Whitehurst etMal.
ARTICLE
10.1177/1087057105284340
Muscarinic
Acetylcholine Receptor Ligands
2
Discovery and Characterization of Orthosteric and
Allosteric Muscarinic M2 Acetylcholine Receptor Ligands
by Affinity Selection–Mass Spectrometry
1
1
1
1
CHARLES E. WHITEHURST, NAIM NAZEF, D. ALLEN ANNIS, YONGMIN HOU, DENISE M. MURPHY,
PETER SPACCIAPOLI, ZHIPING YAO, MICHAEL R. ZIEBELL, CLIFF C. CHENG,
GERALD W. SHIPPS JR., JASON S. FELSCH, DAVID LAU, and HUW M. NASH
Screening assays using target-based affinity selection coupled with high-sensitivity detection technologies to identify smallmolecule hits from chemical libraries can provide a useful discovery approach that complements traditional assay systems.
Affinity selection–mass spectrometry (AS-MS) is one such methodology that holds promise for providing selective and
sensitive high-throughput screening platforms. Although AS-MS screening platforms have been used to discover smallmolecule ligands of proteins from many target families, they have not yet been used routinely to screen integral membrane
proteins. The authors present a proof-of-concept study using size exclusion chromatography coupled to AS-MS to perform a
primary screen for small-molecule ligands of the purified muscarinic M2 acetylcholine receptor, a G-protein-coupled receptor.
AS-MS is used to characterize the binding mechanisms of 2 newly discovered ligands. NGD-3350 is a novel M2-specific
orthosteric antagonist of M2 function. NGD-3366 is an allosteric ligand with binding properties similar to the allosteric antagonist W-84, which decreases the dissociation rate of N-methyl-scopolamine from the M2 receptor. Binding properties of the
ligands discerned from AS-MS assays agree with those from in vitro biochemical assays. The authors conclude that when
used with appropriate small-molecule libraries, AS-MS may provide a useful high-throughput assay system for the discovery
and characterization of all classes of integral membrane protein ligands, including allosteric modulators. (Journal of
Biomolecular Screening 2006:194-207)
Key words: affinity selection, mass spectrometry, G-protein-coupled receptor, allosteric
INTRODUCTION
G
-PROTEIN-COUPLED RECEPTORS (GPCRs) present many
challenges to practitioners of modern drug discovery due to
their structural and functional complexities.1,2 Most drugs presently on the market impart their agonistic or antagonistic actions
by binding to the orthosteric site of GPCRs, this being the interaction site where physiologic ligands directly bind.3 Although this
binding mode can provide good drug efficacy, it can also result in
problems associated with reduced receptor subtype selectivity,
nonphysiologic magnitudes of receptor signaling, and inappropriate tissue selectivity.4 Recent reports have suggested that smallmolecule allosteric modulators of GPCRs that bind outside of the
NeoGenesis Pharmaceuticals, Inc., Cambridge, MA.
1
These authors contributed equally to the work.
Received Dec 14, 2004, and in revised form Sep 23, 2005. Accepted for publication Oct 10, 2005.
Journal of Biomolecular Screening 11(2); 2006
DOI: 10.1177/1087057105284340
orthosteric binding site, thereby leaving the orthosteric site intact
for binding physiologic ligand, can have greater receptor subtype
selectivity and may have greater therapeutic potential compared to
orthosteric ligands.5,6 Thus, there is much interest in new
technologies applicable to the discovery of allosteric modulators
of GPCRs.
Advancements in cell-based reporter and screening methodologies now allow the discovery of all classes of GPCR ligands, including allosteric modulators. Because they are considered to be
more physiologically relevant, cell-based functional assays are
currently the preferred format for most primary high-throughput
screening (HTS) campaigns.6,7 Although non-cell-based biochemical assays remain essential for studying the biophysical nature of
receptor-ligand interactions, these assays are less commonly used
in primary HTS campaigns in GPCR drug discovery. This is in part
because of inherent limitations associated with formatting
biochemical assays for high-throughput platforms. Two typical obstacles are obtaining sufficient quantities of a suitable membrane
preparation, purified receptor, or signaling component (G-proteins)
or a fluorescently or isotopically labeled probe. Furthermore, these
194 www.sbsonline.org
© 2006 Society for Biomolecular Sciences
Downloaded from jbx.sagepub.com at PENNSYLVANIA STATE UNIV on September 17, 2016
Muscarinic M2 Acetylcholine Receptor Ligands
screens are usually based on equilibrium binding and displacement of detectable probes that bind to the orthosteric site of the target receptor, and therefore they are inherently biased toward detection of orthosteric hits.5,7 Based on the allosteric ternary complex
model,8,9 it is known that efficacious ligands that bind to
nonorthosteric sites on a receptor can show negligible activities in
standard equilibrium-based displacement assays. Therefore, depending on how primary screens are formatted, allosteric hits can
be overlooked unless carefully sought for, or thorough follow-up
assays are performed.5,8,9
Screening technologies based on affinity selection from smallmolecule libraries, particularly affinity selection–mass spectrometry (AS-MS) techniques, hold promise for overcoming some of the
above limitations of non-cell-based biochemical assays.10-13 When
appropriately formatted, these assays are highly selective and sensitive and can provide a way of discovering many classes of smallmolecule ligands (agonist, antagonist, orthosteric, and allosteric)
in a single screening campaign. Primary hits are usually identified
and validated solely based on reproducible binding to the target
protein and not on a secondary event, such as enzymatic activity or
competitive displacement of a reporter. Although AS-MS screening technologies have been successfully used to discover and characterize small-molecule ligands to proteins of various drug target
classes,10-15 these techniques have not yet been routinely applied to
screening of integral membrane proteins.
One strategy for AS-MS-based screening involves combining a
purified protein target with a library of potential ligands, thus allowing any protein-ligand complexes to form, then separating the
target and bound ligands away from nonbound ligands by a rapid
(< 15 sec) size exclusion chromatography (SEC) step.14 After this
separation step, the protein target and any bound ligands are
immediately transferred to a reverse-phase chromatography (RPC)
column, where the protein is denatured, causing bound ligands to
dissociate, which are then eluted into a high-resolution mass spectrometer for detection. This particular implementation of AS-MS
bears the advantages of having all reaction components in solution,
so no immobilization of the protein target or small-molecule library
is necessary. No radiochemical or fluorescent tags are required,
and any buffers and salts, metal ions, cofactors, or detergents necessary for proper protein folding and stability may be included.
Also, as is the case with most pure affinity selection–based screening methods that are not dependent on a reporter ligand displacement assay, this form of AS-MS enables the potential discovery of
ligands that interact with all binding surfaces of the target protein.
In 1 AS-MS-based screening campaign, one may discover a variety of ligands having unique activities and binding modes.14
In addition to its application as a screening tool, AS-MS has
also been applied as a non-radiation-based analytical tool for characterizing the biophysical properties of receptor-ligand interactions.16 For example, quantitative AS-MS can be used to estimate
the equilibrium binding affinity of a ligand for a target. Because
AS-MS permits the detection of multiple ligands bound to a target
protein in a single mixture, it can also be used to analyze competi-
tive equilibrium binding between differing ligands, allowing both
relative affinity ranking of ligands and classification of binding
mode as directly competitive or noncompetitive.16 Ligands that are
noncompetitive with orthosteric-site binding ligands have the potential of being allosteric modulators. Therefore, assay methods
based on AS-MS may provide both screening and analytical technology platforms, enabling routine discovery and direct biophysical characterization of leads having orthosteric as well as allosteric
binding modes.
