Open Collections

Abstract

The DNA damage response (DDR) network protects cells from genome instability that can induce mutagenesis, cell death, or tumorigenesis. Tumor cells frequently experience replication stress and genomic instability and rely on DDR pathways for survival. This dependence on the DDR is a tumour vulnerability that can be exploited. The DNA repair enzyme MUS81 is non-essential in healthy somatic cells but required for viability in some cancer backgrounds. MUS81 is a structure-selective endonuclease (SSE) that resolves branched DNA intermediates that arise during replication fork restart and homology directed DNA repair. Despite the apparent therapeutic potential of MUS81, attempts to find a drug-like small molecule inhibitor have been unsuccessful. Mutational analysis offers a powerful approach for studying proteins. Mutations can model the inhibition of a protein and direct inhibitor discovery programs. Mutations that cause a dominant lethal phenotype in cancer-associated genetic and phenotypic backgrounds would define a mutant proteoform capable of overcoming functional redundancies. Furthermore, dominant mutations could elucidate the biophysical interactions underlying the mutant phenotype to guide drug design and development. Following this rationale, I performed a high throughput screen on two comprehensive variant libraries of MUS81 in the model organism Saccharomyces cerevisiae for mutations that elicited a dominant lethal sensitivity to DNA damaging agents. I identified 12 residues that could be dominantly mutated in the yeast MUS81 ERCC4 domain. They include residues that coordinate Mg²⁺ for catalysis and residues in the hydrophobic wedge structure, the nexus of which is hK335 (yK411). Through genetic experiments, I determined that the phenotype of mus81-K411A depends on two factors: 1) the presence of MMS4, which is required for MUS81 DNA binding, and 2) the abundance of MUS81 DNA substrates, which is suggestive of a DNA trapping mechanism. These data, in collaboration with computational scientists, were used to direct a ligand-based virtual docking screen. From a compound library of over 500,000, 36 were assayed for biochemical inhibition of hMUS81, and 3 were validated as robust and specific MUS81 inhibitors. Overall, these data establish the utility of mutational analysis to guide drug discovery and design.