Prime genome editing in rice and wheat

Abstract

Prime editors, which are CRISPR-Cas9 nickase (H840A)reverse transcriptase fusions programmed with prime editing guide RNAs (pegRNAs), can edit bases in mammalian cells without donor DNA or double-strand breaks. We adapted prime editors for use in plants through codon, promoter, and editing-condition optimization. The resulting suite of plant prime editors enable point mutations, insertions and deletions in rice and wheat protoplasts. Regenerated prime-edited rice plants were obtained at frequencies of up to 21.8%. Introduction of genome modifications such as substitutions, insertions, and deletions that improve agronomic traits can accelerate crop improvement and breeding 1,2. In plants, nuclease-initiated homology-directed repair (HDR) is limited by low efficiency and the difficulty of DNA template delivery 3-6. Cytosine and adenine base editors (CBEs and ABEs) install C•G-to-T•A and A•T-to-G•C transitions 7-9 , and have been successfully used in plants 4. However, base editors are unable to install transversions, insertions, or deletions 7-10. Prime editing uses engineered Cas9 nickase-reverse transcriptase (RT) fusion proteins paired with a pegRNA that encodes the desired edit 11. The RT domain uses a nicked genomic DNA strand as a primer for the synthesis of an edited DNA flap templated by an extension on the pegRNA. Subsequent DNA repair incorporates the edited flap, permanently installing the programmed edit 11. To optimize prime editing for plants, we first compared three plant prime editor systems (PPEs): PPE2, PPE3, and PPE3b 11 (Fig. 1a). PPE2 consists of a nCas9(H840A) fused to an engineered M-MLV RT, and a pegRNA composed of a primer binding site (PBS) and an RT template 11. PPE3 adds an additional nicking single guide RNA (sgRNA) to cleave the non-edited strand, which facilitates favorable DNA repair. In PPE3b, this nicking sgRNA targets the edited sequence, thereby preventing nicking of the non-edited strand until after editing occurs, resulting in fewer indels in mammalian cells 11. We codon-optimized PPE genes for cereal plants and expressed them using the maize Ubiquitin-1 (Ubi-1) promoter (Fig. 1b). We used the OsU3 (or TaU6) and TaU3 promoters to drive pegRNA and nicking sgRNA transcription, respectively. To test whether other RTs support prime editing, we replaced the engineered M-MLV RT with either the CaMV RT (RT-CaMV) from cauliflower mosaic virus 12 or a retron-derived RT (RT-retron) from E. coli BL21 (ref. 13) (Fig. 1b). We first used our previously described 14 rice protoplast reporter system to test the PPE system for blue fluorescent protein (BFP) to green fluorescent protein (GFP) conversion, which requires changing codon 66 from CAC (histidine) to TAC (tyrosine)

Prime editing is achievable in the tetraploid potato, but needs improvement

“…Very recently, several reports have shown that prime editing can be applied to cereal plants Lin et al, 2020;Tang et al, 2020;Xu et al, 2020a;Xu et al, 2020b;Butt et al;Hua et al). All these studies showed that the efficiency of PPEs greatly varies between target sites and is sensitive to several parameters such as both the PBS and RT sequence.…”

Prime editing primarily induces undesired outcomes in mice

“…The PE system was primarily developed in HEK293T cells and extensive testing at multiple loci showed high levels of installation of the intended mutations at target loci with indel rates typically less than 5%, suggesting that PE-based editing could represent a monumental step forward in the development of safe and effective gene therapies for a wide range of genetic disorders. However, additional testing in other non-primary cell lines, primary cultured mouse neurons, plants, and mice all showed lower overall rates of editing, more variable indel rates, and increased rates of insertions of pegRNA scaffold sequences (Anzalone et al, 2019;Lin et al, 2020;Liu et al, 2020), suggesting that differences in DNA repair mechanisms, cell cycle status, and chromosomal stability could have profound effects on PE outcomes. Currently, it is of great interest to understand whether PE broadly works in vivo for animal model production and therapeutic development (Cohen, 2019;Hampton, 2020;Ravindran, 2019;Urnov, 2020).…”

Parallel genetics of regulatory sequencesin vivo

“…Although indels are applicable to regulatory regions and even coding sequences [55][56][57] , point mutagenesis would enable fi ne mapping of regulatory nucleotides and amino acids. This exciting possibility could be opened up by implementing hyperactive base [58][59][60][61] -or prime editors [62][63][64] . Alternatives to CRISPR-Cas could be developed based on inducible expression of cassette exchange 65,66 or targeted transposons 67,68 .…”