Comparative anti-Diabetic potential of phytocompounds from Dr. Duke's phytochemical and ethnobotanical database and standard antidiabetic drugs against diabetes hyperglycemic target proteins: an in silico validation

Abstract In the current investigation, the antidiabetic potential of 40 phytocompounds from Dr. Dukes phytochemical and ethanobotanical database and three antidiabetic pharmaceuticals from the market comparatively validated against hyperglycemic target proteins. Silymarin, proanthocyanidins, merremoside, rutin, mangiferin-7-O-beta-glucoside, and gymnemic acid exhibited good binding affinity toward protein targets of diabetes among the 40 phytocompounds from Dr.Dukes database over three chosen antidiabetic pharmaceutical compounds. Further these phytocompounds and sitagliptin are validated for its ADMET and bioactivity score to screen its pharmacological and pharmacokinetics properties. Silymarin, proanthocyanidins, rutin along with sitagliptin screened for DFT analysis found that phytocompounds have great Homo–Lumo orbital energies over commercial pharmaceutical sitagliptin. Finally, four complexes of alpha amylase-silymarin, alpha amylase-sitagliptin, aldose reductase-proanthocyanidins, and aldose reductase-sitagliptin screened for MD simulation and MMGBSA analysis, results shown that the phytocompounds silymarin and proanthocyanidins have strong affinities for binding to the binding pockets of alpha amylase and aldose reductase respectively over antidiabetic pharmaceuticals. Our current study proven proanthocyanidins and silymarin act as novel antidiabetic compounds toward diabetic target protein but it require clinical trial to evaluate its clinical pertinence toward diabetic target proteins. Communicated by Ramaswamy Sarma


