Synthesis and biological characterization of 1,4,5,6- tetrahydropyrimidine and 2-amino-3,4,5,6-tetrahydropyridine derivatives as selective m1 agonists

1230

J . Med. Chem. 1997, 40, 1230-1246

Syn th esis a n d Biologica l Ch a r a cter iza tion of 1,4,5,6-Tetr a h yd r op yr im id in e a n d

2-Am in o-3,4,5,6-tetr a h yd r op yr id in e Der iva tives a s Selective m 1 Agon ists

William S. Messer, J r.,* Yahaya F. Abuh, Yang Liu, Sumudra Periyasamy,^†Dan O. Ngur, Michael A. N. Edgar,^‡

Afif A. El-Assadi, Sbeih Sbeih, Philip G. Dunbar,^§Scott Roknich, Taikyun Rho, Zheng Fang, Babatunde Ojo,

Hao Zhang,^|J ames J . Huzl, III, and Peter I. Nagy

Center for Drug Design and Development, Department of Medicinal & Biological Chemistry, College of Pharmacy,

The University of Toledo, 2801 West Bancroft Street, Toledo, Ohio 43606

Received J une 28, 1996^X

Previous studies identified several novel tetrahydropyrimidine derivatives exhibiting muscarinic

agonist activity in rat brain. Such compounds might be useful in treating cognitive and memory

deficits associated with low acetylcholine levels, as found in Alzheimer’s disease. To determine

the molecular features of ligands important for binding and activity at muscarinic receptor

subtypes, the series of tetrahydropyrimidines was extended. Several active compounds were

examined further for functional selectivity through biochemical studies of muscarinic receptor

activity using receptor subtypes expressed in cell lines. Several amidine derivatives displayed

high efficacy at m1 receptors and lower activity at m3 receptors coupled to phosphoinositide

(PI) metabolism in A9 L cells. Four ligands, including 1b, 1f, 2b, and 7b, exhibited marked

functional selectivity for m1 vs m3 receptors. Compound 1f also exhibited low activity at m2

receptors coupled to the inhibition of adenylyl cyclase in A9 L cells. Molecular modeling studies

also were initiated to help understand the nature of the interaction of muscarinic agonists

with the m1 receptor using a nine amino model of the m1 receptor. Several important

interactions were identified, including interactions between the ester moiety and Thr192.

Additional interactions were found for oxadiazoles and alkynyl derivatives with Asn382,

suggesting that enhanced potency and selectivity may be achieved by maximizing interactions

with Asp105, Thr192, and Asn382. Taken together, the data indicate that several amidine

derivatives display functional selectivity for m1 muscarinic receptors, warranting further

evaluation as therapeutic agents for the treatment of Alzheimer’s disease. In addition, several

amino acid residues were identified as potential binding sites for m1 agonists. These data

may be useful in directing efforts to develop even more selective m1 agonists.

Alzheimer’s disease is associated with a loss of the

basal forebrain neurons that release acetylcholine in the

cerebral cortex and hippocampus.^1-4Decreased levels

of acetylcholine may account for the memory deficits and

cognitive dysfunction associated with the early stages

of the progressive disorder. Acetylcholinesterase inhibi-

tors (e.g., physostigmine, tacrine) have been utilized to

elevate acetylcholine levels and treat the symptoms of

Alzheimer’s disease, yet provide limited therapeutic

benefit, in part due to the high incidence of undesirable

side effects and the progressive loss of the target

enzyme.^5,6

Over the past decade, drug development efforts have

focused on identifying muscarinic agonists that directly

activate receptors on postsynaptic neurons.^7,8Four

subtypes of muscarinic receptor have been identified on

the basis of pharmacological studies (designated M₁-

M₄), while molecular biological studies have identified

five muscarinic receptor subtypes (m1-m5). On the

basis of the preferential localization of M₁muscarinic

receptors in the hippocampus and cerebral cortex,^9-14

and their involvement in memory function,^15-18the

search has focused on selective M₁agonists. Such

ligands would selectively activate receptors coupled to

phosphoinositide metabolism in the cerebral cortex and

hippocampus^19-21without decreasing acetylcholine lev-

els by activating central M₂receptors^22-25or activating

peripheral M₂receptors involved in regulating cardiac

function^26,27and M₃receptors associated with increased

gastrointestinal and exocrine gland activity.^27,28

Over the past few years, several novel amidine

derivatives have been synthesized and tested for mus-

carinic agonist activity in rat brain.^19-31Several com-

pounds stimulated muscarinic receptors coupled to PI

metabolism in the cerebral cortex and hippocampus,

including the methyl (1a ), ethyl (1b), and propargyl (1f)

esters of 1,4,5,6-tetrahydropyrimidine, the methyl- and

ethyl-1,2,4-oxadiazole derivatives (2a and 2b, respec-

tively) of 1,4,5,6-tetrahydropyrimidine, and the methyl

(7a , 8a ) and propargyl (7b, 8b) esters of racemic

2-amino-3,4,5,6-tetrahydropyridine and 2-amino-1,4,5,6-

tetrahydropyrimidine, respectively.

^†Present address: Medical College of Ohio, P.O. Box 10008, Toledo,

OH 43699.

The series of amidine derivatives was extended

through replacement of the ester moiety and modifica-

tion of the alkyl functionality and tested for muscarinic

agonist activity in rat brain. Several compounds dis-

playing promising muscarinic agonist activity were

studied further to determine their selectivities for m1

receptors and their potential as therapeutic agents for

the treatment of Alzheimer’s disease. Agonist activity

was determined at m1, m2, and m3 receptors expressed

^‡Present address: Rudolf Magnus Institute For Neurosciences, 100

Universiteitsweg, 3584 CG Utrecht, The Netherlands.

^§Present address: Washington State University, Health Research

and Education Center, W. 601st Avenue, Spokane, WA 99204.

Present address: Georgia Institute Of Technology, School Of

Chemistry and Biochemistry, Gilbert H. Boggs Building, Room 3-8,

Atlanta, GA 30332-0400.

^|Present address: Bristol Myers Squibb Institute, P.O. Box 4000,

H12-03, Princeton, NJ 08543-4000.

^XAbstract published in Advance ACS Abstracts, March 15, 1997.

Selective m1 Agonists

J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 8 1231

Ch a r t 1

Sch em e 1

in A9 L cells through biochemical assays of PI metabo-

lism and cAMP formation. The data indicate that

several amidine derivatives display functional selectiv-

ity for m1 muscarinic receptors, warranting further

evaluation as therapeutic agents for the treatment of

Alzheimer’s disease.

Resu lts

Syn th etic Ch em istr y. The first series of tetrahy-

dropyrimidine esters (1d and 1e) was synthesized by

esterification of the reduced acid according to literature

procedures.^29,30The remaining esters (1g-s) were

prepared by first forming the acid chloride using a

solution of oxalyl chloride in benzene and then reacting

this with the desired alkynols (see Scheme 1).

A series of branched alkyl- and alkynyloxadiazole

derivatives of tetrahydropyrimidine (2c-k ) was also

synthesized using an extension of a variety of literature

procedures.^30,32,33An efficient single-vessel reaction of

the appropriately substituted amide oximes³²with 1

equiv of sodium hydride in refluxing dry THF led to the

corresponding anions, which, after addition of the trityl-

protected tetrahydropyrimidine ester, cyclized to the

5-oxadiazole-substituted tetrahydropyrimidine deriva-

tives. Deprotection of the N-trityltetrahydropyrimidine

oxadiazole with trifluoroacetic acid (TFA) gave the

desired products (2c-k ) as TFA salts (Scheme 2).³⁰

1232 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 8

Messer et al.

Sch em e 2

Sch em e 4

Sch em e 3

1m , and 1r . All oxadiazole derivatives, with the excep-

tion of butynyl (2f), displayed high potency for musca-

rinic receptors in rat brain. The ether (5a and 5b) and

oxime (6a , 6b and 6c) derivatives displayed much lower

potency with only 5b exhibiting an IC₅₀value less than

100 µM.

All compounds with moderate affinity (IC₅₀values <

100 µM) for muscarinic receptors in the rat brain were

tested further for agonist activity through assays of PI

metabolism in slices from rat cerebral cortex. Several

compounds displayed agonist activity in the PI assay,

including the butynyl esters 1g and 1h , the 2-methyl-

butyl oxadiazole derivative 2e, the butynyl oxadiazole

derivative 2f, and the ether and thioether oxadiazoles

2h -k . None of the ether or oxime derivatives exhibited

agonist activity in the rat brain.

Recep tor Bin d in g in Cell Lin es. Cell lines ex-

pressing a single subtype of muscarinic receptor pro-

vided a suitable measure of muscarinic agonist affinity,

activity, and selectivity. Ligand affinity for muscarinic

receptor subtypes was determined through inhibition

of specific [³H]-(R)-QNB binding to m1, m2, or m3

receptors expressed in A9 L cells. Examples from the

tetrahydropyrimidine and tetrahydropyridine series of

ligands were selected for binding studies to permit a

comparison of the relative contributions of the different

functional groups to agonist affinity.

As shown in Table 2, each cell line displayed a

different level of muscarinic receptors, with m3 > m1

> m2. Inhibition studies revealed that carbachol ex-

hibited highest affinity for m2 receptors (see Figure 1),

and binding curves reflected an interaction with mul-

tiple receptor states. All other agonists examined also

displayed the highest affinity for m2 receptors, as can

be seen in Tables 3-5.

As shown in Figures 1-5 and Table 3, the inhibition

curves for carbachol, 1a , and 2a were biphasic at m1

receptors expressed in A9 L cells. Arecoline and 4

appeared to bind to a single population of m1 receptors.

The nonhydrolyzable GTP analog, GTP-γ-S, lowered

agonist affinity at m1 receptors. Binding isotherms for

carbachol, 1a , and 2a in the presence of GTP-γ-S were

consistent with a single low-affinity binding site.

The binding of [³H]-(R)-QNB, carbachol, and arecoline

to m1 muscarinic receptors expressed in A9 L cells was

similar to that reported previously,³⁸and consistent

with similar studies using brain membranes.³⁹At each

4-Cyanobutyne, which served as a starting material

for the oxadiazole 2f, was prepared from the corre-

sponding alcohol according to the method of Brands-

ma.³³3-Butyn-1-ol was reacted with p-toluenesulfonyl

chloride to obtain the tosylate, followed by reaction with

sodium cyanide in DMSO to give the corresponding

nitrile using an extension of literature procedures.³³The

nitrile was converted to its amide oxime,³²followed by

reaction with trityl ester to form the oxadiazole (2f),

employing a method virtually identical to that used for

2a -e (Scheme 2).

The tetrahydropyrimidine ether derivative (5b) was

synthesized from the reduced alcohol. 1,4,5,6-Tetrahy-

dro-5-hydroxypyrimidine^34,35was reacted with propargyl

chloride in the presence of potassium carbonate to give

the propargyl ether derivative as the free base, which

then was converted to the hydrochloride salt (Scheme

3).

A series of tetrahydropyrimidine oxime derivatives

(6a -c) was synthesized from the reduced ketone using

an extension of a literature procedure.^36,37As shown

in Scheme 4, 5-hydroxy-1,4,5,6-tetrahydropyrimidine

was oxidized to 1,4,5,6-tetrahydropyrimidin-5-one by

reacting the alcohol with a suspension of pyridinium

chlorochromate in dichloromethane. The ketone was

then reacted with the desired alkoxylamine according

to literature procedures.^36,37

R ecep t or Bin d in g a n d P I Met a b olism in R a t

Br a in . Muscarinic receptor binding and agonist activ-

ity were examined in rat brain tissues in preliminary

fashion for all new ligands. Receptor binding affinity

was assessed through inhibition of specific [³H]-(R)-QNB

binding to rat brain membranes. Several novel ligands

exhibited high affinity (IC₅₀values < 10 µM) for

muscarinic receptors in rat brain, as shown in Table 1.

Ester derivatives with the highest affinity included 1l,

Selective m1 Agonists

J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 8 1233

Ta ble 1. Binding Affinities and Agonist Activity in a Series of Ester (1), 1,2,4-Oxadiazole (2), Ether (5), and Oxime (6) Derivatives at

Muscarinic Receptors in the Rat Central Nervous System^a

functionality (R)

ligand

IC₅₀(µM) [³H]QNB

PI cortex (% at 100 µM)

carbachol

arecoline

pilocarpine

5.5 ( 1.0

1.0 ( 0.25

7.6 ( 4.4

0.51 ( 0.10

9.2 ( 1.9

2.2 ( 0.28

3.3 ( 0.12

21 ( 2.2

15 ( 3.0

3.3 ( 0.80

13 ( 2.7

14 ( 2.8

14 ( 2.4

72 ( 9.8

>100

470 ( 81

110 ( 21

58 ( 5.6

80 ( 17

aceclidine

1a

1b

1c

1d

1e

1f

1g

1h

1i

1j

1k

1l

1m

1n

1o

1p

1q

1r

methyl

ethyl

isopropyl

2-methylbutyl

3-methylbutyl

propargyl

2-butynyl

butynyl

pentynyl

2-pentynyl

112 ( 11

150 ( 17

7.2 ( 3.0

-3.0 ( 2.0

2.0 ( 3.0

230 ( 35

140 ( 4.1

100 ( 15

-4.0 ( 1.5

-5.8 ( 4.3

5.8 ( 7.0

35 ( 5.6

50 ( 4.3

7.8 ( 1.6

24 ( 12

3-pentynyl

trans-3-methyl-2-penten-4-ynyl

cis-3-methyl-2-penten-4-ynyl

3-penten-1-ynyl

2-methylbutenyl

3-methylpropynyl

3,3-dimethylpropynyl

phenylpropynyl

propenyl

methyl

ethyl

isopropyl

2-methylpropyl

2-methylbutyl

butynyl

4.1 ( 0.63

1.5 ( 0.42

70 ( 20

38 ( 16

43 ( 2.8

>100

22 ( 3

9.3 ( 1.5

12 ( 5.2

2.7 ( 0.69

1.0 ( 0.32

5.3 ( 0.68

1.4 ( 0.12

6.1 ( 0.38

16 ( 6.9

2.3 ( 1.2

4.0 ( 1.5

5.1

23 ( 8.2

21 ( 12

700 ( 99

130 ( 16

60 ( 13

38 ( 11

180 ( 60

130 ( 18

42 ( 13

110

1s

2a

2b

2c

2d

2e

2f

2g

2h

2i

pentynyl

methoxymethyl

ethoxymethyl

ethoxyethyl

(methylthio)methyl

methyl

propargyl

methyl

ethyl

phenylpropynyl

100

2j

2.6 ( 1.3

5.1 ( 4.8

>100

100

180 ( 79

2k

5a

5b

6a

6b

6c

32 ( 21

-3.0

>100

a

Data represent the mean ((sem, where indicated) from one to three experiments, each performed in triplicate. Data for carbachol,

arecoline, pilocarpine, aceclidine, 1a , 1b, 1c, 1f, 2a , and 2b have been reported previously and are shown here for comparison.^29,30

Ta ble 2. Binding of [³H]-(R)-QNB to Muscarinic Receptor

Subtypes Expressed in A9 L Cells^a

receptor subtype

B_max(fmol/mg)

K_d(pM)

m1

m2

m3

280 ( 50

59 ( 7.9

1500 ( 100

43 ( 3.7

24 ( 0.77

68 ( 4.3

a

Data represent the mean ((sem) from three assays each

performed in triplicate.

muscarinic receptor subtype, carbachol displayed a

complex interaction suggesting multiple receptor affin-

ity states. As expected, GTP-γ-S decreased agonist

affinity for m1 receptors, and the magnitude of the shift,

which has been postulated to reflect agonist efficacy,³⁹

differed for each ligand. Carbachol, generally consid-

ered a full agonist, had a 3-fold shift in affinity, while

arecoline displayed a much smaller shift. GTP-γ-S

decreased the affinity of both 1a and 2a in a similar

fashion to carbachol, suggesting that the 1,4,5,6-tet-

rahydropyrimidine derivatives were able to induce m1

receptor coupling to G-proteins, which was confirmed

by studies of PI metabolism. The predicted efficacies

obtained from GTP-γ-S shift assays were in excellent

agreement with the results of the PI metabolism assays.

