CRISPR-Cas9 fusion to dominant-negative 53BP1 enhances HDR and inhibits NHEJ specifically at Cas9 target sites

Abstract

Precise genome editing/correction of DNA double-strand breaks (DSBs) induced by CRISPR-Cas9 by homology-dependent repair (HDR) is limited by the competing error-prone non-homologous end-joining (NHEJ) DNA repair pathway. Here, we define a safer and efficient system that promotes HDR-based precise genome editing, while reducing NHEJ locally, only at CRISPR-Cas9-induced DSBs. We fused a dominant-negative mutant of 53BP1, DN1S, to Cas9 nucleases, and the resulting Cas9-DN1S fusion proteins significantly block NHEJ events specifically at Cas9 cut sites and improve HDR frequency; HDR frequency reached 86% in K562 cells. Cas9-DN1S protein maintains this effect in different human cell types, including leukocyte adhesion deficiency (LAD) patient-derived immortalized B lymphocytes, where nearly 70% of alleles were repaired by HDR and 7% by NHEJ. Our CRISPR-Cas9-DN1S system is clinically relevant to improve the efficiencies of precise gene correction/insertion, significantly reducing error-prone NHEJ events at the nuclease cleavage site, while avoiding the unwanted effects of global NHEJ inhibition.

Fig. 1

Identification of an HDR-enhancing fragment of 53BP1. a Schematic diagram of full-length human 53BP1 protein showing the minimal focus forming region including the oligomerization domain (OD), a glycine-arginine rich (GAR) motif, a tandem Tudor domain, and the ubiquitin-dependent recruitment (UDR) motif. BRCT repeats at the extreme C-terminus are present. b Different truncated 53BP1 proteins, DN1, DN1S, DN2, DN3, and DN4, were tested for their relative expression by western blot using anti-HA antibody in HeLa cells transduced with empty vector (mock) or lentiviral vectors encoding HA-tagged 53BP1 fragments DN1, DN1S, DN2, DN3, or DN4. Actin is shown as a loading control. c, e, g Representative immunofluorescence (IF) images of HA-tagged DN1 and DN1S at nuclear foci. Cells were either exposed to 2 Gy IR or no IR (control) and fixed 2 h later. Scale bars represent 20 μm. c Representative IF images showing co-localization of HA-tagged DN1 or DN1S with endogenous 53BP1 foci. d Number (no.) of HA+ (DN1S) foci, co-localized HA+/endogenous 53BP1+ (co-localization) foci, or endogenous 53BP1+ (E-53BP1) only foci per IR-treated HeLa cell expressing DN1S. Individual quantifications of foci from 50 cells in each group are shown. Red lines are drawn at the mean number of foci per positive cell. e Representative IF images showing HA-tagged DN1 or DN1S and RIF1 recruitment to IR-induced DNA repair foci. f Quantification of cells with ≥3 RIF1 foci in control cells or IR cells with DN1, DN1S or without vector (mock). g Representative IF images showing HA-tagged DN1S and BRCA1 recruitment to DNA repair foci in control cells. h Quantification of the number of cells with ≥3 BRCA1 foci in control cells or IR cells, with DN1, DN1S or without vector (mock). In panels f and g, data are presented as the mean ± SEM of counts of 150 cells each, from three independent fields, and black circles indicate individual counts. Statistics for panels d, f, and g: ANOVA. ns indicates not significant, *p < 0.05, ***p < 0.001, and ****p < 0.0001. Source data is available in Source Data file

