Biochemistry 2003, 42, 13457-13467
13457
Identification of Nicotinic Acetylcholine Receptor Amino Acids Photolabeled by the Volatile Anesthetic Halothane† David C. Chiara,‡ Lawrence J. Dangott,‡,§ Roderic G. Eckenhoff,*,| and Jonathan B. Cohen*,‡ Department of Neurobiology, HarVard Medical School, Boston, Massachusetts 02115, and Department of Anesthesia, UniVersity of PennsylVania, Philadelphia, PennsylVania 19104 ReceiVed July 3, 2003; ReVised Manuscript ReceiVed September 12, 2003
ABSTRACT: To identify inhalational anesthetic binding domains in a ligand-gated ion channel, we photolabeled nicotinic acetylcholine receptor (nAChR)-rich membranes from Torpedo electric organ with [14C]halothane and determined by Edman degradation some of the photolabeled amino acids in nAChR subunit fragments isolated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and highperformance liquid chromatography. Irradiation at 254 nm for 60 s in the presence of 1 mM [14C]halothane resulted in incorporation of ∼0.5 mol of 14C/mol of subunit, with photolabeling distributed within the nAChR extracellular and transmembrane domains, primarily at tyrosines. γTyr-111 in ACh binding site segment E was labeled, while RTyr-93 in segment A was not. Within the transmembrane domain, RTyr213 within RM1 and δTyr-228 within δM1 were photolabeled, while no labeled amino acids were identified within the δM2 ion channel domain. Although the efficiency of photolabeling at the subunit level was unaffected by agonist, competitive antagonist, or isoflurane, state-dependent photolabeling was seen in a δ subunit fragment beginning at δPhe-206. Labeling of δTyr-212 in the extracellular domain was inhibited >90% by d-tubocurarine, whereas addition of either carbamylcholine or isoflurane had no effect. Within M1, the level of photolabeling of δTyr-228 with [14C]halothane was increased by carbamylcholine (90%) or d-tubocurarine (50%), but it was inhibited by isoflurane (40%). Within the structure of the nAChR transmembrane domain, δTyr-228 projects into an extracellular, water accessible pocket formed by amino acids from the δM1-δM3 R-helices. Halothane photolabeling of δTyr-228 provides initial evidence that halothane and isoflurane bind within this pocket with occupancy or access increased in the nAChR desensitized state compared to the closed channel state. Halothane binding at this site may contribute to the functional inhibition of nAChRs.
At clinically effective concentrations, most general anesthetics modulate the function of ligand-gated ion channels in the superfamily that includes nicotinic acetylcholine receptors (nAChRs)1 and serotonin 5-HT3 receptors with cation-selective channels, as well as the GABAA receptors and glycine receptors with anion-selective channels (1-3). † This research was supported by grants from the National Institutes of Health [GM58448 (J.B.C.) and GM51595 (R.G.E.)] and by an award to Harvard Medical School in Structural Neurobiology from the Keck Foundation. * To whom correspondence should be addressed. R.G.E.: Department of Anesthesia, University of Pennsylvania Medical Center, 311 John Morgan Bldg., 3620 Hamilton Walk, Philadelphia, PA 191046112; phone, (215) 662-3766; e-mail, Roderic.Eckenhoff@ uphs.upenn.edu. J.B.C.: Department of Neurobiology, Harvard Medical School, 220 Longwood Ave., Boston, MA 02115; phone, (617) 4321728; fax, (617) 734-7557; e-mail,
[email protected]. ‡ Harvard Medical School. § Current address: Department of Chemistry, Texas A&M University, College Station, TX 77843. | University of Pennsylvania. 1 Abbreviations: Carb, carbamylcholine; dTC, d-tubocurarine; DTT, dithiothreitol; EDTA, ethylenediaminetetraacetate; EndoLys-C, endoproteinase Lys-C; GABAA, γ-aminobutyric acid type A; HPLC, highpressure liquid chromatography; nAChR, nicotinic acetylcholine receptor; OPA, o-phthalaldehyde; PAGE, polyacrylamide gel electrophoresis; SDS, sodium dodecyl sulfate; Tricine, N-tris(hydroxymethyl)methylglycine; Tris, tris(hydroxymethyl)aminomethane; V8 protease, Staphylococcus aureus endopeptidase Glu-C.
