Evolutionary-docking targeting bacterial FtsZ

Abstract

To accurately predict binding of inhibitors to the FtsZ cell division protein of the antibiotic-resistance Staphilococcus aureus pathogen, evolutionary library docking, ligand-efficiency predictions and rank consensus docking strategies have been sequentially applied. Starting from the crystallographic FtsZ bound model of the PC190723 reference ligand, fragments were derived to generate children molecules fitting low docking-scores with low molecular sizes and hydrophobicities using the DataWarrior Build Evolutionary Library. PC190723 fragment children molecules combined with toxicity filters, and consensus ranks with ligand efficiencies and AutoDockVina docking, identified new benzamide and non-benzamide chemotypes with nanomolar docking-scores and improved specificities to continue with anti-FtsZ ligand investigations

New star-shaped ligands generated by evolutionary fitting the Omicron spike inner-cavity

“…The "Build evolutionary library" subprogram included into the DataWarrior (DW) written in Java (dw550win.zip for Windows) was locally downloaded (https://openmolecules.org/datawarrior/download.html) following the details described by DW and our previous work 53 .…”

“…The best 8 fitting molecules are then selected for additional modifications into the next generation. Generations automatically continued until a fitness plateau was reached 53 . The user-defined fitness criteria and weight values were described in the Figure legends.…”

“…The raw children data were then filtered by excluding predicted high and low toxic DW chemical properties (mutagenesis, tumorigenicity, reproductive interference, irritant, and nasty functions). The toxic-filtered children were saved as both *.dwar and special *.sdf (vs3) files maintaining all evolutionary information including the 3D protein cavity docked to children ligand conformers for more detailed visualization in PyMol (using its split_states command) 53 .…”

“…Other computational programs To identify and predict seeSAR leads used for the initial parent definition, the seeSAR vs.10 package (https://www.biosolveit.de/SeeSAR/) was used as described before 20,21 . The DataWarrior (Osiris, vs5.5.0.Idorsia Ph.Ltd, https://openmolecules.org/datawarrior/download.html) 75 and MolSoft (ICM Molbrowser vs3.9Win64bit, https://www.molsoft.com/download.html) were used for docking and manipulating the *.sdf files, as described before 53…”

“…To derive the final list of top-leads, some particular methods were used for this work. For instance, seeSAR leads from our previous work 20,21 were selected to start our previously described sequential rather than previous independent consensus strategy, similar to the one applied in our most recent work 53 . Comparisons using ranks rather than absolute docking-scores, were used as recommended by many other authors for higher prediction accuracies [54][55][56][57][58] ; 59,60 .…”