Exploring new inhibitors of Plasmodium falciparum purine nucleoside phosphorylase

Article abstract of DOI:10.1016/j.ejmech.2010.08.026

Plasmodium falciparum purine nucleoside phosphorylase (PfPNP) has a central role in purine salvage and inhibitors of the enzyme have been shown to have antiplasmodial activity. The enzyme preferentially uses inosine as substrate (K_m = 5 μM, k

Full text of DOI:10.1016/j.ejmech.2010.08.026

European Journal of Medicinal Chemistry 45 (2010) 5140e5149
Contents lists available at ScienceDirect
European Journal of Medicinal Chemistry
journal homepage: http://www.elsevier.com/locate/ejmech
Original article
Exploring new inhibitors of Plasmodium falciparum purine nucleoside
phosphorylase
,
,
*
Huaqing Cui a b, Gian Filippo Ruda b, Juana Carrero-Lérida a, Luis M. Ruiz-Pérez a, Ian H. Gilbert b
**
,
Dolores González-Pacanowska a
,
a Instituto de Parasitología y Biomedicina “López-Neyra”, Consejo Superior de Investigaciones Científicas, Avda. del Conocimiento s/n, 18100-Armilla, Granada, Spain
b College of Life Sciences, University of Dundee, Sir James Black Centre, Dundee DD1 5EH, UK
a r t i c l e i n f o
a b s t r a c t
Article history:
Plasmodium falciparum purine nucleoside phosphorylase (PfPNP) has a central role in purine salvage and
inhibitors of the enzyme have been shown to have antiplasmodial activity. The enzyme preferentially
Received 26 March 2010
Received in revised form
9 August 2010
Accepted 10 August 2010
Available online 14 August 2010
uses inosine as substrate (Km ¼ 5
m
M, kcat/Km ¼ 7.4 ꢀ 104
M
ꢁ1 sꢁ1), but can also use uridine, albeit less
efficiently (Km ¼ 85
m
M, kcat/Km ¼ 306 Mꢁ1 sꢁ1). In an effort to identify new PfPNP inhibitors, two series of
compounds were prepared. Series 1 was based on known human uridine phosphorylase inhibitors whilst
series 2 was uracil equivalents of purine-based PNP transition state inhibitors. These two series of
compounds were assayed for inhibition of both PfPNP activity and growth of P. falciparum. The transition
Keywords:
Malaria
Purine nucleoside phosphorylase
Uridine phosphorylase
Transition state inhibitor
Drug discovery
state analogues were found to be moderate inhibitors of PfPNP (most potent compound, Ki ¼ 6
mM).
Ó 2010 Elsevier Masson SAS. All rights reserved.
1. Introduction
inosine (Km ¼ 5
m
M, kcat/Km ¼ 7.4 ꢀ104 Mꢁ1 sꢁ1) and the catalytic
efficiency is 240-fold higher with the latter [5].
It is estimated that malaria causes 300e500 million clinical
cases and 2 million deaths annually [1]. Plasmodium falciparum is
responsible for the majority of deaths due to malaria. Drug resis-
tance is a major problem for treatment and there is a need for new
drugs particularly those that have a new mode of action [2,3].
P. falciparum cannot synthesize purines de novo, and is entirely
reliant on the salvage of extracellular purines. Therefore, the purine
salvage pathway is a potential target for antimalarial chemotherapy
[4]. An enzyme involved in this pathway is P. falciparum purine
nucleoside phosphorylase (PfPNP), which catalyzes the phospho-
rolysis of inosine to ribose-1-phosphate and hypoxanthine (Fig. 1).
Uridine is also known to be a substrate for this PfPNP, although the
PfPNP is essential for parasite survival. Thus targeted gene
disruption in a P. falciparum strain in order to ablate PNP expression
has resulted in parasites that exhibit growth defects at physiolog-
ical concentrations of hypoxanthine and require exogenous purines
for sustained growth [9]. Chemical validation of PfPNP has also
been performed and transition state analogues developed that
inhibit hPNP and PfPNP and kill P. falciparum in vitro [10].
Immucillin-H (3) (Fig. 2) is a PNP transition state analogue
inhibitor, which inhibits hPNP very potently (Kd ¼ 56 pM) and the
Plasmodium enzyme PfPNP (Ki ¼ 860 pM) slightly less strongly
[11]. Immucillin-H was designed using kinetic isotope measure-
ments which allowed for the identification of the transition state
[7,12]. The study indicated that catalysis proceeds via a classic SN1
nucleophilic substitution reaction in which, at transition state, the
ribose ring forms an oxocarbenium ion (Fig. 1) [8,13]. In these
transition state inhibitors, oxygen is replaced by a basic nitrogen,
which mimics the positive charge of the oxocarbenium ion, and
the distance between the ribose and the base in Immucillin-H is
affinity (Km ¼ 85
m
M, kcat/Km ¼ 306 Mꢁ1 sꢁ1) is lower than for
Abbreviations: ADA, adenosine deaminase; BBB, 5-(m-benzyloxy)benzylbarbi-
turic acid; BBBA, 1-[(2-hydroxyethoxy)methyl]-5-(m-benzyloxy)benzylbarbituric
acid acyclonucleoside; NOESY, nuclear Overhauser effect spectroscopy; PfPNP,
Plasmodium falciparum purine nucleoside phosphorylase; MT-immH, 50-methyl-
thio-immucillin-H; UP, uridine phosphorylase.
ꢀ
1.5 A, which is very close to the distance between the ribose and
the leaving base in the transition state [13]. The substrate speci-
ficity, catalytic site structure, and subunit structure of PfPNP are
distinct from hPNP [6]. Potentially this can be used to design
inhibitors which are selective for PfPNP over hPNP. However, most
* Corresponding author.
** Corresponding author. Tel.: þ34 958181631; fax: þ34 958181632.
E-mail addresses: i.h.gilbert@dundee.ac.uk (I.H. Gilbert), dgonzalez@ipb.csic.es
(D. González-Pacanowska).
0223-5234/$ e see front matter Ó 2010 Elsevier Masson SAS. All rights reserved.
doi:10.1016/j.ejmech.2010.08.026

H. Cui et al. / European Journal of Medicinal Chemistry 45 (2010) 5140e5149
5141
Fig. 1. Inosine phosphorylase catalyzed by purine nucleoside phosphorylase. The reaction is an SN1 type mechanism [6e8].
of the reported transition state inhibitors bind preferentially to
2. Chemistry
hPNP rather than PfPNP [5,10]. There are several exceptions to this:
for example MT-immH (50-methylthio-immucillin-H) gives the
largest selectivity between PfPNP and hPNP so far reported, a 100-
fold difference between PfPNP (Ki ¼ 2.7 nM) and hPNP
(Ki ¼ 303 nM) [14]; although MT-immH still shows nanomolar
potency against hPNP. It is proving challenging to obtain
compounds that are highly selective towards PfPNP. However, it
should be noted that the antiparasitic activity of these PNP
inhibitors may in part be due to simultaneous inhibition of hPNP,
preventing generation of hypoxanthine by the erythrocytes
[5,9,10].
2.1. Series 1
The first series of compounds prepared were barbituric acid
derived inhibitors of human uracil phosphorylase. 5-(m-Benzyloxy)
benzylbarbituric acid (BBB, 1) and 1-[(2-hydroxyethoxy)methyl]-5-
(m-benzyloxy)benzylbarbituric acid (BBBA, 2) were prepared as
shown in Scheme 1. Barbituric acid (6) was coupled to benzaldehyde
or m-benzyloxybenaldehyde to give the derivatives 7 and 8
respectively [20]. Products 7 and 8 were used crude as they con-
tained more than 80% product and the removal of unreacted alde-
In the study we report here, two series of compounds have been
designed as potential inhibitors of PfPNP. In particular, the fact that
PfPNP can use uridine as a substrate has been exploited.
hydes proved problematic. The a, b-enone was reduced with sodium
borohydride. Work was then carried out to introduce the side chain.
