Synthesis, crystal structures and electronic properties of isomers of chloro-pyridinylvinyl-1H-indoles

Article abstract of DOI:10.1016/j.ejmech.2012.04.033

Three isomers of chloro-3-(2-pyridin-3-ylvinyl)-1H-indole were synthesized and tested as inhibitors of human tryptophan 2,3-dioxygenase (hTDO). The crystal structures of two of them were solved by X-ray diffraction. The solubility of the molecules also was determined experimentally. The molecular electrostatic potentials and dipole moments of the three isomers were calculated by ab initio quantum mechanics (HF/6-311G). The single crystal X-ray analyses reveal non-planar structures. This non-coplanarity is retained during docking of the compounds into a model of hTDO, the molecular target of this series. The position of the Cl atom does not significantly affect the electronic delocalization. Nevertheless, the position of the Cl atom produces a local variation of bond lengths inducing different dipole moments for these isomers. Variations in dipole moments are consistent with the different melting points and crystal packings. Differences in aqueous solubilities are best explained by subtle changes in H-bonds resulting from different accessibilities of the indole NH's due to steric effects of the Cl substituent. The non-coplanarity plays an important role in the crystalline packing of the molecules in contrast to the position of the Cl. This study leads to a better understanding of the structural and electronic characteristics of this chemical series and can potentially help to better understand their inhibitory activity.

Full text of DOI:10.1016/j.ejmech.2012.04.033

European Journal of Medicinal Chemistry 54 (2012) 95e102
Contents lists available at SciVerse ScienceDirect
European Journal of Medicinal Chemistry
journal homepage: http://www.elsevier.com/locate/ejmech
Original article
Synthesis, crystal structures and electronic properties of isomers
of chloro-pyridinylvinyl-1H-indoles
a
,
b
a
a
b
Laurence Moineaux , Sophie Laurent , Jérémy Reniers , Eduard Dolusic , Moreno Galleni ,
ꢀ ꢁ
*
b
a
a
a,
**
Jean-Marie Frère , Bernard Masereel , Raphaël Frédérick , Johan Wouters
Namur Medicine & Drug Innovation Center (NAMEDIC Member of NARILIS), Laboratory of Biological and Structural Chemistry, University of Namur (FUNDP), rue de Bruxelles 61,
B-5000 Namur, Belgium
a
b
CIP, ULg, Liège, Belgium
h i g h l i g h t s
g r a p h i c a l a b s t r a c t
<
<
Three chloro-indolic isomers were
synthesized and evaluated on hTDO.
The single-crystal X-ray analyses of
two of them revealed non-planar
structures.
<
<
<
The position of the Cl atom does not
notably
affect
the
electronic
delocalization.
The non-coplanarity plays a more
important role in the crystalline
packing.
Physico-chemical properties led to
an understanding of mechanism of
TDO inhibition.
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 15 November 2011
Received in revised form
Three isomers of chloro-3-(2-pyridin-3-ylvinyl)-1H-indole were synthesized and tested as inhibitors of
human tryptophan 2,3-dioxygenase (hTDO). The crystal structures of two of them were solved by X-ray
diffraction. The solubility of the molecules also was determined experimentally. The molecular elec-
trostatic potentials and dipole moments of the three isomers were calculated by ab initio quantum
mechanics (HF/6-311G).
30 March 2012
Accepted 24 April 2012
Available online 3 May 2012
The single crystal X-ray analyses reveal non-planar structures. This non-coplanarity is retained during
docking of the compounds into a model of hTDO, the molecular target of this series. The position of the Cl
atom does not significantly affect the electronic delocalization. Nevertheless, the position of the Cl atom
produces a local variation of bond lengths inducing different dipole moments for these isomers. Varia-
tions in dipole moments are consistent with the different melting points and crystal packings. Differ-
ences in aqueous solubilities are best explained by subtle changes in H-bonds resulting from different
accessibilities of the indole NH’s due to steric effects of the Cl substituent. The non-coplanarity plays an
important role in the crystalline packing of the molecules in contrast to the position of the Cl. This study
leads to a better understanding of the structural and electronic characteristics of this chemical series and
can potentially help to better understand their inhibitory activity.
Keywords:
Chloro-3-(2-pyridin-3-ylvinyl)-1H-indole
Human tryptophan 2,3-dioxygenase
Physico-chemical properties
Crystal structure
Isomer
Ó 2012 Elsevier Masson SAS. All rights reserved.
1
. Introduction
*
Corresponding author. Tel.: þ32(0)81724569.
* Corresponding author.
E-mail addresses:
johan.wouters@fundp.ac.be (J. Wouters).
*
The essential amino acid
L-tryptophan (L-Trp) is required for the
Laurence.moineaux@fundp.ac.be
(L.
