Discovery of glutathione S-transferase inhibitors using dynamic combinatorial chemistry

Article abstract of DOI:10.1021/ja058049y

Protein-directed dynamic combinatorial chemistry (DCC) relies on reversible chemical reactions that can function under the near-physiological conditions required by the biological target. Few classes of reaction have so far proven effective at generating

Full text of DOI:10.1021/ja058049y

Published on Web 06/09/2006
Discovery of Glutathione S-Transferase Inhibitors Using
Dynamic Combinatorial Chemistry
Baolu Shi, Ross Stevenson, Dominic J. Campopiano,* and Michael F. Greaney*
Contribution from the School of Chemistry, UniVersity of Edinburgh, The King’s Buildings,
West Mains Road, Edinburgh EH9 3JJ, United Kingdom
Received November 27, 2005; E-mail: Michael.Greaney@ed.ac.uk; Dominic.Campopiano@ed.ac.uk
Abstract: Protein-directed dynamic combinatorial chemistry (DCC) relies on reversible chemical reactions
that can function under the near-physiological conditions required by the biological target. Few classes of
reaction have so far proven effective at generating dynamic combinatorial libraries (DCLs) under such
constraints. In this study, we establish the conjugate addition of thiols to enones as a reaction well-suited
for the synthesis of dynamic combinatorial libraries (DCLs) directed by the active site of the enzyme
glutathione S-transferase (GST). The reaction is fast, freely reversible at basic pH, and easily interfaced
with the protein, which is a target for the design of inhibitors in cancer therapy and the treatment of parasitic
diseases such as schistosomiasis. We have synthesized DCLs based on glutathione (GSH, 1) and the
enone ethacrynic acid, 2a. By varying either set of components, we can choose to probe either the GSH
binding region (“G site”) or the adjacent hydrophobic acceptor binding region (“H site”) of the GST active
site. In both cases the strongest binding DCL components are identified due to molecular amplification by
GST which, in the latter system, leads to the identification of two new inhibitors for the GST enzyme.
Introduction
that it can equilibrate in the presence of the target. Since the
library population distribution is under thermodynamic control,
Dynamic Combinatorial Chemistry (DCC) is a powerful
approach to the discovery of small molecule ligands for large
biomolecules.1 The method uses the biomolecular target to direct
the reversible, in situ assembly of a small molecule library and
as such can be considered, in the wider context of target-guided
synthesis (TGS), an umbrella term that covers a variety of
systems that feature small-molecule synthesis orchestrated by
a large biomolecule (usually a protein).2-10 DCC contrasts with
the majority of TGS methods by using reversible reactions to
assemble the prospective ligands. Whereas kinetic methods use
the interior of proteins to direct and accelerate an irreversible
reaction, usually between isolated sets of reactants, a dynamic
combinatorial library (DCL) of compounds is designed such
stabilization of one member through selective binding to the
protein is expected to amplify that species at the expense of
other (nonbinding) species, generating hit structures that can
be identified through analysis of the DCL population distribu-
tion. DCC thus bridges the gap between chemical synthesis of
drug candidates and their biological binding assay, meshing the
two operations into a single process whereby the structure of
the biological target directs the assembly of its own best inhibitor
in situ.
The set of chemical reactions that have been successfully
applied to the construction of DCLs in the presence of a
biological target is currently rather small, centering on CdN
and SsS bond-forming processes.11-14 This lack of diversity
(1) Reviews: (a) Cheeseman, J. D.; Corbett, A. D.; Gleason, J. L.; Kazlauskas,
R. J. Chem.sEur. J. 2005, 11, 1708-1716. (b) Otto, S. Curr. Opin. Drug
DiscoVery DeV. 2003, 6, 509-520. (c) Otto, S.; Furlan, R. L. E.; Sanders,
J. K. M. Curr. Opin. Chem. Biol. 2002, 6, 321-327. (d) Lehn, J.-M.
Chem.sEur. J. 1999, 5, 2455-2463.
(11) CdN bond formation: (a) Huc, I.; Lehn, J.-M. Proc. Nat. Acad. Sci. U.S.A.
1997, 94, 2106-2110. (b) Bunyapaiboonsri, T.; Ramstro¨m, O.; Lohmann,
S.; Lehn, J.-M.; Peng, L.; Goeldner, M. ChemBioChem 2001, 2, 438-
444. (c) Hochgu¨rtel, M.; Kroth, H.; Piecha, D.; Hofmann, M. W.; Nicolau,
C.; Krause, S.; Schaaf, O.; Sonnenmoser, G.; Eliseev, A. V. Proc. Nat.
Acad. Sci. U.S.A. 2002, 99, 3382-3387. (d) Gerber-Lemaire, S.; Popowycz,
F.; Rodriguez-Garcia, E.; Asenjo, A. T. C.; Robina, I.; Vogel, P.
ChemBioChem 2002, 3, 466-470. (e) Ramstro¨m, O.; Lohmann, S.;
Bunyapaiboonsri, T.; Lehn, J.-M. Chem.sEur. J. 2004, 10, 1711-1715.
