Phospholipase C-related but Catalytically Inactive Protein Is Required for Insulin-induced Cell Surface Expression of γ-Aminobutyric Acid Type A Receptors*

The γ-aminobutyric acid type A (GABAA) receptors play a pivotal role in fast synaptic inhibition in the central nervous system. One of the key factors for determining synaptic strength is the number of receptors on the postsynaptic membrane, which is maintained by the balance between cell surface insertion and endocytosis of the receptors. In this study, we investigated whether phospholipase C-related but catalytically inactive protein (PRIP) is involved in insulin-induced GABAA receptor insertion. Insulin potentiated the GABA-induced Cl− current (IGABA) by about 30% in wild-type neurons, but not in PRIP1 and PRIP2 double-knock-out (DKO) neurons, suggesting that PRIP is involved in insulin-induced potentiation. The phosphorylation level of the GABAA receptor β-subunit was increased by about 30% in the wild-type neurons but not in the mutant neurons, which were similar to the changes observed in IGABA. We also revealed that PRIP recruited active Akt to the GABAA receptors by forming a ternary complex under insulin stimulation. The disruption of the binding between PRIP and the GABAA receptor β-subunit by PRIP interference peptide attenuated the insulin potentiation of IGABA. Taken together, these results suggest that PRIP is involved in insulin-induced GABAA receptor insertion by recruiting active Akt to the receptor complex.

The ␥-aminobutyric acid type A (GABA A ) receptors play a pivotal role in fast synaptic inhibition in the central nervous system. One of the key factors for determining synaptic strength is the number of receptors on the postsynaptic membrane, which is maintained by the balance between cell surface insertion and endocytosis of the receptors. In this study, we investigated whether phospholipase C-related but catalytically inactive protein (PRIP) is involved in insulin-induced GABA A receptor insertion. Insulin potentiated the GABA-induced Cl ؊ current (I GABA ) by about 30% in wild-type neurons, but not in PRIP1 and PRIP2 double-knock-out (DKO) neurons, suggesting that PRIP is involved in insulin-induced potentiation. The phosphorylation level of the GABA A receptor ␤-subunit was increased by about 30% in the wild-type neurons but not in the mutant neurons, which were similar to the changes observed in I GABA . We also revealed that PRIP recruited active Akt to the GABA A receptors by forming a ternary complex under insulin stimulation. The disruption of the binding between PRIP and the GABA A receptor ␤-subunit by PRIP interference peptide attenuated the insulin potentiation of I GABA . Taken together, these results suggest that PRIP is involved in insulin-induced GABA A receptor insertion by recruiting active Akt to the receptor complex.
The ␥-aminobutyric acid (GABA) 4 type A (GABA A ) receptors are GABA-gated chloride channels that mediate the majority of fast synaptic inhibition in the central nervous system (1)(2)(3)(4)(5). The perturbation of GABA-GABA A receptors-mediated neurotransmission causes several central nervous system disorders including motor coordination, anxiety, insomnia, schizophrenia, and epilepsy. Additionally, GABA A receptors are important therapeutic drug targets for sedative, anxiolytic, anticonvulsant, and hypnotic agents (1)(2)(3)(4)(5). Therefore, it is important to uncover how synaptic strength is regulated in GABAergic transmission. The GABA A receptors are heteropentamers composed of a combination of 18 GABA A receptor subunits, which are divided into seven subunit classes (␣1-6, ␤1-3, ␥1-3, ␦, ⑀1-3, , and ) based on their sequence homology (1)(2)(3)(4)(5). Each receptor subunit has a similar structure with a large N-terminal extracellular region, which is the binding site for GABA and psychoactive drugs such as benzodiazepines, followed by four hydrophobic transmembrane domains (TM1-4) with a large intracellular loop region between TM3 and 4. This intracellular loop region is a target for proteinprotein interactions, phosphorylation, ubiquitination, and palmitoylation, which control receptor trafficking, stability, and clustering on the synaptic membrane (1)(2)(3)(4)(5). Regulation of the number of receptors on the postsynaptic membrane is one of the key factors for determining synaptic strength, which is maintained by a balance between the insertion and endocytosis of receptors to/from the cell surface. Recently, it was reported that the dephosphorylation of the GABA A receptor ␤or ␥2-subunit triggers endocytosis by facilitating the binding to the 2-subunit of adaptor protein 2 (AP2) complex, a critical component of clathrin-dependent endocytosis (6 -9). On the other hand, it was reported that insulin stimulates GABA A receptor insertion into the cell surface membrane via Akt-mediated phosphorylation of the GABA A receptor ␤-subunit (10 -14).
