Review: automatic particle detection in electron microscopy

Advances in cryoEM and single-particle reconstruction have led to results at increasingly high resolutions. However, to sustain continuing improvements in resolution it will be necessary to increase the number of particles included in performing the reconstructions. Manual selection of particles, even when assisted by computer preselection, is a bottleneck that will become significant as single-particle reconstructions are scaled up to achieve near-atomic resolutions. This review describes various approaches that have been developed to address the problem of automatic particle selection. The principal conclusions that have been drawn from the results so far are: (1) cross-correlation with a reference image ("matched filtering") is an effective way to identify candidate particles, but it is inherently unable to avoid also selecting false particles; (2) false positives can be eliminated efficiently on the basis of estimates of particle size, density, and texture; (3) successful application of edge detection (or contouring) to particle identification may require improvements over currently available methods; and (4) neural network techniques, while computationally expensive, must also be investigated as a technology for eliminating false particles.     

A near atomic resolution structure of a Melanocarpus albomyces laccase

It is becoming routine for cryoEM single particle reconstructions to result in 3D electron density maps with resolutions of 10 Å, but maps with resolutions of 5 Å or better are still celebrated events. The electron microscope has a resolving power to better than 2 Å, 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 × 4K CCD, side entry cold stage) and a completely automated approach, these approaches resulted in a map with a nominal resolution of 5.4 Å (FSC_0.5) in which secondary structure is clearly discernable and the handedness of some of the α-helices in the GroEL structure can be determined.     

Combined EM/X-ray imaging yields a quasi-atomic model of the adenovirus-related bacteriophage PRD1 and shows key capsid and membrane interactions

BACKGROUND: The dsDNA bacteriophage PRD1 has a membrane inside its icosahedral capsid. While its large size (66 MDa) hinders the study of the complete virion at atomic resolution, a 1.65-A crystallographic structure of its major coat protein, P3, is available. Cryo-electron microscopy (cryo-EM) and three-dimensional reconstruction have shown the capsid at 20-28 A resolution. Striking architectural similarities between PRD1 and the mammalian adenovirus indicate a common ancestor. RESULTS: The P3 atomic structure has been fitted into improved cryo-EM reconstructions for three types of PRD1 particles: the wild-type virion, a packaging mutant without DNA, and a P3-shell lacking the membrane and the vertices. Establishing the absolute EM scale was crucial for an accurate match. The resulting "quasi-atomic" models of the capsid define the residues involved in the major P3 interactions, within the quasi-equivalent interfaces and with the membrane, and show how these are altered upon DNA packaging. CONCLUSIONS: The new cryo-EM reconstructions reveal the structure of the PRD1 vertex and the concentric packing of DNA. The capsid is essentially unchanged upon DNA packaging, with alterations limited to those P3 residues involved in membrane contacts. These are restricted to a few of the N termini along the icosahedral edges in the empty particle; DNA packaging leads to a 4-fold increase in the number of contacts, including almost all copies of the N terminus and the loop between the two beta barrels. Analysis of the P3 residues in each quasi-equivalent interface suggests two sites for minor proteins in the capsid edges, analogous to those in adenovirus.     

3D reconstruction of the ATP-bound form of CCT reveals the asymmetric folding conformation of a type II chaperonin.

The type II chaperonin CCT (chaperonin containing Tcp-1) of eukaryotic cytosol is a heteromeric 16-mer particle composed of eight different subunits. Three-dimensional reconstructions of apo-CCT and ATP-CCT have been obtained at 28 A resolution by cryo-electron microscopy. Binding of ATP generates an asymmetric particle; one ring has a slightly different conformation from the apo-CCT ring, while the other has undergone substantial movements in the apical domains. Upon ATP binding the apical domains rotate and point towards the cylinder axis, so that the helical protrusions present at their tips could act as a lid closing the ring cavity.     

Historical background: Why is it important to improve automated particle selection methods?

