Structural biology in situ--the potential of subtomogram averaging

Automated fiducial-based alignment of cryo-electron tomography tilt series in Dynamo

With the advent of modern technologies for cryo-electron tomography (cryo-ET), high-quality tilt series are more rapidly acquired than processed and analyzed. Thus, a robust and fast-automated alignment for batch processing in cryo-ET is needed. While different software packages have made available several approaches for automated marker-based alignment of tilt series, manual user intervention remains necessary for many datasets, thus preventing high-throughput tomography. We have developed a MATLAB-based framework integrated into the Dynamo software package for automatic detection of fiducial markers that generates a robust alignment model with minimal input parameters. This approach allows high-throughput, unsupervised volume reconstruction. This new module extends Dynamo with a large repertory of tools for tomographic alignment and reconstruction, as well as specific visualization browsers to rapidly assess the biological relevance of the dataset. Our approach has been successfully tested on a broad range of datasets that include diverse biological samples and cryo-ET modalities.

Recent technical advances in cellular cryo-electron tomography

Understanding the in situ structure, organization, and interactions of macromolecules is essential for elucidating their functions and mechanisms of action. Cellular cryo-electron tomography (cryo-ET) is a cutting-edge technique that reveals in situ molecular-resolution architectures of macromolecules in their lifelike states. It also provides insights into the three-dimensional distribution of macromolecules and their spatial relationships with various subcellular structures. Thus, cellular cryo-ET bridges the gap between structural biology and cell biology. With rapid advancements, this technique achieved substantial improvements in throughput, automation, and resolution. This review presents the fundamental principles and methodologies of cellular cryo-ET, highlighting recent developments in sample preparation, data collection, and image processing. We also discuss emerging trends and potential future directions. As cellular cryo-ET continues to develop, it is set to play an increasingly vital role in structural cell biology.

Cryo-Electron Microscopy Studies of Biomolecular Structure and Dynamics

The technical innovation of the last decade has provided novel tools that are now transforming the field of biophysics by bringing remarkable atomic level insights into the mechanisms employed by bio-micromachines to sustain life [...].

Optimizing Contrast in Automated 4D STEM Cryotomography

4D STEM is an emerging approach to electron microscopy. While it was developed principally for high-resolution studies in materials science, the possibility to collect the entire transmitted flux makes it attractive for cryomicroscopy in application to life science and radiation-sensitive materials where dose efficiency is of utmost importance. We present a workflow to acquire tomographic tilt series of 4D STEM data sets using a segmented diode and an ultrafast pixelated detector, demonstrating the methods using a specimen of a T4 bacteriophage. Full integration with the SerialEM platform conveniently provides all the tools for grid navigation and automation of the data collection. Scripts are provided to convert the raw data to mrc format files and further to generate a variety of modes representing both scattering and phase contrasts, including incoherent and annular bright field, integrated center of mass, and parallax decomposition of a simulated integrated differential phase contrast. Principal component analysis of virtual annular detectors proves particularly useful, and axial contrast is improved by 3D deconvolution with an optimized point spread function. Contrast optimization enables visualization of irregular features such as DNA strands and thin filaments of the phage tails, which would be lost upon averaging or imposition of an inappropriate symmetry.

Missing Wedge Completion via Unsupervised Learning with Coordinate Networks

Cryogenic electron tomography (cryoET) is a powerful tool in structural biology, enabling detailed 3D imaging of biological specimens at a resolution of nanometers. Despite its potential, cryoET faces challenges such as the missing wedge problem, which limits reconstruction quality due to incomplete data collection angles. Recently, supervised deep learning methods leveraging convolutional neural networks (CNNs) have considerably addressed this issue; however, their pretraining requirements render them susceptible to inaccuracies and artifacts, particularly when representative training data is scarce. To overcome these limitations, we introduce a proof-of-concept unsupervised learning approach using coordinate networks (CNs) that optimizes network weights directly against input projections. This eliminates the need for pretraining, reducing reconstruction runtime by 3-20× compared to supervised methods. Our in silico results show improved shape completion and reduction of missing wedge artifacts, assessed through several voxel-based image quality metrics in real space and a novel directional Fourier Shell Correlation (FSC) metric. Our study illuminates benefits and considerations of both supervised and unsupervised approaches, guiding the development of improved reconstruction strategies.

Missing Wedge Completion via Unsupervised Learning with Coordinate Networks

Cryogenic electron tomography (cryoET) is a powerful tool in structural biology, enabling detailed 3D imaging of biological specimens at a resolution of nanometers. Despite its potential, cryoET faces challenges such as the missing wedge problem, which limits reconstruction quality due to incomplete data collection angles. Recently, supervised deep learning methods leveraging convolutional neural networks (CNNs) have considerably addressed this issue; however, their pretraining requirements render them susceptible to inaccuracies and artifacts, particularly when representative training data is scarce. To overcome these limitations, we introduce a proof-of-concept unsupervised learning approach using coordinate networks (CNs) that optimizes network weights directly against input projections. This eliminates the need for pretraining, reducing reconstruction runtime by 3 - 20× compared to supervised methods. Our in silico results show improved shape completion and reduction of missing wedge artifacts, assessed through several voxel-based image quality metrics in real space and a novel directional Fourier Shell Correlation (FSC) metric. Our study illuminates benefits and considerations of both supervised and unsupervised approaches, guiding the development of improved reconstruction strategies.

Thinner is not always better: Optimizing cryo-lamellae for subtomogram averaging

Cryo-electron tomography (cryo-ET) is a powerful method to elucidate subcellular architecture and to structurally analyze biomolecules in situ by subtomogram averaging, yet data quality critically depends on specimen thickness. Cells that are too thick for transmission imaging can be thinned into lamellae by cryo-focused ion beam (cryo-FIB) milling. Despite being a crucial parameter directly affecting attainable resolution, optimal lamella thickness has not been systematically investigated nor the extent of structural damage caused by gallium ions used for FIB milling. We thus systematically determined how resolution is affected by these parameters. We find that ion-induced damage does not affect regions more than 30 nanometers from either lamella surface and that up to ~180-nanometer lamella thickness does not negatively affect resolution. This shows that there is no need to generate very thin lamellae and lamella thickness can be chosen such that it captures cellular features of interest, thereby opening cryo-ET also for studies of large complexes.

