High-Affinity Epibatidine Binding of Functional, Human α7-Nicotinic Acetylcholine Receptors Stably and Heterologously Expressed de Novo in Human SH-EP1 Cells

Nicotinic acetylcholine receptors (nAChRs) are members of a superfamily of neurotransmitter-gated ion channels that also include GABA_A, glycine, and 5-hydroxytryptamine (serotonin)₃ receptors (Lindstrom, 1996; Lukas, 1998; Lukas et al., 1999; Clementi et al., 2000; Leonard and Bertrand, 2001). A variety of nAChR subtypes have been described in the brain, autonomic ganglia, sensory tissues, and other organs (opus citatum). nAChR subtype diversity is due in part to nAChR assembly as homo- or heteropentamers containing different subunits (α1–α10, β1–β4, γ, δ, and ϵ) encoded by a presently identified family of 17 genes (Sargent, 1993; McGehee and Role, 1995; Lukas et al., 1999; Elgoyhen et al., 2001; Leonard and Bertrand, 2001).

In the central nervous system, a major nAChR subtype that engages in high-affinity binding of α-bungarotoxin (Bgt) is composed of homopentamers of α7 subunits (Chen and Patrick, 1997). Human α7-containing nAChRs (hα7-nAChRs) have been stably expressed in mammalian cell lines and function as Bgt-sensitive ion channels exhibiting rapid kinetics of activation and inactivation (Puchacz et al., 1994; Gopalakrishnan et al., 1995). α7-nAChRs are Ca²⁺-permeable (Couturier et al., 1990; Seguela et al., 1993; Peng et al., 1994; Puchacz et al., 1994; Gopalakrishnan et al., 1995), making it possible that they participate in additional forms of nicotinic signaling other than or in addition to classical excitatory neurotransmission.

Epibatidine (EBDN) is an alkaloid obtained from skin extracts of the poisonous, Ecuadorian frog Epipedobates tricolor (Daly, 1995). EBDN effectively competes with nicotine and cytisine in rat brain preparations and acts as a potent nicotinic agonist (Badio and Daly, 1994; Houghtling et al., 1994, 1995; Alkondon and Albuquerque, 1995; Briggs et al., 1995). EBDN has been promoted as a ligand with selectivity for nAChRs containing α4 or α3 subunits (the accepted nomenclature classifies these as α4*- or α3*-nAChRs, respectively; Lukas et al., 1999). However, because interactions of EBDN with α7-nAChRs are less well understood, we investigated such interactions using native nAChR-null SH-EP1 human epithelial cells as hosts for constitutive, stable, heterologous expression of human α7-nAChRs (hα7-nAChRs).

Materials and Methods

Cell Culture. Cells of the SH-EP1 human epithelial cell line or the SH-SY5Y human neuroblastoma line (kindly provided by Dr. June Biedler, Sloan Kettering Institute for Cancer Research, New York, NY; Biedler et al., 1973) were grown in Dulbecco's modified Eagle's medium supplemented with 10% horse serum, 100 U/ml penicillin, 100 μg/ml streptomycin, and 0.25 μg/ml amphotericin B (all from Invitrogen, Carlsbad, CA) plus 5% fetal bovine serum (Hyclone Laboratories, Logan, UT) in a humidified atmosphere containing 5% CO₂ in air at 37°C (Lukas, 1986; Lukas et al., 1993).

Construction of Human α7/pCEP4 Plasmid and Generation of Stable Cell Lines. cDNA encoding a human nAChR α7 subunit (hα7 subunit; Breese et al., 1997) was excised from the pcDNAI vector at the BamHI restriction site and subcloned into the BamHI site of the plasmid pCEP4 (Invitrogen). Final constructs were verified by restriction mapping and complete sequencing of the insert. SH-EP1 cells (or SH-SY5Y cells) were transfected using electroporation with a Bio-Rad Gene Pulsar at 960 μF and 0.20 kV/cm (t = 28–36 ms) with the pCEP4-hα7 construct. Forty-eight hours after transfection, culture medium was supplemented with 0.25 mg/ml hygromycin. Growth was monitored until ring cloning was used to isolate single, transfected cell colonies, which were then expanded. ¹²⁵I-Labeled α-bungarotoxin (I-Bgt) binding assays (Bencherif and Lukas, 1993) were used to screen stable transfectants for expression of hα7-nAChRs. Three lines (clones 12, 16, and 28) studied in greater detail stably express hα7-nAChRs of indistinguishable properties. The clonal line derived from transfected SH-EP1 cells used for most of the studies described herein was clone 12 and was named the SH-EP1-pCEP4-hα7 line. One of the stably transfected SH-SY5Y cell clones that expressed hα7-nAChRs at high levels was also used for the studies described herein, and this cell line was named the SH-SY5Y-pCEP4-hα7 line. The previously described SH-EP1-Toff-hα7 cell line (Peng et al., 1999) that stably expresses hα7-nAChRs de novo under control of a tetracycline-sensitive promoter and the previously described SH-SY5Y-pCEP4-rα7 cell line (Puchacz et al., 1994) that heterologously and stably expresses rat α7 subunits and high levels of α7-nAChRs above a background of human, native α3*-nAChR and native hα7-nAChR expression also were used in the current studies. Wild-type SH-SY5Y cells were maintained as described previously (Lukas et al., 1993), and rat brains were obtained from Pel-Freeze (Rogers, AR).

