Synthesis, molecular docking and binding studies of selective serotonin transporter inhibitors

Article abstract of DOI:10.1016/j.ejmech.2010.12.018

With the aim of obtaining compounds possessing high SERT selectivity, in the present work we synthesized and studied the inhibition of serotonin (SERT), dopamine (DAT) and norepinephrine (NET) transporters by docking studies and experimental binding measu

Full text of DOI:10.1016/j.ejmech.2010.12.018

European Journal of Medicinal Chemistry 46 (2011) 825e834
Contents lists available at ScienceDirect
European Journal of Medicinal Chemistry
journal homepage: http://www.elsevier.com/locate/ejmech
Original article
Synthesis, molecular docking and binding studies of selective serotonin
transporter inhibitors
,
b
a
a
Susanna Nencetti a , Maria R. Mazzoni , Gabriella Ortore , Annalina Lapucci ,
Janette Giuntini b, Elisabetta Orlandini a, Irene Banti a, Elisa Nuti a,
Antonio Lucacchini b, Gino Giannaccini b, Armando Rossello a
*
a Dipartimento di Scienze Farmaceutiche, Università di Pisa, Via Bonanno 6, 56126 Pisa, Italy
b Dipartimento di Psichiatria, Neurobiologia, Farmacologia e Biotecnologie, Facoltà di Farmacia, Università di Pisa, Via Bonanno 6, 56126 Pisa, Italy
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 2 July 2010
Received in revised form
14 December 2010
Accepted 14 December 2010
Available online 22 December 2010
With the aim of obtaining compounds possessing high SERT selectivity, in the present work we
synthesized and studied the inhibition of serotonin (SERT), dopamine (DAT) and norepinephrine (NET)
transporters by docking studies and experimental binding measurements of a series of 4-(aryl)piperidin-
3-one O-4-benzyl oxime hydrochlorides (1e10) of both
E and Z configuration. E configuration
compounds showed high SERT binding affinities (Ki ¼ 10e98 nM) and high SERT selectivities over both
NET and DAT. The molecular docking studies allowed a rationalization of the molecular basis of drug-
SERT interactions both of the synthesized compounds and paroxetine and fluoxetine used as reference
antidepressant drugs.
Keywords:
Serotonin transporter inhibitors
Neurotransmitters
Ó 2010 Elsevier Masson SAS. All rights reserved.
Piperidine oxime ether derivatives
Molecular docking
1. Introduction
prescribed pharmacological treatment for depression [5e9]. The
breadth of their therapeutic profile has allowed their use for
Monoamine transporters i.e. the serotonin transporter (SERT),
norepinephrine transporter (NET) and dopamine transporter (DAT)
play an important role in maintaining the concentration of biogenic
amine in the central nervous system (CNS) by transporting
monoamines across neuronal membranes into presynaptic nerve
cells. SERT is part of the large family of Naþ/Clꢀ dependent
membrane transporters [1e4] and is a 630-aminoacid protein and
believed to have 12 transmembrane domains.
SERT has great clinical importance as a molecular target for
selective serotonin re-uptake inhibitors (SSRIs) such as fluoxetine,
paroxetine and citalopram (Fig. 1) which exert their antidepressant
action by preventing the binding of serotonin to its recognition site
on the SERT and therefore by enhancing the concentration of this
neurotransmitter in the synaptic cleft. Due to their favourable side
effects profile and safety over a wide-dose range when compared
with traditional tricyclic antidepressants (TCAs), they have domi-
nated the market of antidepressants becoming the most widely
treating panic disorders (PD), post traumatic stress disorders
(PTSD), social phobia, pre-menstrual dysphoric disorder (PMDD),
obsessiveecompulsive disorders (OCD) and anorexia [10,11].
Despite the fact that SSRIs generally possess good tolerability and
have a reasonable safety profile, 2e4 weeks delay for the onset of
action and side effects such as anxiety, sleep disturbance, sexual
dysfunction and gastrointestinal intolerance remain as considerable
barriers to effective therapy [12e14]. During the last decade fluoxe-
tine, paroxetine and other SSRIs have been found to possess
secondary binding properties among which dopamine (DA) and/or
norepinephrine (NE) re-uptake inhibition, so that they are not so
selective as initially thought [15e17]. In fact, only escitalopram, the
active enantiomer (S(þ)) of citalopram used for the treatment of
depression and anxiety disorders, has been reconfirmed as a pure
SSRI [18e21]. Knowledge of the additional properties of SSRIs
sometimes helps the physician in selecting the SSRIs suitable to the
clinical profile of the patient and allows their use in other clinical
settings outside psychiatry. On the contrary, when the secondary
binding properties are considered as undesirable side effects, the use
of drugs with greater selectivity for SERT and without appreciable
secondary binding properties may be recommended [22]. In this
* Corresponding author. Tel.: þ39 050 2219559; fax: þ39 050 2219605.
E-mail address: nencettis@farm.unipi.it (S. Nencetti).
0223-5234/$ e see front matter Ó 2010 Elsevier Masson SAS. All rights reserved.
doi:10.1016/j.ejmech.2010.12.018

826
S. Nencetti et al. / European Journal of Medicinal Chemistry 46 (2011) 825e834
F
2. Chemistry
The general methods for synthesis of E and Z 4-(aryl)piperidin-3-
one O-4-benzyl oxime 1,3,5,7,9 and 2,4,6,8,10, respectively, are
shown in Scheme 1. 1-Benzyl-4-piperidone (11) was reacted with
the appropriate aryl Grignard reagent in tetrahydrofuran (THF) at
reflux temperature to give the 4-piperidinols 12 and 13 in good
yields. Then, the 4-hydroxyl group was removed by dehydratation
with p-toluensulfonic acid (PTSA) in benzene to give the 3,4-dehy-
dro piperidines 14 and 15. Hydroboration with borane-methyl
sulfide complex of the olefinic double bond of 14 and 15 followed by
treatment with hydrogen peroxide and sodium hydroxide gave the
diastereomeric mixture of 3-piperidinols 16 and 17 which, without
separation, were oxidized under Swern conditions to the corre-
sponding keto compounds 18 and 19. The reaction of 18 and 19 with
hydroxylamine hydrochloride gave the E/Z mixture of the oximes 20
and 21 which were alkylated with the appropriate benzyl chlorides
to yield the Z oxime ether isomers 22e26, exclusively. Catalytic
hydrogenolysis of 22e26 in acidic medium gave the E/Z mixture
(1:1.5) of the 4-(aryl)piperidin-3-one O-4-benzyl oxime hydro-
chlorides (1e10) from which the compounds of E (1,3,5,7,9) and Z
(2,4,6,8,10) configuration were then isolated by preparative TLC.
The configuration around the N]C double bond of the couples
of the oxime ether derivatives was assigned on the basis of
a comparison of their 1H NMR spectral data. In the compounds of E
configuration (1,3,5,7,9) the proton linked to the C(4) of the piper-
idine ring (H4) resonates at lower fields with respect to the same
CF3
O
O
O
O
N
H
N
H
Fluoxetine
Paroxetine
F
O
N
NH2
O
N
F3C
O
NC
Fluvoxamine
Citalopram
Fig. 1. Representative structure of some widely used SSRIs.
context the search for highly SERT selective SSRIs still remains an
important goal offering significant advantages with regards to
tolerability.
Heterocyclic compounds carrying piperidine skeleton are
attractive targets of organic synthesis, moreover piperidine-based
compounds with substituents at carbons 3 and 4 have been docu-
mented as potent antidepressant agents [23]. With the aim of
obtaining new SSRIs with enhanced selectivity towards SERT, in the
present work we designed and studied a series of piperidine oxime
ether derivatives of general structure A (Fig. 2).
These new heterocyclic compounds (A) present an arylmethy-
lenoxyimino moiety (ArCH2ON]) in the 3 position and an aryl
substituted moiety in the 4 position of the piperidine ring. The
choice to introduce an oximethereal moiety, as found in fluvoxamine
structure (Fig. 1), as part of the linker between the aminic portion
hydrogen in the compounds of
Z configuration (2,4,6,8,10)
(1.10e1.13 ppm). This fact may be attributed to the paramagnetic
effect of the spatially proximal oximethereal oxygen. On the
contrary in the compounds of Z configuration it is the protons
linked to the C(2) of the piperidine ring (H2 and H02) which, being on
the same side as the oximethereal oxygen, that resonate at lower
fields with respect to the same hydrogen of the corresponding E
isomers [27] (0.16e0.25 and 0.40e0.45 ppm).
and the b ring was due to the observation that the replacement of an
3. Pharmacological evaluation
ArOCH2 group, as found in some SERT inhibitors, with an arylme-
thylenoxyimino (ArCH2ON]C) or aryliminoxymethyl (ArCH]
NOCH2) moiety leads to compounds with improved biological
activity compared to the parent compounds [24e26].
