Applying computational tools to improve stability and reduce aggregate formation of proteins

The goal of this work was to apply two computational tools to address the aggregation problem in multi-domain proteins. The first tool is based on a set of algorithms developed to predict aggregation- or self association-prone sequences of unfolded peptides. This tool was applied to the analysis of aggregates formed from alpha-chymotrypsinogen to frame a hypothesis for the potential pathway to aggregate formation as well as the aggregation-prone region of the molecule.;The second tool, Rosetta Design, is a growing software suite capable of an assortment of protein-related applications. Rosetta Design was first used to analyze a set of known stabilizing and destabilizing mutations for a biopharmaceutical system and then identify an additional stabilizing mutation previously unidentified by other design tools.;Finally, these two tools are applied in tandem to the problem of human gammaD-crystallin aggregate formation. Two mutations were chosen for experimental design, M69Q and S130P. The N-td mutation, M69Q, was chosen due to the large increase in domain stability predicted by Rosetta as well as a notable increase in domain-domain interaction. The C-td mutation, S130P, was chosen primarily because of its location in a consensus aggregation "hot spot" and the tendency for proline to reduce beta-sheet formation. Additionally, the mutation was predicted to increase domain-domain interactions at the expense of a portion of the C-td stability. Preliminary stability experiments for this indicate a slight increase in the stability of the M69Q mutation and a noticeable the loss of stability for S130P. The overriding result though is a remarkable reduction in aggregation rate for both mutants at elevated temperatures approaching the Tm, essentially forming little or no aggregate in a two hour time period with the S130P variant.;These results highlight the strength of computational tools in protein design. By narrowing in silico the sequence space to be tested experimentally from a vast array of candidate mutations, computational design equips researchers a qualitative evaluation of potential mutants and it enables detailed studies that exhaust potential routes to protein stability.