Published on Web 11/01/2008

An Activity-Based Protein Profiling Probe for the Nicotinic Acetylcholine Receptor Mathew Tantama, Wan-Chen Lin, and Stuart Licht* Department of Chemistry, Massachusetts Institute of Technology, Building 16, Room 573B, Cambridge, Massachusetts 02139 Received July 27, 2008; E-mail: [email protected]

Activity-based protein profiling (ABPP) is rapidly becoming one of the essential experimental approaches in understanding biological processes at the systems level.1 ABPP reagents have been prepared for a number of important enzyme classes,2-4 but ion channels are one important class of proteins for which ABPP probes have not been previously reported. Activation and deactivation of ion channels are central to some of the most important processes in neurobiology, such as neuronal excitability and synaptic plasticity.5 A probe molecule that selectively labels a subset of the different activation states of a channel could be used as an activity-based probe. The nicotinic acetylcholine receptor (nAChR) is an ion channel in the Cys-Loop superfamily that becomes cation-permeable upon binding the neurotransmitter acetylcholine. Like other neurotransmitter-gated channels, nAChRs typically undergo desensitization: a transition into a long-lived inactive state in response to prolonged exposure to acetylcholine.6,7 In contrast to the closed states of nAChRs that predominate in the absence of a neurotransmitter, desensitized states typically have very high affinities for acetylcholine and nicotine8,9 and predominate in the presence of a neurotransmitter. ABPP probes could be used to help characterize protein-protein interactions and posttranslational modifications associated with desensitization and reactivation of nAChRs. Such probes would thus be useful for investigating the neurobiology of desensitization in nicotine addiction10,11 and neuromuscular disorders.12 We synthesized a candidate ABPP probe, named BPyneTEA (benzophenone-alkyne-triethylammonium), for state-dependent binding and photolabeling of nAChRs (Figure 1A). This candidate probe combines features of several “parent” structures that selectively bind to open or closed nAChRs.13-15 We therefore characterized its action on nAChRs both electrophysiologically and biochemically to assess the effect of combining these features in a single structure. To test the hypothesis that BPyneTEA can block both open and closed nAChRs, single-channel patch-clamp current recordings were obtained. To ensure that both open and closed states were observable, single-channel currents were recorded from a gain-offunction muscle-type nAChR mutant, RG153S,16 activated using the weak agonist choline. Single-channel activity occurs as clusters of openings and closings that represent the conformational transitions of exactly one nAChR. Clusters are normally terminated by entry into long-lived desensitized states. In the presence of BPyneTEA, however, a cluster of activity may be terminated early by blockade of either closed or open states (Supporting Figure 3A). Direct observation of individual BPyneTEA blockade events at the single-molecule level supports the hypothesis that this molecule binds both the open and closed states (Figure 1B and C). Blockade of the open state truncates open intervals within a cluster,15,17 decreasing the mean open time with increasing BPyneTEA concentration (Figure 1D and Supporting Figure 3B). 15766

9

J. AM. CHEM. SOC. 2008, 130, 15766–15767

Figure 1. Single-channel electrophysiology demonstrates that BPyneTEA

(A) binds and blocks the open and closed states of the nAChR. Singlechannel activity occurs in clusters in the absence of BPyneTEA (B), but clusters are terminated early in the presence of 500 µM BPyneTEA (C). The decrease in the mean open time (D) and the lifetime of the fastest closed time component (E) with increasing BPyneTEA concentration indicate that BPyneTEA binding to the open and closed nAChR has an association rate constant of ∼106 M-1 · s-1.

From the concentration dependence, the association rate constant for open-state blockade is (1.3 ( 0.7) × 106 M-1 · s-1 (best fit value ( standard error). This association rate constant depends on transmembrane voltage; an analysis of the voltage dependence suggests that BPyneTEA binds the nAChR 20 ( 10% into the transmembrane electrical field relative to the cell surface (Supporting Figure 4). Blockade of the closed state truncates closed intervals within a cluster, decreasing the observed mean closed times with increasing BPyneTEA concentration (Figure 1E and Supporting Figure 3B). In this case, there are multiple kinetic components in the closed time distribution, but there is only one that decreases in a BPyneTEA-dependent fashion; this component represents the lifetime of the closed (and blocker-free) state. The association rate constant for binding to the closed state was determined from the BPyneTEA concentration-dependent decrease of the fastest closed time component and is (5 ( 2) × 106 M-1 · s-1 (best fit value ( standard error). Both open and closed states thus bind BPyneTEA with an association rate constant of ∼1 × 106 M-1 · s-1. The difference in blockade rate constants between the open and closed 10.1021/ja805868x CCC: $40.75  2008 American Chemical Society

COMMUNICATIONS 22

be detected. Optimization of the blocker and benzophenone moieties may allow improved selectivity for the closed state; the modular design of the molecule is expected to enable facile synthesis of second-generation probes. The potential utility of a channel-targeted ABPP strategy depends on whether it will be generalizable to a large number of structurally distinct channels. Large changes in pore structure (as judged by accessibility to reactive probes in solution) have been observed for other Cys-Loop receptors such as the serotonin receptor,23 as well as glutamate receptors24 and potassium channels.25 In addition, the many characterized state-selective channel blockers and inhibitors offer a rich set of potential pore-binding groups for ion channel targeted ABPP probes. Ion channels as a class thus share many of the advantages of enzyme active sites as ABPP targets and appear likely to be a generally useful target for ABPP techniques.

