Sp1 and Sp3 Are the Transcription Activators of Human ek1 Promoter in TSA-Treated Human Colon Carcinoma Cells

Background Ethanolamine kinase (EK) catalyzes the phosphorylation of ethanolamine, the first step in the CDP-ethanolamine pathway for the biosynthesis of phosphatidylethanolamine (PE). Human EK exists as EK1, EK2α and EK2β isoforms, encoded by two separate genes, named ek1 and ek2. EK activity is stimulated by carcinogens and oncogenes, suggesting the involvement of EK in carcinogenesis. Currently, little is known about EK transcriptional regulation by endogenous or exogenous signals, and the ek gene promoter has never been studied. Methodology/Principal Findings In this report, we mapped the important regulatory regions in the human ek1 promoter. 5’ deletion analysis and site-directed mutagenesis identified a Sp site at position (-40/-31) that was essential for the basal transcription of this gene. Treatment of HCT116 cells with trichostatin A (TSA), a histone deacetylase inhibitor, significantly upregulated the ek1 promoter activity through the Sp(-40/-31) site and increased the endogenous expression of ek1. Chromatin immunoprecipitation assay revealed that TSA increased the binding of Sp1, Sp3 and RNA polymerase II to the ek1 promoter in HCT116 cells. The effect of TSA on ek1 promoter activity was cell-line specific as TSA treatment did not affect ek1 promoter activity in HepG2 cells. Conclusion/Significance In conclusion, we showed that Sp1 and Sp3 are not only essential for the basal transcription of the ek1 gene, their accessibility to the target site on the ek1 promoter is regulated by histone protein modification in a cell line dependent manner.


Introduction
Phosphatidylethanolamine (PE) is the second most abundant phospholipid in eukaryotic cells [1]. PE can be synthesized by CDP-ethanolamine pathway (Kennedy pathway) and phosphatidylserine decarboxylase (PSD) pathway but the former is known to contribute to the majority of the PE in mammalian cells [2]. Ethanolamine kinase (EK) is the first enzyme in the CDPethanolamine pathway, which catalyzes the phosphorylation of ethanolamine by using ATP to yield phosphoethanolamine and ADP. In human, EK exists as three isoforms, encoded by two separate genes, named ek1 [NCBI Gene ID: 55500] and ek2 [NCBI Gene ID: 55224]. While ek1 codes for a single protein, EK1 (452 amino acids), ek2 undergoes alternative splicing to produce two other EK isoforms, EK2α (386 amino acids) and EK2β (394 amino acids) [3].
High EK activity has been previously shown to correlate with cell growth [4][5][6][7][8]. Induction of EK activity by oncogenes and carcinogens suggested the involvement of EK in carcinogenesis by promoting cell growth and/or survival [4][5][6]. Malewicz et al. [8] showed that overexpression of EK in NIH 3T3 fibroblasts protects fibroblast cells from apoptotic cell death. Increased phosphorylation of ethanolamine and its analogs (methylethanolamine and dimethylethanolamine) in EK overexpressor cells increases cell survival in serum-free medium but decreases the insulin-dependent DNA synthesis [8].
Our preliminary experiments showed that the ek1 promoter activity was significantly induced under serum starvation condition. Serum starvation induced genes transcription through the modulation of chromatin structure by core histones modifications [9][10][11][12]. Acetylation of lysine residues on histone tail "relaxed" the packed chromatin and provides accessibility to transcription machinery that leads to transcriptional activation. Conversely, histone deacetylases (HDACs) catalyze the deacetylation of lysine residues on core histone, allowing compacted chromatin structure to form and repress transcription [13]. Inhibition of HDACs by trichostatin A (TSA) induced the expression of CTP: phosphocholine cytidylyltransferase α (Ctpctα), the second enzyme in the CDP-choline pathway, through the Sp1 binding site of its promoter [14]. The expression of choline kinase α (ckα), the first enzyme in the CDP-choline pathway, was also stimulated by treatment of MCF-7 cells with HDAC inhibitor, suberoylanilide hydroxamic acid (SAHA) [15].
