The molecular 3-D structure of Dvl2-PDZ domain was extracted from the Protein Data Bank (www.pdb.org, PDB CODE: 3CBY). The chemical structure of curcumin was constructed and optimized in the implicit solvent as preparation for docking, and the docking performance was executed employing the CDOCKER module on the platform of Discovery studio 2.5 (Accelrys Inc.). The critical peptide ligand binding site was defined as docking region (24). The conformation search space was limited to a spherical region with the center of (37.6, 3.0, 7.3) and the radius of 9.5 Å. The other parameters were set according to the default values of the module, with all docking procedures using CHARMm forcefield, and a grid extension of 8.0. The ligand partial charge method was also used according to CHARMm. Ten top hits were obtained as docking results. Simulated annealing method was employed for the final conformation treatment: the system was heated to 700 K with 2,000 steps and then cooled to 300 K with 5,000 steps. Finally, the best conformation was selected as the object of analysis according to the values of score.

Binding free energy calculation was performed with MM-PBSA module of Amber12 (25). In the MM-PBSA method, the free energy of the receptor/ligand binding, ΔG_bind, is obtained from the difference between the free energies of the receptor/ ligand complex (ΔG_cpx) and the unbound receptor (ΔG_rec) and ligand (ΔG_lig) as following: (1) Δ G bind = Δ G cpx − ( Δ G rec + Δ G lig )

The binding free energy (ΔG_bind) is evaluated as a sum of the changes of molecular mechanical (MM) gas-phase binding energy (ΔE_MM), solvation free energy (ΔG_sol) and entropic contribution (-TΔS): (2) Δ G bind = Δ E bind − T Δ S (3) Δ E bind = Δ E MM + Δ G sol

Curcumin was obtained from Shanghai University of Traditional Chinese Medicine. All cell culture reagents were purchased from Biowest (USA). MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide) detection kit and Annexin V-FITC apoptosis detection kit were obtained from BestBio (Shanghai, China). Western blot antibodies specific to β-catenin and β-actin were ordered from Cell Signaling Technology (Beverly, MA, USA). BEL-7402 and QGY-7703 human HCC cell lines were provided by Shanghai Institute of Biochemistry and Cell Biology. They were cultured in RPMI-1640 medium supplemented with 10% (v/v) fetal bovine serum, 1 x penicillin-streptomycin solution in a humidified 5% CO₂ atmosphere at 37°C.

Cell proliferation was determined by MTT assay and used according to the manufacturer's instructions. Briefly, BEL-7402 and QGY-7703 cells were seeded into 96-well plates at a density of 5×10³ per well (100 μl). After 24 h, the indicated concentrations of curcumin were added. After incubating for 48 h, cells were washed twice with PBS. MTT medium (10 μl) were then added into each well, at which time cells were incubated for another 4 h. After which, the medium was removed, and 150 μl of dissolution was added into each well. The plate was gently rotated on an orbital shaker for 10 min to dissolve the precipitation completely. The absorbances were detected at 492 nm with a Microplate reader.

BEL-7402 and QGY-7703 cells were treated with curcumin in complete medium for 48 h as previously described. Following the treatment, cells were harvested by trypsin (not containing EDTA) and rinsed twice with PBS at 4°C. Cells were then resuspended in 1X Annexin V binding buffer. A total of 5 μl of Annexin V-FITC solution was added to each tube. All tubes were incubated for 15 min at 4°C in darkness. PI solution (15 μl) was added to each tube. All tubes were incubated for another 5 min at 4°C in darkness. Cells were then analyzed using flow cytometry (Accuri C6, USA).

BEL-7402 and QGY-7703 cells were cultured with curcumin at various concentrations on 6-well plates, and after 24 h the cells were scratched in straight line with tips. The scratched cells were washed off with PBS and the resting cells were cultured in fresh medium: cell migrations were measured at 6, 12 and 24 h, respectively, using ImageJ software.

To detect protein expression and modification in response to treatment with curcumin, HCC cells were plated onto 6-well plates at a density of 2×10⁵ cells/ml and were then treated with various concentrations of curcumin. After incubating for 48 h, cells were lysed in cold RIPA lysis buffer. Total protein were extracted with high-salt buffer (0.5% sodium deoxycholate, 1% SDS, 1 mM sodium orthovanadate, 1 mM β-glycerol phosphate, 1 mM sodium fluoride, 2.5 mM sodium pyrophosphate) that contains a protease inhibitor cocktail (Roche, Nutley, NJ, USA). Protein samples were separated by SDS-PAGE, transferred onto PVDF membrane, and immunoblotted with the corresponding antibodies. The signal was visualized with Enhanced Chemiluminescence Plus (ECL Plus) detection system (Dingguo, China).

