Based on our study on the interactions between the hydrophobic and hydrophilic areas of the Brest cancer active site provided us with significant information for developing inhibitors. Table 1 shows the geometry optimisation and Avogadro tool results for the 23 suggested compounds with varied heterocyclic rings, functional groups, and 2-amino thiazole chain linkers. During the docking research, the best short energy conformers for each ligand that could be discovered were chosen.

The aforementioned conformations were imperilled to Insilco experiments in order to investigate the interactions between residues, H-bonds, and obligatory energy scores. The novel Aurora kinase pattern, PDB 3D structures: 1MQ4, was produced from the protein catalogue (www.rcsb.org) (Behera et al. 2021). Missing residues were restored using the Modeller V 9.23 program19 after hydrogen atoms were added and pre-existing ligands were removed. The cocrystallized ligand was re docked to 1MQ4 active sites to determine the docking restrictions. Proteins were created through the addition of polar hydrogens in line with Kollman’s unified atomic charges. The PDB fixer tool20 was used to analyse the histidine protonation residue states by assigning ND1 to zinc-bound histidines and NE2 to residual histidines.

The ligand categorizer was constructed by accumulating polar hydrogens and employing Gasteiger charges. The Auto grid selection allows users to generate the energetic map using a distance-dependent dielectric continuous utility by identifying an active site and setting the element size to 60 * 60 * 60 points with 0.503 A⁰ spacing. The RMSD scores between the reference and predicted components were used to assess whether the docking simulation properly foretold a close-match docked pose or not (Rosário-Ferreira et al. 2021).

The “Desmond V 5.9 software” was used to conduct simulations of molecular dynamics in order to investigate how the solvent system altered the molecular makeup of the proteinligand complex (Schrodinger 2019-3). The OPLS 4 force field was used for the docked complex’s MDS (ligand 1MQ4). For the sake of performing dynamic forces simulations, the complex’s site has been center filled in an orthorhombic cubic punnet, and TIP3P water molecules and buffers have been inserted at a distance of 10 A⁰ between the box edge and the protein atom (Siva Kumar et al. 2022). The complex type boundary condition box volume has also been estimated, as have counter ions such as Na+ and Cl- that were inserted to randomly neutralise the system. A clinically tested 2-amino thiazole substituted imidazolidinone derivative was used as the reference chemical for comparing simulation results. The anticipated lead complex molecule 1MQ4 has 458 000 A⁰ box volume and 28 252 atoms with 8124 waters for MD simulations. The complex is made up of over 49998 atoms and 9209 fluids. The Desmond procedure was used to diminish the solvated built system before relaxing using the OPLS 4 force field parameters. To keep heavy atoms on the solute in check, the system was simulated using the Berendsen NVT ensemble at a temperature of 10 K. The MDS was performed using the isothermal isobaric ensemble (NPT) at 300 K, 1 atom, and a thermostat relaxation time of 200 ps (NPT).

Using the Nose-Hoover thermostat and Martyne-Tobias-Klein barostat practices, the temperature and compression balances were kept at 300 K and 1 atom, respectively. Every 50 ps, the simulation progress were methodically recorded. Succeeding the replication phase, which needs 100 ns of manufacturing, the NPT ensemble was initiated. The simulated interaction diagram was utilised to investigate the trajectories of the frames, which assisted in the discovery of variations (Khetmalis et al. 2022).

n_tr = 39 n_pred = R² = 0.8429 R²adj = 0.7061 R²-R²adj = 0.0367 LOF = 0.0541 K_xx = 0.3885 Delta_K = 0.0714 RMSE_tr = 0.1762 MAE_tr = 0.1433 RSS_tr = 1.0241 CCC_tr = 0.8525 s = 0.1912 F = 20.2223

