14-3-3ζ Mediates GABAAR Activation by Interacting with BIG1

Most fast synaptic inhibitions in the mammalian brain are mediated by GABAA receptors (GABAARs). An appropriate level of GABAAR expression at the cell surface is essential for neurodevelopment and the efficacy of GABAergic synaptic transmission. We previously reported that brefeldin A–inhibited GDP/GTP exchange factor 1 (BIG1), a binding partner of GABAARs, plays an important role in trafficking GABAARs to the cell surface. However, its regulatory mechanisms remain unknown. In the present study, we identified a new cellular protein, 14-3-3ζ, which can interact with the β subunit of GABAARs and BIG1 both in vitro and in vivo and colocalizes in the soma, dendrites, and axons of hippocampal neurons. Overexpression of 14-3-3ζ-WT increased the surface expression of BIG1 in dendrites and axons, as well as the binding of BIG1 with GABAAR. Depleted 14-3-3ζ with efficacious siRNA attenuated the interaction between BIG1 and GABAARs and resulted in significant decreases in the surface expression levels of BIG1 and GABAAR. GABAAR agonist treatment increased the expression levels of BIG1 and 14-3-3ζ on the surface, indicating that 14-3-3ζ is involved in regulating BIG1-mediated GABAAR surface expression. Depletion of BIG1 or 14-3-3ζ significantly decreased GABAAR expression at the cell surface and suppressed the GABA-gated influx of chloride ions. These data indicate that the combination of 14-3-3ζ and BIG1 is required for GABAAR membrane expression. Our results provide a potential promising therapeutic target for neurological disorders involving GABAergic synaptic transmission.


Introduction
GABA type A receptor (GABA A R) is a ligand-gated Cl --permeable channel activated by γ-aminobutyric acid (GABA), the primary inhibitory neurotransmitter in the central neural system [1]. Located at soma and proximal dendrites of neurons, GABA A R mediates inhibitory synaptic transmission and regulates neural proliferation, migration, differentiation, and maturation during development [2]. Dysfunction of the GABA A R contributes to several neuropsychiatric conditions, including epilepsy [3], mood disorders [2], anxiety and depression [4,5], schizophrenia [5], pain [6,7], and substance abuse [8]. Understanding the mechanisms that regulate the functional expression of GABA A Rs at synapses could help us to understand the causes of these disorders and provide clues for developing alternative drugs with less sedation, amnesia, and dependence.
The combination of different subunits determines the structure and functions of different GABA A R subtypes. Usually, these receptors contain two α subunits, two β subunits, and one other subunit, such as γ, δ, ε, θ, or π subunit. These subunits are assembled into heteropentameric complexes in the endoplasmic reticulum, and then are transported to the Golgi by vesicles. Finally, GABA A Rs are translocated to and expressed on the plasma membrane under the action of trafficking compartments [2]. It is without a doubt that the trafficking of GABA A Rs regulates the membrane expression and therefore impacts greatly on the functions of GABAergic synapses. Though the details are still under investigation, it is becoming clear that protein-protein interactions and many signal factors are involved in the membrane trafficking of GABA A Rs. GABA receptor-associated protein (GABARAP), which belongs to a family of ubiquitin-like proteins, and is the first identified interaction protein with GABA A R [9]. It is enriched in the Golgi apparatus and intracellular vesicles. By interaction with γ subunits of GABA A R, GABARAP facilitates the translocation of GABA A Rs to the membrane surface of hippocampal neurons [10]. The vesicular trafficking factor N-ethylmaleimide-sensitive factor (NSF), which is concentrated in the Golgi and plasma membrane, has also been found to bind directly to GABA A R β subunits. NSF and GABARAP might act together to promote the forward trafficking of GABA A Rs from the Golgi apparatus. Golgi-specific DHHC zinc-finger-domain protein (GODZ), a palmitoyltransferase that selectively restricted to the trans-Golgi network of neurons, can palmitoylate the γ2 subunit of GABA A R [11,12]. Deficiency of GODZ reduces the GABA A R expression at synapses, resulting in the decrease of amplitude and frequency of miniature inhibitory synaptic currents (mIPSCs), as well as whole-cell currents [12]. TRAK-1 and GABA A R-interacting factor 1 (GRIF1; also known as TRAK2), the member of the TRAK family of coiled-coil domain proteins, have also been found to bind with GABA A R and modulate its expression in the brain [13]. Furthermore, brefeldin A-inhibited GDP/GTP exchange factor 2 (BIG2) interacts with a sequence motif in the intracellular loop of GABA A R β subunits [14].
Our previous study showed that BIG1, which is a member of the same family as BIG2, also binds to GABA A R β subunits directly and plays an important role in the vesicular trafficking of GABA A Rs to the plasma membrane [15]. Big1 enables membrane budding of vesicles from the Golgi apparatus, and thereby facilitate proteins to transport to the plasma membrane through the trans-Golgi network [16][17][18][19]. It was found to regulate neurite development in vitro via the PI3K-AKT and ERK signaling pathways in our previous study [20]. Its function was further confirmed by a recent in vivo study. The size of the neocortex and hippocampus was decreased by BIG1 depletion, indicating its essential role in brain development, including survival of deep layer neurons, neuronal polarity, and the formation of axonal tracts [21].
However, little is known about how the GABA A R expression is regulated by BIG1. In the present study, we identified a new BIG1-binding protein 14-3-3ζ. 14-3-3ζ belongs to the 14-3-3 protein family of proteins, which are abundantly expressed in the brain, constituting approximately 1% of soluble brain proteins [22]. 14-3-3 proteins are crucial for various pathophysiologies of various neurological or neuropsychiatric disorders, such as Parkinson's disease, Alzheimer's disease, Creutzfeldt-Jakob disease, Lissencephaly, stroke, and schizophrenia, and are involved in signaling pathways, cell growth, division, adhesion, differentiation, apoptosis, and the regulation of ion channels [23,24]. In the present study, our results indicate that 14-3-3ζ recruits BIG1 in GABA A R trafficking and plays a role in regulating the surface expression of GABA A Rs.

