Ca2+/Calmodulin-Dependent Protein Kinase II (CaMKII) β-Dependent Phosphorylation of GABAB1 Triggers Lysosomal Degradation of GABAB Receptors via Mind Bomb-2 (MIB2)-Mediated Lys-63-Linked Ubiquitination

The G protein-coupled GABAB receptors, constituted from GABAB1 and GABAB2 subunits, are important regulators of neuronal excitability by mediating long-lasting inhibition. One factor that determines receptor availability and thereby the strength of inhibition is regulated protein degradation. GABAB receptors are constitutively internalized from the plasma membrane and are either recycled to the cell surface or degraded in lysosomes. Lys-63-linked ubiquitination mediated by the E3 ligase Mind bomb-2 (MIB2) is the signal that sorts GABAB receptors to lysosomes. However, it is unknown how Lys-63-linked ubiquitination and thereby lysosomal degradation of the receptors is regulated. Here, we show that Ca2+/calmodulin-dependent protein kinase II (CaMKII) promotes MIB2-mediated Lys-63-linked ubiquitination of GABAB receptors. We found that inhibition of CaMKII in cultured rat cortical neurons increased cell surface GABAB receptors, whereas overexpression of CaMKIIβ, but not CaMKIIα, decreased receptor levels. This effect was conveyed by Lys-63-linked ubiquitination of GABAB1 at multiple sites mediated by the E3 ligase MIB2. Inactivation of the CaMKII phosphorylation site on GABAB1(Ser-867) strongly reduced Lys-63-linked ubiquitination of GABAB receptors and increased their cell surface expression, whereas the phosphomimetic mutant GABAB1(S867D) exhibited strongly increased Lys-63-linked ubiquitination and reduced cell surface expression. Finally, triggering lysosomal degradation of GABAB receptors by sustained activation of glutamate receptors, a condition occurring in brain ischemia, was accompanied with a massive increase of GABAB1(Ser-867) phosphorylation-dependent Lys-63-linked ubiquitination of GABAB receptors. These findings indicate that CaMKIIβ-dependent Lys-63-linked ubiquitination of GABAB1 at multiple sites controls sorting of GABAB receptors to lysosomes for degradation under physiological and pathological condition.

Introduction γ-Aminobutyric acid (GABA), the main inhibitory neurotransmitter in the brain, activates the heterodimeric G protein-coupled GABA B receptors, which are assembled from GABA B1 and GABA B2 subunits. GABA B receptors are abundantly expressed in pre-and postsynaptic compartments of inhibitory as well as excitatory neurons to regulate their excitability [1]. GABA B receptors are expressed by the majority of neurons and are involved in the regulation of virtually all important brain functions, such as neuronal network activity, synaptic plasticity, and neuronal development [2][3][4][5]. Accordingly, dysfunction of GABA B receptor signaling has been implicated in a variety of neurological disorders [6][7][8][9]. Khaled Zemoura and Karthik Balakrishnan contributed equally to this work.
For understanding the contribution of GABA B receptors to physiological and pathological mechanisms, the elucidation of its regulation is essential. A main factor regulating GABA B receptor signaling is the dynamic control of their cell surface expression. In this respect, protein degradation is one important mechanism that regulates receptor availability. The two main cellular protein degradation systems, which are proteasomes and lysosomes, contribute to the regulation of cell surface GABA B receptors in different cellular compartments. In the endoplasmic reticulum (ER), the amount of newly synthesized GABA B receptors that are trafficked to the cell surface is determined by proteasomal degradation via the ER-associated degradation (ERAD) machinery [10]. The activity state of the neuron controls the rate of Lys-48-linked ubiquitination of the GABA B2 subunit required for proteasomal receptor degradation [11]. In contrast, GABA B receptors internalized from the cell surface are degraded in lysosomes [12][13][14][15][16]. Sorting the receptors to lysosomes is mediated most likely via the endosomal sorting complex required for transport (ESCRT) machinery [16], which targets ubiquitinated membrane proteins to lysosomes. Accordingly, lysosomal degradation of GABA B receptors depends on Lys-63linked ubiquitination of the GABA B1 subunit at multiple sites mediated by the E3 ubiquitin ligase MIB2 [17]. However, the mechanism that regulates Lys-63-linked ubiquitination of GABA B1 , and thereby lysosomal degradation of the receptors, was unknown. Here, we show that phosphorylation of Ser-867 of GABA B1 by CaMKIIβ regulates the extent of MIB2-mediated K63-linked ubiquitination of GABA B1 and thereby the amount of lysosomal degradation.

Mutation of GABA B1a
Serine 867 in GABA B1a was mutated either to alanine or to aspartate by GenScript.

Culture and Transfection of Cortical Neurons
Primary cultures (co-cultures of neurons and glia cells) of the cerebral cortex were prepared from 18-day-old embryos of Wistar rats as described previously [25]. Neurons were used after 11 to 15 days in culture. Plasmid DNA was transfected into neurons by magnetofection using Lipofectamine 2000 (Invitrogen) and CombiMag (OZ Biosciences) exactly as specified in Buerli et al. [25]. In our hands, this method reliably yields 50-100 transfected neurons, when transfection was performed with 11-13-day-old cultures.

