Name: sample preparation by 3d-correlative focused ion beam milling for high-resolution cryo-electron tomography
Uploaded: 2021-10-25T13:00:00
Duration: 8 min 20 s
Description: Watch this Scientific Journal Video about Sample Preparation by 3D-Correlative Focused Ion Beam Milling for High-Resolution Cryo-Electron Tomography at JoVE.com

We thank Inga Wolf for supporting the IT infrastructure, Florian Beck for computational support, and Oda H. Schi&#248;tz for the critical reading of the manuscript. Funding was provided in part through an Alexander von Humboldt returners fellowship to Philipp S. Erdmann and an EMBO Long-term Fellowship ALTF 764-2014 to Florian Wilfling. Anna Bieber was supported by a Boehringer Ingelheim Fonds Ph.D. fellowship.

<ol>
	<li>Beck, M., Baumeister, W. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Cryo-Electron+tomography:+Can+it+reveal+the+molecular+sociology+of+cells+in+atomic+detail.">Cryo-Electron tomography: Can it reveal the molecular sociology of cells in atomic detail.</a> Trends in Cell Biology. 26 (11), 825-837 (2016).</li><li>Plitzko, M., Villa, E., Schaffer, M., Baumeister, W. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Opening+windows+into+the+cell+focused-ion-beam+milling+for+cryo-electron+tomography.">Opening windows into the cell focused-ion-beam milling for cryo-electron tomography.</a> Current Opinion in Structural Biology. 23, 771-777 (2013).</li><li>Schaffer, M., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Optimized+cryo-focused+ion+beam+sample+preparation+aimed+at+in+situ+structural+studies+of+membrane+proteins.">Optimized cryo-focused ion beam sample preparation aimed at in situ structural studies of membrane proteins.</a> Journal of Structural Biology. 197 (2), 73-82 (2017).</li><li>Tegunov, D., Xue, L., Dienemann, C., Cramer, P., Mahamid, J. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Multi-particle+cryo-EM+refinement+with+M+visualizes+ribosome-antibiotic+complex+at+3.5+&#197;+in+cells.">Multi-particle cryo-EM refinement with M visualizes ribosome-antibiotic complex at 3.5 &#197; in cells.</a> Nature Methods. 18 (2), 186-193 (2021).</li><li>Klein, S., Wachsmuth-Melm, M., Winter, S. L., Kolovou, A., Chlanda, P. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Cryo-correlative+light+and+electron+microscopy+workflow+for+cryo-focused+ion+beam+milled+adherent+cells.">Cryo-correlative light and electron microscopy workflow for cryo-focused ion beam milled adherent cells.</a> Methods in Cell Biology. 162, 273-302 (2021).</li><li>Klein, S., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Post-correlation+on-lamella+cryo-CLEM+reveals+the+membrane+architecture+of+lamellar+bodies.">Post-correlation on-lamella cryo-CLEM reveals the membrane architecture of lamellar bodies.</a> Communications Biology. 4 (1), 1-12 (2021).</li><li>Arnold, J., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Site-Specific+Cryo-focused+ion+beam+sample+preparation+guided+by+3D+correlative+microscopy.">Site-Specific Cryo-focused ion beam sample preparation guided by 3D correlative microscopy.</a> Biophysical Journal. 110 (4), 860-869 (2016).</li><li>Wilfling, F., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=A+selective+autophagy+pathway+for+phase-separated+endocytic+protein+deposits.">A selective autophagy pathway for phase-separated endocytic protein deposits.</a> Molecular Cell. 80 (5), 764-778 (2020).</li><li>. Scientific volume imaging, Nyquist calculator Available from: <a target="_blank" href="https://svi.nl/NyquistCalculator">https://svi.nl/NyquistCalculator</a> (2021)</li><li>Schindelin, J., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Fiji:+an+open-source+platform+for+biological-image+analysis.">Fiji: an open-source platform for biological-image analysis.</a> Nature Methods. 9 (7), 676-682 (2012).</li><li>de Chaumont, F., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Icy:+an+open+bioimage+informatics+platform+for+extended+reproducible+research.">Icy: an open bioimage informatics platform for extended reproducible research.</a> Nature Methods. 9 (7), 690-696 (2012).</li><li>. 3DCT Available from: <a target="_blank" href="https://3dct.semper.space/">https://3dct.semper.space/</a> (2021)</li><li>Klumpe, S., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=A+modular+platform+for+streamlining+automated+cryo-FIB+workflows.">A modular platform for streamlining automated cryo-FIB workflows.