Docking study, synthesis, and in vitro evaluation of fluoro-MADAM derivatives as SERT ligands for PET imaging

Article abstract of DOI:10.1016/j.bmc.2008.08.002

In order to predict affinity of new diphenylsulfides for the serotonin transporter (SERT), a molecular modeling model was used to compare potential binding affinity of new compounds with known potent ligands. The aim of this study is to identify a suitable PET radioligand for imaging the SERT, new derivatives, and their precursors for a C-11 or F-18 radiolabeling, were synthesized. Two fluorinated derivatives displayed good in vitro affinity for the SERT (K_i = 14.3 ± 1 and 10.1 ± 2.7 nM) and good selectivity toward the other monoamine transporters as predicted by the docking study.

Full text of DOI:10.1016/j.bmc.2008.08.002

Bioorganic & Medicinal Chemistry 16 (2008) 9050–9055
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Docking study, synthesis, and in vitro evaluation of fluoro-MADAM derivatives
as SERT ligands for PET imaging
Sylvie Mavel a,b, , Johnny Vercouillie a,b, Lucette Garreau a,b, Tiziana Raguza a,b,c
,
*
Aina Westrheim Ravna d, Sylvie Chalon a,b, Denis Guilloteau a,b,c, Patrick Emond a,b
a INSERM U930, 37000 TOURS, France
b Université François-Rabelais de Tours, 37000 TOURS, France
c CHRU, Hôpital Bretonneau, Service de Médecine Nucléaire, Tours, France
d Department of Pharmacology, Institute of Medical Biology, University of Tromso, Norway
a r t i c l e i n f o
a b s t r a c t
Article history:
In order to predict affinity of new diphenylsulfides for the serotonin transporter (SERT), a molecular mod-
eling model was used to compare potential binding affinity of new compounds with known potent
ligands. The aim of this study is to identify a suitable PET radioligand for imaging the SERT, new deriv-
atives, and their precursors for a C-11 or F-18 radiolabeling, were synthesized. Two fluorinated deriva-
tives displayed good in vitro affinity for the SERT (Ki = 14.3 1 and 10.1 2.7 nM) and good selectivity
toward the other monoamine transporters as predicted by the docking study.
Received 10 March 2008
Revised 25 July 2008
Accepted 1 August 2008
Available online 6 August 2008
Keywords:
SERT
Ó 2008 Elsevier Ltd. All rights reserved.
Diphenylsulfide derivatives
Molecular docking
1. Introduction
which demonstrated good brain penetration, a moderate signal
to noise ratio but in general, a superior pharmacological profile
The serotonin transporter (SERT) modulates serotonin (5-HT)
levels in brain and thus plays a major role in the regulation of
the serotoninergic neurotransmission.1 As reported in literature,
dysfunction of the SERT is involved in various neuropsychiatric dis-
orders such as depression, schizophrenia, mental illness, obsessive
compulsive disorders,2–5 and degenerative pathologies such as Par-
kinson and Alzheimer diseases.6,7 Correlations between SERT den-
sity and these diseases have been established encouraging the
measurement of the SERT concentration. It should be of great inter-
est to develop neuroimaging methods for the SERT quantification
in living human brain. In this purpose, some radioligands from dif-
ferent structure family were synthesized, such as [123I]b-CIT,8
over the previous synthesized ligands.13
However to get higher resolution and sensitivity by PET imag-
ing, and by the way to improve the quantification, a lot of analogs
of diphenylsulfides radiolabeled with carbon-11 have been devel-
oped notably [11C]DASB14 which is widely used as tracer for SERT
imaging in human15 (for depressive disorders see review written
by Meyer).16 Several diphenylsulfides, especially modulated on
the 40-position where Y is an halogen or a fluoroalkyl chain (see
Fig. 1), have been evaluated as potential SERT imaging agents for
PET (see several examples listed by Jarkas et al.17). If SERT quanti-
fication by PET may be assessed by the use of carbon-11 tracers, no
satisfying fluor-18 ligands (more widely used because especially of
its half-life, T1/2 = 109 min) are yet available. In fact, since few
years, numerous fluor-18 radioligands based on diphenylsulfide
[
123I]5-iodo-6-nitroquipazine,9 and 11C]McN5652,10 to image
[
SERT by single photon emission computed tomography (SPECT)
or positron emission tomography (PET). Unfortunately, their use
to assess the SERT quantification has been limited due to their
kinetics and/or important non-specific binding.11,12
Recently, a novel class of compounds, called diphenylsulfide,
has been studied in order to obtain potent SPECT and PET tracers
to image in vivo human SERT. The first diphenylsulfide ligand de-
scribed as a potent SPECT tracer to image SERT was [123I]IDAM
has emerged such as
[ [
18F]F-ADAM (called also 18F]AFA),18
[
18F]ACF,19 18F]AFM,20 18F]AFE,21 or [18F]EADAM.22 Some of them
[
[
have a limited use due to in vivo defluorination which may inter-
fere with the quantification or kinetics not compatible with the
half-life of F-18. In the aim to optimize the development of a suit-
able F-18 PET tracer to image in vivo human SERT, we have decided
to design new structures through a docking study and we have pre-
pared original fluorinated diphenylsulfides based on the MADAM
structure (Fig. 1), compound developed in our laboratory.23,24 MA-
DAM presents in vitro high affinity and selectivity for the SERT (Ki
SERT = 1.65 nM, Ki DAT > 1000 nM, Ki NET = 325 nM). The methyl
group in the 40-position (Fig. 1) seems to be responsible of its good
* Corresponding author. Address: Faculté de Pharmacie, INSERM U930, Labora-
toire de Biophysique Médicale et Pharmaceutique, 31 avenue Monge, 37200 TOURS,
France. Tel.: +33 2 47 36 72 40; fax: +33 2 47 36 72 24.
