|Home | About | Journals | Submit | Contact Us | Français|
A HEK-293 cell line that stably expresses mouse 5-HT3ARs containing a C-terminal extension that confers high affinity binding of alpha-bungarotoxin (αBgTx) was established (αBgTx-5-HT3ARs) and used to purify αBgTx-5-HT3ARs in a lipid environment for use in structural studies using photoaffinity labeling. αBgTx-5-HT3ARs were expressed robustly (60 pmol [3H]BRL-43694 binding sites (~3 µg receptor) per mg protein and displayed the same functional properties as wild-type receptors (serotonin EC50 = 5.3) +/− 0.04 µM). While [125I]αBgTx bound to the αBgTx-5-HT3ARs with high affinity (Kd = 11 nM), application of nonradioactive αBgTx (up to 300 µM) had no effect on serotonin-induced current responses. αBgTx-5-HT3ARs were purified on an αBgTx-derivatized affinity column from detergent extracts in mg quantities and at ~25% purity. The hydrophobic photolabel 3-trifluoromethyl-3-(m-[125I]iodophenyl) diazirine ([125I]TID) was used to identify the amino acids at the lipid-protein interface of purified and lipid-reconstituted αBgTx-5-HT3ARs. [125I]TID photoincorporation into the αBgTx-5-HT3AR subunit was initially mapped to subunit proteolytic fragments of 8 kDa, containing the M4 transmembrane segment and ~60% of incorporated 125I, and 17 kDa, containing the M1–M3 transmembrane segments. Within the M4 segment, [125I]TID labeled Ser451, equivalent to the [125I]TID-labeled residue Thr422 at the lipid-exposed face of the Torpedo nicotinic acetylcholine receptor (nAChR) α1M4 α-helix. These results provide a first definition of the surface of the 5-HT3AR M4 helix that is exposed to lipid and establish that this surface is equivalent to the surface exposed to lipid in the Torpedo nAChR.
The 5-HT3 1 receptor is a member of the Cys-loop superfamily of ligand-gated ion channels (LGICs) and mediates excitatory fast synaptic transmission in the central and peripheral nervous systems (1). Presynaptic localized 5-HT3 receptors modulate the release of several neurotransmitters, including acetylcholine, dopamine, and γ-aminobutyric acid (2, 3) . In common with all members of the Cys-loop family of LGICs, 5-HT3 receptors are assembled as a pentamer of subunits that surround, in a pseudosymetric manner, a cation-selective ion channel (4–6). Of the five subunits (A–E) cloned to date, only the 5-HT3A and 5-HT3B subunits have been demonstrated to have functional significance in the central or peripheral nervous system (1, 7–9). The 5-HT3A subunit is able to assemble as a homomeric pentamer and is the predominant form of the 5-HT3 receptors present in the rodent brain (10). In contrast, the other 5-HT3R subunits (B–E) must coassemble with the 5-HT3A subunit to form functional receptors (9). The 5-HT3A subunit, along with all Cys-loop LGIC subunits, share a common structural architecture that is comprised of a large extracellular N-terminal domain, four transmembrane domains/α-helices (M1–M4), connected by extracellular (M2–M3) and intracellular (M1–M2 and M3–M4) loops, and an extracellular C-terminus (11–12). The five M2 helices are arranged about a central axis orthogonal to the membrane forming the channel lumen, and the M1, M3, and M4 helices form an outer ring that shields M2 from the lipid bilayer.
Due to the abundance and purity of muscle-type nicotinic acetylcholine receptors (nAChRs) in preparations from the electric organ of the Torpedo electric ray, it is the most studied and understood member of the Cys-loop LGIC superfamily, and many of the structural/functional features of the Torpedo nAChR may be generalized to other member of the Cys-loop superfamily. The 5-HT3A subunit shares about 22–30% sequence identity/homology with Torpedo AChR subunits (1, 13) and a variety of studies utilizing different methodologies, including the construction of functional α7AChR/5-HT3AR chimeras, have contributed to the consensus view that the Torpedo AChR, the 5-HT3AR and other Cys-loop LGIC members display a common three-dimensional architecture (11–14). Nevertheless, direct structural information regarding the 5-HT3AR is currently lacking. This is in large part due to the low expression level of 5-HT3ARs (<0.3 µg receptor/mg of tissue) in the central and peripheral nervous system as well in most established mammalian cell lines (14–17).
The establishment of a rich-source of 5-HT3AR protein and a means of purifying the receptor to or near homogeneity are two important steps in the effort to directly study the structure of the 5-HT3AR. We report here: 1) the establishment of an HEK-293 cell line (αBgTx-5-HT3AR) that stably and robustly (3–8 µg receptor/mg of membrane protein) expresses mouse 5-HT3ARs containing a C-terminal extension (pharmatope tag) that confers high affinity binding of the AChR competitive antagonist alpha bungarotoxin (αBgTx; 18–19) that does not bind to native mouse 5-HT3AR; and 2) purification on an αBgTx-derivatized affinity column of αBgTx-5-HT3ARs in mg quantities and at ~25% purity, reconstituted into lipid vesicles. We used these purified receptors to identify the amino acids photolabeled by 3-trifluoromethyl-3-(m-[125I] iodophenyl) diazirine ([125I]TID, the hydrophobic, photoreactive probe that has been used to identify amino acids at the lipid interface of Torpedo and α4β2 nAChRs (20–22).
