Design, synthesis, and electrophysiological evaluation of NS6740 derivatives: Exploration of the structure-activity relationship for alpha7 nicotinic acetylcholine receptor silent activation

Article abstract of DOI:10.1016/j.ejmech.2020.112669

The α7 nicotinic acetylcholine receptor (nAChR) silent agonists, able to induce receptor desensitization and promote the α7 metabotropic function, are emerging as new promising therapeutic anti-inflammatory agents. Herein, we report the structure–activity

Full text of DOI:10.1016/j.ejmech.2020.112669

Journal Pre-proof
Design, synthesis, and electrophysiological evaluation of NS6740 derivatives:
Exploration of the structure-activity relationship for alpha7 nicotinic acetylcholine
receptor silent activation
Maria Chiara Pismataro, Nicole A. Horenstein, Clare Stokes, Marta Quadri, Marco De
Amici, Roger L. Papke, Clelia Dallanoce
PII:
S0223-5234(20)30641-3
DOI:
https://doi.org/10.1016/j.ejmech.2020.112669
EJMECH 112669
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 16 May 2020
Revised Date: 13 July 2020
Accepted Date: 14 July 2020
Please cite this article as: M.C. Pismataro, N.A. Horenstein, C. Stokes, M. Quadri, M. De Amici, R.L.
Papke, C. Dallanoce, Design, synthesis, and electrophysiological evaluation of NS6740 derivatives:
exploration of the structure-activity relationship for alpha7 nicotinic acetylcholine receptor silent
activation, European Journal of Medicinal Chemistry, https://doi.org/10.1016/j.ejmech.2020.112669.
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition
of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of
record. This version will undergo additional copyediting, typesetting and review before it is published
in its final form, but we are providing this version to give early visibility of the article. 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.
© 2020 Elsevier Masson SAS. All rights reserved.

Design, synthesis, and electrophysiological evaluation of
NS6740 derivatives: exploration of the structure-activity
relationship for alpha7 nicotinic acetylcholine receptor silent
activation
Maria Chiara Pismataro a, b, c, Nicole A. Horenstein b, Clare Stokes c, Marta Quadri
a, b, c, Marco De Amici a, Roger L. Papke c, Clelia Dallanoce a, *
a
Department of Pharmaceutical Sciences, Medicinal Chemistry Section “Pietro Pratesi”, University of
Milan, Via L. Mangiagalli 25, 20133 Milan, Italy
b Department of Chemistry, University of Florida, P.O. Box 117200, Gainesville, FL 32611-7200, USA
c
Department of Pharmacology and Therapeutics, University of Florida, P.O. Box 100267, Gainesville,
FL 32610-0267, USA
Keywords:
α7 nicotinic acetylcholine receptor
Electrophysiology
Silent agonist
Partial agonist
1,4-Diazabicyclo[3.2.2]nonane
Structure-activity relationship (SAR)
Abbreviations: ACh, acetylcholine; CAP, cholinergic anti-inflammatory pathway; diEPP, 1,1-
diethyl-4-(phenyl)piperazin-1-ium iodide; Di, PAM-insensitive nonconducting desensitized state; Ds,
PAM-sensitive nonconducting desensitized state; DMAP, 4-(dimethylamino)pyridine; DME, 1,2-
dimethoxyethane; GAT107, (3aR,4S,9bS)-4-(4-bromophenyl)-3a,4,5,9b-tetrahydro-3H-cyclopenta[c]quinoline-8-
sulfonamide; GTP, guanosine triphosphate; Hn+2PnO3n+1, polyphosphoric acid; IP3, inositol triphosphate;
nAChR, nicotinic acetylcholine receptor; NS6740, 1,4-diazabicyclo[3.2.2]nonan-4-yl(5-(3-
(trifluoromethyl)-phenyl)-furan-2-yl)methanone;
p-CF3-diEPP,
1,1-diethyl-4-(4-
(trifluoromethyl)phenyl)piperazin-1-ium iodide; PAM, positive allosteric modulator; Pd(PPh3)4,
tetrakis(triphenylphoshine)palladium(0); PNU-120596, N-(5-chloro-2,4-dimethoxyphenyl)-N'-(5-methyl-
3-isoxazolyl)-urea; s, second; SARs, structure‒activity relationships; SEM, standard error of the
mean; TEA, triethylamine; TNF-α, tumor necrosis factor-α; triEMA, triethylmethylammonium; TsCl,
tosyl chloride.
* Corresponding author.
E-mail address: clelia.dallanoce@unimi.it (C. Dallanoce)
1

A B S T R A C T
The α7 nicotinic acetylcholine receptor (nAChR) silent agonists, able to induce receptor
desensitization and promote the α7 metabotropic function, are emerging as new promising
therapeutic anti-inflammatory agents. Herein, we report the structure–activity relationship
investigation of the archetypal silent agonist NS6740 (1,4-diazabicyclo[3.2.2]nonan-4-yl(5-(3-
(trifluoromethyl)-phenyl)-furan-2-yl)methanone) (1) to elucidate the ligand-receptor interactions
responsible for the α7 silent activation. In this study, NS6740 fragments 11-16 and analogs 17-32
were designed, synthesized, and assayed on human α7 nAChRs expressed in Xenopus laevis
oocytes with two-electrode voltage clamping experiments. All together the structural portions of
NS6740 were critical to engender its peculiar activity profile. The diazabicyclic nucleus was
essential but not sufficient for inducing α7 silent activation. The central hydrogen-bond acceptor
core and the aromatic moiety were crucial for promoting prolonged α7 receptor binding and
sustained desensitization. Compounds 13 and 17 were efficacious partial agonists. Compounds
12, 21, 23-26, and 30 strongly desensitized α7 nAChR and therefore may be of interest for
additional investigation of inflammation responses. We gained key structural information useful
for further silent agonist development.
2

1. Introduction
The α7 nicotinic acetylcholine receptor (nAChR) is a ligand-gated ion channel composed of
five identical α7 subunits, and it is the second most abundant nAChR subtype in the human
brain, where it mainly exists in the hippocampus, cerebral cortex, and other regions associated
with learning, memory, and cognition [1−4]. The α7 nAChR is also expressed in the peripheral
nervous system and in tissues and cells that are not associated with neuronal function [5,6]. The
α7 subtype exhibits unique physiological and pharmacological properties, including high
calcium permeability, rapid desensitization, fast onset and decay kinetics, and a very low
probability of channel opening [7,8]. The distinctive α7 activation and desensitization have been
investigated in the presence and absence of the α7 type II positive allosteric modulator (PAM) N-
(5-chloro-2,4-dimethoxyphenyl)-N'-(5-methyl-3-isoxazolyl)-urea (PNU-120596), and two major
forms of α7 desensitization have been identified [9]. The desensitized Di state is insensitive to
the PAM, whereas the Ds state is sensitive to and destabilized by PNU-120596 and other type II
PAMs. The Ds state is probably associated with the rapid form of desensitization unique to α7
that is connected to channel open states in presence of type II modulators. Several
electrophysiological and calcium imaging studies, together with the analysis of mutant receptors
have shown the ability of α7 nAChRs to operate both via canonical ionotropic cation influx
through the open-channel state, which due to the high calcium permeability of the channel can
lead to calcium-induced calcium release, and via non-conventional metabotropic channel
signaling, possibly through direct Gq protein coupling that promotes the inositol triphosphate
(IP3)-mediated calcium release [10−13]. In neurons, both ionotropic and metabotropic α7
functions have been observed, and it has been suggested that a structural transition of the channel
opening induced by agonist ligand binding enables a metabotropic signaling states which seem to
coincide with channel desensitization [14,15]. However, the metabotropic α7 signal transduction
behavior is most relevant in cell types such as immune cells, where receptor-mediated ion
currents are not observed [16,17]. Indeed, the α7 nAChR modulates the release of pro-
inflammatory cytokines from monocytes, microglia, and macrophages. These anti-inflammatory
effects are not due to classical channel activation through ionotropic conductance [18,19]; rather,
for this receptor a large interactome was identified, members of which may trigger the ability to
operate metabotropically [20]. Based on mutational studies, a G protein-binding region within
the M3-M4 loop has been implicated to be responsible for a direct coupling of the α7 nAChR to
GTP-binding proteins, which enables a downstream calcium signaling response that can persist
beyond the expected time course of α7 channel activation [21]. In microglia for example, the α7
nAChRs attenuate neuroinflammation and regulate TNF-α release through the interaction with
the heterotrimeric G protein subunit α (Gαi) signaling pathway [22].
3