E-mail address: sylvie.mavel@univ-tours.fr (S. Mavel).
0968-0896/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2008.08.002

S. Mavel et al. / Bioorg. Med. Chem. 16 (2008) 9050–9055
9051
O
N
N
N
O
X
NH
1
A
2'
1' S
S
S
F
B
4'
Y
Y
F (CH2)2
II
IDAM
Ia Y = H
X = CH2OH, Y = I
X = NH , Y = I
Ib
Ic
ADAM
MADAM
Y = Cl
Y = Br
X = NH22 , Y = CH3
Id Y = I
Figure 1. Literature SERT ligands: IDAM,13 ADAM,25 MADAM,44 derivatives Ia–d,26 and compound II.27
selectivity. The 20-position tolerates a quite large of groups such as
amine (ADAM25, MADAM), alcohol (IDAM with a hydroxymethyl
group) or ester. The nature of the function of the chemical group
in this 20-position seems to be important for the difference of selec-
tivity observed between the derivatives.23
Furthermore, Kung et al. have studied the impact of alkyl or aryl
acylation of the amino group on the phenyl ring B (Ia–d, Fig. 1) by
in vitro evaluation26: N fluoroalkyl derivatives of ADAM produced
compounds with good binding affinity.
In 2002, Wellsow and Kovar had made by molecular modeling
around 100 predictions of binding affinity which were generated
by CoMFA and CoMSIA.27 They confirmed in silico that functional-
ization in 20-positions should allow large substituent such as ben-
zoyl moiety (see compound II, Fig. 1).
The LUDI program which is carried out in the InsightII environ-
ment (Accelrys Software Inc., San Diego, CA) has a simple scoring
function to predict binding constants for protein–ligand complex
of known three-dimensional structure.30–33 The LUDI program
positions small molecules into protein binding sites in such a
way that hydrogen bonds are formed with the protein, and lipo-
philic groups are placed into hydrophobic pockets. Ionic interac-
tions and the number of rotatable bonds in the ligand are also
taken into account. It was shown that a very significant correlation
between the sum of atom pair potentials and total binding free en-
ergy exists, and the sum is therefore a good measure for estimating
binding affinities.34 The binding affinities Ki are only a predictive
value, and the ‘score’ is proportionate to Ki.
In attempt to develop a ligand-binding model for the SERT, the
LUDI scoring program was used to identify a potential binding’s
site described by Ravna and colleagues,28 by docking different po-
tent ligands which bear different groups on the 20-position (IDAM,
MADAM, and Ia–d, see Fig. 1). Then, from the model obtained, we
have investigated binding interactions of six diphenylsulfides 5–10
in the SERT pocket (Fig. 2) and we have obtained at physiological
pH a ‘score’ (Table 1). To have a comparative level, we studied
some known SERT ligands: IDAM, MADAM, diphenylsulfides Ia–
d26 (Fig. 1), compared to the new derivatives 5–10. A higher LUDI
score represents a higher potential affinity. The predictive affinities
for 6 and 10 were the best compared to the other reference deriv-
atives (Table 1). Their predictive affinity fall into the range of
experimental values reported for potent diphenylsulfides (nano-
molar range affinity). The low affinity of nitro and the N-desmethy-
lated derivatives 7 and 9 and 5 and 8, respectively, correlated with
the fact that LUDI program failed to place these structures into the
binding pocket (‘no score’). Furthermore, the SERT amino acids in-
volved in predictive ligand bindings are listed in Table 1. These re-
sults are in agreement with that of Ravna and colleagues study
where they had shown that citalopram (SERT ligand) interacted,
for example, with Val102, Trp103 (TMH1), Tyr121 (TMH2),
Phe380 (TMH7) and Ile552 (TMH11).28
Based on these results, we have used a model of three-dimen-
sional molecular structure of SERT, obtained from the lactose per-
mease symporter (lac permease) crystal structure.28 We have
docked several potent ligands into putative binding site of SERT
(Fig. 1) to validate the model, and then, we have predicted good po-
tential binding affinity for new derivatives bearing a p-fluor-
obenzoyl or p-fluorobenzyl group on the 20-position on
a
MADAM scaffold. These compounds as well as their precursors
useable for C-11 or F-18 have been synthesized and evaluated
in vitro for their binding potential at the SERT, DAT (dopamine
transporter), and NET (noradrenalin transporter).
2. Results and discussion
2.1. Chemistry
The synthesis of derivatives 1 and 3 has been previously re-
ported in literature.23,29 Reaction of compound 1 with 2,2,2-tri-
chloroethylchloroformate was realized to protect the secondary
amine before acylation with p-fluorobenzoyl chloride. The nitro
function is reduced by using tin chloride in an HCl/methanol/
dimethylformamide solution. The addition of DMF is due to the
insolubility of 1 in classical conditions. Acylation reactions were
performed with derivatives 2 and 3 in THF, with p-(fluoro or
nitro)benzoyl chloride to afford compounds 4, 6, and 7 in high
yields (>90%). Derivative 5 was obtained in 44% yield by regenera-
tion of the secondary amine function by treatment with glacial ace-
tic acid and zinc. The reduction of the amide function of
compounds 5, 6, and 7 occurred with borane–THF complex (1 M)
and gave the fluoro(or nitro)benzylamines 8, 10, and 9, respec-
tively, in almost 70% yield (Scheme 1). All the compounds have
been characterized by 1H, 13C NMR, and mass spectrometry.
