Chemical-biological characterization of a cruzain inhibitor reveals a second target and a mammalian off-target

Article abstract of DOI:10.3762/bjoc.9.3

Inhibition of the Trypanosoma cruzi cysteine protease cruzain has been proposed as a therapeutic approach for the treatment of Chagas' disease. Among the best-studied cruzain inhibitors to date is the vinylsulfone K777 (1), which has proven effective in animal models of Chagas' disease. Recent structure-activity studies aimed at addressing potential liabilities of 1 have now produced analogues such as N-[(2S)-1-[[(E,3S)-1-(benzenesulfonyl)-5- phenylpent-1-en-3-yl]amino]-3-(4-methylphenyl)-1- oxopropan-2-yl]pyridine-4- carboxamide (4), which is trypanocidal at ten-fold lower concentrations than for 1. We now find that the trypanocidal activity of 4 derives primarily from the inhibition of T. cruzi 14-a-demethylase (TcCYP51), a cytochrome P450

Full text of DOI:10.3762/bjoc.9.3

Chemical–biological characterization of a cruzain
inhibitor reveals a second target and
a mammalian off-target
Jonathan W. Choy1,2,3, Clifford Bryant1,2, Claudia M. Calvet4,5,6,
Patricia S. Doyle4,5, Shamila S. Gunatilleke4,5, Siegfried S. F. Leung2,
Kenny K. H. Ang1,2, Steven Chen1,2, Jiri Gut4,5, Juan A. Oses-Prieto2,
Jonathan B. Johnston2, Michelle R. Arkin1,2,4, Alma L. Burlingame2,
Jack Taunton3, Matthew P. Jacobson2, James M. McKerrow4,5,
Larissa M. Podust4,5 and Adam R. Renslo*1,2,4
Full Research Paper
Open Access
Address:
Beilstein J. Org. Chem. 2013, 9, 15–25.
1Small Molecule Discovery Center, University of California San
Francisco, 1700 4th Street, San Francisco, CA, 94158, USA,
2Department of Pharmaceutical Chemistry, University of California
San Francisco, 1700 4th Street, San Francisco, CA, 94158, USA,
3Department of Cellular and Molecular Pharmacology, University of
California San Francisco, 1700 4th Street, San Francisco, CA, 94158,
USA, 4Center for Discovery and Innovation in Parasitic Diseases,
University of California San Francisco, 1700 4th Street, San
Francisco, CA, 94158, USA, 5Department of Pathology, University of
California San Francisco, 1700 4th Street, San Francisco, CA, 94158,
USA and 6Cellular Ultra-Structure Laboratory, Oswaldo Cruz Institute
(IOC), FIOCRUZ, Rio de Janeiro, RJ, Brazil 21040-362
doi:10.3762/bjoc.9.3
Received: 01 October 2012
Accepted: 27 November 2012
Published: 04 January 2013
This article is part of the Thematic Series "Synthetic probes for the study
of biological function".
Guest Editor: J. Aube
© 2013 Choy et al; licensee Beilstein-Institut.
License and terms: see end of document.
Email:
Adam R. Renslo* - adam.renslo@ucsf.edu
* Corresponding author
Keywords:
activity-based probes; Chagas’ disease; cruzain; CYP51;
14-α-demethylase; hybrid drugs; Trypanosoma cruzi
Abstract
Inhibition of the Trypanosoma cruzi cysteine protease cruzain has been proposed as a therapeutic approach for the treatment of
Chagas’ disease. Among the best-studied cruzain inhibitors to date is the vinylsulfone K777 (1), which has proven effective in
animal models of Chagas’ disease. Recent structure–activity studies aimed at addressing potential liabilities of 1 have now
produced analogues such as N-[(2S)-1-[[(E,3S)-1-(benzenesulfonyl)-5-phenylpent-1-en-3-yl]amino]-3-(4-methylphenyl)-1-
oxopropan-2-yl]pyridine-4-carboxamide (4), which is trypanocidal at ten-fold lower concentrations than for 1. We now find that the
trypanocidal activity of 4 derives primarily from the inhibition of T. cruzi 14-α-demethylase (TcCYP51), a cytochrome P450
15

Beilstein J. Org. Chem. 2013, 9, 15–25.
enzyme involved in the biosynthesis of ergosterol in the parasite. Compound 4 also inhibits mammalian CYP isoforms but is
trypanocidal at concentrations below those required to significantly inhibit mammalian CYPs in vitro. A chemical-proteomics ap-
proach employing an activity-based probe derived from 1 was used to identify mammalian cathepsin B as a potentially important
off-target of 1 and 4. Computational docking studies and the evaluation of truncated analogues of 4 reveal structural determinants
for TcCYP51 binding, information that will be useful in further optimization of this new class of inhibitors.
Introduction
The kinetoplastid protozoan Trypanosoma cruzi is the causative led to their preclinical and clinical investigation as potential
agent of Chagas’ disease, a leading cause of heart failure in new treatments for Chagas’ disease [2,15,16].
