Date Published: January 24, 2017
Publisher: Public Library of Science
Author(s): Katherine F. Donovan, Mudra Hegde, Meagan Sullender, Emma W. Vaimberg, Cory M. Johannessen, David E. Root, John G. Doench, Stefan Maas.
CRISPR/Cas9 screening has proven to be a versatile tool for genomics research. Based on unexpected results from a genome-wide screen, we developed a CRISPR/Cas9-mediated approach to mutagenesis, exploiting the allelic diversity generated by error-prone non-homologous end-joining (NHEJ) to identify novel gain-of-function and drug resistant alleles of the MAPK signaling pathway genes MEK1 and BRAF. We define the parameters of a scalable technique to easily generate cell populations containing thousands of endogenous allelic variants to map gene functions. Further, these results highlight an unexpected but important phenomenon, that Cas9-induced gain-of-function alleles are an inherent by-product of normal Cas9 loss-of-function screens and should be investigated during analysis of data from large-scale positive selection screens.
Deciphering the functional consequences of DNA variation is a defining challenge of the genomic era, and CRISPR/Cas9 technology is the most promising and broadly-developed tool for facile genome engineering [1,2]. Previously, we conducted pooled, genome-wide loss-of-function screens in A375 cells, a melanoma line with the BRAF V600E mutation that is sensitive to MAPK pathway inhibition . These positive selection screens utilized vemurafenib, a BRAF inhibitor, and selumetinib, a MEK inhibitor, to identify sgRNAs that induce drug-resistance in cells and therefore enrich in the cell population over time. As is standard for genetic screens, we then combined information from multiple sgRNAs intended to target the same gene to create a gene-level score. This method identified both previously-validated and novel mediators of this drug resistance phenotype [3,4]. However, examination of the sgRNA-level data revealed a curious result, namely that in one genome-wide subpool, an sgRNA targeting the gene MAP2K1 (which encodes the protein MEK1) at the site encoding K59 generated the strongest drug-resistance phenotype for both vemurafenib and selumetinib (Table 1). Another MAP2K1 sgRNA screened in a different subpool scored strongly with selumetinib but not vemurafenib. One would expect that sgRNAs targeting MAP2K1 for gene knockout, a positive regulator in the pathway inhibited by these drugs, would impair the viability of cells, not rescue them from the drug. Indeed, the other sgRNAs targeting this gene did not lead to drug resistance and were instead strongly depleted (Table 1). We hypothesized that this unexpected result was the consequence of NHEJ-mediated repair of the sgRNA cut site, which led to the creation of drug resistant variants of MEK1.