Biochimica et Biophysica Acta 1778 (2008) 2690–2699 Contents lists available at ScienceDirect Biochimica et Biophysica Acta j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / b b a m e m Identifying the binding site(s) for antidepressants on the Torpedo nicotinic acetylcholine receptor: [3H]2-azidoimipramine photolabeling and molecular dynamics studies Mitesh Sanghvi a, Ayman K. Hamouda a, Krzysztof Jozwiak c, Michael P. Blanton a, James R. Trudell d, Hugo R. Arias b,⁎ a Department of Pharmacology and Neuroscience, Texas Tech University Health Sciences Center, Lubbock, TX 79430, USA Department of Pharmaceutical Sciences, College of Pharmacy, Midwestern University, Glendale, AZ 85308-3550, USA c Department of Chemistry, Medical University of Lublin, 20-090 Lublin, Poland d Department of Anesthesia, Stanford University School of Medicine, Stanford, CA 94305-5117, USA b a r t i c l e i n f o Article history: Received 1 May 2008 Received in revised form 18 August 2008 Accepted 21 August 2008 Available online 10 September 2008 Keywords: Torpedo nAChR Noncompetitive inhibitors Tricyclic antidepressants Photolabeling Molecular modeling a b s t r a c t Radioligand binding, photoaffinity labeling, and docking and molecular dynamics were used to characterize the tricyclic antidepressant (TCA) binding sites in the nicotinic acetylcholine receptor (nAChR). Competition experiments indicate that the noncompetitive antagonist phencyclidine (PCP) inhibits [3H]imipramine binding to resting (closed) and desensitized nAChRs. [3H]2-azidoimipramine photoincorporates into each subunit from the desensitized nAChR with ∼ 25% of the labeling specifically inhibited by TCP (a PCP analog), whereas no TCP-inhibitable labeling was observed in the resting (closed) state. For the desensitized nAChR and within the α subunit, the majority of specific [3H]2-azidoimipramine labeling mapped to a ∼ 20 kDa Staphylococcus aureus V8 protease fragment (αV8-20; Ser173-Glu338). To further map the labeling site, the αV8-20 fragment was further digested with endoproteinase Lys-C and resolved by Tricine SDS-PAGE. The principal labeled fragment (11 kDa) was further purified by rpHPLC and subjected to N-terminal sequencing. Based on the amino terminus (αMet243) and apparent molecular weight, the 11 kDa fragment contains the channel lining M2 segment. Finally, docking and molecular dynamics results indicate that imipramine and PCP interact preferably with the M2 transmembrane segments in the middle of the ion channel. Collectively, these results are consistent with a model where PCP and TCA bind to overlapping sites within the lumen of the Torpedo nAChR ion channel. © 2008 Elsevier B.V. All rights reserved. 1. Introduction The main pharmacological action of antidepressants (ADs) is to increase the synaptic concentrations of norepinephrine, serotonin, and/or dopamine by inhibiting the reuptake of these neurotransmitters from the synaptic cleft. Moreover, additional studies have shown that ADs act as noncompetitive antagonists (NCAs) of muscle- and neuronal-type nicotinic acetylcholine receptors (nAChRs) (reviewed in [1]). The muscle-type nAChR, in particular the nAChR isolated from Torpedo californica electric organ, has been studied extensively over the past several decades and serves as the archetype of the Cys-loop ligand-gated ion channel superfamily. This genetically linked receptor superfamily includes both muscle- and neuronal-type nAChRs, type A and C γ-aminobutyric acid, type 3 5-hydroxytryptamine (serotonin), and glycine receptors (reviewed in [2–4]). Previous studies from our laboratory determined the molecular mechanisms underlying the noncompetitive inhibition of AChRs ⁎ Corresponding author. Tel.: +1 623 572 3589; fax: +1 623 572 3550. E-mail address: harias@midwestern.edu (H.R. Arias). 0005-2736/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.bbamem.2008.08.019 elicited by tricyclic antidepressants (TCAs) [5]. These earlier studies further established that imipramine inhibits the binding of the wellcharacterized NCA [piperidyl-3, 4-3H(N)]-(N-(1-(2 thienyl)cyclohexyl)-3,4-piperidine) ([3H]TCP), a structural and functional analog of the dissociative anesthetic phencyclidine (PCP), to both resting and desensitized nAChRs (Kis = 6.7 ± 0.4 and 0.75 ± 0.04 μM, respectively [5]). However, the structural components of the TCA binding sites in conformationally distinct nAChRs have not been characterized in detail. Thus, to further characterize the TCA binding sites, we have employed [3H]2-azidoimipramine (see Fig. 1), a photoreactive analog of imipramine, to covalently tag the TCA binding sites in the Torpedo nAChR ion channel. We choose the Torpedo nAChR as a receptor model mainly because of its high abundance in the electroplaque tissue of T. californica fish compared to cells and neurons expressing neuronaltype nAChRs. The radioligand binding, photoaffinity labeling, and docking and molecular dynamics data suggest that imipramine interacts with the PCP binding sites in the resting (closed) and desensitized states. More specifically, imipramine binds mainly to the M2 transmembrane segment of the Torpedo AChR, which is the principal segment forming M. Sanghvi et al. / Biochimica et Biophysica Acta 1778 (2008) 2690–2699 Fig. 1. Molecular structure of [3H]2-azidoimipramine. The molecule was built using the Builder module of Insight 2005L (Accelrys, San Diego, CA), and is rendered as a spacefilling surface (CPK). The nitrogen atoms from the azido group are colored in blue. the ion channel. These studies in the muscle-type nAChR will aid the modeling of the TCA binding site on neuronal nAChRs. 2. Materials and methods 2.1. Materials [3H]2-azidoimipramine (Fig. 1) (5H-Dibenz(b,f)azepine-5-propanamine, 2-azido-10,11-dihydro-N,N-dimethyl-) (80 Ci/mmol) was synthesized according to Rotman and Pribluda [6] by American Radiolabeled Chemicals (St. Louis, MO). [Benzene ring-3H(N)]-imipramine hydrochloride ([3H]imipramine) (41.3 Ci/mmol) and [piperidyl-3,4-3H(N)]-(N-(1-(2 thienyl)cyclohexyl)-3,4-piperidine) ([3H] TCP) (45 Ci/mmol) were purchased from PerkinElmer Life Sciences Products, Inc. (Boston, MA). Radiolabeled drugs were stored in ethanol at −20 °C. Carbamylcholine hydrochloride (Carb), suberyldicholine dihydrochloride, l-[l-(2-thienyl)cyclohexyl]pyrrolidine (TCP), imipramine hydrochloride, α-bungarotoxin (α-BgTx), polyethylenimine, and Tricine were purchased from Sigma-Aldrich (St. Louis, MO). Phencyclidine hydrochloride (PCP) was obtained through the National Institute on Drug Abuse (NIDA) (NIH, Bethesda, Maryland). Staphylococcus aureus glutamyl endopeptidase (V8 protease) was obtained from MP Biochemicals (Irvine, CA), and Genapol C-100 from Calbiochem (San Diego, CA). Trifluoroacetic acid (TFA) and [1(dimethylamino) naphtalene-5-sulfonamido]ethyltrimethylammonium perchlorate (dansyltrimethylamine) were purchased from Pierce (Rockford, IL). Prestained low range molecular weight standards were purchased from Life Technologies, Inc. (Gaithersburg, MD). 