Identification of potent inhibitors of NEK7 protein using a comprehensive computational approach

NIMA related Kinases (NEK7) plays an important role in spindle assembly and mitotic division of the cell. Over expression of NEK7 leads to the progression of different cancers and associated malignancies. It is becoming the next wave of targets for the development of selective and potent anti-cancerous agents. The current study is the first comprehensive computational approach to identify potent inhibitors of NEK7 protein. For this purpose, previously identified anti-inflammatory compound i.e., Phenylcarbamoylpiperidine-1,2,4-triazole amide derivatives by our own group were selected for their anti-cancer potential via detailed Computational studies. Initially, the density functional theory (DFT) calculations were carried out using Gaussian 09 software which provided information about the compounds' stability and reactivity. Furthermore, Autodock suite and Molecular Operating Environment (MOE) software’s were used to dock the ligand database into the active pocket of the NEK7 protein. Both software performances were compared in terms of sampling power and scoring power. During the analysis, Autodock results were found to be more reproducible, implying that this software outperforms the MOE. The majority of the compounds, including M7, and M12 showed excellent binding energies and formed stable protein–ligand complexes with docking scores of − 29.66 kJ/mol and − 31.38 kJ/mol, respectively. The results were validated by molecular dynamics simulation studies where the stability and conformational transformation of the best protein–ligand complex were justified on the basis of RMSD and RMSF trajectory analysis. The drug likeness properties and toxicity profile of all compounds were determined by ADMETlab 2.0. Furthermore, the anticancer potential of the potent compounds were confirmed by cell viability (MTT) assay. This study suggested that selected compounds can be further investigated at molecular level and evaluated for cancer treatment and associated malignancies.


Density function theory (DFT).
In current study, quantum chemical calculations have been carried out to optimize the structures of title compounds using DFT/B3LYP method with the SVP basis set implemented in Gaussian 09W 50 . All derivatives were optimized in gas and solvent (methanol) phases. The electronic properties 51 of title compounds like E HOMO, E LUMO and energy gap of HOMO and LUMO was calculated. In addition, electrophilicity (Δω±), electronegativity(X), chemical hardness (η), chemical softness (S), chemical potential (μ) and ionization potential (eV) was also calculated. The energetic parameters of all derivatives i.e., dipole moment (Debye), polarizability (α) and optimization energy (hatree) was also calculated in gas and solvent phases as listed in Table 1. The Dipole moment is a global measure of the accuracy of the electron density of a polar molecule. It affects the interactions of a molecule with other molecules as well as electric fields. Dipole moment is the source of understanding and enumerating intermolecular interactions. The values of dipole moment for compounds M7, M8 and M12 were 3.7, 2.62 and 3.8 Debye in gas respectively and these were 4.5, 4.75 and 3.9 Debye in solvent respectively. Polarizability is chief parameter in molecular electronics which corresponds to softness of the compounds.    Table 2.
Another important parameter to determine the reactivity of the compound is global reactivity descriptors. The HOMO/LUMO energies are used to determine these global reactivity parameters. The Global reactivity descriptors like molecular Hardness, Softness, Chemical potential, Electronegativity and Electrophilicity index of the potent derivatives and Dabrafenib is tabulated in Table 2. Compound M7 and M8 showed high value for chemical softness i.e., 5.1. Whereas Dabrafenib showed highest reactivity with softness value of 6.28 in gas phase.
Moreover, FMOs analysis revealed that HOMO orbitals of potent derivatives in gas and solvent phases were localized to benzene ring, nitrogen atom of N-phenylcarbamyl part and piperdinyl ring of the compound M7, M8 and M12 (Fig. 4) which means charge transfer from HOMO to LUMO was due to contribution of π bonds and lone pairs of nitrogen atoms. Whereas, LUMO orbitals of M7 and M12 in gas and solvent phases were localized to N substituted acetamide part of compound which exhibited that electron accepting ability of compounds were due to involvement of substituted aralkyl ring.