endemic regions of Latin America [1]. The parasite is trans-
mitted by the reduviid bug and the disease manifests in an Cruzain is a cathepsin-L-like protease of the papain family
initial acute phase, followed by a chronic phase that can last thought to be important for intracellular replication and differ-
decades and typically culminates in heart failure. The existing entiation of the T. cruzi parasite [17]. A variety of small-mole-
treatment for Chagas’ disease involves extended therapy with cule cruzain inhibitors have been described, the majority of
nifurtimox or benznidazole, both of which are associated with which act irreversibly by reaction with the catalytic cysteine in
undesirable side-effects and have limited efficacy against the the enzyme active site [18-27]. One of the earliest cruzain
chronic stage of the disease [2,3]. This situation has spurred the inhibitors identified and perhaps the best studied to date is the
search for more effective and better tolerated therapeutics [4-6]. vinysulfone K777 (1, Figure 1). This irreversible inhibitor has
Among a number of drug targets being investigated are cruzain demonstrated efficacy in animal models of Chagas’ disease
[7-10], the major cysteine protease active in the parasite, and [28,29] and continues to undergo preclinical evaluation leading
T. cruzi CYP51 (TcCYP51), a 14-α-demethylase enzyme of the towards a possible human clinical trial.
cytochrome P450 family required for ergosterol biosynthesis
[11-14]. TcCYP51 is analogous to the fungal enzyme targeted Despite many favorable properties, some aspects of 1 are
by the azole class of antifungals, and the observation that some suboptimal from a drug-development perspective. For example,
of these drugs (e.g., posaconazole) also inhibit TcCYP51 has compound 1 is known to be a mechanism-based (irreversible)
Figure 1: Chemical structures of vinylsulfone-based cruzain inhibitors 1–4, known TcCYP51 inhibitor 5, dihydro controls 6–8, and truncated
analogues 12 and 13. The “P2” and “P3” side chains of 1 are labeled and bind, respectively, in the S2 and S3 sub-sites of the cruzain active site.
16

Beilstein J. Org. Chem. 2013, 9, 15–25.
inhibitor of CYP3A4, an enzyme responsible for the metabo- significantly more trypanocidal than 1 or 2, and unexpectedly
lism of many drugs, including 1 itself [30]. In pharmacokinetic exert their trypanocidal effects primarily by inhibition of
studies, compound 1 exhibits nonlinear exposure with esca- TcCYP51 rather than cruzain.
lating dose and is known to be a substrate of the drug trans-
Results and Discussion
porter P-glycoprotein (P-gp). Finally, as a basic (protonatable)
drug species, 1 could potentially accumulate in acidic lyso- Structure–activity studies
somes, where mammalian cathepsins (potential off-targets of 1) Our exploration of the P3 position in 1 included the evaluation
are located. The issue of lysosomotropism figured prominently of regioisomeric 2-, 3-, and 4-pyridyl congeners in the context
in the discovery and clinical development of cathepsin K inhibi- of various P2 side chains. In many such analogue series, we
tors for osteoporosis. The first such inhibitor to successfully found that regioisomeric analogues possessed similar cruzain
navigate human clinical trials is odanacatib, which was inten- activities in vitro, while the 4-pyridyl examples consistently
tionally designed as a nonbasic drug species to minimize the demonstrated superior trypanocidal activity against cultured
potential for lysosomotropic behavior [31,32].
T. cruzi parasites. For example, 4-pyridyl analogues (e.g., 4)
exhibited sub-micromolar minimal trypanocidal concentration
We sought to address the question of lysosomotropism by pre- values (MTC = 0.6 μM) while the MTC values for 2-pyridyl
paring analogues of 1 in which the basic piperazine substituent (e.g., 3) and 3-pyridyl analogues were typically ≈10 μM, which
at “P3” (which binds the S3 subsite of cruzain) was replaced was similar to the MTC of 1 (Table 1). The MTC represents the
with nonbasic or weakly basic heterocycles. In our initial struc- minimum concentration of test compound required to
ture–activity study [21], we found that analogue 2 (Figure 1), completely clear T. cruzi parasites from J774 macrophage host
bearing a 2-pyridylamide at the P3 position, possessed cells over a 40-day experiment, with the test compound being
trypanocidal activity that was on par with 1 (Table 1). However, administered during the initial 28 days.
none of the nonbasic analogues examined proved superior to 1
and only 2-pyridyl analogues such as 2 and 3 appeared even The enhanced potency of 4-pyridyl analogues as compared to 1
comparable. We therefore turned to more dramatic structural al- or their regioisomeric analogues was not predictable on the
teration and successfully identified and structurally character- basis of in vitro cruzain activity (Table 1). Nor could the trends
ized a new nonpeptidic cruzain inhibitor chemotype [24]. Most be explained as an effect of lysosomotropism, since enhanced
recently, we returned to reinvestigate nonbasic analogues of 1 potency was observed only for the 4-pyridyl analogues and not
and now report that 4-pyridyl analogues such as 4 (Figure 1) are for 2- or 3-pyridyl analogues, which have similar pKa values.
Table 1: In vitro biochemical and cellular activities of test compounds and controls. (n.a. = not active (cruzain IC50 > 50 μM); BNZ = benzindazole;
POSA = posaconazole).
compound
cruzain activity
TcCYP51 activity
T. cruzi growth inhibition
kinact/Ki
(s−1·M−1)
in vitro KD
(nM)
cellular activity
(Y/N, conc.)a
MTCb (μM)
HCSc
EC90 (μM)
1
118,000
120,000
16,000
67,300
—
>2,000
—
N (1.6 μM)
—
8
0.10
—
2
10
3
>2,000
≤5
N (2.0 μM)
Y (0.2 μM)
Y (5.0 μM)
N (2.0 μM)
N (0.1 μM)
Y (0.1 μM)
—
8
1.85
0.10
—
4
0.6
≤10d
>10
>10
0.25
5e
5
≤5
6
n.a.
>2,000
>2,000
≤5
>10
>10
0.11
0.017
>10
3.9
7
n.a.
8
n.a.
9
81,500
n.a.
—
12
620 ± 260
75 ± 26
—
—
>10
1f
13
n.a.
—
BNZ
POSA
—
—
10
7.2
—
≤5
Y (0.1 μM)
0.003
2.7
aCompound affects ergosterol biosynthesis at indicated concentration as determined by GC/MS analysis. bMinimum effective concentration that
clears J774 host cells of parasites at day 40 of the experiment, following 28 days of treatment. cConcentration that reduces parasite load in C2C12
cells by 90% relative to untreated controls. dConcentrations lower than 10 μM were not examined. eExperiment performed in BESM host cell rather
than J774 cells. fRead at day 12 following 7 days treatment.
