Résumé
Infections of banana and plantain by banana streak viruses (BSV) can occur in the absence of mealybug-mediated transmission, through the activation of infective endogenous BSV sequences (eBSVs). Such infective eBSVs are present in the genome of Musa balbisiana spp, which are important progenitors for breeding improved banana varieties. Once activated by biotic or abiotic stresses, these viral integrants cause spontaneous and systemic infection in both natural and synthetic interspecific hybrids harbouring the M. balbisiana (B) genome. Therefore, the presence of infective eBSVs within B genomes became the main constraint for banana and plantain breeding programs and for exchanging Musa germplasm. A thorough characterisation of eBSV was first perfomed in the seedy diploid M. balbisiana genitor Pisang Klutuk Wulung (PKW) where three BSV species were unearth: Obino l'Ewaï (BSOLV), Goldfinger (BSGFV) and Imové (BSIMV) species. This organisation is simpler in plantains where we only have integrated viral sequences of the species BSOLV and BSGFV. To lift the eBSV constraint, we have developed in plantain, a genome editing approach based on crispr/cas9 technology. We designed crispr to carry out partial or full deletions of both eBSV in order to inactivate or to delete them respectively. We have performed the proof of concept in protoplastic system in collaboration with IAGE, a startup hosted at CIRAD, and we are currently collaborating with IITA in Kenya to release two different varieties of plantain with edited eBSV.