Identification of a novel selective serotonin reuptake inhibitor by coupling monoamine transporter-based virtual screening and rational molecular hybridization

Article abstract of DOI:10.1021/cn200044x

Ligand virtual screening (VS) using the vestibular binding pocket of a three-dimensional (3-D) monoamine transporter (MAT) computational model followed by in vitro pharmacology led to the identification of a human serotonin transporter (hSERT) inhibitor with modest affinity (hSERT K_i= 284 nM). Structural comparison of this VS-elucidated compound, denoted MI-17, to known SERT ligands led to the rational design and synthesis of DJLDU-3-79, a molecular hybrid of MI-17 and dual SERT/5-HT_1Areceptor antagonist SSA-426. Relative to MI-17, DJLDU-3-79 displayed 7-fold improvement in hSERT binding affinity and a 3-fold increase in [³H]-serotonin uptake inhibition potency at hSERT-HEK cells. This hybrid compound displayed a hSERT:hDAT selectivity ratio of 50:1 and a hSERT:hNET (human norepinephrine transporter) ratio of >200:1. In mice, DJLDU-3-79 decreased immobility in the tail suspension test comparable to the SSRI fluvoxamine, suggesting that DJLDU-3-79 may possess antidepressant properties. This proof of concept study highlights MAT virtual screening as a powerful tool for identifying novel inhibitor chemotypes and chemical fragments for rational inhibitor design.

Full text of DOI:10.1021/cn200044x

RESEARCH ARTICLE
pubs.acs.org/acschemicalneuroscience
Identification of a Novel Selective Serotonin Reuptake Inhibitor by
Coupling Monoamine Transporter-Based Virtual Screening and
Rational Molecular Hybridization
†
,‡
,†
§
||
†
‡
Tammy L. Nolan, David J. Lapinsky,* Jeffery N. Talbot, Martín Indarte, Yi Liu, Sankar Manepalli,
†
§
§
,‡
,†
Laura M. Geffert, Mary Ellen Amos, Phillip N. Taylor, Jeffry D. Madura,* and Christopher K. Surratt*
†
Division of Pharmaceutical Sciences, Mylan School of Pharmacy, Duquesne University, 600 Forbes Avenue, Pittsburgh,
Pennsylvania 15282, United States
‡
Departments of Chemistry and Biochemistry, Center for Computational Sciences, Duquesne University, 600 Forbes Avenue,
Pittsburgh, Pennsylvania 15282, United States
§
Department of Pharmaceutical and Biomedical Sciences, Raabe College of Pharmacy, Ohio Northern University, 525 South Main
Street, Ada, Ohio 45810, United States
Institute of Biotechnology and Biomedicine, Universitat Aut ꢀo noma de Barcelona, 08193 Bellaterra, Spain
S Supporting Information
b
ABSTRACT: Ligand virtual screening (VS) using the vestibular
binding pocket of a three-dimensional (3-D) monoamine
transporter (MAT) computational model followed by in vitro
pharmacology led to the identification of a human serotonin
transporter (hSERT) inhibitor with modest affinity (hSERT
Ki = 284 nM). Structural comparison of this VS-elucidated
compound, denoted MI-17, to known SERT ligands led to the rational design and synthesis of DJLDU-3-79, a molecular hybrid of
MI-17 and dual SERT/5-HT1A receptor antagonist SSA-426. Relative to MI-17, DJLDU-3-79 displayed 7-fold improvement in
3
hSERT binding affinity and a 3-fold increase in [ H]-serotonin uptake inhibition potency at hSERT-HEK cells. This hybrid
compound displayed a hSERT:hDAT selectivity ratio of 50:1 and a hSERT:hNET (human norepinephrine transporter) ratio of
>200:1. In mice, DJLDU-3-79 decreased immobility in the tail suspension test comparable to the SSRI fluvoxamine, suggesting that
DJLDU-3-79 may possess antidepressant properties. This proof of concept study highlights MAT virtual screening as a powerful
tool for identifying novel inhibitor chemotypes and chemical fragments for rational inhibitor design.
KEYWORDS: monoamine, neurotransmitter transporter, virtual screening, molecular docking, molecular hybridization, depression
he plasma membrane monoamine transporter (MAT) pro-
teins (the dopamine transporter (DAT), serotonin trans-
many patients and whose use is often limited by adverse effects. As
examples, citalopram (Celexa) is associated with insomnia, parox-
etine (Paxil) is well-known for somnolence and weight gain, sertra-
line (Zoloft) is especially problematic in causing gastrointestinal
disturbances, and fluvoxamine (Luvox) has been associated with
T
porter (SERT), and norepinephrine transporter (NET)) are relevant
to central nervous system (CNS)-based maladies including substance
1
2
abuse and addiction, orthostatic hypotension, attention-deficit
3
4
5
hyperactivity disorder (ADHD), Lesch-Nyhan disease, autism,
dangerous drug interactions. All SSRIs have been linked to sexual
6
,7
8
9À12
depression, and obsessive-compulsive and other anxiety disorders.
MATs represent pivotal targets for readily recognized medications
such as methylphenidate (Ritalin), fluoxetine (Prozac), and amitripty-
line (Elavil). The development of drugs targeting MATs has been via
traditional ligand-based (usually substrate analogue) approaches. Few
chemical scaffolds have been identified; this shortcoming provides
motivation to develop new tools that enable the discovery of
structurally novel lead compounds. The present work illustrates the
rational transformation of a DAT virtual screening (VS) hit ligand to a
serotonin-selective reuptake inhibitor (SSRI) through a molecular
hybridization approach.
dysfunction.
The adverse effects of these drugs are generally
linked to secondary interactions with one or more of the 14 types of
serotonin receptors. Development of an SSRI structurally dissimilar
to those currently available could reduce the likelihood of undesir-
able interactions with these receptors.
The DAT, SERT, and NET are primarily responsible for the
respective clearance of dopamine, serotonin, and norepinephrine
from the synaptic cleft. A sodium gradient is used to power
neurotransmitter uptake by these MATs back into the presynap-
tic neuron, terminating stimulation of postsynaptic receptors.
With respect to the armamentarium of antidepressant drugs,
another SSRI therapeutic may appear to be unnecessary; however,
SSRI drugs currently on the market are characterized by inefficacy in
Received: May 5, 2011
Accepted: June 8, 2011
Published: June 08, 2011
r 2011 American Chemical Society
544
dx.doi.org/10.1021/cn200044x
|ACS Chem. Neurosci. 2011, 2, 544–552

ACS Chemical Neuroscience
RESEARCH ARTICLE
À
One MAT substrate molecule, a Cl ion, and either one or two
+
Na ions are transported across the plasma membrane through a
symport method. The SERT additionally requires antiport move-
+
13À15
ment of one K ion.
The MATs are members of the
neurotransmitter/sodium symporter (NSS) family, which includes
bacterial and eukaryotic transporters. NSS proteins share 20À30%
16
sequence identity and are characterized by 12 transmembrane
domains (TM) and intracellular N- and C-termini. Crystallization of
17
the leucine transporter LeuT, a bacterial NSS member, has
provided a template for generation of three-dimensional (3-D)
MAT computer models for use in structureÀfunction studies.
LeuT-based MAT homology models have proven useful in guiding
site-directed mutagenesis, studies on substrate translocation me-
18À21
chanism, and elucidation of ligand binding sites.
Two discrete ligand binding sites within the MAT proteins
have been established. The primary substrate site, S1, is at the
approximate midpoint of the lipid bilayer and is flanked by gating
residues that alternate access between the intracellular and extra-
17,19,22,23
cellular regions.
A second substrate site, S2, is to the
extracellular side of the external gate of S1, in what is termed the
2
0,22,24,25
extracellular vestibule.
The substrate occupies S2 before its
24,25
translocation to the interior S1 site.
