Next Article in Journal
Revealing the Hidden Mechanism of Enhanced Responsivity of Doped p-i-n Perovskite Photodiodes via Coupled Opto-Electronic Model
Previous Article in Journal
Integrated System Pharmacology Approaches to Elucidate Multi-Target Mechanism of Solanum surattense against Hepatocellular Carcinoma
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Phyto-Computational Intervention of Diabetes Mellitus at Multiple Stages Using Isoeugenol from Ocimum tenuiflorum: A Combination of Pharmacokinetics and Molecular Modelling Approaches

by
Reshma Mary Martiz
1,2,
Shashank M. Patil
1,
Deepika Thirumalapura Hombegowda
3,
Abdullah M. Shbeer
4,*,
Taha Alqadi
5,
Mohammed Al-Ghorbani
6,7,
Ramith Ramu
1,* and
Ashwini Prasad
2
1
Department of Biotechnology and Bioinformatics, School of Life Sciences, JSS Academy of Higher Education and Research, Mysore 570015, India
2
Department of Microbiology, School of Life Sciences, JSS Academy of Higher Education and Research, Mysore 570015, India
3
Department of Pharmacology, School of Life Sciences, JSS Academy of Higher Education and Research, Mysore 570015, India
4
Department of Surgery, Faculty of Medicine, Jazan University, Jazan 45142, Saudi Arabia
5
Department of Biology, Adham University College, Umm Al-Qura University, Makkah 21955, Saudi Arabia
6
Department of Chemistry, College of Science and Arts, Ulla, Taibah University, Madina 41477, Saudi Arabia
7
Department of Chemistry, College of Education, Thamar University, Thamar 425897, Yemen
*
Authors to whom correspondence should be addressed.
Molecules 2022, 27(19), 6222; https://doi.org/10.3390/molecules27196222
Submission received: 11 August 2022 / Revised: 26 August 2022 / Accepted: 14 September 2022 / Published: 22 September 2022

Abstract

:
In the present study, the anti-diabetic potential of Ocimum tenuiflorum was investigated using computational techniques for α-glucosidase, α-amylase, aldose reductase, and glycation at multiple stages. It aimed to elucidate the mechanism by which phytocompounds of O. tenuiflorum treat diabetes mellitus using concepts of druglikeness and pharmacokinetics, molecular docking simulations, molecular dynamics simulations, and binding free energy studies. Isoeugenol is a phenylpropene, propenyl-substituted guaiacol found in the essential oils of plants. During molecular docking modelling, isoeugenol was found to inhibit all the target enzymes, with a higher binding efficiency than standard drugs. Furthermore, molecular dynamic experiments revealed that isoeugenol was more stable in the binding pockets than the standard drugs used. Since our aim was to discover a single lead molecule with a higher binding efficiency and stability, isoeugenol was selected. In this context, our study stands in contrast to other computational studies that report on more than one compound, making it difficult to offer further analyses. To summarize, we recommend isoeugenol as a potential widely employed lead inhibitor of α-glucosidase, α-amylase, aldose reductase, and glycation based on the results of our in silico studies, therefore revealing a novel phytocompound for the effective treatment of hyperglycemia and diabetes mellitus.

1. Introduction

Type-2 diabetes mellitus (T2DM) is a chronic metabolic disease characterized by hyperglycemia, in which the body’s metabolism is disrupted as a result of abnormalities in the insulin levels [1]. Prolonged hyperglycemic conditions lead to diabetes mellitus, which in turn results in the damage, dysfunction, and failure of various organs. Carbohydrate digestive enzymes, such as α-glucosidase and α-amylase, play a crucial role in fueling hyperglycemia by releasing monosaccharides in the course of digestion [2,3,4]. Therefore, the inhibition of carbohydrate digestive enzymes proves to be an essential part of treating diabetes mellitus.
Furthermore, blood glucose levels that are too high cause a significant flow of glucose into the polyol pathway, where it is converted to sorbitol by aldose reductase [5]. Because the metabolism is impaired by the enzyme sorbitol dehydrogenase, sorbitol accumulates in the kidneys, nerves, and retina in diabetes patients. Additionally, sorbitol accumulation causes microvascular problems and a variety of cardiovascular diseases, which the aldose reductase enzyme can successfully prevent [6]. Thus, aldose reductase can also serve as an important target that can be inhibited in order to prevent sorbitol accumulation, which is associated with a number of microvascular and cardiovascular issues [7].
Furthermore, protein glycation in diabetes mellitus causes a partial impairment of activity as a result of prolonged hyperglycemia. The non-enzymatic nucleophilic addition reaction of the carbonyl residue of sugar with the free amino group of proteins forms a reversible Schiff base, which eventually gives rise to a more stable Amadori product. The Amadori products are then subjected to a series of dicarbonyl intermediate-mediated reactions, creating an unspecified class of compounds known as advanced glycation end products (AGEs). These AGEs accumulate in tissues and are the source of micro- and macro-vascular problems in diabetics [8,9]. It has been suggested that changes in lifestyle, such as increased physical activity and consumption of a diet rich in plant-derived foods (e.g., whole grains, fruits, and vegetables) might prevent 90% of T2DM cases [10]. Plant phytochemicals, such as polyphenolic compounds, as well as vitamins, minerals, and dietary fiber, have been linked to the health advantages of plant-derived products. Polyphenols have been demonstrated to lessen the severity of T2DM symptoms (such as fasting and postprandial hyperglycemia) by inhibiting disaccharidases (such as α-amylase and α-glucosidase) in the gut lumen [11].
Ocimum tenuiflorum, commonly known as tulsi, is a fragrant shrub of the basil family Lamiaceae, which is native to the eastern globe tropics, and is said to have originated in north-central India. In Ayurveda, it aids in the treatment of cough, asthma, diarrhea, fever, dysentery, arthritis, eye diseases, indigestion, gastric ailments, etc. The pharmacological advantages of tulsi have been demonstrated in numerous in vitro, animal, and human studies. O. tenuiflorum is therefore a valuable source of phytoconstituents that can be applied in pharmacotherapeutic procedures. Tulsi has been found to have a variety of pharmacological and phytochemical characteristics, including anti-diabetic properties [12,13]. In the current scenario, in silico pharmacology techniques are becoming an essential aspect of the drug development process. Meanwhile, in order to strategically plan our biological investigations, we sought to assess the anti-diabetic potential of O. tenuiflorum phytocompounds utilizing bioinformatics methods. Computational techniques, such as molecular docking, molecular dynamics modelling, and binding free energy calculations, have been proven to give accurate predictions in the field of drug development [14]. When compared to in vitro and in vivo studies, they save a substantial amount of time and money. As a result, we intended to virtually screen the O. tenuiflorum phytochemicals obtained from the Indian Medicinal Plants, Phytochemistry, and Therapeutics (IMPPAT) database as potential inhibitors of α–glucosidase, α–amylase, human aldose reductase, and human serum albumin proteins using molecular docking simulation, molecular dynamics (MD) simulation, binding free energy calculations, and pharmacokinetic analysis. The findings of this study suggest that the phytochemicals from O. tenuiflorum are potent and can act as multiple target inhibitors of T2DM; thus, they should be taken into consideration for further evaluation. Through this study, we aim to identify a single lead potential inhibitor of all the target enzymes used.

2. Material and Methods

2.1. Data Retrieval and ADMET Profiling

The datasets of active phytocompounds were searched using the Indian Medicinal Plants, Phytochemistry and Therapeutics database (https://cb.imsc.res.in/imppat/) (accessed on 10 July 2022) [15]. The phytocompounds were retrieved from NCBI PubChem database (https://pubchem.ncbi.nlm.nih.gov/) (accessed on 10 July 2022). Further, the compounds were screened based on their ADMET properties. To deduce the pharmacokinetic properties and their functions inside the body, the ADMET study was conducted based on previous works of the authors using ADMETlab 2.0 (https://admetmesh.scbdd.com/) (accessed on 10 July 2022) [16,17].