Introduction
A metabolic condition known as diabetes mellitus (DM) is often characterized by a persistent increase in blood glucose levels because the body is unable to use or generate adequate insulin (Gupta et al., 2017).Around 25% of the population is affected by DM, which is one of the primary causes of death and morbidity worldwide (Vo et al., 2016;Salehi et al., 2019).According to Dami� an-Medina et al. (2020), persistent hyperglycemia causes retinopathy, cardiovascular disease, and neuropathy over time.Type 1 insulindependent diabetes (T1DM) and Type 2 non-insulin-dependent diabetes (T2DM) affect about 90% of the population (Gupta et al., 2017).Around 463 million instances were reported in 2019 (Safitri et al., 2020), and by 2040, that figure is expected to rise to 642 million (Gupta et al., 2017).Some of the proteins that are essential in the development of diabetes include alpha-glucosidase, dipeptidyl peptidase IV, Creative protein, ADP ribosyltransferase-sirtuin-6, glutamine fructose-6-phosphate amidotransferase, protein kinase B, and insulin receptor substrate (Rathore et al., 2016).
Current antidiabetic drugs, such as metformin and sitagliptin, have made significant strides in preventing DM by modifying or blocking these proteins.Still, they also have a lot of side effects such as drug resistance, acute kidney damage, and an elevated risk of heart attack (Hu & Jia 2019;Salehi et al., 2019).Therefore, there is an immediate need to develop a natural substitute that is risk-free, efficient, and has fewer adverse effects.(Pandit et al., 2010).Plants, in particular, offer a feasible supply of innovative medicinal molecules that provide a promising alternative in the paradigm shift toward natural resources.Phytochemicals or bioactive compounds from plants, whose enormous structural variety of biological molecules is still barely explored, are one alternative source of therapeutic possibilities (Arumugam et al., 2013).Due to their safety and effectiveness in traditional medicine, herbal extracts are still used by 80% of the population in underdeveloped nations (Veeresham, 2012).Natural substances have also been proven to be one of the main sources of antidiabetic drugs in several in vivo and in vitro investigations (Jugran et al., 2020;Kabir et al., 2014;Supkamonseni et al., 2014).
Dr. Duke's phytochemical and ethnobotanical databases provide thorough searches using scientific or common names for plants, chemicals, bioactivity, and ethnobotany.It can provide search results as spreadsheets or PDFs.Pharmaceutical, nutritional, and biological research, as well as complementary treatments and herbal items, are all of the interest.Characterized antidiabetic compounds from the plant are retrieved using Dr Duke's phytochemical and ethnobotanical database and utilized for the current research study.
The hydrolases class of enzymes, which includes alphaamylase (-1,4 glucan-4-glucanohydrolase; EC 3.2.1.1),is present in microorganisms, plants, and animals (Anitha Gopal & Muralikrishna, 2009).It is a metalloenzyme, meaning that each enzyme molecule needs at least one Ca2þ ion to function and remain stable (Jane� cek & Bal� a� z, 1992).The alpha bonds in alpha-linked polysaccharides, such as those in starch and glycogen, are hydrolyzed by the enzyme-amylase, which is present in saliva and pancreatic juice (Kazeem et al., 2013).As a result, glucose and maltose are produced, which can reach the bloodstream swiftly.It is the main kind of amylase found in people and other mammals (Chaudhury et al., 2017).By hindering the breakdown of starch in the intestine, inhibition of -amylase slows down the digestive process and can be used as an efficient technique for controlling hyperglycemia (Kumar et al., 2011;Subramaniam et al., 2022).
Aldose reductase, an enzyme with a monomeric structure and a mass of around 35,900 Daltons, is a member of the aldo-keto reductase superfamily (Petrash et al., 1994).When the enzyme converts an aldehydic substrate to the corresponding alcohol, such as glucose, to sorbitol, NADPH is reversibly bound.Patients with Type 2 diabetes and cardiorenal complications (So et al., 2008) and microangiopathy have been found to have the ALD2 gene polymorphism (Watarai et al., 2006).It is crucial to remember that the majority of studies conducted worldwide have shown a connection between diabetic complications and AL2 polymorphisms (Chung & Chung, 2003).
Alpha glucosidases are digestive enzymes coming under the category of glycoside enzymes, which are found in the brush border surface membrane of tiny intestine cells, and participate in the final stage of carbohydrate digestion.These enzymes only catalyze the hydrolysis of oligosaccharides' 1,2, 1,4, and 1,6-glucosidic bonds, releasing absorbable monosaccharides (Lyann et al., 2008).
Utilizing computational techniques and molecular docking is one of the comprehensive methods for finding novel ligands with therapeutic potential.A structure-based modelling technique called molecular docking predicts how specific proteins and ligands will interact.Based on a scoring system, they assess the affinity of the interaction and the binding site in a particular receptor.(Dami� an-Medina et al., 2020;Rathinavel et al., 2020).Because of the information available, we have done an in silico comparison between the natural chemical erythrin and the commonly prescribed conventional drugs (sitagliptin, metformin, and phenformin).We also conducted an in silico toxicity assessment of erythrin and drug-likeness prediction research, both of which might be improved.The objectives of the current study are that provide comparative analysis antidiabetic potential of 40 phytocompounds from Dr. Duke's phytochemical and ethnobotanical database and three commercial antidiabetic drugs through In silico virtual screening approach to find novel drug candidate for diabetes.

Ligand selection and preparation for study
Dr. Duke's database for phytochemical and ethnobotany (https://phytochem.nal.usda.gov/phytochem/search)was used to find 40 antidiabetic phytochemicals.Phenformin, metformin and sitagliptin were chosen as the three commercial standard antidiabetic drugs for our investigation.List of chosen compounds for the present study was presented in Supplementary Table 1.All ligands two dimensional structures in SDF file format was retrieved from PubChem database (https://pubchem.ncbi.nlm.nih.gov/)(Berman et al., 2006).All obtained 2D structures were converted into threedimensional PDB structures utilizing "Online SMILES convertor and Structure file generator" tool for further in silico research (Weininger, 1988).

Drug-likeness calculation
Forty antidiabetic phytocompounds chosen from Dr. Duke's phytochemical and ethnobotanical database after searching for keyword antidiabetic compounds in Dr. Duke database showing 56 molecules leaving some inorganic molecules (Cu, Zn, Mn, etc.) chosen 40 valuable antidiabetic phytocompounds with antidiabetic activity and three commercial antidiabetic drugs also chosen for comparative purpose.All ligands were analyzed for drug-likeness parameters using the Swiss ADME online server (Daina et al., 2017).Violations of five different drug-likeness standards, including Lipinski, Ghose, Veber, Egan, and Muegge, are flagged by the Swiss ADME server.SMILES Each compound annotation are obtained from the PubChem database and sent to SWISS ADME for evaluation.Each compound physicochemical characteristics, including molecular weight, rotatable bonds, hydrogen-bond acceptors, hydrogen-bond donors, the number of rings and heteroatoms, TPSA, molar refractivity, WLogp, XLogp, and MLogP values, were calculated and analyzed to identify compounds with novel drug-likeliness characteristics.Compounds that passed screening with no or few breaches of the five drug-likeness requirements were taken next level of In silico virtual screening.

