Correlative Video Light/Electron Microscopy This unit describes newly developed methods that allow the examination of living cells by time-lapse analysis with the subsequent identification of the just-observed organelle under an electron microscope.

UNIT 4.8 BASIC PROTOCOL

Many cellular functions, such as intracellular traffic, cytokinesis, and cell migration crucially depend on rapid translocations and/or shape changes of specific intracellular organelles. To understand how such functions are organized and executed in vivo, it is important to observe in real time in living cells such dynamic structures as a budding transport carrier, an elongating microtubule, or a developing mitotic spindle, but to have the degree of spatial resolution afforded by electron microscopy (EM). Most suitable for this is a conceptually simple, yet powerful, method called correlative video light/electron microscopy (CVLEM), by which observations of the in vivo dynamics and the ultrastructure of intracellular objects can indeed be combined to achieve the above-mentioned result. This unit describes this methodology, illustrates the type of questions that the CVLEM approach was designed to address, and discusses the expertise required for successful application of the technique. The CVLEM procedure includes several stages: (1) transfection of living cells with an appropriate green fluorescent protein (GFP) fusion protein, (2) observation of structures labeled with GFP in living cells, (3) fixation, (4) immunolabeling for EM, (5) embedding, (6) identification of the cell in the resin block, (7) sectioning, and (8) EM analysis. During the first step, cells have to be transfected with the cDNA encoding the GFP fusion protein whereby the structure of interest can be discovered in living cells (Lippincott-Schwartz and Smith, 1997). In this way, it is possible to gain information about the structure, including its dynamic properties (i.e., motility, speed and direction, changes in size and shape) and life cycle. At the end of this stage, it is necessary to stabilize the cell structure by addition of a fixative, allowing one to capture the fluorescent object at the moment of interest. As GFP is not visible under an electron microscope, immunostaining is required to identify the GFP-labeled structure at an EM level. This protocol uses immunogold and immunoperoxidase protocols to perform staining for EM. Usually the immunogold protocol (Burry et al., 1992) is suitable for labeling the vast majority of antigens; the immunoperoxidase protocol should be used only to label antigens residing within small membrane-enclosed compartments, because the electron-dense products of the peroxidase reaction tend to diffuse from the place of antibody binding (Brown and Farquhar, 1989; Deerinck et al., 1994). Once stained, the cells must be prepared for EM by traditional epoxy (or other) embedding techniques, and the cell and the structure of interest must be identified in EM sections. Finding an individual subcellular structure in a single thin section can be complex, and sometimes impossible, simply because most of the cellular organelles are bigger than the thickness of a routine EM section and may lie in a plane different from the plane of a random section. Analysis of serial EM sections of the whole cell is thus required to identify the structure previously observed in vivo. An example of such an identification is shown in Figure 4.8.1. Finally, EM analysis of serial sections can be supported by digital three-dimensional serial reconstruction or high-voltage EM tomography.

Microscopy Contributed by Roman S. Polishchuk and Alexander A. Mironov Current Protocols in Cell Biology (2001) 4.8.1-4.8.9 Copyright © 2001 by John Wiley & Sons, Inc.

4.8.1 Supplement 11

Figure 4.8.1 In vivo dynamics and ultrastructure of an individual Golgi-to-plasmalemma carrier (GPC) studied using CVLEM. The GPC in (a) (arrow) was observed in vivo using time-lapse confocal microscopy. After fixation, the cells were immunoperoxidase labeled and embedded in epoxy resin. The structure in (b) (arrow) was identified as the fluorescent GPC shown in panel a by the computer-aided superimposition of the two images using the CELLocate coordinates [an example of which is shown in (c)]. The general pattern of immunoperoxidase labeling in (b) coincided with the fluorescent pattern of vesicular stomatitis virus glycoprotein–GFP (VSVG-GFP) in (a). Serial sections of the cell were then produced (d through f), and the first section displaying the transport intermediate [(g); arrow] was captured at low magnification (f, g). Note that the sections contain structures that are helpful for identification [e.g., protrusion in (c through f); arrowheads]. (h through q) represent a series of consecutive 200-nm sections containing the GPC (arrow). The field shown in (h through l) is the area of the cell identified by a white box in (g). Three-dimensional reconstruction characterizes the GPC as an elongated saccules with a short tubular protrusion (r, s). Other GPCs identified using the same approach appeared either as tubules [(t); arrow] or saccules with short tubules [(u); arrow]. The bar in panel u represents the following lengths in the other panels: 9 µm (a, b), 31 µm (c), 12.2 µm (d-f), 6 µm (g), 2.1 µm (h-l), 700 nm (m-q), 320 nm (r, s), 1.2 µm (t, u).