We previously identified a new inositol 1,4,5-trisphosphatebinding protein from rat brain lysate by affinity column chromatography (15). Our subsequent studies on the characterization of the protein revealed that 1) it has a domain organization similar to ␦-type phospholipase C (PLC) but has no PLC activity, which is the reason for its name, PRIP (PLC-related but catalytically inactive protein) (16,17). 2) PRIP has two isoforms, PRIP1 and 2, which are expressed mainly in the brain and ubiquitous organs, respectively (18 -20). 3) PRIP knock-out mice are less sensitive to benzodiazepine-type drugs, such as diazepam, suggesting that the cell surface expression of ␥-subunitcontaining GABA A receptors is diminished in these mutant mice (21,22). 4) PRIP facilitates GABA A receptor-associated protein (GABARAP) mediated cell surface expression of ␥2-subunit-containing GABA A receptors by acting as a bridging molecule between GABARAP and receptors (22)(23)(24). 5) PRIP regulates the phosphorylation level of the GABA A receptor ␤-subunit by binding to protein phosphatases (25)(26)(27). 6) PRIP is involved in clathrin-dependent constitutive endocytosis of GABA A receptors (28). We also have reported that PRIP modulates brain-derived neurotrophic factor (BDNF)-induced GABA A receptor endocytosis through the regulation of the receptor phosphorylation level (29). These results suggest that PRIP regulates GABA A receptor function through receptor trafficking, phosphorylation, and endocytosis (30,31).
In this study, we investigated whether PRIP is involved in insulin-induced GABA A receptor insertion. Insulin potentiated the GABA-induced Cl Ϫ current (I GABA ) by about 30% in wildtype (WT) hippocampal neurons but not in neurons derived from PRIP1 and PRIP2 double knock-out (DKO) mice. The phosphorylation level of the ␤-subunit was increased by about 30% in the WT neurons but not in the DKO neurons, which was similar to the changes observed in I GABA . Using an immunoprecipitation assay and a glutathione S-transferase (GST) pulldown assay using brain lysate together with a HEK293 reconstitution system we revealed that PRIP recruited active Akt to GABA A receptors. The disruption of the binding between PRIP and the ␤-subunit by PRIP interference peptide attenuated the insulin-potentiated I GABA . Interestingly, pretreatment with brefeldin A (BFA), an inhibitor of anterograde trafficking from the ER to the Golgi (32, 33) decreased I GABA under insulin treatment. Collectively, these results suggest that PRIP plays an important role in insulin-induced GABA A receptor insertion by recruiting active Akt to the receptor complex.

EXPERIMENTAL PROCEDURES
Chemicals, Plasmids, and Animals-Insulin and okadaic acid were obtained from Wako. Wortmannin, BFA, and crosstide were purchased from Sigma. The PRIP1-(553-565) peptide and its scrambled peptide were described previously (29). Anti-PRIP1 and anti-PRIP2 polyclonal antibodies were described previously (20,21). Anti-Akt polyclonal antibody, antiphospho-Akt (Thr-308 or Ser-473) polyclonal antibodies, and antiinsulin receptor ␤-subunit monoclonal antibody (clone 4B8) were purchased from Cell Signaling. Anti-GABA A receptor ␣1-subunit and anti-GABA A receptor ␥2-subunit polyclonal antibodies were obtained from Alpha Diagnostic International. Anti-GABA A receptor ␤2/3-subunit monoclonal antibody (clone 62-3G1) and anti-N-ethylmaleimide-sensitive factor (NSF) polyclonal antibody were from Upstate. Anti-GST polyclonal antibody was purchased from Santa Cruz Biotechnology. Horseradish peroxidase-conjugated anti-rabbit and anti-mouse IgG were obtained from GE Healthcare. [ 32 P]Orthophosphate (5.55 