MZF-1/Elk-1 Complex Binds to Protein Kinase Cα Promoter and Is Involved in Hepatocellular Carcinoma

In this study, the molecular mechanism of protein kinase C alpha (PKCα) gene regulation in hepatocellular carcinoma (HCC) involving Ets-like protein-1 (Elk-1) and myeloid zinc finger-1 (MZF-1) was investigated. The luciferase reporter assay results revealed that the presence of both MZF-1 and Elk-1 significantly contributed to the upregulation of PKCα gene transcription activity, and the transcriptional activity decreased when the transfection included a DNA-binding-deficient (∆DBD) gene vector of either MZF-1 or Elk-1 DNA-binding deficiency (MZF-1∆DBD or Elk-1∆DBD), thereby indicating that the enhanced expression of PKCα was caused by the binding of MZF-1 and/or Elk-1 with the PKCα promoter. We investigated MZF-1 and Elk-1 to determine whether they bind to each other. The results of immunoprecipitation (IP), Co-IP, chromatin IP (ChIP), and Re-ChIP analyses indicated that Elk-1 can directly bind to the N-terminal region of MZF-1 and MZF-1 can directly bind to the C-terminal region of Elk-1 to form a complex before attaching to the PKCα promoter. Furthermore, when MZF-1∆DBD or Elk-1∆DBD was added to the cells, PKCα expression decreased, and cell proliferation, migration, invasion, and tumorigenicity also decreased. These findings suggest that PKCα expression in HCC could be stimulated by the formation of MZF-1/Elk-1 complex, which directly binds to the PKCα promoter.


Introduction
Protein kinase C alpha (PKCα), one of the proteins in the kinase C family, is generally known to serve an important function in tumorigenesis, invasion, and metastasis [1][2][3]. It has been considered an interesting and challenging target for the development of new therapeutic agents. Although clinical trials have yet to produce a satisfactory result [4], targeting PKCα remains a viable treatment method because of its significant function in tumor cells.
Our recent studies found that two transcription genes, namely, Ets-like protein-1 (Elk-1) and myeloid zinc finger-1 (MZF-1), can regulate PKCα expression [5][6][7]. Thus, these genes are viable targets for new PKCα inhibition strategies. The Elk-1 protein is capable of activating the c-fos promoter with serum response factors (SRFs) and is a target of both extracellular signalregulated kinase (ERK) and JNK MAP kinase cascades [8]. Elk-1 controls the expression of genes involved in cell cycle progression, differentiation, and apoptosis [9][10][11] and also serves a central function in cell response to extra-cellular signals. MZF-1, the other protein in this equation, is a transcription factor of the Kruppel family of zinc finger proteins. MZF-1 is preferentially expressed in myeloid progenitor cells [12] and is also involved in the growth, differentiation, and apoptosis of myeloid progenitors [13][14][15]. Both MZF-1 and Elk-1 gene expressions have been found to be associated with PKCα in liver cancer cells [5][6][7], thereby suggesting that these transcription factors in PKCα cooperate in influencing carcinogenesis. However, the unfamiliar molecular mechanism remains to be investigated. Thus, this study aims to delineate in greater detail the molecular mechanism underlying PKCα gene regulation by Elk-1 and MZF-1.
This study uses the human hepatocellular carcinoma (HCC) model to demonstrate that MZF-1 and Elk-1 directly bind to the PKCα promoter and modulate PKCα performance through the interactive cooperation between the two transcription factors. The mechanism of a possible formation of a complex (MZF-1 and Elk-1), which binds to the PKCα promoter and enhances PKCα expression and cell growth and invasion, is discussed.

Transient transfection
Transfections were performed using Lipofectamine 2000. Cells seeded in a 60 mm dish were cultured in DMEM supplemented with 10% FCS at 37°C for 24 h. After incubation, the cells were rinsed with serum-free MEM before adding 1 mL MEM containing 10 μg/mL Lipofectamine 2000 Transfection Reagent (Invitrogen, Carlsbad, CA) and 0.5 μg to 5 μg of the indicated plasmid. The cells were then incubated at 37 C for 6 h before adding 1 mL MEM supplemented with 20% FCS to the medium. After incubation at 37°C for 18 h, the medium was replaced with fresh 10% FCS-DMEM, and the cells were incubated at 37°C for 24 h. The cells were then lysed for Western blot analysis and RT-PCR.

