Rapid Freezing of Biological Specimens for Freeze Fracture and Deep Etching

I. PRINCIPLES OF RAPID FREEZING
Stabilization of biological structure by the physical process of freezing (cryofixation) forms the starting point for freeze fracture and deep etching (see article by Shotton) and for freeze substitution, cryoultramicrotomy, and cryoelectron microscopy (see articles by Roos et al., and Resch et al.). To avoid ultrastructural damage to the specimen caused by the growth of large ice crystals, rapid freezing is essential. True vitrification (i.e., formation of amorphous, noncrystalline ice) can be achieved only by cooling rates of greater than 2 × 105 °C/s over the critical range of 20 to -100 °C, i.e., cooling over this range in a fraction of a millisecond. Rates of this magnitude can be attained in very thin (<3µm) films of suspended liquid that are plunged rapidly into liquid nitrogen-cooled liquid propane or ethane, and particulate specimens (e.g., viruses) embedded in such frozen thin films may be observed directly in the vitrified state on the cold stage of a cryoelectron microscope. For freeze fracture and deep etching, however, the requirement for a larger specimen size precludes true vitrification throughout the specimen because the maximal cooling rate possible within the sample is limited by the rate of heat conduction through it. Thus, for most practical purposes, the goal when applying cryofixation for freeze fracture and deep etching is to apply cooling conditions that reduce ice crystal size sufficiently that there is no visible distortion of cellular structure. The techniques employed for such rapid freezing may be divided into three groups: conventional rapid freezing techniques, with cooling rates of between 103 and 104°C/s; ultrarapid freezing techniques, with rates in excess of 104°C/s; and hyperbaric freezing in which, although a relatively slow cooling rate is employed, ice crystal nucleation and growth are retarded by high pressure. This article gives a brief account of the principles and practical applications of these techniques. Comprehensive reviews of the field as a whole can be f ound in Robards and Sleytr (1985), Steinbrecht and Zierold (1987), Echlin (1992), and Severs and Shotton (1995).
Working with cryogenic liquids is potentially hazardous. Novices should ensure that they are fully conversant with appropriate safety precautions by consulting their institutional safety advisor. For further guidance articles on safety in Robards and Sleytr (1985), Ryan and Liddicoat (1987), and Steinbrecht and Zierold (1987).

II. CONVENTIONAL RAPID FREEZING FOR FREEZE FRACTURE
A. Chemical Fixation and Glycerination
Ice crystal damage can be avoided at relatively slow cooling rates if the specimen is first infiltrated with a buffered cryoprotectant. Glycerol, used at concentrations of 20-30%, is by far the most commonly used cryoprotectant. Prior fixation with aldehydes (usually glutaraldehyde) is routinely carried out with the aim of minimizing cryoprotectant-induced artifacts, although such aldehyde fixation may itself induce artifacts. A few specimens (e.g., cells of low water content such as yeast or bacteria, or concentrated membrane preparations such as erythrocyte ghosts) may be successfully frozen by conventional methods without prior chemical pretreatment.

B. Mounting of Specimens Prior to Freezing
To enable processing through the various subsequent steps of freeze fracture or deep etching, specimens are first mounted on specially designed supports (Fig. 1). Standard specimen supports are made from metals of high thermal conductivity and are as small as is compatible with ease of handling. The precise design of support used will vary according to the nature of the specimen, the type of freeze-fracture apparatus, and the manner in which fracturing is executed. Careful mounting is critical and should always be done with the aid of a binocular microscope.
FIGURE 1 Examples of supports used for mounting specimens prior to freezing. (a and b) Standard supports
for knife fracturing made either from a gold-nickel alloy 
(Balzers) or from copper (e.g., Reichert Cryofract, 
Polaron). (c) An inexpensive alternative, compatible with 
the same type of specimen table, punched out from a 
copper sheet. (d-f) Mounting systems for fracturing 
specimens in double-replica devices. Different designs of 
these mounts are used in different freeze-fracture plants. 
(d and e) Different types of mount used to sandwich thin 
specimens as shown in f. When a pair of flattopped 
mounts (d) is used, a spacer may be included to prevent 
compression of the sample. (g) Use of a standard 
flat-topped support and PVA for mounting of cells 
cultured on a plastic coverslip. Reproduced
from Severs and Shotton (1995), with permission.
FIGURE 1 Examples of supports used for mounting
specimens prior to freezing. (a and b) Standard supports
for knife fracturing made either from a gold-nickel alloy
(Balzers) or from copper (e.g., Reichert Cryofract,
Polaron). (c) An inexpensive alternative, compatible with
the same type of specimen table, punched out from a
copper sheet. (d-f) Mounting systems for fracturing
specimens in double-replica devices. Different designs of
these mounts are used in different freeze-fracture plants.
(d and e) Different types of mount used to sandwich thin
specimens as shown in f. When a pair of flattopped
mounts (d) is used, a spacer may be included to prevent
compression of the sample. (g) Use of a standard
flat-topped support and PVA for mounting of cells
cultured on a plastic coverslip. Reproduced
from Severs and Shotton (1995), with permission.

