Cucurbitacin-B inhibits cancer cell migration by targeting mortalin and HDM2: computational and in vitro experimental evidence

Abstract Cancer metastasis, a highly complex process wherein cancer cells move from the primary site to other sites in the body, is a major hurdle in its therapeutics. A large array of synthetic chemotherapeutic molecules used for the treatment of metastatic cancers, besides being extremely expensive and unaffordable, are known to cause severe adverse effects leading to poor quality of life (QOL) of the patients. In this premise, natural compounds (considered safe, easily available and economic) that possess the potential to inhibit migration of cancer cells are deemed useful and hence are on demand. Cucurbitacin-B (19-(10→9β)-abeo-10-lanost-5-ene triterpene, called Cuc-B) is a steroid mostly found in plants of Cucurbitaceae family. It has been shown to possess anticancer activity although the molecular mechanism remains poorly defined. We present evidence that Cuc-B has the ability to interact with mortalin and HDM2 proteins that are enriched in cancer cells, suppress wild type p53 function and promote cancer cell migration. Computational analyses showed that Cuc-B interacts with mortalin similar to MKT077 and Withanone, both have been shown to reactivate p53 function and inhibit cell migration. Furthermore, Cuc-B interacted with HDM2 similar to Y30, a well-known inhibitor of HDM2. Experimental cell and molecular analyses demonstrated the downregulation of several proteins, critically involved in cell migration in Cuc-B (low non-toxic doses)-treated cancer cells and exhibited inhibition of cell migration. The data suggested that Cuc-B is a potential natural drug that warrants further mechanistic and clinical studies for its use in the management of metastatic cancers. Communicated by Ramaswamy H. Sarma


Introduction
Cancer, a syndrome of diseases resulting from uncontrolled cell proliferation, is rapidly increasing in incidence and is expected to rise from the current 18 million new cases/year to �30 million by 2030 (Bray et al., 2018).Although there have been tremendous advances in cancer diagnosis, drug discovery and therapeutics in recent decades, cancer incidence and mortality remain alarming.Among the three treatment options, i.e., surgery, radiation and chemotherapy, the latter is extremely expensive, not readily available and faces the challenges of adverse side effects and drug resistance.In addition, cancer metastasis poses additional hurdles as cells migrate from the primary site to secondary sites and cannot be treated with local surgery and radiation.Cancer metastasis is a highly complex and multi-step process in which cancer cells detach from the primary tumour site, migrate through the bloodstream and colonise a distant site in the body.The process requires multiple factors that determine the ability of cancer cells to acquire (i) migratory properties to leave the primary site, (ii) the ability to proteolytically digest surrounding connective tissue, (iii) properties to enter lymphatic or blood vessels to travel to a distant site in the body, and (iv) the ability to proliferate at the new site to establish a tumour (Dhaliwal & Shepherd, 2022;Zhao et al., 2022).Epithelial to mesenchymal transition (EMT) is a key regulator of cancer metastasis.It involves the transformation of epithelial cells (tightly associated and with apico-basal polarity) into mesenchymal cells (spindle-shaped and lacking polarisation).While epithelial cells express E-cadherin protein, it is downregulated in mesenchymal cells, which are characterised by the expression of N-cadherin, fibronectin and vimentin and are considered reliable biomarkers of EMT (Fedele et al., 2022;Rubtsova et al., 2022).As current treatment options for metastatic cancer are very limited and often encounter drug resistance, the identification of natural compounds is warranted.
