**Pharmacophore-Based Virtual Screening, Quantum Mechanics Calculations, and Molecular Dynamics Simulation Approaches Identified Potential Natural Antiviral Drug Candidates against MERS-CoV S1-NTD**

**Thamer A. Bouback 1,†, Sushil Pokhrel 2,†, Abdulaziz Albeshri <sup>1</sup> , Amal Mohammed Aljohani <sup>1</sup> , Abdus Samad 3,4 , Rahat Alam 3,4 , Md Saddam Hossen 3,4,5 , Khalid Al-Ghamdi <sup>1</sup> , Md. Enamul Kabir Talukder 3,4, Foysal Ahammad 1,3,4,\* , Ishtiaq Qadri 1,\* and Jesus Simal-Gandara 6,\***


**Abstract:** Middle East respiratory syndrome coronavirus (MERS-CoV) is a highly infectious zoonotic virus first reported into the human population in September 2012 on the Arabian Peninsula. The virus causes severe and often lethal respiratory illness in humans with an unusually high fatality rate. The N-terminal domain (NTD) of receptor-binding S1 subunit of coronavirus spike (S) proteins can recognize a variety of host protein and mediates entry into human host cells. Blocking the entry by targeting the S1-NTD of the virus can facilitate the development of effective antiviral drug candidates against the pathogen. Therefore, the study has been designed to identify effective antiviral drug candidates against the MERS-CoV by targeting S1-NTD. Initially, a structure-based pharmacophore model (SBPM) to the active site (AS) cavity of the S1-NTD has been generated, followed by pharmacophore-based virtual screening of 11,295 natural compounds. Hits generated through the pharmacophore-based virtual screening have re-ranked by molecular docking and further evaluated through the ADMET properties. The compounds with the best ADME and toxicity properties have been retrieved, and a quantum mechanical (QM) based density-functional theory (DFT) has been performed to optimize the geometry of the selected compounds. Three optimized natural compounds, namely Taiwanhomoflavone B (Amb23604132), 2,3-Dihydrohinokiflavone (Amb23604659), and Sophoricoside (Amb1153724), have exhibited substantial docking energy >−9.00 kcal/mol, where analysis of frontier molecular orbital (FMO) theory found the low chemical reactivity correspondence to the bioactivity of the compounds. Molecular dynamics (MD) simulation confirmed the stability of the selected natural compound to the binding site of the protein. Additionally, molecular mechanics generalized born surface area (MM/GBSA) predicted the good value of binding free energies (∆G bind) of the compounds to the desired protein. Convincingly, all the results support the potentiality of the selected compounds as natural antiviral candidates against the MERS-CoV S1-NTD.

**Citation:** Bouback, T.A.; Pokhrel, S.; Albeshri, A.; Aljohani, A.M.; Samad, A.; Alam, R.; Hossen, M.S.; Al-Ghamdi, K.; Talukder, M.E.K.; Ahammad, F.; et al. Pharmacophore-Based Virtual Screening, Quantum Mechanics Calculations, and Molecular Dynamics Simulation Approaches Identified Potential Natural Antiviral Drug Candidates against MERS-CoV S1-NTD. *Molecules* **2021**, *26*, 4961. https://doi.org/10.3390/ molecules26164961

Academic Editors: Giovanni N. Roviello and Caterina Vicidomini

Received: 24 April 2021 Accepted: 7 May 2021 Published: 17 August 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

**Keywords:** MERS-CoV; S1-NTD; pharmacophore modeling; ADME; quantum mechanical calculation; DFT; FMO; MM/GBSA; molecular docking; Molecular Dynamics Simulation; HOMO; LUMO

#### **1. Introduction**

Coronavirus (CoVs) is a member of the family Coronaviridae, containing a singlestranded positive-sense RNA genome [+ssRNA] that has a length between ~27 kb to 32 kb. The virus causes illness ranging from the common cold to more severe diseases in humans and animals [1]. Genetically diverse coronaviruses cycles in nature among its three principal hosts, which are the natural host (bats, mice), intermediate host (camels, masked palm civets, swine, dogs, and cats), and humans [2]. Four common human coronaviruses (HCoVs) NL63 and 229E (α-CoVs); OC43 and HKU1 (β-CoVs) circulate widely in the human population, each capable of causing severe disease ranging from common colds to self-limiting upper respiratory infections in immunocompetent people.

