Gene regulation via excitation and BDNF is mediated by induction and phosphorylation of the Etv1 transcription factor in cerebellar granule cells

Abstract

In maturing postnatal cerebellar granule cells, the Etv1/Er81 transcription factor is induced by sequential activity-dependent mechanisms through stimulation of AMPA and NMDA receptors, voltage-dependent Nav1.2 Na⁺ channels, and voltage-dependent Ca²⁺ channels. Etv1 then up-regulates a battery of maturation genes involved in the cerebellar circuitry. In this process, BDNF is also induced and participates in the up-regulation of these maturation genes. Using cultures of granule cells, we addressed how the activity-dependent and BDNF signaling mechanisms converge on the regulation of the representative NR2C NMDA receptor and Tiam1 maturation genes. BDNF up-regulated both the NR2C and Tiam1 genes via the TrkB-Erk cascade and this up-regulation was blocked not only by inhibition of the activity-dependent signaling mechanisms but also by suppression of Etv1 expression with Etv1 siRNA. Importantly, Etv1 was selectively phosphorylated by Erk1/2 in the BDNF signaling cascade, and the inhibition of this phosphorylation abrogated the BDNF-induced up-regulation of the NR2C and Tiam1 genes. The luciferase reporter assays in combination with mutations of MEK and Etv1 indicated that the Erk-mediated, phosphorylated Etv1 interacted with the Ets motifs of the NR2C promoter sequence and that phosphorylation at both serine 94 and a cluster of threonines and a serine (Thr139, Thr143, and Ser146) of Etv1 was indispensable for the BDNF-mediated activation of the NR2C promoter activity. This study demonstrates that the NR2C and Tiam1 maturation genes are synergistically controlled by the activity-dependent induction of Etv1 and its phosphorylation by the BDNF signaling cascade.

In the developing cerebellum, granule cells proliferate, postmitotically differentiate in the external granular layer, and then migrate inwardly into the internal granular layer, where they form refined synaptic connections with mossy fibers (1). During these developmental processes, the resting membrane potential of the maturing granule cells significantly shifts from a relatively depolarized state to a more hyperpolarized one (2–4). This shift of the membrane potential plays a key role in activity-dependent regulation of the expression of a set of maturation genes in maturing granule cells (5, 6). The hyperpolarized state of maturing granule cells enhances responsiveness to glutamatergic excitation and leads to the Nav1.2 Na⁺ channel-mediated action potential, thereby stimulating Ca²⁺ entry via voltage-dependent Ca²⁺ channels (7). Ets variant gene 1 (Etv1/Er81), which is a transcription factor of the ETS family, is then induced by this sequential activity-dependent signaling mechanism and commonly up-regulates the battery of maturation genes, including those of the NMDA receptor subunit NR2C and Tiam1 proteins in both primary cultures and the developing cerebellum (7). During cerebellar development, BDNF and its TrkB receptor are highly expressed in granule cells at the internal granular layer after migration (8, 9), and the activation of the TrkB receptor and its downstream MEK-Erk signaling cascade up-regulates the NR2C and other maturation genes (10, 11). Furthermore, the induction of NR2C is markedly reduced in cerebellar granule cells of the TrkB knockout mice in vivo (10). The BDNF-TrkB-Erk cascade thus contributes to the up-regulation of the NR2C and other maturation genes in maturing granule cells (10, 11). However, how the BDNF-Erk signaling cascade and the activity-dependent sequential events converge to control up-regulation of these maturation genes remains to be clarified.

Granule cells in primary culture are capable of recapitulating many properties characteristic of maturing granule cells in vivo (6, 7, 12, 13). The present study addressed how the expression of maturation genes is controlled by the BDNF-Erk signaling and activity-dependent mechanisms in cultured granule cells. Here we report that Etv1 is induced by the activity-dependent excitation of granule cells, specifically phosphorylated by Erk1/2, and plays a key role in up-regulation of the maturation genes.

Results

Synergistic Up-Regulation of Maturation Genes by BDNF and Cell Excitation.