Target proteins retrieval
Three clinically significant type II diabetic target proteins were chosen for our In silico docking analysis.Alpha-amylase (PDB ID 1B2Y), Aldose reductase (PDB ID 1US0) and Alpha glucosidase (PDB ID 5NN4) are the critical Type II diabetic target proteins essential for carbohydrate metabolism playing crucial role in diabetes progression.These target proteins PDB IDs were used to get their 3D crystal structures from the Protein Data Bank at the Research Collaboratory for Structural Bioinformatics (www.rcsb.org)(Allouche, 2011).Each target protein was prepared for docking after removing all bound ligands, ions, and water molecules.Further energy minimization of target proteins was done by assigning hydrogen atoms charges through Autodock tools.

Molecular docking
For the docking analysis, AutoDock Vina (Version 4) was used, and AutoDock tools performed the calculations.For a molecular docking investigation, three clinically significant type II diabetic target proteins (Alpha-amylase, Aldose reductase, and Alpha glucosidase) were chosen (Trott & Olson, 2010).A grid map for docking protein-binding pockets was made using Autogrid.The appropriate grid box size was determined for the x, y, and z points of dimension for each target human host protein.Additional docking parameters, such as docking assessment (100 times), population size (150), energy evaluation (maximum number of 250,000), generations (maximum number of 27,000), rate of mutations (0.02), rate of cross-over (0.8), and others, were set to default values using the autotor feature of the AutoDock tool.The UCSF Chimera suite v1.14's receptor-ligand interaction tool was used to evaluate the docking results poses and 2D interaction plots of viral target proteins with ligands (Pettersen et al., 2004).

ADMET calculations
PreADMET (http://preadmet.bmdrc.org)online server was used to examine pharmacokinetic properties of three topscored antidiabetic phytocompounds from Dukes database, and one top-scored compound from standard drugs against

Bioactivity score prediction
Drug score values show a compound's overall potential as a drug candidate.The bioactivity score for six anti-diabetic phytocompounds from Dr.Dukes database and one anti-diabetic standard drug sitagliptin are validated against human receptors such as GPCRs, ion channels, kinases, nuclear receptors, proteases and enzyme inhibitors were calculated using molinspiration software version 2011.06.

DFT study
DFT analysis for three shortlisted antidiabetic compounds from Dukes database and one commercial standard antidiabetic drug was carried out to find molecular electrostatic potential (MEP).MEP is useful to identify compounds electrophilic and nucleophilic reactive sites considered as important functional groups of chosen ligand.Highest-occupied and lowest-unoccupied molecular orbitals (HOMO and LUMO) of compounds having high-and low-electron density regions useful to calculate the energy gap between them.The DFT calculations for the chosen antidiabetic phytocompounds were carried out using the Gaussian software 09 and a hybrid functional B3LYP (Becke's three parameters exchange potential and Lee Yang Parr correlation functional) using a typical triple split valence basis set 6-3IG �� .To determine the stability and chemical reactivity of the compounds, the essential parameters HOMO, LUMO, ionization potential, electron affinity, electronegativity, electronic chemical potential, molecular hardness, and softness were determined.

