Binding to large enzyme pockets: Small-molecule inhibitors of trypanothione reductase

Article abstract of DOI:10.1002/cmdc.201402032

The causative agents of the parasitic disease human African trypanosomiasis belong to the family of trypanosomatids. These parasitic protozoa exhibit a unique thiol redox metabolism that is based on the flavoenzyme trypanothione reductase (TR). TR was ide

Full text of DOI:10.1002/cmdc.201402032

CHEMMEDCHEM
FULL PAPERS
DOI: 10.1002/cmdc.201402032
Binding to Large Enzyme Pockets: Small-Molecule
Inhibitors of Trypanothione Reductase
Elke Persch,[a] Steve Bryson,[b] Nickolay K. Todoroff,[c] Christian Eberle,[a] Jonas Thelemann,[a]
Natalie Dirdjaja,[d] Marcel Kaiser,[e] Maria Weber,[a] Hassan Derbani,[d] Reto Brun,*[e]
Gisbert Schneider,*[c] Emil F. Pai,*[b] R. Luise Krauth-Siegel,*[d] and FranÅois Diederich*[a]
The causative agents of the parasitic disease human African
trypanosomiasis belong to the family of trypanosomatids.
These parasitic protozoa exhibit a unique thiol redox metabo-
lism that is based on the flavoenzyme trypanothione reductase
(TR). TR was identified as a potential drug target and features
a large active site that allows a multitude of possible ligand
orientations, which renders rational structure-based inhibitor
design highly challenging. Herein we describe the synthesis,
binding properties, and kinetic analysis of a new series of
small-molecule inhibitors of TR. The conjunction of biological
activities, mutation studies, and virtual ligand docking simula-
tions led to the prediction of a binding mode that was con-
firmed by crystal structure analysis. The crystal structures re-
vealed that the ligands bind to the hydrophobic wall of the
so-called “mepacrine binding site”. The binding conformation
and potency of the inhibitors varied for TR from Trypanosoma
brucei and T. cruzi.
Introduction
Despite the fact that the number of newly reported cases of
human African trypanosomiasis (HAT, sleeping sickness) de-
creased from above 37000 in 1998 to around 7000 in 2012,
this parasitic disease remains a serious public health prob-
lem.[1–3] History has shown that even less than 5000 new cases,
as in the 1960s, can rapidly lead to a new epidemic if control
of trypanosomiasis is decreased in priority.[4] In addition to
broad active case detection campaigns and research on new
diagnostics and vector control activities, the development of
new drug candidates plays a central role.[3,5] Current drugs for
the treatment of HAT show severe side effects, require a cost-
intensive administration, and suffer from the occurrence of re-
sistance.[6]
HAT is transmitted by the tsetse fly and caused by two sub-
species of Trypanosoma brucei (T. brucei), T. brucei rhodesiense
and T. brucei gambiense. These parasites belong to the family
of trypanosomatids, as do the agents of leishmaniasis (Leish-
mania spp.) and Chagas’ disease (T. cruzi).[7] Trypanosomatids
stand out, due to their unusual redox metabolism based on
the flavoenzyme trypanothione reductase (TR), which catalyzes
the reduction of trypanothione disulfide (TS2; 1) (Figure 1) to
the dithiol trypanothione. TR plays a central role in the defense
of the parasites against oxidative stress. The enzyme is essen-
tial for trypanosomatids, and the development of potent inhib-
itors could lead to new drugs for three neglected diseases:
HAT, Chagas’ disease, and the various forms of leishmaniasis.[8,9]
The mammalian counterparts of trypanothione and TR are glu-
tathione and glutathione reductase (GR), respectively.[10,11]
Despite an overall sequence identity of 40% between
human GR (hGR) and TR, both enzymes show mutually exclu-
sive selectivity for their disulfide substrates, and the design of
selective inhibitors for one of the enzymes has already been
achieved.[12–14] A key feature in the discrimination between sub-
strates and putative ligands is the charge distribution in the
active sites.[15] A crucial difference between the disulfide sub-
strate binding sites of hGR and TR is the replacement of Ala34,
Arg37, and Asn117 in hGR by Glu18, Trp21, and Met113 in TR.
Importantly, Glu18 introduces a net negative charge into the
[a] E. Persch, Dr. C. Eberle, J. Thelemann, M. Weber, Prof. Dr. F. Diederich
Laboratorium fꢀr Organische Chemie, ETH Zꢀrich
Vladimir-Prelog-Weg 3, 8093 Zurich (Switzerland)
E-mail: diederich@org.chem.ethz.ch
[b] Dr. S. Bryson, Prof. Dr. E. F. Pai
Departments of Biochemistry, Medical Biophysics & Molecular Genetics
University of Toronto, Medical Sciences Building, #5358
1 King’s College Circle, Toronto, ON, M5S 1A8 (Canada)
and
The Campbell Family Institute for Cancer Research
University Health Network, 101 College Street, Toronto, ON, M5G 1L7
(Canada)
E-mail: pai@hera.med.utoronto.ca
[c] Dr. N. K. Todoroff, Prof. Dr. G. Schneider
Institut fꢀr Pharmazeutische Wissenschaften, ETH Zꢀrich
Vladimir-Prelog-Weg 4, 8093 Zurich (Switzerland)
E-mail: gisbert.schneider@pharma.ethz.ch
[d] N. Dirdjaja, H. Derbani, Prof. Dr. R. L. Krauth-Siegel
Biochemie-Zentrum Heidelberg (BZH), Universitꢁt Heidelberg
Im Neuenheimer Feld 328, 69120 Heidelberg (Germany)
E-mail: luise.krauth-siegel@bzh.uni-heidelberg.de
[e] Dr. M. Kaiser, Prof. Dr. R. Brun
Swiss Tropical & Public Health Institute
Socinstrasse 57, 4002 Basel (Switzerland)
and
University of Basel, Petersplatz 1, 4003 Basel (Switzerland)
E-mail: Reto.Brun@unibas.ch
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cmdc.201402032.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 0000, 00, 1 – 13
&
1
&
ÞÞ
These are not the final page numbers!

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
Figure 2. Structure of 1-[1-(benzo[b]thien-2-yl)cyclohexyl]piperidine (BTCP; 2)
and inhibitor 3, which comprises the pattern of the diaryl sulfide platform
and BTCP.
larger (~817 ꢁ3). Indeed, most currently known potent TR in-
hibitors are of greater molecular size or even form protein
complexes with a 1:2 (TR/ligand) stoichiometry.[18,24–28]
Herein we describe the design and synthesis of novel BTCP
scaffold-based TR inhibitors, their biological activity, and analy-
sis of their binding modes to TR by mutation studies, molecu-
lar docking, and X-ray crystallography. As large binding sites
allow multiple possible ligand orientations, careful elucidation
of the ligand binding mode is crucial for further rational struc-
ture-based drug design. In addition, inhibitor efficacy and the
molecular binding mode toward TRs from different parasite
species was investigated.
Results and Discussion
Figure 1. a) Crystal structure of T. cruzi TR in complex with TS2 and flavin ad-
enine dinucleotide (FAD) (resolution: 2.4 ꢁ; PDB ID: 1BZL[16]). Only the isoal-
loxazine ring of FAD is shown; Cenzyme =grey, O=red, N=blue, S=yellow,
Ligand design
C
TS2 =yellow, CFAD =blue. b) Structure of trypanothione disulfide (TS2).
Based on the binding mode analysis of BTCP-analogous inhibi-
tors by molecular modeling using the computer program
MOLOC,[29] we initially proposed the Z-site of TR (Figure 1) as
the binding site.[18] This hydrophobic region was defined in
1991 by El-Waer et al. and includes the side-chains of Phe396’,
Pro398’, and Leu399’ (Figure 1).[30–32] Several published molec-
ular modeling and docking approaches suggest the Z-site as
the relevant binding pocket and discuss the importance of an
electrostatic interaction between the inhibitor and residues
Glu466’ and Glu467’.[17,33–36]
The structure-based design of the new BTCP analogue inhib-
itors was similarly built on the binding mode predefined by
the salt bridge between the protonated tertiary amine and
Glu467’, and the inhibitor’s cyclohexyl ring, which establishes
the orientation of the aromatic substituent toward the Z-site.
The derivatives shown in Table 1 were synthesized in order to
investigate proper filling of the Z-site in terms of polarity and
size (compounds 4–7) and to optimize the postulated
amide···p stacking with the flat peptide walls of the pocket
(compounds 8–13).[37] Compounds 8a and 10–13a were sub-
stituted with five-membered heteroaromatic linkers (Ar1) fea-
turing different electronic properties and bend angles between
the two scaffold exit vectors.[38,39] All compounds were de-
signed with a polar surface area (PSA) of <80 ꢁ2 (see Support-
ing Information (SI)) to enhance their probability for crossing
active site of TR, thereby promoting binding of the positively
charged substrate TS2 (Figure 1).[15,16]
Previously, we confirmed a favorable energetic contribution
to binding from attractive hydrogen bonding or ion pairing in-
teractions between diaryl sulfide-based inhibitors and the neg-
atively charged Glu18 in the active site of TR.[17] We further
conjugated diaryl sulfides to analogues of 1-[1-(benzo[b]thien-
2-yl)cyclohexyl]piperidine (BTCP; 2),[18] which was discovered as
a remarkably low-molecular-weight TR inhibitor by Fairlamb
and co-workers (Figure 2).[19,20] This conjugation led to ligands
with Kic (competitive inhibitory constant) values against T. cruzi
TR within the sub-micromolar range (3: Kic =510ꢀ
100 nm).[18,21] On the other hand, the molecular mass of these
conjugates became rather large (3: 673 gmol )
ꢁ1 [22] and the syn-
theses undesirably long. The BTCP scaffold alone features sev-
eral distinct advantages: This remarkably small ligand (Mr:
301 gmolꢁ1; volume ~366 ꢁ3) binds to the large active site of
TR (volume ~22ꢂ20ꢂ28=12320 ꢁ3) with a Kic value in the
low micromolar range (Kic =1 mm).[19,20] It has been shown in
other work to cross the blood–brain barrier, which could be of
importance in combating the second stage of HAT.[23] In com-
parison, the volume of the natural substrate (TS2) is twofold
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 0000, 00, 1 – 13
&
2
&
ÞÞ
These are not the final page numbers!

