Computational and pharmacophore-based study of Camellia sinensis phytochemicals targeting BRAF in melanoma

Abstract

Melanoma is the deadliest kind and most invasive form of skin cancer caused by exposure to UV light and mutation in BRAF signaling protein. Plant-derived phytochemicals have several anticancer properties and used as therapeutic option since ancient era. The purpose of this research is to look at their potential for BRAF protein inhibition in melanoma through molecular docking, pharmacophore modeling, pharmacokinetics, molecular dynamic stimulation and quantum mechanics studies using the phytochemicals from green tea (Camellia sinensis). Drug discovery research was done through in series of computational approach, where protein BRAF (PDB ID: 4MNF) preparation, compounds retrieval and pharmacology properties analysis were done. To verify the stability and affinity of protein–ligand binding, post-docking, pharmacophore modelling, and dynamics stimulation were carried out in the virtual screening programs including AutoDock Vina, MOE, SwisDock and Schrodinger. During the molecular docking experiment of 248 phytochemicals of tea, theaflagallin exhibited a greater binding potential (− 10.8 kcal/mol) compared to the control drug plixorafenib (− 11 kcal/mol) and other top two compounds epigallocatechin 3-O-cinnamate (− 10.3 kcal/mol), and epicatechin gallate (− 9.8 kcal/mol). Post-docking validation confirmed robust ligand binding with accurate pose predictions, while structure-based pharmacophore modeling of BRAF identified key hydrogen-bonding, aromatic, and hydrophobic features essential for effective inhibition by the studied phytochemicals. The chosen phytochemicals are found to be non-toxic, anti-carcinogenic, anti-mutagenic, anti-neoplastic, TP53 expression enhancer, hepatoprotectant, cardioprotectant and possess drug-like properties based on the pharmacological and biological activity studies. In addition, molecular dynamics simulations over 200 ns demonstrated stable binding of all ligands to BRAF, with theaflagallin exhibiting the highest structural stability and strongest interactions, surpassing the control drug Plixorafenib in compactness, polar surface exposure, and binding consistency. The MM-GBSA free energies for these compounds were − 66.55 kcal/mol, − 74.40 kcal/mol, and − 100.57 kcal/mol respectively. The energy gaps between the HOMO and LUMO provided insight into the reactivity of these compounds. Considering the present results, the reported three phytochemicals studied in this study may be used as potential drugs to cure melanoma in near future after executing in-vitro and in-vivo studies.

Introduction

The most aggressive and lethal type of skin cancer is melanoma. Despite making up only 5% of all skin cancer cases, it is the cause of most skin cancer-related fatalities¹. More than 9,000 people die from skin cancer each year in the US, where it is the leading cause of mortality from cancer and the sixth most common cause of cancer overall. By 2040, it is predicted that there would be 510,000 new cases (about a 50% increase) and 96,000 fatalities (almost a 68% increase) worldwide per year if the current rate of melanoma incidence stays constant².

Cutaneous melanoma is a neural crest neoplasia that occurs when melanocytes undergo malignant transformation. However, the pigment that affects skin color, melanin, is produced by melanocytes, which are found in the basal layer of the epidermis. Mutations, aberrations, translocations, and deletions can cause malignant transformation, leading to numerous forms of melanomas³. Melanoma risk factors include a family history of disease, several benign or atypical nevi, and a prior melanoma. Additional risk factors include immunosuppression, sun sensitivity and exposure to UV light⁴. The activation of mitogen-activated protein kinase (MAPK) pathway occurs mainly due to the result of BRAF mutation, which is associated with about 50% of melanomas⁵. As part of the extracellular-related kinase (ERK) pathway, which promotes cell growth and proliferation, BRAF is a proto-oncogene that generates a serine/threonine protein kinase⁶. Point mutations in BRAF usually result in one amino acid substitution, which keeps the pathway in an active state all the time⁷. About half of all metastatic melanomas are caused by the V600E mutation, the most prevalent BRAF mutation that converts valine into glutamic acid^8,9. V600K, which causes valine to become lysine, is the second most frequent mutation. Other uncommon replacements are V600M, V600D, and V600R¹⁰.

Surgical and radiation therapies are standard approaches in cancer treatment; however, they frequently lead to undesirable side effects and challenges like drug resistance. This has led to a rising interest in using natural products as potential therapeutic agents^11,12,13,14. Natural sources with chemopreventive properties are increasingly valued in cancer treatment because they can intervene in the process of carcinogenesis^15,16. Among these natural products, tea is particularly notable for its associated health benefits^17,18,19. It has a wide range of health benefits against many diseases such as obesity, diabetes melitus, cardiovascular disorders etc.²⁰. Tea has been widely studied for the anticancer effects. Numerous epidemiological studies and animal research have shown that green tea and other teas may help protect against several types of cancer, including lung, prostate, skin, liver, and breast cancers²¹. Extensive in vivo and in vitro evidence suggests that tea can inhibit cancer initiation and progression, and its combination with anticancer drugs or other compounds may enhance therapeutic efficacy^{22,23,24,25,26}.

The production of a novel pharmaceutical substance is a laborious and intricate procedure. Finding a suitable lead chemical is a crucial step in the drug development process²⁷. Computational drug discovery approaches have gained much attention in recent times due to their ability to speed up the process in regards of time, cost, and manpower²⁸. In silico analysis predicts the possibility of a compound being formulated as a medication, thereby facilitating the selection of a suitable lead compound²⁹. Numerous recent studies show that BRAF affects both melanoma initiation and development, as well as key tumor maintenance activities. This makes BRAF an important therapeutic target for treating melanoma³⁰. That is why BRAF inhibition is the most opted method for drug development to inhibit melanoma. For the treatment of metastatic or incurable melanoma, the sole FDA-approved medication in the United States is vemurafenib, which inhibits the BRAF protein with mutation V600E³¹. Another BRAF inhibitor, plisoferanib, binds to BRAF in its monomeric state and prevents it from dimerizing³². Due to CYP3A4-dependent metabolism, Plixorafenib must be used in combination with the CYP3A inhibitor cobicistat in = order to achieve effective exposure. In addition, cobicistat co-treatment adversely affects hepatotoxicity and prevents MEK inhibitors from being rationally combined with other drugs, particularly those that are substrates of CYP3A4^33,34. Due to spontaneous mutation of BRAF protein, identification of more inhibitors will be needed to specific binding with selective drug molecules. Besides, existing BRAF inhibitor medications have shown effectiveness, but they also encounter obstacles like drug resistance and negative side effects.

In this study, we investigated the anticancer potential of phytochemicals found in green tea (Camellia sinensis).We initially screened 248 unique phytochemicals derived from plants with known anticancer properties to evaluate their binding affinity to B-Raf kinase (V600E), a key target in melanoma³⁵. Tea-derived compounds, notably theaflagallin, epigallocatechin 3-O-cinnamate, and epicatechin gallate, demonstrated significant inhibitory potential, with theaflagallin achieving an impressive binding affinity of − 15 kcal/mol, comparable to standard drugs. These findings underscore the potential of tea compounds as promising agents in melanoma therapy, supporting the focus on tea in this research. This study also examined the pharmacological properties of these substances, with a focus on toxicity, drug-likeness, bioactivities, administration, distribution, metabolic activity, and excretion. These compounds were further investigated using molecular dynamics simulation studies to identify their anti-carcinogenic properties. The compounds were found to be strong drug candidates with good pharmacological properties and showed minimal adverse effects and toxicity. Findings of this study can contribute to the ongoing research for a cure for cutaneous melanoma, providing a glimmer of hope to the patients suffering from this complex ailment.

Results

Identification of active sites of target protein

Theaflagallin (CID: 73818214), epigallocatechin 3-O-cinnamate (CID: 21629801), epicatechin gallate (CID: 107905), and control drug plixorafenib (CID: 90116675), all had their chemical structures redrawn using ChemDraw professional (version 16.0) (Fig. 1). A pocket-like region made up of several amino acids is an active site of protein; it facilitates appropriate ligand binding and can either boost or hinder the activity. Active sites 1 (AS1) and 2 (AS2) are the two active sites of the BRAF enzyme (PDB ID: 4MNF). The eleven amino acids in AS1 are illustrated in Fig. 2 including CYS532, TRP531, PHE583, ALA481, ASN580, LEU514, THR529, ASP594, GLU501, and VAL471. In contrast, active site AS2 contains just one amino acid: LYS591.

Fig. 1

Chemical structures of (A) Theaflagallin (B) Epigallocatechin 3-O-cinnamate (C), Epicatechin gallate, (D) control drug Plixorafenib and (E) Co-crystalized ligand (2-{4-[(1E)-1-(hydroxyimino)-2,3-dihydro-1H-inden-5-yl]-3-(pyridin-4-yl)-1H-pyrazol-1-yl}ethanol) by using ChemDraw, version 22.0 (https://revvitysignals.com/products/research/chemdraw).

Fig. 2

Active site illustration of the BRAF protein (PDB ID: 4MNF) of co-crystalized ligand via Biovia Discovery Studio Visualizer v21 software (https://discover.3ds.com/discovery-studio-visualizer-download).

