Novel potent and selective central 5-HT3 receptor ligands provided with different intrinsic efficacy. 1. Mapping the central 5-HT3 receptor binding site by arylpiperazine derivatives

Article abstract of DOI:10.1021/jm970645i

Synthesis and pharmacological evaluation of a series of condensed quinoline and pyridine derivatives bearing a N-methylpiperazine moiety attached to the 2-position of the quinoline or pyridine nucleus are described. 5-HT receptor binding studies revealed subnanomolar affinity for the 5-HT₃ receptor subtype in some of the compounds under study. The most active compound (5b) displayed a K(i) value about 1 order of magnitude higher than that of quipazine along with a higher selectivity. The potential 5-HT₃ agonist/antagonist activity of four selected compounds was assessed in vitro on 5-HT₃ receptor-dependent [¹⁴C]guanidinium uptake in NG 108-15 cells. Compound 5j acted as a 5-HT₃ agonist in this assay with an EC₅₀ value close to that reported for quipazine, while 5b was a partial agonist with an EC₅₀ value of about 0.25 nM, and compound 5c possessed antagonist properties with an IC₅₀ value (?8 nM) in the same range as those of previously characterized 5-HT₃ receptor antagonists. Qualitative and quantitative structure-affinity relationship studies carried out by making use of theoretical molecular descriptors allowed to elucidate the role of the main pharmacophoric components and to develop a model for the interaction of the 5-HT₃ ligands related to quipazine with their receptor.

Full text of DOI:10.1021/jm970645i

728
J . Med. Chem. 1998, 41, 728-741
Novel P oten t a n d Selective Cen tr a l 5-HT3 Recep tor Liga n d s P r ovid ed w ith
Differ en t In tr in sic Effica cy. 1. Ma p p in g th e Cen tr a l 5-HT3 Recep tor Bin d in g
Site by Ar ylp ip er a zin e Der iva tives
Andrea Cappelli,*,‡ Maurizio Anzini,‡ Salvatore Vomero,‡ Laura Mennuni,§ Francesco Makovec,§ Edith Doucet,
Michel Hamon, Giancarlo Bruni,| Maria R. Romeo,| M. Cristina Menziani,† Pier G. De Benedetti,† and
Thierry Langer∇
Dipartimento Farmaco Chimico Tecnologico, Universita` di Siena, Via Banchi di Sotto 55, 53100 Siena, Italy, Rotta Research
Laboratorium S.p.A., Via Valosa di Sopra 7, 20052 Monza, Italy, Neurobiologie Cellulaire et Fonctionnelle, INSERM U288,
Faculte´ de Me´dicine Pitie´-Salpeˆtrie`re, 91, Boulevard de l’Hoˆpital, 75634 Paris Cedex 13, France, Istituto di Farmacologia,
Universita` di Siena, Le Scotte 6, 53100 Siena, Italy, Dipartimento di Chimica, Universita` degli Studi di Modena,
Via Campi 183, 41100 Modena, Italy, and Institute of Pharmaceutical Chemistry, University of Innsbruck, Innrain 52A,
A-6020 Innsbruck, Austria
Received September 25, 1997
Synthesis and pharmacological evaluation of a series of condensed quinoline and pyridine
derivatives bearing a N-methylpiperazine moiety attached to the 2-position of the quinoline or
pyridine nucleus are described. 5-HT receptor binding studies revealed subnanomolar affinity
for the 5-HT3 receptor subtype in some of the compounds under study. The most active
compound (5b) displayed a Ki value about 1 order of magnitude higher than that of quipazine
along with a higher selectivity. The potential 5-HT3 agonist/antagonist activity of four selected
compounds was assessed in vitro on 5-HT3 receptor-dependent [14C]guanidinium uptake in
NG 108-15 cells. Compound 5j acted as a 5-HT3 agonist in this assay with an EC50 value close
to that reported for quipazine, while 5b was a partial agonist with an EC50 value of about 0.25
nM, and compound 5c possessed antagonist properties with an IC50 value (≈8 nM) in the same
range as those of previously characterized 5-HT3 receptor antagonists. Qualitative and
quantitative structure-affinity relationship studies carried out by making use of theoretical
molecular descriptors allowed to elucidate the role of the main pharmacophoric components
and to develop a model for the interaction of the 5-HT3 ligands related to quipazine with their
receptor.
Ch a r t 1
In tr od u ction
The wide variety of actions exerted by serotonin (5-
hydroxytryptamine, 5-HT) has led to the demonstration
of the existence of several 5-HT receptors, and at least
14 serotonin receptor subtypes have been so far identi-
fied.1 Among them the 5-HT3 receptor subtype has a
special place because it belongs to the ligand-gated ion
channel receptor family.2 Recently, there has been
increased interest in the search for new 5-HT3 receptor
antagonists: e.g., ondansetron (1) (Chart 1), the best
known 5-HT3 receptor antagonist, is on the market as
an antiemetic agent to prevent the vomiting associated
with anticancer chemotherapy. Furthermore, other
possible therapeutic uses in anxiety, schizophrenia, drug
abuse, and age-associated memory impairment are the
subject of intense investigations.3,4
It has been suggested that the stimulation of the
5-HT3 receptor in the central nervous system (CNS)
enhances the release of dopamine from rat striatal
slices5 and that of cholecystokinin from the cerebral
cortex and nucleus accumbens,6 and it also inhibits the
release of acetylcholine from the entorhinal cortex.7
‡ Dipartimento Farmaco Chimico Tecnologico, Universita` di Siena.
§ Rotta Research Laboratorium S.p.A.
INSERM U288.
However, the potential therapeutic interest of the 5-HT3
agonists still remains to be evaluated because of the lack
of potent and selective compounds.
| Istituto di Farmacologia, Universita` di Siena.
† Universita` degli Studi di Modena.
∇ University of Innsbruck.
S0022-2623(97)00645-6 CCC: $15.00 © 1998 American Chemical Society
Published on Web 02/05/1998

Mapping the Central 5-HT3 Receptor Binding Site
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 5 729
Sch em e 1a
Quipazine (2a ) was earlier reported to behave as a
5-HT3 antagonist in peripheral models,8 but subsequent
studies revealed its 5-HT3 agonist action in the [14C]-
guanidinium accumulation test applied to NG 108-15
cells.9 It is noteworthy that the 5-HT3 receptor ex-
pressed by NG 108-15 cells is very similar to that
present in rat cortex.10 Thus, quipazine is very probably
a 5-HT3 receptor agonist in the CNS.
Very recently, we reported the synthesis and the
pharmacological characterization of a new class of
pyrroloquinoxaline derivatives (3) related to quipazine
which behaved as 5-HT3 receptor agonists in both
[14C]guanidinium accumulation in NG 108-15 cells and
cortical acetylcholine release in freely moving rats, while
they showed partial agonist or antagonist properties in
the Bezold-J arish reflex test.11 Furthermore, in a
previous paper we described the synthesis and the
5-HT3 receptor antagonist properties of a new class of
quipazine derivatives (4).12 In the context of such work
we developed a comparative molecular field analysis
(CoMFA) model able to predict nanomolar 5-HT3 recep-
tor affinity for compounds 5a ,13 5b,14 and 5c (predicted
affinities: 5a , Ki ) 10 nM; 5b, Ki ) 11 nM; 5c, Ki ) 9.9
nM),15 and we proposed a tentative interaction model
which implies different but overlapping binding domains
for both antagonists and agonists at the 5-HT3 recep-
tor.16
a
Reagents: (i) (a) pyrrolidine, C2H5OC2H5, TiCl4, C6H6, (b)
C6H5NCO, CHCl3, (c) PPA; (ii) POCl3; (iii) amine (N-methyl-
piperazine, morpholine, etc.).
Sch em e 2a
The aim of the present work was two-fold: (a) to
obtain new 5-HT3 receptor ligands with improved
selectivity versus the main 5-HT receptor subtypes
which are known to interact with quipazine derivatives
and (b) to map the arylpiperazine interactions at the
central 5-HT3 receptor.
a
Reagents: (i) PPA; (ii) (a) (CF3SO2)2O, Na2CO3, CH2Cl2, (b)
N-methylpiperazine.
Thus, we report in this paper the synthesis and 5-HT3
receptor affinities of several quipazine derivatives in
series 5 and 6. The effects of different heteroaromatic
portions on the binding to the 5-HT3 receptor sites are
discussed along with the effect obtained by replacing
the N-methylpiperazine moiety with different N-sub-
stituted piperazines or different heterocycles. The
5-HT3 receptor intrinsic efficacy of the selected com-
pounds was assessed in vitro on 5-HT3 receptor-depend-
ent [14C]guanidinium uptake in NG 108-15 hybrid cells.9
Theoretical descriptors derived by means of the program
CODESSA17 and ad hoc-defined size and shape descrip-
tors18 have been employed for deciphering, on a quan-
titative ground, the molecular features responsible for
the recognition of the 5-HT3 receptors. Finally, a
comprehensive model for the interaction of the 5-HT3
receptor ligands related to quipazine with their receptor
is proposed.
Sch em e 3a
a
Reagents: (i) HCtC-COOCH3, NH3, CH3OH; (ii) (CF3SO2)2O,
Na2CO3, CH2Cl2; (iii) N-methylpiperazine.
ride. The quaternary compound 5bb was obtained by
methylation of the terminal piperazine nitrogen of 5b
with iodomethane.
Ch em istr y
Compounds 5a -c, 5ba ,bc,bd ,be,bf, and 6a 19 were
prepared starting from the corresponding imidoyl chlo-
rides (9a ,20 9b,21 9c, or 2-chloropyridine) by reaction
with the appropriate amine. Imidoyl chloride 9c was
prepared following the procedure summarized in Scheme
1. Bicyclo[3.2.1]octan-2-one (7) was converted, following
the procedure described by White et al.,22 into the
corresponding pyrrolidine enamine which was reacted
with phenyl isocyanate and then cyclized into quinoli-
none 8. The latter compound was readily converted into
imidoyl chloride 9c by refluxing in phosphorus oxychlo-
Compound 5d was synthesized following the proce-
dure described in Scheme 2. Cyclization of amide 10
under the conditions described by Rigby et al.23 ap-
peared to be slow even in refluxing toluene, while it
proceeded rapidly when it was run in polyphosphoric
acid (PPA) at 140 °C, and 2(1H)-pyridone 11 was
obtained in high yield. The latter was converted in two
steps into the corresponding piperazinyl derivative 5d .
Compounds 5e,g-k were prepared according to the
procedure described in Scheme 3. The procedure for
2(1H)-pyridone synthesis originally developed by