Figure 2. Photolabeling nAChRs expressed in live cells in the absence

(closed) or presence (desensitized) of acetylcholine (ACh) demonstrates that BPyneTEA preferentially binds closed receptors. Closed receptors are labeled ∼2-fold more efficiently than desensitized receptors at concentrations of 250 µM and 50 µM. At 10 µM BPyneTEA, differential labeling was not statistically significant.

states is not statistically significant (unpaired, two-sided t test, p ) 0.068). The kinetic studies also allow estimation of upper limits for dissociation constants of BPyneTEA binding to closed and open states: <∼20 µM for the closed state and <∼80 µM for the open state (Supporting Figure 3B). To test whether BPyneTEA selectively labels the closed (but activatable) state of the nAChR compared to the inactive desensitized state, we carried out photolabeling of nAChRs expressed in live HEK293 cells in the presence or absence of the desensitizing agonist acetylcholine. Closed or desensitized nAChRs were photolabeled with BPyneTEA, and copper(I)-catalyzed [3 + 2] cycloaddition (i.e., “click” chemistry, as adapted for bioconjugation18) of an azide-functionalized biotin was carried out to biotinylate the photolabeled receptors.19 Biotinylated receptors were captured on streptavidin-coated beads, and nAChRs were visualized by Western blotting with an antibody against the nAChR R subunit. Quantification of the captured nAChRs (normalized for expression levels) shows that, at BPyneTEA concentrations g50 µM, the closed state is labeled more efficiently than the desensitized state by a factor of ∼2 (Figure 2 and Supporting Table 1). At 10 µM BPyneTEA, weak labeling is observed, but its state selectivity is not statistically significant. Labeling was not observed in the absence of BPyneTEA or UV irradiation (Supporting Figure 1). Selectivity for closed states compared to desensitized states is likely to be a crucial parameter in determining the utility of probes for investigation of nAChR desensitization in ViVo. Because desensitization occurs primarily from the open state and is the thermodynamic minimum for the agonist-bound channel,20 only channel populations that spend most of their time in the closed state will remain activatable. The selectivity of BPyneTEA for closed over desensitized conformations is modest (∼2-fold) but high enough that comparison of subproteomes using mass spectrometry is expected to be feasible. The use of trypsin-catalyzed 18O labeling of peptides for relative quantification of subproteomes by mass spectrometry21 has allowed enrichments/depletions of <2-fold to

Acknowledgment. This work was supported by the Beckman Foundation and the MIT Department of Chemistry. Supporting Information Available: Synthetic, electrophysiology, and biochemical methods; discussion of blockade models and voltagedependence of blockade; statistical analysis of live-cell labeling; labeling results in the absence of light or BPyneTEA. This material is available free of charge via the Internet at https://pubs.acs.org. References (1) Jessani, N.; Cravatt, B. F. Curr. Opin. Chem. Biol. 2004, 8, 54–59. (2) Liu, Y. S.; Patricelli, M. P.; Cravatt, B. F. Proc. Natl. Acad. Sci. U.S.A. 1999, 96, 14694–14699. (3) Yee, M.; Fas, S. C.; Stohlmeyer, M. M.; Wandless, T. J.; Cimprich, K. A. J. Biol. Chem. 2005, 280, 29053–29059. (4) Salisbury, C. M.; Cravatt, B. F. Proc. Natl. Acad. Sci. U.S.A. 2007, 104, 1171–1176. (5) Kandel, E. R.; Schwartz, J. H.; Jessell, T. M. Principles of Neural Science, 4th ed.; McGraw-Hill Health Professions Division: New York, 2000. (6) Quick, M. W.; Lester, R. A. J. J. Neurobiol. 2002, 53, 457–478. (7) Wilson, G. G.; Karlin, A. Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 1241– 1248. (8) Sine, S. M.; Quiram, P.; Papanikolaou, F.; Kreienkamp, H. J.; Taylor, P. J. Biol. Chem. 1994, 269, 8808–8816. (9) Heidmann, T.; Bernhardt, J.; Neumann, E.; Changeux, J. P. Biochemistry 1983, 22, 5452–5459. (10) Mansvelder, H. D.; Keath, J. R.; McGehee, D. S. Neuron 2002, 33, 905– 919. (11) Giniatullin, R.; Nistri, A.; Yakel, J. L. Trends Neurosci. 2005, 28, 371– 378. (12) Elenes, S.; Ni, Y.; Cymes, G. D.; Grosman, C. J. Gen. Physiol. 2006, 128, 615–27. (13) Akk, G.; Steinbach, J. H. J. Physiol. 2003, 551, 155–168. (14) Garcia, G.; Chiara, D. C.; Nirthanan, S.; Hamouda, A. K.; Stewart, D. S.; Cohen, J. B. Biochemistry 2007, 46, 10296–10307. (15) Neher, E.; Steinbach, J. H. J. Physiol. 1978, 277, 153–176. (16) Sine, S. M.; Ohno, K.; Bouzat, C.; Auerbach, A.; Milone, M.; Pruitt, J. N.; Engel, A. G. Neuron 1995, 15, 229–239. (17) Colquhoun, D.; Hawkes, A. G. In Single-Channel Recording; Sakmann, B., Neher, E., Eds.; Plenum Press: New York, 1995; pp 397-482. (18) Wang, Q.; Chan, T. R.; Hilgraf, R.; Fokin, V. V.; Sharpless, K. B.; Finn, M. G. J. Am. Chem. Soc. 2003, 125, 3192–3193. (19) Speers, A. E.; Adam, G. C.; Cravatt, B. F. J. Am. Chem. Soc. 2003, 125, 4686–4687. (20) Auerbach, A.; Akk, G. J. Gen. Physiol. 1998, 112, 181–197. (21) Yao, X.; Freas, A.; Ramirez, J.; Demirev, P. A.; Fenselau, C. Anal. Chem. 2001, 73, 2836–2842. (22) Ramos-Fernandez, A.; Lopez-Ferrer, D.; Vazquez, J. Mol. Cell. Proteomics 2007, 6, 1274–1286. (23) Panicker, S.; Cruz, H.; Arrabit, C.; Slesinger, P. A. J. Neurosci. 2002, 22, 1629–1639. (24) Sobolevsky, A. I.; Beck, C.; Wollmuth, L. P. Neuron 2002, 33, 75–85. (25) Liu, Y.; Jurman, M. E.; Yellen, G. Neuron 1996, 16, 859–867.