Cucurbitacin-B (19-(10!9b)-abeo-10-lanost-5-ene triterpene) (referred to as Cuc-B), a naturally occurring bitter-tasting steroidal bioactive, commonly found in the plants of the Cucurbitaceae family, has been recognised as having potent anticancer properties (Garg et al., 2018).It has been shown to modulate the Wnt/b-catenin, p53, NFjB, PI3K/AKT, RAS/RAF/MAPK and JAK/STAT signalling pathways, resulting in the suppression of cancer cell proliferation (Garg et al., 2020;Luo et al., 2018;Zeng et al., 2021).However, their toxicity to normal cells and tissues has been a concern (Garg et al., 2018;Shukla et al., 2016).To overcome this toxicity, polymeric micelles, cocktails and structural analogues have been tested (Thoennissen et al., 2009;Zheng et al., 2014).We have previously reported that a combination of Cuc-B and withanone (Wi-N) selectively kills cancer cells.It was shown to induce senescence in lung cancer cells and was characterised by an increase in p53 and CARF proteins and a decrease in lamin A/C, CDK2, CDK4, cyclin D, cyclin E, phosphorylated RB.A decrease in cancer cell migration was also observed, mediated by a decrease in mortalin, hnRNP-K, VEGF, MMP2 and fibronectin (Shukla et al., 2016).Cuc-B has been shown to interact with mortalin, a member of the heat shock protein 70 family, which is involved in a variety of intracellular mechanisms such as protein chaperoning and folding, intracellular trafficking, ATP synthesis, mitochondrial genesis, stress protection, anti-apoptotic, pro-proliferative and cell survival (Tang et al., 2018;Wang et al., 2020).Overexpression of mortalin has been frequently detected in a variety of cancers (Wadhwa et al., 2002) and has been shown to (i) inhibit p53-mediated tumour suppression, (ii) activate epithelial-to-mesenchymal transformation and (iii) contribute to cancer cell stemness (Wadhwa et al., 2002;Wang et al., 2020).The wild-type p53 tumour suppressor protein phosphorylates, translocates from the cytoplasm to the nucleus and transcriptionally activates its downstream proteins (p21, NOXA and Bax), leading to senescence/apoptosis (Na et al., 2016;Yoon et al., 2022).Mortalin interacts with p53 and sequesters it in the cytoplasm to inhibit its transcriptional activation function (Capuozzo et al., 2022;Thomas et al., 2022).Phosphorylation of p53 on serine and threonine residues has been shown to play a key role in its apoptotic function (Lu et al., 2011a;Sung et al., 2018).Such phosphorylation is mediated by kinases (HIPK2 and DYRK2) that are degraded by HDM2, an E3 ubiquitin ligase (often overexpressed in cancer cells) that also causes proteasomal degradation of wild-type p53 (Patil & Bihari, 2022;Rinaldo et al., 2007;Smeenk et al., 2011).Thus, HDM2 antagonists lead to upregulation of p53 as well as its phosphorylation and are considered as cancer therapeutic molecules.Given the anticancer activity of Cuc-B, we investigated its effect on proteins involved in cancer cell migration.Cells treated with low non-toxic doses of Cuc-B showed downregulation of these proteins and inhibition of cell migration, invasion and clustering properties.We report computational and experimental analyses suggesting that Cuc-B is a potential natural drug that warrants further mechanistic and clinical studies for its use in the treatment of metastatic cancers.

