Functional uncoupling between Ca²⁺ release and afterhyperpolarization in mutant hippocampal neurons lacking junctophilins

Abstract

Junctional membrane complexes (JMCs) composed of the plasma membrane and endoplasmic/sarcoplasmic reticulum seem to be a structural platform for channel crosstalk. Junctophilins (JPs) contribute to JMC formation by spanning the sarcoplasmic reticulum membrane and binding with the plasma membrane in muscle cells. In this article, we report that mutant JP double-knockout (JP-DKO) mice lacking neural JP subtypes exhibited an irregular hindlimb reflex and impaired memory. Electrophysiological experiments indicated that the activation of small-conductance Ca²⁺-activated K⁺ channels responsible for afterhyperpolarization in hippocampal neurons requires endoplasmic reticulum Ca²⁺ release through ryanodine receptors, triggered by NMDA receptor-mediated Ca²⁺ influx. We propose that in JP-DKO neurons lacking afterhyperpolarization, the functional communications between NMDA receptors, ryanodine receptors, and small-conductance Ca²⁺-activated K⁺ channels are disconnected because of JMC disassembly. Moreover, JP-DKO neurons showed an impaired long-term potentiation and hyperactivation of Ca²⁺/calmodulin-dependent protein kinase II. Therefore, JPs seem to have an essential role in neural excitability fundamental to plasticity and integrated functions.

Functional communication between cell-surface and intracellular channels is an essential feature of excitable cells (1). During initiation of contraction in striated muscle cells, the activation of cell-surface dihydropyridine receptor (DHPRs) channels opens ryanodine receptors (RyRs) and triggers Ca²⁺ release from the sarcoplasmic reticulum via either the “Ca²⁺-induced Ca²⁺ release” or the “voltage-induced Ca²⁺ release” mechanism (2). The functional couplings between the channels take place in junctional membrane complexes (JMCs), designated as the “triad junction” in skeletal muscle, “diad” in cardiac muscle, and “peripheral coupling” in immature striated and smooth muscles (3, 4). Recent studies indicated that junctophilin (JP) subtypes, namely JP-1–JP-4, contribute to JMC formation in muscle cells (5, 6). In JP-1 knockout mice with perinatal lethality, mutant skeletal muscle shows deficiency of triad junctions and insufficient contraction probably caused by impaired communication between DHPRs and RyRs (7). In JP-2 knockout embryos showing cardiac arrest, mutant cardiac myocytes exhibit deficiency of peripheral couplings and arrhythmic Ca²⁺ signaling probably caused by functional uncoupling between DHPRs and RyRs (5). In the brain, both JP-3 and JP-4 are expressed in similar discrete neuronal sites and may collaboratively contribute to JMC formation (8, 9). However, the role of neural JP subtypes is largely unknown. Using knockout mice lacking both JP-3 and JP-4 (JP-DKO mice), we report their essential contributions to the tuning of excitability and plasticity in hippocampal pyramidal neurons.

Results

Generation of JP-DKO Mice Bearing Lethality.

JP-4 knockout mice were generated by using standard gene-targeting methods, and blot analyses confirmed that the mutation introduced is a null mutation in the gene (see Fig. 6, which is published as supporting information on the PNAS web site). The JP-4 knockout mice, showing no significant abnormalities in development, appearance and reproduction, were crossed with JP-3 knockout mice (9) to produce JP-DKO mice. Under normal housing conditions, JP-DKO mice showed severe growth retardation and lethality 3–4 weeks after birth. This lethality is caused by a feeding defect because JP-DKO mice were obviously rescued when the diet was switched from normal dry pellets to hydrated paste (Fig. 1A). This phenotype may be due to a defect in saliva secretion. The detection of normal pilocarpine-induced saliva secretion in mutant mice (K. Obara, M. Ishizuka, I. Saito, M.N., and H.T., unpublished observation) suggested that neural circuits controlling the salivary gland may be severely damaged in JP-DKO mice. Moreover, most of mature JP-DKO mice that survived (>90%) were infertile.

Fig. 1.