Although EK is the key enzyme in cellular phospholipid synthesis and it has been implicated in cell growth and carcinogenesis, the regulation of human ek gene at the transcriptional level has never been described. This study aims to characterize human ek1 promoter by identifying the cis-acting element that regulates ek1 basal expression. Since ek1 promoter is rich in GC sequences with several Sp-family transcription factor binding sites, the role of Sp proteins in the ek1 transcription regulation was investigated. The molecular mechanism that underlies the TSA-mediated induction of ek1 promoter was also studied.

Chromatin immunoprecipitation (ChIP)
Chromatin immunoprecipitation (ChIP) assay was performed using Pierce Agarose ChIP Kit (Thermo Fisher Scientific, USA) according to the manufacturer's protocol. HCT116 and HepG2 cells were cultured with or without TSA (1 μM) for 24 hours. The DNA-protein complexes were immunoprecipitated with 8 μg of Sp1 antibody (Millipore, USA), 8 μg of Sp3 antibody (Millipore, USA) and 8 μg of RNA polymerase II (RNA pol II) antibody (Thermo Fisher Scientific, USA). DNA recovered from samples was PCR amplified using ek1-154-5' and ek1-3' primers followed by agarose gel electrophoresis. The intensities of the bands of interest were quantitated with Image J 1.42 program [23]. All the PCR products were gel-purified and sequenced to confirm that the PCR products were ek1 promoter.

Statistical analysis
Statistical analysis was performed using the Student's t-test or one-way ANOVA, followed by Tukey Honestly Significant Differences (HSD) post-hoc test. The analysis was performed using the PASW Statistics 18, Release Version 18.0.0 (SPSS, Inc., USA). Data were presented as mean ± SEM of at least three independent experiments.

Results
Human ek1 promoter is a TATA-less, CpG island containing promoter The sequence of 1967 bp upstream region of the human ek1 gene (between nucleotides -1966 and +1; +1 refers to the A of the ATG translation start site) was retrieved by combining the database information from NCBI (http://www.ncbi.nlm.nih.gov/) and BLAT server [24]. Analysis of the putative promoter sequence using MatInspector 8.0 and TFSEARCH revealed several potential transcription factor binding sites for Sp-family proteins, Ets, GATA, CREB, MZF1, AML-1a, STATx, Lyf-1, Nkx-2, SRY, Oct-1, S8, and CdxA (Fig 1). The most striking feature of the ek1 promoter is the lack of a typical CAAT or TATA box, which is a common characteristic of GC-rich promoters. Analysis using CpGIS was thus performed to identify potential CpG island in the ek1 promoter. The program predicted one CpG island at the region between -529 and -1 with a GC content of 56.3% and an observed/expected ratio of 0.694.
Activities of ek1 promoter in HepG2, MCF-7 and HCT116 cells The functional activity of the pGL4.10-ek1(-1966/+1) reporter construct was assessed in HepG2, HCT116 and MCF-7 cells. The cloned ek1 promoter markedly increased the luc2 reporter gene expression in all three cell lines as compared to the promoterless control vector (Fig 2). This result showed that the -1966/+1 ek1 promoter was active in all the cell lines tested with the highest promoter activity detected in HCT116 cells (Fig 2). Hence, HCT116 cells were used as the model cell line for all subsequent ek1 promoter analysis in this study.
Identification of important regulatory regions in the ek1 gene promoter recovered the activity to about 70% of the full length promoter, indicating the presence of at least one inhibitory element in this region (between -1183 and -665). Truncation of the region between -665 to -229 did not cause any significant change of promoter activity. The region downstream of the -229 position was further truncated to map the minimal promoter region for initiating ek1 gene transcription. As shown in Fig 3B, deletion of the ek1 promoter from -229 to -154 increased its activity by more than seven times compared to the -229 promoter construct, indicating the presence of at least one repressive element in this region. Subsequent truncations (-100 and -69 promoter constructs) did not show any dramatic change of promoter activity. Thus, the reporter construct containing the -69/+1 ek1 promoter fragment was taken as the minimal promoter region for conferring basal transcription activity of ek1gene. Bioinformatics prediction of transcription factor binding sites in the -69/+1 ek1 basal promoter region revealed two putative Sp and one CREB binding sites at positions -69/-60, -40/-31, and -27/-19, respectively. The importance of the two putative Sp binding sites in regulating the ek1 transcription was examined by single or double mutations of the Sp sites in the pGL4.10-ek1(-69/+1) reporter construct. As shown in Fig 3C, single mutation of the Sp(-40/-31) site significantly decreased the promoter activity to approximately 6% of the wild type promoter construct. However, mutation of the Sp(-69/-60) site had no apparent effect on the promoter activity. Double mutation of these two Sp sites significantly decreased the promoter activity to approximately 5% of the wild type promoter construct. The results showed that the Sp(-40/-31) binding site was a functional activator element and it was essential in modulating the basal ek1 promoter activity.