Severe acute respiratory syndrome coronavirus (SARS-CoV), SARS-CoV-2, and MERS-CoV are other known human beta CoVs (β-CoVs) capable of causing epidemics [3,4]. They are zoonotic in characters and infections with the virus resulting in various clinical severity featuring respiratory and extra-respiratory manifestations [5]. The number of global people infected with coronavirus has risen rapidly, which began with the pandemic of SARS-CoV in 2003, followed by the MERS-CoV in 2012 and, most recently, the SARS-CoV-2 outbreaks with a fatality rate of ~10%, ~35%, and 0.1% to over 25%, respectively [6,7]. Among the animal coronaviruses, MERS-CoV has the highest fatality rate in humans and animals, but effective antiviral candidates are not available to treat the infection caused by the pathogen.

Receptor recognition is an initial and key step of a virus to infections into the host cells. MERS-CoV class I membrane fusion trimeric S glycoprotein of the virion can recognize host receptor that mediates entry into the cells. The S trimer of highly pathogenic MERS-CoV can recognize the host cellular receptor dipeptidyl peptidase 4 (DPP4), resulting in membrane fusion and viral entry [8]. The trimeric ectodomain segment of MERS-CoV-S protein can be divided into two subunits, the first one is the receptor binding S1 subunit and another one is the membrane-fusion S2 subunit. The receptor binding S1 subunit of the virus can also be divided into two independent domains, namely the N-terminal domain (S1-NTD) and C-terminal domain (S1-CTD), which can function as a receptor-binding domain (RBD) for the S protein [9]. The S1-NTD of MERS-CoV can identify specific sugar moieties upon primary attachment and help in the prefusion-to-post fusion transition, which is critical in determining tissue tropism, host ranges, and cross-species infection [10]. Therefore, targeting the S1-NTD of the MERS-CoV S protein can inhibit the primary attachment to the host and block the prefusion-to-post fusion transition and will be an effective prophylactic against the virus [11].

The S1-NTD targeting natural compounds with potent inhibitory activity can be a focus on the improvement of therapeutic interventions of the virus [12]. Many studies reported different neutralizing antibodies and chemically synthesized compounds as drug candidates previously, for example, folic acid showed activity against NTD from the mammalian expression medium [10,13]. Sometimes, this type of antibody can induce resistance against the virus and chemically synthesized compounds can causes adverse side effect of the host [14]. Natural compounds having low toxicity and side effect can be developed as antiviral candidates by targeting S1-NTD that will be novel therapeutics for MERS-CoV [15]. We thus sought to identify potential natural antiviral drug candidates against the MERS-CoV by targeting S1-NTD.

Nowadays, computer-aided drug design (CADD) has become an effective and powerful technique in different therapeutic development. The technique has helped to overcome the long-term and expensive process that costs billions of dollars previously during drug design and development [16]. The importance of the in-silico drug design technique is greater than ever before in the modern drug design process [17]. Therefore, the study

utilized different in-silico technique includes pharmacophore modeling, virtual screening, molecular docking, ADMET, QM calculation, MD simulation, and MM/GBSA to identify effective and potential natural drug candidates against MERS-CoV.

