In silico evaluations of phytochemicals from Withania somnifera exhibiting anticancer activity against NAD[P]H-quinone oxidoreductase

Abstract

Background

Cancer is a leading cause of death globally and in the US, prompting research into medicinal plants with anticancer properties. Withania somnifera, or Ashwagandha, is one such plants, known for its diverse pharmacological effects. Withaferin A and Viscosalactone B are two compounds found in Ashwagandha with known anticancer activity. The protein NQO1, overexpressed in various cancers, was the focus of this study.

Hypothesis and aim

We hypothesize that specific phytochemicals in Withania somnifera can effectively interact with and inhibit the NQO1 protein, thereby exhibiting anticancer properties. This study aims to identify these interactions using in silico approaches.

Methodology

CFDT was performed using the Gaussian 16 program package, followed by QSAR analysis of the compounds in the PASS online web server. The Schrodinger suite was used to carry out ligand and protein preparation, molecular docking, and molecular dynamic simulation to analyse the interaction of these compounds with NQO1 and ADME studies. Protox-II and SWISSADME tools were used to predict the toxicity and blood–brain barrier permeability of the phytochemicals.

Results and conclusion

CDFT and frontier molecular orbital analyses predicted the stability and reactivity of all the selected molecules. QSAR analysis predicted the biological activity and toxicity of the compounds. Withaferin A exhibited the highest glide gscore (−4.953 kcal/mol) and demonstrated 6 hydrogen bond interactions with NQO1, suggesting its potential as an anticancer agent. Conceptual density functional theory-based analysis suggested the strong electrophilicity of the ligands, further supporting their potential anticancer activities. Viscosalactone B, another phytochemical from Ashwagandha, also showed interactions involving 6 hydrogen bonds with NQO1, with a glide gscore of (−4.593 kcal/mol). Molecular dynamic simulations validated the stability of the Withaferin A-NQO1 complex. ADME-T properties predicted high oral absorption for the selected ligands, indicating that Withaferin A could be a viable orally administered drug.

Introduction

Cancer being the second most common cause of mortality in the US and a significant global public health issue. The multitude of cancer cases and deaths are in increasing order this decade. According to GLOBOCAN 2020, there were19,292,789 new instances of cancer and 9,958,133 cancer-related deaths worldwide in 2020. ¹ Female breast cancer has surpassed lung cancer (11.4%) as the most prevailing diagnosed cancer, with an estimated 2.3 million new cases (11.7%), followed by colorectal (10.0 %), prostate (7.3%), and stomach (5.6%) cancers. ² Since the standard treatments for tumors are expensive and have a higher rate of side effects, researchers are now inclined towards bioactive compounds, such as active proteins, polyphenols, phytosterols, biogenic amines, carotenoids, etc., extracted from plants have been demonstrated to improve a variety of conditions, including diabetes, cancer, cardiovascular disease, and gut, immune system, and neurodegenerative disorders. ³

Various plants with anticancer activity are being used to develop drugs. Withania somnifera (W. somnifera) is a small woody shrub commonly known as ‘‘Winter cherry’’ or ‘‘Indian Ginseng’’, otherwise called ‘Ashwagandha’ in Sanskrit, belonging to the family of Solanaceae with a height of 0.5–2 m. ⁴ It grows in dry parts of sub-tropical regions like Rajasthan, Punjab, Haryana, Uttar Pradesh, Gujarat, Maharashtra, and Madhya Pradesh. It develops as a small shrub (35–75 cm) with a central stem from which branches radiate outward in the shape of stars (stellate) and is tomentose—a thick matting of woolly hair. The fruit is mature when orange-red and contains milk-coagulating characteristics, while the blooms are tiny and green. The plant’s tuberous roots, which are long and brown, are utilised for its medicinal properties. ⁵ Ashwagandha helps maintain appropriate feeding of the tissues, notably muscle and bones, while supporting the normal function of the adrenals and reproductive system. This well-known Ayurvedic herb is utilised in many tonics and formulations. ⁶ W. somnifera is a plant with extensive medicinal importance. It is known for its antibacterial, ⁶ antifungal, ⁷ anti-inflammatory, ⁸ anti-diabetic, anti-cancer ⁹ properties. It is also being studied to treat neurodegenerative diseases. ¹⁰ The plant contains various compounds such as alkaloids, nitrogen, sugar, and hydrocarbons. The GC-MS analysis of the extracts of W. somnifera have reported the presence of phytochemical compounds such as Butane, 1,1- diethoxy-2- methyl, Dodecanoic acid, Amyl Nitrite. ¹¹ W. somnifera contains several phytochemicals that are proven to be medicinally important, a few of them are listed below- Withanolide A, 12-deoxy withastramonolide, Withaferin A, Dihydrowithanolide D,^12,13 Withanoside IV, Viscosalactone B, Withanoside V, ¹⁴ Withanoside VII, ¹⁵ 27-hydroxywithanone, ¹⁶ 3β-methoxy-2,3-dihydroxyWithaferin A, ¹⁷ Ixocarpalactone A, Withanone, ¹⁸ Azetidin-2-one 3,3-dimethyl-4-(1-aminoethyl), Hexadecanoic acid, methyl ester, Withanoside XIII, Withagenin A diglucoside, Withanolide glycosides 1–5. ¹⁹

