3D HDO-CLEM: cellular compartment analysis by correlative light-electron microscopy on cryosections.

Fundamental to obtaining a depth-understanding of the function and structure of cells is the ability to study and correlate their molecular topography with the ultrastructural morphology, for example, to visualize the position of a given protein relative to a given cell compartment and its morphology. Standard fluorescence light microscopy (FLM) relies on simple sample preparations, and localizes proteins in living or fixed cells with a resolution in the range of few hundred nanometers, allowing large field of view. However, FLM is unable to visualize the unlabeled cellular context. On the other hand, electron microscopy (EM) techniques reveal protein topology with the resolution in a range of a few tens of nanometer, retains the cellular context, but can only be applied on a limited field of view. Therefore, both approaches present shortcomings, in terms of field of view, statistical output, resolution, sample preparation, and context analysis, that can likely complement each other. To bridge the gap between FLM imaging and EM, several laboratories have developed methods for correlative light-electron microscopy (CLEM). In a nutshell, CLEM enables one to investigate the same exact region of interest utilizing the two microscope platforms, and thereby virtually combine their capabilities. We describe a protocol based on immunolabeling of Tokuyasu cryosections that allows correlation of LM and EM images with excellent preservation of cellular ultrastructure. We will refer to this method as high-data-output CLEM (HDO-CLEM). The major benefits of HDO-CLEM are the possibility to (1) correlate several hundreds of events at the same time, (2) perform three-dimensional (3D) correlation, (3) immunolabel both endogenous and recombinantly tagged proteins at the same time, and (4) combine the high data analysis capability of FLM with the high precision of transmission EM in a CLEM hybrid morphometric analysis. We have identified and optimized critical steps in sample preparation, defined routines for sample analysis and retracing of regions of interest, developed software for semi/fully automatic 3D FLM reconstruction and set the basis for a hybrid light/EM morphometry approach