GBq/ml) and [␥-32 P]ATP (185 MBq/ml, specific activity: 111 TBq/mmol) were purchased from PerkinElmer. Construction of the mammalian expression vectors for the Myc-and FLAG-tagged GABA A receptor subunit (␣1, ␤2, and ␥2S) was described previously (34). Briefly, the Myc or FLAG tag was introduced between amino acids 4 and 5 of the mature form of each receptor subunit. For the mammalian GST fusion protein expression vector pcDNA3.1(-)/GST1, GST was amplified using primers M-81, 5Ј-AAA AAG CTA GCC ACC ATG TCC CCT ATA CTA GG-3Ј (underlining denotes the NheI site) and M-82, 5Ј-AAA AAC TCG AGA TCG ATA CCG TCG ACC TCG A-3Ј (underlining denotes the XhoI site) and pGST4 as a template. The PCR products were digested using NheI/XhoI and cloned into the same sites of pcDNA3.1(-). The rat PRIP1 (rPRIP1) was amplified using primers M-85, 5Ј-AAA AAC TCG AGC ATG GCT GAG GGC GCG GCT A-3Ј (underlining denotes the XhoI site) and M-86, 5Ј-AAA AAA AGC TTT CAC AAC TTC CCG TTC TCT TC-3Ј (underlining denotes the HindIII site) and pcMT31 (16) as a template. The PCR products were digested using XhoI/ HindIII and cloned into the same sites of pcDNA3.1(-)/GST1 to produce pcDNA3.1(-)/GST1-rPRIP1. The PRIP1 expression plasmid pSG5/rPRIP1 was described previously (16). The mammalian expression vector for Akt pECE/Akt was kindly provided by Dr. U. Kikkawa (Kobe University, Japan) (35). The generation of the DKO mice was described previously (22,29). The handling of the mice and all procedures were approved by the Animal Care Committee of Kyushu University, according to the guidelines of the Japanese Council on Animal Care.
Electrophysiology-Electrophysiological measurements were performed in acutely isolated hippocampal CA1 pyramidal neurons using the conventional whole cell patch-clamp technique. The acutely dissociated neurons were prepared from postnatal day 10 -14 WT or DKO mice, as described previously (36). All recordings were performed under voltage clamp conditions at a holding potential of Ϫ50 mV and a patch-clamp amplifier (EPC-7plus, HEKA Instruments Inc). All experiments were performed at a room temperature of 22-25°C. The ionic composition of patch pipette solution containing 80 mM KCl, 70 mM potassium methanesulfonate, 4 mM ATP-Mg, 2 mM EGTA, 1 mM MgCl 2 , 10 mM HEPES, and adjusted pH to 7.2 with Tris-base. Extracellular solution containing 150 mM NaCl, 2.5 mM KCl, 2 mM CaCl 2 , 1 mM MgCl 2 , 10 mM HEPES, and 10 mM glucose. The pH was adjusted to 7.4 with Tris-base. Reagents dissolved in extracellular solution were applied by using the Y-tube perfusion system, which allows rapid exchange of the solution surrounding a cell (37,38). All data are expressed as the means Ϯ S.D.
Cell Culture and Transfection-HEK293 cells were grown in Dulbecoo's modified Eagle's medium (DMEM) containing 10% fetal bovine serum supplemented with 100 units/ml penicillin and 0.1 mg/ml streptomycin. The cells were maintained at 37°C in a humidified 5% CO 2 incubator. Plasmid transfection was performed using the calcium phosphate method as described elsewhere (39) or Lipofectamine 2000 (Invitrogen) according to the manufacturer's protocol. Briefly, 1.5 g of each GABA A receptor subunit (␣1, ␤2, and ␥2S) with or without 2.5 g of pSG5/rPRIP1 and/or pECE/Akt were transfected into 7.5 ϫ 10 5 cells. For the GST pull-down assay, 1.0 g of pcDNA3.1(-)/GST1 or pcDNA3.1(-)/GST1-rPRIP1 was transfected with 2.5 g of pECE/Akt. After 48 h of incubation, the cells were used for each experiment. Cortical neurons were prepared from postnatal day 0 (P0) WT or DKO mice, as described previously (21,29) and were cultured for 14 -18 days in vitro (DIV) before the experiments.
