In our research, we investigated the molecular interaction between O. cumana F race (from Spain) and three susceptible sunflower genotypes. The screening for root infection with germinated O. cumana seeds (IN23 genotype), was done in a rhizotron system (plexiglass plates) at the secondment partner institution, INRAE-Toulouse (supervised by Stéphane Muños). The bio-assay included three biological replicates of each genotype with twelve individual plants for both infected and control treatments. The interaction was examined at four particular stages: (1) O. cumana seeds germinated and established contact without vascular attachment yet (pre-haustorial phase); (2) Vascular connection established (haustorial phase); (3) young tubercule (nutrient organs of O. cumana) established on the root tissue of sunflower and (4) Systemic sunflower root tissue above the tubercule. At Keygene N.V. these samples were used for mRNA isolation, library preparation and RNA-sequencing.
Using a large scale transcriptomics analysis (RNA-Seq), we performed a genome wide differential expression analysis using the proprietary computational lead discovery pipeline – KeySeeQTM. The KeySeeQTM pipeline included: (1) mRNA and miRNA isolation, library preparation and Illumina sequencing; (2) quality check of raw sequence data and absence of biases; (3) mapping of transcriptome reads onto reference genomes XRQr.2 Helianthus annuus (Badouin et al. 2017) and O. cumana (IN23-F race); (4) the quantification of expression profiles and (5) differential expression analysis (DeSeq2). The transcriptome profiling was based on the search for differentially expressed (DE) genes between conditions across samples with the main focus on upregulated genes. We further refined the best gene candidates based on following criteria: (1) upregulated genes with min Fold Change value of 2; (2) p-value of ≤ 0.05; (3) commonality analysis for three genotypes across four stages to downscale number of candidate genes. Final candidate selection was done by considering the functional annotation of the genes and expression at the first two stages described above.
This resulted in 37 selected putative susceptibility genes which were targeted for mutagenesis breeding using KeyPoint® breeding technology in a highly ethylmethanesulfonate (EMS) mutagenized M2 sunflower population for the genotype HA304 with a mutation frequency of ~0.04 variants/kb/plant that was already available at Keygene. A total of 6048 individuals of this M2 population were grown over the summer of 2019, harvested and processed to obtain M3 seed batches per M2 individual in the period of September 2019 to February 2020 and leaf material was collected from each M2 individual for subsequent screening for mutants. From the total of 37 genes selected from RNA-Seq analysis, three genes dropped out from the KeyPoint® mutagenesis pipeline during the set-up of the assays, leaving 34 genes to be screened. The total number of mutations identified was 311, from which 196 were functional mutations (62% of the total identified mutations). These mutations were ranked according to the following criteria: (1) Selection for EMS point mutations (G/A and C/T transitions); (2) Selection based on the quality check automatically performed by the KeyPoint® mutagenesis pipeline (high, moderate and low quality); (3) Selection based on the type of the mutation (e.g. stop codon, splice donor variants, missense variants); (4) Functional validation of the mutation and (5) Technical validation of mutations (validation of SNPs by MiSeq-Nano sequencing). For the functional and technical validation procedure, we only targeted stop gained mutations, splice donor variants and missense variants. Loss-of-function variants (LoF) include frameshifting and stop variants that are of particular interests because of their potentially profound impact on the mRNA transcript and translated protein of the respective gene. The total number of stop gained mutations was 11, distributed along eight candidate genes. Splice donor variants yielded two mutations in two targeted genes, whereas the total number of missense variants corresponded to 181, distributed along 32 prioritized genes.
We successfully identified individuals with alleles that have an expected result in reduced or absence of parasitism. Plants with technically validated mutations were planned to be phenotyped with a bio-assay (rhizotron) system with O. cumana (F race-IN23) at the secondment partner institution at INRAE-Toulouse, in the period August-September, 2020. Due to the COVID-19 pandemic, this second secondment was not possible to realize. Instead, bio-assays are currently running at Keygene N.V. and are planned to be finalized by December 2020. Ultimately, possible pre-breeding material coming from this bio-assay, will be used in breeding programs to ensure a more sustainable resistance to O. cumana and consequently maintain the production of vegetable oil from sunflowers. The expected impact is significant since the pre-breeding material would be the first non-GM durable-resistant sunflower lines to O. cumana which also provides a crucial spotlight on the European market.
