SEARCH

SEARCH BY CITATION

Keywords:

  • α4β2 nicotinic receptor;
  • α7 nicotinic receptor;
  • degradation;
  • localization;
  • RIC-3 expression;
  • transmembrane segment

Abstract

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

The RIC-3 protein acts as a regulator of acetylcholine nicotinic receptor (nAChR) expression. In Xenopus laevis oocytes the human RIC-3 (hRIC-3) protein enhances expression of α7 receptors and abolishes expression of α4β2 receptors. In vitro translation of hRIC-3 evidenced its membrane insertion but not the role as signal peptide of its first transmembrane domain (TMD). When the TMDs of hRIC-3 were substituted, its effects on nAChR expression were attenuated. A certain linker length between the TMDs was also needed for α7 expression enhancement but not for α4β2 inhibition. A combination of increased α7 receptor steady state levels, facilitated transport and reduced receptor internalization appears to be responsible for the increase in α7 membrane expression induced by hRIC-3. Antibodies against hRIC-3 showed its expression in SH-SY5Y and PC12 cells and its induction upon differentiation. Immunohistochemistry demonstrated the presence of RIC-3 in rat brain localized, in general, in places where α7 nAChRs were found.

Abbreviations used
α1GlyR

α1 glycine receptor

α-Bgt

α-bungarotoxin

ACh

acetylcholine

COS cells

Cercopithecus aethiops cells

EGFP

enhanced green fluorescent protein

EGFR

epidermal growth factor receptor

hRIC-3

human RIC-3

nAChR

nicotinic acetylcholine receptor

PBS

phosphate-buffered saline

PFA

p-formaldehyde

SDS-PAGE

sodium dodecyl sulphate–polyacrylamide gel electrophoresis

TMD

transmembrane domain

Nicotinic acetylcholine receptors (nAChR) constitute a gene family of membrane proteins that mediate fast synaptic transmission in nerve and muscle cells (reviewed by Lindstrom 2003; Sine and Engel 2006). They share the same oligomeric structure, being composed of five homologous subunits enclosing the ion pore. Assembly and trafficking of nAChRs is a complex process in which specific proteins are expected to be involved. The RIC-3 protein could be one of them, as it shows differential effects when co-expressed with several ligand-gated receptors (Halevi et al. 2003; Ben-Ami et al. 2005; Castillo et al. 2005; Cheng et al. 2005; Lansdell et al. 2005; Williams et al. 2005). Thus, in Xenopus leavis oocytes, human RIC-3 (hRIC-3) increases surface expression of α7 nAChRs by enhancing its inefficient transport to the cell membrane, whereas it acts as a barrier for other nAChR subtypes (α4β2 and α3β4) preventing its transport (Halevi et al. 2003; Castillo et al. 2005). Previously, we suggested that the first transmembrane domain (TMD) of hRIC-3 could act as a signal peptide, according to a prediction program. Here, we show that this does not seem to be the case, at least ‘in vitro’. We also have shown that the N-terminal region of hRIC-3, which contains two TMDs and a linker of 70 amino acids is needed for its action (Castillo et al. 2005). Here, we present a more detailed study on the functional relevance of these domains. In addition, we demonstrate that the presence of hRIC-3 strongly increases the incorporation rate of α7 nAChRs to the oocyte surface and decreases the rate of receptor internalization. Finally, by using antibodies against hRIC-3, we show its expression in neuronal cell lines, its induction by differentiation stimuli and its localization in the rat CNS.

Materials and methods

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Plasmid constructions

To generate the RIC-3 deletion mutants we annealed single-stranded oligonucleotides with the desired sequences and proper single strand ends which could be easily ligated to the ends generated by restriction enzymes either present in the original cDNA sequence or introduced by PCR. A similar strategy was used to substitute the first and/or second TMDs of RIC-3 either by the bovine α7 signal peptide or the epidermal growth factor receptor (EGFR) TMD. Thus, in construct α7/RIC-3, amino acids 1–29 of RIC-3 corresponding to its first TMD, MAYSTVQRVALASGLVLALSLLLPKAFL, were replaced by the sequence MRGSLCLALAASILHVSLQ that corresponds to the α7 signal peptide. In construct EGFR1/RIC-3, amino acids 3–26 of RIC-3 (YSTVQRVAL ASGLVLALSLLLPKA) were replaced by the sequence IATGMVG ALLLLLVVALGIGLFMR of the EGFR. In construct EGFR2/RIC-3, amino acids 94–114 of RIC-3 (GLMGQIIPIYGFGIFLYILYI), were replaced by the sequence IATGMVGALLLLLVVALGIG of the EGFR. Finally, in construct RIC-3 (α1 glycine receptor; α1GlyR) amino acids 37–90 were substituted by an artificial link made of the sequence RASLPKVSYVRASLPKVSYVK that corresponds to a duplication of the sequence linking the M2 and M3 TMDs of the α1GlyR, a protein that we have previously shown is not affected by RIC-3 (Halevi et al. 2003). DNAs of human neuronal nAChRs subunits (α7, α4 and β2) and hRIC-3 and its mutants were inserted into the pSP64T vector (Krieg and Melton 1984) or a derivative thereof. Capped mRNA was synthesized in vitro using SP6 RNA polymerase and the mMESSAGE mMACHINE kit (Applied Biosystems, Madrid, Spain).

Protein translation ‘in vitro

The RIC-3 mRNA was translated with the TNT lysate system (Promega Biotech Iberica, Barcelona, Spain) and canine pancreatic microsomal membranes (Promega) were used to study membrane insertion and peptide cleavage, as described by the manufacturer. The membrane fraction was recovered by centrifugation for 20 min at 108 000 g through a 10% sucrose solution and analyzed by sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS–PAGE).

Oocyte expression

Defoliculated Xenopus leavis oocytes were injected with 10 ng of total subunit cRNA and 5 ng of hRIC-3 cRNA in 50 nL of sterile water. Lower amounts (1 ng and less) of hRIC-3 RNA resulted in the attenuation of the RIC-3 effects, i.e. reduction of, both, α7 expression increase and α4β2 inhibition. All experiments were performed within 2–3 days after cRNA injection. Wild-type α4, β2 and α7 mRNAs were injected into oocytes from the same frog every time a hRIC-3 mutant was tested. Consequently, the effects of mutants on receptor expression were expressed as a percentage of expression observed in the same experiment with only receptor subunits.

