Synthesis and muscarinic receptor pharmacology of a series of 4,5,6,7-tetrahydroisothiazolo[4,5-c]pyridine bioisosteres of arecoline

Article abstract of DOI:10.1016/S0968-0896(99)00033-4

A series of O- and ring-alkylated derivatives of 4,5,6,7-tetrahydroisothiazolo[4,5-c]pyridin-3-ol was synthesized via treatment of appropriately substituted 4-benzylamino-1,2,5,6-tetrahydropyridine-3-carboxamides with hydrogen sulfide and subsequent ring closure by oxidation with bromine. The muscarinic receptor affinity as well as estimated relative efficacy and subtype selectivity of this series of bicyclic arecoline bioisosteres were determined using rat brain membranes and a number of tritiated muscarinic receptor ligands. The effects at the five cloned human muscarinic receptor subtypes of a selected series of chiral analogues, with established absolute stereochemistry, were studied using receptor selection and amplification technology (R-SAT((TM))). The potency, relative efficacy, and receptor subtype selectivity of these compounds were related to the structure of the O-substituents and the position and stereochemical orientation of the piperidine ring methyl substituents. Copyright (C) 1999 Elsevier Science Ltd.

Full text of DOI:10.1016/S0968-0896(99)00033-4

Bioorganic & Medicinal Chemistry 7 (1999) 795±809
Synthesis and Muscarinic Receptor Pharmacology of a Series of
4,5,6,7-Tetrahydroisothiazolo[4,5-c]pyridine Bioisosteres of
Arecoline
Henrik Pedersen,a Hans Brauner-Osborne,b Richard G. Ball,c Karla Frydenvang,b
Eddi Meier,a Klaus P. Bùgesùa and Povl Krogsgaard-Larsenb,*
aMedicinal Chemistry Research, H. Lundbeck A/S, 9 Ottiliavej, DK-2500 Valby-Copenhagen, Denmark
bDepartment of Medicinal Chemistry, The Royal Danish School of Pharmacy, 2 Universitetsparken, DK-2100 Copenhagen, Denmark
cMerck Research Laboratories, Division of Merck & Co., Inc., Rahway, NJ 07065-0900, USA
Received 17 August 1998; accepted 19 October 1998
AbstractÐA series of O- and ring-alkylated derivatives of 4,5,6,7-tetrahydroisothiazolo[4,5-c]pyridin-3-ol was synthesized via
treatment of appropriately substituted 4-benzylamino-1,2,5,6-tetrahydropyridine-3-carboxamides with hydrogen sul®de and sub-
sequent ring closure by oxidation with bromine. The muscarinic receptor anity as well as estimated relative ecacy and subtype
selectivity of this series of bicyclic arecoline bioisosteres were determined using rat brain membranes and a number of tritiated
muscarinic receptor ligands. The e€ects at the ®ve cloned human muscarinic receptor subtypes of a selected series of chiral analo-
gues, with established absolute stereochemistry, were studied using receptor selection and ampli®cation technology (R-SAT2). The
potency, relative ecacy, and receptor subtype selectivity of these compounds were related to the structure of the O-substituents
and the position and stereochemical orientation of the piperidine ring methyl substituents. # 1999 Elsevier Science Ltd. All rights
reserved.
Introduction
densities of postsynaptic M1 sites,8 an observation simi-
lar to ®ndings in AD patients.8,9 It still is uncertain
which subtypes of the cloned mAChRs parallel the M1
and M2 receptor sites, but it has been shown by immu-
noprecipitation that the forebrain and the cerebral cor-
tex as well as the hippocampus possess a majority of the
m1 mAChR subtype.10±12
Accumulating evidence supports the `cholinergic
hypothesis',1 postulating that the documented de®cit in
central cholinergic transmission causes the learning and
memory impairments seen in patients with Alzheimer's
disease (AD) and senile dementia of the Alzheimer type
(SDAT).2±4 It has been shown that basal forebrain
acetylcholine neurons degenerate in AD patients5 and
that the memory dysfunction can be mimicked to some
extent by muscarinic acetylcholine receptor (mAChR)
antagonists1,5 or by destruction of the nucleus basalis of
Meynert (part of basal forebrain).6 In AD patients, loss
of the presynaptic marker enzyme ChAT in the cortex
and hippocampus correlates with the severity of the
disease,5,7 and this ®nding can be mimicked by lesions in
the basal forebrain.8
The cholinergic communication between the basal fore-
brain and the cortex/hippocampus decays as the disease
progresses, and this leads to the growing memory and
cognitive problems. In light of these observations, there
is a therapeutic interest in agonists acting at the post-
synaptic M1 receptor,13,14 antagonists for the M2 auto-
receptor,8 or, ideally, a drug combining these two
pharmacological e€ects.15 Partial agonists have been
shown to have less predisposition than full agonists to
cause receptor desensitization,16±18 making partial ago-
nist at the M1 receptor more interesting from a ther-
apeutic point of view.
In the mapping of pathophysiological changes in the
brains of AD or SDAT patients, alterations of the den-
sities of pharmacologically characterized (M1 and M2)
and cloned (m1±m5) mAChRs have been extensively
studied. Thus, basal forebrain lesions in experimental
animals lead to loss of M2 receptor sites but unaltered
In order to de®ne mAChR agonist pharmacophore(s)
relevant to AD/SDAT and to design mAChR ligands of
therapeutic interest in these diseases, we have previously
described the synthesis and pharmacology of annulated
bicyclic muscarinic agonists and partial agonists bioiso-
sterically derived from arecoline (1) as exempli®ed by
*Corresponding author. Tel.: +45-35-37-67-77-511; fax: +45-35-37-
22-09; e-mail: nordly@medchem.dfh.dk
0968-0896/99/$ - see front matter # 1999 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(99)00033-4

796
H. Pedersen et al. / Bioorg. Med. Chem. 7 (1999) 795±809
3-methoxy-5-methyl-4,5,6,7-tetrahydroisoxazolo[4,5-c]-
pyridinium (2)19±21 and 3-methoxy-5-methyl-6,7-di-
hydro-4H-thiopyrano[3,4-d]isoxazol-5-ium (3)22 halides
(Fig. 1).
previously.24 The syntheses of the 6- and 7-methylated
analogues of 12, compounds 23 (Scheme 2) and 30
(Scheme 3), respectively, were based on the correspond-
ing b-oxo amides 21 and 28, respectively, as key inter-
mediates. In Scheme 2, the syntheses of the target
molecules, 8c,d and 9c, via 21 and 23 are outlined.
Methyl crotonate (14) was re¯uxed with benzylamine in
methanol to give the amino ester 15 in nearly quantita-
tive yield. The benzyl group was removed by treatment
with ammonium formate in methanol, in the presence of
palladium on activated carbon, in 96% yield. Treatment
of the resulting aminoester 16 with acrylonitrile gave 17
in 95% yield, which was regioselectively cyclized with
potassium tert-butoxide in toluene. The product, b-oxo
nitrile 18, was isolated as the ethyl carbamate in 35%
yield. Treatment of 18 with sulphuric acid did not a€ord
the desired amide 21, and a di€erent approach had to be
used. Compound 18 was converted into the 1,3-dioxane
19 in nearly quantitative yield. After oxidative hydro-
lysis of 19 with basic hydrogen peroxide and subsequent
acid hydrolysis of the 1,3-dioxane ring, 21 was obtained
in 39% yield. The keto amide 21 was converted into the
corresponding benzyl enamine 22 by treatment with
benzylamine in re¯uxing xylene in a yield of 76%. The
3-isothiazolol unit was then established by treating
enamine 22 with excess hydrogen sul®de in dimethyl-
formamide (DMF) and subsequent treatment of the
crude product with bromine in ethyl acetate to give 23
in 31% yield, following a previously described proce-
dure.24 In the ®rst step of the conversion of 23 into the
®nal products, 8c,d and 9c, a tert-butyloxycarbonyl
(Boc) group was substituted for the ethoxycarbonyl
group of 23 to give 24 (Scheme 2). Thus, after removal
of the latter group, using hydrogen bromide in glacial
acetic acid, the resulting amine salt was treated with
potassium carbonate and (Boc)2O in a water:tetra-
hydrofuran (THF) mixture. Under these conditions, the
3-hydroxy group also became acylated, but the 3-car-
bonate group formed was selectively cleaved in metha-
nol containing catalytic amounts of potassium
carbonate.25 The overall yield of these de- and repro-
tection reactions was 33%. The O-alkylations of 24
Structure±activity studies on a series of 3-alkoxy-
5,6,7,8-tetrahydro-4H-isoxazolo[4,5-c]azepines, includ-
ing 4±6 (Fig. 1), emphasized the determinative in¯uence
of the structure of the 3-alkoxy group on the pharma-
cology of these compounds.23 In order to further eluci-
date the importance of the structure of the `bottom part'
of this class of conformationally restricted mAChR
ligands, we have now synthesized and pharmacologi-
cally characterized a series of 3-alkoxy-4,5,6,7-tetra-
hydroisothiazolo[4,5-c]pyridines (7±11) (Fig. 1).
Synthesis
Methyl 3-hydroxy-4,5,6,7-tetrahydroisothiazolo[4,5-c]-
pyridine-5-carboxylate 12, which is the precursor for
compounds 7a±c (Scheme 1) was prepared as described
Scheme 1. Reagents and conditions: (i) HBr/AcOH (33%); (ii) (Boc)2O,
K2CO3, K2CO3 (1%)/CH3OH; (iii) CH2N2 (a), (C2H5)2SO4,
.
(C4H9)4N HSO4, NaOH (b); (iv) KOH/CH3OH; (v) CHꢀCCH2Br,
Figure 1. Structures of a number of known (1±6) and new (7±11)
agonist, partial agonist, or antagonist ligands at central mAChRs.
.
K2CO3, (C4H9)4N HSO4; (vi) HCl, ether, NaOH.

H. Pedersen et al. / Bioorg. Med. Chem. 7 (1999) 795±809
797
.
Scheme 2. Reagents and conditions: (i) C6H5CH2NH2, C2H5OH; (ii) NH4 HCOO, Pd-C (5%); (iii) CH2=CHCN, CH3OH; (iv) KOC(CH3)3,
toluene; (v) ClCOOC2H5, K2CO3; (vi) HOCH2CH2CH2OH, TsOH, toluene, re¯ux; (vii) H2O2, NaOH, C2H5OH; (viii) HCl, H2O, THF; (ix)
.
C6H5CH2NH2, toluene, re¯ux; (x) H2S, DMF, Br2; (xi) HBr/AcOH (33%); (xii) (Boc)2O, K2CO3; (xiii) CHꢀCCH2Br, K2CO3, (C4H9)4N HSO4 (c),
CH2ˆCHCH2Br, K2CO3 (d); (xiv) HCl, ether; (xv) HCOOH, CH2O, re¯ux.
Scheme 3. Reagents and conditions: (i) CH2ˆCHCN, C2H5OH; (ii) KOC(CH3)3, toluene, ClCOOC2H5; (iii) H2SO4 (85%), 60±70 ꢁC; (iv)
.
.
C6H5CH2NH2, xylene, re¯ux; (v) H2S, DMF, Br2; (vi) (CH3)2SO4, (C4H9)4N HSO4, NaOH (a), CH2ˆCHCH2Br, (C4H9)4N HSO4, K2CO3 (d); (vii)
.
KOH/CH3OH; (viii) HCOOH, CH2O, re¯ux; (ix) HBr/AcOH (33%); (x) (Boc)2O, K2CO3; (xi) CNCH2Cl, (C4H9)4N HSO4, K2CO3; (xii) HCl, ether,
NaOH; (xiii) CHꢀCCH2Br, (C4H9)4N HSO4, K2CO3; (xiv) HCOOH, CH2O, re¯ux.
.

