Identification of a novel 4-aminomethylpiperidine class of M3 muscarinic receptor antagonists and structural insight into their M3 selectivity

Article abstract of DOI:10.1021/jm051205r

Identification of a novel class of potent and highly selective M ₃ muscarinic antagonists is described. First, the structure-activity relationship in the cationic amine core of our previously reported triphenylpropionamide class of M₃ selective antagonists was explored by a small diamine library constructed in solid phase. This led to the identification of M₃ antagonists with a novel piperidine pharmacophore and significantly improved subtype selectivity from a previously reported class. Successive modification on the terminal triphenylpropionamide part of the newly identified class gave 14a as a potent M₃ selective antagonist that had > 100-fold selectivity versus the M₁, M ₂, M₄, and M₅ receptors (M₃: K _i = 0.30 nM, M₁/M₃ = 570-fold, M ₂/M₃ = 1600-fold, M₄/M₃ = 140-fold, M₅/M₃ = 12000-fold). The possible rationale for its extraordinarily higher subtype selectivity than reported M₃ antagonists was hypothesized by sequence alignment of multiple muscarinic receptors and a computational docking of 14a into transmembrane domains of M₃ receptors.

Full text of DOI:10.1021/jm051205r

J. Med. Chem. 2006, 49, 5653-5663
5653
Articles
Identification of a Novel 4-Aminomethylpiperidine Class of M3 Muscarinic Receptor
Antagonists and Structural Insight into Their M3 Selectivity
Yufu Sagara,* Takeshi Sagara, Minaho Uchiyama, Sachie Otsuki, Toshifumi Kimura, Toru Fujikawa, Kazuhito Noguchi, and
Norikazu Ohtake
Tsukuba Research Institute, Banyu Pharmaceutical Co., Ltd., 3 Okubo, Tsukuba, Ibaraki 300-2611, Japan
ReceiVed NoVember 30, 2005
Identification of a novel class of potent and highly selective M3 muscarinic antagonists is described. First,
the structure-activity relationship in the cationic amine core of our previously reported triphenylpropionamide
class of M3 selective antagonists was explored by a small diamine library constructed in solid phase. This
led to the identification of M3 antagonists with a novel piperidine pharmacophore and significantly improved
subtype selectivity from a previously reported class. Successive modification on the terminal triphenylpro-
pionamide part of the newly identified class gave 14a as a potent M3 selective antagonist that had >100-
fold selectivity versus the M1, M2, M4, and M5 receptors (M3: Ki ) 0.30 nM, M1/M3 ) 570-fold, M2/M3 )
1600-fold, M4/M3 ) 140-fold, M5/M3 ) 12000-fold). The possible rationale for its extraordinarily higher
subtype selectivity than reported M3 antagonists was hypothesized by sequence alignment of multiple
muscarinic receptors and a computational docking of 14a into transmembrane domains of M3 receptors.
Table 1. Triphenylpropionamide Class of M3 Antagonists
Introduction
Acetylcholine (ACh) exerts a variety of pharmacological
activities via nicotinic and muscarinic receptors distributed
peripherally and centrally. There are five muscarinic receptor
subtypes (M1-M5) identified so far.1 Although these receptor
subtypes have been proposed to participate in a number of
physiologic actions,2 the precise roles of each receptor subtype
in the specific muscarinic actions of ACh remain to be
elucidated. Recent developments of transgenic mice lacking each
of the M1-M5 receptors help to clarify some of their functions;3
however, pharmacological studies using truly subtype-selective
ligands are also important. Furthermore, muscarinic receptors
seem to be involved in many diseases and nonselective
muscarinic receptor ligands have been used as therapeutic
agents. However, these therapeutic agents often cause undesir-
able side effects partly due to their nonselective profiles or the
wide distribution of the muscarinic receptors.4
Since M1-M5 receptors are homologous across receptor
subtypes as well as across mammalian species,5 identification
of truly selective antagonists or agonists can still be challenging.
Efforts are continued to identify selective muscarinic ligands
in pharmaceutical research because of needs for both pharma-
cological tools to clarify the roles of each receptor subtype and
better therapeutic agents.6
binding affinity (Ki, nM)a
(subtype selectivity over M3 receptors)
compd
M3
M1
M2
M4
M5
1
99b
>2500c
(>25)
3100b
(31)
>1700c
(>17)
740b
>5000c
(>51)
(1.0)
2a
2b
23b
>2500c
(>110)
1300 ( 0
(870)
6400b
(280)
>5000c
(>220)
3500 ( 310
(2300)
(1.0)
1.5 ( 0.1
(1.0)
(32)
270 ( 15 57 ( 2
(180) (38)
atropine 0.50 ( 0.01 0.25 ( 0.02 1.5 ( 0.1 0.34 ( 0.02 0.54 ( 0.03
(1.0) (0.50) (3.0) (0.68) (1.1)
a Ki values are expressed as the mean ( standard error of the mean
(SEM), for the compounds which were tested three or more times. b When
only two separate experiments were performed, values are expressed as
the mean of duplicate determinations, and the variances between determina-
tions were less than (11%. c Values are expressed as the mean of duplicate
determinations.
We previously discovered a novel triphenylpropionamide
class of M3 antagonists with good M3 binding affinity and
remarkable subtype selectivity by a combinatorial library
approach.7 Among them, the antagonist possessing a unique
L-Pro-D-Pro spacer, represented by 1, was subjected to further
chemical modification and L-(R)-hydroxyproline(Hyp)-D-Pro
analogues 2a and 2b were identified as M3 antagonists with
improved subtype selectivity (Table 1).8
antagonists with L-Pro-D-Pro spacer and the identification of
14a with a new 4-aminomethylpiperidine core structure that had
sub-nanomolar M3 affinity and >100-fold selectivity versus the
M1, M2, M4, and M5 receptors. In addition, an attempt to
understand the structural basis of the selective recognition of
M3 receptors by 14a was also made using a computational
modeling technique.
In this paper, we will describe the results of structural
alteration in the diamine core structure of the novel M3
Chemistry
Structural alteration in the diamine core part of 1 was explored
by constructing a small library of analogues with diverse diamine
structures.
* To whom correspondence should be addressed. Tel.: +81-29-877-
2000. Fax: +81-29-877-2029. E-mail: yufu_sagara@merck.com.
10.1021/jm051205r CCC: $33.50 © 2006 American Chemical Society
Published on Web 08/19/2006

5654 Journal of Medicinal Chemistry, 2006, Vol. 49, No. 19
Sagara et al.
Scheme 1. Construction of Diamine Library Using Solid Phase Parallel Synthesisa
a Reagents and conditions: (a) see Supporting Information; (b) PyBOP, DIEA, DMF; (c) 50% DCM-TFA.
To maximize the efficiency in exploring structural diversity
in the diamine core part, we designed high-throughput chemistry
as shown in Scheme 1.
tection of both tert-butoxycarbonyl and tert-butyl groups of 13
by treatment with TFA afforded 14a. Finally, either reductive
amination15 or alkylation of 14a with appropriate alkylating
reagents provided the N-alkylated compounds 15a-15g. Ana-
logues 14b and 14c were prepared by similar procedure using
the appropriate tert-butoxycarbonyl-protected diamine instead
of 4-aminomethyl-1-tert-butoxycarbonylpiperidine. Chlorine-
substituted triphenylpropionamide compounds (52a-52c) and
nonsubstituted triphenylpropionamide compounds (2a, 51a-
51c) were also prepared similarly using commercially available
tris(4-chlorophenyl)propionic acid or triphenylpropionic acid,
respectively, instead of 10, in combination with appropriate tert-
butoxycarbonyl-protected diamines.
Tris(4-fluorophenyl)propionic acid (10) used in the prepara-
tion of 14a-14c and 15a-15g was synthesized as shown in
Scheme 3. Condensation reaction of tris(4-fluorophenyl)carbinol
(16) with cyanoacetic acid proceeded cleanly to give tris(4-
fluorophenyl)propionitrile (17) according to the procedures
slightly modified from the original methods reported in the
synthesis of tris(4-chlorophenyl) propionitrile.16 Hydrolysis of
17 gave the target compound 10 in quantitative yield with high
purity.17
The library synthesis was started by solid phase immobiliza-
tion of diverse diamine reagents, as it allows us to operate
selective protection and deprotection of diamine functionalities.
2-Chlorotrityl chloride resin (Cl-Trt(2-Cl)-resin)9 was found to
proceed immobilization of the selected diamine set effectively.10
The solution phase synthesis of monoprotected diamines by
protective groups such as Boc, Cbz, or benzyl group usually
requires lengthy and laborious steps involving conversion of
other functional groups into amine groups. However, solid phase
immobilization of diamines makes it easier to selectively protect
one of the diamine functionalities to leave the other group as
the site for amide condensation with (2R)-1-([(2S)-1-(3,3,3-
triphenylpropanoyl)pyrrolidine-2-yl]carbonyl)pyrrolidine-2-car-
boxylic acid (3) in Scheme 1. The amide condensation reaction
proceeded cleanly in the PyBOP/diisopropylethylamine (DIEA)/
DMF system.11 Finally, the target molecules 4 were cleaved
off from the resin by acid treatment. After cleavage from the
resin, most of the samples except a few showed more than 90%
purity by LC-MS and HPLC spec.12
Subsequently, replacement of the L-Pro-D-Pro spacer in the
novel diamine analogues by L-Hyp-D-Pro spacer and further
modification in the terminal triphenylpropionamide part and
N-substituents of the diamine core part were performed to give
the analogues with the optimal selectivity profiles. All the
reactions for the analogue synthesis were performed by con-
ventional solution phase chemistry.
