Esculetin Suppresses Proliferation of Human Colon Cancer Cells by Directly Targeting β-Catenin Free

The Wnt family of secreted glycoproteins is highly conserved and regulates many biologic processes including development and disease (1–6). In the embryonic development process, appropriate activation of Wnt signaling regulates cell proliferation, differentiation, and determination of cell fate (7–11). Inappropriate activation of the Wnt signaling pathway is implicated in human diseases, including various human cancers (1, 2, 6, 12–14). In particular, deregulation of the Wnt signaling pathway is a critical event in colon carcinoma tumorigenesis (15–18). Thus, the Wnt signaling pathway is considered a key therapeutic and preventive target for cancer (19, 20).

In the absence of active Wnt signaling, the β-catenin destruction complex, which is composed of glycogen synthase kinase-3β (GSK-3β), adenomatous polyposis coli (APC), and Axin, catalyzes the phosphorylation of β-catenin leading to its proteosomal degradation. Activation of the canonical Wnt signaling pathway through the formation of a ligand-activated receptor complex inhibited the formation of the β-catenin destruction complex, which allows β-catenin to accumulate and subsequently translocate to the nucleus. Nuclear β-catenin directly binds with the T-cell factor (Tcf)/lymphoid enhancer factor (LEF) family, and these interactions stimulate transcription of Wnt target genes, whose promoters contain binding sites of the Tcf-transcription factor (3, 21–23).

Colorectal cancer (CRC) is the third most commonly diagnosed cancer in males and the second most frequent cancer reported in females (24). In the majority of colon cancers, the canonical Wnt/β-catenin pathway is constitutively active (15–18). Approximately 90% of colon cancers exhibit mutation of the APC or Axin genes, which leads to a disruption of the β-catenin destruction complex and accumulation of β-catenin. Multiple mutations lead to the nuclear accumulation of β-catenin and subsequent formation of a nuclear β-catenin–Tcf transcription complex (17, 25, 26). This leads to the inappropriate activation of its target genes, including c-myc (27) and cyclin D1 (28, 29), and also plays an essential role in proliferation of colon cancer cells. Therefore, disrupting the β-catenin–Tcf complex and inhibiting the nuclear function of β-catenin is considered to be therapeutically beneficial and could be useful for preventing colon cancer.

Here, we performed virtual structure-based screening of a natural product compound library using the crystal structure of the β-catenin–Tcf complex (PDB ID:1JPW). Esculetin (6,7-dihydroxy-2-chromenone), a derivative of coumarin that is present in many medicinal plants (30–33), was identified as a potential inhibitor of β-catenin–Tcf-mediated transcription. Esculetin exhibits many pharmacologic effects, including inhibiting lipooxygenase (34, 35) and acting as an anticoagulant (36). In addition, inhibition of cancer cell growth by esculetin has been reported previously (37–39). Although esculetin has shown antiproliferative effects in cancer cells, the molecular mechanism by which this occurs has not been investigated carefully. In the present study, we demonstrated that esculetin exhibits inhibitory effects against human colon cancer cells by directly targeting β-catenin, leading to the disruption of the β-catenin–Tcf complex. These results suggested that esculetin has a significant therapeutic and preventive potential against human colon cancer.

Esculetin and anti-β-actin were purchased from Sigma-Aldrich. McCoy's 5A and RPMI1640 medium were obtained from Thermo Fisher Scientific. Basal Medium Eagle (BME), gentamicin, penicillin/streptomycin, and l-glutamine were obtained from Invitrogen. [γ-³²P]-ATP and the chemiluminescence detection kit were obtained from Amersham Pharmacia Biotech. CNBr-Sepharose 4B beads were purchased from GE Healthcare. The MTS solution was purchased from Promega. Antibodies against β-catenin and cyclin D1 were obtained from Santa Cruz Biotechnology. Antibodies against c-Myc and Tcf4 were purchased from Cell Signaling Technology.

The Myc-tagged β-catenin TM mutant, in which Lys312, Gly307, Lys345, and Asn387 was changed to Glu, Val, Glu, and Ala, respectively, was generated using the QuickChange II site directed mutagenesis kit (Stratagene).

