Multi-Targeted Molecular Docking, Pharmacokinetics, and Drug-Likeness Evaluation of Okra-Derived Ligand Abscisic Acid Targeting Signaling Proteins Involved in the Development of Diabetes

Diabetes mellitus is a global threat affecting millions of people of different age groups. In recent years, the development of naturally derived anti-diabetic agents has gained popularity. Okra is a common vegetable containing important bioactive components such as abscisic acid (ABA). ABA, a phytohormone, has been shown to elicit potent anti-diabetic effects in mouse models. Keeping its anti-diabetic potential in mind, in silico study was performed to explore its role in inhibiting proteins relevant to diabetes mellitus- 11β-hydroxysteroid dehydrogenase (11β-HSD1), aldose reductase, glucokinase, glutamine-fructose-6-phosphate amidotransferase (GFAT), peroxisome proliferator-activated receptor-gamma (PPAR-gamma), and Sirtuin family of NAD(+)-dependent protein deacetylases 6 (SIRT6). A comparative study of the ABA-protein docked complex with already known inhibitors of these proteins relevant to diabetes was compared to explore the inhibitory potential. Calculation of molecular binding energy (ΔG), inhibition constant (pKi), and prediction of pharmacokinetics and pharmacodynamics properties were performed. The molecular docking investigation of ABA with 11-HSD1, GFAT, PPAR-gamma, and SIRT6 revealed considerably low binding energy (ΔG from −8.1 to −7.3 Kcal/mol) and predicted inhibition constant (pKi from 6.01 to 5.21 µM). The ADMET study revealed that ABA is a promising drug candidate without any hazardous effect following all current drug-likeness guidelines such as Lipinski, Ghose, Veber, Egan, and Muegge.


Introduction
Diabetes is one of the most prevalent epidemics, affecting almost 382 million people worldwide. According to the International Diabetes Federation (IDF) report, it is alleged that approximately 1.3 million people die from diabetes every year. IDF suggests that around 629 million people will have diabetes by 2045 worldwide [1].

Results & Discussion
Molecular docking and virtual screening are fast, economical, and reliable approaches for identifying both a potential druggable protein target as well as a novel drug (lead molecule) through rational drug designing (RDD) or computer-aided drug design