MAT inhibitors have been
22,26,27
21,28À30
associated with the S1
and S2
1
sites and, apparently, the
3
region between the sites. SSRIs and psychostimulants are most
often associated with the S1 site and the tricyclic antidepressants
with the S2 site or general vestibular region, but exceptions have
26,30
been reported.
To date, no LeuT protein-inhibitor ligand cocrystal structures
have been obtained in which the inhibitor occupies the S1 site,
adding to the difficulty in modeling MATÀinhibitor interactions.
Consequently, de novo, unbiased docking of inhibitors to MAT
computational models of the type utilized in ligand VS is currently
limited to the extracellular vestibule. Screening of thousands to
millions of potential ligands in a structural library using a receptor
molecular model is a powerful, rapid, and inexpensive tool for
32,33
discovery of lead compounds.
Recently, the vestibular pocket of
a MAT computational model was employed for VS of a chemical
34
library, identifying novel inhibitor chemotypes. The work de-
scribed herein couples the VS process with traditional medicinal
chemistry drug design to yield a novel SSRI.
’
RESULTS AND DISCUSSION
The vestibular binding pocket of a previously described LeuT-
34
based rDAT homology model was used for VS of the entire
ENAMINE virtual screening collection, a database containing ap-
proximately 3 million compounds. Utilizing MOE (Molecular
35
Operating Environment) Affinity dG scoring and visual inspection
that focused on optimizing intramolecular interactions, 10 com-
pounds from the database were selected and purchased for initial
hMAT pharmacological evaluation. These VS “hits” (Figure A in the
Supporting Information) were tested at 10 μM final concentration in
one-point binding inhibition assays that displaced the radiolabeled
Figure 1. Pharmacological one-point (10 μM) binding assay of virtual
3
screening hits at (A) hDAT-N2A using [ H]-WIN 35,428 and mazindol
125
(
10 μM) for nonspecific binding, (B) hSERT-HEK293 using [ I]-RTI-
5
5 and paroxetine (10 μM) for nonspecific binding, and (C) hNET-
125
3
125
HEK293 using [ I]-RTI-55 and desipramine (10 μM) for nonspecific
binding. Data represent n = 3 independent experiments. **P < 0.002 vs total
binding in that assay; ***P < 0.0001 vs total binding in that assay.
cocaine analogues [ H]-WIN 35,428 (hDAT) or [ I]-RTI-55
hSERT, hNET). Three of the 10 VS compounds, coded MI-15,
(
MI-17, and MI-20, displayed significant radioligand binding inhibi-
tion of 40% or more at all three MATs (Figure 1). Tetrahydropyr-
idinyl compound MI-17 ((()-1, Figure 2) was the most potent
MAT inhibitor in the screening assay and was more thoroughly
characterized.
affinity was 4.6-fold higher than that for hDAT and 19-fold higher
than that for hNET (Table 1). The potency of MI-17 in
3
inhibiting uptake of cognate [ H]-neurotransmitter was over
MI-17 affinity at hSERT, measured by displacement of radi-
olabeled cocaine analogue, was a modest 284 nM; the hSERT
3-fold higher at hSERT than at hDAT and hNET. Given the
sequence similarity among the three hMATs, the discovery of a
5
45
dx.doi.org/10.1021/cn200044x |ACS Chem. Neurosci. 2011, 2, 544–552

ACS Chemical Neuroscience
RESEARCH ARTICLE
modest hSERT ligand by VS using a DAT homology model was
not surprising. Interestingly, SciFinder Scholar reveals no primary
literature associated with MI-17 ((()-1); rather, it is simply available
in milligram quantities from limited chemical sources. A litera-
Briefly, epoxide (()-4 was prepared in 97% yield by benzylating
racemic epichlorohydrin with 2-chlorobenzyl alcohol analogous to
42
that described by Sun et al. The epoxide was then opened with
43
3-(1,2,3,6-tetrahydropyridin-4-yl)-1H-indole (5) similar to that
36
44
ture survey of known SERT inhibitors indicated that SSA-426
2, Figure 2), a Wyeth Pharmaceuticals dual SERT/5-HT receptor
described by Chan et al. to provide target (()-3 in good yield
(
(95%). Pharmacological evaluation of (()-3 validated the rational
attempt to modify VS hit compound MI-17 to improve hSERT
selectivity, binding affinity, and substrate uptake inhibition potency
(Table 1). While only a 1.6-fold decrease in hDAT binding affinity
was observed for (()-3 relative to lead compound (()-1, hSERT
affinity increased 6.8-fold with the same modification. Thus, hSERT
selectivity over hDAT increased from 4.6-fold to 50-fold. Addition-
1A
37
antagonist, has high affinity for rat SERT (K = 2.34 ( 0.59 nM).
The tetrahydropyridine indole portion of 2 appears to be important
for SERT binding, as several high-affinity SERT ligands containing
i
3
8À40
this chemical signature have been reported.
Apart from the conserved tetrahydropyridine ring and 2,3-
dioxypropyl N-substituent (boxed regions of Figure 2), compounds
125
3
1
and 2 appear structurally dissimilar. However, flexible alignment
ally, inhibition of [ I]-RTI-55 binding and [ H]-norepinephrine
illuminated their chemical similarities and a common 3-D arrange-
ment of certain structural features (Figure 3a). In the top-scoring low
energy conformation alignment, the tetrahydropyridine moieties
were aligned as were the phenol and indole rings, while the alcohol
and ether oxygen atoms of 1 overlapped the dioxane ether oxygen
atoms of 2. Additional docking studies were carried out to further
explore the possible ligand-binding conformations of these com-
pounds within hSERT (Figure 3b). In agreement with alignment
studies, the top-scoring poses suggested that (()-1 and 2 occupy the
vestibular pocket in similar orientations. In particular, the phenol of
uptake at hNET were undetectable for (()-3 (K and IC values >
i
50
10 μM), rendering an hSERT to hNET selectivity of over 200-fold.
The in vitro SERT-selective properties of DJLDU-3-79 ((()-
3) were next tested for translatability to an in vivo model. The tail
suspension test (TST) is a well-established prototypical rodent
45,46
model of learned helplessness
that is strongly correlated to
drugs that possess antidepressant-like activity in humans. This
assay is based in part on findings that drugs that possess antide-
pressant activity in humans reduce the immobility time of the
animal in this test. Because of their in vitro characteristics ana-
logous to therapeutically relevant antidepressants that block
uptake of monoamine neurotransmitters involved in the regula-
tion of mood, (()-1 and (()-3 were screened in mice for
antidepressant-like effects in the TST.
(
R)-1 and the indole of 2 docked deep in the vestibular pocket,
20,24
essentially in the S2 substrate site,
while the variable tetrahy-
dropyridine N-substituents were accessible to the extracellular
environment (Figure 3c and d).
Taking into account the suggested binding modes and common
structural motifs for (()-1 and 2, and the established importance of
Mice treated with (()-3 produced a dose-dependent decrease
in immobility time in the TST that was comparable in effect to
the SSRI fluvoxamine and the norepinephrine selective reuptake
inhibitor desipramine (Figure 4). Compound (()-3 did not alter
general locomotor behavior (Figure 5A). These data suggest that
(()-3 could have potent and efficacious antidepressant-like
activity. In contrast, (()-1 dose-dependently increased immo-
bility times in the TST in two different preparations: 2% DMSO
vehicle (Figure 4C) or 55% DMSO vehicle in the oxalate form
36
indoles in SERT recognition, DJLDU-3-79 ((()-3) was rationally
41
designed as a molecular hybrid of MI-17 ((()-1) and SSA-426 (2)
(
Figure 2). The hypothesis was that replacing the phenol in (()-1
with the indole in 2 would improve hSERT selectivity and affinity.
The synthesis of hybrid compound (()-3 is depicted in Scheme 1.
(
data not shown). Unlike (()-3, (()-1 significantly altered
spontaneous activity and gross behaviors (Figure 5B), particu-
larly at doses greater than 3 mg/kg, which potentially accounts
for increased immobility observed in the TST. Thus, (()-1 has
potentially debilitating effects on animal activity and gross
behavior such that its use in behavioral-based paradigms may
be contraindicated.
