Genome engineering using CRISPR-Cas9 nucleases relies on the creation of double-strand breaks (DSBs) in target DNA, which are repaired via non-homologous end joining (NHEJ) or homology-directed repair (HDR). However, DSBs are cytotoxic and can lead to undesirable genomic rearrangements, including large deletions, and even cell death. To mitigate these risks, base editing and prime editing systems have been developed. These systems utilize Cas9 nickases (nCas9s) instead of Cas9 nucleases, avoiding the generation of DSBs. nCas9s are created by mutating key catalytic amino acid residues in the RuvC or HNH domain of *S. pyogenes* Cas9 (SpCas9). nCas9 (D10A) nicks the target strand, while nCas9 (H840A) nicks the non-target strand. Base editors fuse deaminases to dead Cas9 (dCas9), often replacing dCas9 with the more efficient nCas9 (D10A). Prime editors (PEs), which enable various types of genome editing, consist of nCas9 (H840A) fused to Moloney murine leukemia virus (M-MLV) reverse transcriptase (RT). They are guided by pegRNAs containing a primer-binding site (PBS) and an RT template. While PEs offer advantages, they exhibit relatively high frequencies of unwanted indels, reducing the purity of the desired edits. This research aimed to investigate the off-target effects of nCas9 nickases, particularly nCas9 (H840A), and to improve the fidelity of prime editing by engineering the nCas9 component.

The use of CRISPR-Cas9 technology for genome editing has revolutionized biological research and holds immense promise for therapeutic applications. However, the inherent off-target effects and potential for large deletions and chromosomal rearrangements associated with the generation of double-strand DNA breaks have prompted efforts to develop improved genome editing techniques that minimize these issues. Base editing, employing deaminases fused to catalytically inactive Cas9, allows for single base changes without DSBs. Prime editing, which combines a nickase Cas9 with reverse transcriptase, offers a more versatile approach capable of various types of edits. However, both base and prime editing systems exhibit off-target effects that require improvement. Previous studies have investigated the specificity of different Cas9 variants and attempted to enhance their accuracy. The current study builds on this prior work by focusing on the characterization and engineering of Cas9 nickases within the context of prime editing to reduce unwanted indel formation.

The researchers employed several methods to characterize the activity and specificity of Cas9 nickases and to engineer improved versions for use in prime editing. In vitro plasmid digestion assays were performed to assess the cleavage patterns of wild-type Cas9, nCas9 (D10A), and nCas9 (H840A) using agarose gel electrophoresis. Digenome-seq, a genome-wide method based on whole genome sequencing, was used to identify off-target cleavage sites of the nCas9 variants. Additional mutagenesis was performed on nCas9 (H840A) to create variants with reduced DSB formation. The engineered nCas9 variants were then incorporated into PE2 and PE3 prime editor systems. HEK293T, HeLa, and K562 cells were transfected or nucleofected with plasmids encoding the engineered prime editors and guide RNAs. Targeted deep sequencing was used to quantify the frequencies of intended edits and unwanted indels at multiple genomic loci. Statistical analyses (Student's t-test) were used to compare the performance of the different prime editor versions. The study also incorporated engineered pegRNAs (epegRNAs) to evaluate whether the improvements in nCas9 could further enhance the efficiency and purity of prime editing.

The study revealed that nCas9 (H840A), which is commonly used in prime editing, is not a true nickase and can induce DSBs in vitro and in vivo. Digenome-seq confirmed that nCas9 (H840A) produced significantly more DSBs genome-wide compared to nCas9 (D10A). Introducing the N863A mutation into nCas9 (H840A) significantly reduced its DSB-inducing activity. Further mutagenesis, incorporating N854A mutation in addition to H840A and N863A, resulted in nCas9 variants with minimal DSB formation and greatly reduced unwanted indels in various prime editing systems. Specifically, nCas9 (H840A + N854A) incorporated into PE2 and PE3 systems dramatically increased the frequency of correct edits while having minimal impact on the unwanted indels. When combined with engineered pegRNAs (epegRNAs), nCas9 (H840A + N854A) within the ePE3 system displayed the highest editing purity (up to 96.5%) across various edit types (substitutions, insertions, and deletions) in HEK293T, HeLa, and K562 cell lines.

This study addresses the limitations of the commonly used nCas9 (H840A) in prime editing by demonstrating its propensity for DSB formation, a major source of unwanted indels. The successful engineering of novel nCas9 variants that exclusively generate nicks eliminates this problem and substantially improves the fidelity of prime editing. The findings highlight the importance of careful characterization and optimization of Cas9 variants to enhance the precision and efficacy of genome editing tools. The use of Digenome-seq provided a comprehensive assessment of genome-wide off-target effects, confirming the effectiveness of the engineering strategy. The improvement in editing purity achieved with the engineered nCas9 variants is critical for therapeutic applications, where minimizing off-target effects is crucial for safety. Future studies should investigate whether combining the improved PE variants with other advanced strategies, such as optimized pegRNAs, can further enhance editing efficiency and purity.

This study demonstrated that the commonly used nCas9 (H840A) in prime editing can create unwanted DSBs, leading to a high frequency of indels. Through structure-guided mutagenesis, the authors generated improved nCas9 variants that minimize DSB formation and significantly reduce unwanted indels in PE2, PE3, and ePE3 systems. The resultant prime editing systems, particularly those incorporating nCas9 (H840A + N854A), exhibit a substantially higher purity of intended edits, bringing prime editing closer to clinical applications. Future research could focus on exploring other mutations in the HNH domain and combining these enhanced PEs with improved pegRNAs for further optimization.

The study primarily focused on in vitro and in vivo experiments using specific cell lines. Further studies are necessary to validate the findings in a broader range of cell types and organisms. While the engineered nCas9 variants showed significant improvements, there may be other limitations related to delivery efficiency, accessibility of target sites, and potential for other off-target effects not detected by the assays used. Additional long-term studies are warranted to assess potential long-term effects of these modifications.