Baumketner A. New aspects of protein folding and aggregation: Insights from theory and computer simulations

Proteins are one of the three main types of biological molecules. Following the decoding of the human genome, the problem of how proteins fold and interact with one another has gained a sense of urgency. Theoretical approaches play a key role in the studies of proteins, as they offer a number of unique advantages over the alternative methods. In this dissertation, we present the results of a number of our theoretical projects aimed specifically at the problems of protein folding and aggregation. The specific tasks were chosen using the following criteria. First, we asked that certain experimental information be available for the problem under study, sufficient to lead to the formulation of physical hypotheses. Second, experimental approaches to the chosen problem had to meet significant challenges. These could be, for instance, difficulties stabilizing the object of the study long enough to perform measurements; insufficient resolution of the method; or ambiguity in the interpretation of the results. We use a wide range of theoretical models in our study that vary greatly in their resolution depending on the specific problem. While one interacting particle is used to represent an entire protein in the studies of collective properties of proteins in aqueous solutions, more advanced models are employed to study protein folding and aggregation. Our main results can be summarized as follows. A non-spherical model is introduced for lysozyme that describes well both the phase diagram and the structural functions of this protein seen experimentally. It is shown that systems interacting via a globally repulsive potential with a local minimum at short distances form equilibrium clusters. A detailed thermodynamic analysis of the clusters shows that a) at small densities, clusters are stabilized by entropy, and b) at high densities, by enthalpy. Folding of proteins mediated by molecular chaperonin GroEL/ES is studied by computer simulations. A repulsive and an attractive chaperonin cavity are considered. The effect of the repulsive cavity is found to be strongly influenced by the level of frustration in the protein folding free energy landscape. A new model for simulations of protein aggregation is introduced. The model uses all-atom architecture but treats electrostatic and non-polar interactions in approximate ways. For alanine polypeptides, the tendency toward aggregation with the growing size of the peptide chain is correctly reproduced. Effect of external electric field on folding and aggregation is studied. Recovery stroke in myosin motor protein is studied by computer simulations. A reduced model is introduced that is able to reproduce correctly this structural transition. The effect of E22Q mutation on the adsorption of Alzheimer’s disease peptide Aβ on amyloid fibrils is studied in computer simulations within all-atom approaches. It is shown that this effect is entropic and leads to accelerated reaction. A microscopic model is built for fibrils made by 11-25 fragment of Aβ. The registry of β-sheets in the model is controlled by two C- and N-terminal groups and, in agreement with experiment, changes when pH is varied.