Email updates

Keep up to date with the latest news and content from BMC Pharmacology and BioMed Central.

Open Access Highly Accessed Research article

A web server for predicting inhibitors against bacterial target GlmU protein

Deepak Singla1, Meenakshi Anurag2, Debasis Dash2 and Gajendra PS Raghava1*

Author Affiliations

1 Institute of Microbial Technology, Chandigarh, India

2 G. N. R. Knowledge Centre for Genome Informatics, Institute of Genomics and Integrative Biology (IGIB), New Delhi, India

For all author emails, please log on.

BMC Pharmacology 2011, 11:5  doi:10.1186/1471-2210-11-5

Published: 6 July 2011

Abstract

Background

The emergence of drug resistant tuberculosis poses a serious concern globally and researchers are in rigorous search for new drugs to fight against these dreadful bacteria. Recently, the bacterial GlmU protein, involved in peptidoglycan, lipopolysaccharide and techoic acid synthesis, has been identified as an important drug target. A unique C-terminal disordered tail, essential for survival and the absence of gene in host makes GlmU a suitable target for inhibitor design.

Results

This study describes the models developed for predicting inhibitory activity (IC50) of chemical compounds against GlmU protein using QSAR and docking techniques. These models were trained on 84 diverse compounds (GlmU inhibitors) taken from PubChem BioAssay (AID 1376). These inhibitors were docked in the active site of the C-terminal domain of GlmU protein (2OI6) using the AutoDock. A QSAR model was developed using docking energies as descriptors and achieved maximum correlation of 0.35/0.12 (r/r2) between actual and predicted pIC50. Secondly, QSAR models were developed using molecular descriptors calculated using various software packages and achieved maximum correlation of 0.77/0.60 (r/r2). Finally, hybrid models were developed using various types of descriptors and achieved high correlation of 0.83/0.70 (r/r2) between predicted and actual pIC50. It was observed that some molecular descriptors used in this study had high correlation with pIC50. We screened chemical libraries using models developed in this study and predicted 40 potential GlmU inhibitors. These inhibitors could be used to develop drugs against Mycobacterium tuberculosis.

Conclusion

These results demonstrate that docking energies can be used as descriptors for developing QSAR models. The current work suggests that docking energies based descriptors could be used along with commonly used molecular descriptors for predicting inhibitory activity (IC50) of molecules against GlmU. Based on this study an open source platform, http://crdd.osdd.net/raghava/gdoq webcite, has been developed for predicting inhibitors GlmU.