Attempts to react compound 1 with 2-acetoxyethyl acetoxymethyl
ether were unsuccessful [21]. However the reactionwas successfully
completed using (2-acetoxyethoxy)methyl bromide following
a literature procedure [22]. (2-Acetoxyethoxy)methyl bromide 12
was prepared [23] by stirring acetyl bromide and 1,3-dioxolane
together at room temperature, and then separating the required
product by distillation at 113 ꢂC. Compound 1 was first silylated
using bis(trimethylsilyl)acetamide and then stirred with the (2-
acetoxyethoxy)methyl bromide to give the required ester 13. The
ester 13 was deprotected to give the final product 2.
First, human uridine phosphorylase (UP) has been considered as
target for cancer therapy and many nucleoside analogues have been
evaluated as inhibitors of the enzyme [15e17]. As PfPNP shows
significant activity with uridine and similarity to human uridine
phosphorylase, a potent human uridine phosphorylase inhibitor
5-(m-benzyloxy)benzylbarbituric acid acyclonucleoside (BBBA 2,
Ki ¼ 1.1 nM) [15] and its intermediate-5-(m-benzyloxy)benzylbar-
bituric acid (BBB 1, Ki ¼ 2.3
mM) [15] were made to test for activity
against PfPNP (Fig. 2). If active and selective, these could provide
a starting point for a more focused discovery programme.
Second, different generations of purine-based transition state
inhibitors of human and Plasmodium PNP have been reported
[11,13,18]. Compound 4 (Fig. 2) belongs to a simplified series of
purine-based PNP transition state inhibitors [19] and has a Kd of
5.5 nM for hPNP, although binding is weaker than that of immu-
cillin-H 3 (Kd ¼ 6 pM). In contrast, the substrate inosine has a Kd of
10,000 nM for hPNP [19]. We were interested to replace the purine
base of the transition inhibitors with a uracil base to generate a new
series of compounds and investigate if they show activity against
PfPNP. Therefore uracil containing derivatives of compound 5 were
designed, prepared and tested for inhibition against PfPNP.
Compound 5 is the general structure for the proposed transition
state inhibitor with the purine replaced by uracil. In our study, we
prepared a series of inhibitors (18e22 and 24, 25) which exhibit an
alkyl chain length of two to four carbon atoms between these two
nitrogen atoms.
2.2. Series 2
The second series of inhibitors, compounds 18e22 and 24, 25,
were prepared as shown in Scheme 2. Uracil was stirred with
1,2-dibromoethane, 1,3-dibromopropane or 1,4-dibromobutane [Br
(CH2)nBr (n ¼ 2e4)] in Cs2CO3 in DMF at 40 ꢂC for 2 days. Chro-
matographic purificationwas used to separate the final products. The
dibromoalkanes could alkylate either N1 or N3 of the uracil. In order
to investigate this, NOESY spectrawere run on compounds 15e17 and
23. Coupling could be observed for compounds 15e17 between the
proton of N6 and the protons of alkyl chain (H10 and H20), but not for
compound 23 (Fig. 3). Therefore, in compound 23, alkylation had
occurred on the undesired N3. Compounds 15e17 and 23 were then
reacted with pyrrolidine or l-prolinol in ethanol at 150 ꢂC under
microwave conditions to give the final compounds 18e22, 24 and 25.
Fig. 2. Four selected known inhibitors and one proposed new inhibitor: (1) 5-(m-benzyloxy)benzylbarbituric acid, BBB; (2) 5-(m-benzyloxy)benzylbarbituric acid acyclonucleoside,
BBBA; (3) Immucillin-H; (4) Immucillin-H analogue; (5) Proposed new inhibitors, based on compound 4.

5142
H. Cui et al. / European Journal of Medicinal Chemistry 45 (2010) 5140e5149
Scheme 1. The synthetic pathway for BBB 1 and BBBA 2. Reagents and conditions: (a) benzaldehyde or m-benzyloxybenzaldehyde, acetic acid, reflux, RT, 85e90%; (b) NaBH4, EtOH,
2 h, RT, 90%; (c) reflux, 70 ꢂC, 29%; (d) (i) chloromethylsilane, hexamethyldisilazane, reflux, 80 ꢂC, overnight, Ar; (ii) (2-acetoxyethoxy)methyl bromide, aluminium chloride, 1,2-
dichloroethane, 48 h, RT, Ar, 13%; (e). 0.2 M NaOMe in MeOH, RT, 71%.
Scheme 2. The synthetic pathway of uracil analogues. Reagents and conditions: (a) Br(CH2)nBr (n ¼ 2-4), CsCO3, DMF, 40 ꢂC, 48 h, 13e24%; (b) Pyrrolidine or l-Prolinol, ethanol,
150 ꢂC, 1 min, microwave, 33e41%.

H. Cui et al. / European Journal of Medicinal Chemistry 45 (2010) 5140e5149
5143
M)
Table 1
3. Results and discussion
Evaluation of compounds of series 1 on PfPNP activity.
3.1. Inhibition of PfPNP
Compounds
PfPNP IC50
(
m
M)
hUP Ki (m
1
2
9
>1000
>1000
>1000
2.3a
0.0011b
NRc
Recombinant PfPNP protein [5] was expressed as a soluble
protein in Escherichia coli BL21 (DE3) cells. Purified recombinant
PfPNP appeared as a single band with a subunit molecular weight of
approximately 29 kDa in SDSePAGE. PfPNP assays were performed
as previously described [5,6,14]. Kinetic parameters of the His6
tagged enzyme were similar to those previously reported [5]. The
Km and kcat values were obtained by a non-linear regression fit of
the data to the MichaeliseMenten equation (Supplementary data).
a
From reference [16].
From reference [17].
Not reported.
b
c
substituents of the barbituric acid might affect this hydrogen bond
network and therefore perturb the stabilization of the loop domain.
PNP structures also show binding interactions to the ribose ring (for
PfPNP, 50-OH group with His7b from subunit B and Glu184 with the
20 and 30 hydroxyl groups). However, compound 4 is active against
human PNP, and does not establish riboseePNP interactions, so
these may not have a significant role in its effective inhibition.
Compounds 18e22, 24 and 25 were designed to mimic the
transition state inhibitors of PNPs. The PfPNP inhibition assay
revealed that they showed reasonable inhibitory activities against
The values found for inosine were 5.9 m
M and 2.1 sꢁ1 respectively
(Supplementary data). Previous reported values for kcat range from
0.34 to 1.7 sꢁ1 whilst the reported Km value for inosine is 5
mM
(Kicska et al. [5] and Lewandowicz et al. [11]).
Compounds 1, 2, 9, 18e22, 24 and 25 were tested for inhibition
of PfPNP. The three human UP inhibitors e 1, 2, 9 e (Table 1)
showed no inhibition of PfPNP even at 1000 mM. Compound 2 was
reported to have a Ki against human UP of 1.1 nM [17]. The binding
mechanism of these selected human UP inhibitors to human UP has
not been elucidated, but was studied in E. coli UP [17]. The crystal
structure revealed that the large 5-substituent on the barbituric
acid could be enclosed in a hydrophobic pocket of E. coli UP and the
benzyloxy moiety could form additional hydrophobic interactions
with Met 234 [17].