2.2. Molecular Docking Simulation

The 3D X-ray crystal structures of the target proteins required for the study, α-glucosidase, α-amylase, human serum albumin (HSA), and human aldose reductase (HAR), were retrieved from the RCSB Protein Data Bank (https://www.rcsb.org/) (accessed on 10 July 2022), and their PDB IDs are IDHK, 1AO6, and 1IEI, respectively. The protein sequence of Saccharomyces cerevisiae α-glucosidase MAL-32 obtained from UniProt (UniProt ID: P38158) was used to build a protein model using SWISS-MODEL. The model was created using the X-ray crystal structure of S. cerevisiae isomaltase (PDB ID: 3AXH), which showed a 72% identical and an 84% comparable sequence at a resolution of 1.8 Å. The model was evaluated and found to be stable in the authors’ previous works [7,18]. The pre-preparation of the proteins and ligands and virtual screening of compounds were performed based on Patil et al. (2021) [19]. The binding site was predicted according to the literature available on the RCSB PDB database. The grid box was placed on the binding pockets of the respective target proteins. The size of the grid box was maintained as constant, with different coordinates (Table 1). The molecular docking protocol was validated according to a previous study, where the same proteins (homology-built model of α-glucosidase, α-amylase, and HAR) were used for the in silico experiments [7]. In the case of HSA, the protocol was validated using the literature available on RCSB PDB database [20]. Concurrently, the 3D structures of the ligands were obtained from PubChem in SDF format and were later converted into PDBQT format using OpenBabel 2.3.1 [21,22]. Finally, the prepared protein and ligand compounds were docked using AutoDock Vina 1.1.2, along with their controls. For α-glucosidase and α-amylase, acarbose was considered as a control. Meanwhile, for human serum albumin (HSA) and human aldose reductase (HAR), aminoguanidine and quercetin were considered as controls, respectively. The selection of the control drugs was based on the previous works of the authors [7]. The virtual screening and interaction studies were performed using AutoDock Vina 1.1.2 and BIOVIA Discovery Studios Visualizer 2021, respectively, based on Kumar et al. (2021) [23].

2.3. Molecular Dynamics Simulation

Based on the interaction analysis of the compounds, the best docked conformation was selected for the dynamic investigation using GROMACS-2018.1, which is a biomolecular software package [24]. The molecular dynamics simulation study was conducted in order to understand the complexes’ stability, flexibility, and their conformational changes according to the time interval. Based on the work of Patil et al. (2021a) [18] and Patil et al. (2021b) [19], the simulation was performed for 100 ns. The simulation was carried out using a nanosecond scale, and the pdb2gmx program protein was assigned with the CHARMM36 force field to obtain the protein topology, whereas the SwissParam server (https://www.swissparam.ch/) (accessed on 12 July 2022) [25,26] was used to obtain the ligand topology. Furthermore, the system was solvated using a TIP3 water model with a 10 Å cubic box. An appropriate number of Na+ and Cl counter ions were added to neutralize the whole system, and the concentration of 0.15 M was added to maintain the salt concentration. By using the steepest descent algorithm, the energy minimization of 50,000 steps was performed on the system. Furthermore, the system was equilibrated in two phases, including the NVT and subsequent NPT ensemble (1000 ps each), with a 310 K temperature and 1 bar pressure [26,27]. The MD trajectories obtained were the root mean square deviation (RMSD), root mean square fluctuation (RMSF), radius of gyration (Rg), SASA (solvent accessible surface area), and the ligand hydrogen bonds. The MD trajectories were plotted and analyzed using XMGRACE, based on the previous studies by the authors [28,29].

2.4. Binding Free Energy Calculation

The binding free energy calculation of the complex was estimated using the mechanics/Poisson–Boltzmann surface area (MM–PBSA) approach, using the g_mmpbsa program, which is a GROMACS plugin. The quantitatively estimated of MM–PBSA was performed according to the study conducted by Martiz et al. (2022) [30]. The calculation was performed using the last 50 ns frames, which were extracted from the MD trajectory [18,19].

2.5. Druglikeness, Pharmacokinetics, and PASS Analysis of the Representative Compounds

Details related to the druglikeness and pharmacokinetics of the representative compounds (isoeugenol, acarbose, quercetin, and aminoguanidine) were retrieved from the previous analysis (virtual screening using molecular docking and ADMET profiling). In addition, the pharmacological activity prediction using the PASS online tool (http://www.way2drug.com/passonline/) (accessed on 15 July 2022) was performed on the representative compounds. The PASS server evaluates whether the provided chemical compound(s) can have a specific pharmacological effect [31]. The outcomes were numerical and classified into “Pa” and “Pi,” where “Pa” is symbolizes potential activity, while “Pi” indicates the potential inactivity of the given compound. The compounds that are considered acceptable for a particular pharmacological activity have comparatively greater Pa values than Pi values (Pa > Pi) [18,19]. In this study, parameters such as α-glucosidase inhibition, α-amylase inhibition, AGE-related disorder treatment, and HAR inhibition were assessed.

3. Results and Discussion

3.1. Virtual Screening through ADMET and Molecular Docking Simulation

Prior to performing the molecular docking, the in silico druglikeness and toxicity predictions were carried out in order to understand their biological activities and toxic effects. The screening results of all the phytocompounds of O. tenuiflorum are given in the Supplementary Materials (Supplementary Table S1).
Meanwhile, the ADMET screening results of the 26 selective compounds are given in Table 2. The predicted outcomes showed that most of the phytocompounds satisfy Lipinski’s rule of five, which is a commonly used criteria for classifying the compounds as drugs [32]. The oral bioavailability (OB) and blood–brain barrier (BBB) showed a better permeation. OB is one of the most significant pharmacokinetic features in addition ADME properties. Whereas, in the case of the TPSA, the compounds with <140 Å TPSA value were considered as more flexible and could interact better with the target protein [33]. As evident from Table 2, the values of the selected properties were well within range, and the molecules showed excellent percentages of human oral absorption.
After the ADMET screening, the docking study was carried out for the 26 selected phytocompounds. Table 3 displays the docking results of the compounds with α-glucosidase, α-amylase, HSA, and HAR as their target proteins. From Table 3, it can be concluded that all of the molecules have significantly lower docking scores (the more negative the docking score is, the better the binding is). Out of all the phytocompounds docked, isoeugenol was selected as a single multi-protein inhibitor based on its pharmacokinetic properties and binding efficiency, since our aim was to discover this type of inhibitor.
In this regard, all the pharmacokinetic parameters, the binding affinity, total number of intermolecular interactions, and total number of hydrogen bonds were taken into consideration. Isoeugenol had the highest docking score in comparison to the other docked complexes and in the case of all the protein targets. The π-π stacking interactions, halogen bonding, hydrogen bonding, and aromatic hydrogen bonding were the typical interactions observed. Further analysis, in order to understand the differences in the docking scores, was carried out using MD simulations.
In the case of α-glucosidase, isoeugenol had the better binding affinity and was bound within the inhibitor binding site of the protein. In comparison with the control, acarbose, isoeugenol formed a higher number of bonds, as presented in Figure 1. The isoeugenol complex formed a total of nine intermolecular interactions, which included four hydrogen bonds with Asp68, Arg439, Glu276, and Asp214. A single electrostatic bond was formed with Asp349. Hydrophobic π-π stacked bond bounds were formed via Phe177, whereas Tyr71, Phe157, and Phe177 formed π-alkyl with the ligand. The predicted binding interaction results are in accordance with the previous works [17,21,34]. Meanwhile, the control compound, acarbose, formed a total of seven intermolecular bonds, which is less than the isoeugenol compound, of which six were hydrogen bonds formed via Asn241, Arg439, Asp408, Pro309, and His239. A hydrophobic π-sigma bond was formed between acarbose and His279, whereas Thr307 and Asp349 were found to have formed an unfavorable acceptor–acceptor bond. The visualization of the binding interaction of isoeugenol and acarbose with α-glucosidase is given in Figure 1.
Based on the α-amylase-bound isoeugenol and acarbose docking study, isoeugenol was predicted to bind within the inhibitory binding site. A total of seven intermolecular bonds were predicted, of which four were hydrogen bonds via Arg398, Thr11, and Asp402. It also formed hydrophobic pi-pi-shaped bonds via Phe335 and pi-alkyl via Pro4 and Arg398. Meanwhile, a total four hydrogen bonds were formed between the protein and acarbose, and one unfavorable bond was formed with His331. The docking results were found to be in accordance with the previous studies [23,29,35]. The binding interactions of isoeugenol and acarbose with α-amylase are visualized in Figure 2.
Meanwhile, in the case of HAR, both isoeugenol and quercetin were bound within the inhibitory pocket of protein. Based on the predicted complex, isoeugenol formed a total of 13 bonds, of which Cys298 formed a hydrogen bond. The hydrophobic pi-sigma and pi-pi stacked bonds were formed via Trp111. The alkyl and pi-alkyl bonds were formed with Val47, Cys80, Leu300, Trp20, Trp79, Trp111, and Phe122. In comparison, quercetin had eight intermolecular interactions, of which one hydrogen bond and two donor–donor and acceptor–acceptor unfavorable bonds were formed via Tyr309 and Cys298, thus indicating that isoeugenol might form a better stable complex than quercetin. The docking results were found to be in accordance with the previous studies [7]. The binding interactions of isoeugenol and quercetin with HAR are visualized in Figure 3.
In the case of HSA, the isoeugenol compounds were predicted to result in more non-bonded interactions (8) when compared to aminoguanidine (7). A total of eight inter-molecular interactions were predicted, of which Glu354 and Arg209 formed hydrogen bonds, while hydrophobic alkyl and pi-alkyl were formed with Lys212, Val216, Lys351, Ala213, Leu327, and Ala350. These residues are present in the vicinity of fatty acid site 4 (FA4), which accommodates the methylene tails of lipids bound to this site. In addition, the ligands occupied the same binding site as the co-crystallized inhibitor ligand ibuprofen, in accordance with the previous study [20]. According to this study, binding in the polar patch of the binding pocket induces the conformational changes in the protein. In comparison, aminoguanidine formed an unfavorable donor–donor bond between the ligand and Arg197. The binding interactions of isoeugenol and quercetin with HSA are visualized in Figure 4.
Thus, based on the overall study, using different target proteins, the isoeugenol compound was predicted to have a better stability during complex formation, and the compound was identified as binding within the inhibitory binding pocket, suggesting that it might act as an inhibitor drug. The results obtained were in accordance with the previous studies [36,37].