BRL-43694[9-methyl-3H] ([3H]BRL-43694; 85 Ci/mmol) and [125I]Tyr54-α-Bungarotoxin ([125I]αBgTx; 102 Ci/mmol) were obtained from Perkin Elmer Life Sciences, Inc. (Boston, MA). 5-Hydroxy [3H] tryptamine trifluoroacteate ([3H]5-HT; 113 Ci/mmol) and 3-trifluoromethyl-3-(m-[125I]iodophenyl) diazirine ([125I]TID; ~10 Ci/mmol) were obtained from GE Biosciences (Piscataway, NJ). MDL7222 and serotonin (5-HT) were from Sigma-Aldrich (St. Louis, MO), trypsin (TPCK-treated) from Worthington; Staphylococcus aureus glutamyl endopeptidase (V8 protease) from MP Biochemicals (Solon, OH), trifluoroacetic acid (TFA) from Thermo Fisher Scientific Inc (Rockford, IL); sodium cholate and CHAPS from USB Corporation (Cleveland, OH), protease inhibitor cocktail set III and Genapol C-100 from Calbiochem. Prestained low range molecular weight standards were from Bio-Rad Laboratories (Hercules, CA). Dulbecco’s Modified Eagle’s Medium/Ham’s F-12 50/50 mix (DMEM/Ham’s F-12) and minimum essential media (MEM) were obtained from Mediatech, Inc (Herndon, VA), synthetic lipids from Avanti Polar lipids, Inc. (Alabaster, AL), non radioactive αBgTx from Biotium, Inc. (Hayward, CA). Reversed-phase HPLC columns (Brownlee Aquapore C4 column, 100 × 2.1 mm, BU-300) were obtained from Perkin Elmer Life Sciences Inc. (Boston, MA) and Centriprep-10 concentrators from Amicon Inc. (Beverly, MA).
Mouse 5-HT3A receptor cDNAs, provided by Dr. David Julius (UCSF, San Francisco), were subcloned into pCDNA3.1. The α-bungarotoxin binding (pharmatope) sequence was added to the carboxyl terminal of the mouse 5-HT3A by primer addition PCR using the QuickChange system (Stratagene, USA). A stop codon and an αBgTx pharmatope sequence were first added using a first set of primers (Supplementary Figure S1). Subsequently, another αBgTx pharmatope sequence and a 10 amino acid glycine-asparagine repeat linker sequence were added with a second set of primers (Supplementary Figure S1).
HEK-293 cells were seeded at a density of 105 cells per dish and transfected in 35 mm culture dishes with the αBgTx-5-HT3AR construct using polyethyleneimine at a nitrogen to phosphate ratio of five. After 48 hours, cells were treated with 800 µg/mL Geneticin (G-418) in cell culture medium [Dulbecco’s modified eagle’s media (DMEM) with 10% fetal bovine serum] for an additional 2 days (maintained in a humidified incubator at 37°C in 5% CO2). Cells were then dissociated, counted using Trypan Blue exclusion and plated onto 96-well trays at a dilution of 0.5 viable cells per well (~ 1 cell in every other well). Individual wells were visually screened for the presence of single cells. Clones from single cells were expanded at confluence for microscopic evaluation and semi-quantitative characterization by Texas Red-αBgTx binding. Cells were incubated for 10 min with the αBgTx conjugate (10 µg/mL in Dulbecco’s phosphate buffered saline (DPBS), 1% bovine serum albumin) and washed twice with DPBS to remove excess label. Cells were imaged using a Zeiss epifluorescence microscope and spectrofluorometrically in lysates (0.1% Triton in PBS) after binding. The clone displaying the highest level of stable expression was selected and expanded for further experiments.
The whole-cell configuration of the patch-clamp technique was used to record currents mediated by 5-HT3A ion channels from voltage-clamped cells (23). Cells were held at −60mV and continuously washed by perfusion with extracellular solution containing (in mM): NaCl, 150; KCl, 2.5; MgCl2,2.5; CaCl2, 2.5; glucose, 10; and HEPES, 10 (pH adjusted to 7.4 with NaOH, and osmolarity adjusted to ~350 mosm using sucrose). Agonist was locally applied by rapid superfusion exchange using an SF-77B Warner Instuments fast step perfusion system (Warner Instruments, Hamden, CT). Pipettes were pulled from thin wall borosilicate glass tubing using a Flaming-Brown P-97 two stage puller (Sutter Instruments, Novato, CA). The electrodes were filled with intracellular solution containing (in mM) CsCl, 150; BAPTA 0.2; Mg Cl2, 1; Mg ATP, 3; HEPES, 10, GTP 0.3 (pH adjusted to 7.2 with CsOH, and osmolarity adjusted to ~320 mosm using sucrose). Currents were monitored with an Axopatch 200B (Axon Instruments, Foster City, CA), low pass filtered at 1 kHz, digitized at 12.5 kHz using a Digidata 1320A interface, and stored using pCLAMP8 software. Experiments were performed at room temperature 20–24°C. Data were analyzed using cursor-based measurement with Clampfit 8.2 software to measure peak current amplitudes relative to the pre-drug baseline current. Concentration-response curves were plotted and data were fit using a variable-slope sigmoidal dose response curve in Prism v4.03 (GraphPad Software, La Jolla, CA).
