GATA and FOG2 transcription factors differentially regulate the promoter for Kv4.2 K⁺ channel gene in cardiac myocytes and PC12 cells

Abstract

Objective: Kv4.2 subunits are major components of transient outward K⁺ channels in cardiac myocytes and somatodendritic A-type channels in neurons. To identify molecular mechanisms underlying transcriptional regulation of Kv4.2 gene, we have isolated and characterized the promoter for the rat Kv4.2 gene. Methods: PCR-based amplification of cDNA end (5′RACE) and RNase protection assays were used to determine transcription start sites. Transient transfection of Kv4.2 promoter-luciferase reporter constructs into neonatal cardiac myocytes and PC12 cells was employed to measure activity of the Kv4.2 promoter. Cotransfection of expression vectors for the transcription factors, GATA and/or FOG2, was performed to determine the effects of these transcription factors on the Kv4.2 promoter. Results: Transcription of the gene initiates at 552 bp upstream from the translation initiation site in the brain and heart. Deletion analysis revealed that the ∼200-bp fragment encompassing this start site drives significant transcription in neonatal cardiac myocytes and PC12 cells. The transcription factors GATA4 and 6 differentially enhance activity of the minimum promoter in the two cell types: GATA4 produces a larger increase than GATA6 in cardiac myocytes, whereas the latter results in a more substantial enhancement in PC12 cells. Furthermore, the coregulator of GATA, FOG2, markedly suppresses the GATA-induced increase in myocytes, but enhances it in PC12 cells. The use of GATA mutants that are incapable of forming complexes with FOG2 indicates that the formation of GATA–FOG complexes is required for the FOG2-induced suppression in myocytes, but not for the FOG2-mediated enhancement in PC12 cells. Conclusion: These results indicate that GATA and FOG2 transcription factors use distinct mechanisms to control the expression of Kv4.2 gene in cardiac myocytes and PC12 cells. The lack of a GATA-binding consensus sequence in the Kv4.2 minimum promoter suggests that these transcription factors indirectly influence the channel gene transcription.

Voltage-gated potassium channel, Transcription factor, Promoter

1. Introduction

Excitable cells vary in the pattern of electrical activity they spontaneously produce. The cloning of a wide variety of voltage-gated ion channels suggests that this diversity of electrophysiological phenotype is primarily determined by the assortment of channel genes expressed by each cell type. Specifically, K⁺ channels are very diverse groups of ion channels, and distinct K⁺ channels differentially influence electrical properties of cells [1–3]. For example, ventricular cardiac myocytes contain at least four types of voltage-gated K⁺ currents that differ in gating properties and sensitivities to drugs [2]. The inward rectifier sets the resting membrane potential, whereas three distinct outward currents (Ito, Iks and Ikr) influence various phases of action potential repolarization. Thus, cell type-specific regulation of K⁺ channel gene expression plays pivotal roles in determining electrophysiological phenotypes.

The Kv4 family genes are related to the Drosophila Shal and encode channels that are activated at subthreshold of action potential and rapidly inactivate [4–6]. Studies with antisense oligonucleotides [7] and a dominant-negative subunit [8] demonstrated that the channels encoded by these genes carry significant portions of the transient outward K⁺ current in cardiac myocytes and the somatodendritic A-type current in some neurons. The ventricular transient outward K⁺ current is important for the early phase of repolarization and thus influences the shape and duration of action potential, whereas the somatodendritic A-type current affects propagation and frequency of action potential in neurons [9]. The mammalian Kv4 family consists of three members, Kv4.1, 4.2 and 4.3. The expression of these genes shows overlapped but distinct spatial and temporal patterns in the brain and heart. For example, all three genes are highly expressed in the dentate granule cells in the hippocampus, while only Kv4.2 mRNA is abundant in the CA1 pyramidal neurons [10–12]. Likewise, Kv4.2 mRNA, but not Kv4.3 message, exhibits differential distribution across the left ventricular wall of the adult rat heart [13]. Moreover, the expression of Kv4.2 and Kv4.3 genes is altered under pathological conditions. In the brain, drug-induced seizure activity or chronic convulsive stimuli reduces Kv4.2 mRNA level in hippocampal neurons [11,14,15]. Similarly, prenatal exposure to methylazoxymethanol produces disorganized pyramidal-like cells [16]. Marked reductions in Kv4.2 mRNA and transient current are found in these heterotopic neurons. In the heart, Kv4.2 and Kv4.3 mRNAs and proteins are decreased in failing and hypertrophied myocardium [17]. Thus, the expression of Kv4 channel genes is cell-type specifically controlled, and dysregulated expression of these genes is associated with pathological conditions.

