The 36 phytoconstituents extracted from Moringa oleifera reported in literature along with dinaciclib (reference ligand) were docked with CDK-2 and were ranked in order of their affinities towards the protein (CDK-2) (Fig. 1). A potent CDK-2 inhibitor dinaciclib was used as a reference ligand. The ligands with potency greater than the dinaciclib were selected for further analysis as shown in Fig. 2.
Overall, only Chlorogenic acid (1), quercetin (2), ellagic acid (3), niazirin (4) and kaempferol (5), out of 36 screened ligands showed strong interactions with the CDK-2 protein with good docking score as shown in Table 1. All ligands including dinaciclib exhibited both polar and nonpolar, hydrogen and hydrophobic interactions, pi-pi stacking, salt bridges as well as cation-π contacts with the target protein residues.
The affinities of ligands ranged from -6,717 to -9.119 kcal/mol. Chlorogenic acid (1), quercetin (2), ellagic acid (3), niazirin (4), kaempferol (5) had affinities -9.119, -7.501, -7.138, -6.835 and -6.717 kcal per mol respectively and greater than the the binding affinity of dinaciclib -5.445 kcal/mol. However, glide ligand efficiency which is percentage/potency efficiency index (PEI) of these compounds ranges from -0.318 to -0.365 and it is higher than -0.188, efficiency of dinaciclib. The efficiency refers to the ability of the compounds to produce desired clinical effects, which could be beneficial during optimization. The ligand efficiency determines the optimal interactions between the ligand and the receptor. It is a ratio of Gibb free energy (∆G) to non-hydrogen atoms of compound . It measures the binding energy of each atom of ligand bound to receptor or enzyme. The difference in binding affinities is because of different functional groups like side chains and hydroxyl groups in the structure of the ligands. In molecular docking analysis ligands were ranked on the basis of their best binding poses and higher affinity values in negative and were further analyzed to determine their stability in the protein binding pocket.
Hydrogen bonding and hydrophobic interactions
Hydrogen bonds and hydrophobic interactions were visualized using PyMOL-2.5.2 software. The dinaciclib reference ligand (a potent CDK2 inhibitor) formed two conventional hydrogen bonds and 16 hydrophobic interactions. Chlorogenic acid (1) showed five conventional and one aromatic H-bond along with 21 hydrophobic interactions. Quercetin (2) exhibited two conventional and one aromatic H-bond and 19 hydrophobic interactions. Ellagic acid (3) showed four conventional and one aromatic H-bond as well as 19 hydrophobic interactions while niazirin (4) showed four conventional hydrogen bonds together with 19 hydrophobic interactions. Kaempferol (5) showed two conventional and three aromatic H-bond as well as 19 hydrophobic interactions. No aromatic H-bonds were observed in niazirin; however, it showed pi-pi stacking between two residues Phe-4 and Tyr-77 as shown in Table 2
Polar and non-polar interactions of ligands
The polar and non-polar interactions were visualized using PyMOL-2.5.2 and key amino acid residues were identified. Ala-31, Ala-144, Gln-85, Gly-11, Gly-13, Lys-33, Leu-134, Phe-80, Phe-82, Val-18 and Val-64 showed non-polar contacts with all ligands while Glu-81 showed non-polar contacts with all ligands except niazirin . Asp-145 showed polar contacts with all the ligands. The reference ligand Dinaciclib showed polar interactions with Gln-131 and Lys-9 while non-polar contacts with Asp-86, Asp-92, Glu-12, Glu-162, Gly-11, Thr-158, Thr-160, Thr-165, Tyr-159, Tyr-168, Trp-167, Lys-88, Lys-89, Lys-129, Val-163 and Val-164 as shown in Fig. 3. Chlorogenic acid (1) showed polar interactions with Asp-145, His-84, Lys-89 and Ile-10 while non-polar contacts with Asn-132, Asp-86, Glu-8, Lys-9, Lys-20, Leu-148 and Leu-298 as shown in Fig. 4a. Quercetin (2) showed polar contacts with Asp-145 and His-84 while non-polar contacts with Asn-132, Asp-86, Gln-131, Ile-10, Leu-83, Lys-89 and Leu-298 as shown in Fig. 4b. Chlorogenic acid (1) and Quercetin (2) are overlapped in the active site of CDK-2 protein as shown in Fig. 5.