2.3. In vitro affinity, selectivity, and lipophilicity
As we have previously mentioned,23 the N-desmethylation at
1-position does not improve the affinity for the SERT as proved by
the low affinity of 5 and 8 ((Ki > 1000 nM) and a dimethylamino-
methyl group at this position is important for the cognition in the
SERT binding site. Nitro derivatives 7 and 9 did not present any
affinity for the SERT (Ki > 1000 nM). The two new fluorinated SERT
radiotracers, N,N-dimethyl-2-[2-(N-4-fluorobenzamide)-4-tolyl-
thio]-benzylamine6 andtheN,N-dimethyl-2-[2-(N-4-fluorobenzyl-
amine)-4-tolylthio]benzylamine 10 displayed good affinity for the
2.2. Molecular modeling
SERT as predicted by the docking study (Ki = 14.3
1 and
10.1 2.7 nM, respectively, Table 1). Moreover these two com-
pounds are selective for the SERT as no DAT and NET affinities were
obtained. (Ki DAT > 1000 nM and Ki NET > 1000 nM). Concerning the
The rapid and accurate calculation of binding free energy of a
putative protein–ligand complex is difficult to evaluate but impor-
tant to consider in the structure-based drug design.

9052
S. Mavel et al. / Bioorg. Med. Chem. 16 (2008) 9050–9055
N
N
NH2
NO2
H
OCOCH2CCl3
S
S
a, b
2
1
O
N
N
N
R1
NH
R1
NH
R1
NH2
S
S
S
R2
R2
c
e
1
1
: R = OCOCH2CCl3 R2 = F
8 : R1 = H
R2 = F
2
4
: R = OCOCH2CCl3
d
3 : R1 = CH3
5 : R1 = H
R2 = F
9
1
: R = CH3 R2 = NO2
1
R2 = F
10 : R1 = CH3 R2 = F
6
: R = CH3
7 : R1 = CH3
R2 = NO2
Scheme 1. Synthesis of the fluoro-MADAM derivatives and their corresponding 11C (R1 = H) and 18F (R2 = NO2) precursors. Reagents and conditions: (a) ClOCOCH2CCl3, reflux,
2 h; (b) MeOH, DMF, HCl, SnCl2, below 10 °C, then rt overnight; (c) THF, pyridine, p-(fluoro or nitro)benzoyl chloride, reflux, 2 h; (d) AcOH/Zn, rt 48 h; (e) THF, borane–THF
complex (1 M), reflux 5 h, rt overnight.
lipophilicities, the measured logP7.4 from a reverse-phase HPLC
method (see Table 2) gave a value greater than 4.5 for 10
(logP7.4 = 4.70) indicating that this compound is probably less suit-
able for crossing the blood–brain barrier. Compound 6 showed an
optimum logP7.4 at 3.0 to get a good balance between brain pene-
trance and non-specific binding (2 < logP7.4 < 3.4).35,36 For these
derivatives, clogP (ChemDrawÒ) wasnotable topredictthelipophil-
icity, on the other hand, by molecular modeling (AccelrysÒ), the logD
predictions were quite accurate.
with the in vitro data suggesting their potential application as
imaging agents of serotoninergic nerve. Our work showed that this
docking binding model is suitable to identify chemical structure
with potential SERT affinity, especially to exclude from synthesis
campaign compounds without SERT potency. In addition, because
compounds 6 and 10 exhibit good in vitro SERT affinities, labeling
and in vivo evaluation as PET tracers for the SERT imaging will be
undertaken.
4. Experimental
4.1. Chemistry
3. Conclusion
A docking study into the putative binding site of a 3D model of
human serotonin transporter shown that fluorobenzamide or fluo-
robenzylamine on 20-position on a diphenylsulfide scaffold should
induce a good binding affinity for the SERT. This result is consistent
General: NMR spectra were recorded on a Bruker DPX Advance
200 spectrometer (200 MHz for 1H, 50.3 MHz for 13C). CDCl3 was
used as solvent; chemical shifts are expressed in ppm relative to
Figure 2. Binding model of SERT with residues which bind, in predictive studies, with 6 (in green) and 10 (in red).

