Accelrys discovery studio visualizer 4

The protein sequence of human NDUFAF5 was retrieved from UniProt and modeled using I-TASSER, due to the unavailability of a template with sufficient homology [32 (link)]. The structural quality of the model was assessed by MolProbity for the Ramachandran Plot [33 (link)] and ProSA [34 (link)]. A molecular dynamic simulation for the NDUFAF5 model was carried out with water as the solvent for 10,000 picoseconds using GROMACS under default parameters [35 (link)]. The PDB protein and molecules were prepared by adding hydrogen and were converted to PDBQT format by Open Babel [36 (link)]. A grid box with dimensions of 40 × 40 × 40 Å (41.89, 37.08, 34.245) with a spacing of 0.375 Å was constructed around the docking area using Autodock 4.2 software (Scripps Research, San Diego, CA, USA) [37 (link)]. Molecules were docked using Vina with exhaustiveness grade 8, with up to nine poses saved per molecule. The docking procedure was carried out for the unchanged conformation of the receptor and flexible ligand molecules. The lowest-energy conformations were selected, and the interactions between the ligand and NDUFAF5 were analyzed. Accelrys Discovery Studio Visualizer 4.0 (Accelrys, San Diego, CA, USA) was used for interaction visualization.

The MD trajectories and structures of each system were interpreted using the utilities of GROMACS software and visualized structural representation by using Visual Molecular Dynamics (VMD) software^{44 (link)}, Accelrys Discovery Studio Visualizer 4.0 (Accelrys Software Inc.) and LigPlot^{+45 (link)} version 2.2.5. SigmaPlot 12.5 (Systat Software, San Jose, CA) was used to generate all plots of the various parameters.

AutoDock Vina (Scripps Research Institute, San Diego, CA, USA) was used to screen 135,755 compounds in the database by docking [20 (link)]. For the docking studies, the X-ray crystallographic structure of the DprE1 complex with Ty38c (PDB Code: 4P8K) from the Protein Data Bank (PDB) was used. The PDB protein and molecules were prepared by adding hydrogen and missing residues in Sybyl-X 2.1 and converted to pdbqt format by Openbabel [21 (link)]. A grid box with dimensions of 30 × 30 × 30 Å (38.972, 12.580, 10.750) with a spacing of 0.375 Å was constructed around the docking area using Autogrid 4.2 software [22 ]. Molecules were docked using Vina with exhaustiveness grade 8, with up to nine poses saved per molecule. The docking procedure was carried out for the unchanged conformation of the receptor and flexible ligand molecules. Ty38c was redocked in the 4P8K model to validate the docking algorithms of AutoDock Vina (the root-mean-square deviation value is 0.844). The lowest energy conformations were selected and the ligand interactions with DprE1 were determined. Accelrys Discovery Studio Visualizer 4.0 (Accelrys, San Diego, CA, USA) was used for interaction visualization. Through inspection of the top docking poses, 30 compounds were selected from the top 200 compounds based on the binding affinity (Table S1).