Ric-3 Promotes Functional Expression of the Nicotinic Acetylcholine Receptor α7 Subunit in Mammalian Cells*

Expression of functional, recombinant α7 nicotinic acetylcholine receptors in several mammalian cell types, including HEK293 cells, has been problematic. We have isolated the recently described human ric-3 cDNA and co-expressed it in Xenopus oocytes and HEK293 cells with the human nicotinic acetylcholine receptor α7 subunit. In addition to confirming the previously reported effect on α7 receptor expression in Xenopus oocytes we demonstrate that ric-3 promotes the formation of functional α7 receptors in mammalian cells, as determined by whole cell patch clamp recording and surface α-bungarotoxin binding. Upon application of 1 mm nicotine, currents were undetectable in HEK293 cells expressing only the α7 subunit. In contrast, co-expression of α7 and ric-3 cDNAs resulted in currents that averaged 42 pA/pF with kinetics similar to those observed in cells expressing endogenous α7 receptors. Immunoprecipitation studies demonstrate that α7 and ric-3 proteins co-associate. Additionally, cell surface labeling with biotin revealed the presence of α7 protein on the plasma membrane of cells lacking ric-3, but surface α-bungarotoxin staining was only observed in cells co-expressing ric-3. Thus, ric-3 appears to be necessary for proper folding and/or assembly of α7 receptors in HEK293 cells.

Nicotinic acetylcholine receptors (nAChRs) 1 are members of the neurotransmitter-gated ion channel superfamily. They are widely expressed in the central and peripheral nervous system (1) where they influence numerous cellular and physiological processes. At least 17 different genes that code for nAChR subunits have been identified (2,3), and they assemble as pentamers in different combinations to form a diverse set of nAChR subtypes (4,5). The simplest case is the homopentameric complex such as that formed by the nAChR ␣7 subunit. The ␣7 receptor, for which ␣-bungarotoxin (␣-Bgt) is a specific and high affinity antagonist, is one of the most abundant re-ceptor subtypes in the mammalian brain (6,7). The high Ca 2ϩ permeability of the ␣7 receptor (8) suggests an involvement in the activation of Ca 2ϩ -dependent events in neurons such as transmitter release, participation in signal transduction, and a variety of modulatory effects (9). In addition, ␣7 receptors have been implicated in a number of diseases such as schizophrenia, Alzheimers, and Parkinsons disease (1, 10 -12).
Heterologous expression of the ␣7 subunit in Xenopus oocytes results in homooligomeric, ␣-Bgt-sensitive receptors that activate and inactivate quickly and are highly permeable to Ca 2ϩ (8,13,14), similar to the properties of ␣7 nAChRs in neuronal cells. Although there have been reports of successful functional expression in some mammalian cell lines (15)(16)(17)(18), measurable levels of functional receptors have been difficult to achieve in multiple cell types and this phenomenon appears to be host-cell dependent (19). The reasons for poor heterologous surface expression in these cells are not well understood. Strategies to increase the number of functional receptors on the cell surface, including alteration of culture conditions (20), the generation of ␣7-5HT3 chimeras (21), and site-directed mutagenesis (22) have met with some success. However, these strategies have resulted in only a modest increase in the number of functional receptors or the generation of non-native receptors, which are not ideal for drug discovery. Consequently, there is continued interest in identifying cellular factors that influence the expression of functional ␣7-containing nAChRs.
A screen to identify genes necessary for nAChR function in Caenorhabditis elegans was recently described (23). The search for suppressors of a dominant mutation in the nAChR subunit DEG-3 led to the identification of mutations in ric-3 (resistant to inhibitors of cholinesterase), and subsequent work demonstrated that ric-3 is required for the maturation of multiple nAChRs in oocytes (23). Recent work indicates that ric-3 is a member of a conserved gene family (24). The human homolog (hric3) has diverse effects on co-expressed receptors, including the enhancement of ␣7-mediated whole cell current amplitudes as well as the reduction of ␣4␤2 and ␣3␤4 currents in oocytes. In this study we further examined the effects of hric3 on ␣7 receptors. In addition to demonstrating increased current amplitudes when co-expressed with ␣7 receptors in Xenopus oocytes, we show an association between ␣7 and hric3 proteins and demonstrate that hric3 promotes the formation of functional ␣7 receptors on the surface of mammalian cells. We also present evidence that ␣7 protein can be detected on the surface of HEK293 cells lacking hric3 and those levels do not change significantly in the presence of hric3, thus implicating hric3 as a mediator of folding and/or assembly of nAChR ␣7 receptors.

