The crystal structures of human mortalin (PDB ID: 4KBO) and p53 (PDB ID: 1OLG) were obtained from Protein Data Bank (PDB) (31), while the ligand, ARC (Compound ID: 5472440) was from PubChem Compound (32) database. ArgusLab 4.0.1 (M 2004), a freely available molecular modelling, graphics and drug design package, was used for docking analyses. The structures of protein and the ligand molecules were prepared and the receptor grid around the binding site residues selected from the receptor was defined. The grid resolution was kept as 0.4 Å and an exhaustive search docking was performed. The docking process was carried out using the 'ArgusDock' docking engine, calculation type as 'Dock' and the ligand was kept in rigid mode. The scoring function used was an empirical scoring function, 'AScore', which takes into account van der Waals energy, hydrophobic component, hydrogen bond and deformation penalty. The parameter file, 'AScore.prm' was used to compute the binding energies. A total of 150 docking poses were generated, ranked according to the scoring function, and the highest scoring pose was used in further analyses.

MD simulations were carried out using the GROMACS package (33,34). The force field Gromos43a1 (35) was used for ARC-docked mortalin structure (33,36,37). The GROMACS topology file was generated using the antechamber python parser interface (ACPYPE) script. The docked protein structure was solvated in a cubic box. Water molecules and appropriate counter-ions were added to neutralize the system. The solvated system was minimized using steepest descent and conjugate gradient methods until the force on each atom was less than 100 kJ/mol/nm. These geometry minimized systems were used for 10 nsec for carrying out isobaric (constant pressure-temperature) MD simulations. The temperature and pressure of the system was maintained at 300 K and 1 atmosphere pressure, respectively, with a time constant of 5 psec. A 2-fsec time step was used for integrating the equations of motion. Particle Mesh Ewald summation method along with periodic boundary conditions were also applied throughout to calculate the electrostatic potential between partial charges on atoms. Visual Molecular Dynamics (VMD) version 1.9.2 was used to calculate the root mean square deviations (RMSD) and hydrogen bond dynamics.

HT1080 (fibrosarcoma), A549 (lung carcinoma) and U2OS (osteosarcoma) human cell lines were purchased from DS Pharma Biomedical Co., Ltd., Osaka, Japan, and cultured in complete Dulbecco's modified Eagle's medium (DMEM; Life Technologies) medium using normal cell culture conditions with 10% fetal bovine serum as a supplement at 37°C, 5% CO₂ and 95% air in a humidified incubator. The cultured cells were maintained for 6–10 passages. ARC (Wako Pure Chemical Industries, Ltd., Osaka, Japan), GPSE and GPSE:50%γCD complex were dissolved in dimethylsulfoxide (DMSO), and directly added to the cell culture medium to obtain the working concentrations (as indicated in the respective figures).

The complex of green propolis supercritical extract (GPSE) with γCD was prepared by a conventional kneading method. GPSE and γCD were mixed in a small amount of water. The slurry was kneaded until it became a homogeneous paste. After freeze-drying of the paste, the GPSE-γCD complex thus obtained was used in the present study. The levels of artepillin in GPSE, and its complex with γCD, were determined by high-performance liquid chromatography (HPLC) using the Shimadzu HPLC system (LC-2010C; Shimadzu Corp., Kyoto, Japan) as previously described (38). A Phenomenex HPLC column [Luna 5u C18(2) 100A: 4.60 mm I.D. ×150 mm] was used and the fractionation was performed at 40°C using solution A: H₂O (0.5% acetic acid) and solution B: Acetonitrile with gradient program as follows: Isocratic 70% A (0–5 min), linear gradient 70% → 0% A (5–30 min); flow rate, 1 ml/min; injection volume, 10 µl. Detection was performed at 320 nm (Shimadzu HPLC system LC-2010C; Shimadzu Corp.).

Cytotoxicity assay was performed using 3-(4,5-dimethylthiazol-2-yl)-2, 5-diphenyltetrazolium bromide (MTT) assay (Life Technologies/Thermo Fisher Scientific, Waltham, MA, USA) in which cell viability was observed by the conversion of yellow MTT by the mitochondrial dehydrogenases of living cells into purple formazon (39). The statistical significance of the results was determined from 3–4 independent experiments including triplet or quadruplet sets in each experiment. IC₅₀ values were calculated by the linear regression methods in MS excel.

The cells were seeded in 12-well plates and treated with various concentrations of ARC (100–400 µM, as indicated in Fig. 2C), GPSE and GPSE-50%γCD (0.5%). Morphological changes were observed under a phase contrast microscope (Nikon Eclipse TE300; Nikon, Tokyo, Japan) after 48 h.