As an example, preparation of 14a and its N-alkylated tertiary
amine derivatives 15a-15g is shown in Scheme 2.13 Methyl
esterification of commercially available D-benzyloxycarbonyl-
proline 5 followed by deprotection of benzyloxy group (Z) and
subsequent amidation with N-R-benzyloxycarbonyl-O-tert-butyl-
L-4-trans-hydroxyproline (Z-Hyp(tBu)-OH) gave 8. This monoa-
mide 8 was transformed to diamide 11 by deprotection and
condensation with tris(4-fluorophenyl)propionic acid (10). Hy-
drolysis of 11 and subsequent coupling with 4-aminomethyl-
1-tert-butoxycarbonylpiperidine14 afforded triamide 13. Depro-
Results and Discussion
We chose 33 diamine building blocks from commercially
available reagents and the diamines easily prepared from
commercially available precursors to display maximum struc-
tural diversity (Table 2). The 33 crude products prepared in the
solid phase amide condensation reaction were assayed at 1 µM
to measure their inhibition (%) of M3 binding of N-methylsco-
polamine (NMS). As shown in Table 2, the piperidine deriva-
tives (29, 30, 32, and 37) and cyclohexyldiamine derivatives
(39 and 41) displayed high inhibitory activities (>50%), and
their Ki values were determined by independent titration assays.
Among them, the compounds 30, 32, 37, and 41 showed a
higher Ki value than that of the reference compound 1.
M3 binding affinities of the newly identified diamine core
part in 30, 32, 37, and 41 were further examined by testing
their L-Hyp-D-Pro spacer analogues. Because of the difficulty

Identification of a NoVel Class of M3 Antagonists
Journal of Medicinal Chemistry, 2006, Vol. 49, No. 19 5655
Scheme 2. Preparation of 14a and Its N-Alkylated Compounds 15a
a Reagents and conditions: (a) DMAP, MeOH, WSC, CHCl3 (78%); (b) H2, Pd(OH)2, MeOH; (c) Z-Hyp(tBu)-OH, WSC, HOBt, CHCl3 (50%, two
steps); (d) 10, WSC, HOBt, DIEA, CHCl3 (86%, two steps); (e) NaOH (aq), MeOH; (f) 4-aminomethyl-1-t-butoxycarbonylpiperidine, WSC, HOBt, CHCl3
(85%, two steps); (g) TFA (neat) (quant.); (h) aldehyde, NaBH3CN-ZnCl2, MeOH, rt; or RX, K2CO3, CH3CN, heat.
Scheme 3. Preparation of Tris(4-fluorophenyl)propionic Acid
improvements in M3 binding affinity by 10-fold, when compared
to nonsubstituted analogues 51a-51c. More interestingly,
subtype-selectivity profiles of 14a, 14b, and 14c were greatly
improved, and 14a could be the most selective M3 antagonist
among the ones reported so far.18
10a
The effect of piperidine N-substituents of 14a on M3 binding
affinity and subtype selectivity was also examined (Table 4).
To avoid significant increase of the molecular weight and
hydrophobicity of the molecules, relatively small substituents
were examined.
a Reagents and conditions: (a) cyanoacetic acid, zinc chloride, AcOH
at 130 °C, 6.5 h (61%); (b) conc HCl, conc H2SO4, AcOH at 130 °C, 44 h
(quant.).
All the tested N-substituted derivatives retained good M3
binding affinity and selectivity. The cyclopropylmethyl (15c)
and cyclopentylmethyl (15e) derivatives exhibited higher M3
binding affinity than 14a, but they showed no improvements
in their selectivity profiles.
As the result of chemical modification in the diamine core
part, the terminal triphenylpropionamide part, and the piperidine
N-substituents, five compoundss14a, 14c, 15b, 15c, and 52as
were identified to show sub-nanomolar M3 binding affinity and
>100-fold selectivity versus the M1, M2, M4, and M5 receptors.
in separation of the diastereomers of the 1,2-cyclohexanediamine
derivative, we only succeeded in preparation of 51a, 51b, and
51c and tested their M3 binding affinities and subtype selectivi-
ties (Table 3). These derivatives showed higher M3 binding
affinities and selectivity than the reference compound 2a with
3(S)-aminomethyl piperidine structure. Especially, 51a with a
4-aminomethylpiperidine diamine core part was found to
demonstrate >100-fold selectivity versus M1, M2, M4, and M5
receptors.
Optimization of the triphenylpropionamide part of the newly
identified diamine analogues was next carried out (Table 3). In
comparison with 51a-51c, introduction of chlorine atom onto
each of the phenyl rings generally increased M3 binding affinity
in all diamine derivatives (52a-52c), and especially 52a
exhibited stronger M3 binding affinity than 51a by 10-fold.
However, 52b and 52c had slightly lower subtype selectivity.
In the case of fluorine-substituted triphenylpropionamide ana-
logues, all three compounds 14a, 14b, and 14c showed
Among them, 14a with the highest M3 selectivity was
subjected to further evaluation using an in vitro system with
rat tissues19 to obtain the information on the mechanism of
action. In the isolated rat trachea, 14a (0.3-3 nM) effectively
antagonized the carbachol-induced responses in a dose-depend-
ent manner (Figure 1). Schild analysis yielded a slope of 1.08
( 0.16, suggesting a competitive antagonism for this compound
and a pA2 value of 9.65 ( 0.10 in good agreement with the
binding value20 (Figure 1, inset). Compound 14a also exhibited

5656 Journal of Medicinal Chemistry, 2006, Vol. 49, No. 19
Sagara et al.
Table 2. Results of One-Point Screening of Diamine Library (Percent Inhibition of M3 Binding of NMS) and Ki Values of the Selected Compoundsa
aSingle determination for all values (inhibition percent and M3 Ki).
no intrinsic activity at the receptor by compound alone. In the
isolated rat atria, 14a showed less potent inhibition of carbachol-
induced bradycardia, with a pA2 value of 6.69 ( 0.13 and a
slope of 1.04 ( 0.12 (Figure 2). These results indicated that
14a was a competitive antagonist at M2 and M3 receptors, and
the functional selectivity of this compound was 910-fold greater
for tracheal M3 receptors over cardiac M2 receptors in rat, which
was in good agreement with the selectivity observed in the
binding assay.
Modeling of Selective Recognition in M3 Receptors. To
understand the structural basis of the observed selectivity of
our M3 antagonists, the selective binding mode of this new class
in M3 receptors was hypothesized. To this end, we first analyzed
amino acid sequence difference among muscarinic receptor
subtypes (Figure 3). Although extremely high sequence identity
in transmembrane (TM) regions was observed among mus-
carinic subtypes (73-83% identity), three unique amino acid
residues, Asn138, Lys523, and Asn527 in M3 receptor, were
identified in the upper region of TM3 and TM7. As these three
residues are completely nonconserved among the predicted TM
domains of muscarinic receptor subtypes, it might be possible
to assume that the novel antagonist class may specifically
recognize these residues to achieve the observed good subtype
selectivity.
To validate the hypothesis computationally, modeling of the
M3 receptor and its docking with a novel antagonist 14a was
attempted. The TM domains of the M3 receptor were modeled
using an X-ray crystal structure of bovine rhodopsin as a
template. In the M3 receptor model, the three amino acid residues
identified by multiple sequence alignment were found on the
inner surface of TM3 and TM7 to create a potential binding
pocket for small ligand molecules. Therefore, we hypothesized
that these three residues might have M3 specific interactions
with antagonist molecules bound in the cleft surrounded by
seven TM helices.
To obtain further insights for the hypothesis, a highly selective
M3 antagonist 14a was docked into the receptor model. The
docking experiment was carried out by the following steps. (1)
Assuming that the cationic N center of 14a interacts with the
carboxylate group of Asp148 in TM3 through an ionic bond,21,22
a variety of conformations of 14a were generated and manually
docked into the receptor model. (2) The conformations of 14a

Identification of a NoVel Class of M3 Antagonists
Journal of Medicinal Chemistry, 2006, Vol. 49, No. 19 5657
Table 3. Binding Profiles of the Compounds with the Newly Identified Diamine Core Parts and Various Substituents on the Triphenylpropionamide
Part
a Ki values are expressed as the mean ( standard error of the mean (SEM), for the compounds which were tested three or more times. b When only two
separate experiments were performed, values are expressed as the mean of duplicate determinations, and the variances between determinations were less
than (21%. c Values are expressed as the mean of duplicate determinations.
displaying reasonable interactions with three residues, Asn138,
Lys523, and Asn527, were prioritized. (3) The resultant com-
plex structures were then minimized and subjected to visual
inspections without further modification. As a result, 14a
gave a limited number of binding conformations in the receptor
model to meet the above criteria, and an example of the final
complex model is shown in Figure 4. In the complex model, in
addition to the salt bridge between the cationic N center of 14a
and the carboxylate group of Asp148, formation of a hydrogen
bonding interaction between the hydroxy group of Hyp and the
amide group in the side chain of Asn527 and interactions
between two 4-fluorophenyl groups in the triphenylpropiona-
mide part of 14a and the side chains of Asn138 and Lys523
were observed. As discussed previously, introduction of hydroxy
group to the amino acid spacer part (substitution of L-Pro with
L-Hyp) and fluoro atoms to the triphenylpropionamide part
significantly improved the binding affinity and subtype selectiv-
ity of the novel antagonist class. All three interactions predicted
between 14a and M3 receptor in our docking model might
explain the observed structure-activity relationship (SAR).
