Abstract
Background: Human African trypanosomiasis is a fatal disease prevalent in approximately 36 sub-Saharan countries. Emerging reports of drug resistance in Trypanosoma brucei are a serious cause of concern as only limited drugs are available for the treatment of the disease. Pteridine reductase is an enzyme of Trypanosoma brucei.
Methods: It plays a critical role in the pterin metabolic pathway that is absolutely essential for its survival in the human host. The success of finding a potent inhibitor in structure-based drug design lies within the ability of computational tools to efficiently and accurately dock a ligand into the binding cavity of the target protein. Here we report the computational characterization of Trypanosoma brucei pteridine reductase (Tb-PR) active-site using twenty-four high-resolution co-crystal structures with various drugs. Structurally, the Tb-PR active site can be grouped in two clusters; one with high Root Mean Square Deviation (RMSD) of atomic positions and another with low RMSD of atomic positions. These clusters provide fresh insight for rational drug design against Tb-PR. Henceforth, the effect of several factors on docking accuracy, including ligand and protein flexibility were analyzed using Fred.
Results: The online server was used to analyze the side chain flexibility and four proteins were selected on the basis of results. The proteins were subjected to small-scale virtual screening using 85 compounds, and statistics were calculated using Bedroc and roc curves. The enrichment factor was also calculated for the proteins and scoring functions. The best scoring function was used to understand the ligand protein interactions with top common compounds of four proteins. In addition, we made a 3D structural comparison between the active site of Tb-PR and Leishmania major pteridine reductase (Lm-PR). We described key structural differences between Tb-PR and Lm-PR that can be exploited for rational drug design against these two human parasites.
Conclusion: The results indicated that relying just on re-docking and cross-docking experiments for virtual screening of libraries isn't enough and results might be misleading. Hence it has been suggested that small scale virtual screening should be performed prior to large scale screening.