Q_2loo = 0.7558 R²-Q_2loo = 0.1101 RMSE_cv = 0.2105 MAE_cv = 0.1702 PRESS_cv = 1.4625 CCC_cv = 0.7954 Q_2LMO = 0.6076 R_2Yscr = 0.1289 Q_2Yscr = -0.2219 RMSE_AVYscr = 0.3239 RMSE_ext = 0.3697

MAE_ext = 0.3167 PRESS_ext = 0.8199 R²_ext = 0.6726 Q₂-_F1 = -0.2977 Q_2-F2 = -0.3627 Q_2-F3 = -0.1322 CCC_ext = 0.6877 r²_{m aver} = 0.2370 r²_{m delta} = 0.4976

Using alkyl linkers on imidazolidinone and a 2-amino thiazole ring on the tail, the 2-amino thiazole imidazolidin-4-one core structure was investigated. ESI the dataset chemicals used to create the QSAR models are shown in Table 1 along with the best leads discovered. Twentythree generated compounds were exposed to the descriptor deviousness tools of Chem DesChemoPy services using the lead structure as a template to determine the appropriate EstateVSA5, PSA, MoRSEP3, MATSp5, and RDFC24 values (Fig. 3). In most cases, the MoRSE descriptors (molecular representation of structure by electron diffraction) included weights for volume, atomic charges, Sanderson’s electronegativity, and other variables in addition to the scattering parameter, which ranged from 1 to 23. (Table 1) displays the molecular descriptor values and structural details of 23 of the best-designed compounds along with estimates of their projected bioactivity at nano and sub-nano molar scales.

Figure 3. 

Leads generated from QSAR and docking studies (compounds 1, 7, 9, 13, 19) structural stability analysis upon ligand docking.

Applicability domain. Hat diagonal values versus standardized residuals.

The calculated novel derivatives with greatest pIC₅₀ standards remained exposed to molecular docking studies for necessary empathy and H-bond interfaces estimate. Molecular docking was performed to investigate the potentiality of the chosen compounds, to engage the active cavity of Aurora Kinase in a way that would disrupt its tumorigenic activity. Many poses were observed and the poses with least score for each compound were considered. Interestingly, all the selected compounds except 2, 3, 10, 11, 12, 15 and 17 showed good interaction with the target in terms of glide score and energy; in particular than the reference compounds Tamoxifen and Arimidex. Results of the binding scores and the interactions with the amino acid residues are shown in Table 2. The active site pocket lies within the kinase domain of the target protein, which is essential for its function. Compound 19 having imidazolidionone and sulfonyl groups showed a least glide score of -9.488KJ/mol^-1 with the formation of five hydrogen bonds. The hydroxyl benzene group of compound 19 fits perfectly into the hydrophobic core of the active site sharing two hydrogen bonds with the residues Glu211 and Ala213. The core thiazole group shared two hydrogen bonds with Asp274 and Lys258 respectively. Moreover, an attractive non covalent pi-pi interaction was observed between the core group and Trp277 of the target. Compound 9 with substituted Aromatic P- Hydroxy benzaldehyde interacted through five hydrogen bonds with the binding site residues of the target Aurora Kinase possessing a docking score of -7.894KJ/mol^-1. It was noted that the bromo and hydroxy phenyl groups of both the compounds (Compound 19 and 9) are located deeply inside the hydrophobic core due to their hydrophobic nature.

The superlative conformation of Derivatives 7- 1MQ4 multifaceted from Glide demonstrated obligatory empathy as -7.653KJ/mol^-1. The sulfonyl and oxo groups of the compound showed interactions with the residues Lys143, Lys162, Asn261 and Asp274 through hydrogen bonding. Compared to the previous two compounds, this occupied different binding mode in the active pocket. The core thiazole imidazolidinone group shared four hydrogen bonds with the residues Lys143, Lys162, Asn261 and Asp274. In addition, a pi-pi interaction with the residue Trp277 was observed. Compound 1 shared five hydrogen bonds with the residues Lys162, Lys258, Asn261, Glu211 and Ala213. The thiazole group of the core component and the carbonyl sulphide of R1 shared hydrogen bonds with a single residue Lys 162. Interestingly, compound 19, 9 and 1, 7 showed similar binding modes respectively.

