Regulatory molecules for the 5-HT3 receptor ion channel gating system

Article abstract of DOI:10.1016/j.bmc.2007.02.054

Substituted benzoxazole derivatives which possess a nitrogen-containing heterocycle at C2 are selective partial agonists of the 5-HT₃ receptor. Alteration of substituents on the benzoxazole nucleus affords both agonist-like and antagonist-like compounds, and uniquely modifies the function of the 5-HT₃ receptor ion channel gating system. SAR and corroborative computational docking study for these partial agonists successfully explained structure and function of the 5-HT₃ receptor.

Full text of DOI:10.1016/j.bmc.2007.02.054

Bioorganic & Medicinal Chemistry 15 (2007) 3515–3523
Regulatory molecules for the 5-HT3 receptor
ion channel gating system
Satoshi Yoshida,* Takashi Watanabe and Yasuo Sato*
Pharmaceutical Research Department, Meiji Seika Kaisha Ltd, 760, Morooka-Cho, Kohoku-ku, Yokohama 222-8567, Japan
Received 4 January 2007; revised 23 February 2007; accepted 27 February 2007
Available online 7 March 2007
Abstract—Substituted benzoxazole derivatives which possess a nitrogen-containing heterocycle at C2 are selective partial agonists of
the 5-HT3 receptor. Alteration of substituents on the benzoxazole nucleus affords both agonist-like and antagonist-like compounds,
and uniquely modifies the function of the 5-HT3 receptor ion channel gating system. SAR and corroborative computational docking
study for these partial agonists successfully explained structure and function of the 5-HT3 receptor.
ꢀ 2007 Elsevier Ltd. All rights reserved.
1. Introduction
binding protein (AChBP) possesses high amino acid se-
quence homology with LGIC subunits, and its 3D struc-
ture has been determined at high-resolution from
crystallographic data.
The serotonin type 3 (5-HT3) receptor is one of the li-
gand-gated ion channel (LGIC) receptors that mediate
critical synaptic transmission in both the peripheral
and central nervous systems.1 The nicotinic acetylcho-
line receptor (nACh), GABAA receptor, and ionotropic
glycine receptor are classified into this family, and func-
tional failure of these receptors causes a variety of
chronic diseases. For decades, small-molecular ligands
which interact with and modulate the functions of LGIC
receptors have been recognized as candidate therapeutic
agents. For example, 5-HT3 receptor antagonists, such
as granisetron (3, Figure 1) and alosetron (4, Figure
1), have widely been used for the treatment of chemo-
therapy-induced emesis and irritable bowel syndrome
(IBS).2–4 However, the structures and functions of
LGIC receptors remain obscure. Because the crystal
structure of LGIC receptors has not been determined,
the following two approaches have mainly been used
to investigate their structure and function. (1) Struc-
ture–activity relationship (SAR) studies of LGIC recep-
tor ligands:5 SAR data are particularly valuable to
investigate receptor functions, which are difficult to pre-
dict even from crystallographic data. (2) Modeling based
on homology with related proteins:6,7 the acetylcholine
We have reported that the benzoxazole derivatives 1
(Figure 1), which possess a nitrogen-containing heterocy-
cle at C2, are selective partial agonists of the 5-HT3 recep-
tor.8–11 The most distinctive feature of our compounds is
the substituent-dependence of their effects on the 5-HT3
receptor ion channel, that is, modification of substituents
(R1, R2) on the rigid benzoxazole nucleus of 1 affords both
agonist-like and antagonist-like compounds. As in the
case of the endogenous agonist serotonin (5-HT, 2, Figure
1), an agonist-like compound possessing high intrinsic
Keywords: 5-HT3 receptor; LGIC; Partial agonist; Ion channels;
Molecular modeling.
*
Corresponding authors. Tel.: +81 45 541 2521; fax: +81 45 541 0667
(S.Y.); e-mail addresses: satoshi_yoshida@meiji.co.jp; yasuo_sato
@meiji.co.jp
Figure 1. 5-HT3 receptor ligands.
0968-0896/$ - see front matter ꢀ 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2007.02.054

3516
S. Yoshida et al. / Bioorg. Med. Chem. 15 (2007) 3515–3523
activity (ia) stabilizes the open state of the ion channel and
allows influx of Na+ into the cytoplasm. On the other
hand, an antagonist-like molecule with low ia decreases
the frequency of channel-opening and the permeability
of ions. Herein, we present SAR data for our 5-HT3 recep-
tor partial agonists corroborated with a computational
docking study between our ligands and a homology-mod-
eled guinea pig 5-HT3 receptor. The results not only pro-
vide a basis for rational design of novel 5-HT3 receptor
ligands with high affinity and desirable ia, but also offer
information about structural and functional aspects of
the LGIC receptors. In particular, the substrate-depen-
dence of the ia of these partial agonists provides clues to
the roles of the amino acid residues in the ligand-binding
region of the receptor.
no acid residue of LGIC receptors, because all reported
crystal structures of AChBP complexed with nACh
ligands indicate a close contact between the ammonium
moiety of nACh ligands and a carbonyl-oxygen of
Trp147 in the AChBP, which is highly conserved among
LGIC receptors.14 Numerous SAR studies and muta-
tional studies on LGIC receptors have demonstrated
the importance of the ammonium moiety for interaction
with the receptor, and the crucial role of the conserved
Trp residue for ligand binding.10,15,16 Based on the
presumption, the ligand location was constrained during
calculation such that the quaternary ammonium salt of
the ligand was positioned near the carbonyl oxygen of
Trp183 in the 5-HT3 receptor binding region; this residue
corresponds to Trp147 in the AChBP.
Table 1 summarizes the activity of the compounds. The
ia data of 1a–m in Table 1 indicate that modification of
the C5 substituent (R1) moderately varied the agonist/
antagonist-like characteristics of the compounds. Intro-
ducing an electron-donating group, such as CH3,
OCH3 or NH2 (compds 1b, 1f, and 1g), at C5 afforded
partial agonists with high ia (>0.7), but introduction of
an electron-withdrawing substituent, NO2, Cl, or Br
(1e, 1i, and 1j), yielded partial agonists with lower ia
(around 0.5 or less). The variable ia of the compounds
depending on the C5 substituent may be dependent upon
whether the C5 substituent directly contacts the receptor
amino acid residues that play a role in ion channel gat-
ing, or whether it indirectly influences the electrostatic
character of a distant functional group having contacts
with such residues. The pD2 data of these compounds
indicate that alteration of the C5 substituent can change
the agonist potency of the compounds for the 5-HT3
receptor over 10-fold. Small lipophilic C5 substituents,
such as CH3, F, Cl, and Br (compds 1b and 1h–j,
Table 1), notably increased the pD2 of the compounds.
