Toward atomic resolution structural determination by cryoEM

Hong Zhou
University of California, Los Angeles (UCLA)
California NanoSystems Institute and MIMG

Recent advances have given cryo-electron microscopy and single-particle reconstruction (cryoEM) an increasingly important role in determining the structures of macromolecular complexes (>150 kDa / 10 nm) approaching atomic resolution. At this resolution, the majority of amino-acid backbones and some bulky side chains are resolved in addition to the secondary structural elements of α-helices, β-sheets and loops. Such structural data provide valuable constraints for building pseudo-atomic, or atomic-resolution models of supramolecular machines through integrative cryoEM-bioinformatics modeling means. CryoEM provide exciting opportunities to determine the 3D structures of sub-cellular assemblies that are either too large or too heterogeneous to be investigated by conventional X-ray crystallographic or NMR methods.

Many challenges exist in pushing the resolution limit of cryoEM toward atomic resolution. I will discuss practical considerations on sample preparation; choices of electron source, voltage, specimen cooling temperature (liquid nitrogen vs. liquid helium), and recording media (films vs CCD); image processing and corrections; choices of reconstruction and post-reconstruction processing methods; and structural interpretation and atomic model building.

I will present the three-dimensional structure of cytoplasmic polyhedrosis virus (CPV) at 3.88-Å resolution by cryoEM. CPV is unique within Reoviridae in having a turreted single-layer capsid contained within polyhedrin inclusion bodies, yet being fully capable of cell entry and mRNA transcription and processing. Notably, this structure was obtained by averaging about 13,000 carefully selected CPV images recorded at liquid nitrogen specimen temperature using 300keV electrons on a 16-megapixel CCD camera. Our map clearly shows the turns and deep grooves of α- helices, the strand separation in ß-sheets, and densities for loops and many bulky side-chains; thus prompting atomic model building from cryoEM maps for the first time. Despite of no recognizable sequence homologies between CPV proteins and other viral proteins with known atomic structures, our models show the conserved folds of CPV proteins to their functional/structural homologs in other viruses of the Reoviridae, as well as new conformations and insertional domains. For example, a dramatic helix-to-β-hairpin conformational change has been observed between the two states of the capsid shell proteins, probably resulting from genome replication in the early stage of viral assembly. We have discovered an mRNA release hole coupled with the mRNA capping machinery in a way unique to CPV. Finally, we have identified a new β-strand-rich polyhedron-binding domain at N-terminal of the turret protein, a highly sought-after structure due to its potential in nanobiotechnology applications.

Back to Image Analysis Challenges in Molecular Microscopy