JA805868X

J. AM. CHEM. SOC.

9

VOL. 130, NO. 47, 2008

15767

An Activity-Based Protein Profiling Probe for the Nicotinic Acetylcholine Receptor Mathew Tantama, Wan-Chen Lin, Stuart Licht* Department of Chemistry, Massachusetts Institute of Technology, Building 16, Room 573B, Cambridge, MA 02139

Supporting Information: 1. Synthetic Methods 2. Electrophysiology Methods 3. Live-Cell Photolabeling and Blotting Methods 4. Kinetic Analysis of Blockade Under Conditions of Clustered Single-Channel Activity 5. Woodhull Analysis of the Voltage-Dependence of Open-Channel Blockade 6. Live-Cell Photolabeling Ratios and Statistical Tests 7. Proposed Mechanism for BPyneTEA State-Selective Blockade 1. Synthetic Methods General Methods. 1-Amino-11-azido-3,6,9-trioxaundecane was purchased from Toronto Research Chemicals (North York, ON, Canada). Other chemicals were purchased from Sigma-Aldrich (Milwaukee, WI) as reagent grade and were used as supplied. Anhydrous methylene chloride and acetonitrile were packed in Sure/Seal™ bottles and were transferred under nitrogen with a syringe. 1H and 13C NMR spectra were recorded on a Varian Mercury 300 spectrometer. Chemical shifts () were calibrated to the residual solvent peak and were expressed in parts per million (ppm). Coupling constants (J) were reported in Hz. High resolution electrospray mass spectra were obtained by a Bruker Daltonics APEXIV 4.7 Tesla FT-ICR mass spectrometer. Analytical thin-layer chromatography (TLC) was performed using silica 60 F254precoated glass plates (EMD Chemicals Inc., Gibbstown, NJ). Compounds were visualized by staining with aqueous potassium permanganate. Normal-phase column chromatography was carried out on Merck® silica gel 60 (70-230 mesh, Aldrich). O

O

O

DCC, HOBT OH

H2N

O

DMF, 15 h, 58%

N H

NH2

NH2 BPyneNH2

O iodoacetic anhydride CH3CN, 4 h, 60%

O

O N H

BPyneI

I

N H

O 25% (v/v) NEt3/acetone 12 h, 93%

O

O N H

N H

N+

I

BPyneTEA

Scheme S1. Synthesis of BPyneTEA.

