Ets and GATA Transcription Factors Play a Critical Role in PMA-Mediated Repression of the ckβ Promoter via the Protein Kinase C Signaling Pathway

Background Choline kinase is the most upstream enzyme in the CDP-choline pathway. It catalyzes the phosphorylation of choline to phosphorylcholine in the presence of ATP and Mg2+ during the biosynthesis of phosphatidylcholine, the major phospholipid in eukaryotic cell membranes. In humans, choline kinase (CK) is encoded by two separate genes, ckα and ckβ, which produce three isoforms, CKα1, CKα2, and CKβ. Previous studies have associated ckβ with muscle development; however, the molecular mechanism underlying the transcriptional regulation of ckβ has never been elucidated. Methodology/Principal Findings In this report, the distal promoter region of the ckβ gene was characterized. Mutational analysis of the promoter sequence and electrophoretic mobility shift assays (EMSA) showed that Ets and GATA transcription factors were essential for the repression of ckβ promoter activity. Supershift and chromatin immunoprecipitation (ChIP) assays further identified that GATA3 but not GATA2 was bound to the GATA site of ckβ promoter. In addition, phorbol-12-myristate-13-acetate (PMA) decreased ckβ promoter activity through Ets and GATA elements. PMA also decreased the ckβ mRNA and protein levels about 12 hours after the promoter activity was down-regulated. EMSA further revealed that PMA treatment increased the binding of both Ets and GATA transcription factors to their respective DNA elements. The PMA-mediated repressive effect was abolished by chronic PMA treatment and by treatment with the PKC inhibitor PKC412, but not the PKC inhibitor Go 6983, suggesting PKCε or PKCη as the PKC isozyme involved in the PMA-mediated repression of ckβ promoter. Further confirmation by using PKC isozyme specific inhibitors identified PKCε as the isozyme that mediated the PMA repression of ckβ promoter. Conclusion/Significance These results demonstrate the participation of the PKC signaling pathway in the regulation of ckβ gene transcription by Ets and GATA transcription factors.

Introduction expression can be elucidated by characterization of the ckb promoter. Previous studies in various mammalian cells showed that phorbol esters stimulate the incorporation of choline into PC [13,14]. Phorbol 12-myristate 13-acetate (PMA) is a direct activator of protein kinase C (PKC), and stimulates both the cellular uptake of radiolabeled choline and its incorporation into PC [15,16]. Previously, we isolated a 2000 bp human ckb promoter that was repressed by PMA treatment [17]. In this report, we localized the repressive effect of PMA to the 22000/21886 region upstream of the ATG translation start site, which is bound by Ets and GATA transcription factors. We also demonstrate that PMA exerts its effect on the ckb promoter through a PKC-dependent pathway.

Materials and Methods
In silico analysis of the ckb promoter region The 2000 bp upstream region of the ckb gene (transcript NM_005198) was analyzed using MatInspector 8.0 [18] and TFSEARCH [19] to identify putative transcription factor binding sites. CpG islands within the ckb promoter (the 2000 bp region upstream of the ckb gene) were identified using CpGplot [20] and CpGIS [21]. CpGplot defines a CpG island as a DNA region with an observed/ expected ratio .0.60, a length .200 bp, and GC content .50% [20,22]. CpGIS defines a CpG island as a sequence having an observed/expected ratio .0.65, a length .500 bp, and GC content .55% [21].

Site-directed mutagenesis
Mutation of the Ets and GATA consensus binding sites was carried out by onestep PCR mutagenesis using Platinum Pfx DNA polymerase (Invitrogen, USA). Primers used to introduce the mutations are shown in Table 1. The pGL4.10mut(Ets), pGL4.10-mut(GATA), and pGL4.10-mut(Ets/GATA) mutant constructs were verified by DNA sequencing.