17

Beilstein J. Org. Chem. 2013, 9, 15–25.
Instead, we considered that additional target(s) may be respon- the ligand and surrounding residues (within 5 Å of ligand) using
sible for the surprising potency of the 4-pyridyl analogues. PRIME [35]. Finally, binding scores were computed by using
Specifically, we were aware that a 4-pyridyl ring comprises the both Glide XP and the MM/GMSA method. Compound 4 was
putative heme-binding moiety in a new class of TcCYP51 predicted to bind in a similar fashion as (R)-5, with the
inhibitors represented by compound 5 (Figure 1). Other struc- 4-pyridyl ring chelating the heme-iron atom and the tolyl ring at
tural similarities of 4 and 5 suggested that compound 4 could P2 contacting many of the same residues (e.g., Try103, Phe110)
conceivably bind TcCYP51.
predicted to interact with the tryptophan ring of (R)-5
(Figure 2A). This same hydrophobic site in TcCYP51 binds the
To test the hypothesis that 4 may also target TcCYP51, we fluoroaryl rings of fluconazole and posaconazole in co-crystal
examined the binding of this compound to TcCYP51 using a structures [14]. The predicted binding mode of the enantiomer
UV–vis spectroscopic binding assay described previously [33]. (S)-5 was described previously [16] and is distinct from that
Indeed, compound 4 bound TcCYP51 with an estimated proposed for 4 and (R)-5.
KD ≤ 5 nM, a value comparable to the binding affinity of the
known TcCYP51 inhibitor 5 [16]. 2-Pyridyl analogue 3 did not Thus, computational docking provides a conceptual picture of
measurably bind TcCYP51 (KD > 2,000 nM, Table 1), whereas how compound 4 – notionally a cruzain inhibitor – might also
the corresponding 3-pyridyl congener (not shown) binds about bind TcCYP51. Interestingly, this is not the first time that potent
100-fold more weakly (KD ≈ 500 nM) than 4. These findings TcCYP51 binding has been discovered in a molecule originally
were thus consistent with our hypothesis that the 4-pyridyl ring intended for a different target. Buckner and Gelb unexpectedly
in 4 is involved in binding TcCYP51. The 2-pyridyl ring system found that the human protein farnesyltransferase (PFT) inhibitor
in 3 is presumably unable to chelate heme in TcCYP51 due to tipifarnib exerts its antitrypanosomal effects through inhibition
steric hindrance from the immediately adjacent amide linkage. of TcCYP51 [36]. Subsequently, these researchers succeeded in
divorcing PFT activity from TcCYP51 inhibition in the tipi-
Computational docking studies
farnib scaffold, producing new lead compounds with
We next employed computational docking and a model derived compelling properties [37-39].
from the crystal structure of TcCYP51 to compare predicted
binding modes of 4 and (R)-5. The two ligands were docked by Inhibition of mammalian CYPs
using the induced-fit docking protocol with Glide XP [34], and A concern with any inhibitor of TcCYP51 is the potential for
the models were further refined by minimizing the energies of cross reactivity with mammalian cytochrome P450 (CYP)
Figure 2: Computational docking models. (A) Predicted binding modes of 4 and (R)-5 bound to TcCYP51. For 4, the ligand, the protein, and the heme
group are shown in green, pink, and grey, respectively. For (R)-5, the ligand, the protein, and the heme group are shown in cyan, purple, and white,
respectively. Heme-iron chelation and hydrophobic binding interactions dominate in the models. (B) Predicted binding models of truncated analogues
12 and 13 to TcCYP51. For 12, the ligand, the protein, and the heme group are shown in orange, light green, and light grey, respectively. For 13, the
ligand, the protein, and the heme group are shown in magenta, dark green, and dark grey, respectively.
18

Beilstein J. Org. Chem. 2013, 9, 15–25.
enzymes, especially those CYPs involved in drug metabolism, 4 involves reversible binding of the parent molecule, while the
like CYP3A4. To assess this risk, we evaluated the inhibitory inhibition of CYP3A4 conferred by 1 is dependent on initial
activities of 4 and 1 across a panel of relevant mammalian CYP conversion to a reactive metabolite. Whatever the explanation,
enzymes (Table 2). Both 4 and 1 inhibited all CYPs in the reversible inhibition of CYP enzymes (as with 4) is clearly
panel, with IC50 values generally in the low micromolar range. preferable to irreversible inhibition from a drug-safety perspec-
Although compound 4 did inhibit CYP3A4, the potency of inhi- tive.
bition (IC50 = 0.8 μM) was less than that exhibited by the anti-
fungal drug ketoconazole (IC50 = 0.086 μM). It should be noted Inhibition of TcCYP51 in live parasites
that the substrate-derived IC50 values from the CYP panel are We next sought to better define the relative importance of
not directly comparable to the KD values for binding to TcCYP51 and cruzain inhibition in the antitrypanosomal effects
TcCYP51. What can be said is that the antitrypanosomal effects of compound 4. Since the 2-pyridyl analogue 3 was found to not
of 4 are realized at concentrations (EC90 = 0.1 μM, bind TcCYP51, this compound could serve as a control for the
MTC = 0.5 μM) well below the in vitro potency of the com- cruzain-derived (and/or other cysteine-protease-derived) effects
pound across the CYP panel (average IC50 ≈ 7 μM). Com- of 4. To provide controls lacking activity against cysteine
pound 4 thus possesses reasonable selectivity with regard to off- proteases, we reduced the vinylsulfone function in analogues 1,
target CYP inhibition, and represents a reasonable starting point 3, and 4 to afford the dihydro analogues 6–8 (Figure 3). As
from which further improvements in selectivity may be under- expected, these analogues were devoid of any detectable
taken.
cruzain inhibitory activity (IC50 > 50 μM, Table 1). Com-
pounds 3, 4, 7 and 8 thus comprised a set of analogues with
complementary activity profiles against the two putative targets:
4 (cruzain and TcCYP51 inhibition), 8 (TcCYP51 inhibition
only), 3 (cruzain inhibition only), and 7 (neither activity).
