Synthesis and pharmacological characterization of O-alkynyloximes of tropinone and N-methylpiperidinone as muscarinic agonists

Article abstract of DOI:10.1021/jm9708588

A number of O-alkynyloximes of tropinone and N-methyl-4-piperidinone have been synthesized and evaluated for muscarinic activity. The affinities of these oximes were tested in homogenates of cerebral cortex, heart, and submandibulary glands from rats using [³H]pirenzepine and [³H]-N- methylscopolamine as radioligands. The oximes bind to the cortical muscarinic receptors with pK(i) values varying from 3 to 7. Higher binding affinities were observed for the O-alkynyl tropinone oximes than the corresponding piperidinone analogues. Binding to the muscarinic sites in the heart and submandibulary glands was also observed but with lower affinities. Good M₁ subtype selectivity (10-fold or greater) was observed with some oximes (26a, 28a, 32a) at the cortical sites. These oximes also attenuated scopolamine- induced impairment of the water mask task in mice. Functional assays for M₃ activity on the rat aorta showed that all oximes possessed M₃ agonist action but M₂ agonist activity was not observed at the endothelium-denuded rabbit aorta. Analysis of the quantitative structure-activity relationship (QSAR) indicated that the Connolly surface area is an important determinant of activity, accounting for 70% of the variation in cortical binding affinity among the oximes.

Full text of DOI:10.1021/jm9708588

3220
J . Med. Chem. 1998, 41, 3220-3231
Syn th esis a n d P h a r m a cologica l Ch a r a cter iza tion of O-Alk yn yloxim es of
Tr op in on e a n d N-Meth ylp ip er id in on e a s Mu sca r in ic Agon ists
Rong Xu, Meng-Kwoon Sim, and Mei-Lin Go*
Departments of Pharmacy and Pharmacology, National University of Singapore, 10 Kent Ridge Crescent,
Republic of Singapore 119260
Received December 23, 1997
A number of O-alkynyloximes of tropinone and N-methyl-4-piperidinone have been synthesized
and evaluated for muscarinic activity. The affinities of these oximes were tested in homogenates
of cerebral cortex, heart, and submandibulary glands from rats using [3H]pirenzepine and [3H]-
N-methylscopolamine as radioligands. The oximes bind to the cortical muscarinic receptors
with pKi values varying from 3 to 7. Higher binding affinities were observed for the O-alkynyl
tropinone oximes than the corresponding piperidinone analogues. Binding to the muscarinic
sites in the heart and submandibulary glands was also observed but with lower affinities. Good
M1 subtype selectivity (10-fold or greater) was observed with some oximes (26a , 28a , 32a ) at
the cortical sites. These oximes also attenuated scopolamine-induced impairment of the water
mask task in mice. Functional assays for M3 activity on the rat aorta showed that all oximes
possessed M3 agonist action but M2 agonist activity was not observed at the endothelium-
denuded rabbit aorta. Analysis of the quantitative structure-activity relationship (QSAR)
indicated that the Connolly surface area is an important determinant of activity, accounting
for 70% of the variation in cortical binding affinity among the oximes.
In tr od u ction
in their electrostatic potentials. The propargyl ether (1)
of 1-azabicyclo[2.2.1]heptane 3-aldoxime was found to
be a high-affinity partial agonist with a good central
selectivity (Chart 1). The oxime analogue (2) of UH5
(3), related to the nonselective muscarinic agonist
oxotremorine, demonstrated 5-fold greater m1 selectiv-
ity than UH5.10 The methyl oxime (4) of 1-azabicyclo-
[2.2.2]octan-3-one has been reported to possess modest
M3 agonistic selectivity.11 Tecle et al.12 reported good
M1 selectivity in the phenylpropargyl ether (5), thus
showing that it is possible to synthesize large and long
muscarinic agonists with good potency and efficacy.
Considerable research efforts have been directed to
the development of muscarinic agonists for the treat-
ment of Alzheimer’s disease (AD).1 The rationale for
this thrust is based on two complementary themes, viz.,
the cholinergic hypothesis of memory dysfunction2-4 and
the muscarinic regulation of amyloid metabolism.5 Both
hypotheses suggest that there is a role for centrally
acting muscarinic agonists in retarding the progression
of AD dementia.
Four subtypes of the muscarinic receptor have been
identified on the basis of pharmacological studies (des-
ignated M1-M4), while molecular biological studies have
identified five muscarinic receptor subtypes (m1-m5).6
The muscarinic receptor subtypes are located at differ-
ent sites in the brain, with high levels of M1 (m1)
receptors in the cortex and hippocampus, areas associ-
ated with memory and learning. Development of M1-
selective agonists as a substitution therapy in AD has
been pursued by several investigators. However, at-
taining receptor subtype selectivity among muscarinic
agonists, so critical to the development of clinically
useful candidates, has proven to be an illusive goal. No
agonist is known to discriminate between muscarinic
receptor subtypes on an affinity basis, by more than 1
order of magnitude.7 In addition, screening for selective
M1 receptor agonists is critically dependent on several
tissue factors which may vary between tissues and cell
lines.8 As a result, poor correlations between in vitro
and in vivo tests are usually observed.
In the present study, a series of substituted oximes
of N-methyl-4-piperidinone and 3-tropinone bearing the
alkynyl unsaturation in the side chain have been
synthesized by facile methods. Unlike previously syn-
thesized agonists which possess rigid 1-azabicyclic ring
systems,9,11,12 these compounds possess the more flexible
piperidine and tropane rings. Variations in the side
chain have also been undertaken to reveal the effect of
chain length, branching, and location of unsaturation
on activity. None of the compounds are chiral, a feature
which greatly facilitated their synthesis and separation.
The receptor subtype selectivities of these oximes have
been evaluated by investigating their ability to displace
[3H]pirenzepine from muscarinic receptors in rat cere-
bral cortex and [3H]-N-methylscopolamine from rat
heart ventricle and submandibulary glands. In addi-
tion, the muscarinic agonist potencies have been evalu-
ated in functional assays on the rat aorta and the
endothelium-denuded rabbit aorta for M3 and M2 activi-
ties, respectively. The effect of selected oximes on the
mnemonic responses of mice in the modified Morris
water maze was also investigated.
Several investigators have designed oxime ethers as
muscarinic agonists with promising results. Bromidge
and co-workers9 proposed the oxime ether as an alter-
native bioisostere to the ester and 1,2,4-oxadiazole
functionalities, based on the broad similarities observed
S0022-2623(97)00858-3 CCC: $15.00 © 1998 American Chemical Society
Published on Web 07/21/1998

O-Alkynyloximes as Muscarinic Agonists
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 17 3221
Ch a r t 1. Structures of Compounds 1-5
dibulary gland contains equal proportions of M3 and
M1,21 and it is assumed that this also exists in the rat.
Table 2 gives the receptor binding affinities (ex-
pressed as pKi) of the oximes for the various muscarinic
receptor subtypes. For comparative purposes, the pKi
and Hill coefficients of some muscarinic ligands were
also determined. These include the nonspecific musca-
rinic antagonists atropine and N-methylscopolamine,
pirenzepine (M1 antagonist), McN-A-343 (M1 agonist),
and the nonselective agonist oxotremorine. It was
generally observed that the muscarinic agonists had
lower binding affinities than the antagonists and Hill
coefficients which were less than 1, which may indicate
heterogeneity of binding sites. The rank order observed
at [3H]pirenzepine-labeled cortical sites was atropine >
pirenzepine > oxotremorine > McN-A-343. At the [3H
]-N-methylscopolamine labeled sites of the heart ven-
tricles and submandibulary glands, the sequence N-
methylscopolamine > atropine > oxotremorine > McN-
A-343 was observed.
As seen from Table 2, the pKi values of the oximes
ranged from 4 to 7 for the M1 subtype. Lower binding
affinities were observed for the M2 (pKi ) 4-6) and M3
(pKi ) 3-6) subtypes. Like other muscarinic agonists,
the oximes had Hill coefficients which were significantly
less than unity, which suggested that binding to mul-
tiple affinity states of the muscarinic receptor may have
occurred.
The selectivity of the oximes for the muscarinic
subtypes was assessed from the ratios of their Ki values
at the various sites (Table 3). A large number of oximes
(25b, 26a ,b, 27b, 28a , 31a , 32a ,b) showed at least a
10-fold selectivity for the M1 over M3 subtype, but fewer
oximes (26a , 27a , 28a , 32a ) were at least 15-fold more
selective for the M1 over M2 subtype. Only two oximes
(26b, 34a ) had a 15-fold or more greater selectivity for
M2 than M3.
Because of its outstanding M1 selectivity, the binding
affinity of the oxime 26a which has a but-2-ynyl
(-CH2CtCCH3) side chain was investigated in the
presence of 30 µM Gpp(NH)p. Ligands with good M1
selectivity have been noted to possess small Gpp(NH)p-
induced shifts.22 Table 4 gives the binding parameters
of 26a and two other less selective oximes (24a , 29a ) in
the presence of Gpp(NH)p. The binding affinities of the
less selective oximes (24a , 29a ) for the cardiac and
submandibulary sites were observed to decrease, but
little change was observed in their binding affinities for
cortical sites. As a result, there was a significant
increase in the Ki ratios for M3/M1 and M3/M2 subtype
selectivity in the presence of Gpp(NH)p for 24a and 29a .
Ch em istr y
The syntheses of the O-substituted oximes of tropi-
none and N-methyl-4-piperidinone are outlined in
Scheme 1. The intermediate N-alkynylphthalimides
(6-14) are synthesized from the reaction of diethyl
azodicarboxylate and triphenylphosphine (to form a
betaine), an alcohol, and the nucleophilic N-hydroxyph-
thalimide.13,14 Hydrazinolysis of the intermediate N-
(alkynyloxy)phthalimides was attempted by various
methods such as refluxing/stirring with hydrazine
hydrate in ethanol or pyridine.13,15 It was eventually
found that vigorous agitation with neat hydrazine
hydrate16 gave the O-alkynylhydroxylamines (15-23)
in satisfactory yields (40-70%). The latter were con-
densed with the ketone (3-tropinone or N-methyl-4-
piperidinone) by stirring in methanol at room temper-
ature.17 The resulting O-alkynyloximes of 3-tropinone
(24a -32a ) and N-methyl-4-piperidinone (24b-32b), as
well as 33a ,b and 34a ,b, were converted to the water-
soluble hydrochloride salts for pharmacological evalu-
ation. The physical data of the O-alkynyloximes of
3-tropinone and N-methylpiperidin-4-one are given in
Table 1. The syntheses of O-(1-ethyl-2-propynyl)oximes
(28a ,b), O-(2-pentynyl)oximes (29a ,b), O-(3-pentynyl)-
oximes (30a ,b ), and O-(1-methyl-3-butynyl)oximes
(32a ,b) of 3-tropinone and N-methyl-4-piperidinone,
N-methyl-4-piperidin-3-one O-methyloxime (33b), and
N-methyl-4-piperidinone oxime (34b) have not been
previously reported. The remaining oximes (24a ,b-
27a ,b, 31a ,b, 33a , 34a ) have been mentioned previ-
ously17 but without supporting physical or pharmaco-
logical data.
Resu lts
In the case of oxime 26a , little change in binding
affinity at all three sites (cortical, cardiac, and subman-
dibulary) was noted in the presence of Gpp(NH)p, and
no marked incease in subtype selectivity was observed.
The small Gpp(NH)p shifts of 26a in the cortex, heart,
and submandibulary glands suggest that it is a ligand
with M1 selectivity.
In Vitr o F u n ction a l Assa ys. Functional in vitro
assays were carried out on the endothelial intact rat
aorta (muscarinic M3 receptors)23 and the de-endothe-
lialized rabbit aorta (muscarinic M2 receptors).24 ACh
induces a dose-dependent relaxation of the phenyleph-
rine-contracted endothelium intact rat aortic ring.
Recep tor Bin d in g Assa ys. M1 receptor binding
affinity was assessed from the ability of the oximes to
displace [3H]pirenzepine from its muscarinic binding
sites on the rat cerebral cortex. M2 and M3 receptor
binding affinities were assessed from the concentration
of oxime required to displace [3H]-N-methylscopolamine
from muscarinic sites in the rat heart and submandibu-
lary glands, respectively. The cerebral cortex in rat
expresses a mixture of muscarinic receptor subtypes (M1
40%, M2 30%, M3 5%, and M4 20%) as found from
immunoprecipitation,18,19 of which the M1 subtype is
predominant. The rat heart expresses a homogeneous
population of the M2 subtype.20 The rabbit subman-

