Rigidified acetylcholine mimics: Conformational requirements for binding to neuronal nicotinic receptors

Article abstract of DOI:10.1016/S0960-894X(03)00728-5

Rigidified derivatives have been designed and synthesized assuming the g+t conformer of acetylcholine (N-C-C-O=+60°, C-C-O-C=180°) as active conformation for binding to cytisine sensitive neuronal nicotinic receptors. The SAR of the compounds evaluated, along with those of more flexible analogues, support the g+t conformer hypothesis and highlight the stringent steric limitation of this nicotinic receptor sub-type. Compound 3e has low μM affinity for cytisine sensitive nicotinic receptor binding sites while being selective with regard to the α-bungarotoxin sensitive subclass. We also report few compounds with μM affinity for the α-bungarotoxin sensitive subclass.

Full text of DOI:10.1016/S0960-894X(03)00728-5

Bioorganic & Medicinal Chemistry Letters 13 (2003) 3847–3851
Rigidified Acetylcholine Mimics: Conformational Requirements
for Binding to Neuronal Nicotinic Receptors
Gerald Villeneuve,a,* Danielle Cecyre,b Helene Lejeune,b Marc Drouin,a Ruoxi Lana
and Remi Quirionb
a
´
Departement de Chimie, Universite de Sherbrooke, 2500 boul. de l’Universite, Sherbrooke, Quebec, Canada J1K 2R1
Centre de Recherche de l’Hoˆpital Douglas, McGill University, 6857 boul. Lasalle, Verdun, Quebec, Canada H4H 1R3
´
´
´
b
´
Received 2 April 2003; revised 8 July 2003; accepted 12 July 2003
Abstract—Rigidified derivatives have been designed and synthesized assuming the g+t conformer of acetylcholine (N–C–C–
O=+60ꢀ, C–C–O–C=180ꢀ) as active conformation for binding to cytisine sensitive neuronal nicotinic receptors. The SAR of the
compounds evaluated, along with those of more flexible analogues, support the g+t conformer hypothesis and highlight the strin-
gent steric limitation of this nicotinic receptor sub-type. Compound 3e has low mM affinity for cytisine sensitive nicotinic receptor
binding sites while being selective with regard to the a-bungarotoxin sensitive subclass. We also report few compounds with mM
affinity for the a-bungarotoxin sensitive subclass.
# 2003 Elsevier Ltd. All rights reserved.
Acetylcholine (Ach), the first neurotransmitter to have
been characterized, acts on several classes and sub-
classes of receptors. The nicotinic receptors have been
demonstrated to be ligand-gated ion channel while
muscarinic receptors belong to the superfamily of G-
proteins link receptors. Receptor cloning revealed the
existence of five different muscarinic receptors.1 The
case of neuronal nicotinic receptors is far more complex.
Most nicotinic receptors are believed to be constituted
by the non-covalent association of five proteins sub-
units. Upto now, eight neuronal a subunits have been
cloned (a2–a9) along with three non-a subunits (b2–
b4).2 Although the number of possible combinations is
enormous, it seems that some may predominate. The
pharmacological receptor responsible for a-bungar-
otoxin (a-BGT) binding in the brain is probably mainly
composed of five a7 subunits.3 On the other hand, the
receptors having high affinity for both cytisine (Cyt) and
nicotine (Nic) mostly involve a combination of a4 and b2.4
level in the brain. Tacrine and Donepezil are homo-
logated substances for Alzheimer’s disease treatment
which mainly act via this mechanism.6 Efforts should
also be given to developselective brain nicotinic ago-
nists that could stimulate for example, presynaptic
positive nicotinic autoreceptors.7 Our approach is rela-
ted to that goal and to better understand the steric and
electronic requirements needed to ensure high affinity
and selectivity for neuronal nicotinic receptors. We pre-
sent herein the rationale design, synthesis and receptor
binding affinity of these compounds.