Immunoprecipitation, GST Pull-down, and Western Blotting-Cell lysates were prepared from cortical neurons or plasmidtransfected HEK293 cells using ice-cold lysis buffer containing 50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 5 mM EDTA, 1 mM EGTA, 1% Triton X-100, phosphatase inhibitors (50 mM NaF, 10 mM Na 4 P 2 O 7 , 20 mM ␤-glycerophosphate, and 1 mM Na 3 VO 4 ), and protease inhibitors (1 mM phenylmethylsulfonyl fluoride, 100 M (p-amidinophenyl)methanesulfonyl fluoride hydrochloride, 10 g/ml leupeptin, 10 g/ml pepstatin A, and 3.4 g/ml aprotinin). The mouse whole brain lysates of WT or DKO mice were also prepared using the same buffer. In the case of co-precipitation of NSF, 0.5 mM ATP was added to the lysis buffer. The lysates were subjected to immunoprecipitation using the indicated antibodies. For the GST pull-down assay, 20 l of glutathione-Sepharose TM 4B (GE Healthcare) were added to cell lysates expressing the GST fusion protein. The immunocomplex was washed five times with 1 ml of ice-cold lysis buffer containing phosphatase inhibitors. The lysates and immunocomplexes were subjected to SDS-polyacrylamide gel electrophoresis (PAGE) and then transferred onto polyvinylidene fluoride membrane. Western blotting was performed using the indicated antibodies, and signals were detected using the ECL plus Western blotting detection system (GE Healthcare) and LAS3000 mini (Fuji Film). 32 P Labeling of Cultured Neurons-Cultured cortical neurons (DIV 14 -18) were incubated with 1 ml of phosphate-free DMEM for 1 h and then labeled with 7.4 MBq/ml of [ 32 P]orthophosphate for 4 h at 37°C. The neurons were stimulated with 500 nM insulin for the indicated times at 37°C. The cells were then washed twice with ice-cold phosphate-buffered saline and extracted with 500 l of ice-cold lysis buffer containing phosphatase inhibitors and protease inhibitors. The cell lysates were subjected to immunoprecipitation using an anti-GABA A receptor ␤2/3-subunit monoclonal antibody. The immunocomplexes were washed five times with 1 ml of ice-cold lysis buffer containing phosphatase inhibitors and subjected to SDS-PAGE. Phosphorylated proteins were detected by autoradiography using a Bio-Image analyzer BAS2500 (Fuji Film). shows the amplitude of I GABA normalized to that seen without insulin. All data are represented as means Ϯ S.D. Significance was determined using the Student's t test (**, p Ͻ 0.01, compared with the results from DKO). B, phosphorylation of the ␤-subunit in response to insulin stimulation. The cultured cortical neurons (DIV. 14 -18) of the WT or DKO mice were metabolically labeled with [ 32 P]orthophosphates for 4 h. The neurons were stimulated with 500 nM insulin for the indicated time, and then the cell lysates were subjected to immunoprecipitation using an anti-GABA A receptor ␤2/3-subunit antibody. The immunocomplexes were separated by SDS-PAGE and then subjected to autoradiography. Phosphorylated bands were detected using BAS2500. The autoradiograph represents one of four independent experiments. The other experiments gave similar results. The graph shows quantitative data concerning the phosphorylation of the GABA A receptor ␤2/3-subunit of WT (closed circles) or DKO (open circles) neurons. As mentioned above, 32 P incorporation was analyzed, because the phosphospecific antibody currently available recognizes the di-phosphorylated ␤3-subunit at both Ser-408 and Ser-409 (29,41), and insulin causes a single phosphorylation at Ser-410 or Ser-409 of ␤2or ␤3-subunit, respectively (11,12). Data are represented as means Ϯ S.D. (n ϭ 4). Significance was determined using the Student's t test (*, p Ͻ 0.05; **, p Ͻ 0.01, compared with the results from DKO). FEBRUARY 12, 2010 • VOLUME 285 • NUMBER 7

JOURNAL OF BIOLOGICAL CHEMISTRY 4839
Akt Kinase Assay-The Akt kinase assay was described previously (40). Briefly, immunocomplexes created using an anti-Akt polyclonal antibody were washed once with ice-cold 1 ml of Akt kinase assay buffer (50 mM Tris-HCl, pH 7.5, 10 mM MgCl 2 , and 1 mM dithiothreitol) and then resuspended in 30 l of the same buffer containing 100 M peptide substrate, crosstide, and 10 M [␥-32 P]ATP (37 kBq/reaction). After incubation for 30 min at 30°C, the reaction was stopped by adding 10 l of 300 mM H 3 PO 4 . The reaction products were spotted onto peptide binding paper (Whatman P81 cation exchange paper) and then washed three times with 75 mM H 3 PO 4 to remove nonspecific radioactivity. After drying, the paper was subjected to liquid scintillation counting. Data are expressed as means Ϯ S.D.

RESULTS
Insulin Potentiates GABA-induced Cl Ϫ Current and Phosphorylation of the GABA A Receptor ␤-Subunit in WT but Not DKO Neurons-In this study, we investigated whether PRIP is involved in insulin-induced GABA A receptor insertion. We first investigated the effect of insulin on I GABA using acutely isolated hippocampal CA1 neurons from either WT or DKO mice. I GABA was increased maximally by 30% during 15 min of 1 M insulin stimulation in WT neurons (Fig. 1A), which is consistent with previous reports by other groups (10 -13). In the DKO neurons, however, no insulin effect was observed (Fig. 1A).
