The multiobjective based design, synthesis and evaluation of the arylsulfonamide/amide derivatives of aryloxyethyl and arylthioethyl-piperidines and pyrrolidines as a novel class of potent 5-HT7 receptor antagonists

Article abstract of DOI:10.1016/j.ejmech.2012.07.043

An analysis of the virtual combinatorial library was used for refining a pilot set of 34 derivatives and designing a targeted 38-member library of the arylamide and arylsulfonamide derivatives of aryloxyethyl- and arylthioethyl- piperidines and pyrrolidines. All compounds 24-95 were synthesized according to an elaborated parallel solid-phase method and were biologically evaluated for their affinity for 5-HT₇R. Additionally, the targeted library members were tested for 5-HT_1A, 5-HT₆, and D₂ receptors. Selected compounds of particular interest were examined for their intrinsic activity at 5-HT₇R in vitro employing a cAMP assay. The study allowed us to identify compound 68 (4-fluoro-N-(1-{2-[(propan-2-yl) phenoxy]ethyl}piperidin-4-yl) benzenesulfonamide) as a potent 5-HT₇R ligand (K_i = 0.3 nM) with strong antagonistic properties (K _b = 1 nM) and a 1450-fold selectivity over 5-HT_1ARs.

Full text of DOI:10.1016/j.ejmech.2012.07.043

European Journal of Medicinal Chemistry 56 (2012) 348e360
Contents lists available at SciVerse ScienceDirect
European Journal of Medicinal Chemistry
journal homepage: http://www.elsevier.com/locate/ejmech
Original article
The multiobjective based design, synthesis and evaluation of the arylsulfonamide/
amide derivatives of aryloxyethyl- and arylthioethyl- piperidines and pyrrolidines
as a novel class of potent 5-HT7 receptor antagonists
,
b
a
b
a
Pawe1 Zajdel a , Rafa1 Kurczab , Katarzyna Grychowska , Grzegorz Sata1a , Maciej Paw1owski ,
*
Andrzej J. Bojarski b
a Department of Medicinal Chemistry, Jagiellonian University Medical College, 9 Medyczna Street, 30-688 Kraków, Poland
b Department of Medicinal Chemistry, Institute of Pharmacology, Polish Academy of Sciences, 12 Sme˛tna Street, 31-343 Kraków, Poland
a r t i c l e i n f o
a b s t r a c t
Article history:
An analysis of the virtual combinatorial library was used for refining a pilot set of 34 derivatives and
designing a targeted 38-member library of the arylamide and arylsulfonamide derivatives of arylox-
yethyl- and arylthioethyl- piperidines and pyrrolidines. All compounds 24e95 were synthesized
according to an elaborated parallel solid-phase method and were biologically evaluated for their affinity
for 5-HT7R. Additionally, the targeted library members were tested for 5-HT1A, 5-HT6, and D2 receptors.
Selected compounds of particular interest were examined for their intrinsic activity at 5-HT7R in vitro
employing a cAMP assay. The study allowed us to identify compound 68 (4-fluoro-N-(1-{2-[(propan-2-yl)
phenoxy]ethyl}piperidin-4-yl) benzenesulfonamide) as a potent 5-HT7R ligand (Ki ¼ 0.3 nM) with strong
antagonistic properties (Kb ¼ 1 nM) and a 1450-fold selectivity over 5-HT1ARs.
Received 9 May 2012
Received in revised form
20 July 2012
Accepted 26 July 2012
Available online 4 August 2012
Keywords:
Solid-phase synthesis
BAL linker
Virtual combinatorial library
Virtual screening
Ó 2012 Elsevier Masson SAS. All rights reserved.
Long-chain arylpiperazines
Alicyclic amines
Arylsulfonamides
Aryloxyalkyl-piperidines
Aryloxyalkyl-pyrrolidines
5-HT7 antagonists
1. Introduction
efficient at exploring the chemical space and rapidly acquiring
meaningful structureeactivity relationship data for GPCR-oriented
Over the recent decades high-throughput chemistry methods
have been extensively used in pharmaceutical industry and
academic research for identifying and optimizing potential lead
compounds. The initial ideology behind the solid-phase synthesis
(SPS) underwent transformation from the synthesis of vast
combinatorial libraries, counting even 10 000 members and
showing a low structural diversity and a high attrition rate, to
generation of designed libraries with improved physicochemical
attributes. Nowadays, solid-phase chemistry often extends the
fragment-based drug discovery approach, and aims at combining
building blocks or fragments found in drugs or clinical candidates.
Several examples show that the solid-phase chemistry has been
drug discovery projects [1e3].
A wide variety of different computational approaches are used
to support the design of combinatorial libraries, their analysis, and
identification of the most attractive subsets. Virtual combinatorial
libraries (VCL) are usually created using two main strategies;
a reactant-based and a product-based design [4e6]. As regards to
the former, optimization is focused solely on reagent (building
blocks) selection, achieved for example by clustering or similarity
searching methods. In the case of product-based approaches, all
possible compounds are virtually synthesized from available reac-
tants, and the selection of a final combinatorial library is often
supported by virtual screening (VS) methods for scoring and
ranking of compounds. The obtained product ranking list is next
analyzed to select reactants for the real (often solid-supported)
synthesis. The product-based approach is computationally more
expensive but seems more promising due to the possibility of
* Corresponding author. Tel.: þ48 12 620 54 50; fax: þ48 12 620 54 05.
E-mail address: mfzajdel@cyf-kr.edu.pl (P. Zajdel).
0223-5234/$ e see front matter Ó 2012 Elsevier Masson SAS. All rights reserved.
http://dx.doi.org/10.1016/j.ejmech.2012.07.043

P. Zajdel et al. / European Journal of Medicinal Chemistry 56 (2012) 348e360
349
optimizing properties of the targeted library. Recently, dedicated
software for a focused library design have been developed: COLI-
BREE (Combinatorial Library Breeding) [7], LoFT (Library optimizer
using Feature Trees) [8], GLARE (Global Library Assessment of
Reagents) [9] and PICCOLO (PICking by COmbinatorial Library
Optimization) [10], but the entire process can also be controlled
semi-automatically using different tools.
A growing body of preclinical and clinical data supports the
hypothesis that ligands affectingthe 5-HT7 receptors (5-HT7Rs) may
help develop new therapies for the treatment of the affective
disorders [11e13]. Identified in 1993, the 5-HT7R is the latest addi-
tion to the 5-HT subtypes, and belongs to the family of G protein-
coupled receptors. Its distribution in the hypothalamus (particu-
larly in the suprachiasmatic nucleus), the thalamic and cortical
regions has been associated with the control of circadian rhythms
and involvement in mood regulation; the presence of 5-HT7Rs in the
hippocampus highlights their potential involvement in learning and
memoryas well as in emotional processes [14]. More recently, Abbas
et al. and Sarkisyan et al., have gathered strong evidence that the
antidepressant-like effects of well-known atypical antipsychotics
amisulpride and aripiprazole are mediated by 5-HT7R antagonism
[15,16]. 5-HT7R has also shown some potential for pain control,
prophylaxis or treatment of migraine and seizures [17].
and simultaneously employed partial rigidification in the part cor-
responding to the alkylene spacer of LCAP. At the same time, this
allowed us to explore the second structural homology by diversi-
fying the length between an amide/sulfonamide bond and a basic
nitrogen atom via introduction of three different cores: 3-
aminopyrrolidine, 4-aminopiperidine and 4-aminomethylpipe-
ridine (Fig. 2). Further modification consisted of the bioisosteric
replacement of an amide bond with a sulfonamide bond. Last but not
least, to additionally understand the configurational requirements
for compounds binding to 5-HT7R, we investigated enantiomers of
the selected representatives of the 3-aminopyrrolidine core.
The number of compounds that can be synthesized using an
elaborated protocol is proportional to the number of accessible
building blocks in the commercial chemical space. Therefore, we
developed an integrated VCL-VS protocol to support the solid-phase
methodology and enable the selection of the most promising
building blocks(which are simultaneously diverse and representative)
to prioritize the synthesis and to increase the probability of identifi-
cation of 5-HT7 receptor ligands. Our project involved affinity deter-
mination of the synthesized compounds for the predefined 5-HT7R, as
well as for 5-HT1A, 5-HT6, and dopamine D2 receptors, and functional
evaluation of the selected library representatives for 5-HT7Rs.
Several excellent reviews thoroughly have dealt with the
progress in the development of 5-HT7R ligands and have also
adequately described the proposed pharmacophore models for 5-
HT7R antagonism and agonism [17e19]. Arylsulfonamides, which
are analogs of SB-269970 and SB-656104, hold a pre-eminent
position in the group of 5-HT7R ligands. Another distinguished
class is represented by long-chain arylpiperazines (LCAP) with
tetrahydrobenzindole, arylketone and amide terminal fragments.
However, these compounds are often devoid of selectivity over 5-
HT1A, 5-HT2A or D2 sites [20e22]. Such obstacles were recently
overcome by Leopoldo et al. [23], who described a successful appli-
cation of 2-substituted phenylpiperazine, leading to generation of the
highly potent and selective 5-HT7R agonists LP-44 and LP-211 (Fig.1).
For several years our research group has been developing solid-
phase methodologies for the synthesis of new arylpiperazine-
based selective or multireceptor ligands of 5-HT1A, 5-HT2A, and 5-
HT7 receptors with potential antidepressant, anxiolytic, and anti-
psychotic properties [24e26]. In the present study we expanded our
SPS technology platform for the synthesis of the arylamide and
arylsulfonamide derivatives of aryloxyethyl- and arylthioethyl-
piperidines and pyrrolidines as novel 5-HT7R ligands. In comparison
with the classic LCAP, we replaced an arylpiperazine fragment with
a flexible aryloxy-/arylthio-ethyl, substituted tertiary amine core
2. Chemistry e library synthesis
The synthesis of the designed arylamide and arylsulfonamide
derivatives of aryloxyethyl- and arylthioethyl- piperidines and
pyrrolidines was carried out in two stages. First, starting with the
selected commercial phenols and thiols, we synthesized building
blocks BB2 e aryloxylethyl bromides and arylthioethyl bromides
for a solid-phase approach (for details see Supporting Information).
Treatment of the respective phenol derivatives (1e11) with 1,2-
dibromoethane in refluxing acetone in the presence of potassium
carbonate yielded aryloxylethyl bromides 22{1e11} (Scheme 1,
Pathway A). Alternatively, the respective arylthiols (12, 13) were
treated with an excess of 2-chloroethanol in a 10% solution of
sodium hydroxide to give arylthioethanol derivatives (14, 15). The
latter were readily converted into arylthioethyl bromide analogs
(22{12e13}) upon treatment with phosphorus tribromide (Scheme
1, Pathway B).
H
N
R1
N
classic LCAP
O
N
R
H
O
N
O
S
O
S
Cl
O
R
HO
O
O
N
N
N
H
N
R1
4-aminomethylpiperidine core
N
N
O
O
SB-656104
SB-269970
K (5-HT7)
= 3.2 nM
K (5-HT7) = 1.2 nM
K (5-HT1A) > 5 µM
i
i
K (5-HT1A) = 250 nM
i
i
H
N
R1
R
O
4-aminopiperidine core
3-aminopyrrolidine core
NC
N
H
H
N
N
N
3
S
N
3
O
N
O
N
O
N
HN
LP-44
LP-211
R
R1
K (5-HT7) = 0.22 nM
i
K (5-HT7)
i
= 0.58 nM
i
i
O
K (5-HT1A) = 52.7 nM
K (5-HT1A) = 188 nM
Fig. 2. General structures of the classic long-chain arylpiperazine vs the designed
flexible aryloxy-/arylthio-ethyl analogs.
Fig. 1. The structure of model 5-HT7R ligands.