Other C5 substituents with greater steric hindrance, Et
(1c) and tBu (1d), and with hydrophilicity, OCH3(1f),
NH2(1g), CF3(1k), COCH3(1l), and CO2Et (1m), did
not increase pD2. Both the low pD2 and the weak
5-HT3 antagonism of these compounds reflect their weak
affinity for the receptor. These groups may undergo
steric and/or electrostatic repulsion with amino acid res-
idues in the binding site. On the other hand, the observa-
tion that small lipophilic C5 substituents afford
compounds with both superior pD2 and potent antago-
nism indicates that the C5 substituent may fit in a
narrow, hydrophobic groove of the binding region in
the receptor. In the case of C7 substituents (R2 compds
1n–r, Table 1), the type of functional group on C7 has
a marked effect on ia, but little effect on pD2. Introduc-
tion of a CH3 group at C7 provided an agonist-like com-
pound 1n with high ia, while Et and Cl groups at the
same position afforded antagonist-like compounds 1o
and 1r with low ia (ia = 0.15 for 1o and 0.24 for 1r) with-
out any change of pD2. Compound 1p, which has an
exceptionally bulky substituent (i-Pr), showed unmea-
surably low pD2 and ia, but showed good antagonism
(81%) at 10ꢀ5 M; this suggests that 1p can be classified
as a full antagonist. Furthermore, the low ia of the C7
chlorinated compound 1r suggests that not only the ste-
ric, but also the electrostatic characteristics of the C7
2. Results and discussion
Compounds 1a–z were synthesized by means of the pro-
cedures illustrated in Scheme 1. All compounds substi-
tuted at the 5 or/and 7 positions of the benzoxazole
ring were synthesized from o-nitrophenols. Reduction
of 5b–z by catalytic hydrogenation followed by ring for-
mation with carbon disulfide afforded the cyclized thiols
7b–z in good yield. Coupling of 7b–z with N-methylpip-
erazine was achieved by means of the following methods:
(i) Coupling of N-methylpiperazine with 2-chlorobenz-
oxazole obtained by the reaction of 2-mercaptobenzox-
azole with phosphorus pentachloride. (ii) Heating a
mixture of 2-mercaptobenzoxazole and N-methylpiper-
azine in a nonpolar solvent. Compound 1a was obtained
by the treatment of commercially available 2-chloro-
benzoxazole with N-methylpiperazine.
All of the biological data were obtained from contraction
tests using isolated guinea pig ileum:9 The activity of each
compound for the 5-HT3 receptor is represented by the
parameter pD2 and the intrinsic activity (ia). The pD2 va-
lue indicates the order of agonist potency, expressed as
the inverse logarithm of the half-maximal effective con-
centration (EC50) of a partial agonist, that is, the concen-
tration at which the biological–response curve reaches
50% of the maximal response. The value of ia represents
the relative efficacy of a partial agonist as compared to
the full agonist 5-HT (ia = 1.0, Table 1), that is, this
parameter relatively quantifies the intensity of agonist-in-
duced signal transduction. To avoid overlooking com-
pounds that possess affinity for the receptor, but exhibit
a low ia (antagonist-like molecules), ability to block the
effect of 10ꢀ5 M 5-HT (5-HT3 antagonism) was also mea-
sured. The binding profile of representative compounds
with 5-HT3 receptor and the selectivity over other recep-
tors are presented in supplementary data (Tables S1 and
S2). Homology model building of the guinea pig 5-HT3
receptor was carried out with Insight II (Accelrys, San
Diego, CA) and Cerius2 (Accelrys) on the basis of the
crystal structure of AChBP.12,13 Docking simulations
were performed with Cerius2 based on the crystal struc-
tures of AChBP complexed with nicotinic acetylcholine
(nACh) ligands.12,13 We hypothesized that the character-
istic ammonium moiety of LGIC ligands (protonated
form) commonly interacts with the highly conserved ami-

S. Yoshida et al. / Bioorg. Med. Chem. 15 (2007) 3515–3523
3517
Scheme 1. Synthesis of 1a–z. (a) H2, Pd/C, EtOH; (b) H2, Pt on sulfide carbon, EtOH; (c) CS2, KOH, EtOH, reflux; (d) i.PCl5, toluene, reflux;
ii.N–Me piperazine, toluene; (e) N–Me piperazine, toluene, reflux; (f) N–Me piperazine, chloroform.
Table 1. Agonist activity of 1a–z for the 5-HT3 receptor
Compound
R1
R2
ia SEMa
pD2 SEMa
Antagonist %
1a
1b
1c
1d
1e
1f
H
H
0.62 0.12
0.72 0.04
5.01 0.13
6.00 0.05
<4.0
82
87
63
47
79
74
73
91
92
89
36
22
28
89
90
81
57
90
94
99
96
91
94
100
90
92
—
100
CH3
Et
H
b
H
—
b
t-Bu
NO2
OCH3
NH2
F
H
—
<4.0
H
0.52 0.05
0.78 0.05
0.76 0.05
0.64 0.04
0.50 0.03
0.48 0.03
6.10 0.10
4.78 0.09
4.97 0.08
5.57 0.06
6.07 0.15
6.21 0.04
<4.0
H
1g
1h
1i
H
H
Cl
Br
H
1j
H
b
1k
1l
CF3
COCH3
CO2Et
H
H
—
b
—
H
<4.0
<4.0
b
1m
1n
1o
1p
1q
1r
1s
1t
H
—
0.58 0.06
0.15 0.04
CH3
Et
5.48 0.06
5.48 0.13
<4.0
H
b
H
i-Pr
OCH3
Cl
—
b
—
H
<4.0
H
0.24 0.15
0.62 0.12
0.26 0.02
0.26 0.06
0.56 0.09
0.24 0.07
0.17 0.03
0.14 0.03
0.62 0.07
1.00
5.50 0.13
6.32 0.13
6.46 0.06
6.22 0.08
6.13 0.04
6.70 0.04
6.82 0.20
6.51 0.04
6.16 0.03
5.34 0.02
<4.0
CH3
CH3
CH3
CH3
Cl
CH3
Et
1u
1v
1w
1x
1y
1z
2
Cl
OCH3
CH3
Et
Cl
Cl
Cl
OCH3
Cl
b
—
3
a Standard error of the mean (n = 5).
b No agonistic response was observed at 10ꢀ4 M.
