12-O-Tetradecanoylphorbol-13-Acetate Induces Up-Regulated Transcription of Variant 1 but Not Variant 2 of VIL2 in Esophageal Squamous Cell Carcinoma Cells via ERK1/2/AP-1/Sp1 Signaling

The membrane-cytoskeleton link organizer ezrin may be the most “dramatic” tumor marker, being strongly over-expressed in nearly one-third of human malignancies. However, the molecular mechanisms of aberrant ezrin expression still need to be clarified. Ezrin, encoded by the VIL2 gene, has two transcript variants that differ in the transcriptional start site (TSS): V1 and V2. Both V1 and V2 encode the same protein. Here, we found that 12-O-tetradecanoylphorbol-13-acetate (TPA) induced over-expression of human VIL2 in esophageal squamous cell carcinoma (ESCC) cells. Furthermore, VIL2 V1 but not V2 was up-regulated after TPA stimulation in a time-dependent manner. AP-1 and Sp1 binding sites within the promoter region of VIL2 V1 acted not only as basal transcriptional elements but also as a composite TPA-responsive element (TRE) for the transcription of VIL2 V1. TPA stimulation enhanced c-Jun and Sp1 binding to the TRE via activation of the ERK1/2 pathway and increased protein levels of c-Jun, c-Fos, and Sp1, resulting in over-expression of VIL2 V1, whereas the MEK1/2 inhibitor U0126 blocked these events. Finally, we showed that TPA promoted the migration of ESCC cells whereas MEK1/2 inhibitor or ezrin silencing could partially inverse this alteration. Taken together, these results suggest that TPA is able to induce VIL2 V1 over-expression in ESCC cells by activating MEK/ERK1/2 signaling and increasing binding of Sp1 and c-Jun to the TRE of the VIL2 V1 promoter, and that VIL2 is an important TPA-induced effector.


Introduction
Ezrin is a member of the Ezrin-Radixin-Moesin (ERM) cytoskeleton-associated protein family, which primarily acts as physical and functional links between the plasma membrane and cytoskeleton [1,2]. Ezrin is essential for many cellular processes including the determination of cell polarity, formation of surface structure, cell adhesion, cell-cell interaction, cytokinesis, migration/invasion, and signal transduction [3,4]. The aberrant expression of ezrin is associated with the development and progression of several human cancers, such as hepatocellular carcinoma [5,6], lung cancer [7,8], breast carcinoma [9], pancreatic adenocarcinoma [10], colon cancer [11], osteosarcoma [12,13], rhabdomyosarcoma [14,15], and endometrial cancer [16]. We previously showed that the overexpression of ezrin in esophageal squamous cell carcinoma (ESCC) may be involved in the growth and invasiveness of ESCC cells and ezrin expression can serve as a biomarker that predicts the prognosis of ESCC patients [17,18]. However, although much is known about the functional role of ezrin in cancer development and progression, the biochemical mechanism of ezrin up-regulation has not been thoroughly investigated.
Being encoded by the VIL2 gene, ezrin has two transcript variants-variant 1(V1) and variant 2 (V2)-that differs in the transcriptional start site, but both V1 and V2 encode the same protein, termed as ezrin. Previous studies showed that a range of cytokines, including interleukin 2 (IL-2), IL-8, IL-10, and insulin-like growth factor 1, inhibited ezrin expression in human colon cancer cells, whereas epidermal growth factor and IL-11 enhanced ezrin expression [19]. TNFα enhanced both ezrin expression and phosphorylation in human endothelial cells, which promoted its nuclear translocation [20]. In mouse rhabdomyosarcomas, VIL2 has been suggested to be a downstream target of the homeoprotein transcription factor sineoculis homeobox homolog 1 (Six1), which binds to the VIL2 promoter and regulates its transcription [15,21]. Also, VIL2 is regulated by epigenetic modifications such as histone modification and DNA methylation in its promoter region, and up-regulation of VIL2 associated with histone 'active codes' (i.e., acetyl-H3-K9 and tri-methyl-H3-K4) and unmethylated CpG islands within its promoter region [14]. In ESCC cells, we previously found specificity protein 1 (Sp1) and activator protein 1 (AP-1, a c-Jun/c-Fos heterodimer) co-regulated VIL2 promoter activity and ezrin basic expression [22]. Furthermore, 12-O-tetradecanoylphorbol-13-acetate (TPA), a tumor promoter, could lead to the malignant transformation of human embryonic esophageal mucosa cells to ESCC cells, in which ezrin was overexpressed obviously, suggesting TPA might be an inducer of VIL2 overexpression in ESCC cells [23,24]. These findings of the observation of dramatic overexpression of ezrin in various cancer cells prompt us to explore the induced mechanisms of ezrin over-expression in ESCC.
