Novel 5-(quinuclidin-3-ylmethyl)-1,2,4-oxadiazoles to investigate the activation of the α7 nicotinic acetylcholine receptor subtype: Synthesis and electrophysiological evaluation

Article abstract of DOI:10.1016/j.ejmech.2018.10.015

α7 nicotinic acetylcholine receptors (nAChRs) are relevant therapeutic targets for a variety of disorders including neurodegeneration, cognitive impairment, and inflammation. Although traditionally identified as an ionotropic receptor, the α7 subtype showed metabotropic-like functions, mainly linked to the modulation of immune responses. In the present work, we investigated the structure-activity relationships in a set of novel α7 ligands incorporating the 5-(quinuclidin-3-ylmethyl)-1,2,4-oxadiazole scaffold, i.e. derivatives 21a-34a and 21b-34b, aiming to identify the structural requirements able to preferentially trigger one of the two activation modes of this receptor subtype. The new compounds were characterized as partial and silent α7 nAChR agonists in electrophysiological assays, which allowed to assess the contribution of the different groups towards the final pharmacological profile. Overall, modifications of the selected structural backbone mainly afforded partial agonists, among them tertiary bases 27a-33a, whereas additional hydrogen-bond acceptor groups in permanently charged ligands, such as 29b and 31b, favored a silent desensitizing profile at the α7 nAChR.

Full text of DOI:10.1016/j.ejmech.2018.10.015

Accepted Manuscript
Novel 5-(quinuclidin-3-ylmethyl)-1,2,4-oxadiazoles to investigate the activation of
the α7 nicotinic acetylcholine receptor subtype: Synthesis and electrophysiological
evaluation
Marta Quadri, Almin Silnović, Carlo Matera, Nicole A. Horenstein, Clare Stokes,
Marco De Amici, Roger L. Papke, Clelia Dallanoce
PII:
S0223-5234(18)30877-8
10.1016/j.ejmech.2018.10.015
EJMECH 10801
DOI:
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 2 August 2018
Revised Date: 28 September 2018
Accepted Date: 8 October 2018
Please cite this article as: M. Quadri, A. Silnović, C. Matera, N.A. Horenstein, C. Stokes, M. De
Amici, R.L. Papke, C. Dallanoce, Novel 5-(quinuclidin-3-ylmethyl)-1,2,4-oxadiazoles to investigate
the activation of the α7 nicotinic acetylcholine receptor subtype: Synthesis and electrophysiological
evaluation, European Journal of Medicinal Chemistry (2018), doi: https://doi.org/10.1016/
j.ejmech.2018.10.015.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIPT
Novel 5-(quinuclidin-3-ylmethyl)-1,2,4-oxadiazoles to investigate the
activation of the α7 nicotinic acetylcholine receptor subtype: synthesis
and electrophysiological evaluation
a, b
a
a
c
Marta Quadri , Almin Silnović , Carlo Matera , Nicole A. Horenstein , Clare
b
a
b
a,
Stokes , Marco De Amici , Roger L. Papke , Clelia Dallanoce *
a
Department of Pharmaceutical Sciences, Medicinal Chemistry Section “Pietro Pratesi”,
University of Milan, Via L. Mangiagalli 25, 20133 Milan, Italy
b
Department of Pharmacology and Therapeutics, University of Florida, P.O. Box 100267,
Gainesville, FL 32610-0267, USA
c
Department of Chemistry, University of Florida, P.O. Box 117200, Gainesville, FL 32611-
7
200, USA
Keywords:
α7 nicotinic acetylcholine receptor
Electrophysiological studies
Partial agonist
Silent agonist
1
,2,4-Oxadiazoles
Quinuclidine
Abbreviations: nicotinic acetylcholine receptor, nAChR; half maximal effective concentration, EC ;
50
positive allosteric modulator, PAM; PNU-120596, PNU; PAM-sensitive nonconducting desensitized
state, D ; PAM-insensitive nonconducting desensitized state, D .
s
i
*
Corresponding author.
E-mail address: clelia.dallanoce@unimi.it (C. Dallanoce)
1

ACCEPTED MANUSCRIPT
A B S T R A C T
α7 nicotinic acetylcholine receptors (nAChRs) are relevant therapeutic targets for a variety of
disorders including neurodegeneration, cognitive impairment, and inflammation. Although
traditionally identified as an ionotropic receptor, the α7 subtype showed metabotropic-like
functions, mainly linked to the modulation of immune responses. In the present work, we
investigated the structure-activity relationships in a set of novel α7 ligands incorporating the 5-
(quinuclidin-3-ylmethyl)-1,2,4-oxadiazole scaffold, i.e. derivatives 21a-34a and 21b-34b,
aiming to identify the structural requirements able to preferentially trigger one of the two
activation modes of this receptor subtype. The new compounds were characterized as partial and
silent α7 nAChR agonists in electrophysiological assays, which allowed to assess the
contribution of the different groups towards the final pharmacological profile. Overall,
modifications of the selected structural backbone mainly afforded partial agonists, among them
tertiary bases 27a-33a, whereas additional hydrogen-bond acceptor groups in permanently
charged ligands, such as 29b and 31b, favored a silent desensitizing profile at the α7 nAChR.
2

1
. Introduction
ACCEPTED MANUSCRIPT
The α7 nicotinic acetylcholine receptor (nAChR) is characterized by low probability of
channel opening, rapid concentration-dependent desensitization, and high calcium permeability,
which differentiates it from other nicotinic receptor subtypes [1−3]. Besides being widely
distributed in the central and peripheral nervous systems, the α7 nAChR is present in non-
neuronal cells including immune cells, adipocytes, and intestinal and lung cells [4]. Investigation
of this receptor subtype evidenced its crucial roles in neuroprotection [5] and neurogenesis [6],
cognitive processes [7], and modulation of inflammatory responses [8,9]. Agonists and partial
agonists of the α7 receptor displayed favorable profiles for beneficial treatment of disorders with
neurological, neurodegenerative, and neuroinflammatory components [10]. A number of α7 full
and partial agonists including GTS-21 (1) [11,12], BMS-933043 (2) [13], TC-5619 (3) [14−16],
PNU-282987 (4) [17], EVP-6124 (5) [18], MEM3454 (6) [19], and AR-R17779 (7) [20] have
moved into clinical trials (Fig. 1A). We designed and tested spiroisoxazoline (R)-ICH-3 (8) (Fig.
1
B), which behaved as a potent and selective α7 agonist and was able to reduce pain and protect
nervous tissue in animal models of oxaliplatin-dependent neuropathy. Preclinical eADMET
studies and preliminary pharmacokinetic assays indicated the promising drug-like profile of 8 in
view of its potential application to CNS disorders [21−23]. Moreover, other scaffolds such as the
quinuclidinyl-3(1,2,3-triazole) nucleus (Fig. 1A) provided nicotinic ligands (9-13) characterized
by nanomolar binding affinity for the α7 receptor subtype and promising activity in cognitive
impairment models [24−26].
A growing number of literature reports support ionotropic-independent functions of the α7
nAChRs and their involvement in the regulation of inflammatory and immune responses [27].
First evidenced in various non-neuronal cells including microglia and macrophages, these
metabotropic-like mechanisms have been further clarified by recent studies on direct coupling of
the α7 nAChR subtype to G proteins, both G and G , through the M3-M4 loop [28−32]. In
α
βγ
addition, α7 ligands able to selectively induce non-ionotropic activation of the receptor by
promoting and stabilizing its desensitized conformations showed interesting anti-inflammatory
properties and were proposed as promising therapeutic agents [33−35]. Identified as α7 silent
agonists, these compounds show very little partial agonism on their own and favor the positive
allosteric modulator (PAM)-sensitive nonconducting D state. Indeed, the desensitizing
s
properties of an α7 silent agonist are revealed by co-application of a type II PAM, which evokes
channel current activation [36,37]. Different silent compounds (Fig. 1B) such as NS-6740 (14)
[
38], KC-1 (15) [36], ASM-024 (16) [39], diEPP (17) [40], p-CF -diEPP (18) [41], the sulfonium
3
salt 19 [42], and the spirocyclic quinuclidine derivatives 20a and 20b [43] were identified, and
some of them proved to be effective in in vivo anti-inflammatory chronic pain models [44,45].
3