730 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 5
Cappelli et al.
Sch em e 4a
Molecu la r Descr ip tor s a n d Sta tistica l Tr ea t-
m en t of Da ta . A large number of global and fragment
descriptors (constitutional, topological, electrostatic,
geometrical, and quantum-chemical) were generated for
each compound within the framework of the CODES-
SA17 program. The ad hoc-defined molecular size and
shape parameters18a described above were added as
external descriptors.18b The program was used in this
work to accomplish a preselection of descriptors on the
basis of their statistical significance and, owing to the
limited number of compounds considered, to obtain
simple regression models.
The search for the best correlation equation was
achieved by means of a heuristic method, which ac-
complished a preselection of descriptors on the basis of
their statistical significance.17,31 Default values for
control parameters and criteria were used: minimum
squared correlation coefficient for the one-parameter
correlation to be considered significant, R2
) 0.1;
min
a
Reagents: (i) MCPBA, CHCl3; (ii) POCl3; (iii) N-methyl-
t-test value for the descriptor to be considered significant
in the one-parameter correlation, t1 ) 1.5; significant
intercorrelation level, rsig ) 0.8. As a final result, the
program lists the 10 correlations found with the highest
F-test values. The quantitative structure-affinity re-
lationship (QSAR) models reported in this paper were
selected on the basis of the best statistical parameters
and the largest diversity of the descriptors involved.
piperazine.
Kozikowski et al.24 was slightly modified25 and applied
to the required cyclic ketones 12e,g-k (see Experimen-
tal Section) to obtain pyridones 13e,g-k . These com-
pounds were transformed into the piperazinyl deriva-
tives 5e,g-k in two steps involving the reaction of the
pyridones 13e,g-k with trifluoromethanesulfonic an-
hydride to give the corresponding imidoyl triflates
14e,g-k which, without further purification,26 reacted
with N-methylpiperazine to give 5e,g-k . Compound
6b was prepared following the same reaction sequence
as that described above for compounds 5e,g-k starting
from 1-adamantyl methyl ketone.
The procedure for the synthesis of compound 5f is
shown in Scheme 4. Commercial benzo[h]quinoline (15)
was transformed into the corresponding N-oxide deriva-
tive 1627 which reacted with phosphorus oxychloride to
give the 2-chloro derivative 1727 along with a minor
amount of its 4-chloro isomer 18 which was separated
by flash chromatography. Reaction of 17 with N-
methylpiperazine gave the corresponding piperazinyl
derivative 5f.
Defin ition s of th e Descr ip tor s Used .32 (1) EL
-
EH is the energy difference between the lowest un-
occupied (LUMO) and the highest occupied (HOMO)
molecular orbitals. A decrease in the energy gap usually
leads to easier polarizability of the molecule. (2) SL(N)
is the nucleophilic superdelocalizability on the pro-
tonated terminal piperazine nitrogen atom. It charac-
terizes the propensity to behave as a hydrogen bond
donor. (3) SH(Nh) is the electrophilic superdelocaliz-
ability on the nitrogen atom of the heteroaryl moiety.
It characterizes the propensity to behave as a hydrogen
bond acceptor. For this purpose, the relevant orbital
for 5a is the HOMO-1, whereas the HOMO molecular
orbital is localized on the carbon atoms of the conjugate
system. The energy gap between the two orbitals is only
0.2 eV. (4) ∑fa - SH is the sum of the electrophilic
superdelocalizability on the atoms of the pyridine ring
of the heteroaryl moiety. (5) ∑fa′ - SH is the sum of
the electrophilic superdelocalizability on the atoms of
the benzene ring of either the quinoline nucleus or the
saturated ring in an analogous position, or even of the
substituent of ligands 6b,c. (6) ∑fc - SH is the sum of
the electrophilic superdelocalizability on the atoms of
the additional ring fused to the bicyclic system. (7) ∑f
- SH is the sum of the electrophilic superdelocalizability
on the atoms of the entire heteroaryl moiety. These
indices are used to parametrize π-π drug-receptor
interactions at specific atoms in unsaturated molecules.
Com p u ta tion a l Meth od s
Geom etr y Op tim iza tion . Conformational analysis
of the N4-protonated form of some simple arylpiperazine
derivatives was recently performed,28 and the resulting
absolute minimum structure of quipazine was consid-
ered in this study as the starting geometry. The
substituents were assembled within the Editor module
of the QUANTA 4.1 program.29 Then, the protonated
structures were fully optimized by means of molecular
orbital calculations (AM1),30 using the MOPAC 6.0
(QCPE 455) program.
Molecu la r Su p er im p osition . The most structur-
ally different ligands which show the highest affinity
for the 5-HT3 receptor (compounds 5b,j) were chosen
for the construction of the reference supermolecule and
were superimposed by a rigid fit procedure over the
piperazine moiety. The QUANTA 4.1 molecular model-
ing software was used for molecular comparisons,
matching, and computation of van der Waals volumes.
(8) HASA is the area-weighted surface charge of the
hydrogen-bonding acceptor atoms in the molecule.31 (9)
faa′ - WNSA-3 is the surface-weighted charged partial
surface area of the heteroaryl moiety.33 WNSA-3 is
calculated according to the following formula: WNSA-3
) [∑(-SAi)(Q- )]TMSA/1000, where -SAi is the surface
i
area contribution of the ith negative atom in the
molecule, Q- is the partial atomic charge for the ith
i
negative atom, and TMSA is the total molecular surface

Mapping the Central 5-HT3 Receptor Binding Site
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 5 731
area. (10) Vin and Vout are ad hoc-defined size and shape
descriptors18 and represent, respectively, the intersec-
tion and the outer van der Waals volume of the ligands
considered with respect to a supermolecule constituted
by the most structurally different ligands showing the
highest binding affinity for the 5-HT3 receptor (com-
pounds 5b,j). (11) Vdif is computed according to the
following formula: Vdif ) Vin - Vout/Vsup, where Vsup is
the resulting van der Waals volume of the reference
supermolecule. According to the ligand-receptor com-
plementarity paradigm, this approach assumes that the
volume obtained by superimposing the most structurally
different ligands which show the highest affinities for
the same receptor might reflect the overall shape and
the conformational flexibility of the high-affinity recep-
tor binding site.
structure of NMQ (as in compound 5b) seems to confer
the best stereoelectronic features for the interaction of
these quipazine derivatives with the 5-HT3 receptor. A
comparative inspection of the electronic indices listed
in Table 3 shows an average reduction of 5.6 kcal/mol
in the energy gap (EL - EH) of these ligands with respect
to NMQ. This effect can be interpreted as an increase
in their polarizability and, hence, in their potentiality
of establishing attractive short-range interactions with
the receptor area approaching the c-edge of the quino-
line nucleus of NMQ. Moreover, a slightly higher
propensity to behave as a hydrogen-bonding acceptor
(SH(Nh)) is also observed for the quinoline nitrogen atom
of compounds 5a -c with respect to NMQ through a
perturbation of the electronic cloud localized on the
pyridine ring of the heteroaryl moiety (∑fa - SH).
(b) Mod ifica tion of th e Ben zen e Rin g of th e
Qu in olin e Nu cleu s of NMQ. It is interesting to
observe that both the deletion of the fused benzene ring
of NMQ (compare NMQ with 6a ) and its saturation
(compare NMQ with 5e and 5b with 5d ) lead to a
significant decrease of the 5-HT3 affinity suggesting that
this ring plays a key role in the interaction of this class
of compounds with the receptor. In fact, the trend of
the dynamic reactivity indices seems to suggest a dual
role of the fused benzene ring for (1) a specific interac-
tion (π-π or π-charge) with suitable amino acid residues
of the receptor (compare ∑fa′ - SH of NMQ and 5b with
that of 5e and 5d ) and (2) an indirect but substantial
effect on the capability of establishing hydrogen-bonding
interactions of the nitrogen atom of the quinoline moiety
(compare SH(Nh) of NMQ and 5b with that of 5e and
5d ).
Resu lts a n d Discu ssion
Qu a lita tive Str u ctu r e-Affin ity Rela tion sh ip s.
The arylpiperazine structure can be formally divided
into two substructures: (a) the piperazine moiety which
incorporates a nitrogen atom protonated at physiological
pH and (b) the aryl moiety which, in the case of most of
the 5-HT3 ligands, is a nitrogen heterocycle linked to
the piperazine moiety by a “pseudoamidinic” bond.
As the first step of this work we selected the optimized
N-methylpiperazine substructure,11,12 and the hetero-
aryl moiety was modified with the aim of exploring the
role of this substructure in the interaction with 5-HT3
receptor. Subsequently, the optimized heteroaryl moi-
ety was retained, and the effects of replacement of the
N-methylpiperazine moiety with either different N-
substituted piperazines or related heterocycles were
assessed.
The newly synthesized compounds were tested for
their potential ability to displace [3H]granisetron specif-
ically bound to 5-HT3 receptors in rat cortical mem-
branes (in comparison with the reference compounds
quipazine (2a ) and unlabeled granisetron) following
well-established protocols.34 The results of the binding
studies along with the Ki values reported in the litera-
ture35 for N-methylquipazine (NMQ, 2b) and pyridine
derivatives 6c,d (which are included for comparison) are
summarized in Tables 1 and 2. A detailed picture of
the reactivity determinants of some of these ligands is
reported in Table 3. The descriptors have been com-
puted, within the framework of the AM1 Hamiltonian,30
on the protonated ligands, and their definition is given
in the Computational Methods section.
E ffect s of t h e Mod ifica t ion of t h e H et er oa r yl
Su bstr u ctu r e. (a) In tr odu ction of Addition al Rin gs
a t th e c-Ed ge of th e Qu in olin e Nu cleu s of NMQ.
The most intriguing result emerging from these studies
is that the fusion of a benzo ring, a cyclohexane ring, or
a saturated bicyclic system on the c-edge of the quinoline
nucleus of NMQ (leading to compounds 5a -c, respec-
tively) enhances the affinity for 5-HT3 receptors to a
variable extent. In fact, benzo derivative 5a shows a
5-HT3 receptor affinity slightly higher than that re-
ported for NMQ, while 5c is found to be slightly more
potent than 5a , and compound 5b shows an affinity
about 1 order of magnitude higher than NMQ. There-
fore, the incorporation of a cyclohexane ring into the
It is noteworthy that the halogen atom of compounds
6c,d is able to mimic the condensed benzo ring of NMQ
only to a limited extent. Although an underestimation
of the halogen effects might be expected since d-orbitals
on the halogen atoms are neglected in the AM1 calcula-
tion, the trend in the binding affinity of these ligands
is well accounted for by ∑f - SH and SH(Nh). These
indices show that the introduction of the halogen atoms
on the pyridine ring increases both the ability to donate
electron density through orbitals and the propensity of
the pyridine nitrogen atom to behave as a hydrogen-
bonding acceptor.
(c) In tr od u ction of Ad d ition a l Rin gs a t h - a n d
f-Ed ges of th e Qu in olin e Nu cleu s. The benzofusion
on the h-edge of the quinoline nucleus of NMQ and 5e
(leading to compounds 5f,g, respectively) significantly
decreases the affinity for the 5-HT3 receptor suggesting
a limited tolerance to large substituents in the area of
the receptor protein surrounding the h-edge of the
quinoline nucleus of NMQ. This consideration appears
to be also supported by the low affinity shown by
compounds 5h ,i and 6b. However, the comparison
between compounds 5g and 5h suggests that the phenyl
ring in the 6-position of the pyridine ring plays a role
in the interaction with the receptor similar to that of
the fused benzo ring of NMQ. A comparison of the ∑fc
- SH index of 5g with ∑fa′ - SH of NMQ reveals
similarity in the reactivity of their π-electron. More-
over, the unsaturation of the ring makes the basicity of
the quinoline nucleus comparable with that of NMQ
(compare SH(Nh) for compounds 5h ,i and NMQ).

732 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 5
Cappelli et al.
Ta ble 1. Preparative, Analytical, and 5-HT3 Receptor Binding Affinity Data of Compounds 5a -k and 6a ,b: Effects of Modification of
the Heteroaryl Substructure
c
a
,
,
K
b
d
a
b
Yields are not optimized. Values in parentheses are the yields as salt. Recrystallization solvents: E, ethyl acetate; M, methanol; D,
diethyl ether. c Analyses for the elements indicated were within (0.4% of theoretical values. Each value is the mean ( SEM of three
d
determinations. e See ref 35.
In this respect, the high affinity shown by compound
5j appears to be particularly interesting. This ligand
is nearly equipotent to NMQ, is about 1 order of
magnitude more potent than its higher homologue 5g,
and seems to combine the best electronic features for
the interaction together with a geometry which forces
the phenyl ring into a position in space favorable for
interaction with the 5-HT3 receptor.
On the contrary, the low affinity shown by compound
5k (compare 5k with 5e) suggests that in the area of
the receptor protein interacting with the f-edge of the
quinoline nucleus of NMQ there is a limited tolerance

