Design, Synthesis and Biological Evaluation of Novel Inhibitors of Trypanosoma brucei Pteridine Reductase1

Article abstract of DOI:10.1002/cmdc.201000450

Genetic studies indicate that the enzyme pteridine reductase1 (PTR1) is essential for the survival of the protozoan parasite Trypanosoma brucei. Herein, we describe the development and optimisation of a novel series of PTR1 inhibitors, based on benzo[d]im

Full text of DOI:10.1002/cmdc.201000450

MED
DOI: 10.1002/cmdc.201000450
Design, Synthesis and Biological Evaluation of Novel
Inhibitors of Trypanosoma brucei Pteridine Reductase 1
Daniel Spinks, Han B. Ong, Chidochangu P. Mpamhanga, Emma J. Shanks,
David A. Robinson, Iain T. Collie, Kevin D. Read, Julie A. Frearson, Paul G. Wyatt, Ruth Brenk,
Alan H. Fairlamb, and Ian H. Gilbert*[a]
Genetic studies indicate that the enzyme pteridine reductase 1
(PTR1) is essential for the survival of the protozoan parasite Try-
panosoma brucei. Herein, we describe the development and
optimisation of a novel series of PTR1 inhibitors, based on ben-
zo[d]imidazol-2-amine derivatives. Data are reported on 33
compounds. This series was initially discovered by a virtual
screening campaign (J. Med. Chem., 2009, 52, 4454). The inhibi-
tors adopted an alternative binding mode to those of the nat-
ural ligands, biopterin and dihydrobiopterin, and classical in-
hibitors, such as methotrexate. Using both rational medicinal
chemistry and structure-based approaches, we were able to
derive compounds with potent activity against T. brucei PTR1
(Kai pp =7 nm), which had high selectivity over both human and
T. brucei dihydrofolate reductase. Unfortunately, these com-
pounds displayed weak activity against the parasites. Kinetic
studies and analysis indicate that the main reason for the lack
of cell potency is due to the compounds having insufficient
potency against the enzyme, which can be seen from the low
K
m to Ki ratio (Km =25 nm and Ki =2.3 nm, respectively).
Introduction
Human African trypanosomiasis (HAT) is a serious health prob-
lem in sub-Saharan Africa, with an estimated 50000 new infec-
tions each year, and over 60 million people in 36 countries are
at risk of infection.[1] HAT is a progressive and ultimately fatal
disease. The causative agents of HAT are the protozoan para-
sites Trypanosoma brucei gambiense and T. b. rhodesiense,
which are transmitted by the bite of a tsetse fly. In the initial
stage of the infection, parasites multiply in the blood and lym-
phatic systems of the host, causing fever, headaches and joint
pains. Eventually they cross the blood–brain barrier (BBB) to
invade the central nervous system (CNS). Once the infection
penetrates the CNS, it is very difficult to treat, causing the clas-
sical symptoms of mental deterioration leading to coma and
death, which give the disease its more commonly recognisable
name of ‘sleeping sickness’. The current drugs to treat HAT are
inadequate, due to poor efficacy, side effects and the require-
ment for parenteral administration, which is not appropriate
for a rural African setting.[2]
Figure 1. Known DHFR inhibitors: methotrexate (a); trimethoprim (b); pyri-
methamine (c).
in Leishmania to DHFR inhibitors.[6–8] Data from our laboratory,
however, indicates that PTR1 may be a drug target in its own
right in T. brucei, as PTR1 knockdown by RNA interference in
the bloodstream form of T. brucei results in loss of viability in
culture and loss of virulence in animal models of infection.[9]
Folate metabolism has been successfully used as a drug
target in a number of diseases such as cancer, bacterial infec-
tions and malaria. In particular, the enzyme dihydrofolate re-
ductase (DHFR) is a clinically validated drug target in some dis-
eases.[3,4] Figure 1 shows examples of known DHFR inhibitors.
Although folate metabolism is a potential drug target in Trypa-
nosoma and Leishmania,[5] known DHFR inhibitors are not
potent inhibitors of parasite growth. One possible explanation
involves pteridine reductase 1 (PTR1), an NADPH-dependent
enzyme that not only carries out the reduction of biopterin to
dihydrobiopterin and dihydrobiopterin to tetrahydrobiopterin,
but also the reduction of dihydrofolate to tetrahydrofolate.[6,7]
Thus, PTR1 serves as a possible by-pass/resistance mechanism
[a] D. Spinks, Dr. H. B. Ong, Dr. C. P. Mpamhanga, Dr. E. J. Shanks,
Dr. D. A. Robinson, I. T. Collie, Dr. K. D. Read, Prof. J. A. Frearson,
Prof. P. G. Wyatt, Dr. R. Brenk, Prof. A. H. Fairlamb, Prof. I. H. Gilbert
Drug Discovery Unit, Division of Biological Chemistry & Drug Discovery
College of Life Sciences, University of Dundee
Sir James Black Centre, Dundee, Scotland, DD1 5EH (UK)
Fax: (+44)1382-386-373
E-mail: I.H.Gilbert@dundee.ac.uk
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cmdc.201000450.