DeepETPicker: Fast and accurate 3D particle picking for cryo-electron tomography using weakly supervised deep learning

To solve three-dimensional structures of biological macromolecules in situ, large numbers of particles often need to be picked from cryo-electron tomograms. However, adoption of automated particle-picking methods remains limited because of their technical limitations. To overcome the limitations, we develop DeepETPicker, a deep learning model for fast and accurate picking of particles from cryo-electron tomograms. Training of DeepETPicker requires only weak supervision with low numbers of simplified labels, reducing the burden of manual annotation. The simplified labels combined with the customized and lightweight model architecture of DeepETPicker and accelerated pooling enable substantial performance improvement. When tested on simulated and real tomograms, DeepETPicker outperforms the competing state-of-the-art methods by achieving the highest overall accuracy and speed, which translate into higher authenticity and coordinates accuracy of picked particles and higher resolutions of final reconstruction maps. DeepETPicker is provided in open source with a user-friendly interface to support cryo-electron tomography in situ.

Characterization of Complexes and Supramolecular Structures by Electron Microscopy

Recent advancements in cryo-electron microscopy (cryo-TEM) have enabled the determination of structures of macromolecular complexes at near-atomic resolution, establishing it as a pivotal tool in Structural Biology. This high resolution allows for the detection of ligands and substrates under physiological conditions. Enhancements in detectors and imaging devices, like phase plates, improve signal quality, facilitating the reconstruction of even smaller macromolecular complexes. The 100-kDa barrier has been surpassed, presenting new opportunities for pharmacological research and expanding the scope of crystallographic analyses in the pharmaceutical industry. Cryo-TEM produces vast data sets from minimal samples, and refined classification methods can identify different conformational states of macromolecular complexes, offering deeper insights into the functional characteristics of macromolecular systems. Additionally, cryo-TEM is paving the way for time-resolved microscopy, with rapid freezing techniques capturing snapshots of vital structural changes in biological complexes. Finally, in Structural Cell Biology, advanced cryo-TEM, through tomographic procedures, is revealing conformational changes related to the specific subcellular localization of macromolecular systems and their interactions within cells.

Advanced Techniques for Detecting Protein Misfolding and Aggregation in Cellular Environments

Protein misfolding and aggregation, a key contributor to the progression of numerous neurodegenerative diseases, results in functional deficiencies and the creation of harmful intermediates. Detailed visualization of this misfolding process is of paramount importance for improving our understanding of disease mechanisms and for the development of potential therapeutic strategies. While in vitro studies using purified proteins have been instrumental in delivering significant insights into protein misfolding, the behavior of these proteins in the complex milieu of living cells often diverges significantly from such simplified environments. Biomedical imaging performed in cell provides cellular-level information with high physiological and pathological relevance, often surpassing the depth of information attainable through in vitro methods. This review highlights a variety of methodologies used to scrutinize protein misfolding within biological systems. This includes optical-based methods, strategies leaning on mass spectrometry, in-cell nuclear magnetic resonance, and cryo-electron microscopy. Recent advancements in these techniques have notably deepened our understanding of protein misfolding processes and the features of the resulting misfolded species within living cells. The progression in these fields promises to catalyze further breakthroughs in our comprehension of neurodegenerative disease mechanisms and potential therapeutic interventions.

Charting the molecular landscape of the cell

Biological function of macromolecules is closely tied to their cellular location, as well as to interactions with other molecules within the native environment of the cell. Therefore, to obtain detailed mechanistic insights into macromolecular functionality, one of the outstanding targets for structural biology is to produce an atomic-level understanding of the cell. One structural biology technique that has already been used to directly derive atomic models of macromolecules from cells, without any additional external information, is electron cryotomography (cryoET). In this perspective article, we discuss possible routes to chart the molecular landscape of the cell by advancing cryoET imaging as well as by embedding cryoET into correlative imaging workflows.

Virus structures revealed by advanced cryoelectron microscopy methods

Before the resolution revolution, cryoelectron microscopy (cryo-EM) single-particle analysis (SPA) already achieved resolutions beyond 4 Å for certain icosahedral viruses, enabling ab initio atomic model building of these viruses. As the only samples that achieved such high resolution at that time, cryo-EM method development was closely intertwined with the improvement of reconstructions of symmetrical viruses. Viral morphology exhibits significant diversity, ranging from small to large, uniform to non-uniform, and from containing single symmetry to multiple symmetries. Furthermore, viruses undergo conformational changes during their life cycle. Several methods, such as asymmetric reconstruction, Ewald sphere correction, cryoelectron tomography (cryo-ET), and sub-tomogram averaging (STA), have been developed and applied to determine virus structures in vivo and in vitro. This review outlines current advanced cryo-EM methods for high-resolution structure determination of viruses and summarizes accomplishments obtained with these approaches. Moreover, persisting challenges in comprehending virus structures are discussed and we propose potential solutions.

Dynamics of upstream ESCRT organization at the HIV-1 budding site

In the late stages of the HIV-1 life cycle, membrane localization and self-assembly of Gag polyproteins induce membrane deformation and budding. Release of the virion requires direct interaction between immature Gag lattice and upstream ESCRT machinery at the viral budding site, followed by assembly of downstream ESCRT-III factors, culminating in membrane scission. However, molecular details of upstream ESCRT assembly dynamics at the viral budding site remain unclear. In this work, using coarse-grained (CG) molecular dynamics (MD) simulations, we investigated the interactions between Gag, ESCRT-I, ESCRT-II, and membrane to delineate the dynamical mechanisms by which upstream ESCRTs assemble templated by late-stage immature Gag lattice. We first systematically derived "bottom-up" CG molecular models and interactions of upstream ESCRT proteins from experimental structural data and extensive all-atom MD simulations. Using these molecular models, we performed CG MD simulations of ESCRT-I oligomerization and ESCRT-I/II supercomplex formation at the neck of the budding virion. Our simulations demonstrate that ESCRT-I can effectively oligomerize to higher-order complexes templated by the immature Gag lattice both in the absence of ESCRT-II and when multiple copies of ESCRT-II are localized at the bud neck. The ESCRT-I/II supercomplexes formed in our simulations exhibit predominantly columnar structures, which has important implications for the nucleation pathway of downstream ESCRT-III polymers. Importantly, ESCRT-I/II supercomplexes bound to Gag initiate membrane neck constriction by pulling the inner edge of the bud neck closer to the ESCRT-I headpiece ring. Our findings serve to elucidate a network of interactions between upstream ESCRT machinery, immature Gag lattice, and membrane neck that regulate protein assembly dynamics at the HIV-1 budding site.