We next investigated insulin-induced phosphorylation of the GABA A receptor ␤-subunit using cortical neurons from each genotype because insulin-induced membrane insertion of GABA A receptors is accompanied by the phosphorylation of the ␤-subunit (11,12). For this purpose, cultured neurons were metabolically labeled with [ 32 P]orthophosphate for 4 h and then stimulated with 500 nM insulin for 5, 15, or 30 min. The GABA A receptor ␤-subunits were precipitated using an anti-GABA A receptor ␤2/3-subunit antibody, followed by separation by SDS-PAGE and autoradiography. As shown in Fig.  1B, phosphorylation of the GABA A receptor ␤-subunit was increased by about 30% after 5 min of insulin stimulation and continued for 30 min in WT, but no such increase in phosphorylation was observed in the DKO neurons. These results suggest that PRIP participates in the insulin-induced phosphorylation of the GABA A receptor ␤-subunit, leading to the insulin-induced potentiation of I GABA .
Two possible explanations for the low phosphorylation of the ␤-subunit observed in DKO neurons are higher activity of phosphatases or lower activity of kinases. We previously reported that PRIP participates in the regulation of the phosphorylation level of the GABA A receptor ␤-subunit by acting as a scaffolding protein for protein phosphatases (PP1 and PP2A) (27,29). So, we investigated the effects of protein phosphatase inhibitors on the insulin-induced potentiation of I GABA . We pretreated neurons with 10 M okadaic acid for 15 min, which inhibits both PP1 and PP2A at this concentration (42) and then measured the effect of insulin on I GABA . Pretreatment with okadaic acid had no effect on I GABA in the presence of insulin stimulation with DKO neurons (Fig. 2A, right panel). If higher phosphatase activity is responsible for the low phosphorylation of the ␤-subunit, okadaic acid would have increased I GABA . WT neurons exhibited a marginal increase of I GABA insulin potentiation, but the effect was not significant (Fig. 2A, left  panel). These results indicate that the PRIP deficiency caused the failure of kinase(s) action, rather than the regulation of phosphatases.
It is well known that insulin activates the phosphatidylinositol 3-kinase (PI 3-kinase)-Akt signaling pathway (43,44) and the Akt-mediated phosphorylation of the GABA A receptor ␤-subunit, and the potentiation of miniature inhibitory postsynaptic currents (mIPSCs) is also reported to require the process (11,12). We pretreated neurons with 100 nM wortmannin, a potent PI 3-kinase inhibitor (45) for 15 min prior to insulin stimulation. Consistent with previous reports (12,13), pretreatment with wortmannin completely blocked the insulin potentiation of I GABA in WT neurons (Fig. 2B, left panel), while wortmannin had no effect on the I GABA in DKO neurons (Fig. 2B, right  panel), confirming that the PI 3-kinase signaling pathway was required in our experiments.
Akt Activation following PI 3-Kinase Activation in Response to Insulin Stimulation-To find reasonable explanations for the failure of the phosphorylation of the GABA A receptor ␤-subunit in response to insulin stimulation observed in the DKO neurons, we investigated whether PRIP deficiency impaired the insulin signaling pathway. For this purpose, we first examined the expression level of molecules involved in insulin signaling. Cortical neurons from WT or DKO mice were cultured until 14 -18 DIV. After serum starvation for 4 h, the neurons were stimulated with 500 nM insulin for 5, 15, or 30 min, and cell lysates were prepared, followed by Western blotting using relevant antibodies. As shown in Fig. 3A, PRIP deficiency had no effect on the expression levels of the insulin receptor ␤-subunit, Akt, or several GABA A receptor subunits, and insulin stimulation for up to 30 min also had no effect on the expression of these molecules.