(f) Bugaut, A.; Toulme´, J.-J.; Rayner, B. Angew. Chem., Int. Ed. 2004, 43,
3144-3147. (g) Zameo, S.; Vauzeilles, B.; Beau, J.-M. Angew. Chem.,
Int. Ed. 2005, 44, 965-969. (h) Bugaut, A.; Bathany, K.; Schmitter, J.-
M.; Rayner, B. Tetrahedron Lett. 2005, 46, 687-690.
(2) Rideout, D. Science 1986, 233, 561-563.
(3) Nguyen, R.; Huc, I. Angew. Chem., Int. Ed. 2001, 40, 1774-1776.
(4) Nicolaou, K. C.; Hughes, R.; Cho, S. Y.; Winssinger, N.; Smethurst, C.;
Labischinski, H.; Endermann, R. Angew. Chem., Int. Ed. 2000, 39, 3823-
3828.
(5) Nicolaou, K. C.; Hughes, R.; Cho, S. Y.; Winssinger, N.; Labischinski,
H.; Endermann, R. Chem.sEur. J. 2001, 7, 3824-3843.
(6) Erlanson, D. A.; Braisted, A. C.; Raphael, D. R.; Randal, M.; Stroud, R.
M.; Gordon, E. M.; Wells, J. A. Proc. Natl. Acad. Sci. U.S.A. 2000, 97,
9367-9372.
(7) Lewis, W. G.; Green, L. G.; Grynszpan, F.; Radic´, Z.; Carlier, P. R.; Taylor,
P.; Finn, M. G.; Sharpless, K. B. Angew. Chem., Int. Ed. 2002, 41, 1053-
1057.
(8) Mocharla, V. P.; Colasson, B.; Lee, L. V.; Ro¨per, S.; Sharpless, K. B.;
Wong, C.-H.; Kolb, H. C. Angew. Chem., Int. Ed. 2005, 44, 116-120.
(9) Bourne, Y.; Kolb, H. C.; Radic´, Z.; Sharpless, K. B.; Taylor, P.; Marchot,
P. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 1449-1454.
(10) Krasinski, A.; Radic´, Z.; Manetsch, R.; Raushel, J.; Taylor, P.; Sharpless,
K. B.; Kolb, H. C. J. Am. Chem. Soc. 2005, 127, 6686-6692.
(12) S-S bond formation: (a) Ramstro¨m, O.; Lehn, J.-M. ChemBioChem 2000,
1, 41-48. (b) Krishnan-Ghosh, Y.; Balasubramanian, S. Angew. Chem.,
Int. Ed. 2003, 42, 2171-2173. (c) Whitney, A. M.; Ladame, S.; Bala-
subramanian, S. Angew. Chem., Int. Ed. 2004, 43, 1143-1146. (d)
Hotchkiss, T.; Kramer, H. B.; Doores, K. J.; Gamblin, D. P.; Oldham, N.
J.; Davis, B. G. Chem. Commun. 2005, 4264-4266. (e) Ladame, S.;
Whitney, A. M.; Balasubramanian, S. Angew. Chem., Int. Ed. 2005, 44,
5736-5739.
(13) Thioester formation: Larsson, R.; Pei, Z.; Ramstro¨m, O. Angew. Chem.,
Int. Ed. 2004, 43, 3716-3718.
9
10.1021/ja058049y CCC: $33.50 © 2006 American Chemical Society
J. AM. CHEM. SOC. 2006, 128, 8459-8467
8459

A R T I C L E S
Shi et al.
Scheme 1. DCL Composed of EA Analogues and GSH, 1a
a Glutathione, 1, adds to the five enones 2a-e to set up a DCL that has an equitable distribution of thiol conjugates 3a-e and starting enones (chromatogram
a). The reversibility of the DCL is proven by adding an additional enone component, 2f, to the pre-equilibrated library and observing the incorporation of
adduct 3f (chromatogram b). The pH control of the DCL is illustrated by pre-equilibration, acidification to pH 4, and subsequent addition of a large excess
of enone 2f, which is not incorporated into the library as the conjugate addition reaction has been switched off (chromatogram c).
is unsurprising, as the successful reaction must meet two sets
of exacting criteria; it must proceed under the near physiological
conditions required by the protein target as well as fulfilling
the thermodynamic condition of reversibility inherent to the
DCC process. An increase in the number of chemical reactions
available for DCC is necessary to enable greater functional group
diversity and increased applicability of DCLs as a target-guided
synthesis methodology. With this in mind, we have recently
introduced the conjugate addition of thiols to enones as a new
method for DCL synthesis.15 The reaction is well-suited to DCC
being fast, freely reversible, responsive to pH change, and
proceeding in water under mild conditions with no external
reagents (Scheme 1). In addition, the choice of the tripeptide
glutathione (γ-Glu-Cys-Gly, GSH, 1) as the thiol in our
preliminary studies creates DCLs that are biologically relevant
and amenable to targeting by a variety of proteins, the most
relevant being the enzymes responsible for mediating the
conjugate addition chemistry of GSH in the cell: the glutathione
S-transferases (GSTs).
prostaglandins, and base propenals, making them more water
soluble and easily eliminated from the cell. The GSTs are
potential drug targets in cancer therapy, where resistance to
chemotherapeutic drugs has been directly correlated with the
overexpression of GSTs in tumor cells, and parastic diseases
such as malaria and schistosomiasis.18 We now describe the
application of our GSH-based DCLs to the discovery of GST
inhibitors.