Molecular dynamic simulation analysis
Four docking complexes (1B2Y-5213, 1B2Y-4369359, 1US0-107876 and 1US0-4369359) the highest binding affinity and the highest docking score were chosen for molecular dynamics simulation analysis.Complexes were subjected to molecular dynamics simulation for 1-150 ns using the Desmond v3.6 module from the Schrodinger suite (Bowers et al., 2006).Due to the complexity of molecular and condensed phase structures, it is crucial to study their molecular and atomic mobility to comprehend critical physicochemical processes (Mart� ınez-Archundia et al., 2020).The Protein Preparation Wizard feature of Schrodinger Maestro suite was used to preprocess the protein-ligand complexes using default settings in order to run MD simulations (Madhavi Sastry et al., 2013).The Maestro "System Builder" tool was used to prepare all of the systems.The solvent model was determined to be Transferable Intermolecular Interaction Potential 3 Points (TIP3P) with an orthorhombic box of 10 � 10 � 10.By introducing the proper amount of counter ions (Na or Cl), the solution was made electrically neutral.The simulations for the protein-ligand complexes were conducted using the OPLS 2005 force field parameters (Shivakumar et al., 2010).Prior to the simulations, the model systems were loosened.All MD simulations utilised the NPT ensemble at a temperature and pressure of 300 K and 1 atm, respectively.In order to simulate physiological circumstances, salt (NaCl) was added at a concentration of 0.15 M. A 50 ns simulation time was used.The Coulombic interactions had a cut-off radius of 9.The Nos e-Hoover chain thermostat (with a relaxation time of 1 ps) and the Martyna-Tobias-Klein chain barostat (with a relaxation period of 2 ps) were used, respectively, to control the temperature and pressure.The RESPA integrator was used to determine the nonbonded forces, with recordings being made every 2 fs.The trajectories were saved at intervals of 50 ps for analysis, and the stability of the simulations was assessed by looking at the evolution of the proteins' and ligands' root-mean-square deviations (RMSD).For each of our complexes, MD simulations were performed three times with the same settings.On a Dell Precision T5810 Workstation equipped with an Nvidia GeForce GTX 1070 8GB graphics processor, MD simulations were performed (GPU).

MM-GBSA calculations
By submitting complicated docking poses, post-docking reduction was carried out utilizing the MM-GBSA method (Genheden & Ryde, 2015).The flexibility of the residues was fixed at a maximum of 5 from the ligand.For the MM-GBSA analysis, fixed default settings, including the OPLS3 force field, dielectric constants, and the VSGB salvation model, were used.The following formula was used to determine the MM-GBSA DG (Bind): complex is the estimated energy contribution from the optimized complex ligand-receptor, while ligand and receptor are the calculated energy contributions from the optimized free ligand and free receptor, respectively.A stronger relationship is indicated by more negative DG (bind) values.

Drug-likeness calculation
With relation to bioavailability, the drug-likeness technique empirically assesses the compounds based on their physicochemical and structural properties to find viable oral therapeutic candidates.The SWISS ADME online tool is used to examine the pharmacological and pharmacognostic profile of the selected phytocompounds based on five different rule-based filters: Lipinski, Ghose, Verber, Egan, and Mugge.Drug similarity was analyzed to ascertain whether the phytocompounds acquire ADME and water-soluble characteristics (Daina et al., 2017).Using the ADME web server, drug affinity or drug likeliness for antidiabetic phytocompounds and conventional medications were analyzed.The results are interpreted in Supplementary Table 2. 43 different substances in total, including 40 polyphenolic phytocompounds and 3 Standard Drugs, were investigated for their effects on the diabetes protein targets.Veber's and Lipinski's two discrete rules were used to separating the unique drug molecules from the examined chemicals.Twenty-four phenolic compounds and three common medications out of 40 phenolic compounds demonstrated no deviation from the Lipinski rule.A total of 3 common medications and 26 phenolic compounds displayed no deviations from Vebar's rule.A total of 15 phenolic compounds and 2 standard drugs showed zero violations against Ghose rule.A total of 24 phenolic compounds and 3 standard drugs showed zero violations against Egan rule.A total of 8 phenolic compounds and 2 standard drugs showed zero violations against Muegge rule.Only the compounds with common infractions are taken for additional in silico research, while the compounds with maximum violations are separated.These molecules are suitable for use as drugs if, for example, they have the following properties: molecular weight between 200 and 600 Da, XLOGP between À 2 and þ5, TPSA 150, number of rings 7, number of carbon atoms > 4, number of heteroatoms > 1, number of rotatable bonds 15, hydrogen-bond acceptor 10, and hydrogen-bond donor 5, respectively (Sindhu et al., 2022).To study drug transport, use topological polar surface area (TPSA), the total of polar atom surfaces in a molecule.High-TPSA chemicals are delivered, but low TPSA compounds are not.Our investigation identifies promising medications with a TPSA value of 140 or less and up to ten rotatable bonds.Various research studies on the drug-likeliness of natural products revealed that natural compounds acting in accordance with Lipinski and Veber's rule function as a drug candidate against diabetes.All of the phytocompounds in the current investigation are within the standard limits set forth for each rule, indicating a higher rate of absorption.Similarly to this, if the number of hydrogen-bond providers and acceptors exceeds what is considered acceptable, the medication molecule's ability to pass the cell membrane would be hampered.The current study's findings fall within the permissible range, allowing phytocompounds to pass through cell membranes.