Herein, we firstly investigated the effects of HGF, IL-6, PDGF, testosterone, TGF, TPA and VEGFC stimulation on VIL2 transcription in ESCC cells, and found that TPA could up-regulate the transcription of VIL2 V1, but not V2, through ERK1/2/AP-1/Sp1 signaling, resulting in the enhancement of cell mobility.

Cell culture and transfection
Cell lines used in this study and related cell culture general information were listed in the S1 Table. Mainly, EC109 cells were cultured in Dulbecco's modified Eagle's medium (GIBCO) supplemented with 10% new-born bovine serum (Excell Biology Inc). KYSE180 cells were maintained in 1640 medium (Thermo Scientific) containing 10% fetal bovine serum (Thermo Scientific). Both cell lines were incubated at 37°C in a humidified atmosphere containing 5% Table 1. Primers used in this study.

Primers
Sequences Names of constructs Primers for luciferase-reporter gene constructs Primers for qRT-PCR  (C) Alteration of VIL2 V1 or V2 mRNA level in response to inducers treatment. The total RNA extract (1 μg) from ESCC cells treated with or without inducers for 24 h, and subjected to qRT-PCR analysis using gene-specific primers. The relative value from the control group was considered equal to one arbitrary unit. Each point represents the mean ± SD. The data are representative of at least two independent experiments.

RT-PCR and qRT-PCR assay
Total RNA was extracted and purified using TRIzol (Invitrogen) according to the manufacturer's protocol. Subsequently, 1 μg total RNA was reverse transcribed into cDNA using Prime-Script RT reagent Kit with gDNA Eraser (TaKaRa) according to the manufacturer's protocol.
The target genes of VIL2 V1, VIL2 V2, and β-actin were amplified with the specific primers shown in Table 1. Quantitative RT-PCR(qRT-PCR) was performed in 10 μl total reaction mixture; denaturation occurred at 95°C for 30 s followed by 40 cycles with 5 s at 95°C and 30 s at 64°C. Relative levels of the VIL2 V1 or V2 mRNA were normalized to that of β-actin mRNA. The relative value from the DMSO-treated control group was considered equal to one arbitrary unit. All experiments were repeated at least three times with n = 3 samples per experiment.

Dual-luciferase reporter assay
Cells were seeded into 96-well plates at the density of 1.5×10 5 cells/ml and cultured for 12-24 h until grown to 50-80% confluence. They were then co-transfected with a firefly luciferaseexpressing plasmid (1μg), and a renilla luciferase-expressing plasmid (20ng) (pRL-TK, Promega) as an internal control, using Attractene Transfection Reagent (QIAGEN) according to the manufacturer's protocol. Twenty-four hours later, cells were treated with TPA for another 24 h and then harvested using passive lysis buffer (Promega). Luciferase activity was analyzed using the Dual-Luciferase Reporter Assay System (Promega) according to the manufacturer's protocol. Values for each group are expressed as the mean ± standard deviation (SD) for three separate experiments.