3.2. Molecular Dynamics Simulation

Molecular dynamics simulation was performed to provide insight into the protein–ligand stability and protein structural flexibility of the docked complexes. The simulations of isoeugenol, along with the respective controls (acarbose for α-glucosidase and α-amylase, aminoguanidine for HSA, and quercetin for HAR), which bound to the targets α –glucosidase, α-amylase, HSA, and HAR, respectively, were carried out using the docked structure as a starting geometry [38]. Figure 5 represents the plot of the trajectories of the isoeugenol and acarbose complexes bound to α-glucosidase, along with apo-protein. Throughout the simulation, the RMSD values of the complexes of isoeugenol and acarbose, as well as protein α-glucosidase, show periodic variations, and isoeugenol was found within the inhibitor binding site. The isoeugenol complex was found to be stable after 80 ns, whereas fluctuation was found throughout the simulation in the case of the acarbose complex. The RMSF plot was analyzed to discern each residue’s fluctuations during the period of the simulation (100 ns). Both the complexes that bound to protein were on par, with almost similar pattern. The protein model was shown to be relatively stable at both the N- and T-terminals. The isoeugenol complex showed lower fluctuations, indicating that its interaction may be superior. The radius of gyration (Rg) of the complexes and apo-protein was determined, since it represents the structural compactness of the structure. The Rg and SASA values showed similar patterns throughout the experiment, with no fluctuations. Based on the hydrogen bond analysis, it can be predicted that structural re-agreement may have occurred during the simulation, as the number of hydrogen bonds increased when compared to the docking process; thus, it can be predicted that the isoeugenol complex has better stability compared to the acarbose complex. The simulation results complemented those of recent works that used the same protein model of α-glucosidase [18,23]. The MD simulation analysis of isoeugenol and acarbose demonstrated that both the complexes are found within the inhibitor binding site and formed persistent contacts, which may contribute to the stability of the complexes (Figure 5). Table 4 depicts both isoeugenol and acarbose complexed with α-glucosidase and the MD trajectory values.
In the case of the α-amylase-bound complexes, the RMSD plot indicates that the isoeugenol complex is bound within the inhibitor binding site, whereas a much higher deviation can be seen in the case of the acarbose complex, which might indicate that isoeugenol is more stable than the acarbose complex. Both complexes and the apo-protein showed more or less identical oscillation patterns in the RMSF evaluation. The higher fluctuation can be seen in the loop region of the structures that were studied, which indicates that there might be a chance of high mobility. Meanwhile, compared to the isoeugenol complex, both the acarbose complex and apo-protein showed high fluctuation, which may suggest that the instability with the structure. To understand the structure compactness and the stability of the complexes formed, both Rg and SASA were evaluated. Based on the evaluation, it was observed that the Rg values of both the complexes, isoeugenol and acarbose, showed similar pattern. Thus, it can be said that the complexes were compact throughout the simulation. Furthermore, to understand if any structural rearrangement occurred within the complexes, the ligand H-bond was analyzed. Based on the H-bond plot analysis, he acarbose complex showed the same number of hydrogen bonds as the isoeugenol complex, which was in accordance with our previous study. The outcomes of the MD simulation of α-amylase were found to be in accordance with the previous studies [21,26]. The graphical representation of the MD simulation plot is shown in Figure 6, and the trajectory values are given in Table 5.
The MD simulation plot analysis of the HAR-bound isoeugenol and quercetin complexes is shown in Figure 7. The RMSD plot illustrates that both the complexes are bound to the protein within the inhibitory site. Based on the plot, it can be said that isoeugenol bound to HAR stabilized after 20 ns, whereas quercetin showed a slight variation throughout the simulation. Thus, based on the RMSD evaluation, it may be said that the isoeugenol complex is more stable compared to quercetin complex. The simulation result is in accordance with the authors’ previous studies [7,39]. The RMSF plot shows high fluctuations between the residues, and both the complexes, as well as the apo-protein, showed similar patterns throughout the simulation. The Rg values of both the apo-protein and isoeugenol complex show a similar pattern, which might indicate the better compactness of the structure, whereas, based on the SASA plot, it can be observed that all the structures, the apo-protein, isoeugenol complex, and quercetin complex, yielded values that are on par and show a rather similar pattern. Finally, based on the H-bond analysis, it can be seen that the isoeugenol bound complex has a maximum of seven hydrogen bonds, compared to the quercetin docked complex, which indicates that the structure may have undergone structural rearrangement (Figure 7). The trajectory values of the MD simulation are given in Table 6.
In the case of HSA, the RMSD plot analysis showed that both the complexes and apo-protein showed rather similar patterns of variation throughout the MD simulation. Both the complexes are bound within the inhibitory site of the protein. Similar high fluctuations were seen at the terminal region of the RMSF plot, which indicates that there might be a chance of high mobility between the residues (480–590). However, based on the RMSF plot, overall, the fluctuations of the protein-aminoguanidine complex were found to be greater. The Rg value was evaluated, indicating the compactness of the structure during the complex formation, and based on the plot, it can be observed that the isoeugenol complex may have a better compactness compared with the aminoguanidine complex. A similar pattern can be seen even in the SASA plot. Finally, based on H-bond analysis, it can be predicted that structural rearrangement might have taken place. The results of the MD simulation for HSA were found to be in accordance with the previous studies [40,41]. The graphical visualization of the MD trajectory plot and values are given in Figure 8 and Table 7, respectively.

3.3. Binding Free Energy Calculations

Based on the free binding energy calculations, it can be predicted that van der Waal’s energy and the binding energies had substantial impacts on the complex formation. Based on the energy calculation, the predicted results were, mostly, energetically viable. According to the predicted results, isoeugenol bound to the α-glucosidase complex (−224.811 kJ/mol) showed the highest binding free energy when compared with all the other complexes (Table 8). Van der Waal’s energy and binding free energy were shown to be the primary contributors to the formation of the complexes when compared to the other energies. Furthermore, when compared to the protein-control complexes, the protein-isoeugenol complexes were found to have higher (more negative) binding free energies, which indicates that the protein-control complexes have a weaker interaction and binding affinity compared to the protein-isoeugenol complexes. This study result revealed a similar pattern to that observed in previous studies that performed binding free energy calculations for α-glucosidase, α-amylase, HAR [7], and HSA [42]. The predicted values of the energies calculated are summarized in Table 8 and were obtained using the MMPBSA technique.

