Journal of Structural Biology 152 (2005) 149–156 www.elsevier.com/locate/yjsbi 1047-8477/$ - see front matter  2005 Elsevier Inc. All rights reserved. doi:10.1016/j.jsb.2005.08.004 Dose tolerance at helium and nitrogen temperatures for whole cell electron tomography Luis R. Comolli, Kenneth H. Downing ¤ Life Science Division, Lawrence Berkeley National Laboratory, USA Received 19 April 2005; received in revised form 28 July 2005; accepted 17 August 2005 Available online 13 September 2005 Abstract Electron tomography is currently the only method that allows the direct three-dimensional visualization of macromolecules in an unperturbed cellular context. In principle, tomography should enable the identiWcation and localization of the major macromolecular complexes within intact bacteria, embedded in amorphous ice. In an eVort to optimize conditions for recording data that would bring us close to the theoretical limits, we present here a comparison of the dose tolerance of Caulobacter crescentus cells embedded in amorphous ice at liquid helium versus liquid nitrogen temperature. The inner and outer cell membranes, and the periodic structure of the S-layer of this Gram-negative bacterium provide ideal features to monitor changes in contrast and order as a function of dose. The loss of order in the S-layer occurs at comparable doses at helium and nitrogen temperatures. Macroscopic bubbling within the cell and the plastic support develops at both temperatures, but more slowly at helium temperature. The texture of the bubbles is Wner in initial stages at helium temperature, giving an impression of contrast reversal in some parts of the specimen. Bubbles evolve diVerently in diVerent organelles, presumably a consequence of their diVerent chemical composition and mechanical properties. Finally, the amorphous ice “Xows” at helium temperature, causing changes in the relative positions of markers within the specimen and distorting the cells. We conclude that for cryo-electron tomography of whole cells liquid nitrogen temperature provides better overall data quality.  2005 Elsevier Inc. All rights reserved. Keywords: Electron tomography; Cryo-microscopy; Radiation damage 1. Introduction Cryo-electron tomography is currently the only tech- nique that allows direct visualization of intact, fully hydrated cells and subcellular structures at a resolutions suYcient to identify individual macromolecular complexes (Baumeister, 2002; Koster and Klumperman, 2003; Meda- lia et al., 2002; Steven and Aebi, 2003). The technique has the potential to reach a resolution of 2–3 nm, which would allow mapping the major macromolecular complexes within whole cells. Tomography has already been used to visualize protein distribution and cytoskeletal structures in a few reports (Kurner et al., 2005; Medalia et al., 2002), but it still needs development to reach its full potential. One of the main limits to resolution is the low signal-to-noise ratio (SNR) that results from rapid damage to biological mate- rial by the electron beam. Increasing the exposure that the specimen can tolerate before being severely damaged by the beam would be one of the simplest ways to improve the SNR. It has long been recognized that cooling the specimen to liquid nitrogen temperature provides a signiWcant improve- ment in the resistance to the eVects of radiation damage (Hayward and Glaeser, 1979; International Experimental Study Group, 1986). The most common measure for speci- men damage has been the fading of diVraction spots from crystalline specimens. Nitrogen cooling allows about a fac- tor of 5–10 increase over room temperature in the exposure before a given fractional loss of diVraction intensity occurs. There has been a hope, and some evidence, that further cooling to liquid helium temperature would yield a * Corresponding author. Fax: +1 510 486 6488. E-mail address: khdowning@lbl.gov (K.H. Downing).