Molecular electrostatic potential. The MEP (molecular electrostatic potential) shows the electron density of a molecule and it is used to identify sites of positive and negative electrostatic potentials for nucleophilic and electrophilic attacks 52 . Electrostatic potential is widely used to determine distribution of electronic charge. www.nature.com/scientificreports/ It is very important tool to identify electrophilic and nucleophilic attack on a molecule. It is also powerful tool to identify biological recognition process and hydrogen bonding interactions. Surface and contours provide an idea about interaction of different geometries. Electrostatic potential and electron density of compounds M7, M8, M12 and Dabrafenib are shown in the Fig. 5. It can be seen that electron density is localized uniformly throughout titled molecules but as per ESP bar data, negative ESP is localized only to specific areas of a molecule. This result is obvious because ESP co-relates with partial charges and electronegativity of a molecule.  www.nature.com/scientificreports/ Electrostatic potential values are indicated by different colors as shown in ESP bar. Highly negative electrostatic potential is represented by red color which is related to electrophilic reactivity whereas blue color is indicating highly positive potential relating to nucleophilic reactivity and green color elaborates the zero potential region 53 . It can be seen that highly negative potential was localized to triazole ring of compound M7 and M12. Whereas, compound M8 showed negative potential at oxygen atoms. These findings are important to relate electrophilic and nucleophilic reactivity of a compound.

Molecular docking.
In-silico molecular modeling studies are important part in drug discovery and drug development process; these studies are carried out in order to evaluate binding interactions between top ranked hits and targeted protein. So, the selected Phenylcarbamoylpiperidine-1,2,4-triazole amide derivatives were docked into active pocket of NEK7 using MOE and Autodock which evaluated their binding capacity. Among them, six compounds with binding affinity ranges from − 25.81 to − 31.38 kJ/mol had shown better and comparable binding energies in both docking software. Interestingly compounds with best ligand-protein docked complex also showed good docking scores ( Table 3).

Interpretation of protein-ligand interactions.
Commercially and academically available modelling software's was used for docking of all compounds within active pocket of NEK7 protein. Docking results were sorted out and a compound with highest binding energy and best docking pose was selected for further interpretation. Docking protocol used in current study was validated by re-docking of co crystal ligand ADP ( Fig. 6) with NEK7 protein using MOE and Autodock. ADP was observed to produce good hydrophobic and Hydrophilic interactions with important amino acid residues of activation loop i.e., LYS63, LEU111, GLU82, GLY43, SER46 and GLY112. The Molecular docking experiment in our current study revealed that certain compounds, such as M7, M8, and M12 produced strong bonding and non-bonding interactions with amino acid residues of activation loop i.e., ALA61, ALA165, ASP118, GLY117, ALA116, ARG121, LYS63, GLY43, VAL48, GLY41, LEU111, LEU113, PHE168, ASP115, ALA114, ASP179, ILE40, ASN166, and ILE40.
Docking conformations of compound M7 revealed strong hydrophobic and hydrophilic interactions with active-site amino acid residues. Following amino acid residues were implicated in bonding and non-bonding interactions with M7; GLN44, SER46, LYS63, ILE40, ALA61, ASP115, LEU111, GLU112, ILE95, LEU113, ALA114, VAL48, ASP179, ARG42, GLY41, VAL48, PHE45, and PHE168. The Acetamide part of the compound is substituted with benzene ring bearing ethyl group at ortho position. The hydrophobic interactions between the substituted benzene ring and amino acid residues possessed significant importance. The benzene ring was involved in π-sigma interaction with PHE45, whereas the ethyl group at ortho position was exposed to van der Waals interactions with amino acid residues of active site. In addition, the ethyl group made significant contribution in stabilizing the protein-ligand complex by donating electrons and exerting a positive mesomeric effect (+ M). The 1,2,4 triazolyl-3-thiol ring was engaged in π-alkyl and carbon hydrogen bonding with VAL48 and GLY41 respectively. The π-alkyl interaction contribute significantly in stabilizing the protein-ligand complexes. Furthermore, the Phenyl ring of N-phenylcarbamyl part was implicated in strong alkyl and π-alkyl interactions www.nature.com/scientificreports/ with ALA61, LEU111, and ALA114. LEU111 is a component of the protein's gate keeper residues, which maintain active conformation of the NEK7 protein. Compound M7 formed three hydrogen bonds with distinct amino acids of the active site. The acetamide part of M7 was involved in hydrogen bonding with SER46, which is the part of DFG motif. Furthermore, electronegative