[125I]-α-bungarotoxin binding assays

Specific surface expression of 125I-α-bungarotoxin (α-Bgt) binding sites was tested with 10 nmol/L [125I]-α-Bgt as described (Garcia-Guzman et al. 1994). Briefly, oocytes were incubated with 10 nmol/L [125I]-α-Bgt for 2 h at 18°C. At the end of the incubation, unbound [125I]-α-Bgt was removed, oocytes were washed and bound radioactivity was counted. Non-specific binding was determined using non-inoculated oocytes.

Rates of α7 nAChR surface expression and turnover

Delivery of newly assembled receptors to the cell membrane was assessed by [125I]-α-Bgt binding. For this purpose, nAChRs already present at the oocyte membrane were blocked with 100 nmol/L unlabelled α-Bgt and then the usual binding assay was carried out at designated time points. nAChR turnover was deduced from 125I-α-Bgt degradation and release into oocyte-containing medium as described (Devreotes and Fambroug 1975; Christianson and Green 2004). Oocytes expressing nAChRs were labelled with 125I-α-Bgt, washed as usual and further incubated in 24-well plates at 18°C. At designated times, media were collected and replaced. At the end of the experiment, collected samples and oocytes were counted and turnover expressed as a fraction of 125I-α-Bgt bound at the beginning of the experiment.

Electrophysiological recordings

Electrophysiological recordings were carried out as previously described (Garcia-Guzman et al. 1994). Functional expression of each construct was estimated as the peak ionic current evoked by 1 s application of 1 mmol/L ACh at −80 mV, and no correction for desensitization was made. About 7–10 oocytes were tested every time a mutant was studied. All experiments were performed at room temperature (22°C).

Antibodies

A polyclonal antibody against hRIC-3 was generated by using as immunogen a protein containing amino acids 160–368 of hRIC-3 fused to glutathione S-transferase. The appropriate construct was prepared by inserting the corresponding DNA sequence into pGEX-KG expression vector (Guan and Dixon, 1991). After bacterial expression and purification the fusion protein was used to immunize hens for antibody production in their eggs. IgY was purified by a modification of the EGGstract® IgY Purification System (Promega). This antibody was shown to recognize hRIC-3 by inmunofluorescence and in western blots of Cercopithecus aethiops (COS) cells transfected with the hRIC-3 DNA. A stock solution of antibody (2 mg protein/mL) was diluted as indicated in the following procedures.

Western blot

Cells in culture or oocytes were homogenized in Lysis buffer consisting of 20 mmol/L sodium phosphate, pH 7.5, 0.1 mmol/L EDTA, 1 mmol/L phenylmethylsulphonyl fluoride, 1 mmol/L dithiothreitol and 0.5% Triton X-100. Detergent extracts were cleared by centrifugation (15 min at 13 000 g and 4°C). Proteins were separated by 10% SDS–PAGE. After the transfer, nitrocellulose membranes (Schleicher & Schuell, Dassel, Germany) were blocked 30 min at 20°C with 5% dry milk in phosphate-buffered saline (PBS)–0.2% Tween 20, and incubated with 1 : 100 dilutions in PBS–0.2% Tween 20 and 1% dry milk for 1 h of the following antibodies: anti-enhanced green fluorescent protein (EGFP) (Living Colors® A.v.Peptide Antibody; Clontech, Madrid, Spain), anti-Actin [Actin (H-300); Santa Cruz Biotechnology Inc., Santa Cruz, CA, USA] and anti-RIC-3. After incubation with the secondary antibody at 22°C for 1 h, detection was carried out with the SuperSignal®West Pico Trial Kit (Pierce, Bonn, Germany) in a Luminescent Image Analyzer LAS-1000 Plus (Fujifilm, Duesseldorf, Germany).

Cell culture, transfection and differentiation

COS-1 cells were grown in Dulbecco’s modified Eagle’s medium supplemented with 10% foetal calf serum and antibiotics. Expression plasmids for fusion proteins of the EGFP and hRIC-3 were transfected with Lipofectamine™ transfection reagent (Invitrogen, Barcelona, Spain) into COS-1 cells according to the manufacturer’s instructions. Forty-eight hours post-transfection, transfected cells were lysed for western blot analysis.

SH-SY5Y neuroblastoma cells were grown in minimum essential medium + GLUTAMAX medium (Gibco, Barcelona, Spain) supplemented with 10% foetal calf serum and antibiotics. To induce differentiation all-trans-retinoic acid (SERVA Electrophoresis, Heidelberg, Germany) was added to the medium the day after plating at a final concentration of 1 μmol/L. Cells were always differentiated by treatment for 7 days with all-trans-retinoic acid before experiments (Encinas et al. 2000).

PC12 cells were grown in Dulbecco’s modified Eagle’s medium supplemented with 10% heat-inactivated horse serum and 5% heat-inactivated foetal calf serum and antibiotics. Cells were maintained at 37°C in a saturated humidity atmosphere containing 95% air and 5% CO2. Cells were treated with 50 ng/mL human nerve growth factor-beta (Sigma-Aldrich, Madrid, Spain) for 7 days before experiments (Avila et al. 2003).

Confocal microscopy studies of RIC-3 expression in SH-SY5Y cells

Cells were fixed with 4%p-formaldehyde (PFA) in PBS during 20 min. Then cells were permeabilized with 0.2% Triton X-100 in 3.6% PFA during 10 min and washed twice with 1% bovine serum albumin in PBS during 10 min. Labelling of RIC-3 in permeabilized cells was performed by overnight incubations with 1 : 100 dilutions of polyclonal anti-RIC-3 antibody in PBS. After extensive washes, secondary rabbit anti-chicken antibodies coupled to FITC (1 : 200 dilution; Sigma-Aldrich) were incubated during 2 h and rinsed with PBS. Observation of RIC-3 was performed with an Olympus Fluoview FV300 confocal laser system (Olympus Corp., Tokyo, Japan) mounted in a BX-50 WI up-right microscope incorporating a 100× LUMPlan FI water-immersion objective (Olympus Corp.). This system allows for z-axis reconstruction with theoretical z slice of about 0.5-μm thick. Analysis of fluorescent images was performed with the public domain program ImageJ[http://rsb.info.nih.gov/ij] with Plugins for region of interest measurement, image average and comparison of multiple channel images.

