Rhodotorula (Rhodosporidium) toruloides is an oleaginous basidiomycete yeast with great biotechnological potential. Capable of accumulating lipid up to 76 % of its dry biomass and well suited to the metabolism of lignocellulosic hydrolysate, it is a good candidate for production of advanced biofuels as well as a host of other potential roles in industry. However, molecular genetic tools for manipulation of this yeast are lacking and its high genomic GC content can make routine cloning difficult. Agrobacterium tumefaciens-mediated transformation of R. toruloides CBS 14 was demonstrated, and plasmid vectors were developed for transformation of R. toruloides, including elements for Saccharomyces cerevisiae in-yeast assembly. In-yeast assembly is robust to the manipulation of GC-rich DNA and of large plasmids. Using these vectors and an EGFP reporter, a screen to identify inducible promoters was performed, and promoters from the genes NAR1, ICL1, CTR3, and MET16 identified. These promoters have independent induction/repression conditions and different levels and rates of induction. Minimal inducible promoters were determined, which are as small as 200 bp. As well as showing tight regulation of the EGFP marker, the NAR1 promoter was able to drive conditional rescue of a leu2 mutant strain. In parallel, as a proof of principle for production of advanced biofuels, hydrocarbon biosynthesis pathways were expressed in R. toruloides and analysed by GC-MS. After co-expression of Synechococcus elongatus fatty acyl-ACP reductase and fatty aldehyde decarbonylase, and E. coli ferredoxin and ferredoxin reductase, production of the alkane heptadecane was observed. To increase the availability of free fatty acids (FFA) for production of hydrocarbons by other pathways, Thermomyces lanuginosus lipase 2 was expressed, resulting in a 1.3-fold increase in the concentration of FFAs.

Phakopsora pachyrhizi Syd.& P.Syd. est le plus important fléau du soja (Glycine max (L.) Merrill). Introduit au Brésil dans les années 2000, ce champignon s’est rapidement répandu sur les deux continents Américains. Seule l’utilisation de fongicides associée à des pratiques culturales strictes permet de maintenir le niveau de production. L’utilisation répétitive de ces fongicides ainsi que la plasticité génétique de ce champignon ont rapidement entraîné une diminution d’efficacité de certaines molécules. Par ailleurs, la plupart des résistances verticales identifiées dans les ressources naturelles du soja restent inefficaces contre certains isolats du champignon. La compréhension des mécanismes de l’immunité des plantes permet de proposer des solutions biotechnologiques pour le contrôle des maladies. L’utilisation antérieure du système barnase/barstar induisant une mort cellulaire artificielle, a permis de générer des pommes de terre résistantes à Phytophtora infestans. Cette technologie est basée sur l’expression de la barnase une ribonucléase toxique pour les cellules, et la barstar un inhibiteur de la barnase. Il a été proposé d’évaluer ce système pour le contrôle de P. pachyrhizi. Le point critique de cette approche est de trouver le bon rapport de l’expression des gènes barnase/barstar. Pour ce faire la barnase sera placée sous le contrôle d’un promoteur induit par le pathogène, permettant une régulation spatiotemporelle. La recherche de tels promoteurs a été effectuée en utilisant des données transcriptomiques et bibliographiques. Des sojas stables exprimant les différentes fusions promoteur:GFP ont été créées afin d’étudier l’ expression spatiotemporelle de ces promoteurs en présence du champignon. Les promoteurs pGmCHIT1 (de G. max) et pgst1 (de Solanum tuberosum) contrôlant respectivement l’expression d’une chitinase et d’une glutathione-S-transférase ont été identifiés comme induits par le pathogène. L’impact de différents stress sur ces deux promoteurs a été évalué. Les constructions génétiques « barnase/barstar » comprenant les différentes combinaisons des promoteurs ont été générées. Nicotiana benthamiana a été utilisé pour exprimer transitoirement les construits et évaluer leur phytotoxicité en absence du pathogène. Un seul construit contenant le promoteur gst1 s’est avéré non phytotoxique. Il a été transféré avec succès dans le soja. Ces sojas n’ont pas montré de gain de tolérance à la rouille. Une proposition d’amélioration du système barnase/barstar est discutée afin de mieux cerner les possibilités et les limites de ce système pour le contrôle de la rouille du soja / Phakopsora pachyrhizi Syd.& P.Syd, the fungus responsible for Asian soybean rust, is the most devastating soybean (Glycine max (L.) Merrill) pathogen. First observed in the 2000s in Latin America, the pathogen has spread throughout the Americas. The control of this pathogen depends on the use of fungicides and strict agricultural practices. The repetitive use of the 3 classes of fungicides and the genome plasticity of the pathogen have led to a decreased efficacy of certain molecules. Although vertical resistance genes have been mapped in the soybean germplasm, most of them are not effective against all Asian soybean rust isolates. A deeper understanding of plant immunity facilitates the development of biotechnological approaches for plant disease control. Artificial cell death was previously developed to control Phytophthora infestans development in potato. The technology was based on a barnase ribonuclease that is highly toxic to the plant cell and that consequently needed to be expressed only in the presence of the pathogen. The lethal expression of barnase was counterbalanced by barstar, a highly specific inhibitor of barnase. We propose to evaluate this technology in soybean to control P. pachyrhizi. The key objective is the modulation of the ratio of barnase/barstar based on the identification of an adequate inducible promoter to control the expression of barnase. The previous literature and transcriptomic data were used to identify candidate promoters for barnase expression. Stable transgenic soybean expressing the different promoter:GFP fusions were generated to test the spatiotemporal activity of the promoters in the presence of the pathogen. pGmCHIT1 (from G. max) and pgst1 (from Solanum tuberosum) promoters controlling a chitinase and a glutathione-Stransferase, respectively, were identified as induced by soybean rust. The impacts of different stresses on these promoters were evaluated. Molecular constructs with different promoters driving the barnase and barstar gene combination were generated. Nicotiana benthamiana was used to evaluate construct toxicity in the absence of the pathogen. One single construct containing the promoter pgst1 was shown to be non-phytotoxic. This construct was successfully introduced in soybean plants. The generated soybeans were challenged with rust, but no protection was observed. Based on these results, we discuss how to improve the barnase/barstar system to control soybean rust