4.1. Molecular docking
The nineteen structures were selected by docking sudies and represented in Figure 3. The compounds belonged to the different classes of natural products such as alkaloids (NP-19, NP-4, NP-15, NP-20), flavonoids (NP-5, NP-7, NP-9, NP-3), quinones (NP-6, NP-10), coumarins (NP12, NP-13, NP-18), and some other natural products-derived.
Among alkaloids, compound NP-20 (ZINC08765174) possessed the best docking score of -11.488 Kcal/mol due to hydrogen bond and π-π stacking interactions of Gly143, Cys145, Glu146, Hie41 (Figure 4), and Glide score -11.488 kcal/mol, and IFD scores -673.42 kcal/mol against the protease. This compound was also the most active one among the entire present compounds in the databases. Free binding energy was calculated for the best docking score results by MM/GBSA. The free binding energy of NP-4 was the lowest (-46.14 kcal/mol) in this class of natural products.
As mentioned before, some flavonoids have been identified as inhibitors of the coronavirus; four of them were shown in Figure 3. Among the flavonoids, NP-3 (ZINC96114284) was the best inhibitor of the main protease with docking score of -10.743 Kcal/mol. This ligand interacted with the active sites of the main protease, including Cys145, Gly143, and Thr190 through hydrogen bond interactions with the docking score value of -10.743 Kcal/mol, Glide score of -10.752 Kcal/mol, and IFD score of -669.73 Kcal/mol. The free binding energy of 3 compounds in this group was lower than -36 Kcal/mol.
Five compounds comprised quinone and coumarin skeletons out of the identified hit compounds. NP-6 (ZINC77257242) is an anthraquinone derivative which showed docking score of -10.102 Kcal/mol due to hydrogen bond interactions of the ligand with Glu166, Gln186, Gly143, Cys145, Thr26, and Hie26 amino acids. The Glide score and IFD score values of this ligand were -10.279 and -666.89 Kcal/mol, respectively. The compound NP-10 (ZINC77257242) possessed the lowest free binding energy of -66.04 Kcal/mol and the amount of the free binding energy for other compounds in this class was lower than -36 Kcal/mol.
Some other natural products-derived have been identified as hit compounds. The compound NP-1 (ZINC03839114) was showing hydrogen bond interactions with Gln189, Glu166, His163, Leu141, and Cys145 amino acids with the docking score of -11.132 Kcal/mol, Glide score of -11.132 Kcal/mol, and IFD score of -674.28 Kcal/mol. The free binding energy was calculated for these compounds and it was lower than -36 Kcal/mol for all of the ligands in this class. The free binding energy of the compound NP-2 (ZINC03841676) was the lowest (-53.07 Kcal/mol).
Among the entire compounds, the free binding energy of 7compounds, NP-1, NP-2, NP-4, NP-8, NP-10, NP-11, and NP-16 showed lower amounts than the others. The free binding energy of the compound NP-10 (ZINC06092274) was the lowest (-66.04 Kcal/mol). Eventually, 21 compounds showed free binding energy of lower than -36 Kcal/mol. The results of molecular docking and ligand-protein interaction for the references such as peramivir, laninamivir, and hydroxychloroquine were introduced in Figure 7.
4.2. Absorption, Distribution, Metabolism, and Excretion (ADME)
Drug likeness of 36 natural products (data of all 36 compounds have been listed in supplementary Table 1) was predicted by the QikProp panel of Schrödinger maestro suite. Lipinski’s rule of five was used to test the bioavailability characteristics (ADME) of these compounds. According to the rule, molecules with Molecular Weight ≤ 500, hydrogen bond donors ≤ 5 and acceptors ≤ 10, calculated octanol-water partition coefficient, and log P ≤ 5 possess good membrane permeability [20]. In this study, these properties for the lead compounds were estimated.
All compounds possess at least one hydrogen bond donor and four hydrogen bond acceptors. For pharmacokinetic property, polar surface area (PSA), central nervous system (CNS) activity, and percent oral absorption were predicted. Results of in silico pharmacokinetic study on these natural products showed good pharmacokinetic properties.
The hydrophilicity of all compounds was determined by calculating the log P-value. It has been suggested that the log P value must be less than five, and high log P results in poor absorption. According to this study, the partition coefficient (QPlog Po/w) was within the permissible range for the selected natural products. Calculated PSA was within the range of 7.0-200.0 Å. CNS activity of these natural products was evaluated. According to the results, all the selected natural products were classified as CNS-inactive compounds.
The human oral absorption percentage of 19 compounds was in the appropriate range of 81 to 100%. Compounds NP-2, NP-5, NP-8, NP-11, NP-13, NP-18, and NP-20 showed 100% oral absorption. Also, entire hit compounds showed over 50% oral absorption. It can be observed in Table 1 that no violations of Lipinski’s rule (polar surface area, molecular weight, number of hydrogen donors and acceptors) were found for hit compounds and all mentioned properties were within the allowed range, thereby indicating their potential as a drug-like molecule.
Aqueous solubility (logS) has ever been one of the most significant factors in the brain/blood partition coefficient and percent human oral absorption. Compounds distribution of HTVS results are based on the relations between brain/blood, octanol/water partition coefficient, aqueous solubility, and percent absorption were represented in Figure 8. Approximately, 99% of HTVS results possessed predicted QPlogBB permissible range (-3 to 1.2), 70% of compounds had up to 80% human oral absorption and about 99% were within the considered range of predicted QPlogPo/w (-2 to 5). Increasing in the QPlogPo/w values of the compounds causes increasing in the absorption percentage. Also, direct relations between QPlogPo/w, QPlogBB, and absorption percentge, and opposite relation between QPlogPo/w and QPlogScan be seen.