Development of potent purine-derived nitrile inhibitors of the trypanosomal protease TbcatB

Article abstract of DOI:10.1021/jm070760l

Human African trypanosomiasis (HAT), a major health concern in sub-Saharan Africa, is caused by the protozoan parasite Trypanosoma brucei. Recent studies have shown that a cathepsin B like protease, TbcatB, is essential to the survival of T. brucei in vit

Full text of DOI:10.1021/jm070760l

J. Med. Chem. 2008, 51, 545–552
545
Development of Potent Purine-Derived Nitrile Inhibitors of the Trypanosomal Protease TbcatB
Jeremy P. Mallari,†,‡ Anang A. Shelat,‡ Terri Obrien,§ Conor R. Caffrey,§ Aaron Kosinski,‡ Michele Connelly,‡
Michael Harbut,| Doron Greenbaum,| James H. McKerrow,§ and R. Kiplin Guy*,†,‡
Graduate Program in Chemistry and Chemical Biology and Department of Cellular and Molecular Pharmacology, UniVersity of California,
San Francisco, California 94143-2280, Department of Chemical Biology and Therapeutics, St. Jude Children’s Research Hospital,
Memphis Tennessee 38105, and Department of Pharmacology, UniVersity of PennsylVania, Philadelphia, PennsylVania 19146
ReceiVed June 26, 2007
Human African trypanosomiasis (HAT), a major health concern in sub-Saharan Africa, is caused by the
protozoan parasite Trypanosoma brucei. Recent studies have shown that a cathepsin B like protease, TbcatB,
is essential to the survival of T. brucei in vitro (Mackey, Z. B.; O’Brien, T. C.; Greenbaum, D. C.; Blank,
R. B.; McKerrow, J. H. J. Biol. Chem. 2004, 279, 48426-48433). Herein, we describe the first inhibitors
of TbcatB, a series of purine nitriles. The compounds are potent trypanocides, killing the parasite with a
high degree of selectivity over a panel of three human cell lines. In addition, a predictive model of trypanocidal
activity was developed on the basis of potency against TbcatB and various calculated physical property
descriptors.
Introduction
several transporters to take up purines and purine analogues
from its host.21–24 The purine nitrile scaffold was chosen for
development as a trypanocide based on its established activity
as a protease inhibitor25 and its potential to take advantage of
purine-scavenging mechanisms to improve potency and selective
toxicity against the parasite. In this report, purine nitriles were
synthesized and evaluated for activity against TbcatB. Additionally,
specificity against human cathepsins L and B was assessed and
the general toxicity of these compounds was evaluated against a
panel of mammalian cell lines (HEK 293, HEP G2, and Raji).
Finally, a predictive model of trypanocidal activity was developed
on the basis of potency against TbcatB and various calculated
physical property descriptors.
It is estimated that more than 50 000 people are infected with
human African trypanosomiasis (HATa) and that 60 million
more are at risk of infection.2 The drugs currently in clinical
use are associated with toxicity, emerging resistance, and
impractical dosing regimens.3–5 Effective vaccination is
unlikely because of the antigenic variation of the parasite.6
One potential strategy for discovering small-molecule thera-
peutics is to target T. brucei’s cysteine proteases. Early research
demonstrated that irreversible peptidyl protease inhibitors are
trypanocidal,7 and over the past 2 decades there has been
significant progress toward validation of the parasite’s cysteine
protease activity as a therapeutic target.8–12 Until recently, the
presumed target of these inhibitors was a protease known
interchangeably as brucipain, trypanopain, and rhodesain, which
accounts for the majority of the parasite’s total cathepsin
activity.13 However, recently a cathepsin B like protease known
as TbcatB was shown to be essential for parasite survival in
both in vitro1 and in vivo models (Jim McKerrow, personal
communication). Although the function of TbcatB is not well
understood, RNAi studies showed that even a partial loss of
TbcatB activity leads to accumulation of transferrin, enlargement
of the endosome, and death of the parasite.1 In contrast, RNAi
knockdown of brucipain in vitro produces no phenotype,
suggesting that TbcatB may be the relevant protease target.1 In
this report, we describe the initial testing and optimization of a
purine-derived nitrile scaffold designed to target TbcatB.
Although alkyl nitriles and aryl nitriles are well established
inhibitors of cysteine proteases,14–20 the activity of purine-
derived nitriles is relatively unexplored. T. brucei expresses
Design
In the initial report of the scaffold, the authors described a
crystal structure that indicated that the 6-amino substituent binds
at the S2 site of cathepsin K.25 Because of the structural
similarity within the papain-like protease family, we reasoned
that the scaffold might similarly orient within the active site of
TbcatB. The sequence homology of the TbcatB S2 site to the
S2 sites of the human cathepsins (B, C, F, H, K, L, O, S, V, W,
and X) was determined. These studies indicated the S2 site of
TbcatB was most similar to that of cathepsin L, largely because
of the residue forming the bottom of the S2 pocket (Gly245 in
TbcatB and Ala245 in cathepsin L). These studies suggested
that larger hydrophobic alkyl and aromatic ligands are most
likely to form productive interactions at the S2 site,26–29 which
is consistent with combinatorial substrate-profiling studies of
TbcatB (Jim McKerrow, personal communication). The first
generation of inhibitors was designed to explore this hypothesis.
* To whom correspondence should be addressed. Phone: (901) 495-5714.
Fax: (901) 495-5715. E-mail: kip.guy@stjude.org.
Chemistry
†
Graduate Program in Chemistry and Chemical Biology, University of
Purine nitriles were synthesized by the general route (Scheme
1) previously described.25 Briefly, 1 was reacted with the
appropriate amines (X) in 2-butanol to install the 6-amino
substituent. The crude reaction mixture was concentrated in
vacuo and resuspended in DMF with K2CO3. Alkylation at N9
was accomplished by heating this mixture with the desired alkyl
bromides (Y) to give intermediates 2{X, Y}. Purification was
accomplished by flash chromatography, and overall yield for
California.
