Phosphorylation of Ser¹⁹²⁸ mediates the enhanced activity of the L-type Ca²⁺ channel Ca_v1.2 by the β₂-adrenergic receptor in neurons

Abstract

The L-type Ca²⁺ channel Ca_v1.2 controls multiple functions throughout the body including heart rate and neuronal excitability. It is a key mediator of fight-or-flight stress responses triggered by a signaling pathway involving β-adrenergic receptors (βARs), cyclic adenosine monophosphate (cAMP), and protein kinase A (PKA). PKA readily phosphorylates Ser¹⁹²⁸ in Ca_v1.2 in vitro and in vivo, including in rodents and humans. However, S1928A knock-in (KI) mice have normal PKA-mediated L-type channel regulation in the heart, indicating that Ser¹⁹²⁸ is not required for regulation of cardiac Ca_v1.2 by PKA in this tissue. We report that augmentation of L-type currents by PKA in neurons was absent in S1928A KI mice. Furthermore, S1928A KI mice failed to induce long-term potentiation in response to prolonged theta-tetanus (PTT-LTP), a form of synaptic plasticity that requires Ca_v1.2 and enhancement of its activity by the β₂-adrenergic receptor (β₂AR)–cAMP–PKA cascade. Thus, there is an unexpected dichotomy in the control of Ca_v1.2 by PKA in cardiomyocytes and hippocampal neurons.

INTRODUCTION

Ca_v1.2 is the most abundant L-type Ca²⁺ channel in mammalian brain, heart, and vascular smooth muscle (1–4) regulating neuronal excitability, cardiac contractility, and vascular tone. Mutations in Ca_v1.2 affect many organs, as illustrated by the cardiac arrhythmias, autistic-like behavior, and developmental abnormalities in Timothy syndrome patients, who have gain-of-function mutations in Ca_v1.2 (5).

In the brain, L-type Ca²⁺ currents control neuronal excitability (6, 7), gene expression (8–11), and long-term potentiation (LTP) of synaptic strength. LTP refers to the persistent increase in synaptic strength that a synapse undergoes when it experiences a period of high-frequency activation (12–14). LTP is a key cellular correlate of learning and memory. In particular, L-type Ca²⁺ currents conducted by the voltage-gated calcium channel Ca_v1.2 are important for LTP induced by the widely used 1-s-long, 200-Hz tetanus (15, 16) or by a 90- to 180-s-long, 5- to 10-Hz tetanus mimicking the endogenous ~7-Hz theta rhythm, which we call prolonged theta-tetanus LTP (PTT-LTP) (17). Ca²⁺ influx through Ca_v1.2 is also important for long-term depression (LTD) induced by metabotropic glutamate receptor (mGluR) signaling (18). LTD is a persistent decrease in synaptic strength induced by a prolonged stimulation of a synapse with a frequency in the range of 1 Hz (18).

Norepinephrine is important for wakefulness, behavioral acuity, and various forms of learning, especially in novel or emotionally charged situations (19–23). Norepinephrine activates adenylyl cyclase (AC)–cyclic adenosine monophosphate (cAMP)–protein kinase A (PKA) signaling by the β₁- and β₂-adrenergic receptors (β₁AR and β₂AR) (24). This signaling cascade stimulates Ca²⁺ influx through Ca_v1.2 into neurons (25–28) and cardiomyocytes (29, 30).

Vertebrate Ca_v1.2 is composed of the pore-forming α₁1.2 subunit, a so-called α₂-δ subunit, and a β subunit (2, 31). α₁1.2 consists of a cytosolic N terminus, the four homologous domains I to IV, each with six transmembrane segments, and the long C terminus of ~660 residues. Injection of active PKA into cells exogenously expressing α₁1.2 without the other subunits increases channel activity (32), suggesting that PKA acts directly on α₁1.2. In purified preparations of Ca_v1.2 channels, Ser¹⁹²⁸ in the distal C-terminal domain is the most readily detectable site of PKA phosphorylation (33–36). βAR stimulation with isoprote-renol (ISO) results in a pronounced increase in Ser¹⁹²⁸ phosphorylation in isolated cardiomyocytes (37) and in the brain in vivo (38) [see also (39, 40)]. However, functional studies in transfected nonmuscle cells indicate that phosphorylation of Ser¹⁷⁰⁰ in the proximal C-terminal domain increases Ca_v1.2 channel activity, whereas phosphorylation of Ser¹⁹²⁸ in the distal C-terminal domain does not (41). Moreover, studies of Ca_v1.2 regulation in virally transfected cardiac myocytes and in ventricular myocytes dissociated from knock-in (KI) mice in which Ser¹⁹²⁸ has been mutated to alanine (S1928A KI mice) show that βAR stimulation of L-type currents is normal when Ser¹⁹²⁸ is mutated to alanine (42, 43).

In the brain and heart, the β₂AR forms a signaling complex with Ca_v1.2 that also contains AC, the heterotrimeric guanine nucleotide–binding protein (G protein) G_s, an A-kinase anchoring protein (AKAP), and PKA (10, 25, 28, 29, 35, 38, 41, 44–47). This complex allows highly localized and thereby effective and specific signaling from the β₂AR to Ca_v1.2 in neurons (25) and cardiomyocytes (29, 48) [for a review, see (24)]. The β₂AR binds to α₁1.2 residues 1923 to 1942, and this interaction is strictly required for receptor-mediated regulation of Ca_v1.2 (17). Furthermore, stimulating the β₂AR leads to its temporary displacement from Ca_v1.2, creating a refractory period of ~5 min, during which neither phosphorylation nor channel activity of Ca_v1.2 can be augmented again by restimulation of the β₂AR (17). When we tested whether such a refractory period for sequential β₂AR stimulation would be observed in S1928A KI mice, we discovered that βAR stimulation did not augment neuronal L-type currents in these mice. In further contrast to cardiomyocytes, we found that, in neurons, L-type currents were exclusively augmented by β₂AR but not by β₁AR stimulation. Finally, we demonstrated that this Ser¹⁹²⁸-dependent regulation of L-type Ca²⁺ currents was a key requirement for the induction of stable PTT-LTP by theta rhythm stimulation in the presence of a βAR agonist. Collectively, our results demonstrated a bifurcation of PKA signaling to Ca_v1.2 channels and define a new phosphorylation site that is required for LTP. Furthermore, they revealed that phosphorylation of a single site, Ser¹⁹²⁸, not only uncouples Ca_v1.2 regulation from β₂AR signaling upon repeated stimulation but also mediates the channel-stimulating response to β₂AR signaling in hippocampal neurons.

