Cryo-EM--the first thirty years

Thirty years ago, in December 1981, The Journal of Microscopy published a very short paper entitled ‘Vitrification of pure water for electron microscopy’. It turned out to be important for the development of cryo-electron microscopy and it contributed to reverse, from foe to friend, the status of water in electron microscopists’ minds. This change has brought obvious gains. The future will tell how many more are still to come.

Compression and crevasses in vitreous sections under different cutting conditions

Compression and crevasses are common cutting artefacts in cryo-ultramicrotomy of vitreous sections. They can be reduced or suppressed under optimal cutting conditions. In the present study, compression and thickness were measured for different cutting speeds and knife angles. It was found that compression decreased with feed and that crevasses appeared only above a certain thickness. The optimal feed for vitreous sections was between 50 and 80 nm. The thickness, calculated by two independent methods, was quantitatively related to feed and compression.

Cryoelectron microscopy of vitrified sections: a new challenge for the analysis of functional nuclear architecture

Cryoelectron microscopy of vitrified sections has become a powerful tool for investigating the fine structural features of cellular compartments. In the present study, this approach has been applied in order to explore the ultrastructural morphology of the interphase nucleus in different mammalian cultured cells. Rat hepatoma, Chinese hamster ovary and Potorus kidney cells were cryofixed by high-pressure freezing and the cryosections were examined at low temperature by transmission electron microscopy. Our results show that while the contrast of nuclear structural domains is remarkably homogeneous in hydrated sections, some of them can be recognised due to their characteristic texture. Thus, condensed chromatin appears finely granular and the perichromatin region contains rather abundant fibro-granular elements suggesting the presence of dispersed chromatin fibres and of perichromatin fibrils and granules. The interchromatin space looks homogeneous and interchromatin granules have not been identified under these preparative conditions. In the nucleolus, the most striking feature is the granular component, while the other parts of the nucleolar body, which appear less contrasted, are difficult to resolve. The nuclear envelope is easily recognisable with its regular perinuclear space and nuclear pore complexes. Our observations are discussed in the context of results obtained by other, more conventional electron microscopic methods.

An oscillating cryo-knife reduces cutting-induced deformation of vitreous ultrathin sections

A new oscillating cryo-knife for producing uncompressed vitreous sections is introduced. The knife is a modified cryo diamond knife that is driven by a piezo translator. Optimal setting for the oscillation was found to be in the inaudible frequency range of 20-25 kHz. Yeast cells and polystyrene spheres were used as model systems to describe compression in the vitreous sections. We found that compression could be reduced by a factor of about 2 when the knife was oscillating. When the oscillator was turned off, sections were compressed by 40-45%. However, only 15-25% compression was obtained when the knife was oscillating. In some cases completely uncompressed sections of yeast cells were produced. It was also found that the amount of compression depends on the specimen itself and on its embedding medium. With the results shown here, we demonstrate that the oscillating knife can produce high-quality vitreous sections with minimum cutting artefacts.

Amorphous solid water produced by cryosectioning of crystalline ice at 113 K

Amorphous solid (vitreous) water can be obtained by a number of methods, including quick freezing of a very small volume of pure water, low pressure condensation of water vapour on a cold substrate or transformation of hexagonal ice (the ice which is naturally formed) under very high pressure at liquid nitrogen temperature. Larger volumes can be vitrified if cryoprotectant is added or when samples are frozen under high pressure. We show that a sample of 17.5% dextran solution or mouse brain tissue, respectively, frozen under high pressure (200 MPa) into cubic or hexagonal ice can be transformed into vitreous water by the very process of cryosectioning. The vitreous sections obtained by this procedure differ from cryosections obtained from vitreous samples by the irregular aspect of the sections and by small but significant differences in the electron diffraction patterns. For the growing community of cryo-ultramicrotomists it is important to know that vitrification can occur at the knife edge. A vitreous sample is considered to show the best possible structural preservation. The sort of vitrification described here, however, can lead to bad structural preservation and is therefore considered to be a pitfall. Furthermore, we compare these sections with other forms of amorphous solid water and find it similar to high density amorphous water produced at very high pressures (about 1 GPa) from hexagonal ice and annealed close to its transformation temperature at 117 K.

