Fluorescence resonance energy transfer to probe human M1 muscarinic receptor structure and drug binding properties Brigitte Ilien,* Christelle Franchet,* Philippe Bernard,�,1 Se´verine Morisset,*,2 Claire Odile Weill,*,3 Jean-Jacques Bourguignon,� Marcel Hibert� and Jean-Luc Galzi* *De´partement Re´cepteurs et Prote´ines Membranaires, CNRS UPR 9050, Illkirch, France �Laboratoire de Pharmacochimie de la Communication Cellulaire, UMR CNRS/ULP 7081, IFR 85, Illkirch, France Abstract Human M1 muscarinic receptor chimeras were designed (i) to allow detection of their interaction with the fluorescent ant- agonist pirenzepine labelled with Bodipy [558/568], through fluorescence resonance energy transfer, (ii) to investigate the structure of the N-terminal extracellular moiety of the receptor and (iii) to set up a fluorescence-based assay to identify new muscarinic ligands. Enhanced green (or yellow) fluorescent protein (EGFP or EYFP) was fused, through a linker, to a receptor N-terminus of variable length so that the GFP barrel was separated from the receptor first transmembrane domain by six to 33 amino-acids. Five fluorescent constructs exhibit high expression levels as well as pharmacological and func- tional properties superimposable on those of the native receptor. Bodipy-pirenzepine binds to the chimeras with similar kinetics and affinities, indicating a similar mode of interaction of the ligand with all of them. From the variation in energy transfer efficiencies determined for four different receptor-ligand complexes, relative donor (EGFP)-acceptor (Bodipy) distances were estimated. They suggest a compact architecture for the muscarinic M1 receptor amino-terminal domain which may fold in a manner similar to that of rho- dopsin. Finally, this fluorescence-based assay, prone to min- iaturization, allows reliable detection of unlabelled competitors. Keywords: assay miniaturization, bodipy-pirenzepine, enhanced green (yellow) fluorescence protein, fluorescence resonance energy transfer, hM1 muscarinic receptor. J. Neurochem. (2003) 85, 768–778. Fundamental understanding of receptor function and discov- ery of new medicines require efficient procedures for analysing the dynamics of both ligand–receptor interactions and subsequent functional responses. With the advent of very sensitive fluorescence technol- ogies (Hovius et al. 2000), new approaches to study G-protein-coupled receptor (GPCR) structure (Chollet and Turcatti 1999), trafficking (Kallal and Benovic 2000; Gicquiaux et al. 2002) and signalling (Milligan 1999), as well as GPCR-protein (Angers et al. 2000; Rios et al. 2001) and GPCR-ligand (Chollet and Turcatti 1999; Vollmer et al. 1999; Ghanouni et al. 2001; Palanche´ et al. 2001; Valen- zuela-Fernandez et al. 2001) interactions, have been devel- oped. In these studies, we showed that functional EGFP-tagged NK2 and CXCR4 receptors allow real-time recordings of ligand binding in living cells, through fluorescence resonance energy transfer (FRET) with a fluorescent agonist. The location, at the extracellular membrane surface, of the tachykinin or chemokine binding domain in these receptors (Class Ib of GPCRs; Bockaert and Pin 1998) is certainly a favourable parameter for measurement of efficient energy transfer. Received September 2, 2002; revised manuscript received December 18, 2002; accepted January 23, 2003. Address correspondence and reprint requests to Dr Brigitte Ilien, De´partement Re´cepteurs et Prote´ines Membranaires, CNRS UPR 9050, Ecole Supe´rieure de Biotechnologie de Strasbourg, Boulevard Se´bastien Brant, 67412 Illkirch, France. E-mail:
[email protected] 1The present address of Philippe Bernard is GreenPharma S.A., 3 alle´e du Titane, 45100 Orleans, France. 2The present address of Se´verine Morisset is Centre Paul Broca, INSERM U109, 2ter, rue d’Ale´sia, 75014 Paris, France. 3The present address of Claire Weill is CEA-Grenoble, DRDC/CBCRB, 17 avenue des Martyrs, 38054 Grenoble, France. Abbreviations used: Bo-NKA, Bodipy-neurokinin A; Bo-PZ, Bodipy- pirenzepine; EGFP (EYFP), enhanced green (yellow) fluorescent protein; FRET, fluorescence resonance energy transfer; GPCR, G protein-coupled receptor; HEK cells, human embryonic kidney cells; mAChR, muscarinic acetylcholine receptor; NKA, neurokinin A; NMS, N-methyl scopolam- ine; PZ, pirenzepine; QNB, quinuclidinyl benzilate; wt, wild-type. Journal of Neurochemistry, 2003, 85, 768–778 doi:10.1046/j.1471-4159.2003.01717.x 768 � 2003 International Society for Neurochemistry, J. Neurochem. (2003) 85, 768–778 This work is devoted to muscarinic acetylcholine receptors (mAChRs; M1 to M5 subtypes) which are physiologically important members of the GPCR superfamily (Caulfield 1993) and which, together with the other bioamine receptors and rhodopsin, belong to Class Ia receptors (Bockaert and Pin 1998). All these receptors share a unique property, i.e. a ligand-binding pocket deeply buried within the transmem- brane core of the protein. Taking advantage of (i) the possibility of fusing EGFP to the human M1 muscarinic receptor N-terminus without any detectable alterations in drug binding or functional properties (Weill et al. 1999a,b) and (ii) the commercial availability of a fluorescent derivative (Bodipy [558/568]) of the M1-select- ive antagonist pirenzepine, we decided to examine whether such a buried location of the binding domain was compatible with FRET monitoring of receptor–ligand interactions. Special attention was paid to spectral properties of donor (EGFP or EYFP) and acceptor (Bodipy group of pirenzepine) partners, and to separation between GFP and hM1 first transmembrane helix (six to 33 residues) as critical para- meters for FRET efficacy. A second aspect of this work dealt with the molecular architecture of the receptor amino-terminal tail. Indeed, X-ray crystallography of bovine rhodopsin (Palczewski et al. 2000) describes this domain as a highly organized structure and raises the question of its possible conservation throughout GPCRs, especially in bioamine receptors which display close structural and functional homology with rhodopsin (Bockaert and Pin 1998; Okada et al. 2001). For this purpose, we fused EGFP to an hM1 receptor N-terminus of variable length (three to 20 residues), determined donor-acceptor separation from FRET efficiency for Bodipy-pirenzepine (Bo-PZ) binding to each receptor chimera, and compared their variation with those theoretically expected from the suppression of corres- ponding peptide stretches in an extended conformation. Finally, we show that real-time monitoring, on living cells, of Bo-PZ binding to chimeric receptors is saturable and reversible, and we present this FRET procedure as an interesting alternative to radioligand binding assays, prone to miniaturization with potential application to drug discovery. Experimental procedures Materials [3H]Quinuclidinyl [phenyl-4–3H] benzilate (49 Ci/mmol; [3H]QNB) and [N-methyl-3H] scopolamine (78 Ci/mmol; [3H]NMS) were from Amersham Pharmacia Biotech (Piscataway, NJ, USA). Atropine sulphate, pirenzepine dihydrochloride, McN-A-343 chloride, oxo- tremorine sesquifumarate and carbachol chloride were purchased from Sigma-Aldrich (St Louis, MO, USA). Indo-1 acetoxymethyl ester and Bodipy [558/568] pirenzepine dihydrochloride were from Molecular Probes (Eugene, OR, USA). cDNA constructs, expression and selection The wild-type hM1 receptor gene, as well as EGFPL-hM1 and EGFPL-(D11)hM1 constructs (previously referred to as Chim A2 and Chim A1, respectively), were obtained and inserted as a NotI– Xho1 fragment into the Bluescript KS+ vector (Stratagene, La Jolla, CA, USA), as described (Weill et al. 1999a,b). The EYFPL-hM1 receptor construct was designed by substituting the EGFP sequence of EGFPL-hM1 for the EYFP sequence (taken from pEYFP-C3, Clontech). Fusion of the cDNA encoding the 31 amino-acid-long peptide (chicken alpha 7 nicotinic signal peptide) with the EGFP (residue position 4) or the EYFP (residue position 1) sequence was achieved through a BsrGI restriction site (Weill et al. 1999a). In these chimeras (Fig. 1), EGFP (or EYFP) was fused, via a 13 amino-acid sequence (TLGMDELYKYSDL; long linker), to the hM1 or (D11)hM1 receptor N-terminus through a BglII restriction site. Obtention of EGFPS-wild-type (wt) hM1 and EGFPS-(D17)hM1 constructs (Fig. 1) involved several steps: (i) linker shortening (SDL; short linker) was achieved through introduction of a BglII site, upstream to the pre-existing site in EGFPL-(D11)hM1, by PCR amplification (Elongase Amplification System; Gibco, Rock- ville, MD, USA) of a new NotI-BglII fragment (signal peptide- EGFPS sequence); (ii) introduction of a BglII site at the 5¢- end of the wthM1 gene, with or without concomitant deletion of 17 residues from the receptor N-terminus, was achieved through PCR amplification of a BglII-XhoI segment using appropriate primers and wthM1-KS as the template; (iii) ligation of the NotI-BglII insert with a BglII-XhoI insert (encoding for the wthM1 or for the (D17)hM1 sequence) into a KS vector led to the EGFPS-wthM1 and EGFPS-(D17)hM1 receptor constructs, respectively. Fig. 1 Alignment of sequences connecting GFP to the hM1 receptor N-terminus. EGFP (or EYFP) last residue (I230) was fused through a 3–13 amino-acid-long linker (sequence under brackets) to the receptor N-terminus (bold characters) at residue position N2, I13 or G19. First residues of the TM1 transmembrane domain are underlined. FRET monitoring of BoPZ binding to M1 receptors 769 � 2003 International Society for Neurochemistry, J. Neurochem. (2003) 85, 768–778 All final constructs were subcloned into the pCEP4 (Invitrogen, Carlsbad, CA, USA) vector for expression in eukaryotic cells and sequenced before use. HEK 293 cells, grown in minimal essential medium (GibcoBRL) complemented with 10% fetal calf serum, 2 mM glutamine and antibiotics, were transfected by calcium phosphate precipitation and selected by 400 lg/mL hygromycin-B (Calbiochem–Novabiochem, San Diego, CA, USA). Radioligand binding and intracellular calcium release [3H]-QNB and [3H]-NMS binding were performed at 37�C (1 h) on intact HEK 293 cells, suspended in HEPES–bovine serum albumin (BSA) buffer (10 mM HEPES, 137.5 mM NaCl, 1.25 mM MgCl2, 1.25 mM CaCl2, 6 mM KCl and 10 mM glucose, pH 7.4; supple- mented with 0.1 mg/mL BSA) as previously reported (Weill et al. 1999a). Receptor quantification and competition experiments were performed at a fixed (250 pM) [3H]-QNB concentration, while saturation experiments were performed using increasing (20– 300 pM) radioligand concentrations. Agonist-induced intracellular calcium release was monitored at 37�C on Indo 1-loaded HEK cells suspended in HEPES–BSA buffer. Following excitation at 338 nm, fluorescence of Ca2+–Indo 1 complexes was recorded at 400 nm as previously described (Weill et al. 1999a). Spectroscopy UV-visible absorbance spectroscopy measurements were made using a Cary 1E (Varian, Les Ulis, France) spectrophotometer at ambient temperature. Fluorescence from cell suspensions (placed in a 1 mL quartz cuvette with magnetic stirring and maintained at 21�C in a thermostatted cuvette holder) was recorded on a Fluorolog 2 (SPEX, Jobin Yvon Horiba, Longjumeau, France) spectrofluorim- eter equipped with a 450 W Xe lamp, a double-grating excitation monochromator and a single-grating emission monochromator. Data were stored using the DM3000 software provided with the spectrofluorimeter. Real-time fluorescence monitoring of ligand–receptor interactions Experiments were performed on cells stably expressing the different chimeric receptors or on non-transfected (control) HEK 293 cells, suspended in HEPES–BSA buffer (typically at 3 · 106 cells/mL). Time-based recordings of the fluorescence emitted at 510 nm (excitation at 470 nm; EGFP-expressing cells) or at 530 nm (excitation at 510 nm; EYFP-expressing cells) were performed at 21�C and sampled every 0.2–3 s, depending on experiments. Fluorescence binding measurements were initiated by adding Bo-PZ, at different concentrations, to the 1 mL cell suspension. For competition experiments, EGFPS-(D17)hM1-expressing cells were pre-incubated for 3 min in the absence or presence of various concentrations of unlabelled drugs. Bo-PZ (80 nM) was added and fluorescence was recorded until equilibrium was reached (1000 s). Competition assays on EGFP-fused rat NK2 receptors using Bodipy [530/550]-labelled Neurokinin A (Bo-NKA) at 50 nM were performed as previously described (Vollmer et al. 1999). Data were analysed using Sigma plot 4.16 (SPSS Science, Chicago, IL, USA) and Kaleidagraph 3.08 software (Synergy Software, Reading, PA, USA). Estimation of donor-acceptor distances The distance R (A˚) between the EGFP (EYFP) fluorophore and the Bodipy moiety of bound pirenzepine was calculated using Fo¨rster’s (1948) equation: R ¼ R0(1/E)1)1/6, where R0 is the distance for 50% transfer efficiency E. R0 was calculated from: R0 ¼ 9790 (j2n)4FDJ)1/6. j2 is a geometric factor that accounts for the relative orientation in space of the donor emission and acceptor absorption transition dipoles (Lakowicz 1999). N, the refractive index of the medium, has been taken to be 1.4. The quantum yield FD of the donor (0.66 for EGFP and 0.63 for EYFP) was taken from the literature (Tsien 1998). J, the spectral overlap integral for the combined emission of EGFP (or EYFP) and absorbance of Bodipy, was calculated as previously described (Vollmer et al. 1999). The efficiency of fluorescence energy transfer (E) was defined as the fractional decrease in EGFP (EYFP) fluorescence due to Bo-PZ binding and was expressed by: E ¼ 1 ) FDA/FD, where FDA and FD are specific donor fluorescence emission in the presence or absence of saturating (200 nM) concentrations of ligand, respectively. Specific EGFP (or EYFP) fluorescence in the absence (FD) or presence (FDA) of ligand was determined by subtracting autoflu- orescence of non-transfected HEK cells (in the absence or presence of ligand) from total fluorescence of hM1 chimera- expressing cells. Assay miniaturization and drug discovery EGFPS-(D17)hM1-expressing cells, suspended in HEPES–BSA buffer, were distributed (90 000 cells/192 lL/well) using a Biomek 2000 robot (Beckman Coulter, Villepinte, France) into 96-well polystyrene plates (Cliniplates; Thermo Labsystems, Issy-les- Moulineaux, France) previously filled (4 lL/well) with Bo-PZ (80 nM final concentration) and a compound to be tested (at 10 lM). After 15 min at room temperature, fluorescence of cells was recorded at 510 nm (excitation at 465 nm) using a multi-label counter (Victor 2; BD Biosciences, Le Pont de Claix, France). Total FRET amplitude was defined as the difference between fluorescence measured in the presence or the absence of atropine (100 nM). Drug inhibition was calculated as the decrease (in percent) in total FRET amplitude. Miniaturized FRET assays on EGFP-rNK2 receptors proceeded similarly, with the exception that receptor-expressing cells were incubated on ice with 50 nM Bo-NKA prior well-sampling (70 000 cells/well). After 10 min of incubation at room temperature, total FRET amplitude was determined as the difference between fluorescence measured in the presence or absence of the NK2 antagonist SR48968 (1 lM). A total of 2000 compounds from the Chemical Library of the School of Pharmacy of Strasbourg (IFR 85) was screened in this way. In addition, a reference plate containing 70 different compounds of known pharmacological properties was designed to test the FRET assay for its screening safety. Selected molecules are known to interact as either agonists or antagonists with various receptors (a- and b-adrenergic, adenosine, peptidergic such as endothelin or somatostatin, vasoactive intestinal peptide, P2Y purinergic, prosta- noid, etc.) that are endogenously expressed in HEK 293 cells (see http://www.tumor-gene.org/GPCR/gpcr.html for a list) or to interfere with G-protein coupling, cyclic nucleotide and calcium signalling, lipid or nitric oxide pathways, phosphorylation or ionic channel modulation, or to exert detergent effects. 770 B. Ilien et al. � 2003 International Society for Neurochemistry, J. Neurochem. (2003) 85, 768–778 Results Design of potential FRET partners Four EGFP-hM1 chimeras were constructed by fusing EGFP (last amino-acid residue I230), via a long (EGFPL) or a short (EGFPS) linker, to the hM1 receptor N-terminus (Fig. 1). The receptor sequence was either identical to that of the wthM1 receptor (EGFPS-wthM1), mutated (S4A, P11D and N12L point mutations; Weill et al. 1999a) for suppression of potential glycosylation sites (EGFPL-hM1), or truncated by 11 (EGFPL-(D11)hM1; Weill et al. 1999b) or 17 (EGFPS- (D17)hM1) amino-acids. Thus, the connecting sequence between the EGFP’s last residue (I230) and the receptor first TM1 residue (P22) varied from six to 33 residues, allowing us to explore the impact of EGFP proximity on FRET parameters. Bodipy [558/568]-pirenzepine (Bo-PZ; Fig. 2a), a fluor- escent derivative of the high affinity M1 muscarinic antag- onist, exhibits several interesting properties. The Bodipy moiety displays high quantum yield (close to 0.7) and extinction coefficient (75 000 cm)1 M)1), is not sensitive to medium polarity or pH variation, and presents spectral properties (558 nm and 568 nm maximal excitation and emission wavelengths, respectively) appropriate to its use as the energy transfer acceptor. The spectral integral overlap (J ) between EGFP emission (maximal at 509 nm) and Bo-PZ absorbance spectra was calculated to be 1.5 · 10)13 cm3 M)1 (see Experimental procedures). When EYFP (514 nm and 527 nm maximal excitation and emission wavelengths, respectively) is used as energy donor instead of EGFP, the spectral overlap (J ¼ 2.54 · 10)13 cm3 M)1) is larger, as expected from spectra shown in Fig. 2b. Expression and pharmacological characterization of five fluorescent hM1 chimeras Muscarinic receptor expression was quantified by specific [3H]QNB binding on whole HEK 293 cells, either transiently or stably expressing the different recombinant cDNAs (Table 1). Transient expression levels of the different muscarinic plasmids substantially increased upon selection with hygro- mycin-B. Shortening of the receptor N-terminus in EGFPL- (D11)hM1 was accompanied by a twofold decrease in expression, while suppression of 10 residues in the EGFP- receptor linker led to a substantial increase (EGFPS-wthM1). All five chimeras displayed similar drug-binding affinity properties, as assessed from [3H]QNB saturation and com- petition experiments (Table 1). Interestingly, the introduct- ion, via a spacer, of the Bodipy [558/568] group into pirenzepine leads to a slight increase in affinity of the compound. Bo-PZ interacts with a homogeneous population of receptor sites as assessed from slope factors close to unity. High affinity agonist binding to mAChRs on intact cells was estimated from competition experiments using [3H]NMS (at 37�C) or [3H]QNB (at 12�C) as radioligands (Weill et al. 1999a). Ki-values for McN-A-343, competing with either [3H]NMS on EGFPL-(D11)hM1 (3.5 lM) receptors or [3H]QNB on EGFPS-wthM1 (26 lM) and on EGFPS- (D17)hM1 (30 lM) chimeras, are comparable with those previously reported (Weill et al. 1999a) for this partial agonist on the wthM1 receptor (2.5 and 15 lM for [3H]NMS and [3H]QNB binding, respectively). Agonist-induced intracellular calcium release assays were performed using McN-A-343, which allowed selective activation of M1 receptors on HEK 293 cells (Weill et al. 1999a). On stimulation with this agonist, each receptor chimera leads to a rapid and dose-dependent Ca2+ response with EC50-values close to that reported for McN-A-343 at the wthM1 receptor (EC50 ¼ 0.5 ± 0.1 lM), expressed under similar conditions (Weill et al. 1999a). Detection of receptor–ligand interactions by FRET on living cells FRET experiments were carried out by exciting a cell suspension at EGFP maximal excitation wavelength 450 500 550 600 0.0 0.2 0.4 0.6 0.8 1.0 1.2 (b) (a) N or m al iz ed fl uo re sc en ce Wavelength (nm) N HN N O N N N H N B N S O F F O Bodipy [558/568] Pirenzepine Fig. 2 (a) Chemical structure of Bodipy [558/568] pirenzepine. (b) Fluorescence spectral properties for donor and acceptor pairs. Normalized emission spectra of EGFP- (d, excitation at 470 nm) and EYFP- (s, excitation at 495 nm) fused receptors are shown together with absorbance (solid line) and emission (dashed line; excitation at 485 nm) spectra for Bo-PZ. Fluorescence of cell suspensions was corrected for background fluorescence from non-transfected cell samples (Experimental