In this report, the application of AS-MS to 3 phases of GPCR
drug discovery research using the muscarinic M2 acetylcholine receptor (AChR) as a model system is demonstrated. First, we apply
AS-MS to validate the structural integrity of the orthosteric pocket
of detergent-solubilized purified GPCRs that will subsequently be
used for small-molecule screening. Second, AS-MS is used as an
assay platform to screen a diverse combinatorial small-molecule
library. We present the discovery of 2 novel ligands that bind to the
purified receptor, the antagonist NGD-3350 and the allosteric
ligand NGD-3366. Finally, the binding modes of new ligands to
the receptor are characterized using AS-MS, and results are compared to those obtained using in vitro biological and biochemical
assays. Our findings suggest that assays based on AS-MS may be
routinely applicable to the discovery and characterization of
multiple classes of ligands of purified GPCR targets.
EXPERIMENTAL PROCEDURES
Materials
Atropine and digitonin were purchased from Calbiochem (San
Diego, CA). Scopolamine hydrobromide trihydrate (> 98% purity), N-methyl-scopolamine (NMS), R(–)-quinuclidinyl benzilate
methiodide (QNB), McN-A-343, 3-isobutyl-1-methylxanthine
(IBMX), FLAG-M1 antibody-agarose, and FLAG-M2 antibody
were purchased from Sigma-Aldrich (St. Louis, MO). [2H]3Iodomethane (99.5% atom-2H) was purchased from Cambridge
Isotope Labs (Andover, MA). Forskolin, W-84 dibromide,
pirenzepine dihydrochloride, arecaidine but-2-ynyl ester tosylate
(ABuTs), oxotremorine-M, AF-DX-116, and AF-DX-384 were
purchased from Tocris (Ellisville, MO). EDTA-free protease inhibitor tablets used for protein purification were purchased from
Roche Applied Science (Indianapolis, IN). Purified bovine
cyclooxygenase-1 (COX-1) was purchased from Cayman Chemical (Ann Arbor, MI). Human M1, M3, M4, and M5 expressing membrane preparations were purchased from Perkin Elmer Life and
Analytical Sciences (Boston, MA). Human M1, M3, M4, and M5
containing membrane preparations and [N-methyl- 3H]scopolamine ([3H]-NMS) were purchased from Perkin Elmer Life
and Analytical Sciences (Boston, MA).
Affinity purification of porcine M2 AChR
The porcine M2 AChR (pM2) cDNA (Genbank Accession
NM_214261), engineered with a honeybee melittin signal se-
Journal of Biomolecular Screening 11(2); 2006
Downloaded from jbx.sagepub.com at PENNSYLVANIA STATE UNIV on September 17, 2016
www.sbsonline.org 195
Whitehurst et al.
quence and FLAG-M1 epitope tag positioned upstream of the
starting methionine and inserted in insect expression plasmid
pVL1392, was kindly provided by Dr. E. G. Peralta. Baculovirus
production and Sf9 insect cell infections were performed following standard protocols as outlined by manufacturers (BD Biosciences, San Diego, CA; Orbigen, Inc., San Diego, CA). Membrane
preparations made from infected Sf9 cells typically yielded 10 to
20 pmol/mg of active receptor as determined by standard
radioligand binding assays with [3H]-N-methylscopolamine ([3H]NMS). Membrane preparations adjusted to 2 mg/mL were
solubilized at 4 °C in TBS-A (50 mM TRIS-Cl, pH 7.5, 150 mM
NaCl) supplemented with 1% digitonin and protease inhibitors.
The lysate was clarified by centrifugation at 100,000g for 30 min
followed by 0.65 micron filtration and was diluted 5-fold in TRISbuffered saline (TBS), supplemented with 5 mM CaCl2, and then
passed at < 0.5 mL/min through FLAG-M1 antibody-agarose resin
by fast protein liquid chromatography (FPLC) overnight at 4 °C.
The resin was extensively washed with TBS-A and supplemented
with 0.05% digitonin and 2 mM CaCl2, and pM2 was eluted with 2
column volumes of TBS-A containing 5 mM EDTA, 0.1 mg/mL
FLAG peptide, and 0.05% digitonin. Fractions containing pM2, as
identified by Western blotting with an anti-FLAG-M2 antibody or
specific [3H]-NMS binding, were pooled and directly concentrated
to greater than 20 µM by centrifugal filtration and aliquotted and
stored at –80 °C. The concentration of active pM2 in a preparation
was estimated, taking into account purity and AS-MS-based saturation binding data using NMS as the ligand (see Fig. 2B). Active
pM2 concentrations used in AS-MS experiments are detailed in the
figure legends.
AS-MS screening and mode-of-binding characterization
Details of the AS-MS screening method, mass encoding, and
combinatorial chemical libraries used in this study have been described in detail elsewhere.14,16,17 To summarize, the screening
process involves combining a soluble protein (1-10 µM) and a
mass-encoded small-molecule library (approximately 1500
compounds, each at 1 µM) in a suitable buffer, leading to the formation of protein-ligand complexes that are then separated from
nonbinding library members by a rapid, low-temperature (< 15 sec
at 4 °C) SEC step. This fast separation ensures that relatively
weakly bound ligands (Kd < 10 µM) with moderate dissociate rates
(koff < 0.1 sec–1) are captured. Using an integrated high-performance liquid chromatography (HPLC) system, the protein-ligand
complex is then immediately transferred to an RPC column maintained at 60 °C and pH < 2 to promote dissociation of the ligands.
The ligands are subsequently eluted into a Waters Q-TOF highresolution quadrupole time-of-flight mass spectrometer (Manchester, UK), and tandem (MS/MS) mass spectral data are analyzed by automated software algorithms that ultimately identify
bound ligand structures based on their unique molecular weights
using the mass-encoding data that were associated with the input
libraries used for screening. For the pM2 screen, the affinity selec-
196 www.sbsonline.org
tion step consisted of mixing 1 µL of purified protein (2 µM active
pM2) with 1 µL of mass-encoded library mixture at 5 mM total library concentration, followed by incubation at 25 °C for 30 min
before cooling to 4 °C, and autosampler injection to the SEC for
protein separation from the library pool, ligand dissociation from
the complex, and detection by MS as described above. The affinity
selection formulation consisted of detergent-solubilized pM2 at 1
µM final active concentration in 50 mM TRIS-Cl (pH 7.5), 150
mM NaCl, 0.05% digitonin, and 2.5% DMSO, and each compound library used for affinity screening consisted of a 2500-member mixture at 2.5 mM total compound concentration (approximately 1 µM per compound). As a positive control, the detection of
atropine by the AS-MS platform was used throughout the screen.
