Cell culture

A549 adenocarcinoma human alveolar basal epithelial cell line (ATCC, USA) was obtained as a gift from Prof. Alaina Ammit, Woolcock Institute of Medical Research, Sydney, Australia. The cells were grown in a RPMI media (Sigma-Aldrich, USA) containing 5% foetal bovine serum (Novogen, Australia), 1% penicillin and streptomycin (Gibco, New York) and were maintained at 37 °C in an incubator with 5% CO2. Cells were constantly checked for mycoplasma contamination, and all experiments were conducted in mycoplasma-negative cells.

Preparation of berberine-loaded liquid crystal nanoparticles (LCNPs)

Berberine-loaded liquid crystalline nanoparticles were formulated as previously reported, using the ultrasonication technique (Paudel 2022). Briefly, 5 mg berberine was completely dissolved in 200 mg melted MO at 70 °C. This berberine–MO mixture was then mixed with the surfactant solution (20 mg P407 dissolved in 4.8 mL water), obtaining a coarse dispersion. This dispersion was size-reduced using a probe sonicator (Labsonic® P, Sartorius, Germany), with the amplitude maintained at 80, and 5-s on and 5-s off-cycles for 5 min. This resulted in the production of 1 mg/mL berberine-loaded MO-LCNPs, with a lipid content of 40 mg/mL and a lipid:surfactant ratio of 1:10 w/w (Wadhwa et al. 2021).

Cytotoxicity (cell proliferation) assay of berberine LCNPs (BBR LNCPs) in A549 cells

A549 cells were plated in a 6-well plate at a density of 2 × 105 cells/well. Cells were treated with or without 5 μM of BBR LCNPs for 24 h. The dose of 5 μM of BBR LCNPs was decided from our previously published study showing significant anti-proliferative (MTT assay) and anti-migratory activity of berberine against the A549 cell line at that concentration (Paudel 2022).

RNA extraction and assessment of gene expression level by real-time PCR

RNA was extracted using the TRIzol® method. Extracted RNA was reverse transcribed to cDNA using a high-capacity DNA reverse transcription kit (Applied Biosystems, RI, USA). Real-time PCR analysis was performed using TaqMan primers and probes (Applied Biosystems) to measure the relative expression of genes CTTNB1 (forward primer: CTTGGAATGAGACTGCTG; reverse primer: AGAGTGAAAAGAACGATAGC), Axin1 (forward primer: CCGACCTTANATGAAGATGAG; reverse primer: CAGGATCCATACCTGAACTC). The quantitative expression of these genes was calculated through the 2∧-Ct method and was compared to the respective reference gene (GAPDH). The relative transcript abundance in the treated groups was calculated by comparing the treated groups with untreated controls (Wadhwa et al. 2021).

Immunofluorescence (IF)

Coverslips were sterilised and coated with placenta collagen (Sigma cat. no. C5533) for 30 min and washed twice with 1 ml 1 × PBS to allow cell adhesion. Then, 2 × 105 A549 cells were seeded on the cover slip and placed in a 6-well plate. The following day, cells were treated with 5 μM BBR LCNPs for 24 h. The media was then removed, and cells were washed with sterile 1 × PBS. Cells were fixed with ice-cold methanol for 20 min, followed by blocking with 2% bovine serum albumin (BSA) for 1 h at room temperature. The cells were then incubated with anti-beta catenin (ab32572) antibody at 1:250 dilution overnight at 4 °C, followed by incubation with goat anti-rabbit IgG H&L Alexa Fluor® 647 (ab150083) for 1 h at room temperature (1:500 dilution), protected from light. Finally, the slides were cover slipped with DAPI fluoromount G (ProSciTech). The microscopic images at random fields were then captured with a fluorescence microscope (Zeiss microscope, Germany) at 40 × magnification (Paudel and Kim 2020).

Protein extraction and quantification

RIPA buffer (Sigma-Aldrich) containing phosphatase and protease inhibitor cocktail (Roche, USA) was added to the 6-well plates containing treated and untreated cells, and the lysate was collected in a microcentrifuge tube. Proteins were extracted from lysed cells upon vortex treatment and incubation on ice for 15 min. Soon after, the mixture was sonicated at 30% amplitude, 3 times for 2 s. The mixture was then incubated again on ice for an additional 15 min. The lysate was then centrifuged for 30 min at 18,000 g at 4 °C. Finally, the supernatant was collected, and the proteins were quantified using a PierceTM BCA Protein Array Kit (Thermo Scientific) following the manufacturer’s instructions (Wadhwa et al. 2021).