Immunohistochemistry

Immunohistochemistry was performed as described previously (Domínguez del Toro et al. 1994; Ortiz et al. 2005). Briefly, a total of five adult Wistar rats were used in this study, being housed under normal light–dark conditions (12 : 12) and with free access to food and water. Animals were deeply anaesthetized with chloral hydrate and perfused through the left ventricle with 0.1 mol/L phosphate buffer (pH 7.35) followed by freshly prepared 4% PFA in the same buffer. Brains were removed, post-fixed in the same fixative for 6–8 h and transferred to a 30% sucrose solution in PBS for at least 3 days. Coronal sections (35 μm) were obtained in a cryostat. They were pre-incubated in 5% normal goat serum at 4°C for 1 h, transferred to the anti-RIC-3 antibody diluted 1 : 1000 in 0.1% Triton X-100 in PBS and incubated at 4°C for 16 h. This was followed by PBS washes and incubation in a biotinylated goat anti-chicken antibody (Vectastain, ABC kit; Vector Labs Inc., Burlingame, CA, USA) in 0.1% Triton X-100 in PBS at 4°C for 1 h. The tissue was then washed in PBS and incubated in an avidin–biotinylated peroxidase complex according to the manufacturer’s instructions (Vectastain, ABC kit). Finally the sections were washed in PBS and visualized with 0.05% 3,3′-diaminobenzidine in PBS and 0.009% H2O2. Staining specificity was tested by omission of the anti-RIC-3 antibody during the corresponding incubation, or by using a pre-immune fraction. In these cases, no labelling was observed.

Results

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

In vitro translation of hRIC-3

Previous analysis predicted that the first putative TMD of hRIC-3 (see Fig. 1a) could act as a signal peptide (Castillo et al. 2005). To confirm it we carried out in vitro translation of hRIC-3 and studied its processing in the presence of microsomal membranes. When analyzed by SDS–PAGE, the in vitro translated hRIC-3 protein migrated as a single band with mobility slightly lower than the expected from its molecular weight (Fig. 1b, lane 1). Further recovery of hRIC-3 in the membrane fraction suggested that hRIC-3 was indeed inserted in the membrane (Fig. 1b, lane 2). However, we detected a single band with the same mobility than the one obtained in the absence of microsomal membranes, indicating that no signal peptide cleavage had occurred (Fig. 1b lane 3). By contrast, when the first TMD of hRIC-3 was substituted by the signal peptide of the bovine nAChR α7 subunit (see Fig. 1a, chimaeric protein α7/RIC-3), an additional band of smaller size was detected (Fig. 1b, lane 4, arrowhead) presumably because of the cleavage of the α7 signal peptide. As the first TMD of hRIC-3 is larger than the α7 signal peptide and cleavage of the latter was detected as a slight shift in band size, we should have been able to detect cleavage of the first TMD of RIC-3, if this had occurred. This was not the case and we conclude that, at least in vitro, the hRIC-3 protein keeps its first TMD during the translation process.

image

Figure 1. In vitro translation of hRIC-3 in the presence of microsomal membranes. (a) Amino acid sequences of the N-terminal region of hRIC-3, bovine α7 nAChR subunit and the chimaeric protein (α7/RIC-3) in which the first TMD of hRIC-3 was substituted by the signal peptide of the bovine α7 subunit (both in bold). (b) Autoradiography of 35S-labelled hRIC-3 protein. One microlitre of total reaction volume (10 μL) was directly dissolved in sample buffer without further processing and analyzed in lane 1. The hRIC-3 protein synthesized in the presence of microsomal membranes was recovered in the membrane fraction, as shown in lane 2 (arrow). Band intensity in this lane is lower than in lane 1 because the amount of reaction volume applied represents only 1/10 of the one applied in lane 1. A comparison of the co-translational processing of hRIC-3 and the chimaeric protein α7/RIC-3 is shown in lanes 3 and 4. Signal peptide cleavage of α7/RIC-3 (lane 4) but not of RIC-3 (lane 3) is evidenced by the appearance of a band of lower size (arrowhead). Notice that we chose an amount of microsomal membranes that would allow the processing of just a fraction of the α7/RIC-3 protein, so that the difference between the processed and unprocessed protein could be observed. In the case of the native hRIC-3 protein, even an increase in the amount of microsomal membranes did not yield any signal peptide cleavage. Molecular weight markers (in kDa) are indicated at the left.

Download figure to PowerPoint

Elements at the N-terminal region of hRIC-3 that are required for its function

The most relevant structural elements present at the N-terminal region of hRIC-3 are the two TMDs and the loop in between (amino acids 1–116) as well as a coiled-coil domain (amino acids 137–169). As it has been shown that the coiled-coil domain is not needed for RIC-3 function whereas the two TMDs and the loop are required (Ben-Ami et al. 2005; Castillo et al. 2005), we asked about the relevance of the latter elements. For this purpose, we studied the effects of several hRIC-3 constructs on the expression of α4β2 and α7 nAChRs (Fig. 2).

image

Figure 2.  Effects of different hRIC-3 constructs on nAChRs. Expression of nAChRs in the presence of the indicated hRIC-3 constructs was tested by measuring whole cell currents (open boxes) elicited by application of 1 mmol/L ACh in oocytes expressing human α4β2 and α7 receptors. Expression was also measured by α-Bgt binding to oocytes expressing α7 receptors (closed boxes, notice the different scale at the bottom). Data represent mean ± SD of 30–50 oocytes (n = 2–4 frogs) and are expressed as percentage of controls obtained the same day and in the same conditions but without co-expression of the hRIC-3 constructs. Asterisks represent statistically significant differences with respect to hRIC-3 (p < 0.05), that were calculated by one-way anova test and by Bonferroni’s multiple comparison test. The scheme on the left depicts each construct with the transmembrane segments and the coiled-coil domain shown as open and closed boxes respectively. The TMDs of α7 and EGFR are represented by boxes filled with vertical and horizontal lines respectively.