Molecular docking
Drug discovery relies heavily on docking, and further developments will enable more accurate forecasts.Docking data are crucial in determining the final ligands to use and the study's next steps.A high binding affinity suggests complex elements are more likely to interact strongly.Targeted and blind docking are both possible; if a ligand attaches during blind docking to the target's specific active site, which has already been established through experimentation, it indicates a high likelihood of a potential interaction both in vitro and in vivo.With the aid of AutoDock vina, the most effective type II diabetes target proteins (Alpha-amylase, Aldose reductase, and Alpha glucosidase) were molecularly docked with a total of 40 phenolic compounds and three conventional medicines.Table 1 displays the docking scores of each chemical against 1B2Y, 1US0, and 5NN4.All of the regular medications and chosen phenolic compounds showed excellent rates of binding affinity with the target proteins for diabetes.The docking interactions of the ligands with the 1B2Y target molecule showed binding affinities ranging from À 1.0 to À 10.5 kcal/mol, with Silymarin and Berberine-Iodide recording the lowest-and highest-binding affinities, respectively.Docking results and their images are depicted in Supplementary material, Figures 1-12.
Proanthocyanidins and berberine-iodide recorded the highest-and lowest-binding affinities, respectively, of À 8.3 and À 0.8 kcal/mol, during the docking interactions of the ligands with the 1US0 target molecule.The ligands' binding affinities to the 5NN4 target molecule ranged from À 9.8 kcal/mol to À 1.0 kcal/mol during docking interactions, with Merremoside and Berberine-Iodide recording the highest-and lowest-binding affinities, respectively.The standard drug sitagliptin, which has a docking score of À 9.4 kcal/mol, was chosen for more in silico research.Selected phenolic compounds interact with proteins at binding sites and have high docking scores against type II diabetic target proteins.
Table 2 lists the top docking-scored drugs and significant interactions against type II diabetic target proteins.Following a CastP search, it was discovered that some of the binding pockets in 1B2Y, 1US0, and 5NN4 and the most recent research molecular docking results show that the following residues of screened ligands match with CastP predicted pocket: Arg195, Asp197, His201 with 1B2Y-5213 favoured by the formation of four hydrogen bond and eight alkyl bonds in Leu162, Ala198, Lys200, Glu233, Ile235, His299, Asp300, His305.Gln63, Thr163, Asp197, Glu233, and Asp300 with 1B2Y-6918448 favoured by the formation of five hydrogen bond and one alkyl bond in Trp59.Tyr151, Arg161 with 1B2Y-91617592 favoured by the appearance of two hydrogen bonds and one alkyl bond in Thr163.Gln63, Lys200 with 1B2Y-Glu38, Asn39, Ser594369359 favoured by the formation of four hydrogen bond and eight alkyl bonds in Trp59, Tyr62, Leu162, Leu165, Ala198, His201, Glu233, and Ile235.Asn39, Trp56, Ser59 with 1US0-107876 were favoured by the formation of three hydrogen bonds and one alkyl bond in Trp48.HGlu38, His40, His41, Lys66 with 1US0-5213 favored by the formation of four hydrogen bond and two alkyl bonds in Trp48, Thr64.Glu38,