Immunoblotting

Equal amounts of the extracted proteins were loaded on SDS-PAGE gels and transferred to a PVDF membrane. The membrane was blocked with 5% BSA and incubated with primary rabbit anti-beta catenin antibody (ab 32,572) at a 1:7500 dilution overnight at 4 °C, followed by incubation with goat anti-rabbit IgG at a 1:10,000 dilution for 1 h. Images were acquired using a chemiDoc imaging system. Then, the membrane was stripped and blocked again with 5% BSA. It was then incubated with mouse anti-beta actin antibody (ab8226) at 1:10,000 dilution at room temperature for 3 h followed. After this, the membrane was incubated with anti-mouse IgG at 1:10,000 dilution, and images were again acquired using a chemiDoc imaging system (whole blots were added in the supplementary file). The densitometric analysis of proteins was performed by quantifying pixel density using ImageJ software.

Statistical analysis

The values are represented as mean ± SEM. GraphPad Prism (version 9.3) was used to perform statistical analyses. Statistical comparisons were done by unpaired, two-tailed student’s t-test. A value of p < 0.05 was considered statistically significant.

Molecular docking study

For the molecular docking study, the three-dimensional structure of human beta catenin (PDB ID: 1JDH) and berberine (PubChem CID 2353) was retrieved from the Protein Data Bank (PDB) and PubChem, respectively (Graham et al. 2001; Berman et al. 2000; Kim et al. 2016). Before molecular docking, the protein structure was prepared by adding hydrogens, optimising geometry and minimising energy with the protein preparation wizard of Schrodinger (Sastry et al. 2013). Similarly, the ligand geometry was optimised with the LigPrep module of Schrodinger (Sastry et al. 2013). The molecular docking study was performed using Vina wizard of PyRx software (Dallakyan and Olson 2015). The grid was generated around the β-catenin residues Phe253, His260, Asn261, Leu264, Asn290, Lys292, Phe293, Ile296, Asp299, Gln302, Tyr306, Gly307, Lys312, Lys335, Arg342, Lys345, Val346, Arg376, Arg386, Asn387, Asp390, Gly422, Ser425, Asn426, Cys429, Asn430, Lys435, Arg469, His470, Ser473, Arg474, Lys508, Gly512, Asn516 and Leu519. These residues were selected because they were reported to be involved in the interaction between β-catenin and human TCF4, in a study in which the β-catenin–hTcf4 complex was characterised through X-Ray crystallography (Graham et al. 2001). The coordinates of X, Y and Z were set to 73.8, − 19.75 and − 15.46, respectively. The exhaustiveness of the algorithm is set to 8. After molecular docking, the top docking pose of the ligand was selected based on the docking score and subjected to molecular dynamics simulation.

Molecular dynamics study

The molecular dynamics (MD) simulation was run using the Desmond software (Bowers, et al. 2006). For MD simulation, the system was prepared by adding water and neutralised by the addition of sodium and chloride ions. The berberine–β-catenin complex was hydrated in an orthorhombic box using a three-point water model (TIP3P) (Bowers, et al. 2006). The implicit solvent system was used to represent solvent. After that, the system was energy minimised. Finally, the minimised system was subjected to 50 ns of MD simulations in isothermal–isobaric ensemble (NPT) at high temperature. Fifty ns of MD simulations were deemed sufficient for the MD simulation, considering the high-temperature settings, the use of the implicit solvent system and the fact that the backbone chains of carbon atoms showed stability throughout the 50 ns. The final trajectory was analysed with manual inputs and root mean square deviation (RMSD), protein–ligand contact histogram, protein–ligand contact diagram and root mean square fluctuation (RMSF) calculated using the simulation interaction diagram utility of maestro.

MMGBSA energy calculation

After MD simulation, to estimate the binding affinity of the berberine–β-catenin complex, the binding-free energy (dG_Bind) was estimated using the thermal_mmgbsa.py script of Schrodinger (Genheden and Ryde 2015). The calculation was performed on 100 complexes obtained by extracting every 10th frame from the last 10 ns of the stable trajectory. After MMGBSA calculation, an average binding-free energy was reported.