Download figure to PowerPoint

To explore the role of the two TMDs we substituted the first one either by the α7 signal peptide (construct α7/RIC-3, see also Fig. 1) or the TMD of the EGFR (construct EGFR1/RIC-3). The second TMD was also substituted by the TMD of the EGFR (construct EGFR2/RIC-3). In all cases, the effects of hRIC-3 were attenuated (Fig. 2). Thus, whereas inhibition of α4β2 expression by hRIC-3 was almost total (3% of the expression observed without hRIC-3), in the presence of α7/RIC-3, EGFR1/RIC-3 and EGFR2/RIC-3 inhibition was less prominent (15%, 32% and 30% of the control without hRIC-3 respectively). When both TMDs were simultaneously substituted (construct α7/EGFR2/RIC-3), less inhibition was observed (64% of α4β2 expression without hRIC-3). The differences observed respect to wild-type hRIC-3 were statistically significant in all cases. Thus, for inhibition of α4β2 nAChR expression, both TMDs appear to be important, especially if they are simultaneously substituted. In the latter case, however, a residual inhibitory activity of RIC-3 remains, as the expression observed in the presence of α7/EGFR2/RIC-3 (64%) is statistically different of the control in which RIC-3 was absent. Western blot analysis of oocyte lysates (Fig. 3, lanes 1–5) showed that all hRIC-3 constructs were expressed at comparable amounts, although construct α7/RIC-3 (lane 2) appeared to be expressed at a slightly lower level. Despite the small decrease in expression, this construct produced the closest effect to the produced by the wild-type hRIC-3 protein.

image

Figure 3.  Expression of different RIC-3 constructs in oocytes. Xenopus oocytes expressing the indicated hRIC-3 constructs (see Fig. 2) were lysed and an equal amount of protein separated by SDS–PAGE. Following immunoblotting with hRIC-3 antibodies, a prominent band was detected in all cases. The faint band at the bottom appears to be unspecific, as it was also observed in non-injected oocytes. Lanes 6 and 7 were originally at the left of the gel but were displaced to the right just for the sake of clarity. In this case, the faint band at the bottom was used as reference for a proper alignment. Notice that all RIC-3 proteins migrated slower than expected, probably as a result of their hydrophobicity and/or acidic nature (theoretical pI is around 5 for all of them). This experiment was replicated with oocytes from another frog, yielding the same result. Molecular weight markers (in kDa) are indicated at the right.

Download figure to PowerPoint

The increase in α7 expression previously observed with hRIC-3 (564% and 235% in α-Bgt binding and currents respectively) was less conspicuous when the chimaeric constructs were used instead, especially when the first TMD was exchanged (Fig. 2). Both, constructs α7/RIC-3 and EGFR1/RIC-3, produced similar effects, a slight increase in α-Bgt binding (∼130% of the control) and currents (∼150% of control), that was not statistically different of the expression observed in the absence of RIC-3 constructs. The substitution of the second TMD of hRIC-3 (construct EGFR2/RIC-3) was less harmful for hRIC-3 function, as currents were similar to the ones observed with the wild-type protein (258%). α-Bgt binding was also increased (248%), although at a lower extent than in the case of the wild-type protein. Substitution of both TMDs (construct α7/EGFR2/RIC-3) produced the same effects than the single substitution of the first TMD, the differences respect to the control without RIC-3 having no statistical significance. Therefore, the first TMD appears to be more important to promote α7 nAChR expression. As previously noticed with the effect on α4β2 expression, the slightly lower amount of construct α7/RIC-3 (Fig. 3, lane 2) does not seen to justify its attenuated effect on α7 expression, as the other constructs, which were expressed at higher levels, produced similar results.

We have shown previously that deletion of the loop linking the two TMDs (construct △37–90) did not abolish the inhibition of α4β2 expression observed with the whole hRIC-3 protein (Castillo et al. 2005). As expected, partial deletions of the loop (△34–54, △47–76 and △58–90) did not modify the mentioned inhibition (Fig. 2). Therefore, this region of hRIC-3 is not needed for its inhibitory action on α4β2 nAChRs. The insertion of an unrelated linker between the two TMDs (construct RIC-3 α1GlyR) also confirmed this conclusion. In the case of α7 nAChRs, the expression increase observed with hRIC-3 was not observed with construct △37–90, suggesting that the region between the two TMDs is important for this effect. Notice that this construct was expressed even at higher levels than wild-type RIC-3 (Fig. 3, lane 6), therefore its lack of effect would not be because of decreased expression. By contrast, the increase in α7 expression was restored in the presence of constructs △34–54, △47–76 and △58–90, suggesting the need for just a certain number of residues. These residues, however, must be from the RIC-3 linker, as an unrelated linker of similar length (construct RIC-3 α1GlyR) was unable to mimic the effects observed with the wild-type protein and the smaller deletions (Fig. 2), in spite of its prominent expression (Fig. 3, lane 7).

Effects of hRIC-3 on the surface trafficking of α7 nAChRs

We have previously shown that the enhancement of α7 nAChR expression by RIC-3 occurs, at least, at two levels: increasing the number of mature receptors and facilitating its transport to the membrane (Castillo et al. 2005). Further analysis of this process was carried out. For this purpose, we considered that the elevated surface expression of α7 nAChRs induced by hRIC-3 could be the result from increased receptor delivery to the oocyte surface but also from decreased turnover from the plasma membrane. To distinguish between these possibilities we measured, both, incorporation of newly synthesized receptors to the membrane and receptor internalization, either in the presence or absence of hRIC-3. The arrival of α7 nAChRs at the oocyte surface was monitored by [125I] α-Bgt binding to newly inserted receptors, as the pre-existing ones were previously blocked by unlabelled α-Bgt. It was evident that the rate of receptor incorporation was much higher in the presence of hRIC-3 (Fig. 4a). Besides, the rate of receptor turnover from the oocyte was also measured (Fig. 4b). Turnover rate was clearly slower when hRIC-3 was present. Therefore, it seems that not a single process but a combination of increased α7 receptor steady state levels (Castillo et al. 2005), facilitated transport and reduced degradation is the responsible for the increase in α7 expression induced by hRIC-3.

image

Figure 4.  Effect of hRIC-3 on the insertion and degradation of α7 nAChRs at the oocyte membrane. (a) Insertion of newly synthesized nAChRs into the oocyte surface as measured by α-Bgt binding. Data are shown for α7 nAChRs in the absence (closed circles) and presence (open circles) of hRIC-3, and represent the mean ± SE of 30 oocytes for each time point (n = 3 frogs). The discontinuous line represents the linear regression of the data. The values of the first point, indicated as t = 2 h, correspond to the binding assay carried out immediately after blocking with cold toxin. As the incubation with 125I-α-Bgt is carried out over 2 h, newly synthesized receptors arriving to the membrane during this time should be available for toxin binding. (b) Surface turnover rates of nAChRs deduced from 125I-α-Bgt release into oocyte culture medium. Data are shown for α7 nAChRs in the absence (closed circles) and presence (open circles) of hRIC-3, and represent the mean ± SE of 30 oocytes (n = 2 frogs). Data are expressed as a percentage of 125I-α-Bgt bound at the beginning of the experiment.