Culture and Transfection of HEK 293 Cells
HEK (human embryonic kidney) 293 cells were cultured in Dulbecco's modified Eagle's medium (DMEM, Gibco Life Technologies) containing 10% fetal bovine serum (Gibco Life Technologies) and penicillin/streptomycin (Gibco Life Technologies). Plasmids were introduced into HEK 293 cells using the polyethylenimine method according to the jet-PEI protocol (Polyplus Transfection).

Immunocytochemistry and Confocal Laser Scanning Microscopy
Immunofluorescence staining was done as described previously [13,26]. For detection of total GABA B receptors, neurons were fixed for 15-20 min at room temperature with 4% paraformaldehyde followed by permeabilization with 0.2% Triton X-100 and immunostaining. For analysis of cell surface GABA B receptors, living neurons were incubated with antibodies recognizing the extracellularly located N-terminal domain of GABA B1 or GABA B2 for 1 h at 4°C. After washing, the neurons were incubated with fluorophore-labeled secondary antibodies for 1 h at 4°C and subsequently fixed with 4% paraformaldehyde.
Images of stained neurons were recorded by laser scanning confocal microscopy (LSM 510 Meta, LSM 700 or LSM 710; Zeiss). Five to eight optical sections spaced by 0.3 μm were taken with a ×40, ×63, or ×100 plan-fluar oil differential interference contrast objective (Zeiss) at a resolution of 1024 × 1024 pixels. Total and cell surface fluorescence signals were quantified using the ImageJ software as described previously [26]. To determine cell surface expression of GABA B receptors, all optical sections were merged into one image. Then, the inner as well outer border of somatic staining was carefully outlined and the mean intensity values were measured. Values for cell surface signals were obtained by subtracting the values of the inner border from those of the outer border. For analysis of total GABA B receptor expression, the somata of neurons were carefully outlined and the mean intensity values were measured. For each image, background staining was determined in an area containing no specific signal and subtracted from the mean intensity values. All values were normalized to the analyzed cell area.
Proteins overexpressed in neurons were either tagged with GFP or the HA epitope, which were used to identify transfected neurons and to judge the protein expression level. In case of overexpression studies, transfected neurons with comparable expression of the protein of interested were included into the analysis.

In Situ Proximity Ligation Assay
In situ proximity ligation assay (PLA) is an antibody-based technology for the detection of protein-protein interactions and posttranslational modifications of proteins [27,28]. We applied in situ PLA for the analysis of GABA B receptor K48linked or K63-linked ubiquitination, serine phosphorylation, and CaMKII/GABA B receptor interaction as well as for the co-localization with the endosomal marker proteins EEA1, Rab7, and Rab11. In situ PLA was performed using Duolink PLA probes and detection reagents (Sigma-Aldrich) according to the manufacturer's instructions as described previously [26]. Quantification was done by counting individual in situ PLA spots using the Image J software. The optical sections of each image stack were merged into one image and the number of maxima was determined after setting an appropriate noise tolerance (noise tolerance was kept constant for all images of an experiment). The number of spots was normalized to the area analyzed and to the expression level of GABA B receptors. In some experiments, signals were too abundant to resolve individual spots precisely. In those cases, the fluorescence intensity of signals was determined.

Internalization Assay
Neurons were placed on ice and cell surface GABA B receptors were labeled with GABA B2 antibodies for 1 h at 4°C. After extensive washes, neurons were incubated in 37°C warm medium for exactly 1, 2, 3, 5, or 10 min to permit receptor endocytosis. Endocytosis was terminated by replacing the medium with ice-cold buffer. After incubation with secondary antibody for 1 h at 4°C, the neurons were fixed with 4% PFA and remaining cell surface GABA B receptors were determined by laser scanning confocal microscopy.

Western Blot Analysis
For Western blot analysis, neuron/glia co-cultures were grown for 13-14 days on 6-cm culture dishes plated with 450,000 cells obtained from the cerebral cortex of E18 rat embryos. After incubation with KN93, cultures were immediately washed twice with ice-cold PBS, harvested, and homogenized by sonication. After protein determination using the Bradford protein assay (BioRad), the samples were incubated with Laemmli sample buffer (BioRad) for 2 h at 37°C and aliquots containing 30 μg protein were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) using 7.5% mini-gels (Mini Protean 3, BioRad). Proteins were transferred onto nitrocellulose membranes in a Mini Trans-Blot Module (BioRad) at 365 mA for 120 min using 192 mM glycine, 25 mM Tris, 0.1% SDS, 20% methanol as transfer buffer. After blotting, the transferred proteins were stained with Amidoblack and immediately imaged using the E-box VX2 gel imager (Vilber). For immunodetection, the blots were blocked for 1-2 h in PBST (PBS pH 7.4, 0.05% Tween 20) containing 5% nonfat dry milk at room temperature, followed by incubation with antisera overnight at 4°C in PBST containing 5% non-fat dry milk. The blots were then washed five times for 10 min with TBST and incubated with secondary antibodies conjugated to horseradish peroxidase for 1 h at room temperature. Following extensive washing (see a b o v e ) , im m u n o r e a c t i v i t y w a s d e te c t e d b y t h e chemoluminescence method (SuperSignal West Dura, Thermo Scientific) using a Fujifilm LAS-1000 imager. Immunoreactivity was quantified with the Image Studio software (LI-CORE Biosciences) and normalized to total protein in the corresponding lanes (determined by Amidoblack staining, see above).

