Efficient knockout of phytoene desaturase gene using CRISPR/Cas9 in melon

CRISPR/Cas9 system has been widely applied in many plant species to induce mutations in the genome for studying gene function and improving crops. However, to our knowledge, there is no report of CRISPR/Cas9-mediated genome editing in melon (Cucumis melo). In our study, phytoene desaturase gene of melon (CmPDS) was selected as target for the CRISPR/Cas9 system with two designed gRNAs, targeting exons 1 and 2. A construct (pHSE-CmPDS) carrying both gRNAs and the Cas9 protein was delivered by PEG-mediated transformation in protoplasts. Mutations were detected in protoplasts for both gRNAs. Subsequently, Agrobacterium-mediated transformation of cotyledonary explants was carried out, and fully albino and chimeric albino plants were successfully regenerated. A regeneration efficiency of 71% of transformed plants was achieved from cotyledonary explants, a 39% of genetic transformed plants were successful gene edited, and finally, a 42–45% of mutation rate was detected by Sanger analysis. In melon protoplasts and plants most mutations were substitutions (91%), followed by insertions (7%) and deletions (2%). We set up a CRISPR/Cas9-mediated genome editing protocol which is efficient and feasible in melon, generating multi-allelic mutations in both genomic target sites of the CmPDS gene showing an albino phenotype easily detectable after only few weeks after Agrobacterium-mediated transformation.

Melon (Cucumis melo) belongs to Cucurbitaceae family and is an important plant because of its specific biological properties and economic value of its fruit. In fact, 31 million tons of melon were produced worldwide in 2016 14 , being an important crop in Mediterranean and East Asian countries. The recent melon genome publication by Garcia-Mas et al. 15 and the versatility of CRISPR/Cas9 allows to study and explore gene functions in melon. However, to our knowledge, CRISPR/Cas9 application in melon has not been reported. Therefore, the aim of this study was to demonstrate, for the first time, the applicability of CRISPR/Cas9 system on melon by performing a gene knockout of the melon phytoene desaturase gene (CmPDS) in protoplasts and plants.

Results
Target selection and vector construction for CRISPR/Cas9 system. To test the efficacy of CRISPR/ Cas9 system in melon, we chose to disrupt phytoene desaturase gene of Cucumis melo (CmPDS) which has a single copy, located in the chromosome 7, with reference MELO3C017772.2 in ICUGI data base. The genomic sequence of this reference is 10,443 bp in size, with 14 exons (Fig. 1A). Target sites in CmPDS were designed using Benchling 16 and two target sites were selected: gRNA1 and gRNA2, in exon 1 and 2 respectively (Fig. 1A). Both gRNAs were cloned into one binary vector (pHSE-CmelPDS) carrying the promoter for the Cas9 gene (CaMV 35S), the Cas9 gene, AtU6-26p and AtU6-29p promoters and gRNA scaffold (Fig. 1C). Potential off-target sites were searched using CRISPR-OFFinder 17 , and no potential off-target were detected with 0 to 2 mismatches (Table S1) 18 .
Targeted mutagenesis in melon protoplasts. CRISPR/Cas9 vector pHSE-CmelPDS was tested in protoplasts to validate the functionality of Cas9 via transient expression by PEG-mediated protoplast transfection. A total of 24 protoplast cell colonies were Sanger-sequenced. According to the number of mutated colonies detected, the target efficiency for pHSE-CmelPDS in melon protoplasts was 25% for both gRNAs. Moreover, most of the mutations analyzed were substitutions and two nucleotides insertions ( Fig. 3 and Table S2). These results suggested that the vector was viable to be used in melon for plant regeneration.