The best conformation of compound 13 - 1MQ4 complex from Glide displayed binding affinity as -7.520KJ/mol^-1. Similar to compound 1, the thiazole group of the core component and the carbonyl sulphide of R1 shared hydrogen bonds with a single residue Lys 162. Apart from that the residues Asp274, Asn261 and Lys143 were found to have H-bonded interactions. Compound 23 also showed good interaction with a docking score of -7.520 KJ/mol. Four hydrogen bonds involving the residues Asn261, Asp274 and Lys143 were observed with Asp274 forming two hydrogen bonds.

The compound 4 with substituted Aromatic cinnamaldehyde interacted with ALA 213 and

Oxo amino group interact with LYS 143, ASN 261 and ASP274 by the hydrogen bonding. The compound 22 with amino thiazole group moiety interacted with LYS 162, ASP 274 and Oxo and amino group interacted LYS 143, ASN 261, via H-bond. Bromo benzene formed pi-pi interactions with Trp277. Also it is evident from the table and figures, that all the above mentioned compounds showed better binding affinity than the control drug Tamoxifen. In addition, compounds 14, 8, 5, 20, 16, 21, 6 and 18 also showed good binding affinity than tamoxifen as evident from (Table 2) and their interaction are shown in Suppl. material 1: fig. S1.

Although molecular docking analyses have revealed the complex binding mechanisms (protein-inhibitor), molecular dynamics simulation can detect even the slightest disagreement. To find out as much as possible about the atomic specifics in the solvent system, the compounds 1, 7, 9, 13, and 19 with the best interactions and energy docked conformation were studied (Fig. 3). For the MDS investigation, the active conformation with the best H-bond interactions, docking score and energy values were taken into account. We further looked into the simulation after a 100 ns simulation was run to determine the stability of the protein-inhibitor combination. Several parameters such as RMSD and RMSF of the protein and ligand, their respective torsion profiles and the histogram representing the ligand-protein

To study the stability of the bound complexes, the RMSD and RMSF calculations were performed. RMSD reflects the mobility of an atom during the MD simulation trajectory; with higher RMSD corresponding to higher mobility, whereas, lower RMSD corresponds to lower mobility. The RMSD plot indicates that the compounds 19, 9 and 7 stabilized shortly after commencing the simulation and found no major deviations throughout the simulation. Also a similar pattern of fluctuation was observed for both protein and ligand except for compound 13. For compound 19, the ligand RMSD varied from 0.91 to 2.48Å, whereas protein RMSD showed variation from 1.20Å to 2.19Å (Fig. 4). There was a small increase and decrease in RMSD of the protein and compound 19 respectively around 40–45 ns of simulation time. An initial drift was observed till 30 ns for compound 1, after which certain stabilization could be observed. In case of Compound 9-AurK complex, the protein showed a small decrease in RMSD between 48–51, suggesting minor conformational changes in the protein Compound 13 showed an increase in RMSD till 60 ns and started stabilizing further, whereas the protein undergone minor conformational change initially which might be due the flexibility of the residues. Moreover, compound 19 and compound 1 showed minimal deviation than the control drug tamoxifen.

Figure 4. 

The root mean square deviation (RMSD) of 1MQ4 protein during 100 ns MD, Compounds 1, 7, 9, 13, 19.