RNA Isolation and Northern Blot Analysis. Total cytoplasmic RNA was extracted from transfected cells using the Fast Track (Invitrogen) technique. RNA was separated on 1% formaldehyde/agarose gels and transferred to nylon blotting membrane by pressure blotting. Blots were probed with ³²P-radiolabeled, full-length, human nAChR α7 subunit cDNA (Bencherif et al., 1995).

Immunocytochemistry. To visualize nAChR α7 subunits, SH-EP1-pCEP4-hα7 cells were plated onto 22 × 22-mm glass coverslips and grown for 2 days. Cells were then rinsed with phosphate-buffered saline (PBS), fixed with 4% paraformaldehyde for 10 min, and rinsed again before blocking for 20 min with PBS containing 4% normal horse serum, 1% bovine serum albumin, and 0.4% Triton X-100. Mouse monoclonal antibody 306 (Sigma-Aldrich, St. Louis, MO), targeting an intracellular epitope (due to a lack of suitable anti-α7 antibodies targeting extracellular epitopes, immunofluorescent assessment of cell surface α7 subunit expression was not a viable option), was added overnight at 4°C in blocking solution, or primary antibody was omitted from control samples. Cells were rinsed again and incubated for 30 min with 1% bovine serum albumin-PBS containing biotinylated secondary antibody (horse anti-mouse IgG; Vector Laboratories, Burlingame, CA) and then for 30 min with avidin-Alexa 488 (Molecular Probes, Eugene, OR), with PBS rinses between steps. Staining was visualized using epifluorescence (Olympus IX70; Olympus, Melville, NY).

¹²⁵I-Labeled α-Bungarotoxin Binding Autoradiography. Cultured cells grown for 2 days on Lab Tek II CC2 8-chambered, multi-well slides (Nalge-Nunc International, Rochester, NY) were rinsed at room temperature in PBS for 5 min and then incubated in fresh binding buffer (118 mM NaCl, 4.8 mM KCl, 2.5 mM CaCl₂, 1.2 mM MgSO₄, and 20 mM HEPES, pH 7.5 with NaOH) containing 5 nM I-Bgt (±10 mM nicotine to define nonspecific/total radioligand binding to cell-surface sites) for 60 min at 22°C. After radioligand incubation, samples were rinsed twice in 1× binding buffer at 22°C, once in 1× buffer at 4°C for 5 min, once in 0.1× buffer at 4°C for 10 s, and then in 5 mM HEPES at 4°C for 10 s. Subsequently, slides were fixed in 4% paraformaldehyde (50:50 solution: 8% paraformaldehyde and 0.2 M phosphate buffer, pH 7.2) for 20 min at 22°C before being rinsed briefly in H₂O and air-dried at 22°C. Samples were then subjected to electronic isotope counting using an Instant Imager (PerkinElmer Life and Analytical Sciences, Boston, MA) to quantify specific (total minus nonspecific) radioligand binding (i.e., integrating counts across defined areas of sections or cultures containing equivalent amounts and densities of cells). After radioligand binding quantification, samples were dipped in Kodak NTB-3 nuclear track emulsion and allowed to dry overnight at 22°C in a dust-free environment. Slides were then stored at 4°C for exposure periods ranging from 2 to 5 days (showing a linear grain development response while revealing sites of low level radioligand binding). After exposure, slides were developed for 3 min at 22°C with Kodak D-19 developer and fixed for 3 min at 22°C with Kodak fixer. Slides were counter-stained with 0.5% cresyl violet acetate and serially dehydrated with 50, 75, 85, and 95% ethanol. After ethanol treatments, slides were dipped twice in xylene for 5 min each at 22°C and dry-mounted with Permount. Images were captured using an Olympus IX70 inverted microscope and MagnaFire camera and software and then stored and maximized for Hi Gauss clarity and sharpness using ImagePro Plus (Media Cybernetics, Inc., Silver Spring, MD).

Immunoprecipitation and Western Blot Analysis. Membrane preparations (see below) from SH-EP1-pCEP4-hα7 cells were suspended in Ringer's buffer (150 mM NaCl, 5 mM KCl, 1.8 mM CaCl₂, 1.3 mM MgCl₂, and 33 mM Tris, pH 7.4) supplemented with 1% Triton X-100, incubated for 10 min at room temperature, and centrifuged at 45,000g for 10 min at 4°C. The supernatant containing detergent-solubilized hα7-nAChRs was incubated for 1 h at room temperature and overnight at 4°C with Omnisorb (Calbiochem, La Jolla, CA) previously complexed with goat anti-α7 subunit antisera (C-20, sc-1447; Santa Cruz Biotechnology, Inc., Santa Cruz, CA). The sample was subjected to centrifugation at 5000g for 10 min to recover immunoaffinity-purified hα7-nAChRs in the pellet, which was extracted with sodium dodecyl sulfate-polyacrylamide gel electrophoresis buffer. The immunoisolate was then subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred via electroblotting to a nylon membrane. The membrane was incubated with rabbit anti-α7 subunit antisera (H-302, sc-5544; Santa Cruz Biotechnology, Inc.), rinsed free of unbound antibody, and probed with ¹²⁵I-labeled protein A (Amersham Biosciences Inc., Piscataway, NJ). The labeled immunoisolate was visualized using an Instant Imager, and its mobility was determined relative to mobilities of molecular mass standards (high mass standards; Bio-Rad, Hercules, CA).

