Ubiquitin-dependent lysosomal targeting of GABA_A receptors regulates neuronal inhibition

Abstract

The strength of synaptic inhibition depends partly on the number of GABA_A receptors (GABA_ARs) found at synaptic sites. The trafficking of GABA_ARs within the endocytic pathway is a key determinant of surface GABA_AR number and is altered in neuropathologies, such as cerebral ischemia. However, the molecular mechanisms and signaling pathways that regulate this trafficking are poorly understood. Here, we report the subunit specific lysosomal targeting of synaptic GABA_ARs. We demonstrate that the targeting of synaptic GABA_ARs into the degradation pathway is facilitated by ubiquitination of a motif within the intracellular domain of the γ2 subunit. Blockade of lysosomal activity or disruption of the trafficking of ubiquitinated cargo to lysosomes specifically increases the efficacy of synaptic inhibition without altering excitatory currents. Moreover, mutation of the ubiquitination site within the γ2 subunit retards the lysosomal targeting of GABA_ARs and is sufficient to block the loss of synaptic GABA_ARs after anoxic insult. Together, our results establish a previously unknown mechanism for influencing inhibitory transmission under normal and pathological conditions.

The strength of inhibitory synapses is a major determinant of neuronal processing and excitability and can be determined by the number of γ-amino butyric acid type A receptors (GABA_ARs) found at synaptic sites (1–3). Moreover, in a number of neuropathologies, including status epilepticus and in vitro and in vivo models of cerebral ischemia, the rapid downmodulation of synaptic GABA_ARs has been reported to contribute to pathological disinhibition, excitotoxicity, and cell death (4–8). However, the molecular mechanisms and signaling pathways that regulate GABA_AR trafficking, under normal or pathological conditions, remain unclear.

GABA_ARs undergo significant rates of endocytosis, a process that can rapidly modify receptor number at synapses (9). Upon internalization, receptors are either reinserted into the plasma membrane or targeted for degradation (10). The balance between these two processes is critical for determining the number of receptors expressed at synapses and hence neuronal excitability (11). However, the regulatory mechanisms that control this endocytic sorting of synaptic GABA_ARs remain unknown.

The trafficking of membrane proteins can be regulated via the covalent attachment of the 76 amino acid (aa) peptide, ubiquitin, or ubiquitin-like proteins to lysine residues within target proteins (12, 13). A number of studies on the role of ubiquitination in regulating synaptic function under normal conditions and in pathological conditions such as ischemia have focused on the ubiquitin-proteasome system (14–18). However, within the endocytic pathway, mono-ubiquitination (or the addition of short ubiquitin chains) is a key signal that results in the specific targeting of cargo to late endosomes for subsequent lysosomal degradation (12, 19, 20). The significance of ubiquitination in determining the endocytic fate and degradation of neurotransmitter receptors remains to be established.

In the present study we investigated the role of GABA_AR ubiquitination in regulating GABA_AR endosomal trafficking. We show that the GABA_AR γ2 subunit, which plays a key role in synaptic targeting, can preferentially target internalized GABA_ARs to the lysosome for degradation. We demonstrate that this process is regulated by an amino acid motif within the intracellular loop of the γ2 subunit, which is a target for ubiquitination and that mutation of this motif retards the lysosomal targeting of GABA_ARs. Inhibition of lysosomal activity or the trafficking of ubiquitinated proteins to lysosomal compartments increases the accumulation of GABA_ARs at synapses and the efficacy of synaptic inhibition. Furthermore, we demonstrate that the loss of surface GABA_ARs observed in an in vitro model of ischemia can be inhibited by blocking ubiquitination of the lysosomal targeting motif in the γ2 subunit. Thus, our results demonstrate that GABA_AR ubiquitination is a key mechanism for regulating the strength and plasticity of synapses and the efficacy of neuronal inhibition under both normal and pathological conditions.

Results

Lysosomal Activity Regulates the Accumulation of Synaptic GABA_ARs and the Inhibitory Synaptic Response.