Re-use of this article is permitted in accordance with the Terms and Condi-
tions set out at http://onlinelibrary.wiley.com/journal/10.1002/(ISSN) 1860–
7179/homepage/2452_onlineopen.html
302
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2011, 6, 302 – 308

Inhibitors of Trypanosoma brucei Pteridine Reductase 1
These observations are strongly indicative that TbPTR1 is a
promising drug target alone or in combination with DHFR in-
hibitors for developing a novel treatment for HAT. However, for
PTR1 to be considered a truly viable drug target, further chem-
ical evidence of essentiality and druggability is required.
associated with solubility problems. In addition, most of the
known PTR1 inhibitors are not particularly selective and could
also inhibit human DHFR, causing toxicity issues.[10,12] There-
fore, we recently reported the virtual screening of a fragment
library to identify new scaffolds with a lower PSA.[13] The key
compound series identified from this process was based on
the aminobenzimidazole scaffold (Figure 2). X-ray crystallogra-
phy demonstrated that this scaffold bound in three different
binding modes, depending on the substituents; two of these
were in the pterin binding site (Figure 2a and b, compounds 1
and 2, respectively), whilst in the other binding mode, the
Known inhibitors of PTR1 are based on diaminopteridine
(methotrexate) and diaminopyrimidine (trimethoprim and pyri-
methamine) templates (Figure 1).[10,11] Whilst these templates
have a very strong binding interaction with the enzyme, they
also have a high polar surface area (PSA), which may cause
problems with blood–brain barrier (BBB) permeability and is
Figure 2. Depending on the substitution pattern, the aminobenzimidazole scaffold adopts three different binding modes. Crystal structures of the inhibitors
are shown as follows:[13] a) With only a small substitution in the 6-position, the inhibitor forms direct hydrogen bonds with the phosphate group of the co-
factor NADP+. b) The unsubstituted scaffold forms water-mediated hydrogen bonds (red spheres) with the phosphate group of NADP+. c) When the scaffold
is substituted at N1, the compound binds in an area of the binding site that is perpendicular to the canonical binding site.[13] d) Hydrophobic pockets accessi-
ble by substitution of compound 11. The surfaces of the pockets for substitutions at the 7- and 4-positions of 11 are coloured yellow and green, respectively.
e) X-ray crystal structure of compound 32 (shown in magenta) showing use of the 7’ hydrophobic pocket and the edge–face interaction of the 7-phenyl
group with Trp221.
ChemMedChem 2011, 6, 302 – 308
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
303

I. H. Gilbert et al.
MED
ligand occupied an area of the active site perpendicular to the
canonical binding site (Figure 2c, compound 11). This part of
the PTR1 binding site possesses pronounced differences com-
pared to the DHFR active site and thus selectivity over DHFR
was readily achieved. Two hydrophobic pockets are located
close to the benzimidazole moiety in the canonical binding
site, and can be explored for compound optimisation (Fig-
ure 2d). Substituting the 7-position of 11 led to 32 (Figure 2e)
Table 1. Activity of N1- and N2-substituted compounds against TbPTR1.
Compd
R1 group
R2 group TbPTR1 Kiapp [mm][a]
2[b]
3[b]
4
5[b]
6
H
Et
H
H
H
H
288
>200
>60
24
C(=O)Ph
CH2CH2Ph
CH2Ph
H
16
7
8
9
CH2-(2-chlorobenzene)
CH2-(3-chlorobenzene)
CH2-(4-chlorobenzene)
CH2-(2,5-dichlorobenzene)
CH2-(3,4-dichlorobenzene)
CH2-(3,5-dichlorobenzene)
CH2-(2-methylbenzene)
CH2-(2-fluorobenzene)
CH2-(4-methylbenzene)
CH2-(4-tert-butylbenzene)
CH2-(3-trifluoromethylbenzene)
CH2-(4-bromobenzene)
CH2-pyrid-2-yl
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
10
7.5
5.4
3.7
0.4
8.8
15
21
15
4.2
6.3
3.1
53
54
6.1
2.8
>60
with an inhibition constant in the low nanomolar range (Kiapp
=
7 nm). Despite this potency and favourable physicochemical
properties, no activity against trypanosomes in vitro was ob-
served. The reasons for this remain unclear.
