The chaperone protein 14-3-3eta interacts with the nicotinic acetylcholine receptor alpha 4 subunit. Evidence for a dynamic role in subunit stabilization.

By using the large cytoplasmic domain of the nicotinic acetylcholine receptor (AChR) alpha4 subunit as a bait in the yeast two-hybrid system, we isolated the first cytosolic protein, 14-3-3eta, known to interact directly with neuronal AChRs. 14-3-3eta is a member of a family of proteins that function as regulatory or chaperone/ scaffolding/adaptor proteins. 14-3-3eta interacted with the recombinant alpha4 subunit alone in tsA 201 cells following activation of cAMP-dependent protein kinase by forskolin. The interaction of 14-3-3eta with recombinant alpha4 subunits was abolished when serine 441 of the alpha4 subunit was mutated to alanine (alpha4(S441A)). The surface levels of recombinant wild-type alpha4beta2 AChRs were approximately 2-fold higher than those of mutant alpha4(S441A)beta2 AChRs. The interaction significantly increased the steady state levels of the alpha4 subunit and alpha4beta2 AChRs but not that of the mutant alpha4(S441A) subunit or mutant alpha4(S441A)beta2 AChRs. The EC50 values for activation by acetylcholine were not significantly different for alpha4beta2 AChRs and alpha4(S441A)beta2 AChRs coexpressed with 14-3-3eta in oocytes following treatment with forskolin. 14-3-3 coimmunopurified with native alpha4 AChRs from brain. These results support a role for 14-3-3 in dynamically regulating the expression levels of alpha4beta2 AChRs through its interaction with the alpha4 subunit.

Neuronal nicotinic acetylcholine receptors (AChR) 1 are a family of ligand-gated, cation-selective, homo-or heteropentameric ion channels expressed in the peripheral and central nervous system (1,2). A multitude of neuronal AChR subtypes assembled from different combinations of ␣2-␣9 and ␤2-␤4 subunits have been identified (3,4). Of these, the ␣4␤2 AChR is widely expressed in the central nervous system and represents Ͼ80% of the high affinity [ 3 H]nicotine-binding sites in mammalian brain (5). Our understanding of their physiological roles comes most recently from gene knock-out studies in mice. Mice in which the ␣4 subunit gene has been deleted lack [ 3 H]nicotine-or [ 3 H]epibatidine-binding sites in their brain and exhibit reduced antinociceptive effects of nicotine (6). Mice in which the ␤2 subunit gene has been deleted also show little [ 3 H]nicotine binding in their brains, lose their sensitivity to nicotine in passive avoidance tasks (7), and show attenuated self-administration of nicotine (8) suggesting that ␣4␤2 AChRs have a role in mediating addiction to nicotine. The normal and pathophysiological functions mediated by ␣4␤2 AChRs are of significant importance to human health. Some inherited forms of epilepsy, such as the autosomal dominant nocturnal frontal lobe epilepsies, are caused by ␣4␤2 AChRs harboring at least two separate mutations within their ␣4 subunit (9 -12). Most recently, ␣4␤2 AChRs, among other ␤2 subunit-containing AChRs, have been implicated in neuronal survival during aging, as surmised from the neurodegeneration observed in ␤2-subunit knock-out mice (13).
The ␣4 subunit, like the other AChR subunits, consists of an extracellular N-terminal domain, followed by three transmembrane domains (M1-M3), a large cytoplasmic domain, a fourth transmembrane domain (M4), and a short extracellular C terminus. The large cytoplasmic domain is highly divergent among the various subunits, and this sequence divergence presumably provides the diversity necessary for different AChR subtypes to interact directly with cytosolic proteins of different function. To identify such proteins associated with ␣4␤2 AChRs, we used the large cytoplasmic domain of the ␣4 subunit as a bait to screen a mouse brain cDNA yeast two-hybrid library. Here we describe the isolation of a known protein termed 14-3-3. The 14-3-3 proteins family consists of seven isoforms (␤, ␥, , , ⑀, , and ) that function as intracellular regulators or chaperone/scaffolding/adaptor proteins in diverse cellular functions (14). The binding of 14-3-3 to most of their protein targets are mediated by a phosphoserine or phosphothreonine residue within a consensus binding site motif or within sequences closely resembling it (15).
We found that 14-3-3 interacted with the recombinant AChR ␣4 subunits alone following activation of PKA. The interaction was mediated by serine 441 of the ␣4 subunit within a motif similar to a known consensus binding site motif for 14-3-3 proteins. The interaction significantly increased the steady state levels of the ␣4 subunit alone and ␣4␤2 AChRs. The surface levels of recombinant wild-type ␣4␤2 AChRs were ϳ2-fold higher than those of mutant ␣4 S441A ␤2 AChRs. 14-3-3 coimmunopurified with native ␣4 AChRs from brain suggesting its interaction with native ␣4 AChRs is physiologically relevant. These results support a possible role for 14-3-3 in dynamically regulating the steady state levels of ␣4␤2 AChRs through its interaction with the ␣4 subunit in the ER/Golgi compartments, following activation of PKA.

