Conformational analysis of tandospirone in aqueous solution: Lead evolution of potent dopamine D4 receptor ligands

Article abstract of DOI:10.1016/S0960-894X(01)00167-6

The significant contribution of folded conformation (2) of the anxiolytic tandospirone (1) in aqueous solution was verified by dynamic ¹H NMR. A structurally rigid mimic of 2 was designed and synthesized to evaluate the implication of 2 towards neuroleptic receptor binding. The designed structures provided a new rigid scaffold for dopamine D₄ ligands.

Full text of DOI:10.1016/S0960-894X(01)00167-6

Bioorganic & Medicinal Chemistry Letters 11 (2001) 1141–1144
Conformational Analysis of Tandospirone in Aqueous Solution:
Lead Evolution of Potent Dopamine D4 Receptor Ligands
Tamiki Nishimura, Jun-etsu Igarashi and Makoto Sunagawa*
Sumitomo Pharmaceuticals Research Division, 3-1-98 Kasugadenaka, Konohanaku, Osaka 554-0022, Japan
Received 22 December 2000;accepted 26 February 2001
Abstract—The significant contribution of folded conformation (2) of the anxiolytic tandospirone (1) in aqueous solution was veri-
1
fied by dynamic H NMR. A structurally rigid mimic of 2 was designed and synthesized to evaluate the implication of 2 towards
neuroleptic receptor binding. The designed structures provided a new rigid scaffold for dopamine D4 ligands. # 2001 Elsevier
Science Ltd. All rights reserved.
Introduction
and the importance of these conformations for bioactivity
have been shown in both cases.
Tandospirone (1, Fig. 1),1 developed as an anxiolytic
drug, is an aryl-piperazine compound that binds to both
5-HT1A and dopamine D4 receptors.2 Unlike buspirone,
it has weak binding to the dopamine D2 receptor3 and
binds to neither the benzodiazepine nor GABA receptor.4
Conformational changes of 1 were further examined in
the present study by means of dynamic 1H NMR spectro-
scopy to estimate the activation energy of the con-
formational change. The result of the dynamic NMR
prompted us to design structurally simulated rigid ana-
logues of the folded conformation (2), to study the
contribution of 2 to the anxiolytic activity of 1.
The pharmacological profile of anxiolytic drugs is con-
sidered to be dependent on selectivity in receptor bind-
ing, for which molecular modeling studies have
commonly described an aromatic ring binding site, a
nitrogen binding site and a hydrophobic accessory
binding site as primary binding sites.5
Kinetics
1
Dynamic H NMR studies of 1 were conducted in buf-
fer solution.9 For a kinetic run, a sample was placed in a
Arrangement of those parts is different among receptor
types and it is difficult to identify those relative
geometries for the aryl-piperazines because of structural
flexibility.
1
pre-calibrated H NMR probe that was preset to a spe-
cific temperature. After thermal equilibration of the
1
sample, the H NMR spectra were recorded at different
temperatures ranging from 25 to 50 ꢀC.
In a previous paper,6 we studied a conformational ana-
1
lysis of 1 in aqueous media using H NMR together
The resonances of the alkyl chain part of 1 should be
affected by the temperature due to conformational
with molecular modeling and showed that there is a
significant contribution of a folded conformer (2, Fig. 1)
of 1 which is stabilized by hydrophobic interactions
between the bicyclo ring moiety and the pyrimidine ring
in aqueous media.
Folded conformations in aqueous solution have also
been reported for Taxol7 and substance P antagonists8
*Corresponding author. Tel.: +81-6-6466-5282;fax: +81-6-6466-
5483;e-mail: sunagawa@sumitomopharm.co.jp
Figure 1. Chemical structure of tandospirone (1) (left) and its folded
conformation (2) (right).
0960-894X/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(01)00167-6

