Carbachol dimers with primary carbamate groups as homobivalent modulators of muscarinic receptors

Article abstract of DOI:10.1016/j.ejphar.2020.173183

Although agonists and antagonists of muscarinic receptors have been known for long time, there is renewed interest in compounds (such as allosteric or bitopic ligands, or biased agonists) able to differently and selectively modulate these receptors. As a continuation of our previous research, we designed a new series of dimers of the well-known cholinergic agonist carbachol. The new compounds were tested on the five cloned human muscarinic receptors (hM_1–5) expressed in CHO cells by means of equilibrium binding experiments, showing a dependence of the binding affinity on the length and position of the linker connecting the two monomers. Kinetic binding studies revealed that some of the tested compounds were able to slow the rate of NMS dissociation, suggesting allosteric behavior, also supported by docking simulations. Assessment of ERK1/2 phosphorylation on hM₁, hM₂ and hM₃ activation showed that the new compounds are endowed with muscarinic antagonist properties. At hM₂ receptors, some compounds were able to stimulate GTPγS binding but not cAMP accumulation, suggesting a biased behavior. Classification, Molecular and cellular pharmacology.

Full text of DOI:10.1016/j.ejphar.2020.173183

Journal Pre-proof
Carbachol dimers with primary carbamate groups as homobivalent modulators of
muscarinic receptors
Rosanna Matucci, Cristina Bellucci, Maria Vittoria Martino, Marta Nesi, Dina Manetti,
Jessica Welzel, Ulrike Bartz, Janine Holze, Christian Tränkle, Klaus Mohr, Angelica
Mazzolari, Giulio Vistoli, Silvia Dei, Elisabetta Teodori, Maria Novella Romanelli
PII:
S0014-2999(20)30275-2
DOI:
https://doi.org/10.1016/j.ejphar.2020.173183
EJP 173183
Reference:
To appear in: European Journal of Pharmacology
Received Date: 9 October 2019
Revised Date: 23 March 2020
Accepted Date: 8 May 2020
Please cite this article as: Matucci, R., Bellucci, C., Martino, M.V., Nesi, M., Manetti, D., Welzel, J.,
Bartz, U., Holze, J., Tränkle, C., Mohr, K., Mazzolari, A., Vistoli, G., Dei, S., Teodori, E., Romanelli, M.N.,
Carbachol dimers with primary carbamate groups as homobivalent modulators of muscarinic receptors,
European Journal of Pharmacology (2020), doi: https://doi.org/10.1016/j.ejphar.2020.173183.
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 Published by Elsevier B.V.

Author Contributions:
RM, CB, MN, CT, KM, ET, and MNR designed the experiments and analyzed the data. MVM,
DM, SD performed the synthetic work. RM, CB, MN, JH, JW, UB performed the biological
experiments. AM, GV designed and performed docking studies. RM, MNR, CT, GV wrote the
manuscript. All authors reviewed the manuscript and agreed with its content

Carbachol dimers with primary carbamate groups as homobivalent modulators of
muscarinic receptors.
Rosanna Matucci,a Cristina Bellucci,b Maria Vittoria Martino,b Marta Nesi,a Dina Manetti,b
Jessica Welzel,c Ulrike Bartz,d Janine Holze,c Christian Tränkle,c Klaus Mohr,c Angelica
Mazzolari, e Giulio Vistoli, e Silvia Dei,b Elisabetta Teodori,b Maria Novella Romanelli b
aDepartment of Neuroscience, Psychology, Drug Research and Child’s Health, Section of
Pharmacology and Toxicology, University of Florence, Viale G. Pieraccini 6, 50139 Firenze,
Italy.
bDepartment of Neuroscience, Psychology, Drug Research and Child’s Health, Section of
Pharmaceutical and Nutraceutical Sciences, University of Florence, Via Ugo Schiff 6, 50019
Sesto Fiorentino, Italy.
cDepartment of Pharmacology & Toxicology, Institute of Pharmacy, University of Bonn,
Gerhard-Domagk-Str. 3, 53121 Bonn, Germany.
dDepartment of Natural Sciences, H-BRS University of Applied Sciences, von-Liebig-Str. 20,
53359 Rheinbach, Germany.
eDepartment of Pharmaceutical Sciences, University of Milan, Via Mangiagalli 25, 20133
Milano, Italy.
Corresponding author
Rosanna Matucci, Department of Neuroscience, Psychology, Drug Research and Child’s
Health, Section of Pharmacology and Toxicology, University of Florence, Viale G. Pieraccini
6, 50139 Firenze, Italy. Tel +39 055 2758280 e-mail: rosanna.matucci@unifi.it
1

Classification
Molecular and cellular pharmacology
Abstract
Although agonists and antagonists of muscarinic receptors have been known for long time,
there is renewed interest in compounds (such as allosteric or bitopic ligands, or biased
agonists) able to differently and selectively modulate these receptors. As a continuation of our
previous research, we designed a new series of dimers of the well-known cholinergic agonist
carbachol. The new compounds were tested on the five cloned human muscarinic receptors
(hM1–5) expressed in CHO cells by means of equilibrium binding experiments, showing a
dependence of the binding affinity on the length and position of the linker connecting the two
monomers. Kinetic binding studies revealed that some of the tested compounds were able to
slow the rate of NMS dissociation, suggesting allosteric behavior, also supported by docking
simulations. Assessment of ERK1/2 phosphorylation on hM1, hM2 and hM3 activation
showed that the new compounds are endowed with muscarinic antagonist properties. At hM2
receptors, some compounds were able to stimulate GTPγS binding but not cAMP
accumulation, suggesting a biased behavior.
Keywords
Muscarinic acetylcholine receptor (mAChR), carbachol dimers, allosteric modulation
Chemical compounds studied in this article:
2