For conventional knife fracture of cell suspensions (see article by Shotton), a droplet of concentrated cell suspension is placed on the central raised portion of a cleaned flat-topped specimen support (Figs. 1a and 1c). Care should be taken to avoid bubbles and to avoid getting liquid on the rim of the holder (which will prevent it fitting the specimen table after freezing). Tissue blocks for knife fracture are conventionally mounted in similar supports that have a central well (Fig. 1b). The tissue is held securely in the well with a portion protruding for subsequent fracture. Flat-type holders without wells can also be used as mounts for tissue pieces; in this case, polyvinyl alcohol (PVA) mounting medium is recommended to attach the sample firmly to the holder. PVA mounting medium consists of 20-30% PVA in phosphate-buffered saline (PBS) containing 20-33% glycerol. A simple recipe is to dissolve the PVA powder (mean molecular mass 10kDa) to 45% in PBS by prolonged heating at below 100°C in a double boiler and then to dilute with half the volume of glycerol to give a solution containing 30% PVA and 33% glycerol. PVA is conveniently applied from a syringe or using a sharpened applicator stick and should be kept at 4°C or, for prolonged storage, at -20°C.

As an alternative to fracturing by knife, specimens may be fractured by being broken apart in a hinged double-replica device. This dictates use of a pair of mounts between which the specimen is sandwiched (Figs. 1d-1f). The double--mount principle was originally devised for making complementary P- and E-face replicas, but is often convenient for routine preparation.

Various techniques have been developed for mounting cultured cells for freeze fracture. A particularly versatile approach, suitable for cells grown on plastic coverslips, is that of Pauli et al. (1977). A piece of the coverslip is mounted cell side down on a droplet of PVA placed on a standard flat-topped support, leaving a portion of the coverslip projecting horizontally by about 0.5mm (Fig. l g). Fracturing is done by raising the tip of the knife from below the coverslip; this flips the coverslip off the frozen PVA, directing the fracture plane through the cells.

C. Freezing
Conventionally fixed and cryoprotected specimens are normally frozen by simple immersion in a liquid cryogen. However, direct immersion of specimens into liquid nitrogen (which at atmospheric pressure is at its boiling point, -196°C) leads to formation of an insulating layer of evaporated nitrogen gas around the specimen, preventing rapid cooling (the leidenfrost effect). This problem is avoided by using nitrogen slush, a mixture of solid and liquid nitrogen at its freezing point (-210°C).

A simple way to make nitrogen slush is to fill a small, well-insulated styrofoam container with liquid nitrogen, place it in a desiccator, and evacuate it using a water pump or rotary pump until, through loss of latent heat of evaporation, the nitrogen ceases to bubble and solidifies. After a further 30s, the vacuum is released, and some of the solid nitrogen melts, giving a slush at -210°C. By repeating this process several times, the entire volume of nitrogen is brought down to the same temperature before removal for use. Slush can be made similarly using a standard vacuum evaporator unit or a commercially available slusher. Once made, the slush must be used immediately. The mounted specimen, held in fine forceps, is inserted swiftly into the nitrogen slush and is then transferred into a separate container of liquid nitrogen for storage or further processing. After a few minutes the solid nitrogen will have melted and a new batch of slush must be prepared.

The traditional alternative method for freezing specimens is to immerse them manually into a secondary cryogenic liquid cooled to near its freezing point using liquid nitrogen as the primary cryogen. Such a secondary cryogenic liquid ideally combines the properties of high thermal conductivity, a high heat capacity, a freezing point close to the boiling point of liquid nitrogen, and a large temperature difference between its freezing and boiling points. The nonflammable refrigerant gases freon 22 (chlorodifluoromethane, CHC1F2; melting point -160°C, boiling point -40°C) and freon 12 (CCl2F2; melting point -155°C, boiling point -30°C) fit these requirements and were, for many years, used as the standard secondary cryogens for the freezing of cryoprotected specimens. However, freons are no longer used due to their deleterious effects on the earth's ozone layer. The cooling rates obtainable with nitrogen slush are comparable to those achieved with the fluorocarbon cryogens and are certainly adequate for the freezing of cryoprotected specimens.