All cell lines were purchased from the American Type Culture Collection and were cytogenetically tested and authenticated before the cells were frozen. Each vial of frozen cells was thawed and maintained in culture for a maximum of 8 weeks. HCT116 human colon cancer cells were cultured in McCoY's 5A medium supplemented with 10% (v/v) FBS (Atlanta Biologicals) and HCT15 and DLD-1 human colon cancer cells were cultured in RPMI1640 medium supplemented with 10% (v/v) FBS.

Cells were seeded (2 × 10³ cells per well) in 96-well plates and incubated for 12 hours and then treated with different doses of esculetin. After incubation for 72 hours, 20 μL of Cell Titer96 Aqueous One Solution (Promega) were added and cells were incubated for 1 hour at 37°C in a 5% CO₂ incubator. Absorbance was measured at 492 nm.

To activate Sepharose 4B beads, esculetin and Sepharose 4B powder (0.3 g) were suspended in 1 mmol/L HCl. Then the coupled solution (0.1 mol/L NaHCO₃, pH 8.3 and 0.5 mol/L NaCl) was added and rotated overnight at 4°C. The mixture was washed with coupling buffer, and transferred to 0.1 mol/L Tris–HCl buffer (pH 8.3). The excess of uncoupled esculetin was removed by washing with 0.1 mol/L acetate buffer (pH 4.0) and 0.1 mol/L Tris–HCl buffer (pH 8.0) containing 0.5 mol/L NaCl.

Proteins (500 μg) of HCT116, HCT15, and DLD-1 cells extracted with reaction buffer were mixed with Sepharose 4B beads (as a negative control) or esculetin-Sepharose 4B beads (100 μL) in reaction buffer [50 mmol/L Tris, pH 7.5, 5 mmol/L EDTA, 150 mmol/L NaCl, 1 mmol/L dithiothreitol (DTT), 0.01% Nonidet P-40, 2 mg/mL BSA, 0.2 mmol/L phenylmethylsulfonylfluoride (PMSF), and 1× protease inhibitor mixture]. After incubation with gentle rocking overnight at 4°C, the beads were washed 5 times with buffer (50 mmol/L Tris, pH 7.5, 5 mmol/L EDTA, 150 mmol/L NaCl, 1 mmol/L DTT, 0.01% Nonidet P-40, and 0.02 mmol/L PMSF) and binding was visualized by Western blotting.

In brief, cells (8 × 10³ per well) suspended in BME supplemented with 10% FBS were added to 0.3% agar with different doses of esculetin in a top layer over a base layer of 0.6% agar. The cultures were maintained at 37°C in a 5% CO₂ incubator for 2 weeks and then colonies were counted under a microscope using the Image-Pro Plus Software version 4 program (Media Cybernetics).

Cell lysates were prepared with lysis buffer (10 mmol/L Tris, pH 7.5, 150 mmol/L NaCl, 5 mmol/L EDTA, 1% Triton X-100, 1 mmol/L DTT, 0.1 mmol/L PMSF, 10% glycerol, and protease inhibitor cocktail tablet). The protein concentration was measured using the bicinchoninic acid assay (Pierce Biotechnology). A horseradish peroxidase-conjugated secondary antibody (Pierce Biotechnology) was used and the signal was detected with chemiluminescence reagent (Amersham).

Total RNA was extracted from cells using the TRizol reagent (Tel-Test Inc.) according to the manufacturer's instructions. cDNA was synthesized using the Superscript pre-amplification system (Life Technologies Inc.). PCR reactions were performed using the following conditions: 94°C for 5 minutes and 25 to -28 amplification cycles at 94°C for 30 seconds, the appropriate annealing temperature for 45 seconds, and 72°C for 30 seconds, and a final extension at 72°C for 10 minutes. The PCR primers were GAPDH (forward: 5′-CTCAGACACCATGGGGAAGGT-3′) and reverse: 5′-TGATCTTGAGGCTGTTGTCATA-3′); c-myc (forward: 5′-TGTCAAGAGGCGAACACACAACGTC-3′ and reverse: 5′- ATCTTTCAGTCTCAAGACTCAGCCA-3′); and cyclin D1 (forward: CCTGTCCTACTACCGCCTCA and reverse: 5′- TCCTCCTCTTCCTCCTCCTC-3′).