The study of ligands selective for the α7 receptor evidenced the presence of common
ACCEPTED MANUSCRIPT
pharmacophoric elements in the structures of most full-partial [20,21,25,46−49] as well as silent
agonists [36,38−43], including a basic amine, a central group with hydrogen-bond acceptor
properties and a lipophilic aryl moiety extending from the amine (Fig. 1A and B).
Fig. 1. Structures of selected α7 nAChR full and partial agonists 1-13 (Panel A) and silent agonists 14-20 (Panel B).
The possibility of targeting the α7 nAChR by means of dual distinct mechanisms, such as
the above mentioned ionotropic and metabotropic-like activation modes, encouraged us to design
4

new compounds containing the general pharmacophoric features highlighted above. Our goal
ACCEPTED MANUSCRIPT
was to identify novel α7 ligands and focus on the structural requirements able to tune the full,
partial, and silent agonist profiles. For the design of the new derivatives we selected the rigid
quinuclidine basic nucleus, which is a pivotal moiety in the molecular skeleton of a variety of α7
nAChR ligands [24,50]. Moreover, based on our previous studies on spirocyclic quinuclidinyl-
2
∆
-isoxazoline analogs [21,43,48], such as 8, 20a, and 20b which showed full and silent
agonism, respectively, we planned to increase the distance between the positive ionizable/ionized
quinuclidine ring system and the hydrogen-bond acceptor moiety (Fig. 2). Thus, we explored the
insertion of a methylene group between the azabicyclic nucleus and the heterocyclic ring to
enhance ligand flexibility and thereby the chance for additional interactions with the receptor
protein. Thus, in the new designed derivatives 21a-34a and 21b-34b (Fig. 2), the basic
quinuclidine nucleus was connected through a methylene bridge to a 1,2,4-oxadiazole
heterocyclic moiety, bearing at the 3-position a meta-substituted phenyl or a heterocyclic ring in
analogy to the previously studied α7 ligands related to 20a and 20b. The meta-position of the
phenyl ring was probed with different substituents to investigate the α7 binding pocket and tune
the electrophysiological activity by favoring additional interactions within the orthosteric
recognition domain (Table 1). Substituents were selected to cover a variety of possible
connections, spanning from hydrogen bond to halogen bond and π-π interactions.
Fig. 2. Conception of novel 5-(quinuclidin-3-ylmethyl)-1,2,4-oxadiazole derivatives 21a-34a and 21b-34b (for
structures see Table 1).
5

Herein we describe the synthesis of target compounds as tertiary amines 21a-34a and their
ACCEPTED MANUSCRIPT
corresponding quaternary salts 21b-34b and discuss the changes in the α7 nAChR activation
mode following the introduction of different groups on the 5-(quinuclidin-3-ylmethyl)-1,2,4-
oxadiazole scaffold.
2
. Results and discussion
2
.1. Chemistry
The 5-(quinuclidin-3-ylmethyl)-1,2,4-oxadiazole derivatives 21a-34a and 21b-34b were
synthesized as racemates according to a one-pot cyclization reaction of a quinuclidine methyl
ester intermediate and the appropriate amidoximes under basic conditions [51,52]. The synthetic
route for the preparation of the methyl ester derivative 38 is outlined in Scheme 1. Commercially
available quinuclidin-3-one 35 was reacted with trimethyl phosphonoacetate in dry
tetrahydrofuran and in the presence of sodium hydride [53]. The Wittig reaction afforded in
excellent yield the alkene 36 as a 1:1 mixture of the (E) and (Z) isomers, which was reduced by
hydrogenation over palladium on carbon to give intermediate 37. Subsequent protection of the
quinuclidine ring nitrogen atom with a 1.0 M solution of borane-THF complex yielded the 3-
acetic acid methyl ester quinuclidine-N-borane 38 [21,43].
Scheme 1. Synthetic route to key intermediate 38. Reagents and conditions: a) Trimethyl phosphonoacetate (1.2
equiv), NaH (1.2 equiv), dry THF, 0 °C to rt, 1 h, then 35 (1 equiv), rt, 12 h; b) H , 10% Pd/C, MeOH, rt, 1 h; c) 1
2
M BH ·THF complex (1 equiv), dry THF, 0 °C to rt, 1.5 h.
3
The amidoximes 47, 48, 51 were purchased, while their analogs 39-46, 49, 50, 52 were
quantitatively synthesized by refluxing the corresponding nitrile precursors with hydroxylamine
in ethanol (Scheme 2), according to a standard literature procedure [52].
6

ACCEPTED MANUSCRIPT
Scheme 2. General procedure for preparation of amidoximes 39-46, 49, 50, 52. Reagents and conditions: a)
NH OH·H O (4 equiv), EtOH, 90 °C, 1 h. Amidoximes 47, 48 and 51 are commercially available.
2
2
Key intermediates 53-65 were prepared by employing the quinuclidine-N-borane methyl
ester 38 and the suitable amidoximes in the presence of cesium carbonate or sodium hydride as
base (Scheme 3). Cesium carbonate did not require anhydrous conditions and allowed the
reaction between methyl ester and amidoxime to proceed smoothly and in high yields affording
the 1,2,4-oxadiazole intermediates 53-56, 58-60, 43, 65, 34 (Method A). Conversely,
amidoximes 43, 47, 48, 50 did not react efficiently under these conditions and were therefore
treated with sodium hydride in the presence of 3Å molecular sieves in dry tetrahydrofuran to
give the corresponding derivatives 57, 61, 62, 64 (Method B). The borane-protecting group was
quantitatively removed by treatment of 53-65 with trifluoroacetic acid in acetone and led to the
quinuclidinyl free bases 21-33. Compound 34 was directly prepared in excellent yield by
reacting the unprotected quinuclidine methyl ester 37 with methyl amidoxime 52. On the other
hand, the base-promoted cyclization reaction applied to 37 with the other amidoximes proceeded
in poor yields and required tricky purification steps. Finally, the free bases 21-34 were converted
into the corresponding fumarates 21a-34a or quaternary methyl ammonium derivatives 21b-34b
by treatment with fumaric acid or methyl iodide in methanol, respectively.
7

ACCEPTED MANUSCRIPT
Scheme 3. General synthetic route to 5-(quinuclidin-3-ylmethyl)-1,2,4-oxadiazoles 21a-34a and 21b-34b. Reagents
and conditions: a) 39-52, 44-46, 49, 51 (3 equiv), Cs CO (3 equiv), dry THF, 50°C, 20 min, then 38 (1 equiv.), 0°C
2
3
to 80°C, 12 h; b) 43, 47, 48, 50, 52 (3 equiv), 3Å molecular sieves, dry THF, rt, 30 min, then NaH (3 equiv), rt to
0°C, 20 min and then 38 (or 37 for 52) (1 equiv.), rt to 80°C, 12 h; c) CF COOH (5 equiv), acetone, 0°C to rt, 12 h;
5
3
d) fumaric acid (1 equiv), MeOH, rt, 2 h; e) CH I (8 equiv), MeOH, rt, 12 h.
3
2
.2. Electrophysiology studies
The new 5-(quinuclidin-3-ylmethyl)-1,2,4-oxadiazoles 21a-34a and 21b-34b were assayed
on human α7 nAChRs expressed in Xenopus laevis oocytes with two-electrode voltage clamping.
To investigate the partial agonism, compounds were initially tested at the standard concentration
of 30 µM, and the resulting responses were normalized to the control ACh 60 µM (Table 1).
Data were analyzed considering both peak amplitude and net charge of the evoked responses,
and the ratios of peak to net charge were used to correlate the tested concentration of 30 µM to
the compound EC50 (half maximal effective concentration) [54]. Moreover, the newly
synthesized derivatives were tested in co-application with 10 µM N-(5-Chloro-2,4-
dimethoxyphenyl)-N'-(5-methyl-3-isoxazolyl)-urea (PNU-120596) to further characterize their
activity at the α7 nAChRs (Table 2). Type II PAMs, like PNU-120596 (PNU), are known to
modulate compound agonism by interacting at the receptor allosteric binding site and affecting
the equilibria of receptor states. Indeed, α7 PAMs induce prolonged receptor activation by
increasing receptor opening probability and destabilizing the desensitized PAM-sensitive Ds
receptor conformational state. These two mechanisms result in potentiated evoked responses
8

compared to those induced by application of the agonists alone. Destabilization of the Ds
ACCEPTED MANUSCRIPT
receptor state and conversion to the kinetically coupled PAM-dependent conductive state
induced by type II PAMs allow identification of the so-called α7 silent agonists [36]. Silent
compounds are intrinsically weak α7 agonists able to promote receptor desensitization, which
can be revealed by co-application of a type II PAM. The threshold for silent agonism was set to
one tenth of the response evoked by 60 µM ACh applied alone. Compounds characterized by net
charge responses greater than this value are considered partial agonists. Previous experiments
[
55] identified 10 µM PNU to yield optimal potentiation of a range of concentrations of ACh-
evoked currents over a range of ACh concentrations. Therefore, the target derivatives were tested
at 30 µM in co-application with 10 µM PNU, and the evoked net-charge responses were
normalized to receptor responses to 60 µM ACh applied alone. The degree of potentiation
achieved in the presence of the PAM is quantified as PNU effectiveness and calculated as the
ratio of responses evoked in co-application with PNU to responses evoked by the compound
applied alone.
The detailed α7 nAChR activity profiles of the new derivatives applied alone or in co-
application with PNU are reported in Tables 1 and 2, respectively. Derivatives 21-34 displayed
varying degrees of partial and silent agonism, mainly depending on both the nature of the
substituents on position 3 of the 1,2,4-oxadiazole ring and the presence or absence of the
ammonium group.
9