2.2. Preparation of nAChR-rich membranes and affinity-purified nAChRs reconstituted in lipid vesicles nAChR-rich membranes were isolated from frozen T. californica electric organs obtained from Aquatic Research Consultants (San Pedro, CA) by differential and sucrose density gradient centrifugation, as described previously [7]. The final membrane suspensions in ∼ 38% sucrose were stored at −80 °C. Total AChR membrane protein was determined using the bicinchoninic acid (BCA) protein assay (Pierce Chemical Co., Rockford, IL). Specific activities of these membrane preparations were determined by the decrease in dansyltrimethylamine (6.6 μM) fluorescence produced by the titration of 2691 suberyldicholine into receptor suspensions (0.3 mg/ml) in the presence of 100 μM PCP. The activities ranged from 0.9 to 1.2 nmol of suberyldicholine binding sites/mg total protein (0.45– 0.60 nmol AChR/mg protein). Fluorescence titrations were carried out in 5-mm quartz cuvettes using an Olis DM245 spectrofluorimeter (Bogart, GA). Dansyltrimethylamine excitation and emission wavelengths were 295 and 546 nm, respectively. To reduce straylight effects, a 530 nm cutoff filter was placed in the path of the emission wavelength. nAChRs were affinity-purified using a bromoacetylcholine bromide-derivatized Affi-Gel 10 column (Bio-Rad) as described previously [8]. Briefly, the affinity column was prepared by coupling 50 mL of Affi-gel 10 to cystamine, reduction of the cystamine disulphide, and then sulfhydryl coupling to bromoacetylcholine bromide (1.5 g). The column was then equilibrated with ∼15 column volumes of lipid (asolectin, a crude soybean lipid extract) in 1% cholate in vesicle dialysis buffer (VDB) (100 mM NaCl, 0.1 mM EDTA, 0.02% NaN3, 10 mM MOPS, pH 7.4) (0.2 mL/min; N15 h). The solubilized material was slowly applied to the affinity column (0.3 mL/min, ∼24 h, at 4 °C). The column was then washed extensively with asolectin–lipid solution (0.2–0.9 mg/mL lipid) in 1% cholate in VDB (15 column volumes; N15 h). This extensive wash ensures complete exchange of endogenous lipids for the asolectin–lipid mixture [8]. nAChRs were eluted from the column using the asolectin–lipid solution (0.2 mg/mL) containing 10 mM carbamylcholine (Carb). Peak protein fractions (A280 × 0.6; see Ref. [8]) were pooled and the lipid–protein molar ratio adjusted to 400 to 1 except were specified. To remove Carb and reconstitute nAChRs into membranes containing asolectin–lipid, pooled fractions were dialyzed against 2 L of VDB (4 d with buffer change once every day). The reconstituted nAChRs were aliquoted (0.25 mg per tube) and stored at −80 °C. 2.3. Imipramine- and PCP-induced inhibition of [3H]imipramine binding to nAChR in different conformational states We studied the effect of imipramine and PCP on maximal [3H] imipramine binding to the Torpedo nAChR. In this regard, nAChR native membranes (0.3 μM nAChR) were suspended in binding saline buffer (50 mM Tris–HCl, 120 mM NaCl, 5 mM KCl, 2 mM CaCl2, 1 mM MgCl2, pH 7.4), with 10 nM [3H]Imipramine in the presence of 1 mM Carb (desensitized state), or alternatively with 75 nM [3H]Imipramine in the presence of 1 μM α-bungarotoxin [α-BgTx; resting (closed) state], and preincubated for 30 min at room temperature (RT). α-Bungarotoxin is a competitive antagonist that maintains the nAChR in the resting (closed) state [9]. Since imipramine inhibits [3H]TCP binding to nAChRs in the resting (closed) (∼ 6 μM) and desensitized (∼ 0.7 μM) states [5], the nonspecific binding for the imipramine-induced inhibition of [3H]imipramine binding experiments was determined in the presence of 100 μM (desensitized) or 500 μM imipramine (resting/closed), respectively. For the PCPinduced inhibition of [3H]imipramine binding experiments, the nonspecific binding was determined in the presence of 50 μM (desensitized) or 100 μM PCP (resting/closed) according to Arias et al. [10]. The total volume was divided into aliquots, and increasing concentrations of imipramine or PCP (i.e., 0.1 nM–500 μM) were added to each tube and incubated for 2 h at RT. nAChR-bound radioligand was then separated from free ligand by a filtration assay using a 48-sample harvester system with GF/B Whatman filters (Brandel Inc., Gaithersburg, MD), previously soaked with 0.5% polyethylenimine for 30 min. The membrane-containing filters were transferred to scintillation vials with 3 mL of Bio-Safe II (Research Product International Corp, Mount Prospect, IL), and the radioactivity was determined using a Beckman SL6500 scintillation counter (Beckman Coulter, Inc., Fullerton, CA). The concentration–response data were curve-fitted by nonlinear least-squares analysis using the Prism software (GraphPad Software, 2692 M. Sanghvi et al. / Biochimica et Biophysica Acta 1778 (2008) 2690–2699 San Diego, CA). The corresponding IC50 values were calculated using the Hill equation:
 θ = 1= 1 + ð½LŠ=IC50 ÞnH ð1Þ where θ is the fractional amount of the radioligand bound in the presence of inhibitor at a concentration [L] compared to the amount of the radioligand bound in the absence of inhibitor (total binding). IC50 is the inhibitor concentration at which θ = 0.5 (50% bound), and nH is the Hill coefficient. The nH values were summarized in Table 1. The observed IC50 values from the competition experiments described above were transformed into inhibition constant (Ki) values using the Cheng–Prusoff relationship [11]: i n hh i o 3 Ki = IC50 = 1 + H imipramine =Kdimipramine ð2Þ where [[3H]imipramine] is the initial concentration of [3H]imipramine and Kimipramine is the dissociation constant for [3H]imipramine (0.7 μM d in the desensitized state and 6.4 μM in the resting (closed) state [5]). The calculated Ki values for imipramine and PCP were summarized in Table 1. 2.4. Photolabeling of Torpedo nAChRs with [3H]2-azidoimipramine Affinity-purified Torpedo nAChRs reconstituted in asolectin–lipid membranes were adjusted to 0.25 mg/mL in VDB and supplemented with 1.1 mM oxidized glutathione to serve as an aqueous scavenger. Membrane aliquots (0.25 mg analytical scale; 12 mg preparative scale) were then incubated with [3H]2-azidoimipramine (from an ethanolic stock solution; final concentrations are 0.174 μM and b1% ethanol) either in the presence of 400 μM Carb (desensitized state) or 2 μM αBgTx (resting/closed state). The total volume was divided in different glass test tubes in absence (total photolabeling) or in the presence of 50 μM TCP (nonspecific photolabeling). After 1 h in the dark at RT, samples were irradiated at 312 nm (Spectroline EN-280L, Spectronics, Westbury, NY) for 10 min at a distance of b1 cm. Photolyzed membrane suspensions were then pelleted, solubilized in electrophoresis sample buffer, and subjected to SDS-PAGE. 2.5. SDS-polyacrylamide gel electrophoresis SDS-PAGE was performed according to the method of Laemmli [12] with analytical (1.0 mm thick) and preparative (1.5 mm) separating gels comprised of 8% polyacrylamide/0.33% bis-acrylamide. Following electrophoresis, nAChR subunits were visualized by staining with Coomassie Blue R-250 [0.25% (w/v) in 45% methanol, 10% acetic acid, 45% distilled water] and destaining (25% methanol, 10% acetic acid, 65% distilled water). For fluorography, stained gels were impregnated with fluor (Amplify; GE Biosciences, Piscataway, NJ) for 30 min, dried and exposed at −80 °C to Eastman Kodak X-Omat LS film (typical exposure time = 2–4 weeks). 