Lethality and irregular hindlimb reflex in JP-DKO mice. (A) Lethality at weaning period in JP-DKO mice. JP-DKO mice survived the embryonic and neonatal periods but died within 5 weeks after birth. JP-DKO mice were obviously rescued by paste-food feeding. (B) Abnormal foot-clasping reflex in JP-DKO mice. Control mice opened their hindlimbs when suspended by the tail. However, JP-DKO mice crossed their legs when hung upside-down.

When normal mice were suspended by the tail, they struggled and kicked, and their lower limbs remained opened. By contrast, mature JP-DKO mice always crossed their legs when hung upside-down (Fig. 1B). We never detected such an irregular response in JP-3(+/−) JP-4(+/−) (JP-DHE), JP-3 knockout, or JP-4 knockout mice. This aberrant behavior, called the “foot-clasping reflex” (10), has been reported in several mutant animals bearing neurological defects.

Normal Morphology in JP-DKO Brain.

In addition to having a regular brain size, mature JP-DKO mice retained normal layer organization and cell density in the cerebellum and cerebrum (see Fig. 7A, which is published as supporting information on the PNAS web site), and we found no abnormalities in basic histology and cytology in the mutant brain. Our electronmicroscopic observations also revealed no ultrastructural defects in the hippocampal and cerebellar regions from the JP-DKO brain. The normal expression of major endoplasmic reticulum (ER)-related proteins suggested the absence of ER stress responses in the JP-DKO forebrain (see Fig. 8A, which is published as supporting information on the PNAS web site). Together, these observations indicate that the JP-DKO brain has no major pathological defects.

Our previous studies demonstrated that JP subtypes contribute to the formation of JMCs that contain RyRs and have a gap size of ≈12 nm in cardiac and skeletal muscle cells (5–7). Because both JP-3 and JP-4 are abundantly expressed in hippocampal pyramidal cells (8), we examined JMCs in hippocampal neurons from JP-DKO mice. The soma and dendritic regions of control hippocampal neurons contained JMCs with various gap sizes, and no differences were noted in JP-DKO neurons (Fig. 7 B–D). In recent studies, various types of JMCs have been predicted in neurons; for example, Homer family proteins bind with both cell surface glutamate receptors and Ca²⁺ release channels on the ER to generate neural JMCs (11). It may be that the standard electron microscope analysis cannot distinguish JP-mediated JMCs from those with different molecular components in neurons.

Impaired Memory in JP-DKO Mice.

JP-DKO mice were examined in terms of behavioral tasks. At the beginning of the open-field task (Fig. 2A), JP-DKO mice showed a significant hypolocomotion activity compared with control mice, whereas no significant difference was found among genotypes after a habituation period. Therefore, JP-DKO mice may have an impaired exploratory activity. Neither JP-3 nor -4 knockout mice showed abnormalities in the open-field task (data not shown). In the Y-maze test (Fig. 2B), low counts for arm entry confirmed the impaired exploratory behavior of JP-DKO mice. Poor alternation of the entry was also observed in JP-DKO mice, suggesting an impaired short-term spatial memory. Next, JP-DKO mice were examined in a multiple-trial passive avoidance test by using a cage composed of lighted and dark chambers (Fig. 2C). During training trials, a mouse was placed in the lighted chamber and received an electric shock when it entered the dark chamber to keep it constantly in the lighted chamber. After 3 days, step-through latency to the dark chamber was recorded in retention trials. JP-DKO mice showed a remarkably longer latency than control mice in the first training session, which is probably due to the impaired exploratory behavior, but retained normal performance in learning the task during acquisition trials. The retention trial showed that the latency of JP-DKO mice was remarkably shorter than those of controls, indicating an impaired long-term memory. Neither JP-3 nor -4 knockout mice showed abnormalities in this passive-avoidance test (data not shown). Because of their poor swimming ability, we could not analyze JP-DKO mice in the Morris water maze (F. Jia, K. Yanai, M.N., H.T., unpublished observation).

Fig. 2.

