The optimization of a novel selective antagonist for human M2 muscarinic acetylcholine receptor

Article abstract of DOI:10.1016/j.bmcl.2020.127632

Muscarinic acetylcholine receptors (mAChRs) comprise five distinct subtypes denoted M₁ to M₅. The antagonism of M₂ subtype could increase the release of acetylcholine from vesicles into the synaptic cleft and improve postsynaptic functions in the hippocampus via M₁ receptor activation, displaying therapeutic potentials for Alzheimer's disease. However, drug development for M₂ antagonists is still challenged among different receptor subtypes. In this study, by optimizing a scaffold from virtual screening, we synthesized two focused libraries and generated up to 50 derivatives. By measuring potency and binding selectivity, we discovered a novel M₂ antagonist, ligand 47, featuring submicromolar IC₅₀, high M₂/M₄ selectivity (~30-fold) and suitable lipophilicity (cLogP = 4.55). Further study with these compounds also illustrates the structure–activity relationship of this novel scaffold. Our study could not only provide novel lead structure, which was easy to synthesize, but also offer valuable information for further development of selective M₂ ligands.

Full text of DOI:10.1016/j.bmcl.2020.127632

Bioorganic&MedicinalChemistryLetters30(2020)127632
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
The optimization of a novel selective antagonist for human M2 muscarinic
acetylcholine receptor
Miaomiao Lia,h, Chen Huangb,c,h, Xingyu Wud, Fan Dingb, Zhoumi Hub, Yan Zhue, Lanxue Zhaoa,
⁎
⁎
Lina Houa, Hongzhuan Chena,f, Hao Wanga, Jianrong Xua,g, , Dewei Tangb,
a Department of Pharmacology and Chemical Biology, Shanghai Jiao Tong University School of Medicine, Shanghai, 200025, China
b Department of Nuclear Medicine, Renji Hospital, School of Medicine, Shanghai Jiao Tong University, 160 Pu Jian Road, Shanghai 200127, China
c College of Medical Imaging, Shanghai University of Medicine and Health Sciences, Shanghai 201318, China
d College of Pharmacy, Shanghai University of Medicine and Health Sciences, Shanghai 201318, China
e School of Medical Instrument and Food Engineering, University of Shanghai for Science and Technology, Shanghai 200093, China
f Institute of Interdisciplinary Integrative Biomedical Research, Shanghai University of Traditional Chinese Medicine, Shanghai 201203, China
g Academy of Integrative Medicine, Shanghai University of Traditional Chinese Medicine, Shanghai 201203, China
A R T I C L E I N F O
A B S T R A C T
Keywords:
Muscarinic acetylcholine receptors (mAChRs) comprise five distinct subtypes denoted M1 to M5. The antagonism
of M2 subtype could increase the release of acetylcholine from vesicles into the synaptic cleft and improve
postsynaptic functions in the hippocampus via M1 receptor activation, displaying therapeutic potentials for
Alzheimer’s disease. However, drug development for M2 antagonists is still challenged among different receptor
subtypes. In this study, by optimizing a scaffold from virtual screening, we synthesized two focused libraries and
generated up to 50 derivatives. By measuring potency and binding selectivity, we discovered a novel M2 an-
tagonist, ligand 47, featuring submicromolar IC50, high M2/M4 selectivity (~30-fold) and suitable lipophilicity
(cLogP = 4.55). Further study with these compounds also illustrates the structure–activity relationship of this
novel scaffold. Our study could not only provide novel lead structure, which was easy to synthesize, but also
offer valuable information for further development of selective M2 ligands.
Muscarinic acetylcholine receptor
Antagonist
Structure–activity relationship
Selective
Alzheimer’s disease
G protein-coupled receptors (GPCRs) constitute the largest family of
proteins targeted by Food and Drug Administration approved drugs.1
The muscarinic acetylcholine receptors (mAChRs) belong to α-branch
of class A GPCRs and they are ubiquitously distributed in human or-
gans, regulating a variety of physiological functions including heart-
beats, smooth muscle contraction, glandular secretion and many fun-
damental functions for central nervous system (CNS).2 mAChRs
comprise five distinct subtypes denoted M1 to M5. Three of these sub-
types (M1, M3 and M5) are coupled to G proteins of the Gq/11 family,
while the other two subtypes (M2 and M4) are coupled to G proteins of
the Gi/o family.3 Currently, drugs targeting mAChRs are generally de-
veloped for the treatment of various diseases, including chronic
obstructive pulmonary diseases, Alzheimer’s disease, Parkinson’s dis-
ease, overactive bladder syndromes and other diseases like cancer,
diabetes, cardiovascular diseases, pain and inflammation.3–5
M1 to M5 subtypes feature different tissue distribution patterns in
human. For example, M1, M2 and M4 subtypes are the major mAChRs
subtypes expressed in the brain, where M3 and M5 subtypes are ex-
pressed at significantly lower levels.6 In neurons, the M2 subtype is the
largely presynaptic receptor associated with axons and inhibits neu-
ronal excitability, which results in the negative feedback of neuro-
transmitter release.7 At synapses, the antagonism of the M2 subtype
could increase the release of acetylcholine (ACh) from vesicles into the
synaptic cleft and improve postsynaptic functions in the hippocampus
Abbreviations: mAChRs, muscarinic acetylcholine receptors; GPCRs, G protein-coupled receptors; CNS, central nervous system; GIRK, G protein-coupled inwardly
rectifying K+ channels; ACh, acetylcholine; PAM, positive allosteric modulator; BQCA, benzyl quinolone carboxylic acid; NAM, negative allosteric modulator; BBB,
blood-brain barrier; SARs, structure-activity relationships; LOO, leave-one-out; QSAR, quantitative structure-activity relationships; QNB, quinuclidinyl benzylate; SP,
standard precision
⁎ Corresponding authors at: Department of Pharmacology and Chemical Biology, Shanghai Jiao Tong University School of Medicine, Shanghai, 200025, China and
Academy of Integrative Medicine, Shanghai University of Traditional Chinese Medicine, Shanghai 201203, China (J. Xu) and Department of Nuclear Medicine, Renji
Hospital, School of Medicine, Shanghai Jiao Tong University, 160 Pu Jian Road, Shanghai 200127, China (D. Tang).
E-mail addresses: janker.xu@gmail.com (J. Xu), acetdw@126.com (D. Tang).
h These authors contributed equally to this work.
https://doi.org/10.1016/j.bmcl.2020.127632
Received 31 July 2020; Received in revised form 14 October 2020; Accepted 16 October 2020
Availableonline24October2020
0960-894X/©2020ElsevierLtd.Allrightsreserved.