350
P. Zajdel et al. / European Journal of Medicinal Chemistry 56 (2012) 348e360
XH
O
Boc
Boc
Boc
Br
N
N
i
N
H2N
R
R
H2N
Pathway A
H2N
Boc
1 - 13
22{1-11}
1
2
3
Boc
ii
Pathway B
R
N
N
H2N
H2N
S
S
4
OH
Br
5
iii
R
Fig. 3. The diversity of core amines 17{1e5}.
14 - 15
22{12-13}
different substituted aryl and heteroaryl sulfonyl and acyl chlorides
(BB1 queries), and the second contained general formula for ary-
loxyalkyl- and arylthioalkyl halides, as well as for unfunctionalized
phenols and thiols (BB2 queries). The latter two were transformed
to their alkylated halide derivatives by means of Pathway A
(Scheme 1), by simple combination of a given building block with
a modified linker using CombiGlide [29] application from Schrö-
dinger. Possible structural variations, such as aromatic ring size
changes, migration of heteroatoms in ring systems and substitution
pattern of any R-groups, were encoded by a symbolic chemical
terms language applied in Instant JChem database manager [30].
Scheme 1. The synthesis of aryloxylethyl bromides (22{1e11}) and arylthioethyl
bromides (22{12e13}). Reagents and conditions: (i) 1,2-dibromoethane, K2CO3,
(CH3)2CO KI, 60 ꢀC, 48e72 h; (ii) 2-chloroethanol, 10% NaOH, 0 ꢀC / 105 ꢀC; (iii) PBr3,
CH2Cl2 anh., 30 min, argon.
The solid-phase synthesis was performed according to a five-
step protocol using a BAL-linker functionalized polystyrene resin
(Scheme 2). In the initial step of the synthesis, Boc-protected
amines (17{1e5}, Fig. 3) were anchored to a solid support by one-
pot reductive amination.
The obtained secondary amines 18{1e5} were subsequently
coupled with acyl 19{1e7} and sulfonyl 19{8e15} chlorides (Fig. 4),
yielding resin-bound amides and sulfonamides of the general
structure 20. Treatment with acyl chlorides was conducted in DMF
and the Hunig base, while with sulfonyl chlorides was carried out in
CH2Cl2 in the presence of TEA [27]. Then, selective BOC removal was
accomplished using Burgess’s methodology involving treatment of
the resin with a mixture of trimethylsilyltrifluoromethane sulph-
onate (TMSOTf) and 2,6-lutidine [28].
The obtained secondary amines 21 were further alkylated with
aryloxylethyl bromides and arylthioethyl bromides (22{1e13}, Fig. 5)
in the presence of 1,8-diazabicyclo [5,4,0]udec-7-ene (DBU) in DMF.
The reaction was conducted at 60 ꢀC for 24 h. Treatment of the resin-
bound compounds of the general structure 23 with a mixture of tri-
fluoroacetic acid/CH2Cl2 (90/10;v/v)yieldedthefinal products 24e95.
3.2. Building blocks database searching
The in-house 5.5 ꢁ 106 building blocks (and intermediates)
member library based on stock-available resources from 26
commercial vendors, was build. It was screened using substructure
searching algorithm and defined automatic queries with tools
applied in Instant JChem. The detailed results are shown in
Table SI-2.
3.3. Building blocks filtering
In order to reduce and refine the obtained sets of building
blocks, three filters were applied. The first limited BBs to those
which simultaneously satisfied simple structural rules, i.e.:
molecular weight ꢂ250 Da, the number of rotatable bonds ꢂ4, and
the number of rings ꢂ2. Then, based on medicinal/organic chem-
istry know-how, libraries of unwanted and reactive functional
R-groups were defined (e.g. amines, alcohols, carboxylic acids,
esters) and were successively used to reduce the BB datasets. The
resultant building block sets were hierarchically grouped using
Molprint2D fingerprint and Tanimoto metric (Canvas) [31], yielding
76 and 259 clusters of BB1 and BB2, respectively. Next, centroids
and a few of the structurally most diverse cluster representatives
(1e5 BBs, proportionally to cluster size) were selected. Additionally,
the database was supplemented with building blocks used in the
synthesis of a pilot series of compounds giving 257 representatives
of BB1 set and 555 agents of BB2 set (Fig. 6).
3. Virtual combinatorial library design and analysis
3.1. Queries definition
Based on the proposed chemical reaction pathway, two sets of
reagent queries were defined (Fig. 6). The first described pattern for
Boc
Boc
O
N
N
i
ii
HN
Ar
N
H
X
n
m
n
m
16
18{1-5}
20{1-5, 1-15}
BAL-PS resin
iii
Y
N
NH
iv
R
R
3.4. Virtual combinatorial library (VCL) generation
Ar
Ar
N
Ar
N
X
X
n
m
n
m
The selected building block databases BB1 and BB2 were itera-
tively combined with each core using a set of algorithms applied in
CombiGlide [29]. In this way, all possible combinations (i.e. about
428 000 virtual compounds) were produced and used as an input
database for the Virtual Screening Protocol.
21{1-5, 1-17}
23{1-5, 1-17, 1-13}
v
Y
N
H
N
X
n
m
X = CO, SO2 m = 0,1
Y = O, S n = 0,1
24 - 95
3.5. Virtual screening (VS) protocol
Scheme 2. The solid-phase synthesis of arylamides and arylsulfonamides of the ary-
loxyethyl- and arylthioethyl- derivatives of 3-aminopyrrolidine, 4-aminopiperidine
and 4-aminomethylpiperidine. Reagents and conditions: (i) amine diversity reagent
17{1e5}, NaBH3CN, 1% AcOH/DMF, 60 ꢀC, 24 h; (ii) ArCOCl 19{1e7}, DIEA, DMF, RT,
2 ꢁ 2 h or ArSO2Cl 19{8e15}, TEA, CH2Cl2, RT, 2 ꢁ 2 h; (iii) TMSOTf, 2,6-lutidine, CH2Cl2,
2 ꢁ 30 min; (iv) alkylating agent 22{1e13}, DBU, KI, DMF, 60 ꢀC, 24 h; (v) TFA/CH2Cl2
(90/10; v/v), 2 h.
Basically, the applied protocol followed the hierarchical scheme
reported previously [32], and consisted of four filtering phases:
physicochemical, ADME, 3D pharmacophore and docking. At first,
the Lipinski rules of 5, the Veber rules and the strongest basic
pKa > 5 were applied (Calculator Plugins, JChem) [33]. Since

P. Zajdel et al. / European Journal of Medicinal Chemistry 56 (2012) 348e360
351
O
O
O
O
F
O
O
O
F
F
N
Cl
Cl
O
Cl
Cl
Cl
2
3
4
1
5
O
N
O
O
O O
S
O
O
O
S
S
S
S
S
Cl
Cl
Cl
Cl
Cl
Cl
7
8
9
6
10
O
O
O
O
O
O
O
O
O O
Cl
S
S
F
S
F
S
F
O
S
Cl
Cl
Cl
Cl
Cl
F
F
O
12
13
14
11
15
Fig. 4. The diversity of building blocks BB1 e acyl chlorides 19{1e7} and sulfonyl chlorides 19{8e15}.
successive stages were based on 3D structures, all possible
stereoisomers were generated using LigPrep [34] from Schrödinger.
The calculated ADME descriptors (QikProp) [35], i.e. the number of
3.7. Ranking scheme
The final results were analyzed using two different ranking
approaches, i.e. BB-based and product-based. In the former, the
ranking of building blocks was generated on the basis of the pres-
ence of a given BB in the final set of VCL. A substructure searching of
final VCL in Instant JChem was performed for each building block
used for its generation. Unique compounds containing queried
structure were counted. Following the same idea, the occurrence of
each building block in vendor’s stocks was estimated. The product-
based ranking list of compounds was made by merging various
scoring functions, such as docking scoring function (Glide Score),
SVM-RBF decision function and rankings of BBs occurrence. For
each of those parameters, separate rankings were created and
merged into one using the MIN rule of data fusion [37,38].
reactive
functional
groups
(0e2),
aqueous
solubility
(ꢃ6.5e0.5 mole/liter), gut-blood barrier (>500 nm/s) and blood-
brain-barrier penetration coefficient (ꢃ3.0e1.2), were used to
filter off compounds with unfavorable profiles. In the next phase,
linear combination of six different pharmacophore models, devel-
oped on previously classified diverse groups of 5-HT7 antagonists
[36], were applied. A compound was accepted if it matched at least
one of the pharmacophore models used. The remaining set of
molecules was docked (Glide at SP accuracy level) to six different
conformations of the previously published rhodopsin-based 5-HT7
homology models [36], with a spatial constrain on ionic interaction
between a ligand and Asp3.32 side-chain.
3.6. Post-docking filter
4. Results and discussion
The developed post-docking filter (see experimental) was used to
reduce the final subset to unique compounds with a correct binding
mode. The subset of unique compounds contained only the highly
classified stereogenic form of the representatives of
3-aminopyrrolidine derivatives. The elaborated classification models
were based on active and inactive 5-HT7R ligands, and involved
structural interaction fingerprints (SIFt) and support vector machine
(SVM) method. Compounds whose docking pose had a negative value
of the decision function were deleted from the final set.
Our initial studies were directed towards structural analogs of
the arylamide derivatives of classic long-chain arylpiperazines in
which an arylpiperazine pharmacophore was replaced with a flex-
ible aryloxy-/arylthio-ethyl amine fragments (Fig. 2). A pilot 34-
member library was evaluated for 5-HT7Rs and those compounds
displayed high to low affinity ranging from 5 to 2674 nM (Table 1).
The primary focus being on the amine core fragments (4-
aminomethylpiperidine, 4-aminopiperidine, and 3-aminopyrro-
lidine) which indicated that a larger distance between the amide/
Cl
S
O
O
Cl
O
O
Br
Br
Br
Br
Br
Br
Br
Br
2
3
4
1
O
Cl
Cl
O
O
O
NH
O
7
8
6
5
Cl
O
O
F
S
O
S
Cl
Br
Br
Br
Br
Br
10
11
12
9
13
Fig. 5. The diversity of building blocks BB2 e N-alkylating agents aryloxylethyl bromides 22{1e11} and arylthioethyl bromides 22{12e13}.

