[5 Visualizing Intracellular Events In Vivo by Combined Video Fluorescence and 3D Electron Microscopy

By Alexander A. Mironov, Galina V. Beznoussenko, Alberto Luini, and Roman S. Polishchuk


The combination of the capability of in vivo fluorescence video microscopy with the power of resolution of electron microscopy (EM) has been described. This approach is based on such an association of two techniques. An individual intracellular structure can be monitored in vivo, typically through the use of markers fused with green fluorescent protein (GFP), and a ''snapshot'' of its three-dimensional (3-D) ultrastructure and especially tomographic reconstruction can then be taken at any chosen time during its life cycle. The pitfalls and potential of this approach are discussed.


Many cellular functions are very fast and some events within these functions could be extremely rare. Rapid translocations or shape changes of specific intracellular organelles include intracellular traffic, cytokinesis, and cell migration, among others. To understand how such functions, that is, a budding transport carrier, an elongation of microtubule or a developing mitotic spindle, fusion and subsequent detachment of two membrane compartments, are organized and executed in vivo, it is necessary to apply the degree of spatial resolution afforded by EM. However, routine EM deals with an average image of the structure of interest. Moreover, EM usually examines only sections, which do not allow 3-D view. Thus, one should initially observe the dynamic structures in real time in living cells and then examine the same structures using EM combined with tomography.

The most suitable methodology to achieve this is conceptually simple, yet powerful; we refer to this as correlative video-light EM (CVLEM), by which observations of the in vivo dynamics and ultrastructure of intracel-lular objects can indeed be combined to achieve this result. We also illustrate here the kinds of questions that the CVLEM approach was designed to address, as well as the particular know-how that is important for the successful application of this technique.

Correlative light-EM (CLEM) was developed several years ago, and it has been used in cases in which the analysis of immunofluorescently labeled structures need a better-than-light-microscopy resolution (Powell et al, 1998; Svitkina and Borisy, 1998; Tokuyasu and Maher, 1987). CLEM can also be combined with microinjection (Kweon et al., 2004). Despite its potential, CLEM has not been used very often in the past, probably because the ability to correlate two static images, one fluorescent and one under EM, is of interest only in a limited number of situations, and in particular for the examination of the cell cytoskeleton (Svitkina et al., 2003). CLEM has also been used to characterize the structures that are formed by cells to facilitate the degradation of the extracellular matrix (Baldassarre et al., 2003), and to confirm the direct fusion of an endosome and a lysosome (Bright et al., 2005).

However, the most important gains from CLEM come, we believe, from its combination with the kind of dynamic observations obtainable from GFP video microscopy in living cells (i.e., from its use in CVLEM). The CVLEM approach is potentially valuable in any area of cell biology where the elucidation of the 3-D ultrastructures of individual dynamic cellular objects at times of choice can be informative (Mironov et al., 2000; Polishchuk et al., 2000). For example, the growth of a subset of microtubules can be visualized in vivo (Perez et al., 1999). By CVLEM, it is now possible to study at EM resolution the environment and the interactions of the microtubule tips with other cytoskeletal elements, or intra-cellular organelles, at various stages of the tip growth. In addition to cytoskeletal dynamics, other fields where the application of CVLEM can be easily imagined are those of cell division and cell-cell interactions (although the specific questions to be addressed are best left to the specialists). CVLEM has been successfully applied to the characterization of the ultrastructure of membrane carriers transporting secretory proteins to (Marra et al., 2001; Mironov et al., 2003), through (Mironov et al., 2001), and from (Polishchuk et al., 2000, 2004) the Golgi complex, to the analysis of endocytic structures (Caplan et al., 2002), and to the morphology of the Golgi complex in mitosis (Altan-Bonnet et al., 2003).

One limitation, and at the same time, attraction, of CVLEM, is its complexity. The use of this technique is demanding, and to master its various steps requires a whole array of skills. However, microscopy is developing quickly both in the field of living-cell imaging and in the field of EM, and new powerful technologies are rapidly becoming available in more user-friendly versions. This should make the use of CVLEM more appealing to a number of cell biologists. Through the use of fluorescent proteins of different colors, several marker proteins can now be observed simultaneously (Ellenberg et al., 1999). This will allow the analysis of the interactions between different organelles and organelle subdomains (Ellenberg et al., 1999; Pollok and Heim, 1999). Combining these and other methods with the quickly developing EM tomography (Ladinsky et al., 1994) will increase the subtlety and range of the questions that can be answered by the CVLEM approach. While we await the microscopy of the future, which it is hoped will be endowed with the ''magical'' power to show cellular structures with EM resolution in vivo in real time, CVLEM offers a useful chance to look deeper inside living cells.

