Combining flow-MRI and modeling approach to assess sap flux in tomato plant architecture in response to water deficit
Variations of tomato fruit weight and composition throughout the cycle of production make the management of fruit yield and quality complex. These variations are linked to the fluctuation of water and carbon available for fruit growth and depend on the genotype and environmental conditions. Some structural functional plant models can predict the concentration of water and carbon in the plant architecture and the consequences on fruit growth. Obtaining measures of sap fluxes in situ is difficult. Evaluating and validating these models to understand fruit growth relations in a fluctuating environment is our goal. For that, many techniques exist but only a few non-invasive approaches are feasible. Nuclear magnetic resonance imaging (MRI) is a direct and non-invasive method that allows the study of plant water status and sap transport in large potted plants. The objective of my work is to estimate the anatomical and functional properties of the conductive tissues and to quantify the sap fluxes at different levels of the tomato plant architecture, using a combination of approaches: MRI, histological observations, and plant modeling. We perform MRI experiments on an Agilent scanner working at 9.4T using inflow and outflow sensitive spin echo pulse sequences. We use a novel flow-MRI method, taking advantage of inflow slice sensitivity, called the flip-flop method, to measure the water fluxes at different plant levels in response to water deficit. Histological measurements are used to identify the nature of the conductive tissues in which we measure fluxes with MRI. This technique allows us to quantify the surface of conductive tissues, depending on the environment or on the genotype, at the pedicel and stem levels. Then, the measurements found with these two techniques are compared to the predictions of a structural functional tomato plant model. The combination of techniques and the confrontation with the model predictions allow us to obtain quantitative data at the vessel level to improve model predictions.
Jeanne Simon, Unité Plantes et Systèmes de Culture Horticoles, INRA, Domaine St-Paul, Site Agroparc, 84914 Avignon, France, and Laboratoire Charles Coulomb (L2C), Université de Montpellier, CNRS, 34095 Montpellier, France, e-mail: email@example.com
The article is available in Chronica Horticulturae