EXPERT expands prime editing efficiency and range of large fragment edits

Abstract

Prime editing systems (PEs) hold great promise in modern biotechnology. However, their editing range is limited as PEs can only modify the downstream sequences of the pegRNA nick. Here, we report the development of the extended prime editor system (EXPERT) to overcome this limitation by using an extended pegRNA (ext-pegRNA) with modified 3' extension, and an additional sgRNA (ups-sgRNA) targeting the upstream region of the ext-pegRNA. We demonstrate that EXPERT can efficiently perform editing on both sides of the ext-pegRNA nick, a task that is unattainable by canonical PEs. EXPERT exhibits prominent capacity in replacing sequences up to 88 base pairs and inserting sequences up to 100 base pairs within the upstream region of the ext-pegRNA nick. Compared to canonical PEs such as PE2, the utilization of the EXPERT strategy significantly enhances the editing efficiency for large fragment edits with an average improvement of 3.12-fold, up to 122.1 times higher. Safety wise, the use of ups-sgRNA does not increase the rates of undesirable insertions and deletions (indels), as the two nicks are on the same strand. Moreover, we do not observe increased off-target editing rates genome-wide. Our work introduces EXPERT as a PE tool with significant potential in life sciences.

Fig. 1. EXPERT expands the editing range of canonical PEs.

a Schematics of canonical PEs and EXPERT. The deep blue thick-lined area (downstream region of the pegRNA nick) represents the editable region for canonical PEs and EXPERT. On the other hand, the green thick-lined area (upstream region of the pegRNA nick) is uneditable by canonical PEs, whereas the region marked by the light blue thick-lined area can be edited by EXPERT. nCas9, Cas9 nickase (H840A); RT, reverse transcriptase. b Frequencies of intended edits and indels introduced by EXPERT for two different edits, which canonical PEs cannot perform. PBS, primer binding site. RTT, reverse transcriptase template. Bars represent the mean of n = 3 independent biological replicates. Data are presented as mean ± s.d. c EXPERT performs simultaneous editing on both sides regions of the pegRNA nick at the VEGFA_1 site. Bars represent the mean of n = 3 independent biological replicates. Data are presented as mean ± s.d. All sequencing data were collected from transfection-positive cells. Source data are provided as a Source Data file.

Fig. 2. Enhancing EXPERT editing efficiency through optimizing the distance between ups-sgRNA and ext-pegRNA, introducing mismatches on ext-pegRNA, or optimizing PBS length.

a Schematic diagram illustrating the modification of the stop codon in the 293T-reporter cell line to restore mCherry function under DCNs of varying sizes (top). Frequencies of edits introduced by EXPERT under DCNs of varying sizes were quantified using flow cytometry (bottom). Bars represent the mean of n = 3 independent biological replicates. b Frequencies of intended edits and indels introduced by EXPERT under DCNs of varying sizes were quantified at VEGFA_1 and HEK4_1 sites. Bars represent the mean of n = 3 independent biological replicates. Data are presented as mean ± s.d. c Diagram of the hybridization of the 3′ Flap of the original DNA strand to complementary ext-pegRNA. The hybridization can hinder the reverse transcription process of RT enzymes, leading to editing failure. The bold red line area denotes the editing region. "m" indicates the distance between the editing region and ext-pegRNA nick. d The pattern of introducing mismatch on ext-pegRNA region that hybridizes with the 3′ Flap of the original DNA strand. e Frequencies of intended edits and indels introduced by EXPERT under different numbers of mismatches on the ext-pegRNA region that hybridizes with the 3′ Flap. Bars represent the mean of n = 3 independent biological replicates. Data are presented as mean ± s.d. f The product purity (intended edits: indels ratio) introduced by EXPERT under different numbers of mismatches on the ext-pegRNA region. Bars represent the mean of n = 3 independent biological replicates. Data are presented as mean ± s.d. g The impact of PBS length on editing efficiency of EXPERT. Bars represent the mean of n = 3 independent biological replicates. Data are presented as mean ± s.d. All sequencing data were collected from transfection-positive cells. Source data are provided as a Source Data file.