S. Mavel et al. / Bioorg. Med. Chem. 16 (2008) 9050–9055
9053
Table 1
Inhibition constants of diphenylsulfide derivatives for the SERT and for human cloned monoamine transporters (DAT, NET); predictive binding of SERT–ligand complex via LUDI
score, and SERT protein residues interacted in the binding site
Compound
Ki (nM) SERT
LUDI score pH 7.4
Residues of SERT
IDAM
0.1b
1.65c
34a
364
243
131
242
154
Trp103, Tyr121, Ile552
Val102, Tyr121, Ile552, Phe380
Trp103, Tyr121, Pro560
Phe380, Phe556, Leu563
Phe556, Leu563
MADAM
Ia (X = H)
Ib (X = Cl)
Ic (X = Br)
0.9a
8.0a
Ki (nM)e
Ki (nM)e
c
Id (X = I)
5
6
7
8
9
10
193a
—
—
—
—
hDAT
hNET
d
>500e
>1000
>1000
>1000
>2000
>1000
>1500
>2500
>5200
>1000
>1000
>1900
>1200
14.3 1e
>100e
413
Trp103, Pro106, Ser559
—
—
—
d
—
—
—
d
>100e
d
60 17e
10.1 2.7e
521
Val102, Phe380, Gly384, Phe556
a
Literature inhibition constants using [125I]IDAM.26
Literature inhibition constants using [125I]IPT.42
b
c
Literature inhibition constants using [3H]paroxetine.23
‘—’: no score obtained by LUDI studies.
d
e
Inhibition constants (Ki) were obtained from means SD of four separate determination each in triplicate.
7.13 (m, 1H), 7.13–7.32 (m, 5H). 13C NMR d: 21.4, 34.0, 50.3,
75.1, 95.7, 109.9, 115.0, 120.0, 125.4, 126.2, 127.0, 128.0, 133.4,
135.9, 137.1, 141.5, 148.6, 154.8.
Table 2
Lipophilicity measurements and predictions for compounds 6 and 10 and references
derivatives (ADAM and MADAM)
Compound Experimental
Discovery Studio, AccelrysÒ
ChemDrawÒ
logP7.4
logDd
clogPe
4.1.2. N-Methyl-2-[2-(N-4-fluorobenzamide)-4-
tolylthio]benzylamine (5)
ADAM
MADAM
6
2.52a
2.46b
2.99c
4.70c
2.29
2.20
3.94
4.54
4.51
3.59
5.46
5.90
To a solution of 2 (694 mg, 1.6 mmol) in THF (16 mL) and pyri-
dine (252 mg, 3.2 mmol) was added at 0 °C 4-fluorobenzoyl chlo-
ride (349 mg, 2.2 mmol). The reaction mixture was heated to
reflux for 2 h, the solvents were removed under reduced pressure
and gave 4 as a crude compound. The crude 4 was dissolved to a
solution of acetic acid (13 mL) and zinc (2 g, 30 mmol) and stirred
for 2 days. The solution was filtrated on Celite, diluted with water
(30 mL) and basified with concentrated NaOH. Extraction with
methylene chloride (2ꢀ 30 mL) and flash chromatography (petro-
leum ether/EtOAc/MeOH/TEA 5:5:0.5:0.5) afforded compound 5
as a beige solid in a 44% total yield. 1H NMR d: 1.56 (s, 1H), 2.49
(s, 3H), 2.52 (s, 3H), 3.95 (s, 2H), 6.82 (dd, 1H, J = 7.1 Hz, J = 2.0
Hz), 7.01–7.23 (m, 5H), 7.37 (dd, 1H, J = 6.8 Hz, J = 1.7 Hz), 7.49–
7.62 (m, 3H), 8.53 (d, 1H, J = 1.1 Hz), 8.93 (s, 1H). 13C NMR d:
21.6, 35.9, 53.8, 115.4 (d, 2C, J = 22.1 Hz), 116.6, 121.2, 125.4,
125.9, 126.7, 128.0, 129.1 (d, 2C, J = 9.1 Hz), 129.2, 130.9 (d, 1C,
J = 2.5 Hz), 134.8, 136.3, 137.4, 139.4, 141.4, 164.0, 164.6 (d,
J = 253.1 Hz). MS: m/e = 380 (M+ꢁ, 2), 150 (26), 123 (64), 120
(100), 95 (39), 42 (20).
10
a
n-Octanol/0.02 M phosphate buffer partition.22
n-Octanol/0.02 M phosphate buffer partition.43
Reverse-phase HPLC experiments.
b
c
d
e
Calculated logD, Discovery Studio, QSAR protocol, 2008, AccelrysÒ
Calculated clogP, ChemDraw Ultra 10.0, 2005, CambridgeSoft.
.
TMS as an internal standard. Mass spectra were obtained on a CG-
MS Hewlett Packard 5989A spectrometer (electronic impact at
70 eV). The thin-layer chromatographic (TLC) analyses were per-
formed using Merck 60 F254 silica gel plates. Flash chromatography
was used for routine purification of reaction products using silica
gel (230–400 mesh). For flash chromatography petroleum ether,
ethyl acetate (EtOAc), methanol (MeOH), and triethylamine (TEA)
were used. Visualization was accomplished under UV. All chemi-
cals and solvents were of commercial quality and were purified fol-
lowing standard procedures. Elemental analyses of new
compounds were within 0.4% of theoretical values.
