Development of transient gene silencing tools in grapevine

Establishment of functional genomics in Vitis ssp. is still challenging, due to the lack of reliable high-throughput tools for grapevine transformation. For instance, A. tumefaciens-mediated stable transformation is possible, but it is time consuming and the transformation efficiency is rather low. The aim of this work is the development of four different post-transcriptional gene silencing strategies for functional genomics in Vitis riparia cv. Gloire de Montpellier for future studies on the function of genes putatively related to pathogen resistance, being this species resistant to many pathogens such as Plasmopara viticola. The first strategy is based on the direct administration of small interfering RNAs (siRNAs), different in length, and dsRNAs to V. riparia and N. benthamiana leaves, to test their effectiveness in downregulation of target genes. In particular, different populations of short RNAs (sRNAs) against the PR-10 gene family were supplied to V. riparia leaves while others designed against the phytoene desaturase gene (PDS) were administered to N. benthamiana leaves. In addition, siRNAs or 2’-O-methylated synthetic siRNAs against the Green Fluorescent Protein (GFP) gene were delivered to N. benthamiana plants stably transformed with the GFP marker gene (GFP16C plants), to test their activity in gene silencing induction against this gene. Unfortunately, results obtained by direct delivery of different small RNAs to V. riparia and N. benthamiana plants were controversial and not easily interpretable. More evidences are still needed to determine the role of siRNA characteristics and concentration, and to find the best delivery strategy for the induction of post-transcriptional gene silencing in plants. The second approach aimed to test whether the silencing signal can spread systemically from grapevine hairy roots to the aerial parts of the plant, resulting in the downregulation of the target gene. Some systemic spreading of the silencing signal to the non-transgenic shoots of transformed plants was already observed in Lotus japonica, Medicago truncatula and Arabidopsis thaliana plants, though the silencing in the leaves was not complete. V. vinifera hairy roots have been obtained by A. rhizogenes infection; the ability of the silencing signal to spread to shoots has not been investigated in grape. In our study, V. riparia leaves, cuttings and plantlets were infected with A. rhizogenes ARqua1, transformed with a pRedRoot vector harbouring a marker gene encoding the red fluorescence protein (DsRed1), to find the optimal transformation conditions and to evaluate V. riparia susceptibility. Then, V. riparia infections were carried out using A. rhizogenes strain ARqua1 transformed with a pRedRoot vector containing a 35S hairpin construct against the grapevine PDS gene to evaluate possible plant phenotypic alterations. iii The third approach is based on transient transformation of V. riparia leaves through syringemediated vacuum application. Preliminary experiments were conducted using the A. tumefaciens EHA105 strain transformed either with a pBin19 vector harbouring a hairpin sequence against the Vitis riparia PDS gene (35S:PDShp) or with an empty pBin19 vector, as a negative control. Severe mechanical damage was observed in all infiltrated leaves; to evaluate if this damage was related to the Vitis genotype or to the technique, agroinfiltrations were performed in leaves and plantlets of different V. vinifera cultivars. Two different A. tumefaciens strains (EHA105 and C58C1) transformed with pBin61 vector carrying a GUSi expression cassette (35S:GUSi) were used for transient transformation by syringe or pump-mediated vacuum application. This experiment permitted the evaluation of the optimal transformation conditions, difference in efficiency between the bacterial strains and Vitis susceptibility. The last approach consists in the development of a VIGS system using a GALV (Grapevine algerian latent virus)-based vector. Preliminary infections carried out with in vitro-produced viral RNAs revealed that GALV infective transcripts could infect and spread systemically. Considering that the procedure of infection with in vitro transcripts was not reliable enough due to low infection efficiency we decided to consider the standardization of viral delivery to the plants by agroinfection, producing a true VIGS vector. The molecular manipulation strategy for the production of a GALV-based VIGS vector was then developed. All these four strategies contributed to the pool of available in vivo tools for functional genomics of the valuable grapevine crop. They also opened several exciting research avenues to pursue in the near future.