To elucidate molecular mechanisms controlling Kv4.2 gene transcription, we isolated and characterized the rat Kv4.2 gene promoter. In this paper, we show that the rat Kv4.2 gene uses a single transcription start site in the brain and heart. Furthermore, GATA and FOG transcription factors appear to use distinct mechanisms to regulate the minimum channel promoter with ∼200 bp in cardiac myocytes and PC12 cells.

2. Methods

2.1. Constructions

To obtain genomic clones for the rat Kv4.2 gene, a P1 genomic library was screened by PCR with primers 5′-TCTGGAGCTACAACAACAGGTCG-3′ (21–43) and 5′-CATCTCCTGGCACATCTAAACG-3′ (243–264) [18]. The obtained clones were further examined by restriction mapping and southern blot analyses with the 5′ end (1–539) and the 3′ end (1643–2241) fragments of the rat Kv4.2 cDNA [18] as probes. The ∼5.2-kb EcoRI–BamHI fragment that was positive with the 5′ end probe, but not with the 3′ end probe, was subcloned and sequenced (GenBank, AF486814). Luciferase reporter constructs were made by inserting various portions of the obtained 5′ flanking region into pGL3 basic vector (Promega) by using endogenous enzyme sites or PCR-based methods.

The rat Kv4.3 promoter was previously obtained in our laboratory [19]. The rat ANF promoter was generated by PCR with primers (5′-CAGAATTCTTTAGACCTGTATCATGTTGGCTTCC-3′ and 5′-GTGGATCCGGGCACGATCTGATGTTTGC-3′ where underlined sequences were mutated to introduce a BamHI site) using genomic DNA from rat GH₃ cells. GATA4, GATA6 and FOG2 cDNAs were generously supplied by Drs. M. Nemer (IRCM, Quebec, Canada), M. Futai (Osaka University, Japan) and E.M. Olson (University of Texas Southwestern Medical Center, Texas), respectively. GATA2, GATA3 and Nkx2.5 cDNAs were obtained as EST clones (GenBank BF312548, BM011864 and REHAA44) from ATCC. Single amino acid substitutions were introduced by using a two primer-based mutagenesis (QuickChange XL, Stratagene). EST clones and mutated clones were verified by DNA sequencing. The whole coding regions of transcription factors were subcloned into pcDNA3 (Invitrogen).

2.2. RACE analysis and RNase protection assay

Total RNAs were isolated from various tissues of Sprague–Dawley adult female rats (10–12 weeks old) by one-step extraction with acid phenol chloroform guanidium thiocyanate. Transcription start sites of the rat Kv4.2 gene was determined by a CAP-limited and RNA ligase-based 5′RACE (Invitrogen) with primers: 5′-AGTCTGGTTGATACAGGGTCGGTATATAGA-3′ (151–181) and 5′-CACCTCCTGGCACATCTAAACGCACAGG-3′ (237–264) [18]. The second-cycle PCR products with brain and heart RNAs were subcloned and sequenced.

RNase protection assays were performed with an antisense RNA probe corresponding to the XbaI–BstEII fragment of the 5′ flanking region (3039–3649, AF486814), as described previously [20,21]. Briefly, RNA probe (∼1 × 10⁵ cpm) was annealed with 10 μg of total RNA isolated from adult rat brain, heart or skeletal muscle in a solution containing 80% formamide, 0.4 M NaCl, 1 mM EDTA and 40 mM PIPES–NaOH (pH 6.4) at 50 °C overnight. Unprotected portions of RNA probe were digested with RNases A and T1. Protected fragments were recovered and separated on a 4% denaturing polyacrylamide gel with [³²P]end-labeled markers.

