Upregulation of the neuron-specific K+/Cl- cotransporter expression by transcription factor early growth response 4

Abstract

The expression of the neuron-specific K+/Cl- cotransporter (KCC2) is restricted to the CNS and is strongly upregulated during neuronal maturation, yielding a low intracellular chloride concentration that is required for fast synaptic inhibition in adult neurons. To elucidate the mechanisms of KCC2 gene regulation, we analyzed the KCC2 (alias Slc12a5) promoter and proximal intron-1 regions and revealed 10 candidate transcription factor binding sites that are highly conserved in mammalian KCC2 genes. Here we focus on one of these factors, early growth response 4 (Egr4), which shows a similar developmental upregulation in CNS neurons as KCC2. KCC2 luciferase reporter constructs containing the Egr4 site (Egr4(KCC2)) were strongly induced by Egr4 overexpression in neuro-2a neuroblastoma cells and in cultured neurons. Egr4-mediated induction was decreased significantly by point-mutating the Egr4(KCC2). Insertion of Egr4(KCC2) into the KCC2 basal promoter in the endogenous reverse, but not in the opposite, orientation reestablished Egr4-mediated induction. Electrophoretic mobility shift assay confirmed specific Egr4 binding to Egr4(KCC2). Interference RNA-mediated knock-down of Egr4 and a dominant-negative isoform of Egr4 significantly inhibited KCC2 reporter induction and endogenous KCC2 expression in cultured neurons. Together, the results indicate an important role for Egr4 in the developmental upregulation of KCC2 gene expression.

Figure 1.

Analysis of conserved TF binding sites in the mouse KCC2 regulatory region. a, Shown are 10 TF binding sites that are highly conserved in the mouse, rat, human, and chimpanzee KCC2 regulatory region: two each of the Sp1 and AP2 sites and one site each of AP1, Mef2, E-box, Egr4, NRSE, and Block. The positions of the consensus sequences are indicated relative to the previously characterized transcription start site (+1). Nucleotides that differ from the consensus sequences are shown in boxes. Highly conserved nucleotides according to the matrix of TF binding sites appear in bold. Abbreviations for nucleotides in consensus sequences include n, for any nucleotide; s, for g/c; w, for a/t; y, for c/t; r, for a/g; m, for a/c; k, for g/t. b, Single-stranded (forward) DNA sequence of the mouse KCC2 promoter region, exon-1, part of intron-1, and exon-2 fused to the firefly luciferase cDNA. A corresponding schematic drawing of the promoter is depicted on the right. The previously identified consensus TF binding sites, with their locations on either the sense (+) or the antisense (−) DNA strand, are boxed. The position for the previously identified transcription start site is indicated with an arrow. KCC2 exon-1 and exon-2 sequences are shown in capital letters; coding regions also are in bold, and the translated protein sequence is presented above the corresponding nucleotide strand.

Figure 2.

KCC2 and Egr4 proteins are upregulated synchronously in the hippocampus (Hc) and cerebellum (Cb) during postnatal development. KCC2 and Egr4 protein levels in mouse hippocampus and cerebellum were measured by Western blot (top) at P2, P15, and P30 and are shown relative to β-tubulin (bottom; n = 3 mice for each time point).

Figure 3.

Egr4, but not Egr1, strongly induces the activity of KCC2 reporter constructs containing the Egr4^KCC2 site. a, Basal activity of KCC2 promoter constructs in N2a cells. Activity of the KCC2(−1398/+42) construct, containing a 1.4 kb KCC2 promoter sequence upstream of the transcription start site, is ∼12 times stronger than the activity of the promoterless pGL3-Basic vector. Activities of KCC2(−1398/+42) and KCC2(−309/+42) constructs are similar, implying that the proximal promoter region is mostly sufficient to provide a basal level of KCC2 promoter activity. Note the significantly lower activity of the KCC2(−1398/+5445) construct, which may be explained by the presence of unknown repressor elements in the proximal part of intron-1 (Uvarov et al., 2005). b, Egr4 overexpressed in N2a cells strongly (from 15- to 22-fold) induces luciferase activity of previously described KCC2 reporter constructs (gray bars). Egr4 does not induce activity of the KCC2(−180/+42) construct lacking the Egr4^KCC2 site. Egr1 expression (open bars) moderately upregulates activity of the KCC2(−1398/+5445) construct but has no significant effect on the activity of other KCC2 reporters. To take into account different basal levels of activity for KCC2 reporter constructs, we present the results as fold inductions of Egr4 or Egr1 overexpression in N2a cells in comparison with empty vector transfection. c, Simultaneous coexpression of Egr1 and Egr4 does not prevent induction of the KCC2(−1398/+42) construct. However, activity of the KCC2(−1398/+42) construct is upregulated slightly by increasing amounts of Egr1 plasmid transfected into N2a cells (*p < 0.05; Student's t test; n = 3). Error bars indicate SEM.

