Expressions of CCAAT/Enhancer-binding Proteins β and δ and Their Activities Are Intensified by cAMP Signaling as Well as Ca2+/Calmodulin Kinases Activation in Hippocampal Neurons*

The transcription factor, AplysiaCCAAT enhancer-binding protein (ApC/EBP), plays a crucial role in long term facilitation, a synaptic mechanism of long term memory inAplysia. To gain a clue to whether the mammalian C/EBP family of transcription factors are also involved in long term memory, we examined how C/EBP activities in hippocampal neurons can be modulated in response to cAMP and Ca2+, crucial inductive signals for memory formation. As a result, stimulation of either cAMP or Ca2+ signals in hippocampal neurons was found to enhance mRNA expressions and DNA binding activities of C/EBPβ and C/EBPδ. Furthermore, it is indicated that CaM kinases have essential roles for increasing the expression and DNA binding activities of C/EBPβ in hippocampal neurons activated by membrane depolarization. Overexpression of constitutively active calcium/calmodulin-dependent kinase IV was found to directly stimulate either C/EBPβ-dependent or C/EBPδ-dependent transcription, reinforcing the evidence that C/EBP family members contribute to Ca2+-dependent transcription. Thus, these results suggest that C/EBPβ and C/EBPδ may be involved in the transcription-dependent phase of memory formation by increasing the expression of both the DNA binding and the transcriptional activities under the direction of cAMP and/or Ca2+signaling in hippocampal neurons.

Memory has two phases: short term memory and long term memory. A number of pharmacological studies have demonstrated that the stabilization of long term memory requires the synthesis of new proteins and RNAs (1,2). This requirement suggests that transcription is critical for this process. Among a number of transcription factors expressed in neuronal cell nuclei, the Ca 2ϩ /cAMP-responsive element-binding protein (CREB) 1 has been implicated as being essential to long term memory of Aplysia, Drosophila, and mice (3)(4)(5)(6)(7). In Aplysia, long term facilitation is a basic synaptic mechanism for a nonassociative learning response. The long term synaptic modification is characterized by a consolidation period during which gene expression is required. During this phase, the transcription factor Aplysia CCAAT enhancer-binding protein (ApC/ EBP) is found to be increasingly induced in response to cAMP signals (8). There is a CRE site in the 5Ј-untranslated region of ApC/EBP, suggesting that CREB can control the expression of ApC/EBP (8). Furthermore, blocking the function of ApC/EBP either by antisense oligonucleotides or by specific antibody inhibits long term facilitation selectively without affecting the short term processes (8). These data indicate that induction of another transcription factor ApC/EBP by CREB is also essential for long term facilitation, in addition to CREB activation (8).
In mammals, seven members of C/EBP family of transcription factors have been cloned molecularly and their biological roles extensively analyzed in a variety of systems (9 -26). However, much less is known about the expression and function of C/EBP family members in the neurons of mammalian brain, except that C/EBP␣ mRNA was detected in mouse hippocampus, cerebellum, and cortex by in situ hybridization (27). It was recently shown that two CRE sites in the core promoter of C/EBP␤ gene are mostly recognized by CREB in liver and in several cell lines (28). Transfection experiments with promoter constructs where the CREB sites were mutated further indicated that these sites are important to maintain both basal promoter activity and C/EBP␤ inducibility through CREB (28). However, in several brain regions including the hippocampus, little information is available about the induction of C/EBP family members by cAMP and Ca 2ϩ , both of which are known to activate CREB in neuronal cells such as hippocampal neurons (29 -33).
