NEW YORK (GenomeWeb) – CRISPR/Cpf1, the alternative genome editing system being touted by its discoverers at the Broad Institute and the Massachusetts Institute of Technology, performs genome editing in human cells as efficiently as CRISPR/Cas9 and is highly specific, according to two recently published studies.
Last month, scientists from South Korea's Seoul National University, led by senior author Jin-Soo Kim, published the first independent reports characterizing the on- and off-target activities of CRISPR/Cpf1. His lab suggested that two Cpf1 proteins were able to successfully cause indel mutations in HEK293T cell lines at rates of approximately 20 percent, compared to approximately 32 percent for SpCas9, the gold standard CRISPR/Cas9 system. The study, published in Nature Biotechnology, also suggested that Cpf1 was extremely sensitive to single base pair mismatches in the target region of the genome.
Earlier this week, another study, also published in Nature Biotechnology, reported similar findings. Led by first authors Benjamin Kleinsteiver and Shengdar Tsai and senior author Keith Joung, scientists from Harvard University and Massachusetts General Hospital performed experiments in HEK293 and U2OS cells.
"You get very high rates of mutagenesis when using Cpf1 to generate indels across a range of different sites," Joung told GenomeWeb. Moreover, "the genome-wide specificity of Cpf1 appears to be very, very high," he said.
"There is a lot of excitement about this enzyme," Joung said. Like Cas9, Cpf1 is an RNA-guided DNA-cleaving nuclease; however, it uses a shorter, one-component CRISPR RNA (crRNA), as opposed to Cas9's two-component system that is typically engineered into a single RNA construct. Cpf1 also showed promise for multiplex genome editing, he said.
"The question was, 'How robust is its on-target efficiency and what's the genome-wide specificity like?'" Joung said. "With this work, we're very convinced that this is an enzyme worth pursuing, both for research and possibly therapeutic use."
Both labs looked into Cpf1 proteins taken from Acidaminococcus sp BV3L6 (AsCpf1) and Lachnospiraceae bacterium ND2006 (LbCpf1), the same proteins that were initially highlighted in the study marking that CRISPR system's discovery by MIT and Broad scientist Feng Zhang and the National Center for Biotechnology Information's Eugene Koonin.
Cpf1 has garnered interest because it is a natural two-component (protein and RNA) system and recognizes a different protospacer-adjacent motif (PAM) than Cas9. While Cas9 can only cut targets proximal to an NGG PAM, Cpf1 generally recognizes a TTN PAM, allowing it to be used in thymine-rich regions . Cpf1 targets a 27-base pair region downstream of the PAM.
Both Kim and Joung assayed the Cpf1 proteins to validate their robustness before moving forward. The biggest differences between their studies were in the number of guide RNAs used to target each gene in determining on-target efficiency — 8 for Kim, 20 for Joung — and the method developed by each lab for detecting off-target edits. Kim's lab has developed an in vitro method called Digenome-seq, while Joung's lab uses Guide-seq, a targeted sequencing method.
"Specificity was one of the big questions left unanswered [by Zhang's original study]," Joung said. "It's nice that the two papers reinforce each other."
"We found AsCpf1 and LbCpf1 both performed extremely well," Kleinsteiver added. "In many cases, there were no off-targets that we could detect."
In addition to genome-wide specificity, both Kim and Joung looked at Cpf1's tolerance for mismatches at various positions in the 27-base pair target region contained in the Cpf1 guide RNA.
"It's another way of looking at the specificity of the enzyme," Joung said. "Instead of the DNA target having the wrong base, you're deliberately mismatching on the guide RNA side."
For one of the genes Kim's lab looked at, DNMT1-4, Cpf1 did not tolerate single mismatches within the first 17 base pairs of the target region. "Double mismatches at positions 1 to 18 resulted in an almost complete loss of Cpf1 activity. For the other two sites, single mismatches at most positions were tolerated by Cpf1, whereas double mismatches at positions 1 to18 led to substantial loss of Cpf1 activity," the authors wrote.
Joung's analysis showed that Cpf1 was tolerant of single base pair mismatches at either end of the target region and somewhat in the middle, but there were two troughs in between where a mismatch would prevent off-target edits almost completely.
Kleinsteiver added that AsCpf1, compared to LbCpf1, was "consistently more specific, which is useful information for anyone who's interested in using these nucleases."
AsCpf1 and LbCpf1 are "at least as good as wildtype SpCas9," which notoriously induces off-target edits, "and probably quite a bit better," Joung said. "Having said that, we aren't yet sure if they're as good as high-fidelity Cas9" — a variant of SpCas9 that Joung's lab engineered to be more specific than the wild-type SpCas9. "But we believe that they're closer to hi-fi SpCas9 than wildtype," he said.
The sensitivity of Cpf1 to single-base mismatches opens up the possibility of allele-specific genome editing of heterozygous alleles, Joung said. Cpf1 could be used to this end if the base pair difference between the alleles fell into one of the regions of the target sequence that was intolerant to mismatches.
"If you could position a base different between two alleles in one of those regions, perhaps you could use this system to distinguish between two alleles differing by only one base," Joung said.
Though the Kim and Joung labs only looked at the ability of Cpf1 to cause indel mutations via the non-homologous end joining pathway, Joung said Cpf1 should be "perfectly capable" of inducing homology-directed repair and incorporate larger DNA constructs into edit sites.
If future studies confirm this, Cpf1 will be well on its way to joining SpCas9 in the gene editor's toolkit. As Joung said, these studies give the green light for development of the Cpf1 enzyme.
For example, scientists have developed a whole suite of additional applications using CRISPR/Cas9, like protein fusions for gene expression and structural genomic research, and the same could be done for Cpf1.
"Those types of applications are worth exploring," Joung said.