Molecular docking analysis

The Camellia sinensis plant’s 248 compounds are docked with the BRAF protein, and the ligands that had the most affinity for binding with the BRAF protein were chosen. Using the, Autodock Vinav1.2.5, MOE v23(Molecular Operating Environment) and SwissDock v21, 248 phytochemicals were molecularly docked with the BRAF protein. All 248 chemicals have binding affinities ranging from − 7.5 to − 10.8 kcal/mol (Table S1). Three phytochemicals epicatechin gallate, epigallocatechin 3-O-cinnamate, and theaflagallin that had higher binding affinities were selected for additional study. A zero (0) RMSD was used to calculate the docking scores of the selected compounds and the control medication.

Among all compounds have shown a greater binding affinity near the control drug plixorafenib, Plixorafenib, a well-known BRAF inhibtor that is often used to treat melanoma, serves as the study’s control drug. The binding affinities (− 11 kcal/mol) of plixorafenib and selected phytochemicals are shown in Table 1.

Table 1 Non-bonding interactions between BRAF V600 and the top 3 phytochemical compounds and control drug.

Protein–ligand interactions interpretation after molecular docking analysis

It was investigated how each distinct ligand interacted with the target protein BRAF using the Biovia Discovery Studio Visualizer. With certain amino acid residues, theaflagallin forms four typical hydrogen bonds such as LYS483, ILE527, THR529, and CYS532 with BRAF. The necessary protein and the amino acid TRP531 have formed three Pi-Pi T-shaped bonds. Furthermore, as shown in Fig. 3A and Table 1, four Pi-alkyl bonds have been formed at the locations of ILE463, VAL471, ALA481, and LYS483. One typical hydrogen bond is formed by epigallocatechin 3-O-cinnamate at the CYS532 location, according to an evaluation of the characteristics of its interaction with BRAF. In PHE583 and TRP531, two Pi-Pi T-shaped bonds were discovered. Three Pi-Alkyl bonds have been observed in LYS483, ALA481, and VAL471, as shown in Fig. 3B and Table 1.

Fig. 3

Figure indicate visualize binding modes of the selected three compounds including (A) theaflagallin (CID: 73818214), (B) epigallocatechin 3-O-cinnamate (CID: 21629801), (C) epicatechin gallate (CID: 107905), control drug (D) plixorafenib (CID: 90116675) and co-crystalized ligand (E) 2-{4-[(1E)-1-(hydroxyimino)-2,3-dihydro-1H-inden-5-yl]-3-(pyridin-4-yl)-1H-pyrazol-1-yl}ethanol (CID: 11717001) against BRAF (PDB ID: 4MNF) protein. All 3D figures were generated by using Biovia Discovery Studio Visualizer v21 (https://discover.3ds.com/discovery-studio-visualizer-download).

At SER536, SER465, ASP594, and CYS532 positions, epicatechin gallate displays four typical hydrogen bonds. Furthermore, five pi-Alkyl bonds are seen at places ILE463, ALA481, VAL471, ILE527, and ILE527. A pi-pi T-shaped bond is also produced at position PHE383. The interactions stated above are shown in Table 1 and Fig. 3C. Four typical hydrogen bonds are present at places in the control drug plixorafenib including GLY534, ASN580, SER465 and CYS532. At position 527, the amino acid isoleucine makes a single bond with a fluorine halogen atom and at position PHE583, a solitary pi-pi T-shaped bond was found. Furthermore, at locations LYS483 (3.96 Å), ALA481 (4.10 Å), and LEU514 (5.34 Å), three pi-alkyl bonds have formed (Fig. 3D and Table 1). Finally, the co-crystalized ligand formed hydrogen bond via GLU501 and GLN530 residues while VAL471, ALA481, LYS483, LEU514, ILE527, TRP531 and CYS532 residues contributed to hydrophobic bond formation (Fig. 3E and Table 1). By comparing interaction profiles, we observed several common interactions with key residues including SER465 and CYS532 by forming hydrogen and ILE527, PHE583, LYS483, and ALA481 residues by forming hydrophobic bonds that suggest a similar binding pattern of compounds and control drug. These interactions may be critical for the activity of the drugs and could indicate a shared mechanism of action.

Validation of docking protocol via self-docking

To assess the reliability of the molecular docking protocol, a self-docking validation was performed by re-docking each compound into the binding site of its respective target protein and comparing the docked poses with the original co-crystallized ligands. As illustrated in Fig. 4, the co-crystallized ligands are shown in green, while the re-docked compounds are shown in gray. Docking accuracy was evaluated based on the root-mean-square deviation (RMSD) between the docked pose and the crystallographic conformation. Theaflagallin, Epigallocatechin 3-O-cinnamate, Epicatechin gallate, and Plixorafenib achieved excellent agreement in their best-docked conformations, with RMSD values of 0.00 Å, indicating highly accurate pose prediction. Acceptable poses were observed for all compounds with RMSD values between 2.434 and 2.851 Å. In contrast, the least accurate poses had RMSD values exceeding 3.0 Å, with Plixorafenib showing the highest deviation at 5.758 Å. These results confirm that the docking protocol is generally robust and reliable for predicting ligand binding modes, particularly for the top-ranked poses.

Fig. 4

Self-docking validation of four ligands including Theaflagallin, Epigallocatechin 3-O-cinnamate, Epicatechin gallate, and Plixorafenib. Co-crystallized ligand conformations are displayed in green, and the corresponding re-docked compounds are shown in gray color. Each row represents a pose category based on RMSD: Good Pose (< 2.0 Å), Acceptable Pose (2.0–3.0 Å), and Bad Pose (> 3.0 Å). RMSD values are labeled above each pair. The 3D figures were generated by using PyMOL molecular graphics systems, version 1.7.4.5 Edu (https://pymol.org/2/).

Structure based pharmacophore modelling of BRAF protein and four ligands

For therapeutic targeting, particularly in melanoma, the BRAF pharmacophore model (PDB ID: 4MNF) in Fig. 5A demonstrates the crucial structural characteristics of co-crystalized ligand molecule needed for efficient further ligand binding and inhibition. Important secondary structures including α-helices (red ribbons) and β-sheets/loops (yellow ribbons) are present in the binding pocket, which is created by the kinase domain and provides a clearly defined scaffold for ligand interaction. The pocket’s aromatic residues (F1:Aro to F6:Aro) stabilize the ligand by π-cation interactions and π-π stacking. Furthermore, to improve the binding, hydrogen bond donor residues engage specifically with the ligand’s hydrogen bond acceptor sites (F2:Acc and F3:Acc, represented by blue mesh spheres).

Fig. 5

Structure-based pharmacophore model of BRAF protein structure with showing (A) key co-crystalized ligand–protein interactions generated by LigandScout, version 4.5 (https://ligandscout.com/). Aromatic features (orange spheres) and hydrogen bond acceptor features (blue mesh spheres) highlight the critical interaction points within the protein binding site. Protein secondary structures are shown as ribbons (yellow and red). Besides, F1, F4, F5, and F6 represents aromatic functional groups while F2, and F3 represents hydrogen bond acceptor functional groups. (B) Epicatechin gallate, (C) Epigallocatechin 3-0- cinnamate, (D) Theaflagallin, and (E) Plixorafenib (Control drug). Left panels show 3D pharmacophore features such as H-bond donors (green), acceptors (red), aromatic rings (purple), and hydrophobic regions (yellow). While right panels depict 2D feature interactions.

On the other hand, pharmacophore models were generated to understand the key molecular features responsible for the interaction of selected ligands with the active site of the BRAF kinase. The ligands Epicatechin gallate, Epigallocatechin 3-0-cinnamate, Theaflagallin, and Plixorafenib a known BRAF inhibitor (Standard Drug) were analyzed based on their 3D alignment with pharmacophoric features such as hydrogen bond donors (HBD), hydrogen bond acceptors (HBA), aromatic rings (AR), and hydrophobic regions (HY). The control ligand displayed a well-balanced pharmacophore consisting of HBA, HBD, and AR, consistent with known ATP-competitive BRAF inhibitors. In addition, Theaflagallin (Fig. 5B) demonstrated the highest pharmacophoric profile with multiple hydrogen bond donors and acceptors distributed across the scaffold, along with aromatic and hydrophobic groups that align well with known BRAF interaction patterns. Similarly, Epicatechin gallate (Fig. 5C) also showed a configuration, although with fewer hydrophobic features. In contrast, Epigallocatechin 3-0-cinnamate (Fig. 5D) displayed a less dense pharmacophore model, with moderate hydrogen bonding potential and limited aromatic overlap, suggesting lower specificity for the BRAF binding site. Besides, the control drug Plixorafenib also showed key features including two hydrogen bond acceptors, multiple hydrophobic regions, and aromatic rings. Interactions via fluorine and sulfonamide groups indicate strong binding affinity, supporting Plixorafenib’s effectiveness as a BRAF inhibitor (Fig. 5E). The development of targeted treatments to treat BRAF-driven cancers is guided by these characteristics, which specify the pharmacophoric profile required for effective BRAF inhibition by our reported phytochemicals.

Physiochemical and drug-likeness properties prediction

The compound’s drug-likeness provides insight into its potential as a therapeutic molecular candidate. Table 2 shows that the compounds theaflagallin, epicatechin gallate, and epigallocatechin 3-O-cinnamate all successfully met the drug-likeness criteria. Except for one nonconformity for theaflagallin and epicatechin gallate, the top three compounds also met Lipinski’s rule of five. Furthermore, plixorafenib, theaflagallin, epigallocatechin 3-O-cinnamate, and epicatechin gallate all had a decent bioavailability score of 0.55.

