Effects of substitutions and indels on protein function

Mutations in protein-coding genes can lead to loss of function and disease. However, distinguishing between benign and pathogenic mutations, and understanding the underlying mechanisms of variant effects, remain significant challenges. This hinders accurate diagnosis and treatment of diseases and restricts our comprehension of protein function. This thesis examines the impact of missense mutations and insertions and deletions (indels) on specific aspects of protein function. The first study focuses on the DNA mismatch repair protein MLH1, which forms a heterodimer complex with PMS2. Loss-of-function mutations in MLH1 lead to Lynch syndrome, predisposing individuals to various cancers. We utilized multiplexed assays in yeast to assess the effects of all missense variants in the C-terminal domain of MLH1 on protein stability and interaction with PMS2. We found that reduced cellular protein abundance was a major mechanism of loss-of-interaction variants. Additionally, we identified a previously unrecognized region in MLH1 that conferred both improved and decreased interaction with PMS2. Our data correlated with computational predictions and clinical observations, providing mechanistic insights into MLH1 variant effects. In the second study, we used a yeast-based protein folding sensor termed CPOP to probe the folding stability of all single-residue insertions and deletions of human DHFR. Our findings showed that indels were best tolerated in loop regions and near the protein termini. Several indel variants exhibited folding defects at elevated temperatures, and unstable indels could often be rescued using a small molecule DHFR inhibitor. AlphaFold2 and Rosetta predicted the effects of some indels, suggesting that these computational tools could assist in interpreting clinically observed indel variants and in applications of protein engineering. In conclusion, this thesis characterizes the effects of fundamentally distinct mutation types on specific aspects important for protein function. The results provide novel insights into the mechanisms of variant effects, enhancing our understanding of protein function and potentially paving the way for developing personalized medicine.