Transient transfection and luciferase assay
Transfection was performed using Lipofectamine 2000 (Invitrogen/Life Technologies) according to the manufacturer's instructions. Briefly, MCF-7 cells were plated in 100 mL of medium/well on a 96-well plate at a density of 1.5610 4

Western detection of CKb
Thirty micrograms of cell lysate were separated on 10% SDS-PAGE and electroblotted at 13 volts for 2 hr onto nitrocellulose membrane. After the transfer step, the nitrocellulose membrane was immersed in blocking buffer (10 mM Tris-HCl, pH 7.5 containing 5% (w/v) skim milk, 150 mM NaCl and 0.1% (v/v) Tween 20) for 1 hr with gentle agitation at room temperature. Next, the membrane was incubated with 1:1000 diluted CKb antibody [2] in the blocking buffer at 4˚C for overnight. The membrane was washed 3 times with Tris-buffered saline for 1 hr before incubation with 1:5000 diluted HRP-conjugated rabbit IgG secondary antibody for 1 hr at room temperature. Subsequently, the membrane was washed 3 times for 30 min with Tris-buffered saline and the signal was detected with SuperSignal West Pico chemiluminescent substrate (Thermo Scientific, USA) and Fusion FX luminescence detector system (Vilber Lourmat, France). For loading control, the same membrane was stripped with stripping buffer (200 mM glycine, pH 2.5 containing 0.1% (w/v) SDS and 1% (v/v) Tween 20) and re-probed with 1:1000 diluted b-actin antibody (Abcam, USA). Signal intensities on the blots were analyzed by using ImageJ version 1.49b software (downloaded from http://imagej.nih.gov/ij/). The integrated densities of bands were measured in triplicate and the average values were then normalized to the corresponding integrated densities of b-actin signals.

Electrophoretic mobility shift assays (EMSAs)
Nuclear extract from MCF-7 cells was prepared using the NE-PER nuclear and cytoplasmic extraction kit (Pierce, USA) according to the manufacturer's protocol. All DNA probes used in EMSAs were synthesized by 1st Base (Malaysia) and labeled with biotin using a biotin 39 end DNA labeling kit (Pierce).
Complementary DNA probes were annealed with each other at a 1:1 molar ratio by heating at 95˚C for 5 min, followed by a gradual decrease to room temperature using a thermocycler set to decrease 1˚C per cycle. EMSAs were performed using the LightShift chemiluminescent EMSA kit (Pierce). The binding reaction mixture contained 200 fmol biotin-labeled probe, 10 mg of nuclear extract, 16 binding buffer, 2.5% glycerol, 5 mM MgCl 2 , 1 mg poly(dI-dC), 2 mg BSA, and 0.05% NP-40. For the competition assay, 100-fold unlabeled double-stranded DNA was added into the reaction mixture. DNA probes used in EMSAs are listed in Table 2.
For the supershift assay, 5 mg of GATA2 (Abcam, United Kingdom) or GATA3 (Millipore, USA) antibody was added into the reaction mixture and incubated on ice for 30 min prior the addition of the biotinylated probe. DNA-protein complexes were resolved by electrophoresis on a 6% (w/v) non-denaturing polyacrylamide gel in 0.56 Tris-borate-EDTA (TBE) at 4˚C. The gel was run until the bromophenol blue dye has reached 3/4 of the length of the gel for EMSA or until the dye has completely migrated out of the gel for supershift assay (to increase the possibility of detecting slower-migrating complexes). The biotinlabeled DNA was electro-blotted onto a Biodyne B nylon membrane (Pierce) and the membrane was cross-linked by a UV transilluminator (Spectronics, USA) at 312 nm for 15 min. The signal was then developed with a chemiluminescent nucleic acid detection module (Pierce). The images were developed by exposing the signal to X-ray film for 2-5 min or detected with Fusion FX luminescence detector system (Vilber Lourmat, France).