Results and Discussion
We initially designed a biased DCL to examine the compat-
ibility of our conjugate addition methodology with GST and to
see whether the enzyme could effectively amplify the best
binding component from a DCL. Since DCL synthesis requires
large amounts of highly purified enzyme, we selected GST from
the helminth worm Schistosoma japonica (SjGST, 26kDa
monomer) as our target transferase, as it is well characterized
and amenable to recombinant overexpression in E. coli as an
affinity-tagged construct.19 The DCL is shown in Scheme 2 and
consists of GSH, 1, three tripeptide GSH analogues 4-6, and
the enone ethacrynic acid, 2a (EA). The tripeptide analogues
were synthesized using standard solid-phase synthesis methods
and differ from GSH at the γ-glutamyl residue. As a result, the
three analogues are expected to be poor fits for the GSH binding
region of the GST active site, as the γ-glutamyl residue is
thought to be critical for binding,20 thus biasing the DCL toward
The GSTs are a large family of dimeric enzymes responsible
for cell detoxification, thereby protecting the cell from cytotoxic
and oxidative stress.16,17 They catalyze the conjugation of GSH
to a wide variety of xenobiotic electrophiles such as quinones,
(14) Enzymatic aldol reaction: (a) Lins, R. J.; Flitsch, S. L.; Turner, N. J.; Irving,
E.; Brown, S. A. Angew. Chem., Int. Ed. 2002, 41, 3405-3407. (b) Lins,
R.; Flitsch, S. L.; Turner, N. J.; Irving, E.; Brown, S. A. Tetrahedron 2004,
60, 771-780.
(15) Shi, B.; Greaney, M. F. Chem. Commun. 2005, 886-888.
(16) Mannervik, B.; Danielson, U. H. CRC Crit. ReV. Biochem. 1988, 23, 283-
337.
(18) Mahajan, S.; Atkins, W. M. Cell. Mol. Life Sci. 2005, 62, 1221-1233.
(19) Waugh, D. S. Trends Biotechnol. 2005, 23, 316-320.
(20) Andersson, C.; Mosialou, E.; Adang, A. E.; Mulder, G. J.; van Der Gen,
A.; Morgenstern, R. J. Biol. Chem. 1991, 266, 2076-2079.
(17) Hayes, J. D.; Flanagan, J. U.; Jowsey, I. R. Annu. ReV. Pharmacol. Toxicol.
2005, 45, 51-88.
9
8460 J. AM. CHEM. SOC. VOL. 128, NO. 26, 2006

GST-Directed DCLs
A R T I C L E S
Scheme 2. DCL Composed of GSH Analogues and EA, 2aa
a The four tripeptides 1, 4-6 undergo reversible conjugate addition to EA, 2a, at pH 7.5. Chromatogram (a) shows the blank DCL after 1 h; chromatogram
(b) shows the DCL after 10 min in the presence of SjGST; chromatogram (c) shows the DCL after pre-equilibration and then addition of SjGST, after 2 days;
chromatogram (d) shows the same system as (c) after 6 days. Although each EA adduct is formed as a mixture of diastereoisomers upon conjugate addition,
giving a total of eight adducts in the DCL, the diasteromeric adducts are not separable under the LC conditions.
the GSH adduct 7. The enone ethacrynic acid is a known
inhibitor of GSTs and provides a motif for further structural
elaboration.
DCL and that the amplified compound 7 is the thermodynami-
cally preferred binder.
The expected adduct 7 has been amplified from 35% of total
conjugate concentration to 92% at equilibrium, a large ampli-
fication that ought to reflect correspondingly large differences
in binding affinity between 7 and those peptides lacking the
γ-glutamyl residue. To verify that amplification and binding
affinity are indeed correlated in this system, we quantified the
inhibitory activity of 7 and 10 for SjGST using the standard
chlorodinitrobenzene (CDNB) GST inhibition assay.21 Measured
IC50 values of 0.32 µM for 7 and 88 µM for 10 differ by over
2 orders of magnitude, demonstrating that the extent of
molecular amplification for each component in the biased DCL
can be clearly related to their binding affinity for the protein
target. Similar strong amplification was also observed in slightly
larger DCLs based upon seven starting tripeptides (vide infra).
Control experiments using bovine serum albumin as the protein
target in place of SjGST produced no changes to the blank DCL
composition, indicating that the active site of SjGST is directing
the amplification of the DCL. The success of these biased DCLs
The blank DCL, assembled in the absence of any enzyme,
equilibrates in 1 h to give the distribution of each of the four
conjugates 7-10 shown in chromatogram 2a (Scheme 2). Upon
incubation of the same components with 0.8 equiv of SjGST
for 10 min, the DCL collapsed to essentially one adduct, the
expected GS-EA, 7 (chromatogram 2b), and remained un-
changed thereafter. The rapid equilibration in the presence of
SjGST indicates that the enzyme is accelerating the conjugate
addition, as would be expected given its catalytic function as a
transferase in GSH biochemistry. To verify that the composition
shown in chromatogram 2b represented true equilibrium and
not a metastable distribution generated by target-accelerated
synthesis in the presence of the enzyme, we added SjGST to
the pre-equilibrated DCL shown in chromatogram 2a. Strong
amplification of adduct 7 was again observed, but at a far slower
rate. After 2 days 7 was present in 55% of total adduct
concentration (chromatogram 2c), and after 6 days the DCL
had equilibrated to an identical distribution to that of chromato-
gram 2b, where adduct 7 is by far the dominant species. These
results indicate that catalysis of the conjugate addition reaction
by SjGST does not effect the equilibrium distribution of the
(21) (a) Habig, W. H.; Pabst, M. J.; Jakoby, W. B. J. Biol. Chem. 1974, 249,
7130-7139. (b) Burg, D.; Filippov, D. V.; Hermanns, R.; van der Marel,
G. A.; van Boom, J. H.; Mulder, G. J. Bioorg. Med. Chem. 2002, 10, 195-
205.
