmol)
and p-toluenesulfonic acid monohydrate (436 mg, 2.3 mmol) were
suspended in 10 mL of dry toluene and heated to reflux (125 °C)
for 14 h. After cooling to room temperature, the solution was
neutralized with saturated sodium bicarbonate to pH ) 7. The
solution was acidified by dropwise addition of concentrated HCl
to pH ) 3-4. The precipitate was extracted into ethyl acetate (3
× 100 mL) and the solution was washed with brine (25 mL × 2).
The combined organic extracts were dried over anhydrous sodium
sulfate and the solvent was removed in vacuo. Silica gel column
chromatography (Silicagel 60 (Wako), CH2Cl2/ MeOH, 9:1) af-
forded 6-hydroxy-9-(2-methyl-4-nitrophenyl)-3H-xanthen-3-one (4)
(19.6 mg, crude). The identity of the product was confirmed by
HPLC, and ESI-Tof-MS. A mixture of 92.8 mg of crude 4,
Na2S·9H2O (598 mg, 2.5 mmol), and NaSH ·nH2O (440 mg) was
heated in H2O (50 mL) for 3 h under reflux. After cooling, the
mixture was acidified with AcOH and the resulting dark red
precipitate was collected, washed with cooled water and dried in
Vacuo, affording 28.4 mg of the title compound (3.0% yield in two
1
steps, dark red solid). H NMR (300 MHz, CD3OD) δ: 7.66 (2H,
d, J ) 9.2 Hz), 7.26 (2H, d, J ) 2.3 Hz), 7.18 (2H, d, J ) 9.2, 2.3
Hz), 7.12 (1H, d, J ) 8.3 Hz), 6.93-6.95 (1H, m), 6.90 (1H, dd,
J ) 8.3, 2.2 Hz), 1.99 (3H, s) 13C NMR (100 MHz, CD3OD) δ:
173.2, 166.9, 160.7, 150.2, 139.0, 134.7, 132.4, 122.5, 121.4, 118.6,
118.2, 114.3, 103.6, 20.2 HRMS (ESI+) Calculated for C20H16NO3
[M + H]+; 318.1130, found; 318.1113 (-1.74 mmu).
DNAT-Me. A suspension of 3,4-dinitrobenzoic acid (42.1 mg,
0.20 mmol) in 5 mL of thionyl chloride was refluxed at 80 °C for
3 h. Then, 80 mL of toluene was added to the reaction mixture
and evaporated in Vacuo to dryness three times. The residue was
dissolved to 4 mL of CH2Cl2, and cooled to 4 °C. A THF solution
(2.5 mL) of 2-Me-4-NH2-TokyoGreen (5) 17.3 mg (0.055 mmol)
and DIEA (75 µL) was added dropwise, and the reaction mixture
was stirred under an Ar atmosphere for 1.5 h at room temperature.
The solvent was evaporated to dryness. The residue was dissolved
in AcOEt and the resulting solution was washed with water and
brine, dried over anhydrous Na2SO4, and filtered. The filtrate was
evaporated to dryness. The residue was redissolved in 6 mL of
methanol, and 10 mL of water (containing 1% triethylamine) was
9
14542 J. AM. CHEM. SOC. VOL. 130, NO. 44, 2008

Highly Sensitive Fluorogenic Substrates for GST
A R T I C L E S
averages of three independent determinations at three different
concentrations of enzyme in the range where concentration-initial
rate proportionality was maintained. The determined values of
specific activity are shown at Table 2.
Kinetic Parameter Measurement. NNBA. GST assays were
performed at 37 °C. NNBA in the concentration range of 3-100
mM (each with 1% DMSO as a cosolvent), 1.0 mM GSH, and
0.241 µg/mL hGSTP1 were used. The activity was determined
spectrophotometrically at 381 nm (product absorption, ꢀ ) 3 mM-1
cm-1) based on the reported method.45 Michaelis-Menten (MM)
plots were approximated with KaleidaGraph (Synergy Software,
Reading, PA). From the resulting MM plots, kcat and KM were
obtained. Second-order rate constants were calculated from the
results in the absence of GST.
mBBr. GST assay was performed at 37 °C. mBBr in the
concentration ranging 2.5-25 µM (each with 1% EtOH as a
cosolvent), 1.0 mM GSH, and an appropriate concentration of GSTs
were used in PBS (pH 7.4). The activity was determined spectro-
photometrically at 470 nm (excited at 400 nm) within 100 s after
addition of the substrate.
DNAFs. GST assays were performed at 37 °C. Assay conditions
were the same as described for GST specific activity measurement
except that initial rate values were derived from the averages of
three different DNAF concentrations in the range of 0.5-1.5 µM.
Because quenching of fluorescence owing to the inner filter effect
occurs at high concentrations of fluorogenic substrate, we calculated
approximate kcat/KM values directly. The apparent kcat/KM was
derived using the MM relationship at low DNAF concentrations
(0.5-1.5 µM): V/([E]*[S]) ) kcat/KM, when KM > [S]. Both
substrates displayed linear rate vs substrate concentration behavior
in this concentration region.
cultured in Dulbecco’s modified Eagle’s medium (Invitrogen)
containing 10% fetal bovine serum and penicillin/streptomycin.