M. Li, et al.
Bioorganic&MedicinalChemistryLetters30(2020)127632
Fig. 1. The structures of M2 subtype antagonists.
via M1 subtype activation, displaying therapeutical potentials for Alz-
heimer’s disease.4,8–11
antagonists, methoctramine23–25 and tripitramine26–28, with cLogP va-
lues larger than 7, are also limited in their ability to cross the blood–-
brain barrier (BBB) (Fig. 1). Therefore, there is still an urgent need to
develop novel antagonists for M2 subtype with high selectivity, binding
affinity, drug potential and suitable lipophilicity (3 < cLogP < 5).
At present, the structures of mAChRs have been determined using X-
ray crystallography and Cryo-EM, providing new opportunities for the
discovery of subtype-selective ligands.29–34 Novel selective M2 antago-
nists can be developed based on these structures and the structure–-
activity relationships (SARs). In this study, by optimizing from a non-
selective antagonist with a novel scaffold obtained from virtual screen,
we performed a two-round optimization and synthesized up to 50 dif-
ferent derivatives. By measuring the binding specificity and drug po-
tential of each compound, we discovered a novel M2 antagonist fea-
turing a high M2/M4 selectivity, suitable lipophilicity as well as the
drug potentials. Further SARs have also been elucidated with the fo-
cused library developed in this research.
Human M1 to M5 subtypes share 64 to 82% sequence identity and 82
to 92% sequence similarity in the transmembrane region, but they vary
significantly at the N-terminal and extracellular regions. Although M2
receptor is an attractive target for Alzheimer’s disease, it is still chal-
lenging to discover novel subtype-specific ligands. By using allosteric
strategy, several subtype-selective mAChRs ligands have been success-
fully developed, such as the M1 subtype positive allosteric modulator
(PAM) benzyl quinolone carboxylic acid (BQCA) , and the M5 receptor
negative allosteric modulator (NAM) ML375.12,13 Despite significant
progress through these strategies, the orthosteric site still showed va-
luable potentials for subtype-selective ligands, such as the latest
achievements on both M2 and M3 subtypes.14–17
Currently, several antagonists for M2 mAChRs like SCH 72,788,18
BIBN19,20 and AFDX38415,21,22 have been developed (Fig. 1), but are
still limited on their M2/M4 selectivity. Other two M2 specific
2