Cryo-negative staining reduces electron-beam sensitivity of vitrified biological particles

Beam damage is the main resolution-limiting factor when biological particles are observed by cryoelectron microscopy in a thin vitrified solution film. Furthermore, the low contrast of the specimen frequently makes observation difficult and limits the possibility of image processing. Cryo-negative staining, in which the particles are vitrified in a thin layer of concentrated ammonium molybdate solution, makes it possible to visualize the particles with a much better signal-to-noise ratio (SNR) while keeping the specimen in a good state of preservation. We have observed the Escherichia coli GroEL chaperonin, prepared in a native vitrified solution and by cryo-negative staining after electron exposure from 1000 to 3000e(-)/nm(2). We have compared the resulting three-dimensional models obtained from these different conditions and have tested their fit with the atomic model of the protein subunit obtained from X-ray crystallography. It is found that, down to 1.5-nm resolution, the particles appear to be faithfully represented in the cryo-negatively stained preparation, but there is an approximately 10-fold increase of SNR compared with the native vitrified preparation. Furthermore, for the same range of irradiation and down to the same resolution, the particles seem unaffected by beam damage, whereas the damage is severe in the native vitrified particles.

Bilayers of nucleosome core particles

Among the multiple effects involved in chromatin condensation and decondensation processes, interactions between nucleosome core particles are suspected to play a crucial role. We analyze them in the absence of linker DNA and added proteins, after the self-assembly of isolated nucleosome core particles under controlled ionic conditions. We describe an original lamellar mesophase forming tubules on the mesoscopic scale. High resolution imaging of cryosections of vitrified samples reveals how nucleosome core particles stack on top of one another into columns which themselves align to form bilayers that repel one another through a solvent layer. We deduce from this structural organization how the particles interact through attractive interactions between top and bottom faces and lateral polar interactions that originate in the heterogeneous charge distribution at the surface of the particle. These interactions, at work under conditions comparable with those found in the living cell, should be of importance in the mechanisms governing chromatin compaction in vivo.

DNA in human and stallion spermatozoa forms local hexagonal packing with twist and many defects

In human and other mammal sperm nuclei, DNA is packed in a highly condensed state, the structure of which remains unsolved. Cryoelectron microscopy of vitrified sections provides a first direct view of the local arrangement of the nucleoprotamine filament. DNA aligns in parallel in layers and its orientation rotates along a single-twist direction as in a cholesteric liquid crystal. The structure contains numerous defects, which introduce locally double-twist configurations. Destruction of the SS bonds with dithiotrehitol relaxes the twist and favors the extension of the hexagonal close packing of the filaments, though keeping constant their interfilament distance.

The cell in absence of aggregation artifacts

Eduard Kellenberger understood that the conventional resin-embedding, he helped to develop (Ryter, A., Kellenberger, E., 1958. L'inclusion au polyester pour l'ultramicrotomie. J. Ultrastruct. Res. , 2, 200-214), was prone to aggregation artifacts (Kellenberger, E., 1987. The response to biological macromolecules and supramolecular structures to the physics of specimen cryo-preparation. In: Steinbrecht, R.A., Zierold, K. (Eds.), Cryo-techniques in Biological Electron Microscopy, Springer, Berlin, pp. 35-63). He was instrumental in developing various methods to overcome this limitation, for instance, by using low temperature-embedding and partially hydrophilic resins (Carlemalm, E., Garavito, R.M., Villiger, W., 1982. Resin development for electron microscopy and an analysis of embedding at low temperature. J. Microstruct., 126, 123-143; Villiger,W., 1993. Low temperature-embedding with Lowicryl resins. In: Robards, A.W., Wilson, A.J. (Eds.), Procedures in electron microscopy, Wiley, Chichester, UK, pp. 16:7.3-16:7.6). In principle, cryo-electron microscopy of vitreous sections is free of any aggregation artifact since the material remains fully hydrated and is free of chemical fixation or staining. The method is technically difficult still, but recent progress has made it amenable to routine practical applications. We compare here electron microscopical aspects of Zea mays meristem cells prepared by: (1) conventional resin-embedding and sectioning; (2) low temperature-embedding and sectioning of freeze substituted samples; and (3) cryo-sections of vitrified samples. The appearance of the extra-cellular space, the cytoplasm and the nucleoplasm are very different in conditions (1) and (3). They appear as compact, irregular and well delineated structures in conventional resin sections, whereas they are more diffuse and homogeneous in the vitreous sections. In the resin sections, the material seems to form a complex matrix, whereas it looks more like a thick soup in the vitreous sample. Low temperature-embedding (condition 2) shows an intermediate appearance. We suggest that regardless of the difference due to staining and different sectioning conditions, the other image differences are the consequence of aggregation artifacts in the resin-embedded specimens.

Making science in a fractal landscape

There were times when any progress in science was associated with a social benefit. Public support of science was growing continuously without many questions being asked. These times have gone and the public asks now what science is good for and scientists are not yet well prepared to reply. A new social contract between scientists and society should be elaborated based on knowledge, tolerance and transparency.