HEK-αBgTx-5-HT3AR cells were grown in 140 mm culture dishes at 37°C in a humidified incubator at 5 % CO2 (~150 dishes per week; Greiner Bio-One, Germany) and maintained in a medium consisting of a 1:1 mixture of DMEM and Ham’s F12 medium, supplemented with 10% fetal bovine serum, 100 units/mL penicillin G/streptomycin, and 250 µg/mL G-418 as a selection agent. Typically, cells were treated with 100 µM 5-HT 24 h prior to harvesting to enhance receptor expression (Supplementary Figure S2). At ~80% confluence, the cells were harvested by gentle scraping and washed with vesicle dialysis buffer (VDB: 100mM NaCl, 0.1mM EDTA, 0.02% NaN3, 10mM MOPS, pH 7.5) containing protease inhibitor cocktail III (0.1 µL per mL), and then the final cell pellets were stored at −80°C.
For membrane isolation, HEK-αBgTx-5-HT3AR cell pellets (from 150 dishes) were thawed and homogenized (glass potter) in 100 mL VDB in the presence of protease inhibitor cocktail set III (0.1 µL/mL). The membranes were pelleted by centrifugation (39,000g for 1 h at 4°C), then resuspended in 35 mL VDB with protease inhibitor cocktail III (0.1 µL per mL). The protein concentration was determined by Lowry protein assay (24) and stored at −80°C. Membrane from approximately 700 dishes were collected over a 4–6 week period.
Equilibrium binding of [3H]BRL-43694, [3H]5-HT and [125I]αBgTx to HEK-αBgTx-5-HT3AR cell membranes was measured using a centrifugation assay. Membranes at 0.166 mg/mL in VDB (final volume 150 µL) were incubated for 1.5 h at RT with increasing concentrations of [3H]BRL-43694 (0.8–40 or 1.6–83 nM), [3H]5-HT (2.4–120 nM) or [125I]αBgTx (2–90 nM). Bound radioligand was separated from free by centrifugation (39,000g for 1 h) and quantified by liquid scintillation counting in a Packard 1900 TR counter ([3H]BRL-43694 and [3H]5-HT) or by γ counting in a Packard Cobra II counter ([125I]αBgTx). Non-specific binding was determined in the presence of 50 µM MDL7222 ([3H]BRL-43694 and [3H]5-HT) or 5 µM αBgTx ([125I]αBgTx). Total, nonspecific, and specific cpm were converted to pmol bound radioligand / mg of protein, and free cpm was converted to nM radioligand. Curve fitting and parameter estimations were performed using Graphpad Prism v5.0 software.
αBgTx-5-HT3ARs were affinity-purified using an αBgTx-derivatized Sepharose 4B column originally developed for purification of the Torpedo nAChR from electric organ tissue (25). Briefly, 2 g of CNBR-activated Sepharose 4B (GE Biosciences) was washed with 1 mM HCL (> 400 mL) then amino coupled to 10 mg (1.25 µMoles) of αBgTx in 40 mL of 0.1 M NaHCO3, 0.5 M NaCL, pH 8.3. Coupling was monitored by A280nm and allowed to proceed to completion (>95%, ~ 15 h at 4°C), then excess amino-reactive sites were blocked by ethanolamine (1 g). The affinity column was equilibrated with ~15 column volumes of 0.2 mg/mL asolectin, (a crude soybean lipid extract) in 1% cholate in VDB. Approximately 2–3 g of HEK-αBgTx-5-HT3AR membranes (equivalent to ~700 dishes; ~50–80 nmol 5-HT3A receptor) were solubilized by adding an equal volume of 2% CHAPS (final concentration 2 mg/mL protein; 1% CHAPS), and stirred for 5 h at 4°C. Insoluble materials were pelleted by centrifugation (91,500g for 1 h) and supernatant was dialyzed against 8 volumes of 1% cholate in VDB for 5 h at 4°C then slowly applied to the affinity column (0.7 mL/min, ~24 h, at 4°C). The column was then extensively washed with 0.2 mg/mL asolectin in 1% cholate in VDB (15 column volumes; > 15 h) to remove non-receptor proteins. αBgTx-5-HT3ARs were then eluted from the column by either overnight incubation with 25 mL of 0.2 mg/mL asolectin in 1% cholate in VDB containing either 50 µM αBgTx or 0.5 M sodium chloride (NaCl) and collection of the column eluent or by slow application of αBgTx/NaCl (0.2 mL/min) and collection of 2.4 mL fractions (protein content assessed by A280 nm). To remove detergent and thereby reconstitute αBgTx-5-HT3ARs into membranes containing the defined lipid mixture, pooled eluent was dialyzed against 2 L of VDB (4 d, with buffer changes every 24 h) and purified receptors stored at −80°C. The purity of αBgTx-5-HT3ARs was assessed by SDS-PAGE, by Western analysis (5-HT3AR antibody: GTX13897 from GeneTex, Inc., San Antonio, TX), by protein sequencing and by photoaffinity labeling with [3H]5-HT.
