Genome editing tools have rapidly been adopted by plant scientists for gene function discovery and crop improvement. The current technical challenge is to efficiently induce precise and predictable targeted point mutations valuable for crop breeding purposes. Cytidine base editors (CBEs) are CRISPR/Cas9 derived tools recently developed to direct a C-to-T base conversion. Stable genomic integration of CRISPR/Cas9 components through Agrobacterium-mediated transformation is the most widely used approach in dicotyledonous plants. However, elimination of foreign DNA may be difficult to achieve, especially in vegetatively propagated plants. In this study, we targeted the acetolactate synthase (ALS) gene in tomato and potato by a CBE using Agrobacterium-mediated transformation. We successfully and efficiently edited the targeted cytidine bases, leading to chlorsulfuron-resistant plants with precise base edition efficiency up to 71% in tomato. More importantly, we produced 12.9% and 10% edited but transgene-free plants in the first generation in tomato and potato, respectively. Such an approach is expected to decrease deleterious effects due to the random integration of transgene(s) into the host genome. Our successful approach opens up new perspectives for genome engineering by the co-edition of the ALS with other gene(s), leading to transgene-free plants harboring new traits of interest.

Genome editing has become a major tool for both functional studies and plant breeding in several species. Besides generating knockouts through the classical CRISPR-Cas9 system, recent development of CRISPR base editing holds great and exciting opportunities for the production of gain-of-function mutants. The PAM requirement is a strong limitation for CRISPR technologies such as base editing, because the base substitution mainly occurs in a small edition window. As precise single amino-acid substitution can be responsible for functions associated to some domains or agronomic traits, development of Cas9 variants with relaxed PAM recognition is of upmost importance for gene function analysis and plant breeding. Recently, the SpCas9-NG variant that recognizes the NGN PAM has been successfully tested in plants, mainly in monocotyledon species. In this work, we studied the efficiency of SpCas9-NG in the model moss Physcomitrella patens and two Solanaceae crops (Solanum lycopersicum and Solanum tuberosum) for both classical CRISPR-generated gene knock-out and cytosine base editing. We showed that the SpCas9-NG greatly expands the scope of genome editing by allowing the targeting of non-canonical NGT and NGA PAMs. The CRISPR toolbox developed in our study opens up new gene function analysis and plant breeding perspectives for model and crop plants.

Targeted insertion of transgenes in plants is still challenging and requires further technical innovation. In the present study, we chose the tomato DFR gene involved in anthocyanin biosynthesis as a landing pad for targeted transgene insertion using CRISPR-Cas9 in a two-step strategy. First, a 1013 bp was deleted in the endogenous DFR gene. Hypocotyls and callus of in vitro regenerated plantlets homozygous for the deletion were green instead of the usual anthocyanin produced purple colour. Next, standard Agrobacterium-mediated transformation was used to target transgene insertion at the DFR landing pad in the dfr deletion line. The single binary vector carried two sgRNAs, a donor template containing two homology arms of 400 bp, the previously deleted DFR sequence, and a NptII expression cassette. Regenerating plantlets were screened for a purple-colour phenotype indicating that DFR function had been restored. Targeted insertions were identified in 1.29% of the transformed explants. Thus, we established an efficient method for selecting HDR-mediated transgene insertion using the CRISPR-Cas9 system in tomato. The visual screen used here facilitates selection of these rare gene targeting events, does not necessitate the systematic PCR screening of all the regenerating material and can be potentially applied to other crops.

Summary: Resistance to potyviruses in plants has been largely provided by the selection of natural variant alleles of eukaryotic translation initiation factors (eIF) 4E in many crops. However, the sources of such variability for breeding can be limited for certain crop species, while new virus isolates continue to emerge. Different methods of mutagenesis have been applied to inactivate the eIF4E genes to generate virus resistance, but with limited success due to the physiological importance of translation factors and their redundancy. Here, we employed genome editing approaches at the base level to induce non‐synonymous mutations in the eIF4E1 gene and create genetic diversity in cherry tomato (Solanum lycopersicum var. cerasiforme). We sequentially edited the genomic sequences coding for two regions of eIF4E1 protein, located around the cap‐binding pocket and known to be important for susceptibility to potyviruses. We show that the editing of only one of the two regions, by gene knock‐in and base editing, respectively, is not sufficient to provide resistance. However, combining amino acid mutations in both regions resulted in resistance to multiple potyviruses without affecting the functionality in translation initiation. Meanwhile, we report that extensive base editing in exonic region can alter RNA splicing pattern, resulting in gene knockout. Altogether our work demonstrates that precision editing allows to design plant factors based on the knowledge on evolutionarily selected alleles and enlarge the gene pool to potentially provide advantageous phenotypes such as pathogen resistance.