EXPERIMENTAL PROCEDURES
Isolation of Hric3 Coding Sequence-Sequences encoding the hric3 subunit were isolated by standard PCR techniques. Briefly, total adult brain RNA (Clontech, Palo Alto, CA) was used as the template for first strand cDNA synthesis using random hexamers and a Retroscript kit (Ambion, Austin, TX). An initial set of oligonucleotide primers was designed based on the hric3 sequences contained in the GenBank TM data base (GenBank TM accession number NM_024557). A sense strand 23-mer, TGCGACCACCGTGAGCAGTCATG (corresponds to hric3 nt Ϫ20 to 3), and an antisense 24-mer, CTGAGGAGAGAGAGGTCACC-TTGG (corresponds to hric3 nt 1142 to 1165), were used in amplification reactions with human adult brain cDNA and KOD Hotstart DNA polymerase (Novagen, Madison, WI). Reactions were performed at 94°C for 5 min followed by 30 cycles of 94°C for 20 s, 55°C for 20 s, and 72°C for 1.5 min and an additional cycle of 72°C for 7 min. A second sense strand 37-mer, CTGAATTCGCCACCATGGCGTACTCCACAGT-GCAGAG (contains hric3 nt 1 to 23 preceded by a ribosome binding site and an EcoRI restriction site), and a second antisense strand 34-mer, CTCTCGAGGAGTAATGGATACTTCAGACTGGCTG (contains hric3 nt 1111 to 1136 followed by an XhoI site), were used in a nested amplification reaction with the original PCR product. Following subcloning the reamplified PCR product was sequenced to confirm its identity.
PCR-Total RNA was isolated from cell lines using TRIzol reagent (Invitrogen, Carlsbad, CA) according to the manufacturer's instructions and first strand cDNA was synthesized as described above. Three primer pairs based on the cloned human sequence and two primer pairs based on partial rat sequences in GenBank (see Table I) were used in amplification reactions performed at 94°C for 5 min followed by 30 cycles of 94°C for 20 s, 58°C for 20 s, and 72°C for 60 s. For human templates, primers to glyceraldehyde-3-phosphate dehydrogenase were used in amplification reactions to show the presence of cDNA.
DNA Constructs and Expression-For initial expression studies the full-length hric3 cDNA was subcloned into the EcoRI-XhoI sites of pcDNA3.1(ϩ). For biochemical studies a hric3 cDNA construct containing the hemagglutinin (HA) tag sequence, YPYDVPDYAL, at its COOH terminus was generated by standard PCR techniques. A full-length human ␣7 sequence was reported previously (25). The insert was subcloned into the EcoRI-XhoI sites of pcDNA3.1(ϩ) for transient expression studies and the BamHI-XhoI sites of pcDNA5/FRT for the development of a stable cell line using the flp-in system (Invitrogen, Carlsbad, CA). A cell line stably expressing the nAChR␣7 subunit, designated A7-3, was developed in HEK293 cells according to the manufacturer's instructions. For some biochemical studies an ␣7 cDNA construct containing the HA epitope tag at a unique HindIII site (nt 82) near the amino terminus of the ␣7 coding sequence was generated. For transient expression studies in the A7-3 stable cell line cells plated in 10-cm dishes were transfected with a total of 24 g of a hric3-containing expression plasmid or the bacterial expression plasmid pGEM7z. For experiments in which both ␣7 and hric3 were transiently expressed, HEK293 cells were transfected with ␣7 and hric3 DNAs in a 1:1 ratio. All transfections were performed with Lipofectamine 2000 (Invitrogen) and transfection efficiency, estimated to be 70 -90%, monitored by expression of a green fluorescent protein construct. Cells were examined for protein expression 24 -48 h after transfection.