10
11[b]
12
13
14
15
16
17
18
19
20
21
22
23
Herein, we discuss in detail the medicinal chemistry pro-
gramme around the aminobenzimidazole derivative adopting
the alternative binding mode and structure–activity relation-
ships (SAR) of the series. In addition, we discuss the antiparasit-
ic activity of these compounds and the implications for chemi-
cal validation of the target.
CH2-pyrid-3-yl
CH2-naphth-1-yl
CH2-naphth-2-yl
CH2-(3,4-dichlorobenzene)
Results
C(=O)CH3
[a] Data for each compound was determined in duplicate. Kai pp values
were calculated using BatchKi software (BioKin); see Shanks et al. for
more details.[14] [b] Data for these compounds have been reported previ-
ously.[13]
Initial hit exploration
Initial work concerned systematic
SAR studies around compound 11,
Figure 3. Modifications made
guided by structure-based design,
crystallographically determined
to the aminobenzimidazole
template to explore struc-
ture–activity relationships.
binding modes and activity data.
The key substitutions made are
shown in Figure 3.
(compound 6) appeared to give more potent inhibition than a
two-carbon spacer (compound 5); phenyl, pyridine and naph-
thyl derivatives appeared tolerated (compounds 6, 19–22); lip-
ophilic and chlorine substituents on the phenyl ring (com-
pounds 7–18) gave a distinct SAR but none of the compounds
were more potent than or at least equipotent to compound
11. When N2 of 11 was acetylated, the compound was essen-
tially inactive (compound 23 in Table 1).
N1 and N2 substituents
Compounds were either purchased or synthesised using the
general method outlined in Scheme 1. Compounds were made
by treating the 2-aminobenzimidazole under mild basic condi-
R3 substituent
Preliminary optimisation of the N1-position revealed that the
3,4-dichlorobenzyl substituent (compound 11) was the most
favourable group in that position. Based on the crystallograph-
ically determined binding mode of 11 (Figure 2c), modelling
predicted further enhancements in activity could be achieved
in particular by substituents on the 4- and 7-positions of the
aminobenzimidazole filling hydrophobic pockets close to the
scaffold (Figure 2d).[13] The general synthetic route employed,
shown in Scheme 2, relies on the alkylation of the benzimida-
zole N1 in the final step, affording mixtures of 4- and 7- substi-
tuted benzimidazoles, which were then separated by chroma-
tographic methods. Scheme 2 shows the route employed to
make the 4- and 7-phenyl benzimidazole analogues. The 4’
and 7’ alkoxy analogues were synthesised from corresponding
alkylation of the commercially available 2-amino-3-nitrophe-
nol,[15,16] reduction of the nitro group[17] and cyclisation.[18]
Other synthetic methods were attempted, which would selec-
tively give access to either the 4- or 7-substituted products,
but these methods either failed or were lower yielding than
Scheme 1. Synthetic route to N1-substituted 2-aminobenzimidazoles. Re-
agents and conditions: a) KOH, EtOH, R1X, 208C, 18 h, 47–88%.
tions with the appropriate alkylating agent. Reaction condi-
tions were optimal at room temperature, where the reactions
were generally clean with substitution occurring exclusively on
the N1 nitrogen. Recovered product yields were in the range
of 47 to 88% of the theoretical maximum. In this way ana-
logues were obtained to explore SAR around the N1-position.
Key SAR data are shown in Table 1.
Compounds 3 and 4, where the N1 nitrogen is substituted
with ethyl and benzoyl groups respectively, were inactive.
However, compounds with an aromatic substituent attached
to a flexible linker showed inhibition. A one-carbon spacer
304
www.chemmedchem.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2011, 6, 302 – 308

Inhibitors of Trypanosoma brucei Pteridine Reductase 1
30) increased potency by almost 15-fold. The even more bulky
and rigid phenyl and O-benzyl groups decreased potency in
the 4-postion (26 and 27) while showing increased potency in
the 7-position (31 and 32) affording the most potent com-
pound in this series (compound 32), which has a Kai pp value of
7 nm.
Parasite activity of PTR1 inhibitors
Unfortunately, cell efficacy did not track the PTR1 enzyme po-
tency (Table 2) and no significant improvements in efficacy
against T. brucei were observed going from compound 11 to
compound 32. In addition, some toxicity was evident with
some compounds against the mammalian MRC5 cell line in
the micromolar range. Mammalian cells do not possess PTR1,
so off-target effects may be involved. Inhibition of mammalian
DHFR can be discounted, since none of these compounds in-
hibited this enzyme. Possible reasons for the poor trypanocidal
activity were investigated.
Scheme 2. Synthetic route to compounds 27 and 32. Optimisation of R3.