oxygen atom of acetamide part was forming a hydrogen bond with LYS63. Another hydrogen bonding was formed between ALA114 and the electropositive hydrogen atom of N-phenylcarbamyl component of compound M7. These hydrogen bonds have significant effect on stabilizing the protein-ligand complex. Moreover, the piperdinyl ring was involved in hydrophobic interactions with ILE40 and VAL48. Furthermore, ASP115, PHE168, ASP179, ILE95, GLU112, LEU113, GLN44, VAL65, GLY43, and ARG42 were also involved in 10 van der Waals interactions with compound M7. These interactions were playing important role in altering the protein active conformation and stabilizing the protein-ligand complex. The docking scores of MOE and Autodock was found to be − 29.56 and − 29.66 kJ/mol, respectively. The docked conformation of compound M8 with NEK7 exhibited potent interactions with the following amino acids: LYS63, GLY41, ILE40, ALA61, ASP115, PHE168, LEU111, GLU112, ILE95, LEU113, ALA114, VAL48, ASP179, and ASN166. The N-substituted acetamide part of compound M8 has an ethyl substituted benzene ring. Ethyl group was present at para position of benzene ring. Although the ethyl group is an electron donor, but its substitution at the para position might have decreased the binding score and binding affinity of compound M8 in comparison to M7. However, ethyl group was involved in hydrophobic interactions with www.nature.com/scientificreports/ ALA114, ILE95 and PHE168 respectively. The benzene ring was involved in important hydrophobic interactions i.e., π-π T-shaped interaction with PHE168, π-alkyl and alkyl-alkyl interactions with ALA61, LEU113 and ILE40 respectively. Another significant interaction was π-donor hydrogen bonding was observed between benzene ring and ALA114. Moreover, It was observed that three important hydrophilic interactions were produced by M8 i.e., Carbon hydrogen bond was formed between electronegative oxygen atom of phenylcarbamyl part and LYS63 residue of active site. The electropositive hydrogen atom of acetamide and phenylcarbamyl part were involved in formation of strong hydrogen bonding with ASP115 and ASP179, respectively. Hydrogen bonding is important molecular interactions which maintain stability of the protein-ligand complex. The hydrophobic interactions were also significant in maintaining stability of the complex. Among hydrophobic interactions, it was noticed that 1,2,4 triazolyl-3-thiol ring formed π-lone pair bonding with ILE40. Whereas, piperdinyl ring of compound M8 was producing π-alkyl interaction with VAL48 residue. All these interactions ultimately disturbing the active conformation of NEK7 protein. Furthermore, 4 van der Waals interaction were also observed with LEU111, GLU112, ASN166 and GLY41. Docking score of M8 compound was found to be − 29.33 and − 28.36 kJ/mol obtained from MOE and Autodock, respectively. Docked conformation of compound M12 showed potent interactions and formed most stable complex through bonding and non-bonding interactions with following amino acids; ALA61, ILE95, ALA114, LEU111, PHE168, VAL48, ASN166, SER46, PHE45, VAL65, CYS79, LEU180, LYS63, ASP179, ILE40, LEU113 and GLU112. Compound M12 exhibited excellent binding energy and binding affinity (ki) of 3.12 µM. It might be due to substitution of benzene ring along with two methyl groups at ortho position. It is discussed that position of alkyl froup was playing great role in determining the inhibitory potential of the compound. Two methyl groups was imparting strong mesomeric effect (+ M) by donating electrons and increasing the resonance of benzene ring. It was observed that aromatic ring present in acetamide part of compound was involved in strong π-sulfur interaction with CYS79. π-sulfur interaction formed between sulfur atom of amino acid and π-electronic cloud of aromatic ring. Meanwhile same aromatic ring was also involved in π-alkyl interaction with VAL65. Second aromatic ring present in N-phenylcarbamyl part was also involved in major stabilizing electrostatic interaction. It was involved in 4 π-alkyl interactions with ALA61, ILE95, ALA114 and LEU111 respectively. Another important interaction was π-anion bonding with ASP179. It was observed that important amino acid residues of NEK7 protein was engaged in interactions by these two aromatic rings. In addition, single carbon hydrogen bond was formed between ethyl group of 1,2,4 triazolyl ring and ASP179. The triazolyl ring was also involved in two π-cation interactions with LYS63 and ASP179. The single hydrogen bond was observed between oxygen atom of acetamide part and LYS63. All these interactions were corresponding to stabilized protein-ligand complex. In addition, piperdinyl ring of M12 was producing π-alkyl interaction with VAL48 residue. Total 8 van der Waals interactions was observed with ASN166, SER46, PHE45, LEU180, PHE168, ILE40, LEU113 and GLU112. Docking scores of M12 was found to be − 29.70 and − 31.38 kJ/mol from MOE and Autodock respectively. Docking scores of compound M12 was comparable to standard Dabrafenib.
