Neuronal activity increases the phosphorylation of the transcription factor cAMP response element-binding protein (CREB) in rat hippocampus and cortex.

Activity-mediated gene expression is thought to play an important role in many forms of neuronal plasticities. We have used pentylenetetrazol-induced seizure that produces synchronous and sustained neuronal activity as a model to examine the mechanism(s) of gene activation. The transcription factor CREB (Ca2+/cAMP response element-binding protein) is thought to be necessary for long-term memory formation both in invertebrates and vertebrates. When phosphorylated on Ser133 either by cAMP-dependent protein kinase and/or Ca2+/calmodulin-dependent protein kinases, CREB increases transcription of genes containing the CRE (cAMP response element) sequence. Using an antibody that detects Ser133-phosphorylated CREB protein, we show that CREB phosphorylation is maximal between 3 and 8 min after the onset of seizure activity and declines slowly both in the hippocampus and the cortex. The total amount of CREB protein did not change at the time points examined. The increased phosphorylation of CREB protein is preceded by an increase in the amount of cAMP, suggestive of cAMP-dependent protein kinase activation, in the hippocampus and activation of Ca2+/calmodulin-dependent protein kinases in the cortex. Subsequent to CREB phosphorylation, the expression of the CRE-containing gene, c-fos, and the AP-1 complexes (heterodimers of Fos and Jun family members) is increased. These findings support the role of CREB-mediated gene expression in activity-dependent neuronal plasticities.

Activity-mediated gene expression is thought to play an important role in many forms of neuronal plasticities. We have used pentylenetetrazol-induced seizure that produces synchronous and sustained neuronal activity as a model to examine the mechanism(s) of gene activation. The transcription factor CREB (Ca 2؉ /cAMP response element-binding protein) is thought to be necessary for long-term memory formation both in invertebrates and vertebrates. When phosphorylated on Ser 133 either by cAMP-dependent protein kinase and/or Ca 2؉ / calmodulin-dependent protein kinases, CREB increases transcription of genes containing the CRE (cAMP response element) sequence. Using an antibody that detects Ser 133 -phosphorylated CREB protein, we show that CREB phosphorylation is maximal between 3 and 8 min after the onset of seizure activity and declines slowly both in the hippocampus and the cortex. The total amount of CREB protein did not change at the time points examined. The increased phosphorylation of CREB protein is preceded by an increase in the amount of cAMP, suggestive of cAMP-dependent protein kinase activation, in the hippocampus and activation of Ca 2؉ / calmodulin-dependent protein kinases in the cortex. Subsequent to CREB phosphorylation, the expression of the CRE-containing gene, c-fos, and the AP-1 complexes (heterodimers of Fos and Jun family members) is increased. These findings support the role of CREB-mediated gene expression in activity-dependent neuronal plasticities.
Neuronal activity plays a critical role in many forms of plasticities such as learning and memory (1,2). Studies on the formation of long-term memory indicate that the induction of immediate-early genes is often associated with memory storage (3,4). The immediate-early gene products are thought to activate late-effector genes, which alter the structural and functional properties of nerve cells (5)(6)(7)(8). How neuronal activity is coupled to alteration in gene expression is poorly understood. However, previous studies have shown that phosphorylation of the transcription factor CREB 1 appeared to be important in mediating the expression of several immediate-early genes (9 -11). Moreover, CREB protein has been implicated in the formation of long-term memory. For example, blockade of CREB by injection of CRE sequence containing oligonucleotides impedes long-term facilitation, a correlate of memory in the marine mollusc Aplysia californica (12,13). "Knock out" mice, with a deletion in the CREB gene, and Drosophila melanogaster, expressing a repressor form of CREB protein, show deficits in long-term memory without any effect on short-term memory (14,15). In addition, overexpression of wild type CREB in Drosophila facilitates the formation of long-term memory (16). These studies indicate a role for CREB-mediated transcription in long-term memory rather than for transient or short-term memory.