III. ULTRARAPID FREEZING
Ultrarapid freezing techniques opened a new chapter in biological ultrastructure research, permitting the examination of specimens that had been frozen directly from the living state, without prior chemical treatment. There are four principal methods for the ultrastructural preservation of biological specimens by extremely rapid freezing in the absence of cryoprotectants: plunging, spraying, jetting, and metal block-impact freezing (reviewed by Gilkey and Staehelin, 1986). With all these methods, good ultrastructural preservation is confined to a 10- to 20-µm surface layer. Further from the surface, the inherently low thermal conductivity of biological tissue limits the rate of heat loss and leads unavoidably to the growth of large ice crystals and consequent ultrastructural damage, similar to that observed if noncryprotected samples are frozen directly by standard immersion freezing.

A. Plunge Freezing
By optimizing conditions for the traditional approach of immersing specimens in a liquid nitrogencooled cryogen, cooling rates can be improved sufficiently to permit observation of a well-frozen structure in the absence of chemical cryoprotection. Numerous pneumatic, solenoid-operated and spring-driven devices incorporating these features have been developed; one example is illustrated in Fig. 2a. Key conditions for efficient cooling are that the specimen should have a maximal surface to volume ratio and be mounted in thin supports of low mass, and that its entry velocity into the cryogen should be high (hence the term plunge freezing). The stirred cryogen should fill a deep container to the brim so that the specimen does not undergo precooling before entry and completes its cooling over the critical range from 20°C to -100°C while still in motion through the liquid.

FIGURE 2 Examples of rapid freezing apparatus. (a) An example of a solenoid-operated plunge-freezing apparatus [adapted from Escaig (1982), with permission]. (b) A metal block slam-freezing device (Heuser version). The specimen is mounted on a detachable freezing head, which plugs into the plunger, the highly polished pure copper block is cooled from below with a spray of liquid helium, and the plunger is allowed to fall onto the block under gravity. The original device was set up to permit stimulation in nerve muscle preparations. From Heuser et al. (1979), with permission.
FIGURE 2 Examples of rapid freezing apparatus. (a) An example of a solenoid-operated plunge-freezing apparatus [adapted from Escaig (1982), with permission]. (b) A metal block slam-freezing device (Heuser version). The specimen is mounted on a detachable freezing head, which plugs into the plunger, the highly polished pure copper block is cooled from below with a spray of liquid helium, and the plunger is allowed to fall onto the block under gravity. The original device was set up to permit stimulation in nerve muscle preparations. From Heuser et al. (1979), with permission.

Liquid propane (C3H8; melting point -189°C, boiling point -42°C) and liquid ethane (C2H6; melting point -172°C, boiling point -89°C), cooled with liquid nitrogen to close to their freezing points, are the most efficient cryogens for plunge freezing; they are, however, potentially hazardous and need special care in handling. 1
 

1 Extreme care must always be taken to eliminate any possibility of explosion hazard when working with liquified flammable gases, as ignition of even a small volume of liquid can have devastating consequences. The flash point of liquid propane is -104°C; that of liquid ethane is -130°C. These gases create explosive mixtures in air at concentrations above 3% (ethane) or 2.2% (propane). Beware also that below -183°C oxygen will condense from the air, forming a potentially explosive mixture. All work involving liquifled flammable gases must be undertaken within the confines of an extraction fume cupboard suitable for flammable vapours, and naked flames and electrical switches that might generatesparks must be totally excluded from the work area. The liquified cryogens should not be stored, but should be safely discarded after each experiment either by evaporation within the fume cupboard or, if direct access to outdoors is available, by carefully pouring the liquid onto the ground at a site distant from people, cars, buildings, and other man-made installations. Liquified gases should never be poured down a drain.

The usual mounting method is to sandwich the specimen between a pair of thin supports; not only can these mounts be adapted readily to fulfill the criteria given earlier, but, upon their separation during fracturing in a double-replica device, the fracture plane tends conveniently to follow a superficial well-frozen layer of the specimen, adjacent to the support. If a mechanical plunge-freezing machine is not available, simple manual plunge freezing of specimens into propane under optimized conditions is always worth trying. Sandwich-mounted cell and membrane suspensions, in particular, lend themselves to this approach, and satisfactory results have been obtained in some animal and plant tissues (Severs and Green, 1983; Galway et al., 1995). The entry velocity attainable by manual plunging will not be as high or as reproducible as that achieved using a mechanical device, but the method is simple, has negligible cost, and, with practice, can give satisfactory results.