The method of dissociation constant (Kd) determinations from
saturation binding and the method of affinity competition experiment (ACE50) ranking of ligand affinities and characterization of
the mode of binding using the same AS-MS system have been described in detail elsewhere.16 Briefly, these experiments are based
on equilibrium affinity selection, whereby a constant concentration of test ligands and serially increasing concentrations of a
known competitor ligand (titrant) are incubated together with the
target protein and then analyzed by AS-MS. The ability of the
titrant to displace the target-bound ligands reveals the binding site
classification (competitive vs. noncompetitive) and relative binding
affinity ranking of these ligands. This is because the resulting MS
signals of the ligands and the competitor from such experiments reflect the equilibrium concentrations of each protein-ligand complex, thereby yielding information about the equilibrium dissociation constant (Kd) of each reaction component and allowing
affinity ranking. A plot of the ratio of the MS responses for the
titrant and ligand will yield a straight line for directly competitive
ligands or yield a hyperbolic curve for mixed-competitive or noncompetitive ligand-ligand interactions.16
2
Synthesis of [ H]3-N-methylscopolamine
iodide and other compounds
[2H]3-N-methylscopolamine iodide ([2H]3-NMS) was synthesized according to the literature method18 for the parent compound
with [2H]3-iodomethane substituted for bromomethane. Briefly,
0.94 g (2.15 mmol) scopolamine hydrobromide trihydrate was dissolved in 2.5 mL water and made alkaline by combining with 0.4 g
(10 mmol) sodium hydroxide in 1.5 mL water, and the free base
was extracted with diethyl ether. The aqueous layer was saturated
with sodium chloride and again extracted with ether, the combined
extracts were dried over magnesium sulfate and filtered, and the
ether was removed by rotary evaporation to yield a colorless oil. To
this was added 0.65 mL [2H]3-iodomethane (11 mmol), immediately yielding a colorless precipitate. The flask was stoppered and
allowed to stand in the dark at room temperature for 48 h, and then
the precipitate was collected, washed with cold ether, and
recrystallized from 5 mL absolute ethanol to yield 108 mg [2H]3NMS iodide (20%) as colorless crystals. The product was analyzed
Journal of Biomolecular Screening 11(2); 2006
Downloaded from jbx.sagepub.com at PENNSYLVANIA STATE UNIV on September 17, 2016
Muscarinic M2 Acetylcholine Receptor Ligands
by reverse-phase chromatography with MS, diode-array UV, and
evaporative light-scattering detection and demonstrated to be of
> 98% chemical purity and > 99% isotopic purity. MS (ESI-TOF),
expected for [2H]3-C18H21NO4+ 321.1891, found 321.1823. Other
compounds used in this study were prepared by the NeoGenesis
Medicinal Chemistry Group and demonstrated to be > 95% pure
by reverse-phase chromatography with MS, diode-array UV, and
evaporative light-scattering detection.
Analysis of the influence of allosteric compounds
on the dissociation rate of N-methylscopolamine
from pM2 by AS-MS
For each dissociation experiment, NMS was preloaded onto
purified pM2 for 60 min at 25 °C in a 40-µL binding reaction consisting of 8 µM pM2 and 6 µM NMS in TBS-A (50 mM TRIS-Cl
[pH 7.5], 150 mM NaCl, and 0.05% digitonin) containing 2.5%
DMSO. This was mixed 1:1 with 200 µM allosteric compound
(NGD-3366 or W-84 dibromide) predissolved in TBS-A (2.5%
DMSO), and the reaction was incubated for 90 min at 25 °C. This
was then mixed 1:1 with TBS-A containing 400 µM [2H]3-NMS
and 2.5% DMSO such that the final experimental sample contained 2 µM active pM2, 1.5 µM NMS, 50 µM allosteric compound, 200 µM [2H]3-NMS, and 2.5% DMSO, and this was immediately chilled to 4 °C, and 2-µL aliquots were sampled every 7
min for AS-MS analysis. kq is the association rate constant of the
quencher [2H]3-NMS and is dependent on the dissociation of the
preloaded NMS (koff).
Radioligand-based binding and dissociation assays
The Kd of [3H]-NMS for purified pM2 was determined by incubating 1.5 × 10–3 µM active pM2 in 100 µL of binding buffer (50
mM HEPES, pH 7.5, 100 mM NaCl) with a range of [3H]-NMS
(81 Ci/mmol) up to 125 nM. Nonspecific binding was determined
by inclusion of 2 µM atropine, and the residual radioactivity present in the 2 µM atropine sample was subtracted from the total signal to obtain the specific binding values. After mixing for 30 min at
25 °C, protein was passed through a 1-mL Sephadex G-50 column
to remove free unbound radioligand, and bound [3H]-NMS was
measured by liquid scintillation counting on a Wallac 1450. Kd and
nonspecific binding was calculated using the GraphPad Prism
equation “saturation binding with ligand depletion,” setting the
VOL constant to 0.1 mL and SPEACT constant to 6.35 cpm/fmol.
Human M2 AChR (hM2)–containing membranes were prepared
from the recombinant Chinese hamster ovary (CHO)–K1 cell
line, NGTZ-30, that was engineered to express human M2 AChR
(10-20 pmol/mg membrane) with a FLAG-M1 epitope tag at the
N-terminus and a 6-histidine tag at the C-terminus. The binding
properties of the modified hM2 AChR expressed on NGTZ-30
were not substantially altered from the native receptor, as determined by affinity rank ordering of various M2 antagonists in com-
petition against [3H]-NMS and as determined by the ability of the
cell line to respond normally to M2 agonists using standard cAMP
assays. For competition binding assays, membrane receptor was
adjusted to 40 × 10–6 µM in binding buffer and incubated with 0.3
nM [3H]-NMS and compounds at varying concentrations for 3 h at
25 °C, collected onto GF/B filter plates, and washed 8 times with
ice-cold TBS-B, and bound [3H]-NMS was measured by scintillation counting on a Wallac 1450. Dissociation rate experiments
were performed by premixing hM2 membranes (40 pM) with 0.3
nM [3H]-NMS for 60 min at 25 °C, incubating with allosteric ligands for 10 min at 25 °C, chasing with 2 µM atropine for set times
until reactions were stopped by rapid filtration and washing over
Millipore GF/B filter plates under vacuum filtration, and measuring bound [3H]-NMS by scintillation counting on a Wallac 1450.
Analyses and curve fittings of biochemical data were performed
using GraphPad Prism software.
Cell-based cAMP assay
Changes in cAMP levels in the hM2-expressing cell line
NGTZ-30 were measured using a homogeneous time-resolved
fluorescence assay (HTRF) following the manufacturer recommendations (CIS-US Incorporated, Bedford, MA). Briefly, cells
were seeded at 10,000 cells/well in a flat-bottom 96-well culture
plate in 1:1 F12/Dulbecco’s modified Eagle’s medium (DMEM)
containing 3.1 g/L glucose, 2.5 mM L-glutamine, 55 mg/L sodium
pyruvate, and 10% Fetaclone II (Hyclone, Logan, UT) and were
equilibrated by overnight culture at 37 °C and 5% CO2.
Preincubation with 1 mM IBMX for 60 min was performed before
addition of test compounds or carrier (0.2% DMSO), in combination with 30 nM oxotremorine-M and 2 µM forskolin, for a stimulation period of 30 min before lysis and HTRF analysis. Normalized ratiometric values were determined following the
manufacturer’s recommendations using the following equation:
∆F% = [(SR – SRneg)/SRneg] × 100, where SR = (A665 nm/A620
nm) and SRneg = sample ratio (SR) of the negative control without
fluorescent cAMP conjugate. Data are expressed as percent inhibition of oxotremorine-M-induced down-regulation of forskolinstimulated increases in intracellular cAMP of cell line NGTZ-30.