‡
St. Jude Children’s Research Hospital.
Department of Cellular and Molecular Pharmacology, University of
§
California.
|
University of Pennsylvania.
a Abbreviations: ANOVA, analysis of variance; DTT, dithiothreitol; FBS,
fetal bovine serum; HAT, human African trypanosomiasis; HIC, hydro-
phobic interaction chromotography; MHC, major histocompatibility com-
plex; TbcatB, cathepsin B like protease in Trypanosoma brucei.
10.1021/jm070760l CCC: $40.75
2008 American Chemical Society
Published on Web 01/04/2008

546 Journal of Medicinal Chemistry, 2008, Vol. 51, No. 3
Scheme 1. Synthesis of Purine-Derived Nitrilesa
Mallari et al.
a Conditions: (i) X-NH2, 2-butanol, 60°C, 3-12 h; (ii) Y-Br, K2CO3, DMF, 60°C, 10-16 h; (iii) NaCN, DMSO, µW, 180°C, 160 W, 5-20 min.
Table 1. Compound Numbering
the two reactions was 25-60%. Subsequent reaction with
sodium cyanide in DMSO with microwave acceleration afforded
the target nitriles 3{X, Y}, which were purified by preparative
C18 chromatography with a resulting yield of 30-70%. Purity
of target compounds was confirmed by LCMS on both C4 and
C18 columns.
Determining the Activity of Purine Inhibitors 3{X, Y}
Chemset 3{X, Y} was tested in vitro to determine inhibitory
activity against TbcatB, cathepsin L, and cathepsin B (Table
1). All compounds were also assayed to determine antiprolif-
erative activity against the cultured bloodstream form of T.
brucei. Finally, cytotoxicity was evaluated in HEK 293, Raji,
and HEP G2 cell lines to determine a therapeutic index. A subset
of compounds was tested for time-dependent activity against
TbcatB (Figure S1 in Supporting Information). These experi-
ments indicate these inhibitors’ activity is not time-dependent,
suggesting that these compounds act as reversible inhibitors.
This observation is further supported by activity-based probe
labeling experiments (data not shown) and is consistent with
previous reports indicating that nitrile-based inhibitors may
display either time-dependent or -independent behavior.17,20
Results and Discussion
This initial series produced several compounds with moderate
potency against TbcatB that inhibited parasite proliferation
(Table 2). Although modestly active in the proliferation assay,
compounds with a 2-methylbutyl, hexyl, or R-methylbenzyl
substituent at the 6-amino position were not inhibitors of TbcatB.
This discrepancy may be due to off-target effects in the parasite
or insolubility under the protease assay conditions. Cyclohexyl
and benzyl moieties at the 6-amino position were well tolerated
by the protease, and all displayed trypanocidal activity in the
low micromolar range. As the most amenable to medicinal
chemistry, the benzyl moiety was chosen for further optimiza-
tion. A second compound series was synthesized to optimize
interactions at the 6-aminobenzyl ring and to explore structural
requirements at the N9 position.
The first round of optimization produced several compounds
active against TbcatB and that were trypanocidal (Table 3).
Replacing the benzyl group with a phenylethyl group led to
decreased potency against the protease. Various substitution
patterns at the benzyl position were tolerated by the protease,
with the 4-Cl and 3,4-Cl2 substitutions resulting in potent
compounds in both the enzyme and proliferation assays. For
several compounds, the introduction of a hydroxyl group at the
end of an alkyl linker at the N9 position led to significantly
improved potency against TbcatB. On the basis of this observa-
tion, compound 3{12, 1} was selected for further optimization
at N9.
In this second round of optimization, various oxygen-
containing moieties were introduced at the N9 position (Table
4). It was observed that a variety of functional groups (alcohols,

Purine-DeriVed Nitrile Inhibitors
Table 2. First-Generation Inhibitors
Journal of Medicinal Chemistry, 2008, Vol. 51, No. 3 547
compd
X
Y
IC50 vs TbcatB, µM
EC50 vs T. brucei, µM
3{1, 13}
3{1, 17}
3{2, 11}
3{2, 14}
3{2, 17}
3{3, 12}
3{3, 13}
3{3, 15}
3{3, 17}
3{4, 12}
3{4, 13}
3{4, 17}
3{5, 13}
3{7, 12}
3{7, 13}
3{7, 15}
3{7, 17}
3{8, 12}
3{8, 13}
3{8, 17}
3{10, 12}
3{10, 13}
3{10, 15}
3{10, 17}
2-methylbutyl
2-methylbutyl
cyclohexyl
cyclohexyl
cyclohexyl
hexyl
hexyl
hexyl
hexyl
R-methylbenzyl
R-methylbenzyl
R-methylbenzyl
4-MeO-benzyl
3-Me-benzyl
3-Me-benzyl
3-Me-benzyl
3-Me-benzyl
benzyl
benzyl
benzyl
4-Cl-benzyl
4-Cl-benzyl
4-Cl-benzyl
4-Cl-benzyl
cyclopentyl
benzyl
propyl
cyclohexyl
benzyl
butyl
cyclopentyl
cyclohexylmethyl
benzyl
butyl
cyclopentyl
benzyl
cyclopentyl
butyl
cyclopentyl
cyclohexylmethyl
benzyl
butyl
cyclopentyl
benzyl
14 ( 3
5.0 ( 0.6
8.3 ( 0.8
8.3 ( 0.5
2.3 ( 0.2
>25
1.1 ( 0.2
3.4 ( 0.2
12 ( 1
1.4 ( 0.2
0.33 ( 0.05
0.44 ( 0.06
1.5 ( 0.5
3.6 ( 0.7
0.7 ( 0.1
1.5 ( 0.3
3.4 ( 0.6
1.3 ( 0.2
0.8 ( 0.2
0.27 ( 0.04
1.1 ( 0.2
2.2 ( 0.7
0.48 ( 0.06
1.7 ( 0.1
1.7 ( 0.2
2.0 ( 0.2
0.85 ( 0.09
1.2 ( 0.2
>10
>25
>25
>25
>25
>25
>25
7 ( 1
11 ( 2
>25
>25
3.3 ( 0.7
14 ( 4
>25
6 ( 1
10 ( 1
19 ( 4
>25
butyl
cyclopentyl
cyclohexylmethyl
benzyl
>25
1.5 ( 0.3
are used to treat African trypanosomiasis in humans and cattle,
respectively.30
Although the most potent TbcatB inhibitors are generally the
most potent trypanocides, there were discrepancies that com-
plicated a direct correlation between IC50 values in the protease
and parasite proliferation assays (Figure 2). This complication
was not unexpected, as antiproliferative activity is likely to be
a function of multiple variables. To evaluate the relationship
between inhibition of TbcatB and parasite proliferation and to
determine the potential significance of other parameters, a
regression model of trypanocidal activity was developed. An
unbiased analysis was carried out to identify parameters
displaying a statistically significant correlation with trypanocidal
activity (see Supporting Information). Following an iterative
model-building process, the final model contained three descrip-
tors (standard errors in parentheses):
T. brucei log10(IC50) ) [0.158(0.030)][TbcatB log10(IC50)] -
[0.286(0.031)][SlogP_VSA9] - [0.202(0.040)][a_count] -
[0.159(0.028)]
Figure 1. Lowest energy pose of 3{12, 2} (space filling representation,
colored by atom type) bound to a homology model of TbcatB.