substituent might influence ia. To test this conjecture,
compound 1q, which carries a methoxy group at C7,
was also synthesized. If the electrostatic character of
the C7 group is important for ia, introduction of a
strongly electron-donating substituent, in contrast to
the electron-withdrawing Cl group, should raise the ia
of the partial agonists. However, 1q showed only weak
antagonism (57%) at 10ꢀ5 M and agonistic activity was
not detected. In order to improve the potency of the
C7-substituted derivatives, we examined 5,7-disubsti-
tuted partial agonists. The pD2 and ia data of com-
pounds 1s–z show that they possess greater potency
than C7-substituted compounds, while the C7 substitu-
ent effect is retained. A bulky ethyl group (1t,1x) or an
electronegative Cl group (1u,1y) at C7 lowered the ia
in both the C5 CH3 and the C5 Cl series, but the small,
electron-donating CH3 group retained the ia of 1s. Com-
pound 1w, which possesses a Cl atom at C5 and a CH3
group at C7, exhibited a lower ia (0.24) than we had
expected. However, the substituent effect of the CH3
group is still retained (compare the ia of 1w with that
of 1x or 1y). Finally, the introduction of a methoxy
group at the C7 position of 5,7-disubstituted derivatives
greatly enhanced the intrinsic activity of 1v (ia = 0.56)
and 1z (ia = 0.62) with high pD2, as we had anticipated.
This result demonstrates that the electrostatic character
of the C7 substituent influences the ia of partial agonists
in a similar manner to that of the C5 substituent. The
modest impact of the C7 substituent on pD2 and the
strong influence on ia imply that the receptor amino acid
residues that interact with the C7 substituent play little
role in ligand binding, but have an important role in

3518
S. Yoshida et al. / Bioorg. Med. Chem. 15 (2007) 3515–3523
ion channel gating, and in particular, sterically bulky C7
substituents are likely to show a greater interaction with
the gating amino acid residues.
that the introduction of those groups into the benzoxaz-
ole ring of 1a lowers its LUMO level, and intensifies the
p–p interaction between 1a and electron-rich Tyr153.
The exceptionally good pD2 of the NO2 (1e) group de-
spite its polarity can also be explained by the compact-
ness and strongly electron-withdrawing character of
Figure 2a and b shows the result of docking of 1a with
the homology-modeled guinea pig 5-HT3 receptor. The
quaternary ammonium group of 1a forms hydrogen
bonds with the oxygen functions of Trp183 and
Asp228, and has a cation-p interaction with Phe235.
The aromatic ring of 1a is intercalated, and is involved
in hydrophobic interaction with Leu184 and Tyr153.
The piperazine ring also interacts hydrophobically with
the aromatic side chains of Trp183 and Phe235. The O1
oxygen of 1a can form a hydrogen bond with Ser231.
The N3 nitrogen atom is directed toward Asn138,
Ile139, Tyr141, and may form hydrogen bonds with
these residues through water. This model can explain
the SAR data for our compounds. The structurally
inflexible b-sheets involving Val142 and Tyr143 are
located extremely close to the C5 substituent, and form
a narrow hydrophobic space. This is consistent with the
strong preference for a small substituent at C5 for high
pD2. In contrast, bulky and/or hydrophilic C5 substitu-
ents would have a repulsive interaction with these amino
acid residues on the b-sheets. The benzoxazole nucleus
of 1a sandwiched between Leu184 and Tyr153 may have
a p–p interaction with the side-chain aromatic ring of
Tyr153. The SAR study on the C5 substituent showed
that small electron-withdrawing substituents, such as
F, Cl or Br, are favorable for high pD2. We conjecture
this group. We previously proposed
a ‘3-point
pharmacophore model’ for agonist binding to the 5-
HT3 receptor.9 This model consists of an aromatic ring,
a nitrogen atom in a heteroaryl group, and terminal
amine as essential functions. Our docking model has al-
lowed us to visualize putative interactions between these
ligand structural features and amino acid residues in the
5-HT3 receptor model. In considering the variation of ia
of our partial agonists, the observed interaction between
O1 oxygen and Ser231 is particularly important. Ser231
corresponds to Cys191 on loop C of AChBP, and this
residue is associated with distinct conformational
changes upon binding of agonists as compared with
antagonists (Fig. 2c). This domain appears to play a crit-
ical role in ion channel gating by the LGIC recep-
tors.12,13 The introduction into ligands of C5 and C7
substituents with different electrostatic character alters
the electron density of the O1 oxygen and the strength
of the hydrogen bond between O1 oxygen and Ser231:
C5 and C7 substituents with electron-donating character
increase the electron density of the O1 oxygen and form
a stronger hydrogen bond with the OH group of Ser231.
This interaction stabilizes the ‘closed’ conformation
(agonist binding form of AChBP, in which the ion
Figure 2. (a and b) Docking model of 1a with 5-HT3 receptor ligand binding domain. (c) Agonist and antagonist binding forms of the homology-
modeled 5-HT3 receptor. The ligand binding domain (light blue dotted frame), ‘closed’ conformation (green), and ‘open’ conformation (magenta) are
shown.

S. Yoshida et al. / Bioorg. Med. Chem. 15 (2007) 3515–3523
3519
channel is activated; Fig. 2c).12 On the other hand, sub-
stituents with electron-withdrawing character decrease
the electron density of the O1 oxygen atom and weaken
the hydrogen bond. Binding of these ligands rather sta-
bilizes the ‘open’ conformation (antagonist-binding
form of AChBP, in which the ion channel is deactivated;
Fig. 2c).12 Our model indicates that a substituent at C7
can also interact with loop C. The above SAR data show
that a sterically bulky C7 substituent dramatically low-
ers the ia of the partial agonists. We can interpret this
in terms of steric repulsion between the bulky C7 substi-
tuent and the amino acid residues on loop C, causing the
receptor to favor the ‘open’ conformation. Thus, our
findings indicate that the variable ia of our partial ago-
nists principally originates from the diverse character of
the interactions between the compounds and loop C of
the 5-HT3 receptor.
J = 8 Hz, piperazine –CH2– ·2), 7.02 (1H, t, J = 7 Hz,
benzoxazole 5-H), 7.18 (1H, t, J = 7 Hz, benzoxazole
6-H), 7.26 (1H, d, J = 7 Hz, benzoxazole 7-H), 7.37
(1H, d, J = 7 Hz, benzoxazole 4-H). MS (EI) m/z 217
(M+). Anal. Calcd for (C12H15N3O): C, 66.34; H, 6.96;
N, 19.34. Found: C, 66.46; H, 7.01; N, 19.08.
4.1.2. 5-Chloro-2-(4-methyl-1-piperazynyl)benzoxazole (1i).