ACCEPTED MANUSCRIPT
Table 1
Structures and electrophysiological data for the partial agonist properties of derivatives 21a-34a and 21b-34b.
a
a
Compd
R
Relative response
Net charge
Relative response
Peak
Ratio
4.47
Peak
Net charge
Ratio
1.06
2
1
2.047 ± 0.348
0.458 ± 0.022
0.067 ± 0.012
0.063 ± 0.007
2
2
2
2
2
2
2
3
4
5
6
7
1.202 ± 0.330
0.920 ± 0.151
1.211 ± 0.118
0.144 ± 0.036
2.107 ± 0.699
1.415 ± 0.215
0.320 ± 0.027
0.274 ± 0.009
0.245 ± 0.011
0.055 ± 0.008
0.434 ± 0.020
0.511 ± 0.031
3.76
3.36
4.94
2.62
4.85
2.77
0.293 ± 0.046
0.440 ± 0.070
0.866 ± 0.251
0.077 ± 0.031
0.079 ± 0.010
0.119 ± 0.024
0.151 ± 0.011
0.177 ± 0.013
0.239 ± 0.017
0.035 ± 0.004
0.077 ± 0.005
0.129 ± 0.017
1.94
2.49
3.62
2.20
1.03
0.92
2
2
3
3
3
8
9
0
1
2
1.975 ± 0.375
0.867 ± 0.143
0.553 ± 0.106
2.143 ± 0.508
2.064 ± 0.325
0.401 ± 0.020
0.483 ± 0.025
0.703 ± 0.247
0.575 ± 0.038
0.508 ± 0.029
4.93
1.80
0.79
3.73
4.06
0.051 ± 0.011
0.023 ± 0.002
0.034 ± 0.013
0.010 ± 0.002
0.103 ± 0.012
0.075 ± 0.009
0.048 ± 0.019
0.047 ± 0.029
0.014 ± 0.001
0.127 ± 0.026
0.68
0.48
0.72
0.71
0.81
3
3
3
4
1.299 ± 0.226
0.026 ± 0.005
0.602 ± 0.041
0.042 ± 0.008
2.16
0.62
0.069 ± 0.015
0.003 ± 0.001
0.112 ± 0.016
0.001 ± 0.002
0.62
3.00
a
“
4
Relative response” refers to receptor response to a 30 µM application of test compound. The values are the mean ± SEM for N ≥
experiments and are relative to the response of the receptor to 60 µM control applications of ACh. Details are given in the
experimental section.
1
0

ACCEPTED MANUSCRIPT
Table 2
Structures and electrophysiological data for the silent agonist properties of derivatives 21a-34a and 21b-34b.
Relative
response
Potentiated
response
PNU
effectiveness
Relative
response
Potentiated
response
PNU
effectiveness
Compd
R
a
a
a
a
2
2
2
2
1
2
3
4
0.458 ± 0.022
0.9 ± 0.3
0.6 ± 0.1
0.5 ± 0.1
0.5 ± 0.0
2
2
2
2
0.063 ± 0.007
0.3 ± 0.1
0.7 ± 0.1
0.3 ± 0.0
0.3 ± 0.0
5
5
2
1
0.320 ± 0.027
0.274 ± 0.009
0.245 ± 0.011
0.151 ± 0.011
0.177 ± 0.013
0.239 ± 0.017
2
5
0.055 ± 0.008
0.1 ± 0.0
2
0.035 ± 0.004
0.1 ± 0.0
3
2
2
6
7
0.434 ± 0.020
0.511 ± 0.031
77.3 ± 17.7
1.0 ± 0.1
178
2
0.077 ± 0.005
0.129 ± 0.017
1.2 ± 0.3
1.8 ± 0.4
16
14
2
2
3
3
3
8
9
0
1
2
0.401 ± 0.020
0.483 ± 0.025
0.703 ± 0.247
0.575 ± 0.038
0.508 ± 0.029
1.6 ± 0.3
13.7 ± 3.1
25.5 ± 15.9
32.3 ± 12.3
5.0 ± 1.0
4
0.075 ± 0.009
0.048 ± 0.019
0.047 ± 0.029
0.014 ± 0.001
0.127 ± 0.026
2.2 ± 0.5
3.9 ± 0.9
2.6 ± 0.8
27.9 ± 15.8
0.2 ± 0.0
29
81
28
36
56
10
55
1993
2
3
3
3
4
0.602 ± 0.041
0.042 ± 0.008
1.7 ± 0.4
5.1 ± 0.3
3
0.112 ± 0.016
0.001 ± 0.002
2.4 ± 0.5
0.1 ± 0.0
21
121
100
a
The values are the mean ± SEM for N ≥ 4 experiments. All data are relative to the net charge response of the receptor to 60 µM control
applications of ACh. “Relative response” refers to receptor response to a 30 µM application of test compound. Potentiated response
refers to receptor response with 30 µM test compound and 10 µM PNU-120596 co-application. Details are given in the experimental
section.
1
1

ACCEPTED MANUSCRIPT
2
.2.1. Partial agonists
According to the experimental protocol described above, we characterized as partial agonists
those compounds inducing net-charge responses greater than 0.1 ACh controls. The majority of
the tested derivatives belonging to a and b series displayed weak to strong partial agonist
activation of the α7 nAChR (Table 1, Fig. 3). Compounds of the a series showed both higher
agonist activity and potency compared to their corresponding methylated analogs of the b series
as evidenced by their peak-to-net-charge ratios (Table 1). As it can be observed after an initial
inspection of the data, the nature of the substituent on the oxadiazole ring deeply affects the
resulting responses at the α7 nAChR subtype.
Fig. 3. Agonist activity of the new 5-(quinuclidin-3-ylmethyl)-1,2,4-oxadiazole derivatives at the α7 nAChR.
Compounds of both a (tertiary amines) and b (quaternary ammonium salts) series are reported in ascending order of
response magnitude. The vertical axis refers to the net-charge response evoked by compounds (30 µM), relative to
ACh controls. Yellow bars represent net charge responses lower than 0.50 relative to ACh controls; red bars
represent responses higher than 0.50 relative to ACh controls. Experimental values are the average of at least four
independent measurements, and the error bars reflect the standard deviation of the mean.
As far as the halogen-containing derivatives 21-24 are considered, although they all behaved
as partial agonists, tertiary amines of the a series and quaternary ammonium salts of the b series
showed opposite trends of α7 activation. The increment in halogen size, from fluorine to iodine,
promoted a significant reduction of partial agonism in the first series, while keeping potency
almost constant. On the other hand, the same modifications resulted in a significant increase of
partial agonism and potency in the b series, with the permanently positively charged
quinuclidine. Interestingly, the trifluoromethyl-substituted analog 25 failed to evoke any α7-
1
2

mediated response in both series. Generally, the trifluoromethyl group is considered a
ACCEPTED MANUSCRIPT
bioisosteric surrogate of the methyl group. However, in our case, the two groups gave rise to
quite different properties of the resulting ligands: while the 3-CF -substituted derivatives 25a
3
and 25b showed no activation, the m-tolyl derivative 26a displayed partial agonism at the α7
nAChR. Introduction of hydrogen-bonding groups, such as in the 3-OMe-substituted analogs
2
7a, 27b and 28a, resulted in the partial activation of the receptor subtype under investigation.
Both unsubstituted phenyl derivatives 33a and 33b displayed partial agonism at the α7 receptor,
but they differed in the magnitude of the evoked responses, with the tertiary amine 33a showing
a 5-fold greater receptor activation compared to the ammonium analog 33b. The effects of a
heteroaromatic ring substitution on the 1,2,4-oxadiazole moiety was evaluated by insertion of
pyridine, furan, and thiophene (analogs 29-32). These structural variations caused relevant
differences within the two series of compounds. In the a series, the heteroaromatic rings evoked
partial agonism with a wide range of potencies compared to ACh; on the contrary, the same
substitution pattern resulted in no or negligible activation if the methylated quinuclidine salts of
the b series are taken into account. Finally, replacement of the substituted aromatic ring with an
aliphatic moiety, i.e. the two methyl analogs 34a and 34b, resulted in a lack of any agonist
activation of the α7 subtype, thus suggesting the crucial role of the aromatic ring in that position
to gain the wanted receptor activation.
The compounds showing partial agonism in the two sets of nicotinic ligands were further
investigated to assess the degree of α7 receptor desensitization induced by their application at 30
µM (Table 2, Fig. 4). Partial and full agonists of the α7 nAChRs are known to induce rapid
desensitization of the receptor, which prevents its further activation upon subsequent application
of orthosteric ligands. At least two distinct types of desensitization have been identified and
distinguished for their sensitivity (D ) or insensitivity (D ) to type II PAMs like PNU-120596
s
i
(
PNU) [55,56]. In the presence of a suitable allosteric modulator, D is converted into an open
s
and conductive state, detectable as an ion flux through the channel. On the other hand, D is not
i
affected by PAMs and no current is measured. By co-applying 10 µM PNU, we investigated the
ratio of D and D induced by each compound at the given concentration of 30 µM; partial
s
i
agonists with low PNU responsiveness emerged as better stabilizers of D over D . In spite of
i
s
evoking significant channel activation, some of the studied compounds, i.e. 21a-24a, 27a, 22b-
4b, and 32b, were unable to desensitize the α7 subtype at 30 µM, with responses in co-
2
application with the PAM equal or lower than the ACh controls. The poor desensitization (Ds)
observed for these derivatives may be ascribed to peculiarities of partial agonist PAM-
potentiated responses. Indeed, such responses are often characterized by an inverted U-shaped
curve, with large responses over a limited concentration range and sharp decrease in channel
opening at lower and higher agonist concentrations [34]. The low channel opening in the
1
3

