Cell-specific effects on surface α7 nicotinic receptor expression revealed by over-expression and knockdown of rat RIC3 protein

Authors


Address correspondence and reprint requests to Dr. Ralph H. Loring, Department of Pharmaceutical Sciences, 140 The Fenway, Mailstop 148 TF, Northeastern University, 360 Huntington Avenue, Boston, MA 02115, USA. E-mail: r.loring@neu.edu

Abstract

We tested whether surface α7 nicotinic acetylcholine receptor expression is dependent on an endogenous chaperone named Resistance to Inhibitors of Cholinesterase 3 (RIC3) by comparing RIC3 protein in rat GH4C1 and human SH-EP1 cells, which express strikingly different surface receptor levels following α7 transfection. Cloned rat RIC3 exists in at least two isoforms because of an ambiguous splice site between exons 4 and 5. Both rat isoforms permit surface α7 expression in SH-EP1 and human embryonic kidney (HEK) cells measured by α-bungarotoxin binding. Contrary to expectations, endogenous RIC3 protein expression determined by immunoblots did not differ between untransfected GH4C1 or SH-EP1 cells. siRNA against rat RIC3 exon 4 and shRNA against exons 2, 5 and 6 knocked down transfected rat RIC3 expression in SH-EP1 cells and simultaneously blocked toxin binding. However, no RNAi construct blocked binding when co-transfected with α7 into GH4C1 cells. shRNA against rat exons 2 and 5 knocked down rat RIC3 protein transfected into GH4C1 cells with a time course suggesting a protein half-life of a few days. These results suggest GH4C1 cells may possess unknown chaperone(s) allowing high surface α7 expression in the absence of known RIC3 splice variants.

Abbreviations used
αBGT

alpha-bungarotoxin

BSA

bovine serum albumin

cDNA

complementary DNA

eGFP

enhanced Green Fluorescent Protein

GAPDH

glyceraldehyde 3-phosphate dehydrogenase

HRP

horseradish peroxidase

IDT

Integrated DNA Technology

IgG

immunoglobulin G

mRNA

messenger RNA

nAChRs

nicotinic acetylcholine receptors

pRep4

Invitrogen episomal expression plasmid with a Roux sarcoma virus promoter and mammalian cell hygromycin resistance (pRep9 is similar, but confers mammalian neomycin resistance)

RIC3

resistance to inhibitors of cholinesterase 3

RIPA

Radioimmunoprecipitation Assay

RNAi

RNA interference including both siRNA and shRNA constructs

shRNA

short hairpin RNA

siRNA

small interfering RNA

TBS-0.2%T

Tris-buffered saline with 0.2% Triton X-100

Surface expression of α7 nicotinic acetylcholine receptors (nAChRs) is highly cell-type dependent in heterologous expression systems (Cooper and Millar 1997; Kassner and Berg 1997; Sweileh et al. 2000). As these receptors are implicated in schizophrenia and Alzheimer's disease (Leiser et al. 2009; Wallace and Porter 2011), as well as regulation of inflammation and immune function (Rosas-Ballina and Tracey 2009), considerable interest continues in expressing α7 receptors in cell lines for developing subtype-selective drugs via high throughput screening. RIC3 (Resistance to inhibitors of cholinesterase 3, Nguyen et al. 1995) largely overcame this cell-dependent difficulty when Treinin's laboratory showed that RIC3 discovered in C. elegans is a chaperone for α7 nAChRs (Halevi et al. 2002) allowing surface expression in non-permissive cells.

Two cell lines have a long history for studying heterologously expressed surface α7 receptors: GH4C1 cells, derived from rat pituitary (Osborne and Tashjian 1981), and SH-EP1 cells, derived from human neuroblastoma (Sonnenfeld and Ishii 1982). Neither cell line expresses endogenous mRNA for known nicotinic receptor subunits (Peng et al. 1999; Quik et al. 1996), simplifying the interpretation of transfection with receptor subunit genes. In some situations, SH-EP1 cells allow human α7 surface expression (Peng et al. 1999; Schroeder et al. 2003), but so low that we consider SH-EP1 to be relatively ‘non-permissive’. However, culturing SH-EP1 cells in cycloheximide or at low temperature (Schroeder et al. 2003), or co-transfecting human RIC3 with human α7 (Valles et al. 2009) significantly enhances surface receptor expression. In contrast, rat GH4C1 cells readily express either rat α7 (Cooper and Millar 1998; Quik et al. 1996) or chick α8 (Cooper and Millar 1998), which like α7, demonstrates significant cell-line dependence for expression. In a side-by-side comparison, GH4C1 cells expressed more than 10-fold higher surface receptor than SH-EP1 cells when the gene for rat α7 was introduced into both cell lines via adenovirus (Sweileh et al. 2000). However, both the ‘non-permissive’ SH-EP1 (Nehul Shah, unpublished result) and ‘permissive’ GH4C1 cells (Lansdell et al. 2005; Williams et al. 2005) express low levels of RIC3 mRNA. Therefore, we sought to test the hypothesis that GH4C1 cells express more RIC3 protein than SH-EP1 cells to explain this cell-type dependence for surface α7 receptor expression. As the predicted rat RIC3 gene had not been shown to be functional, we characterized the rat RIC3 gene. What follows is an investigation into cell-type specific differences in surface α7 nAChR expression measured by [125I]-α-bungarotoxin (αBGT) binding following α7 transfection with or without over-expression or knockdown of rat RIC3 protein.