Several studies have proved the properties as mentioned earlier of W. somnifera. In a study conducted by Mishra et al. (2020) evaluated the antibacterial activity of methanolic extract of the whole plant; the extract demonstrated the effective zone of inhibition against Escherichia coli and Klebsiella pneumonia. ¹³ In an in-vitro study evaluating the the antibacterial activity of various extracts (hexane, ethyl acetate, methanol and aqueous) of W. somnifera against B. cereus, E. coli, S. marcescens, S. marcescens was the most susceptible organism against all the extracts of the plant with a maximum zone of inhibition. ²⁰ In another in-vitro study, researchers extracted a monomeric glycoprotein from the roots of W. somnifera named WSG (Withania somnifera glycoprotein) which exhibited strong antifungal activity against Aspergillus flavus, F. oxysporum, and F. verticilloides. ⁷

An in-vivo study was conducted to evaluate W. somnifera as an effective anti-inflammatory agent, the study used zebra fishes which were implanted with stainless steel. The effect of W. somnifera was investigated by Acridine orange/Ethidium bromide (AO/EtBr) staining and Histopathological analysis. The study not only proved W. somnifera as an effective anti-inflammatory agent but cytokines also greatly enhanced cytokine expression. ⁸

Four human cancer cell lines’ in vitro cytotoxicity was assessed against five cell lines of four different tissues. These were PC-3 and DU-145 from prostrate, HCT-15 from colon, A-549 from lung and IMR-32 of neuroblastoma origin, the leaves, stem and root extracts of W. somnifera using 50% ethanol, were found effective as they prevented cell proliferation and caused cell death. ²¹ The root extracts of W. somnifera which contained Withaferin A- an active compound, was evaluated against pro-metastatic protein of vimentin (Vimentin is an intermediate filament protein and is part of the epithelial to mesenchymal transition (EMT) program to promote metastasis) in MCF10A human mammary epithelial cells, it was established that the 1 µM concentration of Withaferin A can prevent cell proliferation with minimal cytotoxic effects. ²² The aqueous root extracts of W. somnifera were tested against MDA-MB-231 breast cancer cells the extracts were proven to be cytotoxic to MDA MB 231 breast cancer cells with (IC50) at 0.19 mg/mL for 72 h. ²³ A previous in-vivo study reported, Withaferin A induced arrest of the G2/M phase of the cell cycle in prostate cancer cell lines (PC-3 and DU-145) when treated for 48 h in addition to G2/M arrest. It also caused upregulation of phosphorylated Wee-1, phosphorylated histone H3, p21, and Aurora B and downregulation of cyclins (A2, B1, and E2) and a significant reduction in phosphorylated Cdc2 (Tyr15). ²⁴ In a study, the researchers observed that pretreatment of cells with Withaferin A inhibited cell adhesion, migration, and invasion of A549 and H1299 cells in non-small cell lung cancer. Withaferin A suppressed TGFβ1 and TNFα-induced EMT in both cell lines. ²⁵

Viscosalactone B, one of the important phytochemicals extracted from W. somnifera displayed antiproliferative activity against following cell lines PC3, DU145, C42 B, PC3/MDVR, DU145/MDVR, and C42 B/MDVR with IC50 values of 1.17, 0.72, 3.86, 2.06, 0.96 and 1.15 μM, respectively in the In-vitro study conducted. Viscosalactone B also acts as an LSD-1 inhibitor, which is one of the significant targets for preventing prostate cancer. ²⁶ The water extracts of W. somnifera were proven to have cytotoxic effects on certain cancer cells in in-vivo study which used HT1080 cells that caused lung tumors in mice. Further investigation of the cancer cells showed the presence of triethylene glycol, the cells also showed the activation of cancer-inhibiting proteins such as p53 and pRB. Additional analysis reported a decreased cyclin B1 and an increase in cyclin D1, indicating the growth arrest of the cancer cells. ²⁷

An in-silico study was carried out utilizing Autodock Vina involving Pristimerin, ixocarpalactone A, Viscosalactone B and zhankuic acid A phytochemicals derived from W. somnifera were docked with EGFR, HER2, estrogen and NF-jB receptors, to check the affinity of the docked complex. The highest docking energy was exhibited by was pristimerin with a vina score of −13.6 against EGFR. ²⁸ Density functional theory (DFT) provides quantum mechanical modeling to investigate the electronic structure of the molecule crucial for understanding reactivity, stability, and interactions with biological targets ²⁹ in its ground state configuration. By employing DFT, researchers have been able to predict the activity of phytochemicals prior to in vitro or in vivo testing, saving time and resources in drug discovery processes. DFT calculations act as a bridge between theoretical predictions and experimental validations, opening new frontiers in the design and optimization of phytochemicals for medical and industrial applications. ³⁰ Quantum mechanical calculations provide a precise description of reactivity indices within the framework of conceptual density functional theory (CDFT).^31–37 Some of the CDFT-based global reactivity indices generally used are electronegativity (χ), electrophilicity index (ω), hardness (η), softness (S), chemical potential (μ), etc.^38–46 CDFT and its efficacy have aided in gaining insight into the molecular properties and behavior of phytochemicals. Computational studies performed using Viscosalactone B to find out its potential as LSD1 inhibitor, according to docking studies, it formed hydrogen bonds with the Thr11, Lys14, and Arg8 residues of LSD1. ²⁶ To investigate the impact on constitutively active BCR-ABL oncogenic signaling, which results in uncontrolled proliferation and prevention of apoptosis in Chronic Myeloid Leukaemia. MD and MDS were performed. It was discovered that there is interaction between Withaferin A and Withanone, a closely similar withanolide, at both catalytic and allosteric sites of the ABL. In comparison to the therapeutically utilised medicines Imatinib (−78.11 ± 5.21) and Asciminib (−54.00 ± 6.45), the predicted binding energies for Withaferin A were greater at the catalytic site (−82.19 ± 5.48) and allosteric site (−67.00 ± 4.96). At the allosteric site, Withanone had a lower binding energy (−42.11 ± 10.57) than Asciminib. It was discovered that the ligand-induced conformational changes and interaction resembled those of the medications asciminib and imatinib. ⁴⁷

The present study aimed to investigate the interaction between Withania somnifera phytochemicals and the NQO1 protein using in silico approaches, including the Schrodinger suite and Density Functional Theory (DFT) studies. The objective was to identify potential phytochemicals that could serve as anticancer agents for treating various cancers.

