The Ubiquitin-like Protein Plic-1 Enhances the Membrane Insertion of GABAA Receptors by Increasing Their Stability within the Endoplasmic Reticulum*

γ-Aminobutyric acid receptors (GABAAR) are the major sites of fast inhibitory neurotransmission in the brain, and a critical determinant for the efficacy of neuronal inhibition is the number of these receptors that are expressed on the neuronal cell surface. GABAARs are heteropentamers that can be constructed from seven subunit classes with multiple members; α, β, γ(1–3), δ, ϵ(1–3), θ, and π. Receptor assembly occurs within the endoplasmic reticulum, and it is evident that transport-competent combinations exiting this organelle can access the cell surface, whereas unassembled subunits are ubiquitinated and subject to proteasomal degradation. In a previous report the ubiquitin-like protein Plic-1 was shown to directly interact with GABAARs and promote their accumulation at the cell surface. In this study we explore the mechanisms by which Plic-1 regulates the membrane trafficking of GABAARs. Using both recombinant and neuronal preparations it was apparent that Plic-1 increased the stability of endoplasmic reticulum resident GABAARs together with an increase in the abundance of poly-ubiquitinated receptor subunits. Furthermore, Plic-1 elevated cell surface expression levels by selectively increasing their rates of membrane insertion. Thus, Plic-1 may play a significant role in regulating the strength of synaptic inhibition by increasing the stability of GABAARs within the secretory pathway and thereby promoting their insertion into the neuronal plasma membrane.

GABA A Rs 3 are Cl Ϫ -selective ligand-gated ion channels and are the major mediators of fast synaptic inhibition in the brain. GABA A Rs are also the major sites of action for anxiolytics, sedatives, and anti-convulsants, including both barbiturates and benzodiazepines (1,2). These receptors are heteropentamers that can be assembled from seven subunit classes, ␣, ␤, ␥(1-3), ␦, ⑀(1-3), , and , providing the structural basis for receptor structure (1,2). A combination of molecular, biochemical, and genetic approaches suggests that, in the brain, the majority of benzodiazepine-sensitive synaptic receptor subtypes are composed of ␣, ␤, and ␥2 subunits (3). In contrast, receptors that incorporate ␣4/5 and ␦ subunits are extrasynaptic and mediate tonic inhibition (4). The number of GABA A receptors on the neuronal cell surface is a critical determinant for the efficacy of synaptic inhibition and, at steady state, is determined by the rates of receptor insertion and removal from the plasma membrane (5).
It is evident that GABA A Rs are assembled within the endoplasmic reticulum (ER) and then transported to the plasma membrane for insertion while misfolded or unassembled receptor subunits are rapidly targeted for ER-associated degradation (6 -9). Proteins targeted for ER-associated degradation are modified by ubiquitination, resulting in their recognition and degradation by the proteasome (10). In addition, neuronal activity can regulate the abundance of cell surface GABA A Rs by altering subunit ubiquitination, a process that in turn controls receptor insertion into the plasma membrane and subsequent accumulation at postsynaptic inhibitory specializations (11).
The ubiquitination and proteasomal degradation of proteins is subject to modulation by regulatory proteins that contain an N-terminal ubiquitin-like domain and one or two ubiquitinassociated domains (12). One member of this class of proteins termed Plic-1 (protein linking integrin-associated protein to cytoskeleton-1) (13) binds to GABA A Rs. These interactions are dependent upon the ubiquitin-associated domain of Plic-1 that binds to conserved amino acids within the intracellular domains of GABA A R ␣ and ␤ subunit isoforms. Moreover, this interaction has been demonstrated to be of significance in mediating the cell surface accumulation of GABA A Rs as revealed using dominant negative peptides (14).
Here we have begun to explore the mechanisms by which Plic-1 regulates GABA A R cell surface stability. We demonstrate that Plic-1 increases the stability of ER resident GABA A R ␤3 subunits and also elevates the levels of a ubiquitinated ␤3 species. As a result, Plic-1 increases the insertion of GABA A Rs into the neuronal cell surface without affecting their endocytic sorting. Therefore, Plic-1 regulation of the ubiquitin-dependent, proteasomal degradation of GABA A Rs may provide a dynamic mechanism regulating the efficacy of inhibitory synaptic transmission.

EXPERIMENTAL PROCEDURES
Antibodies-Rabbit anti-Myc IgGs were obtained from Santa Cruz Biotechnology. Monoclonal anti-HA IgGs were obtained from Roche Applied Science. Rabbit polyclonal anti-␤3 IgGs have been described previously (15,16). Secondary horseradish peroxidase-conjugated antibodies were from Jackson Immunological Laboratories and protein A horseradish peroxidase was from Pierce.
