Visualization of alpha-helical features in a density map constructed using 9 molecular images of the 1.8 MDa icosahedral core of pyruvate dehydrogenase

On bias, variance, overfitting, gold standard and consensus in single-particle analysis by cryo-electron microscopy

Single-particle analysis (SPA) by cryo-electron microscopy comprises the estimation of many parameters along its image-processing pipeline. Overfitting observed in SPA is normally due to misestimated parameters, and the only way to identify these is by comparing the estimates of multiple algorithms or, at least, multiple executions of the same algorithm.

Cryo-electron microscopy (cryoEM) has become a well established technique to elucidate the 3D structures of biological macromolecules. Projection images from thousands of macromolecules that are assumed to be structurally identical are combined into a single 3D map representing the Coulomb potential of the macromolecule under study. This article discusses possible caveats along the image-processing path and how to avoid them to obtain a reliable 3D structure. Some of these problems are very well known in the community. These may be referred to as sample-related (such as specimen denaturation at interfaces or non-uniform projection geometry leading to underrepresented projection directions). The rest are related to the algorithms used. While some have been discussed in depth in the literature, such as the use of an incorrect initial volume, others have received much less attention. However, they are fundamental in any data-analysis approach. Chiefly among them, instabilities in estimating many of the key parameters that are required for a correct 3D reconstruction that occur all along the processing workflow are referred to, which may significantly affect the reliability of the whole process. In the field, the term overfitting has been coined to refer to some particular kinds of artifacts. It is argued that overfitting is a statistical bias in key parameter-estimation steps in the 3D reconstruction process, including intrinsic algorithmic bias. It is also shown that common tools (Fourier shell correlation) and strategies (gold standard) that are normally used to detect or prevent overfitting do not fully protect against it. Alternatively, it is proposed that detecting the bias that leads to overfitting is much easier when addressed at the level of parameter estimation, rather than detecting it once the particle images have been combined into a 3D map. Comparing the results from multiple algorithms (or at least, independent executions of the same algorithm) can detect parameter bias. These multiple executions could then be averaged to give a lower variance estimate of the underlying parameters.

New software tools in EMAN2 inspired by EMDatabank map challenge

EMAN2 is an extensible software suite with complete workflows for performing high-resolution single particle analysis, 2-D and 3-D heterogeneity analysis, and subtomogram averaging, among other tasks. Participation in the recent CryoEM Map Challenge sponsored by the EMDatabank led to a number of significant improvements to the single particle analysis process in EMAN2. A new convolutional neural network particle picker was developed, which dramatically improves particle picking accuracy for difficult data sets. A new particle quality metric capable of accurately identifying "bad" particles with a high degree of accuracy, no human input, and a negligible amount of additional computation, has been introduced, and this now serves as a replacement for earlier human-biased methods. The way 3-D single particle reconstructions are filtered has been altered to be more comparable to the filter applied in several other popular software packages, dramatically improving the appearance of sidechains in high-resolution structures. Finally, an option has been added to perform local resolution-based iterative filtration, resulting in local resolution improvements in many maps.

Resolution advances in cryo-EM enable application to drug discovery

The prospect that the structures of protein assemblies, small and large, can be determined using cryo-electron microscopy (cryo-EM) is beginning to transform the landscape of structural biology and cell biology. Great progress is being made in determining 3D structures of biological assemblies ranging from icosahedral viruses and helical arrays to small membrane proteins and protein complexes. Here, we review recent advances in this field, focusing especially on the emerging use of cryo-EM in mapping the binding of drugs and inhibitors to protein targets, an application that requires structure determination at the highest possible resolutions. We discuss methods used to evaluate the information contained in cryo-EM density maps and consider strengths and weaknesses of approaches currently used to measure map resolution.

Validation of 3D EM Reconstructions: The Phantom in the Noise

Validation is a necessity to trust the structures solved by electron microscopy by single particle techniques. The impressive achievements in single particle reconstruction fuel its expansion beyond a small community of image processing experts. This poses the risk of inappropriate data processing with dubious results. Nowhere is it more clearly illustrated than in the recovery of a reference density map from pure noise aligned to that map—a phantom in the noise. Appropriate use of existing validating methods such as resolution-limited alignment and the processing of independent data sets (“gold standard”) avoid this pitfall. However, these methods can be undermined by biases introduced in various subtle ways. How can we test that a map is a coherent structure present in the images selected from the micrographs? In stead of viewing the phantom emerging from noise as a cautionary tale, it should be used as a defining baseline. Any map is always recoverable from noise images, provided a sufficient number of images are aligned and used in reconstruction. However, with smaller numbers of images, the expected coherence in the real particle images should yield better reconstructions than equivalent numbers of noise or background images, even without masking or imposing resolution limits as potential biases. The validation test proposed is therefore a simple alignment of a limited number of micrograph and noise images against the final reconstruction as reference, demonstrating that the micrograph images yield a better reconstruction. I examine synthetic cases to relate the resolution of a reconstruction to the alignment error as a function of the signal-to-noise ratio. I also administered the test to real cases of publicly available data. Adopting such a test can aid the microscopist in assessing the usefulness of the micrographs taken before committing to lengthy processing with questionable outcomes.