From the beginnings to multidimensional light and electron microscopy of virus morphogenesis

Individual functional viral morphogenesis events are often dynamic, short, and infrequent and might be obscured by other pathways and dead-end products. Volumetric live cell imaging has become an essential tool for studying viral morphogenesis events. It allows following entire dynamic processes while providing functional evidence that the imaged process is involved in viral production. Moreover, it allows to capture many individual events and allows quantitative analysis. Finally, the correlation of volumetric live-cell data with volumetric electron microscopy (EM) can provide crucial insights into the ultrastructure and mechanisms of viral morphogenesis events. Here, we provide an overview and discussion of suitable imaging methods for volumetric correlative imaging of viral morphogenesis and frame them in a historical summary of their development.

Bringing Structure to Cell Biology with Cryo-Electron Tomography

Recent advances in cryo-electron microscopy have marked only the beginning of the potential of this technique. To bring structure into cell biology, the modality of cryo-electron tomography has fast developed into a bona fide in situ structural biology technique where structures are determined in their native environment, the cell. Nearly every step of the cryo-focused ion beam-assisted electron tomography (cryo-FIB-ET) workflow has been improved upon in the past decade, since the first windows were carved into cells, unveiling macromolecular networks in near-native conditions. By bridging structural and cell biology, cryo-FIB-ET is advancing our understanding of structure-function relationships in their native environment and becoming a tool for discovering new biology.

Ot2Rec: A semi-automatic, extensible, multi-software tomographic reconstruction workflow

Electron cryo-tomography is an imaging technique for probing 3D structures with at the nanometer scale. This technique has been used extensively in the biomedical field to study the complex structures of proteins and other macromolecules. With the advancement in technology, microscopes are currently capable of producing images amounting to terabytes of data per day, posing great challenges for scientists as the speed of processing of the images cannot keep up with the ever-higher throughput of the microscopes. Therefore, automation is an essential and natural pathway on which image processing-from individual micrographs to full tomograms-is developing. In this paper, we present Ot2Rec, an open-source pipelining tool which aims to enable scientists to build their own processing workflows in a flexible and automatic manner. The basic building blocks of Ot2Rec are plugins which follow a unified application programming interface structure, making it simple for scientists to contribute to Ot2Rec by adding features which are not already available. In this paper, we also present three case studies of image processing using Ot2Rec, through which we demonstrate the speedup of using a semi-automatic workflow over a manual one, the possibility of writing and using custom (prototype) plugins, and the flexibility of Ot2Rec which enables the mix-and-match of plugins. We also demonstrate, in the Supplementary Material, a built-in reporting feature in Ot2Rec which aggregates the metadata from all process being run, and output them in the Jupyter Notebook and/or HTML formats for quick review of image processing quality. Ot2Rec can be found at https://github.com/rosalindfranklininstitute/ot2rec.

Polyribosomes of circular topology are prevalent in mammalian cells

Polyribosomes, the groups of ribosomes simultaneously translating a single mRNA molecule, are very common in both, prokaryotic and eukaryotic cells. Even in early EM studies, polyribosomes have been shown to possess various spatial conformations, including a ring-shaped configuration which was considered to be functionally important. However, a recent in situ cryo-ET analysis of predominant regular inter-ribosome contacts did not confirm the abundance of ring-shaped polyribosomes in a cell cytoplasm. To address this discrepancy, here we analyzed the cryo-ET structure of polyribosomes in diluted lysates of HeLa cells. It was shown that the vast majority of the ribosomes were combined into polysomes and were proven to be translationally active. Tomogram analysis revealed that circular polyribosomes are indeed very common in the cytoplasm, but they mostly possess pseudo-regular structures without specific inter-ribosomal contacts. Although the size of polyribosomes varied widely, most circular polysomes were relatively small in size (4-8 ribosomes). Our results confirm the recent data that it is cellular mRNAs with short ORF that most commonly form circular structures providing an enhancement of translation.

Preparation of hybrid samples for scanning electron microscopy (SEM) coupled to focused ion beam (FIB) analysis: A new way to study cell adhesion to titanium implant surfaces

The study of the intimate connection occurring at the interface between cells and titanium implant surfaces is a major challenge for dental materials scientists. Indeed, several imaging techniques have been developed and optimized in the last decades, but an optimal method has not been described yet. The combination of the scanning electron microscopy (SEM) with a focused ion beam (FIB), represents a pioneering and interesting tool to allow the investigation of the relationship occurring at the interface between cells and biomaterials, including titanium. However, major caveats concerning the nature of the biological structures, which are not conductive materials, and the physico-chemical properties of titanium (i.e. color, surface topography), require a fine and accurate preparation of the sample before its imaging. Hence, the aim of the present work is to provide a suitable protocol for cell-titanium sample preparation before imaging by SEM-FIB. The concepts presented in this paper are also transferrable to other fields of biomaterials research.

Quantitative Cryo-Electron Tomography

The three-dimensional organization of biomolecules important for the functioning of all living systems can be determined by cryo-electron tomography imaging under native biological contexts. Cryo-electron tomography is continually expanding and evolving, and the development of new methods that use the latest technology for sample thinning is enabling the visualization of ever larger and more complex biological systems, allowing imaging across scales. Quantitative cryo-electron tomography possesses the capability of visualizing the impact of molecular and environmental perturbations in subcellular structure and function to understand fundamental biological processes. This review provides an overview of current hardware and software developments that allow quantitative cryo-electron tomography studies and their limitations and how overcoming them may allow us to unleash the full power of cryo-electron tomography.

Better, Faster, Cheaper: Recent Advances in Cryo-Electron Microscopy

Cryo-electron microscopy (cryo-EM) continues its remarkable growth as a method for visualizing biological objects, which has been driven by advances across the entire pipeline. Developments in both single-particle analysis and in situ tomography have enabled more structures to be imaged and determined to better resolutions, at faster speeds, and with more scientists having improved access. This review highlights recent advances at each stageof the cryo-EM pipeline and provides examples of how these techniques have been used to investigate real-world problems, including antibody development against the SARS-CoV-2 spike during the recent COVID-19 pandemic.