#### **2. Results**

#### *2.1. Results of Pharmacophore Modeling Molecules* **2021**, *26*, x FOR PEER REVIEW 3 of 26

Pharmacophore can be defined as an ensemble of common steric and electronic chemical features that indicates a compound-specific mode of action to the active site of a targeted biological macromolecule. The pharmacophore features can be observed during ligand–protein interaction and helps in screening a large chemical database for retrieving novel scaffolds as a lead compound [18]. **2. Results**  *2.1. Results of Pharmacophore Modeling*  Pharmacophore can be defined as an ensemble of common steric and electronic chemical features that indicates a compound-specific mode of action to the active site of a

To identify novel scaffolds as a lead compound against MERS-CoV S1-NTD, two different pharmacophore models were generated based on the protein PDB ID: 5VYH and 6PXH in complex with folic acid (FOL409) and dihydro-folic acid (DHF428), respectively. The ligand (FOL409) in complex with the protein 5VYH generated a total of 14 pharmacophore features includes two aromatic ring (AR), eight hydrogen bond acceptor (HBA), and four hydrogen bond donor (HBD) features, where complex of DHF428 and 6PXH produced 10 pharmacophore features includes one AR, one hydrophobic (H), four HBA, and four HBD features shown in Figure 1A,B. This two-pharmacophore model was aligned and merged to interpolate overlapping features, which generated a total of 20 pharmacophore features including three AR, one H, two negative ionizable area (NI), 10 HBA, and four HBD features shown in Figure 1C. The overlapped and duplicate pharmacophore features from the aligned pharmacophore models have been removed to optimize and relaxed the geometrical confirmation of the model. After removing the duplicate pharmacophore features a total of 11 features includes three HBD, one H, three HBA, three AR, and one NI feature were selected for further study shown in Figure 1D. Exclusion volume coat generated during the pharmacophore modeling process has not been considered in the study. targeted biological macromolecule. The pharmacophore features can be observed during ligand–protein interaction and helps in screening a large chemical database for retrieving novel scaffolds as a lead compound [18]. To identify novel scaffolds as a lead compound against MERS-CoV S1-NTD, two different pharmacophore models were generated based on the protein PDB ID: 5VYH and 6PXH in complex with folic acid (FOL409) and dihydro-folic acid (DHF428), respectively. The ligand (FOL409) in complex with the protein 5VYH generated a total of 14 pharmacophore features includes two aromatic ring (AR), eight hydrogen bond acceptor (HBA), and four hydrogen bond donor (HBD) features, where complex of DHF428 and 6PXH produced 10 pharmacophore features includes one AR, one hydrophobic (H), four HBA, and four HBD features shown in Figure 1A,B. This two-pharmacophore model was aligned and merged to interpolate overlapping features, which generated a total of 20 pharmacophore features including three AR, one H, two negative ionizable area (NI), 10 HBA, and four HBD features shown in Figure 1C. The overlapped and duplicate pharmacophore features from the aligned pharmacophore models have been removed to optimize and relaxed the geometrical confirmation of the model. After removing the duplicate pharmacophore features a total of 11 features includes three HBD, one H, three HBA, three AR, and one NI feature were selected for further study shown in Figure 1D. Exclusion volume coat generated during the pharmacophore modeling process has not been considered in the study.

**Figure 1.** The overall pharmacophore model generated during the study. Herein, (**A**) pharmacophore features generated from 5VYH (PDB)-Folic Acid (yellow) complex interaction, (**B**) 6PXH (PDB)-Dihydrofolic Acid (red) complex interaction, (**C**) merge pharmacophore features, and (**D**) final pharmacophore features utilized for virtual screening. The hydrogen **Figure 1.** The overall pharmacophore model generated during the study. Herein, (**A**) pharmacophore features generated from 5VYH (PDB)-Folic Acid (yellow) complex interaction, (**B**) 6PXH (PDB)-Dihydrofolic Acid (red) complex interaction, (**C**) merge pharmacophore features, and (**D**) final pharmacophore features utilized for virtual screening. The hydrogen bond donor (HBD) features have shown in green, hydrogen bond acceptor (HBA) in red, negative ionizable area (NI) in red astricts, aromatic ring (AR) in blue, and hydrophobic (H) features in yellow color.