798
H. Pedersen et al. / Bioorg. Med. Chem. 7 (1999) 795±809
Table 1. Alkylation reactions and resolutions after formation of the 4,5,6,7-tetrahydroisothiazolo[4,5-c]pyridin-3-ol ring system
Starting
compound
Alkylating agent
Yield (%) of
N-deprotected product
Yield (%) of
optical resolution
Yield (%) of
N-methylated product
12
12
13
24
CH2N2
7a,HCl (45%)
(C2H5)2SO4
CHꢀCCH2Br
CHꢀCCH2Br
7b, maleate (29%)
7c, hemifumarate (55%)
8c, HCl (45%)
9c, oxalate (32%)
( )-9c, oxalate (33%)
(+)-9c, oxalate (31%)
( )-8c, fumarate (18%)
(+)-8c, hemifumarate (24%)
24
30
30
CH2ˆCHCH2Br
(CH3)2SO4
8d, fumarate (32%)
10a, fumarate (18%)
10d, fumarate (36%)
CH2ˆCHCH2Br
11d, oxalate (40%)
(
)-10d, hemi-l-tartrate (11%)
(+)-10d, hemi-d-tartrate (8%)
31
31
CHꢀC±CH2Br
NꢀCCH2Cl
10c, fumarate (41%)
10e, fumarate (33%)
11c, oxalate (57%)
(R)-( )-11c, oxalate (56%)
(S)-(+)-11c, oxalate (57%)
(R)-( )-10c, fumarate (22%)
(S)-(+)-10c, fumarate (26%)
were performed using the appropriate alkyl halides in
DMF in the presence of potassium carbonate and tet-
rabutylammonium hydrogen sulphate. In both cases,
parallel alkylations at N-2 were observed, but no
attempts were made to purify and characterize these by-
products. The Boc-protected and O-alkylated products
were deprotected by hydrogen chloride in ether to give
the corresponding secondary amines 8c,d. Compound
8c was N-5-methylated by a standard Eschweiler±
Clarke reaction involving heating in formic acid in the
presence of formaldehyde. The yields of these reactions,
starting with compound 24, and the salts of these C-6-
methylated ®nal products isolated are listed in Table 1.
The syntheses of the C-7-methylated target compounds,
10a,c±e and 11c,d, are outlined in Scheme 3. The reac-
tion sequence for the preparation of the key inter-
mediates, b-oxo amide 28 and 3-isothiazolols 30 and 31,
are with a few exceptions analogous with those shown
in Scheme 2. Thus, 28 was synthesized from 2-methyl-3-
aminopropionitrile 25.26 Addition of acrylonitrile gave
dinitrile 26, which under Thorpe±Ziegler cyclization
conditions, using tert-butoxide in toluene, and sub-
sequent acid hydrolysis and protection of the secondary
amine gave the desired b-oxo nitrile 27 in 87% yield.
Hydrolysis of the nitrile group of 27 gave b-oxo amide
28 in 64% yield. The results of the transformations of
3-isothiazolols 30 and 31 into the ®nal products con-
taining secondary amino groups (10a,c±e) and those
containing N-methyl tertiary amino groups (11c,d) are
listed in Table 1.
On the basis of the results of the pharmacological stu-
dies, it was decided to subject compounds 8c, 10c, and
10d to optical resolution procedures. Subsequently, the
enantiomers of 8c and 10c were transformed into the
respective enantiomers of the N-5-methylated analo-
gues, 9c and 11c, using Eschweiler±Clarke reaction
conditions. The optical resolutions of 8c and 10c were
accomplished by precipitation of diastereomeric salts
with (2S,3S)-(+)-O,O0-dibenzoyltartaric acid [d-(+)-
dibenzoyltartaric acid] and (2R,3R)-( )-O,O0-diben-
Figure 2. Perspective drawings27 of the two conformationally di€erent
cations of compound (S)-(+)-10c crystallized as a salt with d-(+)-
dibenzoyltartaric acid. Spheres representing the isotropic or equivalent
isotropic displacement parameters of the non-hydrogen atoms are
shown at the 50% probability level Hydrogen atoms in calculated
positions are represented by spheres of arbitrary size.

H. Pedersen et al. / Bioorg. Med. Chem. 7 (1999) 795±809
799
zoyltartaric acid [l-( )-dibenzoyltartaric acid]. Com-
Receptor binding
pound 10d was resolved by precipitation of diastereo-
meric salts with d-( )- and l-(+)-tartaric acids. The
results of these resolution procedures are shown in
Table 1. The optical purity was determined by HPLC,
The compounds 5±11 were studied in receptor binding
assays using rat brain and rat brain stem membranes
and tritiated quinuclidinyl benzilate ([3H]QNB, pir-
enzepine ([3H]PZ), and oxotremorine-M ([3H]Oxo-M)
as ligands. On the basis of these radioligand binding
studies, M2/M1 index21 and agonist index21,28 values
were calculated (Table 2). The former index is indicative
of selectivity for the M1 receptor site, higher M2/M1
index values predicting greater M1 selectivity.21 The
agonist index values, on the other hand, have been
shown to be predictive of relative ecacy of compounds
at mAChRs,21,28 and in our test setup, values above 100
and between 100 and 10 are considered indicative of full
and partial mAChR agonism, respectively, whereas
values below 10 are predictive of antagonism.29 It must
be emphasized that these indexes are empirical21,28 and
should only be used to predict e€ects at mAChRs in
relative terms within series of structurally related com-
pounds. On the basis of calculated M2/M1 index and
agonist index values (Table 2) a few compounds of par-
ticularly interesting apparent pharmacological pro®le
were selected for studies using the ®ve cloned human
mAChRs (m1±m5) using the functional assay, receptor
selection and ampli®cation technology (R-SAT2)23,30±32
(see subsequent section).
1
using a chiral stationary phase, or by H NMR spec-
troscopy, using the chiral shift reagent (R)-( )-2,2,2-
tri¯uoro-1-(9-anthryl)ethanol which had sucient
enantioselective in¯uence on the chemical shifts of the
methyl groups at C-6 of the enantiomers of 8c or the
methyl groups at C-7 of the enantiomers of 10c and 10d.
The absolute con®guration of (R)-( )- and (S)-(+)-10c
and, thus of (R)-( )- and (S)-(+)-11c was established
by an X-ray crystallographic analysis.
X-ray crystallographic analysis
The X-ray analysis established the absolute con®gura-
tion of (+)-10c, crystallized as a salt with d-(+)-
dibenzoyltartaric acid, to be S. This salt is a 2:1 complex
of (S)-(+)-10c to the d-(+)-dibenzoyltartrate dianion
together with an ethanol molecule of solvation from the
recrystallization. Thus, the absolute con®guration of
(S)-(+)-10c was based on the salt with an anion of
known absolute stereochemistry. The geometry of
cationic (S)-(+)-10c with atomic labeling is shown in
Figure 2.
Table 2. Receptor binding data of muscarinic ligands using rat brain and heart tissues
Receptor binding
[3H]Oxo-Ma
(brain)
[3H]PZb
(brain)
[3H]QNBc
(brain)
[3H]QNBd
(brain stem)
Agoniste
index
M2/M1f
index
Compound
Ki values (nM)
Oxotremorine
Arecoline (1)
5
0.27
1.3
0.35
0.24
12
20
4.4
14
35
650
7.1
10
1000
240
34
14
19
37
57
7.7
1.2
6.0
84
4.4
5.0
3.1
7.1
23
10
31
4.3
2.8
9.0
130
1100
19
3.6
42
2.7
1.0
370
33
50
92
85
250
210
24
20
54
110
13
11
20
480
850
54
92
192
15
0.1
0.06
0.4
0.1
0.4
0.1
1.5
6.6
4.5
6.8
3.7
3.1
17
9.0
1.3
3.0
2.2
6.5
2.5
1.6
3.2
5.2
14
6
22
7a
7b
7c
8c
2300
290
230
120
100
330
310
34
28
81
290
33
23
20
26
42
49
430
54
52
8.6
9.1
8.0
4.8
5.6
18
15
17
25
47
(
)-8c
11
41
65
(+)-8c
8d
9c
6.1
1.6
7.7
17
1.3
0.49
1.1
3.7
7.0
9.2
28
(
)-9c
(+)-9c
10a
10c
(R)-( )-10c
(S)-(+)-10c
10d
18
18
36
32
160
61
32
7.0
6.0
5.3
15
15
13
(
)-10d
(+)-10d
10e
11c
(R)-( )-11c
(S)-(+)-11c
11d
3.7
2.7
13
35
190
160
11
20
9.4
180
160
15
7.0
23
17
a[3H]Oxo-M binding on rat brain homogenate.
b[3H]PZ binding on rat brain homogenate.
c[3H]QNB binding on rat brain homogenate.
d[3H]QNB binding on rat brain stem homogenate. For details see ref 43.
eAgonist index=Ki (QNBbrain)/Ki (Oxo-M).
fM2/M1 index=Ki (QNBbrain stem)/Ki (PZ).