EXPERIMENTAL PROCEDURES
Constructs-The rat ␣4 and ␤2 AChR subunit cDNA clones were generously provided by Dr. Stephen Heinemann (Salk Institute, San Diego, CA). All constructs were made by the polymerase chain reaction (PCR) using appropriate pairs of forward and reverse synthetic oligonucleotide primers (Life Technologies, Inc.) and Pfu Turbo (Stratagene, Inc., San Diego, CA). All DNA sequence analysis was done using the ThermoSequenase radiolabeled terminator cycle sequencing kit (Amersham Pharmacia Biotech). For all primers, the restriction enzyme sites are underlined. The cDNA sequence corresponding to the large cytoplasmic domain (amino acids 302-561) of the rat ␣4 subunit was amplified using the forward primer 5Ј-GGG GAA TTC GTG CAC CAC CGC TCG CCA CGC-3Ј and the reverse primer 5Ј-CCC GGA TCC TCA CTT CAC CGA GAA GT C AG TGT C-3Ј by PCR and subcloned into the EcoRI-BamHI sites of the vector pLexA (CLONTECH Laboratories, Inc., Palo Alto, CA) to form the ␣4 bait. The nested deletions of the ␣4 cytoplasmic domain were generated by PCR using the forward primer 5Ј-GGG GAA TTC GTG CAC CAC CGC TCG CCA CGC-3Ј and nested reverse primers 5Ј-GGG GGA TCC TCA GGT GCC TCC CGC CTT GAG  CAC-3Ј; 5Ј-GGG GGA TCC TCA CAG GGA GGT CGG GGA GCT  GGT-3Ј; 5Ј-GGG GGA TCC TCA TTC TTG GGA GCT GGG CAC ATG-3Ј; 5Ј-GGG GGA TCC TCA GGC CTT CTC AAC CTC TGA TGT-3Ј;  5Ј-GGG GGA TCC TCA TGA CAG ACC TTG GTT GCA GAT-3Ј; 5Ј-GGG  GGA TCC TCA GTT GTC TTT GAC CAC AGA GGG-3Ј and were subsequently subcloned into the EcoRI-BamHI sites of the pLexA. The mutated subunit ␣4 S441A was generated by mutagenesis using the following two primers: 5Ј-C AAA GCC AGG TCC CTG GCT GTC CAG CAT GTG CCC-3Ј and 5Ј-GGG CAC ATG CTG GAC AGC CAG GGA CCT GGC TTT G-3Ј in conjunction the QuikChange site-directed mutagenesis kit (Stratagene Inc., San Diego, CA). The mouse 14-3-3, the rat ␣4 subunit, and the rat ␤2 subunit cDNAs were generated by amplification of the full clone by PCR using forward primer 5Ј-GGG  AAT TCG CCA CCA TGG GGG ATC GAG AGC AG-3Ј and reverse  primer 5Ј-GGG TCT AGA TCA GTT GCC TTC TCC TGC TTC TTC-3Ј  for 14-3-3; forward primer 5Ј-GGG AAT TCG CCA CCA TGG CCA  ATT CGG GCC CCG GG-3Ј and reverse primer 5Ј-GGG TCT AGA TCA  GCA AGC AGC CAG CCA GGG AGG-3Ј for the ␣4 subunit; and forward  primer 5Ј-GGG AAT TCG CCA CCA TGC TGG CTT GCA TGG CCG  GG-3Ј and reverse primer 5Ј-GGG TCT AGA TCA CTT GGA GCT GGG  AGC TGA GTG-3Ј for the ␤2 subunit. The amplified cDNAs were ligated into the EcoRI-XbaI sites of the mammalian cell expression vector pEF6/myc-His A (Invitrogen, Carlsbad, CA).
Yeast Two-hybrid Library Screen-Yeast two-hybrid screens were carried out according to a standard protocol (CLONTECH Laboratories Inc., Palo Alto, CA). The ␣4 bait plasmid pLexA and the p8op-LacZ reporter gene plasmid were first transformed in EGY48 yeast cells followed by transformation of the library of brain cDNA plasmids. Approximately 2 ϫ 10 6 yeast cells cotransformed with the bait and cDNAs from a premade mouse brain cDNA Matchmaker LexA library (CLONTECH Laboratories Inc., Palo Alto, CA) were screened. Positive clones were selected for their ability to grow on plates lacking leucine, tryptophan, histidine, and uracil and assayed for ␤-galactosidase activity on media supplemented with X-gal. Plasmids containing the brain cDNAs were isolated from positive yeast cells, and their nucleotide sequences were determined by manual DNA sequencing using the ThermoSequenase radiolabeled terminator cycle Kit (Amersham Pharmacia Biotech). Seven clones of 14-3-3 that interacted with the ␣4 bait were characterized. These clones contained slightly different cDNA sizes but all had the full-length cDNA of 14-3-3 as determined from limited sequence analysis of their 5Ј ends. One of these clones containing the full-length 14-3-3 subunit was used in all subsequent work.
Antibodies-The anti-14-3-3 mouse monoclonal antibody that crossreacts with multiple isoforms of 14-3-3 and the anti-␤2 subunit that cross-reacts with ␤2 subunits on immunoblots were obtained from Santa Cruz Biotechnology (Santa Cruz, CA). Protein G affinity-purified anti-␣4 subunit (299) and anti-␤2 subunit (295) mAbs were generously provided by Dr. Lindstrom (University of Pennsylvania, Philadelphia). The goat anti-mouse and anti-rat horseradish peroxidase-conjugated Abs were obtained from Pierce. mAbs were coupled to Actigel ALD beads at a concentration of 0.5 mg/ml of gel using the manufacturer's instructions (Sterogene Bioseparations Inc., Carlsbad, CA).

Expression of Recombinant AChR Subunits in Human Embryonic
Kidney tsA201 Cells-Human tsA201 cells (16), a derivative of the human embryonic kidney cell line 293, were cultured at 37°C in 6-well plates in Dulbecco's modified Eagle's medium (Life Technologies, Inc.) supplemented with 10% fetal bovine serum, 50 g/ml penicillin, and 50 g/ml streptomycin. Cells were transfected using LipofectAMINE 2000 (Life Technologies, Inc.) at 90% confluency (ϳ10 6 cells/well) with various combinations of cDNAs as per the manufacturer's instructions and utilized after ϳ48 h. The cDNAs were cloned into the vector pEF6/myc-His but lacked the myc-His tag because of the presence of the endogenous stop codon present in each cloned cDNA. AChR subunit assembly was found to be more efficient at 30°C than at 37°C as described previously (17), and hence the experiments were performed with cells incubated at 30°C following transfection. Expression and Analysis of the ␣4 Subunit-To study the effect of 14-3-3 on the ␣4 subunit alone, tsA201 cells seeded (400,000 per well) in 12 wells were incubated at 37°C. The next day the cells were cotransfected with ␣4 or ␣4 S441A with or without 14-3-3 and incubated at 30°C (the DNA concentrations and ratios were kept constant by using the pEF6A vector DNA). The transfected cells were treated 24 h with or without forskolin (10 M). After washing once with ice-cold PBS, the cells were solubilized in 500 l of the following lysis buffer: 50 mM NaCl, 30 mM triethanolamine, pH 7.5, 5 mM EGTA, 5 mM EDTA, 1 mM benzamidine, 5 g/ml aprotinin, 5 g/ml leupeptin, 5 g/ml pepstatin, and 2% Nonidet P-40. After shaking for 3 h at 4°C, the lysates were centrifuged for 15 min at 18,000 ϫ g, and 20 l of the supernatant were analyzed by SDS-PAGE.