Chromatin immunoprecipitation (ChIP)
Chromatin immunoprecipitation (ChIP) assay was performed using Pierce Agarose ChIP Kit (Thermo Fisher Scientific, USA) according to the manufacturer's instructions. MCF-7 cells were treated with 1% (v/v) formaldehyde (Sigma-Aldrich, USA) for 10 min to induce protein-DNA cross-linking. Glycine solution (16) was added to the mixture, incubated for 5 min and the cells were washed with PBS twice. The cells were harvested by scraping, then pelleted and resuspended in 400 mL of lysis buffer. Samples were centrifuged and the nuclear pellet was mixed with 400 mL of micrococcal nuclease (MNase) solution at 37˚C for 15 min. The reaction was stopped by adding 40 mL of MNase Stop Solution and the samples were centrifuged to prepare the supernatants (digested chromatin) for immunoprecipitation. The DNA-protein complexes were immunoprecipitated at 4˚C for overnight with 10 mg of GATA2 or GATA3 antibody. Immunoprecipitation with 2 mL of pre-immune normal rabbit IgG was used as control. DNA recovered from the immunoprecipitated samples was PCR amplified by using ckb-2000-59 and ckb-Distal-39 primers (Table 1) followed by agarose gel electrophoresis. The positive PCR product was verified by sequencing.

Results
The human ckb promoter is a TATA-less, CpG island-containing promoter A 2000 bp region 59 of the human ckb gene was first analyzed in silico. For convenient description hereafter, the nucleotide A at the ATG translational start site of the ckb gene was designated +1, and nucleotides upstream of +1 were assigned negative numbers. The putative promoter sequence was analyzed for potential transcription factor binding sites using MatInspector 8.0 and TFSEARCH (Fig. 1). The promoter region contains several potential binding sites for Sp1, Ets, GATA factors, SREBP, MZF1, NF-kB, AP-1, and E2F (Fig. 1). The analysis showed that the ckb promoter contains neither a CAAT box nor a recognizable consensus TATA box in close proximity to the transcription start site, which is typical of GC-rich promoters. The presence of numerous Sp1 binding sites indicates that the ckb promoter contains a high GC content. Thus, CpGPlot and CpGIS were used to identify potential CpG island(s) in the ckb promoter. Two potential CpG islands were identified by CpGPlot (Fig. 2), spanning the region between 2989 and 2713 (percent GC content: 59.21% and observed/expected ratio: 0.73) and the region between 2666 and 256 (percent GC content: 71.52% and observed/expected ratio: 0.94). CpGIS analysis identified one potential CpG island covering both the putative CpG islands that were predicted by CpGPlot. This CpG island is 1085 bp long, stretches from 21085 to 21, and has a GC percentage of 65.6% and an observed/expected ratio of 0.864.

Repressive regions are identified at the ckb promoter
To identify the important cis-acting regulatory elements in the ckb promoter, various fragments of the ckb promoter were cloned upstream of the firefly luciferase reporter vector. An initial deletion analysis of the ckb promoter was performed with four ckb promoter constructs [pGL4.10-ckb(22000/21), pGL4.10-ckb(21477/21), pGL4.10-ckb(2914/21), and pGL4.10-ckb(2519/21)] that contain promoter fragments of different lengths. Two repressive regions were identified in the first deletion analysis (Fig. 3A). Deletion of the 59 sequence from the parental promoter construct to position 21477 resulted in a dramatic increase (809%) in promoter activity compared to that of the full-length 2000 bp promoter. Deletion of the region between 2519 to 2914 increased the promoter activity by 66%. This result showed that there is at least one inhibitory element located within the region of 22000/21477 and another within 2914/2519. The deletion analysis also showed that the basal promoter activity of ckb was dependent on the promoter region between 2519 and 21. Subsequently, another series of deletion mutants were constructed to narrow down the region between 22000 and 21477 that contained the inhibitory element(s). As shown in Fig. 3B, deletion of the sequence between 22000 and 21886 increased the promoter activity by 800%, demonstrating that this region significantly repressed the transcription of the ckb promoter. Our subsequent experiments focused on the Ets (21954/21966) and GATA (21950/21959) binding elements located in this Ets and GATA elements are responsible for the repression of ckb promoter activity Bioinformatic analysis revealed that a canonical Ets binding site and a GATA transcription factor binding site were located within the 22000/21886 region (Fig. 1). Therefore, point mutations were introduced in these binding sites to investigate the importance of specific sequences in the repression of the ckb promoter. Mutation of the Ets element increased ckb promoter activity to 513% of the wild-type promoter activity, while disruption of the core sequence of the GATA binding site markedly increased the promoter activity to 703% of the wild type (Fig. 4). Results from these experiments were consistent with the deletion analysis, in which the removal of both the Ets and GATA binding sites (59 deletion from 22000 to 21886) caused a significant loss of ckb promoter repression. Our data suggest that both the GATA and Ets sites located between 22000 and 21886 are negative regulatory elements and these two factors may interact to repress ckb promoter activity.