Materials and methods

Chemicals and reagents

αBGT was purchased from Biotoxins Inc. (St. Cloud, FL, USA) and anti-human/mouse/rat RIC3 primary antibodies (sc-54143 and sc-134787) and horseradish peroxidase (HRP)-conjugated secondary antibodies were obtained from Santa Cruz Biotechnologies (Santa Cruz, CA, USA).

RIC3 clones

Primers for cloning full-length rat ric3 were based on sequence XM_001072861–(PREDICTED: Rattus norvegicus similar to resistance to inhibitors of cholinesterase 3 homolog (LOC687147)). The forward primer included a NheI site 5′ to the open reading frame with the sequence TTTTTAGCTAGCATGGCGTACTCCACAGTACA. The reverse primer included an XhoI site four bases past the stop codon (5′ to 3′, TTTTTTCTCGAGGCTTTCACTCAAAACCCTGG), synthesized by Integrated DNA Technologies (IDT, Coralville, IA, USA). Messenger RNA (Stratagene Absolutely mRNA kit, Agilent Technologies, Santa Clara, CA, USA) purified from flash-frozen rat hippocampus, cerebellum, and cortex was copied into complementary DNA (cDNA) using an Invitrogen Superscript II cDNA kit (Life Technologies, Grand Island, NY, USA). PCR amplicons were cloned into Invitrogen PCDNA3.1/V5-His TOPO TA plasmid, subcloned into Invitrogen pRep4 episomal plasmid using NheI and XhoI restrictions sites, and then sequenced. Human RIC3 isoformA was purchased from Origene (Catalog # TC112180; Rockville, MD, USA) and subcloned into Invitrogen pRep4 plasmid and sequenced.

Cell lines and culture conditions

HEK293 and GH4C1 were obtained from ATCC (Manassas, VA, USA), while SH-EP1 was a generous gift of Ron Lucas. Cells were cultured as previously described (Sweileh et al. 2000).

siRNA and shRNA

Small interfering RNA (siRNA) active against both human and rat RIC3 was designed by providing stretches of conserved DNA sequences to the siRNA design software of IDT. We found only one suitable siRNA sequence consisting of CAACAGAGUGGGACCUAAUGGUGAG (sense strand) complexed with UAGUUGUCUCACCCUGGAUUACCACUC (anti-sense), which was synthesized by IDT. Four commercial rat RIC3 short hairpin RNAs (shRNAs) (Locus ID = 687147) directed against exons 5 and 6 were purchased from Origene in a plasmid expressing red-fluorescent protein (pRFP-C-RS vector, catalog #TF708275), but we used shRNAs with the sequences GGCAAGTTCATTGACACATCTCCAGAGAA and CAGATGGCTACAGTGAGCAAGAGGAAGC against exons 5 and 6, respectively. A custom shRNA designed against exon 2 (sequence ATGGGCCAGATCATTCCAATCTATGGCTT, Origene) was produced in a green fluorescent protein plasmid (pGFP-B-RS). Scrambled versions of siRNA and shRNA were included as controls by the respective companies.

α7 and RIC3 transfections

We previously described (Lee et al. 2009) subcloning rat α7 and enhanced green fluorescent protein (eGFP) into pRep4 plasmids (Invitrogen). pRep4 is an episomal plasmid in mammalian cells that allows for antibiotic selection of transfected cells without requiring recombination into the mammalian genome. Cells were plated into 96-, 24-, or 6-well plates at seeding densities of 10 000 cells/well, 50 000 cells/well, and 200 000 cells/well, respectively, at least 3 h before transfection and 2–3 days prior to binding assays or protein extraction. Transfections were performed using 2 μL Fugene HD (Roche, Indianapolis, IN, USA) per 1 μg DNA. When transfecting multiple DNA constructs, the ratios of components were kept constant across groups, and empty pRep4 or pRep9 vectors substituted for missing components to maintain equivalent amounts of RIC3, α7, and RNA interference (RNAi) or reporter DNA constructs for each condition in a given experiment. Transfection efficiency was monitored by expression of either eGFP in pRep4, or in the case of shRNA constructs, fluorescent proteins expressed by the shRNA vector.