1 and 3 have been synthesized in the laboratory as previously
4.1.3. N,N-Dimethyl-2-[2-(N-4-fluorobenzamide)-4-
tolylthio]benzylamine (6)
described.23,29
To a solution of 3 (545 mg, 2 mmol) in THF (20 mL) and pyridine
(0.32 mL 4 mmol) was added at 0 °C 4-fluorobenzoyl chloride
(349 mg, 2.2 mmol). The reaction mixture was heated to reflux
for 2 h, and solvents were removed under reduced pressure. The
crude product was purified by flash chromatography (petroleum
ether/EtOAc/TEA 7:2:1). Compound 6 was obtained as a beige solid
in 90% yield. 1H NMR d: 2.32, (s, 6H), 2.49 (s, 3H), 3.60 (s, 2H), 6.76–
6.84 (m, 1H), 6.99–7.20 (m, 5H), 7.29–7.37 (m, 1H), 7.52–7.64 (m,
3H), 8.55 (d, 1H, J = 1.2 Hz), 8.93 (s, 1H). 13C NMR d: 21.8, 45.1 (2C),
62.4, 115.6 (d, 2C, J = 21.6 Hz), 117.1, 121.0, 125.5, 125.6, 126.5,
128.3, 129.2 (d, 2C, J = 8.6 Hz), 130.2, 131.1, (d, J = 2.5 Hz), 136.2,
136.6, 136.7, 136.9, 141.6, 164.1, 164.8 (d, J = 253.0 Hz). MS: m/
e = 394 (M+ꢁ, 3), 165 (26), 164 (21), 150 (48), 134 (65), 132 (26),
123 (100), 95 (42), 58 (48), 46 (25), 44 (25). Anal. Calcd for
4.1.1. N-Methyl-N-(3,3,3-trichloropropanoyloxy)-2-[2-amino-4-
tolylthio]benzylamine (2)
Compound 1 (647 mg, 2.24 mmol) was heated in refluxing
2,2,2-trichloroethylchloroformate (3 mL, 21.79 mmol) for 2 h. The
excess of 2,2,2-trichloroethylchloroformate was removed under re-
duce pressure. The crude product was poured in methanol (16 mL),
dimethylformamide (16 mL), and HCl solution (33%) (8 mL). The
solution was cooled and SnCl2 (1.41 g, 7 mmol) was added by por-
tion below 10 °C. The reaction mixture was stirred at ambient tem-
perature overnight, treated with water (20 mL), basified with 10 N
NaOH solution (to pH 10), and extracted with AcOEt (2ꢀ 20 mL).
The organic phases were dried, removed under vacuo, and the res-
idue was purified by flash chromatography (petroleum ether/
EtOAc/TEA: 8/2/1) to give 2 in 52% yield. 1H NMR d: 2.34 (s, 3H),
3.07 (s, 3H), 4.22 (s, 2H), 4.76–4.87 (m, 4H), 6.61–6.71 (m, 1H),
C23H23FN2OS: C, 70.02; H, 5.88; N, 7.10. Found: C, 69.76; H, 5.86;
N, 7.14.

9054
S. Mavel et al. / Bioorg. Med. Chem. 16 (2008) 9050–9055
4.1.4. N,N-Dimethyl-2-[2-(N-4-nitrobenzamide)-4-
tolylthio]benzylamine (7)
solvent, the crude product was purified by flash chromatography
(petroleum ether/TEA 9.5:0.5). Derivative 10 was obtained as a yel-
low solid in 71% yield. 1H NMR d: 2.29 (s, 3H); 2.31 (s, 6H), 3.60 (s,
2H), 4.30 (d, 2H, J = 5.8 Hz), 6.09 (t, 1H, J = 5.8 Hz), 6.36 (d, 1H,
J = 1.2 Hz), 6.55 (dd, 1H, J = 7.7 Hz, J = 1.2 Hz), 6.85–7.04 (m, 5 H),
7.10–7.21 (m, 2H), 7.23–7.31 (m, 1H), 7.48 (d, 1H, J = 7.7 Hz). 13C
NMR d: 21.8, 45.3 (2C), 46.2, 62.3, 111.2, 112.2, 115.2 (d, 2C,
J = 21.1 Hz), 117.6, 125.2, 127.9 (d, 2C, J = 8.6 Hz), 128.1 (2C),
130.1, 134.9 (d, J = 2.5 Hz), 136.9, 137.5 (2C), 141.2, 148.6, 161.6
(d, J = 244.6 Hz). MS: m/e = 380 (M+ꢁ, 8), 226 (49), 212 (32), 211
(26), 194 (26), 165 (100), 164 (58), 150 (28), 134 (78), 132 (37),
109 (66), 58 (46). Anal. Calcd for C23H25FN2S: C, 72.60; H, 6.62;
N, 7.36. Found: C, 72.30; H, 6.63; N, 7.38.
To a solution of 3 (200 mg, 0.73 mmol) in THF (5 mL) and pyri-
dine (0.09 mL, 1.1 mmol) was added at 0 °C 4-nitrobenzoyl chlo-
ride (132 mg, 0.715 mmol). The reaction mixture was heated to
reflux for 2 h, the solvents were removed under reduced pressure.
Compound 7 was obtained after flash chromatography (petroleum
ether/EtOAc/TEA 7:2:1) as a yellow powder in quantitative yield.
1H NMR d: 2.32 (s, 6H), 2.51 (s, 3H), 3.60 (s, 2H), 6.75–6.83 (m,
1H), 7.05–7.23 (m, 3H), 7.30–7.38 (m, 1H), 7.60 (d, 1H,
J = 7.8 Hz), 7.68 (d, 2H, J = 8.8 Hz), 8.24 (d, 2H, J = 8.8 Hz), 8.54 (s,
1H), 9.02 (s, 1H). 13C NMR d: 21.8, 45.1 (2C), 62.5, 117.6, 121.2,
123.7 (2C), 125.7, 126.1, 126.5, 128.0 (2C), 128.4, 130.4, 136.0,
136.7, 139.0, 140.4, 141.7, 141.7, 149.6, 163.0. MS: m/e = 421
(M+ꢁ, 3), 165 (40), 164 (27), 150 (47), 134 (79), 132 (39), 120
(59), 92 (32), 91 (29), 58 (100), 46 (31), 44 (68), 42 (34). Anal. Calcd
for C23H23N3O3S: C, 65.54; H, 5.50; N, 9.97. Found: C, 65.30; H,
5.49; N, 10.00.
