Space-time Imaging of Organelles and Endomembranes Dynamics

Jean Salamero
Scientific keywords: bioimage informatics, high resolution microscopy, image based modeling., Membrane endocytosis and recycling, organelles biogenesis

Imaging Membrane traffic at the pertinent scale. The group develops a workflow of optics, biophysical and mathematic approaches mainly devoted to high spatio-temporal studies of molecular dynamics and mechanisms in membrane traffic. His main focus is on the quantitative and temporal deciphering of the very late steps of vesicle recycling at the plasma membrane. In this context a particular attention is paid to the coordination of the Rab11A platform (Rab11A/Rab11FIP2/myosinVB)  and its interaction with cytoskeleton elements.

We already adapted or developed diverse photonic approaches aimed to determine the dynamic and topological characterization of Rab domains in conjunction with both a cargo specificity study (Langerine/TfR) and of multi-Rab complexes (Rab11A/RCP/Rab4; Rab6A or A’/Rab6IP1/Rab11A), in living cells. Our work is now fully dedicated to decipher the dynamic coordination and organization of molecular complexes at the single cell level. Dedicated “hybrid” technologies have been developed to perform this task. These techniques (i.e. High res SIM+ Q-scanTIRFM; Boulanger et al. 2014) are used to define the spatial positioning (3D high res) of a molecule or a group of molecules within the living cell at the best temporal regime and we are now classifying the 3D dynamical behavioral of exocytic vesicles, in diverse conditions (Basset et al. TPI, submitted).

Modelling dynamics in single cell biology: toward integrated analysis. Together with the SERPICO Team at Inria, Rennes, we have proposed original formalisms and user-friendly algorithms for intracellular traffic estimation (Network Tomography, image denoising (bias-variance trade-off), space-time object detection and background subtraction). We shall pursue this research in 2 complementary directions:

-From the experimental view point, molecular (RNA interference), mechanical (micro-patterning) and optical (FRAP, photoactivation) perturbations combined with multi-parametric image acquisition (TIRF, 4D, FLIM) allow one to quantify the molecular interactions in the cell. The issue we now wish to address is to link this type of  information with meaningful events and signals as they are detected in space and time in order to evaluate correlations (ICS, RICS, optical flow… [P. Roudot et al., (in preparation])

-Another aspect is to analyse and compare these events using original metrics (e.g. geodesics metrics from valuated Network Tomography-based graphs) and simulation methods (e.g. Monte-Carlo methods, “Cross-Entropy”).We have now built a dynamic model able to statistically mimic observed processes (Pécot et al. 2014 Epub ahead; and Fig. 2). Ideally, the simulation and real data should be “similar” as much as possible, both statistically and visually. From simulations, we determined meaningful features to be matched to real image data. Iteratively, we use description parameters from the simulation to extract statistical information, from larger sets of real data with the ultimate goal to predict intracellular behaviours.