Dot and western blot analysis

Total protein was extracted from cells grown in 6-well or 24-well plates using radioimmunoprecipitation assay (RIPA) buffer (Tris 50 mM, NaCl 150 mM, sodium dodecyl sulfate 0.1%, Na Deoxycholate 0.25%, Triton X100 1%) supplemented with phenylmethylsulfonyl fluoride (1 mM) to inhibit protease activity. Cell lysates sat on ice for 30 min, and then centrifuged at 14 000 g for 10 min at 4°C. Supernatant was then aliquoted into microcentrifuge tubes and stored at −20°C until further use. The Pierce BCA assay (Thermo Fisher Scientific Inc., Rockford, IL, USA) was used to calibrate cell extract protein in mg equivalents of Bovine Serum Albumin (BSA), and then OD280 determined total protein levels prior to blotting on nitrocellulose membranes. After blotting 300 ng protein, the membrane was blocked for at least 2 h at 20°C with 5% BSA in TBS-0.2%T (TBS-0.2%T-5% BSA: 20 mM Tris HCl, 150 mM NaCl, 0.2% Triton X-100, 5% BSA), then incubated for 1 h with anti-RIC3 antibody diluted 1 : 400 in 5% BSA in TBS-0.2%T. After incubation with the HRP-conjugated secondary antibody diluted 1 : 1200 in TBS-0.2%T at 20°C for 1 h, dots were visualized using the SuperSignal West Pico Chemiluminescent Substrate Kit (Thermo Fisher Scientific Inc.). Luminescence was monitored using a Kodak In-Vivo Imaging System using exposure times ranging from 20 s to 2 min. For Western blots, protein extracts were prepared in loading buffer (Cell Signaling, Danvers, MA, USA) and separated on 4–20% acrylamide gels and transferred to nitrocellulose membranes (Thermo Fisher Scientific). After washing twice in water, membranes were blocked overnight in TBS-0.2%T/BSA (TBS with 0.2% Triton X-100, 5% BSA) at 20°C, then incubated with primary antibodies in TBS-0.2%T/BSA for 1 h at 20°C. A 1 : 300 dilution was used for RIC3 primary antibody, and 1 : 1000 was used for rabbit anti-glyceraldehyde 3-phosphate dehydrogenase (GAPDH) primary antibody (Cell signaling). The membranes were washed three times with TBS-0.2%T prior to incubation for 30 min with a 1 : 1250 dilution of HRP-linked donkey anti-goat immunoglobulin G (IgG) for RIC3 detection or a 1 : 3000 dilution of HRP-linked anti-rabbit IgG for GAPDH detection. Membranes were washed with TBS-0.2%T three times and once with TBS before being developed using SuperSignal West Pico chemiluminescent substrate (Thermo Fisher Scientific).

Surface receptor binding assay

Assays were performed as described in Lee et al. 2009. Briefly, quadruplicate cell samples plated in 24 or 96-multiwell dishes were washed with 200 or 100 μLphosphate-buffered saline , respectively, and incubated with 10 nM [125I]-αBGT in (200 or 100 μL, respectively) Hank's buffered saline plus 0.1% BSA for 3 h at 4°C before washing three times in Hank's buffered saline plus 0.1% BSA and then counting in a gamma counter. Non-specific binding was determined in the presence of 1 μM αBGT.

shRNA effects in cells expressing RIC3

GH4C1 and SH-EP1 cells were transfected with α7 and/or RIC3 as above. eGFP was included in control cultures to assess transfection efficiency. Three days following transfection, cells were exposed to 400 ug/mL hygromycin for 3 days, followed by treatment with 100 ug/mL hygromycin for 21 days, with media change every 2–3 days. Binding assays and dot-blot analysis were performed to quantify surface α7 expression and validate the presence of RIC3. SH-EP1 and GH4C1 cells expressing transfected RIC3 protein but lacking α7 were transfected with shRNA constructs or corresponding scrambled shRNAs together with the α7 plasmid. All shRNA plasmids encode fluorescent proteins to enable assessment of transfection efficiency. Two days following shRNA transfection, cells were exposed to 100 ug/mL hygromycin and either 10 ug/mL puromycin or blasticidin as indicated in the results. Binding assays and dot-blots were performed to investigate surface α7 expression and RIC3 protein levels.

Statistical analysis

Specific binding is calculated as the mean [125I]-αBGT binding in quadruplicate samples in the absence (total binding) minus the mean binding in quadruplicate samples in the presence of 1 μM αBGT (non-specific). The error is calculated as the square root of the sum of each of the standard deviations squared for total and non-specific binding, respectively. anova analysis of specific binding across all experimental conditions determined if any experiment showed significant binding variance, and post hoc t-tests determined the level of significance between selected pairs of conditions.

Results

PCR experiments suggest that SH-EP1 and GH4C1 cells express similar levels of RIC3 mRNA compared with the ‘housekeeping gene’ glyceraldehyde phosphate dehydrogenase (GAPDH, Figure S1). We therefore cloned rat RIC3 from rat hippocampus and cerebellum mRNA to further characterize the rat gene.

