The first de novo designed inhibitors of Plasmodium falciparum dihydroorotate dehydrogenase

Article abstract of DOI:10.1016/j.bmcl.2005.09.045

The de novo molecular design program SPROUT has been applied to the X-ray crystal structures of Plasmodium and human dihydroorotate dehydrogenase, respectively. The resulting design templates were used to prepare a series of molecules which, in keeping wi

Full text of DOI:10.1016/j.bmcl.2005.09.045

Bioorganic & Medicinal Chemistry Letters 16 (2006) 88–92
The first de novo designed inhibitors of Plasmodium falciparum
dihydroorotate dehydrogenase
Timo Heikkila¨,a,b Srinath Thirumalairajan,a Matthew Davies,a Mark R. Parsons,b
A. Glenn McConkey,c Colin W. G. Fishwicka,* and A. Peter Johnsona,*
aSchool of Chemistry, University of Leeds, Leeds, England LS2 9JT, UK
bSchool of Biochemistry and Microbiology, University of Leeds, Leeds, England LS2 9JT, UK
cSchool of Biology, University of Leeds, Leeds, England LS2 9JT, UK
Received 27 May 2005; revised 17 September 2005; accepted 19 September 2005
Available online 19 October 2005
Abstract—The de novo molecular design program SPROUT has been applied to the X-ray crystal structures of Plasmodium and
human dihydroorotate dehydrogenase, respectively. The resulting design templates were used to prepare a series of molecules which,
in keeping with predictions, showed useful levels of species-selective enzyme inhibition.
Ó 2005 Elsevier Ltd. All rights reserved.
Malaria continues to represent a major threat to world
health infecting between 300 and 500 million people
annually and causing up to two million deaths.1 The dis-
ease results from infection by parasites belonging to the
Plasmodium species and is transmitted by the female
mosquitoes of the Anopheles genus. Of the four species
of parasite that infect humans, Plasmodium falciparum
is responsible for the majority of fatalities. Although
prophylactic treatments are available, widespread resis-
tance to commonly employed anti-malarial drugs (e.g.,
chloroquine,2 pyrimethamine3 and atovaquone4) is
widespread. There is, therefore, an urgent need for the
development of new drugs that can control infection
which can also act at previously unexploited biological
targets.
Dihydroorotate dehydrogenase (DHODH) is the
fourth enzyme in the pyrimidine biosynthetic pathway
and catalyses the oxidation of dihydroorotate (DHO)
to orotate in the presence of the co-factors flavin
mononucleotide (FMN) and ubiquinone (CoQ)
(Scheme 1). The human version of this enzyme
(hDHODH) is the target of a number of inhibitors
with proven efficacy in the treatment of arthritis and
leflunomide, a pro-drug that is metabolised to the ac-
tive DHODH inhibitor, A77-1726, is approved for clin-
ical use.5–9 Additionally, random high-throughput
screening of chemical libraries has been used to identify
selective inhibitors of Escherichia coli,10 Helicobacter
pylori11 and P. falciparum12 PfDHODH, respectively.
Additionally, Boa et al., have recently shown that
selective inhibitors of PfDHODH can be developed
from existing inhibitors.13 The PfDHODH enzyme is
thus an attractive target for the development of new
anti-malarial drugs.
De novo pyrimidine biosynthesis represents an attractive
and potentially selective target for the development of
new therapeutics against P. falciparum. Unlike human
cells, which can both synthesise and salvage pyrimidine
bases, P. falciparum lacks any pathway for the salvage
of preformed pyrimidine bases or nucleosides and relies
completely on a de novo biosynthesis pathway.
As part of a continuing structure-based synthesis and
biological evaluation program for the discovery of new
enzyme inhibitors and receptor antagonists, respective-
ly, we have previously described the computer-aided
molecular design, synthesis and biological evaluation
of novel inhibitors of the bacterial enzyme MurD,14
and of the human NK2 receptor,15, respectively. Here,
we report the application of our de novo molecular de-
sign program SPROUT16 to the design of novel inhibi-
tors of PfDHODH.
Keywords: Medicinal chemistry; Oxoreductases; Inhibitors; Molecular
recognition; Molecular modelling.
*
Corresponding authors. Tel.: +44 113 3436510; fax: +44 113
3436565; e-mail addresses: colinf@chem.leeds.ac.uk; a.p.johnson@
chem.leeds.ac.uk
0960-894X/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.09.045