Correlative Video Light/Electron Microscopy

Materials Cells of interests cDNA encoding an appropriate GFP fusion protein Fixative A (see recipe) Fixative B (see recipe) PBS (APPENDIX 2A) Blocking solution (see recipe) Primary antibody to label structure of interest Monovalent Fab fragments of secondary antibody conjugated with horseradish peroxidase (HRP; Rockland) or Nanogold (Nanoprobes) 1% (w/v) glutaraldehyde in 0.2 M HEPES buffer, pH 7.3

4.8.2 Supplement 11

Current Protocols in Cell Biology

Gold enhancement mixture (see recipe) DAB solution (see recipe) 2% (w/v) OsO4 (Electron Microscopy Sciences) in water 3% (w/v) potassium ferrocyanide in 0.2 M cacodylate buffer 0.2 M cacodylate buffer, pH 7.4 (see recipe) 50%, 70%, 90%, and 100% (v/v) ethanol Epoxy resin (see recipe) 8-mm resin cylinder prepared in advance from a cylindrical mold 35-mm MatTek petri dishes with CELLocate coverslip and map of CELLocate grid (MatTek) Inverted fluorescence microscope Multiphoton microscope, laser-scanning confocal microscope, or digitalized fluorescence inverted microscope capable of acquiring a time-lapse series of images by computer 60°C oven Ultramicrotome with sample holder, glass knife, and diamond knife Eyelashes Pick-up loop (Agar) Adjustable-angle laboratory clamps Slot grids covered with carbon-formvar supporting film (Electron Microscopy Sciences, Agar) Electron microscope Software for three-dimensional reconstruction from serial sections Additional reagents and equipment for transfection (APPENDIX 3A) Transfect cells 1. Plate cells of interest at 50% to 60% confluence on a 35-mm MatTek petri dish with the CELLocate coverslip attached to its bottom. The CELLocate coverslip contains an etched grid with coordinates that allow the localization of the cell of interest at any step in the preparation. To facilitate the process of locating the visualized cell of interest in the resin block, the cells should be plated at 50% to 60% confluence.

2. Transfect cells with cDNA encoding an appropriate GFP fusion protein using any method described in APPENDIX 3A. If electroporation is used, steps 1 and 2 should be reversed. The authors have successfully used procollagen I–mannosidase II–, and vesicular stomatitis virus glycoprotein (VSVG)–GFP constructs for transfection.

Observe labeled structures in living cells 3. Place the dish under an inverted fluorescence microscope, select a transfected cell of interest, and identify its position relative to the coordinates of the CELLocate grid. 4. Draw (or photograph) the position of the cell on the map of the CELLocate grid. 5. Observe the dynamics of the GFP-labeled structures in the selected living cell using a multiphoton microscope, laser-scanning confocal microscope, or digitalized fluorescence inverted microscope that allows acquisition of a time-lapse series of images by a computer.

Microscopy

4.8.3 Current Protocols in Cell Biology

Supplement 11

Fix cells 6. At the moment of interest, while still acquiring images, add fixative A to the cell culture medium at a ratio of 1:1 fixative/medium. Fixation usually induces the fast fading of GFP fluorescence and blocks the motion of labeled structures in the cell.

7. Stop acquiring images and keep cells in fixative for 5 to 10 min. During this time it is useful to acquire a Z series of images of the cell.