Materials and methods

In this study, a library of ligands, comprising of Withaferin A, Withanolide A, Withasomnine, Anaferine, Viscosalactone B, Withanone, Withanolide E, Withanolide B and Withanolide D from Withania somnifera with previously reported bioactivities such as anticancer, anti-oxidant and anti-inflammatory properties was prepared and subjected to conceptual DFT calculations. The ligands were retrieved from the PubChem database and were docked against the active site of the protein downloaded from the Protein data bank using Schrodinger’s computer-aided drug discovery tools for in silico research. All the interactions between the ligand and the protein were evaluated on the Schrödinger Suite (version-13.7).

Computational details

The pre-optimization of all the ligand geometries and frequency calculations were done using the PM6 semiempirical method. ⁴⁸ The geometries were re-optimized using MN12SX/Def2TZVP/H2O level of theory as it has been demonstrated that it facilitates the validation of the 'Koopmans in DFT’ (KID) procedure.^49,50 The SMD solvation model was used for this purpose. ⁵¹ All the ligand geometries were at the minimum of the potential energy surface. Gaussian 16 program package was used for these calculations. ⁵²

Theoretical background of CDFT

The CDFT parameters can be derived using the theoretical concepts of DFT. Employing the finite difference approximation method, hardness (η) ⁵³ is defined as: $\eta ={\left(\frac{{\partial }^{2}E}{\partial N^2}\right)}_{\nu \left(\overrightarrow{r}\right)}={\left(\frac{\partial \mu }{\partial N}\right)}_{\nu \left(\overrightarrow{r}\right)\hspace{0.17em}}\approx IP-EA$ (1)

E is the total energy of the system having N electrons. µ, and ν( $\overrightarrow{r}$ ) are known as the chemical and external potentials, respectively ⁵⁴ . The energy needed to remove an electron from its lowest energy level is denoted by the term vertical ionization potential (IP). Electron affinity (EA), on the other hand, describes the energy change when a single electron is added. Using Koopmans’ theorem ⁵⁵ IP and EA are defined as: $IP\approx {-E}_{HOMO}$ $EA\approx {-E}_{LUMO}$

Furthermore, η can be expressed as: $\eta \approx E_{LUMO}-E_{HOMO}$ (2)

Softness (S) is defined as: $S=\frac{1}{2\eta }$ (3)

Pauling invented the term electronegativity ( $\chi$ ) ³⁹ which is formulated as the negative of chemical potential ( $\mu$ ). It can be represented as: $\chi =-\mu \hspace{0.17em}{=-\left(\frac{\partial E}{\partial N}\right)}_{\nu \left(\overrightarrow{r}\right)}\approx \frac{IP+EA}{2}\approx -\frac{\left(E_{HOMO}+E_{LUMO}\right)}{2}$ (4)µ is the Lagrange multiplier related to the normalization constraint of DFT.

Another reactivity descriptor, electrophilicity index (ω) ⁴⁰ that measures the information related to stability, reactivity, and bonding of the molecular system, is formulated as: $\omega =\frac{{\mu }^2}{2\hspace{0.17em}\eta }=\frac{{\chi }^2}{2\hspace{0.17em}\eta }$ (5)

The ability of a species to accept a fractional charge in comparison to its donating capacity can be elucidated through the concept of net electrophilicity, ⁵⁵ which is expressed as $∆{\omega }^{\pm }={\omega }^{+}-\left(-{\omega }^{-}\right)={\omega }^{+}+{\omega }^{-}$ (6)ω⁺ and ω⁻ denote the electro-accepting and donating powers, ⁵⁶ respectively and can be defined as the tendency of a species to accept or donate fractional amount of charge in certain chemical environment. ${\omega }^{+}=\frac{{\left({\mu }^{+}\right)}^2}{2{\eta }^{+}}=\frac{{EA}^2}{2\left(IP-EA\right)}$ (7) ${\omega }^{-}=\frac{{\left({\mu }^{-}\right)}^2}{2{\eta }^{-}}=\frac{{IP}^2}{2\left(IP-EA\right)}$ (8)

PASS prediction

QSAR analysis was performed to predict the potential ligands based on the relationship between structure and biological activity. The PASS prediction (Prediction of Activity Spectra for Substances) was performed on the selected ligands using the PASS-Way2Drug server (https://www.way2drug.com/PASSOnline/predict.php) using canonical smiles obtained from the PubChem server.^57,58 The results were obtained in the format of Pa (probability to be active) and Pi (probability to be inactive). The Pa was filtered to be over 0.3, indicating a higher probability of biological activity of the compounds. The PASS prediction study predicted both the possible biological activities and the possible adverse effects of the selected ligands.

Ligand preparation

The ligand molecules (Table 1) were downloaded from https://pubchem.ncbi.nlm.nih.gov .The ligand molecules were then configured using the LigPrep module in the Maestro interface. The suitable ionization state was predicted at pH 8.0, types of tautomers were produced, and proper chirality was estimated using the Epik module. Finally, the molecules’ structure was minimized with energy using the OPLS4 force field. ⁵⁹

Table 1.

The 2D structures of all the nine ligands were downloaded from PubChem database and represented along with their PubChem CID. Molecules marked in red denote the presence of oxygen atoms, and those marked with green denote hydrogen atoms.

