Using Hansch's approach, introduced by Hansch and Fujita, we evaluated the QSAR of 46 substituted benzoxazole derivatives. Before employing biological activity data in a QSAR analysis, experimentally determined MIC values are converted to –log MIC or pMIC (in micromole) to make all values positive. Various molecular descriptors such as log P, molar refractivity (MR), Kier's molecular connectivity (0χ, 0χV, 1χ, 1χv, 2χv) and shape (k1, kα1, kα2,kα3) topological indices, Randic topological index (R), Balaban topological index (J), Total energy (TE), Wiener topological index (W), lowest unoccupied molecular orbital (LUMO) and energies of highest occupied molecular orbital (HOMO), electronic energy and dipole moment (µ) were calculated for substituted benzoxazole derivatives and values of selected descriptors are displayed in Table 3.
QSAR model for antimicrobial activity against Bacillus subtilis:
The preliminary investigation was carried out in terms of correlation of molecular descriptors with antimicrobial activity of Bacillus subtilis [Table 4]. The initial correlation r = 0.902 was observed with the electronic parameter R (Eq….1) which is a valid one QSAR model.
The equation comes out as:
pMICBS = -0.276 R – 5.065 (Eq…..1)
n = 46, r = 0.902, q2= 0.795, SD= 0.237, F= 6.932
Where, n = number of data points, r = correlation coefficients, r2 = obtained by leave one out method, q2= cross validated, SD= standard deviation or standard error of the estimate, F= fischer statistics
The value of r was close to 1 and value of q² was also more than 0.5 which indicated that the QSAR model was valid one. Further the validity of QSAR model was demonstrated by plotting observed against predicted activity (Fig. 1). The observed values were plotted against residual values to determine the presence of the systemic error (Fig. 2). The propagation of residuals on both sides of zero revealed that the construction of the QSAR model was free of systemic errors.
QSAR model for antimicrobial activity against Escherichia coli:
The preliminary investigation was carried out in terms of correlation of molecular descriptors with antimicrobial activity of Escherichia coli [Table 5]. The initial correlation r = 0.818 was observed with the electronic parameter 1χ V (Eq….2).
The equation comes out as:
pMICEC = -0.427 1χ V + 5.665 (Eq….2)
n = 46, r = 0.818, q2 = 0.637, SD = 0.340, F = 2.948
But as the value of r was not close to 1 so, the model was not ideal. Therefore, two outliers (compound no. 23, 31) were identified and removed which improves the value of r = 0.856 and in process of improving the q2 value. Further the validity of QSAR model was demonstrated by plotting observed against predicted activity (Fig. 3). The observed values were plotted against residual values to determine the presence of the systemic error (Fig. 4).
pMICEC = -0.421 1χ V + 5.598 (Eq…3)
n = 44, r = 0.856, q2 = 0.702, SD = 0.294, F = 8.466
QSAR model for antimicrobial activity against Salmonella typhi:
The preliminary investigation was carried out in terms of correlation of molecular descriptors with antimicrobial activity of Salmonella typhi [Table 6]. The initial correlation r = 0.860 was observed with the electronic parameter R (Eq…4) which is a valid one QSAR model.
pMICST = 0.270 R + 5.180 (Eq….4)
n = 46, r = 0.860, q2 = 0.707, SD = 0.289, F = 1.350
The value of r was close to 1 and value of q² was also more than 0.5 which indicated that the QSAR model was valid one. Further the validity of QSAR model was demonstrated by plotting observed against predicted activity (Fig. 5). The observed values were plotted against residual values to determine the presence of the systemic error (Fig. 6).
QSAR model for antimicrobial activity against Klebsiella pneumoniae:
The preliminary investigation was carried out in terms of correlation of molecular descriptors with antimicrobial activity of Klebsiella pneumoniae [Table 7]. The initial correlation r = 0.838 was observed with the electronic parameter R (Eq..5).
pMICKP = -0.295 R + 5.569 (Eq…5)
n =46, r = 0.838, q2 = 0.664, SD = 0.346, F = 2.731
But as the value of r was not close to 1 so, the model was not ideal. Therefore, five outliers (compound no. 24, 15, 14, 10, 6) were identified and removed which improves the value of r = 0.911 and in process of improving the r value, the correlation value of R and value of q2 also increased (Eq….6). Further the validity of QSAR model was demonstrated by plotting observed against predicted activity (Fig. 7). The observed values were plotted against residual values to determine the presence of the systemic error (Fig. 8).
