Electron Cryomicroscopy of Biological Assembly
Wah Chiu
National Center for Macromolecular Imaging, Baylor College of Medicine, Houston, TX 77030 (wah@bcm.tmc.edu)
A typical biological process involves tens to hundreds of proteins to act in coordinated and concerted manner. Understanding their mechanism of actions requires the 3-dimensional photos of the complex at different physiological stages. This spatial and temporal information is valuable for designing new therapeutic strategy to prevent or cure diseases related to the assemblies. Typically, a biological assembly has a molecular size close to or over one million Daltons. A hybrid of technologies in structural and digital biology is being used to determine the structures of biological assemblies. In particular, electron cryomicroscopy (cryo-EM) is playing an ever increasingly critical role in such structural determinations.
The structural resolution of cryo-EM of biological assemblies has been improved through the years mostly due to the innovations in image reconstruction methods. For instance, single particle electron cryomicroscopy has been shown to be capable of resolving the structure of GroEL at 6 Å resolution (figure 1), where alpha helices and beta sheets of the individual subunits are unambiguously recognizable (figure 2). The cryo-EM structure motif also matches that seen in the crystal structure though there are domain movements in the intermediate and apical domains of the subunit between the two structures (1). However, when we attempt to visualize the reaction product of GroEL together with the GroES and the protein substrate, the situation becomes more complex because we are identifying not only the structure of a single reaction product but also a mixture of structures of the reactants/products as well as multiple conformers of each of the molecules. These have presented a great research challenge in sorting out the images of structurally heterogeneous particles before a higher resolution reconstruction can be emerged.
|
|
Figure 1. 6 Å cryo-EM map of GroEL reconstructed from ~40000 single particle images (1). |
Figure 2. Single subunit of GroEL superimposed between cryoEM map (rendering) and crystal structure (ribbon). |
Another type of biological assembly which is amenable to cryo-EM analysis is the crystalline organelle such as acrosomal bundle purified from Limulus sperm. This bundle has been regarded as a prototypic biological spring with an unusual structure containing ~100 filaments of actin with scruin cross-linkers, having an axial ratio of 600:1. The stoichiometry of actin and scruin is 1:1. Before activation, the bundle is bent, twisted and coiled. It is a novel actin-based machine that extends itself 60 µm in five seconds without the action of myosin motor proteins or actin polymerization (2). Energy for extension is stored in the initial coiled state. On average the filaments in the coil are overtwisted by 0.23° per subunit. Untwisting the filaments is coupled with the uncoiling of the bundle into the final extended state (3).
The extended acrosomal bundle can be readily purified and diffracts to 7 Å resolution in the electron cryomicroscope. Because of a small twist along the bundle, we can only treat a small segment of the bundle as a single crystal. A unit cell (146x146x765Å, g = 120˚ in a pseudo-hexagonal lattice) contains 4.7 MDa. We have developed a set of novel computational processing procedures to index the diffraction patterns computed from the images (4) and to perform a series of non-crystallographic averaging of the subunits within the asymmetric unit of the crystal in order to enhance its signal to noise ratio. A 9.5 Å density map was computed with about 16,000 unique reflections, merged from 500,000 individual reflections from 1000 segments in 153 different bundle images on about 90 different micrographs (5).
This is the highest resolution of any cryo-EM reconstruction containing an F-actin filament. We can identify each actin subunit and both domains (E and S) of scruin, and secondary structure elements of both molecules. The 9.5 Å cryo-EM map of the acrosomal bundle shows that actin and scruin are distorted from a standard F-actin in an unique and unpredictable way 5. This structural organization may result from the competing requirements to pack scruin cross-linkers in the bundle and to maintain actin filament integrity. This variability in structural organization allows filaments to pack into a highly ordered and rigid bundle in the extended state and suggests a mechanism for storing and releasing energy between coiled and extended states without disassembly.
Acknowledgement: This research has been supported by NIH grants (P41RR02250 and P01GM064692). The GroEL work is carried out by Steven Ludtke and Donghua Chen in collaboration with David Chuang and JL Song at UTSW Medical School; the acrosomal bundle project was carried out by Michael Schmid and Misha Sherman in collaboration with Paul Matsudaira at MIT.
References
(1). Ludtke, S. J., Chen, D. H., Song, J. L., Chuang, D. T. & Chiu, W. Structure (Camb) 12, 1129-36 (2004).
(2). Tilney, L. G. J Cell Biol 64, 289-310 (1975).
(3). DeRosier, D., Tilney, L. & Flicker, P. J Mol Biol 137, 375-89 (1980).
(4). Schmid, M. F. J Struct Biol 144, 195-208 (2003).
(5). Schmid, M. F., Sherman, M. B., Matsudaira, P. & Chiu, W. Nature accepted pending for revision (2004).