Ligand name	Pubchem CID	2D structure
Withaferin A	265237
Withanolide A	11294368
Withasomnine	442877
Anaferine	443143
Viscosalactone B	57403080
Withanone	21679027
Withanolide E	301751
Withanolide B	14236711
Withanolide D	161671

Protein preparation and receptor grid generation

Crystal structure of human NAD[P]H-quinone oxidoreductase (NQO1) protein was retrieved from RCSB protein data bank (PDB ID-1D4A, resolution of 1.70 Å and sequence length of 273), followed by removing other solvents and water molecules bound to the protein and then the protein preparation was carried out. Sitemap tool was utilised to identify the potential ligand binding sites. The sitemap identified with the best site score (>1) was selected for receptor grid generation using the Glide module in Maestro (13.7) Schrodinger suite. ⁵⁹

Molecular docking

NQO1 up-regulation has resulted in many types of cancers including breast, uterine cervix carcinoma, lung, and cholangiocarcinoma cells. Overexpressed NQO1 results in development of favourable position for cell proliferation by shielding them from oxidative stress and chemotherapeutic agents leading to cancer progression and resistance to drugs (31), Molecular docking was carried out using the selected 9 ligands against NQO1 using Ligand docking panel in Schrodinger suite (v 13.7).

Validation of molecular docking method

Validation of molecular docking method was performed. Using the above-mentioned docking protocol, the native ligand (FAD 601) present in the protein was re-docked with the protein to ascertain the binding conformations. ⁶⁰ This conformation of the ligand was compared to the co-crystallized structure, and the RMSD values were measured. The RMSD values between the co-crystal ligands and the re-docked ligand should ideally be less than 1Å.

Molecular dynamic simulations (MDS)

The stability of Withaferin A-NQO1 complex was then validated with MDS using DESMOND module (Schrodinger v13.7). Withaferin A-NQO1 complex consisted 37199 atoms, which was subjected to MDS for 100ns, at 300k, under pressure 1.01325 bar. A 10 × 10 × 10 Ǻ orthorhombic box was created around the ligand-protein complex using the system builder workflow, and the TIP3P water model system was applied with OPSL4 forcefield. Na and Cl ions were added to balance the total charge on the system. The output of the molecular dynamics studies was studied in detail using the Simulation Interactions Diagram Report of the Desmond software.

Prediction of ADME-T properties and BOILED-Egg plot analysis

SWISSADME tool (https://www.swissadme.ch/) was employed to develop the BOILED-Egg plot of the selected ligand library. The canonical smiles formula of each ligand was added to the website and BOILED-Egg plot was generated. ADME properties were predicted using the ADME workflow available in the Maestro suite. The Protox-II tool was used to calculate the toxicity of all the compounds, including their acute hepatotoxicity, cytotoxicity, carcinogenicity, mutagenicity, and immunotoxic properties (https://tox.charite.de/protox2). ⁶¹

Results

CDFT-based study

Table 2 depicts the predicted values of global reactivity descriptors, such as electronegativity (χ), chemical hardness (η), softness (S), electrophilicity index (ω), and net electrophilicity ( $∆{\omega }^{\pm }$ ) along with IP and EA for the selected ligands. All the reactivity indices mentioned here, were assessed using Koopman’s approximation in order to acquire a more thorough comprehension of the chemical characteristics of the phytochemicals. Depending on the values of HOMO and LUMO, we have calculated all the global reactivity descriptors based on the equations mentioned in the theoretical background section. Comparing the η value for all the ligands, it was observed that η is highest (6.46325 eV) for Anaferine indicating the least reactivity of the ligand. Conversely, Withanolide B was found to possess the highest reactivity (η = 4.91411 eV) compared to all selected ligands. From Table 2, it is evident that except Withasomnine, all the ligand molecules possess the value of ω > 1.5 eV. The reactivity and efficiency of the drug molecules depends on high value of electrophilicity index. Comparing ω⁺ for the selected ligands, the tendency for electron acceptance was found to be highest for Withanolide B and least for Withasomnine. Conversely, the ligand Withaferin A was found to be highly capable for electron donation based on its high value of ω⁻. The net electrophilicity ( $∆{\omega }^{\pm }$ ) value was found to be highest for ligand Withaferin A.

Table 2.

Predicted global reactivity descriptors for the selected ligands. (All units are in eV, S in eV⁻¹).

Mol. NO.	IP	EA	χ	η	S	ω	ω +	ω -	$∆{\omega }^{\pm }$
161671	6.92476	1.95813	4.44144	4.96662	0.10067	1.98590	0.38600	4.82745	5.21345
265237	7.07306	1.94861	4.51083	5.12445	0.09757	1.98535	0.37049	4.88132	5.25181
442877	5.82678	0.88383	3.35530	4.94295	0.10115	1.13880	0.07902	3.43432	3.51333
443143	7.76287	1.29962	4.53124	6.46325	0.07736	1.58838	0.13066	4.66190	4.79257
11294368	6.92068	1.94670	4.43369	4.97397	0.10052	1.97605	0.38095	4.81464	5.19559
14236711	6.86408	1.94997	4.40702	4.91411	0.10175	1.97613	0.38688	4.79391	5.18079
21679027	6.88448	1.94453	4.41451	4.93996	0.10122	1.97247	0.38271	4.79722	5.17993
57403080	6.89646	1.50778	4.20212	5.38867	0.09279	1.63842	0.21094	4.41306	4.62401
301751	6.91170	1.89174	4.40172	5.01996	0.09960	1.92981	0.35644	4.75816	5.11460

The calculated global reactivity descriptors for the selected ligands are given in Table 2.