We initially investigated whether the extent of lysosomal GABA_AR degradation is a functionally relevant mechanism for regulating the size and number of postsynaptic GABA_ARs and hence the strength of synaptic inhibition. For this, acute coronal cortical slices from P21 rats were treated with the lysosomal protease inhibitor leupeptin (200 μM, 3 h). Slices were fixed and stained with antibodies against GABA_ARs and the inhibitory presynaptic marker VIAAT (21) and imaged using confocal microscopy (Fig. 1 A and B). Leupeptin treatment resulted in an increase in the mean postsynaptic GABA_AR cluster size (Ctrl: 0.45 ± 0.04 μm²; Leu: 0.63 ± 0.02 μm²) revealed by a rightward shift in the cumulative cluster area distributions (Fig. 1 C and D). In addition, we found a significant increase in the number of synaptic GABAergic clusters per cell (Ctrl: 15.2 ± 1.2; Leu: 27.3 ± 2.4) (Fig. 1E). In contrast, clusters that were not found to have a presynaptic counterpart were unaltered in size (Ctrl: 0.40 ± 0.05 μm²; Leu: 0.45 ± 0.04 μm²) or number (Ctrl: 3.8 ± 0.7; Leu: 3.8 ± 0.9).

Fig. 1.

Functional effects of disrupting lysosomal degradation (A–E) GABA_AR cluster analysis on control and leupeptin treated cortical slices. High resolution confocal images were taken of 25-μm sections across the cells bodies (identified by DAPI nuclear staining, blue). (A and B) Pseudocolor images of GABA_AR clusters. Clusters can be identified by changes in intensity (blue to pink to white) over a central point. (A' and B') Merged images show GABA_AR clusters (green) apposed to presynaptic VIAAT positive terminals (red). (C) Cumulative plot showing the cluster size distributions under control and leupeptin treated conditions. (D) Average GABA_AR cluster size in control and leupeptin treated cells. (E) Average number of GABA_AR clusters in control and leupeptin treated cells (n = 11–12 cells; ***, P < 0.001). (F) Summary data plot of the effect of leupeptin on eIPSC density (n = 12, P < 0.001 unpaired t test). (G) Representative cumulative plots and traces (Inset) of mIPSC amplitude for representative cells from cortical slices under control and leupeptin treated conditions. (H) Bar plot summary showing the effect of leupeptin on mIPSC amplitude (n = 7, ***, P < 0.001). (I) Bar plot summary showing the effect of leupeptin on mIPSC frequency (n = 7, ***, P < 0.001). (J) Representative cumulative plots of mIPSC inter-event interval of representative cells under control and leupeptin treated conditions.

To assess whether this increase in the GABA_AR cluster size and number resulted in modified synaptic inhibition, we compared the properties of miniature or evoked inhibitory postsynaptic currents (mIPSCs or eIPSCs) between control neurons and those treated with leupeptin using whole cell recording. In agreement with our labeling data, leupeptin treatment resulted in a significant increase in the eIPSC current amplitude (Ctrl: 223.6 ± 16 pA; Leu: 412.5 ± 26 pA) (Fig. 1F). In addition, leupeptin treatment caused a sustained increase in mIPSC amplitude (41.6 ± 5.1pA) compared with untreated control cells (25.3 ± 2.2 pA) (Fig. 1 G and H) revealed by the rightward shift in the cumulative amplitude relationship for mIPSCs. Leupeptin treatment also caused a significant enhancement in mIPSC frequency (Ctrl: 4.6 ± 0.3 Hz; Leu: 19.6 ± 1.2 Hz) (Fig. 1 I and J). However, leupeptin did not alter mIPSC kinetics (Table S1). In contrast, the mean amplitudes (Ctrl: 7.5 ± 0.6 pA; Leu: 8.9 ± 0.6 pA) and frequency (Ctrl: 2.5 ± 0.19 Hz; Leu: 2.7 ± 0.18 Hz) of miniature excitatory postsynaptic currents (mEPSC) or the eEPSC current amplitude (Ctrl: 108.8 ± 7.7 pA; Leu: 118 ± 8.9 pA) were unaffected by leupeptin treatment (Fig. S1). Thus, blockade of lysosomal GABA_AR degradation results in an increase in the number and size of synaptic GABA_AR clusters leading to increased synaptic inhibition.

