Screening of epegRNAs and ngRNAs targeting the HBG1/2 promoters in K562 cells. (A) Schematic representation of the β-globin locus on chromosome (chr) 11, including the HBG1 and HBG2 genes and their promoters (prom) (top). HPFH or HPFH-like mutations (bold) disrupt LRF or BCL11A repressor BSs (red boxes) or generate KLF1, TAL1, or GATA1 activator (yellow boxes) BSs. PAM-disrupting mutation (PAMm) and BCL11A BS-deletion (Δ2B: ΔCC, Δ5B: ΔTGACC) present in pegRNA1/epegRNA1 and pegRNA10/epegRNA10.1/epegRNA10.2 are displayed below the promoter sequence. Schematic representation of epegRNA5 (bottom left) and epegRNA10.1 (bottom right) targeting the HBG1/2 promoter region. The epegRNA is composed (5′-3′) of a spacer, a scaffold, a primer binding site (PBS), a reverse transcription template (RTT) containing the edits (ie, base substitutions and deletion indicated in colors and black), and the tevopreQ₁ motif. Upon annealing of the spacer to the target DNA, the Cas9n induces a DNA single-strand break (ie, nick, arrowhead) liberating the opposite strand (3′ flap), which anneals to the PBS of the epegRNA and primes the reverse transcription of the RTT. The interconversion of the original 5′ flap by the newly synthesized, edited 3′ flap and the degradation of the 5′ flap are essential steps to install the DEs into the target region. (B-C) Frequency (%) of InDels generated by transfecting plasmids coding for (B) pegRNAs or (C) ngRNAs and Cas9 (pegRNA1 and ngRNA1.1) or Cas9-SpRY (pegRNA2 to pegRNA9 and ngRNA5.1 to ngRNA5.9) nucleases. Although pegRNA1 was designed to install KLF1, TAL1, and PAMm mutations (to avoid retargeting of the region), pegRNA2 to pegRNA5 have a shorter RTT to install only the KLF1 BS. (D) Percentage of NGS reads containing the DE (3 mutations: PAMm, KLF1, and TAL1 BSs; PAMm_K_T), the partial edits (1 or 2 mutations; K or PAMm_K), or InDels generated using epegRNA1, ngRNA1.1, and PE3max or PE5max in the top 30% green fluorescent protein–positive (GFP⁺) cells. (E) Percentage of NGS reads containing the DE (2 mutations: KLF1 and TAL1; K_T), the partial edit (1 mutation; K), or InDels generated using epegRNA5, with ngRNA5.5 (left) or ngRNA5.6 (right) and PE3max or PE5max (in their PAM-less SpRY version) in the top 30% GFP⁺ cells. (F-G) Frequency (%) of InDels generated by transfecting plasmids coding for (F) pegRNAs or (G) ngRNAs and Cas9 (pegRNA10 and ngRNA10.1 to ngRNA10.5) or Cas9-SpRY (pegRNA11 to pegRNA14) nucleases. (H) Percentage of NGS reads containing the DE (3 modifications: GATA1 BS, BCL11A-5-nt BS deletion [ΔTGACC from –114 to –118 upstream of HBG1/2 TSS], and KLF1 BS; G_Δ5B_K), DE with mutations, or InDels induced by epegRNA10.1 with ngRNA10.3 or ngRNA10.5 in the top 30% GFP⁺ cells (left). The sequences of the PBS and RTT of epegRNA10.1 are aligned on the most frequent NGS reads. The KLF1 and GATA1 BSs are indicated in gray and bold, respectively, and the desired base conversions are in lower case. The 5 base-pair (5-bp) deletion of the edited 3′-flap facilitates the annealing of its GCC trinucleotide (underlined) to the closest complementary motif (red square) leading to DE with mutations (pink) with or without scaffold incorporation (pink and bold) (right). (I) Percentage of NGS reads containing the DE (3 modifications: GATA1 BS, BCL11A-BS 2-nt deletion [ΔCC –117/–118 upstream of HBG1/2 TSS], and KLF1 BS; G_Δ2B_K), DE with mutations, the partial edit (G_Δ2B) or InDels induced by epegRNA10.2 with ngRNA10.3 or ngRNA10.5 in the top 30% GFP⁺ cells (left). The sequences of the PBS and RTT of epegRNA10.2 are aligned on the most frequent NGS reads. The KLF1 and GATA1 BSs are indicated in gray and bold, respectively, and the desired base conversions are in lower case (right). (J) Percentage of NGS reads containing the DE (2 modifications: GATA1 and KLF1 BSs; G_K), the partial edits (GATA1 or KLF1 BSs), or InDels induced by epegRNA10.3 with ngRNA10.3 or ngRNA10.5 in the top 30% GFP⁺ cells (left). The sequences of the PBS and RTT of epegRNA10.3 are aligned on the most frequent NGS reads. The KLF1 and GATA1 BSs are indicated in gray and bold, respectively, and the desired base conversions in lower case (right). The long 3′-flap facilitates the annealing of its GCC trinucleotide (underlined) to the expected complementary motif, preventing the formation of the DE with unintended mutations. (K) Percentage of NGS reads containing the DEs, the partial edits, or InDels and frequency (%) of 4.9-kb deletion measured by droplet digital PCR (ddPCR) (as previously described⁶ using the primers and probes described in supplemental Table 2) generated using epegRNA10.2 or epegRNA10.4 with ngRNA10.5 into the HBG1/2 promoters in the top 30% GFP⁺ cells. (L-M) Top 5 (L) epegRNA10.4- and (M) ngRNA10.5-dependent off-target DNA sites, as identified by GUIDE-seq⁶ in K562 cells. epegRNA and ngRNA were coupled with a Cas9 nuclease equivalent to the Cas9n included in the PE except for carrying the WT amino acid allowing nuclease activity. The protospacer targeted by each epegRNA or ngRNA and the PAM are reported in the first line, followed by the on- and off-target sites and their mismatches with the on-target (highlighted in color). For each target, the abundance (ie, the total number of unique alignments associated with the target site), chromosomal coordinates (Human GRCh38/hg38), type of region, gene and targeted strand, and number of mismatches with the epegRNA or ngRNA protospacer are reported. In total, we identified 14 off-targets for epegRNA10.4 including only 5 with an abundance >3, and 44 off-targets for ngRNA10.5 including 27 with an abundance >3. Most of them have a high number of mismatches (≥5) and map to nonexonic regions. (B-C,F-G) Samples mock-transfected with Tris-EDTA buffer were used as controls. PCR products were subjected to Sanger sequencing (as previously described⁶ using the primers and probes described in supplemental Table 2), and InDels were measured using the TIDE software.⁷ Bars represent the mean ± standard deviation (SD) of 3 biologically independent replicates. Statistical significance was assessed between mock and treated samples (comparison between mock and each of the treated condition) using an unpaired Kruskal-Wallis test with multiple comparisons (Dunn’s Correction) and displayed for ∗∗P < .01, ∗P < .05, or not significant. (D-E,H-K) NGS analysis was performed as previously described⁶ using the primers and probes described in supplemental Table 2. A customized Python pipeline was used to align NGS reads to a reference amplicon sequence and count desired and partial edits, and other InDels. The “Desired edit” and “Partial edits” categories do not contain InDels and are stacked. InDels are located at the nick induced by the epegRNA or the ngRNA. Bars represent the mean ± SD of n = 3 for the panels D,E,H-J or n = 3 to 4 for the panels K-L biologically independent replicates. Statistical significance was assessed between the different groups using unpaired Kruskal-Wallis test with multiple comparisons (Dunn’s correction): data are not significantly different. WT, wild-type.