Exploring cryo-plasma FIB/SEM volume imaging for biological specimens

In the rapidly evolving field of biological imaging, a recent study has explored the untapped potential of serial cryogenic plasma-focused ion beam scanning electron microscopy (cryo-serial pFIB/SEM) to visualize biological samples with unprecedented clarity and detail.

Cryo-serial plasma FIB/SEM represents a leap forward in biological imaging. It allows scientists to examine cells, tissues, and even organelles in a way that maintains their natural, hydrated state. 

The researchers behind this investigation, primarily from the Rosalind Franklin Institute (UK), explored the application of this powerful tool to image vitrified biological specimens. By using cryogenic temperatures, they preserved the native, hydrated state of biological samples while maintaining structural integrity without the need for staining. Cryo conditions prevent ice crystal formation and structural damage, allowing for more accurate imaging.

Their work uncovers specific challenges, such as low contrast and sample artifacts, while offering tangible solutions, including the use of plasma ion sources to enhance image clarity and reduce curtaining. These innovations open new doors for in situ structural biology at the mesoscale (10 nm to 10 μm), providing an unprecedented level of detail for studying cellular architecture and interactions.

Why cryo-serial pFIB/SEM matters

Traditional electron microscopy often requires fixation, staining [1,2], or dehydration of samples, which can alter the true biological structures [3]. By contrast, cryo-FIB/SEM allows for imaging biological samples in their frozen, hydrated state without staining.

However, imaging unstained cryogenic samples has been fraught with challenges, including low contrast and artifacts such as curtaining, which distort image quality. To address these, the researchers introduced plasma ion sources. Cryo-serial pFIB/SEM builds on the qualities of cryo-FIB/SEM by using plasma ion sources, which reduce artifacts and enhance resolution, offering even greater potential for high-quality imaging of complex biological samples. 

Overcoming imaging challenges

One of the key difficulties in imaging biological samples is the low contrast due to the light atomic composition of most biomolecules. This lack of contrast is often exacerbated by the inherent noise in electron microscopy [4]. The researchers in this study explored different plasma ion sources - argon, xenon, nitrogen, and oxygen, to determine which would produce the clearest images. Their experiments showed that argon and xenon gases performed best for cryo-volume EM, as they minimized curtaining artifacts and produced smoother milled surfaces, which is crucial for high-resolution imaging of biological structures in their native state. 

The study also explored the possibility of perpendicular imaging, where the sample is imaged at a 90-degree angle to the milled surface, rather than the standard 52-degree configuration. Combining perpendicular imaging with plasma ion sources and cryo samples significantly improved contrast and resolution for biological structures, including thicker samples like bacterial outer membranes and cellular organelles. 

Curtaining score for the different plasma sources at different currents

For each gas and current setting, Chlamydia trachomatis-infected HeLa cells that had been plunge-frozen were subjected to milling. A total of 15 windows, each measuring 2 × 2.5 × 2 μm3, were milled at various currents using an acceleration voltage of 30 kV for xenon, oxygen, and nitrogen, and 20 kV for argon gas. The plasma incidence angle was set at 18°. Scanning electron microscopy (SEM) images were captured at a 90° angle to the focused ion beam (FIB). (A) Sample images from these data illustrate minimal curtaining (left, oxygen, 213 pA) and significant curtaining (right, nitrogen, 2.2 nA), which appear as vertical lines. The arrow and curly brackets highlight the location of a curtain or curtain group. Scale bar: 1 μm. (B) A graph depicting the curtaining score as a function of current (see Materials and methods). Each point represents the mean value with the standard error. The trend line connects the data points. n=15 for each condition. (C) Sample images of windows created using oxygen. The arrow and curly brackets are used as in (A). Scale bar: 1 μm.

Advances in automation and targeting specific regions

To push the limits of cryo-serial pFIB/SEM further, the researchers developed an automated workflow for serial sectioning and imaging. This innovation allows them to image large volumes of biological material with minimal manual intervention, accelerating the data collection process. Automation is particularly important when imaging thick tissue sections, where traditional methods would require days to complete a full volumetric scan.