Atomic resolution cryo electron microscopy of macromolecular complexes

Single-particle cryo electron microscopy (cryoEM) is a technique for determining three-dimensional (3D) structures from projection images of molecular complexes preserved in their "native," noncrystalline state. Recently, atomic or near-atomic resolution structures of several viruses and protein assemblies have been determined by single-particle cryoEM, allowing ab initio atomic model building by following the amino acid side chains or nucleic acid bases identifiable in their cryoEM density maps. In particular, these cryoEM structures have revealed extended arms contributing to molecular interactions that are otherwise not resolved by the conventional structural method of X-ray crystallography at similar resolutions. High-resolution cryoEM requires careful consideration of a number of factors, including proper sample preparation to ensure structural homogeneity, optimal configuration of electron imaging conditions to record high-resolution cryoEM images, accurate determination of image parameters to correct image distortions, efficient refinement and computation to reconstruct a 3D density map, and finally appropriate choice of modeling tools to construct atomic models for functional interpretation. This progress illustrates the power of cryoEM and ushers it into the arsenal of structural biology, alongside conventional techniques of X-ray crystallography and NMR, as a major tool (and sometimes the preferred one) for the studies of molecular interactions in supramolecular assemblies or machines.

JADAS: a customizable automated data acquisition system and its application to ice-embedded single particles

The JEOL Automated Data Acquisition System (JADAS) is a software system built for the latest generation of the JEOL Transmission Electron Microscopes. It is designed to partially or fully automate image acquisition for ice-embedded single particles under low dose conditions. Its built-in flexibility permits users to customize the order of various imaging operations. In this paper, we describe how JADAS is used to accurately locate and image suitable specimen areas on a grid of ice-embedded particles. We also demonstrate the utility of JADAS by imaging the epsilon 15 bacteriophage with the JEM3200FSC electron cryo-microscope, showing that sufficient images can be collected in a single 8h session to yield a subnanometer resolution structure which agrees with the previously determined structure.

A test-bed for optimizing high-resolution single particle reconstructions

It is becoming routine for cryoEM single particle reconstructions to result in 3D electron density maps with resolutions of approximately 10A, but maps with resolutions of 5A or better are still celebrated events. The electron microscope has a resolving power to better than 2A, and thus should not be a limiting factor; instead the practical limitations in resolution most likely arise from a combination of specimen preparation methods, data collection parameters, and data analysis procedures. With the aid of a highly automated system for acquiring images, coupled to a relational database to keep track of all processing parameters, we have taken a systematic approach to optimizing parameters affecting the resolution of single particle reconstructions. Using GroEL as a test-bed, we performed a series of 3D reconstructions where we systematically varied the number of particles used in computing the map, the accelerating voltage of the microscope, and the electron dose used to acquire the images. We also investigated methods for excluding unacceptable or "bad" particles from contributing to the final 3D map. Using relatively standard instrumentation (Tecnai F20, 4K x 4K CCD, side entry cold stage) and a completely automated approach, these approaches resulted in a map with a nominal resolution of 5.4A (FSC(0.5)) in which secondary structure is clearly discernable and the handedness of some of the alpha-helices in the GroEL structure can be determined.

Structures of the human pyruvate dehydrogenase complex cores: a highly conserved catalytic center with flexible N-terminal domains

Dihydrolipoyl acetyltransferase (E2) is the central component of pyruvate dehydrogenase complex (PDC), which converts pyruvate to acetyl-CoA. Structural comparison by cryo-electron microscopy (cryo-EM) of the human full-length and truncated E2 (tE2) cores revealed flexible linkers emanating from the edges of trimers of the internal catalytic domains. Using the secondary structure constraints revealed in our 8 A cryo-EM reconstruction and the prokaryotic tE2 atomic structure as a template, we derived a pseudo atomic model of human tE2. The active sites are conserved between prokaryotic tE2 and human tE2. However, marked structural differences are apparent in the hairpin domain and in the N-terminal helix connected to the flexible linker. These permutations away from the catalytic center likely impart structures needed to integrate a second component into the inner core and provide a sturdy base for the linker that holds the pyruvate dehydrogenase for access by the E2-bound regulatory kinase/phosphatase components in humans.

