Cryo-electron tomography of cells: connecting structure and function

Structural insight into transmissive mutant huntingtin species by correlative light and electron microscopy and cryo-electron tomography

Vitrification of Tokuyasu-style immuno-labelled sections for correlative cryo light microscopy and cryo electron tomography

Advances in cryogenic electron microscopy (cryoEM) single particle analysis have revolutionized structural biology by facilitating the in vitro determination of atomic- and near-atomic-resolution structures for fully hydrated macromolecular complexes exhibiting compositional and conformational heterogeneity across a wide range of sizes. Cryogenic electron tomography (cryoET) and subtomogram averaging are rapidly progressing toward delivering similar insights for macromolecular complexes in situ, without requiring tags or harsh biochemical purification. Furthermore, cryoET enables the visualization of cellular and tissue phenotypes directly at molecular, nanometric resolution without chemical fixation or staining artifacts. This forward-looking review covers recent developments in cryoEM/ET and related technologies such as cryogenic focused ion beam milling scanning electron microscopy and correlative light microscopy, increasingly enhanced and supported by artificial intelligence algorithms. Their potential application to emerging concepts is discussed, primarily the prospect of complementing medical histopathology analysis. Machine learning solutions are poised to address current challenges posed by "big data" in cryoET of tissues, cells, and macromolecules, offering the promise of enabling novel, quantitative insights into disease processes, which may translate into the clinic and lead to improved diagnostics and targeted therapeutics.

Volume electron microscopy (Volume-EM) has transformed structural cell biology by enabling nanometre-resolution imaging across cellular and tissue scales. Serial-section TEM, Serial Block-Face Scanning Electron Microscopy (SBF-SEM), Focused Ion Beam Scanning Electron Microscopy (FIB-SEM) and multi-beam SEM now routinely generate terabyte-scale volumes that capture organelles, synapses and neural circuits in three dimensions, while cryogenic Volume-EM extends this landscape by preserving vitrified, fully hydrated specimens in a near-native state. Together, these room-temperature and cryogenic modalities define a continuum of approaches that trade off volume, resolution, throughput and structural fidelity, and increasingly interface with correlative light microscopy and cryo-electron tomography. In parallel, advances in computation have turned Volume-EM into a data-intensive discipline. Multistage preprocessing pipelines for alignment, denoising, stitching and intensity normalisation feed into automated segmentation frameworks that combine convolutional neural networks, affinity-based supervoxel agglomeration, flood-filling networks and, more recently, diffusion-based generative restoration. Weakly supervised and self-supervised learning, multi-task objectives and human–AI co-training mitigate the scarcity of dense ground truth, while distributed storage and streaming inference architectures support segmentation and proofreading at the terascale and beyond. Open resources such as COSEM, MICRONS, OpenOrganelle and EMPIAR provide benchmark datasets, interoperable file formats and reference workflows that anchor method development and cross-laboratory comparison. In this review, we first outline the physical principles and imaging modes of conventional and cryogenic Volume-EM, then describe current best practices in data acquisition and preprocessing, and finally survey the emerging ecosystem of AI-driven segmentation and analysis. We highlight how cryo–Volume-EM expands the field towards native-state structural biology, and how multimodal integration with light microscopy, cryo-electron tomography (cryo-ET) and spatial omics is pushing Volume-EM from descriptive imaging towards predictive, mechanistic, cross-scale models of cell physiology, disease ultrastructure and neural circuit function.

Correlative light and electron microscopy (CLEM) is a powerful tool for investigating cellular structure and function at the molecular level. However, while electron microscopy is often performed to great advantage at cryogenic temperatures, this is not always the case for light microscopy. One key challenge is the lack of cryo-compatible immersion objectives. In recent years, multiple cryoimmersion light microscopy (cryo-iLM) approaches have been described, but these techniques have never been used in correlative approaches. Here we present a novel workflow for correlative cryoimmersion light microscopy and electron cryomicroscopy (cryo-iCLEM). Cryo-electron tomography conducted before and after cryo-iLM reveals that cryo-iCLEM maintains ultra-thin, electron-transparent samples mechanically intact and does not degrade the ultrastructural preservation achieved through plunge-freezing. For cryo-iLM, the sample is first embedded in a viscous immersion medium at cryogenic temperatures and examined with a custom cryo-immersion objective. After cryo-iLM, the immersion medium is dissolved in liquid ethane, allowing for subsequent cryo-EM imaging. We further show that cryo-iCLEM can be used on FIB-lamellae, demonstrating that mechanically sensitive samples remain undamaged. Embedding the sample in the immersion fluid reduces contamination and thus allows data acquisition over many hours. Samples can therefore be examined in detail with the advantage of low bleaching rates of fluorophores at cryogenic temperatures. In the future, we hope that our approach can help improve the performance of many advanced light microscopy techniques when they are applied in the context of cryo-CLEM.