Computational experiments
The 3D structures for mortalin, p53 and HDM2 were downloaded from the Protein Data Bank (PDB) [ID: 4KBO,4MZR and 1YCR,respectively].The nucleotide-binding domain of mortalin consisted of the p53 interaction site from residues 253 to 282, whereas the mortalin-binding domain on p53, present in the tetramerisation domain, consisted of residues 323 to 337.The tetramerisation domain of p53 bound to HDM2 (human MDM2) showed interactions at a site comprising residues Thr26, Leu54, Gly58, Met62 and Glu72.Using the protein preparation wizard of Maestro, pre-processing such as the addition of missing disulfide bonds, conversion of selenomethionines to methionines, deletion of all water molecules and missing hydrogen atoms were added, terminals were capped and missing side chains and loops were checked (none found) (Bowers et al., 2006;Friesner et al., 2006;Madhavi Sastry et al., 2013;Schr€ odinger, 2020).The SDF structures of Cuc (CID 5281316), MKT077 (CID 6912334; a known inhibitor of mortalin) and Y30 (CID 137350181; a known inhibitor of HDM2) were downloaded from the PubChem database.Ligands were prepared using the LigPrep module of the Schrodinger Suite (Schr€ odinger, 2020) for optimisation of bond order, generation of ionisation states and energy minimisation.The lattices were generated at the respective interaction domains in each protein with a length of 20 Å and the ligands were docked using the extraprecision glide module of the Schrodinger suite (Harder et al., 2016).The best docked complexes and the apoproteins were further subjected to MD simulation using the TIP4P water model as the system.The Desmond package in the Schrodinger suite's maestro was used to study the stability of the protein-ligand system (Madhavi Sastry et al., 2013).Firstly, the systems were built using Desmond's system builder with the OPLS3e force field; for solvation a predefined TIP4P water model was selected.In the boundary conditions option, an orthorhombic periodic boundary was set to give the shape and size of the box buffered at 10 Å, and then ions were added to each system to balance the charge.After the construction of the solvated protein-ligand complex systems, the energy of the prepared systems was minimised by performing a 100 ps low-temperature (10 K) Brownian motion MD simulation (NVT ensemble) to remove steric hindrance (Friesner et al., 2006;Grover et al., 2012).Furthermore, the minimised systems were equilibrated in seven steps in NVT and NPT ensembles using the 'relax model system before simulation' option in the Desmond Schrodinger suite.Finally, molecular dynamics simulations were performed with the periodic boundary condition in the NPT ensemble.The pressure and temperature of the systems were maintained at 1 atmospheric pressure (using the Martyna-Tobias-Kelin barostat) and 300 K temperature (using the Nose-Hoover chain thermostat), respectively.The production run of 200 ns was performed with the configuration saved at every 200 ps interval.The MD trajectories were analysed after molecular dynamics simulations using the Desmond Simulation event analysis tool (Madhavi Sastry et al., 2013).The Root Mean Square Deviation (RMSD) of the protein-ligand complex over the entire simulation trajectory with respect to its docked structure was obtained from the simulation event analysis module and compared with the known inhibitor.The number of hydrogen bonds between the ligands and the protein throughout the simulation time was calculated using default parameters.The radius of gyration, solvent accessible surface area and RMSD of the ligands were also calculated and compared to investigate their flexibility, binding within the active pocket of the protein and stability throughout the simulation (Schr€ odinger, 2020).From the trajectories, a total of 100 structures were extracted from the duration of 0-200 ns of the simulation time.To observe the structural changes due to the binding of small molecules, we superimposed the average structure of the simulated apoprotein with the protein-ligand complex.The average structure obtained from these 100 extracted structure complexes was used to calculate the MM/GBSA free binding energy using the 'prime MM-GBSA' of the Maestro Schrodinger suite (Friesner et al., 2006;Harder et al., 2016).