Immunoisolation of Recombinant AChRs in tsA201 Cells-tsA201 cells were washed twice with ice-cold PBS containing 50 mM NaF and 1 mM sodium orthovanadate lysed in 1 ml of lysis buffer (50 mM NaCl, 30 mM triethanolamine, pH 7.5, 5 mM EGTA, 5 mM EDTA, 50 mM NaF, 1 mM phenylmethylsulfonyl fluoride, 1 mM benzamidine, 2 mM sodium vanadate, 10 mM p-nitrophenyl phosphate, 25 g/ml aprotinin, 25 g/ml leupeptin, 25 g/ml pepstatin, 0.3 M okadaic acid, 1 mM sodium tetrathionate, 1 mM N-ethylmaleimide, 50 M phenylarsine oxide, and 1% Nonidet P-40), and agitated vigorously for 2 h at 4°C. After centrifugation at 18,000 ϫ g for 15 min, the clear supernatant from each sample (ϳ1 ml) was incubated with 10 l of mAb-coupled beads (that were preblocked with 5% non-fat milk for 30 min) for 24 h. The beads were then washed 8 times with ϳ800 l of solubilization buffer and eluted with sample buffer (lacking ␤-mercaptoethanol to avoid reduction of the disulfide linkage of the IgG chains) at 60°C for 30 min, and then ␤-mercaptoethanol was added to the eluted samples that were then boiled for an additional 5 min prior to analysis by SDS-PAGE.
Immunoisolation of Native AChRs from Rat Brain-Frozen rat brains were homogenized in 10 volumes of homogenization buffer (50 mM NaCl, 30 mM triethanolamine, pH 7.5, 5 mM EGTA, 5 mM EDTA, 50 mM NaF, 1 mM phenylmethylsulfonyl fluoride, 1 mM benzamidine, 2 mM sodium vanadate, 10 mM p-nitrophenyl phosphate) using a homogenizer (OMNI International, Warrenton, VA). The homogenate was centrifuged at 100,000 ϫ g in a Beckmann 50.2 Ti rotor for 30 min at 4°C. The membrane pellet was further briefly homogenized and then extracted with 3 volumes of a solubilization buffer (homogenization buffer containing 1% Nonidet P-40 and 25 g/ml aprotinin, 25 g/ml leupeptin, 25 g/ml pepstatin, 0.1 M okadaic acid, 1 mM sodium tetrathionate, 1 mM N-ethylmaleimide, 50 M phenylarsine oxide) for 2 h at 4°C. The clear supernatant obtained after centrifugation of the pellet at 18,000 ϫ g for 30 min was used for all subsequent immunoisolation procedures. Detergent-solubilized brain extracts (typically 10 ml) thus obtained were incubated with ϳ25 l of mAb-coupled Actigel ALD beads (that were preblocked with 5% non-fat milk for 30 min) at 4°C for 72 h. In initial experiments, to ensure that the binding observed was specific, we first determined the number of successive washes of the mAb beads necessary for the complete removal of unbound 14-3-3 proteins that are abundant in brain extracts and many of whose isoforms cross-react with the anti-14-3-3 mAb used in the immunoblotting experiments. The beads thus were typically washed 10 times with ϳ800 l of solubilization buffer and eluted with sample buffer (lacking ␤-mercaptoethanol to avoid reduction of the disulfide linkage of the IgG chains) at 60°C for 30 min, and then ␤-mercaptoethanol was added to the eluted samples prior to analysis by SDS-PAGE.
Immunoblot Analysis-The proteins bound to the Ab beads were eluted with protein sample buffer and fractionated by SDS-PAGE. The proteins were electroblotted onto polyvinylidene difluoride membrane (IMMUN-BLOT; Bio-Rad), and the membranes were incubated with diluted (typically 1/200 to 1/1000) primary Abs in phosphate-buffered saline solution containing 0.1% Tween and 5% non-fat milk. The binding of the primary mAbs was detected using appropriate secondary Abs conjugated to horseradish peroxidase in conjunction with a chemiluminescence detection kit (SuperSignal, Pierce). To reduce nonspecific binding, the blots were typically cut in half, and the top half was probed with the anti-␣4 subunit mAb and the bottom half with the anti-14-3-3 mAb thus eliminating the need for sequential reprobing of the blots.
Enzyme-linked Immunoassay for Cell Surface AChRs-Cell surface ␣4␤2 AChRs were measured as described previously (17). Briefly, 48 h after transfection, tsA201 cells plated in 12-well plates (0.5 ϫ 10 6 cells/well) were washed once in PBS and then blocked with PBS containing 3% BSA, and the cells were incubated for 1 h with an anti-␤2 subunit mAb (295) in PBS containing 3% BSA at room temperature. After four washes with PBS the cells were fixed with formaldehyde (3%) for 10 min, washed three times with PBS, and blocked again for 10 min. The cells were then incubated with horseradish peroxidase-conjugated goat anti-rat secondary Ab for 1 h in the presence of 3% BSA, washed six times with PBS, and incubated with 300 l of the HRP substrate 3,3Ј,5,5Ј-tetramethylbenzidine (Sigma) for 1 h. The absorbance of the supernatant was then measured at 655 nm in a Beckman spectrophotometer.
Expression in Xenopus Oocytes-cDNAs were subcloned into the vector pSP64T (Invitrogen) with a modified polylinker. cRNAs from linearized cDNA templates were synthesized in vitro using SP6 RNA polymerase in conjunction with reagents from the mMessage mMachine kit (Ambion, Austin, TX). Xenopus oocytes were prepared for injection as described previously (18). Oocytes were injected with 20 ng of cRNAs for the ␣4 and ␤2 subunits and 40 ng of 14-3-3 per oocyte and incubated for 3-7 days at 16 -18°C in 50% L-15 medium (Life Technologies, Inc.) containing 10 mM HEPES buffer, pH 7.5.