9
J. AM. CHEM. SOC. VOL. 128, NO. 26, 2006 8461

A R T I C L E S
Shi et al.
Scheme 3. A DCL Made from GSH and 14 EA Analogues
encouraged us to extend the thiol addition methodology to the
discovery of new GST inhibitors through the synthesis of larger,
nonbiased DCLs.
obtained when SjGST was added to the pre-equilibrated DCL,
indicating that the thermodynamic reversibility inherent to the
thiol conjugate addition in the absence of any target is
maintained in the presence of GST. To ascertain whether the
hit structures generated from DCC were in fact active inhibitors
of SjGST, we synthesized the amplified adducts 12a and 12n,
a nonamplified adduct 12b, and the depreciated adduct 12f
separately and measured their IC50 values (Figure 2).
All GSTs contain two recognition areas for GS-R conju-
gates: a highly conserved GSH binding site and a hydrophobic
binding region for the electrophilic substrate, called the H-site,
which varies across the different GST isoforms. The design of
inhibitors which target the H-site is a promising strategy in terms
of isozyme discrimination,18 so we elected to use DCC to
explore the H-site of SjGST with a series of EA analogues. We
prepared a DCL with reversed stoichiometry from that in
Scheme 2, whereby n EA analogues react with GSH to afford
n GS-EA adducts. The EA analogues were prepared from EA
via EDC-mediated amide bond formation at the carboxylic acid
group, producing a series of compounds 11a-n that display a
variety of randomly selected functional groups (Scheme 3). The
fourteen EA analogues were equilibrated with GSH in both the
absence and presence of SjGST (2 equiv). Figure 1 shows the
LC traces of each DCL, along with an overlay for comparison
purposes. The large peak eluting at ∼5.9 min in the amplified
trace is due to trace amounts of SjGST still present in the
reaction mixture. We verified that all of the expected GSH
conjugates were present in the DCL using ESI-MS analysis and
then looked to identify only those components that have been
amplified by the enzyme.
Looking at the overlay trace in the area of the GSH conjugates
we can see that a single peak has been amplified, marked with
a star at RT ) 4.8 min. This amplification is at the expense of
the first peak at RT ) 3.3 min, which is reduced in intensity.
MS analysis of the amplified peak indicated four adducts, 12a,
d, m, and n, while the reduced peak is due to adduct 12f.
Deconvolution studies established that each of the adducts
12a, m, and n were being amplified in the presence of SjGST,
whereas 12d was not.22 Identical amplification results were
Both piperidine (12a) and leucine (12n) derivatives of EA
inhibit SjGST at the low micromolar/high nanomolar level, with
the piperidine derivative 12a being slightly more potent (IC50
) 0.61 µM versus IC50 ) 1.40 µM). The more polar lysine
amide 12f is, as expected, the least active, with an IC50 value
more than 10-fold lower than 12a at 8.20 µM. The morpholine
amide 12b lies halfway between 12a and 12f with an IC50 value
of 4.30 µM, approximately 7-fold lower than the piperidine
amide. These figures indicate that the extent of DCL amplifica-
tion not only reflects the relative binding affinities of DCL
components for the protein target but is also discriminatory
across an order of magnitude range of IC50 values. The
piperidine and leucine amides 12a and 12n are amplified from
the library at the expense of the lysine amide 12f, which being
an approximately 10-fold weaker binder is removed from the
equilibrium composition of the DCL. The reversibility inherent
to the thiol conjugate addition is enabling the SjGST target to
amplify relatively weak-binding compounds from a nonbiased
selection of small molecules, demonstrating that this TGS
methodology can be applied in a medicinal chemistry lead-
finding context.
As EA is known to inhibit GSTs both on its own and as a
GSH conjugate,23 we were interested in acquiring IC50 data for
some of the enone starting materials. We assayed the five enones
2a, 11a, 11b, 11f, and 11n, the corresponding Michael acceptors
(23) Oakley, A. J.; Rossjohn, J.; Lo Bello, M.; Caccuri, A. M.; Federici, G.;
(22) See Supporting Information.
Parker, M. W. Biochemistry 1997, 36, 576-585.
9
8462 J. AM. CHEM. SOC. VOL. 128, NO. 26, 2006

GST-Directed DCLs
A R T I C L E S
Figure 1. HPLC chromatograms for the targeted DCL of EA analogues. Chromatogram (a) shows the blank DCL distribution after 1 h, chromatogram (b)
shows the DCL composition in the presence of SjGST after 30 min and chromatogram (c) shows an overlay of (a) and (b). The two dotted lines indicate the
peak that is reduced in intensity on addition of the enzyme, at RT ) 3.3 min, along with the peak that is amplified upon enzyme addition, at RT ) 4.8 min.