3H incorporation into individual polypeptides was also quantified by liquid scintillation counting of excised gel slices [13]. For either analytical or preparative [3H]2-azidoimipramine photolabeling experiments, nAChR α-subunits were fragmented by limited “in gel” digestion with S. aureus V8 protease [13–15]. Following electrophoresis, stained 8% acrylamide gels were soaked in distilled water overnight and nAChR α-subunit band from each labeling conditions [+/− Carb, +/+ Carb and TCP in the desensitized state; and − and + α-BgTx in the resting (closed) state] were excised and transferred directly to the well(s) of either a 1.0 mm (analytical) or 1.5 mm (preparative) thick mapping gel, composed of a 5-cm-long 4.5% acrylamide stacking gel and a 11-cm-long 15% acrylamide separating gel. For analytical labelings, each gel slice was overlaid with 6 μg S. aureus V8 protease in overlay buffer (5% sucrose, 125 mM Tris–HCl, 0.1% SDS, pH 6.8) and electrophoresed at 60 V constant voltage for ∼ 3 h and then at ∼ 6 mA constant current overnight. After Table 1 Affinity of phencyclidine and imipramine for the [3H]imipramine binding site in the resting (closed) and desensitized nAChRs NCA Desensitized state PCP Imipramine Ki, μMa 0.41 ± 0.02 0.85 ± 0.05 Resting state nHb 1.15 ± 0.05 0.96 ± 0.05 Ki, μMa 1.2 ± 0.3 3.8 ± 0.4 nHb 0.80 ± 0.10 0.82 ± 0.08 Data obtained from the plots of Fig. 2. a Inhibition constant. b Hill coefficient. Coomassie Blue R-250 staining (1 h) and destaining (2–3 h), the analytical gels were impregnated with fluor, dried and exposed to film (up to 12 weeks). Preparative 8% acrylamide gels (1.5 mm thick) were soaked in distilled water overnight and the nAChR α−subunit excised as a strip (∼15 cm). The gel strip was transferred to 15% acrylamide mapping gel and overlaid with 250 μg V8 protease in overlay buffer. Following electrophoresis, staining, and soaking in distilled water overnight, the following subunit proteolytic fragment band was excised: αV8-20 (αSer173-Glu338; using the nomenclature of Blanton and Cohen [16]). The excised proteolytic fragment was then isolated by passive diffusion into 25 mL of elution buffer [0.1 M NH4HCO3, 0.1% (w/v) SDS, 1% β-mercaptoethanol, pH 7.8] for 4 days at RT with gentle shaking [16]. The gel suspensions were then filtered (Whatman N 1 paper) and concentrated using Centriprep-10 concentrators (10 kDa cutoff, Amicon, Billerica, MA). Excess SDS was removed by acetone precipitation (N85% acetone at −20 °C overnight). 2.6. Proteolytic digestions For digestion with Endoproteinase Lys-C (EndoLys-C), acetone precipitated V8 protease subunit fragments from each condition were resuspended in 200 μL of 15 mM Tris, 0.1% (w/v) SDS, pH 8.1, and then incubated with EndoLys-C (1.5 U in 100 μL EKC buffer) for 5–7 h at RT. Aliquots (∼ 10%) of each EKC digest were resolved on an analytical Tricine SDS-PAGE gels and processed for fluorography (up to 10 weeks exposure). The bulk of the material from each EKC digest was separated on individual preparative scale (1.0 mm thick) Tricine SDS-PAGE gels. Following staining and destaining of each preparative Tricine gel, selected fragments were excised based upon the results of fluorographs of analytical Tricine gels and with the aid of low range molecular weight standards. The selected 11 kDa (αEKC-11) and 6 kDa (αEKC-6) fragments were then isolated by passive elution into 5 mL of elution buffer over 4 days at RT. The gel suspensions were filtered, concentrated (Centriprep-3, Amicon), and further purified by rpHPLC. 2.7. HPLC purification Proteolytic fragments of the nAChR α-subunit were purified by rpHPLC on a Shimadzu LC-10A binary HPLC system, using a Brownlee Aquapore C4 column (100 × 2.1 mm). Solvent A was 0.08% trifluoroacetic acid (TFA) in water and Solvent B was 0.05% TFA in 60% acetonitrile/40% 2-propanol. A nonlinear elution gradient at 0.2 mL/min was employed (25% to 100% Solvent B in 100 min) and fractions were collected every 2.5 min (42 fractions/run). The elution of peptides was monitored by absorbance at 210 nm. 50 μL aliquots of each collected fraction were counted for radioactivity using liquid scintillation counting. 2.8. Sequence analysis Amino terminal sequence analysis was performed on a Beckman Instruments (Porton) 20/20 automated protein sequencer using gas phase cycles (Texas Tech Biotechnology Core Facility). Pooled HPLC M. Sanghvi et al. / Biochimica et Biophysica Acta 1778 (2008) 2690–2699 fractions of αEKC-11 and αEKC-6 fragments were dried by vacuum centrifugation, resuspended in 20 μL 0.1% SDS and immobilized on chemically modified glass fiber disks (Beckman Instruments), which were used to improve the sequencing yields of hydrophobic peptides. Peptides were subjected to at least 10 sequencing cycles. 2.9. Docking and molecular dynamics of imipramine and PCP in the nAChR ion channel The membrane domain of the nAChR is formed by 20 α-helical transmembrane segments, where each subunit incorporates four segments (M1, M2, M3 and M4). The M2 segments are relatively amphiphilic and are exposed to the centrally located pore permeable to water and cations. Since the numbering of amino acid residues varies greatly between members of the Cys-loop superfamily, the residues in M2 of nAChR subunits are referred to here using the prime nomenclature (1′ to 20′), corresponding to residues Met243 to Glu262 in the Torpedo nAChR α-subunit. As a target for modeling the binding of imipramine and PCP in the present study we used a structural model of the pore region of nAChR based on the cryo-electron microscopy structure of the Torpedo nAChR at ∼4 Å resolution [17,18] (PDB ID 2BG9). Computational simulations were divided into three steps. In the first step, imipramine and PCP molecules were prepared using HyperChem 6.0 (HyperCube Inc., Gainesville, FL). Sketched molecules were optimized using the semiempirical method AM1 (Polak-Ribiere algorithm to a gradient lower than 0.1 kcal/Å/mol) and then transferred to the second step: ligand docking. The Molegro Virtual Docker (MVD 2008.2.4.0 Molegro ApS Aarhus, Denmark) was used for docking simulations of flexible ligands into the rigid target nAChR model. In this step the complete 2BG9 structure was used [17,18]. The docking space was limited and centered on the middle of the ion channel and extended enough to ensure covering of the whole channel domain for sampling simulations (docking space was defined as a sphere of 21 Å in diameter). The actual docking simulations (settings: numbers of runs = 100; maximal number of interactions = 10,000; maximal number of poses = 10) were performed and the pose representing the lowest value of the scoring function (MolDockScore) for each ligand was selected for further simulations. In the next step, molecular dynamics were performed using the Yasara 6.10.18 package (Yasara Biosciences, Graz, Austria). Complexes of 2BG9 with imipramine or PCP were edited to provide coordinates for the membrane domain only. For each subunit, M1– M4 transmembrane helices plus connecting loops were left in the system, all other parts of the protein were removed. Thus, we obtained two models representing complexes between the nAChR channel domain and imipramine or PCP. These models were further inserted into periodic boxes (88 Å × 88 Å × 68 Å), solvated with water molecules using Yasara default algorithm and fixing constrains for backbone atoms were assigned. In all further simulations the AMBER99 force field for both protein and ligand structures was used (Yasara BioSciences) with the cutoff 7.86 Å and particle-mesh Ewald longrange function for electrostatic interactions. The initial complexes were pre-optimized with steepest descent method followed by 500 steps of simulated annealing. Finally, the actual 0.8 ns molecular dynamics was performed using the following parameters: temperature = 298 K; multiple timesteps = 1 fs for intramolecular and 2 fs for intermolecular forces, PressureControl — waterprobe (0.99 g/ml) ensemble. Snapshots of the simulations were saved every 5 ps. Results of dynamics simulations were characterized by calculation of the binding energy defined here as the difference between the energy of the complete complex system and a summation of energy of the ligand and the energy of the hydrated protein alone. All these calculations were performed with the same force field settings. 