That is, carbachol and the two 1,4,5,6-tetrahydropyri-

midine derivatives appeared to be efficacious agonists

at m1 receptors with significantly higher activity than

either arecoline or 4.

F igu r e 1. Inhibition of [³H]-(R)-QNB binding to m1, m2, and

m3 receptors expressed in A9 L cells by carbachol in the

absence of GTP-γ-S (GTP[S]). Also shown is the inhibition of

the binding [³H]-(R)-QNB to m1 receptors by carbachol in

presence of 300 µM GTP-γ-S. Data represent the mean ((sem)

from three experiments, each performed in triplicate.

Carbachol displayed the highest affinity for m2 recep-

tors (vs m1 or m3 receptors), in agreement with previous

studies.⁴⁰The percentage of m2 receptors with high

agonist affinity suggests more efficient coupling of m2

receptors to G-proteins than either m1 or m3 receptors.

In contrast, agonist binding to low-affinity sites was

comparable at m1, m2, and m3 receptors. These data

1234 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 8

Messer et al.

Ta ble 3. Inhibition of [³H]-(R)-QNB Binding to m1 Receptors Expressed in A9 L Cells by Several Muscarinic Agonists in the Absence

and Presence of 300 µM GTP-γ-S^a

-GTP-γ-S

% low

+GTP-γ-S

ligand

% high

K_iHigh (nM)

739 ( 668

K_ilow (µM)

% binding

K_i(µM)

carbachol

arecoline

16.5 ( 3.13

83.4 ( 3.13

99.4 ( 1.17

98.5 ( 0.642

76.5 ( 5.25

68.5 ( 7.43

76.6 ( 3.45

76.8 ( 2.15

69.1 ( 1.70

67.5 ( 2.80

19.9 ( 6.60

3.33 ( 0.554

0.589 ( 0.0550

23.6 ( 1.47

2.88 ( 0.104

9.01 ( 2.07

15.5 ( 3.73

2.31 ( 0.333

17.4 ( 1.57

99.8 ( 0.841

101 ( 0.820

100 ( 0.325

97.9 ( 0.726

37.1 ( 15.4

5.45 ( 0.900

1.05 ( 0.144

26.8 ( 1.73

4

1a

1f

1g

1h

2a

2g

23.5 ( 5.25

31.5 ( 7.43

23.4 ( 3.45

23.2 ( 2.15

30.9 ( 1.70

32.5 ( 2.80

359 ( 95.7

41.7 ( 32.9

14.0 ( 2.64

30.9 ( 15.0

21.1 ( 4.61

138 ( 108

98.3 ( 0.529

2.50 ( 0.657

a

Data represent the mean ((sem) percentage of binding sites and K_ivalues for each site. K_ivalues were calculated using IC₅₀values

obtained by nonlinear curve fitting, accordingv to the best fit (p < 0.05) of the binding data to a one- or two-site model. The data are from

three separate binding assays, each determined in triplicate.

Ta ble 4. Inhibition of [³H]-(R)-QNB Binding to m2 Receptors Expressed in A9 L Cells by Several Muscarinic Agonists^a

ligand

% high

K_H(nM)

% medium

K_M(nM)

% low

K_L(µM)

carbachol

1a

2a

arecoline

4

43 ( 6.7

0.81 ( 0.34

40 ( 5.0

70 ( 2.4

66 ( 5.0

65 ( 2.7

60 ( 2.4

46 ( 31

17 ( 1.9

30 ( 2.4

34 ( 5.0

35 ( 2.7

40 ( 2.4

12 ( 5.9

14 ( 3.6

3.3 ( 0.58

1.1 ( 0.52

2.4 ( 0.33

0.94 ( 0.22

0.46 ( 0.28

0.56 ( 0.054

0.087 ( 0.047

a

Data represent the mean ((sem) percentage of binding sites and K_ivalues for each site. K_ivalues were calculated using IC₅₀values

obtained by nonlinear curve fitting, according to the best fit (p < 0.05) of the binding data to a one-, two-, or three-site model. The data

are from three separate binding assays, each determined in triplicate.

Ta ble 5. Inhibition of [³H]-(R)-QNB Binding to m3 Receptors

Expressed in A9 L Cells by Several Muscarinic Agonists^a

ligand

% high

K_H(nM)

56 ( 38

% low

K_L(µM)

carbachol

1a

2a

19 ( 3.8

80 ( 3.8

93 ( 0.44

95 ( 1.2

100 ( 1.6

71 ( 10

8.4 ( 0.72

3.3 ( 0.92

0.57 ( 0.11

0.87 ( 0.25

0.46 ( 0.072

arecoline

4

29 ( 10

16 ( 8.0

a

Data represent the mean ((sem) percentage of binding sites

and K_ivalues for each site. K_ivalues were calculated using IC₅₀

values obtained by nonlinear curve fitting, according to the best

fit (p < 0.05) of the binding data to a one- or two-site model. The

data are from three separate binding assays, each determined in

triplicate.

F igu r e 3. Inhibition of [³H]-(R)-QNB binding to m1, m2, and

m3 receptors expressed in A9 L cells by compound 4 in the

absence of GTP-γ-S (GTP[S]). Also shown is the inhibition of

the binding [³H]-(R)-QNB to m1 receptors by 4 in presence of

300 µM GTP-γ-S. Data represent the mean ((sem) from three

experiments, each performed in triplicate.

agonists solely on the basis of receptor affinity. Thus,

future drug development efforts should focus on en-

hancing functional selectivity rather than increasing

ligand potency. In addition, it will be important to

examine agonist affinity for m4 and m5 receptors.

F u n ction a l Activity a n d Selectivity. Activity at

m1 muscarinic receptors was examined using an A9 L

cell line expressing human m1 receptors. Classical

muscarinic agonists such as carbachol and oxotremo-

rine-M, which contain a quaternary amine, stimulated

PI metabolism in A9 L cells, while tertiary amines (e.g.,

oxotremorine, arecoline, and arecedaine propargyl ester

(3b)) displayed lower efficacy at m1 receptors (see Table

6). Both 1a and 2a stimulated PI metabolism to

maximal levels comparable to that of carbachol as

shown in Figure 6A. Compound 1a was much less

potent than carbachol, while 2a was even more potent.

As shown in Figure 6C and Table 6, 1f stimulated m1

receptors to a maximal level comparable to carbachol,

F igu r e 2. Inhibition of [³H]-(R)-QNB binding to m1, m2, and

m3 receptors expressed in A9 L cells by arecoline in the

absence of GTP-γ-S (GTP[S]). Also shown is the inhibition of

the binding [³H]-(R)-QNB to m1 receptors by arecoline in

presence of 300 µM GTP-γ-S. Data represent the mean ((sem)

from three experiments, each performed in triplicate.

suggest the differences in agonist potencies between

receptor subtypes may relate to the coupling of receptors

to different G-proteins. The finding that each agonist

bound with the highest affinity to m2 receptors suggests

that it will be difficult to develop selective muscarinic

Selective m1 Agonists

J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 8 1235

forskolin-stimulated adenylyl cyclase activity by greater

than 20%. The propargyl esters 1f and 7b were much

less effective in reversing the effects of forskolin on

cAMP levels.

The availability of cell lines stably expressing indi-

vidual muscarinic receptor subtypes provides useful

tools for exploring agonist activity and functional se-

lectivity. Depending upon the level of receptor expres-

sion, however, agonist responses in cell lines may not

reflect accurately the response observed in brain tis-

sue,⁴¹the relevant tissue for developing muscarinic

agonists to treat Alzheimer’s disease. The low activity

of arecoline in A9 L cells expressing either m1 or m3

receptors is consistent with data obtained for PI me-

tabolism in the CNS^20,42-44and cAMP formation in A9

L cells expressing m1 receptors.⁴⁵Furthermore, the

effects of carbachol and arecoline on cAMP accumulation

observed in A9 L cells are comparable to those observed

previously for brain tissue.^19,46Overall, these data

suggest that A9 L cells provide a suitable paradigm for

evaluating muscarinic receptor agonist activity and

selectivity, although further studies are necessary to

determine the level of agonist activity and subtype

selectivity of the responses in brain.

F igu r e 4. Inhibition of [³H]-(R)-QNB binding to m1, m2, and

m3 receptors expressed in A9 L cells by compound 1a in the

absence of GTP-γ-S (GTP[S]). Also shown is the inhibition of

the binding [³H]-(R)-QNB to m1 receptors by 1a in presence

of 300 µM GTP-γ-S. Data represent the mean ((sem) from

three experiments, each performed in triplicate.

The activity of 1,4,5,6-tetrahydropyrimidine deriva-

tives in A9 L cells expressing m1 receptors was consis-

tent with responses observed previously at muscarinic

receptors coupled to PI metabolism in the rat central

nervous system.^29,30Moreover, the agonist activity of

compounds 1a , 1f, and 2a at m1 receptors suggests that

they should be considered fully efficacious muscarinic

agonists (class A compounds according to the classifica-

tion proposed by Fisher).^20,43In rat hippocampal slices,

however, the maximal level of PI metabolism produced

by 1a is only a fraction of that produced by the

corresponding oxadiazole derivative (2a ).^29,30In those

studies, the M₁antagonist pirenzepine blocked the PI

response elicited by low doses of either 1a or 2a . The

M₂- and M₃-selective antagonists, AF-DX 116 and p-F-

hexahydrosiladifenidol, respectively, were less potent in

blocking agonist activity. On the basis of the compa-

rable efficacy displayed in the present studies, the

higher efficacy of 2a in rat hippocampus may be due to

activation of other muscarinic receptor subtypes.

F igu r e 5. Inhibition of [³H]-(R)-QNB binding to m1, m2, and

m3 receptors expressed in A9 L cells by compound 2a in the

absence of GTP-γ-S (GTP[S]). Also shown is the inhibition of

the binding [³H]-(R)-QNB to m1 receptors by 2a in presence

of 300 µM GTP-γ-S. Data represent the mean ((sem) from

three experiments, each performed in triplicate.

Taken together, the assays of functional activity at

and selectivity for muscarinic receptor subtypes indi-

cated a few agonists with high activity and functional

selectivity for m1 muscarinic receptors. In particular,

compounds 1b, 1f, and 2b exhibited functional selectiv-

ity for m1 receptors, with reasonable potency and

efficacy at m1 receptors, and lower efficacy at m2 and/

or m3 receptors. It should be noted that these com-

pounds compared favorably with xanomeline, an m1

agonist currently under clinical investigation for the

treatment of Alzheimer’s disease. In particular, com-

pounds 1b, 1f, and 2b showed higher efficacy than

xanomeline at m1 receptors expressed in A9 L cells.

while 1b and 2b displayed intermediate activity. Com-

pound 1f was approximately as potent as carbachol in

stimulating m1 receptors. Other tetrahydropyrimidine

derivatives, including 1g, 1h , and 2k , also exhibited

high efficacy at m1 receptors, although the butynyl ester

(1h ) and oxadiazole (2f) exhibited much lower potency.

Functional muscarinic receptor selectivity was exam-

ined using the A9 L cell line expressing a high level of

m3 receptors (see Figure 6B,D and Table 7). Carbachol

stimulated m3 receptors coupled to PI metabolism in

A9 L cells, although the maximum level of stimulation

was roughly half that found for m1 receptors expressed

in the same cell line. Overall, 1f, 1b, 2b, and 7b

displayed markedly lower efficacy than carbachol (see

Figure 6D), as did xanomeline in one experiment.

Agonist potencies at m3 receptors were comparable to

those found for m1 receptors.

Molecu la r Mod elin g. The relative energies and

molecular volumes for the different conformations of 1a

and arecoline are indicated in Table 9 and Figure 7A,B.

Molecular mechanics calculations using the modified

MM2 force field⁴⁷revealed three minimum energy

conformations for arecoline, yet six unique conformers

for 1a . Similar results were obtained using the modified

AMBER force field,⁴⁸except that two and five conforma-

tions were observed for arecoline and compound 1a ,

On the basis of the relative activity and selectivity of

compounds for m1 receptors, a few compounds were

evaluated for activity at m2 receptors coupled to the

inhibition of adenylyl cyclase activity in A9 L cells (see

Table 8). Carbachol, xanomeline, and 2a inhibited

1236 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 8

Messer et al.

Ta ble 6. Potency and Efficacy for a Series of Ligands at m1 Muscarinic Receptors Expressed in A9 L Cells^a

ligand

R group

EC₅₀(µM)

S_max(%)

carbachol

oxo-M

oxotremorine

pilocarpine

arecoline

3b

40 ( 15

630 ( 110

380 ( 36

120 ( 27

150 ( 18

180 ( 20

140 ( 30

98 ( 29

0.70 ( 0.30

0.27 ( 0.19

6.9 ( 1.4

27 ( 2.7

0.22 ( 0.07

12 ( 6.4

5.7 ( 2.3

4

xanomeline

180 ( 24

1a

1b

1f

1g

1h

1l

methyl

ethyl

200 ( 51

44 ( 5.2

33 ( 6.6

50 ( 6.5

560 ( 420

660 ( 110

330 ( 75

590 ( 92

430 ( 20

340 ( 56

propargyl

2-butynyl

butynyl

trans-3-methyl-2-penten-4-ynyl

cis-3-methyl-2-penten-4-ynyl

1m

80 ( 57

58 ( 1.1

2a

2b

2f

methyl

ethyl

butynyl

(methylthio)methyl

5.4 ( 1.4

12 ( 1.2

110 ( 14

14 ( 1.5

640 ( 150

330 ( 52

360 ( 39

210 ( 21

2k

7a

7b

methyl

ethyl

41 ( 5.1

116 ( 3.5

400 ( 56

220 ( 20

8a

8b

methyl

ethyl

160 ( 73

300 ( 160

210 ( 31

190 ( 10

a

Data represent the mean ((sem) from one to three experiments, each performed in duplicate.