General anesthetics enhance agonist action for most of the receptors with anion-selective channels, while they inhibit noncompetitively nAChRs. Members of this superfamily are composed of five homologous subunits arranged pseudosymmetrically about a central axis that is the ion channel. Each subunit has a large N-terminal segment that contributes to the receptor extracellular domain and four transmembrane segments (M1-M4). The agonist binding sites are contained in the extracellular domain at subunit interfaces. The M2 segments from each subunit are R-helical and contribute to the lumen of the ion channel, possibly with additional contributions from the M1 segments, while amino acids from the M3 and M4 segments contribute to the lipid-protein interface (4, 5). In the GABAA receptor, mutational analyses have identified three positions, in the M1, M2, and M3 segments, which modulate the enhancing actions of alcohol and volatile anesthetics (6, 7) and are hypothesized to contribute to an anesthetic binding pocket. In muscle nAChRs, substitutions at position M2-10′ modulate the inhibitory potency of long chain alcohols and isoflurane (8-10), while for neuronal nAChRs, substitutions at position M2-15′, a position identified in GABAA receptors as a potential contributor to an anesthetic binding site, as well as substitutions of the amino acids linking the M2 and M3 segments have been identified as determinants of volatile anesthetic sensitivity (11, 12).
10.1021/bi0351561 CCC: $25.00 © 2003 American Chemical Society Published on Web 10/29/2003
13458 Biochemistry, Vol. 42, No. 46, 2003 In the absence of atomic-resolution structures of these receptors in the presence of anesthetics, it is difficult to decide whether the positions where substitutions alter anesthetic potency contribute directly to anesthetic binding sites or are involved in the transduction mechanism and allosterically modulate anesthetic potency. Photoaffinity labeling provides a complementary approach to identifying amino acids contributing directly to drug binding sites (reviewed in refs 13 and 14). For the nAChR, photoreactive agonists and antagonists have provided extensive identification of the amino acids contributing to the agonist binding sites and to the ion channel (15, 16), and [3H]azioctanol, a general anesthetic containing a photoreactive diazirine, reacted with high efficiency with RGlu-262 (M2-20′) at the C-terminal (extracellular) end of the M2 ion channel domain (17). Halothane (2-chloro-2-bromo-1,1,1-trifluoroethane), a clinically important volatile anesthetic, produces anesthesia with an EC50 of 0.3 mM and inhibits nAChRs with IC50 values varying from 0.1 mM for some neuronal nAChR subtypes to 0.8 mM for muscle nAChRs (18). The carbon-bromine bond of halothane is unstable under UV irradiation, and the resultant radical intermediate reacts with fatty acids (19) and with poly(L-lysine) (20). In addition, [14C]halothane can be photoincorporated into amino acids within soluble and integral membrane proteins (21), and individual photolabeled aromatic amino acids (tryptophans) have been identified within one of the fatty acid/drug binding sites of serum albumin (22), apomyoglobin (23; also His), and within the retinal binding pocket of rhodopsin (24). For Torpedo nAChR-rich membranes equilibrated with [14C]halothane at anesthetic doses, brief UV irradiation resulted in covalent incorporation of 14C into each nAChR subunit, with the level of photolabeling reduced by higher concentrations of nonradioactive halothane or isoflurane, but not by a nAChR agonist (carbamylcholine) or a competitive antagonist (Rbungarotoxin) (25). In this report, we use protein chemistry techniques to identify some of the nAChR amino acids photolabeled with [14C]halothane. EXPERIMENTAL PROCEDURES Materials. nAChR-rich membranes were isolated from fresh Torpedo californica electric organs as described previously (26). Membranes contained 1-1.4 nmol of [3H]ACh binding sites per milligram of protein and were stored until they were used at a concentration of 4-8 mg of protein/ mL at -80 °C in 38% sucrose and 0.02% sodium azide. [14C]Halothane (50 mCi/mmol) was obtained as a custom synthesis from New England Nuclear and was used shortly after being received to reduce the level of contamination from β-degradation products. The material was routinely verified to have >95% of the radioactivity in the halothane HPLC peak prior to being used. Nonradioactive halothane and isoflurane were from Halocarbon Laboratories (Hackensack, NJ) and Anaquest, Inc. (Madison, WI). Carbamylcholine and d-tubocurarine were from Sigma. 1-Azidopyrene was from Molecular Probes. Staphylococcus aureus glutamyl endopeptidase (V8 protease) was purchased from ICN Biomedical, endoproteinase Lys-C (EndoLys-C) from Roche Biochemical, and TPCK-treated trypsin from Worthington Biochemical.