The present docking model is also in good agreement with
our hypothesis for subtype selectivity. In the complex model,
the terminal tripheylpropionamide part and Hyp-Pro spacer part
connecting the diamine core part and the terminal tripheylpro-
pionamide part of 14a were predicted to make reasonable
interactions with all three nonconserved residues among the

5658 Journal of Medicinal Chemistry, 2006, Vol. 49, No. 19
Table 4. Introduction of Piperidine N-Substituents to 14aa
Sagara et al.
a Ki values are expressed as the mean ( standard error of the mean (SEM), for the compounds which were tested three or more times. b When only two
separate experiments were performed, values are expressed as the mean of duplicate determinations, and the variances between determinations were less
than (18%. c Values are expressed as the mean of duplicate determinations.
Figure 2. Inhibition by 14a on carbachol-induced bradycardia in
isolated rat right atria. Data represent the mean ( SEM of four
experiments per each dose. The inset shows the Schild regression
analysis of the data.
Figure 1. Inhibition by 14a on contractile response induced by
carbachol in isolated rat trachea. Data represent the mean ( SEM of
four experiments per each dose. The inset shows the Schild regression
analysis of the data.
One of the possible approaches to achieve this is the site-directed
mutagenesis experiment. The mutant M3 receptors altering the
aforementioned nonconserved residues may give useful infor-
mation when their binding profiles with our M3 selective
antagonists are carefully examined. These trials in combination
with further chemical modification of this novel M3 antagonist
class may lead to rational identification of selective muscarinic
receptor antagonists and agonists.
muscarinic receptor subtypes, Asn138, Lys523, and Asn527,
simultaneously. As these substructures are not found in the
reported M3 antagonist class, they may interact specifically with
these residues to establish highly selective binding to M3
receptors.
As the complex model is preliminary and not enough to
explain the details of the SARs in our novel lead class, further
experimental research is required to validate our hypothesis.

Identification of a NoVel Class of M3 Antagonists
Journal of Medicinal Chemistry, 2006, Vol. 49, No. 19 5659
alignment of multiple muscarinic receptor subtypes and com-
putational modeling of selective recognition of M3 receptor by
our antagonist 14a. Although the model was preliminary and
not enough for clear understanding of the mechanism, it was
postulated that nonconserved residues among the muscarinic
receptor subtypes could be specifically recognized by highly
selective M3 antagonists. Finally, as this class of M3 antagonists
is potent and highly subtype-selective, pharmacological evalu-
ation of this class of compounds to clarify the roles of M3
receptors and, furthermore, examination of their potential as lead
structures for orally available drugs are also in progress. These
results will be reported elsewhere.
Experimental Section
General Information. All reagents and solvents were of
commercial quality and used without further purification unless
otherwise noted. Melting points were determined with a Yanaco
MP micromelting point apparatus and were not corrected. 1H NMR
spectra were obtained on a Varian Gemini-300 with tetramethyl-
silane as an internal standard. Mass spectrometry was performed
with a JEOL JMS-SX 102A. TLC was done using Merck Kieselgel
F254 precoated plates. Silica gel column chromatography was carried
out on Wako gel C-200. The purity of all target compounds was
checked using reversed phase HPLC (C18 column: YMC-Pack Pro
C18, 4.6 × 250 mm, s-5um) with two diverse solvent systems:
system Ascompounds were eluted using a linear gradient of 10-
90% B/A over 30 min at a flow rate of 1.0 mL/min, where solvent
A was aqueous 0.016% H2SO4 and solvent B was CH3CN; system
Bscompounds were eluted using a linear gradient of 20-95% B/A
over 30 min at a flow rate of 1.0 mL/min, where solvent A was
aqueous 0.016% H2SO4 and solvent B was CH3OH.
Figure 3. Alignment of the sequences of bovine rhodopsin (OPSD-
_BOVIN) and the human muscarinic receptors (ACM_HUMAN).
General Procedure for the Reaction of Resin Bound Amines
with the Carboxylic Acid 3 and Cleavage of Product 4 from
Solid Support. The resin bound amine (ca. 20 mg, ca. 0.020 mmol)
was suspended in DMF (300 µL) in a reserver with a frit, and
carboxylic acid 3 (29.8 mg, 0.060 mmol), PyBOP (31.2 mg, 0.060
mmol), and DIEA (20.9 µL, 0.12 mmol) were added. The mixture
was left on the orbital shaker for 5 h. DMF (200 µL) was added to
the mixture, and the mixture was left on the orbital shaker for
additional 18 h. Reaction completion was determined by Kaiser
ninhydrin test.23 The resin was filtered, washed successively with
DMF, H2O, MeOH, and dichloromethane (DCM) (3 times each),
and dried in vacuo. The dried resin was then treated with 50%
TFA-DCM (1 mL) for 40 min at room temperature, and the elute
was concentrated under reduced pressure. The residue was dissolved
in MeOH (1 mL), again concentrated, and dried in vacuo to obtain
crude sample of 4. A total of 28 out of 33 samples showed higher
than 90% purity by LC-MS and HPLC spec, and even the lowest
purity was found to be 71%. The products were estimated as a
TFA salt (coefficient ) 1) and were dissolved in DMSO for the
binding assay. Because of high-throughput synthesis, it was difficult
to determine the precise yields of the reaction, but generally 0.020
mmol of the starting materials (resin bound amines) gave greater
than 0.015 mmol (75% yield) of the target molecule (4).
Methyl (2R)-(Benzyloxycarbonyl)pyrrolidine-2-carboxylate
(6). To a solution of D-benzyloxycarbonylproline 5 (7.5 g, 30 mmol)
and DMAP (82 mg, 0.67 mmol) in CHCl3 (20 mL) and MeOH (10
mL) was added WSC-HCl (6.9 g, 36 mmol), and the mixture was
stirred at room temperature for 1 h. The reaction was quenched by
adding aqueous NaHCO3 solution, and the mixture was extracted
with CHCl3. The organic phase was washed with H2O, dried over
Na2SO4, and evaporated. The residue was purified by silica gel
column chromatography (Hex-EtOAc, 6:1-4:1 elution) to give 6
(6.20 g, 23.5 mmol, 78%) as a colorless oil.
Figure 4. Three-dimensional model for binding of 14a (magenta) in
M3 receptor. Amino acid residues involved in ligand binding are
indicated.
Conclusion
A novel triphenylpropionamide class of M3 selective antago-
nists with L-Pro-D-Pro spacer was chemically modified in a
stepwise manner. In the first step, we turned our attention to
the modification of an important diamine core part. Using solid
phase synthesis, we identified the compounds with the novel
piperidine pharmacophores showing significantly improved M3
binding affinity and subtype selectivity.
Introduction of chlorine or fluorine atoms onto the phenyl
rings of the terminal triphenylpropionamide part further en-
hanced M3 binding affinity and subtype selectivity. Among
them, 14a achieved sub-nanomolar M3 binding affinity and
>100-fold selectivity versus the M1, M2, M4, and M5 receptors.
The functional assay using isolated rat tissue suggested that 14a
was a competitive antagonist, and high M2/M3 functional
selectivity was also observed.
Methyl (2R)-Pyrrolidine-2-carboxylate (7). A suspension of
6 (6.20 g, 23.5 mmol) and Pd(OH)2 (Degussa type, 320 mg) in
MeOH (10 mL) was stirred vigorously at room temperature for 4
h under H2 atmosphere. After adding Pd(OH)2 (200 mg), the
The mechanism for the observed selectivity of these novel
M3 antagonists was hypothesized by amino acid sequence

5660 Journal of Medicinal Chemistry, 2006, Vol. 49, No. 19
Sagara et al.
1
suspension was stirred under the same condition for 15 h. The
reaction mixture was filtered by Whatmann GF/F and the filtrate
was evaporated to give the crude product of 7 (4.48 g) as a colorless
oil.
