Protein Kinase C Phosphorylation Regulates Membrane Insertion of GABAA Receptor Subtypes That Mediate Tonic Inhibition*

Tonic inhibition in the brain is mediated largely by specialized populations of extrasynaptic receptors, γ-aminobutyric acid receptors (GABAARs). In the dentate gyrus region of the hippocampus, tonic inhibition is mediated primarily by GABAAR subtypes assembled from α4β2/3 with or without the δ subunit. Although the gating of these receptors is subject to dynamic modulation by agents such as anesthetics, barbiturates, and neurosteroids, the cellular mechanisms neurons use to regulate their accumulation on the neuronal plasma membrane remain to be determined. Using immunoprecipitation coupled with metabolic labeling, we demonstrate that the α4 subunit is phosphorylated at Ser443 by protein kinase C (PKC) in expression systems and hippocampal slices. In addition, the β3 subunit is phosphorylated on serine residues 408/409 by PKC activity, whereas the δ subunit did not appear to be a PKC substrate. We further demonstrate that the PKC-dependent increase of the cell surface expression of α4 subunit-containing GABAARs is dependent on Ser443. Mechanistically, phosphorylation of Ser443 acts to increase the stability of the α4 subunit within the endoplasmic reticulum, thereby increasing the rate of receptor insertion into the plasma membrane. Finally, we show that phosphorylation of Ser443 increases the activity of α4 subunit-containing GABAARs by preventing current run-down. These results suggest that PKC-dependent phosphorylation of the α4 subunit plays a significant role in enhancing the cell surface stability and activity of GABAAR subtypes that mediate tonic inhibition.

␥-Aminobutyric acid type A receptors (GABA A Rs) 2 constitute the major inhibitory ligand-gated receptors in the adult central nervous system and are responsible for both phasic and tonic forms of inhibition (1). These receptors are pentameric, anion-selective ion channels that can be assembled from eight subunit classes: ␣(1-6), ␤(1-3), ␥(1-3), ␦, ⑀, , , and (1-3) (2)(3). This large number of receptor subunits provides the basis for a significant degree of heterogeneity of GABA A R structure and function. However, previous studies suggest that in the brain, the majority of phasic inhibition is dependent upon a few GABA A Rs subunits, namely the ␣, ␤, and ␥2 subunits located within synaptic sites (2,3). In the adult brain, these receptors are specific targets for brief exposures to high concentrations of GABA, resulting in short lived, but significant, hyperpolarization. In contrast, tonic inhibition is characterized by a sustained reduction in the cell's input resistance, effectively reducing the probability of action potential generation (1,4). Tonic inhibition is the result of persistent activation by GABA A Rs consisting primarily of ␣, ␤, and ␦ subunits located within peri-or extrasynaptic sites (1). With respect to specific brain regions, extrasynaptic GABA A Rs that mediate tonic inhibition in the thalamus and dentate gyrus of the hippocampus are composed of the ␣4 and ␤2/3 subunits with or without the ␦ subunit (5)(6)(7)(8)(9). Verification of the role that the ␣4 subunit plays in mediating tonic inhibition comes from ␣4 subunit knock-out mice, which have substantially lower levels of tonic inhibition in these brain areas (10,11).
Changes in tonic inhibition associated with GABA A Rs containing the ␣4 subunit have been implicated in a number of normal and pathological states in which the thalamus and the hippocampus play a role. It is apparent that tonic inhibition is essential for dynamically regulating the neuronal output, frequency of firing, and gain control of neurotransmission (12)(13)(14)(15)(16)(17)(18)(19). In addition, extrasynaptic GABA A Rs have been further shown to be targets for a wide range of endogenous and pharmacological agents, such as neurosteroids, anesthetics, ethanol, and anticonvulsants (20 -23). Finally, modifications in the efficacy of tonic inhibition arise under pathological conditions including stress, fragile X syndrome, aberrant brain activity associated with menstrual cycle, postpartum depression, schizophrenia, and temporal lobe epilepsies (20, 24 -28).
Little is known about the endogenous mechanism by which neurons control the functional properties of GABA A R subtypes that mediate tonic inhibition. It has long been established that a direct relationship exists between the number of synaptic GABA A Rs at the cell surface and the strength of inhibition at the synapse (29,30). Therefore, modulating the insertion and removal rate of GABA A Rs into or from the cell membrane has a marked affect on the amplitude of inhibitory synaptic currents (31). One way in which modulation occurs is via posttranslational modifications of the synaptic GABA A R. Specifically, the phosphorylation of key residues on synaptic GABA A R subunits regulates the extent to which the GABA A R will interact with protein complexes responsible for endocytosis from and insertion to the cell membrane; however, the significance of these regulatory processes for subtypes that mediate tonic inhibition remains largely unknown (32,33). Phosphorylation plays a role in regulating the functional expression of GABA A Rs containing ␣4 subunits. Our results reveal that the ␣4 subunit is phosphorylated on Ser 443 within the intracellular loop between transmembrane domains 3 and 4 (TM3 and TM4) in a protein kinase C (PKC)dependent manner. Activating PKC also resulted in higher steady state cell surface accumulation of GABA A Rs containing the ␣4 subunit that was dependent on enhanced insertion into the plasma membrane when expressed in a mammalian cell line. Consistent with this, PKC-dependent phosphorylation of Ser 433 produced a robust enhancement in GABA-induced currents in this expression system. Finally, we also observed that PKC activity increased both the phosphorylation and cell surface stability of the ␣4 subunit in hippocampal slices, a phenomenon that should be correlated with an increase in tonic inhibition. Together, these experiments establish a crucial role for PKC in regulating the functional expression of GABA A R subtypes that mediate tonic inhibition via direct phosphorylation of the ␣4 subunit.