Chiara et al. Photolabeling nAChR-Rich Membranes with [14C]Halothane. nAChR-rich membranes were resuspended at a concentration of 1.5 mg of protein/mL in deoxygenated Torpedo physiological saline [250 mM NaCl, 5 mM KCl, 3 mM CaCl2, 2 mM MgCl2, and 5 mM sodium phosphate (pH 7.0)] and equilibrated under argon for 30 min on ice prior to photolabeling. Reagents were assembled in Teflonstoppered quartz 2 mL cuvettes to achieve the concentrations listed in Results, with [14C]halothane being the final addition. Samples were mixed and equilibrated for 1 min before being irradiated for 60 s with an Oriel low-pressure Hg(Ar) pencil calibration lamp at 5 mm with constant mixing. The dominant peak from this source is at 254 nm. The cuvette contents were diluted 10-fold and washed iteratively by centrifugation. To visualize nAChR subunits resolved by preparative SDS-PAGE without staining and destaining the gel, after photolysis with [14C]halothane, membrane suspensions were photolabeled with 1-azidopyrene, a fluorescent photoreactive hydrophobic compound (27), and then pelleted and washed as described above. Gel Electrophoresis. Photolabeled nAChR-rich membranes were separated by SDS-PAGE as described previously (28). After electrophoresis, the unstained gel was visualized under UV light, and the nAChR subunits were excised on the basis of the fluorescence of the incorporated 1-azidopyrene. The subunit bands were eluted, filtered, concentrated, acetone precipitated, and resuspended as described previously (29). For fluorography, gels were treated with Amplify (Amersham Pharmacia Biotech), dried, and exposed to Kodak X-OMAT AR film at -80 °C. For quantitation of the 14C in gel slices, slices were soaked in 5 mL of gel cocktail [90% toluene, 10% TS-2 tissue solubilizer (Research Products International Corp.) with 12.6 mM 2,5-diphenyloxazole and 770 µM 1,4bis(5-phenyloxazol-2-yl)benzene] for 3-5 days and counted. Enzymatic Digestions. In-gel digestions of the nAChR R subunit with S. aureus V8 protease were carried out as described previously (30). Solution digestions with V8 protease (1:1, w:w, 25 °C, 2-3 days) were performed in storage buffer [10 mM NaPO4, 1 mM DTT, 1 mM EDTA, and 0.1% SDS (pH 7.0)]. For digestion with EndoLys-C (0.5-1 unit/digest, 3 weeks, 25 °C), the subunits and/or fragments were resuspended in 25 mM Tris, 0.5 mM EDTA, and 0.1% SDS (pH 8.6). Trypsin digests (1:1, w:w, 2-3 days, 25 °C) were performed in storage buffer supplemented with 0.5% Genapol C-100 (Calbiochem). ReVersed-Phase HPLC. HPLC separations and/or purifications of nAChR subunit fragments were performed using either Waters 510 pumps controlled by a 680 gradient controller or an Agilent 1100 Binary HPLC system with a degasser and column heating compartment (set at 40 °C). Separations were achieved using a Brownlee Aquapore butyl 7 µm, 100 mm × 2.1 mm column with a C-2 guard column. HPLC solvents are noted in the figure legends, and gradients are included in the figures. N-Terminal Sequence Analysis. Edman degradation was performed on an Applied Biosystems 477 gas-phase sequencer modified to analyze one-third of each cycle with a 120A amino acid analyzer and to collect two-thirds of each cycle for scintillation counting. Data shown in the figures are the actual counts per minute (cpm) and backgroundsubtracted picomoles (determined by peak height) of the