°C; HPLC tR ) 15.3 min (system A), 21.8 min (system B); H
NMR (CDCl3) δ 0.80-2.08 (m, 12H), 2.28-2.45 (m, 2H), 2.46-
2.57 (m, 2H), 2.98-3.12 (m, 4H), 3.32-3.40 (m, 1H), 3.44 (d, J
) 15.1 Hz, 1H), 3.59 (dd, J ) 4.1, 10.8 Hz, 1H), 3.69 (d, J ) 15.1
Hz, 1H), 3.77-3.87 (m, 1H), 4.40-4.51 (m, 2H), 4.53-4.61 (m,
1H), 6.96 (dd, JHF ) 8.8 Hz, JHH ) 8.8 Hz, 6H), 7.08-7.20 (m,
1H), 7.15 (dd, JHF ) 5.2 Hz, JHH ) 8.8 Hz, 6H); HRMS Calcd for
C37H42N4O4F3 (M + H)+: 663.3158, Found: 663.3154. Anal.
(C37H41N4O4F3‚2.5H2O) C, H, N.
Tris(4-fluorophenyl)propionitrile (17). A ZnCl2 solution (1 M
Et2O solution, 16.9 mL, 16.9 mmol) was coevaporated three times
with toluene. Cyanoacetic acid (5.41 g, 63.6 mmol) was also
coevaporated three times with toluene in another flask. A mixture
of tris(4-fluorophenyl)methanol (16, 10.0 g, 31.8 mmol), ZnCl2,
and cyanoacetic acid in acetic acid (10.9 mL, 191 mmol) was
refluxed at 130 °C for 6.5 h. After the suspension was cooled to
room temperature, H2O (ca. 30 mL) was added, and the mixture
was extracted with CHCl3. The organic phase was dried over Na2-
SO4 and evaporated. The residue was purified by silica gel column
chromatography (Hex-EtOAc, 8:0-8:1 elution) and crystallized
from Hex-EtOAc to give 17 (5.56 g, 16.5 mmol, 52%) as a white
solid.
Tris(4-fluorophenyl)propionic Acid (10). To a suspension of
17 (5.56 g, 16.5 mmol) in acetic acid (50 mL) were added
concentrated hydrocholoric acid (20 mL) and concentrated sulfuric
acid (10 mL), and the mixture was refluxed at 130 °C for 44 h.
The reaction mixture was cooled slowly to 0 °C, and the resultant
precipitate was filtered off and washed with H2O. The precipitate
was dissolved in CHCl3, and the organic layer was washed with
H2O (×2), dried over Na2SO4, and evaporated to give 10 (5.26 g,
14.8 mmol, 90%) as a pale yellow solid.
Compounds 2a, 51a-51c, 52a-52c, 14b, and 14c were prepared
according to the synthesis of 14a, using the appropriate triphenyl-
propionic acids (commercially available triphenylpropionic acid or
commercially available tris(4-chlorophenyl)propionic acid), and
appropriate diamine core parts ((3S)-3-aminomethyl-1-tert-butoxy-
carbonylpiperidine,24 4-aminomethyl-tert-butoxycarbonylpiperi-
dine,14 or 4-aminoethyl-tert-butoxycarbonylpiperidine.25).
(2R)-1-[(2S,4R)-4-Hydroxy-1-(3,3,3-triphenylpropanoyl)pyr-
rolidine-2-yl]carbonyl-N-((3S)-3-piperidinylmethyl)pyrrolidine-
2-carboxamide (2a). mp 92-95 °C; HPLC tR ) 14.6 min (system
A), 20.5 min (system B); 1H NMR (CDCl3) δ 0.93-1.08 (m, 1H),
1.32-1.50 (m, 1H), 1.51-2.05 (m, 10H), 2.09 (t, J ) 11.1 Hz,
1H), 2.23-2.32 (m, 1H), 2.51 (dt, J ) 2.3, 9.6 Hz, 1H), 2.57-
2.68 (m, 1H), 2.74 (d, J ) 10.9 Hz, 1H), 2.85-3.02 (m, 3H), 3.30-
3.33 (m, 1H), 3.42 (d, J ) 14.6 Hz, 1H), 3.58 (dd, J ) 4.1, 11.0
Hz, 1H), 3.80-3.87 (m, 1H), 3.90 (d, J ) 14.6 Hz, 1H), 4.30-
4.38 (m, 2H), 4.56 (d, J ) 6.9 Hz, 1H), 7.17-7.32 (m, 16H); HRMS
Calcd for C37H45N4O4 (M + H)+: 609.3441, Found: 609.3420.
(2R)-1-[(2S,4R)-4-Hydroxy-1-(3,3,3-triphenylpropanoyl)pyr-
rolidine-2-yl]carbonyl-N-(4-piperidinylmethyl)pyrrolidine-2-car-
boxamide (51a). mp 95.0-98.5 °C; HPLC tR ) 14.1 min (system
A), 19.9 min (system B); 1H NMR (CDCl3) δ 1.18-1.72 (m, 6H),
1.75-2.09 (m, 5H), 2.19-2.30 (m, 1H), 2.57-2.76 (m, 3H), 2.86
(d, J ) 10.6 Hz, 1H), 3.00-3.12 (m, 1H), 3.18-3.40 (m, 3H),
3.52 (d, J ) 14.8 Hz, 1H), 3.57 (dd, J ) 4.0, 11.0 Hz, 1H), 3.78
(d, J ) 14.8 Hz, 1H), 3.79-3.86 (m, 2H), 4.30-4.40 (m, 2H),
4.50-4.59 (m, 1H), 7.13-7.32 (m, 15H), 7.33-7.41 (m, 1H);
HRMS Calcd for C37H45N4O4 (M + H)+: 609.3441, Found:
609.3420.
Methyl (2R)-1-([(2S,4R)-4-tert-Butoxy-1-(benzyloxycarbonyl)-
pyrrolidin-2-yl]carbonyl)pyrrolidine-2-carboxylate (8). To a
solution of 7 (4.48 g) and N-R-benzyloxycarbonyl-O-tert-butyl-L-
4-trans-hydroxyproline (7.55 g, 23.5 mmol) in CHCl3 (30 mL) were
added WSC-HCl (5.39 g, 28.2 mmol) and HOBt (4.31 g, 28.2
mmol), and the mixture was stirred for 22 h at room temperature.
The reaction was quenched by adding aqueous NaHCO3, and the
mixture was extracted with CHCl3. The organic phase was washed
with saturated aqueous NH4Cl, dried over Na2SO4, and evaporated.
The residue was purified by silica gel column chromatography
(CHCl3 elution) to give 8 (5.10 g, 11.8 mmol, 50% from 6) as a
colorless oil.
Methyl (2R)-1-([(2S,4R)-4-tert-Butoxy-1-pyrrolidin-2-yl]car-
bonyl)pyrrolidine-2-carboxylate (9). The suspension of 8 (5.10
g, 11.8 mmol) and Pd(OH)2 (Degussa type, 400 mg) in MeOH (12
mL) was stirred vigorously at room temperature for 3 h under H2
atmosphere. After adding Pd(OH)2 (200 mg), the suspension was
stirred under the same condition for 3 h. The reaction mixture was
filtered by Whatmann GF/F, and the filtrate was evaporated to give
the crude product of 9 (3.54 g) as a white solid.
Methyl (2R)-1-[((2S,4R)-4-tert-Butoxy-1-[3,3,3-tris(4-fluo-
rophenyl)propanoyl]pyrrolidine-2-yl)carbonyl]pyrrolidine-2-
carboxylate (11). To a solution of 9 (2.84 g), 10 (3.39 g, 9.52
mmol), and DIEA (1.99 mL, 11.4 mmol) in CHCl3 (40 mL) were
added WSC-HCl (2.18 g, 11.4 mmol) and HOBt (1.74 g, 11.4
mmol), and the mixture was stirred at room temperature for 13 h.
The reaction was quenched by adding aqueous NaHCO3, and the
mixture was extracted with CHCl3. The organic phase was washed
with saturated aqueous NH4Cl, dried over Na2SO4, and evaporated.
The residue was purified by silica gel column chromatography
(CHCl3 elution) to give 11 (5.17 g, 8.12 mmol, 86% from 8) as a
white foam.
(2R)-1-[((2S,4R)-4-tert-Butoxy-1-[3,3,3-tris(4-fluorophenyl)-
propanoyl]pyrrolidine-2-yl)carbonyl]pyrrolidine-2-carboxylic Acid
(12). To a solution of 11 (3.14 g, 4.93 mmol) in MeOH (10 mL)
was added aqueous NaOH solution (4 mol/L, 3 mL), and the
mixture was stirred at room temperature for 40 min. The reac-
tion mixture was acidified to pH ) 1 with aqueous HCl and
extracted with CHCl3. The organic phase was dried over Na2SO4
and evaporated to give the crude product 12 (3.24 g) as a white
foam.
(2R)-1-((2S,4R)-4-tert-Butoxy-1-[3,3,3-tris(4-fluorophenyl)pro-
panoyl]pyrrolidine-2-yl)carbonyl-N-[4-(tert-butoxycarbonyl)pi-
peridinylmethyl]pyrrolidine-2-carboxamide (13). To a solution
of 12 (3.24 g) and 4-aminomethyl-tert-butoxycarbonylpiperidine14
(1.16 g, 5.42 mmol) in CHCl3 (20 mL) were added WSC-HCl
(1.13 g, 5.92 mmol) and HOBt (0.91 g, 5.9 mmol), and the mixture
was stirred at room temperature for 22 h. The reaction was quenched
by adding aqueous NaHCO3, and the mixture was extracted with
CHCl3. The organic phase was washed with saturated aqueous NH4-
Cl, dried over Na2SO4, and evaporated. The residue was purified
by silica gel column chromatography (CHCl3-MeOH, 500:3-500:6
elution) to give 13 (3.43 g, 4.19 mmol, 85% from 12) as a white
foam.