In order to determine the inhibition potential of compound with reference to standard FDA approved inhibitor, we have docked Dabrafenib within activation loop of NEK7. Docking score from both software's was obtained as − 33.22 and − 33.51 kJ/mol, respectively. The 2D and 3D interactions were visualized using discovery studio visualizer. Docked conformation of Dabrafenib showed potent hydrophobic and hydrogen boding interaction with primary amino acid residues of active site. Hydrophobic interactions were included π-cation interaction with ARG50, halogen interaction with ILE40, one π-alkyl, two π-π T shaped and one π-sulfur interaction with PHE168. Moreover, aromatic rings of Dabrafenib were involved in π-donor hydrogen bonding with ALA114. In addition to these interactions, standard was producing three conventional hydrogen bonds with ASP115, GLU112 and ASP179. Other amino acids involved in 6 van der Waals interactions were LEU113, ILE95, LEU111, GLY117, ALA165 and LYS38. Most probable 2D and 3D interactions of compound M7, M8, M12 and standard Dabrafenib is shown in Fig. 7.
SeeSAR analysis. SeeSAR analysis of most potent derivatives was carried out which provided virtual display of binding affinities. The structural components of compounds which were contributing favorably were indicated as blue colored coronas whereas those components having negative impact on binding affinities were shown as red colored coronas. Structural components having no contribution were colored as colorless coronas. Size of corona is predictive of contribution of structural component 54 . SeeSAR visualizations of potent derivatives (Fig. 8), which shows that approximately all structural features were contributing favorably but only thiol group of 1,2,4 triazolyl ring and piperdinyl ring was contributing negatively (indicated by red coronas) due to high desolvation energy. Approximately all structural features of Dabrafenib show positive contribution (blue coronas). In addition, Hyde energies of favorable coronas (blue colored) for compound M7, M8 and M12 were comparable to standard Dabrafenib i.e., − 23.43 kJ/mol.

Molecular dynamic simulation. Analysis of RMSD and RMSF.
Calculating the RMSD of each system's backbone served as a preliminary study of the trajectories. Figure 9 shows the RMSD of NEK7 and the NEK7-M12 complex as a function of time. Throughout the simulation period, both systems showed small variances but stayed rather steady. During MD modelling, the RMSD of both systems stayed less than 2 Å, showing that they are highly stable under aqueous environments 55 . The average RMSDs of the NEK7 and NEK7-M12 complexes were found to be 1.27 Å and 1.22 Å, respectively.