Cell Cultures and Transfection
All experimental procedures involving animals were purchased from Southern Medical University Animal Center (Guangzhou, Guangdong, China). Hippocampal neurons were obtained from E18 Sprague-Dawley (SD) rats as described previously [25,26]. After aseptically removing the hippocampus from the skull, the tissue was freed from the meninges and incubated with 2 mg/ml papain (Sigma) for 15 min at 37 °C. Dissociated neurons were plated on 0.1 mg/ml poly-d-lysine-coated culture dishes (6×10 4 cells/ cm 2 ) in 5% FBS DMEM for 8 h at 37 °C. Subsequently, the medium was removed, and neurobasal medium containing 1% B27 supplement (Life Technologies) was added. Experiments were carried out on neurons after 7-10 days in vitro (DIV). C6 cells, HT-22 cells, or SH-SY5Y cells were cultured in DMEM (Gibco) supplemented with 10% fetal bovine serum (Gibco). Transient transfection of hippocampal neurons (7 days) or SH-SY5Y cells was carried out using Lipofectamine RNAMAX (Invitrogen) according to the manufacturer's instructions. Plasmid transfection was carried out using Lipofectamine 2000 as previously described [26].

Western Blot Analysis
Whole cell lysates were extracted from cultured neurons or from transfected HT-22 cells in lysis buffer (Beyotime) supplemented with Complete TM protease inhibitor. Brain lysates were prepared by homogenization with a glass homogenizer in the same lysis buffer. Samples of equal volumes were mixed with a 2× sample loading buffer. Equal amounts of the corresponding samples were separated by 8-15% SDS-PAGE and transferred to nitrocellulose membranes (Millipore). After blocking with 5% nonfat milk in TBS containing 0.1% Tween 20, the membrane was incubated with appropriately diluted primary antibodies overnight at 4 °C, followed by horseradish peroxidase-conjugated secondary antibodies for 1 h at room temperature. The immunoreactive bands were detected by ECL (Thermo Scientific) and visualized with a Las4000 (GE Healthcare). The intensity of the blots was quantified using MULTI GAUGE software according to the manufacturer's instructions.

Coimmunoprecipitation Assays
Rat hippocampi or cultured rat hippocampal neurons were lysed in weak intensity PRIP buffer (Beyotime) supplemented with Complete TM protease inhibitor. The homogenate was centrifuged at 10,000 × g for 15 min at 4 °C. Supernatants (300 μg of proteins in 300 μl of buffer) were incubated overnight at 4 °C with 3 μg of anti-BIG1 antibodies or control IgG and 50 μl of protein A/G beads (50% slurry, Pierce). The beads were washed six times with cold lysis buffer. The bound proteins were eluted and analyzed by Western blotting with anti-BIG1, anti-GABA A R β2,3, or anti-14-3-3ζ antibodies.
As described previously, the nanospray liquid chromatography (LC)-MS/MS analysis was used to identify the proteins immunoprecipitated with BIG1 in the lysates of five rat (postnatal day 1) hippocampi [27]. In brief, 1 mg of total proteins was immunoprecipitated with BIG1 antibodies or normal rabbit IgG (control), and then were separated and stained in 10% SDSPAGE gel with Coomassie blue R-250.
Then the gel lanes were destained and divided into segments based on molecular size markers. Proteins in each gel block were reduced, alkylated with iodoacetamide, and trypsinized. Extracted peptides were dried and dissolved in 0.1% formic acid for LC-MS/MS analysis using the LCQ Deca XP mass spectrometer and the Thermo Finnigan ProteomeX workstation (LCQ Deca XP MS). The software BioWorks 3.1 (Thermo Finnigan) was used for database search.