#### *2.2. Molecule Library Preparation*

Virtual screening can be defined as a cheminformatics technology that utilizes different computational techniques to screen a large number of molecules and identify the structures of interest for biological assays [19]. The accuracy of a cheminformatics model depends on the data mining process that is related to database preparation. Therefore, to accurately mine the database, a total of 11,295 natural compounds have been retrieved from the Ambinter, and a library has prepared for virtual screening. The geometry of all the molecular structures has been optimized by conforming MMFF94 force field available at the LigandScout tool and a molecular library has been prepared [20]. The library prepared through the software has further utilized for the virtual screening process.

#### *2.3. Active Compounds Identification and Decoy Set Generation*

Validation of a pharmacophore model is essential before a large database screening process can provide reliable outcomes on a real-life project. The SBPM can be validated through known active compounds together with inactive compounds called "decoys". Ideally, active compounds for model validation should be selected based on experimental data [19]. Therefore, 12 experimentally active compounds against MERS-CoV S protein have been identified and retrieved from the ChEMBL database. The active compounds have been selected based on their half maximal inhibitory concentration IC<sup>50</sup> (nM) value shown in Figure 2. The active compounds then submitted into the DUDE-E decoy database and a total of 1326 decoys correspondence to active compounds has been retrieved. The geometry of the compounds has also been optimized by using the MMFF94 force field and converted into the LDB file format through LiganScout software. *Molecules* **2021**, *26*, x FOR PEER REVIEW 5 of 26 the excellence of the model [19]. The enrichment factor in the pharmacophore model provides an idea about the number of active compounds found from a specific model compared to hypothetically active compounds found from a randomly screened model. The EF factor can range from 1 to >100, where 1 indicates the number of randomly sorted molecules and >100 indicates the least number of compounds need to screen in vitro to find a large number of active compounds. The AUC and EF values found in the study were 0.74 and 1.1, respectively, indicating good discrimination ability and robustness of the pharmacophore model shown in Figure 3.

**Figure 2.** List of active compounds identified against MERS coronavirus S1-NTD protein. The IC50 value and correspondence ChEMBL identity for each compound has also been provided. **Figure 2.** List of active compounds identified against MERS coronavirus S1-NTD protein. The IC<sup>50</sup> value and correspondence ChEMBL identity for each compound has also been provided.

#### *2.4. Pharmacophore-Based Virtual Screening*

In-silico virtual screening is a type of computational approach by which molecules with desired properties can retrieve structures with similar properties from large molecule libraries. During the drug design and development process, this technique helps to identify small molecules as hits and further optimization as lead candidates [21]. Furthermore, this process can help to reduce the assay-to-lead attrition rate that has excluded time and expensive experiments require during the drug design and development process. A specific 3D pharmacophoric pattern searching approach to screen large molecule libraries is now being considered as the next step in the drug design process. Therefore, the study utilized a 3D pharmacophore models-based virtual screening process to identify hit compounds against the targeted protein. The structure-based virtual screening process retrieved 32 active compounds as hits with a geometric fit score of 65.46 to 67.75, where the number of conformations generated during the screening was a minimum of eight and a maximum of 25 shown in Table S1.

#### *2.5. Pharmacophore Model Performance Analysis*

To determine the performance of the pharmacophore model, the ROC curve generated during the virtual screening process has been analyzed. Receiver-operating characteristic (ROC) is a simple and useful graphical tool for evaluating the accuracy of a statistical model. The ROC curve in the virtual screening process provides information regarding the discrimination ability of the model from active to inactive (decoy) set [19]. The overall summary of the model accuracy can be calculated from the Area Under the Curve (AUC) that represents the degree of discrimination ability. The AUC value ranges between 0.0 to 1.0, where a value between 0 to 0.5 indicates random chance of discrimination, 0.51 to 0.7 indicates acceptable, 0.71 to 0.8 indicates good, and 0.81 to 1.0 indicates the excellence of the model [19]. The enrichment factor in the pharmacophore model provides an idea about the number of active compounds found from a specific model compared to hypothetically active compounds found from a randomly screened model. The EF factor can range from 1 to >100, where 1 indicates the number of randomly sorted molecules and >100 indicates the least number of compounds need to screen in vitro to find a large number of active compounds. The AUC and EF values found in the study were 0.74 and 1.1, respectively, indicating good discrimination ability and robustness of the pharmacophore model shown in Figure 3. *Molecules* **2021**, *26*, x FOR PEER REVIEW 6 of 26