Fig. 2

Cas9-DN1S blocks NHEJ locally, reducing toxicity associated with global NHEJ inhibition. a Representative immunofluorescence (IF) images showing HA-tagged DN1S or dCas9-DN1S/gRNA and endogenous 53BP1 recruitment to DNA repair foci. Scale bar represents 20 μm. b Quantification of the number of cells with ≥1 HA+ (DN1S) foci. c Quantification of the number of HeLa cells with ≥3 53BP1+ foci transduced with lentivirus for the following: mock (empty vector), DN1S, dCas9, dCas9-DN1S, and dCas9-DN1S/gRNA. b, c the data are presented as the mean ± SEM of three counts of 150 cells each from independent fields, except dCas9-DN1S/gRNA in b, which shows six counts of 150 cells from independent fields. Black circles indicate individual counts. Statistics: ANOVA, ***p < 0.001, ****p < 0.0001. d Representative IF images showing co-localization of dCas9 or dCas9-DN1S with CENPB gRNA to CENPB in centromeres. dCas9-DN1S with CD45 gRNA was used as a control. Scale bar represents 10 μm. e GFP+ HeLa cells with stable integration of the EJ5-GFP NHEJ reporter, after exposure to lentiviral constructs expressing DN1S or Cas9-DN1S, and lipofection with the I-SceI plasmid. I-SceI induced DSB repair via NHEJ at defined sites in the reporter allows for GFP expression. GFP+ cells were quantified by flow cytometry and data normalized to GFP+ cells in mock controls. Viability of (f) Jurkat cells and (g) K562 cells 48 h after DN1S or empty vector transduction, and 24 h after Cas9 or Cas9-DN1S RNP electroporation with AAV donor template delivery. Viability was quantified by flow cytometry using eFluor 780 fixable viability dye. a–e Cas9 denotes SpCas9. f–g Cas9 denotes SaCas9; e–g data are presented as the mean ± SEM of three independent transfections or electroporations. Black circles indicate individual data points. Statistics: ANOVA. ns indicates not significant, **p < 0.01, ***p < 0.001, ****p < 0.0001. Source data is available in Source Data file

Fig. 3

Cas9-DN1S stimulates HDR at different target genes in multiple cell lines. a Bar plots showing the HDR editing efficiency of SpCas9 or SpCas9-DN1S at the AAVS1 and LMO2 loci in 293T cells, the CD45 locus in K562 cells, and the CCR5 locus in Jurkat cells. SpCas9 or SpCas9-DN1S and the donor templates were delivered through the plasmid system in 293T and K562 cells. SpCas9 or SpCas9-DN1S were delivered by ribonucleoprotein (RNP) and the CCR5-GFP donor by rAAV6 in Jurkat cells. HDR efficiency was determined by the percentage of GFP+ cells. The data are presented as the mean ± SEM of three independent electroporations. Black circles indicate individual data points. Statistics: Unpaired t tests, one tailed. *P < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001. b Stacked bar plots showing the NHEJ and HDR editing efficiency of SaCas9 or SaCas9-DN1S at the CD45 and AAVS1 loci in LCL cells, at the AAVS1 locus in K562 cells, and at the CD45 and AAVS1 loci in Jurkat cells. SaCas9 or SaCas9-DN1S was delivered as RNP and donor templates were delivered via rAAV6 vectors. HDR efficiency was quantified by the percentage of GFP+ cells. NHEJ efficiency at the AAVS1 locus was determined by amplifying the edited locus by PCR and performing TIDE assay. When targeting the CD45 locus, NHEJ efficiency was determined by the CD45- population by flow cytometry, and confirmed by TIDE assay. The data are presented as the mean ± SEM of 3–5 independent electroporations. Statistics: Unpaired t-tests, one tailed comparing NHEJ (magenta asterisks) or HDR (green asterisks) separately: **p < 0.01, ***p < 0.001, and ****p < 0.0001. Source data is available in Source Data file

Fig. 4

CD18 insertion at the AAVS1 locus in LAD patient-derived B lymphocytes. a Representative flow cytometry plots of EBV-immortalized LAD patient B-lymphocytes, electroporated for targeted integration of CD18 at the AAVS1 locus, showing fractions of HDR+ cells. Donor template (DT) was delivered via rAAV6. b Stacked bar plots showing quantification of HDR efficiency by flow cytometry and NHEJ efficiency by TIDE assay of SaCas9 or SaCas9-DN1S with AAVS1-CD18 rAAV6 donor template. The data are presented as the mean ± SEM of three independent electroporations. Statistics: unpaired t tests, one tailed comparing HDR (green asterisks) or NHEJ (magenta asterisks): **p < 0.01, ****p < 0.0001. c Relative frequencies of bi-allelic HDR detected by sorting the GFP^dim+ and GFP^bright+ cell populations as gated, based on the density and MFI of CD18 by flow cytometry. The sorted GFP^bright+ population was confirmed to be bi-allelic HDR by PCR (shown in Supplementary Fig. 14). The data are presented as the mean ± SEM of three independent electroporations. Black circles indicate individual data points. Statistics: unpaired t-tests, one tailed: ****p < 0.0001. Source data is available in Source Data file