Self-assembly of beta-amyloid 42 is retarded by small molecular ligands at the stage of structural intermediates

Assemblyof the amyloid-beta peptide (Abeta) into fibrils and its deposition in distinct brain areas is considered responsible for the pathogenesis of Alzheimer's disease (AD). Thus, inhibition of fibril assembly is a potential strategy for therapeutic intervention. Electron cryomicroscopy was used to monitor the initial, native assembly structure of Abeta42. In addition to the known fibrillar intermediates, a nonfibrillar, polymeric sheet-like structure was identified. A temporary sequence of supramolecular structures was revealed with (i) polymeric Abeta42 sheets during the onset of assembly, inversely related to the appearance of (ii) fibril intermediates, which again are time-dependently replaced by (iii) mature fibrils. A cell-based primary screening assay was used to identify compounds that decrease Abeta42-induced toxicity. Hit compounds were further assayed for binding to Abeta42, radical scavenger activity, and their influence on the assembly structure of Abeta42. One compound, Ro 90-7501, was found to efficiently retard mature fibril formation, while extended polymeric Abeta42 sheets and fibrillar intermediates are accumulated. Ro 90-7501 may serve as a prototypic inhibitor for Abeta42 fibril formation and as a tool for studying the molecular mechanism of fibril assembly.

Tightness of random knotting

Long polymers in solution frequently adopt knotted configurations. To understand the physical properties of knotted polymers, it is important to find out whether the knots formed at thermodynamic equilibrium are spread over the whole polymer chain or rather are localized as tight knots. We present here a method to analyze the knottedness of short linear portions of simulated random chains. Using this method, we observe that knot-determining domains are usually very tight, so that, for example, the preferred size of the trefoil-determining portions of knotted polymer chains corresponds to just seven freely jointed segments.

Brownian dynamics simulation of DNA condensation

DNA condensation observed in vitro with the addition of polyvalent counterions is due to intermolecular attractive forces. We introduce a quantitative model of these forces in a Brownian dynamics simulation in addition to a standard mean-field Poisson-Boltzmann repulsion. The comparison of a theoretical value of the effective diameter calculated from the second virial coefficient in cylindrical geometry with some experimental results allows a quantitative evaluation of the one-parameter attractive potential. We show afterward that with a sufficient concentration of divalent salt (typically approximately 20 mM MgCl(2)), supercoiled DNA adopts a collapsed form where opposing segments of interwound regions present zones of lateral contact. However, under the same conditions the same plasmid without torsional stress does not collapse. The condensed molecules present coexisting open and collapsed plectonemic regions. Furthermore, simulations show that circular DNA in 50% methanol solutions with 20 mM MgCl(2) aggregates without the requirement of torsional energy. This confirms known experimental results. Finally, a simulated DNA molecule confined in a box of variable size also presents some local collapsed zones in 20 mM MgCl(2) above a critical concentration of the DNA. Conformational entropy reduction obtained either by supercoiling or by confinement seems thus to play a crucial role in all forms of condensation of DNA.

Unexpected property of trehalose as observed by cryo-electron microscopy

Trehalose is an agent useful in maintaining the integrity of many biological systems submitted to various stresses. It is also presumed to improve specimen preparation for electron microscopy and to reduce beam damage. Here we study the effect of trehalose on the preparation and observation by cryo-electron microscopy of thin vitrified films of biological suspensions. We observe that trehalose, as compared to sucrose, can indeed reduce electron beam damage to biological particles, as determined from the dose necessary for the onset of bubbling. Surprisingly, we also find that the contrast of biological particles is higher in a vitrified solution of trehalose than in one of sucrose. This effect can be explained if the water evaporation during the specimen preparation is less in the presence of trehalose than with sucrose, but we do not yet understand the underlying reasons since the evaporation properties of both sugars are similar at a macroscopic level. We conclude that trehalose is truly a remarkable substance and that more investigation is needed in order to fully understand its properties, and that the addition of ca. 3-5% trehalose to biological suspensions is a simple and useful method to reduce commonly arising drying artefacts and water evaporation in the thin film vitrification method.

Electron beam-induced changes in vitreous sections of biological samples

Nonpretreated high pressure frozen samples of Zea mays, cartilage and human erythrocytes were cryosectioned and observed at 110K in a cryoelectron microscope. Changes induced by medium doses of electron irradiation (< 10 ke nm-2) are described. After some ke nm-2, the most conspicuous cutting artefacts are erased to a large extent and the visibility of the cell organelles is improved. The sections, compressed in the cutting direction by the sectioning process, shrink once more, in the same direction, when irradiated. This shrinkage depends on the section support and on how the section is adsorbed to it. Shrinkage is not uniform: it is most pronounced in mitochondria, condensed chromatin and nucleolus. This differential shrinkage improves the visibility of major structures on the section and, as a result, 'nicer' images are recorded. However, this apparent improvement is a beam-induced artefact that must be paired with a loss of high resolution information.