TSA activates basal ek1 promoter activity via Sp(-40/-31) site
In our preliminary experiments (data not shown), we observed that serum starvation (0.1% FBS) activated the ek1 promoter and the Sp(-40/-31) site was required for this effect to take place. Since serum starvation could affect genes expression through the modulation of chromatin modification [9,10], we postulate that the accessibility of transcription factors to the Sp (-40/-31) site responsible for activating ek1 promoter is modulated by chromatin modification.  Histone acetylation/deacetylation is one of the mechanisms to "unravel" chromatin by neutralization of the basic charge of the lysine [25]. To investigate the role of histone acetylation/deacetylation in regulating ek1 transcription, we treated the pGL4.10-ek1(-69/+1) reporter plasmid transfected-HCT116 cells with a well known HDAC inhibitor, trichostatin A (TSA). TSA significantly activated ek1 promoter in a concentration-dependent manner with the upregulation of the promoter activity started at 0.25 μM and the maximum effect was observed at 1 μM of TSA, which produced 2.3 folds higher promoter activity than the DMSO control ( Fig  4A). The effect of TSA treatment duration shown in Fig 4B showed that the ek1 promoter activity did not significantly increase until 24 hours after TSA treatment with an increase of about 1.6 folds compared to control.
To To investigate the effect of HDAC inhibition by TSA on the endogenous expression of the ek1 gene, the ek1 mRNA level in the cells treated or untreated with TSA was determined by real-time PCR. All the RNA samples used for cDNA synthesis were intact and pure as indicated by the prominent 28S and 18S ribosomal RNA bands and the ratios of OD 260/280 from 1.8 to 2.0. The PCR efficiency of ek1 primers was 89%. Based on the data from triplicate samples from three independent experiments, exposure of HCT116 cells to 1 μM of TSA for 24 hours resulted in moderate but significant increase of ek1 gene expression by about 0.5 fold compared with the DMSO treated control cells (Fig 4D).

TSA induced the binding of Sp proteins and RNA polymerase II on ek1 promoter
Subsequently, the effect of TSA treatment on the binding of Sp proteins to the Sp(-40/-31) site of ek1 promoter was also investigated using ChIP assay. In this study, ChIP assay was performed by using Sp1 and Sp3 specific antibodies as these two transcription factors are expressed in the same cells and are indistinguishable in their DNA-binding specificity [26]. The ek1-154-5' and ek1-3' primers used for PCR of the immunoprecipitated DNA are flanking the Sp(-40/-31) response element. The result showed in Fig 5A shows that TSA treatment for 24 hours increased the binding of both Sp1 and Sp3 on the GC-rich ek1 minimal promoter as compared to control.