Membrane Preparation and I-Bgt Binding Assays. Confluent, stably transfected or untransfected cells were rinsed with ice-cold Ringer's buffer supplemented with 0.1 mg/ml sodium azide, mechanically dislodged and harvested from dishes, and homogenized using a Polytron (setting 65, 45 s). Rat brain tissue was similarly homogenized. Homogenates were centrifuged at 45,000g for 10 min at 4°C, and pellets containing membrane fractions including all but the smallest microsomes were resuspended in ice-cold Ringer's buffer. I-Bgt (Amersham Biosciences Inc.; diluted with 500 nM unlabeled Bgt to achieve a typical specific activity of 45–65 cpm/fmol) binding to membranes was carried out as described previously, using a centrifugation-based protocol and gamma detection (Tracor 1191 gamma-counter; Lukas, 1984). Incubations were carried out in a final volume of 200 μl containing 50 μl of I-Bgt at a given, indicated concentration (typically 40 nM) and 100 μl of cell membrane suspension for 1 h at room temperature in the presence or absence of unlabeled drugs. Nonspecific binding was defined using samples supplemented with 1 μM native Bgt and was subtracted from total binding in other samples to compute specific binding. For I-Bgt binding saturation studies, I-Bgt final concentrations were varied from 10 pM to 40 nM.

I-Bgt Binding to Cell Surface Sites. Based on a previously described assay (Bencherif and Lukas, 1993), cells in culture, sometimes treated over different periods with different concentrations of (–)-nicotine, were rinsed twice with Ringer's buffer before being incubated with 10 nM I-Bgt for 1 h at room temperature in the absence or presence of 1 μM unlabeled Bgt. Binding was terminated by removal of medium followed by three 1-ml rinses in Ringer's buffer. Cells were harvested in 500 μl to 1 ml of 0.1 N NaOH and 0.1% SDS, with shaking, and gamma counting was conducted to quantitate bound I-Bgt.

[³H]EBDN Binding Assays. Membranes prepared as described above were used for ³H-labeled (±)-epibatidine ([³H]EBDN; Amersham Biosciences Inc.) binding assays (modified after Houghtling et al., 1994, 1995) by incubation for 1 to 3 h (or for the indicated time for association rate studies) at room temperature in a final volume of 200 μl containing 50 μl of [³H]EBDN at a given, indicated concentration (typically 4 nM but varying for saturation analyses, for some competition assays, and for kinetics studies) and 100 μl of cell membrane suspension. Samples were supplemented with 50 μl of the specified drug at the indicated concentration for competition assays, with 50 μl of buffer to define total binding, or with 50 μl of nicotine (to a final concentration of 300 μM) to define nonspecific binding. Nonspecific binding was subtracted from total binding in other samples to compute specific binding. Dissociation rate studies were done using bulk (large-volume) master stock samples that were incubated with 1 nM [³H]EBDN for 3 h (other studies done using an initial concentration of 4 nM are not illustrated) and divided into 200-μl aliquots before being diluted 21-fold in assay buffer (physical dilution) or supplemented to 1, 2, or 4 μM(±)-EBDN (chemical dilution) for specified periods before assay termination. To accomplish assay termination, free [³H]EBDN was separated from bound [³H]EBDN using polyethyleneimine-coated GF/B glass fiber filtration and a Whatman manifold (Lukas, 1990), and filters were incubated while shaking overnight in liquid scintillation fluid before sample [³H]EBDN was quantified using liquid scintillation counting.

Whole-Cell Current Recording. Nicotine-evoked currents were recorded in the whole-cell configuration using an Axopatch 200B amplifier (Axon Instruments Inc., Union City, CA). Currents were digitized and recorded using pCLAMP software (Axon Instruments Inc.). Cells were continuously superfused with an external bath solution containing 140 mM NaCl, 5 mM KCl, 1 mM MgCl₂, 1.8 mM CaCl₂, 10 mM HEPES, and 10 mM glucose (pH 7.4, 290–300 mmol/kg). Nicotine (3, 30, 100, and 300 μM) was delivered for 1 s once every 30 s via a multibarrel fast perfusion system (Warner Instrument, Hamden, CT). Patch pipettes were pulled from borosilicate capillary glass tubing using a Flaming/Brown micropipette puller (P97; Sutter Instrument Company, Novato, CA) and filled with an internal pipette solution consisting of 130 mM CH₃SO₃K, 10 mM KCl, 1 mM MgCl₂, 0.5 mM CaCl₂, 5 mM EGTA, 10 mM HEPES, 3 mM Mg²⁺-ATP, and 0.3 mM Na-GTP (pH 7.2, ∼280 mmol/kg). The resistances of the patch pipettes when filled with internal solution ranged between 3 and 6 MΩ. Experiments were conducted at room temperature.

Data Analysis. Parameters (dissociation constant, K_D, and maximum binding level, B_max) for radioligand binding were determined from nonlinear curve-fitting analysis of binding saturation isotherms (GraphPad Software, Inc., San Diego, CA). In competition experiments, 50% inhibition constants (IC₅₀ values), levels of maximum and minimum binding as a percentage of control specific binding, and Hill coefficients (n_H) were determined from plots of specific binding as a function of the logarithm of the concentration of competing ligand (GraphPad Software, Inc.) using the standard logistic equation for one or two binding sites. In some cases, the Hill coefficient was fixed at –1 for radioligand binding competition assays. In cases where specific radioligand binding was incompletely blocked, the extent of inhibition as a percentage of specific binding in control samples was determined based on fits to the logistic equation. However, in most cases, levels of maximum or minimum binding were within 1 standard error of 100 or 0%, respectively, of control specific binding. Mean ± S.E.M. values are provided unless otherwise noted.