Hydrophobic pockets are colored green, polar pockets are purple, and
exposed surfaces are red.
where a_count is the total number of atoms including hydrogens
and SlogP_VSA9 is the sum of van der Waals surface area from
atoms with an SlogP parameter greater than 0.4. SlogP is an
atom-based method to estimate the log(octanol/water) partition
coefficient.31 Atoms assigned SlogP values of >0.4 are highly
hydrophobic. All calculated physical property descriptors were
generated in MOE (Chemical Computing Group, version
2006.08).
In the final model, 7 of 56 compounds were flagged as outliers
(3{1, 17}, 3{2, 11}, 3{2, 17}, 3{3, 15}, 3{7, 15}, 3{8, 9}, 3{10,
15} (Figure 3 and Supporting Information Table S1). Removal
of these points followed by model optimization increased
ethers, diols) were tolerated by the protease and that the length
of the alkyl linker could also be varied without strongly affecting
activity. This strategy produced the best TbcatB inhibitors of
the series, with 3{12, 2} being the most potent (270 nM, Table
4). Figure 1 shows the lowest energy pose of 3{12, 2} bound to
a homology model of TbcatB (see Supporting Information for
modeling details). This subset of inhibitors also yielded the three
most potent trypanocides, with compound 3{12, 4} as the most
active (60 nM, Table 4). These potencies are comparable with
those of suramin (130–360 nM) and berenil (20–100 nM), which

548 Journal of Medicinal Chemistry, 2008, Vol. 51, No. 3
Table 3. Second-Generation Inhibitors
Mallari et al.
compd
X
Y
IC50 vs TbcatB, µM
IC50 vs T. brucei, µM
3{13, 10}
3{13, 12}
3{13, 17}
3{5, 12}
3{5, 17}
3{6, 2}
3{6, 12}
3{6, 17}
3{8, 2}
phenylethyl
phenylethyl
phenylethyl
4-MeO-benzyl
4-MeO-benzyl
4-Me-benzyl
4-Me-benzyl
4-Me-benzyl
benzyl
pyridine-3-ylmethyl
butyl
benzyl
butyl
benzyl
3-hydroxypropyl
butyl
benzyl
3-hydroxypropyl
tetrahydropyran-2-yl methyl
4-MeO-benzyl
3-CF3-benzyl
4-hydroxybutyl
butyl
10 ( 2
16 ( 4
>25
4.4 ( 0.4
1.5 ( 0.1
2.4 ( 0.2
6.0 ( 0.8
4.5 ( 0.6
4.2 ( 0.5
>25
0.39 ( 0.05
1.0 ( 0.1
0.9 ( 0.2
0.55 ( 0.08
0.56 ( 0.06
2.9 ( 0.2
>25
12 ( 3
11 ( 1
11 ( 2
2.8 ( 0.5
18 ( 2
7.1 ( 0.6
3.3 ( 0.3
>25
3{8, 9}
benzyl
benzyl
benzyl
0.24 ( 0.03
0.40 ( 0.06
0.26 ( 0.03
0.8 ( 0.2
3{8, 16}
3{8, 18}
3{9, 3}
4-F-benzyl
4-F-benzyl
4-F-benzyl
4-Cl-benzyl
4-Cl-benzyl
4-Cl-benzyl
4-Cl-benzyl
4-Cl-benzyl
4-Br-benzyl
4-Br-benzyl
4-Br-benzyl
3,4-Cl2-benzyl
3,4-Cl2-benzyl
3,4-Cl2-benzyl
3{9, 12}
3{9, 17}
3{10, 10}
3{10, 2}
3{10, 3}
3{10, 16}
3{10, 18}
3{11, 12}
3{11, 17}
3{11, 19}
3{12, 1}
3{12, 12}
3{12, 17}
0.7 ( 0.2
benzyl
0.9 ( 0.1
pyridine-3-ylmethyl
3-hydroxypropyl
4-hydroxybutyl
4-MeO-benzyl
3-CF3-benzyl
butyl
0.5 ( 0.1
0.57 ( 0.03
4.8 ( 0.5
23 ( 4
>25
8.4 ( 0.9
>25
>25
1.4 ( 0.2
4.1 ( 0.5
15 ( 4
0.31 ( 0.06
0.60 ( 0.09
0.27 ( 0.06
0.34 ( 0.06
0.54 ( 0.08
1.1 ( 0.2
benzyl
phenylpropyl
2-hydroxyethyl
butyl
0.42 ( 0.05
0.6 ( 0.1
0.18 ( 0.02
0.26 ( 0.04
benzyl
Table 4. Optimization at N9
compd
X
Y
IC50 vs TbcatB, µM
IC50 vs T. brucei, µM
3{12, 1}
3{12, 2}
3{12, 5}
3{12, 4}
3{12, 8}
3{12, 6}
3{12, 7}
3,4-Cl2-benzyl
3,4-Cl2-benzyl
3,4-Cl2-benzyl
3,4-Cl2-benzyl
3,4-Cl2-benzyl
3,4-Cl2-benzyl
3,4-Cl2-benzyl
2-hydroxyethyl
3-hydroxypropyl
2,3-dihydroxypropyl
5-hydroxypentyl
2-(methoxymethoxy)ethyl
3-methoxypropyl
1.4 ( 0.2
0.27 ( 0.02
1.2 ( 0.2
0.93 ( 0.03
0.8 ( 0.1
2.5 ( 0.3
3.4 ( 0.3
0.6 ( 0.1
0.11 ( 0.01
0.45 ( 0.1
0.06 ( 0.01
0.32 ( 0.04
0.15 ( 0.02
0.22 ( 0.03
3,3-dimethoxypropyl
Figure 2. TbcatB inhibition and trypanocidal activity.