Are the light-harvesting I complexes from Rhodospirillum rubrum arranged around the reaction centre in a square geometry?

The basic photosynthetic unit containing the reaction centre and the light-harvesting I complex (RC-LHI) of the purple non-sulphur bacterium Rhodospirillum rubrum was purified and reconstituted into two-dimensional (2D) membrane crystals. Transmission electron microscopy using conventional techniques and cryoelectron microscopy of the purified single particles and of 2D crystals yielded a projection of the RC-LHI complex at a resolution of at least 1.6 nm. In this projection the LHI ring appears to have a square symmetry and packs in a square crystal lattice. The square geometry of the LHI ring was observed also in images of single isolated particles of the RC-LHI complex. However, although the LHI units are packed identically within the crystal lattice, a new rotational analysis developed here showed that the reaction centres take up one of four possible orientations within the ring. This fourfold disorder supports our interpretation of a square ring symmetry and suggests that a hitherto undetected component may be present within the photosynthetic unit.

Geometry and physics of catenanes applied to the study of DNA replication

The concept of ideal geometric configurations was recently applied to the classification and characterization of various knots. Different knots in their ideal form (i.e., the one requiring the shortest length of a constant-diameter tube to form a given knot) were shown to have an overall compactness proportional to the time-averaged compactness of thermally agitated knotted polymers forming corresponding knots. This was useful for predicting the relative speed of electrophoretic migration of different DNA knots. Here we characterize the ideal geometric configurations of catenanes (called links by mathematicians), i.e., closed curves in space that are topologically linked to each other. We demonstrate that the ideal configurations of different catenanes show interrelations very similar to those observed in the ideal configurations of knots. By analyzing literature data on electrophoretic separations of the torus-type of DNA catenanes with increasing complexity, we observed that their electrophoretic migration is roughly proportional to the overall compactness of ideal representations of the corresponding catenanes. This correlation does not apply, however, to electrophoretic migration of certain replication intermediates, believed up to now to represent the simplest torus-type catenanes. We propose, therefore, that freshly replicated circular DNA molecules, in addition to forming regular catenanes, may also form hemicatenanes.

Sedimentation and electrophoretic migration of DNA knots and catenanes

Various site-specific recombination enzymes produce different types of knots or catenanes while acting on circular DNA in vitro and in vivo. By analysing the types of knots or links produced, it is possible to reconstruct the order of events during the reaction and to deduce the molecular "architecture" of the complexes that different enzymes form with DNA. Until recently it was necessary to use laborious electron microscopy methods to identify the types of knots or catenanes that migrate in different bands on the agarose gels used to analyse the products of the reaction. We reported recently that electrophoretic migration of different knots and catenanes formed on the same size DNA molecules is simply related to the average crossing number of the ideal representations of the corresponding knots and catenanes. Here we explain this relation by demonstrating that the expected sedimentation coefficient of randomly fluctuating knotted or catenated DNA molecules in solution shows approximately linear correlation with the average crossing number of ideal configurations of the corresponding knots or catenanes.