HCT116 and HCT15 cells were treated with the indicated concentration of esculetin for 24 hours and then disrupted with lysis buffer (50 mmol/L Tris, pH 8, 250 mmol/L NaCl, 5 mmol/L EDTA, 0.1% NP-40, 10% glycerol, and 1× protease inhibitor cocktail). Cell lysates were cleared by centrifugation, and immunoprecipitations were performed by incubating overnight with anti-β-catenin. Protein A/G Plus Agarose (Santa Cruz Biotechnology) was added and the solution was incubated for 3 hours at 4°C. Unbound proteins were removed by washing 4 times with lysis buffer. Bound proteins were harvested by boiling in sample buffer, and resolved by electrophoresis in 8% SDSPAGE. β-catenin and Tcf4 proteins were visualized using a chemiluminescence reagent (Amersham).

Transient transfection was performed using jetPEI (VWR), and assays to detect firefly luciferase and Renilla activities were performed according to the manufacturer's instructions (Promega). Briefly, cells were seeded into 96-well plates and cotransfected with 50 ng of the Renilla luciferase internal control gene and 100 ng of the TOP-flash luciferase reporter construct containing three tandem Tcf consensus binding sites upstream of luciferase cDNA, or the FOP-flash luciferase reporter construct, a plasmid with mutated Tcf binding sites. After 12 hours of transfection, cells were incubated with the indicated concentration of esculetin for another 24 hours. Luciferase and Renilla activities were measured using substrates included in the reporter assay system (Promega). The luciferase activity was normalized to Renilla activity.

Xenopus laevis embryos were obtained by artificial fertilization. Vitelline membranes were removed by immersing embryos in a 2% cysteine solution (pH 8). Embryos were injected with 250 pg of β-catenin mRNA alone or together with the indicated concentration of esculetin and then cultured to stage 40 in 67% Leibovitzs L-15 medium (Invitrogen) containing BSA (1 mg/mL), 7 mmol/L Tris–HCl (pH 7.5), and gentamicin (50 μg/mL). Xenopus embryos at stage 40 do not have vertebrae, but have the notochord, which is a precursor of the backbone.

Female athymic nude mice (6- to 7-week old) were purchased from Central Lab Animal Inc. (Seoul, Korea) and maintained under “specific pathogen-free” conditions on the basis of the guidelines established by the University of Seoul National University (Seoul, Korea) Institutional Animal care and Use Committee (SNU120106-4). Mice were divided into four groups: (i) untreated vehicle group (n = 10); (ii) 20 mg esculetin per kilogram of body weight (n = 10); (iii) 100 mg esculetin per kilogram of body weight (n = 10); (iv) no cells and 100 mg esculetin per kilogram of body weight (n = 10). HCT116 cells (2.5 × 10⁶ cells/100 μL) suspended in serum-free McCoy 5A medium were injected subcutaneously in the flank of each animal. Esculetin or vehicle was administered intraperitoneally 3 times per week for 11 days. Dimethyl sulfoxide (DMSO; 4%) and polyethylene glycol (40%) were diluted with PBS buffer and used as the vehicle. At the end of the experiment, mice were sacrificed, and tumors were extracted. Tumor volume was calculated from measurements of two diameters of the individual tumor base using the following formula: tumor volume (mm³) = (length × width × height × 0.52).

For histopathologic examination, paraffin sections (4 μm) were stained with hematoxylin and eosin. Immunohistochemical staining for Ki-67 (Lab Vision Corporation), cyclin D1 (Santa Cruz Biotechnology), or c-Myc (Novous Biologicals) was performed using the indirect avidin–biotin-enhanced horseradish peroxidase method (Vector Laboratories). For quantitation, each slide was scanned to obtain an overall impression of the staining patterns and 10 representative ×200 power (Ki-67 and cyclin D1) or ×400 (c-Myc) photomicrographs were taken with a digital camera, avoiding gross necrotic areas. The positively stained cancer epithelial cells within each photomicrograph were counted. Counting the total number of cancer cells was aided with the Image Pro+ image-processing program. The Ki-67 and cyclin D1 indices were based on the counting of approximately 7,000 total cells per tumor slide.

