Field Emission Scanning Electron Microscopy and Visualization of the Cell Interior
Resolution in scanning electron microscopy (SEM) has improved dramatically in recent years so that for the majority of biological material, no significant differences exist in resolution between SEM and conventional transmission electron microscopy (TEM). High brightness sources (field emission) and novel final lens configurations have resulted in instrument resolutions of 0.5 to l nm, allowing direct, in situ, threedimensional visualization of surface detail at molecular resolution. As all this technology relies on field emission sources of the electron beam, either by "cold" field emission or thermally assisted "Schottky" field emission, we refer to it as FESEM.
Surface imaging allows bulk samples to be examined without limitation of specimen thickness. Visualization of intracellular surfaces requires some means of access, such as isolation of cell fractions or macromolecules, or in situ, via fracture, or sectioning techniques. Cell-free systems, e.g., in vitro nuclear formation, allow biological interfaces such as developing nuclear envelopes to be imaged directly (Goldberg et al., 1992). True three-dimensional (3D) surface visualization can be achieved by tilting the specimen to make stereo pairs, and accurate surface measurements can be made from computerized 3D reconstructions. The surfaces can be characterized further by immunogold labeling, which can be unequivocally localized by the strong backscatter signal of the gold probes. For specimens that are thin enough to allow electron penetration, a scanning TEM (STEM) image can also be obtained readily and displayed simultaneously alongside the secondary electron image, producing complementary information from the transmitted beam/specimen interactions. The use of low accelerating voltages in FESEM has also been shown to be of advantage, reducing charging and penetration of the electron beam, but maintaining a high-resolution information content. High-pressure freezing, freeze substitution, and examination of cryohydrated specimens may all be used for FESEM (Muller and Hermann, 1990; Walther, 2003) but can be considered specialized and are not covered in this article, although we do describe the use of cryoultramicrotomy and cryoabrasion as techniques to access internal surfaces within the cell prior to conventional imaging by FESEM. Basically, we deal with techniques that rely on chemical preservation, followed by dehydration, critical point drying, and coating. Conventional SEM coating (up to 20nm thickness) with sputtered gold completely obscures fine surface detail in HRSEM and must be replaced by high-resolution coating. We routinely coat with a 1- to 2-nm film of chromium or tungsten, which has a grain size of 0.3 to 0.5 nm (Apkarian et al., 1990).
II. MATERIALS AND INSTRUMENTATION
It is crucial to visit manufacturers' demonstrations with the material that will be investigated to ascertain that suitable performance can be assured from the chosen equipment.
III. PROCEDURES A. Exposing Surfaces within the Cell
1. Subcellular Fractionation
Organelles and macromolecules can be isolated by standard procedures, possibly requiring subsequent modifications in the light of HRSEM visualization, which are beyond the scope of this article. Basically, the specimens must be undamaged by osmotic shock, proteolysis, or unsuitable isolation buffers. They must also be clean. The surface of organelles should, for instance, be free of attached cytoskeletal remnants or cytoplasmic contamination. Where the specimens are available as purified macromolecules or viruses, they may be deposited on carbon-coated TEM grids in the conventional manner and viewed by HRSEM. In this situation, TEM negative staining will usually be replaced by fixation for SEM and air drying replaced by critical-point drying followed by chromium coating. If a STEM detector is available, the virus/macromolecule can be recognized as a transmitted "reference image" after this protocol and compared directly with the secondary electron (SEM) image. The thin coating of chromium applied for the secondary electron imaging does not interfere with the STEM imaging.
Adhering Sample to Support
Many cell components naturally adhere to glass coverslips, silicon chips, or carbon support film on grids. Glass coverslips may be a useful initial prepar ative substratum, as they can be checked in the phasecontrast microscope for the density and distribution of specimens and for the progression of various protocols such as detergent extraction of cytoplasm. Once the isolation protocol is established, coverslips should be replaced with silicon chips as specimen substrates, as silicon is a conductive substratum in contrast to glass, an insulator, which can generate problems with charging in the SEM. Tissue culture cells will grow in identical fashion on silicon as they do on glass or plastic, and isolated cytosol or organelles will also adhere naturally to silicon in the same way as they do to glass. If samples are fixed in suspension it may be necessary to coat the support with poly-L-lysine to facilitate adherence. Different samples may require slight modification, but the basic technique is as follows (Fig. 1).
Poly-L-lysine: Make a fresh 1-mg/ml solution of poly-L-lysine in sterile distilled water; use within 24h.