Results & Discussion
Molecular docking and virtual screening are fast, economical, and reliable approaches for identifying both a potential druggable protein target as well as a novel drug (lead molecule) through rational drug designing (RDD) or computer-aided drug design (CADD). RDD or CADD is now being used to annotate and evaluate large pharmacological libraries swiftly.
This study applies molecular docking-based virtual screening to identify a promising target for T2DM. Based on literature review and available crystal structures of proteins  Figure 1. The plausible molecular/atomic interactions of ABA with these proteins were investigated in this in silico study.
ABA was found to have a binding energy of −8.1 kcal/mol, −7.3 kcal/mol, −7.3 kcal/mol, −7.3 kcal/mol, −6.8 kcal/mol, −6.6 kcal/mol, −6.6 kcal/mol, 6.3 kcal/mol, and −6.2 kcal/mol, with 11β-HSD1, GFAT, PPAR-gamma, SIRT6, glucokinase, aldose reductase, glycogen synthase kinase-3, Pyruvate dehydrogenase kinase (PDK), and Tyrosine kinase proteins, respectively ( Table 1). The binding energy (Kcal/mol) would be used to correlate and investigate the binding affinity of various ligands or inhibitors with their corresponding protein target. In general, the lower the binding energy, the greater the ligand's affinity for the receptor protein will be. As a result, the ligand with the highest affinity can be taken forward as a candidate for further research. Intracellular conversion of metabolically inert cortisone to active cortisol using NADPH as a co-factor is carried out by the 11β-HSD1 enzyme [12][13][14][15]. Cortisol increases hepatic glucose production by inducing genes involved in gluconeogenesis and glycogenolysis in the liver. Cortisol promotes pre-adipocyte differentiation into mature adipocytes, resulting in adipose tissue hyperplasia. By modulating cortisone/cortisol levels, selective inhibition of this enzyme can be a novel treatment for T2DM and hyperlipidemia [16][17][18][19]. Obesity, diabetes, wound healing, and muscular atrophy are glucocorticoid-related disorders, and inhibiting 11β-HSD1 has many therapeutic values, including T2DM and hyperlipidemia. The X-ray crystallography investigations of the crystal structure of an inhibitor molecule (4aS,8aR)-3-(cyclohexylamino)-4a,5,6,7,8,8a-hexahydrobenzo[e] [1,3,4]oxathiazine 1,1-dioxide (C13H22N2O3S) docked with human and murine 11β-HSD1 proteins revealed that its cyclohexyl-NH interacts with the key active site residue Tyr183 [17]. Furthermore, one of the sulfonyl oxygen atoms forms a hydrogen bond to the main-chain nitrogen of Ala172 of 11β-HSD1 protein. Meanwhile, the side chain Tyr177 of the human 11β-HSD1 enzyme typically forms Van der Waals interaction to the same inhibitor molecule [17].
Interestingly in our in silico analysis, we found that key residues of active site viz.,  (Figure 2A-D). This interaction showed the lowest binding energy of −8.1 kcal/mol and the highest inhibition constant of 6.01 µM, respectively ( Table 1). The pi-pi stacking interaction by LEU171, TYR177, LEU217, ILE218, and ALA223 with other pi-cation and pi-alkyl interactions may help in stabilizing the ABA bounded with the active residues of the 11β-HSD1 enzyme ( Figure 2D). Based on these very similar binding patterns and docking complex analysis, we can say that ABA is the true, potent inhibitor of the human 11β-HSD1 enzyme, thus possibly help in controlling T2DM and hyperlipidemia.

Abscisic Acid Significantly Binds and Inhibits Glutamine: Fructose-6-Phosphate Amidotransferase (GFAT) Effectively
Glutamine-fructose-6-phosphate amidotransferase (GFAT) is a rate-limiting enzyme in the hexosamine biosynthetic pathway a key regulator in T2DM [20][21][22]. In mammals, glucose integration via the hexosamine biosynthetic pathway is regarded as a cellular nutrition sensor. This pathway is one of the strategies by which hyperglycemia induces peripheral insulin resistance [23,24]. In vitro and in vivo studies indicate the association of hyperactivity of human GFAT with insulin resistance, thus qualifying it as a promising candidate for T2DM [25].
We found a similar interaction pattern of ABA with the active site amino acid residues of GFAT (PDB ID-2ZJ4), suggesting ABA can potentially inhibit GFAT activity. Five key amino acid residues of GFAT viz., CYS373, THR375, GLN421, SER422, and THR425 were strongly forming hydrogen bonds of bond length 2.87Å, 3.10Å, 3.08Å, 3.14Å and 2.94 Å, respectively, suggesting a strong binding to the active pocket ( Figure 3A-E).