Download figure to PowerPoint

Induction of RIC-3 upon cell differentiation

Polyclonal antibodies directed against hRIC-3 were obtained from hen eggs using as immunogen a fusion protein made of glutathione S-transferase and a large portion of the C-terminal domain of RIC-3 (amino acids 160–368). According to gel electrophoresis, IgY was purified to almost homogeneity by modifying a commercially available protocol. The purified antibody was shown to recognize specifically the RIC-3 protein in western blots and immunohistochemistry of different preparations.

As shown in Fig. 5a, the antibody recognized the hRIC-3 protein (lane 4) in a lysate of COS cells transfected with a plasmid coding for a fusion protein made of EGFP and hRIC-3. An anti-green fluorescent protein antibody recognized the same band (lane 6), whereas an antibody fraction extracted from eggs from the same animal before hRIC-3 immunization yielded no immunoreaction (lane 2). Finally, no reaction was observed in untransfected COS cells (lanes 1, 3 and 5), indicating that they do not express endogenous RIC-3.

image

Figure 5.  RIC-3 protein is expressed in PC12 and SH-SY5Y cells and up-regulated upon differentiation. (a) COS cells expressing a hRIC-3/EGFP fusion protein (lanes 2, 4 and 6) were lysed and their proteins separated by SDS–PAGE. Control untransfected cells were also used (lanes 1, 3 and 5). Following immunoblotting with hRIC-3 (lanes 3 and 4) or EGFP (lanes 5 and 6) antibodies, a single molecular species was detected. A hRIC-3 pre-immune antibody fraction yielded no signal (lanes 1 and 2). (b) Proteins from lysates of PC12 (lanes 7–10) and SH-SY5Y (lanes 11–14) cells were separated by electrophoresis and immunoblotted with RIC-3 antibody (lanes 9, 10, 13 and 14). The pre-immune antibody was used as control (lanes 7, 8, 11 and 12). Extracts from differentiated cells (lanes 10 and 14) showed stronger immunoreaction than the ones from undifferentiated cells (lanes 9 and 13). The amount of protein applied to each well was equivalent in all cases. An anti-actin antibody (lanes 15–17) was used to confirm it. Molecular weight markers (in kDa) are indicated at the right of each panel.

Download figure to PowerPoint

When lysates of rat pheochromocytoma PC12 and human neuroblastoma SH-SY5Y cells were immunoblotted and analyzed with the anti-hRIC-3 antibody (Fig. 5b) a band was detected (lanes 9 and 13 respectively). Given the specificity of the antibody, previously shown with transfected COS cells (Fig. 5a), and the mobility of the band, which is close to the expected for the size of the RIC-3 protein (41 kDa), we deduced that these cell lines endogenously express RIC-3 or an antigenically related protein. Interestingly, immunoreaction was strongly potentiated when both cell types were differentiated (lanes 10 and 14). This induction also evidenced the presence of a double band, particularly clear in the case of PC12 cells (lane 10). The pre-immune antibody did not detect any protein (lanes 7, 8, 11 and 12). We also analyzed the muscle cell line C2C12. In this case the anti-RIC-3 antibody was unable to show immunoreaction, even upon cell differentiation (not shown), a process known to induce the expression of muscle-type nAChRs. The increase in RIC-3 expression induced by differentiation was also observed by immunohistochemistry (Fig. 6). Undifferentiated SH-SY5Y cells showed weak labelling (Fig. 6b), whereas the differentiated ones exhibited much stronger immunofluorescence (Fig. 6d).

image

Figure 6.  Immunofluorescence of RIC-3 in SH-SY5Y cells. Transmitted light (a and c) and fluorescent (b and d) images of undifferentiated (a and b) and differentiated (c and d) SH-SY5Y cells. Confirming the immunoblotting results (see Fig.5), labelling of RIC-3 is much more prominent in differentiated cells. Scale bar: 20 μm.

Download figure to PowerPoint

Localization of RIC-3 in the rat brain

Light microscopic immunohistochemical observation revealed that the RIC-3 protein is broadly distributed in the rat brain with a moderate labelling intensity. Table 1 illustrates the main areas where RIC-3 expression was detected, indicating representative nuclei or neuronal groups as well as their respective intensities.

Table 1.   RIC-3 distribution and labelling intensity in significant areas in the rat CNS
  1. −, not detected; +, weak signal; ++, moderate signal; +++, intense signal; s.c., scattered cells.

1. Cortex
 Layer I+
 Layer II+
 Layer III+
 Layer IV+
 Layer V+(+)
 Layer VI+
2. Amygdaloid complex++
3. Basal ganglia
 Striatum (CPu)+
 Substantia innominata+
 Globus pallidus++
 Substantia nigra pars compacta++
 Substantia nigra pars reticulatas.c.
4. Hippocampus
 Dentate gyrus++
 CA1/CA2++
 CA3++
5. Thalamus
 Medial habenula+++
 Dorsolateral geniculate nucleus++
 Lateral geniculate nucleus++
 Medial geniculate nucleus++
 Reticular nucleus+
6. Hypothalamus
 Lateral hypothalamic area+
 Mammillary nuclei++
7. Midbrain
 Superior colliculus+
 Inferior colliculus+
 Periaqueductal gray+
 Oculomotor nucleus (III)++
 Dorsal raphe++
 Red nucleus magnocellular++
 Ventral tegmental area+
8. Brainstem
 Mesencefalic trigeminal nu.+
 Locus coeruleus+
 A5 and A1 cells+
 Trapezoid body++
 Superior olive+
 Trigeminal nucleus (V)++
 Abducens nucleus (VI)++
 Facial nucleus (VII)++
 Gigantocellular+
 Hypoglossal nucleus (XII)++
 Dorsal cochlear nucleus++
 Inferior olive+++
 Solitary complex++
9. Cerebellum
 Purkinje cell layer
 Granular layer+(+)
 Molecular layer
 Deep nuclei+++