S1

Synthesis of BPyne-NH2. To a cooled (0 ˚C) solution of 4-pentynoic acid (196 mg, 2 mmol), 4,4’-diaminobenzophenone (638 mg, 3 mmol), and 1-hydroxybenzotriazole (HOBT; 270 mg, 2 mmol) in anhydrous DMF (20 mL) was added N,N’dicyclohexylcarbodiimide (DCC; 3 mL of 1.0 M solution in CH2Cl2, 3 mmol) dropwise under an atmosphere of dry N2. The mixture was stirred at 0 ˚C for 15 min, and the reaction was allowed to proceed at ambient temperature for 16 h. The reaction mixture was quenched by cold water (2 mL) at 0 ˚C and then filtered. The filtrate was concentrated by rotary evaporation and purified by silica gel chromatography (2:1 EtOAc:hexanes) to give BPyne-NH2 (337 mg, 58%) as a light-yellow solid. TLC (3:1 EtOAc:hexanes) Rf 0.40; 1H NMR (acetone-d6, 300 MHz):  9.49 (s, 1H), 7.79 (d, J = 8.7 Hz, 2H), 7.67 (d, J = 8.7 Hz, 2H), 7.61 (d, J = 8.9 Hz, 2H), 6.73 (d, J = 8.9 Hz, 2H), 5.53 (s, 2H), 2.66 (m, 2H), 2.55 (m, 2H), 2.39 (t, J = 2.7 Hz, 1H); 13C NMR (acetone-d6, 75 MHz):  193.7, 170.6, 154.0, 143.0, 134.9, 133.3, 131.3, 126.6, 119.1, 113.8, 83.9, 70.4, 36.6, 14.9; HRMSESI (m/z) calcd for C18H17N2O2 [M + H]+: 293.1285; found: 293.1287. Synthesis of BPyne-I. BPyne-NH2 (132 mg, 0.45 mmol) and iodoacetic anhydride (320 mg, 0.9 mmol) were dissolved in anhydrous CH3CN (10 mL) under an atmosphere of dry N2. The mixture was stirred at room temperature for 4 h, cooled to 0 ˚C for 0.5 h, and then filtered. The precipitate was washed with ice-cold CH3CN (2 × 15 mL) and dried under high vacuum to give BPyne-I (82 mg, 40%) as a white solid. The combined filtrate was concentrated, washed with MeOH (2 × 5 mL), and dried under high vacuum to provide more BPyne-I (41 mg, 20%). 1H NMR (DMF-d7, 300 MHz):  10.83 (s, 1H), 10.47 (s, 1H), 7.917.79 (m, 8H), 4.02 (s, 2H), 2.81 (t, J = 2.7 Hz, 1H), 2.70 (m, 2H), 2.58 (m, 2H); 13C NMR (DMF-d7, 75 MHz):  194.6, 171.3, 168.5, 144.5, 144.2, 133.9, 133.3, 132.2, 132.1, 119.5, 119.4, 84.4, 71.6, 36.8, 15.1, 1.3; HRMS-ESI (m/z) calcd for C20H18IN2O3 [M + H]+: 461.0357; found: 461.0369. Synthesis of BPyneTEA. BPyne-I (60 mg, 0.13 mmol) was dissolved in acetone (15 mL) and triethylamine (5 mL) was added to reach a final concentration of 25% (v/v). The reaction was allowed to proceed at room temperature for 12 h, and white precipitate formed slowly over the course of reaction. The supernatant was removed, and the precipitate was washed by diethyl ether (3 × 10 mL) and dried by a steady flow of dry N2. The solid was further dried under high vacuum to give BPyneTEA (55 mg, 75%) as a white powder. The supernatant and ether washes were combined and another 10 mL of diethyl ether was added, allowing the precipitation of BPyneTEA from the combined solution. The precipitate was collected, washed by diethyl ether (3 × 5 mL), and dried to provide additional BPyneTEA (13 mg, 18%). 1H NMR (CD3OD, 300 MHz):  7.797.77 (m, 8H), 4.22 (s, 2H), 3.70 (q, J = 7.2 Hz, 6H), 2.65 (m, 2H), 2.56 (m, 2H), 2.31 (t, J = 2.4 Hz, 1H), 1.40 (t, J = 7.2 Hz, 9H); 13C NMR (DMF-d7, 75 MHz):  194.6, 171.3, 164.7, 144.3, 133.7, 132.0 (2 peaks), 121.6, 119.3, 84.4, 71.6, 59.4, 55.0, 36.8, 15.1, 8.4; HRMSESI (m/z) calcd for C26H32N3O3 [M  I]+: 434.2438; found: 434.2457.

S2

O HN

NH O OH

S

HN

EDC, HOBT

O

NH H N

DMF, 15 h, 91% H2N

O

O

O

S

O

O

O

N3

O

N3

BiotinN3

Scheme S2. Synthesis of Biotin-N3. Synthesis of Biotin-N3. To a cooled (0 ˚C) mixture of biotin (122 mg, 0.5 mmol), Nethyl-N’-(3-dimethylaminopropyl)carbodiimide hydrochloride (EDC, 144 mg, 0.75 mmol), and HOBT (101 mg, 0.75 mmol) in anhydrous DMF (5 mL) was added 1-amino11-azido-3,6,9-trioxaundecane (55 mg, 0.25 mmol, diluted in 1 mL of anhydrous DMF) dropwise. The reaction mixture was stirred at 0 ˚C for 0.5 h and was then allowed to warm to ambient temperature. The mixture was stirred at room temperature for 24 h, concentrated by rotary evaporation, and then diluted with MeOH/CHCl3 (5:95, 15 mL). The organic phase was washed with 1 N HCl (aq) (5 × 10 mL), neutralized with saturated NaHCO3 (aq) (3 × 10 mL), and re-extracted with CHCl3 (5 × 10 mL). The combined organic phase was washed with brine (5 × 10 mL), dried over anhydrous Na2SO4, concentrated by rotary evaporation, and purified by silica gel chromatography (CHCl3  17:83 MeOH:CHCl3) to give biotin-N3 (101 mg, 91%) as a white solid. TLC (1:5 MeOH:CHCl3) Rf 0.40; 1H NMR (CD3OD, 300 MHz):  4.50 (m, 1H), 4.31 (dd, J = 8.0, 4.4 Hz, 1H), 3.723.58 (m, 10H), 3.54 (t, J = 5.4 Hz, 2H), 3.463.32 (m, 4H), 3.21 (m, 1H), 2.93 (dd, J = 12.6, 5.0 Hz, 1H), 2.71 (d, J = 12.6 Hz, 1H), 2.22 (t, J = 7.4 Hz, 2H), 1.801.54 (m, 4H), 1.521.48 (m, 2H); 13C NMR (CD3OD, 75 MHz):  176.2, 166.2, 71.8 (2 peaks), 71.7, 71.4, 71.3, 70.7, 63.5, 61.8, 57.2, 51.9, 41.2, 40.5, 36.9, 29.9, 29.6, 27.0; HRMSESI (m/z) calcd for C18H32N6NaO5S [M  Na]+: 467.2047; found: 467.2045. 2. Electrophysiology Methods HEK 293 cells were transiently transfected with a 2:1:1:1 mass ratio of adult mouse muscle , , , and  AChR cDNAs using calcium phosphate precipitation.1 Single-channel patch-clamp recordings were obtained in the cell-attached mode2. The pipette and bath solutions were Dulbecco’s phosphate buffered saline (DPBS, in mM): 137 NaCl, 2.7 KCl, 0.9 CaCl2, 0.5 MgCl2, 6.6 Na2HPO4, 1.5 KH2PO4, pH 7.3. For experiments at high choline concentration, the  subunit contained the gain-of-function G153S mutation,1,3 and the pipette solution included 1 mM choline and varying amounts of BPyneTEA. For voltage-dependence experiments wild-type receptor activity was evoked with 1 µM acetylcholine and varying amounts of BPyneTEA were included in the pipette solution. The membrane potential was usually -30 to -40 mV, and the command potential was varied from 0 to 200 mV during recording. Single-channel currents were recorded with an Axopatch 200B amplifier (Axon Instruments, Foster City, CA) through a low-pass Bessel filter at 10 kHz, and data were digitized at a sampling rate of 20 kHz using a NI 6040 E Data Acquisition Board (National Instruments, Austin, TX). The S3