Chromatin Immunoprecipitation (ChIP) Assay
The ChIP assay was performed following the methodology described by Hsieh et al. [5] The SK-Hep-1 cells were treated with or without indicated sense or antisense ODN (5 μM), and cells were harvested and cross-linked with 1% formaldehyde for 10 min. The reaction was then terminated by the addition of 0.25 M Glycine. The procedures thereafter were performed according to Hsieh et al.

Re-ChIP of chromatin-bound proteins
The methodology described by Reid et al. [16] was followed. Precipitated complexes were eluted from primary immunoprecipitates, pooled from three or four reactions, and incubated with 30 μL of ChIP elution buffer (50 mM NaHCO 3 , 1% SDS). The samples were shaken for 30 min at room temperature, then spun down and collected as supernatants. The complexes were eluted twice, and both elutions were combined. The pooled elutions were diluted 1:10 in buffer (1% Triton X-100, 5 mM EDTA, 150 mM NaCl, 25mM Tris, pH 8) containing a protease inhibitor mixture (Sigma). Further Re-ChIP of supernatants and analysis of the results were performed as described above for primary ChIP immunoprecipitations.

RNA isolation and RT-PCR analysis
Total RNA was isolated from cell specimens using the guanidinium thiocyanate-phenol method. The extract integrity was assessed by 1.5% agarose gel electrophoresis, and RNA was visualized by ethidium bromide staining. The total amount of RNA was determined spectrophotometrically. RT-PCR assay was performed according to De Petro et al. [17] with slight modifications. An aliquot of total RNA (1 μg) was use for RT. The RT product (2 μL) was diluted with the PCR buffer (50 mM KCl, 10 mM Tris-HCl, and 2 mM MgCl 2 ) to a final volume of 50 μL, containing 0.5 μM dNTPs (final concentration, 0.8 mM) and 0.5 unit of Super-Therm Taq DNA polymerase (Southern Cross Biotechnology, Cape Town, South Africa). PCR was performed on a GeneAmp PCR system 2400 (Applied Biosystems, Foster City, CA). The ODN primers used in RT-PCR were as described previously [18]. The PCR products were analyzed by 1.5% agarose gel electrophoresis and direct visualization after SYBR Green I (Cambrex Bio Science Rockland, Inc., Rockland, ME) staining. The agarose gel was scanned and analyzed using the Kodak Scientific 1D Imaging System (Eastman Kodak Company, New Haven, CT).

Cell proliferation/migration/invasion assay
The cell proliferation, migration, and invasion assays were performed according to Hsieh et al. [19].

Tumorigenicity Assay in Nude Mice
Female BALB/c nude mice, four weeks to six weeks of age, were purchased from National Health Research Institute (Taipei, Taiwan, ROC) and housed in a dedicated nude mouse facility with micro-isolator caging. All procedures involving laboratory animal use were in accordance with the guidelines of the Instituted Animal Care and Use Committee of China Medical University (IACUC, CMU) for the care and use of laboratory animals and all experimental procedures were approved by IACUC-CMU.
The SK-Hep-1 cells were first treated with 5 μg FLAG or MZF-1ΔDBD vector, detached by trypsinisation 48 h later, and then washed in triplicate in serum-free DMEM. 1 × 10 7 cells in 100 μL volume was injected subcutaneously into the right posterior flank of mice by use of a 1 mL syringe with 24-gauge needle 6 . The mice were randomly assigned to a treatment group: FLAG (5 μg, 0.1 mL/mouse) and MZF-1ΔDBD (5 μg, 0.1 mL/mouse). Each group had five mice, and the experiment was repeated twice. The tumor volume was calculated by the formula: 0.5236 × L1(L2) 2 , where L1 is long diameter, and L2 is short diameter. The inhibitory rate of tumor growth was calculated by the formula: (tumor volume FLAG -tumor volume MZF-1ΔDBD / tumor volume FLAG ) × 100%.

Statistical analysis
Data were expressed as mean ± SEM and analyzed by analysis of variance. Student's t-test was used in two-group comparisons. P<0.05 was considered statistically significant.