PDS-edited plants phenotype.
Alteration of CmPDS gene expression function was manifested as albino and dwarf plants (Fig. 2). Some cotyledons only regenerated albino shoots, meanwhile others regenerated green shoots and after a while exhibited mosaicism or regenerated secondary albino shoots. Complete albino plants exhibited a high level of dwarfism, having a reduced leaf area, a growth lesser than 0.5 cm, and impossibility to perform in vitro propagation. A total of 77 plants were regenerated from 958 Agrobacterium-mediated transformed cotyledonary explants. The 22.07% of the regenerated plants showed albino phenotype, 4 complete albino plants and 13 chimeric albinism were detected. Albino plants did not survive more than 3 months after regeneration.   Table S3). In 113 sequenced colonies from albino phenotype plants analyzing both gRNAs, the target mutation www.nature.com/scientificreports www.nature.com/scientificreports/ efficiency was 45% and 42% for gRNA1 and gRNA2, respectively. Up to 91% of mutations were substitution events, 7% were insertions, and 2% deletions. Further analysis of the CmPDS gene mutations revealed substitutions, insertions and deletion events upstream and downstream of the target region for both gRNAs. Multiple alleles were found in all albino and chimeric plants, showing a high level of chimerism for both target sequences (Table S3). In silico analysis of codon change due to substitutions in plants revealed that 28.56% of substitutions caused an introduction of a STOP codon, 60.71% caused a change in amino acid (AA) codon, and a 10.71% caused no AA codon change.