adjusted R2 from 0.42 to 0.75 and decreased residual standard
error (rse, in log units) from 0.35 to 0.19 (df ) 45). Cross-
validated root-mean-squared error (rmse, in log units) was 0.20,
approximately equal to the final model rse, suggesting the

Purine-DeriVed Nitrile Inhibitors
Journal of Medicinal Chemistry, 2008, Vol. 51, No. 3 549
in intracellular accumulation could result from transporter
mediated uptake.
To confirm that the purine nitrile scaffold is able to access
and inhibit TbcatB in an intact parasite, activity-based probe
experiments were performed (data not shown). However, the
reversible nature of the scaffold and the relatively fast off-rate
of these compounds precluded the use of this method, which
relies on competition with a nonspecific irreversible inhibitor
to detect direct interaction between TbcatB and the inhibitor in
vivo.
Protease Specificity. Cathepsins B and L are widely ex-
pressed in a variety of tissues in mammals.32 Although cathe-
psins B and L are not vital in a mouse model, cathepsin L
knockout mice display problems that include abnormal skin
morphology and impaired MHC class II mediated antigen
presentation.33,34 The S2 site of TbcatB is most similar to that
of cathepsin L, whereas the complete sequence of TbcatB is
most homologous to that of cathepsin B. This suggests that the
two proteases are likely candidates for off-target effects. To
evaluate the potential utility of TbcatB directed inhibitors, we
determined the specificity of this compound series for TbcatB
relative to human cathepsins B and L.
Figure 3. Experimental trypanocidal activity versus predicted try-
panocidal activity. Outliers are plotted in blue. The red line has unity
slope and represents perfect predictions.
absence of overfitting. All covariates were statistically significant
by ANOVA (p < 10-5) and predictive as assessed by cross-
validation.
SlogP_VSA9 is a measure of hydrophobicity and may act as
a proxy for membrane permeability. The a_count descriptor is
more difficult to interpret. In general, increasing the number of
atoms augments molecular weight and surface area, thereby
decreasing membrane permeability. On the other hand, a_count
includes hydrogen atoms, so the parameter would increase by
substituting a heteroatom for unsaturated carbon. In light of the
compounds studied in this series, it is reasonable to interpret
a_count as capturing another aspect of hydrophobicity. In the
proposed model, then, activity against the parasite increases with
both hydrophobicity and potency vs TbcatB, suggesting that
trypanocidal activity is determined by membrane permeability
and activity against the protease target.
However, this does not rule out activity due to off-target
effects. There are compounds in this series inactive against
TbcatB that still display trypanocidal activity. This observation
is not necessarily inconsistent with the proposed model,
considering that increased membrane permeability may also
increase activity against other intracellular targets. We hypoth-
esize that although the best trypanocides are generally the best
TbcatB inhibitors, the high membrane permeability of some
inhibitors in this series may allow compounds with poor activity
against TbcatB to produce off-target effects in the cultured
parasites. It is interesting to note that in this inhibitor series,
six compounds contained a cyclohexyl moiety, and of these six,
five were outliers in this model. This may suggest that the
cyclohexyl ring is involved in an off-target effect or that
there is something peculiar about this group in the context of
the calculations used for this model. The discrepancies between
the two activities may also be due to differences in intracellular
inhibitor concentrations in the parasite assay. Because these
inhibitors are based on a purine template, significant differences
The preference of TbcatB’s S2 binding site for large
hydrophobic ligands is quite similar to those of cathepsins B
and L. Therefore, it is not surprising that there was no significant
selectivity between TbcatB and cathepsin L (Figure 4). However,
this compound series was generally highly selective for TbcatB
relative to human cathepsin B with only three compounds
displaying IC50 values below 10 µM against the human protease
(Figure 4). It is interesting to note that all three of these
compounds have a 3-hydroxypropyl substituent at the N9
position. Changing the propyl linker to an ethyl or pentyl
completely abolished activity against cathepsin B, as did
replacement of the hydroxyl group with a methyl. This result
suggests that the 3-hydroxypropyl moiety is making a specific
interaction necessary for effective inhibition of cathepsin B,
although TbcatB and cathepsin L do not share this requirement.
