ABSTRACT The success of land plants can be attributed to the evolution of beneficial associations between plant roots and soil microbes. Root-microbe mutualisms extend the range of plant nutrient acquisition delivered through the hyphal network of mycorrhiza, an ancient and widespread plant symbiosis, or by the more recent adaptive innovation of nitrogen-fixing nodule symbioses. A plant’s genetic toolkit governs its selection of beneficial symbionts and the developmental extent of these intimate interactions. However, the evolutionary origins and function for only a few symbiotic signaling components have been explored. The central aim of this dissertation is to resolve the evolutionary events that contributed two, novel genetic components for establishing root symbioses, NPF1B and NPF1C. The Medicago truncatula (Mt) LATD/NIP/NPF1.7C transporter functions in root and nodule meristems and is a member of the large NPF1 gene subfamily. Here, I propose that LATD/NIP’s role in establishing nitrogen-fixing symbioses is derived from the ancient mycorrhizal signaling pathway. I used a comparative phylogenomic approach to investigate the evolutionary origins of the NPF1 gene across flowering plants and then asked whether diversifying or purifying selection forces influenced NPF1 gene retention. I postulated that such gene retention correlates with the adaptive traits of mycorrhizal or nitrogen-fixing root nodule symbiosis; to test this I measured trait correlation within my dataset. I found that the NPF1 phylogeny is comprised of five well-supported angiosperm clades, A, B, C, D1 and D2, that arose by successive duplications and have unequal gene retention. NPF1B is present as a single copy gene or lost entirely, while the other major NPF1 clades expanded to multiple genes within angiosperms. The NPF1A, B and C genes are under strong purifying selection while the NPF1D genes display positive, diversifying selection. My data revealed a statistically significant correlation of NPF1A, B, C, and D2, but not NPF1D1, gene retention with the ability of a species to form mycorrhizal associations. Additionally, the retention of the NPF1B, C, D1, D2, but not NPF1A, genes within a species is statistically correlated with its ability to form nitrogen-fixing symbiosis. Supporting this correlation, NPF1B genes are expressed in plant root tissues with and without mycorrhizal fungi yet available datasets failed to detect NPF1B expression in nodule tissues whereas the NPF1C genes are expressed in both symbiotic and non-symbiotic plant root tissues. In support of functional conservation, expression of legume LATD/NIP cDNAs from Cicer arietinum (Ca) and Lotus japonicus (Lj) restored, in part, the root and nodule defects of the Mtlatd mutant and resulted in the formation of peculiar hybrid lateral root-nodule structures while, in wild-type M. truncatula, significantly augmented root development. In L. japonicus, the disruption of LATD/NIP alters the number of lateral roots and nodules My thesis data support the hypothesis for an ancestral NPF1 gene function in establishing mycorrhizal associations in angiosperms and, consequent to the monocot-eudicot divergence, co-opted this function for accommodating nitrogen-fixing symbioses in eudicots. Successive duplications then yielded the NPF1B and NPF1C genes that, by neofunctionalization and natural selection, further refined their roles in root organogenesis and symbiosis; a prerequisite for the evolution of nodule organs.

L'azote est essentiel au développement de toutes les cellules vivantes. Il est un des facteurs limitant de la croissance végétale. La seule source d'azote abondante est l'atmosphère contenant 80 % de diazote mais cette forme n'est assimilable que par certains procaryotes. Ces microorganismes sont capables de fixer l'azote atmosphérique sous leur forme libre ou en symbiose avec des plantes. Ainsi, ils fournissent à leur plante partenaire des substrats azotés, sous forme d'ammoniaque, tandis qu'en retour celle-ci fournit à la bactérie des substrats carbonés issus de sa photosynthèse. Il s'agit d'une association à bénéfices réciproques. Il existe deux grands types de symbiose fixatrice d'azote : la symbiose rhizobienne, impliquant diverses Protéobactéries et la symbiose actinorhizienne impliquant une Actinobactérie, Frankia. Les bactéries pénètrent les cellules des plantes pour former un nouvel organe, la nodosité dans laquelle va avoir lieu la fixation d'azote. Les bases moléculaires à l'origine de la symbiose rhizobienne sont très bien caractérisées tandis que celles de la symbiose actinorhizienne restent en grande partie inconnue, de par l'absence d'outils génétiques. Toutefois, les premières étapes de mise en place de la symbiose présentent des similarités. Les deux bactéries sont capables d'induire la déformation du poil racinaire en sécrétant un facteur déformant, le facteur Nod pour la plupart des symbioses rhizobiennes et un facteur encore non caractérisé dans le cas de la symbiose actinorhizienne. La problématique de mes travaux de thèse est de savoir si le dialogue moléculaire s'établissant entre la plante et la bactérie est basé sur des composants universels. Ce travail a utilisé deux approches. Une approche ciblée visait à mettre en évidence la fonction. Une approche non-ciblée par le biais des puces transcriptomiques chez Frankia a permis de comparer l'expression génétique entre des conditions de vie libre et des conditions de vie symbiotique. Enfin, une dernière approche a concerné les composés aromatiques chez Frankia. Il s'agissait d'établir si Frankia était capable de cataboliser différents composés aromatiques. En effet, beaucoup d'entre eux sont impliqués dans les interactions plante-bactérie, notamment dans les réactions de défense de la plante / Nitrogen is essential for cells development. It's one of the limiting factors of plant growth. The only abundant source of this component is the atmosphere which contains 80 % of dinitrogen, but this form can only be assimilated by some prokaryotes. These microorganisms are able to fix atmospheric nitrogen under freeliving condition or in symbiosis with some plants. Thus, they provide nitrogen substrates to the plant in the form of ammonium, and in return the plant provides carbon substrates from photosynthesis. It is an association with reciprocal profits for both partners. There are two major nitrogen-fixing symbioses: rhizobial symbiosis, which involves various Proteobacteria and actinorhizal symbiosis, which involves the Actinobacterium, Frankia. Bacteria enter plant root cells and develop a new organ, the nodule where nitrogen fixation takes place. Molecular bases are well characterized for rhizobial symbiosis, whereas little is known about the actinorhizal symbiosis. This fact is in part due to absence of genetic tools for Frankia. However, early steps of the interaction show some similarities. These two bacteria are able to induce root hair deformation by secreting a deforming factor, Nod factor in most rhizobial symbioses and a noncharacterized factor in the actinorhizal symbiosis. The aim of this thesis was to determine if molecular dialogue between plant and bacteria is based on universal components. This work used two approaches. One was targeted on nodC-like gene from Frankia alni ACN14a. We tried to characterize their function. Another used trancriptomic microarrays in Frankia. This technique allowed us to compare transcripts from 2 conditions: free-living cells and symbiosis. A last approach focused on aromatic compounds in Frankia. We wanted to determine if Frankia was able to use different aromatic compounds to grow. Indeed, a lot of aromatic compounds are involved in plant-bacteria interaction such as plant defense