3.4. Druglikeliness, Pharmacokinetic, and PASS Analysis of the Representative Compounds

In terms of the druglikeness properties, all the molecules except acarbose were found to be in accordance with Lipinski’s rule of five. During the analysis of the pharmacokinetic properties, acarbose and quercetin were found to violate the oral bioavailability parameter. However, studies have shown that both acarbose [43,44] and quercetin [45] possess minimal bioavailability. This makes the drugs effectively unsuitable for oral consumption. moreover, both of these compounds were predicted to cause liver injuries. This arises due to their toxicity and carcinogenic properties [46,47]. Therefore, ADMET profiling of the compounds revealed that both acarbose and quercetin are unfavorable for oral consumption. However, isoeugenol was predicted to be successful in all the investigations conducted. These reports indicate that isoeugenol may be used in in vitro and in vivo investigations. However, phytocompounds could be used as alternative therapeutics due to their minimal adverse effects on the human metabolism [48,49,50]. Table 9 depicts the druglikeness and pharmacokinetic properties of the representative compounds, whereas Figure 9 shows their pharmacokinetic mapping. The radar diagrams demonstrate that all the compounds were found to be within the acceptable boundary regarding the druglikeness properties, except for acarbose.
In addition, PASS pharmacological action predictions were also conducted to examine properties including α-glucosidase inhibition, α-amylase inhibition, AGE-related disorder treatment, and HAR inhibition. Isoeugenol was found to have more ‘Pa’ values than ‘Pi’, which indicates its positive activity in regard to all the parameters. However, acarbose was not predicted to exhibit a pharmacological action for the treatment of AGE-related disorders. In addition, aminoguanidine was found to be inactive in regard to all the parameters investigated. Moreover, quercetin was found to be inactive in terms of α-amylase inhibition and the treatment of AGE-related disorders. These results indicate that isoeugenol may act as a potential lead compound, a hypothesis which requires thorough investigation using in vitro and animal models. The results from the PASS analysis of the representative compounds are depicted in Table 10.

4. Conclusions

Plant-based antidiabetic drug development has been a stumbling block, with few promising results, as most of the drugs have yet to pass the stage of clinical trials. Conventional chemotherapeutics have been widely used in the pharmaceutical industry due to their rapid action, mechanism, and economic viability. In the current investigation, we carried out a combination of in silico investigations of the phytocompounds of O. tenuiflorum and proposed isoeugenol as a potential inhibitor of α-glucosidase, α-amylase, human serum albumin, and human aldose reductase. Isoeugenol showed the highest probability of all the phytocompounds to act as a multi-target inhibitor of all the target enzymes mentioned above. This can reduce the biological enzymatic activity of the target enzymes, which could bring about a decline in hyperglycemia. The MD simulations and binding free energy calculations, through which the binding of isoeugenol with the drug targets was validated, supported the docking results. To summarize, isoeugenol has the potential to act as a multi-target inhibitor that can be used to treat diabetes mellitus at various stages of the disorder. In the near future, isoeugenol could be assessed using in vitro, in vivo, and then clinical trials in order to discover its potential as an antidiabetic drug targeting different stages of diabetes mellitus.

Supplementary Materials

The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/molecules27196222/s1, Table S1: ADMET screening of O. tenuiflorum compounds obtained from IMPPAT database.

Author Contributions

R.R. planned and conceptualized the manuscript. R.M.M., S.M.P. and A.P. were involved in the data analysis and method development. S.M.P., R.M.M. and D.T.H. contributed to the original draft preparation and writing. A.M.S., T.A. and M.A.-G., were involved in the supervision and editing. All authors have read and agreed to the published version of the manuscript.

Funding

This research received no external funding.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

Not applicable.

Acknowledgments

All the authors thank the JSS Academy of Higher Education and Research (Mysore, Karnataka, India) for their kind support and encouragement and for providing the necessary facilities. The authors would like to thank the Deanship of Scientific Research at Umm Al-Qura University for supporting this work (grant code: 22UQU4281952DSR03).

Conflicts of Interest

All authors declare no conflict of interest.