presence of a type II PAM indicates D to be predominant over D , while at the apex of the
i
s
ACCEPTED MANUSCRIPT
potentiated response, D prevails over D . The tested partial agonists displayed short lasting
s
i
desensitization of the α7 nAChR since post-control ACh was not significantly inhibited after
compound application (data not shown).
Fig. 4. Activity of the new 5-(quinuclidin-3-ylmethyl)-1,2,4-oxadiazole tertiary amines of the a series (Panel A) and
quaternary ammonium salts of the b series (Panel B) at the α7 nAChR. The vertical axis refers to the net-charge
response evoked by compounds (30 µM) applied alone (red bars) or co-applied with 10 µM PNU-120596 (green
bars), relative to ACh controls. Experimental values are the average of at least four independent measurements, and
the error bars reflect the standard deviation of the mean.
Derivatives 28a, 31a, 32a, and 33a showed similar partial agonist net charge responses
(0.401 ± 0.020, 0.575 ± 0.038, 0.508 ± 0.029, and 0.602 ± 0.041, respectively) (Table 1) and for
all four compounds the peak-to-net-charge ratio indicated an EC50 value lower than 30 µM.
Despite all compounds being potentiated by co-application of 10 µM PNU, differences emerged
1
4

in the magnitude of responses (Table 2), with the furan derivative 31a displaying a 20-fold
ACCEPTED MANUSCRIPT
higher potentiated response (32.3 ± 12.3) than that shown by compounds 28a and 33a (1.6 ± 0.3
and 1.7 ± 0.4) and 6-fold higher than that of the corresponding thiophene isosteric analog 32a
(5.0 ± 1.0). We further investigated these compounds at a lower concentration and tested them at
3
µM on α7 nAChRs alone and co-applied with 10 µM PNU (Fig. 5). At 3 µM, 28a, 31a, 32a
and 33a retained their partial agonist behavior, with a noticeably decreased response affecting
derivative 31a exclusively (0.15 ± 0.03 at 3 µM vs 0.58 ± 0.04 at 30 µM). Indeed, PAM co-
application uncovered the hypothesized inverted U-shaped responses for compounds 28a, 32a,
and 33a by displaying evoked currents of 4.4 ± 1.5, 13.4 ± 1.8, and 17.4 ± 3.0, respectively.
Therefore, application of tested compound at 3 µM resulted in a 3-fold enhancement of
potentiated response for 28a and 32a, and a 10-fold increase for 33a compared to the potentiated
evoked currents at 30 µM.
Fig. 5. Activity of the new 5-(quinuclidin-3-ylmethyl)-1,2,4-oxadiazoles derivatives 28a, 31a-33a at the α7 nAChR.
A) The vertical axis refers to the net-charge response evoked by compounds (3 and 30 µM), relative to ACh
controls. B) The vertical axis refers to the net-charge response evoked by compounds (3 and 30 µM) co-applied with
1
0 µM PNU-120596, relative to ACh controls. Experimental values are the average of at least four independent
measurements, and the error bars reflect the standard deviation of the mean.
Two partial agonist derivatives, 30a and 31a, were selected for representative
electrophysiological traces to show the effect of PNU co-application at the α7 nAChR and stress
the differences in partial agonist responses evoked at this subtype (Fig. 6). The peak-to-net-
charge analyses implied an EC value close to 30 µM for 30a (peak/net charge = 0.8) and lower
5
0
than 30 µM for 31a (peak/net charge = 3.7). The different potencies of the two compounds are
1
5

highlighted by the corresponding response shapes, broad and longer for 30a while sharp and
ACCEPTED MANUSCRIPT
shorter for 31a.
Fig. 6. Electrophysiological responses of partial agonists 30a and 31a. The multi-cell averaged data traces represent
the averaged normalized responses (see methods) of Xenopus laevis oocytes expressing human α7 nAChRs to the
application of the compounds 30a (Panel A) and 31a (Panel B) at a probe concentration of 30 µM and in co-
application with 10 µM PNU-120596, relative to averaged responses from the same cells to 60 µM ACh controls
applied 4 minutes prior to experimental application. The ACh and drug-evoked responses were scaled in amplitude
to the responses evoked in co-application with PNU-120596. The inset (Panel C) shows at a higher scale the
superimposition of the traces to the application of 30a and 31a at 30 µM to better highlight the different partial
agonist response of the two compounds.
2
.2.2. Silent agonists
Some analogs of the a and b series showed no activation of the α7 nAChR or induced very
weak responses at 30 µM. Various derivatives with poor or no α7 channel activation and the
ability to induce receptor desensitization as silent agonists have been reported in the recent
literature [34,36,40−43]. To study their behavior as α7 receptor desensitizers, derivatives 21-34
were tested in co-application with the type II PAM PNU-120596 (PNU) (Table 2, Fig. 7).
Tertiary amines 21a-34a and their corresponding N-methyl ammonium salts 21b-34b displayed
1
6

quite distinct properties at 30 µM. Among the tertiary amines, the 3-methyl substituted analog
ACCEPTED MANUSCRIPT
4a, the only derivative characterized by the absence of an aromatic or a heteroaromatic moiety,
3
emerged as the sole α7 silent agonist (30 µM: 0.042 ± 0.008), with a PNU-potentiated net-charge
response of 5.1 ± 0.3. On the other hand, quaternary ammonium derivatives of the b series
provided various compounds with silent activity at 30 µM. Derivatives 26b and 28b-30b
displayed weak potentiated responses in co-application with PNU, with net-charge current lower
than 4. Worth noting, the furan derivative 31b emerged as a stronger α7 desensitizer as revealed
by PAM co-application (27.9 ± 15.8).
Fig. 7. Activity of the new 5-(quinuclidin-3-ylmethyl)-1,2,4-oxadiazoles silent derivatives 34a, 26b, 28b-31b at the
α7 nAChR (a series: tertiary amines; b series: quaternary ammonium salts). The vertical axis refers to the net-charge
response evoked by compounds (3 µM, 10 µM and 30 µM) co-applied with 10 µM PNU-120596, relative to ACh
controls. Experimental values are the average of at least four independent measurements, and the error bars reflect
the standard deviation of the mean.
Analogously to partial agonists, silent ligands are α7 desensitizers and selectively stabilize
the D and D receptor conformational states in different ratios. Depending on the induced ratio
s
i
of D and D at a given concentration, each ligand is more or less responsive to PNU application,
s
i
thus resulting in concentration-response inverted U-shaped curves. To investigate their silent
activity over a wider range of concentrations and understand where 30 µM is on the U-shaped
PNU responses, compounds 34a, 26b, 28b, and 31b were further tested at lower concentrations
(3 and 10 µM). The lower concentrations resulted in the lack of potentiated response for 34a,
thus suggesting either preferable stabilization of D or a closed state at 3 and 10 µM. On the other
i
hand, the three permanently charged ligands showed a variety of results over the tested
concentrations. While behaving as a very weak silent agonist at 3 and 30 µM, the m-tolyl
derivative 26b evoked a potentiated response of 10.9 ± 2.8 when co-applied with PNU at 10 µM,
1
7

