Synthesis, biological evaluation, and three-dimensional in silico pharmacophore model for σ1 receptor ligands based on a series of substituted benzo[d ]oxazol-2(3H)-one derivatives

Article abstract of DOI:10.1021/jm900366z

Novel benzo[d ]oxazol-2(3H)-one derivatives were designed and synthesized, and their affinities against σ receptors were evaluated. On the basis of 31 compounds, a three-dimensional pharmacophore model for the σ₁ receptor binding site was devel

Full text of DOI:10.1021/jm900366z

5380 J. Med. Chem. 2009, 52, 5380–5393
DOI: 10.1021/jm900366z
Synthesis, Biological Evaluation, and Three-Dimensional in Silico Pharmacophore Model for
σ1 Receptor Ligands Based on a Series of Substituted Benzo[d ]oxazol-2(3H )-one Derivatives
Daniele Zampieri,‡ Maria Grazia Mamolo,*,‡ Erik Laurini,‡ Chiara Florio,† Caterina Zanette,† Maurizio Fermeglia,§
Paola Posocco,§ Maria Silvia Paneni,§ Sabrina Pricl,*,§ and Luciano Vio‡
‡Department of Pharmaceutical Sciences, Piazzale Europa 1, University of Trieste, 34127 Trieste, Italy, †Department of Life Sciences,
Section of Pharmacology, University of Trieste, 34127 Trieste, Italy, and §Department of Chemical, Environmental,
and Raw Materials Engineering (DICAMP), Molecular Simulation Engineering (MOSE) Laboratory, Piazzale Europa 1,
University of Trieste, 34127 Trieste, Italy
Received March 23, 2009
Novel benzo[d ]oxazol-2(3H )-one derivatives were designed and synthesized, and their affinities against
σ receptors were evaluated. On the basis of 31 compounds, a three-dimensional pharmacophore model
for the σ1 receptor binding site was developed using the Catalyst 4.9 software package. The best 3D
pharmacophore hypothesis, consisting of one positive ionizable, one hydrogen bond acceptor, two
hydrophobic aromatic, and one hydrophobic features provided a 3D-QSAR model with a correlation
coefficient of 0.89. The best hypothesis was also validated by three independent methods, i.e., the Fisher
randomization test included in the CatScramble functionality of Catalyst, the leave-one-out test, and
activity prediction of an additional test set. The achieved results will allow researchers to use this 3D
pharmacophore model for the design and synthesis of a second generation of high affinity σ1 ligands, as
well as to discover other lead compounds for this class of receptors.
Introduction
in neuromotor diseases such as dystonia and dyskinesia;11
indeed, σ receptors are highly expressed in brain areas asso-
ciated with the control of movement of facial muscles.11
The σ1 receptor subtype has been purified and cloned from
several animal species and man.12,13 Its primary sequence is
now available and shows remarkable homology with sterol
C8-C7 isomerase from fungi. The σ1 receptors exert a
modulatory role on neurotransmitter systems such as dopa-
minergic, serotoninergic, and muscarinic systems5,14,15 and on
the NMDA-stimulated neurotransmitters’ release.16 More-
over, σ1 receptors are involved in neuroprotective and
antiamnesic activities,17 modulation of opioid analgesia,18
and attenuation of cocaine-induced locomotor activity and
toxicity.19 In addition, σ1 antagonists have been shown to be
effective against negative symptoms of schizophrenia without
producing extrapyramidal side effects typical of traditional
neuroleptics.20,21
The σ binding sites were originally defined and classified as
opioid receptor subtypes.1 Later investigations demonstrated
that σ receptors were distinct from opioid and phencyclidine
analogues, and since then, at least two distinct σ receptor
subtypes, designated σ1 and σ2,2 have been pharmacologically
characterized.3-5 The cellular and anatomical distribution
of σ receptors is not restricted to the central nervous system
(CNSa),6-8 but extends to peripheral tissues such as blood
vessels, adrenal glands, testicles, ovaries, and immune sys-
tem.9 There is increasing evidence that σ receptors have a
neuromodulatory role in the CNS and are involved in the
etiology of various psychiatric diseases10 such as anxiety,
schizophrenia, and depression. The correlation between the
extent of toxic effects on CNS of some antipsychotic com-
pounds and their σ receptor affinity is also documented.
Furthermore, σ receptors may also have a potential role
On the other hand, the molecular identity of the σ2 receptor
subtype has not been fully determined,12,13 although a number
of studies have presented evidence linking σ2 receptors to
potassium channels and intracellular calcium release in NCB-
20 cells.22,23 Unlike the σ1 subtype, σ2 receptors may contribute
to the acute side effects of typical neuroleptic drugs, and
σ2 antagonists are known to attenuate extrapyramidal effects,
dystonic reactions, and tardive dyskinesia,2,14,22,24,25 suggesting
their potential use in the treatment of psychoses.20,21 Further-
more, σ2 receptors are involved in the regulation of cell
proliferation and maintenance of cell viability. They are highly
expressed in several tumoral cell lines,26,27 where σ2 agonists
produce morphological changes and apoptosis. The σ2 recep-
tor agonists promote Ca2þ release from endoplasmatic reticu-
lum and mitochondrial stores,28 with subsequent cell death by
*To whom correspondence should be addressed. For M.G.M.
(chemistry and biological sections): phone, þ390405583719; fax,
þ3904052572; e-mail, mamolo@units.it. For S.P. (computational and
molecular modeling sections): phone, þ390405583750; fax,þ39040569823;
e-mail, sabrina.pricl@dicamp.units.it.
a Abbreviations: 3D-QSAR, three-dimensional quantitative struc-
ture-activity relationship; CNS, central nervous system; NMDA, N-
methyl D-aspartate; NCB-20, mouse neuroblastoma Chinese hamster
brain hybrid cell line; PET, positron-emission tomography; SPECT,
single photon emission computed tomography; PTZ, (þ)-pentazocine;
NANM, (þ)-N-allylnormetazocine; DMAP, 4-dimethylaminopyridine;
TFA, trifluoroacetic acid; PC, positive charge; NC, negative charge;
HBD, hydrogen bond donor; HBA, hydrogen bond acceptor; HBAl,
hydrogen bond acceptor lipid; HYAr, hydrophobic aromatic; HY,
generic hydrophobic; Ar, ring aromatic; PI, positive ionizable; rmsd,
root-mean-square deviation; ERG2, ether-a-go-go related gene 2; SEM,
standard error of the mean.
r
pubs.acs.org/jmc
Published on Web 08/12/2009
2009 American Chemical Society

Article
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 17 5381
Scheme 1. Synthesis of Compounds 2a-k and 3a-k
Figure 1. Structures of compounds 1a-j.
caspase-independent apoptosis.27 Apoptosis may also be in-
duced in tumoral cells by regulation of the sphingolipid path-
way.29 Therefore, σ2 agonists may be useful as novel anticancer
agents and as imaging agents in cancer diagnosis by positron
emission tomography (PET)30 and single photon emission
computed tomography (SPECT).31,32
To date, well-known compounds characterized by a certain
degree of selectivity for the σ1 receptor subtype include (þ)-
benzomorphans such as (þ)-pentazocine (PTZ) and (þ)-N-
allylnormetazocine (NANM, SKF-10047), while haloperidol
and 1,3-di-(2-tolyl)guanidine show high affinity for both
receptor subtypes.22
In our previous work33 we described the synthesis and the
affinities for the σ receptors of a series of substituted benzo-
oxazolone derivatives 1a-j (Figure 1), in which the benzyl
group was variously substituted at the phenyl ring. The
benzooxazolone derivatives 1a-j have been designed accord-
ing to the σ1 receptor model proposed by Glennon,34-36 under
the following assumptions: (i) the benzooxazolone moiety
may interact with a primary hydrophobic site corresponding
to Glennon’s phenyl “B” region;34-36 (ii) the 4-methylpiper-
idin-1-yl spacer links the basic nitrogen atom to the benzoox-
azolone moiety; (iii) the substituted N-benzyl moiety may
bind the secondary hydrophobic phenyl “A” region of the
σ1 receptorial model, modulating the binding affinity of the
compounds for the σ receptors.
Table 1. Physical and Chemical Data for Compounds 2a-k and 3a-k
The relevant results indicated that the substituents on the
phenyl ring can modulate the σ1 and σ2 binding affinities of
these compounds. Indeed, all molecules show, to various
degrees, a preference for σ1 receptor sites, with the unsubsti-
tuted derivative and the corresponding para-substituted deri-
vatives exhibiting the highest affinity. Particularly, the best
result was reached with the 4-chloro substituted compound,
with a Ki(σ1) value of 0.1 nM and a Ki(σ2)/Ki(σ1) selectivity
ratio of 4270.
In this work we synthesized (Scheme 1) a series of 3-[(N-
benzyl-N-ethylamine)alkyl]benzo[d ]oxazol-2(3H )-one deri-
vatives 2a-k and 3a-k (Table 1), with various substitutions
at the benzene ring, to generate an adequate number of
compoundsfor the purpose ofdevelopinga three-dimensional
(3D) pharmacophore model for σ1 receptor ligands.
Since the benzooxazolone moiety is considered to form a
good interaction with a putative hydrophobic region in these
receptors, this group has been maintained in the new mole-
cular series. Moreover, the electronegative atoms of the
oxazolone moiety may further contribute to the binding
affinity. Effectively, electronegative atoms such as O or S
are frequently present in very potent σ1 ligands, bridging the
aromatic component and the classical alkyl or cycloalkyl
intermediate spacer linked to the basic nitrogen atom.13,36
On the basis of a molecular modeling study of σ1 receptor
ligands, Gund et al.37 also concluded that there could be a
secondary binding region that may surround the oxygen or
sulfur atom of the molecules. In the other half of the molecule,
the piperidin-4-ylmethyl core has been replaced by an alkyl
compd
R
n
mp (°C)a yield (%)
formula
Anal.b
2a
2b
2c
2d
2e
2f
H
3
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
4
180-182
129-131
166-168
192-194
145-147
168-170
179-183
167-169
156-158
195-197
127-130
145-147
115-117
168-170
162-164
127-129
154-156
98-100
30
33
27
30
15
18
21
10
39
41
53
30
23
23
46
15
19
55
11
49
42
40
C20H22N2O6 C, H, N
C20H21ClN2O6 C, H, N
2-Cl
3-Cl
4-Cl
C20H21ClN2O6 C, H, N
C20H21ClN2O6 C, H, N
2-OCH3
3-OCH3
4-OCH3
2-CH3
3-CH3
4-CH3
2,4-(CH3)2
H
C21H24N2O7
C21H24N2O7
C21H24N2O7
C21H24N2O6
C21H24N2O6
C21H24N2O6
C22H26N2O6
C, H, N
C, H, N
C, H, N
C, H, N
C, H, N
C, H, N
C, H, N
C, H, N
2g
2h
2i
2j
2k
3a
3b
3c
3d
3e
3f
C
21H24N2O6
2-Cl
3-Cl
C21H23ClN2O6 C, H, N
21H23ClN2O6 C, H, N
C21H23ClN2O6 C, H, N
C
4-Cl
2-OCH3
3-OCH3
4-OCH3
2-CH3
3-CH3
4-CH3
2,4-(CH3)2
C22H26N2O7
C22H26N2O7
C22H26N2O7
C22H26N2O6
C22H26N2O6
C22H26N2O6
C23H28N2O6
C, H, N
C, H, N
C, H, N
C, H, N
C, H, N
C, H, N
C, H, N
3g
3h
3i
212-214
169-171
138-140
103-105
3j
3k
a The melting points refer to the compounds as oxalate salts. b All
compounds were analyzed to be within (0.3% of the theoretical values.
spacer, endowed with greater conformal freedom. The choice
of a propyl or butyl chain is dictated by the fact that both fall
within the range of optimal distances between the basic
nitrogen atom and hydrophobic primary site in the receptor

