Why use Cas9:sgRNA libraries for functional genomic screens?
As Cas9 target specificity is determined by short 20bp sequences of the sgRNA, large scale oligo synthesis of guide sequences suggests a new way to interrogate gene function at a genome-wide scale. We showed that delivery Cas9:sgRNAs libraries using lentiviruses can facilitate both positive and negative loss-of-function screening in mammalian cells (Shalem*, Sanjana* et al., Science 2014, Sanjana*, Shalem*, et al, Nature Methods 2014). When Cas9 induced double strand breaks are targeted to coding sequences, error prone repair will introduce a frame shift with high frequency resulting in a loss-of-function allele. While shRNAs result in a wide range of expression values for the target gene, using CRISPR, the population of cells which are homozygously knocked out are expected to provide higher screening sensitivity, especially in cases where reduced expression retains gene function.
In addition, we show that GeCKO screening results in extremely high modification efficiency, low off target modification for optimized sgRNAs, high consistency between unique sgRNAs targeting the same genes and high validation rate of screen hits. These results demonstrate the promise of Cas9 mediated screening to revolutionize functional genomics!
Choice of vector system
Before starting you will need to get familiar with the delivery system. Depending on the number and type of cell lines used for screening one can choose between two general library delivery approaches: the single and dual vector system.
In the two vector system you will first transduce cells with Cas9, expand a cell population (or select clones) and then transduce the sgRNA library using the second vector. Using the one vector system, you will transduce a large number of cells (see protocols) with both Cas9 and the sgRNA in a single vector. The disadvantage of the later approach is that the currently available all-in-one vector results in low titer which might be especially problematic in cells which are more difficult to transduce.
How do I get a GeCKO library for my own lab?
The GeCKO library and lentiviral CRISPR reagents are available on Addgene. Click here to request it.
Custom library design, synthesis and cloning
You can design your own custom spacer library targeting any set of genetic elements of interest as detailed in our step-by-step cloning protocol. We recommend to incorporate minimization off-target effects in the design using the most recent information in http://tools.genome-engineering.org/.
Duration of puromycin selection
Following transduction of the sgRNA library (with or without Cas9) an important question is how long to perform puromycin selection. The minimal duration is the one at which a non-transduced control population shows 100% death. This will ensure that only transduced cells will be used to screening. Extension of this time may be used to ensure a higher rate of genomic modification. For example, we observed almost complete modification for most loci that we examined after 7 days (see supplementary figures). Yet this parameter will most likely vary across target sites and cell types and we recommend to check it in each model system.
Screen background and off target effects
Every screen will have some degree of sgRNA background depletion/enrichment determined by the strength of selective pressure applied. This means that depletion/enrichment of sgRNAs can be observed despite having no effect on the selected phenotype. For example, if you apply a drug that kills only 95% of cells then the remaining surviving cells will be read out despite providing no selective advantage. To discriminate between true and false positive hits, there are two principal methods: transduction replicates and reagent redundancy.
Transduction replicates: We recommend to include biological replicates of the viral transduction step in addition to expanded cell populations post-transduction. Depending on the background level of each screen, requiring consistency between transduction replicates will exponentially reduce false positive rate with the number of replicates. Infection replicates will not eliminate false positive hits which are due to sgRNA off target modification!
Reagent redundancy: A key library feature which is used to discriminate true and false positive hits is the incorporation of multiple unique sgRNAs targeting the same genetic elements (if possible). This controls for a few things: low efficacy or off target effects of single guides, and for null background depletion/enrichment. Genes that display high depletion/enrichemnt across multiple sgRNAs are most likely true hits and are expected to be consistent across transduction replicates.
Following selection, cell cultures are collected for readout of sgRNA representation. Readout is done by amplifying the integrated lentiviral construct from genomic DNA using PCR and next generation sequencing to evaluate relative representation for each unique sgRNA. It is important to carefully follow the calculation of cells/DNA amounts/sequencing reads (see supplementary methods in Shalem*, Sanjana*, et al., Science 2014) in all steps from transduction to next generation sequencing to ensure no loss of representation between steps. Analysis is based on following changes in sgRNA representation at different time points, thus multiple time points are required, e.g. before and after selection or comparison of the initial plasmid pool before virus production.
the GeCKO team