Fluorescent derivatives of AC-42 to probe bitopic orthosteric/allosteric binding mechanisms on muscarinic M1 receptors

Article

pubs.acs.org/jmc

Fluorescent Derivatives of AC-42 To Probe Bitopic Orthosteric/

Allosteric Binding Mechanisms on Muscarinic M1 Receptors

Sandrine B. Daval,^†,§Celine Valant,^‡,∥Dominique Bonnet,^‡Esther Kellenberger,^‡Marcel Hibert,^‡

́

Jean-Luc Galzi,^†and Brigitte Ilien*^,†

^†Unite

Strasbourg, BP 10413, 67412 Illkirch, France

^‡Laboratoire d’Innovation Ther

apeutique, UMR 7200 CNRS, Faculte

́

Biotechnologie et Signalisation Cellulaire, UMR 7242 CNRS, Ecole Super

́

ieure de Biotechnologie de Strasbourg, Universite

́

de

́

de Pharmacie, Universite de Strasbourg, 74 Route du Rhin, BP

́

24, 67401 Illkirch, France

S

* Supporting Information

ABSTRACT: Two fluorescent derivatives of the M1 muscarinic

selective agonist AC-42 were synthesized by coupling the

lissamine rhodamine B fluorophore (in ortho and para

positions) to AC42-NH₂. This precursor, prepared according

to an original seven-step procedure, was included in the study

together with the LRB fluorophore (alone or linked to an alkyl

chain). All these compounds are antagonists, but examination of

their ability to inhibit or modulate orthosteric [³H]NMS binding

revealed that para-LRB-AC42 shared several properties with AC-42. Carefully designed experiments allowed para-LRB-AC42 to

be used as a FRET tracer on EGFP-fused M1 receptors. Under equilibrium binding conditions, orthosteric ligands, AC-42, and

the allosteric modulator gallamine behaved as competitors of para-LRB-AC42 binding whereas other allosteric compounds such

as WIN 51,708 and N-desmethylclozapine were noncompetitive inhibitors. Finally, molecular modeling studies focused on

putative orthosteric/allosteric bitopic poses for AC-42 and para-LRB-AC42 in a 3D model of the human M1 receptor.

The properties of the M1-preferring agonist N-desmethyl-

clozapine⁹and of a growing series of novel agonists with

unprecedented high functional selectivity at human M1

receptors, such as AC-42^10,11and its AC-260584¹²or 77-LH-

28−1¹³analogues, TBPB,¹⁴and the more recent LuAE51090¹⁵

and VU0357017¹⁶compounds (Figure 1), raise the exciting

possibility that they may provide useful therapeutics for

cognitive impairment.^14,17,18They exhibit functional selectivity

toward subsets of signaling pathways¹⁹and appear to stabilize

receptor conformations less prone to downstream regulatory

programs than orthosteric agonists.^20,21

These novel compounds are commonly referred to as

allosteric agonists on the basis of their insensitivity to the

mutation of the critical Tyr381 (residue 6.51, using the

Ballesteros and Weinstein convention²²) orthosteric residue

and of their ability to decrease the dissociation rate of the

orthosteric antagonist [³H]N-methylscopolamine. Note that N-

desmethylclozapine only fulfills the first criterion for an

allosteric interaction at the M1 subtype. The competitive^13,23,24

or noncompetitive^11,12,14,15character of the novel agonists,

when tested against orthosteric ligands in equilibrium binding

or functional interaction studies, is still a matter of debate, as

true competition is often difficult to distinguish from very

strong negative binding cooperativity.⁸Most of them bind to all

muscarinic receptors with similar affinity, whereas they display

INTRODUCTION

■

Despite a well-known implication of muscarinic acetylcholine

receptors (mAChRs) in a number of physiopathological

processes,^1,2fine-tuning of M1−M5 receptor subtypes activity

through the use of highly selective ligands remained an almost

insurmountable challenge until the development of novel

compounds able to modulate or stimulate these receptors

through allosteric binding sites.³⁻⁵Topographically distinct

from the conserved orthosteric binding domain where the

endogenous agonist acetylcholine and most classical ligands

bind in a competitive manner, allosteric sites are probably

numerous and diverse (though yet poorly resolved) across the

five receptor subtypes.^4,5In addition to offering the potential

for selective receptor targeting, allosteric ligands may stabilize

different receptor conformational and functional states (relative

to orthosteric ligands) and therefore instigate a distinct

repertoire of receptor signaling and regulatory properties.^6,7

Allosteric modulators, without detectable effect on their own,

reveal their impact on receptor properties when binding

concomitantly with an orthosteric ligand. The formation of

such ternary complexes is accompanied by conformational

rearrangements that result in alterations in ligand affinity,

potency, efficacy, subtype, and/or functional selectivity.^6,8

A

range of muscarinic modulators has been described,⁴with their

therapeutic potential best illustrated by novel selective allosteric

enhancers of acetycholine binding (and function) at M1 or M4

receptors.^3,5

Received: October 8, 2011

Published: February 14, 2012

2125

dx.doi.org/10.1021/jm201348t | J. Med. Chem. 2012, 55, 2125−2143

Journal of Medicinal Chemistry

Article

Figure 1. Chemical structures of novel muscarinic agonists and of well-known allosteric modulators. Selected compounds are cited in this paper.

Some of them (AC-42, NDMC, gallamine, brucine and Win 51,708) were also used as pharmacological tools in several experiments presented in this

paper.

weak efficacies at M2−M5 subtypes, suggesting that activation

mechanisms may vary among subtypes.²⁵Mutagenesis mapping

studies on the M1 receptor, combined with binding and

activation profiling of various muscarinic agonists, provide

evidence for the existence of multiple activation switches, some

of which can be selectively exploited by novel agonists whereas

others are used by all agonist classes.^12,23,26

Altogether, these observations tend to indicate that the novel

agonists may explore the well-conserved orthosteric binding

pocket and gain functional M1 selectivity from additional

interactions with less conserved residues and/or from the

ability to elicit subtype-specific conformational rearrange-

ments.^23,24It has been recently suggested that these agonists

might be bitopic in that they may recognize epitopes within the

orthosteric site and part of an allosteric site.²⁴Yet ligand

docking into M1 receptor homology models provides a

confounding picture of the domains explored by these

compounds: AC-42 and TBPB share a binding pocket distinct

from the orthosteric site,²⁶AC-42 and 77-LH-28-1 better adopt

a bitopic binding mode in extending down into the orthosteric

site as well as up toward a less buried allosteric site.²⁴Whether

these compounds explore a subsidiary binding pocket over-

lapping with²⁴or distinct from²³the “common” allosteric site

used by modulators such as gallamine remains unclear.

fluorescence resonance energy transfer (FRET) based binding

studies.^27,28Functional experiments were first undertaken using

HEK cells stably expressing a series of EGFP-fused human M1

receptors, with progressive deletion of the receptor N-terminus,

in order to examine the role of this domain as a possible

determinant of AC-42 agonism and to define the pharmaco-

logical nature and potency of the various compounds.

Equilibrium and kinetic [³H]NMS binding studies were

conducted on living cells expressing the EGFP(Δ17)hM1

receptor to unravel the orthosteric or allosteric nature of the

fluorescent derivatives. Using the same receptor chimera, we

carefully designed the experimental conditions for FRET

measurements, evaluated the potential of the LRB derivatives

as FRET probes, focusing on the binding parameters of the

para-LRB-AC42 isomer under equilibrium and kinetic con-

ditions and on the ability of several prototypical compounds to

modulate its affinity properties. Finally, molecular modeling

studies were undertaken to get structural insights on putative

binding domains for AC-42 and para-LRB-AC42 in the M1

receptor. Altogether, we describe a fluorescent bitopic tracer

that shares many properties with its parent AC-42 compound

and overlaps both the orthosteric site and the allosteric

gallamine site. Such a combination of [³H]NMS and FRET-

based binding assays proves useful to study bitopic ligands and

may help to discover new muscarinic allosteric agents.

In this work, we report on the synthesis and pharmacological

characterization of two lissamine rhodamine B (LRB)

derivatives (ortho and para isomers) of AC-42 to address

some of these still open questions. We aimed at obtaining a

fluorescent agonist, an allosteric tracer, or a bitopic ligand

amenable to be investigated through radioligand based and

RESULTS

■

Two lissamine rhodamine B derivatives of AC-42 (ortho-LRB-

AC42 and para-LRB-AC42) were synthesized in order to

examine their binding properties at M1 muscarinic receptors

2126

dx.doi.org/10.1021/jm201348t | J. Med. Chem. 2012, 55, 2125−2143

Journal of Medicinal Chemistry

Article

Figure 2. Chemical structures of AC-42 and lissamine rhodamine B derivatives.

Table 1. Spectroscopic Parameters of Fluorophores and of Their Conjugates

absorption

fluorescence

a

MeOH

buffer

MeOH

buffer

donor/acceptor pair

spectral overlap

Forster radius

̈

molecule

ε × 10⁻³(M⁻¹·cm⁻¹

)

λ_max(nm) ε × 10⁻³(M⁻¹·cm⁻¹

)

λ_max(nm) λ_max(nm) λ_max(nm) (cm³·M⁻¹) × 10¹³

(Å)

b

LRB-SO₂Cl

88

568

556

583

LRB-SO₃Na⁺

564

569

574

579

582

587

−

para-LRB-n-butyl

para-LRB-AC42

ortho-LRB-AC42

125

52

561

563

88

1.8 0.1 (n = 3)

nd

51.2 0.4 (n = 3)

nd

88

(33−42)

a

Spectral overlap (J) and Forster radius (R ) values for each donor−acceptor pair are from measurements in Hepes buffer and are determined as

̈

0

reported in Supporting Information Procedures. Values are the mean

SE for three independent determinations. nd, not determined.

b

Spectroscopic parameters for LRB sulfonyl chloride.⁶²

a

Scheme 1. Synthesis of Para- and Ortho-LRB-AC42 Derivatives 2a,b

a

Reagents: (i) (a) lissamine rhodamine B sulfonyl chloride (LRB-SO₂Cl), NEt₃, CHCl₃, rt, overnight; (b) purification by RP-HPLC (solvent A,

H₂O/TFA 0.1%; solvent B, CH₃CN/TFA 0.1%).

and their potential as allosteric tracers under FRET conditions

(Figure 2). The lissamine rhodamine B fluorophore (LRB-

NH₂synthetic precursor were also included in this study to

evaluate the contribution of individual structural moieties to

overall properties of the fluorescent AC-42 derivatives. The

lissamine rhodamine B dye was selected for its excellent

SO₃Na⁺), its close derivative with an aliphatic butyl tail

−

mirroring that of AC-42 (para-LRB-n-butyl), and the AC42-

2127

dx.doi.org/10.1021/jm201348t | J. Med. Chem. 2012, 55, 2125−2143

Journal of Medicinal Chemistry

Article

a

chemical and photochemical stability and for appropriate

spectroscopic properties as an energy acceptor from excited

EGFP (Supporting Information Figure S1, Table 1).

Scheme 3. Compound 8 Preparation

Chemistry. Previous structure−activity relationship studies

have shown that subtle modifications of the 2-methylaceto-

phenone moiety of AC-42 led to variability in agonist potency

and receptor subtype selectivity.^12,13,29We therefore decided to

introduce the fluorophore into AC-42 by modifying its more

tolerant n-butyl side chain. This was done by reacting lissamine

rhodamine B sulfonyl chloride with AC42-NH₂1 (Scheme 1).

All our attempts to access to AC42-NH₂1 following

protocols previously reported for AC-42 synthesis (U.S. Patent

US 6,528,529 B1) failed. An original seven-step process was

thus designed, starting from the commercially available o-

tolylmagnesium bromide (Scheme 2). Condensation of the

latter with succinic anhydride led to the acid 3³⁰which was

subsequently converted into methyl ester 4. Protection of the

a

Reagents: (i) BzOCOCl, NaOH 10%, 0 °C to rt, 2 h, rt; (ii)

C₆H₅N(SO₂CF₃)₂, THF, LDA, −78 to 0 °C, 3 h; (iii) PdCl₂(PPh₃)₂,

CuI, NEt₃, THF, rt, 6 h; (iv) H₂, PtO₂, EtOH, AcOH cat., rt, 65 psi,

overnight.

the presence of oxalyl chloride. A subsequent alkylation with t-

BuOH led to the N-Boc protected compound 17 which was

a

Scheme 4. Compound 13 Preparation

a

Scheme 2. Synthesis of AC42-NH₂1

a

Reagents: (i) ClCOCOCl, CH₂Cl₂, reflux, overnight; (ii) t-BuOH,

toluene, 40 °C, overnight; (iii) DEAD, PPh₃, THF, rt, overnight; (iv)

LiOH, H₂O/THF (1:3), rt, 3 h.

submitted to a Mitsunobu reaction with the commercially

available pent-4-yn-1-ol 18 to provide alkyne 19. The final basic

hydrolysis of the ethyl ester enabled removal of the ethyl

oxalate group to obtain the N-Boc protected alkyne 13 in a 52%

overall yield.

a

Reagents: (i) succinic anhydride, THF, −78 to 0 °C, then 2 h, rt; (ii)

H₂SO₄, MeOH, reflux, overnight; (iii) ethylene glycol, p-TsOH,

benzene, 2 days; (iv) LiAlH₄, ether, 0 °C, 1 h; (v) MeSO₂Cl, NEt₃, 2

h, rt; (vi) NEt₃, CH₃CN, sealed tube, 110 °C ; (vii) 1 N HCl/Et₂O.

The incorporation of lissamine rhodamine B (mixed

isomers) onto the alkyl chain of AC42-NH₂1 was performed

using commercial lissamine rhodamine B sulfonyl chloride

(LRB-SO₂Cl), in the presence of Et₃N in CH₂Cl₂. The reaction

yielded two fluorescent isomers, para-LRB-AC42 2a and ortho-

LRB-AC42 2b (Scheme 1), which were isolated by preparative

reverse-phase high-performance liquid chromatography (RP-

HPLC). The two regioisomers were unambiguously charac-

terized by ¹H NMR. Indeed, in agreement with previous

reports on the identification of LRB isomers,^36,37the H5

proton of the lissamine moiety (Scheme 1) in ortho-LRB-AC42

is downfield the corresponding one in para-LRB-AC42 (8.02

ppm versus 7.92 ppm, respectively). The synthesis of para-

LRB-n-butyl 20 followed the same procedure, starting from the

n-butylamine (Scheme 5). The para isomer was isolated by RP-

HPLC as the major product of the reaction. Purity and identity

of the compounds were checked by analytical RP-HPLC, liquid

chromatography−mass spectrometry (LC−MS), and high-

resolution mass spectrometry (HRMS).

ketone group was performed in the presence of ethylene glycol

in acidic catalysis to form acetonide 5.³¹The methyl ester was

then reduced into alcohol 6 in the presence of LiAlH₄in dry

ether. Conversion of the alcohol into methyl sulfonate 7 was

performed in the presence of methanesulfonyl chloride. The

nucleophilic displacement of mesylate moiety by the piperazine

derivative 8 gave access to compound 9. A subsequent

treatment with HCl enabled the simultaneous deprotection of

the ketone and the amino groups, leading to AC42-NH₂1

obtained as a chlorhydrate salt in a 5% overall yield.

The synthesis of piperazine 8 involved a four-step process

starting from commercially available 4-piperidone 10 (Scheme

3). Protection of the amino group with a benzyloxycarbonyl

moiety was followed by the formation of triflate 12 in the

presence of N-phenyltrifluoromethanesulfonimide.³²A Sonoga-

shira coupling reaction with alkyne 13 gave access to

compound 14^33,34which was treated in the presence of H₂/

PtO₂to perform simultaneously the hydrogenolysis of the Cbz

protecting group and the hydrogenation of the alkyne moiety.