Frontier molecular orbital analysis

The molecular orbitals (MOs), i.e., HOMO (Highest occupied molecular orbital) and LUMO (Lowest unoccupied molecular orbital), play crucial roles in determining chemical reactivity and stability. A small energy gap (ΔE) signifies a substantial electron density transfer from the electron donor to the electron acceptor of the molecule. Conversely, a large ΔE indicates low chemical reactivity and high stability of the molecule. The illustration of HOMO and LUMO of the selected ligands are depicted in Figure 1. The molecular orbitals and corresponding energy gaps are calculated using the optimized structure of the ligands. In Withanolide D ligand, the HOMO is mainly localized on the α,β-unsaturated lactone moiety of the molecule, whereas, the LUMO is concentrated over the α,β-unsaturated ketone with hydroxy group at γ-position as well as the epoxy-oxygen atom. The HOMO-LUMO energy gap (∆E) is 4.97 eV. The same HOMO, LUMO orbitals are found in case of Withanolide E ligand as well (except the absence of γ-hydroxy group). But in this case, slightly higher energy i.e., 5.02 eV is required to transfer the electrons from HOMO to LUMO. In Withanolide A, Withanolide B, and Withanone ligands, the HOMO is located on the same position as Withanolide ligand. The LUMO of Withanolide A is mainly centered on the cyclohexanone moiety, whereas the LUMO of Withanolide B and Withanone are located on α,β-unsaturated ketone moiety of the corresponding molecule respectively. The ligand Withanolide B having the lowest ∆E value (4.91 eV) shows highest reactivity, also confirmed from CDFT analysis. For Withaferin A, the HOMO is located in the entire ligand molecule, whereas, the LUMO is confined over the α,β-unsaturated ketone moiety of the molecule. A small lobe is found over epoxy-oxygen atom as well. In Withasomnine ligand, both the MOs are distributed over the entire molecule. For the ligand Anaferine, the HOMO is located over the whole molecule, whereas, the LUMO is concentrated on the center of the ligand molecule. It shows highest stability of the ligand molecule having ∆E is 6.46 eV. For the ligand Viscosalactone B, the HOMO electron clouds are located on the cyclohexanone with β, γ-dihydroxy group, and the LUMO is distributed over the α,β-unsaturated lactone moiety of the ligand molecule.

Figure 1.

The HOMO and LUMO of the selected ligands viz., Withanolide D (a, b), Withaferin A (c, d), Withanolide E (e, f), Withasomnine (g, h), Anaferine (i, j)), Withanolide A (k, l), Withanolide B (m, n), Withanone (o, p), Viscosalactone B (q, r)) at MN12SX/Def2TZVP/H2O level of theory (isosurface = 0.02).

PASS prediction

The prediction of activity spectra for the compounds was performed for all selected ligands, which provided the Pa and Pi values. This prediction identified 11 intended biological activities, which were antineoplastic effect (including breast and lung cancers), antioxidant and hepatoprotective capacity, ability to enhance TP53 and APOA1 expression, inhibitors of HIF1A expression, vasoprotection, membrane integrity agonist and mucomembranous protection (Table 3). In addition, seven adverse effects were also identified (Table 4). Higher values of Pa indicated a higher probability of intended bioactivity/adverse effect, and Pa > Pi suggests the biological activity to be possible. Threshold of Pa >0.7 indicated a highly signifi-cant activity, Pa = 0.3 – 0.7 indicates moderate activity, while Pa <0.3 indicated a low activity. ⁵⁸

Table 3.

The PASS prediction results showing the biological activities of all selected ligands; Pa (probability ‘to be active’) and Pi (probability ‘to be inactive’); These bioactivity activities were not predicted within Withasomnine and Anaferine. Pa value above 0.3 indicated that the compound had potential in silico biological anticancer activity.

	Withaferin A		Withanolide A		Withanolide B		Withanolide D		Withanolide E		Withasomnine		Anaferine		Viscosalactone B		Withanone
	Pa	Pi	Pa	Pi	Pa	Pi	Pa	Pi	Pa	Pi	Pa	Pi	Pa	Pi	Pa	Pi	Pa	Pi
Antineoplastic	0.916	0.005	0.901	0.005	0.897	0.005	0.887	0.005	0.846	0.007	0.454	0.086	-	-	0.915	0.005	0.876	0.005
Antineoplastic (breast cancer)	0.722	0.005	0.341	0.044	0.583	0.012	0.383	0.035	0.393	0.033	-	-	-	-	0.598	0.010	0.584	0.012
Antineoplastic (lung cancer)	0.552	0.011	-	-	0.389	0.025	0.373	0.027	0.538	0.012	-	-	-	-	0.490	0.014	0.362	0.029
Antioxidant	0.623	0.004	-	-	0.446	0.009	0.488	0.007	-	-	-	-	-	-	0.896	0.003	-	-
Hepatoprotectant	0.904	0.002	0.361	0.040	0.648	0.009	0.818	0.004	0.460	0.024	-	-	-	-	0.988	0.001	0.571	0.015
TP53 expression enhancer	0.452	0.114	-	-	0.453	0.113	0.319	0.194	-	-	-	-	-	-	0.478	0.101	0.412	0.134
HIF1A expression inhibitor	0.400	0.101	0.349	0.135	0.341	0.141	0.406	0.098	-	-	-	-	-	-	0.420	0.090	0.423	0.089
APOA1 expression enhancer	0.367	0.098	-	-	-	-	0.312	0.170	-	-	-	-	-	-	0.338	0.131	-	-
Vasoprotector	-	-	-	-	-	-	-	-	-	-	0.321	0.134	-	-	0.354	0.106	-	-
Membrane integrity agonist											0.363	0.157	0.494	0.104
Mucomembranous protector													0.595	0.102

Table 4.