The study suggests the potential for cryo-serial pFIB/SEM as a tag-free localization tool to identify regions of interest (ROIs) within 3D volumes. This is crucial for in situ structural biology applications, especially when researchers need to target specific areas of tissue for more detailed analyses, such as cryo-ET. Cryogenic fluorescence microscopy, typically used for targeting, lacks the axial resolution necessary to reliably locate ROIs in thick tissue samples, making this new tool a significant improvement. 

Applications across biological specimens

The versatility of cryo-serial pFIB/SEM imaging was demonstrated across various biological samples. The researchers successfully imaged bacterial cells, human cells, and mammalian tissues, each presenting unique challenges and insights.

1. Bacterial cells: Rhodospirillum rubrum

In one of the most exciting findings, the researchers imaged Rhodospirillum rubrum, a species of purple bacteria known for its photosynthetic capabilities. The researchers used cryo-volume EM, milling away layers of the sample to image each one with SEM. Using cryo-FIB/SEM, they then observed developing and mature chromatophores, small vesicles that contain the photosynthetic machinery. This level of detail allowed them to capture the early stages of chromatophore development, a process previously unseen in such clarity. The imaging revealed that chromatophores begin budding from the inner membrane, providing new insights into how these structures form and organize.

R. rubrum image stacks acquired using argon serial (pFIB)/SEM

(A) Comprehensive view of a section from a serial pFIB/SEM volume, depicting vitrified R. rubrum. An enlarged view of the bacterial features is shown on the right, with highlighted areas indicating nascent chromatophores (green), mature chromatophores (pink), and storage granules (brown). The 'X' marks indicate dead/dying bacteria or debris. (B) A slice of the volume overlaid with the volume rendering post-segmentation. The outer membrane is colored red, and the mature chromatophores are green. The number associated with each segmented bacterium represents the ratio of the number of chromatophores to the total surface area occupied by the bacteria in each slice. The slice was processed using a Gaussian filter with a 2-pixel radius.

2. Mammalian cells: HeLa cells

For human HeLa cells, the team used cryo-volume EM to achieve high-resolution imaging of organelles such as the mitochondria, centrosomes, and nuclear compartments. The ability to visualize membrane contact sites (MCS) between the endoplasmic reticulum and mitochondria allowed for a deeper understanding of intracellular interactions and networks. This detailed view of subcellular structures in their native state underscores the power of cryo-serial pFIB/SEM for cellular biology.

HeLa cells imaged using serial (pFIB)/SEM

(A–E) Serial pFIB/SEM volume of a HeLa cell, milled with argon and imaged at a 52° angle to the surface by SEM. (A) A magnified region of interest, displaying nuclear pore complexes (NPCs, arrows) and endoplasmic reticulum (ER). (B) An overview where the nucleus, mitochondria (mito), multivesicular body (MVB), and centriole (red box) are clearly distinguishable. (C) A close-up of the centriole identified in (B), with its three-dimensional (3D) rendering shown in (D). (E) This HeLa cell exhibited two centrosomes, each with two centrioles and associated pericentriolar matrix (PCM). The green line indicates the distance between the centers of the two centrosomes, while the yellow lines show the distances between the centrioles within each pair. (A–C) The slices were processed using a 2-pixel radius mean filter in Fiji [5]. For (E), a band pass filter was applied, also in Fiji.

3. Mouse brain and heart tissue

The study also extended to complex tissue samples, including mouse brain and heart tissue. In the brain, the researchers used cryo-volume EM to visualize synapses, neuronal cells, and their surrounding environment, capturing the intricate connections between neurons and glial cells. In the heart tissue, they imaged cardiomyocytes, revealing the organization of sarcomeres and mitochondria. This level of detail is crucial for understanding tissue-level functions and disorders, particularly in neurodegenerative and cardiovascular diseases. 

