Cryo-EM

Cryo-Transmission Electron Microscopy (cryo-EM) is studies the structure of frozen hydrated biological samples by transmission electron microscopy. Generally, several images of the frozen samples are recorded, sometimes at different sample tilt angles, and are computationally combined into one merged data set that can show the high-resolution structure of the sample in three dimensions.

Cryo-EM can be divided into three domains: Electron tomography, single-particle cryo-EM, and electron crystallography. The first studies cellular environments and usually is limited to a resolution of a few nanometers. The second studies individual larger proteins or protein complexes, and is now reaching 4Å resolution in favorable cases (GroEL, large viruses). The third, electron crystallography, studies the structure of membrane proteins that have been reconstituted into lipid membranes and two-dimensionally crystallized. Electron crystallography has been used to determine the structure of 8 membrane proteins at atomic resolution (< 4Å) so far, and has produced 3D membrane protein maps that reached 1.9Å resolution (See the list of available structures from electron crystallography). Our laboratory is employing all three methods. We also develop methods with the aim of being able to rapidly determine the 3D structure of membrane proteins, using primarily electrons as probes.

Membrane Proteins

Membrane proteins constitute over 25% of predicted proteins, and represent the majority of today’s drug targets in pharmaceutical research. Nevertheless, only ~300 structures of membrane proteins have been determined, compared to more than 37,000 available structures of soluble proteins.

One of the possible applications of cryo-EM is the study of membrane protein structures. Membrane proteins are of central importance for health and disease. While the high-resolution structure of many membrane proteins can efficiently be determined by XRD or NMR, the structure and arrangement of larger membrane protein complexes or the dynamic conformation of certain membrane protein systems in the biological membrane are best studied by analyzing them in the lipid membrane environment.

Membrane protein structure research strongly benefits from pooling resources for expression, purification and structural analysis. To this end, we collaborate with the Membrane Protein Expression Center (MPEC) and are part of the Center for Structures of Membrane Proteins (CSMP). While MPEC expresses and purifies membrane proteins, CSMP focuses on the structure determination of prokaryotic membrane proteins. Both centers are headed by Robert Stroud, UCSF. These centers combine the methods of structure determination by X-ray, NMR, Mass-Spec, cryo-EM and Bioinformatics.

With Jon Scholey, UCD, we have analyzed the kinesin-like motor protein-dependent growth of sensory cilia in C. elegans neurons, using high-pressure freezing, freeze-substitution, serial sectioning and 3D reconstruction from serial section alignment and electron tomography 3D reconstructions (Evans et al., JCB 2006; Tao et al., CB, 2006). With Diana Myles, UCD, we have used electron tomography to analyze the role of the membrane protein CD9 in microvillar membrane function and shape in the context of oocyte fertility (Runge et al., Dev. Biol., 2007). With Crina Nimigean, Cornell University, NY, we have characterized cyclic-nucleotide gated potassium channels by single particle EM and electron crystallography and obtained a first direct characterization of the orientation of the voltage sensor “paddles” in the membrane embedded conformation (Chiu et al., Structure 2007). In collaboration with Senyon Choe, Salk Institute, San Diego, we also study a number of other membrane protein systems, some of which are cell-free expressed. While cell-free expression is a promising and productive method that is affordable when the ingredients are produced locally, the resulting membrane protein was never exposed to lipids, so that the functional and structural analysis when reconstituted into lipid membranes is of interest in addition to XRD or NMR characterizations. In collaboration with Joe Mindell, NIH, Washington, we study the E. coli chloride/proton antiporter.

Membrane Protein Complexes

We also studies protein complexes, partly in collaboration with other laboratories. Lenin Dominguez in our group produces preparations from inner mitochondrial membranes that are enriched in the putatively dimeric F-ATPase-Synthasome complexes and are now beginning to study these by electron tomography and single-particle 3D volume averaging, in order to elucidate the putative higher-order organization of these enzymes.

Complexes of Cytosolic Proteins

In collaboration with Wolf Heyer and Steve Kowalczykowski, UCD, we also study DNA binding proteins responsible for chromosome maintenance and repair. Other projects in collaboration with Scholey and Kent McDonald concern molecular motors, and in collaboration with Roland Riek, ETHZ, Switzerland, we study prion proteins. We also work on the characterization of the interaction of nanoparticles with biological tissue (with Kent Pinkerton, UCD). In collaboration with Paul Fitzgerald (UC Davis), we study intermediate filaments. These projects mainly employ single particle cryo-EM and electron tomography.