The PASS prediction results showing the adverse and toxic effects of all selected ligands; Pa (probability ‘to be active’) and Pi (probability ‘to be inactive’). Pa value above 0.3 indicated that the compound had predicted adverse and toxic effects.

	Withaferin A		Withanolide A		Withanolide B		Withanolide D		Withanolide E		Withasomnine		Anaferine		Viscosalactone B		Withanone
	Pa	Pi	Pa	Pi	Pa	Pi	Pa	Pi	Pa	Pi	Pa	Pi	Pa	Pi	Pa	Pi	Pa	Pi
Inflammation															0.320	0.204	0.344	0.182
Toxic	0.856	0.018	0.569	0.067	0.686	0.046	0.764	0.033	0.843	0.020					0.858	0.018	0.837	0.021
Pain	0.490	0.103					0.433	0.125	0.412	0.138					0.622	0.062	0.435	0.125
Twitching											0.751	0.069	0.855	0.015
Shivering											0.716	0.072	0.771	0.049
Dizziness	0.452	0.114													0.314	0.222	0.601	0.061
Nausea															0.369	0.161	0.396	0.144

Molecular docking

Molecular docking is crucial in drug discovery and development as it helps in understanding how a potential drug binds to its target protein. Analysis of molecular docking of the various compounds with the protein, showed, the best interactions with Withaferin A and Viscosalactone B with 6 hydrogen bonds (TYR (155), GLY (150), PHE (106), HIE (11), PHE (17)) and the glide g score of −4.953 Kcal/mol and PHE (17), TYR (155), PHE (106), GLY (150), HIE (11) amino acid residues of NQO1 protein with glide gscore of −4.593 Kcal/mol respectively. The other complexes of NQO1 and Withasomnine (1 hydrogen bond, −4.682 Kcal/mol), Withanolide D (3 hydrogen bonds, −4.692 Kcal/mol), Withanolide E (1 hydrogen bond, −3.604 Kcal/mol), Withanolide B (3 hydrogen bonds, −4.528 Kcal/mol), Withanolide A (4hydrogen bonds, −5.053 Kcal/mol), Anaferine (2 hydrogen bonds, −3.442 Kcal/mol) revealed successful bonding and interactions. The docking results for the bioactive compounds from W. somnifera suggest strong and stable binding affinities with NQO1, supporting their potential therapeutic applications in cancer treatment. The favorable glide g scores and multiple hydrogen bonds underline the effectiveness and stability of these interactions, providing insights into their mechanisms of action and guiding further experimental validation. The interactions are visualized in Figure 2.

Figure 2.

2D interaction diagrams of ligands with the protein (NQO1). The binding affinity (Kcal/mol) and amino acids involved in the protein and ligand interaction are mentioned. Withaferin A and Viscosalactone B exhibited the strongest interactions with NQO1, each forming six hydrogen bonds with amino acid residues TYR 155, GLY 150, PHE 106, HIE 11, and PHE 17. Their glide g scores were −4.953 Kcal/mol and −4.593 Kcal/mol, respectively. Other compounds also showed significant interactions: Withasomnine (1 hydrogen bond, −4.682 Kcal/mol), Withanolide D (3 hydrogen bonds, −4.692 Kcal/mol), Withanolide E (1 hydrogen bond, −3.604 Kcal/mol), Withanolide B (3 hydrogen bonds, −4.528 Kcal/mol), Withanolide A (4 hydrogen bonds, −5.053 Kcal/mol), and Anaferine (2 hydrogen bonds, −3.442 Kcal/mol).

Validation of molecular docking

The co-crystal prebound ligand (FAD 601) in the protein was re-docked to the protein using the same docking protocol, and the obtained confirmation was recorded. The re-docked ligand and cocrystal ligand conformations were compared, which gave an RMSD value of 0.849 Å (Figure 3). The re-docked ligand with the protein has a glide score of −11.309 Kcal/mol.

Figure 3.

Co-crystal ligand (FAD 601) superimposed on itself post docking. The ligand in red if the cocrystal ligand prior to docking, and the ligand in green is the re-docked cocrystal ligand.

Molecular dynamic simulation

On accordant with the docking results, Withaferin A was opted for MD simulation. After carrying out the simulation for a total duration of 100 ns, the stability of the protein (NQO1), ligand (Withaferin A), and the protein– ligand complex was as follows-

Root mean square deviation (RMSD)

RMSD is a quantitative measure which helps us understand the similitude of two superimposed atomic coordinates. In this study, the protein RMSD value ranged between 0.75 and 2.0 Å, with one outlying peak at 2.25 Å indicating that the selected protein is fairly stable throughout the 100ns simulation period. Ligand RMSD is stabilized (Figure 4)

Figure 4.

Protein-ligand RMSD of NQO1.

Root mean square fluctuation (RMSF)

The average deviation of atoms in the protein from their respective reference positions over a certain length of time is shown by the RMSF. With varying high peak at 4.0 Å, the point of contact between amino acids was largely stable between 0.5 and 2.0 Å (Figure 5).

Figure 5.

RMSF of NQO1 protein.

Ligand root mean square fluctuation (RMSF)

The Ligand RMSF is used to characterize the variations in the ligand atom positions, indicating the interaction of ligand fragements with protein. According to ligand RMSF, the atom count was 72, with some fluctuations making it a fairly stable interaction (Figure 6).

Figure 6.

RMSF of ligand (Withaferin A).

