Laboratory cryo soft X-ray microscopy H.M. Hertz a,⇑ , O. von Hofsten a , M. Bertilson a , U. Vogt a , A. Holmberg a , J. Reinspach a , D. Martz a , M. Selin a , A.E. Christakou a , J. Jerlström-Hultqvist b , S. Svärd b a Biomedical and X-Ray Physics, Dept. of Applied Physics, KTH Royal Inst. of Technology/Albanova, 10691 Stockholm, Sweden b Dept. of Cell and Molec. Biol., Uppsala University, 75105 Uppsala, Sweden article info Article history: Available online 20 November 2011 Keywords: X-ray microscopy Cryo fixation Laboratory Parasites abstract Lens-based water-window X-ray microscopy allows two- and three-dimensional (2D and 3D) imaging of intact unstained cells in their near-native state with unprecedented contrast and resolution. Cryofixation is essential to avoid radiation damage to the sample. Present cryo X-ray microscopes rely on synchrotron radiation sources, thereby limiting the accessibility for a wider community of biologists. In the present paper we demonstrate water-window cryo X-ray microscopy with a laboratory-source-based arrangement. The microscope relies on a k = 2.48-nm liquid-jet high-brightness laser-plasma source, nor- mal-incidence multilayer condenser optics, 30-nm zone-plate optics, and a cryo sample chamber. We demonstrate 2D imaging of test patterns, and intact unstained yeast, protozoan parasites and mammalian cells. Overview 3D information is obtained by stereo imaging while complete 3D microscopy is provided by full tomographic reconstruction. The laboratory microscope image quality approaches that of the syn- chrotron microscopes, but with longer exposure times. The experimental image quality is analyzed from a numerical wave-propagation model of the imaging system and a path to reach synchrotron-like expo- sure times in laboratory microscopy is outlined. Ó 2011 Elsevier Inc. All rights reserved. 1. Introduction Determination of the structure of intact unstained cells with high spatial resolution is important for cell-biology. X-rays have the necessary absorption and scattering properties for two- and three-dimensional imaging of such thick (5–10 lm) objects. The two major X-ray imaging methods are lens-based soft X-ray microscopy (XRM) (Sakdinawat and Attwood, 2010) and lens-less hard X-ray coherent diffraction imaging (CDI) (Chapman and Nugent, 2010). Both require significant X-ray doses to provide the necessary signal-to-noise ratio for high-resolution imaging, making sample preparation and sample damage important issues. XRM (Uchida et al., 2009; Schneider et al., 2010) has shown 50–70 nm isotropic resolution in whole intact unstained cryo-fixed cells. The emerging hard X-ray CDI technique (Giewekemeyer et al., 2010; Jiang et al., 2010; Nelson et al., 2010) shows potential for sim- ilar detail but is still typically applied to freeze dried or fixed cells. Both methods can obtain higher resolution for smaller samples or smaller regions of interests within samples, albeit at the price of higher dose. Unfortunately, both methods presently rely on large-scale accelerator-based X-ray facilities, synchrotrons or free-electron lasers, thereby limiting the accessibility for a wider community of biologists. Present non-X-ray high-resolution laboratory-scale methods, i.e., electron (Medalia et al., 2002) and super-resolution optical (Hell, 2009) microscopy, provide excellent resolution (tens of nanometers and below) but are essentially lim- ited to thin objects due to electron scattering and long exposure times, respectively, and are thereby presently not applicable to in- tact cells. Thus, a laboratory-scale method that allows imaging of intact cells in their native or near-native hydrated state with a res- olution of tens of nanometers has potential to fill an important void. X-ray microscopy (XRM) in the water window (k = 2.3–4.4 nm; E = 284–540 eV) has demonstrated high-resolution imaging of in- tact cells (Sakdinawat and Attwood, 2010; Kirz et al., 1995). Opti- cally, the resolution in two-dimensional (2D) imaging is presently determined by the zone plate fabrication rather than the wave- length to typically 10–20 nm (Chao et al., 2005; Vila-Comamala et al., 2009). The contrast relies on the differential absorption be- tween carbon (protein, lipids, etc.) and water, i.e., no staining is necessary. Detection of a statistically significant feature requires a certain signal-to-noise (cf. Rose criterion (Bushberg et al., 2002)) and, thus, a certain number of photons. Typically a water- window XRM image requires 10 6 Gy for absorption imaging of 50 nm protein object in 10 lm of water under realistic assump- tions (Kirz et al., 1995; Schneider, 1998). This can be lowered 0.5–1 order of magnitude by phase imaging. Hydrated biological samples show structural changes already at 10 4 Gy (Schneider, 1998). Chemical fixation allows up 10 6 Gy while cryo fixation has been show to provide stable samples for 10 10 Gy (Schneider 1047-8477/$ - see front matter Ó 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.jsb.2011.11.015 ⇑ Corresponding author. Fax: +46 8 55378466. E-mail address: hertz@biox.kth.se (H.M. Hertz). Journal of Structural Biology 177 (2012) 267–272 Contents lists available at SciVerse ScienceDirect Journal of Structural Biology journal homepage: www.elsevier.com/locate/yjsbi