As illustrated in Figs. 1 and 2, the CVLEM procedure includes several stages: (1) observation of the structures labeled with green fluorescent protein (GFP) in living cells (Figs. 1A, 2A); (2) immuno-labeling and embedding for EM (Fig. 1B); (3) identification of the cell on the resin block and cutting of serial sections (Fig. 1C); (4) EM analysis and structure identification (Figs. 1D, 2B-E); and (5) digital 3-D reconstruction of the structure of interest (Figs. 1E, 2F). During the first step, the cells are transfected with the cDNA encoding the GFP fusion protein of choice and the fluorescence of the associated structures in living cells is followed (Lippincott-Schwartz and Smith, 1997). In this way, it is possible to gain information about the dynamic properties of these structures (i.e., motility speed and direction, changes in size and shape, etc.). At the end of this stage, the cells are killed by the addition of fixative, capturing the fluorescent object at the moment of interest. As GFP is not visible under EM, immuno-staining allows for the identification of the GFP-labeled structure at the EM level. The immuno-gold and immuno-peroxidase protocols to perform this staining for EM are described below. Usually, the immunogold protocol (Burry et al., 1992) is suitable for the labeling of the large majority of antigens, while immuno-peroxidase only allows the labeling of antigens that reside within small membrane-enclosed compartments, because the electron-dense product of the peroxidase reaction tends to diffuse from the actual location of the antibody binding (Brown and Farquhar, 1989; Deerinck et al., 1994). Once stained, the cells must be prepared for EM by traditional epoxy embedding, and the cell and structure of interest must be identified in sections under EM. The finding of individual subcellular structures in single thin sections can be complex, and sometimes even impossible, simply because most of the cellular organelles are bigger than the thickness of the section and lie along a plane that is different from that of the section. So an analysis of serial sections from the whole cell is required for the identification of the structure(s) observed previously in vivo. An example of this identification is given in Fig. 2A-E. Finally, the EM analysis of serial sections can be supported by high-voltage EM tomography and/or digital 3-D reconstruction (Fig. 2F).

A GFP-based time-lapse confocal microscopy followed by chemical fixation.

A GFP-based time-lapse confocal microscopy followed by chemical fixation.

B lmmunoperoxidase labeling and embedding in resin.
in sections for EM.

E Digital 3D reconstruction.

Fig. 1. The main steps in the CVLEM procedure. A. The structure of interest (circled; in this case a transport carrier) is monitored in vivo using an appropriate marker tagged with GFP and time-lapse confocal microscopy. The cell is then fixed at a time chosen by the

Method Steps

Observation and Fixation of Living Cells Materials

Cells of interest, DNA and transfection reagents;

MatTek petri dishes with CELLocate cover slip (MatTek Corporation, Ashland, MA);

HEPES buffer (0.2 M). Dissolve 4.77 g HEPES in 100 ml distilled water and add 1 N HCl to provide a pH of ^7.2-7.4;

Fixative (0.1% glutaraldehyde-8% paraformaldehyde). Dissolve 8 g paraformaldehyde powder in 50 ml HEPES buffer, stirring and heating the solution to 60°. Add drops of 1 N NaOH to clarify the solution. Add 1.25 ml 8% glutaraldehyde and 50 ml HEPES buffer;

4% paraformaldehyde. Dissolve 4 g paraformaldehyde powder in 100 ml HEPES buffer, stirring and heating the solution to 60°. Add drops of 1 N NaOH to clarify the solution. The cells are plated for CVLEM on MatTek petri dishes that have CELLocate cover slips attached to their base. The CELLocate cover slips have etched grids with coordinates that allow the cells of interest to be found easily through all of the steps of the procedure. Transfect the cells with the cDNA of the GFP fusion protein of choice using any method available in your laboratory. As illustrated in Fig. 3, after a transfected cell has been chosen and located on the CELLocate grid, its position is drawn on the map of the CELLocate grid (available from MatTek; see the example of the map in Fig. 3A). This living cell can then be observed for the GFP-labeled structures using confocal or light microscopy, which allows the grabbing of a time-lapse series of images by a computer. At the moment of interest, the fixative is added to the cell culture medium while still grabbing images (fixative: medium volume ratio of 1:1). The fixation usually induces the fast fading of the GFP fluorescence and blocks the motion of the labeled structures in the cells. Particular attention must also experimenter (e.g., during budding, translocation, or fusion; in this case during translocation). B. The cell is labeled by the immuno-HRP technique with an antibody against the GFP-tagged protein marker. The patterns of peroxidase labeling and of GFP fluorescence must coincide. C. The cell in A and B is identified in the resin block by using the system of spatial coordinates (also described in Polishchuk et al., 2000), and serial sections are cut. D. The carrier previously monitored in vivo (circled) is identified in each section by using specific cellular structures as spatial landmarks (see example in Fig. 2). E. The images in the serial sections are used for the 3-D computer-aided reconstruction.

0 0

Post a comment