thus proving that this concentration is at the apex of the above-mentioned inverted U PAM curve
ACCEPTED MANUSCRIPT
for 26b. The 4-OMe-phenyl analog 28b induced comparable weak PNU-potentiated responses at
1
0 and 30 µM and no potentiation at the lowest 3 µM concentration. We confirmed that the
permanently charged furan derivative 31b is the most potent α7 silent agonist in the whole group
of analogs, achieving an optimal concentration for PNU potentiation at 30 µM. Interestingly, the
furan motif is also present in the α7 prototypical silent agonist NS-6740 (Fig. 1B) and was
associated to hydrogen-bonding acceptor properties within the receptor binding pocket [36]. Our
data on compound 31b, together with those collected on 28b-30b, indicate a suitable hydrogen-
bonding moiety as a structural requirement to trigger the α7 nAChR silent activation. Indeed, in
the entire set of derivatives, compounds with substituents unable to give a hydrogen-bonding
interaction, i.e. halogens or methyl groups, were characterized by higher partial agonist
responses and lower receptor desensitization at 30 µM (Table 2).
2
.3. Structure-activity relationships
The electrophysiological data from whole-cell two-electrode voltage-clamp recordings
allowed us to perform a structure-activity relationship (SAR) for the agonist, either partial or
silent, profile of the set of novel 5-(quinuclidin-3-ylmethyl)-1,2,4-oxadiazole analogs 21a-34a
2
and 21b-34b under investigation. If the results on spirocyclic quinuclidinyl-∆ -isoxazoline
derivatives are considered [43], lengthening of the distance between the positive ionized
quinuclidinyl group and variation of the hydrogen-bond acceptor heterocyclic moiety (Fig. 2)
mostly led to partial agonists at the α7 nAChR while preventing its silent activation. Data on the
unsubstituted phenyl derivatives 33a and 33b suggested that partial agonism is the default profile
at α7 nAChRs for ligands containing the 3-phenyl-5-(quinuclidin-3-ylmethyl)-1,2,4-oxadiazole
scaffold, since the majority of the tested compounds behaved as moderate to strong partial
agonists. However, the nature of the substituents on the phenyl ring, together with its
replacement with heteroaromatic moieties, allowed us to tune the electrophysiological profile
towards silent agonism or led to inactive compounds. A detailed analysis of the new derivatives
evidenced six silent agonists (34a, 26b and 28b-31b) of the α7 receptor subtype and four
compounds (21b, 25a, 25b and 34b) devoid of the ability to activate the receptor.
We compared the agonist net-charge responses for tertiary amines of the a series and
quaternary ammonium salts of the b series (Fig. 8). With the exception of the two pairs 25 and
3
4, characterized by the lowest agonist responses, the resulting scatter plot indicated an inverse
relationship between the net-charge responses obtained with the two series when the compounds
1
8

have the same substituent on the oxadiazole ring, thus suggesting that multiple features may
ACCEPTED MANUSCRIPT
concur in determining the magnitude of the α7 agonism.
Fig. 8. Agonist activity relationship at the α7 nAChR of compounds of the a and b series. The traces represent the
averaged normalized responses (see methods) of Xenopus laevis oocytes expressing human nAChRs to the
application of the compound 21a-34a and 21b-34b at a probe concentration of 30 µM. The horizontal axis
represents the α7 agonist activity of tertiary amine derivatives 21a-34a, expressed as net charge. The vertical axis
represents the α7 agonist activity of quaternary ammonium derivatives 21b-34b, expressed as net charge.
Experimental values are the average of at least four independent measurements, and the error bars reflect the
standard deviation of the mean.
Halogen-substituted ligands were endowed with moderate partial agonism, with increasing
size of the halogen atom inducing opposite trends in the a and b series. Methyl- and
trifluoromethyl-substituted analogs failed to display bioisosterism, since CF -containing
3
derivatives 25a and 25b were inactive while CH -containing ligands 26a and 26b behaved as
3
partial and silent agonists, respectively. The presence of a polar methoxy group on the phenyl
ring strongly promoted partial agonist activity at the α7 receptor, with the exception of 28b
which showed a good silent agonist profile at the tested (10 and 30 µM) concentrations.
A relevant difference among tertiary amines and their corresponding permanently charged
analogs emerged by comparing the data on derivatives which incorporate a heteroaromatic
1
9

moiety, which gave rise to partial agonism in the a subgroup or to the absence of any receptor
ACCEPTED MANUSCRIPT
activation in the b subgroup. However, excluding the thiophene analog 32b, these ligands, i.e.
2
9b, 30b, and 31b, displayed good silent agonist properties, highlighting the relevant role of a
further hydrogen-bonding acceptor group, in addition to the 5-(quinuclidin-3-ylmethyl)-1,2,4-
oxadiazole scaffold, in triggering receptor desensitization. Therefore, as a general observation
and in accordance with previously reported results [43], the presence of a methyl ammonium
function is most effective in affording compounds possessing a silent profile at α7 nAChRs.
However, the absence of an aromatic substituent on the 5-(quinuclidin-3-ylmethyl)-1,2,4-
oxadiazole skeleton coupled with the presence of a tertiary amine core caused silent activity in
compound 34a. These results are in agreement with previous data known for amine-containing
silent agonism pharmacophores [41,43], thus evidencing that a permanent positive charge on the
nitrogen nucleus is an important but not essential requirement for α7 silent receptor activation.
Finally, the m-trifluoromethyl-phenyl substituted analogs 25a and 25b, besides lacking
partial agonist activity (0.055 ± 0.008 and 0.035 ± 0.004, respectively, Table 1), did not induce
desensitization of the α7 receptor subtype and displayed no potentiation in co-application with
PNU (0.1 ± 0.0 and 0.1 ± 0.0, respectively, Table 2). These results underline the differences of
2
the present set of quinuclidinyl-1,2,4-oxadiazoles compared to the spirocyclic quinuclidinyl-∆ -
isoxazolines recently reported by our research group, where the m-trifluoromethyl-phenyl
substituted derivative showed the properties of a silent agonist [43]. We further investigated the
quaternary ammonium salt 25b at the α7 nAChR co-applied with ACh to assess its antagonist
activity (Fig. 9, Panel A). In the presence of the tested compound, ACh showed a residual
activity of only 10% at the α7 receptor subtype and the post-control ACh was still significantly
inhibited (60% residual activity). Once identified as an antagonist at the α7 nAChR, derivative
2
5b was tested at the ganglionic α3β4 nAChR to evaluate its selectivity towards the two
nicotinic receptor subtypes (Fig. 9, Panel B). It displayed relevant antagonist activity at the α3β4
receptor as indicated by the inhibition of ACh-evoked channel activation. However, the
antagonist component of 25b at the α3β4 subtype was brief in duration, as proved by the absence
of ACh post-control inhibition. The ACh co-application experiments indicate 25b as an
antagonist of the nAChRs devoid of subtype specificity.
2
0

ACCEPTED MANUSCRIPT
Fig. 9. Electrophysiological antagonist responses of compound 25b at the α7 (Panel A) and the α3β4 nAChR (Panel
B). The traces represent the averaged normalized responses (see methods) of Xenopus laevis oocytes expressing
human nAChRs to the application of the compound 25b at a probe concentration of 30 µM and in co-application
with 60 µM or 100 µM ACh for α7 and α3β4 nAChRs, respectively. The responses are reported as relative to
averaged responses from the same cells to 60 µM and 100 µM ACh controls applied 4 minutes prior to experimental
application. The ACh and drug-evoked responses were scaled in amplitude to the responses evoked by ACh
controls.
3
. Conclusion
Selective activation of the α7 nAChR subtype is a therapeutic option for a variety of
disorders including cognitive impairment, neurodegeneration, and inflammation. A dual
mechanism of activation, involving either the classic ionotropic channel opening or an
alternative metabotropic signaling pathway, has been reported for this receptor subtype. Among
the ligands that mediate ion channel activation, partial agonists retain robust efficacy in
preclinical models over a wider dose-exposure range than agents with a high level of intrinsic
efficacy. Besides, silent agonists, which promote α7 receptor desensitization over activation by
inducing non-conducting ligand-bound conformational states, appeared to be therapeutically
favorable for peripheral pain and inflammatory diseases. Therefore, compounds characterized by
a partial or silent agonist profile at the α7 nAChR may represent better drug candidates for α7-
related disorders.
In the present study, we aimed at selecting a specific modulation of signal transduction of
α7 nAChRs between a dependent and an independent channel activation pathway. To such an
end, we planned suitable chemical modifications on the 5-(quinuclidin-3-ylmethyl)-1,2,4-
2
1

