Activation of Calcium/Calmodulin-dependent Protein Kinase IV in Long Term Potentiation in the Rat Hippocampal CA1 Region*

The importance of well characterized calcium/calmodulin-dependent protein kinase (CaMK) II in hippocampal long term potentiation (LTP) is widely well established; however, several CaMKs other than CaMKII are not yet clearly characterized and understood. Here we report the activation of CaMKIV, which is phosphorylated by CaMK kinase and localized predominantly in neuronal nuclei, and its functional role as a cyclic AMP-responsive element-binding protein (CREB) kinase in high frequency stimulation (HFS)-induced LTP in the rat hippocampal CA1 region. CaMKIV was transiently activated in neuronal nuclei after HFS, and the activation returned to the basal level within 30 min. Phosphorylation of CREB, which is a CaMKIV substrate, and expression of c-Fos protein, which is regulated by CREB, increased during LTP. This increase was inhibited mainly by CaMK inhibitors and also by an inhibitor for mitogen-activated protein kinase cascade, although to a lesser extent. Our results suggest that CaMKIV functions as a CREB kinase and controls CREB-regulated gene expression during HFS-induced LTP in the rat hippocampal CA1 region.

Introduction of a dominant negative CaMKIV or dominant negative CREB into cultured cerebellar Purkinje neurons resulted in defects in the late phase of long term depression in response to glutamate (28). CaMKIV-deficient mice also showed decreased potentiation at a later stage of rat hippocampal CA1 LTP (26). These reports suggest that CaMKIV functions as a CREB kinase and regulates CREB-dependent gene expression in neurons, which may in turn be linked to a protein synthesis-dependent long lasting synaptic plasticity.
To understand the molecular mechanisms of synaptic plasticity and function of CaMKIV, it is essential to know how CaMKIV activity changes following application of conditioning stimuli such as high frequency stimulation (HFS) to induce LTP and how CREB phosphorylation is regulated by CaMKIV and other protein kinases under these conditions. In the present study, we examined CaMKIV activation with simultaneous CREB phosphorylation and stimulation of gene expression during HFS-induced LTP in the rat hippocampal CA1 region.

EXPERIMENTAL PROCEDURES
Materials-The following chemicals and materials were obtained from the indicated sources: [␥-32 P]ATP and 125 I-protein A from PerkinElmer Life Sciences; calmidazolium (R24517) and H89 from Sigma; KT5720 from Alexis Biochemicals (San Diego, CA); KN93 and KN92 from Seikagaku (Tokyo, Japan); U0126 and U0124 from Calbio-* This work was supported by grants-in-aid for scientific research and for scientific research on priority areas from the Ministry of Education, Science, Sports and Culture of Japan and by the Human Frontier Science Program. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Electrophysiology-Electrophysiological experiments were performed basically according to a previous report (7). The brains were rapidly removed from ether-anesthetized male Wistar rats (5-7 weeks old), and the hippocampi were dissected out. Transverse hippocampal slices (450 m thick) prepared using a McIlwain tissue chopper were incubated in continuously oxygenized (95% O 2 , 5% CO 2 ) artificial cerebrospinal fluid (ACSF) at room temperature for at least 2 h. After a 2-h recovery period, a slice was transferred to an interface recording chamber and perfused at a flow rate of 2 ml/min with ACSF warmed at 30°C. The field EPSP (fEPSP) was evoked by 0.05 Hz test stimuli through a bipolar stimulating electrode placed on the Schaffer collateral/commissural pathway and recorded from the stratum radiatum of CA1 using a glass electrode filled with 3 M NaCl. In the case of CA3 LTP, a stimulating electrode was placed on the mossy fibers in the dentate region, and the recordings were made from the stratum lucidum of CA3. The recording was performed using a single-electrode amplifier (CEZ-3100, Nihon Kohden, Tokyo, Japan), and the maximal value of the initial fEPSP slope was collected and averaged for every 1 min (3 traces) using an A/D converter (PowerLab 200; ADInstruments, Castle Hill, Australia) and a personal computer. The stimulus intensity was adjusted to evoke a fEPSP of 1.0-mV amplitude. After the stable base-line recording was obtained, a HFS of 100 Hz frequency and 1-s duration was applied two times with a 20-s interval between times, and then the recording was continued for the indicated periods in the figures. For the control, slices recorded for the same periods without HFS were prepared. The following inhibitors for protein kinases were dissolved in Me 2 SO as a stock solution at the following concentrations and stored at Ϫ20°C until use: calmidazolium, KN93, and KN92, 50 mM; U0126 and U0124, 10 mM; and H89, 10 mM. The solutions were diluted in ACSF just prior to use at the following concentrations: calmidazolium, KN93, and KN92, 50 M; U0126 and U0124, 10 M; and H89, 5 M. KT5720 was dissolved in methanol at a concentration of 1 mM as a stock solution and then diluted in ACSF at a concentration of 1 M. The slices were treated with these solutions before and during the recording periods, as described in the figure legends. After the electrophysiological recordings, slices were transferred to a plastic Petri dish cooled on ice, and the CA1 areas were dissected out, immediately frozen in liquid nitrogen, and stored at Ϫ80°C until the biochemical assay.