A significant number of studies indicate that α7 nAChR-dependent signals can be activated
via vagus nerve stimulation as part of the cholinergic anti-inflammatory pathway (CAP), leading
to regulation of cytokine production and related tissue injury response, supporting the idea that
CAP is a robust physiological mechanism [23]. It is interesting to consider that α7-activating
nicotinic agents might be used as molecular tools to investigate CAP activity and prevent or
modulate the development of cytokine-mediated diseases. The distinction between ionotropic
and metabotropic receptor activities becomes essential to the development of therapeutics
targeting α7 nAChRs, and efforts should be made to identify compounds that are able to
selectively regulate α7 function [24]. In this direction, there is a growing interest in silent
agonists [25−32], a class of ligands that promote entry of the receptor into a desensitized state
with respect to ionotropic activity. Thus they may represent an alternative therapeutic strategy to
modulate anti-inflammatory responses [20,25−27]. A number of α7 silent agonists have been
described so far: NS6740 (1) [28], KC-1 (2) [29], (R)-47 (3) [30], triEMA (4) [31], TEA (5) [31],
diEPP (6) [31] and its analog p-CF3-diEPP (7) [27,32], the sulfonium salt 8 [33], the spirocyclic
quinuclidine 9 and its analogues [34], and the quinuclidine oxadiazole ammonium salt 10 [35]
(Fig. 1). Silent agonists are able to induce receptor desensitization that responds to type II PAMs
[9]. In the absence of PAMs, silent agonists appear to operate almost entirely through the
metabotropic mode and exert little or no ionotropic activity.
Fig. 1. Structures of known α7 nAChR silent agonists.
4

The archetypal silent agonist 1,4-diazabicyclo[3.2.2]nonan-4-yl(5-(3-(trifluoromethyl)-
phenyl)-furan-2-yl)methanone (NS6740, 1, Fig. 1) was first published in a model study for
cognitive improvement, in which it not only failed to enhance avoidance learning but
antagonized the effects of an efficacious α7 agonist. Although 1 displayed little efficacy at α7
nAChR when applied alone, its agonist properties were revealed in combination with the type II
PAM PNU-120596 [28]. Indeed, NS6740 is a very weak partial agonist that produces little
channel activation on its own but has been shown to stabilize long-lived nonconducting receptor
states that are sensitive to activation by allosteric modulators (Ds). The α7 receptor
desensitization induced by NS6740 is so remarkably prolonged that when the PAM was applied
alone to the receptor after NS6740, the effect persisted even after an hour of washout from the
bath [36]. NS6740 is one of the most promising anti-inflammatory compounds to target the α7
nAChR. It has been proven to reduce the lipopolysaccharide-stimulated secretion of tumor
necrosis factor-α (TNF-α) from rat cultured microglia [19] and is also effective in reducing
inflammation in mouse models of peripheral neuropathy and tonic inflammatory pain [26]. The
time and concentration dependence of these analgesic effects is consistent with NS6740
induction of non-conducting receptor states, thus indicating that signaling pathways required for
anti-inflammatory responses are fulfilled by α7 receptors in the desensitized conformation [26].
The structure of NS6740 shows a protonatable nitrogen atom within the 1,4-
diazabicyclo[3.2.2]nonane system, a hydrogen-bond acceptor heterocyclic moiety, and an
aromatic ring (Fig. 2B). Data suggest that the silent binding site is an extension of the orthosteric
binding site [31,36]; however, some aspects of the ligand-receptor interaction for α7 silent
modulation are yet to be clarified. Additionally, recent data [37] have indicated that some silent
agonists may bind to an extracellular allosteric site, shared with ago-PAMs like GAT107 that can
couple to the transmembrane PAM site to produce PAM-dependent activation of receptors in the
Ds states. The pharmacophoric features of NS6740, which are common to other silent agonists
(Fig. 1), make NS6740 capable of generating multiple interactions in the silent agonist binding
sites, allowing long lasting effects on the conformational equilibrium of receptor states.
Expanding our previous research on the α7 nAChRs [29,31−46], we broadly investigated the
structure-activity relationships (SARs) of the lead structure 1 with the goal to elucidate the
pharmacophore features for silent agonism and the ligand-receptor interactions that are able to
induce the α7 desensitized states. Human α7 nAChRs were expressed in Xenopus laevis oocytes,
and two electrode voltage clamping was employed to establish the electrophysiological profile of
the new compounds in the absence and presence of a type II PAM. We determined channel
activation, inhibition of acetylcholine (ACh) responses after the compounds were applied alone,
and responses potentiated by the PAM PNU-120596. These insights would allow the design of
new compounds which may be useful for future studies to deepen the α7 metabotropic signaling
5

and to selectively target the inflammatory responses derived from this non-conventional α7
function, while importantly minimizing the ionotropic activity.
2. Results and Discussion
2.1. Design
On this ground, we synthesized NS6740 fragments (compounds 11-16) (Fig. 2A) and
analogs (compounds 17-32) (Fig. 2B) to deduce SARs associated with their electrophysiological
properties, paying particular attention to the propensity of compounds to desensitize the receptor
with minimal channel activation.
Fig. 2. Chemical modifications performed on the lead compound NS6740 (1). (A) Structures of
compounds 11-16 derived from NS6740 skeleton fragmentation. (B) New compounds 17-32 obtained by
manipulations on the three NS6740 pharmacophoric substructures.
Following a fragmentation approach, the NS6740 skeleton was dissected in molecular
portions (11-16, Fig. 2A) to clarify the influence of each structural element and confirm the
6

proposed pharmacophore interaction features. We synthesized compound 11, lacking the
diazabicyclic system, to evaluate the role of the protonatable nitrogen atom. Furthermore, we
prepared compounds 12-16, characterized by progressive removal of structural parts, i.e.
trifluoromethyl group (12), phenyl ring (13), furan nucleus (14), carbonyl function (15), and
ethyl chain (16), which could imply loss or impairment of crucial interactions with the silent
binding site. Fragments 15 and 16 retain the positively charged moiety as the only residual
pharmacophore element. Then, for an alternate SAR study, we designed and synthesized a series
of novel NS6740 analogs (compounds 17-32, Fig. 2B), focusing on structural modifications of
the three pharmacophoric portions of its scaffold. Concerning the investigation of the NS6740
positively charged center, we examined the effect of quaternization at the 1,4-
diazabicyclo[3.2.2]nonane nitrogen atom tertiary amine (17) since the permanent charge could
generate stronger ionic interactions and lead to different conformational receptor states. Then the
regioisomeric 1,3-diazabicyclo[3.2.2]nonane system (18) was introduced to evaluate how the
distance between the positively charged nitrogen and the carbonyl group could influence the
proper ligand orientation responsible for inducing receptor desensitization. We also investigated
the different spatial and electrostatic interactions due to the smaller enantiomeric (S,S)- and
(R,R)-2,5-diazabicyclo[2.2.1]heptane nucleus ((S,S)-19 and (R,R)-19)), characterized by a
secondary amine nitrogen. Focusing next on the central portion of NS6740 with hydrogen
bonding capability, we explored the contribution of the carbonyl group by removing it in
derivative 20. Lacking the amide bond, this modification introduced also more flexibility into the
NS6740 together with an additional protonatable nitrogen atom. Finally, we explored chemical
modifications of the pharmacophoric aryl moiety, initially focusing on the meta-substitution. The
influence of the trifluoromethyl group was evaluated by replacement with its bioisosteric methyl
group (21) and by shifting it from the meta to para position (22). The trifluoromethyl group was
also replaced by fluorine (23), chlorine (24), bromine (25), and iodine (26), to assess the effect of
halogen atoms with increasing size. The influence of polar substituents, such as hydroxy (27),
amino (28), and nitro (29) groups, was investigated as well. Then we tested the replacement of
the trifluoromethyl phenyl moiety by the heterocyclic furan ring (30) and by bulky aromatic
groups, namely naphthalene (31) and isoquinoline (32), to explore the effect of the increased
steric hindrance on ligand-receptor interactions.
2.2. Chemistry
The diazabicyclic systems 16 and 37 were synthesized by applying minor changes and
optimizations to a patented procedure [47] (Scheme 1). The commercially available 3-
7