Carbachol chloride (PubChem CID: 5831); Acethylcholine chloride (PubChem CID: 6060);
Hexamethonium chloride (PubChem CID: 93550); Gallamine triethiodide (PubChem CID:
6172); McN-A-343 chloride (PubChem CID: 5926); Atropine sulphate (PubChem CID:
64663).
1. Introduction
Five different muscarinic acetylcholine receptor (mAChR) subtypes are known (M1-M5)
which are widely distributed inside and outside the CNS and involved in many physiological
processes. Muscarinic ligands are mainly used in clinical settings to treat chronic obstructive
pulmonary disease, overactive bladder and Sjögren’s syndrome, but selective agonists or
antagonists of these proteins may be potentially useful in several other disorders or conditions
(Kruse et al., 2014; Wess et al., 2007).
Muscarinic receptors can be modulated by means of ligands interacting with the orthosteric or
allosteric site(s) whose topography has been determined from recent crystal structures of
ligand-complexed M1-M4 receptors (Haga et al., 2012; Kruse et al., 2012; Kruse et al., 2013;
Thal et al., 2016). Modulation from an allosteric site is particularly attractive since these
regions are much less conserved than the orthosteric one for the endogenous neurotransmitter;
with allosteric ligands therefore it should be possible to achieve the subtype-selectivity which
has been elusive with orthosteric modulators (Bock et al., 2018). The location of an allosteric
site near the orthosteric one led to the design of dualsteric/bitopic ligands (i.e. compounds
which can simultaneously bind to both sites) obtained by hybridizing non-selective orthosteric
ligands with subtype-selective allosteric modulators. These bivalent ligands can display
interesting properties, such as subtype selectivity, functional selectivity, and protean agonism
(Bock et al., 2018; De Min et al., 2017; Schrage and Kostenis, 2017; Valant et al., 2012). A
3

muscarinic agonist, 4-[[(3-chlorophenyl)carbamoyl]oxy]-N,N,N-trimethylbut-2-yn-1-
ammonium chloride (McN-A-343) (Mitchelson, 2012), was recently recognized as a bitopic
ligand (May et al., 2007; Valant et al., 2008). This compound showed about 10-fold higher
activity on G15-coupled M2 receptors compared with Gi-coupled M2 receptors, displaying
functional selectivity. Some hybrids composed of the orthosteric agonist xanomeline and the
allosteric agonist 1-[3-(4-butyl-1-piperidinyl)propyl]-3,4-dihydro-2(1H)-quinolinone (77-LH-
28-1) showed different abilities to engage Gq or β-arrestin via M1 mAChRs, compared to both
lead compounds (Bonifazi et al., 2014). A recent paper proposed that the extent of closure of
the extracellular allosteric binding site influences the intracellular coupling to distinct
signaling pathways, suggesting a mechanistic explanation of biased agonism (Bermudez et al.,
2017). However, stimulus bias has been displayed also by orthosteric muscarinic ligands (see
refs Gregory et al., 2010 and Pronin et al., 2017 as examples).
In a previous paper we described a series of carbachol dimers (compounds 1a-f, 2a-f, Fig. 1);
kinetic binding studies and docking simulations have suggested a bitopic behavior for some of
them (Matucci et al., 2016). Compounds 1a-f and 2a-f are symmetric dimers in which the two
monomers are connected through the carbamic nitrogen atoms by means of a methylene chain
of variable length. In this paper we report the activity of their isomers, having the two agonist
units linked by a polymethylene spacer connecting the cationic nitrogen atoms (compounds
3a-e and 4a-e, Fig. 1); tertiary amines and ammonium derivatives were both prepared. In
addition, a hybrid compound has been synthesized (5) where a 10-methylene chain connects
the carbamic nitrogen atom of one monomer with the basic nitrogen atom of the second one.
To verify the importance of the presence of both agonist units, compound 6 and the synthetic
intermediates 7d,e were also tested. The activity of the compounds was measured in binding
and functional studies on CHO cells expressing the five hM1-5 receptors; docking methods
4

were used to rationalize the outcome of binding experiments. The results were compared to
those found for the previously synthesized 1a-f and 2a-f.
FIG. 1 NEAR HERE
2. Materials and methods
2.1. Drugs
The following drugs were used: carbachol chloride, acetylcholine chloride, examethonium
chloride, McN-A-343 chloride, gallamine triethiodide, atropine sulphate salt monohydrate and
(-)-scopolamine methylbromide purchased from Sigma-Aldrich SRL, Milano, Italy; [3H]N-
methylscopolamine chloride specific activity range 2,590–3,200 GBq/mmol and
[35S]guanosine-5’-γ-thiotriphospate specific activity 46.25 TBq/mmol from Perkin-Elmer Life
and Analytical Science, Monza, Milano, Italy. Pertussis toxin (PTX) was purchased from
Biotrend (Cologne, Germany). All other reagents were purchased from Sigma-Aldrich SRL
(Milano, Italy) unless stated otherwise. Compounds 3a-e, 4a-e, 5, 6 and 7d-e were prepared
as reported in Appendix A.
2.2 Biological studies
Equilibrium radioligand binding assays, dissociation kinetic assays, guinea-pig ileum
preparations, and the Extracellular Signal-Regulated Kinase (ERK1/2) Phosphorylation
Assays were performed as previously described (Matucci et al., 2016). The organs used in this
article derive from programs for sharing organs and tissues of animals sacrificed in the animal
5

house facility of the University of Florence, in compliance with the principle of reduction.
The experiments described was approved from Institutional Animal Care and Use Committee
(IACUC) of the University of Florence and from the Italian Ministry of Health (Authorization
No. 54/2014-B), issued under the previous Italian Law Decree (No. 116/1992) on animal
testing. With regard to data analysis for compounds 3a, 3b and 4a, pIC0.5 values (i.e. the
concentration at the inflection point of the respective inhibition curves) are reported instead of
pKi because these compounds did not completely displace [3H]NMS equilibrium binding at
all muscarinic receptors subtypes studied, thus suggesting a non-competitive behavior. In all
other cases, data from these experiments were fitted to a parametric function to derive best
estimates of the IC50 and slope factor; IC50 values were then converted to binding constant Ki
according to Cheng-Prusoff equation.
These parameters are presented as mean ± S.E.M. of at least three experiments, each one
performed in duplicate, unless otherwise noted.
2.2.1 Data analysis:
Data generated from binding assays were analysed using Prism 5.02 (GraphPad Software Inc.,
San Diego, CA). Data points were fitted to models using nonlinear regression equations.
Some equations, regarding dissociation kinetic assays, were derived by Dr Nigel Birdsall and
then formulated to be introduced into GraphPad Prism. According to the protocol developed
by Lazareno and Birdsall (1995), one point kinetic data were analyzed as a function of time
and then in order to obtain estimates of the affinity of an allosteric agent for the [3H]NMS-
occupied receptor (Kocc) in a single step (equation 1).
koff/(1+Kocc × X)
(1)
koff is the dissociation rate constant of the radioligand in the absence of X; X refers to the log
6