QuB software suite (www.qub.buffalo.edu) was used to adjust for baseline drift, idealize events, and analyze data.4-6 OriginLab (OriginLab Corporation, Northampton, MA) and Matlab (The Mathworks, Natick, MA) were also used for non-linear least squares fitting and figure preparation. Open times in Figure 1 were measured as mean open times within clusters of openings; this was feasible because only one component was present in the open time distribution. In Figure 1 D, each data point represents one recording, and the error bar is the standard error of the mean open time within the recording. Because more than one component was observed in the closed time distribution, the time constants corresponding to closed time components of interest were calculated by maximum likelihood fitting within the QuB software suite. In Figure 1 E, each data point represents the fast closed component from one recording, determined by fitting the closed-time distribution in QuB. Errors for the fitted closed lifetimes are not reported by QuB, but are generally negligible compared to the patch-to-patch (experimental) variation shown by the scatter in the data. 3. Live-Cell Photolabeling and Blotting Methods AChR-expressing HEK 293 cells were grown in 35 mm dishes. Cells were washed with DPBS prior to addition of 1 mL DPBS containing BPyneTEA with or without 0.2 µM acetylcholine. Cells were irradiated7 for 1 hour at 4°C at 365 nm wavelength using a UVL-56 (UVP) handheld lamp at a distance of 5 cm and intensity of 1350 µW/cm2. The labeling solution was removed, and cells were gently dissociated and collected in 2 mL DPBS. For each experimental condition, cells from four 35 mm dishes were typically pooled. Cells were pelleted for 10 minutes at 1000 x g, and the supernatant was aspirated. Cells were lysed for 3 hours with gentle agitation at 4°C in 500 µL lysis buffer (50 mM HEPES, pH 8, 150 mM NaCl, 1% Triton-X100, 1X Roche EDTA-free protease inhibitor cocktail).8 Lysates were pelleted for 5 minutes at 10000 x g to remove insoluble material, and cleared lysates were split into equal aliquots. As a control for variability in expression levels, solubilized receptors from one aliquot of cleared lysate were captured with -bungarotoxin-functionalized Sepharose beads for 20 hours at 4°C. -bungarotoxin (Biotium, Hayward, CA) was coupled to CNBr-activated Sepharose (GE Healthcare, Piscataway, NJ) according to the manufacturer’s instructions. Following capture, beads were washed with wash buffers containing 50 mM HEPES, pH 8, 1% Triton-X100, 0.1 mg/mL BSA and NaCl at increasing concentrations of 150, 250, 500, and 1000 mM. Beads were eluted with SDS loading buffer9 for 1 hour at 25°C for SDS-PAGE and Western blotting. To assess photolabeling, a second aliquot of cleared lysate was brought to a final concentration of 0.1% SDS, 5% t-butanol, 100 µM tris[(1-benzyl-1H-1,2,3-triazol-4yl)methyl]amine (Sigma, St. Louis, MO), 1 mM tris-(2-carboxyethyl)phosphine, 100 µM biotin-azide reagent, 1 mM CuSO4 and subjected to click chemistry conditions for 3 hours at 25°C with stirring.10 Following the click reaction, lysates were dialyzed in 4 liters of 50 mM HEPES, pH 8, 150 mM NaCl, 0.1% SDS, 1% Triton-X100 for 3 hours with one dialysis buffer change to remove excess biotin-azide reagent. Biotinylated, soluble receptors were then captured with streptavidin-functionalized agarose beads (Thermo Scientific, Rockford, IL) for 12 hours at 4°C. Following capture, beads were