Mapping the Central 5-HT3 Receptor Binding Site
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 5 733
Ta ble 2. 5-HT3 Receptor Binding Affinities and Nucleophilic Superdelocalizabilities on the Terminal Piperazine Nitrogen Atom of
Compounds 5b,ba -bf: Effects of Modification of the Piperazinyl Substructure
compd
R
Ki (nM) ( SEMa
SL(N) (eV-1
)
5b
5ba
5bb
5bc
5bd
5be
5bf
2a
4-methyl-1-piperazinyl
morpholino
4,4-dimethylpiperazinium-1-yl
1-piperazinyl
4-ethyl-1-piperazinyl
4-benzyl-1-piperazinyl
4-methyl-1-homopiperazinyl
0.23 ( 0.002
>17000
0.0417
13 ( 1.1
0.0401
0.0433
0.0404
0.0297
0.0405
0.0436
1.7 ( 0.6
1.9 ( 0.2
11 ( 2.1
4.8 ( 0.7
1.8 ( 0.3
0.35 ( 0.06
granisetron
a
Each value is the mean ( SEM of three determinations.
Ta ble 3. 5-HT3 Binding Affinities and Theoretical Descriptors of the Protonated Form of the Ligands Reported in Table 1
EL - EH SL(N)
(eV-1 (eV-1
SH(Nh)
(eV-1
∑fa - SH ∑fa′ - SH ∑fc - SH ∑f - SH
(eV-1 (eV-1 (eV-1 (eV-1
HASA faa′ - WNSA-3
Vin
(Å3)
Vout
(Å3)
compd
pKi
)
)
)
)
)
)
)
(Å2)
(Å2)
Vdif
2a
2b
2c
5a
5b
5c
5d
5e
5f
8.74
8.52
7.24
8.72
9.64
9.08
6.96
7.23
6.7
7.6071
7.6302
8.1125
7.3386
7.4383
7.4074
7.8286
8.039
0.0436 0.00639
0.0419 0.00648
0.0435 0.00718
0.0419 0.00670
0.0417 0.00762
0.0417 0.00770
0.0414 0.000305
0.0417 0.00000695
0.0418 0.000437
0.0417 0.00186
0.0416 0.0000239
0.0416 0.00155
0.0416 0.00283
0.0416 0.0000723
0.0422 0.0000251
0.0418 0.000470
0.0426 0.000681
0.0425 0.000493
0.0220
0.0240
0.0520
0.0240
0.0190
0.0195
0.0410
0.0460
0.0145
0.0205
0.0460
0.0140
0.0195
0.0195
0.0670
0.0560
0.0620
0.0490
0.0560
0.0550
0.0240
0.0280
0.0470
0.0470
0.0210
0.0210
0.0310
0.0225
0.0215
0.0200
0.0225
0.0230
0.0780 41.6916
0.0790 22.4493
0.0760 23.2511
0.0810 24.8546
0.0780 26.4581
0.0800 23.2511
0.0690 16.0352
0.0670 17.6387
0.0820 19.8511
0.0800 18.4405
-5.722
-5.941
193.13
208.63
213.63 16.5
247.75
264.5
287.5
264.75 15.13 0.744
209.75 12.63 0.5875
2.25 0.5689
2.87 0.6133
0.5876
4.13 0.7261
-4.8064
-4.5851
-4.8512
-4.2586
-1.1223
-2.5778
-4.1577
-2.5246
-2.0343
-2.3992
-2.6689
-2.4163
-3.5771
-3.6804
-3.1531
-3.432
0.0290
0.0120
0.0135
0.00800
0
0
0.7884
0.8569
7.061
0.0365
0.0370
0.00250
0.0460
0.0400
0.0385
231.13 20
230.25 28.5
0.6293
0.6013
44.88 0.5577
5g
5h
5i
6.7
7.2663
8.0012
7.2498
7.4124
7.1557
8.5464
8.122
6.18
<6.26
7.85
6.08
7.09
<6.26
7.42
7.32
0.0700
9.4282
232
212.25 62.13 0.4475
241.75 0.7206
213.88 44.12 0.506
168.63 1.37 0.4985
222.63 76.37 0.4359
0.0800 11.8335
0.0820 21.6475
0.0810 20.8458
0.0670 26.4581
0.0710 10.4229
0.0750 17.6387
0.0730 22.4493
5j
0
5k
6a
6b
6c
6d
0.0140
0.0130
0.0240
8.3273
8.3327
182.5
183.75
4.63 0.5302
4 0.5358
to large substituents. The superimposition among the
active ligands NMQ and 5j, the less active 5g,k , and
the inactive 5i accounts for these observations (Figure
1).
nitrogen, but it is not protonated at physiological pH).
Since 5ba was found to be at least 5 orders of magnitude
less active than 5b (see Table 2), the crucial role of the
protonated terminal piperazine nitrogen in the interac-
tion with the receptor became rapidly apparent. Similar
results have been reported in the literature about the
replacement of the terminal piperazine NH of qui-
pazine35 and related arylpiperazines36 either by an
isosteric methylene group or by an oxygen atom. It is
In conclusion, the strict geometric prerequisites for
the interaction of the phenyl ring with the suitable
amino acid residue(s) in the receptor strongly support
the assumption that an aromatic specific interaction is
involved in this area of the receptor binding site.
Moreover, the differences between the structure-affin-
ity relationships (SAFIR) concerning the phenyl ring in
the 4-position of the quinoline nucleus of 5-HT3 receptor
antagonists 4 (where compounds with the phenyl group
restrained with different geometries showed very simi-
lar 5-HT3 receptor affinities)12 and those concerning the
phenyl ring in the 6-position of the pyridine nucleus of
compounds 5f-j suggest that the phenyl group in these
two series of compounds binds different receptor areas,
thereby reflecting their differences in functional proper-
ties (compounds 4 are antagonists12 and 5j is an agonist
at the central 5-HT3 receptor).
noteworthy that the estimated value of the free energy
37
contribution (∆Gbind
)
of the protonated nitrogen atom
of 5b to the binding of 5-HT3 receptors (>6.6 kcal/mol,
as an absolute value) is in the range of the assumed
charge-assisted hydrogen bond interactions.38
Quaternization (i.e., N-methylation) of the terminal
piperazine nitrogen of compound 5b leads to a 56-fold
decrease of 5-HT3 receptor affinity, in agreement with
previous observations on other arylpiperazine deriva-
tives.35,36,39 This compound is likely to retain the
electrostatic component in its binding to the receptor,
but it loses the hydrogen-bonding component (∆Gbind
)
Effects of Mod ifica tion s of th e P ip er a zin yl Su b-
str u ctu r e. To establish whether the nitrogen atom of
5b binds the receptor in its protonated form, the
terminal piperazine tertiary nitrogen of 5b was replaced
by an isosteric oxygen atom as in compound 5ba (oxygen
retains the hydrogen-bonding acceptor properties of
2.4 kcal/mol, as an absolute value) because its terminal
piperazine nitrogen cannot be protonated.
The replacement of the methyl group on the terminal
piperazine nitrogen of 5b by a hydrogen atom (as in
5bc), by an ethyl group (as in 5bd ), or by a bulkier
benzyl group (as in 5be) invariably leads to a drop in

734 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 5
Cappelli et al.
Ta ble 4. Regression Models for the Correlation between
Theoretical Molecular Structure Descriptors and 5-HT3
Receptor Binding Affinity
SD of
regression
coeff
regression
coeff
descriptor
(1) R ) 0.839, n ) 16, F ) 33.2, s ) 0.596
intercept
6.763
0.207
SH(Nh)
271.648
47.128
R ) 0.942, n ) 13, F ) 86.5, s ) 0.411;
Omitted 6c,d , 6-Nitroquipazine
intercept
6.612
334.620
0.159
35.974
SH(Nh)
(2) R ) 0.824, n ) 11, F ) 19.1, s ) 0.644;
Omitted 5k , 6a ,c,d , Quipazine
intercept
HASA
3.776
0.194
0.923
0.044
(3) R ) 0.823, n ) 13, F ) 22.9, s ) 0.646;
Omitted 5d ,f, 6-Nitroquipazine
intercept
faa′ - WNSA-3
5.105
-0.716
0.580
0.149
(4) R ) 0.795, n ) 11, F ) 15.5, s ) 0.761;
Omitted 6c,d , NMQ, Quipazine
intercept
4.742
98.884
0.735
25.144
∑fa′ - SH
(5) R ) 0.841, n ) 13, F ) 26.61, s ) 0.611;
Omitted 5d , NMQ, Quipazine
intercept
Vdif
2.469
8.017
0.986
1.554
for the interaction with the suitable amino acid residue
within the receptor binding site.
F igu r e 1. Mapping of the aromatic interactions within the
5-HT3 receptor binding site by superimposition of active
ligands NMQ and 5j (red) to the less active 5g,k and inactive
5i (blue). At the top of the figure the ligands are split according
to their affinity, and front and side views are reported.
Qu a n t it a t ive St r u ct u r e-Affin it y R ela t ion sh ip
An a lysis. The ability of the CODESSA17 program to
provide a fast and adequate search for the best correla-
tion equations for a given property has been exploited
to obtain linear QSAR models. All theoretical descrip-
tors derived within CODESSA were considered with this
goal in view, together with ad hoc-defined size and
shape descriptors.18
A limited set of descriptors can provide a quantitative
rationalization of the binding properties of these ligands
(Table 3). The QSAR models obtained are reported in
Table 4. The descriptors involved emphasize (a) the role
of the quinoline nitrogen atom as hydrogen-bonding
acceptor with respect to a suitable amino acid residue
of the receptor (eqs 1 and 2), (b) the electrostatic
component (eq 3) and the perturbation effect (eq 4) on
the electron density localized on the heteroaryl moiety,
and (c) the dispersive and steric interaction realized by
the heteroaryl moiety (eq 5).
the 5-HT3 receptor affinity. Thus, the methyl group
appears to be optimal (in the limited series of substit-
uents considered) for the interaction of these central
5-HT3 partial agonists (see Biological Activity section)
with their receptor. This SAFIR trend appears to be in
good agreement with that observed for analogous 5-HT3
agonists11 and antagonists;12,39 thus, though a common
mode of binding for arylpiperazine derivatives showing
different intrinsic efficacy at the central 5-HT3 receptor
may be assumed, ad hoc molecules need to be designed
to test this hypothesis. On the contrary, the discrepancy
between the above-mentioned SAFIR trend and that
reported for a series of piperazinylpyrrolothienopyrazine
derivatives36 may be related to the different receptor
sources used in the respective studies.
The reactivity determinants reported in Table 2 only
roughly account for the variation in the binding affinity
of these ligands. According to these indices, substitution
of the protonated nitrogen of quipazine results in a
decrease of the N-H+ acidity, the most significant
reduction being registered for 5be. Therefore, to explain
the positive contribution of the methyl group linked to
the terminal piperazine nitrogen on the binding affinity,
other effects have to be invoked, such as lipophilic
interactions with minimal steric demand.
Finally, the replacement of the N-methylpiperazine
moiety of 5b by the higher homologue N-methyl-
homopiperazine leads to a 21-fold decrease in 5-HT3
receptor affinity (compare 5b with 5bf). The ring
distortion seems to prevent the protonated nitrogen
from reaching the optimal distance and/or directionality
The crucial role of the piperazine protonated nitrogen
is missing from this scenario. In fact, the values of the
nucleophilic index SL(N) (reported in Table 3) being
almost constant in this series of ligands are unable to
describe the variation of the ligand binding affinities
quantitatively. Therefore, in conformity to a general
mode of interaction of bioamine ligands with their
receptors, a long-range electrostatic interaction has to
be assumed as a preliminary recognition step.18a Once
the main docking has been accomplished, the binding
affinity may be modulated by the optimization of the
short-range intermolecular interactions and dispersion
contributions of the heteroaryl moiety.
The data in Table 4 show that the best correlation is
obtained by the SH(Nh) descriptor. This index explains
70% of the variation in the binding affinity, when the

