In silico investigations of natural-product tethered 1,4-naphthoquinones as inhibitors of Staphylococcus aureus

초록

A series of 46 compounds with 1,4 naphthoquinone moieties substituted at positions -2 and 3- showing inhibitory activity against Staphylococcus aureus, were collected from the literature. The molecular descriptors of the compounds were calculated using PaDEL and the ChemDes online server for the development of multiple linear regression (MLR) based predictive 2D QSAR models using QSARINS v2.2.4. The MLR models were developed taking two, three, four, and five descriptors on genetic algorithm (GA) up to ten thousand. The models were evaluated on the parameters of fitting criteria and internal and external validations to assess their robustness, reliability, and predictive capability. The statistically best model using five descriptors ALogP, MATS5e, VR2_DzZ, VE2_Dzs, and MIC1 showed the correlation coefficient (R-2 = 0.8955) and cross-validated R-2 (Q(LOO)(2) = 0.8564). The model showed lower values of root mean squared error (RMSEtr = 0.0886), mean absolute error (MAE(tr) = 0.0656), and greater concordance correlation coefficient (CCC = 0.9449), showing good fitness and prediction accuracy. The developed QSAR model was applied to an in-house library of 72 virtual molecules having natural and heterocyclic moieties as substituents at 2- and 3-positions of 1,4-naphthoquinone. Thirty-four virtual derivatives showing high predicted MIC (<100 mu M) were explored for their binding affinity against five target proteins playing a vital role in S. aureus pathogenesis. The compounds showing overall significant binding affinity with S. aureus proteins and favorable physicochemical and ADMET predictions were subjected to 100 ns molecular dynamics simulation studies. We propose 1,2,3-triazole and uracil-substituted 1,4-naphthoquinone derivatives for further in vivo studies.

키워드