Figure 2. Identification of high-frequency resistance-conferring mutations.
(a) Hypergeometric distribution-based clustering (red = high; black = low similarity) of 8 clones processed for RNA-Seq. Clones collapsed to 5 independent groups. (b, d) Sanger sequencing traces of DNA for the genomic locus of the kinesin-5 A133P mutation in genome-edited HeLa cells (b) or HCT116 ispinesib-resistant clone I11 (d) after selection in 100nM ispinesib. Arrow indicates site of mutation. (c) LD50 values measured for parental HeLa (black) and genome-edited A133P mutation-carrying cells (red). LD50 values: 2.3 – 6.4 nM (parental), >1000 nM (HeLa A133P). Error bars (s.d.) were generated from 3 Wst1 assays done in parallel. Two independent sets of experiments were performed. (e) Schematic highlighting key steps of DrugTargetSeqR. Genetically heterogenous cancer cells are treated with the drug of interest, resistant clones are isolated, and those resistant to MDR substrates excluded. The remaining clones along with the parental cells are processed for RNA-Seq. Mutations present in each drug-resistant clone, but absent in the parental (untreated) cell, are analyzed to identify genes that are frequently mutated in independent groups. CRISPR/Cas9 and drug-based selection is then used to isolate otherwise drug sensitive cells with these mutations.