1142
T. Nishimura et al. / Bioorg. Med. Chem. Lett. 11 (2001) 1141–1144
changes if this rate is slow enough in the NMR time
scale. The spectral lines of protons b are broad and
complex at 25 ꢀC as shown in Figure 2. When the tem-
perature was increased gradually, the signal pattern of b
became simpler. The chemical shift difference between
the two protons at b were found to be different by more
than 0.005 ppm at 25 ꢀC;however, the resonances were
essentially the same at 50 ꢀC. This change of the pattern
by temperature suggested two protons at the b position
became nearly equivalent when conformation of 1
became flexible from 2 to an extended conformation.
Protons at a, c, and d of 1 were hardly analyzed in a
variable temperature because of multiplicity and over-
The structural search was carried out using MDDR-
3D11 database and Cambridge Crystallographic Data-
base12 to find 234 and 32 compounds, respectively.
Highly flexible compounds with more than three rota-
table bonds were eliminated and remaining compounds
were categorized to 18 groups based on similarity. Then,
each of the groups were manually evaluated for both
geometric and steric fitness to conformer 2, and mezila-
mine (3), a known dopamine D2 ligand13 was selected as
a preliminary lead structure.
The target compound to be synthesized (4) was designed
by replacing the methyl group attached to the sulfur
atom with a phenyl group because a molecular model-
ing study showed the methyl group is somewhat small
to mimic the bulky bicyclo ring moiety of 2. The X-ray
crystallographic data12 and modeling showed that the
phenyl group of 4 attached to the pyrimidine ring via
the sulfur atom is oriented perpendicular to the plane of
a pyrimidine ring (Fig. 4).
1
lapping. Because of the complex pattern of H NMR
resonances, we obtained only an approximate activation
energy of the conformational change, 6–7 kcal/mol.10
This observation confirmed the considerable contri-
bution of 2 in the conformational space of 1 in aqueous
solution.
Related analogues (16, 21a and 21b) were synthesized as
shown in Scheme 2 varying the bulkiness of each
hydrophobic moiety in 4 to examine further changes in
receptor affinities.
Design of Conformational Mimics
As a practical method to study how a specific con-
formation of a compound contributes to its biological
activities, we performed a structural search using a
query which fills our hypothesis for designing con-
formational mimics. Our assumption consists of three
critical points that are an aromatic ring, a basic nitro-
gen, and a hydrophobic region in relation to a pharma-
cophore of 1 as shown in Figure 3.
Chemistry
The designed compound (4) was synthesized as shown
in Scheme 1. Methylphenylsulfonium ylide (5) was pre-
pared by the reaction of barbituric acid with methyl-
phenyl sulfoxide in the presence of acetic anhydride.
The methyl group was removed by treatment of 5 with
POCl3 to yield 6. In the case of the related diphenyl
sulfonium ylide compound, no reaction occurred and
the starting ylide was recovered quantitatively. The
methyl group at the sulfur atom of 5 was eliminated as
An aromatic ring and a basic nitrogen in the structure
of 1 forming an aryl-piperazine moiety often regarded
as a mimic structure of 5-hydroxy-tryptamine (5-HT),
while a hydrophobic part is needed for higher 5-HT1A
affinity.5 The acceptable distances in the query were
acquired from the structure of the folded conformation (2).
1
Figure 2. Expanded regions of H NMR spectra of tandospirone (protons b and e) at variable temperatures.
Figure 3. Query for 3D database search. A, B, and C denote carbon or
nitrogen, carbon, and hydrophobic moiety, respectively.
Figure 4. Structures of mezilamine (3) and designed compounds (4)
with 3D structure from molecular modeling.