3222 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 17
Sch em e 1. Synthetic Route to Oximes 24a ,b-32a ,ba
Xu et al.
a
(a) Alcohol ROH, diethyl azodicarboxylate, triphenylphosphine; (b) hydrazine hydrate; (c) tropinone/N-methyl-4-piperidinone, methanol,
room temperature.
Relaxation was observed upon cumulative addition of
the oximes, indicating that these compounds are ago-
nists. The concentration of the oximes required to
produce 50% relaxation of phenylephrine-contracted
endothelium intact rat aortic rings (ED50) ranged from
10-5 to 10-6 M, and they were significantly less potent
than oxotremorine (Table 5). There was a reasonable
linear correlation between the binding assay pKi values
Scopolamine-treated mice also improved their perfor-
mance significantly over time (F ) 34.60, p < 0.001),
but they still took a significantly longer time to perform
the task compared to the control mice (F ) 34.78, p <
0.001). Mice which had received scopolamine and 26a ,
28a , or 32a performed significantly better (shorter time)
than those receiving scopolamine alone: 26a , F )13.68,
p < 0.001; 28a , F ) 27.35, p < 0.001; 32a , F ) 14.82, p
< 0.001. Similar observations were made with mice
treated with McN-A-343 (F ) 15.58, p < 0.001) but not
in mice treated with scopolamine + 33a (F ) 1.00, p >
0.05). No differences in latencies were observed be-
tween control mice and mice treated with scopolamine
and 26a , 28a , and 32a (p > 0.05).
and the pKED values obtained from the M3 functional
50
assay, as shown in eq 1:
pKED ) 0.34((0.63) pKi + 3.55((0.30)
(1)
50
n ) 22, r2 ) 0.59, SE ) 0.23, F ) 29.29
A dose-related study was also carried out in which
the latency time in mice that received various doses of
oxime 28a (0.2, 1, 2, or 5 mg/kg) combined with
scopolamine (1 mg/kg) was determined. There were no
significant differences in the responses of control mice
and mice which received 28a (2 mg/kg) alone (F ) 0.00,
p > 0.05). This was also true for mice which received
higher doses of 28a (5, 2 mg/kg) combined with scopo-
lamine (F ) 1.63, p > 0.05). Responses of mice receiving
lower doses of 28a (0.2, 1 mg/kg) did not differ signifi-
cantly from each other (F ) 0.116, p > 0.05) but were
significantly different from responses obtained in sco-
polamine-treated mice over 4 days (1.0 mg/kg: F )
11.40, p < 0.01; 0.2 mg/kg: F ) 8.91, p < 0.05).
In another set of experiments, the acute actions of
scopolamine in control and oxime-treated mice were
investigated. Two groups of mice were treated with
saline on day 1, and their performance in the water
maze task was monitored daily. There were no signifi-
cant differences in the responses of these two groups of
mice during this period (F ) 0.00, p > 0.05). On day 6
(or day 16 for 26a ), one group of mice was given
scopolamine and saline while another group was given
scopolamine and drug (26a , 28a , 32a , 33a ). It was
observed that the group of mice treated with scopola-
mine and 26a , 28a , or 32a performed significantly
better than the control group treated with scopolamine
alone (p < 0.01, evaluated by two-tailed t-test and
Mann-Whitney U-test). In contrast, no significant
difference was detected between control mice and mice
treated with 33a . The results show that the selected
tropinone oximes do counteract acute scopolamine-
In endothelium-denuded rabbit aortic rings which
have been submaximally precontracted with 10-7
M
phenylephrine, addition of ACh increased the tone in a
dose-dependent manner.24 The M1-selective agonist
McN-A-343 did not alter the contractile response of the
aortic rings up to 10-2 M, but the propargyl ester of
arecoline (APE), an M2 and M3 receptor agonist, in-
creased the contractile tone of the rings at concentra-
tions of 10-4 M. However, McN-A-343, but not arecoline,
relaxed the precontracted denuded aortic rings with an
ED50 of 3.53 × 10-7 M. Like McN-A-343, the oximes
relaxed the precontracted endothelium-denuded aortic
rings, but at much higher ED50 values (Table 5). Thus,
the oximes do not appear to be agonists at the M2
receptor sites of the endothelium-denuded rabbit aorta.
There was a poor correlation (r2 ) 0.13) between the Ki
values for M2 binding and the pKED from functional
50
assay.
Beh a vior a l Testin g in Mice. Selected oximes (26a ,
28a , 32a ) which showed good M1 binding affinities and
selectivities were tested for their ability to attenuate
scopolamine-induced impairment of the water escape
performance of mice in the modified Morris swim maze.
The effects of McN-A-343 and the methyloxime of
tropinone (33a ) which has poor M1 binding affinity were
also investigated for comparison. The variation in
latency (time taken for mice to reach platform) over time
in mice treated daily with scopolamine, saline (control),
or scopolamine + drug (tested at 0.2 mg/kg) was
determined. Control mice treated with saline showed
a constant performance by day 3 (F ) 15.67, p < 0.001).

O-Alkynyloximes as Muscarinic Agonists
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 17 3223
Ta ble 1. Physical Data of O-Alkynyloxime Hydrochlorides of Tropinone (24a -34a ) and N-Methyl-4-piperidinone (24b-34b)
%
compd yield
mp (°C)
184-186
IR (KBr cm-1
)
elemental anal.
accurate mass
192.1271
(C11H18N2O ) 192.1262)
1H NMR (CD3OD, δ)
24a
24b
67
61
3260.30 (υtCH) C,H,N
2122.13 (υCtC)
1663.15 (υCdN)
4.66-4.67 (d, 2H, -O-CH2),
2.87-2.89 (t, 1H, tCH), 2.86
(s, 3H, NCH3), 3.31-2.62
(m, 10H, tropinyl H)
4.65 (d, 2H, -OCH2), 3.30-3.39
(m, 4H, piperidinyl H, at
C2, C6), 2.93 (s, 3H, NCH3),
2.85-2.87 (t, 1H, tCH),
2.67-2.71 (m, 4H, piperidinyl
H at C3, C5)
147.150
3290.79 (υCtN) C,H,N
2122.49 (υCtC)
1645.80 (υCdN)
166.1107
(C9H14N2O ) 166.1106)
25a
25b
26a
26b
37
28
64
51
174-176
84.86
3277.30 (υtCH) C,N
2112.49 (υCtC) H: calcd, 8.09
1629.52 (υCdN) found, 8.05
206.1421
(C12H18N2O ) 206.1419)
4.68-4.70 (m, 1H, -OCH-),
2.83 (s, 3H, NCH3), 2.44-2.45
(m, 1H, tCH), 2.03-4.31
(m, 10H, tropinyl H) 1.48-1.50
(d, 3H, CH3)
3292.72 (υtCH)
C
180.1262
(C10H16N2O ) 180.1262)
4.82-4.87 (m, 1H, -OCH-),
2.85 (s, 3H, NCH3), 2.47-2.45
(d, 1H, tCH), 2.46-4.32
(m, 8H, piperidinyl H), 1.49-1.51
(d, 3H, -CH3)
1.48-1.50 (d, 3H, -CH3), 4.59-4.62
(q, 2H, -OCH2), 2.86
(s, 3H, NCH3), 1.80-1.82
(t, 3H, tCCH3), 1.86-3.35
(m, 10H, tropinyl H)
4.58-4.60 (q, 2H, -O-CH2), 2.93
(s, 3H, NCH3), 3.29-3.48 (m, 2H,
piperidinyl H at C2, C6), 1.80-1.82
(t, 3H, tCCH3), 2.66-2.88
(m, 4H, piperidinyl H at C3, C5)
4.16-4.19 (m, 2H, -OCH2), 2.83
(s, 3H, NCH3), 2.51-2.57 (m, 2H,
-CH2-), 2.27-3.32 (m, 10H,
tropinyl H), 1.86-1.98 (t, 1H, tCH)
4.12-4.17 (t, 2H, -OCH2), 2.85
(s, 3H, NCH3), 2.80-3.77 (m, 8H,
piperidinyl H), 2.50-2.58 (d, 2H,
-CH2-Ct), 1.97-1.98 (t, 1H, tCH)
4.66-4.71 (m, 1H, -O-CH), 2.86
(s, 3H, NCH3), 2.48-2.49 (d, 1H,
tCH), 1.76-1.98 (m, 2H, CH2),
1.00-1.04 (t, 3H, CH3), 2.27-3.32
(m, 10H, tropinyl H)
2120.21 (υCtC) H: calcd, 7.91
1657.37 (υCdN) found, 7.86
N: calcd, 12.93
found, 12.86
2262.27 (υCtC) C,H
172-174
170-172
206.1412
1641.94 (υCdN) N: calcd, 10.74 (C12H18N2O ) 206.1419)
found, 10.87
2226.27 (υCtC) C,H
180.1261
1647.73 (υCdN) N: calcd, 12.93 (C10H16N2O ) 180.1262)
found, 12.66
27a
27b
28a
77
54
77
157-158
150-152
168-169
3238.73 (υtCH) C,N
2127.92 (υCtC) H: calcd, 7.89
1649.66 (υCdN) found, 7.85
206.1417
(C12H18N2O ) 206.1419)
3269.58 (υtCH) C,H
2127.92 (υCtC) N: calcd, 12.93 (C10H16N2O ) 179.1184)
1657.37 (υCdN) found, 12.78
179.1193
3223.30 (υtCH) C,H
2120.21 (υCtC) N: calcd, 10.91 (C13H20N2O ) 220.1576)
1648.66 (υCdN) found, 10.97
220.1568
28b
29a
29b
30a
75
32
52
77
165-167
3246.44 (υtCH) H,N
2112.49 (υCtC) C: calcd, 57.26 (C11H18N2O ) 194.1419)
1665.08 (υCdN) found, 56.77
194.1408
4.64-4.69 (m, 1H, -OCH-), 2.89
(s, 3H, NCH3), 2.60-3.71 (m, 8H,
piperidinyl H), 2.48-2.49 (d, 1H,
tCH), 1.79-1.85 (m, 2H, CH2),
0.99-1.04 (t, 3H, CH3)
4.66-4.71 (m, 2H, -OCH2) 2.83
(s, 3H, NCH3), 2.21-2.28
143-146
2235.92 (υCtC)
C
220.1588
(C13H20N2O ) 220.1576)
1649.66 (υCdN) H: calcd, 8.24
found, 8.28
(m, 2H-CH2-), 1.90-3.94 (m, 10H,
tropinyl H), 1.12-1.17 (m, 3H, CH3)
N: calcd, 10.91
found, 10.81
not determined
2235.91 (υCtC)
H
194.1426
4.63-4.72 (m, 2H, -O-CH2),
2.86 (s, 3H, NCH3), 2.00-3.70
(m, 8H, piperidinyl H), 2.21-2.27
(m, 2H, -CH2-), 1.13-1.18
(t, 3H, CH3)
(too hydroscopic) 1645.79 (υCdN) C: calcd, 57.26 (C11H18N2O ) 194.1419)
found, 57.58
N: calcd, 12.14
found, 11.98
212-214
2259.05 (υCtC) C,H
1641.94 (υCdN) N: calcd, 8.24
found, 8.20
220.1577
(C13H20N2O ) 220.1576)
4.07-4.12 (t, 2H, -OCH2),
2.85 (s, 3H, NCH3), 2.44-2.49
(m, 2H, -CH2-Ct), 1.84-3.97
(m, 10H, tropinyl H), 1.76-1.78
(t, 3H, -CH3)
30b
31a
31
32
178-180
108-110
2259.05 (υCtC) C,H
194.1428
4.06-4.11 (t, 2H, -OCH2), 2.88
(s, 3H, NCH3), 2.44-2.49 (m, 2H,
-CH2-), 2.43-3.69 (m, 8H,
piperidinyl H), 1.77-1.78 (t, 3H, CH3)
4.12-4.17 (m, 2H, -OCH2-), 2.83
(s, 3H, NCH3), 2.23-2.30 (m, 2H,
-CH2-Ct), 2.00-3.97 (m, 10H,
tropinyl H), 1.96-1.98 (t, 1H, tCH),
1.83-1.90 (m, 2H, -CH2-)
1665.08 (υCdN) N: calcd, 12.14 (C11H18N2O ) 194.1419)
found, 12.05
3238.73 (υtCH) C,H
2120.21 (υCtC) N: calcd, 10.91 (C13H20N2O ) 220.1576)
1625.52 (υCdN) found, 10.97
220.1575