Identification of a Sorting Motif in the GABA_AR γ2 Subunit That Facilitates Lysosomal Targeting of GABA_ARs.

Having established that lysosomal degradation of GABA_ARs is an important determinant of the number of surface and synaptic GABA_ARs, we next set out to identify the molecular mechanisms that mediate lysosomal targeting of GABA_ARs. Because the γ2 subunit is an important determinant for the trafficking of GABA_ARs, and their entry into the endocytic pathway (22–25), we examined the possible role that this subunit plays in receptor lysosomal targeting. We compared the lysosomal targeting in HEK-293 cells of GABA_ARs composed of αβ and αβγ2 subunits modified with N-terminal 9E10 reporters. To identify late endocytic/lysosomal compartments we coexpressed GABA_ARs with GFP-Rab7, a marker for late endosomal and lysosomal structures (26, 27). To specifically label receptors in the endocytic pathway, live cells were labeled with myc antibody for 30 min at 4 °C and then placed at 37 °C to allow for internalization. After fixation and permeabilization, internalized receptors were visualized and their distribution in confocal slices showing maximum receptor internalization was compared with that for GFP-Rab7. Endocytosed αβγ2 receptors exhibited a marked increase in targeting to late endosomes/lysosomes as demonstrated by a substantially increased colocalization with GFP-Rab7 compared with those composed of αβ subunits (αβ: 43.3 ± 4.1%; αβγ2: 67.8 ± 5.2%) (Fig. 2 A–C). This suggests a significant role for the γ2 subunit in regulating the targeting of GABA_ARs to the late endosomal/lysosomal compartments.

Fig. 2.

γ2 subunit dependent lysosomal targeting of GABA_ARs. (A and B) Representative images of internalized GABA_ARs (red) comprised of α1β3 (A) or α1β3γ2 (B) in HEK293 cells coexpressed with the late endosomal/lysosomal marker GFP-Rab7 (green). Areas of overlap are seen in yellow. (C) Summary data of the colocalization levels between internalized GABA_ARs and GFP-Rab7 (n = 14–16, ***, P < 0.001). (D–F) Representative images of internalized β3 subunits or β3-γ2 chimeras C1–C2 (red) in HEK293 cells coexpressed with the late endosomal/lysosomal marker GFP-Rab7 (green). Areas of overlap are seen in yellow. (G) Schematic diagram of β3-γ2 chimeras C1–C5. (H) Summary data of the percent colocalization levels of internalized C1–C5 and GFP-Rab7 (n = 8–12, **, P < 0.01).

To identify the domain of the γ2 subunit that mediates the lysosomal targeting of GABA_ARs we exploited the ability of GABA_AR β3 subunits to assemble into pentameric cell surface homomeric ion channels, a property that depends on residues within the N terminus of this subunit (28, 29). As measured by antibody labeling, internalized β3 subunits showed a similar distribution to αβ receptors with respect to GFP-Rab7 (33.8 ± 3.7%) (Fig. 2 D and H). In contrast, a chimeric receptor (C1) in which the major intracellular domain and transmembrane domain 4 of the β3 subunit were exchanged for the corresponding domains of the γ2 subunit (Fig. 2G) exhibited enhanced targeting to lysosomes, as measured by colocalization with GFP-Rab7 similar to levels seen with αβγ2 subunit containing receptors (63.3 ± 3.6%) (Fig. 2 E and H). Thus, the intracellular domain and/or TM4 of the γ2 subunit are sufficient to mediate the lysosomal targeting of GABA_ARs.