B

A

C

D

F igu r e 6. (A) Stimulation of PI metabolism by carbachol, arecoline, 1a , 2a , and 4 in A9 L cells expressing m1 muscarinic receptors.

(B) Stimulation of PI metabolism by carbachol, arecoline, 1a , 2a , and 4 in A9 L cells expressing m3 muscarinic receptors. (C)

Stimulation of PI metabolism by carbachol, 1b, 1f, and 2b in A9 L cells expressing m1 muscarinic receptors. (D) Stimulation of

PI metabolism by carbachol, 1b, 1f, and 2b in A9 L cells expressing m3 muscarinic receptors. Data represent the mean stimulation

above basal levels from at least three assays for each concentration of ligand, each performed in triplicate. The curves were

generated by iterative fit to the equation: Y ) (S_maxx)/(x + EC₅₀) from the data shown in the figure.

respectively, within the 21 kJ /mol energy interval above

the global minimum. The double bond in the 1,2,5,6-

tetrahydropyridine system may limit the number of

stable conformations in the arecoline series of com-

pounds, while the 1,4,5,6-tetrahydropyrimidine system

permits more energy minima upon rotation around the

carbon-carbon bond between the tetrahydropyrimidine

ring and the ester or 1,2,4-oxadiazole moiety. In addi-

tion to conformational constraints, the N-methyl group

in arecoline and 4 may contribute to the relatively low

Selective m1 Agonists

J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 8 1237

Ta ble 7. Potency and Efficacy for a Series of Muscarinic

the amidine-like charge distribution in producing higher

muscarinic agonist activity can be addressed by syn-

thesizing derivatives combining these molecular fea-

tures.

Agonists at m3 Muscarinic Receptors Expressed in A9 L Cells^a

ligand

EC₅₀(µM)

S_max(%)

carbachol

aceclidine

xanomeline

9.9 ( 2.6

10 ( 3.7

11

350 ( 16

130 ( 8.6

50

The interaction of muscarinic agonists with the m1

receptor was examined through a series of docking

studies using the nine amino acid model of the m1

receptor binding site developed by Nordvall and Hack-

sell.⁵⁰Molecular modeling studies, in combination with

site-directed mutagenesis studies, have identified sev-

eral amino acids believed to be involved in ligand

binding to muscarinic receptors. In particular, an

aspartate residue (Asp105 for the m1 receptor) appears

critical for interaction of the onium head group of

acetylcholine and carbachol with muscarinic receptor

subtypes. In addition, site-directed mutagenesis stud-

ies^51,52and molecular modeling approaches^50,53have

implicated threonine, tyrosine, and asparagine residues

as potential hydrogen bond donors for the ester moiety

of muscarinic ligands. As shown in Figure 8 and Table

10, docking studies (using the program Sybyl)⁵⁴revealed

hydrogen bond interactions between simple esters (such

as arecoline and compound 1a ) and the Asp105 and the

Thr192 residues, yet not with the Asn382 side chain.

In contrast, both 1,2,4-oxadiazole derivatives (2a and

4) could interact via a three-point contact with the

aspartate, threonine, and asparagine residues.

1a

1b

1f

50 ( 9.0

17 ( 12

18 ( 6.6

220 ( 8.4

32 ( 5.3

100 ( 9.7

2a

2b

2h

2k

1.1 ( 0.24

10 ( 1.5

250 ( 8.3

56 ( 1.9

110 ( 5.5

98 ( 13

11 ( 2.4

5.8 ( 2.9

arecoline

4

12 ( 3.3

2.8 ( 1.7

96 ( 4.8

40 ( 4.0

7a

7b

14 ( 2.9

12 ( 9.1

140 ( 5.9

31 ( 5.2

a

Data represent the mean ((sem) from one to three experi-

ments, each performed in duplicate.

Ta ble 8. Inhibition of Forskolin-Induced cAMP Production in

A9 L Cells Expressing m2 Receptors^a

ligand

IC₅₀(µM)

% inhibn

carbachol

xanomeline

1a

1f

0.10 ( 0.06

7.0 ( 3.0

1.6 ( 1.7

0.78 ( 0.55

0.11 ( 0.010

1.2 ( 0.51

1.2 ( 0.59

0.44

24 ( 3.0

44 ( 5.0

13 ( 6.0

12 ( 3.0

21 ( 7.0

12 ( 3.0

16 ( 2.0

10

2a

arecoline

4

7b

An important finding from the present studies is that

isosteric replacement of an ester with the 1,2,4-oxadia-

zole group increases the affinity for m1 receptors,

without dramatically affecting coupling to second mes-

senger systems. The data suggest that the oxadiazole

moiety does not enhance efficacy at m1 receptors, but

may be useful in increasing agonist potency for m1

receptors. The utility of the 1,2,4-oxadiazole moiety in

enhancing ligand potency relative to simple esters could

thus be explained by an additional interaction of the

oxadiazole ring with the asparagine residue. This

concept of hydrogen bonding to Thr192 and Asn382 also

fits with the finding of low potencies and efficacies for

ether and oxime derivatives. The oxygen is in the wrong

position for hydrogen bonding in the ether derivatives,

and its charge is reduced in oximes by the nitrogen,

thereby reducing the ability to hydrogen bond to Thr192.

This hypothesis is particularly attractive since recent

site-directed mutagenesis studies indicate that the

asparagine moiety is not as important for the binding

and activity of small muscarinic agonists (e.g., acetyl-

choline, carbachol, and pilocarpine) to m3 muscarinic

receptors, yet is critical for the high-affinity interaction

of larger ligands such as N-methylscopolamine.⁵²Ad-

ditional studies of the binding of these ligands to mutant

receptors are necessary to verify the potential hydrogen-

bonding interaction of the oxadiazole moiety with the

asparagine residue of muscarinic receptor subtypes.

a

Data represent the mean ((sem) IC₅₀values and percentage

inhibition as determined from one to four separate experiments

for each ligand, each performed in duplicate.

Ta ble 9. Relative Energy Values and Molecular Volumes for

Different Conformations of 1a and Arecoline^a

ligand

1a i

1a ii

1a iii

1a iv

1a v

1a vi

rel energy (kJ /mol)

mol vol (Å³)

0

131.9

130.9

131.4

132.0

130.8

131.5

152.8

152.3

8.53

9.37

10.5

14.6

16.1

0

arecoline i

arecoline ii

arecoline iii

12.5

12.6

a

Data were obtained for the protonated species of each ligand.

efficacy of the 1,2,5,6-tetrahydropyridine derivatives.

The methyl group may prevent close interaction be-

tween the negative Asp105 and the protonated site of

the ligand (see below). The arecoline conformers had a

larger molecular volume than the conformers of 1a , due

primarily to the additional methyl group on arecoline

(see Table 9). The methyl group, however, may be

advantageous in the interaction with residues in the

binding pocket other than Asp105.⁴⁹

Color codes of partial atomic charges, calculated by

molecular orbital methods and taken at the van der

Waals surface, indicated similar patterns over the ester

as well as oxadiazole groups for both the tetrahydropy-

ridine and tetrahydropyrimidine systems as seen in

Figure 7C. The positive charge region was more ex-

tended for the 1,4,5,6-tetrahydropyrimidine ring system

as compared to the 1,2,5,6-tetrahydropyridine ring. The

charge distribution of the amidine system may confer

higher agonist activity, since both 1a and 2a display

higher efficacy than their 1,2,5,6-tetrahydropyridine

analogs. The relative merits of the conformationally

flexible system, the increase of molecular volume, and

Taken together, experimental data and molecular-

modeling results suggest that the tetrahydropyrimidine

moiety is an important determinant for agonist activity

at m1 receptors. Three structural factors may contrib-

ute to agonist efficacy at m1 muscarinic receptors:

conformational flexibility, molecular volume, and the

charge distribution of the amidine system. By the same

analysis, the 1,4,5,6-tetrahydropyrimidine system seems

to be a better starting point than the 1,2,5,6-tetrahy-

dropyridine system in the development of efficacious

ligands.

1238 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 8

Messer et al.

F igu r e 7. (A) Six minimum energy conformations of 1a as determined from molecular mechanics calculations using the MM2

force field and the multiconformer submode as implemented in the program MacroModel (v 3.0). (B) Three minimum energy

conformations of arecoline as determined using similar methods. (C) Color-coded van der Waal’s surfaces for the lowest energy

conformations of arecoline, 1a , 2a , and 4. Data were generated from semiempirical MNDO partial atomic charges. Charges

greater than +0.25 are indicated in blue, charges close to zero are in white, and negative charges less than -0.25 are in red.

Su m m a r y

coupled to PI metabolism, confirming previous observa-

tions in brain tissue.^29,30In addition, the data revealed

an important contribution of the 1,2,4-oxadiazole moiety

in enhancing potency (yet not efficacy). Preliminary

modeling studies suggested that the enhanced potency

could be due to an additional interaction of the oxadia-

zole ring with the asparagine residue believed to be

involved in ligand binding to muscarinic receptors.^50,53

Finally, compounds 1b, 1f, 2b, and 7b displayed func-

tional selectivity for m1 muscarinic receptors as indi-

The results of the present study provide new infor-

mation regarding the activity and selectivity of the

tetrahydropyrimidine derivatives. In addition, impor-

tant functional features were identified for agonist

potency, activity, and selectivity for muscarinic receptor

subtypes. Perhaps of greatest importance is that the

methyl (1a ) and propargyl (1f) esters and the methyl-

1,2,4-oxadiazole (2a ) tetrahydropyrimidine derivatives

displayed high efficacy at m1 muscarinic receptors

Selective m1 Agonists

J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 8 1239

F igu r e 8. Stereoviews of some binding modes obtained by docking of muscarinic ligands to a nine amino acid model of the m1

muscarinic receptor.⁵⁰The aspartate (top), threonine (left), and asparagine (right) residues are coded by atom type, as are the

individual ligands, with carbon as white; hydrogen, blue; oxygen, red; and nitrogen, blue. Potential hydrogen bond interactions

are shown by dashed yellow lines. (A) Docking of arecoline to the m1 receptor model revealing two hydrogen bond interactions.

(B) Docking of 1a to the m1 receptor model revealing two hydrogen bond interactions. (C) Docking of 4 to the m1 receptor model

revealing three hydrogen bond interactions. (D) Docking of 2a to the m1 receptor model revealing four hydrogen bond interactions,

including a hydrogen bond to the carbonyl oxygen of the polypeptide backbone.

sented uncorrected. Elemental analyses were performed by

M-H-W Laboratories, Phoenix, AZ, and by the Chemical

Department of the University of Toledo. Analytical results

indicated by elemental symbols were within (0.4% of the

theoretical values, except where noted for compounds 1e and

6c, for which carbon values were off by 0.5 and 0.8%,

respectively.

cated by full agonist activity at m1 receptors and lower

activity at m3 receptors. Compounds 1f and 7b also

displayed relatively low activity at m2 receptors. These

compounds warrant further evaluation as selective

muscarinic agonists for the treatment of Alzheimer’s

disease.

5-(2-Meth ylbu tyloxyca r bon yl)-1,4,5,6-tetr a h yd r op yr i-

m id in e Hyd r och lor id e (1d ). 1,4,5,6-Tetrahydropyrimidine-

5-carboxylic acid hydrochloride²⁹(0.5 g, 3 mmol) was dissolved

in anhydrous 3-methyl-1-butanol (50 mL), and thionyl chloride

(0.2 mL, 3 mmol) was added dropwise with stirring at room

temperature. The solution was refluxed over a period of 24 h

and then evaporated to dryness in vacuo. The resulting crude

residue was recrystallized from methanol-diethyl ether to give

0.4 g (89%) of the orange liquid of 5-(2-methylbutyloxycarbo-

nyl)-1,4,5,6-tetrahydropyrimidine (1d ) as the hydrochloride

salt: ¹H NMR (D₂O) δ 0.9 (d, 6H), 1.2 (q, 2H), 1.5 (m, 1H), 3.1

(m, 1H), 3.5 (d, 4H), 3.9 (t, 2H), 7.9 (s, 1H). Anal. (C₁₀H₁₉N₂O₂-

Cl) C, H, N.

Exp er im en ta l Section

Syn th etic Ch em istr y. Compounds were synthesized uti-

lizing reagent grade chemicals commercially available from

Aldrich Chemical Co. and Fisher Scientific without further

purification. Compounds 1a , 1b, 1f, 2a , 2b, 4, 5a , 7a , 7b, 8a ,

8b, and xanomeline were synthesized according to literature

procedures.^29-31,34,55Note that both 7a and 7b were synthe-

sized as racemates. Nuclear magnetic resonance spectra were

obtained on a Bruker ACF 300 MHz NMR in deuteriochloro-

form, deuteriomethanol, or deuterium oxide, using either TMS

or TSP as an internal standard. Mass spectra were performed

on a Hewlett-Packard 5890 spectrometer. TLC was performed

on Kodak Chromatogram sheet 13181 silica gel with

fluorescent indicator (F254). Melting points were taken on an

Electrothermal digital melting point apparatus and are pre-

a

(R,S)-5-(3-Meth ylbu tyloxyca r bon yl)-1,4,5,6-tetr a h yd r o-

p yr im id in e Hyd r och lor id e (1e). 1,4,5,6-Tetrahydropyri-

midine-5-carboxylic acid hydrochloride²⁹(0.5 g, 3 mmol) was

1240 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 8

Messer et al.

Ta ble 10. Relative Energies and Hydrogen Bonding

Interactions of Four Ligands with a Theoretical Model of the

m1 Muscarinic Receptor^a

recrystallized from ethanol-diethyl ether to give 0.4 g (90%)

of a viscous liquid of (R,S)-5-(2-methylbutyloxycarbonyl)-

1,4,5,6-tetrahydropyrimidine (1e) as the hydrochloride salt: ¹H

NMR (D₂O) δ 0.8 (t, 3H), 0.9 (s, 3H), 1.0 (m, 1H), 1.3 (m, 2H),

3.09 (m, 1H), 3.55 (d, 4H), 4.0 (m, 2H), 7.9 (s, 1H). Anal.

(C₁₀H₁₉N₂O₂Cl) H, N; C: calcd, 51.2; found, 50.6.

H-bond

docking

rel energy

ligand

1a i

orientation (kJ /mol) Asp105 Thr192 Asn382

0

90

17.3

34.4

x

5-(2-Bu t yn yloxyca r b on yl)-1,4,5,6-t et r a h yd r op yr im i-

d in e Hyd r och lor id e (1g). 1,4,5,6-Tetrahydropyrimidine-5-

carboxylic acid chloride³¹(0.4 g, 2.2 mmol) was esterified in

2-butyn-1-ol (20 mL), employing a method virtually identical

to that used for 1l. The crude residue obtained was recrystal-

lized from 2-propanol-diethyl ether to give off-white crystals

(0.3 g, 61%) of 5-(2-butynyloxycarbonyl)-1,4,5,6-tetrahydropy-

180

270

0

0.00

18.1

8.87

xx

x

1a ii

90

15.3

17.0

7.26

20.0

35.3

x

180

270

0

x

1

rimidine (1g) as the hydrochloride salt: mp 139-140 °C; H

1a iii

1a iv

1a v

NMR (D₂O) δ 1.5 (t, 3H), 3.1 (m, 1H), 3.5 (d, 4H), 3.9 (q, 2H),

90

7.9 (s, 1H). Anal. (C₉H₁₃N₂O₂Cl) C, H, N.