To date, there has been only one report of a MAT model-
34
based VS effort that yielded active compounds, in part due to
the complexity of obtaining discrete transporter conformations
at high resolution. Even reports of ligand-based VS approaches
are few. Considering that MAT inhibitors serve to increase
synaptic levels of monoamine neurotransmitter, modulation of
Figure 2. Structural comparison of virtual screening hit MI-17 ((()-1)
with dual SERT/5HT1A receptor antagonist SSA-426 (2), leading to the
rational design of DJLDU-3-79 ((()-3) via molecular hybridization.
47,48
Table 1. MI-17 ((+)-1) and DJLDU-3-79 ((+)-3) Apparent hMAT Binding Affinities and Neurotransmitter Uptake Inhibition
a
Potencies
binding affinity (K
i
, nM)
selectivity ratio
SERT/DAT SERT/NET
substrate uptake inhibition potency (IC50, nM)
b
c
d
b
c
d
DAT
SERT
NET
DAT
SERT
NET
e
e
e
(
()-1
()-3
1298 ( 36
284 ( 66
5300 ( 1341
4.6
50
19
3774 ( 407
3436 ( 245
1167 ( 26
441 ( 62
3764 ( 446
e,f
(
2129 ( 177
37 ( 4
>10 000
>200
>10 000
a
Values are mean ( standard error for 3À5 independent experiments, each performed in duplicate. Stably transfected hDAT-N2A, hSERT-HEK293,
b
3
3
c
125
and hNET-HEK293 cells were used. K
5
hDAT in the same assay. P < 0.05 for (()-1 vs (()-3 in the binding assay.
i
determined using [ H]-WIN 35,428; IC50 determined using [ H]-dopamine. K
i
determined using [ I]-RTI-
3
d
125
3
e
5; IC50 determined using [ H]-serotonin. K determined using [ I]-RTI-55; IC50 determined using [ H]-norepinephrine. P < 0.05 for hSERT vs
i
f
5
46
dx.doi.org/10.1021/cn200044x |ACS Chem. Neurosci. 2011, 2, 544–552

ACS Chemical Neuroscience
RESEARCH ARTICLE
Figure 3. (a) Flexible alignment showing the similar 3-D arrangement for low-energy conformations of (()-1 (cyan, R-enantiomer pictured) and 2
(
orange). (b) High-scoring docking poses of (()-1 (cyan, R-enantiomer pictured) and 2 (orange) in the vestibular S2 binding site of hSERT (beige
cylinders) are superposed; the S1 substrate pocket is defined by a serotonin molecule (magenta). (c) Residues predicted for the hSERT vestibular S2
pocket. Residues within 4.5 Å of (+)-1 (cyan) and 2 (orange) are pictured. The color code indicates residues identical among the MATs and LeuT
(green), residues shared among only the MATs (yellow), and nonidentical residues (gray). TM 11 was removed in this panel to facilitate viewing of the
relevant side chains. (d) Comparison of predicted S2 pocket residues among hSERT, hDAT, hNET, and LeuT. Colors correspond to panel (c). The
local sequence identities for hSERT are 23% between hSERT and LeuT, 38% between hSERT and hDAT, and 42% between hSERT and hNET. Figures
63
were generated using PyMOL.
Scheme 1. Synthesis of MI-17/SSA-426 Hybrid Compound DJLDU-3-79 ((()-3)
postsynaptic monoamine receptors is another therapeutic
avenue. Starting with rhodopsin, findings from two decades of high
resolution structureÀfunction studies of the G protein coupled
validation using VS of a structural database enriched with classic
54,56
ligands for each receptor.
In a unique and intriguing app-
roach, molecular models based on the adrenergic receptor crystal
49
receptors (GPCRs) are available. Just as crystallization of LeuT
launched credible MAT computational modeling, the success in
obtaining X-ray structures of the β and β adrenergic recep-
structures have been used to map potentially druggable allosteric
57
pockets of the receptor. As more is elucidated connecting
specific monoamine receptor subtypes to depression and other
CNS disease states, it is anticipated that VS-guided lead com-
pound discovery will regularly employ a combination of mono-
amine transporter and receptor targets.
1
2
5
0,51
tors
and a dopamine D3 receptorÀantagonist drug cocrys-
5
2
tal has stimulated especially reliable model building of the
monoamine GPCRs and their subsequent use as VS tools. Novel
5
-HT2A and 5-HT2C ligands have been identified via VS, with the
In summary, virtual screening of the vestibular (S2) binding site
of a LeuT-guided DAT computational model provided hMAT
inhibitor MI-17 ((()-1), possessing modest binding affinity for
hDAT, hSERT, and hNET. Subsequent computational chemistry
53À55
latter including compounds with low nanomolar affinity.
Active conformations of adrenergic and serotonergic GPCRs
have been modeled using molecular dynamics followed by
5
47
dx.doi.org/10.1021/cn200044x |ACS Chem. Neurosci. 2011, 2, 544–552

ACS Chemical Neuroscience
RESEARCH ARTICLE
Figure 4. Behavioral effects of MI-17 ((()-1) and DJLDU-3-79 ((()-3) in the tail suspension test. (a) Acute administration (30 min) of the selective
reuptake inhibitors of serotonin, fluvoxamine (FLVX; 10 mg/kg, i.p.), and norepinephrine, desipramine (DSP; 3 mg/kg, i.p.), decreased immobility
times compared to water vehicle-treated (Veh) control animals (n = 7À9). (b) DJLDU-3-79 dose-dependently decreased immobility times compared to
DMSO (100%; 0.002 mL/g) vehicle-treated (Veh) animals, indicating antidepressant-like effects (n = 8À14). (c) In contrast, MI-17 dose-dependently
increased immobility times compared to DMSO (2%; 0.01 mL/g) vehicle-treated (Veh) animals (n = 8À10). Data are presented as the mean ( SEM
and were analyzed by one-way ANOVA with Tukey’s posthoc test with **P < 0.01, ***P < 0.001 vs the appropriate vehicle control.
Figure 5. Locomotor effects of DJLDU-3-79 ((()-3) and MI-17 ((()-1). Locomotor activity was measured in animals 30 min following
administration of (a) DJLDU-3-79 or (b) MI-17 and the appropriate DMSO vehicle: 100% at 0.002 mL/g or 2% at 0.01 mL/g, respectively. Behavioral
responses were measured as the distance traveled (centimeters) and ambulatory time and resting time (seconds). Data are presented as the mean ( SEM
(
n = 6À10), and were analyzed by one-way ANOVA with Tukey’s posthoc test with **P < 0.01 vs vehicle-treated (Veh) control.
efforts and structural comparison of this VS hit to known SERT
ligands led to the rational design and synthesis of a molecular hybrid
with improved in vitro and in vivo hSERT pharmacological proper-
ties. This work highlights the utility of transporter-based virtual
screening as a tool for identifying novel MAT inhibitor chemotypes
and chemical fragments that may be coupled with traditional
medicinal chemistry, toward developing clinical candidates for a
host of CNS-based disorders.
addition of explicit hydrogen atoms, and rebalancing of protonation
states by deprotonating strong acids and/or protonating strong bases at
pH 7. The most feasible tautomer for each compound was retained as
guided by the MOE software in-house “rules”. Compounds with
unconstrained chiral centers were removed from the databases (to
ensure chiral purity) as well as compounds with reactive, mutagenic,
and/or carcinogenic functional groups beyond what is stipulated by
Lipinski’s Rule of Five. These operations were performed using the
sdfilter and sdwash svl commands within MOE. Finally, the SDF files
with 2D coordinates were imported into a MOE molecular database,
partial charges were added, and energy minimized using the Merck
molecular force field MMFF94x with a conjugated gradient/truncated
Newton optimization algorithm (convergence criterion = 0.05 kcal/mol,
ε = 1). These druglike, nonreactive, and enantiomerically pure databases
containing one low energy conformation for each compound were
employed in the docking experiments. The chemical structures and
Affinity dG values for compounds MI-11 through MI-20 can be found in
the Supporting Information (Figure A).