2. "In Situ" Exposure of Intracellular Surfaces Dry Fracture
This is a simple but extremely effective way of exposing internal surfaces in both tissues and cells. After fixation, dehydration, and critical-point drying, merely gently press the surface of the specimen to a square of double-sided tape mounted to a second silicon chip and pull away without shearing, coat both chips as normal, and examine in the SEM. The fracture will remove material on the surface of the adhesive and leave fractured material "in situ.'" This technique may be enhanced by pretreatment with detergent (0.5% Triton X-100, 2-3min for tissue culture cells), either alone or mixed with the primary fixative (2.0% paraformaldehyde and 0.1% glutaraldehyde), and subsequently refixed as described (Allen et al., 1998).
These methods involve sectioning of embedded specimens followed by exposure of internal surfaces by removal of the supporting material. This may vary among epoxy resins, various waxes, and even ice. Resins that require corrosive solvents for removal will tend to be prone to surface etching. A mixture of 50% propylene oxide and 50% sodium methoxide (dissolve 2g NaOH pellets in 100ml absolute methanol) will remove most resins.
3. Cryo Methods to Expose Internal Surfaces for FESEM
Surface imaging by FESEM may be achieved by isolating the cellular component of interest, such as mitochondria, but this approach does not allow access to the interior structure of such an organelle (see later). One way to expose such surfaces is to freeze the cells or tissue and cut cryo sections, which themselves can be viewed in the SEM, to "cryoplane" and expose the whole blockface in the SEM, or to "cryoabrade" the surface of the frozen sample and expose surface features in a different way. Samples are then thawed and processed for FESEM as normal. This gives a crosssectional view but with much greater depth of information than in a resin-embedded thin section viewed in the TEM because the sections can be very thick, they are resin free, and there is a greater depth of focus. Information can also be gathered quite simply in 3D simply by taking stereo pairs, which also allows computerized 3D reconstruction and measurement of the surface. This method is also compatible with immunogold labelling
Cryomicrotomy is an adaptation of the widely used "Tokyasu" technique (Tokyasu, 1986, 89) for immuno-TEM.
4% paraformaldehyde in PBS
2M sucrose in PBS
Cryo ulramicrotome (e.g., Leica Ultracut R with FCS cryo attachment)
These problems are associated with the Tokuyasu technique.
When processing the sample remaining on the pin for SEM, the specimen always detaches from the pin. This leaves a very small specimen that is easily lost during the dehydration and CPD steps. An additional problem is the attachment of the sample to a silicone chip after CPD. The sample is very fragile and it is not always easy to identify the "planed" side. Adhere the sample to the chip using a thin smear of silver dag. It is possible for the sample to flip over during this process. Also, because of the irregular shape of the sample, ensuring good contact and therefore a good earth path from specimen to chip can be tricky. It is easy to submerge the specimen in too much dag.
Specific Protocol for Mitochondrial Isolation and Exposure of Internal Structure by FESEM
Isolate mitochondria using the differential centrifugation method (Gottlieb et al., 2003). Harvest and place cells
(2.5-5 × 108) on ice for 15min, centrifuge at 500g for 5 min at 4°C, wash with ice-cold PBS, and subsequently wash with ice-cold mitochondrial isolation buffer (MIB) (200mM mannitol, 70mM sucrose, 1mM EGTA, 10mM HEPES, 0.5 mg/ml BSA; pH 7.4). Resuspend cells in ice-cold MIB and then homogenize in a syringe-driven cell disruptor. Spin the lysate at 800g for 10min at 4°C. Remove supematants and spin at 10,000g for 10min at 4°C. Add fixative (3% glutaraldehyde) to the pellets and keep samples at 4°C for 1h. Remove the fixative carefully and infiltrate the pellet with sucrose/PVP solution overnight (Tokuysau, 1989). During this process, leave the pellet undisturbed. Then carefully excise small (>1mm2) pieces of sample from the pellets and mount onto aluminium plunge freezing pins (Leica Microsystems, Milton Keynes). Mount the pins into a plunge freeze unit (Leica CPC) and freeze in liquid propane at a temperature of -182°C. Transfer the frozen specimen and pin under LN2 into a cryo ultramicrotome (Leica Ultracut S with FCS attachment). Using a diamond trimming knife (Diatome Cryotrim 45), trim several semithin sections (350 nm) from the sample in order to remove surface sucrose. Cut further semithin sections (350-400nm) from the sample using a diamond cryo knife. Collect each section on a sucrose loop according to the "Tokuyasu" technique (Tokuyasu, 1986). Thaw the frozen sections onto 5-mm silicone chips and process for SEM as follows.
Aligning the cryo abrasive pad with the specimen is very tricky and must be done with care. It is very easy for the protruding abrasive shards of the wet and dry paper to embed themselves into the frozen block. Also, if the section advance is too great, the sample block can be literally ripped from the specimen pin. A few micrometres must be shaved off the sample face in order to ensure that all surface sucrose has been removed. This only leaves a few micrometres of wellfrozen vitrified sample to work with. Once again, the sample size is very small and is easily damaged or lost in subsequent processing and mounting steps.