4.2. Molecular modeling
The calculations and simulations were performed on a PC with
linux workstation using the software modules Builder, Homology
and LUDI in the InsightII environment (vers. InsightII 2005, Accel-
rysÒ Software Inc., San Diego, CA). The SERT model is described by
Ravna and colleagues.28,37 The different structures were improved
by energy minimization and were also optimized using AM1 meth-
od. The center of the investigations in the binding site is obtained
as the centroid of citalopram in the model of Ravna and col-
leagues.28 Interaction sites were calculated within a radius of
7.0 Å. The fit was achieved with a maximum rms deviation of
0.45 Å from the interaction sites for each structure.
4.1.5. N-Methyl-2-[2-(N-4-fluorobenzylamine)-4-
tolylthio]benzylamine (8)
To a solution of compound 5 (228 mg, 0.60 mmol) in THF (2 mL)
under nitrogen atmosphere was added drop by drop borane–THF
(1 M, 1.9 mL) at 0 °C. The mixture was heated to reflux for 5 h, stir-
red at room temperature overnight, and quenched with HCl solu-
tion (10 N, 0.1 mL). The residue was then dissolved in water
(5 mL), basified with NaOH solution (to pH 10), and extracted with
methylene chloride (2ꢀ 10 mL). After evaporation of the solvent,
the crude product was purified by flash chromatography (petro-
leum ether/EtOAc/TEA 5:5:0.5), 8 was obtained as a beige solid in
70% yield. 1H NMR d: 1.62 (s, 1H), 2.32 (s, 3H), 2.50 (s, 3H), 3.94
(s, 2H), 4.32 (d, 2H, J = 5.6 Hz), 5.74 (t, 1H, J = 5.6 Hz), 6.43 (d, 1H,
J = 1.1 Hz), 6.58 (dd, 1H, J = 7.7 Hz, J = 1.1 Hz), 6.97–7.11 (m, 5H),
7.13–7.22 (m, 2H), 7.28–7.35 (m, 1H), 7.45 (d, 1H, J = 7.7 Hz). 13C
NMR d: 21.6, 35.8, 46.1, 53.6, 110.9, 111.1, 114.9 (d, 2C,
2J = 21.6 Hz), 117.7, 125.2, 127.0, 127.5, 128.1 (d, 2C, 3J = 8.1 Hz),
129.0, 134.7 (d, 4J = 2.5 Hz), 136.2, 137.1, 137.2, 141.2, 148.5,
161.4 (d, 1J = 244.6 Hz). MS: m/e = 366 (M+ꢁ, 22), 335 (8), 240
(12), 226 (20), 212 (25), 150 (87), 120 (100), 109 (67), 83 (10),
42 (15).
For the lipophilicities, logPprediction was calculated with the
QSAR protocol in the Discovery Studio environment (AccelrysÒ
Software Inc., San Diego, CA).
4.3. In vitro binding studies
4.3.1. In vitro binding assays for SERT
Compounds were tested in duplicate in competition with
[3H]MADAM with a crude membrane fraction of homogenate of
rat brain cortex. For these studies, each sample contained 0.2 mL
of [3H]MADAM at a concentration of 50 pM (Kd = 50 pM), 0.2 mL
of competitors at various concentrations ranging from 10ꢂ5 M to
10ꢂ10 M, 0.5 mL containing 60
lg of membrane protein prepara-
tion in a total volume of 1 mL in the tris–HCl buffer (50 mM Tris,
120 mM NaCl, 5 mM KCl, pH 7.4). Non-specific binding was deter-
mined with 10ꢂ6 nM paroxetine. Samples were incubated at 22 °C
for 90 mn, filtered on glass fiber filters (GF/B, Whatman), and
washed with ice cold buffer and the residual radioactivity was
measured with a beta counter (LKB, rack Beta 1215). The IC50 val-
ues were determined graphically for each compound and the Ki
values were calculated according to Cheng and Prusoff.38 The re-
sults (mean of four determinations) are expressed as inhibition
constants (Ki) and are summarised in Table 1.
4.1.6. N,N-Dimethyl-2-[2-(N-4-nitrobenzylamine)-4-
tolylthio]benzylamine (9)
Compound
9 was prepared from derivative 7 (253 mg,
0.60 mmol) using the procedure described above. Compound 9
was obtained after purification by flash chromatography (petro-
leum ether/TEA 9.5:0.5) as a beige solid in 72% yield. 1H NMR d:
2.27 (s, 3H); 2.35 (s, 6H), 3.65 (s, 2H), 4.45 (d, 2H, J = 6.3 Hz),
6.24 (d, 1H, J = 1.0 Hz), 6.57 (dd, 1H, J = 7.7, J = 1.0 Hz), 6.73 (t,
1H, J = 6.3 Hz), 7.08–7.22 (m, 6 H), 7.55 (d, 1H, J = 7.6 Hz), 8.05 (d,
2H, J = 8.7 Hz). 13C NMR d: 21.7, 45.3 (2C), 46.0, 62.4, 110.9,
112.7, 117.8, 123.4 (2C), 125.4 (2C), 127.9 (2C), 128.9, 130.3,
136.6, 137.5 (2C), 141.1, 146.6, 147.6, 147.9. MS: m/e = 407 (M+ꢁ,
10), 273 (23), 136 (80), 134 (55), 120 (29), 58 (100), 44 (43). Anal.