The two cloned isoforms of rat RIC3 correspond to accession numbers AM422212 and AM422213. AM422212 (S-rRIC3) encodes a 366 amino acid protein that differs from AM422213 (S+rRIC3) by three bases because of an ambiguous splice site between exons 4 and 5 resulting in the loss of a serine residue at position 173 relative to AM422213 (Fig. 1). Rat RIC3 shows greater than 85% homology with mouse and human. Although Roncarati and co-workers posted accession numbers AM422212 and AM422213 in 2007, no evidence published to date shows that these splice variants are functional in heterologous expression systems. Both rat RIC3 isoforms support rat α7 surface toxin binding in the HEK293 cell line (Figure S2). Two Santa Cruz Biotechnology anti-RIC3 antibodies (W-16 and H-282, catalog numbers sc-54143 and sc-134787, respectively) showed conformationally dependent binding to both rat RIC3 isoforms and full-length human RIC3 (Fig. 2). Boiling protein extracts from cells transfected with S-rRIC3, S+rRIC3, or human RIC3 prevented dot-blot labeling by both antibodies. Furthermore, although antibody to GAPDH shows significant staining on western blots, neither RIC3 antibody shows significant labeling on western blots under similar conditions, with H-282 antibody giving significantly higher background staining. Therefore, we used dot-blots of 0.3 μg cell protein samples to monitor RIC3 protein expression using W-16 antibody.

Figure 1.

Rat gene name for resistance to inhibitors of cholinesterase 3 (RIC3) open reading frame and translated protein. Underlined amino acids indicate transmembrane domains 1** and 2 and the coiled-coil domain, which are each labeled above the corresponding DNA. Underlined DNA indicates the shRNA and siRNA sequences used in this article. The six exons are labeled above the DNA, and inverted triangles indicate the splice boundaries between exons. The box indicates the serine residue and corresponding DNA at the ambiguous splice site between exons 4 and 5. **Transmembrane domain 1 in human or mouse RIC3 may or may not be cleaved from the mature protein as a signal sequence (Cheng et al. 2007; Castelan et al. 2008; Wang et al. 2009).

Figure 2.

Gene name for resistance to inhibitors of cholinesterase 3 (RIC3) labeling with commercial antibodies is blocked by thermal or chemical denaturation. Top Left: Protein lysates obtained from SH-EP1 cells ~72 h following transfection were boiled or not prior to dot blotting with Santa Cruz Biotechnology W-16 antibody diluted 1 : 400. Buffer is Radioimmunoprecipitation Assay (RIPA) buffer without cells blotted on the membrane, UT SH-EP1 signifies extract of untransfected cells, S+rRIC3 is extract from cells transfected with S+rat RIC3-pRep4, S-rRIC3 is extract from cells transfected with S-rat RIC3-pRep4, and hRIC3 is extract from cells transfected with human RIC3 in pRep4 plasmid. Bottom Left: Adding blocking peptide (0.67 μg/mL) prevents W-16 staining of RIC3 from transfected cells. Top Right: western blot of cell extracts from SH-EP1 cells transfected with either S- or S+ rRIC3 (these samples were boiled in loading buffer, but similar results were obtained when samples were not boiled). Inset: western blot of glyceraldehyde 3-phosphate dehydrogenase (GAPDH) to show protein loading. Bottom right: Samples labeled with Santa Cruz Biotechnology H282 antibody shows similar effects of boiling.

Untransfected SH-EP1 or GH4C1 cells did not show either [125I]-αBGT binding or RIC3 antibody staining to dot-blots of cell protein extract (Fig. 3). While α7 transfection alone increased toxin binding in GH4C1 cells, no binding was detected in SH-EP1 cells following α7 transfection alone. However, when either isoform of rat RIC3 was co-transfected with α7 into SH-EP1 cells, toxin binding increased significantly (Fig. 3), which correlated with a strong signal for rat RIC3 immunostaining. In GH4C1 cells, co-transfection of rat RIC3 with α7 did not significantly alter toxin binding compared with transfection with α7 transfection alone, as toxin binding occurred with or without RIC3 immunostaining.

Figure 3.

S+ and S-rRIC3 promote α7 receptor toxin binding in SH-EP1 cells but are not required for similar expression in GH4C1 cells. fMoles surface toxin binding per well are plotted above each corresponding dot-blot extracted from sister cultures. UT = untransfected. Transfection with the indicated components (plus empty vector when needed to keep DNA constant) occurred ~72 h prior to assays, performed as described in the methods. *p < 0.01, **p < 0.001 compared with control.

