The Calcium Sensor Protein Visinin-like Protein-1 Modulates the Surface Expression and Agonist Sensitivity of the α4β2 Nicotinic Acetylcholine Receptor*

The calcium sensor protein visinin-like protein-1 (VILIP-1) was isolated from a brain cDNA yeast two-hybrid library using the large cytoplasmic domain of the α4 subunit as a bait. VILIP-1 is a myristoylated calcium sensor protein that contains three functional calcium binding EF-hand motifs. The α4 subunit residues 302–339 were found to be essential for the interaction with VILIP-1. VILIP-1 coimmunopurified with detergent-solubilized recombinant α4β2 acetylcholine receptors (AChRs) expressed in tsA201 cells and with native α4 AChRs isolated from brain. Coexpression of VILIP-1 with recombinant α4β2 AChRs up-regulated their surface expression levels ∼2-fold and increased their agonist sensitivity to acetylcholine ∼3-fold. The modulation of the recombinant α4β2 AChRs by VILIP-1 was attenuated in VILIP-1 mutants that lacked the ability to be myristoylated or to bind calcium. Collectively, these results suggest that VILIP-1 represents a novel modulator of α4β2 AChRs that increases their surface expression levels and agonist sensitivity in response to changes in the intracellular levels of calcium.

Neuronal nicotinic acetylcholine receptors (AChRs) 1 are members of a gene superfamily of ligand-gated ion channels. In vertebrates, neuronal AChRs are composed of subunits ␣2-␣10 and ␤2-␤4 (for review, see Ref. 1). AChR subunits have a large cytoplasmic domain between their third and fourth transmembrane domain whose amino acid sequence is highly divergent among the various subunits (2). The full functional unit of AChRs, like those of the N-methyl-D-aspartate receptors (3), is likely to include proteins that associate with this large cytoplasmic domain and modulate AChR functions.
To identify proteins associated with ␣4 AChRs, we used bait consisting of the large cytoplasmic domain of the ␣4 subunit to screen a mouse brain cDNA yeast two-hybrid library. In this paper, we describe the isolation of VILIP-1, a member of the visinin-like protein family of calcium sensor proteins, by such a screen. VILIP-1 is a member of a superfamily of neuronal calcium sensor proteins. This superfamily has been classified into the five subfamilies termed group I-V. The recoverins belong to group I, the frequenins and neuronal calcium sensor (NCS-1) to group II, the VILIPs, hippocalcin, and neurocalcins to group III, NCS-2 to group IV, and guanylyl cyclase-activating proteins and GC-inhibiting proteins to group V (for review, see Ref. 4).
The VILIP family is comprised of three members, VILIP-1, VILIP-2, and VILIP-3 (5)(6)(7)(8)(9)(10). The members of this family contain 4 EF-hand motifs, of which only EF-hand 2, 3, and 4 are thought to be functional because EF-hand 1 lacks two oxygencontaining side chain residues crucial for binding calcium. A glycine residue at the second position on the polypeptide chain is myristoylated. Interestingly, within most, but not all members of this calcium sensor protein family, the myristoyl moiety is sequestered and exposed through a rapid conformational change that unmasks it in response to alterations in cellular levels of calcium, thus facilitating its association with membranes (11,12). This calcium-dependent conformational switch has been termed the "calcium-myristoyl switch." In this paper we show that VILIP-1 interacts with both native and recombinant ␣4␤2 AChRs. Coexpression of VILIP-1 with recombinant ␣4␤2 AChRs increased their surface expression ϳ2-fold. This effect was significantly diminished for VILIP-1 mutants that are deficient in their ability to be myristoylated or bind calcium. Coexpression of VILIP-1 with recombinant ␣4␤2 AChRs also increased their agonist-sensitivity ϳ3-fold. Collectively, these results support a novel role for VILIP-1 as an AChR-associated protein that modulates the surface expression levels and functional properties of ␣4␤2 AChRs in response to changes in the intracellular levels of calcium.