This procedure illustrates the general synthetic method
for 1b, 1d, 1e, and 1n. 2-Amino-4-chlorophenol
(6i; 10 g, 70 mmol) was refluxed for 8 h with potassium
hydroxide (4.7 g, 84 mmol) and carbon disulfide
(100 ml) in ethanol (150 ml). The reaction mixture was
concentrated in vacuo. 1 N aqueous hydrochloric acid
(100 ml) and ethyl acetate (200 ml) were added to the
residue. The organic layer was washed with water, then
dried over MgSO4, and concentrated in vacuo. The
11.5 g of 5-chloro-2-mercaptobenzoxazole (7i) was
obtained as a yellow powder and used in the next reac-
tion without further purification. Phosphorus pentachlo-
ride (1.35 g, 6.5 mmnol) was added to the solution of 7i
(1 g, 5.4 mmol) in dry toluene (50 ml) at ambient tem-
perature. The reaction mixture was refluxed for 1 h,
then cooled in an ice bath. 1-Methylpiperazine (5.4 g,
54 mmol) was added dropwise to the mixture, and stir-
red for 30 min at 0 ꢁC. The reaction mixture was di-
luted with chloroform (100 ml), then washed with
water. The organic layer was dried over MgSO4 and
concentrated in vacuo. The residue was chromato-
graphed on silica gel with chloroform–methanol (20:1)
to afford 1i (1 g, 67% from 5i) as a white needle, mp
118–119 ꢁC (acetone–hexane). NMR (CD3OD) d: 2.34
(3H, s, CH3–), 2.57 (4H, t, J = 5 Hz, piperazine
–CH2– ·2), 3.71 (4H, t, J = 5 Hz, piperazine –CH2– ·2),
7.03 (1H, d, J = 7 Hz, benzoxazole 6-H), 7.22 (1H, s,
benzoxazole 4-H), 7.26 (1H, d, J = 7 Hz, benzoxazole
7-H). MS (EI) m/z 251 (M+). Anal. Calcd for
(C12H14N3OCl): C, 61.26; H, 6.00; N, 17.86. Found:
C, 61.21; H, 5.99; N, 17.87.
3. Conclusion
In summary, we have shown that synergy between
biochemical and computational docking studies can
greatly aid interpretation of SAR data for 5-HT3 recep-
tor ligands. Our results will be useful for the rational
design of 5-HT3 receptor partial agonists with desirable
ia and good potency, and demonstrate the effectiveness
of docking calculations with experimental constraints.
The characteristic substrate-dependence of the ia of
our partial agonists allowed us to elucidate the roles of
various amino acid residues in the ligand binding region,
as well as throwing light on the structural features of the
5-HT3 receptor. We are currently using a similar
approach to design novel ligands for other LGIC
receptors.
4. Experimental
4.1. Chemistry
4.1.3. 5-Methyl-2-(4-methyl-1-piperazinyl)benzoxazole
(1b). Obtained as yellow plate, 73% yield from 2-ami-
no-p-cresol, mp 63–64 ꢁC (methanol–ether). NMR
(CDCl3) d: 2.35 (3H, s), 2.39 (3H, s), 2.52 (4H, t,
J = 5 Hz), 3.71 (4H, t, J = 5 Hz), 6.82 (1H, d,
J = 8 Hz), 7.11 (1H, d, J = 8 Hz), 7.15 (1H, s). MS
(EI) m/z 231 (M+). Anal. Calcd for (C13H17N3O-1/8
H2O): C, 66.86; H, 7.44; N, 17.99. Found: C, 66.77;
H, 7.40; N, 18.00.
Elemental analysis data were recorded on VarioEL
(Elementar). NMR spectra were obtained on JEOL
GX-400 FT-NMR spectrometers. The following abbre-
viations are used: s = singlet, d = doublet, t = triplet,
q = quartet,
quin = quintet,
m = multiplet,
and
br = broad. MS were measured with Hitachi M-80B
and JEOL JMS-700 instruments. Serotonin (2) and
granisetron (3) are commercially available products.
4.1.1.
2-(4-Methyl-1-piperazynyl)benzoxazole
(1a).
4.1.4. 5-tert-Butyl-2-(4-methyl-1-piperazynyl)benzoxaz-
ole (1d). Obtained as yellow oil, 62% yield from 2-ami-
no-4-tert-butylphenol. NMR (CDCl3) d: 1.39 (9H, s),
2.42 (3H, s), 2.60 (4H, t, J = 4 Hz), 3.81 (4H, t,
J = 4 Hz), 7.15 (1H, dd, J = 2, 8 Hz), 7.24 (1H, d,
J = 8 Hz), 7.50 (1H, d, J = 2 Hz). MS (EI) m/z 274
(M++1). Anal. Calcd for (C16H23N3O1/3H2O) C,
68.79; H, 8.54; N, 15.04. Found: C, 68.81; H, 8.67;
N, 15.13.
2-Chlorobenzoxazole (8; 1.4 g, 9 mmol) was added to
the solution of 1-methylpiperazine (1 g, 10 mmol) in
chloroform (100 ml) at 0 ꢁC. The reaction mixture was
stirred for 30 min and then the reaction quenched in
ice-water (150 ml). The resulting mixture was extracted
with ethyl acetate (200 ml), and the extract was dried
over MgSO4. After concentration in vacuo, the residue
was chromatographed on silica gel with chloroform–
methanol (20:1) and then recrystallized from water-ace-
tone to give 1a (1.8 g, 85%) as a white needle, mp 37–
38 ꢁC. NMR (CDCl3 d: 2.37 (3H, s, CH3–), 2.54 (4H,
t, J = 8 Hz, piperazine –CH2– ·2), 3.73 (4H, t,
4.1.5. 2-(4-Methyl-1-piperazynyl)-5-nitrobenzoxazole
(1e). Obtained as yellow solid, 5% yield from 2-ami-
no-4-nitrophenol, mp 112–113 ꢁC (Et2O–hexane).

3520
S. Yoshida et al. / Bioorg. Med. Chem. 15 (2007) 3515–3523
NMR (CDCl3) d: 2.37 (3H, s), 2.56 (4H, t, J = 5 Hz),
3.80 (4H, t, J = 5 Hz), 7.32 (1H, d, J = 9 Hz), 8.13
(1H, d, J = 2 Hz), 8.18 (1H, dd, J = 2, 9 Hz). MS (EI)
m/z 263 (M++1). Anal. Calcd for (C12H14N4O3) C,
54.96; H, 5.38; N, 21.36. Found: C, 54.98; H, 5.23;
N, 21.07.
(4H, t, J = 5 Hz), 4.37 (2H, q, J = 7 Hz), 7.26 (1H, d,
J = 8 Hz), 7.80 (1H, dd, J = 2, 8 Hz), 8.03 (1H, d,
J = 2 Hz). MS (TSP) m/z 290 (M++1). Anal. Calcd for
(C15H19N3O3-1/3H2O): C, 61.00; H, 6.71; N, 14.23.