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
Table 1. Inhibition of recombinant T. cruzi and T. brucei TR by the BTCP analogues.
T. cruzi TR
T. brucei TR
Compd
R
Ar1
Ar2
Inh. [%][a]
Kic [mm]
Inh. [%][a]
K
ic [mm]
[b]
[b]
4
5
6
7
H
H
H
Ph
5-cyanobenzo[b]thien-2-yl
5-methoxybenzo[b]thien-2-yl
7-cyclopropylbenzo[b]thien-2-yl
benzo[b]thien-2-yl
28
35
32
65[c]
–
–
–
16
23
15
–
–
–
–
–
[b]
[b]
[b]
[b]
[b]
1.5ꢀ0.5
8a
8b
8c
8d
9c
9d
H
H
H
H
Ph
Ph
1H-indol-5-yl
81
69
30[c]
15[c]
53[c]
21[c]
4ꢀ1[e]
65
53[d]
–
–
–
6ꢀ1
benzo[b]thien-5-yl
5-methyloxazol-2-yl
4-methylthiazol-2-yl
5-methyloxazol-2-yl
4-methylthiazol-2-yl
6ꢀ2[e]
7ꢀ2[e]
10ꢀ1
–
–
–
–
>40
4ꢀ1
~20
–
10a
10b
10e
10 f
10g
10h
10i
H
H
H
H
H
H
H
H
H
1H-indol-5-yl
benzo[b]thien-5-yl
phenyl
4-cyanophenyl
pyrid-4-yl
79
72[d]
33
21
16
11
5
4ꢀ0.5
3ꢀ1
43
12ꢀ2
34[d]
none
none
none
none
none
none
17
10ꢀ3
[b]
–
–
–
–
–
–
–
–
[b]
–
–
–
–
[b]
[b]
pyrid-3-yl
[b]
pyrimidin-5-yl
pyridazin-4-yl
Br
[b]
10j
10k
5
14
–
[b]
–
11 a
12a
H
H
1H-indol-5-yl
1H-indol-5-yl
71
67
6ꢀ0.5
7ꢀ0.5
40
33
11ꢀ2
25ꢀ3
13a
H
1H-indol-5-yl
74
5ꢀ0.5
35
19ꢀ3
Parent compound[18]
20
H
benzo[b]thien-2-yl
81
4ꢀ1
–
–
[a] Percent inhibition by 40 mm inhibitor in the presence of 40 mm TS2, unless otherwise stated. [b] Not measured due to solubility issues. [c] 12 mm inhibi-
tor. [d] 30 mm inhibitor. [e] Mixed-mode inhibition, Kiu =uncompetitive inhibition constant: T. cruzi TR: Kiu(8a)=64ꢀ9 mm, Kiu(8b)=65ꢀ9 mm; T. brucei TR:
Kiu(8b)=47ꢀ14 mm.
Enzymatic assays
the blood–brain barrier and, therefore, to potentially combat
the central nervous system stage of HAT.[40]
All compounds were subjected to kinetic analyses with TRs of
the two parasite species, T. cruzi and T. brucei (Table 1; see SI
for experimental details). The substituted benzo[b]thiophenes
4–6 displayed substantially lower inhibition (28–35%) toward
T. cruzi TR than their parent compound (20), which, under
identical conditions (40 mm of each inhibitor and TS2), inhibits
the enzyme by 81%. Inhibitory constants for these compounds
were not determined, due to their low solubility in the assay
buffer. Replacement of the benzo[b]thiophene substituent by
two connected five-membered rings (8c,d) did not significantly
improve the inhibitory potency, yielding Kic values for competi-
tive inhibition of 10 and >40 mm, respectively. The introduc-
tion of an additional phenyl group at the C3 position of the
pyrrolidine ring (7, 9c–d) increased the activity twofold relative
to the respective proton-substituted inhibitors (8c–d, 20). With
a Kic value of 1.5ꢀ0.5 mm, ligand 7 was the most potent inhibi-
tor of the entire series. Substitution of the terminal five-mem-
Synthesis
The synthesis of BTCP analogues was achieved by two strat-
egies. The first approach involved the formation of an N,N-
acetal-type structure by attack of 1H-benzo[d][1,2,3]triazole on
commercially available 1-pyrrolidino-1-cyclohexene (14), fol-
lowed by substitution with the respective lithiated (hetero)aro-
matic compound to give tertiary amines 4–6, 8k, 10k–12k,
10b, and 13a (Scheme 1).[41,42] The targeted biarylic structures
were afforded by subsequent Suzuki cross-coupling. In the
second strategy, primary amines 15 and 16c–d were built in
three steps starting from cyclohexanone (17) and the respec-
tive organolithium species (Scheme 2). Alkylation with either
1,4-dibromobutane (18) or the 2-phenyl derivative 19 provided
compounds 7, 8c–d, and 9c–d.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 0000, 00, 1 – 13
&
3
&
ÞÞ
These are not the final page numbers!

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
cyclic rings at Ar2 (8a–b, 10a–b, 11–13a) retained
good activity, with Ki values between 3 and 7 mm.
Comparison among the series of inhibitors with dif-
ferent heteroaromatic linkers (8a, 10–13a) revealed
that inhibition of T. cruzi TR is hardly influenced by
the alteration of the respective bend angle and elec-
tronic properties.
Nearly all inhibitors, which were subjected to de-
tailed kinetic analysis (Table 1), inhibited TR from
both species competitively. In general, the ligands
showed slightly lower activity toward T. brucei TR
than toward T. cruzi TR, and small variations in the
ligand structure had a greater influence on the po-
tency of the inhibitors.
Mutation studies
The tertiary amine center in BTCP and its analogues
is likely protonated under physiological conditions.[43]
We assumed that a charge-assisted hydrogen bond
between the protonated inhibitors and the side
chain of Glu467’ could be relevant for the anticipat-
ed binding mode in the Z-site. Alternatively, if the li-
gands bind on the opposite side of the large cavity,
the side chain of Glu18 would be a suitable hydro-
gen bond acceptor.[17,25,44,45]
Scheme 1. First route to BTCP analogues. Reagents and conditions: a) 1H-benzo[d]-
[1,2,3]triazole, THF, 258C, then Ar1-Li, ꢁ85 to ꢁ508C, 29–85% (see SI); b) Ar2 boronates,
[PdCl2(dppf)], Cs2CO3, 1,4-dioxane/H2O, 80–908C, 57–91%. THF=tetrahydrofuran,
dppf=1,1’-bis(diphenylphosphino)ferrocene.
To shed light on the binding mode of the BTCP an-
alogues, the single mutants Glu18Gln, Glu18Ala,
Glu467’Gln, and Glu467’Ala of T. brucei TR were pre-
pared. All four mutant protein species largely re-
tained catalytic activity, yielding kcat values between
56 and 68 sꢁ1, compared with 78 sꢁ1 for the wild-
type enzyme (Table 2). In contrast, some KM values
were affected more strongly. Glu467’ does not seem
to play a crucial role in TS2 recognition, as reflected
by the only slightly increased KM values of 24 mm and
28 mm for the Glu467’Gln and Glu467’Ala mutants
(wild-type: KM =8–11 mm). In contrast, replacement of
Glu18 resulted in KM values of 35 mm and 115 mm, re-
spectively, for the Glu18Gln and Glu18Ala mutants.
As expected, the replacement of Glu18 by an Ala res-
idue induced a stronger effect on substrate binding
than the mutagenesis to Gln, which can be ascribed
to the loss of hydrogen bond formation with the N8
amide of TS2 (Figure 1).[16,46] Glu467’ is not involved
in hydrogen bonding to the substrate.
In contrast to TS2, the binding affinity of inhibitor
10b, designated by its Kic value, is not affected at all
by the mutagenesis of Glu467’. Inhibition of the
Glu18Gln and Glu18Ala mutants by 10a and 10b is
significantly weaker than that of the wild-type
enzyme. Introduction of a glutamine side chain at
position 18 caused a threefold increase in the inhibi-
tion constant of 10a (Kic(10a)=36 mm), very similar
Scheme 2. Second route to BTCP analogues. Reagents and conditions: a) Cl-Ar1-Ar2, tBuLi,
Et2O, ꢁ788C, 5–17%; b) NaN3, TFA, CH2Cl2, 0–258C; c) LiAlH4, Et2O, 0–258C, 37–51%; d) KI,
iPr2NEt, CH3CN, 908C, 47–73%; e) KI, iPr2NEt, CH3CN, 808C, 40–74%. TFA=trifluoroacetic
acid.
bered heteroaromatic ring Ar2 by a series of six-membered
rings resulted in significantly decreased inhibition (10e–j, 5–
33% inhibition). In contrast, the introduction of fused hetero-
to the rise in the KM value for TS2. Remarkably, the inhibition
constant for the Ala mutant is also only about threefold higher
(Kic(10a)=38 mm), while the KM value for TS2 is increased more
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 0000, 00, 1 – 13
&
4
&
ÞÞ
These are not the final page numbers!

