Figure 2.

Modulation of binding affinity and specificity by guide RNA variants. (A) Schematic of dCas9 bound to a single-guide RNA with a two nucleotide truncation from its 5′- end (tru-gRNA, purple). (B) Schematic and proposed mechanism of dCas9 bound to a single-guide RNA with 5′- end extension that forms a hairpin with the PAM-distal binding segment of its targeting region (hp-gRNA, blue). After a PAM site is bound and strand invasion of the DNA by the guide RNA has initiated, the hairpin is opened upon binding to a full protospacer and full strand invasion can occur. If there are PAM-distal mismatches at the target site, then it is more energetically favourable for the hairpin to remain closed and strand invasion is hindered. (C) Single-site binding affinities (K_A) for dCas9 with tru-gRNA (purple, n = 257) along the engineered DNA substrate (see Figure 1D). Dashed line shows the site-specific affinities of dCas9-sgRNA for comparison. The binding distribution of dCas9 with tru-gRNAs exhibits distinct peaks in its affinity exactly at the protospacer sites with 10 PAM-distal mismatches and 5 PAM-distal mismatches, demonstrating that it does not have increased binding specificity relative to full sgRNAs (see Table 1). (D) Single-site binding affinities (K_A) for dCas9 with guide RNAs with 5′- hairpins that overlap the nucleotides complementary to the last 6 (hp6-gRNA, blue) or 10 (hp10-gRNA, green) PAM-distal nucleotides of the protospacer. The specific peaks at the sites with 5 and 10 distal mismatches are significantly flattened, with dCas9 and hp10-gRNA exhibiting substantially decreased affinity for off-target sites (22% drop relative to dCas9 with tru-gRNA in their maximum observed off-target binding affinity, with a decrease of up to 57% at sites with 10 PAM-distal mismatches). The peaks in affinity at the full protospacer sites imply that the hairpins indeed open upon full invasion. n = 243 for hp6-gRNA and n = 212 for hp10-gRNA.