</a> bioRxiv. , 444745 (2021).</li><li>Wolff, G., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Mind+the+gap:+Micro-expansion+joints+drastically+decrease+the+bending+of+FIB-milled+cryo-lamellae.">Mind the gap: Micro-expansion joints drastically decrease the bending of FIB-milled cryo-lamellae.</a> Journal of Structural Biology. 208 (3), 0 (2019).</li><li>Buckley, G., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Automated+cryo-lamella+preparation+for+high-throughput+in-situ+structural+biology.">Automated cryo-lamella preparation for high-throughput in-situ structural biology.</a> Journal of Structural Biology. 210 (2), 107488 (2020).</li><li>Tacke, S., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=A+streamlined+workflow+for+automated+cryo+focused+ion+beam+milling.">A streamlined workflow for automated cryo focused ion beam milling.</a> bioRxiv. , 963033 (2020).</li><li>Zachs, T., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Fully+automated,+sequential+focused+ion+beam+milling+for+cryo-electron+tomography.">Fully automated, sequential focused ion beam milling for cryo-electron tomography.</a> eLife. 9, e52286 (2020).</li><li>. tools3dct Available from: <a target="_blank" href="https://github.com/hermankhfung/tools3dct">https://github.com/hermankhfung/tools3dct</a> (2021)</li><li>Mastronarde, D. N. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Automated+electron+microscope+tomography+using+robust+prediction+of+specimen+movements.">Automated electron microscope tomography using robust prediction of specimen movements.</a> Journal of Structural Biology. 152 (1), 36-51 (2005).</li><li>Yang, J. E., Larson, M. R., Sibert, B. S., Shrum, S., Wright, E. R. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=CorRelator:+Interactive+software+for+real-time+high+precision+cryo-correlative+light+and+electron+microscopy.">CorRelator: Interactive software for real-time high precision cryo-correlative light and electron microscopy.</a> Journal of Structural Biology. 213 (2), 107709 (2021).</li><li>Hagen, W. J. H., Wan, W., Briggs, J. A. G. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Implementation+of+a+cryo-electron+tomography+tilt-scheme+optimized+for+high+resolution+subtomogram+averaging.">Implementation of a cryo-electron tomography tilt-scheme optimized for high resolution subtomogram averaging.</a> Journal of Structural Biology. 197 (2), 191-198 (2017).</li><li>B&#228;uerlein, F. J. B., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=In+situ+architecture+and+cellular+interactions+of+polyQ+inclusions.">In situ architecture and cellular interactions of polyQ inclusions.</a> Cell. 171 (1), 179-187 (2017).</li><li>Toro-Nahuelpan, M., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Tailoring+cryo-electron+microscopy+grids+by+photo-micropatterning+for+in-cell+structural+studies.">Tailoring cryo-electron microscopy grids by photo-micropatterning for in-cell structural studies.</a> Nature Methods. 17 (1), 50-54 (2020).</li><li>Shaner, N. C., Steinbach, P. A., Tsien, R. Y. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=A+guide+to+choosing+fluorescent+proteins.">A guide to choosing fluorescent proteins.</a> Nature Methods. 2 (12), 905-909 (2005).</li><li>Kaufmann, R., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Super-resolution+microscopy+using+standard+fluorescent+proteins+in+intact+cells+under+cryo-conditions.">Super-resolution microscopy using standard fluorescent proteins in intact cells under cryo-conditions.</a> Nano Letters. 14 (7), 4171-4175 (2014).</li><li>Booy, F. P., Pawley, J. B. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Cryo-crinkling:+what+happens+to+carbon+films+on+copper+grids+at+low+temperature.">Cryo-crinkling: what happens to carbon films on copper grids at low temperature.</a> Ultramicroscopy. 48 (3), 273-280 (1993).</li><li>Krull, A., Buchholz, T. O., Jug, F. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Noise2void-Learning+denoising+from+single+noisy+images.">Noise2void-Learning denoising from single noisy images.</a> Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition. 2019, 2124-2132 (2019).