The last two decades have been a fascinating time in the field of imaging in biology. After going through an unnecessary competition between the new and fashionable confocal microscopy and the old and complicated electron microscopy, a few scientists managed to convince their peers that one should benefit from combining both of these extremely powerful imaging modalities to take advantage of what they can offer the best. CLEM therefore became a method of choice, and one can witness today that what it offers goes beyond a simple addition of LM and EM. Indeed, by bridging high-end technologies on both sides of the resolution spectrum, CLEM is now a powerful tool for structure–function studies and will most certainly continue to serve as a valuable tool in biology to link the molecular biology with physiology.

The purpose in CLEM experiments is to combine the advantages of both imaging modalities. LM on its side offers a wide field of observation that can be adapted to a large variety of sample size, from cultured cells to full organisms. With the possibility to operate in open atmosphere, the microscopist can control the environment of his/her specimen even for long periods of time. As a consequence, the study of living samples is possible. We are now at the apogee of the use of Fluorescent Proteins (FP)-tagged proteins, which permits the fine localization of proteins in living cells or organisms. Moreover, with the breakthrough of functional imaging technologies, light microscopes can be used to monitor ion homeostasis (Ca²⁺, H⁺, K⁺, Cl⁻), voltage variations of plasma membranes, molecules diffusion in subcellular compartments, and even protein–protein interactions, conformational switch, or enzymatic activity. With the super-resolution light microscopy, a fine localization of discrete object or rare events is now possible.

Electron microscopy, on the other hand, offers high-resolution imaging of the cell ultrastructure. Affinity labeling allows a precise localization of molecules at the subcellular level; electron tomography, combined with powerful computing and image analysis, gives the possibility to achieve a close to atomic view of macromolecules in situ; serial imaging with transmission EM (TEM) or with scanning EM (SEM) and the so-called “slice and view” strategy can now render cell and tissue volumes with great precision. Moreover, the imaging conditions and processes for preparing the samples are extremely conservative and can preserve the full cellular integrity. As a consequence, the microscopist has access to a very complete and complex amount of data, in complement to the object of interest in the focus of his/her particular research. The challenge is now to compile this data in a meaningful correlation, and great efforts are employed worldwide to cover the entire resolution spectrum on individual samples.

There are different ways of defining CLEM, but it generally refers to those techniques that focus the analysis in LM and in EM on the same specimen. When the EM data acquisition can be considered as the final step in the process, CLEM can thus be defined according to the step, in the sample preparation process, at which the LM is performed (see Fig. 1).

The first successful attempts to correlate fluorescent signals with ultrastructure consisted in performing immunofluorescent labeling on resin sections and then immunogold fine localization on the same or on an adjacent section [1, 2]. The fluorescent signal is then used to screen large areas of the sample, to then focus on the given regions of interest. The same strategies have been used recently, with a dramatic improvement in the precision of object relocation, either by a fiducial-based targeting [3] or by the use of super-resolution microscopy [4]. Interestingly, these authors have been recording the specimen fluorescence that has been preserved during the sample preparation process and were therefore able to image the fluorescent signal directly from the sections prepared for TEM. The overlap with the ultrastructure was then accurate enough to focus on the ultrastructure of the pointed region/object of interest.

When the LM recordings are performed on living samples, one of the biggest challenges for correlation is to track the position of the object of interest throughout the sample preparation procedure for electron microscopy. Besides the sample morphology changes that are often associated to this preparation, finding back one object in its environment is very often referred as trying to find a needle in a haystack. Different strategies have been employed to perform this correlation on adherent cultured cells, and they all share the implementation of coordinates or landmarks that allow the registration of the precise location of the targeted cell or signal. The important point in these techniques is the preservation of these space cues throughout the sample preparation steps, in order to clearly see them on the polymerized block faces for precise trimming and sectioning.

Already in the late 1960s, successful attempts have been made to perform CLEM on cultured cells and to correlate live cell imaging with ultrastructural analysis [5, 6]. The strategy to locate the position of a single cell of interest was to mark the coverslip with a diamond scoring system mounted on the microscope (see ZH Price [6]). The procedure consisted in two marking steps, the first one scoring the non-cell-bearing side of the glass while the cells were still growing and being recorded by LM and the second step scoring the cell-bearing side after fixation. This last step ensured the preservation of the marks and therefore of the position of the cell of interest after the embedding step and subsequent removal of the culture substrate. A similar strategy was described recently, where a pulsed laser was used to mark and etch the surface of a glass coverslip bearing cultured cells [8].

Another way of recording precise positions is to add coordinates to the culture substrate before plating the cells. This has been first developed by acid etching of the glass surface [9]. Later, similar commercial solutions were offered with CELLocate dishes (Eppendorf—registered patent in 1993; see ref. [10]) and now with gridded Mattek or Ibidi culture dishes and coverslips. With this approach, the surface on which the cells are growing presents an embossed pattern that is transferred to the block face after polymerization of the resin. This approach was successfully and elegantly performed on frozen samples for immunogold analysis using the Tokuyasu technique [11].

Deposition of coordinates to the culture substrate by carbon evaporation also gave satisfying results in many applications. Substrates such as glass coverslips [9], Aclar [12], sapphire [13], and Formvar films [14] have been used. Finally, metal grids, such as EM grids, carrying intrinsic coordinates that have been used either for direct cell culture [15, 16] or in addition to cell-bearing sapphire discs [17].

When high-pressure freezing (HPF) is chosen as a fixation method, a few strategies have been proposed to add coordinates to cultured cells. They consist of carbon deposition on the substrate, addition of a metal grid prior to the fixation or engraving of coordinates either by hand with a fine needle or by computer-driven etching methods using either a laser or a focused ion beam [12, 13, 17, 18]. Depending on the high-pressure-freezing machine, etched sapphire discs are now commercially available and can be used efficiently in CLEM experiments.

In this chapter, we describe a strategy that was developed for CLEM on cultured cells, to be fixed by high-pressure freezing. This method has been successfully used for various cell types and applications [18–20] and allowed us to efficiently correlate fluorescent microscopy data from living cells with ultrastructural analysis.