Glutathionyl Transferase Catalyzed Addition of Glutathione to COMC: A New Hypothesis for Antitumor Activity

Article abstract of DOI:10.1021/ol025650h

matrix presented Data are presented indicating that the potent antitumor activity of 2-crotonyloxymethyl-(4R,5R,6R)-4,5,6-trihydroxy-2-cyclohexenone (COTC) and 2-crotonyloxymethyl-2-cyclohexenone (COMC) is not likely the result of glyoxalase I inhibition,

Full text of DOI:10.1021/ol025650h

ORGANIC
LETTERS
2002
Vol. 4, No. 7
1209-1212
Glutathionyl Transferase Catalyzed
Addition of Glutathione to COMC: A
New Hypothesis for Antitumor Activity
,‡
,†
Diana S. Hamilton,† Zhebo Ding,‡ Bruce Ganem,* and Donald J. Creighton*
Department of Chemistry and Biochemistry, UniVersity of Maryland, Baltimore
County, Baltimore, Maryland 21228, and Department of Chemistry and Chemical
Biology, Baker Laboratory, Cornell UniVersity, Ithaca, New York 14853-1301
bg18@cornell.edu
Received January 31, 2002
ABSTRACT
Data are presented indicating that the potent antitumor activity of 2-crotonyloxymethyl-(4R,5R,6R)-4,5,6-trihydroxy-2-cyclohexenone (COTC)
and 2-crotonyloxymethyl-2-cyclohexenone (COMC) is not likely the result of glyoxalase I inhibition, as has long been assumed. An alternative
hypothesis is presented, based on the finding that COMC is a substrate for human glutathionyl transferase, which produces a transient, highly
electrophilic glutathionylated 2-exomethylenecyclohexanone that can covalently modify proteins and nucleic acids.
The Streptomyces metabolite 2-crotonyloxymethyl-(4R,5R,6R)-
4,5,6-trihydroxy-2-cyclohexenone (COTC, 1a, Scheme 1)1,2
Compound 1a is somewhat more active (IC50 ) 0.5-19 µM)
than 1b (IC50 ) 3-44 µM).3 In 1975, Takeuchi et al.
attributed the antitumor activity of 1a to its putative
glutathione (GSH) adduct 2a, which was proposed to inhibit
the enzyme glyoxalase I (GlxI).1
Scheme 1
Glyoxalase I plays a pivotal role in detoxifying intracellular
methylglyoxal, which is formed during normal carbohydrate
metabolism.4 Indeed, certain inhibitors of human GlxI, the
most potent of which are enediol transition state analogue
inhibitors, have been shown to retard the growth of both
murine and human tumors in culture and in tumor-bearing
mice by causing the accumulation of intracellular methyl-
glyoxal.5,6
While a priori quite plausible, the “prodrug” hypothesis
of Takeuchi et al. for the antitumor action of COTC and
and its simpler synthetic analogue 2-crotonyloxymethyl-2-
cyclohexenone (COMC, 1b)3 both exhibit potent antitumor
activity against murine and human tumors in culture.
(3) Aghil, O.; Bibby, M. C.; Carrington, S. J.; Double, J.; Douglas, K.
T.; Phillips, R. M.; Shing, T. K. M. Anti-Cancer Drug Des. 1992, 7, 67-
82.
(4) Creighton, D. J.; Pourmotabbed, T. In Molecular Structure and
Energetics: Principles of Enzyme ActiVity; Liebman, J. F., Greenberg, A.,
Eds.; VCH Publishers: New York 1988; Vol. 9, pp 353-386.
(5) Kavarana, M. J.; Kovaleva, E. G.; Creighton, D. J.; Wollman, M.
B.; Eiseman, J. L. J. Med. Chem. 1999, 42, 221-228.
(6) Sharkey, E. M.; O’Neill, H. B.; Kavarana, M. J.; Wang, H.; Creighton,
D. J.; Sentz, D. L.; Eiseman, J. L. Cancer Chemother. Pharmacol. 2000,
46, 156-166.
† University of Maryland.
‡ Cornell University.
(1) Takeuchi, T.; Chimura, H.; Hamada, M.; Umezawa, H.; Yoshka, H.;
Oguchi, N.; Takahashi, Y.; Matsuda, A. J. Antibiot. 1975, 28, 737-742.
(2) Chimura, H.; Nakamura, H.; Takita, T.; Takeuchi, T.; Umezawa, M.;
Kato, K.; Saito, S.; Tomisawa, T.; Iitaka, Y. J. Antibiot. 1975, 28, 743-
748.
10.1021/ol025650h CCC: $22.00 © 2002 American Chemical Society
Published on Web 03/03/2002