In rostral structures, cortex immunostaining presented a weak labelling. Pyramidal cells from different cortical areas appeared weakly labelled. In all cases, staining was restricted to the perikarya. Hippocampus showed a moderate labelling observed in dentate gyrus, CA1 and CA3 neurons (Fig. 7a–c). In the thalamus, neurons were moderately labelled, except in the Medial habenula, where they appeared intensely labelled. Regarding the basal ganglia, we observed a weak to moderate labelling in the striatum and globus pallidus and a moderate labelling in the substantia nigra neurons. A moderate intensity of the immunoreactivity to RIC-3 was also observed in the hypothalamus, mainly in the mammillary nuclei and in the midbrain, where neurons were clearly stained in motor related groups, such as oculomotor nucleus, red nucleus magnocellular and even in the dorsal raphe. In the brainstem, immunoreactivity to RIC-3 showed higher intensities, ranging from moderate to intense. Signal was moderate in the Dorsal cochlear nucleus and in the solitary complex (Fig. 7f), and also in motor nuclei, such as trigeminal (V), abducens (VI), facial (VII) and hypoglossal (XII) nuclei (Fig. 7g). Highest intensity was found in the inferior olive (Fig. 7f). In the cerebellum, it is noteworthy the lack of signal in the Purkinje cell layer and the weak to moderate labelling found in the granular layer (Fig. 7d). Neurons in the deep cerebellar nuclei appeared intensely labelled (Fig. 7e).

image

Figure 7.  Immuno-colocalization of the RIC-3 protein in rat brain coronal sections. (a) General view of the Hippocampus. (b). Detail showing the labelling at the CA1 region. (c) Detail showing the labelling at the CA3 region. (d) Coronal section showing the labelling in the cerebellum at the level of the granular layer, outstanding the absence of staining at the Purkinje cell layer (arrow). (e) Detail showing the staining in the deep cerebellar nuclei. (f–g) General view of two coronal sections at the brainstem. DG, dentate gyrus; ML, molecular layer; GL, granular cell layer; Sol, solitary complex; IO, inferior olive; VII, facial motor nucleus. Scale bars: (a, f and g) 1000 μm; (b, c and d) 100 μm; and (e) 200 μm.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Assembly of nAChRs occurs in a multistep process, which includes both, folding reactions and post-translational modifications following polypeptide synthesis and insertion into the endoplasmic reticulum membrane. Our group and others have shown that the RIC-3 protein might be involved in these processes. In this work, we have tried to approach different molecular and cellular aspects of the action of RIC-3 on nAChRs.

The function of the N-terminal region of RIC-3

As the first TMD of hRIC-3 (and of the homolog protein in other species) is adjacent to the N-terminus, it could play the role of a leader peptide, as we suggested previously with the aid of a prediction program (Castillo et al. 2005). On the other hand, in the case of the RIC-3 protein present in Caenorhabditis elegans and Drosophila, the first TMD is preceded by 33 and 14 amino acids, respectively, hence its role as a typical leader peptide might be discarded. To solve this apparent discrepancy we carried out in vitro synthesis of hRIC-3 in the presence of microsomal membranes and showed that no processing of the resulting protein had occurred (Fig. 1). By contrast, a chimaeric protein with the α7 subunit leader peptide was indeed processed and, therefore, it appears that the first TMD of hRIC-3 remains as a component of the mature protein during its synthesis ‘in vitro’. Western blots of the different RIC-3 constructs expressed in oocytes (Fig. 3) show that the protein containing the signal peptide of the α7 subunit (construct α7/RIC-3) has larger mobility than the wild-type RIC-3 protein, suggesting that efficient cleavage of the former but not of the later has also occurred during its expression in oocytes. A less prominent band of smaller size was also observed with constructs α7/RIC-3, EGFR2/RIC-3 and α7/EGFR2/RIC-3 (Fig. 3, lanes 2, 4 and 5). This band might be protein with the first TMD cleaved, if we assume that the larger band represent unprocessed protein. However, this does not seem to be the case because this smaller band migrates at about the same position than the RIC-3(△37–90) protein, whose theoretical molecular weight (35 600 kDa) is clearly smaller than the expected for cleaved α7/RIC-3 protein (38 000 kDa).

In contrast with the apparent lack of processing of the first TMD of RIC-3 that we observed ‘in vitro’, Cheng et al. (2007) have found that a chimaera made of the first TMD of RIC-3 and a soluble fluorescent protein (DsRed) was directed to the endoplasmic reticulum and secreted, suggesting that the first TMD of RIC-3 acts as a cleavable signal sequence. As it happens with our constructs α7/RIC-3, EGFR1/RIC-3 and α7/EGFR2/RIC-3, this demonstration of cleavage should be interpreted with caution, as it has been shown that many random sequences can functionally replace the secretion signal sequence (Kaiser et al. 1987). Moreover, in the experiment of Cheng et al. (2007) the first TMD of RIC-3 was isolated from its context at the whole RIC-3 protein, what could modify its behaviour regarding the signal sequence cleavage machinery, as the presence of downstream sequence elements, such as TMDs (the second TMD in the case of RIC-3), can influence the functionality of a signal sequence (Rehm et al. 2001). Another explanation for the discrepancy is that Cheng et al. (2007) examined cleavage in intact cell preparations whereas our study was performed ‘in vitro’.