Guinea pig atrial assay
Compounds were tested by MDS Pharma Services (Bothell,
WA) in a guinea pig atrial assay measuring inhibition of
methacholine-induced negative inotropy.19 Atria excised from
Duncan Hartley guinea pigs were placed in 10-mL baths of
McEwen’s buffer, pH 7.4, at 34 °C, and after equilibration and addition of compounds, inhibition of the methacholine-induced
inotropic response was measured over an assessment period of 5
min. Data were normalized to vehicle (0.1% DMSO), and greater
than 50% inhibition was considered significant in this assay
system.
Journal of Biomolecular Screening 11(2); 2006
Downloaded from jbx.sagepub.com at PENNSYLVANIA STATE UNIV on September 17, 2016
www.sbsonline.org 197
Whitehurst et al.
FIG. 1. Characterization of affinity-purified pM2. (A) Protein was analyzed by SDS-PAGE followed by Coomassie staining (lanes 2-4) or by antiFLAG-M2 Western blotting (lanes 6, 7). Molecular weight standards are in kDa (lanes 1, 5), and monomeric (expected 51.6 kDa) and oligomeric forms of
the receptor are indicated by [pM2]1 and [pM2]n, respectively. (B) Affinity selection–mass spectrometry (AS-MS)–based recovery of antagonists (1 µM atropine, 1 µM QNB, 1 µM AFDX-116, and 10 µM pirenzepine) bound by 1 µM pM2 in the presence of DMSO (control) or compound libraries (libraries A,
B, C, D, and E) containing 2500 compounds each at 1 µM per compound. Data are mean and standard error of duplicate samples. (C) AS-MS-based recovery of 10 µM atropine and agonists McN A343 and ABuTs bound by 1 µM pM2 in the presence of DMSO (CONT) or compound libraries (A-C) consisting
of approximately 2500 compounds each at 1 µM. Control membrane protein (cyclooxygenase-1 [COX-1]) formulated similarly to pM2 was included as a
specificity control with DMSO in place of the library (CONT). Data are the mean and standard error of duplicate samples.
RESULTS
Characterization of purified pM2 AChR
by traditional and AS-MS binding assays
We chose the muscarinic M2 AChR for the proof-of-concept
screening because it is relatively well characterized and amenable
to purification and a variety of agonists, and antagonists were commercially available for binding studies. At the initiation of this
study, a suitable human cDNA clone was not available; therefore,
we used a porcine cDNA to develop an expression system. The
porcine M2 AChR (pM2) is highly homologous to human M2 (hM2)
with 97.4% identity and has nearly identical binding affinities for
muscarinic ligands.20 A honeybee melittin signal sequence and
FLAG M1 epitope tag was engineered at the N-terminus of the
open reading frame of pM2 to enhance proper membrane targeting
and folding of the receptor.21 Consistent with a previous report that
the active M2 receptor can be expressed in the baculovirus expression system,22 we produced Sf9 cell membranes yielding > 10
pmol of pM2 per milligram of membrane protein. The melittin signal sequence was efficiently cleaved to reveal the FLAG-M1
epitope that was subsequently used for affinity purification of
digitonin-solubilized receptor by anti-FLAG M1-agarose and
fluid phase liquid chromatography. A final receptor purity of approximately 70% to 80% and concentration of > 20 µM was routinely obtained (Fig. 1A). The monomeric form of the receptor
([pM2]1) had the expected molecular weight of approximately 51.6
kDa.23
198 www.sbsonline.org
Purified integral membrane proteins in detergent micelles may
become unfolded and structurally denatured during screening. To
determine if receptor stability would be maintained during the critical AS and SEC steps of the screening process, we first performed
control binding experiments by premixing purified pM2 with randomly selected combinatorial libraries (2500 compounds each at 1
µM) into which known M2 antagonists were spiked. Concentrations of each known ligand were chosen to provide adequate MS
signal response, the magnitude of which is dependent on affinity as
well as MS sensitivity of each compound. The MS signal of atropine, QNB, AFDX-116, or pirenzepine was equivalent when combined with pM2 alone or spiked into libraries, indicating that the receptor is fully functional throughout the screening process (Fig.
1B). The finding that agonists McN-A343 and ABuTs are also detected by the same experimental method (Fig. 1C) further supports
this conclusion. Depending on the control antagonist and agonist
or specific library used in these preliminary assays, slight augmentation or attenuation of the MS response signals was observed,
which may result from library-specific stabilization or perturbation effects on the GPCR-ligand complex at the AS, SEC, or RPCMS (ion suppression) stage of the assay process. The observed
binding is specific because the same agonists were not detected
when premixed with another integral membrane protein, COX-1
(Fig. 1C).
We further characterized the purified pM2 by performing a
radioligand binding assay using Sephadex G-50 gel filtration to
separate unbound from bound [3H]-NMS. Saturation binding re-
Journal of Biomolecular Screening 11(2); 2006
Downloaded from jbx.sagepub.com at PENNSYLVANIA STATE UNIV on September 17, 2016
Muscarinic M2 Acetylcholine Receptor Ligands
3
FIG. 2. Characterization of affinity-purified pM2 by saturation binding experiments. (A) Dissociation constant (Kd) of [ H]-NMS, as determined by
–3
3
saturation binding to 1.5 × 10 µM of purified pM2 and separation of free radioligand from pM2-[ H]-NMS complexes by small-scale Sephadex G-50 gel
filtration. Each data point is the mean and standard error of 3 independent experiments. (B) Affinity selection–mass spectrometry (AS-MS)–based Kd of
N-methyl-scopolamine (NMS) binding to purified pM2 and estimation of active receptor in the preparation. Data are the mean and standard error of duplicate samples. Note that this experiment was conducted with an estimated initial pM2 concentration at 2.5 µM, assuming 100% purity and using the total
protein concentration of the preparation. However, pM2 purity was approximately 80%, as shown in Figure 1A, and as indicated by this experiment, active
pM2 in the purified preparation was approximately 1 µM. (C) ACE50 assay demonstrating the direct binding competition between NMS and AF-DX-384
on 1 µM active pM2 by AS-MS. Data are the mean and standard error of duplicate samples.
veals a Kd for [3H]-NMS of 3.8 (± 1.7) nM (Fig. 2A), a value 10- to
20-fold higher than expected for intact pM2 in membrane. This
shift in Kd is likely due to slight alterations in receptor structure, as
has been reported for other solubilized GPCRs present in detergent
micelles.24-28 It has been previously shown that AS-MS can be used
to characterize ligand-receptor binding affinities.16 Therefore, we
performed a saturation binding experiment using 1 µM of pM2 and
NMS concentrations ranging from 0 to 4 µM using AS-MS to
measure bound NMS. As shown in Figure 2B, NMS binding to
pM2 was saturable, and the binding affinity was similar to that obtained by the Sephadex G-50-based method (Kd = 6.2 ± 4 nM), further suggesting that pM2 activity remains unchanged during the
AS-MS process.