Abscisic Acid Significantly Binds and Inhibits Glutamine: Fructose-6-phosphate Amidotransferase (GFAT) Effectively
Glutamine-fructose-6-phosphate amidotransferase (GFAT) is a rate-limiting enzyme in the hexosamine biosynthetic pathway a key regulator in T2DM [20][21][22]. In mammals, glucose integration via the hexosamine biosynthetic pathway is regarded as a cellular nutrition sensor. This pathway is one of the strategies by which hyperglycemia induces peripheral insulin resistance. [23,24]. In vitro and in vivo studies indicate the association of hyperactivity of human GFAT with insulin resistance, thus qualifying it as a promising candidate for T2DM [25].
The C-terminal 40 kDa isomerase domain of GFAT (residues Gln313-Glu680) contains the active site (near the CF helix) and is involved in converting fructose-6-phosphate (Fru6P) to glucosamine-6-phosphate (GlcN6P) utilizing ammonia (NH3) as a substrate. The X-ray diffraction pattern shows that the bound ligand AGP   (Table 1), and three alkyl interactions by LEU556, LYS675 and VAL677 residues with (1S)-1-hydroxy-2,6,6-trimethyl-4-oxocyclohex-2-en-1-yl ring helps in stabilizing the ABA bound with the active site. As a result, we may conclude that ABA is an effective inhibitor of the GFAT enzyme.   PPAR-gamma is a major transcriptional factor (TF) that regulates adipogenesis, insulin sensitivity, and glucose homeostasis in humans [26,27]. The drug rosiglitazone, which acts as a ligand of PPAR-gamma, is an excellent insulin sensitizer, improving glucose absorption and lowering hyperglycemia and hyperinsulinemia [28][29][30].
The decreased ability of PPAR-gamma to bind DNA in response to rosiglitazone manifested the receptors' inability to activate transcription. The PPARs are also potential therapeutic targets that could treat atherosclerosis, inflammation, and hypertension. Studies showed that in the crystal structures of PPAR-gamma and rosiglitazone complex, binding pockets of the intact PPAR-gamma receptor interact with the rosiglitazone, especially with the Gln193, Tyr189, Leu196, Ala197, Tyr192, Glu203, Lys201, Arg202, Asp166, Lys336, Asn335, Asp337, Leu237, Phe347, Val248, Glu351, and Tyr250 residues [30].
Our investigation discovered that crucial binding pocket residues of the human PPARgamma, TYR(A)189 and TYR(D)250, form two hydrogen bonds with the two oxygen atoms of the 3-Methylpenta-2,4-dienoic acid substructure of ABA ( Figure 4A-D).
Our study revealed that binding pocket residues (GLN111, HIS131 of chain C) of the hexameric human SIRT6 protein make two hydrogen bonds with the one oxygen atom of the 3-Methylpenta-2,4-dienoic acid (C 6 H 8 O 2 ) substructure of ABA using molecular docking and pose prediction analysis ( Figure 5A-D). Interestingly, four other residues of SIRT6 like ALA51, ARG63, ILE183 and LEU184 of chain C interact via Van der Waals forces with other atoms of ABA ( Figure 5C,D).
Our study revealed that binding pocket residues (GLN111, HIS131 of chain C) of the hexameric human SIRT6 protein make two hydrogen bonds with the one oxygen atom of the 3-Methylpenta-2,4-dienoic acid (C6H8O2) substructure of ABA using molecular docking and pose prediction analysis ( Figure 5A-D). Interestingly, four other residues of SIRT6 like ALA51, ARG63, ILE183 and LEU184 of chain C interact via Van der Waals forces with other atoms of ABA ( Figure 5C,D).

The Binding Pattern of ABA with Glucokinase
Glucokinase is the most abundant hexokinase in the liver, and it plays a critical role in blood glucose homeostasis because it has strong control over hepatic glucose disposal and serves as the glucose sensor for insulin secretion in pancreatic β-cells [35]. Glucokinase is currently regarded as a promising target of anti-hyperglycemic medicines to control T2DM, but this protein target's mode of inhibition or activation is not fully understood.
There is a single published report on the crystal structure of human glucokinase (PDB ID-4IXC) complexed with alpha-D-glucopyranose and (2S) These small molecules are regarded as activators of human glucokinase, but their exact mechanism of action and key amino residues involved in the interaction are not published yet.
Our docking and binding analysis exhibit a good binding pattern (binding energy = −6.8 Kcal/mol) of ABA with human glucokinase (PDB ID-4IXC) protein ( Figure 6A,B), which is mediated through three hydrogen bonds forming residues ASN83, ARG85, and GLY229 and twelve via Van-der-Waals forces ASP78, GLY80, GLY81, PHE84, MET107, SER151, LYS169, ASP205, GLY227, THR228, GLY410, and SER445 ( Figure 6C,D). Glucokinase is the most abundant hexokinase in the liver, and it plays a critical role in blood glucose homeostasis because it has strong control over hepatic glucose disposal and serves as the glucose sensor for insulin secretion in pancreatic β-cells [35]. Glucokinase is currently regarded as a promising target of anti-hyperglycemic medicines to control T2DM, but this protein target's mode of inhibition or activation is not fully understood.
There is a single published report on the crystal structure of human glucokinase propanamide. These small molecules are regarded as activators of human glucokinase, but their exact mechanism of action and key amino residues involved in the interaction are not published yet.