2.3. Cell cultures

Neonatal cardiac myocytes were prepared from 1-day-old Sprague–Dawley rat pups following the previously described method [19]. The animal investigation conforms with the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication No. 85–23, revised 1996). Briefly, ventricular tissues were obtained by removing the top part of the hearts containing atria and other tissues. The ventricular tissues were dissected into small pieces and dissociated by repeatedly incubating and pipetting at room temperature in a trypsin-containing solution. The dissociated cells were kept in an ice-cold bovine serum albumin-containing solution during this process. The cells were spread on 100 mm plastic dishes in minimum essential medium supplemented with 5% calf serum and incubated for 1 h to remove non-myocytes that were easily attached to the plastic surface. The unattached myocytes were spread in 60 mm plastic dishes at ∼1 × 10⁶ cells/dish in the same serum-containing medium. The cells were cultured with 5% calf serum overnight and then in the serum-free medium.

Non-myocytes, largely consisting of cardiac fibroblasts, were obtained by growing cells that were attached to 100 mm plastic dishes during myocyte preparation in the serum-containing medium for an additional 2 days. These cells were passed into 60-mm dishes 1 day before transfection and maintained in the serum-containing medium for overnight. The medium was then switched to the one without serum.

PC12 and NIH3T3 cells were originally obtained from ATCC and maintained in our laboratory. These cells were cultured in Dulbecco's Modified Eagle Medium supplemented with 10% fetal bovine serum. Cells were passed into 6-well plastic plates at ∼2 × 10⁵/well for transfection.

2.4. Luciferase reporter gene assays

Two days after preparation, neonatal myocytes were transfected with reporter and/or expression constructs by calcium phosphate precipitation for 2 h in the presence of 5% calf serum [19] or lipofectamine 2000 following the manufacturer's protocol (Invitrogen). Two transfection methods gave identical results in GATA and FOG2 responses. A luciferase reporter construct containing channel promoter (2.5 μg) and the renilla luciferase reporter construct with the Herpes simplex virus thymidine kinase promoter (0.5 μg, pRL-tk, Promega) were used for a 60-mm dish. For testing the effects of transcription factor, 2 μg of expression construct(s) was included. Total amounts of expression constructs were kept constant by supplementing the empty vector. Since the activity of the thymidine kinase promoter was significantly affected by the expression of GATA transcription factors, we used pRL-SV containing the SV40 minimum promoter or omitted pRL-tk for testing the effects of these transcription factors. The results normalized with pRL-SV were not significantly different from those obtained without normalization. The transfected cells were maintained in 5% calf serum-containing medium overnight, and then cultured in the serum-free medium for an additional day.

Cardiac non-myocytes were transfected with reporter and expression constructs using lipofectamine 2000 (Invitrogen). Transfected cells were cultured in the presence of 5% calf serum overnight, and in its absence for one additional day. PC12 and NIH3T3 cells were transfected using lipofectamine-plus (Invitrogen) with 0.5 μg Kv4.2 or 0.2 μg ANF promoter reporter construct and 0.2 μg pRL-tk or pRL-SV for 1 well of a 6-well plate. Expression constructs for transcription factors (total 1μg) were included in some experiments. Transfected cells were cultured in the standard serum-containing medium for 1 day prior to harvesting.

Transfected cells were harvested with ice-cold phosphate-buffered saline, and dual luciferase activities were determined according to the manufacturer's protocol (Promega). Statistical analysis of the data was performed by one-sample singed test for effects by individual transcription factors and by two-tailed Student t-test for FOG2 effects on GATA-induced changes. P<0.05 was considered significant.