Co-transfection of siRNA designed against both human and rat RIC3 with RIC3 plasmid prevented RIC3 protein expression in both SH-EP1 and GH4C1 cells (Fig. 4). Transfection of SH-EP1 cells with α7-pRep4 alone showed little [125I]-αBGT binding, but α7 transfection with either human or S-rat RIC3 allowed significant binding. However, co-transfection with 30 nM siRNA blocked protein expression by either type of RIC3 in SH-EP1 cells. Dot-blot analysis failed to show the presence of human RIC3 in untransfected SH-EP1 cells, but transfection with either human or S-rat RIC3 showed significant staining (Fig. 4 left) that correlates well with the toxin binding. Furthermore, co-transfection with 30 nM anti-RIC3 siRNA blocked dot-blot staining for either human or S-rat RIC3 (in the presence of α7) (Fig. 4 left). These data are consistent with rat or human RIC3 protein expression conferring significant toxin binding to the otherwise non-permissive (or poorly permissive) SH-EP1 cells. However, permissive GH4C1 cells did not show RIC3 protein expression unless transfected with either human or rat RIC3. As expected, GH4C1 cells transfected with α7 showed significant [125I]-αBGT binding, but this was not much changed by either co-transfection with human or S-rRIC3 with or without RIC3 siRNA (Fig. 4 right). Dot-blot staining failed to show RIC3 protein expression in untransfected GH4C1 cells, but transfected human or S-rRIC3 showed clear expression that was blocked by co-transfection with siRNA (Fig. 4 right).

Figure 4.

Gene name for resistance to inhibitors of cholinesterase 3 (RIC3) knockdown by siRNA against exon 4 blocks surface α7 receptor toxin binding in SH-EP1 cells but has no effect on receptor expression in GH4C1 cells. In SH-EP1 cells (left), a decrease in RIC3 protein correlates with a reduction in binding. However, in GH4C1 cells (right), binding persists despite a decrease in RIC3 protein levels and in the absence of transfected RIC3 (α7). SH-EP1 and GH4C1 cells were transfected with the indicated components (plus empty vector when needed to keep DNA constant) ~72 h prior to binding assay and dot-blotting. *p < 0.01, **p < 0.001 comparing conditions as shown.

Co-transfection of RIC3 plasmid with RIC3 siRNA is expected to prevent protein expression. However, siRNA can knock down expression of pre-existing RIC3 protein only if the siRNA acts for a period several times longer than the half-life of the protein in a cell. To estimate the extent of this problem, we transfected SH-EP1 cells with S-rRIC3 in pRep4 plasmid, and selected for transfected cells by treating the cultures for 24 days with 400 ug/mL hygromycin. pRep4 is an epsisomal vector expressing hygromycin resistance that replicates during cell division, while the siRNA gets diluted during each cell division in the rapidly dividing SH-EP1 cells. Further transfection of these cells with α7 and RIC3 siRNA only partially blocked toxin binding after 3 days relative to (i) transfection with either α7 and scrambled siRNA into SH-EP1 cells similarly transfected with S-rRIC3 or (ii) α7 and S-rRIC3 transfected into untransfected SH-EP1 cells (Figure S3). Therefore, to knock down pre-existing RIC3 protein, we used a short hairpin vector that expresses both the RNAi of interest and a fluorescent protein in a plasmid that allows for antibiotic selection. Origene provided four shRNAs directed against exons 5 and 6, and the construct against exon 5 (Fig. 1) showed the largest difference (6 bases) from human RIC3. SH-EP1 cells were transfected with human or S-rRIC3 in the presence or absence of rat RIC3 shRNA anti-exon 5 or scrambled shRNA. Rat RIC3 shRNA against exon 5 reduced S-rRIC3 protein while decreasing toxin binding to SH-EP1 cells co-transfected with α7 and S-rRIC3, but had no effect on cells transfected with human RIC3 (Fig. 5 left, binding assay). In the same cells, anti-exon5 shRNA blocked S-rRIC3 protein expression but had no effect on human RIC3 protein expression (Fig. 5 left dot-blots), suggesting that the shRNA action is not because of off-target effects. An Origene shRNA construct directed against exon 6 had similar effects to the construct against exon 5, (Figure S4).

Figure 5.

Surface α7 toxin binding persists in GH4C1 cells following gene name for resistance to inhibitors of cholinesterase 3 (RIC3) knockdown by shRNA against exon 5. Left: In SH-EP1 cells, toxin binding is abolished following RIC3 knockdown. SH-EP1 cells were transfected with α7 and human or S-rRIC3 in the presence or absence of rat RIC3 shRNA against rat exon 5 or scrambled shRNA after selection with puromycin. Top left shows the fMoles specific toxin binding, with dot-blot results showing immunostaining below the x axis. shRNA against exon 5 abolished both toxin binding and ric3 immunoreactivity, but scrambled shRNA (scrRNA) had no effect. However, shRNA had no effect on either toxin binding or immunoreactivity induced by human RIC3. Furthermore, adding blocking peptide to the dot-blots blocked all immunostaining. Right: GH4C1 wild-type cells and GH4C1 cells previously transfected with S-rRIC3 (S-rRIC3) were simultaneously transfected with α7 and either shRNA against exon 5 or scrambled shRNA. After 8 days treatment with 10 μg/mL puromycin for the cells transfected with shRNA against exon 5 or with scrambled shRNA, transfected cells were plated in 96-well plates at 10 000 cells/well. The GH4C1 cells transfected with shRNA against exon 5 show no RIC3 immunostaining, but do show toxin binding, while scrambled shRNA had no effect in either assay. *p < 0.01 and **p < 0.001 compared with control or as indicated.