References

  1. Patil, S.M.; Shirahatti, P.S.; Ramu, R. Azadirachta indica A. Juss (neem) against diabetes mellitus: A critical review on its phytochemistry, pharmacology, and toxicology. J. Pharm. Pharmacol. 2022, 74, 681–710. [Google Scholar] [CrossRef] [PubMed]
  2. Ramu, R.; Shirahatti, P.S.; Zameer, F.; Ranganatha, L.V.; Prasad, M.N. Inhibitory effect of banana (Musa sp. var. Nanjangud rasa bale) flower extract and its constituents Umbelliferone and Lupeol on α-glucosidase, aldose reductase and glycation at multiple stages. S. Afr. J. Bot. 2014, 95, 54–63. [Google Scholar] [CrossRef]
  3. Ramu, R.; Shirahatti, P.S.; Zameer, F.; Lakkappa Dhananjaya, B.; Prasad, M.N. Assessment of in vivo antidiabetic properties of umbelliferone and lupeol constituents of banana (Musa sp. var. Nanjangud Rasa Bale) flower in hyperglycaemic rodent model. PLoS ONE 2016, 11, e0151135. [Google Scholar]
  4. Ramu, R.; Shirahatti, P.S.; Th, D.; Bajpe, S.N.; SreepathI, N.; Patil, S.M.; Mn, N.P. Investigating Musa paradisiaca (Var. Nanjangud rasa bale) pseudostem in preventing hyperglycemia along with improvement of diabetic complications. J. Appl. Biol. Biotechnol. 2022, 10, 56–65. [Google Scholar] [CrossRef]
  5. Patil, S.M.; Martiz, R.M.; Satish, A.M.; Shbeer, A.M.; Ageel, M.; Al-Ghorbani, M.; Ranganatha, L.; Parameswaran, S.; Ramu, R. Discovery of Novel Coumarin Derivatives as Potential Dual Inhibitors against α-Glucosidase and α-Amylase for the Management of Post-Prandial Hyperglycemia via Molecular Modelling Approaches. Molecules 2022, 27, 3888. [Google Scholar] [CrossRef]
  6. Tang, W.; Martin, K.A.; Hwa, J. Aldose reductase, oxidative stress, and diabetic mellitus. Front. Pharmacol. 2012, 3, 87. [Google Scholar] [CrossRef]
  7. Maradesha, T.; Patil, S.M.; Al-Mutairi, K.A.; Ramu, R.; Madhunapantula, S.V.; Alqadi, T. Inhibitory Effect of Polyphenols from the Whole Green Jackfruit Flour against α-Glucosidase, α-Amylase, Aldose Reductase and Glycation at Multiple Stages and Their Interaction: Inhibition Kinetics and Molecular Simulations. Molecules 2022, 27, 1888. [Google Scholar] [CrossRef]
  8. Singh, V.P.; Bali, A.; Singh, N.; Jaggi, A.S. Advanced Glycation End Products and Diabetic Complications. Korean J. Physiol. Pharmacol. 2014, 18, 1–14. [Google Scholar] [CrossRef]
  9. Rungratanawanich, W.; Qu, Y.; Wang, X.; Essa, M.M.; Song, B.-J. Advanced glycation end products (AGEs) and other adducts in aging-related diseases and alcohol-mediated tissue injury. Exp. Mol. Med. 2021, 53, 168–188. [Google Scholar] [CrossRef]
  10. Hossain, U.; Das, A.K.; Ghosh, S.; Sil, P.C. An overview on the role of bioactive α-glucosidase inhibitors in ameliorating diabetic complications. Food Chem. Toxicol. 2020, 145, 111738. [Google Scholar] [CrossRef]
  11. Sharma, P.; Hajam, Y.A.; Kumar, R.; Rai, S. Complementary and alternative medicine for the treatment of diabetes and associated complications: A review on therapeutic role of polyphenols. Phytomed. Plus 2021, 2, 100188. [Google Scholar] [CrossRef]
  12. Mondal, S.; Mirdha, B.R.; Mahapatra, S.C. The science behind sacredness of Tulsi (Ocimum sanctum Linn.). Indian J. Physiol. Pharmacol. 2010, 53, 291–306. [Google Scholar]
  13. Mohan, L.; Amberkar, M.V.; Kumari, M. Ocimum sanctum linn.(TULSI)-an overview. Int. J. Pharm. Sci. Rev. Res. 2011, 7, 51–53. [Google Scholar]
  14. Jyothi, M.; Khamees, H.A.; Patil, S.M.; Ramu, R.; Khanum, S.A. Microwave-Assisted Synthesis, Characterization, Docking Studies and Molecular Dynamic of Some Novel Phenyl Thiazole Analogs as Xanthine Oxidase Inhibitor. J. Iran. Chem. Soc. 2022, 19, 3919–3933. [Google Scholar] [CrossRef]
  15. Mohanraj, K.; Karthikeyan, B.S.; Vivek-Ananth, R.P.; Chand, R.P.B.; Aparna, S.R.; Mangalapandi, P.; Samal, A. IMPPAT: A curated database of Indian Medicinal Plants, Phytochemistry And Therapeutics. Sci. Rep. 2018, 8, 4329. [Google Scholar] [CrossRef]
  16. Patil, S.M.; Manu, G.; Shivachandra, J.C.; Kumar, K.A.; Vigneswaran, J.; Ramu, R.; Shirahatti, P.S.; Ranganatha, V.L. Computational screening of benzophenone integrated derivatives (BIDs) targeting the NACHT domain of the potential target NLRP3 inflammasome. Adv. Cancer Biol. Metastasis 2022, 5, 100056. [Google Scholar] [CrossRef]
  17. Sajal, H.; Patil, S.M.; Raj, R.; Shbeer, A.M.; Ageel, M.; Ramu, R. Computer-aided screening of phytoconstituents from Ocimum tenuiflorum against diabetes mellitus targeting DPP4 inhibition: A combination of molecular docking, molecular dynamics, and pharmacokinetics approaches. Molecules 2022, 27, 5133. [Google Scholar] [CrossRef]
  18. Patil, S.M.; Martiz, R.M.; Ramu, R.; Shirahatti, P.S.; Prakash, A.; Kumar, B.R.P.; Kumar, N. Evaluation of flavonoids from banana pseudostem and flower (quercetin and catechin) as potent inhibitors of α-glucosidase: An in silico perspective. J. Biomol. Struct. Dyn. 2021, 7, 1–15. [Google Scholar] [CrossRef]
  19. Patil, S.M.; Martiz, R.M.; Ramu, R.; Shirahatti, P.S.; Prakash, A.; Chandra, J.S.; Ranganatha, L.V. In silico identification of novel benzophenone-coumarin derivatives as SARS-CoV-2 RNAdependent RNA polymerase (RdRp) inhibitors. J. Biomol. Struct. Dyn. 2021, 11, 1–17. [Google Scholar]
  20. Sugio, S.; Kashima, A.; Mochizuki, S.; Noda, M.; Kobayashi, K. Crystal structure of human serum albumin at 2.5 Å resolution. Protein Eng. Des. Sel. 1999, 12, 439–446. [Google Scholar] [CrossRef]
  21. Kumar, V.; Shetty, P.; Chandra, K.S.; Ramu, R.; Patil, S.M.; Baliga, A.; Rai, V.M.; Udupi, V.; Poojary, V.; Poojary, B.; et al. Potential fluorinated anti-MRSA thiazolidinone derivatives with antibacterial, antitubercular activity and molecular docking studies. Chem. Biodivers. 2022, 19, e202100532. [Google Scholar] [CrossRef] [PubMed]
  22. Patil, S.M.; Maruthi, K.R.; Bajpe, N.S.; Vyshali, V.M.; Sushmitha, S.; Chagalamari, A.; Ramith, R. Comparative molecular docking and simulation analysis of molnupiravir and remdesivir with SARS-CoV-2 RNA dependent RNA polymerase (RdRp). Bioinformation 2021, 7, 932–939. [Google Scholar]
  23. Kumar, V.; Ramu, R.; Shirahatti, P.S.; Kumari, V.C.; Sushma, P.; Mandal, S.P.; Patil, S.M. α-Glucosidase; α-Amylase inhibition; kinetics and docking studies of novel (2-chloro-6-(trifluoromethyl) benzyloxy) arylidene) based rhodanine and rhodanine acetic acid derivatives. Chem. Sel. 2021, 6, 9637–9644. [Google Scholar] [CrossRef]
  24. Abraham, M.J.; Murtola, T.; Schulz, R.; Páll, S.; Smith, J.C.; Hess, B.; Lindahl, E. GROMACS: High performance molecular simulations through multi-level parallelism from laptops to supercomputers. SoftwareX 2015, 1, 19–25. [Google Scholar] [CrossRef]
  25. Patil, S.M.; Sujay, S.; Tejaswini, M.; Sushma, P.; Prithvi, S.; Ramu, R. Bioactive peptides: Its production and potential role on health. Innov. Food Sci. Emerg. Technol. 2020, 7, 167–182. [Google Scholar]
  26. Patil, S.M.; Al-Mutairi, K.A.; Firdose, N.; Ramu, R.; Martiz, R.M.; Ashwini, P. Pharmacoinformatics based screening discovers swertianolin from Lavandula angustifolia as a novel neuromodulator targeting epilepsy, depression, and anxiety. S. Afr. J. Bot. 2022, 149, 712–730. [Google Scholar] [CrossRef]
  27. Kumari, V.C.; Patil, S.M.; Shirahatti, P.S.; Sujay, S.; Tejaswini, M.; Ranganatha, L.V.; Jayanthi, M.K.; Ramu, R. The current status and perspectives for the emerging pandemic: COVID-19. Int. J. Pharm. Pharm. Sci. 2020, 12, 1–12. [Google Scholar] [CrossRef]
  28. Gurupadaswamy, H.D.; Ranganatha, V.L.; Ramu, R.; Patil, S.M.; Khanum, S.A. Competent synthesis of biaryl analogs via asymmetric Suzuki–Miyaura cross-coupling for the development of anti-inflammatory and analgesic agents. J. Iran. Chem. Soc. 2022, 19, 2421–2436. [Google Scholar] [CrossRef]
  29. Ganavi, D.; Ramu, R.; Kumar, V.; Patil, S.M.; Martiz, R.M.; Shirahatti, P.S.; Sathyanarayana, R.; Poojary, B.; Holla, B.S.; Poojary, V.; et al. In vitro and in silico studies of fluorinated 2,3-disubstituted thiazolidinone-pyrazoles as potential α-amylase inhibitors and antioxidant agents. Arch. Pharm. 2021, 12, e2100342. [Google Scholar] [CrossRef]
  30. Martiz, R.M.; Patil, S.M.; Abdulaziz, M.; Babalghith, A.; Al-Areefi, M.; Al-Ghorbani, M.; Kumar, J.M.; Prasad, A.; Nagalingaswamy, N.P.M.; Ramu, R. Defining the Role of Isoeugenol from Ocimum tenuiflorum against Diabetes Mellitus-Linked Alzheimer’s Disease through Network Pharmacology and Computational Methods. Molecules 2022, 27, 2398. [Google Scholar] [CrossRef]
  31. Filimonov, D.; Lagunin, A.A.; Gloriozova, T.A.; Rudik, A.; Druzhilovskii, D.S.; Pogodin, P.V.; Poroikov, V.V. Prediction of the Biological Activity Spectra of Organic Compounds Using the Pass Online Web Resource. Chem. Heterocycl. Compd. 2014, 50, 444–457. [Google Scholar] [CrossRef]
  32. Benet, L.Z.; Hosey, C.M.; Ursu, O.; Oprea, T.I. BDDCS, the Rule of 5 and drugability. Adv. Drug Deliv. Rev. 2016, 101, 89–98. [Google Scholar] [CrossRef] [PubMed]
  33. Martiz, R.M.; Patil, S.M.; Ramu, R.; Jayanthi, M.K.; Ashwini, P.; Ranganatha, L.V.; Khanum, S.A.; Silina, E.; Stupin, V.; Achar, R.R. Discovery of novel benzophenone integrated derivatives as anti-Alzheimer’s agents targeting presenilin-1 and prese-nilin-2 inhibition: A computational approach. PLoS ONE 2022, 17, e0265022. [Google Scholar] [CrossRef]