Method Development

Sample Preparation: We work on a new computerized membrane protein crystallization device, which will feature on-line monitoring of the sample state, and use this in-situ knowledge to control the crystallization process (temperature, sample concentration, detergent dialysis speed). While the Engel lab has pioneered crystallization attempts by diluting the detergent, our plan is to provoke crystal formation by concentrating the prep while removing detergent. Our preliminary results indicate that controlled concentration is a strong factor in the production of high-quality crystals.

Aberration-corrected Imaging: In collaboration with Nigel Browning, UCD and LLNL, and Angus Kirkland, Oxford, UK, we evaluated the usability of aberration corrected TEM imaging for biological samples, a work that resulted in direct images of Si [110] at 1.1 Å, and of Paraffin at 1.6 Å, all under cryo-low dose conditions (Evans et al., Ultramic., in press). While aberration corrected TEM imaging is not interesting in the general case for biological samples, our work paves the path to aberration corrected imaging TEM in combination with a phase plate for optimal contrast of biological specimens.

Phase contrast STEM: One of the main bottlenecks in biological electron microscopy is caused by the beam-induced resolution loss when imaging tilted cryo-EM samples. This affects all types of cryo-EM that have to tilt the sample. In collaboration with Browning, we now have collected data that prove that STEM imaging of tilted frozen hydrated cryo-EM samples is not affected from that resolution loss. The problem with conventional STEM, however, is the absence of phase contrast and a high electron dose or dose-rate. We have developed the theory and computer simulations, and are in the process of implementing high-resolution low-dose phase-contrast STEM of tilted biological cryo-EM samples (James Buban). This development involves aberration corrected illumination, multi-channel multi-ring bright-field and dark-field detectors, while ultrafast beam and instrument control for dynamic focusing on tilted sample planes allow recording of the data. We have data that show that the simultaneous collection of the BF and HAADF signals from an aberration corrected STEM allows access to the phase-contrast signal at good signal-to-noise ratio per beam-damage.

Computer Image Processing: We have developed a software package for the computer image processing of 2D crystal images of membrane proteins (2dx.org, Bryant Gipson). Our software 2dx is originally based on the MRC software (with agreement of Richard Henderson, the main creator of that software), and features a user-friendly interface, optionally fully automatic image processing, merging and 3D structure reconstruction. Our software system is used in the life sciences and also in the materials sciences fields, who now also starts using low-dose data-collection schemes, and requires signal averaging. In order to enable the software system to deal with badly-ordered 2D crystals and eventually with loosely reconstituted or native membranes, we have in collaboration with Niko Grigorieff, Brandeis University, MA, developed a 2D maximum-likelihood-based software module that runs under the 2dx environment (Zeng et al., JSB 2007). We are now extending this for 3D capabilities, where we include a non-linear measure to further eliminate reference-bias, and we correlate the local distribution of alignment parameters with those of neighboring particles from the membrane, to further reduce the effect of reference bias (“The tilt geometry will not change abruptly, but smoothly over the sample”). We also work on extending the 2dx software package with a Fiehnup-Gerchberg-Saxton HIO algorithm for phase extension and diffractive imaging (collaboration with John Spence, Arizona).

Workshop on Electron Crystallography

Together with Ben Hankamer, University of Queensland, Austroalia, and Tom Walz, Harvard Med. School and HHMI, Boston, MA, USA, we organize a biennial Workshop on Electron Crystallography of Membrane Proteins. The one-week workshop is aimed at PhD students, PostDocs, and beyond, who are interested in the structure determination of membrane proteins (or other 2D protein crystal arrays) by electron crystallography. The workshops feature lectures in the morning, practicals in the afternoon, student poster session before dinner, and science talks in the evening. Practicals concern membrane protein crystallization by different methods, sample preparation for cryo-EM, data collection by cryo-EM imaging and electron diffraction, and computer image processing. The workshop features ~20 lecturers and is limited to ~20 students. A first workshop took place at UC Davis in Aug. 6-11, 2006. The next workshop is scheduled for Sept. 7-13, 2008.

Networks

See: CSMP

and: MPEC

Funding

The Stahlberg laboratory is supported by grants from the National Science Foundation (NSF BIO), the National Institutes of Health (NIGMS), the California Breast Cancer Research Program, and the UC Cancer Research Coordinating Committee.