Total secondary structure elements

Protein secondary structure elements (SSE) plot indicates the positional changes that have occurred in the protein throughout the simulation period which showed 27.24% helix, 11.03 % strands and total SSE being 38.27% (Figure 7).

Figure 7.

NQO1 protein SSE plot.

Ligand interactions

In the desired trajectory (0.00 to 100.00 ns), amino acid residue TRP (106), PHE (105) and HIS (11) established hydrogen bonds with Withaferin A molecule, while PRO (102), ILE (192) and ILE (111) established hydrophobic bonds for almost 30% of the simulated time (Figure 8). A timeline represents the interactions and contacts between protein and ligand molecule. The top panel indicates total number of specific contacts the protein has made with the ligand over the total period of the trajectory. The bottom panel indicates which of the protein residues interact with the ligand in each trajectory frame (Figure 9).

Figure 8.

Protein-ligand contact.

Figure 9.

Timeline representation of protein-ligand contact.

Chemical absorption, distribution, metabolism, excretion, and toxicity (ADME-T) report

Qikprop module of Schrodinger suite was utilised for ADME study, all the ligands obeyed Lipinski’s rule of 5 except Viscosalactone B where the total number of hydrogen bond acceptors were exceeding by 1 and all of the ligands were found to have more than 75% of human oral absorption capacity (Table 5). Furthermore, SWISSADME tool was also used along with Schrodinger suite to predict the gastro intestinal (GI) absorption and blood–brain barrier permeability of the selected ligands. BOILED-Egg plot was generated where the dots in the white region suggests that the ligand has high probability of being passively absorbed by the GI-tract and the dots in the yellow region (yolk) suggests the ligands can penetrate through brain-blood barrier. (Figure 10), the plot suggests that Withaferin A was in the white region it can be absorbed into the GI tract. The toxicity prediction done using Protox-II webserver based on the lethal dosage calculations, these compounds were further classified according to their toxicity, which ranged from Class II to IV, with LD50 (lethal dosage) values between 7 mg/kg to 800 mg/kg,the results of which are tabulated in (Table 6).

Table 5.

ADME property prediction of the selected ligands.

Ligands	% oral absorption	Molecular weight	Donor hb	Acceptor hb	logP	GI absorption	BBB
Withaferin A	86.413	470	1	9.4	0.354	High	No
Withanolide A	94.524	470.6	2	8.5	0.568	High	No
Withasomnine	100	184.24	0	1	0.327	High	Yes
Anaferine	79.743	223.34	0	3	0.217	High	Yes
Viscosalactone B	76	488.61	2	11.1	0.137	High	No
Withanone	97	470.6	2	8.5	0.571	High	No
Withanolide E	100	486.6	3	9.25	0.398	High	No
Withanolide B	100	454.6	1	7.75	0.691	High	No
Withanolide D	90.387	470.6	2	9.45	0.425	High	No

Figure 10.

Boiled egg plot of the selected phytochemicals. 1-Withaferin A, 2- Withanolide A, 3- Withasomnine, 4- Anaferine, 5-Viscosalactone B, 6-Withanone, 7-Withanolide E, 8- Withanolide B, 9- Withanolide D.

Table 6.

Toxicity prediction of the Ligands.

Ligand	Toxicity class	LD50 mg/kg	Hepatotoxicity	Immunotoxicity	Cytotoxicity	Carcinogenicity	Mutagenicity
Withaferin A	3	300	Inactive	Active	Active	Inactive	Inactive
Withanolide A	2	34	Inactive	Active	Active	Inactive	Inactive
Withasomnine	4	800	Inactive	Inactive	Inactive	Active	Inactive
Anaferine	4	338	Inactive	Inactive	Inactive	Inactive	Inactive
Viscosalactone B	4	333	Inactive	Active	Active	Inactive	Inactive
Withanone	2	7	Inactive	Active	Active	Inactive	Inactive
Withanolide E	2	7	Inactive	Active	Active	Active	Inactive
Withanolide B	2	34	Inactive	Active	Active	Active	Inactive
Withanolide D	3	300	Inactive	Active	Active	Inactive	Inactive

Discussion

One of the most severe diseases in modern times, cancer claims many lives each year. Science has advanced to the point where several medications, in addition to diagnostic techniques, have been developed to control specific cancers and partially helped. ⁶² An alternate strategy is obviously required, given the side effects of traditional chemotherapeutic regimens followed by control cancer. Biologically active compounds that are naturally occurring molecules often function by controlling molecular pathways linked to the development and spread of cancer. Increase in antioxidant status, inactivation of carcinogens, prevention of proliferation, apoptosis causing cell cycle arrests, and immunoregulation are a few of the specific ways. ⁶³

W. somnifera, has a history of medicinal use in treating various illnesses such as, asthma, diabetes, ulcer, fever, hepatitis, eyesores, arthritis, heart problems, and hemorrhoids. The alkaloids present in the plant are known to help in bone strengthening and lower back pain treatment. Along with its antifungal, antibacterial activity, ⁶⁴ the plant also possesses anti-cancer activities. In a study involving cell lines extracts of W. somnifera demonstrated apoptosis against MDA-MB-231 human breast cancer cells. ⁶⁵ Root extracts of W. somnifera were reported to have anticancer activity against MCF-7 breast cancer cell lines in an in vitro study conducted. ⁶⁶ Nael Abutaha, et al. (2015), reported that the crude extracts of W. somnifera demonstrated potent tumoricidal activity against HepG2 cell lines, further the dichloromethane fractions also reported similar results against HepG2 cells which caused shrinkage and contraction of cytoplasmic components in addition to loss of cellular integrity. ⁶⁷ A former in-vitro study involving cytotoxic effects of the withanolide glycosides extracted from W. somnifera, Withagenin A diglucoside, exhibited promising cytotoxic activity mediated by extrinsic and intrinsic pathways, Withagenin A diglucoside increased the expression of caspase proteins and decreased the Bcl-2. ⁶⁸ Subsequently a study reported Withaferin A was found effective against several cancer cell lines including DLBCL cell lines LY-3, LY-10, SudHL-6, a Burkitt’s lymphoma, Raji, and a mantle cell lymphoma, it caused cell cycle arrest at G2/M phase in SudHL-6 cells, apoptosis in diffuse large B cell lymphoma. ⁶⁹ In research conducted employing Withaferin A along with TT fields as a treatment to Glioblastoma, it was demonstrated that this combination notably inhibited the growth of GBM cells (U87-MG, GBM2, and GBM39). ⁷⁰ Another in-vitro study confirmed Withaferin A prevented the growth of MCF-7 and MDA-MB-231 breast cancer cells by causing hyperpolarization in mitochondrial membranes. ⁷¹

