Regulation of cpg15 by signaling pathways that mediate synaptic plasticity

Abstract

Transcriptional activation is a key link between neuronal activity and long-term synaptic plasticity. Little is known about genes responding to this activation whose products directly effect functional and structural changes at the synapse. cpg15 is an activity-regulated gene encoding a membrane-bound ligand that regulates dendritic and axonal arbor growth and synaptic maturation. We report that cpg15 is an immediate-early gene induced by Ca(2+) influx through NMDA receptors and L-type voltage-sensitive calcium channels. Activity-dependent cpg15 expression requires convergent activation of the CaM kinase and MAP kinase pathways. Although activation of PKA is not required for activity-dependent expression, cpg15 is induced by cAMP in active neurons. CREB binds the cpg15 promoter in vivo and partially regulates its activity-dependent expression. cpg15 is an effector gene that is a target for signal transduction pathways that mediate synaptic plasticity and thus may take part in an activity-regulated transcriptional program that directs long-term changes in synaptic connections.

Fig. 1

cpg15 is an IEG induced by synaptic activity through activation of NMDA receptors and L-type VSCCs. (A) Northern blots of total RNA prepared from cortical cultures treated for 9 h at 14 div. Cultures were untreated (−) or treated with TTX (1 µM), PTX (50 µM), or KCl (50 mM) in the absence or presence of EGTA (2 mM), CPP (10 µM), or nifedipine (Nif, 5 µM) as indicated above. (B) Quantification of cpg15 mRNA levels shown in A normalized to a GAPDH loading control. cpg15 expression in TTX-treated cultures was designated 1. Cultures treated with PTX or KCl and CPP showed significantly higher cpg15 expression compared to TTX-treated cultures (*P < 0.01 versus TTX, ANOVA and SNK post hoc test; n = 3). PTX-induced cpg15 expression was Ca²⁺-dependent and required NMDA receptors and L-type VSCCs. KCl-induced cpg15 expression required L-type VSCCs but not NMDA receptors. (C) Activity-dependent cpg15 expression is protein synthesis independent. Cultures were treated with TTX or PTX in the absence or presence of cycloheximide (CHX). cpg15 expression was analyzed as in B. PTX-induced cpg15 expression was not significantly effected by the presence of cycloheximide (*P < 0.05, ANOVA and SNK post hoc test; n = 2), thus defining cpg15 as an IEG.

Fig. 2

A 1.9-kb cpg15 promoter fragment mediates activity-dependent reporter gene expression. (A) Schematic diagram of the cpg15 genomic structure and the cpg15-Luc reporter construct. Closed boxes indicate the three exons of the cpg15 gene. The cpg15-Luc plasmid contains 1.6 kb of cpg15’s promoter region and 0.3 kb of its 5′-untranslated region fused to the luciferase reporter gene (Luc). (B) The 1.9-kb cpg15 upstream fragment drives both activity-independent and activity-dependent transcription. Cortical cultures were transfected with the cpg15-Luc reporter plasmid, with the pGL3-promoter vector carrying an SV40 promoter (SV40prom-Luc), or with the promoterless pGL3-basic vector (Luc only) at 6 div. At 14 div, cultures were treated with TTX or PTX for 9 h. Promoter activity measured as firefly luciferase activity was normalized to Renilla luciferase activity from the cotransfected plasmid pRL-TK. Luciferase activities shown are relative to cpg15-Luc transfected cells treated with TTX. Luciferase activity in the presence of the cpg15 promoter was higher than in its absence. PTX treatment further increased the luciferase activity from cpg15-Luc as compared to TTX treated cells (*P < 0.01, ANOVA and SNK post hoc test; n = 6). (C) PTX-induced cpg15 promoter activity requires Ca²⁺ influx through NMDA receptors and L-type VSCCs. Cortical cultures transfected with cpg15-Luc were stimulated with PTX for 9 h in the absence or presence of the indicated pharmacological agents. The PTX-induced increase in luciferase activity driven by the cpg15 promoter was blocked by EGTA, CPP, and nifedipine (*P < 0.01 versus TTX, ANOVA, and SNK post hoc test; n = 6).

