Glidescore

Prior to molecular docking calculations, we constructed the IFNAR2 model structure using the X-ray crystallographic structure of IFNAR2 (PDB ID: 3S9D) [25 (link)]. IFN-α2 and water molecules were removed, and the protein structure was prepared using Schrödinger Protein Preparation Wizard of maestro module (Schrödinger, New York, NY, USA) [26 ]. Docking center was set to the loop that included Tyr39 and Tyr43. We used a compound library for docking simulations provided by Namiki Shoji (Tokyo, Japan). The compound library was filtered using Lipinski’s rule of five, resulting in exclusion of compounds with reactive groups and selection of approximately 300,000 compounds at random. Alternative protonation states of each compound as well as chiral forms were generated for the 7 ± 2 pH range using the LigPrep module and ionization penalties were calculated using the Epik panel (Schrödinger, NY, USA) at pH 7 [27 ]. The docking simulations were performed using GOLD with GoldScore (The Cambridge Crystallographic Data Center, Cambridge, UK) [28 (link)], Glide (SP mode) with GlideScore (Schrödinger, NY, USA) [29 ], and Molecular Operating Environment (MOE) Docking score (S score) (Chemical Computing Group, Montreal, Canada) [30 ].

Computer-aided compound docking was performed with GLIDE (GlideScore version SP5.0 Schrödinger, LLC, New York, NY, 2021, re. 2021-2) with standard precision, the Maestro 12.8.117 portal. The receptor grid was generated from the EGFR(T790M/V948R) kinase domain from Chain D of PDB ID 8FV4 (compound 2) with the omitted ligand using the Protein Preparation Wizard^{68 (link)}. Compound 4 was prepared with LigPrep (OPLS4 force field, Epik pH = 7.0 ± 2). The best binding poses were ranked on the basis of the lowest docking and GlideScore values^{69 (link),70 (link)}.

The BDNF dimer model was re-constructed using the template of the X-ray structure of the BDNF/NT3 heterodimer complex (PDB: 1BND). Our approach to docking BDNF pro-peptides on the surface of mature BDNF dimer models utilized two steps that took into account several levels of structural flexibility and scoring criteria: (1) rigid-body docking of BDNF pro-peptide to identify the binding position; and (2) refinement of the BDNF pro-peptide/mature BDNF dimer model complex by re-docking with greater structural flexibility and Glide score (Schrödinger, LLC). To obtain the initial binding position of an extended BDNF pro-peptide on the surface of a mature BDNF dimer model, we used ZDOCK (ver.2.3) 16 with the option of sampling at 6-degree rotational steps. Using fast Fourier transform, ZDOCK searches for all possible binding orientations of a BDNF pro-peptide along the surface of a mature BDNF dimer model, optimizing desolvation, shape complementarity and electrostatics. To account for structural flexibility in the refinement step, the Glide Induced Fit Docking (IFD) protocol17 (Schrödinger, LLC) was utilized, followed by iteratively combining rigid-receptor docking (Glide) and protein remodeling by side chain searching and minimization (Prime) techniques. Finally, the best model was rescored according to the binding energy that was calculated using Glide.