Ets and GATA3 transcription factors bind to the ckb distal promoter
EMSAs with a biotin-labeled DNA probe were used to assess the binding of the Ets and GATA repressive elements (between 22000 and 21886 in the ckb promoter) by transcription factors from nuclear extracts from MCF-7 cells. Fig. 5A shows that the biotin-labeled DNA probe produced two slowly migrating shifted complexes, indicating two different nuclear proteins were bound to the probe. Mutations of either Ets, GATA or both elements in the promoter sequence markedly reduced the formation of the upper shifted complex, suggesting that the protein component of the complex was Ets-and GATA-related. The binding specificity of Ets and GATA in the shifted complexes was verified by competition binding assays with DNA probes containing consensus binding sequences for Ets [25] and Ets/GATA [26]. The core motif, (A/T)GATA(A/G) in the Ets/GATA probe is recognized by six members (GATA1 to GATA6) of GATAfamily transcription factors [27,28]. The competition assay presented in Fig. 5B showed that the upper shifted complex was nearly eliminated by a 100-fold molar excess of unlabeled Ets consensus probe, while both shifted complexes were significantly reduced when the unlabeled Ets/GATA consensus probe was included in the binding reaction. These results demonstrated that the upper shifted complex was due to the Ets transcription factor and that both Ets and GATA transcription factors were bound to the repressive elements on the ckb distal promoter.
Next, GATA2 and GATA3 antibodies were used in a supershift assay to determine if any of these two GATA transcription factors was involved in the formation of complex with the ckb promoter. As shown in Fig. 5C, GATA2 antibody did not change the mobility and intensity of the DNA-protein complexes. The addition of GATA3 antibody into the EMSA did not produce other slower-migrating complex. However, it reduced the intensity of both the original upper and lower shifted complexes possibly by disrupting the interaction of GATA3 with ckb promoter. The results showed that GATA3 but not GATA2 was bound to the ckb distal promoter. Unbound probes were not observed in Fig. 5C because the gel was run for extended duration in order to increase the possibility of detecting other slower-migrating complexes. Similar results (reduced intensity of shifted complexes) were also obtained when the gel was not over-run and the free probes were still visible in an earlier supershift experiment with GATA3 antibody (S1 Figure).
Chromatin immunoprecipitation assay was performed to confirm the intracellular binding of GATA3 transcription factor to the ckb distal promoter. Protein-DNA complexes were immunoprecipitated with GATA2 or GATA3 antibodies, followed by PCR amplification using the primers flanking the GATA binding site of the ckb promoter. Fig. 5D showed that immunoprecipitate obtained with GATA3 antibody yielded a prominent PCR product at the expected size of 150 bp, whereas immunoprecipitates obtained with pre-immune rabbit IgG and GATA2 antibody produced much lower amount of amplified products. Sequencing of the PCR product obtained with GATA3 immunoprecipitate confirmed the specificity of the reaction. These results further support the binding of GATA3 to the GATA binding site in ckb distal promoter.