2693 3. Results 3.1. Imipramine- and PCP-induced inhibition of [3H]imipramine binding to the resting (closed) and desensitized nAChRs To further localize the site of interaction of ADs and to determine the relative associative properties of TCAs in different conformational states of the nAChR, the PCP- and imipramine-induced inhibition of [3H]imipramine binding to Torpedo nAChRs was examined (Fig. 2). The high concentrations of Torpedo nAChRs in these native membrane provides a high level of specific [3H]imipramine binding which cannot be obtained with cells (or cell membranes) expressing neuronal nAChRs. Experiments were carried out in the presence of 1 mM Carb (desensitized state) or 1 μM α-BgTx (resting/closed state). The results show that imipramine (Fig. 2A) and PCP (Fig. 2B) completely eliminate specific [3H]imipramine binding to the nAChR in a concentration-dependent fashion in either the desensitized or the resting (closed) state. From these experiments, we obtained the respective imipramine and PCP IC50 values, which were transformed to absolute Ki values according to the Cheng–Prusoff equation [Eq. (2)]. The imipramine Ki values were 0.85 ± 0.05 and 3.8 ± 0.4 μM in the desensitized and resting (closed) state, respectively (see Table 1). Therefore, it seems likely that imipramine binds with ∼5-fold higher affinity to the desensitized nAChR than to the resting (closed) nAChR. These values are nearly identical to those obtained by [3H]TCP competition experiments (∼ 0.7 and ∼6.4 μM, respectively; [5]). The observed nH values were close to unity (see Table 1), indicating that the binding is noncooperative, and thus, suggesting the existence of a single imipramine binding site on each conformational state. Fig. 2. (A) Imipramine- and (B) PCP-induced inhibition of [3H]imipramine binding to desensitized (○) and resting (closed) (□) nAChRs. nAChR-rich membranes (0.3 μM nAChR) were equilibrated (2 h) with [3H]imipramine (75 nM in the resting (closed) state and 10 nM in the desensitized state) and carbamylcholine (1 mM) or α-BgTx (1 μM), in the presence of increasing concentrations of imipramine or PCP (i.e., 0.1 nM–500 μM). The nAChR membranes were then filtered and the radioactivity of the filters was measured as described under “Materials and methods”. For experiments in (A), the nonspecific radioactivity was determined with 100 (desensitized state) or 500 μM imipramine (resting state). For experiments in (B), the nonspecific radioactivity was determined with 50 (desensitized state) or 100 μM PCP (resting state). Shown is the mean ± SD of two experiments performed in triplicate. The IC50 and nH values were obtained by nonlinear least-squares fit according to Eq. (1). The Ki values for imipramine and PCP were calculated using the Cheng–Prusoff relationship [Eq. (2)]. 2694 M. Sanghvi et al. / Biochimica et Biophysica Acta 1778 (2008) 2690–2699 Since imipramine completely displaces [3H]TCP binding to the nAChR in a steric fashion, we had suggested that the imipramine binding site overlaps the PCP locus in the resting (closed) and desensitized states [5]. Supporting this conclusion is the fact that the calculated PCP Ki values in the desensitized (0.41 ± 0.02 μM) and resting (closed) (1.2 ± 0.3 μM) states (see Table 1) are practically the same as those previously determined in the desensitized (∼ 0.25 μM; [19]) and in the resting (closed) (∼ 0.83 μM; [10]) state, respectively. Further, from these results, and considering that the observed nH values are close to unity (1.15 ± 0.05 in the desensitized state, and 0.80 ± 0.10 in the resting (closed) state; see Table 1), we can conclude that PCP displaces [3H]imipramine from its high-affinity binding site in a mutually exclusive (steric) manner when the receptor is in either the desensitized or resting (closed) state. Thus, it is likely that the locus for TCAs overlaps, at least partially, the PCP binding site in both states. 3.2. Photoincorporation of [3H]2-azidoimipramine into nAChR subunits Initial photolabeling experiments with [3H]2-azidoimipramine (Fig. 1) were designed to characterize the extent of photoincoproration into nAChR subunits in either the desensitized (in the presence of 400 μM Carb) or resting (closed) state (in the presence of 2 μM αBgTx), as well as to assess how the addition of TCP affects the extent of [3H]2-azidoimipramine photoincorporation into nAChR subunits. Affinity-purified nAChRs were utilized for [3H]2-azidoimipramine photolabeling experiments in order to accurately assess the extent of subunit labeling without the confusion of potential labeling of nonAChR polypeptides present in nAChR-rich membranes. These affinitypurified nAChRs (reconstituted in asolectin–lipid vesicles) are functionally indistinguishable from receptors present in native nAChR-rich Torpedo membranes [8]. For labelings in the absence of agonist, αBgTx was added to stabilize the resting (closed) state [9]. Affinitypurified nAChRs reconstituted in asolectin–lipid membranes (0.25 mg/ml) were supplemented with 1.1 mM oxidized glutathione (aqueous scavenger) and equilibrated with 0.174 μM [3H]2-azidoimipramine either in the presence of Carb, both Carb and TCP, or α-BgTx. After UV-light irradiation, nAChRs subunits were separated by SDSPAGE, and the extent of photoincorporation of [3H]2-azidoimipramine into each nAChR subunit was monitored by fluorography following electrophoresis. A representative fluorograph of an 8% polyacrylamide gel (Fig. 3A) demonstrates that there is appreciable photoincorporation of [3H]2-azidoimipramine into individual nAChR subunits either in the presence of Carb (+/− lane) or in the presence of both Carb and TCP (+/+ lane). The overall labeling pattern and the relative incorporation of [3H]2-azidoimipramine into individual nAChR subunits were clearly reduced by the inclusion of TCP, indicating that this photolabeling is TCP-sensitive (analytical labelings were repeated at least three times with nearly identical results). For photolabeling experiments conducted in the absence of agonist, where the receptor is mainly in the resting (closed) state, the photoincorporation of [3H]2azidoimipramine into α-subunits is significantly reduced by ∼50% (see Fig. 4B) in the presence of α-BgTx (the majority of receptors are in the resting/closed state), whereas the extent of photolabeling in the other subunits is not significantly affected (b10%). Further addition of TCP had no effect on the extent of [3H]2-azidoimipramine incorporation into any subunit (data not shown). These result suggests that there is minimal specific [3H]2-azidoimipramine photolabeling in the resting (closed) state. The α-BgTx-sensitive labeling in the α-subunit is best explained by [3H]2-azidoimipramine photoincorporation into residues within the agonist binding sites and thus, this labeling is eliminated when the agonist binding sites are occupied by α-BgTx. To quantify the extent of photoincorporation into each nAChR subunit (in the presence of Carb, as well as both Carb and TCP) the gel bands were excised