Immunohistochemistry-After the electrophysiological experiments, the slices were fixed at 4°C overnight in 50 mM Tris-HCl, pH 7.5, containing 4% paraformaldehyde and 0.1% glutaraldehyde. Slices of 50-m thickness were prepared from 450-m thick slices in Tris-azide buffer (50 mM Tris-HCl, pH 7.5, and 0.01% sodium azide) using a tissue slicer, Vibratome 1000 (Pelco International, Redding). The thin slices were incubated in Tris-azide buffer containing 1% bovine serum albumin and 0.2% Triton X-100 for 2 h at room temperature and then incubated with the first antibody (anti-pT196, 1:200). Bound antibodies were visualized using FITC-conjugated anti-rabbit IgG and analyzed by laser confocal microscopy (Olympus, Tokyo, Japan). Quantification of nuclear CaMKIV phosphorylation was performed as follows. The intensities of the nuclear and cytosolic fluorescence of neurons with FITC were compared in 35-40 neurons, using the Image program, that are located in a 100 ϫ 100-m area between the stimulation and recording electrodes in the CA1 pyramidal cell layer. The number of the neurons in which the intensities of the nuclei are brighter than those of the cytosols was counted. Neurons, which have higher intensities in the nuclei, were defined as positive. The independent experiments of different hippocampal slices were performed four times, and statistical analysis was made.
Statistical Evaluation-The values are the means Ϯ S.E. Comparison between two experimental groups was made by the unpaired Student's t test. For multiple comparisons, one-way analysis of variance with Sheffe's correction was used, and p values of Ͻ 0.05 were considered to have statistical significance.

Activation of CaMKIV, MAPK, and CREB during LTP
Induction-We first examined the changes of CaMKIV activity during LTP. Application of HFS to Schaffer-collateral pathways reliably induced the potentiation of fEPSP slope, which lasts over 60 min in the CA1 region (166 Ϯ 17% of base line at 60 min after HFS; Fig. 1A). CaMKIV activity was determined 3, 10, 30, and 60 min after HFS application and was significantly increased within 3 min (201 Ϯ 12% of control) and 10 min (127 Ϯ 13% of control), returning to basal levels within 30 min after HFS (Fig. 1B). This result was confirmed by immunoblot analysis using the anti-pT196 antibody (21) to assess levels of CaMKIV phosphorylation at Thr 196 (Fig. 1C), an essential site for activation of CaMKIV by CaMK kinase. This result is summarized by statistical analysis in Fig. 1D. Phosphorylation of Thr 196 was significantly increased 3 min after HFS (226 Ϯ 44% of control) and then returned to the basal level. CaMKIV protein levels were unchanged 3, 10, and 30 min after HFS, although a slight but significant increase (118 Ϯ 7.6% of control; n ϭ 6, p Ͻ 0.05) was observed 60 min after HFS and is consistent with the previously reported result (29). We also examined the phosphorylation of p42 mitogen-activated protein kinase (MAPK)/extracellular signal-regulated protein kinase 2 (ERK2) and CREB, and phosphorylation of both was detected with phosphospecific antibodies (Fig. 1C, summarized in Fig. 1D). Phosphorylation of p42 MAPK was significantly increased 3 min after HFS (180 Ϯ 15% of control) and then returned to basal levels. Because HFS selectively activated p42 MAPK but not p44 MAPK, changes in phosphorylation of p42MAPK were examined in the present study. This result is consistent with the previous reports (7, 30) that activity of p42 MAPK is transiently increased by HFS in the rat hippocampal CA1 region. In contrast, CREB phosphorylation detected with an anti-phosho-Ser 133 antibody increased 3 min (151 Ϯ 14% of control), 10 min (132 Ϯ 10% of control), 30 min (151 Ϯ 24% of control), and 60 min (169 Ϯ 29% of control) after HFS, and the increased phosphorylation was sustained during LTP (Fig. 1D). Although the total amounts of p42 MAPK were unchanged 3 and 10 min after HFS, a slight but significant increase was observed 30 min (116 Ϯ 6.2% of control; n ϭ 7) and 60 min (121 Ϯ 5.9% of control; n ϭ 6) after HFS (p Ͻ 0.05). The protein levels of CREB were not significantly changed at any time point after HFS (data not shown).