Figure 2. Inhibition profiles and IC50 values for amplified adducts 12a and 12n, nonamplified adduct 12b, and depreciated adduct 12f along with GS-EA,
7, for comparison purposes.
to the five GSH conjugates we had previously investigated. The
inhibition data are shown in Figure 3 and follow a similar trend
to that of the conjugates, with 2a ≈ 11a > 11n > 11b > 11f.
Additionally, the quantitative IC50 values are similar for each
enone and its corresponding GSH conjugate, a common
observation across a number of GST isozymes.23-25 The absence
of any significant augmentation of binding in the case of the
GS-EA conjugates relative to the EA analogues suggests that
the EA moiety of the conjugate may bind in different orienta-
tions and/or positions within the H-site relative to the parent
(24) Ploemen, J. H. T. M.; van Ommen, B.; van Bladeren, P. J. Biochem.
Pharmacol. 1990, 40, 1631-1635.
(25) Awasthi, S.; Srivastava, S. K.; Ahmad, F.; Ahmad, H.; Ansari, G. A. S.
Biochim. Biophys. Acta 1993, 1164, 173-178.
9
J. AM. CHEM. SOC. VOL. 128, NO. 26, 2006 8463

A R T I C L E S
Shi et al.
Figure 3. Inhibition profiles and IC50 values for enone starting materials 11a, b, f, and n along with EA, 2a.
compound. This phenomenon has been observed in the crystal
structures of EA and GS-EA bound into the active site of
GSTP1-123 and is perhaps to be expected, given that the two
binding fragments of GSH and EA are linked in the conjugate
by a single covalent bond which offers little flexibility for the
EA group to explore analogous optimal binding modes to the
individual enones.
of DCC as a TGS method in medicinal chemistry will likely be
realized early in the drug discovery process, where libraries of
weak-binding compounds can be rapidly screened to produce
lead structures for further development. The ability of SjGST
to amplify the low micromolar inhibitors 12a and 12n from a
nonbiased thiol conjugate DCL, at the expense of compound
12f which differs in IC50 by only a single order of magnitude,
illustrates this discovery process in a microcosm and demon-
strates the power of TGS in combining the chemical synthesis,
discovery, and biological binding assay of enzyme inhibitors
into a single operation. Future work will extend our GSH-based
DCL methodology to the discovery of novel inhibitors for a
range of human GST isoforms.
It is important to note that GST assays use the non-natural
CDNB as a substrate and the exact nature of binding of this
substrate to SjGST is not known. Moreover, the details of GSH-
EA binding to SjGST are also lacking since a high-resolution
structure of the SjGST/GS-EA complex has not been deter-
mined. A model structure of the SjGST GSH EA Michaelis
complex was recently proposed based on structures of SjGST
bound to a series of glutathione derivatives (glutathione sul-
fonate, S-hexyl glutathione, and S-2-iodobenzyl glutathione).26
This model identifies a series of residues (Arg103, Tyr111,
Ser107, Gln 204) that could interact with the EA carboxylic
acid group and proposes that the enzyme must undergo
conformational changes to allow conjugate addition of the GSH
thiol onto the enone of EA. Interestingly, our identification of
the amplified amide compounds 12a and 12n from our DCLs
indicates that the carboxylic acid group of EA is not essential
for binding in the SjGST H site and may be successfully
extended without significant loss of inhibitory activity.
Experimental Section
General. All reagents were purchased from commercial suppliers
and used as received. Dichloromethane (DCM) was distilled from CaH2.
TLC was performed with Merck aluminum plates silica gel 60 F254
.
Melting points were recorded on a Gallenkamp melting point instrument
and are uncorrected. 1H and 13C NMR spectra were recorded on a
Bruker DPX360 MHz NMR spectrometer unless otherwise specified.
Chemical shifts δ are quoted in ppm relative to the CDCl3 signal as
reference. Coupling constants are given in Hz. Optical rotations were
measured on a POLAAR 20 automatic polarimeter using a 2 mL cell.
Mass spectra were obtained from the EPSRC National Mass Spec-
trometry Service Centre, University of Wales, Swansea.
Analytical HPLC was performed on an HP1100 instrument and
analytical HPLC-MS on a Waters 2790 HPLC with a micromass
Platform II quadrupole mass spectrometer. Enzyme concentrations were
determined by Bradford assay on a UNICAM UV-visible spectrometer
with bovine serum albumin as a standard.27 Enzyme activity assays
were conducted on a CARY 300 SCAN UV-visible spectrometer.