Capturing actin assemblies in cells using in situ cryo-electron tomography

Actin contributes to an exceptionally wide range of cellular processes through the assembly and disassembly of highly dynamic and ordered structures. Visualizing these structures in cells can help us understand how the molecular players of the actin machinery work together to produce force-generating systems. In recent years, cryo-electron tomography (cryo-ET) has become the method of choice for structural analysis of the cell interior at the molecular scale. Here we review advances in cryo-ET workflows that have enabled this transformation, especially the automation of sample preparation procedures, data collection, and processing. We discuss new structural analyses of dynamic actin assemblies in cryo-preserved cells, which have provided mechanistic insights into actin assembly and function at the nanoscale. Finally, we highlight the latest visual proteomics studies of actin filaments and their interactors reaching sub-nanometer resolutions in cells.

Validation analysis of EMDB entries

The Electron Microscopy Data Bank (EMDB) is the central archive of the electron cryo-microscopy (cryo-EM) community for storing and disseminating volume maps and tomograms. With input from the community, EMDB has developed new resources for the validation of cryo-EM structures, focusing on the quality of the volume data alone and that of the fit of any models, themselves archived in the Protein Data Bank (PDB), to the volume data. Based on recommendations from community experts, the validation resources are developed in a three-tiered system. Tier 1 covers an extensive and evolving set of validation metrics, including tried and tested metrics as well as more experimental ones, which are calculated for all EMDB entries and presented in the Validation Analysis (VA) web resource. This system is particularly useful for cryo-EM experts, both to validate individual structures and to assess the utility of new validation metrics. Tier 2 comprises a subset of the validation metrics covered by the VA resource that have been subjected to extensive testing and are considered to be useful for specialists as well as nonspecialists. These metrics are presented on the entry-specific web pages for the entire archive on the EMDB website. As more experience is gained with the metrics included in the VA resource, it is expected that consensus will emerge in the community regarding a subset that is suitable for inclusion in the tier 2 system. Tier 3, finally, consists of the validation reports and servers that are produced by the Worldwide Protein Data Bank (wwPDB) Consortium. Successful metrics from tier 2 will be proposed for inclusion in the wwPDB validation pipeline and reports. The details of the new resource are described, with an emphasis on the tier 1 system. The output of all three tiers is publicly available, either through the EMDB website (tiers 1 and 2) or through the wwPDB ftp sites (tier 3), although the content of all three will evolve over time (fastest for tier 1 and slowest for tier 3). It is our hope that these validation resources will help the cryo-EM community to obtain a better understanding of the quality and of the best ways to assess the quality of cryo-EM structures in EMDB and PDB.

Cryogenic electron microscopy approaches that combine images and tilt series

Cryogenic electron microscopy can be widely applied to biological specimens from the molecular to the cellular scale. In single-particle analysis, 3D structures may be obtained in high resolution by averaging 2D images of single particles in random orientations. For pleomorphic specimens, structures may be obtained by recording the tilt series of a single example of the specimen and calculating tomograms. Where many copies of a single structure such as a protein or nucleic acid assembly are present within the tomogram, averaging of the sub-volumes (subtomogram averaging) has been successfully applied. The choice of data collection method for any given specimen may depend on the structural question of interest and is determined by the radiation sensitivity of the specimen. Here, we survey some recent developments on the use of hybrid methods for recording and analysing data from radiation-sensitive biological specimens. These include single-particle reconstruction from 2D images where additional views are recorded at a single tilt angle of the specimen and methods where image tilt series, initially used for tomogram reconstruction, are processed as individual single-particle images. There is a continuum of approaches now available to maximize structural information obtained from the specimen.

Electron microscopy for imaging organelles in plants and algae

Recent developments in both instrumentation and image analysis algorithms have allowed three-dimensional electron microscopy (3D-EM) to increase automated image collections through large tissue volumes using serial block-face scanning EM (SEM) and to achieve near-atomic resolution of macromolecular complexes using cryo-electron tomography (cryo-ET) and sub-tomogram averaging. In this review, we discuss applications of cryo-ET to cell biology research on plant and algal systems and the special opportunities they offer for understanding the organization of eukaryotic organelles with unprecedently resolution. However, one of the most challenging aspects for cryo-ET is sample preparation, especially for multicellular organisms. We also discuss correlative light and electron microscopy (CLEM) approaches that have been developed for ET at both room and cryogenic temperatures.

Super-Resolution Microscopy for Structural Cell Biology

Super-resolution microscopy techniques, and specifically single-molecule localization microscopy (SMLM), are approaching nanometer resolution inside cells and thus have great potential to complement structural biology techniques such as electron microscopy for structural cell biology. In this review, we introduce the different flavors of super-resolution microscopy, with a special emphasis on SMLM and MINFLUX (minimal photon flux). We summarize recent technical developments that pushed these localization-based techniques to structural scales and review the experimental conditions that are key to obtaining data of the highest quality. Furthermore, we give an overview of different analysis methods and highlight studies that used SMLM to gain structural insights into biologically relevant molecular machines. Ultimately, we give our perspective on what is needed to push the resolution of these techniques even further and to apply them to investigating dynamic structural rearrangements in living cells. Expected final online publication date for the Annual Review of Biophysics, Volume 51 is May 2022. Please see http://www.annualreviews.org/page/journal/pubdates for revised estimates.

Subtomogram analysis: The sum of a tomogram's particles reveals molecular structure in situ

Cryo-electron tomography is uniquely suited to provide insights into the molecular architecture of cells and tissue in the native state. While frozen hydrated specimens tolerate sufficient electron doses to distinguish different types of particles in a tomogram, the accumulating beam damage does not allow resolving their detailed molecular structure individually. Statistical methods for subtomogram averaging and classification that coherently enhance the signal of particles corresponding to copies of the same type of macromolecular allow obtaining much higher resolution insights into macromolecules. Here, I review the developments in subtomogram analysis at Wolfgang Baumeister's laboratory that make the dream of structural biology in the native cell become reality.