For photolabeling with [3H]5-HT, 100 µg aliquots of αBgTx-5-HT3AR membranes were incubated with 100 nM [3H]5-HT in the absence or presence of 10 µM MDL7222, photolyzed at 312 nm (FisherBiotech FBUVM-80) for 7 min at a distance of < 1cm, pelleted by centrifugation (39,000g for 1 h at 4°C), resolved by SDS-PAGE and processed for fluorography (26). For analytical labelings with [125I]TID, 100 µg of affinity-purified, lipid-reincorporated αBgTx-5-HT3ARs (in 1 mL VDB) were incubated with ~ 2 µM [125I]TID (~10 Ci/mmol; GE Biosciences) in the absence or presence of 100 µM 5-HT for 1 h at room temperature and under reduced light conditions. For preparative labelings, ~2 mg of affinity-purified αBgTx-5-HT3ARs (0.5 mg/mL) were incubated with 12 µM [125I]TID (~600 µCi) under the same conditions. The samples were irradiated with a 365 nm hand-held UV lamp (Spectroline EN-280L) for 15 minutes at a distance of less than 1 cm and then pelleted by centrifugation (39,000g for 1 h at 4°C). Pellets were resuspended in electrophoresis sample buffer (12.5 mM Tris-HCl, 2% SDS, 8% sucrose, 1% glycerol, 0.01% bromophenol blue, pH 6.8) and the polypeptides were resolved by SDS-PAGE.
SDS-PAGE was performed according to (27) with separating gel comprised of 8% polyacrylamide/0.33 bisacrylamide. Following electrophoresis, gels were stained for 1 h with Coomassie Blue R-250 (0.25% (w/v) in 45% methanol, 10% acetic acid, 45% H2O), and destained (25% methanol, 10% acetic acid, 65% H2O) to visualize bands. Gels were then dried and exposed to Kodak X-OMAT LS film with an intensifying screen at −80°C (1–2 d exposure). After autoradiography, the band corresponding to the [125I]TID-labeled 5-HT3A subunit was excised from each condition (−/+ 5-HT), soaked in overlay buffer (5% sucrose, 125 mM Tris-HCl, 0.1% SDS, pH 6.8) for 30 min and transferred to the wells of a 15% acrylamide “mapping” gel (28). Each gel slice was overlaid with 10 µg (analytical labeling) or 100 µg (preparative labeling) of S. aureus V8 protease in overlay buffer. After electrophoresis, the gels were stained for 1 h with Coomassie Blue R-250, destained, and either prepared for autoradiography (analytical labeling) or soaked in distilled water overnight (preparative labeling). The 125I-containing bands were excised from the preparative gels, and the peptides were retrieved by passive diffusion into 25 mL of elution buffer (0.1M NH4HCO3, 0.1% (w/v) SDS, 1% β-mercaptoethanol, pH 7.8) for 4 d at room temperature with gentle mixing. The eluates were filtered to remove gel pieces and then concentrated using Centriprep-10 concentrators (10 kDa cutoff, Amicon, final volume < 150 µL). Samples were then either directly purified by reversed-phase HPLC or acetone precipitated (>85% acetone at −20°C overnight) to remove excess SDS and then subjected to further proteolytic digestion.
For digestion with trypsin, acetone-precipitated subunit fragments were suspended in 60 µL 0.1M NH4HCO3, 0.1% SDS, pH 7.8, and then the SDS content was diluted by addition of 225 µL 0.1M NH4HCO3 and 35 µL Genapol C-100 (final concentrations: 0.02% (w/v) SDS, 0.5% Genapol C-100, pH 7.8). Trypsin was added at a 200% (w/w) enzyme to substrate ratio and the digestion allowed to proceed for 4–5 d at room temperature.
HPLC was performed on a Shimadzu LC-10A binary HPLC system, using a Brownlee Aquapore C4 column (100 × 2.1mm). Solvent A was comprised of 0.08% trifluoroacetic acid (TFA) in water and Solvent B was comprised of 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, shown as dotted line in the figures) and fractions were collected every 2.5 min (42 fractions/run). The elution of peptides was monitored by the absorbance at 210 nm and the amount of 125I associated with each fraction was determined by γ-counting (5 minute/fraction) in a Packard Cobra II γ-counter.