A current trend in single-particle electron microscopy is to compute three-dimensional reconstructions with ever-increasing numbers of particles in the data sets. Since manual--or even semi-automated--selection of particles represents a major bottleneck when the data set exceeds several thousand particles, there is growing interest in developing automatic methods for selecting images of individual particles. Except in special cases, however, it has proven difficult to achieve the degree of efficiency and reliability that would make fully automated particle selection a useful tool. The simplest methods such as cross correlation (i.e., matched filtering) do not perform well enough to be used for fully automated particle selection. Geometric properties (area, perimeter-to-area ratio, etc.) and the integrated "mass" of candidate particles are additional factors that could improve automated particle selection if suitable methods of contouring particles could be developed. Another suggestion is that data be always collected as pairs of images, the first taken at low defocus (to capture information at the highest possible resolution) and the second at very high defocus (to improve the visibility of the particle). Finally, it is emphasized that well-annotated, open-access data sets need to be established in order to encourage the further development and validation of methods for automated particle selection.     

ResLog plots as an empirical metric of the quality of cryo-EM reconstructions

Compared to the field of X-ray crystallography, the field of single particle three-dimensional electron microscopy has few reliable metrics for assessing the quality of 3D reconstructions. New metrics are needed that can determine whether a given 3D reconstruction accurately reflects the structure of the particles from which it was derived or instead depicts a plausible though incorrect structure due to coarse misalignment of particles. Here an empirical procedure is presented for differentiating between a reconstruction with well-aligned particles and a reconstruction with grossly misclassified particles. For a given dataset, 3D reconstructions are computed from subsets of particles with decreasing numbers of particles contributing to the reconstruction. A plot of inverse resolution vs. the logarithm of the number of particles (a “ResLog” plot) provides metrics for the reliability of the reconstruction and the overall quality of the dataset and processing. Specifically, the y-intercept of a regression line provides a measure of the relative accuracy of the particle alignment and classification, and the slope is an indicator of the overall data quality including the imaging conditions and processing steps. ResLog plots can also be used to optimize conditions for data collection and reconstruction parameters. Although resolution estimates can vary by method of calculation, ResLog-derived parameters are consistent whether calculated by Fourier shell correlation or Fourier neighbor correlation, or a new coordinate-based metric that serves as a yardstick for structures where atomic coordinates are available. ResLog plots could become part of a standard set of parameters to be included in 3D reconstruction reports.     

DNA packaging and delivery machines in tailed bacteriophages

Several symmetric and asymmetric reconstructions of bacteriophage particles have recently been determined using electron cryo-microscopy and image reconstruction, and X-ray crystal structures of phage particles and particle-associated gene products have also been solved. In the past two years, the asymmetric structures of four different phages, T7, epsilon15, P22 and phi29, were determined at resolutions sufficient to visualize details of the machinery for DNA packaging and delivery, as well as the organization of the double-stranded DNA within the particles. Invariably, the portals, through which DNA enters and leaves the particle, have 12-fold symmetry, occupy a pentavalent site in the capsid and, along with tail machine accessory proteins attached to it, are fixed in a specific orientation relative to the rest of the capsid.     

Unnatural amino acid incorporation onto adenoviral (Ad) coat proteins facilitates chemoselective modification and retargeting of Ad type 5 vectors

Surface modification of adenovirus vectors can improve tissue-selective targeting, attenuate immunogenicity, and enable imaging of particle biodistribution, thus significantly improving therapeutic potential. Currently, surface engineering is constrained by a combination of factors, including impact on viral fitness, limited access to functionality, or incomplete control over the site of modification. Here, we report a two-step labeling process involving an initial metabolic placement of a uniquely reactive unnatural amino acid, azidohomoalanine (Aha), followed by highly specific chemical modification. As genetic modification of adenovirus is unnecessary, vector production is exceedingly straightforward. Aha incorporation demonstrated no discernible impact on either virus production or infectivity of the resultant particles. "Click" chemical modification of surface-exposed azides was highly selective, allowing for the attachment of a wide range of functionality. Decoration of human adenovirus type 5 (hAd5) with folate, a known cancer-targeting moiety, provided an ～20-fold increase in infection of murine breast cancer cells (4T1) in a folate receptor-dependent manner. This study demonstrates that incorporation of unnatural amino acids can provide a flexible, straightforward route for the selective chemical modification of adenoviral vectors.     