GH4C1 cells previously transfected with α7 and S-rRIC3 and selected with hygromycin were again transfected with shRNA against exon 5 and then treated with puromycin (10 μg/mL to remove untransfected cells) on day 2 for 8 days for a total of 10 days post-ransfection (Fig. 5 right). Antibody staining is lost by this treatment, suggesting that turnover of S-rRIC3 protein is complete by 10 days (Fig. 5 right) but binding is only partially blocked by siRNA at 3 days in SH-EP1 cells (Figure S3). In contrast, toxin binding to GH4C1 cells transfected with α7 is unaffected by treatment with shRNA against exon 5 and puromycin (Fig. 5 right). Scrambled shRNA had no effect on either toxin binding or antibody staining in GH4C1 cells previously transfected with S-rRIC3 (Fig. 5 right).

One issue regarding knockdown of the known isoforms of rat RIC3, is that unknown splice variants of RIC3 could exist in GH4C1 cells that are not recognized by either of the commercially available antibodies with specificity for protein sequence encoded by exons 5 or 6. However, functional RIC3 splice variants generally include the second RIC3 transmembrane domain encoded within exon 2 (unfortunately, exons 1 and 3 do not have sequences amenable to forming shRNAs). With the help of Origene technical support, we designed a custom shRNA against exon 2 of rat RIC3. Knockdown of rat RIC3 immunostaining with shRNA against rat RIC3 exon 2 completely blocked toxin binding in SH-EP1 cells, but had no effect on binding in GH4C1 cells that had been transfected with S-rRIC3 (Fig. 6). shRNA constructs against either exons 2 or 5 also knocked down expression of S+rRIC3 protein (Figure S5).

Figure 6.

Transfection with shRNA against rat exon 2 decreases both surface α7 receptor toxin binding and rRIC3 protein levels in SH-EP1 cells, but not toxin binding in GH4C1 cells previously transfected with S-rRIC3. Left: Rat gene name for resistance to inhibitors of cholinesterase 3 (RIC3) shRNA targeting exon 2 concomitantly decreases levels of rat RIC3 protein levels and toxin binding in transiently transfected SH-EP1 cells. In contrast, scrambled shRNA has no effect against rat RIC3 promoting toxin binding, nor does shRNA against rat exon 2 have any effect on human RIC3 enhancement of toxin binding. Right: GH4C1 cells transfected with α7 and shRNA for 10 days in the presence of 10 μg/mL blasticidin continue to produce surface toxin binding sites despite knockdown of previously transfected rat RIC3 protein (S-rRIC3), while scrambled shRNA had no effect on immunostaining. **p < 0.001 compared with control or as indicated.

Discussion

In this study, we investigated the consequences of RIC3 over-expression and knockdown on surface α7 nAChR expression detected by 125I-αBGT binding in permissive (GH4C1) and non-permissive (SH-EP1) cell lines. Nguyen et al. discovered RIC3 by screening for C. elegans mutants that could survive exposure to the anti-cholinesterase aldicarb (Nguyen et al. 1995; Miller et al. 1996). Halevi et al. (2002) subsequently demonstrated that RIC3 is necessary for surface expression of multiple nAChR subtypes in C. elegans and enhances expression of homomeric rat α7 nAChR in frog oocytes. RIC3 also regulates heteromeric nAChR expression and significantly modulates 5HT3A receptor expression (Castillo et al. 2005, 2006; Cheng et al. 2005; Lansdell et al. 2005; Walstab et al. 2010a, b), but so far has no reported effect on surface expression of ionotropic glutamate or GABA receptors. RIC3 is a resident protein in the endoplasmic reticulum (Cheng et al. 2007), including endoplasmic reticulum in neuronal dendrites (Alexander et al. 2010), where it plays a key role regulating the folding, assembling, and/or trafficking of a variety of multimeric receptors (Millar 2008; Treinin 2008).