The NOESY interaction between the ortho proton of the phenyl ring and the
N1-CH2 group in 32 is indicated. Reagents and conditions: a) NH2OMe, tBuOK,
CuCl, DMF, 1 h RT, 20% (as in References [19,20]); b) EtOH, Sn(II)Cl2, 1408C,
MW, 10 min, >97%; c) BrCN in MeCN, EtOH 708C, 1.5 h, 63%; d) 3,4-dichlor-
obenzylbromide, EtOH, KOH, 208C, 18 h.
To determine whether there were any differences between
recombinant and endogenous PTR1, the potency of 32 was
measured in clarified lysates of T. brucei using a specific HPLC-
based assay instead of the cytochrome c-coupled assay that is
only suitable for studies on the purified enzyme (Figure 4). Ad-
equate extraction of the parasites was confirmed when trypa-
nothione reductase activity in the lysates (63.3ꢀ1.7 nmol
minꢁ1 mgꢁ1) was found to be in good agreement with previ-
ously published data.[22] The potency of compound 32 was
found to be virtually identical against PTR1 in cell lysates and
the purified recombinant enzyme (IC50 =0.88ꢀ0.13 nm, Hill
slope=0.5, and 0.99ꢀ0.13 nm, Hill slope 0.73, respectively).
However, due to the shallower Hill slope in the cell extract,
higher concentrations of inhibitor are required to completely
inhibit PTR1 activity (~1 mm). The low Hill slopes may be due
to PTR1 being a tetramer, and there being negative coopera-
tivity on binding to the individual subunits.
producing the final product mixtures and separating these by
chromatography. The absolute regiochemistry of the isolated
final compounds was assigned on the basis of the 1H NMR
spectra, the NOESY spectra, and X-ray crystallography, where
we have reported a complex with the enzyme.[13] In the NOESY
spectra for 31 and 32, we were able to see an interaction be-
tween the protons on the 7-substituent with the N1-CH2-
benzyl protons as indicated in Scheme 2.
Biological data for compounds with substituents in the 4- or
7-positions is shown in Table 2. All compounds with substitu-
ents in the 4- or 7-position are selective for TbPTR1 over
human and TbDHFR. Small substituents in the 4- or 7-position
(compounds 24, 28, 29 and 33) did not lead to a significant
improvement in potency against TbPTR1 compared to 11. The
slightly larger propoxy substituent in the 4-position (com-
pound 25) gave similar potency to the parent compound 11,
whereas the same substituent in the 7-position (compound
In order to achieve activity against the parasite, these inhibi-
tors need to pass through the cell membrane to reach PTR1 in
the cytosol. In general, the com-
pounds have reasonable proper-
ties for cellular penetration at
Table 2. Activity of R3-substituted compounds against TbPTR1, TbDHFR and human DHFR, and T. brucei and
MRC5 cells. R1 in all compounds is 3,4-dichlorobenzyl.
physiological pH: molecular
weights less than 400, one hy-
drogen-bond acceptor and two
or three hydrogen-bond donors.
The calculated logD values are
in the range of 3.5 to 4.5, and
the experimental values are in
the range of 3.0 to >4 (Table 3).
Therefore, the compounds are
within acceptable ranges for cel-
lular penetration, albeit at the
higher end of lipophilicity. How-
ever, given the high protein
binding of these compounds, it
is possible that only the free
fraction (1–3%) is able to reach
Compd
R3
TbPTR1
Kai pp [mm]
HsDHFR
IC50 [mm]
TbDHFR
IC50 [mm]
T. brucei[a]
EC50 [mm]
MRC5
EC50 [mm]
11[b]
24
25
26
H
4-Cl
4-O(CH2)2CH3
4-OBn
4-Ph
4-OMe
7-Cl
7-O(CH2)2CH3
7-OBn
7-Ph
7-OMe
5,6-dimethyl
0.8
2.3
0.31
16
>60
0.46
0.51
0.047
0.098
0.007
0.65
3.9
>50
>50
>50
>50
>30
>50
>50
>50
>50
>50
>50
>30
>50
>30
>30
>50
>30
>30
>30
>30
>50
>50
>30
>30
11
ND[d]
ND
>30
ND
ND
ND
9.6
6.7
27
ND
ND
>30
ND
ND
ND
21
27
28[c]
29[b]
30[b]
31
17
32[b]
33
9.9
ND
ND
>30
ND
ND
34
[a] In vitro cell assay conditions used were as reported previously.[21] [b] Some data for these compounds have
been reported previously.[13] [c] Contaminated with ~25% of compound 33. [d] ND=Not determined.