Calculating the RMSF provided more insight into the MD simulation results. Figure 10A shows the RMSF of Cα atoms in each NEK7 residue in the absence and presence of M12. The RMSF of virtually all NEK7 residues was less than 2 Å, indicating that both systems are generally stable 56 . Some residues in NEK7 alone and in the NEK7-M12 complex showed higher\volatility, which might be owing to their terminal location or because they  www.nature.com/scientificreports/ ADMET properties. Determination of biochemical process from drug administration to elimination plays important role in lead optimization. Most of the compounds had excellent pharmacological activity but cannot be accepted due to their poor absorption, distribution, metabolism, excretion, and toxicity issues, which are known as ADMET properties. An ideal drug candidate should be administered and absorbed properly into systemic circulation and must be non-toxic and eliminate without affecting the biological activity. These processes are seems to be distinct but they are closely interrelated, so determination of ADMET properties have prime importance in drug discovery process. Since traditional methods were time consuming and many researches till 1990 went into vain due to appearance of undesirable effects in the middle of drug discovery process. So, physicochemical properties, absorption, distribution, metabolism, toxicity, excretion and medicinal properties of 1,2,4 triazole acetamide derivatives were calculated through online web server ADMETlab 2.0. It is an integrated online platform for accurate and comprehensive prediction of ADMET properties. ADME factors  www.nature.com/scientificreports/ that were considered were physicochemical properties, blood brain barrier(BBB), Caco-2 permeability, volume of distribution(VD), P-glycoprotein (PGP) substrate, P-glycoprotein (PGP) substrate, plasma protein binding, Human intestinal absorption (HIA), MDCK permeability, Clearance (CL), Half-life (T1/2), eye corrosion, eye irritation, respiratory toxicity, AMES toxicity, carcinogenicity and synthetic accessibility score. Total 15 derivatives were subjected ADMET study using online web server ADMETlab 2.0. It was observed that all compounds had a positive computed value of human intestinal absorption (HIA), indicating that they can penetrate the intestinal membrane more easily. Furthermore, compound M7 and M12 had a greater value for HIV than standard Dabrafenib. A substance with a positive blood brain barrier value has a higher lipophilicity profile and can easily absorb from plasma membranes. The calculated value for the blood brain barrier (BBB) and blood placental barriers (BPB) were shown to have a high likelihood of being BBB positive. In terms of plasma glycoprotein (PGP) substrate and inhibitor, it was discovered that the output value of all compounds had probability of being a PGP substrate or inhibitor specifically compound M7, M8 and M12. PPB (plasma protein binding) is a significant component in determining drug safety, since compounds with a high PPB value (> 90%) have a narrow therapeutic index, whereas treatments with a low PPB value are considerably safer. Only compound M2 had low PPB values in this investigation, indicating that these drugs have a broad therapeutic index. PPB values of more than 93 percent were found in all compounds, indicating that they have a narrow therapeutic index. Comprehensive ADMET properties are given in "Supplementary Data S1". It was observed that physicochemical properties of all compounds were meeting the criteria of drug like rule i.e., Lipinski rule of five as listed in Table 4.

Cell viability assay.
To support the in-silico studies, the preliminary screening of the most potent derivatives was carried out using in-vitro cell viability assay (MTT assay). The Human HepG2 liver cancer cells were treated with four different concentrations of the selected compounds i.e., 5 µM, 10 µM, 15 µM and 20 µM. The  www.nature.com/scientificreports/ concentrations were selected based on the predicted inhibitory values obtained during docking studies. A linear response of cell death was observed by each derivative. Derivative M12 showed maximum cell death justifying the computational studies where this compound was found best inhibitor of NEK7. The single concentration taxol was used as positive control. The results were calculated by comparing with the total activity control (without inhibitor) i.e., un-treated cells. The % viability graph was generated via graph pad Prism software and is given in the Fig. 11.

Conclusion
The current work used a thorough in-silico strategy to find strong and selective NEK7 protein inhibitors. The chemical reactivity and stability of the compounds were measured using density functional theory studies. In addition, FMOs and MEP were computed, confirming the chemical reactivity of the compounds. Two docking software's, MOE and Autodock, were used to conduct molecular docking investigations in order to improve the accuracy of in-silico molecular modelling. The majority of compounds produced stable protein-ligand complexes with high binding energies, specifically compound M7, M8, and M12 were forming the most stable complexes with the highest binding energies. The most stable compound with the highest binding energy was then subjected to MD simulation experiments, which provided information on the complex's stability over time. MD simulations demonstrated that the complex was in stable state. To establish the physicochemical attributes and toxicity profile of compounds, thorough ADMET studies were conducted, which investigated the compounds' toxicity, drug similarity, and synthetic accessibility. At the end, the results were supported by cell viability assay where the M12 was identified as the best inhibitor of NEK7. Conclusively, all the compounds had a low toxicity profile and acceptable synthetic accessibility score. NEK7 is a potential and selective target for 1,2,4 trizole acetamide derivatives, according to findings of current study.