The structural integrity of a purified soluble protein can also be
inferred from competitive binding experiments between differing
ligands. The AS-MS-based ACE50 method allows one to distinguish whether ligands are directly competitive for a single binding
site or bind to differing sites on a receptor.16 Using the ACE50
method, we tested whether increasing concentrations of
orthosteric-binding NMS premixed with pM2 would compete off
binding of the orthosteric M2/M4 antagonist AF-DX-384.27 As
shown in Figure 2C, binding of AF-DX-384 was competed off by
increasing concentrations of NMS, and a plot of the MS response
ratio (signal from NMS divided by signal from AF-DX-384) versus the NMS concentration revealed a distribution with a linear fit,
signifying that these ligands were directly competing for the same
binding site on pM2, consistent with their known binding properties, as shown by conventional radioligand binding assays.16,29
From these various binding experiments, we concluded that the
purified pM2 preparation retained sufficient stability and binding
activity during the AS-MS assay to be used for a screening campaign using AS-MS to discover new M2 ligands from smallmolecule combinatorial libraries.
Discovery of novel pM2 AChR ligands by screening
small-molecule combinatorial libraries using AS-MS
We conducted a screen of 244 library pools (356,474 compounds total) consuming 5 nmol (250 µg) of the purified pM2
AChR. A total of 48 primary ligands were identified and confirmed as pM2-specific ligands in rapid follow-up AS-MS binding
experiments. The compounds were then synthesized in multimilligram scale, purified, and used in additional biochemical and
AS-MS-based assays to characterize binding. Of the 48 ligands, 24
were validated as biochemically active hits based on competitive
displacement of [3H]-NMS from insect membrane preparations
expressing pM2 (data not shown). Ligands were prioritized based
on ACE50 binding analyses using AS-MS to identify those compounds that were most resistant to displacement from purified pM2
by increasing concentrations of the competitive antagonist atropine. The results of this ACE50 experiment were previously reported.16 Three compounds—NGD-3346, NGD-3348, and NGD3350—were chosen for further study (Fig. 3A). These ligands are
structural congeners, and of all the compounds tested, NGD-3350
shows the highest binding affinity to pM2.16 The AS-MS-based
binding experiments revealed that all the ligands show saturable
single-site binding to pM2 like the M2-specific control antagonist
AF-DX-11629,30 (Fig. 3B). Kd values of 0.4, 2.9, 2.1, and 0.7 µM
Journal of Biomolecular Screening 11(2); 2006
Downloaded from jbx.sagepub.com at PENNSYLVANIA STATE UNIV on September 17, 2016
www.sbsonline.org 199
Whitehurst et al.
FIG. 3. M2 AChR binding affinities of AF-DX-116, NGD-3346, NGD-3348, and NGD-3350. (A) Chemical structures of NGD-3346, NGD-3348,
and NGD-3350. (B) Kd of ligands as determined by saturation binding to purified pM2 (1 µM active) and directly measuring ligand recoveries by affinity
selection–mass spectrometry (AS-MS). Data are the mean and standard error of duplicate samples. (C) Apparent affinity constants (Ki) of ligands for
3
membrane hM2 as measured by [ H]-NMS (0.3 nM) displacement competition binding experiments. Data are the mean and standard error of duplicate
samples.
were obtained for AF-DX-116, NGD-3346, NGD-3348, and
NGD-3350, respectively. Indirect affinity constants (Ki) of AFDX-116, NGD-3346, 3348, and 3350 were also determined by
measuring competitive displacement of [3H]-NMS from the human receptor (hM2) expressed on CHO-K1 cell membranes (cell
line NGTZ-30). The resultant Ki constants correspond well with
the Kd values derived from the AS-MS assays (Fig. 3C). NGD3350 is therefore the tightest binding M2 ligand (Ki = 0.2 µM) discovered from this proof-of-concept screen, consistent with our
previous observation.16
NGD-3350 is an M2-selective AChR orthosteric antagonist
The ability of NGD-3350 to inhibit the binding of [3H]-NMS to
hM2 expressed in membrane preparations (Fig. 3C) suggested that
it might be directly competitive with compounds binding to the
orthosteric antagonist binding site of the M2 AChR. To test this hypothesis, we performed AS-MS-based ACE50 assays to study the
mode of binding of NGD-3350. Figure 4A shows that NMS inhibits binding of NGD-3350 to purified pM2, and NGD-3350 inhibits
binding of AF-DX-116 (Fig. 4B), as well as AF-DX-384 and the
classic muscarinic antagonist QNB (Fig. 4C). The linear plots of
titrant/competitor response ratios and titrant concentrations shown
200 www.sbsonline.org
in Figure 4D-F indicate direct competition. The data confirm that
NGD-3350 is directly competitive with orthosteric antagonists of
pM2 (Fig. 4D-F).
The receptor-binding selectivity of NGD-3350 was also examined. Like the antagonists atropine and AF-DX-116, at 25-µM
concentrations, NGD-3346, NGD-3348, and NGD-3350 all displace 1 nM [3H]-NMS from binding to membrane preparations
that contain the hM2 AChR but do not displace 1 nM [3H]CGP12177 from binding to membrane preparations that contain
the human β2-adrenergic receptor (Fig. 5A). When assayed for the
ability to displace [3H]-NMS from the other muscarinic AChR
subtypes, NGD-3350 shows a 38-, > 100-, 34-, and 27-fold selectivity for M2 over the M1, M3, M4, and M5 subtypes, respectively
(Fig. 5B). The selectivity of NGD-3350 for M2 versus M1 and M4
subtypes is approximately 5-fold greater than that of AF-DX-116
(Fig. 5C).
NGD-3350 is a functional antagonist
of the muscarinic M2 AChR
Agonism of the M2 AChR induces association with Gαi/o proteins that directly inhibit cellular adenylyl cyclase activity.31 Using
the NGTZ-30 cell line, we tested NGD-3346, NGD-3348, and
Journal of Biomolecular Screening 11(2); 2006
Downloaded from jbx.sagepub.com at PENNSYLVANIA STATE UNIV on September 17, 2016
Muscarinic M2 Acetylcholine Receptor Ligands
FIG. 4. NGD-3350 is directly competitive with classic orthosteric M2 antagonists. (A) ACE50 experiment showing displacement of NGD-3350 from 1
µM active pM2 by titrant N-methyl-scopolamine (NMS). (B) ACE50 experiment showing displacement of AF-DX-116 from 1 µM active pM2 by titrant
NGD-3350. (C) ACE50 experiment showing displacement of R(–)-quinuclidinyl benzilate methiodide (QNB) and AF-DX-384 from pM2 by titrant NGD3350. (D-F) Plots showing linear correlations between increasing titrant concentration and titrant/competitor response ratios for experiments in A to C.
Linearity demonstrates directly competitive binding between NGD-3350 and known M2 antagonists. All data in Figure 4 represent mean and standard error of duplicate samples.
NGD-3350 for agonist and antagonist activities. No agonist activity, assayed as hM2-dependent attenuation of forskolin-stimulated
cAMP production, was detected for these compounds. However,
when assayed for antagonism of hM2-dependent oxotremorineinduced attenuation of forskolin-stimulated cAMP production,
control antagonist AF-DX-116 and NGD-3346, NGD-3348, and
Journal of Biomolecular Screening 11(2); 2006
Downloaded from jbx.sagepub.com at PENNSYLVANIA STATE UNIV on September 17, 2016
www.sbsonline.org 201
Whitehurst et al.