Immunofluorescence
Cultured hippocampal neurons plated on glass coverslips (4×10 4 cells/cm 2 ) were treated as described and fixed with immunostaining fixative (Beyotime) for 15 min. After blocking with PBS containing 10% goat serum for 30 min at room temperature (RT), the cells were labeled with cell surface GABA A R β2,3 with the appropriate antibodies for 1 h, followed by permeabilization and incubation with antibodies against BIG1, 14-3-3ζ, or MAP-2 at RT for 1 h. After washing, the cells were incubated with the appropriate secondary antibodies conjugated to Alexa Fluor 488 and Alexa Fluor 594 (1:500 dilution) at RT for 1 h. Then, coverslips were mounted on microscope slides using ProLong Gold antifade reagent (Invitrogen) and sealed with clear nail varnish. Images were acquired using a confocal microscope (Zeiss 710). Fluorescence intensity was quantified with ImageJ software (NIH) in a minimum of 20 cells per slide.

Cell-Surface Protein Isolation
Biotinylated cell-surface proteins were isolated according to the manufacturer's protocol. Briefly, after experimental treatment as described, hippocampal neurons or HT-22 cells grown in 60-mm dishes were rinsed three times with ice-cold PBS and incubated (30 min at 4 °C) in 10 ml of PBS containing 0.25 mg/ml EZ-link Sulfo-NHS-SS-Biotin (Pierce) before the reaction was quenched. Cells were collected and homogenized in 200 μl of lysis buffer (Pierce). Samples of homogenates (300 μg of protein in 300 μl) were incubated with 100 μl of immobilized NeutrAvidin Gel (Pierce) at RT for 1 h as described previously [28] and were then washed, after which bound biotinylated proteins were eluted in 100 μl of 2× LDS sample buffer containing 50 mM DTT. Samples (20 μl) were separated by electrophoretic separation by 8% SDS-PAGE for immunoblotting.

Preparation of Membrane and Cytosolic Fractions
Isolation of the membrane and cytosol was carried out using a Membrane and Cytoplasmic Protein Extraction Kit according to the manufacturer's instructions (Beyotime, Haimen, China). Briefly, cells used for fractionation were subjected to mechanical homogenization or the freeze-thaw method with liquid nitrogen after experimental treatment as described. Then, the nuclei and precipitates from a few unbroken cells are removed by low-speed centrifugation. The cell membrane and the supernatant-containing cytosolic protein were obtained by high-speed centrifugation respectively. The membrane protein was extracted from the precipitate by the optimized membrane protein extraction reagent.

MQAE Fluorescence Measurements
MQAE, a "nonratiometric" chloride ion (Cl − )-quenched fluorescent indicator, was used for intracellular Cl − to detect the GABA-gated influx of chloride ions, similar to our previous study [15]. Briefly, after washing five times with a warm Cl --free-NO 3 solution (containing 95 mM NaNO 3 , 2.5 mM KNO 3 , 1.8 mM Ca(NO 3 )2, and 10 mM Mops, pH 7.2) as described previously [29], SH-SY5Y cells were cultured in 35-mm confocal dishes labeled with the Cl --sensitive dye MQAE (5 mM) in the dark for 30 min. The cells were rinsed three times, warm Cl --free-NO 3 solution was added, and then the cells were subjected to microscopic inspection (Zeiss 710). The warm Krebs-HEPES (20 mM HEPES, 128 mM NaCl, 2.5 mM KCl, 2.7 mM CaCl 2 , 1 mM MgCl 2 , 16 mM glucose, pH 7.4) was replaced with the cells before the MQAE fluorescence measurements were made. MQAE fluorescence was excited with 405-nm lasers using a confocal microscope, and live cell images were captured for 15 ms every 5 s. The fluorescence intensity was measured according to the cell contour. GABA (100 μM) was perfused during MQAE fluorescence measurements. The normalized value of fluorescence intensity, which shows the changes in cytoplasmic concentrations of Cl -([Cl -]i), was calculated as F/F 0 , and the subscripts "F 0 " or "F" indicate the time just before or just the start of application of GABA (100 μM) perfusion. Blank control images were analyzed with a vehicle (Krebs-HEPES). Data are presented as the mean ± SEM.

Statistical Analysis
Data are expressed as the mean ± standard error of the mean (SEM). All raw data were processed by the authorized software SPSS 13.0. One-way ANOVA was used to compare the mean values of the groups, and two-way ANOVA was used to examine the significance among multiple groups. Statistical significance was accepted at P < 0.05.