800
H. Pedersen et al. / Bioorg. Med. Chem. 7 (1999) 795±809
Within
the
group
of
3-alkoxy-4,5,6,7-tetra-
response to the standard full agonist carbachol, and a
hydroisothiazolo[4,5-c]pyridines without substituents in
the tetrahydropyridine ring, compounds 7a±c (Fig. 1),
only 7a shows the characteristics of a full mAChR ago-
nist (agonist index=192) (Table 2), whereas 7c shows
the highest degree of M1 selectivity (M2/M1 index=1.5).
This receptor selectivity, which is of particular pharma-
cological interest, could be increased by incorporation
of methyl groups at various sites of the molecule of 7c.
Thus, the C-6 methyl substituted analogue of 7c, com-
pound 8c, appears to show at least 4 times higher M1
selectivity (M2/M1 index=6.6) than 7c. Whereas
methylation of 7c at C-7, to give 10c, results in a slightly
increased M1 selectivity, N-methylation of 10c markedly
increases M1 selectivity, 11c showing an M2/M1 index
value of 14, which is about an order of magnitude
higher than that of the parent compound, 7c.
slope not signi®cantly di€erent from unity). A similar
analysis of the e€ects of the 10c enantiomers at m3 did,
on the other hand, show signs of insurmountable
antagonism (statistically signi®cant suppression of
maximal responses at higher antagonist concentrations)
(Fig. 4).
At all ®ve receptors, except m3, (S)-(+)-10c was slightly
more potent than (R)-( )-10c. Both compounds were
In agreement with earlier observations21,22 substitution
of a propargyloxy group for the methoxy group of 7a,
to give 7c, increases mAChR anity in the [3H]Oxo-M
binding assay, assumed to re¯ect the anity for the
agonist conformation of the mAChRs.33 Similarly,
compound 10c, containing an O-propargyl group, is
about an order of magnitude more potent than the cor-
responding O-methyl analogue, 10a, in this binding
assay (Table 2). Whereas the O-alkyl analogue, 10d, is
slightly weaker, the O-cyanomethyl analogue, 10e, is
markedly weaker than 10c. The compounds 8c, 10c, and
10d were selected for optical resolution, and the enan-
tiomers of these compounds were obtained via dia-
stereomeric salt formation using enantiomers of tartaric
acid or dibenzoyltartaric acid as resolving agents. The
receptor anity data and calculated agonist index and
M2/M1 index values for these ®ve pairs of enantiomers
are compared with the respective data for the parent
racemates in Table 2. In all cases, a remarkably low
degree of stereoselectivity was observed.
E€ects on cloned human mAChRs
The e€ects of the enantiomeric pairs of 10c and 11c at
cloned human m1±m5 were studied using the functional
assay, R-SAT2.23,30±32 In agreement with the receptor
binding data (Table 2), all four compounds were potent
mAChR ligands, showing EC50 (agonists) or Ki
(antagonists) values in the nanomolar range (Table 3,
Fig. 3). Both enantiomers of 10c showed partial agon-
ism at m1, m2, and m4, (R)-( )-10c having the highest
relative ecacy at all three receptors. Both compounds
showed higher relative ecacy at m2 and m4 than at
m1, a tendency which has been seen for a number of
standard compounds (Table 3), and which, at least par-
tially, is an inherent result of the R-SAT2 technology.30
Compared to the standard compound 6,23 both enan-
tiomers of 10c showed lower relative ecacy at all ®ve
human mAChRs, but at m3 and m5 neither compound
displayed any detectable activity. A Schild analysis34
revealed that (R)-( )-10c as well as (S)-(+)-10c showed
antagonist e€ects at m3 and m5 (Table 3). The Schild
analysis of data from studies on the m5 receptor was
consistent with both enantiomers of 10c acting as com-
petitive antagonists (no suppression of maximal
Figure 3. Pharmacological pro®le of (R)-( )-10c and (S)-(+)-10c
using the cloned human mAChRs as determined by R-SAT: (A) m1,
(B) m2, and (C) m4. % Response indicates responses relative to max-
imal responses by the full agonist carbachol. Each datapoint is the
mean value‹SEM of a representative experiment performed in tripli-
cate (m1) or quadruplicate (m2 and m4).

H. Pedersen et al. / Bioorg. Med. Chem. 7 (1999) 795±809
801
slightly more potent at m2 and m4 than at m1, m3, and
Discussion
m5, a trend also seen for a number of standard com-
pounds such as carbachol, arecoline (1), oxotremorine,
5, and 6 (Table 3). This can be partially explained by a
higher receptor reserve/G-protein coupling of m2 and
m4 in R-SAT.30,35
The naturally occurring alkaloid, arecoline (1) is an
agonist at mAChRs and has been shown to improve
cognition when administered to AD patients36 and to
enhance learning in young humans and aged non-
human primates.37,38 The pharmacological e€ects of 1
are, however, shortlived, probably due to hydrolysis of
the ester group, and the adverse e€ects of 1 after
administration in man may re¯ect lack of M1 receptor
selectivity and non-optimal relative agonist ecacy at
subtypes of central mAChRs. As attempts to develop
mAChR agonists or partial agonists more suitable for
therapeutic use, we have previously synthesized and
pharmacologically described di€erent series of con-
formationally restricted bicyclic analogues of 1 contain-
ing the ring systems of 2,19,20 3,22 and 4±6.23
The N-methylated analogues of the enantiomers of 10c,
(R)-( )-11c and (S)-(+)-11c, showed no agonist e€ects
at m1, m3, or m5. A Schild analysis of data from studies
of e€ects of these enantiomers at m1 and m5 was con-
sistent with competitive antagonism by both com-
pounds with no suppression of maximal responses and
with slopes not signi®cantly di€erent from unity. Simi-
lar analyses on the e€ects of (R)-( )-11c and (S)-(+)-
11c at m3 did, however, show signi®cant inhibition of
maximal responses at higher antagonist concentrations,
thus indicating insurmountable antagonism. The (R)-
enantiomer was the more potent compound at all three
receptors.
Structure±activity studies on these previously described
series of mAChR ligands led to the conclusion that the
Table 3. Pharmacological parameters of muscarinic agonists and antagonists at the ®ve cloned human muscarinic receptor subtypes determined by
receptor selection and ampli®cation technology (R-SAT)
Receptor response (% of maximal carbachol response)
m1
m2
m3
m4
m5
EC50(m2)/EC50(m1)
Compound
pEC50 or pKB
5.9‹0.1 (100%)
5.5‹0.1 (86‹3%) 7.6‹0.0 (105‹0%) 6.5‹0.1 (66‹9%) 7.0‹0.2 (70‹3%) 6.2‹0.0 (77‹2%)
Carbachola
Arecoline (1)a
5.3‹0.0 (100%)
7.1‹0.0 (100%)
7.1‹0.1 (100%)
6.5‹0.0 (100%)
0.016
0.078
0.049
0.43
0.048
0.24
0.57
Oxotremorinea 6.4‹0.1 (75‹10%) 7.9‹0.3 (105‹6%) 6.7‹0.1 (66‹5%) 7.5‹0.2 (102‹3%) 7.3‹0.0 (74‹2%)
5a
6a
7.1‹0.1 (42‹2%) 7.4‹0.1 (85‹6%) 6.4‹0.2 (34‹6%) 7.2‹0.1 (89‹5%) 5.8‹0.2b
6.9‹0.1 (59‹9%) 8.3‹0.2 (86‹2%) 6.9‹0.1 (70‹8%) 7.5‹0.1 (105‹4%) 6.4‹0.2 (36‹5%)
(R)-( )-10c
(S)-(+)-10c
(R)-( )-11c
(S)-(+)-11c
6.1‹0.0 (39‹3%) 6.7‹0.1 (80‹11%) 6.6‹0.1c
6.6‹0.0 (14‹2%) 6.9‹0.2 (67‹5%) 6.4‹0.1c
6.7‹0.1 (94‹6%) 6.1‹0.3d
6.9‹0.1 (34‹6%) 6.6‹0.2d
6.7‹0.0d
6.6‹0.2d
nd
nd
6.3‹0.1c
6.1‹0.2c
nd
nd
6.3‹0.2d
6.0‹0.1d
aFrom ref 23.
bpKB value calculated from IC50 by the Cheng±Pruso€ equation.44
cSchild analysis shows signs of insurmountable antagonism (depressions of maximal response at higher antagonist doses); pKB values calculated as
described by Kenakin.53
dpKB values calculated from Schild analysis.34
nd, not determined. Data represent the mean (‹SEM) of two to ®ve experiments. Antagonism values are shown in italics.
Figure 4. Pharmacological pro®le of (R)-( )-10c using the cloned human mAChRs as determined by R-SAT: (A) m3 and (B) m5. % Response
indicates responses relative to maximal responses by the full agonist carbachol. At higher concentrations of (R)-( )-10c there was a statistically
signi®cant supression of maximal responses (p<0.05). Each datapoint is the mean value‹SEM of a representative experiment performed in tripli-
cate.

802
H. Pedersen et al. / Bioorg. Med. Chem. 7 (1999) 795±809
structure of the `top part' of the molecules, including the
structure of the alkoxy groups and the substitution at
the cationic centers, largely determine the pharmacol-
ogy of the compounds.21 Although compounds 4±6 did
not show degrees of M1 selectivity or relative agonist
ecacies markedly di€erent from those of their lower
ring homologues containing tetrahydropyridine rings,
their 10±20 times higher potency promoted the idea that
the pharmacological parameters of such mAChR
ligands might be optimized by structural manipulation
of the `bottom part' of the molecules, assumed to inter-
act with a lipophilic part of the receptors.39 In the new
compounds, 7±11 (Fig. 1), we have therefore varied not
only the structure of the `top part' of the molecules, but
we have also introduced, in a stereospeci®c manner,
methyl groups into di€erent positions of the `bottom
part'. The isoxazole ring of previously described com-
pounds, exempli®ed by compounds 2±6, has been
replaced by an isothiazol ring in order to increase the
lipophilicity of the `bottom part' of the molecules (Fig. 1).
Five of the new compounds, 8c, 9c, 10c, 10d and 11c,
were resolved, and the absolute stereochemistry of the
enantiomers of 10c and 11c was established by an X-ray
analysis. Generally, these pairs show a remarkably low
degree stereoselectivity in the binding (Table 2) and
functional (Table 3) assays. This may re¯ect that the
chiral centers of these compounds are located at sites
(C-6 or C-7), which, as previously postulated,39 interact
with very lipophilic and perhaps not particularly chiral
parts of the mAChR sites.
Some of the compounds described in this paper show
pharmacological pro®les, which make them interesting
candidates for clinical studies in, for example, AD
patients. These aspects will not be described and dis-
cussed in the present paper.
Experimental
Chemistry
As in previous structure±activity studies on mAChR
ligands, M1 selectivity and relative agonist ecacy were
estimated on the basis of M2/M1 and agonist index
values, respectively, derived from mAChR binding data
using di€erent radioligands21,28,29 (Table 2). Two pairs
of enantiomers, (R)-( )- and (S)-(+)-10c and (R)-( )-
and (S)-(+)-11c were further characterized pharmaco-
logically and compared with a series of standard
mAChR ligands using the ®ve cloned human receptors
m1±m5 (Table 3). The relative order of M1 selectivity
has been estimated using the M2/M1 index (Table 2) and
determined using the EC50 (m2)/EC50 (m1) ratio
(Table 3). Higher values of these two series of data
indicate higher M1 selectivity, and the relative order of
these two sets of data actually are similar, though not
identical. As examples, the relative order of M1 selec-
tivity of 5 and 6 and of (R)-( )-10c and (S)-(+)-10c are
identical.
Column chromatography (CC) was performed on silica
gel 60, 230±400 mesh, ASTM. Evaporations were per-
formed under vacuum at approximately 20 mmHg.
Melting points were determined in capillary tubes and
are not corrected. 1H NMR Spectra were recorded at 80
MHz on a Bruker WP 80 DS spectrometer or at
250 MHz on a Bruker AC 250 spectrometer with TMS
as an internal standard; where nothing is stated, the
spectrum was recorded at 80 MHz. Chemical shifts are
expressed in ppm values and are relative to internal
TMS. The following abbreviations are used for multi-
plicity of NMR signals: s=singlet, d=doublet, t=tri-
plet, q=quartet, qui=quintet, m=multiplet. NMR
signal corresponding to acidic protons are omitted.
Contents of H2O in crystalline compounds was deter-
mined by Karl Fischer titration. Microanalyses, Karl
Fischer titrations and optical rotations were performed
by Lundbeck Analytical Department, and results
obtained for microanalyses were within ‹0.4% of the
theoretical values unless otherwise stated.
All of the new compounds, 7±11, are potent inhibitors
of the binding of mAChR radioligands, and in the
[3H]Oxo-M agonist binding assays, Ki values are typi-
cally found in the low nanomolar range (Table 2). With
the exception of 7a, all of the compounds show agonist
index values below 100, indicating partial agonism
(values between 100 and 10) or antagonism (values
below 10),29 a majority of compounds showing the
characteristics of low-ecacy partial agonists (Table 2).
In the functional assays, the enantiomers of 10c show
partial agonist or antagonist e€ects, whereas the enan-
tiomers of 11c show antagonism of the receptors tested
(Table 3). This suggests that the compounds pre-
ferentially interact with the antagonist conformation(s)
of the mAChRs,33 and may explain why most of the
new compounds show comparable e€ects in the
mAChR agonist binding assay using [3H]Oxo-M as the
radioligand (Table 2). In other series of structurally
related mAChR ligands, as exempli®ed by 2,21 3,22 and
4±6,23 which typically show higher degrees of relative
agonist ecacies, structural changes of the compounds
typically have more marked in¯uence on agonist bind-
ing anities.
3-Methoxy-4,5,6,7-tetrahydroisothiazolo[4,5-c]pyridine (7a)
hydrochloride. To a solution of methyl 3-hydroxy-
4,5,6,7-tetrahydroisothiazolo[4,5-c]pyridine-5-carbox-
ylate (12)24 (1.60 g, 7.5 mmol) in Et2O (50 mL) and
EtOH (2 mL) was added excess diazomethane. The
mixture was stirred at room temperature for 1 h, and
excess diazomethane was destroyed by addition of
AcOH. The mixture was evaporated and the residue was
eluted from silica gel (CC) with toluene/EtOAc 1/1. The
isolated O-methyl isomer (0.89 g, 3.9 mmol) was dis-
solved in MeOH (9 mL) containing KOH (2.0 g), and
the mixture was re¯uxed for 20 h. The reaction mixture
was evaporated. The residue was dissolved in H2O
(30 mL) and extracted with CHCl3 (3Â50 mL). Standard
workup of the combined organic phases yielded an oil
which crystallized as the hydrochloric salt from Et2O/
EtOAc. Yield: 0.70 g (45%); mp 234±35 ꢁC. H NMR
1
(DMSO-d6) 2.85±3.20 (m, 2H), 3.18±3.48 (m, 2H), 3.99
.
(s, 2H), 3.95 (s, 3H). Anal. (C7H10N2OS HCl) H; C:
calcd, 40.66; found, 42.79. N: calcd, 13.55; found 12.51.

