Directed evolution of CRISPR-Cas9 to increase its specificity

The use of CRISPR-Cas9 as a therapeutic reagent is hampered by its off-target effects. Although rationally designed S. pyogenes Cas9 (SpCas9) variants that display higher specificities than the wild-type SpCas9 protein are available, these attenuated Cas9 variants are often poorly efficient in human cells. Here, we develop a directed evolution approach in E. coli to obtain Sniper-Cas9, which shows high specificities without killing on-target activities in human cells. Unlike other engineered Cas9 variants, Sniper-Cas9 shows WT-level on-target activities with extended or truncated sgRNAs with further reduced off-target activities and works well in a preassembled ribonucleoprotein (RNP) format to allow DNA-free genome editing.

The successful integration of EMX1 into the genomic DNA of the BW25141 strain was confirmed by colony PCR amplification of the Tn7 attachment site in the glmS gene. A 678 bp product was observed in the absence of an EMX1 insertion whereas a 1404 bp product was observed in the presence of an EMX1 insertion. Eight different colonies were picked; amplification of DNA from #1 and #6 resulted in single bands of 1404 bp. #1 was selected for plasmid removal by incubation at 42℃ overnight.

1) DMD
On-target and off-target activities of WT-Cas9 and the pooled library from two different screenings at a target in the human DMD gene in HEK293T cells.
On-target and off-target activities of three different Cas9 variants (Clone-1, 2 and 3) at a target in the human DMD gene in HEK293T cells.

2) EMX1
On-target and off-target activities of WT-Cas9 and the pooled library from two different screenings at a target in the human EMX1 gene in HEK293T cells.
On-target and off-target activities of three different Cas9 variants (Clone-1, 2 and 3) at a target in the human EMX1 gene in HEK293T cells. Supplementary Figure 5.
The amino acid mutations in Clone-1, Clone-2 and Clone-3 are listed below.
Various mutant forms of Cas9 were made by introducing site directed mutations into WT-Cas9. The indel frequencies were measured in HEK293T cells at the DMD and EMX1 targets. Indel % Indel % sgRNAs were transcribed by the U6 promoter and thus must start with a guanine at the 5' end, which may (GX19) or may not (gX19 or gX20) complement the on-target sequences. We tested six sgRNAs with mismatched 5' guanines (gX19) and six with matching 5' guanines (GX19). On-target and off-target activities of Cas9 variants compared to WT-Cas9 using sgRNAs variable lengths targeting rest of 10 target sites out of 12 targets tested. Results for FANCF01 and AAVS were shown in Figure 2a as examples. Specificity ratios were determined by dividing indel frequencies at on-target sites by those at respective off-target sites. sgRNAs with a matched guanine at the 5' terminus (GX18 or GX19) and those with a mismatched guanine (gX17, gX18, gX19 or gX20) are indicated. N/A (Not Available): Specificity ratios were not calculated when on-target activities were less than 70% of the WT plus 20mer guide sequence.
Error bars indicate s.e.m. (n=3) Western blot was performed to check the protein expression level of Cas9 variants in HEK293T cells. Cas9 proteins were detected by HA tag. GAPDH were control.
On-target and off-target activities were measured in HeLa cells using the same sites tested in HEK293T cells. Four different targets were tested. Indel frequencies were measured using targeted deep sequencing. On-target Off-target Tolerance of Cas9 variants for sgRNAs containing mismatches relative to HBB02, VEGFA and FANCF01 targets. HBB02 was targeted with gX19 sgRNAs and VEGFA was targeted with gX20 sgRNAs. FANCF01 was targeted with 2bp mismatched GX19 sgRNAs. Indel frequencies were measured using targeted deep sequencing. Error bars indicate s.e.m. (n = 3) On-target and off-target activities of xCas9-3.7 compared to WT-Cas9 and Sniper-Cas9 using sgRNAs targeting 10 target sites. Specificity ratios were determined by dividing indel frequencies at on-target sites by those at respective off-target sites. sgRNAs with a matched guanine at the 5' terminus (GX19) and those with a mismatched guanine (gX19 or gX20) are indicated. N/A (Not Available): Specificity ratios were not calculated when on-target activities were less than 70% of the WT plus 20mer guide sequence. Error bars indicate s.e.m. (n=3).

Supplementary Figure 13.
Validation of 10 candidate off-target sites for WT-Cas9 and Sniper-Cas9 targeting AAVS, DMD, FANCF01 and HBB04. Candidate sites had received the top rank of cleavage score in Digenome-sequencing experiments and included fewer than 5 mismatches relative to the ontarget site.