EXPERIMENTAL PROCEDURES
Antibodies and Expression Constructs-Polyclonal rabbit anti-␣4 and anti-␦ antibody was graciously provided to us by Dr. Verena Tretter and Dr. Werner Sieghart from Medical University Vienna. ␤3 and phospho-␤3(phospho-S408A/ S409A) antibodies were designed by the Moss laboratory (34). Peroxidase-conjugated IgG secondary antibody was from Jackson ImmunoResearch Laboratories. Fluorescently labeled ␣-bungarotoxin (␣-Bgt) was purchased from Invitrogen. Wild-type and mutant ␣4 subunit and wild-type ␤3 cDNAs were cloned into the mammalian cytomegalovirus (CMV) promoter for transgene expression. For fluorescence experiments, ␣4 subunit protein was conjugated with red fluorescent protein and the Bgt binding site (BBS) (35). Briefly, dsRed monomer fluorescent protein (RFP) was introduced after the 4th amino acid of the mature ␣4 subunit followed by the BBS sequence (WRYYESSLEPYPD). The RFP and the BBS were separated by a 12-alanine/1-proline linker ( RFP-BBS ␣4 and RFP-BBS ␣4 S443A (mutant described below)). All constructs were generated using standard molecular biology cloning techniques and sequenced fully.
Site-directed Mutagenesis-Mutation of the ␣4 subunit was carried out using the QuikChange site-directed mutagenesis kit (Stratagene). The mutagenesis primers used to introduce an alanine in place of a serine at site 443 were CCTTTGCG-GTCGGCGGCTGCTCGCCCGGCATTT and AAATGC-CGGGCGCAGCCGCCGACCGCAAAGG. All mutations were verified by DNA sequencing.
HEK293 Cell Culture and Transfection-Human embryonic kidney (HEK293) cells were cultured in medium composed of Dulbecco's modified Eagle's medium supplemented with 10% calf serum and 1% penicillin/streptomycin at 37°C in a humidified 5% CO 2 atmosphere. Cells were electroporated (110 V, Bio-Rad Gene Pulser Xcell) with equal ratios of cDNA encoding for GABA A receptor subunits along with GFP cDNAs (in pCDM8). Cells were used 24 -72 h after transfection. Successful transfection of the cells was determined by fluorescence microscopy to identify GFP-labeled cells.
Hippocampal Slice Preparation-Hippocampal slices (350 m thick) from 10 -11-week-old C57BL/6 mice were prepared with a microslicer (Leica VT1000S) and pooled in icecold oxygenated artificial cerebral spinal fluid (ACSF Cell Lysis and Immunoprecipitation-Samples collected from either COS-7 cell cultures or hippocampal slices were lysed in lysis buffer containing the following: 20 mM Tris-HCl (pH 8.0), 150 mM NaCl, 5 mM EDTA, 10 mM NaF, 2 mM Na 3 VO 4 , 10 mM sodium pyrophosphate, 1% Triton X-100, and 0.1% SDS. In addition, the following protease inhibitors were added: 250 g/ml 4-(2-aminoethyl) benzenesulfonyl fluoride hydrochloride; 10 g/ml leupeptin, 1 g/ml pepstatin, and 10 g/ml antipain. Samples were then sonicated and spun at 16,000 ϫ g. The supernatant was collected and then subjected to a protein assay using a standard Bradford protocol. 100 -200 g of protein were loaded per immunoprecipitation sample along with 3 g of indicated antibody and 40 l of protein A-Sepharose beads (1:1 slurry) (GE Healthcare). Samples were allowed to conjugate for 18 -24 h at 4°C with constant agitation. The beads were precipitated by centrifuga-tion at 500 ϫ g and washed once with ice-cold Buffer A (20 mM Tris-HCl (pH 8.0), 150 mM NaCl, 5 mM EDTA, 10 mM NaF, 2 mM Na 3 VO 4 , 10 mM sodium pyrophosphate, and 1% Triton X-100 and protease inhibitors), two times with Buffer B composed of Buffer A supplemented with 500 mM NaCl, and once again with Buffer A. After the final wash, the beads were resuspended in 25 l of sample buffer and subjected to SDS-PAGE.
Whole-cell COS-7 Cell and Hippocampal Slice Metabolic 32 P Labeling-COS-7 cells were transfected and incubated as described above. Cells were initially incubated in 2 ml of phosphate-free DMEM (Invitrogen) for 30 min at 37°C. Following this incubation, cells were labeled with 0.5 mCi/ml [ 32 P]orthophosphoric acid for 4 h in phosphate-free DMEM. Hippocampal slices were prepared as described above. Slices were individually transferred to polypropylene tubes containing 2 ml of fresh ACSF; gassed with a mixture of 95% O 2 , 5% CO 2 ; and maintained in a 30°C water bath. Labeling was performed by adding 0.5 mCi/ml [ 32 P]orthophosphoric acid for 1 h. For both COS-7 cells and hippocampal slices, samples were treated with drugs where indicated after the labeling period, followed by the cell lysis and immunoprecipitation procedure described above. Results were attained by SDS-PAGE followed by autoradiography.