Table 2 Physiochemical and drug-likeness properties of top 3 phytochemicals of Camellia sinensis.

Pharmacokinetics properties prediction

Conducting an examination into the pharmacokinetic (PK) properties is essential to determine the pharmacological potential of prospective phytochemicals. Table 3 displays the characteristics of absorption, distribution, metabolism, and excretion. The top three medication candidates, theaflagallin, epigallocatechin 3-O-cinnamate, and epicatechin gallate, have favourable absorption values and demonstrate strong Caco-2 permeability. Furthermore, they have satisfactory excretion and metabolism characteristics.

Table 3 Pharmacokinetics profile of the top three potential candidates derived from SwissADME, admetSAR, and pKCSM web servers.

Toxicity properties prediction

Utilizing in-vitro, in-vivo, and in-silico toxicity prediction techniques, the level of toxicity of the suggested medications is ascertained. These studies used admetSAR, pKCSM, and ProTox-II online servers as indicated in Table 4. Prior to being used as a potential therapy, the FDA advises that all biomolecules undergo testing to ensure hERG safety. Hazardous cardiac rhythm abnormalities have been associated with the obstruction of hERG channels. The cardiac toxicity of the top three drug candidates theaflagallin, epicatechin gallate, and epigallocatechin 3-O-cinnamate is predicted using the pred-hERG web tool. The pred-hERG model predicted that all phytochemicals were non-cardiotoxic. In addition, CLC-Pred is an online program that use the structural formula of cancer and non-transformed cell lines to predict the cytotoxicity of different chemical substances. Table S2 presents the computational forecast of the toxicity of cancer cell lines. CLC-Pred predicts that the top three drug candidates, theaflagallin, epigallocatechin 3-O-cinnamate, and epicatechin gallate, have modest efficacy as a control agent, plixorafenib, against several melanoma cell lines including M19-MEL, SK-MEL-1, and A-375.

Table 4 Prediction of toxicity profile analysis of reported three compounds through ProTox-II server.

According to toxicity profiles, the top three phytochemicals are non-toxic, non-carcinogenic, free of Ames toxicity, skin sensitivity, hepatotoxicity, immunotoxicity, nephrotoxicity, cytotoxicity, and cardiac toxicity.

Determination of substance activity spectra

PASS Online, established by Way2Drug in Moscow, Russia, can predict over 4000 biological activities. Table 5 displays the most favorable outcomes obtained from the expected actions of the top three medication candidates. The molecule is very likely to display these characteristics if the probable activity (Pa) values are greater than Pa > 0.5 and the probable inactivity (Pi) scores are close to 0. According to PASS analysis, our drug candidates all chosen phytochemicals Theaflagallin, Epigallocatechin 3-O-cinnamate, and Epicatechin gallate exhibit TP53 expression enhancement, antimutagenic, anticarcinogenic, chemopreventive, antineoplastic, hepatoprotective, cardioprotective, and other various drug-like activities, as detailed in Table 5.

Table 5 Estimation of activity spectra for substances (PASS) through Way 2 Drug server.

Biological activity of the drug candidates

The Molinspiration Chemoinformatics tools were used to calculate the anticipated bioactivity scores of theaflagallin, epigallocatechin 3-O-cinnamate, epicatechin gallate, and the control medication plixorafenib. The top three drug candidates show action in the presence of an intracellular signaling regulator G-protein-coupled receptor (GPCR) ligand, kinase inhibitors, and protease inhibitors. Additionally, plixorafenib is also active, as shown by the bioactivity score in Table 6. This research also discovered that the leading therapeutic candidates have similar dynamic and high-scoring properties as the control compound, plixorafenib. The levels of enzyme inhibition indicate the activity of both the proposed treatment and the control group. Drugs also have a physiological impact and possess physiological activity.

Table 6 Biological activity score of the top screened phytochemicals and control drug.

Quantitative structure activity relationship (QSAR) analysis via PIC50

Table 7 summarizes the QSAR model results, which show important chemical descriptors associated with biological activity. With greater Chiv5 (3.841) and bcutm1 (4.618) values, as well as notable contributions from MRVSA9 (32.713) and MRVSA6 (71.572), CID:90116675 has the lowest projected activity (pIC50 = 5.78) among the compounds. In contrast, even though its structural properties are similar, CID:73818214 has the highest activity (pIC50 = 4.83). The significance of descriptors like polar surface area (PEOEVSA5), GATSv4, and molecular diameter in affecting inhibitory potency is highlighted by these findings, which offer important information for ligand structure optimization in drug design.

Table 7 Quantitative structure activity relationship (QSAR) analysis of four reported phytocompounds.

Molecular dynamics simulation analysis

Using a 200 ns MD simulation, we examined the intricate structures of the three compounds that were chosen for our investigation based on molecular docking, structure-based pharmacophore model, QSAR model and other properties analysis. Finding out how well-suited they were to the protein, and its active site cavity was the aim of this investigation. The MD simulation results have been characterized by RMSD, RMSF, SASA, MolSA, PSA intramolecular bond, and MM/GBSA analysis.

Root mean square deviation analysis (RMSD)

The RMSD analysis of the BRAF protein in complex with Epicatechin gallate (Fig. 6A) over a 200 ns simulation showed initial fluctuations followed by stabilization around 3.0–4.5 Å, indicating sustained structural stability. Subcomponent analysis such as Cα, backbone, sidechain, and all heavy atoms also demonstrated consistent behavior, with the Cα RMSD remaining below 3.0 Å. In contrast, the Plixorafenib (Control) complex (Fig. 6B) stabilized within 50 ns and maintained lower fluctuation over a 200 ns simulation, with RMSD values of 3.5–4.5 Å. Both complexes exhibited stable conformations, although epicatechin showed slightly greater flexibility over a longer trajectory. In addition, the RMSD analysis for the BRAF-Epigallocatechin 3–0-cinnamate complex (Fig. 6C) across 200 ns revealed that the protein reached equilibrium after 200 ns, maintaining RMSD values between 2.5 and 3.5 Å. Protein subcomponents, including the backbone, Cα atoms, sidechains, and all heavy atoms, remained stable with minimal deviations throughout the simulation. The ligand RMSD fluctuated around 4–6 Å, indicating consistent binding with moderate flexibility. In the case of the BRAF- Theaflagallin complex (Fig. 6D), protein RMSD stabilized quickly (100 ns) and remained within 2.0–3.0 Å. Notably, the backbone RMSD reached slightly higher values (3.5–4.5 Å), while the Cα and sidechain components showed steady trajectories. Besides, Theaflagallin RMSD remained below 5 Å, suggesting stable binding over the entire simulation.

Fig. 6

Graphs illustrate the RMSD values of the BRAF protein (PDB ID: 4MNF) and four ligands over 200 ns simulation run. (A) BRAF-epicatechin gallate, (B) BRAF-plixorafenib (control), (C) BRAF-epigallocatechin 3-0- cinnamate, and (D) BRAF-theaflagallin complexes. The X-axis on the left indicates protein RMSD while the Y-axis on the right indicates ligand RMSD value. Blue lines show total BRAF RMSD; orange, green, red, and purple lines represent RMSD values of all heavy atoms, backbone, Cα atoms, and sidechains, respectively, were generated by RStudio Version: 2025.05.1-513 (https://posit.co/download/rstudio-desktop/).

Besides, the ligand trajectories were monitored at 50, 100, 150, and 200 ns, and the spatial positioning within the binding pocket was evaluated. As shown in Fig. 7A–D, all ligands remained stably bound to the catalytic site throughout the simulation, with no significant dissociation or displacement. At 200 ns, the ligands were still firmly localized within the active site, confirming the stability of the ligand–protein interactions and supporting the docking predictions. Based on the RMSD analysis, Theaflagallin appears to be the best ligand in terms of complex stability with BRAF protein.

Fig. 7

Time-dependent binding stability of protein–ligand complexes including (A) epicatechin gallate, (B) plixorafenib (control), (C) epigallocatechin 3-0- cinnamate, and (D) theaflagallin complexes during 200 ns MDS. Snapshots at 50, 100, 150, and 200 ns are shown from left to right figures were generated by Maestro 2025.1 in Schrodinger (https://www.schrodinger.com/products/maestro). Ligands are highlighted with yellow circles, demonstrating their persistent localization within the active site throughout the simulation, thereby confirming stable binding at 200 ns.

Root mean square fluctuation analysis (RMSF)

The RMSF graph illustrates the root-mean-square fluctuation (RMSF) values for four compounds including Epicatechin gallate, Epigallocatechin 3-0-cinnamate, Theaflagallin, and Plixorafenib (Control)) across various protein residues in Fig. 8. RMSF values provide insight into the movement of specific regions within the protein during the simulation, with higher values signifying greater fluctuation.