Table 2: In vitro inhibition of important mammalian CYP enzymes.
IC50 (μM)
compound
1A2
2C9
2C19 2D6
3A4
1
24
22
—
32
5.5
—
7.6
3.4
—
26
2.7
—
1.7
4
0.8
ketoconazole
0.086
Given the similar IC50 values for 1 and 4 against CYP3A4, we
were curious to determine whether 4 is an irreversible inhibitor
of this enzyme, as is the case for 1 [30]. Irreversible inhibition
is typically assessed by measuring the activity of microsomal
CYPs following pre-incubation with or without NADPH.
Consistent with earlier studies [30], compound 1 exhibited irre-
versible inhibition of CYP3A4 as reflected in a significantly
lower IC50 value with NADPH pre-incubation (Table 3). In
Figure 3: Synthesis of the additional control compound 6–8, the
reduced forms of analogues 1, 3, and 4 respectively.
contrast, compound 4 showed behavior typical of reversible Compounds 3, 4, 7, and 8 were evaluated for potency against
inhibition, with no NADPH-dependent shift in the IC50 value. intracellular T. cruzi parasites by using two different assays.
In the case of CYP2C19, both compounds were found to be re- The reported EC90 values (Table 1) represent compound
versible inhibitors. These results suggest that CYP inhibition by concentrations required to reduce parasite numbers in C2C12
host cells by 90% as compared to untreated controls, as deter-
mined by using a high-content imaging-based screening (HCS)
Table 3: Mechanism of inhibition studies for compounds 1 and 4.
approach [33,40]. This high-throughput assay provides a rapid
measure of the initial acute effects of test compound on para-
site viability. The more laborious MTC assay identifies com-
pound concentrations that clear parasites from the host cell, as
determined ca. two weeks after the conclusion of a four-week
course of treatment. This MTC assay therefore provides a
measure of trypanocidal action that cannot be drawn from the
more rapid HCS assay. We judge that MTC values are more
representative of the therapeutic drug levels that would likely be
These data suggest that inhibition of CYP3A4 by 1 is irreversible in
nature.
2C19
3A4
IC50 (μM)
IC50 (μM)
compound
+NADPH −NADPH +NADPH −NADPH
1
4
5.54
8.22
0.0059
1.08
0.300
0.170
0.117
0.046
19

Beilstein J. Org. Chem. 2013, 9, 15–25.
required to produce efficacy in an animal model of Chagas’
disease.
The antitrypanosomal effects of compounds 3, 4, 7, and 8 were
in general agreement with their in vitro activities against the
two putative targets (Table 1). Analogue 7, devoid of either
activity in vitro, showed no effects on T. cruzi parasites in either
the HCS or MTC assay. Analogue 3, possessing primarily
cysteine-protease-derived effects, was effective in both assays
and equipotent to 1 in the MTC assay. Putatively dual-targeted
analogue 4 was about 10-fold more potent than 1 in the MTC
assay and equipotent by HCS. Most unexpectedly, we found
that compound 8, which lacks any cruzain-derived effects of 4,
was equipotent to 4 by HCS and 2–4 times more potent than 4
in the MTC assay.
The in vitro and cell-based activities of 4 and 8 suggest
TcCYP51 as a relevant target of these compounds. To assess
inhibition of TcCYP51 in live parasites we analyzed the sterol
composition of intracellular T. cruzi parasites treated with test
compounds 3–8, 1, or posaconazole as a positive control. The
analysis was performed by employing GC/MS as reported
previously for compound 5 [33]. The GC/MS trace for unin-
fected host cells establishes that the additional peaks observed
in infected cells are of T. cruzi origin (peaks labeled a-i,
Figure 4). Treatment with the known TcCYP51 inhibitor
posaconazole produces an increase in the relative abundance of
TcCYP51 substrates lanosterol (f) and eburicol (h) and accord-
ingly, a reduction in the abundance of downstream sterols such
as fecosterol (e) and cholesta-7,24-dien-3β-ol (a), among others.
Treatment with 1 had little effect on sterol composition as
expected, whereas treatment with compound 4 or 8 produced
effects very similar to those observed in posaconazole treated
parasites (Figure 4 and Supporting Information File 1). The
other test compounds evaluated (3, 6, 7) produced no signifi-
cant change in lipid composition, as expected since these com-
pounds do not inhibit TcCYP51 in vitro (Supporting Informa-
tion File 1). Test compounds were necessarily studied at
concentrations below their MTC, so as to retain a population of
viable parasites for analysis.
Figure 4: GC/MS analysis of lipid extracts from T. cruzi parasites
treated with test compounds. DMSO and K777 (1) were used as nega-
tive controls; posaconazole served as a positive control. The analysis
of 4 was performed concurrently with other CYP51 inhibitors described
recently [33] and, thus, the spectra for the controls shown above are
reproduced from the earlier report. Spectra of lipid extracts from para-
sites treated with 3, 6, 7, and 8 are provided in Supporting Information
File 1. Uninfected host cell panel (top) demonstrates that chromato-
graphic peaks labeled a to i in subsequent panels are of T. cruzi origin.
These peaks are assigned as a - cholesta-7,24-dien-3β-ol, [M]•+ = m/z
454; b - cholesta-8,24-dien-3β-ol (zymosterol), [M]•+ = m/z 470; c -
24-methyl-7-en-cholesta-en-3β-ol, [M]•+ = m/z 472; d - ergosta-7,24-
diene-3β-ol (episterol), [M]•+ = m/z 470; e - ergosta-8,24-diene-3β-ol
(fecosterol), [M]•+ = m/z 470; f - lanosterol, [M]•+ = m/z 498; g -
4-methylepisterol, [M]•+ = m/z 484; h - eburicol, [M]•+ = m/z 512; i -
24-ethyl-7,24(24’)-encholestadien-3β-ol, [M]•+ = m/z 484.