14-3-3ζ Is a Binding Partner of BIG1 in Neurons
Our previous study demonstrated that BIG1 interacted with GABA A R to mediate its trafficking to the cell surface [15]. In the present study, a coimmunoprecipitation experiment was performed to identify possible modulators that recruit BIG1 and regulate GABA A R surface targeting. The extracts of five rat hippocampi were immunoprecipitational analyzed with either anti-BIG1 antibodies or control IgG. The immunoprecipitated products of BIG1 were subjected to SDS-PAGE electrophoresis and Coomassie R-250 staining (Fig. S1, supporting information (SI)). By nanospray LC-MS/MS analysis, 14-3-3ζ in the 36-26-kDa region was of special interest to us as a potential binding partner of BIG1 (Table S1). To verify this speculation and to determine the role of 14-3-3ζ in GABA A R membrane expression, a BIG1 IP experiment was performed. As shown by the immunoblotting in Fig. 1a, 14-3-3ζ was present in the precipitation of anti-BIG1 antibodies but not control IgG, indicating an interaction of 14-3-3ζ and BIG1. The content of 14-3-3ζ in the supernatant after IP showed no significant change. This could be due to the higher concentration of 14-3-3ζ in brain tissue. Consistent with our previous studies, our present results confirmed that BIG1 interacted with GABA A R β2/3 subunits. Cultured hippocampal neurons were transfected with 14-3-3ζ-EGFP plasmid to determine the association of BIG1 with 14-3-3ζ and GABA A R. Microscopy images showed that BIG1 and 14-3-3ζ were colocalized in the soma, dendrites, and axons of neurons (Fig. 1b). Similar to previous results, GABA A R colocalized with BIG1 in the plasma membrane and Golgilike apparatus (Fig. 1c), while no colocalization of GABA A R and 14-3-3ζ was observed (Fig. 1d). These results imply that 14-3-3ζ may regulate GABA A R by interacting with BIG1.

14-3-3ζ Is Essential for BIG1-Mediated GABA A R Cell Surface Expression
To further verify the role of 14-3-3ζ in regulating BIG1mediated GABA A R membrane targeting, EGFP-tagged wild-type 14-3-3ζ (14-3-3ζ-WT-EGFP) was overexpressed in HT-22 cells, a mouse hippocampal neuronal cell line, by transient transfection. An 8~12% SDS-PAGE gel was used to extract the membrane surface proteins (S) and total contents (T) of HT-22 cells separately. Western blot band analysis showed that the expression levels of BIG1 and GABA A R were increased on the cell surface (S) by overexpression of 14-3-3ζ-WT-EGFP (p<0.01 for surface BIG1, p<0.05 for surface GABA A Rs, one-way ANOVA), while their total contents (T) remained intact (Fig. 2a). Similar to HT-22 cells, BIG1 was colocalized with 14-3-3ζ-WT-EGFP in the soma, dendrites, and axons of hippocampal neurons (Fig. 2b). To investigate the effect of 14-3-3ζ on the interaction of BIG1 and GABA A R, HT-22 cells were treated with EGFP vector or 14-3-3ζ-WT-EGFP plasmids for 24 h, and then the fresh homogenates were immunoprecipitated with either anti-EGFP antibodies or control IgG, followed by immunoblotting. As shown in Fig. 2c and d, compared with EGFP overexpression, overexpression of 14-3-3ζ-WT-EGFP increased the binding of BIG1 with GABA A R, indicating an important role of 14-3-3ζ in the interaction of BIG1 and GABA A R. Together with our data from hippocampal neurons, these results suggest that 14-3-3ζ is involved in regulating BIG1-mediated GABA A R surface expression.

14-3-3ζ Depletion Attenuated the Interaction Between BIG1 and GABA A R
Since no selective inhibitor for 14-3-3ζ is available, a specific small interfering RNA of 14-3-3ζ was used to investigate the effect of 14-3-3ζ depletion on the interaction of BIG1 and GABA A R. Three 14-3-3ζ siRNA sequences (F04, F06, and F08) and a negative control sequence (NC) were transfected into rat C6 cells derived from a rat glial tumor induced by N-nitrosomethylurea, and the protein level of 14-3-3ζ was examined by western blot. As shown in Fig. 3a , the expression of 14-3-3ζ was significantly reduced by these siRNAs (p<0.001 for all three siRNAs). In particular, the sequence of F06 reduced 14-3-3ζ by approximately 95%. Here, F06 was used to downregulate 14-3-3ζ expression in hippocampal neurons. Our previous study found that BIG1 depletion can reduce the expression of GABA A R on the neuron surface and suppress synaptic strength [15]. Here, we found that the surface expression levels of BIG1 and GABA A R on hippocampal neurons (7 DIV) were decreased significantly after 14-3-3ζ depletion (p<0.001 for surface BIG1, p<0.01 for surface GABA A R, one-way ANOVA), but the total BIG1 and GABA A R contents remained unchanged (Fig. 3b). Coimmunoprecipitation experiments were performed with BIG1 antibody after 14-3-3ζ depletion. Our results revealed that more than 30% of GABA A R β2,3 subunits were decreased in the precipitation of anti-BIG1 antibodies by 14-3-3ζ depletion, suggesting that 14-3-3ζ is required for BIG1-mediated GABA A R surface expression (Fig. 3c, d).