(2R)-1-((2S,4R)-4-Hydroxy-1-[3,3,3-tris(4-fluorophenyl)pro-
panoyl]pyrrolidine-2-yl)carbonyl-N-(4-piperidinylmethyl)pyr-
rolidine-2-carboxamide (14a). TFA (4 mL) was added to 13 (3.40
g, 4.15 mmol) at 0 °C, and the mixture was stirred at room
temperature for 1 h. To the reaction mixture was added TFA (2
mL), and the mixture was stirred at room temperature for an
additional 2 h. The reaction mixture was basified to pH ) 10 with
aqueous NaOH solution and extracted with CHCl3. The organic
phase was washed with H2O (×2), and dried over Na2SO4. The
solvent was evaporated to give the crude product 14a (2.67 g,
including H2O) as a white foam. The analytical sample was prepared
by drying the crude product in vacuo at 45 °C: mp 144.5-146.0
(2R)-1-[(2S,4R)-4-Hydroxy-1-(3,3,3-triphenylpropanoyl)pyr-
rolidine-2-yl]carbonyl-N-(4-piperidinylethyl)pyrrolidine-2-car-
boxamide (51b). HPLC tR ) 14.1 min (system A), 19.9 min
1
(system B); H NMR (CDCl3) δ 1.02-2.40 (m, 15H), 2.52-2.64
(m, 2H), 2.65-2.78 (m, 1H), 2.85 (d, J ) 10.8 Hz, 1H), 3.02-
3.11 (m, 2H), 3.12-3.24 (m, 1H), 3.29-3.40 (m, 1H), 3.45 (d, J
) 14.6 Hz, 1H), 3.58 (dd, J ) 4.0, 10.9 Hz, 1H), 3.78-3.82 (m,
1H), 3.83 (d, J ) 14.6 Hz, 1H), 4.32-4.40 (m, 2H), 4.52-4.59
(m, 1H), 7.12-7.31 (m, 16H); HRMS Calcd for C38H47N4O4 (M
+ H)+: 623.3597, Found: 623.3586.

Identification of a NoVel Class of M3 Antagonists
Journal of Medicinal Chemistry, 2006, Vol. 49, No. 19 5661
(2R)-1-[(2S,4R)-4-Hydroxy-1-(3,3,3-triphenylpropanoyl)pyr-
rolidine-2-yl]carbonyl-N-((3R)-3-piperidinylmethyl)pyrrolidine-
2-carboxamide (51c). mp 86-89 °C; HPLC tR ) 14.3 min (system
A), 20.1 min (system B); 1H NMR (CDCl3) δ 0.88-1.05 (m, 1H),
1.17-1.43 (m, 2H), 1.45-2.08 (m, 9H), 2.21 (t, J ) 11.0 Hz, 1H),
2.27-2.36 (m, 1H), 2.46-2.58 (m, 2H), 2.78-2.99 (m, 3H), 3.00-
3.12 (m, 1H), 3.30-3.40 (m, 1H), 3.44 (d, J ) 14.7 Hz, 1H), 3.62
(dd, J ) 4.0, 10.9 Hz, 1H), 3.78-3.85 (m, 1H), 3.86 (d, J ) 14.7
Hz, 1H), 4.30-4.40 (m, 2H), 4.55-4.62 (m, 1H), 7.15-7.38 (m,
16H); HRMS Calcd for C37H45N4O4 (M + H)+: 609.3441,
Found: 609.3416.
(2R)-1-((2S,4R)-4-Hydroxy-1-[3,3,3-tris(4-chlorophenyl)pro-
panoyl]pyrrolidine-2-yl)carbonyl-N-(4-piperidinylmethyl)pyr-
rolidine-2-carboxamide (52a). mp 138-140 °C; HPLC tR ) 17.9
min (system A), 25.3 min (system B); 1H NMR (CDCl3) δ 0.83-
2.08 (m, 12H), 2.29-2.45 (m, 2H), 2.46-2.58 (m, 2H), 2.98-
3.16 (m, 4H), 3.30-3.38 (m, 1H), 3.39 (d, J ) 15.0 Hz, 1H), 3.63
(dd, J ) 4.1, 10.8 Hz, 1H), 3.71 (d, J ) 15.0 Hz, 1H), 3.76-3.86
(m, 1H), 4.38-4.51 (m, 2H), 4.58 (d, J ) 5.9 Hz, 1H), 7.05-7.20
(m, 1H), 7.12 (d, J ) 8.8 Hz, 6H), 7.24 (d, J ) 8.8 Hz, 6H); HRMS
Calcd for C37H42N4O4Cl3 (M + H)+: 711.2272, Found: 711.2261.
(2R)-1-((2S,4R)-4-Hydroxy-1-[3,3,3-tris(4-chlorophenyl)pro-
panoyl]pyrrolidine-2-yl)carbonyl-N-(4-piperidinylethyl)pyrro-
lidine-2-carboxamide (52b). mp 124-127 °C; HPLC tR ) 17.9
min (system A), 25.4 min (system B); 1H NMR (CDCl3) δ 0.95-
1.88 (m, 10H), 1.90-2.00 (m, 2H), 2.01-2.12 (m, 2H), 2.28-
2.38 (m, 1H), 2.46-2.60 (m, 3H), 2.98-3.11 (m, 3H), 3.20-3.38
(m, 2H), 3.40 (d, J ) 15.4 Hz, 1H), 3.63 (dd, J ) 4.1, 11.0 Hz,
1H), 3.71 (d, J ) 15.4 Hz, 1H), 3.73-3.82 (m, 1H), 4.39-4.52
(m, 2H), 4.53-4.60 (m, 1H), 6.98-7.05 (m, 1H), 7.10 (d, J ) 8.8
Hz, 6H), 7.23 (d, J ) 8.8 Hz, 6H); HRMS Calcd for C38H44N4O4-
Cl3 (M + H)+: 725.2428, Found: 725.2430.
Method A (Reductive Amination): (2R)-1-((2S,4R)-4-Hy-
droxy-1-[3,3,3-tris(4-fluorophenyl)propanoyl]pyrrolidine -2-yl)-
carbonyl-N-(1-methyl-4-piperidinylmethyl)pyrrolidine-2-car-
boxamide (15a). To a solution of 14a (202 mg, 0.305 mmol) and
formaldehyde (37% aqueous solution 200 µL) in MeOH (2 mL)
was added NaBH3CN-ZnCl2 (0.3 mol/L, 1.22 mL),15 and the
mixture was stirred at room temperature for 20 min. The reaction
was quenched by adding saturated aqueous NaHCO3 solution, and
extracted with EtOAc. The organic phase was washed with brine,
dried (NaSO4), and evaporated. The residue was purified by
preparative TLC (CHCl3-MeOH-28% aqueous NH3 solution, 8:1:
0.1 elution) to give 15a (184 mg, 0.272 mmol, 89%) as a white
foam: mp 140-142 °C; HPLC tR ) 15.5 min (system A), 21.6
min (system B); 1H NMR (CDCl3) δ 1.12-2.20 (m, 13H), 2.31 (s,
3H), 2.32-2.39 (m, 1H), 2.48-2.58 (m, 1H), 2.83-2.94 (m, 2H),
2.99-3.10 (m, 2H), 3.35-3.42 (m, 1H), 3.45 (d, J ) 15.0 Hz,
1H), 3.62 (dd, J ) 4.1, 11.0 Hz, 1H), 3.73 (d, J ) 15.0 Hz, 1H),
3.78-3.88 (m, 1H), 4.40-4.53 (m, 2H), 4.59 (d, J ) 5.9 Hz, 1H),
6.92-7.03 (m, 6H), 7.08-7.25 (m, 7H); HRMS Calcd for
C38H44N4O4F3 (M + H)+: 677.3315, Found: 677.3312. Anal.
(C38H43N4O4F3‚H2O) C, H, N.