Amino terminal sequence analysis of 5-HT3AR subunit fragments (e.g. V8-17 and V8-8) was performed on a Beckman Instruments (Porton) 20/20 automated protein sequencer using gas phase cycles (Texas Tech Biotechnology Core Facility). Pooled HPLC fractions were dried by vacuum centrifugation, resuspended in a small volume (20 µL) of 0.1% SDS and immobilized on chemically modified glass fiber disks (Beckman Instruments). Peptides were subjected to at least 10 sequencing cycles. Alternatively, [125I]TID labeled peptides were sequenced on an Applied Biosystems PROCISE 492 protein sequencer configured to utilize 1/6 of each cycle of Edman degradation for amino acid identification/quantification and collect the other 5/6 for 125I counting. The pooled HPLC fractions were diluted 3-fold with 0.1% TFA in distilled water (to reduce organic concentration) and loaded onto PVDF disks using Prosorb sample preparation cartridges (Applied Biosystems No. 401950). Before sequencing, filters were processed as recommended by the manufacturer. To determine the amount of the sequenced peptide, the pmol of each amino acid in a detected sequence was quantified by peak height and fit to the equation f(x) = I0Rx, where I0 was the initial amount of the peptide sequenced (in pmol), R was the repetitive yield, and f(x) was the pmol detected in cycle x. Ser, His, Trp, and Cys were not included in the fits due to known problems with their accurate detection/quantification. The fit was calculated in SigmaPlot 11 (SPSS) using a non-linear least-squares method and Figures containing 125I release profiles include this fit as a dotted line. Quantification of 125I incorporated into a specific residue was calculated by (cpmx−cpm(x−1))/5IoRx.
Two important prerequisites for direct structural (e.g. hydrophobic photoaffinity labeling) studies of the 5-HT3AR are: 1) a rich source of receptor protein and 2) a means of further purifying the receptor to or near homogeneity. In the absence of a natural rich source of 5-HT3ARs, HEK-293 cells that stably or transiently express the 5-HT3AR have been used as an in vitro system to study structural and functional aspects of the 5-HT3AR (29–31). The addition of a C-terminal αBgTx-binding sequence (i.e. αBgTx-pharmatope tag; 18–19) to the 5-HT3AR, provides a means of isolating αBgTx-5-HT3ARs using an αBgTx-derivatized affinity column. Here we have constructed a HEK-293 cell line that stably expresses the mouse 5-HT3AR with a C-terminal αBgTx-pharmatope tag. Receptors with a single αBgTx pharmatope tag added did not show evidence of pharmatope labeling when intact cells were exposed to fluorescently-derivatized αBgTx. To extend the binding site further from the C-terminus, and prevent steric hindrance of binding, we added a 10 amino acid glycine-asparagine repeat linker and a second αBgTx pharmatope tag prior to the first tag. This construct (Supplementary Figure S1) was incorporated into stably expressing HEK 293 cells, as described in experimental procedures, and used in the remainder of the experiments.
The functional properties of αBgTx-5-HT3ARs were examined by whole-cell patch clamp electrophysiological recordings. Application of 30 µM serotonin (5-HT) produced inward currents at a −60 mV holding potential that rapidly decayed in the continued presence of agonist (Figure 1A). Cells stably expressing αBgTx-5-HT3ARs exhibited average current density of 393.1 +/− 40.2 pA/pF when activated by a maximally-effective concentration of agonist (30 µM). Concentration-response data revealed that agonist potency and Hill slopes were similar for the WT and αBgTx-5-HT3ARs receptors (EC50: WT = 5.4 +/− 0.05 µM, αBgTx-5-HT3AR = 5.3 +/− 0.04 µM; Hill slope: WT = 1.64, αBgTx-5-HT3AR = 1.75) (Figure 1B). Application of 10–300 µM unlabeled αBgTx did not produce a consistent concentration-dependent effect on current mediated by αBgTx-5-HT3ARs (Figure 1C).
The expression level of αBgTx-5-HT3ARs in HEK-293 cell membranes was assessed using the equilibrium binding of the 5-HT3R competitive antagonist [3H]BRL-43694 (Figure 2), the agonist [3H]5-HT (Supplementary Figure S2) and [125I]αBgTx (Figure 2) which binds to the C-terminal αBgTx-pharmatope (18,19). For a given membrane preparation, [3H]BRL-43694 and [125I]αBgTx binding yielded identical Bmax values (60 pmol of binding sites/mg protein) and Kd values of 1.93 ± 0.15 and 11.0 ± 1.1 nM respectively. If αBgTx binds to one pharmatope site per 5-HT3AR subunit and to five sites per pentamer, then the expression level of receptor is ~12 pmol (3.4 µg)/mg membrane protein (each 150 mm culture dish contains ~8 µg receptor). The Kd value for [3H]BRL-43694 (1.9) nM) as well as that for [3H]5-HT (28 ± 2.6 nM; Supplementary Figure S2) are consistent with the reported equilibrium binding affinities for these ligands (5, 32).