GABA A R Activation Increases Membrane-Associated 14-3-3ζ and BIG1 in Hippocampal Neurons
It is clear that the overactivation of GABA A R rapidly increases the expression of GABA A R on the cell membrane surface [30,31]. To determine whether 14-3-3ζ is involved in this process, GABA A Rs on cultured hippocampal neurons were treated with GABA (50 μM) for 15 min, and the effects on BIG1 and 14-3-3ζ were examined. By using an 8~12% SDS-PAGE gel, membrane-associated and cytoplasmic BIG1 and 14-3-3ζ were extracted and assessed separately. As shown by the immunobands in Fig. 4a, the membrane-associated BIG1 and 14-3-3ζ, including their expression on the membranes of organelles and cytomembranes, were increased (p<0.01 for BIG1 and 14-3-3ζ, one-way ANOVA) after GABA treatment, whereas their expression levels in the cytoplasm were decreased (p<0.01 for BIG1 and p<0.001 for 14-3-3ζ, one-way ANOVA). This result implies that BIG1 and 14-3-3ζ are involved in the membrane expression of activated GABA A R. We also examined the effects of the agonist and antagonist of GABA A Rs on the expression of BIG1 and 14-3-3ζ by examining the immunofluorescence intensity. When muscimol (50 μM, 15 min), a selective agonist Fig. 1 14-3-3ζ coimmunoprecipitated and colocalized with BIG1 in hippocampal neurons. a After IP of fresh rat hippocampus homogenate or rat hippocampal neuron lysates with anti-BIG1 antibodies or control IgG, samples of precipitated proteins and lysates before and after IP were subjected to Western blotting with anti-BIG1, anti-GABA A R β2,3, or anti-14-3-3ζ antibodies. b Hippocampal neurons (7 DIV) were transfected with EGFP-14-3-3ζ-WT plasmid and then fixed and immunostained with antibody against BIG1 (scale bar, 20 μm). c, d Hippocampal neurons were reacted with anti-GABA A R β2,3 antibody after fixation. The cells were then permeabilized and stained with anti-BIG1 or anti-14-3-3ζ antibody (scale bar, 20 μm) of GABA A Rs, was applied to cultured hippocampal neurons, GABA A R expression was significantly increased compared with the effect of the vector (DMSO) (Fig. 4b, p<0.05, oneway ANOVA). However, this increase in GABA A R expression could be abolished by the coapplication of bicuculline, an antagonist of GABA A Rs (Fig. 4b, p<0.001, one-way ANOVA). Coexpression of BIG1 was increased by muscimol (Fig. 4b,  p<0.01, one-way ANOVA), whereas it was decreased by the coapplication of bicuculline (50 μM for 15 min) (Fig. 4b,  p<0.01, one-way ANOVA). In the same way, we also examined the expression of 14-3-3ζ under the administration of muscimol and muscimol-bicuculline coapplication. As shown in Fig. 4 c, muscimol administration significantly increased the expression of 14-3-3ζ, while this increase was completely abolished by the coapplication of bicuculline. These results suggest that the expression of GABA A Rs was changed with 14-3-3ζ simultaneously. These data indicate that when GABA A Rs were activated, more 14-3-3ζ and BIG1 were recruited, resulting in an increase in GABA A Rs on the cell surface.