5382 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 17
Table 2. Physical and Chemical Data for Compounds 6a-c and 7a-c
Zampieri et al.
compd
R
mp (°C)
yield (%)
formula
Anal.a
6a
6b
6c
7a
7b
7c
H
138-140
188-190
162-164
110-112
137-141
112-114
66
70
63
75
72
68
C19H22N2O
C, H, N
C, H, N
C, H, N
C, H, N
C, H, N
C, H, N
4-Cl
4-CH3
H
C
19H21ClN2O
C20H24N2O
20H24N2O
C
4-Cl
4-CH3
C20H23ClN2O
C21H26N2O
a All compounds were analyzed to be within (0.3% of the theoretical values.
Scheme 2. Synthesis of Compounds 6a-c and 7a-c
model of Glennon.34-36 All compounds obtained were sub-
sequently tested in order to assess their affinity and their
selectivity toward σ receptors.
As mentioned previously, the crystal structure of both
σ1 and σ2 receptors remains unsolved to date. In the absence
of a reliable 3D model of the target structure, ligand-based
molecular modeling tools can be successfully employed
to devise structural requirements crucial for receptor bind-
ing. The 3D quantitative structure-activity relationship
(3D-QSAR) pharmacophore modeling is such an approach
and constitutes a consolidated technique in drug design and
discovery.

Article
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 17 5383
Table 3. Binding Affinities of Compounds 2a-k, 3a-k, [3H]-(þ)-Pentazocine (σ1), and [3H]-DTG (σ2) Binding Sites in Rat Liver Homogenatea
Ki(σ1) (nM)b
nH
Ki(σ2) (nM)b
nH
σ2/σ1 ratio
b
b
compd
n
R
2a
2b
2c
2d
2e
2f
3
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
H
223 ( 47
2492 ( 115
3233 ( 447
1147 ( 206
1947 ( 136
8550 ( 500
83 ( 17
901 ( 107
1835 ( 159
871 ( 159
1833 ( 204
2.6 ( 1.5
843 ( 217
302 ( 112
7.1 ( 1.5
594 ( 96
297 ( 34
21 ( 3.8
1.12 ( 0.23
1.13 ( 0.07
1.67 ( 0.34
1.09 ( 0.19
1.57 ( 0.14
1.75 ( 0.16
0.95 ( 0.17
0.99 ( 0.10
1.71 ( 0.22
1.06 ( 0.18
1.87 ( 0.29
0.45 ( 0.09
1.62 ( 0.58
0.87 ( 0.32
0.72 ( 0.10
0.87 ( 0.13
1.69 ( 0.32
0.76 ( 0.11
0.40 ( 0.13
1.03 ( 0.07
0.95 ( 0.09
0.88 ( 0.14
1.30 ( 0.20
0.55 ( 0.07
1121 ( 603
456 ( 154
1119 ( 722
456 ( 154
192 ( 130
2130 ( 295
2602 ( 714
1347 ( 99
2716 ( 1847
991 ( 180
2064 ( 11
120 ( 40
37.4 ( 16
NDc
0.80 ( 0.21
0.92 ( 0.28
0.60 ( 0.28
0.66 ( 0.39
2.04 ( 0.51
0.95 ( 0.07
0.96 ( 0.13
1.16 ( 0.29
1.03 ( 0.39
1.17 ( 0.20
2.41 ( 0.03
0.80 ( 0.23
0.55 ( 0.14
5.0
2-Cl
3-Cl
4-Cl
0.18
0.35
0.40
0.10
0.16
31.3
1.5
2-OCH3
3-OCH3
4-OCH3
2-CH3
3-CH3
4-CH3
2,4-(CH3)2
H
2g
2h
2i
1.5
1.1
2j
2k
3a
3b
3c
3d
3e
3f
1.1
46.2
0.04
NDc
5.1
NDc
0.6
2-Cl
3-Cl
4-Cl
36.2 ( 6.5
NDc
1.12 ( 0.20
2-OCH3
3-OCH3
4-OCH3
3-CH3
4-CH3
2,4-(CH3)2
187 ( 13
20.8 ( 4.3
31.8 ( 5.3
22.9 ( 2.8
6.94 ( 5.2
327 ( 166
130 ( 46
235 ( 71
1.57 ( 0.15
1.11 ( 0.23
0.96 ( 0.14
1.08 ( 0.13
0.34 ( 0.09
0.88 ( 0.25
1.50 ( 0.67
0.85 ( 0.19
3g
3i
1
102 ( 7.4
97 ( 7.3
239 ( 27
15 ( 3
180 ( 22
5.7 ( 1
0.31
0.20
0.03
22
3j
3k
(þ)-pentazocine
DTG
haloperidol (12)
0.72
41
a nH = Hill coefficient. b Standard error of the mean. c ND: not determined.
The major goal this paper is the generation of a predictive
pharmacophore model for σ1 ligands. To develop the model,
we resorted to the HypoGen method implemented in the
Catalyst software package.38 Starting with a training set of
31 σ1 ligands, a pharmacophore model (also called a
hypothesis) able to quantitatively correlate the estimated
affinities with the corresponding measured values was gener-
ated. The model was then validated by statistical means and
by its ability to predict the affinity of a further ensemble of
different compounds (test set). Overall, we verified that our
3D-QSAR pharmacophore model was predictive not only
withinthetraining set butalso for the test set compounds, with
acceptable errors.
The substituted 1-benzyl-N-phenylpiperidine-4-carbox-
amide derivatives 6a-c and the corresponding N-benzyl deri-
vatives 7a-c were prepared starting from the commercially
available 4-piperidinecarboxylic acid which was transformed
into the corresponding N-boc-4-piperidinecarboxylic chloride
by treatment with SOCl2 under N2 (Table 2 and Scheme 2).
A solution of aniline and triethylamine in CH2Cl2 was added
directly to the reaction mixture with a catalytic amount of
DMAP. The reaction was carried out at room temperature
under N2 flux to afford N-phenyl-1-(tert-butoxycarbonyl)-
piperidine-4-carboxamide 8, which was deprotected with
TFA to yield the N-phenylpiperidine-4-carboxamide 10.
Following the same route described above and using
benzylamine, N-benzyl-1-(tert-butoxycarbonyl)piperidine-4-
carboxamide 9 was prepared, from whichN-benzylpiperidine-
4-carboxamide 11 was obtained after deprotection with TFA.
From the piperidine-4-carboxamides 10 and 11, the corre-
sponding 1-benzyl derivatives 6a-c and 7a-c were obtained
by alkylation with the opportune benzyl chlorides.
Chemistry
The substituted 3-[3-(N-benzyl-N-methylamino)propyl]-
benzo[d ]oxazol-2(3H)-one derivatives2a-k and 3-[4-(N-ben-
zyl-N-methylamino)butyl]benzo[d ]oxazol-2(3H)-one deriva-
tives 3a-k were synthesized (Scheme 1) by treating an
acetonitrile solution of benzo[d ]oxazol-2(3H)-one 4 with
1-bromo-3-chloropropane and 1,4-dibromobutane, respec-
tively, in the presence of K2CO3 and a catalytic amount of
KI. From the obtained 3-(3-chloropropyl) and 4-(4-bro-
mobutyl)benzo[d ]oxazol-2(3H)-one intermediates 5a,b, the
substituted N-benzyl derivatives were obtained by heating
at reflux with N-methylbenzylamines in acetonitrile in the
presence of K2CO3 and a catalytic amount of KI.
Results and Discussion
On the basis of the interesting Ki(σ1) values of a series of
benzo[d]oxazolone derivatives 1a-j (Figure 1) we described in
previous work,33 we synthesized a series of new benzooxazolone
derivatives 2a-k and 3a-k (Table 1) in which the 4-methylpi-
peridin-1-yl spacer linking the benzooxazolone moiety to the
benzyl group was replaced by the 3-(N-methylamino)propyl

