Toward Compositional Contrast by Cryo-STEM

Electron microscopy (EM) is a versatile tool for studying matter at various scales. However, the high vacuum and charged irradiation necessitate careful specimen stabilization, particularly for biological samples sensitive to dehydration and radiation damage. Cryogenic fixation, or vitrification, avoids dehydration and preserves a native state, but the samples remain delicate. While cryo-EM has significantly advanced resolution from membranes to atoms, its potential for compositional analysis remains largely untapped. Conventional cryo-transmission electron microscopy (cryo-TEM) suffers from contrast limitations, especially at low spatial frequencies. Soft X-ray cryo-tomography offers quantitative contrast based on differential absorption, inspiring the exploration of similar quantitative approaches in cryo-EM. Scanning transmission electron microscopy (STEM), with its incoherent elastic scattering sensitive to atomic number (Z), presents an opportunity to achieve compositional contrast in cryo-EM. This paper reviews the application of STEM and its various detection modes to biological cryo-EM and tomography.

The paper references several key publications demonstrating the feasibility and advantages of cryo-STEM tomography (CSTET). These include studies on whole bacteria, where CSTET visualized internal structures and provided phosphorus detection in polyphosphate bodies. Another study used CSTET to image mitochondrial calcium stores in intact cells. Furthermore, single-particle cryo-STEM analysis successfully detected and located individual zinc and iron ions bound to heavy-chain ferritin. These prior works lay the foundation for the current account, highlighting the growing interest and successful applications of STEM in biological cryo-EM.

The authors discuss the principles of STEM, emphasizing the importance of illumination convergence and detector configuration. Different STEM modes—bright field (BF), annular dark field (ADF), and high angle annular dark field (HAADF)—are defined by the integration of scattered electrons over specific angular ranges. The contrast in amorphous materials is directly related to atomic number, offering compositional information. They compare STEM with TEM, noting that STEM avoids the contrast inversions and low-frequency limitations of TEM's phase contrast. The paper describes CSTET, a technique combining STEM with tomography, to obtain three-dimensional reconstructions. They detail experimental approaches, including using small condenser apertures for extended depth of field, the use of BF and ADF detectors simultaneously, and the variation of camera length and inner cutoff angle for compositional analysis. Single-particle analysis in ADF-STEM is also used to visualize isolated metal ions. In addition, the application of 3D deconvolution methods for enhancing tomographic reconstruction is explored. Finally, the methods used for analysis of image intensities and the estimation of ion concentrations are thoroughly detailed.

CSTET offered superior contrast for low-resolution features and axial resolution compared to conventional cryo-TEM tomography. The unipolar contrast in STEM facilitates straightforward interpretation of density. The radiation damage threshold was significantly higher for STEM than TEM in cryo-tomography, likely due to the intermittent nature of STEM illumination. BF-STEM was found to be advantageous over ADF-STEM for thick specimens due to its insensitivity to resolution loss from multiple scattering. Analysis of annular dark field scattering data, combined with known scattering cross-sections, allowed for a crude estimation of phosphorus concentration in bacterial polyphosphate bodies. The IBF-STEM mode enabled density estimation of mitochondrial matrix granules, identified as calcium phosphate by EDS, indicating their significance in ion buffering. The application of 3D deconvolution improved the visibility of high-density granules and membrane structures, enabling enhanced morphological interpretation. Finally, single-particle ADF-STEM successfully detected isolated zinc and iron ions bound to heavy-chain ferritin, demonstrating the possibility of visualizing single atoms in cryo-EM.

The findings demonstrate the significant potential of STEM to complement and extend the capabilities of conventional cryo-EM. The advantages of STEM in achieving compositional contrast and in its ability to image thicker specimens make it a powerful tool for investigating cellular structures and processes where both morphology and composition are essential. The higher radiation damage threshold for STEM is particularly important for cryo-tomography. The ability to estimate ion concentrations directly from image intensity offers new avenues for quantitative analysis of cellular composition. The use of 3D deconvolution enhances the interpretability of tomographic reconstructions, improving both morphological and compositional information. The detection of single metal ions expands the possibilities for studying the interaction of metals with biological macromolecules at the atomic level. The differences between STEM and TEM contrast mechanisms are highlighted, emphasizing the need for appropriate interpretation based on the imaging mode and sample characteristics. The authors suggest that advanced phase-imaging techniques in STEM will be crucial for achieving higher resolutions in cryo-EM of biological macromolecules.

This Account showcases the power of STEM in cryo-EM of biological specimens, offering superior contrast and resolution, particularly for thicker samples. The method allows for compositional analysis alongside morphological studies, providing a richer understanding of cellular structure. Future work will focus on improving dose efficiency, developing advanced phase-imaging techniques, and refining quantitative analysis methods to fully exploit the potential of STEM in biological cryo-EM. The combination of various STEM detection modes, advanced image processing, and potentially faster data acquisition hardware will open new opportunities in cryo-EM.

The current methods for estimating ion concentrations from image intensity are still relatively crude, requiring further refinement for accurate quantification. The resolution achieved in single-particle cryo-STEM analysis, while sufficient for localizing metal ions, is lower than that typically obtained in cryo-TEM for protein structure determination. The application of 3D deconvolution, while beneficial, can introduce parameter-dependent distortions in intensity histograms, complicating quantitative interpretation. Further research is needed to fully optimize experimental parameters and to develop more sophisticated image processing techniques for more accurate and precise analysis.