Synthesis of Library Components. A. General Protocol for
Peptide Syntheses: Tripeptides 4-6, 13-15 were prepared using
standard Fmoc solid-phase synthesis methods. The preparation of
H-Val-Cys-Gly-OH (4) is illustrative: Fmoc-Gly-Wang (0.6 mmol/g)
was swollen in DCM at rt for 10 min. The resin was drained and treated
with 20% piperidine in dry DMF. The reaction was agitated with
nitrogen at rt and monitored by Kaiser test. After completion, the resin
was washed with DMF, DCM, and MeOH and dried in vacuo. The
Conclusions
We have established the conjugate addition reaction of thiols
to enones as a new reaction for enzyme-directed DCC. Using
the enzyme SjGST, we could investigate either the GSH- or
the H-regions of the enzyme active site by varying DCL
stoichiometry between n thiols/one enone and one thiol/n
enones. DCL amplification was successful for both approaches,
with the amplified compounds corresponding to the best
inhibitors as measured by a competition assay. While the degree
of amplification from the EA-based DCL is quantitatively far
lower than that observed in the biased GSH-based DCLs, it
represents a more significant application of DCC. The power
(26) Cardoso, R. M. F.; Daniels, D. S.; Bruns, C. M.; Tainer, J. A. Proteins
2003, 51, 137-146.
(27) Bradford, M. Anal. Biochem. 1976, 72, 248-254.
9
8464 J. AM. CHEM. SOC. VOL. 128, NO. 26, 2006

GST-Directed DCLs
A R T I C L E S
resulting H-Gly-Wang resin was then swollen in DCM at rt for 10 min
and drained. Fmoc-Cys(Trt)-OH (5 equiv), HBTU (4.9 equiv), and
HOBt‚H2O (5 equiv) were mixed in dry DMF and added to the resin.
The mixture was agitated with nitrogen, while DIPEA (10 equiv) was
added. The reaction was monitored by Kaiser test. After completion,
the resin was washed with DMF, DCM, and MeOH and dried in vacuo
to afford Fmoc-Cys(Trt)-Gly-Wang. Fmoc removal from the cysteine
residue was accomplished as described for the deprotection of Gly,
followed by coupling to Fmoc-Val under the same conditions as those
used for coupling to cysteine with the exception that DIPEA was added
to the mixture of peptide, HBTU, and HOBt‚H2O for the activation.
Fmoc removal from Fmoc-Val-Cys(Trt)-Gly-Wang was carried out as
before, and the resin finally dried in vacuo over KOH. Cleavage and
deprotection were accomplished by adding aqueous TFA solution
(94.5% TFA, 2.5% H2O, 2.5% EDT, 1% TIS) to the resin followed by
gentle filtration and washing with TFA solution (×2). All of the filtrate
was collected and added dropwise to a 10-fold excess of cold ether in
a centrifuge tube. After centrifugation, ether was carefully decanted,
and the centrifugation was repeated twice. Water was then added to
the white residue, and the aqueous solution was transferred to a
separatory funnel. The aqueous phase was washed with DCM 3 times
and then lyophilized to give the pure tripeptide 4 (11 mg, 35%) as a
metal-affinity resin, rather than glutathione-affinity resins which are
typically used to purify GST isoforms. This strategy thus provided large
amounts of SjGST free from large amounts of glutathione (GSH) which
could interfere with dynamic combinatorial library (DCL) synthesis.
The Schistosoma japonica gst gene was amplified by PCR using
pGEX6P-1 (Amersham Biosciences, Genbank Accession Number
U78872) plasmid DNA as a template with primers GST-FOR (5′ GGA
AAC AAG CTT CAT GAC CCC TAT 3′) which incorporated a BspHI
site (in bold) and GST-REV (5′ GAA CTT CGG GGA TCC CAT
GGG CCC 3′) which incorporated a BamHI site (in bold). The PCR
products were cloned into plasmid pUC19, sequenced to confirm the
fidelity of the cloned DNA, and then cloned into plasmid pET-6His
plasmid DNA digested with appropriate restriction enzymes (NcoI and
BamHI for the pET plasmid and BspHI and BamHI for pUC19/SjGST).
Clones containing the correct insert were isolated and sequenced to
confirm the identity of the pET-6His-SjGST plasmid.
B. Expression and Purification of GST. Overexpression of
Schistosoma japonica GST was achieved by transforming E. coli
BL21(DE3) (Novagen) cells with the plasmid pET-6His-SjGST. A
single colony was used to inoculate 50 mL of LB broth supplemented
with ampicillin (100 µg/mL) and grown overnight at 37 °C with
shaking. This overnight seed culture was then used to inoculate 2 L of
fresh growth medium and grown at 37 °C to OD600 ) 0.8 before
induction with isopropyl thio-â-D-galactoside (IPTG) (1.0 mM final
concentration). After a further 3 h agitation at 37 °C, the cells were
harvested by centrifugation (6000 g for 15 min at 4 °C) and frozen at
-20 °C for 16 h. The defrosted cells were resuspended in 50 mL of
binding buffer (10 mM Tris/HCl, pH 7.5, 10 mM NaCl, containing
one tablet of Complete Protease Inhibitor Cocktail (Roche)) and gently
agitated for 1 h at 4 °C. The cells were disrupted by sonication (15
pulses of 30 s at 30-s intervals, on ice), and the cell debris was removed
by centrifugation at 27 000 g for 20 min at 4 °C after which the cell
lysate supernatant was filtered through a 0.45-µm membrane prior to
chromatography.