An activity-based probe reveals an off-target
of 1 and 4
We next sought to evaluate the cysteine-protease-related effects
of the various test compounds in T. cruzi parasites. To do this,
we designed and synthesized the “clickable” activity-based
probe 9 in which a propargyl group (replacing methyl in 1)
serves as a chemical handle for conjugation to TAMRA- or
biotin-containing reagents (10 and 11, respectively, Figure 5).
Probe 9 was found to be equipotent to 1 against cruzain in vitro
and retained similar effects against T. cruzi parasites in both the
20

Beilstein J. Org. Chem. 2013, 9, 15–25.
HCS and MTC assays. Thus, the cysteine protease target(s) of 9 pretreatment with 10 μM of compound 1, 3, or 4 successfully
in parasite and host cell can reasonably be assumed to be the blocked labeling of the ≈35 kDa band by probe 9, thus indi-
same as for 1 and close analogues such as 4.
cating that these compounds also react with this target
(Figure 6). As expected, the nonelectrophilic dihydro forms of 1
and 4 (i.e., compounds 6 and 8) did not compete for labeling by
9. Taken together, these results strongly suggest that com-
pounds 1, 3 and 4 react irreversibly with the ≈35 kDa protein in
a process involving the electrophilic vinylsulfone moiety.
Figure 5: Chemical structures of compound 9, a “clickable” activity-
based probe based on 1 and complementary azide-containing
reagents 10 and 11.
Independently, another group recently reported the synthesis of
9 and its use to identify putative targets of 1 in the related para-
site Trypanosoma brucei [41]. Our efforts to similarly identify
targets of 1 in T. cruzi were complicated by the presence of a
host-cell protein that was apparently a major target of 9. In a
typical experiment, intracellular T. cruzi amastigotes were
treated with 9 for 1 hour, followed by cell lysis, “click” reac-
tion with TAMRA azide 10, and separation/visualization by
SDS-PAGE. Regardless of the host cell employed (J774
macrophage, or C2C12), only one prominently labeled band at
≈35 kDa was observed in these experiments. This band was
attributed to a host-cell protein as it appeared also in analogous
experiments employing uninfected cells. In fact, we could not
conclusively identify any unique bands of parasitic origin in our
experiments, although such bands might well have escaped
Figure 6: Competitive labeling of host cell (C2C12) proteins. Intact
cells were labeled with probe 9 following a competitive pre-incubation
step with compounds 1, 4, 3, 8, or 6. After cell lysis, protein adducts of
9 were conjugated to rhodamide-based dye 10. The gel image at the
top shows rhodamine fluorescence. The gel image at the bottom is of
the coomassie stained gel. A successfully competed band is observed
at ≈35 kDa, and this was subsequently identified as cathepsin B.
detection due to lower abundance and labeling below the limit Chemical proteomics
of fluorescence detection. We next applied mass spectrometric analysis to identify the
≈35 kDa band, which was an apparent target of the electro-
The discovery of a potential mammalian off-target of probe 9 philic inhibitors described above. To enrich for this protein,
(and presumably also of 1) was of considerable interest, so we C2C12 cells were labeled with 9 as before and then reacted with
explored this finding further. To determine if this protein was the biotin azide reagent 11, followed by biotin capture onto
also a target of 1 and 4, we conducted competition experiments streptavidin beads. A base-cleavable ester function was intro-
in C2C12 cells. Hence, pre-incubation of cells with competitor duced in the linker of 11, and this allowed enriched proteins to
compound at either 1 μM or 10 μM for one hour was followed be released from beads by treatment with sodium hydroxide.
by treatment for one hour with 9, followed by cell lysis, conju- The liberated proteins were separated by SDS-PAGE, and the
gation to 10, separation (SDS-PAGE), and detection by relevant band at ≈35 kDa extracted from the gel. An in-gel
rhodamine fluorescence as before. In these experiments, trypsin digest [42] was followed by UPLC separation of the
21

Beilstein J. Org. Chem. 2013, 9, 15–25.
Defining a new lead scaffold for TcCYP51
tryptic peptides and MS/MS analysis using a hybrid linear ion-
trap-Orbitrap mass spectrometer. Tandem mass spectra acquired inhibition
were searched against the UniProtKb database employing The similar cellular potencies of 4 and its reduced form 8
ProteinProspector; four MS/MS spectra corresponding to the suggest that cruzain inhibition plays a relatively minor role in
same peptide sequence were identified (Figure 7). This the trypanocidal action of 4. To a first approximation, the
sequence was found to correspond to the tryptic peptide span- cruzain- and TcCYP51-derived effects of 4 should be similar to
ning resides S264-R281 from mouse cathepsin B (uniprot those of its close analogues 3 (MTC ≈ 10 μM) and 8 (MTC ≈
P10605). Significantly, this peptide was not found in analogous 0.25 μM), respectively. Unless the effects of inhibiting both
experiments where pre-incubation with 1 or 4 (at 10 μM) targets are synergistic, which is not supported by the data, there
preceded labeling with 9, nor in an experiment in which 9 was would appear to be little benefit gained by combining a rela-
not added. Thus, cathepsin B is very likely the host-cell protein tively weak cruzain-derived effect with a much more potent
target of compounds 1 and 4 identified in the competition insult conferred by TcCYP51 inhibition. Moreover, it now
experiments with compound 9.