B. Spray Freezing
Spray freezing is essentially a version of immersion freezing in which the specimen size is reduced to microscopic droplets and, as such, this method is only suitable for suspensions of single cells, organelles, or membranes (Bachmann and Schmitt-Fumian, 1973). The low thermal mass of the individual droplets permits satisfactory freezing in the absence of cryoprotection. Minute droplets of the suspension are sprayed into liquid nitrogen-cooled propane. The propane is then warmed to -85°C and removed using a vacuum pump. The frozen droplets are mixed with butyl benzene at -85°C and transferred to a standard flat-topped specimen support, which is finally immersed in liquid nitrogen to solidify the specimen. Although, as a freezing method for particulate specimens, spray freezing was superseded by other more straightforward methods (e.g., plunge freezing a sandwich-mounted suspension into propane), it forms an integral part of quenched flow devices designed to capture very rapid dynamic cellular events, on a millisecond time scale, at defined intervals after rapid mixing cells with a stimulating agent. A spray-freezing device based on this quench flow principle is available commercially (Balzers Model SFD 010). After spraying the stimulated cells into vigorously stirred propane and evaporating the cryogen, samples may be processed by freeze substitution in methanol containing glutaraldehyde and then either embedded and sectioned or rehydrated and refrozen for freeze fracture (Knoll, 1995).

C. Jet Freezing
In jet freezing, instead of moving the specimen rapidly through the cryogen, the reverse is done; cold propane is squirted at high velocity onto the stationary specimen (for review, see Gilkey and Staehelin, 1986). Most versions of commercially available equipment have dual jets, which simultaneously squirt liquid propane at each side of a sample (typically a cell or membrane suspension) sandwich mounted between a pair of thin copper supports. A double-jet machine from Balzers (Model JFD 030) incorporates a thermostatically controlled specimen environmental chamber, allowing freezing of the specimen from any chosen starting temperature between +10 and +90°C, a feature of particular importance for lipid research. As an alternative, one-sided propane jet freezers may be constructed with the aid of workshop facilities and have been used successfully in a number of laboratories. The safety precautions outlined in Section III,A apply equally to this freezing method.

D. Metal Block Freezing
The metal block freezing technique, also known as slam freezing or metal mirror freezing, has seen wide application in cell biology. The principle of the technique is to bring the biological specimen into rapid and firm contact with the highly polished surface of a pure copper block, cooled by liquid helium (boiling point 4K; -269°C) to a temperature of ~18K (-255°C). Variations on this theme include the use of a silver block in place of copper and cooling by means of liquid nitrogen instead of liquid helium. Various designs of slam freezing machine have been developed, notably by Heuser et al. (1979) and Escaig (1982), and made available commercially. In the Heuser-type apparatus (Fig. 2b), the specimen is fixed beneath the lower end of a vertical plunger, which is then allowed to fall under the influence of gravity. As the specimen falls, a shutter (which protects the liquid helium-cooled copper block from condensation) is opened, the specimen strikes the copper block, and freezing of the surface layer of the specimen is completed within 2 ms. The Escaig device is more sophisticated, using an electromagnetic plunger to propel the specimen, and protection of the helium-cooled block under vacuum until the instant the shutter opens just prior to specimen contact. Both machines are designed to ensure that the specimen does not bounce on impact, but remains applied firmly to the block until removed into liquid nitrogen. Specimen-mounting systems for slam freezing vary according to the apparatus employed, the requirements of the specimen, and its subsequent processing. Most incorporate features to cushion the specimen from the full force of the impact and to limit its flattening, e.g., by mounting within a spacer ring on a piece of rubber foam or fixed lung tissue.

If a sophisticated automated slam freezer is not available, an effective alternative is to use a simple, hand-held copper block (Dempsey and Bullivant, 1976). The block, fitted with a handle, is cooled in liquid nitrogen and is then raised just above the liquid surface. The upper surface is wiped with absorbent tissue and the specimen is then rapidly pushed onto it manually before being dropped into the nitrogen.

