Exploring the Binding Mechanism between Human Profilin (PFN1) and Polyproline-10 through Binding Mode Screening

Leili Zhang; David R. Bell; Binquan Luan; Ruhong Zhou

PPI structure prediction is a very challenging problem computationally due to the complexity associated with the underlying multibody interactions. 11 There are a few promising methods developed already, mostly for protein-protein docking, including ClusPro 12 , HADDOCK 13 , pyDockWeb 14 , and GRAMM-X 15 . These methods have proven successful in certain systems, especially when proteins are rigid and/or small; however, none of them could provide reliable criteria to determine the most trustworthy prediction. Every year the Critical Assessment of PRediction of Interactions (CAPRI) contest 16 is held multiple times to test and improve various methods and protocols. Within the CAPRI framework, the contestants can submit up to 10 structures (with preferences), and the final scores are assessed based on how many hits are predicted successfully. Generally since there is no crystal structure for comparison, in past practices people used multiple models and compared the predictions, or used other empirical scoring functions to narrow down the selections. [17] [18] Unfortunately, conflicting and confusing results from different models are frequently obtained, which complicates further assessments of the PPI structures. To overcome this shortcoming, we explore the idea of applying free energy calculations, conventional molecular dynamics (MD) simulations, and steered molecular All rights reserved. No reuse allowed without permission.