Molecular modeleling was used to further explore this
hypothesis. Like cathepsin B, the prime side of TbcatB’s active
site cleft is occluded by a peptide loop that is unique among
members of the papain superfamily of proteases. This region
adopts nearly identical conformations in the homology model
of TbcatB and two reported crystal structures of human
cathepsin B except around residues 121–123 (Gly-Glu-Gly in
human cathepsin B). In TbcatB, a phenylalanine insertion and
the lack of flanking glycine residues were predicted to force
the main chain (Gln-Phe-Asn-Phe) away from the active site,
thereby opening space around the putative location of the purine
N9 position (Figure 5). The cathepsin B sequence affords
multiple conformations: in 1gmy, Glu122 clashes with the N9
substitutent, whereas the same residue in 1huc is directed into
the solvent and the Gly121 carbonyl projects into the prime
side of the active site cleft. We hypothesized that the 3-hy-
Figure 4. Inhibitor specificity between TbcatB and human cathepsins B and L.

550 Journal of Medicinal Chemistry, 2008, Vol. 51, No. 3
Mallari et al.
deeper S2 site of TbcatB should be able to accommodate
substituents at the 6-amino position with more steric bulk and
a positively ionizable functional group. This hypothesis will be
explored in future studies.
It is possible that the trends in inhibition observed for TbcatB
and cathepsin L may be due to differences in the inhibitors’
reactivities at the nitrile warhead. However, there are several
examples of compounds with a wide range of potencies in the
protease assays that would be expected to be similarly electrophilic
at the nitrile moiety. This suggests that binding affinity plays a
primary role in determining potency, although electrophilicity
cannot be ruled out as a contributing factor.
Toxicity Studies. General cytotoxicity of each inhibitor was
evaluated by screening each at 25 µM in cultures of Raji (a
lymphoblastoid cell line derived from a Burkitt’s lymphoma),
HEK 293 (a human embryonic kidney cell line), and HEP G2
(a liver cell line derived from a human hepatoblastoma). Of
the human cell lines, HEK 293 was by far the most sensitive.
Therefore, a cellular therapeutic index for each inhibitor was
determined by calculating the ratio of the EC50 values against
HEK 293 and T. brucei (Figure 7). At 25 µM the first generation
of inhibitors caused some toxicity, but second-generation
compounds produced almost none. No inhibitors displayed EC50
values below 20 µM in HEK 293 cultures, with six compounds
exhibiting greater than a 100-fold therapeutic index for T. brucei
over any of the mammalian cell lines. These data suggest an
adequate therapeutic index. It is notable that as potency against
TbcatB increased, parasites were killed more selectively relative
to the HEK293 line. This suggests that increasing potency
against TbcatB may reduce off-target effects. Together, these
data further support the proposal that targeting TbcatB is a viable
strategy to develop therapies for HAT.
Figure 5. A hydrogen bond (dotted green line) between the hydroxyl
moiety on the 3-hydroxyproply substituent of 3{12, 2} and the Gly121
carbonyl stabilizes the 1huc S′ loop conformation of human cathepsin
B. The oxygen–oxygen distance is 2.81Å. The 1gmy and Tbcatb
conformations are depicted in brown and blue wire representations,
respectively.
droxylpropyl group stabilizes the open conformation observed
in 1huc by hydrogen-bonding to this carbonyl. Smaller or larger
N9 substituents lacking this interaction would be unfavorable
for cathepsin B binding. Indeed, a hydrogen bond between the
hydroxyl moiety in 3{12, 2} and the Gly121 carbonyl was
observed in predicted structures following minimization of one
conformation of the ligand within the 1huc active site, although
an exhaustive enumeration of starting conformers and the role
of bound waters in the active site were not explored.
Although good selectivity was observed for most inhibitors
against cathepsin B, specificity against cathepsin L was not
achieved in this series. However, a comparison of the S2 pockets
from the homology model of TbcatB and the 1mhw structure
of human cathepsin L suggests steric and electronic differences
that can be exploited to achieve specificity (Figure 6). The
Conclusions
This paper describes the development of a series of purine-
derived nitriles representing the first reported inhibitors of
TbcatB. Homology modeling techniques were used to guide
Figure 6. Electrostatic potential surfaces of the S2 pockets in TbcatB and catL reveal exploitable steric and electronic differences. (1) The lowest
energy pose of 3{12, 2} bound to the homology model of TbcatB shows the 3,4-dichlorobenzyl side chain packed against solvent exposed Asp73.
The deep S2 pocket is hydrophobic in the vicinity of Gly188 and bounded by an electronegative region marked by Asp165 and Asp234. (2) The
S2 pocket of catL is occluded by Met161 and tends to be more electropositive. In addition, Leu69 replaces Asp73 on the other side of the pocket.
Compound 3{12, 2} is superimposed for reference.
Figure 7. Activity summary sorted by potency against TbcatB. Selective toxicity is defined as (EC50 HEK 293)/(EC50 T. brucei).

Purine-DeriVed Nitrile Inhibitors
Journal of Medicinal Chemistry, 2008, Vol. 51, No. 3 551
(Amersham) in an Akta Explorer Box-900 chromatography system
(Amersham). HIC loading buffer was run through the column, and
protein was eluted by a linear decrease in ammonium sulfate from
0.7 to 0 M over 90 min at a flow rate of 1 mL/min. Fractions were
assayed for proteolytic activity as described below, and active
fractions were pooled. The pooled fractions were buffer-exchanged
into anion-exchange loading buffer (50 mM MES, pH 6.5) and
concentrated in an Amicon centrifugal filtration device (Millipore).
This solution was then loaded onto a Mono Q 5/50 GL anion
exchange column (Amersham) in an Akta Explorer Box-900
chromatography system. Anion exchange loading buffer was run
through the column, and protein was eluted by a linear increase in
NaCl concentration from 0 to 0.5 M over 55 min at a flow rate of
1 mL/min. Fractions were assayed for proteolytic activity, and active
fractions were checked for purity by SDS-PAGE and pooled.