seems likely that electrophilic compounds such as 1 and 4 may
be partially consumed in nonproductive reactions with host-cell
The identification of cathepsin B as a relevant cellular off-target proteases (e.g., cathepsin B) and/or other cytosolic nucleo-
of 1 and 4 is potentially significant. On the one hand, the HCS philes (e.g., glutathione). This possibility is supported by our
EC90 values of 1 and 4 are at least 10-fold lower than the competitive labeling experiments (Figure 6) and by in vitro
concentrations of these compounds used in the competition studies employing physiological concentrations of glutathione
experiments. Thus, one might expect to achieve effects on para- (Supporting Information File 1). With regard to the inhibitor
site viability before significant inhibition of cathepsin B is chemotypes covered here, there appears to be little rationale for
conferred. On the other hand, the MTC for 1 (8 μM) lies targeting both cruzain and TcCYP51. On the other hand, the
squarely in the range at which the compound effectively surprising potency of analogue 8 does suggest this as a new lead
competes for cathepsin B labeling by 9. Thus, if micromolar scaffold for the development of novel TcCYP51 inhibitors.
concentrations of 1 are indeed required to achieve a therapeutic
effect in animals, one might well be concerned about the effects We next sought to define the minimal pharmacophore within 8
on host cathepsin B. Thus, the experiments with 9 identified a required for inhibition of TcCYP51 in vitro and anti-
potential off-target while also providing an experimental means trypanosomal effects in whole cells. We therefore synthesized
for testing the effects of new analogues on this off-target in a truncated analogues of 8, such as 12 and 13 (Figure 1). These
relevant, cellular context.
compounds retain the 4-pyridyl ring and neighboring tolyl side
Figure 7: MS/MS spectrum of the tryptic peptide S264-R281 from mouse cathepsin B, identified in pull-down experiments employing compound 9 in
C2C12 cells. Observed sequence ions are labeled (m = oxidized methionine).
22

Beilstein J. Org. Chem. 2013, 9, 15–25.
chain of 8 while dispensing with those substituents further the use of cell-based counter assays that can detect action at
removed from the putative heme-binding moiety. Interestingly, specific targets or signaling pathways of interest. Activity-based
the truncated analogues 12 and 13 bound TcCYP51 signifi- probes can serve as useful tools to verify on-target action during
cantly more weakly than 4 or 8 (Table 1), suggesting that side the course of chemical optimization campaigns.
chains relatively far removed from the 4-pyridyl ring nonethe-
less play an important role in binding.
Supporting Information
Computational docking of 12 and 13 provided some insight into
The Supporting Information features a table with
the observed binding trends. Analogue 13 adopts a docking
experimentally determined and computationally predicted
pose very similar to 4 with respect to the 4-pyridyl and tolyl
binding affinities, additional GC/MS spectra from
ring systems. When compared to the poses for 4 or 13, the tolyl
lipid-analysis studies, time courses for reaction of
ring in 12 projects much less deeply into the aromatic pocket
compounds 1 and 6 with glutathione in vitro, and synthetic
formed by Phe110 and Tyr103 (Figure 2). Neither 12 nor 13
schemes for analogues 4, 9, 11, 12, and 13, as well as
form interactions with more distal residues (e.g., Leu208,
experimental procedures.
Pro210) that are predicted to form productive contacts with 4.
Supporting Information File 1
Thus, a larger number of hydrophobic contacts and better orien-
tation of some side chains may explain the binding trends for 4,
12, and 13. Interestingly, the rank-order binding affinities of 4,
12, and 13 were correctly predicted by the MM-GBSA method
applied to the binding models of these compounds (Supporting
Information File 1). This suggests that such models could serve
Figures, schemes, and experimental procedures.
[http://www.beilstein-journals.org/bjoc/content/
supplementary/1860-5397-9-3-S1.pdf]
to aid in the design of new TcCYP51 inhibitors derived from Acknowledgements
this scaffold. The authors acknowledge research support from the Sandler
Foundation (to ARR), NIH RO1 AI095437 (to LMP), Conselho
The antitrypanosomal activities of analogues 12 and 13 could Nacional de Desenvolvimento Cientifico e Tecnologico and
be correlated with their in vitro binding affinities for TcCYP51 FIOCRUZ (to CMC). JBJ was supported by NIH RO1 AI74824
(Table 1). Hence, analogue 13 (KD = 75 nM) shows reduced (to Prof. Ortiz de Montellano, UCSF). Mass spectrometry was
antitrypanosomal activity when compared to 8 (KD ≈ 5 nM). provided by the Bio-Organic Biomedical Mass Spectrometry
Still weaker-binding analogue 12 (KD = 620 nM) exhibited no Resource at UCSF (A.L. Burlingame, Director) and supported
antitrypanosomal effect at the highest concentration examined by the Biomedical Technology Research Centers program of the
(10 μM). Thus, compound 13 can be considered to represent a NIH National Institute of General Medical Sciences, NIH
“minimal pharmacophore” that retains reasonable affinity for NIGMS 8P41GM103481. MPJ is a consultant to Schrodinger,
TcCYP51 in vitro while also conferring an effect on T. cruzi LLC.
parasites in culture. Future work will focus on further refining
References
1. de Souza, W. Microbes Infect. 2007, 9, 544–545.
the in vitro and cellular potency of this scaffold, with com-
pound 13 serving as a chemical departure point.
doi:10.1016/j.micinf.2006.12.014
2. Urbina, J. A. Mem. Inst. Oswaldo Cruz 2009, 104 (Suppl. 1), 311–318.
Conclusion
Structure–activity studies are often conducted with the under-
doi:10.1590/S0074-02762009000900041
3. Castro, J. A.; de Mecca, M. M.; Bartel, L. C. Hum. Exp. Toxicol. 2006,
lying assumption that molecular mechanisms are the same
25, 471–479. doi:10.1191/0960327106het653oa
within congeneric analogue series. This assumption is rein-
forced when activity in biochemical assays can be correlated
with cell-based activity. Of course perfect correlation is rarely
observed, even when a series is in fact “on-target”. Especially
perilous is the construction of mechanistic hypotheses based
solely on the correlation of in vitro biochemical assay data with
gross phenotypic endpoints such as parasite growth inhibition or
cell death. As demonstrated here, even seemingly trivial struc-
tural changes within a congeneric SAR series can produce
analogues with disparate molecular mechanisms of action.