Gly173 is fluctuated to around 2.70 Å upon binding of Compound 19. Secondary structural analyses showed a loop to stand conversion around the 180^th residue, induced by compound 19. In case of Compound1-AurK complex, fluctuations were observed in regions of Thr287 to Gly291. From the ligand RMSF graph, the carbonyl and the thiazole groups of compound 19 are exposed towards the surface, which is also evident from the 2D plot (from docking). The benzyl group of compound 1 showed higher fluctuations than the other regions, suggesting their exposure to the surface. The residues Thr287 and Thr288 have fluctuated a bit upon binding of compound 7. Notably, these fluctuations lie under the permissible range of 1–3Å. No major residual fluctuations were observed in Compound 9-AurK complex. However, of the residues GLY 142, LYS 143, PHE 144, Glu 170, LYS 171 and ALA 172 showed an increase in flexibility (>3Å) interaction with compound 13. Ligand RMSF (Fig. 5) also confirmed the exposure of the bromo-benzyl group towards the surface as evidenced from the 2D molecular docking interaction (Fig. 7).

Figure 5. 

The root mean square fluctuation (RMSF) of 1MQ4 protein during 100 ns MD, representing local changes along the protein chain for Compounds 1, 7, 9, 13, 19.

Interactions histogram showed the interactions lasting over different simulation time and the connections that past additional than 30% of the simulation time are measured. It shows that most of the interactions of the docked pose were retained during simulation. Water-bridge, Hydrogen bonding following hydrophobic interactions at stable areas LEU 139, GLY 142, LYS 143, PHE 144, GLY 145, VAL 147, ALA 160, LYS 162, LYS 162, LEU 164, GLU 211, ALA 213, LYS 258, GLU 260, ASN 261, LEU 263, ASP 274 and TRP 277 were found in compound 19 during MDS (Fig. 6). Hydrogen bonds shared by the residues GLU 211, ALA 213, LYS162, ASN 261 and ASP 274 were retained. In addition, an aromatic amino acid PHE144 interacts with the keto group of the compound for 41% of simulation time. Compound 1 showed an additional hydrogen bond with LYS 162 and it was found that the compound was capable to interact retain interaction with the lysine residues compared to other interactions. Similar to Compound 1, compound 7 formed an additional hydrogen bond with Lys162, accompanied by the thiazole and keto group of the compound. However it lost the hydrogen bonds shared by other residues during docking. Unlike compound 1 and 7, compound 9 is capable of preserving the hydrogen bonds shared by the residues Asparagine and arginine. It was also noted that ASP 274 shared two hydrogen bonds with the two N-H groups of the compound, maintaining the same for more than 75% simulation time. To summarize all the above mentioned compounds were found to have good interaction with the active pocket at the kinase domain of aurora kinase, disrupting its action, thereby inhibiting cell growth and induce apoptosis. Overall, it can be concluded that the compounds in particular compound 19, compound 1, compound 9 and compound 7 can be used as a potential inhibitor in targeting the enzyme aurora kinase upon experimental validation.

Figure 6. 

The plot represents the hydrogen bonding interactions of compounds 1, 7, 9, 13, 19 with esteem to deposits of 1MQ4 throughout 100 ns MD simulation.

Figure 7. 

Two-dimensional diagram of Compounds 1, 7, 9, 13, 19 interactions (top) during 100 ns MD simulation and hit molecule compound 19_1MQ4 interaction (bottom).

Estimation of binding energy using the molecular mechanics-generalized born surface area (MMGBSA) system helps in identification of ligands that bind effectively. The bound complexes identified by molecular docking and MDS were further discovered by performing MMGBSA binding free energy calculations. The key contributors for binding namely Hbonding, lipophilic interactions, electrostatic interactions and Van der Waals energy were also estimated from the trajectories attained throughout MDS and depicted in (Table 3). The allosteric site of the protein Aurora kinase is predicted to be more preferable for the binding of the chosen ligands (Lokhande et al. 2022). The corresponding binding free energies of the compounds are -125.18, -102.40, -96.91, -88.51 and -84.88 Kcal/mol. for molecules 19, 7, 9, 1 and 13 respectively indicating the binding affinity in the order of Mol19 > Mol7 > Mol9 > Mol1 > Mol13.