180

270

0

9.25

x

xx

x

20.2

23.2

31.5

41.9

27.0

16.3

6.54

32.9

14.1

8.69

3.12

20.4

9.05

0.00

3.09

5-(Bu t yn yloxyca r b on yl)-1,4,5,6-t e t r a h yd r op yr im i-

d in e Hyd r och lor id e (1h ). 1,4,5,6-Tetrahydropyrimidine-5-

carboxylic acid chloride³¹(2 g, 11 mmol) was esterified in

3-butyn-1-ol (10 mL), employing a method virtually identical

to that used for 11. Recrystallization of the residue from

2-propanol-diethyl ether yielded reddish-brown crystals (0.6

g, 80%) of 5-(butynyloxycarbonyl)-1,4,5,6-tetrahydropyrimidine

(1h ) as the hydrochloride salt (hygroscopic): ¹H NMR (D₂O) δ

2.1 (t, 1H), 2.4 (m, 2H), 3.1 (m, 1H), 3.5 (d, 4H), 4.1 (t, 2H),

7.8 (s, 1H). Anal. (C₉H₁₃N₂O₂Cl) C, H, N.

5-(3-P en t yn yloxyca r bon yl)-1,4,5,6-t et r a h yd r op yr im i-

d in e Hyd r och lor id e (1i). 1,4,5,6-Tetrahydropyrimidine-5-

carboxylic acid chloride³¹(0.4 g, 2.2 mmol) was esterified in

2-pentyn-1-ol (20 mL), employing a method virtually identical

to that used for 1l. Recrystallization of the residue from

2-propanol-diethyl ether yielded off-white crystals (0.25 g,

68%) of 5-(3-pentynyloxycarbonyl)-1,4,5,6-tetrahydropyrimi-

dine (1i) as the hydrochloride salt: mp 179-180 °C; ¹H NMR

(D₂O) 1.9-1.7 (t, 3H), 3.1 (m, 1H), 3.3-3.2 (q, 2H), 3.4 (d, 4H),

4.0 (d, 2H) 7.9 (s, 1H). Anal. (C₁₀H₁₅N₂O₂Cl) C, H, N.

90

180

270

0

x

90

180

270

0

x

xx

x

2a i

90

x

180

270

0

x

xx

2a ii

x

90

180

270

0

28.4

xx

x

17.1

17.8

30.2

28.4

4.2

2.64

6.76

25.8

7.76

13.6

27.4

22.4

30.6

6.70

0.00

5.30

2a iii

2a iv

arecoline i

arecoline ii

4 i

x

90

x

180

270

0

5-(2-P en t yn yloxyca r bon yl)-1,4,5,6-t et r a h yd r op yr im i-

din e (1j). 1,4,5,6-Tetrahydropyrimidine-5-carboxylic acid chlo-

ride³¹(0.4 g, 2.2 mmol) was esterified in 3-pentyn-1-ol (20 mL),

employing a method virtually identical to that used for 1l.

Recrystallization of the residue from 2-propanol-diethyl ether

yielded pale yellow crystals (0.25 g, 62%) of 5-(2-pentynyloxy-

carbonyl)-1,4,5,6-tetrahydropyrimidine (1i) as the hydrochlo-

90

180

270

0

xx

x

90

180

270

0

x

1

ride salt: mp 158-159 °C; H NMR (D₂O) δ 1.9 (s, 3H), 3.2

90

xx

(m, 1H), 3.4-3.2 (t, 2H), 3.5 (d, 4H), 4.0 (d, 2H), 7.8 (s, 1H).

180

270

0

Anal. (C₁₀H₁₅N₂O₂Cl) C, H, N.

41.6

x

5-(P e n t yn yloxyca r b on yl)-1,4,5,6-t e t r a h yd r op yr im i-

d in e Hyd r och lor id e (1k ). 1,4,5,6-Tetrahydropyrimidine-5-

carboxylic acid chloride³¹(0.4 g, 2.2 mmol) was esterified in

4-pentyn-1-ol (20 mL), employing a method virtually identical

to that used for 1l. Recrystallization of the residue from

2-propanol-diethyl ether yielded pale yellow crystals (0.25 g,

72%) of 5-(pentynyloxycarbonyl)-1,4,5,6-tetrahydropyrimidine

(1i) as the hydrochloride salt: mp 156-157 °C; ¹H NMR (D₂O)

δ 2.2 (t, 1H), 3.1 (m, 1H), 3.2 (m, 2H), 3.4 (m, 2H), 3.6 (d, 4H),

3.7 (d, 2H), 7.9 (s, 1H). Anal. (C₁₀H₁₅N₂O₂Cl) C, H, N.

tr a n s-5-(3-Meth yl-2-p en ten -4-yn yloxyca r bon yl)-1,4,5,6-

tetr a h yd r op yr im id in e Hyd r och lor id e (1l). 1,4,5,6-Tet-

rahydropyrimidine-5-carboxylic acid hydrochloride²⁹(1 g, 6

mmol) was suspended in a solution of oxalyl chloride (1.5 mL,

17 mmol) in benzene (10 mL), heated with stirring under reflux

for 2.5 h, and then evaporated to dryness in vacuo after cooling,

to give an orange-yellow residue. The last traces of oxalyl

chloride were removed by adding 10 mL of benzene to the

residue, and then the residue was evaporated to dryness in

vacuo to give a crude residue of 1,4,5,6-tetrahydropyrimidine-

5-carboxylic acid chloride (1.6 g).³¹A mixture of the acid

chloride (0.4 g, 2.2 mmol) and trans-3-methyl-2-penten-4-yn-

1-ol (10 mL, excess) was stirred at room temperature over-

night. The reaction mixture was taken up in water (100 mL),

stirred at room temperature for 2 h, and filtered. The dark-

brown residue obtained on removal of solvents in vacuo was

recrystallized (ethanol-ether) to give yellow crystals (0.3 g,

64%) of trans-5-(3-methyl-2-penten-4-ynyloxycarbonyl)-1,4,5,6-

tetrahydropyrimidine (1l) as the hydrochloride salt: mp 144-

145 °C; ¹H NMR (D₂O) δ 1.8 (m, 3H), 3.09 (m, 1H), 3.44 (s,

22.0

37.5

45.3

41.2

21.0

0.00

36.5

25.2

x

90

180

270

0

4 ii

x

xx

x

90

180

270

0

x

7b i

>100

90

51.8

26.6

0.0

40.5

61.2

39.9

23.1

20.4

x

180

270

0

7b ii

90

180

270

0

x

7b iii

7b iv

90

6.64

x

180

270

0

90

180

270

33.7

50.5

15.0

39.5

56.4

x

8.06

a

Data were obtained for the protonated species of each ligand.

Hydrogen bond interactions (x) were defined by H-O or H-N

distances less than 3.000 Å. Two hydrogen bonds to the same

residue are indicated as xx.

esterified in racemic 2-methyl-1-butanol (50 mL) by a method

virtually identical to that used for 1d . The crude product was

Selective m1 Agonists

J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 8 1241

1H), 3.50 (d, 4H), 4.18 (d, 2H), 5.9 (t, 1H), 7.80 (s, 1H). Anal.

(C₁₁H₁₅N₂O₂Cl) C, H, N.

residue obtained on removal of solvents in vacuo was recrys-

tallized from 2-propanol and diethyl ether to give off-white

crystals (0.3 g, 75%) of 5-(propenyloxycarbonyl)-1,4,5,6-tet-

rahydropyrimidine as the hydrochloride salt: mp 150-151 °C;

¹H NMR (D₂O) 3.1 (m, 1H), 3.5 (d, 4H), 3.8 (d, 2H), 5.1-5.2

(dd, 2H), 5.7-5.9 (m, 1H), 7.8 (s, 1H). Anal. (C₈H₁₃N₂O₂Cl)

C, H, N.

cis-5-(3-Met h yl-2-p en t en -4-yn yloxyca r b on yl)-1,4,5,6-

tetr a h yd r op yr im id in e Hyd r och lor id e (1m ). 1,4,5,6-Tet-

rahydropyrimidine-5-carboxylic acid chloride³¹(0.4 g, 2.2

mmol) was esterified in cis-3-methyl-2-penten-4-yn-1-ol (15

mL) by a method virtually identical to that employed to

synthesize 1l. The dark-brown crude residue was recrystal-

lized from ethanol-diethyl ether to give a brown viscous oil

(0.4 g, 67%) of cis-5-(3-methyl-2-penten-4-ynyloxycarbonyl)-

1,4,5,6-tetrahydropyrimidine (1m ) as the hydrochloride salt:

¹H NMR (D₂O) δ 1.8 (m, 3H), 3.09 (m, 1H), 3.46 (s, 1H), 3.50

(d, 4H), 4.15 (d, 2H), 5.8 (t, 1H), 7.90 (s, 1H). Anal.

(C₁₁H₁₅N₂O₂Cl) C, H, N.

1-(Tr ip h en ylm eth yl)-5-(3-(2-m eth ylbu tyl)-1,2,4-oxa d ia -

zol-5-yl)-1,4,5,6-tetr ah ydr opyr im idin e. 1-(Triphenylmethyl)-

5-(3-(2-methylbutyl)-1,2,4-oxadiazol-5-yl)-1,4,5,6-tetrahydro-

pyrimidine was synthesized according to literature procedures.³⁰

Sodium hydride (60% dispersion in mineral oil, 104 mg, 2.6

mmol) and 4-methylpentamidoxime³²(130 mg, 2.6 mmol) were

suspended in dry THF (20 mL) in an oven-dried round-bottom

flask under a nitrogen atmosphere at 0 °C. After 60 min of

stirring, the ice bath was removed and the gray suspension

refluxed for 60 min to give a white suspension. 1-(Triphenyl-

methyl)-5-(methyloxycarbonyl)-1,4,5,6-tetrahydropyrimi-

dine³⁰(1 g, 2.6 mmol) solution in dry THF (30 mL) was added

all at once, and reflux continued for 18 h. The solvents were

evaporated to dryness in vacuo, and the residue was taken

up in water (20 mL). The aqueous suspension was extracted

exhaustively with chloroform, and the organic fractions were

dried (MgSO₄). The residue obtained after removal of solvents

in vacuo was chromatographed (silica, chloroform-methanol,

9:1) to yield a yellow semisolid residue of the crude cyclized

product (1.2 g, 37%, R_f) 0.38).

5-[3-(2-Isop r op yl)-1,2,4-oxa d ia zol-5-yl]-1,4,5,6-tetr a h y-

d r op yr im id in e Tr iflu or oa ceta te (2c). 5-[3-(2-Isopropyl)-

1,2,4-oxadiazol-5-yl]-1,4,5,6-tetrahydropyrimidine trifluoroac-

etate (2c) was prepared by a method virtually identical to that

employed to synthesize 2e, with the exception that 2-methyl-

propamidoxime³²(266 mg, 2.6 mmol) was used. After forma-

tion of the desired trifluoroacetate salt, the residue was

recrystallized from ethanol-diethyl ether to yield 100 mg

(38%) of a viscous reddish oil: ¹H NMR (D₂O) δ 1.1 (d, 6H),

2.1 (m, 1H), 3.10 (m, 1H), 3.65 (d, 4H), 7.8 (s, 1H); MS m/ z

194 (M⁺of free base). Anal. (C₁₁H₁₅F₃N₄O₃) C, H, N.

5-[3-(2-Met h ylp r op yl)-1,2,4-oxa d ia zol-5-yl]-1,4,5,6-t et -

r a h yd r op yr im id in e Tr iflu or oa ceta te (2d ). 5-[3-(2-Meth-

ylpropyl)-1,2,4-oxadiazol-5-yl]-1,4,5,6-tetrahydropyrimidine tri-

fluoroacetate (2d ) was prepared by a method virtually identical

to that used for 2e, with the exception that 3-methylpenta-

midoxime³²(302 mg, 2.6 mmol) was used. After formation of

the desired trifluoroacetate salt, the residue was recrystallized

from ethanol-diethyl ether to yield 100 mg (39%) of a viscous

reddish oil: ¹H NMR (D₂O) δ 0.9 (d, 6H), 2.0 (m, 1H), 2.50 (d,

2H), 3.10 (m, 1H), 3.70 (d, 4H), 7.9 (s, 1H); MS m/ z 208 (M⁺

of free base). Anal. (C₁₂H₁₇F₃N₄O₃) C, H, N.

cis-5-(3-P en ten -1-yn yloxyca r bon yl)-1,4,5,6-tetr a h yd r o-

p yr im id in e Hyd r och lor id e (1n ). 1,4,5,6-Tetrahydropyri-

midine-5-carboxylic acid chloride³¹(0.5 g, 3 mmol) was ester-

ified in cis/ trans-2-penten-4-yn-1-ol (5 g, 60 mmol) by a method

virtually identical to that used to synthesize 1l. Recrystalli-

zation of the residue from methanol-diethyl ether yielded

beige crystals (0.3 g, 78%) of cis-5-(3-penten-1-ynyloxycarbo-

nyl)-1,4,5,6-tetrahydropyrimidine (1n ) as the hydrochloride

1

salt: mp 180-181 °C; H NMR (D₂O) δ 2.15 (s, 1H), 3.09 (m,

1H), 3.46 (s, 1H), 3.50 (d, 4H), 4.15 (d, 2H), 5.8 (t, 1H), 7.80 (s,

1H). Anal. (C₁₀H₁₃N₂O₂Cl) C, H, N.

5-(2-Meth ylbu ten yloxyca r bon yl)-1,4,5,6-tetr a h yd r op y-

r im id in e Hyd r och lor id e (1o). 1,4,5,6-Tetrahydropyrimi-

dine-5-carboxylic acid chloride³¹(0.5 g, 2.7 mmol) was esterified

in 3-methyl-2-buten-1-ol (20 mL) by a method virtually identi-

cal to that used to synthesize 1l. Recrystallization of the crude

residue from ethanol-diethyl ether yielded a brown viscous

oil (0.4 g, 59%) of 5-(2-methylbutenyloxycarbonyl)-1,4,5,6-

tetrahydropyrimidine (1o) as the hydrochloride salt: ¹H NMR

(D₂O) δ 0.9 (s, 6H), 1.5 (m, 1H), 3.1 (m, 1H), 3.5 (d, 4H), 3.9

(m, 2H), 7.8 (s, 1H). Anal. (C₁₀H₁₇N₂O₂Cl) C, H, N.