The hippocampus in mammalian brain is involved in the normal formation of long term declarative memory (34). Here, we have found that a prime C/EBP transcription factor in the hippocampus is C/EBP␤. To clarify whether C/EBP␤ activity in hippocampal neurons is further modulated by the stimulation of cAMP or Ca 2ϩ signals, which are crucial inducers for memory formation (3)(4)(5)(6)(7)(35)(36)(37)(38)(39), we used cultured hippocampal neurons for detailed biochemical analyses (40). We have found that the expression and DNA binding activities of C/EBP␤ and ␦ are enhanced by the stimulation of cAMP or Ca 2ϩ signals in cultured hippocampal neurons. Our results also suggest that CaMKIV activated by Ca 2ϩ signal not only induces expression of C/EBP members, but also directly enhances C/EBP-dependent gene transcriptions. Therefore, our study supports the possibility that both C/EBP␤ and C/EBP␦ may be involved in long term plasticity in mammalian brain.

EXPERIMENTAL PROCEDURES
Hippocampal Neuron Cultures-Hippocampal neuron cultures were done as described by Baranes et al. (40). To stimulate cAMP signals, water-soluble forskolin (50 M; Research Biochemicals International) was added to 14-day-old cultures. To induce membrane depolarization of hippocampal neurons in the culture, the cultures were treated as described by Bito et al. (33). When used, kinase inhibitor KN93 (30 M, Calbiochem), KN92 (30 M, Calbiochem), and KN62 (10 M, Calbiochem) were present during the preincubation period of 30 min prior to depolarization and during depolarization. Effects of cAMP signal stimulation and membrane depolarization on gene expressions or DNA binding activities were measured 3-4 h after stimulation.
Immunofluorescence Analysis-Rat hippocampal neurons were grown on poly-D-lysineand laminin-coated 12-mm glass coverslips in 35-mm dishes (Corning) under the same conditions described by Baranes et al. (40). Cells were washed with Tyrode's solution (37°C) once and fixed with 4% paraformaldehyde solution for 15 min at 37°C. They were then washed with PBS three times and permeabilized with 0.25% Triton X-100 in PBS at 37°C for 5 min, and the nonspecific bindings were blocked with 10% goat serum, 0.1% Triton X-100, 20 mM glycine in PBS at 37°C for 30 min. Double immunolabeling was performed by incubating cells overnight with a mouse monoclonal anti-MAP2 antibody (Sigma, 1:100) and rabbit polyclonal anti-C/EBP␤ or anti-C/EBP␦ antibodies (Santa Cruz, 1 g/ml). Following three washes with PBS, the cells were incubated with Cy3-conjugated goat antimouse IgG (Cedarlane Laboratories Ltd, 1:100) and fluorescein-conjugated goat anti-rabbit IgG (Cedarlane Laboratories Ltd, 1:100) in PBS at room temperature for 1 h. The coverslips were washed five times with PBS, mounted, and examined by fluoromicroscopy.
(Sp)-cAMPS Treatment of Hippocampal Slice-(Sp)-cAMPS stimulation of hippocampal slices was done as described by Huang et al. (41). Hippocampal slices were incubated in perfusion solutions containing 50 M (Sp)-cAMPS ((Sp)-cAMPS stimulation) or perfusion solutions without (Sp)-cAMPS (control) for 30 min. After 30 min of incubation, slices were maintained in the perfusion solution for 2 h. Then, RNAs were prepared from those slices for RNase protection assay.
RNase Protection Assay-A 220-bp rat C/EBP␤ cDNA NcoI fragment, 342-bp mouse C/EBP␤ cDNA NcoI-PstI fragment covering the leucinezipper domain, 298-bp mouse C/EBP␦ cDNA NcoI-XhoI fragment, and 490-bp rat C/EBP␣ cDNA NcoI-NotI fragment were subcloned into pBluescript SK (Stratagene). Templates for preparing Zif268 cRNA (180 bp is protected), actin cRNA (250 bp), and cyclophilin cRNA probe (100 bp) were purchased from Ambion. Antisense cRNA probes were synthesized as described previously (42). Total RNAs were prepared from tissues or cultured hippocampal neurons using the acid phenol extraction method of Chomczynski and Sacchi (43) with RNAzol B solution purchased from Biotecx Laboratories, Inc. and performed according to manufacturer's protocol. RNase protection assays were performed by hybridizing 25 g of total RNA from the hippocampus or 2.5 g of total RNA from cultured hippocampal neurons with 1 ϫ 10 5 cpm of each labeled cRNA at 68°C for more than 1 h. RNase digestion and analysis were performed as described previously (42).