Following this approach, piperazine 8 was obtained in a 44%

overall yield.

Altogether, the synthetic route we developed for AC42-NH₂

led to the exchange of the AC-42 butyl chain for an

aminopentyl one and subsequently to the introduction of a

more flexible five-carbon atom linker between the AC-42

piperidine ring and the sulfonamide group of the fluorescent

AC-42 derivatives.

N-Boc protected pent-4-yne-1-amino 13 was obtained

following an original four-step process, starting from

commercially available ethyl oxamate 15 (Scheme 4).³⁵First,

the conversion of the latter into isocyanate 16 was achieved in

Spectroscopic Characterization and Solubility Meas-

urements. Spectroscopic properties of the fluorescent

derivatives were examined in MeOH and Hepes buffer

(Table 1). Their absorbance and fluorescence spectra are

2128

dx.doi.org/10.1021/jm201348t | J. Med. Chem. 2012, 55, 2125−2143

Journal of Medicinal Chemistry

Article

a

which may both act as ligand solubilizers. Thus, the solubilities

listed above refer to bottom estimates for compound solubility.

Integrity of the M1 Receptor N-Terminus and AC-42

Agonism. We have already shown that shortening the human

M1 muscarinic receptor N-tail, and its subsequent fusion to

EGFP, does not alter receptor pharmacology or functionality

but markedly improves energy transfer efficacy arising from

binding of a fluorescent antagonist.²⁷Because previous

mutagenesis studies proposed indirect evidence for AC-42 to

bind to an ectopic site, involving the M1 receptor N-

terminus,¹⁰we wondered whether such fluorescent receptors

would be appropriate for studying the properties of AC-42

derivatives. To this end, we selected constructs with EGFP

fused to a full length receptor (EGFP-wthM1) or to receptors

truncated by 11 (EGFP(Δ11)hM1) or 17 (EGFP(Δ17)hM1)

residues at their N-terminus. The three fluorescent chimera,

together with the wild-type hM1 receptor taken as a control,

were stably expressed in HEK cells and tested for global

([³H]QNB binding) or cell surface ([³H]NMS binding)

receptor expression and for agonist-evoked calcium responses.

AC-42 was found to stimulate all these receptors with

comparable EC₅₀values (Table 2). Although subtle differences

may rely on variations in receptor density, potencies are within

the range of values reported for AC-42 (0.15−1.8 μM,

depending on cells lines and functional tests) at the wild-type

M1 receptor.^{10,12,13,15,19,23,26}In line with these studies, AC-42

exhibits partial agonism at all receptor constructs (compared to

carbachol) and shows no detectable activity (even at 10 μM) in

nontransfected HEK cells with endogenous expression of the

M3 subtype.

Thus, our data markedly contrast with previous conclusions

based on the observation of altered AC-42 functionality

following replacement of the first 45 amino acids of the M1

receptor with an equivalent M5 sequence.¹⁰Note that a major

part of the M1 first transmembrane domain was exchanged in

this M1/M5 chimera. As neither the fusion of EGFP nor the

deletion of the M1 receptor N-terminus was detrimental to AC-

42 agonist properties, the EGFP(Δ17)hM1 construct was

selected for all subsequent experiments.

Functional Properties of AC-42 Derivatives. We next

examined the functional impact of the introduction of an

aminomethyl group onto the AC-42 butyl chain (such as in

AC42-NH₂) and of its derivation by a bulky fluorophore (such

as in para- and ortho-LRB-AC42). These compounds are no

longer capable (when tested at 20 or at 37 °C) of promoting a

calcium response in EGFP-wthM1 and EGFP(Δ17)hM1

expressing cells or in nontransfected HEK cells. Instead, para-

Scheme 5. Para-LRB-n-butyl 20 Preparation

a

Reagents: n-butylamine, Et₃N, CH₂Cl₂, rt, overnight.

moderately sensitive to the environment or to chemical

derivatization. In aqueous buffer, the ortho and para LRB

conjugates display superimposable spectra but their molecular

extinction coefficients are lower than in MeOH. The three LRB

derivatives are endowed with appropriate spectroscopic proper-

ties for an energy acceptor from excited EGFP (Supporting

Information Figure S1): they display negligible absorbance at

EGFP’s excitation wavelength (set at 470 nm), negligible

emission at EGFP’s emission wavelength (set at 510 nm),

appreciable overlap of their absorbance spectra with EGFP’s

emission spectrum, and large molecular extinction coefficients.

The last properties dictate a substantial spectral overlap integral

with EGFP and a great Forster radius (R ) for the para-LRB-n-

̈

0

butyl and para-LRB-AC42 compounds (Supporting Informa-

tion Procedures). As expected for molecules bearing the same

fluorophore, the para-LRB-derivatives share overall spectro-

scopic and acceptor properties with RR(17)PZ.³⁸This FRET

tracer for M1 receptors was obtained through coupling of the

succinimidyl ester of the Rhodamine Red-X fluorophore to

functionalized pirenzepine, an M1 selective antagonist.

Ortho-LRB-AC42 was found to photofade in Hepes buffer to

variable extent depending on the dilution factor of DMSO

stock solutions. This precluded robust determination of its

molecular extinction coefficient and R₀(Table 1, values in

parentheses) and prompted us to evaluate the solubility limits

of all fluorescent derivatives in aqueous buffer through

quantitative HPLC analysis (Supporting Information Proce-

dures). AC-42 and AC42-NH₂are perfectly soluble at 100 μM,

as is para-LRB-AC42 at concentrations up to 30 μM. In

contrast, para-LRB-n-butyl and surprisingly ortho-LRB-AC42

are much less soluble in Hepes buffer and begin to precipitate

at concentrations beyond 2 μM. One should emphasize that

solubility measurements were carried out without cells or BSA

Table 2. Functional Parameters for AC-42 in HEK Cells with Stable Expression of Wild Type or EGFP-Fused hM1 Receptors

b

c

receptor expression (fmol/10⁶cells)

calcium response

EC₅₀(μM)

a

receptor construct

N tail length (residues)

[³H]QNB [³H]NMS

E_max(%)

none (HEK cells)

wthM1

7

2

5

2

NR

21

20

9

262 25

940 50

585 65

710 44

195 37

nd

0.74 0.03

0.21 0.08

0.72 0.16

0.45 0.08

72

58

57

67

4

3

4

7

EGFP-wthM1

EGFP(Δ11)hM1

EGFP(Δ17)hM1

nd

3

690 40

a

b

The length of the N terminus (number of residues) of the wild-type hM1 receptor and of the various constructs is indicated. Receptor expression

levels (mean SE values for three independent determinations) are from [³H]QNB and [³H]NMS binding assays conducted on intact cells at

c

saturating radioligand concentrations. n.d., not determined. Measurements of AC-42 evoked calcium responses allowed the determination (mean

SD values from two separate determinations) of EC₅₀and E_max(normalized relative to the maximum response of cells to 10 μM carbachol)

parameters. NR: no response.

2129

dx.doi.org/10.1021/jm201348t | J. Med. Chem. 2012, 55, 2125−2143

Journal of Medicinal Chemistry

Article

LRB-AC42 and AC42-NH₂fully abolish (Figure 3), in a dose-

dependent manner (IC₅₀values of 0.3 0.1 and 3.7 1.4 μM,

proceeded in physiological Hepes buffer at 20 °C in order to

minimize receptor and ligand internalization and for 22 h to

ensure equilibrium in the presence of allosteric ligands.⁴⁰The

apparent equilibrium dissociation constant (95 7 pM, mean

SE; n = 8) for specific [³H]NMS binding to EGFP(Δ17)-

hM1 cells coincided well with K_dvalues reported for this

radiotracer on M1-enriched membrane preparations.⁴⁰⁻⁴²

Displacement experiments (Figure 4) were usually performed

at two radioligand concentrations (about 1K_dand 10K_d) in

order to better distinguish competitive from negatively

cooperative modes of interaction.^4,6Slope factors for all

inhibition curves did not statistically differ from 1. Binding

parameters are listed in Table 3.

As shown in Figure 4A, atropine fully displaces specific

[³H]NMS binding at low and high tracer concentrations, as

expected for competition at the orthosteric receptor site. N-

desmethylclozapine exhibits a similar competitive-like behavior

and a K_iin full agreement with published data.^9,12,23

The allosteric modulator gallamine displays a typical

submaximal inhibition (with plateau levels depending on

[³H]NMS concentration) consistent with a saturable and

negatively cooperative mode of interaction. Analysis according

to the allosteric ternary complex model⁸yielded estimates of

gallamine equilibrium dissociation constant K_xfor the classical

allosteric site on free M1 receptors and of the cooperativity

factor α as a measure of its impact on [³H]NMS binding

affinity (Table 3). Both parameters are in good agreement with

previous reports on M1 membrane preparations,^11,42,43as are

those determined here for brucine (Table 3), another allosteric

modulator.^44,45

Figure 3. Inhibition by AC-42 derivatives of carbachol-induced

calcium responses in EGFP(Δ17)hM1 cells. Response amplitudes are

expressed as percentages of a control response to 0.5 μM carbachol

□

●

and are plotted as a function of para-LRB-AC42 ( ), AC42-NH₂( )

◆

and ortho-LRB-AC42 ( ) concentration. This typical experiment was

performed at 20 °C as reported in the Experimental Section.

respectively; n = 2), the robust M1 response elicited by 0.5 μM

carbachol (EC₅₀of 0.25 μM) in EGFP(Δ17)hM1 cells.

Blockade by ortho-LRB-AC42 is partial, with a 20% maximal

inhibition at 1 μM that progressively vanishes at higher doses.

Such a behavior is most probably related to the poor solubility

of this derivative. The LRB-SO₃Na⁺fluorophore was inactive

−

at 1 mM (not shown).

The calcium response elicited by 0.5 μM carbachol in

EGFP(Δ17)hM1 cells is essentially M1 in nature, as activation

of endogenous M3 receptors in nontransfected HEK cells

requires much higher carbachol (EC₅₀close to 15 μM)

concentrations.³⁹As neither para-LRB-AC42 nor AC42-NH₂is

able to inhibit (even at 10 μM) the strong M3 calcium response

elicited by 20 μM carbachol in HEK cells, one may conclude

that these AC-42 derivatives still exhibit a M1 over M3 selective

functional antagonism.

AC-42 displays an apparent competitive mode of interaction

(Figure 4B) and an affinity constant (Table 3) in close

agreement with previous reports.^20,23,24Interestingly, such

properties are shared by AC42-NH₂. The introduction of a

−

butyl chain onto the almost inactive LRB-SO₃Na⁺fluorophore

confers to para-LRB-n-butyl the ability to inhibit, though

partially, [³H]NMS binding in the micromolar concentration

range. Similar features, mirroring negative binding cooperativity

with [³H]NMS, apply to ortho-LRB-AC42 (Figure 4C).

Unfortunately, given their poor solubility in physiological

Hepes buffer, it was impossible to define the exact nature of

their interaction with M1 receptors or to provide reasonable

estimates of their binding parameters. Most importantly,

however, as shown in Figure 4C, the para-LRB-AC42 derivative

[³H]NMS Inhibition Binding Studies at EGFP(Δ17)hM1

Receptors. The properties of AC42-NH₂and of the LRB

family of compounds were further examined through

equilibrium binding experiments, using [³H]NMS as the

orthosteric site probe. Incubation of EGFP(Δ17)hM1 cells

Figure 4. Inhibition of [³H]NMS binding to EGFP(Δ17)hM1 receptors by orthosteric and allosteric compounds, AC-42, and lissamine rhodamine

B derivatives. Competition-type experiments were performed at low (80−110 pM; dashed lines) and high (800−950 pM; solid lines) [³H]-NMS

concentrations in the absence (B₀value) or the presence (B values) of unlabelled (open symbols) or fluorescent (closed symbols) drugs. A: Atropine

−

+

◇

□

■

○

△

▽

▼

▲

●

, NDMC , gallamine ; B: AC-42 , AC42-NH₂, para-LRB-nbutyl , LRB-SO₃Na

; C: para-LRB-AC42 , ortho-LRB-AC42 . Data

are presented as B/B₀ratios (mean SE values for three to eight separate experiments). Error bars are shown when they exceed the symbols. Curve

fitting was carried out using the standard Hill equation with the exception of gallamine data which were analysed according to the allosteric ternary

complex model (Supplemental Procedures). Corresponding parameters are listed in Table 3.

2130

dx.doi.org/10.1021/jm201348t | J. Med. Chem. 2012, 55, 2125−2143

Journal of Medicinal Chemistry

Article

Table 3. Binding Parameters for a Series of Compounds at EGFP(Δ17)hM1 Receptors As Afforded from Equilibrium and Off-

a

Rate [³H]NMS Binding Assays

b

c

equilibrium assay

off-rate assay

compd

pK_i

pK_x

log α

pEC_50diss

E_max,diss

d

NMS

10.02 0.04

9.24 0.04

atropine

gallamine

5.30 0.04

4.59 0.01

1.10 0.05

0.10 0.02

3.76 0.02

100

5

brucine

nd

N-desmethylclozapine

AC-42

7.22 0.05

5.76 0.07

5.71 0.06

6.00 0.08

nd

AC42-NH₂

para-LRB-AC42

nd

4.6 0.1

79

5

e

(4.4 0.1)

nd

(100)

LRB-SO₃⁻Na⁺

<3

a

Data are the mean SE values for three to nine independent experiments performed in duplicate as described in Experimental Section. Data

b

analyses are detailed in Supporting Information Procedures. Equilibrium dissociation constants (−log M) for the binding of competitive-like

(orthosteric, K_ivalues) and noncompetitive (allosteric, K_xvalues) compounds to free receptors are indicated together with the cooperativity factor

c

(α) governing the allosteric interaction between a modulator and [³H]NMS. pEC_50dissrefers to the drug concentration (−log M) that induces half-

maximal reduction of control [³H]NMS dissociation rate. E_max,dissdefines the maximal extent (%) of inhibition of [³H]NMS dissociation rate. nd: not

d

e

determined. pK_dvalue from [³H]NMS saturation experiments. Parameters are from Figure 5B with constrained curve fitting.

(which is not hampered by such solubility limitations) fully

displaces specific [³H]NMS binding from the EGFP(Δ17)hM1

receptor at low and high tracer concentrations.

other dissociation traces fit a single exponential time course.

The para-LRB-AC42 (k_off= 0.013

0.001 min⁻¹) and

0.001 min⁻¹) compounds

obs

= 0.028

AC42-NH₂(k_off

obs

Special attention was then paid to the mode of interaction of

AC-42, AC42-NH₂, para-LRB-AC42, and NDMC, which all

inhibited [³H]NMS binding down to nonspecific levels. All our

attempts to distinguish strict competition from strong negative

cooperativity, such as the comparison of inhibition levels at low

and high tracer concentrations, the use of individual versus

global curve fitting to both models, or the statistical

examination of model relevance, unfortunately failed. For

example, the increase in [³H]NMS concentration (Figure 4C)

resulted in a rightward shift of the IC₅₀value for para-LRB-

AC42 that is about 2-fold lower than that predicted for a

competitive interplay. Such an observation might be interpreted

as the manifestation of a ceiling effect on inhibition magnitude,

reflecting negative binding cooperativity between [³H]NMS

and the fluorescent compound.^6,8In turn, the competitive

model was preferred over the allosteric one when submitting

para-RR-AC42 curve pairs to a global fitting procedure.