H. Pedersen et al. / Bioorg. Med. Chem. 7 (1999) 795±809
803
3-Ethoxy-4,5,6,7-tetrahydroisothiazolo[4,5-c]pyridine (7b)
maleate. A solution of 1224 (1.50 g, 7.0 mmol), tetra-
butylammonium hydrogen sulphate (0.1 g, 0.31 mmol)
and NaOH (0.6 g, 15 mmol) in CH2Cl2 (6 mL) and H2O
(6 mL) was stirred for 10 min. Et2SO4 (1 mL, 7.63 mmol)
was added and the mixture was re¯uxed for 16 h. Con-
centrated NH4OH (10 mL) was added and re¯ux was
continued for 1 h. The mixture was diluted with CH2Cl2
(100 mL) and the phases were separated. Conventional
workup of the organic phase gave an oil (2.5 g) which
was eluted from silica gel (CC) with EtOAc/heptane
(1/2) to give 1.1 g (4.5 mmol, 65%) of the protected O-
ethyl derivative which was deprotected analogously to
the preparation of 7a. Yield of 7b: 0.40 g (50%); mp
163±164 ꢁC (EtOAc); 1H NMR (DMSO-d6) 1.30 (t, 3H),
2.90±3.25 (m, 2H), 3.25±3.55 (m, 2H), 4.05 (s, 2H), 4.40
with 2 M NaOH solution and the free base of 7c was
isolated in a conventional manner. The hemifumarate
was crystallized from Me2CO. Yield: 0.60 g (55%); mp
186±188 ꢁC; H NMR (DMSO-d6) 2.75±3.25 (m, 4H),
1
3.55 (t, 1H), 3.75 (s, 2H), 5.0 (d, 2H), 6.50 (s, 1H). Anal.
.
(C9H10N2OS 0.5C4H4O4) C, H, N.
Methyl (RS)-3-(N-benzylamino)butyrate (15). A solution
of 14 (1188 g, 11.9 mol) and benzylamine (980 g, 9.2 mol)
in EtOH (1 L) was heated to re¯ux for 16 h. Removal of
the solvent gave the title product as an oil, which was
used for the synthesis of 16 without further puri®cation.
Yield: 1890 g (99%). 1H NMR (CDCl3, 250 MHz) d
1.20 (d, 3H),1.80 (s, 1H), 2.45 (dd, 2H), 3.20 (sextet,
1H), 3.70 (s, 3H), 3.80 (s, 2H).
.
(q, 2H), 6.05 (s, 2H). Anal. (C8H12N2OS C4H4O4) C, H,
N.
Methyl (RS)-3-aminobutyrate (16). To a solution of
crude 15 (485 g, ca. 2.34 mol) in MeOH (1.7 L) was
added ammonium formate (269 g, 3.5 mol) and 5% pal-
ladium on charcoal (59 g). The mixture was re¯uxed for
3 h. Filtration and removal of the solvent gave the title
product as an oil, which was used for the synthesis of 17
without further puri®cation. Yield: 232 g (94%). 1H
NMR (CDCl3, 250 MHz) 1.20 (d, 3H), 2.80 (s, 2H),
2.40±2.55 (m, 2H), 3.25±3.70 (m, 1H), 3.75 (s, 3H).
tert-Butyl 3-hydroxy-4,5,6,7-tetrahydroisothiazolo[4,5-c]-
pyridine-5-carboxylate (13). A solution of 1224 (6.40 g,
30 mmol) in 33% HBr in AcOH (100 mL) was stirred
for 72 h. The mixture was evaporated, and EtOH
(10 mL) was added and evaporated. The residue was
dissolved in H2O (60 mL) and was washed with CH2Cl2
(2Â25 mL). The aqueous phase was cooled on an ice
bath and K2CO3 (7 g, 51 mmol) was added below 20 ꢁC.
The aqueous phase was washed with CH2Cl2 (100 mL),
and was cooled to 5 ꢁC. A solution of (Boc)2O (10 g,
52 mmol) in THF (100 mL) was added below 15 ꢁC.
After addition the mixture is stirred for 16 h. The phases
were separated, and the aqueous phase was extracted
with Et2O (4Â30 mL). The combined organic phases
were washed with saturated NaCl solution (3Â30 mL).
Drying over MgSO4 and activated carbon and eva-
poration yielded 9.5 g of a semicrystalline material
(mostly diBoc-derivative), which was dissolved in
MeOH (200 mL). Anhydrous K2CO3 (3.5 g, 25 mmol)
was added, and the mixture was stirred for 14 h. MeOH
was removed, and the residue was dissolved in H2O
(50 mL) and washed with Et2O (3Â30 mL). The aqueous
phase was acidi®ed to pH 6 with concentrated HCl and
extracted with EtOAc (4Â30 mL). The EtOAc phases
were washed with saturated NaCl solution (2Â30 mL),
dried over MgSO4 and evaporated to yield 7.0 g of
crude product.
Methyl rac-3-amino-N-(2-cyanoethyl)butyrate (17). A
solution of crude 16 (248 g, ca. 2.11 mol) and acryloni-
trile (140 mL, 2.13 mol) in MeOH (490 mL) was heated
to re¯ux for 60 h. Removal of the solvent gave the title
compound as an oil, which was used for the synthesis of
1
18 without further puri®cation. Yield: 318 g (89%). H
NMR (CDCl3, 250 MHz) 1.20 (d, 3H), 2.00 (s, 1H),
2.25±2.75 (m, 4H), 2.80±3.45 (m, 3H), 3.75 (s, 3H).
Ethyl rac-3-cyano-6-methyl-4-oxopiperidine-1-carboxylate
(18). A mixture of potassium tert-butoxide (213 g,
1.90 mol) in toluene (2.2 L) was heated to 70 ꢁC with
stirring. Crude 17 (318 g, ca. 1.87 mol) was added while
keeping the temperature at 70±80 ꢁC. After the addition
the mixture was heated to re¯ux for 30 min. After cool-
ing to 10 ꢁC on an ice bath the potassium enolate was
®ltered o€ and was dissolved in ice H2O (2 L) (the ®l-
tercake must be kept wet with toluene to avoid decom-
position). K2CO3 (310 g, 2.25 mol) was added, and ethyl
chloroformate (196 mL, 2.05 mol) was added dropwise
at 10 ꢁC. After the addition the mixture was stirred for
16 h at room temperature. EtOAc (1 L) was added and
the organic phase was worked up in the usual manner,
giving the 4-enolcarbonate of 18 as an oil (167 g,
0.59 mol). The aqueous phase was acidi®ed with 4 M
HCl and was worked up with EtOAc in the usual man-
ner giving the title compound (109 g, 0.52 mol). The 18-
4-enol carbonate was dissolved in MeOH (1 L) contain-
ing K2CO3 (50 g) and the mixture was stirred for 16 h at
room temperature. The solvent was removed, and the
residue was dissolved in H2O (1 L). The aqueous phase
was acidi®ed with 4 M HCl and was worked up with
EtOAc in the usual manner giving the title product
(99 g, 0.47 mol), which was used for the synthesis of 19
without further puri®cation. Total yield: 208 g (53%). A
3-Propargyloxy-4,5,6,7-tetrahydroisothiazolo[4,5-c]pyridine
(7c) hemifumarate. To a solution of crude tert-butyl 3-
hydroxy-4,5,6,7-tetrahydroisothiazolo[4,5-c]pyridine-5-
carboxylate 13 (1.50 g, ca. 5.6 mmol) in DMF (30 mL)
was added tetrabutylammonium hydrogen sulphate
(0.1 g, 0.31 mmol), K2CO3 (2.0 g, 14.5 mmol) and 3-bro-
mopropyne (1.5 mL, 16.8 mmol) and the mixture was
stirred for 4 h at 70 ꢁC. The solvent was removed, and
the residue was dissolved in Et2O (100 mL). The organic
phase was washed with saturated CaCl2 solution
(3Â50 mL) and was then worked up in a conventional
manner. The residue was dissolved in Et2O (20 mL) and
a concentrated solution of HCl in Et2O (20 mL) was
added. After stirring for 3 h at room temperature the
solvent was removed and the oily residue was dissolved
in H2O (50 mL). The aqueous phase was made basic
ꢁ
1
sample was crystallized from EtOAc: Mp 63±65 C. H
NMR (CDCl3, 250 MHz) 1.60 (d, 3H), 1.70 (t, 3H), 2.80