To determine the localization of activated CaMKIV in neurons, we undertook an immunohistochemical study using the anti-pT196 antibody. In control slices, moderate immunoreac-tivity was observed in the cytosol ( Fig. 2A) but was not competed by absorbance with the peptide antigen (21) (Fig. 2C), indicating that it was nonspecific. The strong nonspecific immunoreactivity in the cytosol would probably mask any small amount of CaMKIV activation in the cytosol. Neuronal nuclei were not stained ( Fig. 2A). By contrast, strong immunoreactivity was observed in the nuclei of neurons in HFS-applied slices (Fig. 2B), and that immunoreactivity was absent when the antibody was absorbed with the peptide antigen (Fig. 2C), indicating that it was specific. Although the antibody used in the present study was specific, it seems sometimes that the antibody lacks specificity against nondenatured protein immunocytochemistry. Therefore, the detected immunoreactivity may be called CaMKIV-like immunoreactivity. These results suggest that CaMKIV is activated in the nuclei of CA1 pyramidal neurons by HFS application. Specific immunoreactivity was also observed in the dendrites of neurons (Fig. 2B), because its expression was not limited to nuclei (19). We observed slight expression of CaMKIV in the cytosols and dendrites of cultured neurons (data not shown).
Effect of Calmodulin Antagonist and CaMK Inhibitor on LTP and CREB Phosphorylation-Because it was reported that the major CREB kinase was CaMKIV in response to neuronal stimulation in cultured rat hippocampal neurons (25), we addressed whether CaMKIV functioned as a CREB kinase during LTP. To inhibit CaMKIV activity, we used calmidazolium, a calmodulin antagonist, and KN93, a CaMK inhibitor. Both inhibitors have been previously shown to inhibit CaMKIV activity in cultured rat hippocampal neurons (20). The application of calmidazolium (Fig. 3A) or KN93 (Fig. 3D) inhibited induction of LTP, possibly by inhibiting CaMKII (6, 7). Calmidazolium (n ϭ 7) and KN93 (n ϭ 8) did significantly inhibit CaMKIV phosphorylation without any effect on p42 MAPK phosphorylation (Fig. 3, B and E). HFS-induced CREB phosphorylation was, however, significantly inhibited by both compounds 3 and 30 min after HFS (Fig. 3, B, C, E, and F). By contrast, KN92, an inactive compound similar to KN93, affected neither LTP induction nor CREB phosphorylation (Fig.  3, D and F), indicating specific inhibition by KN93. These results demonstrate a positive correlation between CaMKIV activation and CREB phosphorylation and are consistent with the idea that CaMKIV functions as a CREB kinase during HFS-induced LTP.
Effect of MEK Inhibitor on LTP and CREB Phosphorylation-We next examined the effect of MAPK on HFS-induced CREB phosphorylation. The MAPK cascade is thought to be involved in CREB phosphorylation through the activation of p90 ribosomal S6 kinase 2 (RSK2) (31). We previously reported that a high concentration of PD098059, an MEK inhibitor, inhibited LTP induction (7). However, the effect of this inhibitor was positively correlated with inactivation of CaMKII, and therefore we concluded that PD098059 inhibited LTP through inactivation of CaMKII rather than MAPK (7). In this study, we examined the effect of U0126 (32), which specifically inhibits MEK at a lower concentration than PD098059. Under the conditions used, U0126 did not affect the induction of HFSinduced LTP (Fig. 4A), a finding consistent with other reports (33,34). 10 M U0126 dramatically inhibited p42 MAPK phosphorylation without any effect on CaMKIV phosphorylation (Fig. 4B). To further inhibit p42 MAPK, 20 and 30 M U0126 was used, and inhibition of CaMKIV phosphorylation was not observed. On the other hand, CREB phosphorylation was not inhibited 3 min after HFS but was significantly inhibited 30 min after HFS (Fig. 4C). These results suggest that not only CaMKIV but also MAPK pathways function as CREB kinases in HFS-induced LTP. It should be noted that CaMKIV functions as a CREB kinase at a relatively early stage of HFSinduced LTP, whereas MAPK functions at a later point during the time course.
