Freeze-fracturing at low temperatures: I. A device for fracturing biological specimens at 77–10 K under high vacuum

Standard freeze-etching or freeze-cleaving is performed at 173 K in a vacuum of 133 μPa or at 77 K under liquid nitrogen with subsequent transfer of the specimen into a vacuum chamber. It has been suggested that the frequent poor resolution of morphological details, the poor complementarity of innermembrane protein particles and the semi-crystalline substructures in biomembranes are caused by structural distortion or plastic deformation due to sheer forces which occur even at 77 K during fracturing or cleaving. In addition, water contamination and radiant heat damage occurring during replication introduce artefacts to the structural record. These artefacts could be avoided or reduced by lowering the temperature at which fracturing or cleaving and shadowing is carried out, to about 10 K. Therefore, a device for cleaving biological specimens at 15–10 K under high vacuum was constructed. To allow the use of existing equipment, the device was built into a standard Balzers 301 vacuum unit, where the specimen transfer is done via an airlock system which allows hoar frost contamination free transport of the specimen holder onto the specimen table. To reduce or prevent the condensation of water and other residual gases in the vacuum onto the freshly cleaved specimen surface at 10 K, the specimen is surrounded by two cooled surfaces of 6 and 20 K. All condensable gases outside those shielding shrouds will condense on these surfaces before reaching the specimen. This makes it possible to work at a high vacuum of 3 μPa outside the cooled shrouds, which can be reached with standard turbomolecular pumps. The actual vacuum within the cooled shrouds is estimated to be approximately 13 nPa. Residual gas analysis before and during replication reveals equal conditions to ultra high vacuum systems. An analysis of the yeast cell paracrystalline plasmalemma structure shows that the topographic resolution of the crystalline arrays has been improved by working at 12 K. However, plastic deformation still occurs under these conditions. This observation points to the possibility that what is described as plastic deformation, for at least some membrane proteins, may be a loss of resilience at low temperatures.     

Plastic deformation during freeze-cleavage: a review

Examples of plastic deformation are illustrated from a wide range of specimens that have been prepared for electron microscopic examination by fracturing at low temperature prior to replication of the frozen surface. Deformation has been shown to occur in some specimens when fractured at temperatures as low as 4 K. Plastic deformation is recognized in non-biological polymers such as polystyrene and polyacrylate latex spheres, as well as in similar biological molecules such as poly-β-hydroxybutyrate (PHB). However, it is also demonstrated that plastic deformation occurs widely in more complex biological systems, including membranes and protein macromolecules. The interpretation of the structure of fracture faces of frozen membranes, and particularly the lack of complementarity on opposing fracture faces, is discussed in relation to deformation artifacts. It is concluded that very considerable energy must be dissipated as heat during the cleavage process. In the case of some of the latex spheres, the glass transition temperature (T_g) of the bulk polymer can be more than 200 K above the cleavage temperature, and yet plastic deformation still occurs. Once a molecule has deformed, its appearance in the final replica may be significantly changed by heating during evaporation of the replica. An empirical attempt has been made to define the factors leading to the ‘survival’ of a deformed particle. Although the evidence in this review has been drawn from freeze-fracture and freeze-etching studies, it is emphasized that the process of cleaving at low temperature—whether in freeze-fracture techniques or in cryoultramicrotomy—is essentially the same. Therefore the interpretation of structure in ultra-thin frozen sections will also need to allow for the possibility of deformation artefacts.     

Bean-induced loss of organic mass under electronmicroprobe conditions

Techniques for obtaining complementary replicas have already been shown to be valuable in aiding the interpretation of electron micrographs of replicas of specimens prepared by freeze-etching and freeze-fracturing techniques, and in the recognition of artefacts. This paper describes a simple and efficient method of obtaining complementary pairs of replicas of all types of specimen. Ordinary hollow brass rivets are used as specimen holders and are frozen in an end-to-end position using a special pair of forceps. Up to six rivets are placed in a device consisting of a hinged plate held together with clamps against the force of small springs. When the clamps are released, the pair of rivets are separated, fracturing the specimens. The device is easily adapted for use in any type of freeze-etching or high vacuum apparatus. On the example of Saccharomyces cerevisiae the application of the technique to the detection of artefacts in complementary replicas are described. It was shown that the fibrils observed on cross-fractured cell walls are produced by plastic deformation of cell wall components.     