oxadiazole core skeleton, then synthesized and tested the two series of structurally related
ACCEPTED MANUSCRIPT
ligands 21a-34a and 21b-34b. Overall, the set of novel derivatives showed a pharmacological
2
profile which differs from that of the previously investigated quinuclidinyl-∆ -isoxazoline
spirocyclic compounds [21,43]. As a general trend, in the new group of nicotinic ligands,
increasing both ligand flexibility and distance between the cationic center and the functionalized
heterocyclic ring shifted the activity profile towards partial agonism, whereas the more
2
constrained spirocyclic quinuclidinyl-∆ -isoxazolines preferentially induced the α7 desensitized
PAM-sensitive D state.
s
However, some of the investigated derivatives (34a, 26b, and 28b-31b) displayed a silent
agonist profile at the target nicotinic receptor subtype. In agreement with previous data available
for the silent α7 derivative NS-6740 (14), we found that an additional hydrogen-bond acceptor
group within the binding pocket favors the silent desensitizing profile at the α7 nAChR, such as
in ligands 28b, 29b, 30b, and 31b, incorporating a 4-OMe-phenyl ring, a 3- or 4- substituted
pyridine, and a furan ring, respectively. Conversely, derivatives bearing groups in the same
position unable to give hydrogen-bonding interactions, i.e. halogen and methyl in both a and b
series, did not induce receptor PAM-sensitive desensitization.
4
. Experimental section
4
.1. Chemistry
All reagents and solvents were purchased from Sigma-Aldrich Srl (Milan, Italy) and were
used without further purification. All reactions were carried out under inert atmosphere of argon
or nitrogen and monitored by thin-layer chromatography (TLC) on commercial aluminum plates
precoated with silica gel 60 (F-254, Merck) or with aluminum oxide (F-254, Fluka).
Visualization was performed with UV light at 254 nm. Spots were further evidenced by spraying
with a dilute alkaline potassium permanganate solution or a phosphomolybdic acid solution and,
for tertiary amines, with the Dragendorff reagent. Glassware was oven-dried or flame-dried prior
to use. The synthesized compounds were purified on glass flash chromatography columns
packed with silica gel (230–400 mesh particle size, pore size 60 Å, Merck). Melting points of
solid products were measured in capillary tubes with a model B 540 Büchi apparatus and are
1
13
1
uncorrected. H NMR and C NMR spectra were recorded with a Varian Mercury 300 ( H,
13
3
00.063; C, 75.451 MHz) spectrometer at 20 °C. Abbreviations used for peak multiplicities are
given as follows: s (singlet), bs (broad singlet), d (doublet), t (triplet), q (quartet), m (multiplet).
Coupling constants (J) are given in Hz, chemical shifts (δ) are expressed in parts per million
1
13
(
ppm) and calibrated for H using TMS and for C using residual deuterated solvent as internal
2
2

standard (CDCl , 77.16 ppm; CD OD, 49.00 ppm; DMSO-d , 39.52 ppm). ESI mass spectra
3
3
6
ACCEPTED MANUSCRIPT
were obtained on a Varian 320 LC-MS/MS instrument. Data are reported as mass-to-charge ratio
m/z) of the corresponding positively charged molecular ions. The purity of final compounds was
(
determined using an Agilent Technologies 1200 series HPLC instrument. The HPLC analyses
were performed using two different methods: method 1 for tertiary amines 21a-34a and method
2
for quaternary ammonium salts 21b-34b. Method 1: Analysis was performed on a
LiChroCART 250-4 LiChrosher 100 RP-18 analytical column (5 µm) operating at 25 °C. The
injection volume of the sample was 10 µL. Elution was carried out with 3 mM ammonium
formate buffer containing 0.1% formic acid as mobile phase A and methanol as mobile phase B
(A/B 10:90) at a flow rate of 1 mL/min. Method 2: Analysis was performed on a Supelco
Supelcosil LC-SCX analytical column (3 × 250 mm, 5 µm) operating at 40 °C. The injection
volume of the sample was 10 µL. Elution was carried out with 50 mM monobasic potassium
phosphate buffer (pH = 3) containing 20 mM potassium chloride as mobile phase A and
acetonitrile as mobile phase B (A/B 60:40) at a flow rate of 1 mL/min. All final compounds
tested for biological activity showed > 95% purity (detection by UV absorption at 207 nm).
4
.1.1. 2-(1-Azabicyclo[2.2.2]oct-3-ylidene)-acetic acid methyl ester (36)
A stirred solution of trimethyl phosphonoacetate (7.74 mL, 48.0 mmol, 1.2 equiv) in dry
tetrahydrofuran (750 mL) under an argon atmosphere at 0 °C was treated portionwise with
sodium hydride 60% dispersion in mineral oil (1.92 g, 48.0 mmol, 1.2 equiv). After stirring for 1
h at room temperature (rt), quinuclidin-3-one 35 (5.00 g, 40.0 mmol, 1 equiv) was added
dropwise via a syringe to the mixture and stirring was continued overnight at rt. The reaction
completion was confirmed by TLC (dichloromethane/methanol 9:1) and the mixture was cooled
to 0 °C and quenched with deionized water (100 mL). The aqueous layer was extracted with
dichloromethane (3 × 300 mL) and the organic layers were collected and treated with 1 N HCl
(500 mL). The acidic aqueous phase was separated and basified to pH = 9 by portionwise
addition of potassium carbonate at 0 °C and then extracted with dichloromethane (3 × 300 mL).
The pooled organic phases were dried over anhydrous Na SO , filtered and concentrated in
2
4
vacuo to obtain compound 36 as a 1:1 mixture of (E) and (Z) stereoisomers, which was used
without further purification in the next reaction step. Yellow, very viscous oil; yield: 7.25 g,
1
1
00%; R = 0.40 (dichloromethane/methanol 9:1). H NMR (300 MHz, CDCl ) δ (ppm): 5.63 (t,
f
3
J = 2.5 Hz, 1H), 5.61 (t, J = 1.7 Hz, 1H), 3.94–3.85 (m, 3H), 3.66 (s, 3H), 3.65 (s, 3H), 3.47–
1
3
3
.41 (m, 2H), 2.98–2.71 (m, 8H), 2.48–2.38 (m, 1H), 1.84–1.56 (m, 8H). C NMR (75 MHz,
CDCl ) δ (ppm): 169.6, 167.9, 166.8, 166.3, 110.6, 110.3, 56.4, 50.7, 47.4, 47.2, 33.5, 27.1, 26.1,
3
2
3

1
3
2
6.1. Note: in the C NMR spectrum, the following pairs of peaks, 169.6 and 167.9, 166.8 and
ACCEPTED MANUSCRIPT
1
66.3, 110.6 and 110.3, represent the same carbon in the two stereoisomers E and Z.
4
.1.2. 1-Azabicyclo[2.2.2]octane-3-acetic acid methyl ester (37)
To a solution of 2-(1-azabicyclo[2.2.2]oct-3-ylidene)-acetic acid methyl ester (Z/E mixture)
6 (7.25 g, 40.0 mmol, 1 equiv) in methanol (500 mL), 10% Pd/C (725 mg) was added under an
3
argon atmosphere. The resulting mixture was stirred at rt for 1 h under a balloon of hydrogen.
The catalyst was filtered off over a celite pad, the celite was washed with methanol and the
solvent was removed under vacuum. The residue was dissolved in dichloromethane (200 mL)
and the solution was treated with 1 N HCl (200 mL). The organic layer was separated, and the
residual aqueous phase was washed with dichloromethane (2 × 200 mL) and then basified to pH
=
9 with Na CO and extracted with dichloromethane (3 × 200 mL). The combined organic
2 3
layers were dried over anhydrous Na SO , filtered and concentrated in vacuo to obtain
2
4
1
compound 37. Colorless oil; yield: 7.33 g, 100%; R = 0.19 (dichloromethane/methanol 7:3). H
f
NMR (300 MHz, CDCl ) δ (ppm): 3.67 (s, 3H), 3.22–3.07 (m, 1H), 2.91–2.71 (m, 4H), 2.43–
3
1
3
2
.29 (m, 3H), 2.19–2.02 (m, 1H), 1.72–1.52 (m, 4H), 1.48–1.32 (m, 1H). C NMR (75 MHz,
CDCl ) δ (ppm): 173.3, 54.7, 51.6, 47.7, 46.8, 38.5, 32.6, 28.1, 25.2, 21.2.
3
4
.1.3. Borane 1-azabicyclo[2.2.2]octane-3-acetic acid methyl ester complex (38)
A 1.0 M solution of BH in tetrahydrofuran (40 mL, 40.0 mmol, 1 equiv) was added dropwise
3
via a syringe under an argon atmosphere to a stirred solution of 1-azabicyclo[2.2.2]octane-3-
acetic acid methyl ester 37 (7.33 g, 40.0 mmol, 1 equiv) in dry tetrahydrofuran (500 mL) at 0 °C.
After stirring for 1.5 h at rt (TLC in dichloromethane/methanol 7:3), the mixture was quenched
by addition of a saturated aqueous solution of NaHCO . Dichloromethane was added and the
3
organic layer was separated. The aqueous layer was then extracted two more times with
dichloromethane. The collected organic phases were dried over anhydrous Na SO , filtered and
2
4
concentrated under reduced pressure affording the N-boranyl derivative 38. Colorless solid;
1
yield: 7.65 g, 97%; mp 87–88 °C; R = 0.30 (cyclohexane/ethyl acetate 4:1). H NMR (300 MHz,
f
CDCl ) δ (ppm): 3.62 (s, 3H), 3.31–3.14 (m, 1H), 3.03–2.82 (m, 4H), 2.54–2.41 (m, 1H), 2.41–
3
1
3
2
1
.22 (m, 3H), 1.85–1.67 (m, 4H), 1.67–1.47 (m, 1H). C NMR (75 MHz, CDCl ) δ (ppm):
3
72.1, 59.7, 53.5, 52.9, 51.8, 37.8, 31.5, 26.3, 24.8, 20.1.
4
.1.4. General procedure for the synthesis of amidoximes (39-52)
In a pressurized sealed vial, aqueous hydroxylamine 50% w/w (4 equiv) was added to a
stirred solution of the appropriate carbonitrile (1 equiv) in absolute ethanol. The resulting
2
4