Averaging tens to hundreds of icosahedral particle images to resolve protein secondary structure elements using a Multi-Path Simulated Annealing optimization algorithm

Accurately determining a cryoEM particle's alignment parameters is crucial to high resolution single particle 3-D reconstruction. We developed Multi-Path Simulated Annealing, a Monte-Carlo type of optimization algorithm, for globally aligning the center and orientation of a particle simultaneously. A consistency criterion was developed to ensure the alignment parameters are correct and to remove some bad particles from a large pool of images of icosahedral particles. Without using any a priori model, this procedure is able to reconstruct a structure from a random initial model. Combining the procedure above with a new empirical double threshold particle selection method, we are able to pick tens of best quality particles to reconstruct a subnanometer resolution map from scratch. Using the best 62 particles of rice dwarf virus, the reconstruction reached 9.6A resolution at which four helices of the P3A subunit of RDV are resolved. Furthermore, with the 284 best particles, the reconstruction is improved to 7.9A resolution, and 21 of 22 helices and six of seven beta sheets are resolved.

Structure determination of clathrin coats to subnanometer resolution by single particle cryo-electron microscopy

Clathrin triskelions can assemble into lattices of different shapes, sizes and symmetries. For many years, the structures of clathrin lattices have been studied by single particle cryo-electron microscopy, which probed the architecture of the D6 hexagonal barrel clathrin coat at the molecular level. By introducing additional image processing steps we have recently produced a density map for the D6 barrel clathrin coat at subnanometer resolution, enabling us to generate an atomic model for this lattice [Fotin, A., Cheng, Y., Sliz, P., Grigorieff, N., Harrison, S.C., Kirchhausen, T., Walz, T., 2004. Molecular model for a complete clathrin lattice from electron cryomicroscopy. Nature 432, 573-579]. We describe in detail here the image processing steps that we have added to produce a density map at this high resolution. These procedures should be generally applicable and may thus help determine the structures of other large protein assemblies to higher resolution by single particle cryo-electron microscopy.

Molecular structure of a 9-MDa icosahedral pyruvate dehydrogenase subcomplex containing the E2 and E3 enzymes using cryoelectron microscopy

The pyruvate dehydrogenase multienzyme complexes are among the largest multifunctional catalytic machines in cells, catalyzing the production of acetyl CoA from pyruvate. We have previously reported the molecular architecture of an 11-MDa subcomplex comprising the 60-mer icosahedral dihydrolipoyl acetyltransferase (E2) decorated with 60 copies of the heterotetrameric (alpha(2)beta(2)) 153-kDa pyruvate decarboxylase (E1) from Bacillus stearothermophilus (Milne, J. L. S., Shi, D., Rosenthal, P. B., Sunshine, J. S., Domingo, G. J., Wu, X., Brooks, B. R., Perham, R. N., Henderson, R., and Subramaniam, S. (2002) EMBO J. 21, 5587-5598). An annular gap of approximately 90 A separates the acetyltransferase catalytic domains of the E2 from an outer shell formed of E1 tetramers. Using cryoelectron microscopy, we present here a three-dimensional reconstruction of the E2 core decorated with 60 copies of the homodimeric 100-kDa dihydrolipoyl dehydrogenase (E3). The E2E3 complex has a similar annular gap of approximately 75 A between the inner icosahedral assembly of acetyltransferase domains and the outer shell of E3 homodimers. Automated fitting of the E3 coordinates into the map suggests excellent correspondence between the density of the outer shell map and the positions of the two best fitting orientations of E3. As in the case of E1 in the E1E2 complex, the central 2-fold axis of the E3 homodimer is roughly oriented along the periphery of the shell, making the active sites of the enzyme accessible from the annular gap between the E2 core and the outer shell. The similarities in architecture of the E1E2 and E2E3 complexes indicate fundamental similarities in the mechanism of active site coupling involved in the two key stages requiring motion of the swinging lipoyl domain across the annular gap, namely the synthesis of acetyl CoA and regeneration of the dithiolane ring of the lipoyl domain.

Electron cryomicroscopy of single particles at subnanometer resolution

Electron cryomicroscopy and single-particle reconstruction have advanced substantially over the past two decades. There are now numerous examples of structures that have been solved using this technique to better than 10 A resolution. At such resolutions, direct identification of alpha helices is possible and, often, beta-sheet-containing regions can be identified. The most numerous subnanometer resolution structures are the icosahedral viruses, as higher resolution is easier to achieve with higher symmetry. Important non-icosahedral structures solved to subnanometer resolution include several ribosome structures, clathrin assemblies and, most recently, the Ca2+ release channel. There is now hope that, in the next few years, this technique will achieve resolutions approaching 4 A, permitting a complete trace of the protein backbone without reference to a crystal structure.