To identify the region within the γ2 subunit responsible for the late endosomal/lysosomal targeting of GABA_ARs, we constructed a set of β3/γ2 subunit chimeras that serially replaced sequences within C1 with those from the β3 subunit (Fig. 2G). Using live cell antibody labeling experiments we followed the extent of lysosomal targeting of internalized β3/γ2 chimeras by their colocalization with GFP-Rab7. Compared with C1, chimeras C2, C3, C4, and C5 all showed reduced levels of colocalization with GFP-Rab7 (71.7 ± 3.4% compared with 31.8 ± 2.5%, 37.7 ± 6.4%, 38.6 ± 4.4%, 25.9 ± 4.3%, respectively) (Fig. 2 E, F, and H, and Fig. S2). Thus, residues 318–337 in the γ2-subunit intracellular domain, proximal to TM3 contain the late endosomal/lysosomal targeting sequence because replacement of this sequence with that of β3 abrogated the enhanced colocalization of the β3/γ2 chimeras with GFP-Rab7.

Blockade of Ubiquitin-Dependent Lysosomal Sorting Increases GABA_AR Functional Expression.

Ubiquitination of lysine residues can act as a signal for sorting membrane proteins to the lysosome for degradation (12). We noted that residues 318–337 of the γ2 subunit contain a number of conserved lysine residues (Fig. S3), suggesting that the late endosomal/lysosomal targeting of γ2 subunit containing GABA_ARs may be dependent on ubiquitination. The ubiquitin dependent sorting of membrane proteins to the lysosome depends on a highly conserved machinery that specifically recognizes ubiquitinated membrane proteins in the early endosome and selects them for late endosomal/lysosomal sorting (30). The hepatocyte growth factor regulated substrate (Hrs) is a key component of this initial recognition machinery and we took advantage of GFP-2FYVE, a dominant negative construct known to block the lysosomal targeting of ubiquitinated proteins by disrupting Hrs function (31). We first assessed the ability of this construct to modulate the trafficking of GABA_ARs by assessing its effect on steady state surface levels of these receptors in transfected neurons. As measured by cell surface biotinylation, neurons transfected with GFP-2FYVE exhibited a 68.2 ± 11.1% increase in the cell surface expression levels of GABA_ARs compared with those expressing GFP alone (Fig. 3 A and B).

Fig. 3.

Functional effects of expressing GFP-2FYVE. (A) Cell surface biotinylations of cortical cells expressing GFP or GFP-2FYVE. (B) Bar plot summary showing GABA_AR surface levels in cells expressing GFP or GFP-2FYVE. Levels are represented as percentage-normalized to those observed in GFP control (n = 11 independent experiments, ***, P < 0.001 paired t test). (C) Representative cumulative plots and traces (Inset) of mIPSC amplitude of representative cells expressing GFP (open circles) or GFP-2FYVE (closed circles). (D) Bar plot summary showing the effect of GFP-2FYVE expression on mIPSC amplitude (n = 8–9, ***, P < 0.001). (E) Representative cumulative plots of mIPSC inter-event interval of representative cells expressing GFP (open circles) or GFP-2FYVE (closed circles). (F) Bar plot summary showing the effect of GFP-2FYVE expression on mIPSC frequency (n = 8–9, ***, P < 0.001).

We further investigated whether the increase in cell surface accumulation of GABA_ARs in the presence of GFP-2FYVE alters the properties of mIPSCs using whole cell recording. Similar to our results using leupeptin, expression of GFP-2FYVE in cultured hippocampal neurons resulted in a 66.3% increase in the mean mIPSC amplitude compared with cells expressing GFP (GFP-2FYVE: 50.4 ± 2.4 pA, n = 9; GFP: 30.3 ± 1.2 pA, n = 8) (Fig. 3 C and D). In addition, expression of GFP-2FYVE also caused a significant enhancement in mIPSC frequency compared with control cultures expressing GFP alone (GFP-2FYVE: 7.9 ± 0.5 Hz, n = 9; GFP: 3.2 ± 0.6 Hz, n = 8) (Fig. 3 C, E, and F). As for leupeptin treatment, channel kinetics were unaffected by GFP-2FYVE expression (Table S2). These results demonstrate that ubiquitin-dependent targeting of GABA_ARs to lysosomal compartments plays a key role in regulating the number of GABA_ARs at synaptic sites together with the efficacy of neuronal inhibition.