We then examined Akt activation in response to insulin stimulation by monitoring the Akt phosphorylation at the Thr-308 and Ser-473 residues using antiphospho-Akt antibodies. As shown in Fig. 3B, the phosphorylation of Akt, an index of Akt activation was increased at 5 min stimulation and sustained for 30 min, which did not differ between the WT and DKO neurons. Wortmannin completely  (n ϭ 3). Significance was determined by Student's t test (*, p Ͻ 0.05; **, p Ͻ 0.01, compared with the result before insulin stimulation), but no difference was detected between WT and DKO. FEBRUARY 12, 2010 • VOLUME 285 • NUMBER 7 blocked the increase in phospho-Akt (results not shown). The activity of Akt was biochemically assayed: immunoprecipitates of anti-Akt antibody attached to WT or DKO neurons stimu-lated with insulin for 15 min were subjected to an Akt kinase assay in vitro using crosstide as a substrate and [␥-32 P]ATP. As shown in Fig.  3C, the immunoprecipitates from the neurons stimulated with insulin exhibited an ϳ2-fold increase of 32 P radioactivity incorporation, and there was no significant difference between the genotypes. The results indicate that Akt kinase activation, which is probably responsible for insulin-induced phosphorylation of the GABA A receptor ␤-subunit (11,12), was not impaired by PRIP deficiency.

PRIP Facilitates Complex Formation between GABA A Receptors and
Akt-Because PRIP deficiency had no effect on insulin-induced Akt activation but caused the impairment of the insulin-induced phosphorylation of the GABA A receptor ␤-subunit, we hypothesized that PRIP might function as a scaffolding molecule that makes Akt more accessible to the GABA A receptor ␤-subunit. To examine this possibility, we performed a co-immunoprecipitation assay using brain lysates. The brain lysates prepared from WT or DKO mice were immunoprecipitated using an anti-GABA A receptor ␤2/3-subunit and then analyzed by Western blotting using anti-PRIP1, anti-PRIP2, and anti-Akt antibodies. Assessment by Western blotting of the amount of immunoprecipitated GABA A receptor ␤2/3-subunits was not possible because the corresponding bands overlapped with that for the immunoglobulin heavy chain used for the immunoprecipitation. However, we confirmed in advance that the antibody we used was able to precipitate the GABA A receptor ␤-subunit using a HEK293 reconstitution system with GABA A receptor subunits in combination with the [ 35 S]methionine pulse-chase technique (results not shown). Consistent with our previous reports (27,28), PRIP1 and -2 in the WT brain lysates were co-immunoprecipitated with the GABA A receptor ␤-subunit (Fig. 4A, right panel). There were no corresponding bands for PRIP1 or 2 in the immunocomplexes produced from the DKO brain lysates (Fig. 4A, right panel). The amount of Akt co-precipitated with the GABA A receptor ␤2/3-subunit was much greater in the WT lysates than in the DKO lysates (Fig. 4A, right panel), indicating that PRIP promotes complex formation between the ␤2/3-subunit and Akt. It is noteworthy that PRIP deficiency caused no effect on the direct binding between GABA A receptor ␤2/3subunit and NSF, one of the ␤-subunit-binding proteins (46) (Fig. 4A, right panel). We next investigated whether insulin affects this complex formation, using a cultured reconstitution system. We exogenously expressed Myc-tagged GABA A receptor subunits (␣1, ␤2, and ␥2S) and Akt, with or without PRIP1 in HEK293 cells, which intrinsically contain trace amounts of PRIP1 and 2. After insulin stimulation for 5, 15, or 30 min, the cell lysates were subjected to immunoprecipitation using an anti-Myc antibody, followed by Western blotting for Akt. The amount of immunoprecipitated GABA A receptors was not apparent for the same reason as mentioned above. A small amount of Akt was seen in fractions co-precipitated with GABA A receptors in the nonstimulated cells. Insulin stimulation only increased the amount of Akt co-precipitated with GABA A receptors in the PRIP-expressing cells (Fig. 4B, the left and right panels show typical blots and a summary of multiple experiments, respectively), suggesting that PRIP facilitates complex formation between the GABA A receptor and Akt under insulin stimulation.
We next examined the direct binding between PRIP1 and Akt using an in vivo GST pull-down assay. Genes for GST or GST-rat(r)PRIP1 were transfected with Akt into HEK293 cells. After insulin stimulation for 15 min, GST alone or GST-rPRIP1 was precipitated from the cell lysates using glutathione-conjugated beads, followed by Western blotting for Akt and phospho-Akt. As shown in Fig. 4C, GST-rPRIP1, but not GST, bound to Akt when the cells were stimulated with insulin. Taken together, these results suggest that PRIP recruits phosphorylated (active) Akt to GABA A receptors by forming a ternary complex under insulin stimulation. Thus, PRIP might be implicated in Akt-dependent phosphorylation of GABA A receptors, leading to their insertion into the cell surface membrane.