Fig. 8

Graphs represent the Root Mean Square Fluctuation (RMSF) analysis of BRAF protein residues bound to various ligands over a 200 ns MD simulation and figure was generated by RStudio Version: 2025.05.1-513 (https://posit.co/download/rstudio-desktop/). The graph compares the atomic fluctuations of residues for complexes with Epicatechin gallate (green), Epigallocatechin 3-O-cinnamate (black), Theaflagallin (purple), and Plixorafenib (red, control). Notable residue fluctuations are annotated, particularly around residues 624–633, which indicate significant ligand-induced dynamic changes.

Across the simulation, most regions of the protein displayed RMSF values below 4 Å, indicating relatively stable secondary structures. However, a pronounced spike in flexibility was observed between residue indices 624–633 for control and theaflagallin complexes, with notable variations in intensity among the other ligands. The theaflagallin-bound complex showed the highest fluctuation in this region, peaking at GLN628 with an RMSF value above 8 Å, and showing elevated mobility in adjacent residues including VAL624, ILE625, ARG626, MET627, ASP629, ASN631, and LYS630. Similarly, the Plixorafenib-bound (control) complex exhibited elevated fluctuations at PRO632 and surrounding residues such as ASP629, LYS630, ASN631 and TYR633, but the peak amplitude was slightly lower than that of theaflagallin. In contrast, the epicatechin gallate and epigallocatechin 3-O-cinnamate complexes maintained comparatively lower RMSF values in the same region, with fluctuations remaining mostly below 3 Å. Outside the 624–633 region, all complexes exhibited minor fluctuations in flexible loop regions such as residues near 479, 539, 659, and 689, but these were generally consistent across all ligands. The reduced RMSF in the epicatechin-based ligand complexes suggests enhanced structural rigidity and ligand-induced stabilization of the protein backbone, particularly in the active or functionally important regions.

Radius of gyration (Rg) and solvent accessible surface area (SASA) analysis

In (Fig. 9A), the graph illustrates the radius of gyration (Rg) measured over time for four different compounds: CID: 107905, CID: 21629801, CID: 73818214, and a control compound, CID: 90116675. The radius of gyration analysis was carried out to understand how tightly the protein–ligand complexes stayed folded during the 200 ns molecular dynamics simulations. All compounds showed stable Rg values between 4.0 and 4.5 Å, suggesting that the overall protein structure remained compact and did not undergo any major unfolding. Among the compounds, Plixorafenib (control, 4.0 Å) and Epicatechin gallate (4.2 Å) stayed slightly more compact, while Theaflagallin (4.3 Å) and Epigallocatechin-3-O-cinnamate (4.5 Å) showed marginally higher values, reflecting subtle differences in flexibility caused by the bound ligands.

Fig. 9

Dynamics plot of (A) radius of gyration (Rg) and (B) solvent-accessible surface area (SASA) for the BRAF protein in complex with Epicatechin gallate (green), Epigallocatechin-3-O-cinnamate (black), Theaflagallin (purple), and Plixorafenib (red, control) over 200 ns molecular dynamics simulations and figures were generated by RStudio Version: 2025.05.1–513 (https://posit.co/download/rstudio-desktop/).

The Solvent Accessible Surface Area (SASA) values over a 200 (ns) period for four compounds shown in (Fig. 9B). The solvent-accessible surface area analysis provided how much of the protein surface remained exposed to the solvent over the course of the simulation. Theaflagallin had the smallest SASA (55 Ų), indicating a tightly packed structure, followed by Epicatechin gallate (85 Ų) and Plixorafenib (100 Ų). Epigallocatechin-3-O-cinnamate had the highest SASA (110 Ų), pointing to a slightly more open and flexible conformation. However, the SASA results were consistent with the Rg analysis, confirming that all complexes remained stable with only minor differences in compactness and flexibility.

Analysis of polar surface area (PSA) and molecular surface area (MolSA)

The polar surface area was monitored to evaluate the stability of polar regions and their exposure to the solvent throughout the 200 ns simulation in Fig. 10A. All four protein–ligand complexes showed stable PSA values with only minor fluctuations. Among the ligands, Theaflagallin maintained the highest PSA (360 Å²), followed by Epicatechin gallate (300 Å²) and Epigallocatechin-3-O-cinnamate (290 Å²), indicating greater polar surface exposure. Plixorafenib (control) exhibited the lowest PSA (190 Ų), reflecting a more compact arrangement of polar regions within the protein structure.

Fig. 10

Dynamics plot of (A) polar surface area (PSA) and (B) molecular surface area (MoISA) for the BRAF protein in complex with Epicatechin gallate (green), Epigallocatechin-3-O-cinnamate (black), Theaflagallin (purple), and Plixorafenib (red, control) during 200 ns molecular dynamics simulations and figures were generated by RStudio Version: 2025.05.1-513 (https://posit.co/download/rstudio-desktop/).

On the other hand, MoISA analysis was used to assess the molecular surface characteristics and conformational stability of the complexes. The values remained largely constant over the simulation period, confirming the absence of significant structural rearrangements (Fig. 10B). Epigallocatechin-3-O-cinnamate (385 Ų) and Plixorafenib (375 Ų) had the highest MoISA values, suggesting a slightly larger molecular footprint, while Epicatechin gallate (345 Ų) and Theaflagallin (330 Ų) displayed comparatively lower surface areas, consistent with a more compact structure. These results confirm that all protein–ligand complexes maintained stable surface properties throughout the simulation.

Intramolecular bonds analysis

Figure 11 illustrates the detailed interaction profiles of the protein–ligand complexes, including hydrogen bonds, ionic interactions, hydrophobic (non-covalent) contacts, and water bridges, over a 200 ns molecular dynamics simulation. A higher interaction fraction indicates a more persistent contact throughout the simulation, which is often associated with stronger and more stable ligand binding. Among the selected compounds, Theaflagallin (CID: 73,818,214) exhibited strong and sustained hydrogen bonding with Thr529 and Ser539, showing interaction fractions of 1.5 and 1.6, respectively. Epicatechin gallate (CID: 107905) formed notable hydrogen bonds with Gln530 and Cys532, with interaction fractions of 0.8 and 1.3, indicating stable interaction patterns. Likewise, Epigallocatechin 3-O-cinnamate (CID: 21629801) formed hydrogen bonds with Thr529, Gln530, and Cys532, with interaction fractions of 0.9, 1.0, and 0.5, respectively. In comparison, the control drug Plixorafenib (CID: 90116675) showed interactions at different residues, but the selected phytochemicals formed more frequent and diverse bonds including hydrogen, hydrophobic, ionic, and water bridge interactions. These bonds were maintained consistently throughout the 200 ns simulation period, indicating robust and stable binding interactions with the target protein.

Fig. 11

The bar charts represented the protein ligands interactions through numerous types of bonds at 200 ns simulation running time and figures were generated by Maestro 2025.1 in Schrodinger (https://www.schrodinger.com/products/maestro). Our reported compounds including (A) theaflagallin (CID: 73818214), (B) epicatechin gallate (CID: 107905), (C) epigallocatechin 3-O-cinnamate (CID: 21629801), and (D) control drug plixorafenib (CID: 90116675).

Analysis of thermal (MM/GBSA)

The “Post Simulation Thermal MMGBSA” chart illustrates the binding free energy profiles for several phytochemicals (CID: 73818214, CID: 21629801, and CID: 107905) and a control drug (CID: 90116675) across different interaction types: van der Waals, solvation, lipophilic, covalent, hydrogen bonding, and Coulomb shown in (Fig. 12). All selected phytochemicals show similar predicted thermal binding free energies to the control, with total binding energies ranging close to the control’s − 102.32 kcal/mol. epicatechin gallate (CID: 107905) exhibits the most comparable binding energy to the control at − 100.57 kcal/mol, indicating a similarly strong binding affinity. The other phytochemicals, theaflagallin (CID: 73818214) and epigallocatechin 3-O-cinnamate (CID: 21629801), display slightly weaker binding energies (− 66.55 and − 74.40 kcal/mol, respectively), reflecting a moderate but still effective binding potential (Table 8). Each compound’s energy profile suggests that these phytochemicals have interaction characteristics comparable to those of the control.

Fig. 12

Thermal (MM/GBSA) analysis of control drug plixorafenib (CID: 90116675), and three components theaflagallin (CID: 73818214), epigallocatechin 3-O-cinnamate (CID: 21629801), epicatechin gallate (CID: 107905) indicated by grey, yellow, blue and green colours, respectively, and the figure was generated by Microsoft Excel 365.

Table 8 Negative energy calculation of four compounds in kcal/mol via thermal (MM/GBSA) analysis.

Quantum mechanics calculation

In the DFT calculations, the compounds theaflagallin (CID: 73818214), epigallocatechin 3-O-cinnamate (CID: 21629801), epicatechin gallate (CID: 107905), and control drug plixorafenib (CID: 90116675) yielded HOMO and LUMO energy values of − 0.27598 and − 0.19555, − 0.29366 and − 0.20923, − 0.29730 and − 0.17020, and − 0.29704 and − 0.18897 a.u., respectively shown in (Fig. 13 and Table 9). For the HLG, hardness, and softness energies, CID: 90116675 (control) and CID: 73818214 showed values of 0.10807 and 0.08043, 0.05403 and 0.04021, and 18.50823 and 24.86943 eV, respectively (Table 9). Additionally, epigallocatechin 3-O-cinnamate and epicatechin gallate produced HLG, hardness, and softness values of 0.08443 and 0.1271, 0.04221 and 0.06355, and 23.69106 and 15.73564 eV, respectively (Table 9). The DFT calculations indicate that CID: 73818214 has the lowest HOMO–LUMO gap (HLG) and greatest softness, suggesting it is the most reactive of the compounds. In contrast, CID: 107905, with the highest HLG and hardness, appears to be the least reactive. The control compound, CID: 90116675, demonstrates moderate reactivity and stability, providing a practical reference point.