Moreover, while some equilibrium binding studies pointed to

an allosteric behavior of AC-42,^11,12recent reports indicate the

competitive model as the most reasonable one to interpret AC-

42^23,24and NDMC^12,23equilibrium binding data. Accordingly,

we assumed AC-42, AC42-NH₂, para-LRB-AC42, and NDMC

interactions with the EGFP(Δ17)hM1 receptor to be

competitive-like and provide their K_ivalues in Table 3.

significantly slow [³H]NMS dissociation, while ortho-LRB-

AC42 (k_{off obs}= 0.041 0.002 min⁻¹) and para-LRB-n-butyl

(not shown) effects are not significant (mean SD values, n =

2).

Figure 5B depicts the concentration dependence of the

retardation effects of the allosteric modulator gallamine and of

para-LRB-AC42 on [³H]NMS dissociation from EGFP(Δ17)-

hM1 receptors. In agreement with previous reports,^40,42

gallamine completely blocks [³H]NMS dissociation from the

receptor (E_max= 100%), with an EC_{50 diss}close to 170 μM

reflecting its affinity for [³H]NMS-occupied receptors (Table

3). Para-LRB-AC42 data also satisfactorily fit an inhibition

curve with a slope close to unity, but reduction of [³H]NMS

dissociation rate is limited (E_max= 80%) with half-maximal

retardation at 25 μM (Table 3). To our knowledge, such dose-

dependency studies have not been reported yet for AC-42

probably because it combines a modest effect on [³H]NMS

dissociation rate and a limited solubility at high concen-

trations.^12,20,24Taking this factor as a valid reason for the

inability of para-LRB-AC42 to fully prevent [³H]NMS

dissociation at concentrations above 100 μM, data were

reanalyzed (Figure 5B, dashed line), focusing on a 1−50 μM

window with E_maxand slope values constrained to 100% and

1%, respectively. The nominal 100 μM para-LRB-AC42

concentration was found to correspond to 50 μM freely

diffusing molecules in Hepes buffer (HPLC solubility measure-

[³H]NMS Dissociation Kinetic Studies. Examination of

[³H]NMS dissociation kinetics is regarded as a stringent test to

detect an allosteric modulation of mAChRs.^40,46Indeed, the

binding of allosteric ligands causes conformational changes in

the receptor, most often manifested as an alteration in off-rate

kinetics of preformed orthosteric ligand−receptor complexes.

Such studies were thus undertaken to get further information

on the binding nature of the AC-42 derivatives.

Figure 5A allows comparison of the ability of several

compounds (taken at 100 μM) to alter the [³H]NMS

dissociation rate from EGFP(Δ17)hM1 receptors. Atropine

(10 μM) was systematically included in the assays to prevent

[³H]NMS reassociation to the orthosteric site. The control

(atropine alone, k_{off control}= 0.044 0.001 min⁻¹; n = 4) and all

ments). A slightly higher EC₅₀

determined.

(40 μM) was then

diss

Altogether and in contrast with equilibrium studies, the

kinetic data support the idea that AC42-NH₂and para-LRB-

AC42 bind to an allosteric site on [³H]NMS-occupied M1

receptors, as do AC-42^11,12,24and 77-LH-28-1²⁴molecules.

Toward FRET Recordings of Para-LRB-AC42 Binding

to EGFP(Δ17)hM1 Receptors. Figure 6A illustrates a FRET

experiment performed at 20 °C on HEK cells expressing the

EGFP(Δ17)hM1 receptor, using para-LRB-AC42 as the

fluorescent tracer, with real-time recordings of EGFP

fluorescence emission at 510 nm. Upon ligand addition (a, 5

2131

dx.doi.org/10.1021/jm201348t | J. Med. Chem. 2012, 55, 2125−2143

Journal of Medicinal Chemistry

Article

extinction is minored by a para-LRB-AC42 emission

component. Although para-LRB-AC42 displays (Supporting

Information Figure S1) negligible absorbance and emission at

EGFP’s excitation (470 nm) and emission (510 nm)

wavelengths, respectively, direct excitation at 470 nm of

micromolar amounts of tracer leads to significant emission at

510 nm. After correction of this interference, which is

proportional to the para-LRB-AC42 concentration, fluores-

cence extinction and fluorescence recovery now define an

identical FRET signal amplitude.

In summary, specific para-LRB-AC42 binding to the

EGFP(Δ17)hM1 receptor was defined, under equilibrium

binding conditions, as the difference in fluorescence intensity

of cells incubated with the tracer, in the presence or the absence

of a saturating concentration of atropine, or under kinetic

conditions, as the amplitude of atropine-promoted fluorescence

recovery. Such FRET signals are not detected on HEK cells

stably expressing non-M1 EGFP-fused GPCRs or with pure

soluble EGFP, thereby confirming their muscarinic receptor

specificity.

Despite the problems listed above, AC-42 and AC42-NH₂

are pivotal compounds that need to be included in the present

FRET study to help characterize fluorescent tracer properties.

The procedure we developed to afford a safe determination of

their binding parameters is illustrated with AC42-NH₂(Figure

6B). First, the amplitude of the artifact resulting from its

addition to EGFP(Δ17)hM1 cells (or to HEK cells) was found

to be proportional to its concentration (Figure 6B inset).

Second, competition-type studies were performed using para-

LRB-AC42 (5 μM) and increasing concentrations of the AC-42

analogue. They are presented as occupancy curves, with 0 and

100 referring to fluorescence levels of EGFP(Δ17)hM1 cells

incubated with the tracer, in the absence or presence of 20 μM

atropine, respectively. As shown in Figure 6B, fluorescence

recovery versus AC42-NH₂concentration appears as a

nonsaturable process and clearly exceeds the atropine-defined

specific FRET signal amplitude. Fitting these raw data

according to the Hill equation superimposed with a linear

drift accounting for AC42-NH₂artifact allowed extraction of a

typical occupancy curve with a slope close to 1, half-maximal

Figure 5. Effect of AC-42 derivatives on [³H]NMS dissociation rate.

Measurement of [³H]-NMS dissociation rate from EGFP(Δ17)hM1

receptors was carried out at 25 °C as described in the Experimental

Section. A: Dissociation started with the addition of 10 μM atropine

□

○

alone ( ) or combined with 100 μM of either AC42-NH₂( ), ortho-

3

◆

●

LRB-AC42 ( ) or para-LRB-AC42 ( ). Specific [ H]-NMS binding

(B value) remaining at various dissociation time points is normalized

to initial binding (B₀value) at time 0. B/B₀ratios are mean SD

values for two to four independent experiments. Best fits to a

monoexponential decay are shown (Supplemental Procedures). B:

Concentration-effect curves are plotted from [³H]-NMS dissociation

studies performed in the presence of 10 μM atropine and of increasing

●

△

concentrations of gallamine ( ) or para-LRB-AC42 ( ). Observed

dissociation rates (k_{off obs}) are normalized to control rates (k_{off control}

)

measured with atropine alone. Data are mean SD values for three

independent experiments. Curve fitting to the Hill equation provided

EC_{50 diss}and E_{max diss}parameters (Table 3). When selecting a 1−50 μM

para-LRB-AC42 concentration window, curve fitting proceeded with

maximal amplitude and slope constrained to 1 (dashed line).

occupation (equivalent to an IC₅₀) at 10

1 μM, and a

maximal amplitude (102 3%) nonsignificantly different from

100%. Finally, very similar parameters were determined after

correction of each raw data point for the fluorescent artifact

measured at the same AC42-NH₂concentration on naive cells

(Figure 5B inset). The latter procedure was used routinely to

quantify and correct undesirable fluorescence artifacts due to

some chemicals.

Association and dissociation kinetics for para-LRB-AC42

binding to EGFP(Δ17)hM1 cells are very fast (Figure 6A).

This enabled us to monitor, over 50 s periods of time, the

impact on cell fluorescence of the addition of cumulative doses

of tracer (Figure 6C). The concentration-dependent fluo-

rescence extinction (with ligand emission artifact visible at high

tracer concentration) was fully prevented by cell preincubation

with atropine (20 μM). Specific para-LRB-AC42 binding, as

defined by the difference in fluorescence levels between both

assays, reaches a maximum at 3 μM tracer, featuring a saturable

process.

μM) to the cell suspension, fluorescence decreases almost

instantaneously to a plateau value featuring stable binding

equilibrium. Subsequent addition of AC-42 (b, 30 μM) induces

a fast fluorescence recovery, indicative of para-LRB-AC42

dissociation, reaching a level that exceeds that achieved when

using atropine (20 μM). A key observation was the measure-

ment of an identical increment of fluorescence when adding

AC-42 to naive (c) or atropine-treated (d) cells. The origin of

such an artifact remains unclear. It is reproduced well with

AC42-NH₂and exhibits a strict concentration dependency

(Figure 6B inset), is independent of the nature and the

expression levels of EGFP-fused receptors, is related somehow

to HEK cell constituents, and is negligible in Hepes buffer. As a

consequence, AC-42 and AC42-NH₂cannot be taken as safe

compounds to validate the receptor specificity of FRET signals,

under equilibrium or kinetic conditions.

Another point highlighted in Figure 6A is the apparent

discrepancy between the amplitudes of fluorescence extinction

at equilibrium and of fluorescence recovery upon addition of

atropine. A series of controls, performed with Hepes buffer or

nontransfected HEK cells, indicate that actual fluorescence

Finally and to further confirm the muscarinic specificity of

the FRET signal associated with para-LRB-AC42 binding, the

experiment was repeated using propylbenzilylcholine mustard

(PrBCM) to selectively alkylate the Asp105 (3.32) receptor

2132

dx.doi.org/10.1021/jm201348t | J. Med. Chem. 2012, 55, 2125−2143

Journal of Medicinal Chemistry

Article

Figure 6. Definition of the specificity of FRET signals associated with para-LRB-AC42 binding to EGFP(Δ17)hM1 receptors. A: Real-time

monitoring of fluorescence variations (data sampled every 0.5 sec) of EGFP(Δ17)hM1 cells upon addition (arrow a) of 5 μM para-LRB-AC42 and

○

(arrow b) of either atropine (20 μM; thick trace) or AC-42 (30 μM; thin trace). Two other experiments ( ) point to a very similar artefactual

increase in fluorescence (16500 cps) following the addition of 30 μM AC-42 to naive (arrow c) or para-LRB-AC42 and atropine-treated (arrow d)

cells. Specific binding of para-LRB-AC42, as defined by the atropine-sensitive fluorescence recovery (double headed arrow), corresponds to a FRET

signal with 24.5% extinction of cell fluorescence. After deduction of its artefact, AC-42 was found to displace up to 80% (---) of specific para-LRB-

AC42 binding. B: Validation of a competition experiment using AC42-NH₂and para-LRB-AC42 under FRET conditions. EGFP(Δ17)hM1

expressing cells are incubated for 1 h at 20 °C with 5 μM para-LRB-AC42 in the absence or the presence of either atropine (20 μM) or increasing

concentrations of AC42-NH₂. Fluorescence levels at equilibrium are measured at 510 nm as reported in the Experimental Section. The AC42-NH₂

□

concentration-dependent increase in fluorescence ( ) is normalized in percent to the amplitude for atropine-prevented fluorescence extinction

(specific FRET signal; 100%). Data fitting (--) to the Hill equation derived for saturation superimposed with a linear drift to account for AC42-NH₂

●

artefact allowed to extract a typical inhibition curve ( ) for para-LRB-AC42 binding, with an IC₅₀(μM) of 10.2 1.2 and a maximal amplitude (%)

of 101.8 2.9, and to define a slope (0.55 0.1) for the linear drift (not shown). A parallel experiment (insert), performed on the same cells in the

absence of para-LRB-AC42, provided a direct measurement of AC42-NH₂fluorescence artefact. Data are normalized as detailed above and fitted

■

according to a linear relationship (slope factor 0.58 0.01). C: Receptor alkylation by PrBCM (30 nM) and para-LRB-AC42 binding. Control ( ,

□

● ○

) or PrBCM-treated ( , ; Experimental Section) EGFP(Δ17)hM1 cells, taken at an identical density, were preincubated (open symbols) or not

(black symbols) with 20 μM atropine for 5 min at 20 °C. Thereafter, cell samples were examined for their ability to generate a FRET signal upon

sequential addition (arrows) of cumulative doses of para-LRB-AC42 (final concentrations indicated in μM). Fluorescence at 510 nm was monitored

in real time (1 point per second; 20% of points are shown). Specific FRET signals (atropine-protected fluorescence extinction in %) measured with 5

μM para-LRB-AC42 were respectively of 25 and 2.8 for control and PrBCM-treated cells, respectively.

residue.⁴⁷Indeed, this conserved amino acid plays a well-known

critical role in the binding of orthosteric ligands to all aminergic

GPCRs.⁴⁸Its mutation has also been shown to abolish signaling

by novel muscarinic agonists such as AC-42, 77-LH-28-1, and

NDMC.^12,23Upon treatment with PrBCM (30 nM), we found

that EGFP(Δ17)hM1 cells display a marked reduction (90−

95%) in their ability to elicit a calcium response to AC-42 (10

μM), to bind [³H]NMS (0.7 nM), or to promote a FRET

signal in the presence of the fluorescent pirenzepine derivative

RR(17)PZ³⁸at 200 nM. As shown in Figure 6C, M1 receptor

alkylation by PrBCM also almost completely abolishes (up to

90% at saturating tracer concentrations) the atropine-sensitive

FRET signal promoted by para-LRB-AC42 binding.

low affinity LRB-SO₃Na⁺binding was fitted according to a

linear relationship with tracer concentration.

−

Assuming that cell surface receptors account for 90% to

overall fluorescence of EGFP(Δ17)hM1 cells (ratio of B_max

values for [³H]NMS and [³H]QNB binding, Table 2), maximal

fluorescence extinction (27%, measured at saturating para-LRB-

AC42 concentrations) may be converted into specific EGFP

extinction (30%) to provide the energy transfer efficacy E

(0.30). This enabled us to estimate the actual distance R (59.2

0.7 Å) separating EGFP from the LRB fluorophore of the

tracer, using the R = R₀(1/E − 1)^1/6Forster relationship⁴⁹and

R₀values listed in Table 1 (Supporting Information

Procedures).

̈

We finally focused on para-LRB-AC42 equilibrium binding

properties through competition experiments (Figure 7B)

performed on EGFP(Δ17)hM1 cells using a set of well-

known orthosteric and allosteric muscarinic compounds. NMS,

atropine, AC-42, and AC42-NH₂(not shown) all follow a

competitive-like mode of interaction. The allosteric modulators

gallamine and brucine (not shown) also fully inhibit specific

para-LRB-AC42 binding at high concentrations. Interestingly,

Win 51,708,⁵⁰an allosteric modulator known to interact on M1

receptors at a site distinct from that of gallamine, and

NDMC,^12,23an atypical allosteric agonist possibly overlapping

the [³H]NMS orthosteric binding domain, both exhibit weak

negative binding cooperativity with para-LRB-AC42. The

potent and M1 selective allosteric MT7 toxin⁴⁵was totally

inactive even at 10 nM (not shown). Quantitative analysis of

the data was done using appropriate sets of equations for

competitive or cooperative interactions to define the

parameters listed in Table 4. The affinity constants of drugs

Equilibrium Binding Properties of Para-LRB-AC42 at

EGFP(Δ17)hM1 Receptors. Saturation experiments were

performed on a 0.25−10 μM concentration range of para-

LRB-AC42 (Figure 7A). Essentially similar results were

obtained whether EGFP(Δ17)hM1 cells (preincubated or not

for 5 min with 20 μM atropine) were submitted to cumulative

tracer addition and checked in real time for fluorescence levels

at binding equilibrium (as in Figure 6C) or incubated for 4 h at

20 °C with increasing tracer concentrations in separate test

tubes. Data fitting to the empirical Hill equation derived for

saturation allowed the determination (mean SE values, n = 5)

of the K_d(0.8 0.1 μM), maximal FRET amplitude (26.5

1.4%), and slope factor (1.06 0.08) for specific binding of

para-LRB-AC42 to EGFP(Δ17)hM1 receptors. Note that the

K_dfor the fluorescent tracer coincides well with its affinity

estimated from [³H]NMS competition studies (Table 3),

thereby confirming the validity of FRET measurements. Very

2133

dx.doi.org/10.1021/jm201348t | J. Med. Chem. 2012, 55, 2125−2143

Journal of Medicinal Chemistry

Article

Taken together, these observations constitute a strong and

direct evidence for a bitopic nature of the tracer, with

simultaneous occupancy of both the orthosteric and the

gallamine allosteric receptor sites.