In the course of optimizing the conditions for culturing HEK-αBgTx-5-HT3AR cells, we tested whether or not serotonin treatment would result in increased receptor expression such as that observed with nicotine treatment for α4β2 nicotinic acetylcholine receptors expressed in HEK-293 cells (22). Treatment of HEK-αBgTx-5-HT3AR cells with 100 µM serotonin 24 h prior to cell harvesting enhanced the level of αBgTx-5-HT3AR expression by 2.83 ± 0.28-fold (N = 4; Supplementary Figure S3). Additional studies are in progress to determine if this ‘serotonin-induced receptor upregulation’ is truly analogous to the nicotine-induced upregulation of neuronal nAChRs observed both in cell culture as well as in the brains of smokers (33–34, 22).
For receptor solubilization we tested detergents that can be readily removed by dialysis (CHAPS and sodium cholate) following receptor purification, allowing reconstitution of purified receptors into lipid. [3H]BRL-43694 binding affinity was the same for αBgTx-5-HT3ARs in membranes and in lipid vesicles after solubilization in either 1% CHAPS or cholate (data not shown), but CHAPS gave a slightly higher level of solubulization (50–65% of the total protein). Therefore αBgTx-5-HT3ARs were solubilized in 1% CHAPS and then purified as described under Experimental Procedures on an αBgTx-affinity column similar to that used previously to purify AChRs (25,35). Figure 3A shows a typical column elution profile in which receptors were eluted with 0.5 M NaCl, but similar column yields were seen for receptors eluted with 50 µM αBgTx. Peak protein fractions were pooled and dialyzed against VDB to remove detergent and reconstitute 5-HT3ARs into membrane vesicles.
When an aliquot (~100 µg) of affinity-purified αBgTx-5-HT3ARs was analyzed by SDS-PAGE, multiple Coomassie-Blue stained bands were visible (Figure 3, lane 1). However the αBgTx-5-HT3A subunit band was identified as a prominent Coomassie-Blue stained band migrating with an apparent molecular mass of ~ 65 kDa by immunoblot (Figure 3, lane 2), by [3H]5-HT photoaffinity labeling (Figure 3, lanes 3–4), and by protein sequencing (Figure 5–Figure 6). On the basis of a densiometric scan of the Coomassie Blue-stained gel, we estimated that the αBgTx-5-HT3AR preparation was ~25% pure. For a typical column run, starting with ~700 dishes (~2.5 g HEK-αBgTx-5-HT3AR membrane protein containing ~75 nmol/ 21 mg receptor), our yield following detergent solubilization and affinity purification was ~10% (7 mg total protein at 25% purity). Although the αBgTx-5-HT3ARs was not purified to homogeneity by this protocol, the αBgTx-affinity column purifications were reproducible (3 runs) and yielded αBgTx-5-HT3ARs in sufficient quantity (~2–7 mg) and purity (20–27%) to enable structural studies (e.g. hydrophobic photolabeling).
The hydrophobic photoreactive probe [125I]TID was employed to map the lipid-exposed domains and to determine amino acid residues in the 5-HT3A subunit that are in contact with membrane lipid. [125I]TID is a small hydrophobic photoreactive compound that partitions efficiently (> 95%) into the lipid bilayer and upon exposure to UV light (365 nm), [125I]TID is activated and covalently tags protein regions that are exposed to the lipid bilayer (21,22). When purified αBgTx-5-HT3ARs (100 µg) were photolabeled with [125I]TID in the absence or presence of 100 µM 5-HT and the polypeptides were resolved by SDS-PAGE, an autoradiogram of the dried gel revealed [125I]TID photoincorporation into the 5-HT3AR subunit and in other unidentified polypeptides also visible by Coomassie stain (Figure 4A). The extent and pattern of labeling was unaffected by addition of 100 µM 5-HT (Figure 4A, + lane).
To further map the sites of [125I]TID labeling within the 5-HT3AR subunit, the labeled 5-HT3AR subunit band was excised from the stained 8% gel and subjected to “in-gel-digestion” (Cleveland gel) with S. aureus V8 protease as described in the Experimental Procedures. Based upon the autoradiograph of the mapping gel (Figure 4B) [125I]TID was photoincorporated into αBgTx-5-HT3AR subunit proteolytic fragments with apparent molecular masses of 17 KDa (V8-17K) and 8 KDa (V8-8K; Figure 4B). Addition of agonist (5-HT) had no significant effect (<5%) on the extent of [125I]TID photolabeling of either proteolytic fragment (Figure 4B, + lane). Based on γ-counting of excised gel bands, approximately 60% of the total labeling was localized in the V8-8K fragment and 40% into V8-17K.
