Development of a versatile nuclease prime editor with upgraded precision

Abstract

The applicability of nuclease-based form of prime editor (PEn) has been hindered by its complexed editing outcomes. A chemical inhibitor against DNA-PK, which mediates the nonhomologous end joining (NHEJ) pathway, was recently shown to promote precise insertions by PEn. Nevertheless, the intrinsic issues of specificity and toxicity for such a chemical approach necessitate development of alternative strategies. Here, we find that co-introduction of PEn and a NHEJ-restraining, 53BP1-inhibitory ubiquitin variant potently drives precise edits via mitigation of unintended edits, framing a high-activity editing platform (uPEn) apparently complementing the canonical PE. Further developments involve exploring the effective configuration of a homologous region-containing pegRNA (HR-pegRNA). Overall, uPEn can empower high-efficiency installation of insertions (38%), deletions (43%) and replacements (52%) in HEK293T cells. When compared with PE3/5max, uPEn demonstrates superior activities for typically refractory base substitutions, and for small-block edits. Collectively, this work establishes a highly efficient PE platform with broad application potential.

Fig. 1. Design and optimization of uPEn.

a Schematics of PE-nuclease induced DNA repair pathways and the impact by 53BP1-inhibitory ubiquitin variants. Upon Cas9-mediated double-stranded DNA cleavage, the 3′-extension of the pegRNA directs the reverse transcriptase (RT) domain to generate a 3′-overhang structure with the upstream end of the break. Within this overhang, the sequence for insertion is marked in blue, whereas the ensuing segment having the potential to base-pair with the downstream DNA end is marked in orange. Hypothetically, the 53BP-1-inhibory ubiquitin variants could mitigate the non-homologous end-joining repair for such an editing intermediate, therefore potentiating precise insertional edits. The blue segment in the precisely repaired DNA corresponds to that within the 3′ overhang intermediate above. b The frequency of all RT-driven edits, accurate edits and indels engaged by PE-nuclease following the co-transfection of Ub (wt), G08 and G08 (I44A) ubiquitin variants at LSP1, SEC61B and RUNX1 site in HEK293T cells. All RT-driven edits consist of desired edits and imprecise edits containing the programmed sequence. Accurate edits represent desired edits. Values and error bars reflect the means and standard deviation (s.d.) of three biological replicates. c The ratio of accurate edits relative to all RT-driven edits in experimental groups with different ubiquitin variants at LSP1, SEC61B and RUNX1 sites. The darker color indicates a higher the percentage of accurate edits. d Schematics of the configuration of three versions of uPEn editing system. PEn, PE-nuclease (Cas9 in grey and RT in light brown). uPEn1, PE-nuclease supplemented with an UbvG08 (I44A) ubiquitin variant that is indicated by the shape in blue. uPEn2, PE-nuclease fused with UbvG08 (I44A) by a 34aa linker (green line). uPEn3, PE-nuclease linked with UbvG08 (I44A) by P2A (purple line). Note that within the illustration for pegRNA, the orange, blue and pink segments represent the spacer, the RTT and the PBS segements, respectively. e The editing efficiency of 34-bp fragment insertion at FANCF, UBE3A, SHANK3-1 and SHANK3-2 sites in HEK293T cells by PEn, uPEn1, uPEn2, and uPEn3. Note that the “Unintended edits” values include those of the direct indels and the imprecise edits containing the programmed sequences (also used in the ensuing figures). Values and error bars reflect the means and s.d. of three biological replicates. f Prime editing efficiencies for insertions by the uPEn systems normalized to the efficiency by PEn at the sites in (e). The mean editing frequency induced by PEn for each locus was set to 1, and other samples were normalized correspondingly. The center line shows medians of all data points and the box limits correspond to the upper the lower quartiles, while the whiskers extend to the largest and smallest values. n = 4 (sites) for each group. The P values (directly marked on the graph) were determined by two-tailed one-sample Student’s t-tests. Source data are provided as a Source Data file.