5384 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 17
Zampieri et al.
Table 4. Experimental and Estimated Affinity Values of the Training
Test Compounds
a training set was derived with 31 compounds from our series,
considering structural diversity and the widest possible cover-
age of the in vitro affinity range (see Table 4). To validate the
developed pharmacophore model, we then predicted the
σ1 affinity of the test set of piperidine-4-carboxamide deriva-
tives 6a-c and 7a-c and the reference σ1 ligands haloperidol
12, trifluperidol 13, and ifenprodil 14. Compounds 6a-c and
7a-c were synthesized (Scheme 2), their Ki(σ1) values were
determined, and their estimated σ1 affinities were compared
with the experimental data (Table 5).
3D Pharmacophore Modeling. In this work we developed
a three-dimensional pharmacophore model in order to have
a tool to design a second generation of σ1 receptor ligands.
It is a widely accepted that 3D pharmacophore modeling
is a well-behaved approach to quantitatively explore the
common chemical characteristics among a considerable
number of different structures. However, a computed phar-
macophore model can only be as good as the informa-
tion that it contains. To achieve a quality model, at least
three must-obey rules should be respected in three-dimen-
sional quantitative structure-activity relationship (3D-
QSAR) generation: (i) the training set must include a wide
population (at least 16 items) of diverse compounds covering
at least 4 orders of magnitude of activity; (ii) the most
active compound should be included in the training set;
(iii) all biological data must be obtained by homogeneous
procedures.39,40
In our case, a training set consisting of 31 compounds from
our series was prepared by considering structural diversity
and the widest possible coverage of the in vitro affinity range
(see Table 4). The molecules in our training set were selected
according to the following criteria: (i) the training set should
contain structures from each series of active compounds;
(ii) the training set should cover the molecular bioactivities
(Ki) as widely as possible. Should there be only one com-
pound with maximum or minimum order of bioactivity in a
series, then this compound was assigned to the training set.
The HypoGen algorithm allows a maximum of five fea-
tures to be considered in the pharmacophore generation
process. Accordingly, from the 11 features available in the
Catalyst features dictionary (see Experimental Section for
details), we excluded all those that clearly did not match the
chemistry of the molecules of the training set, such as positive
charge (PC) and negative charge (NC), as all ligands were
considered in their neutral form. Also, preliminary runs
including the hydrogen bond donor (HBD) and hydrogen
bond acceptor lipid (HBAl) features confirmed that these
features were never used in the generation of the pharmaco-
phore models, even though they were present and properly
mapped on several molecules of the training set. Thus, the
HBD and HBAl features were removed from the list. More-
over, as most of the molecules in the training set possess
both hydrophobic aromatic and hydrophobic aliphatic
groups, the specific hydrophobic aromatic (HYAr) and
hydrophobic aliphatic (HYAl) features were both selected.
The more generic hydrophobic feature (HY) was also chosen
to optimize the substituents mapping on the phenyl ring. In
summary, the following five chemical features were taken
into account for hypothesis generation with HypoGen:
hydrogen bond acceptor (HBA), hydrophobic aromatic
(HYAr), hydrophobic aliphatic (HYAl), ring aromatic
(Ar), and positive ionizable (PI).
Ki(σ1) (nM)
compd
R
experimental
estimated
errora
1a
1b
1c
1d
1e
1f
H
3.58
60.9
0.098
258
18
60
5.0
-1.0
2.8
2-Cl
4-Cl
0.27
280
160
1
2,4-Cl
2-CH3
4-CH3
2,4-CH3
2,4,6-CH3
4-Ph
1.1
5.4
29.8
3.07
30.2
6210
394
-3.1
7.0
1g
1h
1i
210
660
230
670
550
6700
1110
960
3600
880
56
-9.4
-1.7
-1.5
2.5
1j
naphthyl
H
2-Cl
1017
223
2a
2b
2c
2d
2e
2f
2492
3233
1147
1947
8550
83
2.7
3-Cl
4-Cl
-2.9
-1.2
1.8
2-OCH3
3-OCH3
4-OCH3
2-CH3
3-CH3
4-CH3
2,4-(CH3)2
H
-9.7
-1.5
2.2
2g
2h
2i
901
2000
4000
410
1400
24
1835
871
2.2
2j
-2.1
-1.3
9.2
2k
3a
3b
3c
3d
3e
3f
1833
2.6
843
2-Cl
3-Cl
140
450
16
-6.0
1.5
302
4-Cl
7.1
2.3
2-OCH3
3-OCH3
4-OCH3
3-CH3
4-CH3
2,4-(CH3)2
594
297
260
140
42
-2.3
-2.1
2.0
3g
3i
21
102
100
270
860
-1.0
2.8
3j
3k
97
239
3.6
a Values in the error column represent the ratio of the estimated to
experimental affinity, or its negative inverse if the ratio is less than 1.
and 4-(N-methylamino)butyl groups, respectively. The aim was
to verify if these modifications might produce compounds
retaining affinity for σ receptor sites. All compounds were tested
to evaluate their Ki values against both σ1 and σ2 subtypes
(Table 3).
From the results obtained so far (Table 3), it appears that
the substituents on the phenyl ring strongly modulate the
σ1 and σ2 binding affinity of these compounds, butylene
derivatives 3a-k having higher affinity than the correspond-
ing propylene derivatives 2a-k, as in the case of compounds
1a-j.33 The butylene intermediate chain determines the opti-
mal distance between the primary B and secondary A hydro-
phobic centers proposed by the Glennon’s model.34-36 The
compound with the highest affinity is the butylene derivative
3a (R=H), with Ki(σ1) value of 2.6 nM and an interesting
selectivity ratio Ki(σ2)/Ki(σ1) = 46.2. The para-substituted
compound 3d (R=4-Cl), instead, still maintains an appreci-
able σ1 affinity but has a selectivity ratio (Ki(σ2)/Ki(σ1)=5.1)
lower than those of the unsubstituted derivatives.
Unlike the derivatives previously described, compound 3g
(R=4-OCH3) has no selectivity, showing a moderate affinity
toward both receptors. On the other hand, compounds 3b and
3i-k show preference toward the σ2 receptor. Specifically,
compound 3k (R = 2,4-(CH3)2) has the highest σ2 affinity
(Ki(σ2)=6.94 nM) and selectivity ratio (Ki(σ2)/Ki(σ1)=0.03)
of the series.
For the development of a 3D quantitative structure-
activity relationship (3D-QSAR) pharmacophore model,
In total, 10 hypotheses were generated by the Hypo-
Gen algorithm, all characterized by 5 features. The total

Article
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 17 5385
Table 5. Experimental and Estimated Affinity Values of the Test Set Compounds
Ki(σ1) (nM)
compd
R
experimental
estimated
errora
6a
6b
6c
7a
7b
7c
12
13
14
H
48.1
45.0
105
22.5
12.9
69.4
5.7
210
140
200
37
4.4
3.1
4-Cl
4-CH3
H
1.9
1.6
4-Cl
4-CH3
5.5
76
-2.3
1.1
2.2
4.6
19
-2.6
5.8
0.8
2.0
9.5
a Values in the error column represent the ratio of the estimated to experimental affinity, or its negative inverse if the ratio is less than 1.
hypothesis cost of these 10 best models varies between
124.1 for the best ranked model (Hypo1) to 151.2 for the
lowest ranked one (Hypo10). Such a confined difference
(27 bits) reflects both the homogeneity of the generated
hypotheses and the adequacy of the molecular training set.
The difference between the null and the fixed costs, which
should be higher than 70 to guarantee a robust correlation,
is 91 in our case. This corresponds to a chance of true
correlation in the data greater than 90%.41 Furthermore,
in all the generated hypotheses the total costs are much closer
to the fixed cost (103.1) than to the null cost (194.1),
indicating that meaningful models are obtained. Finally,
the root-mean-square deviations (rmsd) and the correlation
coefficients (F) between estimated and experimental affinities
range from 1.126 to 1.844 and from 0.896 to 0.566, respec-
tively. As all the generated pharmacophores map the mole-
cules of the training set in a similar way, the first model
(Hypo1), characterized by the highest cost difference, the
lowest rmsd, and the best F values, was selected for further
analysis.
Hypo1 contains one hydrogen bond acceptor, two hydro-
phobic aromatic features, one hydrophobic feature, and one
positive ionizable group. The affinities of the 31 compounds
estimated using Hypo1 are reported in Table 4, along with
the experimental values and the relevant errors (expressed as
the ratio of estimated to experimental values). This table
clearly shows that 24 out of 31 molecules in the training set
have errors less than 4 while the remaining 7 have errors less
than 10. Figure 2A-C illustrates the selected Hypo1 phar-
macophore model, while parts D, E, and F of Figure 2 show
the mapping of compounds 1c, 1d, and 2h onto Hypo1,
respectively.
In compound 1c, as seen in Figure 2D, the aromatic ring of
the benzooxazolone moiety matches one of the HYAr fea-
tures; the other HYAr feature is nicely overlapped by the
additional phenyl ring. The carbonyl group and the basic
nitrogen atom of the pyperidine ring match the HBA and PI
functions, while the chlorine atom on the monosubstituted
phenyl ring maps the remaining HY feature. Quite an
analogous mapping is observed for 1d (Figure 2E). The
estimated affinities for 1c and 1d are 0.27 and 280 nM, while
the corresponding experimental affinities are 0.098 and
258 nM, respectively.
Figure 2F is an example of a pharmacophore mapping of a
compound that is less active than the former two. Compound
2h does not map all the features encoded in Hypo1. In fact, 2h
maps the two HYAr functions, again by means of the two
phenyl rings; the PI feature is still overlapped by the nitrogen
atom of the cyclic/linear bridging moiety, and the HBA
function is located over the carbonyl oxygen. However, it
does not map the HY function. According to this partial
mapping, this compound is predicted to be less active.
A critical step in automated pharmacophore generation
is model validation, especially in those cases where the
model has been generated for the purpose of predicting the
activity of external sets of compounds or, as in our case,
of estimating the activity of newly conceived molecular
entities prior to their synthesis. The first method we used
to check the robustness of our correlation was the prediction
of the affinity of a further set of molecules, also called the test
set, composed of six additional molecules from our series and
three compounds taken from the literature42 (see Table 5).
The aim of this validation was to verify if our 3D phar-
macophore model was able to predict the experimentally

5386 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 17
Zampieri et al.
Figure 2. Geometrical relationships (A, B) among the features of the top-scoring pharmacophore Hypo1 (C, parallel glaze stereoview), and
pharmacophore mapping of 1c (D), 1d (E), and 2h (F) in the training set. The hypothesis features are portrayed as mashed spheres, color-coded as
follows: red, PI; light blue, HYAr; pink, HY, light green, HBA. HBA is actually represented as a pair of spheres (the smaller sphere represents the location
˚
of the HBA atom on the ligand and the larger one the location of an HB donor on the receptor). Selected distances (A) and angles (deg) are labeled.

Article
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 17 5387
Figure 3. Pharmacophore mapping of 7b (A) and 12 (B) in the test set. The hypothesis features are portrayed as mashed spheres, color-coded as
follows: red, PI; light blue, HYAr; pink, HY, light green, HBA.
Table 6. Output Parameters of the 10 Lowest Cost Hypotheses Result-
ing from the Statistical Evaluation According to the CatScramble
Validation Procedure for the σ1 Receptor Ligands
determined affinity values of the test set compounds. Inter-
estingly, a good correlation coefficient (0.882) was observed
when a regression analysis was performed by mapping the
hypothesis
F
rmsd
total cost
test set onto the features of the best pharmacophore hypoth-
esis Hypo1. The predicted and the experimental Ki values for
the test set along with the respective errors are shown in
Table 5. The average error in predicting the affinity of the test
set molecules is 2.5. Given the inherent simplicity of the
pharmacophoric approach and considering the intrinsic
variability of the biological responses, we can conclude that
the ability of the present 3D pharmacophore model to
predict the affinity of this series of σ1 receptor ligands is
quite satisfactory. Figure 3 shows the mapping of two test set
molecules (7b and 12, respectively) to the σ1 receptor phar-
macophore model. For both compounds, the two phenyl
groups match the HYAr functions, the chlorine atom fits the
HY feature, the carbonyl oxygen atom provides the HBA
function, and the basic nitrogen is located over the PI
feature.
Notwithstanding these good results, a second test was
performed to check the statistical significance of the 3D
pharmacophore model Hypo1, based on a randomization
procedure. This was derived from the Fisher method using
the CatScramble program available in the Catalyst suite of
programs. According to the validation procedure, the ex-
perimental affinities of the compounds in the training set
were scrambled randomly and the resulting new training sets
were used for a number of new HypoGen runs. The para-
meters used in running these calculations were the same
employed in the initial HypoGen calculation, and since a
98% confidence level was selected, 49 random hypothesis
runs were performed. The results clearly indicate that ran-
domization produced hypotheses with no predictive values
similar or close to the corresponding Hypo1. Indeed, none of
the outcome hypotheses had a lower cost score, better
correlation, or smaller root-mean-square deviation than
the initial hypothesis. Table 6 lists the first 10 lowest total
score values of the resulting 49 hypotheses for our test set
molecules. In conclusion, there is a 98% chance for the best
hypothesis to represent a true correlation in the training set
affinity data for the present classes of compounds.
1
0.801
0.744
0.764
0.625
0.613
0.611
0.588
0.578
0.566
0.501
1.126
1.142
1.492
1.456
1.768
1.778
1.783
1.812
1.823
1.844
1.899
0.896
149.3
151.6
152.3
154.8
157.1
158.2
160.0
161.3
165.1
168.9
124.1
2
3
4
5
6
7
8
9
10
Hypo1
affinity of each excluded molecule is correctly predicted by
the corresponding one-missing hypothesis. The value of
F, the feature composition of the pharmacophore, and the
quality of the predicted affinity of the excluded molecule
were used as measures for the assessment of the statistical
test. For each of the 31 new hypotheses generated according
to this method, we did not obtain meaningful differences
between Hypo1 and each hypothesis resulting from the
exclusion of one compound at a time.
Overall, the 3D pharmacophore model derived in the
present work is quite simple, and it is in perfect agreement
with another pharmacophore model for σ1 receptors pre-
viously reported by Glennon et al.,34 in which a basic
nitrogen is placed between two hydrophobic sites. Impor-
tantly, the primary hydrophobic site of Glennon’s model is
˚
located at an optimum distance of 7-9 A from the basic
nitrogen. This corresponds well to our aromatic and alipha-
˚
tic hydrophobic feature spheres (8.5 and 7.0 A from the PI
feature, respectively; see Figure 2C). The second hydropho-
˚
bic site, reported at a distance of 2.5-3.9 A from the basic
nitrogen in Glennon’s model, is matched by the remaining
˚
HYAr, which is located at 3.6 A from the PI feature in our 3D
pharmacophore. Also, our model compares well, both in
terms of feature type and geometrical characteristics, with a
more general pharmacophore model recently obtained by
Laggner et al. for σ1 and the ERG2 protein.42 As a conclud-
ing comment, we emphasize that only five-feature pharma-
cophore hypotheses were generated by Catalyst according to
our procedure. This is significant for a number of reasons,
first, because most prior published pharmacophores for
σ1 receptors contain only four features (as reviewed above),
second, because our five-feature hypothesis does indeed
Finally, a further statistical test, the leave-one-out meth-
od, which consists of recomputing the hypothesis by exclud-
ing from the training set one molecule at a time, was carried
out. Basically, this test is performed to verify whether or not
the correlation is strongly dependent on one particular
compound in the training set. The test is positive if the