Lysosomal Targeting of GABA_ARs Is Regulated by Ubiquitination of the γ2 Subunit.

Our results indicate that ubiquitination of GABA_ARs within the endocytic pathway can determine the lysosomal degradation of GABA_ARs and our mapping experiments suggested that the lysine residues within the γ2 subunit lysosomal targeting motif are likely to be the site of ubiquitination. To analyze this further, we compared the ubiquitination of GABA_ARs containing WT γ2 subunits or a construct in which K325, K328, K330, K332, K333, K334, and K335 within the lysosomal targeting motif were all mutated to arginine residues (γ2^K7R). The ubiquitination of these constructs was tested in HEK cells expressing α1β3^9E10γ2 (or α1β3^9E10γ2^K7R) and HA-tagged ubiquitin. Cells were lysed, and the γ2 subunit was immunoprecipitated with anti-myc and immunoblotted with anti-HA and anti-9E10. This revealed significant levels of ubiquitination for the WT γ2 subunit, which was largely abolished for the γ2^K7R construct despite similar levels of γ2 subunit immunoreactivity. Therefore, these experiments suggest that the principle sites of ubiquitination within the γ2 subunit intracellular domain lie between residues 325 and 335 (Fig. 4D).

Fig. 4.

GABA_AR ubiquitination is required for late endosomal/lysosomal targeting. (A and B) Representative images of internalized α1β3γ2 (A) or α1β3γ2^K7R (B) in HEK293 cells coexpressing the late endosomal/lysosomal marker GFP-Rab7. (C) Summary data of the colocalization levels of internalized GABA_ARs and GFP-Rab7. (D) The γ2^K7R subunit mutant shows decreased levels of ubiquitination compared with the wild-type γ2 GABA_AR subunit. (E and F) Typical traces of evoked GABAergic currents in HEK-293 cells expressing receptors composed of α1β3γ2 or α1β3γ2^K7R clamped at −70 mV and treated with 20 μM Leupeptin. (G) The magnitude of I_GABA measured at 5-min intervals was then normalized to that evident at time 0 (5 min after break-in) for cells expressing α1β3γ2 or α1β3γ2^K7R receptors under control conditions and leupeptin treatment (n = 7–11 cells).

We then tested the role of the respective lysine residues in facilitating sorting of αβγ2 receptors to the late endosome/lysosome. Live antibody labeling experiments showed that mutation of the lysine residues within the endocytic sorting motif did not alter receptor internalization (Fig. 4B). In contrast, receptors expressing γ2^K7R exhibited reduced levels of late endosomal/lysosomal targeting compared with those expressing wild-type γ2, as assessed by colocalization with GFP-Rab7 fluorescence (αβγ2: 67.7 ± 3.8%; αβγ2^K7R: 36.3 ± 3.5%) (Fig. 4 A–C).

In addition to comparing the lysosomal targeting of GABA_ARs we examined the effects of blocking lysosomal degradation on their functional expression using patch clamp recordings from transfected HEK-293 cells (17, 32). To modulate lysosomal degradation we introduced leupeptin into expressing cells via intracellular dialysis with the patch-pipette. The magnitude of GABA activated membrane currents (I_GABA) induced by 5 μΜ GΑΒΑ (ΕC₅₀) (17, 32) were analyzed over a time course of 30 min and normalized to that evident at time 0. In cells expressing receptors incorporating γ2 subunits leupeptin induced a significant increase in I_GABA, which at 30 min reached 207.8 ± 10.6% of that seen at zero time (Fig. 4 E and G). In contrast in cells expressing γ2^K7R potentiation of I_GABA by leupeptin was insignificant (111.3 ± 8.9%) (Fig. 4 F and G). In agreement with our imaging studies, leupeptin did not significantly alter I_GABA for cells expressing receptors composed of αβ subunits (107.0 ± 10.4%) (Fig. S4). Therefore, our observations suggest that multiple lysine residues between amino acids 318–337 within the intracellular domain of the γ2 subunit play a critical role in regulating the late endosomal/lysosomal trafficking of GABA_ARs.