The TMDs of hRIC-3 seem to be important for its action on α4β2 and α7 nAChRs, as their replacement attenuated the effects observed with the wild-type protein (Fig. 2). This was particularly evident when both TMDs were simultaneously substituted. Accordingly, it seems that the TMDs are not only needed for a proper localization of the hRIC-3 protein at the secretory pathway, but also contain specific amino acid residues essential for the function of hRIC-3. A different situation occurs with the region between the two TMDs. In the case of α4β2 nAChRs because a large deletion in this loop (△37–90) of hRIC-3 does not affect its inhibition of α4β2 expression (Fig. 2). On the other hand, and although this same deletion abolished the activation of α7 expression induced by hRIC-3, this effect was restored when different small loop regions were present (deletions △34–54, △47–76 and △58–90) but not when an unrelated linker of approximately the same length was used (construct α1GlyR). These results suggest a certain redundancy in this region, as different loop sequences restore α7 activation. Moreover, we were not able to detect a clear preference for a determined region of the loop and, therefore, the only condition is the presence of a certain loop sequence, but not of unrelated sequences, separating both TMDs. It is also interesting to notice the slightly different requirements for the action of hRIC-3 on α4β2 and α7 nAChRs. Regarding the TMDs, the first one appears to be more important for α7 activation whereas both TMDs contribute equivalently to α4β2 inhibition. In the case of the loop, it is not needed in order to get α4β2 inhibition whereas a certain amount of loop residues is needed for α7 activation. These facts suggest that the interaction of hRIC-3 with these nAChR subtypes is different, and/or take place at different cell locations and/or through different protein complexes in which hRIC-3 might be involved.

α7 nAChR trafficking is affected by hRIC-3

We have previously shown that the mechanism that hRIC-3 uses to induce higher expression of α7 nAChR expression does not seem to rely on a single receptor biogenesis step (Castillo et al. 2005). In fact, hRIC-3 increased the total number of receptors as well as the proportion of them present at the oocyte surface. The latter could be the consequence of more efficient transport, as we suggested previously, but also of delayed receptor internalization. We studied these processes in more detail (Fig. 4) and it is evident that the presence of hRIC-3 not only accelerates the arrival of new receptors to the membrane, but also reduces their internalization rate, suggesting that, in general, hRIC-3 increases the stability of α7 nAChRs. A question then arises: are α7 nAChRs more stable because they maintain a direct and permanent interaction with hRIC-3 on the cell membrane? Coimmunoprecipitation experiments have provided evidence that hRIC-3 associates with unassembled receptor subunits (Lansdell et al. 2005) and that this association is probably transient (Cheng et al. 2005). In addition, hRIC-3 has not been detected at significant levels on the cell surface of transfected cells (Castillo et al. 2005; Cheng et al. 2005) or in cells expressing endogenous RIC-3 (Fig. 6). In co-transfection experiments, we observed colocalization of hRIC-3 with α4β2 nAChRs inside the cells but not with α7 (Castillo et al. 2005). Consequently, we favour the possibility that RIC-3, alone or forming part of a larger protein complex, might modify α7 subunits during its transient interaction and that this modification stabilizes α7 nAChRs at the surface.

Presence of RIC-3 in neurons

RIC-3 transcripts have been found in brain (Halevi et al. 2003) and in some ‘neuronal’ cell lines, like SH-SY5Y and PC12 (Williams et al. 2005). Here, we have shown that the corresponding RIC-3 protein is also present in SH-SY5Y and PC12 cells. Two closely located bands were observed, especially in PC12 cells, that might be the consequence of some post-translational modification and/or reflects the existence of different isoforms of the RIC-3 protein (Halevi et al. 2003). Another possibility would be that the doublet represents RIC-3 protein with its first TMD partially cleaved, if we assume that the latter plays the role of an inefficient signal sequence able to generate two populations of processed and unprocessed protein. However, only a few examples of this signal inefficiency have been reported (García et al. 1988; Shaffer et al. 2005).

Differentiation of SH-SY5Y and PC12 cells strongly increased RIC-3 expression. Further experiments will be needed to clear the mechanism that activates RIC-3 expression upon cell differentiation. This activation might be relevant for the regulation of nAChR expression that occurs during cell differentiation. The degree and direction of this regulation depend on receptor and cell types (for e.g. see Rogers et al. 1992; Henderson et al. 1994; Halvorsen et al. 1995) and even on potential heterogeneities within a cell line; notice, for instance, the contrast between the results of Rogers et al. (1992) and Henderson et al. (1994). Therefore, a variety of factors might be involved in nAChR regulation during differentiation, and RIC-3 could be one of them.

Immunostaining for the RIC-3 protein in the rat brain was broad, as expected from studies related to the localization of its mRNA in the mouse brain (Halevi et al. 2003). In general, we found a good correspondence with the distribution of RIC-3 transcripts except in a few places. Thus, we detected RIC-3 in all areas of the hippocampus, including the dentate gyrus, where no RIC-3 transcripts were found. Also, transcripts were found in the Purkinje cell layer, but we observed no labelling for RIC-3 in any Purkinje cell. These discrepancies could be because of low levels of protein or RNA that precluded their detection and/or to the use of different species (mouse and rat). Regarding the localization of α7 nAChRs (Domínguez del Toro et al. 1994), we have found good correspondence with the distribution of the RIC-3 protein, although, in general, labelling intensity of the latter was weaker. A discrepancy was observed in the cerebellum, as localization of α7 protein in Purkinje cells (Domínguez del Toro et al. 1997) is not correlated by the presence of RIC-3. Given the contradictory results regarding the existence of functional α7 nAChRs in these cells (reviewed by De Filippi et al. 2005), a possibility is that, despite the presence of α7 protein, the lack of RIC-3 hinders the expression of functional α7 nAChRs at the cell membrane. This situation, that also might be occurring in other places, would mimic the one observed in most cell lines transfected with α7 cDNA, which were unable to express functional α7 nAChRs unless the RIC-3 protein was present (Williams et al. 2005).

Acknowledgements

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

This work was supported by Grants from the Ministry of Education and Science of Spain and FEDER (BMC2002-00972, SAF2002-00209, SAF2002-02731, SAF2005-00534 and SAF2005-02045) and Generalitat Valenciana (GRUPOS03/038). FC was the recipient of a pre-doctoral fellowship (No. 153783) of the Consejo Nacional de Ciencia y Tecnología from México. We thank Susana Gerber for her expert technical assistance.