Moineaux),
biosynthesis of proteins and is a precursor for several biologically
0223-5234/$ e see front matter Ó 2012 Elsevier Masson SAS. All rights reserved.
doi:10.1016/j.ejmech.2012.04.033

96
L. Moineaux et al. / European Journal of Medicinal Chemistry 54 (2012) 95e102
important compounds, including 5-hydroxytryptamine (5-HT,
serotonin), kynurenic acid or nicotinamide [1e3]. Tryptophan
hydroxylase controls 5-HT synthesis in the brain. In fact, this
enzyme has a large capacity for processing its substrate, trypto-
phan, and it has been shown that higher brain tryptophan levels
cause an increase in the 5-HT levels in both whole-brain and brain
709W92, Fig. 1) [5,34,35] belonging to the 3-(2-(pyridyl)ethenyl)
indole class described by the Wellcome group as combined TDO/5-
HT (serotonin) reuptake inhibitors for antidepressant therapy
seemed to be the most interesting hits. 680C91 is an excellent
in vitro inhibitor (K
i
of around 30 nM) on liver-extracted TDO, and at
10 M, has no activity on 5-HT reuptake, various 5-HT receptors,
m
extracellular fluid. Indeed,
antidepressant but its efficacy is weak because of its rapid catab-
olism [4].
L
-tryptophan may be used as an
IDO and monoamine oxidases A and B [35]. However, in vivo, the
compound has no pronounced effect on tryptophan levels in rat,
presumably because of bioavailability/metabolism problems
[31,36,37].
The
L-kynurenine pathway is the major catabolic route of
L
-tryptophan metabolism and the key controlling enzymes are
As part of our interest in the design and synthesis of original
TDO inhibitors [38], we describe in this paper the structure and
physico-chemical properties of three isomers of chloro-pyr-
idinylvinyl-1H-indole, confirm they inhibit TDO and propose
a rationale for this inhibition by docking simulations.
tryptophan 2,3-dioxygenase (TDO) and indoleamine 2,3-
dioxygenase (IDO). Both are cytosolic heme dioxygenases that
catalyze the oxidative cleavage of the C
2 3
eC bond of the indole ring
of -tryptophan. This reaction is the first and rate-limiting step of
L
the kynurenine pathway of tryptophan catabolism, which eventu-
þ
ally leads to the formation of nicotinamide dinucleotide (NAD ),
2. Experimental section
a process regarded as the primary biological function of TDO [5e8].
Although hTDO and hIDO catalyze the same reaction and share
relatively conserved active site regions, amino acid sequences do
not exhibit more than a 10% level of identity [9], their physiological
functions are distinct. TDO is homotetrameric and almost exclu-
2
.1. Single X-ray diffraction
Crystals of all three ((E)-chloro-3-(2-pyridin-3-ylvinyl)-1
H-indoles) were obtained by slow evaporation of solutions of the
samples at room temperature. Only crystals of (1) and (3) allowed
good quality structure determination by crystallography. X-ray
measurements were performed on a Gemini Ultra R system
sively found in the liver [10]. It is highly specific for L-tryptophan
and some of its derivatives substituted in the 5- and 6-positions of
the indole ring [11]. On the other hand, IDO is monomeric and
extrahepatic and shows activity toward a larger collection of
substrates, including serotonin, tryptamine, 5-hydroxytryptophan,
and melatonin. The two enzymes also have different inducers; IDO
is induced by inflammatory stimuli such as interferon-g, and TDO
by tryptophan, glucocorticoids, and kynurenine [7,12].
Kynurenine pathway metabolites have been implicated in
a number of diseases ranging from neurological disorders, such as
cerebral malaria and multiple sclerosis, to cataract formation
(
(
4-circle kappa platform, Ruby CCD detector) using Mo Kl radiation
l
¼ 0.71073 Å). After mounting and centering the single-crystal on
the diffractometer, cell parameters were estimated from a pre-
experiment run and full data sets were collected at room
temperature. Structures were solved by direct methods with the
SHELXS-97 program and then refined on F using the SHELXL-97
software [39]. E-map provided positions for all non H-atoms. The
full-matrix least-squares refinement was carried out on F using
anisotropic temperature factors for all non H-atoms. The H-atoms
were located from DF-maps, and then their positions were refined
using a riding model with isotropic thermal parameters taken as 1.2
times temperature factors for their parent-atoms. The ORTEPs of
these isomers were obtained by the PLATON [40] program.
The quality of the crystallographic data for 6-chloro-3-(2-
pyridin-3-ylvinyl)-1H-indole was too poor to allow a good struc-
ture determination.
Crystal data for (1): crystallization by slow evaporation from
a tolueneemethyl acetate solution (1:1). Crystal dimensions:
0
2
2
[13,14]. Recent findings have implicated IDO-mediated tryptophan
catabolism in immune tolerance, including immune suppression in
maternal fetal tolerance and the immune escape of cancers
[15e20].