The structure of PfPNP was analysed to try and explain the lack
of activity of the UP inhibitors. The nucleoside binding site of PfPNP
is a hydrophobic pocket where the inosine base is oriented by
PfPNP, with Ki values ranging from 6
particular compounds 24, 18 and 19 showed good inhibition
(Ki ¼ 6 M, 27 M and 59 M respectively).
mM to 136 mM (Table 2). In
m
m
m
In order to try and interpret our results, we carried out some
molecular modelling. We modeled the inhibitors into the active site
of the PfPNP (Fig. 4). Compounds 18 and 24 were superimposed
within the active site of the protein 1NW4 [6] (active site between
chains A & B), by superposition of the uracil ring and the purine ring
of Immucillin-H. The uracil ring was superimposed on the hypo-
xanthine ring of the Immucillin-H in order to maintain, as far as
possible, the H-bonding network, which is almost certainly driving
much of the binding energy (Table 3). The pyrrolidine ring could
then extend from the base-binding site towards Ser91. This is the
residue that binds with the basic nitrogen in Immucillin-H and
presumably also the pyrrolidine ring in compound 4. Hydrogen
p-stacking and van der Waals interactions [4]. The flexible benzy-
loxy moiety of the human UP inhibitors tested may form steric
clashes with the binding pocket, if the barbituric acid core binds in
a similar manner to purine (Fig. 4E). Additionally, there is a flexible
loop in PfPNP which is stabilized by a network of hydrogen bonds
with the purine moiety [4]. The large flexible benzyloxy
Fig. 3. The identification of compounds 15e17 with 23 by NOESY NMR. The coupling between the H6 and the proton of the alkyl chains (H10 or H20) is marked in the red circle. (For
interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article).

5144
H. Cui et al. / European Journal of Medicinal Chemistry 45 (2010) 5140e5149
Fig. 4. Computer modelling study of 18 (yellow) and 24 (blue) in the active centre of PfPNP. A. Crystal structure of immucillin-H (magenta) in the active site of PfPNP [6], the
hydrogen bonds are shown in the figure. B. The predicted binding mode of compound 18 (yellow) in the active site of PfPNP; C. The predicted binding mode of compound 24 (blue)
in the active site of PfPNP; D. Comparison of the predicted binding modes of compounds 18 (yellow), 24 (blue) to Immucillin-H (magenta) in the active centre of PfPNP. The distance
ꢀ
between C4 and basic N in immucillin-H was measured to be 3.6 A. E. The human UP inhibitor BBBA (2) (green) was superimposed in the active centre of PfPNP based on the similar
binding mode as Immucillin-H (light blue), and the barbituric acid core occupied the position of the purine base. However, the BBBA showed a steric crash in the active centre of
PfPNP. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article).
atoms were added to the protein residues and the ligand, and the
enzymeeligand structures were optimized using the Moloc force
field [24,25]. For optimization the protein was kept rigid, except for
the hydrogens, the ligands and the water molecules within the
active site, which were considered flexible (Fig. 4A and B). Whilst
this model is speculative, in the absence of structural confirmation,
it provides a reasonable explanation of the kinetic data.
The modelling showed that both compounds could fit
adequately in the active site and that there is some potential for
further inhibitor design. In both cases the uracil base is predicted to
be oriented in the nucleoside binding pocket by a p-stacking
similar to the purine base of Immucillin-H. The N1 of the uracil base
in compound 18 and the N3 of the uracil base in compound 24
roughly superimpose with the C4 of Immucillin-H (Fig. 4C). Addi-
tionally the basic nitrogen of the pyrrolidine rings in immucillin-H
and compounds 18 and 24 appeared very close to each other. The
interactions between the ligands and the PfPNP were compared
and are shown in Table 3 and Fig. 4. The interactions between
Immucillin-H and PfPNP are from the crystal structure, and for
compounds 18 and 24, from our model.
Table 3
Table 2
Interactions between the different compounds and PfPNP.
Evaluation of compounds of series 2 on PfPNP and Plasmodium falciparum growth
in vitro.
Immucillin-H
Compound 18
Compound 24
Compounds
Log Pa
tPSAa
PfPNP Ki
(IC50) (
EC50
(m
M)
Atom
distance (A)
H-bond to N7 from D206
H2O mediated
H-bond to
H2O mediated
H-bond to
b
ꢀ
m
M)
O4 from D206
H-bond to N3
from H2O
O4 from D206
H-bond to N1
from H2O
18
19
20
21
22
24
25
ꢁ0.4
ꢁ0.3
ꢁ0.8
0.2
ꢁ0.4
ꢁ0.4
ꢁ0.9
53
53
73
53
73
53
73
27 (141)
59 (230)
136 (712)
105 (524)
113 (592)
6 (42)
>500
>1000
375
>1000
>1000
>500
3.7
4.8
4.8
6.1
6.1
3.7
3.7
H-bond to 50 OH from H7b
H-bond to 20 OH from M183
H-bond to 20 OH from E184
H-bond to 30 OH from E184
H-bond to basic N from S91
H2O mediate H-bond to
N7 from 50 OH
75 (366)
>1000
Chloroquine
0.03
a
Calculated log P and tPSA by the software of Chem Draw 12.0.
H-bond to O6 from H2O
H-bond to N1 from H2O
b
Calculated atom distance between the N-cationic centre and the N-1 (18e22) or
N-3 (24 and 25) atom of the uracil base by Chem Draw 12.0.

H. Cui et al. / European Journal of Medicinal Chemistry 45 (2010) 5140e5149
5145
ꢀ
ꢀ
Fig. 5. The distance between the N-cationic centre and the nitrogen of the uracil base in various inhibitors of this study. The distance measured is as follows 18 (3.7 A), 19 (4.8 A), 21
ꢀ
ꢀ
(6.1 A), 24 (3.7 A). The distance measured for other compounds was 20 ¼ 19, 22 ¼ 21, 25 ¼ 24. Measurements were performed using Chem Draw 12.0.
It is thought that the key factors affecting the binding of the
transition state analogues to PfPNP are: (i) the presence of the
N-cationic centre of the pyrrolidine ring which binds to and mimics
the intermediate oxocarbenium ion; (ii) the relative orientation of
the N-cationic centre and the purine ring e the reported optimum
distance between the 7-position of the purine and the N appears to
be in the range 1.5e3.0 A; (iii) the presence of the hydroxymethyl
substituent on the pyrrolidine ring which can H-bond with the
PfPNP [13,19].
could be due to a number of reasons. One possibility is low cellular
permeability, as the compounds are calculated to have a low log P.
Alternatively the level of enzyme inhibition is insufficient to effi-
ciently deplete PNP activity intracellularly. Furthermore it is not
clear if inhibition of human PNP is also required for antiparasitic
activity [5,10], as, at least in some circumstances, Plasmodium
parasites appear to also exploit host enzymes for the conversion of
adenosine and inosine into hypoxanthine before translocation
across the parasite plasma membrane. In this study, our
compounds may not be active against hPNP.
ꢀ
We therefore decided to compare the distances between the
uracil ring and N-cationic centre of the pyrrolidine ring with
activity. The distance was measured from N1 for the N1-substituted
derivatives (18e22) and from the N3 for the N3-subtituted deriv-
atives (24 and 25) (Fig. 5 and Table 3). Our modelling showed that
the N1 of the uracil base in compound 18 and the N3 of the uracil
base in compound 24 roughly superimposed with the C4 of
Immucillin-H (Fig. 4C). The corresponding distance for immucillin-
4. Conclusion
PfPNP has been shown to present unique properties and
inhibitors of the Plasmodium enzyme exhibit antiparasitic activity
[5,6,10]. Based on the observation that uridine can be a substrate of
the enzyme, we have tested several human uridine phosphorylase
inhibitors against PfPNP and designed several new uracil based
analogues to mimic the PNP transition state. The results show that
while human UP inhibitors are inactive against PfPNP, the uracil
based transition state analogues gave significant inhibition with Ki
values in the micromolar range. However, these compounds do not
have significant antiparasitic activity and therefore are not suitable
for further development as antimalarials.
ꢀ
H is from the 4-position of the purine ring and is 3.6 A. There
appears to be some relationship between this distance and inhi-
bition of PfPNP. Thus the most potent inhibitors were 18 and 24,
which had a very similar distance to that of immucillin-H; this is
also seen in the overlap of predicted inhibitor conformations
(Fig. 4D).