Beer and Reich have proposed that the onnium head
and the carbonyl oxygen of Ach interact with the nico-
tinic receptor.8 Cyt can serve as an unambiguous phar-
macophore template. Using the amide carbonyl dipole
and the secondary amine of Cyt to define a three points
pharmacophore, a relatively good alignment could be
achieved with (+) anatoxin (Ana).9 For Nic and (ꢁ)
epibatidine (Epi),10 the ring center to the pyridine
nitrogen was defined as the corresponding dipole. Ach
can adapt well to that pharmacophore if it adopt the
g+t conformation with t2 (N–C–C–O)=+60ꢀ, t1 (C–
C–O–C)=180ꢀ and t0 (CH2–O–C–O)=0ꢀ. The aligned
molecules are presented in Figure 1.
Nicotinic and muscarinic receptors have been shown to
be the most affected in diseases associated with severe
memory impairment such as in Alzheimer’s disease.5 An
avenue explored was to develop specific acetylcholine
esterase inhibitors in order to increase acetylcholine
Accordingly we attempted to design rigidified analogues
of the g+t conformer of acetylcholine. Our rationale is
presented in Figure 2. Firstly, a cyclic ketone will be
*Corresponding author. Tel.: +819-846-1720; fax: +819-821-8070;
e-mail: geraldvilleneuve@yahoo.ca
0960-894X/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0960-894X(03)00728-5

3848
G. Villeneuve et al. / Bioorg. Med. Chem. Lett. 13 (2003) 3847–3851
Figure 1. Aligned nicotinic receptor ligands according to Beer and
Reich pharmacophore hypothesis. From top, left to right: Nic, Cyt
Epi, Ach Ana and overlay of Ach with Cyt.
Scheme 1. Synthesis of rigidified Ach analogues: (a) C2O2Cl2, DMSO,
DCM, ꢁ50 ꢀC 30 min, then TEA: (b) LDA/THF ꢁ78 ꢀC with ketones
2, 7, 10, 15 or 21 2 h, then 1 (1 equiv) 1–2 h; (c) CH3I, MeOH, rt, 12 h;
(d) PTSA 2 equiv, CH3CN, 20 h; (e) Na (2 equiv) in liquid NH3.
related N-alkyl-pyrrolidine carboxaldehydes.13 Further-
more, the aldol condensation did not induce significant
epimerization since obtained products had high optical
rotations and the diastereomeric ratio for the reaction
of 1a with (+) camphor was 90% (as assessed by
GCMS and 13C NMR). Products were Kugelrhor dis-
tilled such that GC-MS provided purity and identity of
compounds.14
Figure 2. The molecular design of rigidified g+t Ach mimics.
used to mimic the ester bond. Then an E double bond
will be used to lock t1 at 180ꢀ (t). Finally the incor-
poration of the amine within a ring would create
restrain rotation around t2 such that the favored dihe-
dral angle will be close to +60ꢀ (g+) for the S con-
figuration at C2.
More flexible analogues and surrogates were prepared
for the sake of comparison and to try to delineate the
active portion of (ꢁ)lobeline (Lob).15 Their preparations
are presented in Scheme 2. Firstly, the corresponding
The preparation of rigidified acetylcholine derivatives is
shown in Scheme 1. The method relied on the aldol
condensation of the lithium enolate of a ketone with N-
methyl-pyrrolidine carboxaldehyde (1a) or N-methyl-
piperidine carboxaldehyde (1b) in THF at ꢁ78 ꢀC. The
condensation occurred usually with subsequent dehy-
dration to provide mainly the desired E enone with few
Z enone. The E configuration was confirmed by X-ray
crystallography for compound 9 which by the same
token provided the observed t2 value of +117ꢀ.11 The Z
enone was always more polar than the E enone and
purification was possible by flash or layer chromato-
1
graphy. In H NMR, the chemical shift of the E olefin
proton was always more upfield. In one case, the dial-
kylated ketone 24 was obtained but owing to its much
higher polarity, it could be easily separated. Starting
material 1a–d were obtained from the corresponding
commercial alcohols using Swern oxidation technique.12
Measured [a]2D0 was consistent to values reported for
Scheme 2. Reagents and conditions: (a) PPh3, CCl4, CH3CN, rt to
reflux; (b) KF/Al2O3, CH3CN, rt, 24 h; (c) CH3I, MeOH, rt, 12 h; (d)
27, 1 equiv nBuLi, THF or Et2O, TMEDA, ꢁ78 ꢀC, 30 min, then
26b,c; (e) 1 equiv HCl, C6H5I(CF3CO2)2, MeOH, H2O, 1 h.