Molecular Modeling. The three-dimensional structure of

the human M1 muscarinic receptor was modeled by homology

to the dopamine D3 receptor, which is the most similar GPCR

available in the Protein Data Bank to date (about 25% overall

sequence identity and 36% sequence identity in the seven

transmembrane domains (7TMs) with the M1 receptor).

Because the position of the conserved cysteine in the second

extracellular loop (ECL2) differs in M1 and D3 receptors, this

part of the model was reconstructed in order to bury

hydrophobic side chains and expose hydrophilic ones while

producing nondistorted secondary structure elements. A short

molecular dynamic simulation revealed that the extracellular

loops are prone to conformational variations whereas the

rotameric state of residues in each of the transmembrane

helices (TM) is rather well-defined (Figure S2, Supporting

Information).

Although our model is not accurate at the atomic resolution,

it allows a rough delineation in the M1 receptor of three

regions enriched in hydrophobic and negatively charged

residues. These regions constitute known and putative

anchoring spots for positively charged, orthosteric, and

allosteric muscarinic ligands (Figure 8). A selection of key

residues in each region is shown in Figure 8A; it includes

Trp101 (3.28) and Phe77 (2.56), both reported in the

literature for their (concerted) contribution to AC-42 binding

and/or function.^23,24,26The first region is located within the

7TMs of the protein (Figure 8B), displays a high content in

aromatic residues (Trp101, Tyr106, Tyr381, Tyr404, Tyr408),

and exposes the side chains of Asp105 (3.32) and Asn382

(6.52) near its center. It corresponds to the orthosteric site.^51,52

In the model, its access is obstructed by bulky residues in TM2

(Tyr82, Tyr85) and TM7 (Trp400) and by ECL2. Noteworthy,

ECL2 is constrained by the first and third extracellular loops

(ECL1 and ECL3, respectively) and covalently linked to TM3,

thereby restricting the opening of the orthosteric site whatever

the loop conformation (i.e., even if the modeled ECL2

conformations is partly wrong). The second region is located

at the receptor surface close to the N-terminal end of the

protein chain. It involves the extracellular parts of TM1, TM2,

TM3, and TM7 and part of ECL2 (Figure 8C, green). It

Figure 7. Binding properties of fluorescent tracers and of a series of

muscarinic compounds at EGFP(Δ17)hM1 receptors as monitored by

FRET experiments. A: EGFP(Δ17)hM1 cells were incubated at 20 °C

●

for 4 h with increasing concentrations of para-LRB-AC42 ( ) or LRB-

−

SO₃Na⁺

(

△

), in the absence or the presence of 20 μM atropine.

Specific binding of the tracers (defined as an atropine-sensitive

fluorescence extinction at 510 nm) is expressed in percent of control

cell fluorescence. Data fitting to the empirical Hill equation derived for

saturation (Supplemental Procedures) allowed the determination of

−

the binding parameters for para-LRB-AC42. Specific LRB-SO₃Na⁺

binding was fitted according to a linear relationship with tracer

concentration (---). B: Competition experiments were performed

using EGFP(Δ17)hM1 cells incubated (4 h at 20 °C) with 2 μM para-

●

▼

LRB-AC42 and increasing concentrations of NMS ( ), atropine ( ),

◆

□

▲

AC-42 ( ), gallamine ( ), N-desmethylclozapine ( ) or Win 51,708

○

( ). Data from typical experiments, performed in duplicate, are

presented as B/B₀ratios with specific binding B₀taken as the

difference in fluorescence levels of samples incubated with the tracer in

the absence or the presence of 20 μM atropine. Best fits to the most

appropriate (competitive or allosteric) model are shown (Supple-

mental Procedures). Parameters are listed in Table 4.

for their individual sites on the EGFP(Δ17)hM1 receptor, as

determined from radioligand or from FRET equilibrium

binding studies, are in remarkable overall agreement.

Table 4. Parameters for the Interaction of Various Compounds at EGFP(Δ17)hM1 Receptors As Defined from Equilibrium

a

Para-LRB-AC42 Binding Studies Performed under FRET Conditions

b

3

compd

pK_d

pK_i

pK_x

log α

slope

pK_{[ H]NMS}− pK

FRET

para-LRB-AC42

AC-42

6.10 0.05

1.06 0.08

1.02 0.04

0.96 0.04

1.02 0.06

1.05 0.06

1.04 0.04

0.91 0.09

1.02 0.06

1.02 0.05

−0.10 0.13

−0.29 0.11

−0.18 0.08

0.25 0.1

0.32 0.11

0.23 0.08

0.24 0.07

0.30 0.08

nd

6.05 0.04

5.89 0.02

9.77 0.06

8.92 0.07

5.07 0.04

4.35 0.06

AC42-NH₂

NMS

atropine

gallamine

brucine

NDMC

6.92 0.03

5.03 0.05

0.74 0.06

0.92 0.02

WIN 51,708

a

Affinity constants (−log M, mean values SE from three or more independent measurements) are from saturation (K_dvalues) and competition

(K_i, K_x, and α values) studies, as detailed in the Experimental Section. Competition experiments were performed at various para-LRB-AC42

concentrations (1× to 6× its K_dvalue) and analyzed according to a competitive (K_i) or allosteric (K_x, α) model, when appropriate (Supporting

Information Procedures). pK _H[NMS]values are from Table 3. nd: not determined.

b

3

2134

dx.doi.org/10.1021/jm201348t | J. Med. Chem. 2012, 55, 2125−2143

Journal of Medicinal Chemistry

Article

Figure 8. Three-dimensional model of the muscarinic M1 receptor. A: Position of key aromatic residues and solvent accessible charged residues with

no vicinal counter charges in the protein chain. The 7TMs are represented by cylinders. The residue side chains are displayed as capped sticks and

labeled in red if negatively charged. B: Orthosteric site. The 7TMs are represented by cylinders. The transmembrane cavity is delimitated by its

solvent excluded surface colored in magenta. C: Allosteric site. The protein is represented by its solvent excluded surface, which is colored in green

and orange around the negatively charged residues and otherwise colored in white. In A, B, and C, top and bottom panels show views from the

extracellular side or in the plane of the plasma membrane, respectively. The pictures were made using Sybyl-X (Tripos, Inc., St. Louis, Missouri,

USA).

contains Trp400 (7.35) and two negatively charged residues,

Glu397 and Glu401 (7.36). The third region is superficial too

and resides near the extracellular parts of TM3 and TM4

(Figure 8C; orange). It includes the negatively charged Asp99

(3.26), surrounded by aromatic and hydrophobic residues from

ECL1 (Trp91) and ECL2 (Trp164, Tyr166). The three regions

are clearly interconnected via common or bridging amino acids.

The two superficial domains may be defined as allosteric in

nature with respect to the buried orthosteric site. Moreover,

they contain several residues (Asp99, Trp101, Trp400, for

example) involved in gallamine binding.⁴²They surround the

route leading to the central binding site and may thus also

participate in the primary docking step of orthosteric ligands.⁵³

This M1 model provides the structural foundations for

further discussion of our experimental data on AC-42 and its

fluorescent derivatives.

supported by the current view (and our M1 model) that

muscarinic orthosteric and allosteric sites are in proximity and

even on the access way to each other,^48,54is to consider these

compounds as bitopic, i.e., able to recognize epitopes within

both orthosteric and allosteric sites and to interact with them

simultaneously.²⁴How robust is the interpretation of point

mutagenesis studies regarding the definition of topographically

distinct or overlapping binding domains? Do novel agonists

exhibit different binding poses on free or [³H]NMS-occupied

receptors? What are the structural determinants of these

molecules responsible for such hybrid properties ?

Within this very challenging context, traceable novel agonists

may be regarded as tools of choice to get direct information on

their interaction with the M1 receptor. Tritiated AC-42 could

have been an ideal candidate, as it retains the structure and the

activity of the original molecule. However, given its micromolar

affinity, high nonspecific binding levels, fast dissociation

kinetics, and other difficulties are expected, as was encountered

with the [³H]dimethyl-W84 radioalloster.⁵⁵We thus consid-

ered the possibility of fluorescent derivatives of AC-42 to

behave as tracers enabling direct monitoring of binding events

through fluorescence resonance energy transfer (FRET) from

EGFP-fused M1 receptors. This technology is now well-

admitted as a powerful alternative to radioligand binding assays,

prone to fine, temporal, and quantitative description of complex

binding mechanisms.^7,27,28,56The EGFP(Δ17)hM1 construct,

with EGFP fused to a truncated receptor N-terminus,²⁷was

DISCUSSION AND CONCLUSION

■

There is no doubt that novel M1 agonists, such as AC-42 and

its analogues, together with TBPB and NDMC, display a mixed

orthosteric/allosteric binding behavior. On one hand, they

exhibit competitive-like properties when tested against classical

orthosteric compounds in both equilibrium binding and

functional assays.^12,13,23,24On the other hand, their ability to

retard [³H]NMS dissociation and their particular pattern of

response to mutations of key orthosteric residues are taken as

evidence for allosteric properties.^{10−12,14,23,26}An emerging idea,

2135

dx.doi.org/10.1021/jm201348t | J. Med. Chem. 2012, 55, 2125−2143

Journal of Medicinal Chemistry

Article

Figure 9. Hypothetical binding modes of AC-42 and para-LRB-AC42 to the M1 muscarinic receptor. A: Upside-down positions of AC-42 within the

7TMs cavity. The orthosteric site is circled in magenta; the allosteric site is dashed in green-orange to reconcile the two regions depicted in Figure

8C. B: Two possible bitopic binding modes for para-LRB-AC42. For a sake of clarity, ligands are presented as CPK colored sticks and hydrogen

atoms are omitted. In the representation, the scale of the receptor is slightly smaller than that of the ligand. In our three-dimensional model, the

distances between negatively charged residues range from 13 Å to 17 Å.

validated here as a suitable tool for examining AC-42

properties. Indeed, and in contrast with previous observa-

tions,¹⁰neither the fusion of EGFP nor the deletion of the

receptor N-terminus significantly affects AC-42 agonism and

affinity properties. Moreover, the EGFP(Δ17)hM1 chimera

shares with the wild type hM1 receptor superimposable affinity

and modulatory properties for a series of prototypical

orthosteric and allosteric molecules.

para-LRB-AC42 binding takes place on M1 muscarinic

receptors.

What do we learn about the interaction of para-LRB-AC42

(and AC-42) with EGFP(Δ17)hM1 receptors ? From the

orthosteric point of view ([³H]NMS binding studies), para-

LRB-AC42 may be regarded as an unlabeled compound that

shares several properties with AC-42: it binds to [³H]NMS sites

in a competitive manner, exhibits an affinity comparable to that

of AC-42, and slows [³H]NMS dissociation in a concentration-

dependent manner. As the presence of the fluorophore in the

para position is almost neutral in terms of affinity, and preserves

the allosteric character of the molecule at [³H]NMS occupied

receptors as well as M1/M3 selectivity, it is tempting to suggest

that the contribution of LRB to overall para-LRB-AC42 binding

energy is negligible and that it does not significantly alter

proper positioning of the AC-42 moiety on the M1 receptor.

Under FRET conditions, para-LRB-AC42 exhibits the

potential of a micromolar affinity tracer. Moreover, it provides

the first direct evidence for a bitopic binding mode on free M1

receptors: its interaction with the EGFP(Δ17)hM1 receptor is

inhibited by both orthosteric and allosteric ligands, as well as by

AC-42, in a competitive manner. Specific para-LRB-AC42

binding to M1 receptors is abolished (together with AC-42

function and specific [³H]NMS binding) following covalent

binding of PrBCM to the orthosteric Asp105 (3.32) receptor

residue. Whether this bulky affinity label precludes a direct

interaction of the fluorescent probe with Asp105 or sterically

hinders its access to surrounding residues remains to be

clarified. Importantly, the allosteric binding component of para-

LRB-AC42 clearly targets the gallamine site, as Win 51,708 and

NDMC display a typical noncompetitive binding behavior.

Such a view provides additional support to the indirect

information by Avlani et al.²⁴on the competitive nature of

77-LH-28-1 and C₇/3-phth (an allosteric modulator that binds

to the gallamine site⁵⁷) at the [³H]NMS occupied M1 receptor.

Refined kinetic investigations, such as real-time FRET

monitoring of para-LRB-AC42 binding properties, were

Four fluorescent compounds, i.e., the LRB moiety alone

(LRB-SO₃Na⁺) or substituted in the para position with a butyl

−

chain (para-LRB-n-butyl 20) and its two isomers linked to

AC42-NH₂(ortho-LRB-AC42 2a and para-LRB-AC42 2b),

were synthesized and characterized. Solubility measurements,

supported by biological evidence (partial inhibition in func-

tional and binding experiments), pointed to serious limitations

in the use of para-LRB-n-butyl and ortho-LRB-AC42. Affinity

parameters for para-LRB-n-butyl could not be determined with

great precision, but the presence of a butyl chain in the para

position (and the concomitant neutralization of a SO₃⁻charge)

clearly confers to this molecule the ability to bind to M1

receptors in the micromolar concentration range, as opposed to

the millimolar affinity of LRB-SO₃Na⁺.

−

AC-42, AC42-NH₂, and para-LRB-AC42 display fairly similar

affinity constants for the receptor, although the modifications

we introduced within AC-42 converted the compounds into

antagonists. One explanation is that the exchange of the AC-42

butyl chain for a pentyl one and the concomitant addition of a

second protonated amino group in AC42-NH₂may disrupt an

essential ionic bond (presumably between the positively

charged AC-42 piperidine nitrogen and the receptor Asp105

carboxylate) and/or alter a preferential orientation of the AC-

42 aliphatic chain within the receptor, two critical aspects for

AC-42 agonism.^23,24Such explanations probably do not apply

to para-LRB-AC42, as its hardly basic sulfonamido group

remains neutral at physiological pH.