The orientation and the presence of hydroxyl groups of the
pyrrolidine ring is another factor that affects activity. Analysis of
crystal structures of PNPs in general indicates that there are
H-bonding interactions between the ribose ring and the protein.
The crystal structure of PfPNP-Immucillin-H revealed that the 50
hydroxyl group of the ribose forms a hydrogen bond with the
His7b [4] and the 20- and 30-OH groups bind with Glu184 [6]. This
is exemplified in compound 4 (Kd of 5.5 nM for hPNP); addition of
a hydroxymethyl group to the 50-position and a hydroxyl to the
30-position, gives a compound with Kd of 8.5 pM for hPNP [26].
Interestingly in the case of the PfPNP, the crystal structure shows
a cavity around the 50-hydroxyl group, which is not present in the
corresponding human enzyme, which may indicate possibilities
for selective inhibition. In our case, the hydroxymethyl group
actually slightly decreases activity when compared 24 to 25 and 19
to 20.
5. Experimental procedures
5.1. Enzyme purification and inhibition assays
The coding sequence of PfPNP was cloned in pET 28a and
expression was performed in E. coli BL21 (DE3) cells, by induction
with 1 mM IPTG at 37 ꢂC for 4 h. Cell pellets were suspended in
buffer A (50 mM NaH2PO4, pH 8.0, 300 mM NaCl). The cells were
lysed by sonication, and the cell extract was cleared by centrifu-
gation at 15,000 rpm for 30 min. The supernatant was loaded onto
a nickel affinity chromatography column and was eluted with
buffer B (50 mM NaH2PO4, pH 8.0, 300 mM NaCl, 25e500 mM
imidazole) at 4 ꢂC. Fractions containing purified PfPNP were pooled,
desalted on a PD10 column (GE Healthcare) and resuspended in
50 mM NaH2PO4, pH 7.5.
3.2. Antiplasmodial activity
Activity was determined in a reaction mixture of final volume
1 ml containing 50 mM NaH2PO4, pH 7.5, 2.5 mg xanthine oxidase
Unfortunately none of the compounds showed any significant
inhibition of growth of the parasites, even those showing low
micromolar Ki values (see Supplementary data). The lack of activity
(60 mU), 2 m
g (68.7 nM) PfPNP, and inosine at 25 ꢂC. The reaction
was initiated with enzyme and no evidence for two-state inhibition
was obtained for any of the compounds tested. In this coupled

5146
H. Cui et al. / European Journal of Medicinal Chemistry 45 (2010) 5140e5149
assay, inosine is converted to hypoxanthine which is oxidized to
uric acid by xanthine oxidase and the reaction is followed spec-
trophotometrically at 293 nM. Product formation was calculated
using the molar extinction coefficient for uric acid:
12.9 mMꢁ1 cmꢁ1. kcat and Km values were calculated by fitting the
5.4. Chemistry
5.4.1. General
Chemical and solvents were purchased from the SigmaeAldrich
Chemical Company, Fluka, VWR, Acros, Fisher Chemicals and Alfa
Aesar. 1H NMR, 13C NMR and COSY NMR were recorded on a Bruker
Avance DPX 500 spectrometer (1H at 500.1 MHz and 13C at
data obtained at different inosine concentrations (0.5e200
the MichaeliseMenten equation. IC50 values for PfPNP were based
on reaction rate measurements with inosine as substrate (25 M)
mM) to
m
125.8 MHz). Chemical shift (d) are expressed in ppm. Signal splitting
and variable inhibitor concentrations. For IC50 determination, at
least five concentrations were tested. Ki values were obtained
assuming competitive inhibition and using Dixon plots. Essentially
the reciprocal of the reaction velocity is plotted against the
concentration of the inhibitor, and the value of Ki is determined
from the slope of the plot and the intercept on the y-axis. The Dixon
plot uses the following equation:
patters are described as singlet (s), broad singlet (bs), double (d),
triplet (t), quarter (q), multiplet (m). Low resolution electrospray
(ES) mass spectra were recorded either on a Applied Biosystem
Mariner API-TOF biospectrometry Workstation spectrometer or on
a Bruker MicroTof mass spectrometer, run in a positive ion mode,
using either methanol, methanol/water (95:5), or water/acetonitrile
(1:1) þ 0.2% formic acid as the mobile phase. High resolution elec-
trospray measurements were performed on a Bruker Daltonics
MicrOTOF mass spectrometer. Column-chromatography was
carried out using Silica gel 60 fromFluka. Thin layerchromatography
(TLC) was carried out on Merck silica gel 60 F254 plates using UV
light or PMA for visualization. TLC data are given as the Rf value with
the corresponding eluent system specified in brackets.
ꢀ
ꢁ
1
v
Km
Ki½SꢃVmax
1
Vmax
Km
½Sꢃ
¼
½Iꢃ þ
1 þ
where v ¼ reaction velocity, Vmax ¼ limiting value of the velocity of
the reaction obtained at high substrate concentrations, Km ¼ Mi-
chaelis constant for inosine (5.9
m
M), [S] ¼ substrate concentration
(25
mM), [I] ¼ inhibitor concentration.
5.4.2. 5-Benzylidenebarbituric acid (7) [23]
Acetic acid (7.5 ml) solution of barbituric acid (128 mg, 1 mmol)
and 3-benzylaldehyde (144 mg, 1.2 mmol) was refluxed at
97e100 ꢂC for 2 h. The yellow reaction solution mixture was
concentrated to around 3 ml and the resulting yellow suspension
left at RT overnight. The solid was separated by filtration, washed
with ice-cold methanol (3 ꢀ10 ml) and dried at 110 ꢂC for 2 h. The
final product (194 mg, 90%) was yield as a white solid. Reverse
phase TLC (50% MeOH/H2O) Rf ¼ 0.55; LCMS (ESþ): m/z (%), 217
(100) [M þ H]þ.
5.2. In vitro assays
In vitro activity against the erythrocytic stages of P. falciparum
was determined by using an SYBR green assay [27]. The parasite
P. falciparum 3D7 was cultured using standard methods, and
synchronized using 5% sorbitol as previously described.
Compounds were dissolved in DMSO at 100 mM and added to 48 h
post-synchronization parasite cultures incubated in RPMI 1640
medium with hypoxanthine (150
(12.5
mM), NaHCO3 (0.2%), gentamycin
5.4.3. 5-(m-Benzyloxy)benzylidenebarbituric acid (8) [16]
Barbituric acid (128 mg, 1 mmol) was coupled with 3-benzy-
loxy-benzaldehyde (270 mg, 1.2 mmol) following the same proce-
dure of 7. The final product (273 mg, 85%) is yield as a light yellow
solid. Reverse phase TLC (50% MeOH/H2O) Rf ¼ 0.52; LCMS (ESþ):
m/z (%), 340 (100) [M þ NH3]þ, 323 (75) [M þ H]þ.
m
g/ml), Albumax (0.5%), human serum (2%) and washed
human red cells Oþ at 5% haematocrit (0.3% parasitaemia). Chlo-
roquine was used as standard drug. Experiments were carried out
at least twice independently and the different concentrations were
tested in duplicate. After 48 h of growth, 100
(Molecular Probes) in lysis buffer (Tris 20 mM, pH.7.5, EDTA 5 mM,
saponin 0.008%, tritonX-100 0.08%; 0.2 l of SYBR-green/ml of lysis
ml of SYBR green I
m
5.4.4. 5-Benzylbarbituric acid (9) [23]
buffer) was added to each well, and mixed and after 1 h of incu-
bation in the dark at RT, fluorescence was measured with excitation
and emission wavelength bands centred at 485 and 530 nm. The
percent inhibition of each compound at each concentration was
determined. EC50 values were calculated from hyperbolic or
sigmoidal doseeresponse curves using Sigmaplot 10.0.