Genetic and biochemical manipulations of the small ribosomal subunit from Thermus thermophilus HB8

Crystals of the small ribosomal subunit from Thermus thermophilus diffract to 3A and exhibit reasonable isomorphism and moderate resistance to irradiation. A 5A MIR map of this particle shows a similar shape to the part assigned to this particle within the cryo-EM reconstructions of the whole ribosome and contains regions interpretable either as RNA chains or as protein motifs. To assist phasing at higher resolution we introduced recombinant methods aimed at extensive selenation for MAD phasing. We are focusing on several ribosomal proteins that can be quantitatively detached by chemical means. These proteins can be modified and subsequently reconstituted into depleted ribosomal cores. They also can be used for binding heavy atoms, by incorporating chemically reactive binding sites, such as -SH groups, into them. In parallel we are co-crystallizing the ribosomal particles with tailor made ligands, such as antibiotics or cDNA to which heavy-atoms have been attached or diffuse the latter compounds into already formed crystals.     

A two step approach for semi-automated particle selection from low contrast cryo-electron micrographs

Over recent years advances in cryo-electron microscopy for the study of macromolecular structure have resulted in resolutions in the range 10-15 A becoming routine. With this drive for increased resolution comes the need to collect larger datasets, commonly >10,000 particle images. Manual selection of particles from micrographs is often difficult and with such large numbers of particles now involved it is also laborious and a common bottleneck. Automated methods do exist but are normally restricted to specific samples or data, i.e., spherical particles, no aggregation, high contrast, and low noise. A two step approach has been developed that remains general and can be applied to low contrast, high noise micrographs of small molecules. Specifically, application of the approach is presented using micrographs of Escherichia coli RNA polymerase, which due to low contrast and the relatively small size of the molecule prove difficult to pick manually. To test the automated approach, independent reconstructions of RNA polymerase were carried out using manual and automatically picked data. The two reconstructions are shown to be comparable and the reconstruction from the automatically picked dataset is at a higher resolution, due to an increase in the number of particles picked.     

Structural basis of the iron storage function of frataxin from single-particle reconstruction of the iron-loaded oligomer

The mitochondrial protein frataxin plays a central role in mitochondrial iron homeostasis, and frataxin deficiency is responsible for Friedreich ataxia, a neurodegenerative and cardiac disease that affects 1 in 40000 children. Here we present a single-particle reconstruction from cryoelectron microscopic images of iron-loaded 24-subunit oligomeric frataxin particles at 13 and 17 A resolution. Computer-aided classification of particle images showed heterogeneity in particle size, which was hypothesized to result from gradual accumulation of iron within the core structure. Thus, two reconstructions were created from two classes of particles with iron cores of different sizes. The reconstructions show the iron core of frataxin for the first time. Compared to the previous reconstruction of iron-free particles from negatively stained images, the higher resolution of the present reconstruction allowed a more reliable analysis of the overall three-dimensional structure of the 24-meric assembly. This was done after docking the X-ray structure of the frataxin trimer into the EM reconstruction. The structure revealed a close proximity of the suggested ferroxidation sites of different monomers to the site proposed to serve in iron nucleation and mineralization. The model also assigns a new role to the N-terminal helix of frataxin in controlling the channel at the 4-fold axis of the 24-subunit oligomer. The reconstructions show that, together with some common features, frataxin has several unique features which distinguish it from ferritin. These include the overall organization of the oligomers, the way they are stabilized, and the mechanisms of iron core nucleation.     

Adenovirus binds to rat brain microtubules in vitro.