Fig. 13

The structural visualisation and energy values of HOMO and LUMO of the control drug (A) Control drug plixorafenib (CID: 90116675) and selected three components (B) theaflagallin (CID: 73818214), (C) epigallocatechin 3-O-cinnamate (CID: 21629801), and (D) epicatechin gallate (CID: 107905). All figures were generated by using GaussView 6.0.16. software (https://gaussian.com/gaussview6/).

Table 9 The control medication and three chosen compounds’ energy of HOMOs, LUMOs, gaps, hardness, and softness (all measured in Hartree units).

Discussion

Globally, the incidence of melanoma has increased dramatically during the last few decades³⁶. As frequency of skin cancer increasing and the elevated ratio of adverse effects induced by commercialized anticancer medication, it is urgently needed to develop natural anticancer drugs with fewer side effects. An earlier investigation revealed that cytidine analogues with antibacterial and anticancer activities were shown in silico studies³⁷. This current study employs publicity available techniques to evaluate bioactivity, drug-likeness, molecular docking and simulation of tea-derived compounds. The objective is to determine the potential therapeutic and anticancer activity of these compounds against melanoma and assess their suitability as biomolecules. Through the exclusion of biomolecules that may have an unfavourable pharmacological profile, the use of in silico interpretations of pharmacological spectra has not only aided in the first phases of research but has also shown a novel direction towards the most promising regions³⁸. While our study focuses specifically on BRAF V600E, we recognized that melanoma could harbour other clinically relevant BRAF mutations, such as V600K, V600R, and non-V600 variants. These mutations, though less common, exhibit distinct biological behaviours and may show differential responses to targeted therapies. For instance, BRAF V600K mutations are associated with older age, chronic sun damage, and unique clinical outcomes compared to V600E³⁹. Additionally, rare non-V600 mutations, such as K601E and L597Q, have demonstrated sensitivity to MEK inhibitors rather than BRAF inhibitors⁴⁰. While the scope of our current analysis does not include these variants, the mechanistic insights and therapeutic implications described here may provide a foundation for their investigation. Future studies should aim to extend this approach to include diverse BRAF mutations, enabling more comprehensive and personalized therapeutic strategies in melanoma.

A molecular docking method predicts the probable affinities and binding interactions between a ligand and a target⁴¹. Due to their structural diversity, biological activity, and advantageous pharmacokinetic characteristics, phytochemicals derived from several medicinal plants are an important part of current drug design and development with provide an array of lead molecules for pharmaceutical research^42,43. We selected 248 compounds from the Camellia sinensis plant and performed docking experiments with the BRAF enzyme (PDB ID: 4MNF) in our research. The control medication, plixorafenib, was employed to compare the outcomes of the major phytochemicals obtained from Camellia sinensis. Only three of the 248 phytochemicals were chosen as possible therapeutic candidates because they exhibited a strong binding affinity and met the Lipinski criteria. These compounds will now undergo further investigation to validate their pharmacological characteristics. According to Table 2, the top-ranked drug candidates, including theaflagallin, epigallocatechin 3-O-cinnamate, and epicatechin gallate have binding scores of − 11.1, − 10.3, and − 10 kcal/mol, respectively. With a binding affinity of 10.6 kcal/mol, theaflagallin binds to the BRAF enzyme more strongly than Plixorafenib control drug. Studies show that plixorafenib and other BRAF-containing dimer selective inhibitors are still effective against BRAF mutations, including V600 and non-V600 alterations⁴⁴. El Badri et al. stated that plixorafenib was less effective against BRAF V600E and G469A mutant rats, which was consistent with reduced in vitro pERK inhibition and reported in vivo metabolic vulnerability⁴⁵.

Additionally, other FDA-approved targeted therapies like vemurafenib and dabrafenib have transformed melanoma treatment, they come with certain limitations and can cause side effects, including skin rash, fatigue, and cardiovascular issues⁴⁶. Phytochemicals, particularly those derived from natural sources such as green tea and black tea, present potential advantages with multi-target approaches⁴⁷. In contrast, FDA-approved drugs like vemurafenib, dabrafenib, and others drugs are designed to target specific molecular alterations, which may limit their effectiveness against diverse cancer cell populations⁴⁸. From toxicological analyses of our proposed drug candidates also indicate they are non-toxic and safe for human use. In addition, the top 3 drug candidates showed a higher binding affinity than vemurafenib (− 10.5 kcal/mol) and a comparable binding affinity to dabrafenib (− 11.1 kcal/mol).

Figure 4 illustrates how strong hydrogen and hydrophobic bonds are formed during molecule interactions, which is the explanation behind this. Many hydrophobic atoms in the targeted molecules can enhance the drug-target interaction⁴⁹. The specificity of ligand binding depends on the presence of hydrogen bonds⁵⁰. According to our interaction study, hydrophobic and hydrogen bonding play a crucial role in designing a drug against BRAF protein. In our study, we identified several key active site residues in the BRAF protein. Specifically, our findings highlighted the presence of VAL471, ALA481, LEU514, THR529, TRP531, and CYS533. These residues are essential to BRAF function, according to earlier study⁵¹. The C-lobe of the BRAF V600E receptor (residues 535-717) oversees substrate protein binding, whereas the N-lobe (residues 457-530) contains the ATP binding motif. According to recent studies, substances that selectively attach to the BRAF V600E residues through ILE463, VAL471, ALA481, LYS483, LEU514, ILE527, THR529, TRP531, CYS532, HIS539, LYS578, ASN581, and PHE583 are essential for inhibiting BRAF V600E and its anti-proliferative effects⁵². It has been demonstrated that medications that bind to the BRAF V600E receptor within certain residue regions impair its function. Black tea contains a polyphenolic component called theaflagallin, which has drawn interest because of possible anticancer effects⁵³. Our best compound theaflagallin forms four conventional hydrogen bonds with specific amino acid residues such as CYS532, THR529, ILE527, and LYS483 with BRAF V600E protein. Among them, THR529, ILE527, and LYS483 are in the region that engage ATP binding site while CYS532 is slightly outside the ATP binding site. Transketolase and glucose-6-phosphate dehydrogenase activity were both suppressed by epicatechin gallate, a potentially effective chemotherapeutic drug for the treatment of cancer⁵⁴. In our analysis, epicatechin gallate exhibits four conventional hydrogen bonds at the following SER465, CYS532, SER536 and ASP594 residues in both ATP and outside ATP region. According to Su BN et al. study, epigallocatechin 3-O-cinnamate can prevent the development of cancer by inhibiting the MCPyV LT protein function⁵⁵. Epigallocatechin 3-O-cinnamate with BRAF has been assessed and it has been discovered that it forms a single conventional hydrogen bond at the CYS532 residue in the C-lobe region (outside of ATP binding site). Similar in vitro research by Lambo et al. showed that ASP479, ASN512, TRP531, LEU577, THR529, CYS532, GLU533, and ASP587 residues, together with thiazole-2-carboxylic acid, cinnamic acid, theanine, and protocatechuic acid generated strong hydrogen bonds with the BRAF V600E binding pocket⁵⁶. Besides, the self-docking results also demonstrated that the docking protocol could reliably reproduce experimentally observed binding poses⁵⁷. Top-ranked poses of Theaflagallin, Epigallocatechin 3-O-cinnamate, and Epicatechin gallate perfectly matched their co-crystallized conformations (RMSD = 0.00 Å), while others achieved acceptable deviations (2.4–2.9 Å). Although some lower-ranked poses showed higher RMSD values, the protocol proved generally robust for predicting accurate ligand binding modes.

By accurately differentiating between active and inactive drugs, the pharmacophore model showed strong predictive power after calculating the binding pose via molecular docking. Through pharmacophore modeling, our results are consistent with earlier research that highlights the crucial importance of contacts in kinase inhibition. Several recent studies have identified key pharmacophore elements including two hydrogen bond acceptors, one or more hydrogen bond donors, aromatic rings, and hydrophobic regions aligned with the ATP-binding cleft of BRAF protein^58,59. Abdel-Mohsen et al.⁵⁹ demonstrated that the model performed well overall, with an F1 score of 0.502, confirming that it is useful for detecting strong inhibitors of BRAF. Similar study by Dain et al.⁶⁰ found several natural drug candidates through in silico structure based pharmacophore modelling against BRAF mutant protein. In our study, CID: 73818214 provides this pattern robustly, featuring multiple HBD and HBA groups, aromatic rings, and hydrophobic sites, making it a promising candidate. Compounds exhibiting such multivalent pharmacophoric profiles have shown high inhibitory potential in kinase-targeted therapy due to their conformational flexibility and multi-point anchoring according to Zhou et al.⁶¹. The pharmacophore models’ importance in directing the development of selective kinase inhibitors and offering a dependable framework for drug discovery in targeting proteins such as BRAF is further supported by these findings.