To a suspension of compound 6 (216 mg, 1 mmol) in ethanol
(20 ml) was added NaBH4 (113 mg, 3 mmol). The reaction mixture
was stirred at RT for 2 h. The final suspension was filtered and
washed with ethanol (20 ml) and DCM to give compound 9
(196 mg, 90%) as a white solid. Reverse phase TLC (50% MeOH/H2O)
Rf ¼ 0.55; 1H NMR (500 MHz, d2-H2O)
d
7.04e7.24 (m, 5H, HePh
172.7
and CH-3), 3.45 (s, 2H, CH2); 13C NMR (125 MHz, d2-H2O)
d
(C4), 166.5 (C2), 153.1 (C6), 142.0, 129.95, 128.97, 128.5, 128.4, 128.2,
127.7, 125.6 (CePh), 89.1 (C3), 27.5 (CH2); LCMS (ESþ): m/z (%) 219
(100) [M þ H]þ, 437 (56) [2M þ H]þ; HRMS (ESþ): calcd for
C11H10N2O3 [M þ H]þ 219.0764 m/z, found 219.0763 m/z.
5.3. Modelling
3D structures for compounds 18 and 24 were drawn using
Chem3D (ChembridgeSoft, Cambridge, UK). These structures were
then manually docked into PfPNP active site (pdb 1NW4 between
chains A & B) by superposition with the co-crystallised inhibitor
(Immucillin-H) using PyMol (Schrodinger, LLC). Subsequently the
inhibitor Immucillin-H was removed and polar hydrogen atoms
were added to polar atoms and their positions were minimized
using the MAB force field [24,25] as implemented in Moloc (Gerber
Molecular Design, Switzerland) with the compounds present in
order to obtain a hydrogen-bonding network optimized for ligand
binding. Water molecules were kept in the active site during the
optimization and protein residues were kept stationary whilst
ligand, waters and hydrogen atoms were considered flexible. The
optimized binding poses were then inspected using PyMol.
5.4.5. 5-(m-Benzyloxy)benzylbarbituric acid (1) [16]
Compound 8 (322 mg, 1 mmol) was reacted with NaBH4 in
ethanol to give compound (1) (290 mg, 90%) as a white solid.
Reverse phase TLC (50% MeOH/H2O) Rf ¼ 0.60; 1H NMR (500 MHz,
d2-H2O)
d
7.16e7.32 (m, 5H, HePh), 7.06 (t, J ¼ 7.9 Hz, 1H, HePh),
6.76 (d, J ¼ 7.6 Hz, 1H, HePh), 6.66 (dd, J ¼ 3.4 Hz, 1H, HePh), 6.56
(s, 1H, HePh), 4.93 (s, 2H, OCH2), 3.35 (s, 2H, CH2); 13C NMR
(125 MHz, d2-H2O)
d 166.3 (C4), 157.6 (C2), 153.0 (C6), 143.91
(CePh), 136.8, 129.4, 128.6, 128.0, 127.5, 121.2, 113.8, 113.0 (CePh),
88.8 (C3), 70.1 (OCH2), 27.34 (CH2); LCMS (ESþ): m/z (%) 325 (100)
[M þ H]þ, 649 (25) [2M þ H]þ; HRMS (ESþ): calcd for C18H17N2O4
[M þ H]þ 325.1183 m/z, found 325.1185 m/z.

H. Cui et al. / European Journal of Medicinal Chemistry 45 (2010) 5140e5149
5147
5.4.6. (2-Acetoxyethoxy)methyl bromide (12) [23]
DMSO)
d
11.35 (s, 1H, NH), 7.66 (d, J ¼ 7.8 Hz, 1H, H6), 5.58 (dd,
Acetyl bromide (8.32 g, 0.0676 mol) was added dropwise to
1,3-dioxolane (5.01 g, 0.0676 mol) with stirring at RT. After addition
of 1/3 of the acetyl bromide, the temperature was raised to reflux
(70 ꢂC) and the remaining acetyl bromide was added over a 15 min
period. After refluxing for 4 h, the solution was subjected to frac-
tional distillation. At 113 ꢂC, the product (3.8 g, 29%) was distilled as
J ¼ 3.3 Hz, 1H, H5), 4.06 (t, J ¼ 6.3 Hz, 2H, NCH2), 3.71 (t, J ¼ 6.3 Hz,
2H, CH2); 13C NMR (125 MHz, d6-DMSO)
d 163.6 (C4), 150.8 (C2),
145.8 (C6), 100.7 (C5), 48.8 (NCH2), 30.7 (CH2); LCMS (ESþ): m/z (%)
139 (100) [M ꢁ HBr]ꢁ, 218 (33) [M þ H]þ.
5.4.11. 1-(3-Bromopropyl)uracil (16)
a liquid. 1H NMR (500 MHz, d1-CHCl3)
CH2), 3.82 (m, 2H, CH2), 2.03 (s, 2H, CH3); 13C NMR (125 MHz,
d1-CHCl3)
170.6 (CO), 75.7, 75.6 (CH2), 69.3, 69.24, 69.17, 69.1
d
5.66 (s, 2H, CH2), 4.23 (t, 2H,
Uracil (1.12 g, 10 mmol) and 1,3-dibromopropane (1 ml,
10 mmol) reacted and yield compound 16 550 mg (24%) as a white
solid. TLC (10% MeOH/DCM) Rf ¼ 0.33; 1H NMR (500 MHz,
d
(CH2), 62.3, 62.12, 62.09 (CH2), 20.7 (CH3).
d6-DMSO)
d
11.26 (s, 1H, NH), 7.62 (d, J ¼ 7.8 Hz, 1H, H6), 5.56 (dd,
J ¼ 3.3 Hz, 1H, H5), 3.77 (t, J ¼ 6.9 Hz, 2H, NCH2), 3.53 (t, J ¼ 6.6 Hz,
5.4.7. 1-[(2-Acetoxyethoxy)methyl]-5-(m-benzyloxy)
benzylbarbituric acid (13) [23]
2H, BrCH2), 2.13 (qt, J ¼ 6.7 Hz, 2H, CH2); 13C NMR (125 MHz,
d6-DMSO)
d 163.7 (C4), 150.9 (C2), 145.6 (C6), 101.0 (C5), 46.5
Compound 1 (2 g, 6.17 mmol) and chlorotrimethylsilane (1.6 ml,
18 mmol)were heated inhexamethyldisilazane(HMDS)(100 ml)and
refluxed overnight (75e80 ꢂC) in Ar. The excess HMDS was removed
by evaporation (100 ꢂC) and the resulting crystalline persilylated
compound 1 was dissolved in dry 1,2-dichloroethane (20 ml).
Compound 12 (1.2 g, 6.17 mmol) and a catalytic amount of anhydrous
aluminium chloride (40 mg) were added to the reaction, and the
reaction mixture was allowed to stir for 48 h at RT under anhydrous
condition. The reaction mixture was then cooled and carefully
neutralized (pH ¼ 7) with a saturated Na2CO3. The organic layer was
separated and the aqueous phase was extracted with chloroform
(3 ꢀ 100 ml). The organic layerand extractswere combined and dried
over anhydrous MgSO4. The dried organic layer was evaporated to
dryness and the resulting residue was dissolved in a minimal amount
ofDCMandappliedtoasilicagelcolumn, Thecolumnwaselutedwith
2% MeOH/DCM to give compound 13 (352 mg, 13%) as a white solid.
TLC (10% MeOH/DCM) Rf ¼ 0.43; LCMS (ESþ): m/z (%) 458 (100)
[M þ NH3]þ, 441 (70) [M þ H]þ; HRMS (ESþ): calcd for C23H25N2O7
[M þ H]þ 441.1656 m/z, found 441.1654 m/z.
(NCH2), 31.32 (BrCH2), 31.29 (CH2); LCMS (ESþ): m/z (%) 153 (100)
[M ꢁ HBr]ꢁ, 232 (24) [M þ H]þ.