Fig. 3

Three cpg15 CREs bind CREB, CREM, and an unknown factor in brain nuclear extracts. The −1.58-kb EGR RE binds EGR1, EGR3, and an unknown factor. (A) In EMSAs, ³²P-labeled oligonucleotides containing the CRE consensus or each of the CRE-like sequences in the cpg15 promoter (at −1.56 kb, −0.84 kb, and +0.04 kb) bind two factors from mouse brain nuclear extracts and form DNA/protein complexes with similar electrophoretic mobility (marked by closed arrowheads). The binding is specific, as it can be competed away with excess unlabeled consensus CRE oligonucleotide (w), but not with the same oligonucleotide containing two point mutations (m). (B) EMSAs were done for the consensus and cpg15 CREs in the presence of antibodies against CREB, CREM, ATF-1, or control IgG. Anti-CREB antibody caused a supershift of the higher mobility DNA/protein complex seen with all four oligonucleotides (open arrowheads). A similar supershift was caused by anti-CREM antibody, but not by antibodies against ATF-1 or control IgG. (C) EMSAs using EGR RE consensus or each of the cpg15 EGR REs show that only the consensus and −1.58-kb EGR RE bind factors in mice brain nuclear extracts. Binding to these two factors is specific as seen by competition with wild-type and not mutant EGR RE oligonucleotide (closed arrowheads). A third specific band observed with EGR RE consensus could not be resolved with the −1.58-kb EGR RE site due to a comigrating nonspecific band. (D) EMSAs were done for the consensus and the −1.58-kb EGR RE site in the presence of antibodies against EGR1, EGR2, EGR3, or control IgG. Anti-EGR1 antibody supershifted the slowest migrating DNA/protein complex, whereas anti-EGR3 antibody supershifted the fastest migrating complex (open arrowheads). Although the slowest migrating band was not resolved with the −1.58-kb EGR RE site, addition of anti-EGR1 antibody resulted in a supershift similar to that seen with the consensus EGR RE.

Fig. 4

CREB is involved in activity-dependent regulation of the cpg15 promoter. (A) Dominant negative CREB mutants block activity-dependent transcription driven by the cpg15 promoter. Cortical cultures were cotransfected with cpg15-Luc and A-CREB or K-CREB dominant negative CREB expression plasmids or a control EGFP expression plasmid. Cultures were treated with TTX or PTX for 9 h, and cpg15 promoter activity was determined by luciferase assay as in Fig. 2. Luciferase activities shown are relative to TTX-treated cells transfected with cpg15-Luc and the EGFP expression plasmid. A-CREB and K-CREB both significantly reduced activity-dependent luciferase expression driven by the cpg15 promoter (*P < 0.01, ANOVA and SNK post hoc test; n = 10–11). (B) cpg15 CREs play a role in activity-dependent transcription driven by the cpg15 promoter. Cortical cultures were transfected with the cpg15-Luc plasmid carrying the wild-type cpg15 promoter (top) or the same plasmid with individual point mutations in the −1.56-kb CRE, −0.84-kb CRE, +0.04-kb CRE, and −1.58-kb EGR RE, or with a combination of point mutations. The luciferase reporter fused to the SV40 promoter served as a control (SV40, bottom). Schematic diagrams of these reporter plasmids are shown on the left. Intact binding sites are indicated by closed boxes, and mutated sites are marked with an X. Transfected cells were treated with TTX or PTX for 9 h, and cpg15 promoter activity assayed as described for Fig. 2. The effect of mutating each site was analyzed using a two-factor ANOVA with combined data from single and multiple mutations. Mutation of the −1.56-kb CRE significantly increased cpg15 promoter activity in PTX-treated cultures (P < 0.001, n = 10–15), whereas mutation of the +0.04-kb CRE significantly decreased its activity (P < 0.01, n = 10–15). (C) Endogenous CREB binds to the cpg15 promoter in vivo. DNA-binding proteins in cortical cultures were cross-linked to chromatin with formaldehyde, sonicated, and then subjected to immunoprecipitation with anti-CREB antibody or control IgG. After reversing the cross-links, presence of promoter fragments in the immunoprecipitates was examined by PCR with primers covering the cpg15 −1.56-kb CRE, cpg15 +0.04-kb CRE, or the GAP-43 promoter as a negative control. Input chromatin (0.5%) was used as a positive control for the PCR reaction. The +0.04-kb CRE cpg15 promoter region was present in the chromatin immunoprecipitated by the anti-CREB antibody, indicating that it binds CREB in vivo.