The CREB/ATF (activating transcription factor) family of proteins binds to the CRE sequence located in the promoter regions of Ca 2ϩ /cAMP-inducible genes (17)(18)(19)(20). Increases in intracellular cAMP result in the activation of protein kinase A (PKA) by the dissociation of the catalytic from the regulatory subunits. Prolonged increase in cAMP causes translocation of the catalytic subunits into the nucleus (21). It has also recently been shown that isozymes of Ca 2ϩ /calmodulin-dependent protein kinase (CaMK IV and certain isoforms of CaMK II) are localized to the nucleus (22,23). Once activated, PKA or CaMK can phosphorylate CREB on Ser 133 (20, 24 -26). Phosphorylated CREB protein then binds to CREB-binding protein (27,28), and this complex is thought to induce transcription of genes containing CRE sequences such as c-fos (7), tyrosine hydroxylase (29), enkephalin (18), somatostatin (17), and vasoactive intestinal peptide (30).
Many models of neuronal activity, including seizure, produce long-lasting changes in both the structure and function of the brain (31)(32)(33). Several of these studies have demonstrated that c-fos, fra-1, nur/77, c-jun, prodynorphin, enkephalin, and nerve growth factor are induced during stimulation of intact nervous systems, supporting the idea that these genes are important regulators of nerve cell responses in vivo (34 -37). For example, long-term potentiation in the hippocampus, a proposed model for spatial learning and memory, is associated with increased expression of zif/268 (6). Direct electrical stimulation as well as kindling stimulation increase c-fos mRNA (38,39). c-Fos also is increased in the paraventricular and supraoptic nuclei involved in thirst control following 24-h water deprivation (40) and in the dorsal horn neurons of the spinal cord following peripheral sensory stimulation (40,41). Taken together, these in vivo studies are consistent with the idea that alteration in gene expression may underlie longlasting changes in brain function. Many of the genes induced by neuronal activity have CRE-like sequences in their pro-moter regions and therefore are likely to be induced by CREB phosphorylation (7).
Using pentylenetetrazol (PTZ)-induced seizure to induce maximal neuronal activity (42), we have examined the phosphorylation of the CREB protein. We have utilized an antibody that specifically detects CREB when phosphorylated on Ser 133 (10). In this study, we report that the phosphorylation of CREB is maximal between 3 and 8 min after the onset of seizure both in the hippocampus and cortex and declines slowly thereafter. This enhanced phosphorylation of CREB can be reduced by pretreatment with sodium pentobarbital, a GABA receptor agonist which blocks neuronal activity (43). Preceding CREB phosphorylation, we observed an increased activity of autophosphorylated CaMK in the cortex and enhanced concentrations of intracellular cAMP in the hippocampus. Finally, we report that subsequent to CREB phosphorylation, the expression of a CRE-containing gene c-fos and AP-1 complexes (heterodimers of Fos and Jun family members) is increased. These data suggest that neuronal activity stimulates PKA and CaMK, resulting in the phosphorylation of the CREB protein and induction of CRE-containing genes.

EXPERIMENTAL PROCEDURES
Materials-PhosphoCREB and CREB antibodies were purchased from UpState Biotechnology (Lake Placid, NY). c-Fos antibodies were bought from Santa Cruz Biotechnology (Santa Cruz, CA). Induction of Seizure Activity-All procedures involving the use of animals were performed under the guidelines of the National Institutes of Health's Guide for Care and Use of Laboratory Animals and approved by the Institutional Animal Care and Use Committee. Animals were injected intraperitoneally with 55 mg/kg of PTZ prepared in saline (42). Control animals were injected with saline. As an additional control, some animals were intraperitoneally injected with 30 mg/kg of sodium pentobarbital 15 min prior to PTZ injection.
Preparation of Nuclear Extracts-At various times following injection, animals were sacrificed by a guillotine. The cortical (parietal and half of the frontal cortex) and the hippocampal tissues were quickly removed in oxygenated ice-cold artificial cerebrospinal fluid (10 mM HEPES, pH 7.2, 1.3 mM NaH 2 PO 4 , 3 mM KCl, 124 mM NaCl, 10 mM dextrose, 26 mM NaHCO 3 , 2 mM CaCl 2 , and 2 mM MgCl 2 ). Sample preparation was carried out at 4°C. The tissues were separately homogenized (10 strokes) in 5 volumes of a buffer containing 15 mM HEPES, pH 7.2, 0.25 M sucrose, 60 mM KCl, 10 mM NaCl, plus protease inhibitors (1 mM EGTA, 5 mM EDTA, and 1 mM PMSF), and phosphatase inhibitors (2 mM NaF, 2 mM NaPP i , and 5 M microcystin-LR) in a Dounce homogenizer using a loose pestle. The cells were then pelletized at 2,000 ϫ g for 10 min. To lyse the cells, the pelletized material was incubated in 5 volumes of 10 mM HEPES, pH 7.2, 1.5 mM MgCl 2 , 10 mM KCl, 1 mM PMSF, 5 M microcystin-LR, 2 mM NaF, and 2 mM NaPP i for 5 min. The cell suspension was homogenized (7 strokes) in a Dounce homogenizer using a tight pestle. The nuclei were pelletized at 4,000 ϫ g for 10 min. The nuclei were lysed in 1 bed volume of 100 mM HEPES, pH 7.2, 1.5 mM MgCl 2 , 1 mM EDTA, 0.8 M NaCl, 25% glycerol, 2 mM NaF, 2 mM NaPP i , 1 mM PMSF, and 5 M microcystin-LR by gentle rocking for 30 min. Cell debris and genomic DNA were removed by centrifugation at 14,000 ϫ g for 30 min. The supernatant solution was removed and frozen.