concentration of the allosteric ligand and log Kocc is its log affinity constant for the occupied
receptor. This equation can also be used to directly estimate log Kocc by analyzing the amount
of [3H]NMS remaining at certain times at different concentrations of the allosteric modulators
(see Lazareno and Birdsall, 1995).
2.3. GTPγS binding assay
GTPγS binding assay was performed as reported in previous papers (see ref Jäger et al.,
2007). Curves were fitted with a “four parameter logistic function”; the respective curve
slopes were not different from unity (F-test, P>0.05).
2.4. cAMP-accumulation
Quantification of agonist-induced rise of intracellular cAMP was performed using CHO-hM2
cells pretreated with 100 ng/ml-1 PTX for 16-22 h, as described previously using a HTRF-
cAMP dynamic kit (Cisbio, Bagnols-sur-Cèze, France) following the manufacturer´s
instructions. cAMP content was detected dispensing 50,000 cells per well in a buffer
containing Hanks´ balanced salt solution with 20 mM HEPES and 1 mM of the
phosphodiesterase inhibitor IBMX. Fluorescence was quantified on a Mithras LB 940
multimode reader (Berthold Technologies, Bad Wildbad, Germany). Levels of cAMP were
normalized to the amount of cAMP generated by 100 µM Acetylcholine (ACh).
2.5. Computational studies
7

With a view to investigating the molecular recognition of the proposed molecules, docking
simulations were performed by applying the same computational protocols already reported in
the previous study (Matucci et al., 2016), focusing attention on the hM2 subtype. In detail,
docking calculations involved the resolved structures in both active (PDB Id: 4MQT) and
inactive (PDB Id: 3UON) conformations. The inactive structure was simulated retaining the
co-crystallized QNB inhibitor in order to mimic the experimental conditions of the kinetic
studies. Briefly, docking simulations were performed using PLANTS and focusing the search
on a sphere large enough to encompass orthosteric and allosteric binding sites at the same
time. For each ligand, 10 poses were generated using the ChemPLP scoring function and the
obtained results were analyzed by ReScore+ (Vistoli et al., 2017).
3. Results
Compounds 3a-e and 4a-e, 5, 6 and 7d-e have been synthesized using standard methods
(Appendix A). Derivatives 3a-e and 4a-e represent structural isomers of the previously
described 1a-f and 2a-f, differing for the position of attachment of the polymethylene spacer;
it was envisaged that with this new arrangement, the carbammic moiety could be free to
establish the key contacts within the binding site, which appeared to be precluded for 1 and 2
(Matucci et al., 2016).
3.1. Equilibrium binding affinity
The affinity of compounds 3a-e, 4a-e, 5, 6 and 7d-e for the five human muscarinic receptor
subtypes (hM1-hM5), measured in equilibrium binding experiments, is reported in Table 1 as
pKi or pIC0.5 and illustrated in Fig. S1 carbachol (CCh) and McN-A-343 have been taken as
reference compounds.
8

TABLE 1 NEAR HERE
All the compounds were able to displace the radioligand [3H]NMS binding at all muscarinic
receptor subtypes, in a concentration-dependent way and with different inhibitory propensity.
Symmetric dimers 3c-e and 4b-e completely displaced specific radioligand binding at all
receptor subtypes, while 3a,b and 4a did not, even at millimolar concentrations, giving pIC0.5
values <4 at some receptor subtypes (hM2, hM4 and hM5 for 3a, hM3, hM4 and hM5 for 4a,
hM4 for 3b). When n ≥7, structural dimerization generally improved binding affinity with
respect to carbachol, with the exception of compound 3c on hM2; among the shorter
derivatives, 4b (n = 5) showed pKi values higher than carbachol only on hM1 and hM3.
The affinity of bases 3a-e regularly increased with increasing the length of the chain in all
receptor subtypes (Fig. S1A). In contrast, for methiodides 4a-e (Fig. S1B), in some cases the
increment brought by two methylene units was small or absent. This was the case for 4c-4d (n
= 7 and 9, respectively) at hM4 and hM5, while at hM3 affinity was slightly decreased.
Nevertheless, 4e (n = 11) was the most potent compound at each subtype, being equipotent
with 4d on the hM2 receptor
Contrary to what happened in the previously synthesized 1 and 2 series, where methiodides
had higher affinity measures than the corresponding bases, for compounds 3 and 4 a
permanent positive charge increased affinity only for a spacer length of n=7: affinity
measures of 4c were higher than those of 3c at all subtypes. For compounds with n= 3, 5, 9
and 11 the affinity measures of bases were equal or higher than those of methiodides, apart
from those cases when affinity measures were <4.
In general, symmetric dimers 3a-e and 4a-e did not show subtype selectivity of binding; only
for the longer compounds (3d-e, 4d-e, n = 9, 11) was the difference in affinity measures on
9

some subtypes higher than one order of magnitude. With the exclusion of compounds 3a-b
and 4a, whose affinity measures could not be calculated for all receptor subtypes, symmetric
dimer compounds showed the lowest apparent affinity measure at hM3 and hM5 receptors
(Table 1, Fig. S1A,B).
The contribution to affinity of the second carbamate moiety was investigated by testing
compounds 6, 7d and 7e. N-undecyl derivative 6 showed similar binding constants at all
subtypes (pKi values ranging from 5.82±0.08 on hM5 to 6.26±0.04 on hM1), with pKi values
intermediate between those of 3c and 3d. With respect to 6, compounds 7d and 7e possessed
an additional basic moiety, which did not improve affinity but was actually detrimental for it
at hM3. With respect to bis-carbamate 3d, monocarbamate 7d showed lower affinity on hM1
and hM2 but not on hM3 and hM5, while the affinity of 7e was much lower than that of 3e on
all subtypes (Table 1, Fig. S1C). This behavior suggested that the second carbachol unit gave
an important contribution to the interaction of the long derivative 3e (n=11) with all subtypes,
while for 3d (n = 9) its role was less important and limited to the hM1 and hM2 subtypes.
The hybrid compound 5, carrying a 10-methylene chain as a linker, showed similar binding
constants at all subtypes (pKi values ranging from 6.33 at hM2 to 6.90 at hM1) (Table 1, Fig.
S1C). Compared to the homodimers having similar linker’s length, the pKi values of 5 were
intermediate between those of 3d (n=9) and 3e (n=11) at hM1 and hM4 but not at hM2, hM3
and hM5; with respect to the previously described 1d (n=9, pKi in the range 6.53-7.89) and 1e
(n=11, pKi in the range 6.29-7.16), the binding measure of 5 was lower. Compound 5 showed
the lowest pKi value on hM2, its profile resembling that of compounds 1-2 (Matucci et al.,
2016) and not that of 3-4; the latter showed the lowest affinity measures at hM3 and hM5
receptors.
3.2. Kinetic binding studies
10

