Unbiased compound-protein interface mapping and prediction of chemoresistance loci through forward genetics in haploid stem cells

Forward genetic screens in haploid mammalian cells have recently emerged as powerful tools for the discovery and investigation of recessive traits. Use of the haploid system provides unique genetic tractability and resolution. Upon positive selection, these screens typically employ analysis of loss-of-function (LOF) alleles and are thus limited to non-essential genes. Many relevant compounds, including anti-cancer therapeutics, however, target essential genes, precluding positive selection of LOF alleles. Here, we asked whether the use of random and saturating chemical mutagenesis might enable screens that identify essential biological targets of toxic compounds. We compare and contrast chemical mutagenesis with insertional mutagenesis. Selecting mutagenized cells with thapsigargin, an inhibitor of the essential Ca(2+) pump SERCA2, insertional mutagenesis retrieved cell clones overexpressing SERCA2. With chemical mutagenesis, we identify six single amino acid substitutions in the known SERCA2-thapsigargin binding interface that confer drug resistance. In a second screen, we used the anti-cancer drug MG132/bortezomib (Velcade), which inhibits proteasome activity. Using chemical mutagenesis, we found 7 point mutations in the essential subunit Psmb5 that map to the bortezomib binding surface. Importantly, 4 of these had previously been identified in human tumors with acquired bortezomib resistance. Insertional mutagenesis did not identify Psmb5 in this screen, demonstrating the unique ability of chemical mutagenesis to identify relevant point mutations in essential genes. Thus, chemical mutagenesis in haploid embryonic stem cells can define the interaction of toxic small molecules with essential proteins at amino acid resolution, fully mapping small molecule-protein binding interfaces.