Phosphopeptide Mapping and Phosphoamino Acid Analysis-To perform phosphopeptide mapping, gel slices from 32 P labeling experiments were excised from SDS-polyacrylamide gels and washed and digested with 0.1 mg/ml trypsin and subjected to two-dimensional mapping, first by electrophoresis and then by thin layer chromatography (TLC). The resulting plate was then visualized by autoradiography (37). For phosphoamino acid analysis, phosphoproteins from gel slices were hydrolyzed using 6 N HCl. The resulting phosphoamino acids, along with phosphoamino acid standards, were separated by TLC and visualized by autoradiography (37).

Metabolic [ 35 S]Methionine
Labeling-Transfected COS-7 cells were incubated in methionine-free DMEM for 20 min and then pulsed with 0.5 mCi/ml [ 35 S]methionine (PerkinElmer Life Sciences) for 30 min. Cells were washed and incubated in complete DMEM/F-12 with an excess amount of unlabeled methionine for the indicated time periods (chase). Cells were lysed and subjected to immunoprecipitation as described above.
COS-7 Cell and Hippocampal Slice Cell Surface Biotinylation Assay-For transfected COS-7 cells, cultures were washed once with ice-cold PBS and then incubated in 2 ml of ice-cold PBS containing 1 mg/ml NHS-SS-biotin (Pierce) for 20 min in order to label surface proteins with biotin. After labeling, the biotin was quenched by incubating cells in PBS containing 25 mM glycine and 10 mg/ml bovine serum albumin (BSA) (38,39). Cells were then lysed in lysis buffer and sonicated. For hippocampal slice experiments, slices were incubated in ACSF described above at 30°C for 1 h for recovery before experimentation. Slices were then placed on ice and incubated for 30 min with 1 mg/ml NHS-SS-biotin. Excess biotin was removed by washing slices three times in icecold ACSF and lysed as described above (40). For both COS-7 cells and hippocampal slices, insoluble material was removed by centrifugation. The supernatant lysates were incubated with NeutrAvidin beads (Pierce) for 18 -24 h at 4°C. Bound material was eluted with sample buffer and subjected to SDS-PAGE and then immunoblotted with indicated antibodies. Blots were then quantified using the CCD-based FujiFilm LAS 300 system.
Fluorescent BBS Cell Membrane Insertion Assay-COS-7 cells were transfected with RFP-BBS ␣4 or RFP-BBS ␣4 S443A and the ␤3 subunit. All surface proteins expressing the BBS were blocked with 10 g/ml unlabeled ␣-Bgt for 15 min at 18°C. The cells were then washed extensively to remove unbound ␣-Bgt. Newly inserted RFP-BBS ␣4 or RFP-BBS ␣4 S443A was labeled with 2 g/ml Alexa 647-conjugated ␣-Bgt and fixed immediately with 4% paraformaldehyde after the indicated time points (35). Confocal images of fluorescently labeled COS-7 cells were collected using a ϫ60 objective, acquired with Nikon acquisition software, and analyzed with MetaMorph.
Patch Clamp Electrophysiology-Cells were superfused, at a rate of 2 ml/min, with an extracellular solution containing 140 mM NaCl, 5 mM KCl, 1.2 mM MgCl 2 , 2.5 mM CaCl 2 , 10 mM HEPES, 11 mM glucose and adjusted to pH 7.4 with NaOH. Borosilicate glass patch pipettes (resistance 2-5 megaohms) were filled with an internal solution containing 140 mM KCl, 2 mM MgCl 2 , 0.1 mM CaCl 2 , 1.1 mM EGTA, 10 mM HEPES, 2 mM ATP (Mg 2ϩ salt), adjusted to pH 7.4 with KOH. GABA was applied once every 120 s via a fast step perfusion system (Warner Instruments, Hamden, CT). All experiments were carried out at 32-33°C using a recording chamber and in-line perfusion heaters (Warner Instruments). Phorbol esters were applied to the cell either internally via the electrode solution or superfused into the recording chamber.
Data Acquisition and Analysis-For biochemical and immunofluorescent experiments, data are presented as means Ϯ S.E. Statistical analysis was performed by using Student's t test where a p value of Ͻ0.5 is considered significant. For electrophysiological experiments, currents were recorded with an Axopatch 200B amplifier (Molecular Devices, Sunnyvale, CA), filtered at 2 kHz and digitized at 20 kHz with a Digidata 1320A data acquisition system (Molecular Devices), and analyzed using either Clampfit (pClamp, Molecular Devices) or GraphPad Prism version 4 software (GraphPad Software, Inc., San Diego, CA). Statistical analysis was performed by using one-way ANOVA with a Bonferroni posttest with statistical significance set at p Ͻ 0.05. All data are expressed as mean Ϯ S.E.