352
P. Zajdel et al. / European Journal of Medicinal Chemistry 56 (2012) 348e360
Fig. 6. A scheme of the applied VCL-VS protocol.
sulfonamide group and the basic nitrogen atom negatively influ-
enced the binding to 5-HT7R.
aryloxy-/arylthio-ethyl fragment negatively influenced the binding
of compounds 45 and 50 to 5-HT7Rs. On the other hand, an iso-
propyl and a phenyl substituent in an ortho position accommodated
better in the receptor pocket; of the ortho-substituted building
blocks, 22{3} containing tert-butyl moiety proved to be less
preferred (e.g. 35 vs 36). Interestingly, this data is consistent with
that reported by Leopoldo et al. who described the enhanced 5-HT7
receptor affinity of compounds containing ortho-phenyl- and ortho-
isopropyl- piperazines [23]. Another interesting finding was that
arylsulfonamide derivatives containing an ortho-methylsulfanyl
substituent displayed high affinity for 5-HT7Rs.
In general, arylamide derivatives from the 24e57 set displayed
low to moderate affinity for the 5-HT7Rs (Ki5-HT7 >100 nM). The only
exceptions were pyrrolidyn-3-yl amides which contained
a biphenyl-oxy fragment (46, 48). Bioisosteric replacement of an
amide bond with sp2 hybridization and planar configuration with
sulfonamide bond with sp3 hybridization and tetrahedral configu-
ration dramatically increased the affinity of compounds for 5-HT7
receptors. All the arylsulfonamide derivatives displayed high affinity
for the 5-HT7Rs (Ki < 100 nM).
It was found that the introduction of substituents in the meta
position of the phenyl ring (e.g. BB2 e 22{6} and 22{8}) in the
In an attempt to better understand the structural requirements
for the investigated flexible analogs of arylpiperazines for the

P. Zajdel et al. / European Journal of Medicinal Chemistry 56 (2012) 348e360
353
Table 1
5-HT1AR, 5-HT6R and D2 dopamine receptors (Table 3). Compounds
58e95 displayed diversified affinity for 5-HT7, their Ki value ranging
from 0.3 nM to 2375 nM, high to low affinity for 5-HT1A (9e903 nM)
and comparable affinity for D2 receptors (10e1059 nM). The
compounds tested displayed the lowest affinity for 5-HT6 sites,
ranging from 91 to 2759 nM. Within the targeted library, some
compounds were found as highly potent 5-HT7R ligands with
acceptable selectivity over other receptors (Table 3).
Structureeactivity relationship studies within the target library
confirmed the results of virtual screening by showing that the
sulfonamide bond with sp3 hybridization was more beneficial than
the amide in the interaction with 5-HT7Rs. It is noteworthy that, the
highly ranked p-F-benzenesulfonamide fragment (19{11}) turned
out to be advantageous to 5-HT7R affinity (e.g. 61, 68, 81, 82, 83, 92).
Furthermore, introduction of building blocks with an additional
fluoro atom (19{12} or 19{13}) either only slightly decreased the
affinity for 5-HT7Rs (56 vs 84 and 86) or did not influence it at all
(83 vs 87). This was also in accordance with the frequency scoring
presented in Table 2.
The binding data of the pilot 34-member library representatives (24e57) for 5-HT7R.
Compda {Core,BB1,BB2} Ki 5-HT7 [nM] Compda {Core,BB1,BB2} Ki 5-HT7 [nM]
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
{1,1,4}
{1,2,3}
{1,2,8}
{1,2,10}
{1,3,5}
{1,5,7}
{1,7,2}
{1,15,2}
{2,1,2}
{2,1,13}
{2,2,12}
{2,3,3}
{2,3,7}
{2,5,5}
{2,5,13}
{2,7,4}
{2,7,7}
1188 ꢄ 132
1500 ꢄ 116
2674 ꢄ 178
409 ꢄ 36
415 ꢄ 22
334 ꢄ 17
407 ꢄ 28
27 ꢄ 2
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
{2,9,4}
{2,9,7}
{2,12,13}
{3,1,2}
{3,1,6}
{3,1,7}
{3,1,8}
{3,2,7}
{3,2,12}
{3,3,8}
{3,3,10}
{3,5,12}
{3,6,1}
{3,9,2}
{3,9,4}
{3,11,7}
{3,12,5}
17 ꢄ 2
12 ꢄ 1
93 ꢄ 6
204 ꢄ 17
3372 ꢄ 410
85 ꢄ 5
1104 ꢄ 75
59 ꢄ 4
311 ꢄ 34
329 ꢄ 21
815 ꢄ 64
1672 ꢄ 145
391 ꢄ 14
873 ꢄ 97
865 ꢄ 66
1144 ꢄ 120
276 ꢄ 19
427 ꢄ 18
709 ꢄ 86
2224 ꢄ 252
1108 ꢄ 87
161 ꢄ 6
7 ꢄ 1
5 ꢄ 0.5
5 ꢄ 0.4
97 ꢄ 11
a
Physicochemical data for compounds with Ki(5-HT7)
>
90 nM is given in
Table SI-1.
Furthermore, it was difficult to determine precisely the most
favorable amine core for the binding to 5-HT7Rs by defining the
optimal distance between the H-bond donor group and the posi-
tively charged nitrogen atom. Nevertheless, the 4-amino-
methylpiperidine core with a low conformational restriction was
less preferential. That finding was in agreement with VS results
where compounds with 4-aminomethylpiperidine moiety (17{1})
amounted to 18% of the final set.
interaction with 5-HT7 receptors, we designed and synthesized
a targeted library of compounds 58e95. To achieve this, we made
use of the biological results obtained for the pilot set (Table 1), and
analysis of the commercially available BB chemical space. The
search of building block libraries of 26 vendors with defined
queries (Fig. 6) resulted in 1526 BB1 and 11066 BB2. Their combi-
nation with 3 cores would give a virtual library of about 51 ꢁ 106
compounds (without stereoisomers), hence, a BB filtering was
applied prior to VCL generation. The remaining blocks were clus-
tered, and only several representatives (including centroids) were
used to obtain the VCL of 428 000 compounds (0.84% of the initial
chemical space). In the successive stage, the previously elaborated
multistep virtual screening protocol [32] (physicochemical filter,
ADME filter, 3D pharmacophore models, docking protocol) allowed
a 5-fold reduction in the number of compounds.
As already demonstrated in the pilot series, the affinity of
compounds 58e95for 5-HT7Rsalsodepended cruciallyuponthekind
of substitution in the aryloxyethyl fragment. Being highly ranked, the
unsubstituted phenoxy building block (22{9}) yielded compounds
with either high (82) or moderate (e.g. 67 and 79) 5-HT7R affinity.
Surprisingly, the most highly scored building block 22{11} (Table 2)
containing a fluoro atom in the meta position turned out to be very
unfavorable for the interaction with 5-HT7R (and other receptors, as
well). That was much below expectations compared to the data pre-
sented by Volk et al. who described the meta-fluorophenylpiperazine
derivative as a 5-HT7R ligand with Ki ¼ 11 nM [21]. On the other hand,
the considerably less widely populated ortho-isopropylphenoxy
fragment (22{2}) was found to behighly beneficial for5-HT7R affinity;
all the arylsulfonamides containing that substituent displayed good
affinity for 5-HT7 sites (Ki < 60 nM), of which 68 showed outstanding
affinity (Ki ¼ 0.3 nM). The importance of the substituent in the ortho
position of the aryloxyethyl fragment was additionally evidenced by
the high affinity of compounds with methylsulfanyl- and phenyl-
substituted building blocks, 22{4} and 22{7}, respectively (Tables 2
and 3). The above analysis showed that computational results were
generally in good agreement with the biological data reinforcing the
relevance of its application in building blocks filtering.
In regards to the data on the affinity of compounds 58e95 for 5-
HT1ARs, it was demonstrated that of the secondary amine cores 17
{1e3}, 3-aminopyrrolidine moiety (17{3}) was preferential for those
sites. Importantly enough, compounds 64, 81, 89 containing a meth-
ylsulfanyl substituent in the ortho position (22{4}) were always clas-
sified as potent 5-HT1A ligands, regardless of other building blocks
used. This finding shows that flexible aryloxyethyl amine fragment,
substituted with electron-donating moiety in the ortho position,
preserves, e.g., ortho-methoxyphenylpiperazine (a well-known 5-
HT1A privileged structure) high affinity for 5-HT1A sites. Irrespective
of piperazine opening, the influence of the substitution pattern,
similar to that observed in the classic LCAP, supports the concept that
aryloxyalkyl piperidines and pyrrolidines may be regarded as arylpi-
perazine biomimetics. This hypothesis may be further confirmed by
compound 77 with a 2,3-dichlorophenoxy fragment (22{1}), which,
like 2,3-diCl-phenylpiperazines showed high affinity for 5-HT1ARs
Since only one constraint, i.e. an ionic interaction of a ligand
with the crucial Asp3.32 side chain, was implied in docking, an
automated procedure for selecting poses matching the common
binding mode was developed and used. This was based on SIFt, the
machine learning method (support vector machine e SVM) and the
tool was trained on
a
set of active 5-HT7R ligands (Ki(5-
M) with
< 300 nM) and inactive compounds (Ki(5-HT7R) > 5
m
HT7R)
diverse structures. In the final set of 38000 compounds, the
frequency of BBs was analyzed (Table 2) and used as a clue for
selecting reagents for the targeted library.
Since the experimental data is of primary importance, a number
of building blocks not preferable to the pilot series were excluded.
From the previously applied set of acyl chlorides, we chose only 19
{5} and additionally, the well-scored benzoyl chlorides 19{4} and 19
{1}. We extended the set of sulfonyl chlorides by applying several
highly (19{8}, 19{13}) and moderately (19{10}, 19{14}) positioned
representatives (Table 2). Regarding to the choice of alkylating
agents (22{1e13}), we excluded a few non-tolerated fragments from
the pilot series (e.g. 22{3}, 22{6}, 22{8}), as well as arylthioethyl
bromides. On the other hand, we introduced two representatives of
the most highly scored aryloxyethyl building blocks (22{9} and 22
{11}). Several of the top-ranked BBs were unaffordable or caused
unpredictedsynthetic problems, as in the case of 8-(2-chloroethoxy)
quinoline, the second-scored representative of the BB2 set (Table 2).
The latter compound was readily cyclized to the corresponding 2,3-
dihydropyrido [1,2,3-de]-1,4-benzoxazinium salt via the intra-
molecular quaternization of a quinoline heteroatom [39].
All the members of the target library were biologically evaluated
for their affinity for 5-HT7Rs and due to the structural homology
and involvement in the pathology of psychiatric disorders for the

354
P. Zajdel et al. / European Journal of Medicinal Chemistry 56 (2012) 348e360
Table 2
The frequency and ranking of building blocks BB1 and BB2 in a set of 38 000 compounds. BB in shades were used for generation of the targeted library.
BB1
BB2
Structure
Rank
1
Freq.
430
BB avail.
4
Structure
Rank
1
Freq.
248
BB avail.
4
O
O
HO
F
19a
19a
19a
22{11}
22b
S
S
N
Cl
N
O
O
O
N
HO
2
4
428
419
2
4
2
5
228
206
7
7
Cl
N
O
O
HO
S
22{9}
Cl
N
HN
O
O
S
HO
S
19{8}
5
7
402
392
11
22{10}
22{5}
75
82
140
134
6
7
Cl
Cl
O
O
HO
S
Cl
19{11}
9
F
Cl
HO
O
O
19a
19a
13
23
366
336
2
3
22{7}
22{1}
89
131
117
7
5
S
S
Cl
F
Cl
O
HO
Cl
N
109
Cl
F
O
S
F
F
HO
19{4}
27
34
315
308
3
8
22{4}
158
161
82
80
2
1
Cl
HS
O
O
O
S
Cl
19{13}
22{13}
F
O
HO
HS
N
F
F
S
22a
166
174
185
194
75
70
64
57
1
3
1
1
Cl
19{12}
19{9}
19a
35
307
280
266
167
9
13
1
N
CN
Cl
O
O
Cl
S
S
43
22{12}
22a
Cl
Cl
HO
O
O
O
H
N
S
S
55
Cl
N
F
N
HO
F
19{3}
102
7
22a
Cl
CF3
O
HO
HO
Cl
19{1}
19{7}
115
130
142
122
1
7
22{8}
217
225
42
41
4
2
O
O
N
22a
S
Cl
N