Materials. All other techniques and commercial sources for reagents were as indicated previously (Lukas, 1984, 1986, Bencherif and Lukas, 1993; Peng et al., 1999).

Results

Screening of Transfectants. In our previous studies, we demonstrated that SH-EP1 human epithelial cells do not express any endogenous human nAChR subunits as messenger RNA or as radioligand binding sites (Lukas et al., 1993). Colonies derived from SH-EP1 cells transfected with the pCEP4 plasmid containing the hα7 subunit cDNA fragment were selected based on their resistance to hygromycin, subcloned by limiting dilution, and propagated. Screening studies using I-Bgt binding revealed that 20 of the 24 clones expressed I-Bgt binding sites de novo. Absolute levels of expression of I-Bgt binding sites varied across and within these clones for reasons that were not systematically investigated. However, three clones showing the most stable, high-to-moderate levels of expression were selected for further study (clones 12, 16, and 28). Unless otherwise noted, studies reported here were done using clone 12, the cell line named SH-EP1-pCEP4-hα7.

Transgene Expression as Message and Protein. Northern blot analyses using a panel of full-length rat or human nAChR subunit cDNAs as probes confirmed lack of expression of any nAChR subunit as messenger RNA in untransfected SH-EP1 cells. However, Northern analysis using full-length hα7 subunit cDNA as a probe showed that a 2100-base pair mRNA is expressed in transfected SH-EP1-pCEP4-hα7 cells, but not in untransfected SH-EP1 cells (data not shown). Western analysis of immunoaffinity-purified hα7-nAChRs identified an anti-α7 subunit immunoreactive protein migrating with an apparent molecular mass of 53 kDa (based on mobility of the set of mass standards used). No such band was not evident upon immunoprecipitation of extracts from untransfected cells, but a polypeptide from SH-EP1-pCEP4-hα7 cell extracts of the same mobility also was evident after α-cobratoxin affinity purification (data not shown).

Immunofluorescence studies using a monoclonal antibody targeting a cytoplasmic epitope of the α7 subunit indicated specific staining [no staining was evident in untransfected SH-EP1 cells (data not shown) or in the absence of primary antibody (Fig. 1a)] of virtually every SH-EP1-pCEP4-hα7 cell (Fig. 1b). Staining was more intense in perinuclear areas, but generally filled the cytoplasm, even in processes emanating from soma.

hα7-nAChR Expression as α-Bungarotoxin Binding Sites. Cell surface I-Bgt-receptor binding microautoradiography (Fig. 1, c and e) conducted in the presence of nicotine to define nonspecific binding showed minimal silver halide grain development over cell bodies and very low levels of background grain development. By contrast, diffuse, heightened grain development over the entire slide and very dense grain development over cell bodies characterized the labeling pattern in SH-EP1-hα7 cell samples exposed to I-Bgt only (Fig. 1, d and f), but not in untransfected SH-EP1 cells (data not shown). Silver grain development was evident at many SH-EP1-hα7 cell-cell boundaries (i.e., on cells that were in proximity) as well as on solitary cells. More than 80% of SH-EP1-hα7 cells showed labeling (∼30% of cells were densely labeled), but the remaining cells showed no or very low surface I-Bgt binding despite positive selection for transfectants and immunostaining evidence for expression of at least internal pools of α7 subunit epitope in all cells.

I-Bgt saturation assays showed that the established, transfected clonal lines possess moderate-to-high levels (B_max value of ∼0.3–3.5 pmol/mg membrane protein) of high-affinity binding sites presumed to represent hα7-nAChRs (Fig. 2). The apparent K_D value of ∼7 nM for the data set shown is a preequilibrium value likely underestimating the true K_D value due to slow association of I-Bgt with its sites and slow dissociation of I-Bgt from radioligand-receptor complexes. The specific, apparent K_D value for any experiment varied somewhat (∼2–7 nM), apparently reflecting differences in biological activity of radiolabeled material across commercial I-Bgt preparations. Although not studied systematically, variation in B_max values reflected differences in cell passage (sometimes higher passage leads to loss of nAChR expression) and/or cell plating density at the time of harvest for assay (highest expression occurs for cells near confluence), as we have noted in studies of other naturally or heterologously expressed nAChRs in human cell lines. No specific I-Bgt binding was evident in untransfected SH-EP1 cells. Studies of the cellular distribution of I-Bgt binding sites indicated that ∼62% of the total number of I-Bgt binding sites isolated in membrane fractions (which are obtained under conditions where binding sites on very small microsomes might escape detection) are typically expressed on the cell surface of SH-EP1-pCEP4-hα7 cells (Fig. 3).

Screening for Functional hα7-nAChR Expression. Whole-cell current recordings using transfected SH-EP1 cells (clone 28) revealed functional hα7-nAChRs mediating inward currents stimulated by (–)-nicotine (Fig. 4; see Zhao et al., 2001, 2003 for results of other studies of functional hα7-nAChRs in SHEP1-pCEP4-hα7 cells) that were not evident in parallel studies of untransfected SH-EP1 cells (data not shown). The inward current responses in SH-EP1-hα7 cells were transient, reflecting the rapid activation and desensitization of hα7-nAChRs.