Electrophoretic mobility shift assay NE-PER Nuclear and Cytoplasmic Extraction Reagents, LightShift Chemiluminescent Electrophoretic mobility shift assay (EMSA) Kit, and Chemiluminescent Nucleic Acid Detection Module were purchased from Thermo Scientific. Purified biotin-labeled probes were purchased from Sangon Biotech. Cells were treated with DMSO or TPA, and then nuclear extracts were isolated. For binding assays, 3 μg nuclear extracts were incubated with 20 fmol probes at room temperature for 20 min in 20 μl reaction buffers. Electrophoreses, transfer to a nylon membrane, and UV cross-linking were performed according to the manufacturer's protocol.
For confirming the specificity of the transcription factor binding to the probe, the antibodies against c-Jun or Sp1 (Santa Cruz) and 3 μg nuclear extracts were incubated for 20 min at room temperature before adding the purified biotin-labeled probe for another 20 min at room temperature. Oligonucleotide probes used in EMSA were named as wild-type (WT), Am, Sm, or Sam, and their sequences corresponded to the human VIL2 V1 promoter region from -87 to -46 bp ( Table 2).

Chromatin Immunoprecipitation (ChIP)
Cells were grown to 60-70% confluence in 10-cm plates. After starved in serum-free medium for 12 h, cells were treated with or without TPA for another 6 h. ChIP-IT kit (Active Motif) was used in Chromatin Immunoprecipitation assay according to the manufacture's protocol.
Briefly, cells were cross-linked with 1% formaldehyde for 10 min at room temperature and then washed and harvested in ice-cold PBS containing protease inhibitor PMSF (500 μM). After sonication, sheared chromatin was pre-cleared with protein G beads and then 10 μl of the pre-cleared chromatin was transferred to a microcentrifuge tube, which was set as "Input DNA". Then sonicated chromatin preparation was incubated overnight with 3 μg of antibodies against IgG, Sp1 and c-Jun, respectively. Purified DNA was used for PCR analysis. The human VIL2 promoter-specific primers and negative control primers were used in our previous work [22], and all of them were shown in Table 1.

Western blotting analysis
Total cell lysates collected from ESCC cells were prepared in 1× Laemmli Sample Buffer (Bio-Rad). Equal amounts of protein were loaded and separated by SDS-PAGE and transferred to a PVDF membrane (Roche). Blots were blocked in blocking buffer and incubated with antibodies against Sp1, c-Jun, c-Fos, p-ERK, ERK, ezrin, T567 ezrin or β -actin. Specific immunoreactive bands were detected by luminol reagent.

Cell migration assays
Wound-healing assay and Boyden chamber cell migration assay were used to evaluate cell migration ability. For wound-healing assays, cells were inoculated onto the plates and grown to confluence for 24 h. And then the cells were starved in serum-free medium for 12 h, and a scratch made across the monolayer using a sterile pipette tip. Wound closure information was imaged at 0, 6, 12, 24 and 48 h with a 40×objective (Leica), respectively. In Boyden chamber cell migration assay, 4 × 10 4 cells were seeded onto the top chamber and the bottom chamber was filled with medium containing 10% fetal calf serum. The membranes were fixed and stained by Giemsa reagent 24 h later and migrated cells were quantified by counting 10 random fields under a light microscope (40×). The mean value was calculated from data obtained from three separate chambers.

Statistical analysis
Statistical analysis was performed using SPSS software. All data are reported as mean ±SD. Student's t-tests were used to compare groups. Statistical significance was set at p< 0.05.

Results
TPA induces the transcription of VIL2 V1 but not V2.