Protease concentration was determined by active site titration with
CA074.
Enzyme Inhibition Assays. All IC50 values were determined
by sigmoidal curve-fitting with GraphPad Prism software.
TbcatB. An amount of 100 µL of TbcatB (final assay concentra-
tion, 35 nM) in assay buffer (0.05 M sodium acetate, pH 5.5, 4
mM DTT, 0.5 mM EDTA) was added to each well of a black flat-
bottom polystyrene 96-well plate (Corning). An amount of 2 µL
of inhibitor in DMSO was added, and the mixture was incubated
at room temperature for 20 min before the addition of 100 µL of
Z-Phe-Arg-AMC (final assay concentration, 10 µM; Bachem)
substrate in assay buffer. VO was determined by monitoring
fluorescence (λex ) 350 nm, λem ) 460 nm) on an Envision plate
reader (Perkin-Elmer) for 10 min. All data were collected in
triplicate.
Cathepsin L. An amount of 12 µL of human liver cathepsin L
(final assay concentration, 0.21 ng/mL; Biomol) in assay buffer
(0.05 M sodium acetate, pH 5.5; 4 mM DTT; 0.01% Triton X-100;
0.5 mM EDTA) was added to each well of a black polystyrene
384-well assay plate (Corning). An amount of 240 nL of inhibitor
in DMSO was added, and the mixture was incubated at room
temperature for 20 min before the addition of 12 µL of Z-Phe-
Arg-AMC substrate (final assay concentration, 10 µM) in assay
buffer. VO was determined by monitoring fluorescence (λex ) 350
nm, λem ) 460 nm) on an Envision plate reader (Perkin-Elmer) for
10 min. All data were collected in quadruplicate.
Cathepsin B. An amount of 12 µL of human liver cathepsin L
(final assay concentration, 0.56 ng/mL; Biomol) in assay buffer
(0.05 M sodium acetate pH 5.5; 4 mM DTT; 0.01% Triton X-100;
0.5 mM EDTA) was added to each well of a black polystyrene
384-well plate (Corning). An amount of 240 nL of inhibitor (240
nL) in DMSO was added, and the mixture was incubated at room
temperature for 20 min before the addition of 12 µL of Z-Arg-
Arg-AMC substrate (final assay concentration, 10 µM; Bachem)
in assay buffer. VO was determined by monitoring fluorescence (λex
) 350 nm, λem ) 460 nm) on an Envision plate reader (Perkin-
Elmer) for 10 min. All data were collected in quadruplicate.
Cell Proliferation Assays. All IC50 values were determined by
curve-fitting with GraphPad Prism software.
inhibitor design at the 6-amino position, and substituents at the
6-amino and N9 positions were optimized. This process yielded
several inhibitors with submicromolar potency against TbcatB
and T. brucei. A regression model of trypanocidal activity was
developed on the basis of potency against TbcatB and various
measures of hydrophobicity. Several compounds with a high
degree of selectivity for TbcatB versus human cathepsin B were
discovered. Modeling offered insight into achieving specificity
against cathepsin L. In addition, the general toxicity of these
compounds was evaluated in three mammalian cell lines, and a
high degree of selectivity against the parasite was demonstrated.
Experimental Section
Synthesis. Amines and alkyl bromides were purchased from
Sigma Aldrich, TCI America, and Matrix Scientific. Sodium
cyanide, potassium carbonate, and 2,6-dichloropurine 1 were
purchased from Sigma Aldrich. All reaction solvents were anhy-
drous. LCMS data were collected on a Waters system using XTerra
C18 columns (MeOH/H2O, 1% HCOOH).
Inhibitors were synthesized as previously described25 with minor
modifications. Briefly, 2,6-dichloropurine 1 (0.25 g, 1.3 mmol) was
suspended in 3.9 mL of 2-butanol, followed by the dropwise
addition of the appropriate amine (X) (3.9 mmol). The mixture was
stirred for 3-12 h at 60 °C and monitored by thin layer
chromatography (silica gel 60). The crude product mixture was then
concentrated under vacuum and resuspended in 5 mL of anhydrous
DMF. K2CO3 (solid, 0.54 g, 3.9 mmol) was added, followed by
dropwise addition of the appropriate alkyl bromide (Y) (5.2 mmol).
The mixture was stirred for 12-18 h at 60 °C and monitored by
LCMS. The crude mixture was concentrated under vacuum and
purified by flash chromatography (30:1 DCM/MeOH v/v) using
40M columns (Biotage) on a Biotage SP1 flash chromatography
instrument to give intermediates 2{X, Y} with an overall yield of
25-60%.
An amount of 0.05 mmol of 2{X, Y} was dissolved in 0.5 mL
of DMSO in a 1 mL microwave reaction vial (Biotage). Sodium
cyanide (0.35 mmol) was added. The vial was sealed and heated
to 180 °C for 10 min in a Biotage Initiator microwave until the
reaction was complete as shown by LCMS. The crude product was
purified by reverse-phase HPLC (Waters XTerra preparative C18
column, ACN/H2O, 1% HCOOH) to afford target nitrile 3{X, Y}.
The yield ranged from 30% to 70%.
Cloning of TbcatB. The gene encoding the TbcatB open reading
frame was isolated from T. brucei cDNA and cloned into the TOPO
TA vector (Invitrogen) as described.1 By use of this as a PCR
template, the cDNA encoding the TbcatB zymogen was amplified,
purified, and ligated into the pPICZalpha B vector (Invitrogen) for
eventual methanol-induced expression in Pichia pastoris using
methods previously described.13 Cloning was unidirectional, using
the following primers: forward 5′-ATACTCGAGAAAAGAG-
TAAACGCCGCCCTCGTTGCT-3′, containing a restriction endo-
nuclease site for Xho I and a Kexin II cleavage site, and reverse
5′-ATAGCGGCCGCTCACGCCGTGTTGGGTGCAAGAGG-3′,
containing a restriction site for Not I immediately preceding a
translation termination codon.