M. Li, et al.
Bioorganic&MedicinalChemistryLetters30(2020)127632
Fig. 2. Structure, activities and cLogP of lead compound 1, 2 and 47.
was characterized to be an antagonist at high concentration. Because
AJ-292 showed a novel scaffold for mAChRs antagonists and had var-
ious sites for further optimization, we took it as the lead and carried out
the following optimization.
IC50 is a quantitative measure that indicates how much of a parti-
cular inhibitory substance is needed to inhibit a given biological process
by 50% and here refers to the potency to antagonize functional re-
sponse to the agonist ACh on mAChRs. In this study, by optimizing the
lead compound 1, we synthesized up to 50 compounds and measured
the binding affinity and specificity for each compound. The lead com-
pound features a scaffold comprising a carbazole and a piperidine ring.
A simple modification on the piperidine ring produced the compound 2,
which did not show a dramatic IC50 increase when compared to parent
compound 1 (Fig. 2). In order to discover the novel ligand with higher
IC50 and a suitable lipophilicity, an initial optimization was performed
to investigate the role of carbazole for the drug potential, by using
different two-ring amines as shown in Scheme 1 and Table 1. With this
optimization, two desired compounds 11 and 17 was discovered. Both
of these two ligands have a submicromolar IC50 and increased binding
affinities for M1 to M4 subtypes when compared to the parent com-
pound 1.
Scheme 1. Synthetic approach for the ligand optimization to setup the focused
library 1. i) different amines, DMF (3% K2CO3), 130 °C, 30 min; ii) different
amines, DMF(3% K2CO3), 130 °C, 30 min.
Interestingly, the modification from the carbazole to tetra-
hydroquinoline dramatically improve the IC50 by more than 100 fold,
as indicated by the IC50 of 1, 11 and 17. On the other side, the opti-
mization using indoline did not yield the higher potential ligands as
tetrahydroquinoline, as indicated by the comparison between 7, 8, 15,
18 and 9, 10, 11, 14, 16, 17. Similarly, substitution of tetra-
hydroquinoline with tetrahydronaphthyridine will also lower the scaf-
fold binding affinity to all the four subtypes of mAChRs, as indicated by
the comparison between 11, 17 and 13, 19 in Table 1. Interestingly, the
simple substitution of piperidine with pyrrolidine in the scaffold did not
affect the IC50 significantly as shown in Table 1 by the direct
By using Glide module in Schrodinger, this study screened 1.5
million compounds in commercial databases, including SPECS and
Chemdiv. The first-round screening was ranked by G score and the top
10,000 compounds were reranked under extra precision mode. Based
on CHO-K1/M2/Gα15 cell line, Ca2+ mobilization assay was applied
for the top 184 compounds in the ranking, of which 66 compounds
were purchased from SPECS and 118 compounds were from Chemdiv.
Five of them were identified as agonists of M2 receptors, such as AG-
205–6. However, only AJ-292 (lead compound 1) in SPECS database
3

M. Li, et al.
Bioorganic&MedicinalChemistryLetters30(2020)127632
Table 1
Structure and activity of the focused library 1.
Entry
R1
R2
pIC50 M2
pKi M1
< 4
pKi M2
4.02
pKi M3
< 4
pKi M4
< 4
cLogP
5.45
1
4.28
0.08
0.08
0.05
2
4.73
0.01
4.89
0.05
4.64
4.66
0.12
4.59
0.06
5.86
7
< 4
< 4
5.03
5.13
6.36
< 4
NA
< 4
< 4
4.39
5.44
6.38
< 4
< 4
< 4
4.28
4.96
< 4
< 4
4.06
5.05
5.91
3.86
3.35
2.57
3.76
3.33
8
9
0.05
0.06
0.06
< 4
4.55
5.54
0.06
0.07
0.08
0.08
0.07
0.07
10
11
0.07
0.06
0.08
0.07
12
13
5.01
< 4
0.05
4.36
< 4
0.09
4.31
4.39
0.28
0.06
4.35
4.22
0.10
0.07
4.04
4.23
0.08
0.07
4.49
2.29
14
15
16
17
4.61
< 4
5.31
6.21
0.06
4.75
4.00
4.35
5.54
0.08
0.08
0.07
0.04
4.73
4.67
5.61
6.41
0.07
0.05
0.05
0.05
4.18
4.10
< 4
5.13
0.09
0.08
4.67
4.13
5.20
5.86
0.07
0.08
0.05
0.06
3.74
2.79
3.20
2.68
0.07
0.06
0.07
18
19
4.12
< 4
0.06
< 4
4.30
4.36
0.07
0.09
4.02
4.07
0.09
0.08
< 4
< 4
2.59
1.73
4.03
0.10
Value denoted as Mean
SEM.
4

M. Li, et al.
Bioorganic&MedicinalChemistryLetters30(2020)127632
Scheme 2. Synthetic approach for the ligand opti-
mization to setup the focused library 2. i) 1) DMF,
1% K2CO3, MW, 140 °C, 30 min; 2) 6 M NaOH, MW,
120 °C, 30 min; ii) HATU, DIPEA, DMF, MW, 120 °C,
25 min.
Table 2
Structure and activity of the focused library 2.
Entry
R1
pIC50 M2
pKi M1
5.20
pKi M2
5.49
pKi M3
5.07
pKi M4
5.28
cLogP
3.86
22
5.50
0.06
0.06
0.07
0.06
0.09
0.07
0.07
0.07
0.05
23
5.62
0.07
5.94
5.95
5.45
5.95
4.38
24
25
4.66
4.53
0.07
0.07
4.56
4.48
0.05
0.04
5.32
5.21
0.08
0.06
4.26
4.27
0.06
0.06
4.75
4.73
0.07
0.04
4.38
1.87
26
27
4.82
4.97
0.08
0.05
4.63
4.94
0.03
0.04
4.88
5.44
0.07
0.05
4.48
4.50
0.06
0.05
4.77
5.14
0.04
0.04
1.52
2.29
28
6.12
0.07
5.84
0.05
6.21
0.04
5.28
0.05
5.76
0.04
2.29
29
30
< 4
< 4
4.19
6.24
0.09
0.10
< 4
< 4
3.29
3.00
6.11
0.07
0.10
5.98
0.06
0.05
5.34
0.05
6.05
0.05
0.03
31
4.15
4.29
4.78
0.06
< 4
4.30
1.90
(continued on next page)
5