Radioligand Binding
For radioligand binding, 450,000 cells derived from the cerebral cortex of E18 rat embryos were plated onto 6-cm polylysine-coated culture dishes and kept in culture for 14 days. After drug incubation, cells were immediately washed two times with ice-cold PBS and harvested. Cells were resuspended in 500 μl binding buffer containing protease inhibitors (50 mM Tris pH 7.4, 2.5 mM CaCl 2 , complete mini, Sigma-Aldrich) and homogenized using a Potter S homogenizer (B. Braun Biotech International). Aliquots of the homogenate were incubated with 6.7 nM [ 3 H]CGP 54626 in binding buffer for 90 min at room temperature. The incubation was terminated by rapid filtration onto glass fiber filters using a 12channnel semi-automated cell harvester (Scatron) and washed with ice-cold binding buffer. Non-specific [ 3 H]CGP 54626 binding was determined in the presence of 10 μM CGP 56999A. The radioactivity retained by the filters was determined by liquid scintillation counting using a Tricarb 2500 liquid scintillation analyzer.

Statistics
The statistical analyses were done with GraphPad Prism 5.
The tests used and p values are given in the figure legends. Differences were considered statistically significant when p < 0.05.

CaMKIIβ Controls Expression of GABA B Receptors
Sustained activation of glutamate receptors, which occurs in cerebral ischemia, downregulates GABA B receptors by preferentially sorting constitutively internalized receptors to lysosomes for degradation instead of recycling them to the cell surface [29][30][31][32]. Phosphorylation of GABA B1 at Ser-867 by CaMKII plays a key role in this mechanism [29]. However, it was unknown whether CaMKII is also involved in the regulation of GABA B receptor expression under normal physiological conditions. Therefore, we tested whether inhibition of CaMKII by KN93 affects total expression of GABA B receptors in cultured cortical neurons. Incubation of neurons with KN93 transiently increased the expression of GABA B1 and GABA B2 , which peaked at about 7.5 min as tested by Western blotting (Fig. 1a). This was confirmed by radioligand binding experiments on homogenates prepared from the cultures using the GABA B receptor antagonist [ 3 H]CGP 54626 (Fig. 1b). Finally, we monitored cell surface expression of GABA B receptors by staining living neurons at 4°C with antibodies directed against the extracellular located N-terminal domain of the receptors. In line with the experiments on total receptor expression, incubation of neurons with KN93 for 7.5 min considerably increased the level of GABA B receptors at the cell surface (148 ± 64% of control; Fig. 1c). The cause for the transient nature of blocking CaMKII on cell surface expression of the receptors is currently unclear. It might well be that a so far unknown homeostatic response of the neurons downregulates the increased cell surface receptors to normal levels. However, our results clearly indicate that the expression of the receptors in the plasma membrane is indeed controlled by CaMKII under basal conditions. CaMKIIα and CaMKIIβ are the main CaMKII isoforms expressed in neurons [33]. We therefore tested which isoform is involved in the regulation of GABA B receptors by transfecting neurons with either CaMKIIα, CaMKIIβ, or their f u n c t i o n a l l y i n a c t i v e m u t a n t s ( C a M K I I α ( D N ) , CaMKIIβ(DN)) and determined cell surface expression of GABA B receptors using GABA B2 antibodies. While transfecting CaMKIIα or CaMKIIα(DN) did not affect cell surface expression of GABA B receptors, overexpression of CaMKIIβ decreased (43 ± 17% of EGFP-transfected control neurons, Fig. 2a) and its nonfunctional mutant CaMKIIβ(DN) increased cell surface levels of receptors (186 ± 29% of EGFP-transfected control neurons, Fig. 2b).
It was previously shown that CaMKII interacts with GABA B1 [29]. Next, we tested whether CaMKIIβ or CaMKIIα interacts with GABA B receptors under basal conditions by overexpressing either CaMKIIα or CaMKIIβ in neurons and testing for changes in interaction levels with GABA B receptors using in situ PLA with antibodies against GABA B1 and CaMKII (Fig. 2b). In line with our experiments described above, overexpression of CaMKIIα did not affect the level of GABA B receptor/CaMKII interaction (PLA signal EGFP-transfected control, 100 ± 32%; CaMKIIα, 80 ± 30%), but transfection of CaMKIIβ significantly increased the extent of interaction (216 ± 58% of control). The interaction of CaMKIIβ with GABA B receptors was activity-dependent as blocking CaMKII with KN93 reduced the interaction level in neurons transfected with CaMKIIβ (PLA signal. 110 ± 40% of control).
These results indicate that CaMKIIβ interacts with GABA B receptors in an activity-dependent manner to regulate the cell surface expression of the receptors.