Biotinylation-HEK-293 cells or cortical neurons were chilled on ice for 5 min and then washed twice in PBS plus 1 mM CaCl 2 and 0.5 mM MgCl 2 (PBS-CM) at 4°C. Cells were incubated for 15 min at 4°C in 1.5 mg/ml sulfosuccinimidyl 2-(biotinamido)-ethyl-1, 3-dithiopropionate (NHS-SS-Biotin) (Pierce) dissolved in PBS-CM. To quench unreacted biotin, neurons were washed three times (10 min each wash, at 4°C) in PBS-CM plus 75-100 mM glycine and then washed twice in PBS and either lysed in RIPA buffer (50 mM Tris, pH 8, 150 mM NaCl, 1% Nonidet P-40, 0.5% deoxycholate, 0.1% SDS, 2 mM EDTA, and a mammalian protease inhibitor mixture; Sigma) or returned to a 37°C incubator. Protein concentrations were determined using the micro BCA protein assay kit (Pierce), and equal amounts of protein were added to immobilized avidin (neutravidin; Pierce) for 2 h at 4°C. Avidin beads were washed twice, for 15 min at 4°C, in high salt (500 mM NaCl) RIPA buffer, followed by a 15-min wash at 4°C, in (150 mM NaCl) RIPA buffer. Precipitated biotinylated proteins were resolved by SDS-PAGE, and ␤3 was detected using immunoblotting with rabbit anti-␤3, followed by peroxidase-conjugated antirabbit IgGs and detection with ECL. Blots were imaged using the Fujifilm LAS-3000 imaging system and bands quantified with Fujifilm Multi Gauge software.
Deglycosylation-Endoglycosidase H (Endo H) and peptide N-glycosidase-F (PNGaseF) were purchased from New England BioLabs. Transfected HEK-293 cells expressing ␣1 M ␤ 3 with or without Plic-1 were lysed in RIPA buffer plus protease inhibitor mixture (Sigma). Total protein (10 g) from lysates was denatured in 0.5% SDS and 40 mM dithiothreitol for 10 min at room temperature and then incubated for 1 h with Endo H at 37°C in 1ϫ reaction buffer (0.05 M sodium citrate, pH 5.5). N-Glycosidase-F digestion of 10 g of denatured total protein was performed in 0.05 M sodium phosphate, pH 7.5, and 1% Nonidet P-40 for 1 h at 37°C. Digestion was terminated by the addition of Laemmli sample buffer. Protein products were resolved by SDS-PAGE, and the M ␤ 3 subunit was detected by immunoblotting.
Immunoprecipitation-Transfected HEK-293 cells were lysed in 100 l of 1% SDS, 25 mM Tris, pH 7.4. Lysates were then diluted 10-fold, with 900 l of RIPA buffer (lacking SDS) plus 10 mM N-ethylmaleimide (Sigma) and mammalian protease inhibitor mixture (Sigma). Lysates were then sonicated in 1.5 ml of microcentrifuge tubes for 10 s at 10 amplitude microns (Sanyo Soniprep 150). Following centrifugation at 14,000 rpm to pellet insoluble material, lysates were precleared with nonspecific rabbit IgGs and protein A-Sepharose for 1 h at 4°C. Lysates were then incubated for 1 h at 4°C with 3 g of rabbit anti-Myc (Santa Cruz Biotechnology) and then a further 1 h with protein A-Sepharose. Precipitated immunocomplexes were washed, alternatively, in low salt (150 NaCl) and high salt (350 mM NaCl) RIPA buffer for 20 min at 4°C. Immunoprecipitated antigen was resolved on a 10% SDS-PAGE gel and immunoblotted with mouse anti-HA, followed by horseradish peroxidase-conjugated donkey anti-mouse IgGs and detection with ECL (Super-Signal; Pierce). Blots were stripped in 62.5 mM Tris-HCl, pH 6.7, 2% SDS, 100 mM ␤-mercaptoethanol at 60°C for 30 min and reprobed with rabbit anti-Myc, followed by protein A-horseradish peroxidase (Pierce) and detection with ECL.
