Cryo Electron Tomography: Unique Capability for Structural Biology InvestigationsCryo electron tomography's (CET) ability to visualize three dimensional biological structures - ranging in size from molecular to cellular - fills a critical gap between techniques with atomic resolution, such as x-ray diffraction (XRD) and nuclear magnetic resonance (NMR), and conventional light microscopy. However, it is CET's ability to investigate biological structures in their unperturbed, native context that makes it an indispensable tool in the currently exploding field of structural biology. Introduction The recent decoding of the human genome, and the realization that we can now read the entire blueprint of the vastly complex biological machine that is a living organism-human or any other, has had a tremendous symbolic and practical impact on the public consciousness in general, and the scientific community in particular. Since that time, biologists have focused increasing attention on understanding the structure and function of the machine's components - the proteins encoded by the genetic blueprint.

The Strength of CET in Structural Biology

One approach, call it brute force, in many ways analogous to the approach taken in the decoding effort itself, simply seeks to determine the structure of each of the large but finite number of individual proteins that constitute the entire proteome, perhaps one to two million in a human. Though this basic structural information is certainly fundamental to a complete understanding, it is only the lowest level in the hierarchy of structural complexity. Few proteins work alone. Most function in concert with others as components of supramolecular complexes that may be persistent or quite transient. At a still higher level, the cellular context in which these complexes function cannot be regarded as a simple sack of cytoplasm filled with randomly colliding complexes. It is itself a highly structured environment throughout which components are manufactured, regulated, transported and consumed among at myriad functional localities. Clearly, the operation of the machine can be more easily understood by observations of the assembled components.

Compare it to building a watch. We have the parts list, the genetic code. With sufficient effort we may eventually describe each component in atomic detail. But how much easier is it to comprehend the workings of the entire machine if we can look at a fully assembled watch. This is the strength of CET in structural biology. Alone among experimental techniques, it permits observations of fully assembled biological machines from the macromolecular cogs and pulleys to the cellular systems for synthesis and distribution (as shown in figures 1 - 3).

Vitrification Essential for CET

As its name implies, cryo electron tomography uses electron images of frozen samples to construct a three dimensional model of the imaged volume. The images, provided by a high resolution transmission electron microscope (TEM), are two dimensional projections acquired as the sample is rotated incrementally about an axis perpendicular to the viewing direction. A computer creates the 3D model from the projections with reconstruction techniques familiar to most of us from their use in medical imaging applications such as CAT scans and MRI. Key to the value of CET is its use of a cryogenic freezing process known as vitrification. The process freezes the hydrated sample so rapidly that water molecules do not crystallize, instead forming an amorphous solid (vitreous ice) that does little or no damage to delicate molecular structures. CET can resolve structures down to a few nanometers, sufficient for tertiary and quaternary protein morphology. In hybrid approaches to structural analysis, atomic scale structures determined by XRD and NMR are fitted to larger scale CET models to enhance the level of structural detail. CET samples may be pristine biological material, or selectively stained with electron dense materials to emphasize particular features. More specific emphasis to particular proteins or domains can be achieved with immuno labeling techniques using gold or other high contrast nanoparticles.

Investigation of Morphological Transformations

The ability of XRD and NMR to achieve atomic resolution is due in large part to its use of a composite signal acquired from a collection of presumably identical structures. In contrast, CET examines an individual structure. Since the sample is quite literally frozen in time, CET can distinguish morphological differences among structures with identical atomic compositions, such as protein molecules, with identical peptide sequences. By examining a collection of frozen molecular instances, CET can be used to investigate morphological transformations that occur as a result of interactions, the dynamic behavior of flexible proteins, the formation of transient complexes in signaling pathways, and more. CET eliminates XRD's requirement for crystalline samples, avoiding the possibility of structural distortions induced by crystallization, and allowing the examination of virtually any specimen, including flexible and insoluble proteins that are notoriously difficult to crystallize. It also does not suffer the increasing difficulty of NMR analysis with larger molecular size. CET does face at least two significant challenges. Because the signal originates from a single molecule it is weak relative to random variations in its intensity (shot noise). The delicate nature of molecular structures and the high energy of the interrogating radiation do not allow increases in beam intensity or exposure time without inflicting significant damage in the sample. Careful dose management and automated acquisition procedures can help, but this remains a fundamental limitation.

Correlative Microscopy Increasingly Important

The second issue relates to the crowded environment of the cell. We would like to look at distinct structures clearly contrasted against a featureless background. However, in an intact biological system there is no empty space and little fundamental difference between the atoms that compose the structures of interest and those that surround it. It is rather like looking into a very big bag of marbles, or perhaps like looking for specific configurations of bubbles in a bucket of foam. There are techniques that reduce the difficulty. We have mentioned staining and immuno labeling. Another area of growing interest is correlative microscopy, using an optical technique such as fluorescence microscopy to locate labeled structures for subsequent high resolution CET analysis. Finally, the digital nature of tomographic data lends itself well to computational search and fit routines using previously defined templates to locate specific structures within the sample volume. TEM technology has advanced dramatically in recent years. Perhaps the greatest development has been the incorporation of aberration correctors, which has pushed image resolution well below the Angstrom barrier - 0.5 Angstrom resolution was just announced by the TEAM (Transmission Electron Aberration- corrected Microscope) project, a joint effort by the U.S. Department of Energy, FEI Company and CEOS GmbBH.

Improvements in TEM

A significant contribution to improved performance has also come from better instrument stability - mechanical, electronic and environmental. Sophisticated automation at all levels from setup and alignment, through operation, data acquisition, and sample preparation has made TEM much faster, easier, more repeatable and more reliable. Some very interesting progress has been made in sample preparation. FEI's Vitrobot provides automatic vitrification fluid suspensions on a TEM sample grid. Focused ion beam (FIB) based preparation procedures have greatly improved the ease, speed, and reliability of preparing thin samples from bulk specimens, including frozen material. Another area of interest is the FIB based preparation of cylindrical tomography samples that permit the acquisition of 2D images through a complete, 360 degree range of rotation. The sheer volume of information potentially available in a high resolution tomogram is mind boggling. Theoretically, one could capture the entire proteome of a cell, including all of its functional complexes and higher level structures. As one researcher put it, one good tomogram can make a whole career - or perhaps several. With this kind of potential, there is little doubt that CET will play a critical role in the future of structural biology.