RESULTS

Increase in L-type Ca²⁺ currents upon βAR stimulation depends on Ser¹⁹²⁸ in neurons

In neurons, phosphorylation of Ser¹⁹²⁸ upon βAR stimulation displaces the β₂AR from Ca_v1.2 for ~5 min. A second application of the βAR agonist ISO only increases L-type currents in neurons if added more than 5 min after the first application has ended (17). In neuronal cultures from the hippocampal CA1, CA2, and CA3 regions, augmentation of L-type currents by ISO is much more apparent in cell-attached recordings, which typically yield increases of >200% (17, 25, 27), than in whole-cell recordings, which only show increases of 10 to 20% (27, 28). The weak ISO effect in whole-cell recordings from neurons is much smaller than in whole-cell recordings from cardiomyocytes (43). To functionally test whether Ser¹⁹²⁸ phosphorylation mediates this refractory period of L-type current augmentation, we recorded single-channel L-type currents from hippocampal neurons cultured from wild-type and S1928A KI mice by cell-attached patch-clamping (Fig. 1A), using conditions as previously described (17, 25).

Fig. 1. βAR augmentation of L-type currents in neurons requires Ser¹⁹²⁸.

(A) Representative single-channel recordings from hippocampal neurons from wild-type and litter-matched S1928A KI mice at 7 to 14 days in vitro upon depolarization from −80 to 0 mV without (left, black traces) and with 1 μM ISO (right, red traces) in the patch pipette. Arrows (←) indicate the zero-current level (that is, the closed channel). (B) Summary plot of NPo recorded from neurons from wild-type and S1928A KI mice with vehicle (H₂O) or 1 μM ISO added to the patch pipette solution. For statistical analysis, NPo was determined for each recording and pooled under each condition for comparison (*P < 0.05, unpaired t test for experiments in wild-type cells and Mann-Whitney test for experiments in S1928A cells). Mean NPo values are 0.12 ± 0.03 for wild-type control (n = 10), 0.53 ± 0.06 for wild-type ISO (n = 12), 0.12 ± 0.03 for S1928A control (n = 11), and 0.08 ± 0.04 for S1928A ISO (n = 11) (n, number of cell-attached patch recordings pooled across three to five independent culture experiments).

We determined open probability (Po) for all L-type channels within each patch (NPo) during depolarizing pulses from −80 to 0 mV with ω-conotoxins GVIA and MVIIC in the patch pipette to blocked non–L-type Ca²⁺ currents (Fig. 1B) (17). In neurons from wild-type mice, the addition of ISO to the patch pipette significantly increased NPo (Fig. 1, A and B). This increase was also evident in neurons prepared from only the hippocampal CA1 region, avoiding the CA2 and CA3 regions, from wild-type mice (fig. S1). In neurons from S1928A KI mice, however, the NPo value under control conditions was not significantly changed by ISO application (Fig. 1). Given that cardiac L-type currents measured in S1928A KI mice were significantly increased by βAR signaling (43), this result in neurons was unexpected. Thus, we confirmed under our experimental conditions that there was no difference in basal L-type currents between cardiomyocytes from wild-type and S1928A KI mice (fig. S2, A and B), and potentiation by ISO was equally strong for both genotypes (fig. S2, C to G).

Signaling by the β₂AR but not the β₁AR augments L-type currents in neurons

In cardiomyocytes, L-type currents are mainly enhanced by β₁ARs with a much smaller contribution by β₂ARs (29, 37, 48–52). To test the role of these two receptors in regulating neuronal Ca_v1.2, we applied ISO (a nonselective βAR agonist) during cell-attached recordings in wild-type rat hippocampal neurons in the absence or presence of CGP20712 (a β₁AR-specific antagonist) or ICI118551 (a β₂AR-specific antagonist) (53). Blocking β₂AR with ICI118551 completely blocked the ISO-induced increase in NPo (Fig. 2, A and B, and fig. S1), whereas blocking β₁AR CGP20712 had no effect (Fig. 2, A and B).

Fig. 2. β₂AR but not β₁AR signaling augments L-type currents in neurons.

(A) Representative single-channel recordings from wild-type (WT) rat hippocampal neurons. The patch pipette solution contained vehicle (H₂O; control; −) or 1 μM ISO (+; red traces) plus 100 nM CGP20712 (+CGP) or 100 nM ICI118551 (+ICI). Arrows throughout the figure (←) indicate the zero-current level (closed channel). (B) Summary plot for data in (A). For statistical analysis, the NPo value was determined for each recording and pooled under each condition for comparison (*P < 0.05; unpaired t test was used to compare statistical significance in wild-type cells under control conditions with those treated with CGP. Mann-Whitney test was used in cells treated with ICI). Average NPo values are 0.10 ± 0.03 for wild type – ISO (n = 10), 0.39 ± 0.06 for wild type + ISO (n = 12), 0.11 ± 0.04 for CGP – ISO (n = 9), 0.48 ± 0.05 for CGP + ISO (n = 11), 0.09 ± 0.03 for ICI – ISO (n = 12), and 0.10 ± 0.03 for ICI + ISO (n = 15) (n, number of cell-attached patch recordings from three independent culture experiments).