5
8
’
METHODS
3
3
3
Materials. [ H]-WIN 35,428, [ H]-dopamine, [ H]-serotonin,
3
125
[
(
H]-norepinephrine, and [ I]-RTI-55 were purchased from PerkinElmer
Foster City, CA). (()-1 was purchased from Enamine (Kiev, Ukraine).
The N2A-hDAT cell line was a gift from Dr. Margaret Gnegy. HEK293
cells stably transfected with hNET or hSERT were prepared in collabo-
ration with Dr. Mads Larsen and Dr. Susan Amara (University of Pittsburgh,
Pittsburgh, PA).
Molecular Modeling. DAT Virtual Screening. Generation of the
DAT homology model used in the virtual screening has been previously
Flexible Alignment of Ligands. Flexible alignment of (()-1 and 2 was
performed using a stochastic conformational method in which the
conformational space of each ligand is assessed while simultaneously
exploring the alignment space of the compounds. Partial charges were
calculated using MMFF94x prior to the conformational search. The
unconstrained stereogenic center of (()-1 was free to invert, affording
both (R)- and (S)-enantiomers. For the alignment, an iteration limit =
200, a failure limit = 20, and the default atom-based similarity terms were
used. A score was assigned to each alignment based on the average strain
20
described.3 Virtual screening was carried out in MOE as previously
4
described, with the exception that the entire Enamine virtual screening
collection was curated and employed. An account was created at the
Enamine Web site (http://www.enamine.net/). High-throughput
screening virtual libraries 1 to 6 and Advanced Collections 1 and 2 were
downloaded as SDF files. Compounds were subjected to a “wash”
procedure that took into account removal of counterions and salts,
5
48
dx.doi.org/10.1021/cn200044x |ACS Chem. Neurosci. 2011, 2, 544–552

ACS Chemical Neuroscience
RESEARCH ARTICLE
PO , 5.6 mM glucose)
energy and a similarity measure that considers molecular feature overlap.
An energy cutoff = 15 kcal/mol and an rmsd tolerance for heavy atoms =
1.3 mM CaCl
2
, 1.2 mM MgSO
4
, 1.2 mM KH
2
4
supplemented with 50 μM ascorbic acid (KRH/AA). Cells were
3
0.5 Å were set in order to retain only low energy alignments sufficiently
incubated for 15 min at room temperature with 1 nM [ H]-WIN
1
25
distinct from previously generated alignments. Six alignments were
generated ranging in score from À68.277 to À59.613; lower values
indicate a better alignment.
SERT Modeling. The primary sequence of hSERT was downloaded
from the Uniprot database (P31645) and along with LeuTAa (2a65)
35,428 for hDAT or 1 nM [ I]-RTI-55 for hNET supplemented with
tropolone along with increasing concentrations of drug or 10 μM
mazindol (hDAT) or 10 μM desipramine (hNET) (total volume of
500 μL). Following incubation, cells were washed twice with 1 mL of
KRH/AA buffer and then treated with 1 mL of 1% SDS with gentle
shaking at room temperature for 1 h. Cell lysates were transferred into
5 mL of scintillation fluid for radioactivity analysis using a liquid
scintillation analyzer. For saturation binding assays, data were analyzed
with GraphPad Prism 5.0 software. IC50 values were generated and
converted to K values using the Cheng-Prusoff equation (K = IC /(1+
5
9
loaded into Discovery Studio (DS) 2.5.1. The sequence alignment
60
described by Celik et al. was employed with minor modifications
Figure B in the Supporting Information). Manual modifications
(
included movement of the gaps in the TM3 and EL2 loop region, in
EL2 loop, in beta strand 2, and in TM7. These modifications increased
overall sequence identity and produced a more favorable phi-psi plot.
Models were generated using several different alignments, and the
Figure B alignment was found to produce the best hSERT model. This
model varied relative to the Beuming et al. alignment principally in
terms of EL4, which is thought to be important to the S2 binding site.
Specifically, EL4 of the Beuming-based model projected into the S2 site,
hindering ligand occupancy. The “build homology” protocol available in
DS was used to construct 20 models of hSERT. Based on the lowest
discrete optimized potential energy (DOPE) score generated by DS,
which evaluates the relative stability of a model by comparing it to other
generated models, five models were picked for further refinement. The
stereochemical quality of these models was evaluated using the PDB
validation server. From this analysis, two of the five models were
evaluated further for their fitness in a 3-D environment by using the
i
i
50
([RTI-55]/K RTI-55)).
d
hSERT Membrane Pharmacology. hSERT binding affinities
were obtained by displacement of [ I]-RTI-55 in membrane binding
assays. Membranes were prepared from stable HEK293-hSERT cells
1
25
1
6
grown at 37 °C in a 5% CO
environment on 150 Â 25 mm plates. At
2
95% confluence (3 days of growth), cells were washed twice with 10 mL
of cold phosphate-buffered saline (DPBS). An additional 10 mL of
DPBS was added, and cells were harvested by scraping and transferred to
cold centrifuge tubes (15 mL), followed by centrifuging for 10 min at
low speed (700g). After removal of the supernatant, the cell pellet was
resuspended in 500 μL of cold TE buffer (50 mM Tris, 1 mM EDTA,
pH 7.5). Following centrifuging for 30 min at 100 000g at 4 °C, the
supernatant was discarded and the pellet was frozen for later use or
resuspended in ice-cold binding buffer (50 mM Tris, pH 7.5, 100 mM
NaCl). Each sample was analyzed for protein content using the Bradford
protein assay. For competition binding, membranes were incubated
“
verify protein” (Profiles-3D) protocol. The Profiles-3D score calculated
for a model should be close to the high score generated, indicating that
the model has a reasonable fold. The chosen model had a score
comparable to the high score and good spatial overlap with LeuTAa
Following the addition of hydrogen atoms and partial charges to the
model, hydrogen atoms were minimized, keeping heavy atoms fixed to
remove unwanted steric effects.
SERT Docking. Conformations were enumerated for each compound
using a stochastic method followed by docking into the hSERT
vestibular pocket using MOE. A wall constraint (radius = 7 Å) encom-
passing alpha spheres placed by the MOE Site-Finder tool was used to
define the site for docking. Docking into the hSERT model was carried
out using the Proxy Triangle placement method and a force field
refinement. Docking poses were scored using the Affinity dG scoring
method of MOE.
1
25
at room temperature with [ I]-RTI-55 (0.1 nM concentration) and
increasing concentrations of cold competitor (1 fM to 1 μM) or 10 μM
paroxetine to measure nonspecific binding. Reactions were carried out in
12 Â 75 mm borosilicate glass tubes with gentle shaking at room
temperature for 1 h and terminated by rapid filtration through GF/B
filters (Schleicher and Schuell, Keene, NH) presoaked in 0.5% poly-
ethylenimine solution (v/v). Filters were washed twice with 5 mL of cold
50 mM Tris buffer and transferred to vials. Radioactivity was determined
using a Beckman gamma counter. Data were analyzed with GraphPad
.
Prism 5.0 software. IC50 values were generated and converted to K
values using the Cheng-Prusoff equation (K = IC /(1+ ([RTI-55]/K
d
i
i
50
RTI-55)).
Substrate Uptake Inhibition Assays. A substrate uptake in-
hibition screening assay was performed in parallel with the one-point
competition-binding assay. Stably transfected N2A-hDAT, HEK293-
hNET, and HEK293-hSERT cells were grown to >90% confluence on
24-well plates. Cells were washed and then preincubated with inhibitor
(10 μM) for 10 min at room temperature, followed by addition of 10 nM
In Vitro MAT Binding Screen of VS “Hit” Compounds.
Compounds identified from the virtual screening were dissolved in
100% DMSO to a concentration of 10 mM, and an initial one-point
competition binding assay was conducted at 10 μM final concentration.