</li><li>Gorelick, S., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=PIE-scope,+integrated+cryo-correlative+light+and+FIB/SEM+microscopy.">PIE-scope, integrated cryo-correlative light and FIB/SEM microscopy.</a> eLife. 8, 1-15 (2019).</li><li>. Delmic METEOR Available from: <a target="_blank" href="https://www.delmic.com/en/products/cryo-solutions/meteor">https://www.delmic.com/en/products/cryo-solutions/meteor</a> (2021)</li><li>. iFLM Available from: <a target="_blank" href="https://assets.thermofisher.com/TFS-Assets/MSD/Datasheets/iflm-aquilos-datasheet-ds0366.pdf">https://assets.thermofisher.com/TFS-Assets/MSD/Datasheets/iflm-aquilos-datasheet-ds0366.pdf</a> (2021)</li><li>Mahamid, J., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Liquid-crystalline+phase+transitions+in+lipid+droplets+are+related+to+cellular+states+and+specific+organelle+association.">Liquid-crystalline phase transitions in lipid droplets are related to cellular states and specific organelle association.</a> Proceedings of the National Academy of Sciences of the United States of America. 116 (34), 16866-16871 (2019).</li><li>Scher, N., Rechav, K., Paul-Gilloteaux, P., Avinoam, O. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=In+situ+fiducial+markers+for+3D+correlative+cryo-fluorescence+and+FIB-SEM+imaging.">In situ fiducial markers for 3D correlative cryo-fluorescence and FIB-SEM imaging.</a> iScience. 24 (7), 102714 (2021).</li><li>Mahamid, J., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=A+focused+ion+beam+milling+and+lift-out+approach+for+site-specific+preparation+of+frozen-hydrated+lamellas+from+multicellular+organisms.">A focused ion beam milling and lift-out approach for site-specific preparation of frozen-hydrated lamellas from multicellular organisms.</a> Journal of Structural Biology. 192, 262-269 (2015).</li><li>Schaffer, M., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=A+cryo-FIB+lift-out+technique+enables+molecular-resolution+cryo-ET+within+native+Caenorhabditis+elegans+tissue.">A cryo-FIB lift-out technique enables molecular-resolution cryo-ET within native Caenorhabditis elegans tissue.</a> Nature Methods. 16 (8), 757-762 (2019).</li><li>Wu, G. -. H., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Multi-scale+3D+cryo-correlative+microscopy+for+vitrified+cells.">Multi-scale 3D cryo-correlative microscopy for vitrified cells.</a> Structure. 28 (11), 1231-1237 (2020).</li><li>Liu, B., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Three-dimensional+super-resolution+protein+localization+correlated+with+vitrified+cellular+context.">Three-dimensional super-resolution protein localization correlated with vitrified cellular context.</a> Scientific Reports. 5, 13017 (2015).</li><li>Weisenburger, S., Jing, B., Renn, A., Sandoghdar, V. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Cryogenic+localization+of+single+molecules+with+angstrom+precision.">Cryogenic localization of single molecules with angstrom precision.</a> SPIE NanoScience + Engineering. 8815, (2013).</li><li>Tuijtel, M. W., Koster, A. J., Jakobs, S., Faas, F. G. A., Sharp, T. H. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Correlative+cryo+super-resolution+light+and+electron+microscopy+on+mammalian+cells+using+fluorescent+proteins.">Correlative cryo super-resolution light and electron microscopy on mammalian cells using fluorescent proteins.</a> Scientific Reports. 9 (1), 1-11 (2019).</li></ol>

The authors declare no conflicts of interest.

1. Critical steps in the protocol
Optimization of cell culture and grid plunging parameters is fundamental for this workflow. At the beginning of a project, it is worth investing time to optimize tagging strategies, the distribution of cells and fiducial beads, and test different grid preparation and blotting parameters. Working with an optimally plunge-frozen sample will significantly facilitate downstream processing.
As for any TEM experiment, vitreous samples are required. For large mammalian cells such as HeLa, 1-2 cells per grid square are preferable, but cells may still be vitreous at higher density. Optionally, vitrification can be improved in mammalian cells (e.g., HEK293, HeLa) by incubating them with 2.5-10% (v/v) glycerol added to the culture medium 10 min before plunging23. If available, grid patterning may be used to ensure perfect placement and distribution of the cells, thereby improving vitrification and later correlation24.