Computational studies. Comprehensive in-silico investigations were performed with the NEK7 protein to examine the anti-cancer potential of 1,2,4 triazole acetamide derivatives. Molecular docking research, extensive density functional theory investigations, including global and local reactivity descriptors, chemical softness, chemical hardness, and electrostatic potential and Molecular dynamic simulations were among the computational studies. ADMET characteristics were computed using ADMETlab 2.0 to establish the ADME profile of all derivatives 58 .
Density function theory (DFT). The density functional theory was used to optimize the gas phase geometries of the investigated compounds in both gas and solvent phases (DFT). The molecular geometry parameters, frontier molecular orbital (FMO), global and local reactivity descriptors, and molecular electrostatic potential were all collected using DFT (MEP). All of these computations were done with the Gaussian09 59 software, which used the Becke-3-Parameter-Lee-Yang-Parr 50 (B3LYP) method. For all calculations, the SVP basis sets were employed. Gauss View 6 was used to visualize the output files 60 .   49 . The Autodock suite was used to prepare the protein, which involved removing water molecules, adding polar hydrogen atoms, and assigning partial charges to each atom. Autodock tools were also used to prepare missing residues. During the protein production procedure, complexed adenosine diphosphate (ADP) and het atoms were removed from NEK7. Following protein preparation, previously synthesized 1,2,4 triazole acetamide derivatives were selected as testing ligands and docked with NEK7. The active site was selected using rectangular coordinates based on the co-crystal ligand ADP. Grid box XYZ dimensions were set to − 12.348, − 33.512, and − 48.605, respectively. The search spacing was adjusted at 0.55 Å in order to cover the largest number of amino acids in the activation loop. The number of points in the X, Y, and Z dimensions, on the other hand, were set to 60, 60, and 60, respectively. The number of docking postures was set to 100 with a population size of 300 to cover all potential ligand binding locations. Following the establishment of the necessary parameters, ligands were docked with protein in order to identify probable hits and generate reliable binding poses. The binding poses of the compound with the highest docking score were visualized using Discovery Studio Visualizer 17.2. MOE's protein preparation module was used to inspect the protein structure for any missing atoms or residues and make any necessary adjustments. MOE's protein preparation process comprised applying gas tier charges through the MMFF94x forcefield, adding hydrogen atoms, removing water molecules, 3D protonation of the structure, and minimizing the protein structure to a chosen gradient. The binding site of NEK7 was defined using the original co-crystal ligand (ADP). The dummy atoms were generated at the binding location using MOE's site finder program. The docking algorithm was configured to use the triangle matcher placement approach 61 . Furthermore, the induced fit refinement approach was applied to generate docking positions. The triangular match algorithm was programmed to create 100 poses, but the induced fit refinement approach produced ten poses. Furthermore, the default GBVI/WSA dG technique was used as a docking function in MOE 62 . Autodock and MOE docking conformations with the lowest binding energies were chosen for further study.
Preparation of ligands. Ligand databases were created by generating the structures of N-alkyl/aralkyl/aryl acetamide triazole derivatives from Saima, Muzaffar et al. 39 . These molecules were previously tested for their ability to inhibit the 15-lipoxygenase enzyme. Chemdraw Ultra 12.0 63 was used to draw the structures of all derivatives based on the IUPAC name of the compounds. The aryl ring was replaced with an ethyl group, and 3D optimization and energy reduction were performed at 0.1 gradient using Chem3D pro 12.0 64 . Compounds were saved to SDF format after energy reduction. These SDF files were then transformed into the appropriate format for the docking programme, such as PDBQT for Autodock and SMILES for SeeSAR analysis. These modifications were carried out with the help of the OpenBabel GUI 65  Visualization. Autodock docking conformations were displayed using Discovery studio visualizer 17.2 66 .