Found: C, 60.97; H, 6.41; N, 14.30.
4.1.12. 7-Ethyl-2-(4-methyl-1-piperazinyl)benzoxazole (1o).
Obtained as brown oil, 48% yield from 2-ethyl-6-nitrophe-
nol.19 NMR (CDCl3) d: 1.31 (3H, t, J = 8 Hz), 2.36 (3H,
s), 2.54 (4H, t, J = 4 Hz), 2.81 (2H, q, J = 8 Hz), 3.73 (4H,
t, J = 4 Hz), 6.86 (1H, d, J = 7 Hz), 7.09 (1H, t, J = 7 Hz),
7.20 (1H, d, J = 7 Hz). MS (EI) m/z 245 (M+).
4.1.6. 7-Methyl-2-(4-methyl-1-piperazinyl)benzoxazole
(1n). Obtained as yellow oil, 70% yield from 2-ami-
no-6-methylphenol. NMR (CDCl3) d: 2.36 (3H, s),
2.42 (3H, s), 2.53 (4H, t, J = 5 Hz), 3.73 (4H, t,
J = 5 Hz), 6.83 (1H, d, J = 8 Hz), 7.06 (1H, t,
J = 8 Hz), 7.19 (1H, d, J = 8 Hz). MS (EI) m/z 231
(M+). HR-MS Calcd for C13H18N3O: 232.1450.
Found: 232.1448.
4.1.13.
7-Isopropyl-2-(4-methyl-1-piperazinyl)benzox-
azole (1p). Obtained as yellow oil, 66% yield from 2-iso-
propyl-6-nitrophenol.20 NMR (CDCl3) d: 1.35 (6H, d,
J = 7 Hz), 2.36 (3H, s), 2.54 (4H, t, J = 5 Hz), 3.23
(1H, m), 3.73 (4H, t, J = 5 Hz), 6.88 (1H, m), 7.10
(1H, t, J = 8 Hz), 7.20 (1H, dd, J = 1, 8 Hz). FAB-MS
m/z 260 (M++1).
4.1.7. 2-(4-Methyl-1-piperazinyl)-5-trifluoromethylbenz-
oxazole (1k). This procedure illustrates the general syn-
thetic method for 1c, 1f, 1l, 1m, 1o, 1p, 1q, 1s, 1t, and 1v.
2-Nitro-6-trifluoromethylphenol (0.88 g, 5.75 mmol)
was dissolved in ethanol (30 ml), and 10% palladium–
carbon (0.09 g) was added to the solution. The reaction
mixture was stirred under a hydrogen atmosphere for
24 h, and the palladium–carbon was removed by filtra-
tion. The solution was concentrated in vacuo, then the
crude 2-amino-6-trifluoromethylphenol was treated as
described for the preparation of 1i to afford 1k as color-
less needle, 21% yield from 2-nitro-4-trifluoromethyl-
phenol, mp 72–73 ꢁC (Et2O–hexane). NMR (CDCl3) d:
2.36 (3H, s), 2.54 (4H, t, J = 5 Hz), 3.75 (4H, t,
J = 5 Hz), 7.30 (2H, s), 7.57 (1H, s). MS (TSP) m/z
286 (M++1). Anal. Calcd for (C13H14N3OF3): C,
54.74; H, 4.95; N, 14.73. Found: C, 54.63; H, 5.09; N,
14.49.
4.1.14.
7-Methoxy-2-(4-methyl-1-piperazinyl)benzox-
azole (1q). Obtained as brown oil, 89% yield from 2-
methoxy-6-nitrophenol.21 NMR (CDCl3) d: 2.35 (3H,
s), 2.52 (4H, t, J = 5 Hz), 3.74 (4H, t, J = 5 Hz), 3.97
(3H, s), 6.63 (1H, m), 7.00 (1H, m), 7.10 (1H, t,
J = 9 Hz). MS (EI) m/z 247 (M+).
4.1.15. 5,7-Dimethyl-2-(4-methyl-1-piperazinyl)benzox-
azole (1s). Obtained as yellow oil, 73% yield from 2,4-di-
methyl-6-nitrophenol.22 NMR (CDCl3) d: 2.34 (6H, s),
2.36 (3H, s), 2.50 (4H, t, J = 4 Hz), 3.70 (4H, t,
J = 4 Hz), 6.60 (1H, s), 7.00 (1H, s). FAB-MS m/z 246
(M++1). HR-MS Calcd for C14H20N3O(M++1):
246.1609. Found: 246.1608.
4.1.8. 5-Ethyl-2-(4-methyl-1-piperazinyl)benzoxazole (1c).
Obtained as yellow oil, 98% yield from 4-ethyl-2-nitrophe-
nol.17 NMR (CDCl3) d: 1.24 (3H, t, J = 7 Hz), 2.35 (3H, s),
2.52 (4H, t, J = 5 Hz), 2.68 (2H, q, J = 7 Hz), 3.71 (4H, t,
J = 5 Hz), 6.85 (1H, dd, J = 1, 8 Hz), 7.14 (1H, d,
J = 8 Hz), 7.19 (1H, s). MS (EI) m/z 245 (M+). HR-MS
Calcd for C14H20N3O(M++1): 246.1606. Found: 246.1616.
4.1.16. 7-Ethyl-5-methyl-2-(4-methyl-1-piperazinyl)benz-
oxazole (1t). Obtained as yellow oil, 85% yield from 4-
ethyl-2-methyl-6-nitrophenol. NMR (CDCl3) d: 1.31
(3H, t, J = 8 Hz), 2.34 (3H, s), 2.36 (3H, s), 2.54 (4H,
t, J = 4 Hz), 2.81 (2H, q, J = 8 Hz), 3.73 (4H, t,
J = 4 Hz), 6.60 (1H, s), 7.00 (1H, s). MS (TSP) m/z
260 (M++1).
4.1.9. 5-Methoxy-2-(4-methyl-1-piperazinyl)benzoxazole
(1f). Obtained as yellow oil, 92% yield from 4-meth-
oxy-2-nitrophenol. NMR (CDCl3) d: 2.35 (3H, s), 2.52
(4H, t, J = 5 Hz), 3.71 (4H, t, J = 5 Hz), 3.80 (3H, s),
6.58 (1H, dd, J = 3, 9 Hz), 6.92 (1H, d, J = 3 Hz), 7.12
(1H, d, J = 9 Hz). MS (TSP) m/z 248 (M++1). HR-MS
Calcd for C13H18N3O2(M++1): 248.1399. Found:
248.1390.