8. Wash with 2 ml fixative B and leave the cells in this fixative for 30 min. Immunolabel cells for EM For EM with Nanogold: 9a. Wash cells three times for 5 min with 2 ml PBS. 10a. Incubate cells 30 min in 2 ml blocking solution. 11a. Incubate cells overnight with primary antibody diluted in blocking solution. The dilution depends on the antibody; in general, the concentration of antibodies for EM should be 5- to 10-fold higher than for immunofluorescence.

12a. Wash cells six times for 2 min with 2 ml PBS. 13a. Dilute Nanogold-conjugated Fab fragments of the secondary antibody 1:50 (v/v) in blocking solution and add it to the cells; incubate 2 hr. 14a. Wash cells six times for 2 min with 2 ml PBS. 15a. Fix cells 5 min with 1 ml of 1% glutaraldehyde in 0.2 M HEPES buffer, pH 7.3. 16a. Wash cells three times for 5 min with 2 ml PBS and then again three times with distilled water. 17a. Incubate cells 6 to 10 min with 0.5 ml gold enhancement mixture. 18a. Wash cells three times for 5 min with 2 ml distilled water. For EM with HRP: 9b. Wash cells three times for 2 min with 2 ml PBS. 10b. Incubate cells 30 min in 1 ml blocking solution. 11b. Incubate cells overnight with primary antibody diluted in blocking solution. The dilution depends on the antibody; in general, the concentration of antibodies for EM should be 5- to 10-fold higher than for immunofluorescence.

12b. Wash cells six times for 2 min with 2 ml PBS. 13b. Dilute HRP-conjugated Fab fragments of the secondary antibody 1:50 (v/v) in blocking solution and add it to the cells; incubate 2 hr. 14b. Wash cells six times for 2 min with 2 ml PBS. 15b. Fix cells 5 min with 1 ml of 1% glutaraldehyde in 0.2 M HEPES buffer, pH 7.3. 16b. Wash cells three times for 5 min with 2 ml PBS. Correlative Video Light/Electron Microscopy

17b. Incubate the cells with 1 ml DAB solution. 18b. Wash cells three times for 2 min with 2 ml PBS.

4.8.4 Supplement 11

Current Protocols in Cell Biology

Embed cells 19. Incubate cells in 0.5 ml of a 1:1 (v/v) mixture of 2% OsO4 and 3% potassium ferrocyanide in 0.1 M cacodylate buffer on ice for 1 hr. 20. Wash cells with 2 ml distilled water. 21. Dehydrate cells in a series of ethanol solutions: 50% (10 min), 70% (10 min), 90% (10 min), and three times in 100% (10 min each). 22. Incubate cells 1 to 2 hr in 1:1 (v/v) epoxy resin/ethanol, room temperature. 23. Incubate cells in pure epoxy resin for 1 hr at room temperature and then overnight (≥12 hr) in a 60°C oven. Isolate cell of interest in the resin block 24. Put a droplet of fresh epoxy resin on the block where the examined cell is located. Use the fresh resin as glue to attach the block of embedded cells to the flat base of an 8-mm resin cylinder. Return the assembly to the 60°C oven for an additional 16 hr. 25. Carefully lift the resin from the petri dish and cover glass. It is easiest to remove the resin by gently bending the cylinder to and fro. If the CELLocate cover glass cannot be detached from the embedded cells, incubate the sample in a plastic tube with hydrofluoric acid (HF; Sigma) for 30 to 60 min to dissolve the glass. Do not use a glass tube, as it wll be dissolved by the HF. Wash with water and monitor the completeness of glass dissolution under a stereomicroscope. If glass is not dissolved, repeat treatment with HF and wash again with water. Incubate sample 60 min in 0.1 M PBS or 0.2 M HEPES buffer, pH 7.3, to neutralize the HF, wash with water, and dry. If desired, embedded specimens can be stored for an unlimited time at room temperature before sectioning and analysis.