CaMKII Triggers Sorting of GABA B Receptors to Lysosomal Degradation
It had been suggested that phosphorylation of GABA B1 Ser-867 by CaMKII may mediate internalization of the receptors [29]. We therefore tested whether inhibition of CaMKII by KN93 affects the internalization of GABA B receptors. However, blocking CaMKII with KN93 did affect neither the kinetics (half live; control, 1.2 min; KN93, 1.6 min) nor the extent (control, 64 ± 19%; KN93, 65 ± 20%) of receptor internalization (Fig. 3a).
To gain insight into the mechanism that is affected by CaMKII inhibition, we tested the co-localization of GABA B receptors with the endosomal marker proteins EEA1 (marker for early endosomes), Rab7 (marker for late endosomes and lysosomes), and Rab11 (marker for recycling endosomes) [34] using in situ PLA. Inhibition of CaMKII activity in neurons with KN93 did not affect the co-localization of GABA B receptors with EEA1 (94 ± 26% of control; Fig. 3b), supporting our observation of an unchanged internalization rate of the receptors. Instead, blocking CaMKII decreased the co-localization of GABA B receptors with Rab7 (39 ± 28% of control; Fig. 3b) and increased the co-localization of GABA B receptors with Rab11 (145 ± 32% of control; Fig.  3b). Thus, inhibition of CaMKII appears to prevent recruitment of GABA B receptors to the lysosomal degradation pathway (Rab7) and increase recycling of the receptors (Rab11). This observation indicates that basal CaMKII activity is involved in targeting GABA B receptors to lysosomes for degradation.

CaMKIIβ Promotes Lys-63-Linked Ubiquitination of GABA B Receptors
The results so far support the hypothesis that CaMKIIβ regulates lysosomal degradation of GABA B receptors. We had previously shown that lysosomal degradation of GABA B receptors depends on Lys-63-linked ubiquitination of GABA B1 at multiple sites via the E3 ligase MIB2 [17]. It was therefore likely that CaMKIIβ controls Lys-63-linked ubiquitination of GABA B receptors. To test this hypothesis, we first analyzed the effect of various ubiquitin mutants on the downregulation of cell surface GABA B receptors induced by overexpression of CaMKIIβ. Overexpression of CaMKIIβ in neurons reduced cell surface GABA B receptors to 43 ± 18% of control neurons transfected with EGFP (Fig. 4). Additional overexpression of wild-type ubiquitin (Ub, 36 ± 25% of control) and a ubiquitin mutant that permits only Lys-63linked ubiquitination (Ub(K63), 35 ± 18% of control) as well as a mutant that prevents Lys-48-linked ubiquitination (Ub(K48R), 39 ± 24% of control) did not affect CaMKIIβ-induced downregulation of the receptors (43 ± 18% of control). However, overexpressing a ubiquitin mutant in which all lysine residues were mutated to arginine to entirely block polyubiquitination (Ub(KO), 109 ± 22% of control) as well as a ubiquitin mutant that specifically prevents Lys-63-linked ubiquitination (Ub(K63R), 124 ± 40% of control) completely prevented downregulation of the receptors or even increased cell surface expression of the receptors (Fig. 4). These results indicate that downregulation of GABA B receptors induced by overexpression of CaMKIIβ is mediated by Lys-63-linked ubiquitination. Next, we tested for direct CaMKIIβ-induced Lys-63-linked ubiquitination of GABA B receptors using in situ PLA and antibodies directed against GABA B1 and Lys-63-linked ubiquitin. Inhibition of CaMKII with KN93 significantly reduced basal Lys-63-linked ubiquitination of the receptor (69 ± 43% of control, Fig. 5a). In addition, overexpression of the functionally inactive CaMKIIβ mutant CaMKIIβ(DN) in neurons considerably reduced Lys-63-linked ubiquitination of the receptors (46 ± 24%, Fig. 5b) as compared to neurons transfected with wild-type CaMKIIβ. In contrast, overexpression of the functionally inactive mutant of CaMKIIα (CaMKIIα(DΝ)) had no effect on the level of Lys-63-linked ubiquitinated GABA B receptors as compared to wild-type CaMKIIα. Finally, overexpression of neither CaMKIIα(DN) nor CaMKIIβ(DN) affected Lys-48-linked ubiquitination of the receptors, which is required for their proteasomal Fig. 3 CaMKII triggers sorting of GABA B receptors to lysosomes. a Blocking CaMKII activity does not affect internalization of GABA B receptors. GABA B2 antibody-labeled cell surface receptors of living neurons were incubated for the indicated time intervals at 37°C to permit endocytosis. Subsequently, remaining cell surface GABA B2labeled receptors were quantified by immunofluorescence microscopy. The immunofluorescence signals of neurons at time 0 (not exposed to 37°C) were set to 100%. Data were fitted by nonlinear regression to one phase decay. The data represent the mean ± SD of 35-46 neurons per time point from two independent experiments. b Blocking CaMKII activity decreased the colocalization of GABA B receptors with Rab7 and increased the colocalization with Rab11. Neurons were either treated or not with KN93 and subjected to in situ PLA using antibodies directed against GABA B1 and Rab11 (marker for recycling endosomes), Rab7 (marker for the lysosomal pathway), or EEA1 (marker for early endosomes), respectively. Left, representative images (scale bar 5 μm). Right, quantification of PLA signals (white dots in images). The PLA signals were normalized to the cell area. The control condition was set to 100%. The data represent the mean ± SD of 30 neurons (Rab11) and 40 neurons (Rab7, EEA1) from two independent experiments. ****p < 0.0001, ns p > 0.05; two-tailed unpaired t test degradation (Fig. 5c). These results indicate that CaMKIIβ is involved in the regulation of Lys-63-linked ubiquitination of GABA B receptors.