Electrophysiological Recordings-Currents were measured using standard two-microelectrode voltage-clamp amplifier (Oocyte Clamp OC-725C) as described previously (18). Electrodes were filled with 3 M KCl and had resistances of 1.0 -2.0 megohms for the voltage electrode and 0.5-0.1 megohm for the current electrode. All records were digitized at 200 Hz with MacLab software and hardware (AD Instruments). Data were analyzed using KaleidaGraph. The recording chamber was perfused at a flow rate of 10 ml/min with ND-96 solution (96 mM NaCl, 5 mM KCl, 1.8 mM CaCl 2 , 1 mM MgCl 2 , 5 mM HEPES, pH 7.6).
Immunohistochemistry-Cells were fixed with an ice-cold mixture of 50% methanol, 50% acetone for 5 min, washed 3 times with 2 ml of PBS, and blocked using PBS containing 2% BSA for 30 min. Cells were then incubated with diluted primary anti-␣4 mAb (1/2000 dilution) in PBS containing 4% NGS with gentle shaking for 1 h at 4°C. Cells were washed 3 times for 15 min per wash in PBS, incubated with diluted (1/1000) fluorophore-conjugated anti-rat secondary Ab in PBS containing 4% NGS for 1 h at room temperature, and fixed again with an ice-cold mixture of 50% methanol, 50% acetone for 5 min and washed 3 times with PBS. Cells were then incubated with diluted primary anti-14-3-3 mAb (1/200 dilution) in PBS containing 4% NGS with gentle shaking for 1 h at 4°C. Cells were washed 3 times for 15 min per wash in PBS and then incubated with diluted (1/1000) fluorophore-conjugated anti-mouse secondary Ab in PBS containing 4% NGS for 1 h at room temperature. The cells were then washed 3 times for 15-min periods in PBS and then used for immunofluorescence microscopy. Nontransfected cells were processed in parallel as controls for nonspecific staining. The secondary Abs used were goat anti-rat Alexa Fluor 488-conjugated Ab (catalog number A-11006) and the goat anti-mouse Alexa Fluor 546-conjugated Ab (catalog number A-11030) from Molecular Probes, Eugene, OR.
Confocal Microscopy-Confocal microscopy was accomplished using a Noran Instruments Odyssey XL confocal laser scanning microscope (Noran Instruments, Middleton, WI). Cells were scanned using 1-m steps in the z axis, alternating between fluorescein isothiocyanate and rhodamine filters. The resulting confocal images were captured on Silicon Graphics workstations. Images were pseudocolored and processed using Adobe Photoshop software.

FIG. 1. Mapping interaction of 14-3-3 within the ␣4 cytoplasmic domain.
A, residues 413-450 of the ␣4 cytoplasmic domain mediate interaction with 14-3-3. Six nested C-terminal deletions of the ␣4 cytoplasmic domain were tested for their ability to interact with 14-3-3 protein in the yeast two-hybrid system. Positive clones were determined by the ability of yeast cells to grow on plates lacking leucine, tryptophan, and histidine and for their ability to turn blue on media supplemented with X-gal. Media lacking histidine and tryptophan (his Ϫ trp Ϫ ) select for the presence the plasmids carrying the bait and the library protein. Media lacking histidine, tryptophan, and leucine (his Ϫ trp Ϫ leu Ϫ ) additionally select for interaction between the bait protein and the interaction protein.
B, serine 441 is essential for interaction of 14-3-3 with the cytoplasmic domain of the ␣4 AChR subunit. 14-3-3 interacts with the wild-type ␣4 bait but not a mutant ␣4 bait in which residue serine 441 (underlined) is mutated to alanine within a putative 14-3-3 consensus binding-site motif RSLSVQ.

14-3-3 Interacts with the Large Cytoplasmic Domain of the ␣4 Subunit in the Yeast Two-hybrid System-
The large cytoplasmic domain corresponding to amino acids 302-561 of the rat AChR ␣4 subunit was used as a bait to screen ϳ2 ϫ 10 6 clones of a mouse brain cDNA LexA yeast two-hybrid library. The large cytoplasmic domain extends from the third to the fourth transmembrane domain of the AChR ␣4 subunit. Multiple clones of 14-3-3 that interacted with the ␣4 bait were obtained. A full-length clone of 14-3-3 was chosen for further characterization.
To delineate the site where 14-3-3 interacts with the AChR ␣4 subunit cytoplasmic domain, a series of C-terminal nested deletions of the cytoplasmic loop was created as LexA fusion protein baits and tested for their ability to interact with the 14-3-3 clone in the yeast two-hybrid system. The interaction was determined by both the ability of transformed yeast cells to grow on media lacking leucine, tryptophan, histidine, and uracil and by their ability to turn blue on media supplemented with X-gal. As controls, the LexA protein alone was used as a bait. We mapped the interaction of 14-3-3 to residues 413-450 (Fig. 1A). A putative motif, RSXSXP, in which the underlined serine residue is phosphorylated, has been demonstrated previously to be important for the binding of 14-3-3 proteins to some of its target proteins. A sequence that closely resembles this motif, RSLSVQ, occurs within the region 413-450 that was found to be essential for the interaction of 14-3-3 with the ␣4 cytoplasmic domain bait in the yeast two-hybrid system.
Based on the demonstration that the interaction of 14-3-3 with target proteins is mediated by the recognition of a phosphoserine, we mutated serine 441 to alanine within this consensus binding site motif in the ␣4 cytoplasmic domain. Mutating serine 441 to alanine nearly completely abolished interaction of 14-3-3 with the ␣4 cytoplasmic domain bait in the yeast two-hybrid system (Fig. 1B). These results suggested that the high affinity binding of 14-3-3 to the ␣4 cytoplasmic domain is mediated by serine 441 within a specific consensus binding site motif.
