Identification of novel inhibitors of UDP-Glc 4′-epimerase, a validated drug target for african sleeping sickness

Article abstract of DOI:10.1016/j.bmcl.2006.08.091

Novel inhibitors of Trypanosoma brucei and mammalian UDP-Glc 4′-epimerase were identified by screening a small library of natural products and commercially available drug-like molecules. The inhibitors possess low micromolar potency against the T. brucei and human enzymes in vitro, display a degree of selectivity between the two enzymes, and are cytotoxic to cultured T. brucei and mammalian cells.

Full text of DOI:10.1016/j.bmcl.2006.08.091

Bioorganic & Medicinal Chemistry Letters 16 (2006) 5744–5747
Identification of novel inhibitors of UDP-Glc 40-epimerase, a
validated drug target for african sleeping sickness
Michael D. Urbaniak,a Jioji N. Tabudravu,b Aichi Msaki,a Kathy Mansfield Matera,a
Ruth Brenk,a Marcel Jasparsb and Michael A. J. Fergusona,*
aDivision of Biological Chemistry and Molecular Microbiology, School of Life Sciences, University of Dundee,
Dundee DD1 5EH, Scotland, UK
bMarine Natural Product Laboratory, Department of Chemistry, University of Aberdeen, Aberdeen AB24 3EU, Scotland, UK
Received 25 April 2006; revised 22 August 2006; accepted 23 August 2006
Available online 7 September 2006
Abstract—Novel inhibitors of Trypanosoma brucei and mammalian UDP-Glc 40-epimerase were identified by screening a small
library of natural products and commercially available drug-like molecules. The inhibitors possess low micromolar potency against
the T. brucei and human enzymes in vitro, display a degree of selectivity between the two enzymes, and are cytotoxic to cultured
T. brucei and mammalian cells.
ꢀ 2006 Elsevier Ltd. All rights reserved.
Trypanosoma brucei, a protozoan parasite transmitted
by the bite of the Tsetse fly, is the causative agent of
African sleeping sickness in humans and the related dis-
ease Nagana in cattle. T. brucei cells are able to survive
in the blood of the host by virtue of a dense surface coat
of variant surface glycoprotein that protects the parasite
from the complement pathway and undergoes antigenic
variation to evade specific immune responses, making
the production of a vaccine unfeasible.1 Sleeping sick-
ness is invariably fatal if not treated and is responsible
for upwards of 30,000 deaths per year in Sub-Saharan
Africa.2 Current treatments are expensive, toxic, and
difficult to administer, leaving an urgent need for new
therapeutic agents.
essential for survival of the parasite in both the dis-
ease-relevant bloodstream form and the Tsetse-fly dwell-
ing procyclic form.8,9 The high-resolution crystal
structures of human GalE (hGalE) and TbGalE show
that the enzymes share a common topology, and despite
the low level of sequence identity (33%), many protein–
ligand interactions are conserved.10,11 Selective inhibi-
tion of the parasite enzyme may be possible due to a
non-conservative substitution in the UDP-binding pock-
et from Gly237 in hGalE to Cys266 in TbGalE.11 How-
ever, such selectivity may not be essential, as deficiency
in the human enzyme is either asymptomatic or leads
to a rare form of galactosaemia,12 a chronic condition
that can be partially controlled by diet and that may
be tolerable over a short period. Therefore, we believe
that TbGalE is a suitable target for the development
of drugs to treat African sleeping sickness.
Bloodstream form T. brucei contains many galactose-
containing glycoproteins,3–5 yet unlike mammalian cells,
the parasite hexose transporters are unable to uptake
galactose directly.6,7 Instead, T. brucei must obtain its
galactose via the action of the enzyme UDP-glucose
40-epimerase (GalE) that interconverts UDP-glucose
(UDP-Glc) and UDP-galactose (UDP-Gal).8 The gener-
ation of a T. brucei GalE (TbGalE) conditional null
genetic mutant has demonstrated that this enzyme is
We are not aware of any precedent for developing inhib-
itors of UDP-Glc 40-epimerase as drug leads; indeed,
there are few reported inhibitors of GalE. The substrate
analogue UDP-4-deoxy-4-fluorogalactose is reported to
bind Escherichia coli GalE (Kd = 1.1 mM),13 and a com-
petitive inhibitor of hGalE (1–143) has been identified
from a combinatorial chemistry library of 1338 uri-
dine-based Schiff bases (IC50 = 10 lM),14 but neither
are sufficiently potent or drug-like to be pursued as lead
compounds. In order to identify novel and potent inhib-
itors that are not substrate analogues, we decided to
Keywords: Epimerase; UDP-Glc 40-epimerase; UDP-glucose; UDP-
galactose; Trypanosoma brucei; Trypanosome.
*
Corresponding author. Tel.: +44 1382 384219; fax: +44 1382
348896; e-mail: m.a.j.ferguson@dundee.ac.uk
0960-894X/$ - see front matter ꢀ 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2006.08.091

