Age-dependent requirement of AKAP150-anchored PKA and GluR2-lacking AMPA receptors in LTP

Abstract

Association of PKA with the AMPA receptor GluR1 subunit via the A kinase anchor protein AKAP150 is crucial for GluR1 phosphorylation. Mutating the AKAP150 gene to specifically prevent PKA binding reduced PKA within postsynaptic densities (>70%). It abolished hippocampal LTP in 7-12 but not 4-week-old mice. Inhibitors of PKA and of GluR2-lacking AMPA receptors blocked single tetanus LTP in hippocampal slices of 8 but not 4-week-old WT mice. Inhibitors of GluR2-lacking AMPA receptors also prevented LTP in 2 but not 3-week-old mice. Other studies demonstrate that GluR1 homomeric AMPA receptors are the main GluR2-lacking AMPA receptors in adult hippocampus and require PKA for their functional postsynaptic expression during potentiation. AKAP150-anchored PKA might thus critically contribute to LTP in adult hippocampus in part by phosphorylating GluR1 to foster postsynaptic accumulation of homomeric GluR1 AMPA receptors during initial LTP in 8-week-old mice.

Figure 1

Genotyping of AKAP150 D36 mice. The AKAP150 C-terminal coding sequence was amplified by PCR from tail DNA. The resulting 327 bp product (+/+) was treated with XbaI to detect the restriction site that had been introduced by the creation of the stop codon that truncated the last 36 amino acids of AKAP150. The D36 but not the WT allele gave rise to a 208 bp fragment and a 119 bp fragment (−/−). Arrows indicate from top 327, 208, and 119 bp bands.

Figure 2

PKA is reduced in PSDs from D36 mice. (A) Subcellular fractions were prepared from forebrains of WT and D36 mice that were 8 weeks or older by differential centrifugation (P2), sucrose gradient centrifugation (enriched for synaptosomes, Syn), extraction with Triton X-100 and subsequent sucrose gradient centrifugation to remove presynaptic material (Tx-I), and a second extraction with Triton X-100 before ultracentrifugation to obtain purified PSDs (PSD). A 10 μg weight of total protein for each fraction (quantified by bicinchoninic acid protein assay) was separated by SDS–PAGE in parallel lanes for subsequent immunoblotting to monitor enrichment of PSD proteins with increasing purification of PSDs. Panels have been rearranged for clarity. (B) Immunoblot film signals were quantified by densitometry (see Materials and methods). For each individual protein, the signal was normalized against the WT signal from the same experiment. Dark (WT) and light (D36) bars reflect means±s.e.m. from three independent experiments. Five different WT and five different D36 mice were used for each experiment. S.e.m. values for WT signals were obtained by comparison with the WT signal average from all three experiments for that protein. PSD-95, GluR1, and AKAP150 levels are similar in WT and D36 PSD fractions, but RIIα and RIIβ are strongly reduced in D36 (asterisk: P<0.01; t-test).

Figure 3

Activity-induced but not basal phosphorylation of GluR1 S845 is reduced in D36 mice. (A, B) Forebrains from 1, 4, and 8-week-old WT and D36 mice were solubilized with 1% deoxycholate in the presence of phosphatase inhibitors and cleared by ultracentrifugation before immunoprecipitation of GluR1, and immunoblotting with a S845 phosphospecific antibody, and after stripping with a general anti-GluR1 antibody (A, top and bottom panels, respectively). Immunosignals were quantified by densitometry (B). All signals were normalized within one experiment to the 4-week WT value. Bars reflect means±s.e.m. from three independent experiments with three different WT (dark bars) and three different D36 mice (light bars). S.e.m. for 4-week WT signals was obtained by comparison of the individual signals with the 4-week WT signal average from all experiments. Comparable amounts of S845-phosphorylated and total GluR1 were present in WT versus D36 mice at all age groups. (C, D) To chemically induce LTP, acute forebrain slices from WT and D36 mice were treated with forskolin followed by 5 min incubation with 30 mM K⁺ and 0 Mg²⁺. For control, slices were treated for 15 min with vehicle (0.01% DMSO) followed by 5 min incubation with standard ACSF. Immunoprecipitation, immunoblotting (C), and densitometry were as above (D). All signals were normalized within one experiment to the WT control value. Bars reflect means±s.e.m. from three independent experiments with three different WT (dark bars) and three different D36 mice (light bars). Induction of S845 phosphorylation by forskolin/high K⁺ was significantly blunted in D36 mice (^*P<0.05 compared with WT treated with forskolin; t-test).

Figure 4

LTP is affected in 7–12 but not 4- to 5-week-old D36 mice. LTP was induced by a single tetanus (1 s/100 Hz; arrows in B) in CA1 of acute hippocampal slices from D36 and littermate-matched WT control mice. (A) Example fEPSP recordings before (dashed lines) and 60 min after LTP induction (solid lines) from mice 4–5 (A1) and 7–12 weeks old (A2). (B) Averages of the complete time courses are shown from mice 4–5 (B1) and 7–12 weeks old (B2). At 4–5 weeks LTP is comparable in hippocampal slices from D36 and control mice (145±4% s.e.m. versus 152±11% at 55–60 min, open and closed circles, respectively). However, at 7–12 weeks LTP is nearly absent in D36 mice (116±3%, open circles; control mice, 149±9%, closed circles; this difference is highly significant: P<0.001, t-test).