3. Results

3.1. The rat Kv4.2 gene uses the same transcription start site in heart and brain

Genomic clones containing the 5′ flanking region of the rat Kv4.2 gene were obtained by screening a rat genomic P1 library by PCR with primers corresponding to the 5′ end of the Kv4.2 cDNA clone [18]. The obtained P1 clones contained 25- to 45-kb inserts. The ∼5.2-kb EcoRI–BamHI fragment containing the 5′ end of cDNA was subcloned for further analysis. Sequencing of this fragment indicated that it consists of ∼3.7-kb 5′ flanking and ∼1.5-kb coding regions (Fig. 1A). The sequence of the 5′ flanking region of the rat Kv4.2 gene exhibits significant similarity to that of the human counterpart (GenBank AC004946). The overall nucleotide identify of the region from the translation start site to ∼2 kb upstream from this site between human and rat is ∼80%. No apparent similarity is seen in the region further upstream.

Fig. 1

Structure of the 5′ flanking region of the rat Kv4.2 gene. (A) The restriction map of the 5.2-kb 5′ flanking region of the rat Kv4.2 gene (top) and sequence of the region between SmaI and the translation initiation methionine (bottom) are shown. An arrow indicates the determined transcription start site. (B) RACE analysis was performed with brain, heart and skeletal muscle RNAs as starting materials. The second cycle nested PCR products were separated on an agarose gel with molecular size markers. (C) RNase protection assays were performed with an RNA probe corresponding to the XbaI–BstEII region. Yeast RNA was used as a negative control.

To determine the transcription start sites of the Kv4.2 gene, 5′RACE analysis was performed with total RNAs isolated from adult rat brain, heart and skeletal muscle (Fig. 1B). The nested PCR with brain and heart RNAs produced single bands. Cloning and sequencing of these products revealed that transcription of the Kv4.2 gene starts at 552 bp upstream from the translation initiation site in both brain and heart (Fig. 1A). There is no TATA or CAAT in the region near this start site. To verify this determined site, we performed RNase protection assays with a probe encompassing this start site (Fig. 1C). Single protected fragments were detected with RNAs from the brain and heart, but not the skeletal muscle. The size of these protected fragments is in good agreement with the start site determined by 5′RACE analysis. Furthermore, the intensity of signals obtained by the two methods is consistent with the abundance of Kv4.2 mRNA in these tissues. Thus, the rat Kv4.2 gene utilizes the same single transcription start site in the brain and heart.

3.2. The ∼200-bp 5′ flanking region of Kv4.2 gene drives significant promoter activity

To identify regions important for transcription of the Kv4.2 gene, we generated a series of luciferase reporter gene constructs that contain various portions of the 5′ flanking region (Fig. 2). Each Kv4.2 promoter reporter gene was transfected with the normalization control pRL-tk containing the Herpes simplex virus thymidine kinase promoter into neonatal cardiac myocytes (Fig. 2A). The longest Kv4.2 construct showed very weak reporter gene expression. This was not due to its large size, because other reporter genes containing an insert with a similar size resulted in only ∼30% reductions in luciferase activity. Eliminating a ∼2-kb fragment from the 5′ end significantly increased reporter gene expression. Further deletion from the 5′ end up to −122 did not cause a significant reduction in the promoter activity. Similarly, deleting from the 3′ end to +87 did not affect reporter gene expression. These results indicate that the ∼200-bp fragment (−122 to +87) acts as the minimum promoter for the Kv4.2 gene in neonatal cardiac myocytes.

Fig. 2

Deletion analysis of the rat Kv4.2 promoter. Various portions of the Kv4.2 5′ flanking region (filled bars) were fused to the upstream of the luciferase gene (dotted bars) of pGL3. Luciferase activity was normalized with the activity of cotransfected pRL-tk, and is presented as fold increases compared to that obtained with the reporter vector without any promoter (pGL3). Columns and error bars represent the means and S.E.M., respectively. N≥6 for each construct in neonatal cardiac myocytes (A) and PC12 cells (B).