5388 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 17
Zampieri et al.
represent a more stringent and significant model, and last,
because it illustrates that Catalyst identified sufficient evi-
dence in this structural data set to add or distinguish an
additional feature or nuance not seen previously.
solid. Yield 4.01 g (85%); mp 62-64 °C. IR (Nujol): 1791 cm-1
.
1H NMR (CDCl3-TMS) ppm (δ): 2.30 (m, 2H, CH2-CH2-CH2,
J=5.9-6.6 Hz); 3.63 (t, N-CH2, J=6.6 Hz); 4.00 (t, 2H, CH2-Cl,
J=5.9 Hz); 7.00-7.30 (m, 4H arom). MS: m/z 212 [MHþ] 214
[MHþ þ 2].
3-(4-Bromobutyl)benzo[d ]oxazol-2(3H)-one (5b).44 This inter-
mediate was obtained in an analogous way using 1,4-dibromo-
butane. Yield 2.99 g (60%); mp 42-45 °C. IR (Nujol): 1747
Conclusions
In this work we discussed how, from three series of newly
synthesized compounds characterized by a broad range of
affinity toward σ1 receptors, we derived a three-dimensional
pharmacophore model with quantitative predictive ability for
these classes of molecules. The best generated pharmacophore
model (Hypo1) consists of five features: two hydrophobic
aromatic, one hydrophobic aliphatic, one hydrogen bond
acceptor, and one positive ionizable group. Hypo1 reasonably
predicts the affinity of the test set molecules with a correlation
coefficient of 0.896 and shows the best statistical significance
among all the generated models. Two validation tests, the
Fisher test and the leave-one-out test, confirmed the statistical
validity of our simple but effective 3D pharmacophore,
excluding any possibility of a chance correlation between
experimental and predicted affinity values. Finally, the model
was predictive not only for the training set compounds but
also for a test set of nine additional molecules, three of which
were taken from the literature and not structurally related to
our series.
Compared with other drug discovery tools, the pharmaco-
phore approach hasthe significant advantagethatitis fast and
able to predict the activity of quite a large number of
molecules in a relatively short time. Given the reasonable
predictive ability of our model, we expect to exploit it in the
development and optimization of our promising series of
compounds. In particular, we will use this 3D pharmacophore
to estimate the potential affinity of virtual libraries of newly
designed, second generation σ1 receptor ligands prior to
synthesis and biological testing.
1
cm-1. H NMR (CDCl3-TMS) ppm (δ): 1.80 (m, 4H, CH2-
CH2-CH2-CH2); 3.63 (m, N-CH2); 4.00 (m, 2H, CH2-Br);
7.10-7.40 (m, 4H arom). MS: m/z 270 [MHþ] 272 [MHþ þ 2].
3-[3-(N-Benzyl-N-methylamino)propyl]benzo[d ]oxazol-2(3H)-
one (2a). A mixture of 3-(4-chloropropyl)benzo[d ]ossazol-
2(3H)-one (0.60 g, 2.84 mmol), N-methylbenzylamine (0.28 g,
2.27 mmol), anhydrous K2CO3 (2.19 g, 15.9 mmol), and a
catalytic amount of KI was dissolved in 50 mL of CH3CN, and
the solution was refluxed and monitored by TLC. The inorganic
salts were filtered off, and the solvent was evaporated under
reduce pressure. Distilled water (50 mL) was added, and the
residue was extracted 3 times with CHCl3 (3 ꢀ 100 mL). The
organic phase was separated and dried over anhydrous Na2SO4.
The filtered solution was concentrated under reduced pressure
and the remaining oil was treated with an equimolar amount of
oxalic acid in absolute ethanol to yield the oxalate salt, which was
filtered and washed with cold ethanol. Yield 0.26 g (0.20 g of free
base, 30%); mp 180-182 °C. IR (Nujol): 1772, 2697 cm-1. 1H
NMR (free base, CDCl3-TMS) ppm (δ): 2.01 (m, 2H, CH2-
CH2-CH2, J=6.7-7.3 Hz); 2.21 (s, 3H, N-CH3); 2.49 (t, 2H,
CH2-N(CH3)-CH2-Ar, J=6.7 Hz); 3.51 (s, 2H, N-CH2-Ar); 3.93
(t, 2H, N-CH2-(CH2)2-, J=7.3 Hz); 6.95-7.40 (m, 9H, arom).
MS: m/z 297 [MHþ].
Compounds 2b-k were synthesized following the same route
described above for compound 2a.
3-[3-[N-(2-Chlorobenzyl)-N-methylamino]propyl]benzo[d]oxazol-
2(3H)-one (2b). IR (Nujol): 1764, 2721 cm-1. 1H NMR (free base,
CDCl3-TMS) ppm (δ): 1.97 (m, 2H, CH2-CH2-CH2, J= 6.6-
7.3 Hz); 2.15 (s, 3H, CH3); 2.46 (t, 2H, CH2-N(CH3)-CH2-Ar, J=
6.6 Hz); 3.51 (s, 2H, N-CH2-Ar); 3.84 (t, 2H, N-CH2-(CH2)2-,
J=7.3 Hz); 6.89-7.37 (m, 8H, arom). MS: m/z 331 [MHþ] 333
[MHþ þ 2].
3-[3-[N-(3-Chlorobenzyl)-N-methylamino]propyl]benzo[d]oxazol-
2(3H)-one (2c). IR (Nujol): 1767, 2673 cm-1. 1H NMR (free base,
CDCl3-TMS) ppm (δ): 1.95 (m, 2H, CH2-CH2-CH2, J=6.6-
7.3 Hz); 2.15 (s, 3H, CH3); 2.45 (t, 2H, CH2-N(CH3)-CH2-Ar,
J=6.6 Hz); 3.43 (s, 2H, N-CH2-Ar); 3.86 (t, 2H, N-CH2-(CH2)2-,
J=7.3 Hz); 6.92-7.30 (m, 8H, arom). MS: m/z 331 [MHþ] 333
[MHþ þ 2].
Experimental Section
Unless otherwise noted, starting materials and reagents were
obtained from commercial suppliers and were used without
purification. Melting points were determined with a Buchi 510
capillary apparatus and are uncorrected. Infrared spectra in
Nujol mulls were recorded on a JaskoFT 200spectrophotometer.
Proton nuclear magnetic resonance (1H NMR) spectra were
determined on a Varian Gemini 200 spectrometer, and the
chemical shifts are reported as δ (ppm) in CDCl3 solution.
Coupling constants J are expressed in hertz (Hz). Reaction
courses and product mixtures were routinely monitored by
thin-layer chromatography (TLC) on silica gel precoated F254
Merck plates. ESI-MS spectra were obtained on a PE-API I
spectrometer by infusion of a solution of the sample in MeOH.
Elemental analyses (C, H, N) were performed on a Carlo Erba
1106 analyzer and were within (0.3 of the theoretical value.
Synthesis of Benzooxazolone Derivatives. 3-(3-Chloropropyl)-
benzo[d ]oxazol-2(3H)-one (5a).43 A mixture of benzo[d ]oxazol-
2(3H)-one 4 (3.0 g, 22.22 mmol) and K2CO3 (7.7 g, 55.55 mmol)
was dissolved in 50 mL of CH3CN and the solution was refluxed
for 10 min. 1-Bromo-3-chloropropane (8.7 g, 55.55 mmol) and a
catalyticamount of KI were added, and the mixturewas stirred for
an additional 3 h at reflux temperature. The inorganic salts were
filtered off, and the solvent was evaporated under reduced pres-
sure. Distilled water (50 mL) was added, and the residue was
extracted 3 times with CHCl3 (3 ꢀ 100 mL). The organic phase
was separated and dried over anhydrous Na2SO4. The filtered
solution was concentrated under reduced pressure, and the re-
maining oil was crystallized from n-hexane to afford a light-yellow
3-[3-[N-(4-Chlorobenzyl)-N-methylamino]propyl]benzo[d]oxazol-
2(3H)-one (2d). IR (Nujol): 1770, 2677 cm-1. 1H NMR (free base,
CDCl3-TMS) ppm (δ): 2.05 (m, 2H, CH2-CH2-CH2, J=6.6-
7.3 Hz); 2.17 (s, 3H, N-CH3); 2.45 (t, 2H, CH2-N(CH3)-CH2-Ar,
J=6.6 Hz); 3.46 (s, 2H, N-CH2-Ar); 3.90 (t, 2H, N-CH2-(CH2)2-,
J=7.3 Hz); 6.95-7.30 (m, 8H, arom). MS: m/z 331 [MHþ] 333
[MHþ þ 2].
3-[3-[N-(2-Methoxybenzyl)-N-methylamino]propyl]benzo[d]oxa-
zol-2(3H)-one (2e). IR (Nujol): 1771, 2673 cm-1. 1H NMR (free
base, CDCl3-TMS) ppm (δ): 1.87-2.00 (m, 2H, CH2-CH2-CH2,
J=6.6-7.3 Hz); 2.14 (s, 3H, CH3); 2.43 (t, 2H, CH2-N(CH3)-CH2-
Ar, J=6.6 Hz); 3.45 (s, 2H, N-CH2-Ar); 3.75 (s, 3H, OCH3);3.87(t,
2H, N-CH2-(CH2)2-, J=7.3 Hz); 6.79-7.25(m, 8H, arom). MS:m/
z 327 [MHþ].
3-[3-[N-(3-Methoxybenzyl)-N-methylamino]propyl]benzo-
[d ]oxazol-2(3H)-one (2f). IR (Nujol): 1770, 2676 cm-1 1H
.
NMR (free base, CDCl3-TMS) ppm (δ): 1.83-1.97 (m, 2H,
CH2-CH2-CH2, J=6.6-7.3 Hz); 2.12 (s, 3H, CH3); 2.40 (t,
2H, CH2-N(CH3)-CH2-Ar, J=6.6 Hz); 3.39 (s, 2H, N-CH2-
Ar); 3.74 (s, 3H, OCH3); 3.84 (t, 2H, N-CH2-(CH2)2-, J=7.3
Hz); 6.70-7.21 (m, 8H, arom). MS: m/z 327 [MHþ].
3-[3-[N-(4-Methoxybenzyl)-N-methylamino]propyl]benzo[d ]-
oxazol-2(3H)-one (2g). IR (Nujol): 1771, 2685 cm-1. 1H NMR