MZF-1 and Elk-1 directly bind to the PKCα promoter
A previous study has identified that poorly differentiated cell lines contain higher expressions of MZF-1, Elk-1, and PKCα genes. To check if the enhanced level of both Elk-1 and MZF-1 were synergistically responsible for the increase in transcriptional activity of PKCα promoter, a luciferase assay was performed in the same cell-lines. S1A Fig shows that the transcriptional activities in the poorly differentiated HA22T/VGH and SK-Hep-1 cells were higher than those in the well-differentiated HepG2, Hep3B, and Huh-7 cells. The activities were consistent with the PKCα, Elk-1, and MZF-1 results in Western blot, in which the expressions in HA22T/VGH and SK-Hep-1 cells were also higher than in other HCC cell lines.
When the MZF-1 vector was transfected to the Huh-7 cells, both MZF-1 and Elk-1 expressions increased, and the transcriptional activities also proportionally increased (S1B and S1C Fig). S1D Fig confirms that in all five HCC cell lines tested, enhanced Elk-1 or MZF-1 increased transcriptional activity. The enhancement of both Elk-1 and MZF-1 were greater than that of Elk-1 or MZF-1 alone in terms of transcriptional activity. These results confirmed previous findings that PKCα expression is regulated by the two transcription genes [5][6][7].
To determine whether transcriptional activity can be regulated by the binding activity of MZF-1 and Elk-1 to PKCα promoters, a mutated form of DNA-binding-deficient (ΔDBD) mutant was used. Fig 1A shows that when HepG2 and Huh-7 cells were transfected with either full-length MZF-1 or Elk-1, the transcriptional activity increased compared with their respective control groups (with the empty pcDNA 3.1/Myc-His vector). In particular, when both transcriptional factors were used, the activity significantly increased, which is consistent with previous findings 3 . However, when the transfection included a ΔDBD gene vector of either MZF-1 (MZF-1ΔDBD) or Elk-1 (Elk-1ΔDBD), the transcriptional activity decreased, thereby indicating that the enhanced expression of PKCα is caused by the binding of MZF-1 or/and Elk-1 with the PKCα promoter.
Furthermore, we modified the above mentioned reporter vectors containing Elk-1 and MZF-1 binding sequences in the PKCα promoter by performing minor alterations, as follows: for promoter mut MZF-1, GGG was replaced with TTT in the MZF-1 binding site; for promoter mut Elk-1, CCT was replaced with AAG at the Elk-1 binding site; and for promoter mut MZF-1/Elk-1, both changes were made. The vectors with mutation in the MZF-1 binding site showed a slight increase in transcriptional activity, and the vectors with mutation in the Elk-1 binding site showed a slight decrease in transcriptional activity, whereas the vectors with both MZF-1 and Elk-1 mutations showed a significant decrease in transcriptional activity (Fig 1B). Similar findings were observed in cells with overexpressed MZF-1, overexpressed Elk-1. However, when the cells treated with both overexpressed MZF-1 and Elk-1, the transcriptional activity showed a significant decrease in all mutation types. The mutations in one gene binding site do not completely abolish the transcriptional activity, but the mutations in two gene binding sites nearly abolish the gene expression. Both binding sites function in a gene promoters [20], thereby indicating that the failure of Elk-1 and MZF-1 to bind and a form of mutative PKCα promoter results in lower transcriptional activity.

MZF-1 and Elk-1 form a complex and bind to the PKCα promoter region
To determine whether MZF-1 and Elk-1 form a complex, a co-IP process was performed with SK-Hep-1 cells. Antibody-bound fractions from the process were subjected to SDS-PAGE and then immunoblotted using the antibody of their counterparts (Fig 2A). The result of IP using the Elk-1 antibody revealed the presence of MZF-1 protein in the immunocomplex, which was in contrast with non-specific IgG IP and was similar to the MZF-1 antibody in Elk-1. Furthermore, when the cells were transfected with c-Myc-Elk-1ΔDBD vector, the MZF-1 protein was observed in the immunoprecipitated complex with Myc antibody (Fig 2B). When the cells were transfected with the FLAG-MZF-1ΔDBD vector, Elk-1 protein was observed in the immunoprecipitated complex with the FLAG antibody. Therefore, the presence of MZF-1 and Elk-1 in all cells indicated that Elk-1 binds to the N-terminal region of MZF-1, whereas MZF-1 binds to the C-terminal region of Elk-1 during the formation of a complex.
To confirm that endogenous PKCα promoter can be bound by endogenous Elk-1 and MZF-1 complex, chromatin immunoprecipitation (ChIP) and re-chromatin immunoprecipitation (Re-ChIP) assays were performed. The PKCα promoter fragment was amplified from the immunoprecipitated complex with either Elk-1 or MZF-1 antibodies (Fig 2C). In a Re-Chip assay, the amplified products were obtained by first using MZF-1 and then Elk-1 antibody. In another Re-ChIP assay, the order of antibodies used was reversed. In Fig 2D, when the cells were transfected with antisense Elk-1, the quantity of products of the PKCα promoter fragment amplified from the immunoprecipitated complex with MZF-1 antibodies was reduced. No change was observed in the cells transfected with sense Elk-1. Similarly, when the cells were transfected with antisense MZF-1, the quantity of products amplified from the immunoprecipitated complex with Elk-1 antibodies was also reduced, and again, no change was observed in cells transfected with sense MZF-1. These findings indicated the cooperative interaction between Elk-1 and MZF-1, which form a complex and bind to the promoter region of PKCα in SK-Hep-1 cells.