Kv4.2 gene is highly expressed in the brain and heart. Therefore, we also examined these deletion promoter reporter constructs in neuron-like PC12 cells. RT-PCR analysis found significant Kv4.2 mRNAs in PC12 cells (data not shown). Similar to the reporter gene expression in cardiac myocytes, the longest construct (−3162 to +592) showed very weak reporter gene expression in PC12 cells, and eliminating the 5′ end 2-kb fragment enhanced the expression (Fig. 2B). Likewise, the ∼200-bp region encompassing the start site (−122 to +87) drove significant reporter gene expression in these cells. However, deletion of the region between −432 and −285 caused a significant reduction in the promoter activity in PC12 cells, suggesting the presence of a PC12 cell-specific enhancer in this region. Taken together, the Kv4.2 5′ flanking region between −122 and +87 drives significant promoter activity in cardiac myocytes and PC12 cells.

3.3. GATA4 and 6 differentially activate the Kv4.2 promoter in cardiac myocytes and PC12 cells

GATA4, 5 and 6 transcription factors are important for cardiac development and function [22]. The ontogenic pattern of Kv4.2 gene expression in the heart resembles those of GATA4 and 6. These findings raise the possibility that these cardiac transcription factors might control transcription of the Kv4.2 gene. We tested this possibility by cotransfecting expression constructs for GATA 4 or 6 transcription factors with Kv4.2 promoter reporter genes into neonatal cardiac myocytes (Fig. 3A). To examine cell-type specificity of the transcription factor effects, we also used PC12 cells for these experiments (Fig. 3B). The two GATA factors differed in the ability to activate the channel promoter in the two cell types. In cardiac myocytes, GATA4 produced a larger increase (∼5-fold) in the Kv4.2 promoter activity than GATA 6 (∼2.5-fold). Increasing the amount of a GATA6 expression construct by four times did not cause further enhancement (data not shown). The GATA-induced increase in reporter gene expression was similar in all three Kv4.2 constructs that contained different lengths of the 5′ flanking region. To confirm that the observed GATA-induced up-regulation indeed occurs in cardiac myocytes, we also tested the effects of these factors in non-myocyte preparation, mostly consisting of cardiac fibroblasts. The Kv4.2 minimum promoter drove very weak luciferase expression in these cells: the normalized channel promoter activity was less than 1/20 compared to those in myocytes or NIH3T3 cells (see below). Furthermore, GATA4 did not produce a significant change in channel promoter activity (94.7±18.1%, N = 4). Thus, these transcription factors regulate the channel promoter in cardiac myocytes, but not fibroblasts.

Fig. 3

Differential activation of the rat Kv4.2 promoter by GATA4 and 6 in neonatal cardiac myocytes and PC12 cells. Kv4.2 luciferase reporter constructs with different lengths (−1094 to +592, −285 to +592 or −122 to +87) were transfected with the indicated GATA factor expression construct or empty vector (pcDNA3, None) into neonatal cardiac myocytes (A) or PC12 cells (B). Columns and error bars indicate the mean and S.E.M., respectively. N≥10 for myocytes, and N = 8 for PC12 cells. *P<0.05 compared to none (one-sample singed test).

In contrast to the preferential activation by GATA4 in cardiac myocytes, GATA6 led to a larger enhancement of promoter activity than GATA4 in PC12 cells (Fig. 3B). Again, the three constructs with different lengths of the 5′ flanking region similarly responded to GATA4 or 6. Since GATA2 and 3 are involved in the differentiation of several types of neurons, we also tested whether these factors might affect Kv4.2 promoter [23–29]. However, the two factors did not produce any significant effects (Fig. 3B). The GATA4/6-induced increases in the two cell types were specific for the Kv4.2 promoter, since these transcription factors at the same or four times larger amounts produced no significant changes in the rat Kv4.3 [19] or SV40 minimum promoters (data not shown). Thus, GATA4 and 6 transcription factors differentially regulate the activity of the Kv4.2 promoter in cardiac myocytes and PC12 cells. These results also indicate that the minimum channel promoter is sufficient for the GATA-induced regulation.