COMC has never been tested. Recently, an authentic sample
of 2a was prepared by independent synthesis and shown to
be a modest inhibitor of GlxI.7
that 2a competitively inhibited human erythrocyte GlxI, with
a dissociation constant Ki ) 183 ( 6 µM.7
We have now prepared and tested 2b against human
erythrocyte GlxI and report that, like 2a, it is several orders
of magnitude weaker than other, mechanism-based, GlxI
inhibitors that display potent antitumor activity.5,6 To estab-
lish an alternative mode of tumor toxicity for COTC and
COMC, we have developed a short and efficient synthesis
of 1b and have now demonstrated that it is a substrate for
glutathionyl transferase, an enzyme that is widely distributed
in mammalian tissue.
Here we show that glutathionyl transferase catalyzes a
conjugate addition of GSH to 1b with concomitant elimina-
tion of crotonic acid leading to 3-glutathionyl-2-exomethyl-
enecyclohexanone 3 (Scheme 2), a reactive intermediate that
Scheme 2
Figure 1. Reciprocal plot of the velocity of the glyoxalase I
reaction (∆OD240) versus the concentration of GSH-methylglyoxal-
thiohemiacetal (S) in the absence and presence of different
concentrations of 2b. In each kinetic run, the concentration of free
GSH was maintained at 0.2 mM by varying the total concentration
of GSH and methylglyoxal on the basis of the dissociation constant
of the hemithioacetal (Kdiss ) 2.2 mM).9 Conditions: 50 mM
phosphate buffer, pH 7.0, 25 °C.
has now been isolated and characterized for the first time.
Trapping studies with amino acids and nucleotides support
the hypothesis that 3 can react with and covalently modify
reactive groups on proteins and DNA that may be critical to
cell viability.
Scheme 3 depicts a simple and efficient synthesis of 1b
that is superior to the earlier reported seven-step procedure.3
Baylis-Hillman reaction of 2-cyclohexenone 4 with form-
aldehyde afforded 2-hydroxymethyl-2-cyclohexenone 5,8
which was then crotonylated to 1b following the literature
procedure.3
To investigate the mechanism of formation of 2b, the
nonenzymatic reaction of 1b with GSH was monitored
spectrophotometrically and followed a simple first-order
decay (Figure 2, trace A). No intermediate species was
detectable. However, in the presence of human placental
glutathione transferase (GSTP1-1)10 the reaction rate profile
Scheme 3
(a) DMAP, CH2O, THF, rt 65%; (b) crotonic anhydride, pyridine,
DMAP, rt 92%.
The GSH conjugate 2b was prepared from 1b by adapting
the procedure for making 2a from 1a.7 The NMR spectrum
of 2b featured the expected glutathionyl resonances and the
downfield resonance (δ 7.12, triplet) characteristic of H3 in
â,γ-unsubstituted 2-cyclohexenones.
Kinetic studies with human erythrocyte GlxI (sodium
phosphate buffer, pH 7.0, 25 °C) indicated that 2b was a
competitive inhibitor of the enzyme, with a dissociation
constant Ki ) 107 ( 1 µM. We have previously reported
Figure 2. Spectrophotometrically determined rates of reaction of
GSH (1.03 mM) with 1b (0.05 mM) (A) in the absence of GSTP1-1
(k ) 0.070 ( 0.0002 min-1); (B) in the presence of 1.8 units of
GSTP1-1 (k1 ) 0.882 ( 0.055 min-1; k2 ) 0.633 ( 0.053 min-1);
and (C) in the presence of 2.4 units of GSTP1-1 (k1 ) 1.33 (
0.037 min-1; k2 ) 0.685 ( 0.024 min-1). Conditions: 100 mM
phosphate buffer, 0.05 mM EDTA, pH 6.5, 25 °C.
1210
Org. Lett., Vol. 4, No. 7, 2002