BBS ␤3 Cell Surface Labeling/Receptor Insertion Assay-Hippocampal neurons (9 DIV) were transfected with equimolar amounts of cDNAs encoding BBS ␤3 (23) and Plic-1 (14) or empty vector (pRK5) using Effectene (Qiagen) according to the manufacturer's instructions. Following transfection, neurons were grown for a further 24 h. To measure steady state cell surface levels of GABA A R incorporating BBS ␤3 subunits, neurons were labeled with 1 g/ml rhodamine-conjugated ␣-bungarotoxin (Invitrogen) for 15 min at 15°C. Neurons were washed three times in PBS at 15°C and then fixed. To measure the membrane insertion of neuronal GABA A Rs incorporating BBS ␤3 subunits in the presence or absence of Plic-1, neurons were labeled with 10 g/ml unlabeled ␣-bungarotoxin for 15 min at 15°C to block cell surface receptors. The cells were washed three times in PBS at 15°C followed by a 4-min incubation at 37°C with 1 g/ml rhodamine-conjugated ␣-bungarotoxin (Invitrogen). All incubations were performed in the presence of 200 M tubocurarine (Sigma) to block ␣-bungarotoxin (Bgt) binding to endogenous acetylcholine receptors (23,24). Cells were fixed in 4% paraformaldehyde, and confocal images were collected using a ϫ60 objective lens acquired with Olympus FluoView Version 1.5 software; the same image acquisition settings for BBS ␤ 3 with or without Plic-1 were used. These images were analyzed using MetaMorph (Universal Imaging Corp.) imaging software. First, a three-dimensional reconstruction of an imaged neuron was made from a series of Z sections, and then the average fluorescence intensity of rhodamine-Bgt staining was measured along 30 m of two proximal dendrites/neuron, after subtraction of background fluorescence.
GABA A Rs composed of these subunits because they reproduce many of the physiological and pharmacological properties of their neuronal counterparts (1, 2). Following transfection, HEK cells were lysed and equal amounts of detergent-soluble extracts were resolved by SDS-PAGE and subjected to immunoblotting with anti-Plic-1 and anti-␤3 IgGs. Although abundant Plic-1 immunoreactivity was seen in cells transfected with the Plic-1 construct, significant levels were absent in cells transfected with empty vector (Fig. 1A). Consistent with this we have previously demonstrated that HEK-293 only express low endogenous levels of Plic-1 (14). In cells expressing Plic-1, total expression levels of the ␤3 subunit were increased by 67.2 ϩ 4.5% compared with control ( Fig. 1A). Using biotinylation to label and isolate cell surface proteins, it was also apparent that Plic-1 increased the cell surface accumulation of GABA A Rs relative to control cells (Fig. 1A), which is consistent with previous observations (14).
The increase in GABA A R cell surface levels on expression with Plic-1 may be a consequence of decreased endocytosis, increased recycling of GABA A R from an endocytic pool, or an increase in insertion from the secretory pathway. To distinguish among these possibilities, we compared the cell surface degradation of GABA A Rs in the presence or absence of Plic-1. HEK cells expressing ␣1␤3␥2 with or without Plic-1 were labeled with cell-impermeant NHS-SS-Biotin and incubated for varying time periods at 37°C. The levels of remaining biotinylated ␤3 subunits were then measured by immunoblotting of the avidin-purified fraction with rabbit anti-␤3 IgGs. These experiments revealed similar levels of degradation of cell sur-face GABA A Rs in the presence or absence of Plic-1 over a 20-h time period (Fig. 1B).
The Abundance of ER-retained GABA A Rs Is Increased by Plic-1-The results in Fig. 1 suggest that Plic-1 primarily acts to increase GABA A R cell surface abundance by modifying receptor stability within the secretory pathway. Given that GABA A Rs are assembled from their constitutive subunits within the ER, we assessed the significance of Plic-1 in regulating receptor stability within this intracellular compartment. For these experiments, we examined the sensitivity of N-linked glycans within the GABA A R ␤3 subunit to digestion with Endo H. It is well known that transmembrane proteins residing within the ER carry high mannose N-linked glycans that are sensitive to cleavage by Endo H (26,27). In contrast, transmembrane proteins that have exited the ER have mature N-linked glycans that are Endo H-insensitive. In our experiments, detergent-soluble extracts from HEK cells expressing ␣1␤3 with or without Plic-1 were digested with Endo-H and then immunoblotted with rabbit anti-␤3 IgGs. Endo H-resistant and -sensitive ␤3 was detected in control cells and those expressing Plic-1 ( Fig. 2A). For comparison, samples were also treated with N-glycosidase-F, which cleaves all N-linked glycans, resulting in the detection of a single band of 52 kDa ( Fig. 2A) (17,28). We measured the levels of Endo H-sensitive ␤3 subunits in the absence or presence of Plic-1. Plic-1 increased the amount of Endo H-sensitive GABA A R ␤3 subunits by 32.5 ϩ 5.4% compared with the ␤3 subunit co-expressed with empty vector (Fig. 2B). These results suggest that Plic-1 stabilizes ER resident ␤3 subunits.