While IDO is a well-established target for cancer immuno-
therapy [21e28], it was only recently reported that TDO is also
expressed in many tumor cells including melanoma, colorectal,
bladder, hepatic, breast and lung cancers in a murine model of
cancer [29]. This work demonstrated that TDO expression in
tumors has an effect similar to the expression of IDO, preventing
tumor rejection by locally degrading tryptophan. A recent study
confirmed another important mechanism of TDO-mediated
tumoral immune suppression, namely the release of the
tryptophan catabolite kynurenine, which then binds to the aryl
hydrocarbon receptor (AHR) [30]. Recently, we contributed to
a study describing a small-molecule TDO inhibitor which enables
reversal of tumoral immune resistance in vivo [31]. These findings
underline the importance of the precise TDO catalytic mechanism
and fully justify the search for potent inhibitors.
,20 ꢀ 0,10 ꢀ 0,28 mm. Monoclinic C2/c, a ¼ 15.784(2) Å,
b ¼ 9.352(1) Å, c ¼ 17.080(2) Å, V ¼ 2492.3(5) Å, Z ¼ 8,
3
ꢁ1
D
calc ¼ 1.3576 g/cm ,
m
(Mo K
l
) ¼ 0.288 mm , F(000) ¼ 1056,
ꢂ
ꢂ
q
min ¼ 3.4 to
q
max ¼ 27.9 , R ¼ 0.0344, wR2 ¼ 0.072, observed data
(I > 2
s
I) ¼ 1257, total ¼ 2599, R(int) ¼ 0.035, S ¼ 0.793.
Crystal data for (3): crystallization by slow evaporation from
tolueneeethyl acetate solution (1:1). Crystal dimensions:
,20 ꢀ 0,40 ꢀ 0,35 mm. Triclinic Pꢁ1, a ¼ 11.081(1) Å, b ¼ 13.128(1) Å,
a
0
c ¼ 14.863(1) Å, V ¼ 1862.9(3) Å, Z ¼ 6 (3 molecules in the a.u.),
3
ꢁ1
Until recently, only a handful of compounds based on the indole
D
calc ¼ 1.3623 g/cm ,
m
(Mo K
l
) ¼ 0.289 mm , F(000) ¼ 792,
or
b-carboline scaffolds were reported as exhibiting TDO inhibitory
ꢂ
ꢂ
q
min ¼ 3.3 to
q
max ¼ 32.6 , R ¼ 0.0636, wR2 ¼ 0.119, observed data
activity [5,32e35]. Among these, fluoroindoles (680C91 and
(I > 2
s
I) ¼ 4184, total ¼ 12,288, R(int) ¼ 0.05, S ¼ 0.845.
Fig. 1. a, Structures of the leads. 680C91: R ¼ pyrid-3-yl; 709W92: R ¼ pyrid-4-yl; b, general synthetic procedure for chloro-pyridinylvinlyl-1H-indoles. (1), 5-Cl; (2), 6-Cl; (3), 7-Cl-
isomer.

L. Moineaux et al. / European Journal of Medicinal Chemistry 54 (2012) 95e102
97
2
.2. Computational methods
GraphPadPrism 5.0 program [43]. All determinations were per-
formed in duplicate and the presented data are average values.
The solutions and dilutions of the different inhibitors were
prepared in 100% DMSO. The final concentration of DMSO in the
assay mixture was 5% and it was verified that this concentration did
not affect the enzymatic activity. The enzyme studied is
a recombinant form of human tryptophan 2,3-dioxygenase (hTDO)
produced in Escherichia coli and purified to homogeneity.
In order to model the structure, geometry optimization was
performed at the rhf/6-311G level of theory using the Gaussian suite
of programs [41]. Molecular electrostatic potential (MEP), suppress
HOMOeLUMO topologies and energies, and dipole moment were
computed. Gaussview allowed the visualization of dipole moments,
suppress HOMOeLUMO topologies and MEPs.
2.3. Solubility evaluation
2.5. Molecular modeling
The solubility was evaluated using Multiscreen HTS-PCF filter
Sequence alignment between human TDO (obtained from the
Swiss-Prot database) [44] and Ralstonia metallidurans (rmTDO (PDB
code 2NOX)) [45] was performed using BLASTP [46e48]. A model of
human TDO was thus developed by in silico mutation of only a few
amino acids from rmTDO. To take into account protein flexibility,
the resulting model was minimized using the MINIMIZE module
included in SYBYL 8.0 program [49].
plates from Millipore according to a protocol adapted from the
Millipore application note (“Quantitative method to determine
drug aqueous solubility: optimization and correlation to standard
methods” Thomas Onofrey and Greg Kazan http://www.millipore.