Enhanced Lysosomal Targeting Underlies the Deficits of GABA_AR Cell Surface Expression After Anoxia.

Protein ubiquitination is increased in the brain upon ischemic injury (33). In addition, several studies have reported compromised synaptic inhibition and a reduction in cell surface GABA_AR expression during ischemia (4, 7, 8). Although the cellular mechanisms responsible for this pathological loss of GABA_AR functional expression remain unknown, one possible explanation is altered receptor trafficking (4). Therefore, we assessed whether ubiquitin dependent trafficking of GABA_ARs plays a role in the deficit of GABA_AR surface expression after ischemia. To examine this, we imaged cultured hippocampal neurons transfected with either ^GFPγ2 or ^GFPγ2^K7R under control conditions or during oxygen and glucose deprivation (OGD) to induce anoxia. We first investigated whether ^GFPγ2 and ^GFPγ2^K7R were equally targeted to synaptic sites. We found 75.96 ± 4.5% of GABA_ARs clusters expressing ^GFPγ2 and 79.07 ± 2.8% of GABA_ARs expressing ^GFPγ2^K7R were found apposed to VIAAT labeled presynaptic inhibitory terminals (Fig. S5). Having established that both ^GFPγ2 and ^GFPγ2^K7R constructs were equally targeted to synaptic sites, we imaged the fluorescence of individual GFP clusters over time in 12–16DIV cultured hippocampal cells under control and OGD conditions. Analysis of ^GFPγ2 subunit containing clusters revealed a significant loss in cluster fluorescence over time under OGD conditions compared with control (OGD t = 30 min: 31.1 ± 2.8%; control t = 30 min: 6.8 ± 5.7%; P < 0.001, n = 7–8) (Fig. 5A, C, and D, and Fig. S6). In contrast, cells expressing ^GFPγ2^K7R did not exhibit a significant loss of cluster fluorescence during OGD treatment (OGD t = 30 min: 6.43 ± 6.6%; control t = 30 min: 13.29 ± 3.0%; P > 0.05, n = 7–8) (Fig. 5 B–D and Fig. S6). Thus, ubiquitin-dependent lysosomal targeting of GABA_ARs results in the loss of synaptic GABA_ARs during OGD treatment. In accordance with these results, the reduction in GABA_AR cluster fluorescence observed during OGD was also blocked by leupeptin treatment (Fig. S7). These results provide evidence that modified ubiquitin dependent trafficking of GABA_ARs in the endocytic pathway may underlie deficits in receptor cell surface stability after ischemia.

Fig. 5.

Oxygen glucose deprivation results in a ubiquitination dependent decrease in GABA_AR surface levels. (A and B) 25-μm dendritic segment of 12–16 DIV hippocampal cells expressing GFP-tagged γ2 (A) and γ2^K7R (B) imaged over time under OGD conditions. Arrowheads point to GFP clusters. (Scale bar: 5 μm.) (C) GFP cluster fluorescence was monitored over time and normalized to t = −10 min. (D) Summary data plot of normalized cluster fluorescence (F/F₋₁₀) at t = 30 min for cells expressing ^GFPγ2 and ^GFPγ2^K7R (n = 7–8, ***, P < 0.001).