Article
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 17 5389
(free base, CDCl3-TMS) ppm (δ): 1.98 (m, 2H, CH2-CH2-
CH2, J = 6.7-7.3 Hz); 2.20 (s, 3H, CH3); 2.43 (t, 2H, CH2-
N(CH3)-CH2-Ar, J = 6.7 Hz); 3.45 (s, 2H, N-CH2-Ar); 3.84
(s, 3H, OCH3); 3.91 (t, 2H, N-CH2-(CH2)2-, J=7.3 Hz); 6.85-
7.30 (m, 8H, arom). MS: m/z 327 [MHþ].
1.75-1.90 (m, 2H, CH2-CH2-CH2); 2.15 (s, 3H, CH3);
2.36-2.43 (t, 2H, CH2-N(CH3)-CH2-Ar, J=6.6 Hz); 3.42 (s,
2H, N-CH2-Ar); 3.79-3.86 (t, 2H, N-CH2-(CH2)3-, J=7.3 Hz);
6.93-7.29 (m, 8H, arom). MS: m/z 345 [MHþ] 347 [MHþ þ 2].
3-[4-[N-(2-Methoxybenzyl)-N-methylamino]butyl]benzo[d]oxazol-
3-[3-[N-Methylamino-N-(2-methylbenzyl)]propyl]benzo[d]oxazol-
2(3H)-one (3e). IR (Nujol): 1770, 2673 cm-1 1H NMR
.
1
2(3H)-one (2h). IR (Nujol): 1766, 2723 cm-1. H NMR (free
(CDCl3-TMS) ppm (δ): 1.53-1.90 (m, 4H, CH2-CH2-CH2-
CH2); 2.20 (s, 3H, CH3); 2.45 (t, 2H, CH2-N(CH3)-CH2-Ar, J=
6.6 Hz); 3.48 (s, 2H, N-CH2-Ar); 3.79 (s, 3H, CH3); 3.87 (t, 2H,
N-CH2-(CH2)3-, J=7.3 Hz); 6.83-7.30 (m, 8H, arom). MS: m/z
341 [MHþ].
base, CDCl3-TMS) ppm (δ): 1.82-1.96 (m, 2H, CH2-CH2-
CH2, J=6.6-7.3 Hz); 2.14 (s, 3H, CH3); 2.30 (s, 3H, CH3); 2.44
(t, 2H, CH2-N(CH3)-CH2-Ar, J=6.6 Hz); 3.40 (s, 2H, N-CH2-
Ar); 3.78 (t, 2H, N-CH2-(CH2)2-, J=7.3 Hz); 6.86-7.23 (m, 8H,
arom). MS: m/z 311 [MHþ].
3-[4-[N-(3-Methoxybenzyl)-N-methylamino]butyl]benzo[d ]oxa-
3-[3-[N-Methylamino-N-(3-methylbenzyl)]propyl]benzo[d]oxazol-
zol-2(3H)-one (3f). IR (Nujol): 1769, 2674 cm-1 1H NMR
.
1
2(3H)-one (2i). IR (Nujol): 1770, 2676 cm-1. H NMR (free
(CDCl3-TMS) ppm (δ): 1.54-1.91 (m, 2H, CH2-CH2-CH2-
CH2); 2.19 (s, 3H, CH3); 2.44 (t, 2H, CH2-N(CH3)-CH2-Ar, J=
6.6 Hz); 3.49 (s, 2H, N-CH2-Ar); 3.79 (s, 3H, CH3); 3.86 (t, 2H,
N-CH2-(CH2)3-, J=7.3 Hz); 6.81-7.28 (m, 8H, arom). MS: m/z
341 [MHþ].
3-[3-[N-(4-Methoxybenzyl)-N-methylamino]butyl]benzo[d ]oxa-
zol-2(3H)-one (3g). IR (Nujol): 1774, 2724 cm-1. 1H NMR (free
base, CDCl3-TMS) ppm (δ): 1.53-1.90 (m, 4H, CH2-CH2-CH2-
CH2, J = 6.7-7.3 Hz); 2.18 (s, 3H, CH3); 2.40 (t, 2H, CH2-
N(CH3)-CH2-Ar, J=6.7 Hz); 3.42 (s, 2H, N-CH2-Ar); 3.86 (m,
5H, OCH3 and N-CH2-(CH2)3-, J=7.3 Hz); 6.84-7.30 (m, 8H,
arom). MS: m/z 341 [MHþ].
base, CDCl3-TMS) ppm (δ): 1.84-1.98 (m, 2H, CH2-CH2-
CH2, J=6.6-7.3 Hz); 2.12 (s, 3H, CH3); 2.26 (s, 3H, CH3); 2.40
(t, 2H, CH2-N(CH3)-CH2-Ar, J=6.6 Hz); 3.38 (s, 2H, N-CH2-
Ar); 3.84 (t, 2H, N-CH2-(CH2)2-, J=7.3 Hz); 6.91-7.18 (m, 8H,
arom). MS: m/z 311 [MHþ].
3-[3-[N-Methylamino-N-(4-methylbenzyl)]propyl]benzo[d]oxazol-
1
2(3H)-one (2j). IR (Nujol): 1766, 2675 cm-1. H NMR (free
base, CDCl3-TMS) ppm (δ): 1.85-1.99 (m, 2H, CH2-CH2-
CH2, J=6.6-7.3 Hz); 2.14 (s, 3H, CH3); 2.27 (s, 3H, CH3); 2.42
(t, 2H, CH2-N(CH3)-CH2-Ar, J=6.6 Hz); 3.43 (s, 2H, N-CH2-
Ar); 3.82 (t, 2H, N-CH2-(CH2)2-, J=7.3 Hz); 6.92-7.20 (m, 8H,
arom). MS: m/z 311 [MHþ].
3-[3-[N-(2,4-Dimethylbenzyl)-N-methylamino]propyl]benzo[d ]-
oxazol-2(3H)-one (2k). IR (Nujol): 1777, 2725 cm-1. 1H NMR
(free base, CDCl3-TMS) ppm (δ): 1.97 (m, 2H, CH2-CH2-CH2,
J=6.7-7.3 Hz); 2.21 (s, 3H, N-CH3); 2.33 (s, 3H, CH3); 2.35
(s, 3H, CH3); 2.50 (t, 2H, CH2-N(CH3)-CH2-Ar, J=6.7 Hz);
3.43 (s, 2H, N-CH2-Ar); 3.86 (t, 2H, N-CH2-(CH2)2-, J=7.3
Hz); 6.95-7.30 (m, 7H, arom). MS: m/z 325 [MHþ].
3-[4-(N-Methylamino-N-(2-methylbenzyl)]benzo[d]oxazol-2(3H)-
one (3h). IR (Nujol): 1769, 2668 cm-1. 1H NMR (CDCl3-TMS)
ppm (δ): 1.49-1.90 (m, 4H, CH2-CH2-CH2-CH2); 2.15 (s, 3H,
CH3); 2.33 (s, 3H, CH3); 2.41 (t, 2H, CH2-N(CH3)-CH2-Ar, J=
6.6 Hz); 3.41 (s, 2H, N-CH2-Ar); 3.79 (t, 2H, N-CH2-(CH2)3-, J=
7.3 Hz); 6.89-7.25 (m, 8H, arom). MS: m/z 325 [MHþ].
3-[4-(N-Methylamino-N-(3-methylbenzyl)]benzo[d]oxazol-2(3H)-
1
one (3i). IR (Nujol): 1770, 2666 cm-1. H NMR (CDCl3-TMS)
3-[4-(N-Benzyl-N-methylamino)butyl]benzo[d ]oxazol-2(3H)-
one (3a). A mixture of 3-(4-bromobutyl)benzo[d ]ossazol-2(3H)-
one (0.25 g, 0.926 mmol), N-methylbenzylamine (0.089 g,
0741 mmol), anhydrous K2CO3 (0.72 g, 5.19 mmol), and a
catalytic amount of KI was dissolved in 50 mL of ACN, and
the solution was refluxed and monitored by TLC. The inorganic
salts were filtered off, and the solvent was evaporated under
reduce pressure. Distilled water (50 mL) was added, and the
residue was extracted 3 times with CHCl3 (3 ꢀ 100 mL). The
organic phase was separated and dried over anhydrous Na2SO4.
The filtered solution was concentrated under reduced pressure
and the remaining oil was treated with an equimolar amount of
oxalic acid in absolute ethanol to afford the oxalate salt, which
was filtered and washed with cold ethanol. Yield 0.12 g (0.09 g of
ppm (δ): 1.50-1.91 (m, 4H, CH2-CH2-CH2-CH2); 2.15 (s, 3H,
CH3); 2.31 (s, 3H, CH3); 2.38 (t, 2H, CH2-N(CH3)-CH2-Ar, J=
6.6 Hz); 3.40 (s, 2H, N-CH2-Ar); 3.81 (t, 2H, N-CH2-(CH2)3-, J=
7.3 Hz); 6.92-7.22 (m, 8H, arom). MS: m/z 325 [MHþ].
3-[4-(N-Methylamino-N-(4-methylbenzyl)]benzo[d]oxazol-2(3H)-
1
one (3j). IR (Nujol): 1769, 2674 cm-1. H NMR (CDCl3-TMS)
ppm (δ): 1.51-1.92 (m, 4H, CH2-CH2-CH2-CH2); 2.16 (s, 3H,
CH3); 2.33 (s, 3H, CH3); 2.39 (t, 2H, CH2-N(CH3)-CH2-Ar, J=
6.6Hz);3.43(s,2H,N-CH2-Ar); 3.82 (t, 2H, N-CH2-(CH2)3-, J=7.3
Hz); 6.93-7.23 (m, 8H, arom). MS: m/z 325 [MHþ].
3-[4-[N-(2,4-Dimethylbenzyl)-N-methylamino]butyl]benzo[d]oxa-
zol-2(3H)-one (3k). IR (Nujol): 1779, 2724 cm-1. 1H NMR (free
base, CDCl3-TMS) ppm (δ): 1.50-1.88 (m, 4H, CH2-CH2-CH2-
CH2, J=6.7-7.3 Hz); 2.17 (s, 3H, N-CH3); 2.32 (s, 6H, 2 ꢀ CH3);
2.42 (t, 2H, CH2-N(CH3)-CH2-Ar, J=6.7 Hz); 3.40 (s, 2H, N-CH2-
Ar); 3.82 (t, 2H, N-CH2-(CH2)3-, J=7.3 Hz); 6.90-7.30 (m, 7H,
arom). MS: m/z 341 [MHþ].
Synthesis of Carboxamide Derivatives. N-Phenyl-1-(tert-but-
oxycarbonyl)piperidine-4-carboxamide (8).45 To a mixture of
N-Boc-4-piperidinecarboxylic acid (2.12 g, 9.22 mmol), pyridine
(1.90 mL, 23,6 mmol) and CH2Cl2 (15 mL) and SOCl2 (0.80 mL,
11.0 mmol) were added, under N2 at room temperature, while
stirring. After 25 min, a solution of aniline (0.95 g, 10.2 mmol),
Et3N (4.50 mL, 32.3 mmol), and a catalytic amount of DMAP in
CH2Cl2 (15 mL) was added dropwise. The reaction was mon-
itored by TLC. After 14 h, the organic phase was washed with
1 N HCl (2 ꢀ 20 mL) and distilled water (2 ꢀ 20 mL), dried with
Na2SO4, and concentrated in vacuum to give 1.70 g (60.1%) of
a light-brown solid; mp 140-142C. IR (Nujol): 1656, 1685,
3257 cm-1. 1H NMR (CDCl3-TMS) ppm (δ): 1.44 (s 9H, CH3,
free base, 30%); mp 145-150 °C. IR (Nujol): 1772, 2668 cm-1
.
1H NMR (free base, CDCl3-TMS) ppm (δ): 1.50-2.00 (m, 4H,
CH2-CH2-CH2-CH2, J=6.7-7.3 Hz); 2.20 (s, 3H, N-CH3); 2.40
(t, 2H, CH2-N(CH3)-CH2-Ar, J=6.7 Hz); 3.45 (s, 2H, N-CH2-
Ar); 3.85 (t, 2H, N-CH2-(CH2)3-, J=7.3 Hz); 6.90-7.40 (m, 9H,
arom). MS: m/z 311 [MHþ].
Compounds 3b-k were synthesized following the same route
describe above for compound 3a.
3-[4-[N-(2-Chlorobenzyl)-N-methylamino]butyl]benzo[d ]oxazol-
2(3H)-one (3b). IR (Nujol): 1770, 2674 cm-1 1H NMR
.
(CDCl3-TMS) ppm (δ): 1.49-1.90 (m, 4H, CH2-CH2-CH2-
CH2); 2.13 (s, 3H, CH3); 2.39 (t, 2H, CH2-N(CH3)-CH2-Ar, J=
6.6Hz);3.40(s, 2H, N-CH2-Ar); 3.81 (t, 2H, N-CH2-(CH2)3-, J=7.3
Hz); 6.90-7.25(m, 8H, arom). MS:m/z345 [MHþ] 347 [MHþ þ 2].
3-[4-[N-(3-Chlorobenzyl)-N-methylamino]butyl]benzo[d ]oxazol-
2(3H)-one (3c). IR (Nujol): 1766, 2670 cm-1 1H NMR
.
0
0
(CDCl3-TMS) ppm (δ): 1.49-1.90 (m, 4H, CH2-CH2-CH2-
CH2); 2.14 (s, 3H, CH3); 2.39 (t, 2H, CH2-N(CH3)-CH2-Ar, J=
6.6Hz);3.40(s, 2H, N-CH2-Ar); 3.82 (t, 2H, N-CH2-(CH2)3-, J=7.3
Hz); 6.91-7.27(m, 8H, arom). MS:m/z345 [MHþ] 347 [MHþ þ 2].
3-[4-[N-(4-Chlorobenzyl)-N-methylamino]butyl]benzo[d]oxazol-
Boc); 1.60-1.93 (m, 4H, H3,3 -H5,5 , pip.); 2.36 (m, 1H, H4, pip.);
0
0
2.76(m, 2H, H2-H6, pip.); 4.17 (m, 2H, H2 -H6 , pip.); 7.05-7.52 (m,
6H, arom þ NH disappearing on deuteration). MS: m/z305 [MHþ].
Compound 9 was synthesized in analogous way using ben-
zylamine instead of aniline. Yield: 1.52 g (52%).
2(3H)-one (3d). IR (Nujol): 1768, 2676 cm-1
.
1H NMR
N-Benzyl-1-(tert-butoxycarbonyl)piperidine-4-carboxamide (9).
IR (Nujol): 1634, 1681, 3251 cm-1 1H NMR (CDCl3-TMS)
.
(CDCl3-TMS) ppm (δ): 1.50-1.65 (m, 2H, CH2-CH2-CH2);