com/techpublications/tech1/an1730en00) at pH 1.0 (HCl 0.1 M,
NaCl 20 mM), 7.4 (phosphate buffer 50 mM, NaCl 30 mM) and 9.0
(
boric acid 0.1 M, NaCl 20 mM) and with a final DMSO concentra-
Molecular modeling studies were carried out on a Linux work-
station. The compounds were built using the SKETCH module
implemented in SYBYL (version 8.0) [49]. Docking was performed
using the 3D coordinates of our humanized TDO model with the
help of the automated GOLD program [50] (active site definition:
residues within 10 Å around heme). In order to take protein flexi-
bility into account, the enzymeeinhibitor complexes were opti-
mized using the MINIMIZE module. The minimization process uses
the Powell method with the Tripos force field (dielectric constant
tion of 2.5%. The filtrate was analyzed using an Agilent 1100 Series
LC/MSD Trap system with UV detection at 254 nm equipped with
a C18 Zorbax SB column (100 ꢀ 3 mm; 3.5
mm) and following
ꢁ1
a gradient (flow rate: 0.5 mL min ) of acetonitrile in aqueous
acetic acid 0.1% (v/v) (gradient: from 5 to 95% of acetonitrile in
5
0
min, holding for 3 min, then reversing to 5% of acetonitrile in
.1 min and holding for an additional 5.4 min).
ꢁ1
1r) to reach a final convergence of 0.01 kcal mol
.
2
.4. IC50 determination on hTDO
3. Results and discussion
Enzymatic assays were performed by measuring the rate of
kynurenine formation by a discontinuous colorimetric method. The
reactions were performed by adding 1.3 M (100 l) purified
human TDO (hTDO) in a reaction mixture (900 l) containing
0 mM tris buffer pH 8.0, 20 mM ascorbic acid, 100 g of Bovine
Liver catalase, 5 M hemin, 10 M methylene blue, 100 -tryp-
3
.1. Compound synthesis
m
m
m
The compounds were synthesized according to the procedure
5
m
described in Fig. 1 and detailed elsewhere [38].
m
m
mM L
tophan and different concentrations of potential inhibitors in
ꢂ
a range of 0e100
m
M. The mixture was pre-incubated at 37 C for
3.2. IC50 determination on human TDO
10 min before addition of the enzyme. The reactions were stopped
after 5,10,15 or 20 min by adding 40
00
m
l of 30% trichloroacetic acid to
IC50 values were determined for two isomers of chloro-3-(2-
pyridin-3-ylvinyl)-1H-indole on the purified human TDO. For
ꢂ
2
m
l of the reaction mixture and further incubated at 65 C for
1
5 min to hydrolyze N-formylkynurenine into kynurenine. After
centrifugation at 19,300 g for 10 min, 100 l of supernatant were
added to 100 l of para-dimethylaminobenzaldehyde (Ehrlich’s
isomers (1) and (2), the IC50 values were 10 ꢃ 0.5
5 ꢃ 0.4 M respectively (Fig. 2).
For isomer (3), concentrations higher than 50
obtained because of solubility problems so that it was not possible
to determine an IC50 value by the same method. Thus, a K value was
mM and
m
m
m
mM could not be
reagent) in a 96-well plate. A positive reaction yields a yellow
solution [42]. The amount of formed kynurenine was determined
spectrophotometrically at 480 nm. Data were analyzed with the
i
determined by varying the substrate and inhibitor concentrations
a 100
b 100
80
60
40
20
0
80
60
40
20
0
1
2
3
4
5
2
3
4
5
6
Fig. 2. Residual activity (%) of purified human TDO in presence of isomers (1) (a) and (2) b. The enzymatic assays were performed in duplicate as described in the Experimental
section.

98
L. Moineaux et al. / European Journal of Medicinal Chemistry 54 (2012) 95e102
ꢂ
ꢂ
ꢂ
ꢂ
and fitting the data to the HenrieMichaeliseMenten equation
Fig. 3). This study showed that isomer (3) behaved as a competitive
inhibitor with a K M. On the basis of the same
value of 21 ꢃ 0.3
data, the K M, in good agree-
value was evaluated at 180 ꢃ 1.4
46.49(7) (1), 22.23(7) (3a), 19.01(5) (3b), and 17.17(5) (3c).
Table 1S gives values of selected bond lengths for the two isomers.
These geometries obtained by crystallography (crystal packing
effects) are retained in structures optimized by ab initio methods
(rhf/6-311G) (Table 1S, supplementary data and tables for this
paper are available from supplementary materials). Influence of the
position of the Cl atom in the two isomers is essentially limited to
the C(4)eC(5), C(5)eC(6) and C(6)eC(7) bond lengths. Optimiza-
tion of (3), starting from the three distinct conformers (a, b, c)
observed in the crystal structure converged to a unique structure.
(
i
m
m
m
ment with the value determined in the absence of inhibitor.
To compare with the other two isomers, an IC50 value was
calculated with the help of the Dixon equation [51] for competitive
inhibition (Equation (1)).
ꢀ
ꢁ
s
IC50 ¼ Ki 1 þ K
(1)
m
3
.4. Crystal packing and intermolecular interactions
At a substrate concentration of 100
mM, the calculated IC50 is
3
2 ꢃ 0.02
mM. The inhibitory activity is thus lower than that of the
The crystal packings of (1) and (3a, b, c) are essentially deter-
other two isomers. This is consistent with the docking analysis (see
Sec. 3.6).