To address whether, and if so how, BDNF and the activity-dependent signaling cascades synergistically regulate the expression of the maturation genes, we examined the effects of an inhibitor or blocker of each activity-dependent signaling component on the BDNF-mediated up-regulation of the representative maturation genes, namely the NR2C and Tiam1 genes. The following inhibitors/blockers were used: for glutamate receptors, an AMPA receptor-selective antagonist, 2, 3-dioxo-6-nitro-1,2,3,4-tetrahydrobenzo [f] quinoxaline-7- sulfonamide (NBQX) and an NMDA receptor-selective antagonist, 3-(2-carboxypiperazin-4-yl)-propyl-1-phosphonic acid (CPP) in combination; for Na⁺ channels, either tetrodotoxin (TTX) or siRNA against Na⁺ channel type II mRNA (Nav1.2 siRNA); for Ca²⁺ channels, 4 Ca²⁺ channel blockers in combination: that is, nifedipine (L-type), ω-agatoxin TK (P/Q-type), ω-conotoxin GVIA (N-type), and NNC 55–0396 (T-type) (7, 13) (Fig. 1A). As reported previously (10, 11), BDNF markedly up-regulated NR2C mRNA expression in granule cells cultured for 96 h at the physiological KCl concentration (5 mM KCl). This up-regulation was abolished by the inhibitor or blocker of each of the above activity-dependent signaling components (Fig. 1A). Similarly, the Tiam1 gene was significantly up-regulated by BDNF, and this up-regulation was likewise blocked by all inhibitors and blockers of the signaling components (Fig. 1A). These results demonstrate that the up-regulation of the representative NR2C and Tiam1 maturation genes by BDNF requires the activity-dependent excitation of granule cells. Interestingly, Etv1 mRNA levels progressively and markedly increased during the culture period (Fig. 1B), but BDNF did not appreciably increase the Etv1 mRNA level (Fig. 1B). Furthermore, the expression of the Etv1 mRNA was more markedly inhibited by all five inhibitors (Fig. 1A).

Fig. 1.

Signaling analysis of the activity-dependent and BDNF-induced up-regulation of the NR2C and Tiam1 genes. (A) Granule cells were cultured in serum-containing medium for 24 h and then in serum-free medium with and without 50 ng/mL BDNF for 96 h in the presence or absence of TTX (5 μM), CPP and NBQX (100 μM each), or four types of Ca²⁺ channel blockers (3 μM each). Nav1.2 siRNA and Etv1 siRNA (6 μg each) were electroporated into dissociated granule cells, and the cultures were incubated in serum-containing medium for 24 h and then in serum-free medium for 96 h. The KCl concentration of all cultures was set at 5 mM KCl and (−) represents no addition of any inhibitors or blockers. (B) Granule cells were cultured as in A for the indicated days in vitro (DIV) in the absence or the presence of BDNF or BDNF plus TTX. (C) Etv1 siRNA or scRNA (6 μg each) was electroporated into the cells, and the cultures were incubated in the presence or absence of 50 ng/mL BDNF for 96 h, as in A. (D) Granule cells were cultured as in A and the BDNF-treated cells were simultaneously incubated with 500 nM K252a, 10 μM LY294002, 20 μM U0126, 2 μM U73122, 5 μM SB203580, or 1 μM rapamycin. Indicated mRNA levels were quantified by PCR. All experiments were performed three times. In A, C, and D, the data are expressed as percentages of mRNA levels relative to those of the BDNF-untreated cells (100%). In B, the data are expressed as fold-increase in mRNA levels relative to those of BDNF-untreated cells at DIV 1. Columns and bars in A, C, and D and symbols and bars in B represent the mean ± SEM, respectively. *P < 0.05, ***P < 0.001 vs. BDNF-treated, inhibitor-untreated cells in A and BDNF- and scRNA-treated cells in C.