Surface expression of α7 receptors is highly cell-type dependent in heterologous cell expression systems (Cooper and Millar 1997; Kassner and Berg 1997; Sweileh et al. 2000). HEK cells do not allow significant α7 surface expression measured by 125I-αBGT binding without RIC3 transfection (Williams et al. 2005) (but see however, Gopalakrishnan et al. 1995). In some situations, human SH-EP1 cells allow low α7 surface expression (Peng et al. 1999; Sweileh et al. 2000; Schroeder et al. 2003), but so low that we consider SH-EP1 to be relatively ‘non-permissive’. In contrast, culturing SH-EP1 in cycloheximide or at low temperature (Schroeder et al. 2003), or transfecting with RIC3 (Valles et al. 2009) significantly enhances α7 expression, but usually not as much as that of “permissive” rat GH4C1 cells (Sweileh et al. 2000).

Both the ‘non-permissive’ SH-EP1 (Nehul Shah, unpublished result, and Figure S1) and ‘permissive’ GH4C1 cells (Lansdell et al. 2005; Williams et al. 2005) express low levels of RIC3 mRNA (Figure S1), but we find that neither cell line shows significant levels of RIC3 immunoreactivity with Santa Cruz primary antibodies that recognize human and rat RIC3 proteins, unless the cells are transfected with RIC3 DNA to increase expression. The Santa Cruz antibodies against RIC3 recognize undisclosed epitopes located approximately in exons 5 and 6 of rat RIC3 (Santa Cruz Biotechnology Technical Support) and our results show that these epitopes are disturbed by thermal or chemical denaturation. The antibodies do show staining in extracts from cells transfected with exogenous RIC3, and W16 staining is blocked by a proprietary blocking peptide. However, the existence of multiple human RIC3 splice variants (Halevi et al. 2003; Seredenina et al. 2008) and 11 splice variants in drosophila (Lansdell et al. 2008) raises the possibility that these commercially available antibodies may not recognize certain functional rat RIC3 splice variants. To evaluate which exons in rat RIC3 could be allowing α7 surface expression in GH4C1 cells, we tested the effects of knocking down every known rat RIC3 exon (exons 2, 4, 5, and 6) for which we could design reagents.

RIC3 protein has four major conserved structural features (Halevi et al. 2003): Two transmembrane domains [the first of which may or may not be a signal sequence, see (Castelan et al. 2008) versus (Cheng et al. 2007; Wang et al. 2009)], the linking sequence between the transmembrane domains, and at least one coiled-coil domain. Rat RIC3 protein shares 94% sequence identity with mouse RIC3 and has all the conserved domains. Human RIC3 mRNA shows identical exon splicing to mouse RIC3, but human RIC3 protein shares less identity (87%) with rat RIC3 with two gaps in the amino acid sequence. Interestingly, all three species retain the ambiguous splice site between exons 4 and 5 leading to S+ and S-forms of RIC3 protein. Roncarati and co-workers previously submitted S+ and S-rRIC3 sequences (Accession numbers AM422212 and AM422213, respectively), but this is the first demonstration that both isoforms functionally increase αBGT binding in cell lines.

RIC3 mutants with deleted regions have failed to show a unified mechanism of action for the molecule in promoting nAChR surface expression, and recent data suggests that the C. elegans molecule is intrinsically disordered (i.e. having no ‘well-defined tertiary structure in the native, functional state’- http://iupred.enzim.hu/) but folds into different conformations when interacting with the different receptors that it modulates (Ben-Ami et al. 2009). Earlier studies found that the minimal functional regions allowing Deg2/Des3 nicotinic receptor assembly in C. elegans consists of the two worm RIC3 transmembrane domains and the intervening sequence, but the coiled-coiled domain is not required (Ben-Ami et al. 2005). However, other C. elegans nAChR subtypes require the coiled-coil domain for efficient assembly (Ben-Ami et al. 2009; Biala et al. 2009). In contrast, enhanced human α7 expression in frog oocytes requires some additional human RIC3 C-terminal regions beyond the transmembrane domains, but not the coiled-coil domain (Castillo et al. 2005; Castelan et al. 2008). Drosophila RIC3 has 11 exons (Lansdell et al. 2008), and the coiled-coil domain is not required for enhanced human α7 surface expression, although in drosophila exon 2, a small part of the sequence between the transmembrane domains must be absent for efficient human α7 expression. The homologous rat RIC3 region of two transmembrane domains and the intervening sequence consists of exons 1 and 2. As the sequence and length of rat exon 1 does not lead to interfering RNA constructs, we investigated knocking down rat RIC3 with a short hairpin RNA designed against exon 2. Although shRNA against rat RIC3 exon 2 prevented surface α7 expression in SH-EP1 cells transfected with both α7 and rat RIC3, the same shRNA had no effect on α7 expression in GH4C1 cells. The coiled-coil domain straddles exons 3 and 4 in rat RIC3, similar to both human and mouse RIC3. siRNA against exon 4 (no possible RNAi constructs against exon 3 were found) similarly blocked surface α7 expression in SH-EP1 cells transfected with both α7 and rat RIC3, but had no effect on GH4C1 cell surface α7 expression. Finally, shRNA against either exons 5 or 6 blocked surface α7 expression in SH-EP1 cells transfected with both α7 and rat RIC3, but had no effect on α7 expression in GH4C1 cells. These data suggest that the ‘non-permissive’ SH-EP1 cells require the presence of RIC3 protein for enhanced surface expression of α7 receptors, while the ‘permissive’ GH4C1 cells do not require the presence of any known splice variant of rat RIC3 protein.