5390 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 17
Zampieri et al.
0
0
0
0
ppm (δ): 1.38 (s, 9H, CH3,Boc);1.45-1.77(m,4H,H3,3 -H5,5 ,pip.);
CH3); 3.00 (m, 2H, H2 -H6 , pip.); 3.53 (s, 2H, N-CH2-Ar); 4.51
(s, 2H, Ar-CH2-N-CO); 5.84 (s broad, 1H, NH, disappearing
on deuteration); 7.18-7.45 (m, 9H, arom). MS: m/z 323
[MHþ].
0
2.21 (m, 1H, H4, pip.); 2.65 (m, 2H, H2-H6, pip.); 4.06 (m, 2H, H2 -
0
H6 , pip.); 4.36 (s, 2H, Ar-CH2-N-CO); 5.90 (s broad, 1H, NH,
disappearing on deuteration); 7.08-7.30 (m, 5H, arom). MS: m/z
319 [MHþ].
Pharmacology. Radioligand Binding Assays. Binding assays
were performed on rat liver membranes according to the
methods of Hellewell23 and were slightly modified as pre-
viously described.33 Briefly, for the σ1 receptor assay 250 μg of
rat liver homogenate was incubated for 120 min at 37 °C with
1 nM [3H]-(þ)-pentazocine (PerkinElmer, specific activity
34.9 Ci/mmol) in 50 mM Tris-HCl, pH 8.0, 0.5 mL final volume.
Nonspecific binding was defined in the presence of 10 μM
haloperidol. The reaction was stopped by vacuum filtration
through GF/B glass-fiber filters presoacked with 0.5% poly-
ethylenimine, followed by rapid washing with 2 mL of ice-cold
buffer. The filters were placed in 3 mL of scintillation cocktail,
and the radioactivity was determined by liquid scintillation
counting.
N-Phenylpiperidine-4-carboxamide (10).46 Compound 8 (1.00 g,
3.24 mmol) was deprotected with 2 mL of trifluoroacetic acid at
room temperature for 24 h. The solution was concentrated at
reduced pressure, and the residue was poured into water and
basified with NaOH 10% solution (pH 10). The mixture was
extracted with AcOEt (3 ꢀ 25 mL). The organic phase was dried
with Na2SO4 and concentrated. The obtained solid was used
without further purification. Yield: 0.60 g (89%); mp 105-
107 °C. IR (Nujol): 1655, 3258 cm-1. 1H NMR (CDCl3-TMS)
0
0
ppm (δ): 1.57-1.96 (m, 5H, H3,3 -H5,5 , pip. þ NH disappearing on
deuteration); 2.33 (m, 1H, H4, pip.); 2.57 (m, 2H, H2-H6, pip.); 3.13
0
0
(m, 2H, H2 -H6 , pip.); 7.04-7.50 (m, 6H, arom þ NH disappear-
ing on deuteration). MS: m/z 205 [MHþ].
Compound 11 was synthesized starting from 9 (1.2 g) using
the same procedure. Yield: 0.75 g (91%); mp115-117 °C.
N-Benzylpiperidine-4-carboxamide (11).47 IR (Nujol): 1634,
For the σ2 receptor assay, 150 μg of rat liver homogenate was
incubated for 120 min at room temperature with 3 nM [3H]DTG
(PerkinElmer, specific activity 58.1 Ci/mmol) in 50 mM Tris-
HCl, pH 8.0, 0.5 mL final volume. (þ)-Pentazocine (100 nM)
and haloperidol (10 μM) were used to mask σ1 receptors and to
define nonspecific binding, respectively.
1
3248 cm-1. H NMR (CDCl3-TMS) ppm (δ): 1.50-1.78 (m,
0
0
4H, H3,3 -H5,5 , pip.); 2.11-2.26 (m, 2H, H4, pip. þ NH dis-
appearing on deuteration); 2.52 (m, 2H, H2-H6, pip.); 3.05 (m,
0
0
2H, H2 -H6 , pip.); 4.32 (s, 2H, Ar-CH2-N-CO); 6.13 (s.broad,
1H, NH, disappearing on deuteration); 7.10-7.34 (m, 5H,
arom). MS: m/z 219 [MHþ].
Competition studies were done using at least 11 different
concentrations of the ligand under investigation. As an internal
control, three increasing concentrations of unlabeled (þ)-pent-
azocine (σ1 receptors) or DTG (σ2 receptors) were always
included. The compounds were prepared as 10 mM stock
solutions in 100% DMSO and diluted with Tris-HCl buffer on
the day of the experiment. The final DMSO concentration in the
incubation tubes was maintained at 0.1%.
The competition data for two to four separate determinations
performed in duplicate were averaged by fitting to a four-
parameter curve by means of the SigmaPlot software. Calcu-
lated IC50 values and Hill’s coefficients (nH) are reported
as mean values ( SEM. The corresponding Ki values were
obtained by the Cheng-Prusoff equation, as previously
reported.49
1-Benzyl-N-phenylpiperidine-4-carboxamide (6a). A solution
of 10 (0.17 g, 0.84 mmol), K2CO3 (0.14 g, 1.00 mmol), and benzyl
chloride (0.11 g, 0.84 mmol) in 50 mL of acetone was stirred
under reflux for 5 h. After the mixture was cooled, the inorganic
salt was filtered and the solvent evaporated at reduced pressure.
The residue was washed with distilled water and then with ethyl
ether to afford 6a as a chromatographically pure solid: yield
0.18 g (66%); mp 138-140 °C. IR (Nujol): 1657, 3318 cm-1
.
1H NMR (CDCl3-TMS) ppm (δ): 1.80-2.32 (m, 7H, H2-H3,3
-
0
0 0 0
5,5 -H6-H4, pip.); 3.02 (m, 2H, H2 -H6 , pip.); 3.57 (s, 2H,
H
N-CH2-Ar); 7.06-7.55 (m, 11H, arom þ NH disappearing on
deuteration). MS: m/z 295 [MHþ].
Compounds 6b and 6c were synthesized according to the
procedure reported above. Compounds 7a, 7b, and 7c were also
synthesized following the same route but starting from 11.
N-Phenyl-1-(4-chlorobenzyl)piperidine-4-carboxamide (6b).
IR (Nujol): 1656, 3301 cm-1. 1H NMR (CDCl3-TMS) ppm (δ):
Molecular Modeling. Training and Test Sets. The model
structures of all compounds were built using the Catalyst 2D-
3D sketcher.38 High quality conformational models are crucial
for the development of predictive pharmacophore models.
Accordingly, in this study we employed an ad hoc procedure
to derive molecular conformations, instead of using those
generated by Catalyst, for better quality in covering the low-
energy conformational space. Each molecular structure was
subjected to energy minimization using the generalized
CHARMM force field50 until the gradient dropped below
0.05. The minimized structures were used as the starting point
for subsequent conformational searches. A 10000-step Monte
Carlo torsional sampling conformational search was conducted
for each compound. Unique low-energy conformations within
20 kcal/mol of the corresponding global energy minimum were
collected for each molecule. A conformation was considered
unique only when the maximum displacement of at least
0
0
1.63-2.26 (m, 7H, H2-H3,3 -H5,5 -H6-H4, pip.); 2.93 (m, 2H,
0
0
H2 -H6 , pip.); 3.47 (s, 2H, N-CH2-Ar); 7.06-7.53 (m, 10H,
arom þ NH disappearing on deuteration). MS: m/z 329 [MHþ]
331 [MHþ þ 2].
N-Phenyl-1-(4-methylbenzyl)piperidine-4-carboxamide (6c).
1
IR (Nujol): 1654, 3292 cm-1. H NMR (CDCl3-TMS) ppm
0
0
(δ): 1.62-2.29 (m, 7H, H2-H3,3 -H5,5 -H6-H4, pip.); 2.34 (s,
0
0
3H, CH3); 2.97 (m, 2H, H2 -H6 , pip.); 3.48 (s, 2H, N-CH2-Ar);
7.10-7.53 (m, 10H, arom þ NH disappearing on deuteration).
MS: m/z 309 [MHþ].
N,1-Dibenzylpiperidine-4-carboxamide (7a).48 IR (Nujol):
1634, 3248 cm-1
.
1H NMR (CDCl3-TMS) ppm (δ):
˚
0
0
1.65-2.13 (m, 7H, H2-H3,3 -H5,5 -H6-H4, pip.); 2.84 (m, 2H,
one heavy atom was greater than 0.5 A. A maximum of 250
0
0
H2 -H6 , pip.); 3.40 (s, 2H, N-CH2-Ar); 4.40 (s, 2H, Ar-CH2-N-
CO); 5.75 (s.broad, 1H, NH, disappearing on deuteration);
7.13-7.36 (m, 10H, arom). MS: m/z 309 [MHþ].
unique conformations were recovered for each compound. The
classical conformational search was also carried out using
the Poling algorithm51-53 and the CHARMM force field51
as implemented in the Catalyst program for comparison. The
“best quality” generation option was adopted to select repre-
sentative conformers over a 0-20 kcal/mol interval above the
computed global energy minimum in the conformational space,
and again, the number of conformers generated for each com-
pound was limited to a maximum of 250. Comparing the results
of the two conformational searches, we verified the existence of
considerable differences between the two approaches in gener-
ating conformations for saturated six-member rings such
as piperidine. This group is quite common in drug molecules
N-Benzyl-1-(4-chlorobenzyl)piperidine-4-carboxamide (7b).
IR (Nujol): 1633, 3247 cm-1. H NMR (CDCl3-TMS) ppm
1
0
0
(δ): 1.62-2.20 (m, 7H, H2-H3,3 -H5,5 -H6-H4, pip.); 2.87
0
0
(m, 2H, H2 -H6 , pip.); 3.43 (s, 2H, N-CH2-Ar); 4.37 (s, 2H,
Ar-CH2-N-CO); 5.78 (s broad, 1H, NH, disappearing on
deuteration); 7.14-7.32 (m, 9H, arom). MS: m/z 343 [MHþ]
345 [MHþ þ 2].
N-Benzyl-1-(4-methylbenzyl)piperidine-4-carboxamide (7c).
IR (Nujol): 1636, 3251 cm-1. H NMR (CDCl3-TMS) ppm
1
0
0
(δ): 1.80-2.29 (m, 7H, H2-H3,3 -H5,5 -H6-H4, pip.); 2.41 (s, 3H,