The three isomers inhibited hTDO and the best inhibitory
activity was obtained with the chlorine atom in position 5 on the
indole ring (1). However, these isomers of the chloro-pyr-
idinylvinyl-1H-indole family do not exhibit better inhibitory
mined by two factors: p-stacking and hydrogen bonds (HB).
0
For both isomers (1) and (3), a strong HB links nitrogen (N3 ) of
pyridine (hydrogen bond acceptor, A) with the (N1eH1) of the
indole ring (H-bond donor, DH) of a symmetry-related molecule
(
Fig. 5). The isomers (3a) and (3b) exhibit a weak HB between
0
nitrogen (N3 ) of the pyridine and the (C1eH1) of the indole ring
activities than the reference inhibitor 680C91 (IC50 near 1.3 mM).
[
53,54]. The total number (one DeH and one A per molecule) and
strength (reflected by d, Table 2S, [52]) of these H-bonds is similar
for both isomers. Geometrical parameters used to describe the
stacking interactions are defined in Table 3S which lists the values
of these parameters for our 2 crystal structures.
The crystal packing modes of compounds (1) and (3) are given in
Fig. 5. In the packing of compound (1), the 6-membered ring of the
indole interacts with the 5-membered ring of another indole
3
.3. Molecular conformation
Only a very limited number of substituted vinyl-1H-indoles (e.g.
0
3
-(2 -nitrovinyl)indole (CSD code: GOJDUE), ethyl 3-(1H-indol-3-
yl)acrylate (CSD code: GUGLEA), 3-(3-indolyl)acrylamide (CSD
code: SUVYIR), or (E)-4-(3-indolylvinyl)-N-methylpyridinium
iodide (CSD code: TIKBEU and TIKBIY)) crystal structures have been
determined by crystallography. Efforts to obtain crystals of fluoro-
substituted 3-(2-pyridin-3-ylvinyl)-1H-indole isomers, including
moiety by
pep stacking interactions. There is an additional pep
interaction between the pyridine group and the indole ring. In
contrast, for compound (3), only the 6-membered rings of the
indoles interact.
6
80C91 proved unsuccessful in our hands. In contrast, crystals of
chloro-substituted 3-(2-pyridin-3-ylvinyl)-1H-indoles could be
grown providing valuable structural information on this family of
compounds both in terms of in depth analysis of molecular
conformations and crystal packings but also serving as starting
point for geometry optimizations.
Fig. 4 shows the ORTEP diagram of 5- and 7-chloro-3-(2-
pyridin-3-ylvinyl)-1H-indole isomers ((1) and (3), respectively)
with the atom numbering scheme. There are three molecules in the
asymmetric unit of (3) in contrast to (1).
It is tempting to compare the stabilizing effects of crystal
packing with experimental physico-chemical parameters of the
compounds, in particular melting points (mp) [55] and solubility.
Therefore, melting points were determined. The solubility was
evaluated using Multiscreen HTS-PCF filter plates from Millipore
(Table 1).
Indeed, several H-bond interactions in (3) are similar to those in
compound (1) but analysis of the crystal packing suggests that pep
interactions are stronger for (1). This is consistent with the melting
Both compounds possess an E configuration at the C(10)]C(11)
double bond. Unexpectedly, this study reveals a significant devia-
tion from coplanarity. Indeed, the values of the dihedral angle
between the plane of the pyridine and that of the indole are:
points of these compounds. Compound (1), in which there is
ꢂ
a better stabilization by
p
-stacking, has a higher mp (222e223 C)
ꢂ
than compound (3) (204e205 C) with weaker stacking interac-
tions and similar H-bonds. It can be anticipated that the crystal
packing of (2) would be closer to that of (3) than to that of (1).
Aqueous solubility of the compounds (Table 1) is related to both
the stability of the solid state packing and solvation effects. If one
refers to the stability of the solid solutes (similar H-bonds and
differences in
p-interactions) used to explain the experimental
melting points, the solubilities of (1) and (3) cannot be easily
explained. Indeed, compound (1), with the stronger packing (and
higher mp), is also more soluble that (3) (23 mg/mL for (1)
compared to 2.3 mg/mL for (3) at pH 7.4). The lower solubilities (at
pH 7.4 and 9.0) must thus be explained by solvation effects. As
discussed further (point 3.5.2 MEP) the position of the chlorine
atom close to the indole NH limits access to the H-bond donor and
significantly stabilizes interactions with water molecules. This
steric effect of the Cl atom, which affects aqueous solubility
(mediated by H-bonding to the indole NH) could also explain, in
part, the lower affinity of (3) for TDO.