quinuclidinone hydrochloride 33 was reacted with hydroxylamine, and the resulting oxime 34
underwent Beckmann rearrangement upon treatment with polyphosphoric acid at 130 °C to give
the lactam regioisomers 35 and 36, which were then separated by silica gel column
chromatography and obtained in a 6:1 ratio. Reduction with lithium aluminum hydride yielded
the diazabicyclic derivatives 16 and 37 which, due to their volatility, were isolated as
dihydrochlorides after treatment with a 4 M solution of hydrochloric acid in 1,4-dioxane.
Scheme 1. Synthesis of diazabicyclic systems 16 and 37. Reagents and conditions: (i) NH2OH·HCl (1.15
equiv), CH3COONa (3 equiv), H2O, 70 °C, 2 h; (ii) Hn+2PnO3n+1, 130 °C, 20 h; (iii) 1 M LiAlH4 in THF
(1.2 equiv), 0 °C to reflux, 8 h; (iv) 4 M HCl in 1,4-dioxane (2 equiv), rt, overnight.
The synthesis of the two enantiomeric diazabicyclic systems (S,S)-43 and (R,R)-43 was
performed according to Scheme 2 in a similar manner as described earlier [48]. Trans-4-
hydroxy-L-proline (2S,4R)-38 was N-tosylated by reaction with tosyl chloride in the presence of
sodium carbonate to produce intermediate (2S,4R)-39, which was reduced with borane in THF,
leading to the corresponding diol (2S,4R)-40. Subsequent ditosylation of (2S,4R)-40 in the
presence of DMAP in pyridine afforded the tritosyl derivative (2S,4R)-41 together with
ditosylchloromethyl side product (ratio about 3:1), which were separated by silica gel column
chromatography. The cyclization of (2S,4R)-41 was achieved with benzylamine in MeOH heated
in a sealed tube at 90 °C to form the (1S,4S)-2-benzyl-5-tosyl-2,5-diazabicyclo[2.2.1]heptane
(S,S)-42. Treatment of (S,S)-42 with hydrobromic acid in acetic acid allowed cleavage of the
tosyl group and led to the diazabicyclic moiety (S,S)-43. The enantiomer (R,R)-43 was prepared
starting from cis-4-hydroxy-D-proline (2R,4R)-44, which was converted to methyl ester (2R,4R)-
45 as the hydrochloride by esterification with aqueous hydrochloric acid in methanol at reflux.
Ditosylation of (2R,4R)-45 in pyridine at 0 °C in the presence of trimethylamine, followed by
configurational inversion of the 4-hydroxy moiety of ditosyl intermediate (2R,4R)-46 with
tetraethylammonium acetate tetrahydrate in toluene under refluxing conditions, provided the 4-
acetyl derivative (2R,4S)-47. Saponification of (2R,4S)-47 in THF with a 0.5 M aqueous
potassium hydroxide solution at room temperature directly afforded the N-tosyl-trans-4-
hydroxy-
D-proline (2R,4S)-39, which is the enantiomer of compound (2S,4R)-39. We obtained
8

the (1R,4R)-2-benzyl-2,5-diazabicyclo[2.2.1]heptane (R,R)-43 by carrying intermediate (2R,4S)-
39 through the same reaction sequence as that described for (2S,4R)-39 (Scheme 2).
Scheme 2. Synthesis of enantiomeric diazabicyclic systems (S,S)-43 and (R,R)-43. Reagents and
conditions: (i) TsCl (1.2 equiv), Na2CO3 (2.1 equiv), H2O, 0 °C to rt, 48 h; (ii) 1 M BH3·THF complex
(2
DMAP (0.2 equiv), anhydrous CH2Cl2, 0 °C to rt, 24 h; (iv) benzylamine (10 equiv), MeOH, 90 °C,
overnight; (v) 33% HBr/AcOH, 50 °C, 2 5 h; (vi) 1.25 M HCl in MeOH, 0 °C to reflux, 6 h; (vii) TsCl
‒3 equiv), anhydrous THF, 0 °C to rt, 6‒15 h; (iii) TsCl (2‒5 equiv), anhydrous pyridine (4.8 equiv),
‒
(6.6 equiv), anhydrous pyridine, Et3N (3 equiv), 0 °C, 15 h; (viii) tetraethylammonium acetate
tetrahydrate (1.3 equiv), anhydrous toluene, reflux, 4 h; (ix) 0.5 M KOH in H2O (5 equiv), THF, rt, 3 h.
The synthetic route for the preparation of the 5-arylfuran-2-carbonyl chloride intermediates 62a-
m and 63-65 is outlined in Scheme 3. Methyl 5-bromofuran-2-carboxylate 49 was prepared by
refluxing the commercially available 5-bromofuran-2-carboxylic acid 48 with concentrated
sulfuric acid in methanol. To introduce the planned aromatic substituents, intermediate 49 was
treated with different commercially available aryl boronic acids 50a-g, 50i, 50l, 51-53 under
Suzuki-Miyaura
cross-coupling
reaction
conditions,
using
tetrakis(triphenylphoshine)palladium(0) as the catalyst in the presence of sodium carbonate in
dimethoxyethane or dimethylformamide to afford compounds 54a-g, 54i, 54l, 55-57 in 60-90%
yield [49]. Due to the reactivity of iodine in Suzuki conditions, intermediate 54h was prepared
from the amino derivative 54l in a one-pot procedure via diazonium salt followed by substitution
with potassium iodide [50]. Subsequent saponification of 54a-l, 55-57 yielded the 5-arylfuran-2-
carboxylic acid intermediates 58a-l, 59-61, which, in addition to the commercial meta-nitro
9

phenyl compound 58m, were treated with thionyl chloride at reflux to give the corresponding
carbonyl chloride derivatives 62a-m, 63-65.
Scheme 3. Synthesis of carbonyl chlorides 62a-m and 63-65. Reagents and conditions: (i) conc H2SO4 (4
equiv), MeOH, reflux, overnight; (ii) Pd(PPh3)4 (0.05 equiv), Na2CO3 (4.2 equiv), anhydrous DME or
DMF, reflux, 3
6 h; (iv) 1.5 M NaOH (1.5 equiv), THF, rt, overnight or 2.5 M LiOH (4 equiv), THF, rt, 48 h for 58i; (v)
SOCl2 (23 equiv), 0 °C to reflux, 3 6 h.
‒12 h; (iii) NaNO2 (1.05 equiv), conc H2SO4 (0.2 mL/mmol), KI (1.5 equiv), water, 0 °C,
‒
NS6740 fragments 11-15 were prepared according to Scheme 4. The reaction between the
carbonyl chloride intermediate 62a and diethylamine in dichloromethane afforded the diethyl
amide derivative 11. Treatment of the 1,4-diazabicyclo[3.2.2]nonane dihydrochloride 16 with
carbonyl chloride intermediate 62b, the commercially available furan-2-carbonyl chloride 67 or
acetyl chloride 68, in the presence of cesium carbonate in dry dimethylformamide, provided
compounds 12-14. Subsequent reduction of 14 using lithium aluminum hydride in diethyl ether
gave the N-ethyl diazabicyclic fragment 15. The target compounds 12-15 were converted into
their corresponding fumarates by treatment with fumaric acid in methanol.
10

Scheme 4. Synthesis of NS6740 fragments 11-15. Reagents and conditions: (i) anhydrous TEA (2 equiv),
anhydrous CH2Cl2, rt, 1.5 h; (ii) Cs2CO3 (5 equiv), anhydrous DMF, 0 °C to rt, overnight; (iii) LiAlH4
(2.25 equiv), anhydrous Et2O, 0 °C to reflux, 3 h; (iv) fumaric acid (1 equiv), MeOH, rt, overnight.
The synthesis of NS6740 (1) and the analogs 17-20 is depicted in Scheme 5. The carbonyl
chloride intermediate 62a reacted with the diazabicyclic compound 16 to give 1, which was
converted into the corresponding methyl ammonium salt 17 by treatment with methyl iodide in
methanol. The carbonyl group of 1 was reduced with lithium aluminum hydride in dry diethyl
ether at reflux and yielded compound 20. Compounds 18, (S,S)-19 and (R,R)-19 were
synthesized by acylation reaction of the carbonyl chloride 62a with the 1,3-
diazabicyclo[3.2.2]nonane dihydrochloride 37, and the enantiomeric azabicyclic systems (S,S)-
43 and (R,R)-43 respectively, under basic conditions at room temperature. The acylation step
was followed by deprotection of the benzyl group by catalytic hydrogenation to give the
enantiomers (S,S)-19 and (R,R)-19. The free bases 1, 18, (S,S)-19, (R,R)-19 and 20 were then
treated with fumaric acid to provide the corresponding fumarates. Finally, NS6740 derivatives
21-32 were obtained by acylation reaction of the 1,4-diazabicyclo[3.2.2]nonane dihydrochloride
16 with the appropriate carbonyl chlorides 62c-m, 63-65 under similar conditions as mentioned
above. Again, we converted the analogues 21-32 into the corresponding fumarates by reaction
with fumaric acid in methanol (Scheme 6).
11