Analysis of the Molecular Binding Pattern of Abscisic Acid with Aldose Reductase
Aldose reductase is the rate-limiting enzyme in the polyol pathway. It converts excess D-glucose to D-sorbitol with NADPH as a co-factor [36]. It is crucial in the treatment of diabetic microvascular problems [37]. Aldose reductase is also involved in lipid metabolism.

Analysis of the Molecular Binding Pattern of Abscisic Acid with Aldose Reductase
Aldose reductase is the rate-limiting enzyme in the polyol pathway. It converts excess D-glucose to D-sorbitol with NADPH as a co-factor [36]. It is crucial in the treatment of diabetic microvascular problems [37]. Aldose reductase is also involved in lipid metabolism.

Analysis of Abscisic Acid and Glycogen Synthase Kinase-3 (GSK-3) Docked Complex
GSK-3, a unique multifunctional serine/threonine kinase, is involved in the glycolysis pathway. GSK-3 is active and capable of synthesizing glycogen in its unphosphorylated state. PKB/AKT phosphorylates GSK-3 on serine 9 in response to insulin binding [38]. As a result, it is critical to the insulin signaling pathway's transmission of regulatory and proliferative signals occurring at the cell membrane, potentially modulating blood glucose levels [38].
However, unlike previously mentioned proteins, GSK-3 (PDB ID -3F7Z) did not show a good binding pattern with ABA (binding energy = −6.6 Kcal/mol) ( Figure 8A-D). There were no hydrogen bonds involved, only via Van der Waals forces, alkyl-pi-alkyl interaction, and pi-pi sigma interaction with ABA stabilizes the ABA bounded with the active site ( Figure 8D).

Analysis of Abscisic Acid and Glycogen Synthase Kinase-3 (GSK-3) Docked Complex
GSK-3, a unique multifunctional serine/threonine kinase, is involved in the glycolysis pathway. GSK-3 is active and capable of synthesizing glycogen in its unphosphorylated state. PKB/AKT phosphorylates GSK-3 on serine 9 in response to insulin binding [38]. As a result, it is critical to the insulin signaling pathway's transmission of regulatory and proliferative signals occurring at the cell membrane, potentially modulating blood glucose levels [38].
However, unlike previously mentioned proteins, GSK-3 (PDB ID -3F7Z) did not show a good binding pattern with ABA (binding energy = −6.6 Kcal/mol) ( Figure 8A-D). There were no hydrogen bonds involved, only via Van der Waals forces, alkyl-pi-alkyl interaction, and pi-pi sigma interaction with ABA stabilizes the ABA bounded with the active site ( Figure 8D).

Screening of Pyruvate Dehydrogenase Kinase (PKD) and Abscisic Acid Docked Complex
PKD negatively regulates the mitochondrial pyruvate dehydrogenase complex (PDC) activity by reversible phosphorylation. PDK isoforms are upregulated in obesity, diabetes, heart failure, and cancer and are potential therapeutic targets for these important human diseases [39].
Our analysis showed a poor binding pattern (binding energy = −6.3 Kcal/mol) of ABA with PKD (PDB ID-4MP2) protein ( Figure 9A-E) due to unfavorable repulsion. However, it forms hydrogen bonds, Van der Waals forces, alkyl-pi-alkyl interaction, and pi-pi sigma interaction with ABA ( Figure 9C-E).