CaMKII Regulates Cell Surface Receptors by Promoting MIB2-Mediated Lys-63-Linked Ubiquitination via Phosphorylation of GABA B1 at Ser-867
CaMKII phosphorylates GABA B1 at Ser-867 [29]. To prove that direct phosphorylation of Ser-867 in GABA B1 is involved in the regulation of GABA B receptor cell surface expression, we silenced this phosphorylation site by mutating it to alanine (GABA B1 (S867A)) or rendered the site permanently active by mutating it to aspartate (GABA B1 (S867D)). First, we verified that CaMKIIβ indeed phosphorylates GABA B1 at Ser-867 by Representative images of stained neuronal somata (scale bar, 5 μm). The scatter plot shows quantification of fluorescence intensities. The fluorescence signal of neurons transfected with EGFP was set to 100%. The data represent the mean ± SD of 28-30 neurons from two independent experiments. ***p < 0.001, ns p > 0.05; one-way ANOVA, Dunnett's multiple comparison test in situ PLA using antibodies directed against GABA B1 and phosphoserine. Upon expression in HEK 293 cells, wild-type GABA B receptors displayed a basal level of serine phosphorylation, which was more than threefold enhanced by coexpression with CaMKIIβ (PLA signal wt-GABA B1 : 98 ± 5, GABA B1 +CaMKIIβ: 358 ± 14; Fig. 7a). In contrast, GABA B receptors in which Ser-867 of GABA B1 was inactivated by mutation to alanine, only basal level of serine phosphorylation was detected in the presence of CaMKIIβ (PLA signal GABA B1 (S867A)+CaMKIIβ: 108 ± 6 compared to wt-GABA B1 : 98 ± 5, Fig. 7a). These results confirm that CaMKIIβ specifically phosphorylates GABA B1 at Ser-867.
Next, we tested the effect of inactivating or permanently mimicking phosphorylation of GABA B1 Ser-867 on cell surface expression of the receptors by expressing GABA B1 (S867A) or GABA B1 (S867D) in cortical neurons. Inactivation of GABA B1 Ser-867 considerably increased cell surface expression of receptors containing GABA B1 (S867A) (271 ± 119% of control, Fig. 7b) upon transfection in neurons, whereas receptors containing the phosphomimetic GABA B1 (S867D) mutant were expressed at reduced levels (50 ± 22% of control, Fig. 7b). In line with this finding, receptors containing the CaMKII phosphorylation-deficient GABA B1 (S867A) mutant exhibited reduced K63-linked ubiquitination (47 ± 32% of control, Fig. 7c), while receptors containing the phosphomimetic GABA B1 (S867D) mutant showed a considerably higher level of Lys-63-linked ubiquitination (226 ± 76% of control, Fig. 7c). These findings We previously showed that Lys-63-linked ubiquitination of GABA B1 is mediated by the ubiquitin E3 ligase MIB2 [17]. We therefore tested for the involvement of MIB2 in CaMKIIβ-induced downregulation of GABA B receptors. Overexpression of MIB2 or CaMKIIβ reduced cell surface expression of GABA B receptors in neurons to a similar extent (MIB2: 40 ± 16%, CaMKIIβ: 45 ± 17% of EGFP-transfected control neurons; Fig. 8a). Co-transfection of MIB2 and CaMKIIβ in neurons did not further increase downregulation of the receptors (43 ± 14% of EGFP-transfected control neurons; Fig. 8a), indicating that both enzymes affect GABA B receptors via the same pathway. In line with this observation, Fig. 7 Phosphorylation of GABA B1 Ser-867 controls cell surface expression and Lys-63-linked ubiquitination of GABA B receptors. a CaMKIIβ phosphorylates GABA B1 at Ser-867. HEK293 cells were transfected either with GABA B1 /GABA B2 with or without CaMKIIβ or with GABA B1 (S867A)/GABA B2 and CaMKIIβ. Two days after transfection, the cells were tested for serine phosphorylation by in situ PLA using antibodies directed against GABA B1 and phosphoserine. Left, representative images (scale bar, 10 μm; wt, wild type). Right, quantification of PLA signals (white dots in images). The PLA signals were normalized to the cell area as well as to the expression level of GABA B1 . The data represent the mean ± SD of 50 cells from two independent experiments. ***p < 0.001, ns p > 0.05; one-way ANOVA, Dunnett's multiple comparison test. b, c Inactivation or constitutively mimicking phosphorylation of GABA B1 Ser-867 affects cell surface expression (b) as well as Lys-63linked ubiquitination (c) of the receptors. Neurons were transfected with wild-type HA-tagged GABA B1a (control), with the phosphorylationdeficient HA-tagged GABA B1a (S867A) mutant or with the phosphomimetic HA-tagged GABA B1a (S867D) mutant along with GABA B2 and tested for cell surface expression using HA antibodies (b) as well as for Lys63-linked ubiquitination by in situ PLA with HA and Lys-63 antibodies (c). Left, representative images (scale bars, 5 μm). Right, quantification of fluorescence intensities (b) and PLA signals (c). The data represent the mean ± SD of 23-27 (b) and 18-20 (c) neurons from two independent experiments. *p < 0.05, ***p < 0.001; one-way ANOVA, Dunnett's multiple comparison test Fig. 6 Blocking CaMKII did not affect the expression levels of GABA B1a (K->R) mutants. Neurons were transfected with HA-tagged wild-type GABA B1a or HA-tagged GABA B1a (K->R) mutants along with GABA B2 and tested for cell surface expression of GABA B1 and GABA B2 after treating the neurons with 10 μM KN93 for 7.5 min. Left, representative images of untreated neurons (control, left panels) and of neurons treated with KN93 (right panels, scale bar 5 μm). The corresponding graphs show the quantification of fluorescence signals. The fluorescence intensity of neurons transfected with wild-type GABA B1 from untreated neurons (control) was set to 100%. The data represent the mean ± SD of 45-60 neurons per experimental condition derived from three independent experiments. Please note that the considerably lower increase in GABA B2 cell surface expression as compared with mutant GABA B1 was due to the fact that in the case of GABA B1 , only transfected subunits were detected (HA-tagged) but in the case of GABA B2 , transfected as well as endogenously expressed subunits were detected. ***p < 0.001, ns p > 0.05; two-tailed unpaired t test MIB2 was unable to downregulate cell surface GABA B receptors when the nonfunctional CaMKIIβ mutant CaMKIIβ(DN) was simultaneously overexpressed with MIB2 (CaMKIIβ(DN): 195 ± 42%, CaMKIIβ(DN)+MIB2: 200 ± 43% of EGFP-transfected control neurons, Fig. 8a).
Finally, we tested the hypothesis that CaMKII-mediated phosphorylation of GABA B1 might affect the interaction of MIB2 with GABA B receptors. Upon transfection in HEK 293 cells, MIB2 displayed considerably lower interaction with receptors containing the phospho-deficient mutant GABA B1a (S867A) (25 ± 15% of control, Fig. 8c) and a higher level of interaction with receptors containing the phosphomimetic mutant GABA B1a (S867D) (126 ± 44% of control, Fig. 8c) than wild-type receptors. This result indicates that phosphorylation of GABA B1 at Ser-867 promotes the interaction of GABA B receptors with MIB2.