Basal Phosphorylation of the GABA A Receptor ␣4 Subunit Is
Enhanced by PKC When Expressed in COS-7 Cells-Immunoprecipitation was used to examine the phosphorylation of GABA A receptors in COS-7 cells transiently transected with the GABA A receptor ␣4 and ␤3 subunits. Immunoprecipitation with anti-␣4 from transfected COS-7 cells that had been prelabeled with [ 32 P]orthophosphoric acid under basal conditions yielded a major phosphoprotein at ϳ64 kDa, demonstrating that the recombinant ␣4 subunit is basally phosphorylated (Fig. 1A). A corresponding band was not observed in untransfected COS-7 cells. Previous studies have shown that GABA A receptor subunits are the target of PKC (41)(42)(43). To determine whether the ␣4 subunit is a substrate of PKC, specific kinase activators and inhibitors were utilized. Activation of PKC with 500 nM phorbol 12,13-dibutyrate (PDBu) for 10 min produced a significant increase (p Ͻ 0.05) in ␣4 subunit phosphorylation of 223.7 Ϯ 25.17% (n ϭ 3) compared with control cells treated with DMSO for 10 min (Fig. 1A). An inhibitor of PKC, GF 109203X (GFX) (10 M for 20 min), had little effect on the basal phosphorylation of the ␣4 subunit (74.39 Ϯ 15.38% of control, n ϭ 3). However, treating transfected COS-7 cells with GFX 10 min prior to PDBu treatment prevented the increase in ␣4 subunit phosphorylation (81.1 Ϯ 26.98% of control, n ϭ 3) observed with PDBu treatment alone (Fig. 1A). Peptide mapping and phosphoamino acid analysis revealed that the PKC-dependent phosphorylation of the ␣4 subunit primarily occurs on serine residues (Fig. 1C) within one major phosphopeptide (Fig. 1B, circled). Together these results strongly suggest that PKC enhances basal levels of phosphorylation on the ␣4 subunit.
Ser 443 Is a Major Site for PKC-dependent Phosphorylation of the ␣4 Subunit When Expressed in COS-7 Cells-To further analyze ␣4 subunit phosphorylation, site-directed mutagenesis was utilized to convert candidate serine residues within the ␣4 subunit intracellular domain to alanines. Based on the consensus PKC motif of (R/K)X 1-4 (S/T)X 1-3 (R/K) (44 -45), a mutant version of the ␣4 subunit was produced in which Ser 443 was changed to an alanine (␣4 S443A subunit) ( Fig. 2A). COS-7 cells transfected with wild-type ␣4 or ␣4 S443A and ␤3 subunits were subjected to [ 32 P]orthophosphoric acid labeling and treated with either DMSO or PDBu. PDBu significantly enhanced (p Ͻ 0.05) levels of phosphorylation of wild-type ␣4 336 Ϯ 50.03% of control (Fig. 2B). In contrast to wild type, PDBu did not significantly enhance the phosphorylation of the ␣4 S443A subunit (Fig. 2B). These results strongly suggest FIGURE 1. ␣4 subunit phosphorylation is increased by PDBu, a specific PKC activator. A, untransfected COS-7 cells (UT) or COS-7 cells transfected with GABA A receptor ␣4 and ␤3 subunits were labeled with 0.5 mCi/ml [ 32 P]orthophosphoric acid and then treated with either PDBu (500 nM for 10 min) alone or following pretreatment with GFX (1 M for 10 min), a PKC inhibitor. The ␣4 subunit was immunoprecipitated, subjected to SDS-PAGE, and visualized with autoradiography (top). The level of phosphorylation was normalized to the amount observed in vehicle-treated samples (bottom) (dashed line represents vehicle set at 100%, p Ͻ 0.05). B, phosphopeptide map of the ␣4 subunit. [ 32 P]␣4 immunopurified from transfected COS-7 cells was digested with trypsin, and the resulting phosphopeptides were blotted onto TLC plates and subjected to electrophoresis followed by ascending chromatography. The small arrow indicates the origin. C, the ␣4 subunit was subjected to phosphoamino acid analysis followed by autoradiography. The migration of phosphoserine (pS), phosphothreonine (pT), and phosphotyrosine (pY) standards is indicated. Error bars, S.E. that the major site for PKC phosphorylation within the intracellular domain of the ␣4 subunit is Ser 443 .
PKC-dependent Phosphorylation of Ser 443 Enhances ␣4 Subunit Cell Surface Expression Levels in COS-7 Cells-Previous studies have shown that GABA A subunit phosphorylation dynamically regulates GABA A R cell surface expression (38,46,47). To address the functional consequences of ␣4 subunit phosphorylation, we first measured the effect PKC activation has on ␣4 subunit cell surface expression in transfected COS-7 cells using a biotinylation assay (38,39). This revealed that a 10-min treatment with 500 nM PDBu significantly increased (p Ͻ 0.05) the cell surface levels of the ␣4 subunit by 144.9 Ϯ 21.78%, but the total levels of protein were not altered (Fig. 3A). Treatment with GFX alone had no significant effect, and co-treatment with GFX and PDBu resulted in a modest increase in ␣4 subunit cell surface levels (Fig. 3A). Insignificant amounts of actin protein were pulled down in our biotinylated samples, ensuring that only cell surface proteins were being collected (Fig. 3A). In contrast to these results, PDBu treatment did not significantly increase (p Ͻ 0.05) the cell surface expression level of ␣4 S443A (Fig. 3B). Together, these biochemical experiments indicate that the PKC-dependent phosphorylation of the ␣4 subunit on Ser 443 increases receptor cell surface expression.