PMA represses ckb promoter activity PMA regulates certain Ets and GATA family transcription factors by activating the PKC-mediated mitogen-activated protein kinase (MAPK) pathway [29,30]. This prompted us to examine the effect of PMA on ckb promoter activity. The effects of PMA on PKCs are dependent on the PMA concentration and the duration of treatment. Short-term treatment with a low concentration of PMA activates PKCs, while long-term exposure to a high concentration has an inhibitory effect on PKCs [31,32]. In this study, MCF-7 cells were treated with different concentrations of PMA (10, 20, and 30 ng/mL) for 6 hr after transfection with the wild-type pGL4.10-ckb(22000/21) reporter plasmid. PMA repressed ckb promoter activity in MCF-7 cells starting at 10 ng/mL and reaching a maximal effect at 20 ng/mL (Fig. 6A). However, a 6 hr exposure to 30 ng/mL PMA attenuated the repression of ckb promoter activity. The effect of PMA treatment duration on ckb promoter activity was examined in the next series of experiments, ranging from 6 to 72 hr with a PMA concentration of 20 ng/mL. As shown in Fig. 6B, the maximum repressive effect of PMA occurred at 6 hr after treatment (approximately 38% of the DMSO control). Promoter activity began to increase and was almost equal to that of the DMSO control treatment after 12 hr of PMA treatment. Overall, chronic or extended PMA treatment eliminated the PMA   on the ckb promoter activity, suggesting that PMA-sensitive PKC isozymes are involved in the PMA-mediated repression of the ckb promoter. This result also ruled out the involvement of PKCf, as this isozyme is known to be resistant to chronic PMA treatment [31,33].
The effect of PMA treatment on endogenous ckb gene expression was also investigated. Real-time PCR quantifications of ckb mRNA and Western detection of CKb protein were performed on DMSO (control) and PMA treated (20 ng/mL for 6 to 24 hr) MCF-7 cells. Based on the results shown in Fig. 6C, PMA treatment resulted in significant reduction of ckb mRNA (by 20%) and protein (by 30%) expressions compared to controls after 12 hr. There was no significant difference for mRNA and protein levels between PMA treated samples and controls for the other treatment durations.
PKC412 and PKCe inhibitor peptide abolished the repressive effect of PMA on ckb promoter activity PKC412 (midostaurin, or CGP 41251), an inhibitor of PKCa, 2b, 2c, 2d, 2e, 2g, and 2f [34], completely abolished the PMA-induced down-regulation of the ckb promoter, whereas Go 6983, an inhibitor of PKCa, 2b, 2c, 2d, and 2f [32,35], did not show the same effect (Fig. 7A). The ckb promoter activity was also increased by the individual treatment with PKC412 but not Go 6983. These findings, along with the resistance of PKCf to chronic PMA treatment, suggest that PKCe or PKCg is most likely the PKC isozyme that mediates the repressive effect of PMA in this system. Subsequently, PKCe and PKCg specific inhibitors were used to identify the isozyme involved. The results in Fig. 7B show that PKCe inhibition abolished the PMA repression of ckb promoter. PKCg inhibition did not affect the down-regulation of ckb promoter by PMA treatment. The promoter activity of treatment with PMA and PKCg inhibitor was also significantly lower than treatment with PKCg alone. This further excludes the involvement of PKCg isozyme in the PMA repression of ckb promoter.
Ets and GATA binding is required for PMA-mediated repression of the ckb promoter To investigate the role of the Ets and GATA binding sites in the PMA-mediated repression of the ckb promoter, the wild-type promoter, pGL4.10-ckb(22000/ 21), and constructs in which the Ets and GATA binding sites were mutated were transfected into MCF-7 cells followed by treatment with or without PMA. Fig. 8 shows that PMA treatment decreased the promoter activities of pGL4.10-blot shown is representative of three independent experiments that produced similar results. (C) Supershift assay was performed by using GATA2 or GATA3 antibody in the EMSA. The blot shown is representative of two independent experiments that produced similar results. (D) ChIP was performed with GATA2 or GATA3 antibody and pre-immune normal rabbit IgG as control. Lane M: GeneRuler DNA Ladder Mix, Lane T: total input positive control (unprocessed chromatin), Lane N: pre-immune normal rabbit IgG (negative control).
doi:10.1371/journal.pone.0113485.g005 ckb(22000/21), pGL4.10-mut(Ets), and pGL4.10-mut(GATA) to 38%, 57%, and 68%, respectively, as compared to the control (DMSO treatment). This result indicated the participation of both Ets and GATA in the ckb promoter repression by PMA. The involvement of other transcription factors in PMA-mediated downregulation of ckb promoter activity cannot be ruled out because the mutation of both the Ets and GATA sites did not completely abolish the effect of PMA treatment (Fig. 8). EMSA was performed to investigate the effect of PMA treatment on Ets and GATA binding to the ckb promoter. As shown in Fig. 9, PMA increased the binding of Ets and GATA transcription factors to the ckb promoter and thereby suppressed its activity. Current data provide evidence that supports the Ets and GATA binding proteins as the key transcription factors involved in the PMAmediated down-regulation of the ckb promoter via activation of the PKC signaling pathway.