Signaling by the β₂AR but not the β₁AR promotes phosphorylation of α₁1.2 subunits in hippocampal neurons

The importance of β₂AR for the regulation of Ca_v1.2 in neurons was further confirmed by biochemical analysis (Fig. 3, A and B, top). Treatment of acutely isolated murine forebrain slices with ISO for 5 min significantly increased phosphorylation of Ser¹⁹²⁸ and also Ser¹⁷⁰⁰. These effects were completely blocked by ICI118551 but unaffected by CGP20712. The increase in ISO-induced Ser¹⁹²⁸ phosphorylation and its significant inhibition by ICI118551 were also observed when only the hippocampal CA1 region was used for such analysis (fig. S3). These results indicate that, in neurons in the forebrain and hippocampus, and specifically in those from the CA1 region, βAR signaling of Ca_v1.2 occurs through the β₂AR but not through the β₁AR.

Fig. 3. β₂AR controls phosphorylation of Ca_v1.2 and GluA1, and β₁AR controls phosphorylation of GluN2B.

(A) Forebrain slices from wild-type mice of either sex were treated with ASCF containing vehicle (H₂O; Ctr) or 1 μM ISO plus 100 nM ICI118551 (+ICI) or 100 nM CGP20712 (+CGP), as indicated, for 5 min before solubilization, ultracentrifugation, simultaneous immunoprecipitation (IP) of α₁1.2 plus GluA1, and subsequent immunoprecipitation of GluN2B from supernatants resulting from the α₁1.2/GluA1 immunoprecipitation. Samples were analyzed by sequential immunoblotting for phosphorylated Ser¹⁹²⁸ (pSer¹⁹²⁸), pSer¹⁷⁰⁰, and α₁1.2 (top panels) and for pSer⁸⁴⁵ and GluA1 (middle panels) from the same blots and sequential immunoblot-ting for pSer¹¹⁶⁶ and GluN2B of a second blot using corresponding regions of the blots. (B) For quantification of α₁1.2 phosphorylation, pSer¹⁹²⁸ and pSer¹⁷⁰⁰ signals were nor-malized to total α₁1.2 [*P < 0.05 versus control, analysis of variance (ANOVA) with Bonferroni correction for multiple testing between all experimental conditions; n, number of samples analyzed across three independent experiments]. For quantification of GluA1 phosphorylation, pSer⁸⁴⁵ signals were normalized to GluA1 (*P < 0.05, ANOVA with Bonferroni’s test; n, number of samples analyzed across three independent experiments). For quantification of GluN2B phosphorylation, pSer¹¹⁶⁶ signals were normalized to GluN2B (*P < 0.05, ANOVA with Bonferroni correction for multiple testing; n, number of samples analyzed across three independent experiments).

To confirm the specificity of the antagonists for β₂AR and β₁AR and that both receptors are present and active in the neurons, we also analyzed the effect of these drugs on ISO-induced phosphorylation of AMPA-type glutamate receptor (AMPAR) subunits and N-methyl-D-aspartate (NMDA)–type glutamate receptor (NMDAR) subunits. The AMPAR subunit GluA1 is phosphorylated on Ser⁸⁴⁵ through a β₂AR-dependent pathway (17, 53–55), and the NMDAR subunit GluN2B is phosphorylated on Ser¹¹⁶⁶ through a β₁AR-dependent pathway (Fig. 3).

The AMPAR forms a signaling complex analogous to Ca_v1.2 containing β₂AR, G_s, AC, AKAP, and PKA to enable highly localized signaling by cAMP (54, 55). In this complex, the β₂AR is linked to the AMPAR GluA1 subunit through the scaffold protein PSD-95 and AMPAR subunits known as transmembrane AMPAR regulatory proteins (TARPs) (54). Phosphorylation of GluA1 on Ser⁸⁴⁵ in this complex augments postsynaptic strength by increasing channel activity of the AMPAR (56) and its postsynaptic accumulation (24, 54, 57–67). ICI118551 completely blocked the ISO-induced increase in the phosphorylation of the GluA1 subunit of the AMPAR on Ser⁸⁴⁵, but there was minimal effect of CGP20712 (Fig. 3, A and B, bottom).

Phosphorylation of the NMDAR subunit GluN2B on Ser¹¹⁶⁶ by PKA selectively increases Ca²⁺ permeability of the NMDAR (68, 69). The ISO-induced increase in Ser¹¹⁶⁶ phosphorylation was not blocked by ICI118551, but it was completely eliminated by CGP20712 (Fig. 3, A and B, bottom). Thus, in contrast to Ca_v1.2 and GluA1, these data showed that NMDAR phosphorylation is mediated by the β₁AR and not by the β₂AR. These results could reflect differential β-adrenergic signaling in individual neurons, or the differential effects of β₁AR versus β₂AR signaling could occur in different subcellular regions or different neuronal subtypes, because our forebrain slices consist of a mixture of cells. Furthermore, these results demonstrated that CGP20172 effectively blocked the β₁AR. Therefore, its lack of effect on Ca_v1.2 and GluA1 phosphorylation reflects the requirement for β₂AR activation and not the lack of efficacy of the drug treatment (53).

Induction of PTT-LTP requires channel activity of Ca_v1.2

Extended stimulation of the Schaffer collateral pathway, which originates in the hippocampal CA3 region and projects to the CA1 region, at 5–10 Hz induces PTT-LTP but only if β-adrenergic stimulation occurs at the same time (22, 53, 70, 71). This stimulation pattern is especially relevant because it mimics the naturally occurring theta rhythm (~7 Hz) (72). During the first 5 min of stimulation, which include 3 min for delivery of the 5-Hz tetanus, synaptic transmission shows an initial depression before it recovers from what might be desensitization of the postsynaptic response, and the potentiation develops during the subsequent 5 min (22, 53, 70).