pMICKP = -0.339 R + 6.244 (E…6)
n = 41, r = 0.911, q2 = 0.815, SD = 0.257, F = 7.127
QSAR model for antimicrobial activity against Aspergillus niger:
The preliminary investigation was carried out in terms of correlation of molecular descriptors with antimicrobial activity of Aspergillus niger [Table 8]. The initial correlation r = 0.858 was observed with the electronic parameter 1χ (Eq….7).
pMICAN = -0.264 1χ + 5.144 (Eq……….7)
n = 46, r = 0.858, q2 = 0.702, SD = 0.248, F = 1.697
But as the value of r wa not close to 1 so, the model was not ideal. Therefore, nine outliers (compound no. 46, 39, 38, 28, 27,25, 24, 19, 10) were identified and removed which improves the value of r = 0.927 and in process of improving the r value, the value of q2 also increased (Eq….8). Further the validity of QSAR model was demonstrated by plotting observed against predicted activity (Fig. 9). The observed values were plotted against residual values to determine the presence of the systemic error (Fig. 10).
pMICAN = -0.254 1χ + 4.954 (Eq.…8)
n = 37, r = 0.927, q2 = 0.841, SD = 0.178, F = 8.056
QSAR model for antimicrobial activity against Candida albicans:
The preliminary investigation was carried out in terms of correlation of molecular descriptors with antimicrobial activity of Candida albicans [Table 9]. The initial correlation r = 0.900 was observed with the electronic parameter 1χ (Eq…9).
pMICCA = -0.414 1χ + 7.15 (Eq…9)
n = 46, r = 0.900, q2 = 0.791 SD = 0.360, F = 9.913
But as the value of r was not close to 1 so, the model was not ideal. Therefore, seven outliers (compound no. 28, 22, 14, 10, 9, 3, 2) were identified and removed which improves the value of r = 0.919 and in process of improving the r value, the correlation value of 1χ and value of q2 also increased (Eq..10). Further the validity of QSAR model was demonstrated by plotting observed against predicted activity (Fig. 11). The observed values were plotted against residual values to determine the presence of the systemic error (Fig. 12).
pMICCA = -0.425 1χ + 7.323 (Eq..…10)
n = 39, r = 0.919, q2 = 0.830, SD 0.316, F = 7.462
From the above data, we can conclude that all QSAR models are valid one. Further, the validity of QSAR model has been demonstrated by the comparison of observed, predicted and residual values of each of the organisms taken which is shown in Table 10 and Table 11 respectively.
Molecular Docking
The DNA Gyrase has been validated as the primary molecular target for benzoxazole derivatives as antimicrobial drugs. Here, these compounds were docked into active site of DNA Gyrase (PDB ID: 4KTN) to know ligand–protein interactions. The oxygen atom of compound 26 amide nucleus formed a hydrogen bond with Gln 119, Ser 122 in the 2D ligand interaction diagrammatic view(Fig. 13). Compound 14 amino group established a hydrogen bond with Asn 48(Fig. 14), while compound 13 amino group formed a hydrogen bond with Gln 85 and Ser 122 (Fig. 15). Compound 10 oxygen atoms created a hydrogen bond with the water molecule, and the water molecule interacted with Lys 138 (Fig. 16). The amino group interacted with Ser 122 and the water molecule. Comopund 3 exhibited Pi-Pi stacking interaction with Hie 101 (Fig. 17), while the reference medication (Ciprofloxacin) showed a water molecule interaction (Fig. 18). In negative terms, the docking scores, glide energy, and glide emodel values were shown. The more negative docking score, stronger ligand's affinity for binding to the receptor. All the docking scores are given in Table 12. According to the docking studies, benzoxazole derivatives demonstrated that compounds 26, 14, 13, 10, 3 exhibited promising antimicrobial activity and the docking results were also correlated with the experimental data. Table 13 shows the docking score of best 5 molecules.
ADME Analysis
ADME Analysis mainly used to check the likeliness properties of the drug, All the Pharmacokinetic properties were estimated by SwissADME. The compounds (4, 5, 6, 8, 10, 16, 19, 20 showing 1 voilation of Lipinski rule of five) all remaining derivatives did not violation of any Lipinski rule of five as shown in Table 14. These findings support the compounds' drug-like characteristics, and they are expected to be orally accessible.