**Figure 3.** Showing the ROC curve generated during pharmacophore model-based virtual screening. The curve presents the relationship between sensitivity (true positive fraction to the *Y*-axis) and specificity (false positive fraction to the *x*axis) for every possible cut-off. *2.6. Binding Site Identification and Receptor Grid Generation*  **Figure 3.** Showing the ROC curve generated during pharmacophore model-based virtual screening. The curve presents the relationship between sensitivity (true positive fraction to the *Y*-axis) and specificity (false positive fraction to the *x*-axis) for every possible cut-off.

A binding site can be defined as a specific amino acid (AA) residue in a protein to which ligands can binds and is fundamentally important for guiding drug design. Identification of the location of protein binding sites is essential during molecular docking simulation, which helps to generate enough contact points with the protein and significantly increases the docking efficiency [22]. Binding site is evolved to be optimized to bind a

tion found eight binding site residues in the protein. The eight-binding site residues was resided at TRP44, PRO45, ALA123, GLY128, THR129, ILE140, TRP310, and ALA312 in the S1-NTD protein has been represented in a ball shape and shown in Figure 4.

#### *2.6. Binding Site Identification and Receptor Grid Generation*

A binding site can be defined as a specific amino acid (AA) residue in a protein to which ligands can binds and is fundamentally important for guiding drug design. Identification of the location of protein binding sites is essential during molecular docking simulation, which helps to generate enough contact points with the protein and significantly increases the docking efficiency [22]. Binding site is evolved to be optimized to bind a particular substrate, therefore the binding site of the protein has been identified in this study. Analysis of previously identified complex protein–ligands (PDB: 5VYH) interaction found eight binding site residues in the protein. The eight-binding site residues was resided at TRP44, PRO45, ALA123, GLY128, THR129, ILE140, TRP310, and ALA312 in the S1-NTD protein has been represented in a ball shape and shown in Figure 4. *Molecules* **2021**, *26*, x FOR PEER REVIEW 7 of 26

**Figure 4.** The binding site position of MERS-CoV S1-NTD identified from the protein–ligand complex (PDB ID: 5VYH) **Figure 4.** The binding site position of MERS-CoV S1-NTD identified from the protein–ligand complex (PDB ID: 5VYH) structure. Ball shape 3D representation of the binding site with the grid box shown on the left side in the figure, where 2D binding site position has also been represented on the right side of the figure.

structure. Ball shape 3D representation of the binding site with the grid box shown on the left side in the figure, where 2D

binding site position has also been represented on the right side of the figure. PyRx is a grid-based docking program that requires the definition of receptor grid box size before initiating the molecular docking process. Grid box fixation before the molecular docking process helps to generate more reliable scoring to the ligand poses. Therefore, to identify more reliable ligand poses towards the protein, a receptor grid box with PyRx is a grid-based docking program that requires the definition of receptor grid box size before initiating the molecular docking process. Grid box fixation before the molecular docking process helps to generate more reliable scoring to the ligand poses. Therefore, to identify more reliable ligand poses towards the protein, a receptor grid box with a dimension X = 30.69 (Å), Y = 33.36 (Å), and Z = 43.41(Å) has been generated based on previously identified binding residues position of the protein.

#### a dimension X = 30.69 (Å), Y = 33.36 (Å), and Z = 43.41(Å) has been generated based on previously identified binding residues position of the protein. *2.7. Molecular Docking Simulation*