P. Zajdel et al. / European Journal of Medicinal Chemistry 56 (2012) 348e360
355
Table 2 (continued )
BB1
BB2
Structure
Rank
135
Freq.
116
BB avail.
Structure
Rank
229
Freq.
39
BB avail.
O
O
N
HO
19{5}
19{6}
4
22{2}
22{6}
4
6
Cl
Cl
O
O
O
HO
HO
NH
160
84
7
253
502
32
4
O
19{2}
164
175
81
67
7
8
22{3}
5
Cl
O
O
Cl
S
19{14}
Cl
N
O
O
S
19{10}
19{15}
188
236
44
9
Cl
O
O
O
S
Cl
8
12
O
Freq. e frequency of a given BB in the final set of 38 000 compounds.
BB avail. e BB availability in vendors’ stocks.
a
Building blocks recognized as unaffordable.
Building block with unpredicted chemistry.
b
[40]. On the contrary, the 2,5-dichlorophenoxy fragment (22{5}),
which did not fit the substitution pattern of LCAP for the 5-HT1AR,
displayed low affinity for 5-HT1A sites.
Moreover, an increase of selectivity of the R enantiomer over 5-HT1A
sites was also observed. Of the evaluated enantiomers, compound
91 displayed the highest 5-HT7/5-HT1A selectivity ratio (S1A/7 ¼ 19).
The results obtained for the (R) enantiomers 90 and 92 were in line
with the virtual screening data, more interestingly, they were
consistent with the stereospecificity of SB-269970 for 5-HT7Rs [41].
Since, like aripiprazole and amisulpride, some D2Rs ligands
display high affinity for 5-HT7 sites, we alsotested the targeted library
for D2 receptors. We found that majority of the compounds were
classified as potent D2R ligands. The influence of amine cores and the
kind of arylamide and arylsulfonamide moieties were not very clear.
However, the properties of a substituent at the phenyl ring in the
aryloxyethyl fragment affected differentially the interaction with
D2Rs. Interestingly, a fluoro atom in the meta position was the least
preferential substituent for D2R affinity. It was later found that the
selected (R) enantiomers of pyrrolidyn-3-yl arylsulfonamides
(90e92) displayed a higher 5-HT7/D2 selectivity, thatobservationwas
consistent with the results obtained for 5-HT7/5-HT1A selectivity. On
the other hand, their (S)-counterparts may become templates for
designing multireceptor ligands in this class of derivatives.
As regards to the structural similarity between 5-HT7 and 5-HT1A
receptors, designing selective 5-HT7R ligands is often problematic.
On the basis of the above data we traced structural requirements for
5-HT7/5-HT1A selectivity. Regarding to the arylsulfonamide frag-
ment, introduction of a chlorine atom into position 5 of the thio-
phenesulfonamide core, regardless additional chiral switch from
racemic mixture to respective R/S enantiomers, increased the 5-HT7/
5-HT1ARs selectivity (compound 73 vs compounds 90, 93 and 74 vs
91, 94, respectively). The important element determining 5-HT7/5-
HT1AR selectivity was the kind of a substituent at the phenyl ring
in the aryloxy fragment. As proof, compounds containing an
unsubstituted phenyl ring (82 and 79) displayed low 5-HT7/5-HT1A
selectivity. The same relationship was observed in the case of
compounds containing tetrahydronaphthalenyloxy moiety (75, 83,
87). Although ligands with an ortho-methylsulfanyl substituent
showed high affinity for 5-HT7R (64, 81, 89), their closer examination
revealed that, in contrast to the data concerning arylpiperazines
[23], the modification did not enhance the selectivity over 5-HT1ARs
within the evaluated series. Compounds containing electron-
donating substituents in the ortho position either displayed no 5-
HT7/5-HT1A selectivity (81), or even bound preferentially to 5-HT1A
sites (89, 64). In contrast, library members with isopropyl and phenyl
substituents in the ortho position displayed the highest 5-HT7
preference with outstanding 5-HT7/5-HT1A selectivity ratio for
compound 68 equaling 1453-fold. Thus, like in the case of LCAP [23],
4.1. The functional effect of selected ligands on 5-HT7 receptors
Five selected library representatives were evaluated in a functional
model in CHO cells, which stably expressed the human 5-HT7 receptor.
All the compounds behaved like potent antagonists at h5-HT7R
(Table 4), and methiothepin was used as a reference compound. Inter-
estingly, the most potent compound in the 5-HT7 affinity binding assay
68 displayed the KB value equaling 1 nM, which indicated superior
antagonistic activity compared to the other congeners tested. Further-
more, the potency depended on stereochemistry. Examination of the
antagonistic activity of a pair of enantiomers (91 and 94) provided
additional support for the preference of 5-HT7Rs for R enantiomers.
hydrophobic and
pep interaction properties in the ortho position
accounted better selectivity for the 5-HT7/5-HT1AR.
In the series of pyrrolidyn-3-yl arylsulfonamides we observed
a slight preference for the R enantiomers 90e92 over their S
counterparts 93e95 in respect to their affinity for 5-HT7Rs.

356
P. Zajdel et al. / European Journal of Medicinal Chemistry 56 (2012) 348e360
Table 3
The binding data of the second set of library members (58e95) for 5-HT7Rs and 5-HT1A, 5-HT6, and D2 receptors.
b
Compd
{Core,BB1,BB2}
Ki ꢄ SEM [nM]
S1A/7
5-HT7
5-HT1A
5-HT6
D2
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
{1,9,7}
20 ꢄ 2
2222 ꢄ 180
60 ꢄ 5
101 ꢄ 9
841 ꢄ 58
141 ꢄ 15
290 ꢄ 22
903 ꢄ 48
104 ꢄ 7
18 ꢄ 2
410 ꢄ 26
2759 ꢄ 180
313 ꢄ 26
315 ꢄ 33
2124 ꢄ 194
177 ꢄ 10
147 ꢄ 8
32 ꢄ 3
5
<1
2
22
<1
2
<1
<1
<1
2
1453
45
2
<1
1
<1
5
<1
<1
2
{1,8,11}
{1,11,2}
{1,11,7}
{1,11,11}
{1,14,2}
{1,14,4}
{1,14,11}
{2,1,11}
{2,8,9}
{2,11,2}
{2,12,7}
{3,4,7}
{3,5,2}
{3,5,7}
1059 ꢄ 99
50 ꢄ 3
46 ꢄ 4
839 ꢄ 75
66 ꢄ 5
50 ꢄ 1
560 ꢄ 62
76 ꢄ 9
84 ꢄ 7
51 ꢄ 5
63 ꢄ 4
15 ꢄ 2
79 ꢄ 7
24 ꢄ 1
16 ꢄ 1
16 ꢄ 2
72 ꢄ 5
729 ꢄ 41
67 ꢄ 8
78 ꢄ 6
318 ꢄ 24
10 ꢄ 1
41 ꢄ 3
11 ꢄ 1
72 ꢄ 5
24 ꢄ 3
523 ꢄ 19
18 ꢄ 2
60 ꢄ 4
25 ꢄ 3
36 ꢄ 2
20 ꢄ 1
39 ꢄ 1
16 ꢄ 2
25 ꢄ 1
8 ꢄ 1
13 ꢄ 1
1362 ꢄ 118
46 ꢄ 5
78 ꢄ 6
2195 ꢄ 257
2375 ꢄ 194
155 ꢄ 11
0.3 ꢄ 0.1
7 ꢄ 1
283 ꢄ 33
456 ꢄ 35
237 ꢄ 7
436 ꢄ 21
313 ꢄ 28
48 ꢄ 4
831 ꢄ 39
3785 ꢄ 415
862 ꢄ 62
240 ꢄ 29
199 ꢄ 17
1799 ꢄ 128
1761 ꢄ 186
1335 ꢄ 94
258 ꢄ 20
267 ꢄ 17
102 ꢄ 8
21 ꢄ 3
178 ꢄ 14
41 ꢄ 3
30 ꢄ 3
45 ꢄ 6
{3,8,2}
{3,8,7}
{3,8,10}
{3,8,11}
{3,9,1}
29 ꢄ 4
21 ꢄ 2
10 ꢄ 1
45 ꢄ 3
27 ꢄ 2
13 ꢄ 1
1438 ꢄ 79
14 ꢄ 2
270 ꢄ 29
28 ꢄ 4
1779 ꢄ 140
91 ꢄ 8
{3,9,5}
{3,9,9}
58 ꢄ 5
717 ꢄ 58
65 ꢄ 7
225 ꢄ 31
957 ꢄ 74
504 ꢄ 44
357 ꢄ 26
248 ꢄ 21
96 ꢄ 8
12
<1
2
1
1
121 ꢄ 9
10 ꢄ 1
{3,10,2}
{3,11,4}
{3,11,9}
{3,11,10}
{3,12,7}
{3,12,11}
{3,13,7}
{3,13,10}
{3,14,2}
{3,14,4}
{4,9,2}
{4,9,7}
{4,11,7}
{5,9,2}
{5,9,7}
{5,11,7}
23 ꢄ 1
11 ꢄ 1
10 ꢄ 1
9 ꢄ 1
9 ꢄ 1
16 ꢄ 2
22 ꢄ 1
1
7 ꢄ 0.5
293 ꢄ 17
9 ꢄ 1
92 ꢄ 5
327 ꢄ 23
951 ꢄ 88
515 ꢄ 54
156 ꢄ 17
245 ꢄ 15
185 ꢄ 7
13
<1
9
<1
8
<1
8
19
11
3
230 ꢄ 9
81 ꢄ 4
16 ꢄ 1
10 ꢄ 1
4 ꢄ 0,6
15 ꢄ 1
31 ꢄ 2
5 ꢄ 0.6
49 ꢄ 6
90
91
92
93
94
95
6 ꢄ 0.3
4 ꢄ 0.2
2 ꢄ 0.2
28 ꢄ 2
200 ꢄ 17
202 ꢄ 23
102 ꢄ 4
75 ꢄ 6
22 ꢄ 3
89 ꢄ 7
537 ꢄ 49
551 ꢄ 34
271 ꢄ 17
4 ꢄ 0.3
12 ꢄ 1
147 ꢄ 12
32 ꢄ 2
12
3
10 ꢄ 1.1
18 ꢄ 2
10 ꢄ 1
72 ꢄ 5
7 ꢄ 1
Clozapinea
Olanzapinea
143 ꢄ 11
3442 ꢄ 408
e
185 ꢄ 21
5 ꢄ 0.7
e
a
Standard drugs added as a reference.
Ratio of affinity for 5-HT1A vs 5-HT7 receptors.
b
5. Conclusions
ligands. A few compounds of the series, i.e. 61, 69, 91, displayed high
affinity for 5-HT7Rs and moderate selectivity over 5-HT1A, 5-HT6,
and D2 sites. The arylsulfonamide derivative 68 showed high affinity
for 5-HT7Rs (Ki ¼ 0.3 nM) with high selectivity over 5-HT1AR and
D2R, and was classified as a potent antagonist (Kb ¼ 1 nM). Structure-
activity relationship studies demonstrated that ortho-phenyl or
ortho-isopropyl substitutes at the aryloxyethyl fragment were
preferred for an interaction with 5-HT7Rs. Furthermore, the
sulfonamide bond turned out to be favorable for those sites. It seems
that the replacement of arylpiperazine pharmacophore with flexible
aryloxyethyl piperidine and pyrrolidine fragments opens up the
possibility of exploring a new chemical space and designing new
selective 5-HT7R ligands or multireceptor ligands targeted on
monoamine receptors. Further investigation of this series of deriv-
atives may result in discovering of novel psychotropic agents.
Aiming to develop of novel 5-HT7 receptor antagonists, a 34-
member pilot library of the arylamide and arylsulfonamide deriva-
tives of differently substituted aryloxy- and arylthio-ethyl piperi-
dines and pyrrolidines was synthesized using
a solid-phase
methodology. Successively, a computational approach involving
a combinatorial library design and a multistep virtual screening,
followed bypost-docking filtering andbuilding blockranking within
compounds satisfying the desired 5-HT7R binding pattern allowed
us to identify critical molecular substructures and provided ratio-
nale data for designing the targeted 38-member library of 5-HT7R
Table 4
The receptor binding data and antagonistic activities for 5-HT7 receptors of some
selected library members.
Compd
{Core,BB1,BB2}
5-HT7
6. Experimental
Ki [nM]
Kb [nM]
61
68
69
91
{1,11,7}
{2,11,2}
{2,12,7}
{4,9,7}
13
0.3
7
5.1
1
4.6
7.4
21
6.1. Chemistry
4
Solution and solid-phase organic transformations and resin
washes were carried out at ambient temperature, unless indicated
otherwise. Organic solvents (from Aldrich and Chempur) were of
reagent grade and were used without purification. BAL linker MBHA-
94
{5,9,7}
12
18
e
Clozapine
Methiothepin
e
0.075

