Tools for Targeted Mutagenesis

Gene editing with the CRISPR/Cas9 system is among the hottest tools in biomedical research, fast eclipsing the excitement spawned by RNAi, transcription activator-like effector nucleases (TALENs) and, before them, zinc finger nucleases (ZFNs). The CRISPR/Cas9 system includes a guide RNA (gRNA) that directs the Cas9 nuclease to make a double-stranded break in the DNA at a locus corresponding to the gRNA sequence. This is then repaired by the cellular machinery, often introducing base changes that can inactivate the gene.

Because CRISPR/Cas9 editing does not require any protein engineering, site-specific gene editing has become democratized. In fact, is it possible not only for researchers to easily knock out single genes seriatim, but also to perform high-throughput screens en masse, using libraries to cast a wide net upon the whole genome. Here we explore how researchers are using gene-editing libraries to perform targeted mutagenesis studies.

Pools vs. arrays

Broadly speaking, gRNA libraries come in two distinct flavors. Perhaps the more familiar of these is the pooled library, in which collections of (in this case) gRNAs are introduced to the cells of interest, in theory resulting in the generation of a collection of cells with different gene disruptions. Such libraries typically are used for functional genomic screening—“for target discovery, predominantly, identifying novel genes, novel drug targets, gene/drug associations,” notes Sunitha Sastry, senior product manager at Cellecta.

With pooled screens, “you can cover a lot of ground cheaply and quickly, because you’re basically screening one tube full of hundreds or thousands of clones at a time,” explains Shawn Shafer, functional genomics global market segment manager at MilliporeSigma. By applying a strong selective pressure—drug or environmental perturbation, for example—you can skew the representation of the clones.

The disadvantage of pools is that identifying hits requires deep sequencing followed by a statistical analysis of which gRNAs are under-represented, based on the recipient cells’ competitive growth disadvantage. “That still requires you to go and individually confirm afterwards that that clone was having an effect, so all pooled screens are followed by arrayed screens anyway,” Shafer says.

MilliporeSigma just launched the first-ever arrayed CRISPR whole-genome library, Shafer says, “where every clone is in its own well,” and there are about 34,000 to 35,000 individual wells.

There are, to be sure, already many arrayed CRISPR libraries on the market—from Thermo Fisher Scientific and GE Dharmacon, for example—representing different subsets of the genome, such as the kinome and apoptosis-related proteins. Other vendors are in various stages of creating their own whole-genome arrayed libraries, as well.

What’s in the library?

The raison d'être of a CRISPR library is typically to provide a means of knocking out a collection of genes by systematically delivering gRNA to the target cells.

Sometimes these cells already contain the Cas9 nuclease, but in most instances it needs to be introduced, along with the gRNA. How is this done?

Plasmids are an efficient starting point—two plasmids separately encoding the gRNA and nuclease, or a single plasmid encoding both. These are transfected into the targets, says Jon Chesnut, leader of Thermo Fisher Scientific’s synthetic biology R&D group. Libraries can now be found in a host of formats—as plasmids that are used to create viral particles in a packaging line, for example, as synthetic or in vitro transcribed guide RNA or as ready-to-transduce encapsulated viral particles—each with its own merits.

Plasmids, for example, are very easy to propagate. They have the advantage of being able to incorporate markers such as antibiotic resistance or a fluorescent tag. They can contain cell type-specific or inducible promoters. And they can be a theoretically unlimited source of both guide RNA and Cas9 mRNA—something that has been linked to off-targeting, points out Mark Behlke, chief scientific officer at Integrated DNA Technologies (IDT).

Lentivirus, too, can provide an unlimited source of reagent. And it can be used to readily transduce stem cells, primary cells and otherwise intransigent cell types.

The CRISPR field—especially for those with the appropriate expertise and resources—is shifting toward the use of complexes of gRNA and enzyme in the form of ribonucleoprotein (RNP) complexes (or in some cases, gRNA complexed with mRNA), says Chesnut: “In general, that protein-complex delivery is evolving to be the go-to platform. We’re seeing that reflected in the customer base that we deal with.”

Guide RNA

The term “guide RNA” has intentionally thus far been used agnostically. It can mean the semiduplex hybridization of CRISPR RNA (crRNA) and trans-activating crRNA (tracrRNA)—as it is found in the Streptococcus pyogenes bacteria from which the system was derived, and what Behlke refers to as the “natural two-part system.” Or it can mean a “single guide” RNA (sgRNA or single gRNA), which is a fusion of the two connected by a linker. Libraries are available in both of these configurations.

Guide RNA can be made either by in vitro transcription (IVT) or chemically synthesized, and the providers of both formats sing their respective praises. Thermo Fisher, for example, has benchmarked its IVT sgRNA “against synthetic guides that are available on the market,” Chesnut notes, “and found that ours are by all accounts more efficient and active in creating the cleavage event” and display great stability even at room temperature. On the other hand, synthetic oligo manufacturers point to the ability to modify the RNA, for example so that it is not seen by the cell as foreign.

When using plasmids or virus to create the RNA, there is an advantage to using sgRNA in that it requires only a single promoter and a single transcription unit. But single guides become a disadvantage when chemically synthesizing the RNA, because sgRNAs are very long and thus tend to be more expensive to make and of lower quality than shorter RNAs, says Behlke. Thus many synthetic guide RNA libraries are sold as collections of target-specific short crRNA—which can be complexed with the protein and tracrRNA that are universal for every target, he notes.

Getting it into the cell

As nucleic acids and proteins are the vehicles transiently delivering CRISPR tools to target cells, it stands to reason that your favorite transfection reagents will do the trick. Nonetheless, several vendors offer reagents specifically designed to work with this technology. For example, Thermo Fisher’s Lipofectamine™ CRISPRMAX™ Cas9 Transfection Reagent is optimized for delivering RNPs, Chesnut says. MilliporeSigma’s polymeric TransIT®-CRISPR transfection reagent can “accomplish highly effective transfection of CRISPR/Cas9 DNA, RNA, Cas9-gRNA RNP (ribonucleoprotein) complexes and siRNA,” according to the product bulletin. And MTI-GlobalStem’s DNA-In® CRISPR transfection reagent is optimized for delivery of large plasmids, such as the popular CRISPR/Cas9 all-in-one vector, notes director of marketing Donna Trollinger.

IDT has found that very few cationic lipid formulations work well for RNPs, but if siRNA can be transfected into a cell line using RNAiMAX (their preferred reagent) then the CRISPR RNP complex can likewise be transfected. But the most interesting cell types (such as primary human T-cells, for example) do not work with lipofection, Behlke says, so then it is necessary to resort to electroporation – both Thermo Fisher’s Neon® and Lonza’s Nucleofector™ electroporation systems yield good results in their hands.

Choosing a library

In addition to the above notes, researchers should consider how many guides per gene a library contains. “We have a library from the Moffat lab that has 12 guides per gene, so it’s a really huge library, and it allows you get a lot of data from one screen,” says Joanne Kamens, executive director of Addgene, a nonprofit repository with many pooled CRISPR libraries and “thousands of CRISPR-related plasmids from over 125 depositors.” Placement of the guides may be a consideration, as well as how they are designed—newer algorithms, for example, take into account the role of mismatches and provide other insights to avoid off-target effects and inefficiencies. There are even libraries which, with the appropriate enzyme, can enable inhibition or activation, or facilitate epigenetic changes.

Ultimately, Kamens recommends that researchers look at the literature describing libraries of interest (or the data generated with them) to determine the best fit for what they’re hoping to accomplish.