5-(3-Meth ylp r op yn yloxyca r bon yl)-1,4,5,6-tetr a h yd r o-

p yr im id in e Hyd r och lor id e (1p ). 1,4,5,6-Tetrahydropyri-

midine-5-carboxylic acid chloride³¹(0.3 g, 1.6 mmol) was

esterified in 3-methylbutyn-2-ol (15 mL), employing a method

virtually identical to that used for 1l. The residue was

recrystallized from ethanol-ether to give a red viscous oil (0.2

g, 54%) of 5-(3-methylpropynyloxycarbonyl)-1,4,5,6-tetrahy-

dropyrimidine (1p ) as the hydrochloride salt: ¹H NMR (D₂O)

δ 1.5 (s, 3H), 2.3 (d, 1H), 3.09 (m, 1H), 3.5 (d, 4H), 3.8 (d, 1H),

7.9 (s, 1H). Anal. (C₉H₁₃N₂O₂Cl) C, H, N.

5-(3,3-Dim eth ylp r op yn yloxyca r bon yl)-1,4,5,6-tetr a h y-

d r op yr im id in e Hyd r och lor id e (1q). 1,4,5,6-Tetrahydro-

pyrimidine-5-carboxylic acid chloride³¹(0.2 g, 1.1 mmol) was

esterified in 2-methyl-3-butyn-2-ol (20 mL) by a method

virtually identical to that used to synthesize 1l. Recrystalli-

zation of the residue from methanol-diethyl ether yielded a

red viscous oil (90 mg, 52%) of 5-(3,3-dimethylpropynyloxy-

carbonyl)-1,4,5,6-tetrahydropyrimidine (1q) as the hydrochlo-

ride salt: ¹H NMR (D₂O) δ 1.4 (s, 6H), 2.5 (m, 1H), 3.1 (m,

1H), 3.5 (d, 4H), 7.8 (s, 1H). Anal. (C₁₀H₁₅N₂O₂Cl) C, H, N.

5-(P h en ylp r op yn yloxyca r bon yl)-1,4,5,6-tetr a h yd r op y-

r im id in e Hyd r och lor id e (1r ). 1,4,5,6-Tetrahydropyrimi-

dine-5-carboxylic acid chloride³¹(1 g, 6 mmol) was esterified

in 3-phenyl-2-propyn-1-ol (10 mL) by a method virtually

identical to that used for 1l. The first recrystallization from

ethanol and THF gave beige crystals (50 mg), mp 158-159

°C. ¹H NMR taken in deuteriomethanol indicated this com-

pound was an impurity. The filtrate from the first recrystal-

lization was then evaporated to dryness in vacuo to give an

orange-yellow viscous oily residue. The residue was recrystal-

lized from methanol-diethyl ether and dry THF to give white

crystals (0.4 g, 62%) of 5-(phenylpropynyloxycarbonyl)-1,4,5,6-

tetrahydropyrimidine (1r ) as the hydrochloride salt: mp 117-

5-[3-(2-Met h ylb u t yl)-1,2,4-oxa d ia zol-5-yl]-1,4,5,6-t et -

r a h yd r op yr im id in e Tr iflu or oa ceta te (2e). 1-(Triphenyl-

methyl)-5-(3-(2-methylbutyl)-1,2,4-oxadiazol-5-yl)-1,4,5,6-tet-

rahydropyrimidine was dissolved in trifluoroacetic acid (5 mL)

with stirring, and the resulting yellow solution was warmed

gently overnight for 24 h. The resulting dark solution was

evaporated to dryness in vacuo, and the yellow residue was

triturated with hexane (6 × 10 mL). The hexane layers were

decanted, and the residue was recrystallized (ethanol-ether)

to yield a viscous oily product (0.1 g, 39%) of 5-[3-(2-methyl-

butyl)-1,2,4-oxadiazol-5-yl]-1,4,5,6-tetrahydropyrimidine (2e)

as the trifluoroacetate salt: ¹H NMR (D₂O) δ 0.7 (d, 6H), 1.45

(q, 2H), 2.0 (m, 1H), 2.30 (t, 2H), 3.20 (m, 1H), 3.60 (d, 4H),

8.0 (s, 1H); MS m/ z 338 (M⁺of free base). Anal. (C₁₃H₁₉F₃N₄O)

C, H, N.

3-Bu t yn yl-p -t olu en esu lfon a t e. 3-Butynyl-p-toluene-

sulfonate was prepared utilizing an extension of literature

56

procedures.³³

,

Tosyl chloride (191 g, 0.7 mol) was dissolved

in anhydrous ether (1 L) in a 2-L three-necked round-bottomed

flask equipped with a thermometer, mechanical stirrer, and

powder funnel. 3-Butyn-1-ol (35 g, 38 mL, 0.50 mol) was

added, and the mixture was cooled to between -5 and -10 °C

(bath with dry ice and acetone) with rapid stirring. Freshly

powdered KOH (250 g) was added with vigorous stirring. The

addition was initially at the rate of 5 g of KOH every 2 min,

so as to maintain the temperature between 0 and 5 °C.

Evolution of heat was considerable; hence efficient cooling was

required. After adding the first 50 g of KOH, over 20 min,

1

118 °C; H NMR (CD₃OD) δ 3.2 (m, 1H), 3.6 (d, 4H), 4.9 (s,

2H), 7.3-7.5 (m, 5H), 7.9 (s, 1H). Anal. (C₁₄H₁₅N₂O₂Cl) C,

H, N.

5-(P r op e n yloxyca r b on yl)-1,4,5,6-t e t r a h yd r op yr im i-

d in e Hyd r och lor id e (1s). A mixture of 1,4,5,6-tetrahydro-

pyrimidine-5-carboxylic acid chloride (0.4 g, 2.2 mmol) and

2-propen-1-ol (15 mL) was stirred at room temperature for 24

h. The reaction mixture was then taken up in water (50 mL),

stirred at room temperature for 1.5 h, and filtered. The

1242 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 8

Messer et al.

the remainder was added over an additional 10 min (to

neutralize excess tosyl chloride). The mixture was stirred for

an additional 1 h while maintaining the temperature between

0 and 5 °C. After 1 h of stirring, the mixture was poured into

1.5 L of ice water, and the solid remaining in the flask was

quickly hydrolyzed with the ice water and subsequently added

to the bulk of the solution. After vigorous shaking, the layers

were separated. The organic layer (ether layer) and two

ethereal extracts were combined and dried over anhydrous

magnesium sulfate overnight. The ether was evaporated to

dryness in vacuo while keeping the temperature of the water

bath below 80 °C. The desired 3-butynyl tosylate was obtained

as an orange-yellow liquid in an overall yield of over 90%: ¹H

NMR (CDCl₃) δ 7.8 (2H, d), 7.4 (2H, d), 4.10 (2H, t), 2.6 (2H,

m), 2.45 (3H, s), 2.0 (1H, t).

thyloxy)methyl)-1,2,4-oxadiazol-5-yl)-1,4,5,6-tetrahydropyri-

midine trifluoroacetate (2h ) was prepared by a method virtually

identical to that used for 2e, with the exception that (methy-

loxy)methamidoxime (271 mg, 2.6 mmol) was used. After

formation of the desired trifluoroacetate salt, the residue was

recrystallized from ethanol-diethyl ether to yield 370 mg

(50%) of a brown oil: ¹H NMR (CD₃OD) δ 3.40 (3H, s), 3.80

(1H, m), 3.85 (4H, d), 4.54 (2H, s), 8.10 (1H, s). Anal.

(C₁₀H₁₃F₃N₄O₄) C, H, N.

3-(Meth yloxy)p r op a m id oxim e. 3-(Methyloxy)propami-

doxime was synthesized in a manner virtually identical to that

used for (methyloxy)acetamidoxime, except that 3-(methyloxy)-

propionitrile (6.54 mL, 72 mmol) was used. The resulting

filtrate was evaporated to yield 4.62 g (54%) of a green oil:

¹H NMR (DMSO) δ 2.21 (2H, t), 3.21 (3H, s) 3.47 (2H, t), 5.55

(2H, s).

4-Cya n o-1-bu tyn e. 4-Cyano-1-butyne was prepared from

the corresponding 3-butynyl-p-toluenesulfonate, employing an

extension of literature procedures.^33,56Dry, powdered NaCN

(3.6 g, 73 mmol, excess), was dissolved with stirring in DMSO

(30 mL). 3-Butynyl-p-toluenesulfonate (5 g, 4.5 mmol) was

added slowly because of the exothermic nature of the reaction,

all in a three-necked 1-L round-bottomed flask equipped with

a condenser, thermometer, and magnetic stirrer. After all of

the additions, the mixture was gradually heated to 70 °C using

a temperature-controlled water bath. The reaction mixture

was occasionally cooled during the additions so as to maintain

the temperature within the desired range. After an additional

30 min of stirring at 70 °C, the mixture was cooled to room

temperature and then poured into 400 mL of saturated

aqueous ammonium chloride to give a dark-brown mixture.

The dark-brown mixture was extracted several times with

ether. The ether extracts were washed with aqueous am-

monium chloride (2 × 100 mL), and the extracts were dried

(K₂CO₃). The ether was removed in vacuo, and the remaining

liquid was distilled on a Widmer column to yield 4.1 g (89%)

of off-white liquid product: bp 60-61 °C/10 mmHg (lit.³³bp

60 °C/10 mmHg); ¹H NMR (CD₃OD) δ 2.4 (1H, t), 2.5 (2H, m),

2.6 (2H, t).

5-(3-(Bu t yn -1-yl)-1,2,4-oxa d ia zol-5-yl)-1,4,5,6-t et r a h y-

d r op yr im id in e Tr iflu or oa ceta te (2f). 5-(3-(Butyn-1-yl)-

1,2,4-oxadiazol-5-yl)-1,4,5,6-tetrahydropyrimidine trifluoroac-

etate (2f) was prepared by a method virtually identical to that

used for 2e, with the exception that 4-pentynamidoxime³²(292

mg, 2.6 mmol) was used. After formation of the desired

trifluoroacetate salt, the residue was recrystallized from

ethanol-diethyl ether to yield 120 mg (40%) of viscous reddish

oil: ¹H NMR (D₂O) δ 2.25 (t, 1H), 2.5 (q, 2H), 2.6 (t, 2H), 3.10

(m, 1H), 3.50 (d, 4H), 7.8 (s, 1H); MS m/ z 204 (M⁺of free base).

Anal. (C₁₂H₁₃F₃N₄O₃) C, H, N.

5-(3-((Meth yloxy)eth yl)-1,2,4-oxa d ia zol-5-yl)-1,4,5,6-tet-

r a h yd r op yr im id in e Tr iflu or oa ceta te (2i). 5-(3-((Methy-

loxy)ethyl)-1,2,4-oxadiazol-5-yl)-1,4,5,6-tetrahydropyrimi-

dine trifluoroacetate (2i) was prepared by a method virtually

identical to that used for 2e, with the exception that (methy-

loxy)propamidoxime (307 mg, 2.6 mmol) was used. After

formation of the desired trifluoroacetate salt, the residue was

recrystallized from methanol-diethyl ether to yield 212 mg

(43%) of a yellow oil: ¹H NMR (CD₃OD) δ 2.97 (2H, t, J ) 6),

3.30 (3H, s), 3.75 (2H, t, J ) 6), 3.80 (1H, m), 3.87 (4H, d),

8.11 (1H, s). Anal. (C₁₁H₁₄F₃N₄O₄) C, H, N.

(Eth yloxy)pr opam idoxim e. (Ethyloxy)propamidoxime was

synthesized in a manner virtually identical to that used for

3-(methyloxy)acetamidoxime, except that (ethyloxy)propioni-

trile (7.84 mL, 72 mmol) was used. The resulting filtrate was

evaporated to yield 4.62 g (54%) of a yellow oil: ¹H NMR

(DMSO) δ 1.03 (3H, t), 2.21 (2H, t) 3.35 (2H, q), 3.50 (2H, t),

6.01 (2H, s).

5-(3-((Eth yloxy)eth yl)-1,2,4-oxa d ia zol-5-yl)-1,4,5,6-tet-

r a h yd r op yr im id in e Tr iflu or oa ceta te (2j). 5-(3-((Ethyloxy-

)ethyl)-1,2,4-oxadiazol-5-yl)-1,4,5,6-tetrahydropyrimidine tri-

fluoroacetate (2j) was prepared by a method virtually identical

to that used for 2e, with the exception that (ethyloxy)-

propamidoxime (344 mg, 2.6 mmol) was used. After formation

of the desired trifluoroacetate salt, the residue was recrystal-

lized from methanol-diethyl ether to yield 200 mg (57%) of a

yellow oil: ¹H NMR (CD₃OD) δ 1.03 (3H, t), 2.87 (2H, t), 3.38

(2H, q), 3.74 (7H, m), 8.01 (1H, s). Anal. (C₁₂H₁₇F₃N₄O₄) C,

H, N.

(Met h ylt h io)m et h a m id oxim e. Hydroxylamine hydro-

chloride (5g, 72 mmol) was suspended in methanol (30 mL)

under nitrogen and cooled to 0 °C with stirring. Sodium

methoxide (3.89 g, 72 mmol) suspended in methanol was added

slowly with the temperature maintained at 0 °C. (Methylthio)-

acetonitrile (6.27 g, 72 mmol) in methanol (10 mL) was added

dropwise to the white suspension. The mixture was allowed

to warm to room temperature overnight with stirring. The

mixture was heated to 40 °C with stirring for 6 h, cooled to

room temperature, and vacuum-filtered, and the filtrate was

evaporated in vacuo. The residue was dissolved in ethanol

and filtered to remove traces of remaining sodium chloride.

The filtrate was evaporated to yield 7.36 g (85%) of a green

viscous oil: ¹H NMR (DMSO) δ 1.89 (3H, s), 2.89 (2H, s), 5.38

(2H, s), 8.24 (1H, b).

5-(3-(P en tyn -1-yl)-1,2,4-oxa d ia zol-5-yl)-1,4,5,6-tetr a h y-

d r op yr im id in e Tr iflu or oa ceta te (2g). 5-(3-(Pentyn-1-yl)-

1,2,4-oxadiazol-5-yl)-1,4,5,6-tetrahydropyrimidine trifluoroac-

etate (2g) was prepared by a method virtually identical to that

used for 2e, with the exception that 5-hexynamidoxime³²(328

mg, 2.6 mmol) was used. After formation of the desired

trifluoroacetate salt, the residue was recrystallized from

methanol-diethyl ether to yield 55 mg (18%) of yellow oil: ¹H

NMR (D₂O) δ 1.83 (2H, m), 2.18 (3H, m), 2.76 (2H, t), 3.75

(1H, m), 3.77 (4H, d), 8.02 (1H, s). Anal. (C₁₃H₁₄F₃N₄O₃) C,

H, N.