procedures). FRET monitoring of BoPZ binding to M1 receptors 771 � 2003 International Society for Neurochemistry, J. Neurochem. (2003) 85, 768–778 (470 nm) and by recording fluorescence emission spectra before and after addition of Bo-PZ. Figure 3(a) shows that addition of 100 nM Bo-PZ to a EGFPS-(D17)hM1 cell suspension leads to a decrease in fluorescence emission (around 35% at 510 nm). Such a phenomenon is not observed in non-transfected HEK cells. When autofluores- cence of HEK cells is subtracted from the signal of transfected cells, specific EGFP fluorescence amplitude of EGFPS-(D17)hM1 cells diminishes by 48% upon Bo-PZ addition. In a continuous recording mode (Fig. 3b), a time-depend- ent reduction in fluorescence emission at 510 nm clearly starts with the addition of Bo-PZ to each EGFP-fused chimera. No fluorescence extinction is observed when Bodipy alone, Bodipy [530/550]-Neurokinin A (a fluorescent agonist of the NK2 receptor; Vollmer et al. 1999) or unlabelled pirenzepine is added to the cell suspension (not shown). Moreover, the contribution of receptor internalizat- ion to fluorescence variations can be completely ruled out for several reasons: (i) all experiments were performed at 21�C, Time (sec) Fl uo re sc en ce d ec re as e (% ) 500 550 600 0.0e+00 5.0e+04 1.0e+05 1.5e+05 (a) (b) (c) Fl uo re sc en ce e m is si on (c ps ) Wavelength (nm) 0 200 400 600 40 20 0 EGFPS(∆17)hM1 EGFPS wt hM1 EGFPL(∆11)hM1 Control 0 200 400 600 0 20 40 60 80 100 N or m al iz ed B o- PZ a ss oc ia tio n (% ) Time ( sec) EGFPS(∆17)hM1 EGFPS-wt hM1 EGFPL(∆11)hM1 Fig. 3 Spectral and kinetic detection of Bo-PZ interaction with chimeric muscarinic receptors. (a) Fluorescence emission spectra (excitation at 470 nm) of non-transfected (dots) and of EGFPS- (D17)hM1-expressing (circles) cells were recorded before (open symbols) or 10 min after (closed symbols) addition of 100 nM Bo-PZ to the cell suspension. (b) Continuous recordings of fluorescence (at 510 nm) of cells (3 · 106 cells/mL) stably expressing various EGFP-fused receptors. Bo-PZ (100 nM) or vehicle (DMSO; EGFPS- (D17)hM1 cells; control) were added at time 50 s (arrow). (c) Super- imposition of association traces for Bo-PZ binding to three EGFP-fused receptor chimeras. Binding traces in (b) were normalized with 100% representing maximal fluorescence extinction for each chimera at time 600 s. Bo-PZ (100 nM) was at time 0. Averaged trace best fit for a two- exponential variation of fluorescence is shown (white line). Table 1 Pharmacological parameters for chimeric hM1 muscarinic receptors Receptor Expression Binding affinity parameters Ca2+ response Transient (fmol/106 cells) Stable (fmol/106 cells) [3H]QNB (Kd, pM) Atropine (Ki, nM) Pirenzepine (Ki, nM) Bo-PZ (Ki, nM) McN-A-343 (EC50, lM) EYFPL-hM1 240 644 ± 24 55 1.3 49 31 (0.95) 0.3 (n ¼ 6) EGFPL-hM1 170 a 588 ± 40 61 ± 11a 2.7 ± 0.6a 32.1 ± 0.4a 22 ± 8 (1.02) 0.7 ± 0.2a (n ¼ 4) (n ¼ 4) (n ¼ 3) (n ¼ 3) (n ¼ 2) (n ¼ 3) EGFPL-(D11)hM1 83 ± 9 315 ± 4 57 ± 7 1.7 ± 0.2 57 ± 8 13.5 ± 4.5 (0.97) 0.45 (n ¼ 4) (n ¼ 5) (n ¼ 2) (n ¼ 3) (n ¼ 2) (n ¼ 4) EGFPS-wthM1 470 1850 ± 140 65 ± 5 2.3 ± 0.6 47 ± 8 10.3 ± 0.9 (1.15) 0.71 (n ¼ 3) (n ¼ 2) (n ¼ 2) (n ¼ 2) (n ¼ 2) EGFPS-(D17)hM1 100 640 ± 70 50 ± 5 1.6 ± 0.4 23.5 ± 0.8 8.8 ± 2.7 (0.98) 0.75 (n ¼ 3) (n ¼ 2) (n ¼ 3) (n ¼ 2) (n ¼ 3) Cell receptor quantification was performed before and after hygromycin selection. Kd-values were from Scatchard analyses of saturation data. Ki-values of drugs were determined from IC50-values corrected for radioligand concentration and affinity constant. Mean slope factors for Bo-PZ competition curves are given in italics. EC50-values for McN-A-343 were from non-linear regression analyses of dose–response curves. Data are mean ± SEM values for n independent experiments or mean ± SD values for two separate experiments. aValues are taken from Weill et al. (1999a). 772 B. Ilien et al. � 2003 International Society for Neurochemistry, J. Neurochem. (2003) 85, 768–778 a temperature low enough to prevent receptor internalization; (ii) internalization is an agonist-promoted event not usually triggered by antagonists; and (iii) even at 37�C, a 15 min exposure of the chimera-expressing cells to 100 lM carba- chol did not lead to any detectable decrease in EGFP fluorescence at 510 nm. After a 600 s incubation period, Bo-PZ binding to the EGFPL-(D11)hM1, EGFPS-wthM1 and EGFPS-(D17)hM1 receptor constructs results in an 11%, 15% and 37% reduction in overall fluorescence amplitude, respectively. Under similar conditions, fluorescence extinction of EGFPL- hM1 (7%) and EYFPL-hM1 (10%; monitored at 530 nm with excitation set at 510 nm) chimeras is lower (not shown). Finally, association traces from Fig. 3(b) can be normal- ized according to maximal value at time 600 s to show a nice superimposition of their time-courses (Fig. 3c). Half-time values (s ± SE) for Bo-PZ (100 nM) association to the various chimeras do not differ significantly at 105.3 ± 1.3 [EGFPL-(D11)hM1], 103.5 ± 0.7 [EGFPS-(D17)hM1] and 104.8 ± 1.1 [EGFPS-wthM1], and the averaged trace is best fit (white line) with a two-exponential variation of fluorescence, including fast (kapp value: 0.030 ± 0.004 s )1) and slow (kapp value: 0.0051 ± 0.0001 s )1) Bo-PZ binding components. Examination of Bo-PZ binding properties at EGFPS-(D17)hM1 receptors In order to monitor Bo-PZ binding in real-time at 21�C, we selected cells expressing the EGFPS-(D17)hM1 chimera since they give the most robust FRET signal. Figure 4(a) illustrates three typical features of ligand– receptor interactions. Upon addition of Bo-PZ (100 nM) to EGFPS-(D17)hM1 cells, fluorescence decreases over time and reaches a plateau (42% fluorescence extinction) after about 15 min, reflecting association and binding equilibrium. Bo-PZ binding is fully prevented by atropine (1 lM) added before the fluorescent ligand. Moreover, when added at Bo-PZ equilibrium, atropine (5 lM) induces a slow (half- time value: 26 min) fluorescence recovery, reflecting Bo-PZ dissociation from the receptor. Four independent experiments, performed on a 5000 s time-range, lead to the observation of a biphasic dissociation time-course and to the determination of mean values (± SEM; n ¼ 4) for a fast (2.0 ± 0.4 · 10)3 s)1) and a slow (3.5 ± 0.3 · 10)4 s)1) dissociation rate constant. Similar findings are obtained using the other receptor chimeras (not shown). 