Discussion
To date, CRISPR/Cas9 system has been applied for basic research and trait development in many plant species 19,20 . To our knowledge, this study reports for the first time gene editing by CRISPR/Cas9 in melon (Cucumis melo). We employed this system to target the PDS gene in melon, a key enzyme in the carotenoid biosynthesis pathway involved in at least 20 metabolic pathways, including the inhibition of many genes in carotenoid, chlorophyll, and GA biosynthesis pathways 21 . Successful disruption of PDS gene results in a generation of mutants expected to be photobleached or albino, which allows to phenotypically identify the feasibility of CRISPR/Cas9 in melon. Alteration of PDS was chosen as a visual marker to easily detect the CRISPR/Cas9 genome editing in melon. In order to improve our chances of success we designed two gRNAs and adapted an existing genetic transformation system in cucumber 22 . Chimeric plants, showing albino and green tissue in the same plant or leaf, and albino plants were regenerated from melon cotyledonary explants, as reported in many other species [23][24][25][26][27]  Agrobacterium-mediated transformation is the only reported method for genetic transformation in melon. One of the major issues when transforming species of the Cucurbitacea family is the occurrence of "escapes", which are non-transgenic well-developed shoots regenerated into selective media. In melon, the regeneration rates from cotyledonary explants are usually high (40-70%), nevertheless, transformation rates of the regenerated shoots are usually very low, and is highly genotypic dependent 22,31 . Melon, of the Cucurbitacea family, is a recalcitrant crop for transformation 31 . In our study, 8% of the cotyledonary explants regenerated shoots, among which a high percentage (71%) proved to be transgenics. Although we did not find biallelic mutations in melon plants and albino plants failed to grow in vitro, mutations should be stably inherited in the T1 and T2 generations 32 . The lethal phenotype of PDS mutant plants could be alleviated by optimizing the medium. The use of an additional medium to grow albino plants, increasing sugar composition or increasing cytokinin proportions could allow the survival of albino plants. Furthermore, plant regeneration from cotyledonary explants can induce endoduplication phenomenon that generate tetraploid regenerants 33 . In this experiment we did not analyze the ploidy of the regenerated albino plants because majority of albino plants were also dwarf hence plant material availability was low.
The pHSE-CmelPDS vector contained the Cas9, a gRNA scaffold and both gRNA guides. Promoters from the cauliflower mosaic virus (CaMV35S) and Arabidopsis thaliana (U6 and U3), were used for driving the constitutive expression of Cas9 and both sgRNA (gRNA1 and gRNA2), respectively. Within dicotyledons, when CaMV35S promotor is used the mutation rate depends on the species, from 20 to 100% 26,34 . High transcription levels of Cas9 protein and sgRNA are essential for the activity of the CRISPR/Cas9 system 35 . The expression driver of Cas9 gene is an important factor that can increase the mutation frequency, and the use of a endogenous promoter ensures a high level of transgene expression and allows to increase mutation efficiencies 2 to 7 fold, as showed in soybean and liverwort 28,36 . The endogenous U6 promoter for sgRNA expression has a key role in the CRISPR/Cas9 efficiency too 37 . The regeneration efficiency of mutant plants from cotyledonary explants (1.8%) was very similar to those reported in watermelon 12 . The mutation rate of the sequenced colonies was 42-45% in the analyzed transgenic plants, which is lower than the ones found in other dicotyledons. In protoplast transfected cells, mutation was successful despite low edition frequencies, 25% for each gRNA. In comparison to other species this value is low and could be due either to low transfection efficiency or to a lack of CRISPR/Cas9 editing activity. It is possible that the fact that the promoters used were not endogenous is the cause of the efficiency drop of the CRISPR/Cas9 system. Overall, the use of endogenous promoters for Cas9 and sgRNA expression is the best way to increase the CRISPR/Cas9 efficiency in melon and other cucurbits.
Efficient gene editing in melon presents the possibility to study new gene functions for basic research, and new opportunities for melon productivity by improving biotic stress, melon production and the post-harvest utilization. Towards this, we demonstrated the occurrence of targeted mutagenesis with the CRISPR/Cas9 system in melon protoplasts and plants. The CmPDS knockout system described generates easily detectable albino and dwarf plants and mutation events in a frame of 6-8 weeks. Therefore, it provides a valuable method to facilitate rapid assessment and optimization of CRISPR/Cas9 and other genome-editing technologies in melon.
Extraction of melon protoplast and its transformation. Melon protoplasts were extracted and transfected following the protocol from Yoo et al. 39  Agrobacterium-mediated transformation and plant regeneration. Cucumis melo var. cantaloupensis inbred line "Charentais" (provided by ROCALBA S.A.) was used as the source of explants for genetic transformation. Seed coats were removed, and seeds were surface-sterilized with a 20% sodium hypochlorite solution containing 8 drops·L −1 of Tween 20 for 10 min in agitation and rinsed two times with sterile distilled water. Seedlings were pre-cultured on MS medium supplemented with 3% sucrose for two days. Then, nodal cotyledons were used as explants after removing the embryo. Explants were vacuum-infiltrated with Agrobacterium tumefaciens EHA105 infection solution for 5 min twice and an additional 5 min in a shaker without vacuum. To prepare the Agrobacterium tumefaciens EHA105 infection solution, a single colony containing the selected binary CRISPR/Cas9 vector was picked and put in a 5 mL starter culture of LB containing 50 mg·L −1 Kanamycin and 100 mg·L −1 Rifampicin and cultured for 20 h at 24 °C at 130 rpm. Next, 1 mL of the starter culture was transferred to 45 mL LB culture containing Kanamycin and Rifampicin and cultured until OD 600 reached 0.6. Once the Agrobacterium transformation was carried out, explants were transferred to co-cultivation medium consisting on MS medium supplemented with 3% sucrose and cultured 48 h at 24 °C in the dark. Afterward, explants were rinsed with antibiotic solution composed of 500 mg·L −1 Cefotaxime and 300 mg·L −1 Ticarcillin in distilled water. Then, the explants were transferred into selective medium which was MS supplemented with Sucrose 3%, 2.22 µM BAP, 0.48 µM IAA, 4 µM CuSO 4 ·5H 2 O, 500 mg·L −1 Cefotaxime, 300 mg·L −1 Ticarcillin and 5 mg·L −1 hygromycin B. Every two weeks explants were transferred to fresh selective media. Regenerated plants were transferred to E20A 40 media supplemented with 5 mg·L −1 hygromycin B. Explants on selective medium and plants were cultured under a 16/8 h light/dark photoperiod at 24 °C.
Detection of transgene and CRISPR/Cas9 mutation. Genomic DNA of protoplasts and plants was extracted following the methodology of Doyle and Doyle 41 with minor modifications. The transgene presence in plant (Fig. 1B) was confirmed by PCR using specific primers named pHSE401.SeqF and pHSE401.SeqR (Table S5). A fragment flanking gRNA1 and gRNA2 (846 bp) of the CmPDS gene from transgenic plants and transformed protoplasts was amplified using F-CmPDS and R-CmPDS primers, and the amplified PCR product was gel-purified and cloned into pCR-Blunt II-TOPO vector (Life Technologies). Colonies were Sanger-sequenced using M13F and M13R primers for detecting specific additions, deletions and substitutions. Mutation rate was calculated as the ratio of mutated clonal colonies versus total sequenced colonies.