Immunoblot Analysis-For Western blot analysis, nuclear extracts prepared from cultured hippocampal neurons for gel shift assay were used. Twenty micrograms of each sample were adjusted to give a final solution of 60 mM Tris-HCl, pH 6.8, 2% SDS, 10% glycerol, 0.1% bromphenol blue, and 5% ␤-mercaptoethanol, heated at 100°C for 5 min, electrophoresed through 10% SDS-polyacrylamide gel, and transferred to polyvinylidene difluoride membrane (Millipore). C/EBP␤ was detected with the ECL Western blotting detection system as instructed by the manufacturer (Amersham). The dilution factor for anti-C/EBP␤ antibody was 1:100 (Santa Cruz).
Gel Shift Assay-Nuclear extracts of mice hippocampus were prepared as described by others (30) with some modifications. The tissues were homogenized (10 strokes) in four volumes of a buffer containing 0.25 M sucrose, 15 mM Tris-HCl, pH 7.6, 60 mM KCl, 15 mM NaCl, 5 mM EDTA, 1 mM EGTA, and protease inhibitors (1 ϫ ␣-Complete from Boehringer Mannheim, 0.5 mM phenylmethylsulfonyl fluoride, and 1 mM dithiothreitol) in a Dounce homogenizer. After centrifugation, the pelletized material was resuspended in buffer A (10 mM HEPES, pH 7.9, 10 mM KCl, 0.1 mM EDTA, 0.1 mM EGTA, 1ϫ ␣-Complete, 0.5 mM phenylmethylsulfonyl fluoride, and 1 mM dithiothreitol) by gently pipetting and kept on ice for 15 min. Nonidet P-40 was added to reach a final concentration of 0.6% and vortexed for 10 s. After centrifugation for 60 s, the nuclei were resuspended in 70 l of buffer C (20 mM HEPES, pH 7.9, 25% glycerol, 0.4 M NaCl, 1 mM EDTA, 1 mM EGTA, 1ϫ ␣-Complete, 0.5 mM phenylmethylsulfonyl fluoride, and 1mM dithio-threitol) and rocked on an Eppendorf shaker for 20 min at 4°C. The supernatant solution after centrifugation was frozen as nuclear extract. To prepare nuclear extracts from hippocampal neurons in culture, cells were washed with Tyrode's solution once, then immersed with 300 l of solution A, scraped, and collected into Eppendorf tubes. The following procedure is identical to the above for preparation of tissue nuclear extracts. Protein amounts were quantified by the colorimetric determination using bicinchoninic acid solution (Sigma). For electrophoretic mobility shift assay (EMSA), C/EBP consensus oligonucleotide, 5Ј-TG-CAGATTGCGCAATCTGCA-3Ј (Santa Cruz), and the core promoter sequence of C/EBP␤ gene, 5Ј-GCGGCCGGGCAATGACGCGCACCG-3Ј (28), were 32 P-labeled with [␥-32 P]ATP, using T4 polynucleotide kinase. Labeled DNA was incubated with 5 g of nuclear extract protein at room temperature for 20 min in the binding buffer containing 10 mM Tris, pH 7.5, 50 mM NaCl, 1 mM EDTA, 5% glycerol, and 1 g of poly(dI-dC). DNA-protein complexes were resolved as described previously (44). Supershift assays were performed as described above with the exception that, after the incubation of probes with extracts, 1 g of antibody (TransCruz gel supershift antibodies to C/EBP␣, ␤, ␦, CRP1, Rb, CREB1, ATF3, and ATF4; Santa Cruz) was added and incubated for another 30 min at room temperature.