Discussion
In this study, the promoter region of the ckb gene was isolated and characterized to understand the transcriptional regulation of this enzyme. We have previously shown that a promoter region located 2000 bp upstream of the ckb start codon was transcriptionally active in MCF-7 cells [17]. The ckb promoter is a TATA-less, GC-rich promoter that may belong to the family of housekeeping gene promoters. In general, promoters that lack TATA-binding sites but contain CpG islands contain multiple GC box motifs that serve as binding sites for the Sp1 transcription factor [36,37]. The ubiquitous transcription factor Sp1 serves as a constitutive activator of housekeeping genes by recruiting TATA-binding protein (TBP) to promoters lacking a recognizable consensus TATA box [37]. These observations are in agreement with our preliminary sequence analysis showing that the ckb promoter contains multiple Sp1 binding sites, which reflects the GCrich nature of this region.
In this study, Ets and GATA binding sites located in the ckb distal promoter region were shown to be essential for the repression of this promoter's activity. Several Ets-related and GATA-related transcription factors are downstream transcriptional effectors in the PKC-dependent pathway, which is inducible by PMA [29,30]. PMA stimulates PKC by binding to the C1 region of the PKC regulatory domain [38]. This knowledge prompted us to analyze the transcriptional regulation of the ckb gene by PMA. Two observations supported the involvement of PKC in the PMA-mediated repression of ckb promoter activity: (i) sensitivity to chronic PMA treatment and (ii) inhibition of the PMA-mediated effect by the PKC inhibitor PKC412. PKC is a serine/threonine-specific protein kinase comprising at least 12 related isozymes that can be categorized into three groups [39]. The ''conventional'' PKC members (cPKCs a, BI, BII, and c) are activated by Ca 2+ , phosphatidylserine, and diacylglycerol/phorbol esters. Members of the second group, known as ''novel'' PKCs (nPKCs d, e, g, and h), are unresponsive to Ca 2+ but are stimulated by phosphatidylserine and diacylglycerol/ phorbol esters. The third group, ''atypical'' PKCs (aPKCs l and f), are activated by phosphatidylserine but are unresponsive to both Ca 2+ and diacylglycerol/ phorbol esters. PKCm is a member of the PKC family that does not fit into any of the known PKC subgroups [40]. In the current study, it was deemed important to identify the main PKC isozyme involved in the PMA-mediated repression of the ckb promoter to better elucidate the signaling mechanism that regulates ckb transcription. The PKCh isozyme is not expressed in MCF-7 cells [41]. Therefore, the involvement of PKCh in the transcriptional regulation of ckb can be ruled out because MCF-7 cells were used in this work. All PKC isozymes are sensitive to chronic PMA treatment except PKCf [31,33]. We observed that chronic PMA treatment led to complete abrogation of the PMA-mediated repression of the ckb promoter, demonstrating that PMA-sensitive PKC(s) participated in the negative regulation of the ckb promoter and ruled out the involvement of the PKCf isozyme (because of its PMA resistance). The PKC isozymes responsible for the PMA-Mediated Repression of ckb Promoter suppression of the ckb promoter can be narrowed down to either PKCe or PKCg based on the complete abolishment of the PMA-mediated effect after treatment with PKC412 but not Go 6983, which does not target these two isozymes. Our subsequent experiments using specific inhibitors for the two isozymes revealed that PKCe was the PKC isozyme involved in the PMA-induced suppression of ckb promoter. Previously, Kobayashi et al. [42] reported that PMA treatment of bipolar undifferentiated CG-4 cell line down regulated PKCa and PKCb II more rapidly than PKCe, which was still detectable by Western blot after 12 hr PMA treatment. On the contrary, the results in this study (Fig. 6B) showed that the PMA-mediated repression of ckb promoter activity was abolished after 12 hr PMA treatment, indicating that the PKCe could be down regulated more rapidly by PMA treatment in MCF-7 cells than in the cell line used by Kobayashi et al.