ChemMedChem 2011, 6, 302 – 308
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
305

I. H. Gilbert et al.
MED
this binding mode, the N1 undergoes no hydrogen-bonding
interaction with the protein, so substituents on this position
should not disrupt hydrogen-bonding interactions. However,
compound 3 with an ethyl substituent, failed to give signifi-
cant binding; whilst the ethyl substituent probably did not dis-
rupt binding, it did not give any significant interactions with
the enzyme, especially with Trp221, which is known to contrib-
ute substantially to binding affinity (unpublished results de-
rived from other compound series). The benzoate substituent
of compound 4 also diminished binding affinity. This is most
likely due to a decrease of the pKa of the N3 nitrogen of the
aminobenzimidazole, thereby effecting the hydrogen-bonding
interaction with Asp161 and to the change in geometry of the
substituent caused by the carbonyl group, which does not
allow the compound to form favourable interactions with
Trp221.
As was previously reported for compound 11,[13] the crystal
structure in complex with PTR1 revealed that the chloro-sub-
stituents on positions 3 and 4 fill a hydrophobic pocket
formed by Val206, Trp221, Lys224 and Leu263 (Figure 2c). Re-
moving or replacing one or more of the chlorine atoms and/or
changing the substitution pattern (compounds 6–10, 12–18)
resulted in weaker inhibitors, probably due to less than opti-
mal filling of this hydrophobic pocket. Similarly, the pyridyl an-
alogues (19 and 20) are weaker inhibitors, probably due to the
lack of hydrogen-bonding partners for the pyridyl nitrogen
atoms in the hydrophobic pocket and due to the pyridyl sub-
stituents not filling the hydrophobic pocket as optimally. The
naphthyl derivates (21 and 22) were designed to mimic the
3,4-dichlorophenyl group of 11. Compound 22 is only a slightly
weaker (fourfold) inhibitor than compound 11, whilst com-
pound 21 is tenfold weaker. This may be due to 2-naphthyl de-
rivative occupying the hydrophobic cavity in a slightly more fa-
vourable conformation than compound 21.
Figure 4. Activity of compound 32 against recombinant PTR1 and endoge-
nous PTR1 in a trypanosome lysate: recombinant PTR1 (*); clarified trypa-
nosome lysate (*). The uninhibited rates were 2.5 and 1.7 pmolminꢁ1 mLꢁ1
respectively.
,
Table 3. Physicochemical properties of compounds with trypanocidal ac-
tivity.
Compd MW
logD
PSA Solubility Exptl PB Predicted
Calcd Exptl [ꢁ2]
[mm]
[%]
BBB[b]
11
30
31
32
292
350
398
368
3.0
3.6
4.4
4.7
>4
3.0
>4
3.7
44
53
53
44
200–300
ND
ND
97
99
98
99
+
+
+
+
ND
[a] Calcd, calculated; Exptl, experimental (measured); PSA, polar surface
area; PB, protein binding. [b] Blood–brain barrier (BBB) permeability pre-
dicted by StarDrop (www.optibrium.com).
the target in whole cells. This does not seem to be the case,
since inclusion of either foetal calf serum (10% v/v) or bovine
serum albumin (1 mgmLꢁ1) in the assay only increased the IC50
values against recombinant PTR1 by two- to threefold (2.2ꢀ
0.4 and 2.9ꢀ0.1 nm, respectively).
The acetate derivative 23 was synthesized in order to probe
the flexibility of the binding pocket around Gly205 and
Asp161, which form hydrogen bonds with the amino group of
compound 11 in the binding mode determined by crystallog-
raphy.[13] There was a complete loss of activity of compound
23, indicating that the binding site is rather rigid in that area
and does not accommodate substituents of the amino group.
Our analysis of the crystal structure of PTR1·11 had revealed
two hydrophobic pockets that could be reached by substitu-
ents from the 4- and 7-positions (Figure 2d).[13] The pocket in
the 7-position bounded by residues Trp221, Met213, Val206
and Leu209 appeared to be larger than the pocket in the 4-po-
sition, formed by residues Pro167, Tyr174 and Asn175. Indeed,
smaller substituents at the 4- or 7-position are tolerated in
both pockets (compounds 24, 28, 29 and 33) but do not lead
to a significant increase in binding affinity. The slightly larger
propoxy group of compounds 25 and 30 leads to similar activi-
ty against the PTR1 compared to 11 when attached in the 4-
position, but to an almost 20-fold increase when attached in
the 7-position. A phenyl group in the 4-position (compound
27) leads to a greater than tenfold decrease in potency com-
pared to compound 11, probably caused by steric clashes in
this rather small hydrophobic pocket. In contrast, attachment
Discussion
During this work, we successfully prepared a number of 2-ami-
nobenzimidazoles as new PTR1 inhibitors. This work was
guided using structure-based design and medicinal chemistry
principles. The SAR that we observe here can be explained by
the crystal structure information previously reported on com-
plexes with the enzyme.[13]
Binding modes
All of the compounds that are reported here have a substitu-
ent at the 1-position (R1). Consequently, binding in modes simi-
lar to one or two (Figure 2a and 2b) is highly unlikely, as this
would disrupt the hydrogen-bonding interactions from N1. We
therefore assume that these compounds bind in a similar
manner as compound 11 in binding mode three (Figure 2c). In
306
www.chemmedchem.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2011, 6, 302 – 308

Inhibitors of Trypanosoma brucei Pteridine Reductase 1
of the same group to the 7-position results in an almost 100-
fold increase in potency, probably due to edge–face interac-
tions of this group with Trp221 and displacement of water
molecules from the hydrophobic pocket (Figure 2e). In gener-
al, no enhancement in activity was achieved by accessing the
putative pocket at the 4-position.
