Abstract
Low macroelement availability in most cultivated soils severely limits crop yields in the absence of fertilization. A better understanding of the adaptation of root systems to nutrient-poor soils, and the exploitation of existing genetic diversity in this field, between species and/or varieties, are likely to contribute to the development of new cultivars and new agronomic practices allowing to limit costly and environmentally polluting chemical fertilization inputs. The architecture of the root system and the production of root hairs at the root-soil interface are major determinants of the capacity of the root system to explore the soil and take up nutrient ions. To date, no methodology has been available to phenotype root hairs in a root system considered entirely. In this thesis, I developed a methodology for global, integrative phenotyping of root systems, including root hairs. An original rhizobox-type device was developed, allowing to acquire high resolution images, for which I developed a computerized analysis procedure associating the free software Ilastik for image segmentation, and the softwares WinRHIZOTM and ImageJ for the analysis of global traits characterizing the root development. After validation of the methodology, the root systems of two wheat genotypes, a cultivated emmer wheat cultivar (T.t. dicoccum, cv Escandia), ancestor of durum wheat, and a landrace of durum wheat (T.t. durum, cv Oued Zenati) were compared with each other and with respect to their response to low phosphate (P), nitrogen (N) or potassium (K) availability. In 15-day-old seedlings (roots ca. 30 cm long), N, P or K deficiencies differentially affected plant growth (biomass allocation between roots and leaves, and preferential development of the root system). All three deficiencies were found to result in an increase in the total surface area of the root system, resulting primarily from an increase in the total surface area of root hairs over the entire root system (reflecting an increase in the density and/or length of hairs over the entire system). The rate of increase in total absorptive root hair area was variable between the two varieties and among limiting elements, stronger under N deficiency conditions in the emmer wheat, and P deficiency in the landrace. All the root responses analyzed, including or not the root hairs, revealed a greater developmental plasticity in response to nutrient deficiency in the ancestral variety. A perspective opened by this work would be to compare this plasticity in different wheat varieties recapitulating the domestication and improvement of this species. I also show that the methodology I have developed can be used to phenotype root responses to biotic conditions (presence of Plant Growth Promoting Rhizobacteria).