0 250 500 750 15000 20000 25000 (a) (b) (c) (d) + atropine 2 1 0 2000 4000 6000 Time (sec) Fl uo re sc en ce a t 5 10 nm (c ps ) 11 0 1000 10 20 30 40 [Bo-PZ] , nM FR ET a m pl itu de (% ) -9 -8 -7 -6 -5 -4 -3 -2 -1 0 20 40 60 80 100 R es id ua l F RE T (% ) [Competitor] , logM 0 250 500 750 1000 1250 2500 60 80 100 Fl uo re sc en ce a t 5 10 nm (% ) Time (sec) 300 200 100 50 25 15 8 4 2 Fig. 4 FRET monitoring of Bo-PZ binding properties to EGFPS- (D17)hM1 receptors. (a) Bo-PZ (100 nM) was added at time 50 s (arrow 1) to a 3 · 106 cells/mL suspension and fluorescence was recorded (one point per 0.3 s) until equilibrium. Thereafter, atropine (5 lM) was added (arrow 2) in order to initiate the dissociation step (one point recorded every 3 s). In a control experiment (upper trace), atropine (1 lM) was added prior to Bo-PZ. (b) Time-recordings of the reduction in EGFPS-(D17)hM1 fluorescence amplitude at various Bo-PZ concentrations. The ligand was added at time 0 to a 3 · 106 cells/mL suspension. For Bo-PZ concentrations below 200 nM, after an overall incubation time of 40 min, fluorescence recordings of the last 60 s are shown to verify real equilibrium levels. (c) Occupancy curve for Bo-PZ binding to EGFPS-(D17)hM1 receptors. Amplitudes (mean values in percent ± SD) for fluorescence extinction at equilib- rium obtained in four independent experiments, conducted as in (b), are plotted as a function of Bo-PZ concentration. Best fit to the em- pirical Hill equation derived for saturation is drawn. (d) Competition of muscarinic antagonists and agonists with Bo-PZ for binding to the EGFPS-(D17)hM1 chimera. Various concentrations of atropine (s), pirenzepine (d), oxotremorine (h) and carbachol (j) or vehicle (wa- ter; r) were added to the cell suspension 3 min before Bo-PZ (80 nM). Fluorescence was recorded during 1000 s. Residual FRET amplitude at equilibrium (percent of control in the absence of competitor) is plotted as a function of drug concentration. FRET monitoring of BoPZ binding to M1 receptors 773 � 2003 International Society for Neurochemistry, J. Neurochem. (2003) 85, 768–778 A plot of EGFPS-(D17)hM1 emission intensity versus Bo-PZ concentration (Fig. 4b) shows that fluorescence extinction increases with Bo-PZ concentration and reaches a maximum at 200 nM Bo-PZ. Steady-state fluorescence emission, i.e. binding equilibrium, is reached at all ligand concentrations, providing the incubation period is long enough. Though the fluorescence decay appears to be biphasic, especially at high ligand concentrations (100– 300 nM), control experiments allow the exclusion of inter- ference by cell settling during recordings. Indeed, cell suspension fluorescence remains remarkably stable in the absence of ligand or when atropine is added prior to Bo-PZ (Figs 3b and 4a). Plotting amplitudes (mean values ± SD; n ¼ 4) of fluorescence extinction (at equilibrium) versus Bo-PZ con- centration yields an occupancy curve (Fig. 4c) that best fits with a single category of sites (nH ¼ 1.07 ± 0.13), an apparent KD-value of 13.2 ± 1.3 nM and a maximal FRET amplitude of 44.7 ± 1.5%. Saturation is reached at 200 nM Bo-PZ for all chimeras (not shown), and a very close KD-value (19 ± 4 nM; n ¼ 3) can be estimated for Bo-PZ binding on EGFPL-(D11)hM1 receptors. While incubation temperature (21�C versus 37�C) does not modify FRET amplitudes at equilibrium, Bo-PZ binding at 37�C leads to a slight increase in affinity, with KD-values (nM) of 6.0 and 7.2 for the EGFPS-wthM1 and EGFPS- (D17)hM1 constructs, respectively. Figure 4(d) shows typical competition curves, which were generated using residual FRET amplitudes measured, at equilibrium, after incubation of EGFPS-(D17)hM1 cells in the presence of increasing antagonist or agonist concentra- tions together with Bo-PZ (80 nM). From such an experi- ment, IC50-values can be estimated (atropine: 9.2 nM; pirenzepine: 147 nM; oxotremorine: 4.7 lM and carbachol: 1.07 mM; slope factors were close to unity except for carbachol: 0.65) and corrected for Bo-PZ concentration and affinity to give the corresponding Ki-values (Table 3). Measurements of donor-acceptor separation Fo¨rster radius R0-values (A˚, mean ± SE, n ¼ 4), with a j2 factor set at 2/3 for dynamic random averaging of donor- acceptor transition dipoles (Lakowicz 1999), were calculated to be 53.9 ± 0.2 (EYFP-Bo-PZ pair) and 49.7 ± 0.2 (EGFP- Bo-PZ pair) as described in Experimental procedures. Efficiency of FRET (E) was estimated from the extent of maximal fluorescence extinction determined (i) at equilib- rium (800 s incubation time at 21�C) (ii) at a saturating Bo-PZ concentration (200 nM) and (iii) after correction for cell autofluorescence. Average E-values for the five different fluorescent chimeras are listed in Table 2. Experimental distances between the Bodipy group and the EGFP (or EYFP) fluorophore were then calculated (Table 2). Non-statistically different values are obtained when compar- ing EGFPL- and EYFPL-hM1 receptors, in agreement with the fact that they only differ by a few mutations in the GFP core, leaving an identical position for the fluorophore entity (Yang et al. 1996). Reduction of the linker size per se brings EGFP and Bodipy fluorophores about 6 A˚ closer (EGFPL- hM1 compared with EGFPS-wthM1). Shortening the receptor N-terminus by 11 residues (EGFPL-(D11)hM1) decreased the donor-acceptor distance by 10 A˚ (as compared with EGFPL- hM1), while suppression of 17 residues (EGFPS-(D17)hM1) led to a 15 A˚ reduction in distance (as compared with EGFPS- wthM1). In addition, when the possibility that bound Bodipy exhibits reduced freedom to rotate is taken into consideration (j2 factor equal to 0.476 instead of 2/3; Lakowicz 1999), Table 2 Estimation of donor-acceptor separation by fluorescence resonance energy transfer Receptor Linker length (a.a. number) Receptor N-terminal length (a.a. number) Maximal specific fluorescence extinction (%) FRET efficiency Experimental mean-distance K2 ¼ 2/3 (A˚) Experimental mean-distance K2 ¼ 0.476 (A˚) Long linker EYFPL-hM1 13 20 11.6 ± 1.1 (n ¼ 10) 0.17 ± 0.02 (n ¼ 10) 70.2 ± 1.6 66.3 ± 1.5 EGFPL-hM1 13 20 7.7 ± 0.6 (n ¼ 10) 0.14 ± 0.02 (n ¼ 10) 67.3 ± 1.7 63.6 ± 1.6 EGFPL-(D11)hM1 13 9 12.9 ± 0.3 (n ¼ 20) 0.29 ± 0.02 (n ¼ 20) 57.7 ± 0.9 54.6 ± 0.9 Short linker EGFPS-wthM1 3 20 17.6 ± 0.7 (n ¼ 6) 0.22 ± 0.01 (n ¼ 6) 61.1 ± 0.6 57.8 ± 0.6 EGFPS-(D17)hM1 3 3 44.9 ± 1.0 (n ¼ 9) 0.63 ± 0.01 (n ¼ 9) 45.4 ± 0.3 42.9 ± 0.2 Energy transfer efficiencies (mean-values ± SEM for n independent experiments) are from spectral or kinetic determinations of maximal specific fluorescence extinction obtained with 200 nM Bo-PZ. Interchromophore distances (mean ± SE values) were calculated using Fo¨rster radius values R0 (the dipole j 2 factor set at 2/3 or at 0.476) and experimentally determined mean-FRET efficiencies as detailed in Experimental procedures. 