A Dynamical Model for Insulin Degrading Enzyme Conformational Transition between Closed and Open States

Abstract

Insulin Degrading Enzyme (IDE) is involved in the deconstruction of insulin, amylin, and glucagon peptides involved in controlling blood glucose level and has also been found to break down amyloid-β, a peptide involved in Alzheimer's disease. IDE has two homologous N- and C-terminal domains, IDE-N and IDE-C. For IDE to degrade its substrates, both domains must come together for the catalytic cleft to assume closed conformation. New Cryo-Electron Microscopy (Cryo-EM) data suggests IDE passes through five conformations: completely closed with substrate bound, partially closed with substrate bound, partially open with substrate unbound, and open without substrate. Dynamical details of how these conformations connect are unknown. We investigated this process using long all atom molecular dynamics simulations (MD) of IDE monomer and dimer structures with and without substrate. Our simulations showed that IDE without insulin (pdb-code 4IOF) has larger relative motion between four subdomains (D1, D2, D3, D4) in the N- and C-terminal domains compared to an IDE with insulin (pdb-code 2WBY). Analysis of the insulin inside the catalytic chamber suggested that insulin was stabilized by interactions of its residues Gly1, Tyr14, Glu17 form its chain A and Phe1 and Ph25 from its chain B with IDE residues via a combination of hydrogen bonding and π - π interactions. We used, Essential Dynamics Analysis (EDA) to calculate principal normal modes of IDE motion from our trajectories. These normal modes were then applied as coordinate transformations combined with short MD simulations to gradually open the IDE-closed conformation. We used cross-comparison with Cryo-EM and hydrogen-deuterium exchange data of IDE as a guide to obtain a realistic model of how IDE changes between closed and open conformations.