Figure 5

Basal synaptic transmission is normal in 4–5 and 7- to 12-week-old D36 mice. (A) Paired-pulse facilitation is virtually identical for D36 and control mice within both age groups over the whole range of interstimulus intervals. (B) Input–output curves with the postsynaptic response (initial slope of fEPSP) plotted as a function of the presynaptic fiber volley amplitude are indistinguishable between D36 and control mice within both age groups. (C) Examples of AMPAR mEPSC recordings from WT (C1) and D36 mice (C2). (D) Cumulative fraction (D1), histogram distribution (D1, insert), and overall average (D2) of mEPSC peak amplitudes (and frequencies; not illustrated) are virtually identical in 8-week-old WT versus D36 mice. (E1) Field EPSPs were recorded in the presence of CNQX (10 μM) to block AMPAR and in the absence of Mg²⁺ to eliminate NMDAR block by Mg²⁺. Input–output curves with the postsynaptic NMDAR response (initial slope of fEPSP) plotted as a function of the presynaptic fiber volley amplitude are indistinguishable between D36 and control mice at the age of 8 weeks. (E2) Ratio of NMDAR- to AMPAR-mediated EPSCs in D36 mice (0.61±0.08) is not distinguishable from WT (0.57±0.06) at the age of 8 weeks (right panel). Left panel shows recorded traces at +40 mV; black traces are the evoked EPSCs recorded under control condition; dark gray traces are the AMPAR-evoked EPSCs recorded in the presence of 100 μM D-APV; light gray traces are the NMDAR-evoked EPSCs computed by subtraction of AMPAR evoked EPSCs (dark gray trace) from the control-evoked EPSCs (black trace).

Figure 6

PKA is necessary for LTP in 8 but not 4-week-old WT C57BL/6 mice. (A) Inhibition of PKA with H-89 (20 μM in DMSO) does not affect LTP in hippocampal slices from 4-week-old mice (150±7% s.e.m., at 55–60 min) compared with vehicle control (0.025% DMSO; 148%±4%). (B) In-8-week-old mice, LTP is blocked by H-89 (107±7% compared with 150±11% for vehicle control at 55–60 min: P<0.01; t-test). (C, D) Inhibition of PKA with KT5720 (1 μM in DMSO) does not affect LTP in 4-week-old mice (C; 151±9% at 55–60 min) compared with vehicle control (0.1% DMSO; 155±10%), but inhibits LTP at 8 weeks (D; 110±4 versus 155±10% for vehicle control; P<0.004; t-test). (E, F) Bath application of H89 or KT5720 does not affect baseline fEPSP responses at 4 or 8 weeks over extended periods. (G–I) Field EPSPs were recorded in the presence of CNQX (10 μM) to block AMPARs and in the absence of Mg²⁺ for full NMDARs currents. Increasing stimulus strength increased fEPSPs to the point were population spikes started to appear (upward deflections in I at highest stimulus intensities; shown are fEPSP responses of increasing amplitude with increasing stimulus strength). Input–output curves were not different for any stimulus strength before versus after application of H89 or KT5720 at 4 (G) and 8 weeks (H). (J) At 8 weeks LTP is also blocked by the highly specific membrane-permeable PKA-inhibitory peptide 11R-PKI (Matsushita et al, 2001) (10 μM; 107±3% compared with 148±5% for the protease inhibitor only control: P<0.0001). Arrows indicate tetanus (1 s/100 Hz); black bars show times of drug applications.

Figure 7

GluR2-lacking AMPARs are necessary for LTP in 8 but not 4-week-old WT C57BL/6 mice. (A) Inhibition of GluR2-lacking AMPARs with PhTx-433 (2.5 μM) does not affect LTP in hippocampal slices from 4-week-old mice (140±7% s.e.m., at 55–60 min) compared with control (150±5%). PhTx-433 has also no effect on baseline responses (107±2%). (B) In slices from 8-week-old mice, PhTx-433 inhibits LTP nearly completely (113±5% compared with 146±10% for control at 55–60 min; P<0.02; t-test) but has no effect on baseline (105±9%). (C) Inhibition of GluR2-lacking AMPARs with NASPM (20 μM) does not affect LTP in hippocampal slices from 4-week-old mice (139±5% s.e.m., at 55–60 min, when NASPM is applied during stimulation; 138±6%, when NASPM is applied after stimulation) compared with control (142±4%). NASPM has also no effect on baseline (110±2%). (D) In slices from 8-week-old mice, LTP is inhibited by NASPM (111±7% when NASPM is applied during stimulation; 120±7% when NASPM is applied after stimulation; t-test: P<0.003 and P<0.02, respectively, compared with 150±7% for control at 55–60 min). NASPM has no effect on baseline (106±4%). Arrows indicate tetanus (1 s/100 Hz); bars show times of drug applications starting before (black) or immediately after (open) tetanus.

Figure 8

GluR2-lacking AMPAR are necessary for LTP in 2 but not 3-week-old WT C57BL/6 mice. LTP was induced by two tetani (1 s/100 Hz) 20 s apart in the CA1 area of acute hippocampal slices. (A) In slices from 12- to 14-day-old mice, PhTx-433 (2.5 μM) and NASPM (20 μM) inhibit LTP (111±4% s.e.m. and 119±3%, respectively, compared with 151±4% for control at 55–60 min; in both cases, P<0.00005; t-test). (B) PhTx-433 and NASPM do not affect LTP in hippocampal slices from 20- to 22-day-old mice (143±3 and 151±6%, respectively) compared with control (144±6%).