*2.7. Molecular Docking Simulation*  Molecular docking in CADD is an important technique that helps to determine the bound geometry and interaction between a small molecule and a protein at the atomic level. The technique has become an increasingly important tool for drug discovery due to the ability to screen large compound libraries [23]. The technique also helps to determine the behavior and predict how a protein (enzyme) interacts with small molecules (ligands) Molecular docking in CADD is an important technique that helps to determine the bound geometry and interaction between a small molecule and a protein at the atomic level. The technique has become an increasingly important tool for drug discovery due to the ability to screen large compound libraries [23]. The technique also helps to determine the behavior and predict how a protein (enzyme) interacts with small molecules (ligands) to the binding site of target proteins. To elucidate the ligand–receptor binding mechanism, a molecular docking simulation has been performed in this study. The 32 hits identified

to the binding site of target proteins. To elucidate the ligand–receptor binding mechanism, a molecular docking simulation has been performed in this study. The 32 hits identified

has a range between −6.4 and −9.2 kcal/mol provided in Table S1. Based on the binding affinity top (10%), four compounds Amb6600135 (−9.2 kcal/mol), Amb23604132 (−9.1 kcal/mol), Amb23604659 (−8.6 kcal/mol), and Amb1153724 (−8.1 kcal/mol) with zero upper and lower RMSD value have been chosen for further evaluation listed in Table 1.

previously through the structure-based virtual screening process have been docked to the binding site of the MERS-CoV S1-NTD protein. The docking score found for the 32 hits has a range between −6.4 and −9.2 kcal/mol provided in Table S1. Based on the binding affinity top (10%), four compounds Amb6600135 (−9.2 kcal/mol), Amb23604132 (−9.1 kcal/mol), Amb23604659 (−8.6 kcal/mol), and Amb1153724 (−8.1 kcal/mol) with zero upper and lower RMSD value have been chosen for further evaluation listed in Table 1. *Molecules* **2021**, *26*, x FOR PEER REVIEW 8 of 26 *Molecules* **2021**, *26*, x FOR PEER REVIEW 8 of 26 *Molecules* **2021**, *26*, x FOR PEER REVIEW 8 of 26 *Molecules* **2021**, *26*, x FOR PEER REVIEW 8 of 26

**Table 1.** List of the top four compounds and their chemical name, molecular formula, binding affinity (kcal/mol), and pharmacophore fit score. **Table 1.** List of the top four compounds and their chemical name, molecular formula, binding affinity (kcal/mol), and pharmacophore fit score. **Table 1.** List of the top four compounds and their chemical name, molecular formula, binding affinity (kcal/mol), and pharmacophore fit score. **Table 1.** List of the top four compounds and their chemical name, molecular formula, binding affinity (kcal/mol), and pharmacophore fit score. pharmacophore fit score. **Pharmacophore** 

**Table 1.** List of the top four compounds and their chemical name, molecular formula, binding affinity (kcal/mol), and


#### *2.8. ADME Analysis 2.8. ADME Analysis 2.8. ADME Analysis 2.8. ADME Analysis*

drug-likeness, and medicinal chemistry of selected four compounds.

drug-likeness, and medicinal chemistry of selected four compounds.

drug-likeness, and medicinal chemistry of selected four compounds.

drug-likeness, and medicinal chemistry of selected four compounds.