Competition binding was performed using stably transfected N2A-
3
3
3
3
hDAT and HEK293-hNET whole cells with [ H]-WIN 35,428 and
[ H]-dopamine, [ H]-norepinephrine, or [ H]-serotonin. After 5 min,
cells were washed twice then treated with 5% SDS and shaken for 1 h.
Lysates were transferred to 5 mL of scintillation fluid, and radioactivity
was detected using a liquid scintillation analyzer. Nonspecific radioli-
gand binding was determined using 10 μM concentrations of mazindol,
desipramine, and paroxetine for hDAT, hNET, and hSERT, respectively.
1
25
[
I]-RTI-55, respectively, while hSERT binding was examined using
stably transfected HEK293-hSERT membrane (described below) with
1
25
[
I]-RTI-55. Nonspecific radioligand binding was determined using
0 μM concentrations of mazindol, desipramine, and paroxetine for
1
hDAT, hNET, and hSERT, respectively. Screening results were analyzed
with one-way ANOVA (P < 0.05) with a posthoc Dunnett’s multiple
comparison test.
hNET and hDAT Whole Cell Pharmacology. Compounds
were evaluated using in vitro radioligand binding assays at all three
transporters. Nonspecific radioligand binding was determined using
mazindol, desipramine, and paroxetine for hDAT, hNET, and hSERT,
respectively. Whole-cell competition binding assays were performed for
hDAT and hNET using stable N2A-hDAT or HEK293-hNET cell lines
3
VS compounds demonstrating the ability to decrease net uptake of [ H]-
neurotransmitter in the screening assay were characterized further using
a range of concentrations, typically 0.01À10 000 nM, to determine IC50
values (GraphPad Prism 5, La Jolla, CA).
General Experimental Approach for Organic Synthesis.
Reaction conditions and yields were not optimized. All reactions were
performed in flame-dried glassware under argon unless otherwise noted.
All solvents and chemicals were purchased from Aldrich Chemical Co. or
Fisher Scientific and used without further purification. Flash column
2
grown at 37 °C in a 5% CO environment. Cell monolayers were grown
in 24-well plates to >90% confluence. Cells were washed twice with 1 mL
of KRH buffer (25 mM HEPES, pH 7.3, 125 mM NaCl, 4.8 mM KCl,
61
chromatography was performed according to the method of Still et al.
using Fisher S826-25 silica gel sorbent (70-230 mesh) and eluting
5
49
dx.doi.org/10.1021/cn200044x |ACS Chem. Neurosci. 2011, 2, 544–552

ACS Chemical Neuroscience
RESEARCH ARTICLE
solvent mixtures as specified. Thin-layer chromatography (TLC) was
performed using TLC silica gel 60 F254 plates obtained from EMD
Chemicals, Inc., and compounds were visualized under UV light and/or
followed the National Institute of Health guidelines outlined in “Using
Animals in Intramural Research”.
62
Mouse Tail Suspension Test. As previously described, mice
were individually suspended by the tail from a metal bar elevated 30 cm
using adhesive tape. Behavior was videotaped for 6 min and videos were
later scored for immobility time (seconds) by a blinded observer.
Locomotor Activity. Locomotor activity was assessed as
1
13
I
2
stain. Proportions of solvents used for TLC are by volume. H and
C
NMR spectra were recorded on either a Bruker 400 or 500 MHz
1
13
spectrometer. Chemical shifts for H and C NMR spectra are reported
as parts per million (δ ppm) relative to tetramethylsilane (0.00 ppm) as
an internal standard. Coupling constants are measured in Hz. HRMS
samples were analyzed at Old Dominion University (Norfolk, VA) by
positive ion electrospray on a Bruker 12 T APEX -Qe FTICR-MS
6
2
described in a noninvasive manner using a Opto-Varimax 4 activity
analyzer (Columbus Instruments, Columbus, OH) that measures
animal activity in a cage intersected with photocells projecting infrared
beams 2.5 cm apart and 2 cm above the floor. After a brief habituation
period, animals were treated with either drug or the appropriate vehicle
and activity was measured for 20 min.
1
13
instrument with an Apollo II ion source. On the basis of H and
C
NMR, all compounds were >95% pure.
Synthesis of Molecular Hybrid (()-DJLDU-3-79 ((()-3). A
suspension of 2-chlorobenzyl alcohol (1.4 g, 10 mmol, Aldrich), n-
Bu
13 mL) was initially stirred at room temperature and then cooled to
°C. Racemic epichlorohydrin (3.1 mL, 40 mmol, Aldrich) was then
added dropwise via syringe. The resulting mixture was stirred at 0 °C for
0 min, then the ice bath was removed, and the reaction was allowed to
4
NBr (160 mg, 0.5 mmol, Acros), and 40% (w/v) aq. NaOH solution
’
ASSOCIATED CONTENT
(
0
S
Supporting Information. Figure A: Virtual screening hits
b
MI-11 through MI-20. The Affinity dG docking score is indicated
below each structure in the figure. Figure B: LeuTÀhSERT
polypeptide sequence alignment used in the generation of the
hSERT model. This material is available free of charge via the
Internet at http://pubs.acs.org.
3
stir at room temperature overnight. After 18 h, the mixture was poured
onto ice with the aid of H O and Et O, and then extracted. The organic
2
2
layer was dried (MgSO
4
), filtered, and concentrated to provide 1.91 g
1
(97% yield) of benzyl ether (()-4 as a light brown oil. H NMR
indicated the material was sufficiently pure to take directly on to the next
’
AUTHOR INFORMATION
reaction without further purification. Characterization data for (()-4:
1
R = 0.42 (EtOAc/hexanes, 2:8). H NMR (CDCl , 400 MHz): δ 7.50
f
3
Corresponding Author
(
(
3
dd, 1H, J = 2 Hz, 7.6 Hz), 7.35 (dd, 1H, J = 1.6 Hz, 7.6 Hz), 7.29À7.21
*
(D.J.L.) Mailing address: Division of Pharmaceutical Sciences,
Mylan School of Pharmacy, Duquesne University, 422B Mellon Hall,
00 Forbes Avenue, Pittsburgh, PA 15282. E-mail: lapinskyd@duq.
m, 2H), 4.68 (q, 2H, J = 12.8 Hz), 3.85 (dd, 1H, J = 3.2 Hz, 11.2 Hz),
.52 (dd, 1H, J = 6 Hz, 12 Hz), 3.23 (m, 1H), 2.83 (dd, 1H, J = 4.8 Hz, 4
6
13
Hz), 2.66 (dd, 1H, J = 2.8 Hz, 4.8 Hz). C NMR (CDCl , 100 MHz): δ
1
3
edu. (J.D.M.) Mailing address: Departments of Chemistry and
Biochemistry, Duquesne University, 312 Mellon Hall, 600 Forbes
Avenue, Pittsburgh, PA 15282. E-mail: madura@duq.edu (C.K.S.)
Mailing address: Division of Pharmaceutical Sciences, Mylan School
of Pharmacy, Duquesne University, 411 Mellon Hall, 600 Forbes
Avenue, Pittsburgh, PA 15282. E-mail: surratt@duq.edu.
35.7, 132.9, 129.3, 129.1, 128.8, 126.8, 71.4, 70.3, 50.8, 44.3. HRMS
+
calculated for (C10
tion of tetrahydropyridine 5 (589 mg, 3 mmol) and epoxide (()-4
393 mg, 2 mmol) in DMSO (14 mL) was heated at 77 °C for 5 h.