Mapping the Central 5-HT3 Receptor Binding Site
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 5 735
Ta ble 5. 5-HT Receptor Binding Profile of Compounds 5b,c,j (Ki (nM) ( SEM)a
compd
5-HT3
5-HT1A
5-HT1B
>4000
(>17000)
2300 ( 890
(2800)
5-HT2A
5-HT2C
5-HTtransporter
5b
5c
5j
0.23 ( 0.002
2800 ( 88
(12000)
2400 ( 410
(2900)
5700 ( 180
(410)
3600 ( 800
(2000)
3000 ( 140
(13000)
3700 ( 240
(4500)
2600 ( 130
(190)
1800 ( 480
(1000)
4400 ( 470
(19000)
2800 ( 910
(3400)
12000 ( 3700
(52000)
24000 ( 3200
(29000)
59 ( 1.9
(4.2)
31 ( 2.9
(17)
0.83 ( 0.03
14 ( 1.1
3800 ( 420
16000 ( 5200
(270)
(1100)
2a (quipazine)
1.8 ( 0.3
310b
190b
(170)
(110)
granisetron
5-HT
8-OH-DPAT
mesulergine
2c (6-nitroquipazine)
0.35 ( 0.06
120 ( 34
7.3 ( 1.2
1.2 ( 0.13
1.1 ( 0.098
740 ( 120
0.73 ( 0.083
58 ( 3.2
0.12 ( 0.01
a
Each value is the mean ( SEM of three determinations. Each value in parentheses is the selectivity ratio calculated as the ratio of
b
the Ki value for the indicated site over that for 5-HT3 receptors. See: Schoeffter, P.; Hoyer, D. Interaction of Arylpiperazines with
5-HT1A, 5-HT1B, 5-HT1C, and 5-HT1D Receptors: Do Discriminatory 5-HT1B Receptor Ligands Exist? Naunyn-Schmiedeberg’s Arch.
Pharmacol. 1989, 339, 675-683.
16 ligands with significant affinities listed in Table 3
are considered. A notable improvement in the correla-
tion statistics is obtained by omitting 6-nitroquipazine
and compounds 6c,d , for which the neglect of d-orbitals
in the computation might result in an underestimation
of the effects of the halogen atoms on the pyridine ring.
A satisfactory correlation is also obtained by using
HASA (eq 2). Since the molecular descriptor is com-
puted on the whole ligand, it takes into account all the
possible hydrogen-bonding acceptor centers of the mol-
ecule simultaneously.
The surface-weighted charged partial surface area of
the heteroaryl moiety, faa′ - WNSA-3, might be consid-
ered to rationalize the electrostatic component of the
binding adjustments which modulate the binding affin-
ity after the main docking has been accomplished. A
good correlation (eq 3) is obtained by omitting com-
pounds 5f,d , outliers with overpredicted binding affin-
ity, and 6-nitroquipazine, which is, on the contrary,
underpredicted.
Correlation 4, involving the sum of either the elec-
trophilic superdelocalizability on the atoms of the fused
benzene ring of the heteroaryl moiety or the saturated
ring in an analogous position, ∑fa′ - SH, corroborates
the qualitative conclusions reported in the previous
section on the important role of this portion of the
ligands in the binding to the 5-HT3 receptor. It explains
60% of the variation in the binding affinity, by omitting
the halogen derivatives 6c,d , quipazine, and NMQ
which were found to have been overestimated.
Finally, eq 5 involves the ad hoc-derived size and
shape descriptor Vdif, which, by taking into account in
its formulation both the inner and outer molecular
volumes of the ligands considered with respect to the
reference volume of the supermolecule, indicates that,
in these series of compounds, the binding affinity is
modulated by the molecular shape of the heteroaryl
moiety through the optimization of dispersive and steric
interactions.
studies (summarized in Table 5) show that, while 5j
displays a binding profile very similar to that of qui-
pazine, the highly potent compounds 5b,c are also very
selective, indeed showing selectivity ratios higher than
those of quipazine. These results clearly demonstrate
that among the 5-HT receptor subtypes interacting with
quipazine (with medium to high affinity) only 5-HT3
tolerates additional steric hindrance in the receptor area
corresponding to the c-edge of quipazine.
F u n ction a l Activity. The measurements of [14C]-
guanidinium uptake in the presence of substance P (SP)
appear to be a rapid and reliable method for assessing
functional properties of 5-HT3 receptor ligands in NG
108-15 hybridoma cells.9 Owing to the similarities
existing between 5-HT3 receptors present on NG 108-
15 cells and those of rat cortex,10 relevant inferences
can be drawn from this pharmacological model regard-
ing the functional properties of ligands of the central
5-HT3 receptor.11 Accordingly, the potential 5-HT3
receptor agonist/antagonist activity of three selected
compounds was assessed on the 5-HT3 receptor-depend-
ent [14C]guanidinium uptake in NG 108-15 cells. In the
presence of SP (10 µM), compound 5j increased in a
concentration-dependent manner the uptake of [14C]-
guanidinium into NG 108-15 cells with an EC50 value
(≈20 nM) close to that reported for quipazine,9 indicat-
ing that it acted as a 5-HT3 agonist in this assay. The
maximum increase of the [14C]guanidinium uptake was
similar to that due to 5-HT, and the effect of 5j could
be completely antagonized by ondansetron (1) but was
unaltered in the presence of 5-HT (1 µM). On the other
hand, compound 5b appeared to act as a partial agonist
with an EC50 value of about 0.25 nM, since it stimulated
[14C]guanidinium uptake in the presence of SP (10 µM),
up to a level equal to about two-thirds of that reached
with 5-HT (1 µM) plus SP (10 µM). Indeed, in the
presence of the latter two agents, 5b produced a
decrease in the accumulation of [14C]guanidinium (Fig-
ure 2). Finally, compound 5c was revealed to possess
antagonist properties in this test as it reversed in a
concentration-dependent manner the stimulation effect
of 5-HT (in the presence of SP) on [14C]guanidinium
uptake with an IC50 value (≈8 nM) in the same range
as those of well-characterized 5-HT3 receptor antago-
nists.12
Selectivity. Three representative compounds were
selected among the most potent inhibitors of [3H]-
granisetron specific binding. These compounds were
tested for their possible binding to several 5-HT receptor
subtypes interacting with quipazine derivatives (5-HT1A
,
5-HT1B, 5-HT2A, 5-HT2C) and the 5-HT transporter in
comparison with reference compounds using well-
established protocols.40-43 The results of these binding
Ar ylp ip er a zin e In ter a ction s a t th e Cen tr a l 5-HT3
Recep tor . In conclusion, our work provides evidence

736 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 5
Cappelli et al.
of the quinoline nucleus. Finally, the existence of
limited tolerance to large substituents in the areas
surrounding both the fused benzene ring and the
terminal piperazine nitrogen atom of quipazine com-
pletes this model.
Con clu sion s
Starting from the high affinity and low selectivity
shown by quipazine for 5-HT3 receptors, our work has
resulted in the discovery of novel central 5-HT3 receptor
ligands provided with a full range of intrinsic efficacies
which, by virtue of their selectivity and potency, are
interesting candidates as pharmacological tools to evalu-
ate the role(s) of 5-HT3 receptors in the CNS. Moreover,
qualitative and quantitative structure-affinity relation-
ship studies carried out by means of descriptors capable
of capturing the strict ligand-receptor complementarity
criteria (which determine the binding affinity of those
ligands) allowed the role of the main pharmacophoric
components to be elucidated and a model for the
interaction of the 5-HT3 ligands related to quipazine
with their receptor to be developed. The essential
nature of the arylpiperazine interaction mode toward
the receptor can be summarized as follows: (a) a charge-
assisted hydrogen bond, (b) a hydrogen-bonding interac-
tion, and (c) an aromatic specific interaction. Further-
more, short-range interactions in the receptor area
corresponding to the c-edge of the quinoline nucleus of
quipazine seem to be able to modulate the intrinsic
efficacy of these quipazine derivatives toward the
central 5-HT3 receptor. Finally, the existence of limited
bulk tolerance in the areas surrounding the fused
benzene ring and the terminal piperazine nitrogen atom
of quipazine completes this model.
F igu r e 2. Effects of compound 5b on the 5-HT3 receptor-
dependent [14C]guanidinium uptake in NG 108-15 cells: 9,
with 10 µM SP alone; 0, with SP (10 µM) and 5-HT (1 µM); b,
5b and SP (10 µM); O, 5b, SP (10 µM), and 5-HT (1 µM). Assays
were carried out as described in the Experimental Section, and
the accumulation of [14C]guanidinium is expressed as dpm/
culture dish. Each point is the mean of triplicate determina-
tions with less than 5% variation among them.
Exp er im en ta l Section
F igu r e 3. Arylpiperazine interactions at the central 5-HT3
receptors.
Melting points were determined in open capillaries on a
Gallenkamp apparatus and are uncorrected. Microanalyses
were carried out using a Perkin-Elmer 240C elemental ana-
lyzer. Merck silica gel 60 (70-230 or 230-400 mesh) and
Merck aluminum oxide 90 II-III, 70-230 mesh, were used
for column chromatography. Merck TLC plates, silica gel 60
for the key role played by the protonated piperazine
nitrogen atom, the role of the quinoline nitrogen atom
as a hydrogen-bonding acceptor, and the dual role of
the fused benzene ring of quipazine derivatives in the
interaction with the 5-HT3 receptor.
F
254, were used for TLC. 1H NMR spectra were recorded with
a Bruker AC 200 spectrometer in the indicated solvents (TMS
as internal standard); the values of chemical shifts are
expressed in ppm and coupling constants (J ) in Hz. Mass
spectra (EI, 70 eV) were recorded on either a VG 70-250S
spectrometer (Centro di Analisi e Determinazioni Strutturali,
Universita` di Siena) or a Varian Saturn 3 (Dipartimento
Farmaco Chimico Tecnologico, Universita` di Siena). NMR
spectra and elemental analyses were performed in the Dipart-
imento Farmaco Chimico Tecnologico, Universita` di Siena.
6-(4-Meth yl-1-p ip er a zin yl)p h en a n th r id in e (5a ). A mix-
ture of 6-chlorophenanthridine (9a )20 (0.214 g, 0.1 mmol) with
N-methylpiperazine (5 mL) was heated at 130-140 °C under
argon for 24 h. The reaction mixture was then poured into
ice-water (250 mL) and extracted with chloroform (4 × 15
mL). The organic layer was washed with water (2 × 30 mL),
dried over sodium sulfate, and evaporated under reduced
pressure. Purification of the residue by flash chromatography
with ethyl acetate-triethylamine (8:2) as the eluent gave pure
5a as a pale brown oil (0.24 g) which crystallized on standing.
1H NMR (CDCl3): 2.42 (s, 3H), 2.72 (t, J ) 4.8, 4H), 3.55 (t, J
) 4.7, 4H), 7.47 (t, J ) 7.3, 1H), 7.56-7.65 (m, 2H), 7.76 (t, J
) 7.6, 1H), 7.92 (d, J ) 8.6, 1H), 8.19 (d, J ) 8.0, 1H), 8.41 (d,
J ) 7.8, 1H), 8.54 (d, J ) 8.2, 1H).
Accordingly, a three-component model for the interac-
tion of central 5-HT3 ligands related to quipazine with
their receptor is schematically represented in Figure 3.
High-affinity arylpiperazine interaction with the central
5-HT3 receptor probably requires (a) a charge-assisted
hydrogen bond between the protonated terminal pip-
erazine nitrogen atom and a negatively charged car-
boxylic amino acid residue in the receptor, (b) a hydrogen-
bonding interaction between the heterocyclic nitrogen
atom and a suitable H-bond donor in the receptor, and
(c) an aromatic specific interaction between an aromatic
ring and a suitable amino acid residue in the receptor.
Furthermore, attractive short-range (e.g., van der Waals)
interactions in the receptor area corresponding to the
c-edge of the quinoline nucleus of quipazine may be
assumed to take place. These interactions seem to be
of crucial importance in modulating the intrinsic efficacy
at the central 5-HT3 receptor since quipazine (agonist),
5b (partial agonist), and 5c (antagonist) mainly differ
in the presence and the type of substituent on the c-edge

