Combined machine learning and pharmacophore based virtual screening approaches to screen for antibiofilm inhibitors targeting LasR of Pseudomonas aeruginosa

Abstract

Pseudomonas aeruginosa, a virulent pathogen affects patients with cystic fibrosis and nosocomial infections. Quorum sensing (QS) mechanism plays a crucial role in causing these ailments by mediating biofilm formation and expressing virulent genes. A novel approach to circumvent this bacterial infection is by hindering its QS network. Targeting LasR of las system serves beneficial as it holds the top position in QS system cascade. Here, we have integrated machine learning, pharmacophore based virtual screening, molecular docking and simulation studies to look for new leads as inhibitors for LasR. Support vector machine (SVM) learning algorithm was used to generate QSAR models from 66 antagonist dataset. The top three models resulted in correlation coefficient (R²) values of 0.67, 0.86, and 0.91, respectively. The correlation coefficient (R²_test) values on external test set were found to be 0.62, 0.57, and 0.55, respectively. A four-point pharmacophore model was developed. The pharmacophore hypothesis AAAD_1 was used to screen for potential leads against MolPort database in ZincPharmer. The leads which showed predicted pIC50 value of >8.00 by SVM models were subjected to docking analysis that reranked the compounds based on docking scores. Four top leads namely ZINC3851967 N-[3,5-bis(trifluoromethyl)phenyl]-5-tert-butyl-6-chloropyrazine-2-carboxamide, ZINC4024175 4-Amino-1-[(2R,3S,4S,5S)-3,4-dihydroxy-5-(hydroxymethyl)oxolan-2-yl]-2-oxopyrimidine-5-carbonitrile, ZINC2125703 N-[(5-Methoxy-4,7-dimethyl-2-oxo-2H-chromen-3-yl)acetyl]-beta-alanine, and ZINC3851966 N-[3,5-Bis(trifluoromethyl)phenyl]5-tert-butylpyrazine-2-carboxamide were selected. These compounds were checked for its stability by performing a molecular dynamics simulation for a period of 100 ns. The ADME properties of the leads were also determined. Hence, the compounds identified in this study can be used as possible leads for developing a novel inhibitor for LasR.

Communicated by Ramaswamy H. Sarma

Keywords:

• Pseudomonas aeruginosa
• LasR inhibitor
• molecular descriptors
• QSAR model
• support vector machine
• virtual screening
• pharmacophore modeling
• molecular docking
• molecular dynamics simulation

Acknowledgement

We wholeheartedly express our gratitude to DST-CURIE-AI Facility, Centre for Machine Learning and Intelligence of Avinashilingam Institute for Home Science and Higher Education for Women for their financial support and continuous guidance rendered during the course of this work. Also, we express our heartfelt gratitude to Mr. Vinod Devaraji, Senior Scientist, Schrödinger team, for his timely help rendered for molecular dynamics simulation studies.

Disclosure statement

No potential conflict of interest was reported by the authors.

Data availability statement

The authors confirm that the data supporting the finding of this study are available within the article [and/or] its supplementary materials.

Additional information

Funding

The author(s) reported there is no funding associated with the work featured in this article.