from the 8% acrylamide gel and the radioactivity was measured in a scintillation counter. Based on liquid scintillation counting of excised gel bands, [3H]2-azidoimipramine incorporates into each nAChR subunit, with approximately 25% (n = 3) of the total subunit labeling inhibited by addition of TCP (see Fig. 3. Photoincorporation of [3H]2-azidoimipramine into Torpedo nAChRs in the desensitized state. (A) Affinity-purified nAChRs were equilibrated (1 h) with 0.174 μM [3H]2azidoimipramine in the presence of carbamylcholine (Carb) (+/− lane) and in presence of Carb and TCP (+/+ lane), and then irradiated at 312 nm for 10 min at a distance of b1 cm. After irradiation, polypeptides were resolved by SDS-PAGE and processed for fluorography (4 weeks exposure). Indicated on the left are the mobility of Torpedo nAChR subunits (α, β, γ, and δ). (B) The extent of [3H]2-azidoimipramine photoincorporation into nAChR subunits in the presence of Carb (closed bars) and in the presence of Carb and TCP (open bars). The [3H]2-azidoimipramine-labeled nAChR subunit bands were excised from the dried 8% acrylamide gel and the 3H-radioactivity associated with each band was determined by scintillation counting. The amount of [3H]2-azidoimipramine photoincorporation into α, β, γ, and δ subunits, inhibited by the addition of TCP was 24%, 23%, 18%, and 18%, respectively (results from three separate photolabeling experiments produced qualitatively similar results). M. Sanghvi et al. / Biochimica et Biophysica Acta 1778 (2008) 2690–2699 2695 Fig. 4. (A) Photoincorporation of [3H]2-azidoimipramine into Torpedo nAChRs in the resting (closed) state. Affinity-purified nAChRs were equilibrated (1 h) with 0.174 μM [3H]2azidoimipramine in the absence (− lane) and in presence of α-BgTx (+ lane), and then irradiated at 312 nm for 10 min at a distance of b 1 cm. After irradiation, polypeptides were resolved by SDS-PAGE and processed for fluorography (4 weeks exposure). The [3H]2-azidoimipramine-labeled nAChR subunit bands were excised from the dried 8% acrylamide gel and the 3H-radioactivity associated with each band was determined by scintillation counting. (B) The amount of [3H]2-azidoimipramine photoincorporation into the α−subunit inhibited by the addition of α-BgTx was 54%. Labeling in the β, γ, δ-subunits was not significantly (b10%) affected. Fig. 3B). Given the relatively low affinity (∼ μM) binding of imipramine to the nAChR, the lipophilicity of imipramine, as well as other factors, the presence of a significant level of nonspecific [3H] 2-azidoimipramine incorporation is by no means unexpected. The site(s) of [3H]2-azidoimipramine photoincorporation into the nAChR α-subunit labeled in the presence of Carb, both Carb and TCP (Fig. 3), and α-BgTx (Fig. 4), were initially mapped by proteolytic digestion. The α-subunit from each condition was partially digested Fig. 5. Proteolytic mapping of the site of [3H]2-azidoimipramine incorporation into the nAChR α-subunit in the desensitized state using “in gel” digestion with S. aureus V8 protease. Affinity-purified nAChRs were labeled with 0.174 μM [3H]2-azidoimipramine in the presence of Carb (+/− lane) and in presence of Carb and TCP (+/+ lane). After photolysis (312 nm for 10 min), peptides were resolved by SDS-PAGE (1.0 mm thick, 8% acrylamide). nAChR subunit bands were excised following identification by staining (Coomassie Blue) and transferred to the wells of a 15% acrylamide mapping gel for ‘in gel’ digestion with S. aureus V8 protease. Following electrophoresis, the mapping gel was stained with Coomassie Blue, destained and processed for Fluorography (12 week exposure). The principal [3H]2-azidoimipramine-labeled proteolytic fragment was αV8-20. Fig. 6. Proteolytic mapping of the site of [3H]2-azidoimipramine incorporation into the nAChR α-subunit in the resting (closed) state using “in gel” digestion with S. aureus V8 protease. Affinity-purified nAChRs were labeled with 0.174 μM [3H]2-azidoimipramine in the absence (− lane) and in the presence of α-BgTx (+ lane). After photolysis (312 nm for 10 min), peptides were resolved by SDS-PAGE (1.0 mm thick, 8% acrylamide). nAChR subunit bands were excised following identification by staining (Coomassie Blue) and transferred to the wells of a 15% acrylamide mapping gel for ‘in gel’ digestion with S. aureus V8 protease. Following electrophoresis, the mapping gel was stained with Coomassie Blue, destained and processed for Fluorography (12 week exposure). The [3H]2-azidoimipramine labeling on the principal proteolytic fragment, αV8-20, was lost in the presence of α-BgTx. 2696 M. Sanghvi et al. / Biochimica et Biophysica Acta 1778 (2008) 2690–2699 using S. aureus V8 protease under ‘in gel’ conditions [15,20]. Photoincorporation in the α-subunit was found to be restricted to one principal V8 protease fragment, αV8-20 (Ser173-Glu338) (Figs. 5 and 6). The stretch of the primary sequence of the α-subunit contained within the V8 protease fragment V8-20 includes the transmembrane segments M1, M2, and M3. This result indicates that [3H]2-azidoimipramine photoincorporates into a 20 kDa proteolytic fragment of the α-subunit in presence of Carb (+/− lane; Fig. 5) and that the labeling is significantly reduced by the addition of TCP (+/+ lane; Fig. 5). The site(s) of [3H]2-azidoimipramine photoincorporation into the nAChR α-subunit labeled in the absence of Carb, and in the presence of α-BgTx (Fig. 4), were also mapped by ‘in gel’ digestion with V8 protease and found to be restricted to the αV8-20 fragment (Fig. 6). That α-BgTx essentially eliminated [3H]2-azidoimipramine photoincorporation into αV8-20 is best explained by two possibilities: [3H] 2-azidoimipramine photolabeling is lost when the ion channel is in the resting (closed) conformational state or [3H]2-azidoimipramine photolabeling of the α-subunit is largely restricted to the agonist binding site (in particular Loop C). In order to further localize the site(s) of TCP-sensitive [3H]2azidoimipramine photoincorporation within the α-subunit when the receptor is in the desensitized state, labeled αV8-20 was digested in solution with EndoLys-C. Digestion of αV8-20 with EndoLys-C is known to generate fragments of approximately 11 and 6 kDa starting at αMet243 (the N-terminus of αM2) and at αHis186 (that contains the agonist binding site Loop C; [7]), respectively. When aliquots of the EndoLys-C digestion of [3H]2-azidoimipramine-labeled αV8-20 were resolved by Tricine SDS-PAGE, two labeled fragments were visualized by fluorography with a principal band migrating at an apparent molecular weight of 11 kDa and a minor band migrating at 6 kDa (Fig. 7). The bulk of each digest was purified by rpHPLC (Fig. 8). When the EndoLys-C digests of [3H]2-aziodimipramine-labeled 11 kDa and 6 kDa fragments were fractionated by HPLC, there was a broad 3Hpeak for the fragments labeled in the presence of Carb, whereas the Fig. 8. Reversed-phase HPLC purification of [3H]2-azidoimipramine-labeled fragments from EKC digests of αV8-20. The [3H]2-azidoimipramine-labeled 11 kDa (Panel A) and 6 kDa (Panel B) fragments isolated from exhaustive EKC digests of αV8-20 (see Fig. 6) were resolved by rpHPLC on a Brownlee Aquapore C4 column as described in the “Materials and methods”. The elution of peptides was monitored by absorbance at 210 nm and elution of 3H by scintillation counting of aliquots (50 μL) of each 500 μL fraction in presence of Carb (●) and in presence of Carb and TCP (○). On the basis of recovery of radioactivity, N90% of the material was recovered from the rpHPLC column. HPLC fractions 37–39 (A) or 34–36 (B) were pooled and subjected to 10 automated protein sequencing cycles. 