Štular Tanja, Lešnik Samo, Rožman Kaja, Schink Julia, Zdouc Mitja, Ghysels An, Liu Feng, Aldrich Courtney C, Haupt V Joachim, Salentin Sebastian, Daminelli Simone, Schroeder Michael, Langer Thierry, Gobec Stanislav, Janežič Dušanka, Konc Janez
National Institute of Chemistry , Hajdrihova 19, SI-1000 Ljubljana, Slovenia.
Faculty of Pharmacy, University of Ljubljana , Aškerčeva cesta 7, SI-1000 Ljubljana, Slovenia.
J Med Chem. 2016 Dec 22;59(24):11069-11078. doi: 10.1021/acs.jmedchem.6b01277. Epub 2016 Dec 12.
Drug discovery is usually focused on a single protein target; in this process, existing compounds that bind to related proteins are often ignored. We describe ProBiS plugin, extension of our earlier ProBiS-ligands approach, which for a given protein structure allows prediction of its binding sites and, for each binding site, the ligands from similar binding sites in the Protein Data Bank. We developed a new database of precalculated binding site comparisons of about 290000 proteins to allow fast prediction of binding sites in existing proteins. The plugin enables advanced viewing of predicted binding sites, ligands' poses, and their interactions in three-dimensional graphics. Using the InhA query protein, an enoyl reductase enzyme in the Mycobacterium tuberculosis fatty acid biosynthesis pathway, we predicted its possible ligands and assessed their inhibitory activity experimentally. This resulted in three previously unrecognized inhibitors with novel scaffolds, demonstrating the plugin's utility in the early drug discovery process.
药物研发通常聚焦于单一蛋白质靶点；在此过程中，与相关蛋白质结合的现有化合物常常被忽视。我们描述了ProBiS插件，它是我们早期的ProBiS-配体方法的扩展，对于给定的蛋白质结构，该插件能够预测其结合位点，并且对于每个结合位点，能预测来自蛋白质数据库中相似结合位点的配体。我们开发了一个新的数据库，其中预先计算了约290000种蛋白质的结合位点比较结果，以便快速预测现有蛋白质中的结合位点。该插件能够在三维图形中对预测的结合位点、配体的姿态及其相互作用进行高级查看。使用结核分枝杆菌脂肪酸生物合成途径中的烯酰还原酶InhA作为查询蛋白，我们预测了其可能的配体，并通过实验评估了它们的抑制活性。这产生了三种以前未被识别的具有新型骨架的抑制剂，证明了该插件在早期药物研发过程中的实用性。