To find potential inhibitors of β-catenin, a molecular docking method was developed using the Glide module from Schrödinger Suite 2011 (40, 41) and used to perform the virtual screening. A crystal structure of a human Tcf-4–β-catenin complex (PDB ID:1JPW; refs. 42, 43) was downloaded from the PDB Bank (44) for virtual screening studies. This is an X-ray diffraction structure with a resolution of 2.5 Å. Waters, metals, and Tcf-4 were stripped from the structure, and then hydrogens and atom charges are added to the structure using the Protein Preparation Wizard in Schrödinger suite 2011 with the standard procedure outlined. Two pockets were generated respectively within a 30-ų grid based on the binding site of Tcf-4 with β-catenin. One pocket was centered with Lys312 and the other with Lys435. The 2-D TCMD (Traditional Chinese Medicine Databse) structure database, which consists of approximately 9,000 structures of natural products (45), was first converted to a three-dimensional (3D) structure database using the LigPrep module of the Schrödinger Suite of software and then used for virtual screening. High-throughput virtual screening (HTVS) docking is usually first performed because it is intended for the rapid screening of large numbers of ligands followed by standard and extra precision (SP and XP) docking.

All data are presented as means ± SD of triplicate samples from at least three independent experiments. Differences between means were assessed by ANOVA, and the minimum level of significance was set at P value less than 0.05.

To find potential inhibitors of β-catenin, we performed structure-based in silico screening using a molecular docking method as described in the Materials and Methods section. The crystal structure of β-catenin bound to the Tcf protein has been determined and two major pockets of β-catenin for binding with Tcf were identified previously (42, 43). Based on the crystal structure, we screened a small-molecule library to identify compounds that could possibly bind to the two pockets of β-catenin, sites that are critical for the binding of Tcf to β-catenin. We used in silico screening to select three top-ranked compounds and tested their effect on human colon cancer cell growth. Of the three, esculetin was the most effective (Supplementary Fig. S1). Esculetin (Fig. 1A) is a derivative of coumarin that is present in many medicinal plants (30–33). The computational prediction of the binding affinity between esculetin and β-catenin was predicted to be very good with a score of −6.75 kcal/mol. In addition, computer modeling results predicted that esculetin could directly bind to the Lys312, Gly307, Lys345, and Asn387 residues of β-catenin through hydrogen bonds (Fig. 1B). Therefore, we suggest that esculetin might contribute to the inhibition of the β-catenin–Tcf complex formation by directing interacting with β-catenin.

To confirm the results of the computer docking model, we determined whether esculetin could directly bind to β-catenin. We performed pull-down assays using esculetin-conjugated Sepharose 4B beads with cell lysates of human colon cancer cells. These results showed that esculetin directly bound to β-catenin precipitated from HCT116, HCT15, and DLD1 cell lysates (Fig. 2A). In addition, molecular docking of the esculetin and β-catenin complex indicated that esculetin binds with β-catenin at Lys312, Gly307, Lys345, and Asn387. In order to confirm this prediction, we generated myc-tagged wild-type β-catenin and mutant (TM) β-catenin plasmids in which Lys312, Gly307, Lys345, and Asn387 were replaced with Glu, Val, Glu, and Ala, respectively. Lysates from 293T cells were transfected with myc-tagged wild-type or TM β-catenin and were incubated with esculetin-conjugated Sepharose 4B beads for a pull-down assay. These results indicated that mutant (TM) β-catenin failed to bind esculetin (Fig. 2B), which supports the computer docking model of the predicted binding sites between esculetin and β-catenin.

We then investigated whether the direct binding of esculetin with β-catenin could interrupt the association of β-catenin with Tcf4. Tcf4 in nuclear extracts of esculetin-treated HCT116 and HCT15 cells was coimmunoprecipitated with anti-β-catenin. Indeed, esculetin suppressed the association of β-catenin with Tcf4 in a dose-dependent manner (Fig. 2C). These results suggested that esculetin might be able to attenuate the association of the β-catenin–Tcf4 transcriptional complex by directly interacting with β-catenin.

To confirm the effect of esculetin on the β-catenin–Tcf transcriptional complex in cells, we measured TOP/FOP-luciferase activity in HCT116, HCT15, and DLD1 colon cancer cells. Colon cancer cells were cotransfected with a luciferase reporter gene containing three tandem Tcf consensus binding sites (TOP) or a mutated Tcf binding site (FOP) and the Renilla-luciferase reporter gene as a normalizing transfection control. At 12 hours posttransfection, cells were treated for 24 hours with the indicated concentration of esculetin. TOP or FOP-luciferase activity was measured and normalized to Renilla-luciferase activity. Data are shown as relative values compared with untreated control. Esculetin effectively suppressed the TOP-luciferase activity in a dose-dependent manner, without significantly affecting FOP-luciferase activity (Fig. 3A). These results indicated that esculetin inhibited the β-catenin–Tcf complex signaling, which is consistent with the finding that esculetin disrupted the association of the β-catenin–Tcf transcriptional complex.