Preparation of Samples for Kinase and cAMP Assays-At different time points following the onset of seizure, rats were sacrificed by decapitation. The cortex and hippocampus were immediately dissected while submerged under oxygenated ice-cold artificial cerebrospinal fluid without Ca 2ϩ or Mg 2ϩ . The tissues were homogenized in 10 volumes of 10 mM Tris/HCl (pH 7.4), 1 mM EGTA, 1 mM EDTA, 0.5 mM dithiothreitol, 0.5 mM PMSF, 10 g/ml leupeptine, 0.5 mM isobutylmethylxanthine, and 5 M microcystin-LR. After 20 strokes in a motordriven Teflon-glass homogenizer, 0.6 ml of each sample was added to an equal volume of 12% trichloroacetic acid for cAMP measurements. The rest of the homogenate was aliquoted and centrifuged at 14,000 ϫ g for 30 min at 4°C. The total homogenates, supernatant solutions, and pelletized materials were stored at Ϫ80°C.
For cAMP sample preparation, the 0.6 ml of homogenate mixed with an equal volume of 12% trichloroacetic acid was allowed to sit on ice for 10 min. This was followed by centrifugation at 14,000 ϫ g for 10 min. The supernatant solutions were collected and divided into three aliquots. Aliquots were washed five times with 2.5 volumes of watersaturated ether to remove the trichloroacetic acid. Each wash was performed by vigorous vortexing followed by centrifugation at 14,000 ϫ g for 2 min to separate the two phases. The organic phase was discarded prior to the addition of the next wash. Ether-extracted samples were frozen at Ϫ80°C until ready for use.
Western Blots-For CREB and c-Fos Western blots, nuclear extracts were diluted in water to reduce the salt concentration to approximately 100 mM prior to use. For Western blots examining the ␣ isoform of CaMK II, supernatant solutions were thawed on ice and sonicated briefly to ensure homogeneity. The pellet samples were resuspended in 200 l of homogenization buffer (10 mM Tris/HCl (pH 7.4), 1 mM EGTA, 1 mM EDTA, 0.5 mM dithiothreitol, 0.5 mM PMSF, 10 g/ml leupeptin, 0.5 mM isobutylmethylxanthine, and 5 M microcystin-LR) and sonicated briefly. The amount of protein in each sample was measured using a Micro BCA assay (Pierce). Samples were prepared by boiling in sample buffer, and equal amounts of protein were separated on a 8.5% SDS-polyacrylamide gel electrophoresis. The proteins were transferred to an Immobilon-P membrane (Millipore) using a semi-dry transfer apparatus (Millipore). Western blots were carried out using antibodies at a final concentration of 0.2 g/ml as described previously (44). The blots were visualized using an alkaline phosphatase chemiluminescence detection system (Life Technologies, Inc.), which allowed the production of multiple fluorographs with different lengths of exposure. The quantification of the immunoreactive bands was carried out utilizing a Bio-Rad model GS-670 imaging densitometer. The Western blots were repeated in at least four independent experiments.