Cell viability assays
Cell viability was determined by the MTT (3-(4, 5-dimethylthiazol-2-yl)-2, 5-diphenyltetrazolium bromide)-based viability assay as described previously.Control and treated cells were incubated with MTT at 37 � C for 4 h, followed by the addition of 100 ml of dimethyl sulfoxide (DMSO) (Fujifilm WAKO, Tokyo, Japan) and shaking for 30 min to dissolve the chromogen.The optical absorbance of the chromogen was measured at 570 nm using a microplate reader (Infinite M200 PRO; TECAN, M€ annedorf, Switzerland).Cell viability was determined as a percentage of the control group.To investigate the long-term cytotoxic effect of Cuc-B, the colony formation assay was performed.Five hundred U2OS cells per well were seeded in 6-well plates and allowed to adhere overnight, followed by treatment with Cuc-B with regular medium changes every 3rd day for 15 days.At the end of the experiment, the colonies were fixed with methanol:acetone (1:1) for 5 min, the plates were air dried overnight and the colonies were counted manually.The experiment was repeated three times to determine statistical significance.

Western blotting
Cells (�2 � 10 5 /well) were seeded in 6-well plates, allowed to adhere overnight and then treated with CuC-B.Control and treated cells were lysed with RIPA buffer (Wako, Japan) containing protease inhibitors (Roche Applied Sciences, Mannheim, Germany).Protein concentration was determined using the Pierce BCA protein assay kit (Thermo Fisher Scientific, USA).Normalised proteins were separated by SDSpolyacrylamide gel electrophoresis (SDS-PAGE) followed by transfer to a polyvinylidene difluoride (PVDF) membrane (Millipore, Billerica, MA, USA).The membrane was blocked in 3% bovine serum albumin (BSA) in 0.2% Tween 20 in TBS (TBST) for 1 h, followed by incubation with specific primary antibodies at 4 � C overnight.The membranes were washed with TBST and incubated with secondary antibodies (antirabbit IgG (31460) or anti-mouse IgG (31430); Thermo Fisher Scientific, Waltham, MA, USA) for 1 h at room temperature.b-Actin (detected by antibody 643807; BioLegend, Japan) was used as an internal loading control.Protein bands were detected using Gel Doc Documentation (Bio-Rad, CA, USA) and quantification of protein expression was analysed using ImageJ software (NIH, Bethesda, MD, USA).