Analyzing the Phosphorylation of the ␤3 and ␦ Subunits-In the brain, the ␣4 subunit assembles with the ␤2/3 subunit with or without the ␦ subunit (5-9, 48). Thus, we examined if these subunits are also subject to PKC-dependent phosphorylation. To do so, the ␣4, ␤3, and ␦ subunits were co-expressed in COS-7 cells and then labeled with [ 32 P]orthophosphoric acid. After treatment with 500 nM PDBu for 10 min, the ␦ subunit was isolated via immunoprecipitation after denaturing lysis. Under these conditions, minimal levels of phosphorylation of the ␦ subunit were seen under basal conditions or after the activation of PKC (Fig. 4B). However, robust immunoprecipitation of the ␦ subunit was seen as measured via immunoblotting (Fig. 4B).
Previous studies have revealed that the ␤3 subunit is predominantly phosphorylated on serines 408/409 (Ser 408/409 ) in neurons upon activation of PKC (34,49). To examine if these residues are phosphorylated in COS-7 cells, lysates were immunoblotted with a phosphospecific antibody against these residues, phosphoserines 408/409 (phospho-Ser 408/409 ). PDBu treatment produced a robust enhancement of Ser 408/409 phosphorylation as measured via immunoblotting with Ser(P) 408/409 antibody (Fig. 4A). Thus, this experiment suggests that the ␦ subunit is not a PKC substrate at least when expressed in COS-7 cells and suggests that the principle sites for PKC phosphorylation within GABA A R subtypes that mediate tonic inhibition are Ser 443 in the ␣4 subunit and Ser 408/409 in ␤3.

Mutation of Ser 443 Increases the Rate of Insertion of the ␣4
Subunit on the Cell Membrane-To further evaluate the mechanism underlying PKC-dependent modulation of the ␣4 subunit cell surface stability, we determined what effect mutating the PKC site on the ␣4 subunit has on the level of insertion of ␣4 subunit-containing receptors. To do so, we utilized a Bgt binding assay that has previously been used to measure the rates of insertion of various receptor types (35,50). To analyze the insertion of the ␣4 subunit, we engineered the ␣4 subunit with the BBS peptide, WRYYESSLEPYPD. The BBS is derived from the ␣ subunit of the muscle nicotinic receptor and has been established to bind Bgt with an affinity of ϳ3 nM (51,52). The BBS together with a red fluorescent protein reporter were added to the N-terminal region of both the ␣4 and ␣4 S443A subunit.
Before conducting our insertion assay, we verified that the BBS-tagged versions of our ␣4 constructs were capable of forming a functional GABA A R. To do so, we expressed RFP-BBS ␣4 and non-tagged ␣4 separately with the ␤3 subunit in HEK293 cells and measured the current responses to 1 M and 1 mM GABA. Utilizing this method, we demonstrated that the RFP-BBS ␣4␤3 GABA A R subtype forms a functional receptor and has GABA-mediated currents similar to those of the ␣4␤3 GABA A R subtype (Fig. 5).
To measure the role that Ser 443 plays in regulating the cell surface accumulation of ␣4, RFP-BBS ␣4 and RFP-BBS ␣4 S443A cDNAs were separately transfected into COS-7 cells along with the GABA A ␤3 subunit and subjected to the BBS inser-tion assay. To perform this assay, we first masked the surface RFP-BBS ␣4 and RFP-BBS ␣4 S443A with unlabeled Bgt by incubating the transfected COS-7 cells with native Bgt at 18°C for 15 min. Under these conditions, the unlabeled Bgt completely blocked the existing cell surface population of RFP-BBS ␣4 and RFP-BBS ␣4 S443A subunits. Next, the cells were incubated at 37°C with Alexa 647-Bgt in order to fluorescently label newly inserted RFP-BBS ␣4 and RFP-BBS ␣4 S443A . Visually, it is apparent that there is a higher amount of RFP-BBS ␣4 S443A inserted after 10 min compared with RFP-BBS ␣4 (Fig. 6A). To quantify these results, we calculated the ratio of the level of Alexa 647-Bgt staining to the level of RFP fluorescence. This ratio was significantly higher (p Ͻ 0.05) in COS-7 cells transfected with RFP-BBS ␣4 S443A than in cells transfected with RFP-BBS ␣4 after 10 min of labeling (1.35 Ϯ 0.33 versus 0.78 Ϯ 0.16, n ϭ 3; Fig.  6B). We also tested the rate of endocytosis of the RFP-BBS ␣4 and RFP-BBS ␣4 S443A in a similar BBS assay and found no significant difference between the wild-type and mutant ␣4 subunit (data not shown). Together, these results strongly suggest that phosphorylation of Ser 443 , the major site of PKC-dependent phosphorylation within the ␣4 subunit, regulates the rate of insertion of the ␣4 subunit in COS-7 cells.