Method B (Alkylation): (2R)-N-(1-Cyclopentyl-4-piperidi-
nylmethyl)-1-((2S,4R)-4-hydroxy-1-[3,3,3-tris(4-fluorophenyl)-
propanoyl]pyrrolidine-2-yl)carbonylpyrrolidine-2-carboxam-
ide (15g). To a solution of 14a (19.7 mg, 0.0297 mmol) in CH3CN
(1.5 mL) were added bromocyclopentane (32 µL, 0.297 mmol) and
K2CO3 (12 mg, 0.089 mmol), and the mixture was stirred at 80 °C
for 15 h. The reaction was quenched by adding H2O, and extracted
with CHCl3. The organic phase was washed with diluted aqueous
NaOH solution, dried (Na2SO4), and evaporated. The residue was
purified by preparative TLC (CHCl3-MeOH-28% aqueous NH3
solution, 10:1:0.1 elution) to give 15g (20 mg, 0.027 mmol, 90%)
as a white foam; HPLC tR ) 16.9 min (system A), 22.5 min (system
B); 1H NMR (CDCl3) δ 1.08-2.20 (m, 21H), 2.29-2.65 (m, 3H),
3.02-3.18 (m, 4H), 3.30-3.92 (m, 3H), 3.42 (d, J ) 15.2 Hz,
1H), 3.71 (d, J ) 15.2 Hz, 1H), 4.39-4.67 (m, 3H), 6.90-7.27
(m, 13H); HRMS Calcd for C42H50N4O4F3 (M + H)+: 731.3784,
Found: 731.3799.
(2R)-1-((2S,4R)-4-Hydroxy-1-[3,3,3-tris(4-chlorophenyl)pro-
panoyl]pyrrolidine-2-yl)carbonyl-N-((3R)-3-piperidinylmethyl)-
pyrrolidine-2-carboxamide (52c). mp 132-135 °C; HPLC tR )
1
18.6 min (system A), 25.8 min (system B); H NMR (CDCl3) δ
0.83-1.82 (m, 8H), 1.88-2.10 (m, 4H), 2.13-2.22 (m, 1H), 2.28-
2.40 (m, 2H), 2.48-2.59 (m, 1H), 2.80-2.88 (m, 1H), 2.90-3.07
(m, 2H), 3.08-3.20 (m, 1H), 3.29-3.39 (m, 1H), 3.39 (d, J )
15.3 Hz, 1H), 3.60-3.66 (m, 1H), 3.73 (d, J ) 15.3 Hz, 1H), 3.77-
3.86 (m, 1H), 4.39-4.52 (m, 2H), 4.57 (d, J ) 5.6 Hz, 1H), 7.02-
7.20 (m, 1H), 7.12 (d, J ) 8.9 Hz, 6H), 7.24 (d, J ) 8.9 Hz, 6H);
HRMS Calcd for C37H42N4O4Cl3 (M + H)+: 711.2272, Found:
711.2269.
(2R)-N-(1-Ethyl-4-piperidinylmethyl)-1-((2S,4R)-4-hydroxy-
1-[3,3,3-tris(4-fluorophenyl)propanoyl]pyrrolidine-2-yl)carbo-
nylpyrrolidine-2-carboxamide (15b). Yield 52% (Method B); mp
97-99 °C; HPLC tR ) 15.8 min (system A), 21.6 min (system B);
1H NMR (CDCl3) δ 1.03-2.15 (m, 16H), 2.29-2.48 (m, 3H),
2.50-2.60 (m, 1H), 2.86-2.96 (m, 2H), 2.97-3.10 (m, 2H), 3.30-
3.40 (m, 1H), 3.42 (d, J ) 15.1 Hz, 1H), 3.63 (dd, J ) 3.9, 11.0
Hz, 1H), 3.72 (d, J ) 15.1 Hz, 1H), 3.78-3.88 (m, 1H), 4.40-
(2R)-1-((2S,4R)-4-Hydroxy-1-[3,3,3-tris(4-fluorophenyl)pro-
panoyl]pyrrolidine-2-yl)carbonyl-N-(4-piperidinylethyl)pyrro-
lidine-2-carboxamide (14b). HPLC tR ) 15.3 min (system A),
4.52 (m, 2H), 4.53-4.61 (m, 1H), 6.96 (dd, JHF ) 8.8 Hz, JHH
8.8 Hz, 6H), 7.07-7.20 (m, 1H), 7.16 (dd, JHF ) 5.3 Hz, JHH
)
)
1
21.7 min (system B); H NMR (CDCl3) δ 0.85-1.86 (m, 10H),
8.8 Hz, 6H); HRMS Calcd for C39H46N4O4F3 (M + H)+: 691.3471,
Found: 691.3475.
1.88-2.00 (m, 2H), 2.00-2.10 (m, 2H), 2.27-2.36 (m, 1H), 2.47-
2.60 (m, 3H), 2.98-3.06 (m, 2H), 3.07-3.14 (m, 1H), 3.15-3.28
(m, 1H), 3.30-3.41 (m, 1H), 3.45 (d, J ) 15.5 Hz, 1H), 3.60 (dd,
J ) 4.2, 10.6 Hz, 1H), 3.70 (d, J ) 15.5 Hz, 1H), 3.75-3.85 (m,
1H), 4.41-4.52 (m, 2H), 4.53-4.60 (m, 1H), 6.94 (dd, JHF ) 8.5
Hz, JHH ) 8.5 Hz, 6H), 7.00-7.10 (m, 1H), 7.13 (dd, JHF ) 5.3
Hz, JHH ) 8.5 Hz, 6H); HRMS Calcd for C38H44N4O4F3 (M +
H)+: 677.3315, Found: 677.3314.
(2R)-N-(1-Cyclopropylmethyl-4-piperidinylmethyl)-1-((2S,4R)-
4-hydroxy-1 -[3,3,3-tris(4-fluorophenyl)propanoyl]pyrrolidine-
2-yl)carbonylpyrrolidine-2-carboxamide (15c). Yield 91% (Method
A); mp 103.5-106 °C; HPLC tR ) 16.6 min (system A), 22.3 min
1
(system B); H NMR (CDCl3) δ 0.06-0.20 (m, 2H), 0.48-0.60
(m, 2H), 0.80-0.95 (m, 2H), 1.10-2.13 (m, 11H), 2.22-2.40 (m,
3H), 2.48-2.58 (m, 1H), 2.98-3.20 (m, 4H), 3.30-3.52 (m, 1H),
3.43 (d, J ) 15.3 Hz, 1H), 3.58-3.90 (m, 3H), 3.72 (d, J ) 15.3
Hz, 1H), 4.40-4.64 (m, 3H), 6.96 (dd, JHF ) 8.7 Hz, JHH ) 8.7
Hz, 6H), 7.07-7.26 (m, 1H), 7.15 (dd, JHF ) 5.2 Hz, JHH ) 8.7
Hz, 6H); HRMS Calcd for C41H48N4O4F3 (M + H)+: 717.3628,
Found: 717.3606.
(2R)-1-((2S,4R)-4-Hydroxy-1-[3,3,3-tris(4-fluorophenyl)pro-
panoyl]pyrrolidine-2-yl)carbonyl-N-[(3R)-3-piperidinylmethyl]-
pyrrolidine-2-carboxamide (14c). mp 137.0-139.5 °C; HPLC tR
) 15.8 min (system A), 22.1 min (system B); 1H NMR (CDCl3) δ
0.85-2.10 (m, 12H), 2.12-2.23 (m, 1H), 2.25-2.40 (m, 2H),
2.47-2.58 (m, 1H), 2.80-2.90 (m, 1H), 2.91-3.00 (m, 1H), 3.04-
3.18 (m, 2H), 3.30-3.42 (m, 1H), 3.45 (d, J ) 15.4 Hz, 1H), 3.62
(dd, J ) 4.3, 10.9 Hz, 1H), 3.72 (d, J ) 15.4 Hz, 1H), 3.77-3.85
(2R)-N-(1-Cyclobutylmethyl-4-piperidinylmethyl)-1-((2S,4R)-
4-hydroxy-1-[3,3,3-tris(4-fluorophenyl)propanoyl]pyrrolidine-
2-yl)carbonylpyrrolidine-2-carboxamide (15d). Yield 84% (Method
(m, 1H), 4.40-4.50 (m, 2H), 4.52-4.60 (m, 1H), 6.96 (dd, JHF
8.9 Hz, JHH ) 8.9 Hz, 6H), 7.09-7.20 (m, 1H), 7.15 (dd, JHF
)
)
1
B); HPLC tR ) 17.2 min (system A), 22.9 min (system B); H
5.3 Hz, JHH ) 8.9 Hz, 6H); HRMS Calcd for C37H42N4O4F3 (M +
H)+: 663.3158, Found: 663.3168. Anal. (C37H41N4O4F3‚3H2O) C,
H, N.
NMR (CDCl3) δ 0.82-2.10 (m, 20H), 2.28-2.44 (m, 2H), 2.45-
2.60 (m, 2H), 2.76-2.92 (m, 2H), 2.97-3.08 (m, 2H), 3.30-3.40
(m, 1H), 3.41 (d, J ) 15.1 Hz, 1H), 3.57-3.67 (m, 1H), 3.72 (d,
J ) 15.1 Hz, 1H), 3.78-3.87 (m, 1H), 4.39-4.52 (m, 2H), 4.54-
3.61 (m, 1H), 6.96 (dd, JHF ) 8.8 Hz, JHH ) 8.8 Hz, 6H), 7.06-
N-Alkylated tertiary amine derivatives, 15a-15g, were prepared
according to either method A or method B.

5662 Journal of Medicinal Chemistry, 2006, Vol. 49, No. 19
Sagara et al.
7.20 (m, 1H), 7.15 (dd, JHF ) 5.3 Hz, JHH ) 8.8 Hz, 6H); HRMS
Calcd for C42H50N4O4F3 (M + H)+: 731.3784, Found: 731.3797.
