Evaluating Imatinib's Affinities and Specificities for Tyrosine Kinases Using Molecular Dynamics Simulations

Computational chemistry lets us model intermolecular interactions in ways assays cannot. My project focuses on the multi-kinase interactions of the cancer drug, imatinib. Most cancer drugs target one kinase, but some affect multiple kinases. Imatinib treats chronic myeloid leukemia by targeting ABL kinase. Proteomics data reveals it can interact with other kinases, such as KIT to treat gastrointestinal stromal tumors, but the mechanisms are unknown. Imatinib has different affinities for similar kinases, such as a 3000x difference between ABL and SRC, despite sharing 50% structural homology. Here, I investigate the conformational differences between free and imatinib-bound ABL, KIT, and SRC using Molecular Dynamics simulations to understand the key imatinib-kinase interactions. The alignment analysis shows the docked conformations are similar to co-crystal structures in the Protein Data Bank. Root-mean-square-deviation and fluctuation (RMSD and RMSF) analysis show that all simulations converge at 45 ns, with some regions exhibiting differential flexibility. Hydrogen bond analysis across 100 ns simulations show that ABL has one main H-bond, KIT has three main H-bonds, and SRC has no main H-bonds. All the drug-kinase complexes feature at least 15 key salt bridge interactions relevant for structural stability. The dihedral distributions reveal that most residues adopt a single conformation, but some can adopt multiple, increasing the protein flexibility. The entropy results quantify the protein disorder, revealing KIT and SRC favors the apoprotein while ABL favors the complex. This signifies that broad protein similarity does not govern imatinib binding, instead, it is explained by smaller structural details