Western Blots-Proteins were separated on 8 or 4 -20% gradient acrylamide gels and transferred to nitrocellulose filters (Amersham Biosciences). Blots were blocked overnight at 4°C in PBS-T (PBS containing 0.1% Tween 20) with 5% dried milk followed by incubation with primary antibodies in PBS-T for 3 h at room temperature. The following primary antibodies were used: ␣7 antibody (C-20, Santa Cruz Biotechnologies, Santa Cruz, CA) at a 1:200 dilution, a HA epitope antibody (clone HA-7, Sigma) at a 1:10000 dilution, a p42 MAP kinase antibody (Cell Signaling, Beverly, MA) at a 1:700 dilution, and a COX IV antibody (Molecular Probes, Eugene, OR) at a final concentration of 0.2 g/ml. The blots were washed with PBS-T (3 to 5 changes) followed by incubation for 1 h with a 1:150,000 dilution of horseradish peroxidaselinked sheep anti-goat Ig (Sigma) for ␣7 detection, a 1:3000 dilution of horseradish peroxidase-linked goat anti-rabbit Ig (Amersham Biosciences) for p42 MAP kinase detection, or with a 1:2000 dilution of horseradish peroxidase-linked sheep anti-mouse Ig (Amersham Biosciences) for HA epitope and COX IV detection. Filters were subsequently washed with PBS-T (8 changes) and developed with the Lumi-Glo chemiluminescence substrate reagent (KPL, Gaithersburg, MD).
Immunoprecipitations-Total cell membranes were prepared as described elsewhere (26). Membrane preparation, solubilization, and all subsequent steps were performed at 4°C. Membranes were solubilized in PBS containing 1.0% Triton X-100 and Complete protease inhibitors (Roche Molecular Biochemicals) by rocking at 4°C for 1 h. The lysate was centrifuged at 20,000 ϫ g for 20 min and the supernatant removed to a new tube. To immunoprecipitate the ␣7 subunit, 5 g of ␣7-specific antibody (C-20, Santa Cruz Biotechnology) and 50 l of protein G-Sepharose beads (Amersham Biosciences) were added and incubated overnight. To immunoprecipitate the HA-tagged hric3 subunit 50 l of HA affinity gel (Sigma) was added and incubated overnight. Samples were washed 4 times with 1 ml of chilled PBS containing protease inhibitors, resuspended in SDS-PAGE sample buffer, and analyzed by Western blot as described above.
Cell Surface Biotinylation-Cells on tissue culture dishes were washed 3 times with chilled PBS and incubated in the same buffer containing 0.5 mg/ml sulfo-NHS-LC-biotin (Pierce) for 20 min at 4°C. The reaction was quenched by washing 3 times with chilled PBS containing 100 mM glycine. Cells were lysed and proteins were solubilized in PBS, 1% Triton X-100. Approximately 1 mg of solubilized protein was incubated with 70 l of neutravidin-agarose (Pierce) overnight at 4°C. Beads were washed 4 times and bound protein was eluted by heating in SDS-PAGE buffer. Samples were electrophoresed and analyzed by Western blot with antibodies to ␣7, HA epitope tag, p42 MAP kinase, and COX IV.
Immunofluorescence Staining-Mammalian cells expressing nAChR ␣7 alone, hric3 alone, or nAChR ␣7 ϩ hric3 were plated on poly-Dlysine-coated glass coverslips (BD Biosciences, Bedford, MA) in a 24well plate at a density of ϳ1 to 2 ϫ 10 5 cells/12-mm coverslip. Live, intact cells were stained with Alexa Fluor 488-labeled ␣-Bgt (Molecular Probes) at a 1:500 dilution and a rabbit polyclonal anti-HA antibody (Sigma) at a 1:50 dilution. The secondary antibody was an Alexa Fluor 594-labeled goat anti-rabbit Ig (Molecular Probes) used at a 1:400 dilution. The primary antibody and ␣-Bgt were added directly to the growth media or diluted in PBS/bovine serum albumin and allowed to react with the cells for 1 h. Cells were rinsed 3 times and incubated with the second antibody for 30 min. Cells were rinsed 4 times with 1ϫ PBS and fixed in 4% paraformaldehyde/PBS for 15 min at room temperature. After a final wash with 1ϫ PBS, coverslips were dried and mounted on glass slides for visualization with a Zeiss LSM510 confocal microscope. In other experiments cells were fixed and permeabilized prior to Ab incubations.