Numerous control FRET experiments were performed under

equilibrium and kinetic conditions, taking para-LRB-AC42 as

the tracer and using pure EGFP, nontransfected HEK cells, and

cells overexpressing the EGFP(Δ17)hM1 or other EGFP-fused

receptors. They allowed ligand emission interference and an

unexpected artifact of AC-42 (and AC42-NH₂) on cell

autofluorescence to be detected, quantified, and carefully

corrected. They provided operational criteria for a robust

definition of the specificity of FRET signals. FRET and

[³H]NMS binding experiments provided nicely concordant

affinity parameters for all orthosteric and allosteric compounds

included in this study. Altogether, we demonstrate that specific

hampered by very fast association and dissociation (k_off

=

0.65 0.01 s⁻¹at 20 °C; n = 5) rates. For comparison, the off-

rate of para-LRB-AC42 was about 10 times faster than that

estimated for the nanomolar radioalloster [³H]dimethyl-W84

(0.07 s⁻¹) on the M2 receptor.⁵⁵

What are the orthosteric and allosteric molecular determi-

nants in the para-LRB-AC42 and AC-42 molecules? FRET

measurements of donor/acceptor distances indicate that this

tracer, when bound to the EGFP(Δ17)hM1 receptor, locates

its LRB fluorophore at a greater distance from EGFP (59 Å)

2136

dx.doi.org/10.1021/jm201348t | J. Med. Chem. 2012, 55, 2125−2143

Journal of Medicinal Chemistry

Article

than do a series of fluorescent Bodipy or Rhodamine Red-X

pirenzepine derivatives.³⁸Some of them, including RR(17)PZ,

were already pointed out for their bitopic binding properties:

the pirenzepine moiety binds to the orthosteric site, whereas

the Bodipy (or Rhodamine Red) fluorophore is stacked within

the classical M1 allosteric binding site (at 45−50 Å from

EGFP) and competes with brucine (and gallamine) for binding

to the receptor.³⁸Clearly, the possibility for the fluorophore in

para-LRB-AC42 to occupy the classical gallamine site is

unlikely. The para-LRB-n-butyl and ortho-LRB-AC42 deriva-

tives would have been valuable tools to further dissect the

molecular bases for the bitopic binding pose of the fluorescent

derivative. Unfortunately, these expectations fell short because

of solubility limitations incompatible with the micromolar

affinity of these two compounds. Thus, to gain insight into M1

receptor domains that may accommodate para-LRB-AC42 (and

AC-42), the three-dimensional structure of the receptor was

modeled by homology to the dopamine D3 receptor. As the

M1 model was built from a distant homologue, we felt it was

not reliable enough to conduct an automated docking

approach. Preliminary trials, using different docking tools and

alternative receptor conformations generated by molecular

dynamics, indeed did not predict a consensual meaningful pose

for AC-42 and its derivatives (data not shown). This is the

reason why the coming discussion will not focus on possible

intermolecular interactions but instead will suggest which

binding modes best agree with the geometric and electrostatic

properties of the ligands and the M1 receptor.

According to our three-dimensional model, the orthosteric

site in the 7TMs (with its typical cleft shape) is wide and deep

enough to bury AC-42 and bulkier ligands (Figure 9A). The

volume of the protein cavity is about 540 Å³(and could be even

larger if ECL2 would not close the site entrance), whereas the

volume of AC-42, in an extended conformation, is only about

340 Å³. The degree of penetration and the relative position of

AC-42 within the receptor cavity are hardly predictable because

it is a flexible and pseudosymmetrical molecule. The selection

of the AC-42 active conformer among a profusion of

possibilities requires an accurate definition of the receptor

three-dimensional structure. In addition, the central position of

the positively charged piperidine of AC-42 precludes the

hypothesis of a preferential orientation of this ligand in the

vicinity of the receptor extracellular domains that display an

overall negative electrostatic potential. Accordingly, the

literature reports on different docking orientations and

positions of AC-42 into the M1 muscarinic receptor.^23,24,26

Nevertheless, all the binding poses proposed for AC-42 (two

examples are shown in Figure 9A, right and left) practically

agree with experimental observations based on site-directed

mutagenesis of the M1 receptor. In particular, several aromatic

residues such as Trp101, Phe77, and the less consensual

Tyr404 were shown to be important determinants for AC-42

binding and function, probably by controlling the accessibililty

of the transmembrane cavity.^23,24,26According to our model,

these residues, possibly together with Trp400, constitute a

gateway between the orthosteric and the allosteric sites. If one

considers the additional possibility of AC-42 to interact with

the orthosteric Asp105 residue,^12,23it becomes obvious that the

molecule may overlap both the orthosteric and allosteric sites

and thus exhibit a bitopic binding pose. Several points,

however, need to be clarified. Indeed, given the presence of

important steric constraints in the unoccupied M1 receptor,

one may suspect that extended and embedded binding poses

for AC-42 may not predominate. Moreover, such positions

hardly reconcile with the observation of allosteric binding

properties for AC-42 on [³H]NMS-occupied receptors (this

study and references therein).^11,12

The introduction of the LRB fluorophore onto AC42-NH₂

results in an elongation of about 10 Å of the molecule, thereby

restraining the number of accessible positions for AC-42 within

the cavity (Figure 9B). Both AC-42 and the AC-42 moiety of

para-LRB-AC42 may occupy the same region in the receptor

(Figure 9A right and Figure 9B left) so that the bulky

fluorophore is placed next to ECL2 and ECL3 on top of (or in

partial overlap with) the allosteric site. An alternative binding

mode would correspond to the switch of the AC-42 moiety of

para-LRB-AC42 toward the extracellular allosteric domains of

the receptor (Figure 9B right). The piperidine ring could hence

establish an ionic bond in the gallamine allosteric domain, while

LRB would occupy the opposite side of the allosteric site

(orange or green patches in Figure 8C), leaving a place for the

linker between AC-42 and the fluorophore to explore the top of

the orthosteric site. Although such a more superficial pose

agrees better with very fast para-LRB-AC42 binding kinetics,

both assumptions explain well why the distance between the

LRB and EGFP fluorophores is 10 Å longer than that estimated

by FRET, on the same receptor, using pirenzepine derivatives

known to stack their Bodipy or Rhodamine Red fluorophore

within the gallamine site.

Future mutagenesis mapping studies should provide further

information on the M1 receptor domains that accommodate

AC-42 and the para-LRB-AC42 derivative. Although the

interpretation of such data is often ambiguous regarding the

definition of topographically distinct or overlapping binding

sites, one may expect interesting clues from the effects of the

mutation of the negatively charged amino acids (Asp 99,

Glu397, Glu 401) and of the aromatic Trp101 and Trp 400

residues. Indeed, they may orientate the choice among the

possibilities depicted in Figure 9, help define the allosteric

subdomains occupied by these ligands, give additional support

to or invalidate the possibility of a similar binding pose for AC-

42 and the AC-42 moiety of para-LRB-AC42, and perhaps

allow the extension of FRET experiments using the ortho-LRB-

AC42 and para-LRB-n-butyl tools.

In conclusion, this study provides new pieces of evidence for

a bitopic nature of AC-42 binding on M1 receptors. Our

strategy may be compared with the “fragment approach” that

revealed the bitopic nature of McN-A-343 at M2 receptors,⁵⁸

with the additional possibility here of monitoring the binding

properties of intermediates and final compounds through

FRET. Interpretation of such data is not straightforward, as

isolated or fused fragments do not necessarily occupy the same

binding pockets. Linker orientation, length, and nature are

critical factors for sculpting affinity, potency, and site-targeting

of hybrid molecules.^38,59It remains that dualsteric, hybrid, or

bitopic molecules may combine high affinity (orthosteric

fragment) and subtype-selective (allosteric fragment) receptor

binding to engender functional selectivity.^60,61Their rational

design and development are expected to provide further

insights into ligand binding mechanisms and to improve the

specificity of therapeutic GPCR activation.

EXPERIMENTAL SECTION

■

General Chemistry. [N-Methyl-³H]scopolamine chloride

([³H]NMS, 82 Ci/mmol) and quinuclidinyl [phenyl-4-³H] benzilate

([³H]QNB, 30 Ci/mmol) were from Perkin Elmer Life and Analytical

2137

dx.doi.org/10.1021/jm201348t | J. Med. Chem. 2012, 55, 2125−2143

Journal of Medicinal Chemistry

Article

Sciences (Courtaboeuf, France). Atropine sulfate, NMS bromide,

pirenzepine dihydrochloride, carbachol chloride, N-desmethylcloza-

pine hydrochloride, gallamine triethiodide, and WIN 51,708 hydrate

were purchased from Sigma-Aldrich (Saint-Quentin Fallavier, France).

Indo-1-acetoxymethyl ester was supplied by Molecular Probes

(Invitrogen, Cergy Pontoise, France). Lissamine rhodamine B sodium

(m, 8H), 3.48 (bd, J = 12.2 Hz, 2H), 3.08−3.02 (m, 4H), 2.94−2.87

(m, 4H), 2.41 (s, 3H), 2.00−1.85 (m, 4H), 1.51−1.45 (m, 3H), 1.37−

1.21 (m, 20H). ¹³C NMR (DMSO-d₆, 100 MHz): δ 157.8, 157.5,

157.3, 157.0, 155.0, 147.9, 141.5, 137.3, 137.0, 132.9, 132.6, 131.6,

130.6, 128.6, 128.4, 126.5, 125.9, 125.6, 113.6, 113.4, 95.4, 55.5, 52.0,

45.2, 42.4, 37.5, 35.0, 32.5, 29.1, 28.9, 25.9, 25.2, 20.8, 18.2, 12.4 (bs).

RP-HPLC purity: 95%; t_R= 18.9 min. HRMS: calcd for C₄₈H₆₃N₄O₇S₂

871.4133, found 871.4122

−

sulfonate (LRB-SO₃Na⁺, sulforhodamine B) was from TCI Europe

(Zwijndrecht, Belgium) and lissamine rhodamine B sulfonyl chloride

(mixed isomers) from Acros Organics (Halluin, France). 4-n-Butyl-1-

[4-(2-methylphenyl)-4-oxo-1-butyl]piperidine hydrogen chloride (AC-

42) was a generous gift from Dr. T. Spalding (Acadia Pharmaceuticals

Inc., San Diego, CA) and Pr. A. Christopoulos (Monash University,

Victoria, Australia). Propylbenzilylcholine mustard was generously

provided by Dr. N. Birdsall (NIMR, London, U.K.).

Ortho-LRB-AC42 (2b, purple powder, 10 mg, 10%): ¹H NMR

(DMSO-d₆, 500 MHz) δ 8.40 (d, J = 1.5 Hz, 1H), 8.02 (dd, J = 7.8

and 1.5 Hz, 1H), 7.94 (t, J = 5.6 Hz, 1H), 7.79 (dd, J = 7.4 and 1.2 Hz,

1H), 7.47−7.43 (m, 2H), 7.35−7.30 (m, 2H), 7.08 (dd, J = 9.6 and 2.4

Hz, 2H), 6.99 (d, J = 9.5 Hz, 2H), 6.96 (d, J = 2.4 Hz, 2H), 3.65 (q, J

= 7.2 Hz, 8H), 3.49−3.47 (m, 2H), 3.08−3.05 (m, 4H), 2.92−2.83 (m,

3H), 2.73 (t, J = 6.9 Hz, 1H), 2.41 (s, 3H), 2.00−1.92 (m, 2H), 1.83−

1.80 (m, 2H), 1.33−1.15 (m, 23H). ¹³C NMR (DMSO-d₆, 100 MHz):

δ 157.6, 157.1, 155.2, 155.1, 152.2, 150.4, 148.4, 139.9, 139.6, 137.3,

137.0, 131.6, 131.4, 131.0, 130.3, 129.5, 129.2, 128.6, 125.9, 125.6,

114.2, 113.5, 95.7, 55.5, 52.0, 45.3, 42.2, 37.6, 35.0, 32.5, 29.1, 28.8,

25.9, 25.1, 20.8, 18.2, 12.4 (bs). RP-HPLC purity: >95%; t_R= 18.0

min. HRMS: calcd for C₄₈H₆₃N₄O₇S₂871.4133, found 871.4138.

4-(2-Methylphenyl)-4-oxobutanoic Acid (3). o-Tolylmagnesi-

um bromide (2.0 M in diethyl ether, 5.5 mL) was added to a solution

of succinic anhydride (1 g, 10 mmol) in anhydrous THF (20 mL) at

−78 °C. The mixture was stirred at room temperature for 2 h, then

diluted in CH₂Cl₂, washed with a 1 M HCl solution, 10% NaHCO₃,

and brine solutions, and finally dried over MgSO₄. The solvent was

removed under reduced pressure and the crude product was purified

by flash chromatography using an EtOAc/hexane (2/8, v/v) solution

¹H NMR spectra were recorded on a Bruker Advance spectrometer.

Chemical shifts (δ values) are expressed in ppm relative to internal

solvent peaks, and coupling constants (J) are measured in hertz.

Signals are described as s (singlet), d (doublet), t (triplet), m

(multiplet), and brs (broad singlet). Analytical reverse-phase high

performance liquid chromatography (RP-HPLC) separations were

performed on a C18 Symmetry Shield column (4.6 mm × 150 mm)

using a linear gradient (100% A for 1 min and then 0% to 100% B in

30 min, flow rate of 1 mL·min⁻¹) of solvent B (100% CH₃CN, 0.1%

TFA, v/v) in solvent A (100% H₂O, 0.1% TFA, v/v). Detection was

set at 220 and 254 nm. Purified final compounds eluted as single and

symmetrical peaks (thereby confirming a purity of ≥95%) at retention

times (t_R) given below. Their identity was determined by high-

resolution mass spectra (HRMS) which were acquired on a Bruker

MicroTof mass spectrometer, using electrospray ionization (ESI) and

a time-of-flight analyzer (TOF).

1

as eluant. R_f= 0.15, white solid (1.88 g, 98%). H NMR (300 MHz,

UV−visible absorbance spectroscopy was done with a Cary 1E

spectrophotometer (Varian). Fluorescence measurements were made

using a SPEX Fluorolog 2 (Jobin Yvon Horiba) spectrofluorimeter.

4-[4-(5-Aminopentyl)piperidinyl]-1-o-tolylbutanone, HCl

Salt (1). To a solution of methyl sulfonate 7 (285 mg, 0.95 mmol)

and piperidine derivative 8 (255 mg, 0.95 mmol) in acetonitrile (10

mL) was added triethylamine (198 μL, 1.42 mmol). The mixture was

heated at 110 °C overnight in a sealed tube and concentrated in vacuo

to a residue. Subsequent purification by flash chromatography

(EtOAc/MeOH, 5/95, v/v) gave access to the tert-butyl 5-[1-(3-(2-

o-tolyl-1,3-dioxolan-2-yl)propyl)piperidin-4-yl]pentylcarbamate com-

pound 9 (90 mg, 20%) which was dissolved in a 1 N HCl/Et₂O

solution and stirred for 1 h at room temperature. The solvent was

removed under reduced pressure and the residue triturated in ether to

give 1 as a salt (52 mg, 84%). RP-HPLC purity >95%; t_R= 16.8 min.

¹H NMR (300 MHz, CD₃OD): δ 7.77 (d, J = 8.2 Hz, 1H), 7.39−7.24

CDCl₃): δ 7.69 (d, J = 8.2 Hz, 1H), 7.42−7.25 (m, 3H), 3.24 (t, J =

6.2 Hz, 2H), 2.80 (t, J = 6.2 Hz, 2H), 2.51 (s, 3H).

Methyl 4-(2-Methylphenyl)-4-oxobutanoate (4). A solution of

acid 3 (500 mg, 2.6 mmol) in 30 mL of methanol with sulfuric acid

(50 μL) was refluxed overnight. The solvent was removed under

reduced pressure, then diluted with EtOAc and washed with 10%

NaHCO₃, brine and dried over MgSO₄. After solvent removal under

reduced pressure, the product was isolated by flash chromatography

using an EtOAc/hexane (2/8, v/v) solution as eluant. R_f= 0.45, white

solid (380 mg, 71%). ¹H NMR (200 MHz, CDCl₃): δ 7.74 (d, J = 8.2

Hz, 1H), 7.46−7.30 (m, 3H), 3.74 (s, 3H), 3.26 (t, J = 6.3 Hz, 2H),

2.78 (t, J = 6.3 Hz, 2H), 2.53 (s, 3H).