ADMET calculations
Pharmacokinetic and pharmacological characteristics, including absorption, distribution, metabolism, excretion, and toxicity (ADMET), direct the initial assessment of in vivo efficacy and drug safety.These are the critical attributes for drug design and discovery in pharmaceutical research.The pharmacokinetic characteristics of medicine have a significant impact on both the level of biological activity it has toward its target protein and any negative side effects.Determining if the compounds or ligands have desirable qualities such as oral delivery and absorption, also helps to minimize late-stage failure (Chandrasekaran et al., 2018;Pricopie et al., 2019).Hence, ADMET features are crucial for drug filtering, which evaluates drug-likeness characteristics.
To be a successful product outcome, a drug must satisfy the drug-likeness requirements.Only about 50% of the medications can change these conditions.To assess the medication absorption, with several factors are taken into consideration, including intestinal absorption, permeability through membranes, skin permeability, and P-glycoprotein substrate inhibitor.CYP models for substrate or inhibitory variables determine CNS permeability and the extent of distribution, whereas the blood-brain barrier and other factors affect drug distribution (Kumari et al., 2015).CIR can be used to determine excretion.(Weininger, 1988;Wink, 2015) AMES toxicity, hepatotoxicity, and cutaneous hypersensitivity all contribute to drug toxicity.By comparing each of the aforementioned attributes to its corresponding standard value, each property was investigated.SWISS ADME software was  used to examine the compound (Han et al., 2019) ADMET characteristics (Baez-Santos et al., 2015;Daina et al., 2017) The pharmacokinetic profile of the shortlisted phenolic compounds with standard drugs against diabetic protein targets is analyzed and listed in Table 3.It states that 3082731 showed maximum human intestinal absorption (93.20%) followed by 5213 (92.55%), and the minimum HIA showed in 4369359 (81.05%).In Caco-2 Cell Permeability showed maximum activity in 369359 (21.6827 cm/s) followed by 91617592 (19.8932 cm/sec) and the minimum activity showed in 5213 (4.8446 cm/s).Observing the bioavailability of pure water showed maximum activity in 6918448 (10521.5 mg/l) followed by 4369359 (742.033mg/l) while the minimum activity in 5NN4-3082731 (0.0013 mg/l).The observation of distribution factors 5213 and 91617592 was found to be having higher blood-brain penetration ability, whereas 6918448 showed lower blood-brain penetration ability.
The metabolism properties of CYP_1A2 Inhibitor, CYP_ 2C19 Inhibitor, CYP_2C9 Inhibitor, CYP_2D6CYP_3A4 Inhibitor properties of all the selected phenolic derivatives were also measured and recorded in Table 3.The excretion toxicity like total clearance, showed maximum activity in 91617592 (0.891 log ml/min/kg) followed by 5213 (0.762 log ml/min/kg), and the minimum activity showed in 5280805 (0.382 log ml/min/kg).The AMES toxicity, hERG inhibition, Hepatotoxicity and oral rate acute toxicity were also tested, and the results are represented in Table 3.

Bioactivity score prediction
With better binding selectivity profiles and fewer side effects, novel functional medications can be found using the bioactivity score.It provides details on the drug binding sequence for various protein structures.Inhibitors of kinases, nuclear receptor ligands, enzymes, and GPCR ligands were among the substances against which the phytocompound drug-likelihood qualities were examined.Table 4  findings as bioactivity scores.The bioactivity scores of the phytocompound were categorized as inert (bioactivity score >5.0), moderately active (bioactivity score between 5.0 and 0.00), and extremely active (bioactivity score >0.00) biologically (Mishra et al., 2018).As seen in Table 4, the bioactivity scores of the phytocompounds were between 5.0 and 0.0, which predicted moderate biological activity.Our study results consistent with previous research results (Rathinavel et al., 2019(Rathinavel et al., , 2021) )