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 GTC AGT GTC-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 C-terminal 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, and 5Ј-GGG GGA TCC TCA GTT GTC TTT GAC CAC AGA GGG-3Ј. The N-terminal deletion was generated using the forward primer 5Ј-GGG GAA TTC TGC CGG AGA CTT ATT GAG TCC-3Ј and the reverse primer 5Ј-CCC GGA TCC TCA CTT CAC CGA GAA GTC AGT GTC-3Ј. The PCR products were subsequently subcloned into the EcoRI-BamHI sites of the pLexA. The ␣4 and ␤2 subunit cDNAs were generated by amplification of the full clone by PCR. We used the forward primer 5Ј-GGG AAT TCG CCA CCA TGG CCA ATT CGG GCC CCG GG-3Ј and reverse primer GGG TCT AGA GCA AGC AGC CAG CCA GGG AGG for the ␣4 subunit and the forward primer 5Ј-GGG AAT TCG CCA CCA TGC TGG CTT GCA TGG CCG GG-3Ј and reverse primer 5Ј-GGG TCT AGA CTT GGA GCT GGG AGC TGA GTG-3Ј for the ␤2 subunit. These products were then ligated into the EcoRI-XbaI site of the mammalian cell expression vector pEF6/myc-His A (Invitrogen). All DNA sequence analysis was done using the Thermo-Sequenase-radiolabeled terminator cycle sequencing Kit (Amersham Biosciences). The wild-type VILIP-1 cDNA was amplified from mouse total brain cDNAs using the forward primer 5Ј-GGG GAT CCG CCA CCA TGG GGA AAC AGA ATA GCA AA-3Ј and reverse primer 5Ј-GGG TCT AGA TCA TTT CTG AAT GTC ACA CTG CAG-3Ј and the PCRamplified cDNA ligated into the BamHI-XbaI sites of the mammalian cell expression vector pEF6/myc-HisA (Invitrogen). The expressed protein lacked the Myc-His tag because of the presence of the endogenous stop codon present in each cloned cDNA. Mutagenesis to generate mutations Asp-81 to Val in the EF-hand motif 2, Thr-117 to Ala in the EF-hand motif 3, Thr-167 to Ala in EF-hand motif 4 in the triple mutant (mVILIP-2,3,4EF), and Gly-2 to Ala (mVILIP-1-myr) used the QuikChange site-directed mutagenesis kit (Stratagene). Mutagenesis was performed using the primers 5Ј-GAT GGC ACC ATC GTT TTC CGA GAG TTC-3Ј and 5Ј-GAA CTC TCG GAA AAC GAT GGT GCC ATC-3Ј to mutagenize residue Asp-81 to Val in the EF-hand motif 2, 5Ј-GGT GAC GGC AAG ATC GCC CGA GTG GAG ATG CTG G-3Ј and 5Ј-CCA GCA TCT CCA CTC GGG CGA TCT TGC CGT CAC C-3Ј to mutagenize residue Thr-117 to Ala in the EF-hand motif 3, and 5Ј-GAA CAA AGA TGA CCA GAT TGC ACT GGA TGA ATT CAA AGA AGC TGC-3Ј and 5Ј-GCA GCT TCT TTG AAT TCA TCC AGT GCA ATC TGG TCA TCT TTG TTC-3Ј to mutagenize residue Thr-167 to Ala in EFhand motif 4. The triple mutant (mVILIP-2,3,4EF) was generated by sequential mutagenesis using these primers. The non-myristoylatable mVILIP-myr mutant was generated by mutagenesis of residue glycine 2 to alanine using the primers 5Ј-GGA TCC GCC ACC ATG GCG AAA CAG AAT AGC AAA CTG G-3Ј and 5Ј-CCA GTT TGC TAT TCT GTT TCG CCA TGG TGG CGG ATC C-3Ј. The full characterization of the functional and structural properties of these mutants has been described elsewhere (13).
Yeast Two-hybrid Library Screen-Yeast two-hybrid screens were carried out according to a standard protocol (Clontech). 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 10 ϫ 10 6 yeast cells cotransformed with the bait, and cDNAs from a pre-made mouse brain cDNA Matchmaker LexA library (Clontech) 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 Biosciences). Two clones of VILIP-1 that interacted with the ␣4 bait were characterized. These clones contained slightly different cDNA sizes as determined from limited sequence analysis of their 5Ј ends.
Antibodies-The anti-VILIP-1 rabbit antiserum used for the immunoblots has been previously described (14). Protein G affinity-purified mAbs to the ␣4 subunit (mAb 299) and the ␤2 subunit (mAb 295) were generously provided by Dr. Lindstrom (University of Pennsylvania, Philadelphia, PA). The goat anti-rat and anti-rabbit horseradish peroxidase-conjugated Abs were obtained from Pierce. The Alexa Fluor 488 and Alexa Fluor 546-labeled goat anti-rat or anti-rabbit secondary Abs were obtained from Molecular Probes, Eugene, OR. mAbs including the rat IgG were coupled to Actigel ALD beads at a concentration of 0.5 mg/ml gel using the manufacturer's instructions (Sterogene Bioseparations Inc., Carlsbad, CA).