M. D. Urbaniak et al. / Bioorg. Med. Chem. Lett. 16 (2006) 5744–5747
5745
screen a small library of natural product consisting of 44
compounds isolated from marine invertebrates, and a
small commercial library of 880 structurally diverse
drug-like molecules (Prestwick) against TbGalE.
The activity of GalE can be conveniently measured by
using an excess of UDP-glucose dehydrogenase
(UGD) in a coupled enzyme assay (Fig. 1), enabling
the reaction to be followed by monitoring the produc-
tion of NADH at 340 nm.15 Although E. coli UGD
has previously been cloned and over-expressed in E.
coli,16 and the enzyme is commercially available, we
decided to clone and express the E. coli UGD to provide
a cheap and readily available source of the enzyme for
our screening efforts. The E. coli K12 UGD sequence
(GenBank Accession No. U00096) was PCR amplified
from genomic DNA and cloned into the expression vec-
tor pET15b (Novagen), appending an N-terminal hexa-
histidine tag. Over-expression in E. coli and single step
purification using an immobilized nickel column pro-
duced P100 mg of purified UGD per litre of cell
culture.
Prior to compound screening, the kinetics of the
TbGalE–UGD coupled assay were examined to ensure
that the activity of UGD was in excess; the specific activ-
ity of 6.5 U/mg and Km for UDP-Gal of 95 10 lM ob-
tained were in good agreement with literature values for
TbGalE of 5.9 U/mg and 77 lM, respectively.8 The nat-
ural products were screened in triplicate using the cou-
pled assay with an inhibitor concentration of 50 lM,17
revealing three compounds that showed >75% inhibi-
tion. The commercial compounds were screened once
with an inhibitor concentration of 50 lM,17 identifying
thirteen compounds that showed >75 % inhibition (Z0
0.85).
Figure 2. Structure of the GalE inhibitors selected for further study.
significantly different potencies against the two en-
zymes, which may indicate that the compounds have
distinct interactions with each enzyme. However, only
three of the compounds (1–3) were found to be more
potent against the T. brucei enzyme than the human
enzyme.
To confirm that these initial hits were inhibiting TbGalE
rather than the coupling enzyme UGD, their inhibition
was measured directly using high pH anion exchange
chromatography (HPAEC) to follow the conversion of
UDP-Gal to UDP-Glc by TbGalE.18 Whilst the major-
ity of compounds displayed significant inhibition of
TbGalE at 50 lM, five of the initial hits (all commercial
compounds) showed no significant inhibition. The most
potent seven inhibitors (Fig. 2) that gave greater than
75% inhibition of TbGalE at 50 lM were selected for
further study.
To further investigate the binding of the new inhibitors
we docked the three compounds that showed selectivity
for TbGalE (1–3) into the substrate binding site of the
TbGalE crystal structure.19,20 Compound 2 was too
large to be accommodated without structural rearrange-
ments of the binding site. Compounds 3 and 1 show a
good shape complementarity to the pocket and occupy
roughly the same space as the UDP part of the natural
substrate. In addition, in the predicted binding mode,
compound 1 forms a hydrogen bond to Phe255 and a
salt bridge to Arg268 (Fig. 3). Both interactions are also
found in the crystal structure complex with the sub-
strate. These docking results suggest that at least com-
pounds 1 and 3 can act as competitive inhibitors of
TbGalE.
The potencies of the selected inhibitors against T. bru-
cei- and human-GalE were determined using the
HPAEC assay,18 confirming the low micromolar IC50
values expected (Table 1). All the inhibitors showed
In order to determine if any of the compounds were
covalently modifying TbGalE, compounds were co-in-
cubated with the enzyme prior to mass spectrometric
analysis.21 Whereas compounds 3, 6 and 7 did not form
any covalent adducts, compounds 1, 2, 4 and 5 were
found to form multiple adducts with TbGalE at high
compound concentrations (50 lM), indicating that
non-selective covalent modifications were occurring,
probably due to the intrinsic reactivity of the molecules.
Figure 1. The GalE–UGD coupled enzyme assay. Provided UGD is in
excess, the kinetics of GalE can be measured by following the
production of NADH at 340 nm.