In the current study, the phytochemicals selected from W. somnifera were evaluated against the NAD(P)H: quinone oxidoreductase (NQO1) protein, which is in charge of both quinone (s) detoxification and bioactivation. It is strongly associated with several carcinogenic pathways and aberrantly overexpressed with poor prognosis in various types of cancers. NQO1 upregulation shields cells from oxidative stress and several other harmful quinones. ⁷²

DFT establishes the physical basis for numerous chemical concepts arising from the change in electron density in atoms and molecules. It is an effective method because it fosters straightforward manipulation, quick computation, and an interface for depicting various interactions and vibrational modes, among other things.^73,74 CDFT, a branch of DFT is sketched by the introduction of several global and local reactivity descriptors. We have done the CDFT analysis to check the reactivity of the selected ligand molecules. The response of the system can be determined by calculating the chemical reactivity descriptors. The global reactivity descriptor electronegativity (χ) is introduced as the ability of the atom in a molecule to attract shared pair of electrons towards itself. The term µ is defined as the negative of χ. Having large electronegativity value of the ligand molecule Anaferine indicates high affinity to captivate the electrons. On the other hand, another descriptor, namely, global hardness (η) defines the stability of the system which is basically the energy gap between HOMO and LUMO orbitals of that particular ligand molecule. Comparing the η value for all the ligands, it can be observed that η value was highest (6.46325 eV) for Anaferine indicating the least reactivity of the ligand. Conversely, Withanolide B possesses the highest reactivity (η = 4.91411 eV) compare to all selected ligands. The global softness (S) holds inverse relationship with the η value. Another response function ω is based on the lowering of energy associated with the maximal flow of electron between two species. Through comparing the electrophilicities of several compounds used in Diels-Alder reactions, a theoretical range of values was proposed to categorize them from strong to moderate such as: ω > 1.5 eV for the first, 0.8 < ω < 1.5 eV for the second, and ω < 0.8 eV for the final instance. ⁷⁵ From the values listed in Table 2, it can be observed that except Withasomnine, all the ligand molecules act as potent electrophiles as well as strong anticancer agents. The molecules with a higher tendency to accept the fractional amount of charge are associated with larger ω⁺ values, whereas a higher tendency to donate a fractional amount of charge is reflected in higher values of ω⁻. Comparing ω⁺ for the selected phytochemicals, the tendency for electron acceptance is found to be highest for Withanolide B and least for Withasomnine. Conversely, the ligand Withaferin A is highly capable for electron donation based on its high value of ω⁻. PASS prediction ascertained the predictive bioactive nature of the molecules, indicating a good predictive anticancer effect of Withaferin A.

Although Withaferin A and Viscosalactone B were identified to demonstrate good interactions with NQO1 protein, Withaferin A was further evaluated using MDS. Also, Withasomnine and Withanolide E exhibited the least number of interactions, with just one hydrogen bond. Withanolide E had the lowest glide score of −3.662 kcal/mol. The preciseness of observed interactions is usually dependent on the formation of hydrogen bonds between the ligand and protein amino acid residues. These interactions were found to be in the range of moderate to strong in accordance with the docking glide score and MD simulation, including RMSD and RMSF results. The molecular dynamic simulations of the withaferin A-NQO1 complex revealed a stable protein-ligand complex throughout the duration of the simulation, in an in-silico screening method using CMap technique, among all other drugs, it was reported that Withaferin A, was a potential anti–lung cancer and anti–lung cancer stem-like cell agent which was further validated by SRB assay. ⁷⁶

The ADME and gastrointestinal absorbability of phytochemicals suggests that the selected molecules were orally bioavailable with very high possibilities of intestinal absorption and was further validated by the boiled egg analysis data, in contrast to a clinical trial study that took place with the patients having advanced stage high-grade osteosarcoma, where Withaferin Awas administered to the patients, it was reported that the molecule was well absorbed with very minute side effects, and was not found in the circulation of the patients, this study reported that Withaferin A showed low oral bioavailability. ⁷⁷

Conclusion

The in-silico evaluations suggest that Withaferin A, a phytochemical in W. somnifera, can be a potential anticancer drug with further validation by cell culture and animal studies. Computational findings indicate good binding affinity and interaction of this molecule with the NQO1 protein, along with the prediction of low toxicity of the compound, which could facilitate the development of novel plant-based anticancer drug with lower side effects. The DFT-based approach has been used to compute the stability and reactivity of the ligand molecules. CDFT and frontier molecular orbital analyses predict the highest stability of the Anaferine ligand, having an energy gap of 6.46 eV, whereas the ligand Withanolide B, having the lowest ∆E value (4.91 eV), shows the highest reactivity among all other ligand molecules considered in this study.