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
establishes a hydrogen bond between Glu18 and the
protonated amine of inhibitor 10a, whereby the in-
duced binding pocket accommodates the aromatic
substituent in a similar fashion to the dihydroquina-
zolines (Figure 3a).[45]
Table 2. Kinetic constants for wild-type (WT) T. brucei TR and the Glu467’Gln,
Glu467’Ala, Glu18Gln, and Glu18Ala mutants and respective inhibition constants (Kic)
for BTCP analogues 10a and 10b.[a]
TR species
Kic [mm]
KM [mm] Vmax [Umgꢁ1
]
kcat [sꢁ1
]
kcat/KM [sꢁ1 mmꢁ1
]
10a
10b
To take account of the mutation study results, the
binding site definition of the T. cruzi TR structure was
changed so that an interaction with Glu18 was en-
forced. Two different centers (p1 and p2) with radii of
10 ꢁ were defined as ligand binding sites within the
enzyme active site (Figure 4). More than half of the
docking poses in the T. cruzi TR structure (p1) appa-
rently form a hydrogen bond with Glu18 and place
the biarylic moiety in the groove that is occupied by
WT
12ꢀ2
n.d.
n.d.
10ꢀ3
9ꢀ1
8–11
24ꢀ5
28ꢀ6
35ꢀ4
77ꢀ2
74ꢀ4
63ꢀ3
77ꢀ1
71ꢀ1
78ꢀ2
66ꢀ4
56ꢀ3
68ꢀ1
63ꢀ1
9.8ꢀ3ꢂ106
Glu467’Gln
Glu467’Ala
Glu18Gln
Glu18Ala
2.8ꢀ0.6ꢂ106
2.0ꢀ0.4ꢂ106
1.9ꢀ0.2ꢂ106
5.5ꢀ0.1ꢂ105
11ꢀ2
36ꢀ3 102ꢀ42
38ꢀ3
84ꢀ15 115ꢀ25
[a] Values are the mean ꢀ SD of two independent determinations; n.d.: not deter-
mined.
than tenfold. Although the data for inhibitor 10b exhibited
a greater variance, mainly due to solubility issues, the same
tendency was observed. These results indicate that the BTCP
analogue inhibitors recognize and differentiate between Glu18
and its mutants, predominantly by their charge. Glu is nega-
tively charged under the assay conditions, whereas Gln and
Ala are neutral. While TS2 forms a hydrogen bond with Glu18
and probably also with Gln18, and therefore discriminates be-
tween Gln and Ala, as reflected by the KM values, BTCP ana-
logue inhibitors most likely establish a Coulombic interaction
with Glu18 that cannot be formed with Gln18 or with Ala18.
gGluI of TS2 in the substrate complex (Figure 3b).[16]
Besides a few conformations that involved hydrogen bond-
ing to Tyr110 in the docking setup with the T. cruzi TR struc-
Docking studies
In the absence of structural information for the ligand–protein
complexes, a docking study using the program GOLD (genetic
optimization for ligand docking) was conducted.[47,48] GOLD is
one of the three most used docking programs[49] and has al-
ready been employed for computing binding modes of TR in-
hibitors.[26,50,51] Despite its large active site, TR seems to be an
acceptable target because of its hypothesized active site rigidi-
ty.[52]
In our study, the co-crystal structure of T. cruzi TR with
bound TS2 (PDB ID: 1BZL[16]) served as a docking template. In
addition, the co-crystal structure of T. brucei TR in complex
with a dihydroquinazoline inhibitor (PDB ID: 2WPF[45]) was
chosen, due to the unusual conformation of Met113 that leads
to the formation of a subpocket in the active site, which we
postulated to potentially accommodate BTCP analogue inhibi-
tors. All atoms within a 12 ꢁ radius for the T. cruzi TR structure
and a 15 ꢁ radius for the T. brucei TR structure of the respec-
tive reference ligand were used for the binding site definition,
so that the docking simulation was able to sample the whole
active site for ligand binding modes. The docking simulation
was executed for both enzymes with inhibitor 10a.
The docking simulation for the T. cruzi TR structure generat-
ed the postulated Z-site binding mode within a root-mean-
square deviation (RMSD) of 3–4 ꢁ. The overlay of all predicted
poses suggests that the Z-site is favored (Figure 1SI). No inter-
actions between the protonated amine of BTCP analogue 10a
and Glu18 were postulated. In the docking run of the T. brucei
TR structure, the predicted poses clustered in the induced
binding pocket (Figure 2SI) and suggest a binding mode that
Figure 3. GOLD-predicted binding conformations that were selected in con-
junction with the results of the mutation study: a) Selected pose generated
by a docking simulation on T. brucei TR (PDB ID: 2WPF[43]). b) Selected poses
generated by a docking simulation on T. cruzi TR (PDB ID: 1BZL[16]) with the
ligand binding site center p2; Cenzyme =grey, O=red, N=blue, S=yellow,
C
ligand =green.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 0000, 00, 1 – 13
&
5
&
ÞÞ
These are not the final page numbers!

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
complexes with reasonable statistics (Table 3). The
crystallographic results largely confirmed the predic-
tions by GOLD, within an RMSD of 2–3 ꢁ.
In line with its competitive mode of inhibition, 10a
binds in the trypanothione binding site in both
enzyme complexes. Electron density of the ligand
was observed in both active sites of the homodimeric
proteins but with differences in occupancy. Specifical-
ly, 10a occupies the mepacrine binding site (Fig-
ure 3SI),[44] and the tertiary amine, which is assumed
to be protonated, is located in close proximity to
Glu18 in accordance with the results of the mutation
study. However, distances of 3.8 ꢁ (T. brucei TR) and
4.0 ꢁ (T. cruzi TR) between the oppositely charged
atoms indicate a long-range Coulombic interaction
rather than a strong charge-assisted hydrogen bond.
This correlates with the observation of an only 3.5-
fold increased Kic value of the Glu18Ala mutant rela-
tive to the wild-type enzyme.[53] In both TR species,
Glu467’ is not involved in binding.
Key interactions for binding of 10a to T. brucei TR
are formed with the hydrophobic wall, comprising
Leu17, Trp21, Tyr110, and Met113 (Figure 5).[54]
Met113 occupies the same position as in complex
with the natural substrate; an opening corresponding
to an induced binding pocket, as suggested previ-
ously, was not observed (for overlay, see Fig-
Figure 4. Two different centers (p1 and p2) with a radius of 10 ꢁ were defined as puta-
tive ligand binding sites in the active site of TR and applied in docking simulations with
GOLD (red spheres); Cenzyme =grey, O=red, N=blue, S=yellow (T. cruzi TR; PDB ID:
1BZL[16]).
ture and p2, the majority established an interaction with
Glu18, whereby the poses clustered to three different putative
binding modes. The first mode is the same as observed in the
equivalent setup with p1, occupying the groove of gGluI (Fig-
ure 3b). The other two proposed conformations bind to the
hydrophobic wall, similar to the mepacrine complex.[44]
ure 4SI).[16,45] The side chain of Met113 is involved in S···p
(d(SMet113···thiazole)=3.9 ꢁ), SC-H···p (d(SCH2Met113···indole)=
3.7 ꢁ), and dispersive S···S (d=4.1 ꢁ) interactions with the
indole and thiazole rings.
Further contributions to the binding affinity are made by C-
H···p interactions between Trp21 and the cyclohexyl and pyrro-
lidine moieties, where the latter undergoes additional disper-
sive contacts with Leu17. An overlay of this co-crystal structure
with another structure of T. cruzi TR (PDB ID: 1AOG[55]) shows
that the pyrrolidine ring is placed in a small groove that is oc-
cupied by water in the free enzyme (Figure 5SI). The overlay
also suggests the possibility of a solvent-mediated hydrogen
bond between the protonated pyrrolidine nitrogen atom and
Glu18 in the complex with 10a.
In the T. brucei TR complex, the pyrrole moiety of the indole
substituent is oriented parallel to the amide bond of Gly112
and Met113 and gains energy from amide···p stacking interac-
tions (Figure 5).[37] The overlay of the two active sites reveals
that the ligand can also be complexed with its indole ring
turned by 1808 (Figure 6SI). This conformation is less optimal
for dipolar amide···p stacking interactions, but instead, a C-
H···p interaction with Met113 (Ca) is formed. A third inhibitor
molecule was found in a groove close to Pro42 at a crystal-
packing interface, suggesting that 10a participated in the pro-
tein crystallization process. Most likely, the inhibitor does not
populate this cavity in solution.
Considering the findings of the mutation study, a selection
of the predicted poses led to four hypothetical binding modes
(Figure 3). In the first binding conformation, the aromatic sub-
stituent occupies the gGluI groove; in the second, it is accom-
modated by the induced binding pocket; and in the last two,
it binds to the hydrophobic wall of the mepacrine binding site,
either in a horizontal or a vertical orientation. We docked the
inhibitor using the ChemPLP scoring function and obtained
values between 47 and 66 for the four suggested poses. Re-
scoring of all solutions with the GOLDscore function resulted
in comparably higher values (32 and 40 versus 9 and 13) for
the binding poses at the mepacrine binding site and suggest
a preference for this possibility.
X-ray crystal structure analysis
Co-crystallization by hanging drop vapor diffusion resulted in
crystals of 10a in complex with both T. brucei TR (PDB ID:
4NEV) and T. cruzi TR (PDB ID: 4NEW) with maximum resolution
of 2.5 ꢁ and 2.8 ꢁ, respectively (Figures 5 and 6). The presence
of significant amounts of DMSO, necessary to assure sufficient
solubility of the inhibitor, is a probable reason for somewhat
compromised diffraction patterns. Nevertheless, diffraction
data could be processed and led to molecular models of the
The co-crystal structure of T. cruzi TR with 10a also does not
display an induced binding pocket, as a result of a conforma-
tional shift of the Met113 side chain (Figure 6).[45] The overall
location of the ligand in the mepacrine binding site is main-
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 0000, 00, 1 – 13
&
6
&
ÞÞ
These are not the final page numbers!