Fig. 3. EXPERT enables the insertion and replacement of large DNA fragments at the upstream region of the ext-pegRNA nick.

a Schematic diagram for deletion of the sequence between two nicks and insertion of fragments of different lengths. b Frequencies of fragment insertions of varying lengths (5, 10, 20, 30, 50, 80, 100 bp) and indels introduced by EXPERT at the EMX1 and VEGFA_1 sites, respectively. Bars represent the mean of n = 3 independent biological replicates. Data are presented as mean ± s.d. c Schematics of EXPERT and EXPERT + Helper gRNA. In comparison to EXPERT, EXPERT + Helper gRNA incorporates an additional sgRNA in the middle region between the ups-sgRNA and ext-pegRNA, referred to as the Helper gRNA. d Frequencies of edits introduced by EXPERT and EXPERT + Helper gRNA were quantified using flow cytometry. Bars represent the mean of n = 3 independent biological replicates. Data are presented as mean ± s.d. The P value was calculated using a two-tailed t-test, and no adjustments were made for multiple comparisons. P_{reporter replace 71 bp} = 0.0002, P_{reporter replace 96 bp} = 0.00005; ***P < 0.001; ****P < 0.0001. e Frequencies of intended edits, indels, and product purity introduced by EXPERT and EXPERT + Helper gRNA. Bars represent the mean of n = 3 independent biological replicates. Data are presented as mean ± s.d. All sequencing data were collected from transfection-positive cells. Source data are provided as a Source Data file.

Fig. 4. The EXPERT strategy enhances editing efficiency with high product purity and low off-target effects.

a Frequencies of intended edits and indels introduced by PE2 and EXPERT at multiple loci. Additional mismatches were introduced in the insertion-type edits. Bars represent the mean of n = 3 independent biological replicates. Data are presented as mean ± s.d. b Statistical analysis of normalized editing frequencies, setting the frequencies induced by PE2 as 1. n = 19 editing from independent experiments shown in (a) and supplementary Fig. 9. Data are presented as mean ± s.d. c Statistical analysis of normalized frequencies of indels, setting the frequencies induced by PE2 as 1. n = 19 editing from independent experiments shown in (a) and supplementary Fig. 9. Data are presented as mean ± s.d. d Frequencies of intended edits, indels, and relative product purity introduced by twinPE and EXPERT at different loci. Bars represent the mean of n = 3 independent biological replicates. Data are presented as mean ± s.d. e Numbers of genome-wide base substitutions. Bars represent the mean of n = 3 independent biological replicates. Data are presented as mean ± s.d. The P value was calculated using a two-tailed t-test, and no adjustments were made for multiple comparisons. f Numbers of genome-wide indels. Bars represent the mean of n = 3 independent biological replicates. Data are presented as mean ± s.d. The P value was calculated using a two-tailed t-test, and no adjustments were made for multiple comparisons. All sequencing data were collected from transfection-positive cells. Source data are provided as a Source Data file.

Fig. 5. EXPERT strategy can enhance efficiency across various PE systems and can be applied to different cell types from multiple species.

a Frequencies of intended edits and indels introduced by PE2max, PE3max, PE4max, PE5max and their corresponding EXPERTmax systems. Additional mismatches were introduced in the insertion-type edits. Bars represent the mean of n = 3 independent biological replicates. Data are presented as mean ± s.d. b Frequencies of intended edits and indels introduced by PE2max and EXPERTmax in K562 cells. Additional mismatches were introduced in the insertion-type edits. Bars represent the mean of n = 3 independent biological replicates. Data are presented as mean ± s.d. c Frequencies of intended edits and indels introduced by PE2max and EXPERTmax in Jurkat cells. Additional mismatches were introduced in the insertion-type edits. Bars represent the mean of n = 3 independent biological replicates. Data are presented as mean ± s.d. d Frequencies of intended edits and indels introduced by PE2max and EXPERTmax in Hela cells. Additional mismatches were introduced in the insertion-type edits. Bars represent the mean of n = 3 independent biological replicates. Data are presented as mean ± s.d. e Frequencies of intended edits, indels, and product purity introduced by PE2max and EXPERTmax in N2a cells. Bars represent the mean of n = 3 independent biological replicates. Data are presented as mean ± s.d. f Frequencies of intended edits, indels, and product purity introduced by PE2max and EXPERTmax in PFF cells. Additional mismatches were introduced in the insertion-type edits. Bars represent the mean of n = 3 independent biological replicates. Data are presented as mean ± s.d. g Schematic diagram of complex mutations in CFTR exon 4. The intended edits that carried four mutations were performed in HEK293T cells using PE2max, PE3max, EXPERTmax, and EXPERTmax + nicking sgRNA, respectively. h Frequencies of intended edits, indels, and product purity introduced by PE2max, PE3max, EXPERTmax, and EXPERTmax + nicking sgRNA, respectively. Bars represent the mean of n = 3 independent biological replicates. Data are presented as mean ± s.d. All sequencing data were collected from transfection-positive cells. Source data are provided as a Source Data file.