Screening of Pyruvate Dehydrogenase Kinase (PKD) and Abscisic Acid Docked Complex
PKD negatively regulates the mitochondrial pyruvate dehydrogenase complex (PDC) activity by reversible phosphorylation. PDK isoforms are upregulated in obesity, diabetes, heart failure, and cancer and are potential therapeutic targets for these important human diseases [39].
Our analysis showed a poor binding pattern (binding energy = −6.3 Kcal/mol) of ABA with PKD (PDB ID-4MP2) protein ( Figure 9A-E) due to unfavorable repulsion. However, it forms hydrogen bonds, Van der Waals forces, alkyl-pi-alkyl interaction, and pi-pi sigma interaction with ABA ( Figure 9C-E).

Investigation of the Docked Complex of Tyrosine Kinase with Abscisic Acid
Further, ABA displayed a weak binding pattern (ΔG= −6.2 Kcal/mol) with the human tyrosine kinase (PDB ID-1IR3) protein ( Figure 10A-D) due to unfavorable repulsion.

Investigation of the Docked Complex of Tyrosine Kinase with Abscisic Acid
Further, ABA displayed a weak binding pattern (∆G = −6.2 Kcal/mol) with the human tyrosine kinase (PDB ID-1IR3) protein ( Figure 10A-D) due to unfavorable repulsion. However, it forms hydrogen bonds, Van der Waals forces, alkyl-pi-alkyl interaction, and pi-pi sigma interaction with ABA ( Figure 10D).

Molecules 2021, 26, x FOR PEER REVIEW 15 of 23
However, it forms hydrogen bonds, Van der Waals forces, alkyl-pi-alkyl interaction, and pi-pi sigma interaction with ABA ( Figure 10D).

Computational Pharmacodynamics Screening of Abscisic Acid Ligand
The molinspiration bioactivity score (v2018.03) is calculated and presented (Table 2) for active drug-likeness towards parameters like ion channel modulators, kinase inhibitors, GPCR ligands, nuclear receptor ligands, protease inhibitors, and other enzyme inhibitors with scores for >100,000 average drug-like molecules. The score allows efficient separation of active and inactive molecules. The higher values of bioactivity score of nuclear receptor ligand, enzyme inhibitor, and Ion channel modulator of 1.06, 0.75, and 0.28, respectively, shows that ABA may act as an active inhibitor for different insulin receptor proteins.

Computational Pharmacodynamics Screening of Abscisic Acid Ligand
The molinspiration bioactivity score (v2018.03) is calculated and presented (Table 2) for active drug-likeness towards parameters like ion channel modulators, kinase inhibitors, GPCR ligands, nuclear receptor ligands, protease inhibitors, and other enzyme inhibitors with scores for >100,000 average drug-like molecules. The score allows efficient separation of active and inactive molecules. The higher values of bioactivity score of nuclear receptor ligand, enzyme inhibitor, and Ion channel modulator of 1.06, 0.75, and 0.28, respectively, shows that ABA may act as an active inhibitor for different insulin receptor proteins.