Building the atomic model for the bacterial flagellar filament by electron cryomicroscopy and image analysis

The bacterial flagellar filament is a helical propeller for bacterial locomotion. It is a well-ordered helical assembly of a single protein, flagellin, and its tubular structure is formed by 11 protofilaments, each in either of the two distinct conformations, L- and R-type, for supercoiling. We have been studying the three-dimensional structures of the flagellar filaments by electron cryomicroscopy and recently obtained a density map of the R-type filament up to 4 angstroms resolution from an image data set containing only about 41,000 molecular images. The density map showed the features of the alpha-helical backbone and some large side chains, which allowed us to build the complete atomic model as one of the first atomic models of macromolecules obtained solely by electron microscopy image analysis (Yonekura et al., 2003a). We briefly review the structure and the structure analysis, and point out essential techniques that have made this analysis possible.

Bridging the imaging gap: visualizing subcellular architecture with electron tomography

Transmission electron microscopy is a powerful tool that is used to explore the internal structure of tissues, cells, organelles and macromolecular complexes. By integrating data from a series of images in which the orientation of the specimen is progressively varied relative to the incident electron beam it is also possible to extend electron microscopic imaging into the third dimension. This approach, commonly referred to as electron tomography, has been greatly aided in recent years by advances in technology for imaging specimens at cryogenic temperatures, as well as by substantial progress in procedures for automated data collection and image processing. The intense pace of developments in this field is inspired, in a large part, by the hope that the quality of the data will ultimately be good enough to allow interpretation of tomograms of cells, organelles, bacteria and viruses in terms of the three-dimensional spatial arrangements of the constituent molecules.

High-throughput film-densitometry: an efficient approach to generate large data sets

A film-handling machine (robot) has been built which can, in conjunction with a commercially available film densitometer, exchange and digitize over 300 electron micrographs per day. Implementation of robotic film handling effectively eliminates the delay and tedium associated with digitizing images when data are initially recorded on photographic film. The modulation transfer function (MTF) of the commercially available densitometer is significantly worse than that of a high-end, scientific microdensitometer. Nevertheless, its signal-to-noise ratio (S/N) is quite excellent, allowing substantial restoration of the output to "near-to-perfect" performance. Due to the large area of the standard electron microscope film that can be digitized by the commercial densitometer (up to 10,000 x 13,680 pixels with an appropriately coded holder), automated film digitization offers a fast and inexpensive alternative to high-end CCD cameras as a means of acquiring large amounts of image data in electron microscopy.

CryoEM Structure at 9Å Resolution of an Adenovirus Vector Targeted to Hematopoietic Cells

We report a sub-nanometer resolution cryo-electron microscopy (cryoEM) structural analysis of an adenoviral vector, Ad35F, comprised of an adenovirus type 5 (Ad5) capsid pseudo-typed with an Ad35 fiber. This vector transduces human hematopoietic cells via association of its fiber protein with CD46, a member of the complement regulatory protein family. Major advances in data acquisition and image processing allowed a significant improvement in resolution compared to earlier structures. Analysis of the cryoEM density was enhanced by docking the crystal structures of both the hexon and penton base capsid proteins. CryoEM density was observed for hexon residues missing from the crystal structure that include hypervariable regions and the epitope of a neutralizing monoclonal antibody. Within the penton base, density was observed for the integrin-binding RGD loop missing from the crystal structure and for the flexible beta ribbon of the variable loop on the side of the penton base. The Ad35 fiber is flexible, consistent with the sequence insert in the third beta-spiral repeat. On the inner capsid surface density is revealed at the base of the hexons and below the penton base. A revised model is presented for protein IX within the virion. Well-defined density was assigned to a conserved domain in the N terminus of protein IX required for incorporation into the virion. For the C-terminal domain of protein IX two alternate conformations are proposed, either binding on the capsid surface or extending away from the capsid. This model is consistent with the tolerance of the C terminus for inserted ligands and its potential use in vector retargeting. This structural study increases our knowledge of Ad capsid assembly, antibody neutralization mechanisms, and may aid further improvements in gene delivery to important human cell types.

Structure determination of macromolecular assemblies by single-particle analysis of cryo-electron micrographs

A new generation of electron microscopes equipped with field emission gun electron sources and the ability to image molecules in their native environment at liquid nitrogen or helium temperatures has enabled the analysis of macromolecular structures at medium resolution (approximately 10 angstroms) and in different conformational states. The amalgamation of electron microscopy and X-ray crystallographic approaches makes it possible to solve structures in the 100-1000 angstroms size range, advancing our understanding of the function of complex assemblies. Many new structures have been solved during the past two years, including one of the smallest complexes to be determined by single-particle cryo-electron microscopy, the transferrin receptor-transferrin complex. Other notable results include the near atomic level resolution structure of the nicotinic acetylcholine receptor in helical arrays and an icosahedral virus structure with an asymmetric polymerase resolved.