14-3-3ζ Depletion Blocked GABA-Stimulated Clinflux
When GABA A Rs are activated by GABA, extracellular Cl − goes through GABA A Rs, resulting in Cl − influx and hyperpolarization of neurons. Meanwhile, overactivation of GABA A R leads to an increase in the receptor on the cell surface [30,31]. We have shown that 14-3-3ζ and BIG1 are required for the surface expression of GABA A R. To determine whether 14-3-3ζ and BIG1 also modulate the function of GABA A R and thereby affect Cl − influx, the effects of BIG1 and 14-3-3ζ depletion on GABA-induced Cl − influx were examined in SH-SY5Y cells, in which GABA A R was expressed and siRNA could be transfected. As shown in Fig. 5a, BIG1 siRNA (G05) and 14-3-3ζ siRNA (F06) were transfected into SH-SY5Y cells, resulting in significant decreases in the expression levels of BIG1 (vs. negative control, p<0.001, one-way ANOVA) and 14-3-3ζ (vs. negative control, p<0.001, one-way ANOVA). Cl − influx occurs when GABA A Rs on the cell surface are activated by GABA. The interaction of Cl − and intercellular MQAE resulted in a rapid decrease in fluorescence intensity, by which Cl − influx Fig. 2 Effects of 14-3-3ζ overexpression on the surface expression of GABA A R and BIG1 in HT-22 cells and hippocampal neurons. a Total (T) and cell surface proteins (S) of HT-22 cells 24 hours after transfection with EGFP vector or 14-3-3ζ-WT-EGFP plasmids were isolated and separated in an 8-12% SDS-PAGE gel, followed by immunoblotting with antibodies against BIG1, GABA A R β2,3, or 14-3-3ζ. ICAM1 and GAPDH were used as loading controls for surface and total proteins, respectively. Densitometry quantification was analyzed by ImageQuant TL software. Data are the means ± SEM of at least three independent experiments. *p<0.05, ** p<0.01. b Hippocampal neurons (7 DIV) transfected with EGFP vector or 14-3-3ζ-WT-EGFP plasmids were fixed and immunostained with antibodies against BIG antibody (scale bar, 20 μm). c HT-22 cells were treated as described in a. Total proteins were prepared for IP with anti-EGFP antibody or control IgG. Samples of precipitated proteins and lysates before and after IP were separated in 8-12% SDS-PAGE gels and immunoblotted with BIG1, GABA A R β2,3, or EGFP antibodies. d Densitometry quantification showed that the overexpression of 14-3-3ζ increased the expression of BIG1 (left) and GABA A R (right). Data are the means ± SEM of at least three independent experiments. *p<0.05, **p<0.01 ▸ Fig. 3 14-3-3ζ depletion reduced the interaction of BIG1 and GABA A R. a C6 glioma cell lines were used for small RNA interference sequence screening targeting 14-3-3ζ. C6 cells were incubated with negative control siRNA (NC) or 14-3-3ζ siRNA (F04, F06, F08) for 72 h. Total proteins were extracted and subjected to immunoblotting with antibodies against 14-3-3ζ or tubulin. Densitometry quantification was analyzed by ImageQuant TL software. Data are the means ± SEM of at least three independent experiments. ***p<0.001 vs. NC. b Total (T) and cell surface proteins (S) were isolated from hippocampal neurons (7 DIV) transfected with NC siRNA or 14-3-3ζ (F06) siRNA for 72 h and separated in an 8-12% SDS-PAGE gel, followed by immunoblotting with antibodies against BIG1, GABA A R β2,3, or 14-3-3ζ. ICAM1 and GAPDH were used as loading controls for surface and total proteins, respectively. Densitometry quantification was analyzed by ImageQuant TL software. Data are the means ± SEM of at least three independent experiments. **p<0.01, ***p<0.001 vs. NC. c Lysates from hippocampal neurons (7 DIV) transfected with NC siRNA or 14-3-3ζ (F06) siRNA for 72 h were prepared for IP with anti-BIG1 antibody or control IgG. Samples of precipitated proteins and lysates before and after IP were separated in an 8% SDS-PAGE gel and immunoblotted with anti-BIG1 or anti-GABA A R β2,3 antibodies. d Densitometry quantification of c. Data are the mean ± SEM of at least three independent experiments. **p<0.01 could be detected, reflecting the function of GABA A Rs on the cell surface. As shown by the confocal image in Fig. 5 b (NC+GABA) and movie 2 in the supplementary materials, GABA (100 μM) stimulation reduced the fluorescence intensity in control cells, indicating a fast influx of chloride ions. However, such fluorescent changes could not be observed in control cells without GABA stimulation (Fig. 5b, Blank, and movie 1 in supplementary materials) or in cells with BIG1 (Fig. 5b, G05+GABA, and movie 3 in supplementary materials) or 14-3-3ζ (Fig. 5b, F06+GABA, and movie 4 in supplementary materials) depletion. Figure 5c shows the time course of fluorescent changes after GABA stimulation in these cells. Due to bleaching, the fluorescence of intercellular MQAE was slightly decreased in control cells (Fig. 5c, black line). In negative control RNA-transfected cells, the fluorescence was significantly reduced by GABA stimulation Fig. 4 GABA A R activation increases membrane-associated 14-3-3ζ and BIG1 expression levels in hippocampal neurons. a Cultured neurons (9 DIV) treated with vehicle (DMSO) or GABA (50 μM for 15 min). Cell membrane and cytoplasmic proteins were isolated and separated in a 12% SDS-PAGE gel and immunoblotted with antibodies against BIG1 or 14-3-3ζ. Actin was used as a loading control for membrane and cytoplasmic proteins. Densitometry quantification was analyzed by ImageQuant TL software. Data are the means ± SEM of at least three independent experiments. ** or ## p<0.01, ### p<0.001 vs. NC. b Hippocampal neurons (9 DIV) treated with vehicle (DMSO), muscimol alone (50 μM for 15 min; Musc), or bicuculline (100 μM) plus muscimol (50 μΜ for 15 min, Musc + Bic). The cells were then premeabilized and stained with anti-GABA A R β2,3 (green) antibody or anti-BIG1 (red) antibody (scale bar, 20 μm). The fluorescence intensity was quantified with ImageJ software (NIH) in a minimum of 20 cells per slide. The experiments were repeated at least three times. *p<0.05, ## p<0.01, or ***p<0.001 vs. NC. c Hippocampal neurons were treated as in b and stained with anti-GABA A R β2,3 (green) antibody or anti-14-3-3ζ (red) antibody (scale bar, 20 μm). The fluorescence intensity was quantified with ImageJ software in a minimum of 20 cells per slide. The experiments were repeated at least three times. ** or ## p<0.01 vs. NC (Fig. 5c, red line, p<0.001, two-way ANOVA), indicating a fast Cl − influx and the activation of abandoned GABA A Rs. This GABA-induced decrease in MQAE fluorescence was significantly abolished in cells transfected with G05 (Fig. 5c, blue line, p<0.001, two-way ANOVA) and F06 (Fig. 5c,  green line, p<0.001, two-way ANOVA). These results suggested that the depletion of BIG1 or 14-3-3ζ resulted in the insufficiency of GABA A Rs on the cell surface and suppressed the GABA-gated influx of chloride ions.