Next, by treating cultured granule cells with Etv1 siRNA or scrambled siRNA (scRNA), we examined whether Etv1 would be necessary for the BDNF-mediated up-regulation of the NR2C and Tiam1 genes (Fig. 1C). The Etv1 siRNA corresponding to the DNA-binding domain efficiently abrogated the expression of the Etv1 mRNA (Fig. 1A). The Etv1 siRNA, but not the scRNA, effectively blocked the BDNF-stimulated expression of both NR2C and Tiam1 mRNAs (Fig. 1C). Thus, Etv1 is indispensable for the up-regulation of both genes by BDNF.

BDNF activates TrkB by causing tyrosine phosphorylation at the cytoplasmic domain of this receptor (14, 15). The phosphorylated TrkB then activates several downstream signaling cascades (10, 14, 15). These signaling cascades include the MEK-Erk, the PI3K/Akt, and the phospholipase C (PLC)-γ1 pathways. Therefore, we next examined the signaling mechanisms of BDNF-mediated gene expression in culture at 96 h after treatment with various signaling inhibitors (Fig. 1D). The BDNF-stimulated up-regulation of both NR2C and Tiam1 genes was blocked by K252a, a tyrosine kinase inhibitor. Among the inhibitors tested for the signaling cascades downstream of TrkB activation, inhibitors of PI3K/Akt (LY294002), PLC-γ1 (U73122), and mammalian target of rapamycin failed to inhibit the BDNF-induced up-regulation of NR2C mRNA, and only a MEK-Erk inhibitor, U0126, specifically abrogated it. Furthermore, an inhibitor of p38 MAP kinase (SB203580) was ineffective in blocking the NR2C mRNA up-regulation, indicating that the MEK-Erk cascade among different MAP kinase signaling pathways selectively participates in NR2C up-regulation in the BDNF-TrkB pathway. This specificity of the MEK-Erk cascade was preserved in the induction of both Tiam1 and Etv1 mRNAs, as the signaling inhibitors other than the MEK-Erk inhibitor were ineffective in inhibiting the up-regulation of Tiam1 and Etv1 mRNAs (Fig. 1D). The results indicate that the MEK-Erk signaling cascade plays a key role in up-regulation of the NR2C and Tiam1 maturation genes after stimulation of the BDNF-TrkB pathway.

Phosphorylation of Etv1 via the BDNF-Mediated Erk1/2 Cascade.

The previous study indicated that Etv1 is the primary transcription factor that commonly controls the activity-dependent up-regulation of the maturation genes in granule cells both in vitro and in vivo (7). Therefore, we examined whether MEK-Erk could phosphorylate Etv1 after stimulation of the BDNF-TrkB pathway in cultured granule cells. To address this question, we first examined whether the signaling molecules involved in the BDNF-TrkB pathway were phosphorylated in granule cells and if so, whether such phosphorylation would be selectively blocked by the respective signaling inhibitors (Fig. 2). Erk1/2 was intensely phosphorylated in response to BDNF treatment, and this phosphorylation was abolished by U0126, but not by LY294002 or U73122 (Fig. 2A). In addition, U0126 had no inhibitory effect on the phosphorylation of the p38 MAP kinase (Fig. 2B). In other controls, LY294002 was verified to abrogate selectively the phosphorylation of the target Akt (Fig. 2C). We also confirmed no inhibitory effects on the expression levels of the signaling proteins by the respective inhibitor treatments (Fig. 2). We then examined whether phosphorylation of Etv1 was induced after BDNF treatment by culturing granule cells in the presence and absence of the selective signaling inhibitors (Fig. 2D). Etv1 was considerably phosphorylated by BDNF, and this phosphorylation was selectively blocked by the U0126 MEK-Erk inhibitor, but not by the other inhibitors. The results thus explicitly demonstrate that among BDNF-activated signaling cascades, the Erk1/2 activation results in specific phosphorylation of Etv1 in the granule cells.

Fig. 2.

BDNF-mediated phosphorylation of Etv1 via the Erk1/2 cascade. (A–D) Granule cells were cultured in serum-containing medium for 24 h and then in serum-free medium for 96 h with BDNF. Indicated inhibitors were added 6 h before termination of cell culture. Lysates of the granule cells were subjected to immunoblot (IB) analysis with the indicated antibodies. In the case of Etv1, lysates were immunoprecipitated (IP) and the immunoprecipitates were then immunoblotted. The arrow and arrowhead indicate phosphorylated and nonphosphorylated, respectively; the asterisk marks a nonspecific band.

