Identification of Pseudobaptigenin As a Novel Polyphenol-Based Multi-Target Antagonist of Different Hormone Receptors for Breast Cancer Therapeutics

Abstract

Breast cancer (BC) is one of tmost prevalent cancers in the world and is one of the major reasons for the death of women worldwide. BC is majorly categorized based on the presence or absence of three cell receptors ER, PR and HER2. The latest treatment for BC involves interfering with the production and action of hormones such as estrogen and progesterone. These hormones bind with receptors such as ER and PR and enhance the growth and proliferation of the BC cells. Although the available are effective, the increasing resistance and side effects related to hormonal imbalance are significant and hence there is a need for designing. On the other hand, plant-derivative products have gained a lot of popularity for their promising anti-cancerous activities. Polyphenols are one such group of plant derivatives that have proven to be useful against cancer. In the present study, an in-silico approach was used to search for a polyphenol that can inhibit ER. In this work, a total of 750 polyphenols were taken into consideration. This number was narrowed down to 55, based on their ADMET properties. These 55 polyphenols were then docked to the receptors, ER, PR and HER2. The molecular docking was followed by Molecular Dynamics (MD) simulations. Based on molecular docking and MD simulation results it was concluded that Pseudobaptigenin has the potential to be an inhibitor of ER, PR, and HER2.

1. Introduction

Breast cancer (BC) is one of the major reasons for the death of women worldwide due to its high prevalence. According to GLOBOCAN 2020, female breast cancer has emerged as the most prevailing cancer surpassing lung cancer. The number of new cases for BC in women was 2,261,419 comprising 11.7% of total cancer cases across the world in the year 2020 alone. BC was not limited to only females, but around 2,710 men in the United States were diagnosed with breast invasive carcinoma in the year 2020. In women, BC is the leading cause of cancer-related deaths (Sung et al., 2021). Breast cancer subtypes are defined by immunohistochemistry expression of estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor receptor 2 (HER2). The presence or absence of these receptors defines the subtype of BC and hence the treatment strategies involving interaction with these three receptors. If all three receptors are when not expressed on a cancer cell, then it is called a “Triple Negative” subtype (Onitilo et al., 2009).

Estrogen controls multiple physiological processes including the growth and differentiation, and functioning of the reproductive system. Most of the actions of estrogen are controlled by ER, which belongs to the nuclear hormone receptor family. In the case of breast cancer, the ER is over-expressed and the over-expression of ER supports the maintenance and proliferation of the breast cancer cells. Hence, ER is a very popular molecular target for breast cancer treatment and drug discovery (Duffy, 2006; Shanle & Xu, 2010; Yager & Davidson, 2006). Following is a graph obtained from the UALCAN database (http://ualcan.path.uab.edu/) (Chandrashekar et al., 2017, 2022), conveying that expression of the ESR1 gene (gene coding for ER) is much higher in the BC sample when compared to the normal sample (Figure 1.a.).

Figure 1:

a. Expression of ESR1 in BC based on sample type. b. Expression of PGR in BC based on sample type. c. Expression of erbB2 in BC based on sample type. Data is obtained from the UALCAN database and statistical value is measured in terms of the p-value. ns (p > 0.05); * (p ≤ 0.05); ** (p ≤ 0.01); *** (p ≤ 0.001); **** (p ≤ 0.0001).

Progesterone is an ovarian steroid hormone that is required for appropriate breast development during puberty, as well as lactating and breastfeeding preparation. PR-positive breast cancer refers to breast cancer cells that have progesterone receptors. Progesterone increases the sensitivity of breast cancer cells to the effects of growth factors by upregulating target genes that include signal transduction pathway components (namely, EGFR and EGFR ligands, IRS-2, cyclins D and E, p21) (McCormack et al., 2020). The following graph obtained from the UALCAN database (Chandrashekar et al., 2017, 2022) shows that the expression of the PGR gene (gene coding for PR), is much more in the case of breast cancer sample than in a normal sample (Figure 1.b.).

The human epidermal growth factor receptor-2 (HER2/erbB2) is one of the four transmembrane receptors that govern cell growth and differentiation through signal transduction pathways. Increased stability through dimer formation initiates a variety of signaling pathways leading to cell proliferation and tumorigenesis. HER-2 gene amplification and/or protein overexpression have been identified in 10%–34% of invasive breast cancers (Ross et al., 2009). HER-2 gene amplification in breast cancer has also been associated with increased cell proliferation, cell motility, tumor invasiveness, progressive regional and distant metastases, accelerated angiogenesis, and reduced apoptosis (Moasser, 2007). Data obtained from the UALCAN database (Chandrashekar et al., 2017, 2022) suggests that the expression of the erbB2 gene (gene coding for HER2 receptor) is highly expressed in the breast cancer sample than in the normal sample (Figure 1.c.).

