Influence of Go-like interactions on global shapes of energy landscapes in beta-barrel forming model proteins: inherent structure analysis and statistical temperature molecular dynamics simulation.

Recently, the native-centric, coarse-grained "Go-like" protein model has gained a renewed popularity because of its computational simplicity and successful description of some key aspects of folding thermodynamics and kinetics. In the present paper, the properties of Go-like models are investigated in terms of the potential energy landscape (PEL). The non-native attractions of the beta-barrel forming BLN model 46-mer are scaled with a parameter 0 < or = lambda < or = 1, to make a continuous tuning of the PEL from multi-funneled and energetically frustrated at lambda = 1 to a perfect funnel including only topological frustration at lambda = 0. The combination of the enhanced sampling of "statistical temperature molecular dynamics", STMD (Kim, J.; Straub, J. E.; Keyes, T. Phys. Rev. Lett. 2006, 97, 050601), and extensive inherent structure (IS) analysis, clearly demonstrates the evolution of the topography of the PEL. It is shown that a small amount of Go-like interactions brings about a significant modification in the occupation of low-lying IS by destabilizing the misfolding funnel into a higher energy region, strongly influencing the folding thermodynamics. The alteration of the PEL also induces a dramatic change in the folding mechanism, from a second-order-like collapse transition into a cooperative, first-order-like folding transition, occurring through a transient, intermediate state ensemble characterized by partially structured IS. The appearance of multiple van der Waals loops in the statistical temperature of the Go-like model is associated with the development of the intermediate states.