Metal block freezing using automated slammers reproducibly gives excellent cryofixation in the surface 10-µm layer of the sample, and satisfactory results can, with experience, also be obtained using simple manual-freezing blocks. Deeper regions of noncryoprotected specimens will always be badly damaged by large ice crystals and compression shock, and so, for freeze-fracture replication, the samples have to be fractured with precision by microtome through the well-frozen surface layer.

IV. HIGH.PRESSURE FREEZING
In high-pressure (hyperbaric) freezing, a sandwichmounted specimen is frozen by double-sided jetting with liquid nitrogen while briefly being subjected to a pressure of 2100 bar (Galway et al., 1995; Kiss and Staehelin, 1995). At such high-pressure, the critical cooling rate needed to limit ice crystals to a size below that causing ultrastructural damage is reduced from 10,000°C/s to approximately 100 to 500°C/s This cryoprotective effect is achieved because high-pressure lowers the freezing point and reduces the rate of ice nucleation and growth to a degree equivalent to that achieved by using 20% glycerol for cryoprotection with conventional immersion freezing. Although high pressure is potentially lethal to cells, major artifacts from this source appear to be avoidable. High-pressure freezing requires specialist equipment, available from Balzers Union (Model HPM 010) and from Leica (Model PI 32-165). The former device freezes specimens up to 0.5 mm in thickness and 1 mm3 in volume; the latter use specimens of up to 0.6 mm in thickness and 1.5 mm3 in volume.

The major advantage of high-pressure freezing is that structure is well preserved to a much greater depth than is possible with any of the ultrarapid freezing techniques discussed previously. True vitrification to an average depth of 200 µm has been reported in test specimens [catalase crystals in sugar solutions; Sartori et al. (1993)]; in plant specimens, no ice crystal damage is apparent to a depth of 600µm in planar samples and 1mm in spherical samples (Gilkey and Staehelin, 1986). High-pressure freezing is thus of particular value in the direct freezing of solid tissue specimens.


V. ADVANTAGES OF ULTRARAPID AND HIGH.PRESSURE FREEZING METHODS

Ultrarapid and high-pressure freezing methods offer a multitude of advantages as preparation methods in cell biology. By avoiding the need for chemical fixation, these cryofixation techniques potentially permit the study of cell structure in a condition close to that existing in life. Because one particular instant in a biological process can be captured, the accumulation of intermediate stages, which may occur during slow death in aldehyde fixatives, is avoided. Living specimens can thus be frozen for ultrastructural examination at known intervals after application of a biological stimulus. This has made it possible to use the electron microscope for studies of transient biological events that are completed within a few seconds or even, in favorable instances, within a few milliseconds. The ability to undertake such direct kinetic studies was a significant breakthrough in cell biology, as previously, sequences of such rapid events could only be guessed at indirectly from images of chemically fixed specimens. Metal block impact freezing, spray freezing, plunge freezing, and jet freezing methods have all been adapted to permit timeresolved analysis of rapid events (for review, see Knoll, 1995).

Another important advantage is that ultrarapidfrozen specimens can be subjected to deep etching or freeze drying, a technique in which water molecules are allowed to sublime from the frozen surface of a fractured (or, in some cases, unfractured) specimen before replication (see article by Shotton). Glycerol cannot be sublimed, but by directly freezing specimens in dilute aqueous solutions, the outer surfaces of membranes, extracellular matrix components, and intracellular cytoskeletal elements can be exposed by deep etching or freeze drying. For deep-etch observations of the cytoskeleton and internal membrane surfaces of cells, a compromise has to be made in order to obtain clean views unobscured by cytoplasmic components. Typical procedures for cultured cells attached to a substrate involve lysing them with Triton X-100 or physically tearing them open by peeling off a strip of nitrocellulose membrane that has been allowed to adhere to their dorsal surfaces. This is followed by rinsing in dilute buffer to remove cytoplasmic components, light fixation with aldehydes, and then immersion in 10-15% methanol immediately prior to freezing. The methanol acts as a cryoprotectant, increasing the depth of adequate freezing, and also has the advantage of being volatile under vacuum at -100°C, thus facilitating the etching process. This application is thus quite distinct from studies aiming to preserve structure in the native state, but it is a fundamentally important one, as it provides access to structural information that cannot be obtained by other electron microscopical methods (Heuser, 1981). Deep etching has also been adapted to study macromolecules absorbed to microscopic mica flakes and other substrates (Heuser, 1989).