Hence, all three receptors that are ER, PR and HER2 are playing a significant role in the case of breast cancer, and the data obtained from the UALCAN database also statistically substantiate the fact that these receptors are over-expressed in breast cancer samples and hence can be a good target for the treatment strategies.

The present drugs for the treatment of breast cancer interfere with the production and function of the hormones, which in turn has many side effects such as blood clots, strokes, uterine cancer, and cataracts (Freedman et al., 2011; Parkkari et al., 2003). Dealing with treatment side effects, resistance and recurrence are challenges in the management of BC. Recurrent illness, primarily metastases, affects 30% of early-stage BC patients. As a result, developing novel strategic medicines to properly treat each BC subtype is critical (Pisani et al., 2002). Even though there are several effective anticancer drugs and active inhibitors against various protein targets for the treatment of BC, an increase in resistance coupled with several side effects indicates the need for designing better and more effective therapy. To address the issue of side effects, our approach is to identify natural compounds having anti-BC properties with zero side effects. In this search for natural compounds, Polyphenols are a group of plant secondary metabolites characterized by the presence of multiple phenol units in their chemical structure (Abbas et al., 2016). Polyphenols are reactive species towards oxidation and hence are described as antioxidants (Santos et al., 2005). Polyphenols have found application in the field of biological systems to cure different diseases. Various polyphenols are found to have promising biological activities such as anti-inflammatory, antibacterial, anticancer, antihyperglycemic, and antimutagenic (Daglia, 2012; Scalbert et al., 2005). Due to the propitious properties of polyphenols, we are encouraged to use polyphenols for our studies of polyphenols against BC.

Hence, in the present study, we have evaluated the action of polyphenols against ER, PR, and HER2 receptors. We have assessed the physicochemical properties, specific parameters for drug likeliness, and toxicity of polyphenols using open-source software, to consider them as potential drug candidates for Breast Cancer. We have also shown how a systematic virtual screening and a strategic discovery process involving ADMET tests, toxicity studies, pharmacokinetics studies, molecular docking, and molecular dynamics simulation might lead to a meaningful conclusion.

2. Materials and method

2.1. Retrieval of polyphenols.

For retrieval of polyphenols, Phenol-Explorer 3.6 database (http://phenol-explorer.eu/) was used. The Database classifies polyphenols into 6 major classes and the 6 classes are then further classified into more than 30 sub-classes. In total 750 polyphenols were retrieved from the database (Neveu et al., 2010; Rothwell et al., 2012, 2013). For SMILE notations of the various polyphenols PubChem database (https://pubchem.ncbi.nlm.nih.gov/) was referred (Kim et al., 2021).

2.2. Calculation of molecular properties.

For the calculation of various physicochemical and ADME properties, SwissADME (http://www.swissadme.ch/) web tool was used (Daina et al., 2017). The physicochemical property was followed by the calculation of the Drug Score using DataWarrior software (Sander et al., 2015). Finally For the toxicity study Toxtree software was used (Patlewicz et al., 2008).

2.3. Sorting of polyphenols based on drug-likeness and other physicochemical properties.

For sorting of polyphenols based on drug-likeness properties rules that were considered is Lipinski rule of five (Lipinski et al., 2001), Ghose filter (Ghose et al., 1999), Veber’s rule (Veber et al., 2002), Egan rule (Egan et al., 2000) and Muegge rule (Muegge et al., 2001). Other than the drug-likeness properties, the polyphenols were also sorted based upon some other physicochemical properties that are high GI absorption (Barthe et al., 1999), should be blood-brain barrier permeable (Pardridge, 2005) and should not be a Pgp substrate (Li et al., 2021). The reason behind considering the above-mentioned properties is that these properties are important factors in deciding the overall absorption, distribution and metabolism of the drug in the human body. Apart from these drug-likeness and physicochemical property-based filtering drug scores and in silico toxicity studies were also taken into consideration.

2.4. Retrieval and preparation of lead molecules and target proteins.

For the 3D structure of polyphenols, the PubChem database was referred and for further application in docking the file formats were converted to “.pdbqt” format using Open Babel version 2.4.1 (O’Boyle et al., 2011).

The target proteins (ER, PR and HER2) were searched in the RCSB Protein databank (https://www.rcsb.org/) to obtain the crystal structures of the proteins (Berman et al., 2000). For the structure of ER, PR and HER2, PDB IDs 2IOG (Dykstra et al., 2007), 1E3K (Matias et al., 2000) and 3PP0 (Aertgeerts et al., 2011) were referred respectively.