G. Villeneuve et al. / Bioorg. Med. Chem. Lett. 13 (2003) 3847–3851
3849
Table 1. Radioligand binding results at neuronal nicotinic receptors for enone compounds
Compd
m
n
R
Geom.
Ster.
[3H]Cytisine IC50 (mM)a
[
125I]a-BGT IC50 (mM)b
Epi
Ana
Lob
3a
3b
3c
3d
3e
4d
4e
5a
5b
5c
5d
5e
6e
8a
8b
0.0008 (ꢂ.0001)
0.0045 (ꢂ.001)
0.087 (ꢂ0.026)
0.48 (ꢂ0.16)
0.93 (ꢂ0.43)
11.3 (ꢂ2.6)
23.7 (ꢂ8.2)
0.46 (ꢂ0.16)
207 (ꢂ145)
5.46 (ꢂ3.5)
0.12 (ꢂ0.06)
0.84 (ꢂ0.10)
1.5 (ꢂ0.24)
6.5 (ꢂ1.9)
0.75 (ꢂ0.33)
20.5 (ꢂ7.8)
12.5 (ꢂ4.0)
>100
0.0037
0.004
18.1
2.8
44
nd
nd
1
1
2
2
1
2
1
1
1
2
2
1
1
1
2
2
2
2
2
1
1
2
2
1
2
2
2
2
1
1
1
2
1
2
2
2
2
1
2
1
2
2
2
H
H
H
H
Me
H
Me
H
H
H
H
E
E
E
E
E
Z
Z
E
E
E
E
E
Z
E
E
E
E
Z
E
Z
E
E
E
E
E
S
S
RS
RS
S
RS
S
S
31
nd
111
0.16
0.15
134
207
43
0.21
nd
nd
nd
S
RS
RS
S
S
S
RS
RS
RS
RS
RS
S
Me
Me
9
11
12
13
>100
>100
>100
37
81
5.4
2.4
136 (ꢂ79)
110 (ꢂ33)
35.9(ꢂ16)
52.5 (ꢂ31)
12.2 (ꢂ3.0)
>100
57.7 (ꢂ47)
32.3 (ꢂ23)
194 (ꢂ104)
68.6 (ꢂ8.4)
207 (ꢂ130)
89 (ꢂ15)
14
16a
16b
17b
19a
19b
20
22
23
24
25
S
nd
RS
RS
S
RS
RS
RS
RS
S
115
25.3
436
>100
91
E
E
E
E
nd
nd
nd
nd
2
2
—
—
S
aValues are means of two to three independent experiments using triplicate, standard deviation is given in parentheses.
bValues from a single triplicate experiment (nd, not determined).
saturated ketones were prepared using a 1–3 dithiane
anion16 which was shown to react well with the acti-
vated chloride 26b,c. The resulting alkylated 1–3
dithianes (32) were converted to the ketones using the
protocol described by Stork and Zhao.17 On the other
hand, isosteric amide compounds were prepared by
reacting g or d lactams with activated chlorides 26a,b
under KF catalyzed conditions.18 Owing to the forma-
tion of aziridinum intermediate, the reaction gave also
the ring expanded compounds 30. Normal products and
ring expanded products were separated by silica gel
flash or layer chromatography.
parallel affinity was found for corresponding quaternary
salts 5a, 5b, and 5e. When the selectivity over the a-
BGT-sensitive Ach receptor subclass was considered,
compound 3e cumulates the best affinity/selectivity
characteristics. Although compound 5e has also a good
profile, its quaternary salt nature makes it less interest-
ing in consideration of blood–brain barrier penetration.