HEK293 Electrophysiology-Transient transfections for electrophysiological characterization included pCMVCD4, a human CD4 ϩ expression plasmid, to permit the identification of transfected cells. Prior to recording, cells were washed with mammalian Ringer's solution, incubated for ϳ10 min in a solution containing a 1/1000 dilution of M-450 CD4 ϩ Dynabeads (Dynal Inc., Lake Success, NY) and rewashed with mammalian Ringer's solution to remove excess beads. Expression of functional ␣7 receptors in transfected cells was evaluated 24 -48 h following transfection using the whole cell patch clamp technique. All recordings were performed on single cells at room temperature (19 -24°C). Whole cell currents were recorded using an EPC-9 (HEKA elektronik, Lambrecht, Germany) patch clamp amplifier, low-pass filtered at 1 kHz (Ϫ3 dB, 8-pole Bessel filter), and digitized at a rate of 10 kHz, unless otherwise stated. Pipettes were manufactured from borosilicate glass (TW150, WPI, Sarasota, FL) and had a resistance of 1.1-2.0 M⍀ when filled with internal solution. Series resistances were 2-4 M⍀ prior to compensation. The cell capacitance range was 5.7 to 13.6 pF. The pipette solution contained (in mM): 110 Tris phosphate dibasic, 28 Tris, 11 EGTA, 0.1 CaCl 2 , 4 ATP, 2 MgCl 2 (pH 7.3, adjusted with Tris). The external solution contained (in mM): 120 NaCl, 3 KCl, 2 MgCl 2 , 2 CaCl 2 , 25 glucose, 10 HEPES (pH 7.3, adjusted with Tris). The membrane potential of individual HEK293 cells was held at Ϫ90 mV.
Nicotine was obtained from Sigma. Stock solutions were prepared in water and stored at Ϫ20°C. Nicotine was dissolved in the external solution and applied for 300 ms using a fast application system consisting of a triple-barrel glass pipette attached to an electromechanical switching device (piezo-electric drive, Winston Electronics, Millbrae, CA). The speed of solution exchange between control and nicotinecontaining solution, measured as the open-tip response, displayed a time constant of 0.7 ms, with a steady state reached in less than 1.5 ms.
Xenopus Oocyte Electrophysiology-Sections of ovary were surgically isolated from anesthetized Xenopus frogs (Nasco, Fort Atkinson, WI). Mature females were anesthetized by immersion in a 0.1% tricaine methanesulfonate solution and oocytes were surgically removed. The follicular cell layer was enzymatically removed by gentle shaking with collagenase (Worthington, Type II, 1.7 mg/ml for 90 min, then Sigma, Type II, 1.7 mg/ml for 30 min) in Ca 2ϩ -free Barth's solution. Oocyte injection and incubation methods used in this work were as previously reported (27), with minor changes as indicated below. Oocytes were injected with 50 nl containing 25 ng of ␣7 in vitro synthesized mRNA. In vitro transcribed hric3 RNA was injected in a 1:1 ratio when coinjected with rat or human ␣7. Following injection, oocytes were incubated at 16 -19°C for 4 -7 days in Oocyte Ringer-3 medium containing 50% L-15 medium, 100 g/ml gentamicin, 25 g/ml tetracycline, 4 mM glutamine, and 30 mM Na-HEPES (all from Invitrogen), with pH ad-justed to 7.6 with NaOH. The extracellular recording solution (standard Ringer's) contained (in mM): NaCl (115), KCl (2.5), BaCl 2 (1.8), HEPES (10), atropine (0.001), pH 7.3. Functional expression was examined using the two-electrode voltage clamp technique; membrane potential was held at Ϫ70 mV.
Concentration-response curves were obtained by normalizing the current responses to varying concentrations of ACh to the maximal response observed to saturating concentrations of ACh in each oocyte. Curves were fitted by nonlinear regression to the Hill equation (27). Statistical significance between groups was assessed with a Student's t test (SigmaStat, SPSS Inc.).

Hric3 mRNA Is Present in Neuronal Cells and Absent in
HEK293 Cells-We amplified and subcloned the hric3 coding sequence from human adult brain RNA using primers based on sequence in the GenBank data base. Sequence analysis indicates that the coding sequence isolated here is identical to GenBank TM accession number AY326435 except for the absence of a single serine residue (Ser-173 in AY326435). This polymorphism was described previously (24).
We used PCR to examine hric3 mRNA expression in several cell lines. Primers directed to portions of the hric3 coding sequence (Table I, Fig. 1A) were used to detect transcripts using standard PCR assays. As shown in Fig. 1B, we detected hric3 transcripts in brain tissue and the SH-SY5Y cell line, a human neuronal cell line known to express endogenous ␣7 receptors (28). In contrast, hric3 transcripts were not detected in HEK293 cells with primer pairs covering the majority of the hric3 coding sequence. A primer pair covering the extreme 3Ј end of the coding sequence generated low, but detectable, levels of a fragment of the expected size from HEK293 cells, indicating that low levels of at least a portion of the hric3 transcript may be present in these cells.