All the synthesized compounds of both E (1,3,5,7,9) and Z
(2,4,6,8,10) configuration were evaluated as a racemic mixture for
their ability to interfere with the system of 5-HT, NE and DA
transmission by evaluating their ability to inhibit the binding of
specific radioligands to SERT and NET in rabbit cortical membranes
and to DAT in rabbit striatal membranes. [3H]-Paroxetine and [3H]-
nisoxetine were used as specific radiolabelled ligand for SERT and
NET, respectively while [3H]WIN 35,428 was used to label DAT. The
structure, the Ki values, calculated using the Cheng-Prusoff equa-
tion [28], and selectivity ratio for compounds 1e10 are shown in
Table 1 together with those obtained in the same assays for
fluoxetine and paroxetine used as reference drugs.
The results of binding assays reported in Table 1 show that all the
synthesized compounds (1e10) were able to inhibit [3H]-paroxetine
binding to SERT with Ki values in the nanomolar range. The most
active compounds of this series werethe oxime etherderivatives of E
configuration 1, 5 and 9 and the Z compound 6 with Ki values ranging
from 10.28 (compound 9) to 81.07 nM (compound 6). E configura-
tion compounds 3 and 7 and the Z compound 10 were slightly less
active (Ki values: 98.42, 96.82, and 96.58 nM, respectively) while the
Z compounds 2, 4 and 8 were those with the highest Ki values
(Ki > 100 nM). In the same kind of assay, fluoxetine and paroxetine
showed Ki values of 5.80 and 0.31 nM, respectively.
The R and R1 substituents (F, CF3, Cl) on the
a and b aromatic
moieties of compounds A are the most frequently present in anti-
depressant agents clinically used. Moreover to understand the
molecular basis of compounds 1e10 interactions with monoamine
transporters we performed a docking study using the structural
information about SERT, DAT, NET provided by the models of the
transporters constructed by Ravna et al. [4].
This paper reports the synthesis, the docking studies and the
binding properties of 4-(aryl)piperidin-3-one O-4-benzyl oxime
derivatives of both E and Z configuration (1e10, see Table 1)
towards SERT, NET and DAT.
R
α
4
N
3
O
β
N
1
R
A comparison between the Ki values of E and Z isomers showed
differences between the diastereomeric couples. Compound 3 was
1.4-fold more potent than 4, compound 5 was 3.6 fold more potent
than 6, compound 7 was 4-fold more potent than 8, compounds 1
H
A
Fig. 2. General structure of the new piperidine oxime ether derivatives A.

S. Nencetti et al. / European Journal of Medicinal Chemistry 46 (2011) 825e834
827
R
R
R
O
OH
OH
N
N
N
N
Bn
11
Bn
Bn
Bn
12, R = F
13, R = CF3
14, R = F
15, R = CF3
16, R = F
17, R = CF3
R
R
R
N
O
NOH
O
N
N
N
R1
Bn
Bn
Bn
22, R = F, R1= 4-F
23, R = F, R1 = 4-CF3
24, R = F, R1 = 2-Cl
25, R = CF3, R1 = 4-F
20, R = F
21, R = CF3
18, R = F
19, R = CF3
26, R = CF3, R1 = 4-CF3
h
R
R
O
N
N
O
1
R
N
H
N
H
1
2, R = F, R1 = 4-F
1, R = F, R1 = 4-F
R
4, R = F, R1 = 4-CF3
3, R = F, R1 = 4-CF3
6, R = F, R1 = 2-Cl
8, R = CF3, R1 = 4-F
5, R = F, R1 = 2-Cl
7, R = CF3, R1 = 4-F
10, R = CF3, R1 = 4-CF3
9, R = CF3, R1 = 4-CF3
Scheme 1. a: ArMgBr, an.THF, reflux; b: PTSA, C6H6, reflux; c: BH3(CH3)2S, an.THF; d: H2O2, NaOH ; e: (COCl)2,DMSO, Et3N, CH2Cl2; f: NH2OH HCl/H2O, CH2Cl2; g: NaH, an.THF,
appropriate Ar-CH2X, DMF; h: H2, Pd/C, EtOH, EtOH$HCl.
and 9 were about 10-fold more potent than 2 and 10, respectively.
In Fig. 3 are shown the displacement curves of [3H]paroxetine
binding to cortical membranes by diastereoisomeric couple of
compounds 9 (E), 10 (Z).
In the case of Z configuration derivatives, the data reported in
Table 1 showed that only 6 and 10 possess Ki values lower than
100 nM while compounds 2, 4, 8 exhibited Ki values higher than
100 nM. Moreover, an analysis of the biological data shown inTable 1
revealed that in type A compounds, the chlorine atom in the 2
compounds (DAT/SERT ratio ranging from 7 for compound 8 to 118
for compound 6) since they were better inhibitors of [3H]WIN
35,428 binding to DAT.
In order to better define the selectivity profile of the most active
E configuration compounds (compounds 1,3,7 and 9), their ability
to inhibit 3[H]-nisoxetine binding to NET in rat cortical membranes
was also evaluated. The data reported in Table 1 indicate that these
compounds were very weak inhibitors of 3[H]-nisoxetine binding
to NET with Ki values ranging from 9300 for compound 7 to
position on the benzylic group (
b
ring, Fig. 2) was the substituent
19,500
mM for compound 3. A comparison of Ki values of all the
that confers the best affinity for SERT independently of the substit-
uent (fluorine atom or a trifluoromethyl group) on the aromatic ring
directly linked to the piperidine nucleus (a ring, Fig. 2).
compounds for SERT and NET showed that
compounds possessed high selectivity for SERT with the NET/SERT
ratio ranging from y96 for 7 to 972 for 9.
E
configuration
As regards the ability of synthesized compounds (1e10) to
inhibit [3H]WIN 35,428 binding to DAT, Z configuration compounds
2,4,6,8,10 showed Ki values in the micromolar range while those of
E configuration 1,3,5,7,9 were practically inactive (Table 1).
Most of the compounds meet the “druglike” criteria [29] having
MW < 500 and ClogP values (see the Supporting Information) in the
range deemed suitable for penetration of the blood-brain barrier.
Comparison of the Ki values of all the compounds for SERT and
DAT showed that E configuration compounds possessed high
selectivity for SERT with DAT/SERT ratio ranging from y300 for 1 to
>9000 for 3, 7, 9. Such selectivity was reduced for Z configuration
4. Molecular docking
In order to rationalize activity trends, molecular docking of
paroxetine, fluoxetine and compounds 1e10 was performed in the

828
S. Nencetti et al. / European Journal of Medicinal Chemistry 46 (2011) 825e834
Table 1
Inhibition constant values (Ki)a of oxime ethers 1e10, fluoxetine and paroxetine for inhibition of [3H]paroxetine, [3H]-nisoxetine and [3H]WIN 35,428 binding to SERT, NET and
DAT.
R
α
N
O
β
N
1
R
H
Compdb
R
R1
SERTc
DATd
NETe
NET/SERTf
DAT/SERTf
Ki (nM)
Ki (nM)
Ki (nM)
1
2
3
4
5
6
7
8
9
F
F
F
F
F
F
CF3
CF3
CF3
CF3
Fluoxetine
Paroxetine
GBR12909
Nisoxetine
4-F
4-F
E
Z
E
Z
E
Z
E
Z
E
Z
28.74 ꢁ 7.03
305.90 ꢁ 71.01
98.42 ꢁ 17.14
138.61 ꢁ 20.79
22.60 ꢁ 3.53
81.07 ꢁ 5.30
96.82 ꢁ 1 4.80
396.50 ꢁ 17.60
10.28 ꢁ 3.89
96.58 ꢁ 14.80
5.80 ꢁ 2.90
9460 ꢁ 2160
3190 ꢁ 70
>100,000
6200 ꢁ 270
12,900 ꢁ 1040
3800 ꢁ 600
>100,000
13,400 ꢁ 1570
466
198
496
96
329
10
>1000
45
571
47
>1000
7
>9000
43
670
2480
ntg
4-CF3
4-CF3
2-Cl
2-Cl
4-F
4-F
2-Cl
2-Cl
19,500 ꢁ 2100
ntg
11,200 ꢁ 1300
ntg
9300 ꢁ 1000
ntg
2800 ꢁ 300
>100,000
10,000 ꢁ 1200
972
10
4200 ꢁ 700
4000 ꢁ 1700
769h
ntg
609.00 ꢁ 50.00
80.00 ꢁ 10.00
105
258
0.31 ꢁ 0.018
1.5 ꢁ 0.2
3.5 ꢁ 0.8
a
Compounds tested at concentrations from 1 nM to 10 mM. Values represent the mean ꢁ s.e.m. of 3 separate experiments.