mixture was heated at 90 °C for 1 h. The solvent was then evaporated to dryness providing the
ACCEPTED MANUSCRIPT
desired amidoxime quantitatively, which was used without further purification.
4
.1.4.1. 3-Fluoro-N'-hydroxybenzimidamide (39). According to general procedure,
hydroxylamine (2.29 mL, 37.42 mmol) was reacted with 3-fluorobenzonitrile (1.0 mL, 9.35
mmol) in absolute ethanol (8 mL) at 90 °C for 1 h to afford the desired compound 39. Colorless
1
oil; yield: 1.44 g, 100%; R = 0.51 (cyclohexane/ethyl acetate 1:1). H NMR (CDCl , 300 MHz)
f
3
1
3
δ (ppm): 7.64 (br s, 1H), 7.43–7.28 (m, 3H), 7.18–6.99 (m, 1H), 4.96 (br s, 2H). C NMR
CDCl , 75 MHz) δ (ppm): 162.8 (d, J = 246.4 Hz), 151.9, 134.7 (d, J = 8.1 Hz), 130.4 (d, J =
(
3
8
.2 Hz), 121.7 (d, J = 3.0 Hz), 117.0 (d, J = 21.2 Hz), 113.2 (d, J = 23.4 Hz).
4
.1.4.2. 3-Chloro-N'-hydroxybenzimidamide (40). According to general procedure,
hydroxylamine (2.67 mL, 43.62 mmol) was reacted with 3-chlorobenzonitrile (1.50 g, 10.90
mmol) in absolute ethanol (8 mL) at 90 °C for 1 h to afford the desired compound 40. Colorless
1
oil; yield: 1.86 g, 100%; R = 0.51 (cyclohexane/ethyl acetate 1:1). H NMR ((CD ) SO, 300
f
3 2
MHz) δ (ppm): 9.77 (s, 1H), 7.73–7.69 (m, 1H), 7.65 (dt, J = 6.7, 1.8 Hz, 1H), 7.44–7.35 (m,
1
3
2
H), 5.88 (s, 2H). C NMR ((CD ) SO, 75 MHz) δ (ppm): 149.6, 135.4, 132.9, 130.0, 128.6,
3 2
1
25.1, 123.9.
4
.1.4.3. 3-Bromo-N'-hydroxybenzimidamide (41). According to general procedure,
hydroxylamine (2.02 mL, 32.96 mmol) was reacted with 3-bromobenzonitrile (1.50 g, 8.24
mmol) in absolute ethanol (8 mL) at 90 °C for 1 h to afford the desired compound 41. Colorless
1
oil; yield: 1.77 g, 100%; R = 0.51 (cyclohexane/ethyl acetate 1:1). H NMR (CDCl , 300 MHz)
f
3
δ (ppm): 7.75–7.69 (m, 1H), 7.53–7.45 (m, 2H), 7.20 (dd, J = 9.4, 6.4 Hz, 1H), 4.79 (br s, 2H).
1
3
C NMR (CDCl , 75 MHz) δ (ppm): 151.6, 134.6, 133.2, 130.3, 129.2, 124.6, 122.9.
3
4
.1.4.4. N'-Hydroxy-3-iodobenzimidamide (42). According to general procedure, hydroxylamine
(1.60 mL, 26.2 mmol) was reacted with 3-iodobenzonitrile (1.50 g, 6.55 mmol) in absolute
ethanol (8 mL) at 90 °C for 1 h to afford the desired compound 42. Colorless oil; yield: 1.72 g,
1
1
00%; R = 0.51 (cyclohexane/ethyl acetate 1:1). H NMR (CDCl , 300 MHz) δ (ppm): 8.57 (br
f
3
s, 1H), 7.97 (t, J = 1.7 Hz, 1H), 7.74 (ddd, J = 7.9, 1.7, 1.0 Hz, 1H), 7.58 (ddd, J = 7.8, 1.5, 1.1
1
3
Hz, 1H), 7.12 (t, J = 7.9 Hz, 1H), 4.91 (br s, 2H). C NMR (CDCl , 75 MHz) δ (ppm): 151.6,
3
1
39.1, 135.0, 134.5, 130.4, 125.3, 94.5.
4
.1.4.5. N'-Hydroxy-3-(trifluoromethyl)benzimidamide (43). According to general procedure,
hydroxylamine (430 µL, 7.01 mmol) was reacted with 3-(trifluoromethyl)benzonitrile (234 µL,
2
5

1
.75 mmol) in absolute ethanol (6 mL) at 90 °C for 1.5 h to afford the desired compound 43.
ACCEPTED MANUSCRIPT
1
Pale yellow oil; yield: 350 mg, 98%; R = 0.45 (cyclohexane/ethyl acetate 1:1). H NMR (CDCl ,
f
3
3
7
00 MHz) δ (ppm): 8.77 (br s, 1H), 7.87–7.80 (m, 1H), 7.78–7.70 (m, 1H), 7.65–7.57 (m, 1H),
13
.50–7.41 (m, 1H), 4.87 (br s, 2H). C NMR (CDCl , 75 MHz) δ (ppm): 151.8, 133.4, 131.3 (d,
3
J = 32.6 Hz), 129.4, 129.3, 126.8 (q, J = 3.7 Hz), 124.0 (q, J = 272.3 Hz), 123.1 (q, J = 3.9 Hz).
4
.1.4.6. N'-Hydroxy-3-methylbenzimidamide (44). According to general procedure,
hydroxylamine (3.06 mL, 50.0 mmol) was reacted with 3-methylbenzonitrile (1.50 mL, 12.50
mmol) in absolute ethanol (8 mL) at 90 °C for 1 h to afford the desired compound 44. Colorless
1
oil; yield: 1.80 g, 96%; R = 0.46 (cyclohexane/ethyl acetate 1:1). H NMR (CDCl , 300 MHz) δ
f
3
1
3
(
ppm): 7.50–7.38 (m, 2H), 7.34–7.18 (m, 2H), 4.94 (br s, 2H), 2.37 (s, 3H). C NMR (CDCl ,
3
7
5 MHz) δ (ppm): 152.9, 138.4, 132.5, 130.8, 128.6, 126.7, 123.1, 21.5.
4
.1.4.7. N'-Hydroxy-3-methoxybenzimidamide (45). According to general procedure,
hydroxylamine (1.84 mL, 30.04 mmol) was reacted with 3-methoxybenzonitrile (1.0 g, 7.51
mmol) in absolute ethanol (8 mL) at 90 °C for 1.5 h to afford the desired compound 45;
Colorless solid; yield: 1.22 g, 98%; mp 104–105 °C; R = 0.30 (cyclohexane/ethyl acetate 1:4).
f
1
H NMR (CDCl , 300 MHz) δ (ppm): 7.39 (br s, 1H), 7.16–7.08 (m, 1H), 7.07–7.01 (m, 2H),
3
1
3
6
.85–6.69 (m, 1H), 4.96 (br s, 2H), 3.63 (s, 3H). C NMR (CDCl , 75 MHz) δ (ppm): 159.5,
3
1
52.9, 133.7, 129.6, 118.2, 115.9, 111.2, 55.2.
4
.1.4.8. N'-Hydroxy-4-methoxybenzimidamide (46). According to general procedure,
hydroxylamine (1.84 mL, 30.04 mmol) was reacted with 4-methoxybenzonitrile (1.0 g, 7.51
mmol) in absolute ethanol (8 mL) at 90 °C for 1.5 h to afford the desired compound 46.
Colorless solid; yield: 1.21 g, 97%; mp 121–122 °C; R = 0.30 (cyclohexane/ethyl acetate 1:4).
f
1
H NMR (CDCl , 300 MHz) δ (ppm): 7.57 (d, J = 8.9 Hz, 2H), 6.91 (d, J = 8.9 Hz, 2H), 4.90 (s,
3
1
3
2
H), 3.83 (s, 3H). C NMR (CDCl , 75 MHz) δ (ppm): 161.1, 152.7, 127.4, 125.1, 114.2, 55.5.
3
4
.1.4.9. N'-Hydroxyfuran-2-carboximidamide (49). According to general procedure,
hydroxylamine (50% w/w in water, 790 µL, 12.89 mmol) was reacted with 2-furonitrile (282 µL,
3
.22 mmol) in absolute ethanol (6 mL) at 90 °C for 1.5 h to afford the desired compound 49.
1
Colorless solid; yield: 400 mg, 99%; mp 57–58 °C; R = 0.32 (cyclohexane/ethyl acetate 1:1). H
f
NMR (CDCl , 300 MHz) δ (ppm): 8.67 (br s, 1H), 7.30 (dd, J = 1.7, 0.7 Hz, 1H), 6.69 (dd, J =
3
1
3
3
.4, 0.6 Hz, 1H), 6.29 (dd, J = 3.5, 1.8 Hz, 1H), 5.05 (br s, 2H). C NMR (CDCl , 75 MHz) δ
3
(ppm): 145.9, 145.5, 142.9, 111.6, 108.4.
2
6