Impaired exploratory activity and memory in JP-DKO mice. (A) Abnormal performance of JP-DKO mice in the open-field test. Although similar baseline activities were observed among the genotypes, JP-DKO mice showed hypolocomotion at the beginning of the test. (B) Poor Y-maze performance in JP-DKO mice. The number of entries into the arms (Left) and the percentage of alternation behavior (Right) are indexes of exploratory activity and spatial working memory, respectively. (C) Impaired passive-avoidance memory in JP-DKO mice. JP-DKO mice showed a prolonged latency to enter the dark chamber in the first trial (Left), but no difference among the genotypes was observed in the acquisition trials for learning the task (Center). (Right) In the retention trial, JP-DKO mice showed a reduced latency. In the tests, 7- to 8-week-old mice were used, and the number of mice examined is shown in parentheses. The data represent the mean ± SEM, and statistical differences between JP-DKO and WT mice are marked with asterisks (∗, P < 0.05; ∗∗, P < 0.01 in Student’s t test), whereas no significant difference was observed between WT and JP-DHE mice.

Absence of AHP in JP-DKO Hippocampal Neurons.

To survey neural dysfunction in JP-DKO mice, we performed whole-cell current-clamp recording of CA1 pyramidal cells in hippocampal slices and examined action potentials evoked by direct current injection. A striking difference was found in action potential configuration between control and JP-DKO neurons (Fig. 3A and B). In control neurons, obvious afterhyperpolarization (AHP) was observed at the resting potentials of −40 to 60 mV, and its zero-current potential, designated as the V₀ value in this analysis, was −71 mV corresponding approximately to the K⁺ reversal potential. In accordance with a previous report (12), the observed AHP was completely abolished by apamin as a blocker of small-conductance Ca²⁺-dependent K⁺ (SK) channels, but not by iberiotoxin as an inhibitor of large-conductance Ca²⁺-dependent K⁺ channels. In JP-DKO neurons, apamin-sensitive AHP was completely absent at any resting potentials. Of the SK channel subtypes, both SK2 and SK3 are expressed in the hippocampus, and SK2 knockout mice lack CA1 AHP (13). Because JP-DKO mice retained the normal protein levels of SK2 and SK3 in the brain (Fig. 8A), our observations suggest that the activation of SK channels is impeded under physiological conditions in JP-DKO neurons.

Fig. 3.

Pharmacological characterization of diminished AHP in hippocampal JP-DKO neurons. (A) Whole-cell current-clamp recordings were made from CA1 neurons in hippocampal slices, and action potentials were evoked by depolarizing currents (25-ms duration) before and after apamin application at various resting potentials. (B)AHP amplitudes were measured at 15 ms after the peak of evoked action potential (dashed line), and the values were plotted with the resting potentials recorded. The resting potential showing no currents at 15 ms after the overshoot was estimated from the plotted line and is designated as the V₀ value in this analysis. (C) The V₀ value was variable in response to inhibitors for Ca²⁺ signaling molecules in control neurons but moderately stable in JP-DKO neurons. The drugs used for the bath application were 1 μM apamin (Apa), 100 nM iberiotoxin (IBTX), 50 μM APV, 10 μM cyclopiazonic acid (CPA), and 50 μM ryanodine (Ry). The data were obtained from at least eight slices derived from four mice and represent the mean ± SEM. Statistical differences from the control value (None) in each genotype are indicated by asterisks (∗∗, P < 0.01 in Student’s t test).

Contribution of Ca²⁺ Release for SK Channel Activation and Its Dysfunction in JP-DKO Neurons.

AHP generated by depolarizing currents in control CA1 neurons was abolished by 2-amino-5-phosphonovaleric acid (APV), an inhibitor of NMDA receptors (NMDARs) (Fig. 3C). The extrasynaptic space contains micromolar concentrations of glutamate in the hippocampus (14), and this ambient glutamate, possibly with ambient glycine/d-serine, probably induces a tonic activation of NMDARs with high affinities for the agonists (15, 16). In the hippocampal slices, significant fractions of NMDARs appear to bind with ambient glutamate but are inactive at the resting potential because of the blockage by the extracellular Mg²⁺. It is likely that the somatic action potential removes the Mg²⁺ block of dendritic NMDARs to produce Ca²⁺ influx. Therefore, our observations indicate that the activation of SK channels requires NMDAR-mediated Ca²⁺ influx in CA1 neurons, as reported in ref. 17. In JP-DKO neurons, APV had no significant effect on action potential configuration, suggesting the disconnection of functional communication between NMDARs and SK channels.