In Silico Pharmacokinetics and ADMET Evaluation of Abscisic Acid
The pharmacokinetic properties and drug-likeness data are summarized in Table 3. According to the pharmacokinetic/ADMET properties, ABA showed high (96.712%) human intestinal absorption (HIA) and very low BBB permeability (−0.047 log BB). On the other hand, ABA did not affect Cytochrome P450 isomers (CYP1A2 and CYP2D6). The drug-likeness prediction was also performed using the Lipinski, Ghose, and Veber rules, as well as the bioavailability score. The Lipinski (Pfizer) filter is the pioneer to filter out any drug at the absorption or permeation level that an ideal drug has a molecular weight of less than 500 g/mol, a log P value of less than 5, and a maximum of 5 H-donor and 10 H-acceptor atoms [40]. The drug-likeness requirements are defined as follows by the Ghose filter (Amgen): The computed log P ranges from −0.4 to 5.6, the MW ranges from 160 to 480, the molar refractivity (MR) ranges from 40 to 130, and the total number of atoms ranges from 20 to 70 [41]. Veber (GSK) rule defines drug-likeness constraints as Rotatable bond count ≤ 10 and topological polar surface area (TPSA) ≤ 140 [42]. According to Martin et al., the bioavailability score was implemented to predict the probability of a compound to have at least 10% oral bioavailability in rat or measurable Caco-2 permeability [43]. AMES toxicity (non-mutagenic), hepatotoxicity, or skin sensitization was not found in the ABA. By the overall analysis of Table 3, we conclude that ABA does not violate any existing drug-likeness rules like Lipinski, Ghose, Veber, Egan and Muegge. Meanwhile, ABA has physicochemical, molecular, and ADMET properties between the upper and lower predicted values (Table 3 and Figure 11A,B). small molecules. Here PGP+/− shows the P-glycoprotein substrate positive/negative nature of the molecule under study. The BOILED-Egg's white (white area) predicts that the molecule located in this area may be passively absorbed by the human intestinal tract (HIA). The ABA molecule (red circle) is located at the extreme periphery of BOILED-Egg's yolk (yellow area), which predicts that the ABA molecule may passively permeate through the blood-brain barrier (BBB) but have very low chances of −0.047 log BB, contrary to the chances of ABA being absorbed by the human intestinal tract (HIA) is 96.712% (Table 3 and Figure 11).  (Table 3 and Figure 11 ).

Retrieval and Preparation of Proteins and Ligand
The human proteins related to diabetes mellitus 11-β-hydroxysteroid dehydrogenase (PDB ID-

Boiled-Egg Plot and Radar Graph Analysis
A BOILED-Egg plot predicts the gastrointestinal absorption and brain penetration of small molecules. Here PGP+/− shows the P-glycoprotein substrate positive/negative nature of the molecule under study. The BOILED-Egg's white (white area) predicts that the molecule located in this area may be passively absorbed by the human intestinal tract (HIA). The ABA molecule (red circle) is located at the extreme periphery of BOILED-Egg's yolk (yellow area), which predicts that the ABA molecule may passively permeate through the blood-brain barrier (BBB) but have very low chances of −0.047 log BB, contrary to the chances of ABA being absorbed by the human intestinal tract (HIA) is 96.712% (Table 3 and Figure 11).

Retrieval and Preparation of Proteins and Ligand
The human proteins related to diabetes mellitus 11-β-hydroxysteroid dehydrogenase Before docking, the protein structures downloaded from PDB were analyzed in PyMol software (The PyMOL Molecular Graphics System, Version 2.0 Schrödinger, LLC, San Diego, CA, USA), and the already docked ligands or nucleic acid or heteroatoms or water molecules were removed from the X-ray crystallographic protein-ligand complexes. Then the pure proteins as a receptor were prepared in Swiss-Pdb viewer (version 4.1.0) by optimizing bonded atoms, angles, torsions, non-bonded atoms, and improper atoms of the protein backbone and side chains.

Molecular Docking
The docking calculations were performed using the AutoDock Vina version 4.2 (ADT4.2) software suite [45]. The receptor proteins were solvated with water, and only polar hydrogens were added. The receptor grid boxes (in X, Y, Z dimension) were prepared in the ADT4.2, and the pdbqt files of proteins were generated. Similarly, the ligand was prepared with default parameters, and only Gasteiger charges were added. Flexible Ligand docking was performed applying the Lamarckian Genetic Algorithm with an exhaustiveness value of eight. The contributions of intramolecular hydrogen bonds, hydrophobic, ionic, and Van der Waals interactions between docked protein and ligand complexes were used to determine the free energy (∆G) specifying affinity scoring of the binding. The docking poses were narrowed down using the force field's free binding energy computation. After the docked protein-ligand complexes were created, the binding sites were analyzed to construct a 2D representation of the ligand interaction for each complex.