Cryogenic electron tomography (cryoET) allows visualization of cellular structures in situ. However, anisotropic resolution arising from the intrinsic “missing-wedge” problem has presented major challenges in visualization and interpretation of tomograms. Here, we have developed IsoNet, a deep learning-based software package that iteratively reconstructs the missing-wedge information and increases signal-to-noise ratio, using the knowledge learned from raw tomograms. Without the need for sub-tomogram averaging, IsoNet generates tomograms with significantly reduced resolution anisotropy. Applications of IsoNet to three representative types of cryoET data demonstrate greatly improved structural interpretability: resolving lattice defects in immature HIV particles, establishing architecture of the paraflagellar rod in Eukaryotic flagella, and identifying heptagon-containing clathrin cages inside a neuronal synapse of cultured cells. Therefore, by overcoming two fundamental limitations of cryoET, IsoNet enables functional interpretation of cellular tomograms without sub-tomogram averaging. Its application to high-resolution cellular tomograms should also help identify differently oriented complexes of the same kind for sub-tomogram averaging.

Correlative Light and Electron MIcroscopy

The eukaryotic cell relies on complex, highly regulated, and functionally distinct membrane bound compartments that preserve a biochemical polarity necessary for proper cellular function. Understanding how the enzymes, proteins, and cytoskeletal components govern and maintain this biochemical segregation is therefore of paramount importance. The use of fluorescently tagged molecules to localize to and/or perturb subcellular compartments has yielded a wealth of knowledge and advanced our understanding of cellular regulation. Imaging techniques such as fluorescent and confocal microscopy make ascertaining the position of a fluorescently tagged small molecule relatively straightforward, however the resolution of very small structures is limited. On the other hand, electron microscopy has revealed details of subcellular morphology at very high resolution, but its static nature makes it difficult to measure highly dynamic processes with precision. Thus, the combination of light microscopy with electron microscopy of the same sample, termed Correlative Light and Electron Microscopy (CLEM), affords the dual advantages of ultrafast fluorescent imaging with the high-resolution of electron microscopy. This powerful technique has been implemented to study many aspects of cell biology. Since its inception, this procedure has increased our ability to distinguish subcellular architectures and morphologies at high resolution. Here, we present a streamlined method for performing rapid microinjection followed by CLEM (Fig. 1). The microinjection CLEM procedure can be used to introduce specific quantities of small molecules and/or proteins directly into the eukaryotic cell cytoplasm and study the effects from millimeter to multi-nanometer resolution (Fig. 2). The technique is based on microinjecting cells grown on laser etched glass gridded coverslips affixed to the bottom of live cell dishes and imaging with both confocal fluorescent and electron microscopy. Localization of the cell(s) of interest is facilitated by the grid pattern, which is easily transferred, along with the cells of interest, to the Epon resin used for immobilization of samples and sectioning prior to electron microscopy analysis (Fig. 3). Overlay of fluorescent and EM images allows the user to determine the subcellular localization as well as any morphological and/or ultrastructural changes induced by the microinjected molecule of interest (Fig. 4). This technique is amenable to time points ranging from ≤5 s up to several hours, depending on the nature of the microinjected sample.