5746
M. D. Urbaniak et al. / Bioorg. Med. Chem. Lett. 16 (2006) 5744–5747
Table 1. Potency of the selected GalE inhibitors in vitro
a
a
Compound
In vitro IC50 (lM)
hGalE
14 7.7
Cytotoxicity EC50 (lM)
TbGalE
T. brucei
CHO-K1
1
2
3
4
5
6
7
3.1 0.1
5.6 1.5
17 8.1
15 5.1
14 5.6
13 3.2c
0.62 0.2
3.3 0.27
>100
56 38
>100
12 7.2
75 22
3.6 0.5
2.8 1.6
2.6 2.4
5.2 2.6
0.3 0.25
4.1 2.7
8.5 4.2
0.9 0.6
6.0 2.4
0.6 0.35
6.1 1.8
NDb
0.5 0.6
0.014 0.003
a Values are means of three experiments, standard deviation is given.
b Insufficient material for determination.
c Pre-incubation with TbGalE for 1 h, otherwise IC50 > 100 lM.
Figure 3. Molecular model of compound 1 bound to TbGalE. Atoms are coloured as follows: carbon grey (protein)/yellow (compound 1) nitrogen—
blue, oxygen—red and chlorine—green.
Compound 1 contains an enone moiety that is suscepti-
ble to Michael addition, compound 2 can tautomerise to
an enol sulfate and N,N-dimethylguanidine and com-
pounds 4 and 5 are able to undergo disulfide bond
exchange.
essential to the survival of these cells, and is instead
likely to result from separate off-target effects. Indeed,
many of these compounds are known to possess other
biological activity; ethacrynic acid (1) is a clinically used
loop diuretic, diethylstilbestrol (3) is a carcinogenic oest-
rogen mimic, psammaplin A (4) and the related bisapra-
sin (5) inhibit topoisomerase and histone deactylase
activity, haloprogin (6) is a clinically used topical fungi-
cide and ebselen (7) has entered stage III clinical trials
for ischaemic stroke treatment. The formation of multi-
ple covalent adducts observed for compounds 1, 4 and 5
may account in part for their cytotoxicity, but does not
adequately explain the observed differences in potency
between cell types.
The cytotoxicity of the inhibitors against cultured
bloodstream-form T. brucei and mammalian CHO-K1
cells was also determined (Table 1).22 All the inhibitors
were found to display significant cytotoxicity against
both cell types except for the natural product 2, which
may reflect poor cell permeability or inherent reactivity
of this compound. Unfortunately, only 1 displayed the
desired selective cytotoxicity against T. brucei cells,
and the most potent in vitro inhibitor 7 gave only micro-
molar cytotoxicity. However, for the structurally unre-
lated compounds 1, 3 and 4 the trends in IC50 values
against trypanosomal and mammalian GalE correlate
well with the trends in the EC50 cytotoxicity values
against the respective cultured cells, suggesting that the
two properties are linked.
In conclusion, we have identified novel inhibitors of
UDP-glucose 40-epimerase by screening a small library
of natural products and commercially available com-
pounds. The inhibitors possess low micromolar potency
against the T. brucei and human enzymes in vitro, dis-
play a degree of selectivity between the two enzymes
and are cytotoxic to cultured T. brucei and CHO-K1
cells. However, none of the compounds are promising
drug leads, and the number of intrinsically chemically
The cytotoxicity of the inhibitors against CHO-K1 cells
is somewhat unexpected since it is not clear if GalE is