Ellagic acid  showed polar interactions with Asp-145, Asp-86 and Leu-83 while non-polar interactions with Asn-132, Gln-131, Glu-12, His-84, Ile-10 and Lys-89 as shown in Fig. 6a. Kaempferol  showed polar interactions with Asp-145 and His-84 while non-polar contacts with Asn-132, Asp-86, Gln-131, Glu-81, Ile-10, Lys-89, Leu-83 and Leu-298 as shown in Fig. 6b. Ellagic acid  and Kaempferol  overlapping at the active site of CDK-2 protein is shown in Fig. 7. Niazirin (4) showed polar interactions with Asp-145, Asn-132, Gln-131 and Lys-89 while non-polar interactions with Asp-86, Glu-8, His-84, Ile-10, Lys-20, Lys-129 and Leu-148 as shown in Fig. 8
The residues with the least interactions include Leu-298, Glu-8, Lys-9, Lys-20 and Leu-148.
Asp-92, Glu-12, Gu-162, Thr-158, Thr-160, Thr-165, Tyr-159, Tyr-168, Trp-167, Lys-88, Lys-129, Val-163 and Val-164 were the residues involved in non-polar interactions in dinaciclib only as mentioned in Table 3.
The binding free energy of the ligand protein complex was evaluated by MM-GBSA method in Maestro 13.2 (Schrödinger, LLC, 2022.2) and the compounds with highest binding free energy in negative value were further analyzed in MD simulations using Maestro-Desmond v12.3 Schrödinger software for evaluating the stability of the complex.
The prime MM-GBSA method was used to calculate the binding free energy of ligands and protein complex. All the docked poses were optimized using OPLS 2005 force field feature in prime and Generalized-Born/ Surface Area continuum solvent model was used to calculate energies of complex.
This analysis revealed the binding energy ∆G of dinaciclib) with CDK-2 protein as -36.21 kcal/mol in comparison with best docked ligand chlorogenic acid (1) -38.16 kcal/mol. Quercetin (2) had binding energy less than dinaciclib (reference) -34.02 kcal/mol. Ellagic acid (3) had the highest negative binding energy as -48.91 kcal/mol. Niazirin (4) had binding energy close to the top hit compound that is -38.65 kcal/mol. The binding energy of kaempferol (5) -31.92. Energy calculation by prime analysis gives relative energies of ligands. The ligands with the highest negative energies and binding affinity as mentioned in Table 4 were further selected for analysis in MD simulation.
Molecular dynamic simulation
Molecular Dynamic (MD) simulations were carried out using Desmond Molecular Dynamics Simulation System to check the stability of ligands and optimization of complexes. The trajectories obtained from simulations were analyzed using RMSF and RMSD. RMSF was calculated to check flexibility of ligand and changes in its conformation upon binding. The RMSD measured the displacement of a selection of atoms over this time period. Molecular dynamic simulations were run at normal pressure and temperature for 100 ns and RMSD plots of complex were generated. The protein and ligand contacts were recorded as fraction plots for binding during the simulation.
The RMSF of ellagic acid is shown in Fig. 9a. The RMSD value of ellagic acid-protein complex was highest 4 Å (Fig. 9b). Asp-145 residue of protein forms H-bonds with ligand for 95% along with bridges of water for 46%, Leu-83 forms H-bonds with ligand for 85% along with bridges of water for 30% and Asp-86 forms H-bonds for 39% along with bridges of water for 30% of the simulation time. Similarly, Val-18, Ala-31, Ala-144, Ile-10 and Leu-134 show hydrophobic contacts for 10–50% of the time. Other amino acid residues for stabilizing the complex forming H-bonds and water bridges include His-84, Asn-132, Ile-10, Lys-89, Lys-33, Glu-81, Glu-12, Thr-14, Gln-85 and Lys-129 (Fig. 9d). Figure 9c shows a schematic diagram of several amino acid residues of CDK-2 involved in interactions with ellagic acid.
The RMSF of chlorogenic acid is shown in Fig. 10a. In chlorogenic acid-protein complex equilibrium of system is attained within one sec and average RMSD value of 3.8 Å (Fig. 10b). Asp-145 residue forms H-bonds with ligand for 85–99% along with water bridges for 34%, Leu-83 and Lys-89 form H-bonds for 94% and 71%. Lys-20, Asp-86 and His-84 form H-bonds and bridges of water for 40%, 34% and 30% of the simulation time respectively. Similarly, Phe-80, Phe-82, Ala-31, Val-64, Leu-134, Ile-10, Ala-144 and Val-18 show hydrophobic contacts for 10–20% of the simulation time. The residues involved in ionic interaction include Lys-89, Ile-10 and Lys-20. Other amino acid residues for stabilizing the complex forming water bridges include Asn-132, Glu-8, Lys-9, Lys-20, Lys-33, Gln-85, and Gln-131 (Fig. 10d). Figure 10c shows a schematic diagram of several amino acid residues of CDK-2 involved in interaction with chlorogenic acid.
