2d sketcher

The crystal structure of bovine xanthine oxidase in complex with febuxostat (PDB 1N5X) [41 (link)] was retrieved from the Protein Data Bank (PDB). The amino acid sequence has 90% sequence identity to the human form of the enzyme [50 (link)].
The structure was preprocessed and optimised with the Protein Preparation Wizard in Maestro (Protein Preparation Wizard; Epik, Schrödinger, LLC, New York, NY, USA, 2021; Impact, Schrödinger, LLC, New York, NY, USA; Prime, Schrödinger, LLC, New York, NY, USA, 2021). The hydrogen atoms were added after determining the appropriate bond orders, charges and atom types. Extensive sampling of the rotamers, tautomers and protonation states of titratable amino acids at neutral pH was performed in order to optimise the H-bond network.
Finally, the protein structure was subjected to a restricted minimisation using the Impref module and the OPLS4 force field, with a 0.3 RMSD limit imposed from the original coordinates as a constraint.
Compounds under investigation, namely ALS-1, -8, -15 and -28, were drawn by means of a Maestro 2D-sketcher and prepared with LigPrep (LigPrep, Schrödinger, LLC, New York, NY, USA, 2021) to generate suitable 3D conformations and tautomerisation states at pH 7.0 ± 2.0. The compounds were then energetically minimised using the OPLS4 force field.

Figure 1, Figure 2 and Figure 8 were prepared using Maestro (Schrödinger Release 2022-3; Schrödinger, LLC, New York, NY, USA, 2018). Figure 3 was prepared using ROCKER [15 (link)] (http://www.medchem.fi/rocker). Figure 4 (lower panel) was prepared using PyMOL (version 2.5.0, Schrödinger, LLC). Figure 6 was prepared using GraphPad Prism version 8.4.2 (GraphPad Software, LLC, Boston, MA, USA). The 2D structures in Figure 7 were prepared with 2D Sketcher in Maestro (Schrödinger Release 2018-2; Schrödinger, LLC, New York, NY, USA, USA, 2018).

A 2D molecular model of DRB18 was constructed using an NIH PubChem Sketcher (https://pubchem.ncbi.nlm.nih.gov/edit3/index.html). The 3D model of the compound was constructed by importing the 2D model into the 2D sketcher (Schrödinger). A 4PYP model for hGLUT1 inward open conformation and 5C65 model for hGLUT3 outward open conformation were used to generate homology models for other GLUTs in the respective conformations using SWISS-MODEL [35 (link), 36 (link)]. A protein preparation wizard module (Schrödinger) was used to prepare the GLUTs for docking, and receptor grid preparations were conducted using the Glide module of Maestro (Schrödinger) with default protocols [31 (link), 37 (link), 38 (link)]. Grid box settings including amino acids for the centroid of the grid box have been discussed in detail in supplementary methods. DRB18 was initially docked using Standard Precision mode in Glide. The pose with the lowest Glide score was then used to redock DRB18 using the Induced-fit docking module in Glide. The resultant best docked structure for the compound was selected based on the Glidescore values.