Cryo-negative staining

A procedure is presented for the preparation of thin layers of vitrified biological suspensions in the presence of ammonium molybdate, which we term cryo-negative staining. The direct blotting of sample plus stain solution on holey carbon supports produces thin aqueous films across the holes, which are routinely thinner than the aqueous film produced by conventional negative staining on a continuous carbon layer. Because of this, a higher than usual concentration of negative stain (ca. 16% rather than 2%) is required for cryo-negative staining in order to produce an optimal image contrast. The maintenance of the hydrated state, the absence of adsorption to a carbon film and associated sample flattening, together with reduced stain granularity, generates high contrast cryo-images of superior quality to conventional air-dry negative staining. Image features characteristic of unstained vitrified cryo-electron microscopic specimens are present, but with reverse contrast. Examples of cryo-negative staining of several particulate biological samples are shown, including bacteriophage T2, tobacco mosaic virus (TMV), bovine liver catalase crystals, tomato bushy stunt virus (TBSV), turnip yellow mosaic virus (TYMV), keyhole limpet hemocyanin (KLH) types 1 and 2, the 20S proteasome from moss and the E. coli chaperone GroEL. Densitometric quantitation of the mass-density of cryo-negatively stained bacteriophage T2 specimens before and after freeze-drying within the TEM indicates a water content of 30% in the vitreous specimen. Determination of the image resolution from cryo-negatively stained TMV rods and catalase crystals shows the presence of optical diffraction data to ca. 10 A and 11.5 A, respectively. For cryo-negatively stained vitrified catalase crystals, electron diffraction shows that atomic resolution is preserved (to better than 20 diffraction orders and less than 3 A). The electron diffraction resolution is reduced to ca. 10 A when catalase crystal specimens are prepared without freezing or when they are freeze-dried in the electron microscope. Thin vitrified films of TMV, TBSV and TYMV in the presence of 16% ammonium molybdate show a clear indication of two-dimensional (2-D) order, confirmed by single particle orientational analysis of TBSV and 2-D crystallographic analysis of TYMV. These observations are in accord with earlier claims that ammonium molybdate induces 2-D array and crystal formation from viruses and macromolecules during drying onto mica. Three-dimensional analysis of the TBSV sample using the tools of icosahedral reconstruction revealed that a significant fraction of the particles were distorted. A reconstruction from a subset of undistorted particles produced the characteristic T = 3 dimer clustered structure of TBSV, although the spikes are shortened relative to the structure defined by X-ray crystallography. The 20S proteasome, GroEL, catalase, bacteriophage T2, TMV, TBSV and TYMV all show no indication of sample instability during cryo-negative staining. However, detectable dissociation of the KLH2 oligomers in the presence of the high concentration of ammonium molybdate conforms with existing knowledge on the molybdate-induced dissociation of this molecule. This indicates that the possibility of sample-stain interaction in solution, prior to vitrification, must always be carefully assessed.

Properties of ideal composite knots

The shortest tube of constant diameter that can form a given knot represents the 'ideal' form of the knot. Ideal knots provide an irreducible representation of the knot, and they have some intriguing mathematical and physical features, including a direct correspondence with the time-averaged shapes of knotted DNA molecules in solution. Here we describe the properties of ideal forms of composite knots-knots obtained by the sequential tying of two or more independent knots (called factor knots) on the same string. We find that the writhe (related to the handedness of crossing points) of composite knots is the sum of that of the ideal forms of the factor knots. By comparing ideal composite knots with simulated configurations of knotted, thermally fluctuating DNA, we conclude that the additivity of writhe applies also to randomly distorted configurations of composite knots and their corresponding factor knots. We show that composite knots with several factor knots may possess distinct structural isomers that can be interconverted only by loosening the knot.

Opposite effect of counterions on the persistence length of nicked and non-nicked DNA

Using cryo-electron microscopy we reconstructed the three-dimensional trajectories adopted in cryovitrified solutions by double-stranded DNA molecules in which the backbone of one strand lacked a phosphate at regular intervals of 20 nucleotides. The shape of such nicked DNA molecules was compared with that of DNA molecules with exactly the same sequence but without any single-stranded scissions. Upon changing the salt concentration we observed opposite effects of charge neutralization on nicked and non-nicked DNA. In low salt solutions (10 mM Tris-HCl, 10 mM NaCl) the applied dense nicking caused ca 3.5-fold reduction of the DNA persistence length as compared with non-nicked DNA. Upon increasing the salt concentration (to 150 mM NaCl and 10 mM MgCl2) the persistence length of non-nicked DNA appreciably decreased while that of nicked DNA molecules increased by a factor of 2.

Comparison of slam-freezing and high-pressure freezing effects on the DNA cholesteric liquid crystalline structure

Using in parallel electron microscopy of ultrathin frozenhydrated sections and freeze-fracture replicas, we compare the ultrastructural consequences of two freezing techniques: slam-freezing at liquid helium temperature and high-pressure freezing, on a model system, the DNA cholesteric liquid crystalline phase. Both freezing techniques are able to vitrify DNA liquid crystalline solutions containing up to 85% water, but induce structural rearrangements of the molecular organization. The cholesteric structure is preserved by the slam-freezing method despite the formation of periodic distortions induced by the mechanical compressive stress. In contrast, high-pressure freezing does not preserve the structure of the liquid crystal: the long-range cholesteric stratification disappears, and the local continuous twist between molecules is modified. These results show that vitrification, though necessary, may not be a sufficient token of preservation of the native state of hydrated materials. We discuss the possible origins of the molecular rearrangements that have time to occur in the specimens as a result of the low freezing rate permitted by the high-pressure freezing process.