Because human VIL2 has two variants (V1 and V2, Fig 1A), we first examined their transcription levels in a panel of 27 human cell lines using qRT-PCR. Both V1 and V2 were expressed in all cell lines detected, the mRNA level of V1 was higher than V2 ( Fig 1B). To explore the different induced-expression of VIL2 variants, HGF, IL-6, PDGF, testosterone, TGF, VEGF and TPA were used. Results of qRT-PCR assay revealed that TPA showed the greatest induction of the transcription of VIL2 V1 among those inducible factors, whereas it did not affect the expression of VIL2 V2 (Fig 1C). To further confirm this finding, two cell lines were treated with or without TPA (10 ng/ml) for different time points. Results showed that TPA treatment increased the mRNA level of VIL2 V1 in a time-dependent manner, but not V2 (Fig 2A). Induced expression of TPA on ezrin protein level was also addressed by using western blotting ( Fig 2B). Moreover, to explore whether the increase of VIL2 V1 mRNA level and ezrin protein expression trigged by TPA treatment was linked to transcriptional regulation, we then treated cultured cells with TPA (10 ng/ml) or the combination of TPA (10 ng/ml) and actinomycin D (10 μg/ml, a transcription inhibitor) for 6 h and determined the level of VIL2 V1 by qRT-PCR. VIL2 V1 mRNA was up-regulated after treatment with TPA, but decreased after the combined treatment of actinomycin D and TPA (Fig 2C). These results strongly suggest that TPA-induced elevation of the VIL2 V1 mRNA was initiated by transcriptional activation.
The VIL2 V1 promoter responds to TPA stimulation.
To identify the region that is responsive to TPA, three reporter constructs containing different fragments of the VIL2 promoter region (i.e., the full-length cloned fragment (-1759/+134), V1 promoter fragment (-87/+134), or V2 promoter fragment (-1695/-1148)) were transiently transfected into EC109 cells. After the transfected cells were treated with different dosage of TPA for 24 h, luciferase activity was analyzed. The dose-response curve showed increased luciferase activity of the -87/+134 and -1759/+134 constructs starting at 2.5 ng/ml of TPA and reaching a maximum by 10 ng/ml TPA, but not -1695/-1148 construct (Fig 3A). Because the effect of TPA reached a plateau at 10-40 ng/ml, so 10 ng/ml of TPA was used for subsequent experiments. To further confirm the location of the TRE, we transfected a series of 5'-deleted reporter constructs into ESCC cells, and luciferase activity was analyzed after TPA treatment. The -1759/+134 construct containing the full-length VIL2 promoter region showed the highest luciferase activity, with decreasing activity associated with shorter promoter regions ( Fig 3B).
Notably, the -87/+134 sequence was still responsive to TPA treatment, suggesting that this fragment contains the TRE(s). These findings were confirmed in other four ESCC cell lines (Fig 3C).Taken together, our results suggest that TREs embedded in the VIL2 V1 promoter region modulate the V1expression in response to TPA stimulation.
The Sp1/AP-1 complex is necessary for the response of VIL2 V1 to TPA It is reported that Sp1 and AP-1 are involved in the alteration of target genes transcription in a TPA-dependent stimulation [25,26]. Our previous work also showed that Sp1/AP-1 binding sites located within the V1 promoter region could regulate the basal transcription of VIL2 [22]. Therefore, to detect whether the Sp1/AP-1 binding sites located in the VIL2 V1 promoter and subsequently use qRT-PCR assay to detect the change. The relative value from the DMSO-treated control group was considered equal to one arbitrary unit. Each point represents the mean ± SD, n!3. (B) EC109 and KYSE180 cells were stimulated with DMSO or TPA (10 ng/ml) for the indicated time after starvation for 12 h. Western blot was used to determine the protein level of ezrin upon TPA treatment. Signal intensity for the expression of ezrin was quantified by densitometric scanning and normalized by internal control (β-actin). The data represent the means ± SD of triplicate experiments. (C) Relative RT-PCR analysis. The total RNA extract was prepared from EC109 and KYSE180 cells after stimulation with DMSO or TPA (10 ng/ml) for 6 h, or pretreated with actinomycin D (1 μg/ml) for 1 h before TPA was added (10 ng/ml). The target genes were examined using the specific primers. The densitometry values for VIL2 V1 levels were normalized to the values for β-actin and then presented relative to that of the DMSO-treated control, which was set as 1. The results of a representative experiment were presented as mean ± SD of the two independent samples. *p< 0.05 or **p< 0.01.