Article
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 17 5391
and constitutes a popular molecular scaffold. A survey of the
crystal structures of druglike molecules and protein/ligand
complexes available in the literature and in public databases
reveals that this saturated ring overwhelmingly adopts low-
energy chair conformations. The conformational search con-
ducted with the typical Catalyst settings described above, how-
ever, generated predominantly twisted conformations that
might lead to an incorrect mapping of this functionally impor-
tant group. In comparison, the alternative procedure of con-
formational search produced a considerable number of chair
conformations for this heterocyclic moiety. The main drawback
of this technique, however, is that it takes considerable longer to
generate the relevant conformational models. Nonetheless,
as the spirit of the work was the generation of a predictive
3D pharmacophore model for these classes of compounds,
we considered it worthwhile to use more accurate conforma-
tional models.
On the basis of the conformations for each compound, the
HypoGen module of the Catalyst 4.9 software package was used
to generate three-dimensional pharmacophore models. During
hypotheses generation, the software attempts to minimize a cost
function containing two main terms: the first penalizes the
deviation between the estimated affinities of the training set
molecules and their experimental values, while the second
penalizes the complexity of the hypothesis. The uncertainty
factor for each compound represents the ratio range of uncer-
tainty in the affinity value based on the expected statistical
irregularity of biological data collection. Uncertainty influences
the first step (also called the constructive phase) of the hypoth-
esis generating process. In this work, an uncertainty of 1.1 was
preferred over the default factor of 3.0, as the experimental
affinities of our compounds barely span the required 4 orders of
magnitude.
Briefly, a pharmacophore captures the three-dimensional
arrangement of the structural features shared by all active
molecules that are presumably essential for the desired phar-
macological activity. These features include hydrogen bond
donors (HBD) and acceptors (HBA), hydrophobic groups
(HY), aromatic rings (RA), positively charged/ionizable groups
(PC/PI), and negatively charged/ionizable (NC/NI) moieties. In
addition, shape restraints and excluded volume effects can also
be incorporated in the 3D-QSAR pharmacophores to account
for the framework of the target active site. These popular
modeling techniques find many practical applications in drug
design. For instance, they can be used to align structurally
unrelated lead compounds, thus identifying the groups in
each molecular structure that play an important role for the
corresponding biological activity. Moreover, the nonessential
parts of the molecules can be altered to improve their physico-
chemical or pharmacokinetic properties, and new molecular
scaffolds can be designed to establish novel patent space.
Last but not least, pharmacophore models have been applied
with success in virtual screening to extract molecular entities for
biological activity testing from large, proprietary, or public
databases.54
the best pharmacophore hypothesis. The high correlation
coefficients obtained using the test set compounds revealed
the good correlation between the actual and estimated
affinities and hence the predictive validity of the correspond-
ing 3D hypothesis. The CatScramble validation procedure
is based on Fisher’s randomization test.55 The goal of this
type of validation is to check whether there is a strong
correlation between the chemical structures and the biologi-
cal activity. This is done by randomizing the affinity data
associated with the training set compounds, generating phar-
macophore hypotheses using the same features and para-
meters employed to develop the original pharmacophore
model. The statistical significance is calculated according to
the following formula:
significance ¼ 100 ꢀ ½1 -ð1þx=yÞꢁ
where x is the total number of hypotheses having a total cost
lower than the original (best) hypothesis and y is the total
number of HypoGen runs (initial þ random runs). Thus,
49 random spreadsheets (i.e., 49 HypoGen runs) have to be
generated to obtain a 98% confidence level. Should any
randomized data set result in the generation of a 3D pharma-
cophore with similar or even better cost values, root-
mean-square deviations, and correlation coefficients, then it
is likely that the original hypothesis does reflect a chance
correlation.
Finally, the leave-one-out test checks if the correlation bet-
ween experimental and computed affinities is heavily dependent
on one particular molecule of the training set by recomputing
the pharmacophore model with the exclusion of one molecule at
a time. Accordingly, 31 new training sets were derived, each
composed of 30 molecules, and 31 HypoGen calculations were
performed under the same conditions. For each run, the hypoth-
esis characterized by the lowest total cost was employed to
predict the affinity of the excluded compound and to estimate
the new correlation coefficient.
All PDB structures and 3D hypotheses generated in this work
are available from the authors upon request.
Acknowledgment. Financial support from the Italian
Ministry of University and Scientific Research (MIUR,
Rome) (Project PRIN-2005) is gratefully acknowledged.
The authors thank Dr. Tien Luu for thoroughly reading the
manuscript and her critical comments.
Supporting Information Available: Parameters for the 10 top
hypotheses generated, elemental analysis results of compounds
2, 3, 6, and 7, and table with Hill coefficients for compounds
6a-c and 7a-c. This material is available free of charge via the
Internet at http://pubs.acs.org.
References
(1) Martin, W. R.; Eades, C. E.; Thompson, J. A.; Huppler, R. E.;
Gilbert, P. E. The effects of morphine- and nalorphine-like drugs in
the nondependent and morphine-dependent chronic spinal dog.
J. Pharmacol. Exp. Ther. 1976, 197, 517–532.
An analysis of the functional groups characterizing our
compounds suggested that hydrophobic aromatic (HYAr) and
aliphatic (HYAl) features, hydrogen bond acceptors (HBA),
positive ionizable (PI), and ring aromatic (RAr) features could
effectively map the critical chemical features and hence describe
the σ1 receptor affinity of our compounds. Accordingly, these
five features were selected to constitute the essential information
in the automated hypothesis generation process.
(2) Bowen, W. D. Sigma receptors: recent advances and new clinical
potentials. Pharm. Acta Helv. 2000, 74, 211–218.
(3) Hellewell, S. B.; Bowen, W. D. A sigma-like binding site in rat
pheochromocytoma (PC12) cells: decreased affinity for (þ)-benzo-
morphans and lower molecular weight suggest a different sigma
receptor form from that of guinea pig brain. Brain Res. 1990, 527,
235–236.
Three validation procedures were used to determine the
statistical relevance and the validity of the proposed 3D
pharmacophore models: the test set prediction method, the
CatScramble method, and the leave-one-out procedure. In
this work, the first procedure consisted of a collection of
further, different compounds into a test set and of performing
a regression analysis by mapping the test set molecules onto
(4) Itzhak, Y.; Stein, I. Regulation of sigma receptors and responsive-
ness to guanine nucleotides following repeated exposure of rats to
haloperidol: further evidence for multiple sigma binding sites.
Brain Res. 1991, 566, 166–172.
(5) Quirion, R; Bowen, W. D.; Itzhak, Y.; Junien, J. L.; Mustacchio,
J. M.; Rothman, R. B.; Su, T. P.; Tam, S. W.; Taylor, D. P. A
proposal for the classification of sigma binding sites. Trends
Pharmacol. Sci. 1992, 13, 85–86.