Cryo-shift: reducing domain shift in cryo-electron subtomograms with unsupervised domain adaptation and randomization

MOTIVATION

Cryo-Electron Tomography (cryo-ET) is a 3D imaging technology that enables the visualization of subcellular structures in situ at near-atomic resolution. Cellular cryo-ET images help in resolving the structures of macromolecules and determining their spatial relationship in a single cell, which has broad significance in cell and structural biology. Subtomogram classification and recognition constitute a primary step in the systematic recovery of these macromolecular structures. Supervised deep learning methods have been proven to be highly accurate and efficient for subtomogram classification, but suffer from limited applicability due to scarcity of annotated data. While generating simulated data for training supervised models is a potential solution, a sizeable difference in the image intensity distribution in generated data as compared with real experimental data will cause the trained models to perform poorly in predicting classes on real subtomograms.

RESULTS

In this work, we present Cryo-Shift, a fully unsupervised domain adaptation and randomization framework for deep learning-based cross-domain subtomogram classification. We use unsupervised multi-adversarial domain adaption to reduce the domain shift between features of simulated and experimental data. We develop a network-driven domain randomization procedure with 'warp' modules to alter the simulated data and help the classifier generalize better on experimental data. We do not use any labeled experimental data to train our model, whereas some of the existing alternative approaches require labeled experimental samples for cross-domain classification. Nevertheless, Cryo-Shift outperforms the existing alternative approaches in cross-domain subtomogram classification in extensive evaluation studies demonstrated herein using both simulated and experimental data.

AVAILABILITYAND IMPLEMENTATION

https://github.com/xulabs/aitom.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Into the matrix: current methods for mitochondrial translation studies

In addition to the cytoplasmic translation system, eukaryotic cells house additional protein synthesis machinery in mitochondria. The importance of this in organello translation is exemplified by clinical pathologies associated with mutations in mitochondrial translation factors. Although a detailed understanding of mitochondrial translation has long been awaited, quantitative, comprehensive, and spatiotemporal measurements have posed analytic challenges. The recent development of novel approaches for studying mitochondrial protein synthesis has overcome these issues and expands our understanding of the unique translation system. Here, we review the current technologies for the investigation of mitochondrial translation and the insights provided by their application.

Compressed sensing for electron cryotomography and high-resolution subtomogram averaging of biological specimens

Cryoelectron tomography (cryo-ET) and subtomogram averaging (STA) allow direct visualization and structural studies of biological macromolecules in their native cellular environment, in situ. Often, low signal-to-noise ratios in tomograms, low particle abundance within the cell, and low throughput in typical cryo-ET workflows severely limit the obtainable structural information. To help mitigate these limitations, here we apply a compressed sensing approach using 3D second-order total variation (CS-TV²) to tomographic reconstruction. We show that CS-TV² increases the signal-to-noise ratio in tomograms, enhancing direct visualization of macromolecules, while preserving high-resolution information up to the secondary structure level. We show that, particularly with small datasets, CS-TV² allows improvement of the resolution of STA maps. We further demonstrate that the CS-TV² algorithm is applicable to cellular specimens, leading to increased visibility of molecular detail within tomograms. This work highlights the potential of compressed sensing-based reconstruction algorithms for cryo-ET and in situ structural biology.

•

Compressed sensing (CS-TV²) for cryo-ET using 3D second-order total variation

•

CS-TV² increases signal contrast while retaining high-resolution information

•

Improved subtomogram averaging from CS-TV² reconstructions of small datasets

•

Increased contrast and detail in CS-TV² reconstructions of cellular specimens

Entropy-regularized deconvolution of cellular cryotransmission electron tomograms

Cryo-electron tomography (cryo-ET) allows for the high-resolution visualization of biological macromolecules. However, the technique is limited by a low signal-to-noise ratio (SNR) and variance in contrast at different frequencies, as well as reduced Z resolution. Here, we applied entropy-regularized deconvolution (ER-DC) to cryo-ET data generated from transmission electron microscopy (TEM) and reconstructed using weighted back projection (WBP). We applied deconvolution to several in situ cryo-ET datasets and assessed the results by Fourier analysis and subtomogram analysis (STA).

Imaging Techniques for Detecting Prokaryotic Viruses in Environmental Samples

Viruses are the most abundant biological entities on Earth with an estimate of 1031 viral particles across all ecosystems. Prokaryotic viruses-bacteriophages and archaeal viruses-influence global biogeochemical cycles by shaping microbial communities through predation, through the effect of horizontal gene transfer on the host genome evolution, and through manipulating the host cellular metabolism. Imaging techniques have played an important role in understanding the biology and lifestyle of prokaryotic viruses. Specifically, structure-resolving microscopy methods, for example, transmission electron microscopy, are commonly used for understanding viral morphology, ultrastructure, and host interaction. These methods have been applied mostly to cultivated phage-host pairs. However, recent advances in environmental genomics have demonstrated that the majority of viruses remain uncultivated, and thus microscopically uncharacterized. Although light- and structure-resolving microscopy of viruses from environmental samples is possible, quite often the link between the visualization and the genomic information of uncultivated prokaryotic viruses is missing. In this minireview, we summarize the current state of the art of imaging techniques available for characterizing viruses in environmental samples and discuss potential links between viral imaging and environmental genomics for shedding light on the morphology of uncultivated viruses and their lifestyles in Earth's ecosystems.

End-to-end robust joint unsupervised image alignment and clustering

Computing dense pixel-to-pixel image correspondences is a fundamental task of computer vision. Often, the objective is to align image pairs from the same semantic category for manipulation or segmentation purposes. Despite achieving superior performance, existing deep learning alignment methods cannot cluster images; consequently, clustering and pairing images needed to be a separate laborious and expensive step. Given a dataset with diverse semantic categories, we propose a multi-task model, Jim-Net, that can directly learn to cluster and align images without any pixel-level or image-level annotations. We design a pair-matching alignment unsupervised training algorithm that selectively matches and aligns image pairs from the clustering branch. Our unsupervised Jim-Net achieves comparable accuracy with state-of-the-art supervised methods on benchmark 2D image alignment dataset PF-PASCAL. Specifically, we apply Jim-Net to cryo-electron tomography, a revolutionary 3D microscopy imaging technique of native subcellular structures. After extensive evaluation on seven datasets, we demonstrate that Jim-Net enables systematic discovery and recovery of representative macromolecular structures in situ, which is essential for revealing molecular mechanisms underlying cellular functions. To our knowledge, Jim-Net is the first end-to-end model that can simultaneously align and cluster images, which significantly improves the performance as compared to performing each task alone.