In addition to freeze fracture, deep etching, and cryoelectron microscopy, other key routes to the examination of ultrarapid-frozen specimens are freeze substitution and cryoultramicrotomy. Here the ability to preserve epitopes is of prime importance for immunocytochemical studies (see article by Roos et al.). The complementary application of these approaches, together with freeze-fracture cytochemistry (Severs, 1995; Fujimoto, 1997), has wide application in cell biology today.

Acknowledgment
We thank Stephen Rothery for preparation of the figures.

References
Bachmann, L., and Schmiee-Fumian, W. W. (1973). Spray-freezing and freeze-etching. In "Freeze-Etching: Techniques and Applications" (E. L. Benedetti and E Favard, eds.), pp. 73-80. Soci6t6 Frangaise de Microscopie Électronique, Paris.

Dempsey, G. E, and Bullivane, S. (1976). A copper block method for freezing non-cryoprotected tissue to produce ice-crystal-free regions for electron microscopy. J. Microsc. 106, 251-271.

Echlin, P. (1992). "Low-Temperature Microscopy and Analysis." Plenum, New York.

Escaig, J. (1982). New instruments which facilitate rapid freezing at 8BK and 6K. J. Microsc. 126, 221-230.

Fujimoto, K. (1997). SDS-digested freeze-fracture replica labeling electron microscopy to study the two-dimensional distribution of integral membrane proteins and phospholipids in biomembranes: Practical procedure, interpretation and application. Histochem. Cell Biol. 107, 87-96.

Galway, M. E., Heckman, M. E., Hyde, G. J., and Fowke, L. C. (1995). Advances in high-pressure and plunge-freeze fixation. Methods Cell Biol. 49, 3-19.

Gilkey, J. C., and Staehelin, L. A. (1986). Advances in ultrarapid freezing for the preservation of cellular ultrastructure. J. Electron Microsc. Tech. 3, 177-210.

Heuser, J. (1989). Protocol for 3-D visualization of molecules on mica via the quick freeze, deep etch technique. J. Electron Microsc. Tech. 13, 244-263.

Heuser, J. E. (1981). Preparing biological specimens for stereo microscopy by the quick-freeze, deep-etch, rotary-replication technique. Methods Cell Biol. 22, 97-122.

Heuser, J. E., Reese, T. S., Dennis, J., Jan, Y., Jan, L., and Evans, L. (1979). Synaptic vesicle exocytosis captured by quick freezing and correlated with quantal transmitter release. J. Ceil Biol. 81, 275-300.

Kiss, J. Z., and Staehelin, L. A. (1995). High pressure freezing. In "Rapid Freezing, Freeze Fracture and Deep Etching" (N. J. Severs and D. M. Shotton, eds.), pp. 89-104. Wiley-Liss, New York.

Knoll, G. (1995). Time resolved analysis of rapid events. In "Rapid Freezing, Freeze Fracture and Deep Etching" (N. J. Severs and D. M. Shotton, eds.), pp. 105-126. Wiley-Liss, New York.

Pauli, B. U., Weinstein, R. S., Soble, L. W., and Alroy, J. (1977). Freezefracture of monolayer cultures. J. Cell Biol. 72, 763-769.

Robards, A. W., and Sleytr, U. B. (1985). Low temperature methods in biological electron microscopy. In "Practical Methods in Electron Microscopy" (A. M. Glauert, ed.), Vol. 10. Elsevier, Amsterdam.

Ryan, K. P., and Liddicoat, M. I. (1987). Safety considerations regarding the use of propane and other liquified gases as coolants for rapid freezing purposes. J. Microsc. 147, 337-340.

Sartori, N., Richter, K., and Dubochet, J. (1993). Vitrification depth can be increased more than 10-fold by high pressure freezing. J. Microsc. 172, 55-61.

Severs, N. J. (1995). Freeze-fracture cytochemistry: An explanatory survey of methods. In "Rapid Freezing, Freeze Fracture and Deep Etching" (N. J. Severs, and D. M. Shotton, eds.), pp. 173-208. Wiley-Liss, New York.

Severs, N. J., and Green, C. R. (1983). Rapid freezing of unpretreated tissues for freeze-fracture electron microscopy. Biol. Cell. 47, 193-204.

Severs, N. J., and Shotton, D. M. (1995). "Rapid Freezing, Freeze Fracture and Deep Etching." Wiley-Liss, New York.

Steinbrecht, R. A., and Zierold, K. (1987). "Cryotechniques in Biological Electron Microscopy." Springer-Verlag, Berlin.