BANΔIT: B'-Factor Analysis for Drug Design and Structural Biology

Abstract

The analysis of B-factor profiles from X-ray protein structures can be utilized for structure-based drug design since protein mobility changes have been associated with the quality of protein-ligand interactions. With the BANΔIT (B'-factor analysis and ΔB' interpretation toolkit), we have developed a JavaScript-based browser application that provides a graphical user interface for the normalization and analysis of B'-factor profiles. To emphasize the usability for rational drug design applications, we have analyzed a selection of crystallographic protein-ligand complexes and have given exemplary conclusions for further drug optimization including the development of a B'-factor-supported pharmacophore model for SARS CoV-2 main protease inhibitors. BANΔIT is available online at https://bandit.uni-mainz.de. The source code can be downloaded from https://github.com/FBarthels/BANDIT.

Keywords: B-factor; Bioinformatics; Drug design; Molecular modeling; Protein flexibility.

Figure 1

User interface of the BANΔIT with numbered instruction steps for pairwise analysis of B’‐factor profiles. (1) Choose 1 or 2 protein structures by either upload or fetch‐by‐id. (2) Select the set of atoms to be considered for normalization. (3) Check the options‐toolbar for advanced algorithm settings. (4) For two structures may align them by one of the alignment methods. (5) Select the normalized data that should be displayed. (6) Analyze the B‘‐factor profile plot. Interesting regions can be zoomed by dragging. (7) Analyze the 3D‐models colored by their B’‐factors in the NGLviewer. (8) Save the B‘‐factor profile plot. (9) Save the normalized PDB‐records. (10) Switch to the list‐viewer interface for multiple structure alignment and heatmap visualization

Figure 2

Presentation of the B′‐factor analysis from representative drug design examples. (A) Plot of ΔB′ for PTP1E apo‐structure (PDB: 3LNX) vs. PTP1E in complex with the RA‐GEF ligand (PDB: 3LNY). ΔB′‐values outside the salmon‐colored horizontal box are statistically significant (p<0.05). (B) Superposed crystal structures for PTP1E with and without the RA‐GEF ligand colored by the B’‐factors. (C) Representation of the most rigidified residues in the RA‐GEF2 PTP1E complex. Turquoise: PTP1E apo‐structure; Yellow: PTP1E holo‐structure; Orange: RA‐GEF2 peptide ligand. (D) Plot of ΔB′ for the Src kinase in complex with the conformationally constrained ligand (PDB: 1IS0) vs. Src in complex with the natural ligand (PDB: 1SPS). (E) Superposed crystal structures for the respective Src complexes with both ligands colored by the B’‐factors. (F) The natural phosphopeptide ligand reveals a favourable cation‐π interaction with K60. (G) The cyclopropyl constraint (arrow) leads to a suboptimal cation‐π interaction, which induces increased mobility in the neighboring residues (K60–L64). (H) B’‐factor analysis of SARS CoV‐2 main protease in complex with multiple fragments. Clustering of 22 non‐covalent complexes vs. the non‐liganded apo‐structure (PDB: 5R8T). (I) Development of a B’‐factor‐based pharmacophore hypothesis. (J) Superposition of selected crystal structures from a crystallographic fragment screening to the SARS CoV‐2 main protease. In the center, aromatic structural elements are predominant. Sulfonamides interact with T190, A191 and P168 (PDB: 5R80, 5R81, 5RF1; Carbons colored in orange). DMSO molecules were also found in this S‐4 pocket (PDB: 5REH, 5R82, 5RE9; Carbons colored in magenta). Various polar fragments interact with T45, S46 (PDB: 5REB, 5R7Y, 5R82, 5RGH; Carbons colored in cyan). Pyridine containing ligands show interactions with H163 and E166 (PDB: 5RE4, 5R83, 5R84; Carbons colored in green).