M. D. Urbaniak et al. / Bioorg. Med. Chem. Lett. 16 (2006) 5744–5747
5747
7. Eisenthal, R.; Game, S.; Holman, G. D. Biochem. Biophys.
Acta 1989, 985, 81.
reactive compounds obtained as hits was somewhat
disappointing.
8. Roper, J. R.; Guther, M. L. S.; Milne, K. G.;
Ferguson, M. A. J. Proc. Natl. Acad. Sci. U.S.A.
2002, 99, 5884.
9. Roper, J. R.; Guther, M. L. S.; MacRae, J. I.; Prescott, A.
R.; Hallyburton, I.; Acosta-Serrano, A.; Ferguson, M. A.
J. J. Biol. Chem. 2005, 280, 19728.
10. Thoden, J. B.; Wohlers, T. M.; Fridovich-Keil, J. L.;
Holden, H. M. Biochemistry 2000, 39, 5691.
11. Shaw, M. P.; Bond, C. S.; Roper, J. R.; Gourley, D. G.;
Ferguson, M. A. J.; Hunter, W. N. Mol. Biochem.
Parasitol. 2003, 126, 173.
12. For details, see Online Mendelian Inheritance in Man,
Johns Hopkins University, USA, Number: 230350, URL:
<http://www.ncbi.nlm.nih.gov/omim//>.
13. Chapeau, M.-C.; Frey, P. A. J. Org. Chem. 1994, 59, 6994.
14. Winnans, K. A.; Bertozzi, C. R. Chem. Biol. 2002, 9, 113.
15. Vorgais, C. E.; Lemaire, H.-G.; Wilson, K. S. Protein
Expr. Purif. 1991, 2, 330.
16. De-Luca, C.; Lansing, M.; Crescenzi, F.; Martini, I.; Shen,
G.-J.; O’Regan, M.; Wong, C.-H. Bioorg. Med. Chem.
1996, 4, 131.
Nevertheless, T. brucei GalE is a validated drug target
for African sleeping sickness, and these results demon-
strate that the T. brucei enzyme can be selectively inhib-
ited by small molecules that are not substrate analogues
in vitro, although the mechanism of cytotoxicity of these
inhibitors is unconfirmed. Of the compounds screened,
ethacrynic acid displays the best therapeutic index, but
is clinically used as a loop diuretic and was found to
form multiple covalent adducts with TbGalE in vitro,
limiting its potential as a lead compound. However,
these results show that novel inhibitors of TbGalE
may be readily identified by screening small compound
libraries, demonstrating that high-throughput screening
of larger compound libraries of drug-like and lead-like
molecules against TbGalE using the methodology de-
scribed here is a feasible approach to the discovery of
drug leads for the treatment of African sleeping sickness.
17. GalE–UGD coupled assay: The reaction mixture
(100 mM glycine, pH 8.7, 0.5 mM b-NAD+, 0.2 mM
UDPGal, 0.2 mg/ml UGD and 80 ng/ml TbGalE) was
aliquoted into 96-well plates (198 ll/well) with or without
inhibitor (in 2 ll DMSO), and the reaction monitored at
340 nm for 15 min at 30 ꢁC on a SpectraMax 380
(Molecular Devices).
18. GalE HPAEC assay: The reaction mixture (1 mM Tris,
pH 7.6, 0.02% BSA, 250 lM UDP-Gal, 250 lM b-NAD+
and 2.5 g/mL TbGalE) was incubated at 30 ꢁC for 30 min
with or without inhibitor, quenched with 10-fold excess of
1 mM NaOH and then subjected to HPAEC chromatog-
raphy on a CarboPac PA-1 column (Dionex) using
conditions adapted from Tomiya et al. (Tomiya, N.,
Ailor, E., Lawerence, S. M., Bettenbaugh, M. J., Lee, Y.
C. Anal. Biochem. 2002, 293, 129). The eluant was
monitored at 260 nm, and peaks assigned by comparison
to commercial standards.
Acknowledgments
We thank Judith L. Fridovich-Keil of Emory University
School of Medicine, Atlanta, USA, for the P. pastoris
clones for expression of human GalE, Daan van Aaltan
of the University of Dundee, UK, for access to the
Prestwick Chemical library and Angela Melhert of the
University of Dundee, UK, for conducting the mass
spectrometry. This work was funded by a programme
grant from the Wellcome Trust (071463). JT wishes to
thank The Marine Science Programme at the University
of the South Pacific for sample collection, The Universi-
ty of the South Pacific, the Fiji Government and the
ORS (UK) for funding.
19. Shaw, M. P.; Bond, C. S.; Roper, J. R.; Gourley, D. G.;
Ferguson, M. A.; Hunter, W. N. Mol. Biochem. Parasitol.
2003, 126, 173.
References and notes
1. Cross, G. A. M. BioEssays 1996, 18, 283.
2. World, H. O. African Trypanosomiasis, Fact sheet number
259 ed.; WHO publications: Geneva, 2001.
3. Mehlert, A.; Zitzmann, N.; Richardson, J. M.; Traumann,
A.; Ferguson, M. A. J. Mol. Biochem. Parasitol. 1998, 91,
145.
4. Nolan, D. P.; Geuskens, M.; Pays, E. Curr. Biol. 1999, 9,
1169.
5. Atrih, A.; Richardson, J. M.; Prescott, A. R.; Ferguson,
M. A. J. J. Biol. Chem. 2005, 280, 865.
20. Lorber, D. M.; Shoichet, B. K. Protein Sci. 1998, 7, 938.
21. The reaction mixture (1 mM Tris, pH 7.6, 0.02% BSA and
100 lg/mL TbGalE) was incubated for 30 min with or
without inhibitor (50 lM) prior to analysis by MALDI-
TOF mass spectrometry.
22. Cytotoxicity assay: 200 microlitres of cells (1 · 104 cells/
ml) was aliquoted into 96-well plate and incubated with or
without inhibitor (100 mM to 30 nM) for a further three
days before the number of viable cells was counted using
the Alamar blue assay. For a description of the Alamar
blue assay, see: Raz, B.; Iten, M.; Grether-Buhler, Y.;
Kaminski, R.; Brun, R. Acta Trop. 1997, 68.
6. Tetaud, E.; Barrett, M. P.; Bringaud, F.; Baltz, T.
Biochem. J. 1997, 325, 569.