Prepared and tested as hydrochloride salts.
b
c
Values determined using [3H]-paroxetine as radioligand.
Values determined using [3H]-WIN 35,428 as radioligand.
Values determined using [3H]-nisoxetine as radioligand.
Ratios of Ki values.
d
e
f
g
h
Not tested.
Ki value represents average of at least two independent experiments.
homology models of SERT, DAT and NET constructed by Ravna et al.
and in particular engages an ionic interaction between the proton-
ated N of piperidine and Asp98 (Fig. 4aec). The piperonilic moiety
interacts with Asn177 and with the non-conserved Thr439, in
analogy with the proposed binding mode of 5-HT [2], that allows the
same ionic interactionwith Asp98 and a hydrogen bond between the
hydroxyl group of 5-HT and Thr439. The fluorophenyl moiety of
paroxetine is stabilized through a stacking with Phe341 and inserted
in a pocket defined by Tyr95, Ala169, Ile172, Val343, apart from
Phe341. Also fluoxetine has a similar binding mode, but the CF3
cannot interact with the same strength of the piperonilic moiety of
paroxetine with Asn177 and Thr439 (Fig. 4c).
[4], using LeuT as a template [1]. Compounds were docked in the
central binding pocket 1 of SERT, DAT and NET, corresponding to the
substrate binding pocket of leucine in the LeuTAa crystal structure,
using the GOLD program. The principal ligandereceptor interactions
were analysed, and the best orientation of the newly synthesized
compounds was compared to the paroxetine one. In accordance with
previously reported docking studies on SERT ligands [30e32],
paroxetine occupies a cavity delimited by TM1, TM3, TM6 and TM8,
As shown in Fig. 4a, the new compounds of E configuration seem
to be able to overlap the orientation of paroxetine in the binding
site, with three points of interaction: the piperidine nitrogen with
Asp98, the R substituent (see Table 1) with Tyr95, and the R1 group
(see Table 1) with Asn177 and Thr439. This binding mode is inde-
pendent of their chirality, with the only exception of compound 3
(yellow in Fig. 4a), whose fluorosubstituted
a ring of the R enan-
tiomer directs away from the paroxetine one.
The presence of a bulkier substituent like CF3 in para position,
when also the second phenyl bears a substituent in para position,
makes the contact unfavourable with the binding site surface. As
shown in Fig. 4d, the CF3 of compounds 3 and 7, nevertheless the
effective orientation in SERT of these compounds, creates a steric
clashwiththe atoms of thecavity walls, whose surfacerepresentedin
Fig. 4d is forced by the threefluoromethyl group. This could explain
the higher affinity of compounds 1, 5 and 9 for SERT, with respect to
compounds 3 and 7. Compounds of Z configuration (Fig. 4b) do not
overlap the paroxetine orientation, because their oximic chains lie
Fig. 3. Displacement of [3H]paroxetine binding to cortical membranes by compounds
and 9 (E), 10 (Z). Membranes were incubated in duplicate with [3H]paroxetine in the
presence and absence of increasing concentrations of each compound as described in
Experimental section. Data points represent the means ꢁ s.e.m. of three independent
experiments.

S. Nencetti et al. / European Journal of Medicinal Chemistry 46 (2011) 825e834
829
Fig. 4. Docking orientations of compounds: a) (R)-1 (cyan), (R)-3 (yellow), (R)-5 (purple), (R)-7 (magenta), (R)-9 (green) and (3S,4R)-paroxetine (grey) into SERT on the left; (S)-1
(cyan), (S)-3 (yellow), (S)-5 (purple), (S)-7 (magenta), (S)-9 (green) and (3S,4R)-paroxetine (grey) into SERT on the right; b) (R)-2 (cyan), (R)-4 (yellow), (R)-6 (purple), (R)-8
(magenta), (R)-10 (green) and (3S,4R)-paroxetine (grey) into SERT on the left; (S)-2 (cyan), (S)-4 (yellow), (S)-6 (purple), (S)-8 (magenta), (S)-10 (green) and (3S,4R)-paroxetine (grey)
into SERT on the right; c) (R)-fluoxetine (orange) compared to (3S,4R)-paroxetine (grey) into SERT; d) (S)-3 (yellow) and (S)-5 (purple) in the SERT binding site, whose Van der Waals
surface is represented in grey; e) 1 (green), as a model compound compared to (3S,4R)-paroxetine (grey) into DAT; f) 1 (green), as a model compound compared to (3S,4R)-
paroxetine (grey) into NET. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article).

830
S. Nencetti et al. / European Journal of Medicinal Chemistry 46 (2011) 825e834
generally outof the binding site. The dockingof compounds 2, 4 and8
is very similar, and independent of their enantiomeric configuration.
The ortho-substituted compounds 6 and 10 have a different orien-
tation compared to the other Z derivatives: in both the R and S
configuration compounds 6 assumes in SERT a collapsed orientation
which allows a partial superposition on the paroxetine. This
compound showed an affinity of 81.07 nM, the best value among the
compounds of Z configuration.
5. Conclusion
In the present study we synthesized a series of E and Z 4-(aryl)
piperidin-3-one O-4-benzyl oxime derivatives (1e10) in order to
obtain new selective SERT inhibitors. The inhibition studies of [3H]
paroxetine, 3[H]-nisoxetine and [3H]WIN 35,428 binding to SERT,
NET and DAT showed that all the compounds, and particularly E
configuration compounds (1,3,5,7,9), showed high SERT binding
affinities (Ki ¼ 10e98 nM) and high SERT selectivities since they are
scarcely or practically inactive towards NET and DAT. These results
are in accordance with the fact that only E configuration compounds
can assume in the SERT binding site a favourable orientation in
analogy with the proposed binding mode of 5-HT [2], which is also
the same for paroxetine, while in DAT and NET the disposition of the
compounds could not allow effective interactions.
Thus, analyzing the docking results, the E configuration of our
oxime ethers and the absence of bulky substituents in para position
of the phenyls seem to be determinant for the affinity towards
SERT. On the contrary, the enantiomeric configuration of
compounds 1e10 seems to have no significant effects on the
interaction of these compounds with the SERT binding site.
As regards the docking into DAT (Fig. 4e), all compounds assume
a different orientation in this transporter. Paroxetine interacts with
Asp79 (corresponding to Asp98 in SERT), and its piperonilic ring
occupies the same cavity as in SERT, but the presence of Phe76 and
Ala423 in place of Tyr95 and Thr439 causes a slight shift of the
ligand towards Phe326 in DAT, thus causing the lack of the
hydrogen bond with the hydroxyl group of the Thr439. Further-
more, the fluorophenyl ring prefers another collocation instead of
the interaction with Ser149, which replaces Ala169 of the SERT and
makes the region less accessible. The fluorophenyl ring lies in
a cavity peculiar to DAT, due to the presence of the non-conserved
residues Val152 and Ala480 which open a larger space compared to
Ile172 and Thr497 of SERT, but cannot allow any effective interac-
tion with the binding site (Fig. 4e). Fig. 4e also represents the
general arrangement of compounds 1e10 in DAT (compound 1 is
reported in figure as an example), which is different from the
paroxetine one. The model compound 1 occupies the same cavities
of the transporter without important interactions with residues
except for the ionic bond with Asp79. This kind of binding mode is
in agreement with the affinity trend of all the synthesized
compounds, characterized by a very low affinity for DAT.
