An adenine base editor with expanded targeting scope using SpCas9‐NG v1 in rice

Abstract

The CRISPR/Cas9 genome editing system induces mainly small insertions or deletions at the target site. In contrast, base editors are precise genome editing tools that enable base conversion at a target site without inducing a DNA double-strand break. Two types of base editors have now been developed: cytosine base editors (CBEs)—cytidine deaminase fused with catalytically impaired Cas9—can efficiently convert a C–G base pair to a T–A pair, while adenine base editors (ABEs)—Cas9 nickase (nCas9) plus engineered Escherichia coli adenosine deaminase (TadA)—enable A–T to G–C conversions (Gaudelli et al., 2017; Kim, 2018). Both CBEs and ABEs have been applied successfully in plants such as Arabidopsis, wheat and rice (e.g., Hua et al., 2018, 2019; Li et al., 2018). However, potential base editing targets remain restricted by the requirement for protospacer adjacent motif (PAM) sequences around these sites. Cas9 derived from Streptococcus pyogenes (SpCas9)—currently the Cas9 most widely used for genome editing in many organisms—requires as a PAM an NGG sequence adjacent to the target site. To extend Cas9-recognized sites, several groups have engineered Cas9 variants with amino acid substitutions and modified PAM sequences. Using these Cas9 variants in base editors could expand the scope of base editing sites (Endo et al., 2019; Hu et al., 2018; Hua et al., 2019; Nishimasu et al., 2018). Engineered SpCas9s, SpCas9-NG variants and xCas9 variants that can recognize NG as PAM have been reported recently (Hu et al., 2018; Nishimasu et al., 2018). SpCas9-NG and xCas9 variants can induce mutation at target sites with NG PAMs in rice (Endo et al., 2019; Wang et al., 2019), while the in vitro DNA cleavage activity of SpCas9-NGv1 (reported as ARVRFRR; Nishimasu et al., 2018), which includes a 7-aa mutation (R1335A/L1111R/D1135V/G1218R/E1219F/A1322R/T1337R) in the PAM-interacting domain, is higher than that of xCas9 (Endo et al., 2019; Nishimasu et al., 2018). SpCas9-NGv1 is also available for C–T base conversions as CBEs in rice (Endo et al., 2019). Here, we report the development of novel ABEs using nSpCas9-NGv1 (ABE7.10-nSpCas9-NGv1) and show that SpCas9-NGv1 can induce A–G substitutions at a target site containing NG as PAM in the rice genome. We first examined whether the ABE system can function in rice callus. We synthesized an Arabidopsis thaliana codon-optimized coding region containing TadA-wt-(XTEN)2-TadA*7.10-(XTEN)2 (Gaudelli et al., 2017). Then, A. thaliana codon-optimized nSpCas9 (D10A) and 3×NLS were fused to the C-terminus of the coding region (Figure 1a). We chose four target sites (sgOs-siteG1: chr05_27537337–27537356; sgOs-site2: chr05_27536662–27536681; sgOs-site3: chr02_19223102–19223121; and sgOs-site4: chr11_23944837–23944856) with NGG at their 3′ end as PAM in the rice genome (Figure 1b, c). A binary vector harbouring a single guide RNA (sgRNA) and ABE7.10-nSpCas9 expression cassettes was integrated into the rice genome via Agrobacterium-mediated transformation. After culturing for 1 month, genomic DNA extracted from the transformed calli was subjected to cleaved amplified polymorphic sequences (CAPS) analysis. All four target sites have restriction enzyme sites overlapping the editing window of previously reported ABEs (Gaudelli et al., 2017; Hua et al., 2018). Undigested DNA fragments were detected in transgenic calli, suggesting that ABE7.10-nSpCas9 had indeed induced the desired mutation, and with roughly the same target window as other reported ABEs (Figure 1b). The ratios of mutated callus detected by CAPS analysis at the four different target sites were 81.7% (76/93 calli: site G1), 63.8% (51/80 calli: site G2), 68.5% (63/92 calli: site G3), and 69.3% (61/88 calli: site G4). DNA sequences around the target sites were confirmed using several transgenic calli in which undigested DNA fragments were clearly detected by CAPS analysis. Sequence analysis revealed that ABE7.10-nSpCas9 efficiently induced an A–G substitution at a position −16 to −13 nt upstream from the PAM (Figure 1c). This target window is almost the same as that seen in data reported previously in human and rice cells (Gaudelli et al., 2017; Hua et al., 2018). We detected no substitutions other than from A to G. These results showed that ABE7.10-nSpCas9 can direct A–G substitutions effectively at NGG PAM target sites. The newly engineered SpCas9-NGv1 can recognize NG as PAM, which increases the target sites available to Cas9-mediated genome editing (Endo et al., 2019; Nishimasu et al., 2018). To apply the SpCas9-NGv1 to the ABE system, we constructed ABE7.10-nSpCas9-NGv1 harbouring the nickase type of SpCas9-NGv1 instead of nSpCas9 (Figure 1a). We chose four target sites (sgOs-siteG1; sgOs-siteA: chr06_1770405–1770424; sgOs-siteC: chr04_33074491–33074510; and sgOs-siteT: chr06_2926795–2926814) in the rice genome with NGG, NGA, NGC and NGT, respectively, as PAM sequences (Figure 1d). Genomic DNA was extracted from ABE7.10-nSpCas9-NGv1