doi:10.1371/journal.pone.0124680.g002 promoter region, pGLB-hE(-1759/+134) containing both VIL2 V1 and V 2 promoter region, pGLB-hE(-1695/-1148) containing VIL2 V2 promoter region or empty vector pGL3-basic (pGLB) were co-transfected with pRL-TK into EC109 cells simulated with TPA for 24 h before assay the luciferase activity. The firefly luciferase activity was normalized to Renilla luciferase activity and the relative value from the DMSO-treated control group was considered equal to one arbitrary unit. Each value represents the mean ± SD, n!3. (B) The schematic of the VIL2 promoter 5 0 -deletion constructs used for transient transfections (Left) and the relative luciferase activity of whole cell extracts from EC109 and KYSE180 cells transiently transfected with the constructs listed in the left for 24 h and then treated with DMSO or TPA for another 24 h (Right). The firefly luciferase activity was normalized to Renilla luciferase activity and the relative value from the DMSO-treated cells transfected with pGLB-hE (-1759/+134) was set to 100%. Each value represents the mean ± SD, n!3. (C) TPA-induced activity of VIL2 promoter region in different cells. pGLB-hE(-87/+134) and pGLB-hE(-1695/-1148) Constructs were used and the relative value from the DMSO-induced control group was considered equal to one arbitrary unit. Each value represents the mean ± SD, n!3. The data are representative of at least two independent experiments. *p< 0.05 or **p< 0.01.
doi:10.1371/journal.pone.0124680.g003 region were involved in the TPA-induction, nuclear extracts from EC109 cells treated with TPA were incubated with biotin-labeled oligonucleotides spanning from -87 to -46 bp of the V1 promoter (i.e., WT probe), which contained Sp1 (-75 to -69) and AP-1(-64 to -58) binding site sequences. We found that TPA treatment robustly increased DNA/protein complex formation (Fig 4A). To further investigate whether Sp1 and AP-1 transcription factors were involved in the TPA-induced up-regulation of VIL2 V1, nuclear extracts were prepared from EC109 cells after treatment with TPA for 24 h and incubated with four biotin-labeled oligonucleotide probes: (1) WT probe containing binding sites of both Sp1 and AP-1, (2) Sm probe containing a mutant Sp1 binding site, (3) Am probe containing a mutant AP-1 binding site, or (4) SAm probe containing double-mutated Sp1 and AP-1 binding sites. Although some DNA/protein complex formed in nuclear extracts incubated with WT probe (Fig 4B, lane 1), TPA treatment markedly increased complex formation (Fig 4B, lane 2). Mutation of the AP-1 binding site ( Fig  4B, lane 3) or double mutation of both AP-1 and Sp1 binding sites (Fig 4B, lane 5) abolished DNA/protein complex formation, whereas mutation of the Sp1 binding site only partially reduced complex formation (Fig 4B, lane 4). To validate the specific interaction of Sp1 and AP-1 with VIL2 promoter, cross-linked chromatin was immunoprecipitated with IgG, anti-Sp1 or anti-c-Jun by ChIP assay, respectively. As shown in Fig 4C, TPA stimulation enhanced the binding of Sp1 and c-Jun to the VIL2 V1 promoter region. These results suggest that Sp1/AP-1 complex is important for the response of VIL2 V1 to TPA.