Before being subjected to docking, solvent molecules and ligands were removed from the protein structure using BIOVIA Discovery Studio Visualizer software (Discovery Studio Visualizer, 2017).

2.5. Determination of active site of the proteins.

For the determination of the active site of the proteins CASTp 3.0 server was used (Tian et al., 2018). The search was made using the respective PDB IDs and the list of residues that are present in the active site of the protein was noted.

2.6. Molecular Docking using AutoDock4.

Before molecular docking, the proteins were prepared in AutoDock Tools-1.5.6. As a part of the preparation, the polar hydrogens were added, Kollman charges were added and the atom types were assigned as AD4.

The molecular docking was carried out using AutoDock4 (Morris et al., 2009). For ER (PDB ID: 2IOG) the grid size was 80×52×80ų with points separated by 0.375Å and was cantered at 32.053, −1.571, 22.04. For PR (PDB ID: 1E3K), the grid size was 52×68×68ų with points separated by 0.375Å and was centered at 30.313, −7.611, 1.695. Similarly, for HER2 (PDB ID: 3PP0, only chain A was considered) the grid size was 82×68×78ų with points separated by 0.375Å and was centered at 10.032, 19.369, 21.952. For all ligands, random starting positions, random orientations and torsions were used. The Lamarckian genetic algorithm method was used for minimization using default parameters in AutoDock4. To decide the docked pose of ligand conformation with the lowest energy was considered.

2.7. Molecular dynamics simulation.

Classical Molecular Dynamics simulations were performed using AMBER20 (D.A. Case et al., 2020) using ff14sb force field for the protein (Maier et al., 2015) and gaff2 force field for the ligands (The Amber Force Fields, 2022; Wang et al., 2004). The solution builder feature of CHARMM-GUI was used to generate the force field parameters for the ligand (Jo et al., 2008; Lee et al., 2016). The missing hydrogen atoms were added using the LEaP module of the AMBER20 package (D.A. Case et al., 2020). The protein was then subjected to energy minimization using the steepest descent and conjugate gradient algorithms for 2000 steps. The energy minimized structure was solvated in a rectangular water box containing TIP3P water molecules (Jorgensen et al., 1983). To calculate the electrostatic interactions at a cut the distance of 12Å Particle mesh Ewald method was used (Essmann et al., 1998). Initial minimization and equilibration were carried out to avoid bad contact. This was followed by equilibration using an NVT ensemble at a temperature of 300K for a 500ps time duration. Then the system was equilibrated using an NPT ensemble at a pressure of 1atm for 1ns time duration. 2fs time step was considered for minimization and equilibration production. The production run was then carried out in an NPT ensemble at a temperature of 300K and pressure of 1atm for 100ns. The coordinates were saved at a time interval of 2ps and hence total of 50,000 frames were saved for the time duration of 100ns.

2.8. Trajectory analyses.

VMD was used for visualizing the trajectory (Humphrey et al., 1996). ). The analyses like calculations of backbone root mean square standard deviation (RMSD) with respect to the starting structure, B-Factor, Solvent Accessible Surface Area (SASA) and Radius of Gyration (RoG) were done using the CPPTRAJ module in AMBER (Roe & Cheatham, 2013). The free energy of ligand binding with the receptor was also calculated. For the calculation of free energy two methods are popularly used, which are the molecular mechanics Poisson Boltzmann surface area (MM-PBSA) and molecular mechanics generalized Born surface area (MM-GBSA) methods. Although MM-PBSA is theoretically more rigorous, MM-GBSA is as efficient as MM-PBSA for the calculation of binding affinity. In MM-GBSA approach, the gas-phase energies are calculated using the molecular mechanics force field, which is a sum of van der Waals (VDWAALS) and electrostatic (EEL) energies, whereas the solvation energy includes a polar and a non-polar component. The polar component is calculated using the implicit Solvent Model, i.e. Generalized Born (GB), (EGB) and the non-Polar component depends on solvent accessible surface area (ESURF). Free energies of solvation are estimated by applying Poisson–Boltzmann (PB) calculations for the electrostatic contribution and a surface-area-dependent term for the non-electrostatic contribution to solvation. In our study, the MM-GBSA method was used to calculate the relative free energy between ligand and protein. The calculation was carried out with the help of the MM-GBSA suite of AMBER (Miller et al., 2012). The binding free energy was calculated as:

$$\mathrm{\Delta }{\mathrm{G}}_{\mathrm{bind}}=\mathrm{\Delta }{\mathrm{E}}_{\mathrm{MM}}+\mathrm{\Delta }{\mathrm{G}}_{\mathrm{solv}}-\mathrm{T}\mathrm{\Delta }\mathrm{S}$$

Where $\mathrm{\Delta }{\mathrm{E}}_{\mathrm{MM}}$, $\mathrm{\Delta }{\mathrm{G}}_{\mathrm{solv}}$ and $\mathrm{T}\mathrm{\Delta }\mathrm{S}$ represent molecular mechanics energy change, solvation-free energy and conformational entropy change during binding. A total of 500 snapshots, equally distributed in the last 20 ns of the all the three trajectories (Pseudobaptigenin and ER. HER2 and PR complexes) were selected to calculate the binding free energy. The single trajectory approach was used to calculate the binding free energy between the ligand and the protein. The energetic contribution of each amino-acid residue towards ligand binding was also calculated by following the per-residue energy decomposition modules of MM-GBSA. The graphs for RMSD and B-Factor were plotted using the QtGrace program.