Static and dynamic experiments in cryo-electron microscopy: comparative observations using high-vacuum, low-voltage and low-vacuum SEM

We carried out a unique comparative study between three modes of cryo‐scanning electron imaging: high‐vacuum, low‐voltage and low‐vacuum, using ice cream as a model system. Specimens were investigated both with and without a conductive coating (Au/Pd) and at temperatures for which ice either remains fully frozen (< ?110 °C) or undergoes sublimation (?110 to ?90 °C). At high magnification, high‐vacuum imaging of coated specimens gave the best results for ‘static’ specimens (i.e. containing fully frozen ice). Low voltages, such as 1 kV, could be used for imaging uncoated specimens at high vacuum, although slight ‘classical’ charging artefacts remained an issue, and the reduced electron beam penetration tended to decrease the definition between different microstructural features. However, this mode was useful for observing in situ sublimation from uncoated specimens. Low‐vacuum mode, involving small partial pressures of nitrogen gas, was particularly suited to in situ sublimation work: when sublimation was carried out in low vacuum in the absence of an anti‐contaminator plate, sublimation rates were significantly reduced. This is attributed to a small partial pressure of sublimated water vapour remaining near the specimen surface, enhancing thermodynamic stability.     

Freeze-fracturing at defined temperatures provides information on temperature rise during fracture,and on membrane complementarity

An apparatus is described which allows freeze-fracturing at defined temperatures, followed by immediate immersion of the specimen in liquid nitrogen. Replication is carried out in a Bullivant & Ames (1966) freeze-fracture device. Using yeast as a test specimen, the following results were obtained with the apparatus: (a) It was confirmed that fracturing at temperatures between 243 K and 223 K gave undeformed volcanoes on the PF of the plasma membrane, as shown originally by Steere et al. (1980). (b) Considerable energy is released by the fracturing process, both as shown by thermocouple readings and by the fact that at relatively high fracture temperatures portions of the specimen surface were melted. A temperature rise of 50–70 K was indicated, (c) Under standard conditions, there is a lack of complementarity between the yeast plasma membrane fracture faces, trigonal point particles not being present opposite the corresponding depressions in the array. Fracturing yeast suspended in salt solutions at 203 K demonstrates these particles. Their absence in normal fractures can be explained by ‘secondary fracture’, a concept based on polymer fracture studies.     

Specimen heating during sputter-coating

It is possible to sputter thin films of gold on to surfaces of frozen biological specimens at very low temperatures (< 120 K) without untoward effects from heating. This is achieved by using permanent magnets to confine the plasma and thus to minimize the energy required to give a reasonable sputtering yield. The system described uses only 250 V at 12–15 mA to give 15 nm films within 2–3 min. It is shown, from theory portraying ‘worst-case’ conditions, that the specimen temperature could not increase by more than 6.0 K at equilibrium. Practical results support the theoretical assumptions. Similar considerations have been applied to sputtering at normal ambient temperature where it is shown that appropriate design of simple apparatus and selection of operational conditions can give adequate films in a reasonable time with negligible (< 3 K) temperature rise above the starting temperature.     

Heating of TEM specimens during ion milling

Sample heating during preparation of electron-thin specimens for observation in transmission electron microscopy (TEM) can produce artefacts which invalidate observations. This is particularly true of two-phase materials such as metal matrix composites, for which sample cooling with liquid nitrogen cannot be used to preserve the substructure during milling. A series of experiments is conducted using an age-hardenable aluminium alloy which produces a trace of peak temperature attained by TEM specimens during ion milling. It is shown that peak temperatures of the order of 650 K can be attained using conventional milling parameters; the technique must therefore be used with caution with materials such as metal matrix composites. A simplified one-dimensional heat transfer analysis of the problem is conducted to show that the most critical parameter is heat transfer along the sample holder legs and across interfaces along the heat path. Temperature differentials within the TEM specimen are shown to be less significant, yet these alone are capable of reaching 60 K within a dimpled specimen under usual milling conditions.     