Calcd for C23H25N3 O2S: C, 68.31; H, 5.99; N, 9.37. Found: C, 68.11;
H, 6.00; N, 9.32.
4.3.2. In vitro binding assays for NET and DAT
Candidate compounds were assayed for their affinities to the
monoamine transporters (NET and DAT) in competitive binding
experiments in vitro using cloned human transporters (hNET and
hDAT) expressed on HEK-293 cells and the radioligands [3H]nisox-
etine (NET), and [3H]GBR12935 (DAT), in accordance with the pub-
lished procedures.39
4.1.7. N,N-Dimethyl-2-[2-(N-4-fluorobenzylamine)-4-
tolylthio]benzylamine (10)
To a solution of compound 6 (600 mg, 1.52 mmol) in THF
(3.8 mL) under nitrogen atmosphere was added drop by drop bor-
ane–THF (1 M, 3.8 mL) at 0 °C. The mixture was heated to reflux for
5 h, stirred at room temperature overnight, and quenched with HCl
solution (10 N, 0.2 mL). The residue was then dissolved in water
(10 mL), basified with NaOH solution (to pH 10), and extracted
with methylene chloride (2ꢀ 20 mL). After evaporation of the
4.4. Lipophilicity measurements
An indirect determination of octanol–water partition coeffi-
cients is to obtain logP7.4 by reverse-phase C-18 HPLC studies by
comparison of their retention time (in triplicate) to that of seven
reference compounds with known logP values as previously re-
ported.40,41 A Water XBridge, 5
l
m 4.6 mm ꢀ 150 mm analytical

S. Mavel et al. / Bioorg. Med. Chem. 16 (2008) 9050–9055
9055
14. Houle, S.; Ginovart, N.; Hussey, D.; Meyer, J. H.; Wilson, A. A. Eur. J. Nucl. Med.
2000, 27, 1719.
15. Frankle, W. G.; Slifstein, M.; Gunn, R. N.; Huang, Y.; Hwang, D. R.; Darr, E. A.;
Narendran, R.; Abi-Dargham, A.; Laruelle, M. J. Nucl. Med. 2006, 47, 815.
16. Meyer, J. H. J. Psychiatry Neurosci. 2007, 32, 86.
column was used with methanol/ 0.1 M phosphate buffer (pH 7.4)
(50/50) at a flow rate of 1 mL/min. The standards used to produce a
calibration equation were quinoline, phenol, benzamide, chlor-
promazine, hexachlorobenzene, ADAM, MADAM, dissolved in the
mobile phase. A calibration curve of retention time versus logP7.4
was obtained with an experimental calibration equation
(y = 1.438ꢁLn(retention time) ꢂ 0.7301) with R2 of 0.93.
17. Jarkas, N.; Voll, R. J.; Williams, L.; Votaw, J. R.; Owens, M.; Goodman, M. M. J.
Med. Chem. 2008, 51, 271.
18. Shiue, G. G.; Fang, P.; Shiue, C.-Y. Appl. Radiat. Isot. 2003, 58, 183.
19. Oya, S.; Choi, S. R.; Coenen, H.; Kung, H. F. J. Med. Chem. 2002, 45, 4716.
20. Huang, Y.; Bae, S. A.; Zhu, Z.; Hwang, D. R.; Narendran, R.; Talbot, P. S.; Hackett,
E.; Kegeles, L. S.; Laruelle, M. J. Nucl. Med. 2002, 43S, 358.
21. Huang, Y.; Bae, S. A.; Zhu, Z.; Guo, N.; Roth, B. L.; Laruelle, M. J. Med. Chem. 2005,
48, 2559.
Acknowledgments
22. Jarkas, N.; McConathy, J.; Votaw, J. R.; Voll, R. J.; Malveaux, E.; Camp, V. M.;
Williams, L.; Goodman, R. R.; Kilts, C. D.; Goodman, M. M. Nucl. Med. Biol. 2005,
32, 75.
23. Emond, P.; Vercouillie, J.; Innis, R.; Chalon, S.; Mavel, S.; Frangin, Y.; Halldin, C.;
Besnard, J. C.; Guilloteau, D. J. Med. Chem. 2002, 45, 1253.
24. Chalon, S.; Tarkiainen, J.; Garreau, L.; Hall, H.; Emond, P.; Vercouillie, J.; Farde,
L.; Dasse, P.; Varnas, K.; Besnard, J. C.; Halldin, C.; Guilloteau, D. J. Pharmacol.
Exp. Ther. 2003, 304, 81.
25. Choi, S. R.; Hou, C.; Oya, S.; Mu, M.; Kung, M. P.; Siciliano, M.; Acton, P. D.; Kung,
H. F. Synapse 2000, 38, 403.
26. Choi, S.-R.; Oya, S.; Hou, C.; Kung, H. F. J. Labelled Compd. Radiopharm. 2001, 44,
S190.