3. Results and Discussion

3.1. Retrieval and screening of physicochemical properties of polyphenols

Out of 750 polyphenols, the ADME properties for 4 polyphenols could not be calculated due to their large structure and limitations of handling large structures in SwissADME. Out of these 750 polyphenols, only 239 compounds passed the Lipinski, Ghose, Veber, Egan, and Muegge tests for Druglikeness. Out of these 239, 55 compounds showed high GI Absorption, positive to BBB permeability and are not a substrate of P glycoprotein. This series of filtering the compounds based on their physicochemical and ADMET properties resulted in 55 candidate molecules and hence toxicity study was performed on these 55 compounds using Toxtree (Patlewicz et al., 2008) and DataWarrior (Sander et al., 2015). The parameters assessed by Toxtree showed that most compounds were not particularly toxic. DataWarrior was used to check for properties like Drug Score. A positive drug score indicates that the compound has an overall potential to qualify as a drug, it combines drug-likeness, partition coefficient (cLogP), solubility (logS), molecular weight and toxicity risks in one handy value. All 55 compounds, in this case, presented a positive drug score value; they are further considered for molecular docking. Table 1 lists the number of polyphenols obtained after every step of filtering). Tables S1 and S2 tabulate the results obtained from the SwissADME web tool and toxicity studies for the 55 sorted ligands.

Table 1:

Number of polyphenols after every step of filtering.

Filtering steps	Result
Total compounds from the polyphenol explorer database	750
Total compounds evaluated in SwissADME	746 (= 750–4)
Total compounds qualifying drug-likeness properties (i.e., qualifying Lipinski, Ghose, Veber, Egan, Muegge rules)	239
High GI Absorption, should not be Pgp substrate and should be blood-brain barrier permeable	55
Drug score using DataWarrior	55
Toxicity studies using Toxtree	55
Ligands to be considered for Docking	55

3.2. Investigating protein-ligand interaction using molecular docking

Molecular docking was carried out for the 55 polyphenols against the wild-type ER, PR and HER2 using AutoDock4. The free energies of binding for the series of 55 ligands to each of the three proteins have been tabulated in Table 2.

Table 2:

Docking results of 55 polyphenols against the three receptors.