Post-Docking Protein-Ligand Interaction Analysis
The visualization and analysis of protein-ligand complexes were performed using PyMOL software (The PyMOL Molecular Graphics System, Version 2.0 Schrödinger, LLC, San Diego, CA, USA). The receptor's active sites and interactions with the ligand or drug were determined using the PDBe [46] and PDBsum [47] servers. The protein-ligand complexes were further visualized in the Discovery Studio client v21.1.0.20298, Dassault Systemes Biovia Corp, to show the 2D diagram of ligand-receptor interaction. LigPlot+ and Maestro12.4 (Schrodinger-2020-2) were applied to visualize ligands' exact atomic level interaction with their corresponding receptor atoms.

Calculation of Inhibition Constant
Moreover, it is concluded that ABA may act as a competitive inhibitor that should compete with the known substrates to the active centers of the protein targets relevant to diabetes mellitus. Therefore, it induces a competitive type of inhibition, that inhibitors could bind to only the free enzyme and formed reversible enzyme-inhibitor (E-I) complexes. An enzyme-inhibitor complex's inhibition constant (Ki) is traditionally calculated by the basic equations of enzyme kinetics of the Lineweaver-Burk assay extrapolated on 2D plots. If there is inconsistency in the Lineweaver-Burk plots, non-linear regression of the Michaelis-Menten equation is used to validate the related constants obtained.
Sophisticated arithmetic and analytical in silico algorithms have been proposed to compute the inhibition constant (Ki) parameter, since the Ki principally depends on the binding (or association) constant (Kb) and dissociation constant (Kd) of an enzyme-inhibitor complex, which occurs in opposite directions (ln Kb = −ln Kd).
Therefore, Ki is computationally calculated using the following formula: The binding energy ∆G is in kcal/mol, the universal gas constant R = 1.987 kcal/K/mol, at room temperature (25 • C) T = 273 + 25 = 298 K. Ki is having a unit of mM.

Screening of Ligand Abscisic Acid for Pharmacodynamics Properties
The molinspiration (https://www.molinspiration.com/cgi-bin/properties), (accessed on 29 May 2021), an online screening server based on sophisticated Bayesian statistics, was implemented to analyze the pharmacodynamics properties of ABA. It compares representative ligands' structures and determines physico-chemical properties of a particular molecule for being active molecules. There is no need to know about the target's 3D structure or binding mode. The trained model makes it possible to screen large libraries of hundreds of thousands of molecules in less than an hour to identify molecules with the highest chance of becoming active drugs, pesticides, irritants, or toxic substances. The larger the value of the bioactivity score is, the higher the probability that the particular molecule will be active.