3224 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 17
Xu et al.
Ta ble 1 (Continued)
%
compd yield
mp (°C)
90-91
IR (KBr cm-1
)
elemental anal.
accurate mass
194.1420
(C11H18N2O ) 194.1419)
1H NMR (CD3OD, δ)
31b
37
3231.01 (υtCH) C,H,N
2120.21 (υCtC)
1657.31 (υCdN)
4.12-4.16 (t, 2H, -OCH2), 2.86
(s, 3H, NCH3), 2.38-3.71 (m, 8H,
piperidinyl H), 2.23-2.29
(m, 2H, -CH2-Ct), 1.96-1.97
(t, 1H, tCH), 1.84-1.90
(m, 2H, -CH2-)
32a
32b
33a
33b
74
75
73
142-144
147-149
3254.15 (υtCH) C,H,N
2120.21 (υCtC)
220.1569
(C13H20N2O ) 220.1576)
4.31-4.35 (m, 1H, -OCH-), 2.84
(s, 3H, NCH3), 2.47-2.51
1649.66 (υCdN)
(m, 2H, -CH2-), 2.26-3.99
(m, 10H, tropinyl H), 1.97-1.99
(t, 1H, tCH), 1.30-1.34 (d, 3H, CH3)
4.28-4.34 (t, 1H, -O-CH-),
2.86 (s, 3H, NCH3), 2.44-2.49
(m, 2H, -CH2-), 2.38-3.70
(m, 8H, piperidinyl H), 1.98-2.00
(t, 1H, tCH), 1.31-1.34 (d, 3H, CH3)
3.85 (s, 3H, CH3), 2.83 (s, 3H, NCH3),
1.86-3.88 (m, 10H, tropinyl H)
3223.30 (υtCH) C,H,N
2116.34 (υCtC)
1653.51 (υCdN)
194.1423
(C11H18N2O ) 194.1419)
217-220
1649.65 (υCdN)
C
168.1241
(C9H16N2O ) 168.1263)
(145-147)17
H: calcd, 8.37
found, 8.31
N: calcd, 13.69
found, 13.45
68.7 145-147
1649.86 (υCtN) C,H
142.1114
3.85 (s, 1H2, CH3), 2.85 (s, 3H,
N-CH3), 2.54-3.71 (m, 8H,
piperidinyl H)
D2O: 4.75 (s, 1H, OH), 2.86 (s, 3H, NCH3),
1.84-4.11 (m, 10H, tropinyl H)
4.75 (s, 1H, OH), 2.85 (s, 3H, NCH3),
2.65-3.40 (m, 8H, piperidinyl H)
N: calcd, 15.68 (C7H14N2O ) 142.1106)
found, 15.60
C,H,N
34a
34b
62
217-220
(217-220)17 1657.37 (υCdN)
57.7 247-248 3192.44 (υOH)
3254.15 (υOH)
C,H
1626.52 (υCdN) N: calcd, 17.02
found, 17.13
induced learning deficits in mice in a spatial reference
memory task such as the water maze performance.
Qu a n tita tive Str u ctu r e-Activity Rela tion sh ip s.
Some consideration is paid to the molecular properties
of the oximes, such as lipophilic, electronic, and size
parameters, as these are likely to influence their affini-
ties for the various muscarinic receptor subtypes.
The lipophilicity of the oximes was assessed from the
experimentally determined log kw values and the theo-
retical ClogP values. Under the prevailing experimental
conditions, the log kw values are actually distribution
coefficients obtained at pH 3.2, where the ionized form
of the oxime is predominant. In contrast, the ClogP
values have been determined for the neutral molecule.
Not withstanding this difference, there was a good
correlation (r ) 0.91) between the experimental and
theoretical hydrophobicity values (Table 6).
Various size parameters were determined to assess
the bulk of the oximes. One such parameter is molar
refractivity (MR) which is proportional to the polariz-
ability and volume of a molecule. Since polarizabilities
of most organic molecules tend to show limited varia-
tion, molar refractivity is widely considered as a steric
parameter characterizing the bulk of a substituent.
Indeed, Table 6 shows that it is better correlated (r >
0.97) to other size parameters (surface area, volume)
than to electronic parameters. In series rich in alkyl
substituents, molar refractivity may reflect a hydropho-
bic effect.26 The good correlation between MR and the
hydrophobic parameter log kw (r ) 0.95) observed in this
series may be explained in this manner.
Conventional molecular volume (MV) and solvent-
accessible Connolly surfaces (area A and volume V)27
were determined from the low-energy conformations of
the molecules using SYBYL.
Except for the methyloximes (33a ,b), the remaining
O-substituted oximes possess an electron-withdrawing
alkynyl side chain, the electron-withdrawing effect of
which would vary according to chain length and the
location of unsaturation along the chain. Quantification
of this effect could be done in several ways. One
approach would involve determining the Taft’s polar
substituent constant σ25 of these side chains. Electron-
withdrawing side chains would be associated with large
(positive) σ values. The negative charge (O-) on the
oxime oxygen could also be determined, as this would
be diminished by the presence of an electron-withdraw-
ing side chain. Another suitable electronic parameter
would be the dipole moment of the oxime which reflects
the global polarity of the molecule. As seen from Table
6, these variables (σ, O-, dipole moment) are poorly
correlated to each other. In particular, σ and dipole
moment are very poorly correlated (r ) 0.001) compared
to σ and O- (r ) 0.56).
The shape of the alkynyl substituent may also be
quantified by the STERIMOL set of parameters pro-
posed by Verloop.28 These parameters represent length
(L) and widths (B1-B4) of the substituent, the latter
extending perpendicular to the L axis while being
perpendicular to each other. Table 6 shows that most
but not all STERIMOL parameters are correlated to
each other. For example, the width parameter B1 is
well-correlated to B2 and B3, as are the two longest
width parameters B4 and B5. As expected, little cor-
relation is observed between length (L) and width (B1-
B5) parameters. Correlations with other size/electronic
parameters are also observed among some STERIMOL
parameters, particularly with the length (L) substituent
and the longest width parameter B5.
Because branching in the alkynyl side chain is an
important structural feature in the present series of
oximes, it is important to find a steric parameter which