S. No.	Compound Name	ΔG with ER (kCal/mole)	ΔG with PR (kCal/mole)	ΔG with HER2 (kCal/mole)
1	Isoxanthohumol	−6.95	−6.89	−6.28
2	Bisdemethoxycurcumin	−6.91	−6.05	−6.14
3	Dihydroformononetin	−6.64	−6.67	−5.49
4	Sesamolinol	−6.58	−5.29	−5.91
5	Sativanone	−6.51	−6.46	−5.22
6	Pseudobaptigenin	−6.49	−6.87	−5.84
7	Formononetin	−6.48	−6.77	−5.3
8	Sesamin	−6.37	−5.32	−5.75
9	Sesamolin	−6.3	−5.75	−6.92
10	7,4’-Dihydroxyflavone	−6.27	−6.82	−5.68
11	Episesaminol	−6.19	−5.54	−6.28
12	Isoliquiritigenin	−6.14	−6.24	−5.32
13	Daidzein	−6.12	−6.71	−5.34
14	3’-O-Methylviolanone	−6.11	−5.9	−4.66
15	Dihydroresveratrol	−5.95	−6.11	−5.2
16	Schisandrin C	−5.83	−4.81	−4.63
17	Urolithin B	−5.8	−6.38	−5.69
18	Sesaminol	−5.74	−5.41	−5.52
19	Arctigenin	−5.68	−4.7	−3.57
20	Dihydrobiochanin A	−5.66	−6.68	−5.49
21	Sakuranetin	−5.65	−6.25	−6.2
22	Anhydro-secoisolariciresinol	−5.65	−5.7	−4.45
23	Chrysin	−5.63	−5.84	−5.83
24	Pinocembrin	−5.62	−5.93	−5.92
25	Resveratrol	−5.59	−6.18	−4.94
26	Orobol	−5.56	−6.13	−5.05
27	Trans-Resveratrol	−5.56	−6.14	−4.94
28	Urolithin A	−5.53	−6.12	−5.69
29	O-Desmethylangolensin	−5.51	−6.3	−5.53
30	Pinosylvin	−5.44	−5.97	−5.34
31	2-Dehydro-O-desmethylangolensin	−5.42	−6.37	−5.21
32	Isosakuranetin	−5.41	−5.49	−5.85
33	Pterostilbene	−5.4	−6.19	−5.03
34	3’-Hydroxy-O-desmethylangolensin	−5.39	−6.08	−5.1
35	Dalbergin	−5.31	−5.74	−5.82
36	5-(3’,5’-dihydroxyphenyl)- γ -valerolactone	−5.09	−5.57	−4.83
37	Gomisin M2	−5.06	−3.95	−3.96
38	Dimethylmatairesinol	−4.95	−4.3	−4.1
39	Schisandrol_B	−4.89	−3.71	−4.31
40	5-(3’-Methoxy-4’-hydroxyphenyl)- γ -valerolactone	−4.84	−5.13	−4.31
41	Caffeic_acid_ethyl_ester	−4.81	−4.43	−3.96
42	Bergapten	−4.8	−5.47	−4.7
43	3,4,5,4’-Tetramethoxystilbene	−4.72	−3.79	−4.65
44	Schisandrin	−4.67	−3.67	−3.85
45	Schisanhenol	−4.65	−2.92	−3.72
46	Sinensetin	−4.59	−3.06	−3.9
47	Tetramethylscutellarein	−4.59	−3.44	−5.34
48	4-Hydroxy-3,5,4’-trimethoxystilbene	−4.52	−3.54	−4.39
49	4’-Hydroxy-3,4,5-trimethoxystilbene	−4.4	−3.6	−4.27
50	Tangeretin	−4.36	−3.2	−4.95
51	Xanthotoxin	−4.35	−5.22	−4.7
52	Acetyl_eugenol	−4.33	−4.8	−3.68
53	Isopimpinellin	−4.13	−5.11	−4.51
54	[6]-Gingerol	−3.91	−5.03	−3.34
55	Sinapaldehyde	−3.74	−3.81	−3.93

Apart from this, we also probed the STITCH 5 database (Szklarczyk et al., 2016) for ligands that show possible interactions with ESR1, PGR and HER2 gene and some compounds were found to have a direct and indirect effects on the ESR1 gene (15 Items (Homo Sapiens) - STITCH Network View, 2022). Daidzein was found to bind to ER in the STITCH network. The network obtained from STITCH 5 showing the direct and indirect effect of polyphenols on the various genes is shown in Figure 3. For simplicity and clarity of the figure, we have shown the interaction with 15 compounds and the network for the rest of the compounds has been shown in, Supplementary Figure: 1).

Figure 3:

STITCH 5 result showing the direct and indirect effect of polyphenols on ESR1 (15 Items (Homo Sapiens) - STITCH Network View, 2022).

Keeping in mind that the lesser the ΔG value more the binding affinity between the ligand and protein along with the already existing work regarding these ligands, the ligand Pseudobaptigenin was chosen for further analysis. Pseudobaptigenin was a common ligand in all three receptors and had the best binding affinity and no previous work was found in the literature. Pseudobaptigenin had a ΔG (kCal/mole) of −6.49, −6.87 and −5.84 with ER, PR and HER2 respectively (Figure 2). The structure of Pseudobaptigenin is shown in Figure 4.

Figure 2:

2D ad 3D images of molecular docking interaction profiles of (a) ER- Pseudobaptigenin (b) PR-Pseudobaptigenin and (c) HER2-Pseudobaptigenin.

Figure 4:

a. 2D structure and b. 3D conformation of Pseudobaptigenin.

3.3. Molecular dynamics simulation.

We performed triplicate runs of 100ns of all atomic Molecular Dynamics Simulations for Pseudobaptigenin docked to each of ER, PR and HER2 to observe the behavior of Pseudobaptigenin with ER, PR and HER2 in a dynamic environment. Further To quantify the structural changes in all the three protein complexes, the RMSD, B-factor, Solvent Accessible Surface Area (SASA) and Radius of Gyration (RoG) was computed w.r.t to time. The MM-GBSA analysis was performed to evaluate the binding free energy between ligand and the ER, HER2 and PR complexes. The interaction of ligand with the proteins were also checked and for this average structure of last 10ns of the simulation was considered.