Physico-chemical Properties

Physico-chemical Properties

Physico-chemical Properties

Physico-chemical Properties

*2.8. ADME Analysis*  ADME properties of chemical compounds play an important role in the likelihood of success of a drug. Optimization of the ADME properties can reduce the pharmacokineticsrelated failure in the clinical phases, which is difficult and challenging in the drug development and discovery process [24]. It has been found that early-stage evaluation of ADME can reduce the attrition rates during the clinical drug development phase. Therefore, the study utilized the SwissADME web tool for the early-stage evaluation of ADME properties for selected four compounds. The server evaluated the ADME properties of selected four (Amb6600135, Amb23604132, Amb23604659, and Amb1153724) compounds based on lipophilicity, solubility, pharmacokinetics, medicinal chemistry, and drug-likeness prop-ADME properties of chemical compounds play an important role in the likelihood of success of a drug. Optimization of the ADME properties can reduce the pharmacokineticsrelated failure in the clinical phases, which is difficult and challenging in the drug development and discovery process [24]. It has been found that early-stage evaluation of ADME can reduce the attrition rates during the clinical drug development phase. Therefore, the study utilized the SwissADME web tool for the early-stage evaluation of ADME properties for selected four compounds. The server evaluated the ADME properties of selected four (Amb6600135, Amb23604132, Amb23604659, and Amb1153724) compounds based on lipophilicity, solubility, pharmacokinetics, medicinal chemistry, and drug-likeness properties. ADME properties of chemical compounds play an important role in the likelihood of success of a drug. Optimization of the ADME properties can reduce the pharmacokineticsrelated failure in the clinical phases, which is difficult and challenging in the drug development and discovery process [24]. It has been found that early-stage evaluation of ADME can reduce the attrition rates during the clinical drug development phase. Therefore, the study utilized the SwissADME web tool for the early-stage evaluation of ADME properties for selected four compounds. The server evaluated the ADME properties of selected four (Amb6600135, Amb23604132, Amb23604659, and Amb1153724) compounds based on lipophilicity, solubility, pharmacokinetics, medicinal chemistry, and drug-likeness properties. ADME properties of chemical compounds play an important role in the likelihood of success of a drug. Optimization of the ADME properties can reduce the pharmacokineticsrelated failure in the clinical phases, which is difficult and challenging in the drug development and discovery process [24]. It has been found that early-stage evaluation of ADME can reduce the attrition rates during the clinical drug development phase. Therefore, the study utilized the SwissADME web tool for the early-stage evaluation of ADME properties for selected four compounds. The server evaluated the ADME properties of selected four (Amb6600135, Amb23604132, Amb23604659, and Amb1153724) compounds based on lipophilicity, solubility, pharmacokinetics, medicinal chemistry, and drug-likeness properties. ADME properties of chemical compounds play an important role in the likelihood of success of a drug. Optimization of the ADME properties can reduce the pharmacokineticsrelated failure in the clinical phases, which is difficult and challenging in the drug development and discovery process [24]. It has been found that early-stage evaluation of ADME can reduce the attrition rates during the clinical drug development phase. Therefore, the study utilized the SwissADME web tool for the early-stage evaluation of ADME properties for selected four compounds. The server evaluated the ADME properties of selected four (Amb6600135, Amb23604132, Amb23604659, and Amb1153724) compounds based on lipophilicity, solubility, pharmacokinetics, medicinal chemistry, and drug-likeness properties.

erties. All the compounds have maintained an optimum pharmacokinetics property except the compounds Amb6600135, which has negative Log Po/w value, active P-glycoprotein (P-GP) substrate (Figure S1) and violated the maximum Lipinski's rule of five (RO5) listed in Table 2. On the other hand, the synthesis accessibility of the compound (Amb6600135) was higher (difficult to synthesize) than the other three compounds. Therefore, the com-All the compounds have maintained an optimum pharmacokinetics property except the compounds Amb6600135, which has negative Log Po/w value, active P-glycoprotein (P-GP) substrate (Figure S1) and violated the maximum Lipinski's rule of five (RO5) listed in Table 2. On the other hand, the synthesis accessibility of the compound (Amb6600135) was higher (difficult to synthesize) than the other three compounds. Therefore, the compound has not been considered for further stages of evaluation. All the compounds have maintained an optimum pharmacokinetics property except the compounds Amb6600135, which has negative Log Po/w value, active P-glycoprotein (P-GP) substrate (Figure S1) and violated the maximum Lipinski's rule of five (RO5) listed in Table 2. On the other hand, the synthesis accessibility of the compound (Amb6600135) was higher (difficult to synthesize) than the other three compounds. Therefore, the compound has not been considered for further stages of evaluation. All the compounds have maintained an optimum pharmacokinetics property except the compounds Amb6600135, which has negative Log Po/w value, active P-glycoprotein (P-GP) substrate (Figure S1) and violated the maximum Lipinski's rule of five (RO5) listed in Table 2. On the other hand, the synthesis accessibility of the compound (Amb6600135) was higher (difficult to synthesize) than the other three compounds. Therefore, the compound has not been considered for further stages of evaluation. All the compounds have maintained an optimum pharmacokinetics property except the compounds Amb6600135, which has negative Log Po/w value, active P-glycoprotein (P-GP) substrate (Figure S1) and violated the maximum Lipinski's rule of five (RO5) listed in Table 2. On the other hand, the synthesis accessibility of the compound (Amb6600135) was higher (difficult to synthesize) than the other three compounds. Therefore, the compound has not been considered for further stages of evaluation.