O-Alkynyloximes as Muscarinic Agonists
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 17 3225
Ta ble 2. Affinities (pKi) of Oximes (24a ,b-34a ,b) and Other Muscarinic Ligands for M1, M2, and M3 Muscarinic Receptors in the
Rat Cortex, Heart, and Submandibulary Glands, Respectively, as Determined from Binding Displacement Studies
pKi valuesa for inhibition of
[3H]pirenzepine
[3H]-N-methylscopolamine
[3H]-N-methylscopolamine
compd
R
binding in cortexb
binding in heartb
binding in submandibulary glandb
24a
24b
25a
25b
26a
26b
27a
27b
28a
28b
29a
29b
30a
30b
31a
31b
32a
32b
33a
33b
34a
34b
5.9 (0.1)
(0.57 ( 0.09)
4.3
(0.47 ( 0.05)
5.0
(0.71 ( 0.19)
5.3
(0.52 ( 0.07)
6.3
(0.55 ( 0.03)
5.4 (0.2)
(0.51 ( 0.04)
5.8 (0.1)
(0.79 ( 0.05)
5.8 (0.1)
(0.48 ( 0.06)
6.5 (0.1)
(0.62 ( 0.16)
5.2 (0.1)
(0.49 ( 0.04)
5.3 (0.1)
(0.65 ( 0.03)
5.4 (0.1)
(0.51 ( 0.08)
6.1 (0.1)
(0.65 ( 0.06)
5.7 (0.1)
(0.62 ( 0.06)
6.2 (0.1)
(0.72 ( 0.03)
5.1
(0.62 ( 0.06)
7.0 (0.1)
5.2
5.2
(0.68 ( 0.06)
4.6 (0.1)
(0.70 ( 0.12)
5.2 (0.1)
(0.71 ( 0.14)
4.9
(0.36 ( 0.10)
4.9 (0.1)
(0.72 ( 0.06)
5.1 (0.1)
(0.78 ( 0.06)
4.5
(0.85 ( 0.02)
5.0 (0.2)
(0.54 ( 0.15)
5.2 (0.1)
(0.58 ( 0.08)
5.0 (0.1)
(0.64 ( 0.07)
4.9 (0.1)
(0.74 ( 0.04)
4.8
(0.51 ( 0.03)
5.5 (0.1)
(0.79 ( 0.08)
4.8 (0.1)
(0.70 ( 0.06)
5.2
(0.73 ( 0.05)
5.0
(0.75 ( 0.11)
5.8 (0.1)
(0.39 ( 0.01)
4.7 (0.1)
(0.31 ( 0.03)
5.4 (0.1)
(0.67 ( 0.10)
4.1
(0.28 ( 0.06)
4.8 (0.1)
(0.19 ( 0.01)
3.9 (0.1)
(0.24 ( 0.02)
5.0 (0.1)
(0.46 ( 0.08)
4.4
(0.30 ( 0.09)
5.2
(0.70 ( 0.09)
5.0
(0.80 ( 0.07)
5.5
(0.38 ( 0.03)
4.7 (0.1)
(0.25 ( 0.06)
5.6 (0.1)
(0.59 ( 0.03)
4.4 (0.1)
(0.65 ( 0.07)
5.0
(0.80 ( 0.07)
4.8 (0.1)
(0.33 ( 0.02)
5.5 (0.1)
-CH2CtCH
-C(CH3)HCtCH
-CH2CtCCH3
-CH2CH2CtCH
-C(C2H5)HCtCH
-CH2CtCC2H5
-CH2CH2CtCCH3
-CH2CH2CH2CtCH
-C(CH3)HCH2CtCH
-CH3
(0.43 ( 0.04)
5.5 (0.1)
(0.58 ( 0.04)
3.0
(0.66 ( 0.09)
4.3
(0.66 ( 0.05)
3.3
(0.52 ( 0.04)
5.0 (0.1)
(0.61 ( 0.04)
3.9 (0.2)
(0.48 ( 0.05)
4.4 (0.1)
(0.48 ( 0.18)
4.4 (0.1)
(0.51 ( 0.06)
4.3
(0.32 ( 0.06)
4.0 (0.1)
(0.43 ( 0.04)
4.6
(0.10 ( 0.04)
3.1
(0.45 ( 0.04)
3.3 (0.1)
(0.75 ( 0.07)
3.4
(0.27 ( 0.02)
2.5
-H
(0.73 ( 0.13)
8.4
(0.54 ( 0.06)
8.2
(0.22 ( 0.02)
8.6
atropine
(0.98 ( 0.08)
8.1
(0.88 ( 0.04)
(0.40 ( 0.03)
pirenzepine
(0.91 ( 0.03)
5.3
McN-A-343
5.2
5.6
(0.45 ( 0.02)
6.2
(0.30 ( 0.09)
(0.59 ( 0.11)
(0.29 ( 0.10)
oxotremorine
N-methylscopolamine
6.7 (0.1)
6.4
(0.55 ( 0.08)
9.8 (0.7)
(0.26 ( 0.07)
9.3
(1.05 ( 0.05)
(0.93 ( 0.07)
a
Each value represents the mean (SD) of three independent experiments. Values in parentheses represent the Hill coefficients of the
b
compounds as determined from the Hill plot. The Kd (dissociation constant), Bmax (maximal number of binding sites), and Hill coefficient
of the following radiolabeled ligands for the specified tissue are given as follows: [3H]pirenzepine (cortex), 6.8 nM, 63.5 (fmol/mg of wet
weight), 1.00; [3H]-N-methylscopolamine (heart), 0.22 nM, 36.17 (fmol/mg of wet weight), 1.03; [3H]-N-methylscopolamine (submandibulary
glands), 0.55 nM, 11.9 (fmol/mg of wet weight), 0.98.
could represent it in QSAR. This appears to be the
width parameter B2 which has a very good correlation
(r ) 0.88) to the dummy parameter D (D ) 1 for
branching). B1 and B3 are also alternative indicators,

3226 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 17
Xu et al.
Ta ble 3. Receptor Subtype Selectivities of Oximes
24a ,b-34a ,b and Other Muscarinic Ligands as Determined
from the Ratio of Ki Values Obtained from Binding
Displacement Studies
Interestingly, the tropinone oximes with the highest
binding affinities have branched side chains: -CH(CH2-
CH3)-CtCH (28a ) and -CH(CH3)-CH2CtCH (32a ).
These oximes have larger pKi values than McN-A-343,
determined under similar conditions. However, branch-
ing would result in smaller surface areas (as have been
observed earlier) which is contrary to the relationship
expressed in eq 2. It is also noted that none of the
parameters depicting branching (D, B2, B1, B3) are
present in this equation. A possible reason for this
anomaly may be that the percent change in surface area
brought about by branching is small compared to the
total surface area of the molecule. Moreover, the
number of active branched analogues is small, as the
desirable effect of branching on affinity is observed only
among the tropinone (and not N-methylpiperidinone)
derivatives.
compd
Ki(M2)/Ki(M1)
Ki(M3)/Ki(M1)
Ki(M3)/Ki(M2)
24a
24b
25a
25b
26a
26b
27a
27b
28a
28b
29a
29b
30a
30b
31a
31b
32a
32b
33a
33b
34a
34b
4.2
0.59
4.5
2.2
24
2.0
19
6.2
20
1.8
2.4
4.0
4.5
7.5
12
4.6
0.42
0.46
14
27
35
6.8
25
22
1.5
0.56
5.4
3.0
2.0
16
1.7
29
15
0.10
0.51
1.4
7.6
0.60
0.46
0.63
1.1
0.72
0.72
6.3
1.1
17
0.34
3.7
1.1
0.85
0.23
1.4
0.66
0.27
1.3
1.1
16
1.5
1.8
5.4
0.74
0.71
18
A comparison of pKi values of branched and un-
branched O-substituted tropinone oximes suggests that
for long side chains (viz., pentynyl), branching is favored
(28a > 29a , 32a > 30a ), but this is not so for short side
chains (viz., butynyl), as seen in 26a > 25a . In addition,
for the five-carbon pentynyl side chain, there is a
preference for the location of the triple bond at a distal
position (as in 31a ) compared to the proximal position
(as in 29a ). We compared these results with those
obtained earlier12 from a series of O-alkynyloximes of
the more rigid 1-azabicyclo[2.2.1]heptane ring. In con-
trast to the present findings, proximal (rather than
distal) unsaturation is favorable in those oximes with
side chains of three- to five-carbon side length. Branch-
ing was also observed to diminish binding affinity.
Further lengthening of the side chain beyond a five-
carbon chain length generally resulted in lower activity,
an effect which might also be observed in the present
series of oximes if longer chain analogues were avail-
able. It is clear that the nature of the nitrogen-
containing ring system to which the side chain is
attached has an important bearing on structure-
activity relationships.
2.9
0.14
0.74
0.74
0.90
1.4
1.4
0.32
8.5
atropine
McN-A-343
oxotremorine
0.82
0.33
2.0
but with lesser correlations. The Connolly surfaces (log
V, log A) and molecular volume (log MV) determined
from the energy-minimized conformations of the oximes
are less useful indicators of branching.
Although log A is poorly correlated to branching, a
casual examination shows that the branched side chains
of the oximes 25a ,b, 28a ,b, and 32a ,b generally have
smaller surface areas and larger volumes compared to
unbranched side chains of the same carbon number
(27a ,b, 31a ,b).
Discu ssion
Stepwise multiple linear regression of M1 binding
affinity against the various hydrophobic, size, and
electronic parameters indicated that activity was best
correlated to either molecular refractivity or surface
area (log A) of the oximes. However, cross-validation
of the data indicated that better predictive ability was
obtained with surface area. The regression eq 2 shows
that up to 70% of the binding affinity data could be
explained by this variable alone.
Although previously synthesized rigid ring oximes are
associated with greater binding affinities,12 they have
poor M2/M1 selectivities. The present series of tropinone
oximes is good in that despite their moderate M1 binding
affinities, they have markedly greater M1 (over M2 and
M3) selectivities. Of the three tropinone oximes (26a ,
28a , 32a ) with good M2/M1 and M3/M1 selectivities (at
least 15-fold), two members have branched alkynyl side
chains (28a , 32a ).
pKi(M ) ) 15.54((2.27) log A - 31.27((5.36) (2)
1
Oximes 26a , 28a , and 32a were able to attenuate
scopolamine-induced impairment of the performance of
mice in the modified Morris water maze task. In
comparison, mice treated with an oxime with low M1
binding affinity (33a ) were unable to overcome this
effect of scopolamine.
n ) 22, r2 ) 0.70, SE ) 0.59, F ) 46.48
r2 ) 0.57, SEcv ) 0.69
cv
Equation 2 suggests that an increase in surface area
enhances binding affinity. Indeed, for the same side
chain, greater binding affinity is observed in the larger
size tropinone oxime compared to the corresponding
piperidinone oxime. In addition, lengthening the alky-
nyl side chain from three to five carbon atoms (an
increase in surface area) generally results in improved
binding affinity. For example, pKi increases for 27b
(4C) > 24b (3C) and 31a (5C) > 27a (4C).
Much discussion has centered on the nature of the
muscarinic agonist binding site. Some investigators29
have proposed that the binding site is located in a very
tight pocket with little tolerance for excess volume.
Greater steric bulk is thought to lead to loss in agonist
potency or a shift toward antagonism. On the other
hand, Tecle et al.12 proposed that longer and larger
agonists would ensure maximum contact between the
agonist and the internal surface of the binding cavity,