Mapping the Central 5-HT3 Receptor Binding Site
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 5 737
6-(4-Meth yl-1-p ip er a zin yl)-7,8,9,10-tetr a h yd r op h en a n -
th r id in e (5b). This compound was prepared in a similar way
as 5a starting from 6-chloro-7,8,9,10-tetrahydrophenanthri-
dine (9b)21 (0.2 g, 0.92 mmol) and 6 mL of N-methylpiperazine
(reaction time 24 h at 130-140 °C). Compound 5b (0.23 g)
was obtained as a pale-yellow oil which slowly crystallized on
standing. 1H NMR (CDCl3): 1.71-1.82 (m, 2H), 1.91-2.03 (m,
2H), 2.38 (s, 3H), 2.62 (t, J ) 4.7, 4H), 2.77 (t, J ) 5.6, 2H),
3.10 (t, J ) 6.4, 2H), 3.34 (t, J ) 4.8, 4H), 7.31-7.39 (m, 1H),
7.50-7.58 (m, 1H), 7.79-7.86 (m, 2H).
6-Mor ph olin o-7,8,9,10-tetr ah ydr oph en an th r idin e (5ba).
This compound was prepared in a similar way as 5a starting
from 9b21 (0.22 g, 1.01 mmol) and 5 mL of morpholine (reaction
time 17 h at 130-150 °C) and purified by flash chromatogra-
phy with n-hexane-ethyl acetate (8:2) as the eluent. Com-
pound 5ba (0.26 g, yield 96%) was obtained as a colorless oil
which slowly crystallized on standing (mp 76-77 °C). 1H NMR
(CDCl3): 1.74-1.85 (m, 2H), 1.92-2.05 (m, 2H), 2.79 (t, J )
5.6, 2H), 3.12 (t, J ) 6.4, 2H), 3.30 (t, J ) 4.7, 4H), 3.90 (t, J
) 4.5, 4H), 7.38 (t, J ) 7.8, 1H), 7.56 (t, J ) 7.3, 1H), 7.82-
7.88 (m, 2H). Anal. (C17H20N2O) C, H, N.
1-(7,8,9,10-Tetr a h yd r op h en a n th r id in -6-yl)-4,4-d im eth -
ylp ip er a zin iu m Iod id e (5bb). A mixture of 5b (0.282 g, 1.0
mmol) in 20 mL of diethyl ether with CH3I (0.062 mL, 1.0
mmol) was stirred at room temperature for 3 days. The
precipitate was collected by filtration, washed with diethyl
ether (10 mL), and dried under reduced pressure to yield 0.17
g (40%) of white powder (mp 252-253 °C). 1H NMR (CDCl3):
1.73-1.85 (m, 2H), 1.92-2.05 (m, 2H), 2.76 (t, J ) 5.5, 2H),
3.13 (t, J ) 6.1, 2H), 3.71 (s, 10H), 3.93 (t, J ) 4.7, 4H), 7.40-
7.48 (m, 1H), 7.56-7.64 (m, 1H), 7.80-7.89 (m, 2H). Anal.
(C19H26N3I) C, H, N.
6-(1-P ip er a zin yl)-7,8,9,10-t et r a h yd r op h en a n t h r id in e
(5bc). This compound was prepared in a similar way as 5a
starting from 9b21 (0.22 g, 1.01 mmol) and 1.0 g (11.6 mmol)
of anhydrous piperazine with ethylene glycol (10 mL) as the
solvent (reaction time 17 h at 130-140 °C) and purified by
flash chromatography with ethyl acetate-triethylamine-
MeOH (6:2:2) as the eluent. Compound 5bc (0.21 g, yield 78%)
was obtained as a colorless oil which crystallized on standing
(mp 98-100 °C). 1H NMR (CDCl3): 1.72-1.83 (m, 2H), 1.91-
2.04 (m, 2H), 2.19 (br s, 1H), 2.78 (t, J ) 5.4, 2H), 3.07-3.14
(m, 6H), 3.24-3.29 (m, 4H), 7.36 (t, J ) 8.0, 1H), 7.55 (t, J )
7.3, 1H), 7.80-7.87 (m, 2H). Anal. (C17H21N3) C, H, N.
6-(4-E t h yl-1-p ip er a zin yl)-7,8,9,10-t et r a h yd r op h en a n -
th r id in e (5bd ). This compound was prepared in a similar
way as 5a starting from 9b21 (0.22 g, 1.01 mmol) and 4 mL of
N-ethylpiperazine (reaction time 24 h at 130-140 °C). Com-
pound 5bd (0.25 g, yield 84%) was obtained as a pale-yellow
oil which slowly crystallized on standing (mp 66-67 °C). 1H
NMR (CDCl3): 1.15 (t, J ) 7.1, 3H), 1.71-1.82 (m, 2H), 1.92-
2.04 (m, 2H), 2.51 (q, J ) 7.2, 2H), 2.64 (t, J ) 4.6, 4H), 2.77
(t, J ) 5.6, 2H), 3.10 (t, J ) 6.4, 2H), 3.34 (t, J ) 4.8, 4H),
7.35 (t, J ) 7.4, 1H), 7.54 (t, J ) 7.3, 1H), 7.79-7.86 (m, 2H).
Anal. (C19H25N3) C, H, N.
6-(4-Ben zyl-1-p ip er a zin yl)-7,8,9,10-tetr a h yd r op h en a n -
th r id in e (5be). This compound was prepared in a similar
way as 5a starting from 9b21 (0.22 g, 1.01 mmol) and 0.20 mL
(1.1 mmol) of N-benzylpiperazine with ethylene glycol (3 mL)
as the solvent and 0.12 g (1.1 mmol) of Na2CO3 (reaction time
18 h at 130-140 °C) and purified by flash chromatography
with n-hexane-ethyl acetate (1:1) as the eluent. Compound
5be (0.13 g, yield 36%) was obtained as a pale-yellow oil which
crystallized on standing (mp 107-109 °C). 1H NMR (CDCl3):
1.71-1.82 (m, 2H), 1.92-2.03 (m, 2H), 2.65 (t, J ) 4.6, 4H),
2.76 (t, J ) 5.5, 2H), 3.10 (t, J ) 6.3, 2H), 3.32 (t, J ) 4.7,
4H), 3.61 (s, 2H), 7.23-7.40 (m, 6H), 7.54 (t, J ) 7.2, 1H),
7.79-7.86 (m, 2H). Anal. (C24H27N3) C, H, N.
2.08 (m, 4H), 2.43 (s, 3H), 2.72-2.82 (m, 6H), 3.09 (t, J ) 6.4,
2H), 3.59 (t, J ) 6.1, 2H), 3.67 (t, J ) 4.8, 2H), 7.31 (t, J )
7.3, 1H), 7.52 (t, J ) 7.6, 1H), 7.78 (d, J ) 8.7, 2H). Anal.
(C19H25N3) C, H, N.
2-(4-Meth yl-1-piper azin yl)pyr idin e Dim aleate (6a). This
compound was prepared in a similar way as 5a starting from
2-chloropyridine (0.15 mL, 1.6 mmol) and 4 mL of N-methyl-
piperazine (reaction time 21 h at 130-140 °C). The free base
(0.27 g, 1.52 mmol) was dissolved in methanol-CHCl3 (10 mL),
and maleic acid (0.38 g, 3.3 mmol) was added. The resulting
mixture was stirred at room temperature for 15 min and
concentrated under reduced pressure, and the residue was
recrystallized to give 0.43 g of 6a . 1H NMR (CDCl3): 2.87 (s,
3H), 3.19-3.33 (m, 4H), 3.83-3.98 (m, 4H), 6.32 (s, 4H), 6.69-
6.81 (m, 2H), 7.54-7.64 (m, 1H), 8.21-8.24 (m, 1H).
6-(4-Meth yl-1-piper azin yl)-7,8,9,10-tetr ah ydr o-8,10-eth -
a n op h en a n th r id in e (5c). This compound was prepared in
a similar way as 5a starting from 9c (0.098 g, 0.40 mmol) and
5 mL of N-methylpiperazine (reaction time 20 h at 130-140
°C). Compound 5c (0.12 g) was obtained as a pale-yellow oil
which slowly crystallized on standing. 1H NMR (CDCl3):
1.37-1.45 (m, 1H), 1.75-2.05 (m, 5H), 2.37 (s, 3H), 2.51-2.69
(m, 6H), 3.02-3.44 (m, 5H), 3.83 (br s, 1H), 7.33-7.41 (m, 1H),
7.50-7.58 (m, 1H), 7.86 (d, J ) 8.0, 1H), 7.96 (d, J ) 8.2, 1H).
7,8,9,10-Tet r a h yd r o-8,10-et h a n o-6(5H )-p h en a n t h r id i-
n on e (8). To a 100-mL, three-necked flask fitted with a
condenser, thermometer, and dropping funnel (under argon)
were added 7 (1.24 g, 9.98 mmol), freshly distilled anhydrous
ethyl ether (40 mL), and pyrrolidine (3.3 mL, 39.5 mmol). To
the resulting solution was added, over a 20-30-min period,
TiCl4 (0.55 mL, 5.0 mmol) in an additional 10 mL of dry
benzene. The temperature was kept between 0 and 10 °C
during the addition. Once the TiCl4 addition was complete,
the mixture was allowed to stir at room temperature for 19 h.
The reaction mixture was then filtered, and the solvent was
removed under reduced pressure. The residue was dissolved
into 10 mL of CHCl3, and the resulting solution was cooled at
0 °C. To this was added a solution of phenyl isocyanate (1.1
mL, 10.1 mmol) in CHCl3 (10 mL) dropwise. When the
addition was complete, the reaction mixture was stirred for 2
h at room temperature and the solvent was removed under
reduced pressure. Purification of the residue by chromatog-
raphy (n-hexane-ethyl acetate, 65:35, as eluent) gave an oil
which crystallized on standing. To this material was added
PPA (20 g), and the resulting mixture was heated at 150 °C
with stirring for 15 min. Then the cooled brown mass was
decomposed with ice-water, and the precipitate was extracted
with CHCl3 (3 × 20 mL). The combined organic extracts were
washed with water, dried over sodium sulfate, and concen-
trated under reduced pressure. Purification of the residue by
washing with ethyl acetate gave pure 8 (0.48 g, yield 21%) as
a white solid. Recrystallization from ethyl acetate-methanol
gave an analytical sample melting at 248-249 °C. 1H NMR
(CDCl3): 1.49-1.63 (m, 1H), 1.76-2.05 (m, 5H), 2.52-2.67 (m,
2H), 2.86-2.97 (m, 1H), 3.58 (br s, 1H), 7.18-7.29 (m, 2H),
7.44 (t, J ) 7.3, 1H), 7.80 (d, J ) 8.1, 1H), 10.64 (br s, 1H).
Anal. (C15H15NO) C, H, N.
6-Ch lor o-7,8,9,10-t et r a h yd r o-8,10-et h a n op h en a n t h r i-
d in e (9c). A mixture of 8 (0.25 g, 1.11 mmol) and POCl3 (5
mL) was refluxed for 1 h. Then the cooled reaction mixture
was poured into ice-water, and the precipitate was extracted
with CH2Cl2 (3 × 30 mL). The combined extracts were washed
with water, dried over sodium sulfate, and concentrated under
reduced pressure to obtain pure 9c (0.26 g, yield 96%) as a
white solid. Recrystallization from n-hexane gave an analyti-
cal sample melting at 127-128 °C. 1H NMR (CDCl3): 1.49-
1.64 (m, 1H), 1.79-2.14 (m, 5H), 2.70-2.79 (m, 2H), 3.06-
3.18 (m, 1H), 3.87 (br s, 1H), 7.50-7.69 (m, 2H), 7.97 (d, J )
8.3, 1H), 8.05 (d, J ) 8.2, 1H). MS: m/ z 243 (M+, 100). Anal.
(C15H14NCl) C, H, N.
6-(4-Met h yl-1-h om op ip er a zin yl)-7,8,9,10-t et r a h yd r o-
p h en a n th r id in e (5bf). This compound was prepared in a
similar way as 5a starting from 9b21 (0.22 g, 1.01 mmol) and
2 mL of N-methylhomopiperazine (reaction time 24 h at 130-
140 °C). Compound 5bf (0.21 g, yield 70%) was obtained as a
pale-yellow oil. 1H NMR (CDCl3): 1.70-1.81 (m, 2H), 1.90-
1,2,3,4,7,8,9,10-Octah ydr o-6(5H)-ph en an th r idin on e (11).
A mixture of 1023 (0.25 g, 0.91 mmol) in 14 g of PPA was heated
at 140 °C with stirring for 30 min. The cooled reaction mixture
was then decomposed with ice-water, and the resulting