Expression of TbcatB in the Yeast Pichia pastoris. One liter
of yeast-peptone-dextrose (YPD) medium with 100 µg/mL Zeocin
(Invitrogen) was inoculated with 2.0 mL of P. pastoris recombinant
for TbcatB. The culture was expanded at 29 °C until an A600 value
of 2-3 was reached. The culture was then centrifuged at 1000g
for 10 min at room temperature (Eppendorf A-4-62 rotor). The
resulting pellet was resuspended in 200 mL of buffered minimal
methanol (BMM) induction medium (Invitrogen) containing
0.5% methanol and incubated at 29 °C for 48 h with the addition
of methanol after 24 h. The medium was centrifuged at 1000g for
10 min, and the supernatant was decanted and lyophilized. The
crude lyophilized solid was dissolved in HIC loading buffer (1.5
M ammonium sulfate, 50 mM K2HPO4/KH2PO4, pH 7.0) and
loaded onto a Tricorn 10/100 column packed with Phe15 resin
T. b. brucei. Culture-adapted T. brucei were grown at 37 °C in
5% CO2 in HMI-9 medium (HyClone) supplemented with penicillin/
streptomycin (50 units/mL), 10% heat-inactivated FBS (Omega
Scientific, lot no. 3137), and 10% Serum Plus (JHR Biosciences)
to a density of 1 × 106 cells/mL and then diluted to 1 × 104 cells/
mL. An amount of 100 µL of the diluted culture was added to
each well of a white flat-bottom sterile 96-well plate (Greiner), and
1 µL of inhibitor in DMSO was added. Plates were incubated for
48 h at 37 °C, in 5% CO2, and then equilibrated at room temperature
for 1 h before the addition of 100 µL of Cell Titer Glo (Promega)
to each well. Plates were then shaken on an orbital shaker for 2
min at 500 rpm. Luminescence was read after 8 min on an Envision
plate reader (Perkin-Elmer). All data were collected in quadrupli-
cate. (Note: For this inhibitor series it was noted that IC50 values
varied by as much as 10-fold depending on the lot of FBS used in
the medium.)

552 Journal of Medicinal Chemistry, 2008, Vol. 51, No. 3
Mallari et al.
(13) Caffrey, C. R.; Hansell, E.; Lucas, K. D.; Brinen, L. S.; Alvarez
Hernandez, A.; Cheng, J.; Gwaltney, S. L., 2nd; Roush, W. R.;
Stierhof, Y. D.; Bogyo, M.; Steverding, D.; McKerrow, J. H. Mol.
Biochem. Parasitol. 2001, 118, 61–73.
(14) Hanzlik, R. P.; Zygmunt, J.; Moon, J. B. Biochim. Biophys. Acta 1990,
1035, 62–70.
(15) Greenspan, P. D.; Clark, K. L.; Tommasi, R. A.; Cowen, S. D.;
McQuire, L. W.; Farley, D. L.; van Duzer, J. H.; Goldberg, R. L.;
Zhou, H.; Du, Z.; Fitt, J. J.; Coppa, D. E.; Fang, Z.; Macchia, W.;
Zhu, L.; Capparelli, M. P.; Goldstein, R.; Wigg, A. M.; Doughty, J. R.;
Bohacek, R. S.; Knap, A. K. J. Med. Chem. 2001, 44, 4524–4534.
(16) Greenspan, P. D.; Clark, K. L.; Cowen, S. D.; McQuire, L. W.;
Tommasi, R. A.; Farley, D. L.; Quadros, E.; Coppa, D. E.; Du, Z.;
Fang, Z.; Zhou, H.; Doughty, J.; Toscano, K. T.; Wigg, A. M.; Zhou,
S. Bioorg. Med. Chem. Lett. 2003, 13, 4121–4124.
(17) Robichaud, J.; Oballa, R.; Prasit, P.; Falgueyret, J. P.; Percival, M. D.;
Wesolowski, G.; Rodan, S. B.; Kimmel, D.; Johnson, C.; Bryant, C.;
Venkatraman, S.; Setti, E.; Mendonca, R.; Palmer, J. T. J. Med. Chem.
2003, 46, 3709–3727.
(18) Palmer, J. T.; Bryant, C.; Wang, D. X.; Davis, D. E.; Setti, E. L.;
Rydzewski, R. M.; Venkatraman, S.; Tian, Z. Q.; Burrill, L. C.;
Mendonca, R. V.; Springman, E.; McCarter, J.; Chung, T.; Cheung,
H.; Janc, J. W.; McGrath, M.; Somoza, J. R.; Enriquez, P.; Yu, Z. W.;
Strickley, R. M.; Liu, L.; Venuti, M. C.; Percival, M. D.; Falgueyret,
J. P.; Prasit, P.; Oballa, R.; Riendeau, D.; Young, R. N.; Wesolowski,
G.; Rodan, S. B.; Johnson, C.; Kimmel, D. B.; Rodan, G. J. Med.
Chem. 2005, 48, 7520–7534.
(19) Altmann, E.; Aichholz, R.; Betschart, C.; Buhl, T.; Green, J.; Lattmann,
R.; Missbach, M. Bioorg. Med. Chem. Lett. 2006, 16, 2549–2554.
(20) Loser, R.; Schilling, K.; Dimmig, E.; Gutschow, M. J. Med. Chem.
2005, 48, 7688–7707.
(21) Carter, N. S.; Fairlamb, A. H. Nature 1993, 361, 173–176.
(22) Barrett, M. P.; Zhang, Z. Q.; Denise, H.; Giroud, C.; Baltz, T. Mol.
Biochem. Parasitol. 1995, 73, 223–229.
(23) Maser, P.; Sutterlin, C.; Kralli, A.; Kaminsky, R. Science 1999, 285,
242–244.