The cells (500 cells/well) were seeded in a 6-well plate. Following 24 h of incubation, when the cells had attached to the surface of the dish they were treated with ARC, GPSE or GPSE-50%γCD (0.1%) in culture medium (DMEM). The cultures were maintained for 10–15 days (when colonies formed in the control cultures) with a regular change of medium every 3rd day. The colonies were fixed in acetone: methanol (1:1), and stained with crystal violet (Wako Pure Chemical Industries Ltd.) at the end of experiments.

The cells were seeded on glass coverslips placed in 12-well plates and treated with either ARC (300 µM), GPSE or GPSE-50%γCD (0.5%) in DMEM. After 48 h of treatment, cells were fixed in 4% paraformaldehyde in phosphate-buffered saline (PBS), permeabilized with 0.1% Triton X-100 for 10 min, blocked with 0.2% bovine serum albumin (BSA)/PBS for 1 h and incubated with specific primary antibodies including anti-mortalin (clone 088: raised in our laboratory) (3–5 µg/ml) and p53 (DO-1: SC-126) (3–5 µg/ml) (Santa Cruz Biotechnology, Santa Cruz, CA, USA) at 4°C overnight. The cells were then incubated with either Alexa-488 or Alexa-594 conjugated secondary antibodies (1 µg/ml) (A11034 or A11032, Molecular Probes, Eugene, OR, USA) for 1 h. Counterstaining was performed with Hoechst 33342 (Sigma, St. Louis, MO, USA) in the dark for 10 min. The cells were examined on a Zeiss Axiovert 200 M microscope and analyzed using AxioVision 4.6 software (Carl Zeiss, Oberkochen, Germany).

The cells were plated in 6-well plates and allowed to make monolayers following which a scratch was made using a 100-ml pipette tip. The cells were then washed with PBS and cultured with the control or test compounds [GPSE or GPSE-50%γCD or γCD, (0.1%) as indicated in Fig. 4C]. The movement of the cells to the scratch area was followed for the following 24–48 h. Images were acquired under a phase contrast microscope (Nikon) with a X10 phase objective lens at the 0, 24 and 48 h time-points. The movement was quantitated by measuring the empty surface area in the control and test samples by ImageJ software. The results are expressed as the means ± standard deviation of the mean (SEM). Statistical significance of the data was determined using a Student's t-test. Values of P<0.05, P<0.01 and P<0.001 were considered to indicate significant, very significant and highly significant differences, respectively.

The effects of GPSE and GPSE-50%γCD on in vivo tumor progression were investigated in a nude mouse subcutaneous tumor xenograft model. BALB/c nude mice (4 weeks old, female, 3 mice in each group/average weight 21.8 g) were obtained from Nihon Clea (Tokyo, Japan). The animals were allowed to acclimatize in our laboratory for 1 week. The cells were injected subcutaneously (1×10⁷ suspended in 0.2 ml of growth medium) into the abdomen of the nude mice and the tail vein (1×10⁶ suspended in 0.2 ml of growth medium). GPSE and GPSE 50%γCD were administered (by oral gavage) every alternate day beginning at 1 day after the injection of the cells for 3 weeks. Tumor formation and the body weight of the mice were monitored every alternate day. The volume of the subcutaneous tumors was calculated as V = L × W2/2, where 'L' was the length and 'W' was the width of the tumor, respectively. Statistical significance of the data was calculated from 3 independent experiments (n=3 per experiment). All procedures were carried out in accordance with the Animal Experiment and Ethics Committee, Safety and Environment Management Division, National Institute of Advanced Industrial Science and Technology (AIST), Tsukuba, Japan.

All the experiments were performed in triplicate and data are expressed as the means ± SEM. An unpaired t-test (GraphPad Prism GraphPad Software, San Diego, CA, USA) and one-way ANOVA followed by Tukey's honestly significant difference (HSD) post hoc test were used to determine the degree of significance between the control and experimental samples. Statistical significance was defined as P-values as follows: P<0.05 (significant), P<0.01 (very significant) and P<0.001 (highly significant) for the t-test and P<0.05 for ANOVA followed by a post hoc test, as indicated in the figure legends.

The interaction of mortalin and p53 occurs at N-terminal amino acids residues (253–282) of mortalin and C-terminal residues (323–337) of p53 (12,40–42). Based on this information, we docked ARC with mortalin or p53 or at mortalin-p53 complex interface. ARC interacted with the crystal structure of mortalin and p53 at the binding regions. It showed a docking score of −6.99 kcal/mol with mortalin. The docked complex exhibited multiple hydrogen bonds and hydrophobic interactions (Fig. 1A–a). The ligand formed 3 hydrogen bonds, 2 with Arg 126 and 1 with Thr 271 of mortalin, while residues Glu 132, Ala 195, Tyr 196, Thr 267, Asn 268 and Gly 269 played a role in hydrophobic interactions. The second oxygen atom of ARC formed 2 hydrogen bonds with nitrogen atoms (NH1 and NH2) of Arg 126. The third hydrogen bond was formed between the third oxygen atom of ARC and the third oxygen atom of Thr 271. Together the hydrogen bonds and hydrophobic interactions formed a stable protein-ligand complex.