In order to detect a possible interaction with an allosteric site, the ability of selected
compounds to affect the [3H]NMS dissociation rate was evaluated, using a one-point kinetic
protocol as previously reported (Matucci et al., 2016; Lazareno and Birdsall, 1995); gallamine
and McN-A-343 were taken as reference molecules. Representative graphs are shown in Fig.
2 for 3e and gallamine.
FIG. 2 NEAR HERE
Panels A-B show the increase of the specifically M2 receptor-bound [3H]NMS, due to the
reduction of its dissociation rate by the ligand studied, as a function of time with increasing
concentrations of the modulators. In Fig. 2D the increase of [3H]NMS bound to the receptor is
shown as a function of increasing concentrations of allosteric modulators. The top data point
(shown at log [agent] = -1) represents [3H]NMS binding before dissociation starts. The curves
extrapolate to this value, implying that these compounds could substantially slow down
[3H]NMS dissociation at a sufficiently high concentration. Like gallamine, 3e, taken as an
example, markedly inhibited the dissociation of [3H]NMS from hM2 receptors, giving rise to
an increase in residual [3H]NMS binding at the time point selected (t= 20 min), with the curve
being well defined. Analysis of the kinetic data in Fig. 2A-B, according to Lazareno and
Birdsall (Lazareno and Birdsall, 1995), gave logKocc values which represent the binding
affinity at the NMS-occupied receptor and are reported in Table 2 and in Fig.S2.
TABLE 2 NEAR HERE
11

As expected, gallamine showed preference for the NMS-occupied hM2 subtype (Table 2). All
tested compounds were able to slow the [3H]NMS dissociation rate, but with different potency
depending on their structure and the receptor subtype.
The change in activity of compounds 3a-e with respect to the linker length (Table 2, Fig.
S2A) was similar on all subtypes, 3c being the least active compound (n = 7) and 3d (n = 9)
the most active. Methiodides 4a-e (Table 2, Fig. S2B) showed a similar trend on hM1, on
which the logKocc values for 4d and 4e were similar. On the contrary, at hM2 receptors the
apparent allosteric affinity of methiodides 4a-e regularly increased with chain length, while at
the other subtypes the trend was different and uneven. The longest compound 4e (n = 11) was
the most active quaternary derivative on all subtypes except hM3 and hM5.
A permanent positive charge had no clear-cut effect on radioligand dissociation: for short
compounds 3a,b and 4a,b (n = 3, 5) the difference in logKocc values between bases and
quaternary compounds was very small. A difference greater than 0.5 log unit could be seen
for 3c-4c (n = 7), at all subtypes except hM1, and for 3e-4e (n = 11) only on hM1; at these
subtypes, methiodides 4c and 4e were more potent than the corresponding bases 3c and 3e.
Base 3d (n = 9) was more potent than its quaternary ammonium derivative 4d on all subtypes.
As a general trend, the logKocc values of longer derivatives 3c-e and 4c-e (n = 9-11) for the
occupied receptor were lower at hM3 and hM5 compared to the other receptor subtypes. This
affinity profile was found also for the previously described series 1 and 2, and it was shared
also by hybrid 5 and amines 7d,e (Table 2). Among the tested compounds, 3d appeared to be
the most potent one on all the receptor subtypes.
When n = 9, the presence of both carbamate groups increased the retarding action via the
allosteric site: logKocc estimates for compound 3d were higher than those of 6 and 7d at all
receptor subtypes, with large differences at hM2 and hM4 and, with respect to 6, also at hM1.
When n = 11 (compound 3e), the contribution of the second carbamate moiety was significant
12

with respect to 6 only at hM1 and hM2, as it increased the affinity measure 8.5 and 11 times,
respectively. With respect to 7e, the second carbamate moiety did not affect activity
significantly, but at hM1 receptors it was actually detrimental, since affinity of 3e was 3-fold
lower than that of 7e.
3.3. Functional studies on isolated guinea pig ileum
As done for the previous series (Matucci, 2016), the new compounds were tested for their
functional activity on guinea pig ileum, tissue in which M3 than M2 receptor sites are mainly
present (Barocelli et al., 1993). All compounds, with the exception of 3e and 5, were able to
evoke weak smooth muscle contractions (data not shown) which, however, were not
antagonized by 1 nM atropine (which is able to antagonize the contractions induced by ACh
and CCh), nor by the nicotinic antagonist hexamethonium (30 µM). This behavior could
imply the involvement of a receptor system different from the cholinergic one, which was not
further investigated. Compounds 5, 6, 7d and 7e behaved as weak antagonists (pKB <5), being
able to reduce contractions induced by ACh. Compound 3e was devoid of activity both as
agonist and antagonist, despite the micromolar affinity displayed in equilibrium binding
studies (cf. Table 1) and a measurable binding propensity to NMS-occupied hM1-5 receptors
as quantified by logKocc (cf. Table 2).
3.4. ERK 1/2 phosphorylation
To determine the functional activity of the compounds on muscarinic receptors, selected
molecules (3d-e, 4d-e, 7d-e) were evaluated in ERK 1/2 phosphorylation assays in CHO cells
stably expressing hM1, hM2 and hM3 receptor subtypes. When tested alone up to a 100 ꢀM
13

concentration, none of the compounds displayed agonistic activity (data not shown). Thus,
their antagonistic properties against 10 nM ACh were assessed. The results, reported in Fig.
3A-C, show that a 100 ꢀM concentration of the compounds was able to reduce ACh-induced
ERK 1/2 phosphorylation to different extents depending on the receptor subtype. At the hM1
receptor (Fig. 3A) all tested compounds behaved as antagonists with the exception of 7d,
which was devoid of activity. At the hM2 receptor (Fig. 3B) compounds 3d and 4e effectively
reduced ACh stimulation, while the antagonistic effect of 4d, 3e, 7d and 7e was weaker or
absent. At the hM3 receptor (Fig. 3C) all tested compounds displayed an antagonistic effect,
which was strong for 3d, 4e, 7d and 7e and much weaker, but still significant, for 4d and 3e.
These results highlight for some compounds a certain degree of selectivity: for instance, 7d
was able to strongly antagonize ACh-induced ERK phosphorylation at hM3 but it was
ineffective at hM1 and hM2 receptors.
FIG. 3 NEAR HERE
3.5. Interaction with G proteins via hM2 receptors
Activation of hM2 muscarinic receptors was measured by means of [35S]GTPγS binding
assays, performed according to previously reported protocols (Jäger et al., 2007). Initial pilot
experiments had shown that specific [3H]NMS hM2 equilibrium binding as defined by 10 µM
atropine to define nonspecific binding was concentration-dependently displaced by the test-
compounds 4b-4e to 0% specific binding revealing the hM2 receptor to be a target structure
for them (data not shown).
14