Correlative light and electron microscopy allows features of interest defined by fluorescence signals to be located in an electron micrograph of the same sample. Rare dynamic events or specific objects can be identified, targeted and imaged by electron microscopy or tomography. To combine it with structural studies using cryo-electron microscopy or tomography, fluorescence microscopy must be performed while maintaining the specimen vitrified at liquid-nitrogen temperatures and in a dry environment during imaging and transfer. Here we present instrumentation, software and an experimental workflow that improves the ease of use, throughput and performance of correlated cryo-fluorescence and cryo-electron microscopy. The new cryo-stage incorporates a specially modified high-numerical aperture objective lens and provides a stable and clean imaging environment. It is combined with a transfer shuttle for contamination-free loading of the specimen. Optimized microscope control software allows automated acquisition of the entire specimen area by cryo-fluorescence microscopy. The software also facilitates direct transfer of the fluorescence image and associated coordinates to the cryo-electron microscope for subsequent fluorescence-guided automated imaging. Here we describe these technological developments and present a detailed workflow, which we applied for automated cryo-electron microscopy and tomography of various specimens.

Microanatomy Around the Facial Nerve Pathway for Microvascular Decompression Surgery Investigated with Correlative Light Microscopy and Block-Face Imaging

Correlative light and electron microscopy techniques are powerful tools that allow the identification and analysis of fluorescently labeled cells and structures of interest by an electron microscope (EM).1–3 Nevertheless, employing the methods listed above for the identification of rare cell populations can be very time consuming.4 A recently described method combines imaging of a fluorescently labeled tissue before fixation, with laser etching at the region of interest after embryo fixation.5 While the region of interest in the zebrafish embryo could be better focused on, this technique does not assist the identification of rare cell populations or specific cellular compartments when adjacent morphological landmarks are not present. In this study, we present a procedure based on a recently described method6 that allows effective identification and EM analysis of a small group of cells within the embryo, without relying on distinct morphological markers (Fig. 1). We demonstrate the merit of this procedure using zebrafish primordial germ cells, a population of less than 25 cells within early embryos, which at that developmental stage (8–12 hours post fertilization [hpf]) consist of more than 20,000 cells.7,8 This method relies on preservation of the fluorescent protein signal after fixation, combined with two-photon laser marking within the tissue at defined locations relative to the position where the cells of interest reside. FIG. 1. Schematic workflow of the correlative light and electron microscopy using laser marking. (A) Zebrafish embryo fixed in 2% PFA + 0.2% glutaraldehyde in 0.1 M PHEM buffer for 10 min, deyolked and fixed for an additional 2 h ...

Aclar, a copolymer film with properties very similar to those of tissue culture plastic, is a versatile substrate to grow cells for light (including fluorescence) and electron microscopic applications in combination with both chemical fixation and cryoimmobilization. In this paper, we describe complete procedures to perform correlative light and electron microscopy using Aclar as substrate for the culture of cell monolayers to be finally embedded in plastic. First, we developed straightforward, efficient and flexible ways to mark the surface of the Aclar to create substrates to locate cells first at the light microscopy and then the electron microscopy level. All the methods enable the user to self-design gridded Aclar pieces, according to the purpose of the experiments, and create a large number of substrates in a short time. Second, we confirmed that marked Aclar supports the normal growth and morphology of cells. Third, we validated the correlative light and electron microscopy procedure using Aclar. This validation was done for the high-resolution analysis of endothelial cells using transmission electron microscopy and focused ion beam-scanning electron microscopy in combination with the use of fluorescence, phase contrast and/or bright field microscopy to map areas of interest at low resolution. The methods that we present are diverse, easy to implement and highly reproducible, and emphasize the versatility of Aclar as a cell growth substrate for diverse microscopic applications.