The docking of ARC with p53 exhibited a score of −10.89 kcal/mol. The ligand was observed to form 1 hydrogen bond and various hydrophobic interactions (Fig. 1B–a). The nitrogen atom of Arg 333 was observed to form a hydrogen bond with the third oxygen atom of ARC. Residues Thr 329, Leu 330, Gln 331, Ile 332, Phe 338, Phe 341 and Arg 342 were involved in the hydrophobic interaction with p53. ARC docked in the cavity of p53 forming a stable complex.

The stability of the ARC-docked mortalin structure was further analyzed by simulating the complex for 10 nsec under conditions mimicking the bodily environment. The complex was observed to be stable for a period of 5 nsec from 5 to 10 nsec, as seen in the RMSD trajectory. Changes observed in the interacting residues of mortalin are shown in Fig. 1A–b. One hydrogen bond was observed between the oxygen atom of Arg 127 and the third oxygen atom of ARC, while residues Tyr 196, Ser 266, Thr 267 and Thr 271 were involved in hydrophobic interaction following simulations. Despite the changes in the interaction pattern of ligand-docked protein complex, the structure was more stable following simulations. The RMSD trajectory is shown Fig. 1A–c. A change in the interaction pattern of ARC docked p53 complex was observed (Fig. 1B–b). The number of hydrogen bonds between ARC and p53 increased to 2 post-simulations. One bond was formed between the second oxygen atom of ARC and nitrogen atom (NH1) of Arg 333, while the second was observed to be formed between nitrogen atom (NE) of Arg 337. The number of residues involved in hydrophobic interactions were the same, with 2 residues (Leu 330 and Ile 332) replaced with Phe 328 and Asn 345. The complex was more stable post-simulations. The ARC-docked p53 structure was observed to be stable from 5 to 10 nsec that can be seen in the RMSD trajectory (Fig. 1B–c). Bioinformatics analyses revealed that the ARC docked into the mortalin-p53 complexes (Fig. 1), similar to CAPE, and hence may cause abrogation of these complexes, and the re-activation of p53 function in cancer cells.

We then examined the cytotoxicity of ARC in HT1080 cells and found that it induced a considerable (>20%) decrease in cell viability at concentrations of 250 µM (IC₅₀, 275 µM) (Fig. 2A). Based on such high IC₅₀ values, we considered whether the extract of green propolis could be more effective than purified ARC for the reasons including that: i) it may provide a complex mixture of bioactives with multi-module action; and ii) it may protect the bioactive components from degradation. We prepared the GPSE and its conjugate with γCD (GPSE-γCD) and HPLC analyses revealed that they contained 9.6 and 3.0% ARC, respectively. The IC₅₀ values were 0.2–0.5% for GPSE and GPSE-γCD, whereas they were 250–300 µM for ARC (Fig. 2B). Similar results were obtained with the A549 and U2OS cells (data not shown). Cell morphological analysis revealed that the cytotoxicity induced by 0.5% GPSE and GPSE-γCD was comparable to that induced by 300 µM ARC (Fig. 2C). Long-term colony forming assays revealed that ARC (100 µM) induced a marked decrease in the number and size of the colonies; a more potent decrease was observed by treatment with 0.1% GPSE and GPSE-γCD (Fig. 2D).

We then examined the above-mentioned concentrations of ARC, GPSE and GPSE-γCD, and determined their effects on mortalin-p53 interactions as predicted by bioinformatics analysis (Fig. 1). In several independent experiments, we found that treatment with GPSE and ARC led to the nuclear trans-location of p53 (Fig. 3A), accompanied by an increase in p53 expression (Fig. 3B). We performed antitumor assays using a subcutaneous tumor xenograft model with nude mice. As shown in Fig. 4, mice fed the GPSE or GPSE-γCD exhibited no change in body weight (Fig. 4A) or physical activity (data not shown) with respect to the control group. A reduction in tumour growth was observed in the mice fed GPSE or GPSE-γCD as compared to the control group over the 3 weeks of treatment (Fig. 4B). The tumor growth data were strongly suggestive that the GPSE-γCD induced a more marked suppressive effect on tumor growth, endorsing its potential as an antitumor natural compound. We also performed in vitro wound healing assays to examine whether the GPSE and GPSE-γCD extracts possess anti-migratory activity and to determine whether they can be used to inhibit cancer metastasis. As shown in Fig. 4C and D, a considerable inhibition of cell migration was detected in the cells cultured in the presence of non-toxic doses of GPSE and GPSE-γCD.