Luciferase Assay-Expression vectors for C/EBP␤ and ␦ (gift from Dr. P. F. Johnson) and luciferase reporter vector with four tandem C/EBP binding sites were transfected into HeLa cells with or without the construct for expressing the constitutively active form of CaMKIV (45) through CaPO 4 method (Transfection MBS mammalian transfection kit, Stratagene). Triplicate transfections were done in three independent experiments. Two days after transfection, cells were harvested for luciferase assay and their luciferase activities were examined using the Dual-luciferase reporter assay system (Promega). In the system, the luciferase activities of the reporter were normalized to the Renilla luciferase activities from 20 ng of the cotransfected pRL-SV40 (Promega). The experimental results were represented as the mean Ϯ S.E. of the values, which were calculated by comparing the activities of cotransfection experiments with those obtained by the transfections of the luciferase reporter vector alone.

C/EBP␤ Is a Prime C/EBP in the Mouse Hippocampus-To
know whether any C/EBP family member is expressed in the hippocampus, which is a crucial region for explicit memory formation in vertebrates, we examined the mRNA expressions of C/EBP members in hippocampus by performing RNase protection assay. The result indicated that C/EBP␤ and ␦ transcripts are expressed in the mouse hippocampus (Fig. 1A). It is also shown that C/EBP␤ mRNA is more highly expressed in the mouse hippocampus than C/EBP␦ mRNA (Fig. 1A).
Next, we asked which members of C/EBP family is most active in mouse hippocampus by performing EMSA using high affinity C/EBP binding site as a probe. A retarded band was observed in mouse hippocampus nuclear extracts (Fig. 1B). With the addition of anti-C/EBP␤ antibodies to the binding reaction, the result was a prominent supershift of the retarded band, which was removed by the addition of cold C/EBP binding sites as a competitor (Fig. 1B). Antibodies to C/EBP␦ and ␣ also induced supershifts in EMSA of hippocampus, which are much weaker than the supershift induced by anti-C/EBP␤ antibodies (Fig. 1B). Addition of the antibodies to CRP1, another newly identified C/EBP member to the binding reaction could not induce any supershift in EMSA suggesting that CRP1 is not a component of hippocampal C/EBP (Fig. 1B). Thus, we find that C/EBP␤ is a prime C/EBP in the mouse hippocampus.
C/EBP␤ and ␦ Are Expressed in Cultured Hippocampal Neurons-To examine whether C/EBP␤ is exactly expressed in hippocampal neurons, double immunofluorescence analyses were performed on cultured hippocampal neurons using antibodies to C/EBP family members and anti-MAP2 antibody to locate neurons. As a result, C/EBP␤ was found to be specifically expressed in hippocampal neurons since positive staining was only detected in the nuclei of MAP2 positive cells (Fig. 2, A and  B). C/EBP␦ is also found to be expressed in some hippocampal neurons (Fig. 2, C and D). However, other neurons scarcely express C/EBP␦. C/EBP␦ is found to be rather highly expressed in astrocytes, which are easily distinguished from neurons by remaining unstained by MAP2 (Fig. 2, C and D). This fact is consistent with the previous findings of others (46).

C/EBP␤ Can Be a Downstream Target Gene of CREB in Mouse Hippocampus and Cultured Hippocampal
Neurons-A small region containing two CREB sites in the C/EBP␤ promoter is found to be important in controlling transcription of C/EBP␤ gene in liver and several cell lines (28). To examine the possibility that CREB can also control the expression of C/EBP␤ even in hippocampus and cultured hippocampal neurons, EMSA of nuclear extracts from hippocampus and cultured hippocampal neurons were performed using the region containing first CREB site of C/EBP␤ promoter as a probe. Several retarded bands could be observed in this EMSA suggesting that nuclear factors are actually binding to the region of the C/EBP␤ promoter. Furthermore, the addition of CREB antibody to the EMSA reaction could induce a supershift of the major retarded band demonstrating that main binding activity to the C/EBP␤ promoter region is due to CREB (Fig. 3). Further addition of cold C/EBP␤ promoter probes as competitors weakened the intensity of supershifted bands, thereby demonstrating specific binding reactions in EMSA. The addition of antibodies to other ATF/CREB family members, such as activating transcription factor 3 (ATF3) and activating transcription factor 4 (ATF4), into the EMSA binding reaction could not supershift retarded bands. These results suggest that CREB binds to the crucial region of C/EBP␤ promoter and controls the expression of C/EBP␤ gene in hippocampal neurons.