<p xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" class="first" dir="auto" id="d26882e143">Over the past 30 years, therapeutic monoclonal antibodies (mAbs) have been used successfully as anti-cancer agents [ <a class="xref-link" href="#r1">1</a>]. Specific binding to target antigens induces cytotoxicity by promoting immune responses. Cluster of differentiation 20 (CD20) is a cell-surface receptor expressed in most B cells. Rituximab (RTX) was the first approved anti-CD20 therapeutic mAb for the treatment of B-cell malignancies and autoimmune disorders [ <a class="xref-link" href="#r2">2</a>]. In the meantime, RTX is complemented by human and humanized mAbs such as ofatumumab (OFA) and obinutuzumab (OBZ), respectively. Binding of these mAbs to B cells alter the nanoscale organization of CD20, consequently activating a cytotoxic pathway. However, molecular details on the binding mechanism of mAbs and binding-induced reorganization of endogenous CD20 on the membrane and how this activates the immune system to kill B cells remains poorly understood. To decode the molecular interplay of endogenous CD20 on B cells with therapeutic mAbs, 3D volumetric super-resolution fluorescence imaging methods are needed that provide a high spatiotemporal resolution. <p xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" dir="auto" id="d26882e151">In this work, we have introduced a speed-optimized variant of DNA-points accumulation for imaging in nanoscale topography (DNA-PAINT) [ <a class="xref-link" href="#r3">3</a>]. Using standard imager probes with two identical fluorophores ( <a class="xref-link" href="#fg001">Fig. 1a</a>) that form non-fluorescent H-dimers in the unbound state enables the use of higher probe concentrations and consequently ~15-fold faster 2D imaging, compared to classical DNA-PAINT probes [ <a class="xref-link" href="#r4">4</a>]. For implementing fast 3D imaging of whole cells at virtually molecular resolution, we combined two-dye imager (TDI)-DNA-PAINT with lattice light-sheet (LLS) microscopy ( <a class="xref-link" href="#fg001">Fig. 1b</a>) [ <a class="xref-link" href="#r5">5</a>]. Furthermore, we exploited LLS imaging to elucidate the real time binding interactions of mAbs and CD20 in live B cells in 4D. <div class="fig panel" id="fg001"> <a class="named-anchor" id="fg001"> <!-- named anchor --> </a> <div class="figure-container so-text-align-c"> <img alt="" class="figure" src="/document_file/857f59d2-151c-4143-a6bf-7ea96473d0fd/ScienceOpen/image/292_6725fecdb0b33-ArindamGhoshAPMC2025_fig1.jpg"/> </div> <div class="panel-content"> <div class="label">Fig. 1</div> <div class="caption" id="d26882e172"> <p class="first" dir="auto" id="d26882e173"> <b>a</b>, TDI-DNA-PAINT workflow with dual-labeled imager probes (left). Exemplary fast super-resolved reconstruction of cellular microtubule networks after 5 minutes (right) of data acquisition. Scale 1 μm. <b>b,</b> LLS imaging principle. Created with BioRender. </div> </div> </div>

Correlative light and electron microscopy (CLEM) methods are powerful methods that combine molecular organization (from light microscopy) with ultrastructure (from electron microscopy). However, CLEM methods pose high cost/difficulty barriers to entry and have very low experimental throughput. Therefore, we have developed an indirect correlative light and electron microscopy (iCLEM) pipeline to sidestep the rate-limiting steps of CLEM (i.e., preparing and imaging the same samples on multiple microscopes) and correlate multiscale structural data gleaned from separate samples imaged using different modalities by exploiting biological structures identifiable by both light and electron microscopy as intrinsic fiducials. We demonstrate here an application of iCLEM, where we utilized gap junctions and mechanical junctions between muscle cells in the heart as intrinsic fiducials to correlate ultrastructural measurements from transmission electron microscopy (TEM), and focused ion beam scanning electron microscopy (FIB-SEM) with molecular organization from confocal microscopy and single molecule localization microscopy (SMLM). We further demonstrate how iCLEM can be integrated with computational modeling to discover structure-function relationships. Thus, we present iCLEM as a novel approach that complements existing CLEM methods and provides a generalizable framework that can be applied to any set of imaging modalities, provided suitable intrinsic fiducials can be identified.

Photobleaching is a major challenge in fluorescence microscopy, in particular if high excitation light intensities are used. Signal‐to‐noise and spatial resolution may be compromised, which limits the amount of information that can be extracted from an image. Photobleaching can be bypassed by using exchangeable labels, which transiently bind to and dissociate from a target, thereby replenishing the destroyed labels with intact ones from a reservoir. Here, we demonstrate confocal and STED microscopy with short, fluorophore‐labeled oligonucleotides that transiently bind to complementary oligonucleotides attached to protein‐specific antibodies. The constant exchange of fluorophore labels in DNA‐based STED imaging bypasses photobleaching that occurs with covalent labels. We show that this concept is suitable for targeted, two‐color STED imaging of whole cells.