Herein, we report a hit-finding strategy focused on identifying novel allosteric inhibitors of RIPK1 using a combination of virtual screening, structure-based design, and fragment-based screening. Our virtual screening effort was focused on structure-guided use of pharmacophore and shape-based screening, while structure-based design emphasized the replacement of metabolically labile groups and modifications to the core scaffold. Fragment screening was carried out using a reporter displacement assay followed by computational assessment of hits to remove those with high likelihood of binding to the kinase hinge. X-ray crystallography was used to generate experimental structures of several diverse scaffolds from our screening efforts bound to RIPK1, verifying that our hits were indeed allosteric inhibitors.

Optimization of inhibitors was carried out via structure-guided design using X-ray structures of our hit compounds bound to RIPK1, which enabled hybridization of diverse inhibitor scaffolds and opportunities for property improvement. Our efforts led to the identification of highly potent, exquisitely selective, orally bioavailable inhibitors with excellent pharmacokinetics and in vivo activity in models of neuroinflammation.

Due to limited resolution (usually peptide-level), HDX-MS is generally applied to make qualitative identification of protein binding sites and assess protein conformational dynamics in solution. The limited availability of tools to analyze, deconvolute and interpret HDX-MS data is a fundamental gap limiting the efficient application of HDX-MS. Despite these challenges, HDX-MS has great potential to be combined with other experimental and computational techniques to give feasible quantitative assessment of protein structure and dynamics.

In a collaboration between Janssen and Chemical Computing Group (CCG), we have developed a workflow to detect protein binding sites and determine more reliable protein-ligand binding conformations using HDX-MS. We will present an inhouse project involving protein-protein and protein-peptide interactions as an example to demonstrate the process. Starting from deconvoluting HDX-MS data to residue-level parameters for identifying and visualizing key residues in the binding sies, we then applied protein-protein docking with HDX-MS constraints to predict protein-peptide-protein ternary complex conformations. Finally, we demonstrated the use of MD simulations together with HDX modeling to identify the correct binding pose. A comparison of the HDX-MS determined model to the X-ray structure will be discussed.

In the last year, MT has been fully integrated with MOE and it can be executed locally, through a webservice, or even on-demand via a cloud-based backend. With this integration, MOE users can utilize standard MDB files generated through MOE's simulation and docking infrastructure to make MT available within routine drug discovery campaigns.

Software Available:

• http://www.quantumbioinc.com/s
• https://github.com/quantumbio
• https://svl.chemcomp.com

Literature:

1. Zheng, Z., M.N. Ucisik, and K.M. Merz, Jr., The Movable Type Method Applied to Protein–Ligand Binding. Journal of Chemical Theory and Computation, 2013. 9(12): p. 5526–5538. https://doi.org/10.1021/ct4005992
2. Zheng, Z., O.Y. Borbulevych, H. Liu, J. Deng, R.I. Martin, and L.M. Westerhoff, MovableType Software for Fast Free Energy-Based Virtual Screening: Protocol Development, Deployment, Validation, and Assessment. J Chem Inf Model, 2020. 60(11): p. 5437-5456. https://doi.org/10.1021/acs.jcim.0c00618
3. Westerhoff, L.M. and Z. Zheng, Fast, Routine Free Energy of Binding Estimation Using MovableType, in Free Energy Methods in Drug Discovery: Current State and Future Directions, K.A. Armacost and D.C. Thompson, Editors. 2021, American Chemical Society. p. 247-265. https://pubs.acs.org/doi/book/10.1021/bk-2021-1397
4. Liu, W., Z. Liu, H. Liu, L.M. Westerhoff, and Z. Zheng, Free Energy Calculations Using the Movable Type Method with Molecular Dynamics Driven Protein-Ligand Sampling. J Chem Inf Model, 2022. 62(22): p. 5645-5665. https://doi.org/10.1021/acs.jcim.2c00278