O-Alkynyloximes as Muscarinic Agonists
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 17 3227
Ta ble 4. Affinities (pKi) and Receptor Selectivities of Oximes 24a , 26a , and 29a for M1, M2, and M3 Muscarinic Receptors in the Rat
Cortex, Heart, and Submandibulary Glands, Respectively, as Determined from Binding Displacement Studies Carried Out in the
Presence of Gpp(NH)p
pKia for inhibition of
subtype selectivityb
[3H]pirenzepine
binding in cortex
[3H]NMS
binding in heart
[3H]NMS binding in
submandibulary gland
compd
Ki(M2)/Ki(M1)
Ki(M3)/Ki(M1)
Ki(M3)/Ki(M2)
24a
5.9 (0.1)
(0.73 ( 0.05)
6.0
(0.69 ( 0.12)
5.1
4.6
3.3
18
(4.2)
23
(25)
7.7
320
(4.6)
21
(27)
9.1
18
(1.02 ( 0.07)
4.7 (0.1)
(0.55 ( 0.04)
4.2
(0.34 ( 0.03)
4.7 (0.1)
(0.29 ( 0.01)
4.1
(1.1)
0.88
(1.1)
1.2
26a
29a
(0.92 ( 0.18)
(1.05 ( 0.17)
(0.27 ( 0.01)
(2.4)
(0.56)
(0.23)
a
Each value represents the mean ((SD) of three independent experiments. Values in parentheses represent the Hill coefficients of the
b
compounds as determined from the Hill plot. Values in parentheses represent the ratios of Ki values obtained in the absence of Gpp(NH)p.
Ta ble 5. Concentration of Oximes 24a ,b-34a ,b Required to
Relax Phenylephrine-Precontracted Aortic Rings by 50% (ED50
Values) in Endothelium Intact Rat Aorta and
as agonists at the putative M2 sites on the endothelium-
denuded rabbit aorta.
Stepwise linear regression of various variables against
M2 binding affinities identified the size parameter (log
V) as being superior in accounting for activity (eq 3).
This was further confirmed by cross-validation. The
reliance on size parameters for both M1 and M2 binding
affinities is not unexpected since binding affinities at
these two sites are themselves fairly well-correlated to
each other (r ) 0.84).
Endothelium-Denuded Rabbit Aorta
ED50 (SEM) for n experiments
endothelium intact
rat aorta
endothelium-denuded
rabbit aorta
compd
24a
24b
25a
25b
26a
26b
27a
27b
28a
28b
29a
29b
30a
30b
31a
31b
32a
32b
33a
33b
34a
34b
3.5 (0.4) × 10-6 (8)
6.8 (0.2) × 10-6 (8)
6.9 (0.2) × 10-6 (8)
7.8 (0.3) × 10-6 (8)
3.4 (0.2) × 10-6 (6)
8.7 (0.2) × 10-6 (6)
5.3 (0.3) × 10-6 (8)
7.4 (0.4) × 10-6 (8)
5.5 (0.2) × 10-6 (8)
7.6 (0.3) × 10-6 (8)
1.3 (0.3) × 10-6 (8)
6.2 (0.3) × 10-6 (8)
7.8 (0.3) × 10-6 (8)
6.3 (0.3) × 10-6 (8)
5.9 (0.4) × 10-6 (6)
9.3 (0.4) × 10-6 (6)
5.4 (0.2) × 10-6 (6)
5.0 (0.3) × 10-6 (6)
1.1 (0.2) × 10-5 (8)
1.3 (0.4) × 10-5 (8)
5.3 (0.3) × 10-5 (8)
4.3 (0.2) × 10-5 (8)
7.3 (0.5) × 10-8 (24)
2.2 (0.5) × 10-7 (8)
1.6 (0.4) × 10-6 (12)
9.6 (0.2) × 10-6 (12)
4.4 (0.3) × 10-8 (12)
7.1 (0.4) × 10-6 (6)
1.1 (0.4) × 10-5 (6)
2.6 (0.4) × 10-5 (12)
4.1 (0.4) × 10-5 (6)
4.8 (0.4) × 10-6 (8)
8.7 (0.2) × 10-6 (6)
1.5 (0.4) × 10-5 (9)
2.8 (0.6) × 10-5 (8)
4.1 (0.5) × 10-5 (12)
1.8 (0.4) × 10-5 (9)
3.5 (0.2) × 10-6 (8)
1.2 (0.4) × 10-5 (6)
9.8 (0.3) × 10-6 (9)
1.0 (0.2) × 10-6 (8)
6.5 (0.4) × 10-6 (12)
1.3 (0.3) × 10-5 (12)
2.3 (0.4) × 10-5 (10)
2.5 (0.5) × 10-5 (10)
1.1 (0.2) × 10-5 (8)
2.6 (0.5) × 10-4 (6)
3.2 (0.4) × 10-4 (6)
7.8 (0.3) × 10-5 (6)
pKi(M ) ) 6.49((1.04) log V - 10.30((2.44) (3)
2
n ) 22, r2 ) 0.66, SE ) 0.31, F ) 38.66
r2 ) 0.56, SEcv ) 0.35
cv
In a similar manner, ClogP was identified as the best
variable accounting for 73% of the observed M3 binding
affinities (eq 4):
pKi(M ) ) 0.69((0.09) ClogP - 3.14((0.23) (4)
3
n ) 22, r2 ) 0.73, SE ) 0.42, F ) 53.48
r2 ) 0.74, SEcv ) 0.41
ACh
carbacol
arecoline
McN-A-343
oxotremorine
cv
When pKED values obtained from the M3 functional
3.5 (0.4) × 10-7 (4)
50
assay on the rat aorta endothelium were regressed
against the various independent variables, it was found
that M3 agonist activity was best accounted for by a
combination of two variables, viz., ClogP and the nega-
tive charge on oxygen (O-), with the latter making a
greater contribution to the overall relationship.
thereby achieving greater subtype selectivity. Our
results appear to support the latter hypothesis. The
QSAR for the present series of oximes shows that
binding affinity increases as the surface area of the side
chain is extended from a three- to five-carbon chain
length. However, it is likely that an optimum chain
length exists beyond which binding affinity would
diminish.
pKED (M ) ) 0.20((0.06) ClogP + 12.90((6.08) O- +
50
3
8.18((1.73) (5)
n ) 22, r2 ) 0.68, SE ) 0.21, F ) 20.25
Compared to their binding affinities for the M1
receptor subtype, the oximes generally demonstrated
lower binding affinities for the M2 and M3 receptor
subtypes, with the tropinone oximes showing superior
binding affinities than the piperidinone oximes. There
was no clear-cut preference for the oximes with branched
side chains, although branched analogues such as 32a
and 28a are still among those with the highest binding
affinities at M2 and M3. Functional assays revealed that
all the oximes demonstrated agonistic activity at the
putative M3 sites on the rat aorta but did not behave
r2 ) 0.63, SEcv ) 0.22
cv
Thus, a lower negative charge on oxygen (possibly due
to the presence of an electron-withdrawing substituent)
enhances activity. It is interesting to note that although
it is an electronic parameter, O- is better correlated to
size parameters (B1, B2, B5) (r > 0.7) than to the other
electronic parameters such as σ or dipole moment. The
correlation to width parameters suggests that the shape
(“thickness”) of the alkynyl side chain plays an impor-

3228 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 17
Xu et al.
Ta ble 6. Correlation Matrix (r) for Electronic, Hydrophobic, and Size Parameters Used in Deriving QSAR Equationsa
log
Kw
log
A
CLogP
CLogP 1.00
σ
O-
DMb
0.57
MRb log MV
log V
Db
Lc
B1c
0.78
B2c
0.48
B3c
0.59
(0.04)
0.27
B4c
0.64
B5c
0.91
1.00
-0.15 0.63
0.87
0.89
0.95
0.88
0.86
0.89
0.29
0.70
0.72
(0.002) (0.006)
-0.21 -0.09 0.61
(0.478) (0.733) (0.007)
log Kw 0.91
0.95
0.96
0.10
0.19
0.23
0.21
0.21
0.20
0.29
(0.445) (0.365) (0.415) (0.397) (0.277) (0.435) (0.245)
0.41 -0.17 0.16 0.43 0.21 -0.05 0.13
σ
-0.15 -0.21 1.00
0.56
-0.00 0.07
-0.08 0.03
(0.478)
(0.016) (0.997) (0.771) (0.758) (.907) (0.708) (0.094) (0.500) (0.531) (0.07) (0.616) (0.857) (0.611)
O-
0.63
(0.002) (0.733) (0.016)
0.57 0.61 -0.00 0.30
(0.006) (0.007) (0.997) (0.175)
-0.09 0.56
1.00
0.30
0.56
0.58
0.56
0.60
0.44
(0.039) (0.302)
0.12 0.46
0.23
0.72
0.71
0.46
(0.031) (0.011)
0.41
0.53
0.70
0.66
0.73
0.70
0.76
0.72
(0.175) (0.007) (0.005) (0.007) (003)
DMb
MRb
1.00
0.72
0.69
0.70
0.69
0.12
0.72
1.00
0.99
0.98
0.99
0.29
0.69
0.99
1.00
0.97
0.99
0.34
0.70
0.98
0.97
1.00
0.97
0.17
0.69
0.99
0.99
0.97
1.00
0.36
0.32
0.11
0.67
0.64
0.61
(0.592) (0.034) (0.193) (0.627) (0.059)
0.87
0.95
0.95
0.86
0.96
0.19
0.07
0.56
(0.771) (0.007)
-0.08 0.58
0.29
(0.184)
0.34
(0.125)
0.17
(0.437)
0.36
0.73
0.72
0.81
0.71
0.61
0.44
0.49
(0.003) (0.038) (0.020)
log MV 0.89
0.66
0.57
0.48
0.54
(0.758) (0.005)
(0.023) (0.010) (0.003)
0.35
log A
log V
Db
0.88
0.89
0.29
0.70
0.78
0.48
0.59
0.03
(0.907) (0.007)
0.10 0.60
(0.708) (0.003)
0.41 0.44
0.56
0.46
0.71
(0.006) (0.113) (0.033)
0.68 0.51 0.55
0.62
(0.098)
1.00
(0.016) (0.08) (0.002)
0.88
-0.04 0.77
0.76
-0.06 0.21
(0.445) (0.094) (0.039) (0.592) (0.184) (0.125) (0.437) (0.098)
0.23
(0.365) (0.500) (0.302) (0.034)
0.21
(0.415) (0.531)
0.21 0.43
(0.397) (0.07)
0.27 0.21
(0.863)
(0.800) (0.349)
Lc
-0.17 0.23
0.46
0.73
0.72
0.66
0.48
0.81
0.71
0.68
0.51
-0.04 1.00
(0.863)
0.35
0.08
0.21
0.39
0.38
(0.114) (0.710) (0.348) (0.074) (0.079)
B1c
B2c
B3c
B4c
B5c
0.16
0.72
0.71
0.46
0.32
(0.143) (0.003)
0.11 0.44
(0.627) (0.038) (0.023) (0.113) (0.016)
0.41 0.49 0.54 0.46 0.55
0.61
0.57
(0.006)
0.35
0.77
0.88
0.76
0.35
(0.114)
0.08
(0.710)
0.21
1.00
0.79
0.82
0.46
0.79
0.82
0.46
(0.03)
0.28
0.64
0.50
1.00
0.53
(0.012) (0.206) (0.017)
1.00
0.53
(0.012)
0.28
0.37
(0.090) (0.016)
1.00
0.95
0.51
(0.04) (0.277) (0.626) (0.031) (0.059) (0.020) (0.010) (0.033) (0.08)
(0.348)
-0.06 0.39
0.64
0.72
0.20
(0.435) (0.857) (0.011)
0.29 0.13 0.70
(0.245) (0.611)
-0.05
0.53
0.67
0.66
0.64
0.73
0.61
(0.003)
0.70
0.71
0.76
0.62
0.37
0.95
1.00
(0.002) (0.800) (0.074) (0.03) (0.206) (0.090)
0.72 0.21 0.38 0.64 0.50 0.51
(0.349) (0.079)
(0.017) (0.016)
a
b
Values in parentheses represent two-tailed significance p value. If p value is not given, p ) 0.001 or smaller. DM, dipole moment;
MR, molar refractivity; D, branching in side chain. c STERIMOL parameters indicating length (L) and width (B1-B5) of substituent.
tant role in activity, possibly by modulating the charge
on the oxygen atom. The contribution of ClogP may
reflect the importance of the tropine ring system (which
has a larger ClogP value) over the piperidine ring. In
the majority of cases, the tropinone oximes have better
activity than their piperidinone counterparts.
Equation 6 describes the variables which are best
correlated to the ability of the oximes to induce relax-
ation at the endothelium-denuded rabbit aorta.
An attempt was made to determine the variables
determining selectivity (pKM /M , pKM /M , or pKM /M
)
3
1
2
3
1
2
among the different receptor subtypes. As it turned out,
all the variables showed poor predictability (r2cv < 0.5).
Therefore, no single parameter could be identified from
those presently available which could meaningfully
account for the selectivities observed among the oximes.
Con clu sion
The results of the present study provide information
regarding the muscarinic activity and selectivity of
O-alkynyloximes of tropinone and N-methyl-4-piperidi-
none. Other investigators11,12 have synthesized oxime
derivatives of 1-azabicylo[2.2.2]octane or 1-azabicyclo-
[2.2.1]heptane because these rigid ring structures have
been associated with good muscarinic agonist activity.
The tropinone and N-methylpiperidinone rings have
generally been bypassed as suitable pharmacophores
possibly because of their greater flexibility, arising from
the nonbridgehead position of the basic nitrogen atom.17
Our results show that while the tropinone and N-
methylpiperidine oximes demonstrate lower binding
affinities (Ki ≈ 10-7 M), selected members demonstrated
impressive M1 selectivities which exceeded those ob-
served with the 1-azabicyclic ring systems.12 It is seen
that oximes 26a -28a , 31a , and 32a have at least a 10-
fold selectivity for the M1 subtype over the other
receptor (M2, M3) subtypes. Oximes 26a , 28a , and 32a
were also capable of attenuating scopolamine-induced
impairment of the water maze performance in mice. It
would appear that rigidity in the basic nitrogen ring
system is not an absolute requirement for muscarinic
pKED (M ) ) 0.26((0.06) ClogP + 18.67((6.51) O- -
50
2
0.62((0.11) D + 9.37((1.85) (6)
n ) 22, r2 ) 0.80, SE ) 0.21, F ) 24.35
r2 ) 0.62, SEcv ) 0.28
cv
There are many similarities between the regression
eqs 5 and 6. Both equations contain the variables ClogP
and O-, with the latter variable making the greater
contribution to activity. The main difference is the
presence of the branching parameter D in eq 6. The
negative coefficient of D indicates that branching is not
a favorable structural feature among oximes that induce
relaxation in the endothelium-denuded rabbit aorta.
The role of branching is evident when the activities
of two representative oximes (25a , 31a ) are compared.
The branched oxime 25a (1-methylprop-2-ynyl side
chain) is as active as the unbranched O-pent-4-ynyl-
oxime 31a at the endothelium intact rat aorta (M3).
However, 25a is 10 times less active than 31a at the
endothelium-denuded rabbit aorta (M2).