To identify the V8-8K and V8-17K fragments, the labeled bands were excised from the mapping gel, the 125I-labeled peptides in those bands were further purified by rpHPLC for N-terminal sequence analysis. For V8-17K, the 125I eluted in a single hydrophobic peak (Figure 5A) which contained as the primary sequence a peptide beginning at Val195 of the mouse 5-HT3A subunit (Table 1; Figure 5C and Supplementary Figure S1B). On the basis of the apparent molecular mass (17 kDa) and the likely cleavage sites of V8 protease, the V8-17K fragment is predicted to include the M1 (~Pro223-Pro249), M2 (Val255-Asp274) and M3 (Tyr288-His311) and terminate at Glu339. For V8-8K, sequence analysis of the hydrophobic peak of 125I from the HPLC fractionation (Figure 5B), revealed the presence of a primary sequence beginning at Val424 of the mouse 5-HT3A subunit (Table 1; Figure 5C and Supplementary Figure S1B). On the basis of the apparent molecular mass (8 kDa), the V8-8 fragment is predicted to include M4 (~Leu438-Leu457) and terminate at the C-terminus of the αBgTx-5-HT3A subunit (Supplementary Figure S1B).
To identify individual amino acid residue(s) labeled by [125I]TID within the M4 segment of the 5HT3A subunit, [125I]TID-labeled V8-8K, with an amino terminus (Val424) just 14 amino acids before the beginning of M4 (~Leu438-Leu457), was isolated from a preparative scale labeling (2 mg affinity-purified 5-HT3AR) and further digested with trypsin, and the digest was fractionated by reversed-phase HPLC. Sequence analysis of the hydrophobic peak of 125I (Figure 6) revealed two overlapping 5-HT3A subunit peptides present in approximately equal abundance and beginning at Val424 (5.6 pmol) and Val431 (5.1 pmol,). There were peaks of 125I release in cycles 21 (36 cpm) and 28 (19 cpm) that are consistent with [125I]TID labeling of Ser451 within the M4 segment, with the 125I release in cycle 21 resulting from the peptide starting at Val431 and that in cycle 28 from the peptide starting at Val424. The efficiencies of [125I]TID photoincorporation into Ser451 calculated from the 125I released in cycle 21 and the mass of peptide beginning at Val431 or from the 125I released in cycles 28 and the mass of peptide beginning at Val424 are nearly identical (2.6 and 2.1 cpm/pmol respectively).
We were unable to identify any photolabeled amino acids within the 17 kDa subunit fragment (V8-17K, Val195-Glu-339 and containing M1–M3 transmembrane segments). As for V8-8K, a hydrophobic peak of 125I was recovered from a tryptic digest of V8-17K resolved by rpHPLC, but sequence analysis of that material, which contained similar amounts of 125I as in the sample sequenced from the tryptic digest of V8-8K, revealed no detectable peptide at >1 pmol and no peak of 125I release >4 cpm in 20 cycles of Edman degradation.
The important role that the 5-HT3AR plays in excitatory fast synaptic transmission in the central and peripheral nervous system, its role as a modulator of neurotransmitter release, and its association with numerous neuropsychiatric disorders and conditions including radio-and chemotherapy-induced nausea and vomiting, make the 5-HT3AR a particularly attractive therapeutic target (36–37). A more refined understanding of the molecular structure of the 5-HT3AR will likely contribute to the development of new therapeutic agents as well to our understanding of the structure/function correlates of this receptor and LGICs in general. To this end, we developed a mammalian expression system that produces mouse 5-HT3ARs containing a C-terminal αBgTx-pharmatope tag (αBgTx-5-HT3ARs) that is functionally indistinguishable from wild-type 5-HT3ARs. The robust expression of αBgTx-5-HT3ARs (~60 pmol [3H]BRL-43694 binding sites (3.4 µg receptor)/mg protein; ~ 8 µg receptor per 150 mm dish) and its high affinity binding for αBgTx (~11 nM; Figure 2) enabled us to reproducibly (N=3) purify large quantities (2–7 mg) of 5-HT3ARs in a single step purification on an αBgTx-derivatized affinity column. This yield exceeded those reported for purification of 5-HT3ARs from NCB20 cells (yield = 13 µg; 38), from porcine brain (yield = 2–5 µg; 39), or from virally transfected Sf9 insect cells (yield = 0.2 mg; 5) or mammalian cells (yield, 0.2 mg, 40) and was comparable to that reported (15 mg) using Semliki Forest Virus transfected BHK cells grown in suspension in a bioreactor (41).
Although an additional purification step is expected to further increase the purity of the 5-HT3AR, in this report we established the usefulness of this purified, lipid-reincorporated receptor preparation for structural analyses. For that, we examined the structure of the 5-HT3AR lipid-protein interface by use of [125I]TID, a hydrophobic, photoreactive compound that has been used extensively to study the structure of Torpedo and α4β2 nAChRs (42–44, 21–22). In Torpedo nAChR, [125I]TID photolabeled amino acids at the lipid-protein interface, within the ion channel, and in the delta subunit helix bundle. Though, agonist-induced transition from the resting to the desensitized state has no effect on [125I]TID photolabeling at nAChR lipid-protein interface (21), it inhibits photolabeling within the ion channel (42) and enhances photolabeling within the δ subunit helix bundle (43–44).