Surprisingly, AHP in control mice disappeared in response to cyclopiazonic acid, a sarcoplasmic reticulum/ER Ca²⁺-ATPase inhibitor inducing ER Ca²⁺ depletion (Fig. 3C). In addition, ryanodine, a specific inhibitor of RyR-mediated Ca²⁺ release, also abolished AHP generation. Therefore, RyR-mediated Ca²⁺ release from the ER seems to be essential for SK channel activation in CA1 neurons, and it is rather unlikely that NMDAR-mediated Ca²⁺ influx directly opens SK channels. However, the treatments canceling ER Ca²⁺ release produced no significant effects in JP-DKO neurons. Nevertheless, cultured hippocampal pyramidal neurons from JP-DKO mice retained regular caffeine-evoked Ca²⁺ transients (Fig. 8B). These observations probably suggest that RyR-mediated Ca²⁺ release is impaired under physiological conditions in JP-DKO neurons. This notion was further supported by the data from the extracellular recording of CA1 excitatory postsynaptic potentials (EPSPs) triggered by the orthodromic stimuli of Schaffer collaterals (see Fig. 9, which is published as supporting information on the PNAS web site). In this analysis, apamin- and APV-sensitive AHP was detectable as the positive deflection of population spikes and required ER Ca²⁺ release in control neurons. In the recordings of JP-DKO hippocampal slices, the positive deflection was not detected, and the treatments inducing the dysfunction of ER Ca²⁺ release showed no significant effects on EPSP profile.

Impaired Hippocampal Plasticity in JP-DKO Mice.

Because the impaired memory may imply abnormal hippocampal plasticity in JP-DKO mice, we next analyzed CA1 long-term potentiation (LTP) induced by a high-frequency stimulation (HFS) using the slice preparations. In control mice, HFS caused a stable and long-lasting potentiation of EPSPs as reported in ref. 18. In JP-DKO mice, highly enhanced EPSP potentiation was observed immediately after the same HFS (Fig. 4A). This irregular transient potentiation was mediated by NMDAR activation because the same HFS in the presence of APV induced normal posttetanic potentiation as presynaptic short-term plasticity in JP-DKO mice (Fig. 4B). The increased potentiation of JP-DKO mice was quickly attenuated, and a significantly smaller LTP was observed afterward compared with control LTP (see Fig. 10, which is published as supporting information on the PNAS web site). No differences in posttetanic potentiation and the envelope of depolarization during HFS as indexes for presynaptic excitability were observed between the genotypes (Fig. 4C), suggesting normal presynaptic functions in JP-DKO mice. Therefore, the irregular transient potentiation and impaired LTP were probably caused by postsynaptic abnormalities in JP-DKO mice.

Fig. 4.

Impaired LTP in JP-DKO hippocampal neurons. (A) Abnormal LTP profile in JP-DKO mice. Initial EPSP slopes were normalized in each experiment by using the averaged slope value before HFS (100 Hz for 1 s). At most time points after HFS, significant differences were detected between the genotypes (P < 0.05 in Student’s t test). (B) Normal response of posttetanic potentiation in JP-DKO mice. Short-term plasticity evoked by HFS in the presence of 50 μM APV was analyzed. (C) Normal envelope of depolarization during HFS in JP-DKO mice. Depolarization at 200 ms from the beginning of the tetanus was normalized among the slices by dividing it by the amplitude of baseline EPSPs, as described in ref. 18. (D) Western blot analysis of phosphoproteins in CA1 slices. CA1 slices were prepared before HFS (control) or 60 min after HFS and analyzed with various antibodies. Representative immunoreactive signals produced by specific antibodies are shown. The antibodies to CaMKII detected both α and β isoforms. (E) Summary histogram showing relative immunoreactivity in Western blot analysis. Signal densities were normalized for each antibody by using the values from JP-DHE CA1 slices without the HFS treatment. The data were obtained from at least six slices derived from three mice and represent the mean ± SEM. Statistical differences between JP-DKO and JP-DHE mice are indicated by asterisks (∗, P < 0.05 and ∗∗, P < 0.01 in Student’s t test). Syn, synapsin I; MAR, myristoylated alanine-rich C kinase substrate; DARP, DARPP-32; p, phospho-.