3.3.1. MDS of Pseudobaptigenin and ER complex

To understand the stability of the simulated systems, we computed the RMSD (Root Mean Squared Deviation) of the protein-ligand complexes over time with respect to the starting structure. The plot for the time evolution of the RMSD for the system has been shown in Fig. 5a. From the plot we observe that the RMSD has reached a stable plateau of approximately 2.25Å within 10ns of the simulation. The plateau profile of RMSD indicates that the simulated protein-ligand system is stable for the entire part of the trajectory. We have done our simulations in triplicate and the RMSD plots for the other two replicas are shown in Supp. Fig. 2 and 3. They show similar plateau profiles, indicating stable simulations. While RMSD gives an idea of the stability of the simulated system, we can determine the flexibility of the proteins at a residue level by computing its RMSF (Root Mean Square Fluctuation) or B-factor. We calculated the B-factor of the Pseudobaptigenin and ER complex and plotted it in Fig. 5b. From this plot, we observe that apart from the terminal residues, most of the other residues showed minimum fluctuations. The low value of the B-factor indicates the residues are well placed in the protein structure and their position doesn’t change much over time. We observe similar profiles from the other two replicate simulations and these are shown in Supp Fig. 4 and 5. Thus we can say that the simulated complex is a stable one. To quantify the changes in the surface area and the compactness of the ER complex, the SASA (Solvent Accessible Surface Area) and Radius of Gyration (ROG) of the molecules was computed. The time evolution of radius of gyration of the protein is shown in Fig. 7.a. Initially, the radius of gyration decreased which was followed by a more stable plateau region near 18.4–18.6 Å. This relatively constant compactness of the protein suggests that the system has reached a stable conformation. The plot for the time evolution of the SASA has been shown in Fig. 7.b. From the plot we observe that the SASA value for the complex ranges between 2600–2900 Ų, indicating about the structural stability of the protein.

Figure 5:

a. RMSD graph for Pseudobaptigenin and ER complex. b. B-factor graph for Pseudobaptigenin and ER complex. c. Average structure, interactions of Pseudobaptigenin and ER.

Figure 7:

a. RMSD graph for Pseudobaptigenin and HER2 complex. b. B-factor graph for Pseudobaptigenin and HER2 complex. c. Average structure, interactions of Pseudobaptigenin and HER2.

To understand the molecular details of the binding of Pseudobaptigenin to ER, we calculated and listed the interactions (hydrophobic and hydrogen bonded) of the ligand with ER and this is shown in the Ligplot figure in Fig.5c. We observe that Pseudobaptigenin forms a hydrogen bond with the side chain of Glutamine residue at the 45^th position of ER protein (GLU353 in original PDB file, PDB ID: 2IOG). This Glutamine is a key residue and is present at the active site of ER. Apart from this, there are several hydrophobic interactions that Pseudobaptigenin makes with ER. Along with this the two replicates of the simulation suggests that along with GLU45, the side chain of ARG85 (ARG394 in original PDB file, PDB ID: 2IOG) is also playing an important role in the interaction of protein and ligand (Supplementary Figure 6 and 7). These interactions are responsible for the high binding energy observed between Pseudobaptigenin and ER.

3.3.2. MDS of Pseudobaptigenin and PR complex

The stability of the simulated PR and Pseudobaptigenin complex was determined by calculating the RMSD over time with respect to the starting structure. The plot of the evolution of RMSD with respect to time is shown in Figure 6.a. From the plot it is observed that the RMSD value reached a stable plateau of around 1.4Å within the initial 10ns. The plateau profile of RMSD indicates the protein-ligand complex to be stable for the entire part of the trajectory. We performed the simulation in triplicate and RMSD plots for the rest of the two repetitions showing a similar plateau profile, as shown in Supplementary Figures 8 and 9 hence, indicating stable simulations. Parallelly, the B-factor plot to determine the flexibility of the protein residue was calculated and plotted in Figure 6. b. The B-factor plot shows the fluctuations in residue position to be very low however, it is not the case with the terminal residues. The low value of the B-factor suggests that the ligand is well placed in the protein structure and doesn’t change much over time. Similar B-factor profiles were observed in the two replicates, shown in Supplementary Figures 10 and 11. The time-evolution of the compactness and solvent accessible surface area of the PR complex were computed and shown in the Fig. 7.c. and Fig. 7.d, respectively. The RoG value remained relatively constant ~18 Å, throughout the simulation time, indicating the the system has maintained a stable conformation. The plot for the time evolution of the SASA has been shown in Fig. 7.d. From the plot we observe that the SASA value for the complex ranges between 2400–2800 Ų, suggesting that the structure ahas remained stable during the simulation time. In the case of Pseudobaptigenin and PR interaction, the ligand was found to form a hydrogen bond with the side chain of Lysine residue at 88^th position (LYS769 in original PDB file, chain A of PDB ID: 1E3K) as shown in Figure 6. c. Other than the hydrogen bond there are a few more non-bonded interactions with the residues present in the active site of PR also (MENTION THE BONDS). And the other two replicas of the simulation show hydrogen bonding with the side chain of ASP28 (ASP709 in the original PDB file, chain A of PDB ID: 1E3K) (Supplementary Figure 12) and the side chain of ARG85 (ARG766 in the original PDB file, chain A of PDB ID: 1E3K) (Supplementary Figure 13).