When affinity-purified and lipid-reconstituted αBgTx-5-HT3ARs were labeled with [125I]TID, photolabeling was limited to two subunit fragments: V8-17K, which contains transmembrane segments M1–M2–M3, and V8-8K, which contains the M4 segment. As seen for [125I]TID photolabeling of the affinity-purified α4β2 nAChR, but not for the affinity-purified Torpedo nAChR, addition of agonist did not alter the extent or the pattern of [125I]TID labeling at the level of the 5-HT3AR subunit or subunit fragments. This suggests that the affinity-purified 5-HT3ARs are stabilized in a desensitized state (i.e. unable to undergo agonist-induced conformational changes) or that [125I]TID labeling of the 5-HT3A subunit has no agonist-sensitive component. Additional studies are needed to distinguish between these two possibilities.
At amino acid level, [125I]TID photolabeled Ser451 within the M4 helix of the mouse 5-HT3AR, a residue that corresponds to [125I]TID-labeled α1Thr422 in the αM4 helix of Torpedo nAChR (Figure 7A). Since only a single residue is labeled in the 5-HT3A M4 helix, using the labeling of Ser451 to identify a [125I]TID-labeled face should be taken with a cautionary note. Nonetheless, the labeling of Ser451 suggests that Arg441, Leu444, Val447, Ser451, Leu454 form the lipid exposed interface of 5-HT3AR M4 helix (Figure 7B). Closer examination of the amino acids labeled by [125I]TID in the M4 segments of the 5-HT3AR and nAChR α1 subunit (Figure 7A) reveals a labeling preference for Cys, Ser, and Met residues and a general lack of insertion into aliphatic residues (Val, Leu, etc). This result is consistent with the established relative reactivities of amino acid side chains with trifluoromethylphenylcarbenes (45) and a low affinity interaction. In contrast, as a nAChR noncompetitive antagonist [125I]TID binds with micromolar affinity to the closed channel of the Torpedo receptor and labels a homologous set of aliphatic residues in each subunit (e.g. α1Leu251 and α1Val255) with ~10-fold greater efficiency than residues situated at the lipid-protein interface (42, 44). Given that the labeled (lipid-exposed) face of the 5-HT3AR M4 helix (Figure 7B) is lined predominantly by aliphatic residues (Arg441, Leu444, Val447, Ser451, Leu454), the selective labeling of Ser451 is therefore not unexpected.
While additional work is clearly needed, including the development of protocols to isolate for sequence analysis the individual M1, M2, and M3 transmembrane helices, our results demonstrate the merit of using photolabeling techniques and protein chemistry to study drug binding sites in 5-HT3AR purified from an expression system.
We thank Dr. Guoxiang Luo (NIH, NIAAA) for technical assistance in making the αBgTx-pharmatope tagged mouse 5-HT3AR construct, Dr. Jose-Luis Redondo (Department of Pharmacology and Neuroscience, TTUHSC) for his cell culture assistance and Dr. David C. Chiara (Department of Neuroscience, Harvard Medical School) for assistance with protein sequencing. We also thank Sarah Hiyari and Heidi Hsiao (TTU/HHMI undergraduate fellows) for technical assistance.
†This research was supported in part by American Heart Association South Central Affiliate Grant-In-Aid 0755029Y (M.P.B.), by the South Plains Foundation (M.P.B), by grant GM-58448 from the National Institute of General Medical Sciences (J.B.C.) and by NS-43438 from the National Institute for Neurological Disorders and Stroke (T.K.M.).
1Abbreviations: nAChR, nicotinic acetylcholine receptor; αBgTx, α–bungarotoxin; 5-HT, serotonin; αBgTx-5-HT3AR, mouse 5-hydroxy tryptamine type 3A receptor with a C-terminal α–BgTx-pharmatope tag; HEK-αBgTx-5-HT3AR: LGIC, ligand gated ion channel: human embryonic kidney-293 cells stably expressing αBgTx-5-HT3ARs. DPBS, Dulbecco’s phosphate buffered saline; HPLC, high-performance liquid chromatography; SDS-PAGE, sodium dodecyl sulfate polyacrylamide gel electrophoresis; TFA, trifluoroacetic acid; PTH, phenylthiohydantoin; [125I]TID, 3-trifluoromethyl-3-(m-[125I] iodophenyl) diazirine; Tricine, N-tris(hydroxymethyl)methylglycine; VDB, vesicle dialysis buffer; V8 protease, S. aureus endopeptidase Glu-C
SUPPORTING INFORMATION AVAILABLE
Shown is the amino acid sequence of the αBgTx-5-HT3A subunit (Supplementary Figure S1); the equilibrium binding of [3H]5-HT to αBgTx-5-HT3ARs in HEK-293 cell membranes (Supplementary Figure S2) and the effect of serotonin-treatment during culturing of HEK-αBgTx-5-HT3AR cells on the expression level of αBgTx-5-HT3ARs in HEK-293 cell membranes as measured by the equilibrium binding of [3H]BRL-43694 (Supplementary Figure S3). This material is available free of charge via the Internet at http://pubs.acs.org.