Because Ca²⁺/calmodulin protein kinase II (CaMKII) activation by HFS is essential for hippocampal LTP induction (19, 20), we examined its intrinsic activity in CA1 slices by Western blotting before and after HFS (Fig. 4D). The summary histogram for relative immunoreactivity demonstrates that basal phospho-CaMKII levels were highly increased without a shift in protein content in JP-DKO mice (Fig. 4E). HFS increased phospho-CaMKII levels in control mice but did not further induce its autophosphorylation in JP-DKO mice. The proposed CaMKII hyperactivation under basal conditions was further supported by results on phospho-Ser-831 of glutamate receptor type 1, which is a postsynaptic CaMKII target (21). However, we detected no abnormalities in phosphosynapsin I at Ser-603, a presynaptic CaMKII target (22) in JP-DKO mice, and thus presynaptic CaMKII activity was probably within normal ranges. These observations indicate that CaMKII hyperactivation is produced in the postsynaptic CA1 neurons of JP-DKO mice, and this abnormality probably underlies the impaired LTP described above. In contrast with the observations on CaMKII, JP-DKO mice showed no abnormalities in phospho-PKCα and phospho-myristoylated alanine-rich C kinase substrate as a substrate of PKC isoforms. Moreover, we could not detect any abnormality in the phosphorylation of glutamate receptor type 1 at Ser-845 and DARPP-32 at Thr-34 as cAMP-dependent PK (PKA) targets. Therefore, the activity levels of PKCα and PKA seem to be apparently normal in the CA1 region of JP-DKO mice.

Discussion

The insertions of triplet repeats in the JP-3 locus result in a human genetic disease called HDL2, with clinical symptoms similar to those of Huntington’s disease (23). Although HDL2 bears the features of polyglutamine diseases, the triplet repeats do not encode polyglutamine and seem to simply disrupt the JP-3 gene. The abnormal foot-clasping reflex is shared by JP-DKO and Huntington’s disease model (R6/2) mice (10). The direct cause of the aberrant behavior is unknown; JP-DKO and R6/2 mice probably incur overlapping damages in neural networks responsible for the hindlimb reflex. Therefore, JP-DKO mice will be a useful model system for HDL2 in future studies.

Our previous studies on muscle cells may predict that the role of JPs is generally associated with RyR functions in excitable cell types. The present results demonstrate that the loss of neural JPs leads to abnormalities in hippocampal neurons at the cellular level, as well as an impaired memory at the whole-animal level. Of the RyR subtypes, namely RyR-1–RyR-3, both RyR-2 and RyR-3 are abundantly expressed in CA1 neurons (24), and RyR-3 knockout mice showed a poor memory and impaired CA1 LTP (25–27). Moreover, RyR-3 knockout and JP-DKO mice share abnormal characteristics during LTP induction including an enhanced transient potentiation immediately after HFS and impaired LTP afterward. Therefore, both mutant mice probably bear similar signaling defects in postsynaptic CA1 neurons at least in part, and we can presume impaired RyR functions in JP-DKO neurons.