P. Zajdel et al. / European Journal of Medicinal Chemistry 56 (2012) 348e360
357
resin (loading 1.1 mmol/g) was purchased from Iris Chemicals. The
reagents were from Aldrich, Alfa Aesar, Chembridge, Fluorochem.
Analytical HPLC were run on a Waters Alliance HPLC instrument,
equipped with a ChromolithSpeedROD column (4.6 ꢁ 50 mm).
Standard conditions were eluent system A (water/0.1% TFA), system
B (acetonitrile/0.1% TFA). A flow rate of 5 mL/min and a gradient of
(0e100)% B over 3 min were used. Detection was performed on
a PDA detector.
(3 ꢁ 3 mL), DMF (3 ꢁ 3 mL), MeOH (5 ꢁ 3 mL), and CH2Cl2
(3 ꢁ 3 mL), and dried under low vacuum. The efficacy of the
reaction was confirmed by a positive chloranil test.
6.2.5. Alkylation with aryloxyalkylbromides and
arylthioalkylbromides
The deprotected resin 21 was swelled in CH2Cl2 for 30 min, and to
the resin was added a solution of 0.8 ml of 0.65 M solution of alky-
lating agent 22 ([Diversity reagent 22, 22{1-13}], 0.52 mmol, 6.5 eq,
Fig. 5) in DMF, followed byaddition of DBU (1.04 mmol,13 eq) and KI
(0.08 mmol, 1 eq). The reactors were placed in 60 ꢀC for 24 h. The
resulting resin 23 was drained, washed with DMF (3 ꢁ 1 mL), MeOH
(1 ꢁ1 mL), CH2Cl2 (3 ꢁ 1 mL) and dried under vacuum.
1H NMR and 13C NMR spectra were obtained in a Varian BB 200
spectrometer using TMS (0.00 ppm) in chloroform-d1, and were
recorded at 300 and 75 MHz, respectively; J values are in hertz (Hz),
and splitting patterns are designated as follows: s (singlet),
d (doublet), t (triplet), m (multiplet).
Melting points (mp) were determined with a Büchi apparatus
and are uncorrected.
6.2.6. Cleavage of the final products
Elemental analysis for C, H, and N were carried out by a micro
method using the elemental Vario EI III elemental analyzer. The
results for elemental analyses were found within ꢄ0.4% of the
theoretical values.
LC/MS were carried out on a system consisted of a Waters
Acquity UPLC, coupled to a Waters TQD mass spectrometer. All the
A 1.5 mL of a mixture of TFA/CH2Cl2 (9/1, v/v) was dispensed into
glass vials containing the resin. The cleavage was carried out for
120 min, then the mixture was filtrated and the resin was washed
with a mixture of TFA/CH2Cl2 (8/2; v/v), and the collected filtrates
were evaporated with a stream of argon on Eva parallel evaporator.
Average overall yields of the crude products were between 38
and 57% and were calculated on the basis of the initial loading of
the resin. The LC/MS of the identified compounds 24e95 revealed
an average purity exceeding 82%. The crude residue was re-
dissolved in CH2Cl2 or CH2Cl2/MeOH mixture, and purified using
silica gel columns and mixture CH2Cl2/MeOH, to elute the pure
products 24e95.
analyses were carried out using
a Acquity UPLC BEH C18,
50 ꢁ 2.1 mm column, at 40 ꢀC. A flow rate of 0.3 mL/min and
a gradient of (5e95)% B over 5 min was used. Eluent A: water/0.1%
HCO2H; eluent B: acetonitrile/0.1% HCO2H. Retention times (Rt) are
given in minutes. The UPLC/MS purity of all the test compounds and
key intermediates was determined to be >98%.
Abbreviations used: DIEA, N,N-diisopropylethylamine; EtOAc,
ethyl acetate; TEA, triethylamine.
6.3. Spectroscopic data for selected library members
6.2. Solid-phase synthesis on BAL-MBHA-PS-resin
The compound characterization data for the key final
compounds from Table 3 are summarized below. The synthesis and
characterization data for the remaining final compounds are in the
Supporting Information.
6.2.1. Preparation of amine-bound resin 16 via reductive amination
The dried MBHA-BAL resin was divided into five reactors con-
taining a suspension sodium cyanoborohydride ([NaBH3CN] ¼ 5.5 m-
mol,
5
equiv) and the amine ([Diversity reagent 17, 17
6.3.1. 4-Fluoro-N-({1-[2-(biphenyl-2-yloxy)ethyl]piperidin-4-yl}
methyl) benzenesulfonamide (61)
{1e5}] ¼ 5.5 mmol, 5 equiv, Fig. 3), in a 1% acetic acid in 3 mL of
DMF. The reactors were placed in the oven for 24 h at 60 ꢀC. Then the
resin was drained; washed with 10% AcOH in DMF (1 ꢁ 5 mL) then
with DMF (3 ꢁ 5 mL), MeOH (1 ꢁ 5 mL), and CH2Cl2 (3 ꢁ 5 mL), and
dried under low vacuum.
Yellow oil, 38 mg (54% yield) following chromatographic puri-
fication over silica gel with CH2Cl2/MeOH (9/0.5); initial LC/MS
purity 88%, tR ¼ 3.05. MW 468.58. Monoisotopic Mass 468.2,
[M þ H]þ 469.5. Anal. Calcd for C26H29FN2O3S: C, 65.43; H, 5.72; N,
6.36; found: C, 63.18; H, 5.48; N, 6.32%. 1H NMR (300 MHz, CDCl3)
6.2.2. Preparation of support-bound carboxamide derivatives
1 mL of a solution containing acyl chloride ([Diversity reagent
19, 19{1e7}] ¼ 0.40 mmol, 5 equiv, Fig. 4) in DMF was added to the
resin followed by addition of DIEA (0.88 mmol, 11 equiv). The
reactors were shaken for 2 h at room temperature. The resin was
drained and washed with DMF (3 ꢁ 5 mL), MeOH (1 ꢁ 5 mL), and
CH2Cl2 (3 ꢁ 5 mL), and dried under low vacuum. The procedure
described above was repeated.
d 1.08e1.20 (m, 2H), 1.34e1.42 (m, 1H),1.56e1.60 (m, 2H),1.92e1.99
(d, 2H), 2.68e2.71 (t, 2H), 2.79e2.88 (m, 4H), 4.04e4.08 (t, 2H), 4.47
(b s, 1H), 6.94e6.97 (d, 1H), 7.00e7.05 (td, 1H), 7.15e7.23 (m, 2H),
7.29e7.40 (m, 5H), 7.49e7.54 (m, 2H), 7.83e7.90 (m, 2H). 13C NMR
(75 MHz, CDCl3)
d 26.62, 33.64, 46.76, 53.03, 55.87, 63.18, 112.52,
115.95, 116.44, 122.01, 126.97, 127.79, 128.71, 129.19, 129.24, 129.31,
130.59, 136.00, 136.04, 138.18, 154.10, 163.07, 166.44.
6.3.2. 3-Chloro-N-({1-[2-(2-methylsulfanylphenoxy)ethyl]
piperidin-4-yl}methyl) benzenesulfonamide (64)
6.2.3. Preparation of support-bound sulfonamide derivatives
1 mL of a solution containing sulfonyl chloride ([Diversity
reagent 19, 19{8e15}] ¼ 0.40 mmol, 5 equiv, Fig. 4) in CH2Cl2 was
added to the resin followed by addition of TEA (0.88 mmol, 11 eq).
The reactors were shaken for 2 h at room temperature. The resin
was drained and washed with DMF (3 ꢁ 5 mL), MeOH (1 ꢁ 5 mL),
and CH2Cl2 (3 ꢁ 5 mL), and dried under low vacuum. The procedure
described above was repeated.
Yellow oil, 39 mg (57% yield) following chromatographic purifica-
tion over silica gel with CH2Cl2/MeOH (9/0.5); initial LC/MS purity 90%,
tR ¼ 2.75. C21H27ClN2O3S2, MW 455.03, Monoisotopic Mass 454.1,
[M þ H]þ 455.4. 1H NMR (300 MHz, CDCl3)
d 1.16e1.28 (m, 2H),
1.43e1.48 (m, 1H), 1.65e1.69 (m, 2H), 2.08e2.17 (td, 2H), 2.41 (s, 3H),
2.83e2.87 (t, 4H), 3.00e3.04 (m, 2H), 4.12e4.16 (t, 2H), 4.54e4.56 (m,
1H), 6.79e6.82 (dd, 1H), 6.92e6.98 (td, 1H), 7.08e7.14 (m, 2H),
7.43e7.48(t,1H),7.53e7.57 (d, 1H), 7.71e7.75 (m, 1H), 7.84e7.85(t,1H).
6.2.4. Boc-deprotection protocol
The resin was treated with 1 mL of a freshly prepared solution of
TMSOTf (1.5 M) and 2,6-lutidine (1 M) in dry CH2Cl2. The resin was
shaken for 30 min and was washed with CH2Cl2 (3 ꢁ 1 mL). The
reaction was repeated and the resin was washed with CH2Cl2
6.3.3. 4-Fluoro-N-(1-{2-[(propan-2-yl)phenoxy]ethyl}piperidin-4-
yl) benzenesulfonamide (68)
Yellow oil, 28 mg (44% yield) following chromatographic purifi-
cation over silica gel with CH2Cl2/MeOH (9/0.7); initial LC/MS purity

