NMDA receptor activity downregulates KCC2 resulting in depolarizing GABAA receptor-mediated currents

Abstract

KCC2 is a neuron-specific K(+)-Cl(-) co-transporter that maintains a low intracellular Cl(-) concentration that is essential for hyperpolarizing inhibition mediated by GABA(A) receptors. Deficits in KCC2 activity occur in disease states associated with pathophysiological glutamate release. However, the mechanisms by which elevated glutamate alters KCC2 function are unknown. The phosphorylation of KCC2 residue Ser940 is known to regulate its surface activity. We found that NMDA receptor activity and Ca(2+) influx caused the dephosphorylation of Ser940 in dissociated rat neurons, leading to a loss of KCC2 function that lasted longer than 20 min. Protein phosphatase 1 mediated the dephosphorylation events of Ser940 that coincided with a deficit in hyperpolarizing GABAergic inhibition resulting from the loss of KCC2 activity. Blocking dephosphorylation of Ser940 reduced the glutamate-induced downregulation of KCC2 and substantially improved the maintenance of hyperpolarizing GABAergic inhibition. Reducing the downregulation of KCC2 therefore has therapeutic potential in the treatment of neurological disorders.

Figure 1

The furosemide-sensitive K⁺/Cl⁻ pump KCC2 generates hyperpolarizing GABA-activated currents. Recordings of exogenous GABA (10 μM) application were obtained in I=0 recording mode (A–C). A, Consecutive pulses of GABA (5 s, black bars) spaced 1 min apart. B, Exposing neurons to furosemide (100 μM) abolished the hyperpolarizing effects of GABA application within minutes. The traces shown were not consecutive pulses (see text). C, Hyperpolarizing GABA-activated potentials were quickly reestablished after washout of furosemide. Traces shown are from consecutive GABA pulses spaced at 1 min intervals. All data were obtained in the presence of DNQX, AP5, and TTX.

Figure 2

Glutamate application switches GABA transmission to depolarizing and excitatory. A-top, Consecutive pulses of GABA (10 μM-black bars) spaced 1 min apart. A-bottom, Neurons exhibited sIPSP’s that were blocked by bicuculline (50 μM-black bar). Consecutive traces are overlaid in black and gray. B, Glutamate (20 μM) application for 2 min caused a sustained depolarization. Arrow indicates the barrage of action potentials (APs) fired before the depolarizing block. C, Four consecutive overlaid traces obtained 1 min after the end of the glutamate pulse. Solid bar indicates the presence of DNQX and AP5, the dotted bar indicates the addition of bicuculline. D, Quantification of the average number of APs during each 1 min epoch after the glutamate pulse. E, Consecutive 500 ms GABA pulses (black bars) obtained at 2 min intervals. F, After a 2 min glutamate pulse and a 2 min recovery period, GABA was applied every 2 min. Numbers above each trace indicate the time elapsed from the beginning of the glutamate application. For comparison, the last pulse of GABA (gray) from panel (E) prior to glutamate application is superimposed onto the 20^th min GABA pulse (black) obtained after glutamate exposure. G, The graph represents the time course of the shift in GABA responses before and after glutamate. Time 0 represents the beginning of the glutamate pulse. Glutamate was applied in the absence (squares) or presence of bicuculline (triangles). DMSO (0.1 %, diamonds) was applied alone as a control. Error bars represent the mean ± SEM.

Figure 3

Glutamate shifts E_GABA to more depolarized potentials. A, Recordings of the membrane potential responses to GABA (10 μM) were obtained before (left) and 10 min after (right) application of glutamate (20 μM). B, I–V relationships of GABA-activated currents recorded before (left) and after (right) exposure to glutamate. Currents were recorded at the indicated holding potentials shown to the left of each trace. C, Dot plot of the extrapolated E_GABA values obtained before and after glutamate treatment. Lines connect the E_GABA values obtained for each cell. D, I–V plot of normalized GABA-activated currents obtained before (squares, solid line) and after (triangles, dashed line) glutamate treatment. The calculated shift in E_GABA caused by glutamate is displayed above the linear fits to the data points. Error bars represent the mean ± SEM. All of these data were pooled from interleaved glutamate/control experiments performed in parallel with each glutamate/experimental condition.

Figure 4

Glutamate caused the dephosphorylation of S940 and the degradation of KCC2. A, Cultured neurons were treated with 20 μM glutamate for 5 to 15 min before cell lysis. Proteins from whole cell lysates were analyzed by SDS-PAGE followed by immunoblotting with pS940 and anti-KCC2 antibodies. Full-length blots are provided in Supplementary Fig. 3. B, Biotinylation assay was performed on cultured neurons after a 10 min treatment of 20 μM glutamate. Anti-α-tubulin (Tubulin) antibodies were used to show equal protein loading. (*) Indicates statistically significant differences relative to control conditions as assessed by an unpaired t-test (p < 0.01; n = 4). Error bars represent the mean ± SEM.