(Meth yloxy)eth a m id oxim e. Hydroxylamine hydrochlo-

ride (5g, 72 mmol) was suspended in methanol (60 mL in a

100 mL, three-neck, round-bottom flask fitted with a condenser

and dropping funnel) under nitrogen and cooled to 0 °C with

stirring. Sodium (1.66 g, 72 mmol) was added slowly with the

temperature maintained at 0 °C. (Methyloxy)acetonitrile (5.12

g, 72 mmol) in methanol (10 mL) was added dropwise to the

white suspension. The mixture was allowed to warm to room

temperature overnight with stirring. The mixture was heated

to 40 °C with stirring for 6 h, cooled to room temperature, and

vacuum-filtered, and the filtrate was evaporated in vacuo. The

semisolid residue was dissolved in ethanol and filtered to

remove traces of remaining sodium chloride. The filtrate was

evaporated to yield 3.85 g (51%) of a brown oil: ¹H NMR

(DMSO) δ 3.28 (3H, s), 3.84 (2H, s).

5-(3-((Meth ylth io)m eth yl)-1,2,4-oxa d ia zol-5-yl)-1,4,5,6-

tetr a h yd r op yr im id in e Tr iflu or oa ceta te (2k ). 5-(3-((Me-

thylthio)methyl)-1,2,4-oxadiazol-5-yl)-1,4,5,6-tetrahydropyri-

midine trifluoroacetate (2k) was prepared by a method virtually

identical to that used for 2e, with the exception that (meth-

ylthio)methamidoxime (480 mg, 3.99 mmol) was used. After

formation of the desired trifluoroacetate salt, the residue was

recrystallized from ethanol-diethyl ether to yield 381 mg

(51%) of a brown crystalline solid (mp 89-91 °C): ¹H NMR

(CD₃OD) δ 2.01 (3H, s), 3.70 (1H, m), 3.76 (4H, d), 4.77 (2H,

s), 8.10 (1H, s). Anal. (C₁₂H₁₃F₃N₄O₃) C, H, N.

1,4,5,6-Tetr a h yd r o-5-(p r op yn yloxy)p yr im id in e Hyd r o-

ch lor id e (5b). 5-Hydroxy-1,4,5,6-tetrahydropyrimidine³⁵(2.3

g, 23 mmol) was dissolved in absolute ethanol (50 mL).

Propargyl chloride (1.7 g, 1.6 mL, 23 mmol) was added slowly

via a syringe followed by potassium carbonate (3.1 g, 23 mmol),

and the mixture was heated with stirring under reflux. After

5-(3-((Meth yloxy)m eth yl)-1,2,4-oxa d ia zol-5-yl)-1,4,5,6-

tetr a h yd r op yr im id in e Tr iflu or oa ceta te (2h ). 5-(3-((Me-

Selective m1 Agonists

J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 8 1243

2.5 h, the mixture was filtered, and the filtrate was evaporated

to dryness in vacuo to given an orange-yellow residue. The

hydrochloride salt was obtained by addition of 1 M HCl in

ether to an ethanol solution of the residue, evaporation of

solvents to dryness in vacuo, and recrystallization (ethanol-

ether) to give yellow crystals (2 g, 79%) of 5-(propynyloxy)-

1,4,5,6-tetrahydropyrimidine as the hydrochloride salt

(hygroscopic): ¹H NMR (CD₃OD) δ 1.1 (1H, t), 3.1 (1H, m),

3.45 (4H, d), 4.2 (2H, d), 7.9 (1H, s). Anal. (C₇H₁₁N₂OCl) C,

H, N.

1,4,5,6-Tetr a h yd r op yr im id in -5-on e Hyd r och lor id e. To

a suspension of pyridinium chlorochromate (23 g) in CH₂Cl₂

(800 mL) was added 5-hydroxy-1,4,5,6-tetrahydropyrimidine

(9.8 g).³⁵The resulting mixture was stirred at room temper-

ature for 18 h, Celite was added, and the mixture was filtered.

The filtrate was washed with NaHCO₃solution (2 × 50 mL),

and the aqueous layers were extracted with CH₂Cl₂(3 × 100

mL). The combined organic layers were dried (Na₂SO₄) and

the solvents evaporated in vacuo to give a residue, which was

purified by flash chromatography (silica, 60% ether-hexane

or EtOAc-petroleum ether, 3:7) to give the ketone as the free

base. The hydrochloride salt was obtained by addition of 1 M

HCl in ether to an ethanol solution of the residue; evaporation

of solvents to dryness in vacuo gave a brown residue. The

residue was discolorized and purified with activated charcoal

and then recrystallized from ethanol-diethyl ether to give

white crystals (6.5 g, 57%) of 1,4,5,6-tetrahydropyrimidin-5-

one as the hydrochloride salt: ¹H NMR (CD₃OD) δ 3.5 (s, 4H),

7.9 (s, 1H).

Meth yl-1,4,5,6-tetr a h yd r op yr im id in -5-on e Oxim e Hy-

d r och lor id e (6a ). Methoxylamine hydrochloride (0.3 g, 3

mmol) was suspended in methanol (20 mL) under a nitrogen

atmosphere and cooled to 0 °C. Sodium methoxide, 95% (0.2

g, 3 mmol) suspension in methanol (20 mL), was added slowly

with stirring, and the mixture was allowed to stir for 11 h at

0 °C under anhydrous conditions to generate the free base.

1,4,5,6-Tetrahydropyrimidin-5-one (0.3 g, 3 mmol) solution in

methanol (20 mL) was added dropwise over a period of 1 h.

The reaction mixture was allowed to warm to room temper-

ature, and stirring continued for 18 h. The solvents were

evaporated in vacuo to give a brown residue. The hydrochlo-

ride salt was obtained by addition of 1 M HCl in ether to an

ethanol solution of the residue. Evaporation of solvents to

dryness in vacuo gave a brown residue. The residue was

discolorized and purified with activated charcoal and then

recrystallized from 2-propanol-diethyl ether to give gray

crystals (0.2 g, 51%) of methyl-1,4,5,6-tetrahydropyrimidin-5-

one oxime as the hydrochloride salt: mp 147-148 °C; ¹H NMR

(CD₃OD) δ 3.5 (s, 4H), 3.72 (s, 3H), 7.9 (s, 1H). Anal. (C₅H₁₀N₃-

OCl) C, H, N.

(at a bath temperature ∼50 °C). The remaining liquid was

taken up in 150 mL of an aqueous solution containing 30 g of

NH₄Cl and extracted (50 mL × 10) with THF. The extracts

were combined and concentrated in vacuo, and the remaining

liquid was vacuum distilled to give a yellow liquid (4.0 g, 58%)

of 3-phenyl-2-propyn-1-ol: bp 128-129 °C/10 mm; ¹H NMR

(CD₃OD) δ 4.30 (s, 2H), 7.2 (m, 3H), 7.3 (m, 2H).

N-(P h en ylp r op yn oxy)p h t h a lim id e. N-(Phenylpropyn-

oxy)phthalimide was synthesized from 3-phenyl-2-propyn-1-

ol using an extension of previously reported procedure.⁵⁶

3-Phenyl-2-propyn-1-ol (3.3 g, 25 mmol), N-hydroxyphthalim-

ide (4 g, 25 mmol), and triphenylphosphine (6 g, 25 mmol) were

dissolved in dry tetrahydrofuran (120 mL) and treated with

diethyl azodicarboxylate (DEAD, 4.8 g, 27 mmol). The reaction

mixture became dark red, and the color disappeared after a

few minutes. A slight exothermic effect was observed on

mixing of the reagents. The reaction vessel was glass-

stoppered and then left at room temperature for 24 h. The

residue obtained after evaporation of solvents to dryness in

vacuo was chromatographed (silica, benzene/ether, 9:1) to give

a yellow viscous residue (13 g, R_f) 0.77). Recrystallization

of the residue from ethanol gave an orange viscous liquid (12.7

g, 90%) of pure N-(phenylpropynoxy)phthalimide: ¹H NMR

(CDCl₃) δ 4.35 (s, 2H), 7.25 (m, 3H), 7.35 (m, 2H), 7.90 (s, 4H).

P h en ylp r op yn oxyla m in e. Phenylpropynoxylamine was

obtained through an hydrazinolysis reaction of N-(phenylpro-

pynoxy)phthalimide according to literature procedures.⁵⁶N-

(Phenylpropynoxy)phthalimide (12.6 g, 45 mmol) and hydra-

zine hydrate (1.5 g, 45 mmol) were dissolved in ethanol and

heated with stirring under reflux. After 1 h the reaction

mixture was cooled and poured into a 3% sodium carbonate

solution to dissolve the N-aminophthalimide. The resulting

hydroxylamine was extracted (100 mL × 3) with ether, and

the extracts were dried (MgSO₄). The dried extracts was

adjusted to pH 5 with anhydrous ethereal hydrogen chloride

(1 M ether-HCl). After evaporation of solvents to dryness in

vacuo, the orange-yellow residue obtained was recrystallized

from ethanol and ether. The supernatant ether layer was

decanted, evaporated to dryness in vacuo, and then dried in a

vacuum oven at ∼50 °C for 2 h to give an orange-yellow solid

(8.5 g, 82%) of phenylpropynoxylamine: ¹H NMR (CDCl₃) δ

4.110 (s, 2H), 5.30 (br s, 2H), 7.15 (m, 3H), 7.25 (m, 2H).

(P h en ylp r op yn yl)-1,4,5,6-t et r a h yd r op yr im id in -5-on e

Oxim e Hyd r och lor id e (6c). (Phenylpropynyl)-1,4,5,6-tet-

rahydropyrimidin-5-one oxime hydrochloride (6c) was pre-

pared by a method virtually identical to that used for 6a , with

the exception that phenylpropynoxylamine (0.9 g, 4.9 mmol)

and 1,4,5,6-tetrahydropyrimidin-5-one (0.5 g, 4.9 mmol) were

used. After formation of the hydrochloride salt, the residue

was recrystallized from 2-propanol-diethyl ether to yield 0.4

g (61%) of white crystals: mp 115-116 °C; ¹H NMR (CD₃OD)

δ 3.45 (s, 4H), 4.30 (s, 2H), 7.3 (m, 3H), 7.4 (m, 2H), 7.9 (s,

1H); MS m/ z 227 (M⁺of free base). Anal. (C₁₃H₁₄N₃OCl) H,

N; C: calcd 59.2; found 60.0.

Ma ter ia ls. A9 L cells stably transfected with the m1, m2,

or m3 muscarinic receptor genes were received as a generous

gift from Dr. Mark Brann (University of Vermont). Eagle’s

minimum essential medium (EMEM), Dulbecco’s modified

Eagle’s minimum essential medium (DMEM), and ^L-glutamine

were purchased from Whittaker Bioproducts (Walkersville,

MD). Fetal bovine serum, penicillin G, and streptomycin were

obtained from Sigma (St. Louis, MO). The 0.2 µm filterware

(Nalge, Rochester, NY), plastic scrapers, and tissue culture

flasks were from VWR Scientific.

[³H]-(R)-Quinuclidinyl benzilate ([³H]-(R)-QNB; 44.9 Ci/

mmol) was obtained from Du Pont (Wilmington, DE). [³H]-

myo-Inositol was obtained from New England Nuclear (Boston,

MA) with a specific activity of 12.3-20.0 Ci/mmol. The cAMP

assay kit (cAMP [¹²⁵I] assay system code RPA 509) was

purchased from Amersham. Carbamylcholine chloride (car-

bachol) was obtained from Aldrich (Milwaukee, WI). Arecoline

hydrobromide, atropine sulfate, oxotremorine, and 4-(2-hy-

droxyethyl)-1-piperazineethanesulfonic acid (HEPES) were

obtained from Sigma. Arecedaine propargyl ester was pur-

chased from Research Biochemicals Incorporated. MgCl₂was

obtained from Fisher Scientific. CytoScint was purchased from

ICN Biomedicals, Inc. (Irvine, CA).

Eth yl-1,4,5,6-tetr a h yd r op yr im id in -5-on e Oxim e Hy-

d r och lor id e (6b). Ethyl-1,4,5,6-tetrahydropyrimidin-5-one

oxime hydrochloride (6b) was prepared by a method virtually

identical to that employed to synthesize 6a , with the exception

that ethoxylamine hydrochloride (0.3 g, 3 mmol), sodium

methoxide, 95% (0.2 g, 3 mmol), and 1,4,5,6-tetrahydropyri-

midin-5-one (0.3 g, 3 mmol) were used. After formation of the

hydrochloride salt, the residue was recrystallized from 2-pro-

panol-diethyl ether to yield 0.4 g (54%) of white crystals: mp

1

160-162 °C; H NMR (CD₃OD) δ 1.2 (t, 3H), 3.5 (s, 4H), 4.0

(q, 2H), 7.9 (s, 1H). Anal. (C₆H₁₂N₃OCl) C, H, N.

3-P h en yl-2-p r op yn -1-ol. 3-Phenyl-2-propyn-1-ol was pre-

pared by utilizing an extension of a literature procedure.³³

Diethylamine, 98% (100 mL, dried over powdered KOH), Pd-

(PPh₃)₄catalyst (0.5 g, 0.4 mmol), and CuI (0.5 g, 2.6 mmol)

were placed in a 250 mL, three-necked, round-bottomed flask.

Propargyl alcohol (11 g, 12 mL, 0.2 mol) was added, after which

the temperature rose to 45 °C. Then a mixture of bromoben-

zene (47 g, 32 mL, 0.3 mol) in dry THF (30 mL) (previously

cooled to 5-10 °C) was added portionwise over 30 min, while

the temperature was maintained between 45 and 50 °C (by

controlled addition of bromobenzene solution and occasional

cooling or heating). After 10 min, salt formation began from

the solution. The mixture was gently refluxed for another 1

h, allowed to cool, and filtered to remove insoluble residues.

The residue (salt) was washed with dry THF, filtrates were

combined, and then the mixture was concentrated in vacuo

1244 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 8

Messer et al.

Recep tor Bin d in g in Ra t Br a in . Binding to muscarinic

receptors was performed essentially as previously described.⁵⁷

Binding was determined indirectly by the ability of compounds

to compete with 50 pM [³H]-(R)-quinuclidinyl benzilate in a

suspension of brain membranes. Nonspecific binding was

evaluated by the inclusion of 1000-fold excess atropine in a

separate set of samples. IC₅₀values were determined from

Hill plots of the inhibition data and are reported as means (

sem of three independent experiments performed in triplicate.

P h osp h oin ositid e Tu r n over Assa ys. Intact A9 L cells

expressing either m1 muscarinic receptors (from passages 20-

30) or m3 receptors (from passages 12-23) were incubated

with [

³H]-myo-inositol (0.5 µCi/well) in 24-well plates with a

plating density of 10⁵cells per well. After incubation at 37

°C for 48 h in 0.5 mL of DMEM with 95% air/5% CO₂, the

medium was aspirated and the cells were washed twice with

1 mL of EMEM containing 10 mM LiCl. The medium was

replaced by 0.5 mL of the same medium containing 10 mM

LiCl and incubated at 37 °C for 20 min. The reaction was

initiated by the addition of 25 µL of agonist at appropriate

concentrations (or medium for determination of basal levels)

and allowed to continue for an additional 30 min. For

antagonist experiments, atropine was added simultaneously

with the agonist.