The effects of PMA treatment on ckb mRNA and protein levels were slower (12 hr) and less pronounced than its effect on promoter activity (6 hr). However, both followed the same trend whereby chronic PMA treatment abolished the effects. Since the samples were taken every six hours, lower ckb mRNA and protein levels caused by the strong suppression of promoter activity might be detected between 6 to 12 hr PMA treatments.
The present study shows that ckb promoter activity was regulated by a PMAinduced PKC-dependent pathway. Mutation of Ets and GATA binding elements located on region 21950 to 21966 of the ckb promoter confirmed the importance of these two binding sites for the PMA-induced, PKC-dependent repression of the ckb promoter. The importance of these Ets and GATA elements was further demonstrated by EMSA results showing that PMA treatment increased the binding of both Ets and GATA factors to their respective binding elements on the ckb promoter. It must be noted that a minor participation of other elements at the same promoter region cannot be excluded, as mutation of the Ets/GATA elements did not completely abolish the effect of PMA. Supershift assay was not performed to identify the binding protein of Ets motif in ckb promoter from more than 30 Ets transcription factors in human [43]. However, we predicted that Ets1 and Ets2 are the potential binding proteins of the Ets site based on their overexpression in breast carcinomas like MCF-7 used in this study [44,45] and their regulation by PMA through PKC pathway activation [30,46]. Although this study clearly indicated that Ets and GATA transcription factors are required for PMA-induced repression of the ckb promoter, the underlying mechanism of PKC involvement remains to be established. GATA family transcription factors are regulated by PMA via the MAPK signaling pathway [29,47]. PMA has been shown to directly induce GATA4 phosphorylation via activation of the MAPK family member ERK, without involvement of the PKC signaling pathway [29]. By contrast, Li et al. [47] showed that PMA repressed erythroid differentiation-associated gene (EDAG) expression via the PKC-MAPK-GATA1 pathway. The abolishment of the PMA effect on the ckb promoter by the PKC inhibitors PKC412 and PKCe inhibitor peptide suggested that PMA-induced binding of GATA transcription factors was mediated by MAPKs through a PKCedependent signaling pathway. Similarly, PMA activated PKCe to stimulate endothelin-converting enzyme 1 (ECE-1) activity through an MAPK-dependent Ets-1 pathway [48]. Activation of PKCe by PMA also induced translocator protein (TSPO) gene expression through the increased binding of c-jun and Ets-related GA-binding protein (GABP) to the AP-1 and Ets sites, respectively, on the TSPO promoter [49,50]. The MAPK (Raf-1-MEK1/2-ERK1/2) pathway was identified as the downstream target of PKCe that mediates the PMA effect on TSPO gene expression [50]. These observations support the hypothesis in this study that the binding of Ets and GATA transcription factors to the ckb promoter is induced by PMA through the activation of PKCe and MAPKs.
The activity of Ets family transcription factors is regulated by their interactions with other adjacently located transcription factors [51]. The formation of proteinprotein complexes on the promoter can modulate the transcriptional activation or repression properties of the two partners and allow for crosstalk between different signal transduction pathways [43]. Functional interactions between Ets1/Ets2 and GATA3 have been shown to synergistically increase the activity of the human interleukin-5 promoter in the presence of PMA [52]. Likewise, in this study Etsand GATA-related proteins could synergistically and cooperatively repress ckb promoter activity due to their adjacently located binding sites.
In summary, this is the first report linking transcriptional regulation of ckb to the PKC signaling pathway. We postulate that PMA-induced Ets and GATA transcription factors binding to the region between 21950 and 21966 of the ckb promoter represses its activity via the activation of PKCe in MCF-7 cells.
Supporting Information S1 Figure. Characterization of GATA binding to the ckb promoter by supershift assay. Supershift assay was performed by using 2 or 5 mg of GATA3 antibody in the EMSA. The electrophoresis was stopped when the bromophenol blue dye has reached 3/4 of the length of the gel. The blot shown is representative of two independent experiments that produced similar results. doi:10.1371/journal.pone.0113485.s001 (TIF)