Previous studies have clearly demonstrated that NMDAR-dependent Ca²⁺ influx is required for SK channel activation in CA1 neurons (17). Our present data suggest that RyR-mediated Ca²⁺ release evoked by the Ca²⁺ influx is important for SK channel activation. Thus, we can propose a Ca²⁺-mediated signal cascade in which NMDAR activation to produce Ca²⁺ influx, RyR activation through the Ca²⁺-induced Ca²⁺ release mechanism, and SK channel activation triggered by ER Ca²⁺ release take place successively for AHP generation (Fig. 5). Insufficient Ca²⁺ release proposed in RyR3 knockout CA1 neurons probably leads to an impairment of this signal cascade. However, the proposed JP-mediated JMC formation would support this signal cascade by providing the platform for functional coupling between the channels. The mobility of intracellular Ca²⁺ is tightly restricted by cytosolic buffering effects (1). Therefore, geographic decoupling between the channels under JP-deficient conditions may lead to diminished Ca²⁺ supply for SK channel activation, which is possibly caused by insufficient RyR activation synchronized with NMDAR opening and/or ectopic RyR-mediated Ca²⁺ release at the microdomain level. The abolishment of AHP possibly prolongs the opening of voltage-gated Ca²⁺ channels and facilitates firing frequency in JP-DKO neurons. Moreover, enhanced local Ca²⁺ elevation underneath the plasma membrane can be also predicted by detaching the ER with strong Ca²⁺ buffering from the plasma membrane. These deduced atypical Ca²⁺ signalings possibly result in CaMKII hyperactivation in JP-DKO CA1 neurons. However, it is unclear why both hyper-CaMKII activity and normal PKCα activity are suggested in JP-DKO neurons, given that the activation of both kinases absolutely requires cytosolic Ca²⁺. This discrepancy might reflect their subcellular localization; CaMKII is predominantly localized in the postsynaptic density (28), but cytoplasmic PKCα translocates to the ER and plasma membrane upon its activation (29). Again, it can be proposed that the irregular Ca²⁺ signaling occurs at the microdomain level in JP-DKO CA1 neurons.

Fig. 5.

Proposed signaling cascade for AHP generation in hippocampal neuron and its dysfunction in JP-DKO mice. The hypothetical molecular mechanism underlying AHP generation in CA1 neurons is schematically illustrated. The present results suggest that AHP generation requires Ca²⁺-mediated communication between NMDARs, RyRs, and SK channels. Both JP-3 and JP-4 collaboratively contribute to the proposed JMC formation and probably provide the structural platform for the signal cascade. In JP-DKO neurons, the Ca²⁺-mediated communication is likely to be damaged, and the resulting AHP deficiency probably underlies the abnormal plasticity observed.

In this article, we have focused on the “functional triplet” composed of NMDARs, RyRs, and SK channels in hippocampal neurons. In cerebellar Purkinje cells expressing no functional NMDARs, the activation of SK channels seems to require Ca²⁺ influx through P/Q-type voltage-gated Ca²⁺ channels (30). In the triplet, NMDAR could be replaced by a certain set of cell surface Ca²⁺ channels. In pyramidal neurons of the visual cortex, inositol 1,4,5-trisphosphate receptors are probably involved in SK channel activation (31), and the triplet could be customized by using inositol 1,4,5-trisphosphate receptors instead of RyRs. Moreover, smooth muscle cells seem to possess a functional communication between RyRs and large-conductance Ca²⁺-dependent K⁺ channels (32). It is thus important to examine the predicted functional triplets among cell surface Ca²⁺ channels, Ca²⁺ release channels, and Ca²⁺-activated K⁺ channels in other neural sites, and JP-DKO mice would provide us an ideal model system in future studies.

Materials and Methods

Generation of Knockout Mice.

JP-4 knockout mice were generated as described in ref. 33. Chimeric mice produced with the positive embryonic stem cell clones numbered 132 and 440 were crossed with C57BL/6J mice and could transmit the mutant gene to their offspring. The PCR primers used for genotyping the mutant mice were JP4–1 (GTGTGAGTGGAGGGACCCTGCCATG), JP4–2 (CCTCAGTGGCTGAGCCCTCGATGAG), and JP4–12 (GAAAGCCGGCCCCAGGTGGGCCGATG). JP-DKO and DHE mice were fed with paste food to avoid possible chew effects on the experiments, and WT mice were fed with normal pellet food. All of the mice examined shared the genetic background of the 129-B6 hybrid. Anatomical observations in the mouse brain were carried out as described in refs. 5 and 9.

Behavioral Analysis.

