June 1, 2018 (Vol. 38, No. 11)
Caroline Seydel Contributor GEN
Innovations in Genome Editing on Multiple Fronts
The easy customization of CRISPR gene editing tools has captured the public imagination, but the benchtop reality is never so simple. Beneath every flag of inspiration lies a mountain of perspiration to bring a concept to fruition. The 4th Annual Precision CRISPR Congress gathered innovators from all areas of industry and academia to share their newest developments in genome editing.
Gene editing tools lend themselves to a variety of applications, and in each area, researchers constantly seek ways to improve efficiency. Companies continually introduce larger, better libraries for genome-wide screening They dream up new tricks to adapt this bacterial immune system for use in eukaryotic cells. But behind all the applications lies a basic study of the Cas9 nuclease, the enzyme that cuts the DNA. Cas9 hooks up with a short sequence of gRNA, or guide RNA, which draws the nuclease to complementary sequences within the genome. Selecting the right gRNA sequence is the first challenge facing would-be CRISPR users.
No Dud RNAs
“Some gRNAs just don’t work very well,” said Bradley Merrill, Ph.D., scientific director for the Genome Editing Core at the University of Illinois, Chicago. He and graduate student Ryan Clarke conducted a mutagenesis study with 40 gRNAs, and some of them failed to generate mutations. “We looked at the ones that performed poorly, and we found that they all happened to anneal to the same strand of the target gene,” Dr. Merrill noted.
After making a double-stranded break, Cas9 clings onto the DNA for about 5.5 hours, and this reluctance to disengage can slow down genome editing. Clarke and Dr. Merrill showed that when the gene is transcribed, the RNA polymerase will knock Cas9 off the broken strand—but only if the gRNA is bound to the strand that is actively transcribed. If the gRNA anneals to the nontemplate strand, then Cas9 will stay put—preventing DNA repair machinery from getting at the break, and preventing efficient mutagenesis.
“It’s a rate-limiting step in genome editing,” Dr. Merrill pointed out. It may be possible to get around this limit by modifying Cas9 in some way. “If you can turn it into a multi-turnover enzyme, you can use fewer molecules of Cas9,” he continued. Using less enzyme could reduce the incidences of off-target cutting, for instance, or mitigate the possibility of an immune reaction being mounted against Cas9.
More immediately, the finding will help researchers design gRNAs that actually work. “You can use the direction or the orientation of RNA polymerase to eliminate a lot of bad guides, or duds, simply by paying attention to which strand they’re going to anneal to,” Dr. Merrill asserted.
Building a Better Library
Avoiding dud gRNA sequences gets to the heart of improving CRISPR-based tools for genome-wide screening. “You need to design good single guide RNAs (sgRNAs) in order to have a good screen,” says Paul Diehl, Ph.D., COO, Cellecta. The company offers various pooled sgRNA libraries that cover the entire human genome. To optimize the performance of the screening libraries, the company looked for ways to improve the sgRNA sequences.
Besides the targeting sequence itself, Dr. Diehl pointed out, “there was some indication that the rest of the single-guide sequence that binds to Cas9 could be optimized.” Accordingly, Cellecta tested various modifications of the sgRNA sequences to improve Cas9 recruiting efficiency until the company found a version that boosted the signal strength of the library hits.
“We’re looking for depletion or enrichment of specific guides in the library relative to the whole population,” Dr. Diehl stated. “A fivefold depletion is going to be easier to see than a twofold depletion.”
In addition to its whole human genome knockout library, Cellecta recently introduced CRISPRi (inhibition) and CRISPRa (activation) libraries, which cover the entire genome with five constructs per promoter region. Whereas a standard CRISPR knockout construct stops transcription by directing Cas9 to cut the genomic DNA, these libraries use a modified Cas9 that doesn’t cut. In this way, the constructs turn genes on or off without permanently modifying the genomic DNA. With five sgRNAs per protein coding gene, the library provides built-in redundancy to confirm that the results are real.
“If just one of those gives an effect, it may be something weird going on,” Dr. Diehl noted. “You want to see at least three or four of them to the same target have the same effect on the cells.”