4.1.17. 7-Methoxy-5-methyl-2-(4-methyl-1-piperazinyl)benz-
oxazole (1v). Obtained as yellow oil, 93% yield from 2-
methoxy-4-methyl-6-nitrophenol.23 NMR (CDCl3) d: 2.34
(3H, s), 2.38 (3H, s), 2.51 (4H, t, J = 5 Hz), 3.72 (4H, t,
J = 5 Hz), 3.95 (3H, s), 6.44 (1H, s), 6.80 (1H, s). MS
(TSP) m/z 262 (M++1). HR-MS Calcd for
C14H20N3O2(M++1): 262.1555. Found: 262.1559.
4.1.10. 5-Aceto-2-(4-methyl-1-piperazinyl)benzoxazole (1l).
Obtained as brown oil, 87% yield from 40-hydroxy-30-nit-
roacetophenone. NMR (CDCl3) d: 2.36 (3H, s), 2.54 (4H,
t, J = 5 Hz), 2.62 (3H, s), 3.75 (4H, t, J = 5 Hz), 7.29 (1H,
d, J = 8 Hz), 7.74 (1H, dd, J = 2, 8 Hz).
4.1.18. 5-Chloro-7-methyl-2-(4-methyl-1-piperazinyl)benz-
oxazole (1w). This procedure illustrates the general
synthetic method for 1h, 1j, 1r, 1u, and 1x–z. 4-Chloro-
2-methyl-6-nitrophenol24 (5w, 2.0 g, 10.7 mmol) was
dissolved in ethyl acetate (60 ml), and 5% platinum on sul-
fide carbon (60 mg; Aldrich Chemical Co.) was added to
the solution. The reaction mixture was stirred under
hydrogen atmosphere for 24 h, and platinum on sulfide
carbon was removed by filtration. The solution was con-
centrated in vacuo. The 1.7 g (10.7 mmol) of crude 2-ami-
no-4-chloro-6-methylphenol was refluxed for 8 h with
4.1.11. 5-Ethoxycarbonyl-2-(4-methyl-1-piperazinyl)benz-
oxazole (1m). Obtained as colorless needle, 90% yield
from 4-ethoxycarbonyl-2-aminophenol,18 mp 49–51 ꢁC
(Et2O–hexane). NMR (CDCl3) d: 1.39 (3H, t,
J = 7 Hz), 2.36 (3H, s), 2.53 (4H, t, J = 5 Hz), 3.74

S. Yoshida et al. / Bioorg. Med. Chem. 15 (2007) 3515–3523
3521
potassium hydroxide (4.7 g, 84 mmol) and carbon disul-
fide (33 ml) in ethanol (100 ml). The reaction mixture
was concentrated in vacuo, and 5 N aqueous hydrochloric
acid (18 ml) and ethyl acetate (100 ml) were added to the
residue. The organic layer was washed with water
(100 ml), dried over MgSO4, and concentrated in vacuo.
5-Chloro-2-mercapto-7-methylbenzoxazole (7w) was ob-
tained as a pale-brownish solid and used in the next step
without further purification. 7w (0.2 g, 1.0 mmol) and
methylpiperazine (0.5 ml, 4.5 mmol) were dissolved in
dry toluene (15 ml), and the mixture was refluxed for
16 h. The solvent was removed under reduced pressure,
and the residue was chromatographed on silica gel with
dichloromethane–methanol (10:1) to give 1w (270 mg,
84% from 5w) as yellow solid, mp 62–64 ꢁC (Et2O–hex-
ane). NMR (CDCl3) d: 2.36 (3H, s), 2.37 (3H, s), 2.53
(4H, t, J = 5 Hz), 3.72 (4H, t, J = 5 Hz), 6.81 (1H, d,
J = 2 Hz), 7.14 (1H, d, J = 2 Hz). MS (TSP) m/z 266
(M++1), 268 (M++3). Anal. Calcd for (C13H16N3OCl):
C, 58.76; H, 6.07; N, 15.82. Found: C, 58.73; H, 6.12;
N, 15.69.
d: 1.29 (3H, t, J = 7 Hz), 2.36 (3H, s), 2.53 (4H, t,
J = 5 Hz), 2.78 (2H, q, J = Hz), 3.72 (4H, t,
J = 5 Hz), 6.83 (1H, d, J = 2 Hz), 7.15 (1H, d,
J = 2 Hz). MS (TSP) m/z 280 (M++1), 282 (M++3).
HR-MS Calcd for C14H19N3OCl (M++1): 280.1217.
Found: 280.1193.
4.1.24. 5,7-Dichloro-2-(4-methyl-1-piperazinyl)benzoxaz-
ole (1y). Obtained as yellow needle, 64% yield from 2,4-
dichloro-6-nitrophenol, mp 106–107 ꢁC (Et2O–hexane).
NMR (CDCl3) d: 2.36 (3H, s), 2.53 (4H, t, J = 5 Hz),
3.75 (4H, t, J = 5 Hz), 7.00 (1H, d, J = 2 Hz), 7.18
(1H, d, J = 2 Hz). MS (EI) m/z 285 (M+), 287 (M++2),
289 (M++4). Anal. Calcd for (C12H13N3Ocl2): C,
50.37; H, 4.58; N, 14.68. Found: C, 50.10; H, 4.42; N,
14.44.
4.1.25. 5-Chloro-7-methoxy-2-(4-methyl-1-piperazinyl)benz-
oxazole (1z). Obtained as colorless solid, 70% yield from
4-chloro-2-methoxy-6-nitrophenol.27 NMR (CDCl3) d:
2.32 (3H, s), 2.51 (4H, t, J = 5 Hz), 3.73 (4H, t,
J = 5 Hz), 3.95 (3H, s), 6.61 (1H, d, J = 2 Hz), 6.97
(1H, d, J = 2 Hz). MS (TSP) m/z 282 (M++1), 175
(M++3).
4.1.19. 5-Fluoro-2-(4-methyl-1-piperazinyl)benzoxazole
(1h). Obtained as colorless needle, 74% yield from 4-flu-
oro-2-nitrophenol, mp 116–117 ꢁC (Et2O–hexane).
NMR (CDCl3) d: 2.35 (3H, s), 2.52 (4H, t, J = 5 Hz),
3.72 (4H, t, J = 5 Hz), 6.71 (1H, dt, J = 2, 9 Hz), 7.04
(1H, dd, J = 2, 9 Hz), 7.14 (1H, dd, J = 4, 9 Hz). MS
(EI) m/z 235 (M+). Anal. Calcd for (C12H14N3OF): C,
61.26; H, 6.00; N, 17.86. Found: C, 61.21; H, 5.99; N,
17.87.
4.1.26. 5-Amino-2-(4-methyl-1-piperazinyl)benzoxazole (1g).