Expression of Recombinant AChRs in Human Embryonic Kidney tsA201 Cells-Human tsA201 cells, a derivative of human embryonic kidney cell line 293 were cultured in 6-well plates in Dulbecco's modified Eagle's medium (Invitrogen) supplemented with 10% fetal bovine serum, 50 g/ml penicillin, and 50 g/ml streptomycin at 37°C. Cells were transfected using LipofectAMINE 2000 (Invitrogen) at 90% confluency (ϳ10 6 cells/well) with various combinations of cDNAs as per the manufacturer's instructions. Little ␣4␤2 AChR surface expression was detected after transfection if the cells were incubated at 37°C for up to 72 h. Hence, cells were typically utilized after incubation at 30°C for 24 -48 h. This effect of lowering the temperature to increase recombinant ␣4␤2 AChR subunit assembly efficiency has been previously reported (15,16) and was found to be critical to get detectable levels of AChR surface expression. Hence, cells were treated ϳ16 h after transfection with or without forskolin (10 M) or the PKA inhibitor H-89 (30 M) and incubated at 30°C for ϳ24 h and then for an additional 24 h at a physiological temperature of 37°C before measuring surface AChRs.
Immunoisolation of Native AChRs from Rat Brain and Immunoblot Analysis-Blots containing SDS-PAGE-separated proteins from previously described immunoisolations (15) were reused. Briefly, detergentsolubilized brain extracts (typically 10 ml) were incubated with ϳ20 l of mAb-coupled Actigel ALD bead, and the proteins bound to the Ab beads were eluted with protein sample buffer and fractionated by SDS-PAGE. Membranes containing electroblotted proteins were incubated with diluted (typically 1:5000) primary Abs in phosphate-buffered saline solution containing 0.1% Tween and 5% nonfat 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). The membrane was cut in half, the top half was probed with the anti-␣4 mAb and anti-␤2 mAb, and the bottom half was probed with the anti-VILIP-1 antiserum.
Enzyme-linked Immunoassay for Quantitating Cell Surface AChRs-Cell surface ␣4␤2 AChRs were measured using the methodology previously described (15,16). Briefly, transfected tsA201 cells plated in 12-well plates (0.5 ϫ 10 6 cells/well) and expressing AChRs were washed once in PBS and blocked with PBS containing 2% bovine serum albumin. The cells were incubated for 1 h with an anti-␤2 mAb (295) in PBS containing 2% bovine serum albumin at room temperature. After four washes with PBS, the cells were fixed with formaldehyde (3%) for 10 min and washed 3 times with PBS. The cells were then incubated with horseradish peroxidase-conjugated goat anti-rat secondary Ab for 1 h and washed 6 times and incubated with 500 l of the horseradish peroxidase 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) but 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 previously described (17). Oocytes were injected with cRNAs of the ␣4 and ␤2 subunits (20 ng/subunit) and of VILIP-1 (80 ng) per oocyte and incubated for 3-7 days at 16 -18°C in 50% L-15 medium (Invitrogen) containing 10 mM HEPES buffer, pH 7.5.
Electrophysiological Recordings-Currents were measured using a standard two-microelectrode voltage-clamp amplifier (Oocyte Clamp OC-725C) as previously described (17). Electrodes were filled with 3 M KCl and had resistances of 1.0 -2.0 megaohms for the voltage electrode and 0.5-0.1 megaohm for the current electrode. All records were digitized at 200 Hz with MacLab software and hardware (AD Instruments). Data was 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 CaCl2, 1 mM MgCl2, 5 mM HEPES, pH 7.6).
Primary Cultures of Neurons-Primary cultures of rat cerebellar granule cells were isolated from postnatal day 6 -8 male or female Sprague-Dawley rat pups. After rapid dissection, the cerebellum was immediately immersed in ice-cold calcium-magnesium-free PBS. The tissue was spun at 150 ϫ g for 2 min, the supernatant was gently removed, and cells were dissociated by enzymatic treatment with DNase in calcium-magnesium-free PBS and by repeated trituration through a series of decreasing diameter fire-polished Pasteur pipettes. Cells were resuspended in culture medium consisting of Dulbecco's modified Eagle's medium supplemented with 10% heat-inactivated horse serum, 2 mM glutamine, 1ϫ insulin-transferrin-selenium-S supplement (Invitrogen), 20 mM KCl, 100 IU/ml penicillin, and 100 mg/ml streptomycin. Cells were plated at a density of 2-3 ϫ 10 6 cells/35-mm dish precoated with poly-L-lysine. Ara-C (10 M) was added to the culture medium 16 -18 h after plating to prevent the proliferation of non-neuronal cells. The cells were incubated at 37°C in 5% CO2 atmosphere for 6 -7 days before being used in the experiments.