Sustained Activation of Glutamate Receptors Increases Lys-63-Linked Ubiquitination of GABA B Receptors via Phosphorylation of GABA B1 Ser-867
Glutamate-induced downregulation of GABA B receptors via lysosomal degradation depends on CaMKII-mediated phosphorylation of GABA B1 Ser-867 [29] as well as on Lys-63linked ubiquitination of GABA B1 at Lys-697/698, Lys-892, and Lys-960 [17]. To investigate the role of CaMKII in this process, we first analyzed the effect of sustained glutamate treatment on the expression level of CaMKII as well as on the degree of CaMKII-GABA B receptor interaction. Treatment of neurons for 60 min with glutamate significantly upregulated CaMKII (233 ± 61% of control, Fig. 9a) and considerably increased the interaction of CaMKII with GABA B receptors (165 ± 74% of control, Fig. 9b).
These findings suggest that sustained activation of glutamate receptors induces GABA B1 -Ser-867 phosphorylationmediated K63-linked ubiquitination of GABA B receptors, promoting their lysosomal degradation.

Discussion
The expression of GABA B receptors available for signal transduction at the cell surface critically depends on their rate of degradation. The two main cellular degradation systemsproteasomes and lysosomes-control the abundance of GABA B receptors in distinct cellular compartments in response to the activity state of the neuron [10,11,17]. In the ER, the amount of newly synthetized GABA B receptors destined for trafficking to the plasma membrane is regulated by proteasomal degradation via the ERAD machinery [10], whereas cell surface GABA B receptors are degraded in lysosomes [12][13][14][15][16]. Both degradation pathways require Fig. 8 The ubiquitin E3 ligase MIB2 is involved in CaMKIIβ-mediated regulation of cell surface GABA B receptors. a Overexpression of MIB2 in addition to CaMKII does not enhance downregulation of cell surface GABA B receptors. Neurons were transfected with the indicated plasmid and tested for cell surface expression of GABA B receptors using GABA B2 antibodies 2 days after transfection. Co-expression of MIB2 did not further enhance downregulation of cell surface receptors, indicating the involvement in the same pathway. Upper panels show representative images (scale bars, 5 μm). The graphs depict quantification of the fluorescence signals. The data represent the mean ± SD of 30 neurons per condition derived from two independent experiments. ns p > 0.05, ****p < 0.0001; two-way ANOVA with Bonferroni's multiple comparison test (interaction: F(2,234) = 32.78, p < 0.0001). b MIB2-induced downregulation of cell surface receptors depends on phosphorylation of GABA B1 Ser-867. Neurons were transfected with wild-type HA-tagged GABA B1a (control), with the phosphorylation-deficient HA-tagged GABA B1a (S867A) mutant or with the phosphomimetic HA-tagged GABA B1a (S867D) mutant along with GABA B2 and with or without MIB2. Two days after transfection, neurons were tested for cell surface expression of tagged GABA B1 using HA antibodies. Upper panels show representative images (scale bars, 5 μm). The graph depicts quantification of the fluorescence signals. The data represent the mean ± SD of 30 neurons per condition derived from two independent experiments. ns p > 0.05, ****p < 0.0001; two-way ANOVA, Bonferroni's multiple comparison test (interaction: F(2,174) = 49.09, p < 0.0001). c GABA B1 phosphomutants display altered MIB2 interaction. HEK 293 cells were transfected with wild-type GABA B1 , GABA B1 (S867A), or GABA B1 (S867D) together with GABA B2 and MIB2. Two days after transfection, cells were analyzed for interaction of GABA B receptors with MIB2 by in situ PLA using antibodies directed against GABA B1 and MIB2. Upper panels show representative images (scale bar, 10 μm). The graph depicts quantification of the PLA signals. The data represent the mean ± SD of 35 neurons per condition derived from two independent experiments. **p < 0.01, ***p < 0.001; one-way ANOVA, Dunnett's multiple comparison test ubiquitination of the receptor as targeting signals. Lys-48linked ubiquitination of GABA B2 at Lys-767/771 tags GABA B receptors for proteasomal degradation [10] and Lys-63-linked ubiquitination of GABA B1 at several sites sorts the receptors to lysosomes [17]. Interestingly, Lys-48-and Lys-63-linked ubiquitination appears to be largely segregated to GABA B2 and GABA B1 [10,17], respectively, which might be explained by a selective interaction of the subunits with the respective E3 ligases. However, the mechanisms regulating lysosomal degradation of GABA B receptors in response to changes in the physiological state of the neuron were unclear.