Non-fixed, high-pressure frozen (HPF) mouse brain slice milled with argon and imaged by SEM at 52° to the surface

(A) A representative slice from a wide field of view of the serial plasma focused ion beam (pFIB)/SEM volume. Scale bar: 1 µm. Colored insets highlight regions of interest, including myelin sheaths (yellow), putative nuclear pore complexes (red), and mitochondria (blue). Non-colored insets display synapse morphologies from different slices, with pre- and post-synaptic cells indicated by light pink and magenta arrows, respectively. Scale bar: 500 nm. (B) Magnified slices from the region indicated in pink in (A), showing a neuronal synapse at progressive Z positions from left to right. Scale bar: 500 nm. (C) A three-dimensional (3D) volume rendering of the synapse shown in (B), with pre- (green) and post- (purple) synaptic membranes labeled. The post-synaptic density (red) and pre-synaptic vesicles (yellow) are clearly visible. The presented slices have been filtered using a 2-pixel radius mean filter for clarity.

Moving toward correlative imaging workflows

One of the most promising aspects of this study is the use of cryo-serial pFIB/SEM for milling lamella in correlative imaging workflows. By combining this method with cryo-ET and cryo-correlative light and electron microscopy (cryo-CLEM), as well as integrated fluorescence light microscopes like Delmic's METEOR, researchers can achieve a seamless transition from tissue-level to molecular-level imaging. This correlative approach allows scientists to link observations made at the mesoscale with molecular data obtained from high-resolution techniques like cryo-ET. 

The researchers successfully used cryo-serial pFIB/SEM to mill lamellae (thin sections of biological tissue), for subsequent analysis by cryo-ET. This technique opens up new possibilities for studying protein complexes in their native environment, without the need for chemical fixation or staining.

Computational solutions and machine learning aids

To address ongoing challenges, such as image artifacts caused by charging in lipid-rich areas, the researchers implemented a computational approach to remove distortions. Using machine learning, specifically U-Net-based algorithms, they segmented and cleaned up images to enhance biological feature visibility. This not only improved the overall quality of the datasets but also reduced the manual effort required to analyze large volumes of data. 

Moreover, the team applied automated segmentation techniques to mitochondria within the datasets, achieving high accuracy in identifying these organelles. This approach is a significant step forward in handling the vast amount of data generated by cryo-FIB/SEM, allowing for quicker and more accurate analyses of complex biological samples. 

Correlation of fluorescence, pFIB/SEM, and TEM at cryogenic temperatures

(A and B) R. rubrum were imaged using fluorescence microscopy within the dual-beam microscope chamber prior to the deposition of protective platinum layers. Serial pFIB/SEM was then conducted, followed by the acquisition of TEM images of the lamella (F and G). The colored dots in panels B and D–G represent common features across the different imaging modalities. The green rectangles in A and B denote the regions of interest (ROI) where serial pFIB/SEM was performed. The orange rectangle in D and E highlights the septum on both SEM and TEM, while the red corner indicates the presence of lipid droplets in E and G. The blue outline in (A–C) shows the alignment between the fluorescence and SEM images. The images were not filtered.

Conclusion

The work conducted by the researchers paves the way for future advancements in biological imaging. By optimizing plasma ion sources, refining imaging workflows, and developing new tools for targeting and automation, they have demonstrated that cryo-serial pFIB/SEM can overcome many of the traditional limitations of electron microscopy. This method’s ability to provide high-resolution, volumetric images of biological samples without compromising their native state is a significant contribution to the field of structural biology.

The study’s findings underscore the potential for cryo-serial pFIB/SEM to become a staple in biological research, particularly for studies of protein complexes, cellular structures, and tissue-level interactions. As the technique continues to evolve, it promises to unlock new discoveries and deepen our understanding of the intricate workings of life at the smallest scales.

Based on the study - https://elifesciences.org/articles/83623

References

1. Knott, G., Marchman, H., Wall, D., & Lich, B. (2008). Serial section scanning electron microscopy of adult brain tissue using focused ion beam milling. Journal of Neuroscience, 28(12), 2959–2964. https://doi.org/10.1523/jneurosci.3189-07.2008
2. Watson, M. L. (1958). Staining of Tissue Sections for Electron Microscopy with Heavy Metals. The Journal of Cell Biology, 4(4), 475–478. https://doi.org/10.1083/jcb.4.4.475
3. Thompson, R. F., Walker, M., Siebert, C. A., Muench, S. P., & Ranson, N. A. (2016). An introduction to sample preparation and imaging by cryo-electron microscopy for structural biology. Methods, 100, 3–15. https://doi.org/10.1016/j.ymeth.2016.02.017