M. Li, et al.
Bioorganic&MedicinalChemistryLetters30(2020)127632
Table 2 (continued)
Entry
R1
pIC50 M2
pKi M1
4.85
pKi M2
5.29
pKi M3
4.49
pKi M4
4.93
cLogP
2.52
32
4.72
0.07
0.04
0.03
0.06
0.07
0.08
0.09
0.05
0.03
33
4.76
0.05
4.85
5.29
4.36
4.89
2.52
34
35
4.92
4.69
0.05
0.04
5.35
4.76
0.05
0.03
5.77
5.45
0.05
0.06
4.94
4.49
0.08
0.07
5.24
4.80
0.06
0.07
2.23
3.34
36
37
5.82
6.25
0.05
0.05
6.22
6.52
0.07
0.07
6.39
7.05
0.07
0.07
5.54
6.01
0.07
0.07
6.11
6.54
0.06
0.04
3.86
3.86
38
39
5.74
6.27
0.07
0.07
5.98
6.58
0.05
0.06
6.54
6.94
0.07
0.04
5.54
6.11
0.06
0.09
5.80
6.67
0.06
0.06
2.78
4.04
40
6.53
0.06
6.64
0.05
7.26
0.04
5.76
0.05
0.06
6.80
0.04
4.11
41
42
6.50
6.40
0.06
0.05
6.61
5.95
0.05
0.06
7.01
5.46
0.05
0.08
5.59
< 4
6.59
4.23
0.04
0.77
3.55
3.28
43
44
5.85
6.47
0.06
0.07
5.25
5.03
0.04
0.03
5.48
6.04
0.05
0.05
4.97
4.58
0.07
0.14
4.97
5.39
0.06
0.10
5.01
5.22
45
5.18
0.05
4.85
0.05
5.14
0.06
4.53
0.08
4.85
0.06
2.56
(continued on next page)
6

M. Li, et al.
Bioorganic&MedicinalChemistryLetters30(2020)127632
Table 2 (continued)
Entry
R1
pIC50 M2
pKi M1
5.30
pKi M2
5.64
pKi M3
< 4
pKi M4
4.92
cLogP
4.71
46
6.37
0.06
0.06
0.05
0.09
0.08
0.09
0.10
47
6.32
0.07
5.31
6.21
< 4
4.74
4.55
48
49
50
5.95
5.99
6.01
0.07
0.06
0.07
5.58
6.19
5.12
0.04
0.05
0.05
5.91
6.19
5.48
0.05
0.04
0.05
4.92
5.14
4.59
0.07
0.05
0.09
5.36
5.67
5.06
0.05
0.05
0.06
3.18
3.65
2.56
51
5.86
0.07
5.69
0.04
5.72
0.07
4.89
0.11
5.19
0.04
4.15
Value denoted as Mean
SEM.
comparison between 8, 10, 11, 13 and 15, 16, 17, 19. For this initial
optimization, we discovered two novel ligands (11 and 17) featuring a
100-fold higher IC50 compared to the lead compound 1. The result of
the IC50 in Table 1 also indicates the substitution of carbazole with
tetrahydroquinoline will dramatically improve the IC50 of the scaffold.
In order to discover the novel ligands with higher M2 drug potential,
binding affinity and specificity, further optimization with 11 was also
performed using the synthetic approach shown in Scheme 2. In this
optimization, we took the linker length of ACh into accounts and
elongated the linker between the tetrahydroquinoline and the amine
from 2-carbon to 3-carbon. Furthermore, we also employed a variety of
amines for the construction of a novel focused library as shown in
Table 2. By analyzing the M2 IC50 of the compounds in this focused
library, we discovered several novel ligands featuring the sub-
micromolar binding IC50, like 28, 30, 37, 39, 40, 41, 42, 44, 46, 47
and 50 as shown in Table 2. Direct comparison between data in Table 1
happen for antagonists. By using ACh titration in absence/presence of
100 μM 8, 15 and 47, it was identified that these ligands were an-
tagonists, rather than NAMs (data not shown). In order to elucidate the
binding specificity for each ligand, the binding affinities to each sub-
type (M1 to M4) in Table 1 and Table 2 were also analyzed. The M5
mAChR-expressing cell line was provided by our vender GenScript and
it exhibited quite low expression level, which made us difficult to
characterize.
The inhibition constant, Ki, is reflective of the binding affinity and
could be determined with radioligand competition binding assays. In
the assay, both ACh and atropine were taken as positive controls for
competitive binding and showed pKi as follows: M1: 4.3 (ACh) and 9.5
(atropine), M2: 5.6 (ACh) and 8.5 (atropine), M3: 4.5 (ACh) and 9.0
(atropine), M4: 4.9 (ACh) and 9.0 (atropine). As shown in Fig. 3, by
directly comparing the pKi values for M2 with the other three mAChR
subtypes, we can directly determine the binding specificity for each
compound. In detail, the binding specificity for M2 over M1, M3 and M4
was determined individually for each compound using the pKi values
calculated. By comparing M2 and M1 binding affinity, we noticed li-
gands 9, 10, 16, 17, 44 and 47 demonstrate a significant higher (~8-
fold) binding affinity for M2 over M1, as shown in Fig. 3A. While for M2
over M3 selectivity, ligands 16, 42 and 47 demonstrate more than 50-
fold higher binding specificity, as shown in Fig. 3B. For M2 over M4
selectivity, only ligands 42 and 47 demonstrates more than 16-fold
higher binding selectivity as shown in Fig. 3C. The binding curve in
Fig. 4A and Fig. 4C clearly demonstrate ligand 47 binds specificity to
and Table
2 indicates the 3-carbon linker between the tetra-
hydroquinoline group and the amine groups has an equivalent drug
potential and binding affinities for M1 to M4 when compared with the
two-carbon linker, as indicated by ligands of 17 and 38. In sum, the
second-round optimization of the scaffold in Table 2 offers a variety of
novel M2 subtype ligands with an overall increased mAChRs binding
affinity and M2 IC50 when compared to the ligands in Table 1.
Since the ligands showed antagonism toward M2 mAChR, they
could be antagonists or NAMs. For NAMs, they could shift ACh titration
curve to the right in radioligand binding assay; while it would not
7