Under normal physiological conditions, cell surface GABA B receptors constitutively internalize to early endosomes and recycle to the plasma membrane or are sorted to lysosomes for degradation [12-14, 35, 36]. Sorting the receptors to recycling or degradation must be precisely regulated in order to provide the required number of cell surface receptors for signal transduction under a given physiological condition. However, after prolonged activation of glutamate receptors, GABA B receptors are rapidly downregulated by shifting the recycling/degradation balance towards lysosomal degradation [30][31][32]. This downregulation of GABA B receptors critically depends on CaMKII-mediated phosphorylation of GABA B1 on Ser-867 [29] as well as on Lys-63-linked ubiquitination of GABA B1 mediated by the E3 ligase MIB2 [17]. As the level of CaMKII activity is regulated by the intracellular Ca 2+ concentration, it is in an ideal position to link the increased neuronal activity after prolonged activation of glutamate receptors to lysosomal degradation of GABA B receptors at least under pathological conditions. We, therefore, hypothesized that CaMKII regulates lysosomal degradation also under physiological conditions. Indeed, we found that pharmacologically blocking CaMKII significantly increased cell surface expression of GABA B receptors. This increase was accompanied by a reduced co-localization of the receptors with the late endosomal marker Rab7, indicating that inhibition of CaMKII prevented sorting of GABA B receptors to lysosomes. Our data also imply that a considerable fraction of GABA B receptors is constitutively degraded under basal conditions since blocking CaMKII activity for only 7.5 min was sufficient to induce a significant increase (~150%) of cell surface GABA B receptors.
CaMKII is a large protein complex constituted from 12 catalytically active subunits of different isoforms (CaMKIIα-δ) [33]. In brain, CaMKIIα and CaMKIIβ are the predominant subunits and in forebrain neurons the CaMKII holoenzyme largely consists of nine α subunits and three β subunits [37,38]. Interestingly, we found that overexpression of CaMKIIβ, but not CaMKIIα, increased the level of GABA B receptor/CaMKII interaction in an activitydependent manner (inhibition of CaMKII activity prevented the increase in interaction), resulting in decreased cell surface receptor expression. Consistent with this observation, overexpressing its functionally inactive mutant CaMKIIβ(DN), which was expected to inhibit CaMKIIβ activity and thus lysosomal degradation of the receptors, increased cell surface expression of GABA B receptors. As intracellular Ca 2+ concentrations under basal conditions are low, this finding fits well to the observation that CaMKIIβ exhibits a ninefold higher affinity to Ca 2+ /calmodulin than CaMKIIα [38]. Thus, our data suggest that under normal physiological conditions basal CaMKIIβ activity determines the level of lysosomal degradation of GABA B receptors.
Phosphorylation often regulates ubiquitination of proteins, thereby promoting their degradation [39][40][41]. Several lines of evidence indicate that regulation of cell surface GABA B receptors by CaMKIIβ is conveyed by MIB2-mediated Lys-63-linked ubiquitination of GABA B1 . First, overexpression in neurons of ubiquitin mutants unable to form Lys-63-linkages prevented downregulation of GABA B receptors induced by overexpression of CaMKIIβ, while a mutant that only can form Lys-63-linkages, but not any other kind of linkages, did not inhibit downregulation of the receptors. Second, blocking CaMKII activity pharmacologically or by overexpression of the functionally inactive mutant CaMKIIβ(DN) reduced Lys-63-linked ubiquitination of GABA B receptors. Third, the cell surface expression of three GABA B1a (K->R) mutants with inactivated Lys-63-linked ubiquitination sites, which are indispensable for lysosomal targeting of the receptors [17], remained unaffected by CaMKII inhibition. Fourth, overexpression of the E3 ligase MIB2, which downregulates GABA B receptors via lysosomal degradation [17], neither enhanced CaMKIIβ-mediated downregulation of cell surface receptors nor affected the expression of a GABA B1 mutant with inactivated CaMKII phosphorylation site (GABA B1 (S867A)). Finally, our experiments with the CaMKII Fig. 9 Glutamate-induced downregulation of GABA B receptors. a Prolonged glutamate treatment increased CaMKII expression. Neurons were treated with 50 μM glutamate for 60 min and analyzed after additional 16 h for CaMKII expression. Upper panel shows representative images (scale bar, 5 μm). The data represent the mean ± SD of 45 neurons per experimental condition derived from two independent experiments. ****p < 0.0001, two-tailed unpaired t test. b Prolonged glutamate treatment increased the interaction of CaMKII with GABA B receptors. Neurons were treated with 50 μM glutamate for 60 min and analyzed after additional 16 h for the interaction of CaMKII with GABA B receptors via in situ PLA using CaMKII and GABA B1 antibodies. Upper panel shows representative images (scale bar, 5 μm). The data represent the mean ± SD of 39 neurons per experimental condition derived from two independent experiments. ****p < 0.0001, twotailed unpaired t test. c, d Glutamate-induced downregulation of GABA B receptors is mediated by GABA B1 Ser-867 phosphorylation-induced Lys-63-linked ubiquitination. Neurons transfected with wild-type HA-tagged GABA B 1 a , with the phosphorylation-deficient HA-tagged GABA B1a (S867A) mutant or with the phosphomimetic HA-tagged GABA B1a (S867D) mutant along with GABA B2 were incubated for 60 min in the absence (control) or presence of 50 μM glutamate. Neurons were analyzed for cell surface expression of transfected GABA B1 using HA antibodies (c) or analyzed for Lys-63-linked ubiquitination by in situ PLA using HA and Lys-63 antibodies (d). Upper panels show representative images (scale bars, 5 μm). The graphs depict quantification of the fluorescence signals (c) and in situ PLA signals (d). The data represent the mean ± SD of 28-30 (c) and 28-36 (d) neurons per condition derived from three independent experiments. ns p > 0.05, ****p < 0.0001; two-way ANOVA, Bonferroni's multiple comparison test (c interaction: F(2,170) = 6.27, p < 0.005; d interaction: F(2, 205) = 17.22, p < 0.0001) phosphorylation-deficient and phosphomimetic GABA B1 (S867) mutants indicate that direct phosphorylation of GABA B1 Ser-867 promotes Lys-63-linked ubiquitination and thereby degradation of GABA B receptors. Under normal physiological as well as conditions of over-excitation via prolonged activation of glutamate receptors, the phospho-deficient mutant GABA B (S867A) displayed only marginal Lys-63-linked ubiquitination and increased cell surface expression whereas the phosphomimetic mutant GABA B1 (S867D) exhibited strongly increased Lys-63linked ubiquitination and reduced cell surface expression.
The precise mechanism by which CaMKIIβ promotes Lys63linked ubiquitination of GABA B1 remains unresolved. Phosphorylation of GABA B1 (S867) might be the signal that targets the receptors to an endosomal compartment where they are Lys-63-linked ubiquitinated by MIB2. Alternatively, phosphorylation of GABA B1 (S867) might induce a conformational change in the receptor that uncovers the MIB2 interaction site or exposes the lysine residues for ubiquitination. We favor the latter hypothesis, which is supported by our finding that MIB2 displayed increased interaction with the phosphomimetic mutant GABA B1 (S867D) and reduced interaction with the phosphodeficient mutant GABA B1 (S867A). However, further experimentation is required to reveal the mechanism as to how CaMKIIβ promotes Lys-63-linked ubiquitination of GABA B1 .
In addition to CaMKIIβ controlling cell surface GABA B receptor expression under normal physiological conditions, it rapidly downregulates the receptors upon sustained activation of glutamate receptors [29][30][31][32], a pathological condition that occurs in cerebral ischemia and leads to excitotoxic cell death [42,43]. Downregulation of GABA B receptors is triggered by increased Ca 2+ influx through NMDA receptors and voltagegated Ca 2+ channels [30], leading to enhanced CaMKIImediated phosphorylation of GABA B1 on Ser867 [29]. This causes the re-routing of GABA B receptors from the normal constitutive recycling pathway to lysosomal degradation. Our present results show that this pathological downregulation of GABA B receptors is maintained after removal of glutamate due to the upregulation of CaMKII expression and the increased interaction of CaMKII with GABA B receptors. Therefore, interfering with this pathway to normalize GABA B receptormediated inhibition might be a potential approach to counteract over-excitation and to limit neuronal death.
In conclusion, our data indicate that phosphorylation of GABA B1 at Ser-867 by CaMKIIβ induces MIB2-mediated K63-linked ubiquitination of GABA B1 at multiple sites, which sorts GABA B receptors to lysosomes for degradation. This mechanism is expected to fine-tune cell surface expression of GABA B receptors under physiological conditions and to considerably affect receptor expression in diseases associated with disturbed Ca 2+ homeostasis.