The complete inhibition of PTT-LTP by ICI118551 and not by CGP20712 (53) indicates that ISO acts through β₂AR but not through β₁AR during PTT-LTP. Because β₂AR stimulation prominently augments Ca²⁺ influx through Ca_v1.2 at postsynaptic sites (73), we tested whether Ca_v1.2 and its enhancement by the β₂AR are required for PTT-LTP. We recently found that PTT-LTP is absent in conditional forebrain Ca_v1.2 knockout (KO) mice (17). Ca_v1.2 KO mice could have altered synaptic signaling due to changes in gene expression rather than electrophysiological changes that acutely contribute to PTT-LTP. We used acute pharmacological inhibition of L-type channels and NMDARs, both of which are main conduits for Ca²⁺ influx during many forms of synaptic plasticity (12–14), to evaluate their respective contributions. Whereas the competitive glutamate site antagonist aminophosphonovaleric acid (APV) significantly reduced PTT-LTP, the pore blocker MK801 had a more modest, nonsignificant effect (Fig. 4, A and B). Standard LTP induced by a single 100-Hz tetanus, which strictly requires NMDARs, was fully abrogated by either drug, indicating that both drugs effectively blocked the NMDAR in these experiments (Fig. 4, C and D). The L-type Ca²⁺ channel blocker isradipine significantly reduced PTT-LTP to a degree that appeared to be larger than that by APV or MK801. Although these differences in PTT-LTP did not reach statistical significance (Fig. 4, A and B; P = 0.0806 for ISO + isradipine versus ISO + APV and P = 0.0591 for ISO + isradipine versus ISO + MK801, Mann-Whitney test), they collectively suggested that Ca²⁺ influx through L-type channels plays a more prominent role in PTT-LTP than Ca²⁺ influx through NMDARs under our experimental conditions.

Fig. 4. L-type Ca²⁺ channels mediate PTT-LTP.

Graphs show representative fEPSP initial slopes recorded from hippocampal CA1 and amalgamated data before [basal (Bsl)] and after either a 3-min, 5-Hz tetanus (A and B) or two 1-s, 100-Hz tetani (C and D). Arrowheads mark onset of tetani, gray bars indicate perfusion with 1 μM ISO, and black bars indicate perfusion with 50 μM DL-APV, 100 μM MK801, or 10 μM isradipine (ISRA), as indicated. Insets on top show sample traces immediately before (left) and ~30 min after (right) tetani. Degree of potentiation was determined 30 min after tetanus and compared to baseline before tetanus. PTT-LTP was 159.4 ± 7.4% of baseline for control conditions, 127.6 ± 5.3% for APV, 142.4 ± 8.7% for MK801, and 107.0 ± 4.1% for isradipine. LTP induced by two 100-Hz tetani was 138.9 ± 3.0% for control conditions, 108.6 ± 10.5% for APV, and 105.2 ± 4.0% for MK801 (*P < 0.05; Kruskal-Wallis with Dunn’s test; n, number of slices from three to six mice of either sex used for n number of recordings).

Isradipine, which is also known as PN200-110 when tritiated, is selective for L-type Ca²⁺ channels but binds multiple members of this class of channels, not only Ca_v1.2 (1). Hippocampal neurons also contain the L-type Ca²⁺ channel Ca_v1.3, albeit at a much lower amount than Ca_v1.2 (1, 3). Therefore, the effect of isradipine may have resulted from blocking Ca_v1.3 rather than Ca_v1.2 or by blocking both channels. Mutating Thr¹⁰⁶⁶ to Tyr in the α₁1.2 subunit makes Ca_v1.2 insensitive to dihydropyridines such as isradipine (3). We used T1066Y KI mice (3) to scrutinize whether isradipine inhibited PTT-LTP by blocking Ca_v1.2 rather than Ca_v1.3. Isradipine blocked PTT-LTP in litter-matched wild-type mice but not in homozygous dihydropyridine-insensitive T1066Y KI mice (Fig. 5). These results indicated that induction of PTT-LTP depends on Ca_v1.2 channel activity.

Fig. 5. PTT-LTP requires Ca_v1.2.

Graphs show fEPSP initial slopes recorded from hippocampal slices from litter-matched wild-type mice and isradipine-insensitive (DHP) T1066Y KI mice (A) and amalgamated data (B). Arrowheads mark onset of the 3-min, 5-Hz tetanus, gray bars indicate perfusion with 1 μM ISO, and black bars indicate perfusion with 10 μM isradipine. Insets on top show sample traces immediately before (left) and ~30 min after (right) tetani. Degree of potentiation was determined 30 min after tetanus and compared to baseline before tetanus. Average fEPSP was 114.2 ± 6.8% of baseline in slices from wild-type mice and 160.4 ± 9.2% of baseline in slices from DHP-insensitive, T1066Y KI mice (*P < 0.05; Mann-Whitney test; n, number of slices for the same number of recordings; wild-type mice: three animals; isradipine-insensitive mice: six animals of either sex).

