Genome-scale CRISPR-Cas9 knockout and transcriptional activation screening

Abstract

Forward genetic screens are powerful tools for the unbiased discovery and functional characterization of specific genetic elements associated with a phenotype of interest. Recently, the RNA-guided endonuclease Cas9 from the microbial CRISPR (clustered regularly interspaced short palindromic repeats) immune system has been adapted for genome-scale screening by combining Cas9 with pooled guide RNA libraries. Here we describe a protocol for genome-scale knockout and transcriptional activation screening using the CRISPR-Cas9 system. Custom- or ready-made guide RNA libraries are constructed and packaged into lentiviral vectors for delivery into cells for screening. As each screen is unique, we provide guidelines for determining screening parameters and maintaining sufficient coverage. To validate candidate genes identified by the screen, we further describe strategies for confirming the screening phenotype, as well as genetic perturbation, through analysis of indel rate and transcriptional activation. Beginning with library design, a genome-scale screen can be completed in 9-15 weeks, followed by 4-5 weeks of validation.

Figure 1. Approaches to genetic perturbation: shRNA knockdown, Cas9 knockout, and Cas9 transcriptional activation

Schematic of the mechanisms behind shRNA knockdown, Cas9 knockout, and Cas9 transcriptional activation. ShRNA knockdown begins with processing of the shRNA by Drosha/Dicer machinery and results in degradation of an RNA transcript with a complementary target site by the RNA-induced silencing complex (RISC). Cas9 knockout is accomplished by targeted indel formation at a genomic site complementary to the sgRNA. An indel can result in a frameshift, causing early termination, and either production of non-functional protein or non-sense mediated decay (NMD) of the mRNA transcript. Programmable transcriptional activation can be achieved using dCas9 and activation domains (e.g. VP64/p65/HSF1) to recruit transcriptional machinery to the transcriptional start site of the desired gene target, resulting in upregulation of the target transcript. PAM, protospacer adjacent motif; NHEJ, non-homologous end joining; Pol II, RNA Polymerase II.

Figure 2. Timeline and overview of experiments

Genome-scale Cas9 knockout and transcriptional activation screens begin with the construction of a plasmid library encoding the effector protein and sgRNAs. These plasmid libraries are packaged into lentivirus and then transduced into the cell type of interest to generate stably expressing lines for the screen, along with an accessory transcriptional activator complex (MS2-p65-HSF1) lentivirus for the case of activation screening. A selection pressure is applied depending on the nature of the screen and at given timepoints, genomic DNA is harvested. The sgRNA regions (colored bars) are amplified from genomic DNA and then analyzed by next generation sequencing followed by statistical analyses (e.g. RIGER) to identify candidate genes. Candidate genes are then validated by various forms of analysis, including testing individual sgRNAs for the screening phenotype, indel formation by targeted sequencing, or transcript upregulation by qPCR.

Figure 3. GeCKO and SAM libraries for genome-scale knockout and activation screens

(a) For knockout screening, the GeCKO v2 libraries target the 5′ conserved coding exons of 19,050 human or 20,611 mouse coding genes with 6 sgRNAs per gene. (b) The GeCKO library is available in a 1 vector or 2 vector format. (c) For activation screening, the SAM libraries target the 200bp region upstream of the transcriptional start site of 23,430 human or 23,439 mouse RefSeq coding isoforms with 3 sgRNAs per isoform. (d)The library has to be combined with additional SAM effectors in a 2 vector or 3 vector format. Both libraries select sgRNAs with minimal off-target activity.

Figure 4. Anticipated results for genome-scale knockout and activation screens

We provide data from genome-scale knockout and activation screens for identifying drivers of resistance to the BRAF inhibitor vemurafenib (PLX) in a BRAF^V600E (A375) melanoma cell line^,. (a,b) Box plots showing the distribution of sgRNA frequencies after control (Veh, vehicle) or PLX treatment from n = 2 infection replicates. A significant number of guides are seen enriched and depleted in the PLX day 14 condition, revealing depletion of guides essential for cell growth and enrichment of guides that promote resistance to BRAF inhibitor. Boxes, 25^th to 75^th percentile; Whiskers, 1^st to 99^th percentile. (c,d) Scatterplot showing enrichment of sgRNAs targeting the top candidate genes identified by RIGER (colored dots) compared to other sgRNAs in the library (grey dots) after PLX treatment. Each gene has multiple sgRNAs that are enriched. Inset panel represents the entire dataset. Many of these genes are known tumor suppressors or oncogenes that play a role in PLX4720 resistance. (e,f) The top hits of the screen are seen as distributed across the genome, revealing the necessity of genome-scale screens for identifying drivers of resistance. RIGER P values for candidate enriched genes (colored dots) are significantly lower compared to other genes (grey dots) targeted by the sgRNA library.