The solubility of this series of compounds is higher at low pH (a
2
0 fold increase of solubility of (1) between pH 7.4 and pH 1). This is
Fig. 3. Initial rate of L-Trp oxidation by purified hTDO in the presence of varying
concentrations of 7-chloro-3-(2-pyridin-3-ylvinyl)-1H-indole (3) and L-Trip. The
enzymatic assays were performed as described in the Experimental section.
most probably explained by protonation of the pyridine ring at low
pH values.

L. Moineaux et al. / European Journal of Medicinal Chemistry 54 (2012) 95e102
99
Fig. 4. View of the crystal structure conformations (ORTEP diagram, 50% probability level) for compounds (1) [top] and (3) [bottom]. Numbering is the same for all the molecules.
There are three molecules (a, b, c) in the asymmetric unit of (3).
3
3
.5. Electronic properties
Table 1). Compounds (1) and (2) exhibit a molecular dipole
moments larger than that of compound (3).
.5.1. Dipole moments
This is in agreement with similar H-bond patterns (in the crystal
The molecular dipole moment is a measure of charge distribu-
packing) but distinct orientations of the molecule (due to the
stacking of the rings) in the crystal.
pep
tion in a molecule. The accuracy of the overall distribution of
electrons in a molecule is difficult to quantify, since it involves all
the multipoles. From the present calculations, the total dipole
moment and orientation vectors calculated with the geometries
optimized at the rhf/6-311G level of theory are given in Table 1.
The variations in the dipole moments are consistent with the
different melting points of the solids and to some extent, with their
solubility. Correlation with the solubility is less clear, as already
noted by others [56].
3.5.2. Electrostatic potential around isolated molecules
In order to calculate the molecular electrostatic potentials (MEP)
of these molecules, the geometry of the system was first optimized
at the rhf/6-311G level of theory. Secondly, mapping of the MEP
onto the molecular surface was performed with Gaussview. The
MEP maps (Fig. 4b) reveal that essentially all three non charged
molecules present similar topologies and charge distributions. All
three compounds are characterized by attractive (negative)
potentials (redeyellow) on the N atom of the pyridine and a posi-
tive potential (blue) on the NeH function of the indole ring. The Cl
The influence of the position of the Cl substituent on the
distribution of atomic charges is limited but nevertheless leads to
significant variations in the molecular dipole moments (Fig. 6a,

100
L. Moineaux et al. / European Journal of Medicinal Chemistry 54 (2012) 95e102
Fig. 5. Views of the crystal packing of (1) [top] and (2) [bottom]. Two selected orientations of the cell are presented for each compound in order to better visualize both hydrogen
bonds (highlighted by dotted lines) and -stacking interactions.
p
atom also contributes and generates some attractive potential but
to a much lesser extent than the nitrogen of the pyridine.
This topology is in agreement with the H-bonds determined in
3.6. Modeling and docking
The biological data and stereo-electronic properties derived
from our study of the title compounds were combined with dock-
ing simulations into the active site of TDO and proved useful to
better understand the TDO inhibitory activity of those molecules.
To date, only the 3-D coordinates of TDO from R. metallidurans
(rmTDO) (PDB code: 2NOX) [45] and Xanthomonas Campestris
(xcTDO) [9] (PDB code: 1YW0) have been experimentally obtained.
These two bacterial TDOs present high sequence similarities with
the human form (28e34%) and excellent identities in the catalytic
site area. Sequences of hTDO and rmTDO were aligned using the
BLASTP program [46e48], showing that amino acids essential to
the activity were well preserved around the heme cofactor. Docking
of the three isomers in the active site was performed in the struc-
ture of humanized rmTDO (i.e.: hTDO residues were mutated in the
structure of rmTDO) using the GOLD program [50]. The results are
summarized in Fig. 7.
0
the crystallographic structure. Indeed, the pyridine nitrogen (N3 ) is
an H-bond acceptor which corresponds to the basic (negative) site
in MEP and the (N1eH1) of the indole group is an H-bond donor
which corresponds to the most positive region of the MEP. Inter-
estingly, as already underlined in the discussion of aqueous solu-
bility of our series, the relative position of the chlorine substituent
has a direct effect (steric hindrance and ‘shading’) on the MEP and
in consequence on the ability of the three isomers to form H-bonds.
Table 1
Physico-chemical parameters of (1), (2) and (3): experimental properties (melting
points and solubilities determined at 3 pH’s) and calculated (rhf/6-311G) electronic
properties (molecular dipole moment).
Compound
(1)
(2)
(3)
Solubility (mg/mL)
pH 1
pH 7.4
The geometries of the three isomers in the docking solutions are
consistent with the crystallographic structures. Indeed, the non-
coplanarity, and the trans configuration observed in the crystallo-
graphic study of these compounds is retained in the docking.