Figure 5

The glutamate-induced effects on KCC2 were Ca²⁺-dependent. A, Membrane potential responses to GABA (10 μM) were obtained before (left) and after (right) glutamate (20 μM) and EDTA (5 mM) co-application. B, Dot plot of the extrapolated E_GABA values obtained before and after glutamate/EDTA treatment. Lines connect the E_GABA values obtained for each cell. C, I–V plot of normalized GABA-activated currents obtained before (squares, solid line) and after (triangles, dashed line) glutamate/EDTA treatment. The calculated shift in E_GABA caused by glutamate is displayed above the linear fits to the data points. EDTA was applied only during the glutamate pulse. D, Cultured neurons were treated with EDTA (5 mM) for 15 min before administration of glutamate followed by biotinylation and cell lysis. Biotin-labeled surface proteins were pulled down using avidin. Whole cell lysates and biotinylated portions were analyzed by SDS-PAGE followed by immunoblotting. For all panels, anti-α-tubulin (Tubulin) antibodies were used to show equal protein loading. The ratio of signal given by surface KCC2, pS940:KCC2 and the total amount of KCC2 were quantified in the panels shown to the right. (*) Indicates statistically significant differences relative to control conditions as assessed by an unpaired t-test (p < 0.01; n = 4). Error bars represent the mean ± SEM.

Figure 6

The glutamate-induced effects on KCC2 were mediated by NMDA receptors. A, Membrane potential measurements of GABA (10 μM) application were obtained before (left) and after (right) a 2 min exposure to glutamate (20 μM) and AP5 (50 μM). B, Dot plot of the extrapolated E_GABA values obtained before and after glutamate/AP5 treatment. Lines connect the E_GABA values for each cell. C, I–V plot of normalized GABA-activated currents obtained before (squares, solid line) and after (triangles, dashed line) glutamate/AP5 treatment. The calculated shift in E_GABA caused by glutamate is displayed above the linear fits to the data points. D, Cultured neurons were treated with 50 μM AP5 for 15 min before administration of glutamate followed by biotinylation. Biotin-labeled surface proteins were pulled down using avidin. For all panels, anti-α-tubulin (Tubulin) antibodies were used to show equal protein loading. The ratio of signal given by cell surface KCC2, pS940:KCC2 and the total amount of KCC2 were quantified in the panels on the right. (*) Indicates statistically significant differences relative to control conditions as assessed by an unpaired t-test (p < 0.01; n = 4). Error bars represent the mean ± SEM.

Figure 7

The dephosphorylation of S940 and internalization of KCC2 caused by glutamate is dependent on PP1. A, Cultured neurons were treated with 500 nM okadaic acid (OA) or 20 μM cyclosporin A (CsA) for 15 min before cell lysis. Proteins were analyzed by SDS-PAGE followed by immunoblotting using pS940 or anti-KCC2 antibodies. Signals from pS940:KCC2 between control and drug-treated conditions were quantified and compared in the lower panel. B, Different concentrations of OA (0.03 to 3 μM) were applied to cultured neurons for 15 min before cell lysis. Proteins were analyzed by SDS-PAGE followed by immunoblotting using pS940 and anti-KCC2 antibodies. Signals from pS940:KCC2 in each sample were quantified and normalized to total KCC2 before plotting against their respective log concentration values. C, Cultured neurons were treated with 1 μM OA for 15 min before administration of glutamate followed by biotinylation. Biotin-labeled surface proteins were pulled down using avidin. This surface fraction and whole cell lysates were analyzed by SDS-PAGE followed by immunoblotting. The cell surface amount of KCC2, pS940:KCC2 and the total amount of KCC2 protein were quantified in the panels on the right. (*) Indicates significantly different from control condition (p < 0.01, unpaired t-test, n = 3). Error bars represent the mean ± SEM.

Figure 8

Okadaic acid reduced the glutamate-induced shift in E_GABA. OA was applied 10 min prior to recording and for the entire duration of the experiment, including the glutamate pulse. A, Membrane potential measurements of GABA (10 μM) application were obtained before (left) and after (right) glutamate (20 μM) treatment. B, Dot plot of the extrapolated E_GABA values obtained before and after glutamate treatment. Lines connect the E_GABA values obtained for each cell both before and after glutamate exposure. C, I–V plot of normalized GABA-activated currents obtained before (squares, solid line) and after (triangles, dashed line) glutamate treatment. Error bars represent the mean ± SEM.