804
H. Pedersen et al. / Bioorg. Med. Chem. 7 (1999) 795±809
(dd, 1H), 3.10 (dd, 1H), 3.70±3.90 (m, 1H), 4.00 (dd,
1H), 4.60 (q, 2H), 5.0±5.45 (m, 2H). Anal. (C10H14N2O3)
H, N; C: calcd, 57.12; found, 57.62.
(m, 4H), 4.95 (s, 2H (NH2)), 7.20±7.40 (m, 5H), 10.2
(t, 1H).
Ethyl
(RS)-3-hydroxy-6-methyl-4,5,6,7-tetrahydroiso-
Ethyl rac-7-cyano-11-methyl-1,5-dioxa-9-azaspiro[5.6]-
undecane-9-carboxylate (19). A solution of crude 18
(270.0 g, ca. 1.28 mol), 1,3-propanediol and p-toluene-
sulphonic acid monohydrate (5 g, 0.026 mol) in toluene
(1.15 L) was re¯uxed with a Dean±Stark water separa-
tor for 16 h. After cooling to room temperature, H2O
was added, and the phases were separated. The aqueous
phase was extracted with EtOAc, and the combined
organic phases were washed with dilute sodium hy-
droxide solution and worked up in the usual manner
giving the title product (340.7 g, 1.26 mol, 98%) as an
oil, which was used for the synthesis of 20 without fur-
thiazolo[4,5-c]pyridine-5-carboxylate (23). H2S (93 g,
2.74 mol) was dissolved with cooling in DMF (1.5 L)
and crude 22 (124.5 g, ca. 0.39 mol) was added. The
reaction mixture was stirred for 16 h. Then DMF and
excess H2S was removed by evaporation. The residue
was dissolved in EtOAc (1 L) and washed with saturated
CaCl2 solution (4Â150 mL), dried over MgSO4 and
evaporated to a foamy product, which was dissolved in
EtOAc (0.5 L). To this solution was added Br2 (45 mL,
0.84 mol) in CH2Cl2 with cooling at 20 ꢁC. After addi-
tion the reaction mixture was stirred for 16 h at room
temperature, and evaporated. The residue was dissolved
in CH2Cl2 (0.6 L) and washed with ice cold saturated
NH4Cl solution (3Â0.3 L). The organic phase was dried
over MgSO4 and evaporated to give crude 23 as an oil.
Yield: 29.4 g (0.12 mol, 31%). A sample was crystallized
1
ther puri®cation. H NMR (CDCl3) 1.30 (m, 6H), 1.75
(qui, 2H), 2.00 (dd, 1H), 2.75 (m, 2H), 3.25±4.75 (m,
9H).
from EtOAc: mp 148±150 ꢁC. H NMR (CDCl3) d 1.15
1
Ethyl rac-7-carboxamido-11-methyl-1,5-dioxa-9-azaspiro-
[5.6]undecane-9-carboxylate (20). To a solution of crude
19 (174.4 g, ca. 0.671 mol) in EtOH (720 mL) was added
a solution of NaOH (30 g, 0.75 mol) in H2O (50 mL).
The mixture was heated to 60 ꢁC with stirring, and 30%
H2O2 (560 mL, 9.9 mol) was added keeping the tem-
perature at 60±65 ꢁC. After the addition the mixture was
stirred for 2 h at 60 ꢁC and then left overnight at room
temperature. The mixture was concentrated, and the
residue was dissolved in CH2Cl2. The layers were sepa-
rated and the organic phase was worked up in the usual
manner giving the title product (120.4 g, 0.433 mol,
65%) as an oil, which was used for the synthesis of 21
without further puri®cation. 1H NMR (CDCl3) 1.20
(d, 3H), 1.30 (t, 3H), 1.40±2.25 (m, 2H), 2.45 (dd, 1H),
2.90 (dd, 1H), 3.10±4.65 (m, 10H), 6.70 (broad d, 2H
(NH2)).
(d, 3H), 1.30 (t, 3H), 2.65 (d, 1H), 3.05 (dd, 1H), 3.95 (d,
1H), 4.20 (q, 2H) 4.80 (d, 1H), 4.90±5.05 (m, 1H). Anal.
.
(C10H14N2O3S H2O) C, H, N.
tert-Butyl (RS)-3-hydroxy-6-methyl-4,5,6,7-tetrahydro-
isothiazolo[4,5-c]pyridine-5-carboxylate (24). A solution
of crude 23 (29.3 g, ca. 121 mmol) in 33% HBr in AcOH
(400 mL) was stirred for 72 h. The mixture was evapo-
rated, and EtOH (25 mL) was added and evaporated.
The residue was dissolved in H2O (200 mL) and was
washed with CH2Cl2 (2Â50 mL). The aqueous phase
was cooled on an ice bath and K2CO3 (28 g, 203 mmol)
was added below 20 ꢁC. The aqueous phase was washed
with CH2Cl2 (100 mL), and was cooled to 5 ꢁC. A solu-
tion of (Boc)2O (43.3 g, 208 mmol) in THF (100 mL)
was added below 15 ꢁC. After addition the mixture was
stirred for 16 h. The phases were separated, and the
aqueous phase was extracted with Et2O (4Â100 mL).
The combined organic phases were washed with satu-
rated NaCl solution (3Â100 mL). Drying over MgSO4
and activated carbon and evaporation yielded 38.2 g of
a semicrystalline material (mostly diBoc-derivative),
which was dissolved in MeOH (400 mL). Anhydrous
K2CO3 (14.0 g, 102 mmol) was added, and the mixture
was stirred for 14 h. MeOH was removed, and the resi-
due was dissolved in H2O (200 mL) and washed with
Et2O (3Â100 mL). The aqueous phase was acidi®ed to
pH 6 with concentrated HCl and extracted with EtOAc
(4Â100 mL). The EtOAc phases were washed with
saturated NaCl solution (2Â100 mL), dried over MgSO4
and evaporated to give crude 24. Yield: 10.2 g
Ethyl rac-3-carboxamido-6-methyl-4-oxopiperidine-1-carb-
oxylate (21). To a solution of crude 20 (718.4 g, ca.
2.51 mol) in THF (1 L) was added 6 M HCl (2.62 L)
with cooling and stirring. The mixture was stirred for
2 h at room temperature, and then the THF was
removed, and pH was adjusted to 7.0 with dilute NaOH
solution. The mixture was then worked up in the usual
manner with CH2Cl2 yielding the title product as an oil
(220.7 g, 0.97 mol, 39%), which was used for the synth-
esis of 22 without further puri®cation. 1H NMR
(CDCl3, 250 MHz) 1.20 (d, 3H), 1.30 (t, 3H), 2.05 (d,
1H), 2.70 (dd, 1H), 3.65 (dd, 1H), 4.20 (q, 2H), 4.45 (d,
1H), 4.55±4.70 (m, 2H), 5.45 (broad, 2H (NH2)).
1
Ethyl (RS)-4-benzylamino-3-carboxamido-6-methyl-1,2,5,6-
tetrahydropyridine-1-carboxylate (22). A mixture of
crude 21 (220 g, ca. 976 mmol), benzylamine (130 mL,
1.04 mol), and xylene (mixture of isomers, 1.5 L) was
re¯uxed with a Dean±Stark water separator for 90 min.
The reaction mixture was evaporated under vacuum to
constant weight to give crude 22, which was used for the
synthesis of 23 without further puri®cation. Yield:
(0.040 mol, 33%) as an oil. H NMR (CDCl3) 1.20 (d,
3H), 1.50 (s, 9H), 2.40 (d, 1H), 3.00 (dd, 1H), 3.80 (dd,
1H), 4.70 (dd, 1H), 4.75±5.15 (m, 1H).
(RS)-6-Methyl-3-propargyloxy-4,5,6,7-tetrahydroisothia-
zolo[4,5-c]pyridine (8c) hydrochloride. The title product
was prepared from crude 24 and 3-bromopropyne as
described for the preparation of 7c. Yield: 41%. Mp
249.1 g (0.787 mol, 81%) as an oil. H NMR (CDCl3,
250 MHz) d 1.10 (d, 3H), 1.25 (t, 3H), 2.20 (dd, 1H),
2.45 (dd, 1H), 3.75 (d, 1H), 4.15 (q, 2H), 4.30 4.55
169±171 ꢁC; H NMR (DMSO-d6) 1.45 (d, 3H), 2.65±
1
1
3.25 (m, 2H), 3.30±3.60 (m, 1H), 4.05 (s, 3H), 5.05 (d,
.
2H). Anal. (C10H12N2OS HCl) C, H, N.