Phosphorylation of Ca_v1.2 on Ser¹⁹²⁸ is required for PTT-LTP

Given that the β₂AR and Ca_v1.2 are required for PTT-LTP and that Ser¹⁹²⁸ is needed for augmenting Ca_v1.2 activity, we hypothesized that PTT-LTP depends on phosphorylation of Ser¹⁹²⁸ by PKA downstream of the β₂AR. PTT-LTP was absent 30 to 60 min after a theta burst stimulus in slices from S1928A KI mice, whereas it was readily induced in litter-matched control wild-type mouse slices (Fig. 6). In hippocampal slices from mice of each genotype, after the typical initial reversal of the depression of synaptic transmission seen during the theta tetanus, synaptic strength returned to the baseline level seen before PTT-LTP induction and then increased above the baseline up to 15 min after stimulation (Fig. 6A). However, after 15 min, synaptic strength declined to below control values for S1928A mice, whereas it remained increased for at least 60 min in controls (Fig. 6B).

Fig. 6. PTT-LTP requires phosphorylation of Ca_v1.2 on Ser¹⁹²⁸.

Graphs show fEPSP initial slopes recorded from hippocampal slices from litter-matched S1928A KI mice and wild-type mice (A) and amalgamated data (B). Arrowheads mark onset of the 5 Hz, 3 min tetanus, and gray bars indicate perfusion with 1 μM ISO. Insets on top show sample traces immediately before (left) and ~30 min after (right) tetani. Degree of potentiation was determined 30 min after tetanus and compared to baseline before tetanus. Average fEPSP was 151.3 ± 8.6% of baseline in slices from wild-type mice (*P < 0.05 baseline versus tetanized; Mann-Whitney test) and 85.8 ± 9.3% of baseline in slices from S1928A KI mice (*P < 0.05; Mann-Whitney test; n, number of slices for the same number of recordings; wild-type mice: six animals; S1928A KI mice: three animals of either sex)..

Input-output curves for increased stimulus strength versus field excitatory postsynaptic potential (fEPSP) response strength, an indication of postsynaptic glutamate receptor activity, and paired-pulse facilitation, an indication of presynaptic glutamate release function, did not show any differences between slices from S1928A KI mice and those from litter-matched wild-type control mice over a wide range of values (fig. S4). Accordingly, basal synaptic transmission including presynaptic glutamate release and postsynaptic glutamate receptor activity appear unaltered in S1928A KI mice.

Both the PKA-targeted site Ser¹⁷⁰⁰ and the nearby residue Thr¹⁷⁰⁴ are implicated in the regulation of Ca_v1.2 (41, 74–76). We analyzed PTT-LTP in S1700A/T1704A double KI mice (“STAA mice”). There was no difference in PTT-LTP between slices from STAA mice and litter-matched wild-type mice (fig. S5). In addition, input-output curves for fEPSP responses and paired-pulse facilitation were comparable for these two genotypes (fig. S6). We concluded that induction of PTT-LTP requires increased Ca_v1.2 activity mediated by β₂AR-PKA signaling through Ser¹⁹²⁸ phosphorylation but not through Ser¹⁷⁰⁰ or Thr¹⁷⁰⁴ phosphorylation.

DISCUSSION

Regulation of the L-type Ca²⁺ channel Ca_v1.2 by βAR-cAMP-PKA signaling plays a prominent role in mediating the fight-or-flight response of our heartbeat, as triggered by norepinephrine (29, 30, 77, 78). Signaling by norepinephrine through this pathway also augments arousal, behavioral acuity, and emotionally motivated learning (19–23). A prominent target of norepinephrine action in the brain is Ca_v1.2, which forms a complex with the β₂AR, G_s, AC, AKAPs, and PKA (24–26, 28, 29). In turn, Ca_v1.2 controls neuronal excitability through closely associated Ca²⁺-activated K⁺ channels (6, 7). Ca²⁺ influx through Ca_v1.2 is more effective in regulating gene expression through the cAMP response element–binding protein and nuclear factor of activated T cells than Ca²⁺ influx through other Ca²⁺ channels (8–11), and Ca_v1.2-induced Ca²⁺ influx is important for various forms of synaptic plasticity (15, 16, 18). Yet, the molecular mechanism by which PKA increases ion flux through Ca_v1.2 in the brain and heart remains incompletely defined. We identified Ser¹⁹²⁸ of Ca_v1.2 as an essential PKA-regulated site important for PTT-LTP. Furthermore, we found differential regulation of Ca_v1.2 by the PKA pathway in the brain and heart. Parallel work on the regulation of Ca_v1.2 by PKA in vascular smooth muscle cells (VSMCs) also shows that this regulation is completely abrogated in S1928A KI mice (79).

The findings in neurons and VSMCs are in striking contrast to cardiomyocytes, in which Ser¹⁹²⁸ in the distal C-terminal domain does not seem to be involved; rather, phosphorylation of Ser¹⁷⁰⁰ in the proximal C-terminal domain has been implicated in the increase in channel activity of cardiac Ca_v1.2 by βAR signaling (41, 74–76). In neurons, we found that the β₂AR-specific blocker ICI118551 completely blocked the ISO-induced increase in NPo of Ca_v1.2, whereas blocking β₁AR with CGP20712 had no effect. In contrast, cardiac Ca_v1.2 channels are primarily regulated by β₁AR signaling (29, 37, 48–52). Although β₂ARs are present in rodent and human heart, they do not participate in the increase in Ca_v1.2 channel activity that augments contractility in the flight-or-flight response. Accordingly, the βAR-mediated regulation of Ca_v1.2 in the heart and brain depends on different βAR subtypes and phosphorylation of different sites on Ca_v1.2.

β₂AR regulation of Ca_v1.2 channels in the heart is thought to occur mostly in caveolae and to be important for initiation of cardiac hypertrophy in response to hypertrophic signals (80). The β₁AR and β₂AR engage different downstream effectors in the G protein and arrestin pathways through biased signaling (81, 82), and these different signaling pathways may lead to biased changes in phosphorylation status of Ca_v1.2 channels in the heart and brain. More work is required to determine how these two signaling pathways can engage regulation of Ca_v1.2channels through phosphorylation of Ser¹⁹²⁸ versus Ser¹⁷⁰⁰ and Thr¹⁷⁰⁴. Some possible mechanisms for the different regulation of Ca_v1.2 in the brain and heart include differential subunit composition of the channel, differential proteolytic processing of the channel, and differential association of the channel with AKAPs in the brain and heart.