Prepare EM sections 26. Trim the resin block to give a pyramid of ∼3 × 3 mm size with the cell of interest in its center. 27. Put a sample holder into an ultramicrotome so that the segment arc of the ultramicrotome is in the vertical position. Secure a glass knife in place. 28. Bring the sample as close to the knife as possible. 29. Rotate the knife stage to align the bottom edge of the pyramid parallel to the knife edge. 30. Tilt the segment arc and the knife to adjust the gap between the knife edge and the surface of the sample. The gap is visible as a bright band if all three lamps of the ultramicrotome are switched on. It must be identical in width during the up and down movements of the resin block to ensure that every point of the sample surface containing the cell of interest is at the same distance from the knife edge.

31. Turn the specimen holder 90° to the left or to the right and trim the edges of the resin block to leave a narrow pyramid (≤100 µm wide) with its long axes parallel to the knife edge. The pyramid should be as narrow as possible and the cell of interest should be in its center. An experienced person can trim a pyramid directly with a razor blade.

Microscopy

4.8.5 Current Protocols in Cell Biology

Supplement 11

32. Turn the specimen holder 90° back and lock it in exactly the same position as before. This is very important.

33. Replace the glass knife with a diamond knife and position the latter towards the pyramid. 34. Make serial sections according to the instructions for the ultramicrotome. 35. Stop the motor. Use two eyelashes to divide the band of the sections into pieces suitable in size for collection with a pick-up loop. Each piece should be small enough to fit completely inside the inner circle of the loop without touching it.

36. Pick up a band of sections by touching the pick-up loop to the surface of water containing the band, making sure the band is completely inside the inner circle of the loop and does not touch it. 37. Raise the loop with the sections on it, and fix the loop inside the lab clamps near the stereomicroscope of the microtome. The loop should be visible under the stereomicroscope of the ultramicrotome.

38. Take a formvar-coated slot grid and gently touch sections on the water (do not touch the loop) with the carbon-coated surface of the grid. 39. Very slowly move the slot grid away from the loop. If the movement is slow enough, the water is eliminated from the surface of the supporting film, and only a very small droplet of water remains on the grid, which does not hinder the placement of the grid directly into the grid container. Grids can be stored an unlimited time at room temperature.

Perform EM analysis 40. Place the slot grid under the electron microscope and identify the cell, using the traces of the coordinated grid filled with the resin on the first few sections. 41. Take consecutive photographs (or acquire the images with a computer using a video camera) from the serial sections until the organelle of interest (just observed by light microscopy) is no longer seen. 42. Using software designed for three-dimensional reconstruction, align the images and make a three-dimensional model according to the instructions for the software. REAGENTS AND SOLUTIONS Use deionized or distilled water in all recipes and protocol steps. For common stock solutions, see APPENDIX 2A; for suppliers, see SUPPLIERS APPENDIX.

Blocking solution 100 ml PBS (APPENDIX 2A) containing: 0.5 g BSA 0.1 g saponin 0.27 g NH4Cl Store up to 3 months at 4°C

Correlative Video Light/Electron Microscopy

Cacodylate buffer, 0. 2 M 2.12 g sodium cacodylate 100 ml water Adjust to pH 7.4 with 1 N HCl Store up to 6 months at 4°C

4.8.6 Supplement 11

Current Protocols in Cell Biology

Diaminobenzidine (DAB) solution 10 mg DAB 20 ml 0.1 M Tris⋅Cl, pH ∼7.6 (APPENDIX 2A) Prepare fresh Just before use add 13.3 µl of 30% H2O2 Epoxy resin Combine 20 g epoxy embedding resin (Fluka), 13.0 g dodecenylsuccinic acid (DDSA), and 11.5 g methylnadic anhydride (MNA) in a test tube. Heat in a 60°C oven for 2 to 3 min and then vortex well to mix the components. Add 0.9 g 2,4,6 Tris (dimethylaminomethyl) phenol (DMP-30) and immediately vortex again. Store 1-ml aliquots up to 6 months at –20°C. Fixative A (0.1% glutaraldehyde, 4% paraformaldehyde) Add 1.25 ml of 8% (w/v) glutaraldehyde to 100 ml of fixative B (see recipe). Store up to 2 to 3 days at 4°C. Fixative B (4% paraformaldehyde) Dissolve 4 g paraformaldehyde powder in 100 ml 0.2 M HEPES buffer, pH 7.4, while stirring and heat to 60°C. Add a few droplets of 1 N NaOH to form a clean solution. Store up to 3 months at 4°C. Gold enhancement mixture Mix equal volumes of solutions A (enhancer; green cap) and B (activator; yellow cap) of a Gold-enhance kit (Nanoprobes) and wait 5 min. Add equal volumes of solution C (initiator; purple cap) and then solution D (buffer; white cap) and mix. Prepare ∼400 µl final reagent per dish. A convenient method is to use equal numbers of droplets from each bottle.