Instrumental requirements for high resolution imaging

The resolution of modern transmission electron microscopes reaches the physical limits imposed by lens aberrations and energy width. One of the many conditions to be fulfilled, the alignment of illuminating and imaging beam onto the coma-free objective axis, is particularly discussed here since axial coma cannot be detected by the usual resolution-checking methods. Space consumption of specimen stages prevents the full utilization of the magnetic saturation limit only in the 100 keV range. With higher energies, this handicap is obviated, and some additional advantages can be gained which promote material investigations at atomic resolution, and which are presently utilized in instrumental research projects. High resolution with biological specimens has up to now been unsuccessful because of radiation damage. Optimum utilization of all electrons scattered at the specimen must thus be given priority over optical resolution. Important instrumental requirements are minimum exposure beam control, imaging modes with high collection efficiency, and recording devices with high detection quantum efficiency connected on-line to image processors. A remarkable decrease in beam sensitivity of organic crystals, by more than one order, has been found by cooling the specimen down to 4 K which, by the use of superconducting lenses, can be combined with both ultra high vacuum and the stability requirements for high resolution. Yet up to now, such protection has not been achieved with He cryostates in conventional lenses, perhaps because a temperature increase even of only a few degree K is harmful. Purely magnetic imaging energy filters are about to be developed to a high optical quality but have been employed so far in only a few high resolution instruments. Such filters allow removal of the inelastic background and thus improvement of contrast of images of low-Z specimens, particularly in the dark field mode. Finally, some ‘non-conventional’ projects have made progress. Correction of spherical and chromatic aberration by multipole lenses offers a chance to improve remarkably the resolution in the 100 keV range, to extend the bandwidth of phase contrast transfer and to obtain highly resolved information about inelastic images when an energy filter is also applied. Electron holography provides possibly useful large area phase contrast, particularly if the electron energy is decreased, which may be of great benefit in investigations of unstained specimens.     

Developments of new Lowicryl® resins for embedding biological specimens at even lower temperatures

Two new Lowicryl resins have been developed for embedding biological materials at temperatures down to 210K (hydrophilic K11M) and to 190K (hydrophobic HM23). They have similar properties to Lowicryl K4M and HM20. The new resins were first tested for low temperature applications by the ‘progressive lowering of temperature’ procedure and this shows that the low viscosity of K11M and HM23 is favourable for the infiltration of biological specimens. Hardening is achieved through photo-polymerization at these lower temperatures. These properties make K11M and HM23 suitable for cryosubstitution of rapidly frozen material and it is speculated that the preservation of antigenicity may be further improved.     

A freeze-fracture replication apparatus for biological specimens.

A freeze-fracture apparatus of original design has been constructed which can be fitted onto a standard vacuum evaporator unit. In it, cell suspensions and organized tissue may be processed by inserting a sample into a cylindrical holder. By leaving a small part of the tissue protruding from the holder, pre-selected and aligned portions of the specimen can subsequently be revealed by fracture under vacuum. After rapid freezing, the specimen remains firmly attached to the inner wall of the sample holder, preventing its possible loss during fracturing. A mechanism, in the form of a double-sided converging wedge, which is operated from outside the vacuum chamber, is used to produce a fracture in the specimen. The device gently induces a fracture in the desired part of the tissue and lifts the protruding part of the specimen out of the way. In this way, reasonably flat fracture faces are produced for subsequent replication. As the fracturing mechanism comes into contact only with the outer edges of the specimen, damage and contamination liable to occur when the entire specimen is traversed by a blade, is avoided. In addition the specimen stage is surrounded by a cold metal shroud which acts as an efficient trap for contaminants. In this way, favourable vacuum conditions are produced in the vicinity of the specimen. Such effective enclosing of the specimen also facilitates controlled sublimation of the sample.     