The different substituents introduced on the
moieties played an important role in influencing the potency of
compounds as SERT inhibitors. In particular the docking studies
suggested that the concurrent para-substitution of both
a and b aromatic
a and
b
aromatic moieties with almost one bulkier substituent, makes
unfavourable the contact with the binding site surface determining
a detrimental effect on the affinity. Moreover the enantiomeric
configuration of compounds 1e10 seems to have no significant
effects on the interaction of these compounds with the SERT
binding site.
The most interesting compound in this series is 9 possessing an
affinity for SERT in the same range as fluoxetine and an excellent
SERT selectivity, higher than that of fluoxetine and paroxetine as
well as a ClogP value (see the Supporting Information) in the range
deemed suitable for penetration of the blood-brain barrier.
Furthermore, 9 represents an attractive potential lead to further
optimization by introducing small para substituents on
to engage hydrogen bonds with Asn177 together with small ortho
substituents to better define the SAR of this class of compounds.
b ring able
The third transporter, NET, presents a very high homology with
DAT in the binding site (>80%), and all the residues previously
reported as representative residues for the interaction of DAT with
the ligands are conserved, except for the one corresponding to
Ala423 of DAT and Thr439 of SERT, which in NET is Ser420. The
presence of this serine allows a docking of the ligands similar to the
SERT one (Fig. 4f): the piperonilic moiety of paroxetine engages
a weak interaction with the non-conserved Ser420 but loses the
interaction with Asn153 (Asn177 in SERT) because it shifts about
2.5 Å with respect to its position in the SERT. This different
arrangement is due to the preferred insertion of the paroxetine
fluorophenyl ring between Phe72, lacking in SERT, and Phe323. The
E configuration compounds 1, 3, 5, 7 and 9 are oriented in a similar
manner (Fig. 4f, compound 1 is reported as a model compound), but
could engage only weak polar interactions between the substituent
6. Experimental section
6.1. Chemistry
Melting points were determined on a Köfler hot-stage apparatus
and are uncorrected. IR spectra for comparison between
compounds were recorded with a PerkineElmer mod.1310, as Nujol
mulls in the case of solid substances, or as liquid film in the case of
liquids. 1H NMR spectra were obtained with a Varian Gemini-
200 MHz spectrometer in a ca. 2% solution of CDCl3. Chemical shifts
(d) are reported in parts per million (ppm) downfield from tetra-
methylsilane (TMS) and adjusted according to residual solvent
peaks. Electron impact (EI, 70 eV) mass spectra were obtained on
a HP-5988A mass spectrometer. Reactions were monitored by TLC
on silica gel plates (Merck Silica Gel 60 F254), spots were detected
under UV light. Preparative TLC were performed with silica gel
plates (2 mm, Merck Silica Gel 60 F254). Chromatographic separa-
tions were performed using Merck silica gel (70e230 mesh).
Na2SO4 was always used as the drying agent. Evaporation was
carried out “in vacuo” (rotating evaporator). Elemental analyses
were performed by our analytical laboratory and agreed with
theoretical values to within ꢁ0.4%.
on the phenyl
b ring and Ser420, and between the protonated
amine and Asp79. The presence of Phe72 instead of Tyr95 hampers
further interactions of the substituent on the phenyl
peculiar to the SERT binding mode.
a ring, that are
The dockingperformedinSERT, DATandNET highlightstheroleof
the SERT Thr439 substitution with DAT Ala423 and NET Ser420 in
directing the orientation of the ligands in the binding site and the
importance of the Tyr95/Phe76/Phe72, Ala169/Ser149/Ala145,
Ile172/Val152/Val148, Thr497/Ala480/Ala477 replacements in SERT/
DAT/NET in determininga different stabilization of compounds in the
three transporters. Furthermore, the contribution of Asn177 and/or
Thr439 in the polar interaction with the ligands seems to be deter-
minant for an effective binding in SERT. The lack or the lessening of
these interactionsin DATand NETprecludesanysignificant activityof
the ligands.
6.1.1. 1-Benzyl-4-(4-fluorophenyl)-(12) and 1-benzyl-4-(4-
(trifluoromethyl)phenyl)-(13) piperidin-4-ol
The opportune 4-(aryl)-magnesium bromide was prepared in
the usual manner from 4-fluoro or 4-trifluoromethyl bromo-
benzene (0.015 mol), magnesium turning (0.015 g-atom) and
anhydrous THF (20 mL). The reaction mixture was then diluted
with anhydrous THF (30 mL), refluxed under stirring for 30 min and

S. Nencetti et al. / European Journal of Medicinal Chemistry 46 (2011) 825e834
831
treated dropwise at room temperature with a solution of N-benzyl-
4-piperidone (11) (1.4 g, 0.0075 mol) and anhydrous THF (18 mL).
After total addition, the solution was refluxed under stirring for
15 h, then quenched after cooling by the addition of iced water
(10 mL) and extracted three times with EtOAc. The combined
organic phases were dried and evaporated to yield an oily residue
that was crystallized from hexane to give pure 12 as a pale yellow
solid, or purified by column chromatography on silica gel (EtOAc/
hexane 4:6) to give pure 13 as a white solid. 12: (73%) mp:
atmosphere. A solution of DMSO (5.8 mmol, 0.4 mL) in CH2Cl2
(20 mL) was added dropwise. The reaction mixture was then stirred
for 15 min. A solution of the appropriate alcohol 16 or 17
(2.2 mmol) in CH2Cl2 (20 mL) was added dropwise and then the
reaction mixture was stirred at ꢀ60 ꢂC for 15 min. The reaction
mixture was cooled to ꢀ78 ꢂC and Et3N (8.7 mmol, 1.21 mL) was
added in one portion. The reaction mixture was warmed to room
temperature over 6 h, and then poured into water, mixed, and
extracted three times with Et2O. The organic extracts washed with
water, dried and evaporated gave a crude oil that was purified by
column chromatography on silica gel (EtOAc/hexane 4:6) to give 18
and 19 as solids. 18 and 19 were suddenly converted into hydro-
chloride salts by dissolving in Et2O and treating with Et2O$HCl to
yield pure 18$HCl and 19$HCl. 18$HCl (65%) mp: 170e172 ꢂC; 1H
78e80 ꢂC; 1H NMR [33] (200 MHz, CDCl3):
d
¼ 7.49 (m, 2H), 7.32 (m,
5H), 7.04 (m, 2H), 3.61 (s, 2H), 2.82 (m, 2H), 2.50 (m, 2H), 2.16 (m,
2H), 1.74 ppm (m, 3H); MS (EI, 70 eV) m/z (%): 285 (7); 91 (100)
[M þ H]þ; Anal. calcd for C18H20FNO: C 75.76, H 7.06, N 4.91, found:
C 75.49, H 7.28, N 4.73. 13: (85%) mp: 127 ꢂC; 1H NMR (200 MHz,
CDCl3):
d
¼ 7.63 (m, 4H), 7.32 (m, 5H), 3.64 (s, 2H), 2.86 (m, 2H), 2.52
NMR (200 MHz, CDCl3):
d
¼ 7.27 (m, 5H), 7.07 (m, 2H), 6.98 (m.2H),
(m, 2H), 2.22 (m, 2H), 1.73 ppm (m, 3H); MS (EI, 70 eV) m/z (%): 335
(9); 91 (100) [M þ H]þ; Anal. calcd for C19H20F3NO: C 68.05, H 6.01,
N 4.18, found: C 68.38, H 6.19, N 4.07.