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
The protonated pyrrolidine ring puckers differently in the
T. cruzi TR complex and adopts a geometry that does not
enable hydrogen bonding of the ligand to Glu18. Rather, the
two moieties undergo Coulombic interactions at a similar dis-
tance to that observed in the T. brucei TR structure. The cyclo-
hexyl ring is located closer to Trp21 and undergoes short C-
H···p interactions (d(C···C)=3.3 ꢁ) with the indole side chain.
The thiazole ring is rotated by 1808 and establishes weaker in-
termolecular contacts than in the T. brucei TR structure. It di-
rects the indole substituent toward Ile106 at the bottom of
the binding cleft. A short O-H···p interaction (d(O···C)=2.8 ꢁ) is
observed between the side chain of Ser109 and the indole
ring.
Table 3. X-ray co-crystal structures: data collection and refinement statis-
tics.
Crystal data (PDB ID)
T. brucei
T. cruzi
TR+10a (4NEV)
TR+10a (4NEW)
A) Data collection and processing
l [ꢁ]
0.979
1.54
space group
a [ꢁ]
b [ꢁ]
P43212
117.2
117.2
224.6
P4322
86.8
86.8
c [ꢁ]
151.1
B) Diffraction data
resolution range [ꢁ]
highest-resolution shell [ꢁ]
unique reflections
R (I)sym [%][a]
66.0–2.50
2.55–2.50
56511
17.0–2.80
2.90–2.80
14050
It may be difficult to generalize the structural findings of
these two co-crystal structures for a larger family of BTCP-de-
rived ligands. The finding that even the same inhibitor did not
display an identical binding mode in the TRs from two differ-
ent species suggests that minor differences in both protein
and inhibitor can largely affect the interaction. Different bind-
ing modes for other inhibitors cannot be excluded, in particu-
lar for large ligands, such as BTCP-derived conjugate 3. Also,
the potent phenylated ligands 7 and 9c–d could adopt anoth-
er binding geometry in the large active site. In the binding
modes observed for 10a, there is no room to accommodate
the phenyl ring. Furthermore, given the substantial protein in-
teractions of the indole ring in 10a in the two structures, it is
difficult to explain how compound 8c, with its smaller 5-meth-
yloxazolyl substituent (instead of the indole ring), can have
a similar binding affinity against T. cruzi TR (Table 1).
7.2 (53)
2.8 (11)
completeness [%]
redundancy
I/s (I)
99.8 (100)
8.0 (6.8)
16.4 (3.3)
94.2 (75.4)
1.7 (1.3)
30.0 (7.0)
C) Refinement
programs used
CNS, Refmac,
PHENIX
66.0–2.5
55052
23.4
25.8
CNS, Refmac,
PHENIX
17.0–2.80
14005
23.2
resolution range [ꢁ]
reflections used
[b]
Rwork
Rfree
[c]
27.8
no. of atoms (non-hydrogen) 7667
3787
protein atoms
water molecules
ligand atoms
7368
63
236
3709
0
78
RMSD, angle [8]
RMSD, bond [ꢁ]
0.9093
0.0046
1.83
0.011
Ramachandran plot
most favored regions [%]
additionally allowed regions
[%]
generously allowed regions
[%]
At this point, it becomes clear that the extended cavity of
TR allows numerous binding modes for small-molecule ligands.
This is reflected in the three different binding modes found for
inhibitor 10a for both T. brucei TR and T. cruzi TR. Baiocco et al.
also reported the presence of more than one binding confor-
mation for their azole-based TR inhibitor.[26] Taken together,
these findings confirm that the hydrophobic wall plays a crucial
role in the binding of small-molecule inhibitors in TR. In all
crystal structures of TR–inhibitor complexes known so far, this
protein site is involved in ligand recognition and therefore
should be considered with priority when evaluating potential
binding sites in the absence of structural information.[26,44,45]
94.6
4.5
87.3
10.6
0.93
2.1
mean B factors [ꢁ2]
protein atoms
water molecules
ligand atoms
67.8
55.7
79.3
79.9
N/A
87.2
[a] R(I)sym =[ꢀhꢀi jIi(h)ꢁhI(h)ij/ꢀhꢀiIi(h)]ꢂ100, where hI(h)i is the mean of
the I(h) observation of reflection h. [b] Rwork =ꢀhkl jFoꢁFc j/ꢀhkl jFo j. [c] Rfree
was calculated as shown for Rwork, but on refinement, excluded 5% of
data. N/A: not applicable.
Selectivity, antiparasitic activities, and cytotoxicity
tained, as expected from the fact that T. brucei and T. cruzi TR
exhibit a sequence identity of over 80%.[45] Nevertheless, the
binding mode of 10a in T. cruzi TR features some significant
differences. While, for the two enzyme species, all 25 amino
acid residues that are in contact with bound TS2 are con-
served,[16] Gly112 of T. brucei TR at the binding site of 10a is re-
placed by Glu112 in T. cruzi TR. This bulkier residue with its
highly polar, strong solvation-requiring side chain might cause
the 1808 rotation about the bond connecting the thiazole ring
to the quaternary carbon atom, which orients the indole ring
away from Met113 (Figure 6b). A superimposition of the two
co-crystal structures nicely illustrates the different orientations
of the indole ring (Figure 7SI).
To gain insight into the selectivity of the BTCP analogues for
the parasite TR, molecules 8a and 10–13a were studied
against hGR, the closest related host enzyme (Table 1SI). All
four compounds showed very low or no activity against hGR.
Only 8a displayed some inhibition, and detailed kinetic analy-
sis revealed a Ki value of 141ꢀ11 mm and a non-competitive
mode of inhibition. In comparison, 8a showed mixed inhibi-
tion toward T. cruzi TR, with Kic and Kiu values of 4 mm and
64 mm, respectively.
The in vitro activities of the new BTCP analogues were stud-
ied against the trypanosomatids T. brucei rhodesiense, T. cruzi,
and the malarial parasite Plasmodium falciparum (Table 4). In
contrast to the enzymatic assay, all BTCP analogues inhibited
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 0000, 00, 1 – 13
&
7
&
ÞÞ
These are not the final page numbers!