Stimulation of cAMP Signaling Pathway in Hippocampal Neurons Increases Both mRNA Expressions and DNA Binding Activities of C/EBP␤ and ␦-To know whether CREB activated by cAMP signaling can actually control the expression of C/EBP␤ in hippocampal neurons, cultured hippocampal neu-rons were stimulated by the direct adenylate cyclase activator forskolin and changes of C/EBP␤ transcripts were followed by RNase protection assay. After 4 h of exposure to forskolin, C/EBP␤ mRNA was found to be prominently induced in cultured hippocampal neurons (Fig. 4A). The increase of C/EBP␤ transcript was observed even from 30 min after addition of forskolin to the culture (data not shown). Consistent with the induction of mRNA, immunoblot analysis with anti-C/EBP␤ antibodies showed that the amount of both the 39-and 33-kDa form of C/EBP␤ are increased 4 h after forskolin treatment of the culture (Fig. 4B).
To examine whether the increase of C/EBP␤ mRNA and protein can enhance its DNA binding activity to C/EBP sites, EMSA was performed using both nuclear extract from culture treated with forskolin for 4 h and nuclear extract from control culture. Forskolin treatment of the culture was found to increase the binding of C/EBP family members to the C/EBP site in EMSA (Fig. 4C, lanes 1 and 2). Anti-C/EBP␤ antibodies induced a supershift, and the amount of these supershifted bands were greatly enhanced by forskolin treatment of the culture (Fig. 4C). A supershift by anti-C/EBP␦ antibodies was also found to be increased by forskolin treatment of the culture (Fig. 4C). Unrelated antibodies such as anti-Rb antibodies do not induce any supershifts, and the addition of cold competitors almost erased the shifted bands demonstrating specificity of this EMSA. Thus, it is clear that binding of C/EBP␤ to C/EBP binding sites is intensified in cultured hippocampal neurons after forskolin treatment.
C/EBP␤ and ␦ mRNA Are Induced in Hippocampal Slices by cAMP Signaling-(Sp)-cAMPS is known to stimulate cAMP signaling in hippocampal slices (29,41), which preserve the anatomical relation of neurons in the intact hippocampus. To examine whether stimulation of cAMP signaling can induce the mRNA expression of C/EBP family members even in hippocampal slices, rat and mouse hippocampal slices were treated with (Sp)-cAMPS for 30 min. Two hours after the 30 min of (Sp)-cAMPS treatment, RNAs were prepared from slices and used for RNase protection assay. Both C/EBP␤ and ␦ mRNAs were induced by (Sp)-cAMPS in mouse hippocampal slices (Fig. 5A). In rat slices, both C/EBP␤ and ␣ mRNAs are induced after (Sp)-cAMPS treatment (Fig. 5B). Thus, the increase of C/EBP␤ and ␦ mRNA can be induced by the stimulation of cAMP signaling pathway, even in hippocampal slices. This result suggests that cAMP signal can induce mRNA expressions of C/EBP family members in intact neurons of hippocampal slices. (32,33,47). Our EMSA, which uses core promoter sequence of C/EBP␤ gene as a probe, suggests that CREB can control the expression of C/EBP␤ in hippocampal neurons (Fig. 3). Given that the expression of C/EBP␤ can be controlled by CREB, C/EBP␤ activity is predicted to be enhanced by Ca 2ϩ signals in hippocampal neurons. To examine this possibility, hippocampal neurons in cultures were subjected to membrane depolarization by incubating them with high K ϩ solution for 5 s (33). Four hours after membrane depolarization, nuclear extracts were prepared from both depolarized neurons and control neurons incubated with normal Tyrode's solution. EMSA using a C/EBP binding site as a probe showed that the DNA binding activity of C/EBP␤ is augmented by membrane depolarization of hippocampal neurons (Fig. 6A). The DNA binding activity of C/EBP␦ is also found to be enhanced by membrane depolarization (Fig. 6A). The increase of

FIG. 1. The mRNA expressions of C/EBP family members and C/EBP activities in the mouse hippocampus.