H. Pedersen et al. / Bioorg. Med. Chem. 7 (1999) 795±809
805
(+)-6-Methyl-3-propargyloxy-4,5,6,7-tetrahydroisothia-
zolo[4,5-c]pyridine [(+)-8c] hemifumarate and ( )-6-
methyl-3-propargyloxy-4,5,6,7-tetrahydroisothiazolo[4,5-
c]pyridine [( )-8c] fumarate. To a stirred solution of 8c
base (10.3 g, 49.2 mmol) in EtOH (20 mL) was added a
solution of d-(+)-dibenzoyltartaric acid (4.4 g,
12.3 mmol) in EtOH (20 mL). The mixture was cooled
to 10 ꢁC, and the crystals (10.2 g) were ®ltered o€. Three
recrystallizations from 80% EtOH gave 5.76 g of a
hemi-d-(+)-dibenzoyltartrate salt with mp 180±181 ꢁC.
The base was set free with NaOH yielding 2.82 g, which
was precipitated from EtOH with fumaric acid (1.7 g).
Yield of (+)-8c hemifumarate: 3.86 g (11.8 mmol, 24%).
described for the preparation of 9c in 31% yield. Mp
61±63 ꢁC (from Me2CO, amorphous). [a]2d5 +19.1ꢁ (c
.
0.5, H2O). Anal. (C11H14N2OS C2H2O4 0.275 H2O) C,
.
H; N: calcd, 8.83; found, 8.40.
(RS)-N-(2-Cyanoethyl)-2-methyl-3-aminopropionitrile (26).
A
solution of (RS)-2-methyl-3-aminopropionitrile
(25)26 (1032 g, 12.27 mol) and acrylonitrile (718 g,
13.53 mol) in EtOH (2.15 L) was re¯uxed for 16 h. The
solution was evaporated to give crude 26 as a light oil
1
(1604 g, ca. 11.7 mol, 95.3%). H NMR (CDCl3) 1.35
(d, 3H), 1.50 (s, 1H, NH), 2.55 (t, 2H), 2.70±2.95 (m,
3H), 3.00 (t, 2H).
Mp 180±181 ꢁC. [a]d25 +52.9ꢁ (c 0.1, H2O). Anal.
1
.
.
(C10H12N2OS 0.5C4H4O4 0.91H2O) C, H, N. H NMR
spectroscopy [(+)-8c base (15 mg) and (R)-( )-2,2,2-
tri¯uoro-1-(9-anthryl)ethanol (60 mg)] gave baseline
splitting of the doublet corresponding to the 6-methyl
group. On the basis of peak intensities, enantiomeric
excess was estimated to be 95.0%. To the crude free base
of ( )-8c, isolated from the ®ltrate from the ®rst crys-
tallization of (+)-8c, hemi-d-(+)-dibenzoyltartrate was
added l-( )-dibenzoyltartaric acid monohydrate (5.0 g),
and ( )-8c fumarate was obtained as described above.
Yield: 3.86 g (11.8 mmol, 24%). Mp 155±7 ꢁC. [a]d25 48.6ꢁ
Ethyl rac-3-cyano-5-methyl-4-oxopiperidine-1-carboxyl-
ate (27). A stirred mixture of potassium tert-butoxide
(589 g, 5.26 mol) in toluene (4.6 L) was heated to 80 ꢁC.
Without heating crude 26 (656 g, 4.78 mol) was added at
a rate that kept the mixture at gentle re¯ux. After the
addition, the mixture was re¯uxed for 0.5 h, cooled to
10 ꢁC and ®ltered with suction (the ®ltercake must be
kept wet with toluene to avoid decomposition). The ®l-
tercake was washed with toluene (1 L) and poured on
ice (1.2 kg). The mixture was acidi®ed with concentrated
HCl (1.52 L) and re¯uxed for 0.5 h. The reaction mix-
ture was cooled on an ice bath and was made basic with
28% NaOH solution (2.6 L) keeping the temperature
below 30 ꢁC. Then the mixture was cooled to 10 ꢁC and
ethyl chloroformate (571 g, 5.26 mol) was added at
10 ꢁC. After the addition the mixture was stirred for
0.5 h on the ice bath. Then the organic phase was dis-
charged, and the aqueous phase was washed once with
EtOAc (600 mL), and was then acidi®ed with con-
centrated HCl (600 mL). The product was extracted
with CH2Cl2 (3Â600 mL), and the combined organic
phases were washed once with saturated NaCl solution,
dried over MgSO4 and evaporated to yield crude 27 as
an oil, which solidi®ed on standing (874 g, ca. 4.16 mol,
87%). Mp 59±61 ꢁC; 1H NMR (250 MHz, CDCl3,
recorded at 57 ꢁC, 1/1 mixture of keto-enol forms) 1.10
(two d, 6H), 1.30 (two t, 6H), 2.55 (heptet, 1H), 2.80
(dd, 1H), 2.90±3.20 (m, 1H), 3.30 (dd, 1H), 3.50±3.75
(m, 2H), 4.15±4.35 (m, 3H), 4.45 (ddd, 1H), 4.80 (ddd,
1H). Anal. (C10H14N2O3) H, N; C: calcd, 57.12; found,
57.64.
.
(c 0.1, H2O). Anal. (C10H12N2OS C4H4O4) C, H, N.
Enantiomeric excess of ( )-8c was estimated by 1H NMR
spectroscopy, as described above for (+)-8c, to be 92.0%.
(RS)-3-Allyloxy-6-methyl-4,5,6,7-tetrahydroisothiazolo-
[4,5-c]pyridine (8d) fumarate. The title product was pre-
pared from crude 24 and allyl bromide as described for
ꢁ
1
the preparation of 7c. Yield: 32%. Mp 169±171 C; H
NMR (DMSO-d6) 1.25 (d, 3H), 2.75 (d, 1H), 2.95±3.25
(m, 3H), 3.70 (s, 2H), 4.70 (dt, 2H), 5.15±5.50 (m, 2H),
6.05 (ddd, 1H), 6.55 (s, 2H). Anal. (C10H14N2O-
.
S C4H4O4) C, H, N.
(RS)-5,6-Dimethyl-3-propargyloxy-4,5,6,7-tetrahydroiso-
thiazolo[4,5-c]pyridine (9c) dioxalate. A solution of 8c
(2.0 g, 9.5 mmol) in formic acid (70 mL) and formalde-
hyde (37% solution, 25 mL) was re¯uxed for 4 h. The
reaction mixture was concentrated, and the residue was
dissolved in 2 M NaOH solution (25 mL) and EtOAc
(50 mL). The aqueous phase was extracted with two
50 mL portions of EtOAc, and the combined organic
phases were worked up in the usual manner giving
1.02 g of an oil, which was crystallized as oxalate. Yield:
0.95 g, (3.4 mmol, 32%). Mp 154±156 ꢁC (from
Ethyl rac-3-carboxamido-5-methyl-4-oxopiperidine-1-carb-
oxylate (28). To 2 L of 85% H2SO4, kept at 65 ꢁC, mol-
ten 27 (600 g, 2.85 mol) was added with cooling at 60±
70 ꢁC. After stirring for 0.5 h at 60±70 ꢁC the mixture is
poured on ice (5 kg), CH2Cl2 (2 L), and NaCl (0.5 kg)
with ecient stirring. The phases were separated, and to
the aqueous phase was added 28% NaOH solution
(1 L), and it was extracted with CH2Cl2 (3Â1 L). The
combined organic phases were washed with saturated
NaCl solution (2Â0.5 L) and evaporated to 620 g of
crude 28, which was dissolved in hot EtOAc (600 mL).
Compound 28 crystallized on cooling yielding 418 g
(1.83 mol, 64%). Mp 133±138 ꢁC; 1H NMR (CDCl3,
250 MHz) 1.20 (d, 3H), 1.30 (t, 3H), 2.45±2.60 (m, 1H),
3.40 (dd, 1H), 3.55±3.70 (m, 1H), 4.05 (d, 1H), 4.15 (d,
1H), 4.20 (q, 2H), 5.4±5.7 (broad, 2H (NH2)), 14.1 (s,
1H (enol H)). Anal. (C10H16N2O4) C, H, N.
1
Me2CO); H NMR (DMSO-d6) 1.30 (d, 3H), 2.80 (s,
3H), 2.95±3.30 (m, 2H), 3.40±3.90 (m, 1H), 3.60 (t, 1H),
4.20 (s, 2H), 5.05, (d, 2H). Anal. (C11H14N2O-
.
S 2C2H2O4) C, H, N.
( )-5,6-Dimethyl-3-propargyloxy-4,5,6,7-tetrahydroiso-
thiazolo[4,5-c]pyridine [( )-9c] oxalate. From ( )-8c as
described for the preparation of 9c in 33% yield. Mp
62±64 ꢁC (from Me2CO, amorphous). [a]2d5 19.6ꢁ (c
.
0.5, H2O). Anal. (C11H14N2OS C2H2O4) C, H; N: calcd,
8.97; found, 8.49.
(+)-5,6-Dimethyl-3-propargyloxy-4,5,6,7-tetrahydroiso-
thiazolo[4,5-c]pyridine [(+)-9c] oxalate. From (+)-8c as

806
H. Pedersen et al. / Bioorg. Med. Chem. 7 (1999) 795±809
Ethyl (RS)-4-Benzylamino-3-carboxamido-5-methyl-1,2,5,6-
tetrahydropyridine-1-carboxylate (29). A mixture of 28
(600 g, 2.63 mol), benzylamine (350 mL, 2.8 mol), and
xylene (mixture of isomers, 3.9 L) was re¯uxed with a
Dean±Stark water separator for 45 min. Then more
benzylamine was added, and re¯ux was continued for
another 45 min. On cooling to 20 ꢁC the title product
crystallized. The crystals were ®ltered o€ and washed
with xylene yielding 737.9 g of 29 (2.32 mol, 88%). Mp
157±166 C; H NMR (CDCl3, 250 MHz) 1.20 (d, 3H),
1.25 (t, 3H), 2.65 (m, 1H), 2.95 (t, 1H), 3.70±4.05 (m,
2H), 4.15 (q, 2H), 4.30±4.50 (m, 3H), 5.05 (broad, 2H
(NH2)), 7.20±7.40 (m, 5H), 10.15 (t, 1H). Anal.
(C17H23N3O3) C, H, N.
tallization. The crystals were ®ltered o€ and dried
yielding 14.5 g with Mp 158±160 ꢁC, which was trans-
formed to base (10.42 g) and again precipitated with
0.25 equiv l-(+)-tartaric acid. After a total of 5 pre-
cipitations from EtOH the yield was 2.79 g (9.8 mmol,
11%) of hemi-l-(+)-tartrate salt. Mp 192±3 ꢁC. [a]d25
39.5ꢁ (c 1, MeOH on the free base). Anal.
1
.
(C10H14N2OS 0.5C4H6O6) C, H, N. H NMR spectro-
scopy [( )-10d base (15 mg) and (R)-( )-2,2,2-tri¯uoro-
1-(9-anthryl)ethanol (60 mg)] gave baseline splitting of
the doublet corresponding to the 7-methyl group. On
the basis of the peak intensities enantiomeric excess was
estimated to 95.6%. The ®ltrates from the crystal-
lizations of ( )-10d hemi-l-(+)-tartrate were combined
and the base was isolated (15.2 g). After 5 precipitations
as described above with d-( )-tartaric acid the yield
was 1.95 g (6.8 mmol, 8%) of hemi-d-( )-tartrate salt.
Mp 191±2 ꢁC. [a]2d5: +39.7ꢁ (c=1, MeOH on the free
ꢁ
1
Ethyl (RS)-3-hydroxy-7-methyl-4,5,6,7-tetrahydroisothia-
zolo[4,5-c]pyridine-5-carboxylate (30). H2S (155 g,
4.56 mol) was dissolved with cooling in DMF (2.5 L),
and 29 (205 g, 0.65 mol) was added. The reaction mix-
ture was stirred for 16 h. Then DMF and excess H2S
was removed by evaporation. The residue was dissolved
in EtOAc (1.5 L) and washed with saturated CaCl2
solution (4Â250 mL), dried over MgSO4 and evaporated
to a foamy product, which was dissolved in EtOAc
(0.8 L). To this solution was added Br2 (75 mL, 1.4 mol)
in CH2Cl2 with cooling at 20 ꢁC. After addition, the
reaction mixture was stirred for 16 h at room tempera-
ture, and evaporated. The residue was dissolved in
CH2Cl2 (1 L) and washed with ice cold saturated NH4Cl
solution (3Â0.5 L). The organic phase was dried over
MgSO4 and evaporated to an oil (116 g) from which 30
was crystallized from EtOAc. Yield: 65 g (0.27 mol,
41%). Mp 144±146 ꢁC; 1H NMR (CDCl3, 250 MHz)
1.30 (d, 3H), 1.35 (t, 3H), 3.05±3.20 (m, 2H), 4.10 4.25
(m, 4H), 4.60 (d, 1H). Anal. (C10H14N2O3S) C, H, N.
.
base). Anal. (C10H14N2OS 0.5C4H6O6) C, H, N. Enan-
tiomeric excess was estimated as described above to
90.6%.
(RS)-3-Allyloxy-5,7-dimethyl-4,5,6,7-tetrahydroisothiazolo-
[4,5-c]pyridine (11d) oxalate. A solution of 10d (2.0 g,
9.5 mmol) in formic acid (70 mL) and formaldehyde
(37% solution, 25 mL) was re¯uxed for 4 h. The reac-
tion mixture was concentrated, and the residue was dis-
solved in 2 M NaOH solution (25 mL) and EtOAc
(50 mL). The aqueous phase was extracted with two
50 mL portions of EtOAc, and the combined organic
phases were worked up in the usual manner giving
1.02 g of an oil, which was crystallized as oxalate. Yield:
1.19 g, (3.8 mmol, 40%) of oxalate salt. Mp 162±3 ꢁC
(from Me2CO); 1H NMR (DMSO-d6) 1.25, 2.75 (s, 3H),
2.80 (dd, 1H), 3.20±3.55 (m, 2H), 3.70 (d, 1H), 4.00 (d,
1H), 4.85 (dt, 2H), 5.20±5.55 (m, 2H), 6.05 (ddd, 1H).
.
(RS)-3-Methoxy-7-methyl-4,5,6,7-tetrahydroisothiazolo-
[4,5-c]pyridine (10a) fumarate. The title product was
prepared from 30 and Me2SO4 as described for the pre-
paration of 7b. Yield: 18%. Mp 167±168 ꢁC (from
Anal. (C11H16N2OS C2H2O4) C, H, N.
tert-Butyl (RS)-3-hydroxy-7-methyl-4,5,6,7-tetrahydro-
isothiazolo[4,5-c]pyridine-5-carboxylate (31). A solution
of 30 (366 g, 1.51 mol) in 33% HBr in AcOH (4.5 L) was
stirred for 72 h. The mixture was evaporated, and EtOH
(250 mL) was added and evaporated. The residue was
dissolved in H2O (2.3 L) and was washed with CH2Cl2
(2Â300 mL). The aqueous phase was cooled on an ice
bath and K2CO3 (350 g, 2.53 mol) was added below
20 ꢁC. The aqueous phase was washed with CH2Cl2
(0.5 L), and was cooled to 5 ꢁC. A solution of (Boc)2O
(541 g, 2.6 mol) in THF (1 L) was added below 15 ꢁC.
After addition, the mixture was stirred for 16 h. The
phases were separated, and the aqueous phase was
extracted with Et2O (4Â0.5 L). The combined organic
phases were washed with saturated NaCl solution
(3Â0.5 L). Drying over MgSO4 and activated carbon
and evaporation yielded 477 g of a semicrystalline
material (mostly diBoc-derivative), which was dissolved
in MeOH (4 L). Anhydrous K2CO3 (175 g, 1.27 mol)
was added, and the mixture was stirred for 14 h. MeOH
was removed, and the residue was dissolved in H2O
(2 L) and washed with Et2O (3Â0.5 L). The aqueous
phase was acidi®ed to pH 6 with concentrated HCl and
extracted with EtOAc (4Â0.5 L). The EtOAc phases
were washed with saturated NaCl solution (2Â0.5 L),
1
Me2CO); H NMR (DMSO-d6, 250 MHz) 1.25 (d, 3H),
2.65 (dd, 1H), 3.10±3.25 (m, 1H), 3.30 (dd, 1H), 3.75 (d,
1H), 3.80 (d, 1H), 3.95 (s, 3H), 6.65 (s, 2H). Anal.
.
(C8H12N2OS C4H4O4) C, H, N.
(RS)-3-Allyloxy-7-methyl-4,5,6,7-tetrahydroisothiazolo-
[4,5-c]pyridine (10d) fumarate. The title product was
prepared from 30 and allyl bromide. The alkylation was
performed as described for 7c, and the deprotection as
described for 7a. Yield: 36%. Mp 162±164 ꢁC (from
1
Me2CO); H NMR (DMSO-d6, 250 MHz) 1.25 (d, 3H),
2.70 (dd, 1H), 3.2 (sextet, 1H), 3.35 (dd, 1H), 3.75 (d,
1H), 3.85 (d, 1H), 4.85 (dt, 2H), 5.25 (dd, 1H), 5.40 (dd,
1H), 6.00±6.20 (m, 1H), 6.55 (s, 2H). Anal.
.
(C10H14N2OS C4H4O4) C, H, N.
( )-3-Allyloxy-7-methyl-4,5,6,7-tetrahydroisothiazolo[4,5-
c]pyridine [( )-10d] hemi-L-(+)-tartrate and (+)-3-allyl-
oxy-7-methyl-4,5,6,7-tetrahydroisothiazolo[4,5-c]pyridine
[(+)-10d] hemi-D-( )-tartrate. To a solution of 10d
base (18.75 g, 0.089 mol) in EtOH (50 mL) was added a
solution of l-(+)-tartaric acid (3.34 g, 0.022 mol) in
EtOH (20 mL). The mixture was left overnight for crys-