Discussion
GABA A Rs mediate most fast synaptic inhibition in the mammalian brain, determine neuronal excitability, and play a crucial role in the physiological function of neurons. Appropriately modulating the surface dynamics of GABA A Rs is considered essential for controlling the strength of synaptic inhibition for neuropsychiatric etiology. Trafficking regulation and protein-protein interactions contribute to transport-competent GABA A R transport to and insertion into the plasma membrane after assembly from their component subunits in the endoplasmic reticulum [32]. A number of receptor-interacting proteins have been identified to be involved in regulating GABA A R membrane expression. For example, phospholipase-C-related catalytically inactive proteins (PRIPs) could bind to GABARAP and the intracellular domains of GABA A R β subunits and, more weakly, to γ2 subunits [33], through which PRIP1/2 modulates the trafficking of GABA A Rs to the synaptic membrane. In addition, PRIP1/2 promotes the de novo insertion of GABA A Rs into the plasma membrane by interacting with the serine/threonine kinase Akt [34]. Although our previous results [15] and Charych's study [14] found that BIG1 and BIG2 play important roles in maintaining the surface level of GABA A Rs by interacting with the GABA A R β subunit isoforms, the precise mechanisms are unclear. BIG1 and BIG2 belong to the brefeldin A-inhibited guanine nucleotide-exchange factor family proteins and are required for membrane budding and vesicular transport from the Golgi apparatus [35]. Studies have found that BIG1 contains A-kinase anchoring protein sequences and acts as a scaffold protein for the assembly of protein kinase A (PKA) with additional enzymes, substrates, and other regulators to form a molecular machine for the accurate reception, transduction, and integration of cAMP signals, with others initiated, propagated, and transmitted by chemical, electrical, or mechanical means [36,37]. In the intracellular domains, the GABA A R β subunit has been identified to have a binding motif with the clathrin-adaptor protein 2 (AP2) μ2 subunit [38]. The AP2 μ2 subunit interacts with GABA A R β1-3 and γ2 subunits directly in the brain to modulate GABA A R endocytosis by PKA [39], protein kinase C (PKC) [40], CaMKII [41], and protein kinase B (AKT) [42] phosphorylation and plays an important role in recruiting membrane-associated proteins into clathrincoated pits. However, to mediate the membrane expression of GABA A Rs, it is unclear what proteins are involved in constituting the transport complex with BIG1. Here, we showed that 14-3-3ζ, which has been reported to act as a phosphoserine-binding module and adaptor [23], interacts with BIG1 and the GABA A R β subunit isoforms (Fig. 1). Our data indicate that 14-3-3ζ plays an essential role in GABA A R transport complex constitution (Figs. 2, 3) and contributes to regulating the fast membrane expression, activation, and function of GABA A Rs, thereby affecting the GABA-gated influx of chloride ions (Figs. 4, 5).
14-3-3ζ was found to be a modulator involved in the cell cycle and transcriptional control, signal transduction, intracellular trafficking, and the regulation of ion channels. Initially, 14-3-3ζ described as an activator of neurotransmitter synthesis acts as an essential component of neurodevelopment and a central risk factor in the interaction network of schizophrenia proteins. Deletion of 14-3-3ζ causes neurodevelopment anomalies similar to those seen in neuropsychiatric disorders such as schizophrenia, autism spectrum disorder, and bipolar disorder [23,24]. 14-3-3ζ knockout mice display schizophrenia, and related neuropsychiatric disorders associated with 14-3-3ζ control dopamine transporter stability by interacting with DISC1, which is a schizophrenia risk factor [43,44]. In addition, the other subunits of the 14-3-3 protein family have been found to contribute to GABA receptor function. 14-3-3η mediates GABA B R-Kir3 channel activity through the decoupling of GABA B Rs from GIRK/Kir3 channels during rapid antidepressant stimulation [45]. Climbing fiber activity decreased the expression of 14-3-3-θ, impaired the interaction with protein kinase C-γ (PKC-γ), and reduced serine phosphorylation of the GABA A R γ2 subunit in cerebellar Purkinje cells. Knockdown of 14-3-3-θ or PKC-γ reduced serine phosphorylation of the GABA A R γ2 subunit in N2a cells [46]. In this study, we found that overexpression or depletion of 14-3-3ζ affects the interaction and surface expression between BIG1 and GABA A R β2,3 subunits (Figs. 2, 3).
An increasing number of results have revealed that 14-3-3 is required for efficient cell surface expression of membrane proteins, but the molecular mechanism by which 14-3-3ζ proteins promote cell surface expression remains poorly understood [47]. The essential roles of 14-3-3 in GABA A R trafficking were demonstrated by a recent report in which the autism-related protein PX-RICS forms the PX-RICS/ GABARAP/14-3-3 complex involved in GABA A R transportation during NMDAR-dependent GABAergic inhibitory long-term potentiation (iLTP) [48]. PX-RICS may interact with PI4P and GABARAP and link the N-cadherin-βcatenin cargo involved in the recruitment process to the ER membrane [49]. The 14-3-3ζ/θ heterodimer may act as a linker to mediate the interaction between ER-anchored PX-RICS and the dynein-dynactin motor complex [47]. 14-3-3 proteins allow the dynein-dynactin motor to carry different cargo proteins without canonical ER export motifs due to its substrate diversity [47]. Here, the expression levels of BIG1 and 14-3-3ζ on the cell membrane were increased in a GABA A R activation-dependent manner (Fig. 4), indicating that 14-3-3ζ may also mediate the trafficking and function of GABA A Rs in this way.
GABA A Rs are members of a large family of GABA-gated Cl − /HCO 3 − -permeable ion channels. Measurement of the intracellular Clconcentration ([Cl -]i) is usually used to detect the function and understand the regulatory mechanisms of GABA A Rs. The microelectrode technique is a classical method for the real-time investigation of ion channel functions but is difficult to use, has restricted chloride selectivity and response rate, and can only be performed in neurons with large cell bodies [50]. Here, we used the selectivity quinoline-based chloride (Cl − ) indicator dye MQAE to monitor [Cl − ]i. This dye has advantageous features, including simplicity of dye loading and fast diffusion of the dye inside the cells; in addition, it is not affected in a major way by dye bleaching and phototoxic damage and is insensitive to physiological changes in pH and bicarbonate concentration [51]. MQAE is a useful indicator of GABA-evoked transmembrane Cl − fluxes in pyramidal cells of cortical and hippocampal slices as well as in cerebellar Purkinje neurons using an MQAE-based two-photon microscopy technique [51]. Our previous study also found that the changes in [Cl − ]i correspond to the cell surface quantity of GABA A Rs [15]. In our present study, 14-3-3ζ depletion blocked GABAstimulated Clinflux, similar to the BIG1 deletion (Fig. 5). These data indicate an important role of 14-3-3ζ in regulating the GABA A R-mediated Cl − influx through trafficking the BIG1-GABA A R complex to the cell surface.