T. Heikkila¨ et al. / Bioorg. Med. Chem. Lett. 16 (2006) 88–92
89
O
O
HN
HN
O
N
H
CO2
orotate
O
N
H
CO2
DHO
R
N
R
N
H
N
O
Me
Me
N
O
Me
Me
NH
NH
N
N
H
O
O
FMN
FMNH2
O
O
OH
MeO
MeO
Me
MeO
Me
H
10
H
MeO
10
Me
OH
Me
CoQ
CoQH2
Scheme 1. Reactions catalysed by DHODH.
SPROUT16 is a powerful suite of software modules for
de novo structure-based molecular design. SPROUT
has modules for (i) characterising regions within a pro-
tein, (ii) detecting ÔhotspotsÕ where ligand atoms would
form favourable interactions with the protein, (iii) dock-
ing molecular fragments to these sites, (iv) joining these
fragments with a molecular backbone to give a complete
ligand and (v) ranking the set of predicted ligands by cri-
teria including predicted binding energy, structural com-
plexity and synthetic accessibility.
In order to use a de novo design approach for the dis-
covery of novel PfDHODH inhibitors, we have applied
SPROUT to the recently solved X-ray crystal structure
of PfDHODH the details of which are published else-
where.17 The human and Plasmodium enzymes share a
very similar structural morphology. The previously
reported18 X-ray crystal structure of hDHODH, co-
crystallised with the inhibitor A77-1726, revealed the
presence of an N-terminal a-helical domain that forms
a channel leading to the active site (which contains both
FMN and orotate) and that is thought to facilitate bind-
ing of CoQ. This structure also features a molecule of
A77-1726 bound within this putative Ôubiquinone chan-
nelÕ which makes H-bonding contacts to R136 and
Y356, respectively (Fig. 1).
Figure 1. X-ray structure of human DHODH.
the same region within the Plasmodium-derived enzyme is
much less congested and, unlike that in the human en-
zyme, appears to be able to accommodate inhibitors that
are somewhat cylindrical in overall shape. With this dif-
ference in mind, we wished to explore the possibility of
using SPROUT to design species-selective inhibitors
targeted at this channel.
The X-ray crystal structure of PfDHODH, also contain-
ing a molecule of A77-1726 in the ubiquinone channel,
features analogous positioning of the FMN and orotate
molecules to those found in the human enzyme (Fig. 2).
In order to design structurally simple inhibitors, for
which good affinity for PfDHODH was predicted,
we specified that in addition to use of the hydropho-
bicity within the channel, just two residues, histidine
185 and arginine 265, were to make direct H-bonding
contacts to the designed inhibitors. Both of these res-
idues are highly conserved in all type II dihydrooro-
tate dehydrogenases of eukaryotic organisms studied
to date.
A detailed comparison of the two structures however re-
veals subtle differences, particularly in the dimensions
and topography of the hydrophobic ubiquinone chan-
nels. In particular, the channel within the human en-
zyme is considerably ÔflattenedÕ by the protrusion of a
methyl group from A59 in the region occupied by the
aromatic ring of the bound inhibitor (Fig. 3).
Therefore, the shape of this cavity appears to require
inhibitors of this type to be essentially planar. In contrast,
SPROUT-based de novo ligand design produced
some 20 different small molecule templates. The most

