Abstract:
The weighted ensemble (WE) methodology was incorporated into an iterative ANM-MC (Anisotropic network model- Monte Carlo) simulation algorithm, which has been developed to generating targeted conformational transition pathways of proteins. Through the WE sampling, it is possible to produce unbiased pathways by increasing the number of ANM-MC simulations along the reaction coordinate, which is specified as the root-mean-squared distance to target structure. In this modified algorithm, the structures are deformed along randomly chosen (unbiased) ANM modes (among the slowest five modes) and then their energy is minimized by Monte Carlo (MC) method. Three proteins were studied using 18 Å and 10 Å cutoff distances and distinct successful pathways were obtained for a total of 13 systems (simulations). Various sequences of domain motions (LID and AMP domain closures) were observed for adenylate kinase (AdK) systems. Calmodulin (CaM) systems followed quite similar pathways, whereas glutamine binding protein (GBP) displayed high conformational freedom. Strain energies and dihedral angle changes of each pathway for all systems were calculated to detect possible cracking regions where high strain energy causes local unfolding. Strained regions well matched with other studies for AdK with five out of eight previously reported hinge regions having high strain energies. Two sites for AdK and CaM can be reported for cracking and there was no cracking in GBP motion.