We next addressed the sequential phosphorylation cascades by pursuing time courses of phosphorylation of TrkB, Erk1/2 and Etv1 following BDNF treatment (Fig. 3). BDNF caused rapid phosphorylation of TrkB, and this phosphorylation became pronounced at 5 min and faded out by 24 h after the addition of BDNF (Fig. 3A). Erk1/2 was also rapidly phosphorylated, but this phosphorylation was sustained at least up to 96 h after BDNF had been added (Fig. 3B). Etv1 was more slowly phosphorylated, being maximally phosphorylated 2 h after the addition of BDNF (Fig. 3C). BDNF had no effect on protein levels of TrkB, Erk1/2 and Etv1. Furthermore, BDNF had the ability to elevate the level of phosphorylated Akt in the cultured granule cells (Fig. 3D), although this signaling was not relevant to the BDNF-induced up-regulation of the maturation genes (Fig. 1D). The results indicate that MEK-Erk phosphorylates Etv1 in the signaling cascade downstream of the BDNF-TrkB pathway in the granule cells.

Fig. 3.

Time courses of protein phosphorylation of BDNF-treated granule cells. (A–D) Granule cells were cultured in serum-containing medium for 24 h and then for 96 h in serum-free medium. BDNF (50 ng/mL) was added at the indicated times before termination of cell culture. Cell lysates were prepared and subjected to immunoblot analysis with the indicated antibodies as in Fig. 2. For explanation of symbols, see Fig. 2 legend.

Phosphorylated Etv1 Activates the Promoter of the NR2C Gene.

The NR2C gene possesses a cluster of three Etv1-interacting motifs (Ets motifs) around 600-bp upstream of the transcription initiation site (7). In the expression of the NR2C gene, these motifs are sufficient to induce the Etv1-mediated promoter activity of this gene (7). By performing a luciferase reporter assay with Neuro2A cells, we tested whether the phosphorylated Etv1 could stimulate the promoter activity of the NR2C gene. Three different promoter constructs were prepared (7): the promoter region containing −800 to +200 bp from the transcription initiation site (wtNR), the sequence covering the three Ets motifs (−718 to −518 bp, coreNR), and a mutated coreNR (NR-mut), in which the core Ets motifs had been mutated. The MEK/Erk-mediated activation of these three promoters was assessed by cotransfection of Etv1 with and without constitutively active MEK (c-MEK) or inactive MEK (d-MEK) (16) (Fig. 4). Etv1 stimulated the promoter activity of both wtNR and coreNR genes but not that of the NR-mut gene. The cotransfection with c-MEK significantly stimulated the activity of both wtNR and coreNR (Fig. 4A). In contrast, the activation of the NR-mut gene by Etv1 was negligible regardless of the presence or absence of c-MEK. Furthermore, d-MEK failed to stimulate the promoter activity in both wtNR and coreNR genes more than that elicited by Etv1 alone.

Fig. 4.

Regulation of the promoter function of the NR2C gene by phosphorylated Etv1. (A) Three different NR2C promoter constructs were attached to the luciferase gene. Neuro2A cells were cotransfected with NR2C cDNA constructs and the indicated genes, and the effects of c-MEK and d-MEK on the Etv1-mediated activation of the NR2C promoter activity were analyzed by performing the luciferase assay (n = 3). (B) Neuro2A cells were cotransfected with an expression vector for the Etv1 cDNA or one of the indicated mutant Etv1 cDNAs, the coreNR gene, and either of the indicated MEK genes. The effects of Etv1 mutations were analyzed by performing the luciferase assay (n = 3). The data are expressed as fold-increase in luciferase activities relative to those of cells transfected with NR-mut gene alone. Columns and bars represent the mean ± SEM, respectively. *P < 0.05, **P < 0.01, ***P < 0.001; n.s., not significant.