358
P. Zajdel et al. / European Journal of Medicinal Chemistry 56 (2012) 348e360
87%, tR ¼2.82. MW420.54. Monoisotopic Mass 420.2, [Mþ H]þ 421.1.
Anal. Calcd for C22H23ClN2O3S2: C, 57.07; H, 5.01; N, 6.05; found: C,
57.16; H, 4.99; N, 6.06%.1H NMR (300 MHz, CDCl3)
(ppm) 1.49e1.60
Anal. Calcd for C22H29FN2O3S: C, 62.83; H, 6.95; N, 6.66; found: C,
d
62.71; H, 6.92; N, 6.61%. 1H NMR (300 MHz, CDCl3)
d
1.16e1.21 (m,
(m, 1H), 2.05e2.21 (m, 2H), 2.35e2.40 (dd, 1H), 2.46e2.50 (dd, 1H),
2.73e2.78 (m, 3H), 3.78 (b s, 1H), 3.99e4.05 (t, 2H), 5.05 (b s, 1H),
6.86e6.87 (d, 1H), 6.94e6.96 (d, 1H), 7.02e7.07 (td, 1H), 7.28e7.41
6H), 1.45e1.55 (m, 2H), 1.75e1.81 (m, 2H), 2.17e2.25 (m, 2H),
2.76e2.80 (t, 2H), 2.83e2.88 (m, 2H), 3.16e3.29 (m, 2H), 4.03e4.06
(t, 2H), 4.63e4.66 (d, 1H), 6.77e6.80 (dd, 1H), 6.89e6.94 (td, 1H),
7.09e7.21 (m, 4H), 7.88e7.92 (m, 2H). 13C NMR (75 MHz, CDCl3)
(m, 6H), 7.47e7.51 (m, 2H). 13C NMR (75 MHz, CDCl3)
d 30.65,
51.10, 53.47, 53.82, 58.99, 63.92,112.71,122.32,127.08,127.23,128.08,
129.00, 129.55, 130.89, 131.37, 137.25, 138.32, 139.52, 154.19.
d
22.59, 26.34, 29.61, 52.50, 56.06. 62.57,111.27,116.10,116.40,121.96,
126.17, 126.75, 129.30, 129.71, 136.48, 154.15, 163.17, 166.55.
6.3.9. 5-Chloro-(S)-N-(1-{2-[(propan-2-yl)-phenyloxy]ethyl}
pyrrolidin-3-yl)thiophene-2-sulfonamide (93)
6.3.4. 3,4-Difluoro-N-{1-[2-(biphenyl-2-yloxy)ethyl]piperidin-4-yl}
benzenesulfonamide (69)
Yellow oil, 27 mg (42% yield) following chromatographic
purification over silica gel with CH2Cl2/MeOH (9/0.5); initial LC/
MS purity 69%, tR ¼ 2.93. C19H25ClN2O3S2, MW 428.99 Mono-
isotopic Mass 428.1, [M þ H]þ 429.0. 1H NMR (300 MHz, CDCl3)
Yellow oil, 32 mg (45% yield) following chromatographic puri-
fication over silica gel with CH2Cl2/MeOH (9/0.7); initial LC/MS
purity 82%, tR ¼ 2.84. MW 472.54, Monoisotopic Mass 472.2,
[M þ H]þ 473.1. Anal. Calcd for C25H26F2N2O3S: C, 63.54; H, 5.55; N,
5.93; found: C, 63.58; H, 5.51; N, 5.90%. 1H NMR (300 MHz, CDCl3)
d
1.19e1.21 (dd, 6H), 1.59e1.71 (m, 1H), 2.12e2.25 (m, 1H),
2.31e2.40 (m, 1H), 2.54e2.60 (dd, 1H), 2.68e2.72 (dd, 1H),
2.86e2.90 (t, 2H), 2.92e3.00 (m, 1H), 3.24e3.33 (m, 1H),
3.91e3.96 (m, 1H), 4.02e4.06 (t, 2H), 6.78e6.81 (dd, 1H),
6.86e6.87 (d, 1H), 6.91e6.96 (td, 1H), 7.11e7.17 (td, 1H), 7.20e7.23
(dd, 1H), 7.36e7.37 (d, 1H).
d
1.33e1.46 (m, 2H), 1.66e1.71 (m, 2H), 2.01e2.09 (td, 2H),
2.65e2.72 (m, 4H), 3.10e3.12 (m, 1H), 4.00e4.04 (t, 2H), 4.57e4.59
(d, 1H), 6.93e6.95 (m, 1H), 7.00e7.05 (td, 1H), 7.26e7.38 (m, 6H),
7.47e7.51 (m, 2H), 7.63e7.69 (m, 1H), 7.71e7.75 (m, 1H). 13C NMR
(75 MHz, CDCl3)
d 33.05, 50.84, 52.22, 56.77, 66.95, 112.71, 116.72,
116.97, 118.06, 118.30, 121.18, 123.91, 126.78, 127.75, 128.57, 129.60,
130.86, 131.18, 138.52, 148.32, 148.50, 151.15, 151.32, 151.87, 155.65.
6.3.10. 5-Chloro-(S)-N-{1-[2-(biphenyl-2-yloxy)ethyl]pyrrolidin-3-
yl}thiophene-2-sulfonamide (94)
Yellow oil, 38 mg (55% yield) following chromatographic puri-
fication over silica gel with CH2Cl2/MeOH (9/0.5); initial LC/MS
purity 82%, tR ¼ 2.84. MW 462.02, Monoisotopic Mass 462.1
[M þ H]þ 463.3. Anal. Calcd for C22H23ClN2O3S2: C, 57.07; H, 5.01; N,
6.05; found: C, 57.12; H, 5.02; N, 6.06%. 1H NMR (300 MHz, CDCl3)
6.3.5. 4-Fluoro-N-(1-{2-[2-(methylsulfanyl)phenoxy]ethyl}
pyrrolidin-3-yl)benzenesulfonamide (81)
Yellow oil, 26 mg (42% yield) following chromatographic puri-
fication over silica gel with CH2Cl2/MeOH (9/0.5); initial LC/MS
purity 85%, tR ¼ 2.38. C19H23FN2O3S2, MW 410.52, Monoisotopic
d
(ppm) 1.50e1.61 (m, 1H), 2.02e2.23 (m, 2H), 2.36e2.41 (dd, 1H),
Mass 410.1, [M þ H]þ 411.0. 1H NMR (300 MHz, CDCl3)
d
1.62e1.64
2.47e2.51 (dd, 1H), 2.73e2.77 (m, 3H), 3.79e3.82 (m, 1H),
4.00e4.04 (t, 2H), 5.05 (b s, 1H), 6.86e6.87 (d, 1H), 6.94e6.96 (d,
1H), 7.02e7.07 (td, 1H), 7.26e7.30 (m, 2H), 7.34e7.41 (m, 4H)
(m, 1H), 2.06e2.14 (m, 1H), 2.27e2.34 (m, 1H), 2.40e2.42 (m, 1H),
2.43 (s, 3H), 2.57e2.60 (m, 1H), 2.77e3.03 (m, 3H), 3.87 (b s, 1H),
4.03e4.07 (m, 2H), 5.38 (b s, 1H), 6.77e6.79 (d, 1H), 6.96e7.02 (m,
1H), 7.04e7.14 (m, 4H), 7.79e7.84 (m, 2H).
7.47e7.51 (m, 2H). 13C NMR (75 MHz, CDCl3)
d 30.85, 51.28, 53.45,
53.71, 59.57, 112.69, 122.39, 127.13, 127.28, 128.09, 128.17, 129.05,
129.59, 129.67, 130.92, 131.40, 131.62, 137.19, 138.35, 139.58, 154.15.
6.3.6. 3-Chloro-N-(1-{2-[(propan-2-yl)-phenyloxy]ethyl}
pyrrolidin-3-yl)benzenesulfonamide (88)
6.4. Molecular modeling
Yellow oil, 25 mg (39% yield) following chromatographic purifi-
cation over silica gel with CH2Cl2/MeOH (9/0.7); initial LC/MS purity
80%, tR ¼ 2.93. C21H27ClN2O3S, MW 422.96 Monoisotopic Mass 422.1,
6.4.1. Software
Instant JChem version 5.3.4 was used for structure database
management, search and prediction of simple molecular proper-
ties. The command-line tools of JChem, i.e. Calculator Plugins (used
for calculation of simple molecular descriptors) and Chemical Term
Evaluator [42] (used for analyzing sdf and removing non-matching
molecules) were combined to filter out building blocks and
compounds (VCL) in batch mode. A set of Schrodinger’s applica-
tions were used for generating combinatorial library (CombiGlide),
preparing high quality 3D ligand structures taking into consider-
ation protonation states and stereoisomers (LigPrep), calculating
simple ADME descriptors (QikProp) and automated ligand-receptor
docking (Glide). The Catalyst module from Discovery Studio 2.5 [43]
was used to 3D pharmacophore models generation and screening.
[M þ H]þ 423.1. 1H NMR (300 MHz, CDCl3)
d 1.15e1.22 (m, 7H),
1.55e1.61 (m, 1H), 2.09e2.28 (m, 1H), 2.48e2.53 (q, 1H), 2.60e2.64
(dd, 2H), 2.83e2.87 (m, 1H), 2.90e2.97 (m, 1H), 3.21e3.31 (m, 1H),
3.89 (b s, 1H), 4.00e4.03 (t, 2H), 5.06 (b s, 1H), 6.77e6.80 (dd, 1H),
6.91e6.96 (td, 1H), 7.11e7.17 (td, 1H), 7.20e7.23 (dd, 1H), 7.37e7.42
(t, 1H), 7.47e7.52 (ddd, 1H), 7.71e7.74 (ddd, 1H), 7.85e7.86 (t, 1H).
6.3.7. 5-Chloro-(R)-N-(1-{2-[(propan-2-yl)-phenyloxy]ethyl}
pyrrolidin-3-yl)thiophene-2-sulfonamide (90)
Yellow oil, 32 mg (49% yield) following chromatographic puri-
fication over silica gel with CH2Cl2/MeOH (9/0.5); initial LC/MS
purity 72%, tR ¼ 2.19. C19H25ClN2O3S2, Monoisotopic Mass 428.1,
[M þ H]þ 429.0. 1H NMR (300 MHz, CDCl3)
d 1.19e1.21 (dd, 6H),
1.64e1.74 (m, 1H), 2.14e2.26 (m, 1H), 2.35e2.44 (m, 1H), 2.58e2.63
(dd, 1H), 2.73e2.77 (dd, 1H), 2.89e2.93 (t, 2H), 2.97e3.05 (m, 1H),
3.27e3.33 (m, 1H), 3.92e3.99 (m, 1H), 4.04e4.08 (t, 2H), 6.79e6.82
(dd, 1H), 6.86e6.87 (d, 1H), 6.91e6.97 (td, 1H), 7.12e7.17 (td, 1H),
7.20e7.23 (dd, 1H), 7.36e7.38 (d, 1H).
6.4.2. Post-docking filter
To generate a post-docking filter, a set of 188 well-known diverse
active molecules and of 258 diverse inactive ones of 5-HT7R ligands
and decoys (1589 molecules) were docked (Glide SP mode) to six
homology models. The active and inactive subsets were prepared on
the basis of the data extracted from version 09 of the ChEMBL
database [44] while a decoy set was prepared following the DUD
methodology [45]. The molecules with the inhibition Ki value below
300 nM were regarded as active, while in case of inactive molecules,
the Ki threshold was higher than 5000 nM. Then, for the final poses,
the structural interaction fingerprints (SIFt) [46,47] were generated
6.3.8. 5-Chloro-(R)-N-{1-[2-(biphenyl-2-yloxy)ethyl]pyrrolidin-3-
yl}thiophene-2-sulfonamide (91)
Yellow oil, 39 mg (56% yield) following chromatographic purifi-
cation over silica gel with CH2Cl2/MeOH (9/0.5); initial LC/MS purity
82%, tR ¼ 2.83. MW 462.02, Monoisotopic Mass 462.1 [M þ H]þ463.3.