We extended our analyses to determine whether ric-3 is expressed in additional cell lines known to express functional ␣7 receptors. Primers homologous to the rat ric-3 subunit (Table I, Fig. 1C) were again used in standard PCR assays using cDNA generated from rat brain, PC12, and GH4C1 cells. In all cases PCR products of the expected size were detected (Fig. 1D), indicating the presence of ric-3 transcripts in these cell lines.
␣7 Protein Is Detected on the Surface of HEK293 Cells Lacking Hric3-We developed a HEK-based stable cell line constitutively expressing the ␣7 subunit (A7-3 cells) using the flp-in system. As demonstrated in Fig. 2A, A7-3

cells express ␣7
protein with a pattern similar to that observed in HEK293 cells transiently transfected with the human ␣7 subunit. The presence of the doublet at ϳ60 kDa observed in HEK293 cells was similar to the pattern observed with human recombinant ␣7 expressed in GH4C1 cells 2 and may reflect differential posttranslational processing of the ␣7 protein. Despite the presence of ␣7 protein, patch clamp studies failed to identify detectable currents through ␣7 channels (Fig. 7).
We next examined if the lack of functional expression was because of lack of surface expression of ␣7 receptors. We biotinylated cell-surface proteins and isolated them by binding to streptavidin-linked agarose beads. The presence of ␣7 protein was subsequently detected by Western blot analysis. Fig. 2B shows that in the A7-3 cell line a detectable, albeit a proportionally low, level of ␣7 protein was biotinylated and pulled down with streptavidin beads, indicating its presence on the cell surface. To exclude the possibility of intracellular protein contamination, we also looked for the presence of COX IV, an integral mitochondrial membrane protein, and p42 MAP kinase, a cytosolic protein, in the biotinylation experiment. As demonstrated in Fig. 2B, neither COX IV nor p42 MAP kinase were detected, consistent with the biotinylation of only cell-surface proteins.
␣7 and Hric3 Proteins Co-associate-To determine whether hric3 can form a stable complex with the human ␣7 subunit we performed coimmunoprecipitations using untransfected A7-3 cells, A7-3 cells transiently expressing the hric3-HA fusion protein, or HEK293 cells transiently expressing the hric3-HA protein. Proteins were immunoprecipitated with subunit-specific antibodies from detergent extracts of total membrane fractions. The antibody to the HA epitope tag immunoprecipitated both the ␣7 subunit and the hric3-HA fusion protein from hric3-transfected A7-3 cells (Fig. 3, samples 1 and 5). There was no signal when using the HA epitope antibody with untransfected A7-3 cell extracts (samples 2 and 6). Conversely, the ␣7-specific antibody immunoprecipitated both the ␣7 and hric3 subunits from hric3-transfected A7-3 cells (samples 3 and 7). There was no signal from hric3-transfected HEK293 cells (samples 4 and 8). These results demonstrate that ␣7 and hric3 are able to form a stable complex in HEK293 cells.
Hric3 Does Not Alter ␣7 Expression Levels in HEK293 Cells-We next examined if hric3 co-expression alters the overall expression levels of the ␣7 subunit or the levels of ␣7 found on the cell surface of HEK293 cells. Total cell lysates and biotin-labeled surface proteins prepared from cells expressing ␣7 alone or ␣7 and hric3 were subjected to Western analysis (Fig. 4A). Lanes 1 and 3 of Fig. 4A indicate that the overall expression levels of ␣7 protein did not differ significantly when co-expressed with hric3. In addition, we observed no significant difference in the levels of cell-surface ␣7 when hric3 was coexpressed (Fig. 4A, lanes 2 and 4). Identical results were obtained when both ␣7 and hric3 were transiently expressed in HEK293 cells (data not shown). These experiments indicate that hric3 does not play a direct role in regulating the trafficking of ␣7 receptors to the cell surface.
We also used the biotin labeling technique to determine whether hric3 can be detected on the cell surface. The presence of hric3 protein in lane 2 of Fig. 4B indicates that a significant proportion of the hric3 protein expressed in HEK293 cells was Surface ␣7 Receptors Are Recognized by ␣-Bgt Only in Cells Expressing Hric3-To monitor the ability of ␣7 receptors to bind ␣-Bgt we incubated live, mock-transfected and hric3transfected A7-3 cells with fluorescently labeled ␣-Bgt and an antibody directed to the HA tag on the hric3 subunit in conjunction with a fluorescently labeled second antibody. Fig. 5A shows that there was no discernable ␣-Bgt binding to mocktransfected A7-3 cells despite the presence of ␣7 protein on the cell surface. Immunostaining of whole hric3-transfected cells (Fig. 5B) resulted in ␣-Bgt binding to a subset of the cells, consistent with the overall transfection efficiency. The lack of a hric3 signal under these experimental conditions is expected given the proposed intracellular location of the HA epitope tag. Immunostaining of fixed and permeabilized hric3-transfected cells (Fig. 5C) demonstrated the presence of HA-tagged hric3 protein in the same cells that displayed ␣-Bgt binding. These experiments indicate that in HEK293 cells hric3 co-expression is necessary for binding of ␣-Bgt to ␣7 receptors.