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
Conclusions
Herein we present the elucidation of the binding
mode for BTCP analogue inhibitors in the active site
of T. cruzi TR and T. brucei TR. An extensive SAR study,
based on rational ligand design, led to reconsidera-
tion of the previously proposed Z-site as binding
pocket. These findings were supported by mutational
analyses that revealed an impact of Glu18, but not
Glu467’, on inhibitor binding. The protonated amine
of the ligands and the Glu18 side chain are likely en-
gaged in a Coulombic interaction rather than in
a short hydrogen bond. Docking simulations with the
program GOLD predicted, in conjunction with the
mutagenesis study, four possible binding modes. Two
of these were suggested as most likely, due to their
scoring values, and were subsequently confirmed by
X-ray structure analysis. This demonstrated the utility
of combining a mutagenesis study with virtual dock-
ing. The X-ray co-crystal structures of 10a in complex
with T. cruzi TR and T. brucei TR confirm a binding site
that includes Glu18 and reinforces the presence of
a Coulombic interaction with the protonated tertiary
amine center of the ligand. The biarylic moiety in
compound 10a binds in both structures to the hy-
drophobic wall of the mepacrine binding site but in
a different orientation. This study reveals, for the first
time, that inhibitors of TR behave differently toward
TRs of different parasite species. This is reflected by
both the structural data and the inhibition constants
of the BTCP analogues. Experiments with hGR and L6
myoblast cells as controls corroborated the selectivity
of the BTCP analogues for TR and the parasitic cells.
However, a direct correlation between TR and cell
Figure 5. Crystal structure of T. brucei TR in complex with BTCP analogue 10a (PDB ID:
4NEV): a) Overview of the TS2 binding site of the enzyme with ligand 10a bound to the
hydrophobic wall and Glu18 at the mepacrine binding site. b) Closest contacts between
10a and protein residues; Cenzyme =grey, O=red, N=blue, S=yellow, Cligand =green. The
2FoꢁFc electron density map is drawn at 1s as a green mesh; distances are shown as
dashed lines and are given in ꢁ.
the cell growth of T. brucei rhodesiense more efficiently than
that of T. cruzi. This may at least partially be due to the fact
that African trypanosomes are extracellular parasites whereas
T. cruzi multiplies in a mammalian host cell and thus may be
less accessible by the compounds. Interestingly, the compara-
bly weak TR inhibitors 4–6 had an in vitro activity for T. brucei
rhodesiense that is among the highest of the series (1.7 to
4.4 mm). Correlation of the in vitro results with the kinetic anal-
ysis of the recombinant proteins is not straightforward. This
may be due to several reasons; for instance, the uptake of the
compounds by the intact parasite. It also indicates that TR may
not be the only cellular target of the BTCP analogues. This is
supported by the fact that P. falciparum is inhibited to a similar
degree (2.2 to 29.3 mm) as T. brucei rhodesiense, although the
malarial parasite does not possess a trypanothione-based
redox metabolism.
growth inhibition was not observed, which indicates that the
enzyme might not be the only target of BTCP analogues in
these parasitic protozoa, an assumption that is supported by
the rather low selectivity indices.
Experimental Section
Chemistry
Materials and general methods: In the following, the experimen-
tal details for the syntheses of compounds 5, 10k, and 10 f are de-
scribed as examples for the first synthetic route, and compounds
16c, 8c, and 9c for the second synthetic route. All other synthetic
details, experimental data, and description of the biological assays
and the molecular docking are provided in the Supporting Infor-
mation. Crystallographic figures were prepared using PyMOL.[56]
1-[1-(5-Methoxybenzo[b]thien-2-yl)cyclohexyl]pyrrolidine (5): A
solution of 5-methoxybenzo[b]thiophene[57] (35 mg, 213 mmol) in
THF (1.4 mL) was treated with 1.6 m nBuLi in hexane (133 mL) and
stirred at ꢁ788C for 90 min (solution 1). At the same time, a solu-
tion of 1-(cyclohex-1-en-1-yl)pyrrolidine (14) (23 mL, 142 mmol) in
THF (1.4 mL) was treated with 1H-benzo[d][1,2,3]triazole (17 mg,
142 mmol) at 258C and stirred for 90 min (solution 2). Solution 2
was cooled to ꢁ788C, treated dropwise with solution 1, stirred at
With the exception of compound 7, the cytotoxicity against
L6 myoblast cells was lower (higher IC50 values) than the ob-
served in vitro activity for the parasitic protozoa, yielding selec-
tivity indices (IC50 for L6 cells/IC50 for parasite) for T. brucei rho-
desiense of moderate values up to 24.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 0000, 00, 1 – 13
&
8
&
ÞÞ
These are not the final page numbers!

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
(1.06 mL, 6.61 mmol) and stirred at 258C for 75 min (so-
lution 1).
A
solution of 2-bromothiazole (834 mL,
9.26 mmol) in THF (33 mL) was cooled to ꢁ408C, treated
dropwise with 1 m lithium magnesium 2,2,6,6-tetrame-
thylpiperidin-1-ide dichloride in THF/toluene (Aldrich;
9.92 mL), and stirred at this temperature for 20 min (so-
lution 2). Solution 2 was cooled to ꢁ508C, treated drop-
wise with cooled (ꢁ508C) solution 1, stirred at this tem-
perature for 1 h, allowed to warm to 258C over a period
of 3 h, treated with water, and diluted with EtOAc
(5 mL). After separation of the phases, the aqueous layer
was extracted with EtOAc (3ꢂ20 mL). The combined or-
ganic layers were dried over MgSO4, filtered, and evapo-
rated. Flash chromatography (SiO2; cyclohexane/EtOAc/
NEt3, 99:1:0 to 95:5:1) gave 10k (601 mg, 29%) as pale
yellow crystals: Rf =0.41 (SiO2; cyclohexane/EtOAc/NEt3,
95:5:1); mp: 69–708C; 1H NMR (400 MHz, CDCl3): d=
1.32–1.48 (m, 4H), 1.61–1.69 (m, 4H), 1.67–1.74 (m, 2H),
1.93 (t, J=5.8 Hz, 4H), 2.52 (td, J=6.5, 5.3, 2.5 Hz, 4H),
7.31 ppm (s, 1H); 13C NMR (100 MHz, CDCl3): d=22.26,
23.64, 25.80, 37.65, 45.08, 58.83, 134.31, 139.83,
~
144.19 ppm; IR (ATR): n=3086, 2966, 2938, 2815, 1695,
1505, 1445, 1397, 1125, 1010, 883, 603 cmꢁ1; HRMS (ESI)
m/z (%): calcd for C13H2079BrN2S+ [M+H]+: 315.0525,
found: 315.0525 (100); Anal. calcd for C13H19BrN2S: C
49.53, H 6.07, N 8.89; found: C 49.92, H 6.14, N 8.71.
4-{5-[1-(1-Pyrrolidinyl)cyclohexyl]-1,3-thiazol-2-yl}ben-
zonitrile (10 f): A solution of 10k (30 mg, 95 mmol), (4-
cyanophenyl)boronic acid (15 mg, 105 mmol), and Cs2CO3
(62 mg, 190 mmol) in 1,4-dioxane/water (10:1, 1.1 mL)
was degassed under argon, treated with [PdCl2(dppf)]
(7 mg, 9.5 mmol), heated at 808C for 4 h, and filtered
through a pad of silica gel and MgSO4, eluting with cy-
clohexane/EtOAc/NEt3 (5:5:0.1). Evaporation and flash
chromatography (SiO2; cyclohexane/EtOAc/NEt3, 90:10:0
to 90:10:1 to 80:20:1) gave 10 f (20 mg, 83%) as light-
yellow crystals: Rf =0.24 (SiO2; cyclohexane/EtOAc/NEt3,
Figure 6. Crystal structure of T. cruzi TR in complex with BTCP analogue 10a (PDB ID:
4NEW): a) Overview of the TS2 binding site of the enzyme with ligand 10a bound to the
hydrophobic wall and Glu18 at the mepacrine binding site. b) Closest contacts between
10a and protein residues; Cenzyme =grey, O=red, N=blue, S=yellow, Cligand =green. The
2FoꢁFc electron density map is drawn at 0.5s as a green mesh; distances are shown as
dashed lines and are given in ꢁ.
1
80:20:1); mp: 119–1228C; H NMR (400 MHz, CD2Cl2): d=
Table 4. In vitro activities of the BTCP analogues against T. brucei rhode-
siense (T.b.r.), T. cruzi (T.c.), P. falciparum NF54 (P.f.), and mammalian L6
cells.
this temperature for 20 min, treated with water (3 mL), allowed to
warm to 258C, and diluted with EtOAc (3 mL). After separation of
the layers, the aqueous layer was extracted with EtOAc (3ꢂ10 mL).
The combined organic layers were dried over MgSO4, filtered, and
evaporated. Flash chromatography (SiO2; cyclohexane/EtOAc/NEt3,
9:1:0 to 9:1:0.1) gave 5 (38 mg, 85%) as white crystals: Rf =0.14
(SiO2; heptane/EtOAc, 8:2); mp: 70–718C; 1H NMR (400 MHz,
CDCl3): d=1.38–1.53 (m, 4H), 1.59–1.66 (m, 4H), 1.71 (tt, J=7.0,
2.6 Hz, 2H), 1.97 (ddd, J=13.0, 9.5, 3.5 Hz, 2H), 2.12–2.23 (m, 2H),
2.60–2.67 (m, 4H), 3.87 (s, 3H), 6.94 (dd, J=8.8, 2.5 Hz, 1H), 7.02
(d, J=0.7 Hz, 1H), 7.22 (d, J=2.5 Hz, 1H), 7.65 ppm (dt, J=8.7,
0.6 Hz, 1H); 13C NMR (100 MHz, CDCl3): d=22.74, 23.57, 25.95,
37.40, 45.17, 55.68, 59.83, 105.44, 113.96, 121.25, 122.71, 131.54,
Compd
IC50 [mm][a]
P.f.
SI[b]
(L6/T.b.r.)
T.b.r.
T.c.
L6
4
5
6
7
4.4
1.9
1.7
8.9
1.7
45.2
21.9
16.5
49.8
7.0
9.9
4.4
2.6
29.3
2.3
2.2
2.9
7.4
2.25
3.5
7.8
105
41.1
26.1
4.6
24
22
15
0.5
6
17
4
7
5
3
8
11
6
8a
8b
8c
8d
9c
9d
10a
10b
10k
11 a
12a
10.6
13.5
141.6
104.0
124.1
41.4
28.9
40.0
127.1
28.2
38.3
0.8
4.9
37.6
15.9
23.4
14.0
3.5
3.6
19.8
2.1
39.8
51.1
46.4
21.6
19.0
7.9
89.9
17.4
25.9
~
140.71, 147.77, 157.42 ppm; IR (ATR): n=2934, 2851, 1595, 1565,
1452, 1334, 1233, 1158, 1029, 850 cmꢁ1; HRMS (ESI) m/z (%): calcd
for C19H26NOS+ [M+H]+: 316.1730, found: 316.1736 (8), calcd for
C15H17OS+ [MꢁC4H8N]+: 245.0995, found: 245.1012 (100); Anal.
calcd for C19H25NOS: C 72.34, H 7.99, N 4.44, found: C 72.29, H 8.32,
N 4.06.
8.3
11.8
3.4
13
3
12.8
8.9
[a] Values are the means of two independent assays; individual values
vary by less than a factor of 2. [b] Selectivity index (SI) corresponds to the
ratio of the IC50 values for T. brucei rhodesiense and L6 cells (cytotoxicity).
2-Bromo-5-[1-(pyrrolidinyl)cyclohexyl]-1,3-thiazole (10k): A solu-
tion of 1H-benzo[d][1,2,3]triazole (788 mg, 6.61 mmol) in THF
(33 mL) was treated with 1-(cyclohex-1-en-1-yl)pyrrolidine (14)
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 0000, 00, 1 – 13
&
9
&
ÞÞ
These are not the final page numbers!