Immunohistochemistry-Cultured neurons were fixed with 100% methanol for 5 min, washed 3 times with 2 ml of PBS, and blocked using PBS containing 2% bovine serum albumin for 30 min. Cells were then incubated simultaneously with diluted anti-␣4 subunit mAb (1/2000 dilution) and the anti-VILIP-1 antiserum (1/1000 dilution) in PBS containing 4% normal goat serum with gentle shaking overnight at 4°C. Cells were washed 3 times for 15 min/wash in PBS and incubated with diluted (1/1000) fluorophore-conjugated secondary Ab in PBS containing 2% bovine serum albumin and 4% normal goat serum 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.
Immunofluorescence Microscopy-Fluorescence microscopy was accomplished using a Leica DM RXA automated upright deconvolution microscope. Cells were scanned using 0.2-m steps in the z axis, alternating between fluorescein isothiocyanate and rhodamine filters. The resulting optical section images were captured on Silicon Graphics workstations. Images were pseudocolored and processed using Adobe Photoshop software.

RESULTS
Isolation of VILIP-1 from a Brain Yeast Two-hybrid cDNA Library-We used the large cytoplasmic domain (residues 302-561) of the rat AChR ␣4 subunit as bait in the yeast two-hybrid system to screen ϳ10 ϫ 10 6 of a mouse brain lexA cDNA library. Unique clones thus identified were subject to limited nucleotide sequence (ϳ100 -150 nucleotides) analysis. By homology searches of established nucleotide databases, we identified VILIP-1 as a putative interacting protein.
Residues 302-339 of the ␣4 Subunit Are Essential for the Interaction of VILIP-1 with the ␣4 Cytoplasmic Domain-We used bait containing N-terminal and nested C-terminal deletions of the ␣4 cytoplasmic domain to test their ability to interact with VILIP-1 in the yeast two-hybrid system. We found using the nested C-terminal deletions that residues 302-339 of the ␣4 subunit loop are required for the interaction with VILIP-1. We verified this finding by using bait in which the N-terminal residues 302-339 were deleted and found that deletion of these residues was sufficient to abolish interaction of this bait with VILIP (Fig. 1).
VILIP-1 Interacts with Recombinant ␣4␤2 AChRs in Transfected tsA201 Cells-To examine whether VILIP-1 could associate with recombinant ␣4␤2 AChRs, we coexpressed VILIP-1 with recombinant ␣4␤2 AChRs by transfecting tsA201 cells with their respective cDNAs. We immunoisolated 1% Nonidet P-40-solubilized recombinant ␣4␤2 AChRs from tsA201 cells 48 h after transfection using anti-␤2 subunit mAb-coupled beads. Proteins eluted from these beads were fractionated by SDS-PAGE, and the membrane containing the electroblotted proteins was probed for VILIP-1 using a polyclonal antiserum to VILIP-1. As controls for nonspecific binding of proteins to beads, we used rat IgG-coupled beads. We found that VILIP-1 coimmunoisolated with recombinant ␣4␤2 AChRs, and no immunoreactivity for VILIP-1 was observed with the control rat IgG beads (Fig. 2). The lysates represent ϳ1/250 of the total solubilized protein used in each of the coimmunopurifications. The relatively weaker signal from VILIP-1 compared with the ␣4 and ␤2 signals in the coimmunoprecipitation experiments suggests that only a small fraction of the total VILIP-1 present is associated with the ␣4␤2 AChR complexes under these conditions. It is likely that the detergent used to solubilize the ␣4␤2 AChR complexes significantly affects protein-protein interactions, including that of VILIP-1, with ␣4␤2 AChRs.