Ultracryotomy: a technique for cutting ultrathin sections of unfixed frozen biological tissues for electron microscopy

An apparatus is described which facilitates the sectioning of unfixed and unembedded biological tissues for electron microscopy. The specimen is maintained at the temperature of boiling liquid nitrogen; the knife temperature may be controlled between ?50 and ? 150°C. Under the worst operating conditions specimen drift with respect to the knife is about 5 nm/sec, and suggestions are made on how this drift may be further reduced. We propose to call the apparatus an ‘ultracryotome’. Electron micrographs of a variety of unfixed, unembedded and unstained tissues are shown and commented on. Some developments of experimental electron microscopy which may be expected to develop from use of the ultracryotome are listed.     

Preparation of ultrathin amorphous ice films for cryo-electron microscopy

Electron microscopy of biological macromolecules embedded in vitrified ice films suffers from serious problems caused by excessive inelastic scattering, beam-induced movements of the specimen, deformation of the molecules by adsorption at the water-air interface and insufficient mechanical stability of the films. We have built an environmental chamber to control temperature and humidity independently in order to produce ultrathin water films (< 20 nm) spanning holes with diameters of 20 μm to 1 mm. The surface tension of the water films was reduced by adding lipid monolayers, thus prolonging the usable time for thinning of the film and avoiding adsorption artefacts in the embedded material. After cryofixation in ethane a carbon film was evaporated on each side of the specimen to stabilize the ultrathin ice—lipid film. Mechanical stability and charging effects could successfully be reduced by this preparation method. Collapsing water films could be cryofixed and the shape of the hole was analysed. By the eccentricity of the elliptical holes an estimation could be made of the burst velocity of the rim of the hole and of the cooling rate of the cryofixation process.     

Electron probe x-ray microanalysis of frozen-hydrated biological specimens.

A technique is described for preparing frozen-hydrated bulk samples of biological specimens for electron probe X-ray microanalysis. The method allows reproducible quantitative analyses to be made. Specimens are rapidly frozen, transferred to a vacuum evaporator, fractured under high vacuum at - 180 degrees C and coated with 20 nm of chromium. Transferal to the cryostage of a scanning electron microscope is accomplished without exposure to the atmosphere and without the specimen temperature rising above -120 degrees C. Analyses are made at a temperature of -145 degrees C. Contamination by frost does not occur. Etching and charging of the specimen are eliminated. Specimen charging is shown to be related to temperature. It can be eliminated at low temperature by coating with carbon, aluminium or chromium but consistent elimination could only be achieved with chromium. The chromium coat does not appear to have an absorption effect on quantitative analysis.     

The direct measurement of temperature changes within freeze-fracture specimens during rapid quenching in liquid coolants

We have evaluated the cooling rates of specimens mounted in a variety of freeze-fracture holders when plunged into a series of liquid coolants. These rates were measured using miniature thermocouples placed within the mounted specimens. The most rapid cooling rates were obtained using propane at 83 K as the coolant. When mounted on a newly devised ‘copper sandwich’ holder, specimen cooling rates in excess of 4500 K/s have been recorded. A simple guillotine-like device for quenching freeze-fracture specimens under reproducible conditions is presented.     

A temperature-jump device for time-resolved cryo-transmission electron microscopy.

We describe a temperature-jump device that permits time-resolved studies of thin cryo-transmission electron microscopy specimens. The specimen is rapidly heated to induce a change in microstructure just prior to cryo-fixation. The apparatus consists of a xenon arc lamp equipped with a shutter controlled by timing circuitry, used in conjunction with an environmental specimen preparation chamber. The specimen is heated by exposure to focused light from the lamp, and then plunged into cryogen. Using a thermocouple constructed from an electron microscope grid, we show that temperature jumps of 30-60 K are achieved with exposure times of 150-450 milliseconds. Micrographs of dimyristoyl phosphatidylcholine (DMPC) vesicles and n-docosane films, subjected to these exposures, show that the specimens are still at least 20-30 K above their initial temperature when they contact the cryogen. This method could be applied to a variety of biological and chemical systems which undergo structural changes activated by a rise in temperature.     

Surfaces of cryosections: is cryosectioning ‘cutting’ or ‘fracturing’?