T. brucei is 480 nm of which 98% is present in the tetrahydro
form (Ong and Fairlamb, unpublished results). Assuming that
all of this has to be oxidised to dihydrobiopterin for lethality
and that enzyme inhibition must be maintained at 90% of
normal levels to successfully deplete the tetrahydrobiopterin
levels, then the required free concentration of 32 can be calcu-
lated from Equation (2).
ꢀ
ꢁ
Parasite activity and chemical validation
½Sꢂ
½Iꢂ0:9 ¼ 9 1 þ
Ki
ð2Þ
Km
Several potential reasons for the disappointingly large discrep-
ancy between PTR1 enzyme potency and cell efficacy have
been investigated using compound 32:
Using these assumptions (S=480 nm, Km =25 nm, Ki =
2.3 nm), the predicted concentration of unbound 32 required
to maintain 90% inhibition would be 418 nm, consistent with
the results in Figure 4. A similar calculation for 95% inhibition
yields 883 nm. Combined with the two- to threefold decrease
in potency due to protein binding, a trypanocidal effect would
only be expected in the 1–2 mm range, which is reasonably
consistent with the observed EC50 value of 10 mm in Table 2.
These theoretical calculations provide a plausible explanation
for the 1000-fold decrease in potency between target and cell,
and underline the need for the development of substantially
more potent competitive inhibitors since the current ratio of
Km/Ki (25 and 2.3 nm, respectively) is only 10.
From Equation (2), [I]0.9 is inversely related to Km/Ki. In sharp
contrast, the Km/Ki ratio for the folate analogue methotrexate
against murine DHFR is 260000[24] and 30000 for human
DHFR.[25] In trypanosomes, the only known targets for metho-
trexate are DHFR and PTR1 with Ki values of 0.15 and 3.6 nm,
respectively.[14] Significantly, methotrexate, which is equipotent
with 32 in respect of TbPTR1 inhibition, displays similar cell po-
tency in a genetically engineered cell line lacking DHFR com-
pared to the parental cell line used here (EC50 values of 17.9
and 9.9 mm, respectively). This supports the idea that the poor
cellular potency has more to do with Km/Ki ratios than the spe-
cific physicochemical explanations indicated for this particular
novel series.
*
Physicochemical factors: The physicochemical properties of
all the trypanocidal compounds appear within acceptable
ranges for cellular penetration, albeit at the higher end of lipo-
philicity (Table 3). However, the possibility that compound 32
is effectively sequestered in cellular lipids and therefore unable
to reach its cytosolic target enzyme or is actively effluxed out
of the cells cannot be definitively excluded by our current
studies.
*
Differences between the recombinant and endogenous
enzyme: It is clear that differences between recombinant and
endogenous native PTR1 do not appear to play a significant
role. Interestingly, the consistently low Hill slope observed here
and previously[13] could indicate negative cooperativity restrict-
ing ligand occupancy on all four subunits of the tetramer.[6]
*
Extent of enzyme inhibition: Although the extent to which
PTR1 has to be inhibited to cause growth effects is not known
with certainty, gene knockout experiments indicate that halv-
ing enzyme activity has no effect on growth and RNAi deple-
tion studies suggest >90% knockdown of PTR1 protein is re-
quired to exert a trypanocidal effect.[9] Therefore, it is possible
that compounds have insufficient potency at the enzyme level
to cause an effect at the cellular level. PTR1 has an apparent
Km (Kampp) value of 167ꢀ54 nm using the cytochrome c-coupled
assay[14] used to determine the Kai pp for 32, and the crystal
structure of 32 indicates that this inhibitor is competitive with
respect to the substrate dihydrobiopterin (Figure 2e). Thus, the
inhibition constant (Ki) for 32 can be calculated from Equa-
tion (1) for competitive inhibition.