M. Li, et al.
Bioorganic&MedicinalChemistryLetters30(2020)127632
Fig. 3. The binding affinity comparison among different mAChRs. (A) The binding affinity comparison between M2 and M1 subtypes for all the ligands; (B) The
binding affinity comparison between M2 and M3 subtypes for all the ligands; (C) The binding affinity comparison between M2 and M4 subtypes for all the ligands.
8

M. Li, et al.
Bioorganic&MedicinalChemistryLetters30(2020)127632
Fig. 4. Comparison between lead compound 1 and novel M2 antagonist 47 on their mAChRs subtype selectivity. (A) mAChRs selectivity for compound 1; (B) mAChRs
selectivity for compound 47; (C) Binding selectivity comparison between 1 and 47.
Fig. 5. The evaluation of QSAR performance (A) Scaffold alignment for all the ligands; (B) Experimental vs predicted biological activities.
M2 when compared to the other subtypes of mAChRs. To the contrary,
the lead compound 1 did not show a specific binding profile as com-
pound 47, as shown in Fig. 4A, 4B, and 4C. In summary, the novel
ligand 47 features an 8-fold higher binding affinity than M1, more than
250-fold higher binding affinity than M3 and ~30-fold higher than M4,
demonstrating a significant selectivity for M2 subtype as shown in Fig. 3
and Fig. 4.
(LOO) method of cross-validation was adopted for assessment of the
predictive abilities of the models. The R2 for the regression was 0.7913
and the stability was 0.7210. The p value (5.34 e −-9) also suggested a
greater degree of confidence. Since external test was the gold standard
for QSAR model validation, the reliability of the model was tested by
external test set of 20 compounds. According to the recommended rules
for robust model,35 the RMSE (0.58), Q2 (0.6121) and Pearson’s r
(0.8048) of the model all further confirmed its robustness. Correlation
between the experimental and predicted activities for the data set is
displayed in Fig. 5B.
The cLogP measurement using ChemDraw (Perkin Elmer, USA) was
also performed to evaluate the lipophilicity for each compound. The
result indicates the lead compound 1 has a higher lipophilicity
(cLogP = 5.45) when compared with most of the compounds in Table 1
and Table 2, with a cLogP in a range of 1.5 to 5.5. The desired com-
pound 47 also has a lipophilicity with cLogP equals 4.55, which is in a
suitable range of 3 to 5.
The five conventional Gaussian fields, which contained Gaussian
steric, electrostatic, hydrophobic, H-bond acceptor and H-bond donor
field, illustrated that the substituents on far-end aromatic ring and near-
end aromatic ring were crucial for the activity. To depict the field-based
QSAR model in detail, the compound 47 was placed in the field contour
maps as reference ligand.
The quantitative structure–activity relationships (QSAR) model was
generated based on best conformations and common substructure
alignment. The retrieved conformations of all the molecules from
LigPrep module were superimposed using the common scaffold align-
ment approach of Schrödinger Suite (Fig. 5A). The training set of 30
molecules was randomly selected to build the model. The best model
(PLS factor = 3) offered a good predictive power. The leave-one-out
From the steric and electrostatic contour map (Fig. 6A), we could
observe that one kidney-like region exists at the center of near-end
aromatic ring as illustrated with green contour parts. Thus, the in-
troduction of a bulky moiety on amide carbonyl group would improve
the inhibitory activity, as was shown on the compound 34
9