3
FIG. 5. NGD-3350 binds with high selectivity to the M2 subtype of muscarinic AChR. (A) Assays measuring displacement of 1 nM [ H]-NMS from
3
hM2 AChR and displacement of 1 nM [ H]-CGP12177 from the human β2-adrenergic receptor by 25 µM atropine, NGD-3346, NGD-3348, and NGD3
3350. (B) Apparent affinity constants (Ki) of NGD-3350 for M1 to M5 subtypes of human muscarinic AChR as measured by [ H]-NMS displacement com3
petition binding experiments. (C) Apparent affinity constants (Ki) of AF-DX-116 for M1 to M5 subtypes of human muscarinic AChR as measured by [ H]NMS displacement competition binding experiments. All data points in Figure 5 represent the mean and standard error of duplicate samples.
NGD-3350 all show antagonist activity, with IC50 values of 0.22,
3.2, 4.5, and 0.34 µM, respectively (Fig. 6A). As another measure
of functional antagonism by NGD-3350, the compounds were
tested in the classic guinea pig left atrial assay measuring antagonism of methacholine-induced negative inotropy, which is stimulated through the M2 AChR.19,32 Only NGD-3350 shows significant (> 50%) antagonism at 30 µM (Fig. 6B), and a dose-response
experiment suggests an IC50 value of approximately 7 µM (Fig.
6C).
NGD-3366 binds to the muscarinic M2 AChR
outside of the orthosteric binding pocket
During follow-up ACE50 experiments to characterize additional
ligands identified in our screen, we discovered that NGD-3366
(Fig. 7A) binds to pM2, and its binding is not prevented by saturating concentrations of the orthosteric antagonists NMS or NGD3350 (Fig. 7B,C). Plots of the titrant/competitor response ratios
versus increasing titrant concentrations suggest that ternary complexes are formed between pM2, NGD-3366, and NMS or pM2,
NGD-3366, and NGD-3350 (Fig. 7E,F). In a reciprocal biochemical assay, we found that at high concentrations, NGD-3366 is able
202 www.sbsonline.org
to displace binding of [3H]-NMS to hM2 expressed on NGTZ-30
membranes (Ki = 45 µM; Fig. 7D). These observations suggest
that NGD-3366 might be a weak allosteric antagonist that binds
outside of the orthosteric binding pocket of M2 that is occupied by
NMS.
The allosteric antagonist compound W-84 binds to a common
allosteric site on the M2 AChR that resides above the orthosteric
binding pocket.4 Because of this binding mode, W-84 may form a
“cap” on the receptor, impeding the dissociation of compounds
that are bound in the orthosteric pocket and effectively decreasing
their dissociation rate.33,34 Therefore, we tested whether NGD3366 could bind in a manner analogous to W-84 and alter the dissociation rate of the orthosteric ligand NMS from the M2 AChR.
As shown in Figure 8A, both NGD-3366 and W-84 decrease the
dissociation rate of [3H]-NMS from hM2 AChR expressed on
NGTZ-30 membranes (Fig. 8A).
As an additional test of this binding mode, we performed a similar dissociation experiment using AS-MS. Purified pM2 was
preloaded with NMS, incubated with W-84 or NGD-3366, and
then the dissociation of NMS was assayed over time by chasing
with an excess of the isotopically labeled competitor ligand [2H]3NMS. The mass difference between NMS and [2H]3-NMS is suffi-
Journal of Biomolecular Screening 11(2); 2006
Downloaded from jbx.sagepub.com at PENNSYLVANIA STATE UNIV on September 17, 2016
Muscarinic M2 Acetylcholine Receptor Ligands
FIG. 6. NGD-3350 demonstrates functional antagonism of M2 AChR in cell- and organ-based assays. (A) Antagonism of oxotremorine-induced hM2
AChR-dependent attenuation of forskolin-stimulated cAMP production in Chinese hamster ovary (CHO)-K1 cell line NGTZ-30 by AFDX-116, NGD3346, NGD-3348, and NGD-3350. Data shown are the mean and standard error of duplicate samples. (B) Antagonism of methacholine-induced negative
inotropy of guinea pig left atria by compounds at 30 µM. (C) Dose-response curve of NGD-3350 antagonism of methacholine-induced negative inotropy
of guinea pig left atria. Data in B and C are single data points.
cient to permit independent monitoring of both NMS dissociation
from pM2 (koff) and simultaneous association of the quencher
[2H]3-NMS to pM2 (kq), the latter of which indirectly approximates
the off rate of NMS. As shown in Figure 8B, the AS-MS dissociation rate experiment revealed that both NGD-3366 and W-84 decrease the dissociation rate of NMS from pM2. The corresponding
changes in the rate of [2H]3-NMS association effected by both W84 and NGD-3366 also support this conclusion, and the convergence of signals for [2H]3-NMS at the last 100-min time point indicates that neither W-84 nor NGD-3366 act by irreversibly denaturing the receptor. The dissociation rates measured by AS-MS are 4fold faster than those determined by radioligand-based biochemical assay (Fig. 8B), suggesting that this method does not perfectly
mimic standard biochemical methods using radioligands. This difference may be associated with limitations in AS-MS sampling
speed. There is a time lag of approximately 2 min between sample
retrieval by the HPLC autosampler and the final liquid chromatography/mass spectrometry (LC/MS) ligand detection step, and samples could only be acquired at 7-min intervals. Therefore, koff rates
under 5 min are based purely on curve fitting and are therefore inherently less accurate. Nevertheless, the relative changes in the dissociation rates observed by AS-MS are informative and consistent
with the results from radioligand-based assays, suggesting that
NGD-3366 has a binding mode similar to W-84.
DISCUSSION
In this report, we have demonstrated the use of AS-MS for early
stage hit discovery and ligand characterization at many levels.
First, we have shown that AS-MS-based ligand-binding assays can
be used to validate the structural integrity of the orthosteric site of a
purified GPCR prior to its use in subsequent screening campaigns.
Purified preparations of digitonin-solubilized pM2 were validated
for screening by 3 types of AS-MS experiments: we demonstrated
specific binding of pM2 to known muscarinic M2 AChR agonists
and antagonists spiked into random combinatorial libraries, we
demonstrated that the AS-MS-based Kd of NMS binding to pM2
was similar to that determined by traditional assays using [3H]NMS and Sephadex G-50 gel filtration, and we demonstrated that
NMS behaved classically as an orthosteric competitor of AF-DX384 in the ACE50 assay. AS-MS-based assays (Kd, koff, and ACE50)
may provide an alternative to, or complement, existing traditional
biochemical assay methodologies when one needs to validate that
a purified receptor has expected ligand-binding activity.
Second, we have shown that AS-MS can be used as a discovery
platform for identifying multiple classes of GPCR ligands in a single screening campaign. In our proof-of-concept screen of M2
AChR against 244 library pools (356,474 compounds total), we
consumed under 5 nmol (250 µg) of purified receptor and identified
Journal of Biomolecular Screening 11(2); 2006
Downloaded from jbx.sagepub.com at PENNSYLVANIA STATE UNIV on September 17, 2016
www.sbsonline.org 203
Whitehurst et al.