540.3
23.0
76.0
7.41
473.3
22.3
32.1
5.69
550.8
2.3
6.0
pH 9
Molecular dipole
moment (D)
Orientation vector of
dipolar moment (x,y,z)
3.15
Moreover, the essential binding points (HB interactions and
p-
ꢁ1.61; ꢁ7.18;
0.86
222e223
0.43; 5.61;
0.80
204e208
0.32; ꢁ2.93;
1.12
204e205
stacking interactions) are also observed between the inhibitors and
the active site residues of TDO, for example, the formation of H-
bonds between the eNH of indole and His 72 and between the
ꢂ
Melting point ( C)

L. Moineaux et al. / European Journal of Medicinal Chemistry 54 (2012) 95e102
101
Fig. 6. Representation of selected electronic properties calculated (rhf/6-311G) on optimized geometries of (1) [left], (2) [middle] and (3) [right]: molecular dipole moments (a, top)
and MEP contours (range between 179 and 179 kcal/mol, b, bottom).
nitrogen of pyridine and Thr 271. These hydrogen bonds are similar
to those observed in the crystallographic study performed on the
two (isolated) isomeric chloro-3-(2-pyridin-3-ylvinyl)-1H-indoles.
These molecules interact in the same way as the reference inhibitor
(680C91) [38].
Based on the docking simulations, it is tempting to suggest that
the reduced inhibition of TDO by compound (3) is due to a signifi-
cant displacement of this inhibitor, compared to molecules (1) and
(2) which overlap well. This displacement, induced by steric effects
caused by the Cl atom, could result in a less favorable HB interaction
with His 72. In addition, the significantly reduced value of the
dipole moment of (3) versus (1) and (2) could also explain weaker
pe
p
interactions within the active site of TDO and thus a reduced
affinity for the enzyme. This is consistent with the IC50 analysis.
4. Conclusions
In conclusion, the TDO inhibition potencies of three previously
synthesized isomers of chloro-3-(2-pyridin-3-ylvinyl)-1H-indole
were evaluated. The single-crystal X-ray analyses of two of them
revealed non-planar structures. This non-coplanarity is retained
during docking of the compounds into the active site of a model of
human TDO. The position of the Cl atom does not significantly affect
the electronic delocalization within the molecule but allows local
bond length variations and overall differences for the value and
orientation of the dipole moments. Variations in dipole moments
are consistent with different melting points and crystal packings.
Differences in aqueous solubilities are best explained by subtle
Fig. 7. (a, top), Sequence alignment used for modeling of hTDO: residues of catalytic
site area (4 Å, underlined), conserved residue (pink), similar residues (yellow) and
different residues (blue) and (b, bottom) docked structure of (1) (blue), (2) (green) and
(
3) (pink) into the active site of rmTDO.

1
02
L. Moineaux et al. / European Journal of Medicinal Chemistry 54 (2012) 95e102
[4] D.J. Madge, et al., Biochem. Pharmacol. 49 (1995) 1435e1442.
changes in H-bonds resulting from different accessibilities of the
indole NH due to the steric effects of the Cl substituent. Similarly,
differences in H-bonding to the His 72 residue of the active site of
TDO could, in part, explain the affinities that were measured. This
study also leads to a better understanding of the mechanism of TDO
inhibition by this chemical series from the structural and electronic
points of view and the results obtained with the chloro-(2-pyridin-
[
[
[
5] D.J. Madge, et al., Bioorg. Med. Chem. Lett. 6 (1996) 857e860.
6] S.A. Rafice, et al., Biochem. Soc. Trans. 37 (2009) 408e412.
7] M. Sono, et al., Chem. Rev. 96 (1996) 2841e2888.
[8] T.W. Stone, et al., Nat. Rev. Drug Discovery 1 (2002) 609e620.
9] F. Forouhar, et al., PNAS 104 (2007) 473e478.
[
[
[
10] Y. Kotake, et al., Physiol. Chem. 243 (1936) 237e244.
11] J.M. Leeds, et al., J. Biol. Chem. 268 (1993) 17781e17786.
[12] J.B. Katz, et al., Immunol. Rev. 222 (2008) 206e221.
[
[
[
[
[
[
[
13] O. Takikawa, et al., Adv. Exp. Med. Biol. 527 (2003) 277e285.
14] L.A. Sanni, et al., Am. J. Pathol. 152 (1998) 611e619.
15] A.L. Mellor, et al., Immunol. 13 (2001) 213e218.
16] U. Grohmann, et al., Trends Immunol. 24 (2003) 242e248.
17] D.H. Munn, et al., J. Clin. Invest. 117 (2007) 1147e1154.
18] A.L. Mellor, et al., Nat. Rev. Immunol. 4 (2004) 762e774.
19] S.J. Thackray, et al., Biochemistry 47 (2008) 10677e10684.
3
-ylvinyl)-1H-indoles must now be extrapolated to the fluoro
analogs of 680C91.
Author’s contribution
LM wrote most of the manuscript and performed the structural
[20] C. Uyttenhove, et al., J. Nat. Med. 9 (2003) 1259e1274.