To further confirm the cooperation of Sp1 and AP-1 transcription factors, we analyzed the TPA responsiveness of four reporter constructs: (1) pGLB-hE(-87/+134) containing the V1 promoter, (2) pGLB-hE(-87/+134)Am containing a mutant AP-1 binding site, (3) pGLB-hE (-87/+134)Sm containing a mutant Sp1 binding site, and (4) pGLB-hE(-87/+134)SAm containing both mutant Sp1 and AP-1 binding sites. These constructs were transiently transfected in TPA-or DMSO-treated ESCC cells. Compared with DMSO-treated control group, luciferase activity of the pGLB-hE(-87/+134) construct was significantly increased after TPA treatment ( Fig 4D). When either the AP-1 or Sp1 binding site was mutated, however, luciferase activity was reduced by~90%. When both sites were simultaneously mutated, luciferase activity was almost completely blocked under both DMSO and TPA treatment conditions. Thus, AP-1 and Sp1 binding sites may not only play a critical role in VIL2 basal promoter activity but also function as TREs within the VIL2 V1 promoter in ESCC cells.
TPA promoted cell motility by inducing up-regulation of VIL2 V1 via the ERK1/2/AP-1/Sp1 signaling To explore the involvement of AP-1and Sp1 in TPA-induced expression of VIL2 V1, EC109 and KYSE180 cells were treated with TPA or DMSO for different durations of time and the whole-cell extracts were analyzed by western blotting. The result showed that Sp1, c-Jun, and c-Fos protein expression levels were up-regulated in a time-dependent manner after TPA treatment (Fig 5A). The expression of Sp1 was peaked at 1 h in EC109 cells but 6 h in KYSE180 cells, c-Jun was peaked at 6h in both cells, and c-Fos was observed by 1 h in KYSE180 cells and 6 h in EC109 cells. To determine whether ERK1/2 pathway took part in this process, we further evaluated the expression levels of ERK1/2 and phosphorylation of ERK1/2 (p-ERK1/2) in TPA-treated cells, and found that p-ERK1/2 was enhanced after TPA treatment in both ESCC cell lines, while total ERK1/2 did not. These preliminary findings suggest that MEK/ERK1/2 signaling mediates the regulation of VIL2 V1 by TPA (Fig 5A).
As a further confirmation that ERK1/2 signaling participates in the TPA-mediated transcriptional regulation of VIL2 V1, the cells were pretreated with the specific MEK1/2 inhibitor U0126 for 1 h before TPA was added and the whole-cell extracts were analyzed by western The specific DNA-protein complex was confirmed by using biotin-labeled Sm, Am, or SAm oligonucleotides covering the same segment with probe WT but containing a substituted sequence (lanes 3 through 5), which are listed in Table 2. The probes were incubated with nuclear extract prepared from EC109 cells treated with DMSO or with TPA for 24 h. (C) Sonicated chromatin isolated from EC109 cells was incubated with antibodies against IgG, Sp1 and c-Jun, respectively. ChIP assay showed that TPA treatment enhanced Sp1 and c-Jun bound to the VIL2 promoter region. (D) The effect of site-directed mutagenesis of the AP-1/Sp1 binding sites on the VIL2 V1 promoter activity. The mutagenesis constructs were co-transfected with pRL-TK into EC109 or KYSE180 cells. After transfection for 24 h, the cells were treated with DMSO (control; open bars) or TPA (10 ng/ml; filled bars) for another 24 h before blotting. It showed that U0126 blocked the rapid phosphorylation of ERK1/2, leading to reduced expression of ezrin, and decreased phosphorylation of ezrin at the T567 site (termed here as T567 ezrin), c-Jun, and Sp1 in both cells, but not c-Fos in KYSE180 cells (Fig 5B). Moreover, pretreatment with U0126 decreased the TPA-induced expression of VIL2 V1 as evidenced by both mRNA expression and luciferase activity (S1 Fig). These results indicate that TPA induces the transcription of VIL2 V1 in ESCC cells through activating the ERK1/2/AP-1/ Sp1 signaling pathway.
Finally, to explore the functional role of TPA-induced ezrin in ESCC cells, we detected the motility of KYSE180 cells after TPA treatment by wound healing assay. Results showed that TPA treatment significantly promoted the motility of ESCC cells, and MEK/ERK1/2 inhibitor U0126 completely reversed this change (Fig 5C). The same results were obtained by using Boyden chamber-based cell migration assay in EC109 cells (S2A Fig). Further analysis revealed that silencing of ezrin by siRNA could also lead to restoration of the TPA-promoted cell migration ( Fig 5D; S2B Fig). These data suggested that MEK/ERK1/2/Ezrin might be important effector of TPA-mediated cellular function.