Dose response curves for agonist mediated receptor activation are reported in Fig. 4;
statistical parameters obtained from curves fitted to the mean data values are summarised in
Table S1.
FIG. 4 NEAR HERE
Contrary to what happened in ERK assays, in this system test compounds 4b-e produced
receptor activation, being able to stimulate [35S]GTPγS binding, albeit with low intrinsic
activity. Their maximal effect was about 20 % of the maximum [35S]GTPγS binding produced
by the full agonists carbachol (100 µM) and acetylcholine (100 µM), thus behaving as weak
partial agonists.
To further investigate functional activity, selected compounds (4b-e) were evaluated in a
cAMP-accumulation assay applying CHO cells stably expressing the human M2-receptor
(hM2). In contrast, none of the tested compounds (4b-e) at any concentration was able to
induce a cAMP-accumulation significantly different from baseline on this Gs-signaling
pathway (data not shown). In comparison to the results of the GTPγS binding experiments,
compounds 4b-e exhibited a Gi over Gs signaling bias compared with ACh and CCh.
3.7. Docking studies
As anticipated under Methods, docking simulations were focused on the hM2 receptor by
comparing its active and inactive conformations. When comparing the here obtained docking
poses within the active hM2 conformation with those already published (Matucci et al., 2016),
one might notice a similar effect of the linker length on the observed binding modes. Indeed,
the short derivatives (n < 7) cannot simultaneously occupy both orthosteric and allosteric
15

binding sites and tend to be completely accommodated within the orthosteric binding site,
while the molecules with longer chains appeared to be able to engage both binding sites.
Unlike the previously reported dimers 1 and 2, where the carbachol unit within the orthosteric
site was always unable to elicit the same key interactions stabilized by the carbachol ligand
alone [18], compounds 3 and 4 were almost always able to optimize their contacts within the
orthosteric site.
In more detail, the ligands endowed with short linkers can assume three possible binding
modes since they are accommodated within the orthosteric site, within the allosteric pocket or
in-between. Thus, the computed poses for the shortest derivatives (3a and 4a, n = 3) are
roughly equally distributed between orthosteric and allosteric binding site even though the
poses within the orthosteric pocket always represent the top scores. The intermediate poses
represent an energy disfavored minority. Also the derivatives with n = 5 (3b and 4b) show
similar distributions with the best poses within the orthosteric site (as displayed in Fig. 5A)
even though the intermediate poses progressively increase their relevance and indeed the
ligands with n = 7 (3c and 4c) preferentially assume in-between poses by which they properly
occupies both binding pockets. Similar distributions are observed for longer derivatives even
though all these longer compounds show some poses in which they are accommodated within
the allosteric pocket only. Collectively, one may notice that the short ligands are able to
occupy the allosteric site only in a less than half of the computed complexes which never
correspond to best solution, while the longer derivatives are always able to be inserted within
the allosteric cavity.
To exemplify the above described trends, Fig. 5 compares the best putative complexes as
generated for ligands endowed with short (3b, n = 5, Fig. 5A) and long (3d, n = 9, Fig. 5B)
linker and confirms that in both complexes at least one carbachol unit was properly
16

accommodated within the orthosteric site, while only the 3d is able to insert the second unit
within the allosteric pocket.
FIG. 5 NEAR HERE
The above discussed results are in agreement with the docking simulations based on the
inactive hM2 structure in complex with the co-crystallized quinuclidyl benzylate (QNB)
inhibitor, where only the long derivatives were able to afford satisfactory complexes while
showing more superficial poses compared to the corresponding compounds of set 1 and 2. For
example, Fig. 5C shows the putative complex as computed for 3d which exhibits the highest
log Kocc value. Both carbachol units are able to contact Glu175 and differ for the arrangement
of carbamate functions since one carbamate remains in a superficial region while the second
function approach the orthosteric cavity. Notably, a carbachol unit shows within the allosteric
cavity a pose very similar to that already seen in the active hM2 structure, thus suggesting that
the arrangement of the orthosteric site does not markedly affect the accessibility of the
allosteric cavity. The short derivatives are seen to approach the orthosteric cavity where they
unsuitably interfere with the key residues of the orthosteric binding site thus competing with
the QNB binding. The reported affinity and kinetic data confirm that the involvement of this
empty space around the orthosteric pocket has an overall detrimental role when preventing the
correct accommodation of the orthosteric ligands.
Altogether, the comparison of docking results for the two sets of carbachol dimers suggests
that compounds 3 and 4 are more capable than compounds 1 and 2 to occupy the orthosteric
binding site, a trend which can be ascribed to the different arrangement of the two key
interacting moieties (i.e. the ammonium head and the carbamate function) which only in 3 and
4 parallels the corresponding arrangement of the hM2 interacting residues. In contrast
17