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
1.35–1.52 (m, 4H), 1.61–1.71 (m, 4H), 1.72 (m, 2H), 1.93–2.15 (m,
4H), 2.50–2.65 (m, 4H), 7.66 (s, 1H), 7.69–7.75 (m, 2H), 7.99–
8.08 ppm (m, 2H); 13C NMR (100 MHz, CD2Cl2): d=22.72, 24.17,
26.36, 38.20, 45.39, 58.89, 113.18, 119.11, 126.99, 133.31, 138.40,
12 h. The mixture was diluted with CH2Cl2 (10 mL) and washed
with H2O (10 mL). The phases were separated, and the aqueous
layer was extracted with CH2Cl2 (3ꢂ10 mL). The combined organic
layers were dried over MgSO4, filtered, and evaporated. Flash chro-
matography (SiO2; heptane/EtOAc, 4:1) gave (ꢀ)-9c (68 mg, 60%)
as a brown oil: Rf =0.29 (SiO2; heptane/EtOAc, 4:1); 1H NMR
(400 MHz, CDCl3): d=1.37–1.57 (m, 4H), 1.66–1.81 (m, 3H), 1.96–
2.08 (m, 4H), 2.13 (dtd, J=12.2, 8.2, 4.0 Hz, 1H), 2.37 (d, J=1.2 Hz,
3H), 2.61 (dd, J=8.8, 7.5 Hz, 1H), 2.79 (q, J=8.8 Hz, 1H), 2.88 (td,
J=8.7, 4.0 Hz, 1H), 3.02–3.20 (m, 2H), 6.77 (q, J=1.2 Hz, 1H), 6.85
(d, J=3.8 Hz, 1H), 7.13–7.29 (m, 5H), 7.53 ppm (d, J=3.8 Hz, 1H);
13C NMR (100 MHz, CDCl3): d=11.12, 22.45 and 22.47 (1C), 25.94,
32.88, 37.12 and 37.43 (1C), 43.28, 45.52, 53.35, 59.44, 124.12,
125.18, 126.14, 126.27, 127.42, 128.07, 128.40, 145.35, 148.31,
~
141.92, 143.09, 164.02 ppm; IR (ATR): n=3092, 2928, 2799, 2229,
1607, 1493, 1442, 1267, 1152, 977, 840, 631 cmꢁ1; HRMS (ESI) m/z
(%): calcd for C20H24N3S+ [M+H]+: 338.1685, found: 338.1689 (40),
calcd for C16H15N2S+ [MꢁC4H8N]+: 267.0950, found: 267.0955 (100);
Anal. calcd for C20H23N3S: C 71.18, H 6.87, N 12.45; found: C 70.91,
H 6.89, N 12.18.
1-[5-(5-Methyloxazol-2-yl)thien-2-yl]cyclohexaneamine (16c):
A
solution of alcohol 23 (Scheme 4SI) (440 mg, 1.67 mmol) in CH2Cl2
(5 mL) was treated at 08C with NaN3 (325 mg, 5.01 mmol) and then
with TFA (0.51 mL, 6.68 mmol) over 5 min. The solution was stirred
at 258C for 15 h and treated slowly with H2O (10 mL). The phases
were separated, and the aqueous layer was extracted with EtOAc
(3ꢂ8 mL). The combined organic layers were dried over MgSO4, fil-
tered, and evaporated. The residue was dissolved in Et2O (2 mL),
treated with LiAlH4 (127 mg, 3.34 mmol) at 08C, stirred at 258C for
5 h, and treated dropwise with iPrOH (0.3 mL) at 08C and then
with Rochelle solution (10 mL). The phases were separated, and
the aqueous layer was extracted with CH2Cl2 (2ꢂ15 mL) and
CH2Cl2/iPrOH (3:1, 3ꢂ15 mL). The combined organic layers were
dried over MgSO4, filtered, and evaporated. Flash chromatography
(SiO2; EtOAc/MeOH/NH3, 1:0:0 to 9:1:1) gave amine 16c (161 mg,
37%) as a brown oil: Rf =0.42 (SiO2; EtOAc/MeOH/NH3, 9:1:1);
1H NMR (400 MHz, CDCl3): d=1.50–1.71 (m, 6H), 1.71–1.81 (m, 2H),
1.87–2.02 (m, 2H), 2.34 (d, J=1.3 Hz, 3H), 6.74 (q, J=1.2 Hz, 1H),
6.91 (d, J=3.8 Hz, 1H), 7.42 ppm (d, J=3.8 Hz, 1H); 13C NMR
(100 MHz, CDCl3): d=11.07, 22.55, 25.62, 40.71, 53.89, 122.53,
~
148.36, 157.27 ppm; IR (ATR): n=2931, 2856, 1605, 1494, 1444,
1352, 1159, 1115, 999, 907, 806, 723, 699 cmꢁ1; HRMS (MALDI) m/z
(%): calcd for C24H29N2OS+ [M+H]+: 393.1995, found: 393.1992 (3),
calcd for C14H16NOS+ [MꢁC10H12N]+: 246.0947, found: 246.0937
(100).
Biology
Cloning, heterologous expression, and purification of wild-type and
mutant T. brucei TR species: The site-directed mutagenesis of TR
was performed using the QuikChange site-directed mutagenesis
kit (Stratagene), using the primers depicted in Table 2 (SI). The
pET3a vector with the wild-type gene[58] served as a template.
Competent E. coli Nova blue cells were transformed and grown at
378C in LB medium containing 100 mgmLꢁ1 carbenicillin. After pu-
rification of the plasmid with the Nucleospin plasmid kit (Macher-
ey–Nagel), the mutation was confirmed by DNA sequencing
(MWG-Biotech or GATC). Competent E. coli BL21 (DE3) pLysS cells
were transformed, and the expression was performed as de-
scribed,[58] with the exception that 5 L of terrific broth medium
containing 34 mgmLꢁ1 chloramphenicol and 100 mgmLꢁ1 carbeni-
cillin was used. Recombinant proteins were purified by a procedure
combining previous purification schemes.[21,58,59] The bacteria were
harvested and suspended in 50 mm potassium phosphate (pH 7.0),
1 mm EDTA, containing 0.04% sodium deoxycholate, 10 mm FAD,
2 mm dithiothreitol, 10 mm phenylmethanesulfonyl fluoride,
0.15 mm pepstatin, and 5 mg lysozyme. DNase I was added, and
the cells were disintegrated by sonication. After centrifugation, the
supernatant was kept on ice, the pellet was re-extracted, and both
extracts were combined. Streptomycin (3%) and 35% ammonium
sulfate were added, the suspension was centrifuged, and the pellet
was discarded. In the supernatant, the ammonium sulfate satura-
tion was increased to 70%. The resulting protein pellet was dis-
solved in 5 mm potassium phosphate (pH 7.0) and 0.1 mm EDTA
and dialyzed twice for 2 h at 48C. The solution was loaded onto
a DEAE (diethylaminoethyl)-Sepharose column and washed with
25 mm potassium phosphate (pH 7.0) and 0.5 mm EDTA (buffer A).
After elution with 300 mm KCl in buffer A, the sample was subject-
ed to affinity chromatography on a 2’,5’-ADP-Sepharose column.
Pure TR was eluted by 300 mm NADPH in buffer A as described pre-
viously.[21]
~
124.08, 126.84, 127.99, 148.26, 157.23, 160.05 ppm; IR (ATR): n=
3115, 3063, 2925, 1608, 1519, 1472, 1401, 1373, 1297, 1221, 1163,
1015, 995, 919, 900, 805, 722 cmꢁ1; HRMS (ESI) m/z (%): calcd for
C14H19N2OS+ [M+H]+: 263.1213, found: 263.1210 (100).
5-Methyl-2-{5-[1-(pyrrolidine-1-yl)cyclohexyl]thien-2-yl}oxazole
(8c): A solution of amine 16c (57 mg, 0.22 mmol) in CH3CN (5 mL)
was treated with KI (37 mg, 0.22 mmol), iPr2NEt (76 mL, 0.44 mmol),
and 1,4-dibromobutane (18) (26 mL, 0.22 mmol) and stirred at 908C
for 36 h, while the mixture was treated with 1,4-dibromobutane
(18) (26 mL, 0.22 mmol) and iPr2NEt (76 mL, 0.44 mmol) every 12 h.
The mixture was evaporated, diluted with CH2Cl2 (10 mL), and
washed with H2O (10 mL). The phases were separated, and the
aqueous layer was extracted with CH2Cl2/iPrOH (3:1, 5ꢂ10 mL).
The combined organic layers were dried over MgSO4, filtered, and
evaporated. Flash chromatography (SiO2; EtOAc/MeOH/NH3,
10:0.1:0.1) gave 8c (33 mg, 47%) as a brown oil: Rf =0.41 (SiO2;
EtOAc/MeOH/NH3, 10:0.1:0.1); 1H NMR (400 MHz, CDCl3): d=1.35–
1.49 (m, 4H), 1.57–1.77 (m, 6H), 1.91–2.03 (m, 2H), 2.06–2.18 (m,
2H), 2.36 (d, J=1.2 Hz, 3H), 2.54–2.66 (m, 4H), 6.76 (q, J=1.2 Hz,
1H), 6.84 (d, J=3.8 Hz, 1H), 7.51 ppm (d, J=3.8 Hz, 1H); 13C NMR
(100 MHz, CDCl3): d=11.13, 22.59, 23.50, 25.84, 37.38, 45.13, 59.94,
124.12, 125.76, 126.27, 128.25, 148.34 (2 C), 157.20 ppm; IR (ATR):
~
n=2930, 2855, 1605, 1587, 1442, 1351, 1262, 1163, 1115, 997, 804,
735 cmꢁ1; HRMS (MALDI) m/z (%): calcd for C18H25N2OS+ [M+H]+:
317.1682, found: 317.1686 (10), calcd for C14H16NOS+ [MꢁC4H9N]+:
246.0947, found: 246.0946 (100).
X-ray crystallography
(ꢀ)-5-Methyl-2-{5-[1-(3-phenylpyrrolidin-1-yl)cyclohexyl]thien-2-
yl}oxazole ((ꢀ)-9c): A solution of amine 16c (76 mg, 0.29 mmol) in
CH3CN (5 mL) was treated with KI (48 mg, 0.29 mmol), iPr2NEt
(0.20 mL, 1.16 mmol), and (ꢀ)-19 (169 mg, 0.58 mmol). The solution
was stirred at 808C for 84 h, while the mixture was treated with
(ꢀ)-19 (169 mg, 0.58 mmol) and iPr2NEt (0.20 mL, 1.16 mmol) every
T. brucei and T. cruzi TR co-crystallization, X-ray data collection,
structure determination, and refinement: Proteins for crystalliza-
tion were concentrated to 10 mgmLꢁ1 in 20 mm TRIS, pH 8. Inhibi-
tor 10a was dissolved in DMSO (10 mm) and added in 1 mL ali-
quots to 95 mL of the protein solution, resulting in a final concen-
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 0000, 00, 1 – 13
&10
&
ÞÞ
These are not the final page numbers!