At the end of the incubation period, the reaction was

terminated by aspiration of medium and addition of 0.5 mL

of 5% (v/v) ice-cold trichloroacetic acid (TCA). The wells were

rinsed with 0.5 mL of distilled H₂O, which was added to the

TCA extract. The TCA extract was applied to Dowex-formate

columns (Biorad AG1-X8 resin, formate form, 100-200 mesh).

The columns were washed three times with 3 mL of 5 mM

myo-inositol, and then total [³H]inositol phosphates were

eluted with 1 mL of 1.0 M ammonium formate/0.1 M formic

acid. Then, 0.5 mL of the eluate was counted in 10 mL of

CytoScint on a TM Analytic beta counter.

Assa ys of Cyclic AMP F or m a tion . Measurement of

cAMP levels was carried out according to the procedures

described by Wess et al.⁶⁵Intact A9 L cells expressing m2

muscarinic receptors (from passages 9-14) were incubated in

24-well plates (at a density of 10⁵cells per well) in DMEM

with 95% air/5% CO₂at 37 °C for 24 h. After incubation, the

growth medium was replaced with 250 µL of DMEM contain-

ing 1 mM IBMX, 20 mM HEPES, and the appropriate agonist

concentrations. The reaction was terminated after a 10-min

incubation at room temperature by aspiration of medium and

addition of 250 µL of an ice-cold solution containing 0.1 N HCl

and 1 mM CaCl₂. After 15 min, the solution was aspirated

into appropriately labeled tubes and diluted with 250 µL of

cAMP assay buffer. The cAMP levels were determined by

radioimmunoassay as previously described by Brooker et al.⁶⁶

using the nonacetylation protocol as outlined in the Amersham

cAMP assay kit. A standard solution of cAMP was prepared

by adding 2 mL of deionized water to a stock powder contain-

ing 64 pmol of cAMP to obtain a final concentration of 32 pmol/

mL in 0.05 M acetate buffer containing 0.01% thimerosal.

Serial dilutions of 500 µL/tube were made from the stock

solution using the diluted assay buffer (0.05 M acetate buffer,

pH 5.8 with 0.01% azide). Aliquots (100 µL) from each serial

dilution permitted seven standard levels of cAMP ranging from

25 to 1600 fmol.

Polypropylene tubes were labeled in duplicate for total

counts (TC), zero standard tubes (B_o), standards, and samples.

Assay buffer (100 µL) was added to the zero standard tubes

(B_o). Each standard and the unknown samples (100 µL) were

similarly added to the appropriately labeled tubes. [¹²⁵I]cAMP

(100 µL) was added into all tubes followed by 100 µL of

antiserum in all tubes except the TC. All samples were

vortexed thoroughly, covered with plastic film, and incubated

for 3 h at 2-8 °C. After the incubation, Amerlex-M second

antibody reagent was added (500 µL) into all tubes except the

TC, which was covered and put aside for counting. All the

samples were vortexed and incubated for 10 min at room

temperature. The antibody bound fraction was separated by

centrifugation for 10 min at 1500g, decanting the supernatant

and allowing the samples to drain for 5 min. The radioactivity

in each tube was then determined by counting for 2 min in a

gamma scintillation counter (TM Analytic). The results were

calculated from the average cpm values for each set of replicate

samples and the percent B_o/TC. The percent bound for each

standard and sample was also calculated. A standard curve

was generated by plotting the percent bound for each concen-

tration of cAMP, and the amount of cAMP (fmol/tube) for each

unknown sample was determined. The concentration of

(cAMP/mL or cAMP/10⁵cells) was derived using the appropri-

ate dilution factor and correcting for the sample volume

assayed.

P h osp h oin ositid e Meta bolism in Ra t Cor tica l Slices.

The methods were modified from those described by Brown et

al.⁵⁸as previously reported.^59,60The cerebral cortex was

dissected according to the method of Glowinski and Iversen.⁶¹

In these studies, [³H]inositol was purified prior to use by

passing over a Dowex AGI-X8 anion-exchange column to

remove charged degradation products of [³H]inositol. The

amount of [³H]inositol phosphates formed in the assay was

determined essentially according to Wreggett and Irvine⁶²

except that the separation of inositol phosphates was ac-

complished using an Amersham Super Separator Manifold.

Cell Cu ltu r e. A9 L cells were maintained as monolayer

cultures in 175 cm²polystyrene tissue culture flasks at 37 °C

in a humidified atmosphere of 95% air and 5% CO₂. Growth

media (DMEM supplemented with 10% fetal bovine serum,

50 units/mL penicillin G and 50 µg/mL streptomycin, and 4

mM ^L-glutamine) were changed 24 h after each subculture

(every third day). Media were sterilized by filtration through

0.2 µm filterware. Cell viability was determined by observing

the cells microscopically at a magnification of 100× and

through trypan blue exclusion. Living cells were used from

passages 20-31 of the A9 L cell line expressing m1 receptors,

passages 19-38 of the A9 L cell line expressing m2 receptors,

and passages 12-23 of the A9 L cell line expressing m3

receptors.

Mem br a n e P r ep a r a tion (for Bin d in g Assa ys). As de-

scribed previously by Lee and colleagues,⁶³cells reaching

confluence from two flasks were scraped into 20 mL of cold

medium (4 °C) and collected by centrifugation for 10 min at

1600g at 4 °C. Cells were washed by resuspension of the cell

pellet in 20 mL of ice-cold binding buffer (10 mM HEPES

containing 5 mM MgCl₂, pH 7.4) and collected by centrifuga-

tion for 10 min at 1600g at 4 °C. The pellet was resuspended

in 5 mL of cold buffer and homogenized for 30 s at setting 8 in

a Brinkman Polytron. The resulting suspension was diluted

into a volume of 20 mL and centrifuged for 10 min at 1600g

and 4 °C. The pellet, containing nuclei and large cellular

debris, was discarded, and membranes in the supernatant

were pelleted for 30 min at 20000g and 4 °C. The pellet was

suspended in 4 mL of cold buffer and homogenized again. The

final membrane preparation was used for radioligand binding

assays after its protein concentration was determined through

a modified Lowry assay.

[³H]-(R)-QNB Sa tu r a tion Exp er im en ts. In a fashion

similar to the procedure described by Do¨rje and colleagues,⁶⁴

freshly prepared cell membranes (10-30 µg protein/mL) were

incubated in triplicate with increasing concentrations of [³H]-

(R)-QNB in a final volume of 1 mL at room temperature for 2

h. At each concentration of [³H]-(R)-QNB, 5000-fold excess

atropine (or ^L-hyoscyamine) was used to determine nonspecific

binding. Membranes with bound [³H]-(R)-QNB were collected

by filtration through an M-24R Brandel cell harvester onto

Whatman GF/C filters. Filters were rinsed twice with 5 mL

of ice-cold buffer and transferred to 20 mL scintillation vials.

After drying overnight, vials were filled with 5 mL of CytoS-

cint, and the radioactivity of each vial was counted in a TM

Analytic â-counter for 2 min at 800 K.

In h ibition Assa ys for Agon ists. Assays were initiated

by addition of freshly prepared membranes (30 µg) into

triplicate test tubes containing increasing concentrations of

unlabeled ligands and one concentration of [³H]-(R)-QNB. The

concentration of [³H]-(R)-QNB was chosen based on the affinity

observed in the direct binding assays (30 pM for m1 receptors).

A 5000-fold excess concentration of atropine was used to assess

nonspecific binding. Incubations were carried out in 1 mL

total volume at room temperature for 2 h and were terminated

by filtration as described above.

Selective m1 Agonists

J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 8 1245

Da ta An a lysis. All data points were determined in trip-

licate except where noted and plotted as the mean of at least

three independent experiments. Data from [³H]-(R)-QNB

binding experiments were subjected to a Scatchard analysis

and B_maxand K_dvalues were determined. Data from each

inhibition assay were analyzed by nonlinear regression meth-

ods for best fit to single or multiple binding sites. An F test

was used to determine the model which best fit the experi-

mental data with R set at 0.05. The IC₅₀values and percent-

ages of high-, medium-, and/or low-affinity sites were deter-

mined for each ligand. IC₅₀values were converted to K_ivalues

using the Cheng-Prusoff equation.⁶⁷

Data from each PI assay were analyzed by iterative fit to

the hyperbolic equation Y ) (S_maxx)/(x + EC₅₀). Correlation

coefficients were generally >0.95. Values of EC₅₀and S_max

determined from separate experiments for each agonist were

then compared by one-factor analysis of variance, and indi-

vidual comparisons were made using a Tukey test.

Refer en ces

(1) Davies, P.; Maloney, A. J . F. Selective loss of central cholinergic

neurons in Alzheimer’s disease. Lancet 1976, No. 2, 1403.

(2) Whitehouse, P. J .; Martino, A. M.; Antuono, P. G.; Lowenstein,

P. R.; Coyle, J . T.; DeLong, M. R. Alzheimer’s disease and senile

dementia: loss of neurons in the basal forebrain. Science 1982,

215, 1237-1239.

(3) Coyle, J . T.; Price, D. L.; DeLong, M. R. Alzheimer’s disease: A

disorder of cortical cholinergic innervation. Science 1983, 219,

1184-1190.

(4) Soininen, H.; Kosunen, O.; Helisalmi, S.; Mannermaa, A.;

Palia¨rvi, L.; Talasniemi, S.; Ryyna¨nen, M.; Riekkinen, P., Sr. A

severe loss of choline acetyltransferase in the frontal cortex of

Alzheimer patients carrying apolipoprotein ꢀ4 allele. Neurosci.

Lett. 1995, 187, 79-82.

(5) Christie, J . E.; Shering, A.; Ferguson, J .; Glen, A. I. M.

Physostigmine and arecoline. Effects of intravenous infusions

in Alzheimer presenile dementia. Br. J . Psychiatry 1981, 138,

46-50.

(6) Summers, W. K.; Majovski, L. V.; Marsh, G. M.; Tachiki, K.;

Kling, A. Oral tetrahydroaminoacridine in long-term treatment

of senile dementia, Alzheimer type. N. Engl. J . Med. 1986, 315,

1241-1245.

(7) Hollander, E.; Mohs, R. C.; Davis, K. L. Cholinergic approaches

to the treatment of Alzheimer’s disease. Br. Med. Bull. 1986,

42, 97-100.

(8) Moos, W.; Davis, R. E.; Schwarz, R. D.; Gamzu, E. R. Cognition

activators. Med. Res. Rev. 1988, 8, 353-391.

(9) Wamsley, J . K.; Gehlert, D. R.; Roeske, W. R.; Yamamura, H. I.

Muscarinic antagonist binding site heterogeneity as evidenced

by autoradiography after direct labeling with [³H]-QNB and [³H]-

pirenzepine. Life Sci. 1984, 34, 1395-1402.

(10) Mash, D. C.; Potter, L. T. Autoradiographic localization of M1

and M2 muscarine receptors in the rat brain. Neuroscience

1986, 19, 551.

(11) Corte´s, R.; Palacios, J . M. Muscarinic cholinergic receptor

subtypes in the rat brain. I. Quantitative autoradiographic

studies. Brain Res. 1986, 362, 227-238.

(12) Messer, W. S., J r.; Ellerbrock, B. R.; Smith, D. A.; Hoss, W.

Regional differences in the binding of selective muscarinic

antagonists to rat brain: Comparison with minimum energy

conformations. J . Med. Chem. 1989, 32, 1164-1172.

(13) Buckley, N. J .; Bonner, T. I.; Brann, M. R. Localization of a

family of muscarinic receptor mRNAs in rat brain. J . Neurosci.

1988, 8, 4646-4652.

(14) Wall, S. J .; Yasuda, R. P.; Hory, F.; Flagg, S.; Martin, B. M.;

Ginns, E. I.; Wolfe, B. B. Production of antisera selective for m1

muscarinic receptors using fusion proteins: Distribution of m1

receptors in rat brain. Mol. Pharmacol. 1991, 39, 643-649.

(15) Caulfield, M. P.; Higgens, G. A.; Straughan, D. W. Central

administration of the muscarinic receptor subtype-selective

antagonist pirenzepine selectively impairs passive avoidance

learning in the mouse. J . Pharm. Pharmacol. 1983, 35, 131-

132.

Molecu la r Mod elin g a n d Com p u ta tion a l Ch em istr y.

Computational chemical calculations were performed on a VAX

6420 computer running Still’s program MacroModel (version

3.0)⁴⁷as outlined in previous studies.^12,60Conformational

minima were generated for the protonated species of arecoline

and 1a using the modified MM2 and AMBER force fields and

parameters as implemented in the program. Torsional angles

were explored in 30° increments over the full 360° circle and

the resulting structures then fully minimized to a final RMS

gradient of <0.005 kJ /Å by using the full-matrix Newton-

Raphson method. Structures 2a and 4 were modeled using

similar methods, except that, due to the absence of 1,2,4-

oxadiazole parameters in the MM2 force field, a carbon atom

was substituted for the ring oxygen to conduct a multicon-

former search as described above. The oxygen was then

reinserted for subsequent molecular orbital calculations as

outlined below.

MacroModel files generated from MM2 calculations were

used as input files for Dewar’s MNDO methodology⁶⁸imple-

mented in the program MOPAC.⁶⁹Refined geometries and

partial atomic charges were generated within MOPAC. Out-

put files then were viewed using the MacroModel graphics

subroutine.

Docking studies were initiated on a Silicon Graphics Indigo²

workstation using MacroModel (version 4.5),⁴⁸Sybyl (version

6.1),⁵⁴and MOPAC (version 6.0) software, employing the model

of the m1 muscarinic receptor devised by Nordvall and

Hacksell.⁵⁰The model of the m1 receptor site was modified

to show connectivity to a polypeptide backbone, and charges

were added using the semiempirical method AM1, which

describes hydrogen bond systems well.⁷⁰The charges on each

ligand were added in a similar fashion.

(16) Hagan, J . J .; J ansen, J . H. M.; Broekkamp, C. L. E. Blockade of

spatial learning by the M₁muscarinic antagonist pirenzepine.

Psychopharmacology 1987, 93, 470-476.

(17) Messer, W. S., J r.; Thomas, G. J .; Hoss, W. P. Selectivity of

pirenzepine in the central nervous system. II. Differential

effects of pirenzepine and scopolamine on performance of a

representational memory task. Brain Res. 1987, 407, 37-45.

(18) Messer, W. S., J r.; Bohnett, M.; Stibbe, J . Evidence for

a

Each conformation of the ligands was docked onto the

receptor site, i.e. was fitted into the m1 receptor model. The

use of the TRIPOS forcefield (as implemented in the program

Sybyl) was necessary in order to obtain the oxadiazole param-

eters. Throughout the process, internal energy changes for

the ligands and intermolecular steric and electrostatic energy

terms between the elements of the dimer were considered,

keeping the geometry of the site rigid. Each minimized ligand

conformation was rotated by 90° increments about the C5-C

bond to find different starting positions for docking studies.

The resulting structures were ranked in order of increasing

energy, and potential H-bonding interactions were examined

within Sybyl.

preferential involvement of M₁muscarinic receptors in repre-

sentational memory. Neurosci. Lett. 1990, 116, 184-189.