NUDIM: A non-uniform fast Fourier transform based dual-space constraint iterative reconstruction method in biological electron tomography

Electron tomography, a powerful imaging tool for studying 3D structures of macromolecular assemblies, always suffers from imperfect reconstruction with limited resolution due to the intrinsic low signal-to-noise ratio (SNR) and inaccessibility to certain tilt angles induced by radiation damage or mechanical limitation. In order to compensate for such insufficient data with low SNR and further improve imaging resolution, prior knowledge constraints about the objects in both real space and reciprocal space are thus exploited during tomographic reconstruction. However, direct Fast Fourier transform (FFT) between real space and reciprocal space remains extraordinarily challenging owing to their inconsistent grid sampling modes, e.g. regular and uniform grid sampling in real space whereas radial or polar grid sampling in reciprocal space. In order to solve such problem, a technique of non-uniform fast Fourier transform (NFFT) has been developed to transform efficiently between non-uniformly sampled grids in real and reciprocal space with sufficient accuracy. In this work, a Non-Uniform fast Fourier transform based Dual-space constraint Iterative reconstruction Method (NUDIM) applicable to biological electron tomography is proposed with a combination of basic concepts from equally sloped tomography (EST) and NFFT based reconstruction. In NUDIM, the use of NFFT can circumvent such grid sampling inconsistency and thus alleviate the stringent equally-sloped sampling requirement in EST reconstruction, while the dual-space constraint iterative procedure can dramatically enhance reconstruction quality. In comparison with conventional reconstruction methods, NUDIM is numerically and experimentally demonstrated to produce superior reconstruction quality with higher contrast, less noise and reduced missing wedge artifacts. More importantly, it is also capable of retrieving part of missing information from a limited number of projections.

A novel constrained reconstruction model towards high-resolution subtomogram averaging

MOTIVATION

Electron tomography (ET) offers a unique capacity to image biological structures in situ. However, the resolution of ET reconstructed tomograms is not comparable to that of the single-particle cryo-EM. If many copies of the object of interest are present in the tomograms, their structures can be reconstructed in the tomogram, picked, aligned and averaged to increase the signal-to-noise ratio and improve the resolution, which is known as the subtomogram averaging. To date, the resolution improvement of the subtomogram averaging is still limited because each reconstructed subtomogram is of low reconstruction quality due to the missing wedge issue.

RESULTS

In this article, we propose a novel computational model, the constrained reconstruction model (CRM), to better recover the information from the multiple subtomograms and compensate for the missing wedge issue in each of them. CRM is supposed to produce a refined reconstruction in the final turn of subtomogram averaging after alignment, instead of directly taking the average. We first formulate the averaging method and our CRM as linear systems, and prove that the solution space of CRM is no larger, and in practice much smaller, than that of the averaging method. We then propose a sparse Kaczmarz algorithm to solve the formulated CRM, and further extend the solution to the simultaneous algebraic reconstruction technique (SART). Experimental results demonstrate that CRM can significantly alleviate the missing wedge issue and improve the final reconstruction quality. In addition, our model is robust to the number of images in each tilt series, the tilt range and the noise level.

AVAILABILITY AND IMPLEMENTATION

The codes of CRM-SIRT and CRM-SART are available at https://github.com/icthrm/CRM.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Molecular-scale visualization of sarcomere contraction within native cardiomyocytes

Sarcomeres, the basic contractile units of striated muscle, produce the forces driving muscular contraction through cross-bridge interactions between actin-containing thin filaments and myosin II-based thick filaments. Until now, direct visualization of the molecular architecture underlying sarcomere contractility has remained elusive. Here, we use in situ cryo-electron tomography to unveil sarcomere contraction in frozen-hydrated neonatal rat cardiomyocytes. We show that the hexagonal lattice of the thick filaments is already established at the neonatal stage, with an excess of thin filaments outside the trigonal positions. Structural assessment of actin polarity by subtomogram averaging reveals that thin filaments in the fully activated state form overlapping arrays of opposite polarity in the center of the sarcomere. Our approach provides direct evidence for thin filament sliding during muscle contraction and may serve as a basis for structural understanding of thin filament activation and actomyosin interactions inside unperturbed cellular environments.

An overview of the recent advances in cryo-electron microscopy for life sciences

Cryo-electron microscopy (CryoEM) has superseded X-ray crystallography and NMR to emerge as a popular and effective tool for structure determination in recent times. It has become indispensable for the characterization of large macromolecular assemblies, membrane proteins, or samples that are limited, conformationally heterogeneous, and recalcitrant to crystallization. Besides, it is the only tool capable of elucidating high-resolution structures of macromolecules and biological assemblies in situ. A state-of-the-art electron microscope operable at cryo-temperature helps preserve high-resolution details of the biological sample. The structures can be determined, either in isolation via single-particle analysis (SPA) or helical reconstruction, electron diffraction (ED) or within the cellular environment via cryo-electron tomography (cryoET). All the three streams of SPA, ED, and cryoET (along with subtomogram averaging) have undergone significant advancements in recent times. This has resulted in breaking the boundaries with respect to both the size of the macromolecules/assemblies whose structures could be determined along with the visualization of atomic details at resolutions unprecedented for cryoEM. In addition, the collection of larger datasets combined with the ability to sort and process multiple conformational states from the same sample are providing the much-needed link between the protein structures and their functions. In overview, these developments are helping scientists decipher the molecular mechanism of critical cellular processes, solve structures of macromolecules that were challenging targets for structure determination until now, propelling forward the fields of biology and biomedicine. Here, we summarize recent advances and key contributions of the three cryo-electron microscopy streams of SPA, ED, and cryoET.

Few shot domain adaptation for in situ macromolecule structural classification in cryoelectron tomograms

MOTIVATION

Cryoelectron tomography (cryo-ET) visualizes structure and spatial organization of macromolecules and their interactions with other subcellular components inside single cells in the close-to-native state at submolecular resolution. Such information is critical for the accurate understanding of cellular processes. However, subtomogram classification remains one of the major challenges for the systematic recognition and recovery of the macromolecule structures in cryo-ET because of imaging limits and data quantity. Recently, deep learning has significantly improved the throughput and accuracy of large-scale subtomogram classification. However, often it is difficult to get enough high-quality annotated subtomogram data for supervised training due to the enormous expense of labeling. To tackle this problem, it is beneficial to utilize another already annotated dataset to assist the training process. However, due to the discrepancy of image intensity distribution between source domain and target domain, the model trained on subtomograms in source domain may perform poorly in predicting subtomogram classes in the target domain.