Etv1 has been reported to undergo phosphorylation at several serine and threonine residues (17, 18). So we converted these serine and threonine residues individually or together to alanine and tested these mutated Etv1s for their ability to stimulate the promoter activity of coreNR. Among the many mutated Etv1s tested, those with conversion of serine at residue 94 (Ser94) or a cluster of threonines at residues 139 and 143 and serine at residue 146 (Thr139, Thr143, and Ser146) retained their ability to stimulate the promoter activity, but lost their responsiveness to activation by c-MEK (Fig. 4B). The results thus demonstrate that the specific phosphorylation of Etv1 by MEK-Erk plays a key role in activation of the NR2C gene expression through interaction of the phosphorylated Etv1 with the Ets motifs in the NR2C promoter sequence.

Discussion

In the developing cerebellum, a battery of maturation genes is up-regulated in postmigratory granule cells by both activity-dependent and extracellular signaling mechanisms (6, 7, 10, 11). In this process, the resting membrane potential significantly shifts from a depolarized state to a relatively hyperpolarized one (2–4). Under the depolarized state, high levels of intracellular Ca²⁺ ([Ca²⁺]i) are maintained and induce the expression of BDNF in maturing granule cells (8). However, a sustained increase in [Ca²⁺]i activates Ca²⁺-dependent calcineurin phosphatase, which blocks the BDNF-TrkB-Erk signaling pathway (10). The membrane depolarization also inactivates Na⁺ channels and prevents excitatory synaptic transmission (5, 12). Thus, the hyperpolarizing shift of granule cells during cell maturation is crucial for regulating the expression of maturation genes by both activity-dependent and BDNF-directed signaling mechanisms. However, how the extracellular and intrinsic signaling mechanisms synergistically control the maturation gene expression remained elusive.

The present investigation has revealed that the Etv1 transcription factor plays a key role in converging the BDNF-mediated and activity-dependent regulation of expression of the maturation genes. The up-regulation of the NR2C and Tiam1 maturation genes, as well as their master regulator of the Etv1 gene, was blocked by inhibition of the activity-dependent signaling mechanisms that sequentially operate through stimulation of glutamate receptors, voltage-dependent Na⁺ channels, and voltage-dependent Ca²⁺ channels (7). BDNF significantly up-regulated the NR2C and Tiam1 gene expression via the MEK-Erk cascade but had little effect on the Etv1 gene expression. The Etv1 gene expression is thus mainly controlled by the activity-dependent mechanism. Importantly, the BDNF-mediated up-regulation of the NR2C and Tiam1 genes was blocked not only by inhibition of the activity-dependent signaling mechanisms but also by suppression of the Etv1 expression with Etv1 siRNA. Furthermore, Etv1 was selectively phosphorylated by Erk1/2 in the BDNF-TrkB signaling pathway, and the inhibition of this phosphorylation abrogated the BDNF-induced up-regulation of the NR2C and Tiam1 genes. Etv1 thus controls the expression of the maturation genes by converging regulatory mechanisms at the levels of transcription and protein modification.

A mechanistic model of the regulation of the representative NR2C maturation gene expression by convergence of the BDNF and activity-dependent signaling is summarized as depicted in Fig. 5. In the maturation process, glutamatergic excitation of granule cells elicits action potentials via the activation of Nav1.2 Na⁺ channels, and in turn induces oscillatory Ca²⁺ entry via voltage-dependent Ca²⁺ channels (7). This sequential signaling progressively induces expression of Etv1, which commonly controls the expression of a maturation gene battery in the terminal maturation program (7). In this process, BDNF is synthesized de novo and activates the TrkB-Erk signaling cascade. This signaling cascade enhances phosphorylation of Etv1 and activates the promoter activity of the NR2C gene. Thus, the maturation genes, such as NR2C and Tiam1, are up-regulated by synergistic mechanisms via the activity-dependent induction of Etv1 and its phosphorylation by the TrkB-Erk cascade.

Fig. 5.