S4

washed with wash buffers containing 50 mM HEPES, pH 8, 0.1% SDS, 1% Triton-X100, 0.1 mg/mL BSA and NaCl at increasing concentrations of 150, 250, 500, and 1000 mM. Beads were eluted with reducing SDS loading buffer (2% SDS)9 for 20 minutes at 25°C followed by 5 minutes at 65°C, and the supernatant was removed while hot. Two additional elutions were carried out, and the eluates were pooled. Following SDS-PAGE of the pooled eluates on a 10% gel, protein was transferred to nitrocellulose, and protein transfer was quantified by reversible Ponceau S staining. Western blotting was conducted using a primary mouse anti-AChR  subunit antibody (Clone 26, BD Biosciences, San Jose, CA) at 1:250 dilution and a secondary horseradish peroxidaseconjugated goat anti-mouse antibody (BioRad, Hercules, CA) at 1:15000 dilution. SuperSignal West (Thermo Scientific) substrate was used for chemiluminescent detection. Blots were imaged, and densitometry was conducted using a CCD imaging system. For quantitative densitometry, the AlphaImager2200 software was used, and raw pixel intensities were quantified without image enhancement. The background was subtracted from bands of interest, and photolabeling levels were normalized for expression levels. To perform the background subtraction, the area of interest around a given band was defined and pixel density quantified. Then, an equal area of background immediately above and immediately below was quantified, the two background estimates averaged, and the background value subtracted from the band of interest. The background-subtracted band from the streptavidin eluate, representing the photolabeled protein, was normalized by dividing it by the background-subtracted expression-control band. For any given blot, 2 to 7 exposures were typically taken and densitometry conducted in order to assess the variability in quantification for a single blot. In Figure 2 and Supporting Figure 1, the entire blot image was enhanced in Adobe Photoshop using the AutoLevels option for qualitative visualization; however, the digitally enhanced images were not used for densitometry.

S5

Supporting Figure 1. Western blots of negative controls for state-dependent photolabeling experiments. Blots were developed using a chemiluminescent substrate, and images were acquired with a CCD camera. Densitometry was carried out using the AlphaImager2200 software. Negative controls in the absence of BPyneTEA or in the absence of UV illumination show no labeling. Left panel shows expression levels; right panel shows no receptor captured by streptavidin-beads, indicating that no photolabeling occurred. 4. Kinetic Analysis of Blockade Under Conditions of Clustered Single-Channel Activity Open-channel blockers can be classified as fast, intermediate, or slow blockers according to the magnitude of their unblocking rate constants (assuming a large association rate constant, typically 106 - 108 M-1 · s-1). For fast blockers, the unblocking rate constant is fast compared to the sampling frequency of the recording, so that the mean residence time in the blocked state is shorter than the time resolution of the measurement. As a result, when the channel opens, several blocking and unblocking events occur before a detectable channel closure occurs, leading to a net decrease in the observed open-channel current amplitude. BPyneTEA does not exhibit this phenotype. For intermediate blockers, the unblocking rate constant is of similar magnitude to channel opening and closing rate constants. As a result, when the channel opens, several blocking and unblocking events occur and are resolvable, causing a burst of shortened singlechannel openings within the overall cluster of openings. However, BPyneTEA does not cause resolvable bursts within clusters of activity and does not exhibit this phenotype. Instead, BPyneTEA acts as a slow blocker.

Scheme S3. Simple Blockade Model For Analyzing Early Termination of Clusters For slow blockers, the unblocking rate constant is slow compared to channel opening, resulting in long sojourns in the closed blocked state that are indistinguishable from desensitized sojourns. As a result, an additional path is available to terminate clustered activity, resulting in shortening of clusters. Because the probability of blockade is proportional to the time a channel spends open, longer open events are more likely to be blocked. This leads to a net decrease in the observed mean open time within a cluster, that is = ( + k+desensitize + k+blockopen · [blocker])-1 ~ ( + k+blockopen · [blocker])-1 since k+desensitize << .

S6

The same argument can be applied to blockade of the closed state within in a cluster, and closed-state blockade leads to a net decrease in the closed time: = (´ + k+blockclosed · [blocker])-1. In the case of closed-state blockade, a complication is introduced when dealing with clusters. Because clusters are defined by sojourns in the long-desensitized state, sojourns in a fast-desensitized or “gap” state, shorter-thanaverage sojourns in the long-desensitized state and shorter-than-average sojourns in the blocked state may contaminate the clusters. The closed time component which decreases in magnitude as blocker concentration increases is therefore the closed component of interest for determining blockade rate constants. Kinetic simulations illustrate how a blocker that binds to both the open and the closed states affects the open and closed time distributions. The following parameters were used for kinetic simulations: k+bc=5x106 M-1 s-1, k-bc=50 s-1, ==1000 s-1, k+bo=1x106 M-1 s-1, k-bo=50 s-1, k+desensitize=k-desensitize=100 s-1. The results of the simulation show the decrease in both closed and open dwell times with increasing blocker concentration (Supporting Figure 2) and the appearance of a long-lived closed component.