1
fluffy white solid: H NMR (360 MHz, D2O) δ ) 1.06 (6H, dd, J )
6.5, 6.5 Hz, Val (CH3)2), 2.27 (1H, m, Val CH), 2.91-3.04 (2H, m,
Cys CH2), 3.91 (1H, d, J ) 5.8 Hz, Val CH), 3.97-4.08 (2H, m, Gly
CH2), 4.53 (1H, t, J ) 6.5 Hz, Cys CH); 13C NMR (91 MHz, D2O) δ
) 17.8, 18.6, 26.2, 31.1, 42.4, 56.8, 59.4, 170.4, 172.7, 174.3; HRMS
(ES+) calcd for C10H20O4N3S [M + H]+ 278.1169, found 278.1171.
B. Synthesis of EA derivatives. EA derivatives were synthesized
according to the coupling procedure described in ref 15, followed by
Boc cleavage with TFA. The preparation of 11g is illustrative:
Ethacrynic acid (151.6 mg, 0.50 mmol), HOBt‚H2O (84.2 mg, 0.55
mmol), EDCI (105.4 mg, 0.55 mmol), H-Ser(tBu∼)-OtBu‚HCl (139.6
mg, 0.55 mmol), and DIPEA (428 µl, 2.50 mmol) were dissolved in
DCM (10 mL), and the reaction was left to stir at room temperature
overnight. The reaction was worked up in the usual fashion, followed
by flash column chromatography (SiO2, DCM/EtOAc 9:1) to yield the
tert-butyl-protected amide (156 mg, 62%) as a colorless oil; Rf ) 0.3
(SiO2, DCM/EtOAc 9:1); 1H NMR (360 MHz, CDCl3) δ ) 1.09-
The cell lysate was applied to a HisTrap HP column (5 mL,
Amersham Biosciences) pre-equilibrated with binding buffer (50 mM
Tris/HCl, 0.5 M NaCl, 5 mM imidazole, pH 7.5). The column was
then washed with 20 column volumes of binding buffer (100 mL) before
bound material was eluted using a linear gradient of 0-100% elution
buffer (50 mM Tris/HCl, 0.5 M NaCl, 0.5 M imidazole, pH 7.5, over
20 column columns, 100 mL; see Figure S5.1 in the Supporting
Information). Fractions were analyzed by SDS-PAGE, and those
containing GST were pooled and dialyzed overnight against 4 L of 10
mM Tris/HCl (pH 7.5) at 4 °C. SDS-PAGE analysis suggested that
the protein was greater than 95% pure (Figure S5.2 in the Supporting
Information). Further purification was achieved using MonoQ anion
exchange (1 mL, Amersham Biosciences). The protein was loaded onto
the column, washed with 50 mM Tris, pH 7.5, and then eluted with a
linear 0-1 M NaCl gradient over 20 mL. Fractions were analyzed by
SDS-PAGE, and the purest GST fractions were pooled, concentrated
by ultrafiltration (10 kDa cutoff), and frozen. A 1 mg aliquot of the
isolated 6His-SjGST was analyzed by gel filtration chromatography
(Sephacryl HR-75, Amersham Bioscience) to confirm that it was
dimeric in solution.
t
t
1.16 (12H, m, Ser Butyl + CH3), 1.45 (9H, s, Ser Butyl), 2.45 (2H,
q, J ) 7.4, methylene CH2), 3.54 (1H, dd, Ser CH2, J1 ) 8.7 Hz, J2 )
3.1 Hz), 3.85 (1H, dd, Ser CH2, J1 ) 8.7 Hz, J2 ) 2.7 Hz), 4.54-4.61
(2H, m, OCH2), 4.66 (1H, m, Ser CH), 5.57 (1H, s, enone H), 5.93
(1H, s, enone H), 6.85 (1H, d, J ) 8.5 Hz, Ph H), 7.17 (1H, d, J ) 8.5
Hz, Ph H), 7.58 (1H, d, J ) 8.7 Hz, amide H); 13C NMR (91 MHz,
CDCl3 + CD3OD) δ ) 13.4, 24.4, 28.3, 29.0, 53.8, 63.0, 69.0, 74.2,
83.1, 111.8, 124.1, 128.2, 129.8, 132.4, 135.0, 151.2, 155.6, 167.6,
169.8, 196.7; HRMS (ES+) calcd for C24H34O6NCl2 [M + H]+
502.1758, found 502.1748; [R]20D ) +11° (c ) 2.72, chloroform). The
amide was then taken up into dry DCM (5-10 mL), and TFA (5 mL)
was added. The reaction was allowed to stir under nitrogen at rt for 1
day, after which time it was partitioned between water and DCM and
the organic phase was extracted (×3). Drying over MgSO4 and
concentration in vacuo yielded the acid 11g (66 mg, 54%) as a white
1
solid; mp 144-146 °C; H NMR (360 MHz, CDCl3 + CD3OD) δ )
Electrospray mass spectrometry (ESI-MS) analysis of the pure
enzyme gave the molecular mass of the 6His-SjGST as 28 068 Da ((5
Da), in good agreement with the predicted value of 28 057 Da for the
monomer. The final yield of GST using this method was ∼20 mg per
L of bacterial culture, and this protein was used for all subsequent
library synthesis and inhibition assays.