Figure 4.

Egr4 binds the Egr4^KCC2 element in vitro. a, Three specific complexes (I–III) are detected in EMSA, using labeled oligonucleotides bearing a single copy of the Egr4^KCC2 element and nuclear extracts of N2a cells transfected with Egr4 expression plasmid. Complex I is competed effectively by cold Egr4^KCC2 oligonucleotides. In contrast, unlabeled Egr4^KCC2mut oligonucleotides do not influence the formation of complex I. The addition of Egr4 antibody prevents complex I formation in a dose-dependent manner and results in the formation of a new supershifted complex (indicated by ss). Complex II is affected less by Egr4^KCC2 oligonucleotides and not at all by Egr4^KCC2mut oligonucleotides. Complex III is competed by both wild-type and mutant Egr4^KCC2 oligonucleotides. Note the significant increase in complex III intensity after the addition of the Egr4 antibody. Nonspecific binding is indicated by an asterisk. b, Sequences of Egr4^KCC2 and Egr4^KCC2mut oligonucleotides used in EMSA. The Egr4^KCC2 sequence is highlighted in bold. Mutated nucleotides are shown in small case letters.

Figure 5.

Mutagenesis of the Egr4^KCC2 element decreases Egr4-mediated induction of the KCC2(−309/+42) reporter construct in N2a cells. a, The C nucleotide at position 4 in Egr4^KCC2 that deviates from the Egr consensus sequence is shown in gray. Two G nucleotides modified by site-directed mutagenesis are indicated with an asterisk. b, Mutagenesis of the two G critical nucleotides in Egr4^KCC2 significantly prevents Egr4-induced activation of the KCC2(−309/+42)^Egr4mut reporter. c, Insertion of two Egr4^KCC2 sites in the endogenous (reverse), but not in the opposite (forward), orientation just upstream of the minimal KCC2(−180/+42) promoter enables the reporter to be activated by Egr4 (*p < 0.05; ***p < 0.001; Student's t test; n = 4). Error bars indicate SEM.

Figure 6.

The dominant-negative isoform of Egr4 inhibits Egr4-mediated induction of KCC2(−309/+42) reporter activity in N2a cells. a, Schematic drawing of the Egr4 and DN-Egr4 isoforms. DN-Egr4 retains the DNA binding domain containing three zinc fingers, whereas the N-terminal regulatory region of Egr4 is deleted. b, Increasing the ratio of DN-Egr4 to Egr4 effectively inhibits Egr4-mediated induction of the KCC2(−309/+42) reporter, whereas even the highest amount of DN-Egr4 used has no effect alone on basal reporter activity. Error bars indicate SEM.

Figure 7.

DN-Egr4 downregulates endogenous KCC2 expression in neurons. a, KCC2 mRNA expression is already detectable in mouse hippocampal neurons cultured at 4 DIV and is increased gradually up to 11 DIV. Egr4 mRNA also is detected in the cultures on day 4 and is increased approximately sevenfold by 11 DIV. Quantification of KCC2 and Egr4 mRNA expression was performed by using real-time reverse transcription-PCR and normalized to GAPDH (glyceraldehyde phosphate dehydrogenase). The number of independent experiments for each point is indicated in parentheses. b, Hippocampal neurons were cotransfected at 7 DIV with a GFP expression vector and either the DN-Egr4 or the control empty expression plasmid. After 2 d the cultures were fixed and assayed for KCC2 immunoreactivity. Images of GFP-positive neurons (indicated with arrows) coexpressing either DN-Egr4 or control vector plasmids are shown. Scale bar, 20 μm. c, The endogenous KCC2 expression clearly is reduced in neurons transfected with DN-Egr4, but not with the empty vector. KCC2 immunoreactivity is similar between nontransfected neurons (−) and neurons transfected with the empty vector. The number of cells analyzed in each group is indicated in parentheses. ns, Not significant (*p < 0.05; Student's t test). Error bars indicate SEM.

Figure 8.

Knock-down of Egr4 inhibits KCC2 promoter activity in mature neurons. a, In 5 DIV neurons exogenously expressed Egr4 increased the activity of the KCC2(−309/+42) luciferase reporter (by twofold), whereas DN-Egr4 failed to affect the reporter activity. b, In 10 DIV neurons overexpressed DN-Egr4 suppressed the KCC2(−309/+42) reporter activity in a dose-dependent manner, whereas overexpression of Egr4 failed to increase the KCC2 promoter activity. c, Synthetic Egr4 siRNA duplexes transfected into 8 DIV neurons significantly suppressed (by 26–33% compared with the negative control) the luciferase activity of the KCC2(−309/+42) reporter in 10 DIV neurons (*p < 0.05; **p < 0.01; Student's t test). Error bars indicate SEM.