14-3-3 Interacts with Recombinant ␣4 Subunits in tsA 201 Cells-To test directly if 14-3-3 could interact with full-length unassembled ␣4 subunits alone, we transfected tsA 201 cells with the ␣4 subunit cDNA. Because it has been demonstrated previously that phosphorylation of a consensus binding site motif for 14-3-3 greatly increases its affinity for the site, we also tested if activation of kinases would increase the interaction of 14-3-3 with the recombinant ␣4 subunit. We treated transfected cells coexpressing 14-3-3 and recombinant ␣4 subunits with forskolin (50 M) plus IBMX (1 mM) to activate PKA and PMA (0.1 M) to activate PKC. Activation of PKA (Fig. 2, 3rd lane) but not PKC (Fig. 2, 6th lane) was found to very significantly enhance the interaction of 14-3-3 with the ␣4 subunit. In the presence of the PKA inhibitor H-89 (30 M), the effect of forskolin was significantly attenuated (Fig. 2, 4th  Immunopurified proteins were fractionated by SDS-PAGE, and the top half of the blots were immunoblotted (IB) with the anti-␣4 subunit mAb and the bottom half with the anti-14-3-3 mAb. B, in a separate experiment, cells were transfected with the ␣4, ␤2 cDNAs and with or without the 14-3-3 cDNA. Cells were treated with or without forskolin (10 M), and the immunocomplexes were isolated with anti-␤2 subunit mAb beads. The final immunoblot analysis of all samples eluted from the beads was done the same way as described in Fig. 2A. C, an aliquot of the immunocomplexes tethered to the anti-␤2 mAb beads was separated into two pools, one of which was treated with (ϩ) and the other without (Ϫ) protein phosphatase 1 (PPI) for 1 h at 30°C. The final immunoblot analysis of all samples eluted from the beads was done the same way as described in Fig. 2A. consistent with the idea that PKA-dependent phosphorylation was involved in mediating the interaction.
The preceding results were obtained by immunoisolating recombinant ␣4␤2 AChRs with anti-␣4 mAb beads. Since this mAb binds both assembled and unassembled ␣4 subunits, we were unable to distinguish if 14-3-3 interacted with ␣4 subunits that were unassembled or assembled with the ␤2 subunits. To test if 14-3-3 could interact with assembled ␣4 subunits, ␣4␤2 complexes were isolated with the anti-␤2 mAb beads. The anti-␤2 mAb used (mAb 295) binds the conformationally mature ␤2 subunit only. Reactivity with denatured ␤2 subunits on immunoblots is not observed. We found that activation of PKA by forskolin significantly enhanced the interaction of 14-3-3 with ␣4␤2 AChR complexes immunoisolated with the anti-␤2 mAb beads both in the absence (Fig. 3B, 2nd lane) and the presence of exogenous 14-3-3 (Fig. 3B, 4th lane). We also observed low basal levels of interaction of 14-3-3 with ␣4␤2 AChR complexes even prior to activation of PKA by forskolin (Fig. 3B, 1st and 3rd lanes). These results suggested that both endogenous 14-3-3 and exogenous 14-3-3 associated with the ␣4␤2 AChR complexes. In addition, in three independent experiments we also observed that both in the presence or absence of exogenous 14-3-3, the amount of ␣4 subunits immunoprecipitated by the anti-␤2 mAb beads from detergent extracts of cells treated with forskolin was consistently higher than that from detergent extracts of cells not treated with forskolin. The increased amounts of the ␣4 subunits also correlated with greater amounts of 14-3-3 coimmunoprecipitated with the ␣4␤2 AChR complexes using the anti-␤2 mAb beads suggesting a possible role for 14-3-3 in altering the steady state levels of ␣4␤2 AChR complexes.
To demonstrate further that the interaction of 14-3-3 with the ␣4␤2 AChRs was phosphorylation-dependent, the detergent-solubilized ␣4␤2 AChR immunoisolated complexes from cells treated with forskolin and IBMX were treated with recombinant protein phosphatase I. We observed a significant reduction in the amount of 14-3-3 associated with immunopurified recombinant ␣4␤2 AChRs treated with protein phosphatase I compared with those treated with the buffer alone under identical conditions (Fig. 3C). These results further supported the fact that forskolin-dependent enhancement of the interaction of 14-3-3 with ␣4␤2 AChRs was due to a PKAmediated phosphorylation event.
14-3-3 Fails to Interact with Mutant ␣4 S441A ␤2 AChRs-We examined if 14-3-3 could associate with mutant ␣4 S441A ␤2 AChRs, in which serine 441 of the ␣4 subunit was mutated to alanine, following activation of PKA by forskolin in tsA201 cells. In keeping with the yeast two-hybrid mapping and mutagenesis studies, mutating serine 441 to alanine almost completely abolished interaction of 14-3-3 with mutant ␣4 S441A subunit (Fig. 2, 5th lane) and the mutant ␣4 S441A ␤2 AChRs (Fig. 3A, 5th lane; Fig. 3B, 5th to 8th lanes). These results, in conjunction with the results from previous studies of several other investigators demonstrating that a phosphoserine enhances the interaction of 14-3-3 with its target protein (15), suggested that serine 441 is the most likely target of PKA phosphorylation.