0.06

0.06

0.05

0.05

0.04

0.04

Count/total

0.03

No blocker

0.03

0.02

0.02

0.01

0.01

0.00 -1.0

-0.5

0.0

0.5

0.00 -1.0

-0.5

0.0

0.5

0.08

0.07 0.07

0.06 0.06

0.05

2 mM blocker

0.05

0.04

0.04

0.03

0.03

0.02

0.02 0.01

0.01

0.00 -1.0

-0.5

0.0

0.5

0.00 -1.0

-0.5

0.0

0.5

Duration (log10 ms)

Supporting Figure 2. Closed (left) and open (right) residence time distributions from kinetic simulations of blockade. Dotted lines mark the mean closed and open times in the absence of blocker. Blue arrow: short closed component. Black arrow: long closed component due to blockade events. The long-lived closed component corresponds to the time that the channel remains in blocked states (closed/blocked and open/blocked states may not be kinetically distinguishable). As mentioned above, slow blockade terminates clustered single-channel activity early, but shorter-than-average sojourns in blocked states may contaminate clusters. Therefore, the longest-lived closed component in clusters represents an underestimate of the actual blockade lifetime. Because the unblocking rate constant is proportional to the reciprocal of the mean blockade lifetime, the long-lived closed component provides an upper bound for the total unblocking rate constant, a function of

S7

80 40 0

A

100

200

300

400

BPyneTEA (µM)

500

0.0 -2

Closed Dwells Open Dwells 0 M 1

1.0

-1

0

1

2

3

1.0

0.0 -2

0.0 -2

-1

0

1

-1

0

1

-1

0

0.5

-1

0

1

2

3

1.0

0.0 -2

0.5

-1

0

1

2

3

0.0 -2

2

3

100  M 2

3

200  M

1.0

0.5 0.0 -2

0.5

1.0

0.5

1

2

3

1 1.0

1.0

0.5 0.0 -2

1

120

0.5

0.5

-1

0

1

2

3

0.0 -2

1.0

1.0

0.5

0.5

0.0 -2

-1

0

1

2

3

0.0 -2

-1

0

1

-1

0

1

-1

0

1

300  M 2

3

400  M 2

3

1

1

160

0

1.0

Normalized Frequency

Cluster Length (ms)

1/2

unblocking rate constants for both the open and closed states. Ideally, this component would increase in amplitude as blocker concentration increases; however, because of the presence of contaminating desensitized sojourns this is likely not to be the case in practice with a limited sample of clusters. Still, this complication does not preclude estimation of an upper bound, with the caveat that, by definition, the upper bound may overestimate the true unblocking rate constant, leading to an upper bound for the blocking dissociation constant that overestimates the true dissociation constant. Here, we estimate an lower bound for the residence time of BPyneTEA blockade as 10 ms (Supporting Figure 3B), and therefore the estimated upper bound for the BPyneTEA unblocking dissociation rate constant in 100 s-1. The upper bounds for BPyneTEA affinities to the closed and open nAChR reported in the main text were calculated as (100 s-1)/(5x106 M-1·s-1) = 20 µM and (100 s-1)/(1.3x106 M-1·s-1) = 80 µM, respectively.

1.0

1.0

0.5

0.5

0.0

-2

500  M

0.0

-1

0

1

2

3

-2

2

3

Dwell Time Log 10 (msec)

B

Supporting Figure 3. Effect of increasing concentrations of BPyneTEA on clustered single-channel nAChR G153S activity. (A) Clusters decrease in length with increasing BPyneTEA concentration. (B) Intracluster closed (left) and open (right) time distributions as a function of BPyneTEA concentration (values at right). Arrows point to kinetic components corresponding to unblocked closures (left) and blockade events (right).

5. Woodhull Analysis of the Voltage-Dependence of Open-Channel Blockade

Scheme S4. Simple Open-Channel Block Model For a simple open-channel blockade model, the mean open time is inversely proportionally to the sum of rate constants leaving the open state. We can write the mean S8

open time as = ( + k+block · [blocker])-1. It has been shown previously that the closing rate constant () is voltage-dependent.11 The mean open time increases with hyperpolarization (the transmembrane voltage used in the current experiments range from 0 to -200 mV) according to the relationship (V) = 0 · exp(z··F·V/R·T) where F is Faraday’s constant, R is the gas constant, T is the temperature, z is the gating charge, and  is the electrical distance or the percent of the transmembrane electric field sensed by the gating charge.

A 6

1200

3x10

-1

k+block (M s )

800 600 400 200 0 -0.2

B

6

2x10

-1

-1

 (s )

1000

-0.1

Voltage (mV)

6

1x10

0 -0.2

0.0

C

-0.1

0.0

Voltage (V)

Supporting Figure 4. The voltage dependence of BPyneTEA open-state blockade. (A) In the absence of BPyneTEA (left) decreasing voltage increases open times. In the presence of 500 µM BPyneTEA (right) decreasing voltage decreases open times because of open-state blockade. Woodhull analysis of the voltage dependence gives the inherent voltage dependence of the closing rate constant (B) 0 = 1150 ± 50 s-1,  = 0.25 ± 0.02 (mean ± standard error from nonlinear least squares fitting) and the voltage dependence of BPyneTEA blockade association rate constant (C) k+block0 = 5±1 x 105 M-1 s-1, block = 0.2 ± 0.1. The results demonstrate that BPyneTEA binds approximately 20% within the transmembrane electrical field. For open-channel blockade by a charged group, the voltage dependence of the blockade association rate constant can be similarly described using the Woodhull S9

model.12 In this analysis, we neglect permeation of the blocker as a rare event, meaning that increasing hyperpolarization increases blockade according to the relationship k+block(V) = k+block0 · exp(-zblock · block · F·V/RT). For BPyneTEA, zblock = +1. Hence, we can write = (0 · exp(z··F·V/R·T) + k+block0 · exp(-block · F·V/RT)·[blocker])-1. From experiments at 1 µM acetylcholine using the wild-type receptor, we have a range of as a function of concentration and voltage. Initial fitting of the 3-dimensional surface using non-linear least squares fitting in Matlab (The Mathworks, Natick, MA) did not lead to a stable solution. Therefore, we fitted as a function of [BPyneTEA] at each voltage, providing estimates of  and k+block as a function of voltage. The voltage dependence was then fitted to the appropriate equations from the Woodhull model, providing estimates of the electrical distance. 6. Live-Cell Photolabeling Ratios and Statistical Tests