One explanation of the apparent difference between cell lines could be differences in cell line susceptibility to RNAi methods. However, GH4C1 cells transfected with rat RIC3 to allow detection by immunoblotting lost immunostaining when transfected for 10 days with shRNA against exons 2 or 5, but not if transfected with scrambled shRNA, indicating that the shRNA works similarly in both GH4C1 and SH-EP1 cells. Ten days were the minimum time to allow for sufficient expression of antibiotic selection against cells not transfected with shRNA. Besides showing that RNAi methods are effective in GH4C1 cells, these experiments also provide a rough measure of rat RIC3 protein stability in cells, and suggest an approximate T1/2 of a few days at most. Some other caveats of these experiments include that we do not know the absolute sensitivity of RIC3 detection with our immunoblotting method, that gene knockdown by RNAi methods is never complete, and the dose-response curve for RIC3 effects on α7 receptor surface expression may not be monotonic. A further issue is that we only detect successfully assembled surface α7 receptor by binding but cannot determine the ratio between unassembled receptor subunits and RIC3 protein. It could be that the ratio is more favorable for receptor assembly in GH4C1 than other cell lines even after shRNA knockdown of endogenous Ric3. Nevertheless, these results demonstrate a cell-dependent requirement for RIC3 enhancement of surface α7 receptor expression. Lansdell et al. previously demonstrated a host-cell dependency for drosophila versus human RIC3 in mammalian and insect cell lines: Human RIC3 promoted surface human α7 expression better in mammalian than in insect cells, while drosophila RIC3 showed reverse activity (Lansdell et al. 2008).

It is also possible that RIC3 interacts with other proteins in ‘permissive’ cells so that currently undetectable amounts of RIC3 could contribute to α7 receptor expression. For instance, RIC3 is one out of three chaperones simultaneously required to express C. elegans levamisole-sensitive nAChRs in ooctyes (Boulin et al. 2008). By itself, RIC3 has only a slight effect, but leads to robust levamisole-sensitive nAChR expression when combined with the worm chaperones Unc-74 (Lewis et al. 1980) and Unc-50 (Lewis et al. 1980; Eimer et al. 2007), the latter of which has the mammalian homolog UNCL (Fitzgerald et al. 2000). Conversely, inhibitory proteins may prevent RIC3 actions on α7 receptor expression in ‘non-permissive’ cells, similar to the actions of BATH-42 protein preventing RIC3 actions on levamisole-sensitive nAChRs expression in C. elegans (Shteingauz et al. 2009).

In summary, even though SH-EP1 and GH4C1 cells show similar low levels of RIC3 mRNA expression, and neither cell line shows significant RIC3 protein expression detectable by commercially available antibodies, we find dramatically different cell-specific requirements for RIC3 transfection that allow surface α7 receptor expression. We expected that knocking down RIC3 expression would decrease ‘permissive’ GH4C1 cell surface expression of α7 receptors. Instead, we failed to find any evidence of RIC3 involvement in α7 expression in these cells. In contrast, expression of detectible levels of either human or rat RIC3 protein was obligatory for robust α7 surface expression in the ‘non-permissive’ HEK and SH-EP1 cell lines, and SH-EP1 cells rapidly lost surface expression if RIC3 protein was knocked down. The simplest, but not exclusive explanation is that α7 surface expression in GH4C1 is either dependent on an as yet undiscovered splice variant of rat RIC3 not detectable with our current methods or is dependent on some other chaperone(s) present in GH4C1 cells, but lacking in SH-EP1. Therefore, GH4C1 cells may be a reasonable source of mRNA for cloning additional α7 chaperones. An extrapolation from these results is that a mouse RIC3 knockout may only show a partial phenotype, with some brain regions retaining [125I]-αBGT binding to surface α7 receptors, e.g. in mouse hippocampus dentate gyrus, which shows high levels of [125I]-αBGT binding in the absence of significant RIC3 mRNA expression (Halevi et al. 2003).

Acknowledgements

We thank Sean Sherman, Emily Thomas, Nehul Shah, and Gaurav Gogri for preliminary results that contributed to this project, Ron Lucas for the generous gift of SH-EP1 cells, Rachel Rodgers for statistical consultations, Kyu Won Kim for help with PCR, and Geeta Gwalani for checking rat RIC3 exon structure against DNA sequence data for rat chromosome one. The authors declare no conflicts of interest. Tom Koperniak is a recipient of an American Foundation for Pharmaceutical Education pre-doctoral award.

Ancillary