Search

Electron microscopy of biological structural research has made significant advances. These advances were the stimulus to develop an electron microscope containing most of the desirable features and methods for high resolution cryo-electron microscopy. The new instrument is a modified commercial Philips CM 20 FEG electron microscope equipped with a helium-cooled cryoobjective from Siemens. This paper presents the essential features and research possibilities of this new microscope named “SOPHIE” (acronym for S uperconducting O bjective in a PHI -lips E lectron microscope).

Using cryo-microscopy at liquid-helium temperature we have determined the projection structure of the surface layer of the archaebacterium Sulfolobus shibatae embedded in aurothioglucose to a resolution of 0.85 nm. Image analysis was performed in real space and incorporated a classification of unit cells, thus taking into account the frequent occurrence of twin boundaries. In an attempt to identify resolution-limiting factors we have performed a distortion analysis quantifying the main distortion parameters.

Based on experience gained with the helium-cooled cryomicroscope “Suleika” [I. Dietrich et al., Ultramicroscopy 2 (1977) 241], the requirements for a microscope that also satisfies coherency conditions are developed and discussed.

Cryoelectron microscopy is now proven to be an advantageous tool for structure research of protein crystals. Using a helium-cooled superconducting electron microscope, the following protein crystals have been imaged with high resolution in one projection: crotoxin complex 3.5 Â, purple membrane 2.8 Â, matrix porin OmpF 3.5 Â, surface protein of sulfolobus spec. B12 10 Â. Fine-structure details within the unit cell were resolved, although even at 4.5 K specimen temperature the crystal still suffers considerable radiation damage. While the tolerable dose is only 20 e/Â2, for atomic resolution a dose at least 100 times higher is needed. The ladder to these successes had two steps: First, the image was taken in a special way with a minimum of pre-exposure. Second, the low-dose images were evaluated by a sophisticated image-processing procedure, the main part of which is the cross-correlation averaging. The high resolution structure is in all cases revealed by averaging thousands of low-dose-imaged unit cells.

Electron cryo-microscopy of individual non-crystallized macromolecules (“single particles”) is a very rapid technique for probing the three-dimensional (“3D”) structure of biological macromolecules. By exploiting the different orientations of the macromolecules in the embedding medium, one may extract 3D information from the data without ever collecting tilt series in the microscope. The angular reconstitution approach, designed for this purpose, was recently extended with a number of refinements which allow its use as a routine technique for finding 3D structures of macromolecules with arbitrary pointgroup symmetry, from entirely asymmetric ribosomes to viruses with icosahedral symmetry.The specimen preparation technique associated with the angular reconstitution approach is simple and fast since crystallization experiments are avoided altogether. The vitreous-ice embedding specimen preparation technique remains one of our favorite specimen preparation techniques. We are, however, currently experimenting with specimens embedded in glucose and ammonium molybdate or other heavy-metal salts. Collecting good micrographs can also be quite straightforward since the images are taken from untilted specimens, while exposing each image area only once.

High-resolution cryo-electron microscopy yielded remarkable results in biological molecular structure research. These successes encouraged to proceed in developing even better instruments and procedures. In our laboratory we built the cryo-electron microscope "SOPHIE", a modified Philips CM 20 FEG. SOPHIE stands as an acronym for Superconducting Objective in a PHIlips Electronmicroscope. The main features of this microscope are the following: 1.200kV acceleration voltage2.Schottky field emitter3.Helium cooled superconducting objective The Helium-cooled cryo-objective from the former microscope SULEIKA was adapted into the column of the original Philips CM20 FEG [Fig. 1], This new SOPHIE microscope, thus, combines two important features: a modern computer controlled electron microscope with a high coherent beam source and the proven advantages of the superconducting cryo-objective. The advantage of the field emission gun at 200 kV over a thermionic cathode at 100 kV can be best visualized while comparing the respective phase-contrast-transfer functions at Scherzer focus [Fig.2]. Moreover, SOPHIE is equipped with the small spot scanning procedure.

The task in investigating the structure of crystals is to measure the Fourier coefficients in amplitude and phase. Electron crystallography, i.e. electron diffraction combined with electron-microscopical imaging, is a straigthforward method, because there is a priori no “phase problem”. But radiation damage during electron exposure was for a long time an absolutely unbridgeable chasm on the way to high-resolution structure research of two-dimensional protein crystals by electron microscopy. Now, using cryo-electron microscopy and computer image processing, it has been proven possible to overcome this difficulty and achieve high resolution.Here we present a rough description of the recording procedure. We used a helium-cooled superconducting objective lens with 4.5 K specimen temperature, but the procedure should also be applicable at higher specimen temperatures. The radiation damage is reduced at low specimen temperature, but unfortunately there still remains a remarkable deterioration of the crystal with increasing dose, which means that minimum exposure is advisable.