In order to determine if cryosectioning involves ‘fracturing’ or ‘cutting’ we examined the surfaces obtained in cryosectioning by a metal-replicating procedure commonly used in freeze-fracture microscopy. Platinum-carbon replicas were made of the surfaces of both the sections and the complementary surfaces of the sample stubs from which the sections were cut. When samples of frozen red cells were sectioned at ?120°C with large knife advancements (1 μm), the chips produced did not resemble sections. Membrane fracture faces, produced by splitting of the lipid bilayer, were found in electron micrographs of replicas of the sample stubs. This demonstrates that a cryomicrotome can be used to produce large intact replicas. When dull knives were used with small knife advancements, both smooth and fractured regions were found. The sections produced with dull knives had a snowflake appearance in the light microscope. When sharp knives were used with small advancements (0·1 μm), replicas of the surfaces were free of fracture faces and the sections had a cellophane-like appearance in the light microscope. Therefore, in cryosectioning a different process other than ‘fracturing’ is responsible. This ‘cutting’ process may be micromelting of a superficial layer by the mechanism of melting-point depression from the pressure exerted by the sharp edge of the knife.     

Understanding the deformation behavior of 17-4 precipitate hardenable stainless steel produced by direct metal laser sintering using micropillar compression and TEM

Micropillar compression testing and transmission electron microscopy were used to investigate the deformation mechanisms under compression in 17-4 precipitation-hardening stainless steel, fabricated by direct metal laser sintering. The as-built specimen as well as that aged for 4 h at 866 K contained fractions of retained austenite, while that aged for 60 min at 755 K was fully martensitic. It was observed that the columnar elongated grains underwent martensite/austenite phase changes with changes in the aging temperature. The precipitate phase that developed with aging enhanced the material hardness and yield strength, with values being higher in the case of the specimen aged at a lower temperature for a shorter time. The results showed that the microstructures and properties of 17-4 stainless steel specimens fabricated by DMLS vary significantly from those of specimens produced using conventional methods.     

Early results using high-resolution,low-voltage,low-temperature SEM

Recent advances in the design of the scanning electron microscope (SEM) column, such as the coupling of a field-emission gun to a low-aberration immersion lens and the availability of a high-stability cryo-transfer stage, make low-temperature, low-voltage SEM (LTLVSEM) possible at very high resolution. We have used this combination to obtain results with uncoated biological specimens. The trichocyst from a Paramecium was used as a test specimen to observe the shrinkage of this structure as the temperature is raised from 170 K to room temperature following freeze-drying. High-magnification stereo images were obtained of trichocysts that had been prepared by freezing, freeze-substitution and critical-point drying and which were subsequently viewed by LTLVSEM to reduce beam damage and contamination.     

EM autoradiography of non-sectioned biological material

Methods of EM autoradiography suitable for the study of non-sectioned biological material deposited on a support film from suspension are described. The methods, which include modified procedures for specimen support and emulsion preparation, may also have advantages for the study of any specimen where physical separation of specimen and emulsion is desirable. In a study of the growth of ¹²⁵I-labelled collagen fibrils reconstituted from solution in vitro, these procedures have allowed the production of EM autoradiographs which show a high resolution and are free from stress artefacts and chemography.     

High-pressure scanning electron microscopy of insulating materials: A new approach

A new SEM technique for imaging uncoated non-conducting specimens at high beam voltages is described which employs a high-pressure environment and an electric field to achieve charge neutralization. During imaging, the specimen surface is kept at a stable low voltage, near earth potential, by directing a flow of positive gas ions at the specimen surface under the action of an electric bias field at a pressure of about 200 Pa. In this way charge neutrality is continuously maintained to obtain micrographs free of charging artefacts. Images are formed by specimen current detection containing both secondary electron and backscattered electron signal information. Micrographs of geological, ceramic, and semiconductor materials obtained with this method are presented. The technique is also useful for the SEM examination of histological sections of biological specimens without any further preparation. A simple theory for the charge neutralization process is described. It is based on the interaction of the primary and emissive signal components with the surrounding gas medium and the resulting neutralizing currents. Further micrographs are presented to illustrate the pressure dependence of the charge neutralization process in two glass specimens which show clearly identifiable charging artefacts in conventional microscopy.