3 Fig. 7. Endoproteinase Lys-C digestion of [3H]2-azidoimipramine-labeled V8 protease fragment αV8-20. The V8 protease fragment αV8-20, isolated from nAChRs labeled with 0.174 μM [3H]2-azidoimipramine in the presence of Carb (+/− lane) and in presence of Carb and TCP (+/+ lane), were exhaustively digested with Endoproteinase Lys-C (EKC, 1.5 U per 100 μl EKC buffer) for 5–7 days at RT. Aliquots of the total digests (∼10%) were fractionated by Tricine SDS-PAGE and then subjected to fluorography for 10 weeks. The principal band of 3H evident in EKC digest migrates with an apparent molecular mass of ∼11 kDa (αEKC11) and minor band at ∼6 kDa (αEKC6). H-radioactivity was significantly reduced in those fractions from the sample labeled in the presence of Carb and TCP. The fractions 37 to 39 (Fig. 8A) and 34 to 36 (Fig. 8B) were pooled for the digests of 11 kDa and 6 kDa, respectively, and subjected to 10 automated protein sequencing cycles. For each fragment, sequence analysis revealed the presence of a primary peptide beginning at Met243 (11 kDa fragment) and His186 (6 kDa fragment). Based on the apparent molecular weights of 11 and 6 kDa, estimated from the Tricine gel, the 11 kDa peptide begins at Met243 and would extend to the end of the αV8-20 (Glu338) fragment and includes both the M2 and M3 transmembrane segments. The 6 kDa fragment begins at His186 and extends through the transmembrane segment M1 (Arg209-Tyr234) and terminates at the beginning of the M2 segment (Lys242). Unfortunately, the extent of [3H]2-azidoimipramine incorporation into the M1, M2, and M3 segments was too low to effectively allow for the determination of individually labeled amino acids by amino terminal radio-sequence analysis. Nonetheless, these results suggest that [3H]2-azidoimipramine labels amino acids primarily within the M2 segment with minor labeling in the M1 segment of the α-subunit when the nAChR is in the desensitized state. 3.3. Docking and molecular dynamics of imipramine and PCP in the nAChR ion channel In the present study, molecular modeling was used to build and optimize molecular models of the nAChR ion pore in complex with M. Sanghvi et al. / Biochimica et Biophysica Acta 1778 (2008) 2690–2699 imipramine and PCP molecules. Molegro Virtual Docker generated a series of docking poses and ranked them using energy-based criterion using the embedded scoring function in MolDockScore. 2697 Based on this ranking, the lowest energy pose of the ligand–receptor complex can be selected. The MolDockScore value for the PCP best ranked complex was −77.31 kJ/mol. This pose is shown in Fig. 9C. The molecule interacts with the middle section of the ion channel in the cavity formed between the serine (position 6′) and valine rings (position 13′) (Fig. 9C). Analysis of the complex shows that the docked molecule interacts only with all five M2 helices, provided by each subunit. A very similar mode of binding was observed in previous simulations where the PCP molecule was docked into the channel models of α3β4 [21] and α3β2 [22] nAChR. In contrast, in an earlier docking study using the α7 nAChR as a model [23], PCP mainly bound in a region deeper in the ion channel located between positions 2′ and 6′. Fig. 9A and B show the results of analogous docking of imipramine molecule. In this case the lowest energy of the complex was −100.68 kJ/mol, suggesting that the modeled binding should be higher for imipramine than for PCP. Generally, the orientation of the imipramine within the channel is similar to that observed for PCP. The imipramine lowest energy pose is located in the middle portion of the channel, in the cavity formed between the serine (position 6′) and valine (position 13′) rings. Molecule location allows interaction with each of the M2 helices and, as in the case of PCP docking, there is no evidence suggesting interactions with transmembrane helices other than M2. When best scored complexes (this time systems were limited to membrane domain of the receptor as described in Materials and methods section) were hydrated with explicit waters, optimized, and simulated with molecular dynamics at 298 K for 0.8 ns, the PCP molecule did not significantly changed the position comparing to the starting pose (mean RMSD from starting position = 1.89 Å). PCP interacted exclusively with nearby pore-lining residues at overlapping sites between the serine (position 6′) and valine rings (position 13′), without interacting with the external mouth of the ion channel, which is located beyond position 13′, closer to the extracellular ring (position 29′). In comparison to the PCP simulations, molecular dynamics of the imipramine-nAChR complex showed more significant changes compared to the starting position (mean RMSD value = 3.23 Å). The imipramine molecule reoriented significantly during the period of simulation (0.8 ns), which was reflected by a greater RMSD value. Nevertheless, the molecule still interacted predominantly with residues between positions 6′ and 13′ within the ion channel. Carried simulation did not indicate any possibility of ligand interacting with transmembrane helices other than M2. Amber ForceField allowed to estimate the ligand binding energy (defined here as the difference between the energy of the complete complex system and a summation of the energy of the ligand and the energy of the hydrated protein alone). Comparison of these values for both analyzed ligands suggest a similar binding energy for imipramine in complex with the nAChR ion channel (−35.2 ± 3.8 kcal/ mol) compared to that for PCP (−29.9 ± 2.4 kcal/mol). Binding energy estimations indicate that affinities of both NCAs are quite similar, which is consistent with the Ki values determined experimentally (see Table 1). Fig. 9. Lowest energy complexes of imipramine (A and B) and PCP (C) docked into the Torpedo nAChR model (2BG9.pdb) [17]. The complexes are shown with the δ subunit removed for clarity. The order of the remaining subunits are (from left to right): α, γ, α and β. (A) Transversal view of the AChR ion channel showing the location of the imipramine binding site. The AChR is rendered in secondary structure mode and the ligand is rendered in CPK model. (B and C) Detailed view of imipramine (B) and PCP (C) interacting within the ion channel between the valine (position 13′) and serine (position 6′) rings. The target protein is rendered in semitransparent surface with visible secondary structure. Explicit CPK atoms of residues forming the valine (green) and serine (red) rings. The ligands are rendered in stick mode with hydrogen atoms not shown explicitly. 