Screening of Ligand Abscisic Acid for Pharmacokinetics and Drug-Likeness
The main causes of drug development failure are undesirable pharmacokinetics and toxicity of candidate molecules. Absorption, distribution, metabolism, excretion, and toxicity (ADMET) of chemicals have long been recognized as important considerations in the early stages of CADD.
The pkCSM [48], SwissADME [49] and ADMETLAB2.0 [50] are free web tools to evaluate Pharmacokinetics, drug-likeness, and medicinal chemistry of small molecules based on very extensive experimental data sets. The SMILES format of the molecule was entered, and 2D structure files were generated in SwissADME, pkCSM, and ADMETLAB2.0. The pkCSM is an authentic source (collaboratively developed by Instituto Rene Rachou Fiocruz Minas, The University of Melbourne and University of Cambridge) to predict small-molecule pharmacokinetics using graph-based signatures. Several parameters are analyzed to check the ADMET properties of a small molecule or inhibitor. SwissADME [51] of Swiss Institute of Bioinformatics used to evaluate the pharmacokinetics and drug-likeness ADMET behaviors of compounds [52] employing support vector machine (SVM) algorithm [53] with well-characterized large datasets of known inhibitors/non-inhibitors as well as substrates/non-substrates. ADMETlab 2.0 has a greater capacity to assist medicinal chemists in accelerating the drug research and development process. It allows users to calculate and predict 17 physicochemical parameters, 13 medicinal chemistry measures, 23 ADMET endpoints, 27 toxicity endpoints, and eight toxicophore rules (751 substructures) quickly and easily, allowing them to identify interesting lead compounds for further investigation.
The major target of the study was to examine if the substance in question inhibited the cytochrome P450 (CYP) family's CYP1A2 and CYP2D6 isoforms. Pharmacokinetics parameters such as human intestinal absorption, P-glycoprotein, and the BBB and druglikeness prediction Lipinski, Ghose, and Veber criteria, as well as the bioavailability score, are crucial in judging the molecule [41,42,54]. According to several essential criteria such as molecular weight, LogP, number of HPA, and HBD, the Lipinski, Ghose, and Veber guidelines were used to assess drug-likeness to determine whether a compound is likely to be bioactive.
According to Lipinski's "Rule of 5" most "drug-like" compounds have logP ≤ 5, molecular weight (MW) ≤ 500, number of hydrogen bond acceptors (nHA) ≤ 10, and number of hydrogen bond donors (nHD) ≤ 5 [55]. The molecule that violates more than one of these principles may have problems with bioavailability. The methodology calculates logP (octanol/water partition coefficient) as a sum of fragment-based contributions and correction factors. This approach is quite reliable, and it may be used to analyze almost any organic or organometallic compound.
Topological Polar Surface Area (TPSA) is calculated based on the methodology published by Ertl et al. (2000) depending on the fragment contributions [56]. TPSA is defined as the sum of surfaces of polar atoms (typically oxygens, nitrogens, and linked hydrogens) in a molecule. TPSA is an ideal descriptor characterizing drug absorption, including human intestinal absorption, bioavailability, Caco-2 (human epithelial colorectal adenocarcinoma cell line), monolayers permeability, and BBB penetration. These parameters are quite important in predicting drug transport qualities. The number of rotatable bonds (nRot) is a simple topological parameter that measures molecular flexibility. It is a very good descriptor of the oral bioavailability of drugs [42]. Any single non-ring bond bounded to a non-terminal heavy (i.e., non-hydrogen) atom is termed a rotatable bond. Because of their large rotational energy barrier, amide C-N bonds are not considered.

Conclusions
Effective medications with no cytotoxicity are needed to treat diabetes mellitus, and phytohormones such as ABA are among the best natural extract with no side effects. Molecular docking investigation of ABA with nine different protein targets relevant to diabetes mellitus revealed four potential target proteins perfectly docks with ABA. Docking analysis also revealed that based on binding energy (∆G) and predicted inhibition constant (pKi), 11β-HSD1 (4K1L) showed best binding with ABA, followed by GFAT (2ZJ4), PPARgamma (3DZY), and SIRT6 (3K35), which were equal in inhibition constant. The docking and interaction pattern of ligands were fantastically able to interact with the key residues of the catalytic cavity of the enzyme or located in the very close proximity of the active sites of these proteins. Following all current drug-likeness guidelines such as Lipinski, Ghose, Veber, Egan, and Muegge, the pharmacodynamic and pharmacokinetic features with ADMET study revealed that ABA could be taken best molecule without any hazardous effect. A BOILED-Egg plot and radar graph analysis confirm that all the molecular and physico-chemical properties of ABA are within the upper and lower limit fulfilling all the criteria of an ideal drug. Thus, ABA can be considered a potential candidate for developing a potent anti-diabetic drug and a promising bioactive compound of okra for developing nutraceuticals and functional foods.  Data Availability Statement: All data generated or analyzed during this study are included in this article.

Conflicts of Interest:
The authors declare no conflict of interest. The funders had no role in the design of the study, in the collection, analyses, or interpretation of data, in the writing of the manuscript, or in the decision to publish the results.
Sample Availability: Samples of the compounds are not available from the authors.