Characterizing the specificity and nicking activity of CRISPR endonucleases, Cas9 and Cas12a

The bacterial adaptive immune system is composed of a CRISPR (clustered regularly interspaced short palindromic repeats) locus and CRISPR-associated (Cas) proteins. The effector proteins are CRISPR (cr) RNA-guided nucleases that cleave crRNA-complementary foreign genetic material. Of the two classes of CRISPR-Cas systems, class 2 Cas endonucleases, Cas9 (type II) and Cas12a/Cpf1 (type V) generate double-strand breaks (DSB) at the dsDNA targets. The development of two-component systems consisting of Cas9-single guide RNA (sgRNA) or Cas12a-crRNA and the programmable nature of the guide RNA (gRNA) have enabled precise genome editing in various animal and plant model systems. Despite the many advantages of this technology, a major shortcoming for its use in genome editing is the potential for off-target effects. Streptococcus pyogenes Cas9 (SpCas9) can bind and cleave sequences with partial complementarity to the gRNA, resulting in undesired genome editing. Several studies have been conducted to determine the specificity and off-target effects of Cas9 and Cas12a. However, most of these studies were conducted in eukaryotic settings which limits determining the native specificity and off-target effects of the Cas endonucleases. We investigated the native cleavage specificity of Cas9 and Cas12a by developing an in vitro target library cleavage assay coupled with high-throughput sequencing. The goal of the research was to understand the specificity and cleavage activities of natural and engineered Cas9 and Cas12a variants. We show that both Cas9 and Cas12a have higher mismatch tolerance that previously reported. Cas9 is highly promiscuous and can cleave target sequences with up to five mismatches while Cas12a can cleave target sequences with up to four mismatches. We also show that both Cas9 and Cas12a have sequence-dependent nicking activity where target sequences with mismatches are often nicked and complete DSB does not occur. This nicking activity of both Cas9 and Cas12a were dependent on the type of mismatch and position in the target. Engineered, high-fidelity Cas9 variants have higher nicking activity than the commonly used wild-type SpCas9. In vivo studies often look for DSB or indel formation, while nicks are unaccounted. This may explain the higher in vivo targeting specificity of engineered Cas9 variants and Cas12a. We further characterized an activated, sequence-independent nicking activity of Cas12a. After target cleavage, Cas12a is activated for non-specific nicking of dsDNA substrates similar to previously described trans, single-stranded DNA cleavage. Both these non-specific activities of Cas12a are comparably efficient in vitro. The non-specific nicking activity varies among Cas12a orthologs and with the target activator. Further, multiple nicks within the dsDNA substrate causes degradation. Phages evolve rapidly via mutations in their genomes and if the targeting specificity of the Cas9 and Cas12a is too stringent, phages can easily escape CRISPR-based immunity. The target sequence-dependent nicking and non-specific, activated nicking activities may be beneficial for bacterial immunity as they allow for targeting of mutated target sequences in phages or closely related phages. However, off-target nicking from these activities in genome editing studies could result in unwanted edits. Overall, our results highlight the specificity profiles of Cas9 and Cas12a and demonstrate nicking activities that were not fully characterized before and demonstrate previously unknown nicking activities that may have implications for both CRISPR-based immunity and genome editing