5.4.12. 1-(4-Bromobutyl)uracil (17)
Uracil (1.12 g, 10 mmol) and 1,4-dibromobutane (1.18 ml,
10 mmol) reacted and yield compound 17 (440 mg, 18%) as a white
solid. TLC (10% MeOH/DCM) Rf ¼ 0.33; 1H NMR (500 MHz, d6-DMSO)
d
11.26 (s, 1H, NH), 7.76 (d, J ¼ 7.8 Hz, 1H, H6), 5.56 (d, J ¼ 7.8 Hz, 1H,
H5), 3.69 (t, J ¼ 6.8 Hz, 2H, NCH2), 3.56 (t, J ¼ 6.5 Hz, 2H, BrCH2), 1.78
(qt, J ¼ 6.9 Hz, 2H, CH2), 1.69 (qt, J ¼ 5.8 Hz, 2H, CH2); 13C NMR
(125 MHz, d6-DMSO)
d 163.7 (C4), 150.9 (C2), 145.6 (C6), 100.9 (C5),
46.5 (NCH2), 34.5 (BrCH2), 29.1 (CH2), 27.2 (CH2); LCMS (ESþ): m/z (%)
247 (99) [M þ H]þ, 249 (100) [M þ 3H]þ.
5.4.13. 3-(2-Bromoethyl)uracil (23)
Uracil (1.12 g, 10 mmol) and 1,2-dibromoethane (0.85 ml,
10 mmol) reacted and yield compound 23 (325 mg, 15%) as a white
solid. TLC (10% MeOH/DCM) Rf ¼ 0.25; 1H NMR (500 MHz,
d6-DMSO)
d
7.70 (d, J ¼ 6.6 Hz, 1H, H6), 5.98 (d, J ¼ 6.6 Hz, 1H, H5),
4.69 (t, J ¼ 8.5 Hz, 2H, BrCH2), 4.16 (t, J ¼ 8.5 Hz, 2H, NCH2); 13C NMR
5.4.8. 1-[(2-Hydroxyethoxy)methyl]-5-(m-benzyloxy)
benzylbarbituric acid (2) [23]
(125 MHz, d6-DMSO)
d 160.5,160.3 (C4),154.8 (C2),107.3 (C6),100.7,
100.0 (C5), 66.6 (BrC), 42.3 (NC); LCMS (ESþ): m/z (%) 139 (100)
[M ꢁ HBr]ꢁ, 218 (5) [M þ H]þ.
The ester 13 (250 mg, 0.57 mmol) was dissolved in a solution of
0.2 M sodium methoxide in MeOH. The reaction was stirred at RT
until the disappearance of the starting ester which was observed by
TLC. The solution was neutralized with Dowex ion exchange resin
to pH. 6.0. The residue was filtered and washed twice with MeOH.
The MeOH was evaporated to afford a crude residue, which was
eluted through a reverse phase column to give the final compound
2 (156 mg, 71%) as a white solid. Reverse TLC (50% MeOH/H2O)
5.4.14. General procedures for 18e22 and 24, 25
Compounds 15e17 and 23 (1 mmol, 1 eq) were stirred with
pyrrolidine or l-prolinol (1 mmol, 1 eq) at 150 ꢂC in ethanol under
microwaves irradiation for 1 min. Then ethanol was evaporated and
the residue was purified by chromatography in 20% methanol/DCM
to give the final product.
Rf ¼ 0.38; 1H NMR (500 MHz, d4-MeOH)
d 7.27e7.43 (m, 5H, HePh),
7.08 (t, J ¼ 7.8 Hz, 1H, HePh), 6.95 (d, J ¼ 1.5 Hz, 1H, HePh), 6.89 (d,
J ¼ 7.6 Hz, 1H, HePh), 6.70 (dd, J ¼ 3.3 Hz, 1H, HePh), 5.36 (s, 2H,
NCH2), 5.04 (s, 2H, OCH2), 3.60e3.69 (m, 6H, OCH2, OCH2 and CH2);
5.4.15. 1-((2-Pyrrolidine-1-yl)-ethyl)uracil (18)
Compound 15 (217 mg, 1 mmol) reacted with pyrrolidine (83 ml,
1 mmol) to give product 18 (68 mg, 33%) as a light yellow solid. TLC
13C NMR (125 MHz, d4-MeOH)
d
166.6, 166.3 (C4), 160.1 (C2), 154.6
(20% MeOH/DCM) Rf ¼ 0.02; 1H NMR (500 MHz, d4-MeOH)
d 7.64
(C6), 146.2 (CePh), 139.2 (CePh), 129.6, 129.4, 128.7, 128.6, 122.1,
115.6, 112.8 (CePh), 89.4 (C3), 71.6, 71.2, 70.8 (NCH2, OCH2 and
OCH2), 62.2 (OCH2), 30.0 (CH2); LCMS (ESþ): m/z (%) 416 (100)
[M þ NH3]þ, 399 (53) [M þ H]þ; HRMS (ESþ): calcd for C21H23N2O6
[M þ H]þ 399.1551 m/z, found 399.1539 m/z.
(d, J ¼ 7.8 Hz, 1H, H6), 5.68 (d, J ¼ 7.8 Hz, 1H, H5), 3.98 (t, J ¼ 6.4 Hz,
2H, NCH2), 2.99 (t, J ¼ 5.6 Hz, 2H, NCH2), 2.84 (m, 4H, CH2NCH2),1.90
(m, 4H, CH2CH2); 13C NMR (125 MHz, d4-MeOH)
d 166.8 (C4), 153.0
(C2), 147.3 (C6), 102.4 (C5), 55.3 (NCH2), 55.2 (NCH2), 46.7
(CH2NCH2), 24.29 (CH2CH2); LCMS (ESþ): m/z (%) 210 (100)
[M þ H]þ; HRMS (ESþ): calcd for C10H16N3O2 [M þ H]þ 210.1237 m/
z, found 210.1248 m/z.
5.4.9. General procedures for 15e17 and 23
Uracil (10 mmol, 1 eq) and Br(CH2)nBr (n ¼ 2e4) (10 mmol, 1 eq)
were stirred in DMF (50 ml) with Cs2CO3 (10 mmol, 1 eq) at 40 ꢂC
for 48 h. The residue was evaporated to remove DMF and purified
by chromatography with 5% MeOH/DCM.
5.4.16. 1-((3-Pyrrolidine-1-yl)-propyl)uracil (19)
Compound 16 (230 mg, 1 mmol) reacted with pyrrolidine (83 ml,
1 mmol) to give product 19 (82 mg, 37%) as a light yellow solid. TLC
(20% MeOH/DCM) Rf ¼ 0.03; 1H NMR (500 MHz, d4-MeOH)
d 7.62
5.4.10. 1-(2-Bromoethyl)uracil (15) [28]
(d, J ¼ 7.8 Hz, 1H, H6), 5.66 (d, J ¼ 7.8 Hz, 1H, H5), 3.82 (t, J ¼ 7.1 Hz,
Uracil (1.12 g, 10 mmol) and 1,2-dibromoethane (0.85 ml,
10 mmol) reacted and yield compound 15 (280 mg, 13%) as a white
solid. TLC (10% MeOH/DCM) Rf ¼ 0.31; 1H NMR (500 MHz, d6-
2H, NCH2), 2.54e2.60 (m, 6H, NCH2 and CH2NCH2), 1.93 (qt,
J ¼ 7.3 Hz, 2H, CH2 and CH2CH2); 13C NMR
d
(125 MHz, d4-MeOH)
166.9 (C4), 152.9 (C2), 147.4 (C6), 102.2 (C5), 54.9 (NCH2), 54.04

5148
H. Cui et al. / European Journal of Medicinal Chemistry 45 (2010) 5140e5149
(NCH2), 48.05 (NCH2), 28.82 (CH2), 24.25 (CH2CH2); LCMS (ESþ):
m/z (%) 224 (100) [M þ H]þ; HRMS (ESþ): calcd for C11H18N3O2
[M þ H]þ 224.1394 m/z, found 224.1386 m/z.
d4-MeOH)
d 166.1 (C4), 153.5 (C2), 142.1 (C6), 101.4 (C5), 68.7 (CH),
63.0 (OCH2), 55.8 (NCH2), 53.3 (NCH2), 39.1 (NCH2), 28.2 (CH2), 23.8
(CH2); LCMS (ESþ): m/z (%) 240 (100) [M þ H]þ; HRMS (ESþ): calcd
for C11H18N3O3 [M þ H]þ 240.1343 m/z, found 240.1346 m/z.