A, in the RNase protection assay, the hippocampus mRNAs annealed with and protected by mouse C/EBP␤ cRNA probe (left), mouse C/EBP␦ probe (right), and cyclophilin probe (all samples) were analyzed as described previously (42). B, EMSA of mice hippocampus nuclear extracts shows DNA binding activities of C/EBP members in mouse hippocampus (lane 1). Antibodies to C/EBP␤ (Anti␤), C/EBP␦ (Anti␦), and C/EBP␣ (Anti␣) induced supershifts (lanes 2, 4, and 8), which disappeared with the addition of competitor, cold C/EBP binding sites (lane 3, 5, and 9). C/EBP␤ mRNA was observed in hippocampal neurons 4 h after membrane depolarization, suggesting that the enhanced DNA binding activity of C/EBP␤ results from the increase of C/EBP␤ mRNA (Fig. 6B).
In hippocampal neurons, CaM kinases, probably CaMKIV, can activate CREB after Ca 2ϩ signal reaches nuclei of hippocampal neurons with heightened synaptic activities (33). To determine which signaling pathway is involved in the enhancement of C/EBP␤ activity by membrane depolarization of hippocampal neurons, we depolarized hippocampal neurons in the presence of CaM kinase inhibitor KN93. Inactive form KN92 did not erase the increase of C/EBP␤ activity in neurons that received membrane depolarization treatment (Fig. 7). In contrast, CaM kinase inhibitor KN93 clearly blocked the enhancement of DNA binding activities of C/EBP␤ in depolarized neurons, indicating that CaM kinase pathway, probably CaMKIV activation, is involved in the increase of C/EBP␤ activity by Ca 2ϩ in hippocampal neurons (Fig. 7). Another CaM kinase inhibitor, KN62 also blocked the enhancement of C/EBP␤ binding to DNA after membrane depolarization (data not shown). Both CaM kinases inhibitors also blocked the increase of C/EBP␤ mRNA expression in hippocampal neurons after mem-brane depolarization (data not shown). These results indicate that Ca 2ϩ signal enhances mRNA expressions of C/EBP␤ and ␦, and their DNA binding activities in hippocampal neurons through CaM kinases activation.
C/EBP␤ has in its b-ZIP domain a consensus sequence for CaM kinases phosphorylation. CaMKII can actually phosphorylate C/EBP␤ and enhance C/EBP␤-dependent gene transcription (48). Thus, it may be possible that CaMKIV has a similar positive effect to C/EBP␤-dependent gene transcription. To test this possibility, constructs for expressing the constitutively active form of CaMKIV (45), vectors for expressing C/EBP␤ or ␦ and a reporter construct with C/EBP binding sites were transfected and C/EBP-dependent gene transcriptions were assayed by measuring luciferase activity. As a result, overexpression of active CaMKIV was found to enhance C/EBP␤dependent gene transcription (Fig. 8). Unexpectedly, overexpression of active CaMKIV is found to significantly stimulate C/EBP␦-dependent transcription (Fig. 8). This result suggests that CaMKIV activated by Ca 2ϩ signal can enhance either C/EBP␤-or ␦-dependent gene transcription when these C/EBP family members are expressed within nuclei of hippocampal neurons with heightened synaptic activities.