(2R)-N-(1-Cyclopentylmethyl-4-piperidinylmethyl)-1-((2S,4R)-
4-hydroxy-1-[3,3,3-tris(4-fluorophenyl)propanoyl]pyrrolidine-
2-yl)carbonylpyrrolidine-2-carboxamide (15e). Yield 68% (Method
a table listing the elemental analysis results, spectroscopic data for
the synthetic intermediates 6-13 and 17, and solid phase im-
mobilization of diverse diamine reagents using 2-chlorotrityl
chloride resin (Cl-Trt(2-Cl)-resin). This material is available free
of charge via the Internet at http://pubs.acs.org.
1
B); HPLC tR ) 17.8 min (system A), 23.4 min (system B); H
References
NMR (CDCl3) δ 1.07-2.50 (m, 26H), 2.82-2.98 (m, 2H), 3.00-
3.12 (m, 2H), 3.32-3.40 (m, 1H), 3.44 (d, J ) 15.2 Hz, 1H), 3.62
(dd, J ) 4.3, 10.8 Hz, 1H), 3.72 (d, J ) 15.2 Hz, 1H), 3.78-3.88
(1) (a) Kubo, T.; Fukuda, K.; Mikami, A.; Maeda, A.; Takahashi, H.;
Mishina, M.; Haga, T.; Haga, K.; Ichiyama, A.; Kangawa, K.;
Kojima, M.; Matsuo, H.; Hirose, T.; Numa, S. Cloning, sequencing
and expression of complementary DNA encoding the muscarinic
acetylcholine receptor. Nature 1986, 323, 411-416. (b) Kubo, T.;
Maeda, A.; Sugimoto, K.; Akiba, I.; Mikami, A.; Takahashi, H.;
Haga, T.; Haga, K.; Ichiyama, A.; Kangawa, K.; Matsuo, H.; Hirose,
T.; Numa, S. Primary structure of porcine cardiac muscarinic
acetylcholine receptor deduced from the cDNA sequence. FEBS Lett.
1986, 209, 367-372. (c) Peralta, E. G.; Ashkenazi, A.; Winslow, J.
W.; Smith, D. H.; Ramachandran, J.; Capon, D. J. Distinct primary
structures, ligand-binding properties and tissue-specific expression
of four human muscarinic acetylcholine receptors. EMBO J. 1987,
6, 3923-3929. (d) Bonner, T. I.; Buckley, N. J.; Young, A. C.; Brann,
M. R. Identification of a family of muscarinic acetylcholine receptor
genes. Science 1987, 237, 527-532. (e) Bonner, T. I.; Young, A.
C.; Brann, M. R.; Buckley, N. J. Cloning and expression of the human
and rat m5 muscarinic acetylcholine receptor genes. Neuron 1988,
1, 403-410.
(m, 1H), 4.40-4.52 (m, 2H), 4.53-4.61 (m, 1H), 6.96 (dd, JHF
8.8 Hz, JHH ) 8.8 Hz, 6H), 7.07-7.22 (m, 1H), 7.15 (dd, JHF
)
)
5.2 Hz, JHH ) 8.8 Hz, 6H); HRMS Calcd for C43H52N4O4F3 (M +
H)+: 745.3941, Found: 745.3953.
(2R)-1-((2S,4R)-4-Hydroxy-1-[3,3,3-tris(4-fluorophenyl)pro-
panoyl]pyrrolidine-2-yl)carbonyl-N-[1-(2-propyl)-4-piperidinyl-
methyl]pyrrolidine-2-carboxamide (15f). Yield quant. (Method
1
B); HPLC tR ) 16.1 min (system A), 21.8 min (system B); H
NMR (CDCl3) δ 0.82-2.17 (m, 13H), 1.06 (d, J ) 6.2 Hz, 6H),
2.28-2.37 (m, 1H), 2.48-2.59 (m, 1H), 2.66-2.93 (m, 3H), 2.96-
3.10 (m, 2H), 3.32-3.41 (m, 1H), 3.43 (d, J ) 15.0 Hz, 1H), 3.63
(dd, J ) 4.3, 10.8 Hz, 1H), 3.73 (d, J ) 15.0 Hz, 1H), 3.78-3.88
(m, 1H), 4.38-4.51 (m, 2H), 4.57 (d, J ) 6.7 Hz, 1H), 6.96 (dd,
JHF ) 8.8 Hz, JHH ) 8.8 Hz, 6H), 7.08-7.20 (m, 1H), 7.16 (dd,
JHF ) 5.4 Hz, JHH ) 8.8 Hz, 6H); HRMS Calcd for C40H48N4O4F3
(M + H)+: 705.3628, Found: 705.3633.
(2) Caulfield, M. P.; Birdsall, N. J. M. International union of pharmacol-
ogy. XII. Classification of muscarinic acetylcholine receptors.
Pharmacol. ReV. 1998, 50, 279-290.
Alignment of Muscarinic Receptor Subtypes. Sequence align-
ment of the five human muscarinic receptor subtypes with bovine
rhodopsin was accomplished as shown below. Aligned muscarine
subfamilies were downloaded from GPCR database (http://tinygra-
p.uit.no/).26 The alignment of rhodopsin and all muscurinic receptor
subtypes was generated manually by aligning key class A sequence
motifs (shown in bold in Figure 3). The lengths of the TM helices
were defined as reported by Bondensgaard et al.,27 and the sequence
identity between M3 and rhodopsin sequences in the transmembrane
helical regions is 20.9%.
Receptor Modeling and Docking. All molecular modeling
studies were carried out on a Silicon Graphics Octane 2 workstation.
The homology model of the 7 TM domain for the human M3
receptor was constructed using the homology module in In-
sight2000.28 The bovine rhodopsin crystal structure (PDB entry
1F88)29 was used as template structure for homology modeling.
Loop regions as well as the C- and N-terminus were not included
in the models as these regions are impossible to model accurately.
The side chains were minimized to avoid steric clashes using the
CHARMm force field as implemented in QUANTA.30
(3) Wess, J. Muscarinic acetylcholine receptor knockout mice: novel
phenotypes and clinical implications. Annu. ReV. Pharmacol. Toxicol.
2004, 44, 423-450. Wess, J. Novel insights into muscarinic
acetylcholine receptor function using gene targeting technology.
Trends Pharmacol. Sci. 2003, 24, 414-420. Nakamura, T.; Matusi,
M.; Uchida, K.; Futatsugi, A.; Kusakawa, S.; Matsumoto, N.;
Nakamura, K.; Manabe, T.; Taketo, M. M.; Mikoshiba, K. M3
muscarinic acetylcholine receptor plays a critical role in parasym-
pathetic control of salivation in mice. J. Physiol. 2004, 558, 561-
575.
(4) (a) Eglen, R. M.; Choppin, A.; Dillon, M. P.; Hegde, S. Muscarinic
receptor ligands and their therapeutic potential. Curr. Opin. Chem.
Biol. 1999, 3, 426-432. (b) Andersson, K.-E. Potential benefits of
muscarinic M3 receptor selectivity. Eur. Urol., Suppl. 2002, 1, 23-
28.
(5) For example, the human M3 receptor is 73-83% homologous to the
other subtypes, M1, M2, M4, and M5 in the transmembrane region.
See the details in modeling part of Results and Discussion and
Experimental sections.
(6) Regarding M2 antagonists and M4 antagonists, highly selective ones
have been disclosed so far. See the following recent papers and the
references therein: (a) Palani, A.; Dugar, S.; Clader, J. W.; Greenlee,
W. J.; Ruperto, V.; Duffy, R. A.; Lachowicz, J. E. Isopropyl amide
derivatives of potent and selective muscarinic M2 receptor antagonists.
Bioorg. Med. Chem. Lett. 2004, 14, 1791-1794. (b) Clader, J. W.;
Billard, W.; Binch, H., III; Chen, L.-Y.; Crosby, G., Jr.; Duffy, R.
A.; Ford, J.; Kozlowski, J. A.; Lachowicz, J. E.; Li, S.; Liu, C.;
McCombie, S. W.; Vice, S.; Zhou, G.; Greenlee, W. J. Muscarinic
M2 receptor antagonists: anthranilamide derivatives with exceptional
selectivity and in vivo activity. Bioorg. Med. Chem. 2004, 12, 319-
326. (c) Bo¨hme, T. M.; Keim, C.; Kreutzmann, K.; Linder, M.;
Dingermann, T.; Dannhardt, G.; Mutschler, E.; Lambrecht. G.
Structure-activity relationships of dimethindene derivatives as new
M2-selective muscarinic receptor antagonists. J. Med. Chem. 2003,
46, 856-867. (d) Bo¨hme, T. M.; Augelli-Szafran, C. E.; Hallak, H,;
Pugsley, T.; Serpa, K.; Schwarz, R. D. Synthesis and pharmacology
of benzoxazines as highly selective antagonists at M4 muscarinic
receptors. J. Med. Chem. 2002, 45, 3094-3102.