Sites of [14C]Halothane Photoincorporation in the nAChR detected amino acid in each cycle. The picomoles detected were fit (nonlinear, least squares, SigmaPlot, SPSS Inc.) to the equation f(x) ) I0Rx, where f(x) is the backgroundsubtracted picomoles in cycle x, I0 is the initial amount of peptide sequenced, and R is the repetitive yield. The fit of this equation is included as a dotted line in figures of sequence runs. Cys, Ser, Trp, and His were excluded from this analysis because of known problems with their quantitation via Edman degradation. The specific 14C incorporation (cpm per picomole) in cycle x was calculated as (cpmx cpmx-1)/(2I0Rx). The total amount of peptide sequenced was assumed to be 3I0. HPLC fractions of interest were either drop-loaded onto Beckman peptide supports (no. 290111) placed on a 45 °C heating block or pooled, concentrated by centrifugal evaporation, resuspended in a minimal volume of 0.1% SDS (∼40 µL), and loaded. When samples containing SDS were sequenced, the filters were first treated with gas trifluoroacetic acid (5 min) followed by an ethyl acetate wash (4 min) to remove excess detergent. In some cases, the sequencing run was interrupted and the material on the filter was treated with o-phthalaldehyde (OPA) as described previously (26). OPA reacts with primary amines preferentially over secondary amines (i.e., proline) and can be used at any sequencing cycle to block Edman degradation of peptides not containing an N-terminal proline (31). Models of the Torpedo nAChR. A homology model of the Torpedo nAChR extracellular domain (N-terminus to the beginning of M1 for each subunit) was constructed from the structure of the AChBP (32) by using the Homology module in Insight II (Accelrys) on a Silicon Graphics O2 workstation as described previously (15, 33). The coordinates for the AChBP structure (Protein Data Bank entry 1I9B) and for the recently published structure (34) at 4 Å resolution of the Torpedo nAChR transmembrane domain (Protein Data Bank entry 1OED) were obtained from the Research Collaboratory for Structural Bioinformatics. The percent solvent exposures of the amino acids in the models that were labeled in the nAChR with [14C]halothane were calculated by use of the access_surf function in the NMR module which determines the Connolly surface area of each amino acid in a structure using a 1.4 Å diameter ball. RESULTS Previous photolabeling of nAChR-rich membranes with [14C]halothane demonstrated incorporation of 14C into protein and lipid fractions. All nAChR subunits were labeled, and nAChR subunit labeling was not altered by the presence of agonist [carbamylcholine (Carb)] or antagonist (R-bungarotoxin), but the level of nAChR subunit labeling was reduced 75% by excess nonradioactive halothane, as was the extent of lipid labeling (25). To identify nAChR subunit amino acids photolabeled with [14C]halothane, we carried out labelings on a preparative scale (∼15 mg of protein, 7.5 nmol of nAChR) so that Edman degradation could be used to identify sites of 14C incorporation in nAChR subunit fragments isolated by HPLC and/or SDS-PAGE. To also provide an initial definition of the pharmacological specificity of the photolabeling, membrane suspensions were photolabeled with [14C]halothane in the absence of other drugs; in the presence of Carb, which would occupy the ACh sites and stabilize the nAChR in the desensitized state; in the presence
Biochemistry, Vol. 42, No. 46, 2003 13459
FIGURE 1: [14C]Halothane photoincorporation into nAChR-rich membranes. nAChR-rich membranes (1 nmol of ACh sites/mg of protein, 2 mg/mL) were equilibrated (5 min) with 1 mM [14C]halothane alone (3) or with 7 mM isoflurane (2), 100 µM Carb (5), or 10 µM dTC (4). The suspensions were irradiated for 60 s at 254 nm. Aliquots (50 µg) were separated by SDS-PAGE (8% acrylamide) along with an unlabeled sample (1). Polypeptides were visualized with Coomassie blue (A), and the gel was processed for fluorography (B, 14 day exposure). (C) After fluorography, we cut lane 3 into 2 mm slices to quantitate the distribution of 14C. The bottom of the sample well, the dye front, and the bottom of the gel contained 9400, 8200, and 9600 cpm, respectively. Indicated on the left are the Coomassie-stained bands corresponding to the nAChR subunits and the R subunit of the Na+/K+ ATPase (90K).