FIG. 7. M2 AChR ligand NGD-3366 shows allosteric binding versus N-methyl-scopolamine (NMS) and NGD-3350. (A) NGD-3366 structure. (B)
ACE50 assay showing NGD-3366 binding to 1 µM active pM2 is not directly competed off by NMS. (C) ACE50 assay showing NGD-3366 binding to 1 µM
3
active pM2 is not directly competed off by NGD-3350. (D) Radioligand displacement assay showing that NGD-3366 weakly inhibits [ H]-NMS binding to
membrane-expressed hM2 AChR. (E, F) Plots showing nonlinear relationship between increasing NMS concentration and NMS/NGD-3366 response ratio (E) or NGD-3350 concentration and NGD-3350/NGD-3366 response ratio (F), indicating ternary complex formation between pM2, NMS, and NGD3366 or pM2, NGD-3350, and NGD-3366, as detected by affinity selection–mass spectrometry (AS-MS).
48 primary hits that showed confirmed binding in follow-up ASMS-based binding experiments. Of these, 50% showed activity in
radioligand competition binding assays using pM2-expressing
membranes. Several biologically active antagonists were identified, and one in particular, NGD-3350, displayed submicromolar
affinity and a greater degree of selectivity for the M2 AChR subtype than the classic M2-specific antagonist AF-DX-116. NGD3350 was shown to be a bioactive M2 antagonist in preliminary
cell-based experiments measuring antagonism of oxotremorineinduced attenuation of forskolin-stimulated intracellular cAMP, as
well as in an organ-based experiment using a standard ex vivo
guinea pig atrial assay of M2 function. NGD-3350 may be a useful
molecular probe for future studies of muscarinic M2 AChR
pharmacology.
Finally, we have shown that AS-MS-based assays can be used
to characterize receptor binding properties of newly discovered
ligands. The relative binding affinities and binding modes
204 www.sbsonline.org
(orthosteric vs. allosteric) of initial pM2 hits were characterized by
ACE50 assays using the antagonist NMS as a reference titrant. This
led to the early identification of NGD-3350 as an interesting and
relatively high-affinity orthosteric M2 hit. NGD-3350 was further
characterized by measuring its Kd with AS-MS, by applying
standard biochemical radioligand displacement assays using
M2-expressing membranes, and by performing additional ACE50
assays using other well-defined M2 antagonists (QNB, AF-DX116, and AF-DX-384). In follow-up studies of M2 ligands that
could not be classified as orthosteric binders, we discovered that
NGD-3366 binding to pM2 could not be displaced by the
orthosteric ligand NMS in ACE50 experiments, indicating that a
ternary complex is formed between pM2, NMS, and NGD-3366.
In an equilibrium binding reaction using membrane hM2, NGD3366 is a weak competitor of [3H]-NMS binding. These binding
characteristics of NGD-3366 are consistent with a binding model
that uses the common allosteric pocket of the M2 receptor, perhaps
Journal of Biomolecular Screening 11(2); 2006
Downloaded from jbx.sagepub.com at PENNSYLVANIA STATE UNIV on September 17, 2016
Muscarinic M2 Acetylcholine Receptor Ligands
FIG. 8. Allosteric ligand NGD-3366 decreases the dissociation rate of N-methyl-scopolamine (NMS) from M2 AChR. (A) Measure of the effect of
3
NGD-3366 or W-84 on dissociation of [ H]-NMS from hM2 AChR expressed on membranes derived from cell line NGTZ-30. (B) Simultaneous measure
2
of the effect of NGD-3366 or W-84 on NMS dissociation (koff) from purified pM2 AChR and the linked association of the quencher [ H]3-NMS (kq) to purified pM2 AChR as assayed by affinity selection–mass spectrometry (AS-MS).
in a mode similar to the allosteric antagonist W-84.33,34 Unfortunately, W-84 is not directly detectable in AS-MS due to its
physicochemical properties, and therefore we were not able to perform an ACE50 experiment to directly measure the competition of
W-84 versus NGD-3366 and determine orthosteric or allosteric
binding for these 2 ligands. However, to confirm the proposed
allosteric binding mode of NGD-3366, we performed dissociation
rate studies using both biochemical and AS-MS-based methods,
demonstrating that NGD-3366 retards the dissociation of [3H]NMS from the M2 AChR in a manner similar to the known
allosteric ligand W-84 and confirming that NGD-3366 is a specific
ligand to M2 AChR.
Although we have shown that AS-MS-based screening of
combinatorial libraries can be a potentially useful method for discovering new ligands of integral membrane proteins, we should
highlight the limitations of this methodology. First, this technology platform requires a purified (> 70% purity) soluble active receptor preparation (> 0.5 µM active). Only those receptors that
can maintain an active structure throughout the detergent
solubilization, affinity purification, and concentration or dialysis
steps in the formulation process will be suitable for AS-MSbased applications. Second, there is the liability associated with
the fact that a detergent-solubilized GPCR may never reflect the
true physiological structure of a GPCR embedded in membrane, as
might be expected in most membrane- and cell-based assays. As
advances in integral membrane protein expression, purification,
and biophysical and structural studies progress, so will the opportunities for applying AS-MS-based assays for receptor screening
and characterization. Future targets for AS-MS might be receptors
beyond the GPCR family, for example, ion channels, transporters,
integrins, and cytokine receptors. Third, we should emphasize that
all AS-MS-based assays are dependent on the availability of ligands having suitable physicochemical properties for detection by
MS. Experimental ligands of interest may not always have suitable
properties. Typically, they must appropriately ionize and be routinely detectable at levels down to 100 to 200 fmoles. Highly hydrophobic compounds lacking suitable charged groups are often
not amenable to detection in the AS-MS platform used in this report. Indeed, some of the compounds discovered in our primary
AS-MS-based screen of M2 AChR displayed relatively weak signals in the mass spectrometer that were sufficient for initial hit
identification but too weak for follow-up Kd and ACE50 assays. Obviously, to ensure that weak hits in primary screens are not mistaken as false negatives, they should be tested in alternative assays
not reliant on MS for detection. The usefulness of compounds for
AS-MS-based assays can be easily verified ahead of time by stan-
Journal of Biomolecular Screening 11(2); 2006
Downloaded from jbx.sagepub.com at PENNSYLVANIA STATE UNIV on September 17, 2016
www.sbsonline.org 205
Whitehurst et al.
dard LC/MS analyses. Finally, we should emphasize that the combinatorial small-molecule libraries used in this study were specially designed for use in the AS-MS screening platform.16,17 As is
the case for any high-throughput drug screening platform, the discovery outcome can be biased by the assay methodology used. For
AS-MS-based screening, although using affinity selection may reduce bias because it allows discovery of all classes of ligands
(orthosteric and allosteric), bias is nevertheless introduced at another level because all screening libraries to be used for AS-MS
must contain ligands detectable by MS. Reducing bias while increasing the chances of finding drug-like hits presents a unique
synthesis challenge to those designing and implementing smallmolecule combinatorial libraries for an AS-MS screening platform. Future advancements in chemical library development for
AS-MS platforms may greatly improve upon the current discovery
potential of this discovery technology.
ACKNOWLEDGMENTS
We thank Aisling O’Donovan and William Ho Hong Li for excellent technical assistance and Satish Jindal, Steve Adams, Elliott
Nickbarg, Art Patchett, and Roland Seifert for helpful discussions
and support.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
The state of GPCR research in 2004. Nat Rev Drug Discov 2004;3:577-626.