compounds 3 and 4 are less capable than compounds 1 and 2 to occupy the allosteric binding
site and this can be explained by considering the greater distance between the ammonium
heads in the first set, having the carbachol units linked through their carbamate moieties,
compared to the new dimers.
4. Discussion
This paper reports the characterization of a new series of carbachol dimers (compounds 3a-e
and 4a-e) in which the two monomers are connected through the choline nitrogen atoms. The
new compounds share some features with the previously characterized series 1a-f and 2-a-f: a
lack of selectivity toward one of the five muscarinic receptors, and a progressive increase of
affinity by elongating the linker’s chain, observed in equilibrium binding studies (Fig. S1,
Table 1). However, some differences are also evident: in the new series, quaternarization of
the amine moiety did not generally increase affinity, with the exception of compounds with n
= 7, as pKi values for 4c are higher than those of 3c on all subtypes. In addition, while
compounds belonging to the 1 and 2 series show higher affinity at hM3 and hM5 (Matucci et
al., 2016), compounds 3a-e and 4a-e show the lowest apparent affinity on these subtypes.
Docking studies also highlight some differences. In fact, for compounds 3 and 4 one
carbachol unit in the othosteric site is able to elicit the key interactions engaged by the
carbachol ligand alone (i.e ion-pairing with Asp103 and H-bonds) with a correct geometry,
even when the linker is long enough to allow for a bitopic binding.
Kinetic binding studies also show a different behavior between series of compounds. In fact,
in the old series (1, 2) affinity for the allosteric site in NMS-occupied hM receptors increased
smoothly with the linker’s elongation, with the only exception of the compound with a six-
methylene chain (1f, 2f); the effect was more evident for the tertiary amines 1a-e compared to
18

the quaternary ammonium analogs 2a-e. On the contrary, the increase of logKocc values for
amines 3a-e is not constant (Table 2, Fig. S2A): the jump going from n=7 (3d) to n=9 (3e) is
associated to an increase of apparent affinity of about two orders of magnitude on hM1, hM2
and hM4 mAChRs, and about one on the hM3 and hM5 subtypes. A similar trend is observed
for methiodides 4a-e only on hM1 mAChR. This behavior is consistent with a change in the
binding mode of the compounds, allowing a much better fit within the receptor binding
pocket, at the allosteric site, since the orthosteric one is occupied by NMS.
Functional experiments gave contradicting results. On guinea-pig ileum the compounds
behaved as much weaker agonist with respect to carbachol; however, the evoked contractions
were not antagonized by atropine, nor by hexamethonium. This finding suggests that they do
not depend on the cholinergic transmission; however, an activation of muscarinic receptors
not blocked by muscarinic antagonist (QNB) was reported some years ago by Jakubik et al
(Jakubík et al., 1996). Similarly to compounds 1c-e, new compounds 5, 6, 7d and 7e, behaved
as antagonists, although with lower potency.
Applying the ERK1/2 phosphorylation assay, performed at hM1, hM2 and hM3 AChRs in
CHO cells, selected compounds (3d-e, 4d-e, 7d-e) showed only antagonist properties, being
able to reduce the effect elicited by ACh (Fig. 5). On the contrary, at hM2 receptors
methiodides 4b-e stimulated GTPγS binding with a maximum effect of roughly 20% with
respect to the Emax evoked by CCh, demonstrating a partial agonist effect. This suggests that
structural dimerization of the compounds did not completely abolish an agonistic effect, at
least not at the hM2 subtype, although the intrinsic activity was low. However, differently
from carbachol, these compounds were able to initiate hM2 receptor coupling only with Gi,
but not with Gs proteins. Since ERK1/2 phosphorylation is believed to be an example of
arrestin-dependent signaling in GPCR (Rajagopal et al., 2010), these results seem to suggest
for methiodides 4b-e at hM2 receptors a biased activity toward Gi in comparison to the Gs and
19

β-arrestin pathways. Although binding studies did not highlight subtype selectivity, some
compounds showed a different activity on the three muscarinic receptors tested (hM1, hM2
and hM3) suggesting some degree of functional selectivity.
In conclusion, we have prepared a new series of homodimers of the well-known cholinergic
agonist carbachol. The compounds have been analyzed by means of equilibrium and kinetic
binding studies; docking simulations on hM2 receptor supported the view that some
compounds may bind in a bitopic fashion. Functional experiments showed that
homodimerization, connecting two carbachol units through the choline nitrogen atoms, gave
derivatives which maintained some agonistic activity, being able to stimulate the binding of
GTPγS to the Gi protein. The inability to couple with Gs, or to stimulate ERK1/2
phosphorylation suggests some degree of functional pathway selectivity for these molecules.
A more in-depth investigation on the ligand bias potential of these compounds is underway
and will be reported in due time.
5. Acknowledgments
This work was supported by grants from MIUR – Italy (PRIN 2009, 2009ESXPT2_002).
6. Appendix A:
Supplementary Data. Synthetic procedures and the physicochemical properties of the
compounds synthesized in this study.
7. References
20

Barocelli, E., Chiavarini, M., Ballabeni, V., Bordi, F., Impicciatore, M., 1993. Interaction of
selective compounds with muscarinic receptors at dispersed intestinal smooth muscle cells.
Br. J. Pharmacol. 108, 393-397.
Bermudez, M., Bock, A., Krebs, F., Holzgrabe, U., Mohr, K., Lohse, M.J., Wolber, G., 2017.
Ligand-Specific Restriction of Extracellular Conformational Dynamics Constrains Signaling
of the M2 Muscarinic Receptor. ACS Chem. Biol. 12, 1743-1748.
Bock, A., Schrage, R., Mohr, K., 2018. Allosteric modulators targeting CNS muscarinic
receptors. Neuropharmacology 136, 427-437.
Bonifazi, A., Yano, H., Del Bello, F., Farande, A., Quaglia, W., Petrelli, R., Matucci, R.,
Nesi, M., Vistoli, G., Ferré, S., Piergentili, A., 2014. Synthesis and Biological Evaluation of a
Novel Series of Heterobivalent Muscarinic Ligands Based on Xanomeline and 1-[3-(4-
Butylpiperidin-1-yl)propyl]-1,2,3,4-tetrahydroquinolin-2-one (77-LH-28-1). J. Med. Chem.
57, 9065-9077.
De Min, A., Matera, C., Bock, A., Holze, J., Kloeckner, J., Muth, M., Traenkle, C., De Amici,
M., Kenakin, T., Holzgrabe, U., Dallanoce, C., Kostenis, E., Mohr, K., Schrage, R., 2017. A
new molecular mechanism to engineer protean agonism at a G protein-coupled receptor. Mol.
Pharmacol. 91, 348-356.
Gregory, K.J., Hall, N.E., Tobin, A.B., Sexton, P.M., Christopoulos, A., 2010. Identification
of Orthosteric and Allosteric Site Mutations in M2 Muscarinic Acetylcholine Receptors That
Contribute to Ligand-selective Signaling Bias. J. Biol. Chem. 285, 7459-7474.
Haga, K., Kruse, A.C., Asada, H., Yurugi-Kobayashi, T., Shiroishi, M., Zhang, C., Weis,
W.I., Okada, T., Kobilka, B.K., Haga, T., Kobayashi, T., 2012. Structure of the human M2
muscarinic acetylcholine receptor bound to an antagonist. Nature 482, 547.
Jäger D., Schmalenbach C., Prilla S., Schrobang J., Kebig A., Sennwitz M., Heller E., Tränkle
C., Holzgrabe U., Höltje H.D., Mohr K. 2007. Allosteric small molecules unveil a role of an
21