The inhibition of HDACs activities by TSA treatment has also been reported to enhance the recruitment of RNA polymerase II to certain promoters [27][28][29]. Therefore, the effect of TSA treatment on the binding of RNA polymerase II to the ek1 promoter was also examined by using ChIP assay. As shown in Fig 5B, 24 hours after TSA treatment, the binding of RNA polymerase II was increased significantly compared to control. Based on our results, it can be concluded that the induction of ek1 promoter activity and endogenous expression by TSA treatment is resulted from the changes in the chromatin structure around the Sp(-40/-31) site that consequently increased the local recruitment of Sp1/Sp3 transcription factors and RNA polymerase II to the ek1 minimal promoter region.  Effects TSA treatment on the binding of Sp1, Sp3 and RNA polymerase II to the ek1 minimal promoter region. ChIP analysis was performed to confirm the interaction of (A) Sp proteins and (B) RNA polymerase II with the promoter under 1 μM TSA treatment for 24 hours. PCR amplification products were resolved on 2% (w/v) agarose gel and visualized by EtBr staining. Band intensities were quantitated with Image J 1.42 and the relative intensities (compared to negative control) of PCR products from Sp1 and Sp3 immunoprecipitates were plotted. Each bar represents standard error of means (SEM) from two independent experiments. M: GeneRuler™ DNA Ladder Mix; T: total input sample (unprocessed chromatin); P: positive control (amplified using GAPDH primers) and N: pre-immune normal rabbit IgG (negative control). skeletal muscle and prostate. This suggests that ek1 promoter might be regulated differently according to cell type. To investigate whether the TSA activation of ek1 promoter was cell-line specific, the effect of TSA treatment on ek1 promoter activity was studied in liver cancer cell line, HepG2. As shown in Fig 6, TSA did not significantly induce the ek1 promoter activity as in the case of HCT116 cells.
In order to answer the question of whether the binding of Sp transcription factors was also cell-line specific, ChIP analysis for the binding of Sp1 and Sp3 to the ek1 minimal promoter region in HCT116 and HepG2 cells without any treatment was carried out. Fig 7 shows that PCR products were amplified from Sp1 and Sp3 antibodies precipitated DNA from HepG2 cells (lower panel) but not from HCT116 cells (upper panel). The results showed that in HepG2 cells, Sp1 and Sp3 transcription factors were already bound to the basal ek1 promoter without TSA treatment and this explained why the promoter activity was not affected by TSA as in the case of HCT116 cells.

Discussion
In this work, we aimed to characterize the ek1 promoter and investigate the possibility of ek1 gene transcription regulation by epigenetic mechanism. Based on our sequence analysis, the ek1 gene promoter is a TATA-less, GC-rich promoter which probably belongs to the housekeeping gene family. This assumption is supported by the fact that approximately 76% of human TATA-less core promoters contain multiple Sp1 binding elements in the GC rich sequences [30] and the ubiquitous Sp1 serves as a constitutive activator of housekeeping genes by recruiting TATA-binding protein (TBP) to the promoters without recognizable consensus TATA box [31,32].
The activity of putative ek1 promoter (-1966/+1) was initially assessed in HepG2, HCT116 and MCF-7 cells. Under the same transfection conditions, the luciferase expression level driven by the ek1 promoter varied with the cell line used. The transcription level of luc2 reporter gene in the transfected promoter-luciferase reporter construct generally depends on the level of required transcription factors in the host cell. The levels of certain transcription factors are cell line dependent such as the expression level of Ets-1 in MCF-7 cell is known to be 10% lower than the other cell lines [33]. Therefore, the variation of the ek1 promoter activities in the three cell lines tested might be due to different expression levels of the required transcription factors.
Deletion analysis revealed that the region spanning from -69 to +1 is critical for the ek1 gene promoter basal activity. This region is CpG rich, with two putative binding sites for Sp proteins. However, only one of the Sp binding sites located at -40/-31 is critical as the positive regulatory element for ek1 gene transcription. Sp1 and Sp3 are two closely related members of Sp-family transcription factors based on their similarities in structure and binding specificity to GC or GT motif [34,35]. The binding of ubiquitous Sp1 to TATA-less promoters is critical for gene transcription initiation as it has been demonstrated that the transcription of TATAless promoters requires multisubunit TFIID complex that comprises Sp1 and other coactivators like TBP and TBP-associated factors (TAFs) [36][37][38][39]. Sp1 is suggested to be involved in assembling the basal transcription machinery by interaction with the TBP/TFIID [38]. Sp1 plays an important role to tether the preinitiation complex (PIC) to TATA-less promoters by stabilizing the recruitment of TFIIB and TFIIE into PIC [32]. Since ek1 promoter is a TATAless promoter, the binding of Sp1 to Sp(-40/-31) site in the promoter is pivotal to the transcription initiation of ek1 gene. Mutation introduced at this site completely abolished the ek1 promoter activity possibly due to the disruption of PIC formation on ek1 promoter.