Gel Retardation Assay-Equal amounts of the nuclear extracts were used for gel retardation assays using the 12-O-tetradecanoylphorbol-13-acetate response element (TRE) sequence as a probe for measuring the AP-1 binding. The double-stranded oligonucleotide probe was synthesized using the metallothionein gene TRE sequence 5Ј-GATCTGT-GAGTCAGCGCGA-3Ј and its complement sequence (45). Probe preparation and assays were carried out essentially as described previously (46). For competition with oligonucleotides, the cold competitors were added to the reaction mixture prior to the addition of protein extracts. The oligonucleotides used for competition were 5Ј-GATCATCTCAATT-AGTCAGCAA-3Ј and its complement for TRE, and 5Ј-GGAGCGCGC-CTCGAATGTTCTAGAAAAGGCTGCA-3Ј and its complement for the heat shock element. Supershift assays were performed by preincubating the nuclear extracts with 0.2 g of c-Fos antibodies in the absence of any thiol reagent for 15 min at room temperature. The resulting antibody-antigen complex was then incubated with the probe.
Immunohistochemistry-Hippocampi were removed quickly as described under "Preparation of Nuclear Extracts." The tissues were fixed in cold 4% paraformaldehyde and 15% picric acid in phosphate-buffered saline (PBS) for 8 -10 h. The tissue was then placed in 30% sucrose in PBS overnight at room temperature. Tissues were briefly dried to remove any surface liquid and mounted in optimal cutting temperature (O.C.T., Miles Inc.) cryostat imbedding medium. 50-micron slices were prepared using a cryostat. Slices were washed twice with PBS and incubated with 5 g/ml primary antibody in 2% bovine serum albumin, 0.3% Triton X-100, and 1.5% normal goat serum in PBS at room temperature overnight. Slices were washed four times with PBS, and immunoreactivity was detected using an avidin-biotin detection system from Vector Laboratory essentially as suggested by the supplier.
Kinase Assay-Supernatant and total homogenate samples were thawed on ice and sonicated briefly to ensure homogeneity. Pellet samples were resuspended and sonicated as described under "Western Blots." The protein concentration in each sample was determined using Micro BCA assays with bovine serum albumin as the standard. CaMK activity was measured using a synthetic peptide substrate (autocamtide-3; KKALHRQETVDAL) (47,48). Phosphorylation reactions were initiated by adding 3 g of protein to a 20-l mixture resulting in the following final concentrations: 10 mM HEPES (pH 7.4), 0.5 mM dithiothreitol, 20 M substrate peptide, 50 M ATP, 3 mM EGTA, 5 mM MgCl 2 , and 3 Ci of [␥-32 P]ATP (3,000 Ci/mmol). After a 1-min incubation at 30°C, the reactions were terminated by spotting 15 l of the reaction mixture on P-81 phosphocellulose filters (Whatman). The filters were washed three times, 10 min each, in 75 mM phosphoric acid, rinsed with ethanol, and dried under air flow. The radioactivity on the phosphorylated peptide substrate was quantitated in a scintillation counter by the Cerenkov method.
Measurements of cAMP-cAMP was measured using a radioimmunoassay kit purchased from Amersham. Frozen samples were completely dried in a vacuum centrifuge at ambient temperature. Dried samples were resuspended in diluted assay buffer as recommended by the vendor. cAMP measurements were performed utilizing the acetylation method and compared to a standard curve, which was prepared simultaneously. The amount of cAMP in each sample was normalized to the total amount of protein in the homogenate prior to trichloroacetic acid precipitation. The results from the three aliquots were averaged for each hippocampal or cortical sample. cAMP measurements were repeated in at least three independent experiments.
Statistical Analysis-An analysis of variance was computed to detect a significant main effect for groups (p Ͻ 0.05). This was followed by a Tukey's b post-hoc test to identify specific group differences. Statistical analyses were performed using the integrated optical densities (Western blots), counts per minute (kinase assays), or by comparison to standard curves (cAMP measurements). The data are presented as percent control.

RESULTS
Injection of PTZ Results in the Production of Seizure Activity-PTZ-induced seizure has been used as a model for robust neuronal activity (42,58). As described under "Experimental Procedures," seizure activity was initiated by an intraperitoneal injection of PTZ at a dose of 55 mg/kg body weight (42). We consistently observed the physical manifestations of seizure activity in all animals approximately 58 s after injection of the drug. As previously reported, the activity began with acute facial and forelimb contraction progressing to animals running and bouncing as well as tonic convulsion (49). No seizure activity was observed in either the saline or the sodium pentobarbital controls.