M. Li, et al.
Bioorganic&MedicinalChemistryLetters30(2020)127632
Fig. 6. Field-based QSAR models for the ligands (A) Gaussian steric field: Green (+) and Gaussian electrostatic field: Blue (+), Red (−); (B) Gaussian hydrophobic
field: Yellow (+), White (−); (C) Gaussian hbond acceptor field: Red (+), Magenta (−); (D) Gaussian hbond donor field: Purple (+), Cyan (−). (For interpretation
of the references to colour in this figure legend, the reader is referred to the web version of this article.)
(pIC50 = 4.92) and the compound 40 (pIC50 = 6.53). Besides, the size
of the substitution was not strictly limited. Although the dual aromatic
ring was rigid and larger, the activity was not significantly decreased,
which could be observed on the compound 40 (pIC50 = 6.53) and the
compound 44 (pIC50 = 6.47).
contributed the most to the activity regulation.
Compounds 47 was docked to the orthosteric binding site of the
human M2 receptor X-ray structure in the inactive state in complex with
quinuclidinyl benzylate (QNB) (PDB ID: 3UON). The molecular docking
was performed with the standard precision (SP) approach of Glide in
Schrödinger Suite with default settings using the grid based on the
cocrystallized ligand, QNB. The selected binding pose and interaction
diagram of compound 47 are shown in Fig. 7. The main contact of
compound 47 with the human muscarinic M2 receptor is a hydrogen
bond interaction between the amide nitrogen of the ligand and the
conserved Tyr104 (3.33) (Ballesteros–Weinstein nomenclature) from
the third transmembrane helix, as is typical for orthosteric ligands of
aminergic GPCRs. Moreover, there is pi-pi stacking interaction between
the far-end aromatic ring of the core amide group and the side chain of
Tyr104. In our model, there was no strong interaction with the known
key residues (Asp103 and Asn404) for other ligands, e.g. AFDX 384,
which indicated that these ligands could bind to a partially different
orthosteric site. That could be why the affinity of our ligands were
much lower than AFDX 384 and it was confirmed that compound 47
had a different pose from AFDX 384, when comparing the docking
model with the crystal structure of M2-AFDX 384 complex (PDB ID:
5ZKB).15
From the electrostatic field contour map (Fig. 6A), we noticed that
electropositive group, with blue contour part, at the near-end aromatic
ring might contribute to the activity. The red contour at far-end ni-
trogen indicated that introducing electronegative group or electron
donating group substitution on this aromatic ring would potentially
improve the activity. For example, the compound 9 (pIC50 = 5.03)
bearing with hydroxyl group substitution could be more active than
that of compound 18 with fluoro-substitution (pIC50 = 4.12).
The hydrophobicity contour map, presented in Fig. 6B, is consistent
with the hydrophobicity of the aromatic lid of M2 receptor, which
clearly demonstrated that the amide-aromatic ring area and the far-end
aromatic area were in the solvent-inaccessible hydrophobic pocket. The
yellow contour (Fig. 6B) around these two positions was recognized as a
preference for the hydrophobic group as described in the interaction
mode from docking. The introduction of the hydrophilic moiety at ei-
ther region could result in decreased activity. This trend could be seen
from the compound 11 (pIC50 = 6.36) VS 13 (pIC50 = 3.19) and 25
(pIC50 = 4.53), respectively.
Because Phe181 was the specific residue, which was only located
around the orthosteric site of M2 subtype, we considered that Phe181
should play important roles in the selectivity of compound 47. The
aromatic ring at the near-end of the amide group is surrounded by the
famous aromatic lid, including Tyr403, Phe181 and Tyr177 (Fig. 7A
and 7B). The aromatic lid could change the binding kinetics for both
association and dissociation, which was also accordant with the faster
kinetics of M2 subtype in radioligand binding assay. Comparing to the
interactions of non-selective QNB and tiotropium,29,33,36 interactions
with aromatic lid in M2 receptor could allow the accommodation of
bulkier ligand and the bulkier moiety could contribute to the se-
lectivity, which has been confirmed in another new M2-selective an-
tagonist with crystal structures.15
The red contour maps at the amide carbonyl group showed the
importance of oxygen atom as H-bond acceptor. The magenta contour
maps around linker suggested that H-bond acceptors were disfavored at
these places (Fig. 6C). The purple contour in Fig. 6D suggested that H-
bond donor groups would be beneficial to the amide nitrogen. For ex-
ample, the H-bond was observed between amide nitrogen and hydroxyl
group of Tyr104 side chain, which was vital for the combination of the
ligands with M2 receptor. Besides, the cyan contour maps around the
compound showed that H-bond donors were disfavored, which could be
accordant with the hydrophobic maps. Therefore, field contribution of
the defined features in the final model revealed that H-bond donor field
was the fundamental for the activity and steric field, hydrophobic field
10