Mice were maintained on a 12-h light–dark cycle (light time, 0800–2000 h), and food and water were supplied ad libitum. All mice were tested in the particular behavioral assays during the late light part of the cycle (1630–1800 h). Locomotion activity in an open field was analyzed as described in ref. 26. Each mouse was placed in a novel environment of the testing cage (19 × 26 × 12 cm) with an IR beam sensor (CompACT-AMS; Muromachi Kikai Co., Tokyo). Exploratory locomotion activity was recorded in blocks of 20 min. The step-through-type passive avoidance test was carried out by using an apparatus composed of lighted and darkened compartments as described in ref. 18. When mice in the lighted compartment stepped through the door to the dark compartment, an electric shock (2 mA for 1 s) was delivered through the grid floor (Shock Generator SGS-001; Muromachi Kikai Co.) in the acquisition trial. This training continued until the mice stayed in the lighted compartment for 120 s. After 3 days, step-through latency was recorded until 300 s in the retention trial. The Y-maze test was carried out as described in ref. 26. The maze was composed of three arms (40 cm long, 5 cm wide, and 8 cm high) converged at equal angles. Each mouse was placed at the end of one arm and allowed to move freely during an 8-min session.

Electrophysiological Analysis.

Hippocampal slices (400 μm thick) were prepared by using a tissue slicer and submerged beneath a perfusion medium (125 mM NaCl/2.5 mM KCl/1.3 mM MgSO₄/2.4 mM CaCl₂/1.0 mM NaH₂PO₄/26 mM NaHCO₃/11 mM glucose/0.1 mM picrotoxin) saturated with 95% O₂/5% CO₂ and warmed at 34°C. Whole-cell recordings from CA1 neurons in current-clamp mode were made by using an Axopatch 200A amplifier (Axon Instruments, Union City, CA) as described in ref. 34. Firing responses evoked by depolarizing currents were filtered at 2 kHz, digitalized at 5 kHz, and analyzed by using pclamp software (Axon Instruments) using the internal pipette solution containing 40 mM KCl, 80 mM KCH₃SO₄, 1 mM MgSO₄, 10 mM Hepes, 0.5 mM EGTA, 2 mM ATP-Mg (pH 7.4), and the perfusion medium described above. For the extracellular recording of CA1 EPSPs, a bipolar stimulating electrode was placed in the stratum radiatum, Schaffer collateral/commissural fibers were stimulated at 0.05 Hz, and field-potential recording was made by using a glass electrode containing 3 M NaCl. The data were recorded and digitized by using a single-electrode amplifier (CEZ-3100; Nihon Kohden, Tokyo) and analyzed by using PowerLab 200 (A. D. Instruments, Milford, MA). All experiments were performed at room temperature (24–28°C), as described in refs. 18 and 27. Before and after HFS (100 Hz for 1 s), the CA1 regions were cut out from the hippocampal slices and frozen in liquid nitrogen for Western blot analysis.

Western Blotting.

Hippocampal CA1 slices were homogenized in the buffer (70 μl) containing 50 mM Tris·HCl (pH 7.4), 0.5% Triton X-100, 4 mM EGTA, 10 mM EDTA, 1 mM Na₃VO₄, 40 mM sodium pyrophosphate, 50 mM NaF, 100 nM calyculin A, 50 μg/ml leupeptin, 25 μg/ml pepstain A, 50 μg/ml trypsin inhibitor, and 1 mM DTT. After the removal of insoluble materials by centrifugation (15,000 rpm for 10 min), the samples were subjected to immunoblotting as described in ref. 20. Immunoreactivity was visualized with an enhanced chemiluminescense detection system (Amersham Pharmacia) and analyzed quantitatively by using the National Institutes of Health image program.

Abbreviations

RyR

ryanodine receptor

JMC

junctional membrane complex

JP

junctophilin

DKO mice

double-knockout mice

ER

endoplasmic reticulum

AHP

afterhyperpolarization

SK channel

small-conductance Ca²⁺-dependent K⁺ channel

NMDAR

NMDA receptor

EPSP

excitatory postsynaptic potential

APV

2-amino-5-phosphonovaleric acid

LTP

long-term potentiation

HFS

high-frequency stimulation

CaMKII

Ca²⁺/calmodulin protein kinase II.