P. Zajdel et al. / European Journal of Medicinal Chemistry 56 (2012) 348e360
359
using in-house scripts. The results were stored in a 1D binary string,
wherenine bit patternwas used todescribe the interaction type: any
contact, backbone, side chain, polar, aromatic, hydrophobic inter-
action, hydrogen bond donor/acceptor and charged. On the basis on
those ligand-receptor interaction description formats, training and
testing sets were created by splitting the overall set in a ratio of 1:3,
respectively. Vectors describing the interaction profile of known
actives were labeled “þ1”, whereas those referring to known inac-
tives and decoys were labeled “ꢃ1”. Such representations were used
tocreateinput filestotheSVMlight [48]software, separatelyforeach
receptor conformation. SVM classification models were built using
training sets and a radial base function (RBF), also called the
Gaussian function. This kernel function requires two parameters: C,
2 nM [3H]-LSD (85.2 Ci/mmol for 5-HT6R; 0.6 nM [3H]-5-CT
(39.2 Ci/mmol) for 5-HT7R or [3H]-Raclopride (74.4 Ci/mmol).
Non-specificbindingisdefinedwith10
5-HT7R binding experiments, whereas 10
m
M of5-HT in 5-HT1ARand
mM methiothepine or 1 mM
of (þ)butaclamol were used in 5-HT6R and D2L assays, respectively.
Each compound was tested in triplicate at 7e8 concentrations
(10ꢃ11e10ꢃ4 M). The inhibition constants (Ki) were calculated from
the Cheng-Prushoff equation [50]. Results were expressed as means
of at least two separate experiments.
Membrane preparation and general assay procedures for cloned
receptors were adjusted to 96-microwell format based on protocols
described by us previously [22,51,52].
i.e. a penalty parameter of the error term, and
g, i.e. a gamma
6.5.3. Effects on adenylate cyclase activity
parameter for the radial base function. The easy.py script of LibSVM
[49] was used to automate the search of the above mentioned
parameters by applying a systematic grid search algorithm. Test sets
were used to evaluate the quality of models using different perfor-
mance measures, i.e. recall, precision, enrichment factor and the
Matthews correlation coefficient (MCC). For a given receptor, the
best models were identified and used to rank VCL docking poses. The
obtained values of the decision function were used to select only
those molecules whose interaction pattern was very close to the
known actives and distant from that of the known inactives and
decoys. The results obtained for all the receptors were merged, and
the modified MAX rule of data fusion was applied to further reduce
the obtained matrix. Only those compounds were accepted which
had at least one positive value of the decision function; in the case of
at least two positive values, a pose with the higher value was chosen.
The functional activity of the five selected compounds 61, 68, 69,
91, 94 on intracellular cAMP levels, studied in CHO cells which
stably expressed the human 5-HT7 receptor, was determined at
CEREP (Le Bois l’Eveque, 86600 Celle L’Evescault, France) according
to the previously published methods [53]. cAMP concentration was
measured using HTRF method. Adenylate cyclase activity is
expressed as the percentage of the maximal effect obtained with
300 nM serotonin. The compounds were tested in 5 concentrations
at 10ꢃ4e10ꢃ9 in the h5-HT7 antagonist effect. For the antagonists,
the apparent dissociation constants (KB) were calculated using the
modified Cheng Prusoff equation (KB ¼ IC50/(1 þ (A/EC50A)), where
A
¼
concentration of reference agonist in the assay, and
EC50A ¼ EC50 value of the reference agonist).
6.5. In vitro pharmacology
Acknowledgments
6.5.1. Cell culture and preparation of cell membranes
The authors wish to thank students G. Glanowski and M. Szy-
miec for their technical assistance. This study was partly supported
by the Polish Ministry of Science and Higher Education (MNiSW),
Grant No. N N405 671540 and Funds for Statutory Activity of
Jagiellonian University Medical College. Radioligand binding
experiments were financially supported by the Norwegian Finan-
cial Mechanism as part of the PolisheNorwegian Research Fund,
Grant No. PNRF-103-AI-1/07.
HEK293 cells with stable expression of human serotonin 5-
HT1AR, 5-HT6, 5-HT7bR or dopamine D2LR (prepared with the use
of Lipofectamine 2000) were maintained at 37 ꢀC in a humidified
atmosphere with 5% CO2 and were grown in Dulbeco’s Modifier
Eagle Medium containing 10% dialysed foetal bovine serum and
500 mg/ml G418 sulphate. For membranes preparations, cells were
subcultured in 10 cm diameter dishes, grown to 90% confluence,
washed twice with prewarmed to 37 ꢀC phosphate buffered saline
(PBS) and were pelleted by centrifugation (200 g) in PBS containing
0.1 mM EDTA and 1 mM dithiothreitol. Prior to membrane prepa-
rations pellets were stored at ꢃ80 ꢀC.
Appendix A. Supplementary material
Supplementary data available: experimental procedures, and
spectral data for non-commercial building blocks, spectral data for
target compounds, BB available resources database, and post-
docking procedure description, associated with this article is
available in on-line version at http://dx.doi.org/10.1016/j.ejmech.
2012.07.043.
6.5.2. Radioligand binding assays
Cell pellets were thawed and homogenized in 20 volumes of
assay buffer using an Ultra Turrax tissue homogenizer and centri-
fuged twice at 35 000 g for 20 min at 4 ꢀC, with incubation for
15 min at 37 ꢀC in between. The composition of the assay buffers
was as follows: for 5-HT1AR: 50 mM TriseHCl, 0.1 mM EDTA, 4 mM
References
MgCl2, 10
TriseHCl, 0.5 mM EDTA and 4 mM MgCl2, for 5-HT7bR: 50 mM
TriseHCl, 4 mM MgCl2, 10 M pargyline and 0.1% ascorbate; for
dopamine D2LR: 50 mM TriseHCl, 1 mM EDTA, 4 mM MgCl2,
120 mM NaCl, 5 mM KCl, 1.5 mM CaCl2 and 0.1% ascorbate.
mM pargyline and, 0.1% ascorbate; for 5-HT6R: 50 mM
[1] E. Agai-Csongor, K. Nógrádi, J. Galambos, I. Vágó, A. Bielik, I. Magdó, G. Ignácz-
Szendrei, G.M. Keseru, I. Greiner, I. Laszlovszky, E. Schmidt, B. Kiss, K. Sághy,
J. Laszy, I. Gyertyán, M. Zájer-Balázs, L. Gémesi, G. Domány, Novel sulfon-
amides having dual dopamine D2 and D3 receptor affinity show in vivo
antipsychotic efficacy with beneficial cognitive and EPS profile, Bioorg. Med.
Chem. Lett. 17 (2007) 5340e5344.
[2] P. Zajdel, M. Paw1owski, J. Martinez, G. Subra, Combinatorial chemistry on
solid support in the search for central nervous system agents, Comb. Chem.
High Throughput Screen. 12 (2009) 723e739.
[3] B. Bang-Andersen, T. Ruhland, M. Jørgensen, G. Smith, K. Frederiksen,
K.G. Jensen, H. Zhong, S.M. Nielsen, S. Hogg, A. Mørk, T.B. Stensbøl, Discovery
of 1-[2-(2,4-dimethylphenylsulfanyl)phenyl]piperazine (Lu AA21004): a novel
multimodal compound for the treatment of major depressive disorder, J. Med.
Chem. 54 (2011) 3206e3221.
m
All assays were incubated in a total volume of 200 ml in 96-well
microtitre plates for 1 h at 37 ꢀC, except for 5-HT1AR which were
incubated at room temperature for 1 h. The process of equilibration
is terminated by rapid filtration through Unifilter plates with a 96-
well cell harvester and radioactivity retained on the filters was
quantified on a Microbeta plate reader.
For displacement studies the assay samples contained as radi-
[4] V.J. Gillet, Reactant- and product-based approaches to the design of combi-
natorial libraries, J. Comput. Aided Mol. Des. 16 (2002) 371e380.
oligands: 1.5 nM [3H]-8eOHeDPAT (187 Ci/mmol) for 5-HT1AR;