DFT analysis
With accuracy down to the B3LYP/6-31G � level, density functional theory (DFT), a quantum mechanical technique, can explain the structural and electrical properties of molecules.Through orbital energy calculations, the electronic distribution of the docked selected phytocompounds was theoretically established in the current work.To explore the inhibitory potential of phytocompounds is helpful to frame an understanding of protein-ligand interactions.The highest and lowest occupied molecular orbital energies (HOMO) and the regions with high and low levels of electron density in phytocompounds can be calculated (LUMO), and the results are depicted in Table 5 and Figure 2. One of the examined compounds, rutin (CID 5280805) displayed the lowest energy gap (0.17146), the lowest hardness (0.08571 eV), and the highest binding affinity (11.6672 eV).Antidiabetic compound sitagliptin (CID 43939) exhibits good Homo-Lumo orbital energies next to rutin compound DFT profiles.Phytocomppunds proanthocyanidins (CID 107876) and silymarin (CID 5213) scores next to the sitagliptin drug DFT profile.The maximum binding affinity and best docking score of the ligand with the target site are better supported by the DFT calculations carried out in this report's in-silico screening.To determine a molecule's energy gap, softness, and hardness, the HOMO-LUMO orbital energies protocol was applied as the border orbital energy.This semi-empirical method allowed for the quantitative analysis of the biological activity of chemical compounds.Functional orbitals of ligand molecules that hold the best-visualized structure in HOMO-LUMO orbital energies contribute to a product's higher docking score (Aramice et al., 2016;Menachery et al., 2015).Since the free electrons from HOMO aid in the transfer of charges during the formation of protein ligand complexes, it is crucial that HOMO and LUMO orbitals are localized.Additionally, a molecule's HOMO and LUMO are crucial for exposing the intermolecular interactions that occur between a drug's HOMO and a receptor's LUMO, as well as between a receptor and a drug's HOMO and LUMO.The energy difference between the interacting orbitals is used to calculate the stability of interactions in an inverse manner.Increased HOMO and decreased LUMO energies in the compounds indicate the development of complex interaction stability and the greatest receptor binding.Additionally, relative polarity, electrophilic, and nucleophilic sites of a molecule are graphically provided by electrostatic potential maps, which can be utilized to envision the favourable spots on the compound that involve the development of interactions with receptors.The energy gap between HOMO and LUMO is a key factor in defining a chemical system's electrical conductivity and other molecular characteristics.Our previous research study results shown that phytocompound vanillin from Plectranthus amboinicus (Lour.)Spreng shows less energy gap (À 0.1660 eV) and hardness (0.083 eV), and more softness (12.0481 eV) values are consistent with our current research findings (Sindhu et al., 2022).

Molecular dynamic simulation analysis
For the four shortlisted complexes (aldose reductase-proanthocyanidins, aldose reductase-sitagliptin, and alpha amylasesilymarin), molecular dynamics simulation was carried out to examine the structural stability at 150 ns.The results are shown in Figures 3-6.From trajectory analysis, the RMSD plot was obtained to assess protein stability and the RMSD ligand indicates the protein's binding pockets and the strength of the corresponding ligand.The RMSD ligand demonstrated the protein's binding pockets and the stability of the corresponding ligand.Alpha amylase (PDB ID 1B2Y) -silymarin showed RMSD values for both ligand 2.0 Å and protein 1.5 Å with a small deviation in between 90 and 120 ns simulation period is quite common in globular protein slight structural conformational change (Figure 3a).Whereas alpha amylase-sitagliptin complex display good simulation RMSD with slight deviation (2 Å for protein and 9 Å for ligand) for the period of 40-90 ns (Figure 3b), and it exhibits more deviation in the rest of the simuation period.Similarly, phytocompound proanthocyanidins display satisfactory simulation trajectory with aldose reductase enzyme at 1.2 Å for protein and 3. 0 Å for ligand RMSD values (Figure 3c) on the other hand aldose reductase-sitagliptin complex shown more deviated RMSD values showing overlapping RMSD values between 70 and 100 ns at 1 A � for protein and 32 A � for ligand RMSD values (Figure 3d).More RMSF peaks were recorded in alpha amylase enzyme amino acid residues, which lie in the range of 1-4.5 Å.Three tall RMSF peaks were recorded in alpha amylase enzyme amino acids residues which is due to the binding pocket of enzyme alpha amylase, whereas less number RMSF peaks in recorded in aldose reductase enzyme amino acid residues and their RMSF peak values falls below 1.5 Å except a single peak shown 3.5 Å (Figure 4).
Previous research report about compound from ChEMBL database and its scaffolds showing in vitro aldose reductase inhibition with known IC50 values are validated in silico analysis, revealing stable and strong binding affinity and good simulation trajectories with aldose reductase.Further, it showed good QSAR results in acting as a novel aldose reductase inhibitor (Bakal et al., 2022).Our results are also consistent with the previous research report about two new  benzophenones: garcimangophenones A purified from the pericarps ethyl acetate fraction of Garcinia mangostana ((GM) Clusiaceae) possess stable binding affinity toward binding pocket of alpha amylase.Further MD simulation analysis reveals that a steady simulation trajectory exists between benzophenones derivatives with alpha amylase during entire simulation period (50 ns) (Alhakamy et al., 2022).