All cell lines were maintained at 37 °C in an atmosphere of air
containing 5% CO2.
Microscopic Imaging. For fluorescence imaging of GST activity
in living cells, DNAT-Me was used at a final concentration of 2
µM in Hanks’ buffered salt solution (HBSS), pH 7.4. Imaging
experiments with HuCCT1 cells were conducted at two days after
seeding. Cells were washed with PBS twice and incubated in HBSS,
containing the vehicle or 5 µM CCCP or 50 µM N-ethylmaleimide
(NEM) (0.1% DMSO as a cosolvent) for 15 min at 37 °C. Then,
the medium was replaced with new HBSS solution containing 2
µM DNAT-Me (0.3% DMSO as a cosolvent) in the control
experiment. In the case of CCCP or NEM treatment, the medium
was replaced with new HBSS solution containing 2 µM DNAT-
Me and 5 µM CCCP or 50 µM NEM (0.3% DMSO as a cosolvent)
and incubated for a further 15 min at 37 °C. (In the confocal overlay
experiment with SYTO, HBSS containing 2 µM DNAT-Me, 5 µM
CCCP, and 1 µM SYTO 60-red (0.4% DMSO as a cosolvent) was
used.) Fluorescence images were captured using an Olympus IX
71 with a cooled CCD camera (Coolsnap HQ, Olympus), a xenon
lamp (AH2RX-T, Olympus) with a BP-470-490 excitation filter,
and a BA 510-550 emission filter. In the confocal microscopic
experiment, fluorescence images were captured using an Olympus
Fluoview system withg an IX81 inverted microscope, an argon ion
laser and HeNe laser, and a PlanApo 40×/1.00 objective lens
(Olympus). The excitation wavelength was 488 nm, and the
emission was detected via an FITC filter for DNAT-Me. For SYTO
60-red, the excitation wavelength was 633 nm and the emission
was detected via a Cy-5 filter.
Acknowledgment. We thank our colleagues Suguru Kemmoku,
Tasuku Ueno, Takuji Shoda, and Akira Yastushige for helpful
discussions, Eri Ikeda and Yosuke Kaku for their kind assistance
with molecular cloning technique, and Nae Saito for UPLC-MS
analysis. This study was supported in part by research grants (Grant
Nos. 19021010 and 19205021) to Y.U., and a grant for the
Advanced and Innovational Research Program in Life Sciences to
T.N. from the Ministry of Education, Culture, Sports, Science and
Technology of the Japanese Government, and grants from Hoansha
Foundation to T.N.
Measurement of Released Nitrite. Nitrite released from NNBA
was measured with a NO2/NO3 Assay Kit-C II (Colorimetric) Griess
Reagent Kit (Dojindo, Japan) according to the manufacturer’s
instructions. Briefly, the kinetic measurement assays were done with
NNBA in the concentration range of 6.25-100 µM. In the plateau
phase of the reaction, the reaction mixtures were transferred to a
96-well plate. Using known concentrations of nitrite as a standard,
the concentration of released nitrite was determined with a SH-
8000 (Laboratory) Micro plate reader (Corona Electric Co., Ltd.,
Tsukuba, Japan).
Cells Lines and Culture Conditions. Human cholangiocarci-
noma cell line HuCCT1 and a cell line derived from normal lung,
OUS11, were purchased from the Health Science Research Re-
sources Bank (Osaka, Japan). Human leukemia cell line HL-60 and
human fibrosarcoma cell line HT1080 were purchased from RIKEN
Bioresource Center cell bank (Tsukuba, Japan). HuCCT1 and HL-
60 were cultured in RPMI 1640 medium (Sigma) containing 10%
fetal bovine serum (Invitrogen) and penicillin/streptomycin, and
OUS 11 was cultured in minimum essential medium (Invitrogen)
containing 10% fetal calf serum (Invitrogen) and penicillin/
streptomycin. Human cervical adenocarcinoma cell line HeLa was
Supporting Information Available: Specific activities toward
CDNB and DNAFs; molecular structures of evaluated substrates
for GST; 1H NMR chart of the NNBA before and after
glutathionylation; spectra of DNAFs and DNAT-Me before and
after glutathionylation; fluorescence timecourse of reactions
between DNAF1 and various thiol compounds and lysine
derivatives; inhibitor evaluation; HPLC chromatograms of
DNAT-Me loaded cell lysate; and additional microscopic
image.This material is available free of charge via the Internet
at http://pubs.acs.org.
(45) Morgenstern, R.; Lundqvist, G.; Hancock, V.; Depierre, W. J. Biol.
Chem. 1988, 263, 6671–6675.
JA802423N
9
J. AM. CHEM. SOC. VOL. 130, NO. 44, 2008 14543