Advisable approaches to deal with these uncertainties include
4. Renslo, A. R.; McKerrow, J. H. Nat. Chem. Biol. 2006, 2, 701–710.
doi:10.1038/nchembio837
5. McKerrow, J. H.; Doyle, P. S.; Engel, J. C.; Podust, L. M.;
Robertson, S. A.; Ferreira, R.; Saxton, T.; Arkin, M.; Kerr, I. D.;
Brinen, L. S.; Craik, C. S. Mem. Inst. Oswaldo Cruz 2009, 104
(Suppl. 1), 263–269. doi:10.1590/S0074-02762009000900034
6. Urbina, J. A. J. Mol. Med. (Heidelberg, Ger.) 1999, 77, 332–338.
doi:10.1007/s001090050359
7. McKerrow, J. H.; Engel, J. C.; Caffrey, C. R. Bioorg. Med. Chem. 1999,
7, 639–644. doi:10.1016/S0968-0896(99)00008-5
8. Eakin, A. E.; McGrath, M. E.; McKerrow, J. H.; Fletterick, R. J.;
Craik, C. S. J. Biol. Chem. 1993, 268, 6115–6118.
23

Beilstein J. Org. Chem. 2013, 9, 15–25.
9. McKerrow, J. H.; McGrath, M. E.; Engel, J. C. Parasitol. Today 1995,
11, 279–282. doi:10.1016/0169-4758(95)80039-5
27.Beaulieu, C.; Isabel, E.; Fortier, A.; Massé, F.; Mellon, C.; Méthot, N.;
Ndao, M.; Nicoll-Griffith, D.; Lee, D.; Park, H.; Black, W. C.
Bioorg. Med. Chem. Lett. 2010, 20, 7444–7449.
10.McGrath, M. E.; Eakin, A. E.; Engel, J. C.; McKerrow, J. H.;
Craik, C. S.; Fletterick, R. J. J. Mol. Biol. 1995, 247, 251–259.
doi:10.1006/jmbi.1994.0137
doi:10.1016/j.bmcl.2010.10.015
28.Engel, J. C.; Doyle, P. S.; Hsieh, I.; McKerrow, J. H. J. Exp. Med. 1998,
188, 725–734. doi:10.1084/jem.188.4.725
11.Buckner, F. S.; Joubert, B. M.; Boyle, S. M.; Eastman, R. T.;
Verlinde, C. L. M. J.; Matsuda, S. P. T. Mol. Biochem. Parasitol. 2003,
132, 75–81. doi:10.1016/j.molbiopara.2003.07.004
12.Hankins, E. G.; Gillespie, J. R.; Aikenhead, K.; Buckner, F. S.
Mol. Biochem. Parasitol. 2005, 144, 68–75.
29.Barr, S. C.; Warner, K. L.; Kornreic, B. G.; Piscitelli, J.; Wolfe, A.;
Benet, L.; McKerrow, J. H. Antimicrob. Agents Chemother. 2005, 49,
5160–5161. doi:10.1128/AAC.49.12.5160-5161.2005
30.Jacobsen, W.; Christians, U.; Benet, L. Z. Drug Metab. Dispos. 2000,
28, 1343–1351.
doi:10.1016/j.molbiopara.2005.08.002
13.Podust, L. M.; von Kries, J. P.; Eddine, A. N.; Kim, Y.;
Yermalitskaya, L. V.; Kuehne, R.; Ouellet, H.; Warrier, T.;
Alteköster, M.; Lee, J.-S.; Rademann, J.; Oschkinat, H.;
Kaufmann, S. H. E.; Waterman, M. R. Antimicrob. Agents Chemother.
2007, 51, 3915–3923. doi:10.1128/AAC.00311-07
31.Falgueyret, J.-P.; Desmarais, S.; Oballa, R.; Black, W. C.;
Cromlish, W.; Khougaz, K.; Lamontagne, S.; Massé, F.; Riendeau, D.;
Toulmond, S.; Percival, M. D. J. Med. Chem. 2005, 48, 7535–7543.
doi:10.1021/jm0504961
32.Robichaud, J.; Black, W. C.; Thérien, M.; Paquet, J.; Oballa, R. M.;
Bayly, C. I.; McKay, D. J.; Wang, Q.; Isabel, E.; Léger, S.; Mellon, C.;
Kimmel, D. B.; Wesolowski, G.; Percival, M. D.; Massé, F.;
Desmarais, S.; Falgueyret, J.-P.; Crane, S. N. J. Med. Chem. 2008, 51,
6410–6420. doi:10.1021/jm800610j
14.Chen, C.-K.; Leung, S. S. F.; Guilbert, C.; Jacobson, M. P.;
McKerrow, J. H.; Podust, L. M. PLoS Negl. Trop. Dis. 2010, 4, e651.
doi:10.1371/journal.pntd.0000651
15.Urbina, J. A. Acta Trop. 2010, 115, 55–68.
doi:10.1016/j.actatropica.2009.10.023
33.Gunatilleke, S. S.; Calvet, C. M.; Johnston, J. B.; Chen, C. K.;
Erenburg, G.; Gut, J.; Engel, J. C.; Ang, K. K.; Mulvaney, J.; Chen, S.;
Arkin, M. R.; McKerrow, J. H.; Podust, L. M. PLoS Negl. Trop. Dis.
2012, 6, e1736.