We have found by negative staining electron microscopy that when similar concentrations of adenovirus and reovirus (viruses of about the same diameter, 75 to 80 nm, and density, 1.34 to 1.36 g/cm3) were incubated with a carbon support film containing microtubules, 72% of adenovirus on the grid, but only 32% (equivalent to random association) of reovirus, were associated with microtubules. Similar concentrations of both larger and smaller particles, such as polystyrene latex spheres and coliphage f2, also exhibited a low degree of interaction, viz., 17 to 37%, with microtubules. Moreover, 90% of microtubule-associated adenovirus binds to within +/- 4 nm of the edge of microtubules, but lower fractions (again equivalent to a random association) of the other particles bind to the edge of the microtubules. The mechanism behind this phenomenon, which we denote as "edge binding," is presently obscure; however, it provides us with a second, albeit empirical, method to distinguish between the microtubular association of adenovirus and other particles. We found that edge binding of adenovirus also occurred when adenovirus was initially placed on the carbon support film and then incubated with microtubules and when adenovirus and microtubules were mixed prior to placement on the support. In contrast, reovirus or the other particles prepared by similar techniques exhibited a random amount of edge binding. The binding of adenovirus appears to involve the hexon capsomers of the virion since (i) high resolution electron micrographs showed that the edge of the virus was in contact with the edge of the microtubules, and (ii) adenovirions briefly treated with formamide to remove pentons and fibers bind as efficiently as intact virions. Core structures, which were obtained by further formamide degradation of the virion, do not associate with microtubules. These observations support the hypothesis of Dales and Chardonnet (1973) that the transport of adenovirions within infected cells is mediated by interaction with microtubules.     

Stoichiometry and Topology in Protein Folding

Abstract

Crystals of the small ribosomal subunit from Thermus thermophilus diffract to 3Å and exhibit reasonable isomorphism and moderate resistance to irradiation. A 5Å MIR map of this particle shows a similar shape to the part assigned to this particle within the cryo-EM reconstructions of the whole ribosome and contains regions interpretable either as RNA chains or as protein motifs. To assist phasing at higher resolution we introduced recombinant methods aimed at extensive selenation for MAD phasing. We are focusing on several ribosomal proteins that can be quantitatively detached by chemical means. These proteins can be modified and subsequently reconstituted into depleted ribosomal cores. They also can be used for binding heavy atoms, by incorporating chemically reactive binding sites, such as -SH groups, into them. In parallel we are co-crystallizing the ribosomal particles with tailor made ligands, such as antibiotics or cDNA to which heavy-atoms have been attached or diffuse the latter compounds into already formed crystals.     

Fast automatic particle picking from cryo-electron micrographs using a locally normalized cross-correlation function: a case study

Recent progress in single-particle reconstruction methods and cryo-EM techniques has led to the determination of macromolecular structures with unprecedented resolution. The number of particles that goes into the reconstruction is a key determinant in achieving high resolution. Interactive manual picking of particles from an electron micrograph is a very time-consuming, tedious, and inefficient process. We have implemented a fast automatic particle picking procedure in the SPIDER environment. The procedure makes use of template matching schemes and employs a recently developed locally normalized correlation algorithm based on Fourier techniques. As a test, we have used this procedure to pick 70S Escherichia coli ribosomes from a cryo-electron micrograph. Different search strategies including use of a circular mask and asymmetric masks for different orientations of the particle have been explored, and their relative efficiencies are discussed. The results indicate that the procedure can be optimally used to pick ribosomes in a fully automatic way within the limit of selecting less than 10% false positives while missing about 15% of true positives.     

Replication in simian cells of defective viruses in an SV40-adenovirus "hybrid" population

Butel, Janet S. (Baylor University College of Medicine, Houston, Tex.), and Fred Rapp. Replication in simian cells of defective viruses in an SV40-adenovirus "hybrid" population. J. Bacteriol. 91:278-284. 1966.-An SV40-adenovirus type 7 "hybrid" virus population, previously shown to contain two viruses capable of complementation in green monkey kidney (GMK) cells, has a growth cycle in GMK cells similar to that of adenovirus type 7 in the presence of SV40. Extending previous preliminary results, the addition of adenovirus types 2, 7, or 12 to monolayers of GMK cells enhanced plaque formation by the SV40-adenovirus hybrid by as much as 200-fold. The terminal enhanced plaques, initiated by the hybrid in the presence of helper adenovirus, were found to contain progeny which could induce the synthesis of SV40 tumor antigen but which were coated with the protein of the helper adenovirus, type 2, 7, or 12, respectively. The particle carrying the SV40 tumor antigen determinant, named PARA, is defective in that it cannot direct the synthesis of capsid protein; information for the coat for PARA is supplied by the adenovirus. One-step growth curves of the hybrid virus population in monkey cells revealed that synthesis of both types of particles, adenovirus and PARA, proceeds at a similar rate, with a latent period of 16 to 20 hr being followed by an exponential increase in titer during the following 20 hr. Maximal titers for both particles were obtained 48 hr after inoculation of the cultures. Neither the PARA nor the adenovirus component replicated in GMK cells in the absence of the other.     