Immunostaining
Cells were plated on 18-mm glass coverslips in 12-well plates and incubated overnight to allow adherence.The cells were then treated with Cuc-B and fixed with pre-cooled methanol/acetone (1/1: v/v) at 4 � C for 10 min.The fixed cells were washed three times with PBS and then permeabilized by incubation with PBS-Triton X-100 (PBST, 0.1%) for 10 min.The reaction was blocked with 2% BSA dissolved in PBST at room temperature for 1 h.The coverslips were incubated with specific primary antibodies at 4 � C overnight, washed three times with PBST (10 min each time) and then incubated with secondary antibodies (conjugated with either Alexa-488 or Alexa-594; Molecular Probes, Eugene, OR, USA) for 1 h.The coverslips were washed three times with PBST (10 min each time).The coverslips were extensively washed with PBST and counterstained with Hoechst 33342 (1 lg/mL) (Thermo Fisher Scientific, USA).The coverslips were mounted on glass slides and visualised under a Carl Zeiss microscope (Axiovert 200 M).Quantification of protein expression was performed using ImageJ software.

Statistical analysis
All quantifications were performed using ImageJ software (NIH, Bethesda, Maryland, USA); calculations were performed using Microsoft Office and presented as percentages.Statistical significance was calculated by two-sided unpaired t-test in Microsoft Excel (2016) using the mean, SD and N from at least three independent experiments, and presented as nsp � 0.05, � p < 0.05, �� p < 0.01, ��� p < 0.001.

Cucurbitacin-B has the potential to target both mortalin and HDM2
The small molecules MKT077 and withanone (Wi-N) have previously been shown to bind to the p53 binding pocket of mortalin (Grover et al., 2012).We have previously shown that Cuc-B can also bind to the p53 binding region of mortalin.However, the interaction between Cuc-B and p53 did not involve the mortalin-binding pocket of p53 (Garg et al., 2020).In the present study, we extended the analysis to further investigate the ability of Cuc-B to disrupt the p53-mortalin interaction, together with the inhibition of HDM2, which has not been previously reported.By extending the simulation parameters to 200 ns, we first found that Cuc-B was able to dock to the mortalin binding site of p53.Molecular docking analyses at these parameters confirmed the interaction of Cuc-B with mortalin and showed a docking score of À 5.54 kcal/mol, which was stronger than the docking score (À 1.99 kcal/mol) of MKT077 (inhibitor of mortalin) to mortalin.Although both ligands interacted with the same region (p53 binding domain of mortalin) (Figure 1A, B), there was only a small deviation in the RMSD curve of Cuc-B:mortalin compared to MKT077: mortalin (Figure 1C).This could be attributed to the movement of the long tail of Cuc-B, evident from the 10 ns frames extracted from 100 to 200 ns (Figure 2A), and the RMSD of Cuc-B alone over the MD simulation trajectory with respect to its initial docked conformation (Figure 2B).Both Cuc-B and MKT077 showed binding in the same cavity of mortalin, involving interactions with the key residues also involved in its binding to p53 (Asn221, Ser266, Thr267 and Asp268) (Figure 2C and Supplementary Figure 1).Furthermore, Cuc-B also showed significant docking (dock score ¼ À 4.60 kcal/mol) into the mortalin-binding region of p53 (amino acid residues 323 to 337).As shown in Figure 1D-F, Tyr327 and Phe328 (involved in mortalin binding) and Arg344 and Asn347 (close to the mortalin binding region) showed interaction throughout the simulation period, suggesting stable docking, which was also supported by a stable RMSD.The MM/GBSA binding energy between Cuc-B and mortalin was À 55.52 kcal/mol, which was similar to the binding energy of MKT077 of À 52.22 kcal/mol (Table 1).Furthermore, the interaction fraction diagram shows the hydrogen bonding of residues Asn221 and Thr267 for a maximum fraction of time during the simulation (Figure 1F and Supplementary Figure 1).These data suggest that Cuc-B may be a potent inhibitor of mortalin-p53 interactions, leading to reactivation of its wild-type transcriptional activation function.
We next examined the interaction of Cuc-B with HDM2 (a p53 antagonist that negatively regulates p53 function by promoting its degradation).The p53 binding domain of HDM2 consists of residues Thr26, Leu54, Gly58, Met62 and Glu72.As shown in Figure 3A and Supplementary Figure 1D, Cuc-B was found to dock to HDM2 (docking score-¼ À 5.70 kcal/mol) comparatively more strongly than Y30 (inhibitor of HDM2, docking score ¼ À 4.00 kcal/mol) with a stable and convergent RMSD plot (Figure 3B).The superimposed images of Cuc-B and Y30 with HDM2 showed similar interactions throughout the simulation by forming hydrophobic interactions with residues Leu54, Glu56 and Met62 (Figure 3C, D; Table 1 and Supplementary Figure 1D).The MM/GBSA binding energy between Cuc-B and HDM2 was found to be À 52.44 kcal/mol, which is slightly lower than the known binding energy of the inhibitor Y30 of À 67.00 kcal/ mol (Table 1).These data suggest that in addition to its ability to disrupt the mortalin-p53 interaction, Cuc-B could inhibit HDM2 and reactivate p53 function.