COMMENTARY Background Information Correlative light/electron microscopy (CLEM) was developed several years ago, and has been used when the analysis of immunofluorescently labeled structures required higher resolution than can be achieved using light microscopy (Tokuyasu and Maher, 1987; Powell et al., 1998). In spite of its potential, CLEM has not been used very often, probably because the possibility of correlating two static images–one fluorescent and one electron microscopical–is of interest only in a limited number of situations. The real gains from CLEM come 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 elucidating the three-dimensional ultrastructure of individual dynamic cellular objects at times of choice can be informative (Mironov 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 environ-

ment in which they are growing, as well as the interactions of the microtubule tips with other cytoskeletal elements or intracellular organelles, at various stages of tip growth. In addition to cytoskeletal dynamics, CVLEM could also be useful for studying cell division and cell-cell interactions (although the specific questions to be addressed are best left to the specialists). CVLEM has been applied successfully to characterize the ultrastructure of membrane carriers transporting secretory proteins from the Golgi complex to the cell surface (Polishchuk et al., 2000). 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 fast both in the field of live-cell imaging and in the field of EM, and new powerful technologies are rapidly becoming available in user-friendly versions. This should make the use of CVLEM appealing to a number of cell biologists. Through the use of fluorescent proteins of different colors, two or more marker proteins can

Microscopy

4.8.7 Current Protocols in Cell Biology

Supplement 11

be observed simultaneously (Ellenberg et al., 1999). This will allow 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 electron microscope tomography (Ladinsky et al., 1994) will increase the subtlety and range of the questions that can be answered by CVLEM. While we wait for the microscopy of the future, 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.

Critical Parameters

Correlative Video Light/Electron Microscopy

Taken together, all the steps of CVLEM represent quite a long procedure (see Time Considerations) and require significant effort by a researcher or technician. Hence, it would be especially disappointing to loose such tourde-force experiments due to small problems in specimen handling. To apply CVLEM successfully, several important parameters should always be taken into account by the experimenter. Fixation/labeling protocol. Preliminary experiments should be carried out to understand whether the antibodies selected for labeling of a GFP–fusion protein work with the immunoEM protocol. Many antibodies that give perfect results for immunofluorescence do not work for immuno-EM staining, because the glutaraldehyde used in most EM fixatives tends to cross-link amino groups of antigen epitopes, therefore decreasing the antigenicity of the target protein and restricting the penetration of antibodies across the cytosol. However, a decrease or an absence of glutaraldehyde in the fixative can result in poor preservation of ultrastructure of the intracellular organelles. If immuno-EM labeling is not successful, it may be possible to optimize the concentration of glutaraldehyde in the fixative, or to use the periodate/lysine/paraformaldehyde fixative described by Brown and Farquhar (1989). In addition, it is important to select an immunoperoxidase or immunogold protocol to label the structure of interest. To label epitopes of a GFP–fusion protein located in the cytosol, only the immunogold protocol should be used; for other epitopes, HRP labeling is also suitable. Locating the cell of interest. It is extremely important to be able to find the cell of interest at any step in the CVLEM procedure. Only cells located on the grid of the MatTek petri dish can be selected for time-lapse observations. The position of the cell of interest on the grid should

be noted, otherwise the cell will be difficult to locate again. Low-magnification images showing the field surrounding the cell of interest can greatly help when trimming the resin block around the cell and when locating the cell under the electron microscope. In this case, neighboring cells can be used as landmarks to identify the cell of interest. It is more convenient to work with cells that are plated at a lower confluence (e.g., 50% to 60%). It is also useful to have fluorescent and phase-contrast images of the target cell, because particular structures (e.g., microvilli, pseudopodia, inclusions) can also be used to help find both the cell and the structure of interest. Section thickness. The thickness of serial sections should be ∼50 nm (or less) for very precise three-dimensional reconstruction, ∼80 nm for routine work, or ∼250 nm for electron microscope tomography.