O-Alkynyloximes as Muscarinic Agonists
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 17 3229
Removal of the solvent in vacuo gave a white solid which was
recrystallized from ethanol/benzene. The recrystallized solid
was dissolved in water, made alkaline with solid K2CO3, and
extracted with dichloromethane. After drying over anhydrous
Na2SO4, the organic layer was removed by evaporation under
reduced pressure. A clear yellow oil was obtained which was
purified by passing through a silica gel column eluted with
hexane/ethyl acetate. The eluted oil was treated with ethereal
HCl to give the desired hydrochloride salt of the product. The
yields and physical data of the following synthesized oximes
of tropinone and N-methyl-4-piperidinone are given in Table
1: O-prop-2-ynyloxime (24a ,b), O-(1-methylprop-2-ynyl)oxime
(25a ,b), O-but-2-ynyloxime (26a ,b), O-but-3-ynyloxime (27a ,b),
O-(1-ethylprop-2-ynyl)oxime (28a ,b ), O-pent-2-ynyloxime
(29a ,b ), O-pent-3-ynyloxime (30a ,b ), O-pent-4-ynyloxime
(31a ,b), O-(1-methylbut-3-ynyl)oxime (32a ,b), O-methyloxime
(33a ,b), and unsubstituted oxime (34a ,b).
P h a r m a cology. Ma ter ia ls. [3H]Pirenzepine (73.9 Ci/
mmol) and [3H]-N-methylscopolamine (84.8 Ci/mmol) were
purchased from NEN (Boston, MA). Acetylcholine (ACh)
chloride, phenylephrine HCl, atropine sulfate, N-methylsco-
polamine, and 5′-guanylylimidodiphosphate trisodium salt
[Gpp(NH)p] were obtained from Sigma Chemical Co., St. Louis,
MO. (()-3-Quinuclidinyl benzilate (QNB) was purchased from
Research Biochemicals Inc., Massachusetts. Tissue solubilizer
and scintillation cocktail (Ready Sol-HP) from Beckman were
used. The hydrochloride salts of the oximes were prepared in
distilled water for pharmacological testing.
Mu sca r in ic Recep tor Bin d in g Stu d ies. The binding of
the synthesized compounds to muscarinic M1, M2, and M3
receptor subtypes was determined using the cerebral cortex,
heart, and submandibulary glands of the rat, respectively.
Male SD rats (200-250 g) were purchased from the Laboratory
Animal Centre of the National University of Singapore. The
animals were given free access to food and water and housed
in groups of four under standard conditions of humidity (60%),
room temperature (22 °C), and 12-h light-dark cycles. The
rats were killed by cervical dislocation. The tissues (cerebral
cortex, heart, and submandibulary glands) were removed and
cleaned of adhering tissue on an ice-cold surface.
For the submandibulary glands and cerebral cortex, the
tissues were homogenized with a Potter-Elvejhem homogenizer
in 10 volumes of Na+/Mg2+ HEPES buffer (pH 7.4), containing
NaCl 10 mM, MgCl2 10 mM, and HEPES 20 mM. In the case
of the homogenate from submandibulary glands, centrifugation
was carried out at 2000g for 10 min. The supernatant was
decanted and centrifuged at 20000g for 20 min. In the case
of the cerebral cortex, the homogenate was centrifuged only
once at 20000g for 20 min. For both tissues, the pellet was
resuspended in 10 volumes of Na+/Mg2+ HEPES buffer and
resedimented at 20000g, 10 min. The washing procedure was
repeated once. Finally the pellet was resuspended in 40
volumes of Na+/Mg2+ HEPES buffer and kept in liquid nitrogen
till use.
agonist activity, and acceptable levels of activity and
selectivity are attainable in rings of moderate flexibility.
In conclusion, the present results indicate that the
tropinone oximes are interesting lead compounds for
agonist activity at M1 receptors. Some members of this
class have been shown to bind selectively to M1 receptors
and to exert mnemonic effects in mice. It is particularly
interesting to note that functional assays for M2 activity
on the denuded rabbit endothelium reveal that these
oximes have no M2 agonist action.
Exp er im en ta l Section
Ch em istr y. Melting points were determined on a Gallen-
kamp apparatus and are uncorrected. Mass spectra were
determined on a VG Micromass 7035 E mass spectrometer
(with chemical ionization). IR spectra were recorded on a
Philips PU 9624 FTIR instrument in pressed KBr disks or
neat. 1H NMR spectra were recorded on a Bruker ACF (300
MHz) spectrometer. Chemical shifts are reported in δ (ppm)
relative to Me4Si as an internal standard. Elemental analyses
were performed by the Department of Chemistry, National
University of Singapore, and are in agreement with calculated
values within (0.4%, unless indicated.
Gen er a l Meth od for th e Syn th esis of N-(Alk yn yloxy)-
p h th a lim id es 6-14.13,14 A 0.044-mol (19.15 g) portion of
diethyl azodicarboxylate (40% solution in toluene) was added
dropwise to a stirred solution of the alcohol (0.04 mol),
N-hydroxyphthalimide (0.04 mol, 6.52 g), and triphenylphos-
phine (0.04 mol, 10.50 g) dissolved in tetrahydrofuran (100
mL). An exothermic effect was observed, and the mixture
turned red upon addition. After stirring for 24 h at room
temperature, the solvent was removed in vacuo. Ether was
added to the residue to precipitate triphenylphosphine oxide
and diethyl hydrazinedicarboxylate, which were filtered off.
The filtrate was evaporated under reduced pressure and the
residue applied to a silica gel column which was eluted with
benzene to give the N-alkoxyphthalimide on recrystallization
from ethanol. The following phthalimides were synthesized
in this way: N-(prop-2-ynyloxy)phthalimide (6), N-(1-methyl-
prop-2-ynyloxy)phthalimide (7), N-(but-2-ynyloxy)phthalimide
(8), N-(but-3-ynyloxy)phthalimide (9), N-(1-ethylprop-2-ynyl-
oxy)phthalimide (10), N-(pent-2-ynyloxy)phthalimide (11),
N-(pent-3-ynyloxy)phthalimide (12), N-(pent-4-ynyloxy)phthal-
imide (13), and N-(1-methylbut-3-ynyloxy)phthalimide (14).
Their yields, melting points, and IR and 1H NMR data are
given in Table 1.
Gen er a l Meth od for th e Syn th esis of O-Alk yn ylh y-
d r oxyla m in e (15-23) Hyd r och lor id es.15 A mixture of
N-alkynylphthalimide (0.02 mol) and hydrazine monohydrate
(0.021 mol, 1 mL) was vigorously shaken at room temperature
for 15 min. Diethyl ether (50 mL) was added, and the mixture
was shaken for another 15 min before it was filtered. The
residue was washed again with 3 × 30-mL portions of ether.
Ethereal HCl was added dropwise to the combined ethereal
filtrate to precipitate out the O-substituted hydroxylamine
hydrochloride which was filtered, washed with ether, and
recrystallized from 2-propanol/ether. The following hydroxy-
lamines were synthesized in this way: O-prop-2-ynylhydroxy-
lamine (15), O-(1-methylprop-2-ynyl)hydroxylamine (16), O-but-
2-ynylhydroxylamine (17), O-but-3-ynylhydroxylamine (18),
O-(1-ethylprop-2-ynyl)hydroxylamine (19), O-pent-2-ynylhy-
droxylamine (20), O-pent-3-ynylhydroxylamine (21), O-pent-
4-ynylhydroxylamine (22), and O-(1-methylbut-3-ynyl)hydrox-
ylamine (23). The yields of the HCl salts, their melting points,
and IR and 1H NMR data are given in Table 2. The
hydrochlorides of O-methylhydroxylamine and hydroxylamine
were obtained commercially from Tokyo Kasei Kogyo Pte Ltd.
(Tokyo, J apan).
In the case of the rat heart ventricle, homogenization was
carried out in 5 volumes of 0.25 M sucrose and 25 mM TRIS
buffer (pH 7.5) using a Potter-Elvejhem homogenizer for 30 s.
The homogenate was sedimented at 10000g (20 min), and the
pellet was discarded. The supernatant was decanted and
resedimented at 45000g (30 min). The second pellet was
resuspended in 5 volumes of 0.6 M KCl and 30 mM histidine
at pH 7.0 to solubilize actin and myosin filaments and
resedimented at 45000g (30 min) to yield the crude membrane
vesicles. The yield was approximately 1 mg protein/g heart.
The pellets were washed three times with 5 volumes of Na+/
Mg2+ HEPES buffer (pH 7.4). Finally, the pellet was resus-
pended in 20 volumes of the same buffer and kept in liquid
nitrogen. Protein was determined by the method of Lowry et
al.,30 using bovine serum albumin as standard.
Gen er a l Meth od for th e Syn th eses of Tr op in on e a n d
N-Meth ylp ip er id in on e Oxim e Eth er s 24-34.16 Equimolar
quantities of the ketone (tropin-3-one, N-methylpiperidin-4-
one) and O-substituted hydroxylamine (or hydroxylamine) HCl
were stirred in methanol at room temperature for 18 h.
The muscarinic receptor affinities of the oximes were
determined from competitive experiments in which the ability
of varying concentrations of the test compound to displace a
fixed concentration of a receptor specific radiolabeled ligand
was monitored. [3H]Pirenzepine and [3H]-N-methylscopola-