(19) Gil, D. W.; Wolfe, B. B. Pirenzepine distinguishes between

muscarinic receptor-mediated phosphoinositide breakdown and

inhibition of adenylate cyclase. J . Pharmacol. Exp. Ther. 1985,

232, 608-616.

(20) Fisher, S. K.; Figueiredo, J . C.; Bartus, R. T. Differential

stimulation of inositol phospholipid turnover in brain by analogs

of oxotremorine. J . Neurochem. 1984, 43, 1171-1179.

(21) Fisher, S. K.; Agranoff, B. W. Receptor activation and inositol

lipid hydrolysis in neural tissues. J . Neurochem. 1987, 48, 999-

1017.

(22) Mash, D. C.; Flynn, D. D.; Potter, L. T. Loss of M2 muscarine

receptors in the cerebral cortex in Alzheimer’s disease and

experimental cholinergic denervation. Science 1985, 228, 1115-

1117.

(23) Lapchak, P. A.; Araujo, D. M.; Quirion, R.; Collier, B. Binding

sites for 3H-AF-DX 116 and effect of AF-DX 116 on endogenous

acetylcholine release from rat brain slices. Brain Res. 1989, 496,

285-294.

(24) Hoss, W.; Messer, W. S., J r.; Monsma, F. J ., J r.; Miller, M. D.;

Ellerbrock, B. R.; Scranton, T.; Ghodsi-Hovsepian, S.; Price, M.

A.; Balan, S.; Mazloum, Z.; Bohnett, M. Biochemical and

behavioral evidence for muscarinic autoreceptors in the CNS.

Brain Res. 1990, 517, 195-201.

(25) Richards, M. H. Rat hippocampal muscarinic autoreceptors are

similar to the M2 (cardiac) subtype - Comparison with hippoc-

ampal M1, atrial M2 and ileal M3 receptors. Br. J . Pharmacol.

1990, 99, 753-761.

Ack n ow led gm en t. The work was supported by NIH

research grants NS 01493 and NS 31173, by a Research

Challenge grant from the Ohio Board of Regents, and

by the DeArce Memorial Endowment Fund. The au-

thors are grateful to Drs. Wayne P. Hoss and Graham

J . Durant for their contributions to the work outlined

here. Also, the authors thank Dr. Doug Gage of

Michigan State University and Dr. Sharad Magar for

obtaining mass spectral data.

1246 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 8

Messer et al.

(26) Hammer, R.; Giraldo, E.; Schiavi, G. B.; Monferini, E.; Ladinsky,

H. Binding profile of a cardioselective muscarine receptor

antagonist, AF-DX 116, to membranes of peripheral tissues and

brain in the rat. Life Sci. 1986, 38, 1653-1662.

(27) 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.

(28) Wall, S. J .; Yasuda, R. P.; Li, M.; Wolfe, B. B. Development of

an antiserum against m3 muscarinic receptors - Distribution

of m3-receptors in rat tissues and clonal cell lines. Mol.

Pharmacol. 1991, 40, 783-789.

(29) Messer, W. S., J r.; Dunbar, P. G.; Rho, T.; Periyasamy, S.; Ngur,

D.; Ellerbrock, B. R.; Bohnett, M.; Ryan, K.; Durant, G. J .; Hoss,

W. Synthesis, biochemical activity and behavioral effects of a

series of 1,4,5,6-tetrahydropyrimidines as novel ligands for M₁

receptors. Bioorg. Med. Chem. Lett. 1992, 2, 781-786.

(30) Dunbar, P. G.; Durant, G. J .; Fang, Z.; Rho, T.; Abuh, Y. F.; El-

Assadi, A. A.; Ngur, D.; Periyasamy, S.; Hoss, W.; Messer, W.

S., J r. Design, synthesis, and neurochemical evaluation of 5-(3-

alkyl-1,2,4-oxadiazol-5-yl)-1,4,5,6-tetrahydropyrimidines as M₁

receptor agonists. J . Med. Chem. 1993, 36, 842-847.

(31) Dunbar, P. G.; Durant, G. J .; Rho, T.; Ojo, B.; Huzl, J . J ., III;

Smith, D. A.; El-Assadi, A. A.; Sbeih, S.; Ngur, D. O.; Peri-

yasamy, S.; Hoss, W.; Messer, W. J ., J r. Design, synthesis, and

neurochemical evaluation of 2-amino-5-(alkoxycarbonyl)-3,4,5,6-

tetrahydropyridines and 2-amino-5-(alkoxycarbonyl)-1,4,5,6-tet-

rahydrohydropyrimidines as M₁muscarinic receptor agonists.

J . Med. Chem. 1994, 37, 2774-2782.

(32) Hirotsu, T.; Katoh, S.; Sugasaka, K.; Seno, M.; Itagaki, T.

Binding ability of acetamide oxime with proton, copper (II) and

dioxouranium (VI) in aqueous solutions. J . Chem. Soc., Dalton

Trans. 1986, 1609.

(33) Bandsma, L. In Acetylenic Chemistry, 2nd ed.; Elsevier Press:

New York, 1988; pp 217-280.

(34) Evans, R. F.; Shannon, J . S. Hydropyrimidies. 4. Catalytic

reduction of substituted pyrimidines. J . Chem. Soc. 1965, 1406-

1412.

(35) Spry, D. O.; Aaron, H. S. 5-Hydroxytetrahydropyrimidine and

its 2-methyl homolog. J . Org. Chem. 1966, 31, 3838-3839.

(36) Tecle, H.; Lauffer, D. J .; Mirzadegan, T.; Moos, W. H.; Moreland,

D. W.; Pavia, M. R.; Schwarz, R. D.; Davis, R. E. A rationale for

the design and synthesis of m1 selective muscarinic agonists.

Bioorg. Med. Chem. Lett. 1992, 2, 821-826.

(37) 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-azabicyclo[2.2.1]-3-one oximes as muscarinic

receptor subtype selective agonists. Life Sci. 1993, 52, 505-511.

(38) Brann, M. R.; Buckley, N. J .; J ones, P.; Bonner, T. I. Expression

of a cloned muscarinic receptor in A9 L cells. Mol. Pharmacol.

1987, 32, 450-455.

(39) Potter, L. T.; Ferrendelli, C. A. Affinities of different cholinergic

agonists for the high and low affinity states of the hippocampal

M1 muscarine receptors. J . Pharmacol. Exp. Ther. 1989, 248,

974-978.

(40) Bujo, H.; Nakai, J .; Kubo, T.; Fukuda, K.; Akiba, I.; Maeda, A.;

Mishina, M.; Numa, S. Different sensitivities to agonist of

muscarinic acetylcholine receptor subtypes. FEBS Lett. 1988,

240, 95-100.

(41) Schwarz, R. D.; Davis, R. E.; J aen, J . C.; Spencer, C. J .; Tecle,

H.; Thomas, A. J . Characterization of muscarinic agonists in

recombinant cell lines. Life Sci. 1993, 52, 465-472.

(42) Fisher, S. K.; Klinger, P. D.; Agranoff, B. W. Muscarinic agonist

binding and phospholipid turnover in brain. J . Biol. Chem.

1983, 258, 7358-7363.

(43) Fisher, S. K.; Snider, R. M. Differential receptor occupancy

requirements for muscarinic cholinergic stimulation of inositol

lipid hydrolysis in brain and in neuroblastomas. Mol. Pharma-

col. 1987, 32, 81-90.

(44) Periyasamy, S.; Messer, W. S., J r.; Rocknich, S.; Hoss, W. Alkoxy-

1,2,5-thiadiazole derivatives of arecoline stimulate M1 musca-

rinic receptors coupled to phosphoinositide turnover. Brain Res.

1995, 693, 118-123.

(45) Novotny, E. A.; Brann, M. R. In Fourth International Symposium

on Subtypes of Muscarinic Receptors; Elsevier Trends J ournals:

Wiesbaden, FRG, 1989; p 116.

(46) Olianas, M. C.; Onali, P.; Neff, N. H.; Costa, E. C. Adenylate

cyclase activity of synaptic membranes from rat striatum.

Inhibition by muscarinic receptors. Mol. Pharmacol. 1983, 23,

393-398.

(47) Lipton, M.; Still, W. C. The multiple minimum problem in

molecular modeling. Tree searching internal coordinate con-

formational space. J . Comput. Chem. 1988, 9, 343-355.

(48) Mohamadi, F.; Richards, N. G. J .; Guida, W. C.; Liskamp, R.;

Lipton, M.; Caufield, C.; Chang, G.; Hendrickson, T.; Still, W.

C. Macromodel-An integrated software system for modeling

organic and bioorganic molecules using molecular mechanics.

J . Comput. Chem. 1990, 11, 440.

(49) Ojo, B.; Dunbar, P. G.; Durant, G. J .; Nagy, P. I.; Huzl, J . J .,

III; Periyasamy, S.; Ngur, D. O.; El-Assadi, A. A.; Hoss, W. P.;

Messer, W. S., J r. Synthesis and biochemical activity of novel

amidine derivatives as m1 muscarinic receptor agonists. Bioorg.

Med. Chem. 1996, 4, 1605-1615.

(50) Nordvall, G.; Hacksell, U. Binding-site modeling of the musca-

rinic m1 receptor - a combination of homology-based and

indirect approaches. J . Med. Chem. 1993, 36, 967-976.

(51) Wess, J .; Maggio, R.; Palmer, J . R.; Vogel, Z. Role of conserved

threonine and tyrosine residues in acetylcholine binding and

muscarinic receptor activation. A study with m3 muscarinic

receptor point mutants. J . Biol. Chem. 1992, 267, 19313-19319.

(52) Bluml, K.; Mutschler, E.; Wess, J . Functional role in ligand

binding and receptor activation of an asparagine residue present

in the sixth transmembrane domain of all muscarinic receptors.

J . Biol. Chem. 1994, 269, 18870-18876.

(53) Ward, J . S.; Merritt, L.; Klimkowski, V. J .; Lamb, M. L.; Mitch,

C. H.; Bymaster, F. P.; Sawyer, B. D.; Shannon, H. E.; Olesen,

P. H.; Honore´, T.; Sheardown, M. J .; Sauerberg, P. Novel

functional M₁selective muscarinic agonists. II. Synthesis and

structure-activity relationships of 3-(pyrazinyl)-1,2,5,6-tetrahy-

dro-1-methylpyridines. Construction of a molecular model for

the M₁pharmacophore. J . Med. Chem. 1992, 35, 4011-4019.

(54) TRIPOS, Inc., St. Louis, MO, 1994.

(55) Sauerberg, P.; Olesen, P. H.; Nielsen, S.; Treppedahl, S.;

Sheardown, M. J .; Honore´, T.; Mitch, C. H.; Ward, J . S.; Pike,

A. J .; Bymaster, F. P.; Sawyer, B. D.; Shannon, H. E. Novel

functionally selective M₁muscarinic agonists. Synthesis and

structure-activity relationships of 3-(1,2,5-thiadiazolyl)-1,2,5,6-

tetrahydro-1-methylpyridines. J . Med. Chem. 1992, 35, 2274-

2283.

(56) Grochowski; J urczac. New synthesis of O-alkylhydroxylamines.

Synthesis 1976, 10, 682-684.

(57) Farrar, J . R.; Hoss, W.; Herndon, R. M.; Kuzmiak, M. Charac-

terization of muscarinic cholinergic receptors in the brains of

copper deficient rats. J . Neurosci. 1985, 5, 1083-1089.

(58) Brown, E.; Kendall, D. A.; Nahorski, S. R. Inositol phospholipid

hydrolysis in rat cerebral cortical slices: I. Receptor charac-

terisation. J . Neurochem. 1984, 42, 1379-1387.

(59) Periyasamy, S.; Hoss, W. Kappa opioid receptors stimulate

phosphoinositide turnover in rat brain. Life Sci. 1990, 47, 219-

225.

(60) Hoss, W.; Woodruff, J . M.; Ellerbrock, B. R.; Periyasamy, S.;

Ghodsi-Hovsepian, S.; Stibbe, J .; Bohnett, M.; Messer, W. S.,

J r. Biochemical and behavioral responses of pilocarpine at

muscarinic receptor subtypes in the CNS. Comparison with

receptor binding and low-energy conformations. Brain Res.

1990, 533, 232-238.

(61) Glowinski, J .; Iversen, L. L. Regional studies of catecholamines

in rat brain I: the disposition of [³H]norepinephrine, [³H]-

dopamine and [³H]dopa in various regions of the brain. J .

Neurochem. 1966, 13, 655-669.

(62) Wreggett, K. A.; Irvine, R. F. A rapid separation method for

inositol phosphates and their isomers. Biochem. J . 1987, 245,

655-660.

(63) Lee, N. H.; El-Fakahany, E. E. Allosteric interactions at the m1,

m2 and m3 muscarinic receptor subtypes. J . Pharmacol. Exp.

Ther. 1991, 256, 468-470.

(64) Do¨rje, F.; Wess, J .; Lambrecht, G.; Tacke, R.; Mutschler, E.;

Brann, M. R. Antagonist binding profiles of five cloned human

muscarinic receptor subtypes. J . Pharmacol. Exp. Ther. 1991,

256, 727-733.

(65) Wess, J .; Bonner, T. I.; Do¨rje, F.; Brann, M. R. Delineation of

muscarinic receptor domains conferring selectivity of coupling

to guanine nucleotide-binding proteins and second messengers.

Mol. Pharmacol. 1990, 38, 517-523.

(66) Brooker, G.; Harper, J . F.; Teraski, W. L.; Moylan, R. D.

Radioimmunoassay of cyclic AMP and cyclic GMP. Adv. Cycl.

Nucleotide Res. 1979, 10, 1-33.

(67) Cheng, Y.-C.; Prusoff, W. H. Relationship between the inhibition

constant (K_i) and the concentration of inhibitor which causes

50 per cent inhibition (I₅₀) of and enzymatic reaction. Biochem.

Pharmacol. 1973, 22, 3099-3108.

(68) Dewar, M. J . S.; Thiel, W. Ground states of molecules. The

MNDO method. Approximations and parameters. J . Am. Chem.

Soc. 1977, 99, 4899-4907.

(69) Stuart, J . J . P. QCPE no. 455, Bloomington, IN.

(70) Dewar, M. J . S.; Zoebisch, E. G.; Healy, E. F.; Stuart, J . J . P.

AM1: A new general purpose quantum mechanical molecular

model. J . Am. Chem. Soc. 1985, 107, 3902-3909.

J M960467D

Article Doi

Synthesis and biological characterization of 1,4,5,6- tetrahydropyrimidine and 2-amino-3,4,5,6-tetrahydropyridine derivatives as selective m1 agonists

DOI: 10.1021/jm960467d

Source and publish data:

Authors:

Article abstract of DOI:10.1021/jm960467d

Full text of DOI:10.1021/jm960467d

Products guided by the article

R&D Labs maybe for 936112-78-6

Relevant to this article

Hot Product