To further study if myristoylation and the calcium binding EF-hand motifs were required for the association of VILIP-1 with the ␣4␤2 AChRs, we generated two VILIP-1 mutants. The first mutant, mVILIP-myr, was not myristoylated because the residue at which it is myristoylated (Gly-2) was mutated to Ala. VILIP-1 has four EF-hand motifs, and their structure consists of two ␣-helical segments bridged by a calcium binding loop. The loop is formed by 12-amino acid stretches, 6 of which (designated as X, Y, Z, ϪY, ϪX, ϪZ at positions 1, 3, 5, 7, 9, and 12) participate in the coordination of one calcium ion. The N-terminal-most EF-hand motif 1 has substitutions at the conserved positions that are essential for binding calcium ions. and hence. only three of the four EF-hand motifs are thought to be capable of binding calcium ions. Thus, the second mutant generated was mutated at position ϪX within each of the three functional calcium binding EF-hands (D81V/T117A/T167A) to yield the triple mutant mVILIP-2,3,4EF. Both these mutants, mVILIP-myr and mVILIP-2,3,4EF, also interacted with recombinant ␣4␤2 AChRs albeit with some apparent loss in association (as inferred from the intensity of band on immunoblots in Fig. 2). This result showed that calcium is not essential for the binding of VILIP-1 to ␣4␤2 AChRs. This result is similar to that found for other members of this superfamily of calcium sensor proteins (KChIPs), which also exhibit a lack of calcium dependence for their association with A-type K ϩ channels (18).
Coexpression of VILIP-1 with ␣4␤2 AChRs Increases Their Surface Expression-We examined the cell surface expression of ␣4␤2 AChRs coexpressed with VILIP-1 by measuring their surface expression levels using a previously described enzymelinked immunoassay (15,16). As controls for nonspecific binding of mAbs, we used cells transfected with the vector alone. The surface expression of wild-type ␣4␤2 AChRs coexpressed with VILIP-1 was found to be increased ϳ2-fold over those cells expressing ␣4␤2 AChRs alone. We also examined the ability of both mutants mVILIP-myr and mVILIP-2,3,4EF to increase surface expression of ␣4␤2 AChRs and found that both mVILIP-myr and mVILIP-2,3,4EF did not significantly change the expression levels of ␣4␤2 AChRs when coexpressed with them in tsA201 cells (Fig. 3).
The Increase in Surface Expression of ␣4␤2 AChRs by VILIP-1 Is Not Due to Changes in cAMP Levels or Activation of PKA-VILIP-1 has previously been shown to cause small changes in the intracellular levels of cAMP when expressed as a recombinant protein in cells (19). We attempted to determine if this relatively small increase in cAMP levels or the subsequent activation of PKA could be responsible for the up-regulation of ␣4␤2 AChRs by VILIP-1. We measured the surface expression of ␣4␤2 AChRs in the presence and absence of VILIP-1 in cells pretreated with and without forskolin (10 M) or with and without the PKA inhibitor H89 (30 M). We observed that treatment of cells with forskolin alone up-regulated ␣4␤2 AChR surface expression as previously described (20). In cells coexpressing VILIP-1, VILIP-1 also increased ␣4␤2 AChR surface expression in addition to that observed by forskolin alone (Fig. 4, center panel). The PKA inhibitor H89 failed to block the ability of VILIP-1 to up-regulate ␣4␤2 AChRs (Fig. 4,  right panel). These results support our conclusion that the up-regulation of ␣4␤2 AChR by VILIP-1 occurs through a mechanism that is distinct from that attributed to changes in cellular levels of cAMP or the activation of PKA-dependent phosphorylation.
VILIP-1 Coimmunoprecipitates with Native Detergent-solubilized ␣4␤2 AChRs from Brain-To investigate the physiolog- AChRs were determined in cells coexpressing wild-type VILIP-1 and mutants mVILIP-myr and mVILIP-2,3,4EF. The relative amount of primary anti-␤2 subunit mAb bound to the surface AChRs was quantitated using a horseradish peroxidase-conjugated secondary Ab in conjunction with the horseradish peroxidase substrate (3,3Ј,5,5Ј-tetramethylbenzidine) in a colorimetric assay as described under "Experimental Procedures." The bar graphs represent the normalized levels of AChRs after subtraction of the mean background value obtained from cells transfected with the vector alone. Each measurement was done in duplicate. The error bars represent the S.E. of measurements from three separate experiments.
