Enhancement of prime editing via xrRNA motif-joined pegRNA

Abstract

The prime editors (PEs) have shown great promise for precise genome modification. However, their suboptimal efficiencies present a significant technical challenge. Here, by appending a viral exoribonuclease-resistant RNA motif (xrRNA) to the 3'-extended portion of pegRNAs for their increased resistance against degradation, we develop an upgraded PE platform (xrPE) with substantially enhanced editing efficiencies in multiple cell lines. A pan-target average enhancement of up to 3.1-, 4.5- and 2.5-fold in given cell types is observed for base conversions, small deletions, and small insertions, respectively. Additionally, xrPE exhibits comparable edit:indel ratios and similarly minimal off-target editing as the canonical PE3. Of note, parallel comparison of xrPE to the most recently developed epegRNA-based PE system shows their largely equivalent editing performances. Our study establishes a highly adaptable platform of improved PE that shall have broad implications.

Fig. 1. The xrRNA-joined pegRNAs show enhanced prime editing activities toward a reporter.

a The schematic secondary structure of a representative xrRNA motif. Some long-distance interactions (highlighted by red dotted lines) contribute to the formation of a stable knot-like structure. b Illustration for the fluorescent prime editing reporter system. The translation of EGFP sequence as part of a mRuby-led fusion protein is prevented by a stop codon (TAG, red). Prime editing-mediated of A-to-G edit would allow the expression of mRuby-EGFP fusion protein. The spacer sequence and the PAM for prime editing are underlined. c. The xrRNA motifs from five different viruses: Murray Valley encephalitis (MVE), West Nile virus (WNV), Zika, Dengue (Dengue), and Yellow Fever (YF)) were appended to the 3′ end of pegRNAs that targets the reporter. In results shown in c–f, HEK293T cells were co-transfected with plasmids for PE2, WT or modified pegRNA, and the reporter. EGFP⁺ cells were observed under a fluorescent microscope. Transfection of PE2, a non-targeting pegRNA and the reporter served as the negative control, whereas in the positive control the reporter plasmid was replaced with one encoding a constantly expressing mRuby-EGFP fusion protein. Scale bars, 250 µm. d Following prime editing using WT pegRNA and xr-pegRNAs, the frequencies (%) of EGFP⁺ (relative to mRuby⁺) were measured by flow cytometry. e Following prime editing using WT pegRNA and xr-pegRNAs, the expression of EGFP was determined by Western blot. f The reporter-targeted editing efficiencies were determined by deep-sequencing of DNA prepared from mRuby⁺ cells. The editing frequencies induced by PE with WT pegRNA were set as 100%. In quantitation shown in d and f, data are presented as mean values ±SD, n = 3 biological replicates. Two-tailed Student’s t tests (one-sample test for f) were performed (P values are marked on the graphs, n.s. not significant). The P values [n.s.] not marked on d and f are 0.09 and 0.20, respectively. Source data are provided as a Source Data file.

Fig. 2. The xr-pegRNA enhances prime editing of base conversions at various sites within genomic context.

a HEK293T cells were co-transfected with plasmids for PE2 and WT pegRNAs or different xr-pegRNAs targeting indicated (6) genomic loci for base conversions. Following isolation of genomic DNA from transfected cells (sorted), correct editing rates at each site were determined by deep-sequencing (mean ± SD, n = 3 biological replicates). Reads that only contain the intended edits were counted. b. Results in a is further analyzed by considering editing at all sites (n = 6 sites) as a whole. The editing frequencies induced by PE2 with WT pegRNA were set as 100%. c. Experiments were carried out similar to (a), except that a PE3 strategy was used (mean ± SD, n = 3 biological replicates). d Results in c is further analyzed by considering editing at all sites (n = 6 sites) as a whole. The editing frequencies induced by PE3 with WT pegRNA were set as 100%. In the violin plots shown in b and d, each point represents the averaged editing activity at the particular site. The thicker dotted line shows the medians of all data points, while the thinner dotted lines correspond to quartiles (1st and 3rd). Two-tailed one-sample Student’s t tests were performed. The P values are marked on the graphs (n.s. not significant). Source data are provided as a Source Data file.

Fig. 3. The xrPE shows enhanced performance for base conversions in multiple cell types.