  34. Shivanna, C.; Patil, S.M.; Mallikarjunaswamy, C.; Ramu, R.; Akhileshwari, P.; Nagaraju, L.R.; Sridhar, M.A.; Khanum, S.A.; Ranganatha, V.L.; Silina, E.; et al. Synthesis, characterization, Hirshfeld surface analysis, crystal structure and molecular modeling studies of 1-(4-(Methoxy (phenyl) methyl)-2-methylphenoxy) butan-2-one derivative as a novel α-glucosidase inhibitor. Crystals 2022, 12, 960. [Google Scholar] [CrossRef]
  35. Lolok, N.; Sumiwi, S.A.; Muhtadi, A.; Susilawati, Y.; Hendriani, R.; Ramadhan, D.S.; Levita, J.; Sahidin, I. Molecular docking and molecular dynamics studies of bioactive compounds contained in noni fruit (Morinda citrifolia L.) against human pancreatic α-amylase. J. Biomol. Struct. Dyn. 2021, 2, 1–8. [Google Scholar] [CrossRef] [PubMed]
  36. Isogai, H.; Hirayama, N. In Silico Prediction of Interactions between Site II on Human Serum Albumin and Profen Drugs. ISRN Pharm. 2013, 2013, 10–13. [Google Scholar] [CrossRef] [Green Version]
  37. Ghuman, J.; Zunszain, P.A.; Petitpas, I.; Bhattacharya, A.A.; Otagiri, M.; Curry, S. Structural Basis of the Drug-binding Specificity of Human Serum Albumin. J. Mol. Biol. 2005, 353, 38–52. [Google Scholar] [CrossRef]
  38. Pushpa, V.H.; Jayanthi, M.K.; Rashmi, H.R.; Shivamurthy, V.K.; Patil, S.M.; Shirahatti, P.S.; Ramu, R. New insights on the phytochemical intervention for the treatment of neuropsychiatric disorders using the leaves of Michelia champaca: An in vivo and in silico approach. Pharm Biol. 2022, 60, 1656–1668. [Google Scholar]
  39. Tanawattanasuntorn, T.; Thongpanchang, T.; Rungrotmongkol, T.; Hanpaibool, C.; Graidist, P.; Tipmanee, V. (−)-Kusunokinin as a Potential Aldose Reductase Inhibitor: Equivalency Observed via AKR1B1 Dynamics Simulation. ACS Omega 2020, 6, 606–614. [Google Scholar] [CrossRef]
  40. Soltanabadi, O.; Atri, M.S.; Bagheri, M. Spectroscopic analysis, docking and molecular dynamics simulation of the interaction of cinnamaldehyde with human serum albumin. J. Incl. Phenom. Macrocycl. Chem. 2018, 91, 189–197. [Google Scholar] [CrossRef]
  41. Radibratovic, M.; Minic, S.; Stanic-Vucinic, D.; Nikolic, M.; Milcic, M.; Velickovic, T.C. Stabilization of Human Serum Albumin by the Binding of Phycocyanobilin, a Bioactive Chromophore of Blue-Green Alga Spirulina: Molecular Dynamics and Experimental Study. PLoS ONE 2016, 11, e0167973. [Google Scholar] [CrossRef] [PubMed]
  42. Yu, H.; Li, D.; Xu, F.; Pan, Q.; Chai, P.; Liu, B.; Chen, C. The binding affinity of human serum albumin and paclitaxel through MMPBSA based on docked complex. Mol. Simul. 2016, 42, 1460–1467. [Google Scholar] [CrossRef]
  43. Cyndya, Shibao, Acarbose. In Primer on the Autonomic Nervous System, 3rd ed.; Robertson, D.; Biaggioni, I.; Burnstock, G.; Phillip, A.L.; Julian, F.R.P. (Eds.) Academic Press: London, UK, 2012; pp. 639–641. [Google Scholar]
  44. Rosak, C.; Mertes, G. Critical evaluation of the role of acarbose in the treatment of diabetes: Patient considerations. Diabetes Metab. Syndr. Obes. Targets Ther. 2012, 5, 357–367. [Google Scholar] [CrossRef] [PubMed]
  45. Dabeek, W.M.; Marra, M.V. Dietary Quercetin and Kaempferol: Bioavailability and Potential Cardiovascular-Related Bioactivity in Humans. Nutrients 2019, 11, 2288. [Google Scholar] [CrossRef] [PubMed]
  46. Schlüter, G. Toxicology of acarbose, with special reference to long-term carcinogenicity studies. In Acarbose for the Treatment of Diabetes Mellitus, 1st ed.; Creutzfeldt, W., Ed.; Springer: Berlin, Germany, 1988; pp. 5–14. [Google Scholar]
  47. Chen, R.; Lin, J.; Hong, J.; Han, D.; Zhang, A.D.; Lan, R.; Fu, L.; Wu, Z.; Lin, J.; Zhang, W.; et al. Potential toxicity of quercetin: The repression of mitochondrial copy number via decreased POLG expression and excessive TFAM expression in irradiated murine bone marrow. Toxicol. Rep. 2014, 1, 450–458. [Google Scholar] [CrossRef] [PubMed]
  48. Ramu, R.; Patil, S.M. A Perspective on the Effective Conduction of Functional-based Coaching Program on Diabetic Indonesian Communities. Oman Med. J. 2021, 36, e281. [Google Scholar] [CrossRef]
  49. Patil, S.M.; Shirahatti, P.S.; Ramu, R. Azadirachta indica A. Juss (neem) as a contraceptive: An evidence-based review on its pharmacological efficiency. Phytomedicine 2021, 88, 153596. [Google Scholar] [CrossRef]
  50. Patil, S.M.; Ramu, R.; Shirahatti, P.S.; Shivamallu, C.; Amachawadi, R.G. A systematic review on ethnopharmacology, phytochemistry and pharmacological aspects of Thymus vulgaris Linn. Heliyon 2021, 7, e07054. [Google Scholar] [CrossRef]
Figure 1. Visualization of the docking simulation of the compounds with α-glucosidase. (A,B) Interaction of isoeugenol (purple) visualized in 3D and 2D, respectively. (C,D) Interaction of acarbose (red) visualized in 3D and 2D, respectively.
Figure 1. Visualization of the docking simulation of the compounds with α-glucosidase. (A,B) Interaction of isoeugenol (purple) visualized in 3D and 2D, respectively. (C,D) Interaction of acarbose (red) visualized in 3D and 2D, respectively.
Molecules 27 06222 g001
Figure 2. Visualization of the docking simulation of compounds with α-amylase. (A,B) Interaction of isoeugenol (purple) visualized in 3D and 2D, respectively. (C,D) Interaction of acarbose (red) visualized in 3D and 2D, respectively.
Figure 2. Visualization of the docking simulation of compounds with α-amylase. (A,B) Interaction of isoeugenol (purple) visualized in 3D and 2D, respectively. (C,D) Interaction of acarbose (red) visualized in 3D and 2D, respectively.
Molecules 27 06222 g002
Figure 3. Visualization of the docking simulation of compounds with HAR. (A,B) Interaction of isoeugenol (purple) visualized in 3D and 2D, respectively. (C,D) Interaction of quercetin (red) visualized in 3D and 2D, respectively.
Figure 3. Visualization of the docking simulation of compounds with HAR. (A,B) Interaction of isoeugenol (purple) visualized in 3D and 2D, respectively. (C,D) Interaction of quercetin (red) visualized in 3D and 2D, respectively.
Molecules 27 06222 g003
Figure 4. Visualization of the docking simulation of compounds with HSA. (A,B) Interaction of isoeugenol (purple) visualized in 3D and 2D, respectively. (C,D) Interaction of aminoguanidine (red) visualized in 3D and 2D, respectively.
Figure 4. Visualization of the docking simulation of compounds with HSA. (A,B) Interaction of isoeugenol (purple) visualized in 3D and 2D, respectively. (C,D) Interaction of aminoguanidine (red) visualized in 3D and 2D, respectively.
Molecules 27 06222 g004
Figure 5. Analysis of RMSD, RMSF, Rg, SASA, and the number of hydrogen bonds of isoeugenol- (purple) and acarbose-bound (red) α-glucosidase complexes, as well as apo-protein α-glucosidase (green), at 100 ns. (A) Time evolution of RMSD values of both the complexes along with the protein. (B) RMSF. (C) Radius of gyration (Rg). (D) SASA. (E) Hydrogen bonds.
Figure 5. Analysis of RMSD, RMSF, Rg, SASA, and the number of hydrogen bonds of isoeugenol- (purple) and acarbose-bound (red) α-glucosidase complexes, as well as apo-protein α-glucosidase (green), at 100 ns. (A) Time evolution of RMSD values of both the complexes along with the protein. (B) RMSF. (C) Radius of gyration (Rg). (D) SASA. (E) Hydrogen bonds.
Molecules 27 06222 g005
Figure 6. Analysis of RMSD, RMSF, Rg, SASA, and the number of hydrogen bonds of isoeugenol- (purple) and acarbose-bound (red) α-amylase complexes, as well as apo-protein α-amylase (green), at 100 ns. (A) Time evolution of RMSD values of both the complexes along with the protein. (B) RMSF. (C) Radius of gyration (Rg). (D) SASA. (E) Hydrogen bonds.
Figure 6. Analysis of RMSD, RMSF, Rg, SASA, and the number of hydrogen bonds of isoeugenol- (purple) and acarbose-bound (red) α-amylase complexes, as well as apo-protein α-amylase (green), at 100 ns. (A) Time evolution of RMSD values of both the complexes along with the protein. (B) RMSF. (C) Radius of gyration (Rg). (D) SASA. (E) Hydrogen bonds.
Molecules 27 06222 g006
Figure 7. Analysis of RMSD, RMSF, Rg, SASA, and the number of hydrogen bonds of isoeugenol- (purple) and quercetin- bound (red) HAR complexes, as well as apo-protein α-glucosidase (green), at 100 ns. (A) Time evolution of RMSD values of both the complexes, along with the protein. (B) RMSF. (C) Radius of gyration (Rg). (D) SASA. (E) Hydrogen bonds.
Figure 7. Analysis of RMSD, RMSF, Rg, SASA, and the number of hydrogen bonds of isoeugenol- (purple) and quercetin- bound (red) HAR complexes, as well as apo-protein α-glucosidase (green), at 100 ns. (A) Time evolution of RMSD values of both the complexes, along with the protein. (B) RMSF. (C) Radius of gyration (Rg). (D) SASA. (E) Hydrogen bonds.
Molecules 27 06222 g007
Figure 8. Analysis of RMSD, RMSF, Rg, SASA, and the number of hydrogen bonds of isoeugenol- (purple) and aminoguanidine-bound (red) HSA complex, as well as apo-protein HSA (green), at 100 ns. (A) Time evolution of RMSD values of both the complexes, along with the protein. (B) RMSF. (C) Radius of gyration (Rg). (D) SASA. (E) Hydrogen bonds.