90
T. Heikkila¨ et al. / Bioorg. Med. Chem. Lett. 16 (2006) 88–92
Figure 2. Overlay of X-ray structures of human and Plasmodium
DHODH (Plasmodium structure and contacting residues numbered
and coloured in red; distances of ligand to R265 and H185 indicated).
Figure 4. A SPROUT-designed template within the Ôubiquinone
channelÕ of Plasmodium DHODH.
Figure 5. Conformations of designed inhibitor template.
In order to test this hypothesis, we prepared19 a small
number of amides based upon this general design, which
were either N-unsubstituted (R = H) to provide systems
predominantly populating conformation 1 (via intramo-
lecular H-bonding between the amide NH and carboxyl-
ic acid group), or N-methylated (R = Me) and therefore,
preferring to exist predominantly as conformation 2.
Figure 3. Surfaces of ubiquinone cavities within Plasmodium DHODH
(top) and human DHODH (bottom), coloured according to electro-
static potential.
The binding affinities of these for both Plasmodium and
human DHODH were then established from IC50 values
as described20,21 and are summarised below (Table 1).
As is evident from the data in Table 1, four of the six
inhibitors are active against the Plasmodium enzyme
whereas only the N-unsubstituted systems (entries 2, 4
and 6) show affinity to hDHODH.
attractive of these, for which predicted binding affini-
ties were in the micromolar range or better, consisted
of simple and readily prepared amides of anthranilic
acid, an example of which is shown in Figure 4. As
noted previously, the differing shapes of the ubiquinone
cavities within the Plasmodium and human enzymes
suggest that inhibitors of this type would fit better
within the Plasmodium—as opposed to the human en-
zyme, if they adopt a non-planar arrangement between
the amide unit, attached aryl groups and the carboxylic
acid moiety (2, Fig. 5). This twisting appears to maxi-
mise the interaction of the aromatic rings of the de-
signed inhibitors with the walls of the hydrophobic
cavity. In contrast, planar versions of this type of
inhibitor (1, Fig. 5) are predicted to interact best with
the human enzyme, where the somewhat flattened nat-
ure of the cavity allows close contacting with both fac-
es of the aromatic portions of these inhibitors.
Additionally, in keeping with our design criteria, N-
methyl systems (entries 1 and 3) exhibit a higher affinity
for the Plasmodium enzyme than those found for the N-
unsubstituted cases (entries 2 and 4). Curiously, whereas
the bromonaphthyl-derived N-methyl amide (entry 3)
was active towards PfDHODH, the simpler naphthyl-
derived N-methyl amide (entry 5) was found to be inac-
tive towards both enzymes.
In conclusion, the de novo design program SPROUT
has been used to produce simple molecular templates
for the selective inhibition of Plasmodium and human