Mutation of Ser 443 Increases the Protein Stability of the ␣4 Subunit-To determine the role phosphorylation plays in the production and stability of the ␣4 subunit, COS-7 cells transfected with ␣4 or ␣4 S443A alone were subjected to a [ 35 S]methionine pulse-chase assay. Transfected COS-7 cells were labeled with 100 Ci/ml [ 35 S]methionine for 30 min and chased for 0 and 4 h with excess cold methionine. Cell lysates were then prepared and subjected to immunoprecipitation with anti-␣4 and resolved on SDS-polyacrylamide gels and quantified on a Bio-Rad isotope imager (Fig. 7). Data at 4 h are presented as a percentage of [ 35 S]methionine-labeled protein existing at time 0. ␣4 subunit protein does not reach the cell surface unless a ␤3 subunit is also present (supplemental Fig.  1); therefore, under these conditions, we are measuring the stability of protein that is retained in the endoplasmic reticulum. Using this technique, we determined that 47.40 Ϯ 8.18% of newly synthesized ␣4 subunit protein remained after 4 h (Fig. 7). Interestingly, this reduction was markedly less robust for the ␣4 S443A subunit protein, reducing to 80.20 Ϯ 7.00% of newly synthesized protein after 4 h (Fig. 7). Therefore, the ␣4 S443A subunit is more stable than the wild-type ␣4 subunit, suggesting that PKC phosphorylation of the Ser 443 site regulates the ␣4 subunit protein half-life.
Protein Kinase C Activation Reverses Current Run-down-The functional effects of protein kinase C activation were de- . Analyzing PKC phosphorylation of GABA A R subunits that mediate tonic inhibition. A, COS-7 cells expressing the ␣4 and ␤3 subunits were treated with 500 nM PDBu for 10 min and then immunoblotted with phospho-S408A/S409A (pS408/9) or ␤3 antibodies, and the ratio of pS408/9/␤3 immunoreactivity was determined and normalized to control (dashed line represents vehicle set at 100%; p Ͻ 0.05). B, COS-7 cells expressing ␣4, ␤3, and ␦ subunits were labeled with 0.5 mCi/ml [ 32 P]orthophosphoric acid and treated with 500 nM PDBu for 10 min. The ␦ subunit was isolated by denaturing immunoprecipitation followed by SDS-PAGE (IP/ 32 P). Parallel cultures were immunoprecipitated and immunoblotted with ␦ antibodies (IP/WB ␦). termined by whole-cell patch clamp recording of HEK293 cells. Transient expression of wild-type ␣4␤3 receptors and mutant ␣4 S443A ␤3 receptors resulted in functional channels that had GABA-evoked EC 50 values of 1.7 Ϯ 0.7 and 2.6 Ϯ 0.9 M (n ϭ 3-10), respectively, with both Hill coefficients of 0.8 Ϯ 0.2 (data not shown). To examine the run-down of GABA-mediated currents, 1 M GABA was applied to the cells once every 2 min.
In the absence of PKC activation, the GABA-mediated current amplitude decreased over time but appeared to plateau after 16 min of recording. After 20 min of recording, the GABA-mediated current was 37 Ϯ 12% (n ϭ 5) of the initial response. Inclusion of 100 nM PDBu in the recording pipette solution prevented the GABA-mediated current amplitude run-down. At 20 min after the start of the experiment, the GABA-mediated current amplitude was 97 Ϯ 9% (n ϭ 3) of the initial GABA-mediated response. Similarly, external appli-cation of 100 nM PDBu also reverses GABA-mediated current amplitude run-down with current amplitude being 107 Ϯ 12% (n ϭ 4) compared with that at the start of the experiment (Fig. 8).
When the PKC-inactive phorbol ester, 4-␣-phorbol 12,13didecanoate (100 nM), was included in the intracellular solution of the recording pipette, the GABA-mediated current amplitude run-down was no different from control. Cells rarely remained healthy past 16 min of recording. After 16 min of recording, the GABA-evoked currents were 17 Ϯ 6% (n ϭ 3) of the first current in the presence of 100 nM 4-␣phorbol 12,13-didecanoate compared with 33 Ϯ 10% (n ϭ 7) in control (Fig. 9). At this time point, minimal run-down was observed in cells internally perfused with 100 nM PDBu (98 Ϯ 9% of control, n ϭ 3).
The ␣4 S443A Subunit Mutation Prevents Run-down-GABA-mediated currents in cells expressing ␣4 S443A ␤3 re-FIGURE 6. S443A point mutation increases the rate of insertion of the ␣4 subunit into the cell membrane. A, ␣4-WT and ␣4-S443A DNA constructs were made containing an RFP tag as well as a BBS. COS-7 cells were then co-transfected with either wild-type (␣4-WT-BBS) or S443A mutant (␣4-S443A-BBS) and ␤3 GABA A receptor subunits. Transfected COS7 cells were then incubated with unlabeled Bgt (10 nM for 10 min) at 12°C to block insertion. Cells were then incubated with Alexa 647-conjugated Bgt (10 nM for 10 min) at 37°C. Cells were then fixed, and the level of newly inserted Alexa 647-tagged ␣4 was determined using confocal microscopy and quantified using MetaMorph. B, the graph is presented as a ratio of Alexa 647 fluorescent intensity (newly inserted protein) to RFP fluorescence (total protein) over specific time periods. Ratios for ␣4-WT and ␣4-S443A were then compared with one another (p Ͻ 0.05). Error bars, S.E. ceptors did not show the typical time-dependent run-down phenomena. Unlike that observed with wild-type ␣4-containing receptors in control conditions, after 20 min of recording, the GABA-mediated current from ␣4 S443A containing receptors was 84 Ϯ 13% (n ϭ 4) of the initial GABA-mediated response. In the presence of internal 100 nM PDBu, GABA-mediated current was 98 Ϯ 9% (n ϭ 4) of the initial response after 20 min of recording (Fig. 10).