774 B. Ilien et al. � 2003 International Society for Neurochemistry, J. Neurochem. (2003) 85, 768–778 experimentally determined distances exhibit minor changes and relative distance variations remain similar. Thus, upon interacting with Bo-PZ, the four different EGFP-fused hM1 receptors yield an extended range of FRET efficiencies (14–63%), which corresponds to a narrow range of donor-acceptor separation (45–67 A˚). Specificity of hM1 and rNK2 receptors fluorescence assays and search for new compounds Bo-PZ binding to EGFPS-(D17)hM1 receptors and Bo-NKA binding to EGFP-rNK2 receptors (Vollmer et al. 1999) were tested for their specificity, accuracy, use in drug screening and possible miniaturization. Table 3 Specificity of FRET assays at hM1 and rNK2 receptors and search for new biologically-active compounds Ligands hM1R rNK2R [3H]QNB Bo-PZ Bo-NKA Ki (lM) or inhibition, % (conc) Pharmacological agents Muscarinic ligands Atropine 0.0016 ± 0.0004 0.0013 ± 0.0002 0% (1.2 lM) Pirenzepine 0.023 ± 0.001 0.020 ± 0.003 0% (1.2 lM) Oxotremorine n.d. 0.76 ± 0.06 0% (1.2 lM) McN-A-343 30 15 n.d. Carbachol 215 ± 28 162 ± 20 0% (1.2 lM) Neurokinin ligands SR 48968 n.d. 30% (1 lM) 100% (100 nM) NKA-Bodipy (Bo-NKA) n.d. 0% (50 nM) 0.009 NKA n.d. 0% (1 lM) 100% (1 lM) Other compounds Proadifen 0.6 0.3 0% (1.2 lM) Chloroquine 1.4 1.6 0% (1.2lM) Genistein n.d. 0% (10 lM) 18 Selected bis-cations A 0.02 0.03 0% (1.2 lM) B 45% (1 lM) 0.2 0% (1.2 lM) C 0.12 0.74 0% (1.2 lM) D n.d. 0% (10 lM) 1.3 E n.d. 0% (10 lM) 1.3 FRET assays were performed on cell suspensions expressing EGFPS-(D17)hM1 or EGFP-rNK2 receptors, using either Bo-PZ (80 nM) or NKA- Bodipy (50 nM), respectively. Inhibitory potencies of drugs are given as percentages of total FRET amplitude (at a given concentration) or as Ki- values (mean-values ± SD; n ¼ 3) determined from IC50-values of drugs, corrected for tracer concentration and affinity. Symetric bis-cations (R-(CH2)12-R), named by coded letters, are listed together with the structure of their R moiety. FRET monitoring of BoPZ binding to M1 receptors 775 � 2003 International Society for Neurochemistry, J. Neurochem. (2003) 85, 768–778 Table 3 shows that several muscarinic antagonists and agonists compete with Bo-PZ for binding to EGFPS- (D17)hM1 receptors, with affinities (Ki-values) very similar to those determined from [3H]QNB binding (or taken from the literature on the M1 receptor; Richards 1991; Weill et al. 1999a). None of them (at a micromolar concentration) interfere with Bo-NKA binding to EGFP-rNK2 receptors. Conversely, the more potent tachykinin ligands (SR 48968, Bo-NKA and NKA; Vollmer et al. 1999) are inactive or poor inhibitors in the muscarinic FRET assay. Among 70 drugs known as endogenous receptor ligands or signalling modulators (Experimental procedures), only pro- adifen and chloroquine interact with muscarinic receptors at micromolar concentrations. This observation fully agrees with literature reports on these two compounds (Choo et al. 1986; Schmidt and Oettling 1987). In the rNK2 binding assay, genistein (a protein kinase inhibitor) and digitonin were the unique compounds which affected energy transfer between the receptor and Bo-NKA. In 96-well plates, Bo-PZ binding to EGFPS-(D17)hM1 receptors leads to a 30% decrease in total fluorescence emission at 510 nm. Atropine (100 nM) fully abolishes the FRET signal. Similarly, Bo-NKA added to EGFP-NK2- expressing cells leads to a 25% fluorescence extinction, which is fully reversed by SR48969 or unlabelled NKA (at 1 lM). In order to test whether such miniaturized FRET-based binding assays could be developed to screen for new ligands, 2000 compounds of the Chemical Library of the School of Pharmacy were tested at 1.2 lM (rNK2 assay) or at 10 lM (hM1 assay). About 20 positive hits at hM1 or rNK2 receptors were selected, and confirmed by FRET measure- ments (performed in a cuvette) of dose-dependent inhibition. Five of these compounds (A to E), belonging to a chemical class characterized by two heterocycles linked by a dodeca- nyl group, are presented in Table 3. Their various substituted heterocyclic amidines (listed as moieties) are basic (pKA ‡ 10) and exist as cationic entities at physiological pH. Three compounds (A, B and C) display nanomolar to submicromolar affinities for the hM1 receptor, while D and E interact in the micromolar range with the rNK2 receptor. Again, a good agreement is observed for drug inhibitory potencies (or Ki-values) determined either from [ 3H]QNB- or muscarinic FRET-based binding assays. Our results highlight a marked muscarinic or tachykinin receptor selectivity of these molecules, depending on specific structural features of their heterocyclic moieties: the presence of an additional phenyl ring seems detrimental for high affinity binding to M1 receptors, while it represents an apparently favourable factor for binding selectivity to NK2 receptors. Discussion The aim of the present work was to use ligands from the bioamine group in order to detect ligand–receptor interac- tions, characterize their dynamics, investigate the structure of ligand–receptor complexes and identify new receptor lig- ands. So far we have constructed a series of five different fluorescent hM1 muscarinic receptors, carrying EGFP (or its variant EYFP) linked to the N-terminus of the receptor and separated from the first receptor transmembrane domain by six to 33 amino-acids. Consistent with the fact that transmembrane and cytoplas- mic domains of muscarinic receptors are critical for proper ligand recognition, receptor activation and coupling selec- tivity to G-proteins (Wess 1998), we found that most of the N-terminal end of the hM1 receptor can be deleted without significant effects on ligand binding or signalling by agonists, though expression of the protein is reduced by such a deletion. Very few studies have been devoted to the role of the N-tail of GPCRs. Truncation mutants of the amino terminus of the glucagon (Unson et al. 1995) and the l-opioid (Chaturvedi et al. 2000) receptors display proper processing, folding and membrane targeting, though expression levels are affected. Also in agreement with our results, truncated l-opioid receptors remain functional. Interestingly, this latter work (Chaturvedi et al. 2000), together with a study on chimeric m1/m5 muscarinic receptors (Spalding et al. 2002), pointed to the role of the N-terminal tails of both the l-opioid and muscarinic M1 receptors as determinants of binding affinity or activation selectivity of ‘atypical’ agonists. Such a contribution of the M1 receptor N-tail to the composition of an ‘ectopic’ activation site (Spalding et al. 2002) has not been detected in the present study, probably due to the use of classical muscarinic ligands. Combination of Bo-PZ with EGFP- or EYFP-fused muscarinic receptors led to high affinity, reversible and saturable interaction signals with all receptor constructs. FRET signal specificity was verified by drug binding selectivity (muscarinic vs. tachykinin ligands) and by the absence of false positives, such as molecules interfering with the GPCR signalling cascades. Kinetic determination of off-rate constants for Bo-PZ binding, which were in the range of that determined for [3H]pirenzepine binding (3.3 · 10)4 s)1) to brain mem- branes (Potter