Only when chemicals fulfill specific criteria that are comparable to those of pharmaceuticals during the drug development process are they considered candidates for drug candidacy. According to the Lipinski rule of five, a medication taken orally must to comply with at least four regulatory requirements⁶². Epigallocatechin 3-O-cinnamate cannot violate any rule whereas Theaflagallin and Epicatechin gallate violate only one rule which indicates that our top-listed phytochemicals have satisfactory drug-likeness properties and can be used as drugs in the future. MLOGP, an indicator of a compound’s lipophilicity, is generally optimal within the range of − 0.4 to + 5.0 for oral drugs, supporting favorable bioavailability and membrane permeability⁶³. Theaflagallin’s MLOGP of − 0.70 suggests a higher degree of hydrophilicity, which may influence its absorption and distribution properties in vivo, possibly reducing its bioavailability⁶⁴. Although this MLOGP is slightly outside the ideal range, moderate deviations can sometimes be acceptable depending on a compound’s broader pharmacokinetic profile⁶⁵. In contrast, other compounds with positive MLOGP values are closer to the recommended range, potentially indicating better bioavailability and cellular permeability⁶³. Pharmacokinetic studies can be utilized to predict alterations in the clinical characteristics of a certain disease ⁶⁶. Absorption features have been demonstrated in skin permeability, human gut absorption, and Caco-2 cell permeability. If a drug has a caco-2 cell permeability value more than 0.90, it is considered very permeable. The remarkable intestinal permeability exhibited by each phytochemical suggests that their incorporation into pharmaceutical formulations may enhance their absorption⁶⁷. Models of the blood–brain barrier (BBB), central nervous system (CNS), fraction unbound (human), and vector-directed sprays (human) were used to study the distribution of the selected phytochemicals. It is crucial to consider the drug’s distribution across the blood–brain barrier when assessing the safety of adverse responses caused by medications. Hazardous or poisonous chemicals may have more detrimental effects when they enter the central nervous system and cross the blood–brain barrier⁶⁸. BBB permeability of certain phytochemicals is disregarded since BRAF inhibition does not require BBB crossing, even though many drugs require BBB penetration to have pharmacological effect.

In addition, as previously stated, the medication candidates indicated by top are not substrates of P-gp (P-glycoprotein). As a result, P-gp’s efflux mechanism, which cancer cell lines frequently use to acquire drug resistance, has no effect on them⁶⁹. Approximately 75% of the drugs available for sale are digested by these specific isoforms of CYP enzymes, which play a vital part in the elimination of pharmaceuticals from the body. Pharmacokinetics-based drug-drug interactions of significant importance occur when any of these isoforms are blocked^70,71. The CYP enzymes were not affected by any of our top-ranked drug candidates; however, plixorafenib, the control drug, may inhibit all CYP enzymes, indicating that it may cause drug-drug interactions with other medications that target CYP enzymes. Additionally, the three compounds selected for this study demonstrated evidence of renal clearance, a vital aspect of contemporary drug development that reveals how the drug is eliminated from the body following its effects⁷². Identification of toxic compounds, their levels in the human body after administration, and their consequences all depend on toxicological research⁴¹. According to the study, the three most significant phytochemicals did not exhibit any toxicity that would have an impact on human health, suggesting that they might be used safely. The Ames toxicity test identifies the mutagenicity of these chemicals by observing the positive results obtained in the standard assay model⁷³. On the other hand, the AMES test assay produced negative results for all three compounds selected as the most promising treatment options. The LD50 values of the ligands are a crucial factor to consider when evaluating the toxicity of potential treatment options⁷⁴. In addition, the control drug plixorafenib may demonstrate hepatotoxicity, while the top listed drug candidates may not exhibit hepatotoxicity. Additionally, they do not suppress hERG and do not have any negative effects on the heart. Selectivity is indeed crucial for therapeutic efficacy. We propose that future studies include cytotoxicity assays on both melanoma and normal melanocytes or fibroblasts to assess selectivity. Selective cytotoxicity could reduce off-target effects and enhance the therapeutic index. In future study, we will be included an in vitro study into the Shrimp on the importance of this testing and potential methods to evaluate selectivity, such as differential IC50 assessments across cell lines.

Based on the PASS prediction, our most promising therapeutic candidates include a TP53 expression enhancer, antimutagenic, anticarcinogenic, antioxidant, antineoplastic, hepatoprotectant, cardioprotectant, anti-inflammatory, and testosterone beta-dehydrogenase (NADP +) inhibitor. These findings indicate that the chemicals at the top of our list may have the ability to hinder the breakdown of estrogen and androgen. The compounds also demonstrated robust predictions for NADP⁺ scores, corroborating our findings about the molecule’s efficacy against melanoma and its potential utility in future investigations targeting other dermatological disorders⁷⁵. The bioactive compounds Theaflagallin, Epigallocatechin 3-O-cinnamate, and Epicatechin gallate have demonstrated significant anti-carcinogenic potential in both in vitro and in vivo models. Theaflagallin and its derivatives Epitheaflagallin were shown to inhibit the proliferation of human cancer cells by inducing apoptosis and cell cycle arrest⁷⁶. Similarly, derivatives of Epigallocatechin, such as EGCG and related cinnamate conjugates, have demonstrated significant antiproliferative effects on cancer stem cells and induced apoptosis in vitro⁷⁷. Furthermore, Epicatechin gallate is widely recognized for its capacity to inhibit multiple cancer cell types, including prostate, and breast cancers through mechanisms involving apoptosis and oncogenic pathway modulation^78,79. However, these findings validate the potential of these phytochemicals as promising candidates for anticancer therapeutics, supporting their further investigation in molecular and clinical studies. Resistance remains a major challenge for BRAF inhibitors, as adaptive resistance frequently involves MAPK pathway reactivation⁸⁰. Theaflagallin, our lead compound, may target key pathways related to this resistance. Combining theaflagallin with MEK inhibitors could help counteract MAPK reactivation, addressing one of the prominent resistance mechanisms⁸¹. We will be described resistance mechanisms and how dual inhibition with theaflagallin could potentially provide synergistic effects, enhancing outcomes in resistant melanoma cases in our future study.

Molecular dynamic simulation (MDS), an essential phase in the process of creation of novel drugs, estimates the biomolecular interactions between proteins and ligands using dynamic trajectory analysis⁸². Theaflagallin (CID: 73818214), epicatechin gallate (CID: 107905), epigallocatechin 3-O-cinnamate (CID:21629801), and control drug plixorafenib (CID: 90116675) exhibited a stable profile in several simulated trajectories. RMSD is a tool designed to figure out conformational changes and assess the stability of protein structure. Protein–ligand complexes that exhibit an average RMSD value change of 1–3 Å are appropriate for MD simulation⁸³. These values indicate that the protein maintained structural integrity throughout the simulation. The fluctuations observed are consistent with expected dynamic behavior rather than indicative of significant structural deviations. The RMSD stabilized following the initial equilibration phase, indicating that the simulation converged and the stability suggests that the protein attained a steady-state shape. On the other hand, , the RMSD values for two compounds and control drug was suitably matched to the reference value except compound (CID: 21629801) had very minor modifications that contrasted favorably to the natural protein’s structure shown in Fig. 6A–D. Time-dependent binding stability of protein–ligand complexes demonstrating their persistent localization within the active site throughout the simulation, thereby confirming stable binding at 200 ns Snapshots at 50, 100, 150, and 200 ns are shown in Fig. 7. The root mean square fluctuation (RMSF) computation is the primary method used to assess the average fluctuation of proteins during interactions⁸⁴. A lower RMSF value generally indicates a more stable and rigid region of the protein, while a higher RMSF value points to a more flexible and dynamic region, where the residues show greater movement and conformational variability⁸⁵. Compared to the control medication plixorafenib (CID: 90116675), our chosen phytochemical theaflagallin (CID: 73818214) had a lower RMSF, which is depicted in Fig. 8.

The radius of gyration (Rg) depicts the folding and compactness of the protein; to be more explicit, a lower Rg value denotes greater compactness and a higher Rg value denotes less compactness^86,87. The Rg values of all the chosen phytochemicals complexed with the target proteins are displayed in Fig. 9A, showing greater compactness than the control with protein complex. To figure out the dimensional alterations of druggable small molecules along the simulated trajectories, the solvent accessible surface value (SASA) of the protein–ligand complex was computed⁸⁸. A higher SASA value indicates a less stable structure, while a lower value suggests a more tightly packed complex of water molecules and amino acid residues⁸⁹.The SASA findings visualization in Fig. 9B shows that the protein-selected phytochemical complex has a more stable structure than the drug-control complex. Compared to the control medication values, the PSA and MolSA value visualization from Fig. 10A–B had greater potential indication. Using the targeted protein in MD simulation, ligands established a variety of bonds, such as ionic, water bridge, hydrophobic, and hydrogen bonding. These interactions allowed for tight binding among the ligands and the proteins, and they remained throughout the simulation⁹⁰. Figure 11 illustrates the multiple intramolecular bonds that were established by the phytochemicals and the control medication with distinct amino acid residues.