Lipophilicity Log Po/w (Cons) -5.39 4.37 3.70 0.45

Lipophilicity Log Po/w (Cons) -5.39 4.37 3.70 0.45

Lipophilicity Log Po/w (Cons) -5.39 4.37 3.70 0.45

Lipophilicity Log Po/w (Cons) -5.39 4.37 3.70 0.45

pound has not been considered for further stages of evaluation.

**Properties Amb6600135 Amb23604132 Amb23604659 Amb1153724** 

**Properties Amb6600135 Amb23604132 Amb23604659 Amb1153724** 

**Properties Amb6600135 Amb23604132 Amb23604659 Amb1153724** 

**Properties Amb6600135 Amb23604132 Amb23604659 Amb1153724** 

MW (g/mol) 664.43 568.53 540.47 432.38 Heavy atoms 44 42 40 31 Aro. atoms 15 28 28 16

MW (g/mol) 664.43 568.53 540.47 432.38 Heavy atoms 44 42 40 31 Aro. atoms 15 28 28 16 Rotatable bonds 11 5 4 4

MW (g/mol) 664.43 568.53 540.47 432.38 Heavy atoms 44 42 40 31 Aro. atoms 15 28 28 16 Rotatable bonds 11 5 4 4

MW (g/mol) 664.43 568.53 540.47 432.38 Heavy atoms 44 42 40 31 Aro. atoms 15 28 28 16 Rotatable bonds 11 5 4 4

H-bond donors 8 4 5 6 TPSA (Å2) 337.88 155.89 166.89 170.05

H-bond donors 8 4 5 6 TPSA (Å2) 337.88 155.89 166.89 170.05

H-bond donors 8 4 5 6 TPSA (Å2) 337.88 155.89 166.89 170.05

H-bond donors 8 4 5 6 TPSA (Å2) 337.88 155.89 166.89 170.05

**Table 2.** List of pharmacokinetics (ADME) properties includes physicochemical properties, lipophilicity, water-solubility,

**Table 2.** List of pharmacokinetics (ADME) properties includes physicochemical properties, lipophilicity, water-solubility,

**Table 2.** List of pharmacokinetics (ADME) properties includes physicochemical properties, lipophilicity, water-solubility,

**Table 2.** List of pharmacokinetics (ADME) properties includes physicochemical properties, lipophilicity, water-solubility,


**Table 2.** List of pharmacokinetics (ADME) properties includes physicochemical properties, lipophilicity, water-solubility, drug-likeness, and medicinal chemistry of selected four compounds.

#### *2.9. Toxicity Test*

Analysis of toxicity is an important and one of the main steps in drug design that helps to identify the harmful effects of chemical substances on humans, animals, plants, or the environment. Traditional assessment of compounds toxicity requires in vivo animal model, which is time-consuming, expensive, and subject to be ethical considerations [25]. Therefore, computer-aided in silico toxicity measurement of chemical substances can be considered useful in the drug design process. The study utilized the ProTox-II web server to identify the toxicity of the compound computationally, as it is not time-consuming, non-expensive, and requires no ethical considerations. The three compounds (Amb23604132, Amb23604659, and Amb1153724) selected previously through different screening process have been submitted in the ProTox-II web server that determines the acute toxicity, hepatotoxicity, cytotoxicity, carcinogenicity, and mutagenicity of the compounds listed in Table 3. All three compounds have shown no oral toxicity or organ toxicity effect.

**Table 3.** List of compounds toxicity endpoints includes acute toxicity, hepatotoxicity, cytotoxicity, carcinogenicity, and mutagenicity of selected three compounds.

