Monte Carlo approach to binding and aggregation of disordered peptides

Sammanfattning: In this thesis, theoretical models are applied to study some aspects of intrinsically disordered proteins, i.e. proteins that partly or entirely lack a stable native structure. In particular, focus is on peptide binding and aggregation. Knowledge about these mechanisms is important for understanding how proteins interact with each other and how mutations can affect the function and sometimes give rise to disease. The methods used are based on all-atom Monte Carlo simulations with implicit solvent. In Papers I and II a model is developed for analysis of peptide binding. It is tested on PDZ-domains, structural units that are often found in signaling proteins, which bind to the C-termini of other proteins. They especially recognize specific patterns of short amino-acid sequences. Paper III likewise deals with peptide binding, but the system of interest is the S100B protein, which in the same manner is able to bind various sequences. In this case, the two peptides studied are disordered when they are free, but assume helical structures as they become part of the complex. Since the peptides differ in their final bound configurations, the main question is whether there still are any similarities between their binding processes. Papers IV and V are devoted to questions related to peptide aggregation. They describe the analysis of Abeta42, a peptide that forms miscellaneous aggregates, and which is believed to be involved in the development of Alzheimer's disease. Monomers and dimers of the wild type, as well as three mutants, are examined in order to observe the initial interactions and features of peptide aggregation, and comparisons of the four variants are made aiming to identify structural propensities that could explain their different aggregation properties.