Nicotine Exposure-Induced Up-Regulation of hα7-nAChRs Measured as Toxin Binding Sites. Exposure of SH-EP1-pCEP4-hα7 cells for 2 days to (–)-nicotine resulted in a significant increase (up to 100%) in I-Bgt binding sites (Fig. 5). The effect of nicotine was both dose-dependent (half-maximal effect at ∼100 μM; Fig. 5) and time-dependent (half-maximal increase at about 12 h of drug exposure; data not shown).

Epibatidine Binding of hα7-nAChRs. [³H]EBDN binding saturation assays also showed that SH-EP1-pCEP4-hα7 cells expressed sites with picomolar affinity (macroscopic K_D value of ∼0.6 nM; B_max value up to 3.5 pmol/mg) for this radioligand, which binds in a Bgt-sensitive manner (Fig. 6). No specific binding of [³H]EBDN occurs for untransfected SH-EP1 cells, consistent with their previous characterization as native nAChR-null (Lukas et al., 1993). As was the case for determination of specific I-Bgt binding sites, variability in B_max values reflected influences of cell passage number and/or cell plating density at the time of harvest for assay. However, for paired studies, numbers of [³H]EBDN and I-Bgt binding sites in the same preparations agreed within 10% after correction for radioligand dissociation during sample processing (e.g., 3.3 and 3.5 pmol/mg membrane protein, respectively).

Association rate kinetics studies showed that [³H]EBDN binding to hα7-nAChRs in SH-EP1-pCEP4-hα7 cells can be described as a monoexponential process (Fig. 7, top). Observed rates of [³H]EBDN binding at concentrations of 0.1, 0.3, and 1 nM varied between 6.5 and 9.5 min^–1 nM^–1 (half-times of 4.4–6.4 s). A plot (not shown) of observed association rates versus [³H]EBDN concentration yielded a slope corresponding to k₁ of 3.1 min^–1 nM^–1 and a y-intercept corresponding to k_–1 of 6.3 min^–1, yielding a microscopic _Kd value of 2.0 nM, in reasonably good agreement with the macroscopic K_D value of ∼0.6 nM determined using saturation binding assays.

Independent empirical studies of dissociation kinetics revealed a dependence of the apparent dissociation rate on the method used in the study. A comparatively slow rate of dissociation occurred for physical dilution, in which reaction samples were diluted 21-fold before separation of bound from free radioligand (k_–1^app = 1.3 min^–1; Fig. 7, bottom). Observed rates of dissociation of bound [³H]EBDN were faster under conditions of chemical dilution, in which reaction mixtures were supplemented with unlabeled EBDN to promote radioligand dissociation (k_–1^app = 1.5–4.4 min^–1; Fig. 7, bottom), and in the presence of higher concentrations of unlabeled drug. These findings are hallmarks of negative cooperativity in radioligand binding to interacting sites, possibly reflecting the presence of five possible binding sites on the homopentamer but weakening of binding if more than two sites are occupied at once. Calculations done using empirically derived rates of [³H]EBDN dissociation gave microscopic K_D values between 0.41 nM (for physical dilution-based measures of k_–1) and 1.4 nM (for chemical dilution-based measures of k_–1), also in good agreement with determinations of macroscopic K_D values.

To assess whether [³H]EBDN binds specifically to hα7-nAChRs from other preparations, [³H]EBDN binding saturation assays also were done using membrane fractions from SH-EP1-Toff-hα7 cells (Peng et al., 1999) transfected with hα7 subunit cDNA under control of a tetracycline-regulated promoter and heterologously expressing hα7-nAChRs (Fig. 8). The results showed that heterologously expressed hα7-nAChRs in this cell line also have high affinity for [³H]EBDN (K_D = 0.46 nM). Other [³H]EBDN binding saturation studies were done using membranes from SH-SY5Y-pCEP4-hα7 cells generated for this study and heterologously expressing recombinant hα7 subunits in a natural background of hα7 subunits and human nAChRs containing α3 subunits (α3*-nAChRs; Fig. 8). Specific [³H]EBDN binding in these cells defined using sensitivity to nicotine blockade suggested expression of high-affinity [³H]EBDN binding, and analysis of specific binding defined using sensitivity to 1 μM Bgt indicated expression of sites with a K_D value of 1.2 nM, in close agreement with K_D values for [³H]EBDN binding to hα7-nAChRs expressed de novo in transfected SH-EP1 cells (Fig. 8). Sites showing insensitivity to blockade by Bgt are likely to be naturally expressed human α3*-nAChRs found in SH-SY5Y cells.