The thermoacidophilic archaebacterium Sulfolobus spec. B12 , which is closely related to Sulfolobus solfataricus , possesses a regularly arrayed surface protein (S-layer), which is linked to the plasma membrane via spacer elements spanning a distinct interspace of approximately 18 nm. The S-layer has p3-Symmetry and a lattice constant of 21 nm; three-dimensional reconstructions of negatively stained fragments yield a layer thickness of approximately 6-7 nm.For analysing the molecular architecture of Sulfolobus surface protein in greater detail we use aurothioglucose(ATG)-embedding for specimen preparation. Like glucose, ATG, is supposed to mimic the effect of water, but has the advantage of being less volatile. ATG has advantages over glucose when working with specimens composed exclusively of protein because of its higher density of 2.92 g cm-3. Because of its high radiation sensitivity electromicrographs has to be recorded under strict low-dose conditions. We have recorded electromicrographs with a liquid helium-cooled superconducting electron microscope (the socalled SULEIKA at the Fritz-Haber-lnstitut) with a specimen temperature of 4.5 K and with a maximum dose of 2000 e nm-2 avoiding any pre-irradiation of the specimen.

The light driven proton pump bacteriorhodopsin (bR) occurs naturally as two-dimensional crystals. A three-dimensional density map of the structure, at near atomic resolution, has been obtained by studying the crystals using electron cryo-microscopy to obtain diffraction patterns and high resolution micrographs (1).New methods have been developed for analysing micrographs from tilted specimens, incorporating the methods previously developed for untilted specimens that enable large areas to be analysed and corrected for distortions. Data from 72 images, from both tilted and untilted specimens, have been analysed to produce the phases of 2700 independent Fourier components of the structure. The amplitudes of these components have been accurately measured from 150 diffraction patterns. Together, these data represent about half of the full three-dimensional transform to 3.5 Å. The distribution of the data which is included in the map is shown in fig. 1. For specimen tilts up to around 20° the data is essentially complete. For higher tilts the data is more sparsely sampled, and is at present about half complete.

Many of the most interesting specimens are so severely damaged during the electron microscopical imaging that no useful image emerges from the electron noise. One method of reducing this radiation damage is to cool the specimen, and if this helps it seems reasonable to cool it down as far as possible. This goal is reached most straightforwardly by using the Siemens helium-cooled superconducting objective lens designed by I. Dietrich and co-workers.

The advantage of MgO platelet crystals as specimen supports is shown. The contrast transfer of carbon foil granulation through the thin MgO platelet crystals was measured. The reduction of the contrast due to the crystal of 40 nm thickness was about 50%. Gold particles and ferritin particles were used as test specimens on MgO platelet crystals. The lattice images of both specimens were observed through the crystals. The protein shell of ferritin was recognizable. Some different problems arising with this specimen support are discussed.

Calculated defect diffuse scattering patterns, using the kink defect model employed earlier to explain changes in Bragg electron diffraction intensities for heated n-paraffins, are shown to account best for observed diffuse data from solution-grown and epitaxially grown micro-crystals. The presence of significant diffuse scattering below 15°K, moreover, indicates that such chain kinks are stable in the crystal lattice.

Cross-correlation of micrographs taken under identical optical conditions allows measurement of (a) the optical transfer function free from electron and photographic noise and (b) object change, especially that due to radiation damage. Light optical methods based on superposition diffractograms are discussed. A subtraction procedure is described for measuring the optical transfer function, and a two-step diffraction process is presented for measuring object change. As an example, experiments with radiation damage in carbon foil are described. Here increased stability regarding radiation-induced change was found with increasing dose, making possible measurement of deviations from the “zero dose image”.

There is an urgent interest in analyzing the structure of biomolecules by EM imaging. In order to reduce the quantum noise under the level of the inherent contrast C of the structure imaged, a minimum dose Dm=4/C2r2is required when a structure detail of size r is to be detected. (The factor 4 is just the square of the signal-to-noise ratio, justified by experiments.) Setting r to about .4 nm and C to about 10% the required minimum dose amounts to 2000 e/nm2. Unfortunately at room temperature this minimum dose destroys most biomolecules. One way out of this dilemma seems to be cooling the specimen to He temperature.Experiments in an improved superconducting electron microscope (Siemens system 2) have been performed with the aim of confirming and/or refuting measurements by others in this very well defined experimental situation. The fading of diffraction patterns was used to characterize the radiation damage (Fig. 1), yet the implication for the image formation was introduced by the size of the pixel.