Next, we examined the expression of c-myc and cyclin D1, which are direct target genes of the β-catenin–Tcf complex. We treated HCT116, HCT15, and DLD1 colon cancer cells with esculetin and assessed the protein abundance of c-Myc and cyclin D1. Treatment with esculetin for 24 hours substantially repressed the protein levels of c-Myc and cyclin D1, but had no effect on the levels of β-catenin or Tcf (Fig. 3B). In addition, we examined the mRNA levels of c-myc and cyclin D1 in esculetin-treated human colon cancer cells. Consistent with the protein abundance of c-Myc and cyclin D1, the mRNA level of c-myc and cyclin D1 was attenuated by esculetin (Fig. 3C). Taken together, these results indicated that esculetin potently antagonizes the cellular effects of β-catenin–Tcf-dependent transactivation.

To further confirm the inhibitory effects of esculetin on the Wnt–β-catenin signaling pathway, we determined the effect of esculetin on Wnt–β-catenin signaling in vivo. During early Xenopus development, ectopic expression of β-catenin on the future ventral side leads to duplication of the embryonic body axis. This developmental model has been well established to investigate the regulation of Wnt–β-catenin signaling in vivo (46). Thus, this developmental model pathway provides rigorous methods for determining the biologic effects of compounds on the β-catenin–Tcf pathway. Approximately 90% of embryos injected with 250 pg of β-catenin mRNA showed duplication of ventral–dorsal axis (Fig. 3D). In contrast, coinjection of β-catenin mRNA with 5 pmol of esculetin inhibited the induction of axis duplication, with no contribution from the DMSO vehicle (Fig. 3D). These experiments further support the notion that esculetin inhibits the β-catenin–Tcf-dependent signaling pathway.

To determine the biologic effects of esculetin in cells, HCT116, HCT15, and DLD1 colon cancer cells were treated for 72 hours with various concentration of esculetin and cell viability was assessed by the MTS assay. Esculetin decreased viability of all three cell lines in a dose-dependent manner (Fig. 4A). In addition, we evaluated the effect of esculetin on anchorage-independent cell growth. HCT116, HCT15, and DLD1 colon cancer cells were seeded with esculetin in 0.3% agar and incubated for 3 weeks. Data showed that esculetin significantly suppressed anchorage-independent cell growth in a dose-dependent manner (Fig. 4B). These results indicated that esculetin had antiproliferative and antitumorigenic effects against human colon cancer cells.

To examine the antitumor activity of esculetin in vivo, HCT116 cancer cells were injected into the right flank of individual athymic nude mice. Mice were injected intraperitoneally with vehicle or esculetin at 20 or 100 mg/kg body weight 3 times a week for 2 weeks. Esculetin treatment suppressed xenograft tumor development in mice. Treatment with 20 and 100 mg/kg esculetin significantly inhibited HCT116 tumor size by 44% and 64%, respectively, relative to the vehicle-treated mice (Fig. 5A). All mice, including those treated with only esculetin, were viable at the end of the experiment, and body weight loss was not observed in any mice treated with esculetin. This indicates that the doses used were not overtly toxic to the animals.

Using the xenograft tumor tissues, we examined the effect of esculetin on a tumor proliferation marker, Ki-67. The expression of Ki-67 was markedly inhibited by treatment with esculetin. In addition, we investigated the effect of esculetin on the direct targets of Wnt–β-catenin signaling, cyclin D1, and c-Myc by immunohistochemical analysis of HCT116 xenograft tumor tissues. Expression of cyclin D1 and c-Myc also was suppressed by treatment with esculetin (Fig. 5B). These data suggested that esculetin inhibited HCT116 colon tumor development by suppressing Wnt–β-catenin signaling.

CRC is the third most common cancer in men and women and is the third leading cause of cancer death (39, 40). Thus, the causes and methods of preventing and treating colon cancer are a high priority. In particular, abnormal activation of Wnt signaling has been tightly linked to colon cancer. Therefore, identifying small molecules targeting Wnt signaling is a promising preventive or treatment strategy against cancer.