T. Nishimura et al. / Bioorg. Med. Chem. Lett. 11 (2001) 1141–1144
1143
Scheme 1. Reagents, conditions, and yields: (a) methylphenylsulfoxide, Ac2O/AcOH, 90 ꢀC, 2 h, quantitative yield;(b) POCl 3/dimethyl aniline,
reflux, 24 h, quantitative yield;(c) 40% MeNH 2 aq/methylethyl ketone, 5 ꢀC, 2 h, then ambient temperature, 80% for 6 and 10% for 7;(d) N-methyl-
piperazine, EtOH, reflux, 3 h, 60%.
Scheme 2. Reagents, conditions and yields: (a) Br2/AcOH, rt, overnight, 40%;(b) thiophenol, NaH/DMF, reflux, 8 h, 95%;(c) POCl
reflux, 2 h,
3,
40%;(d) Br 2/AcOH, rt, overnight, 84%;(e) thiophenol, NaH/DMF, rt, 4 h, 91%;(f) POCl 3, reflux, 2 h, 97%;(g) N-methylpiperazine, EtOH, reflux,
24 h, 22%;(h) Br 2/AcOH, rt, overnight, 40%;(i) p-chlorothiophenol, NaH/DMF, reflux, 8 h, 95%;(j) POCl 3, reflux, 2 h, 20%;(k) for 21a: N-
methylpiperazine, 2-butanone, K2CO3, H2O, reflux, 1 h, 55%;for 21b: 4-amino-1-benzylpiperidine, 2-butanone, K2CO3, H2O, reflux, 1 h, 56%.
methyl chloride, whereas removal of phenyl group
might be difficult by the same reaction.14 The reaction
of a primary amine with trichlorinated molecule (6)
gave a mixture of isomers bearing the amino substituent
in either position 2 (7) or 4 (8). These isomers can be
separated easily by silica gel column chromatography.
Major product was found to be the product substituted
in position 2. 2-Methylamino-4,6-dichloro-5-phenylthio
pyrimidine (7) was condensed with an N-methyl piper-
azine to yield 4. The other isomer (8) was treated with
N-methylpiperazine to give a condensed compound,
which was confirmed to show only weak affinity
(Ki>10À6) for the 5-HT1A and dopamine receptors. In
both cases, di-N-methylpiperazine substituted derivatives
were not isolated under the reaction condition.
different pyrimidine carbons, while isomer (7) has two
equivalent pyrimidine carbons. Carbons 4 and 6 of the
pyrimidine ring should be equivalent for 7 and non-
equivalent for 8. Thus, each isomer could be identified.
Compound (7) was also synthesized as shown in Scheme
2. Pyrimidine (9) was prepared by the condensation of
diethyl malonate and N-methyl guanidine. Bromination
of 9 followed by substitution by thiophenol gave 11,
which was chlorinated with POCl3 to give 7. Related
compounds (16, 21a, and 21b) were also synthesized in a
similar manner as shown in Scheme 2.
Results and Discussion
Designed compounds were evaluated in a receptor
binding assay for D2, D4, and 5-HT1A binding affinities
as shown in Table 2.
The structure of each isomer (7 or 8) was determined by
13C NMR spectroscopy. Comparison of the 13C chemi-
cal shifts in pyrimidines, 6, 7, and 8 permits their
identification (Table 1).
Table 2. Receptor binding assay15
Pyrimidine and trichlorinated compound (6) have two
equivalent pyrimidine carbons. The isomer (8) has four
Ki (nM)
Compound
D2
D4
5-HT1A
Table 1. 13C NMR chemical shifts of pyrimidines
1
3
4
16
21a
21b
460a
5.4
188
31
62
62
7.2a
>10,000
250
92
32a
430
Position of C
Pyrimidine
6
7
8
460
2000
1500
160
C(2)
C(4)
C(5)
C(6)
158.8
156.4
121.4
156.4
153.5
167.4
128.3
167.4
167.2
160.7
112.1
160.7
161.7
165.2
125.5
164.5
260
2.3
aSee ref 2.