Epibatidine Binding to Other α7-nAChRs. Other [³H]EBDN binding saturation analyses using SH-SY5Y-pCEP4-rα7 cells heterologously expressing rat α7 subunits in a natural background of hα7 subunits and human α3*-nAChRs revealed Bgt-sensitive [³H]EBDN binding sites with K_D values of ∼4 to 16 nM and Bgt-insensitive [³H]EBDN sites with K_D values of about 120 pM (data summarized but not shown). Additional [³H]EBDN binding saturation studies revealed Bgt-sensitive [³H]EBDN binding in rat brain membranes with an apparent K_D value of ∼8 nM and Bgt-insensitive [³H]EBDN sites with a K_D value of ∼20 pM (data summarized but not shown). Similarly, [³H]EBDN binding saturation studies done using untransfected SH-SY5Y cells indicated expression of Bgt-sensitive [³H]EBDN binding sites with an apparent K_D value of ∼1 nM and Bgt-insensitive [³H]EBDN sites with a K_D value of ∼40 pM (data summarized but not shown). In some cases, two site fits to [³H]EBDN saturation curves (not shown) obtained using nicotine to define specific binding to rat brain, wild-type SH-SY5Y cell, or transfected SH-SY5Y cell preparations allowed clear separation of two classes of binding sites with K_D values like those for Bgt-sensitive and Bgt-insensitive sites. Thus, analyses of specific [³H]EBDN saturation binding studies using wild-type or transfected SH-SY5Y cells were consistent with expression of Bgt-sensitive, lower affinity [³H]EBDN binding sites corresponding to human or rat α7-nAChRs at different levels of expression, either naturally expressed in wild-type cells, or representing a mixture of naturally and heterologously expressed forms in transfected cells. In addition, these analyses were consistent with expression of Bgt-insensitive, higher affinity [³H]EBDN binding sites corresponding to naturally expressed α3*-nAChRs found in wild-type or transfected SH-SY5Y cells. Similarly, saturation analyses using rat brain indicated the presence of Bgt-sensitive, lower affinity [³H]EBDN binding sites corresponding to α7-nAChRs and a predominant population of Bgt-insensitive, higher affinity [³H]EBDN binding sites presumably correlated with α4*-nAChRs. However, heterologous de novo expression of hα7 subunits in transfected SH-EP1 cells gave the clearest indications of Bgt-sensitive [³H]EBDN binding properties of hα7-nAChRs.

Mutual Inhibition of Epibatidine and α-Bungarotoxin Binding to α7-nAChRs. Results indicating that EBDN and Bgt bind at overlapping sites on hα7-nAChRs in transfected SH-EP1-pCEP4-hα7 cells and that hα7-nAChRs have higher affinity for EBDN than for Bgt come from a series of radioligand binding competition studies. Specific I-Bgt binding at 10 nM was inhibited by EBDN with an IC₅₀ value of 13 nM (Fig. 9A) and by Bgt with an IC₅₀ value of 35 nM (Fig. 9C; Table 1). However, whereas all of specific I-Bgt binding was blocked in the presence of Bgt, a subset (∼20%) of specific I-Bgt binding sites were not blocked by EBDN at the highest concentration tested (Fig. 9, A and C; Table 1). Nevertheless, “small drug-insensitive but Bgt-sensitive” I-Bgt binding sites were observed when using other drugs (Fig. 12) and other α7-nAChR expression systems (Peng et al., 1999; see other results in Fig. 9; Table 1). Similarly, specific [³H]EBDN binding at 1 nM to sites in SH-EP1-pCEP4-hα7 cells was completely blocked in the presence of Bgt with an IC₅₀ value of 45 nM (Fig. 10C), but specific [³H]EBDN binding was even more sensitive to autologous (and complete) blockade by EBDN (IC₅₀ value of 3.2 nM; Fig. 10A; Table 1).

Other radioligand binding competition studies showed that specific I-Bgt binding at 10 nM to sites on SH-EP1-Toff-hα7 cells (heterologously expressing hα7-nAChRs) was inhibited by EBDN (IC₅₀ = 14 nM; Fig. 9B; Table 1). Moreover, all specific [³H]EBDN binding sites at 1 nM on SH-EP1-Toff-hα7 cells were blocked by Bgt with an IC₅₀ value of 47 nM (Fig. 10D; Table 1). Similar studies demonstrated that all I-Bgt binding sites in SH-SY5Y cells transfected with either human or rat α7 subunits or in wild-type SH-SY5Y cells were blocked by Bgt (IC₅₀ values of 21, 20, and 26 nM, respectively; Fig. 9, C and D; Table 1). All I-Bgt binding sites on these cells were also blocked by EBDN (IC₅₀ values of 22, 36, and 17 nM, respectively, for SH-SY5Y-pCEP4-hα7, SH-SY5Y-pCEP4-rα7, or wild-type SH-SY5Y cells; Fig. 9, A and B; Table 1). Whereas all specific [³H]EBDN binding sites on these cells were blocked by EBDN (IC₅₀ values of 7.0, 8.4, and 2.8 nM, respectively, for SH-SY5Y-pCEP4-hα7, SH-SY5Y-pCEP4-rα7, or wild-type SH-SY5Y cells; Fig. 10, A and B), only a subset of those [³H]EBDN binding sites were blocked by Bgt in transfected SH-SY5Y cells (56% inhibition and 9.3 nM IC₅₀ value for SH-SY5Y-pCEP4-hα7 cells; Fig. 10C; 35% inhibition and 8.2 nM IC₅₀ value for SH-SY5Y-pCEP4-rα7 cells; Fig. 10D) or in wild-type SH-SY5Y cells (only 6% inhibition; Fig. 10D; Table 1). The remaining sites that were insensitive to Bgt block would correspond to naturally expressed α3*-nAChRs characterized in these cells (Lukas et al., 1993; Ke et al., 1998). Similarly, studies done using rat brain indicated that nearly all I-Bgt binding sites are blocked by EBDN (87% inhibition, 6.7 nM IC₅₀ value; Fig. 9A), whereas all were blocked by Bgt (55 nM IC₅₀ value; Fig. 9C; Table 1). All [³H]EBDN binding sites in rat brain preparations were blocked by EBDN (2.2 nM IC₅₀ value; Fig. 10A), but only a small percentage (8%) were blocked by Bgt (38 nM IC₅₀ value; Fig. 10C; Table 1).