Conclusions
Kiapp
We have established SAR for a series of novel PTR1 inhibitors.
The most potent compounds of this series have appropriate
druglike properties and are highly selective (>7000-fold) for
PTR1 over human or trypanosomal DHFR. Compounds 32 and
30 are the most potent and selective TbPTR1 inhibitors dis-
closed in the literature to date and will hopefully prove to be
useful pharmacological tools for the exploration of the role
PTR1 plays in the survival and growth of these parasites. How-
ever, in order to produce effective drug candidates directed
solely at PTR1, potency will need to be enhanced by at least
another two orders of magnitude.
Ki ¼
ð1Þ
1 þ S=Km
Applying Equation (1) yields a Ki value of 2.3 nm under our
assay conditions, where the substrate concentration (S) is
350 nm. Using the direct HPLC method, the Kampp value for PTR1
is 25 nm,[14] which is sevenfold lower than the value obtained
in the cytochrome c assay. This is due to the latter assay gener-
ating quinonoid dihydrobiopterin, which can then rearrange to
form 7,8-dihydrobiopterin.[23] Subsequent work from this labo-
ratory has established that T. brucei PTR1 can also reduce qui-
nonoid dihydrobiopterin to tetrahydrobiopterin (Ong and Fair-
lamb, unpublished). Thus, the Kampp value determined by the cy-
tochrome c method is a hybrid Kampp value for a mixture of
these substrates. Additional studies from our laboratory indi-
cate that the total intracellular biopterin concentration in
Experimental Section
The chemistry and biology experimental sections are in the Sup-
porting Information.
ChemMedChem 2011, 6, 302 – 308
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
307

I. H. Gilbert et al.
MED
[8] M. L. Cunningham, R. G. Titus, S. J. Turco, S. M. Beverley, Science 2001,
292, 285–287.
Acknowledgements
[9] N. Sienkiewicz, H. B. Ong, A. H. Fairlamb, Mol. Microbiol. 2010, 77, 658–
671.
We would like to acknowledge the Wellcome Trust (WT077705,
WT079838 and WT083481) for funding. The authors would like to
thank Mrs. Gina McKay for performing HRMS analyses and for
assistance with performing other NMR and MS analyses; Mr.
Daniel James, for data management; Mrs. Bhavya Rao and Ms.
Irene Hallyburton for performing the biological assays; Ms. Su-
zanne Norval for the DMPK work.
[10] A. Cavazzuti, G. Paglietti, W. N. Hunter, F. Gamarro, S. Piras, M. Loriga, S.
Alleca, P. Corona, K. McLuskey, L. Tulloch, F. Gibellini, S. Ferrari, M. P.
Costin, Proc. Natl. Acad. Sci. USA 2008, 105, 1448–1453.
[11] L. W. Hardy, W. Matthews, B. Nare, S. M. Beverley, Exp. Parasitol. 1997,
87, 158–170.
[12] L. B. Tulloch, V. P. Martini, J. Iulek, J. K. Huggan, J. H. Lee, G. C. L. , T. K.
Smith, C. J. Suckling, W. N. Hunter, J. Med. Chem. 2010, 53, 221–229.
[13] C. P. Mpamhanga, D. Spinks, L. B. Tulloch, E. J. Shanks, D. A. Robinson,
I. T. Collie, A. H. Fairlamb, P. G. Wyatt, J. A. Frearson, W. N. Hunter, I. H.
Gilbert, R. Brenk, J. Med. Chem. 2009, 52, 4454–4465.
[14] E. J. Shanks, H. B. Ong, D. A. Robinson, S. Thompson, N. Sienkiewicz,
A. H. Fairlamb, J. A. Frearson, Anal. Biochem. 2010, 396, 194–203.
[15] E. M. Doherty, C. Fotsch, A. W. Bannon, Y. X. Bo, N. Chen, C. Dominguez,
J. Falsey, N. R. Gavva, J. Katon, T. Nixey, V. I. Ognyanov, L. Pettus, R. M.
Rzasa, M. Stec, S. Surapaneni, R. Tamir, J. W. Zhu, J. J. S. Treanor, M. H.
Norman, J. Med. Chem. 2007, 50, 3515–3527.
[16] L. Zhang, J. H. Fan, K. Vu, K. Hong, J. Y. Le Brazidec, J. D. Shi, M. Bia-
monte, D. J. Busch, R. E. Lough, R. Grecko, Y. Q. Ran, J. L. Sensintaffar, A.
Kamal, K. Lundgren, F. J. Burrows, R. Mansfield, G. A. Timony, E. H. Ulm,
S. R. Kasibhatla, M. F. Boehm, J. Med. Chem. 2006, 49, 5352–5362.
[17] M. Rangarajan, J. S. Kim, S. P. Sim, A. Liu, L. F. Liu, E. J. LaVoie, Bioorg.