RESULTS

In this article, we adapt a few shot domain adaptation method for deep learning-based cross-domain subtomogram classification. The essential idea of our method consists of two parts: (i) take full advantage of the distribution of plentiful unlabeled target domain data, and (ii) exploit the correlation between the whole source domain dataset and few labeled target domain data. Experiments conducted on simulated and real datasets show that our method achieves significant improvement on cross domain subtomogram classification compared with baseline methods.

AVAILABILITY AND IMPLEMENTATION

Software is available online https://github.com/xulabs/aitom.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Membrane imaging in the plant endomembrane system

Recent studies on membrane imaging in the plant endomembrane system by 2-D/3-D CLSM and TEM provide future perspectives of whole-cell ET and cryo-FIB-aided cryo-ET analysis.

Physics of viral dynamics

Viral capsids are often regarded as inert structural units, but in actuality they display fascinating dynamics during different stages of their life cycle. With the advent of single-particle approaches and high-resolution techniques, it is now possible to scrutinize viral dynamics during and after their assembly and during the subsequent development pathway into infectious viruses. In this Review, the focus is on the dynamical properties of viruses, the different physical virology techniques that are being used to study them, and the physical concepts that have been developed to describe viral dynamics.

In Situ Imaging and Structure Determination of Biomolecular Complexes Using Electron Cryo-Tomography

Electron cryo-tomography (cryo-ET) is a technique that allows the investigation of intact macromolecular complexes while they are in their cellular milieu. Over the years, cryo-ET has had a huge impact on our understanding of how large biomolecular complexes look like, how they assemble, disassemble, function, and evolve(d). Recent hardware and software developments and combining cryo-ET with other techniques, e.g., focused ion beam milling (FIB-milling) and cryo-light microscopy, has extended the realm of cryo-ET to include transient molecular complexes embedded deep in thick samples (like eukaryotic cells) and enhanced the resolution of structures obtained by cryo-ET. In this chapter, we will present an outline of how to perform cryo-ET studies on a wide variety of biological samples including prokaryotic and eukaryotic cells and biological plant tissues. This outline will include sample preparation, data collection, and data processing as well as hybrid approaches like FIB-milling, cryosectioning, and cryo-correlated light and electron microscopy (cryo-CLEM).

Fake It 'Till You Make It-The Pursuit of Suitable Membrane Mimetics for Membrane Protein Biophysics

Membrane proteins evolved to reside in the hydrophobic lipid bilayers of cellular membranes. Therefore, membrane proteins bridge the different aqueous compartments separated by the membrane, and furthermore, dynamically interact with their surrounding lipid environment. The latter not only stabilizes membrane proteins, but directly impacts their folding, structure and function. In order to be characterized with biophysical and structural biological methods, membrane proteins are typically extracted and subsequently purified from their native lipid environment. This approach requires that lipid membranes are replaced by suitable surrogates, which ideally closely mimic the native bilayer, in order to maintain the membrane proteins structural and functional integrity. In this review, we survey the currently available membrane mimetic environments ranging from detergent micelles to bicelles, nanodiscs, lipidic-cubic phase (LCP), liposomes, and polymersomes. We discuss their respective advantages and disadvantages as well as their suitability for downstream biophysical and structural characterization. Finally, we take a look at ongoing methodological developments, which aim for direct in-situ characterization of membrane proteins within native membranes instead of relying on membrane mimetics.

Molecular Logic of Prokaryotic Surface Layer Structures

Most prokaryotic cells are encased in a surface layer (S-layer) consisting of a paracrystalline array of repeating lattice-forming proteins. S-layer proteins populate a vast and diverse sequence space, performing disparate functions in prokaryotic cells, including cellular defense, cell-shape maintenance, and regulation of import and export of materials. This article highlights recent advances in the understanding of S-layer structure and assembly, made possible by rapidly evolving structural and cell biology methods. We underscore shared assembly principles revealed by recent work and discuss a common molecular framework that may be used to understand the structural organization of S-layer proteins across bacteria and archaea.

Despite enormous sequence diversity in surface (S)-layer proteins, structural diversity is much lower than previously thought.

S-layer proteins have a bipartite arrangement with a lattice-forming and an anchoring segment.

Novel structural biology methods are revealing the architectures of S-layers in situ.

S-layer assembly across prokaryotes is tightly coupled to the cell cycle, including the cell division machinery.

Three-dimensional deconvolution processing for STEM cryotomography

The complex environment of biological cells and tissues has motivated development of three-dimensional (3D) imaging in both light and electron microscopies. To this end, one of the primary tools in fluorescence microscopy is that of computational deconvolution. Wide-field fluorescence images are often corrupted by haze due to out-of-focus light, i.e., to cross-talk between different object planes as represented in the 3D image. Using prior understanding of the image formation mechanism, it is possible to suppress the cross-talk and reassign the unfocused light to its proper source post facto. Electron tomography based on tilted projections also exhibits a cross-talk between distant planes due to the discrete angular sampling and limited tilt range. By use of a suitably synthesized 3D point spread function, we show here that deconvolution leads to similar improvements in volume data reconstructed from cryoscanning transmission electron tomography (CSTET), namely a dramatic in-plane noise reduction and improved representation of features in the axial dimension. Contrast enhancement is demonstrated first with colloidal gold particles and then in representative cryotomograms of intact cells. Deconvolution of CSTET data collected from the periphery of an intact nucleus revealed partially condensed, extended structures in interphase chromatin.

Cellular Electron Cryo-Tomography to Study Virus-Host Interactions

Viruses are obligatory intracellular parasites that reprogram host cells upon infection to produce viral progeny. Here, we review recent structural insights into virus-host interactions in bacteria, archaea, and eukaryotes unveiled by cellular electron cryo-tomography (cryoET). This advanced three-dimensional imaging technique of vitreous samples in near-native state has matured over the past two decades and proven powerful in revealing molecular mechanisms underlying viral replication. Initial studies were restricted to cell peripheries and typically focused on early infection steps, analyzing surface proteins and viral entry. Recent developments including cryo-thinning techniques, phase-plate imaging, and correlative approaches have been instrumental in also targeting rare events inside infected cells. When combined with advances in dedicated image analyses and processing methods, details of virus assembly and egress at (sub)nanometer resolution were uncovered. Altogether, we provide a historical and technical perspective and discuss future directions and impacts of cryoET for integrative structural cell biology analyses of viruses.