T. Heikkila¨ et al. / Bioorg. Med. Chem. Lett. 16 (2006) 88–92
91
Table 1. Structures, IC50 and apparent binding affinities21 (lM) of designed inhibitors with PfDHODH and hDHODH
Entry
Structure
IC50 (Pf)
IC50 (h)
Kai pp (Pf)
Kai pp (h)
HO C
2
Me
1
N
42.6 (4.6)
>200
4.9
n.a.
O
HO2C
H
N
2
3
4
5
6
153.5 (13.2)
93.4 (6.4)
142.6 (14.2)
>200
5.0 (1.6)
>200
17.7
10.8
16.4
n.a.
n.a.
0.7
O
HO2C
Me
Br
Br
N
n.a.
1.1
O
HO2C
H
N
8.4 (2.7)
>200
O
HO2C
N
Me
n.a.
1.8
O
HO2C
H
N
>200
13.8 (3.3)
O
Values in brackets refer to error limits; n.a. = not attempted.
5. Fox, R. I.; Hermann, M. L.; Frangou, C. G.; Wahl, G.
M.; Morris, R. E.; Strand, V.; Kirschbaum, B. J. Clin.
Immunol. 1999, 93, 198.
6. Hermann, M. L.; Schleyerbach, R.; Kirschbaum, B. J.
Immunopharmacology 2000, 47, 273.
DHODH, respectively. Subtle differences in the struc-
tures within the targeted binding regions of these en-
zymes, which might provide the basis of this
selectivity, have been identified.
7. Williamson, R. A.; Yea, C. M.; Robson, P. A.; Curnock,
A. P.; Gadher, S.; Hambleton, A. B.; Woodward, K.;
Bruneau, J.-M.; Hambleton, P.; Moss, D.; Thompson, T.
A.; Spinella-Jaegle, S.; Morand, P.; Courtin, O.; Sautes,
C.; Westwood, R.; Hercend, T.; Kuo, E. A.; Ruuth, E. J.
Biol. Chem. 1995, 270, 22467.
8. Greene, S.; Watanabe, S.; Braatz-Trulson, J.; Lou, L.
Biochem. Pharmacol. 1995, 50, 861.
9. Davis, J. P.; Cain, G. A.; Pitts, W. J.; Magolda, R. L.;
Copeland, R. A. Biochemistry 1996, 35, 1270.
10. Marcinkeviciene, J.; Rogers, M. J.; Kopcho, L.; Jiang,
W.; Wang, K.; Murphy, D. J.; Lippy, J.; Link, S.;
Chung, T. D. Y.; Hobbs, F.; Haque, T.; Trainor, G. L.;
Slee, A.; Stern, A. M.; Copeland, R. A. Biochem.
Pharmacol. 2000, 60, 339.
In these proof-of-principal studies, the designed inhibi-
tors show only modest enzyme affinities, but provide
an excellent starting point for further optimisation. We
believe that these results further illustrate the tremen-
dous potential of a de novo design-based approach in
inhibitor discovery. Such an approach is complementary
to the use of high-throughput screening and is particular-
ly attractive where such screening methodology is not
available or where access to large collections of library
compounds of sufficient molecular diversity is limited.
Acknowledgments
11. Copeland, R. A.; Marcinkeviciene, J.; Haque, T. S.;
Kopcho, L. M.; Jiang, W.; Wang, K.; Ecret, L. D.;
Sizemore, C.; Amsler, K. A.; Foster, L.; Tadesse, S.;
Combs, A. P.; Stern, A. M.; Trainor, G. L.; Slee, A.;
Rogers, M. J.; Hobbs, F. J. Biol. Chem. 2000, 275, 33373.
12. Baldwin, J.; Michnoff, C. H.; Malmquist, N. A.; White, J.;
Roth, M. G.; Rathod, P. K.; Phillips, M. A. J. Biol. Chem.
2005, 280, 21847.
The authors gratefully acknowledge the financial sup-
port provided by the Overseas Research Scholarship
fund (to S.T.), UK Engineering and Physical Sciences
Research Council and the Wellcome Trust (studentship
to T.H.).
References and notes
13. Boa, A. N.; Canavan, S. P.; Hirst, P. R.; Ramsey, C.;
Stead, A. M. W.; McConkey, G. A. Bioorg. Med. Chem.
2005, 13, 1945.
1. WHO report, 2002.
2. Wellens, T.; Plowe, C. J. Infect. Dis. 2001, 184, 770.
3. Wu, Y. Proc. Natl. Acad. Sci. U.S.A. 1996, 93, 1130.
4. Srivastava, S. Mol. Microbiol. 1999, 33, 704.
14. Horton, J. R.; Bostock, J. M.; Chopra, I.; Hesse, L.;
Phillips, S. E. V.; Adams, D. J.; Johnson, A. P.; Fishwick,
C. W. G. Bioorg. Med. Chem. Lett. 2003, 13, 1557.

92
T. Heikkila¨ et al. / Bioorg. Med. Chem. Lett. 16 (2006) 88–92
15. Ali, M. A.; Bhogal, N.; Fishwick, C. W. G.; Findlay, J. B.
C. Bioorg. Med. Chem. Lett. 2001, 11, 819.
16. Gillet, V. J.; Newell, W.; Mata, P.; Myatt, G.; Sike, S.;
Zsoldos, Z.; Johnson, A. P. J. Chem. Inf. Comput. Sci. 1994,
34, 207.
Murphy, K. B.; Patterson, T. A. Arch. Biochem. Biophys.
1995, 323, 79) by the standard DCIP assay, using
saturating concentrations of DHO (200 lM) and decyl-
ubiquinone (CoQd), (100 lM) and a range of inhibitor
concentrations (1 nM to 200 mM). IC50 values were
derived from the data by using Graphpad Prism v. 4.0
(Graphpad Inc.).
17. Hurt, D. E.; Widom, J.; Clardy, J. Acta Crystallogr.,
submitted for publication.
18. Liu, S.; Neidhardt, E. A.; Grossman, T. H.; Ocain, T.;
Clardy, J. Structure 2000, 8, 25.
19. Prepared via coupling of methyl 2-amino benzoate with
the appropriate aryl carboxylic acid chlorides followed by
ester hydrolysis.
21. The IC50 values were converted into Ki values by using the
Cheng–Prusoff equation (Cheng, Y.; Prusoff, W. H.
Biochem. Pharmacol. 1973, 22, 3099) assuming competi-
tive inhibition with CoQd (Km values used for CoQd were
13 lM for PfDHODH and 15 lM for HsDHODH,
respectively).
20. The IC50 values were determined as described (Copeland,
R. A.; Davis, J. P.; Dowling, R. L.; Lombardo, D.;