The best docked conformation's 2D and 3D interactions were generated using the Discovery Studio visualizer. www.nature.com/scientificreports/ MOE, on the other hand, has an inbuilt visualizer tool that may be used to visualise 2D and 3D interactions of the best docked conformation. Furthermore, to identify a persuasive rationale for NEK7 inhibitor binding affinity, SeeSAR analysis of the most active and least active ligands was performed using SeeSAR software 67 , which gives a visual depiction of binding affinity. The structural features of ligands that were not contributing favourably to overall binding affinity were indicated by red coloured coronas, whereas structural features of ligands that were contributing favourably were indicated by blue coloured coronas; the larger the corona, the greater the contribution. No-contribution structural elements were not coloured. SeeSAR was used to create a total of ten docking postures for each ligand 54 . Validation. Validation of docking protocol was done by redocking co-crystal ligand and Dabrafenib into the active pocket of NEK7 protein. Assessment was done on the basis of calculated RMSD value. Only those docking poses were considered successful whose RMSD value of docking pose and the experimentally determined conformation of a ligand was less than 2.0 Å 68 . Molecular dynamic simulation. Molecular dynamics (MD) simulations were used to investigate the stability and interaction of the hit molecule (M12) acquired from prior research. Gromacs-2018.1 was used to run the MD simulations, which used the amber99sb-ILDN force field 69,70 . For simulation investigations, the docked complex was used as the starting point. The topology of M12 was created using AmberTools21's Antechamber and the AM1-BCC charge model 71 . TIP3P water model was used to solvate NEK7 and its complex with M12, and then the energy of each system was reduced using steepest descent minimization to eliminate the weak Van der Waals connections. Both systems were then equilibrated in two processes. First, NVT was equilibrated at constant temperature and volume for 1 ns using a V-rescale thermostat 72 . The Parrinello-Rahman barostat was used to produce the second equilibration for NPT at constant pressure and temperature for 1 ns 73 . The equilibration coordinates were used to run a 100 ns MD simulation, with 10,000 frames of each system stored from their individual trajectories. Before analysis, both trajectories were treated to PBC adjustments. Gromacs tools were used to conduct all of the analyses.
ADMET properties. The determination of ADMET characteristics is critical for rolling out unfavorable effects of a drug candidate at the early stages of the drug development process. For the in-silico estimation of ADMET characteristics, efficient and accurate online prediction models were constructed. The online in-silico prediction model ADMET lab 2.0 58 was used to determine ADMET attributes such as absorption, distribution, metabolism, excretion, toxicity, and physicochemical properties of all chemicals, and drug similarity properties were assessed using the Lipinski rule of five 58 . Chemdraw Ultra was used to convert all compounds to SMILES format, and these SMILES structures were then uploaded to the ADMET lab 2.0 web server. It also makes using the JMSE editor to create desired structures easier. Within loading the SMI structure, the submit button was used to submit the data, and it returned ADMET attributes in pdf and spreadsheet format, which could be downloaded after a few minutes.
Cell viability measurement (MTT assay). In order to determine the anticancer effect of the most potent compounds, the Human HepG2 liver cancer cells were cultivated in Dulbecco's Modified Eagle's Medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 100 units/ml penicillin, and 100 g/ml streptomycin and kept at 37 °C in a humidified environment with 5% CO 2 . Cells were treated with compounds dissolved in DMSO with the final concentration of DMSO (0.05%). The MTT assay was performed as discussed earlier 74 . In a brief, HepG2 cells were treated for 48 h with various doses of the potent derivatives. Then 10 µl solution of MTT (5 mg/mL) was added to each well after 48 h of post-incubation. The plate was covered with aluminium foil and kept in the incubator at 37 °C for about 4 h. The medium was taken out and replaced with 150 µL of DMSO to dissolve the remaining purple formazan precipitate that was still there. A microplate reader (Thermo Scientific) was used to record the absorbance at 570 nm. This colorimetric assay looked at how the mitochondrial enzymes of the metabolically active cancer cells reduced MTT to purple formazan, which made the cancer cells look more alive. The Cell viability was determined as a percentage. The experiment was done three times, and the results are shown in standard deviation and mean values.