of d-tubocurarine (dTC), a competitive antagonist; or in the presence of isoflurane, another volatile anesthetic. Figure 1 shows a Coomassie-stained 8% polyacrylamide gel and fluorogram with 50 µg aliquots from suspensions photolabeled with 1 mM [14C]halothane under these four conditions. On the basis of the fluorogram, the presence of Carb, dTC, or isoflurane did not alter incorporation of 14C into the nAChR subunits, and the 14C distribution in the gel was similar to that of the Coomassie stain, with nAChR subunits and other polypeptides photolabeled as well as highmolecular mass aggregates produced by the UV irradiation. When 14C incorporation for the control sample was quantified by liquid scintillation counting of 2 mm gel slice analysis (Figure 1C), there was a background distribution of 14C throughout the gel at ∼1000 cpm/gel slice, with 14C incorporation in the nAChR R subunit at ∼ 2000 cpm above background and in the β, γ, and δ subunits at ∼1000 cpm. Since the 50 µg membrane aliquots contained ∼50 pmol of nAChR R subunit and the specific activity of 14C halothane was 50 Ci/mol, [14C]halothane was incorporated into the nAChR R subunit at ∼50 cpm/pmol or ∼0.5 mol/mol of R subunit. The level of incorporation of [14C]halothane into the nAChR subunits was also estimated from the 14C cpm and protein recovered from the preparative gel bands after the eluted material was precipitated with acetone and resuspended. From 15 mg of nAChR-rich membranes, we recovered 200-400 µg of the β, γ, and δ subunits with a 14 C incorporation of ∼1000-1200 cpm/µg of protein, or ∼0.5 mol of [14C]halothane/mol of subunit. The incorporation of 14C in those subunits was altered by less than 10% when nAChRs were labeled in the presence of Carb or isoflurane,
13460 Biochemistry, Vol. 42, No. 46, 2003 while labeling in the presence of dTC reduced the level of 14 C incorporation by 25-30%. Rather than elution of the gel bands containing the R subunit, they were transferred directly to a second “mapping” gel for digestion with S. aureus glutamyl endopeptidase (V8 protease). To provide an initial map of the 14C distribution within the nAChR R subunit, samples of the labeled R subunit were digested in gel with V8 protease to generate four large fragments of 20 (RV8-20), 18 (RV8-18), 10 (RV8-10), and 4 kDa (RV8-4) (30). RV8-20 begins at Ser-173 and contains ACh binding site segment C as well as the M1-M3 hydrophobic segments. RV8-18 begins at Val-46/Thr-52 and contains ACh binding site segments A and B. RV8-10 begins at Asn-339 and contains M4. RV8-4 begins at Ser-1. The bands corresponding to RV8-20, RV8-18, and RV8-10 were identified by Coomassie stain and excised, and the materials eluted from those bands were purified by reversed-phase HPLC and characterized by N-terminal sequence analysis. On the basis of the 14C associated with the purified R subunit fragments and their picomole yields calculated by sequence analysis and/or protein determination, [14C]halothane was incorporated in RV8-20 at a level of 30-50 cpm/pmol, in RV8-18 at 5-6 cpm/pmol, and in RV8-10 at 12-15 cpm/ pmol. Labeling of nAChRs in the presence of Carb, dTC, or isoflurane did not alter the 14C incorporation in RV8-10. Labeling in the presence of dTC reduced the level of 14C incorporation in RV8-18 by ∼50%, while for Carb, the extent of labeling was not reduced. Difficulties in reliable quantification of the low mass levels that were sequenced precluded a similar analysis for the RV8-20 samples. To determine whether the incorporated 14C was generally stable to the necessary cycles of acid and base, which it was, and whether any amino acids near the amino terminus of the δ subunit were photolabeled, an aliquot of intact δ subunit isolated from nAChR-rich membranes photolabeled with [14C]halothane in the presence of Carb was sequenced for 20 cycles (21 000 cpm loaded, 17 430 cpm left). There was no release of 14C above background (