Discussion
VIL2 has been found to be not only a key component in tumor metastasis but also a diagnostic marker and therapeutic target for numerous cancers [10,12,14,15,27]. In ESCC, we previously found that overexpression of ezrin promoted cell proliferation and invasiveness [17,18]. Moreover, Sp1 and AP-1 activated the transcription of human VIL2 in ESCC cells through binding to the Sp1 site and adjacent AP-1 site of the VIL2 promoter, respectively, involving the MEK/ ERK1/2 signaling pathway [22]. Extracellular factors that induce transcription of VIL2 in ESCC cells, however, have not been clearly identified. Here, we found that TPA promoted the transcription of human VIL2 V1 but not V2 in a time-dependent manner in ESCC cells. Furthermore, we found that Sp1 and AP-1 binding sites located in the core promoter region of VIL2 V1 not only functioned as basal transcriptional elements but also as composite TREs, and the MEK/ERK1/2 signaling pathway enhances the binding of Sp1 and c-Jun to the TRE.
The regulation of gene transcription by combinations of different transcription factors has been widely demonstrated. In particular, Sp1 interacts with different transcription factors such as CCAAT-enhancer-binding proteins (C/EBP) [28], signal transducers and activators of transcription (STAT1) [29], and c-Jun [30], producing synergistic effects on the expression of target genes. Sp1 also cooperates with AP-1 to play a critical role in osteopontin promoter activation downstream of hepatitis C virus-mediated Ca 2+ signaling [31]. Cooperation between Sp1 and AP-1 has also been found to control loricrin expression [32] and to activate the urokinase-type plasminogen activator gene [33]. Moreover, multiple reports indicate that cooperation between functional Sp1 and AP-1 binding sites does not appear to depend on their spacing [30,34]. Based on these findings, we suspect that Sp1 may function as a scaffolding protein for recruiting AP-1 components after TPA stimulation to assure the efficient and orderly completion of VIL2 V1 transcription. The finding that Sp1 directly binds to guanine/cytosine (GC)-rich regions within damaged neuronal endopeptidase promoters, thus providing activating transcription factor 3 (ATF3)/c-Jun/STAT3 with a platform to engage their analysis. The firefly luciferase activities of mutant constructs were normalized to Renilla luciferase activity and shown as a ratio compared to that of the pGLB-hE (-87/+134) construct treated with DMSO, which was set as 1.Each value represents the mean ± SD, n!3. The data are representative of at least two independent experiments. *p < 0.05 or **p < 0.01, compared to the pGLB-hE (-87/+134) construct treated with DMSO; Δ p < 0.05 or ΔΔ p < 0.01, compared to the pGLB-hE (-87/+134) construct treated with TPA.
doi:10.1371/journal.pone.0124680.g004 functional synergy, is also consistent with our hypothesis, as the promoter region of human VIL2 V1 is GC-rich [30]. AP-1 is a heterodimeric protein containing proteins belonging to the Jun (c-Jun, JunD, and JunB) and Fos (c-Fos, FosB, Fra-1, and Fra-2) families [35,36]. In this study, except for c-Jun and c-Fos, we did not examine other AP-1 components after TPA treatment. A previous study demonstrated that the Jun/Fos heterodimer was more stable than the Jun/Jun homodimer due to its longer half-life and stronger DNA binding affinity [37].