Fig. 2. Impacts by the size of the homologous region in pegRNAs on uPEn-dependent editing.

a Schematics of homologous region-containing pegRNAs (HR-pegRNA) with different sizes of the HR (orange). PR, programmed region (blue). Red arrow indicates the position of uPEn-induced DSB. b uPEn (in the uPEn3 format, same throughout this figure) editing efficiencies of using pegRNAs with differently sized HR at EMX1, FAM17A and PRNP sites in HEK293T cells. Values and error bars reflect the mean and s.d. of three biological replicates. c The pattern of prime editing efficiencies with increasing sizes of HR at three sites in (b). Similarly, the values and error bars reflect the mean and s.d. of three biological replicates. d Insertional edits of different sequences with uPEn at multiple sites in HEK293T cells. Values and error bars reflect the mean and s.d. of three biological replicates. Source data are provided as a Source Data file.

Fig. 3. The uPEn mediated more efficient prime editing for insertions, deletions and sequence replacements.

a Schematic diagram illustrating designs and various types of edits (insertions, deletions or replacements) generated by the uPEn system. PR, programmed region (blue); HR, homologous region (orange). Gray fill indicates the base-pairing with the genomic sequence. b Deletional edits of different sequences with uPEn (in the uPEn3 format, same throughout this figure) at multiple sites in HEK293T cells. Values and error bars reflect the mean and s.d. of three biological replicates. c Sequence replacement type of edits with uPEn at multiple sites in HEK293T cells. Values and error bars reflect the mean and s.d. of three biological replicates. d Summary of the insertional (n = 6 sites), deletional (n = 6 sites) and sequence replacement type of (n = 5 sites) editing efficiencies by uPEn in (b, c and 2d). The over efficiencies for all three types of editing (n = 17 sites) are also shown. The center line shows medians of all data points and the box limits correspond to the upper the lower quartiles, while the whiskers extend to the largest and smallest values. Source data are provided as a Source Data file.

Fig. 4. The uPEn mediated efficient base conversions at some largely PE-intractable sites.

a Comparison of base conversion efficiencies and indels induced by PE3max and uPEn (in the uPEn3 format, same throughout this figure) at ten endogenous sites in HEK293T cells. Values and error bars reflect the mean and s.d. of three biological replicates. P values were determined (for precise editing efficiencies between PE3max and uPEn at individual sites) by two-tailed Student’s t-tests. The asterisks on the graph are used to indicate the ranges (**P < 0.005, ***P < 0.0005). The respective P values in accordance to their left-to-right order are: 0.002209, 0.001554, 0.003482, 0.000163, 0.000001, 0.000002, 0.000001, 0.000015, 0.001491, 0.000003. b Summary of the fold change in uPEn efficiency normalized to PE3max at the same target sites in (a). The mean editing frequency induced by PE3max for each locus was set to 1, and other samples were normalized correspondingly. The center line shows medians of all data points and the box limits correspond to the upper the lower quartiles, while the whiskers extend to the largest and smallest values. n = 10 (sites) for each group. The P value (directly marked on the graph) was determined by a two-tailed one-sample Student’s t-test. c Comparison of base conversion efficiencies and indels induced by PE5max and uPEn at six endogenous sites in U2OS cells. Values and error bars reflect the mean and s.d. of three biological replicates. P values were determined by two-tailed Student’s t-tests. The asterisks on the graph are used to indicate the ranges (*P < 0.05, **P < 0.005, ***P < 0.0005). The respective P values in accordance to their left-to-right order are: 0.000054, 0.000025, 0.00265, 0.010059, 0.000194, 0.003916. d Summary of the fold change in uPEn efficiency normalized to PE5max at the same target sites in (c). The mean editing frequency induced by PE5max for each locus was set to 1, and other samples were normalized correspondingly. The box plot was generated with the same settings as in (b). n = 6 (sites) for each group. The P value (directly marked on the graph) was determined by a two-tailed one-sample Student’s t-test. e Summary of base conversion efficiencies of uPEn normalized to PE5max in HeLa cells. The box plot was generated with the same settings as in (b). n = 6 (sites) for each group. The P value (directly marked on the graph) was calculated by a two-tailed one-sample Student’s t-test. Source data are provided as a Source Data file.

Fig. 5. Benchmarking the performances of uPEn for insertion, deletion and replacement.

a–d Comparison of efficiencies for insertion, deletion, replacement by PE2max, PE3max, PE5maxP and uPEn at CDKL5 (a), CXCR4 (b), DNMT1 (c) and RNF2 (d) loci in HEK293T cells. The levels of corresponding unintended edits are shown in grey bars. Values and error bars reflect the mean and s.d. of three biological replicates. Source data are provided as a Source Data file.