DISCUSSION
We have found that C/EBP␤ transcripts are expressed in mouse hippocampus more than another C/EBP member, C/EBP␦ (Fig. 1A). In previous work, C/EBP␣ transcripts were detected in CA1 to CA4 regions of the mouse hippocampus by in situ hybridization analysis (27). However, we found that C/EBP␣ mRNA is hardly detectable in rat hippocampus even by RNase protection assay (data not shown). Although it may simply represent the species difference used in both studies, C/EBP␣ dose not seem to constitute a major C/EBP member even in mouse hippocampus. Our experimental ground for this notion is that the antibodies to C/EBP␣ induce much weaker supershifts in EMSA of mouse hippocampus than the antibodies to C/EBP␤ (Fig. 1B). Thus, supershift experiments have clearly demonstrated that C/EBP␤ is a major C/EBP member active in mouse hippocampus (Fig. 1B). C/EBP␤ is also found to be expressed in cultured hippocampal neurons (Fig. 2, A and  B). Niehof et al. (28) have found that the expression of C/EBP␤ mRNA can be controlled by CREB in various physiological conditions. We have found that CREB in nuclear extracts from both hippocampus and cultured hippocampal neurons can bind a crucial region of C/EBP␤ promoter, which is essential for C/EBP␤ expression (Fig. 3). Thus, CREB may probably serve as a central regulator of C/EBP␤ expression in hippocampal neurons. It has been previously reported that there is modest level of phosphoCREB in the hippocampal neurons of mice, which may explain the existence of C/EBP␤ mRNA and its DNA binding activity in mice hippocampus and cultured hippocampal neurons (30).
To simulate the cAMP signaling pathway, we stimulated  cultured hippocampal neurons by forskolin, a potent adenylate cyclase activator. Within 30 min after forskolin stimulation of the culture, phosphorylation of CREB was significantly induced in the nuclei of hippocampal neurons. 2 It has been recently reported that forskolin increases the endogenous C/EBP␤ mRNA expression in a neuronal cell line, Neuro 217 cells suggesting that forskolin might also induce its mRNA expression in primary neurons in culture (28). We actually observed that level of mRNA, protein, and DNA binding activity of C/EBP␤ are augmented 4 h after forskolin treatment in hippocampal neurons (Fig. 4, A-C). Thus, it is possible that CREB activated by cAMP signaling pathway can induce the expression of C/EBP␤ mRNA in hippocampal neurons. C/EBP␦, another member of C/EBP, was also induced in this culture by forskolin. It was previously demonstrated that C/EBP␦ can be induced by the stimulation of cAMP signals in mouse cortical astrocytes in culture (49). Our immunocytochemistry suggests that C/EBP␦ is moderately expressed in astrocytes, while it is weakly expressed in some hippocampal neurons (Fig. 2, C and  D). Thus, our data suggest that stimulation of cAMP signaling in hippocampal neurons also enhances C/EBP␦ expression through CREB activation in hippocampal neurons.
