Ethanol promotes clathrin adaptor-mediated endocytosis via the intracellular domain of δ-containing GABAA receptors

Abstract

Pharmacological and genetic evidence reveals that GABA(A) receptor (GABA(A)-R) expression and localization are modulated in response to acute and chronic ethanol (EtOH) exposure. To determine molecular mechanisms of GABA(A)-R plasticity in response to in vivo acute EtOH, we measured early time changes in GABA(A)-R subunit localization. Single doses of EtOH (3 g/kg via i.p. injection in rats) produced decreases in surface levels of GABA(A)-R α4 and δ subunits at 5-15 min post-EtOH in hippocampus CA1 and dentate gyrus, verifying our earlier report (Liang et al., 2007). Here we also examined the β3 subunit and its phosphorylation state during internalization. β3 also was internalized during 5-15 min after EtOH exposure, while phosphorylation of β3 was increased, then decreased at later times, ruling out β3 dephosphorylation-dependent endocytosis. As early as 5 min post-EtOH, there is an initial increase in association between the δ subunits with clathrin adaptor proteins AP2-μ2 revealed by coimmunoprecipitation, followed by a decrease in association 15 min post-EtOH. In vitro studies using glutathione S-transferase fused to the δ subunit intracellular domain (ICD) show that two regions, one containing a classical YxxΦ motif and the other an atypical R/K-rich motif, directly and differentially bind to AP2-μ2, with the former YRSV exhibiting higher affinity. Mutating both regions in the δ-ICD abolishes μ2 binding, providing a possible mechanism that can explain the rapid downregulation of extrasynaptic α4βδ-GABA(A)-R following in vivo EtOH administration, in which the δ-ICD increases in affinity for clathrin AP2-μ2 leading to endocytosis.

Figure 1.

A single in vivo EtOH dose causes decreases in extrasynaptic surface GABA_A-R in hippocampal DG and CA1 within 15 min. A–D, Crosslinking experiments show a trending decrease in surface levels of α4, pβ3, and δ subunits as early as 5 min following EtOH dose (A, B), with significant decreases measured 15 min post-EtOH (C, D). t, Total protein (untreated slices); int, internal protein (slices treated with BS3). β3 and γ2 showed no significant changes. Pictures show representative blots of the average changes; graphs (B, D) plot the average of several experiments, N = 5. E, F, GABA_A-R subunit levels at early time points after EtOH treatment. Total GABA_A-R subunit changes in DG and CA1 at times 5–60 min following EtOH (3 g/kg) treatment do not change significantly (N = 4–5).

Figure 2.

EtOH-induced changes in GABA_A-R β3 phosphorylation at early time points. A, Phosphorylation of β3 as identified with sequence-specific phospho-peptide antibody. Initial increases at 5 and 10 min in phosphorylation of β3 subunit compared with control treated at 5 and 10 min (153.10 ± 40.14%, 138.08 ± 29.20%, respectively; N = 4–5). B, Plot of quantified pβ3 level changes following various EtOH time treatments and brain regions. *p < 0.05; **p < 0.001 between 5 min treatment and stated treatment times in each brain region; ***p < 0.05 between control and EtOH-treated rats from same time point treatment; two-way ANOVA.

Figure 3.

GABA_A-R δ and α4 subunit association with μ2 changes following EtOH exposures. A, B, EtOH caused a 32.7% increase in δ association with μ2 at 5 min, and a 55.1% decrease in δ association with μ2 15 min following EtOH treatment. The EtOH-induced changes in association of δ with μ2 are statistically significant from each other (**p < 0.0001), and statistically significant from control treated δ-μ2 associations (two-way ANOVA, Bonferroni post hoc test; *p < 0.05; N = 5). C, D, GABA_A-R α4 subunit associations following EtOH exposure. EtOH caused a 38% increase in α4 association with μ2 at 5 min (not significant, p = 0.05, N = 4; but was significantly different from association of α4 and μ2 at 15 min, p < 0.001). This suggests that, at least at an early time point, these subunits are internalizing with δ-containing GABA_A-R.

Figure 4.

Characterization of δ intracellular domain. A, Putative μ2 binding domains within the GABA_A-R δ subunit ICD. The entire rat δ ICD (aa 316–410) is shown. A classical YxxΦ motif (³⁷⁵YRSV) is located at residues 375–378, while an atypical basic rich motif similar to those seen in GABA_A-R β(1–3) is found at residues 322–334. Both are highlighted. B, Diagram of GST δ ICD constructs in this study: 1, GST only; 2, GST+ δICD full-length residues 316–410 (δ); 3, GST + first half of δ ICD residues 316–359 (δR); 4, GST + second half δ ICD residues 360–410 (δY); 5, GST + δ full-length R/K piece mutated residues 322–326 (δ³²²RKKRK→GAAGA); 6, GST + δ full-length Y375 mutated to A (δ³⁷⁵Y→A); 7, GST + δ full-length, where both the ³²²R/K piece and ³⁷⁵Y are mutated (δ2X mutant). C, Each μ2 binding domain of the δ ICD binds to μ2. GST-only (control) does not bind to μ2, but constructs 2–6 all bind to μ2. This includes all δ ICD constructs containing at least one complete μ2-binding domain (³²²R/K region only shown as δR; ³⁷⁵YRSV-only shown as δY). Only the construct lacking both μ2-binding domains shows no binding to μ2. Therefore, each of the two μ2 binding domains directly bind to μ2. Quantification of binding shows there are no significant differences between mutants and δ ICD full-length, except for the double mutant, which significantly lost all μ2 binding (*p < 0.001; one-way ANOVA). All other constructs are significantly different from the double mutant (p < 0.05; one-way ANOVA). D, The ³⁷⁵YRSV-motif of the δ ICD detects lower [μ2]. The 1 nm, 0.1 μm, and 0.1 mm AP2 fractions were incubated with the GST-δ truncation proteins to see if one region binds at a lower μ2 concentration than the other. The procedure and analysis were similar to those done with the pull-down experiments, with representative blot and cumulative quantified μ2 binding plot shown here. The 1 nm concentration showed no binding, but 0.1 μm showed μ2 binding more to the construct lacking the basic rich region, suggesting this lower-concentration binding is more influenced by the ³⁷⁵YRSV-motif. The difference between binding seemed to be undetectable at higher μ2 concentrations, likely because the basic-rich ³²²R/K motif is now participating in its own μ2 binding (p < 0.05; one-way ANOVA; N = 3).