In all cases within the enone series, the pyrrolidine
compounds (m=1) were 10–20 times more potent than
the corresponding piperidines (m=2) (3a vs 3c, 3b vs 3d,
Table 2. Radioligand binding results at neuronal nicotinic receptors
for lactam compounds
All pure compounds were characterized by 1H, 13C NMR
and GCMS. Free bases were converted to hydrochloride
salts with anhydrous ethereal HCl solution and stored
in a moisture free environment. Quaternary salts were
generally purified by precipitating them out. All hydro-
chlorides and quaternary salts were water soluble. The
compounds were evaluated for their abilities to compete
for specific [3H]cytisine (a4b2 subclass) and [125I]a-bun-
garotoxin (a7 subclass) binding sites in rat forebrain
homogenates.19
Compd
m
n
R1
Ster.
[3H] Cytisine
IC50 (mM)a
[
125I] a-BGT
IC50 (mM)b
28a
28b
28c
28d
29a
29b
29c
29d
30
1
1
2
2
1
1
2
2
—
—
1
2
1
2
1
2
1
2
H
H
H
S
S
RS
RS
S
20.1 (ꢂ0.9)
>100
49.5
>100
2.72
>100
12.2
>100
2.87
>100
>100
87
26.3 (ꢂ1.6)
H
>100
93 (ꢂ18)
>100
Me
Me
Me
Me
H
S
RS
RS
RS
RS
>100
>100
78 (ꢂ18)
31
Me
>100
Binding affinity for the cytisine sensitive neuronal Ach
receptors was obtained in the low mM range for com-
pounds within the enone series (Table 1). Tertiary amine
compounds 3a, 3b and 3e were the best ligands and
aValues are means of two to three independent experiments using
triplicate, standard deviation is given in parentheses.
bValues from a single triplicate experiment (nd, not determined).

3850
G. Villeneuve et al. / Bioorg. Med. Chem. Lett. 13 (2003) 3847–3851
Table 3. Radioligand binding results at neuronal nicotinic receptors for dithiane and ketone compounds
Compd
series
x
R3
[3H]Cytisine
IC50 (mM) 32a
[3H]Cytisine
IC50 (mM) 33a
[3H]Cytisine
IC50 (mM) 34a
b
a
b
c
d
e
f
g
h
i
1
1
1
1
1
1
2
2
2
2
Ph
>100
>100
>100
>100
>100
>100
5.99 (ꢂ1.9)
>100
—
—
>100
—
—
b
2-Furyl
pCl-Ph
2-Pyridyl
2-Thienyl
3-Pyridyl
pCl-Ph
2-Thienyl
Me
b
>100
10.4 (ꢂ4.6)
b
b
b
53.7 (ꢂ23)
42.6 (ꢂ2.1)
59.4 (ꢂ19.2)
>100
—
>100
—
62.6 (ꢂ29.8)
29.0 (ꢂ6.8)
1.80 (ꢂ0.53)c
>100
24.6 (ꢂ0.2)
c
3.7 (ꢂ1.6)
j
p-F-Ph
59.4 (ꢂ31.9)
23.4 (ꢂ9.5)
aValues are means of two to three independent experiments using triplicate, standard deviation is given in parentheses.
bCompound was not prepared.
cBinding for [125I]a-BGT were 33i: 15.0 mM and 34i: 0.17 mM.
5a vs 5c, 8a vs 8b) which supersed the sole effect of the
racemic nature of the piperidine compounds and argue
for a ring size effect similar to that observed for nicotine
versus anabasine.20 Having large substituents nearby
the ketone function (compounds 8–10 and 15–20) or a
large substituent within the area that could interact with
the receptor carbonyl binding subsite (compounds 11–14)
induces significant loss in affinity. This could suggest
that results obtained with compounds 22 and 23 are
inconsistent (22 is not larger than 3e at first glance).