738 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 5
Cappelli et al.
solution was neutralized with solid Na2CO3. The precipitate
was extracted with CHCl3 (4 × 30 mL), and the combined
extracts were washed with water, dried over sodium sulfate,
and concentrated under reduced pressure to give pure 11 as a
white solid (0.154 g, yield 83%) melting at 295-297 °C (lit.23
mp 295-297 °C). 1H NMR (CDCl3): 1.68-1.82 (m, 8H), 2.30-
2.67 (m, 8H), 11.33 (br s, 1H).
7.67 (d, J ) 9.2, 1H), 8.21 (d, J ) 7.36, 1H), 14.15 (br s, 1H).
MS: m/ z 183 (M+, 100). Anal. (C12H9NO‚0.33H2O) C, H, N.
5,6-Dih ydr oben zo[f]qu in olin -3(4H)-on e (13k). This com-
pound was prepared in 68% yield starting from â-tetralone
(12k ). An analytical sample was obtained from ethyl acetate-
methanol as pale-green plates melting at 254-255 °C. 1H
NMR (CDCl3): 2.97 (s, 4H), 6.59 (d, J ) 9.3, 1H), 7.14-7.33
(m, 3H), 7.49 (d, J ) 7.6, 1H), 7.93 (d, J ) 9.7, 1H), 13.66 (br
s, 1H). Anal. (C13H11NO) C, H, N.
Gen er a l P r oced u r e for th e Syn th esis of 4-Meth yl-1-
p ip er a zin yl Der iva tives 5e,g-k . A mixture of the ap-
propriate 2(1H)-pyridone 13e,g-k (1.0 mmol) in CH2Cl2 (20
mL) with anhydrous Na2CO3 (0.26 g, 2.45 mmol) was cooled
at - 60 °C, and trifluoromethanesulfonic anhydride (0.5 mL,
2.97 mmol) was added under an argon atmosphere. The
resulting mixture was stirred at - 60 °C for 10 min and then
at 0 °C for 1 h. The solid was removed by filtration and
washed with CHCl3, and the filtrate was concentrated under
reduced pressure. The residue was diluted with N-methylpip-
erazine (5 mL), and the resulting mixture was heated at 130-
140 °C under argon for a suitable time (1-2 h). The cooled
reaction mixture was poured into ice-water, the precipitate
was extracted with CH2Cl2 (5 × 15 mL), and the combined
extracts were washed with water (2 × 20 mL), dried over
sodium sulfate, and evaporated under reduced pressure.
Chromatographic purification (Al2O3) of the residue with
n-hexane-ethyl acetate (65:35) as the eluent gave analytically
pure title compounds (as free bases).
2-(4-Met h yl-1-p ip er a zin yl)-5,6,7,8-t et r a h yd r oq u in o-
lin e (5e). 1H NMR (CDCl3): 1.69-1.89 (m, 4H), 2.34 (s, 3H),
2.51 (t, J ) 5.2, 4H), 2.62 (t, J ) 5.9, 2H), 2.74 (t, J ) 6.0,
2H), 3.50 (t, J ) 5.1, 4H), 6.43 (d, J ) 8.3, 1H), 7.17 (d, J )
8.4, 1H).
5,6-Dih yd r o-2-(4-m eth yl-1-p ip er a zin yl)ben zo[h ]qu in o-
lin e Ma lea te (5g). The free base (0.19 g, 0.68 mmol) was
dissolved into methanol (10 mL), and maleic acid (0.079 g, 0.68
mmol) was added. The resulting mixture was stirred at room
temperature for 15 min and concentrated under reduced
pressure, and the residue was recrystallized to give crystalline
maleate salt 5g (0.21 g).1H NMR (CDCl3): 2.80-3.09 (m, 9H),
3.38-3.72 (m, 4H), 4.39-4.58 (m, 2H), 6.32 (s, 2H), 6.60 (d, J
) 8.5, 1H), 7.19-7.38 (m, 3H), 7.42 (d, J ) 8.4, 1H), 8.15-
8.20 (m, 1H).
2-(4-Met h yl-1-p ip er a zin yl)-5,6,6a ,7,8,9,10,10a -oct a h y-
d r oben zo[h ]qu in olin e (5h ). 1H NMR (CDCl3): 0.98-1.60
(m, 6H), 1.73-1.94 (m, 4H), 2.14-2.28 (m, 1H), 2.32 (s, 3H),
2.50 (t, J ) 5.0, 4H), 2.64-2.81 (m, 3H), 3.40-3.60 (m, 4H),
6.41 (d, J ) 8.4, 1H), 7.15 (d, J ) 8.3, 1H).
6,7-Dih yd r o-2-(4-m eth yl-1-p ip er a zin yl)-5H-ben zo[6,7]-
cycloh ep ta [1,2-b]p yr id in e (5i). 1H NMR (CDCl3): 2.14-
2.24 (m, 2H), 2.35-2.41 (m, 5H), 2.52-2.59 (m, 6H), 3.61 (t, J
) 5.0, 4H), 6.57 (d, J ) 8.3, 1H), 7.19-7.38 (m, 4H), 7.70-
7.75 (m, 1H).
2-(4-Met h yl-1-p ip er a zin yl)-5H -in d en o[1,2-b]p yr id in e
(5j). 1H NMR (CDCl3): 2.38 (s, 3H), 2.59 (t, J ) 4.9, 4H), 3.68
(t, J ) 4.9, 4H), 3.73 (s, 2H), 6.58 (d, J ) 8.2, 1H), 7.30-7.43
(m, 2H), 7.52 (d, J ) 6.8, 1H), 7.62 (d, J ) 8.6, 1H), 7.97-8.01
(m, 1H).
5,6-Dih yd r o-3-(4-m eth yl-1-p ip er a zin yl)ben zo[f]qu in o-
lin e (5k ). 1H NMR (CDCl3): 2.36 (s, 3H), 2.54 (t, J ) 5.0,
4H), 2.94 (s, 4H), 3.62 (t, J ) 5.0, 4H), 6.59 (d, J ) 8.5, 1H),
7.10-7.29 (m, 3H), 7.56 (d, J ) 7.5, 1H), 7.83 (d, J ) 8.6, 1H).
5,6-Dih yd r o-3-[[(tr iflu or om eth yl)su lfon yl]oxy]ben zo-
[f]qu in olin e (14k ). This compound was isolated as an
intermediate of the synthesis of 5k and purified by flash
chromatography with n-hexane-ethyl acetate (8:2) as the
eluent (colorless thick oil). 1H NMR (CDCl3): 2.96-3.11 (m,
4H), 7.10 (d, J ) 8.4, 1H), 7.28-7.38 (m, 3H), 7.65-7.69 (m,
1H), 8.15 (d, J ) 8.3, 1H). Anal. (C14H10NO3SF3) C, H, N.
6-(1-Ad a m a n t yl)-2-(4-m et h yl-1-p ip er a zin yl)p yr id in e
(6b). 1-Adamantyl methyl ketone (0.45 g, 2.5 mmol) and
methyl propiolate (0.45 mL, 5.0 mmol) were mixed in an
ammonia-saturated methanol solution (30 mL) in a 125-mL
stainless steel vessel, and the reaction mixture was heated at
6-(4-Met h yl-1-p ip er a zin yl)-1,2,3,4,7,8,9,10-oct a h yd r o-
p h en a n th r id in e (5d ). To a mixture of 11 (0.21 g, 1.03 mmol)
and anhydrous Na2CO3 (0.26 g, 2.45 mmol) in CH2Cl2 (20 mL)
cooled to - 60 °C was added trifluoromethanesulfonic anhy-
dride (0.5 mL, 2.97 mmol). The resulting mixture was allowed
to stir at the same temperature for 10 min and at 0 °C for 3 h
and then filtered, and the filtrate was concentrated under
reduced pressure. The residue was diluted with 5 mL of
N-methylpiperazine, and the resulting mixture was heated at
130-140 °C under argon overnight. The reaction mixture was
then poured into ice-water, and the precipitate was extracted
with CH2Cl2 (5 × 10 mL). The combined organic extracts were
washed with brine, dried over sodium sulfate, and evaporated
under reduced pressure. The residue was purified by column
chromatography (Al2O3), and elution with n-hexane-ethyl
acetate (65:35) gave analytically pure 5d (0.25 g) as a colorless
oil which slowly crystallized on standing. 1H NMR (CDCl3):
1.62-1.87 (m, 8H), 2.34 (s, 3H), 2.49-2.64 (m, 10H), 2.78 (br
s, 2H), 3.12 (t, J ) 4.8, 4H).
Gen er a l P r oced u r e for th e Syn th esis of 2(1H)-P yr i-
d on es (13e,g-k ). Commercially available cyclohexanone
(12e), R-tetralone (12g), trans-1-decalone (12h ), 1-benzosub-
erone (12i), 1-indanone (12j), and â-tetralone (12k ) (obtained
from Aldrich) were used as starting material for the synthesis
of the title compounds without further purification. The
appropriate ketone (10 mmol) and methyl propiolate (1.7 mL,
19.1 mmol) were mixed in an ammonia-saturated methanol
solution (50 mL) in a 125-mL stainless steel vessel (Parr model
4754), and the reaction mixture was heated at 150 °C for 10
h. After cooling, the reaction mixture was concentrated under
reduced pressure and the residue flash-chromatographed on
silica gel. Elution with ethyl acetate-methanol (85:15) gave
the title compounds.
5,6,7,8-Tetr a h yd r o-2(1H)-qu in olin on e (13e). This com-
pound was prepared in 55% yield starting from cyclohexanone
(12e). An analytical sample was obtained from ethyl acetate
as colorless prisms melting at 205-206 °C (lit.44 mp 205-206
°C). 1H NMR (CDCl3): 1.68-1.78 (m, 4H), 2.47 (t, J ) 5.0,
2H), 2.67 (t, J ) 5.2, 2H), 6.37 (d, J ) 9.0, 1H), 7.18 (d, J )
9.0, 1H), 12.71 (br s, 1H). Anal. (C9H11NO) C, H, N.
5,6-Dih ydr oben zo[h ]qu in olin -2(1H)-on e (13g). This com-
pound was prepared in 61% yield starting from R-tetralone
(12g). An analytical sample was obtained from ethyl acetate-
methanol as pale-green prisms melting at 204-206 °C. 1H
NMR (CDCl3): 2.64-2.71 (m, 2H), 2.83-2.92 (m, 2H), 6.52 (d,
J ) 9.5, 1H), 7.24-7.44 (m, 4H), 7.86-7.90 (m, 1H), 11.58 (br
s, 1H). Anal. (C13H11NO) C, H, N.
5,6,6a ,7,8,9,10,10a -Octa h yd r oben zo[h ]qu in olin -2(1H)-
on e (13h ). This compound was prepared in 56% yield starting
from trans-1-decalone (12h ). An analytical sample was ob-
tained from ethyl acetate as pale-yellow prisms melting at
200-204 °C. 1H NMR (CDCl3): 1.11-2.04 (m, 11H), 2.16-
2.31 (m, 1H), 2.44-2.63 (m, 2H), 6.36 (d, J ) 9.6, 1H), 7.13-
7.18 (m, 1H), 10.15 (br s, 1H). Anal. (C13H17NO) C, H, N.
6,7-Dih yd r o-5H -b en zo[6,7]cycloh ep t a [1,2-b]p yr id in -
2(1H)-on e (13i). This compound was prepared in 65% yield
starting from 1-benzosuberone (12i). An analytical sample
was obtained from ethyl acetate-methanol as colorless prisms
melting at 204-205 °C. 1H NMR (CDCl3): 2.16-2.28 (m, 4H),
2.57 (t, J ) 6.3, 2H), 6.48 (d, J ) 8.9, 1H), 7.30-7.55 (m, 5H),
11.43 (br s, 1H). Anal. (C14H13NO) C, H, N.
5H -In d en o[1,2-b]p yr id in -2(1H )-on e (13j). This com-
pound was prepared starting from 1-indanone (12j) and further
purified by recrystallization from methanol-CHCl3 (12%
yield). An analytical sample melted at 298 °C dec. 1H NMR
(CDCl3): 3.70 (s, 2H), 6.61 (d, J ) 8.9, 1H), 7.35-7.56 (m, 3H),