5392 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 17
Zampieri et al.
(6) McLean, S.; Weber, E. Autoradiographic visualization of haloper-
idol-sensitive sigma receptors in guinea-pig brain. Neuroscience
1988, 25, 259–269.
(7) Matsumoto, R. R.; Hemstreet, M. K.; Lai, N. L.; Thurkauf, A.;
De Costa, B. R.; Rice, K. C.; Helleweel, S. B.; Bowen, W. D.;
Walker, J. M. Drug specificity of pharmacological dystonia.
Pharmacol., Biochem. Behav. 1990, 36, 151–155.
(28) Vilner, B. J.; Bowen, W. D. Modulation of cellular calcium by
sigma-2 receptors: release from intracellular stores in human SK-
N-SH neuroblastoma cells. Pharmacol. Exp.Ther. 2000, 292, 900–
911.
(29) Crawford, K. W.; Coop, A.; Bowen, W. D. σ2 receptors regulate
changes in sphingolipid levels in breast tumor cells. Eur. J. Phar-
macol. 2002, 443, 207–209.
(8) Wolfe, S. A., Jr.; Culp, S. G.; De Souza, E. B. Sigma-receptors in
endocrine organs: identification, characterization, and autoradio-
graphic localization on rat pituitary, adrenal, testis, and ovary.
Endocrinology 1989, 124, 1160–1172.
(9) Su, T. P.; London, E. D.; Jaffe, J. H. Steroid binding at sigma
receptors suggests a link between endocrine, nervous, and immune
systems. Science 1988, 240, 219–221.
(10) Marrazzo, A.; Prezzavento, O.; Pappalardo, M. S.; Bousquet, E.;
Iadanza, M.; Pike, V. W.; Ronsisvalle, G. Synthesis of (þ)- and
(-)-cis-2-((1-adamantylamino)-methyl)-1-phenylcyclopropane de-
rivatives as high affinity probes for sigma1 and sigma2 binding
sites. Farmaco 2002, 57, 45–53.
(11) Patrick, S. L.; Walker, J. M.; Perkel, J. M.; Lockwood, M.; Patrick,
R. L. Increases in rat striatal extracellular dopamine and vacuous
chewing produced by two sigma receptor ligands. Eur. J. Pharma-
col. 1993, 231, 243–249.
(12) Hanner, M.; Moebius, F. F.; Flandorfer, A.; Knaus, H. G.;
Striessnig, J.; Kempner, E.; Glossmann, H. Purification, molecular
cloning, and expression of the mammalian sigma1-binding site.
Proc. Natl. Acad. Sci. U.S.A. 1996, 93, 8072–8077.
(13) Kekuda, R.; Prasad, P. D.; Fei, J.-Y.; Leibach, F. H.; Ganapathy,
V. Cloning and functional expression of the human type 1 sigma
receptor (hSigmaR1). Biochem. Biophys. Res. Commun. 1996, 229,
553–558.
(30) Caveliers, V.; Everaert, H.; Lahoutte, T.; Dierickx, L. O.; John, C.
S.; Bossuyt, A. Labelled sigma receptor ligands: can their role in
neurology and oncology be extended? Eur. J. Nucl.Med. Mol.
Imaging 2001, 28, 133–135.
(31) Friebe, M.; Mahmood, A.; Bolzati, C.; Drews, A.; Johannsen, B.;
Eisenhut, M.; Kraemer, D.; Davison, A.; Jones, A. G. 99mTc
oxotechnetium(V) complexes amine-amide-dithiol chelates with
dialkylaminoalkyl substituents as potential diagnostic probes for
malignant melanoma. J. Med. Chem. 2001, 44, 3132–3140.
(32) Choi, S. R.; Yang, B.; Plossl, K.; Chumpradit, S.; Wey, S. P.;
Acton, P. D.; Wheeler, K. T.; Mach, R. H.; Kung, H. F. Develop-
ment of a Tc-99m labeled sigma-2 receptor-specific ligand as a
potential breast tumor imaging agent. Nucl. Med. Biol. 2001, 28,
657–666.
(33) Zampieri, D.; Mamolo, M. G.; Laurini, E.; Zanette, C.; Florio, C.;
Collina, S.; Urbano, M.; Azzolina, O.; Vio, L. Substituted benzo-
(d)oxazol-2(3H)-one derivatives with preference for the sigma1
binding site. Eur. J. Med. Chem. 2009, 44, 124–30.
(34) Glennon, R. A.; Ablordeppey, S. Y.; Ismaiel, A. M.; El-Ashmawy,
M. B.; Fischer, J. B.; Howie, K. B. Structural features important for
sigma.1 receptor binding. J. Med. Chem. 1994, 37, 1214–19.
(35) Ablordeppey, S. Y.; Fischer, J. B.; Law, H.; Glennon, R. A.
Probing the proposed phenyl-A region of the sigma-1 receptor.
Bioorg. Med. Chem. 2002, 10, 2759–2765.
(14) Wyrick, S. D.; Booth, R. C. Progress in sigma receptor research.
(36) Glennon, R. A. Pharmacophore identification for sigma-1 (σ1)
receptor binding: application of the “deconstruction-recon-
struction-elaboration” approach. Mini-Rev. Med. Chem. 2005,
5, 927–940.
(37) Gund, T. M.; Floyd, J.; Jung, D. Molecular modeling of σ1
receptor ligands: a model of binding conformational and electro-
static considerations. J. Mol. Graphics Modell. 2004, 22, 221–230.
(38) Catalyst, version 4.9; Accelrys Inc.: San Diego, CA.
(39) Wang, H.; Duffy, R. A.; Boykow, G. C.; Chackalamannil, S.;
Madison, V. S. Identification of novel cannabinoid CB1 receptor
antagonist by using virtual screening with a pharmacophore model.
J. Med. Chem. 2008, 51, 2493–2446.
(40) Guner, O.; Clement, O.; Kurogi, Y. Pharmacophore modeling
and three-dimensional data bases searching for drug desing
using Catalyst: recent advances. Curr. Med. Chem. 2004, 11,
2991–3005.
(41) Sutter, J.; Guner, O. F.; Hoffman, R. D.; Li, H.; Wadman, M.
Effect of Variable Weight and Tolerances on Predictive Model
Generation. In Pharmacophore Perception, Development, and Use
in Drug Design; Guner, O. F., Ed.; International University Line: La
Jolla, CA, 1999; pp 501-511.
Drugs Future 1995, 20, 1033–1044.
(15) Debonnel, G.; de Montigny, C. Modulation of nmda and dopa-
minergic neurotransmissions by sigma ligands: possible implica-
tions for the treatment of psychiatric disorders. Life Sci. 1996, 58,
721–734.
(16) Monnet, F. P.; Debonnel, G.; de Montigny, C. In vivo electro-
physiological evidence for a selective modulation of N-methyl-
D-aspartate-induced neuronal activation in rat CA3 dorsal hippo-
campus by sigma ligands. J. Pharmacol. Exp. Ther. 1992, 261, 123–
130.
(17) Maurice, T.; Lockhart, B. P. Neuroprotective and anti-amnesic
potentials of sigma (σ) receptor ligands. Prog. Neuro-Psychophar-
macol. Biol. Psychiatry 1997, 21, 69–102.
(18) King, M.; Pan, X. Y.; Mei, J.; Chang, A.; Xu, J.; Pasternak, G. W.
Enhanced K-opioid receptor-mediated analgesia by antisense
targeting the σ1 receptor. Eur. J. Pharmacol. 1997, 331, R5–R6.
(19) McCracken, K. A.; Bowen, W. D.; Walker, F. O.; de Costa, B.;
Matsumoto, R. R. Two novel σ receptor ligands, BD1047 and
LR172, attenuate cocaine-induced toxicity and locomotor activity.
Eur. J. Pharmacol. 1999, 370, 225–232.
(20) Modell, S.; Naber, D.; Holzbach, R. Efficacy and safety of an
opiate sigma-receptor antagonist (SL 82.0715) in schizophrenic
patients with negative symptoms: an open dose-range study.
Pharmacopsychiatry 1996, 29, 63.
(21) Huber, M. T.; Gotthardt, U.; Schreiber, W.; Krieg, J. C. Efficacy
and safety of the sigma receptor ligand EMD 57445 (panamesine)
in patients with schizophrenia: an open clinical trial. Pharmaco-
psychiatry 1999, 32, 68–72.
(22) Walker, J. M.; Bowen, W. D.; Walker, F. O.; Matsumoto, R. R.; de
Costa, B.; Rice, K. C. Sigma receptors: biology and function.
Pharmacol. Rev. 1990, 42, 355–402.
(23) Hellewell, S. B.; Bruce, A.; Feinstein, G.; Orringer, J.; Williams,
W.; Bowen, W. D. Rat liver and kidney contain high densities
of sigma 1 and sigma 2 receptors: characterization by ligand
binding and photoaffinity labeling. Eur. J. Pharmacol. 1994, 268,
9–18.
(42) Laggner, C.; Schieferer, C.; Fiechtner, B.; Poles, G.; Hoffmann, R.
D.; Glossmann, H.; Langer, T.; Moebius, F. Discovery of high-
affinity ligands of s1 receptor, ERG2, and emopamil binding
protein by pharmacophore modeling and virtual screening.
J. Med. Chem. 2005, 48, 4754–4764.
(43) Sam, J.; Valentine, J. L.; Aboul-Enein, M. N. Reaction of
3-(choroalkyl)-2-benzoxazolinones with amines. Formation of
3-(aminoalkyl)-2-benzoxazolinones and 5-substituted 2,3,4,5-tetra-
hydro-1,5-benzoxazepines. J. Pharm. Sci. 1971, 60, 1370–1375.
(44) Mokrosz, M. J.; Kowalski, P.; Kowalska, T.; Majka, Z.; Duszynska,
B.; Bojarski, A. J.; Fruzinski, A.; Karolak-Wojciechowska, J.;
Wesolowska, A.; Klodzinska, A.; Tatarczynska, E.; Chojnacka-
Wojcik, E. Structure-activity relationship studies of CNS agents.
Part 38. Novel 1,4-benzoxazin-3(4H)-one, 1,2-benzoxazolin-3-one
and 1,3-benzoxazolin-2,4-dione arylpiperazine derivatives with dif-
ferent 5-HT1A and antagonistic 5-HT2A activities. Arch. Pharm.
1999, 332, 373–379.
(24) Bastianetto, S.; Perrault, G.; Sanger, D. J. Pharmacological
evidence for the involvement of sigma sites in DTG-induced
contralateral circling in rats. Neuropharmacology 1995, 34, 107–
114.
(45) Freund, R.; Mederski, W. W. K. R. A convenient synthetic route to
spiro(indole-3,40-piperidin)-2-ones. Helv. Chim. Acta 2000, 83,
1247–1255.
(25) Jeanjean, A. P.; Laterre, E. C.; Maloteaux, J. M. Neuroleptic
binding to sigma receptors: possible involvement in neuro-
leptic-induced acute dystonia. Biol. Psychiatry 1997, 41, 1010–
1019.
(26) Vilner, B. J.; John, C. S.; Bowen, W. D. Sigma-1 and sigma-2
receptors are expressed in a wide variety of human and rodent
tumor cell lines. Cancer Res. 1995, 55, 408–413.
(46) Lunn, G.; Sansone, E. B. Facile reduction of pyridines with
nickel-aluminum alloy. J. Org. Chem. 1986, 51, 513–517.
(47) Alabaster, V. A.; Campbell, S. F.; Danilewicz, J. C.; Greengrass,
C. W.; Plews, R. M. 2,4-Diamino-6,7-dimethoxyquinazolines.
2. 2-(4-Carbamoylpiperidino) derivatives as alpha 1-adrenoceptor
antagonists and antihypertensive agents. J. Med. Chem. 1987, 30,
999–1003.
(27) Crawford, K. W.; Bowen, W. D. Sigma-2 receptor agonists activate
a novel apoptotic pathway and potentiate antineoplastic drugs in
breast tumor cell lines. Cancer Res. 2002, 62, 313–322.
(48) Atrash, B.; Bradley, M.; Kobylecki, R.; Cowell, D.; Reader,
J. Revolutionizing resin handling for combinatorial synthesis.
Angew. Chem., Int. Ed. 2001, 40, 938–941.

Article
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 17 5393
(49) Mamolo, M. G.; Zampieri, D.; Zanette, C.; Florio, C.; Collina, S.;
Urbano, M.; Azzolina, O.; Vio, L. Substituted benzylaminoalk-
ylindoles with preference for the σ2 binding site. Eur. J. Med. Chem.
2008, 43, 2073–2081.
(50) Brooks, B. R.; Bruccoleri, R. E.; Olafson, B. D.; States, D. J.;
Swaminathan, S.; Karplus, M. J. CHARMM: a program for
macromolecular energy, minimization, and dynamics calculations.
J. Comput. Chem. 1983, 4, 187–217.
(52) Smellie, A.; Kahn, S. D.; Teig, S. L. Analysis of conformational
coverage. 1. Validation and estimation of coverage. J. Chem. Inf.
Comput. Sci. 1995, 35, 285–294.
(53) Smellie, A.; Kahn, S. D.; Teig, S. L. Analysis of conformational
coverage. 2. Applications of conformational models. J. Chem. Inf.
Comput. Sci. 1995, 35, 295–304.
(54) Langer, T.; Krovat, E. M. Chemical-feature based pharmaco-
phores and virtual library screening for discovery of new leads.
Curr. Opin. Drug Discovery Dev. 2003, 6, 370–376.
(51) Smellie, A.; Teig, S. L.; Towbin, P. Poling: promoting conforma-
tional variation. J. Comput. Chem. 1994, 16, 171–187.
(55) Fisher,R.The Design of Experiments; Hafner Publishing: New York, 1966.