DBD-truncated-MZF-1 or DBD-truncated-Elk-1 protein interferes with PKCα expression and inhibits cell proliferation/migration/invasion
In the above experiment, the cooperative interaction of Elk-1 and MZF-1 regulated PKCα expression. When the transfection included a ΔDBD gene vector of either MZF-1ΔDBD or Elk-1ΔDBD, the transcriptional activity decreased compared with their respective control tests (pFLAG-CMV vector or pcDNA-Myc vector). Therefore, we hypothesized that the use of a mutated form of ΔDBD gene can interrupt the endogenous genes when forming a complex and binding to the PKCα promoter, thereby inhibiting gene expression and cell malignancy. When we transfected the MZF-1ΔDBD or Elk-1ΔDBD gene vector to SK-Hep-1 cells, the PKCα expression decreased, but the other PKC isoenzyme expressions, such as those of δ, ε, or ι, were not influenced (Fig 3A and 3B). Moreover, the MZF-1ΔDBD or Elk-1ΔDBD gene vector showed a dose-dependent reduction in cell proliferation in the same cells (Fig 3C and 3D), and the cell doubling time increased from 103.0% and 104.4% in the control (28.4 h) for 1 μM MZF-1ΔDBD and Elk-1ΔDBD to 148.0% or 147.7% for 2.5 μM. Furthermore, cell proliferation was significantly inhibited by 5 μM MZF-1ΔDBD and Elk-1ΔDBD.
To determine whether MZF-1ΔDBD and Elk-1ΔDBD are inhibited in cell migration/invasion, assays were performed on SK-Hep-1 cells 13 . Both MZF-1ΔDBD and Elk-1ΔDBD inhibited cell migration and invasion in cells by around 50% to 59% compared with the FLAG-or c-Myc-vector-transfected cells (Fig 3E to 3H). This effect further demonstrates the importance of the two genes in cell malignant progression.

Effect of MZF-1ΔDBD on the tumorigenicity of SK-Hep-1 Cells
Considering that MZF-1ΔDBD specifically inhibited PKCα expression and inhibited cell proliferation/migration/invasiveness, we further examined the effects of MZF-1ΔDBD on tumor formation. The experiment was performed by injecting 5 × 10 7 FLAG or MZF-1ΔDBD vectorpretreated SK-Hep-1 cells into nude mice, which were then monitored for tumor morphological growth. The FLAG vector-pretreated cells displayed rapid formation and growth of tumors (Fig 4A and 4B). By contrast, the mice injected with MZF-1ΔDBD vector-pretreated cells developed slow-growing tumors. The maximum inhibitory rate of tumor growth was 71.4% ± 9.5% (n = 5), whereas the mean of inhibitory rate of tumor growth from 28 d to 56 d was 70.2% ± 8.6%. The tumor formation time was prolonged from 10 d to 35 d, with a significant difference (P < 0.01) between the MZF-1ΔDBD vector-treated group (28.5 d ± 6.9 d) and the FLAG vector-treated group (16.2 d ± 3.8 d).