4
.1.4.10. Thiophene-2-amidoxime (50). According to general procedure, hydroxylamine (50%
ACCEPTED MANUSCRIPT
w/w in water, 2.63 mL, 42.96 mmol) was reacted with 2-thiophenecarbonitrile (1.0 mL, 10.74
mmol) in absolute ethanol (8 mL) at 90 °C for 1.5 h to afford the desired compound 50.
1
Colorless solid; yield: 1.53 g, 100%; mp 90–96 °C; R = 0.52 (cyclohexane/ethyl acetate 1:1). H
f
NMR (CDCl , 300 MHz) δ (ppm): 7.52–7.08 (m, 3H), 7.01 (dd, J = 4.9, 3.9 Hz, 1H), 4.98 (br s,
3
1
3
2
H). C NMR (CDCl , 75 MHz) δ (ppm): 148.6, 134.9, 127.2, 126.8, 125.3.
3
4
.1.4.11. N'-Hydroxyacetimidamide (52). According to general procedure, hydroxylamine (50%
w/w in water, 25.0 mL, 384.0 mmol) was reacted with acetonitrile (5.0 mL, 96.0 mmol) in
absolute ethanol (50 mL) at 90 °C for 3h. The desired acetamidoxime 52 was recrystallized from
propan-2-ol. Colorless solid; yield: 7.10 g, 100%; mp 134–136 °C; R = 0.54 (CH Cl /MeOH
f
2
2
1
13
4
:1). H NMR ((CD ) SO, 300 MHz) δ (ppm): 8.76 (br s, 1H), 5.33 (br s, 2H), 1.62 (s, 3H). C
3 2
NMR ((CD ) SO, 75 MHz) δ (ppm): 149.6, 16.7.
3
2
4
.1.5. General procedure for the synthesis of boranyl intermediates (53-65)
Method A. Cesium carbonate (3 equiv) was added to a suspension of the appropriate
amidoxime (3 equiv) in tetrahydrofuran. The mixture was heated at 50 °C for 20 min and then
cooled to rt before adding, dropwise via a syringe, a solution of methyl ester (1 equiv) in
tetrahydrofuran. The final mixture was stirred for 12 h at 80 °C (TLC in
dichloromethane/methanol 7:3) and then standard workup was applied.
Method B. In anhydrous conditions, under an atmosphere of argon and in presence of 3Å
molecular sieves, the appropriate amidoxime (3 equiv) was suspended in dry tetrahydrofuran and
stirred at rt for 30 min. 60% NaH dispersion in mineral oil (3 equiv) was then added portionwise
and the resulting reaction mixture was heated at 50 °C for 20 min. A solution of methyl ester (1
equiv) in dry tetrahydrofuran was subsequently added dropwise via a syringe at rt and the final
suspension was stirred for 2 h at 80 °C (TLC in dichloromethane/methanol 7:3). Standard
workup was then applied.
Standard workup. The reaction was quenched with saturated aqueous solution of NaHCO3
and diluted with dichloromethane. The mixture was acidified with 1N HCl (pH = 1), the organic
phase removed and then extracted two more times with 1N HCl. The combined aqueous layers
were basified with Na CO and extracted with dichloromethane (× 3). Alternatively, the reaction
2
3
mixture was directly extracted three times with dichloromethane. Upon complete extraction, the
organic phases were combined, dried over anhydrous Na SO , filtered and concentrated under
2
4
vacuum. The residue was purified by silica gel column chromatography.
2
7

4
.1.5.1. Borane 3-(3-fluorophenyl)-5-(quinuclidin-3-ylmethyl)-1,2,4-oxadiazole complex (53).
ACCEPTED MANUSCRIPT
Following method A, methyl ester 38 (500 mg, 2.54 mmol) was reacted with 3-fluoro-N'-
hydroxy-benzimidamide 39 (1.17 g, 7.61 mmol) and Cs CO (2.48 g, 7.61 mmol). Standard
2
3
workup and silica gel column chromatography gained the desired compound. Pale yellow oil;
1
yield: 589 mg, 77%; R = 0.43 (cyclohexane/ethyl acetate 7:3). H NMR (CDCl , 300 MHz) δ
f
3
(ppm): 7.82–7.75 (m, 1H), 7.69 (ddd, J = 9.4, 2.6, 1.5 Hz, 1H), 7.39 (td, J = 8.0, 5.7 Hz, 1H),
7
.14 (tdd, J = 8.4, 2.6, 1.0 Hz, 1H), 3.39–3.24 (m, 1H), 3.09–2.87 (m, 6H), 2.62 (ddd, J = 13.6,
6
.9, 2.0 Hz, 1H), 2.57–2.43 (m, 1H), 1.97–1.85 (m, 2H), 1.85–1.74 (m, 2H), 1.74–1.59 (m, 1H).
1
3
C NMR (CDCl , 75 MHz) δ (ppm): 177.8, 167.6, 162.9 (d, J = 246.7 Hz), 130.7 (d, J = 8.1
3
Hz), 128.6 (d, J = 8.6 Hz), 123.2 (d, J = 3.2 Hz), 118.4 (d, J = 21.2 Hz), 114.5 (d, J = 23.7 Hz),
5
9.4, 53.6, 52.9, 33.1, 30.1, 26.2, 24.6, 20.0.
4
.1.5.2. Borane 3-(3-chlorophenyl)-5-(quinuclidin-3-ylmethyl)-1,2,4-oxadiazole complex (54).
Following method A, methyl ester 38 (400 mg, 2.03 mmol) was reacted with 3-chloro-N'-
hydroxybenzimidamide 40 (1.04 g, 6.09 mmol) and Cs CO (1.98 g, 6.09 mmol). Standard
2
3
workup and silica gel column chromatography gained the desired compound. Colorless solid;
1
yield: 609 mg, 95%; mp 159–160 °C; R = 0.42 (cyclohexane/ethyl acetate 7:3). H NMR
f
(
CDCl , 300 MHz) δ (ppm): 8.01–7.97 (m, 1H), 7.88 (dt, J = 7.5, 1.5 Hz, 1H), 7.42 (ddd, J = 8.1,
3
2
1
1
1
.0, 1.4 Hz, 1H), 7.35 (t, J = 7.8 Hz, 1H), 3.39–3.24 (m, 1H), 3.09–2.86 (m, 6H), 2.61 (ddd, J =
3.6, 6.9, 2.1 Hz, 1H), 2.57–2.42 (m, 1H), 1.99–1.85 (m, 2H), 1.85–1.73 (m, 2H), 1.73–1.59 (m,
1
3
H). C NMR (CDCl , 75 MHz) δ (ppm): 177.9, 167.6, 135.1, 131.5, 130.4, 128.4, 127.7,
3
25.6, 59.5, 53.7, 53.0, 33.2, 30.2, 26.3, 24.7, 20.1.
4
.1.5.3. Borane 3-(3-bromophenyl)-5-(quinuclidin-3-ylmethyl)-1,2,4-oxadiazole complex (55).
Following method A, methyl ester 38 (300 mg, 1.52 mmol) was reacted with 3-bromo-N'-
hydroxybenzimidamide 41 (982 mg, 4.57 mmol) and Cs CO (1.49 g, 4.57 mmol). Standard
2
3
workup and silica gel column chromatography gained the desired compound. Colorless solid;
1
yield: 490 mg, 89%; mp 155–156 °C; R = 0.42 (cyclohexane/ethyl acetate 7:3). H NMR
f
(
CDCl , 300 MHz) δ (ppm): 8.15 (t, J = 1.8 Hz, 1H), 7.96–7.89 (m, 1H), 7.57 (ddd, J = 8.0, 2.0,
3
1
.0 Hz, 1H), 7.29 (t, J = 7.9 Hz, 1H), 3.38–3.23 (m, 1H), 3.07–2.87 (m, 6H), 2.62 (ddd, J = 13.7,
6
.9, 2.1 Hz, 1H), 2.57–2.41 (m, 1H), 1.99–1.85 (m, 2H), 1.85–1.73 (m, 2H), 1.73–1.58 (m, 1H).
1
3
C NMR (CDCl , 75 MHz) δ (ppm): 177.9, 167.5, 134.4, 130.6, 128.7, 126.1, 123.1, 59.6, 53.7,
3
5
3.1, 33.3, 30.3, 26.4, 24.8, 20.2.
4
.1.5.4. Borane 3-(3-iodophenyl)-5-(quinuclidin-3-ylmethyl)-1,2,4-oxadiazole complex (56).
Following method A, methyl ester 38 (500 mg, 2.54 mmol) was reacted with N'-hydroxy-3-
2
8