In the current study, we identified a natural product, esculetin, as a potent inhibitor of Wnt signaling. Previously, the inhibitory effect of esculetin on cell proliferation in many cell lines has been reported (32, 33, 43, 44). However, the molecular mechanism of the antiproliferation effect of esculetin against colon cancer cells is not clearly understood. In the present study, we demonstrated that esculetin could disrupt the interaction between β-catenin and Tcf by targeting β-catenin, leading to suppression of the proliferation of three different colon cancer cell lines without exhibiting any significant cytotoxic effects on the viability of normal human colon cells (Supplementary Fig. S2). In addition, we demonstrated that esculetin potently antagonized the cellular effects of β-catenin–dependent activity in human colon cancer cells and in an animal model (Fig. 3). Moreover, esculetin effectively decreased viability and inhibited colony formation of human colon cancer cells (Fig. 4) and tumorigenesis in vivo (Fig. 5). A high level of the β-catenin–Tcf transcriptional complex caused by the accumulation of β-catenin is associated with human colon cancer and colon carcinogenesis (9, 14, 16, 41, 42). Overall, our results suggest that esculetin could be a potential chemopreventive agent against colon carcinogenesis. In particular, our results showing that esculetin could effectively inhibit β-catenin–induced morphogenesis in the frog provide strong evidence supporting the potential preventive effect of esculetin.

Recently, a number of existing drugs and natural compounds were reported to be antagonists of Wnt signaling. However, their molecular mechanisms of action and cellular targets are largely unknown, making their applications in cancer prevention and treatment and drug development very limited. In the present studies, we identified esculetin as a potent natural inhibitor of the Wnt–β-catenin signaling pathway and elucidated its molecular mechanism of action. Thus, we suggest that esculetin could be a potent cancer therapeutic and preventive agent. However, several challenges remain to be addressed in the development of clinically useful Wnt pathway inhibitors. We screened many natural compounds that might affect the binding of Tcf to β-catenin. However, biochemical and structural studies showed that the region of β-catenin that binds to Tcf overlaps other binding partners, such as E-cadherin and APC (47, 48). In the future, we will address whether esculetin can disrupt other interactions or whether it is selective for the β-catenin–Tcf protein–protein interaction. In addition, although esculetin, a nontoxic natural compound, appears to be potentially useful in treating or preventing cancer, its efficacy, safety, and toxicity need to be elucidated.

In conclusion, we identified esculetin as a potent inhibitor of Wnt–β-catenin signaling, which acts by targeting β-catenin to effectively suppress proliferation of human colon cancer cells both in culture and in vivo. Overall, these results indicate that esculetin could be a potential therapeutic or preventive and safe agent against human colon cancers.

No potential conflicts of interest were disclosed.

Conception and design: S-Y. Lee, T.G. Lim, H. Chen, S.K. Jung, H-J. Lee, K.W. Lee, Z. Dong

Development of methodology: S-Y. Lee, H. Chen, Z. Dong

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): S-Y. Lee, T.G. Lim, M-H. Lee, A. Shin, Z. Dong

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): S-Y. Lee, T.G. Lim, H. Chen, Z. Dong

Writing, review, and/or revision of the manuscript: S-Y. Lee, T.G. Lim, H. Chen, K.W. Lee, A.M. Bode, Z. Dong

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): S-Y. Lee, H. Chen, H-J. Lee, M-H. Lee, D-J. Kim, A.M. Bode, Z. Dong

Study supervision: H-J. Lee, K.W. Lee, A.M. Bode, Y-J. Surh, Z. Dong

This work was financially supported by The Hormel Foundation and National Institutes of Health (grant nos. CA120388, CA1666011, CA172457, and R37 CA081064), the World Class University Grant, Ministry of Education, Science, Technology, Republic of Korea (grant no. R31-2008-000-10103-0), the National Leap Research Program (No. 2010-0029233) through the National Research Foundation of Korea funded by the Ministry of Education, Science and Technology of Korea, Republic of Korea and by the Global Frontier Project grant (NRF-M1AXA002-2012M3A6A4054949) of National Research Foundation funded by the Ministry of Education, Science and Technology of Korea, Republic of Korea.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.