2698 M. Sanghvi et al. / Biochimica et Biophysica Acta 1778 (2008) 2690–2699 4. Discussion The principal goal of this work was to structurally characterize the TCA binding sites on the Torpedo muscle-type nAChR in the desensitized and resting (closed) states. To this end, receptor statedependent radioligand competition binding, photoaffinity labeling, and docking and molecular dynamics experiments were performed. We used different experimental strategies to identify the TCA binding sites. First, we calculated the Ki and nH values of imipramine and PCP displacement of [3H]imipramine binding to the nAChR as a complement to previously published results demonstrating mutually exclusive binding of [3H]TCP and imipramine [5]. These results serve to localize the TCA site within the pore of the ion channel in the desensitized and resting (closed) conformational states. Second, we used [3H]2-azidoimipramine, a photoactivatable analog of imipramine, to directly identify the TCA binding domains in the nAChR. Finally, we performed docking and molecular dynamics of the Torpedo nAChR ion channel complexed with imipramine or PCP to support our experimental data. We first estimated that imipramine binds the Torpedo AChR with ∼ 5-fold higher affinity to the desensitized state compared to the resting (closed) state by [3H]imipramine competition experiments (see Table 1). This is in accord with previous competition experiments where imipramine inhibits [3H]TCP binding to the desensitized nAChR with ∼10-fold higher affinity than to the resting (closed) nAChR [5]. The result from these experiments suggests that TCAs specifically displace [3H]TCP from its high-affinity site in a mutually exclusive manner in either the resting (closed) or desensitized state. Complementary results indicate that PCP completely inhibits [3H]imipramine binding to nAChRs in both states in a concentration-dependent manner (see Table 1). In fact, Schild-type analyses indicate that imipramine-induced inhibition of [3H]TCP binding to the nAChR in each conformational state is mediated by steric mechanisms [24]. These results support our hypothesis that the TCA binding site overlaps the PCP locus in each conformational state [5,25]. The exact location for the PCP site is still not firmly established (reviewed in [1,25]). Many lines of indirect evidence support a single site for PCP (one on each conformational state) within the ion channel of the muscle-type nAChR [10,26–28]. In the resting (closed) nAChR, the PCP site has been located between the valine ring (position 13′) and the extracellular ring (position 29′) [10,28,29] (reviewed in [1,25]). Whereas in the desensitized state, the PCP binding site is located closer to the middle of the ion channel [19,30] (reviewed in [1,25]). The results from the second set of experiments demonstrate that [3H]2-azidoimipramine photoincorporates into each nAChR subunit in the presence of Carb (desensitized state) with approximately 25% of the labeling inhibitable by addition of TCP (specific labeling). In contrast, TCP has no significant effect on the extent of labeling in the resting (closed) state (data not shown). The lack of any detectable TCP-inhibitable [3H]2-azidoimipramine photoincorporation in the αsubunit in the presence of α-BgTx (resting/closed state) is in accord with the radioligand binding results indicating that the imipramine binding affinity to the resting (closed) nAChR is 5–10 times lower than that in the desensitized state (see Table 1 and Refs. [5,24]). The fact that TCP, the structural and functional analog of PCP, inhibits [3H]2-azidoimipramine photoincorporation in the desensitized state supports the conclusion, obtained by radioligand binding competition experiments (see Fig. 2, Table 1, and Refs. [5,24]), that both imipramine and PCP bind to overlapping sites within the ion channel (see previous paragraph). In the desensitized state, the majority of the specific 3H-labeling within the α-subunit maps the transmembrane segment M2 (although the possibility that specific [3H]2-azidoimipramine labeling resides within the M3 segment cannot be excluded), suggesting that imipramine binds primarily within the ion channel of the receptor. A small amount of TCP-inhibitable [3H]2-azidoimipramine labeling mapped to the M1 transmembrane segment. Whether this labeling represents an additional binding site or labeling of the M1 segment from within the pore of the ion channel has yet to be established. In this regard, previous photolabeling experiments using [3H]ethidium diazide as a probe for the PCP site, showed that, in addition to residues at the M2 transmembrane segment, amino acids at the M1 transmembrane segment were labeled [30]. However, our dynamics studies do not support the possibility that PCP or imipramine interacts with other transmembrane segments than M2 (see Fig. 9). The photolabeling and radioligand binding results obtained experimentally are consistent with the data from our docking and molecular dynamics experiments (Fig. 9). We began simulations with docking of imipramine and PCP molecules into the ion pore. When we applied molecular dynamics simulations to lowest docked energy poses, both ligands sampled overlapping sites within a strip of residues into the ion pore. Fig. 9A and B show that imipramine fits nicely into the ion pore in a domain comprising valine (position 13′) and serine (position 6′) rings. This result is in contrast to an earlier study on the α7 nAChR [23] in which PCP bound deeper in the ion channel, proximal to the two serine rings located at position 2′ and 6′, respectively. On the other hand, Fig. 1 shows that the two positive charges on imipramine are shielded within the center of the molecule and distributed as partial atomic charges over many atoms. As a result, strong short-range electrostatic interactions are prevented. This effect is in contrast to our previous studies with adamantane-amine derivatives in which the single positive charge was exposed on the surface of the molecule [10]. It is interesting that we observed overlapping binding sites for imipramine when the lumen of the ion channel was kept relatively constant by tethering the backbone atoms of the M2 α-helices to their position in the nAChR template (PDB ID 2BG9) [17]. There is general agreement that the structure of the ion channel will change during the transitions between the resting (closed) and desensitized states [18,31,32], and recent molecular dynamics studies of nAChR [33] and the ion channel from the homologous glycine receptor [34] have revealed unexpectedly large fluctuations of the dimensions of the ion pore. These large fluctuations are consistent with the significant experimental differences in [3H]2-azidoimipramine labeling between the resting (closed) and desensitized states. Since the Torpedo nAChR remains the prototype for the study of all members of the Cys-loop ligand-gated ion channel superfamily, it may be possible that ADs inhibit neuronal nAChRs by mechanisms similar to the ones described for the muscle nAChR [5]. Studies of ACh-current generated by rat neuronal α2β4 and mouse muscletype nAChRs expressed in Xenopus laevis oocytes showed that imipramine inhibits both types of nAChRs with similar potency and through similar mechanisms [35]. As was outlined above, both muscle- and neuronal-type nAChRs have homologous structures. Therefore, the evidence presented here may help to further determine the primary structural components of TCA binding sites on the muscle-type as well as neuronal-type nAChRs. We plan to further examine the contribution of α1 M2 and M1 amino acids to imipramine binding and pharmacological action, specifically by site-directed mutagenic amino acid substitutions and electrophysiological recordings to asses the effect of imipramine on receptor function. Acknowledgements This research was supported by intramural grants from Texas Tech University Health Sciences Center (SOM Seed Grant and Roger Alan Valkenaar Estate) (to M.P.B), by research grants from the Science Foundation Arizona and from the Office of Research and Sponsored Programs, Midwestern University (to H.R.A.), by a NIH-NIAAA grant (AA13378) (to J.R.T.), and by the FOCUS research subsidy from the Foundation for Polish Science (to K.J.). The authors thank Xiao Juan Yuan, Jorgelina L. Arias Castillo, and Paulina Iacoban for their technical M. Sanghvi et al. / Biochimica et Biophysica Acta 1778 (2008) 2690–2699 assistance. The authors also thank NIDA (NIH, Bethesda, Maryland, USA) for its gift of phencyclidine. References [1] H.R. Arias, P. Bhumireddy, C. Bouzat, Molecular mechanisms and binding site locations for noncompetitive antagonists of nicotinic acetylcholine receptors, Int. J. Biochem. Cell Biol. 38 (2006) 1254–1276. [2] H.R. Arias, Ligand-gated ion channel receptor superfamilies, in: H.R. Arias (Ed.), Biological and Biophysical Aspects of Ligand-Gated Ion Channel Receptor Superfamilies, Research Signpost, India, 2006, pp. 1–25. [3] C.N. Connolly, K.A. Wafford, The Cys-loop superfamily of ligand-gated ion channels: the impact of receptor structure and function, Biochem. Soc. Transac. 