Conclusions
Receptor dynamic modulation has important implications for generating and maintaining an appropriate receptor cell-surface localization and synaptic function. Stabilization of GABA A Rs on the plasma membrane is likely to be facilitated by multiple mechanisms. Thus, 14-3-3ζ regulates GABA A R activation and surface expression by interacting with BIG1, which might provide an additional mechanism for the control of inhibitory synaptic strength. Alterations in these processes could result in the development of CNS pathologies.
Author Contribution Cuixian Li, Chun Zhou, and Jie Tang contributed to the study conception and design. Material preparation, data collection, and analysis were performed by all authors. The first draft of the manuscript was written by Cuixian Li and Jie Tang. All authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.
Funding This work was supported by the "technology innovation 2030-major projects" on brain science and brain-like computing of the Ministry of Science and Technology of China (No. 2021ZD0202603), and the National Natural Science Foundation of China (grant number 31500841).

Data Availability
The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.

Declarations
Ethics Approval This study was approved by the Institutional Animal Care and Use Committee (IACUC) at Southern Medical University. Principles established by the Committee of Southern Medical University on Animal Care were followed at all times.

Consent for Publication Not applicable.
Research involving Human Participants and/or Animals No human subject was involved in this study. Procedures involving animal subjects were approved by the Institutional Animal Care and Use Committee (IACUC) at Southern Medical University.
Informed Consent Not applicable.

Competing Interests
The authors declare no competing interests.