Compound 1g was obtained by the catalytic hydrogenoly-
sis of 1e (26 mg, 0.1 mmol) using palladium–carbon (3 mg)
in 1:1 mixture of ethyl acetate–ethanol (4 ml), 27% yield
based on 1e. NMR (CDCl3) d: 2.36 (3H, s), 2.56 (4H, t,
J = 5 Hz), 3.64 (4H, t, J = 5 Hz), 6.45 (1H, dd, J = 2,
8 Hz), 6.91 (1H, d, J = 2 Hz), 7.33 (1H, d, J = 8 Hz). MS
(EI) m/z 233 (M++1).
4.1.20. 5-Bromo-2-(4-methyl-1-piperazinyl)benzoxazole
(1j). Obtained as colorless needle, 60% yield from
4-bromo-2-nitrophenol, mp 100–101 ꢁC (Et2O–hex-
ane). NMR (CDCl3) d: 2.36 (3H, s), 2.47 (3H, s),
2.53 (4H, t, J = 5 Hz), 3.74 (4H, t, J = 5 Hz), 6.71
(1H, dt, J = 2, 9 Hz), 7.04 (1H, d, J = 7 Hz), 7.26
(1H, s), 7.28 (1H, d, J = 7 Hz). MS (TSP) m/z 235
(M++1), 238 (M++3). Anal. Calcd for (C12H14N3OBr):
C, 48.67; H, 4.76; N, 14.19. Found: C, 48.48; H, 4.66;
N, 13.99.
4.2. Computational chemistry
4.2.1. Molecular modeling. Agonist/antagonist bound
models of extracellular regions of the homopentameric
guinea pig 5-HT3 receptor were constructed by using
HOMOLOGY module of Insight II and Cerius2
(Accelrys, San Diego, CA): An agonist bound model
was built based on the crystal structure of AChBP
complexed with
a nACh receptor agonist, LOB
4.1.21. 7-Chloro-2-(4-methyl-1-piperazinyl)benzoxazole
(1r). Obtained as brown oil, 87% yield from
2-chloro-6-nitrophenol. NMR (CDCl3) d: 2.36 (3H,
s), 2.54 (4H, t, J = 5 Hz), 3.76 (4H, t, J = 5 Hz),
7.00 (1H, dd, J = 1, 8 Hz), 7.09 (1H, t, J = 8 Hz),
7.23 (1H, dd, J = 1, 8 Hz). MS (EI) m/z 251 (M+),
253 (M++2).
(PDB entry 2BYS).12 An antagonist bound model
was formed from the crystal structure of AChBP with
a
nACh receptor antagonist MLA (PDB entry
2BYR).12 The sequence of the guinea pig 5-HT3 recep-
tor subunit was brought from GenBank
A
(AAC06136). Sequence alignment between extracellu-
lar region of guinea pig 5-HT3 receptor and AChBP
was performed based on reported alignment results7,28
by using Clustal W.29 On the basis of this alignment,
models of the 5-HT3 receptor subunit were built by
Insight II, and each model was then minimized in
Cerius2 using the cff1.02 force field30 by moving side
chains alone. The pentamer models were generated
by superimposing the models onto each protomer of
AChBP and then were minimized by moving side
chains alone.
4.1.22. 7-Chloro-5-methyl-2-(4-methyl-1-piperazinyl)benz-
oxazole (1u). Obtained as yellow plate, 96% yield from
2-chloro-4-methyl-6-nitrophenol,25 mp 58–59 ꢁC (Et2O–
hexane). NMR (CDCl3) d: 2.35 (3H, s), 2.36 (3H, s),
2.53 (4H, t, J = 5 Hz), 3.74 (4H, t, J = 5 Hz), 6.82 (1H,
s), 7.02 (1H, s). MS (TSP) m/z 266 (M++1), 268
(M++3). Anal. Calcd for (C13H16N3OCl): C, 58.76; H,
6.07; N, 15.81. Found: C, 58.88; H, 5.97; N, 15.93.
4.1.23. 5-Chloro-7-ethyl-2-(4-methyl-1-piperazinyl)benz-
oxazole (1x). Obtained as yellow powder, 55% yield
from 4-chloro-2-ethyl-6-nitrophenol.26 NMR (CDCl3)
4.2.2. Docking of 1a. Docking of 1a was carried out
using agonist bound model as docking template, be-
cause 1a has a high ia value (0.62). Initial position

3522
S. Yoshida et al. / Bioorg. Med. Chem. 15 (2007) 3515–3523
of 1a in agonist bound model was determined by
referring binding conformations of nACh receptor
agonists to AChBP,12,13 that is, an ammonium salt
on the piperidinyl moiety, a Ph ring, and a carbonyl
oxygen in LOB’s structure were, respectively, replaced
to the ammonium salt in the C2 piperadinyl substitu-
ent, the Ph ring, and the nitrogen atom in benzoxaz-
ole of 1a. After an initial position of 1a is determined,
energy minimization was carried out by Cerius2 using
the cff1.02 force field with moving 1a and amino
acids’ side chains around 1a.
Supplementary data
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/j.bmc.
2007.02.054.
References and notes
1. Kilpatrick, G. J.; Bunce, K. T.; Tyers, M. B. Med. Res.
Rev. 1990, 10, 441–475.
2. Sanger, G. J.; Nelson, D. R. Eur. J. Pharmacol. 1989, 159,
113–124.
4.3. Biology
3. Butler, A.; Hill, J. M.; Ireland, S. J.; Jordan, C. C.; Tyers,
M. B. Br. J. Pharmacol. 1988, 94, 397–412.
4. Camileri, M.; Northcutt, A. R.; Kong, S.; Dukes, G. E.;
McSorley, D.; Mangel, A. W. Lancet 2000, 355,
1035–1040.
5. For selected SAR based studies on the LGIC receptors,
see: Glennon, R. A.; Dukat, M.; Liao, L. Curr. Top. Med.
Chem. 2004, 4, 631–644; Dukat, M. Curr. Med. Chem.
Cent. Nerv. Syst. Agents 2004, 4, 77–94.
6. Costa, V.; Nistri, A.; Cavalli, A.; Carloni, P. Br. J.
Pharmacol. 2003, 140, 921–931.
7. Suryanarayanan, A.; Joshi, P. R.; Bikadi, Z.; Mani, M.;
Kulkarni, T. R.; Gaines, C.; Schulte, M. K. Biochemistry
2005, 44, 9140–9149.
8. Sato, Y.; Imai, M.; Amano, K.; Iwamatsu, K.; Konno, F.;
Kurata, Y.; Sakakibara, S.; Hachisu, M.; Izumi, M.;
Matsuki, N.; Saito, H. Biol. Pharm. Bull. 1997, 20,
752–755.