Fig. 5

Activity-dependent cpg15 expression is mediated by the CaMK and MAPK pathways. (A) PTX-induced cpg15 expression is dependent on the CaMK and MAPK pathways. Cortical cultures were treated with TTX or PTX in the absence or presence of kinase inhibitors or their inactive analogs: KN93 and KN92 (0.5 µM) for CaMK or U0126 and U0124 (5 µM) for MAPK. cpg15 expression was analyzed by Northern blot hybridization as in Fig. 1. Both KN93 and U0126 significantly reduced the PTX-induced cpg15 mRNA expression (*P < 0.05, ANOVA and SNK post hoc test; n = 3–4). (B) Activity-dependent cpg15 promoter activity is uneffected by dnCaMKIV. Cortical cultures were cotransfected with cpg15-Luc and dnCaMKIV or a control EGFP expression plasmid. Cultures were treated with TTX or PTX for 9 h, and promoter activity was assayed as in Fig. 2. Luciferase activities shown are relative to EGFP transfected cells treated with TTX (n = 8). (C) Activity-dependent cpg15 expression is comparable in dnCaMKIV transgenic mice and wild-type controls. Adult wild-type (WT) and dnCaMKIV transgenic mice were injected with PBS or 25 mg/kg kainate (KA), scored for seizure severity (see Experimental Methods) and sacrificed after 6 h. RNA was extracted from the cerebral cortices, and cpg15, dnCaMKIV, and GAPDH expression levels were examined by Northern blot analysis as in Fig. 1. Normalized cpg15 expression relative to PBS-injected wild-type mice is shown for each animal. (D) Quantification of the results shown in C. cpg15 expression increased significantly by kainate injection in both wild-type and dnCaMKIV transgenic mice (*P < 0.05, ANOVA and SNK post hoc test; n = 2–8). The kainate-induced cpg15 expression was not significantly different between wild-type and dnCaMKIV transgenic mice.

Fig. 6

cpg15 expression is activated by cAMP, but activation of PKA is not sufficient for cpg15 induction. (A) cpg15 induction by forskolin requires concurrent synaptic stimulation. Cortical cultures were treated with TTX, PTX, forskolin (Fsk, 10 µM), CPP or nifedipine for 9 h as indicated on the left. cpg15 expression was analyzed by Northern blot hybridization as in Fig. 1. Forskolin significantly increased cpg15 mRNA expression, but not in the presence of TTX or CPP and nifedipine (*P < 0.01 versus TTX, ANOVA, and SNK post hoc test; n = 3). (B) PKA is not activated by PTX stimulation. Cortical cultures were treated as indicated on the left for 10 min. PKA activity was determined by a kinase assay using the specific peptide substrate Kemptide. Forskolin, but not PTX, significantly increased PKA activity (*P < 0.01 versus TTX, ANOVA, and SNK post hoc test; n =3). PKA activation by forskolin was not blocked by TTX or CPP and nifedipine.

Fig. 7

Two alternative models of activity dependent cpg15 induction. In the sequential activation model (A), NMDA receptor activation by synaptic activity leads to activation of L-type VSCCs and Ca²⁺ influx through these channels, triggering CaMK and MAPK pathways. In the parallel activation model (B), influx of Ca²⁺ from NMDA receptors and L-type VSCCs each trigger local activation of either the CaMK or the MAPK pathways. In both cases, activation of CaMK and MAPK pathways converge on a set of transcription factors, including CREB, that together activate cpg15 transcription.