2
H11ClO )2Na 419.0787, found 419.0794. A solu-
43
(
During the course of the reaction, the mixture changed from a yellow to
orange solution. After cooling to room temperature, the mixture was
poured into H
washed with brine, dried (MgSO
tographed (95:3:2, CHCl /MeOH/Et
()-3 as a brown oil. Characterization data for (()-3: R = 0.36 (CHCl :
2
O then extracted with EtOAc. The organic layer was
), filtered, concentrated, and chroma-
N) to provide 750 mg (95%) of
Author Contributions
T.L.N. conducted SERT and NET pharmacological assays,
performed molecular modeling studies, and wrote the first draft
of the manuscript. D.J.L. devised the molecular hybridization
strategy, carried out all organic synthesis, and wrote portions of
the manuscript. J.N.T. wrote a portion of the manuscript and
designed and oversaw all animal studies; L.M.G., M.E.A., and P.
N.T. collaborated on these studies. M.I. carried out ligand virtual
screening using the DAT computational model he generated
(reported in ref 20) to provide the compounds MI-10 through
MI-20. Y.L. conducted initial DAT and SERT pharmacologic
screening of VS hits, DAT pharmacology of DJLDU-3-79, and
SERT pharmacology of MI-17. S.M. created the SERT molecular
model and contributed to manuscript writing. J.D.M. oversaw
and contributed to the computational model building and
contributed to the manuscript. C.K.S. launched the efforts to
create the computational models, guided and consulted on
pharmacologic assays and data interpretation, and wrote portions
of the manuscript.
4
3
3
(
f
3
1
MeOH, 9:1). H NMR (CDCl , 400 MHz): δ 8.32 (br s, 1H), 7.88 (d,
3
1
7
3
3
2
1
1
H, J = 8.4 Hz), 7.51 (dd, 1H, J = 2.0 Hz, 7.6 Hz), 7.38À7.34 (m, 2H),
.29À7.12 (m, 5H), 6.19 (t, 1H, J = 3.2 Hz), 4.69 (s, 2H), 4.07 (m, 1H),
.62 (qd, 2H, J = 4.4 Hz, 11.6 Hz), 3.40 (dd, 1H, J = 3.2 Hz, 16.4 Hz),
.22 (dd, 1H, J = 3.2 Hz, 16.4 Hz), 2.92 (m, 1H), 2.73 (m, 1H),
13
.68À2.56 (m, 5H). C NMR (CDCl , 100 MHz): δ 136.8, 135.9,
3
33.0, 129.8, 129.3, 129.2, 128.7, 126.8, 125.1, 122.2, 121.4, 120.6, 120.1,
18.8, 117.7, 111.4, 73.3, 70.5, 66.5, 60.5, 53.2, 41.0, 28.9. HRMS
+
calculated for (C23
H25Cl
1
N
2
O
2
)H 397.1677, found 397.1682.
Behavioral Conditions. At weaning, male C57BL/6J mice were
group housed (six animals per cage) with same-sex littermates, in a
temperature- and humidity-controlled vivarium, under a 12 h light/dark
cycle (lights on at 7:00 a.m.). Behavioral testing was conducted during
the light phase of the light/dark cycle between 9:00 a.m. and 12:00 pm.
Animals used in behavioral studies were typically between 10 and 24
weeks of age and were naive to all behavioral treatments prior to testing.
Room luminance was maintained between 70 and 75 foot candles for all
behavioral studies. Unless specified otherwise, drugs were administered
intraperitoneally (i.p.) 30 min prior to behavioral testing. Sterile water
was the vehicle used with fluvoxamine and desipramine; the vehicle
DMSO at 2% (0.01 mL/g) or 100% (0.002 mL/g) was employed with
MI-17 or DJLDU-3-79, respectively. All experimental procedures were
approved by the local Institutional Animal Care and Use Committee and
Funding Sources
This work was supported by NIDA Grants DA026530 (to C.K.S.)
and DA027806 (to J.D.M.), Department of Education equipment
Grants P116Z050331 and P116Z080180 (to J.D.M.), and the
Bower, Bennett, and Bennett Endowed Chair Research Award
(to J.N.T.), and by a Duquesne University Faculty Development
Fund grant (to D.J.L).
5
50
dx.doi.org/10.1021/cn200044x |ACS Chem. Neurosci. 2011, 2, 544–552

ACS Chemical Neuroscience
RESEARCH ARTICLE
’
ACKNOWLEDGMENT
(14) McElvain, J. S., and Schenk, J. O. (1992) A multisubstrate
mechanism of striatal dopamine uptake and its inhibition by cocaine.
Biochem. Pharmacol. 43, 2189–2199.
We thank Drs. Mads Larsen and Susan Amara for collabora-
tion in preparing the hSERT and hNET HEK cell lines. M.I.
would like to thank Drs. Jack Smith and Jaime Hargrave (Univer-
sity of Texas—Houston, School of Biomedical Informatics) for
fellowship support.
(15) Rudnick, G. (2002) In Neurotransmitter Transporters, Structure,
Function, and Regulation (Reith, M. E. A., Ed.), 2nd ed., pp 25À52,
Humana Press, Totowa, NJ.
(16) Beuming, T., Shi, L., Javitch, J. A., and Weinstein, H. (2006) A
comprehensive structure-based alignment of prokaryotic and eukaryotic
neurostransmitter/Na+ symporters (NSS) aids in the use of the LeuT
structure to probe NSS structure and function. Mol. Pharmacol.
’
ABBREVIATIONS
7
0, 1630–1642.
MAT, monoamine transporter protein; hDAT, plasma mem-
brane human dopamine transporter protein; hSERT, plasma
membrane human serotonin transporter protein; hNET, plasma
membrane human norepinephrine transporter protein; SSRI,
selective serotonin reuptake inhibitor; VS, virtual screening;
TST, tail suspension test
(17) Yamashita, A., Singh, S. K., Kawate, T., Jin, Y., and Gouaux, E.
+
À
(
2005) Crystal structure of a bacterial homologue of Na /Cl -depen-
dent neurotransmitter transporters. Nature 437, 215–223.
18) Andersen, J., Olsen, L., Hansen, K. B., Taboureau, O., Jorgensen,
(
F. S., Jorgensen, A. M., Bang-Andersen, B., Egebjerg, J., Stromgaard, K., and
Kristensen, A. S. (2009) Mutational mapping and modeling of the binding
site for (S)-citalopram in the human serotonin transporter. J. Biol. Chem.
2
85, 2051–2063.
’
REFERENCES
(19) Forrest, L. R., Zhang, Y. W., Jacobs, M. T., Gesmonde, J., Xie, L.,
(
1) Amara, S., and Sonders, M. S. (1998) Neurotransmitter trans-
porters as molecular targets for addictive drugs. Drug Alcohol Depend.
1, 87–96.
2) Hahn, M. K., Robertson, D., and Blakely, R. D. (2003) A
Honig, B. H., and Rudnick, G. (2008) Mechanism for alternating access in
neurotransmitter transporters. Proc. Natl. Acad. Sci. U.S.A. 105, 10338–10343.
(20) Indarte, M., Madura, J. D., and Surratt, C. K. (2008) Dopamine
transporter comparative molecular modeling and binding site prediction
using the LeuTAa leucine transporter as a template. Proteins 70, 1033–1046.
(21) Sarker, S., Weissensteiner, R., Steiner, I., Sitte, H. H., Ecker,
G. F., Freissmuth, M., and Sucic, S. (2010) The high-affinity binding site
for tricyclic antidepressants resides in the outer vestibule of the
serotonin transporter. Mol. Pharmacol. 78, 1026–1035.
(22) Beuming, T., Kniazeff, J., Bergmann, M. L., Shi, L., Gracia, L.,
Raniszewska, K., Newman, A. H., Javitch, J. A., Weinstein, H., Gether, U.,
and Loland, C. J. (2008) The binding sites for cocaine and dopamine in
the dopamine transporter overlap. Nat. Neurosci. 11, 780–789.
(23) Boudker, O., and Verdon, G. (2010) Structural perspectives on
secondary active transporters. Trends Pharmacol. Sci. 31, 418–426.
(24) Shi, L., Quick, M., Zhao, Y., Weinstein, H., and Javitch, J. A.
(2008) The mechanism of a neurotransmitter:sodium symporter—
inward release of Na+ and substrate is triggered by substrate in a second
binding site. Mol. Cell 30, 667–677.