(24) De Koning, H. P. Mol. Pharmacol. 2001, 59, 586–592.
(25) Altmann, E.; Cowan-Jacob, S. W.; Missbach, M. J. Med. Chem. 2004,
47, 5833–5836.
(26) Maciewicz, R. A.; Etherington, D. J. Biochem. J. 1988, 256, 433–
440.
(27) Choe, Y.; Leonetti, F.; Greenbaum, D. C.; Lecaille, F.; Bogyo, M.;
Bromme, D.; Ellman, J. A.; Craik, C. S. J. Biol. Chem. 2006, 281,
12824–12832.
(28) Gosalia, D. N.; Salisbury, C. M.; Ellman, J. A.; Diamond, S. L. Mol.
Cell. Proteomics 2005, 4, 626–636.
(29) Puzer, L.; Cotrin, S. S.; Alves, M. F.; Egborge, T.; Araujo, M. S.;
Juliano, M. A.; Juliano, L.; Bromme, D.; Carmona, A. K. Arch.
Biochem. Biophys. 2004, 430, 274–283.
(30) Merschjohann, K.; Sporer, F.; Steverding, D.; Wink, M. Planta Med.
2001, 67, 623–627.
(31) Wildman, S. A; Crippen, G. M. J. Chem. Inf. Comput. Sci. 1999, 39,
868–873.
Toxicity Screens. Raji, HEK 293, and HEP G2 cell lines were
purchased from the American Type Culture Collection (ATCC,
Manassas, VA) and were cultured according to recommendations.
Cell culture media were purchased from ATCC. Cells were
routinely tested for mycoplasma contamination using the MycoAlert
mycoplasma detection kit (Cambrex). Exponentially growing cells
were plated in Costar 96-well custom assay plates, which were black
with clear bottoms (Corning), and incubated overnight at 37 °C in
a humidified 10% CO2 incubator. DMSO inhibitor stock solutions
were added the following day to a final concentration of 25 µM,
0.5% DMSO. Cytotoxicity was determined following a 72 h
incubation using the Alamar blue assay. Fluorescence (λex ) 510
nm, λem ) 590 nm) was measured on an Envision plate reader
(Perkin-Elmer) 4 h after the addition of Alamar blue.
Acknowledgment. This work was supported by the Ameri-
can Lebanese Syrian Associated Charities (ALSAC) and St. Jude
Children’s Research Hospital, National Institute of Allergy and
Infectious Diseases Grant AI35707, Drugs for Neglected
Diseases initiative, and the Sandler Family Supporting Founda-
tion. The TbcatB TOPO TA vector was a kind gift from Zachary
Mackey, and we thank him for all of his additional help during
the T. brucei studies. We also thank Joey Hansell for her
assistance with the TbcatB assays.
Supporting Information Available: Spectroscopic data, LCMS
data, enzyme kinetic data, complete protease assay data for all listed
compounds, and experimental details for modeling studies. This
material is available free of charge via the Internet at http://
pubs.acs.org.
References
(1) Mackey, Z. B.; O’Brien, T. C.; Greenbaum, D. C.; Blank, R. B.;
McKerrow, J. H. J. Biol. Chem. 2004, 279, 48426–48433.
(2) Sleeping Sickness; WHO: Geneva, 2007; http://www.who.int/
mediacentre/factsheets/fs259/en/.
(3) Sleeping Sickness Treatment Schedule; WHO: Geneva, 2007.
whqlibdoc.who.int/hq/2001/WHO_CDS_CSR_EPH_2001.3_fre.pdf.
(4) Brun, R.; Schumacher, R.; Schmid, C.; Kunz, C.; Burri, C. Trop. Med.
Int. Health 2001, 6, 906–914.
(5) Legros, D.; Evans, S.; Maiso, F.; Enyaru, J. C.; Mbulamberi, D. Trans.
R. Soc. Trop. Med. Hyg. 1999, 93, 439–442.
(6) Vickerman, K. Nature 1978, 273, 613–617.
(7) Ashall, F.; Angliker, H.; Shaw, E. Biochem. Biophys. Res. Commun.
1990, 170, 923–929.
(8) Scory, S.; Stierhof, Y. D.; Caffrey, C. R.; Steverding, D. Kinetoplastid
Biol. Dis. 2007, 6, 2.
(9) Nkemgu, N. J.; Grande, R.; Hansell, E.; McKerrow, J. H.; Caffrey,
C. R.; Steverding, D. Int. J. Antimicrob. Agents 2003, 22, 155–159.
(10) Scory, S.; Caffrey, C. R.; Stierhof, Y. D.; Ruppel, A.; Steverding, D.
Exp. Parasitol. 1999, 91, 327–333.
(32) Xing, R.; Addington, A. K.; Mason, R. W Biochem. J. 1998, 332 (Part
2), 499–505.
(33) Deussing, J.; Roth, W.; Saftig, P.; Peters, C.; Ploegh, H. L.;
(11) Troeberg, L.; Morty, R. E.; Pike, R. N.; Lonsdale-Eccles, J. D.; Palmer,
J. T.; McKerrow, J. H.; Coetzer, T. H. Exp. Parasitol. 1999, 91, 349–
355.
(12) Fujii, N.; Mallari, J. P.; Hansell, E. J.; Mackey, Z.; Doyle, P.; Zhou,
Y. M.; Gut, J.; Rosenthal, P. J.; McKerrow, J. H.; Guy, R. K. Bioorg.
Med. Chem. Lett. 2005, 15, 121–123.
Villadangos, J. A. Proc. Natl. Acad. Sci. U.S.A. 1998, 95, 4516–4521.
(34) Nakagawa, T.; Roth, W.; Wong, P.; Nelson, A.; Farr, A.; Deussing,
J.; Villadangos, J. A.; Ploegh, H.; Peters, C.; Rudensky, A. Y. Science
1998, 280, 450–453.
JM070760L