In MMGBSA analysis, the compound with the lowest (most negative) total binding energy score is regarded as having the strongest binding affinity. Its plays a crucial role in drug design by providing an efficient and reliable approach to estimate the binding free energy of a ligand to its target protein. This computational technique integrates molecular mechanics energy terms with solvation and entropy components, enabling a deeper understanding of molecular interactions⁹¹. MM/GBSA calculations (Fig. 12) showed that the three selected compounds had significantly negative binding free energy values, similar to that of the control compound when bound to the protein. This indicates that these compounds interact more strongly with the target protein.

Frontier molecular orbitals, the most significant orbitals for an atom, can be used to assess a compound’s kinetic stability and chemical reactivity. They are called surround molecular orbitals, with the terms highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO)⁹². The energies of the HOMO and LUMO are key factors in assessing a molecule’s reactivity, including its chemical reactivity characteristics⁹³. The difference between the HOMO and LUMO energies, known as the HOMO–LUMO gap, represents the electronic excitation energy^94,95. Compounds with a smaller HOMO–LUMO gap generally have an increased ability to interact with biomolecules, making them more effective in therapeutic applications⁹⁶. Additionally, the gap energy is linked to the hardness and softness attributes of a molecule⁹⁷. In our study, theaflagallin (CID: 73818214) and epigallocatechin 3-O-cinnamate (CID: 21629801) were found to have the highest level of softness and displayed the lowest hardness and HOMO–LUMO gap compared to the control drug shown in Fig. 13. Therefore, our research verified that the top three new compounds theaflagallin, epigallocatechin 3-O-cinnamate, and epicatechin gallate could inhibit BRAF, a potent drug target for Melanoma. Since each candidate complied with all requirements to become a drug, they may all be taken into consideration for doing further research. Specifically, we have suggested potential in vitro methods, such as B16-F10 cell-based assays to study melanoma biology, drug responses, and molecular mechanisms of our computational models. Additionally, we have proposed in vivo approaches, such as genetically eengineered BRAF V600E mice model to further evaluate the melanoma initiation, progression and biological relevance of our findings. Theaflagallin (CID: 73818214) is the most promising option among them after evaluating in vitro and in vivo research aimed at target-specific medication development against Melanoma.

Materials and methods

Phytochemical collection, target selection, preparation, molecular docking, pharmacokinetics, toxicity properties, dynamics simulations and free energy calculation have done by using several online databases. The Fig. 14 represents the overall process of our current study.

Fig. 14

In silico workflow for phytochemical-based drug design targeting BRAF generated by Adobe Illustrator software v24.1 (https://www.adobe.com/products/illustrator.html).

Phytocompounds retrieved and preparation

Phytochemicals that are obtained from naturally occurring medicinal plants can have a wide range of chemical structures that can be used to develop new medications. From Indian Medicinal Plants, Phytochemistry and Therapeutics (IMPPAT) (https://cb.imsc.res.in/imppat/), we get all these compounds. IMPPAT is a manually maintained database that uses cheminformatic approaches to speed up the process of creating pharmaceuticals from natural materials⁹⁸. For virtual screening, the phytochemicals of Camellia sinensis have been identified and separated from the database. Next, the 3D structures and SDF formats of all phytochemicals have been downloaded from the PubChem database, which is publicly accessible at (https://pubchem.ncbi.nlm.nih.gov/)⁹⁹. We used Biovia Discovery Studio Visualizer v21 to combine all of the ligand molecules and save them in an SDF format after obtaining the 2-dimensional SDF format from PubChem database. The combined SDF ligands are then converted to 3D PDB and energy minimization in an Open Babel tool. The MMFF94(s) force field can be used in Open-Babel software to optimize ligand molecule geometry and minimize energy in drug-like compounds¹⁰⁰.

Rationale of protein selection, preparation and active site prediction

From the RCSB Protein Data Bank, the crystal structure of BRAF-V600E bound to GDC0879 was obtained in PDB format (PDB ID: 4MNF) (https://www.rcsb.org/). The protein of interest has a molecular weight of 75.29 kDa and consists of two identical chains, labeled chain A and chain B. In our study, chain A was utilized for further docking procedures¹⁰¹. The protein’s PDB structure was prepared by performing these procedures: Using BIOVIA Discovery Studio Visualizer v19.1.0.18287, the original protein was stripped of unnecessary chains, water molecules, hetatoms, and ligand groups. The energy reduction stage was successfully performed using the Swiss PDB viewer (https://spdbv.unil.ch/). This step is of utmost importance in protein preparation as it ensures optimal binding of ligands at the lowest energy level¹⁰². Active sites of protein were identified by analysis of the binding interaction of co-crystalized ligand molecule in original protein structure (PDB ID: 4MNF) via Biovia Discovery Studio Visualizer v2021 and and SiteMap in the Schrödinger suite was used to confirm it¹⁰³.

Prediction of binding affinity via molecular docking

This study employed molecular docking, a popular virtual screening technique, to predict the binding strength and specific interactions between the target protein and ligands. Using MOE¹⁰⁴, virtual screening was carried out against selected phytochemicals filtering top binding poses according to their pharmacophore fit score¹⁰⁵. AutoDock Vina v1.2.5 was used with default settings, employing a gradient-based search and empirical scoring to estimate binding affinities (kcal/mol)¹⁰⁶. In addition, SwissDock v21 was used to dock top-scoring molecules using the EADock DSS engine, which applies an evolutionary search algorithm and CHARMM-based scoring. Docking was performed in accurate mode, and poses were evaluated based on full fitness and estimated ΔG¹⁰⁷. The default grid box parameters were set and the exhaustiveness value of 50 was used to identify the optimal binding pose for each ligand in different software. Complexes with the most negative docking scores and strongest binding affinities were considered the most promising candidates. Finally, BIOVIA Discovery Studio Visualizer v21 was used to analyze and visualize the protein–ligand interactions.

Validation of docking protocol via self-docking

The co-crystalized ligand 2-{4-[(1E)-1-(hydroxyimino)-2,3-dihydro-1H-inden-5-yl]-3-(pyridin-4-yl)-1H-pyrazol-1-yl}ethanol (CID: 11717001) from the crystal structure (PDB ID: 4MNF) was extracted and re-docked into the binding site using MOE¹⁰⁵, AutoDock Vina¹⁰⁶ and SwissDock¹⁰⁷ software’s. The docked pose was aligned with the original crystallographic ligand in PyMol¹⁰⁸ and the RMSD between them was calculated using the built-in ligand RMSD tool. An RMSD value below 2.0 Å was considered acceptable, confirming the reliability of the docking protocol⁵⁷.

Structure based pharmacophore modelling of BRAF protein

To clarify important interaction characteristics necessary for ligand binding, the BRAF protein’s structure-based pharmacophore modeling was carried out. The Protein Data Bank (PDB ID: 4MNF) provided the crystal structure of BRAF and the AMBER force field in AMBER Tools was used to improve the protein structure after hydrogen atoms were introduced. LigandScout⁴³ was used to create pharmacophore features, such as hydrophobic regions (HyPho), aromatic moieties (Ar), hydrogen bond donors (HBD) and hydrogen bond acceptors (HBA). These features were then improved based on interactions seen in native inhibitors as co-crystalized ligand. A training set of active and inactive chemicals was used to validate the model, and metrics like sensitivity, specificity, and enrichment factor were evaluated. Then our selected four phytocompounds were incorporated individually into the LigandScout for alignment purpose to generate the shared feature pharmacophore model.

Drug-likeness and pharmacokinetics properties analysis

The drug-likeness of the compounds was assessed using Lipinski’s Rule of Five, which evaluates key parameters such as molecular weight, lipophilicity (LogP), hydrogen bond donors (HBD), and hydrogen bond acceptors (HBA). This analysis was performed using the SwissADME platform (http://www.swissadme.ch)¹⁰⁹. Additionally, the pharmacokinetic profiles including gastrointestinal absorption, blood–brain barrier (BBB) permeability, interactions with cytochrome P450 enzymes, and other properties crucial for evaluating the drug-like potential of the compounds were predicted using the same tool.

Toxicity properties analysis

Toxicological profiles indicate that a medicine is safe and effective regardless of whether it has any harmful effects on the human or animal body. In order to prevent any negative effects, evaluating the toxic impact is a highly beneficial phase in the drug research process¹¹⁰. Therefore, the admetSAR, pKCSM, and ProTox-II web servers were utilized to determine the selected phytochemicals’ toxicity profile^111,112. There is evidence that blockage of the hERG K + channel causes fatal cardiac arrhythmias. Therefore, in this study pred-hERG 4.2 online server (http://predherg.labmol.com.br/ ) was used to predict cardiac toxicity for early identification of suspected hERG blockers and non-blockers¹¹³. The Cell Line Cytotoxicity Predictor, also referred to as CLC-Pred, is an online program that predicts the cytotoxicity of different chemicals based on the structural formula of cancer and non-transformed cell lines¹¹⁴. In this study, the CLC-Pred tool (http://www.way2drug.com/cell-line) is used to forecast the cytotoxicity of the top-listed phytochemicals on cell lines. Probable activity (Pa) and probable inactivity (Pi) grades reflected the expected output activity. It was assumed that Pa > 0.5 represented the most activity, Pa < 0.3 represented the lowest activity, and Pa > 0.3 represented the moderate activity¹¹⁵.