a An illustration for the xrPE platform. The joining of an xrRNA motif (Zika) to the 3′ end of pegRNA is shown. A fusion protein of Cas9 H840A nickase and a reverse transcriptase (Moloney Murine Leukemia Virus, M-MLV) is guided by the modified pegRNA to a DNA target. The yellow marks within xrRNA-joined pegRNA indicate an alternative C-G base pair replacing a U-A in the main scaffold, to potentially reduce premature termination. The prime editor nicks the DNA and reverse transcribes using the 3′-extended portion of pegRNA as the template. This is followed by 5′ flap removal and ligation to complete editing on one strand. When supplying another sgRNA to nick the non-edited strand, the cellular DNA repair mechanisms tend to install the desired edit into the genome. b HEK293T cells were transfected with plasmids for canonical PE3 or xrPE for base conversion at 9 individual sites as indicated. Correct editing efficiencies were determined by deep-sequencing (mean ± SD, n = 3 biological replicates). For targets same as those in Fig. 2c, a consistent pattern of activity enhancements is noted. Gray bars next to those for PE3 (red) and xrPE (blue) indicate the indel frequencies associated with each tool. c Results in b and Supplementary Fig. 8a are further analyzed by considering editing at all sites (n = 15 sites) as a whole. The editing frequencies induced by canonical PE3 were set as 100%. d. The experiment similar to b was carried out in N2a cells (base conversions at 9 individual sites). The rates for correct editing and indel formation are shown (mean ± SD, n = 3 biological replicates). e Results in d were further analyzed by considering editing at all sites (n = 9 sites) as a whole. The editing frequencies induced by canonical PE3 were set as 100%. Multiple t tests (two-tailed) were performed in data from b, d. Discoveries were determined using the two-stage linear step-up procedure of Benjamini, Krieger, and Yekutieli, with Q = 1%. When discoveries are made (9/9 in b and 9/9 in d), the exact P values (unadjusted) are shown on the graphs. In the box plots shown in c, e, each data point represents the averaged editing activity at the particular site. 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. Two-tailed one-sample Student’s t tests were performed (with P values marked). Source data are provided as a Source Data file.

Fig. 4. The xrPE shows enhanced performance for precisely introducing small deletions and small insertions in multiple cell types.

a HEK293T cells were transfected with plasmids for canonical PE3 or xrPE targeting 6 individual sites for 3-bp deletions and 3-bp insertions, separately. Correct editing efficiencies were determined by deep-sequencing (mean ± SD, n = 3 biological replicates). Gray bars next to those for PE3 (red) and xrPE (blue) indicate the indel frequencies associated with each tool. The same set of indel data from untreated cells (background) were presented for deletions and insertions. b Results in a is further analyzed by considering editing at all sites (n = 6 sites for deletions and insertions, respectively) as a whole. The editing frequencies induced by canonical PE3 were set as 100%. c. The experiment similar to a was carried out in N2a cells (at 6 individual sites for 3-bp deletions and 3-bp insertions, respectively). The rates for correct editing and indel formation are shown (mean ± SD, n = 3 biological replicates). The same set of indel data from untreated cells (background) were presented for deletions and insertions. d Results in c were further analyzed by considering editing at all sites (n = 6 sites for deletions and insertions, respectively) as a whole. The editing frequencies induced by canonical PE3 were set as 100%. Multiple t tests (two-tailed) were performed in data from a, c. Discoveries were determined using the two-stage linear step-up procedure of Benjamini, Krieger, and Yekutieli, with Q = 1%. When discoveries are made (10/12 in a and 12/12 in c), the exact P values (unadjusted) are shown on the graphs. Otherwise, the comparisons are marked by n.s. not significant, where the corresponding P values are 0.315 (−3 bp in CTLA4) and 0.062 (−3 bp in FANCF), respectively. In the box plots shown in b, d, each data point represents the averaged editing activity at the particular site. 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. Two-tailed one-sample Student’s t tests were performed (with P values marked). Source data are provided as a Source Data file.

Fig. 5. The 3′ xrRNA stabilizes xr-pegRNA to enhance prime editing, resulting in comparable performance as the latest epegRNA-based strategy.

a The stability of in vitro-transcribed WT pegRNAs or the corresponding xr-pegRNAs upon exposure to HEK293T nuclear lysates. A representative agarose gel with samples from biological triplicates is shown. The sizes of the products (nt) are marked on the side. The levels for untreated WT pegRNAs or xr-pegRNAs were considered as 100%. Data presented are from quantitation of band intensity from 3 biological replicates (mean ± SD). The amounts of remaining pegRNAs or xr-pegRNAs were compared (two-tailed Student’s t tests, with P values marked). b Comparison of PE intermediates generated by PE2 with either WT pegRNAs or xr-pegRNAs at EMX1 site in HEK293T cells. The black dotted line represents the end of the full-length RT template (18 nt). In the histogram, the x axis corresponds to the sizes of the 3′ flaps, with the first base downstream of the PE2-induced nick denoted as position +1, while the y axis represents the relative abundance of the reads (percentage of all reads). The bottom box contains pie charts showing percentages of reads with (red/blue) or without (gray) intended edits. The data presented are calculated from an average of three independent biological replicates. c The WT pegRNA, xr-pegRNA, and two mutant xr-pegRNAs bearing either U3C or C21G mutations at the xrRNA domain were used in a PE3 context for base conversion or insertion at the EMX1 sites in HEK293T cells. The editing efficiencies were determined (mean ± SD, n = 3 biological replicates) and compared (two-tailed Student’s t tests, with P values marked). d The efficiencies by PE with xr-pegRNA (xrPE), or epegRNA (containing tevopreQ1 motif and Cr772 scaffold) for nine different genetic modifications attempted in our earlier experiments. Gray bars next to the red/blue-colored bars indicates the indel frequencies in association with a particular experiment group, where the editing efficiency is shown in red or blue (mean ± SD, n = 3 biological replicates). Source data are provided as a Source Data file.