Four different genes encode α₂δ subunits (α₂δ₁ to α₂δ₄) and β subunits (β₁ to β₄), which further undergo differential splicing (83, 84). In cardiomyocytes, β₂ is functionally the most prominent β subunit in defining the biophysical properties of Ca_v1.2 (85). β_2A is phosphorylated by PKA on Ser⁴⁵⁹, Ser⁴⁷⁸, and Ser⁴⁷⁹, which are not present in the β₁, β₃, or β₄ subunit (86). Although there is evidence that Ser⁴⁷⁸ and Ser⁴⁷⁹ are not required for increased Ca_v1.2 activity in virally infected ventricular myocytes (42), this system does not fully reconstitute β-adrenergic regulation of L-type currents, leaving open the possibility that Ser⁴⁷⁸ and Ser⁴⁷⁹ contribute to the increase of L-type currents in the heart. In the brain, all four β subunits are broadly expressed (87–89). Thus, it is conceivable that PKA-regulated phosphorylation of β₂ compensates for the loss of Ser¹⁹²⁸ phosphorylation in S1928A KI mice in cardiomyocytes but not neurons. It is also possible that differential splicing of α₁1.2 plays a role, although all known splice variants are predicted to contain Ser¹⁹²⁸ (90, 91).

In cardiomyocytes, α₁1.2 is proteolytically processed, and this processing and subsequent noncovalent association of the cleaved C-terminus with the channel core are required for PKA-dependent regulation (41, 46). Reconstitution of regulation of a cardiac form of Ca_v1.2 by PKA in cultured nonmuscle cells requires formation of an autoinhibitory signaling complex composed of the core of the Ca_v1.2 channels with the noncovalently bound cleaved portion (41), which makes analysis of PKA-dependent regulation in this system challenging. Ca_v1.2 channels are not as extensively proteolytically processed in hippocampal neurons as in the heart, unless massive Ca²⁺ influx is induced by prolonged activation of NMDA receptors (33, 92). Therefore, differences in proteolytic processing and assembly of the Ca_v1.2 signaling complex in the heart and brain may contribute to the differences in mechanisms of PKA regulation.

Regulation of Ca_v1.2 channels in the heart and brain requires AKAPs to anchor PKA (10, 28, 35, 38, 41, 44–47). Differential regulation of Ca_v1.2 channels can be induced in reconstituted nonmuscle cells by coexpression of different AKAPs (93). Association of Ca_v1.2 channels with different sets of AKAPs in the brain and heart might contribute to differential channel regulation by PKA phosphorylation of distinct sites.

Another important finding was that PTT-LTP depends on Ca_v1.2 and its phosphorylation of Ser¹⁹²⁸ but not Ser¹⁷⁰⁰. PTT-LTP is thought to underlie contextual learning under difficult or stressful situations (22). PTT-LTP requires simultaneous stimulation by prolonged theta tetanus and β₂AR signaling (22, 53, 70). The β₂AR forms two distinct signaling complexes with postsynaptic ion channels, namely, AMPARs and Ca_v1.2, and both are now emerging as critical targets for β₂AR-PKA signaling during PTT-LTP (17, 53, 63).

The β₂AR binds to the scaffold protein PSD-95, which, in turn, binds to the AMPAR (54, 55). In addition, AKAP150, which is associated with GluA1 through SAP97 (94), recruits AC (65, 95), PKA, and the phosphatase PP2B to the AMPAR complex (24, 62, 63, 65, 96–98). The close proximity of all components of this cAMP cascade in this AMPAR complex and the analogous β₂AR-G_s-AC-PKA-Ca_v1.2 complex results in highly localized, selective, and potent augmentation of channel activity of AMPARs and Ca_v1.2 by cAMP (17, 25, 29, 47, 48, 54). Phosphorylation of Ca_v1.2 on Ser¹⁹²⁸ and of the AMPAR GluA1 subunit on Ser⁸⁴⁵, which is also required for PTT-LTP (53), might synergistically augment Ca²⁺ influx into postsynaptic sites through Ca_v1.2 in PTT-LTP (Fig. 7). During PTT-LTP, stimulation of the β₂AR results in phosphorylation of Ser⁸⁴⁵, which increases AMPAR Po (56, 97) and promotes postsynaptic accumulation of GluA1-containing AMPARs (54, 57, 58, 60). Consequently, Na⁺ influx increases, resulting in depolarization of postsynaptic sites during PTT-LTP. In addition, we found that phosphorylation of Ser¹⁹²⁸ increases the Po of Ca_v1.2. In other cell types, Ca_v1.2 phosphorylation at Ser¹⁹²⁸ also shifts the activation threshold for Ca_v1.2 toward more negative membrane potentials (fig. S2) (27, 32, 37, 43, 75, 76). This leftward shift means that less depolarization is required for channel opening. In addition, it further increases Ca²⁺ influx because the driving force for this Ca²⁺ influx is larger at these more negative potentials. The combined effect of Ser⁸⁴⁵ and Ser¹⁹²⁸ phosphorylation would thereby be stronger Ca²⁺ influx through Ca_v1.2.

Fig. 7. Model for enhancement of Ca²⁺ influx through Ca_v1.2 during PTT-LTP.