M. Li, et al.
Bioorganic&MedicinalChemistryLetters30(2020)127632
Fig. 7. Interaction model of compound 47 and M2 subtype.
In conclusion, by using virtual screening, we discovered a novel lead Acknowledgments
compound featuring a specific binding and potential to mAChRs. The
following optimization based on this scaffold yielded a class of novel
ligands featuring submicromolar IC50 for M2 subtype. Among these,
ligand 47 exhibits a specific M2 binding selectivity and high drug po-
tential (IC50). Furthermore, the binding affinity, potential and QSAR
obtained from this study will also give the valuable information in
further development of M2 subtype ligands.
This work was supported by National Natural Science Foundation of
China (81601536 to Dewei Tang, 81503174 to Jianrong Xu, 81973297
to Hao Wang and 81872841 to Hongzhuan Chen) and by Shanghai
Committee of Science and Technology (19JC1413100 to Hao Wang).
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://
doi.org/10.1016/j.bmcl.2020.127632.
Declaration of Competing Interest
References
The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influ-
ence the work reported in this paper.
1. Hauser AS, Chavali S, Masuho I, et al. Pharmacogenomics of GPCR drug targets. Cell.
2018;172:41–54 e19.
11

M. Li, et al.
Bioorganic&MedicinalChemistryLetters30(2020)127632
2. Wess J, Eglen RM, Gautam D. Muscarinic acetylcholine receptors: mutant mice
provide new insights for drug development. Nat Rev Drug Discov. 2007;6(9):721–733.
3. VanPatten S, Al-Abed Y. The challenges of modulating the 'rest and digest' system:
acetylcholine receptors as drug targets. Drug Discov Today. 2017;22(1):97–104.
4. Moran SP, Maksymetz J, Conn PJ. Targeting muscarinic acetylcholine receptors for
the treatment of psychiatric and neurological disorders. Trends Pharmacol Sci.
2019;40(12):1006–1020.
impaired. J Neurosci. 1995;15(2):1455–1462.
20. Doods H, Entzeroth M, Ziegler H, et al. Characterization of BIBN 99: a lipophilic and
selective muscarinic M2 receptor antagonist. Eur J Pharmacol. 1993;242(1):23–30.
21. Hernandez CM, Terry Jr AV. Repeated nicotine exposure in rats: effects on memory
function, cholinergic markers and nerve growth factor. Neuroscience.
2005;130(4):997–1012.
22. Pittel Z, Lazar S, Gez R, Chapman S. Early changes in M2 muscarinic acetylcholine
receptors (mAChRs) induced by sarin intoxication may be linked to long lasting
neurological effects. Neurotoxicology. 2018;65:248–254.
5. Kruse AC, Kobilka BK, Gautam D, Sexton PM, Christopoulos A, Wess J. Muscarinic
acetylcholine receptors: novel opportunities for drug development. Nat Rev Drug
Discov. 2014;13(7):549–560.
23. Watson N, Barnes PJ, Maclagan J. Actions of methoctramine, a muscarinic M2 re-
ceptor antagonist, on muscarinic and nicotinic cholinoceptors in guinea-pig airways
in vivo and in vitro. Br J Pharmacol. 1992;105(1):107–112.
6. Thomsen M, Sørensen G, Dencker D. Physiological roles of CNS muscarinic receptors
gained from knockout mice. Neuropharmacology. 2018;136(Pt C):411–420.
7. Valuskova P, Farar V, Forczek S, Krizova I, Myslivecek J. Autoradiography of 3H-
pirenzepine and 3H-AFDX-384 in Mouse Brain Regions: possible insights into M1,
M2, and M4 muscarinic receptors distribution. Front Pharmacol. 2018;9:124.
8. Korczynska M, Clark MJ, Valant C, et al. Structure-based discovery of selective po-
sitive allosteric modulators of antagonists for the M2 muscarinic acetylcholine re-
ceptor. Proc Natl Acad Sci. 2018;115(10):E2419–E2428.
24. Wess J, Angeli P, Melchiorre C, Moser U, Mutschler E, Lambrecht G. Methoctramine
selectively blocks cardiac muscarinic M2 receptors in vivo. Naunyn Schmiedebergs
Arch Pharmacol. 1988;338(3):246–249.
25. Jakubík J, Zimčík P, Randáková A, Fuksová K, El-Fakahany EE, Doležal V. Molecular
mechanisms of methoctramine binding and selectivity at muscarinic acetylcholine
receptors. Mol Pharmacol. 2014;86(2):180–192.