MM-GBSA analysis
MMGBSA results of screened complexes are presented in Table 6.Among all four screened complexes alpha amylasesilymarin possesses the highest MMGBSA score À 64.65 Kcal/mol followed by aldose reductase-proanthocyanidin complex score À 41.46 Kcal/mol.In contrast, ligand sitagliptin possess lowest MMGBSA score values against alpha amylase and aldose reductase target protein with MMGBSA scores of À 35.43 and À 41.02 Kcal/mol respectively.Our study results were consistent with our previous results (Thirumalaisamy et al., 2022).

Conclusion
The current research study aims at hyperglycemic protein targets of diabetes and its novel inhibitor from phytocompounds without side effects.Molecular docking study results reveal phytocompounds such as silymarin, proanthocyanidins, merremoside, rutin, mangiferin-7-O-beta-glucoside, gymnemic acid from Dr. Duke's phytochemical and ethnobotanical database possess strong binding affinity with all three screened protein targets of diabetes.Further, these compounds have excellent pharmacokinetic Homo-Lumo orbital energies.These compounds displayed good bioactivity score against number of pharmacologically important protein and enzyme molecules screened and validated through in silico approach.Finally, four complexes alpha amylasesilymarin, alpha-amylase-sitagliptin, aldose reductaseproanthocyanidins, aldose reductase-sitagliptin screened for MD simulation and MMGBSA analysis revealed that phytocompounds silymarin and proanthocyanidins possess strong binding affinity and simulation trajectory toward binding pockets of alpha-amylase and aldose reductase.Both silymarin and proanthocyanidin compounds can be used as novel antidiabetic compounds after successful human clinical trials.

Figure 3 .
Figure 3. RMSD of the Ca atoms of the target protein and the ligand over time.The left y-axis shows the variation in the target protein RMSD, and the right y-axis shows the variation in the ligand RMSD over time (a) Alpha amylase (PDB ID 1B2Y)-Silymarin (CID 5213) complex, (b) Alpha amylase (PDB ID 1B2Y)-Sitagliptin (CID 4369359) complex, (c) Aldose reductase (PDB ID 1US0)-Proanthocyanidins (CID 107876) complex, and (d) Aldose reductase (PDB ID 1US0)-Sitagliptin (CID 4369359)complex.

Figure 6 .
Figure 6.Timeline representation of the interactions and contacts (H-bonds, Hydrophobic, Ionic, and Water bridges) the top panel of each image shows the total number of specific contacts the protein makes with the ligand throughout the trajectory.The bottom panel of each image shows which residues interact with the ligand in each trajectory frame.(a) Alpha amylase (PDB ID 1B2Y)-Silymarin (CID 5213) complex, (b) Alpha amylase (PDB ID 1B2Y)-Sitagliptin (CID 4369359) complex, (c) Aldose reductase (PDB ID 1US0)-Proanthocyanidins (CID 107876) h-complex, and (d) Aldose reductase (PDB ID 1US0)-Sitagliptin (CID 4369359)complex.

Table 1 .
Docking score of anti-diabetic compounds and standard drug against potent diabetic protein targets.eachtargets.PreADMET predicts the numerous pharmacokinetic factors involved in phytocompounds ADME behaviour, including absorption, bioavailability, and the therapeutic candidate's metabolic profile.PreADMET software was used to determine the ADME parameters of phytocompounds of Dukes database and one top-scored commercial drug.

Table 2 .
Docking interaction details between anti-diabetic compounds and diabetic target proteins.

Table 3 .
Pharmacokinetic profiles of anti-diabetic compounds and standard drugs.

Table 4 .
Bioactivity score for chosen anti-diabetic compounds.

Table 5 .
DFT indices for anti-diabetic compounds from Duke's database and standard drugs.