Abstract
The emergence of fusion proteins that express the ROS1 kinase domain has become a promising target in non-small-cell lung cancer (NSCLC). Although earlier kinase inhibitors effectively managed ROS1-positive tumors, the rise of point mutations, particularly those beyond the binding pocket, has challenged the inhibitor efficacy. Notably, mutations at residue F2004, which cause cysteine or valine substitution, exhibit intriguing response profiles to the inhibitors. These mutations respond to small molecules that target the active conformation of the kinase (type I) but resist inhibitors that explore the inactive conformation (type II). Our study generates a ROS1 kinase model and uses molecular dynamics simulations to discern structural differentiators of the inactive conformation. A hydrophobic cluster within the active site, involving DFG residue F2103, demarcates the active conformation. We unveil insights from F2004C/V mutations in the ROS1 kinase domain from both the active and inactive states. Notably, the mutations do not perturb the active conformation, resembling wild-type (WT) ROS1. However, in the inactive conformation, the mutations disrupt the flexibility of DFG residue F2103, stabilizing the hydrophobic cluster. Our results provide a model for the inactive conformation of the elusive ROS1 kinase domain, offering pivotal insights into potential differences from the active conformation. Furthermore, our study of F2004C/V mutants proposes a plausible mechanism underlying the type I or II inhibitor response.