Scheme 5. Synthesis of NS6740 derivatives 17, 18, (S,S)-19, (R,R)-19 and 20. Reagents and conditions:
(i) Cs2CO3 (5 equiv), anhydrous DMF, rt, overnight; (ii) MeI (8 equiv), MeOH, rt, overnight; (iii) LiAlH4
(2.25 equiv), anhydrous Et2O, 0 °C to reflux, 2 h; (iv) anhydrous TEA (1.1 equiv), anhydrous 1,4-
dioxane, rt, 3 h; (v) H2, Pd/C (10%), anhydrous MeOH, rt, overnight; (vi) fumaric acid (1 equiv), MeOH,
rt, overnight.
Scheme 6. Synthesis of NS6740 derivatives 21−32. Reagents and conditions: (i) Cs2CO3 (5 equiv),
anhydrous DMF, rt, overnight; (ii) fumaric acid (1 equiv), MeOH, rt, overnight. For structures 62c-m, see
Scheme 3, for structures 21-32, see Fig. 2.
12

2.3. Electrophysiology
The NS6740 fragments (11-16) and the new derivatives (17-32) were assayed on human α7
nAChRs expressed in Xenopus laevis oocytes with two-electrode voltage clamping [51]. The
responses evoked at the α7 nAChR by the tested derivatives were measured both as peak-current
and net-charge values at a holding potential of -60 mV. During drug applications at high
concentrations, the α7 receptor currents were progressively more synchronized prior to the
complete exchange of solution [52]. Therefore, to analyze and compare our results, we used net-
charge values, which are less affected by channel synchronization than peak currents and have
greater physiological relevance and scientific validity. Indeed, high levels of agonist binding site
occupancy induce concentration-dependent desensitization of the α7 receptor subtype, which
affects the relationship between peak current and net charge. For each derivative tested, the net-
charge values represent the average of the responses from at least four different oocytes. For
each oocyte, the test response was normalized to the average of two initial 60 µM ACh pre-
controls (Fig. 3).
As previously described [26,36], NS6740 displays particularly prolonged receptor binding
and a strong ability to stabilize the desensitized states of α7 nAChR. To analyze this
electrophysiological profile of NS6740 and investigate the newly synthesized derivatives 11-32
as potential long-lasting desensitizers, we employed protocol I in which each compound was
applied alone to evaluate its ability to promote channel opening. Then a 60 μM ACh post-control
was delivered to verify the ability of the previously applied test drug to remain in the binding site
and induce a prolonged desensitized state. Next there was an application of 10 μM of the type II
PAM PNU-120596 to reveal the residual desensitized states induced by the test drug.
Additionally, to highlight the induction of short-lasting desensitization, NS6740 and compounds
11-32 were investigated following protocol II in which the test drug was delivered in co-
application with 10 μM of the type II PAM PNU-120596. Short-lasting silent agonists, in fact,
do not cause either channel activation or prolonged desensitization when tested alone, but they
give rise to channel potentiation only in co-application with PNU-120596. In Fig. 3 protocols I
and II are illustrated for 10 μM applications of NS6740. Note that the colored bars above the
applications in Fig. 3 serve as a key to the bar graphs in Fig. 4−7.
13

Fig. 3. Electrophysiological responses of NS6740 (1) at 10 μM. Data for each cell were normalized to the
average of the peak-current amplitude of two initial 60 μM ACh controls. The solid lines are the averages,
and the shaded areas are the standard error of the mean for each of the 10,322 data points. Each trace
segment is 206.44 s long. Protocol I: cells (n = 7) treated first with 10 μM NS6740 (green bar) and then
with 60 μM ACh post-control (red bar), followed by 10 μM PNU-120596 (blue bar), and finally 60 μM
ACh post-control. Protocol II: cells (n = 5) treated first with 10 μM NS6740 plus 10 μM PNU-120596
(yellow bar), followed by 60 μM ACh post-control (purple bar).
The electrophysiological data of NS6740 (1) and compounds 11-32 are illustrated in Fig.
4−7 as net-charge responses, normalized to the average of the initial ACh controls on the vertical
axis of the bar graphs. The ACh post-control net-charge responses of protocol II are not included
in Fig. 5−7 since not relevant to the discussion. For both net-charge and peak-current values see
Supplementary data, Table S1-S4.
Fig. 4 shows the electrophysiological results for NS6740 assayed at 1, 3, 10, and 30 μM. It
displayed very weak partial agonist behavior at all four concentrations tested (Fig. 4A and
Supplementary data, Table S1). Regarding the inhibition of the ACh response, a strong reduction
of the post-control activity was displayed after application of NS6740 at 10 and 30 μM (net
charges 0.19 ± 0.06 and 0.18 ± 0.04, respectively, Fig. 4B) due to the ability of the compound to
promote deep receptor desensitization. Interestingly, the intensely potentiated response given by
PNU-120596 alone was observed even after 3 μM NS6740 application (net charge 123.21 ±
71.01, 111.06 ± 37.18, and 57.41 ± 15.72, respectively at 3 μM, 10 μM, and 30 μM, Fig. 4C),
thus revealing the prolonged residence time of NS6740 in the α7 receptor for drug concentrations
≥ 3 μM. Although 3 μM NS6740 showed a remarkable residual receptor potentiation by PNU-
120596 (Fig. 4C), it did not provide a strong ACh post-control inhibition (Fig. 4B). Presumably
14

only a fraction of receptors are bound by the drug at low concentrations, or the drug may be
bound at only a fraction of the sites. It may be that a lower fraction occupancy of the receptors is
sufficient to generate a residual potentiated current due to prolonged receptor binding without
affecting the ACh post-control response. The direct potentiated response obtained at 30 μM (net
charge 2.73 ± 0.73, Fig. 4D) was much lower (P = 0.0009) than the corresponding residual one
(net charge 57.41 ± 15.72, Fig. 4C). This may be explained by the ability of higher
concentrations of NS6740 to favor the faster stabilization of Di over Ds. Indeed, when the PAM
was delivered few minutes after NS6740 in the first experiment (Fig. 4C), the α7 receptor had
enough time to allow the transition from the Di to the Ds state so that higher potentiation was
observed. This confirmed what was previously described by Papke et al. [26]. Finally, the
reduced ACh post-control response after co-application of 10 or 30 μM NS6740 with 10 μM
PNU-120596 verified the sustained α7 receptor desensitization (Fig. 4E).
Fig. 4. Electrophysiological data of NS6740 (1) at 1, 3, 10, and 30 μM normalized to the average
responses of the first two 60 μM ACh controls. Protocol I: (A) Direct activation by 1; (B) 60 μM ACh
post-control activation after the delivery of 1 indicating residual inhibition; (C) Residual potentiation by
10 μM PNU-120596 after the delivery of 1 and the 60 μM ACh post-control. Protocol II: (D) Direct
potentiated response given by co-application of 1 plus 10 μM PNU-120596; (E) 60 μM ACh post-control
activation.
The new NS6740 analogs 11-32 were analyzed at 10 μM following the previously described
protocols I and II. The electrophysiological profiles of the NS6740 fragments 11-16 are
illustrated in Fig. 5 (see also Supplementary data, Table S2). Fragment 11, in which the basic
nucleus was removed, and fragments 15 and 16, in which only the basic nucleus was retained,
proved to be completely inactive in terms of channel activation, ACh post-control inhibition, and
15