1144
T. Nishimura et al. / Bioorg. Med. Chem. Lett. 11 (2001) 1141–1144
Compound 3, originally reported as an anti-psychotic
drug,13 showed strong D2 affinity and modest 5-HT1A
affinity. However, unlike tandospirone (1), 3 had no D4
affinity.16 Compound 4 designed from the folded con-
formation (2) showed D2, D4, and 5-HT1A affinities.
Compound 4 has a bulky and hydrophobic S-Ph group
as an equivalent of the hydrophobic bicyclo group of 2,
and 3 has a S-Me group in the same position. When the
binding profile of 4 was compared to that of 3, sig-
nificant enhancement of D4 affinity was observed in 4,
whereas D2 affinity of 4 was decreased. In contrast to
this, no significant change was observed in 5-HT1A affi-
nity between 3 and 4. That is, introduction of a bulky
and hydrophobic moiety in 4 contributed to the dopa-
mine D4 affinity. These observations suggested that the
folded conformation (2) of 1 was responsible for the
dopamine D4 affinity.
References and Notes
1. Ishizumi, K.;Kojima, A.;Antoku, F. Chem. Pharm. Bull.
1991, 39, 2288.
2. Ohno, Y.;Kojima, A.;Wakabayashi, J.;Tagashira, R. WO
9837893.
3. (a) Shimizu, H.;Hirose, A.;Tatsuno, T.;Nakamura, M.;
Junki, K. Jpn. J. Pharmacol. 1987, 45, 493. (b) Tatsuno, T.;
Shimizu, H.;Hirose, A.;Tanaka, H.;Kumasaka, Y.;Naka-
mura, M. Pharmacol., Biochem. Behav. 1989, 32, 1049.
4. Shimizu, H.;Karai, N.;Hirose, A.;Tatsuno, T.;Tanaka, H.;
Kumasaka, Y.;Nakamura, M. Jpn. J. Pharmacol. 1988, 46, 311.
5. (a) Hibert, M. F. Actual Chim. Ther. 1989, 16, 37. (b)
Froimowiz, M.;Ra msby, S. J. Med. Chem. 1991, 34, 1707. (c)
Raghupathi, R. K.;Rydelek-Fitzgerald, L.;Teitler, M.;Glen-
non, R. A. J. Med. Chem. 1991, 34, 2633. (d) van Steen, B. J.;
van Wijngaarden, I.;Tulp, M. Th. M.;Soudijin, W. J. Med.
Chem. 1993, 36, 2751.
6. Igarashi, J.;Nishimura, T.;Sunagawa, M. Bioorg. Med.
Chem. Lett. 1997, 7, 1659.
In addition to this, related compounds (16, 21a, and
21b) of 4 also had affinity to the D4 receptor as well as
the D2 and 5-HT1A receptors. Among them, compound
21b showed the strongest D4 affinity and the D4/D2
selectivity was also observed. Several dopamine D4
selective ligands have been reported,17 but most of them
have flexible structures. Compound 4 was considered to
be a new scaffold for structurally rigid dopamine D4
selective ligands.
7. Gao, Q.;Parker, W. L. Tetrahedron 1996, 52, 2291.
8. Desai, M. C.;Vincent, L. A.;Rizzi, J. P. J. Med. Chem.
1994, 37, 4263.
9. A sample was dissolved in 0.6 mL of phosphate buffer
solution (50 mM sodium phosphate, pH 7.4). 1H NMR spectra
were recorded on JEOL A 500 spectrometer. The chemical
shifts (in ppm) were referenced to 2-trimethylsilyl-2,2,3,3-tetra-
deuteriopropionic acid, sodium salt as an internal standard.
The data sets were processed using EDL (JEOL Inc., Tokyo,
Japan) on a VAX 3200 workstation.
10. Kolocouris, A.;Mikros, E.;Kolocouris, N. J. Chem. Soc.,
Perkin Trans. 2 1998, 1701.
11. MDL Drug Data Report-3D;MDL Information Systems,
14600 Catalina Street, San Leandro, CA, USA.
12. Allen, F. H.;Davies, J. E.;Galloy, J. J.;Johnson, O.;
Kennard, O.;Maccrae, C. F.;Mitchell, G. K.;Smith, J. M.;
Watson, D. G. J. Chem. Inf. Comput. Sci. 1991, 31, 187.
13. Gueremym, C.;Audiau, F.;Uzan, A.;Le Fue, G.;Leger,
J. M.;Carpy, A. J. Med. Chem. 1982, 25, 1459.
14. Furukawa, S.;Igarashi, J.;Watanabe, M.;Kinoshita, T.
Chem. Pharm. Bull. 1990, 38, 227.
15. MDS Panlabs, 11804 North Creek Parkway South,
Bothell, WA, USA.
16. 36% inhibition of [3H]-Spiperon at 10À5 M.
17. Hrib, N. J. Drug Future 2000, 25, 587.
Summary
The rigid compound (4) designed from 2, the folded
conformation of 1 in aqueous solution, revealed that
this conformation is mainly responsible for the dopa-
mine D4 affinity.
It was shown that conformational analysis of flexible
drugs in aqueous solvent using NMR measurements
and a three-dimensional database search could provide
an additional way toward lead evolution processes in
drug discovery.