Anticipated Results Once considerations for specific localization of the cell and the organelle have been optimized, the precise three-dimensional structure of the organelle and its connections with other organelles at different moments of its life cycle can be identified.

Time Considerations The entire CVLEM procedure requires 4 to 5 days. During the first day, cells must be transfected with cDNA. The next day, it is possible to make observations in living cells, fix them, and start the immunolabeling. The third day is required to complete the immunolabeling and resin embedding of the cells. During the fourth experimental day, the cell of interest is identified in the resin block and cut into serial sections. On the fifth day, EM analysis is performed. If desired, it is possible to store embedded specimens indefinitely before sectioning and analysis.

Literature Cited Brown, W.J. and Farquhar, M.G. 1989. Immunoperoxidase methods for the localization of antigens in cultured cells and tissue sections by electron microscopy. Methods Cell Biol. 31:553-569. Burry, R.W., Vandre, D.D., and Hayes, D.M. 1992. Silver enhancement of gold antibody probes in pre-embedding electron microscopic immunocytochemistry. J. Histochem. Cytochem. 40:1849-1856. Deerinck, T.J., Martone, M.E., Lev-Ram, V., Green, D.P., Tsien, R.Y., Spector, D.L., Huang, S., and Ellisman, M.H. 1994. Fluorescence photooxidation with eosin: A method for high resolution

4.8.8 Supplement 11

Current Protocols in Cell Biology

immunolocalization and in situ hybridization detection for light and electron microscopy. J. Cell Biol. 126:901-910. Ellenberg, J., Lippincott-Schwartz, J., and Presley, J.F. 1999. Dual-colour imaging with GFP variants. Trends Cell Biol. 9:52-56. Ladinsky, M.S., Kremer, J.R., Furcinitti, P.S., McIntosh, J.R., and Howell, K.E. 1994. HVEM tomography of the trans-Golgi network: Structural insights and identification of a lace-like vesicle coat. J. Cell Biol. 127:29-38. Lippincott-Schwartz, J. and Smith, C.L. 1997. Insights into secretory and endocytic membrane traffic using green fluorescent protein chimeras. Curr. Opin. Neurobiol. 7:631-639. Mironov, A.A., Polishchuk, R.S., and Luini, A. 2000. Visualising membrane traffic in vivo by combined video fluorescence and 3-D-electron microscopy. Trends Cell Biol. 10:349-353. Perez, F., Diamantopoulos, G.S., Stalder, R., and Kreis, T.E. 1999. CLIP-170 highlights growing microtubule ends in vivo. Cell 96:517-527. Polishchuk, R.S., Polishchuk, E.V., Marra, P., Buccione, R., Alberti, S., Luini, A., and Mironov,

A.A. 2000. GFP-based correlative light-electron microscopy reveals the saccular-tubular ultrastructure of carriers in transit from the Golgi apparatus to the plasma membrane. J. Cell Biol. 148:45-58. Pollok, B.A. and Heim, R. 1999. Using GFP in FRET-based applications. Trends Cell Biol. 9:57-60. Powell, R.D., Halsey, C.M., and Hainfeld, J.F. 1998. Combined fluorescent and gold immunoprobes: Reagents and methods for correlative light and electron microscopy. Microsc. Res. Tech. 42:212. Tokuyasu, K.T. and Maher, P.A. 1987. Immunocytochemical studies of cardiac myofibrillogenesis in early chick embryos. II. Generation of α-actinin dots within titin spots at the time of the first myofibril formation. J. Cell Biol. 105:27952801.

Contributed by Roman S. Polishchuk and Alexander A. Mironov Consorzio Mario Negri Sud S. Maria Imbaro (Chieti), Italy

Microscopy

4.8.9 Current Protocols in Cell Biology

Supplement 11