Abstract
Protein misfolding and aggregation are often causative to the onset of numerous neurodegenerative diseases such as Alzheimer's disease, Parkinson’s disease, Huntington’s disease, and Amyotrophic lateral sclerosis (ALS). Further, intracellular aggregate formation and misfolding of biopharmaceuticals in mammalian cells are often obstacles to achieving high-yield production of biopharmaceuticals. Thus, understanding protein aggregation mechanism and preventing protein aggregation is imperative from an industrial and clinical perspective.
This dissertation details two novel semi-rational protein engineering approaches to prevent protein misfolding and aggregation in mammalian cells. In particular, this work investigated the aggregation of human copper-zinc superoxide dismutase (SOD1) as a model protein. Many reports indicate that twenty percent of the familial form of amyotrophic lateral sclerosis (fALS) is caused by misfolding and aggregation of mutant SOD1. In particular, this work concentrates on two particular fALS SOD1 mutants, SOD1A4V and SOD1G93A. The most common fALS-linked mutant SOD1 in the US is one containing alanine to valine mutation at 4th residue (A4V). We hypothesized that intracellular protein aggregation can be suppressed by increasing conformational stability of the protein. Using the computational software RosettaDesign, Phe20 was chosen in SOD1A4V as a key residue responsible for SOD1A4V conformational destabilization. This information was used to rationally develop a pool of candidate mutations at the Phe20 site. In order to evaluate how these mutations affect the relative aggregation propensity of the target protein inside cells, we developed a mammalian cell-based screening assay that directly correlates aggregation/misfolding propensity to mean cellular fluorescence. We performed two rounds of screening to select the most promising variants for further characterization. Three novel SOD1A4V variants, SOD1A4V/F20G, SOD1A4V/F20A and SOD1A4V/F20A/C111S were determined to have a significantly reduced aggregation propensity inside cells, demonstrating that the semi-rational protein engineering strategy to increase conformational stability by an additional mutation(s) can be used to reduce the aggregation propensity inside mammalian cells
The aforementioned strategy is beneficial when the destabilizing factor can be pinpointed easily. However, this approach has the limited utility when no rational design approach to enhance conformational stability is readily available. We hypothesized that the aggregation of a target protein within mammalian cells can be reduced using a protein engineering strategy to increase kinetic stability The aggregation-prone mutant SOD1G93A was used as a model system. Unlike the repulsive interaction between two side chains caused by the A4V mutation in SOD1A4V, the G93A mutation causes the repulsive interaction with the neighboring protein backbone resulting in the backbone shift and multiple side chain displacements and eventually the SOD1G93A aggregation. A panel of SOD1G93A variants containing a mutation at three residues (K36, E40, and K91) of which side chain were significantly displaced upon the G93A mutation were generated and subjected to mammalian cell-based screening. Three novel SOD1G93A variants (SOD1G93A/K36E, SOD1G93A/K91D and SOD1G93A/K91E) with a reduced aggregation propensity inside mammalian cells were identified. Although the conformational stability of these variants is comparable to that of SOD1G93A, the aggregation rate of these variants is substantially lower than that of SOD1G93A supporting that the three variants are kinetically more stable than SOD1G93A. From these studies, we have concluded that the semi-rational protein engineering strategies consisting of computational stability analysis, protein structure inspection, and cell-based screening of protein variant has tremendous utility for understanding and modulating protein aggregation in mammalian cells.