Figure 6:

a. RMSD graph for Pseudobaptigenin and PR complex. b. B-factor graph for Pseudobaptigenin and PR complex. c. Average structure, interactions of Pseudobaptigenin and PR.

3.3.3. MDS of Pseudobaptigenin and HER2 complex

In the case of Pseudobaptigenin and HER2 complex, the RMSD was computed with respect to the starting structure. The plot for RMSD with respect to time is shown in Figure 7. a. From the plot, we observed that first, it fluctuated between 1.5Å and 2.5Å for the initial 30ns then reached a stable plateau of approximately 3Å. The plateau profile indicates that Pseudobaptigenin and HER2 complex is stable for the simulated trajectory. We did the simulation in triplicate and the other two replicas are shown in Supplementary Figures 14 and 15. The two repetitions of the simulation show a similar plateau RMSD plot, hence concluding the stable simulation. To determine the flexibility of protein residues we calculated the B-factor of Pseudobaptigenin and HER2 complex and plotted it in Figure 7. b. The graph was observed to be showing low fluctuations but in the case of terminal residues, some fluctuations were observed. The low B-factor indicates that the ligand is well placed in the protein and the protein structure doesn’t change much. From the other two repetitions, we observe similar B-factor profiles and these are shown in Supplementary Figures 16 and 17. The time-evolution of the radius of gyration and solvent accessible surface area of the HER2 in complex with Pseudobaptigenin were computed and shown in the Fig. 7.e. and Fig. 7.f, respectively. The RoG value remained relatively constant ~19–19.5 Å, throughout the simulation time, indicating the the system has maintained a stable conformation. From the SASA plot (Figg. 7.f.) we observe that the SASA value for the complex ranges between 2800–3200 Ų, suggesting that the structure has remained stable during the simulation time. Comparatively, the HER2 system had higher radius of gyration and solvent accessible surface area than ER and PR complexes, suggesting that HER2 complex was less compact and more flexible than other systems.

To understand the molecular details of the interaction between Pseudobaptigenin and HER2 interaction we calculated and listed the interactions which are shown in the Ligplot figure in Figure 7. c. The ligand was found to be forming a hydrogen bond with the main chain of Methionine residue at the 96^th position (MET801 of original PDB file, chain A of PDB ID: 3PP0). It is conclusive that this interaction will not be affected by any mutation at the 96^th position considering the main chain is taking part in the interaction. Other than the hydrogen bonding there are many other non-bonded interactions also with the residues in the active site of HER2. Apart from this, the two replicates of the simulation suggest that the interaction between Pseudobaptigenin and HER2 is purely non-bonded (Supplementary Figures 18 and 19). The structural changes in the first and last frame of the Pseudobaptigenin-ER complex were calculated by overlaying both the structures in Chimera. The RMSD was found to be 2.782 Å shown in Supplementary Figure 20a. The first and last frame of the Pseudobaptigenin-PR complex were also superimposed and the RMSD was found to be 0.907 Å shown in Supplementary Figure 20b and the RMSD for structures of Pseudobaptigenin-HER2 complex before and after simulation was found to be 2.026 Å shown in Figure 20c. This shows that binding of Pseudobaptigenin with ER protein led to more structural changes as compared to other systems.

3.4. Binding ability of Pseudobaptigenin to ER, HER2 and PR proteins

The binding free energy of Pseudobaptigenin to ER, HER2 and PR proteins was calculated using the MM-GBSA suite of AMBER. It was carried out for 500 frames equally distributed within the last 20ns of all three trajectories. The binding energy of the Pseudobaptigenin to ER was found to be −44.10 +/− 0.39 kCal/mol (Fig. 9.a.), indicating strong binding affinity between Pseudobaptigenin and ER. The other two repetitions simulation also shows the energy to be −40.9137 kCal/mol and −36.8224 kCal/mol, suggesting the same conclusion.

Figure 9:

a. Table showing binding free-energy values of Pseudobaptigenin with ER, PR and HER2 complexes. Per-residue energy decomposition results obtained for b. ER complex. c. PR complex. d. HER complex.

