The butein structure was downloaded from the PubChem database (33) in the 2D format and was subsequently exported on to the Discovery Studio v18 (hereinafter referred to as DS). The compound was minimized by employing the Full Minimization protocol accessible with DS.

Using the Discovery Studio, the B ring of butein was modified to obtain new structures as described in Figure 1 , and three compounds were designed and synthesized.

To elucidate on their binding affinity toward the target DDX3, the molecular docking was performed using the CDOCKER program (34–37) available on DS. The CDOCKER utilized a CHARMm-based molecular dynamics (MD) strategy to dock ligands into a receptor-binding site. Correspondingly, Random Ligand Conformations are generated using high-temperature MD. These conformations are then translated into the binding site. The candidate poses are subsequently created using random rigid-body rotations followed by simulated annealing. A final minimization is then used to refine the ligand poses (34, 38).

The protein for the current study is the human DEAD-box RNA helicase DDX3X bearing the PDB code 2I4I cocrystallized with the AMP (39). The protein was initially prepared by initiating the Clean Protein tool available with the DS. Correspondingly, the heteroatoms and the water molecules were dislodged and were minimized using the Minimization tool available under the Minimize and Refine Protein protocol. The active site was marked for all the atoms and the residues around the cocrystallized AMP (39).

The synthesized butein-like compounds were docked into the active site of the protein permitting the generation of 50 conformers for each compound along with the butein compound. The best pose was selected from the largest cluster demonstrating a higher dock score, read according to the -CDOCKER interaction energy and the key residue interactions.

To a solution of 2′,4′-dihydroxyacetophenone (0.76 g, 5.0 mmol) in dry CH₂Cl₂ (5 mL) triethylamine (2.4 mL, 17.5 mmol) and 4-dimethylamino pyridine (0.06g, 0.5 mmol) were added, and then tertbutyldimethylsilyl chloride (2.11g, 14.0 mmol) dissolved in CH₂Cl₂ (9 mL) was added dropwise. After the completion of addition, the reaction mixture was allowed to stirred overnight at room temperature. The reaction mixture was quenched by the addition of water. The organic fractions were collected, and the aqueous phase was extracted with CH₂Cl₂. The combined organic fractions were washed with water and brine, dried over anhydrous MgSO₄, and concentrated under reduced pressure. The residue was purified by short column chromatography on silica gel using n-hexanes as the eluent to afford compound 2 (1.56 g, 82% yield) as a colorless oil; ¹H NMR (300 MHz, Chloroform-d) δ 7.62 (d, J = 8.6 Hz, 1H), 6.47 (dd, J = 8.6, 2.3 Hz, 1H), 6.32 (d, J = 2.3 Hz, 1H), 2.57 (s, 3H), 1.00 (s, 9H), 0.97 (s, 9H), 0.28 (s, 6H), 0.21 (s, 6H).

To a solution of compound 2 (1.0 mmol) in dry THF (3 ml) under argon was added lithium diisopropylamide (LDA, 1.0 M, 1.1 ml, 1.1 mmol) dropwise at −78°C. The reaction mixture was allowed to warm to −20°C over 30 min, and then cooled to −78°C. The solution of substituted aldehyde (1.0 mmol) in dry THF (3 ml) was added dropwise at −78°C, and the resulting reaction mixture was allowed to warm to room temperature stirred overnight and then quenched by addition of sat. aqueous NH₄Cl. The organic fraction was collected, and the aqueous layer was extracted with diethyl ether. The combined organic fractions were washed with brine, dried over anhydrous MgSO₄, and evaporated under reduced pressure to afford the monohydroxy protected chalcones. The monohydroxy protected chalcones were purified by using CH₂Cl₂/MeOH as eluent.

The monohydroxy protected chalcone derivatives (0.2 mmol) were dissolved in THF (5 ml), and tetrabutylammonium fluoride (TBAF, 1.0 M, 0.4 mmol) was added. The resulting reaction mixture was stirred at room temperature for 30 min. and quenched by addition of water. The organic fraction was collected, and the aqueous layer was extracted with ethyl acetate three times. The combined organic fractions were washed with brine, dried over anhydrous MgSO₄, and evaporated under reduced pressure. The residues were purified column chromatography on silica gel using CH₂Cl₂/MeOH as eluent to afford compounds 3a-3c in 70% to 85% yields.

The compound (3a) was synthesized by following the general procedure using monohydroxy-protected chalcone (0.10 g, 0.23 mmol) dissolved in THF (8 mL) and tetrabutylammonium fluoride (TBAF, 1.0 M, 0.46 mL, 0.46 mmol). The 3a was purified using silica gel column CH₂Cl₂/MeOH 10: 1, v/v. Rf. Value = 0.5. Yellow solid (63 mg, 85%); ¹H NMR (300 MHz, Chloroform-d) δ 13.28 (s, 1H), 7.87 (dd, J = 15.1, 0.6 Hz, 1H), 7.75 (d, J = 8.5 Hz, 1H), 7.23 (d, J = 15.1 Hz, 1H), 7.12 (d, J = 3.9 Hz, 1H), 7.07 (d, J = 3.9 Hz, 1H), 6.46–6.40 (m, 2H), 5.53 (s, 1H).

