Computational Design of Allosteric Pathways in the CXCR3 Receptor

Abstract

Chemokine receptors, a subfamily of G-protein-coupled receptors (GPCRs), orchestrate immune cell trafficking and are promising targets for enhancing adoptive cell therapy. This thesis explores computational methods to engineer allosteric signaling in the CXCR3 chemokine receptor, aiming to improve its G-protein signaling sensitivity to CXCL10, a chemokine prevalent in tumor microenvironments. Integrating molecular dynamics simulations with mutual information analysis, we identified ligand-specific allosteric sites that mediate CXCR3 activation. We reveal that the dynamic binding pose of CXCL10 engages multiple allosteric communication pathways, propagating through transmembrane helices 2, 4, 6, and 7, to influence G-protein coupling. In-silico mutagenesis at allosteric sites and elastic network modeling predicted mutations likely to enhance the ligand-effector allosteric coupling. These predictions were experimentally validated using Bioluminescence Resonance Energy Transfer (BRET) assays to quantify signaling through Gi-protein and β-Arrestin2 pathways. While no mutations increased signal transduction through Gi proteins, one variant emerged as a biased receptor. It exhibited a 26.5% increase in the β-Arrestin2 maximal response and a 19.9% decrease in maximal G-protein signaling compared to wild-type CXCR3 upon CXCL10 stimulation. These findings suggest that the identified transmembrane allosteric sites are sensitive to perturbation and can influence diverse signaling outcomes.

Overall, the integration of computational modeling and experimental assays presented in this work highlights the feasibility of strategically designing receptor variants to tune signaling output. With further research, this approach may lead to more effective strategies in cancer immunotherapy, addressing current limitations in CAR-T cell therapies targeting solid tumors.