However, owing to the non-cyclic nature of 22 and 23,
the enone can assume either the s-Z (t0=0ꢀ) or the s-E
(t0=180ꢀ) conformation. MM2 calculations suggest
that the s-E conformer is 2 Kcal/mol more stable than
the s-Z. The s-E conformer will be largely predominant
and since the later does not permit to achieve the pos-
tulated pharmacophore topology, this may explain the
low affinity of 22 and 23. Finally, when the effect of the
stereochemistry of the double bond was studied alone,
the E compounds were generally 10 times more potent
than the corresponding Z compounds (3d vs 4d, 3e vs
4e, 5e vs 6e) which give support to our molecular design.
Interestingly, the Z compounds 6e and 14 are fairly
potent on a-BGT binding sites with good selectivity
over cytisine sensitive sites.
mylcholine. None of the compound was close to Lob
affinity (Table 1), although 33b and 33e showed some
tendency. Compounds bearing an aryl ketone with
X=2 had moderate affinity but were clearly less good
ligands than the methyl ketones 33i and 34i. Compar-
ison of the affinity of 33i with that of 3a supports the
beneficial effect of an appropriate locking of the con-
formation, even if some steric bulk is created under such
conditions.
Acknowledgements
This work was supported by a conjoint NSERC Nordic-
Merrell-Dow Canadian grant. We would like to thank
Dr. Andre Michel for initiating this project and Pr.
Pierre Deslongchamps for taking after. Mr. Gaston
Boulay is acknowledged for GC–MS and exact mass.
References and Notes
1. Hulme, E. C.; Birdsall, N. J. M.; Buckley, N. J. Annu. Rev.
Pharmacol. Toxicol. 1990, 30, 633.
2. Sargent, P. Annu. Rev. Neurosci. 1993, 16, 403.
3. Seguela, P.; Wadiche, J.; Dineley-Miller, K.; Dani, J. A.;
Patrick, J. W. Eur. J. Neurosci. 1993, 13, 596.
Within the lactam series (Table 2), affinity for the cyti-
sine sensitive subclass was much lower than for the
corresponding enone compounds. In principle the lac-
tam series is isosteric to the enone series but the con-
formation is not as well controlled. In fact, X-ray
crystallography of compound 29c11 showed that the
later adopts a conformation where corresponding t2
and t1 dihedral angles take the values of ꢁ161ꢀ and
ꢁ112ꢀ, respectively; at departure from the putative
pharmacophore. Interestingly, this conformation is
apparently more appropriate for a-BGT sites since sig-
nificant affinity was found for compounds 29c and rela-
ted 28c. The ring size of the lactam ring (n=1) in that
series was rather critical.
4. Flores, C. M.; Rogers, S. W.; Pabreza, L. A.; Wolfe, B. B.;
Kellar, K. J. Mol. Pharmacol. 1992, 41, 31.
5. Coyle, J. T.; Price, D. L.; DeLong, M. Science 1983, 219,
1184.
6. Snape, M. F.; Misra, A.; Murray, T. K.; DeSouza, R. J.;
Williams, J. L.; Cross, A. J.; Green, A. R. Neuropharmacology
1999, 38, 181.
7. Araujo, D. M.; Lapchak, P. A.; Meaney, M. J.; Collier, B.;
Quirion, R. J. Neurosci. 1990, 10, 3069.
8. Beers, W. H.; Reich, E. Nature 1970, 228, 917.
9. Thomas, P.; Stephens, M.; Wilkie, G.; Amar, M.; Lunt,
G. G.; Whiting, P.; Gallagher, T.; Pereira, E.; Alkondon, M.;
Albuquerque, E. X. J. Neurochem. 1993, 60, 2308.
10. Quian, C.; Li, T.; Shen, T. Y.; Libertin-Garahan, L.;
Eckman, J.; Biftu, T.; Ip, S. Eur. J. Pharm. 1993, 250, R13.