References

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  • Avila A. M., Davila-Garcia M. I., Ascarrunz V. S., Xiao Y. and Kellar K. J. (2003) Differential regulation of nicotinic acetylcholine receptors in PC12 cells by nicotine and nerve growth factor. Mol. Pharmacol. 64, 974986.
  • Ben-Ami H. C., Yassin L., Farah H., Michaeli A., Eshel M. and Treinin M. (2005) RIC-3 affects properties and quantity of nicotinic acetylcholine receptors via a mechanism that does not require the coiled-coil domains. J. Biol. Chem. 280, 2805328060.
  • Castillo M., Mulet J., Gutierrez L. M., Ortiz J. A., Castelan F., Gerber S., Sala S., Sala F. and Criado M. (2005) Dual role of the RIC-3 protein in trafficking of serotonin and nicotinic acetylcholine receptors. J. Biol. Chem. 280, 2706227068.
  • Cheng A., McDonald N. A. and Connolly C. N. (2005) Cell surface expression of 5-hydroxy-tryptamine type 3 receptors is promoted by RIC-3. J. Biol. Chem. 280, 2250222507.
  • Cheng A., Bollan K. A., Greenwood S. M., Irving A. J. and Connolly C. N. (2007) Differential subcellular localization of RIC-3 isoforms and their role in determining 5-HT3 receptor composition. J. Biol. Chem. 282, 2615826166.
  • Christianson J. C. and Green W. N. (2004) Regulation of nicotinic receptor expression by the ubiquitin-proteasome system. EMBO J. 23, 41564165.
  • De Filippi G., Baldwinson T. and Sher E. (2005) Nicotinic receptor modulation of neurotransmitter release in the cerebellum. Prog.Brain Res. 148, 309320.
  • Devreotes P. N. and Fambroug D. M. (1975) Acetylcholine receptor turnover in membranes of developing muscle fibers. J. Cell Biol. 65, 335358.
  • Domínguez del Toro E., Juiz J. M., Peng X., Lindstrom J. and Criado M. (1994) Immunocytochemical localization of the α7 subunit of the nicotinic acetylcholine receptor in the rat central nervous system. J. Comp. Neurol. 349, 325342.
  • Domínguez del Toro E., Juíz J. M., Smillie F. I., Lindstrom J. and Criado M. (1997) Expression of α7 neuronal nicotinic receptors during postnatal development of the rat cerebellum. Dev. Brain Res. 98, 125133.
  • Encinas M., Iglesias M., Liu Y., Wang H., Muhaisen A., Ceña V., Gallego C. and Comella J. X. (2000) Sequential treatment of SH-SY5Y cells with retinoic acid and brain-derived neurotrophic factor gives rise to fully differentiated, neurotrophic factor-dependent, human neuron-like cells. J. Neurochem. 75, 9911003.
  • García P. D., Ou J. H., Rutter W. J. and Walter P. (1988) Targeting of the hepatitis B virus precore protein to the endoplasmic reticulum membrane after signal peptide cleavage: translocation can be aborted and the product released into the cytoplasm. J. Cell Biol. 106, 10931104.
  • Garcia-Guzman M., Sala F., Sala S., Campos-Caro A. and Criado M. (1994) Role of two acetylcholine receptor subunit domains in homomer formation and intersubunit recognition, as revealed by α3 and α7 subunit chimeras. Biochemistry 33, 1519815203.
  • Guan K. L. and Dixon J. E. (1991) Eukaryotic proteins expressed in Escherichia Coli: an improved thrombin cleavage and purification procedure of fusion proteins with glutathione S-transferase. Anal. Biochem. 192, 262267.
  • Halevi S., Yassin L., Eshel M., Sala F., Sala S., Criado M. and Treinin M. (2003) Conservation within the RIC-3 gene family. Effectors of mammalian nicotinic acetylcholine receptor expression. J. Biol. Chem. 278, 3441134417.
  • Halvorsen S. W., Jiang N. and Malek R. (1995) Regulation of nicotinic acetylcholine receptors on human neuroblastoma cells during differentiation. Biochem. Pharmacol. 50, 16651671.
  • Henderson L. P., Gdovin M. J., Liu C., Gardner P. D. and Maue R. A. (1994) Nerve growth factor increases nicotinic ACh receptor gene expression and current density in wild-type and protein kinase A-deficient PC12 cells. J. Neurosci. 14, 11531163.
  • Kaiser C. A., Preuss D., Grisafi P. and Botstein D. (1987) Many random sequences functionally replace the secretion signal sequence of yeast invertase. Science 235, 312317.
  • Krieg P. A. and Melton D. A. (1984) Functional messenger RNAs are produced by SP6 in vitro transcription of cloned cDNAs. Nucleic Acids Res. 12, 70577070.
  • Lansdell S. J., Gee V. J., Harkness P. C., Doward A. I., Baker E. R., Gibb A. J. and Millar N. S. (2005) RIC-3 enhances functional expression of multiple nicotinic acetylcholine receptor subtypes in mammalian cells. Mol. Pharmacol. 68, 14311438.
  • Lindstrom J. M. (2003) Nicotinic acetylcholine receptors of muscles and nerves: comparison of their structures, functional roles, and vulnerability to pathology. Ann. NY Acad. Sci. 998, 4152.
  • Ortiz J. A., Castillo M., Domínguez del Toro E., Mulet J., Gerber S., Valor L. M., Sala S., Sala F., Gutierrez L. M. and Criado M. (2005) The cysteine-rich with EGF-like domains 2 (CRELD2) protein interacts with the large cytoplasmic domain of human neuronal nicotinic acetylcholine receptor alpha4 and beta2 subunits. J. Neurochem. 95, 15851596.
  • Rehm A., Stern P., Ploegh H. L. and Tortorella D. (2001) Signal peptide cleavage of a type I membrane protein, HCMV US11, is dependent on its membrane anchor. EMBO J. 20, 15731582.
  • Rogers S. W., Mandelzys A., Deneris E. S., Cooper E. and Heinemann S. (1992) The expression of nicotinic acetylcholine receptors by PC12 cells treated with NGF. J. Neurosci. 12, 46114623.
  • Shaffer K. L., Sharma A., Snapp E. L. and Hedge R. S. (2005) Regulation of protein compartmentalization expands the diversity of protein function. Dev. Cell 9, 545554.
  • Sine S. M. and Engel A. G. (2006) Recent advances in Cys-loop receptor structure and function. Nature 440, 448455.
  • Williams M. E., Burton B., Urrutia A., Shcherbatko A., Chavez-Noriega L. E., Cohen C. J. and Aiyar J. (2005) Ric-3 promotes functional expression of the nicotinic acetylcholine receptor alpha7 subunit in mammalian cells. J. Biol. Chem. 280, 12571263.