Seizure Activity Increases the Phosphorylation of CREB Protein-Phosphorylation of the CREB protein following PTZ injection was examined using an antibody that detects Ser 133 -phosphorylated CREB protein (10). As an initial study, nuclear extracts were prepared 8 min after the injection of PTZ or saline, and the levels of total CREB and phosphorylated CREB were determined by Western blots. Fig. 1A shows that total amount of CREB protein remained unchanged by seizure activity both in the cortex and in the hippocampus. In contrast, the phosphorylation of CREB in the cortex and hippocampus is enhanced at this time point in experimental animals (Fig. 1B). In addition to the 43-kDa phosphoCREB, weakly cross-reacting bands were detected at 40 kDa and around 100 kDa. The identity of these cross-reactive bands is unknown at present. When animals were injected with 30 mg/kg of sodium pentobarbital prior to PTZ injection to block neuronal activity, the increased phosphoCREB signal was reduced both in the hippocampus and the cortex (Fig. 1, B and C). In the cortical samples, the signal for phosphoCREB was lower in comparison to saline controls. However, this difference was not statistically significant. In addition, no significant difference was detected between naive or saline-injected animals (data not shown). Fig.  1C shows the summary of results from several such experiments. The signal for phosphoCREB is increased by approximately 62% (saline ϭ 99.5 Ϯ 16.6%, experimental ϭ 161.8 Ϯ 34.1%, p Ͻ 0.05) in the hippocampus and by approximately 76% (saline ϭ 99.7 Ϯ 13.3%, experimental ϭ 175.8 Ϯ 13.3%, p Ͻ 0.05) in the cortex of experimental animals compared to salineinjected animals. Injection of sodium pentobarbital 15 min prior to PTZ injection blocked the enhanced CREB phosphorylation in both the hippocampus and cortex (Fig. 1C).
Time Course of CREB Phosphorylation-We next examined the time course of CREB phosphorylation. Fig. 2A shows that phosphorylation of CREB protein is increased in the hippocampus as early as 3 min after the onset of seizure activity. The phosphorylation was maximal between 3 and 8 min and declined slowly. At 60 min following the onset of seizure, the phosphorylation was higher compared to the control. A similar time course for CREB phosphorylation was observed in the cortical tissue. The total amount of CREB protein remained unchanged both in the cortex and hippocampus at these time points (Fig. 2B).
Spatial Pattern of CREB Phosphorylation-The spatial pattern of CREB phosphorylation in the hippocampus was investigated by immunohistochemistry. Animals for this study were When examined under high magnification, the signal for CREB immunoreactivity was found to be localized to the nucleus (data not shown). Consistent with the Western blot data, immunohistochemistry did not show any difference in total CREB protein between the control and experimental hippocampi. No staining was detected when the primary antibody was eliminated from the incubation solution (data not shown). Fig. 3B shows the low level of phosphorylated CREB-like proteins present in a control hippocampus. Seizure activity increases the phosphorylation of CREB-like proteins in the CA1, CA3 subfields and to a lesser extent in the dentate gyrus.
Increase of cAMP in the Hippocampus Following PTZ Injection-CREB protein when phosphorylated on Ser 133 either by PKA or CaMK can induce transcription of CRE-containing genes (20,24,26). To examine PKA activation, the amount of cAMP in both the cortical and hippocampal tissues at various time points following the onset of seizure activity were measured using a radioimmunoassay kit as described under "Experimental Procedures." It has been reported that increased levels of cAMP correlate well with enhanced PKA activity (50). The amounts of cAMP in the hippocampal and cortical tissue of a control animal were found to be 23.9 Ϯ 0.35 pmol/mg of protein and 33.3 Ϯ 0.75 pmol/mg of protein, respectively. In the hippocampus, the amount of cAMP was maximal 3-4 min after the onset of seizure and returned to control level by 8 min (Fig. 4A) (control ϭ 100.3 Ϯ 1.7%, 3 min ϭ 138.7 Ϯ 7.0%, 4 min ϭ 134.0 Ϯ 6.8%, p Ͻ 0.05). In the cortex, the level of cAMP did not change significantly over time (Fig. 4B).