Discussion

The regulated endocytosis of ligand-gated ion channels is an important determinant for the efficacy of synaptic transmission (34, 35). Within the endocytic pathway these proteins are recycled or targeted for lysosomal degradation and the impact of this endocytic sorting decision is emerging as an important determinant for neuronal excitability (11, 36). Here, we investigated the role of regulated endocytic sorting with reference to GABA_ARs, the principal sites of fast synaptic inhibition in the brain. We demonstrate that acute inhibition of lysosomal activity increases inhibitory synaptic strength by blocking synaptic GABA_AR degradation. In addition, we find that ubiquitination of a motif within the γ2 subunit is responsible for GABA_AR lysosomal targeting. Furthermore, our results suggest that enhanced ubiquitination dependent degradation of GABA_ARs may directly contribute to the previously reported reduction in the number of synaptic GABA_ARs observed in ischemia.

The GABA_AR γ2 subunit confers important pharmacological, functional, and membrane trafficking properties to GABA_ARs including the selective targeting of GABA_ARs to inhibitory postsynaptic domains (22). In this study, we compared the endocytic fate of GABA_ARs composed of αβ and αβγ2 subunits and found that the γ2 subunit also plays a key role in the endocytic sorting of GABA_ARs, by facilitating the trafficking of GABA_ARs to late endocytic/lysosomal compartments. This may provide a mechanism by which neurons can specifically regulate the endocytic trafficking of synaptic (γ2 subunit containing) GABA_ARs without altering the trafficking of other GABA_ARs (lacking γ2 subunits), which may contribute to tonic inhibition.

Using β3/γ2 subunit chimeras we found that a motif within the γ2 subunit was critically important for the enhanced lysosomal targeting of GABA_ARs. This motif was found to be a direct target for ubiquitination and mutation of seven lysine residues within this motif resulted in reduced levels of GABA_AR ubiquitination and compromised targeting to lysosomes. Therefore, these experiments provide compelling evidence that the lysosomal targeting of GABA_ARs depends on the ubiquitination of lysine residues between residues 317–338 of the γ2 subunit. Ubiquitination of lysine residues within the intracellular loops of the GABA_AR α1 and β3 subunits can target unassembled subunits within the endoplasmic reticulum (ER) for ER associated degradation by the proteosome, but does not appear to be involved in GABA_AR endocytic trafficking (17, 18). Moreover, this proteosomal degradation of GABA_AR subunits within the secretory pathway regulates synaptic inhibition over long time scales (days) (17). In contrast, results in this study reveal a mechanism for more rapid (tens of minutes), γ2 subunit-specific, ubiquitination of GABA_ARs resulting in fast regulation of synaptic inhibition.

We found that blocking lysosomal degradation of GABA_ARs (with leupeptin treatment) or their ubiquitin-dependent lysosomal targeting (using the dominant negative GFP-2FYVE), caused a marked increase in both mIPSC amplitude and frequency suggesting not only an increase in the number of synaptic GABA_ARs but also an increase in the number of synaptic responses. In agreement with these findings, correlative immunofluorescence experiments revealed that blocking lysosomal GABA_AR degradation resulted in an increase both in GABA_AR cluster size and in the number of GABA_AR clusters apposed to VIAAT positive inhibitory presynaptic terminals. The mIPSC amplitude and frequency increase is similar to that observed when surface GABA_AR numbers are increased by dialysing inhibitors of GABA_AR endocytosis into the postsynaptic neuron via the electrophysiological recording pipette (9). We previously suggested that this mIPSC frequency increase reflects the increase in surface GABA_AR number and mIPSC amplitude unmasking mIPSCs that were previously below the threshold for detection in the recordings rather than because of a presynaptic change (9). The changes in GABA_AR cluster number and mIPSC frequency that we report here are similarly likely to be because of the improved detection of existing established synapses (which before blockade of lysosomal targeting contained too few GABA_ARs to be detected) as a result of an increase in the availability of surface GABA_ARs rather than because of presynaptic effects such as increases in the probability of release.