5.4.17. 1-((3-(l-Prolinol)-1-yl)-propyl)uracil (20)
Compound 16 (230 mg, 1 mmol) reacted with l-prolinol (100 ml,
Acknowledgements
1 mmol) to give product 20 (88 mg, 35%) as a light yellow solid. TLC
(20% MeOH/DCM) Rf ¼ 0.03; 1H NMR (500 MHz, d4-MeOH)
d 7.68
Financial support for this research came from the Plan Nacional
(SAF2007-62596), the RICET FIS Network (RD06/0021), the Junta de
Andalucía (CVI-199) and the European Union (ANTIMAL contract
no. LSHP-CT-2005-018834). The authors also want to thank Gina
MacKay for HRMS.
(d, J ¼ 7.8 Hz, 1H, H6), 5.68 (d, J ¼ 7.8 Hz, 1H, H5), 3.60e3.69 (m, 2H,
NCH2), 3.81e3.94 (m, 2H, CH2), 3.33 (qt, J ¼ 1.6 Hz, 1H, NCHH),
3.17e3.22 (m, 1H, NCHH), 3.01e3.06 (m, 1H, CH), 2.64e2.75 (m, 2H,
NCHH and NCHH), 2.01e2.09 (m, 3H, NCCH2 and CHH), 1.74e1.95
(m, 3H, CH2 and CHH); 13C NMR (125 MHz, d4-MeOH)
d 166.9 (C4),
152.9 (C2), 147.6 (C6), 102.1 (C5), 67.5 (CH), 64.8 (OCH2), 55.3
(NCH2), 53.4 (NCH2), 48.0 (NCH2), 28.7 (CH2), 28.6 (CH2), 23.7 (CH2);
LCMS (ESþ): m/z (%) 254 (100) [M þ H]þ; HRMS (ESþ): calcd for
C12H20N3O3 [M þ H]þ 254.1499 m/z, found 254.1500 m/z.
Appendix. Supplementary data
Supplementary data associated with this article can be found in
the online version, at doi:10.1016/j.ejmech.2010.08.026.
5.4.18. 1-((4-Pyrrolidine-1-yl)-butyl)uracil (21)
References
Compound 17 (246 mg, 1 mmol) reacted with pyrrolidine (83 ml,
[1] B. Greenwood, T. Mutabingwa, Malaria in 2002, Nature 415 (2002) 670e672.
[2] Africa Malaria Report, 2003. Available from: http://www.rbm.who.int/
amd2003/amr2003/amr2003/amr_toc.htm.
[3] I. Bathurst, C. Hentschel, Medicines for malaria venture: sustaining antima-
larial drug development, Trends Parasitol. 22 (2006) 301e307.
[4] C. Schnick, M.A. Robien, A.M. Brzozowski, E.J. Dodson, G.N. Murshudov,
L. Anderson, J.R. Luft, C. Mehlin, W.G.J. Hol, J.A. Brannigan, et al., Structures of
Plasmodium falciparum purine nucleoside phosphorylase complexed with
sulfate and its natural substrate inosine, Acta Crystallogr., Sect. D 61 (2005)
1245e1254.
[5] G.A. Kicska, P.C. Tyler, G.B. Evans, R.H. Furneaux, K. Kim, V.L. Schramm,
Transition state analogue inhibitors of purine nucleoside phosphorylase from
Plasmodium falciparum, J. Biol. Chem. 277 (2002) 3219e3225.
[6] W.X. Shi, L.M. Ting, G.A. Kicska, A. Lewandowicz, P.C. Tyler, G.B. Evans,
R.H. Furneaux, K. Kim, S.C. Almo, V.L. Schramm, Plasmodium falciparum purine
nucleoside phosphorylase: crystal structures, immucillin inhibitors, and dual
catalytic function, J. Biol. Chem. 279 (2004) 18103e18106.
[7] A. Lewandowicz, V.L. Schramm, Transition state analysis for human and
Plasmodium falciparum purine nucleoside phosphorylases, Biochemistry 43
(2004) 1458e1468.
[8] A. Chaikuad, R.L. Brady, Conservation of structure and activity in Plasmodium
purine nucleoside phosphorylases, BMC Struct. Biol. 9 (2009) 42.
[9] D.C. Madrid, L.M. Ting, K.L. Waller, V.L. Schramm, K. Kim, Plasmodium falci-
parum purine nucleoside phosphorylase is critical for viability of malaria
parasites, J. Biol. Chem. 283 (2008) 35899e35907.
[10] G.A. Kicska, P.C. Tyler, G.B. Evans, R.H. Furneaux, V.L. Schramm, K. Kim, Purine-
less death in Plasmodium falciparum induced by immucillin-H, a transition
state analogue of purine nucleoside phosphorylase, J. Biol. Chem. 277 (2002)
3226e3231.
[11] A. Lewandowicz, E.A.T. Ringia, L.M. Ting, K. Kim, P.C. Tyler, G.B. Evans,
O.V. Zubkova, S. Mee, G.F. Painter, D.H. Lenz, et al., Energetic mapping of
transition state analogue interactions with human and Plasmodium falciparum
purine nucleoside phosphorylases, J. Biol. Chem. 280 (2005) 30320e30328.
[12] P.C. Kline, V.L. Schramm, Pre-steady-state transition-state analysis of the
hydrolytic reaction catalyzed by purine nucleoside phosphorylase, Biochem-
istry 34 (1995) 1153e1162.
[13] E.A.T. Ringia, P.C. Tyler, G.B. Evans, R.H. Furneaux, A.S. Murkin, V.L. Schramm,
Transition state analogue discrimination by related purine nucleoside phos-
phorylases, J. Am. Chem. Soc. 128 (2006) 7126e7127.
[14] L.M. Ting, W.X. Shi, A. Lewandowicz, V. Singh, A. Mwakingwe, M.R. Birck,
E.A.T. Ringia, G. Bench, D.C. Madrid, P.C. Tyler, et al., Targeting a novel Plas-
modium falciparum purine recycling pathway with specific immucillins, J. Biol.
Chem. 280 (2005) 9547e9554.
[15] S. Cha, Development of inhibitors of pyrimidine metabolism, Yonsei Med. J. 30
(1989) 315e326.
[16] D.J. Guerin, D. Mazeas, M.S. Musale, F.N.M. Naguib, O.N. Al Safarjalani,
M.H. el Kouni, R.P. Panzica, Uridine phosphorylase inhibitors: chemical
modification of benzyloxybenzyl-barbituric acid and its effects on URDPASE
inhibition, Bioorg. Med. Chem. Lett. 9 (1999) 1477e1480.
[17] W.M. Bu, E.C. Settembre, M.H. el Kouni, S.E. Ealick, Structural basis for inhi-
bition of Escherichia coli uridine phosphorylase by 5-substituted acyclour-
idines, Acta Crystallogr., Sect. D 61 (2005) 863e872.
[18] J.M. Mason, A.S. Murkin, L. Li, V.L. Schramm, G.J. Gainsford, B.W. Skelton, A beta-
fluoroamine inhibitor of purine nucleoside phosphorylase, J. Med. Chem. 51
(2008) 5880e5884.
[19] K. El Bissati, R. Zufferey, W.H. Witola, N.S. Carter, B. Ullman, C. Ben Mamoun,
The plasma membrane permease PfNT1 is essential for purine salvage in the
human malaria parasite Plasmodium falciparum, Proc. Natl. Acad. Sci. U.S.A.