5
(
mutation in the human norepinephrine transporter gene (SLC6A2)
associated with orthostatic intolerance disrupts surface expression of
mutant and wild-type transporters. J. Neurosci. 23, 4470–4478.
(
3) Bowton, E., Saunders, C., Erreger, K., Sakrikar, D., Matthies,
H. J., Sen, N., Jessen, T., Colbran, R. J., Caron, M. G., Javitch, J. A.,
Blakely, R. D., and Galli, A. (2010) Dysregulation of dopamine
transporters via dopamine D2 autoreceptors triggers anomalous dopa-
mine efflux associated with attention-deficit hyperactivity disorder.
J. Neurosci. 30, 6048–6057.
(
4) Wong, D. F., Harris, J. C., Naidu, S., Yokoi, F., Marenco, S.,
Dannals, R. F., Ravert, H. T., Yaster, M., Evans, A., Rousset, O., Bryan,
R. N., Gjedde, A., Kuhar, M. J., and Breese, G. R. (1996) Dopamine
transporters are markedly reduced in Lesch-Nyhan disease in vivo. Proc.
Natl. Acad. Sci. U.S.A. 93, 5539–5543.
(
5) Prasad, H. C., Steiner, J. A., Sutcliffe, J. S., and Blakely, R. D.
2009) Enhanced activity of human serotonin transporter variants asso-
ciated with autism. Philos. Trans. R. Soc. London Biol. Sci. 364, 163–173.
6) Hahn, M. K., and Blakely, R. D. (2002) Monoamine transporter
(
(25) Zhao, Y., Terry, D. S., Shi, L., Quick, M., Weinstein, H.,
Blanchard, S. C., and Javitch, J. A. (2011) Substrate-modulated gating
dynamics in a Na(+)-coupled neurotransmitter transporter homologue.
Nature 474, 109À113.
(
gene structure and polymorphisms in relation to psychiatric and other
complex disorders. Pharmacogenomics J. 2, 217–235.
(
Feng, T., Chang, A. S., Ganapathy, V., and Blakely, R. D. (1993)
Antidepressant- and cocaine-sensitive human serotonin transporter”
Molecular cloning, expression, and chromosomal localization. Proc.
Natl. Acad. Sci. U.S.A. 90, 2542–2546.
(26) Sinning, S., Musgaard, M., Jensen, M., Severinsen, K., Celik, L.,
Koldso, H., Meyer, T., Bols, M., Jensen, H. H., Schiott, B., and Wiborg,
O. (2010) Binding and orientation of tricyclic antidepressants within the
central substrate site of the human serotonin transporter. J. Biol. Chem.
285, 8363–8374.
7) Ramamoorthy, S., Bauman, A. L., Moore, K. R., Han, H., Yang-
(27) Thompson, B. J., Jessen, T., Henry, L. K., Field, J. R., Gamble,
K. L., Gresch, P. J., Carneiro, A. M., Horton, R. E., Chisnell, P. J.,
Belova, Y., McMahon, D. G., Daws, L. C., and Blakely, R. D. (2011)
Transgenic elimination of high-affinity antidepressant and cocaine
sensitivity in the presynaptic serotonin transporter. Proc. Natl. Acad.
Sci. U.S.A. 108, 3785–3790.
(28) Singh, S. K., Yamashita, A., and Gouaux, E. (2007) Antidepres-
sant binding site in a bacterial homologue of neurotransmitter trans-
porters. Nature 448, 952–956.
(29) Zhou, Z., Zhen, J., Karpowich, N. K., Goetz, R. M., Law, C. J.,
Reith, M. E. A., and Wang, D. N. (2007) LeuT-desipramine structure
reveals how antidepressants block neurotransmitter reuptake. Science
317, 1390–1393.
(30) Zhou, Z., Zhen, J., Karpowich, N. K., Law, C. J., Reith, M. E. A.,
and Wang, D.-N. (2009) Antidepressant specificity of serotonin trans-
1p 6o ,r t6e 5r 2s –u 6g 5g 7e s. ted by three LeuTÀSSRI structures. Nat. Struct. Mol. Biol.
(
8) Ozaki, N., Goldman, D., Kaye, W. H., Greenberg, B. D., Lappa-
lainen, J., Rudnick, G., and Murphy, D. L. (2003) Serotonin transporter
missense mutation associated with a complex neuropsychiatric pheno-
type. Mol. Psychiatry 8, 933–936.
(
9) Ferguson, J. M. (2001) SSRI antidepressant medications: Adverse
effects and tolerability. Prim. Care Compan. J. Clin. Psychiatry 3, 22–27.
10) Kennedy, S. H., Andersen, H. F., and Lam, R. W. (2006)
(
Efficacy of escitalopram in the treatment of major depressive disorder
compared with conventional selective serotonin reuptake inhibitors and
venlafaxine XR: a meta-analysis. J. Psychiatry Neurosci. 31, 122–131.
(
11) Preskorn, S. H. (1995) Comparison of the tolerability of
buproprion, fluoxetine, imipramine, nefazodone, paroxetine, sertraline,
and venlafaxine. J. Clin. Psychiatry 56 (Suppl 6), 12–21.
(
12) Schellander, R., and Donnerer, J. (2010) Antidepressants:
Clinically relevant drug interactions to be considered. Pharmacology
6, 203–215.
13) Gu, H., Wall, S. C., and Rudnick, G. (1994) Stable expression of
8
(
(31) Andersen, J., Taboureau, O., Hansen, K. B., Olsen, L., Egebjerg,
J., Stromgaard, K., and Kristensen, A. S. (2009) Location of the
antidepressant binding site in the serotonin transporter: Importance
biogenic amine transporters reveals differences in inhibitor sensitivity,
kinetics, and ion dependence. J. Biol. Chem. 269, 7124–7130.
5
51
dx.doi.org/10.1021/cn200044x |ACS Chem. Neurosci. 2011, 2, 544–552

ACS Chemical Neuroscience
RESEARCH ARTICLE
of Ser-438 in recognition of citalopram and tricyclic antidepressants.
J. Biol. Chem. 284, 10276–10284.
(49) Tebben, A. J., and Schnur, D. M. (2011) Beyond rhodopsin: G
protein-coupled receptor structure and modeling incorporating the
beta2-adrenergic and adenosine alpha2A crystal structures. Methods
Mol. Biol. 672, 359–386.
(32) Bailey, D., and Brown, D. (2001) High-throughput chemistry
and structure-based design: survival of the smartest. Drug Discovery
Today 6, 57–59.
(50) Cherezov, V., Rosenbaum, D. M., Hanson, M. A., Rasmussen,
S. G., Thian, F. S., Kobilka, T. S., Choi, H. J., Kuhn, P., Weis, W. I.,
Kobilka, B. K., and Stevens, R. C. (2007) High-resolution crystal
structure of an engineered human beta2-adrenergic G protein-coupled
receptor. Science 318, 1258–1265.
(51) Warne, T., Serrano-Vega, M. J., Baker, J. G., Moukhametzianov,
R., Edwards, P. C., Henderson, R., Leslie, A. G., Tate, C. G., and
Schertler, G. F. (2008) Structure of a beta1-adrenergic G-protein-
coupled receptor. Nature 454, 486–491.
(52) Chien, E. Y., Liu, W., Zhao, Q., Katritch, V., Han, G. W.,
Hanson, M. A., Shi, L., Newman, A. H., Javitch, J. A., Cherezov, V., and
Stevens, R. C. (2010) Structure of the human dopamine D3 receptor in
complex with a D2/D3 selective antagonist. Science 330, 1091–1095.
(53) Ahmed, A., Choo, H., Cho, Y. S., Park, W. K., and Pae, A. N.
(2009) Identification of novel serotonin 2C receptor ligands by sequen-
tial virtual screening. Bioorg. Med. Chem. 17, 4559–4568.