Histone acetylation/deacetylation is mainly regulated by histone acetyltransferases (HATs) and histone deacetylases (HDACs) [13]. HATs acetylate the lysine residues on N-terminal tail of the histones and thus neutralizing their positive charges. This will diminish the electrostatic interactions with the negatively charged DNA and open accessibility to general transcriptional machinery, which subsequently leads to transcriptional activation. Conversely, HDACs remove the acetyl group on lysine residues, allowing more condense chromatin structure to form which causes transcriptional silencing [13]. Therefore, the balance between the activities of HDACs and HATs is important in the regulation of gene transcription. In this study, the role of histone acetylation/deacetylation in the regulation of ek1 gene transcription was studied using HDAC inhibitor TSA. It was found that treatment of HCT116 cells with TSA caused a potent stimulation of ek1 promoter activity. In addition, we showed that the Sp(-40/-31) site located in the GpG-enriched region of ek1 promoter was not only essential for ek1 basal promoter activity but also was important for TSA response. Previous studies have pointed to Sp1 and Sp3 transcription factors as the mediators of HDAC inhibition in the regulation of gene transcription [27,40,41]. Interestingly, the increase in ek1 mRNA level induced by TSA is markedly lower than the TSA-mediated upregulation of ek1 promoter activity. One possible explanation for the discrepancy between ek1 mRNA level and its promoter activity is the involvement of other regulatory region in ek1 promoter or post-transcriptional regulation of ek1 gene expression.
Sp-family of transcription factors consists of four proteins designated Sp1, Sp2, Sp3 and Sp4 which are involved in the expression of a large number of housekeeping genes [26]. Sp1 and Sp3 transcription factors are highly similar in protein structure. They contain two major glutamine-rich activation domains at the N terminus and a conserved Cys2His2 zinc finger domain at the C terminus for DNA binding [34]. Due to these structural similarities, Sp1 and Sp3 share similar affinities for GC (GGGGCGGGC) or GT (GGGTGTGGC) motifs [34,35]. Sp3 can also be directly acetylated by HAT and TSA-induced acetylation of Sp3 switches its function from the transcriptional repressor to the transcriptional activator of transforming growth factor-β receptor gene [42]. Our ChIP analysis showed that TSA induced the binding of Sp1 and Sp3 to the ek1 promoter. We suggest that the activation of ek1 promoter by HDAC inhibitor requires the binding of Sp1 or Sp3 to the ek1 minimal promoter region. Previous study has shown that TSA could induce the recruitment of RNA polymerase II to the transcription factor complex for the initiation of transcription process [27]. Here, the increased binding of RNA polymerase II after TSA treatment could be due to the changes of chromatin structure or the increased binding of Sp1/Sp3 that recruits RNA polymerase II to the TSS for the activation of ek1 transcription. Our results indicate that the activity of ek1 minimal promoter region harboring the Sp(40/-31) site is significantly affected by histone modifications. HDACs play an important role in carcinogenesis by regulating the expression of genes involved in both cancer initiation and progression [13]. The findings from this study provide important link between chromatin modification by HDACs and ek1 gene expression that affects cell survival.
It is worth noting that TSA up-regulation of ek1 promoter was cell-line specific. We showed that, unlike in HCT116, the binding of Sp1 and Sp3 to ek1 basal promoter region had already occurred in HepG2 cells before TSA treatment. The cell-type specificity of TSA effect could be due to differential regulation of ek1 gene in different cell types as evidenced by variable level of ek1 mRNA transcript across multiple tissues [3]. Differential regulation of ek1 promoter was also supported by the results in this study that showed ek1 promoter (-1966/+1) activity was clearly different among the three cell lines used.
In conclusion, this study showed that Sp(-40/-31) site is required for the transcriptional activation of the ek1 basal promoter. The transcriptional regulation of ek1 is associated with chromatin remodeling and Sp1/Sp3/RNA polymerase II accessibility to promoter region around the Sp(-40/-31) site. Sp1 and Sp3 not only function in concert with the basal transcriptional machinery to initiate ek1 gene transcription, their accessibility to the target site on ek1 promoter also depends on histone proteins modification. This study delineates the essential cisacting regulatory element of ek1 promoter and its possible modulation by chromatin structure modification that is useful for the fundamental understanding of ek1 gene transcriptional regulation.