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
[8] A. Schmidt, R. L. Krauth-Siegel, Curr. Top. Med. Chem. 2002, 2, 1239–
1259.
[9] L. S. C. Bernardes, C. L. Zani, I. Carvalho, Curr. Med. Chem. 2013, 20,
2673–2696.
[10] A. H. Fairlamb, P. Blackburn, P. Ulrich, B. T. Chait, A. Cerami, Science
1985, 227, 1485–1487.
[11] G. E. Schulz, R. H. Schirmer, W. Sachsenheimer, E. F. Pai, Nature 1978,
273, 120–124.
[12] S. L. Shames, A. H. Fairlamb, A. Cerami, C. T. Walsh, Biochemistry 1986,
25, 3519–3526.
[13] T. Aboagye-Kwarteng, K. Smith, A. H. Fairlamb, Mol. Microbiol. 1992, 6,
3089–3099.
[14] G. B. Henderson, A. H. Fairlamb, P. Ulrich, A. Cerami, Biochemistry 1987,
26, 3023–3027.
[15] C. H. Faerman, S. N. Savvides, C. Strickland, M. A. Breidenbach, J. A. Po-
nasik, B. Ganem, D. Ripoll, R. L. Krauth-Siegel, P. A. Karplus, Bioorg. Med.
Chem. 1996, 4, 1247–1253.
[16] C. S. Bond, Y. Zhang, M. Berriman, M. L. Cunningham, A. H. Fairlamb,
W. N. Hunter, Structure 1999, 7, 81–89.
[17] B. Stump, C. Eberle, M. Kaiser, R. Brun, R. L. Krauth-Siegel, F. Diederich,
Org. Biomol. Chem. 2008, 6, 3935–3947.
[18] C. Eberle, B. S. Lauber, D. Fankhauser, M. Kaiser, R. Brun, R. L. Krauth-
Siegel, F. Diederich, ChemMedChem 2011, 6, 292–301.
[19] J. L. Richardson, I. R. E. Nett, D. C. Jones, M. H. Abdille, I. H. Gilbert, A. H.
Fairlamb, ChemMedChem 2009, 4, 1333–1340.
tration of 0.5 mm. Adding the inhibitor caused a small amount of
precipitate to form, which was removed by centrifugation. Both
protein–inhibitor complexes were crystallized by the hanging drop
vapor diffusion method, mixing 2 mL of protein with 2 mL of well
solution (0.1 m HEPES, pH 7.5, 2.0m (NH4)2SO4 for T. brucei TR and
0.1 m TRIS, pH 8.0, 0.2 m CaCl2, 20–22% PEG2000-monomethyleth-
er for T. cruzi TR). The crystals were dipped in well solution, plus
20% glycerol, and flash-frozen to ꢁ1808C for data collection. Dif-
fraction data for the T. brucei TR complex were collected at the
Canadian Macromolecular Crystallography Facility, beamline 08B1-
1 (Canadian Light Source). Those for the T. cruzi TR complex were
collected on a Rigaku MicroMax-007 HF generator run at 40 kV and
30 mA and equipped with Rigaku VariMax multilayer optics and
a
mar345 image plate detector. Data were reduced using
Mosflm[60] and HKL2000.[61] The structures were solved by molecular
replacement (Phaser[62]) using PDB structure 2WOI[45] for T. brucei
TR, and 1BZL[16] for T. cruzi TR as search models. The atomic models
were refined using CNS,[63,64] Refmac,[65] and PHENIX.[66] Data collec-
tion and refinement statistics are listed in Table 3. Atomic coordi-
nates and structure factors were deposited in the PDB with the
access codes 4NEV and 4NEW for the T. brucei and T. cruzi TR inhibi-
tor complexes, respectively.
[20] S. Patterson, D. C. Jones, E. J. Shanks, J. A. Frearson, I. H. Gilbert, P. G.
Wyatt, A. H. Fairlamb, ChemMedChem 2009, 4, 1341–1353.
[21] R. L. Krauth-Siegel, C. Sticherling, I. Jçst, C. T. Walsh, E. F. Pai, W. Kabsch,
C. B. Lantwin, FEBS Lett. 1993, 317, 105–108.
Acknowledgements
[22] C. A. Lipinski, F. Lombardo, B. W. Dominy, P. J. Feeney, Adv. Drug Delivery
Rev. 1997, 23, 3–25.
[23] C. Deleuze-Masquefa, M. Michaud-Dutreilh, J. Vignon, J.-M. Kamenka,
Eur. J. Pharm. Sci. 2000, 9, 345–354.
[24] S. Bonse, C. Santelli-Rouvier, J. Barbe, R. L. Krauth-Siegel, J. Med. Chem.
1999, 42, 5448–5454.
[25] A. Saravanamuthu, T. J. Vickers, C. S. Bond, M. R. Peterson, W. N. Hunter,
A. H. Fairlamb, J. Biol. Chem. 2004, 279, 29493–29500.
[26] P. Baiocco, G. Poce, S. Alfonso, M. Cocozza, G. C. Porretta, G. Colotti, M.
Biava, F. Moraca, M. Botta, V. Yardley, A. Fiorillo, A. Lantella, F. Malatesta,
A. Ilari, ChemMedChem 2013, 8, 1175–1183.
[27] B. Chitkul, M. Bradley, Bioorg. Med. Chem. Lett. 2000, 10, 2367–2369.
[28] M. H. Duyzend, C. T. Clark, S. L. Simmons, W. B. Johnson, A. M. Larson,
A. M. Leconte, A. W. Wills, M. Ginder-Vogel, A. K. Wilhelm, J. A. Czecho-
wicz, D. G. Alberg, J. Enzyme Inhib. Med. Chem. 2012, 27, 784–794.
[29] P. R. Gerber, K. Mꢃller, J. Comput.-Aided Mol. Des. 1995, 9, 251–268.
[30] Residues marked with a prime belong to the second subunit of the ho-
modimeric enzyme.
The authors are grateful to the ETH Research Council for the gen-
erous support of the work done at ETH. Research at the Universi-
ty of Heidelberg was supported by the Deutsche Forschungsge-
meinschaft (LKS), and at the University of Toronto by the Canada
Research Chairs Program and a grant from the National Science
and Engineering Council Canada (EFP). The authors thank Dr.
Bruno Bernet (ETH) for proofreading the Experimental Section, Dr.
Luzi Jakob Barandun (ETH) and Dr. Tristan Reekie (ETH) for proof-
reading the entire manuscript, Dr. Johannes Kirchmair for helpful
discussions, Annie Cunningham for technical support, and Shau-
nivan Labiuk for help with data collection. Research described
herein was performed using beamline 08B1-1 at the Canadian
Light Source, which is supported by the Natural Sciences and En-
gineering Research Council of Canada, the National Research
Council Canada, the Canadian Institutes of Health Research, the
Province of Saskatchewan, Western Economic Diversification
Canada, and the University of Saskatchewan.
[31] A. El-Waer, K. T. Douglas, K. Smith, A. H. Fairlamb, Anal. Biochem. 1991,
198, 212–216.
[32] J. Garforth, J. H. McKie, R. Jaouhari, T. J. Benson, A. H. Fairlamb, K. T.
Douglas, Amino Acids 1994, 6, 295–299.
[33] M. O. F. Khan, S. E. Austin, C. Chan, H. Yin, D. Marks, S. N. Vaghjiani, H.
Kendrick, V. Yardley, S. L. Croft, K. T. Douglas, J. Med. Chem. 2000, 43,
3148–3156.
[34] S. Parveen, M. O. F. Khan, S. E. Austin, S. L. Croft, V. Yardley, P. Rock, K. T.
Douglas, J. Med. Chem. 2005, 48, 8087–8097.
Keywords: docking · inhibitors · mutation studies · structure-
based design · trypanothione reductases · X-ray structures
[35] F. A. de Molfetta, R. F. de Freitas, A. B. F. da Silva, C. A. Montanari, J. Mol.
Model. 2009, 15, 1175–1184.
[36] S. K. Venkatesan, A. K. Shukla, V. K. Dubey, J. Comput. Chem. 2010, 31,
2463–2475.
[37] M. Harder, B. Kuhn, F. Diederich, ChemMedChem 2013, 8, 397–404.
[38] L. M. Salonen, M. C. Holland, P. S. J. Kaib, W. Haap, J. Benz, J. L. Mary, O.
Kuster, W. B. Schweizer, D. W. Banner, F. Diederich, Chem. Eur. J. 2012,
18, 213–222.
[1] P. P. Simarro, A. Diarra, J. A. Ruiz Postigo, J. R. Franco, J. G. Jannin, PLoS
Neglected Trop. Dis. 2011, 5, e1007.
[2] Report of a WHO meeting on elimination of African trypanosomiasis
(Trypanosoma brucei gambiense), World Health Organization, Geneva,
2013.
[3] A. Stich, A. Ponte-Sucre, U. Holzgrabe, Lancet Infect. Dis. 2013, 13, 733–
734.
[4] D. Steverding, Parasites Vectors 2008, 1, 3.
[39] J. Bunzen, J. Iwasa, P. Bonakdarzadeh, E. Numata, K. Rissanen, S. Sato,
M. Fujita, Angew. Chem. Int. Ed. 2012, 51, 3161–3163; Angew. Chem.
2012, 124, 3215–3217.
[40] K. M. M. Doan, J. E. Humphreys, L. O. Webster, S. A. Wring, L. J. Sham-
pine, C. J. Serabjit-Singh, K. K. Adkison, J. W. Polli, J. Pharmacol. Exp.
Ther. 2002, 303, 1029–1037.
[5] Research Priorities for Chagas Disease, Human African Trypanosomiasis
and Leishmaniasis, WHO Technical Report Series No. 975, 2012.
[6] M. P. Barrett, D. W. Boykin, R. Brun, R. R. Tidwell, Br. J. Pharmacol. 2007,
152, 1155–1171.
[7] R. L. Krauth-Siegel, H. Bauer, R. H. Schirmer, Angew. Chem. Int. Ed. 2005,
44, 690–715; Angew. Chem. 2005, 117, 698–724.
[41] A. R. Katritzky, Z. Najzarek, Z. Dega-Szafran, Synthesis 1989, 66–69.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 0000, 00, 1 – 13
&
11
&
ÞÞ
These are not the final page numbers!