Plic-1 Increases GABA A R Stability within the ER
Plic-1 with M ␤3 increased the abundance of M ␤3-ubiquitin conjugates by 100 Ϯ 6.8% compared with control cells transfected with empty vector (Fig. 3B). However, Plic-1 did not appear to significantly alter the level of ubiquitination of individual GABA A receptor ␤3 subunits, as the ubiquitin/␤3 signal ratio remained the same (Fig. 3C). Thus, co-expression of Plic-1 with GABA A Rs increases the accumulation of ␤3-ubiquitin conjugates that previous studies have illustrated are enriched within the ER (11).

Plic-1 Enhances the Cell Surface Expression Levels of Neuronal GABA A Rs-To examine the relevance of our studies of recombinant receptors in HEK cells to endogenous GABA A Rs
in neurons, we examined the effects of increased Plic-1 expression on the levels of GABA A Rs in cultured neurons. E18 cortical neurons were transfected with either pRK5-GFP (control) or with pRK5-Plic-1 using nucleofection and then incubated for 3-4 days in vitro. Neurons were then lysed in RIPA buffer, and equal amounts of protein were subjected to SDS-PAGE and immunoblotting with anti-Plic-1 and anti-␤3 IgGs. In cultures nucleofected with pRK5-Plic-1 an increase in the level of Plic-1 expression was evident compared with control (Fig. 4A). Furthermore, Plic-1 increased the total pool of endogenous ␤3 subunits by 26.7 Ϯ 2% compared with that seen in control neurons expressing GFP (Fig. 4A). In contrast, the levels of the ionotropic glutamate receptor GluR1 subunit were unaltered (Fig. 4A).
To determine whether Plic-1 expression regulated the cell surface abundance of GABA A Rs in neurons, we used a biotinylation assay to label and isolate surface receptors. In these experiments cortical neurons were nucleofected with control pRK5-GFP or Plic-1 and incubated for 3-4 days in vitro fol-lowed by biotinylation and isolation of cell surface receptors. Plic-1 increased the cell surface expression levels of GABA A Rs incorporating ␤3 subunits by 27.6% compared with those in control neurons without altering the cell surface levels of the GluR1 subunit (Fig. 4B). We also examined the influence of Plic-1 on the cell surface degradation of GABA A Rs. Nucleofected neurons expressing pRK5-GFP or Plic-1 (3-4 days in vitro) were biotinylated to label cell surface receptors and then incubated for 0 or 24 h at 37°C. Consistent with our recombinant experiments in HEK cells, Plic-1 did not appear to influence the degradation of cell surface ␤3-containing GABA A Rs (Fig. 4, C and  D). Together these results suggest that Plic-1 acts to increase the steady state cell surface levels of GABA A R by selectively increasing their insertion into the plasma membrane.
Plic-1 Enhances the Cell Surface Accumulation of Recombinant ␤3 Subunits in Neurons-To corroborate our biochemical experiments we examined the influence of Plic-1 on the cell   surface accumulation of GABA A Rs using fluorescent labeling of live hippocampal neurons. For these experiments hippocampal neurons (9 DIV) were transfected with pRK5-Plic-1 or empty control vector and the GABA A R ␤3 subunit engineered with an ␣-bungarotoxin binding site (BBS) and a pHluorin reporter incorporated into the N terminus ( BBS ␤3) (11,23). Critically, these modifications do not alter GABA A R assembly or their functional properties (11,23,32). Following transfection, neurons were incubated for 24 h and then labeled with ␣-bungarotoxin conjugated to rhodamine (Rd-Bgt) for 15 min at 15°C in order to label cell surface BBS ␤3 (Fig. 5A). Cell surface BBS ␤ 3 labeling with Rd-Bgt was performed at 15°C to block both GABA A R endo-and exocytosis. Quantitative analysis of cell surface Rd-Bgt fluorescence intensity showed a significant increase (64 Ϯ 12.2%) in BBS ␤ 3 in neurons cotransfected with Plic-1 compared with control (Fig. 5B). Furthermore, not only an increase in cell surface BBS ␤ 3 was observed but also an increase in total BBS ␤ 3 (65 Ϯ 12.1%) pHluorin fluorescence (Fig. 5C) 6A). The insertion of BBS ␤3 increased linearly over time, reaching a plateau at ϳ10 min (Fig. 6B). This increase in rhodamine fluorescence could be specifically blocked by co-incubation with 10 g/ml unlabeled Bgt (data not shown). We therefore used a time point of 4 min to assess the influence of Plic-1 on the insertion of BBS ␤3. Hippocampal neurons were transfected with BBS ␤3 and Plic-1 or control empty vector (pRK5). 24 h later neurons were incubated with Bgt to block existing cell surface BBS ␤3 and then incubated with Rd-Bgt for 4 min to label newly inserted BBS ␤3. These experiments revealed that in the presence of Plic-1 the level of BBS ␤3 insertion increased by 50 Ϯ 10.6% compared with control (Fig. 6, C and D).