The flying cylinder: a new algorithm for filament recognition in noisy stereo images

We present a new method of automatic 3D filament representation which uses stereo micrographs to reconstruct three-dimensional trajectories of filament-like objects as DNA molecules. The method deals with low contrasted and noisy images, as obtained from cryovitrified samples by means of electron microscopy. The three-dimensional information is extracted from skeletizing simultaneously both images of a given stereo-pair, instead of processing them separately. The main advantages of the technique are reproducibility and speed, compared to the reconstruction done by manual registration, i.e., clicking on the stereo micrographs.

Determination of DNA persistence length by cryo-electron microscopy. Separation of the static and dynamic contributions to the apparent persistence length of DNA

Axial deflection of DNA molecules in solution results from thermal motion and intrinsic curvature related to the DNA sequence. In order to measure directly the contribution of thermal motion we constructed intrinsically straight DNA molecules and measured their persistence length by cryo-electron microscopy. The persistence length of such intrinsically straight DNA molecules suspended in thin layers of cryo-vitrified solutions is about 80 nm. In order to test our experimental approach, we measured the apparent persistence length of DNA molecules with natural "random" sequences. The result of about 45 nm is consistent with the generally accepted value of the apparent persistence length of natural DNA sequences. By comparing the apparent persistence length to intrinsically straight DNA with that of natural DNA, it is possible to determine both the dynamic and the static contributions to the apparent persistence length.

Chromatin conformation and salt-induced compaction: three-dimensional structural information from cryoelectron microscopy

Cryoelectron microscopy has been used to examine the three-dimensional (3-D) conformation of small oligonucleosomes from chicken erythrocyte nuclei after vitrification in solutions of differing ionic strength. From tilt pairs of micrographs, the 3-D location and orientation of the nucleosomal disks, and the paths of segments of exposed linker can be obtained. In "low-salt" conditions (5 mM NaCl, 1 mM EDTA, pH 7.5), the average trinucleosome assumes the shape of an equilateral triangle, with nucleosomes at the vertices, and a length of exposed linker DNA between consecutive nucleosomes equivalent to approximately 46 bp. The two linker DNA segments converge at the central nucleosome. Removal of histones H1 and H5 results in a much more variable trinucleosome morphology, and the two linker DNA segments usually join the central nucleosome at different locations. Trinucleosomes vitrified in 20 mM NaCl, 1 mM EDTA, (the salt concentration producing the maximal increase in sedimentation), reveal that compaction occurs by a reduction in the included angle made by the linker DNA segments at the central nucleosome, and does not involve a reduction in the distance between consecutive nucleosomes. Frequently, there is also a change in morphology at the linker entry-exit site. At 40 mM NaCl, there is no further change in trinucleosome morphology, but polynucleosomes are appreciably more compact. Nevertheless, the 3-D zig-zag conformation observed in polynucleosomes at low salt is retained at 40 mM NaCl, and individual nucleosome disks remain separated from each other. There is no evidence for the formation of solenoidal arrangements within polynucleosomes. Comparison of the solution conformation of individual oligonucleosomes with data from physical measurements on bulk chromatin samples suggests that the latter should be reinterpreted. The new data support the concept of an irregular zig-zag chromatin conformation in solution over a range of ionic strengths, in agreement with other in situ (McDowall, A.W., J.M. Smith, and J. Dubochet. 1986, EMBO (Eur. Mol. Biol. Organ.) J.5: 1395-1402; Horowitz, R.A., D.A. Agard, J.W. Sedat, and C.L. Woodcock, 1994. J. Cell Biol. 125:1-10), and in vitro conclusions (van Holde, K., and J. Zlatanova. 1995. J. Biol. Chem. 270:8373-8376). Cryoelectron microscopy also provides a way to determine the 3-D conformation of naturally occurring chromatins in which precise nucleosome positioning plays a role in transcriptional regulation.

DNA-DNA interaction in thin layer analysed by cryo-electron microscopy

Thin vitrified films of 146 base pair double-stranded DNA fragments in semi-dilute solutions are observed by cryo-electron microscopy. Depending on the salt conditions of the buffer and the DNA concentration, the molecules form different characteristic arrangements: they are arranged either locally parallel to each other in random overall orientation or packed perpendicular to the film surface, or else are randomly oriented. These arrangements reflect the weak forces acting between the DNA molecules themselves, as well as between the molecules and the thin film surface.

DNA at the entry-exit of the nucleosome observed by cryoelectron microscopy

Whereas the DNA path inside the nucleosome is well established, it is essentially unknown in the "entry-exit" region, a missing piece in our understanding of the chromatin fiber's folding. The three-dimensional structure of "linker" DNA was investigated here on single nucleosomes reconstituted without H1 on a 256-bp DNA fragment. Cryoelectron microscopy and three-dimensional reconstruction of the DNA axis reveal these nucleosomes as 1.61 +/- 0.15 left-handed superhelical turn particles whose DNA arms bend away from the core particle, thus preventing the occurrence of a crossing in the entry-exit region.