Med. Chem. 2000, 8, 2591–2600.
[18] N. J. Leonard, D. Y. Curtin, K. M. Beck, J. Am. Chem. Soc. 1947, 69, 2459–
2461.
[19] S. Seko, (Sumitomo Chemical Company Ltd.; Osaka, Japan), Process for
Preparing Nitroaniline Derivatives, United States Patent 5466871, 1995.
[20] S. Seko, K. Miyake, N. Kawamura, J. Chem. Soc. Perkin Trans. 1 1999,
1437–1444.
[21] D. Spinks, E. J. Shanks, L. A. T. Cleghorn, S. McElroy, D. Jones, D. James,
A. H. Fairlamb, J. A. Frearson, P. G. Wyatt, I. H. Gilbert, ChemMedChem
2009, 4, 2060–2069.
[22] T. J. Vickers, A. H. Fairlamb, J. Biol. Chem. 2004, 279, 27246–27256.
[23] W. L. F. Armarego, D. Randles, H. Taguchi, Eur. J. Biochem. 1983, 135,
393–403.
Keywords: antiprotozoal agents · drug discovery · pteridine
reductase · structure-based drug design · Trypanosoma brucei
[1] R. Pink, A. Hudson, M. A. Mouries, M. Bendig, Nat. Rev. Drug Discov.
2005, 4, 727–740.
[2] A. H. Fairlamb, Trends Parasitol. 2003, 19, 488–494.
[3] A. Gangjee, H. D. Jain, S. Kurup, Anti-Cancer Agents Med. Chem. 2007, 7,
524–542.
[4] A. Gangjee, H. D. Jain, S. Kurup, Anti-Cancer Agents Med. Chem. 2008, 8,
205–231.
[5] M. Berriman, E. Ghedin, C. Hertz-Fowler, G. Blandin, H. Renauld, D. C.
Bartholomeu, N. J. Lennard, E. Caler, N. E. Hamlin, B. Haas, W. Bohme, L.
Hannick, M. A. Aslett, J. Shallom, L. Marcello, L. H. Hou, B. Wickstead,
U. C. M. Alsmark, C. Arrowsmith, R. J. Atkin, A. J. Barron, F. Bringaud, K.
Brooks, M. Carrington, I. Cherevach, T. J. Chillingworth, C. Churcher, L. N.
Clark, C. H. Corton, A. Cronin, R. M. Davies, J. Doggett, A. Djikeng, T.
Feldblyum, M. C. Field, A. Fraser, I. Goodhead, Z. Hance, D. Harper, B. R.
Harris, H. Hauser, J. Hostetter, A. Ivens, K. Jagels, D. Johnson, J. Johnson,
K. Jones, A. X. Kerhornou, H. Koo, N. Larke, S. Landfear, C. Larkin, V.
Leech, A. Line, A. Lord, A. MacLeod, P. J. Mooney, S. Moule, D. M. A.
Martin, G. W. Morgan, K. Mungall, H. Norbertczak, D. Ormond, G. Pai,
C. S. Peacock, J. Peterson, M. A. Quail, E. Rabbinowitsch, M. A. Rajan-
dream, C. Reitter, S. L. Salzberg, M. Sanders, S. Schobel, S. Sharp, M. Sim-
monds, A. J. Simpson, L. Talton, C. M. R. Turner, A. Tait, A. R. Tivey, S.
Van Aken, D. Walker, D. Wanless, S. L. Wang, B. White, O. White, S.
Whitehead, J. Woodward, J. Wortman, M. D. Adams, T. M. Embley, K.
Gull, E. Ullu, J. D. Barry, A. H. Fairlamb, F. Opperdoes, B. G. Barret, J. E.
Donelson, N. Hall, C. M. Fraser, S. E. Melville, N. M. El-Sayed, Science
2005, 309, 416–422.
[24] J. W. Williams, J. F. Morrison, R. G. Duggleby, Biochemistry 1979, 18,
2567–2573.
[25] J. R. Appleman, N. Prendergast, T. J. Delcamp, J. H. Freisheim, R. L. Blak-
ley, J. Biol. Chem. 1988, 263, 10304–10313.
[6] B. Nare, L. W. Hardy, S. M. Beverley, J. Biol. Chem. 1997, 272, 13883–
13891.
Received: October 14, 2010
[7] A. R. Bello, B. Nare, D. Freedman, L. Hardy, S. M. Beverley, Proc. Natl.
Acad. Sci. USA 1994, 91, 11442–11446.
Revised: November 25, 2010
Published online on December 29, 2010
308
www.chemmedchem.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2011, 6, 302 – 308