103 (2006) 9286e9291.
1 mmol) to give product 21 (97 mg, 41%) as a light yellow solid. TLC
(20% MeOH/DCM) Rf ¼ 0.05; 1H NMR (500 MHz, d4-MeOH)
d 7.63
(d, J ¼ 7.8 Hz, 1H, H6), 5.67 (d, J ¼ 7.8 Hz, 1H, H5), 3.79 (t, J ¼ 7.2 Hz,
2H, NCH2), 2.68e2.70 (m, 4H, CH2NCH2), 2.63 (t, 2H, NCH2),
1.82e1.89 (m, 4H, CH2CH2), 1.73 (qt, J ¼ 7.4 Hz, 2H, CH2), 1.58e1.64
(m, 2H, CH2); 13C NMR (125 MHz, d4-MeOH)
d 167.0 (C4), 153.0 (C2),
147.3 (C6), 102.3 (C5), 56.8 (CH2NCH2), 55.0 (NCH2), 28.0 (CH2),
26.13 (CH2), 24.17 (CH2); LCMS (ESþ): m/z (%) 238 (100) [M þ H]þ;
HRMS (ESþ): calcd for C12H20N3O2 [M þ H]þ 238.1550 m/z, found
238.1561 m/z.
5.4.19. 1-((4-(l-Prolinol)-1-yl)-butyl)uracil (22)
Compound 17 (246 mg, 1 mmol) reacted with l-prolinol (100 ml,
1 mmol) to give product 22 (106 mg, 40%) as a light yellow solid. TLC
(20% MeOH/DCM) Rf ¼ 0.06; 1H NMR (500 MHz, d4-MeOH)
d 7.61
(d, J ¼ 7.8 Hz, 1H, H6), 5.67 (d, J ¼ 7.8 Hz, 1H, H5), 3.79 (t, J ¼ 7.2 Hz,
2H, NCH2), 3.60 (dd, J ¼ 5.2 Hz, 1H, OCHH), 3.50 (dd, J ¼ 5.6 Hz, 1H,
OCHH), 3.16e3.20 (m, 1H, NCHH), 2.93e2.98 (m, 1H, NCHH),
2.60e2.65 (m, 1H, CH), 2.36e2.42 (m, 1H, NCHH), 2.31 (q, J ¼ 8.7 Hz,
1H, NCHH), 1.91e2.00 (m,1H, CHH),1.65e1.84 (m, 5H, CHH, CH2 and
CH2), 1.53e1.62 (m, 2H, CH2); 13C NMR (125 MHz, d4-MeOH)
d 166.9
(C4), 152.9 (C2), 147.3 (C6), 102.3 (C5), 67.94 (CH), 64.91 (OCH2), 56.2
(NCH2), 55.5 (NCH2), 49.6 (NCH2), 28.8 (CH2), 26.2 (CH2), 26.2 (CH2),
23.7 (CH2); LCMS (ESþ): m/z (%) 268 (100) [M þ H]þ.
5.4.20. 3-((2-Pyrrolidine-1-yl)-ethyl)uracil (24)
Compound 23 (217 mg, 1 mmol) reacted with pyrrolidine (83 ml,
1 mmol) to give product 24 (81 mg, 39%) as a light yellow solid. TLC
(20% MeOH/DCM) Rf ¼ 0.02; 1H NMR (500 MHz, d4-MeOH)
d 7.41
(d, J ¼ 7.6 Hz, 1H, H6), 5.72 (d, J ¼ 7.6 Hz, 1H, H5), 4.16 (t, J ¼ 6.6 Hz,
2H, NCH2), 3.01 (t, J ¼ 6.6 Hz, 2H, NCH2), 2.95 (m, 4H, CH2NCH2),1.92
(m, 4H, CH2CH2); 13C NMR (125 MHz, d4-MeOH)
d 166.2 (C4), 153.6
(C2), 142.0 (C6), 101.4 (C5), 55.4 (NCH2), 54.3 (CH2NCH2), 39.2 (CH2),
24.15 (CH2); LCMS (ESþ): m/z (%) 210 (100) [M þ H]þ; HRMS (ESþ):
calcd for C10H16N3O2 [M þ H]þ 210.1237 m/z, found 210.1245 m/z.
5.4.21. 3-((2-(l-Prolinol)-1-yl)-ethyl)uracil (25)
Compound 23 (217 mg, 1 mmol) reacted with l-prolinol (100 ml,
1 mmol) to give product 25 (78 mg, 33%) as a light yellow solid. TLC
(20% MeOH/DCM) Rf ¼ 0.02; 1H NMR (500 MHz, d4-MeOH)
d 7.44 (d,
J ¼ 7.6 Hz, 1H, H6), 5.72 (d, J ¼ 7.6 Hz, 1H, H5), 4.10e4.15, 4.19e4.24
(m, 2H, NCH2), 3.62 (m, 2H, OCH2), 3.52 (m, 1H, NCHH), 3.38 (m, 1H,
NCHH), 3.07 (m, 1H, CH), 2.86e2.91 (m, 1H, NCHH), 2.78e2.83 (m,
1H, NCHH), 2.04e2.13 (m, 1H, CHH), 1.92e1.99 (m, 1H, CHH),
1.83e1.89 (m,1H, CHH),1.73e1.80 (m,1H, CHH); 13C NMR (125 MHz,

H. Cui et al. / European Journal of Medicinal Chemistry 45 (2010) 5140e5149
5149
[20] B.S. Jursic, E.D. Stevens, Mono C-alkylation and mono C-benzylation of
barbituric acids through zinc/acid reduction of acyl, benzylidene, and
alkylidene barbiturate intermediates, Tetrahedron Lett. 44 (2003)
2203e2210.
[21] A. Rosowsky, S.H. Kim, M. Wick, Synthesis and antitumor activity of an acy-
clonucleoside derivative of 5-fluorouracil, J. Med. Chem. 24 (1981)
1177e1181.
[22] G.F. Orr, D.L. Musso, G.E. Boswell, J.L. Kelley, S.S. Joyner, S.T. Davis,
D.P. Baccanari, Inhibition of uridine phosphorylase: synthesis and structur-
eeactivity relationships of aryl-substituted 5-benzyluracils and 1-[(2-
hydroxyethoxy)methyl]-5-benzyluracils, J. Med. Chem. 38 (1995) 3850e3856.
[23] D.L. Levesque, E.C. Wang, D.C. Wei, C.C. Tzeng, R.P. Panzica, F.N.M. Naguib,
M.H. Elkouni, Synthesis of a new class of uridine phosphorylase inhibitors,
J. Heterocycl. Chem. 30 (1993) 1399e1404.
[24] P.R. Gerber, K. Muller, MAB, a generally applicable molecular force field for
structure modelling in medicinal chemistry, J. Comput. Aided Mol. Des. 9
(1995) 251e268.
[25] P.R. Gerber, Charge distribution from a simple molecular orbital type calcu-
lation and non-bonding interaction terms in the force field MAB, J. Comput.
Aided Mol. Des. 12 (1998) 37e51.
[26] E.A. Taylor, K. Clinch, P.M. Kelly, L. Li, G.B. Evans, P.C. Tyler, V.L. Schramm, Acyclic
ribooxacarbenium ion mimicsastransition state analogues of human and malarial
purine nucleoside phosphorylases, J. Am. Chem. Soc. 129 (2007) 6984e6985.
[27] M. Smilkstein, N. Sriwilaijaroen, J.X. Kelly, P. Wilairat, M. Riscoe, Simple and
inexpensive fluorescence-based technique for high-throughput antimalarial
drug screening, Antimicrob. Agents Chemother. 48 (2004) 1803e1806.
[28] B. Lohse, S. Hvilsted, R.H. Berg, P.S. Ramanujam, N-1-alkylated pyrimidine films as
a new potential optical data storagemedium, Chem. Mater. 18 (2006) 4808e4816.