Hric3 Co-expression Enhances ␣7-Mediated Currents in Oocytes-The magnitudes of currents induced by ACh (10 M to 1 mM) were compared between oocytes injected with rat ␣7 mRNA alone (25 ng in 50 nl) or with rat ␣7 and hric3 transcripts (25 ng each, 50 nl total). Oocytes co-injected with rat ␣7 and hric3 showed larger currents in response to 1 mM ACh (1.5 Ϯ 0.5 A, mean Ϯ S.D.) than oocytes injected with rat ␣7 transcript alone (0.68 Ϯ 0.5 A; p ϭ 0.01, t test, Fig. 6A). ACh-induced currents were also larger in oocytes co-injected with human ␣7 and hric3 (6.1 Ϯ 2.9 A) than in those injected with human ␣7 nAChR alone (2.7 Ϯ 1.  1. Distribution of ric-3 transcripts. A and C, location of PCR products (PCR1-PCR3 for human and PCR4-PCR5 for rat, respectively) generated from ric-3 coding regions. Coding regions are denoted by shaded rectangles. B and D, PCR assays were performed as described under "Experimental Procedures" and products were separated on 1.5% agarose gels.

FIG. 2. Immunoblot analysis of ␣7 expression in HEK293 cells.
A, total ␣7 expression in HEK293 cells. Total membrane proteins (10 g/lane) from transiently transfected HEK293 cells and an HEK293based stable cell line were separated on an 8% gel and immunostained with polyclonal antisera specific for the ␣7 subunit. B, cell surface expression of the ␣7 subunit. Biotinylated surface proteins from whole cells were isolated with neutravidin-linked beads and separated on a 4 -20% gradient gel. After protein transfer to nitrocellulose the filter was cut horizontally at the 50-kDa marker and below the 36-kDa marker to allow identification of multiple proteins in the same lane. The upper filter was immunostained with ␣7 antisera, the middle filter was immunostained with p42 MAP kinase antisera, and the lower filter was immunostained with a COX IV mAb. The subunits are denoted with arrows. FIG. 3. Physical association of ␣7 and hric3 subunits. Anti-␣7 or anti-HA epitope antibodies were used to immunoprecipitate proteins from Triton-solubilized cell membrane fractions. For hric3 the immunoprecipitation was performed with hric3-transfected A7-3 cells (samples 1 and 5) or with untransfected A7-3 cells (samples 2 and 6). For ␣7 the immunoprecipitation was performed with hric3-transfected A7-3 cells (samples 3 and 7) or with hric3-transfected HEK293 cells (samples 4 and 8). Proteins were separated on 8% gels and immunostained with the indicated antibodies. Total membranes from untransfected HEK293 cells or hric3-transfected A7-3 cells (first two lanes of each blot) were included on the gels as references.