Schematic model of signaling mechanisms for the Etv1-mediated regulation of the NR2C gene.

The synergistic mechanism underlying the activation of the Etv1 function plays an important role in the functional synaptic assembly of the cerebellar circuit. For example, a switch of NMDA receptor subunits from NR2B to NR2C alters many properties of NMDA receptors and is essential for the functional synaptic transmission involved in cerebellar functions, such as the coordination of complex motor movements (19). Tiam1 also facilitates dendritic development by inducing Rac1-dependent remodeling and protein synthesis (20). The present study thus has demonstrated that the synergistic control of Etv1 by extracellular signaling and activity-dependent intrinsic regulation is crucial in directing the maturation program of the cerebellar circuit.

Materials and Methods

Culture.

All procedures for animal handling were performed according to the guidelines of Osaka Bioscience Institute. Cultures containing ∼90% granule cells were prepared from ICR mice at postnatal day 8. The cells were cultured in serum-containing medium for 24 h and then for 96 h in serum-free medium containing a physiological concentration of KCl (5 mM KCl) in the presence or absence of BDNF (50 ng/mL) (10, 11). Inhibitors or blockers were added to the culture medium for selected times (10). Nav1.2 siRNA, Etv1 siRNA or scRNA (6 μg each) was electroporated into dissociated granule cells as described previously (7), and the electroporated cells were cultured in serum-containing medium for 24 h and in serum-free medium for 96 h. The addition of inhibitors, blockers and siRNAs used did not affect cell viability at least up to 96 h in culture (7, 10, 13). Reagents were obtained from the following sources: TTX, CPP, NBQX, nifedipine, NNC 55–0396, K252a, LY294002, U0126, U73122, SB203580 and rapamycin from Tocris; ω-agatoxin TK and ω-conotoxin GVIA from Peptide Institute; BDNF from Peprotech. siRNAs were obtained from the source described previously (7).

Quantitative PCR and Immunoblot Analysis.

Quantitative PCR analysis was performed by using appropriate primers that gave rise to 60- to 150-bp DNA segments; glyceraldehyde-3-phosphate dehydrogenase mRNA was used as an internal control for normalization of mRNA levels (7, 13). Immunoprecipitation and immunoblot analysis were performed as described previously (10, 13). The antibodies used included anti-TrkB (1:1,000) from BD Transduction Laboratories; antiphospho-TrkB (1:1,000), anti-Erk1/2 (1:2,000), antiphospho-Erk1/2 (1:2,000), anti-Akt (1:2,000), antiphospho-Akt (1:2,000), anti-p38 (1:1,000), and antiphospho-p38 (1:1,000), all from Cell Signaling Technology; and anti-Etv1 (1:1,000) from Abcam. Because antiphospho-Etv1 was not available, Erk1/2-mediated, phosphorylated Etv1 was identified in immunoblot analysis as follows: Etv1 and c-MEK were expressed in Neuro2A cells by cotransfection of these cDNAs and cell lysates were immunoprecipitated by anti-Etv1. Immunoprecipitates were then electrophorased on a gel and immunoblotted by either anti-Etv1 or antiphosphoserine. The upper and lower bands on a gel were thus assigned as phosphorylated and nonphosphorylated Etv1.

Luciferase Assay.

Three NR2C promoter constructs were generated as described previously (7) and attached to the luciferase gene in the pGL4.10 vector (Promega). In the NR-mut construct, the core GGA(A/T) sequence of the three Ets motifs was all mutated to the CCGC sequence (7). The mutated Etv1s, c-MEK, and d-MEK (16, 18) were generated by using a QuikChange Lightning Site-Directed Mutagenesis Kit (Agilent Technologies); in d-MEK, lysine at residue 97 was replaced with methionine. Transfection of Neuro2A cells with cDNAs and measurements of the luciferase activity were performed as described previously (7). Renilla luciferase was used as a transfection control.

Statistical Analysis.

Statistical analysis was conducted by using the unpaired Student t test or one-way ANOVA. Statistical significance was determined with GraphPad Prism software (GraphPad software).