Under basal conditions, phosphorylation of Ser⁸⁴⁵ in the AMPAR GluA1 subunit and of Ser¹⁹²⁸ in the Ca_v1.2 α₁1.2 subunit and Ser¹⁹²⁸ is low. Stimulation of AMPAR- and Ca_v1.2-associated β₂ARs activates G_s-mediated AC signaling and thereby cAMP production and PKA. The resulting Ser⁸⁴⁵ phosphorylation increases Po of the AMPAR, which results in Na⁺ influx and depolarization during synaptic transmission induced by the theta stimulation. Ser¹⁹²⁸ phosphorylation increases Po of Ca_v1.2 and thereby Ca²⁺ influx through this channel. This increase in Ca²⁺ influx through Ca_v1.2 triggers the potentiation of synaptic strength in PTT-LTP via yet-to-be-defined signaling pathways.

The time course of PTP-LTP after the theta tetanus train in S1928A mice initially showed an increase in synaptic strength above aseline, but eventually, postsynaptic responses fell below those of wild-type controls (Fig. 6). It is possible that, in the absence of the ISO-induced increase in Ca_v1.2 activity through Ser¹⁹²⁸ phosphorylation, PKA phosphorylation of Ser⁸⁴⁵ in GluA1 in the AMPAR is sufficient to initially enhance synaptic strength. However, without augmented Ca²⁺ entry through the Ca_v1.2 channels, the neurons ultimately respond with synaptic depression. Such coincidence detection of depolarization through AMPAR and Ca²⁺ entry mediated by Ca_v1.2 channels could be a mechanism to ensure precise capture of incoming signals that lead to LTP.

In contrast to the S1928A KI mice, a decrease in synaptic transmission upon PTT-LTP induction was not seen in slices from S845A KI mice, although PTT-LTP was also impaired in S845A KI slices (53). Perhaps in the absence of increased Ca²⁺ influx through Ca_v1.2 in the S1928A slices, mechanisms that induce NMDAR-dependent LTD prevail, which requires Ser⁸⁴⁵ phosphorylation for temporary insertion of Ca²⁺-permeable AMPARs into the postsynaptic membrane (63). Another potential explanation is that, in S1928A KI mice, the constitutive absence of the increase in Ca_v1.2 activity upon βAR receptor signaling affects gene expression in a way that influences postsynaptic signaling during PTT-LTP. The absence of PTT-LTP in S1928A KI mice is consistent with an essential role of Ser¹⁹²⁸ phosphorylation in the increase in Ca_v1.2 channel activity by β₂AR signaling because PTT-LTP depends on both β₂AR and Ca_v1.2 (17, 53). In summary, we demonstrate that Ser¹⁹²⁸ phosphorylation of Ca_v1.2 is required for augmentation of Ca_v1.2 channel activity by β₂AR stimulation and that this mechanism constitutes a critical component of the molecular mechanism underlying stable PTT-LTP.

MATERIALS AND METHODS

Animals

All procedures involving animals followed the guidelines for the Care and Use of Laboratory Animals of the National Institutes of Health and had been approved by the Institutional Animal Care and Use committees at the University of California, Davis, and the University of Washington. S1928A KI mice were described in (43), and STAA mice were described in (75).

Reagents, peptides, and antibodies

ISO in the form of ISO bitartrate salt, ICI118551, CGP20712, and microcystin LR were from Sigma-Aldrich; protein A–coated beads were from Repligen; polyvinyldene fluoride (PVDF) membranes were from Millipore; horseradish peroxidase (HRP)–coupled protein A and enhanced chemiluminescence (ECL) reagents were from GE Healthcare; EGTA, EDTA, Tween 20, Triton X-100, and tris were from Thermo Fisher Scientific; and ω-conotoxins GVIA and MCVIIC were from Bachem. (S)-(−)-Bay K8644 was from Millipore, and isradipine was from Enzo Life Sciences. Other reagents were from the typical suppliers and of standard quality. Antibodies against α₁1.2 (FP1), pSer¹⁷⁰⁰, pSer¹⁹²⁸, GluA1, and pSer⁸⁴⁵ are exactly as detailed (17). Antibodies against GluN2B and pSer¹¹⁶⁶ are described in (69).

Preparation of brain slices and use for biochemical analysis

Eight- to 12-week-old mice with a genetic background of 50% C57BL/6 and 50% S129/Sv (S1928A KI mice) or mostly C57BL/6 (STAA, back-crossed at least four times with C57BL/6) were decapitated, and brains placed into ice-cold artificial cerebrospinal fluid (ACSF) (127 mM NaCl, 26 mM NaHCO₂, 1.2 mM KH₂PO₄, 1.9 mM KCl, 2.2mMCaCl₂,1mM MgSO₄, and 10 mM D-glucose; 290 to 300 mosmol/kg; saturated with 95% O₂ and 5% CO₂; final pH 7.3). About one-third of the rostral and caudal portions of the brain were trimmed off. Forebrain slices (350 μm thick) containing the hippocampus were prepared with a vibratome (Leica VT 1000A) and equilibrated in oxygenated ACSF for 1 hour at 30°C before transfer to incubation or recording chambers.

Biochemistry

Slices were equilibrated for 30 min at 32°C and treated with the various drugs for 5 min. Slices were processed with a glass Teflon homogenizer in a 10-fold excess (v/w) of immunoprecipitation buffer [1% Triton X-100, 150 mM NaCl, 10 mM EDTA, 10 mM EGTA, and 10 mM tris-HCl (pH 7.4)] (35) containing protease inhibitors [pepstatin A (1 μg/ml), leupeptin (10 μg/ml), aprotinin (20 μg/ml), and 200 nM phenylmethylsulfonyl fluoride, the latter being added immediately before homogenization and again at the start of the immunoprecipitation] and phosphatase inhibitors (2 μM microcystin LR, 1 mM p-nitrophenyl phosphate, 50 mM Na-pyrophosphate, and 50 mM NaF) (38, 99). Nonsolubilized material was removed by ultracentrifugation (250,000g for 30 min) before immunoprecipitation (4 hours at 4°C) with antibodies against α₁1.2 (FP1; 4 μg) plus GluA1 (1 μg) and GluN2B. The protein A beads were washed three times with immunoprecipitation buffer containing 0.05% Triton X-100 plus 0.5% glycerol. SDS–polyacrylamide gel electrophoresis and transfer onto PVDF membranes were followed by blocking [5% dry milk in tris-buffered saline (TBS)], incubation with primary antibodies (5% dry milk in TBS; 1 hour), three washes with 0.05% Tween 20 in TBS, incubation with HRP–protein A (1:10,000 in 5% dry milk in TBS; 1 hour), several washes for 2 hours, and detection of HRP with ECL or ECL plus chemiluminescence reagents. Multiple exposures of increasing length ensured that signals were in the linear range, as previously described (40, 99).