␣4␤4 or ␣3␤2 nAChR subunits. 2 Similarly, untransfected or mock-transfected A7-3 cells (n ϭ 11) failed to show any detectable inward currents upon application of 1 mM nicotine (see Fig. 7). However, following transient transfection of A7-3 cells with a hric3 expression plasmid, we observed robust, agonistinduced inward currents (Fig. 7). All cells that were transfected, as determined by the binding of CD4 ϩ -specific beads, showed clear, detectable currents; mean current density ϭ 42 Ϯ 39 pA/pF (n ϭ 12). Nicotine-induced currents displayed very fast kinetics of activation and desensitization (Fig. 7), a hallmark of ␣7-mediated currents. Thus, hric3 promoted the expression of functional ␣7 receptors in these cells. DISCUSSION There are several examples of the detection of functional, recombinant ␣7 receptors expressed on the surface of mamma- FIG. 4. ␣7 expression levels ؎ hric3 and hric3 surface expression. A, cell surface expression of the ␣7 subunit. Biotinylated surface proteins from A7-3 cells Ϯ hric3 were isolated with neutravidinlinked beads and separated on a 4 -20% gradient gel. Total cell lysate from the same biotinylated cells was run in parallel. After protein transfer to nitrocellulose the filter was cut horizontally at the 50-kDa marker and below the 36-kDa marker to allow identification of multiple proteins in the same lane. The upper filter was immunostained with ␣7 antisera, the middle filter was immunostained with p42 MAP kinase antisera, and the lower filter was immunostained with a COX IV mAb. The subunits are denoted with arrows. B, cell surface expression of hric3. Biotinylated surface proteins from A7-3 Ϯ hric3 were processed as in A but the upper filter was immunostained with the HA tag antibody.  6. Co-expression of rat ␣7 nAChR with hric3 in Xenopus oocytes. A, superimposed traces of current responses to 1 mM ACh applied during the time indicated by the horizontal bar in two different oocytes injected with rat ␣7 (black trace) or rat ␣7 ϩ hric3 (gray trace) transcripts. Oocytes were from the same batch and recorded on the same day. B, the potency for ACh derived from the concentration-response curve obtained in oocytes injected with rat ␣7 ϩ hric3 transcript was 235 M. lian cell lines. Electrophysiological and biochemical techniques as well as ␣-Bgt binding have been used to document the expression of ␣7 in SH-SY5Y, GH4C1, PC12, and SH-EP1 cells (15)(16)(17)19). However, overexpression of recombinant ␣7 in HEK293 cells or the closely related tsa201 cell line has not resulted in sufficient numbers of functional receptors to be detected by these methods 2 (19,29). In contrast, the 5HT3 receptor, another member of the ligand-gated ion channel family that can also assemble as a homooligomer, demonstrates relatively high expression levels and generates large Ca 2ϩ signals in HEK293 cells (30). These data indicate that there are host-specific cellular factors that influence the level of functional ␣7 surface receptors. We show here that co-expression of ␣7 with the hric3 subunit significantly enhances human and rat ␣7-mediated currents in oocytes. In addition, hric3 co-expression results in detectable levels of functional ␣7 receptors in HEK293 cells. We demonstrate that the ability to detect functional ␣7 receptors through whole cell patch clamp studies in HEK293 cells corresponds to the presence of ␣-Bgt binding sites on the cell surface and that the increase in the number of functional receptors is not because of an increase in the amount of protein on the cell surface.
Exogenous ric-3 does not appear to be absolutely necessary for the maturation of ␣7 receptors in Xenopus oocytes. A Xenopus leavis ric-3 homolog has recently been identified by sequence data base searches (24) and expression of this protein in oocytes may account for the ability to detect functional ␣7 receptors when injected only with ␣7 transcripts. In contrast, we demonstrate that HEK293 cells do not express detectable levels of full-length hric3 transcripts and this correlates with the lack of functional human ␣7 receptors. In addition, our data show that the human SH-SY5Y cell line as well as the rat PC12 and GH4C1 cell lines, all cell lines that express native or functional recombinant ␣7 receptors, express significant levels of ric-3 transcripts. Taken together, these data indicate that ric-3 is necessary for the formation of functional nAChR ␣7 receptors in mammalian cells.
We observed a ϳ2-fold increase in both human and rat ␣7mediated currents upon co-expression with hric3. Halevi et al. (24) showed a similar increase in the magnitude of ACh-induced currents when co-expressing human ␣7 and hric3. It is conceivable that the increase in current amplitudes seen with hric3 co-expression in Xenopus oocytes is a result of slower desensitization rather than an increase in the number of functional channels. It is difficult to discern these differences using a two-microelectrode voltage clamp, wherein the channels desensitize much more rapidly than the rate of agonist application. These limitations can be overcome by measuring ␣7 currents in mammalian cells equipped with a fast solution exchange system. Indeed, we measured robust ␣7-mediated currents upon co-expression of hric3 in HEK293 cells. However, a direct comparison to address this issue is complicated by our observation that functional ␣7 receptors are not expressed in mammalian cells lacking ric-3.
Cooper and Millar (19) attempted to express the rat, human, and chick ␣7 subunits in HEK293 cells, but were unable to detect specific surface 125 I-␣-Bgt binding despite the fact that protein could be immunoprecipitated by an ␣7-specific mAb. Our data showing the lack of detectable ACh-induced currents in HEK293 cells expressing ␣7 alone is consistent with those results. In contrast, Gopalakrishnan et al. (31) measured significant levels of 125 I-␣-Bgt binding to cells transfected with a recombinant human ␣7 subunit. The most likely explanation for this discrepancy is the use of different HEK293 cell isolates. It is possible that the HEK293 cell isolate used in that study expressed hric3 at levels high enough to permit the proper folding or assembly of ␣7 subunits.