1.03 (3H, t, J ) 7.4 Hz, CH3), 2.34 (2H, q, J ) 7.4 Hz, methylene
CH2), 3.78-3.93 (2H, m, Ser CH2), 4.49-4.57 (3H, m, OCH2 + Ser
CH), 5.50 (1H, s, enone H), 5.87 (1H, s, enone H), 6.83 (1H, d, J )
8.5 Hz, Ph H), 7.07 (1H, d, J ) 8.5 Hz, Ph H); 13C NMR (91 MHz,
CDCl3 + CD3OD) δ ) 17.2, 28.3, 59.4, 66.9, 72.9, 116.1, 128.1, 132.1,
134.2, 136.2, 138.8, 155.0, 159.7, 172.7, 201.2; HRMS (ES+) calcd
Dynamic Combinatorial Chemistry. Representative procedure for
GSH analogue DCL: The seven GSH tripeptides 1, 4-6, 13-15 (7 ×
8.2 µL, 10 mM aqueous) and ethacrynic acid (0.82 µl, 0.1 M in DMSO)
were added to tris buffer (200 µL, 50 mM, pH 7.5). The DCL was
allowed to stand at rt with occasional gentle shaking and monitored
by HPLC at regular intervals to establish the blank DCL composition
for C16H21O6N2Cl2 [M + NH4]+ 407.0771, found 407.0771; [R]21
+28.1° (c ) 1.32, chloroform/methanol 9:1).
)
D
Protein Synthesis. A. Cloning of SjGST as a His-Tagged Fusion.
The pET-6His-SjGST plasmid was designed to have SjGST fused to a
6His N-terminal tag which facilitates purification of the SjGST on nickel
9
J. AM. CHEM. SOC. VOL. 128, NO. 26, 2006 8465

A R T I C L E S
Shi et al.
Scheme 4. HPLC Analysis of DCL Made from Seven GSH-Based Tripeptides 1, 4-6, 13-15a
a Chromatogram (a) shows the blank DCL, and chromatogram (b) shows the DCL in the presence of SjGST. The GSH-EA adduct 7 is amplified from
35% of total adduct concentration to 88%.
(HPLC conditions: Column: Luna 5 µ C18(2) 30 mm × 4.60 mm,
flow rate 1 mL min-1, wavelength 254 nm, temperature 23 °C, gradient
H2O/MeCN (0.01% TFA) from 80% to 70% over 8 min, then to 40%
over 1 min eventually to 20% over 2 min).
The DCL was then resynthesized in the presence of SjGST (200 µl,
160 µM in 50 mM tris buffer pH 7.5), and the HPLC trace was
compared to that of the blank (Scheme 4). The DCL composition was
identical in the case of addition of SjGST from the start and the case
of SjGST addition to a pre-equilibrated library.
11a-n (14 × 2 µL, 0.1 M in DMSO) were added to tris buffer (3.3
mL, 50 mM, pH 8). The DCL was allowed to stand with occasional
gentle shakes at rt and monitored by HPLC at regular intervals. LC-
MS verified that each of the expected adducts was present in the blank
DCL (Figure S6.1). HPLC conditions: Column: Luna 5 µ C18(2) 30
mm × 4.60 mm, flow rate 1 mL min-1, wavelength 254 nm,
temperature 23 °C, gradient H2O/MeCN (0.01% TFA) from 80% to
5% over 10 min.
The DCL was then resynthesized in the presence of SjGST (1.1 mL,
180 µM in 50 mM tris buffer pH 7.5) and added to tris buffer (2.2 mL,
50 mM, pH 8), and the HPLC traces were compared (Figure 1). The
Representative procedure for EA analogue DCL: Reduced gluta-
thione (20 µL, 10 mM aqueous) and the 14 ethacrynic acid derivatives
9
8466 J. AM. CHEM. SOC. VOL. 128, NO. 26, 2006

GST-Directed DCLs
A R T I C L E S
DCL composition was identical in the case of addition of SjGST from
the start and the case of SjGST addition to a pre-equilibrated library.
Inhibition Assays. To a 500 µL cuvette were added phosphate buffer
(415 µL, 0.1 M pH 6.6), GST (10 µL, 0.148 mg/mL), and inhibitor
(12.5 µL in DMSO), and the solution was mixed well. After incubation
at ambient temperature for 5 min, CDNB (12.5 µl, 40 mM in EtOH)
and GSH (50 µl, 10 mM) were added and quickly mixed well.
Absorbance was measured at 340 nm at 20 °C for 5 min. Prior to each
experiment, the baseline of the UV spectrometer was corrected by
replacing GST solution with phosphate buffer. For each inhibitor, a
group of experiments were conducted by varying inhibitor concentra-
tions (0.02-10 µM). After each experiment absorbance against time
graphs were plotted such that the initial gradients as velocities were
worked out. To calculate the IC50 of each inhibitor, the ratio of these
velocities and the velocity without any inhibitor were plotted against
inhibitor concentrations.
Acknowledgment. We thank the HFSP for the award of a
postdoctoral fellowship to Dr. Shi, AstraZeneca and Pfizer for
research support grants, and the EPSRC mass spectrometry
service at the University of Swansea.
Supporting Information Available: Characterization data for
all compounds and analyses of GST protein purity. This material
is available free of charge via the Internet at http://pubs.acs.org.
JA058049Y
9
J. AM. CHEM. SOC. VOL. 128, NO. 26, 2006 8467