Discussion
The present luciferase reporter data demonstrate that Elk-1 and MZF-1 directly bind to the PKCα promoter and enhance PKCα transcription activity in HCC cells. These findings confirm previous reports that the knockdown of either Elk-1 or MZF-1 can reduce PKCα expression and Elk-1 or MZF-1 over-expression or the overexpression of both can increase PKCα expression in the same cells 3 , thereby suggesting that the cooperation of Elk-1 and MZF-1 may be involved in PKCα gene expression.
Bio-informational analysis results of protein-protein interaction showed that similar functional proteins tend to form a grape-like cluster pattern in cells [21]. Given that the DNA binding sites of Elk-1 and MZF-1 in PKCα promoters are in close proximity to each other [22], we proposed that MZF-1 and Elk-1 may directly interact inside a cluster. Our present data demonstrated that Elk-1 can directly bind to the N-terminal region of MZF-1, whereas MZF-1 can bind to the C-terminal region of Elk-1, and the resulting cluster binds to PKCα promoters. Moreover, when gene knockdown was performed on either MZF-1 or Elk-1, their overall DNA binding capability was compromised. These findings indicated that MZF-1 and Elk-1 do interact cooperatively and affect each other's binding capability and possibly form a complex.
The formation of a complex bond is a reasonable inference. MZF-1 reportedly has an acidic domain (amino acids 60-72 containing six amino acid of aspartates or glutamates) just upstream of the zinc finger regions [23]. This domain probably functions as a transcriptional activation domain, which is a common feature of various transcriptional factors [24,25], and is important in the binding with positively charged heparin-binding domains [26]. Heparinbinding sites often contain clusters (XBX, XBBX, and XBBBX) (B = basic residue; X = nonbasic residue) of basic amino acids [arginine (R) and lysine (K)] [27]. In the Elk-1 protein structure, two sequences may contain the following clusters: one (ARSSRNEYMRS; amino acids 146-156) is in the B domain where three clusters can be observed (ARS, SRN, and MRS); and the other (QKGRKPRD; amino acids 311-318) is in the D domain where three clusters are observed (QKG, GRKP, and PRD). The B domain enables Elk-1 to bind to a dimer of its cofactor, SRF [28], and the D domain enables binding to MAP kinases of ERK, JNK, and p38 families. Meanwhile, the acidic domain (60-72) of MZF-1 on the entire sequence of (1-485) is where direct interaction with the heparin-binding domains (146-156) of Elk-1 on the entire sequence of (1-428) occurs via co-IP (preparing for publication). We therefore suggest that the cooperative interaction between MZF-1 and Elk-1 may be mediated through the acidic and heparinbinding domains, respectively, although the search for the binding sequence on Elk-1 is currently ongoing.
In this study, when we transfected MZF-1 or Elk-1 with a deleted DNA-binding domain to a cell, the expression of PKCα decreased, and the cells exhibited decreased potential proliferation/migration/invasion activity and tumorigenesis, thereby confirming the results of previous reports, in which gene knockdown of PKCα expression by antisense oligonucleotide or shRNA inhibits cell proliferation/migration/invasion [5][6][7]. We therefore propose the following model: more often than not, in tumor cells, the binding of MZF-1 and Elk-1 followed by clustering to the binding site of the PKCα promoter can stimulate PKCα expression. This activity can be inhibited by transfecting a DNA-binding domain truncated MZF-1 to the cell; the a DNA-binding domain truncated MZF-1 will bind to an endogenous Elk-1, or vice versa, thereby decreasing the number of successful endogenous Elk-1 and MZF-1 interactions and their binding to the PKCα promoter. Consequently, PKCα expression and cell growth and invasion decrease. Thus, the information pertaining to the influence of upstream genes of PKCα is expected to undergo further basic and clinical trials with alternative directions in the development of anti-cancer solutions. expression of a 60bp PKCα promoter was observed in five HCC cells when transfected with 60bp PKCα promoter luciferase vector. Expression levels of PKCα, MZF-1 and Elk-1 proteins were estimated by Western blotting in five HCC cells. Huh-7 cells were transiently transfected with 1 μg of pGL3-60bp PKCα promoter luciferase reporter vectors and then the indicated amounts of empty pcDNA3 or MZF-1 expression vectors (B) and empty pcDNA3 or Elk-1 expression vectors (C). The bar below the two graphs displays Western blot of MZF-1 expression and Elk-1 expression. Five HCC cells were transiently co-transfected with 1 μg of 60 bp-PKCα promoter luciferase vector and 2.5 μg of MZF-1 or Elk-1 or MZF-1/Elk-1 expression vectors (D). Transcriptional activity is expressed as fold induction compared with the level obtained with each reporter vector in the absence of expression vectors. The experiments were repeated three times in triplicate. Data are means ± SE. (TIF)