Methyl 3-[2-(2-Methylphenyl)-1,3-dioxolan-2-yl]propanoate

(5). To a solution of ester 4 (380 mg, 1.84 mmol) in benzene (30

mL), in the presence of a catalytic amount of p-toluenesulfonic acid,

was added ethylene glycol (114 mg, 1.84 mmol). The resulting mixture

was refluxed for 2 days. Then the solvent was removed under reduced

pressure and the crude product was diluted in EtOAc, washed with a

brine solution, and dried over MgSO₄. The product was used without

(m, 3H), 3.30−3.27 (m, 2H), 3.19−3.13 (m, 4H), 3.09−2.88 (m, 4H),

2.45 (s, 3H), 2.40−2.37 (m, 2H), 2.12−2.10 (m, 2H), 1.60−1.34 (m,

6H), 1.29−1.14 (m, 5H). HRMS: calcd for C₂₁H₃₅N₂O 331.2744,

found 331.2742.

1

any further purification (460 mg, quantitative yield). H NMR (300

MHz, CDCl₃): δ 7.50 (d, J = 8.2 Hz, 1H), 7.39−7.13 (m, 3H), 4.00−

3.98 (m, 2H), 3.80−3.77 (m, 2H), 3.64 (s, 3H), 3.23 (t, J = 6.3 Hz,

2H), 2.75 (t, J = 6.3 Hz, 2H), 2.51 (s, 3H).

Para-LRB-AC42 (2a) and Ortho-LRB-AC42 (2b), Trifluoroace-

tate Salt. To a solution of lissamine rhodamine B sulfonyl chloride

(72 mg, 125 μmol) in CHCl₃(0.3 M) at 0 °C were added 4-[4-(5-

aminopentyl)piperidinyl]-1-o-tolylbutanone 1 (50 mg, 125 μmol) and

triethylamine (60 μL, 437 μmol). The mixture was stirred overnight at

room temperature, then concentrated under reduced pressure.

Isolation of both ortho and para isomers of lissamine rhodamine B

labeled AC-42 was performed by preparative RP-HPLC on a C18

Symmetry Shield column (19 mm × 300 mm) equilibrated in solvent

A (100% H₂O, 0.1% TFA, v/v). Elution proceeded with a linear

gradient from 20% to 70% of solvent B (100% CH₃CN, 0.1% TFA, v/

v) in solvent A in 40 min at a 10 mL·min⁻¹flow rate. Detection was

set at 220 and 254 nm. Fractions containing the products of interest

were pooled, concentrated, and further checked for purity by analytical

RP-HPLC and HRMS as indicated.

3-[2-(2-Methylphenyl)-1,3-dioxolan-2-yl]propanol (6). To a

solution of ester 5 (500 mg, 2 mmol) in ether (10 mL), stirred at 0 °C,

was added LiAlH₄(100 mg, 2.6 mmol). After 1 h of agitation at room

temperature, the solvent was removed under reduced pressure. The

crude oily product was diluted in EtOAc, washed with water, a brine

solution, and then dried over MgSO₄. Following removal of the

solvent under reduced pressure, the crude product was purified on

flash chromatography using an EtOAc/hexane mixture (5/5, v/v) to

give 6 as a yellow oil (205 mg, 48%). ¹H NMR (300 MHz, CDCl₃): δ

7.50 (d, J = 8.2 Hz, 1H), 7.20−7.14 (m, 3H), 4.06−4.03 (m, 2H),

4.02−3.75 (m, 2H), 3.64 (t, J = 6.2 Hz, 2H), 2.51 (s, 3H), 2.06 (t, J =

6.4 Hz, 2H), 1.95 (bs, 1H), 1.73−1.69 (m, 2H).

3-[2-(2-Methylphenyl)-1,3-dioxolan-2-propyl Methylsulfo-

nate (7). To a solution of alcohol 6 (210 mg, 0.95 mmol) in

CH₂Cl₂(10 mL) at 0 °C were added methanesulfonyl chloride (130

mg, 1.14 mmol) and then triethylamine (115 mg, 1.14 mmol). The

solution was allowed to warm to room temperature and was stirred for

Para-LRB-AC42 (2a, purple powder, 35 mg, 30%): ¹H NMR

(DMSO-d₆, 500 MHz) δ 8.44 (d, J = 1.9 Hz, 1H), 7.92 (dd, J = 7.9

and 1.9 Hz, 1H), 7.81 (t, J = 5.8 Hz, 1H), 7.75 (dd, J = 7.7 and 1.4 Hz,

1H), 7.47 (d, J = 7.9 Hz, 1H), 7.44 (td, J = 7.5 and 1.3 Hz, 1H), 7.34−

7.29 (m, 2H), 7.05−6.98 (m, 4H), 6.93 (d, J = 2.3 Hz, 2H), 3.70−3.60

2138

dx.doi.org/10.1021/jm201348t | J. Med. Chem. 2012, 55, 2125−2143

Journal of Medicinal Chemistry

Article

2 h. Then the mixture was washed with 10% NaHCO₃solution, brine

and dried over MgSO₄. The solvent was removed under reduced

pressure and the product was purified on silica gel column

chromatography using an EtOAc/hexane mixture (2/8, v/v) to

furnish 7 (240 mg, 84%). ¹H NMR (300 MHz, CDCl₃): δ 7.49 (d, J =

8.2 Hz, 1H), 7.27−7.14 (m, 3H), 4.28−4.23 (m, 2H), 4.03−3.99 (m,

2H), 3.78−3.73 (m, 2H), 2.98 (s, 3H), 2.48 (s, 3H), 2.08−2.03 (m,

2H), 1.93−1.84 (m, 2H). ES-MS m/z for C₁₄H₂₁O₅S (M + H)⁺, calcd

301.38, found 301.41.

4-(5-tert-bButoxycarbonylaminopentynyl)piperidine (8).

Compound 14 (995 mg, 2.50 mmol) diluted in ethanol (20 mL),

few drops of acetic acid, and 10% (w/w) PtO₂was stirred overnight

under 65 psi of hydrogen pressure. The solution was then filtered on a

Celite pad that was washed with ethanol. The filtrate was concentrated

under vacuo to provide 8 (660 mg, quantitative yield) that was used

without any further purification. ES-MS m/z for C₁₅H₃₀N₂O₂(M +

H)⁺, calcd 271.42, found 271.27 and 293.39 (M + Na)⁺.

heated at 40 °C overnight. The solvent was concentrated in vacuo to a

residue that was purified by flash chromatography (EtOAc/CH₂Cl₂, 5/

1

95, v/v) to give 17 (7.82 g, 72%). H NMR (200 MHz, CDCl₃): δ

4.37 (q, J = 6.9 Hz, 2H), 1.51 (s, 9H), 1.38 (t, J = 6.9 Hz, 3H).

4-Pentynylamine(oxo)ethyl Acetate Carboxylic Acid tert-

Butyl Ester (19). A solution of DEAD (3.32 g, 19 mmol) in THF (10

mL) was added dropwise to a solution of compound 17 (4.16 g, 19

mmol), 4-pentynol 18 (1.6 g, 19 mmol), and triphenylphosphine (4.99

g, 19 mmol) in THF (15 mL). The solution was stirred overnight at

room temperature and concentrated in vacuo to a residue that was

purified by flash chromatography (heptane/EtOAc, 9/1, v/v) to give

19 (4.57 g, 85%). ¹H NMR (200 MHz, CDCl₃): δ 4.33 (q, J = 6.9 Hz,

2H), 3.80−3.75 (m, 2H), 2.26−2.21 (m, 2H), 1.98−1.96 (m, 1H),

1.86−1.77 (m, 2H), 1.52 (s, 9H), 1.35 (t, J = 6.9 Hz, 3H).

Para-LRB-n-butyl (20). To a solution of lissamine rhodamine B

sulfonyl chloride (100 mg, 173 μmol) in CH₂Cl₂was added n-

butylamine (86 μL, 860 μmol) in the presence of triethylamine (119.5

μL, 860 μmol). The resulting solution was stirred at room temperature

overnight (Scheme 5). Following concentration under reduced

pressure, the para isomer (24.4 mg, 23%) was isolated by preparative

RP-HPLC on a C18 Symmetry Shield column (19 mm × 300 mm)

using a linear gradient of eluent B (5−100% B in 30 min, flow rate of

10 mL·min⁻¹) in eluent A (100% H₂O, 0.1% TFA, v/v). Fractions of

interest were pooled, concentrated, and further checked for purity by

analytical RP-HPLC on an Agilent Eclipse XDB-C18 column (4.6 mm

× 150 mm), using a linear gradient from 5% to 100% solvent B in 15

min (flow rate of 1 mL/min). RP-HPLC purity: >98%; t_R= 11.5 min.

¹H NMR (DMSO-d₆, 500 MHz): δ 8.41 (s, 1H), 7.92 (d, J = 7.5 Hz,

Benzyl-(4-oxopiperidinyl) Acetate (11). 4-Piperidone 10 (4 g,

26 mmol) was added to a solution of NaOH, 10%, and toluene cooled

down at 0 °C. After a few minutes, a solution of benzyl chloroformiate

(4.43 g, 26 mmol) in toluene was slowly added. The solution was

allowed to warm to room temperature and was stirred for 2 h. The

solvent was removed under reduced pressure, and the product was

diluted in chloroform and washed with water (2×), with brine (2×)

and dried over MgSO₄. The solvent was removed under reduced

pressure and the product was purified on silica gel column

1

chromatography with EtOAc/heptane to give 11 (5.88 g, 97%). H

NMR (300 MHz, CDCl₃): δ 7.36 (s, 5 H), 5.20 (s, 2H), 3.80 (t, J =

6.8 Hz, 4H), 2.46 (t, J = 6.8 Hz, 4H).

1H), 7.88 (t, J = 5.6 Hz, 1H), 7.47 (d, J = 8.0, 1H), 7.05−6.93 (m,

6H), 7.35−7.30 (m, 2H), 3.65−3.63 (m, 8H), 2.86 (q, J = 6.3 Hz,

2H), 1.48−1.39 (m, 2H), 1.32−1.19 (m, 14H), 0.85 (t, J = 7.2 Hz,

3H). ¹³C NMR (DMSO-d₆, 100 MHz): δ 155.9, 156.5, 154.4, 147.5,

141.1, 132.3, 132.0, 130.0, 125.9, 125.1, 113.0, 94.8, 44.6, 41.7, 30.5,

18.6, 12.9, 11.9. ES-MS m/z for C₃₁H₃₉N₃O₆S₂(M + H)⁺, calcd

614.23, found 614.20.

Benzyl-(4-{[(trifluoromethyl)sulfonyl]oxy}piperidinyl) Ace-

tate (12). To a −78 °C cooled solution of N-Cbz-piperidone 11 (1

g, 5 mmol) was added LDA (2.75 mL, 5.5 mmol). Following agitation

for 20 min, N-phenyltrifluoromethanesulfonimide (1.87 g, 5.25 mmol)

was slowly added. The resulting mixture was allowed to warm to room

temperature and was stirred for 3 h. The solvent was removed under

reduced pressure and the product was purified on silica gel column

chromatography using an EtOAc/heptane mixture (1/9, v/v) to

Construction of EGFP-Fused Muscarinic M1 Receptors and

Cell Expression. Fluorescent chimera of the human M1 receptor was

essentially as previously reported.²⁷EGFP was fused, via an identical

short linker, to the full-length (EGFP-wthM1) or truncated by 11

(EGFP(Δ17)hM1) or 17 residues (EGFP(Δ17)hM1) receptor N-

terminus. The exchange of the long linker (referred to L in the

previously described EGFP_L(Δ11)hM1 chimera) for a short one was

done through ligation, into a Bluescript KS+ vector, of the NotI-BglII

(signal peptide-EGFP sequence-short linker) and BglII-XhoI (hM1

sequence with a truncated N tail taken from the EGFP_L(Δ11)hM1

construct) fragments, as earlier defined.²⁷The final construct (NotI-

XhoI insert) was subcloned into the pCep4 vector for expression into

eukaryotic cells and sequenced before use. Transfection of HEK 293

cells by calcium phosphate precipitation and hygromycin-B selection

proceeded as previously detailed.²⁷Cells were grown in minimal

essential medium supplemented with 10% fetal calf serum, 2 mM

glutamine, penicillin (100 U/mL), and streptomycin (100 μg/mL)

and incubated at 37 °C in a 5% CO₂humidified atmosphere.

Functional Calcium Responses. Agonist-evoked increases in

intracellular calcium were monitored in HEK cells, with stable

expression of the various receptor constructs. Adherent cells were

loaded with 5 μM Indo-1 AM for 45 min at 37 °C, harvested by rapid

0.05% trypsin/0.02% EDTA treatment (wt/v), washed, and finally

suspended (10⁶cells/mL) in Hepes buffer (10 mM Hepes, 137.5 mM

NaCl, 1.25 mM MgCl₂, 1.25 mM CaCl₂, 6 mM KCl, 10 mM glucose, 1

mg/mL bovine serum albumin; pH 7.4).

1

furnish 12 (1.02 g, 56%). H NMR (300 MHz, CDCl₃): δ 7.38−7.30

(m, 5H), 5.78 (m, 1H), 5.17 (s, 2H), 4.14−4.12 (m, 2H), 3.72 (t, J =

6.8 Hz, 2H), 2.47 (m, 2H).

4-Pentynylaminecarboxylic Acid tert-Butyl Ester (13). A

solution of LiOH (1.15 g, 48 mmol) in H₂O (12 mL) was added to

a solution of compound 19 (4.54 g, 16 mmol) in THF (36 mL) and

stirred at room temperature for three hours. The solvent was removed

under reduced pressure, and the residue was diluted in CH₂Cl₂(40

mL) and washed with 10% NaHCO₃solution (2 × 40 mL) and brine.

The organic layer was dried over Na₂SO₄, filtered, and concentrated in

1

vacuo to give compound 13 (2.75 g, 94%). H NMR (200 MHz,

CDCl₃): δ 4.66 (brs, 1H), 3.30−3.20 (m, 2H), 2.30−2.20 (m, 2H),

2.00−1.96 (m, 1H), 1.80−1.66 (m, 2H), 1.45 (s, 9H).

4-(5-tert-Butoxycarbonylaminopentynyl)-3,6-dihydro-2H-

pyridinbenzyloxycarbonyl (14). Compound 12 (1 g, 2.9 mmol)

was dissolved in THF (20 mL) in the presence of alkyne 13 (795 mg,

4.35 mmol), CuI (60 mg, 0.3 mmol), PdCl₂(PPh₃)₂(210 mg, 0.3

mmol), and triethylamine (460 mg, 4.35 mmol) at room temperature.

After 6 h at room temperature, the mixture was concentrated in vacuo

to a residue. Compound 14 (940 mg, 81%) was isolated by flash

chromatography using an EtOAc/heptane mixture (v/v, 1/9). ¹H

NMR (300 MHz, CDCl₃): δ 7.38−7.34 (m, 5H), 5.92 (m, 1H), 5.15

(s, 2H), 4.69 (bs, 1H), 4.03−3.99 (m, 2H), 3.59−3.54 (m, 2H), 3.24−

3.22 (m, 2H), 2.39−2.35 (m, 2H), 2.32−2.30 (m, 2H), 1.75−1.72 (m,

2H), 1.44 (s, 9H).

Calcium signals were recorded over time as an increase in

fluorescence emission at 400 nm, with excitation set at 338 nm.