The compound (3b) was synthesized by following general procedure using monohydroxy-protected chalcone (0.10 g, 0.24 mmol) dissolved in THF (8 mL) and tetrabutylammonium fluoride (TBAF, 1.0 M, 0.48 mL, 0.48 mmol). The 3b was purified using silica gel column CH₂Cl₂/MeOH 10: 1, v/v. Rf. Value =0.5. Yellow solid (52 mg, 70%); ¹H NMR (300 MHz, Chloroform-d) δ 13.34 (s, 1H), 7.84 (d, J = 8.5 Hz, 1H), 7.59–7.38 (m, 2H), 6.67 (d, J = 3.5 Hz, 1H), 6.48–6.40 (m, 3H), 5.30 (s, 1H).

The compound (3c) was synthesized by following general procedure using monohydroxy protected chalcone (0.03 g, 0.08 mmol), was dissolved in THF (2 mL) and tetrabutylammonium fluoride (TBAF, 1.0 M, 0.16 mL, 0.16 mmol). The 3c was purified using silica gel column CH₂Cl₂/MeOH 10: 1, v/v. Rf. Value = 0.5. Yellow solid (16 mg, 77%) (40); ¹H NMR (300 MHz, Chloroform-d) δ 13.36 (s, 1H), 7.90–7.83 (m, 2H), 7.56 (d, J = 15.5 Hz, 1H), 7.46 (d, J = 6.9 Hz, 2H), 7.23–7.35 (m, 2H), 6.47–6.41 (m, 2H), 5.47 (s, 1H), 2.41 (s, 3H); CAS. No. 1385668-91-6.

Human MCF-7 and MDA-MB-231 cell lines were purchased from the Korean Cell Line Bank (KCLB, Seoul, Korea). The MCF-7 and MDA-MB-231 cell lines were maintained in RPMI-1640 medium (Gibco, Life Technologies, Carlsbad, CA, USA) containing 10% (v/v) fetal bovine serum (FBS, Gibco) and 1% penicillin–streptomycin (Gibco) at 37°C in a humidified atmosphere of 5% CO2 (41).

The cell viability assay was conducted as described before (42) the cells were seeded in 48 well, plated at a density of 5 × 10⁴ cells in 500 µL medium per well for 16 h. The cultured cells were treated with various concentration for 24, 48, and 72 h. After incubation, the cells were added with 55 µl of 5 mg/ml 3-(4, 5-dimethylthiazol-2-yl)-2, 5-diphenyltetrazolium bromide (MTT; Duchefa Biochemie, Haarlem, The Netherlands) solution for 2 h, and the medium was removed, which was followed by lysis with DMSO. The absorbance at 570 nm was measured with PowerWave HT microplate spectrophotometer (BioTek, Winooski, VT, USA).

The Western blot analysis was performed as described before (43). After 48 h treatment of each compounds, the cells were lysed with RIPA buffer (50 mM Tris-HCl pH 7.5, 0.1% SDS, 1% Triton X-100, 150 mM NaCl, 0.5% Sodium deoxycholate, and 2 mM EDTA) containing protease/phosphatase inhibitor cocktail (Thermo Fisher Scientific, Waltham, MA, USA) at 4°C for 1 h. The supernatants were collected and quantified using BCA protein assay kit (Thermo Fisher Scientific) according to the manufacturer’s instructions. To identify the molecular weight, we used Regular Range Protein marker (PM2510; SMOBIO Technology. Inc., Taiwan) and precisionplus protein dual color standard marker (Bio-Rad catalog no 1610374). Protein of 10–20 µg concentration was loaded on 8–12% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE), and then transferred to polyvinylidene difluoride membrane (PVDF, ATTO, Tokyo, Japan). The membranes were blocked with TBS-T buffer (Tris-buffered saline containing 0.1% Tween 20) containing 5% (w/v) skim milk power for 1 h at 25°C, and then incubated with primary antibodies for 16 h at 4°C. After incubation with the secondary antibody for 3 h at 25°C, the membranes was detected using Clarity™ Western ECL Blotting Substrates (Bio-Rad, Hercules, CA, USA).

The data were expressed as mean ± standard error of mean (SEM) and analyzed by GraphPad Prism Version 4.0b (GraphPad Software Inc., La Jolla, CA, USA), for statistical significance using one-way analysis of variance (ANOVA). p < 0.05 was considered as statistically significant. All experiments were performed in triplicates.