Modified membrane trafficking of GABA_ARs is strongly implicated in a reduction in the number of surface and synaptic GABA_ARs, which contributes to the loss of inhibition in status epilepticus (6, 37–39) and ischemic brain injury (4, 8), but the mechanisms have remained unclear. Here, we demonstrate that although anoxic insult reduces the number of wild type GFP-tagged synaptic GABA_ARs, the synaptic expression of GABA_ARs incorporating ^GFPγ2^K7R subunits (and hence no longer ubiquitinated) remained unaffected. Thus, our results suggest that ubiquitin-dependent degradation of GABA_ARs may be a key mechanism that underlies the downmodulation of synaptic inhibition after a pathological insult.

In summary, the lysosomal targeting of GABA_ARs is modulated by the ubiquitination of lysine residues between amino acids 317–328 within the intracellular domain of the γ2 subunit. This process regulates the accumulation of GABA_ARs at inhibitory synapses and the efficacy of neuronal inhibition under basal conditions, and is responsible for deficits in neuronal inhibition after anoxia.

Materials and Methods

Antibodies and Constructs.

For details see SI Materials and Methods.

Cell Culture and Transfections.

Rat hippocampal and cortical neurons were prepared from embryonic day 18 rat brains. All cell types were transfected by nucleofection as described in ref. 10. See SI Materials and Methods.

Brain Slice Preparation and Leupeptin Treatment.

Two hundred micrometers frontal cortex coronal sections were prepared from P21 rats and incubated for 3 h in ACSF (126 mM NaCl, 24 mM NaHCO₃, 1 mM NaH₂PO₄, 2.5 mM KCl, 1 mM MgCl₂, 2 mM CaCl₂, 10 mM glucose, and gassed with 95% O₂/5% CO₂) or ACSF supplemented with 200 μM Leupeptin. After treatment, slices were immediately fixed and stained based on protocols from (40, 41). See SI Materials and Methods.

Immunofluorescence and Confocal Microscopy.

For receptor internalization assays receptors were labeled live with 9E10 anti-myc antibody at 4 °C. Labeled receptors were allowed to internalize for 1 h at 37 °C. Surface antibody was stripped away using 0.2 M acetic acid, 0.5 M NaCl before fixation as described in ref. 34. Internalized receptors were identified using Alexa-fluor594 conjugated goat anti-mouse antibody and confocal microscopy. See SI Materials and Methods.

Image Processing and Quantitation.

Captured confocal images were analyzed using Metamorph software (Universal Imaging Corporation). See SI Materials and Methods.

Live Cell Imaging and Oxygen/Glucose Deprivation.

Transfected 12–15DIV hippocampal cultures were imaged in ACSF under constant perfusion (at a rate of 1.3 mL/min). For OGD treatment ACSF was replaced with OGD buffer (126 mM NaCl, 24 mM NaHCO₃, 1 mM NaH₂PO₄, 2.5 mM KCl, 1 mM MgCl₂, 2 mM CaCl₂, 7 mM sucrose, and gassed with 95% N₂/5% CO₂) (42). Images were captured every 10 min for a total of 30 min and analyzed using ImageJ software. See SI Materials and Methods.

Biotinylation and Ubiquitination Assays.

Cell surface biotinylations were carried out in triplicate on 12–16 DIV cortical cells, as previously described. For ubiquitination assays, Myc-tagged GABA_ARs were immunoprecipitated from transfected HEK cells, as previously described. Precipitates were resolved by SDS/PAGE electrophoresis and analyzed by Western blotting. See SI Materials and Methods.

Electrophysiology.

Whole-cell recordings of mIPSCs from cortical slices, or GFP-2FYVE transfected cultured cortical cells, were performed using standard voltage clamp techniques in the presence of CNQX (20 μM) and APV (40 μM) to block AMPA and NMDA receptors, respectively (43, 44). Membrane currents from HEK-293 cells were measured, as outlined previously at 32 °C (10, 43, 45). Data were collected 5 min after obtaining a stable recording conformation. See SI Materials and Methods.