(54) Isberg, V., Balle, T., Sander, T., Jorgensen, F. S., and Gloriam,
D. E. (2011) G protein- and agonist-bound serotonin 5-HT(2A)
receptor model activated by steered molecular dynamics simulations.
J. Chem. Inf. Model. 51, 315–325.
(
33) Schneider, G., and Bohm, H. J. (2002) Virtual screening and
fast automated docking methods. Drug Discovery Today 7, 64–70.
34) Indarte, M., Liu, Y., Madura, J. D., and Surratt, C. K. (2010)
(
Receptor-based discovery of a plasmalemmal monoamine transporter
inhibitor via high-throughput docking and pharmacophore modeling.
ACS Chem. Neurosci. 1, 223–233.
(
35) (2010) Molecular Operating Environment (MOE), In
010.10, Chemical Computing Group Inc., Montreal, QC, Canada.
36) Butler, S. G., and Meegan, M. J. (2008) Recent developments in
2
(
the design of anti-depressive therapies: targeting the serotonin trans-
porter. Curr. Med. Chem. 15, 1737–1761.
(37) Zhou, D., Stack, G. P., Lo, J., Failli, A. A., Evrard, D. A.,
Harrison, B. L., Hatzenbuhler, N. T., Tran, M., Croce, S., Yi, S.,
Golembieski, J., Hornby, G. A., Lai, M., Lin, Q., Schecter, L. E., Smith,
D. L., Shilling, A. D., Huselton, C., Mitchell, P., Beyer, C., and Andree,
T. H. (2009) Synthesis, potency, and in vivo evaluation of 2-piperazin-1-
ylquinoline analogues as dual serotonin reuptake inhibitors and seroto-
nin 5-HT1A receptor antagonists. J. Med. Chem. 52, 4955–4959.
(38) Herold, F., Chodkowski, A., Izbicki, L., Turlo, J., Dawidowski,
M., Kleps, J., Nowak, G., Stachowicz, K., Dybala, M., Siwek, A., Mazurek,
A. P., Mazurek, A., and Plucinski, F. (2011) Novel 4-aryl-pyrido[1,2-
c]pyrimidines with dual SSRI and 5-HT(1A) activity. part 3. Eur. J. Med.
Chem. 46, 142–149.
(55) Renault, N., Gohier, A., Chavatte, P., and Farce, A. (2010)
Novel structural insights for drug design of selective 5-HT(2C) inverse
agonists from a ligand-biased receptor model. Eur. J. Med. Chem. 45,
5086–5099.
(
39) Herold, F., Izbicki, y., Chodkowski, A., Dawidowski, M., Kr ꢁo l,
(56) Simpson, L. M., Wall, I. D., Blaney, F. E., and Reynolds, C. A.
(2011) Modeling GPCR active state conformations: The beta(2)
-adrenergic receptor. Proteins 79, 1441–1457.
(57) Ivetac, A., and McCammon, J. A. (2010) Mapping the drug-
gable allosteric space of G-protein coupled receptors: a fragment-based
molecular dynamics approach. Chem. Biol. Drug Des. 76, 201–217.
(58) Kazius, J., McGuire, R., and Bursi, R. (2005) Derivation and
validation of toxicophores for mutagenicity prediction. J. Med. Chem.
48, 312–320.
M., Kleps, J., Turzo, J., Wolska, I., Nowak, G., and Stachowicz, K. (2009)
Novel 4-aryl-pyrido[1,2-c]pyrimidines with dual SSRI and 5-HT1A
activity: Part 2. Eur. J. Med. Chem. 44, 4702–4715.
(40) Tomlinson, I. D., Iwamoto, H., Blakely, R. D., and Rosenthal,
S. J. (2011) Biotin tethered homotryptamine derivatives: high affinity
probes of the human serotonin transporter (hSERT). Bioorg. Med. Chem.
Lett. 21, 1678–1682.
(
41) Viegas-Junior, C., Danuello, A., da Silva Bolzani, V., Barreiro,
E. J., and Fraga, C. A. M. (2007) Molecular hybridization: A useful tool in
the design of new drug prototypes. Curr. Med. Chem. 14, 1829–1852.
(59) (2010) Discovery Studio Modeling Environment, 2.5.1 ed.,
Accelrys Software Inc., San Diego.
(
42) Sun, F., Xu, G., Wu, J., and Yang, L. (2006) Efficient lipase-
(60) Celik, L., Sinning, S., Severinsen, K., Hansen, C. G., Moller,
M. S., Bols, M., Wiborg, O., and Schiott, B. (2008) Binding of serotonin
to the human serotonin transporter. Molecular modeling and experi-
mental validation. J. Am. Chem. Soc. 130, 3853–3865.
catalyzed kinetic resolution of 4-arylmethoxy-3-hydroxybutanenitriles:
Application to an expedient synthesis of a statin intermediate. Tetra-
hedron: Asymmetry 17, 2907–2913.
(
43) Foquerna, S., Miralpeix, M., Pages, L., Puig, C., Cardus, A.,
(61) Still, W. C., Kahn, M., and Mitra, A. (1978) Rapid chromato-
graphic technique for preparative separations with moderate resolution.
J. Org. Chem. 43, 2923–2925.
(62) Talbot, J. N., Jutkiewicz, E. M., Graves, S. M., Clemans, C. F.,
Nicol, M. R., Huang, X., Mortensen, R. M., Neubig, R. R., and Traynor,
J. R. (2010) RGS inhibition selectively potentiates serotonin-mediated
antidepressant effects. Proc. Natl. Acad. Sci. U.S.A. 107, 11086–11091.
(63) (2008) PyMOL Molecular Graphics System, Mac OS X ed.,
Schrodinger, LLC, New York.
Anton, F., Cardenas, A., Vilella, D., Aparici, M., Calaf, E., Prieto, J., Gras,
J., Huerta, J. M., Warrellow, G., Beleta, J., and Ryder, H. (2004) Synthesis
and structure-activity relationships of novel histamine H1 antagonists:
Indolylpiperidinyl benzoic acid derivatives. J. Med. Chem. 47, 6326–6337.
(
44) Chan, A. W., Curran, T. T., Iera, S., Sellstedt, J. H., Vid, G.,
Feigelson, G., Ding, Z. (2002) Processes for the synthesis of derivatives of
,3-dihydro-1,4-dioxino-[2,3-f] quinoline, U.S. Patent WO/2002/092602.
45) Cryan, J. F., and Holmes, A. (2005) The ascent of mouse:
2
(
advances in modeling human depression and anxiety. Nat. Rev. Drug
Discovery 4, 775–790.
(
46) Steru, L., Chermat, R., Thierry, B., and Simon, P. (1985) The
tail suspension test: a new method for screening antidepressants in mice.
Psychopharmacology (Heidelberg, Ger.) 85, 367–370.
(47) Kim, C. Y., Mahaney, P. E., McConnell, O., Zhang, Y., Manas,
E., Ho, D. M., Deecher, D. C., and Trybulski, E. J. (2009) Discovery of a
new series of monoamine reuptake inhibitors, the 1-amino-3-(1H-indol-
1-yl)-3-phenylpropan-2-ols. Bioorg. Med. Chem. Lett. 19, 5029–5032.
(48) O’Neill, D. J., Adedoyin, A., Alfinito, P. D., Bray, J. A., Cosmi, S.,
Deecher, D. C., Fensome, A., Harrison, J., Leventhal, L., Mann, C.,
McComas, C. C., Sullivan, N. R., Spangler, T. B., Uveges, A. J.,
TrybulskiE. J., Whiteside, G. T., and Zhang, P. (2010) Discovery of novel
selective norepinephrine reuptake inhibitors: 4-[3-aryl-2,2-dioxido-2,1,3-
benzothiadiazol-1(3H)-yl]-1-(methylamino)butan-2-ols (WYE-103231).
J. Med. Chem. 53, 4511–4521.
5
52
dx.doi.org/10.1021/cn200044x |ACS Chem. Neurosci. 2011, 2, 544–552