Activation of Ca 2ϩ /Calmodulin-dependent Protein Kinase-The highly abundant, brain-specific CaMK II is activated in response to increases in intracellular calcium concentrations (51)(52)(53). Once activated, CaMK II undergoes autophosphorylation resulting in a Ca 2ϩ /calmodulin-independent form of the enzyme (51-53). We measured the Ca 2ϩ /calmodulin-independent activity in the presence of 3 mM EGTA to detect the autophosphorylated enzyme in the total homogenate, supernatant, and pellet fractions. The specific activities of CaMK in the hippocampal total, supernatant, and pelletized fractions (in nmol of phosphate transferred/min/mg protein) of control animals were calculated to be 9.54 Ϯ 2.18, 41.59 Ϯ 4.73, and 2.96 Ϯ 0.77, respectively. In the cortex, the specific activities for CaMK (in nmol of phosphate transferred/min/mg protein) of control animals in the total, supernatant, and pelletized fractions were measured to be 6.02 Ϯ 1.72, 43.92 Ϯ 9.15, and 3.90 Ϯ 0.72, respectively. The increased specific activity of soluble CaMK compared to the total has been reported previously (48). However, the reason for this increase is not yet known. Fig. 5A shows that there was no significant change in the Ca 2ϩ /calmodulin-independent activity of the hippocampal total homogenate in response to PTZ injection. However, a significant increase in the activity was detected in the cortical total homogenate 3 min following the onset of seizure (control ϭ 100.0 Ϯ 20.1%, 3 min ϭ 216.3 Ϯ 33.6%, p Ͻ 0.05) (Fig. 5A). To further examine which subcellular fraction contained the activated CaMK, supernatant and pellet fractions were used for activity measurements. Consistent with the total homogenate data, no significant changes were detected in either the hippocampal supernatant or pellet fractions (Fig. 5, B and C). When the cortical samples were assayed, an increase in Ca 2ϩ /calmodulin-independent activity was detected using the supernatant fraction (control ϭ 100.1 Ϯ 21.8%, 3 min ϭ 179.0 Ϯ 20.4%, p Ͻ 0.05) (Fig. 5B). No significant change was detected in the cortical pellet fractions (Fig. 5C). Control experiments show that the Ca 2ϩ /calmodulinindependent activity could be blocked by more than 99% when the extract was preincubated with 20 M of a CaMK-specific inhibitory peptide (55) with the sequence MHRQEAVD-CLKKFNARRKLKGA (data not shown).
Western Blot Analysis of ␣CaMK II-The ␣ subunit of CaMK II is the predominant brain-specific calcium-calmodulin-dependent protein kinase (54). It has previously been reported that ischemic insults can cause translocation of CaMK II to a particulate fraction resulting in a decrease in specific activity (55). To examine if CaMK II was changing its subcellular distribution (which could contribute to changes in activity), Western blots were carried out using a monoclonal antibody to the CaMK II ␣ subunit (48,56). Fig. 6, A and B, shows representative photographs of Western blots for the supernatant and pellet fractions, respectively. No significant changes in signal intensity were detected in either the hippocampal or cortical samples.
Induction of Late-effector Genes-Once phosphorylated, CREB protein induces the transcription of CRE-containing genes. To investigate if CRE-containing genes are induced following PTZ-induced seizure, we next examined the level of the AP-1 complex, which is a heterodimer of Fos and Jun family members. Consistent with previous reports (57), we found that the binding of the AP-1 complex to an oligonucleotide containing the TRE sequence was enhanced 3 h following PTZ injection (Fig. 7A). The enhanced expression of the AP-1 complex was detected both in cortical and hippocampal nuclear extracts. The enhanced binding could be specifically blocked by the addition of an excess amount of unlabeled TRE-containing oligonucleotides. The addition of an excess amount of unlabeled heat shock element-containing oligonucleotides had little effect on AP-1 binding (Fig. 7B). To determine if c-Fos was present in the retarded band, supershift assays using a c-Fos antibody were carried out. Fig. 7C shows that the addition of 0.5 g of the c-Fos antibody resulted in the appearance of a new slower migrating band and a concomitant decreased intensity of the original retarded band. Control experiments without the c-Fos antibody, using a CREB antibody, or using the samples from saline-injected animals did not produce any specific supershifted bands (Fig. 7C and data not shown). A new, slower migrating band was seen in the samples incubated with the CREB antibody. However, this band was also present in the absence of nuclear extracts, and its migration was different than that of the supershifted material when the c-Fos antibody was used (Fig. 7C). In addition to the supershift, Western blot analysis confirmed that the c-Fos protein was induced in the 3-h PTZ cortical sample (Fig. 7D). The antibody also crossreacted with two unidentified bands in the control and experimental samples, which do not change as a result of seizure activity. Similar results for c-Fos induction were obtained with either the cortical or hippocampal samples. These findings suggest that the increase in the AP-1 binding is due at least in part to the induction of c-fos, a CRE-containing gene. DISCUSSION An understanding of how neuronal activity modifies gene expression may help elucidate mechanisms of long-term neu-ronal plasticities in the brain. Using chemically induced seizure as a model for neuronal activity (42,58), we have identified some of the early biochemical changes that may be responsible for gene induction. Both in the cortex and hippocampus, the phosphorylation of CREB protein is one of the initial events that may participate in gene activation.