Effect of PKA Inhibitor on LTP and CREB Phosphorylation-We examined the effects of H89 and KT5720, inhibitors for PKA, on HFS-induced CREB phosphorylation, because PKA can directly phosphorylate and activate CREB. In our experi-mental conditions, H89 and KT5720 did not inhibit HFS-induced LTP in the CA1 region (Fig. 5A). In contrast, the mossy fiber CA3 LTP was reduced by PKA inhibitors (Fig. 5B), as previously reported (35), indicating that these inhibitors were active under our experimental conditions. Because H89 has been shown to have nonspecific effects on other kinases than PKA, compared with KT5720, we checked its effect on CaMKIV and MAPK phosphorylation. As shown in Fig. 5C, H89 had no effect on CaMKIV and MAPK phosphorylation. CREB phosphorylation was inhibited neither by H89 nor by KT5720 during LTP in the CA1 region (Fig. 5C). The results for CREB phosphorylation are summarized in Fig. 5D. Neither of the PKA inhibitors inhibited CREB phosphorylation induced by LTP at 3 and 30 min after HFS in the CA1 region. These results suggest that the PKA pathway is involved neither in HFSinduced LTP nor in CREB phosphorylation at an early stage and are consistent with the previous report (36).
Stimulation of c-Fos Expression during LTP-To address whether CREB phosphorylation induced by HFS stimulates gene expression, we examined the expression of the c-Fos protein, one of the markers for cyclic AMP-responsive element promoter-mediated gene expression (37). The expression of the c-Fos protein was unchanged 3 and 10 min after LTP induction and then significantly increased 30 (115 Ϯ 10% of control) and 60 (153 Ϯ 13% of control) min after LTP induction (Fig. 6A), suggesting that the phosphorylated CREB stimulated c-Fos expression. We further tested whether c-Fos expression was regulated by CaMKIV and/or MAPK pathways during LTP. The increase in c-Fos expression at 60 min after HFS was inhibited by the addition of calmidazolium or U0126 (Fig. 6B). Treatment of hippocampal slices with calmidazolium strongly inhibited the HFS-induced c-Fos expression, whereas treatment with U0126 moderately but significantly inhibited c-Fos expression (Fig. 6B). The extent of inhibition by calmidazolium and U0126 was positively correlated with levels of CREB phosphorylation shown in Figs. 3 and 4. These results suggest that both CaMKIV and MAPK function as CREB kinases, which in turn regulate CREB-mediated gene expression in HFS-induced LTP. DISCUSSION Here we show that CaMKIV is activated accompanied with phosphorylation of Thr 196 with CREB phosphorylation and stimulation of c-Fos expression during LTP. To our knowledge, this is the first report to show that CaMKIV is activated and involved in stimulation of gene expression during LTP in the hippocampal CA1 region.
The activation of CaMKIV shown in Fig. 1 was transient and returned to the original level in a short time. This activation pattern differed from that of CaMKII, which showed a long lasting activation during LTP (5, 6). In our previous studies, the difference in activation patterns of CaMKII (21, 38) and IV (21) was also observed in cultured hippocampal neurons in response to glutamate receptor stimulation. Because transient activation of CaMKIV was also reported in Jurkat T cells (22) in response to T cell receptor stimulation, it may be common in most cell types. One explanation for the difference in activation of CaMKII and CaMKIV may be that the latter is dephosphorylated by Ca 2ϩ /calmodulin-dependent protein phosphatase 2B (calcineurin; Ref. 21), whereas CaMKII cannot be dephosphorylated directly by calcineurin (39). Elevation of intracellular Ca 2ϩ is required to activate CaMKIV. Such elevation activates calcineurin, which may rapidly dephosphorylate and inactivate phosphorylated CaMKIV, which differs from the persistent activation of CaMKII.