The RMSF of quercetin is shown in Fig. 11a. In quercetin-protein complex equilibrium of system is attained within one sec and average RMSD value of 2.7 Å (Fig. 11b). Glu-81 and Asp-86 residues of protein form H-bonds with ligand for 100% and 99% of the time respectively. Asp-145 forms H-bonds for 78% of the time along with water bridges. Lys-33 forms hydrogen bonds for 51–71% of the simulation time. Similarly, Phe-80, Ala-31, Ile-10, Val-18, Phe-82, Leu-83, Ala-144 and Leu-134 show hydrophobic contacts for 10–30% of the total simulation time. Other amino acid residues for stabilizing the complex forming water bridges and hydrogen bonds include Asp-145, Asn-132, Gly-16, Glu-12, His-84, Ile-10, Lys-89, Gln-85, and Gln-131 (Fig. 11d). Figure 11c shows a schematic diagram of several amino acid residues of CDK-2 involved in interaction with quercetin.
ADME (absorption, distribution, metabolism and excretion) Prediction
Top ranked compounds from molecular docking were analyzed for their pharmacokinetic properties using Swiss ADME server. ADME server predicts the physicochemical properties of compounds including molecular weight, partition co-efficient of octanol/water (log Po/w) & relative absorption in intestine included in Table 5.
According to five rules of Lipinski (RO5) of likeliness of drug for consideration in pre-clinical studies a good drug candidate should have a molecular weight less than or equal to 500 Dalton, rotatable bonds should be less than or equal to 10, hydrogen bond acceptors should be less than or equal to 10 in number, also hydrogen bond donors should be less than or equal to 5 in number and log value (P o/w) should be ≤ 5 [40,41,42].
All of the compounds analyzed are good candidates for drug design, follow Lipinski rule and show no violations except chlorogenic acid as shown in Table 5. Chlorogenic acid exhibits one violation. Number of H-bond donors is 6 in chlorogenic acid. Also, it exhibits low GI absorption compared to other compounds. However, it can be enhanced synthetically by modifying physical properties that will improve the permeability, lipophilicity and absorption of the compounds.
The toxicological properties of compounds were predicted by webserver Pro-Tox II with results summarized in Table 6. The oral toxicity of compounds predicted ranged from 159 mg/kg to 5000 mg/kg. Quercetin (159 mg/kg) was the only compound that belonged to toxicity class III, toxic if swallowed. The compound ellagic acid (2991 mg/kg) belonged to toxicity class IV, harmful if swallowed. All the other compounds such as chlorogenic acid (5000 mg/kg), niazirin (3750 mg/kg), and kaempferol (3919 mg/kg) belonged to class v, could be harmful if swallowed. However, none of the screened compounds were predicted in severe toxic class that is fatal (Class I or II). So, they can be used as lead compounds in the treatment of breast cancer.
Effect of Fraction A (petroleum ether) on Mcf-7 cell lines
The results of MTT assay using petroleum ether fraction of Moringa oleifera leaves extract are shown in Fig. 12a. After 24 h, results were compared with control and no significant reduction was observed in cell viability, rather an exposure of 300 µg/mL of extract resulted in an increase in cell viability by 99.9% and other two dilutions (100 µg/mL, 200 µg/mL) also showed enhanced proliferation as 84.6% and 90.7% respectively.
Effect of Fraction B (ethyl acetate) on Mcf-7 cell lines
The results of MTT assay using ethyl acetate fraction of Moringa oleifera leaves extract are shown in Fig. 12b. After 24 h, results were compared with control significant reduction was observed in cell viability at higher concentrations of extract, 40% at 200 µg/mL and 47.7% at 300 µg/mL respectively. However, no significant reduction was observed at 100 µg/mL and cell viability was increased by 84.6%.
Though it was observed that both extracts exhibited slight cytotoxicity after 24 h, while Mcf-7 cells showed enhanced proliferation. The findings of MTT assay indicate that extracts of Moringa oleifera have not shown any significant anticancer effect on Mcf-7 cell lines.
The statistical significance was calculated using one way analysis of variance (ANOVA) and value of P < 0.05 was taken as statistically significant. The p value for petroleum ether data was found to be 0.0024 and for ethyl acetate it was 0.05. So, the statistical results for petroleum ether fraction were found to be more significant.