3.4. FOG2 differentially regulates the GATA-induced increases in the Kv4.2 promoter activity in cardiac myocytes and PC12 cells

The function of GATA factors is regulated by the U shape-related cofactors FOG1 and 2. Specifically, FOG2 is highly expressed in cardiac and neuronal tissues and influences the activity of GATA factors [30–33]. Therefore, we examined whether FOG2 might influence the observed GATA-induced regulation of the Kv4.2 promoter (Fig. 4). FOG2 itself produced a marginal increase in promoter activity in cardiac myocytes and PC12 cells. However, when both GATA and FOG2 were expressed, the promoter activity was markedly reduced in cardiac myocytes (Fig. 4A). On the contrary, GATA and FOG2 synergistically activated the channel promoter activity in PC12 cells (Fig. 4B). FOG2 did not alter the selectivity of GATA factors: its coexpression with GATA6 produced a larger increase than that with GATA4 in these cells. Thus, FOG2 differentially controls the GATA-induced regulation of the Kv4.2 promoter in cardiac myocytes and PC12 cells.

Fig. 4

Differential effects of FOG2 on the GATA-induced regulation of the rat Kv4.2 promoter in neonatal cardiac myocytes and PC12 cells. The minimum Kv4.2 promoter reporter gene (−122 to +87) was transfected with expression constructs for the indicated combination of transcription factors into neonatal cardiac myocytes (A) or PC12 cells (B). Total amount of expression constructs was kept constant by including an empty vector (pcDNA3) in each transfection. Columns and error bars indicate the mean and S.E.M., respectively. N≥8 for myocytes, and N = 8 for PC12 cells. *P<0.05 compared to none (one-sampled singed test), and ^#P<0.05 compared to GATA4 or 6 (two-tailed t-test).

To elucidate the mechanisms by which GATA and FOG2 regulate the Kv4.2 promoter, we used GATA factors with single amino acid substitutions [30,34]. These mutants lack the ability to associate with FOG2, but are capable of binding to a target DNA to alter transcription of the gene. As a control, we used the ANF promoter that is known to be regulated by GATA and FOG2 [32]. As expected, FOG2 suppressed the GATA4- or GATA4 and Nkx2.5-induced increases in the ANF promoter activity when wild-type GATA4 was expressed, but failed to affect these elevations when GATA4 mutants were used (Fig. 5D). Similar to the effects on the ANF promoter, these GATA4 mutants were still able to increase the activity of the Kv4.2 promoter by themselves, but did not support the FOG2-induced suppression in cardiac myocytes (Fig. 5A). In contrast, coexpression of GATA mutants and FOG2 resulted in a synergistic enhancement of the channel promoter in PC12 cells (Fig. 5B). In addition, we found that wild-type or mutant GATA4s increase the Kv4.2 promoter activity in NIH3T3 cells (Fig. 5C). FOG2 significantly reduced the GATA4-induced increase in these cells. However, unlike its effects in cardiac myocytes, FOG2 inhibited the GATA4-induced increase even when mutant GATA4s were used. Hence, interaction of GATA4 with FOG2 influences the Kv4.2 promoter activity in cardiac myocytes, but not in PC12 or NIH3T3 cells.

Fig. 5

Distinct requirement of GATA–FOG2 complex formation for the regulation of the Kv4.2 promoter in cardiac myocytes and PC12 cells. The minimum Kv4.2 promoter reporter gene (−122 to +87) was transfected with the indicated combination of transcription factor cDNAs into neonatal cardiac myocytes (A), PC12 cells (B) or NIH3T3 cells (C). The ANF promoter reporter gene was used in place of the Kv4.2 construct (D). N≥8 for myocytes, and N = 8 for PC12 and NIH3T3 cells. *P<0.05 compared to none (one-sampled singed test), and ^#P<0.05 compared to GATA4 or 6 (two-tailed t-test). Note that mutant GATAs are capable of supporting FOG2-induced changes in (B) and (C), but not in (A) and (D).