14-3-3 Stabilized the Wild-type ␣4 Subunit but Not the Mutant ␣4 S441A Subunit-Since 14-3-3 bound to the ␣4 subunit alone, we examined if 14-3-3, a chaperone protein, had a role in the early biogenesis of the ␣4 subunit. The ␣4 and the ␣4 S441A subunits were separately cotransfected with or without 14-3-3 cDNA into tsA201 cells. We studied the influence of the presence of 14-3-3 on the ␣4 or ␣4 S441A subunit steady state levels prior to, and following, activation of PKA by forskolin. To ensure differences were not simply due to variability in transfections between wells treated similarly, each condition was independently processed and analyzed in duplicate (indicated by a bar over the two lanes in Fig. 4). Prior to activation of PKA, the presence of 14-3-3 did not significantly alter the steady  4. 14-3-3 stabilizes the ␣4 wild-type subunit and not the mutant ␣4 S441A subunit. Cells were cotransfected with the ␣4 cDNA (1st to 4th lanes) or ␣4 S441A cDNA (5th to 8th lanes), with 14-3-3 cDNA (2nd, 4th, 6th, and 8th lanes) or without 14-3-3 cDNA (vector pEF6A alone) (1st, 3rd, 5th, and 7th lanes). The following day, the cells were treated with or without forskolin (Forsk, 10 M). The bar across the paired lanes represents samples treated under identical conditions from independently transfected and processed samples from a single experiment. Forty eight hours after transfection, the cells were lysed in 500 l of 2% Nonidet P-40 lysis buffer, and the proteins were solubilized for 3 h. The lysates were centrifuged, and 20 l of supernatant were fractionated by SDS-PAGE. The top half of the blots was immunoblotted (IB) with the anti-␣4 subunit mAb and the bottom half with the anti-14-3-3 mAb.

14-3-3 Stabilizes Nicotinic Receptor ␣4 Subunits kolin (50 M) for only 1 h instead of 24 h (data not shown).
These results indicated that a significant increase in the steady state levels of the ␣4 subunit occurred in the presence of 14-3-3 and following activation of PKA. These results, in conjunction with the increase in steady state levels of ␣4␤2 AChRs, support a role for 14-3-3 in regulating the stability of the ␣4 subunit through a PKA-dependent phosphorylation mechanism.
Higher Surface Expression of Wild-type ␣4␤2 AChRs Than Mutant ␣4 S441A ␤2 AChRs-To determine whether the interaction of 14-3-3 with the ␣4 subunits altered the cell surface expression levels of ␣4␤2 AChRs, we monitored these levels for both wild-type ␣4␤2 AChRs and mutant ␣4 S441A ␤2 AChRs using an enzyme-linked immunoassay. Because 14-3-3 interacts with ␣4␤2 AChRs following activation of PKA, we also examined if activation of PKA by forskolin (10 M) altered cell surface expression levels of wild-type ␣4␤2 AChRs and mutant ␣4 S441A ␤2 AChRs. The modified enzyme-linked immunoassay we used has been used previously to measure the surface expression of ␣4␤2 AChRs (17). In our assay we measured the relative amount of the ␤2 subunit in cells treated under the described conditions with an anti-␤2 subunit primary antibody. The amount of ␤2 immunoreactivity was then determined using an HRP-conjugated secondary Ab. The amount of secondary Ab bound to the primary mAb was then determined by measuring HRP enzymatic activity of the conjugated enzyme on a substrate (3,3Ј,5,5Ј-tetramethylbenzidine) whose product is colored blue and whose concentration can then be determined spectrophotometrically. As controls for nonspecific binding of the Abs, we used cells transfected with the vector alone. The surface expression of wild-type ␣4␤2 AChRs was found to be ϳ2-fold higher than the mutant ␣4 S441A ␤2 AChRs (Fig. 5). Following treatment with forskolin (10 M), the wild-type ␣4␤2 AChRs showed a small but statistically significant increase (ϳ20%, n ϭ 7, p Ͻ 0.005) in their cell surface expression levels. In contrast, forskolin did not induce a statistically significant change in the cell surface expression levels of mutant ␣4 S441A ␤2 AChRs. The 2-fold difference between the surface expression levels of the ␣4␤2 AChRs and the ␣4 S441A ␤2 AChRs was observed with two different preparations of cDNAs, making it very unlikely that it was due to differences in transfection efficiencies between the ␣4 subunit cDNA and the ␣4 S441A subunit cDNA due to differences in the quality of the DNA samples. Similar results in the absence of transfected exogenous 14-3-3 (data not shown) are in keeping with our findings that the endogenous 14-3-3 associated with ␣4␤2 AChRs under these conditions.
Immunohistochemical Localization of 14-3-3 and the ␣4 Subunit in Transfected Cells-We compared the distribution of 14-3-3 proteins with that of the ␣4 subunit at the single cell level in transfected cells treated with and without forskolin (10 M). Transfected cells were fixed with methanol/acetone and then sequentially immunostained for the ␣4 subunit followed by staining for 14-3-3 as described under "Experimental Procedures." Antibody binding was then visualized by confocal immunofluorescence microscopy using goat anti-mouse Alexa Fluor 546-conjugated Abs and the goat anti-rat Alexa Fluor 488-conjugated Abs. At the single cell level, diffuse immunostaining for the ␣4 subunit (red, top panel, Fig. 7) was observed throughout the ER/Golgi compartments and the surface membrane. In contrast staining for 14-3-3 was very distinctively different and was confined to the cytosolic region (green, top panel, Fig. 7). Some colocalization was evident in the merged images in the cytosolic region (yellow, top panel, Fig. 7). Following treatment with forskolin, colocalization within the ER/ Golgi compartments was significantly enhanced, but no significant colocalization was evident at the surface membrane (yellow, bottom panel, Fig. 7). As controls, we immunostained nontransfected cells and observed only very weak and diffuse staining with the anti-14-3-3 Ab to endogenous 14-3-3 proteins and no detectable nonspecific staining with the anti-␣4 mAb (not shown). Thus, the colocalization results complemented the coimmunoisolation results and suggested that 14-3-3 interacted with the ␣4 subunit and ␣4␤2 AChRs primarily within the ER/Golgi compartments of cells. It also complemented the results of our functional studies by showing that the reason why no significant change in the functional properties of ␣4␤2 AChRs was observed when they were coexpressed with 14-3-3 was most likely because 14-3-3 did not colocalize with surface ␣4␤2 AChRs.