Cell-attached patch-clamp recording

Primary hippocampal neurons were cultured from Sprague-Dawley rats or wild-type and S1928A KI mice, which had a 50% 129/SV and 50% C57BL/6 genetic background, of either sex, as previously described (100–102). Cell-attached patch-clamp recordings were performed after 7 to 14 days in culture, as previously described (17, 25), on an Olympus IX70 inverted microscope at 22°C. Signals were recorded at 10 kHz and low-pass–filtered at 2 kHz with an Axopatch 200B amplifier and Digidata 1440 digitizer (all from Molecular Devices). Recording electrodes were pulled from borosilicate capillary glass (outer diameter, 0.86 mm) with a Flaming micropipette puller (model P-97, Sutter Instruments) and polished (polisher from World Precision Instruments; resistance, 3.5 to 6.5 megohms). The patch transmembrane potential was zeroed by keeping neurons in high K⁺ extracellular solution containing 145 mM KCl, 10 mM NaCl, and 10 mM Hepes [pH 7.4 (NaOH)] during the recordings. The pipette solution contained 20 mM tetraethylammonium chloride (TEA-Cl), 110 mM BaCl₂ (as charge carrier), and 10 mM Hepes [pH 7.3 (TEA-OH)] plus 1 μM ω-conotoxin GVIA and 1 μM ω-conotoxin MCVIIC to block N- and P/Q-type Ca²⁺ channels, respectively. (S)-(−)-Bay K8644 (500 nM) was added to the pipette solution to promote longer open times. This procedure is routinely applied to augment detection of L-type channels by single-channel recordings. Single-channel activity was recorded during an average of 50 2-s-long pulses from a holding potential of −80 to 0 mV every 5 s for each experimental condition. The single-channel event-detection algorithm of pClamp 10 was used to determine single-channel opening amplitudes and NPo. NPo values were combined for each condition and analyzed with GraphPad Prism 5.

fEPSP recording from brain slices

Recordings were from either sex, as previously described (17, 53). Slices were perfused with a flow rate of 2 ml/min at 30°C, with ACSF equilibrated with 95% O₂ and 5% CO₂ (final pH 7.3). Schaffer collaterals were stimulated every 15 s with a bipolar tungsten electrode, and resulting fEPSPs in CA1 were recorded with a glass electrode filled with ACSF. Signals were amplified with an Axopatch 2B amplifier, digitized with Digidata 1320A, and analyzed with Clampex 9 (Molecular Devices). Stimulus strength was titrated to define maximal response and input-output curves and adjusted to result in ~50% of maximal response. PTT-LTP was induced by a 3-min, 5-Hz tetanus. The average of fEPSP initial slopes from the 5 min preceding the tetanus was set to equal 100% baseline level. The PTT-LTP strength was defined as the average of fEPSP initial slopes obtained between 15 and 45 min after the tetanus. To determine paired-pulse facilitation, fEPSP initial slopes of two consecutive stimuli with the indicated interevent intervals were recorded.

Whole-cell patch-clamp recording from cardiomyocytes

Ventricular myocytes were isolated from mice of either sex, as previously described (Alliance for Cellular Signaling Procedure Protocol PP00000125) (103), maintained at 37°C, and aerated with 98% O₂ and 2% CO₂. Calcium currents (I_Ca) were recorded in whole-cell mode at room temperature from calcium-tolerant myocytes within 1 to 24 hours of isolation. The extracellular solution contained 137 mM NaCl, 1.8 mM CaCl₂, 25 mM CsCl, 0.5 mM MgCl₂, 10 mM Hepes, and 10 mM glucose (pH 7.4). Patch pipettes (1 to 1.5 megohms) were filled with 120 mM CsCl, 10 mM 1,2-bis(2-aminophenoxy)ethane-N,N,N',N'-tetraacetic acid, 1 mM EGTA, 1 mM MgCl₂, 1 mM Na₂GTP, 5 mM phosphocreatine, and 10 mM Hepes (pH 7.2). I_Ca was elicited by depolarizing steps. A prepulse from −80 to −40 mV was used to inactivate fast Na⁺ currents and then stepped to voltages between −40 and +60 mV (10-mV increments). I_Ca was recorded by an Axopatch 200B amplifier (Axon Instruments) and stored on a computer through an analog-digital converter (DIGIDATA 1332A; Axon Instruments). Protocols were controlled by pClamp 9 software (Axon Instruments). I_Ca was measured as the difference between the peak inward current and the current after the test pulse ended. After establishing a stable baseline, the effect of 1 μM ISO on I_Ca was examined.

Statistical analysis

Data are expressed as means ± SEM. Data were analyzed using GraphPad Prism software. Data were assessed for normality of distribution using the Shapiro-Wilk test. Statistical significance was then determined using appropriate paired or unpaired Student’s t test, ANOVA, or nonparametric tests. Statistical significance, denoted by asterisks (*) in figures, was considered at P < 0.05.