Multimodal Nanoscale Tomographic Imaging for Battery Electrodes

Accurate representations of the 3D structure within a lithium‐ion battery are key to understanding performance limitations. However, obtaining exact reconstructions of electrodes, where the active particles, the carbon black and polymeric binder domain, and the pore space are visualized is challenging. Here, it is shown that multimodal imaging can be used to overcome this challenge. High‐resolution ptychographic X‐ray computed tomography are combined with lower resolution but higher contrast transmission X‐ray tomographic microscopy to obtain 3D reconstructions of pristine and cycled graphite‐silicon composite electrodes. This cross‐correlation enables quantitative analysis of the surface of active particles, including the heterogeneity of carbon‐black and binder domain and solid‐electrolyte interphase coverage. Capturing the active particles as well as the carbon black‐binder domain allows using these segmented structures for electrochemical simulations to highlight the influence of the particle embedding on local state of charge heterogeneities.     

Labeling of adenovirus particles with PARACEST agents

Recombinant adenovirus type 5 particles (AdCMVLuc) were labeled with two different bifunctional ligands capable of forming stable complexes with paramagnetic lanthanide ions. The number of covalently attached ligands varied between 630 and 1960 per adenovirus particle depending upon the chemical reactivity of the bifunctional ligand (NHS ester versus isothiocyanide), the amount of excess ligand added, and the reaction time. The bioactivity of each labeled adenovirus derivative, as measured by the ability of the virus to infect cells and express luciferase, was shown to be highly dependent upon the number of covalently attached ligands. This indicates that certain amino groups, likely on the surface of the adenovirus fiber protein where cell binding is known to occur, are critical for viral attachment and infection. Addition of (177)Lu3+ to chemically modified versus control viruses demonstrated a significant amount of nonspecific binding of (177)Lu3+ to the virus particles that could not be sequestered by addition of excess DTPA. Thus, it became necessary to implement a prelabeling strategy for conjugation of preformed lanthanide ligand chelates to adenovirus particles. Using preformed Tm3+- L2, a large number of chelates having chemical exchange saturation transfer (CEST) properties were attached to the surface residues of AdCMVLuc without nonspecific binding of metal ions elsewhere on the virus particle. The potential of such conjugates to act as PARACEST imaging agents was tested using an on-resonance WALTZ sequence for CEST activation. A 12% decrease in bulk water signal intensity was observed relative to controls. This demonstrates that viral particles labeled with PARACEST-type imaging agents can potentially serve as targeted agents for molecular imaging.     

Structure and Composition of a Small Particle Prepared from a Simian Adenovirus.

Mayor, Heather D. (Baylor University College of Medicine, Houston, Tex.), Richard M. Jamison, Liane E. Jordan, and Joseph L. Melnick. Structure and composition of a small particle prepared from a simian adenovirus. J. Bacteriol. 90:235-242. 1965.-When tissue-culture fluids infected with simian adenovirus SV15 are examined in an electron microscope, either as fresh harvests or after treatment with Genetron, typical mature adenovirus particles are found. These are 65 to 70 mmu in diameter, with an icosahedral capsid built from 252 capsomeres. Also present is a population of small polyhedral particles approximately 20 mmu in diameter. These small particles can be separated from the mature virions by ultrafiltration or density gradient centrifugation. The small particles have a density of 1.43 in cesium chloride. They contain protein and double-stranded deoxyribonucleic acid. They appear to possess cubic symmetry of the icosahedral type, with a coat composed of 12 subunits each at the vertex of an icosahedron.