32 (2004) 529–534. [4] H.A. Lester, M.I. Dibas, D.S. Dahan, J.F. Leite, D.A. Dougherty, Cys-loop receptors: new twists and turns, Trends Neurosci. 27 (2004) 329–336. [5] F. Gumilar, H.R. Arias, G. Spitzmaul, C. Bouzat, Molecular mechanisms of inhibition of nicotinic acetylcholine receptors by tricyclic antidepressants, Neuropharmacology 45 (2003) 964–976. [6] A. Rotman, V. Pribluda, Photoaffinity labelling of the serotonin carrier protein in platelets and brain synaptosomes, Biochim. Biophys. Acta 714 (1982) 173–176. [7] S.E. Pedersen, S.D. Sharp, W.S. Liu, J.B. Cohen, Structure of the non-competitive antagonist-binding site of the Torpedo nicotinic acetylcholine receptor. [3H] Meproadifen mustard reacts selectively with α-subunit Glu-262, J. Biol. Chem. 267 (1992) 10489–10499. [8] A.K. Hamouda, M. Sanghvi, D. Sauls, T.K. Machu, M.P. Blanton, Assessing the lipid requirements of the Torpedo californica nicotinic acetylcholine receptor, Biochemistry 45 (2006) 4327–4337. [9] M.A. Moore, M.P. McCarthy, Snake venom toxins, unlike smaller antagonists, appear to stabilize a resting (closed) state conformation of the nicotinic acetylcholine receptor, Biochim. Biophys. Acta 1235 (1995) 336–342. [10] H.R. Arias, J.R. Trudell, E.Z. Bayer, B. Hester, E.A. McCardy, M.P. Blanton, Noncompetitive antagonist binding sites in the Torpedo nicotinic acetylcholine receptor ion channel. Structure–activity relationship studies using adamantane derivatives, Biochemistry 42 (2003) 7358–7370. [11] Y.C. Cheng, W.H. Prusoff, Relationships between the inhibition constant (Ki) and the concentration of inhibitor which causes 50 per cent inhibition (IC50) of an enzymatic reaction, Biochem. Pharmacol. 22 (1973) 3099–3108. [12] U.K. Laemmli, Cleavage of structural proteins during the assembly of the head of bacteriophage T4, Nature 227 (1970) 680–685. [13] R.E. Middleton, J.B. Cohen, Mapping of the acetylcholine binding site of the nicotinic acetylcholine receptor: [3H]nicotine as an agonist photoaffinity label, Biochemistry 30 (1991) 6987–6997. [14] M.P. Blanton, E.A. McCardy, A. Huggins, D. Parikh, Probing the structure of the nicotinic acetylcholine receptor with the hydrophobic photoreactive probes [125I] TID-BE and [125I]TIDPC/16, Biochemistry 37 (1998) 14545–14555. [15] D.W. Cleveland, S.G. Fischer, M.W. Kirschner, U.K. Laemmli, Peptide mapping by limited proteolysis in sodium dodecyl sulfate and analysis by gel electrophoresis, J. Biol. Chem. 252 (1977) 1102–1106. [16] M.P. Blanton, J.B. Cohen, Identifying the lipid–protein interface of the Torpedo nicotinic acetylcholine receptor: secondary structure implications, Biochemistry 33 (1994) 2859–2872. 2699 [17] N. Unwin, Refined structure of the nicotinic acetylcholine receptor at 4 Å resolution, J. Mol. Biol. 346 (2005) 967–989. [18] A. Miyazawa, Y. Fujiyoshi, N. Unwin, Structure and gating mechanism of the acetylcholine receptor pore, Nature 423 (2003) 949–955. [19] O.R. Pagán, V.A. Eterović, M. Garcia, D. Vergne, C.M. Basilio, A.D. Rodriguez, R.M. Hann, Cembranoid and long-chain alkanol sites on the nicotinic acetylcholine receptor and their allosteric interaction, Biochemistry 40 (2001) 11121–11130. [20] M.P. Blanton, E.A. McCardy, M.J. Gallagher, Examining the noncompetitive antagonist-binding site in the ion channel of the nicotinic acetylcholine receptor in the resting state, J. Biol. Chem. 275 (2000) 3469–3478. [21] K. Jozwiak, S. Ravichandran, J.R. Collins, I.W. Wainer, The interactions of noncompetitive inhibitors with an immobilized α3β4 nicotinic acetylcholine receptor investigated by affinity chromatography, QSAR and molecular docking, J. Med. Chem. 47 (2004) 4008–4021. [22] K. Jozwiak, S. Ravichandran, J.R. Collins, R. Moaddel, I.W. Wainer, Interaction of noncompetitive inhibitors with the α3β2 nicotinic acetylcholine receptor investigated by affinity chromatography and molecular docking, J. Med. Chem. 50 (2007) 6279–6283. [23] M.O. Ortells, G.E. Barrantes, Understanding channel blocking in the nicotinic acetylcholine receptor, Receptors Channels 7 (2001) 273–288. [24] C.L. Sullivan, M. Crundsen, H.R. Arias, Characterization of Antidepressant Binding Sites on the Nicotinic Acetylcholine Receptor. 2008 ACCP Spring Practice and Research Forum, Phoenix, AZ, USA, April 5–9, 2008. [25] H.R. Arias, W.R. Kem, J.R. Trudell, M.P. Blanton, Unique general anesthetic binding sites within distinct conformational states of the nicotinic acetylcholine receptor, Int. Rev. Neurobiol. 54 (2002) 1–50. [26] M.J. Eaton, C.G. Labarca, V.A. Eterović, M2 mutations of the nicotinic acetylcholine receptor increase the potency of the non-competitive inhibitor phencyclidine, J. Neurosci. Res. 61 (2002) 44–51. [27] V.A. Eterović, R. Lu, A.E. Eakin, A.D. Rodriguez, P.A. Ferchmin, Determinants of phencyclidine potency on the nicotinic acetylcholine receptors from muscle and electric organ, Cell Mol. Neurobiol. 19 (1999) 745–757. [28] H.R. Arias, E.A. McCardy, E.Z. Bayer, M.J. Gallagher, M.P. Blanton, Allosterically linked noncompetitive antagonist binding sites in the resting nicotinic acetylcholine receptor ion channel, Arch. Biochem. Biophys. 403 (2002) 121–131. [29] H.R. Arias, P. Bhumireddy, G. Spitzmaul, J.R. Trudell, C. Bouzat, Molecular mechanisms and binding site location for the noncompetitive antagonist crystal violet on nicotinic acetylcholine receptors, Biochemistry 45 (2006) 2014–2026. [30] M.B. Pratt, S.E. Pedersen, J.B. Cohen, Identification of the sites of incorporation of [3H]ethidium diazide within the Torpedo nicotinic acetylcholine receptor ion channel, Biochemistry 39 (2000) 11452–11462. [31] E. Bertaccini, J.R. Trudell, E. Lindahl, Normal mode analysis reveals the channel gating motion within a ligand gated ion channel model, Int. Cong. Ser. 1283 (2005) 160–163. [32] P. Purohit, A. Mitra, A. Auerbach, A stepwise mechanism for acetylcholine receptor channel gating, Nature 446 (2007) 930–933. [33] X. Cheng, I. Ivanov, H. Wang, S.M. Sine, J.A. McCammon, Nanosecond-timescale conformational dynamics of the human α7 nicotinic acetylcholine receptor, Biophys. J. 93 (2007) 2622–2634. [34] M.H. Cheng, M. Cascio, R.D. Coalson, Homology modeling and molecular dynamics simulations of the α1 glycine receptor reveals different states of the channel, Proteins 68 (2007) 581–593. [35] H.E. Lopez-Valdés, J. García-Colunga, Antagonism of nicotinic acetylcholine receptors by inhibitors of monoamine uptake, Mol. Psych. 6 (2001) 511–519.