Mapping the Central 5-HT3 Receptor Binding Site
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 5 739
150 °C for 11 h. After cooling, the reaction mixture was
concentrated under reduced pressure and the residue flash-
chromatographed with ethyl acetate as the eluent to give a
solid. To a mixture of this solid in CH2Cl2 (30 mL) and
anhydrous Na2CO3 (0.4 g, 3.8 mmol) cooled at -60 °C was
added trifluoromethanesulfonic anhydride (0.75 mL, 4.5 mmol)
under an argon atmosphere. The resulting mixture was
stirred at -60 °C for 10 min and then at 0 °C for 1 h. The
salt was removed by filtration and washed with CH2Cl2, and
the filtrate was concentrated under reduced pressure. The
residue was diluted with N-methylpiperazine (10 mL), and the
resulting mixture was heated at 110-120 °C under argon for
1 h. The cooled reaction mixture was poured into ice-water,
the precipitate was extracted with CHCl3 (4 × 15 mL), and
the combined extracts were washed with brine (2 × 40 mL),
dried over sodium sulfate, and evaporated under reduced
pressure. Chromatographic purification (Al2O3) of the residue
with n-hexane-ethyl acetate (65:35) as the eluent gave
analytically pure 6b (0.26 g). 1H NMR (CDCl3): 1.77 (br s,
6H), 1.96 (br s, 6H), 2.07 (br s, 3H), 2.35 (s, 3H), 2.53 (t, J )
5.1, 4H), 3.58 (t, J ) 5.0, 4H), 6.44 (d, J ) 8.5, 1H), 6.59 (d, J
) 7.3, 1H), 7.38-7.46 (m, 1H).
Ben zo[h ]qu in olin e 1-Oxid e (16). To a solution of benzo-
[h]quinoline (1.8 g, 10.0 mmol) in 40 mL of CHCl3 cooled at
0-5 °C was added a solution of MCPBA (3.0 g, 15.0 mmol) in
60 mL of CHCl3. The resulting mixture was stirred at room
temperature for 4 days, then washed with a 5% solution of
K2CO3 (6 × 60 mL), dried over sodium sulfate, and concen-
trated under reduced pressure. Purification of the residue by
washing with diethyl ether gave pure 16 (1.7 g, yield 87%) as
a crystalline solid melting at 120 °C (lit.27 mp 123 °C).
buffer, 50 mM, pH 7.4, for 5-HT3 receptors) and centrifuged
according to the procedures indicated in the above-cited
references. The pellet obtained was finally resuspended in the
appropriate incubation buffer (Tris HCl, 50 mM, pH 7.4, for
5-HT1A receptors; Tris HCl, 50 mM, pH 7.7, containing 10 µM
pargyline, 4 mM CaCl2, and 0.1% ascorbic acid for 5-HT1B and
5-HT2C receptors; Tris HCl, 50 mM, pH 7.7, for 5-HT2A
receptors; Hepes HCl, 50 mM, pH 7.4, for 5-HT3 receptors; Tris
HCl, 50 mM, pH 7.4, containing 120 mM NaCl and 5 mM KCl
for 5-HT transporter) just before the binding assay. Membrane
preparations were used the day of the experiment or stored
at -80 °C (for 5-HT transporter binding assay).
[3H]-8-OH-DPAT (sa 137 Ci/mmol, for 5-HT1A; NEN) and
[3H]ketanserin (sa 80.9 Ci/mmol, for 5-HT2A; NEN) binding was
assayed in a final incubation volume of 0.5 mL. Tissue
concentrations and [3H]ligand final concentrations were re-
spectively 4 mg of tissue/sample and 0.5 nM, and 8 mg of
tissue/sample and 0.5 nM.
[3H]Mesulergine (sa 86 Ci/mmol, for 5-HT2C; Amersham) and
[3H]-5-HT (sa 30 Ci/mmol, for 5-HT1B; NEN) binding was
assayed in a final incubation volume of 1.0 mL. Tissue
concentration and [3H]ligand final concentration were 23 mg
of tissue/sample and 1.3 nM ([3H]mesulergine binding) and 14
mg of tissue/sample and 2.3 nM ([3H]-5-HT binding).
[3H]Granisetron (sa 81 Ci/mmol, for 5-HT3; NEN) and [3H]-
paroxetine (sa 29.7 Ci/mmol, for 5-HT transporter; NEN)
binding was assayed in final incubation volumes of 1.0 and
2.0 mL, respectively. Tissue concentration and [3H]ligand final
concentration were 20 mg of tissue/sample and 0.5 nM ([3H]-
granisetron binding) and 2 mg of tissue/sample and 0.1 nM
([3H]paroxetine binding).
2-Ch lor oben zo[h ]qu in olin e (17). A mixture of 16 (0.8 g,
4.1 mmol) in 5 mL of POCl3 was refluxed for 30 min. The
cooled reaction mixture was poured into crushed ice, neutral-
ized with concentrated NH4OH solution, and extracted with
CH2Cl2 (3 × 20 mL). The combined extracts were washed with
water, dried over sodium sulfate, and evaporated under
reduced pressure. The residue was flash-chromatographed
with CH2Cl2 as the eluent to give 17 (0.49 g, yield 56%) as the
less-polar fraction. A sample recrystallized from n-hexane
melted at 109 °C (lit.27 mp 111 °C). 1H NMR (CDCl3): 7.51
(d, J ) 8.6, 1H), 7.65-7.93 (m, 5H), 8.12 (d, J ) 8.2, 1H), 9.20-
9.25 (m, 1H). Anal. (C13H8NCl) C, H, N.
4-Ch lor oben zo[h ]qu in olin e (18). Compound 18 was
isolated from the above flash chromatography as the more-
polar fraction (0.11 g, yield 12%) (white crystalline powder
melting at 83-84 °C). 1H NMR (CDCl3): 7.59 (d, J ) 4.7, 1H),
7.71-7.80 (m, 2H), 7.88-7.95 (m, 2H), 8.12 (d, J ) 9.5, 1H),
8.85 (d, J ) 4.8, 1H), 9.25-9.30 (m, 1H). Anal. (C13H8NCl)
C, H, N.
2-(4-Meth yl-1-p ip er a zin yl)ben zo[h ]qu in olin e Ma lea te
(5f). The title compound was prepared in a similar way as
5a starting from 2-chlorobenzo[h]quinoline (17) (0.22 g, 1.0
mmol) and 8 mL of N-methylpiperazine (reaction time 7 h at
120-130 °C). The free base (0.26 g, 0.94 mmol) was dissolved
in methanol (10 mL), and maleic acid (0.12 g, 1.0 mmol) was
added. The resulting mixture was stirred at room temperature
for 15 min and concentrated under reduced pressure, and the
residue was recrystallized to give 5f (0.32 g) as colorless
prisms. 1H NMR (CDCl3): 2.87 (s, 3H), 2.93-3.21 (m, 2H),
3.43-3.95 (m, 4H), 4.45-4.96 (m, 2H), 6.32 (s, 2H), 7.06 (d, J
) 8.9, 1H), 7.56-7.69 (m, 4H), 7.83-7.89 (m, 1H), 8.06 (d, J
) 8.9, 1H), 9.03-9.08 (m, 1H).
The specific binding of the tritiated ligands was defined as
the difference between total binding and nonspecific binding
in the presence of 10 µM 5-HT (for 5-HT1A and 5-HT1B), 3 µM
ketanserin (for 5-HT2A), 1 µM mianserine (for 5-HT2C), 100 µM
5-HT (for 5-HT3), and 1 µM 6-nitroquipazine (for 5-HT
transporter); it represented 80%, 62%, 90%, 70%, 70%, and
75%, respectively, of the total binding.
Incubations were stopped by rapid filtration under vacuum
through Whatman GF/B filters presoaked in assay buffer or
poly(ethylenimine) at 0.1% (for 5-HT3) or 0.5% (for 5-HT
transporter). Filters were immediately rinsed with 12 mL (3
× 4 mL) of ice-cold buffer using a Brandel M-24R cell
harvester, dried, and immersed in vials containing 8 mL of
Ultima Gold MV (or Filter Count) (Packard) for the measure-
ment of trapped radioactivity with a TRI-CARB 1900TR (or
300C) (Packard) liquid scintillation spectrometer, at a counting
efficiency of about 60%. Competition experiments were ana-
lyzed by the “Allfit” program45 to obtain the concentration of
unlabeled drug that caused 50% inhibition of ligand binding
(IC50). Apparent affinity constants (Ki) were derived from the
IC50 values according to the Cheng and Prusoff equation.46 The
Kd (nM) values of the radiolabeled ligands obtained in satura-
tion isotherms were 2.3 for [3H]-8-OH-DPAT binding, 2.5 for
[3H]-5-HT binding, 0.5 for [3H]ketanserin binding, 2.0 for
[3H]mesulergine binding, 0.6 for [3H]granisetron binding, and
0.09 for [3H]paroxetine binding.
Mea su r em en t of [14C]Gu a n id in iu m Up ta k e in NG 108-
15 Cells. This procedure has been described by Emerit et al.9
Briefly, mouse neuroblastoma × rat glioma hybrid cells of the
NG 108-15 clone were grown in Dulbecco’s modified Eagle’s
medium supplemented with the appropriate nutrients9 for 2
days. The cell layer in each culture dish (35 mm) was then
washed twice with 1.5 mL of buffer A (145 mM NaCl, 5.4 mM
KCl, 1.8 mM CaCl2, 1.0 mM MgCl2, 2.0 mM Na2HPO4, 20 mM
glucose, 20 mM HEPES, pH adjusted to 7.4 with NaOH) and
covered with 1 mL of buffer B (same composition as buffer A
except that [NaCl] was reduced to 135 mM and 10 mM
guanidinium was added) containing 0.20-0.25 mCi of
[14C]guanidinium (sa 59 mCi/mmol; Service des Mole´cules
Marque´es at CEA, 91191 Gif-sur-Yvette, France) and, where
indicated, 1 µM 5-HT, 10 µM SP, and/or eight different
concentrations of each drug to be tested. After 10 min at 37
°C, the assay was stopped by aspiration of the medium, and
In Vitr o Bin d in g Assa ys. Binding assays were performed
as described in refs 40 (5-HT1A receptors), 41 (5-HT1B recep-
tors), 42 (5-HT2A receptors), 41 (5-HT2C receptors), 34 (5-HT3
receptors), and 43 (5-HT transporter).
Male Wistar rats (Charles River, Calco, Italy) were killed
by decapitation, and their brains were rapidly dissected into
various areas (rat hippocampus for 5-HT1A, rat striatum for
5-HT1B, rat pre-frontal cortex for 5-HT2A, rat cortex for 5-HT2C
,
rat cortex and hippocampus for 5-HT3, rat forebrain for 5-HT
transporter). Tissues were homogenized (Polytron PTA 10TS)
in ice-cold Tris HCl, 50 mM, at appropriate pH value (or Hepes