PRIP1-(553-565) Peptide Attenuates Insulin Potentiation of I GABA -We next investigated whether such complex formation is important for the insulin potentiation of I GABA . To address this issue, PRIP1-(553-565) peptide at 3 g/ml, which reduces the binding between PRIP1 and GABA A receptor ␤-subunit in cultured cells (29), was applied into WT neurons through a patch pipette, and then I GABA was measured in the presence of insulin. As shown in Fig. 5, the PRIP1-(553-565) peptide but not the control peptide (scrambled peptide of PRIP1-(553-565)) partially attenuated the insulin potentiation of I GABA , indicating that the association between the ␤-subunit and PRIP is important for making Akt accessible to the receptor ␤-subunit, resulting in the potentiation of I GABA .
BFA Reverses the Effect of Insulin on I GABA -Insulin triggers the activation of Akt, leading to the phosphorylation of GABA A receptors, which are resistant to internalization, by inhibiting its association with AP2 complex (6 -9). Therefore, the apparent potentiation of I GABA observed in WT neurons could have resulted from the inhibition of insulin-induced internalization rather than insulin-induced facilitation of GABA A receptor insertion. To examine this possibility, we pretreated WT neurons with 5 g/ml BFA, an inhibitor of anterograde trafficking from the ER to the Golgi (32, 33) for 15 min and measured the effect of insulin on I GABA . If the assumption is correct, BFA would have had little effect; however, BFA caused further decreases in I GABA below the control level after insulin treatment (Fig. 6, left panel). This effect was not observed under nonstimulated conditions (time: Ϫ6 to 0 min) (Fig. 6, left  panel). BFA had no effect on I GABA in the DKO neurons (Fig. 6,  right panel). The results suggest that insulin induces both GABA A receptor insertion and subsequent endocytosis of the GABA A receptor and that this insertion mainly occurs in WT neurons. Additionally, the result indicates that PRIP is an important molecule in the mechanism that allows insulin to execute its effects on GABA A receptor trafficking.

DISCUSSION
It has been reported that insulin triggers rapid translocation of functional GABA A receptors from the intracellular pool to the cell surface membrane, thus increasing the amplitude of GABA A receptor-mediated mIPSCs (10). The underlying molecular mechanisms have been proposed as follows: insulin elicits tyrosine phosphorylation at residues Tyr-372 and Tyr-379 of the ␤2-subunit of GABA A receptors by unknown kinase(s), and these phosphotyrosines are then recognized by the SH2 domain of p85 (13), a regulatory subunit of PI 3-kinase. The PI 3-kinase bound to GABA A receptors produces phosphatidylinositol 3,4,5-trisphosphate, an upstream activator of serine/ threonine kinase Akt around the receptors, leading to the phosphorylation of the intracellular loop region of the ␤-subunits (Ser-409 in ␤1, Ser-410 in ␤2, or Ser-409 in ␤3), which is essential for the membrane insertion of GABA A receptors (11,12).
The current study was motivated by the finding that no effect of insulin on the potentiation of I GABA was seen in neurons derived from PRIP-deficient mice, indicating that PRIP is involved in process(es) triggered by insulin stimulation. Therefore, our studies exploring the possible mechanisms in which PRIP is implicated have continued to examine each step involved in known insulin signaling pathways (43,44). PRIP deficiency neither perturbs protein expression profiles including those of insulin receptors and Akt nor impairs the activation of Akt in whole cell extract, as assessed by the phosphorylation of residues Thr-308 and Ser-473 and in vitro enzymatic activity using synthetic peptide substrate. However, the phosphorylation of the GABA A receptor ␤-subunit in neurons from DKO mice was not augmented by insulin stimulation. This phenomenon is probably attributed to the fact that Akt is not accessible to GABA A receptor ␤-subunits in the absence of PRIP. Based on the observations, we propose that PRIP functions as a scaffolding protein that presents the active form of Akt to GABA A receptors, enabling insulin signaling to potentiate I GABA (Fig. 7).