PKC Activity Modulates the Phosphorylation and Cell Surface Stability of Endogenous ␣4 Subunit in Hippocampal Slices-To determine the relevance of our recombinant studies, we examined the phosphorylation of the endogenous ␣4 subunit and its effects on cell surface stability of this protein in hippocampal slices. We began by examining the PKC-dependent phosphorylation of the ␣4 subunit. Hippocampal slices were cut from brains dissected from 10 -11-week-old C57BL/6 mice. Slices were labeled with [ 32 P]orthophosphoric acid for 4 h in ACSF continuously being bubbled with a 95% O 2 , 5% CO 2 gas mixture. Toward the end of the 4 h, control slices were treated with DMSO for 10 min, whereas PKC slices were treated with 500 nM PDBu for 10 min. Afterward, slices were lysed and subjected to immunoprecipitation with anti-␣4. Under control conditions, a very faint band at ϳ64 kDa was detected, representing the basal level of ␣4 subunit phosphorylation (Fig. 11A). Treatment of slices with PDBu significantly increased (p Ͻ 0.05) ␣4 subunit phosphorylation to 2016 Ϯ 1260% of control (Fig. 11A).
We further evaluated phosphorylation of the ␣4 subunit in hippocampal slices by performing phosphoamino acid analysis. We determined that PKC activation with PDBu resulted in the phosphorylation of the ␣4 subunit principally on serine   resides (Fig. 11B), similar to what was observed in transfected COS-7 cells (Fig. 1C).
We also examined the effects of PKC activation on the cell surface stability of the ␣4 subunit in hippocampal slices using biotinylation. This revealed that activation with PDBu produced a significant increase (p Ͻ 0.05) in the amount of ␣4 subunit protein on the surface of cells in the hippocampal slice to 293.7 Ϯ 90.70% of DMSO-treated slices (Fig. 11C). Similar to what was observed in our recombinant studies, these results demonstrate that the ␣4 subunit, in its native environment, is phosphorylated in a PKC-dependent fashion and that this post-translational modification increases the targeting of the ␣4 subunit to the cell surface.

DISCUSSION
The hippocampus is responsible for the formation of various types of memory in mammals. Pivotal to this role, the dentate gyrus of the hippocampus acts as an informationprocessing center between afferent connections from the entorhinal cortex and efferent connections to cornu ammonis region 3 of the hippocampus (53). Specifically, the dentate gyrus is responsible for filtering out stimulus-related high frequency firing from the entorhinal cortex and organizing it into a coherent signal that the rest of the hippocampus can utilize (54). Compromising filtering function of the dentate gyrus leads to epileptiform activity in downstream hippocampal structures (55). The filter-like function of the dentate gyrus is due to the intrinsic low excitability of granule cells residing in this region. This low excitability is a result of the high levels of protein expression of the GABA A ␣4 subunit, creating a significant degree of tonic inhibition (10). Therefore, understanding the cellular mechanisms that control the activity of the ␣4 subunit is crucial for better comprehending the complex mechanisms of memory formation as well as other higher brain functions where ␣4 subunit-mediated tonic inhibition has been shown to play a role.
Here we have begun to analyze the endogenous mechanisms that neurons utilize to regulate the efficacy of tonic inhibition, focusing on the possible role that direct phosphorylation may play in these processes. For these studies, we used receptors composed of ␣4␤3 and ␣4␤3␦ subunits because both of these combinations have been demonstrated to be present within the dentate gyrus of the hippocampus as measured using biochemical and electrophysiological experiments (5)(6)(7)(8)(9). We initiated these studies by determining if the ␣4 subunit is phosphorylated by expressing this protein in COS-7 cells and measuring the amount of radiolabeled phosphate that covalently bonds to the ␣4 subunit. This revealed that the ␣4 subunit is phosphorylated basally, and activation of PKC enhances subunit phosphorylation. Peptide mapping and phosphoamino acid analysis showed that PKC-dependent phosphorylation of the ␣4 subunit occurs on a serine residue in one distinct phosphopeptide region. Using site-directed mutagenesis, we determined that the ␣4 subunit is phosphorylated at Ser 443 , an amino acid located in the major intracellular domain between TM3 and TM4. In addition to our recombinant studies, we also showed that PKC leads to high levels of ␣4 subunit serine phosphorylation in hippocampal slices FIGURE 11. PKC increases the level of phosphorylation and cell surface expression of the ␣4 subunit in hippocampal slices. A, hippocampal slices from 10 -11-week-old C57BL/6 male mice were labeled with [ 32 P]orthophosphoric acid and treated with either vehicle or PDBu (500 nM for 10 min). Detergent-soluble extracts were immunoprecipitated with either rabbit IgG or anti-␣4, resolved by SDS-PAGE, and then visualized by phosphorimaging (top). Histograms are presented as 32 P incorporation expressed as a percentage of vehicle-treated control (bottom) (dashed line, p Ͻ 0.05). B, the immunoprecipitated ␣4 subunit from [ 32 P]orthophosphoric acid-treated hippocampal slices was subjected to phosphoamino acid analysis followed by autoradiography. The migration of phosphoserine (pS), phosphothreonine (pT), and phosphotyrosine (pY) standards is indicated. C, hippocampal slices from 10 -11week-old C57BL/6 male mice treated with either vehicle or PDBu (500 nM for 10 min) were labeled with NHS-SS-biotin and detergent-soluble extracts were purified on NeutrAvidin. Cell surface (Surface) and 10% of total fractions (Total) were analyzed by immunoblotting with anti-␣4 (top). Histograms show the proportion of cell surface ␣4 protein expressed as a percentage of vehicle-treated controls (bottom) (dashed line, vehicle set at 100%; p Ͻ 0.05). Error bars, S.E. from adult male mice. We also analyzed the phosphorylation of the ␦ and ␤3 subunits in our study. Consistent with studies on GABA A R subtypes that mediate phasic inhibition, Ser 408/ 409 in the ␤3 subunit were phosphorylated by PKC activity when expressed with ␣4. However, at least in COS-7 cells, only low levels of ␦ subunit phosphorylation were seen. Thus, these results suggest that the primary PKC substrates with GABA A R subtypes that mediate tonic inhibition are Ser 443 in the ␣4 subunit and Ser 408/409 in ␤3.