9. Greenlee W, Clader J, Asberom T, et al. Muscarinic agonists and antagonists in the
treatment of Alzheimer's disease. Farmaco. 2001;56(4):247–250.
26. Gamberini MT, Bolognesi ML, Nasello AG. The modulatory role of M2 muscarinic
receptor on apomorphine-induced yawning and genital grooming. Neurosci Lett.
2012;531(2):91–95.
10. Rowe WB, O'Donnell JP, Pearson D, Rose GM, Meaney MJ, Quirion R. Long-term
effects of BIBN-99, a selective muscarinic M2 receptor antagonist, on improving
spatial memory performance in aged cognitively impaired rats. Behav Brain Res.
2003;145(1–2):171–178.
27. Chiarini A, Budriesi R, Bolognesi ML, Minarini A, Melchiorre C. In vitro character-
ization of tripitramine, a polymethylene tetraamine displaying high selectivity and
affinity for muscarinic M2 receptors. Br J Pharmacol. 1995;114(7):1507–1517.
28. Angeli P, Cantalamessa F, Gulini U, Melchiorre C. Selective blockade of muscarinic
M2 receptors in vivo by the new antagonist tripitramine. Naunyn Schmiedebergs Arch
Pharmacol. 1995;352(3):304–307.
11. Muramatsu I, Uwada J, Yoshiki H, et al. Novel regulatory systems for acetylcholine
release in rat striatum and anti-Alzheimer's disease drugs. J Neurochem.
2019;149(5):605–623.
12. Jakubik J, El-Fakahany EE. Current advances in allosteric modulation of muscarinic
receptors. Biomolecules. 2020;10(2) pii: E325.
29. Haga K, Kruse AC, Asada H, et al. Structure of the human M2 muscarinic acet-
ylcholine receptor bound to an antagonist. Nature. 2012;482(7386):547–551.
30. Kruse AC, Ring AM, Manglik A, et al. Activation and allosteric modulation of a
muscarinic acetylcholine receptor. Nature. 2013;504(7478):101–106.
31. Staus DP, Hu H, Robertson MJ, et al. Structure of the M2 muscarinic receptor–β-
arrestin complex in a lipid nanodisc. Nature. 2020;579(7798):297–302.
32. Thal DM, Sun B, Feng D, et al. Crystal structures of the M1 and M4 muscarinic
acetylcholine receptors. Nature. 2016;531(7594):335–340.
13. Bock A, Schrage R, Mohr K. Allosteric modulators targeting CNS muscarinic re-
ceptors. Neuropharmacology. 2018;136(Pt C):427–437.
14. Liu H, Hofmann J, Fish I, et al. Structure-guided development of selective M3
muscarinic acetylcholine receptor antagonists. Proc Natl Acad Sci.
2018;115(47):12046–12050.
15. Suno R, Lee S, Maeda S, et al. Structural insights into the subtype-selective antagonist
binding to the M2 muscarinic receptor. Nat Chem Biol. 2018;14(12):1150–1158.
16. Del Bello F, Bonifazi A, Quaglia W, et al. Mode of interaction of 1,4-dioxane agonists
at the M2 and M3 muscarinic receptor orthosteric sites. Bioorg Med Chem Lett.
2014;24(15):3255–3259.
33. Vuckovic Z, Gentry PR, Berizzi AE, et al. Crystal structure of the M5 muscarinic
acetylcholine receptor. Proc Natl Acad Sci. 2019;116(51):26001–26007.
34. Maeda S, Qu Q, Robertson MJ, Skiniotis G, Kobilka BK. Structures of the M1 and M2
muscarinic acetylcholine receptor/G-protein complexes. Science.
17. Del Bello F, Bonifazi A, Giorgioni G, et al. Novel potent muscarinic receptor an-
tagonists: investigation on the nature of lipophilic substituents in the 5- and/or 6-
positions of the 1,4-dioxane nucleus. J Med Chem. 2020;63(11):5763–5782.
18. Lachowicz JE, Duffy RA, Ruperto V, et al. Facilitation of acetylcholine release and
improvement in cognition by a selective M2 muscarinic antagonist, SCH 72788. Life
Sci. 2001;68(22–23):2585–2592.
2019;364(6440):552–557.
35. Cappel D, Dixon SL, Sherman W, Duan J. Exploring conformational search protocols
for ligand-based virtual screening and 3-D QSAR modeling. J Comput Aided Mol Des.
2015;29(2):165–182.
36. Burger WA, Sexton PM, Christopoulos A, Thal DM. Toward an understanding of the
structural basis of allostery in muscarinic acetylcholine receptors. J Gen Physiol.
2018;150(10):1360–1372.
19. Quirion R, Wilson A, Rowe W, et al. Facilitation of acetylcholine release and cog-
nitive performance by an M(2)-muscarinic receptor antagonist in aged memory-
12