Supporting Figure 5. Using 250 µM BPyneTEA, closed, non-desensitized nAChRs are photolabeled 2-fold more efficiently than desensitized nAChRs. Supporting Table 1. BPyneTEA State-Selective Photolabeling µM BPyneTEA Photolabeling Ratio 2-Sided t-test Closed/Desensitized p-value* (mean ± sd) 250 1.95 ± 0.37 0.047* 50 1.80 ± 0.26 0.034* 10 1.24 ± 0.19 0.16 * Null hypothesis of non-selective labeling with a photolabeling ration of 1. Significant at the  = 0.05 level. 7. Proposed Mechanism for BPyneTEA State-Selective Blockade Structural differences in the pore region may account for the differential labeling of resting and desensitized states. On the basis of electrophysiological13,14 and solvent accessibility15 studies, the constriction in the pore (or “gate”) that occludes ion access

S10

differs structurally for closed and desensitized states. The gate is deep within the pore in the closed state, but shallower and closer to the extracellular face of the membrane in the desensitized state. The results of the Woodhull analysis of the voltage-dependence of BPyneTEA open-channel blockade suggest that the dominant barrier to blockade is not far from the extracellular side of the channel. This observation provides a possible explanation for the similarity of blockade affinities in closed and open states and the lower affinity in the desensitized state compared to the closed state. If BPyneTEA binds within the nAChR transmembrane pore at a shallow site near the desensitization gate, then BPyneTEA binding to the non-desensitized states would be unobstructed by the closed-state gate, but constriction of the desensitization gate would disfavor binding to the desensitized state. The data do not rule out the possibility that the probe binds the open and/or closed channel outside the conduction pore. However, the hypothesis of shallow pore blockade provides a straightforward explanation for the observation that BPyneTEA rapidly binds both closed and open states of the channel. Further structural and functional studies will be required to test this hypothesis and determine which specific residues are labeled by BPyneTEA.

Supporting Figure 6. Proposed structural mechanism for state-selective blockade of the nAChR by BPyneTEA. BPyneTEA binds the transmembrane pore far from the resting gate (green bars), and binding is not hindered by changes in the resting gate conformations. Thus, BPyneTEA blocks the closed and open (i.e. non-desensitized) states similarly. However, BPyneTEA binds the transmembrane pore near the desensitization gate, which is closer to the extracellular opening. When the nAChR is desensitized, the constriction of the desensitization gate inhibits BPyneTEA binding. Supporting References (1) Salamone, F. N.; Zhou, M.; Auerbach, A. J. Physiol. 1999, 516, 315-330. (2) Hamill, O. P.; Marty, A.; Neher, E.; Sakmann, B.; Sigworth, F. J. Pflug. Archiv 1981, 391, 85-100. (3) Sine, S. M.; Ohno, K.; Bouzat, C.; Auerbach, A.; Milone, M.; Pruitt, J. N.; Engel, A. G. Neuron 1995, 15, 229-239. (4) Qin, F.; Auerbach, A.; Sachs, F. Biophys. J. 1996, 70, 264-280. (5) Qin, F.; Auerbach, A.; Sachs, F. Biophys. J. 1996, 70, Mp432-Mp432. S11

(6) Qin, F.; Auerbach, A.; Sachs, F. Proc. R. Soc. Lond. Ser. B-Biol. Sci. 1997, 264, 375-383. (7) Salisbury, C. M.; Cravatt, B. F. Proc. Natl. Acad. Sci. U.S.A. 2007, 104, 1171-1176. (8) Keller, S. H.; Kreienkamp, H. J.; Kawanishi, C.; Taylor, P. J. Biol. Chem. 1995, 270, 4165-4171. (9) Gallagher, S. R. In Short Protocols in Molecular Biology; Ausubel, F. M., et al., Ed.; John Wiley and Sons: New York, 1999, p 10-10. (10) Speers, A. E.; Cravatt, B. F. Chem. Biol. 2004, 11, 535-546. (11) Auerbach, A.; Sigurdson, W.; Chen, J.; Akk, G. J. Physiol. 1996, 494, 155-170. (12) Woodhull, A. M. J. Gen. Physiol. 1973, 61, 687-708. (13) Purohit, Y.; Grosman, C. J. Gen. Physiol. 2006, 127, 703-717. (14) Leonard, R. J.; Labarca, C. G.; Charnet, P.; Davidson, N.; Lester, H. A. Science 1988, 242, 1578-1581. (15) Wilson, G. G.; Karlin, A. Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 12411248.

S12