Our data indicate that even in the absence of functional receptors, ␣7 protein can be detected on the surface of HEK293 cells stably or transiently expressing ␣7. This is consistent with the work of Rakhilin et al. (29). By immunostaining tsa201 cells, an HEK293 derivative, transiently expressing a HAtagged ␣7 subunit Rakhilin et al. (29) found that a significant number of cells stained positive on the cell surface with an anti-HA mAb, although ␣-Bgt binding sites could not be detected. Additionally, we show that co-expression of hric3 has little effect on the overall ␣7 expression levels or on the amount of ␣7 protein on the cell surface. Based on these data it appears that hric3 does not play a direct role in regulating the transport of ␣7 protein to the cell surface, but more likely facilitates proper folding or assembly of ␣7 complexes.
The mechanism by which ric-3 exerts its effects on ␣7 subunits remains to be determined. Halevi et al. (24) suggest that ric-3 is located inside the cell and plays a role in folding or assembly, processes that take place in the endoplasmic reticulum (ER). In our recombinant expression system biotin labeling experiments indicate that some of the hric3 protein expressed is located on the cell surface. Transmembrane prediction algorithms indicate the presence of two transmembrane domains within the hric3 protein, but such algorithms often cannot distinguish between a signal peptide and a transmembrane domain (32). Indeed, sequence analysis using either SIG-NALP2.0, a signal peptide predictor, or Phobius, a relatively new combined transmembrane topology and signal peptide predictor (33), suggests that hric3 contains a signal peptide. Taken together our data indicate that hric3 can be found as an integral plasma membrane protein and likely has a single transmembrane domain with its NH 3 terminus located on the extracellular surface. Whether hric3 exerts its effects on nAChRs in the ER or the plasma membrane, the COOH-terminal coiledcoil domain following the transmembrane domain is predicted to be located within the cytoplasm. Here it may be available to exert its effects directly or bring together ␣7 subunits and, as yet, unidentified proteins involved in the maturation process.
The list of known accessory/chaperone proteins that mediate folding, assembly, or targeting of nicotinic acetylcholine receptors continues to expand. Calnexin, an ER-resident membrane protein, was found to moderately increase overall expression levels as well as facilitate folding and assembly of the ␣ subunit of the mouse muscle-type nAChR (␣1) in COS and HEK293 cells (34,35). BIP, another ER protein found in many cell types, has been shown to bind newly synthesized mouse muscle nAChR ␣ and ␤ subunits in the muscle-like cell line BC3H-1 and has been proposed to participate in the processing and assembly of subunits and subunit complexes (36). 14-3-3, an- other entity reported to act as a chaperone protein, appears to play a role in subunit stabilization as it increased the steady state levels and consequently the surface expression levels of the ␣4 subunit expressed in tsa201 cells (37). Each of these proteins can contribute to nAChR surface expression. Still other proteins, such as lynx1 (38) and the integral membrane protein VILIP-1 (39), have been shown to increase the surface expression levels as well as the functional properties of nAChRs.
In addition to the accessory proteins described above, PDZcontaining proteins of the PSD-95 family have been shown to associate with native ␣7 receptors in chick ciliary ganglia (40), although their identity was not firmly established. In parallel experiments individual PDZ proteins were tested for their ability to associate with ␣7 expressed in HEK293 cells. Interestingly, there was no significant association of ␣7 with any of the PSD95 family members tested. It is possible that splice variants of those members tested or other untested family members associate with ␣7. Alternatively, the folding/assembly step provided by hric3 may be required for the proper association of ␣7 with PDZ-containing proteins.
The precise mechanisms by which ion channels fold and assemble are unknown. Ion channel complexity, brought about by their generally large size and requirements for proper subunit composition and stoichiometry, cause their assembly to be a relatively slow and inefficient process (41). Experimental evidence indicates that nAChRs assemble in the ER, but recent data indicates that subunit folding continues after formation of the nAChR pentamer (41). It is not known if these late folding events occur in the ER, but they appear to be required for channel function. The precise role that ric-3 plays in this process, including the location within the cell where it exerts its effects, is unclear and requires further exploration. Nevertheless, robust expression of ␣7 receptors in non-neuronal cells for high-throughput screening purposes should now be possible with the co-expression of ric-3. Additional characterization of the interactions between ric-3 and nAChRs will further the understanding of the control of nAChR expression and function.