3.62 (s, 2H), 3.52 (m, 1H), 3.27 (m,1H), 3.06e2.57 (m,3H), 2.19 ppm
(m, 2H); IR (KBr):
n
¼ 1725 cmꢀ1; Anal. calcd for C18H19ClFNO: C
76.30, H 6.40, N 4.94, found: C 76.19, H 6.25, N 4.78. 19$HCl (61%)
mp: 173e175 ꢂC; 1H NMR (200 MHz, CDCl3):
d
¼ 7.60 (m, 9H), 3.62
6.1.2. 1-Benzyl-4-(4-fluorophenyl)-(14) and 1-benzyl-4-(4-
(trifluoromethyl)phenyl)-(15) 1,2,3,6-tetrahydropyridine
(s, 2H), 3.50 (m,1H), 3.23 (m, 1H), 3.02e2.57 (m, 3H), 2.17 ppm (m,
2H); IR (KBr):
H 5.44, N 4.20, found: C 68.63, H 5.29, N 4.37.
n
¼ 1720 cmꢀ1; Anal. calcd for C19H19ClF3NO: C 68.46,
The appropriate alcohol 12, 13 (7.3 mmol) was dissolved in anhy-
drous benzene (50 mL) containing PTSA (11 mmol, 2.1 g). The mixture
was heated to reflux with azeotropic removal of water until no further
water was collected. After cooling at room temperature the solution
was washed with saturated aqueous NaHCO3, dried and evaporated to
yield pure 14 and 15 as yellow solids.14: (96%) mp: 48e50 ꢂC; 1H NMR
6.1.5. 1-Benzyl-4-(4-fluorophenyl) (20) and 1-benzyl-4-(4-
(trifluoromethyl)phenyl) (21) piperidin-3-one oxime
To a stirred solution of 18$HCl or 19$HCl (1.2 mmol) in CH2Cl2
(7 mL), was added a solution of hydroxylamine hydrochloride
(1.4 mmol, 0.096 g) and water (8.3 mL). The reaction mixture was
stirred vigorously at room temperature for 24 h. After separation of
the aqueous phase NaHCO3 was added and then extracted with
CH2Cl2. The combined organic layers were evaporated to give a solid
that was crystallized from EtOH, to give 20 and 21 as white solids. 20
[34] (200 MHz, CDCl3):
d
¼ 7.31 (m, 7H), 6.99 (t, J ¼ 8.8 Hz, 2H), 5.98 (m,
1H), 3.62 (s, 2H), 3.14 (m, 2H), 2.72 (t, J ¼ 5.8 Hz, 2H), 2.53 ppm (m, 2H).
IR (KBr):
n
¼ 1620 cmꢀ1; MS (EI, 70 eV) m/z (%): 267 (48); 91 (100)
[M þ H]þ; Anal. calcd for C18H19FN: C 80.87, H 6.79, N 5.24, found: C
80.63, H 6.91, H 5.12. 15: (72%) mp: 87 ꢂC; 1H NMR (200 MHz, CDCl3):
d
¼ 7.55 (m, 4H), 7.39 (m,5H), 6.11 (m,1H), 3.60 (s, 2H), 3.13 (m, 2H), 2.71
(61%) mp: 187e189 ꢂC; 1H NMR (200 MHz, CDCl3)
d
¼ 7.26 (m, 7H),
(t, J ¼ 5.8 Hz, 2H) 2.48 ppm (m, 2H); IR (KBr):
n
¼ 1620 cmꢀ1; MS (EI,
7.0 (m, 2H), 4.12 (d, J ¼ 14 Hz,1H), 3.71, 3.58 (2d, J ¼ 13.2 Hz, 2H), 3.45
(m,1H), 2.81 (d, J ¼ 14 Hz,1H), 2.40 (m, 2H), 2.22e1.98 ppm (m, 2H);
MS (EI, 70 eV) m/z 298 (8%), 91 (100%); Anal. calcd for C18H19FN2O: C
72.46, H 6.42, N 9.39, found: C 72.65, H 6.61, N 9.25. 21 (55%) mp
70 eV) m/z (%): 317 (48); 91 (100). Anal. calcd for C19H18F3NO: C 71.91, H
5.72, N 4.41, found: C 71.73, H 5.57, N 4.26.
6.1.3. 1-Benzyl-4-(4-fluorophenyl) (16) and 1-benzyl-4-(4-
(trifluoromethyl)phenyl) (17) piperidin-3-ol
180e181 ꢂC; 1H NMR
(m, 2H), 3.47 (m, 1H), 2.72 (d, J ¼ 14 Hz, 1H), 2.38e1.89 (m, 4H); MS
(EI, 70 eV) m/z (%): 340 (9%), 91 (100%); Anal. calcd for C19H19F3N2O C
65.51, H 5.50, N 8.04; found: C 65.34, H 5.32, N 8.27.
d
7.30e6.92 (m, 9H), 4.11 (d, J ¼ 14 Hz,1H), 3.60
A 2 M solution of BH3$(CH3)2S complex in anhydrous THF
(14.6 mmol, 7.4 mL) was added dropwise to a cooled (0 ꢂC) stirred
solution of 14 or 15 (4.8 mmol) in anhydrous THF (22 mL). The
mixture was then stirred at room temperature for 3 days. Water
(0.3 mL) was then carefully added, heated at 40 ꢂC, then NaOH 10%
(1.2 mL) and H2O2 36% (1 mL) added and stirred for 1 h. The reac-
tion mixture was then extracted with Et2O and combined organic
layers washed with water and evaporated to yield an oil that was
chromatographed on silica gel column (EtOAc/hexane 4:6) to give
16, 17 as solids. 16: (66%) mp: 89e91 ꢂC; 1H NMR (200 MHz, CDCl3):
6.1.6. General procedure for preparation of N-benzyl oxime ethers
22e26
To a stirred solution of the oxime 20 or 21 (0.73 mmol) in anhy-
drous THF (10 mL) was added NaH 60% (0.78 mmol) under nitrogen
atmosphere. The mixture was stirred for 1 h at room temperature
then evaporated. The oxime sodium salt of 20 or 21 was dissolved in
anhydrous DMF (3.9 mL) and added dropwise under stirring with
a solution of the opportune benzyl chloride (0.78 mmol) in anhy-
drous DMF (2.3 mL). The mixture was then stirred under nitrogen at
50 ꢂC for 5 days then added with water and extracted with CH2Cl2.
The organic extracts were evaporated to yield a crude oil which was
purified bycolumn chromatographyon silica gel (EtOAc/hexane 3:7)
to yield the Z ether isomer 22e26, exclusively.
d
¼ 7.29 (m, 7H), 7.11 (m, 2H), 3.79 (dt, J ¼ 9.8, 4.4 Hz, 1H), 3.63, 3.65
(2d, J ¼ 13.2 Hz, 2H), 3.18 (dd, J ¼ 10.5, 4.4 Hz, 1H), 2.93 (d,
J ¼ 11.2 Hz, 1H), 2.39 (m, 1H), 1.99 (m, 2H), 1.80 ppm (m, 2H); IR
(KBr):
n
¼ 3468 cmꢀ1; MS (EI, 70 eV) m/z (%): 285 (9); 91 (100);
Anal. calcd for C18H20FNO: C 75.76, H 7.06, N 4.91, found: C 75.92, H
6.93, N 4.82. 17: (42%) mp: 105 ꢂC; 1H NMR (200 MHz, CDCl3):
d
¼ 7.61 (m, 2H), 7.35 (m, 7H), 3.90 (m, 1H), 3.66, 3.58 (2d, J ¼ 13 Hz,
2H), 3.21 (dd, J ¼ 10.6, 4.5 Hz, 1H), 2.97 (d, J ¼ 11 Hz, 1H), 2.51 (m,
6.1.6.1. (Z)-1-Benzyl-4-(4-fluorophenyl)piperidin-3-one
robenzyl oxime (22). (80%) 1H NMR (200 MHz, CDCl3)
(m, 13H), 4.91 (s, 2H), 3.76 (d, J ¼ 14 Hz, 1H), 3.59 (m, 2H); 3.52 (m,
1H), 3.07 (d, J ¼ 14 Hz, 1H), 2.78 (m, 1H), 2.46 (m, 1H), 2.12 ppm (m,
2H); Anal. calcd for C25H24F2N2O: C 76.30, H 6.40, N 4.94, found: C
76.51, H 6.63, N 4.72.
O-4-fluo-
¼ 7.33e6.96
1H), 1.99 (m, 2H), 1.86 ppm (m, 2H); IR (KBr):
n
¼ 3465 cmꢀ1; MS
d
(EI, 70 eV) m/z (%): 335 (10); 91 (100); Anal. calcd for C19H20F3NO: C
68.05, H 6.01, N 4.18,. found: C 68.31, H 6.23, N 4.36.