Additional studies showed that proportions of Bgt-sensitive [³H]EBDN binding sites in membranes from rat brain, SH-SY5Y-pCEP4-rα7, or wild-type SH-SY5Y cells were higher for assays done in the presence of 10 nM [³H]EBDN than in the presence of 1 nM [³H]EBDN (Table 1). These findings indicated that Bgt-insensitive, higher affinity binding sites for [³H]EBDN (K_D values of ∼10 pM for α4*-nAChRs in rat brain or K_D values of ∼100 pM for α3*-nAChRs naturally found in SH-SY5Y cells) were already saturated when assays were done using 1 nM [³H]EBDN, but only a fraction of Bgt-sensitive, lower affinity sites (K_D value of ∼1 nM) for [³H]EBDN corresponding to rat or human α7-nAChRs in these preparations were labeled at 1 nM [³H]EBDN. By contrast, Bgt-insensitive α3*- or α4*-nAChR binding sites remained saturated in the presence of 10 nM [³H]EBDN, but a higher fraction of Bgt-sensitive α7-nAChRs became labeled when assay [³H]EBDN concentrations were increased from 1 to 10 nM, and the proportion of [³H]EBDN binding sites corresponding to Bgt-insensitive, lower affinity α7-nAChRs also increased. Other studies demonstrated that α-cobratoxin also inhibited specific [³H]EBDN (1 nM) binding to these preparations to the same extent as did Bgt (inhibition by 88 ± 9% for SH-EP1-pCEP4-hα7 cells, 57 ± 6% for SH-SY5Y-pCEP4-hα7 cells, 37 ± 5% for SH-SY5Y-pCEP4-rα7 cells, and 5 ± 3% for rat brain; data summarized but not shown).

Similar Pharmacological Profiles for Epibatidine and α-Bungarotoxin Binding Sites on hα7-nAChRs. Another series of radioligand binding competition studies revealed that specific [³H]EBDN binding sites on heterologously expressed hα7-nAChRs in SH-EP1-pCEP4-hα7 cells have a nicotinic pharmacological profile (Fig. 11; Table 2). Rank order inhibitory potency and IC₅₀ values for ligands acting to block [³H]EBDN binding were 0.047 μM methyllycaconitine (MLA) > 0.27 μM 1,1-dimethy-4-phenyl-piperazinium (DMPP) > 0.43 μM nicotine > 0.89 μM cytisine > 5.9 μM carbamylcholine ≥7.1 μM d-tubocurarine. This profile closely matched rank order inhibition potency and IC₅₀ values for nicotinic ligand competition toward specific I-Bgt binding to SH-EP1-pCEP4-hα7 cell membranes of: 0.035 μM MLA > 0.087 μM DMPP > 0.91 μM cytisine ∼ 1.0 μM nicotine > 7.2 μM d-tubocurarine ∼ 8.3 μM carbamylcholine (Fig. 12; Table 2; although “small drug-insensitive but Bgt-sensitive” I-Bgt binding sites were evident, perhaps representing incompletely assembled receptors). These profiles also are like those for blockade of specific I-Bgt binding to naturally expressed hα7-nAChRs found in SH-SY5Y human neuroblastoma cells (Lukas et al., 1993) or heterologously expressed hα7-nAChRs in SH-EP1-Toff-hα7 cells (Peng et al., 1999). Moreover, these findings are consistent with at least some overlap between Bgt, EBDN, and other small drug binding sites on hα7-nAChRs. Lending further support to the notion that both naturally expressed α3*-nAChRs and either naturally or a combination of naturally and heterologously expressed α7-nAChRs are found on wild-type or transfected SH-SY5Y cells, MLA competition toward specific [³H]EBDN binding was best fit by a two-site model for either wild-type SH-SY5Y cells or SH-SY5Y-pCEP4-hα7 cells (data not shown). High-affinity (IC₅₀ value of ∼30 nM) blockade occurred for about 85% of specific [³H]EBDN binding sites on SH-SY5Y-pCEP4-hα7 cells and for about 34% of the sites on wild-type SH-SY5Y cells, whereas the remaining sites were inhibited by MLA with lower affinity (IC₅₀ value of ∼3 μM). The sites with high affinity for MLA that are expressed at higher levels in transfected SH-SY5Y cells would seem to be hα7-nAChRs, whereas the lower affinity MLA sites would seem to be α3*-nAChRs.

Discussion

The major findings of this study are that human nAChR α7 subunit CDNAs introduced into native nAChR-null SH-EP1 human epithelial cells lead to production of human α7 subunit message and protein and that subunits assemble to form functional hα7-nAChRs that have high affinity for the radioligands [³H]EBDN and I-Bgt. As we observed in previous studies using a different expression vector (Peng et al., 1999), heterologously expressed hα7-nAChR are found on the cell surface and undergo nicotine exposure-induced increases in expression measured as membrane-bound radioligand binding sites. Based on several criteria, mutually exclusive binding of unlabeled or radiolabeled EBDN or Bgt occurs to a single class of sites on heterologously expressed hα7-nAChRs.