Interaction of 14-3-3 with Native ␣4␤2 AChRs from Rat Brain-To validate the physiological importance of the interaction of 14-3-3 with the ␣4 subunit in yeast, and with recombinant ␣4␤2 AChRs in transfected cells, we determined if 14-3-3 is associated with native ␣4␤2 AChRs immunopurified from rat brain. Rat brain membranes were solubilized using 1% Nonidet P-40 and the ␣4␤2 AChRs immunopurified using anti-␣4 subunit-specific mAbs and anti-␤2 subunit-specific mAbs. Detergent-solubilized brain membrane extracts were also incubated with beads coupled to a control Ab (rat IgG). The interaction of 14-3-3 with ␣4␤2 AChRs was then detected by immu-noblotting with an anti-14-3-3 mAb. 14-3-3 was found to be associated with complexes of native Nonidet P-40-solubilized ␣4␤2 AChRs (Fig. 8). The significant 14-3-3 immunoreactivity detected with ␣4␤2 AChRs immunopurified with two different mAbs to the ␣4␤2 AChRs compared with the absence of any detectable immunoreactivity among proteins that bind nonspecifically to the control Ab suggested that the association with the complex was specific. This result also strongly supported the idea that the interaction of 14-3-3 with ␣4␤2 AChR is physiologically significant in vivo.

DISCUSSION
The cloning of a multitude of neuronal AChR subunit cDNAs has revealed a great diversity of AChR subtypes whose functions in the nervous system remain enigmatic (19). The large cytoplasmic domain between the third and fourth transmembrane domain is highly divergent among the subunits (20). Some aspects of the roles subserved by the large cytoplasmic domain such as the polarized trafficking of AChR in neurons (21) and the clustering of muscle AChRs at synaptic membrane subsites (22)(23)(24) are known.
Identification of proteins that interact with the cytoplasmic domain is likely to provide a better understanding of proteins involved in the subunit assembly, trafficking, clustering, and functions of AChRs. As a first step toward understanding which proteins interact with the widely expressed neuronal ␣4 AChRs, we used the ␣4 subunit cytoplasmic domain in a yeast two-hybrid screen. In this paper, we describe the identification of the first protein known to interact with the ␣4 subunit, 14-3-3, and the characterization of its interaction with recombinant and native ␣4␤2 AChRs. The results of our study provide novel mechanistic insights into the cellular events that mediate the interaction of 14-3-3 with the AChR ␣4 subunit following activation of PKA and the consequences of this interaction on the stability of the subunit.
14-3-3 proteins have previously been shown to bind to the sequence motif (RSXSX(P), where X ϭ any amino acid) in which phosphorylation of the second serine residue (underlined) is critical for the binding of 14-3-3 (33). The arginine residue at the first position appears to be essential, but other residues are tolerated at the position of the proline residue. The yeast two-hybrid mapping studies using nested deletions of the ␣4 cytoplasmic domain allowed us to localize a potential 14-3-3-binding site (Arg-Ser-Leu-Ser 441 -Val-Gln) between residues 413 and 450. By using site-directed mutagenesis we show that changing serine 441 to alanine nearly completely abolished the interaction of 14-3-3 with the mutated ␣4 bait. Interestingly, a second motif (RSRSIQ) closely resembling the consensus 14-3-3-binding site motif is also present in the ␣4 subunit between residues 459 and 464 but was not essential for interaction with 14-3-3.
We have provided several lines of evidence that suggests that serine 441 is phosphorylated by PKA. By treating transfected tsA201 cells with either forskolin to activate PKA, or PMA to activate PKC, we demonstrated that 14-3-3 binding to the FIG. 6. Functional properties of ␣4␤2 AChRs and ␣4 S441A ␤2 AChRs coexpressed with 14-3-3. Top, currents induced by ACh. Currents induced by 4-s applications of different concentrations of ACh are shown for Xenopus oocytes expressing ␣4␤2 AChRs, ␣4␤2 AChRs ϩ 14-3-3 (treated with forskolin (Forsk)), ␣4 S441A ␤2 AChRs, and ␣4 S441A ␤2 AChRs ϩ14-3-3 (treated with forskolin). The oocytes were clamped at a holding potential of Ϫ70 mV. ACh was applied successively following 4-min wash out periods following each application of ACh. Bottom, concentration/response curves of ACh. Data obtained from 2 to 3 oocytes held at Ϫ70 mV were normalized to the control response induced by 1 mM ACh, averaged, and fit using the Hill equation. The error bars represent the S.E. recombinant ␣4 subunit is most robust following activation of PKA and not PKC, and this effect is attenuated by the PKA blocker H-89. We also showed that there was a very significant reduction in the amount of 14-3-3 associated with the ␣4␤2 AChR complex following treatment with the protein phosphatase protein phosphatase I. Furthermore, 14-3-3 fails to interact with recombinant ␣4 S441A subunits alone, or ␣4 S441A ␤2 AChRs, following treatment of cells with forskolin. In addition, serine 441 of the ␣4 subunit is within a predicted PKA phosphorylation site (as determined by sequence analysis using the phosphobase program from CMS Molecular Biology Resource), and previous studies (15,34) have indicated that the binding of 14-3-3 to most of their other target proteins is mediated by a phosphoserine or phosphothreonine residue.
We observed that activation of PKA significantly enhanced the interaction of 14-3-3 with unassembled ␣4 subunits and with assembled ␣4␤2 AChR complexes. We have, however, failed to detect an increase in association of 14-3-3 with ␣4␤2 AChRs in tsA201 cells following acute or chronic (24 h) exposure of AChRs to nicotine (data not shown). These results suggest that other intracellular processes, other than channel activity, possibly govern the interaction of 14-3-3 with the ␣4 subunit and ␣4␤2 AChRs.
We have provided compelling evidence for a role of 14-3-3 in increasing the stability of the ␣4 subunit and ␣4␤2 AChR under conditions that also correlate well with those that favor interaction of 14-3-3 with the ␣4 subunit. When ␣4 subunits are expressed alone, the wild-type ␣4 and mutant ␣4 S441A subunits did not show significant differences in their steady state levels. However, we observed a very significant increase in the steady state levels of only the wild-type ␣4 subunit and not the mutant ␣4 S441A subunit following activation of PKA by forskolin only in the presence of 14-3-3. Corresponding differences in the steady state levels of the ␣4␤2 AChR and the ␣4 S441A ␤2 AChRs were also observed and strongly suggested that 14-3-3 plays a role in early posttranslational events that govern subunit and ␣4␤2 AChR stability.
