Editor’s note: Although we can now use thousands of CRISPR guides for screening and imaging, generating your own guide library can be a laborious and expensive process, costing several thousands of dollars. In this article, Andy Lane and his colleagues from Rebecca Heald’s lab at UC Berkeley describe a method to create sgRNA guide libraries for less than $100. Andy shares with Benchling how it works and also provides a detailed tutorials on how you can design your own guide RNA libraries.
Want to share your research with Benchling? Contact us.
Most studies thus generate sgRNA libraries using ssDNA oligonucleotide arrays, which create thousands of unique oligo sequences in parallel. These oligos are then cleaved from the arrays, collected as a pooled mixture, and via a PCR step, are converted into dsDNA molecules. These molecules are finally cloned into plasmid vectors for amplification, creating a defined set of known sgRNAs.
While this approach allows researchers to design the exact sgRNAs included in the library, the cost of synthesis increases as the complexity of the library increases, and even smaller libraries made in this way cost thousands of dollars.
In the Heald lab, we were interested in using CRISPR-imaging to image large sections of the genome, tiling hundreds to thousands of sgRNAs across chromosomal regions. When we began to calculate the expense of synthesizing libraries for very large regions, we quickly realized that each experiment was likely to require several thousand dollars’ worth of oligos. This drove us to start thinking about developing a less expensive method that would make it possible to label any target we wanted, routinely and relatively inexpensively.
We named this approach, CRISPR-EATING (Everything Available Turned Into New Guides). The general workflow looks like this:
CRISPR-EATING lets you make sgRNA libraries inexpensively using Streptococcus pyogenes Cas9 (SpCas9) in addition to a set of restriction enzyme and ligation steps on any source of DNA you’d like to target. We’ve been able to use this method to make libraries with thousands of guides for imaging, or libraries targeting 40,000 times in a bacterial genome—all for less than $100 in reagent costs . And in principle, this approach can be used to make very diverse libraries with DNA from all organisms.
CRISPR-EATING relies on a sequence of enzyme digestion and ligation steps that takes a purified source of DNA (e.g. genomic DNA or a cDNA library), extracts PAM-adjacent 20mers, and ligates them upstream of the constant sgRNA guide body and downstream of a promoter sequence.
You can use this protocol for creating a CRISPR-EATING library:
|Genome-wide screens||Genomic DNA||Inexpensive Generates a complex library||Guides might have significant off-target effects|
|Transcriptome-wide screens||cDNA||Inexpensive Generates a complex library||Guides might have significant off target effects|
|Imaging||PCR products covering region to be imaged||Less expensive than array-based synthesis when generating hundreds of guides||PCR step adds $2 to $4 in primer expenses per PCR used, and ~100 PCRs may be required to label megabase regions|
When designing a CRISPR-EATING library, it’s best to weigh in the pros and cons of CRISPR-EATING vs. array-based approaches:
The specifics of this will depend on:
As you can see, understanding the specificity profile of the library you want to generate via CRISPR-EATING can be an important factor in assessing likelihood of success in your application.
We have provided a Python module you can use to list the guides (i.e. 20mers next to the PAM-adjacent restriction sites) and the corresponding specificity scores that will be generated from a given input sample in your genome of interest.
Many researchers are now familiar with using CRISPR to design a single guide RNA towards engineering a cell line or model organism of choice at a single specific locus. But that’s not the only way to harness genome engineering tools.
Four papers were published in 2014 describing the use of hundreds or thousands of sgRNAs in a pooled library to screen for phenotypes, instead of a single sgRNA molecule [2-5]. The applications include using active Cas9 to generate mutations in genes and dCas9 fusion proteins to repress (CRISPRi) or activate (CRISPRa) transcription . For an example on how to carry out CRISPRi screens in yeast, read this Benchling article by Justin Smith at Stanford University; for a broader introduction to using using pooled libraries in screening, check out Addgene’s introduction to this topic.
In addition to screening, pooled libraries can also be designed for live, sequence-specific genome imaging by fluorescence microscopy (CRISPR-imaging): if hundreds to thousands of guides are tiled across a genomic locus and introduced in combination with a catalytically inactive fluorescently labeled Cas9, their target loci can be imaged in living cells [7-8].
In both imaging and screening, the flexibility of CRISPR guide libraries is beginning to revolutionize research and with the rapid pace of discovery of new applications. We hope that by lowering the cost of creating CRISPR guide libraries, researchers can continue to invent new uses for these reagents.