6.1.4. 1-Benzyl-4-(4-fluorophenyl) (18) and 1-benzyl-4-(4-
(trifluoromethyl)phenyl) (19) piperidin-3-one hydrochlorides
A solution of oxalyl chloride (2.9 mmol, 0.25 mL) in CH2Cl2
(13 mL) was cooled to ꢀ78 ꢂC with stirring under a nitrogen
6.1.6.2. (Z)-1-Benzyl-4-(4-(trifluoromethyl)phenyl)piperidin-3-one
O-4-fluorobenzyl oxime (23). (85%); 1H NMR (200 MHz, CDCl3)

832
S. Nencetti et al. / European Journal of Medicinal Chemistry 46 (2011) 825e834
d
¼ 7.64e6.93 (m, 13H), 5.01 (s, 2H), 3.78 (d, J ¼ 13.7, 1H), 3.61 (m,
5.19, N 7.59, found: C 58.71, H, 5.36, N 7.72. 6$HCl 1H NMR
(200 MHz, CDCl3)
J ¼ 15 Hz), 3.67 (m, 1H), 3.56 (d, 1H, J ¼ 15 Hz), 3.15e2.90 (m, 2H),
2.14e2.09 ppm (m, 2H); Anal. calcd for C18H19Cl2FN2O: C 58.55, H
5.19, N 7.59, found: C 58.77, H 5.39, N 7.76.
2H), 3.55 (m, 1H), 3.09 (d, J ¼ 13.7, 1H), 2.80 (m, 1H), 2.49 (m, 1H),
2.12 ppm (m, 2H); Anal. calcd for C26H24F4N2O: C 68.41, H 5.30, N
6.14, found: C 68.63, H 5.19, N 6.27.
d
¼ 7.27e6.99 (m, 8H), 5.12 (s, 2H), 3.99 (d, 1H,
6.1.6.3. (Z)-1-Benzyl-4-(4-fluorophenyl)piperidin-3-one O-2-chlor-
obenzyl oxime (24). (73%) 1H NMR (200 MHz, CDCl3)
d
¼ 7.23 (m,
6.1.7.4. 4-(4-(Trifluoromethyl)phenyl)piperidin-3-one
robenzyl oxime hydrochlorides (7 E, 8 Z). 7$HCl 1H NMR (200 MHz,
CDCl3)
J ¼ 14.5 Hz, 1H), 3.33 (d, J ¼ 14.5 Hz, 1H), 3.15e2.82 (m, 2H),
2.36e1.98 (m, 2H) Anal. calcd for C19H19ClF4N2O: C 56.65, H 4.75, N
6.95, found: C 56.93, H 4.96, N 7.18. 8$HCl 1H NMR (200 MHz,
O-4-fluo-
11H), 6.98 (m, 2H), 5.10 (s, 2H); 3.80 (d, J ¼ 14 Hz, 1H), 3.62 (s, 2H),
3.58 (m, 1H), 3.19 (d, J ¼ 14 Hz, 1H,), 2.79 (m, 1H), 2.51 (m, 1H),
2.10 ppm (m, 2H); Anal. calcd for C25H24FClN2O: C 71.00, H 5.72, N
6.62, found: C 71.19, H 5.59, N 6.77.
d
¼ 7.62e6.98 (m, 8H), 5.16 (s, 2H), 4.79 (m, 1H), 3.50 (d,
6.1.6.4. (Z)-1-Benzyl-4-(4-(trifluoromethyl)phenyl)piperidin-3-one
O-4-fluorobenzyl oxime (25). (66%) 1H NMR (200 MHz, CDCl3)
CDCl3)
d
¼ 7.62e6.96 (m, 8H), 5.04 (s, 2H,), 4.01 (d, J ¼ 15.4 Hz, 1H),
3.66 (m, 1H), 3.58 (d, J ¼ 15.4 Hz, 1H), 3.15e2.88 (m, 2H),
2.17e2.05 ppm (m, 2H). Anal. calcd for C19H19ClF4N2O: C 56.65, H
4.75, N 6.95, found: C 56.87, H 4.91, N 6.74.
d
¼ 7.58e6.99 (m, 13H), 4.91 (2s, 2H), 3.79 (d, J ¼ 14 Hz, 1H,), 3.64 (s,
2H), 3.55 (m, 1H), 3.17 (d, J ¼ 14 Hz, 1H), 2.80 (m, 1H), 2.52 (m, 1H),
2.10 ppm (m, 2H); Anal. calcd for C26H24F4N2O: C 68.41, H 5.30, N
6.14, found: 68.27, H 5.46, 6.29.
6.1.7.5. 4-(4-(Trifluoromethyl)phenyl)piperidin-3-one
obenzyl oxime hydrochlorides (9 E, 10 Z). 9$HCl 1H NMR (200 MHz,
CDCl3)
J ¼ 14.5 Hz, 1H), 3.39 (d, J ¼ 14.5 Hz, 1H), 3.19e2.83 (m, 2H),
2.40e2.03 ppm (m, 2H). Anal. calcd for C19H19Cl2F3N2O: C 54.43, H
4.57, N 6.68, found: C 54.74, H 4.81, N 6.89. 10$HCl 1H NMR
O-2-chlor-
6.1.6.5. (Z)-1-Benzyl-4-(4-(trifluoromethyl)phenyl)piperidin-3-one
O-2-chlorobenzyl oxime (26). (41%) 1H NMR (200 MHz, CDCl3)
d
¼ 7.31e6.97 (m, 8H), 5.08 (s, 2H), 4.77 (m, 1H), 3.64 (d,
d
¼ 7.50e6.95 (m, 13H), 4.93 (s, 2H), 3.80 (d, J ¼ 14 Hz, 1H), 3.64 (s,
2H), 3.57 (m, 1H), 3.13 (d, J ¼ 14 Hz, 1H), 2.80 (m, 1H), 2.52 (m, 1H),
2.11 ppm (m, 2H); Anal. calcd for C26H24F3ClN2O: C 66.03, H 5.12, N
5.92, found: C 65.83, H 5.23, N 5.83.
(200 MHz, CDCl3)
d
¼ 7.28e6.96 (m, 8H), 4.94 (s, 2H), 3.99 (d,
J ¼ 15.4 Hz, 1H), 3.66 (m, 1H), 3.59 (d, J ¼ 15.4 Hz, 1H), 3.19e2.88 (m,
2H), 2.11e2.04 ppm (m, 2H). Anal. calcd for C19H19Cl2F3N2O: C
54.43, H 4.57, N 6.68 found: C 54.71, H 4.39, N 6.87.
6.1.7. General procedure for preparation of the E and Z 4-(aryl)
piperidin-3-one O-4-benzyl oxime ether hydrochlorides 1e10$HCl
To a solution of 22e26 (0.42 mmol) in EtOH anhydrous (25 mL),
was added a solution of EtOH$HCl to pH y 3. The mixture was
shaken under hydrogen at room temperature and atmospheric
pressure for 30 h in the presence of 10% Pd on charcoal (75 mg),
then the catalyst was filtered off and the solution was evaporated to
yield a 1:1.5 mixture of E and Z piperidine hydrochlorides. The pure
E (1,3,5,7,9$HCl) and Z (2,4,6,8,10$HCl) N-unsubstituted piperidine
hydrochlorides were obtained by preparative TLC eluting with
6.2. Molecular docking
Molecular docking was performed in the homology models of
SERT, DAT and NET constructed by Ravna et al. [4], using LeuT as
a template [2]. The ligands were built by means of Maestro [35],
considering the active enantiomers (3S,4R)-paroxetine and (R)-
fluoxetine, and both the enantiomers of compounds 1e10, which
were subjected to a Conformational Search (CS) of 1000 steps in
a implicit water environment using the Macromodel program [36].
The Monte Carlo algorithm was used with the MMFFs forcefield.
The ligands were then minimized using the Conjugated Gradient
method to a convergence value of 0.05 kcal/Å mol, using the same
forcefield and parameters as for the CS.
a
mixture of CH2Cl2:hexane:NEt3 7.5:1.5:1 (1,2,9,10) or
CH2Cl2:hexane:NEt3 6:3:1 (3e8).