H. Pedersen et al. / Bioorg. Med. Chem. 7 (1999) 795±809
807
dried over MgSO4 and evaporated to yield 350 g of
crude product from which 31 could be crystallized from
Et2O. Yield: 235 g (0.869 mol, 58%), mp 143±145 ꢁC; 1H
NMR (CDCl3) 1.25 (d, 3H), 1.50 (s, 9H), 2.95±3.15 (m,
2H), 4.00±4.25 (m, 2H), 4.50 (d, 1H). Anal.
(C12H18N2O3S) C, H, N.
Mp 156±157 ꢁC (from Me2CO). [a]2d5: 41.6ꢁ (c=0.5,
.
H2O). Anal. (C11H14N2OS C2H2O4) C, H, N.
(S)-(+)-5,7-Dimethyl-3-propargyloxy-4,5,6,7-tetrahydro-
isothiazolo[4,5-c]pyridine [(S)-11c] oxalate. From (S)-10c
as described for the preparation of 11d in 57% yield.
Mp 151±3 ꢁC (from Me2CO). [a]d25: +42.5 (c=0.5, H2O).
.
(RS)-3-Cyanomethyloxy-7-methyl-4,5,6,7-tetrahydroiso-
thiazolo[4,5-c]pyridine (10e) fumarate. The title com-
pound was prepared from 31 and chloroacetonitrile as
described for the preparation of 7c. Yield: 33%. Mp
196±197 ꢁC (from EtOAc); 1H NMR (DMSO-d6,
250 MHz) 1.25 (d, 3H), 2.60 (dd,1H), 3.05±3.25 (m, 1H),
3.25 (dd, 1H), 3.70 (d, 1H), 3.80 (d, 1H), 5.30 (s, 2H),
Anal. (C11H14N2OS C2H2O4) C, H, N.
X-ray crystallographic analysis
Compound C40H44N4O11S2; the asymmetric unit con-
sists of two cations (C10H13N2SO+) with one doubly
2
charged d-(+)-dibenzoyltartrate anion (C18H12O8
)
.
6.55 (s, 2H). Anal. (C9H11N3OS C4H4O4) C, H, N.
and an ethanol (C2H5OH) of solvation; Mr=820.94,
monoclinic, space group P21, a=8.7253(7), b=
21.142(2), c=11.876(2) A,=105.28(1), V=2113.3(8) A3,
(RS)-7-Methyl-3-propargyloxy-4,5,6,7-tetrahydroisothia-
zolo[4,5-c]pyridine (10c) fumarate. The title product was
prepared from 31 and 3-bromopropyne as described for
3
Z=2, Dx=1.290g cm
, monochromatized Cu Ka
radiation (l=1.541838 A), m=1.62 mm 1, F(000)=864,
T=294 K. Data were collected on a Enraf±Nonius
CAD4 di€ractometer to a y limit of 75 which yielded
4757 measured (4489 unique) re¯ections. There are 3602
unique, observed re¯ections (with Iꢂ3s(I) as the criter-
ion for being observed) out of the total measured. The
structure was solved by direct methods40,41 and re®ned
using full-matrix least-squares on F (SDP-PLUS).42 The
®nal model was re®ned using 434 parameters and 3602
observed re¯ections. The non-hydrogen atoms were
re®ned with a mixture of anisotropic (41 atoms) and
isotropic (16 atoms) displacement parameters. The
hydrogen atoms were included at their calculated posi-
tions. The ®nal agreement statistics are: R=0.057,
wR=0.073, S=2.17 with (Á/s)max=0.01. The least-
squares weights were de®ned using 1/s2(F). The max-
the preparation of 7c. Yield: 41%. Mp 184±186 ꢁC; H
1
NMR (DMSO-d6) 1.25 (d, 3H), 2.75 (dd, 1H), 3.00±3.50
(m, 3H), 3.55 (t, 1H), 3.80 (s, 2H), 5.00 (d, 2H), 6.50 (s,
.
2H). Anal. (C10H12N2OS C4H4O4) C, H, N.
(S)-(+)-7-Methyl-3-propargyloxy-4,5,6,7-tetrahydroiso-
thiazolo[4,5-c]pyridine [(S)-10c] fumarate and (R)-( )-7-
methyl-3-propargyloxy-4,5,6,7-tetrahydroisothiazolo[4,5-
c]pyridine [(R)-10c] fumarate. To a stirred solution of
10c base (102.5 g, 0.492 mol) in EtOH (200 mL) was
added
a solution of d-(+)-dibenzoyltartaric acid
(44.0 g, 0.123 mol) in EtOH (200 mL). The mixture was
cooled to 10 ꢁC, and the crystals (102.0 g) were ®ltered
o€. Three recrystallizations from 80% EtOH gave 57.6 g
of a d-(+)-dibenzoyltartrate salt with mp 165±166 ꢁC.
The absolute con®guration of this salt was determined
for 10c to be S by an X-ray crystallographic analysis.
The base was set free with NaOH yielding 28.2 g, which
was precipitated from EtOH with fumaric acid (17 g).
Yield: 41.8 g (0.128 mol, 26%). Mp 169±170 ꢁC. [a]d25:
imum peak height in a ®nal di€erence Fourier map is
3
0.38(5) eA
and this peak is without chemical sig-
ni®cance. The atomic coordinates for this structure have
been deposited with the Cambridge Crystallographic
Data Centre. The coordinates can be obtained on
request from the Director, Cambridge Crystallographic
Data Centre, 12 Union Road, Cambridge CB2 1EZ,
UK.
.
+40.7 (c=0.1, H2O). Anal. (C10H12N2OS C4H4O4) C,
H, N. Enantiomeric excess was estimated as described
for ( )-10d to be >98.8%. To the crude free base of
(R)-( )-10c, isolated from the ®ltrate from the ®rst
crystallization of (S)-(+)-10c d-(+)-dibenzoyltartrate
was added l-( )-dibenzoyltartaric acid monohydrate
(50 g), and the ( ) enantiomer isolated as the fumarate
salt as described above. Yield of (R)-10c fumarate:
35.1 g (0.108 mol, 22%). Mp 171±172 ꢁC. [a]d25: 40.4
Receptor radioligand binding
The anity of the compounds for muscarinic receptors
were estimated by their ability to displace [3H]Oxo-M
(0.20 nM) from whole rat brain homogenate, [3H]PZ
(1.0 nM) from whole rat brain homogenate, and
[3H]QNB (0.12 nM) from whole rat brain homogenate
and from rat brainstem homogenate, respectively.
Details have been described previously.43 Inhibitory
constants (Ki values) were estimated from IC50 values
.
(c=0.1, H2O). Anal. (C10H12N2OS C4H4O4) C, H, N.
Enantiomeric excess was estimated, as described for
( )-10d, to be >98.6%.
(RS)-5,7-Dimethyl-3-propargyloxy-4,5,6,7-tetrahydroiso-
thiazolo[4,5-c]pyridine (11c) oxalate. From 10c as
described for the preparation of 11d in 57% yield. Mp
44
using the Cheng±Pruso€ equation: Ki=IC50/(1+s/
KD), where s is the ®xed concentration and KD the dis-
sociation constant of the labeled ligand. The following
KD constants were derived from computer-assisted
Scatchard analyses of binding experiments: [3H]Oxo-M,
0.48‹0.03 nM; [3H]PZ, 1.8‹1.0 nM; [3H]QNB to
homogenate brain, 13.7‹0.9 pM; [3H]QNB to homo-
genate brain stem, 9.7‹0.8 pM. Two complete con-
centration±response curves were determined using ®ve
concentrations of test drug in triplicate (covering 3
135±140 ꢁC (from Me2CO); H NMR (DMSO-d6) 1.30
1
(d, 3H), 2.75 (dd, 1H), 2.80 (s, 3H), 3.25±3.70 (m, 3H),
3.65 (t, 1H), 3.75 (d, 1H), 4.05 (d, 1H), 5.05 (d, 2H)
.
Anal. (C11H14N2OS C2H2O4) C, H, N.
(R)-( )-5,7-Dimethyl-3-propargyloxy-4,5,6,7-tetrahydro-
isothiazolo[4,5-c]pyridine [(R)-11c] oxalate. From (R)-
10c as described for the preparation of 11d in 56% yield.