Estimation of activity spectra for substances (PASS)

The PASS prediction was made using the free web application Pass Online (http://www.pharmaexpert.ru/passonline). For a particular compound, only actions with Pa > Pi was considered feasible. Pa levels between 0.5 and 0.7 showed a moderate chance of experimental pharmacological action, whereas Pa values higher than 0.7 indicated a high potential. If Pa was less than 0.5, there was very little chance of pharmacological activity¹¹⁶.

Biological activity of the potential drug candidates

Molinspiration Chemoinformatics technologies were used to identify G protein-coupled receptors (GPCRs), nuclear receptors, ion channel modulators, kinase inhibitors, protease inhibitors, and enzyme inhibitors characteristics (https://www.molinspiration.com/cgi/properties). As per Roy’s research results, a chemical was classified as highly active if its value was more than or equal to 0.00 (~ 0); moderately active if its value lay between 0.50 and 0.00; and inactive if its value was less than or equal to 0.50 (< 5.0)¹¹⁷.

Quantitative structure activity relationship (QSAR) analysis via PIC50

An extensively utilized quantum chemistry method for drug research and discovery is the Quantitative Structure–Activity Relationship (QSAR). To assess chemical bioactivity and forecast the effectiveness of novel treatment options, QSAR employs molecular structure analysis. The Chemdes website (http://www.scbdd.com/chemdes/) and the multiple linear regression (MLR) standard equation were used to calculate the pIC50 and QSAR values¹¹⁸. The free Chemdes database provided the required information, which included Chiv5, MRVSA9, and PEOEVSA5. The QSAR and PIC50 calculations for the reported ligand were then completed, and an Excel spreadsheet was created using multiple linear regression (MLR). ‘bcutm1’ stands for burden descriptors, whereas ‘MRVSA9,’ ‘MRVSA6,’ and ‘PEOEVSA5’ are MOE-type descriptors included in the data set. An autocorrelation descriptor is referred to as “GATSv4”. ‘J’ and ‘diametert,’ the last two variables, are proposed as topological features for biological compound^119,120.

Molecular dynamics simulation (MDS) studies

The selected ligand molecules’ conformational behavior and protein stability were examined using a dynamics simulation. Following an investigation of binding affinity, drug likeliness, and pharmacokinetic features, three compounds theaflagallin (CID: 73818214), epigallocatechin 3-O-cinnamate (CID: 21629801), epicatechin gallate (CID: 107905), and a control drug plixorafenib (CID: 90116675) were chosen for MDS. The Maestro 2025.1 Premium in Schrodinger was used to execute a 200 ns MDS in a Linux environment (Ubuntu 22.04) to ascertain the binding consistency of the protein–ligand complex structures. Using this framework, the form of the orthorhombic periodic bounding box is partitioned via a space of beyond 3 Å (water boundary) in a preset volume using the established TIP3P aqueous method. To counterbalance the electric power inside its structure, suggested ions, including 0.15 M Na⁺ and Cl^- have been chosen and dispersed at random across its solvent surroundings. A cut-off distance of 12 Å is usually used to calculate the electrostatic interactions in order to minimize computational costs¹²¹.

After building its solvency proteins structures containing agonist combinations, the system’s framework subsequently reduced as well as comfortable utilizing the protocol carried out applying force field constants OPLS 2005 included inside the Desmond package¹²². A single atmospheric pressure of 101325 bar and an overall Nose–Hoover temperature combination of 310 K were maintained for each Isothermal-Isobaric ensemble (NPT) assembly, which employed an isotropic technique and 50 PS grabbing pauses with an efficiency of 1.2 kcal/mol. The fidelity of the MD simulation was evaluated throughout the entire simulation using the Simulations Interaction Diagram (SID) from the Desmond modules of the Schrödinger suite.By employing data from root mean square deviation (RMSD), root mean square fluctuation (RMSF), radius of gyration (Rg), solvent accessible surface area (SASA), intramolecular bond analysis, polar surface area (PSA), and molecular surface area (MolSA), it was possible to assess the relative solidity of specific protein–ligand interaction combinations¹²³.

Free energy (MM/GBSA) calculation

To approximate the complex free energy throughout the 200 ns running duration, the molecular mechanics-generalized born surface area (MM/GBSA) was computed using the thermal_mmgbsa.py python program provided by Schrödinger, which processes a Desmond trajectory file, divides it into discrete snapshots, performs Prime-MMGBSA calculations on each frame, and produces the average computed binding energy¹²⁴. The MM-GBSA calculation was performed using the OPLS3e force field in combination with the default prime protocol¹²⁵. Solvent models are also taken into account, along with the electrostatic and van der Waals interactions. The following equation was used to calculate the binding free energy in MM-GBSA:

$$\Delta {\text{G}}_{{{\text{bind}}}} = {\text{ G}}_{{{\text{complex}}}} - \, \left( {{\text{G}}_{{{\text{receptor}}}} + {\text{ G}}_{{{\text{ligand}}}} } \right)$$

$$\Delta {\text{G}}_{{{\text{bind}}}} = \, \Delta {\text{E}}_{{{\text{MM}}}} + \, \Delta {\text{G}}_{{{\text{GB}}}} + \, \Delta {\text{G}}_{{{\text{SA}}}}$$

$$\Delta {\text{E}}_{{{\text{MM}}}} = \, \Delta {\text{E}}_{{{\text{internal}}}} + \, \Delta {\text{E}}_{{{\text{electrostatic}}}} + \, \Delta {\text{E}}_{{{\text{vdw}}}}$$

where ΔG_bind denotes the total binding free energy during protein–ligand interactions; ΔE_MM signifies the total gas phase energy within the molecular mechanics (MM) force field (OPLS3e), encompassing ΔE_internal from bond, angle, and dihedral terms; ΔE_{electrostatic} and ΔE_vdw represent the electrostatic and van der Waals energies, respectively; ΔG_GB and ΔG_SA refer to the two solvation free energy contributions, specifically the polar electrostatic solvation energy computed via the generalized Born (GB) method and the nonelectrostatic solvation component (nonpolar contribution)¹²⁶.

Quantum mechanics analysis of four compounds

Density functional theory (DFT) was used to implement the Gaussian09W programming package, the Gauss View 6.0 software provided the input files, which were then used to produce the various graphical representations of the compounds⁷⁶. The compounds exhibiting the highest docking score and the compounds found by MDS simulations undergo DFT or QM computations. DFT calculations are typically performed in the gas phase at 0 K. In our study, these calculations were intended to evaluate the intrinsic electronic properties of the ligands, such as frontier molecular orbitals and energy gaps, which are fundamental and independent of solvation to a first approximation. Additionally, thermochemical corrections were applied using frequency analysis at 298.15 K, offering a reasonable approximation to physiological temperature. The DFT was handled using basis sets 6–311 +  + G(d,p) and B3LYP (Becke exchange functional, incorporating Lee, Yang, and Parrs (LYP) correlation function)^29,77. The B3LYP functional paired with the 6–311 +  + G(d,p) basis set serves as a highly effective approach in computational chemistry, enabling detailed studies of complex molecular systems¹²⁷. This approach delivers an optimal combination of accuracy, computational efficiency, and flexibility, making it suitable for various chemical studies¹²⁸. The energies of the surrounding HOMOs and LUMOs were used to calculate the degree of hardness and softness of each chemical. In chemistry, hardness refers to an atom’s resistance to donating or accepting electrons from other atoms or metallic surfaces, whereas softness describes its tendency to participate in such interactions¹²⁹. With the DFT interpretations of Parr and Pearson⁷⁸ and Koopmans’s theorem⁷⁹ in consideration, the hardness (η) and softness (S) were calculated according to the following equation

$$\eta = \frac{{\varepsilon_{{{\text{HOMO}}}} - \varepsilon_{{{\text{LUMO}}}} }}{2}$$

$$s = \frac{1}{\eta }$$

Higher reactivity is associated with a lower hardness (η) and a higher softness (S) value, whereas lower reactivity corresponds to higher hardness (η) and lower softness (S)¹³⁰. Softness (S) represents an atom’s ability to accept electrons, and it is inversely related to hardness (η), meaning that an increase in softness leads to a decrease in hardness, and a higher hardness results in lower softness.

Conclusions

Phytochemicals have been widely acknowledged for their medicinal potential, since several substances produced from plants have been used as the foundation for developing drugs. Three potent compounds were identified from the Camellia sinensis species and assessed for their efficacy against B-Raf kinase, with the goal of creating a successful treatment molecule for melanoma. Molecular docking demonstrates that three phytochemicals including theaflagallin (CID: 73818214), epigallocatechin 3-O-cinnamate (CID: 21629801), epicatechin gallate (CID: 107905) efficiently may be blocked the B-Raf kinase protein receptor by binding in their active sites confirmed via post-valiidation and structure-based pharmacophore model of BRAF protein. Toxicological evaluations also demonstrated that these three compounds exhibit no carcinogenic properties and are well tolerated when administered at appropriate dosages. Furthermore, there were no signs of hepatotoxicity or cardiotoxicity, indicating their potential as efficacious and safe inhibitors of the BRAF enzyme. The stability of these compounds in interacting with the BRAF enzyme was further validated by molecular dynamics simulations, as shown by consistent values of RMSD, RMSF, Rg, SASA, MolSA, and PSA represents the stable binding and no fluctuation during 200 ns simulation run.