Present malarial chemotherapies mainly rely
Present malarial chemotherapies mainly rely on a very few chemotypes or their allied structures such as quinolines, anti-folates and some cyclic endo-peroxides (Nayyar et al., 2012). Over the last few decades, the extensive and repetitive deployment of these drugs has stimulated resistance in the pathogen and has seriously compromised the effectiveness of current antimalarial arsenal (Rastelli et al., 2011). This selection pressure on present chemotherapies has underlined the need of testing/screening for diverse scaffolds from Available Chemical Space (ACS)/New Chemical Entities (NCEs) of small molecules.
The above-mentioned screening can be performed by “actual biochemical” or “virtual in-silico” routes. No doubt, the progress in robotics, automation and miniaturization has greatly improved the efficiency of traditional biochemical screening; nevertheless these methods have become more expensive and are still labor intensive (Durrant and McCammon, 2010).
On the other hand, screening based on “virtual platform” such as molecular docking are particularly suitable for most of the academic setup, with limited time and resources (Shoichet, 2004). Moreover, the increasing availability of 3D structure of proteins/macromolecules and advancements in NMR and X-ray, are expected to assist the development of these virtual platforms one step further. If utilized with caution and proper validation, these methods can be a very fruitful and many previous reports have shown the strength of this approach (Shoichet, 2004, Walter et al., 1998).
Consistent with this theme, we undertook a multilayer screening of xanthones like natural products (Xanthone-NPs) against Pf-DHFR (Plasmodium falciparum dihydrofolate reductase). Pf-DHFR is a well known and validated target in antimalarial chemotherapy. It specifically catalyzes the conversion of dihydrofolate to tetrahydrofolate, which is a precursor to important nitrogenous bases such as thymidine, hence plays an essential role in taselisib and survival of the pathogen (Toyoda et al., 1997a, Abbat et al., 2015, Oliveira et al., 2011, Bhat et al., 2013, Cunha et al., 2005).
The selection of xanthones and related scaffolds was not arbitrary, but was based on the following observations/facts:
Due to their unmatched complexity, abundant structural diversity and broad medicinal spectrum, xanthone like natural products are important for drug discovery efforts. Unfortunately, from last two decades their utility remained underrepresented in drug discovery, because of difficulty in synthesis and structural modification, lower natural availability, low drug likeness, problems in further optimization etc (Harvey et al., 2015, Molinari, 2009).
Proposed multi-layer screening approach uses a collective knowledge of “in silico”, “statistical” and “experimentally driven” results. The one obvious advantage of this multilayer screening approach is that hits are not the results of a mere computational work, but statistical factors such as molecular weight were also taken into account. Finally the interaction footprints of poses, satisfied by both docking and statistical criteria, were compared with that of well known and experimentally validated antifolates, emphasizing interactions with key residues of the protein. We believed that these hits would have a greater chance of success in actual biochemical screening and would be helpful in designing better candidates for future anti-malarial drug discovery ventures.
Materials and methods Filter 1: Computational Docking (in silico origin)
Results and discussion Our group has previously explored the binding cavity of Pf-DHFR against 185 diverse natural products using AutoDock (Kumar et al., 2014). Interestingly, all the xanthones displayed strong binding interactions with hydrophobic lining of binding cavity. Another very important outcome of this study was the spatial overlapping of all the xanthones in the binding cavity with conserved hydrogen-bonding interactions with Asp194 and Gly44. Motivated by these initial findings, we decided to further investigate the subject using a large and diverse library of xanthones itself. A framework based on size normalized score and pose clustering was generated, which was then applied to screen the ZINC database.