DISCUSSION
Plic-1, an established negative regulator of proteasome activity (30), has been previously documented to be associated with GABA A Rs. This interaction is mediated between the C-terminal ubiquitin-associated domain of Plic-1 and a conserved motif within the major intracellular domains of GABA A R receptor ␣ and ␤ subunits (14). Although this interaction is critical for modulating GABA A R receptor functional expression, the underlying mechanisms remain to be established.
Here we have examined the role that Plic-1 plays in regulating the membrane trafficking of GABA A Rs. To do so we compared the expression levels of heterotrimeric GABA A Rs composed of ␣1␤3 and ␥2 subunits in the presence or absence of a Plic-1 expression construct in HEK-293 cells. Co-expression with Plic-1 increased the cell surface and total expression levels of GABA A Rs in these cells. Consistent with this result, overexpression of Plic-1 in cortical neurons also significantly enhanced the cell surface expression of endogenous GABA A Rs containing ␤3 subunits. In both systems this modulatory effect of Plic-1 was independent of alterations in receptor endocytic sorting. Thus, these experiments suggest that Plic-1 acts to increase GABA A R receptor cell surface expression primarily by altering receptor trafficking or stability within the secretory pathway.
To analyze further the significance of Plic-1 for GABA A R receptor stability within the secretory pathway, we assessed its potential role in regulating the abundance of ER-retained GABA A Rs. We focused on this compartment because it represents the site where hetero-oligomers of GABA A Rs are assembled from their constituent subunits (17,33,34). As revealed by Endo H digestion, Plic-1 increased the stability of ER-retained GABA A Rs. In support of our findings, immunofluorescence and immunoelectron microscopy have revealed that Plic-1 is associated with subsynaptic cisternae, the ER, and the Golgi apparatus in fibroblasts and neurons (14). It is emerging that the residence time of GABA A R subunits within the ER is determined by their rates of ubiquitination and subsequent proteasomal degradation (6,11,14). In agreement with this mechanism we have established that Plic-1 increases the accumulation of ubiquitinated GABA A R subunits. Finally, we were also able to directly demonstrate that increasing the expression of Plic-1 in neurons specifically enhanced the insertion of GABA A receptors into the plasma membrane.
Collectively, our results suggest that Plic-1 increases GABA A R cell surface expression levels by increasing their stability within the ER. Given that the assembly of GABA A Rs from their constituent subunits in the ER is inefficient (6), this enhanced stability may increase the production of transportcompetent heteromeric receptors for insertion into the plasma membrane. In agreement with our findings, inhibiting proteasome activity has been shown to increase the abundance of assembled nicotinic acetylcholine receptors, which leads to enhanced insertion into the plasma membrane (35). Significantly, previous studies have revealed that Plic-1 can interact with the GABA A R receptor ␣1-3, ␣6, and ␤1-3 subunits (14), which is suggestive of a conserved role for this protein in regulating the cell surface stability of this structurally diverse family of inhibitory receptors. Consistent with our results with GABA A R, Plic-1 and its yeast homolog Dsk2 are associated with the proteasome and are implicated in regulating ER-associated degradation and protein targeting to aggresomes (29,30,36,37). In addition, Plic proteins appear to have a role in regulating the endocytosis of G protein-coupled receptors and function of heterotrimeric G proteins (25,38).
Finally, chronic perturbation of neuronal activity leads to bidirectional changes in ubiquitin-dependent degradation of GABA A Rs (11). Given the role of Plic-1 in ubiquitin-dependent proteasomal degradation of the GABA A R ␤ 3 subunit, it is tempting to speculate that activity-dependent ubiquitination, and thus turnover, may be regulated by binding of Plic-1 to the ␤ 3 subunit in response to changes in neuronal activity.