To begin ascertaining the functional consequences of phosphorylation on tonic inhibition, we looked at the effect PKC activation had on the cell surface stability of the ␣4 subunit. The activation of PKC leads to a dramatic increase in the amount of ␣4 subunit protein at the cell surface in both transfected COS-7 cells and hippocampal slices, as measured by biotinylation. The Ser 443 phosphorylation site plays a crucial role in this enhancement because mutation of this residue did not result in elevated levels of ␣4 subunit protein at the cell surface of COS-7 cells. At this point, it was clear that Ser 443 is essential in mediating the effects of PKC activity on ␣4 subunit cell surface accumulation. To begin to address the underlying mechanism, we measured the rate of insertion of the ␣4 subunit into the cell membrane using a BBS fluorescent insertion assay. Here we discovered that over a 10-min period, more mutant ␣4 was being inserted into the COS-7 cell membrane than wild type ␣4 subunit. This increased rate of insertion was also paralleled with an increase in stability of newly translated ␣4 S433A subunit compared with wild type ␣4 when expressed alone in COS-7 cells. Given that the ␣4 subunit is retained within the endoplasmic reticulum in homomeric expression, this result suggests that phosphorylation of Ser 443 acts to regulate the stability of the ␣4 subunit in this intracellular compartment, which would be predicted to increase receptor assembly, leading to increased insertion into the plasma membrane.
To investigate this possibility, we measured the amount of protein degradation using an [ 35 S]methionine pulse-chase assay and found that the ␣4 subunit Ser 443 mutant was more stable in the endoplasmic reticulum over a 4-h period. Taking these results together, we see a situation in which the mutant version of the ␣4 subunit is not only degraded less but is inserted faster into the cell membrane. At first glance, this seems at odds with our results showing that Ser 443 is a critical residue for the PKC-dependent phosphorylation of the ␣4 subunit that leads to higher levels of this protein on the cell surface. How is it then that ablating this phosphorylation site prevents ␣4 from being phosphorylated but still causes the mutant protein to be inserted at a faster rate? One answer to this question is that mutation of Ser 443 to an alanine results in phosphorylation mimic of the ␣4 subunit. That is to say that masking the hydroxyl group that is normally found in a serine residue by removing it, which is what we do when we replace this residue with an alanine, is tantamount to masking it by covalently attaching a phosphate group. Both situations may lead to similar protein conformational changes that result in the ␣4 subunit being more stable and therefore being inserted at a faster rate. We can further draw this conclusion from our electrophysiological studies, which suggest a similar occur-rence. In these studies, we found that PKC activation reverses the run down that normally occurs in non-treated HEK293 cells expressing the wild-type ␣4 subunit, as is expected from our biotinylation studies. Interestingly, HEK293 cells expressing the ␣4 phosphomutant do not exhibit any run down, either in the presence or absence of PKC activation. That is to say that the mutant by itself is protected by the run-down effect that is normally observed in the wild-type ␣4 subunit. If this is the case, it is logical to conclude that the reason PKC does not exert an effect on the mutant ␣4 subunit is that the protein is already acting as if it is constitutively phosphorylated and is being inserted into the membrane at a maximal rate.
Here we have for the first time a description of a PKC-dependent mechanism that regulates the activity of a GABA A R subunit that is primarily expressed in extrasynaptic sites. Our laboratory has shown in the past that synaptic GABA A Rs are highly regulated by kinase and phosphatase activity (32). Due to the plethora of kinases and phosphatases that exert an effect on synaptic GABA A Rs and the different brain regions and cell types in which this activity has been observed, phosphorylation regulates synaptic inhibition in a multitude of ways. However, one common facet of phosphoregulation of synaptic GABA A Rs, with respect to cell surface stability, is that it modulates the endocytosis of the receptor. In contrast, we have shown in this study that kinase activity affects the insertion of extrasynaptic GABA A R subtypes. Synaptic and extrasynaptic inhibition are two fundamentally different ways a neuron can regulate its excitability; therefore, it is important that the neuron be able to regulate each form by modulating different cellular mechanisms.
In summary, our studies demonstrate that the ␣4 subunit, a protein critical for tonic inhibition in the dentate gyrus of the hippocampus, is phosphorylated by PKC on Ser 443 . This phosphorylation leads to an increase in the functional expression of the ␣4 subunit-containing GABA A R by increasing its stability and enhancing the rate at which this receptor is inserted into the plasma membrane. Therefore, PKC-dependent phosphorylation of the ␣4 subunit may have profound effects on the efficacy of tonic inhibition mediated by GABA A Rs.