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
[42] A. R. Katritzky, H. Yang, S. K. Singh, J. Org. Chem. 2005, 70, 286–290.
[43] For the experimental pKa value of the related phencyclidine, see: M. A.
Iorio, M. Molinari, A. Scott de Carolis, T. Niglio, Farmaco Sci. 1984, 39,
599–611.
[44] E. M. Jacoby, I. Schlichting, C. B. Lantwin, W. Kabsch, R. L. Krauth-Siegel,
Proteins Struct. Funct. Genet. 1996, 24, 73–80.
[56] The PyMOL Molecular Graphics System, Version 1.5.0.4, Schrçdinger LLC,
2010.
[57] H. L. Aalten, G. van Koten, D. M. Grove, T. Kuilman, O. G. Piekstra, L. A.
Hulshof, R. A. Sheldon, Tetrahedron 1989, 45, 5565–5578.
[58] D. C. Jones, A. Ariza, W.-H. A. Chow, S. L. Oza, A. H. Fairlamb, Mol. Bio-
chem. Parasitol. 2010, 169, 12–19.
[45] S. Patterson, M. S. Alphey, D. C. Jones, E. J. Shanks, I. P. Street, J. A. Frear-
son, P. G. Wyatt, I. H. Gilbert, A. H. Fairlamb, J. Med. Chem. 2011, 54,
6514–6530.
[46] G. B. Henderson, N. J. Murgolo, J. Kuriyan, K. Osapay, D. Kominos, A.
Berry, N. S. Scrutton, N. W. Hinchliffe, R. N. Perham, A. Cerami, Proc. Natl.
Acad. Sci. USA 1991, 88, 8769–8773.
[47] G. Jones, P. Willett, R. C. Glen, J. Mol. Biol. 1995, 245, 43–53.
[48] G. Jones, P. Willett, R. C. Glen, A. R. Leach, R. Taylor, J. Mol. Biol. 1997,
267, 727–748.
[59] F. X. Sullivan, C. T. Walsh, Mol. Biochem. Parasitol. 1991, 44, 145–147.
[60] A. G. W. Leslie, H. R. Powell, Processing Diffraction Data with MOSFLM,
Evolving Methods for Macromolecular Crystallography (Eds.: R. Read, J.
Sussman), Springer, Houten, 2007, pp. 41–51.
[61] Z. Otwinowski, W. Minor, Macromolecular Crystallography, Part A, Meth-
ods in Enzymology (Eds.: C. W. Carter, R. M. Sweet, Jr.), Academic Press,
New York, 1997, pp. 307–326.
[62] A. J. McCoy, R. W. Grosse-Kunstleve, P. D. Adams, M. D. Winn, L. C. Storo-
ni, R. J. Read, J. Appl. Crystallogr. 2007, 40, 658–674.
[49] S. F. Sousa, P. A. Fernandes, M. J. Ramos, Proteins Struct. Funct. Bioinf.
2006, 65, 15–26.
[63] A. T. Brꢃnger, P. D. Adams, G. M. Clore, W. L. DeLano, P. Gros, R. W.
Grosse-Kunstleve, J.-S. Jiang, J. Kuszewski, M. Nilges, N. S. Pannu, R. J.
Read, L. M. Rice, T. Simonson, G. L. Warren, Acta Crystallogr. Sect. D
1998, 54, 905–921.
[64] A. T. Brunger, Nat. Protoc. 2007, 2, 2728–2733.
[65] G. N. Murshudov, A. A. Vagin, E. J. Dodson, Acta Crystallogr. Sect. D
1997, 53, 240–255.
[66] P. D. Adams, P. V. Afonine, G. Bunkꢅczi, V. B. Chen, I. W. Davis, N. Echols,
J. J. Headd, L.-W. Hung, G. J. Kapral, R. W. Grosse-Kunstleve, A. J. McCoy,
N. W. Moriarty, R. Oeffner, R. J. Read, D. C. Richardson, J. S. Richardson,
T. C. Terwilliger, P. H. Zwart, Acta Crystallogr. Sect. D 2010, 66, 213–221.
[50] C. H. T. de Paula da Silva, L. S. C. Bernardes, V. B. da Silva, C. L. Zani, I.
Carvalho, J. Biomol. Struct. Dyn. 2012, 29, 1206–1220.
[51] M. Beig, F. Oellien, R. L. Krauth-Siegel, P. M. Selzer, Trypanosomatid Dis-
eases: Molecular Routes to Drug Discovery, Drug Discovery in Infectious
Diseases, Vol. 4, Wiley-VCH, Weinheim, 2013, pp. 383–404.
[52] S. Bailey, K. Smith, A. H. Fairlamb, W. N. Hunter, Eur. J. Biochem. 1993,
213, 67–75.
[53] For an example of a large increase in Ki value upon mutation of a Glu
to Ala, see: Y. Xu, Y. Nakajima, K. Ito, H. Zheng, H. Oyama, U. Heiser, T.
Hoffmann, U.-T. Gꢄrtner, H.-U. Demuth, T. Yoshimoto, J. Mol. Biol. 2008,
375, 708–719.
[54] T. J. Benson, J. H. McKie, J. Garforth, A. Borges, A. H. Fairlamb, K. T.
Douglas, Biochem. J. 1992, 286, 9–11.
Received: February 17, 2014
[55] Y. Zhang, C. S. Bond, S. Bailey, M. L. Cunningham, A. H. Fairlamb, W. N.
Hunter, Protein Sci. 1996, 5, 52–61.
Published online on && &&, 0000
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 0000, 00, 1 – 13
&12
&
ÞÞ
These are not the final page numbers!

FULL PAPERS
Large protein pockets, several binding
conformations: A series of small mole-
cules was evaluated as inhibitors of try-
panothione reductase (TR). A multidi-
mensional approach comprising SAR,
mutation, and docking studies was
taken to investigate their binding mode,
which was confirmed by X-ray co-crystal
structures. Different binding conforma-
tions were revealed, depending on the
parent species of TR.
E. Persch, S. Bryson, N. K. Todoroff,
C. Eberle, J. Thelemann, N. Dirdjaja,
M. Kaiser, M. Weber, H. Derbani, R. Brun,*
G. Schneider,* E. F. Pai,*
R. L. Krauth-Siegel,* F. Diederich*
&& – &&
Binding to Large Enzyme Pockets:
Small-Molecule Inhibitors of
Trypanothione Reductase
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 0000, 00, 1 – 13
&13
&
ÞÞ
These are not the final page numbers!