conformed to a double exponential decay, composed of a
rapid, enzyme-dependent, initial phase involving 1b followed
by a slower enzyme-independent first-order phase (Figure
2, traces B and C).
This finding was consistent with the mechanism shown
in Scheme 4, whereby an initial, enzyme-catalyzed Michael
That finding indicated that 3 dissociated from the enzyme
prior to reacting with the free thiols in solution and was
consistent with the mechanism in Scheme 4. Adduct 2b (0.5
mM) was stable under the reaction conditions, undergoing
less than 1% conversion to 4 in the presence of cysteine (2.5
mM) over 27 h, as determined by HPLC.
Initial rates of reaction of 1a and 1b with GSTP1-1 were
determined from reciprocal plots of initial velocities (∆OD235
/
min), versus [substrate] in buffered solution at pH 6.5, with
[GSH] ) 1 mM (25 °C). Under those conditions, the
enzyme-catalyzed GSH addition became rate determining
(<0.01 units of transferase in the assay cuvettes). For 1b,
kcat ) 1.2 ( 0.2 s-1 and Km ) 52 ( 10 µM. For 1a, the
individual kinetic constants could not be accurately deter-
mined, although kcat/Km was estimated to be 8.3-fold lower
than that of 1b.
Scheme 4
Brief incubation of a mixture of 1b, GSH, and GSTP1-1
gave rise to a transient intermediate that could be isolated
by reverse-phase HPLC, with a retention time close to that
of synthetic 2b. The kinetic properties of this species suggest
that it is an obligatory intermediate associated with the
enzymatic conversion of 1b to 2b. When this species is
combined with cysteine in buffered solution (pH 7), a new
species is produced that comigrates with authentic 4.
Moreover, incubation of this intermediate species with excess
GSH (1.04 mM) in potassium phosphate buffer (0.1 M, pH
6.5) at 25 °C results in a first order increase in OD235 with
a rate constant of 0.87 ( 0.08 min-1 similar in magnitude
to that associated with the second phase of the transferase-
catalyzed conversion of 1b to 2b, Figure 2. The absorptivity
of the intermediate (ꢀ235 ) 4300 cm-1 M-1) is significantly
less than that of the final product 2b (ꢀ235 ) 7500 cm-1 M-1),
which accounts for the overall shape of the reaction-rate
profile in the presence of transferase (Figure 2).
addition of GSH to 1b afforded the exocyclic enone 3. Once
dissociated from the enzyme, free 3 reacted with GSH
nonenzymatically to form 2b.
When 1b (0.1 mM) was incubated with cysteine (0.5 mM)
and GSH (0.5 mM) in the presence of GSTP1-1 (1.5 units)
for 30 min and the reaction mixture was fractionated by
reverse-phase HPLC, thiol adducts 2b and 4 were isolated.
On the basis of the integrated intensities of the well-resolved
peaks corresponding to 4 (∼9.5 min) and 2b (∼11.5 min),
the product ratio in the presence of GSTP1-1 was identical
to that obtained from a nonenzymatic incubation of 1b with
the same concentrations of cysteine and GSH (Figure 3).
1
The 600-MHz NMR spectrum of the intermediate was
consistent with the structure of 3 (Figure 4). The vinyl proton
resonances at 5.76 and 5.29 ppm were characteristic of
Figure 3. Fractionation of product mixtures in the absence and
presence of GSTP1-1 by reverse-phase HPLC (Waters, µ-Bondapak
C18, 0.78 × 30 cm).
1
Figure 4. The 600-MHz H NMR spectrum of intermediate 3.
Org. Lett., Vol. 4, No. 7, 2002
1211