16.Doyle, P. S.; Chen, C.-K.; Johnston, J. B.; Hopkins, S. D.;
Leung, S. S. F.; Jacobson, M. P.; Engel, J. C.; McKerrow, J. H.;
Podust, L. M. Antimicrob. Agents Chemother. 2010, 54, 2480–2488.
doi:10.1128/AAC.00281-10
34.Sherman, W.; Day, T.; Jacobson, M. P.; Friesner, R. A.; Farid, R.
J. Med. Chem. 2006, 49, 534–553. doi:10.1021/jm050540c
35.Zhu, K.; Shirts, M. R.; Friesner, R. A.; Jacobson, M. P.
J. Chem. Theory Comput. 2007, 3, 640–648. doi:10.1021/ct600129f
36.Hucke, O.; Gelb, M. H.; Verlinde, C. L. M. J.; Buckner, F. S.
J. Med. Chem. 2005, 48, 5415–5418. doi:10.1021/jm050441z
37.Kraus, J. M.; Verlinde, C. L. M. J.; Karimi, M.; Lepesheva, G. I.;
Gelb, M. H.; Buckner, F. S. J. Med. Chem. 2009, 52, 1639–1647.
doi:10.1021/jm801313t
17.Harth, G.; Andrews, N.; Mills, A. A.; Engel, J. C.; Smith, R.;
McKerrow, J. H. Mol. Biochem. Parasitol. 1993, 58, 17–24.
doi:10.1016/0166-6851(93)90086-D
18.Roush, W. R.; Cheng, J.; Knapp-Reed, B.; Alvarez-Hernandez, A.;
McKerrow, J. H.; Hansell, E.; Engel, J. C. Bioorg. Med. Chem. Lett.
2001, 11, 2759–2762. doi:10.1016/S0960-894X(01)00566-2
19.Huang, L.; Lee, A.; Ellman, J. A. J. Med. Chem. 2002, 45, 676–684.
doi:10.1021/jm010333m
20.Greenbaum, D. C.; Mackey, Z.; Hansell, E.; Doyle, P.; Gut, J.;
Caffrey, C. R.; Lehrman, J.; Rosenthal, P. J.; McKerrow, J. H.;
Chibale, K. J. Med. Chem. 2004, 47, 3212–3219.
38.Chennamaneni, N. K.; Arif, J.; Buckner, F. S.; Gelb, M. H.
Bioorg. Med. Chem. Lett. 2009, 19, 6582–6584.
doi:10.1016/j.bmcl.2009.10.029
doi:10.1021/jm030549j
39.Kraus, J. M.; Tatipaka, H. B.; McGuffin, S. A.; Chennamaneni, N. K.;
Karimi, M.; Arif, J.; Verlinde, C. L. M. J.; Buckner, F. S.; Gelb, M. H.
J. Med. Chem. 2010, 53, 3887–3898. doi:10.1021/jm9013136
40.Engel, J. C.; Ang, K. K. H.; Chen, S.; Arkin, M. R.; McKerrow, J. H.;
Doyle, P. S. Antimicrob. Agents Chemother. 2010, 54, 3326–3334.
doi:10.1128/AAC.01777-09
21.Jaishankar, P.; Hansell, E.; Zhao, D.-M.; Doyle, P. S.; McKerrow, J. H.;
Renslo, A. R. Bioorg. Med. Chem. Lett. 2008, 18, 624–628.
doi:10.1016/j.bmcl.2007.11.070
22.Brak, K.; Doyle, P. S.; McKerrow, J. H.; Ellman, J. A.
J. Am. Chem. Soc. 2008, 130, 6404–6410. doi:10.1021/ja710254m
23.Chen, Y. T.; Lira, R.; Hansell, E.; McKerrow, J. H.; Roush, W. R.
Bioorg. Med. Chem. Lett. 2008, 18, 5860–5863.
41.Yang, P.-Y.; Wang, M.; He, C. Y.; Yao, S. Q. Chem. Commun. 2012,
48, 835–837. doi:10.1039/c1cc16178d
doi:10.1016/j.bmcl.2008.06.012
42.Rosenfeld, J.; Capdevielle, J.; Guillemot, J. C.; Ferrara, P.
Anal. Biochem. 1992, 203, 173–179.
24.Bryant, C.; Kerr, I. D.; Debnath, M.; Ang, K. K. H.; Ratnam, J.;
Ferreira, R. S.; Jaishankar, P.; Zhao, D.; Arkin, M. R.; McKerrow, J. H.;
Brinen, L. S.; Renslo, A. R. Bioorg. Med. Chem. Lett. 2009, 19,
6218–6221. doi:10.1016/j.bmcl.2009.08.098
doi:10.1016/0003-2697(92)90061-B
25.Brak, K.; Kerr, I. D.; Barrett, K. T.; Fuchi, N.; Debnath, M.; Ang, K.;
Engel, J. C.; McKerrow, J. H.; Doyle, P. S.; Brinen, L. S.; Ellman, J. A.
J. Med. Chem. 2010, 53, 1763–1773. doi:10.1021/jm901633v
26.Mott, B. T.; Ferreira, R. S.; Simeonov, A.; Jadhav, A.; Ang, K. K.-H.;
Leister, W.; Shen, M.; Silveira, J. T.; Doyle, P. S.; Arkin, M. R.;
McKerrow, J. H.; Inglese, J.; Austin, C. P.; Thomas, C. J.;
Shoichet, B. K.; Maloney, D. J. J. Med. Chem. 2010, 53, 52–60.
doi:10.1021/jm901069a
24

Beilstein J. Org. Chem. 2013, 9, 15–25.
License and Terms
This is an Open Access article under the terms of the
Creative Commons Attribution License
(http://creativecommons.org/licenses/by/2.0), which
permits unrestricted use, distribution, and reproduction in
any medium, provided the original work is properly cited.
The license is subject to the Beilstein Journal of Organic
Chemistry terms and conditions:
(http://www.beilstein-journals.org/bjoc)
The definitive version of this article is the electronic one
which can be found at:
doi:10.3762/bjoc.9.3
25