ical importance of the interaction of VILIP-1 with the ␣4 subunit in yeast cells and with recombinant ␣4␤2 AChRs in transfected cells, we determined if VILIP-1 was present in native ␣4␤2 AChR complexes immunoisolated from rat brain. A previously prepared immunoblot (15) generated by immunoisolating ␣4␤2 AChRs with anti-␣4 mAb-and anti-␤2-mAbcoupled beads and rat IgG Ab (as a control) from 1% Nonidet P-40-solubilized rat brain membranes was probed using an anti-VILIP-1 antiserum. The protein lysate represents ϳ1/5,000 of the total solubilized protein used in the immunopurifications. The absence of signals for the ␣4 subunit and the ␤2 subunit compared with the presence of a signal for VILIP-1 most likely reflects their relative abundance in the detergentsolubilized membrane extracts rather than significant differences in the affinities of the Ab probes for their targets. Our ability to detect immunoreactivity to VILIP-1 on these blots containing immunopurified native ␣4␤2 AChRs complexes (Fig. 6) suggests that VILIP-1 is present in these complexes. As previously mentioned for the recombinant ␣4␤2 AChRs complexes, the relatively weaker signal of VILIP-1 compared with that of the ␣4 and ␤2 signals in the coimmunoprecipitation experiments suggests that a relatively small fraction of the total VILIP-1 is associated with the ␣4␤2 AChR complexes. This is most likely in part due to the deleterious effects of the detergent on the interaction of VILIP-1 with ␣4␤2 AChRs. This explanation would be consistent with our expectation that a multitude of proteins involved in AChR assembly, trafficking, membrane clustering, turnover, and functional modulation must interact with them but are not detected in stoichiometric amounts by silver staining of the proteins even on large scale affinity purification of detergent-solubilized AChRs. 2 Notably, ␣4, ␤2, and VILIP-1 immunoreactivity was absent in immunoisolates with the control rat IgG Ab, suggesting that the presence of VILIP-1 was not due to its nonspecific binding to the beads or to the coupled IgG Ab.
Colocalization of VILIP-1 and ␣4 AChR Subunits in Cultured Cerebellar Granule Cells-Because both ␣4 AChRs (21) and VILIP-1 (9) are expressed in cerebellar granule cells and because of the relative ease with which one can culture these cells, we used them to examine the subcellular distribution of VILIP-1 and AChR ␣4 subunits. One-week-old primary cerebellar granule cell cultures were fixed with 100% methanol and immunostained for VILIP-1 using an anti-VILIP-1 rabbit antiserum and an anti-␣4 subunit rat mAb as described under "Experimental Procedures." Binding of the secondary goat antirat Alexa Fluor 488-conjugated Abs and the goat anti-rabbit Alexa Fluor 546-conjugated Abs was visualized by immunofluorescence deconvolution microscopy. Staining for VILIP-1 was observed primarily near the cytoskeletal matrix and at the cell surface membrane (red) (Fig. 7, left panel). Staining for the ␣4 subunit was observed both at the cell surface membrane and within the cytosol (green) (Fig. 7, middle panel), possibly representing AChRs in endoplasmic reticulum/Golgi compartments. Partial colocalization (yellow) (Fig. 7, right panel) was apparent only at discrete locations at the cell surface mem-2 R. Anand, unpublished observations.

FIG. 4. VILIP-1 up-regulates ␣4␤2
AChRs through a non-cAMP-dependent mechanism. The surface expression levels of ␣4␤2 AChRs was determined in cells coexpressing VILIP-1, mVILIP-myr, and mVILIP-2,3,4EF. The cells were untreated (Control) or pretreated with forskolin (10 M) and the PKA inhibitor H-89 (30 M), and the AChR surface expression levels were measured after 48 h. The measurements in each experiment were done in duplicate, and the error bars represent the S.E. of measurement from at least two separate experiments.

FIG. 5. VILIP-1 increases the agonist-sensitivity of ␣4␤2
AChRs expressed in oocytes. ␣4␤2 AChRs were expressed with and without wild-type VILIP-1 or the mutants VILIP-myr and mVILIP-2,3,4EF from in vitro transcribed cRNAs in Xenopus oocytes. The oocytes were clamped at a holding potential of Ϫ70 mV. Currents elicited by a 4-s application of different concentrations of ACh were recorded with a 4-min wash-out period between each application. Data obtained from 3-4 oocytes from at least two independent experiments were normalized to the control response elicited by 1 mM ACh, averaged, and fit using the Hill equation. The error bars represent the S.E. ACh activated ␣4␤2 AChRs with an EC 50 ϭ 32 Ϯ 7 M (n H ϭ 0.95), ␣4␤2 AChRs ϩ VILIP-1 with an EC 50 ϭ 13 Ϯ 3 M (n H ϭ 0.95), ␣4␤2 AChRs ϩ mVILIP-myr with an EC 50 ϭ 26 Ϯ 8 M (n H ϭ 0.6), and ␣4␤2 AChRs ϩ mVILIP-2,3,4EF with an EC 50 ϭ 7 Ϯ 2 M (n H ϭ 0.46). brane and not within the cytosolic regions. No detectable staining of the secondary Abs to these cultured cells was observed when the primary Abs were omitted (not shown). Thus, these results complement the coimmunoisolation results with recombinant and native ␣4␤2 AChRs and the functional studies and suggested a novel role for VILIP-1 in modulating ␣4␤2 AChR function.