In this study, we show that the phosphorylation of CREB on Ser 133 is maximal between 3 and 8 min after the onset of robust activity both in the hippocampus and cortex. This time course of CREB phosphorylation is consistent with the reports that phosphorylation of CREB in the suprachiasmatic nuclei during light exposure, in the striatum following administration of amphetamine and in the hippocampus following cortical impact injury, is maximally phosphorylated between 5 and 15 min (10,11,59). Our immunohistochemical studies indicate that the increase in hippocampal CREB phosphorylation as a result of chemically induced seizure is predominantly located in the pyramidal cells of the CA3 and CA1 subfields. This is in agreement with the previous reports that activity in the CA1 and CA3 pyramidal cells plays a major role in temporal lobe seizures (60). The increase in CREB phosphorylation detected by Western blots can be blocked by sodium pentobarbital injection prior to PTZ administration. Sodium pentobarbital is a reported GABA agonist that inhibits neuronal activity (43). Thus, the enhanced phosphorylation of the CREB protein is most likely due to the neuronal activity induced by PTZ.
In the hippocampus, we observed that CREB phosphorylation was preceded by an increase in cAMP levels. This is consistent with the finding that electrical stimulation increases cAMP levels in hippocampal slices (61). It has been previously reported that increases in cAMP lead to CREB phosphorylation on Ser 133 via activation of PKA (9). This suggests that the activation of PKA is likely to be responsible for the CREB phosphorylation we observed in the hippocampus. Stimulation of adenylyl cyclase either by Ca 2ϩ /calmodulin and/or by receptor-coupled G proteins increases intracellular concentrations of cAMP (62). Neuronal activity increases intracellular calcium via opening of voltage-gated and transmitter-gated calcium channels (63,64), which would activate calcium-sensitive adenylyl cyclases. It has been observed that calcium-stimulated adenylyl cyclase type I and type VIII are highly enriched in hippocampal neurons (62). It is possible that the increase in cAMP levels we observed in the hippocampus following PTZ administration is due to activation of calcium-sensitive adenylyl cyclase. In addition, the activation of receptor-coupled G proteins during seizure activity in conjunction with Ca 2ϩ /calmodulin could synergistically stimulate adenylyl cyclase (65). Interestingly, we did not detect any activation of a second Ca 2ϩ /calmodulin stimulated enzyme, CaMK, in the hippocampus. However, it has been reported that CaMK II has weaker binding affinity (K d ϭ 45 nM versus 15 nM for adenylyl cyclase) for the Ca 2ϩ /calmodulin complex (66). Future experiments will help determine if calcium-sensitive adenylyl cyclases are indeed activated as a result of PTZ administration.
In contrast to the hippocampus, we observed no significant change in the amount of cAMP in the cortex. The low levels of calcium-stimulated cyclase in the cortical tissue may account for our inability to detect increased cAMP following PTZ administration (67). However, a significant increase in the activity of CaMK was detected. Increases in CaMK activity were detected in both the total homogenate and supernatant fractions as early as 3 min following the onset of seizure activity (Fig. 5, A and B). Interestingly, no change in the activity of the pelletized form of CaMK was observed. Moreover, Western blot analysis did not reveal any change in the levels of ␣CaMK II (Fig. 6, A and B). These results suggest that the increase in CaMK activity we observed is due to autophosphorylation of soluble CaMK rather than a translocation of the insoluble form. The increase in the cAMP level in the hippocampus and the activation of CaMK in the cortex return to control values 8 min after the onset of seizure. However, the CREB phosphorylation remains higher than the control value as long as 60 min. Decreased phosphatase activity such as that of protein phosphatase I, which is known to dephosphorylate CREB (68), or increased activity of other kinases (e.g. protein kinase C) could contribute to this prolonged CREB phosphorylation.