Figure 8. Analysis of RMSD, RMSF, Rg, SASA, and the number of hydrogen bonds of isoeugenol- (purple) and aminoguanidine-bound (red) HSA complex, as well as apo-protein HSA (green), at 100 ns. (A) Time evolution of RMSD values of both the complexes, along with the protein. (B) RMSF. (C) Radius of gyration (Rg). (D) SASA. (E) Hydrogen bonds.
Molecules 27 06222 g008
Figure 9. Pharmacokinetic mapping of the representative compounds: (A) isoeugenol, (B) acarbose, (C) aminoguanidine, and (D) quercetin.
Figure 9. Pharmacokinetic mapping of the representative compounds: (A) isoeugenol, (B) acarbose, (C) aminoguanidine, and (D) quercetin.
Molecules 27 06222 g009
Table 1. The position of the grid box placed on binding pockets of the respective target proteins.
Table 1. The position of the grid box placed on binding pockets of the respective target proteins.
Enzyme TargetsCoordinates of the Grid BoxSize of the Grid Box
xyz
α-glucosidase−17.48 Å−8.62 Å−19.65 Å40 Å × 40 Å × 40 Å
α-amylase103.46 Å37.17 Å19.60 Å40 Å × 40 Å × 40 Å
HSA8.24 Å2.58 Å−14.75 Å40 Å × 40 Å × 40 Å
HAR−5.06 Å0.19 Å9.94 Å40 Å × 40 Å × 40 Å
Table 2. Pharmacokinetics and ADMET screening results of selective compounds of O. tenuiflorum obtained from the ADMETlab 2.0 server.
Table 2. Pharmacokinetics and ADMET screening results of selective compounds of O. tenuiflorum obtained from the ADMETlab 2.0 server.
Sl. NoCompound NamesOB
(OB ≥ 30%)
BBBDHL
(HL < 3 h)
LR5IEP
(Caco-2 Cells)
DILICL (CL > 15 mL/min/kg)MW (100~600)HBA (0~12)HBD
(0~7)
TPSA (0~140)PAINS
1(−)-AlloaromadendrenePassPass0.671Accepted−4.711Negative5.356212.1504258.2000
2(−)-CamphenePassPass0.013Accepted−4.756Negative16.686414.3901120.2300
3(−)-LinaloolPassPass0.040Accepted−4.577Negative13.563204.190000.0000
4(+)-α-PhellandrenePassPass0.077Accepted−4.463Negative9.346136.130000.0000
5(+)-Endo-β-bergamotenePassPass0.493Accepted−4.375Negative9.738154.1401120.2300
6(1S)-1,7,7-Trimethylbicyclo[2.2.1]heptan-2-onePassPass0.617Accepted−4.383Negative12.660136.130000.0000
7(1S,2R,4S)-(−)-Bornyl acetatePassPass0.063Accepted−4.466Negative16.946204.190000.0000
8(E)-β-ocimenePassPass0.243Accepted−4.552Moderate6.063196.1502026.3000
91S-α-PinenePassPass0.701Accepted−4.582Negative13.808152.1201017.0700
102,3-DimethylanilinePassPass0.678Accepted−4.434Negative14.171136.130000.0000
113-CarenePassPass0.114Accepted−4.303Negative15.022136.130000.0000
124-TerpineolPassPass0.583Accepted−4.255Negative10.496121.0901226.0200
13AcetyleugenolPassPass0.132Accepted−4.307Negative16.061136.130000.0000
14α-FenchenePassPass0.447Accepted−4.217Negative14.345154.1401120.2300
15α-TerpineolPassPass0.843Accepted−4.453Moderate8.457206.0903035.5300
16β-caryophyllenePassPass0.099Accepted−4.460Negative10.559136.130000.0000
17β-PinenePassPass0.527Accepted−4.193Negative8.942154.1401120.2300
18Cis-AnetholePassPass0.048Accepted−4.517Negative9.943204.190000.0000
19Cyclo(L-Val-L-Leu)PassPass0.107Accepted−4.460Negative10.097136.130000.0000
20Dehydro-p-cymenePassModerate0.638Accepted−4.440Negative11.146148.090109.2300
21EucalyptolPassPass0.568Accepted−4.344Moderate10.755132.090000.0000
22γ-SelinenePassPass0.352Accepted−4.414Negative8.066154.140109.2300
23Geranyl acetatePassPass0.088Accepted−4.577Negative13.350204.190000.0000
24IsoeugenolPassPass0.506Accepted−4.420Moderate9.707196.1502026.3000
25MyrcenePassModerate0.880Accepted−4.579Negative13.435164.0802129.4600
26PhytosterolsPassPass0.453Accepted−4.402Moderate13.108136.130000.0000
27AcarboseFailFail0.546Rejected−6.149Positive0.373645.2501914321.1700
28AminoguanidinePassModerate0.714Accepted5.448Negative5.85774.0604687.9200
29QuercetinFailPass0.929Accepted−5.204Positive8.284302.04078131.3601
Note: OB: oral bioavailability, BBB: blood–brain barrier, DHL: drug half-life, LR5: Lipinski’s rule of five, IEP: intestinal epithelial permeability (Caco-2 cells), DILI: drug-induced liver injury, CL: clearness, MW: molecular weight, HBA: hydrogen bond acceptor, HBD: hydrogen bond donor, TPSA: topological polar surface area.
Table 3. Name of the compounds, respective binding affinity, and total non-bonded interactions and hydrogen bonds, with their respective target proteins.
Table 3. Name of the compounds, respective binding affinity, and total non-bonded interactions and hydrogen bonds, with their respective target proteins.
Sl. No.Compound NamesBinding Affinity (kcal/mol)Total No. of Intermolecular InteractionsTotal No. of Hydrogen Bonds
AGAMHSAHARAGAMHSAHARAGAMHSAHAR
1(−)-Alloaromadendrene−7.3−6.8−6.3−5.12758----
2(−)-Camphene−5.3−5.4−5.7−6.23337----
3(−)-Linalool−5.8−4.9−5.2−6.2107693---
4(+)-α-Phellandrene−5.8−5.6−5.9−6.55467----
5(+)-Endo-β-bergamotene−7.3−6.1−6.2−7.36567----
6(1S)-1,7,7-Trimethylbicyclo[2.2.1]heptan-2-one−5.9−5.5−5.7−6.32433-2--
7(1S,2R,4S)-(−)-Bornyl acetate−6.9−5.7−6.4−6.423331-12
8(E)-β-ocimene−5.7−5.0−5.8−6.486611----
91S-α-Pinene−5.5−5.4−6.2−6.14969----
102,3-Dimethylaniline−5.1−5.3−5.6−5.93573331-
113-Carene−5.4−5.5−6.0−6.53466----
124-Terpineol−5.8−5.7−6.2−6.263571---
13Acetyleugenol−6.3−5.7−6.5−6.674862121
14α-Fenchene−5.6−5.3−5.8−6.14548----
15α-Terpineol−6.2−6.0−6.2−6.555572---
16β-caryophyllene−7.3−6.0−6.1−7.01123----
17β-Pinene−5.4−5.6−6.3−6.13568----
18Cis-Anethole−5.7−5.4−6.3−5.97466----
19Cyclo(L-Val-L-Leu)−6.6−5.6−6.4−6.34311----
20Dehydro-p-cymene−6.0−5.8−6.1−6.856710----
21Eucalyptol−5.5−5.3−6.2−6.215471-2-
22γ-Selinene−7.3−6.7−6.8−7.41458--11
23Geranyl acetate−6.3−5.3−6.3−7.07757222-
24Isoeugenol−7.6−6.9−6.8−7.4978114421
25Myrcene−5.6−4.9−5.9−6.28699----
26Phytosterols−7.2−6.6−6.8−4.06732----
27Acarbose−8.2−7.4--74--64--
28Aminoguanidine--−8.0---7---3-
29Quercetin---−7.8---8---1
Note: AG: α-glucosidase, AM: α-amylase, HSA: human serum albumin, HAR: human aldose reductase.
Table 4. MD trajectory values of isoeugenol and acarbose complexed with α-glucosidase.
Table 4. MD trajectory values of isoeugenol and acarbose complexed with α-glucosidase.
MD Trajectory ValuesApo-ProteinProtein-Acarbose ComplexProtein-Isoeugenol Complex
RMSD0.30–0.40 nm0.25–0.32 nm0.20–0.25 nm
Rg3.10–3.14 nm2.39–2.45 nm2.39–2.45 nm
SASA350–370 nm2240–250 nm2240–250 nm2
Ligand H-bonds-97
Table 5. MD trajectory values of isoeugenol and acarbose complexed with α-amylase.
Table 5. MD trajectory values of isoeugenol and acarbose complexed with α-amylase.
MD Trajectory ValuesApo-ProteinProtein-Acarbose ComplexProtein-Isoeugenol Complex
RMSD0.20–0.30 nm0.25–0.32 nm0.20–0.25 nm
Rg3.10–3.14 nm2.39–2.45 nm2.39–2.45 nm
SASA350–370 nm2240–250 nm2240–250 nm2
Ligand H-bonds-77
Table 6. MD trajectory values of isoeugenol and quercetin complexed with HAR.
Table 6. MD trajectory values of isoeugenol and quercetin complexed with HAR.
MD Trajectory ValuesApo-ProteinProtein-Quercetin ComplexProtein-Isoeugenol Complex
RMSD0.20–0.36 nm0.20–0.35 nm0.20–0.30 nm
Rg1.70–1.90 nm1.90–1.92 nm1.70–1.90 nm
SASA140–150 nm2140–150 nm2140–150 nm2
Ligand H-bonds-57
Table 7. MD trajectory values of isoeugenol and aminoguanidine complexed with HSA.
Table 7. MD trajectory values of isoeugenol and aminoguanidine complexed with HSA.
MD Trajectory ValuesApo-ProteinProtein-Aminoguanidine ComplexProtein-Isoeugenol Complex
RMSD0.25–0.45 nm0.25–0.45 nm0.25–0.45 nm
Rg2.60–2.61 nm2.60–2.61 nm2.58–2.60 nm
SASA270–280 nm2270–280 nm2270–280 nm2
Ligand H-bonds-67
Table 8. Binding free energy values of the target proteins complexed with ligands.
Table 8. Binding free energy values of the target proteins complexed with ligands.
Protein-Ligand ComplexesTypes of Binding Free Energies
Van Der Waal’s
Energy
(kJ/mol)
Electrostatic
Energy
(kJ/mol)
Polar Solvation Energy
(kJ/mol)
SASA
Energy
(kJ/mol)
Binding
Energy
(kJ/mol)
α-Glucosidase-isoeugenol−224.811−10.38283.618−26.746−186.222
α-Glucosidase-acarbose−134.192−4.81362.125−9.310−90.102
α-Amylase-isoeugenol−218.568−29.89162.172−21.886−180.194
α-Amylase-acarbose−130.161−2.10639.340−9.564−87.109
HSA-isoeugenol−189.601−15.28856.9638−9.149−98.169
HSA-aminoguanidine−150.719−5.12747.498−7.981−81.872
HAR-isoeugenol−184.951−10.10587.107−25.191−99.171
HAR-quercetin−171.669−3.29181.102−8.781−85.768
Table 9. Druglikeness and pharmacokinetic properties of the representative compounds.
Table 9. Druglikeness and pharmacokinetic properties of the representative compounds.
PropertiesIsoeugenolAcarboseAminoguanidineQuercetin
Oral bioavailability (OB ≥ 30%)PassFailPassFail
Blood–brain barrier (BBB)ModerateFailModeratePass
Drug half-life (HL < 3 h)0.8800.5460.7140.929
Lipinski’s rule (LR) of fiveAcceptedRejectedAcceptedAccepted
Intestinal epithelial permeability (Caco-2 cells)−4.579−6.1495.448−5.204
Drug-induced liver injury (DILI)NegativePositiveNegativePositive
Clearness (CL > 15 mL/min/kg)13.4350.3735.8578.284
Molecular weight (MW 100~600)164.080645.25074.060302.040
Hydrogen bond acceptor (0~12)21947
Hydrogen bond donor (0~7)11468
TPSA (0~140)29.460321.17087.920131.360
PAINS0001
Table 10. PASS analysis of the representative compounds.
Table 10. PASS analysis of the representative compounds.
Compoundsα-Glucosidaseα-AmylaseAGE-Related DisorderHAR
PaPiPaPiPaPiPaPi
Isoeugenol0.1070.0250.1840.0650.3710.0110.0640.041
Acarbose0.4510.0090.9430.000----
Aminoguanidine--------
Quercetin0.1390.058----0.4660.003
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Share and Cite