6.1.7.1. 4-(4-Fluorophenyl)piperidin-3-one O-4-fluorobenzyl oxime
hydrochlorides (1 E,
2
Z). 1$HCl: 1H NMR (200 MHz, CDCl3)
d
¼ 7.31e6.96(m, 8H), 5.07 (s, 2H), 4.78(m,1H); 3.52 (d, J ¼ 14.4 Hz,1H),
The minimized ligands were docked into the proteins using
GOLD 3.2 [37]; the region of interest in Gold was defined in such
a manner that it contains all the residues which stay within 10 Å
from Asp98 of SERT, Asp79 of DAT and Asp75 of NET. The ‘allow
early termination’ command was deactivated, and a ‘protein
hydrogen bond constraint’ set to the value 15 was used to specify
that the carboxylic oxygen of Asp98/Asp79/Asp75 of SERT, DAT and
NET respectively, should be hydrogen-bonded to the ligand, but
without specifying to which ligand atom. The default Gold
parameters were used for all remaining variables, and ligands were
submitted to 100 Genetic Algorithm runs with the GoldScore
fitness function. The best docked pose for each ligand, obtained by
clustering the results for 1.5 Å of tolerance, was then used for
further studies. Generally at most two or three clusters of solutions
were generated, the first was the most populated, and the score
difference between the best solution and the second cluster one
had a value of about 20. The docking results were visually evaluated
using UCSF Chimera [38].
3.32 (d, J ¼ 14.4 Hz, 1H), 3.05e2.88 (m, 2H), 2.36e1.99 ppm (m, 2H);
Anal. calcd for C18H19ClF2N2O: C 61.28, H 5.43, N 7.94, found C 61.53, H
5.61, N 8.15. 2$HCl: 1H NMR (200 MHz, CDCl3)
d
¼ 7.27e6.94 (m, 8H),
4.91 (s, 2H), 3.93 (d, J ¼ 15.2 Hz, 1H), 3.63 (m, 1H), 3.49 (d, J ¼ 15.2 Hz,
1H), 3.12e2.83 (m, 2H), 2.13e1.97 ppm (m, 2H); Anal. calcd for
C18H19ClF2N2O: C 61.28, H 5.43, N 7.94, found: C 60.07, H 5.25, N 7.75.
6.1.7.2. 4-(4-Fluorophenyl)piperidin-3-one
benzyl oxime hydrochlorides (3 E, 4 Z). 3$HCl 1H NMR (200 MHz,
CDCl3)
J ¼ 14.5 Hz, 1H), 3.32 (d, J ¼ 14.5 Hz, 1H), 3.13e2.85 (m, 2H),
2.38e2.01 ppm (m, 2H); Anal. calcd for C19H19ClF4N2O: C 56.65, H
4.75, N 6.95, found: C 56.44, H 4.57, N 6.76. 4$HCl 1H NMR
O-4-(trifluoromethyl)
d
¼ 7.61e6.97 (m, 8H), 5.16 (s, 2H), 4.80 (m, 1H), 3.48 (d,
(200 MHz, CDCl3)
d
¼ 7.61e6.94 (m, 8H), 5.03 (s, 2H), 3.99 (d,
J ¼ 15 Hz, 1H), 3.66 (m, 1H), 3.56 (d, J ¼ 15 Hz, 1H), 3.13e2.87 (m,
2H), 2.16e2.09 ppm (m, 2H); Anal. calcd for C19H19ClF4N2O: C
56.65, H 4.75, N 6.95, found: C 56.84, H 4.93, N 7.12.
6.1.7.3. 4-(4-Fluorophenyl)piperidin-3-one O-2-chlorobenzyl oxime
hydrochlorides (5 E,
6.3. Binding studies
6
Z). 5$HCl 1H NMR (200 MHz, CDCl3)
d
¼ 7.27e7.01 (m, 8H), 5.25 (s, 2H), 4.82 (m, 1H), 3.56 (d, 1H,
6.3.1. Animals
J ¼ 14.5 Hz), 3.35 (d, 1H, J ¼ 14.5 Hz), 3.12e2.89 (m, 2H),
2.40e2.05 ppm (m, 2H); Anal. calcd for C18H19Cl2FN2O: C 58.55, H
Cerebral tissue was from adult New Zealand White rabbits
(4e5 kg) obtained from
a commercial source (Charles River

S. Nencetti et al. / European Journal of Medicinal Chemistry 46 (2011) 825e834
833
Laboratories, Inc., Wilmington, MA). Animals were maintained in
standard laboratory conditions and feeding in sawdust-lined cages
and a 12 h light/dark cycle. They were killed by intravenous injec-
tion of a lethal dose of pentobarbital. All procedures conformed to
the guidelines of the International European ethical standards for
the care and use of laboratory animals. All protocols were approved
by the Ethical Deontological Committee for animal experimenta-
tion of the University of Pisa.
protein) in 50 mM TriseHCl buffer, pH 7.4, containing 300 mM NaCl
and 5 mM KCl with 1 nM [3H]nisoxetine (specific activity, 80 Ci/
mmol; PerkineElmer Life Science, Waltham, MA) in a final volume
of 0.5 mL. Incubation was carried out at 4 ꢂC for 4 h. Non-specific
binding was defined in the presence of 10 mM desipramine. Specific
binding was obtained by subtracting non-specific binding from
total binding and approximated to 85e90% of total binding. The
binding reaction was concluded by filtration through Whatman GF/
C glass-fiber filters using a Brandel Harvester (see above). Filters
were washed four times with 5 mL of the ice-cold binding buffer
and placed in vials with 4 mL of a scintillation cocktail. Radioac-
6.3.2. Affinity for SERT, NET, DAT
For SERT binding assays, membranes were prepared from rabbit
frontal
cortex
and
[3H]paroxetine
(specific
activity,
tivity was measured by means of a b-counter (see above). The
15e20 Ci mmolꢀ1
;
PerkineElmer Life Science, Waltham, MA)
affinity of [3H]nisoxetine for NET was assessed by saturation
experiments and Kd value of 1.5 ꢁ 0.07 nM (n ¼ 3) was obtained.
Stock solutions (1 mM) of the test compounds were prepared in
ethanol and then diluted in TriseHCl saline buffer at the required
concentration. Competition binding assays were performed with at
least seven different concentrations of the test compounds.
binding was performed as previously described [39]. The affinity of
[3H]-paroxetine for SERT was assessed by saturation experiments
obtaining a dissociation constant (Kd) value of 56.00 ꢁ 8.00 pM
(n ¼ 3).
Membranes used in DAT binding assays were prepared from
frozen rabbit striatal by homogenization with a polytron homoge-
nizer in a phosphate buffered saline solution (PBS) (25 mM
Na2HPO4/NaH2PO4, 48 mM NaCl, pH 7.7 at 4 ꢂC) containing 320 mM
sucrose. The homogenate was centrifuged at 46,000g for 10 min at
4 ꢂC. The supernatant was discarded and the pellet resuspended by
diluting 1:100 (w:v) in the homogenization buffer. The pellet was
recentrifuged at 46,000g for 10 min and the pellet was resuspended
diluted 1:100 in PBS containing sucrose, divided in 1.0 mL aliquots
and stored frozen at ꢀ80 ꢂC. On the day of the experiment, one or
more aliquots were quickly thawed at 37 ꢂC, centrifuged at 46,000g
for 10 min at 4 ꢂC and the pellet was resuspended in the binding
assay buffer at a 1:200 (w:v) dilution.
6.3.3. Analysis of data
Saturation data were analysed and fitted by the non-linear
regression analysis of the GraphPad Prism (Version 3.00) computer
program (GraphPad Software). The calculated dissociation constant
(Kd) values for [3H]-paroxetine and [3H]-nisoxetine binding to
rabbit cortical membranes were 56.00 ꢁ 8.00 pM (n ¼ 3) and
1.50 ꢁ 0.07 nM (n ¼ 3), respectively, while the calculated Kd value
for [3H]-WIN 35,428, binding to rabbit striatal membranes was
6.10 ꢁ 0.30 nM (n ¼ 3). The inhibition curves of the synthesized
compounds were analysed and fitted by the non-linear regression
analysis of the GraphPad Prism (Version 3.00) computer program
(GraphPad Prism, Inc., San Diego, CA). The derived IC50 values were
used to calculate the inhibition constants (Ki) by the Cheng and
Prusoff equation [28].
Cortical membranes for NET binding assays were prepared by
homogenizing freshly dissected cerebral cortex in 30 vols of ice-cold
50 mM TriseHCl buffer, pH 7.4, containing 120 mM NaCl and 5 mM
KCl. The homogenate was centrifuged at 46,000g for 10 min at 4 ꢂC.
The resulting pellet was suspended in 30 vols of 50 mM TriseHCl
buffer, pH 7.4, containing 120 mM NaCl and 5 mM KCl, incubated at
37 ꢂC for 10 min to remove endogenous norepinephrine and
centrifuged at 46,000g for 10 min at 4 ꢂC. This washing procedure
was repeated twice. The resulting pellet was immediately used in
the binding assay or frozen at ꢀ80 ꢂC until the time of the assay.
