Protein folding thermodynamics and dynamics: where physics, chemistry, and biology meet.

Abstract

As was noted in our recent review 1 the protein folding field underwent a cyclic development. Initially protein folding was viewed as a strictly experimental field belonging to realm of biochemistry where each protein is viewed as a unique system that requires its own detailed characterization – akin to any mechanism in biology. The theoretical thinking at this stage of development of the field was dominated by the quest to solve so-called “Levinthal paradox” that posits that a protein could not find its native conformation by exhaustive random search. Introduction, in the early nineties, of simplified models to the protein folding field and their success in explaining several key aspects of protein folding, such as two-state folding of many proteins, the nucleation mechanism and its relation to native state topology, have pretty much shifted thinking towards views inspired by physics. The “physics”-centered approach focuses on statistical mechanical aspect of the folding problem by emphasizing universality of folding scenarios over the uniqueness of folding pathways for each protein. Its main achievement is a solution of the protein folding problem in principle, i.e. demonstration how proteins could fold. As a result, a “psychological” solution of the Levinthal paradox was found (i.e. it was generally understood that this is not a paradox. after all). The key success of this stage of the field is discovery of the general requirements for polypeptide sequences to be cooperatively foldable stable proteins and realization that such requirements can be achieved by sequence selection. That put the field strongly into the realm of biology (“Nothing in Biology makes sense except in the light of Evolution” (Theodosius Dobzhansky)) The physics-based fundamental approach to protein folding dominated theoretical thinking in the last decade (reviewed in 1-4) and its successes brought theory and experiment closer together At the present stage we seek better understanding of how protein folding problem is actually solved in Nature. In this sense the protein folding field has made a full circle as attention is again focused on specific proteins and details of their folding mechanism. However these questions are asked at a new level of sophistication of both theory and experiment. Understanding of general principles of folding and vastly improved computer power makes it possible to develop tractable models that sometimes achieve atomic level of accuracy. Further, better general understanding of requirement for polypeptide sequences to fold, lead to establishment of direct links between protein folding and evolution of their sequences This development opened an opportunity to employ powerful methods of bioinformatics to test predictions of various folding models, in addition to more traditional tests of models against experiment After all, evolution presents a giant natural laboratory where sequences are designed to fold and function and availability of vast amounts of data certainly calls for its use to better understand folding of proteins at very high resolution. At the same time in vitro experimental approaches progressed to the point that very accurate time- and structure- resolved data are available. A close interaction with experimentalists helps to keep theorists honest by providing detailed tests of theories and simulation results. In this review, which to a great extent reflects the thinking of the author on the subject, we will first summarize basic questions and present simple, coarse-grained models that provide a basis for a fundamental understanding of protein folding thermodynamics and kinetics. Then we will discuss more recent developments (over last five years) that focus on detailed studies of folding mechanisms of specific proteins, and finally we will briefly discuss some outstanding questions and future directions.