3230 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 17
Xu et al.
mine were used as the specific radioligands for the cerebral
cortex (M1) and the heart (M2)/submandibulary gland (M3),
respectively.
test compound was evaluated from a plot of the percent
contractile response of phenylephrine versus concentration of
test compound using a commercial software program, Phar-
macological Calculation System Version 4.2.33
In a typical assay, an aliquot (0.2 mL, containing an
equivalent of 8.0-1.0 mg of protein/mL) of the resuspended
pellets prepared from rat cerebral cortices was incubated at
30 °C with a fixed concentration (0.5 nM) of [3H]pirenzepine
and increasing concentrations of the test compound (10 nM-
10 mM) for 45 min. Nonspecific binding to rat cerebral cortices
was determined from saturation studies in which 0.2-mL
aliquots of the tissue homogenate were incubated with in-
creasing concentrations of [3H]pirenzepine (0.1-5.0 nM) in the
absence/presence of 1 µM atropine in a total volume of 1 mL
for a period of 45 min. The incubation was terminated by
centrifugation (8000g, 3 min) at room temperature. The tube
containing the pellet was washed with 1.5 × 2 mL of saline to
remove free radioactivity, and the final pellet was allowed to
drain dry. Tissue solubilizer (200 µL) was then added, the
mixture was allowed to stand overnight, and radioactivity was
counted after addition of 4 mL of liquid scintillator. All
binding measurements were determined in triplicate.
A similar procedure was followed for determining M2 and
M3 binding affinities. For M2 binding, an aliquot (0.08 mL,
containing an equivalent of 0.45-0.50 mg of protein/mL) of
the resuspended pellet prepared from rat heart ventricles was
incubated at 30 °C with a fixed concentration (0.3 nM) of [3H]-
N-methylscopolamine and increasing concentrations of the test
compound (10 nM-10 mM) for 45 min. An aliquot (0.08 mL
containing an equivalent of 10-12 mg of protein/mL) of the
resuspended pellet prepared from rat submandibulary gland
was used for M3 binding experiments. For the saturation
experiments, 0.08-mL aliquot of the tissue homogenate (heart
ventricle or submandibulary glands) was incubated with
increasing concentrations of [3H]-N-methylscopolamine (0.1-
2.5 nM) in the absence or presence of 1 µM (()-QNB. The
M1, M2, and M3 binding affinities of oximes 24a , 26a , and 29a
were also determined in the presence of 30 µM Gpp(NH)p.
Beh a vior a l Testin g in Mice. Swiss albino mice (male,
20 ( 2 g) were required to perform a water escape task which
was modified from the standard version of the Morris water
maze.34 The task consisted of locating a circular, transparent
platform, measuring 8 cm in diameter, which was located at
the center of a temperature-regulated (25 ( 1 °C) pool
measuring 60 × 60 × 10 cm3. The platform was submerged 1
cm below the water surface, and the water was colored black
to make the platform invisible. The mouse was always
introduced at the same corner of the pool and the time taken
for it to locate and mount the platform (the latency period)
was used as the dependent measure.
In the first experimental setup, mice were divided into three
groups of 9. The control group received saline, another group
received scopolamine (1 mg/kg, ip), and the third group
received drug (26a , 28a , 32a , or 33a at 2 mg/kg, sc) +
scopolamine (1 mg/kg, ip). The drugs were freshly prepared
each day and administered daily to the mice 60 min before
starting the experiment. Each mouse was then subjected to
the water maze task. In addition, one drug (28a ) was
investigated at different dose levels (5-0.2 mg/kg) by the same
procedure.
In the second experimental setup, the mice were divided
into two groups of 10 mice. Each mouse was tested individu-
ally over a period of 5 days or longer. On day 6 (or day 16 in
mice tested for a longer period), one group received saline and
the other group received the drug (26a , 28a , 32a , 33a , or
McN-A-343) at 2 mg/kg, sc. After 30 min, scopolamine (1 mg/
kg, ip) was administered to mice in both groups, and they were
subjected to the water escape task 30 min later. The responses
of the two groups of mice to the same task were evaluated
again the following day.
Sta tistica l An a lysis. The effects of drugs on the daily
performance (latency period) were calculated by analysis of
variance (ANOVA) with a split-plot design (between-within
subjects). The pairwise comparisons were performed by two-
tailed t-test and Mann-Whitney U-test.
Deter m in a tion of Ca p a city F a ctor s (k′) of Oxim es
fr om Rever sed -P h a se HP LC. The hydrophobicity of the
oximes was assessed from their capacity factors (k′) determined
by reverse-phase HPLC on a LiChrosorb R RP-18 (10 µm)
stationary phase, using as mobile phase varying concentrations
of methanol and a buffer (pH 3.2) prepared from 0.175 M acetic
acid and 0.014 M triethylamine in deionized water. Determi-
nations were carried out at 30 °C with a flow rate of 1.5 mL/
min and diode-array detector wavelengths set at 254 and 240
nm. A stock solution (0.1 g/mL) of the oxime was prepared in
methanol. An aliquot (20 µL) of a solution (1 mL) prepared
from the stock solution (0.1 mL), acetone (0.05 mL), and the
mobile phase was injected into the column, and the retention
time was determined (in triplicate). Five different mobile
phase compositions, varying from 50% to 75% (w/w) methanol
content, were used for each compound. The capacity factor
(k′) was determined from log k′ ) log[(Vs - Vo)/Vo], where Vs
and Vo are the retention volumes (retention time × flow rate)
of the oxime and acetone, respectively. Linear regression of
log k′ of each compound against mobile phase composition and
extrapolation to 100% aqueous phase gave log kw of each
compound at pH 3.2.
Molecu la r Mod elin g Meth od s. The following parameters
were determined from the energy-minimized conformations of
the oximes (as free bases) obtained by minimization using the
SYBYL 6.2 (Tripos Associates, St. Louis, MO) Tripos force field
(MAXIMIN2), with calculations continued until the rms gradi-
ent was less than 0.001 kcal mol -1 Å: ClogP, molecular volume
(MV), molecular refractivity (MR), total dipole moment (DM),
Connolly surfaces (volume and surface area, calculated from
MOLCAD in SYBYL), and negative charge on the oxime O
using the Gasteiger Huckel method. Total dipole moment is
calculated from the protonated oximes. The ClogP of the
oximes could not be calculated directly. Modified ClogP values
were obtained by addition of log P values of the tropinone/N-
The displacement data were analyzed using the software
Equilibrium Binding Data Analysis (EBDA)31 to obtain Hill
coefficients and IC50 values for competing ligands. Affinities,
expressed as pKi, the negative logarithm of the inhibition
constants Ki, were calculated from IC50 values using the
Cheng-Prusoff equation.32 The binding parameters (Kd, Bmax
)
of [3H]pirenzepine and [3H]-N-methylscopolamine were also
determined from saturation studies described above.
F u n ction a l in Vitr o Stu d ies. Aortic rings were prepared
from rat and rabbit aortas. The animal (male Sprague-
Dawley rats, 275 ( 25 g, or male rabbits, 1.75 ( 0.25 kg) was
paralyzed by cervical dislocation and sacrificed immediately
by decapitation, and the aortic rings were isolated as de-
scribed.24,25 The endothelium of the rabbit aorta was removed
by gently rubbing the intima with a cotton bud for 30-60 s.
Each aorta was dissected free of adhering tissue, cut into 2-mm
segments, and mounted in an organ bath containing Krebs-
Ringer bicarbonate solution of composition (mM) NaCl 118,
KCl 5, NaHCO3 25, glucose 10, CaCl2 2.5, MgSO4‚7H2O 1.2,
KH2PO4 1.2, and EDTA 0.026, under a resting tension of 1 g
(1.5 g for rabbit aorta) under conditions previously described.24
After an equilibration period of 60 min (at 37 °C, 95% O2-
5% CO2 aeration), phenylephrine (10-7 M) was added to the
bath to produce a contraction of the rat aortic ring which was
sustainable for 15 min. ACh (10-6 M) was then added to
produce a relaxation response which confirmed the intactness
of the endothelium in the aortic ring. After 30 min of
re-equilibration with two washings, phenylephrine (10-7 M)
was added to produce a contraction. Cumulative concentra-
tions of the test compound (10-6-10-4 M) were then added at
2-min intervals to produce a relaxation response. For the
rabbit aortic ring, phenylepherine (10-7 M) was added to the
bath after the equilibration period to produce a submaximal
increase in the tone of the ring. This was followed by the
cumulative additions (10-6-10-4 M) of test compound at 2-min
intervals. Each rat or rabbit aortic ring was used only once
for testing against a test compound. Each test compound was
evaluated using at least three concentrations. The EC50 of each

