A computational workflow for predicting protein oligomer complex structure induced by small-molecule PPI modulators: validation with BMS-1166 induced PD-L1 Homodimers

Protein-protein interactions (PPIs) are pivotal for human signaling pathways, and their aberrations drive diseases such as cancer and autoimmune disorders. PPI modulators are promising therapeutic agents designed to regulate aberrant PPIs by altering or covering the functional domain of target proteins. Although significant progress has been achieved in machine learning-based predictions of binary protein complexes, predicting protein oligomer complexes induced by small-molecule PPI modulators remains challenging due to the increased complexity of multicomponent interactions. In this study, we developed a computational workflow utilizing molecular dynamics (MD) simulations to explore the formation of protein oligomer complexes induced by small-molecule PPI modulators. This workflow integrates extensive MD sampling with a multistep screening strategy to identify candidate conformations of the complexes. We validated the workflow by investigating the dimerization of programmed cell death 1 ligand 1 (PD-L1) induced by the small-molecule PPI modulator BMS-1166. The predicted structures showed strong consistency with experimental crystal structures, confirming the workflow's reliability. Given that PPIs are widely involved in multiple signaling pathways, this adaptable workflow can be applied to various small-molecule PPI modulators, providing a novel method for predicting protein oligomer complexes. Furthermore, it helps elucidate the detailed mechanisms of protein-molecule interactions and offers practical guidance for the rational design of novel PPI modulators.