Synthetic gRNAs for Gene Activation
Pooled libraries provide a powerful tool to sift through thousands of genes at once. Once a screen has uncovered some genes of interest, those genes need to be turned on individually for more detailed study. To that end, Horizon Discovery (formerly Dharmacon) has introduced synthetic gRNAs for CRISPRa. “It’s really a next-generation way to do gene-overexpression studies,” said Louise Baskin, senior product manager for Dharmacon RNA reagents at Horizon Discovery.
The Edit-R system uses a deactivated Cas9 nuclease fused to transcriptional activators. The synthetic gRNA directs the enzyme where to bind, and then, rather than cutting the DNA, it activates transcription. “We can provide this as tubes or arrayed libraries up to and including the whole human genome,” Baskin added. “Researchers can request gRNAs targeting just one or two genes or whole classes of genes, such as all ubiquitin enzymes. That’s really the power of this tool.”
CRISPRa improves on previous gene-activation strategies because it induces the cell to express the gene in its native form. Rather than introducing a gene into the cell based on a consensus sequence, CRISPRa reagents simply switch on the gene that’s already present in the cells being studied. “You’re getting the version of the gene that that cell type is intended to have,” Baskin asserted. “You’re getting more of the real biology.”
Storming the Chromatin Castle: Proxy-CRISPR
Depending on the task at hand, a researcher might look beyond the standard Cas9 (called SpCas9, because it’s derived from Streptococcus pyogenes) and select a Cas9 ortholog from a different bacterium. While the various Cas9 nucleases bring different strengths to the table, they may need some tweaking to boost their performance in eukaryotic cells.
“MilliporeSigma’s R&D team has uncovered some surprisingly effective methods for increasing CRISPR specificity and activity,” said Gregory D. Davis, Ph.D., head of genome engineering R&D at MilliporeSigma. Working with FnCas9 (derived from Francisella novicida), the MilliporeSigma group discovered that in human cells, the nuclease didn’t cut as well as it did in purified DNA. It seemed possible that the chromatin structure could be preventing the nuclease from binding to the target sequence.
To get the chromatin out of the way, MilliporeSigma developed a system called proximal CRISPR targeting, or proxy-CRISPR. By sending in a deactivated SpCas9, designed to bind a nearby sequence, they opened up the chromatin structure, allowing the FnCas9 to move in on its target sequence and cut. “This approach of proximal targeting turned out to work wonderfully to enhance genome editing of many different CRISPR systems,” declared Dr. Davis.
Cas9 homologs can have different protospacer adjacent motif (PAM) sequences, and that could make them desirable if the needed PAM lies adjacent to a gene sequence of interest. But some Cas9 variants need a little help getting through the chromatin thicket. “If a CRISPR complex has the right PAM to cut closely to a disease SNP, but is blocked by chromatin, proxy-CRISPR can assist to open the region to editing,” Dr. Davis insisted. “Furthermore, we also found that proxy-CRISPR can enhance homology directed repair, enabling more efficient integration of donor DNA.”
RNA-Guided Aspects of CRISPR
The flexibility of RNA-guided aspects of CRISPR have enabled rapid innovation of alternative formats to drive genome editing projects with clinical trajectories. One key goal is the minimization of off-target effects. It has been well established for both ZFN- and CRISPR-based genome editing work that protein-based delivery of targeted nucleases offers a significant reduction in off-target cutting (Gaj, et al., 2012; Zuris et al., 2015), according to Gregory D. Davis, Ph.D., head of genome engineering R&D at Millipore Sigma.
“While ZFN protein sequences change for every new genomic target site, the protein parts of CRISPR remain the same with the only molecular change happening within the guide RNA (gRNA) sequence,” says Dr. Davis. “This has allowed the Cas9 protein to serve as a chassis for engineering versions with enhanced specificity.
“MilliporeSigma offers a broad array of CRISPR proteins, with two specifically aimed at minimizing off-target effects: (1) Cas9-D10A nickase (CAS9D10APR) for implementation of paired gRNAs, and (2) the high-fidelity eSpCas9 protein (ESPCAS9PRO) for projects which prefer genome editing with a single gRNA design.”
The primary difference between these two formats is that paired nickases use twice the gRNA sequence to create a double-strand break, enhancing specificity through doubling of the recognition sequence.
“Thus,” he concluded, “if it is difficult to find a highly specific gRNA sequence for eSpCas9 in a certain region, the specificity requirements for gRNA design in paired nickases can be relaxed slightly to find suitable designs to cut within the region you need and maintain high specificity across the larger region of two gRNA sequences.”