extracellular E2/transmembrane helix 7 junction for G protein-coupled receptor activation. J.
Biol. Chem. 282, 34968-34976.
Jakubík, J., Bacáková, L., Lisá, V., el-Fakahany, E.E., Tucek, S., 1996. Activation of
muscarinic acetylcholine receptors via their allosteric binding sites. PNAS 93, 8705-8709.
Kruse, A.C., Hu, J., Pan, A.C., Arlow, D.H., Rosenbaum, D.M., Rosemond, E., Green, H.F.,
Liu, T., Chae, P.S., Dror, R.O., Shaw, D.E., Weis, W.I., Wess, J., Kobilka, B.K., 2012.
Structure and dynamics of the M3 muscarinic acetylcholine receptor. Nature 482, 552.
Kruse, A.C., Kobilka, B.K., Gautam, D., Sexton, P.M., Christopoulos, A., Wess, J., 2014.
Muscarinic acetylcholine receptors: novel opportunities for drug development. Nat. Rev.
Drug. Discov. 13, 549-560.
Kruse, A.C., Ring, A.M., Manglik, A., Hu, J., Hu, K., Eitel, K., Hubner, H., Pardon, E.,
Valant, C., Sexton, P.M., Christopoulos, A., Felder, C.C., Gmeiner, P., Steyaert, J., Weis,
W.I., Garcia, K.C., Wess, J., Kobilka, B.K., 2013. Activation and allosteric modulation of a
muscarinic acetylcholine receptor. Nature 504, 101-106.
Lazareno, S., Birdsall N.J., 1995. Detection, quantitation, and verification of allosteric
interactions of agents with labeled and unlabeled ligands at G protein-coupled receptors:
interactions of strychnine and acetylcholine at muscarinic receptors. Mol. Pharmacol. 48, 362-
378.
Matucci, R., Nesi, M., Martino, M.V., Bellucci, C., Manetti, D., Ciuti, E., Mazzolari, A., Dei,
S., Guandalini, L., Teodori, E., Vistoli, G., Romanelli, M.N., 2016. Carbachol dimers as
homobivalent modulators of muscarinic receptors. Biochem. Pharmacol. 108, 90-101.
May, L.T., Avlani, V.A., Langmead, C.J., Herdon, H.J., Wood, M.D., Sexton, P.M.,
Christopoulos, A., 2007. Structure-Function Studies of Allosteric Agonism at M2 Muscarinic
Acetylcholine Receptors. Mol. Pharmacol. 72, 463-476.
Mitchelson, F.J., 2012. The pharmacology of McN-A-343. Pharmacol. Ther. 135, 216-245.
22

Pronin, A.N., Wang, Q., Slepak, V.Z., 2017. Teaching an Old Drug New Tricks: Agonism,
Antagonism, and Biased Signaling of Pilocarpine through M3 Muscarinic Acetylcholine
Receptor. Mol. Pharmacol. 92, 601-612.
Rajagopal, S., Rajagopal, K., Lefkowitz, R.J., 2010. Teaching old receptors new tricks:
biasing seven-transmembrane receptors. Nat. Rev. Drug. Discov. 9, 373.
Schrage, R., Kostenis, E., 2017. Functional selectivity and dualsteric/bitopic GPCR targeting.
Curr. Opin. Pharmacol. 32, 85-90.
Thal, D.M., Sun, B., Feng, D., Nawaratne, V., Leach, K., Felder, C.C., Bures, M.G., Evans,
D.A., Weis, W.I., Bachhawat, P., Kobilka, T.S., Sexton, P.M., Kobilka, B.K., Christopoulos,
A., 2016. Crystal structures of the M1 and M4 muscarinic acetylcholine receptors. Nature
531, 335.
Valant, C., Gregory, K.J., Hall, N.E., Scammells, P.J., Lew, M.J., Sexton, P.M.,
Christopoulos, A., 2008. A Novel Mechanism of G Protein-coupled Receptor Functional
Selectivity: Muscarinic partial agonist McN-A-343 as a bitopic orthosteric/allosteric ligand. J.
Biol. Chem. 283, 29312-29321.
Valant, C., Robert Lane, J., Sexton, P.M., Christopoulos, A., 2012. The Best of Both Worlds?
Bitopic Orthosteric/Allosteric Ligands of G Protein–Coupled Receptors. Annu. Rev.
Pharmacol. 52, 153-178.
Vistoli, G., Mazzolari, A., Testa, B., Pedretti, A., 2017. Binding Space Concept: A New
Approach To Enhance the Reliability of Docking Scores and Its Application to Predicting
Butyrylcholinesterase Hydrolytic Activity. J. Chem. Inf. Model. 57, 1691-1702.
Wess, J., Eglen, R.M., Gautam, D., 2007. Muscarinic acetylcholine receptors: mutant mice
provide new insights for drug development. Nat. Rev. Drug Discov. 6, 721-733.
Legends to Fig.s
23