geminal vinylic hydrogens and consistent with published
NMR spectra of several closely related 2-methylenecyclo-
hexanones.11 Other resonances in the spectrum corresponded
to those expected for the tripeptide moiety.12 The resonance
at δ 7.12 indicated the presence of 2b, which was formed in
the transferase-independent addition of GSH to 3.
Besides providing insight into the mechanism of addition
of GSH to COMC, the formation of 3 enzymatically from
1b also offers an attractive alternative biological mechanism
for the tumoricidal activity of both COTC and COMC. It
may now be hypothesized that 1a and 1b are enzyme-
activated prodrugs in which the crotonate ester serves as a
leaving group, in a process triggered by glutathionyl trans-
ferase. Methylenecyclohexanones such as 3 are known to
be highly reactive Michael acceptors11 and can function as
carcinostatic agents by reacting with proteins and nucleic
acids critical to cell function. Besides the thiol adducts
documented here, covalent adducts between 1b and model
polynucleic acids in the presence of GSTP1-1 have been
detected by mass spectrometry.13
Comparative data analysis suggested that the potent
antitumor activities of COTC (1a) and COMC (1b) could
not be rationalized by the biological activity of 2a and 2b,
respectively, which are weak competitive inhibitors of human
erythrocyte GlxI. Earlier in vitro studies on enediol analogue
inhibitors of GlxI that inhibited the growth of L1210 and
B16 melanotic melanoma as ester prodrugs revealed Ki values
against GlxI in the submicromolar range.6 In fact, a direct
correlation was observed between the Ki values of the
prodrugs against GlxI and their corresponding antitumor
activities, as measured by IC50 values. For example, the
weakest GlxI inhibitor in that study (Ki ) 0.16 µM) displayed
an IC50 value >100 µM. On the basis of that correlation,
the potent antitumor activity of 1b (IC50 ) 0.5-19 µM) is
inconsistent with the hypothesis of Takeuchi et al. attributing
antitumor activity to the action of 2b as GlxI inhibitor (Ki
) 107 µM).1
Acknowledgment. This work was supported in part by
grants from the NIH (GM 24054 to B.G.; CA 59612 to
D.J.C.) and the U.S. Army Medical Research and Material
Command (to D.J.C.). Support of the Cornell NMR Facility
has been provided by NSF and NIH.
Note Added after Print Publication: Due to a production
error, certain µM concentrations appeared as mM concentra-
tions in the version published on the Web 03/03/2002
(ASAP) and in the April 4, 2002 issue (Vol. 4, No. 7, pp
1209-1212); the correct electronic version of the paper was
published on 04/23/2002 and an Addition and Correction
appears in the May 16, 2002 issue (Vol. 4, No. 10).
(7) Huntley, C. F.; Hamilton, D. S.; Creighton, D. J.; Ganem, B. Org.
Lett. 2000, 2, 3143-3144.
(8) Rezgui, F.; El Gaied, M. M. Tetrahedron Lett. 1998, 39, 5965-5966.
(9) Hamilton, D. S.; Creighton, D. J. J. Biol. Chem. 1992, 267, 24933-
24936.
(10) Predominantly the pi or P1-1 isoform, purchased from Sigma
Chemical Company. Salts and free GSH were removed by ultrafiltration.
Units of transferase activity were determined using 1-chloro-2,4-dinitroben-
zene as substrate: Mannervik, B.; Danielson, U. H. Crit. ReV. Biochem.
1988, 23, 283-337.
Supporting Information Available: Experimental pro-
1
cedures and H and 13C NMR data for the synthesis of 1b.
This material is available free of charge via the Internet at
http://pubs.acs.org.
(11) Tamura, R., Watabe, K., Ono, N., Yamamoto, Y. J. Org. Chem.
1992, 57, 4895-4903.
OL025650H
(12) Rabenstein, D. L.; Keire, D. A. In Coenzymes and Cofactors:
Glutathione; Dolphin, D., Poulson, R., Avramovic, O., Eds,; John Wiley:
New York, 1989; Vol. 3, Part A, pp 67-101.
(13) Fabris, D.; Creighton D. J.; Ganem, B. Unpublished work.
1212
Org. Lett., Vol. 4, No. 7, 2002