Pierce KL, Premont RT, Lefkowitz RJ: Seven-transmembrane receptors. Nat
Rev Mol Cell Biol 2002;3:639-650.
Drews J: Drug discovery: a historical perspective. Science 2000;287:19601964.
Christopoulos A: Allosteric binding sites on cell-surface receptors: novel targets for drug discovery. Nat Rev Drug Discov 2002;1:198-210.
Christopoulos A, May LT, Avlani VA, Sexton PM: G-protein-coupled receptor allosterism: the promise and the problem(s). Biochem Soc Trans
2004;32:873-877.
Soudijn W, van Wijngaarden I, Ijzerman AP: Allosteric modulation of G
protein-coupled receptors: perspectives and recent developments. Drug Disc
Today 2004;9:752-758.
May LT, Christopoulos A: Allosteric modulators of G-protein-coupled receptors. Curr Opin Pharmacol 2003;3:551-556.
Christopoulos A, Kinakin T: G protein-coupled receptor allosterism and
complexing. Pharmacol Rev 2002;54:323-374.
Lazareno S, Birdsall NJM: 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.
Kaur S, McGuire L, Tang D, Dollinger G, Huebner V: Affinity selection and
mass spectrometry-based strategies to identify lead compounds in combinatorial libraries. J Protein Chem 1997;16:505-511.
Kelly MA, McLellan TJ, Rosner PJ: Strategic use of affinity-based mass
spectrometry techniques in the drug discovery process. Anal Chem
2002;74:1-9.
Stockwell BR: Chemical genetics: ligand-based discovery of gene function.
Nat Rev Genet 2000;1:116-125.
206 www.sbsonline.org
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
Falb D, Jindal S: Chemical genomics: bridging the gap between the proteome
and therapeutics. Curr Opin Drug Disc Develop 2002;5:532-539.
Annis DA, Athanasopoulos J, Curran PJ, Felsch JS, Kalghatgi K, Lee WH,
et al: An affinity selection-mass spectrometry method for the identification of
small molecule ligands from self-encoded combinatorial libraries: discovery
of a novel antagonist of E. coli dihydrofolate reductase. Int J Mass Spectrometry 2004;238:77-83.
Coburn CA, Stachel SJ, Yue-Ming L, Rush DM, Steele TG, Chen-Dodson E,
et al: Identification of a small molecule nonpeptide active site β-secretase inhibitor that displays a nontraditional binding mode for aspartyl proteases. J
Med Chem 2004;47:6117-6119.
Annis DA, Nazef N, Chuang CC, Scott MP, Nash HM: A general technique to
rank protein-ligand binding affinities and determine allosteric vs. direct binding site competition in compound mixtures. J Am Chem Soc 2004;126:1549515503.
Makara GM, Nash H, Zheng Z, Orminati JA, Wintner EA: A reagent-based
strategy for the design of large combinatorial libraries: a preliminary experimental validation. Mol Diversity 2003;7:3-14.
Visscher FE: Scopolamine lower-alkyl halide therapeutic composition [patent no. 2,753,288]. Washington, DC: US Patent Office, 1956.
Lambrecht G, Feifel R, Wagner-Roder M, Strohmann C, Zilch H, Tacke R,
et al: Affinity profiles of hexahydro-sila-difenidol analogues at muscarinic receptor subtypes. Eur J Pharmacol 1989;168:71-80.
Ashkenazi A, Peralta G: Muscarinic acetylcholine receptors. In Peroutka SJ
(ed): Handbook of Receptors and Channels: G Protein Coupled Receptors.
Boca Raton, FL: CRC Press, 1994.
Tessier DC, Thomas DY, Khouri HE, Laliberte F, Vernet T: Enhanced secretion from insect cells of a foreign protein fused to the honeybee melittin signal
peptide. Gene 1991;98:177-183.
Weill C, Autelitano F, Guenet C, Heitz F, Goeldner M, Ilien B: Pharmacological and structural integrity of muscarinic M2 acetycholine receptors produced
in Sf9 insect cells. Eur J Pharmacol 1997;333:269-278.
Kubo T, Maeda A, Sugimoto K, Akiba I, Mikami A, Takahashi H, et al: Primary structure of porcine cardiac muscarinic acetylcholine receptor deduced
from cDNA sequence. FEBS Lett 1986;209:367-372.
Peterson G, Schimerlik M: Large scale preparation and characterization of
membrane-bound and detergent-solubilized muscarinic acetylcholine receptor from pig atria. Preparative Biochem 1984;14:33-74.
Smisterova J, Ensing K, Zeeuw R: Evaluation of a purification procedure for
the muscarinic receptor for the purpose of quantitative receptor assays of
anticholinergics. Preparative Biochem 1995;25:197-221.
Loukas S, Mercouris M, Panetsos F, Zioudrou C: Purification to homogeneity
of an active opioid receptor from rat brain by affinity chromatography. Proc
Natl Acad Sci USA 1994;91:4574-4578.
Niznik H, Otsuka N, Dumbrille-Ross A, Grigoriadis D, Tirpak A, Seeman P:
Dopamine D1 receptors characterized with [3H]SCH 23390: solubilization of
a guanine nucleotide-sensitive form of the receptor. J Biol Chem
1986;261:8397-8406.
Senogles SE, Amlaiky N, Falardeau P, Caron MG: Purification and characterization of the D2-dopamine receptor from bovine anterior pituitary. J Biol
Chem 1988;263:18996-19002.
Doods HN, Entzeroth M, Ziegler H, Mayer N, Holzer P: Pharmacological
profile of selective muscarinic receptor antagonists on guinea-pig ileal
smooth muscle. Eur J Pharmacol 1994;253:275-281.
Baskin SI, Thomsen RH: The effects of AF-DX-116, a cardioselective
muscarinic antagonist, on the negative inotropic action of acetylcholine. Res
Commun Chem Pathol Pharmacol 1991;71:3-16.
Journal of Biomolecular Screening 11(2); 2006
Downloaded from jbx.sagepub.com at PENNSYLVANIA STATE UNIV on September 17, 2016
Muscarinic M2 Acetylcholine Receptor Ligands
31.
32.
33.
34.
Kostenis E, Conklin BR, Wess J: Molecular basis of receptor/G protein coupling selectivity studied by coexpression of wild type and mutant m2
muscarinic receptors with mutant G alpha(q) subunits. Biochemistry
1997;36:1487-1495.
Stengel PW, Gomeza J, Wess J, Cohen ML: M2 and M4 receptor knockout
mice: muscarinic receptor function in cardiac and smooth muscle in vitro. J
Pharmacol Exp Ther 2000;292:877-885.
Bender W, Staudt M, Tränkle C, Mohr K, Holzgrabe U: Probing the size of a
hydrophobic binding pocket within the allosteric site of muscarinic acetylcholine M2-receptors. Life Sci 2000;66:1675-1682.
Tränkle C, Weyand O, Voigtländer U, Mynett A, Lazareno S, Birdsall NJM,
et al: 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.
Journal of Biomolecular Screening 11(2); 2006
Downloaded from jbx.sagepub.com at PENNSYLVANIA STATE UNIV on September 17, 2016
Address reprint requests to:
Charles E. Whitehurst, PhD
Schering-Plough Research Institute
840 Memorial Drive
Cambridge, MA 02139
E-mail: [email protected]
www.sbsonline.org 207