This study was funded in part by the EC-FP6-project DiMI,
LSHB-CT-2005-512146. We thank the ‘‘Département d’analyses
Chimiques et S.R.M. biologique et médicale” (Tours, France) for
chemical analyses. The authors thank Nathalie Meheux for her
technical help.
Ki determinations were generously provided by the National
Institute of Mental Health’s Psychoactive Drug Screening Program,
Contract # NO1MH32004 (NIMH PDSP). The NIMH PDSP is directed
by Bryan L. Roth MD, Ph.D. at the University of North Carolina at
Chapel Hill and Project Officer Jamie Driscol at NIMH, Bethesda,
MD, USA.
27. Wellsow, J.; Kovar, K. A.; Machulla, H. J. J. Pharm. Pharm. Sci. 2002, 5, 245.
28. Ravna, A. W.; Sylte, I.; Kristiansen, K.; Dahl, S. G. Bioorg. Med. Chem. 2006, 14,
666.
29. Tarkianen, J.; Vercouillie, J.; Emond, P.; Sandell, J.; Hiltunen, J.; Frangin,
Y.; Guilloteau, D.; Halldin, C. J. Labelled Compd. Radiopharm. 2001, 44,
1013.
References and notes
30. Bohm, H. J. J. Comput. Aided Mol. Des. 1992, 6, 593.
31. Bohm, H. J. J. Comput. Aided Mol. Des. 1992, 6, 61.
32. Bohm, H. J. J. Comput. Aided Mol. Des. 1994, 8, 243.
33. Bohm, H. J. J. Comput. Aided Mol. Des. 1998, 12, 309.
34. Muegge, I.; Martin, Y. C. J. Med. Chem. 1999, 42, 791.
35. Waterhouse, R. N. Mol. Imaging Biol. 2003, 5, 376.
36. Wilson, A. A.; Houle, S. J. Labelled Compd. Radiopharm. 1999, 42, 1277.
37. Ravna, A. W.; Sylte, I.; Dahl, S. G. J. Pharmacol. Exp. Ther. 2003, 307, 34.
38. Cheng, Y.; Prusoff, W. H. Biochem. Pharmacol. 1973, 22, 3099.
39. Huang, Y.; Hwang, D.-R.; Zhu, Z.; Bae, S.-A.; Guo, N.; Sudo, Y.; Kegeles, L. S.;
Laruelle, M. Nucl. Med. Biol. 2002, 29, 741.
40. Brent, D. A.; Sabatka, J. J.; Minick, D. J.; Henry, D. W. J. Med. Chem. 1983, 26,
1014.
41. Waterhouse, R. N.; Mardon, K.; Giles, K. M.; Collier, T. L.; O’Brien, J. C. J. Med.
Chem. 1997, 40, 1657.
42. Kung, M. P.; Hou, C.; Oya, S.; Mu, M.; Acton, P. D.; Kung, H. F. Eur. J. Nucl. Med.
1999, 26, 844.
43. Jarkas, N.; McConathy, J.; Voll, R. J.; Goodman, M. M. J. Med. Chem. 2005, 48,
4254.
1. Lucki, I. Biol. Psychiatry 1998, 44, 151.
2. D’Amato, R. J.; Largent, B. L.; Snowman, A. M.; Snyder, S. H. J. Pharmacol. Exp.
Ther. 1987, 242, 364.
3. Mann, J. J. Neuropsychopharmacology 1999, 21, 99S.
4. Owens, M. J.; Nemeroff, C. B. Depress. Anxiety 1998, 8, 5.
5. Saxena, S.; Brody, A. L.; Ho, M. L.; Alborzian, S.; Maidment, K. M.; Zohrabi, N.;
Ho, M. K.; Huang, S. C.; Wu, H. M.; Baxter, L. R., Jr. Arch. Gen. Psychiatry 2002, 59,
250.
6. Meltzer, C. C.; Smith, G.; DeKosky, S. T.; Pollock, B. G.; Mathis, C. A.; Moore, R.
Y.; Kupfer, D. J.; Reynolds, C. F. Neuropsychopharmacology 1998, 18, 407.
7. Menza, M. A.; Palermo, B.; DiPaola, R.; Sage, J. I.; Ricketts, M. H. J. Geriatr.
Psychiatry Neurol. 1999, 12, 49.
8. Innis, R.; Baldwin, R.; Sybirska, E.; Zea, Y.; Laruelle, M.; al-Tikriti, M.; Charney,
D.; Zoghbi, S.; Smith, E.; Wisniewski, G., et al Eur. J. Pharmacol. 1991, 200, 369.
9. Biegon, A.; Mathis, C. A.; Hanrahan, S. M.; Jagust, W. J. Brain Res. 1993, 619, 236.
10. Suehiro, M.; Scheffel, U.; Ravert, H. T.; Dannals, R. F.; Wagner, H. N., Jr. Life Sci.
1993, 53, 883.
11. Guilloteau, D.; Chalon, S. Curr. Pharm. Des. 2005, 11, 3237.
12. Laruelle, M.; Slifstein, M.; Huang, Y. Methods 2002, 27, 287.
13. Oya, S.; Kung, M. P.; Acton, P. D.; Mu, M.; Hou, C.; Kung, H. F. J. Med. Chem. 1999,
42, 333.
44. Oya, S.; Choi, S. R.; Hou, C.; Mu, M.; Kung, M. P.; Acton, P. D.; Siciliano, M.; Kung,
H. F. Nucl. Med. Biol. 2000, 27, 249.