11. Crystallographic data for structures 9 and 29c have been
deposited to the Cambridge Crystallographic Data Centre as
supplementary publication numbers CCDC 205925 and
205926. Copies of the data can be obtained, free of charge, on
In the ketone series (Table 3), compounds with X=1
were made mainly to try to identify the active portion of
Lob whereas with X=2 they mimic more Ach or carba-

G. Villeneuve et al. / Bioorg. Med. Chem. Lett. 13 (2003) 3847–3851
3851
application to CCDC, 12 Union Road, Cambridge CB2 1EZ,
UK (fax: +44-1223-336033 or e-mail: deposit@ccdc.
cam.ac.uk).
freshly prepared by homogenization in 50 mM Tris buffer, pH
7.4 at 4 ꢀC using a Brinkmann polytron (setting 6, 20 s). The
homogenate was centrifuged for 10 min at 49,000g, the super-
natant was discarded and the pellet washed twice by resus-
pension in fresh buffer and re-centrifugation. The final pellet
was resuspended in incubation buffer, consisting of 50 mM
Tris containing 1 mM MgCl2, 120 mM NaCl, 5 mM KCl and
2 mM CaCl2 (pH 7.4), to obtain concentration of 1.0–3.0 mg
protein/mL. Aliquots (0.2 mL) were added to assay tubes
containing analogue (4–8 concentrations, 10 pM–0.1 mM) and
2 nM [3H]Cytisine or 1.5 nM [125I]a-BGT for a final assay
volume of 0.6 mL. Incubation was performed for 90 min at
4 ꢀC for [3H]Cytisine and for 120 min at 22 ꢀC with BSA (1
mg/mL) for [125I]a-BGT. 10 and 100 mM of Nic were used
respectively to define the non-specific binding. Separation of
bound from free radioactive ligand was performed by rapid
filtration through #32 glass fiber filters (presoaked in 0.1%
polyethyleneimine solution) under reduced pressure using a
Brandell Cell Harvester followed by 3ꢃ4 mL rapid washes
with 50 mM Tris buffer, pH 7.4 at 4 ꢀC. Bound radioactivity
was determined via liquid scintillation.
12. Mancuso, A. J.; Huang, S. L.; Swern, D. J. Org. Chem.
1978, 43, 2480.
13. 1a [a]2D0=ꢁ130ꢀ (c 7.0, CH2Cl2): Linkert, F.; Laschat, S.;
Kotila, S.; Fox, T. Tetrahedron 1995, 52, 955.
1
14. Selected data: 3d: H NMR (CDCl3, 300 MHz) d 6.50 (dt,
J=9.5, 3.8 Hz, 1H, C¼CH), 2.86 (dtd, J=11.2, 3.1, 1.5 Hz,
1H, NCHeqHax), 2.55–2.68 (m, 2H), 2.32–2.50 (m, 3H), 2.11
(s, 3H), 1.93–2.02 (m, 1H), 1.78–1.90 (m, 2H), 1.65–1.78 (m,
3H), 1.56–1.65 (m, 2H), 1.15–1.55 (m, 3H) ; 13C NMR
(CDCl3, 75 MHz) d 200.7, 140.7 (CH), 136.3, 61.9, 56.0, 44.4,
40.2, 30.8, 27.1, 25.6, 23.5, 23.4, 23.2; m/z (GC–MS) 207
(M+), 192 (M+ꢁMe), 98.
15. Barlow, R. B.; Johnson, O. Br. J. Pharmacol. 1989, 98,
799.
16. Corey, E. J.; Seebach, D. Angew. Chem., Int. Ed. Engl
1965, 4, 1075.
17. Stork, G.; Zhao, K. Tetrahedron Lett. 1989, 30, 287
(addition of HCl prevent formation of amine-oxide).
18. Yamawaki, J.; Ando, T. Chem. Lett. 1989, 755.
19. Membrane-enriched homogenates from rat forebrain were
20. Reavill, C.; Jenner, P.; Kumar, R.; Stolerman, I. P.
Neuropharmacology 1988, 27, 235.