4. Discussion

We have isolated and characterized the 5′ flanking region of the rat Kv4.2 gene. Deletion analysis indicated that the ∼200-bp 5′ flanking region drives significant promoter activity in cardiac myocytes and PC12 cells. The activity of this minimum promoter is differentially regulated by GATA and FOG2 transcription factors. GATA4 more dramatically activates the channel promoter than GATA6 in myocytes, whereas the latter transcription factor exhibits a stronger enhancement of promoter activity in PC12 cells. Furthermore, FOG2 markedly suppresses the GATA-induced activation of the promoter in myocytes, but enhances it in PC12 cells. The FOG2-induced suppression in cardiac myocytes is mediated by its association with GATA4, whereas the FOG2-induced enhancement in PC12 cells does not require the complex formation. These results suggest that the same set of transcription factors differentially control the expression of Kv4.2 gene in cardiac myocytes and PC12 cells.

GATA transcription factors bind to the consensus sequence WGATAR [35,36]. Slightly different sequences may also provide a binding site for these factors. Our results indicate that the minimum Kv4.2 promoter is sufficient for supporting the GATA-induced regulation. This region contains one potential element in the anti-parallel strand [CATC (−57 to −54) as a core]. However, mutating TC to AG in this core sequence did not affect the GATA-induced regulation or the basal activity of the channel promoter in cardiac myocytes or PC12 cells (data not shown). Therefore, the GATA-induced changes in Kv4.2 promoter activity may be indirectly mediated by the changes in expression or function of other transcription factors.

The sequence of the 5′ flanking region of the rat Kv4.2 gene exhibits an extensive similar to that of the human gene (GenBank AC004946). The identity between the rat and human sequences is nearly 80% in the region from the determined start site to ∼2 kb upstream from this site. The 5′ untranslated region other than a 30-bp portion near the start site shows rather low conservation between the two species with ∼50% identity. Thus, the region with high conservation may contain elements that are important for transcriptional regulation of the Kv4.2 gene. However, it is important to note that the expression of Kv4.2 gene displays significant differences between the two species: the channel mRNA is high and region-selective in the rat heart, whereas it is very low in the human tissue [13]. These findings raise the possibility that sequence differences in the channel promoter between the two species may be responsible for the distinct expression patterns. Alternatively, the mechanisms that control the channel gene transcription may differ between the two species. Further analyses of the rat and human Kv4.2 gene promoters may resolve these issues.

The present study revealed that GATA6 more dramatically regulates the Kv4.2 promoter than GATA4 in PC12 cells. The data also support the important role of FOG2 in the regulation of the Kv4.2 promoter in neuronal cells. However, the expression of GATA4 or 6 is less significant in neuronal tissues of adult rats, while FOG2 is abundantly expressed in these tissues [33]. This raises the issue of which GATA or GATA-like transcription factors may be involved in the regulation of Kv4.2 gene expression in adult neurons. Many studies have shown that GATA2 and 3 play important roles in the early development and differentiation of the central and peripheral neurons [23–29,37]. In some cases, these GATA-induced responses are regulated by FOG2. Thus, GATA2 or 3, instead of GATA4 or 6, may be the physiological regulator of the Kv4.2 promoter in neurons. However, we found that GATA2 or 3 produces no significant changes in the Kv4.2 promoter activity. Therefore, it is likely that GATA4 and/or 6 are physiological regulators of neuronal Kv4.2 gene expression. Taken together, the identified GATA4/6 and FOG2-induced regulation of Kv4.2 gene expression may constitute a part of the transcription factor-mediated commitment to certain neuronal cell types with distinct electrophysiological phenotypes.