Although we cannot formally exclude the possibility that entities composed of α7 subunits plus other endogenous components could give rise to [³H]EBDN-binding sites, Bgt and EBDN binding profiles for hα7-nAChRs expressed in SH-EP1-pCEP4-hα7 cells match those for other preparations heterologously or naturally expressing hα7-nAChRs or rat α7-nAChRs but using I-Bgt as a radioprobe (Peng et al., 1994, 1997, 1999; Gopalakrishnan et al. 1995). High sensitivities to MLA and DMPP seem to be distinguishing characteristics of human nAChR composed of α7 as opposed to other subunits. The current finding that EBDN inhibits binding of I-Bgt or [³H]EBDN to hα7-nAChR with higher affinity than Bgt suggests that although EBDN binds with higher affinity to α3*- or α4*-nAChR than to α7-nAChR, it can be used as a radioligand for α7-nAChRs. Furthermore, autoradiographic studies and radioligand binding studies using [³H]EBDN and brain or other preparations containing a mixture of nAChR subtypes need to consider interactions with α7-nAChRs.

Functional studies using heterologous or natural expression systems indicated that EBDN is a reasonably potent agonist at α7-nAChRs (see Zhao et al., 2001, 2003, for results of additional studies of functional hα7-nAChRs in SHEP1-pCEP4-hα7 cells). The functional EC₅₀ found for (–)- or (+)-EBDN at human α7-nAChRs heterologously expressed in Xenopus oocytes was 1.1 to 1.2 μM (Gerzanich et al., 1995). Briggs et al. (1995) reported that functional potency for (±)-EBDN was 64-fold higher than that for (–)-nicotine (EC₅₀ values = 1.30 ± 0.11 and 83 ± 10 μM, respectively) for hα7-nAChRs heterologously expressed in oocytes. Alkondon and Albuquerque (1995) reported that EBDN elicited type IA currents presumably mediated by α7-nAChRs in rat hippocampal neurons grown in culture [EC₅₀ values of 2.9 and 4.3 μM for (–)- and (+)-enantiomers, respectively]. It is widely appreciated that nicotinic agonists show higher affinities for many nAChR subtypes when assessed using radioligand binding saturation or competition assays than when assessed based on acute functional potencies. This presumably reflects higher affinity agonist interactions with (or stabilization of) nAChRs in a functionally desensitized compared with lower affinity agonist interactions with nAChRs in a functionally ready state. However, it was not appreciated that the affinity difference would be ∼1000-fold for α7-nAChR interactions with EBDN, although a 3 order of magnitude difference in binding and functional affinities for EBDN also is a feature of human α4β2-nAChRs (Eaton et al., 2003). The likelihood that these findings reflect differences in EBDN binding affinity of recombinant α7-nAChRs heterologously expressed in oocyte as opposed to SH-EP1 cells is diminished because of the findings using rat hippocampal neurons.

IC₅₀ values for unlabeled EBDN or Bgt homologous competition toward 1 nM [³H]EBDN or 10 nM I-Bgt binding, respectively, were ∼3 and ∼30 nM, and application of the Cheng-Prusoff correction (radioligand concentrations were 2–3 times higher than their K_D values as determined in radioligand binding saturation analyses) thus brings radioligand binding homologous competition K_i values much closer to K_D values, but not completely. Commercially provided radioligand specific activities and/or concentrations are not always as represented, possibly confounding accurate determinations of radioligand concentration, but there is little reason to question the purity and stock solution concentrations of unlabeled homologous competitors. We also have noted that ligand depletion is a factor in some of our experiments using more concentrated α7-nAChRs, in that homologous competition of unlabeled EBDN for 3 nM [³H]EBDN binding gives IC₅₀ values of ∼3 nM also. Nevertheless, homologous competition studies serve to calibrate heterologous competition studies. Thus, they show that both Bgt and EBDN are slightly less effective in heterologous competition than in homologous competition studies, possibly suggesting imperfect overlap of binding domains for both ligands. Other than a possibility that there are separate sites interrogated by radioagonist and radiotoxin binding, an alternative explanation would be that Bgt stabilizes a conformer(s) of α7-nAChR that has lower affinity for EBDN than does the conformer(s) stabilized by EBDN and that the converse is also true. The small differences in other ligand apparent affinities for EBDN or Bgt binding sites could also be explained by this circumstance. Interestingly, Gerzanich et al. (1995) found that K_i values for heterologous competition for I-Bgt binding to immunoimmobilized chick α7-nAChR were 590 and 350 nM for (+)- or (–)-EBDN, respectively, and were 9.8 and 3.1 nM, respectively, for competition toward I-Bgt binding to immunoimobilized human α7-nAChR naturally expressed by SH-SY5Y cells [although functional EC₅₀ values for (+)- or (–)-EBDN were only 2-fold higher for chick α7-nAChR than for hα7-nAChR]. This suggests that radioligand binding and functional features of α7-nAChR from different species may differ. However, the K_i values for EBDN competition toward SH-SY5Y hα7-nAChR I-Bgt binding are close to those observed in the present study.

Recent studies on analgesic, locomotor, hypothermic, cardiovascular, and metabolic activity of EBDN indicate the potential importance of it and related compounds in studies targeting nAChRs for therapeutic purposes (Damaj et al., 1994; Levin and Rosecrans, 1994). nAChRs exist as diverse subtypes with unique distributions and pharmacological profiles. However, profiling of nAChRs of defined compositions is necessary toward a complete understanding of these important entities and toward rational nicotinic drug design. The current studies reinforce the use of the SH-EP1 cell line as a host for heterologous expression of specific nAChR subtypes of defined subunit composition and as models for clarification of the physiological roles and therapeutic potential of nAChR ligands, but they also indicate that α7-nAChRs should be considered as targets of EBDN and related compounds.