Identifying the abraded face can be tricky even under a stereomicroscope. The swirled pattern of the specimen pin that has been embossed into the underside of the sample can look very similar to the abraded face, leading to the specimen being mounted pin side up.
All fixatives are ideally made up just before use or at least the same day; both glutaraldehyde and glutaraldehyde-tannic acid solutions should be filtered before use through a 0.22-µm filter. The 1% aqueous uranyl acetate should be stored in a brown bottle. Osmium tetroxide is made by breaking the glass ampoules in which the crystals are delivered, having previously washed them free of label and adhesive under the tap, in a fume cupboard. The ampoules plus crystals are dropped in the correct amount of buffer or distilled water where the osmium dissolves to give the appropriate final concentration. (Note: Osmium is extremely hazardous and appropriate precautions must be observed.) Thiocarbohydrazide or tannic acid solutions should also be made just prior to use (Allen et al., 1988).
Always use glutaraldehyde of EM-grade quality from a high stock concentration (50%) stored in a freezer. Low concentration stock solutions and storage in large bottles at room temperature will reduce the cross-linking properties of the glutaraldehyde.
1. Isolated Proteins and Nucleoproteins
2. Small and Easily Preserved Structures
3. Large and/or Fragile Structures (e.g., Whole Cells, Organelles, Cytoskeletal Preparations, Isolated Cells, or Nuclear Membranes)
5. Preparation of in Vitro-Assembled Organelles for FESEM
Organelles, such as nuclei, endoplasmic reticulum (ER), and Golgi, can be assembled in cell-free extracts. Extracts made from frog eggs are a particularly powerful system for studying the assembly, dynamics, and functions of these organelles. Organelles can be isolated cleanly from the extract and their surfaces can be examined by FESEM. In vitro-assembled nuclei, as well as ER, can be prepared for FESEM as follows.
C. Critical-Point Drying
All traces of water should be removed from ethanol, Arklone, and CO2. Let 100% ethanol and Arklone stand over molecular sieve for more than 24h prior to use. High-purity liquid CO2 (less than 5 ppm water) should be used and passed through a water filter as a precaution.
Critical-point-dried samples should be transferred immediately into the sputter coater to avoid rehydration, and coated samples are best viewed in the microscope directly. However, if it is known that the microscope cannot be accessed, it is better to pause preparations after critical-point drying and store preparations under vacuum.
D. Sputter Coating
F. Immunogold Labeling
The basics of specimen preparation for immunogold labeling are beyond the length limits for this article and are adequately covered elsewhere (see article by Roos et at. for additional information). For immunogold labeling for HRSEM, the following points are important.
1. Size of Probe The choice of probe size is a compromise between sensitivity and subsequent detection. Very small gold probes (around l nm) have minimal steric hindrance and consequently label with maximum sensitivity. One-nanometer gold has been visualized by backscatter imaging in HRSEM (Hermann et al., 1991), but this is at the limits of resolution and is best increased in diameter in situ by silver or gold enhancement to a size at which it can be visualised more easily (around 5-10nm). We have used both 5- and 10-nm gold as a good compromise between sensitivity and localization. Because most modern instruments will discriminate easily between 5- and 10-nm labeling, these can be used together successfully for double-labeling studies.
Using gold probes obviously prohibits gold coating for SEM. In the past, gold-labelled specimens have been coated with carbon, mainly to inhibit charging, but carbon produces a severely limited secondary electron signal and, consequently, little topographical information. We have found that a 1.5-nm chromium coating provides the ideal solution, retaining the full secondary electron-generated surface information, without compromising the detection of gold by backscattered electron detection (Allen and Goldberg, 1993).
In this situation, having found that "mixed" imaging of SE and BSE signals was not satisfactory, we have chosen to collect each signal separately (but simultaneously) and then to superimpose the gold BSE signal onto the secondary signal (retaining register) in Adobe Photoshop, often altering the colour to improve the appearance of label against the monochrome background. In modern instruments with good low kilovolt performance, uncoated or carbon-coated imaging will generate such a strong signal from gold probes that they are observed easily in secondary electron imaging.
Although field emission SEM has been available for some time, it is still a relatively new technique in cell biology. The procedures given here may need to be modified to optimize the preservation of some structures. Probably the most difficult step is exposing recognizable and undamaged intracellular surfaces. Isolation of organelles offers the possibilities of further characterization by other methods, but gives no "in situ" information and may involve extensive biochemical protocols. Resinless sections and dry fracture give in situ information, but only after some initial extraction of the cell. Freeze fracture, followed by frozen hydrated coating and visualization, may alleviate these problems but is limited by the plane of fracture, as the structure of interest may not be exposed. It is also technically difficult and expensive. Osmium etching results in spectacular images of intracellular membranes, but the uncertainty of what is removed makes interpretation difficult. Direct visualization of biological interfaces in cell-free systems (e.g., in vitro nuclear formation) is a particularly promising area (Goldberg et al., 1992, 1997). Considerable fresh structural information has also been demonstrated for nuclear pore complexes and associated structures (Ris, 1991; Goldberg and Allen, 1992, 1996; Kiseleva et al., 1996).