Structure of intracellular mature vaccinia virus observed by cryoelectron microscopy

Intracellular mature vaccinia virus, also called intracellular naked virus, and its core envelope have been observed in their native, unfixed, unstained, hydrated states by cryoelectron microscopy of vitrified samples. The virion appears as a smooth rounded rectangle of ca. 350 by 270 nm. The core seems homogeneous and is surrounded by a 30-nm-thick surface domain delimited by membranes. We show that surface tubules and most likely also the characteristic dumbbell-shaped core with the lateral bodies which are generally observed in negatively stained or conventionally embedded samples are preparation artifacts.

The twist, writhe and overall shape of supercoiled DNA change during counterion-induced transition from a loosely to a tightly interwound superhelix. Possible implications for DNA structure in vivo

A cryo-electron microscopy study of supercoiled DNA molecules freely suspended in cryo-vitrified buffer was combined with Monte Carlo simulations and gel electrophoretic analysis to investigate the role of intersegmental electrostatic repulsion in determining the shape of supercoiled DNA molecules. It is demonstrated here that a decrease of DNA-DNA repulsion by increasing concentrations of counterions causes a higher fraction of the linking number deficit to be partitioned into writhe. When counterions reach concentrations likely to be present under in vivo conditions, naturally supercoiled plasmids adopt a tightly interwound conformation. In these tightly supercoiled DNA molecules the opposing segments of interwound superhelix seem to directly contact each other. This form of supercoiling, where two DNA helices interact laterally, may represent an important functional state of DNA. In the particular case of supercoiled minicircles (178 bp) the delta Lk = -2 topoisomers undergo a sharp structural transition from almost planar circles in low salt buffers to strongly writhed "figure-eight" conformations in buffers containing neutralizing concentrations of counterions. Possible implications of this observed structural transition in DNA are discussed.

Twisting in a crowd

Given our current understanding of nuclear structure, it is difficult to imagine how the nucleus performs its varied functions and controls the traffic of its many components. For example, how can densely packed chromatin be transcribed without the helical nature of the DNA resulting in entangled DNA and RNA and a stalled RNA polymerase? Here, Jacques Dubochet discusses a model of transcription in which DNA rotates around its axis, rather than RNA polymerase rotating around the DNA. Furthermore, he suggests that a view of chromatin as 'liquid' may help in understanding a wide range of nuclear functions.

The apical localization of transcribing RNA polymerases on supercoiled DNA prevents their rotation around the template

The interaction of Escherichia coli RNA polymerase with supercoiled DNA was visualized by cryo-electron microscopy of vitrified samples and by classical electron microscopy methods. We observed that when E. coli RNA polymerase binds to a promoter on supercoiled DNA, this promoter becomes located at an apical loop of the interwound DNA molecule. During transcription RNA polymerase shifts the apical loop along the DNA, always remaining at the top of the moving loop. This relationship between RNA polymerase and the supercoiled template precludes circling of the RNA polymerase around the DNA and prevents the growing RNA transcript from becoming entangled with the template DNA.

Spatial visualization of DNA in solution

An image analysis method is presented which allows for the reconstruction of the three-dimensional path of filamentous objects from two of their projections. Starting with stereo pairs, this method is used to trace the trajectory of DNA molecules embedded in vitreous ice and leads to a faithful representation of their three-dimensional shape in solution. This computer-aided reconstruction is superior to the subjective three-dimensional impression generated by observation of stereo pairs of micrographs because it enables one to look at the reconstructed molecules from any chosen direction and distance and allows quantitative analysis such as determination of distances, curvature, persistence length, and writhe of DNA molecules in solution.

A model for cryosectioning based on the morphology of vitrified ultrathin sections

Electron microscopy of vitrified ultrathin sections allows cell ultrastructure to be studied in the hydrated state. Sectioning of the frozen material is, however, a limiting step, since the cutting forces cause severe mechanical deformation. In order to address this problem, we have investigated the surface of cryosections. It is shown that cryosections have two fundamentally different surfaces. One surface is rough, deformed by cutting-induced deformation lines which are orientated perpendicular to the cutting direction. The other surface, in comparison, is not affected by those deformation lines. Except for knife marks it is smooth. In order to explain the observations, the following model is proposed. The rough relief corresponds to the former block face. Its roughness originates from material that is squeezed out of the section plane when the section is compressed in the cutting direction and bent away from the specimen block. The smooth section surface is the surface in contact with the knife during the sectioning. This contact keeps the surface smooth while imprinting the knife marks.