Electron cryotomography of intact motile cilia defines the basal body to axoneme transition

Greenan et al. use electron cryotomography on intact motile cilia to elucidate how basal bodies template the formation of motile axonemes.

Cells use motile cilia to generate force in the extracellular space. The structure of a cilium can be classified into three subdomains: the intracellular basal body (BB) that templates cilium formation, the extracellular axoneme that generates force, and the transition zone (TZ) that bridges them. While the BB is composed of triplet microtubules (TMTs), the axoneme is composed of doublet microtubules (DMTs), meaning the cilium must convert between different microtubule geometries. Here, we performed electron cryotomography to define this conversion, and our reconstructions reveal identifying structural features of the BB, TZ, and axoneme. Each region is distinct in terms of microtubule number and geometry, microtubule inner proteins, and microtubule linkers. TMT to DMT conversion occurs within the BB, and microtubule geometry changes to axonemal by the end of the TZ, followed by the addition of axoneme-specific components essential for cilium motility. Our results provide the highest-resolution images of the motile cilium to date and reveal how BBs template axonemes.

Towards high-throughput in situ structural biology using electron cryotomography

Electron cryotomography is a rapidly evolving method for imaging macromolecules directly within the native environment of cells and tissues. Combined with sub-tomogram averaging, it allows structural and cell biologists to obtain sub-nanometre resolution structures in situ. However, low throughput in cryo-ET sample preparation and data acquisition, as well as difficulties in target localisation and sub-tomogram averaging image processing, limit its widespread usability. In this review, we discuss new advances in the field that address these throughput and technical problems. We focus on recent efforts made to resolve issues in sample thinning, improvement in data collection speed at the microscope, strategies for localisation of macromolecules using correlated light and electron microscopy and advancements made to improve resolution in sub-tomogram averaging. These advances will considerably decrease the amount of time and effort required for cryo-ET and sub-tomogram averaging, ushering in a new era of structural biology where in situ macromolecular structure determination will be routine.

Gum-Net: Unsupervised Geometric Matching for Fast and Accurate 3D Subtomogram Image Alignment and Averaging

We propose a Geometric unsupervised matching Network (Gum-Net) for finding the geometric correspondence between two images with application to 3D subtomogram alignment and averaging. Subtomogram alignment is the most important task in cryo-electron tomography (cryo-ET), a revolutionary 3D imaging technique for visualizing the molecular organization of unperturbed cellular landscapes in single cells. However, subtomogram alignment and averaging are very challenging due to severe imaging limits such as noise and missing wedge effects. We introduce an end-to-end trainable architecture with three novel modules specifically designed for preserving feature spatial information and propagating feature matching information. The training is performed in a fully unsupervised fashion to optimize a matching metric. No ground truth transformation information nor category-level or instance-level matching supervision information is needed. After systematic assessments on six real and nine simulated datasets, we demonstrate that Gum-Net reduced the alignment error by 40 to 50% and improved the averaging resolution by 10%. Gum-Net also achieved 70 to 110 times speedup in practice with GPU acceleration compared to state-of-the-art subtomogram alignment methods. Our work is the first 3D unsupervised geometric matching method for images of strong transformation variation and high noise level. The training code, trained model, and datasets are available in our open-source software AITom.

Preparing samples from whole cells using focused-ion-beam milling for cryo-electron tomography

Recent advances have made cryogenic (cryo) electron microscopy a key technique to achieve near-atomic-resolution structures of biochemically isolated macromolecular complexes. Cryo-electron tomography (cryo-ET) can give unprecedented insight into these complexes in the context of their natural environment. However, the application of cryo-ET is limited to samples that are thinner than most cells, thereby considerably reducing its applicability. Cryo-focused-ion-beam (cryo-FIB) milling has been used to carve (micromachining) out 100-250-nm-thin regions (called lamella) in the intact frozen cells. This procedure opens a window into the cells for high-resolution cryo-ET and structure determination of biomolecules in their native environment. Further combination with fluorescence microscopy allows users to determine cells or regions of interest for the targeted fabrication of lamellae and cryo-ET imaging. Here, we describe how to prepare lamellae using a microscope equipped with both FIB and scanning electron microscopy modalities. Such microscopes (Aquilos Cryo-FIB/Scios/Helios or CrossBeam) are routinely referred to as dual-beam microscopes, and they are equipped with a cryo-stage for all operations in cryogenic conditions. The basic principle of the described methodologies is also applicable for other types of dual-beam microscopes equipped with a cryo-stage. We also briefly describe how to integrate fluorescence microscopy data for targeted milling and critical considerations for cryo-ET data acquisition of the lamellae. Users familiar with cryo-electron microscopy who get basic training in dual-beam microscopy can complete the protocol within 2-3 d, allowing for several pause points during the procedure.

High-Resolution In Situ NMR Spectroscopy of Bacterial Envelope Proteins in Outer Membrane Vesicles

The cell envelope of Gram-negative bacteria is an elaborate cellular environment, consisting of two lipid membranes separated by the aqueous periplasm. So far, efforts to mimic this environment under laboratory conditions have been limited by the complexity of the asymmetric bacterial outer membrane. To evade this impasse, we recently established a method to modify the protein composition of bacterial outer membrane vesicles (OMVs) released from Escherichia coli as a platform for biophysical studies of outer membrane proteins in their native membrane environment. Here, we apply protein-enriched OMVs to characterize the structure of three envelope proteins from E. coli using nuclear magnetic resonance (NMR) spectroscopy and expand the methodology to soluble periplasmic proteins. We obtain high-resolution in situ NMR spectra of the transmembrane protein OmpA as well as the periplasmic proteins CpxP and MalE. We find that our approach facilitates structural investigations of membrane-attached protein domains and is especially suited for soluble proteins within their native periplasmic environment. Thereby, the use of OMVs in solution NMR methods allows in situ analysis of the structure and dynamics of proteins twice the size compared to the current in-cell NMR methodology. We therefore expect our work to pave the way for more complex NMR studies of bacterial envelope proteins in the native environment of OMVs in the future.