The prepared butein compound was used as the starting structure (parent structure) to computationally design new compounds. Subsequently, modifying the ring B has yielded three compounds as shown in Figure 1 .

The synthesized compounds were assessed for the binding affinity studies using the CDOCKER program. These compounds were docked into the ATP binding pocket of the DDX3 to understand their potentiality along with butein for comparative study. The results have shown that the modified compounds have demonstrated a better to comparable dock score than the parent compound butein as shown in Table 1 .

Further, the compounds have shown interactions with the key residues positioning at the binding pocket as elucidated in Figure 2 . The compound butein has formed hydrogen bond with the residue Arg202, as displayed in Figure 3A . Additionally, the key residue Tyr200 has formed the π-π stacked interaction. The 3a has prompted hydrogen bond interactions with the residues, Thr201 and Gln207, as shown in Figure 3B , along with the π-π stacked interaction by Tyr200. The compound 3b has interacted via hydrogen bonds with the residues, Thr231, Ala232, Gln281, and Glu285, respectively, as represented in Figure 3C . The key residue Tyr200 has participated by π-alkyl interaction holding the compound at the binding pocket. The compound 3c has formed hydrogen bond interactions with the residues, Thr231, Ala232, and Gln281 residues. The residue Thr231 has formed two hydrogen bonds with the ligand as illustrated in Figure 3D . The key residue Tyr200 has demonstrated the π-π stacked interaction. Several other residues have clamped the ligands at the binding pocket as shown in Supplementary 1A–C respectively.

The syntheses of chalcones are illustrated in Scheme 1 . All the chalcones were synthesized by aldol condensation reaction using an appropriate aldehyde and ketone as previously described in the literature method with modification. The dihydroxy ketone was protected by using TBSCl. Dihydroxy protected compound by reacting with aldehyde in the presence of LDA base resulted in the condensation product, which on desilylation by TBAF, the desired chalcones were obtained. The ¹H NMR spectrum data of the three compounds is provided as, Supplementary Figures 2–4 for 3a, 3b, and 3c, respectively.

To evaluate the anti-proliferative effect of 3b and 3c, MTT assays were carried out on the MDA-MB-231 and MCF-7 representing the human breast cancer cell lines. 3b and 3c were treated with various concentration for 24, 48, and 72 h in the MDA-MB-231 and MCF-7, respectively. The cell viability on the MDA-MB-231 and MCF-7 cells was markedly decreased by 3b and 3c treatment in a dose- and time-dependent manner as shown in Figures 4A, B . The IC₅₀ values of 3b were recorded as 58.23 μM and 37.74 μM at 48 h in MCF-7 and MDA-MB-231 cell line, respectively. Also, IC₅₀ values of 3c were shown as 22.72 μM and 20.51 μM at 48 h in MCF-7 and MDA-MB-231 cell line, respectively. These results showed that 3b and 3c have anti-proliferative effect in both human cell lines. Therefore, we used 20 and 30 μM of 3a, 25 and 50 μM of 3b, and 10 and 20 μM of 3c in our subsequent experiments.

Based on cell viability results, we also identified the DDX3 protein expression. The two cell lines, which include MCF-7 and MDA-MB-231 cells were treated with indicated concentration of 3a, 3b, or 3c, respectively. The DDX3 protein expression was decreased dose-dependently in 3b or 3c treated MCF-7 and MDA-MB-231 cell lines as shown in Figures 5A, B , and Supplementary Figure 5 . This data indicated that 3b and 3c could suppress DDX3 protein expression in MCF-7 and MDA-MB-231 cell lines. Since 3a did not show any significant reduction of DDX3 expression, as shown in Supplementary Figure 7 , we have not proceeded with the compound further.

To determine the cell death and cell cycle arrest effect on 3b and 3c in human breast cancer cell lines, we identified the apoptosis and cell cycle related protein expression in 3b and 3c treated MCF-7 and MDA-MB-231 cell lines. The apoptosis marker proteins, such as cleaved PARP and cleaved caspase 3 were increased, and anti-apoptosis protein, BCL-xL, was decreased in compound use to treat both cell lines, as shown Figures 6A, B , and Supplementary Figure 6 . In addition, G2/M cell cycle arrest-related protein including CDK1 and cyclinB1, were inhibited in compound treated both cell lines. Those data suggest that 3b and 3c can induce apoptotic cell death and G2/M phase cell cycle arrest in MCF-7 and MDA-MB-231 cell lines, indicating anti-cancer effect, as shown in Figures 6A, B .

In addition, PI3K/AKT signaling pathway is an important intercellular signal involving apoptotic cell death and cell cycle arrest. Therefore, we examined the phosphorylated PI3K and AKT protein expression in the 3b and 3c treated human breast cancer cells. 3b was not affective to PI3K/AKT protein change. The results showed that inhibition of PI3K and AKT activation was induced in the 3c treated both cell lines, as shown in Figure 6B .