360
P. Zajdel et al. / European Journal of Medicinal Chemistry 56 (2012) 348e360
[5] V.J. Gillet, P. Willett, J. Bradshaw, The effectiveness of reactant pools for
[28] A.J. Zhang, D.H. Russel, J. Zhu, K. Burgess, A method for removal of N-BOC
protecting groups from substrates on TFA-sensitive resins, Tetrahedron Lett.
39 (1998) 7439e7442.
generating structurally-diverse combinatorial libraries, J. Chem. Inf. Comput.
Sci. 37 (1997) 731e740.
[6] E.A. Jamois, M. Hassan, M. Waldman, Evaluation of reagent-based and
product-based strategies in the design of combinatorial library subsets,
J. Chem. Inf. Comput. Sci. 40 (2000) 63e70.
[7] M. Hartenfeller, E. Proschak, A. Schüller, G. Schneider, Concept of combina-
torial de novo design of drug-like molecules by particle swarm optimization,
Chem. Biol. Drug Des. 72 (2008) 16e26.
[29] CombiGlide, Version 2.7, Schrödinger, LLC, New York, NY, 2011.
[30] Instant JChem was used for structure database management, search and
prediction, Instant JChem 5.3.4. ChemAxon, http://www.chemaxon.com, 2010.
[31] Canvas, Version 1.4, Schrödinger, LLC, New York, NY, 2011.
[32] R. Kurczab, M. Nowak, Z. Chilmonczyk, I. Sylte, A.J. Bojarski, The development
and validation of
a novel virtual screening cascade protocol to identify
[8] J.R. Fischer, U. Lessel, M. Rarey, LoFT: similarity-driven multiobjective focused
library design, J. Chem. Inf. Model. 50 (2010) 1e21.
potential serotonin 5-HT7R antagonists, Bioorg. Med. Chem. Lett. 20 (2010)
2465e2468.
[9] J. Truchon, C.I. Bayly, GLARE: a new approach for filtering large reagent lists in
combinatorial library design using product properties, J. Chem. Inf. Model. 46
(2006) 1536e1548.
[10] W. Zheng, S.T. Hung, J.T. Saunders, G.L. Seibel, PICCOLO: a tool for combina-
torial library design via multicriterion optimization, Pac. Symp. Biocomput.
(2000) 585e596.
[33] Calculator Plugins were used for structure property prediction and calcula-
tion, Marvin 5.4.1. ChemAxon, http://www.chemaxon.com, 2010.
[34] LigPrep, Version X.X, Schrödinger, LLC, New York, NY, 2011.
[35] QikProp, Version 3.4, Schrödinger, LLC, New York, NY, 2011.
[36] M. Ko1aczkowski, N. Nowak, M. Paw1owski, A.J. Bojarski, Receptor-based
pharmacophores for serotonin 5-HT7R antagonists-implications to selectivity,
J. Med. Chem. 49 (2006) 6732e6741.
[37] D.L. Hall, S.A.H. McMullen, Mathematical Techniques in Multisensor Data
Fusion, Artech House, Inc, Norwood, MA, 2004.
[38] P. Willett, Enhancing the effectiveness of ligand-based virtual screening using
data fusion, QSAR Comb. Sci. 25 (2006) 1143e1152.
[39] I.R. Kovel’man, A.I. Tochilkin, O.A. Volkova, V.Z. Dubinskii, Synthesis of 2,3-
dihidropyridoi[1,2,3-de]-1,4-benzoxazinium chloride and some of its deriva-
tives substituted in the carbocyclic ring, Pharm. Chem. J. 29 (1995) 350e352.
[40] S. Butini, S. Gemme, G. Campiani, S. Franceschini, F. Trotta, M. Boriello,
N. Ceres, S. Ros, S.S. Coccone, M. Bernetti, M.D. Angelis, M. Brindisi, V. Nacci,
I. Fiorini, E. Novellino, A. Cagnotto, T. Mennini, K. Sandager-Nielsen,
J.T. Andreasen, J. Scheel-Kruger, J.D. Mikkelsen, C. Fattorusso, Discovery of
a new class of potential multifunctional atypical antipsychotic agents tar-
geting dopamine D3 and serotonin 5-HT1A and 5-HT2A receptors: design,
synthesis, and effects on behavior, J. Med. Chem. 52 (2009) 151e169.
[41] P.J. Lovell, S.M. Bromidge, S. Dabbs, M. Duckworth, I.T. Forbes, A.J. Jennings,
F.D. King, D.N. Middlemiss, S.K. Rahman, D.V. Saunders, G.J. Riley, D.R. Thomas,
[11] P.B. Hedlund, The 5-HT7 receptor and disorders of the nervous system: an
overview, Psychopharmacology 206 (2009) 345e354.
[12] O. Mnie-Filali, L. Lambás-Señas, L. Zimmer, N. Haddjeri, 5-HT7 receptor antago-
nists as a new class of antidepressants, Drug News Perspect. 20 (2007) 613e618.
[13] A. Matthys, G. Haegeman, K. Van Craenenbroeck, P. Vanhoenacker, Role of the
5-HT7 receptor in the central nervous system: from current status to future
perspectives, Mol. Neurobiol. 43 (2011) 228e253.
[14] P. Hedlund, J. Sutcliffe, Functional, molecular and pharmacological advances in
5-HT7 receptor research, Trends Pharmacol. Sci. 25 (2004) 481e486.
[15] A.I. Abbas, P.B. Hedlund, X.-P. Huang, T.B. Tran, H.Y. Meltzer, B.L. Roth, Ami-
sulpride is a potent 5-HT7 antagonist: relevance for antidepressant actions
in vivo, Psychopharmacology 205 (2009) 119e128.
[16] G. Sarkisyan, A.J. Roberts, P.B. Hedlund, The 5-HT7 receptor as a mediator and
modulatorofantidepressant-likebehavior, Behav. BrainRes. 209(2010)99e108.
[17] M. Leopoldo, E. Lacivita, F. Berardi, R. Perrone, P.B. Hedlund, Serotonin 5-HT7
receptor agents: structure-activity relationships and potential therapeutic
applications in central nervous system disorders, Pharmacol. Therap. 129
(2011) 120e148.
A
novel, potent, and selective 5-HT(7) antagonist: (R)-3-(2-(2-(4-
[18] M. Leopoldo, E. Lacivita, F. Berardi, R. Perrone, 5-HT(7) receptor modulators:
a medicinal chemistry survey of recent patent literature (2004e2009), Expert.
Opin. Therap. Pat. 20 (2010) 739e754.
[19] V. Pittalà, D. Pittalà, Latest advances towards the discovery of 5-HT7 receptor
ligands, Mini-Rev. Med. Chem. 11 (2011) 1108e1121.
methylpiperidin-1-yl)ethyl)pyrrolidine-1-sulfonyl)
J. Med. Chem. 43 (2000) 342e345.
[42] Chemical term evaluator was used for evaluate in a batch mode a Chemical
Terms expression on molecules, ChemAxon (http://www.chemaxon.com).
[43] Discovery Studio 2.5 Accelrys Inc., San Diego, CA, USA.
phenol
(SB-269970),
[20] R. Perrone, F. Berardi, N.A. Colabufo, E. Lacivita, M. Leopoldo, V. Tortorella,
Synthesis and structure-affinity relationships of 1-[omega-(4-aryl-1-
piperazinyl)alkyl]-1-aryl ketones as 5-HT(7) receptor ligands, J. Med. Chem.
46 (2003) 646e649.
[44] A. Gaulton, L.J. Bellis, A.P. Bento, J. Chambers, M. Davies, A. Hersey, Y. Light,
S. McGlinchey, D. Michalovich, B. Al-Lazikani, J.P. Overington, ChEMBL:
a large-scale bioactivity database for chemical biology and drug discovery,
Nucleic Acids Res. 40 (2012) 1100e1107.
[21] B. Volk, I. Gacsályi, K. Pallagi, L. Poszávácz, I. Gyönös, E. Szabó, T. Bakó,
M. Spedding, G. Simig, G. Szénási, Optimization of (arylpiperazinylbutyl)
oxindoles exhibiting selective 5-HT7 receptor antagonist activity, J. Med.
Chem. 54 (2011) 6657e6669.
[22] P. Zajdel, K. Marciniec, A. Maslankiewicz, G. Sata1a, B. Duszynska, A.J. Bojarski,
A. Partyka, M. Jastrze˛bska-Wie˛sek, D. Wróbel, A. Weso1owska, M. Paw1owski,
Quinoline- and isoquinoline-sulfonamide derivatives of LCAP as potent CNS
multi-receptor-5-HT(1A)/5-HT(2A)/5-HT(7) and D(2)/D(3)/D(4)-agents: the
synthesis and pharmacological evaluation, Bioorg. Med. Chem. 20 (2012)
1545e1556.
[45] N. Huang, B.K. Shoichet, J.J. Irwin, Benchmarking sets for molecular docking,
J. Med. Chem. 49 (2006) 6789e6801.
[46] Z. Deng, C. Chuaqui, J. Singh, Structural interaction fingerprint (SIFt): a novel
method for analyzing three-dimensional proteineligand binding interactions,
J. Med. Chem. 47 (2004) 337e344.
[47] S. Mordalski, T. Kosciolek, K. Kristiansen, I. Sylte, A.J. Bojarski, Protein binding
site analysis by means of structural interaction fingerprint patterns, Bioorg.
Med. Chem. Lett. 21 (2011) 6816e6819.
[48] T. Joachims, Making large-scale SVM learning practical, in: B. Schölkopf,
C. Burges, A. Smola (Eds.), Advances in Kernel Methods e Support Vector
Learning, MIT-Press, 1999. Software available at: http://svmlight.joachims.org/.
[49] C.-C. Chang, C.-J. Lin, LIBSVM: a library for support vector machines. Software
available at: http://www.csie.ntu.edu.tw/wcjlin/lib, 2001.
ꢀ
ꢀ
[23] M. Leopoldo, E. Lacivita, M. Contino, N.A. Colabufo, F. Berardi, R. Perrone,
Structure-activity relationship study on N-(1,2,3,4-tetrahydronaphthalen-1-
yl)-4-aryl-1-piperazinehexanamides,
a
class of 5-HT7 receptor agents. 2,
J. Med. Chem. 50 (2007) 4214e4221.
[50] Y. Cheng, W. Prusoff, Relationship between the inhibition constant (K1) and
the concentration of inhibitor which causes 50 per cent inhibition (I50) of an
enzymatic reaction, Biochem. Pharmacol. 22 (1973) 3099e3108.
ꢀ
[24] P. Zajdel, G. Subra, A.J. Bojarski, B. Duszynska, M. Paw1owski, J. Martinez,
A new class of arylpiperazine derivatives: the library synthesis on SynPhase
lanterns and biological evaluation on serotonin 5-HT(1A) and 5-HT(2A)
receptors, J. Comb. Chem. 6 (2004) 761e767.
ꢀ
[51] A.J. Bojarski, M.T. Ceg1a, S. Charakchieva-Minol, M.J. Mokrosz, M. Mackowiak,
S. Misztal, J.L. Mokrosz, Structureeactivity relationship studies of CNS agents.
Part 9: 5-HT1A and 5-HT2 receptor affinity of some 2- and 3-substituted
1,2,3,4-tetrahydro-beta-carbolines, Pharmazie 48 (1993) 289e294.
ꢀ
[25] P. Zajdel, G. Subra, P. Verdie, E. Gabzdyl, A.J. Bojarski, BDuszynska, J. Martinez,
M. Paw1owski, Sulfonamides with the N-alkyl-N0-dialkylguanidine moiety as
5-HT7 receptor ligands, Bioorg. Med. Chem. Lett. 19 (2009) 4827e4831.
ꢀ
ꢀ
[52] M.H. Paluchowska, R. Bugno, B. Duszynska, E. Tatarczynska, A. Nikiforuk,
T. Lenda, E. Chojnacka-Wójcik, The influence of modifications in imide frag-
ment structure on 5-HT(1A) and 5-HT(7) receptor affinity and in vivo phar-
macological properties of some new 1-(m-trifluoromethylphenyl)piperazines,
Bioorg. Med. Chem. 15 (2007) 7116e7125.
ꢀ
[26] P. Zajdel, K. Marciniec, A. Maslankiewicz, M.H. Paluchowska, G. Sata1a,
ꢀ
A. Partyka, M. Jastrze˛bska-Wie˛sek, D. Wróbel, A. Weso1owska, B. Duszynska,
A.J. Bojarski, M. Paw1owski, Arene- and quinoline-sulfonamides as novel
5-HT7 receptor ligands, Bioorg. Med. Chem. 19 (2011) 6750e6759.
[27] P. Zajdel, G. Subra, M. Paw1owski, J. Martinez, Solid-phase synthesis of aryl-
alkylamine derivatives using protected aminoalcohol building blocks on
SynPhaseÔ Lanterns, QSAR Comb. Sci. 26 (2007) 215e219.
[53] N. Adham, J.M. Zgombick, J. Bard, T.A. Branchek, Functional characterization of
the recombinant human 5-hydroxytryptamine7(a) receptor isoform coupled
to adenylate cyclase stimulation, J. Pharmacol. Exp. Ther. 287 (1998) 508e514.