Membrane depolarization of hippocampal neurons induced the increase of mRNA expression and the DNA binding activity of C/EBP␤ and ␦ (Fig. 6). Since CaM kinase inhibitors KN93 and KN62 blocked the enhancement of C/EBP␤ mRNA and its activity by membrane depolarization of hippocampal neurons (Fig. 7), this indicates that activation of CaM kinases is involved in the induction of mRNA expression and DNA binding activity of C/EBP␤ by membrane depolarization (Fig. 7). Bito et al. (33) recently reported that CaMKIV activated by Ca 2ϩ signal can phosphorylate CREB and enhance its activity in hippocampal neurons in culture. Our data suggest that CREB binds to the essential promoter of C/EBP␤ gene in cultured hippocampal neurons, as well as mice hippocampus (Fig. 3). It is possible that CaMKIV through CREB activation is critically involved in the enhancement of C/EBP␤ expression and its activity in hippocampal neurons activated by membrane depolarization. We also observed that active CaMKIV can enhance both C/EBP␤-and C/EBP␦-dependent gene transcription (Fig.  8). It has been previously shown that CaMKII can augment C/EBP␤-dependent gene transcription by phosphorylating a consensus sequence within the leucine-zipper of C/EBP␤ (48). However, CaMKII is found to be localized in the cytoplasm of the hippocampal neurons, even after synaptic activities of those neurons are intensified by strong direct depolarization (50). In contrast, CaMKIV is constantly localized in nuclei of hippocampal neurons before and after synaptic activation (50). It may be concluded that CaMKII may not be an upstream kinase for C/EBP␤ in hippocampal neurons unless CaMKII phosphorylates newly synthesized C/EBP␤ in cytoplasm. Rather, it is possible that nuclear CaMKIV phosphorylates C/EBP␤ to increase C/EBP␤-dependent gene transcription in nuclei of hippocampal neurons. Though C/EBP␦ dose not have any known phosphorylation consensus sequence for CaM kinases, overexpression of the constitutively active form of CaMKIV significantly stimulated C/EBP␦-dependent gene transcription (Fig. 8). One possibility is that C/EBP␦ is making heterodimer with an endogenous bZIP partner or an unknown 2 K. Yukawa, unpublished results. physiological partner in host cell lines used in transfection assay. As synergy between ␤ and ␦ or ␣ and ␦ to C/EBP-dependent transcription could not be observed in the presence of the constitutively active form of CaMKIV, 2 it is unlikely that C/EBP␦ enhances its transcriptional activity by making heterodimers with endogenous C/EBP␤ or ␣. Alternatively, it may be possible that CaMKIV directly phosphorylates some sequences on C/EBP␦. This alternative might be addressed in the future. Thus, in addition to transcriptional control of C/EBP␤ and ␦ by Ca 2ϩ , Ca 2ϩ signal might strengthen transcriptional activities of C/EBP family members using CaM kinases in hippocampal neurons.
A number of studies have indicated that both Ca 2ϩ and cAMP signaling pathways have crucial roles in the process of learning and memory in many species (3)(4)(5)(6)(7)(35)(36)(37)(38)(39). Our data suggests that both cAMP and Ca 2ϩ signal might enhance the expression of C/EBP␤ and ␦ through CREB activation in hippocampal neurons with heightened synaptic activities during memory formation. Furthermore, CaM kinases may augment the transcriptional activities of C/EBP␤ and ␦ whose expressions are increased in hippocampal neurons during certain phases of long term memory. Even protein kinase A might enhance C/EBP␤ and C/EBP␦ activity since C/EBP␤-and C/EBP␦-dependent gene transcription can be strengthened by the overexpression of the catalytic subunit of protein kinase A. 3 Thus, dual roles of these kinases in enhancing both expression and function of C/EBP family members might play pivotal roles while acting under the direction of CREB in hippocampal neurons for stabilization of long term memory (Fig. 9). Long term potentiation LTP is believed to be a basic synaptic mechanism for some kinds of learning and memory (51)(52)(53). Thomas and Hunt (54) recently observed that C/EBP␤ mRNA levels were dramatically induced in the dentate granule cells 2 h after in vivo LTP induction by in situ hybridization. Besides this observation, they could hardly detect any C/EBP␤ mRNA in intact hippocampus. Their finding accords with our data, which show the induction of C/EBP␤ mRNA by cAMP and Ca 2ϩ in cultured hippocampal neurons, if we consider that Ca 2ϩ and cAMP are crucial signals for induction and maintenance of several kinds of LTP (29,41,52,55). To ask if C/EBP members are critically involved in synaptic plasticities such as hippocampal LTP and long term depression LTD or certain kinds of memory, it may be desirable to analyze mutant mice lacking the C/EBP family of transcription factors (23)(24)(25)(26) or conditional knock-out mice (56).