Protein concentration was determined according to the method
of Lowry et al. [40] using BSA as standard.
Ki values are presented as the means ꢁ s.e.m. of three inde-
pendent experiments performed in duplicate.
Acknowledgements
The authors thank Aina W. Ravna that generously provided the
SERT, NET, DAT models.
Appendix A. Supplementary material
For DAT binding assays, striatal membranes (0.2 mg of protein)
were incubated in 0.5 mL of a TriseHCl saline buffer (50 mM
TriseHCl, pH 7.4, 120 mM NaCl, 5 mM KCl) with 0.5e2 nM [3H]-
WIN 35,428 (specific activity, 84.5 Ci mmolꢀ1; PerkineElmer Life
Science, Waltham, MA) for 2 h at 4 ꢂC. Non-specific binding was
defined in the presence of 10 mM dopamine and was subtracted
from total binding to obtain specific binding. Ligand bound to the
transporter was separated from free ligand by filtration using a 30-
well manifold (Brandel Harvester, Brandel, Gaithersburg, MD) with
glassefiber filters (GF/C) pre-soaked for 2 h at room temperature
with 0.6% (w:v) polyethylenimine. Filters were quickly washed
three times with ice-cold 50 mM TRISeHCl, pH 7.4, placed into
scintillation vials and soaked overnight in 3 mL of Cytoscint ES (MP
Biomedicals Solon, OH). The following day, samples were read by
Supplementary material associated with this article can be
found, in the online version, at doi:10.1016/j.ejmech.2010.12.018.
References
[1] S.K. Yamashita, T. Singh, Y. Jin Kawate, E. Gouaux, Nature 437 (2005)
215e223.
ꢀ
[2] W. Ravna, M. Jaronczyk, I. Sylte, Bioorg. Med. Chem. Lett. 16 (2006) 5594e5597.
ꢀ
[3] M. Jaronczyk, Z. Chilmonczyk, A.P. Mazurek, G. Nowak, A.W. Ravna,
K. Kristiansen, I. Sylte, Bioorg. Med. Chem. 16 (2008) 9283e9294.
[4] A.W. Ravna, I. Sylte, S.G. Dahl, J. Mol. Model. 15 (2009) 1155e1164.
[5] S.G. Buttler, M.J. Meegan, Curr. Med. Chem. 15 (2008) 1737e1761.
[6] L. Iversen, Brit. J. Pharmacol. 147 (2006) 582e588.
[7] P.R. Finely, Ann. Pharmacother 28 (1994) 1359e1369.
[8] M.J. Owens, J. Clin. Psychiatry 65 (Suppl. 4) (2004) 5e10.
[9] P. Pacher, V. Kecskemeti, Curr. Med. Chem. 11 (2004) 925e943.
[10] D. Spinks, G. Spinks, Curr. Med. Chem. 9 (2002) 799e810.
[11] L.L. Iversen, R.A. Glennon., Antidepressants. in: D.J. Abraham (Ed.), Burger’s
Medicinal Chemistry and Drug Discovery, sixth ed., Nervous System Agents,
vol. 6 Wiley, New York, 2003, pp. 483e524.
[12] M.D. Dorsey, S.E. Lukas, S.L. Cunningham, Neuropsychopharmacology 14
(1996) 437e442.
[13] S.A. Montgomery, D.S. Baldwin, A. Riley, J. Affect. Disord. 69 (2002)
119e140.
[14] S.H. Kennedy, Eur. Neuropsychopharmacol. 16 (2006) S619eS624.
[15] S.M. Stahl, J. Clin. Psychiatry 59 (1998) 343e344.
scintillation spectroscopy in
a b-counter (PerkineElmer Life
Science, Waltham, MA). The affinity of [3H]-WIN 35,428 for DAT
was assessed by saturation experiments obtaining a Kd value of
6.1 ꢁ0.3 nM (n ¼ 3).
For NET binding assays, [3H]nisoxetine binding was performed
essentially as described by Tejani-Butt et al. [41]. The cortical
membrane pellet was resuspended in 50 mM TriseHCl buffer, pH
7.4, containing 300 mM NaCl and 5 mM KCl. The binding assay was
performed incubating aliquots of membranes (0.2e0.3 mg of

834
S. Nencetti et al. / European Journal of Medicinal Chemistry 46 (2011) 825e834
[16] M.J. Owens, W.N. Morgan, S.J. Plott, C.B. Nemeroff, J. Pharmacol. Exp. Therap.
283 (1997) 1305e1322.
[31] L. Celik, S. Sinning, K. Severinsen, C.G. Hansen, M.S. Møller, M. Bols, O. Wiborg,
B. Schiøtt, J. Am. Chem. Soc. 130 (2008) 3853e3865.
[17] M.J. Owens, D.L. Knight, C.B. Nemeroff, Biol. Psychiatry 47 (2000) 842e845.
[18] M.J. Owens, D.L. Knight, C.B. Nemeroff, Biol. Psychiatry 50 (2001) 345e350.
[19] R.W. Lam, Int. J. Psych. Clin. Practice 8 (Suppl. 1) (2004) 25e29.
[20] F. Chen, M.B. Larsen, C. Sànchez, O. Wiborg, Eur. Neuropsychopharmacol.
15 (2005) 193e198.
[32] S.M. Cloonan, J.J. Kating, S.G. Butler, A.J. Knox, A.M. Jørgensen, G.H. Peters,
D. Rai, D. Corrigan, D.G. Lloyd, D.C. Williams, M.J. Meegan., Eur. J. Med. Chem.
44 (2009) 4862e4888.
[33] B. Wenzel, D. Sorger, K. Heinitz, M. Scheunemann, R. Schliebs, J. Steinbach,
O. Sabri, Eur. J. Med. Chem. 40 (2005) 1197e1205.
[21] S. Schreiber, C.G. Pick, Eur. Neuropsychopharmacol. 16 (2006) 464e468.
[22] S.M. Stahl, Int. J. Psych. Clin. Practice 8 (Suppl. 1) (2004) 3e10.
[23] S. Källströ, R. Leino, Bioorg. Med. Chem. 16 (2008) 601e605.
[24] A. Balsamo, A. Lapucci, M. Macchia, B. Macchia, A. Martinelli, S. Nencetti,
A. Rossello, T. Cristina, Il Farmaco 49 (1994) 77e82.
[34] C. Morrill, N.S. Mani, Org. Lett. 9 (2007) 1505e1508.
[35] Maestro, ver. 7.5. Schrödinger Inc., Portland, OR, 1999.
[36] Macromodel, ver. 8.5. Schrödinger Inc., Portland, OR, 1999.
[37] G. Jones, P. Willett, R.C. Glen, A.R. Leach, R.J. Taylor, J. Mol. Biol. 267 (1997)
727e748.
[25] A. Balsamo, A. Lapucci, A. Lucacchini, M. Macchia, C. Martini, C. Nardini,
S. Nencetti, Eur. J. Med. Chem. 29 (1994) 967e973.
[38] C. Huang, G.S. Couch, E.F. Pettersen, T.E. Ferrin, Pac. Symp. Biocomput. 1
(1996) 724.http://www.cgl.ucsf.edu/chimera.
[26] Our unpublished results.
[39] S. Nencetti, G.C. Demontis, M.R. Mazzoni, L. Betti, I. Banti, A. Rossello,
A. Lapucci, J. Pharm. Pharmacol. 59 (2007) 1439e1445.
[40] O.H. Lowry., N.J. Rosenbrough, A.L. Farr, R.J. Randall, J. Biol. Chem. 193 (1951)
265e267.
[41] S.M. Tejani-Butt, D.J. Brunswick, A. Frazer, Eur. J. Pharm. 191 (1990)
239e243.
[27] G.J. Karabatsos, N. Hsi, Tetrahedron 23 (1967) 1079e1095.
[28] Y. Cheng, W.H. Prusoff, Biochem. Pharmacol. 22 (1973) 3099e3108.
[29] A. Lipinski, F. Lombardo, B.W. Dominy, P.J. Feeney, Adv. Drug Deliv. Rev. 23
(1997) 3e25.
[30] S. Tavoulari, L.R. Forrest, G. Rudnick, J. Neurosci. 29 (2009) 9635e9643.