We previously reported that PRIP is involved in the regulation of BDNF-induced endocytosis of GABA A receptors (29). In this case, PKC, which directly associates with the GABA A receptor ␤-subunit, is activated by BDNF stimulation and triggers the phosphorylation of the Ser-408 and Ser-409 residues of the ␤3-subunit. These residues are subsequently dephosphorylated by PP2A, which is recruited to the vicinity of the receptors via PRIP (29,41). Thus, BDNF stimulation triggers transient phosphorylation of the ␤-subunit, followed by longlasting dephosphorylation. The 2-subunit of AP2 complex were pretreated with 5 g/ml of BFA, which inhibits anterograde trafficking from the ER to the Golgi apparatus (32, 33) for 15 min and throughout the experiment. The experiment was performed in the same way as that described for Fig. 1A. All data are represented as means Ϯ S.D. The I GABA from either WT or DKO without BFA, which were taken from those shown in Fig. 1A, are also shown as references. Significance was determined using the Student's t test (**, p Ͻ 0.01, from the results obtained in the absence of the drug). Double-headed arrows indicate the time period of insulin stimulation. none, no drug; BFA, brefeldin A. specifically binds to the dephosphorylated form of the ␤-subunit, leading to clathrin-mediated endocytosis of the GABA A receptor (6 -8). Correspondingly, a transient increase and a subsequent long-lasting decrease in I GABA is observed (29,41). On the other hand, in DKO neurons, BDNF caused a gradual increase in the phosphorylation of the ␤-subunit and therefore of I GABA , which lasted for the full 30-min examination period (29). Taken together, these results indicate that different extracellular stimuli evoke phosphorylation of the ␤-subunit of the GABA A receptor at the same residues via different kinases, the level of which is regulated by the balance of activity between kinase(s) and phosphatase(s) especially in the vicinity of GABA A receptors but not inside cells. Therefore, the time courses of the phosphorylation level appear to be dependent on the type of stimuli involved. In either case, PRIP through direct association with the GABA A receptor ␤-subunit, plays an important role in recruiting proteins including the active form of Akt (this study) and protein phosphatases (PP1 and PP2A) (27,29), which regulate the phosphorylation of the GABA A receptor ␤-subunit, leading to the regulation the number of receptors on the cell surface membrane. In fact, we observed a further decrease of insulin-mediated I GABA below the control level in the BFA-treated WT but not DKO neurons. This observation suggests that insulin elicits both the insertion and subsequent endocytosis of GABA A receptors and that the balance shifts to membrane insertion in insulin-stimulated WT neurons. The result also suggests that PRIP is involved in both insulin-induced membrane insertion and endocytosis of GABA A receptors. Other scaffolding molecules such as receptor for activated C kinase-1 (RACK-1) for protein kinase C (47,48), protein kinase A-anchoring protein (AKAP) 79/150 for cAMP-dependent protein kinase A (PKA) (49), and PRIP (27,29) have been reported to determine the specificity of the specific kinase(s) or phosphatase(s) recruited to the vicinity of GABA A receptors. We still do not know the exact molecular mechanisms by which different stimuli regulate the recruitment of kinase(s) and phosphatase(s) to the vicinity of GABA A receptors. The phosphorylation state of the scaffolding molecules may be one of the pathways that regulates the interaction among these molecules. Additionally, the molecular mechanisms by which phosphorylation of ␤-subunit triggers the membrane insertion of GABA A receptors remains largely unknown.
Is there any physiological or pathological relevance of the insulin-induced membrane insertion of GABA A receptors and the involvement of PRIP? It is reported that oxygen-glucose deprivation (OGD), an ischemia-like challenge, decreases the number of cell surface GABA A receptors and thereby leads to excitotoxic cell death in cultured hippocampal neurons. Insulin treatment counteracts the OGD-induced diminishment of the number of cell surface GABA A receptors and thus prevents ischemic cell death (14). Additionally, it is reported that insulin-induced cell surface expression of GABA A receptors leads to membrane hyperpolarization in islet ␣ cells, thereby suppressing glucagon secretion (12), suggesting its involvement in diabetic pathogenesis. Another example is that interleukin-1␤ (IL-1␤) increases in the cell surface expression of GABA A receptors depend on the PI 3-kinase-Akt signaling pathway (50). Patients with sepsis-associated encephalopathy (SAE), a neurological complication in sepsis, have higher plasma levels of IL-1␤, therefore this may contribute to the cognitive dysfunction observed in SAE by altering GABAergic synaptic strength (50). It is possible that PRIP is implicated in such neuronal dysfunction and pathogenesis through the recruitment of active Akt to GABA A receptors, suggesting that PRIP could a therapeutic target.
In conclusion, we showed here that PRIP is implicated in the insulin-induced membrane insertion of GABA A receptors as it recruits active Akt to the vicinity of GABA A receptors. The subsequent complex formation may serve as the molecular basis for the efficient phosphorylation of GABA A receptors through Akt and receptor insertion into the cell surface membrane. Therefore, PRIP is a key factor in the control of the plasticity of GABAergic transmission.