Downregulation of mortalin and HDM2 by cucurbitacin-B
We then investigated the effect of Cuc-B on the expression levels of mortalin and HDM2, which play a key role in p53 function, cancer cell proliferation and migration characteristics.Non-toxic concentrations of Cuc-B were first determined to investigate its effect on cell migration.As shown in Figure 4, U2OS cells treated with 0.01 and 0.02 mM Cuc-B showed no cytotoxicity in either short-term (Figure 4A, B) or long-term (Figure 4C) viability assays.Cells treated with these concentrations of Cuc-B were examined for the expression of mortalin and HDM2 proteins.As shown in Figure 5A, a significant decrease in mortalin protein was observed in Cuc-B treated cells.At the same time, these cells showed a decrease in HDM2 protein (Figure 5B).Furthermore, other proteins including CARF, vimentin and hnRNP-K, which are enriched in metastatic cancer cells with their key role in cell migration, were found to be decreased in Cuc-B treated cells compared to the control.We also performed immunostaining and detected a decrease in Clic-1, N-cadherin, b-catenin, MMP-7, MMP-9 and MMP-2 in Cuc-B treated cells (Figure 6A).Based on the above  computational and experimental data, we predicted that Cuc-B has the ability to inhibit cancer cell migration.

Anti-metastatic potential of cucurbitacin-B
Cells treated with low non-toxic doses of Cuc-B were subjected to cell migration, invasion and clustering assays.As   To rule out a cell line-specific effect of Cuc-B, we examined its effect on five additional cancer cell types.Similar to the effect on U2OS cells, these cell lines also showed dosedependent toxicity with IC50 and IC10-20 in the range of 0.1-0.5 mM and 0.01-0.02mM (Supplementary Figure 2).We also performed molecular analysis in HSC3 cells (mutant p53) and determined the effect of low non-toxic doses of Cuc-B (0.01 and 0.02 mM; Supplementary Figure 2B, C) on cell  migration.Cuc-B treated cells showed a decrease in their ability to migrate and form clusters (Supplementary Figures 3A and 3B).Furthermore, similar to U2OS cells, Cuc-B treated HSC3 cells showed a decrease in mortalin, HDM2 and vimentin immunostaining (key regulators of cell migration).Taken together, these data support the anti-metastatic potential of Cuc-B.
Mortalin has been shown to contribute to cancer metastasis, poor disease prognosis and reduced survival (Na et al., 2016;Wadhwa et al., 2002;Yoon et al., 2022).Knockdown of mortalin by specific shRNA results in impaired proliferation and stemness properties of cancer cells (Lu et al., 2011b;Na et al., 2016;Wadhwa et al., 2002).Based on these findings, targeting mortalin by small natural products was considered useful for cancer treatment.We previously reported that Cuc-B has the ability to bind to mortalin protein (Garg et al., 2020), and in the present study, we performed extensive computational and in vitro biochemical evaluations.Computational analyses revealed that Cuc-B interacts with mortalin with a higher affinity than its known inhibitor MKT077 (Figure 3A-C).The analyses showed that Cuc-B has a high binding affinity for mortalin and interacts at its p53 binding site.It also showed interaction with p53 at its mortalin binding site (residues 323-337) (Figure 1).The MM/GBSA binding energies calculated for Cuc-B: mortalin were À 55.82 kcal/mol, MKT077-mortalin was À 52.22 kcal/mol and Cuc-p53 was À 59.56 kcal/mol (Table 1).These analyses strongly suggested that the interaction of p53 and mortalin could be effectively abrogated and promote the translocation of p53 to the nucleus, allowing it to reactivate its function, as previously reported (Garg et al., 2020;Lu et al., 2011b;Na et al., 2016).Furthermore, computational analyses revealed the ability of Cuc-B to interact with and inhibit HMD2, similar to Y30 (HDM2 inhibitor) (Figure 3A, B).Cuc-B showed stable interactions with the p53 binding domain of HDM2 (residues Leu54, Gly58, Met62 and Glu72), similar to Y30 (Figure 3D), although the MM/GBSA binding energy of Cuc-HDM2 (À 52.44 kcal/mol) was weak compared to Y30-HDM2 (À 67.00 kcal/mol) (Table 1).We also performed simulations of the p53, mortalin and HDM2 apoproteins for 200 ns (Supplementary Figure 4A-C).The average structures from the simulation were obtained and superimposed on the average representative structures of the cucurbitacin-bound proteins.As shown in Supplementary Figure 4B and C, no significant changes were observed in mortalin and HDM2 after binding to the Cuc-B.The RMSD was less than 2 Å for both complexes after superimposing the structures.However, there is a deviation in the p53 chain due to the free-hanging helix loop motif at the mortalin binding site (Supplementary Figure 4A).This change could be attributed to the use of the p53 monomer instead of the tetrameric structure, which is also evident from the slight fluctuation observed in the RMSD plot (Supplementary Figure 4D).
In addition to the computational analyses, we performed expression analyses in cells treated with low non-toxic concentrations of Cuc-B, as determined by short and long-term viability assays.It was confirmed that the low concentrations, such as 0.01 and 0.02 mM, of Cuc-B did not affect cell proliferation and were therefore suitable for determining the effect on cell migration characteristics.Cells treated with these low concentrations of Cuc-B showed down-regulation of several proteins involved in cell migration and metastasis of cancer cells.These included mortalin, HDM2, CARF, vimentin, hnRNP-K, Clic-1, N-cadherin, b-catenin, MMP7, MMP9, MMP2.Finally, cell migration, invasion and clustering assays showed that Cuc-B significantly inhibited cell migration.Reduction of mortalin has been shown to impair cancer cell migration and metastasis via inhibition of EMT signalling (Na et al., 2016;Yoon et al., 2022).Decrease in HDM2 could also cause polymorphic effects including increase in p53 activity and inhibition of cell migration mediated by decrease in CARF and associated inhibition of EMT and b-catenin signalling (Cheung et al., 2009;Huang, 2010;Kalra et al., 2018;Kamrul et al., 2007).miR-708 and Soyasapogenol-A have been shown to target CARF resulting in inhibition of cancer cell migration (Kaul et al., 2021;Omar et al., 2020).Zhang et al (Zhang et al., 2022) recently demonstrated enhanced anti-tumour immunity through up-regulation of CD4 and CD8 and inhibition of cancer metastasis by Cuc-B in a mouse model.Piao et al. (2018) showed inhibition of cell migration by Cuc-B in HUVEC cells mediated by inhibition of vascular endothelial growth factor receptor 2. Mortalin has been shown to regulate hnRNP-K and VEGF (Ryu et al., 2014), supporting the results of the present study.Taken together, Cuc-B is proposed as an anti-metastatic compound and may be useful in the treatment and/or prevention of cancer metastasis.

