NEW YORK – Researchers in the US and the Netherlands have found that type I CRISPR-Cas systems can be used for genome engineering in eukaryotic cells, despite the complexity of their Cascade RNA-guided system.
Type I CRISPR systems are the most abundant adaptive immune systems in bacteria and archaea, but they've been largely avoided for editing applications in eukaryotic cells. Target interference in these systems relies on a multi-subunit, RNA-guided system called the CRISPR-associated complex for antiviral defense (Cascade), which recruits Cas3 for target degradation. In a study published on Monday in Nature Biotechnology, the researchers described their adaptation of Cascade to achieve RNA-guided gene editing in multiple human cell lines with high specificity and efficiencies of up to about 50 percent.
"Expression of the full Cascade-Cas3 complex in human cells resulted in targeted deletions of up to about 200 kilobases in length," the authors wrote. "Our work demonstrates that highly abundant, previously untapped type I CRISPR-Cas systems can be harnessed for genome engineering applications in eukaryotic cells."
The Cascade in Escherichia coli (EcoCascade) was the first RNA-guided DNA targeting complex to be biochemically characterized, and it represents the hallmark multi-subunit effector complex in type I CRISPR systems. EcoCascade is composed of a 61-nucleotide CRISPR-RNA (crRNA) and five separate Cas proteins. Several recent studies have elucidated the mechanistic details of EcoCascade's targeting reaction, and several research groups have succeeded in using Cascade for synthetic transcriptional regulation and genome editing in bacterial species.
For this study, the researchers aimed to explore the abundance and diversity of type I systems in nature, and the potential for applications that are not accessible with class 2 CRISPRs, such as Cas9. Type I systems require coordinated action of Cascade for DNA targeting and Cas3 for processive DNA degradation. The researchers engineered Cascade to make precise double-strand breaks (DSBs) by fusing the complex to a non-specific nuclease domain from FokI, a bacterial restriction endonuclease naturally found in Flavobacterium okeanokoites. They eventually succeeded in generating FokI–EcoCascade nucleases that catalyzed robust DSB formation in a PAM-, seed sequence- and paired gRNA-dependent fashion.
They then designed 16 paired gRNAs targeting therapeutically relevant Homo sapiens genes and purified the corresponding FokI-EcoCascade complexes. After incubating a plasmid substrate with FokI-EcoCascade, the researchers observed site-specific DNA cleavage of each target sequence. When they subsequently added intact FokI-EcoCascade RNPs to HEK293 cells, they observed up to about 4 percent editing efficiency. Among the 16 target sites tested, editing was largely restricted to sites containing 30-bp interspacer lengths.
Further experiments showed that the researchers were able to develop a simplified expression system to reconstitute an elaborate, 11-subunit RNA-guided nuclease in eukaryotic cells with just two molecular components that are similar in size to Cas9 and single-guide RNA plasmids. This two-plasmid nucleofection approach also enabled them to significantly increase the scope of engineering by exploring parameter space around complex assembly and targeting.
Next, the researchers looked at additional Cascade homologs in order to uncover variants with enhanced biochemical activity or target flexibility in eukaryotic cells. They selected 11 additional Type I-E systems and generated polycistronic and paired gRNA expression vectors analogous to the E. coli subtype. They found a Cascade homolog from Pseudomonas sp. S-6–2 that induced editing efficiencies that were approximately tenfold higher than the E. coli homolog.
They then selected Cascade homologs from E. coli, Pseudomonas sp. S-6-2 (PseCascade), and Streptococcus thermophilus (SthCascade) for more detailed characterization and screened them across a panel of 96 genomic target sites.
"FokI-PseCascade consistently yielded, on average, about 15 to 25 percent editing efficiencies within a 30 to 33-bp interspacer window, and some targets exhibited up to about 40 to 50 percent indels," the authors wrote. "Similar trends were observed with the other homologs."
In examining the specificity of FokI-PseCascade editing activity, the researchers noted that a targetable site can be expected to occur about every 30 bp in the human genome, whereas potential off-target sites with optimal interspacer distances will be rare. They nucleofected HEK293 cells with plasmids expressing either FokI-PseCascade and one of four high-efficiency paired gRNAs or Cas9 and a gRNA previously shown to induce substantial off-target activity. They observed more than 250 Cas9 off-targets, ranging from 11,000 to about 41,000 reads, compared to two off-targets, at most, across any of the FokI-PseCascade paired gRNAs.
"Taken together, these data show that FokI-Cascade can be used for high-specificity eukaryotic genome editing," the authors concluded. "To our knowledge, this is the first report demonstrating functional, expression-based reconstitution of class 1 CRISPR-Cas systems in mammalian cells, and we anticipate that our streamlined plasmid designs will similarly allow engineering of other diverse Cascade variants, which contain fewer protein components and use unique PAM requirements, or even the RNA- and DNA-targeting effector complexes from type III CRISPR-Cas systems."