T. D. Allen, S. Rutherford, and S. Murray are supported by CRUK and M. W. Goldberg is supported by a Wellcome Lectureship. The mitochondrial pellets were supplied by Dr. E Gottlieb (Beatson Institute).
Allan, V. J., and Vale, K. (1994). Movement of membrane tubules along microtubules in vitro, J. Cell. Sci. 107, 1885-1895.
Allen, T. D., and Goldberg, M. W. (1993). High resolution SEM in cell biology. Trends Cell Biol 3, 203-208.
Allen, T. D., Jack, E. M., and Harrison, C. (1988). Three dimensional structure of human metaphase chromosomes determined by scanning electron microscopy. In "Chromosomes and Chromatin" (K. W. Adolph, ed.), Vol. 11, pp. 52-70. CRC Press, Boca Raton, FL.
Allen, T. D., Rutherford, S. A., Bennion, G. R., Wiese, C., Riepert, S., Kiseleva, E., and Goldberg, M. W. (1998). Three dimensional surface structure analysis of the nucleus. Methods Cell Biol. 53, 125-138.
Apkarian, R. P., Gutekunst, M. I., and Joy, D. C. (1990). High resolution SEM study of enamel crystal morphology. Electron Microsc. Tech. 14, 70-78.
Goldberg, M. W., and Allen, T. D. (1992). High resolution scanning electron microscopy of the nuclear envelope: Demonstration of a new regular, fibrous lattice attached to the baskets of the nucleoplasmic face of the nuclear pores. J. Cell Biol. 119, 1429- 1440.
Goldberg, M. W., and Allen, T. D. (1996). The nuclear pore complex and lamina: Three dimensional structures and interactions determined by field emission in lens scanning EM. J. Mol Biol. 257, 848-865.
Goldberg, M. W., Blow, J. J., and Allen, T. D. (1992). The use of the field emission in-lens scanning electron microscope to study the steps of assembly of the nuclear envelope in vitro. J. Struct. Biol. 108, 257-26.S.
Goldberg, M. W., Wiese, C., Allen, T. D., and Wilson, K. L. (1997). Dimples, pores, star rings and thin rings on growing nuclear envelopes: Evidence for structural intermediates in nuclear pore complex assembly. J. Cell Sci. 110, 409-420.
Gottlieb, E., Armour, S. M., Harris, M. H., and Thompson, C. B. (2003). Mitochondrial membrane potential regulates matrix configuration and cytochrome c release during apoptosis. Cell Death Differ. 10, 709-717.
Hermann, R., Schwartz, H., and Muller, M. (1991). High precision immunostaining electron microscopy using Fab fragments coupled to ultra-small colloidal gold. J. Struct. Biol. 107, 38-47.
Kiseleva, E., Goldberg, M. W., Daneholt, B., and Allen, T. D. (1996). RNP export is mediated by structural reorganisation of the nuclear pore basket. J. Mol. Biol. 260, 304-311.
Muller, M., and Hermann, H. (1990). Towards high resolution SEM of biological objects. In "Proceedings 12th International Congress on Electron Microscopy" (L. D. Peachy, D. R. Williams, eds.). Vol. 3, pp. 4-5. San Francisco Press, San Francisco.
Newmeyer, D. D., and Wilson, K. L. (1991). Egg extracts for nuclear import and nuclear assembly reactions. In "Methods in Cell Biology" (B. K. Kay, H. B. Peng, eds.), Vol. 36, pp. 608-635. Academic Press, San Diego.
Reipert, S., Reipert, B. M., and Allen, T. D. (1994). Preparation of isolated nuclei from K562 haemopoietic cell line for high resolution scanning electron microscopy. Microsc. Res. Tech. 29, 54-61.
Ris, H. (1991). The three dimensional structure of the nuclear pore complex as seen by high voltage electron microscopy and high resolution low voltage scanning electron microscopy. EMSA Bull. 21, 54-56.
Tokyasu, K. T. (1986). Application of cryomicrotomy to immunocytochemistry. J. Microsc. 143, 139-149.
Tokuyasu, K. T. (1989). Use of poly(vinylpyrrolidone) and poly(vinyl alcohol) for cryoultramicrotomy. Histochem. J. 21, 163-171.
Walther, P. (2003). Recent progress in freeze-fracturing of high pressure frozen samples. J. Microsc. 212, 34-43.
© 2018 Biocyclopedia | All rights reserved.