Direct visualization of supercoiled DNA molecules in solution

The shape of supercoiled DNA molecules in solution is directly visualized by cryo-electron microscopy of vitrified samples. We observe that: (i) supercoiled DNA molecules in solution adopt an interwound rather than a toroidal form, (ii) the diameter of the interwound superhelix changes from about 12 nm to 4 nm upon addition of magnesium salt to the solution and (iii) the partition of the linking deficit between twist and writhe can be quantitatively determined for individual molecules.

Evaporation during preparation of unsupported thin vitrified aqueous layers for cryo-electron microscopy

Evaporation of water cannot be fully avoided when an unsupported thin vitrified film of an aqueous suspension is prepared for cryo-electron microscopy. This results in increasing concentration of solute which could affect the observed material. We have quantitatively studied this effect by measuring the contrast of polystyrene spheres in a metrizamide solution. The drying effect is generally negligible when specimens are prepared on a hydrophilic perforated support but it is frequently important when hydrophobic films are used instead. A flow of humid air, double blotting with minimal exposure of the thin liquid film to the atmosphere, or an automatic plunger optimizing the blotting conditions are simple methods for reducing drying effects. With this third device acting on a hydrophilic supporting film, the increase of solute concentration is limited to less than 20%.

Interactions between actin and myosin filaments in skeletal muscle visualized in frozen-hydrated thin sections

For the purpose of determining net interactions between actin and myosin filaments in muscle cells, perhaps the single most informative view of the myofilament lattice is its averaged axial projection. We have studied frozen-hydrated transverse thin sections with the goal of obtaining axial projections that are not subject to the limitations of conventional thin sectioning (suspect preservation of native structure) or of equatorial x-ray diffraction analysis (lack of experimental phases). In principle, good preservation of native structure may be achieved with fast freezing, followed by low-dose electron imaging of unstained vitrified cryosections. In practice, however, cryosections undergo large-scale distortions, including irreversible compression; furthermore, phase contrast imaging results in a nonlinear relationship between the projected density of the specimen and the optical density of the micrograph. To overcome these limitations, we have devised methods of image restoration and generalized correlation averaging, and applied them to cryosections of rabbit psoas fibers in both the relaxed and rigor states. Thus visualized, myosin filaments appear thicker than actin filaments by a much smaller margin than in conventional thin sections, and particularly so for rigor muscle. This may result from a significant fraction of the myosin S1-cross-bridges averaging out in projection and thus contributing only to the baseline of projected density. Entering rigor incurs a loss of density from an annulus around the myosin filament, with a compensating accumulation of density around the actin filament. This redistribution of mass represents attachment of the fraction of cross-bridges that are visible above background. Myosin filaments in the "nonoverlap" zone appear to broaden on entering rigor, suggesting that on deprivation of ATP, cross-bridges in situ move outwards even without actin in their immediate proximity.

Length and shape variants of the bacteriophage T4 head: mutations in the scaffolding core genes 68 and 22

The shape and size of the bacteriophage T4 head are dependent on genes that determine the scaffolding core and the shell of the prohead. Mutants of the shell proteins affect mainly the head length. Two recently identified genes (genes 67 and 68) and one already known gene (gene 22), whose products are scaffold constituents, have been investigated. Different types of mutants were shown to strongly influence the proportion of aberrantly shaped particles. By model building, these shape variants could be represented as polyhedral bodies derived from icosahedra, through outgrowths along different polyhedral axes. The normal, prolate particle is obtained by elongation along a fivefold axis. The mutations of the three core genes (genes 67, 68, and 22) affect the width mainly by lateral outgrowths of the prolate particle, although small and large isometric particles are also found. Many of the aberrant particles are multitailed, suggesting a correlation between tail attachment sites and shape.

Head structure of bacteriophages T2 and T4

The length-to-width ratios of bacteriophage T2 and T4 heads and stereometric angles specifying the prolate icosahedral T2 capsid were evaluated on electron micrographs recorded from samples prepared by a variety of methods. The copy numbers of the major capsid protein, gp23*, of T2 and T4 phages were compared by quantitative gel electrophoresis. Taken together, the resulting values are most compatible with triangulation numbers T = 13 and Q = 21 for both T2 and T4, thus confirming the previously proposed capsid architecture of T4 revealed by indirect measurements and thereby eliminating the repeatedly reported discrepancy between T2 and T4 in favor of a common Q number of 21 corresponding to 960 copies of gp23*.