Increasing the genome-targeting scope and precision of base editing with engineered Cas9-cytidine deaminase fusions

Abstract

Base editing induces single-nucleotide changes in the DNA of living cells using a fusion protein containing a catalytically defective Streptococcus pyogenes Cas9, a cytidine deaminase, and an inhibitor of base excision repair. This genome editing approach has the advantage that it does not require formation of double-stranded DNA breaks or provision of a donor DNA template. Here we report the development of five C to T (or G to A) base editors that use natural and engineered Cas9 variants with different protospacer-adjacent motif (PAM) specificities to expand the number of sites that can be targeted by base editing 2.5-fold. Additionally, we engineered base editors containing mutated cytidine deaminase domains that narrow the width of the editing window from ∼5 nucleotides to as little as 1-2 nucleotides. We thereby enabled discrimination of neighboring C nucleotides, which would otherwise be edited with similar efficiency, and doubled the number of disease-associated target Cs able to be corrected preferentially over nearby non-target Cs.

Figure 1. SaBE3, SaKKH-BE3, VQR-BE3, EQR-BE3, and VRER-BE3 mediate efficient base editing at target sites containing non-NGG PAMs in human cells

a, Base editor architectures using S. pyogenes and S. aureus Cas9, and recently characterized Cas9 variants with alternate or relaxed PAM requirements^,. b–f, HEK293T cells were treated with the base editor variants shown as described in the Methods. The percentage of total DNA sequencing reads (with no enrichment for transfected cells) with C converted to T at the target positions indicated are shown. The PAM sequence of each target tested is shown below the X-axis. The charts show the results for SaBE3 and SaKKH-BE3 at genomic loci with NNGRRT PAMs (b), SaBE3 and SaKKH-BE3 at genomic loci with NNHRRT PAMs (c), VQR-BE3 and EQR-BE3 at genomic loci with NGAG PAMs (d) and with NGAH PAMs (e), and VRER-BE3 at genomic loci with NGCG PAMs (f). Values and error bars reflect the mean and standard deviation of at least two biological replicates.

Figure 2. Base editors with mutant cytidine deaminase domains exhibit narrowed editing windows

a–c, HEK293T cells were transfected with plasmids expressing mutant base editors and an appropriate sgRNA. Three days after transfection, genomic DNA was extracted and analyzed by high-throughput DNA sequencing at the indicated loci. The percentage of total DNA sequencing reads (without enrichment for transfected cells) with C changed to T at the target positions indicated are shown for the EMX1 site, HEK293 site 3, FANCF site, HEK293 site 2, site A, and site B loci. a, Cytidine deaminase mutations narrow the width of the editing window. See Supplementary Figure 6 for the characterization of additional mutations. b, Effect of cytidine deaminase mutations that narrow editing window width on genomic loci containing multiple Cs within the canonical BE3 editing window (positions 4–8). Combining mutations has an additive effect on narrowing the editing window. YE1 = W90Y + R126E; EE = R126E + R132E; YE2 = W90Y + R132E; YEE = W90Y + R126E + R132E. c, YE1-BE3, YE2-BE3, EE-BE3, and YEE-BE3 alter the product distribution of base editing, producing predominantly singly-modified products, in contrast with BE3, even when multiple Cs are present in positions 4–8. Values and error bars reflect the mean and standard deviation of at least two biological replicates.

Figure 3. Genetic variants in ClinVar that can be corrected in principle by the base editors developed in this work

The NCBI ClinVar database of human genetic variations and their corresponding phenotypes (see Ref. 5) was searched for genetic diseases that in theory can be corrected by base editing. a, The base editors with altered PAM specificities developed in this study substantially increases targetable loci among all pathogenic T→C or A→G mutations in the ClinVar database. b, Improvement in base editing targeting scope among all pathogenic T→C or A→G mutations in the ClinVar database through the use of base editors with narrowed activity windows. BE3 is assumed to edit Cs in positions 4–8 with comparable efficiency as shown in Fig. 2. YE1-BE3, EE-BE3, YE2-BE3, and YEE-BE3 are assumed to edit with a preference of C5 > C6 > C7 ≈ C4. The blue fractions denote the proportion of pathogenic T→C or A→G mutations that can be edited by BE3 without comparable editing of other Cs (left), or that can be edited by BE3 or by one of BE3 mutants without comparable editing of other Cs (right).