receptor potentiation (Fig. 5A‒D). Derivative 12, which lacks the meta trifluoromethyl group,
showed increased partial agonist properties (net charge 0.54 ± 0.02, Fig. 5A) compared to the
parent compound 1 (net charge 0.19 ± 0.06, brown bar) (P = 0.0003). Interestingly, 12 was able
to strongly inhibit the ACh post-control (net charge reduced to 0.14 ± 0.07, Fig. 5B) while
producing a large residual potentiated response (net charge 18.33 ± 5.13, Fig. 5C), thus proving
to be effective at inducing the Ds state. Fragment 13, without the phenyl moiety, displayed
efficacious partial agonist behavior with a net charge of 0.71 ± 0.04 (≈ 70% of ACh response,
Fig. 5A) and in co-application with PNU-120596 provided a direct PAM-potentiation (Fig. 5D)
in line with what is expected for a partial agonist. The lack of residual potentiated response to
PNU-120596 (Fig. 5C) by compound 13 indicated a failure to remain in the receptor long
enough to stabilize a prolonged non-conductive state. Fragment 14, retaining the carbonyl group
and the diazabicyclic system, exhibited short-lasting silent agonist characteristics. Indeed, 14 did
not give channel opening when tested alone (Fig. 5A) yet induced potentiation in co-application
with PNU-120596 (Fig. 5D), thus proving its silent agonist behavior. The absence of the
aromatic portion of the NS6740 pharmacophore resulted in a short-lasting interaction with the
binding site that was evidenced by the full ACh post-control response (Fig. 5B) and by the lack
of a residual PNU-120596-potentiated response (Fig. 5C).
Fig. 5. Electrophysiological data of fragments 11-16 at 10 μM. Protocol I: (A) Channel opening given by
11-16; (B) 60 μM ACh post-control activation after the delivery of 11-16; (C) Residual potentiated
response given by application of 10 μM PNU-120596 after delivery of 11-16 and the 60 μM ACh post-
control. Protocol II: (D) Direct potentiated response given by 11-16 plus 10 μM PNU-120596.
16

The electrophysiological profiles of compounds 17-20 are reported in Fig. 6 (see also
Supplementary data, Table S3). The methylammonium salt of NS6740 17 displayed efficacious
partial agonist behavior with a net charge of 0.79 ± 0.03 (≈ 80% of ACh response, Fig. 6A). It
was not able to stabilize a prolonged non-conductive state and displayed a very small residual
potentiated response (net charge 1.39 ± 0.86, Fig. 6C). Although the response was not strong, it
induced a PAM-potentiation in line with what is expected for a partial agonist in co-application
with PNU-120596 (Fig. 6D). In comparison with NS6740, compound 17 had an increased
channel opening ability, a shorter residence time in the receptor binding site, and a reduced
effectiveness in stabilizing the α7 desensitized state. The 1,3-diazabyciclo[3.2.2]nonane
regioisomer of NS6740 18 proved to be completely inactive in all the test conditions. The pair of
enantiomers (S,S)-19 and (R,R)-19 exhibited large potentiated responses (15.22 ± 2.95 and 43.70
± 3.30, respectively) in co-application with PNU-120596 (Fig. 6D). Given the absence of agonist
responses (Fig. 6A) as well as of ACh post-control inhibitions and residual potentiations (Fig. 6B
and Fig. 6C, respectively), we conclude that (S,S)-19 and (R,R)-19 behaved as silent agonists
with readily reversible binding. The deletion of the carbonyl group (compound 20) was
detrimental to the activity, since this derivative behaved as a very weak partial agonist (Fig. 6A)
without affecting the ACh post-control activity (red bar), and did not show any residual receptor
PAM potentiation (Fig. 6C). Additionally, even in co-application with PNU-120596, 20 showed
a negligible potentiation.
Fig. 6. Electrophysiological data of derivatives 17-20 at 10 μM. Protocol I: (A) Channel opening given by
17-20; (B) 60 μM ACh post-control activation after the delivery of 17-20; (C) Residual potentiated
17

response given by application of 10 μM PNU-120596 after delivery of 17-20 and the 60 μM ACh post-
control. Protocol II: (D) Direct potentiated response given by 17-20 plus 10 μM PNU-120596.
The profiles of compounds 21-32 at the α7 nAChR are summarized in Fig. 7 (see also
Supplementary data, Table S4). Generally, the replacement of the original meta trifluoromethyl
group on the aromatic ring conferred partial agonist activity to the new derivatives 21-32,
associated with a greater propensity of channel opening than the parent NS6740 (Fig. 7A),
whereas both residual and direct PNU-120596-related responses were higher (Fig. 7C 0.0012 < P
< 0.0445 and Fig. 7D 0.00005 < P < 0.0101, respectively) for the lead compound than for 21-32.
In detail, the m-methyl derivative 21 behaved as a weak partial agonist (Fig. 7A), displayed
strong ACh post-control inhibition (net charge 0.20 ± 0.04, Fig. 7B), and a residual potentiated
response (net charge 30.74 ± 8.77, Fig. 7C) higher (P = 0.0283) than the direct potentiation in
co-application with PNU-120596 (Fig. 7D). Moving the trifluoromethyl group from the meta to
para position (compound 22) caused the loss of the peculiar NS6740 electrophysiological
profile. 22 promoted more channel opening (net charge 0.40 ± 0.03, Fig. 7A), approximately 2-
fold greater than that of the lead 1 (brown bar in Fig. 7A). Moreover, 22 did not show strong
inhibition of ACh post-control (net charge 0.64 ± 0.06, Fig. 7B), with residual and direct
potentiation (net charges 3.23 ± 0.65 and 6.25 ± 1.40, Fig. 7C and Fig. 7D, respectively) much
lower than the parent compound 1. The introduction of halogen atoms in the meta position of the
phenyl ring, fluorine (23), chlorine (24), bromine (25), and iodine (26), gave weak partial agonist
responses ranging from 25% to 50% of ACh control (Fig. 7A). Moreover, 23-26 showed strong
ACh post-control inhibition, with almost total absence of the ACh response after the application
of the chlorine-containing derivative 24 (net charge 0.04 ± 0.01, Fig. 7B). The large residual
PAM-potentiated responses (Fig. 7C) suggested that halogen derivatives 23-26 induced
prolonged α7 desensitization. Regarding the polar substituents, the m-hydroxy derivative 27
displayed weak partial agonist response (Fig. 7A), inhibition of ACh post-control (Fig. 7B), and
remarkable residual PAM-potentiation (Fig. 7C). The m-amino compound 28 behaved as a
typical short-lasting silent agonist, displaying very low agonist response (net charge 0.10 ± 0.02,
Fig. 7A) and low ACh post-control inhibition (net charge 0.50 ± 0.04, Fig. 7B). Moreover, PAM
potentiation was observed only in co-application with PNU-120596 (net charge 24.62 ± 3.30,
Fig. 7D). Conversely, the nitro-containing derivative 29 behaved as a weak partial agonist with
weak ACh post-control inhibition (Fig. 7A and Fig. 7B, respectively). The residual observed
channel PNU-102596 potentiation was relatively low (net charge 4.78 ± 2.07, Fig. 7C), while the
potentiation with co-application reflected what is expected for a partial agonist (net charge 12.19
± 1.68, Fig. 7D). The furan derivative 30 acted as partial agonist with almost 50% of ACh
response (Fig. 7A). The most interesting feature of 30 was its relatively long binding to the
receptor, evidenced by considerable ACh post-control inhibition and large residual PAM-
18

potentiated response (Fig. 7B and Fig. 7C, respectively). Regarding the bulkier naphthalene (31)
and isoquinoline (32) derivatives, both displayed low agonist activity (net charges 0.15 ± 0.02
and 0.11 ± 0.02, respectively in Fig. 7A) and low ACh post-control inhibition (net charges 1.18 ±
0.05 and 0.94 ± 0.03, respectively in Fig. 7B). Besides, the residual PNU-120596 potentiation
(Fig. 7C) was completely absent. Surprisingly, in co-application with PNU-120596 (Fig. 7D), 31
did not show any direct potentiated response, while 32 gave a noticeable response (net charge
20.44 ± 11.77), which makes it a short-lasting silent agonist.
Fig. 7. Electrophysiological data of the NS6740 derivatives 21-32 at 10 μM. Protocol I: (A) Channel
opening given by 21-32; (B) 60 μM ACh post-control activation after the delivery of 21-32; (C) Residual
potentiated response given by application of type II PAM PNU-120596 after delivery of 21-32 and the 60
μM ACh post-control. Protocol II: (D) Direct potentiated response given by 21-32 plus 10 μM PNU-
120596.
Among the studied compounds, fragment 13 and methylammonium derivative 17
behaved as efficacious partial agonists of the α7 nAChR, showing an increased ability to
promote channel opening compared to 1 without inducing prolonged desensitized states. In
19