Conclusions
Cancer is an extremely complex and deadly disease.Despite recent advances in its diagnosis and treatment, cancer incidence and mortality continue to increase.Cancer metastasis is a major obstacle to its therapy and is regulated by a large number of oncogenic proteins, including mortalin and HDM2.Identifying natural products with anti-metastatic activities and understanding their molecular mechanism(s) of action are important avenues for cancer treatment and prevention.Although Cuc-B was previously reported to possess anticancer activity, the mechanism of action/molecular signalling pathways have not been clearly elucidated (Garg et al., 2018(Garg et al., , 2020;;Luo et al., 2018;Shukla et al., 2016;Thoennissen et al., 2009;Zeng et al., 2021).Our computational analyses revealed that Cuc-B has the ability to interact with mortalin, p53 and HDM2 in a manner that could activate the tumour suppressor activities of p53 and induce growth arrest in cancer cells (Garg et al., 2018(Garg et al., , 2020)).In addition, protein analysis of Cuc-B treated cells showed down-regulation of mortalin, HDM2 and several other proteins involved in cancer cell migration, suggesting that it has anti-migration/metastasis activities.Indeed, a reduction in cell migration, invasion and clustering characteristics of cancer cells treated with Cuc-B was observed.This was supported by the downregulation of proteins (b-catenin, hnRNP-K, vimentin and MMPs) that play a key role in cancer cell migration/metastasis (Supplementary Figure 5).Taken together, it is suggested that while high doses of Cuc-B could be used to induce growth arrest/apoptosis in cancer cells, low doses possess anti-migration potential and could therefore be recruited as a natural drug for treatment/prevention of cancer metastasis (Supplementary Figure 5).

Figure 1 .
Figure 1.Interactions of Cuc-B with mortalin at its mortalin-p53 binding site.(A) Interactions of Cuc-B with mortalin (at its p53 binding site) in the average structure obtained from the simulation trajectory.(B) Interactions of MKT077 with the p53 binding region of mortalin in the average structure obtained from the simulation trajectory.(C) RMSD plot of Cuc and MKT077 bound mortalin throughout the simulation.(D) Interaction of Cuc-B with the mortalin binding region of p53 as observed in the average structure obtained from the simulation trajectory.(E) RMSD plot showing stable interactions of Cuc-B with p53 throughout the simulation.(F).Interaction fraction of mortalin residues involved in binding with Cuc-B.

Figure 2 .
Figure 2. (A) The binding of the Cuc-B tail to mortalin (as visualised during the simulation from the frames extracted from the trajectory between 110 and 200 ns) could be the reason for the deviation in the RMSD during this time period.(B) RMSD plot of the Cuc-B ligand throughout the 200 ns simulation.(C) Superimposed structure showing Cuc-B and MKT077 interacting within the same region of mortalin (p53 binding domain of mortalin).
shown in Figure7A, B, a significant reduction in cell migration ability was observed in Cuc-B treated cells.There was a dose-dependent reduction in cell invasion (Figure7C, D) and clustering (Figure7E, F) characteristics.These results confirmed that low non-toxic doses of Cuc-B have anti-metastatic potential.

Figure 3 .
Figure 3. Cuc-B docked to HDM2 with a higher docking score than Y30, a known HDM2 inhibitor.(A) Interaction of Cuc-B with the p53 binding domain of HDM2 in the average structure obtained from the simulation trajectory.(B) RMSD plot of Cuc-B and Y30-bound HDM2 showing a stable interaction throughout the simulation.(C) Superimposed image showing Cuc-B and Y30 interacting with the same region of HDM2 (p53 binding domain of HDM2).(D) Interaction of Y30 with the p53-binding domain of HDM2 in the average structure obtained from the simulation trajectory.

Figure 4 .
Figure 4. Dose-dependent cytotoxicity of Cuc-B on human osteosarcoma cells.Cells treated with serially increasing concentrations of Cuc-B showed no effect on viability up to 0.02 mM.(A).A decrease in viability was observed at 0.04, 0.08 and 0.16 mM of Cuc-B.Cells cultured in the presence of 0.01 and 0.02 mM of Cuc-B showed no toxicity as observed by cell morphology (48 h post treatment) (B) and colony forming assay (12 days post treatment) (C).

Figure 7 .
Figure 7. Cuc-B treated U2OS cells showed a decrease in cell migration, invasion and clustering ability.Migration (A and B), invasion (C and D) and clustering (E and F) of control and treated cells as determined by wound scratch assay, invasion assay and clustering assay, respectively, are shown.Quantitation from three independent experiments is shown in B, D and F. Standard deviation and statistical analysis of Cuc-B treated vs. control (DMSO treated) cells are shown.NS ¼ not significant, � p < 0.05 (significant), �� p < 0.01 (highly significant) or ��� p < 0.001 (very highly significant).

Table 1 .
Overall summary of the interactions of Cucurbitacin-B with Mortalin, p53 and HDM2.