9. Yamada, M.; Sato, Y.; Kobayashi, K.; Konno, F.;
Soneda, T.; Watanabe, T. Chem. Pharm. Bull. 1998, 46,
445–451.
10. Sato, Y.; Yamada, M.; Yoshida, S.; Soneda, T.; Ishikawa,
M.; Nizato, T.; Suzuki, K.; Konno, F. J. Med. Chem.
1998, 41, 3015–3021.
4.3.1. Contraction Test.9 Male Hartley guinea pigs
weighing 500–600 g were sacrificed by exanguination,
and the ileum was excised. Pieces (about 20 mm) of
ileal longitudinal muscles were placed in a 5 ml organ
bath containing Krebs–Henseleit solution aerated with
95% O2 and 5% CO2 at 37 ꢁC. The composition of
the solution was as follows (mM): NaCl 118, KCl
4.7, KH2PO4 1.19, MgSO4 1.2, CaCl2 2.54, NaH-
CO325, and glucose 10. The solution contained ritans-
erin (10ꢀ7 M) to block the 5-HT2 receptor-mediated
response. The preparations were allowed to equili-
brate for at least 30 min under 0.5 g of tension. After
equilibration, the preparations were repeatedly ex-
posed to 3 · 10ꢀ7 M 5-HT for 60 min to desensitize
the 5-HT4 receptor. Test compounds were added to
the bath, and contractions were recorded isometri-
cally. The potencies of agonists were expressed as
pD2 values, that is, the negative logarithm of the mo-
lar concentration which produced 50% of the maxi-
mum
contraction
obtained
from
individual
concentration–response curves. The ia of test com-
pounds was expressed as the ratio between the maxi-
mum response to a test compound and that to 5-HT
(10ꢀ5 M). To evaluate antagonistic activities of the
partial agonists, the contractile responses after a sin-
gle application at 10ꢀ5 M of 5-HT were obtained in
the absence and presence of the test compounds for
20 min. The vehicle-treated preparation was used for
a time-matched control. The 5-HT3 antagonism of
test compounds were defined as the ratio of the con-
tractile response for 5-HT in the presence and absence
of compounds.
11. Yoshida, S.; Shiokawa, S.; Kawano, K.; Ito, T.; Muraka-
mi, H.; Suzuki, H.; Sato, Y. J. Med. Chem. 2005, 48,
7075–7079.
12. Hansen, S. B.; Sulzenbacher, G.; Huxford, T.; Marchot,
P.; Taylor, P.; Bourne, Y. EMBO J. 2005, 24,
3635–3646.
13. Celie, P. H. N.; van Rossum-Fikkert, S. E.; van Dijk, W.
J.; Brejc, K.; Smit, A. B.; Sixma, T. K. Neuron 2004, 41,
907–914.
14. Celite, P. H. N.; Klaassen, R. V.; van Rossum-Fikkert, S.
E.; van Elk, R.; van Nierop, P.; Smit, A. B.; Sixma, T. K.
J. Biol. Chem. 2005, 280, 26457–26466.
15. Spier, A. D.; Lummis, S. C. R. J. Biol. Chem. 2000, 275,
5620–5625.
4.3.2. 5-HT3 receptor binding assay. The assay was per-
formed according to the method of Kilpatrick et al.31
Brain cortices were isolated from male Wistar rats
(250–300 g), and a membrane fraction was prepared by
standard techniques. The membrane fraction (0.04 mg)
was incubated with 0.2 nM [3H]GR65630 for 60 min at
22 ꢁC. Membranes were collected by filtration and
washed three times. The radioactivity on the filters was
counted to determine [3H]GR65630 specifically bound.
Nonspecific binding was estimated in the presence of
1 mM ICS205-930. For obtaining Ki values, assays were
carried out six times at each dose and displacement
curves were fitted by nonlinear regression. IC50 values
were obtained directly. Ki values were calculated from
IC50 values by using the equation of Cheng and
Prusoff.32
16. Beene, D. L.; Brandt, G. S.; Zhong, W.; Zacharias, N. M.;
Lester, H. A.; Dougherty, D. A. Biochemistry 2002, 41,
10262–10269.
17. Kano, F.; Yamane, K.; Yamashita, K.; Hosoe, K.;
Kiyoshi, W. JP Patent, 64–6279, 1989.
18. Papenfuss, T. (Hoechst A-G, Fed. Rept. Ger.). US patent,
3929864, 1975.
19. Melvin, J. A.; Wendell, P. C. J. Am. Chem. Soc. 1943, 65,
2395–2399.
20. Kano, F.; Yamane, K.; Yamashita, K.; Hosoe, K.;
Watanabe, K. Eur. Pat. Appl. 1988, 253340.
21. Dwyer, C. L.; Holzapfei, C. W. Tetrahedron 1998, 54,
7843–7848.
22. Cross, G. G.; Fisher, A.; Henderson, G. N.; Smyth, T. A.
Can. J. Chem. 1984, 62, 1446–1451.
23. Ford, R. E.; Knowles, P.; Lunt, E.; Marshal, S. M.;
Penrose, A. J.; Ramsden, C. A.; Summers, A. J. H.;

S. Yoshida et al. / Bioorg. Med. Chem. 15 (2007) 3515–3523
3523
Walker, J. L.; Wright, D. E. J. Med. Chem. 1986, 29,
538–549.
24. Gibson, G. P. J. Chem. Soc. 1925, 127tl, 42–48.
25. Clewly, R. G.; Cross, G. G.; Fischer, A.; Henderson, G.
N. Tetrahedron 1989, 5, 1299–1310.
28. Brejc, K.; van Dijk, W. J.; Klaassen, R. V.; Schuurmans,
M.; van der Oost, J.; Smlt, A. B.; Sixma, T. K. Nature
2001, 411, 269–276.
29. Thompson, J. D.; Higgins, D. G.; Gibson, T. J. Nucleic
Acid Res. 1994, 22, 4673–4680.
26. Schroetter, E.; Weuffen, W.; Herudek, D. Pharmazie 1974,
29, 374.
30. Hwang, M. J.; Stockfisch, T. P.; Hagler, A. T. J. Am.
Chem. Soc. 1994, 116, 2515–2525.
27. Backstrom, R.; Honkanen, E.; Pippuri, A.; Kairisalo, P.;
Pystynen, J.; Heinola, K.; Nissinen, E.; Linden, I.;
Mannisto, P. T.; Kaakkola, S.; Pohto, P. J. Med. Chem.
1989, 32, 841–846.
31. Kilpatrick, G. J.; Jones, B. J.; Tyers, M. B. Nature 1987,
330, 746–748.
32. Cheng, Y.-C.; Prusoff, W. H. Biochem. Pharmacol. 1973,
22, 3099–3108.