MDPI and ACS Style

Martiz, R.M.; Patil, S.M.; Thirumalapura Hombegowda, D.; Shbeer, A.M.; Alqadi, T.; Al-Ghorbani, M.; Ramu, R.; Prasad, A. Phyto-Computational Intervention of Diabetes Mellitus at Multiple Stages Using Isoeugenol from Ocimum tenuiflorum: A Combination of Pharmacokinetics and Molecular Modelling Approaches. Molecules 2022, 27, 6222. https://doi.org/10.3390/molecules27196222

AMA Style

Martiz RM, Patil SM, Thirumalapura Hombegowda D, Shbeer AM, Alqadi T, Al-Ghorbani M, Ramu R, Prasad A. Phyto-Computational Intervention of Diabetes Mellitus at Multiple Stages Using Isoeugenol from Ocimum tenuiflorum: A Combination of Pharmacokinetics and Molecular Modelling Approaches. Molecules. 2022; 27(19):6222. https://doi.org/10.3390/molecules27196222

Chicago/Turabian Style

Martiz, Reshma Mary, Shashank M. Patil, Deepika Thirumalapura Hombegowda, Abdullah M. Shbeer, Taha Alqadi, Mohammed Al-Ghorbani, Ramith Ramu, and Ashwini Prasad. 2022. "Phyto-Computational Intervention of Diabetes Mellitus at Multiple Stages Using Isoeugenol from Ocimum tenuiflorum: A Combination of Pharmacokinetics and Molecular Modelling Approaches" Molecules 27, no. 19: 6222. https://doi.org/10.3390/molecules27196222

Article Metrics

Back to TopTop