et al. 1988), indicates the occurrence of a fast and a slow dissociation component. Bo–PZ association with the fluorescent chimeras is probably also complex as it reveals a biphasic time-course taking place over several hundreds of seconds. Further work will be required to characterize these binding components and to test whether they reflect (i) the existence of distinct conformational states of the M1 receptor (with different affinities for Bo-PZ), (ii) interconversion according to a two-step isomerization pro- cess (Ja¨rv et al. 1979; Luthin and Wolfe 1984) involving, or not, ligand translocation from a peripheral to a central receptor site (Jakubik et al. 2000), or (iii) binding of a second 776 B. Ilien et al. � 2003 International Society for Neurochemistry, J. Neurochem. (2003) 85, 768–778 ligand molecule according to the two-site tandem model (Jakubik et al. 2000). To these ends, the use of a non- separative Bo-PZ binding assay with real-time monitoring of binding events and, probably, higher sensitivity towards transient binding states certainly represents an advantage over classical filtration radioligand assays. With regard to FRET efficiencies, we found that the shorter the length of the domain linking EGFP to the hM1 receptor the higher the efficiency of energy transfer. Com- pared with NK2 tachykinin receptor interacting with Bo-NKA (60% EGFP fluorescence extinction, 29 amino- acids separating GFP from receptor TMI; Vollmer et al. 1999), energy transfer taking place between the M1 receptor and Bo-PZ reaches comparable efficiency (63%) only when most of the N-tail of the receptor is deleted (EGFPS- (D17)hM1 construct). This is consistent with the notion that peptides bind to the receptor surface while bioamines interact with a more buried domain. Using rhodopsin-derived models, Bo-PZ docking into the hM1 receptor suggests that the pirenzepine moiety is surrounded by residues identified by site-directed mutagen- esis to contribute to pirenzepine and ACh binding (Bourdon et al. 1997), while the Bodipy group is constrained within a hydrophobic pocket delineated by amino-acids from trans- membrane helices I, II and VII (unpublished data). Although its precise location remains to be verified, i.e. by mutagenesis experiments combined with binding and FRET efficiency measurements, the fluorophore probably elicits a number of additional interactions that may account for an increase in binding affinity of Bo-PZ (compared with pirenzepine alone). FRET monitoring of Bo-PZ binding to the five fluorescent muscarinic receptors yielded comparable equilibrium param- eters and kinetic behaviours whatever the length of the receptor N-terminus. This indicates that Bo-PZ recognizes a binding domain that is common to, and presumably identical in, all constructs and that it probably does not interact with the receptor N-terminus either when bound or during access to its binding site. A surprising finding was that removal of 11–17 residues (i.e. a peptide stretch up to 30–50 A˚ long in an extended structure) from the hM1 receptor N-terminus results in a limited shortening (10–15 A˚) in donor-acceptor distance. As the Bo-PZ binding domain is probably identical in the four fluorescent M1 chimera, and as similar distance variations were measured, assuming dynamic random averaging (j2 ¼ 2/3) or a static range (j2 ¼ 0.476) of donor-acceptor dipole orientations, small Bodipy to EGFP distance varia- tions associated with hM1 N-tail deletions strongly suggest the occurrence of a well-defined folded structure of the receptor amino-terminal domain. Interestingly, atomic reso- lution of the extracellular regions of rhodopsin has shown that its N-terminus and the three exoloops associate to form a compact structure, sitting on the transmembrane core with exoloop 2 folding deeply into the centre of rhodopsin (Palczewski et al. 2000). According to this rhodopsin X-ray structure, we propose that two domains of the hM1 N-terminus (Thr3 to Pro6 and Ala8 to Pro11; wthM1 receptor nomenclature) may correspond to b-strand homologues of the two anti-parallel strands of rhodopsin (Gly3 to Pro7 and Asn8 to Pro12). A single (Pro7) residue is inserted in the presumed hM1 b-turn and should not disrupt the b-sheet fold that may associate the two strands from the N-terminus to two strands from the e2 loop. Finally, miniaturization and automation on multi-well plates of the present muscarinic (or tachykinin) binding assay, and its application to drug screening, is illustrated here by a selection of five compounds belonging to a set of 20 positive hits from 2000 compounds tested on the EGFPS- (D17)hM1 and on the EGFP-rNK2 chimera. These five compounds share a bis-cation-like structure and display altogether nanomolar to micromolar affinities and good selectivities for hM1 versus NK2 receptors. Symmetrical bis-cations deriving from heterocyclic ami- dines exhibit various pharmacological properties, such as enzyme inhibition (Bourguignon 1996; Qin et al. 2000) or ionic channel blocking (Chen et al. 2000). Moreover, many of them are known to interact with mAChRs, as found for other symmetrical bis-quaternary ammonium ligands (Bour- guignon 1996; Holzgrabe and Mohr 1998). The availability of new fluorescent muscarinic ligands (agonists and allosteric effectors) should aid characterization of receptor conformational states, and localization and dissection of their modes of interaction with accessory and/ or modulatory sites (Jakubik et al. 2000; Birdsall et al. 2001) present on the muscarinic M1 receptor protein. Acknowledgements We thank Drs C. Lugnier, N. Frossard, K. Takeda, R. Andria- ntsitohaina, J.-S. Re´my and J. Zwiller for kindly providing pharmacological agents, and Dr J. Garwood for critical reading of the manuscript. Dr B. Didier was very helpful in organizing access to compounds from the Chemical Library of the School of Pharmacy (Strasbourg). This work was supported by the Centre National de la Recherche Scientifique, the Institut National de la Sante´ et de la Recherche Me´dicale, the Universite´ Louis Pasteur, the Association pour la Recherche contre le Cancer (to J-LG and to IFR 85), the Ministe`re de l’Enseignement Supe´rieur et de la Recherche (fellow- ship to CW and funds for screening platform), the Lilly ULP-CNRS convention (fellowship to SM) and the Groupement d’Inte´reˆt Public Hoechst-Marion-Roussel (funds for screening platform). References Angers S., Salahpour A., Joly E., Hilairet S., Chelsky D., Dennis M. and Bouvier M. (2000) Detection of beta 2-adrenergic receptor dime- rization in living cells using bioluminescence resonance energy transfer (BRET). Proc. Natl Acad. Sci. USA 97, 3684–3689. FRET monitoring of BoPZ binding to M1 receptors 777 � 2003 International Society for Neurochemistry, J. Neurochem. (2003) 85, 768–778 Birdsall N. J. M., Lazareno S., Popham A. and Saldanha J. (2001) Multiple allosteric sites on muscarinic receptors. Life Sci. 68, 2517–2524. Bockaert J. and Pin J.