Fig. 1. Structure of the reference molecules and newly synthesized compounds.
Fig. 2. Representative graphs for one-point kinetic assay. Panels A-D show the course of
[3H]NMS dissociation from the hM2 receptor in the presence of increasing concentrations of
the selected ligands 3e (A) and gallamine (B) added in combination with atropine (10 µM) as
a function of time. The incubation time point selected in each case was 20 min, around 2-3
times the koff of [3H]NMS at hM2 receptors. Please notice that the control curve (black) in
panel A lays below the green curve. Panel D: representative curves of residual specific
binding obtained by transposing the respective binding levels from one-point kinetic assays
for 3e and gallamine at [3H]NMS-occupied M2 receptors to illustrate the promoting effect of
allosteric retardation on specific [3H]NMS receptor binding by a four parameter logistic curve
fitting. The data point at log [agent] = -1 represents the [3H]NMS bound in the absence of
added atropine (10 µM) and “allosteric” ligand.
Fig. 3. Determination of ERK1/2 phosphorylation in intact cells stably transfected with (A)
hM1, (B) hM2 and (C) hM3 receptors. Concentrations used were the following: acetylcholine
(10 nM), tested compounds (100 ꢀM) and atropine (10 ꢀM). Data are illustrated as a
percentage of the response mediated by 10% serum and are presented as mean ± S.E.M. of
three to four experiments, each one performed in quadruplicate. Parameters were statistically
evaluated with one-way ANOVA followed by Tukey’s Multiple Comparison Test (vs. ACh as
the control). *** P<0.0001; ** P<0.001; *P<0.05.
Fig. 4. Stimulation of M-receptor mediated [35S]GTPγS binding by compounds 4b-e in CHO-
hM2 membranes, in comparison to ACh and CCh. [35S]GTPγS binding in CHO-hM2
membranes is plotted versus increasing (log)-concentrations of the respective ligands
24

(abscissa). Emax values are expressed as a percentage of the maximal [35S]GTPγS binding
induced by CCh being 100%. Experiments were performed with a final protein concentration
of 40 ꢀg ml-1 and a [35]GTPγS concentration of 0.07 nM. Incubation took place at 30 °C for
one hour. Data are means ± S.E.M. from three to eight independent experiments performed in
quadruplicate or triplicate. Error bars are only visible when exceeding the symbols. Curves
were fitted with a “four parameter logistic function” and the curve slope not different from
unity (F-test, P>0.05). Numerical estimates of selected parameters obtained from the curve
analyses are listed in Table S1.
Fig. 5. Main interactions stabilizing the putative complexes for 3b (A) and 3d (B) within the
active hM2 structure as well as for 3d (C) within the inactive hM2 structure in complex with
QNB (depicted in yellow).
25

Table 1. Parameters describing estimated equilibrium binding affinity of the tested
compounds (see structures shown in fig.1) for human cloned muscarinic receptors expressed
in CHO-K1 cells membranes. Results are expressed as inhibition binding constants, pKi, or as
pIC0.5 (in round brackets), when the compound did not fully displace the radioligand. Values
are reported as means of 3-4 experiments ± S.E.M.
Compound n
X
hM1
hM2
hM3
(4.02±0.16) (<4)
(<4)
(4.46±0.06) (5.07±0.04) (4.05±0.20) (<4)
hM4
hM5
3
3
5
5
7
7
9
9
NMe
(4.30±0.16) (<4)
(<4)
3a
NMe2I (4.37±0.06) (4.69±0.07) (<4)
NMe
NMe2I 4.97±0.06
NMe 5.53±0.03
NMe2I 5.99±0.03
NMe 6.73±0.04
NMe2I 6.28±0.03
7.42±0.06
11 NMe2I 6.95±0.04
(<4)
4a
(4.00±0.11)
3b
5.00±0.06
5.64±0.05
6.13±0.06
6.68±0.06
6.75±0.08
7.45±0.06
6.82±0.06
6.33±0.07
6.00±0.04
6.30±0.03
6.16±0.06
5.92±0.07
4.68±0.07
5.09±0.05
5.41±0.03
5.50±0.03
5.18±0.05
6.45±0.04
6.12±0.03
6.54±0.10
6.25±0.07
5.75±0.09
5.46±0.07
4.36±0.10
5.18±0.07 4.16±0.08
5.48±0.09 4.80±0.04
5.80±0.03 5.11±0.04
6.16±0.04 5.33±0.03
5.74±0.05 5.08±0.04
7.15±0.06 6.32±0.04
6.38±0.04 5.80±0.04
6.66±0.10 6.43±0.13
6.07±0.05 5.82±0.08
6.08±0.07 6.00±0.01
5.90±0.06 5.75±0.01
5.20±0.07 4.16±0.09
4b
3c
4c
3d
4d
11 NMe
3e
4e
-
-
-
-
6.90±0.06
6.26±0.04
6.31±0.01
6.47±0.21
4.42±0.10
5
-
6
9
7d
11 -
7e
-
-
Carbachol

Table 2. Log affinity estimates (log Kocc) of the tested compounds (see structures shown in
fig.1) at the indicated [3H]NMS-occupied muscarinic receptor subtypes calculated as
described in ref.18. Values are reported as means ± S.E.M. of at least three experiments
performed in duplicate.
Compound
n
X
hM1
hM2
hM3
hM4
hM5
3
3
5
5
7
7
9
NMe
NMe2I
NMe
NMe2I
NMe
NMe2I
3.77±0.04 3.27±0.16 3.65±0.04 3.92±0.02 3.51±0.09
3.89±0.13 3.06±0.20 3.79±0.01 3.67±0.26 3.68±0.01
3.59±0.17 3.48±0.20 3.48±0.20 3.72±0.07 3.40±0.15
3.64±0.11 3.42±0.11 3.96±0.01 3.90±0.04 3.85±0.04
3.37±0.52 3.37±0.14 2.96±0.13 3.52±0.22 3.24±0.05
3.57±0.01 4.17±0.01 3.56±0.01 4.20±0.01 3.92±0.01
5.32±0.14 5.79±0.12 4.14±0.23 5.36±0.12 4.16±0.07
3a
4a
3b
4b
3c
4c
NMe
3d
9
NMe2I
NMe
NMe2I
.
-
-
-
-
4.73±0.12 4.65±0.06 4.09±0.08 3.84±0.09 3.70±0.15
4.28±0.18 4.96±0.07 4.15±0.08 4.79±0.06 4.15±0.06
4.91±0.07 5.18±0.05 3.82±0.08 4.67±0.04 4.07±0.10
5.02±0.07 5.30±0.07 3.80±0.10 4.81±0.07 3.72±0.09
3.35±0.30 3.91±0.15 3.93±0.07 4.42±0.12 4.03±0.07
4.98±0.13 4.57±0.08 4.01±0.04 4.36±0.06 3.87±0.12
4.80±0.15 4.51±0.26 4.23±0.06 4.52±0.10 4.26±0.06
4.15±0.03 5.16±0.07 4.20±0.05 4.56±0.08 4.30±0.07
3.57±0.26 3.39±0.15 3.42±0.16 3.64±0.15 3.68±0.15
4d
3e
4e
5
6
7d
7e
11
11
-
-
9
11
-
Gallamine
McN-A-343