Inhibitory compounds were preincubated with cells for 5 min before

carbachol challenge. After baseline correction, peak amplitudes were

normalized to maximal fluorescence levels determined in the presence

of 20 μM digitonine. Data for M1-transfected cells are expressed, as

appropriate, as a percentage of the maximal response to 10 μM

carbachol or of a control response to 0.5 μM carbachol obtained in the

absence of inhibitor. Unless otherwise stated, calcium signals were

Ethyl Oxoisocynate Acetate (16). The oxalyl chloride (66.7

mmol) was added to a suspension of ethyl oxamate 15 (6.5 g, 55.6

mmol) in CH₂Cl₂(80 mL) and refluxed overnight. The solvent was

reduced in vacuo to isocyanate 16 that was used without any further

purification (7.15 g, 90%).

tert-Butyl(oxo)ethyl Acetate Carbamate (17). A solution of

tert-butanol (3.75 g, 51 mmol) in toluene (20 mL) was added to a

solution of isocyanate 16 (7.17 g, 50 mmol) in toluene (50 mL) and

2139

dx.doi.org/10.1021/jm201348t | J. Med. Chem. 2012, 55, 2125−2143

Journal of Medicinal Chemistry

Article

recorded at 20 °C from M1-overexpressing cells and at 37 °C when

using nontransfected HEK cells.

Equilibrium binding assays consisted of incubating cells with

increasing (saturation experiments) or fixed (competition experi-

ments) concentrations of a fluorescent tracer, in the absence or the

presence of various concentrations of unlabeled drugs. Incubation

proceeded for 4 h at 20 °C (unless otherwise stated) in test tubes (1

mL final volume) before transfer of the medium into the quartz

cuvette for fluorescent readouts of EGFP emission.

Time-based monitoring of tracer binding consisted of continuous

fluorescence recordings of EGFP(Δ17)hM1 cells before and following

the addition of the fluorescent probe (4 μL of a 250-fold concentrated

DMSO stock). Once equilibrium was reached, tracer dissociation was

initiated by adding atropine (20 μM), alone or in the presence of other

compounds, to the cell suspension and fluorescence was recorded over

time until full recovery of EGFP emission.

Measurements of Receptor Expression Levels. HEK cells

stably expressing the various hM1 constructs were checked for

receptor density through quantification of specific binding at a single

saturating concentration of [³H]QNB (0.3 nM, 2 h at 37 °C) or

[³H]NMS (0.75 nM, 2 h at 20 °C) or determination of B_maxvalues

from [³H]NMS (0.02−1.2 nM) saturation experiments conducted on

EGFP(Δ17)hM1 cells. K_dvalues for [³H]QNB and [³H]NMS binding

to hM1 receptors are 50 ²⁷and 95 pM, respectively.

Alkylation of Muscarinic Receptors by Propylbenzilylcho-

line Mustard. A 50 μM solution of PrBCM in Hepes buffer (deprived

of BSA) was incubated at 37 °C for 30 min to allow cyclization of the

compound into the aziridinium ion. Thereafter, activated PrBCM was

added to EGFP(Δ17)hM1 cells (2 × 10⁶cells/mL) to a final 30 nM

concentration and allowed to incubate for 40 min at 37 °C. Buffer

without PrBCM was added to control cells. After a 3-fold dilution of

the medium with Hepes buffer at 20 °C, centrifugation at 1000 rpm

for 5 min, and supernatant removal, cell pellets were washed by

resuspension in 200 vol of buffer and centrifuged again. Final cell

suspensions, at an appropriate density, were kept on ice until further

use.

The FRET signal associated with specific binding of a fluorescent

probe to M1 receptors was defined as the difference in fluorescence

levels of cells incubated with the tracer alone or together with 20 μM

atropine.

Molecular Modeling. Sequence alignment and homology

modeling were performed using MOE 2009.10 with default settings

(Chemical Computing Group Inc., Montreal, Canada). Constraints

were manually set to guide the sequence alignment in order to avoid

the introduction of gap in the TM regions and to fix the disulfide

bridge between Cys98 in TM3 and Cys178 in ECL2 (sequence

alignment is provided in Supporting Information Figure S3). A first

model of the full M1 receptor (residues Gly29-Cys435) was obtained

by homology to the dopamine D3 template (PDB entry 3PBL, chain

A). This model was edited using SYBYL-X, version 1.3 (Tripos, Inc.,

St. Louis, MO, U.S.), to replace the ECL2 loop by fragments extracted

from the structures of the human β2 adrenergic (ADRB2) and

adenosine 2A (AA2A) receptors. More precisely, the Leu162-Gln177

segment of the M1 model was replaced by the Gln170-Cys190

segment of the A chain of the 2RH1 PDB file and the Cys178-Pro187

segment of the M1 model was replaced by the Cys166-Asn175

segment of the A chain of 3EML PDB file. The chimeric structure

constituted the template for the second round of M1 homology

modeling using MOE 2009.10. The third intracellular loop (ICL3) of

the M1 was removed using SYBYL-X, version 1.3, and the termini

were capped with ACE and NME residues. The M1 model was energy

minimized using AMBER9 (University of California, San Francisco,

CA, USA) (Supporting Information Figure S2).

[³H]NMS Binding Assays. Nearly confluent cells were harvested

using a Versene solution (PBS + 5 mM EDTA), washed with Hepes

buffer, and pelleted by centrifugation at 1000 rpm for 5 min before

suspension at an appropriate dilution. [³H]NMS binding was carried

out in Hepes buffer at 20 °C, using whole cells stably expressing the

EGFP(Δ17)hM1 receptor. Separation of bound from free radioactivity

was achieved by filtration through Whatman GF/B glass-fiber filters

presoaked with 0.2% polyethyleneimine. Tubes and filters were rinsed

three times with ice-cold 25 mM Tris-HCl buffer (pH 7.4) before

determination of bound radioactivity by liquid scintillation counting.

Nonspecific binding was determined in the presence of 10 μM

atropine.

Competition-type experiments were performed on EGFP(Δ17)-

hM1 cells (typically 60 000 per assay) with 0.1 nM [³H]NMS, unless

otherwise stated, and various concentrations of test compounds.

Incubation lasted for 22 h at 20 °C in a 1 mL final volume. Under such

conditions, total bound [³H]NMS did not exceed 10% of total added

radioligand and nonspecific binding was low (less than 5% of total

binding). When adding drugs diluted in DMSO, we verified that the

highest solvent concentration in the incubation medium (1%, v/v) did

not affect tracer binding. Data are expressed as B/B₀ratios, where B₀

(set as 1) and B refer to specific [³H]NMS binding in the absence and

the presence of competitors, respectively.

Dissociation-type experiments were performed at 25 °C using

EGFP(Δ17)hM1 cells, according to a previously described protocol.⁵⁰

A 100-fold concentrated cell suspension (60 000 cells per 10 μL) was

preincubated for 30 min with 3 nM [³H]NMS in the absence (total

binding) or the presence (nonspecific binding) of 10 μM atropine,

then kept on ice. The 10 μL aliquots of both prelabeled cell batches

were added to a series of test tubes filled with 1 mL Hepes buffer (at

25 °C) containing 10 μM atropine alone or combined with a test

compound at various concentrations. At given dissociation time points

(usually set between 1 and 3 [³H]NMS dissociation half-lives), the

reaction was stopped by rapid filtration, allowing the quantification of

residual specific [³H]NMS binding (B values). Specific [³H]NMS

binding a time 0 (B₀) was determined in parallel through direct

filtration of extemporaneously diluted (in ice-cold buffer) prelabeled

cell samples (10 μL). All off-rate experiments, presented as B/B₀

versus dissociation time, fitted a monoexponential time course with a

characteristic k_offrate constant.

The 3D structures of AC-42 and para-LBR-AC42 were generated

using Corina, version 3.1 (Molecular Network GmbH, Erlangen,

Germany), from a 2D sketch drawn using Marvinsketch (Marvin-

Beans, version 5.3.01, ChemAxon, Budapest, Hungary). Ternary amine

groups were assigned a positive charge, sulfate groups a negative

charge. The flexible alignment of molecules was performed using

SurflexSim, version 2.6 (BioPharmics LCC, San Mateo, CA, U.S.).

ASSOCIATED CONTENT

■

S

* Supporting Information

Figure S1 showing normalized absorption and emission spectra

of EGFP and para-LRB-AC42 in Hepes buffer; Figure S2

showing structural variation of the M1 receptor; Figure S3

showing sequence alignment for M1 homology modeling;

procedures for solubility measurements, estimation of donor−

acceptor distances through FRET, and biological data analyses.

This material is available free of charge via the Internet at

http://pubs.acs.org.

Fluorescence Resonance Energy Transfer (FRET) Based

Binding Assays. Data were collected from cell suspensions (10⁶

cells/mL) placed in a 1 mL quartz cuvette (with magnetic stirring)

maintained at 20 °C, using a Fluorolog 2 spectrofluorimeter (Horiba

Jobin-Yvon). Specific binding of fluorescent tracers to EGFP(Δ17)-

hM1 expressing cells was monitored, in real-time or at equilibrium,

through atropine-dependent variations in EGFP emission at 510 nm

(excitation set at 470 nm).

AUTHOR INFORMATION

■

Corresponding Author

*Phone: +33 368 85 47 38. Fax: +33 368 85 48 29. E-mail:

brigitte.ilien@unistra.fr.

Present Addresses

^§Quintiles, Parc d’Innovation, 67404 Illkirch, France.

2140

dx.doi.org/10.1021/jm201348t | J. Med. Chem. 2012, 55, 2125−2143

Journal of Medicinal Chemistry

Article

^∥Drug Discovery Biology Laboratory, Department of Pharma-

cology, Monash University, Clayton, Victoria 3800, Australia.

Scolnick, E. M.; Conn, P. J. N-Desmethylclozapine, an allosteric

agonist at muscarinic M1 receptor, potentiates N-methyl-^D-aspartate

receptor activity. Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 13674−13679.

(10) Spalding, T. A.; Trotter, C.; Skjaerbaek, N.; Messier, T. L.;

Currier, E. A.; Burstein, E. S.; Li, D.; Hacksell, U.; Brann, M. R.

Discovery of an ectopic activation site on the M1 muscarinic receptor.

Mol. Pharmacol. 2002, 61, 1297−1302.

(11) Langmead, C. J.; Fry, V. A. H.; Forbes, I. T.; Branch, C. L.;

Christopoulos, A.; Wood, M. D.; Herdon, H. J. Probing the molecular

mechanism of interaction between 4-n-butyl-1-[4-(2-methylphenyl)-4-

oxo-1-butyl]-piperidine (AC-42) and the muscarinic M1 receptor:

direct pharmacological evidence that AC-42 is an allosteric agonist.

Mol. Pharmacol. 2006, 69, 236−246.

(12) Spalding, T. A.; Ma, J. N.; Ott, T. R.; Friberg, M.; Bajpai, A.;

Risso Bradley, S.; Davis, R. E.; Brann, M. R.; Burstein, E. S. Structural

requirements of transmembrane domain 3 for activation by the M1

muscarinic receptor agonists AC-42, AC-260584, clozapine, and N-

desmethylclozapine: evidence for three distinct modes of receptor

activation. Mol. Pharmacol. 2006, 70, 1974−1983.

(13) Langmead, C. J.; Austin, N. E.; Branch, C. L.; Brown, J. T.;

Buchanan, K. A.; Davies, C. H.; Forbes, I. T.; Fry, V. A. H.; Hagan, J.

J.; Herdon, H. J.; Jones, G. A.; Jeggo, R.; Kew, J. N. C.; Mazzali, A.;

Melarange, R.; et al. Characterization of a CNS penetrant, selective M1

muscarinic receptor agonist, 77-LH-28-1. Br. J. Pharmacol. 2008, 154,

1104−1115.

Notes

The authors declare no competing financial interest.

ACKNOWLEDGMENTS

■

We thank Dr. T. Spalding, ACADIA Pharmaceuticals Inc. (San

Diego, CA, U.S.) and Pr. A. Christopoulos (Monash University,

Victoria, Australia) for the generous gift of AC-42. Dr. N.

Birdsall (NIMR, London, U.K.) kindly provided us with

propylbenzilylcholine mustard. We are grateful to Patrick

Wehrung and Cyril Antheaume for NMR, HR-MS, and HPLC

analyses and to Patrick Gizzi (TechMed platform; UMR 7242

CNRS, Illkirch, France) for solubility measurements. The

skillful technical assistance of Step

Utard was greatly appreciated.

This work was supported in part by CNRS, INSERM, and

the Universite de Strasbourg.

́ ́ ́

hanie Riche and Valerie

́

S.B.D. was a recipient of a grant from the Association

Nationale de la Recherche Technique (Convention CIFRE

564/2005 with Prestwick Chemical).

C.V. was supported by a fellowship from the Minister

̀

e de

l'Education Nationale, de l'Enseignement Superieur et de la

́

(14) Jones, C. K.; Brady, A. E.; Davis, A. A.; Xiang, Z.; Bubser, M.;

Tantawy, M. N.; Kane, A. S.; Bridges, T. M.; Kennedy, J. P.; Bradley, S.

R.; Peterson, T. E.; Ansari, M. S.; Baldwin, R. M.; et al. Novel selective

allosteric activator of the M1 muscarinic acetylcholine receptor

regulates amyloid processing and produces antipsychotic-like activity

in rats. J. Neurosci. 2008, 28, 10422−10433.

Recherche (MENSER).

ABBREVIATIONS USED

■

AC-42, 4-n-butyl-1-[4-(2-methylphenyl)-4-oxo-1-butyl]-

piperidine hydrogen chloride; BSA, bovine serum albumin;

ECL, extracellular loop; EGFP, enhanced green fluorescent

protein; FRET, fluorescence resonance energy transfer; HEK,

human embryonic kidney; hM1, human M1 muscarinic

receptor subtype; LRB, lissamine rhodamine B; mAChR,

muscarinic cholinergic receptor; NDMC, N-desmethylcloza-

pine; NMS, N-methylscopolamine; PrBCM, propylbenzilylcho-

line mustard; QNB, quinuclidinyl benzilate; TM, trans-

membrane domain

(15) Sams, A. G.; Hentzer, M.; Mikkelsen, G. K.; Larsen, K.;

Bundgaard, C.; Plath, N.; Christofferson, C. T.; Bang-Andersen, B.

Discovery of N-{1-[3-(3-oxo-2,3-dihydrobenzo[1,4]oxazin-4-yl)-

propyl]piperidin-4-yl}-2-phenylacetamide (Lu AE51090): an allosteric

muscarinic M1 receptor agonist with unprecedented selectivity and

precognitive potential. J. Med. Chem. 2010, 53, 6386−6397.

(16) Lebois, E. P.; Bridges, T. M.; Lewis, L. M.; Dawson, E. S.; Kane,

A. S.; Xiang, Z.; Jadhav, S.; Yin, H.; Kennedy, J. P.; Meiler, J.;

Niswender, C. M.; Jones, C. K.; Conn, P. J.; Weaver, C. D.; Lindsley,

C. W. Discovery and characterization of novel subtype-selective

allosteric agonists for the investigation of M1 receptor function in the

central nervous system. ACS Chem. Neurosci. 2010, 1, 104−121.

(17) Bradley, S. R.; Lameh, J.; Ohrmund, L.; Son, T.; Bajpai, A.;

Nguyen, D.; Friberg, M.; Burstein, E. S.; Spalding, T. A.; Ott, T. R.;

Schiffer, H. H.; Tabatabaei, A.; McFarland, K.; Davis, R. E.; Bonhaus,

D. W. AC-260584, an orally bioavailable M1 muscarinic receptor

allosteric agonist, improves cognitive performance in an animal model.

Neuropharmacology 2010, 58, 365−373.

(18) Buchanan, K. A.; Petrovic, M. M.; Chamberlain, S. E. L.;

Marrion, N. V.; Mellor, J. R. Facilitation of long-term potentiation by

muscarinic M1 receptors is mediated by inhibition of SK channels.