O-Alkynyloximes as Muscarinic Agonists
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 17 3231
(10) Sanders, K. B.; Thomas, A. J .; Pavia, M. R.; Davis, R. E.;
Coughenour, L. L.; Myers, S. L.; Fisher, S.; Moos, W. H.
Cholinergic agents: aldehyde, ketone and oxime analogues of
the muscarinic agonist UH5. BioMed. Chem. Lett. 1992, 2, 803-
808.
(11) Plate, R.; Plaum, M. J . M.; de Boer, T.; Andrews, J . S. Synthesis
and muscarinic M3 pharmacological activities of 1-azabicyclo-
[2.2.2]octan-3-one oxime derivatives. BioMed. Chem. 1996, 4,
239-245.
(12) Tecle, H.; Lauffer, D. J .; Mirzadegan, T.; Moos, W. H.; Moreland,
D. W.; Pavia, M. R.; Schwarz, R. D.; Davis, R. E. Synthesis and
SAR of bulky 1-azabicylco[2.2.1]-3-one oximes as muscarinic
receptor subtype selective agonists. Life Sci. 1993, 52, 505-511.
(13) Grochowski, E.; J urczak, J . A new synthesis of O-alkylhydroxy-
lamines. Synthesis 1976, 682-684.
(14) Mitsunobu, O. The use of diethyl azodicarboxylate and triph-
enylphosphine in synthesis and transformation of natural
products. Synthesis 1981, 1-28.
(15) Leclerc, G.; Bieth, N.; Schwartz, J . Synthesis and â-adrenergic
blocking activity of new aliphatic oximes ethers. J . Med. Chem.
1980, 23, 620-624.
(16) Galons, H.; Fiet, J .; Combet-Farnoux, C.; Miocque, M.; Bram,
G. Organic syntheses without solvent: preparation of alkoxyph-
thalimides and of alkoxylamines. Mol. Cryst. Liq. Cryst. Inc.
Nonlin. Opt. 1988, 161, 521-528.
(17) Lauffer, D. J .; Moos, W. H. Azabicyclo and azacyclo oxime and
amine cholinergic agents and pharmaceutically acceptable salts
thereof. European Patent Appl. EP 0445 731 A1, 1991, 1-32.
(18) Wolfe, B. B.; Yasuda, R. P. Development of selective antisera
for muscarinic cholinergic receptor subtypes. Ann. N. Y. Acad.
Sci. 1995, 151, 186-193.
(19) Levey, A. I.; Kitt, C. H.; Simonds, W. F.; Price, D. L.; Brann, M.
R. Identification and localization of muscarinic acetylcholine
receptor proteins with subtype-specific antibodies. J . Neurosci.
1991, 11, 3218-3226.
methylpiperidinone fragment and the O-substituted side chain
for each oxime. Verloop’s STERIMOL substituent parameters
(L, B1-B5) were calculated using the computational chemistry
program Physical Properties! Pro Revision 2.3 (CSW, Fairfield,
CA). L refers to the substituent length measured along the
axis formed by the bond between the substituent and the atom
to which it is attached. B1-B5 are radii extending perpen-
dicular to L1 and are also perpendicular to each other. B1
and B5 give the minimum and maximum widths of the
substituents, respectively. As defined in the program, the
following pairs (B1, B4) and (B2, B3) are in opposite directions.
Sta tistica l Meth od s. Multiple linear regression analyses
were carried out using SPSS for Windows (SPSS Inc., Chicago,
IL). The following statistical parameters were determined for
each regression equation: 95% confidence interval variables,
correlation coefficient r, measure of explained variance r2,
Fisher significance ratio F, and standard error SE. Cross-
validated r2 and SE were determined using the QSAR module
of SYBYL 6.2.
Ack n ow led gm en t. This work has been supported
by a grant (RP920376) from the National University of
Singapore. Rong Xu gratefully acknowledges the Na-
tional University of Singapore for granting her a
research scholarship.
Su p p or tin g In for m a tion Ava ila ble: Tables containing
physical data of 6-14 and 15-23 and QSAR parameters of
24a ,b-34a ,b (4 pages). Ordering information is given on any
current masthead page.
(20) Peralta, E. G.; Ashkenazi, A.; Winslow, J . W.; Smith, D. H.;
Ramachandran, J .; Capon, D. J . Distinct primary structure,
ligand binding properties and tissue-specific expression of four
human muscarinic acetylcholine receptors. EMBO J . 1987, 6,
3923-3929.
Refer en ces
(1) (a) Growden, J . H. Muscarinic agonists in Alzheimer’s Disease.
Life Sci. 1997, 60, 993-998. (b) Muller, D. M.; Mendla, K.;
Farber, S. A.; Nitsch, R. M. Muscarinic receptor agonists increase
the secretion of the amyloid precursor protein ectodomain. Life
Sci. 1997, 60, 985-991. (c) J aen, J .; Barrett, S.; Brann, M.;
Callahan, M.; Davis, R.; Doyle, P.; Eubanks, D.; Lauffer, D.;
Lauffer, L.; Lipinski, W.; Moreland, D.; Nelson, C.; Raby, C.;
Schwarz, R.; Spencer, C.; Tecle, H. In vitro and in vivo evaluation
of the subtype selective muscarinic agonist PD 151832. Life Sci.
1995, 56, 845-852. (d) J aen, J .; Moos, W. H.; J ohnson, G.
Cholinomimetics and Alzheimer’s Disease. BioMed. Chem. Lett.
1992, 2, 777-780.
(2) (a) Whitehouse, P. J .; Price, D. L.; Sruble, R. G.; Clark, A. W.;
Coyle, J . T.; DeLong, M. R. Alzheimer’s Disease and senile
dementia: Loss of neurons in the basal forebrain. Science 1982,
215, 1237-1239. (b) Davies, P. Neurotransmitter related en-
zymes in senile dementia of the Alzheimer’s type. Brain Res.
1979, 171, 319-327.
(3) (a) Coyle, J . T.; Price, D. L.; DeLong, M. R. Alzheimer’s
Disease: A disorder of cortical cholinergic innervation. Science
1983, 219, 1184-1190. (b) Davis, P.; Maloney, A. F. J . Selective
loss of central cholinergic neurons. Lancet 1976, 2, 1403. (c)
Perry, E. K.; Perry, R. H.; Blessed, G.; Tomlinson, B. E. Necropsy
evidence of central cholinergic deficits in senile dementia. Lancet
1977, 1, 189.
(4) (a) J ohns, C. A.; Haroutunian, V.; Davis, B. M.; Horvath, T. B.;
Thomas, B.; Mohs, R. C.; Davis K. L. Acetylcholinesterase
inhibitors in Alzheimer’s Disease and animal models. Proc. Meet.
Int. Study Group Treat. Mem. Disord. Assoc. Aging, 3rd;
Cambridge, MA, 1984; pp 349-373. (b) Huff, F. J .; Mickel, S.
F.; Corkin, S.; Growden, J . H. Cognitive functions affected by
scopolamine in Alzheimer’s Disease and normal aging. Drug.
Dev. Res. 1988, 12, 271-278. (c) Bartus, R. T.; Dean, R. L.; Beer,
B. An evaluation of drugs for improving memory in aged
monkeys: implications for clinical trials in humans. Psycho-
pharmacol. Bull. 1983, 19, 168-184.
(21) Dorje, F.; Levey, A. I.; Brann, M. R. Immunological detection of
muscarinic receptor subtype proteins (m1-M5) in rabbit periph-
eral tissues. Mol. Pharmacol. 1991, 40, 459-462.
(22) Watson, M.; Yamamura, H.; Roeske, W. R. [3H]Pirenzepine and
(-)-[3H]quinuclidinyl benzilate binding to rat cerebral cortex and
cardiac muscarinic cholinergic sites. Characterization and regu-
lation of agonist binding to putative muscarinic subtypes. J .
Pharmacol. Exp. Ther. 1986, 237, 411-418.
(23) Sim, M. K.; Lim, B. C. Presence of an endothelial esterase in
the rat aorta: effects on the actions of ester and nonester
muscarinic antagonists. Endothelium 1993, 1, 109-114.
(24) J aiswal, N.; Lambrecht, G.; Mutschler, E.; Tacke, R.; Malik, K.
U. Pharmacological characterization of the vascular muscarinic
receptors mediating relaxation and contraction in rabbit aorta.
J . Pharmacol. Exp. Ther. 1991, 258, 842-850.
(25) Hansch, C.; Leo, A. Substituent Constants for Correlation
Analysis in Chemistry and Biology; J ohn Wiley & Sons: New
York, 1979; pp 69-167.
(26) Tute, M. S. History and Objectives of Quantitative Drug Design.
In Comprehensive Medicinal Chemistry; Hansch, C., Sammes,
P. G., Taylor, J . B., Eds.; Pergamon Press: New York, 1990; Vol.
4, pp 1-28.
(27) Connolly, M. L. Solvent accessible surfaces of proteins and
nucleic acids. Science 1983, 221, 709-713.
(28) Verloop, A.; Hoogenstraaten, W.; Tipker, J . Development and
Application of New Steric Substituent Parameters in Drug
Design. In Drug Design; Ariens E. J ., Eds.; Academic Press: New
York, 1976; Vol. 7, pp 165-207.
(29) Shapiro, G.; Floersheim, P.; Amstutz, R.; Boddeke, H.; Bolliger,
G.; Cottens, S.; Enz, A.; Gmelin, G.; Gull, P.; Supavilai, P.
BioMed. Chem. Lett. 1992, 2, 815-820.
(30) Lowry, O. H.; Rosebrough, N. J .; Far, A. L.; Randall, R. J . Protein
measurements with the Folin Phenol Reagent. J . Biol. Chem.
1951, 193, 265-275.
(31) McPherson, G. A. Analysis of radioligand binding experiments.
A collection of computer programs for the IBM PC. J . Pharmacol.
Methods 1985, 14, 213-228.
(5) Nitsch, R. M.; Slack, B. E.; Wurtman, R. J .; Growdon, J . H.
Release of Alzheimer amyloid precursor derivatives stimulated
by activation of muscarinic acetylcholine receptors. Science 1992,
258, 304-307.
(6) Caulfield, M. D. Muscarinic receptors - characterization, cou-
pling and function. Pharmacol. Ther. 1993, 58, 319-379.
(7) Eglen, R. M.; Watson, N. Selective muscarinic receptor agonists
and antagonists. Pharmacol. Toxicol. 1996, 78, 59-68.
(8) Kenakin, T. P.; Bond, R. A.; Bonner, T. I. Definition of pharma-
cological receptors. Pharmacol. Rev. 1992, 44, 351-362.
(9) Bromidge, S. M.; Brown, F.; Cassidy, F.; Clark, M. S. G.; Dabbs,
S.; Hawkins, J .; Loudon, J . M.; Orlek, B. S.; Riley, G. J . A novel
and selective class of azabicyclic muscarinic agonists incorporat-
ing an N-methoxy imidoyl halide or nitrile functionality. BioMed.
Chem. Lett. 1992, 2, 791-796.
(32) Cheng, Y. C.; Prusoff, W. H. Relationship between the inhibition
constant (Ki) and the concentration of inhibitor which causes
50% inhibition (I50) of enzymatic reaction. Biochem. Pharmacol.
1973, 22, 3099-3108.
(33) Tallarida, R. J .; Murray, R. B. Manual of Pharmacologic
Calculations with Computer Programs; Springer Verleg: New
York, 1987.
(34) Morris, R. G. M. Spatial localization does not require the
presence of local cues. Learn. Motiv. 1981, 12, 239-249.
J M9708588