[
[
[
21] S. Löb, et al., Nat. Rev. Cancer 9 (2009) 445e452.
22] A.J. Muller, et al., Nat. Med. 11 (2005) 312e319.
23] A.J. Muller, et al., Expert Opin. Ther. Targets 9 (2005) 831e849.
(Crystallography, molecular modeling) and electronic analyses (dipolar
moment, MEP). She also contributed to the determination of the
solubilities of compounds. SL produced hTDO (overexpression and
purification) and performed the enzymatic assays and IC50 determi-
nation. JR and ED performed the synthesis and chemical character-
ization of the compounds. MG and JMF supervised the enzymatic work.
BM, RF and JW supervised and coordinated the chemical work.
[24] A.J. Muller, et al., Curr. Cancer Drug Targets 7 (2007) 31e40.
[25] A. Macchiarulo, Amino Acids 37 (2009) 219e229.
[
[
[
26] U. Röhrig, et al., J. Med. Chem. 53 (2010) 1172e1189.
27] E. Dolu ꢀs i ꢁc , et al., Bioorg. Med. Chem. 19 (2011) 1550e1561.
28] E. Dolusic, et al., Eur. J. Med. Chem. 46 (2011) 3058e3065.
ꢀ ꢁ
[29] B. Van den Eynde, et al., WO2010008427 (2010).
30] C.A. Opitz, et al., Nature 478 (2011) 197e204.
[31] L. Pilotte, et al., Proc. Natl. Acad. Sci. 109 (2012) 2497e2502.
[
Conflict of interest
[
[
[
[
[
[
32] M. Civen, et al., J. Biol. Chem. 235 (1960) 1716e1718.
33] N. Eguchi, et al., Arch. Biochem. Biophys. 232 (1984) 602e609.
34] M. Salter, et al., Biochem. Pharmacol. 49 (1995) 1435e1442.
35] M. Salter, et al., Neuropharmacology 34 (1995) 217e227.
36] M. Salter, et al., Bioorg. Med. Chem. Lett. 6 (1996) 857e860.
37] S.N. Young, Raven Press, New York, 1986, 49.
None.
Acknowledgments
[38] E. Dolu
ꢀs i ꢁc , et al., J. Med. Chem. 54 (2011) 5320e5334.
The authors thank Jérémy Maury who participated in the
synthesis and crystallization of the compounds used in this study
and Bernadette Norberg for assistance in the crystallographic part
of the work. Financial contribution from the Fonds de la Recherche
Scientifique-FNRS for the award of a Télévie grant to L. M.
[39] G.M. Sheldrick, SHELX-97: Program for the Solution of Crystal Structures,
University of Göttingen, Göttingen (Germany), 1998.
[40] A.L. Speck, University of Utrecht, Utrecht, The Netherlands, 2001.
[41] M.J. Frisch, et al., Gaussian 03. Revision B.04, Gaussian. Inc., Wallingford CT,
004.
2
[
[
[
[
[
[
42] E. Alegre, et al., Anal. Biochem. 339 (2005) 188e189.
43] GraphPadPrism. Version 5.03, GraphPad Software, SanDiego CA, 2003.
44] H.M. Berman, et al., Nat. Struct. Biol. 10 (2003) 980.
45] Y. Zhang, et al., Biochem. 46 (2007) 145e155.
46] C.L. Pierri, et al., Biochim. Biophys. Acta 1804 (2010) 1695e1712.
47] S.F. Altschul, et al., Nucleic Acids Res. 25 (1997) 3389e3402.
(
7.4.543.07F) is greatly appreciated.
Appendix A. Supplementary data
Supplementary data related to this article can be found online at
doi:10.1016/j.ejmech.2012.04.033.
[48] R. Lopez, et al., Nucleic Acids Res. 31 (2003) 3795e3798.
[
[
[
49] SYBYL, Version 8.0, Tripos Inc., St. Louis, U.S.
50] G. Jones, et al., J. Mol. Biol. 267 (1997) 727e748.
51] M. Dixon, et al., Biochem. J. 55 (1952) 170e171.
References
[52] G.R. Desiraju, et al., in: , IUCr. Monographs on Crystallography, 9, 2001, p. 5.
53] B.G. McGaughey, et al., J. Biol. Chem. 273 (1998) 15458e15463.
[54] J.P. Gallivan, et al., Proc. Natl. Acad. Sci. 96 (1999) 9459e9464.
nꢁ ska, et al., Pol. J. Chem. 64 (1990) 813e817.
[
[1] N.P. Botting, Chem. Soc. Rev. 24 (1995) 401e412.
[2] R. Schwarcz, Curr. Opin. Pharmacol. 4 (2004) 12e17.
[3] K. Schröcksnadel, et al., Clin. Chim. Acta 364 (2006) 82e90.
[55] E. Salwi
[56] A.R. Katritzky, et al., Cryst. Growth Des. 1 (2001) 261e265.