The phosphorylation of the ␣4 subunit at serine 441 by PKA and its subsequent interaction with 14-3-3 alters cell surface ␣4␤2 AChRs by increasing the ␣4 subunit and ␣4␤2 AChR steady state levels. In keeping with such a role for 14-3-3, we observed a correlation between higher cell surface expression levels of wild-type ␣4␤2 AChRs and its ability to bind 14-3-3 and lower surface expression levels of the mutant ␣4 S441A ␤2 AChRs and their inability to bind 14-3-3. Furthermore, we observed a small but significant increase in their cell surface expression levels following treatment with forskolin. In contrast, forskolin did not induce a significant change in the cell surface expression levels of mutant ␣4 S441A ␤2 AChRs. Similar results in surface expression levels following treatment with forskolin were observed in the absence of exogenous 14-3-3 and were most probably due to the observed ability of endogenous 14-3-3 proteins to interact with ␣4␤2 AChRs.
Previously, it has been reported that activation of PKA by forskolin results in an ϳ200% increase in cell surface expression of recombinant human ␣4␤2 AChRs expressed in tsA201 cells (35). However, we do not observe such a large increase in surface expression of rat ␣4␤2 AChRs expressed in tsA201 cells. We suggest that this difference perhaps reflects differences in the growth conditions and species-specific differences (human versus rat) that might also affect the intrinsic efficiency of subunit assembly. The rather small but statistically significant increase (ϳ20%) in surface expression levels of the   FIG. 7. Colocalization of 14-3-3 and AChR ␣4 subunit in transfected tsA 201 cells. Transfected cells were fixed and processed for immunohistochemistry. Immunofluorescence was detected by confocal microscopy as described under "Experimental Procedures." Top panel, transfected with ␣4 ϩ ␤2 ϩ 14-3-3 cDNAs. Bottom panel, transfected with ␣4 ϩ ␤2 ϩ 14-3-3 cDNAs and treated with forskolin (10 M). The images are 1-m thick optical sections through single cells in culture. ␣4 immunoreactivity was visualized by the binding of Alexa Fluor 488conjugated secondary Abs (red), and 14-3-3 immunoreactivity was visualized by the binding of Alexa Fluor 546-conjugated secondary Abs (green). Superimposition of the two images is shown in the 3rd panel (yellow). Original magnification of all images was ϫ 600.
wild-type ␣4␤2 AChRs following treatment with forskolin is consistent with the idea that when subunit assembly was efficient, PKA-dependent phosphorylation only marginally contributes to further increases in surface expression.
The role of phosphorylation in regulating subunit assembly and cell surface expression is better characterized for muscletype AChRs (36 -43). In muscle-type AChRs, pulse-chase experiments and immunofluorescent microscopy indicate that AChR subunit assembly is complete in the ER following which AChR oligomers move rapidly through the Golgi membrane onto the plasma membrane (37). Interestingly, it has been demonstrated that both the ␥ and ␦ subunits are phosphorylated in vivo, and the ␦ subunits are more highly phosphorylated in the unassembled than in the assembled state indicating that phosphorylation precedes assembly and that phosphorylation/dephosphorylation mechanisms control the AChR subunit (36). Furthermore, using Torpedo AChR subunits expressed in mouse fibroblasts, it has been demonstrated previously that cAMP-induced increase in expression of cell surface AChRs is due to phosphorylation of the unassembled ␥ subunit assembly (37). But the underlying mechanism by which this phosphorylation increases the efficiency of subunit assembly and increased surface AChR expression has not been elucidated.
We have demonstrated that phosphorylation of the unassembled ␣4 subunit and the subsequent association of 14-3-3 with it increases its steady state levels in nonneuronal cells. This mechanism is consistent with such a proposed role for 14-3-3 in regulating the turnover of the plasma membrane H ϩ -ATPase (44). In addition, both PKA (45) and 14-3-3 isoforms (46) have been demonstrated previously to be localized appropriately to the ER/Golgi compartments to participate in such a process. Our results do not identify which exact isoform(s) of 14-3-3 is associated with the native ␣4 AChR subunit because the anti-14-3-3 mAb we used cross-reacts with several members of the 14-3-3 family. The family of 14-3-3 proteins consists of closely related members that do not show measurable differences in their affinities for a consensus binding site motif in vitro (33), although their binding in vivo is regulated by modulating their expression levels (47)(48)(49) and by phosphorylation of the 14-3-3 proteins themselves (50 -52). Thus, the identity of the particular isoform(s) of 14-3-3 that binds the native ␣4 AChR subunit in vivo remains to be determined. Because 14-3-3 can dimerize and thus simultaneously bind two different proteins, further experimentation will be needed to establish if other proteins are also involved in this process.
Finally, we would like to point out a possible pathophysiological significance of our work. It is well established that chronic intake of nicotine in smokers increases the expression levels of ␣4␤2 AChRs in their brains (53). Because schizophrenics are very heavy smokers, a recent study showing reduced levels of [ 3 H]nicotine-binding sites, mostly to ␣4␤2 AChRs (5), in the brains of schizophrenics compared with those of normal smokers (54) suggests that schizophrenics could have a possible defect in processes that regulate cell surface expression levels of their ␣4␤2 AChRs. However, it is likely that schizophrenics have deficits that are not limited to reduced levels of ␣4␤2 AChRs. Interestingly, recent genetic analyses of allelic frequencies of a variable number of tandem repeats in the 5Ј-noncoding region of the 14-3-3 gene suggests that it is a potential susceptibility gene for schizophrenia, particularly for early-onset schizophrenia (55). It has been reported previously that the 14-3-3 gene has a cAMP-response element-binding site in its promoter (56) and as such its expression levels are likely to be regulated by changes in cellular levels of cAMP through the activation of the transcription factor cAMP-re-sponse element-binding protein. Thus we speculate that 14-3-3 could have a broader role in regulating the excitability of neurons in an activity-dependent manner by modulating the levels of other proteins necessary for adaptive changes within specific neural networks. An understanding of the physiological significance of the interaction of 14-3-3 with native AChR ␣4 subunits might be better understood by modulation of the expression levels of 14-3-3 in vivo.