In the Wild West of CRISPR genome editing, Cas9 has established itself as a wrangler of animal models that takes out any sequence with an sgRNA bounty on its head. But now there’s a new sheriff in town with marksman aim. CRISPR from Prevotella and Francisella 1 (Cpf1) was brought forth by the pioneering lab of Feng Zhang and shot to the spotlight as not only a potential CRISPR/Cas9 patent disruptor, but also a nuclease with some important differences from Cas9.
Here’s a refresher on how Cpf1 stacks up against Cas9:
- Cpf1 is made for delivery. Not only is it smaller, but it uses a single RNA guide (cRNA), while the sgRNA used for Cas9 is a fusion of two distinct RNAs (tracrRNA and crRNA). This means a smaller sgRNA is required for Cpf1 genome editing.
- Cpf1 recognizes a T-rich protospacer adjacent motif (PAM), as opposed to the G-rich PAM of Streptococcus pyogenes Cas9 (SpCas9), opening up new targeting possibilities.
- When it cuts, Cpf1 produces sticky ends instead of the blunt ends Cas9 produces, which could make for easier non-homologous end joining (NHEJ)-based gene knock-ins.
The Zhang lab also showed that Cpf1 systems from Lachnospiraceae bacterium (LbCpf1) and Acidaminococcus sp. (AsCpf1) could efficiently edit human cells by targeting DNMT1. Now, three new papers from South Korea bring Cpf1 to maturity by bringing forth an analysis of off-target effects and knockout mice.
Off-Target Effects in Human Cells: Cpf1 Ribonucleoproteins (RNPs) Hit the Mark
One of the biggest challenges currently facing genome editing is off-target effects, and a number of techniques have emerged to study them. By utilizing these techniques, the team tested Cpf1 and provided some much-needed direct comparison to Cas9 function in human cell lines.
- When compared across ten chromosomal sites, SpCas9 has a higher on-target mutation frequency than LbCpf1 and AasCpf1, which have frequencies similar to Staphylococcus aureus Cas9 (SaCas9).
- Cpf1 is highly specific but it isn’t perfect: it doesn’t tolerate mismatches in the 5’ proximal-PAM region, but single and double mismatches can exist in the 3’ PAM-distal region.
Next, the team compared Cpf1 to Cas9 by targeting them both to the same site and analyzing the off-target effects with Digenome-seq, where they found that both LbCpf1 and AasCpf1 cleave at fewer off-target sites than SpCas9. Thus, SpCas9 appears to produce more on- and off-target effects than Cpf1.
After determining the above, the team made their precision breakthrough, as senior author Jin-Soo Kim shares, “To reduce off-target effects, we introduced preassembled Cpf1-crRNA RNPs into cells, assuming that Cpf1 RNPs, similar to Cas9 RNPs, would cleave target sites immediately after transfection and would be degraded rapidly by endogenous enzymes that break the proteins, thereby reducing off-target effects without sacrificing on-target effects.” This modification produced such an improvement that, as first author Daesik Kim puts it, “Notably, both LbCpf1 and AsCpf1 targeted to a certain site cleaved only the on-target site in the entire human genome.” Ultimately, the study shows that using pre-assembled recombinant RNPs makes Cpf1 a marksman genome editor that can take on SpCas9 in vitro.
Cpf1 Generates Knockout Mice
As if all the above data wasn’t enough, the second paper extends the power of Cpf1 from in vitro to in vivo, and demonstrates Cpf1 knockout mice using LbCpf1 and AsCpf1. These mice had Trp53 knocked out by embryonic microinjection. They exhibited frequent mosaicism (63% for AsCpf1 and 46% for LbCpf1) and some minimal off-target effects, which could be due to the delivery choice of of mRNA instead of RNP.
Finally, the third paper takes an interesting twist on the delivery route for AsCpf1. Their clever approach utilized electroporation instead of microinjection to deliver RNPs to mouse embryos, which allows for RNP delivery in up to 50 embryos simultaneously in a lightning speed of under 5 minutes.
Using their new approach, the team targeted Foxn1, which controls the growth of skin hairs, as well as Tyr, which is involved in pigment production. These two genes made for easily observable phenotypes to help identify successful editing. First author Junho Hur shares, “The data showed that Cpf1 RNP’s delivery to mouse embryos resulted in knocking out the intended genetic functions with high mutation frequencies, 64% and 33% respectively. We transplanted the mutant mouse embryos into surrogate mothers and obtained mice with targeted mutations. The mutations resulted in hairless and white-haired mouse respectively.”
But the team wasn’t done there, as Hur adds, “To investigate whether Cpf1 had off-target effects, we performed whole genome sequencing using genomic DNA isolated from one Foxn1 mutant mouse and its wild-type sibling. The sequence analysis showed that no off-target mutation occurred. Targeted deep sequencing of other mutants also showed no off-target mutation.”
Thus, electroporation of Cpf1 RNPs offers a promising new avenue for the generation of animal models. Taken together these 3 papers show that Cpf1 is a precise CRISPR genome editing system that could give Cas9 a run for its money.