-P. (1998) Use of a G-protein-coupled receptor to communicate. A success during evolution. C. R. Acad. Sci. Paris 321, 529–551. Bourdon H., Trumpp-Kallmeyer S., Schreuder H., Hoflack J., Hibert M. and Wermuth C. G. (1997) Modelling of the binding site of the human m1 muscarinic receptor: experimental validation and refinement. J. Comput. Aided Mol. Des. 11, 317–332. Bourguignon J.-J. (1996) Identical and nonidentical twin drugs, in The Practice of Medicinal Chemistry (Wermuth C. G., ed.), pp. 261– 293. Academic Press, London. Caulfield M. P. (1993) Muscarinic receptors—Characterization, coupling and function. Pharmac. Ther. 58, 319–379. Chaturvedi K., Shahrestanifar M. and Howells R. D. (2000) l-Opioid receptor: role for the amino terminus as a determinant of ligand binding affinity. Molec. Brain Res. 76, 64–72. Chen J. Q., Galanakis D., Ganellin C. R., Dunn P. M. and Jenkinson D. H. (2000) bis-Quinolinium cyclophanes: 8,14-diaza-1,7(1, 4)- diquinolinacyclotetradecaphane (UCL 1848), a highly potent and selective, nonpeptidic blocker of the apamin-sensitive Ca(2+)-ac- tivated K(+) channel. J. Med. Chem. 43, 3478–3481. Chollet A. and Turcatti G. (1999) Biophysical approaches to G protein- coupled receptors: Structure, function and dynamics. J. Comput. Aided Mol. Des. 13, 209–219. Choo L. K., Malta E. and Mitchelson F. (1986) Investigation of the antimuscarinic and other actions of proadifen in-vitro. J. Pharm. Pharmacol. 38, 898–901. Fo¨rster T. (1948) Zwischenmolekulare Energiewanderung und Fluores- zenz. Ann. Phys. (Leipzig) 2, 55–75. Ghanouni P., Steenhuis J. J., Farrens D. L. and Kobilka B. K. (2001) Agonist-induced conformational changes in the G-protein-coupling domain of the b2 adrenergic receptor. Proc. Natl Acad. Sci. USA 98, 5997–6002. Gicquiaux H., Lecat S., Gaire M., Dieterlen A., Me´ly Y., Takeda K., Bucher B. and Galzi J.-L. (2002) Rapid internalization and recycling of the human neuropeptide Y Y1 receptor. J. Biol. Chem. 277, 6645–6655. Holzgrabe U. and Mohr K. (1998) Allosteric modulators of ligand binding to muscarinic acetylcholine receptors. DDT 3, 214–222. Hovius R., Vallotton P., Wohland T. and Vogel H. (2000) Fluorescence techniques: shedding light on ligand–receptor interactions. Trends Pharmacol. Sci. 21, 266–273. Jakubik J., El-Fakahany E. E. and Tucek S. (2000) Evidence for a tan- dem two-site model of ligand binding to muscarinic acetylcholine receptors. J. Biol. Chem. 275, 18836–18844. Ja¨rv J., Hedlund B. and Bartfai T. (1979) Isomerization of the muscarinic receptor-antagonist complex. J. Biol. Chem. 254, 5595–5598. Kallal L. and Benovic J. L. (2000) Using green fluorescent proteins to study G-protein-coupled receptor localization and trafficking. Trends Pharmacol. Sci. 21, 175–180. Lakowicz J. R. (1999) Principles of Fluorescence Spectroscopy. Kluwer Academic/Plenum Publishers, New York. Luthin G. R. and Wolfe B. B. (1984) [3H]Pirenzepine and [3H]Quinuc- lidinyl benzilate binding to brain muscarinic cholinergic receptors. Differences in measured receptor density are not explained by dif- ferences in receptor isomerization. Mol. Pharmacol. 26, 164–169. Milligan G. (1999) Exploring the dynamics of regulation of G protein- coupled receptors using green fluorescent protein. Br. J. Pharma- col. 128, 501–510. Okada T., Ernst O., Palczewski K. and Hofmann P. (2001) Activation of rhodopsin: new insights from structural and biochemical studies. Trends Biochem. Sci. 26, 318–324. Palanche´ T., Ilien B., Zoffmann S., Reck M.-P., Bucher B., Edelstein S. J. and Galzi J.-L. (2001) The neurokinin A receptor activates calcium and cAMP responses through distinct conformational states. J. Biol. Chem. 276, 34853–34861. Palczewski K., Kumasaka T., Hori T., Behnke C. A., Motoshima H., Fox B. A., Le Trong I., Teller D. C., Okada T., Stenkamp R. E. et al. (2000) Crystal structure of rhodopsin: a G protein-coupled recep- tor. Science 289, 739–745. Potter L. T., Ferrendelli C. A. and Hanchett H. E. (1988) Two affinity states of M1 muscarine receptors. Cell. Mol. Neurobiol. 8, 181– 191. Qin D., Sullivan R., Berkowitz W. F., Bittman R. and Rotenberg S. A. (2000) Inhibition of protein kinase C alpha by dequalinium analogues: dependence on linker length and geometry. J. Med. Chem. 43, 1413–1417. Richards M. H. (1991) Commentary. Pharmacology and second mes- senger interactions of cloned muscarinic receptors. Biochem. Pharmacol. 42, 1645–1653. Rios C. D., Jordan B. A., Gomes I. and Devi L. A. (2001) G-protein- coupled receptor dimerization: modulation of receptor function. Pharmac. Ther. 92, 71–87. Schmidt H. and Oettling G. (1987) Chloroquine is a muscarinic antag- onist. Binding and dose–response studies with chick embryo cells. Eur. J. Pharmacol. 133, 83–88. Spalding T. A., Trotter C., Skjærbæk N., Messier T. L., Currier E. A., Burstein E. S., Li D., Hacksell U. and Brann P. R. (2002) Dis- covery of an ectopic activation site on the M1 muscarinic receptor. Mol. Pharmacol. 61, 1297–1302. Tsien R. Y. (1998) The green fluorescent protein. Annu. Rev. Biochem. 67, 509–544. Unson C. G., Cypess A. M., Kim H. N., Goldsmith P. K., Carruthers C. J. L., Merrifield R. B. and Sakmar T. P. (1995) Characterization of deletion and truncation mutants of the rat glucagon receptor. Seven transmembrane segments are necessary for receptor trans- port to the plasma membrane and glucagon binding. J. Biol. Chem. 270, 27720–27727. Valenzuela-Fernandez A., Palanche´ T., Amara A., Magerus A., Altmeyer R., Delaunay T., Virelizier J.-L., Baleux F., Galzi J.-L. and Arenzana-Seisdedos F. (2001) Optimal inhibition of X4 HIV iso- lates by the CXC chemokine stromal cell-derived factor 1a requires interaction with cell surface heparan sulfate proteoglycans. J. Biol. Chem. 276, 26550–26558. Vollmer J.-Y., Alix P., Chollet A., Takeda K. and Galzi J.-L. (1999) Subcellular compartmentalization of activation and desensitization of responses mediated by NK2 neurokinin receptors. J. Biol. Chem. 274, 37915–37922. Weill C., Galzi J.-L., Chasserot-Golaz S., Goeldner M. and Ilien B. (1999a) Functional characterization and potential applications for enhanced green fluorescent protein- and epitope-fused human M1 muscarinic receptors. J. Neurochem. 73, 791–801. Weill C., Ilien B., Goeldner M. and Galzi J.-L. (1999b) Fluorescent muscarinic EGFP-hM1 chimeric receptors: design, ligand binding and functional properties. J. Recept. Signal Transduct. Res. 19, 423–436. Wess J. (1998) Molecular basis of receptor/G-protein-coupling selec- tivity. Pharmac. Ther. 80, 231–264. Yang F., Moss L. G. and Phillips G. N. (1996) The molecular structure of green fluorescent protein. Nat. Biotechnol. 14, 1246–1251. 778 B. Ilien et al. � 2003 International Society for Neurochemistry, J. Neurochem. (2003) 85, 768–778