The per-residue energy decomposition analysis for the ER complex revealed that residues - LEU38, LEU41, ALA42, LEU79, LEU83, ARG86, ARG110, GLU111, ILE116, GLY213, HID216 and MET220 positively contributed towards ligand binding, however, GLU45 negatively contributed to the ligand binding as shown in Fig 9b. Among these residues, LEU79, ARG83 and MET220 have shown a greater affinity towards the ligand as they have energy contribution of more than −1.5 kCal/mol and were also involved in hydrophobic interactions with the ligand as per the Ligplot results (Fig. 5.c.).

The binding free energy of Pseudobaptigenin to PR protein was found to be −42.37 +/− 0.39 kCal/mole. The other two repeat sets of simulation also show the energy to be −16.4986 kCal/mole and −22.9362 kCal/mole, concluding the strong binding affinity between ligand and protein. Thus, from our triplicate simulation data, we can conclude that the binding of the ligand to the protein results in the formation of a stable complex. The energetic contribution made by amino-acid residues in binding with Pseudobaptigenin are shown in Fig. 9.c. From this analysis, it was found that VAL17, ILE18, GLN44, and ARG85 residues have shown greater contribution towards ligand binding as the energy contribution by these residues was greater than 1.5 kCal/mol. Other residues, which have shown affinity to ligand binding in the PR complex were PRO15, LEU40, TRP51, LEU77, TYR96, PHE97, ALA98, PRO99, and LYS141. The binding free energy of the ligand to the HER2 complex was found to be −32.71 +/− 0.71 kCal/mol. The other two repetitions of the simulation show the binding energy of ligand and protein to be −26.0811 kCal/mole and −31.9566 kCal/mole hence, concluding strong binding affinity between Pseudobaptigenin and HER2. The overall trajectory results suggest a stable complex of Pseudobaptigenin and HER2 that too with a strong binding affinity between Pseudobaptigenin and HER2.

According to the per-residue energy decomposition analysis, PHE26, GLY160 and LEU161 were the major contributors to ligand binding. Other important contributors to ligand binding in the HER2 system were, LEU50, THR54, GLU61, LYS176, VAL177, and PRO178. However, in comparison with other complexes, it was found to have the least binding with the Pseudobaptigenin (−32.71 kCal/mol) than the ER-Pseudobaptigenin system (−41.10 kCal/mol) and PR-Pseudobaptigenin system (−42.37 kCal/mol). This observation was in agreement with the docking results where the binding affinity of Pseudobaptigenin with PR protein was highest (−6.87 kCal/mol), followed by ER complex (−6.49 kCal/mol) and least for HER2 complex (−5.84 kCal/mol).

The MM-GBSA analysis revealed that van der Waal energy is a major contributor to ligand binding in all three systems. This energy is associated with the attractive and repulsive interactions between atoms in a protein-ligand complex. Its higher contribution to ligand binding, suggests that the protein-ligand complex is stabilized by hydrophobic interactions, which was evident from the ligand interaction diagrams too.

4. Conclusion

Recently, polyphenols have gained a lot of attention for their application in curing different diseases. The polyphenols were checked for effect against BC, Estrogen receptor, Progesterone receptor and Human Epidermal Growth Factor Receptor 2 in particular. Based on ADME properties, physicochemical properties and drug-likeness rules the polyphenols were sorted, then these were studied for toxicity and the drug score was calculated using DataWarrior and Toxtree software. After this, the polyphenols were docked with ER, PR and HER2. A site-specific docking was carried out in AutoDock4. Then for further verification and in search of novel ligand literature survey was conducted keeping binding energy in mind and based on that “Pseudobaptigenin” ligand was chosen as a novel ligand and further taken for MDS studies. MDS was carried out for 100ns in AMBER 20. The MDS studies of Pseudobaptigenin with the three receptors (ER, PR, and HER2) show that the three simulations are stable, the ligand is well placed in the protein and the protein residues do not fluctuate much with respect to its original position. It also reveals that Pseudobaptigenin has a strong binding affinity with all three receptors and the important residues are also being identified in the three receptors that are interacting with Pseudobaptigenin. If we compare then, the Pseudobaptigenin-PR interaction is most promising followed by Pseudobaptigen-ER and then Pseudobaptigenin-HER2. Hence, the results obtained were very encouraging, the in-silico results obtained pave the path for future experimental studies on using Pseudobaptigenin as a drug targeting ER, PR and HER2 in breast cancer therapy.

Figure 8:

a. Radius of Gyration plot for Pseudobaptigenin and ER complex. b. SASA plot for Pseudobaptigenin and ER complex. c. Radius of Gyration plot for Pseudobaptigenin and PR complex. d. SASA plot for Pseudobaptigenin and ER complex. e. Radius of Gyration plot for Pseudobaptigenin and HER2 complex. f. SASA plot for Pseudobaptigenin and HER2 complex.

Availability of supporting data-

Databases used in the study was mentioned in the manuscript.