Since the receptor structure is only a homology model, docking
was conducted by hand followed by minimization with the
CHARMm program, using the CHARMm force field. The following
criteria were employed to achieve meaningful docking modes: (1)
The ligands would form ionic interaction between the cationic N
and Asp 148 in TM3. (2) No steric clashes would happen between
any two atoms. (3) The ligands would have most favorable
conformations to make interactions with the key residues Asn138
in TM3, Lys523, and Asn527 in TM7. Side chain orientations were
adjusted to avoid steric clashes between receptor and ligand atoms
by minimization.
Binding Assay. According to the reported method,19 the binding
affinities were determined by inhibition of specific binding of [3H]-
NMS using membranes from CHO cells expressing cloned human
M1-M5 receptors.
(7) Sagara, Y.; Sagara, T.; Mase, T.; Kimura, T.; Numazawa, T.;
Fujikawa, T.; Noguchi, K.; Ohtake, N. Cyclohexylmethylpiperidi-
nyltriphenylpropioamide: a selective muscarinic M3 antagonist
discriminating against the other receptor subtypes J. Med. Chem.
2002, 45, 984-987.
Functional Assay. According to the reported method,19 antago-
nistic activities using isolated rat tissues were evaluated.
(8) Sagara, Y.; Kimura, T.; Fujikawa, T.; Noguchi, K.; Ohtake, N.
Identification of novel muscarinic M3 selective antagonists with a
conformationally restricted Hyp-Pro spacer. Bioorg. Med. Chem. Lett.
2003, 13, 57-60.
Acknowledgment. We are grateful to Ms. Nami Sakaizumi,
Mr. Hirokazu Ohsawa, Ms. Chihiro Sato, Ms. Aya Mitsuya,
Mr. Taro Yamashita, and Ms. Miyoko Miyashita for technical
support and Ms. Sandy Camburn for her valuable assistance in
preparation of the manuscript.
(9) Cl-Trt(2-Cl)-resin (1% divinylbenzene, 100-200 mesh, 1.2 mmol/
g) was purchased from Novabiochem.
(10) See Supporting Information for detailed information on solid phase
immobilization of diverse diamine reagents using 2-chlorotrityl
chloride resin (Cl-Trt(2-Cl)-resin) and its general experimental
procedures.
Supporting Information Available: A table listing the degree
of purity for all target compounds (area percent and retention time),

Identification of a NoVel Class of M3 Antagonists
Journal of Medicinal Chemistry, 2006, Vol. 49, No. 19 5663
(11) Coste, J.; Le-Nguyen, D.; Castro, B. PyBOP: A new peptide coupling
reagent devoid of toxic byproduct. Tetrahedron Lett. 1990, 31, 205-
208.
(12) See Experimental Section for detailed information.
(13) All new compounds were characterized by 1H NMR, RP-HPLC, and
mass spectra.
(14) Prugh, J. D.; Birchenough, L. A.; Egbertson, B. A simple method of
protecting a secondary amine with tert-butyloxycarbonyl (BOC) in
the presence of a primary amine. Synth. Commun. 1992, 22, 2357-
2360.
(15) Kim, S.; Oh, C. H.; Ko, J. S.; Ahn, K. H.; Kim, Y. J. Zinc-modified
cyanoborohydride as a selective reducing agent. J. Org. Chem. 1985,
50, 1927-1932.
(16) Schoor, M. Reaction of triphenylcarbinols with cyanoacetic acid.
Justus Liebigs Ann. Chem. 1963, 661, 157-164.
(17) The best hydrolysis reaction condition was found to be the following
after examination of several conditions: heating of 17 in the solution
consisting of conc HCl:conc H2SO4:AcOH ) 2:1:5 at 130 °C for 44
h.
(22) Wess, J.; Maggio, R.; Palmer, J. R.; Vogel, Z. Role of conserved
threonine and tyrosine residues in acetylcholine binding and mus-
carinic receptor activation. A study with m3 muscarinic receptor point
mutants. J. Biol. Chem. 1992, 267, 19313-19319. Fraser, C. M.;
Wang, C.-D.; Robinson, D. A.; Gocayne, J. D.; Venter, J. C. Site-
directed mutagenesis of m1 muscarinic acetylcholine receptors:
conserved aspartic acids play important roles in receptor function.
Mol. Phamacol. 1989, 36, 840-847. Curtis, C. A. M.; Wheatley,
M.; Bansal, S.; Birdsall, N. J. M.; Eveleigh, P.; Pedder, E. K.; Poyner,
D.; Hulme, E. C. Propylbenzilylcholine mustard labels an acidic
residue in transmembrane helix 3 of the muscainic receptor. J. Biol.
Chem. 1989, 264, 489-495. Kurtenbach, E.; Curtis, C. A. M.; Pedder,
E. K.; Aitken, A.; Harris, A. C. M.; Hulme, E. C. Muscarinic
acetylcholine receptors. Peptide sequencing identifies residues in-
volved in antagonist binding and disulfide bond formation. J. Biol.
Chem. 1990, 265, 13702-13708.
(23) Kaiser, E.; Colescold, D. C.; Bossinger, C. D.; Cook, P. Color test
for detection of free terminal amino groups in the solid phase
synthesis of peptides. Anal. Biochem. 1970, 34, 595-598.
(24) Hilpert, K.; Ackermann, J.; Banner, D. W.; Gast, A.; Gubernator,
K.; Hadva´ry, P.; Labler, L.; Mu¨ller, K.; Schmid, G.; Tschopp, T. B.;
Waterbeemd, H. Design and synthesis of potent and highly selective
thrombin inhibitors. J. Med. Chem. 1994, 37, 3889-3901.
(25) Askew, B. C.; Bednar, R. A.; Bednar, B.; Claremon, D. A.; Cook, J.
J.; McIntyre, C. J.; Hunt, C. A.; Gould, R. J.; Lynch, R. J.; Lynch,
J. J., Jr.; Gaul, S. L.; Stranieri, M. T.; Sitko, G. R.; Holahan, M. A.;
Glass, J. D.; Hamill, T.; Gorham. L. M.; Prueksaritanont, T.; Baldwin,
J. J.; Hartman, G. D. Non-peptide glycoprotein IIb/IIIa inhibitors.
17. Design and synthesis of orally active, long-acting non-peptide
fibrinogen receptor antagonists. J. Med. Chem. 1997, 40, 1779-1788.
(26) Kristiansen, K.; Dahl, S. G.; Edvardsen, Ø. A database of mutants
and effects of site-directed mutagenesis experiments on G protein-
coupled receptors. PROTEINS: Struct. Funct. Genet. 1996, 26, 81-
94.
(18) The solubility of 14a in assay buffer (50 mM Tris-HCl, 10 mM
MgCl2, 1 mM EDTA, pH 7.4, see ref 19 for the detailed assay
condition) was tested, and 14a was found to be soluble in the buffer
at the concentration of at least 85 µM. As the final (highest) assay
concentration was 10 µM, so 14a was thought to be soluble at any
concentration during the assay.
(19) Hirose, H.; Aoki, I.; Kimura T.; Fujikawa, T.; Numazawa, T.; Sasaki,
K.; Sato, A.; Hasegawa, T.; Nishikibe, M.; Mitsuya, M.; Ohtake,
N.; Mase, T.; Noguchi, K. Pharmacological properties of (2R)-N-
[1-(6-aminopyridin-2-ylmethyl)piperidin-4-yl]-2-[(1R)-3,3-difluoro-
cyclopentyl]-2-hydroxy-2-phenylacetamide: a novel muscarinic an-
tagonist with M2-sparing antagonistic activity. J. Pharmacol. Exp.
Ther. 2001, 297, 790-797.
(20) pKi(M3) value of 14a: 9.5.
(21) The results of the functional assay using isolated rat tissues suggested
that 14a could be competitive with carbachol, a representative
muscarinic agonist possessing a cationic amino core. The binding of
small molecule antagonists and agonists such as carbachol to
muscarinic receptors appears to be mediated by an ion-ion interaction
between the amine moiety of the ligands and an aspartic residue in
TM3 of the receptors (Asp148 in the human M3 receptors) according
to the literatures (see ref 22). Therefore, it might be reasonable to
assume that ligand-receptor interaction begins with the interaction
of the cationic amine moiety of 14a and the carboxylate group of
Asp148 in human M3 receptors.
(27) Bondensgaard, K.; Ankersen, M.; Thogersen, H.; Hansen, B. S.;
Wulff, B. S.; Bywater, R. P. Recognition of privileged structures by
G-protein coupled receptors. J. Med. Chem. 2004, 47, 888-899.
(28) Insight2000; Accelrys, Inc.: San Diego, CA.
(29) Palczewski, K.; Kumasaka, T.; Hori, T.; Behnke, C. A.; Motoshima,
H.; Fox, B. A.; Le Trong, I.; Teller, D. C.; Okada, T.; Stenkamp, R.
E.; Yamamoto, M.; Miyano, M. Crystal structure of rhodopsin: A
G protein-coupled receptor. Science 2000, 289, 739-745.
(30) Quanta/CHARMm; Accelrys, Inc.: San Diego, CA.
JM051205R