particular, compounds 17 and 13 displayed comparable efficacy with Imax values of 0.81 ±
0.01 and 0.71 ± 0.02, respectively, relative to 60 μM ACh. Compound 17 was twice as potent
as 13, with EC50 values of 322 ± 18 nM and 764 ± 52 nM, respectively. Concentration-
response curves for 13 and 17 are provided in Fig. 8.
Fig. 8. Concentration-response curves for the α7 nAChR responses evoked by derivatives 13 and 17 at
0.1, 0.3, 1, 3, 10, and 30 μM measured as net charge. Data points represent the mean ± SEM from at least
four oocytes and are normalized to 60 μM ACh control responses.
2.4. Structure-Activity Relationship (SAR) analyses of the synthesized compounds
Electrophysiological assays were utilized to evaluate the activity profiles of the synthesized
compounds (Fig. 2). The SAR analyses summarized in Fig. 9 focused on effects of the
modifications of the lead compound NS6740 related to the three pharmacophoric regions,
namely the positively charged diazabicyclic system, the central hydrogen-bond acceptor
heterocyclic moiety, and the terminal phenyl group. The key features we are trying to pin down
are low channel activation associated with prolonged receptor binding and stabilization of Ds
states. However, some derivatives are classifiable as partial agonists based on the greater
efficacy in channel opening when applied alone, but they may also display an NS6740-like
feature because of their ability to induce a sustained receptor desensitization. Initially, we
investigated the effect of the NS6740 skeleton fragmentation (Fig. 2A). The focused series of six
compounds 11-16 provided the first insight to the key moieties and functionalities that play
important roles for inducing α7 silent activation. Removal of the basic diazabicyclic system with
a residual diethylamide group in derivative 11 led to complete loss of activity. This result
indicated that the interactions established by the protonatable nitrogen atom within the receptor
binding site are essential for conferring to NS6740 its peculiar activity profile. When the
20

trifluoromethyl group in meta position on the phenyl ring of 1 was removed, the resulting
compound 12 appeared to be a partial agonist with desensitizing properties. This indicates the
relevant role of this substituent in silencing the agonist response, although it is not strictly
required to stabilize the desensitized states. Interestingly, further removal of the phenyl group
afforded compound 13, which showed a very efficacious partial agonist activity, which is
associated with a short-lasting receptor binding, thus proving the central role of the phenyl
portion for inducing the desensitized states. The additional deletion of the furan ring further
destabilized the receptor binding. Indeed, a short-lasting silent interaction was maintained as
long as the carbonyl group was retained as a pharmacophore feature in compound 14, while
compounds 15 and 16 were completely devoid of activity, demonstrating that the diazabicyclic
system as the positively charged center is an essential but not sufficient pharmacophoric feature
for silent activation. On the whole, these results suggest that all the structural portions of NS6740
are required to confer its peculiar profile. In particular, the basic nucleus along with the hydrogen
bond acceptor moiety and the aromatic moiety are associated with its ability to engender a strong
desensitization. Additionally, the presence of the trifluoromethyl group seems to be related to the
low channel opening that is typical of the very weak partial agonist 1.
We next moved to exploring the distinct elements of the key scaffold (Fig. 2B). Concerning
the 1,4-diazabicyclo[3.2.2]nonane system, the introduction of a permanent positive charge led to
a strong partial agonist (compound 17), although the desensitizing properties were compromised.
Interestingly, the regioisomeric 1,3-diazabicyclo[3.2.2]nonane derivative 18 was completely
inactive, demonstrating the critical role of the distance between the protonatable nitrogen atom
and the carbonyl group to maintain a correct orientation in the binding domain for stimulating
the receptor desensitization. Moreover, smaller dimensions of the basic nucleus and the
introduction of a secondary amine (compounds (S,S)-19 and (R,R)-19) gave a much shorter
receptor binding duration.
Then we found that the deletion of the pharmacophoric carbonyl group (compound 20) was
detrimental to the α7 Ds state. The loss of activity could be related to the absence of hydrogen
bond interaction between the carbonyl group and the receptor, but also to different
conformations that the more flexible derivative 20 may assume, as well for the presence of a
second competitive protonatable center which may alter the interaction within the receptor
binding site.
Finally, we examined the possible influences of substituents at the meta position of the phenyl
moiety. Various groups with different sizes and electronic properties were used to replace the
original trifluoromethyl group. In general, the introduction of different meta substituents
determined an increase of partial agonist activity in terms of channel opening, whereas the PAM
potentiation was differently affected depending on the substituent. The bioisosteric methyl group
21

(compound 21) as well as the halogen atoms (compounds 23-26) and the hydroxyl group
(compound 27) maintained relevant desensitizing properties, while the introduction of the
electron-withdrawing nitro group (compound 29) and the electron-donor amino substituent
(compound 28) shortened the permanence in the receptor binding pocket, suggesting different
interactions established in the binding site. When we moved the original trifluoromethyl group
from the meta to para position (compound 22), we observed greater channel opening, which is
associated with diminished stabilization of non-conducting receptor states. This could suggest a
different orientation of the phenyl moiety in the receptor binding site. Finally, the phenyl moiety
was replaced by different aromatic groups. Interestingly, the introduction of the furan ring in the
place of the phenyl moiety (compound 30) maintained a desensitizing profile, whereas the more
hindered naphthyl- and isoquinoline-containing analogs (compounds 31 and 32, respectively) did
not preserve prolonged receptor binding and resulted in very weak partial agonists devoid of the
ability to induce a sustained receptor desensitization. Therefore, we deduced that bulky
substituents in the terminal aryl region are detrimental to establishing a stable receptor
interaction. Fig. 9 summarizes the main structure–activity findings.
Fig. 9. Summary of the structure‒activity relationships for the synthesized NS6740 derivatives.
3. Conclusion
In the present work, we extensively investigated the SARs of NS6740 (1), a very weak
partial agonist able to promote strong α7 nAChR desensitization through induction of stable non-
conducting states, which appear to be strictly related to the metabotropic activation mode of this
receptor channel. The NS6740 skeleton was analyzed using a fragmentation approach and
22

systematic molecular modifications performed at the main pharmacophore moieties, namely, the
positively charged center, the hydrogen bond acceptor moiety, and the aromatic nucleus. Overall,
the three relevant structural portions of NS6740 were found to be essential to the peculiar
electrophysiological behavior of this nicotinic ligand. Indeed, modifications of the basic nucleus
led to loss of the NS6740 activity profile, and deletion of the central carbonyl group was
detrimental to both the stabilization of the receptor binding and the subsequent sustained
induction of receptor desensitization. The effects of the trifluoromethyl substituent on the phenyl
ring were found to be critical in the meta more than in the para position. Moreover, insertion of
halogen atoms and the methyl group in the meta position generally increased the partial agonist
behavior. However, these ligands maintained prolonged receptor binding and sustained
desensitization, whereas more polar groups produced a shorter residence time and a weaker
stabilization of non-conducting states. Aromatic moieties bulkier than the phenyl ring decreased
the binding stability, thus abolishing the prolonged stabilization of the receptor desensitization.
Among the new compounds, fragment 12, meta-methyl 21 and meta-halogen derivatives 23-
26, and the furan analogue 30 emerged as novel α7 receptor desensitizers, overall displaying a
partial agonist activity higher than that shown by 1, but still associated with prolonged receptor
binding and strong desensitization. Fragment 13 and methylammonium derivative 17 behaved
also as efficacious partial agonists.
In summary, among the designed compounds, we identified a group of new promising
desensitizing agents, which will be further studied to investigate the α7 metabotropic G-protein
signaling and to evaluate their ability to modulate anti-inflammatory and analgesic responses.
The results herein discussed reinforce the potential of a metabotropic activation of the α7 nAChR
as an example of biased agonism for a receptor channel, whereby downstream activity is
mediated by a mechanism other than the canonical channel opening. This may represent an
approach for the treatment of pathological conditions such as inflammatory diseases and
neuropathic pain.
23

