Unravelling protein stability and aggregation

Proteins are complex macromolecules that are fundamental to all living species. The stability of proteins are governed by non-covalent forces within the protein and between the protein and its surrounding environment. At high concentrations proteins are susceptible to self-association, which may lead to amyloid fibril formation – a process linked to many diseases, such as Alzheimer’s disease, Parkinson’s disease and type II diabetes. Fibril formation is believed to be a generic property of proteins and the formation of amyloid fibrils occur at protein concentrations exceeding the critical aggregation concentration.The objective of this thesis was to study protein stability and aggregation and the noncovalent forces governing these processes. We showed that the EF-hand protein secretagogin is highly stable at physiological conditions, and that the protein structure is further stabilized by Ca2+ binding. Secretagogin contains intra- and intermolecular disulfide bonds and the protein forms dimers and oligomers in a redox-dependent manner. Interestingly, the disulfide bond does not contribute to the stability implying that the disulfide bond is not well accommodated in the most stable structure of the protein. In order to derive stabilized protein variants we applied an in vivo screening method based on the thermodynamic linkage between fragment complementation and protein stability. This method relies on the split GFP system and the thermodynamic linkage between protein fragment reconstitution and the stability of the intact protein. Using this method we derived a stabilized variant of the sweet tasting protein monellin, with retained sweetness, from a random library. Furthermore, we investigated the influence of sequence and solution conditions on the aggregation behaviour of the amyloid β42 peptide (Aβ42), involved in Alzheimer’s disease. In vivo, Aβ peptides of many different lengths are found and we showed that the aggregation is accelerated for an N-terminally truncated peptide, Aβ5-42, compared to the full length Aβ42 peptide. While the elongation rate is very similar between the two species the nucleation steps are significantly accelerated for the truncated peptide, which is due to the lower net charge and the shorter length of the nonamyloidogenic part of the truncated peptide. Thus, we can establish that the N-terminus of the Aβ42 peptide is important in governing primarily the secondary nucleation rate, even though the N-terminus is not part of the fibril forming core. We addressed the influence of the solution conditions on the aggregation mechanism of Aβ42 by adding denaturants to the aggregation process. Addition of non-ionic urea decrease the aggregation rate of Aβ42, while addition of ionic GuHCl accelerates the aggregation at low concentrations and decreases the rate at high concentrations. Finally, we applied our knowledge of the aggregation mechanism of Aβ42 to derive single chain antibody fragments that specifically inhibit the surface catalyzed nucleation step, with the aim to reduce the amount of toxic oligomeric species