DISCUSSION
It is becoming increasingly evident that calcium binding and calcium sensor proteins play a significant role in coupling changes in intracellular levels of calcium to modulation of different types of voltage-and ligand-gated channels. Recently, different members of the superfamily of calcium sensor proteins such as KChIP (18), frequenin (22), and NCS-1 (23) were shown to modulate the A-type K ϩ channels. Similarly, modulation of voltage-gated calcium channels by NCS-1 was recently reported (24). Another calcium-binding protein, calmodulin, has a well established role in regulating the functional properties of N-methyl-D-aspartate receptors (25,26) and voltagegated calcium channels (27,28). However, despite the well recognized permeability of neuronal AChRs to calcium, it is not known if AChRs are also regulated by calcium sensor proteins.
In a broad yeast two-hybrid screen designed to identify cytosolic proteins that interact with ␣4 AChRs, we discovered the protein VILIP-1. In this paper we presented experimental evidence for the modulation of ␣4␤2 AChRs by the calcium sensor protein VILIP-1. VILIP-1 coimmunoprecipitates with both recombinant and native ␣4␤2 AChRs and shows partial colocalization with native ␣4 subunits in cultured cerebellar granule cells. Coexpression of VILIP-1 with ␣4␤2 AChRs in tsA201 cells increases the surface expression of ␣4␤2 AChRs ϳ2-fold. Coexpression of VILIP-1 with recombinant ␣4␤2 AChRs increases their agonist sensitivity ϳ3-fold. These results suggest a novel mechanism by which changes in intracellular levels of calcium can alter the expression levels and functional properties of AChRs.
The surface expression of ␣4 AChRs can be modulated by different mechanisms. The binding of nicotine to ␣4 AChRs up-regulates their surface expression levels both in vitro (20, 29 -31) and in vivo (32)(33)(34)(35). Our results illustrate a novel mechanism by which ␣4␤2 AChR surface expression can be upregulated by an endogenous cytosolic protein. The exact mechanism by which the association of VILIP-1 with ␣4␤2 AChRs increases their surface expression remains to be established. However, we have provided experimental evidence that the previously reported ability of recombinant VILIP-1 to increase basal levels of cAMP or the coupled PKA activity is not responsible for the increase in ␣4␤2 AChR surface expression. This is because we observe that VILIP-1 increases the surface expression of ␣4␤2 AChR even in cells that have been pretreated with forskolin and in cells that have been pretreated with the PKA inhibitor H-89. Because we observe little detectable colocalization of VILIP-1 with ␣4 subunits in the endoplasmic reticulum/ Golgi membranes, it is unlikely that VILIP-1 is involved in the trafficking of ␣4 AChRs to the surface membrane. The colocalization at the surface membrane instead favors the possibility that VILIP-1 alters the turnover of ␣4 AChRs at the surface membrane when it associates with them after the activation of its calcium-myristoyl switch and translocation to the membrane in response to changes in intracellular levels of calcium. The ability of VILIP-1 to increase the surface expression of membrane proteins it associates with is conserved among other members of this superfamily because KChIP also increases the surface density of A-type K ϩ channels (18).
The VILIP-1 mutants lacking the myristoyl moiety and the functional EF-hand motifs were attenuated in their ability to modulate the ␣4␤2 AChR response. This was inferred from the Hill fits that gave significantly lower n H for both single-site and two-site fits, suggesting that the heterogeneous responses were from modulated and unmodulated ␣4␤2 AChRs, possibly because of the loss in each of the VILIP-1 mutant's affinity for the ␣4␤2 AChR. We observed that ACh activated ␣4␤2 AChR coexpressed with mVILIP-2,3,4EF with an EC 50 that was significantly lower than the EC 50 for activating ␣4␤2 AChR coex- Calcium ions entering through activated ␣4␤2 AChRs directly or indirectly trigger the calcium-myristoyl switch of VILIP-1. Exposure of the myristoyl moiety leads to recruitment of cytosolic VILIP-1 to the membrane. Association of VILIP-1 with ␣4␤2 AChRs changes their constitutive turnover rate and their agonist sensitivity.