FOG2 has been shown to produce distinct effects on GATA-induced regulation of several promoters depending on cellular context. For example, FOG2 enhances and suppresses the GATA-induced activation of the α-MHC promoter in COS cells and cardiac myocytes, respectively [32]. However, the mechanisms by which FOG2 causes the opposite changes in the GATA-induced regulation remain unknown. In this study, we found that FOG2 also produced opposite effects on the GATA-induced regulation of the Kv4.2 promoter in different cell types: FOG2 inhibits the GATA-induced activation of the channel promoter in cardiac myocytes and NIH3T3 cells, whereas the coregulator enhances it in PC12 cells. Our data with GATA mutants further indicate that FOG2 eliminates the GATA-induced activation in cardiac myocytes by interacting with GATAs. In contrast, the FOG2-induced suppression in NIH3T3 cells and enhancement in PC12 cells do not require its association with GATAs. These results suggest that GATAs and FOG2 act independently to influence the channel promoter in the latter two cell types. However, FOG2 itself resulted in only small decrease and increase in promoter activity in NIH3T3 and PC12 cells, respectively. Therefore, the FOG2-induced regulation requires the presence of GATAs. The simplest explanation for these phenomena is that FOG2 can indirectly influence the activity of GATAs, in addition to its direct binding to these factors. In this regard, FOG2 is a relatively large protein with seven zinc fingers and may act as a scaffold for various transcription factors. FOG2-interacting transcription factors present in each cell type may determine the outcome of its influence on the channel promoter.

In addition to the opposite effect of FOG2 on the GATA-induced regulation, we also observed that the size and direction of the GATA or FOG2 effects on the Kv4.2 promoter differ in various fibroblastic cells. For example, GATA4 caused 5–7-fold increases in the Kv4.2 promoter in NIH3T3, but less than 50% elevation in HEK293 (data not shown). Similarly, GATA4 did not produce any appreciable induction of the channel promoter activity in cardiac non-myocytes, mostly consisting of fibroblasts. Likewise, FOG2 by itself increased and decreased the channel promoter in CHO-K1 and NIH3T3 or HEK293, respectively. Therefore, fibroblasts from distinct origins are quite different in transcriptional capacity.

Activation of GATA4 is associated with cardiac hypertrophy [38]. Since Kv4.2 mRNA and proteins are decreased in hypertrophied myocardium, exogenous expression of GATA4 might be expected to decrease activity of the Kv4.2 promoter. However, we found that GATA4 enhances Kv4.2 promoter activity in neonatal ventricular myocytes. Our data also show that Kv4.2 and the hypertrophy marker ANF promoters are similarly regulated by GATA4 and FOG2 in myocytes. A similar discrepancy is seen with the regulation of the α-MHC gene: its expression is decreased in hypertrophied myocardium, whereas GATA4 enhances its promoter [32]. We also found that GATA mutants and FOG2 similarly regulate Kv4.2 and α-MHC promoters in cardiac myocytes and NIH3T3 (unpublished observation). These results suggest that a common mechanism other than GATA4 activation may be involved in the disease-associated down-regulation of Kv4.2 and α-MHC genes.

The shape and duration of cardiac action potential dramatically change during development. For example, rat ventricular myocytes exhibit pronounced action potential shortening in the early development [39]. This change in action potential duration parallels with appearance of the transient outward K⁺ current. Since the levels of GATA4 and 6 dramatically increase during early development, the observed GATA-induced activation of Kv4.2 promoter may significantly contribute to the developmental increase in the transient outward K⁺ channels. In addition to the developmental change, cardiac Kv4.2 channel expression is dynamically regulated by hormones [40–44] and exhibits a daily fluctuation [45]. Although other factors, such as KChIP auxiliary subunits, may significantly influence the expression of functional Kv4.2 channels, changes in Kv4.2 gene transcription can lead to alterations in the number of functional channels, which in turn influence the shape and duration of action potentials. Thus, the identified regulation of the channel promoter by GATA and FOG2 may also participate in some of these physiological adaptive changes in electrical properties of cardiac tissues.

Acknowledgements

We thank Drs. M. Nemer for GATA4 cDNA, M. Futai for GATA6 cDNA and E.M. Olson for FOG2 cDNA. This work was supported by a grant from the NIH HL63123 (to KT).

Rat Kv4.2 genomic sequence has been deposited to GenBank (Accession number AF486814).