808
H. Pedersen et al. / Bioorg. Med. Chem. 7 (1999) 795±809
decades). IC50 values were estimated from handdrawn
log concentration±response curves.
for the Macintosh computer. Statistical signi®cance was
evaluated by t-test using the program StatView 1.0
(Abacus Concepts) for the Macintosh computer. p 0.05
was considered statistically signi®cant.
In a series of n determinations the variance of the log
ratio (VarR) between the double determinations was
determined according to: VarR=(1/2n)Æ(log Ri)2, where
Ri is the ith ratio and n is the number of observations.
The VarR is equivalent to the square of the standard
deviation of the log ratio (SDR2). The following stan-
dard deviations were obtained: [3H]Oxo-M, 1.5
(n=100); [3H]PZ, 1.6 (n=100); [3H]QNB, 1.5 (n=100).
Acknowledgements
This work was supported by grants from the Danish
Medical (PharmaBiotec Neuroscience Center) and
Technical Research Councils and the Lundbeck Foun-
dation. The secretarial assistance of Mrs. A. Nordly and
the technical assistance of Ms. M. Glarù are gratefully
acknowledged. Finally we wish to acknowledge Dr.
Bruce R. Conklin for providing the Gq-i5 construct.
R-SAT2 is patented by Acadia Pharmaceuticals Inc.,
which we acknowledge for allowing us to use this assay.
E€ects on cloned human mAChRs: receptor selection and
ampli®cation technology
R-SAT was performed as described earlier.23,30±32 NIH
3T3 cells (ATCC no. CRL 1658) were grown in a 37 ꢁC
humidi®ed CO2 atmosphere in Dulbecco's Modi®ed
Eagle's Media supplemented with 4500 mg/L glucose,
862 mg/L l-alanyl-l-glutamine, 50 units/mL penicillin
G, 50 units/mL streptomycin (Gibco, Paisley, Scotland)
and 10% calf serum (HyClone, Utah, USA).
References
1. Bartus, R. T.; Dean III, R. L.; Beer, B.; Lippa, A. S. Sci-
ence 1982, 217, 408.
2. Perry, E. Br. J. Psychiat. 1988, 152, 737.
3. Smith, C. J.; Perry, E. K.; Perry, R. H.; Candy, J. M.;
Johnson, M.; Bonham, J. R.; Dick, D. J.; Fairbairn, A.; Bles-
sed, G.; Birdsall, N. J. M. J. Neurochem. 1988, 50, 847.
4. Wurtman, R. J.; Corkin, S.; Growdon, J. H.; Ritter-
Walker, E., Eds. Advances in Neurology, Vol. 51. Alzheimer's
Disease; Raven Press: New York, 1990.
One day prior to transfection, cells were plated in a
density of 2Â106 cells per 15 cm of tissue culture dish.
Cells were transfected by calcium phosphate±DNA pre-
cipitation as described by Wigler et al.45 using 10 mg
receptor DNA,46,47 10 mg p-SV-û-galactosidase DNA
(Promega, WI, USA) and 40 mg salmon sperm DNA
(Sigma, MO, USA). Cells transfected with m2 or m4
receptor DNA were further cotransfected with 10 mg of
the chimeric G-protein, Gq-i5.48
5. Coyle, J. T.; Price, D. L.; DeLong, M. R. Science 1983, 219,
1184.
6. Bartus, R. T.; Flicker, C.; Dean III, R. L.; Fisher, S.; Pon-
tecorvo, M.; Figueiredo, J. Prog. Brain Res. 1986, 70, 345.
7. Perry, E. K.; Tomlinson, B. E.; Blessed, G.; Bergmann, K.;
Gibson, P. H.; Perry, R. H. Br. Med. J. 1978, ii, 1457.
8. Mash, D.; Flynn, D. D.; Potter, L. T. Science 1985, 228, 1115.
9. Aubert, I.; Araujo, D. M.; Cecyre, D.; Robitaille, Y.;
Gauthier, S.; Quirion, R. J. Neurochem. 1992, 58, 529.
10. Levey, A. I.; Kitt, C. A.; Simonds, W. F.; Price, D. L.;
Brann, M. R. J. Neurosci. 1991, 11, 3218.
Media was exchanged the day after transfection. Two
days after transfection, cells were spilt into two 96-well
plates, and ligands were added into a total volume of
200 mL/well.
11. Wall, S. J.; Yasuda, R. P.; Hory, F.; Flagg, S.; Martin, B.
M.; Ginns, E. I.; Wolfe, B. B. Mol. Pharmacol. 1991, 39, 643.
12. Wall, S. J.; Yasuda, R. P.; Li, M.; Ciesla, W.; Wolfe, B. B.
Dev. Brain Res. 1992, 66, 181.
13. Pomara, M.; Bagne, C. A.; Stanley, M.; Yarbrough, G. G.
Biol. Psychiat. 1986, 10, 553.
14. Mitchelson, F. Pharmacol. Ther. 1988, 37, 357.
15. Whitehouse, P. J. Acta Neurol. Scand. 1993, Suppl. 149, 42.
16. Hu, J.; Wang, S. Z.; el-Fakahany, E. E. J. Pharmacol.
Exp. Ther. 1991, 257, 938.
17. Pontzer, N. J.; Crews, F. T. J. Pharmacol. Exp. Ther.
1990, 253, 921.
18. Thompson, A. K.; Fisher, S. K. J. Pharmacol. Exp. Ther.
1990, 252, 744.
19. Sauerberg, P.; Larsen, J. J.; Falch, E.; Krogsgaard-Larsen,
P. J. Med. Chem. 1986, 29, 1004.
20. Sauerberg, P.; Fjalland, B.; Larsen, J. J.; Bach-Lauritsen,
T.; Falch, E.; Krogsgaard-Larsen, P. Eur. J. Pharmacol. 1986,
130, 125.
21. Krogsgaard-Larsen, P.; Falch, E.; Sauerberg, P.; Freed-
man, S. B.; Lembùl, H. L.; Meier, E. In Subtypes of Muscari-
nic Receptors III; Levine, R. R.; Birdsall, N. J. M.; North, R.
A.; Holman, M.; Watanabe, A.; Iversen, L. L., Eds.; Elsevier:
Amsterdam, 1988; pp 69±74.
Agonist data from R-SAT experiments were ®tted to the
four parameter equation:49,50
D ‡ ꢀA D†
R ˆ
b
1 ‡ ꢀx=c†
For agonists: A=minimum response, D=maximum
response and c=EC50; for antagonists: A=maximum
response, D=minimum response and c=IC50
(R=response, x=ligand concentration and b=slope at
EC50 or IC50). The b value was set to unity.
Ki values were calculated from IC50 values by the
44
method of Cheng and Pruso€
McKinney et al.51 and Craig.52
as described by
Antagonists that were unable to fully inhibit a carba-
chol response were subjected to Schild analysis.34 Schild
analysis is invalid if the maximal agonist response is
suppressed by the addition of antagonist, which is a sign
of noncompetitive antagonism. Noncompetitive
antagonists were analyzed as described by Kenakin.53
22. Sauerberg, P.; Falch, E.; Meier, E.; Lembùl, H. L.;
Krogsgaard-Larsen, P. J. Med. Chem. 1988, 31, 1312.
23. Brauner-Osborne, H.; Ebert, B.; Brann, M. R.; Falch, E.;
Krogsgaard-Larsen P. J. Med. Chem. 1995, 38, 2188.
Curves were generated by nonweighted least-squares ®ts
using the program KaleidaGraph 2.1 (Abelbeck Software)

H. Pedersen et al. / Bioorg. Med. Chem. 7 (1999) 795±809
809
24. Krogsgaard-Larsen, P.; Mikkelsen, H.; Jacobsen, P. et al.
J. Med. Chem. 1983, 26, 895.
39. Lagersted, A.; Falch, E.; Ebert, B.; Krogsgaard-Larsen, P.
Drug Design & Discovery 1993, 9, 237.
25. Meyers, A. I.; Tomioka, K.; Roland, D. M.; Comins, D.
Tetrahedron Lett. 1978, 1375.
26. Dickey, J. B. U. S. Patent 2 659 739, 1953.
27. Johnson, C. K. ORTEPII. ORNL-5138; Oak Ridge
National Laboratory: Tennessee, 1976.
28. Freedman, S. B.; Beer, M. S.; Harley, E. A. Eur. J. Phar-
macol. 1988, 156, 133.
29. Moltzen, E. K.; Pedersen, H.; Bùgesù, K. P.; Meier, E.;
Frederiksen, K.; Sanchez, C.; Lembùl, H. H. J. Med. Chem.
1994, 37, 4085.
40. Sheldrick, G. M. SHELXS-86. Program for Crystal
Structure Solution; University of Gottingen: Germany, 1986.
41. Sheldrick, G. M. Acta Crystallogr. 1990, A46, 467.
42. Frenz, B. A. and Associates Inc. SDP Structure Determi-
nation Package; College Station and Enraf-Nonius: TX, and
Delft: The Netherlands, 1985.
43. Arnt, J.; Lembùl, H. L.; Meier, E.; Pedersen, H. Eur. J.
Pharm. 1992, 218, 159.
44. Cheng, K.-C.; Pruso€, W. H. Biochem. Pharmacol. 1973,
22, 3099.
30. Brauner-Osborne, H.; Brann, M. R. Eur. J. Pharmacol.
1996, 295, 93.
45. Wigler, M.; Silverstein, S.; Lee, L.-S.; Pellicer, A.; Cheng,
Y.-C.; Axel, R. Cell 1977, 11, 223.
31. Burstein, E. S.; Spalding, T. A.; Hill-Eubanks, D.; Brann,
M. R. J. Biol. Chem. 1995, 270, 3141.
46. Bonner, T. I.; Buckley, N. J.; Young, A. C.; Brann, M. R.
Science 1987, 237, 527.
32. Burstein, E. S.; Spalding, T. A.; Brauner-Osborne, H.;
Brann, M. R. FEBS Lett. 1995, 363, 261.
47. Bonner, T. I.; Young, A. C.; Brann, M. R.; Buckley, N. J.
Neuron 1988, 1, 403.
33. Brauner-Osborne, H.; Ebert, B.; Brann, M. R.; Falch, E.;
Krogsgaard-Larsen, P. Eur. J. Pharmacol. 1996, 313, 145.
34. Arunlakshana, O.; Schild, H. O. Br. J. Pharmacol. Che-
mother. 1959, 14, 48.
35. Messier, T. L.; Dorman, C. M.; Brauner-Osborne, H.;
Eubanks, D.; Brann, M. R. Pharmacol. Toxicol. 1995, 76, 308.
36. Christie, J. E.; Shering, A.; Ferguson, J.; Glen, A. I. M.
Br. J . Psychiat. 1981, 138, 46.
37. Sitaram, N.; Weingartner, H.; Gillin, J. C. Science 1978,
201, 274.
38. Bartus, R. T.; Dean, R. L.; Beer, B. Neurobiol. Aging
1980, 1, 145.
48. Conklin, B. R.; Farfel, Z.; Lustig, K. D.; Julius, D.;
Bourne, H. R. Nature 1993, 363, 274.
49. Healy, M. J. R. Biochem. J. 1972, 130, 207.
50. Rodbard, D.; Hutt, D. M. In Radioimmunoassay and
Related Procedures in Medicine; International Atomic Energy
Agency: Vienna 1974; Vol. 1, pp 165±192.
51. McKinney, M.; Anderson, D. J.; Vella-Rountree, L.;
Connolly, T.; Miller, J. H. J. Pharmacol. Exp. Ther. 1991, 257,
1121.
52. Craig, D. A. Trends Pharmacol. Sci. 1993, 14, 89.
53. Kenakin, T. Pharmacological Analysis of Drug-Receptor
Interaction; 2nd ed., Raven Press: New York, 1993.