740 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 5
Cappelli et al.
et al. reported the synthesis and the 5-HT3 receptor affinity of
this compound; see: Castan, F.; Schambel, P.; Enrici, A.;
Rolland, F.; Bigg, D. C. H. New Arylpiperazine Derivatives with
High Affinity for 5-HT3 Receptor Sites. Med. Chem. Res. 1996,
6, 81-101.
the cell layer was washed three times with 1.5 mL of ice-cold
buffer C (same composition as buffer A except that NaCl was
replaced by choline chloride). The cells were then dissolved
in 0.5 mL of 0.4 M NaOH, and the resulting extracts were
transferred to scintillation vials. The culture dishes were
further rinsed with 0.5 mL of 1 M HCl and then 0.5 mL of 0.4
M NaOH, which were added to the vials. Each mixture (1.5
mL) was supplemented with 10 mL of the scintillation fluid
Aquasol (New England Nuclear, Les Ulis, France) for radio-
activity counting at 50% efficiency. All assays were performed
in triplicate.
Under these conditions, [14C]guanidinium accumulation in
NG 108-15 cells was 4-5 times higher in the presence of both
1 µM 5-HT and 10 µM SP than in their absence (basal uptake).
5-HT3 receptor antagonists (zacopride, ondansetron, tro-
pisetron, etc.) completely prevented the stimulatory effect of
5-HT (with SP) (see ref 9 for details).
(15) Affinity prediction was done using our CoMFA model previously
described in ref 12. Compounds were built up starting from
structure fragments used in our previous study and minimized
to the next local minimum. The same alignment rule as in ref
12 was used for pKi calculation, and other parameters were set
in accordance to those described earlier.
(16) Cappelli, A.; Donati, A.; Anzini, M.; Vomero, S.; De Benedetti,
P. G.; Menziani, M. C.; Langer, T. Molecular Structure and
Dynamics of Some Potent 5-HT3 Receptor Antagonists. Insight
into the Interaction with the Receptor. Bioorg. Med. Chem. 1996,
4, 1255-1269.
(17) Katritzky, A. R.; Lobanov, V. S.; Karelson, M. QSPR: The
Correlation and Quantitative Prediction of Chemical and Physi-
cal Properties from Structure. Chem. Soc. Rev. 1995, 279-287.
(18) (a) De Benedetti, P. G.; Cocchi, M.; Menziani, M. C.; Fanelli, F.
Theoretical Quantitative Size and Shape Activity Selectivity
Analyses of 5-HT1A serotonin and R1-Adrenergic Receptor Ligands.
J . Mol. Struct. (THEOCHEM) 1994, 305, 101-110 and refer-
ences cited therein. (b) Menziani, M. C.; De Benedetti, P. G.;
Karelson, M. Theoretical Descriptors in Quantitative Structure-
Affinity and Selectivity Relationship Study of Potent N4-
Substituted Arylpiperazine 5-HT1A Receptor Antagonists. Bioorg.
Med. Chem., in press.
(19) Ten Hoeve, W.; Kruse, C. G.; Luteyn, J . M.; Thiecke, J . R. G.;
Wynberg, H. Direct Substitution of Aromatic Ethers by Lithium
Amides. A New Aromatic Amination Reaction. J . Org. Chem.
1993, 58, 5101-5106.
(20) Arcus, C. L.; Coombs, M. M.; Evans, J . V. Reactions of Organic
Azides. Part V. The Schmidt Reaction with Fluorenones. The
Structure of the Intermediate in the Ketonic Schmidt Reaction.
J . Chem. Soc. 1956, 1498-1506.
(21) Bose, A. K.; Manhas, M. S.; Rao, V. V.; Chen, C. T.; Trehan, I.
R.; Sharma, S. D.; Amin, S. G. Heterocyclic Compounds IV.
Synthesis of Some Mono- and Diazaphenanthrene Derivatives.
J . Heterocycl. Chem. 1971, 8, 1091-1094.
(22) White, W. A.; Weingarten, J . J . A Versatile New Enammine
Synthesis. J . Org. Chem. 1967, 32, 213-214.
Ack n ow led gm en t . Thanks are due to Italian
MURST (40% and 60% of funds) and to French INSERM
and DRET for financial support. M.C.M. and P.G.D.B.
are especially grateful to Prof. Mati Karelson for the
CODESSA program. Prof. Stefania D’Agata D’Ottavi’s
careful reading of the manuscript is also acknowledged.
Refer en ces
(1) Hartig, P. R. Molecular Pharmacology of Serotonin Receptors.
Experientia 1994, 71P (Suppl.), 93-102.
(2) Kilpatrick, G. J .; Bunce, K. T.; Tyers, M. B. 5-HT3 Receptors.
Med. Res. Rev. 1990, 10, 441-475 and references therein.
(3) Van Wijngaarden, I.; Hamminga, D.; Van Hes, R.; Standaar, P.
J .; Tipker, J .; Tulp, M. Th. M.; Mol, F.; Olivier, B.; De J onge, A.
Development of High-Affinity 5-HT3 Receptor Antagonists.
Structure Affinity Relationships of Novel 1,7-Annelated Indole
Derivatives. 1. J . Med. Chem. 1993, 36, 3693-3699 and refer-
ences therein.
(4) Costall, B.; Naylor, R. J .; Tyers, M. B. The Psychopharmacology
of 5-HT3 Receptors. Pharmacol. Ther. 1990, 47, 181-202.
(5) Blandina, P.; Goldfarb, J .; Green, J . P. Activation of a 5-HT3
Receptor Releases Dopamine from Rat Striatal Slices. Eur. J .
Pharmacol. 1988, 155, 349-350.
(6) Paudice, P.; Raiteri, M. Cholecystokinin Release Mediated by
5-HT3 Receptors in Rat Cerebral Cortex and Nucleus Accum-
bens. Br. J . Pharmacol. 1991, 103, 1790-1794.
(23) Rigby, J . H.; Balasubramanian, N. Preparation of Highly
Substituted 2-Pyridones by Reaction of Vinyl Isocyanates and
Enamines. J . Org. Chem. 1989, 54, 224-228.
(24) Kozikowski, A. P.; Reddy, E. R.; Miller, C. P. A Simplified Route
to a Key Intermediate in the Synthesis of the Chinese Nootropic
Agent Huperzine A. J . Chem. Soc., Perkin Trans. 1 1990, 195-
197.
(7) Barnes, J . M.; Barnes, N. M.; Costall, B.; Naylor, R. J .; Tyers,
M. B. 5-HT3 Receptors Mediate Inhibition of Acetylcholine
Release in Cortical Tissue. Nature 1989, 338, 762-763.
(8) (a) Ireland, S. J .; Tyers, M. B. Pharmacological Characterization
of 5-Hydroxytryptamine-Induced Depolarization of the Rat
Isolated Vagus Nerve. Br. J . Pharmacol. 1987, 90, 229-238. (b)
Round, A.; Wallis, D. I. Further Studies on the Blockade of 5-HT
Depolarizations of Rabbit Vagal Afferent and Sympathetic
Ganglion Cells by MDL 72222 and Other Antagonists. Neuro-
pharmacology 1987, 26, 39-48.
(9) Emerit, M. B.; Riad, M.; Fattaccini, C. M.; Hamon, M. Charac-
teristics of [14C]Guanidinium Accumulation in NG 108-15 Cells
Exposed to Serotonin 5-HT3 Receptor Ligands and Substance
P. J . Neurochem. 1993, 60, 2059-2067.
(10) Bolanos, F. J .; Schechter, L. E.; Miquel, M. C.; Emerit, M. B.;
Rumigny, J . F.; Hamon, M.; Gozlan, H. Common Pharmacologi-
cal and Physicochemical Properties of 5-HT3 Binding Sites in
the Rat Cerebral Cortex and NG 108-15 Clonal Cells. Biochem.
Pharmacol. 1990, 40, 1541-1550.
(11) Campiani, G.; Cappelli, A.; Nacci, V.; Anzini, M.; Vomero, S.;
Hamon, M.; Cagnotto, A.; Fracasso, C.; Uboldi, C.; Caccia, S.;
Consolo, S.; Mennini, T. Novel and Highly Potent 5-HT3 Receptor
Agonists Based on a Pyrroloquinoxaline Structure. J . Med.
Chem. 1997, 40, 3670-3678.
(12) Anzini, M.; Cappelli, A.; Vomero, S.; Giorgi, G.; Langer, T.;
Hamon, M.; Merahi, N.; Emerit, B. M.; Cagnotto, A.; Skorupska,
M.; Mennini, T.; Pinto, J . C. Novel, Potent, and Selective 5-HT3
Receptor Antagonists Based on the Arylpiperazine Skeleton:
Synthesis, Structure, Biological Activity, and Comparative Mo-
lecular Field Analysis Studies. J . Med. Chem. 1995, 38, 2692-
2704.
(13) Compound 5a was reported in a patent by Barber et al. to exhibit
antidepressant activity: Barber, H. J .; J ones, D. H. 6-(1-
Piperazinyl)phenanthridines. Ger. Patent 1,802,639, 1969; Chem.
Abstr. 1969, 71, 49979f.
(25) J anin, Y. L.; Bisagni, E. Synthesis of Some Benzo[h]quinoline
Derivatives. J . Heterocycl. Chem. 1993, 30, 1129-1131.
(26) The imidoyl triflate 14k was isolated and characterized (see
Experimental Section).
(27) Iwai, I. Polarization of Aromatic Heterocyclic Compounds. XCIV.
Benzo[h]quinoline and its Derivatives. J . Pharm. Soc. J pn. 1951,
71, 1288-1291; Chem. Abstr. 1952, 46, 5587g.
(28) Cocchi, M.; Fanelli, F.; Menziani, M. C.; De Benedetti, P. G.
Conformational Analysis and Theoretical Quantitative Size and
Shape-Affinity Relationships of N4-Protonated N1-Arylpiper-
azine 5-HT1A Serotoninergic Ligands. J . Mol. Struct.
(THEOCHEM) 1997, 397, 129-145.
(29) QUANTA/CHARMm, 1995, Molecular Simulations Inc., 16 New
England Executive Park, Burlington, MA 01803-5297.
(30) Dewar, M. J . S.; Zoebisch, E. G.; Healey, E. F.; Stewart, J . J . P.
AM1: A New General Purpose Quantum Mechanical Molecular
Model. J . Am. Chem. Soc. 1985, 107, 3902-3909.
(31) Katritzky, A. R.; Mu, L.; Lobanov, V. S.; Karelson, M. Correlation
of Boiling Points with Molecular Structure. 1. A Training Set of
298 Diverse Organics and a Test Set of 9 Simple Inorganics. J .
Phys. Chem. 1996, 100, 10400-10407.
(32) Karelson, M.; Lobanov, V. S.; Katritzky, A. R. Quantum-
Chemical Descriptors in QSAR/QSPR Studies. Chem. Rev. 1996,
96, 1027-1043 and references cited therein.
(33) Stanton, D. T.; J urs, P. C. Development and Use of Charged
Partial Surface Area Structural Descriptors in Computer-
Assisted Quantitative Structure-Property Relationship Studies.
Anal. Chem. 1990, 62, 2323-2329.
(34) Nelson, D. R.; Thomas, D. R. [3H]-BRL 43694 (Granisetron) a
Specific Ligand for 5-HT3 Binding Sites in Rat Brain Cortical
Membranes. Biochem. Pharmacol. 1989, 38, 1693-1695.
(35) Dukat, M.; Abdel-Rahman, A. A.; Ismaiel, A. M.; Ingher, S.;
Teitler, M.; Gyermek, L.; Glennon, R. A. Structure-Activity
Relationships for the Binding of Arylpiperazines and Arylbigu-
anides at 5-HT3 Serotonin Receptors. J . Med. Chem. 1996, 39,
4017-4026.
(14) Compound 5b was included in a grant application (Programma
Galileo 1995) which we presented in October 1994 to the
Evaluation Committee. While our work was in progress, Castan

Mapping the Central 5-HT3 Receptor Binding Site
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 5 741
(36) Rault, S.; Lancelot, J .-C.; Prunier, H.; Robba, M.; Renard, P.;
Delagrange, P.; Pfeiffer, B.; Caignard, D.-H.; Guardiola-
Lemaitre, B.; Hamon, M. Novel Selective and Partial Agonists
of 5-HT3 Receptors. Part 1. Synthesis and Biological Evaluation
of Piperazinopyrrolothienopyrazines. J . Med. Chem. 1996, 39,
2068-2080.
(40) Hall, M. D.; El Mestikawy, S.; Emerit, M. B.; Pichat, L.; Hamon,
M.; Gozlan, M. [3H]8-Hydroxy-2-(di-n-propylamino)tetralin Bind-
ing to Pre- and Postsynaptic 5-Hydroxytryptamine Sites in
Various Regions of the Rat Brain. J . Neurochem. 1985, 44,
1685-1696.
(41) Peroutka, S. J . Pharmacological Differentiation and Character-
ization of 5-HT1A, 5-HT1B, and 5-HT1C Binding Sites in Rat
Frontal Cortex. J . Neurochem. 1986, 47, 529-540.
(42) Leysen, J . E.; Niemegeers, C. J . E.; Van Neuten, J . M.; Laduron,
P. M. [3H]Ketanserin (R41468), a Selective 3H-Ligand for Sero-
tonin2 Receptor Sites. Mol. Pharmacol. 1982, 21, 301-314.
(43) Plenge, P.; Mellerup, E. T.; Nielsen, M. Inhibitory and Regula-
tory Binding Sites on the Rat Brain Serotonin Transporter:
Molecular Weight of the [3H]Paroxetine and [3H]Citalopram
Binding Proteins. Eur. J . Pharmacol. 1990, 189, 129-134.
(44) Meyers, A. I.; Garcia-Munoz, G. The Synthesis of 2-Piridones
from Cyclic Cyano Ketones. A New Aromatization Procedure for
Dihydro-2-pyridones. J . Org. Chem. 1964, 29, 1435-1438.
(45) De Lean, K. W.; Munson, P. J .; Rodbard, D. Simultaneous
Analysis of Families of Sigmoidal Curves: Application to Bio-
assay, Radioligand Assay and Physiological Dose-Response
Curves. Am. J . Physiol. 1978, 235, E97-E102.
(37) A rough estimate of the free energy contribution of functional
group X to the binding process can be calculated by the
equation: ∆Gbind(X) ) -RT ln[Ki(compd A)/Ki(compd B)] (1),
derived from ∆Gbind ) -RT ln Ki (2). In eq 1, R is the gas
constant, T is the absolute temperature (298 K), and Ki(compd
A)/Ki(compd B) is the ratio between the Ki values of two
compounds differing only in the considered functional group.
Equation 1 computes 1.36 kcal/mol for each order of magnitude
in the difference in the binding affinity expressed as Ki values.
We are aware that this kind of approach suffers from some
limitations (e.g., it assumes that one of the two functional groups
compared is inert toward the receptor and no significant
difference is present in the conformational and molecular
dynamics properties of the compounds; this indeed may appear
an oversimplifying assumption); therefore the free energy con-
tributions estimated with this method are used only for com-
parative purposes. See: Dietrich, S. W.; Bolger, M. B.; Kollman,
P. A.; J orgensen, E C. Thyroxine Analogues. 23. Quantitative
Structure-Activity Correlation Studies of in Vivo and in Vitro
Thyromimetic Activities. J . Med. Chem. 1977, 20, 863-880.
(38) Burley, S. K.; Petsko, G. A. Weakly Polar Interaction in Protein.
Adv. Protein Chem. 1988, 39, 125-189.
(46) Cheng, Y.; Prusoff, W. H. Relationship between the Inhibition
Constant (Ki) and the Concentration of Inhibitor Which Causes
50 Per Cent Inhibition (IC50) of an Enzymatic Reaction. Biochem.
Pharmacol. 1973, 22, 3099-3108.
(39) Anzini, M.; Cappelli, A.; Vomero, S.; Cagnotto, A.; Skorupska,
M. 6-(1-Piperazinyl)-7H-indeno[2,1-c]quinoline Derivatives with
High Affinity and Selectivity for 5-HT3 Serotonin Sites. Med.
Chem. Res. 1993, 3, 44-51.
J M970645I