DNA targeting specificity of RNA-guided Cas9 nucleases

Abstract

The Streptococcus pyogenes Cas9 (SpCas9) nuclease can be efficiently targeted to genomic loci by means of single-guide RNAs (sgRNAs) to enable genome editing. Here, we characterize SpCas9 targeting specificity in human cells to inform the selection of target sites and avoid off-target effects. Our study evaluates >700 guide RNA variants and SpCas9-induced indel mutation levels at >100 predicted genomic off-target loci in 293T and 293FT cells. We find that SpCas9 tolerates mismatches between guide RNA and target DNA at different positions in a sequence-dependent manner, sensitive to the number, position and distribution of mismatches. We also show that SpCas9-mediated cleavage is unaffected by DNA methylation and that the dosage of SpCas9 and sgRNA can be titrated to minimize off-target modification. To facilitate mammalian genome engineering applications, we provide a web-based software tool to guide the selection and validation of target sequences as well as off-target analyses.

Figure 1

Optimization of guide RNA architecture for SpCas9-mediated mammalian genome editing. (a) Schematic of bicistronic expression vector (PX330) for U6 promoter-driven sgRNA and CBh promoter-driven human codon-optimized S. pyogenes Cas9 (hSpCas9) used for all subsequent experiments. The sgRNA consists of a 20-nt guide sequence (blue) and scaffold (red), truncated at various positions as indicated. (b) SURVEYOR assay for SpCas9-mediated indels at the human EMX1 and PVALB loci. Arrowheads indicate the expected SURVEYOR fragments (n = 3). (c) Northern blot analysis for the four sgRNA truncation architectures, with U1 as loading control. (d) Both wild-type (WT) or nickase mutant (D10A) of SpCas9 promoted insertion of a HindIII site into the human EMX1 gene. Single-stranded oligonucleotides, oriented in either the sense or antisense direction relative to genome sequence, were used as homologous recombination templates (Supplementary Fig. 3). (e) Schematic of the human SERPINB5 locus. sgRNAs and PAMs are indicated by colored bars above sequence; methylcytosine (Me) are highlighted (pink) and numbered relative to the transcriptional start site (TSS, +1). (f) Methylation status of SERPINB5 assayed by bisulfite sequencing of 16 clones. Filled circles, methylated CpG; open circles, unmethylated CpG. (g) Modification efficiency by three sgRNAs targeting the methylated region of SERPINB5, assayed by deep sequencing (n = 2). Error bars indicate Wilson intervals (Online Methods).

Figure 2

Single-nucleotide specificity of SpCas9. (a) Schematic of the experimental design. sgRNAs carrying all possible single base-pair mismatches (blue Ns) throughout the guide sequence were tested for each EMX1 target site (target site 1 shown as example). (b) Heatmap representation of relative SpCas9 cleavage efficiency by 57 single-mutated and 1 nonmutated sgRNA each for four EMX1 target sites (aggregated from Supplementary Table 5). For each EMX1 target, the identities of single base-pair substitutions are indicated on the left; original guide sequence is shown above and highlighted in the heatmap (gray squares). Modification efficiencies (increasing from white to dark blue) are normalized to the original guide sequence. Sequence logo representation of the same data can be found in Supplementary Figure 7. (c) Heatmap for relative SpCas9 cleavage efficiency for each possible RNA:DNA base pair, compiled from aggregate data from single-mismatch guide RNAs for 15 EMX1 targets (Supplementary Fig. 8). Mean cleavage levels were calculated for the 10 PAM-proximal bases (right bar) and across all substitutions at each position (bottom bar); positions in gray were not covered by the 469 single-mutated and 15 unmutated sgRNAs tested (Supplementary Table 5). (d) SpCas9-mediated indel frequencies at targets with all possible PAM sequences, determined using the SURVEYOR nuclease assay. Two target sites from the EMX1 locus were tested for each PAM (Supplementary Table 4). (e) Histogram of distances between 5′-NRG PAM occurrences within the human genome. Putative targets were identified using both strands of human chromosomal sequences (GRCh37/hg19).

Figure 3

Multiple mismatch specificity of SpCas9. (a–c) SpCas9 cleavage efficiency with guide RNAs containing consecutive mismatches of 2, 3 or 5 bases (a), or multiple mismatches separated by different numbers of unmutated bases for EMX1 targets 1, 2, 3 and 6 (b, c). Rows represent each mutated guide RNA; nucleotide substitutions are shown in white cells; gray cells denote unmutated bases. All indel frequencies are absolute and analyzed by deep sequencing from two biological replicas. Error bars indicate Wilson intervals (Online Methods).

Figure 4

SpCas9-mediated indel frequencies at predicted genomic off-target loci. (a, b) Cleavage levels at putative genomic off-target loci containing two or three individual mismatches (white cells) for EMX1 target 1 and target 3 are analyzed by deep sequencing. List of off-target sites are ordered by median position of mutations. Putative off-target sites with additional mutations did not have detectable indels (Supplementary Table 8). The Cas9 dosage was 3 × 10⁻¹⁰ nmol/cell, with equimolar sgRNA delivery. Error bars indicate Wilson intervals (Online Methods). (c, d) Indel frequencies for EMX1 targets 1 and 3 and selected off-target loci (OT) as a function of SpCas9 and sgRNA dosage, (n = 2, Wilson intervals). 400 ng to 10 ng of Cas9-sgRNA plasmid corresponds to 7.1 × 10⁻¹⁰ to 1.8 × 10⁻¹¹ nmol/cell. Cleavage specificity is measured as a ratio of on- to off-target cleavage.