Exploring the Inner Space of Cells by Cryoelectron-Tomography W. Baumeister , M. Beck, A. Briegel, M. Cyrklaff, Th. Keil, J. Kürner, V. Lucic, O. Medalia, St. Nickell, and J. Plitzko Max-Planck-Institute of Biochemistry, Am Klopferspitz 18, D-82152 Martinsried, Germany Electron Tomography (ET) is uniquely suited to obtain three-dimensional (3D) images of large pleiomorphic structures, such as supramolecular assemblies, organelles, or even whole cells. While the principles of ET have been known for decades, its use has gathered momentum only in recent years. Technological advances (namely computer controlled transmission electron microscopes and large area CCD cameras) have made it possible to develop automated data acquisition procedures. This, in turn, allowed to reduce the total electron dose to levels low enough for studying radiation sensitive biological materials embedded in vitreous ice. As a result, we are now poised to combine the power of high-resolution 3-D imaging with the best possible preservation of the specimen [1,2]. In the past, ET has mainly been used to examine thin sections of plastic-embedded materials. This approach has provided valuable insights into cellular architecture, but it falls short of revealing the macromolecular organization inside cells. Chemical fixation, staining with contrast enhancing heavy atom compounds, dehydration and plastic embedding affect the specimen significantly and make the interpretation of such tomograms at the molecular level very problematic, if not impossible. The use of high-pressure freezing instead of chemical fixation improves specimen preservation significantly but it does not eliminate the problems arising from the intricate interactions between heavy atom stains and molecular structures. Obviously, cryo-sectioning of vitrified material is the method of choice for large (> 1µm) objects (cells, tissues), but it remains a challenging task despite recent progress [3]. It is often possible to isolate macromolecular complexes, virus particles or organelles using procedures that maintain their structural integrity. Such nanoscale specimens are suitable for direct analysis by ET [4,5]. The resolution obtained allows the docking of high resolution component structures obtained by X-ray crystallography or NMR. Hybrid approaches of this kind can be used to generate pseudoatomic maps of assemblies that are too large or variable for direct high-resolution structural studies. ET of frozen-hydrated whole prokaryotic cells or thin eukaryotic cells grown directly on EM girds provides 3-D images of macromolecular structures unperturbed and in their functional environment [6,7]. Currently resolution is limited to 4-6 nm but with instrumental advances, such as liquid He cooling and CCD cameras optimized for intermediate voltage TEMs, we are now entering the realm of molecular resolution (2-4 nm) [8]. High resolution tomograms of organelles or cells contain vast amount of information; essentially they are 3-D images of the cell’s entire proteome and they should ultimately enable us to map the spatial relationships of macromolecules in a cellular context, the ‘interactome’. However, it is no trivial task to retrieve this information because of the poor signal-to-noise ratio of such tomograms and the crowded nature of the cytoplasm and many organelles. Denoising procedures can help to combat noise and to facilitate visualization, but advanced pattern recognition methods are needed for detecting and identifying with high fidelity specific macromolecules based on their structural signature (size and shape, for example). Provided that high- or medium-resolution structures of the 152 Microsc Microanal 10(Suppl 2), 2004 DOI: 10.1017/S1431927604880218 Copyright 2004 Microscopy Society of America