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
with the Dragendorff reagent for tertiary amines and quaternary ammonium salts, or with an
ethanolic solution of ninhydrin for aminoacids. 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
uncorrected. Rotary power determinations (sodium D line, 589) were carried out with a Jasco P-
1
1010 Polarimeter coupled with a Huber thermostat. H NMR and 13C NMR spectra were
recorded with a Varian Mercury 300 (1H, 300.063; 13C, 75.451 MHz) spectrometer at 20 °C.
Abbreviations used for peak multiplicities are given as follows: s (singlet), br (broad signal), d
(doublet), t (triplet), q (quartet), quint (quintet), m (multiplet), dd (doublet of doublets), td (triplet
of doublets), qd (quartet of doublets), dt (doublet of triplets), ddt (doublet of doublet of triplets),
ddd (doublet of doublets of doublets), qdd (quartet of doublets of doublets). Coupling constants
(J) are given in Hz, chemical shifts (δ) are expressed in parts per million (ppm) and calibrated for
13
1H using TMS and for C using residual deuterated solvent as internal standard (CDCl3, 77.16
ppm; CD3OD, 49.00 ppm; DMSO-d6, 39.52 ppm; (CD3)2CO, 29.84 ± 0.01 ppm, 206.26 ± 0.13
ppm). The purity of compounds was determined to be over 95% (except for 20, 92%) by UPLC–
MS analysis performed on a Waters SYNAPTG2-Si system equipped with a Waters ACQUITY
UPLC HSS T3 column, 1.8 μm, 2.1 mm × 100 mm. The injection volume of the sample was 10
μL, and the flow rate was 0.4 mL/min. The mobile phase consisted of 0.1 % formic acid in water
(phase A), and acetonitrile (phase B). The time program elution was as follows: 0.0–1.0 min A/B
90:10, 1.0–3.0 min A/B 10:90, 3.0–7.0 min A/B 10:90, 7.0–7.1 min A/B 90:10, 7.2–10.0 min
A/B 90:10 at 40 °C. For UV detection was used an ACQUITY PDA detector at 280 and 300 nm.
For the high-resolution mass spectrometry (HRMS) measurements, electrospray ionization (ESI)
mass spectra were obtained with a capillary voltage of 2.0 kV at a resolution of 40000–50000
units in positive-ion mode. Data are reported as mass-to-charge ratio (m/z) for the lead
compounds 1 and the target compounds 11-32.
24

The synthesis of intermediate compounds (S,S)-43 and (R,R)-43 (Scheme 2) and 62a−m, 63,
64, 65 (Scheme 3) is reported in the Supplementary Data.
4.1.1. Preparation of compounds 16 and 37
4.1.1.1. Quinuclidine-3-one oxime (34)
To a stirred solution of hydroxylamine hydrochloride (16 g, 230.5 mmol, 1.15 equiv) and
sodium acetate (48.4 g, 589.4 mmol, 3 equiv) in water (150 mL) was added dropwise
quinuclidine-3-one hydrochloride 33 (32.4 g, 200.5 mmol, 1 equiv) dissolved in water (100 mL)
and the reaction was heated at 70 °C for 2 h. The mixture was subsequently cooled to rt, and
Na2CO3 was added portionwise, bringing the pH of the reaction mixture to 10. The resulting
white precipitate was filtered off under reduced pressure and the aqueous layer was extracted
with a mixture of CH2Cl2/i-PrOH 9:1 (3 × 200 mL). The organic layers were dried over
anhydrous Na2SO4, filtered and concentrated under reduced pressure. Both the filtered and
extracted solid residues were combined to afford 34 as a white amorphous solid (28 g, 100%
1
yield); mp = 220–222 °C. Rf = 0.43 (CH2Cl2/MeOH 4:1). H NMR (300 MHz, DMSO-d6): δ
10.22 (s, 1H), 3.38 (s, 2H), 2.84–2.61 (m, 4H), 2.42 (quint, J = 3.1 Hz, 1H), 1.78–1.56 (m, 4H).
13C NMR (75 MHz, DMSO-d6): δ 162.21, 51.80, 46.73 (2C), 28.32, 26.38 (2C).
4.1.1.2. Synthesis of 35 and 36
The oxime intermediate 34 (38 g, 271.1 mmol) was added portionwise over 6 h to
mechanically stirred polyphosphoric acid (95 mL) heated at 130 °C so as to obtain a light brown
gummy mixture. The reaction was stirred at 130 °C overnight and the mixture’s color changed to
dark brown. The acidic mixture was then cooled to rt and ice-water (2 L) and 10 M NaOH were
added to adjust the pH to 12−13. The aqueous phase was saturated with NaCl and extracted with
a mixture of CH2Cl2/i-PrOH 4:1 (6 × 700 mL). The combined organic layers were dried over
anhydrous Na2SO4, filtered, and concentrated under reduced pressure. The resulting crude was
purified by silica gel column chromatography (CH2Cl2/MeOH 4:1) to afford the two
regioisomers 35 as light brown amorphous solid (9.23 g, 24% yield) and 36 as light brown
amorphous solid (1.64 g, 4.3% yield).
4.1.1.2.1. 1,4-Diazabicyclo[3.2.2]nonan-3-one (35). Mp = 211−213 °C. Rf = 0.39
(CH2Cl2/MeOH 4:1). 1H NMR (300 MHz, CDCl3): δ 6.91 (s, br 1H), 3.69 (s, 2H), 3.49–3.37 (m,
1H), 3.09 (qdd, J = 14.5, 9.2, 5.6 Hz, 4H), 2.21–2.06 (m, 2H), 1.95-1.81 (m, 2H). 13C NMR (75
MHz, CDCl3): δ 178.67, 62.30, 46.25 (2 C), 44.14, 30.20 (2 C).
25

4.1.1.2.1. 1,3-Diazabicyclo[3.2.2]nonan-4-one (36). Mp = 122−123 °C. Rf = 0.41
(CH2Cl2/MeOH 95:5). 1H NMR (300 MHz, CDCl3): δ 7.12 (br s, 1H), 5.82–5.72 (m, 1H), 5.70–
13
5.61 (m, 1H), 3.12–3.02 (m, 5H), 2.70–2.61 (m, 2H), 2.24–2.13 (m, 2H). C NMR (75 MHz,
CDCl3): δ 173.95, 61.56, 52.91 (2C), 50.59, 26.28 (2C).
4.1.1.3. 1,4-Diazabicyclo[3.2.2]nonane dihydrochloride (16)
To a suspension of 35 (5.59 g, 39.9 mmol, 1 equiv) in anhydrous THF (70 mL) at 0 °C under
an anhydrous and inert atmosphere was added dropwise a 1.0 M solution of LiAlH4 in THF (47.9
mL, 47.9 mmol, 1.2 equiv). The mixture suspension was heated to refluxing temperature and a
light-yellow solution was obtained. After refluxing for 8 h, the reaction was complete, and a
white solid was crashed out. The excess of LiAlH4 was quenched with EtOAc (50 mL) and few
drops of MeOH at 0 °C, producing an additional white precipitate, which was removed by
filtration. The yellow solution filtrate was subsequently treated with a 4.0 M solution of HCl in
1,4-dioxane (22.2 mL, 88.6 mmol, 2 equiv) and stirred at rt overnight. The corresponding
hydrochloride salt 16 was obtained as a brown precipitate, which was isolated by vacuum
filtration and recrystallized from MeOH/Et2O (4:1) to afford 16 as light brown prisms (4.2 g,
1
47.6% yield); mp = 307 °C (decomposed). H NMR (300 MHz, CD3OD): δ 4.14 (quint, J = 3.8
13
Hz, 1H), 3.81–3.73 (m, 2H), 3.68–3.42 (m, 6H), 2.40 (td, J = 8.0, 3.8 Hz, 4H). C NMR (75
MHz, CD3OD): δ 51.76, 49.22, 46.69 (2C), 38.77, 21.34 (2C). HRMS (ESI) m/z calcd for
C7H15N2 127.1235 [M + H]+, found 127.1233.
4.1.1.4. 1,3-Diazabicyclo[3.2.2]nonane dihydrochloride (37)
This compound was prepared starting from the minor regioisomer 36 (700 mg, 4.99 mmol, 1
equiv) in anhydrous THF (9 mL) according to the method described for 16 to afford 37 as light
brown prisms (600 mg, 45% yield); mp = 117−119 °C. 1H NMR (300 MHz, CD3OD): δ 6.02 (d,
J = 10.4 Hz, 1H), 5.77 (d, J = 10.4 Hz, 1H), 3.95–3.82 (m, 2H), 3.65–3.41 (m, 7H), 2.66–
2.46(m, 2H). 13C NMR (75 MHz, CD3OD): δ 53.70, 51.92, 50.86 (2C), 35.21, 23.43 (2C).
4.1.2. Preparation of compounds 11−15
4.1.2.1. N,N-Diethyl-5-(3-(trifluoromethyl)phenyl)furan-2-carboxamide (11)
To a solution of carbonyl chloride 62a (515 mg 1.88 mmol, 1 equiv) in anhydrous CH2Cl2 (10
mL) was added dropwise at rt triethylamine (0.5 mL, 3.60 mmol, 2 equiv), followed by
diethylamine 66 (0.19 mL, 1.88 mmol, 1 equiv). After stirring for 1.5 h at rt, the mixture was
diluted with water (6 mL) and extracted with EtOAc (3 × 15 mL). The combined organic layers
26