The temporal association between kinase activity and CREB phosphorylation suggests that PKA and CaMK phosphorylate CREB in the hippocampus and the cortex, respectively, during seizure activity. However, due to the nature of our biochemical assays to measure changes in cAMP and CaMK, we are unable to localize these changes to any specific cell type or cellular compartment. Localized increases in PKA or CaMK activity may be more important in terms of gene regulation than overall changes in activity measured in tissue homogenates. For example, several protein kinases bind to specific anchoring proteins (e.g. PKA binds to AKAP (a kinase anchoring protein)), which are localized to specific subcellular compartments allowing cells to respond in disparate ways to distinct extracellular stimuli (69). In addition, it has been reported that calcium influx through L-type voltage-sensitive calcium channels activates immediate-early genes while calcium influx via NMDA subtypes of glutamate receptors can lead to both apoptotic and necrotic cell death possibly via the induction of different sets of genes (70 -72).
As stated previously, PKA phosphorylates CREB protein on Ser 133 resulting in its activation. CaMK II has also been reported to phosphorylate CREB on Ser 133 as well as on Ser 142 (20). In vitro transcription experiments show that when phosphorylated on Ser 133 by CaMK II, CREB can increase the transcription of a CRE-containing reporter gene (20,25). However, recent experiments using cell lines overexpressing CaMK II show that phosphorylation of Ser 142 may suppress transcriptional activation by CREB (26,73). It has also been reported FIG. 7. Gel-retardation assays for AP-1 binding and Western blots for c-Fos. Animals were injected with either saline or PTZ and sacrificed 3 h after the onset of seizure activity. A, the amount of AP-1 complexes in the hippocampus and the cortex. B, the specificity of AP-1 binding was determined by competition with unlabeled oligonucleotides containing either the heat shock element or TRE sequence as described under "Experimental Procedures." C, supershift assays using either 0.5 g of c-Fos or CREB antibody and the PTZ 3-h cortical sample. The arrowhead indicates the migration of the supershifted band. Control experiments without the nuclear extract are also shown. D, photograph of a representative Western blot using cortical samples from saline and PTZ-injected animals and the same c-Fos antibody employed for supershift assays. The 62-kDa c-Fos protein is indicated by the open arrowhead. Similar results for both the supershift assays and Western blots were obtained using the hippocampal samples. that phosphorylation of the residue Ser 133 of CREB by CaMK IV enhances its transcriptional activity (26,73). It is possible that several isoforms of CaMK are activated in response to neuronal activity. Unfortunately, the kinase assay system we have used may not distinguish between isoforms of CaMK. Therefore, we cannot evaluate the contribution of CaMK IV to the increase in CaMK activity we observed. Moreover, the antibody we have employed to detect phosphoCREB recognizes only the phosphorylation of Ser 133 . We therefore cannot determine if Ser 142 is phosphorylated in vivo during seizure activity.
Subsequent to CREB phosphorylation, the expressions of the AP-1 complex and c-Fos were enhanced in both the hippocampus and the cortex (Fig. 7, A and D). Due to multiple covalent post-translational modifications, c-Fos migrates as a smeared band around 62 kDa (Ref. 74 and Fig. 7D). The presence of the c-Fos protein in the AP-1 complex was identified by supershift assays using a c-Fos antibody. However, the intensity of the supershifted band did not appear to be proportional to the decrease seen in the originally retarded band. This suggests that the c-Fos antibody, in addition to causing a supershift, also affects binding of the AP-1 complex to the oligonucleotide probe. Since all the original band was not supershifted by the c-Fos antibody, other c-Fos-related proteins, such as Fra(s) (Fos-related antigens), could be induced as a result of seizure and contribute to the total AP-1 signal we detected in Fig. 7A (57).
At present, we cannot establish a causal link between kinase activation, CREB phosphorylation and AP-1 induction. For example, the c-fos gene and the AP-1 complexes can be induced by multiple signal transduction pathways (8). However, the temporal relationships between kinase activation, CREB phosphorylation, and the subsequent induction of c-Fos and the AP-1 complex suggest that CREB phosphorylation is important for activity-dependent gene expression. Future studies using specific kinase inhibitors or antisense oligonucleotides may help in identifying the causal relationships between kinase activation, CREB phosphorylation, and the subsequent increases in c-Fos and AP-1 complexes. Furthermore, identification of effector genes induced by phosphorylated CREB or the AP-1 complexes following neuronal activity would help elucidate the molecular basis of long-term plasticities in the brain.