FIG. 6. VILIP-1 coimmunoprecipitates with native detergentsolubilized ␣4␤2 AChRs from brain. 1% Nonidet P-40 detergentsolubilized ␣4␤2 AChRs were coimmunopurified from rat brain. Proteins eluted from specific mAb beads (mAb 299 to the ␣4 subunit; mAb 295 to the ␤2 subunit) and control Ab beads (rat IgG) were fractionated by SDS-PAGE and immunoblotted with the anti-VILIP-1 antiserum, anti-␣4 mAb, and anti-␤2 antiserum. The lysates represent ϳ1/5000 of the total solubilized protein used in each of the coimmunopurifications. pressed with mVILIP-myr. We have recently observed that mVILIP-myr is poorly associated with membranes even in the presence of calcium in contrast to the association of mVILIP-2,3,4EF with membranes both in the presence and absence of calcium (13), possibly because the myristoyl moiety is constitutive exposed due to structural changes caused by mutations in the EF-hand motifs. Based on these results we concluded that the calcium binding EF-hand motifs of VILIP-1 have a structural role in triggering the calcium-myristoyl switch of VILIP-1, but its ability to remain associated with membranes critically requires the presence of the myristoyl moiety. Thus, the differences between the EC 50 values for activation of the ␣4␤2 AChR coexpressed with the two VILIP-1 mutants most likely reflected individual differences in their ability to associate with membranes (loss for mVILIP-myr and retention for mVILIP-2,3,4EF).
Neuronal AChRs in the central nervous system are the primary mediators of addiction to nicotine. Addiction to nicotine due to repetitive activation of AChRs is thought to change the functional properties of the AChRs themselves and the functional circuitry of the central nervous system by a complex sequence of molecular and cellular processes similar to those activated by other drugs of addiction (36 -38). Because AChRs exhibit significant permeability to calcium, it is reasonable to expect that some of these changes are effected via calcium entry through nicotine-gated AChRs. It was recently demonstrated that chronic low doses of nicotine can also up-regulate recombinant ␣4␤2 AChR function by increasing their agonist sensitivity (39). Interestingly, this result was attributed to a possible interaction of ␣4␤2 AChRs with cytosolic proteins in the host cells (39). However, little is known about the endogenous intracellular proteins that can increase the agonist sensitivity of ␣4␤2 AChRs. The ability of VILIP-1 to increase the agonist sensitivity of ␣4␤2 AChRs could account for the observations of that chronic nicotine alters ␣4␤2 AChR function. The increased agonist sensitivity of ␣4␤2 AChRs may represent one mechanism by which chronic exposure to low doses of nicotine (as occurs in the central nervous system neurons of nicotine addicts) augments synaptic transmission involving ␣4␤2 AChRs where they, in turn, modulate the release of multiple neurotransmitters including dopamine.
The illustration in Fig. 8 shows a model that incorporates key features of our findings. In this model, repetitive exposure to nicotine would increase calcium entry into neurons due to activation of nicotine-gated AChRs, which in turn would trigger the calcium-myristoyl switch of VILIP-1 and cause it to translocate to the membrane in the vicinity of activated AChRs. The association of VILIP-1 with the AChRs would have two consequences, 1) to increase AChR levels possibly by decreasing their turnover and 2) to stabilize them in a state with higher agonist sensitivity. It is possible that behaviors that lead to accumulation of AChRs in this state, including smoking tobacco, become reinforced because of subsequent functional adaptations of neural networks to functioning with AChRs that have higher agonist sensitivity.
We have previously demonstrated that a chaperone protein identified as 14-3-3 has a dynamic role in regulating the expression levels of ␣4␤2 AChRs through a phosphorylation-dependent interaction with the ␣4 subunit (15). In addition, an endogenous prototoxin identified as lynx1 was shown to decrease the agonist sensitivity of ␣4␤2 AChRs and slow their recovery from desensitization (40). The ability of VILIP-1 to modulate both the surface expression and the agonist sensitiv-ity of ␣4␤2 AChRs adds it a growing list of accessory proteins that associate with neuronal AChRs and modulate their biogenesis and functions.