As I previously mentioned, I am interested in discovery of a new allosteric inhibitor of the Methionine tRNA synthetase of the tropical parasite Leishmania major. In order to achieve this goal, I have performed a virtual screen of the Leishmania major Methionyl tRNA synthetase protein (PDB ID 6SWX) at an allosteric site crystallized with an inhibitor. I began by removing the cognate ligand and redocking the co-crystallized ligand using UCSF DOCK 6.6. As expected, the X-ray conformation of the ligand in the binding site was correctly predicted. The compound scored -49.5 kJ per mol within the binding site. Next, I began to search for new molecules to score within the binding site. I chose a library of neutrally charged, anodyne molecules with molecular weights between 325 and 400. By limiting molecular weight, molecular weight bias will also be curtailed. In the case of allosteric inhibitors, my intuition is that molecular weight bias can be particularly misleading. An allosteric inhibitor is determined more by the quality of contacts rather than the raw number of interacts.
After scoring more than 7.7 million compounds, several interesting in silico hits emerged. Ultimately, the ZINC tranche which I scored in the binding site did not contain the methyl-pyrazolo-pyrimidine motif, or really any similar to it among the top 500 hits. This may be an artifact of the tranche I chose, but in either case it’s probably a good thing since it will lead to more diverse scaffolds which can bind to the key protein of the parasite. The pyrimidine of the cognate ligand, which usually hydrogen bonds with the backbone of the methionyl tRNA synthetase, is replaced with isoxadiazoles, oxdiazole, traizoles, diazines, and several other interesting heteroaromatic rings amongst the hit compounds. This makes sense since the authors of the X-ray structure discuss their hit discovery as limited to adenosine analogs, an interesting but limited section of molecular space. With the more diverse chemical scaffolds employed during in silico screening, more diverse interaction profiles are possible. Several new hydrogen bonds were observed among the hit compounds, an encouraging sign. If you’re interested in learning more about this work, especially for the purposes of testing these compounds, send me an email. Until next time, cheers.
P.S. I’d like to give a huge thanks to the NSF and NYSTAR for funding the Seawulf supercomputing cluster used for this work (https://www.nsf.gov/awardsearch/showAward?AWD_ID=1531492&HistoricalAwards=false)