were dried over anhydrous Na2SO4 and evaporated under reduced pressure. The crude residue
was purified by silica gel column chromatography (cyclohexane/EtOAc 7:3) to afford the target
compound 11 as a yellow solid (290 mg, 50% yield); mp = 45−48 °C. Rf = 0.45
(cyclohexane/EtOAc 3:2). 1H NMR (300 MHz, (CD3)2CO): δ 8.11–8.05 (m, 2H), 7.76–7.66 (m,
2H), 7.20 (d, J = 3.6 Hz, 1H), 7.11 (d, J = 3.6 Hz, 1H), 3.77–3.45 (m, 4H), 1.48–1.15 (m, 6H).
13C NMR (75 MHz, (CD3)2CO): δ 159.29, 153.61, 149.91, 131.94, 131.75 (q, J = 32.2 Hz),
130.94, 128.48, 125.45 (q, J = 3.5 Hz), 125.10 (q, J = 271.8 Hz), 121.39 (q, J = 3.3 Hz), 118.29,
109.16, 42.61 (2C), 14.31 (2C). UPLC tR = 2.57 min; HRMS (ESI) m/z calcd for C16H17NO2F3
312.1211 [M + H]+, found 312.1210.
4.1.2.2. General procedure for the synthesis of 12−14
To a stirred suspension of 16 (1 equiv) in anhydrous DMF (0.2 M) preheated at 60 °C for 40
min under inert atmosphere, Cs2CO3 (5 equiv) was added portionwise at rt to reach pH 8−9. The
mixture was subsequently cooled down to 0 °C and the appropriate carbonyl chloride (1.2 equiv)
was added portionwise. The reaction was then warmed to rt and stirred overnight. Upon
completion, the mixture was diluted with water and extracted repeatedly with CHCl3/i-PrOH
(4:1). The combined organic layers were washed with brine (3 ×), dried over anhydrous Na2SO4,
filtered and concentrated under vacuum to obtain compounds 12-14 as tertiary bases, which were
converted to fumarate salts. The tertiary base (1 equiv) was dissolved in MeOH (0.2 M) and
fumaric acid (1 equiv) was added. After stirring overnight at rt, the solvent was removed under
reduced pressure and the crude salt was recrystallized from an appropriate solvent to give the
desired final fumarate salts 12-14 × C4H4O4.
4.1.2.2.1. 1,4-Diazabicyclo[3.2.2]nonan-4-yl(5-phenylfuran-2-yl)methanone fumarate (12).
The title compound was prepared according to the general procedure from 16 (180 mg, 0.90
mmol, 1 equiv), Cs2CO3 (1.47 g, 4.52 mmol, 5 equiv), and carbonyl chloride 62b (224 mg, 1.08
mmol, 1.2 equiv). After standard workup and subsequent reaction with fumaric acid (66 mg, 0.57
mmol, 1 equiv), the crude salt was recrystallized from MeOH to provide the pure fumarate salt
1
12 × C4H4O4 as a white solid (180mg, 53% yield); mp = 174−176 °C. H NMR (300 MHz,
CD3OD): δ 7.82–7.75 (m, 2H), 7.49–7.33 (m, 3H), 7.22 (d, J = 3.6 Hz, 1H), 6.97 (d, J = 3.7 Hz,
13
1H), 6.71 (s, 2H), 4.91 (s, br 1H), 4.28 (s, br 2H), 3.59–3.44 (m, 6H), 2.48–2.13 (m, 4H). C
NMR (75 MHz, CD3OD): δ 170.45 (2C), 160.90, 157.61, 147.35, 135.89 (2C), 130.90, 130.08
(2C), 129.96, 125.56 (2C), 120.84, 107.90, 55.84, 47.14, 24.86. UPLC tR = 2.40 min; HRMS
(ESI) m/z calcd for C18H21N2O2 297.1603 [M + H]+, found 297.1609.
27

4.1.2.2.2. 1,4-Diazabicyclo[3.2.2]nonan-4-yl(furan-2-yl)methanone fumarate (13). The title
compound was prepared according to the general procedure from 16 (180 mg, 0.90 mmol, 1
equiv), Cs2CO3 (1.47 g, 4.50 mmol, 5 equiv), and carbonyl chloride 67 (148 mg, 1.1 mmol, 0.11
mL, 1.2 equiv). After standard workup and subsequent reaction with fumaric acid (65 mg, 0.56
mmol, 1 equiv), the crude salt was recrystallized from MeOH/Et2O (4:1) to provide the pure
1
fumarate salt 13 × C4H4O4 as a white solid (100 mg, 53% yield); mp = 180−182 °C. H NMR
(300 MHz, CD3OD): δ 7.72 (dd, J = 1.7, 0.7 Hz, 1H), 7.11 (d, J = 3.4 Hz, 1H), 6.69 (s, 2H), 6.62
(dd, J = 3.5, 1.8 Hz, 1H), 4.87–4.80 (m, 1H), 4.23 (s, br 2H), 3.52 (dt, J = 15.9, 6.8 Hz, 6H),
2.42–2.10 (m, 4H). 13C NMR (75 MHz, CD3OD): δ 171.27 (2C), 161.01, 148.34, 146.35, 136.10
(2C), 118.38, 112.63, 56.36, 46.96, 24.28. UPLC tR = 0.87 min; HRMS (ESI) m/z calcd for
C12H17N2O2 221.1290 [M + H]+, found 221.1288.
4.1.2.2.3. 1-(1,4-Diazabicyclo[3.2.2]nonan-4-yl)ethanone fumarate (14). The title compound
was prepared according to the general procedure from 16 (150 mg, 0.75 mmol, 1 equiv), Cs2CO3
(1.20 g, 3.80 mmol, 5 equiv), and carbonyl chloride 68 (200 mg, 1.50 mmol, 0.06 mL, 1.2
equiv). After standard workup and subsequent reaction with fumaric acid (54 mg, 0.46 mmol, 1
equiv), the crude salt was recrystallized from EtOH/Et2O (4:1) to provide the pure fumarate salt
14 × 0.75 C4H4O4 as a light brown solid (150 mg, 64% yield); mp = 169−170 °C. 1H NMR (300
MHz, CD3OD): δ 6.68 (s, 1.5H), 4.80–4.71 (m, 0.5H), 4.46–4.37 (m, 0.5H), 4.06–3.97 (m, 1H),
3.96–3.87 (m, 1H), 3.52–3.37 (m, 6H), 2.20–2.10 (m, 4H), 2.15 (s, 3H). 13C NMR (75 MHz,
CD3OD): δ 171.86, 171.62 (2C), 136.21 (2C), 56.83, 55.90, 47.08, 46.82, 46.26, 30.67, 23.65,
22.03. HRMS (ESI) m/z calcd for C9H17N2O 169.1341 [M + H]+, found 169.1341.
4.1.2.3. 4-Ethyl-1,4-diazabicyclo[3.2.2]nonane fumarate (15)
To a suspension of 14 (300 mg, 1.78 mmol, 1 equiv) in anhydrous Et2O (14 mL) was added
dropwise a suspension of LiAlH4 (152 mg, 4 mmol, 2.25 equiv) in anhydrous Et2O (4 mL) at 0
°C under inert and anhydrous atmosphere. After refluxing for 3 h, the reaction was quenched
with sodium sulfate decahydrate at 0 °C, the mixture was filtered, and the filtrate was
concentrated under vacuum. The resulting residue compound 15 as free base (1 equiv) was
reacted with fumaric acid (206 mg, 1.78 mmol, 1 equiv) at rt overnight. After evaporation of the
solvent under reduced pressure, the crude salt was recrystallized from MeOH/Et2O (4:1) to
provide the pure fumarate salt 15 × 2 C4H4O4 as a white solid (165 mg, 25% yield); mp =
1
134−135 °C. H NMR (300 MHz, CD3OD): δ 6.71 (s, 4H), 3.74–3.65 (m, 1H), 3.52–3.44 (m,
2H), 3.37–3.22 (m, 6H), 3.10 (q, J = 7.2 Hz, 2H), 2.44–2.28 (m, 2H), 2.19–2.01 (m, 2H), 1.29 (t,
J = 7.2 Hz, 3H). 13C NMR (75 MHz, CD3OD): δ 170.22 (2C), 135.82 (2C), 55.75, 52.30, 52.22,
28