The BEL-7402 cells and QGY-7703 cells were treated with curcumin at 10, 20, 40 and 80 μM for 48 h as previously described. Total RNAs were extracted from the cells using a commercially available RNA-Bee isolation kit (Tel-Test). Standard reverse transcriptions were performed with 500 ng of total RNA using TIANScript RT kit (Tiangen, Beijing, China). RT-PCRs were performed by using 1 μl of cDNA template, 10 pmol of primers, and a PCR premix (1 unit of Taq DNA polymerase, 250 mM dNTPs, 10 mM Tris-HCl, 40 mM KCl and 1.5 mM MgCl₂; Tiangen). The following primers were used in the PCR reactions: c-myc forward, 5′-ctaccctctcaacgacagcag-3′; reverse, 5′gtgtgttcgcctcttgacatt-3′; VEGF forward, 5′-gcagaatca tcacgaagtggt-3′; reverse, 5′-catttgttgtgctgtaggaagc-3′; cyclin D1 forward, 5′-atctacaccgacaactccatcc-3′; reverse, 5′-gcattttggagag gaagtgttc-3′; β-actin forward, 5′-agagctacgagctgcctgctg-3′; reverse, 5′-agtacttgcgctcaggagga-3′.

The amplified products obtained from β-actin specific primers served as internal controls. PCRs were conducted using Bio-Rad T-100 (Bio-Rad, Hercules, CA, USA) with a denaturation step at 94°C for 5 min; 30 cycles of 94°C for 30 sec, 62°C for 30 sec and 72°C for 30 sec; and a final extension at 72°C for 10 min. PCR amplifications were verified to be in the linear range.

Data were presented as mean ± SD for three separate experiments. One-way ANOVA was employed for statistical analysis using SPSS 17.0. P<0.05 was considered statistically significant.

The details of interaction between curcumin and dishevelled2 (Dvl2) PDZ domain is shown in Fig. 1A. According to the figure, the two separated hydrogen atoms of -NH₂ group of ARG338 formed hydrogen bonds with the C=O group in curcumin and 2 (-NH₂) groups of ARG338 with positive charge had cation-π interaction with an aromatic ring of curcumin. The backbone -NH groups of ILE282 and GLY284 formed hydrogen bonds with the C=O and OH in curcumin, respectively. The configuration of curcumin fits into a hydrophobic pocket of Dvl2-PDZ domain that is mainly constructed by LEU278, LGY279, ILE280, SER281, VAL283, PRO313, ASP331, VAL334, LEU337, VAL341 and HIS342 in a shape-shape complimentary pattern, as shown in Fig. 1B. To evaluate the binding free energy of curcumin with Dvl2-PDZ domain, 3 nsec MD simulation was performed for each complex. We also calculated binding free energy of Dvl2-PDZ with a well-known inhibitory N1 peptide, pep-N1 (WKDYGWIDG) (26). Table I shows free energies of each items of protein in complex with curcumin and peptide based on the heavy atom RMSD of the last 2 nsec of each simulation. The RMSD values prove that the conformation of both complexes were stable.

To determine potential cytotoxic and anti-proliferative effects of curcumin, the human HCC cell lines BEL-7402 and QGY-7703 were cultured with curcumin at various concentrations (0–80 μM). Cell viability was then determined by MTT assay. Results showed that treatment with curcumin led to a significant dose-dependent inhibition of HCC cell growth in vitro (Fig. 2A). The half maximal (50%) inhibitory concentrations (IC₅₀) for BEL-7402 and QGY-7703 cells were approximately 32 and 33 μM, respectively.

Induction of cell apoptosis was confirmed by Annexin V-FITC staining in BEL-7402 and QGY-7703 cells. Results showed that treatment with curcumin led to significant dose-dependent apoptosis in HCC cells in vitro. According to Fig. 2B, the upper right quadrant represents late apoptosis, and the lower right quadrant represents early apoptosis. Increasing concentrations of curcumin, 0, 10, 20, 40 and 80 μM, respectively, were added to BEL-7402 and QGY-7703 cells for 48 h. As a result, the rates of cell apoptosis were 5.8, 9.3, 13.1, 32.8 and 38.0% for BEL-7402 and 7.7, 13.4, 15.8, 25.0 and 40.8% for QGY-7703, respectively. Thus, as the concentration of curcumin increased, the rates of apoptosis for BEL-7402 and QGY-7703 cells were also increased. The standard deviations were calculated based on three independent experiments.

The inhibitory effect of curcumin for HCC cell migration was confirmed by cell migration test. After 24 h, cells migrated to the scratch area, which achieved healing in control group without the curcumin treatment. When treated with curcumin, especially at 80 μM, cells grew and migrated at a much slower pace (Fig. 3A). The width of the healewound was significantly different between groups treated with curcumin and the control group after 6 h, and the differences increased further after 12 and 24 h (Fig. 3B).

After being treated with curcumin, expression of β-catenin in BEL-7402 and QGY-7703 cells decreased in a dose-dependent pattern as shown by western blot analysis. As a result, β-catenin activity in these cells was reduced in general (Fig. 4A). Exposure to curcumin resulted in a dose-dependent decrease in the expression of cyclin D1 mRNA in both BEL-7402 and QGY-7703 cells, as demonstrated by RT-PCR experiments. Furthermore, the expressions of c-myc and VEGF was also significantly reduced at transcriptional level (Fig. 4B).