Although activation of CaMKIV was transient, the present study suggests that CaMKIV acts as a CREB kinase. Experiments with a calmodulin antagonist and protein kinase inhibitors demonstrated that CaMKs are mainly involved in phosphorylation of CREB, with a slight phosphorylation by MAPK, and that PKA has no effect on phosphorylation of CREB. Among CaMKs, the contribution of CaMKII to CREB phosphorylation may be excluded, because c-Fos expression was stimulated during LTP in the present study, and phosphorylation of CREB by CaMKII reportedly had inhibitory effects on gene expression (40). These results were consistent with a recent report (26) that showed reduced CREB phosphorylation and no significant potentiation of fEPSP at 45 min after LTP in hippocampal slices of CaMKIV-deficient mice with normally expressed CaMKII. On the other hand, the c-fos promoter the contains cyclic AMP-responsive element and other motifs such as the serum response element. Because the serum response element was reported to be involved in activation of c-fos promoter by CaMKII (41), stimulation of c-Fos protein expression may also occur during LTP. Because we have no selective inhibitors for CaMK subtypes, it was not possible to identify which subtypes of CaMK are involved in HFS-induced stimulation of c-Fos protein expression.
Involvement of MAPK cascade in hippocampal LTP (7,30,33,34) and learning and memory (42) was recently reported. Consistent with these reports, we confirmed that MAPK was transiently activated in CA1 LTP. It should be noted that MAPK seems to function as a CREB kinase at a later stage of LTP, although MAPK activation was transient. Although CaMKIV and MAPK are transiently activated, such activation in the context of a signaling cascade would be sufficient to trigger changes in gene expression. CaMKIV can directly phosphorylate CREB, which may enable CaMKIV to function as a CREB kinase at a relatively early stage in LTP. MAPK phosphorylates RSK2, which in turn phosphorylates CREB, and the time lag required for the signal transduction to occur may be too great to allow the MAPK cascade to function as a CREB kinase at an early rather than a late stage of LTP. Because c-Fos expression observed 60 min after HFS was inhibited by both inhibitors for CaMKIV and MAPK, activation of both enzymes observed during LTP is enough to be involved in gene expression.
To form long term memory, newly synthesized synaptic components and the formation of a new circuit such as morphologically changed spines may be required (43). This indicates that the stimulation of gene expression and consequent protein synthesis are needed for the formation of memory. In view of the observation on CREB phosphorylation and stimulation of c-Fos expression by CaMKIV and MAPK in the present study, activation of both enzymes may be associated with LTP maintenance and memory formation.
Recent reports showed that PKA is activated during LTP in the CA1 region (44) and is involved in the late phase of LTP (45) and CREB-stimulated gene expression (46). These reports differed from results presented here, because the PKA inhibitors had no effect on LTP induction and CREB phosphorylation. These discrepancies could be due to several reasons. First, there were differences in conditions used to induce LTP in ours and previous studies. In the previous studies, more than twice repetitive 100 Hz trains with 5-10-min intervals (44 -46) or low frequency stimulation of 5 Hz (33, 34) were used to induce LTP, which was blocked by inhibitors for PKA (33, 44 -46) and MAPK (33,34), whereas here, LTP induced by two trains of 100 Hz stimulation with a 20-s interval was not inhibited by these inhibitors (33). There is no doubt that conditional changes in the experiments would bring about different results. Second, other reports showed that PKA is involved in LTP maintenance 3 h after HFS but not involved in LTP induction (36,45). Because the effect of a PKA inhibitor was examined at the early stage of LTP in the present study, the results were consistent with these reports. Third, it has been reported that protein phosphatases, especially protein phosphatase 1, are involved in LTP induction in the PKA pathway (47). In our experimental conditions, however, it seems that PKA may not be activated, FIG. 6. Expression of c-Fos during LTP. A, c-Fos expression was examined during the time course of LTP. The samples were collected at the indicated time points after HFS, and c-Fos expression was determined with a bio-imaging analyzer for autoradiograms. *, p Ͻ 0.05; **, p Ͻ 0.01. The data from six experiments are shown. B, effects of protein kinase inhibitors on c-Fos expression. All of the samples were collected 60 min after HFS. The slices were treated with no drug (control) and 50 M calmidazolium (CMZ) for 2-3 h and with 10 M U0126 for 1-2 h before and throughout the recording periods. The data from five (calmidazolium) and eight (U0126) experiments are shown. because the PKA inhibitors had no effect on HFS-induced LTP and CREB phosphorylation. In other words, phosphorylation of inhibitor 1 by PKA may not have occurred under the conditions of the present study.
In conclusion, the present study showed that CaMKIV activation was associated with LTP induction and had a positive correlation to CREB phosphorylation and up-regulation of c-Fos expression. Further study will be required to clarify how the actual targets of CaMKIV and MAPK function to alter gene expression required to form long lasting memory.