Sp1 and AP-1 are recruited and bind to their respective binding sites within target promoter regions through signal-transducing kinase cascades, leading to the transcription of corresponding genes. Velpula et al. [38] and Zhu et al. [39] found that Sp1 was activated by MEK/ERK for different genes in various cell types. Tissue factor-factor/VIIa/protease-activated receptor 2-induced PKCα and ERK1/2 signaling could phosphorylate c-Jun, which subsequently enhanced c-Jun expression [40]. In human articular cartilage, c-Fos was activated by ERK1/2 signaling upon transactivation of the IL-11 gene [41]. Also, previous studies showed that activated ERK1/2 phosphorylated Sp1 and AP-1 and enhanced their binding to the VIL2 promoter, resulting in activation of VIL2 basal transcription [22]. In the present study, TPA stimulation was also found to activate the MEK/ERK1/2 pathway, leading to the up-regulation of Sp1 and AP-1expression. This suggests that a potential mechanism of TPA-induced transcription of VIL2 V1 is the activation of the MEK/ERK1/2 pathway, which leads to enhanced expression of Sp1, c-Fos, and c-Jun and promotion of their binding to TRE within the promoter region of VIL2 V1, ultimately resulting in the up-regulation of VIL2 V1 in ESCC cells. However, why only VIL2 V1 but not V2 is responsive to TPA treatment is unclear.
The key point of inducing morbidity and mortality of tumors is their fast migration. TPA has been reported to promote cancer cell motility through regulating the expression and function of S100A14, cyclooxygenase-2 (COX-2) or matrix metalloproteinase-9 (MMP-9) [42,43].
Here, we showed that ERK1/2/AP-1/Sp1/ezrin signaling was involved in TPA-mediated motility of ESCC cells, suggesting that VIL2 might be an important downstream effector of TPA. Another point should be highlighted was that our results showed TPA not only up-regulated the expression of ezrin, but also increased the phosphorylation of ezrin at T567 site. Phosphorylation of ezrin at T567 was found to be necessary for the activation of ezrin, and was important for the migration process [44,45]. TPA could induce the activation of certain kinases such as PKCα, which might influence the process of protein phosphorylation [46,47]. Therefore, we proposed that phosphorylation of ezrin might participate in the TPA-mediated alterations of cell behaviors. Studies for the precise mechanisms are underway.
In summary, we show for the first time that TPA stimulates transcriptional expression of human VIL2 V1 but not V2 via the ERK1/2-dependent AP-1/Sp1 pathway in ESCC cells, leading to the enhancement of cell motility. Considering the importance of ezrin in the invasion and metastasis of ESCC cells, a full understanding of the mechanisms involved in TPA regulation of ezrin is urgently needed for future clinical applications. (A) qRT-PCR assay for the message RNA level of VIL2 V1. The cells were stimulated with TPA as the prior experiment. The relative expression was normalized to the DMSO-treated control. Each value represents the mean ± SD, n!3. (B) Reporter gene assay of VIL2 V1 promoter activity. The pGLB-hE(-87/-134) construct was co-transfected with pRL-TK into ESCC cells for 24 h, then the cells were stimulated for 24 h by DMSO, TPA(10 ng/ml), or pretreated for 1 h with U0126 (10 μM) before added TPA. The reporter gene activity was measured. The firefly luciferase activity was normalized to Renilla luciferase activity. Each value represents the mean ± SD, n!3. The data are representative of at least two independent experiments. Ã p< 0.05 or ÃÃ p< 0.01.

Supporting Information
(TIF) S2 Fig. TPA increased cell mobility through up-regulation of ezrin via MEK/ERK1/2 pathway in EC109 cells. (A) 24-well Boyden chamber-based cell migration assay was used to determine the alterations of cell migration after being treated with TPA (10 ng/ml) or TPA and U0126. (B) 24-well Boyden chamber-based cell migration assay was employed to detect the effect of ezrin knockdown on the TPA-mediated cell migration. Left, western blotting analysis for the ezrin silencing in the TPA-treated cells; Right, cell migration assay. Representative tumor cells migrated were photographed (40×), data represent mean ± SD of triplicates. (TIF) S1 Table. Cell lines used in this study and cell culture general information. (PDF)