Neuron 2010, 68, 948−963.

(19) Thomas, R. L.; Mistry, R.; Langmead, C. J.; Wood, M. D.;

Challiss, R. A. J. G protein coupling and signalling pathway activation

by M1 muscarinic acetylcholine receptor orthosteric and allosteric

agonists. J. Pharmacol. Exp. Ther. 2008, 327, 365−374.

REFERENCES

■

(1) Eglen, R. M. Muscarinic receptor subtype pharmacology and

physiology. Prog. Med. Chem. 2005, 43, 105−136.

(2) Wess, J.; Eglen, R. M.; Gautam, D. Muscarinic acetylcholine

receptors: mutant mice provide new insights for drug development.

Nat. Rev. Drug Discovery 2007, 6, 721−733.

(3) Digby, G. J.; Shirey, J. K.; Conn, P. J. Allosteric activators of

muscarinic receptors as novel approaches for treatment of CNS

disorders. Mol. BioSyst. 2010, 6, 1345−1354.

(4) Birdsall, N. J. M.; Lazareno, S. Allosterism at muscarinic

receptors: ligands and mechanisms. Mini-Rev. Med. Chem. 2005, 33,

523−543.

(5) De Amici, M.; Dallanoce, C.; Holzgrabe, U.; Trankle, C.; Mohr,

̈

K. Allosteric ligands for G protein-coupled receptors: a novel strategy

with attractive therapeutic opportunities. Med. Res. Rev. 2010, 30,

463−549.

(20) Thomas, R. L.; Langmead, C. J.; Wood, M. D.; Challiss, R. A. J.

Contrasting effects of allosteric and orthosteric agonists on M1

muscarinic acetylcholine receptor internalization and down-regulation.

J. Pharmacol. Exp. Ther. 2009, 331, 1086−1095.

(6) May, L. T.; Leach, K.; Sexton, P. M.; Christopoulos, A. Allosteric

modulation of G protein-coupled receptors. Annu. Rev. Pharmacol.

Toxicol. 2007, 47, 14.1−14.51.

(7) Maillet, E. L.; Pellegrini, N.; Valant, C.; Bucher, B.; Hibert, M.;

Bourguignon, J.-J.; Galzi, J.-L. A novel, conformation-specific allosteric

inhibitor of the tachykinin NK2 receptor (NK2R) with functionally

selective properties. FASEB J. 2007, 21, 2124−2134.

(8) Ehlert, F. J. Estimation of the affinities of allosteric ligands using

radioligand binding and pharmacological null methods. Mol.

Pharmacol. 1988, 35, 187−194.

(9) Sur, C.; Mallorga, P. J.; Wittmann, M.; Jacobson, M. A.;

Pascarella, D.; Williams, J. B.; Brandish, P. E.; Pettibone, D. J.;

(21) Davis, A. A.; Heilman, C. J.; Brady, A. E.; Miller, N. R.;

Fuerstenau-Sharp, M.; Hanson, B. J.; Lindsley, C. W.; Conn, P. J.; Lah,

J. J.; Levey, A. I. Differential effects of allosteric M1 muscarinic

acetylcholine receptor agonists on receptor activation, arrestin 3

recruitment, and receptor downregulation. ACS Chem. Neurosci. 2010,

1, 542−551.

(22) Ballesteros, J. A.; Weinstein, H. Integrated methods for the

construction of three dimensional models and computational probing

2141

dx.doi.org/10.1021/jm201348t | J. Med. Chem. 2012, 55, 2125−2143

Journal of Medicinal Chemistry

Article

of structure−function relationships in G-protein coupled receptors.

Methods Neurosci. 1995, 25, 366−428.

(23) Lebon, G.; Langmead, C. J.; Tehan, B. G.; Hulme, E. C.

Mutagenic mapping suggests a novel binding mode for selective

agonists of M1 muscarinic acetylcholine receptors. Mol. Pharmacol.

2009, 75, 331−341.

(24) Avlani, V. A.; Langmead, C. J.; Guida, E.; Wood, M. D.; Tehan,

B. G.; Herdon, H. J.; Watson, J. M.; Sexton, P. M.; Christopoulos, A.

Orthosteric and allosteric modes of interaction of novel selective

agonists of the M1 muscarinic acetylcholine receptor. Mol. Pharmacol.

2010, 78, 94−104.

(25) Gregory, K. J.; Hall, N. E.; Tobin, A. B.; Sexton, P. M.;

Christopoulos, A. Identification of orthosteric and allosteric site

mutations in M2 muscarinic acetylcholine receptors that contribute to

ligand-selective signalling bias. J. Biol. Chem. 2010, 285, 7459−7474.

(26) Jacobson, M. A.; Kreatsoulas, C.; Pascarella, D. M.; O’Brien, J.

A.; Sur, C. The M1 muscarinic receptor allosteric agonists AC-42 and

1-[1′-(2-methylbenzyl)-1,4′-bipiperidin-4-yl]-1,3-dihydro-2H-benzimi-

dazol-2-one bind to a unique site distinct from the acetylcholine

orthosteric site. Mol. Pharmacol. 2010, 78, 648−657.

(27) Ilien, B.; Franchet, C.; Bernard, P.; Morisset, S.; Weill, C. O.;

Bourguignon, J.-J.; Hibert, M.; Galzi, J.-L. Fluorescence resonance

energy transfer to probe human M1 muscarinic receptor structure and

drug binding properties. J. Neurochem. 2003, 85, 768−778.

(28) Ilien, B.; Glasser, N.; Clamme, J.-P.; Didier, P.; Piemont, E.;

Chinnappan, R.; Daval, S. B.; Galzi, J.-L.; Mely, Y. Pirenzepine

promotes the dimerization of muscarinic M1 receptors through a

three-step binding process. J. Biol. Chem. 2009, 284, 19533−19543.

(29) Vanover, K. E.; Veinbergs, I.; Davis, R. E. Antipsychotic-like

behavioural effects and cognitive enhancement by a potent and

selective muscarinic M1 receptor agonist, AC-260584. Behav. Neurosci.

2008, 122, 570−575.

(30) Astles, P. C.; Brealey, C.; Brown, T. J.; Facchini, V.;

Handscombe, C.; Harris, N. V.; McCarthy, C.; McLay, I. M.; Porter,

B.; Roach, A. G.; Sargent, C.; Smith, C.; Walsh, R. J. A. Selective

endothelin A receptor antagonists: discovery and structure−activity

relationships of a series of 4-phenoxybutanoic acid derivatives. J. Med.

Chem. 1998, 41, 2732−2744.

(31) Ward, H. R.; Sherman, P. D. Carbonyl participation in the

solvolysis of ketone derivatives. The observation and isolation of

intermediates. J. Am. Chem. Soc. 1968, 90, 3812−3817.

(40) Lazareno, S.; Birdsall, N. J. M. 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.

1995, 48, 362−378.

(41) Jakubik, J.; Bacakova, L.; El-Fakahany, E. E.; Tucek, S. Positive

cooperativity of acetylcholine and other agonists with allosteric ligands

on muscarinic acetylcholine receptors. Mol. Pharmacol. 1997, 52, 172−

179.

(42) Matsui, H. S.; Lazareno, S.; Birdsall, N. J. M. Probing of the

location of the allosteric site on m1 muscarinic receptors by site-

directed mutagenesis. Mol. Pharmacol. 1995, 47, 88−98.

(43) Lazareno, S.; Popham, A.; Birdsall, N. J. M. Allosteric

interactions of staurosporine and other indolocarbazoles with N-

[methyl-³H]scopolamine and acetylcholine at muscarinic receptor

subtypes: identification of a second allosteric site. Mol. Pharmacol.

2000, 58, 194−207.

(44) Gharagozloo, P.; Lazareno, S.; Popham, A.; Birdsall, N. J. M.

Allosteric interactions of quaternary strychnine and brucine derivatives

with muscarinic acetylcholine receptors. J. Med. Chem. 1999, 42, 438−

445.

́

(45) Fruchart-Gaillard, C.; Mourier, G.; Marquer, C.; Menez, A.;

Servent, D. Identification of various allosteric interaction sites on M1

muscarinic receptor using ¹²⁵I-Met35-oxidized muscarinic toxin 7. Mol.

Pharmacol. 2006, 69, 1641−1651.

(46) Kostenis, E.; Mohr, K. Two-point kinetic experiments to

quantify allosteric effects on radioligand dissociation. Trends

Pharmacol. Sci. 1996, 17, 280−283.

(47) Kurtenbach, E.; Curtis, C. A. M.; Pedder, E. K.; Aitken, A.;

Harris, A. C. M.; Hulme, E. C. Muscarinic acetylcholine receptors.

Peptide sequencing identifies residues involved in antagonist binding

and disulfide bond formation. J. Biol. Chem. 1990, 265, 13702−13708.

(48) Congreve, M.; Langmead, C. J.; Mason, J. S.; Marshall, F. H.

Progress in structure based drug design for G protein-coupled

receptors. J. Med. Chem. 2011, 54, 4283−4311.

(49) Forster, T. Zwischenmolekulare energiewanderung und

̈

fluoreszenz. Ann. Phys. (Leipzig) 1948, 2, 55−75.

(50) Lazareno, S.; Popham, A.; Birdsall, N. J. M. Analogs of WIN

62,577 define a second allosteric site on muscarinic receptors. Mol.

Pharmacol. 2002, 62, 1492−1505.

(51) Hulme, E. C.; Lu, Z. L.; Saldanha, J. W.; Bee, M. S. Structure

and activation of muscarinic acetylcholine receptors. Biochem. Soc.

Trans. 2003, 31, 29−34.

(32) Wustrow, D. J.; Wise, L. D. Coupling of arylboronic acids with a

partially reduced pyridine derivative. Synthesis 1991, 11, 993−995.

(33) Scott, W. J.; Pena, M. R.; Sward, K.; Stoessel, S. J.; Stille, J. K.

Palladium-catalyzed olefination of vinyl triflates. J. Org. Chem. 1985,

50, 2302−2308.

(52) Peng, J. Y.; Vaidehi, N.; Hall, S. E.; Goddard, W. A The

predicted 3D structure of the human M1 muscarinic acetylcholine

receptor with agonist and antagonist bound. ChemMedChem 2006, 1,

878−890.

(34) Crisp, G. T.; Jiang, Y.-L.; Pullman, P. J.; De Savi, C. Elaboration

of the side-chain of amino acid derivatives by palladium catalysed

couplings. Tetrahedron 1997, 53, 17489−17500.

(53) Jakubik, J.; El-Fakahany, E. E.; Tucek, S. Evidence for a tandem

two-site model of ligand binding to muscarinic acetylcholine receptors.

J. Biol. Chem. 2000, 275, 18836−18844.

́

(35) Berree, F.; Michelot, G.; Le Corre, M. N-Boc ethyl oxamate: a

new nitrogen nucleophile for use in Mitsunobu reactions. Tetrahedron

Lett. 1998, 39, 8275−8276.

(54) Gregory, K. J.; Sexton, P. M.; Christopoulos, A. Allosteric

modulation of muscarinic acetylcholine receptors. Curr. Neuro-

pharmacol. 2007, 5, 157−167.

(36) Bonnet, D.; Riche, S.; Loison, S.; Dagher, R.; Frantz, M. C.;

Boudier, L.; Rahmeh, R.; Mouillac, B.; Haiech, J.; Hibert, M. Solid-

phase organic tagging resins for labeling biomolecules by 1,3-dipolar

cycloaddition: application to the synthesis of a fluorescent non-

peptidic vasopressin receptor ligand. Chem.Eur. J. 2008, 14, 6247−

6254.

(55) Trankle, C.; Weyand, O.; Voiglander, U.; Mynett, A.; Lazareno,

̈

S.; Birdsall, N. J. M.; Mohr, K. Interactions of orthosteric and allosteric

ligands with [³H]dimethyl-W84 at the common allosteric site of

muscarinic M2 receptors. Mol. Pharmacol. 2003, 64, 180−190.

́

(56) Palanche, T.; Ilien, B.; Zoffmann, S.; Reck, M.-P.; Bucher, B.;

(37) Corrie, J. E. T.; Davis, C. T.; Eccleston, J. F. Chemistry of

sulforhodamine−amine conjugates. Bioconjugate Chem. 2001, 12, 186−

194.

Edelstein, S.; Galzi, J.-L. The neurokinin A receptor activates calcium

and cAMP responses through distinct conformational states. J. Biol.

Chem. 2001, 276, 34853−34861.

(38) Tahtaoui, C.; Parrot, I.; Klotz, P.; Guillier, F.; Galzi, J.-L.; Hibert,

M.; Ilien, B. Fluorescent pirenzepine derivatives as potential bitopic

ligands of the human M1 muscarinic receptor. J. Med. Chem. 2004, 47,

4300−4315.

(39) Weill, C.; Galzi, J. L.; Chasserot-Golaz, S.; Goeldner, M.; Ilien,

B. Functional characterization and potential applications for enhanced

green fluorescent protein- and epitope-fused human M1 muscarinic

receptors. J. Neurochem. 1999, 73, 791−801.

(57) Christopoulos, A.; Sorman, J. L.; Mitchelson, F.; El-Fakahany, E.

E. Characterization of the subtype selectivity of the allosteric

modulator heptane-1,7-bis-(dimethyl-3′-phtalimidopropyl) ammonium

bromide (C₇/3-phth) at cloned muscarinic acetylcholine receptors.

Biochem. Pharmacol. 1999, 57, 171−179.

(58) Valant, C.; Gregory, K. J.; Hall, N. E.; Scammels, P. J.; Lew, M.

J.; Sexton, P. M.; Christopoulos, A. A novel mechanism of G protein-

coupled receptor functional selectivity. Muscarinic partial agonist

2142

dx.doi.org/10.1021/jm201348t | J. Med. Chem. 2012, 55, 2125−2143

Journal of Medicinal Chemistry

Article

McN-A-343 as a bitopic orthosteric/allosteric ligand. J. Biol. Chem.

2008, 283, 29312−29321.

(59) Narlawar, R.; Lane, J. R.; Doddareddy, M.; Lin, J.; Brussee, J.;

Ijzerman, A. P. Hybrid ortho/allosteric ligands for the adenosine A1

receptor. J. Med. Chem. 2010, 53, 3028−3037.

(60) Antony, J.; Kellershohn, K.; Mohr-Andra, M.; Kebig, A.; Prilla,

̈

S.; Muth, M.; Heller, E.; Disingrini, T.; Dallanoce, C.; Bertoni, S.;

Schrobang, J.; Trankle, C.; Kostenis, E.; Christopoulos, A.; Holtje, H.

̈

D.; Barocelli, E.; De Amici, M.; Holzgrabe, U.; Mohr, K. Dualsteric

GPCR targeting: a novel route to binding and signaling pathway

selectivity. FASEB J. 2009, 23, 442−450.

(61) Valant, C.; Sexton, P. M.; Christopoulos, A. Orthosteric/

allosteric bitopic hybrids. Going hybrid at GPCRs. Mol. Interventions

2009, 9, 125−135.

(62) Molecular Probes. http://www.probes.com/handbook/.

2143

dx.doi.org/10.1021/jm201348t | J. Med. Chem. 2012, 55, 2125−2143

Article Doi

Fluorescent derivatives of AC-42 to probe bitopic orthosteric/allosteric binding mechanisms on muscarinic M1 receptors

DOI: 10.1021/jm201348t

Source and publish data:

Authors:

Article abstract of DOI:10.1021/jm201348t

Full text of DOI:10.1021/jm201348t

Products guided by the article

R&D Labs maybe for 1197279-47-2

Relevant to this article

Hot Product