While specific cells can be selected during the workflow, too few cells that show the biological feature of interest will significantly reduce overall throughput. To improve correlation in POI-positive cells, sufficiently bright fluorophores should be used. This is especially important at endogenous expression levels. We found that under cryo-conditions, mVenus often performed better than EGFP due to its increased brightness25 and the hypsochromic shift, which keeps it suitable for standard GFP filter setups under cryo-conditions26. For non-point-like target structures, the trade-off between wavelength and localization accuracy (Abbe diffraction limit) should also be considered.
Efficient 3D-correlation also requires that grids are mechanically stable and are handled with great care. While standard gold or copper grids with carbon support may be used, the success rate may significantly be increased by using more rigid SiO2 films depending on the project. However, it has not yet been conclusively determined whether (a) mechanical stability or (b) matching thermal expansion coefficients (substrate vs. film) to reduce cryo-wrinkling27, is the most crucial factor for successful 3D correlation. Moreover, for picking up fragile Au grids, polydimethylsiloxane-coated dishes may be used5.
In addition to ensuring sample stability, a careful choice of FLM imaging parameters is necessary for obtaining high-quality fluorescence stacks that are suitable for optimal targeting during FIB milling. In this regard, testing different denoising28 or deconvolution techniques on the FLM data is also advised, as it may considerably improve the localization of fiducials and cellular signals. When correlating the fluorescence signal to FIB-SEM images, a good sampling of fiducial beads is important. They should be well distributed around the cells and possibly at different z&#160;heights. It is also good practice to validate the consistency of the correlation by checking the predicted vs. actual positions of beads that were deliberately left out of the fiducial model but can clearly be correlated by eye. 3DCT&#39;s RMSE values should also always be considered to check the registration consistency.
Since the deposition of milled material and residual water from the FIB-SEM chamber (i.e., recontamination) increases the effective lamella thickness by adding amorphous material to both sides of it, keeping fine-milled lamellas in the microscope for a prolonged time generally reduces TEM data quality due to additional electron scattering events. Accordingly, milling is most often performed in a two-step fashion: first, all positions are milled roughly (i.e., to about 800 nm), and then finely (to ~150-250 nm), and the grid is immediately unloaded after the last lamella has been completed. Better correlation success may, however, be achieved by processing the positions of interest in a site-wise manner, hence performing rough and fine milling on the same lamella directly after one another since this leaves no time for bending or deformation. This, however, reduces the maximum number of lamellas that can be produced per grid depending on the recontamination rate of the system. For a rate of 20 nm/h, 4-6 lamellas are produced within 1-1.5 h.
Movement of the entire grid or the rough-milled lamellas &#62;300 nm will result in poor or unsuccessful correlation (see also limitations discussed below). It should therefore be checked regularly, e.g., by comparing IB images before, during, and after FIB milling. Sites that show significant movement (&#62;300 nm) should be discarded. Optimize the sample preparation (i.e., choice of grid type, cell density, and plunging parameters; see protocol section 1) and milling strategy to avoid these movements. Lamella bending can significantly be reduced by site-wise milling as described in step 3.6 and reducing the lamella width. As mentioned before, while stress relief cuts15 have been designed to reduce lamella bending, they often result in a concerted movement of the de-coupled lamella, thereby effectively preventing correlation. Integrated FLM systems may be used to solve this problem.
2. Modifications and troublehooting of the method
It is highly advised to perform a thorough characterization of the sample in live-cell imaging before going to cryo-conditions. Optimizing the cellular samples, treatment schemes, and knowing what kind of signal to expect before entering the cryo-workflow can substantially improve its success rate.
In the workflow presented here, a stand-alone fluorescence microscope with a cryo-stage is used to image the samples, followed by a transfer of the grids into the focused ion beam microscope. However, it has been tested on systems where a fluorescence microscope is integrated into the FIB-SEM chamber, and therefore no sample transfer is required to acquire fluorescence images29,30,31. Using such integrated systems, positions of interest can be imaged during and after FIB milling to check for the presence of the target fluorescence signal without increasing the risk of contaminating the final lamellas. It is, however, important to keep in mind the optical parameters of the used microscopes, as, e.g., a low NA objective will limit the precision with which fiducial beads and target signals can be localized. Nonetheless, integrated FLM setups will help to also better deal with slight deformations of grids and lamellas, as FLM stacks can continuously be updated and compared to up-to-date SEM and IB views.
As an alternative to fluorescence imaging of the lamella between FIB milling and TEM data acquisition, post-TEM correlation can be used to verify correct placement and milling of the lamellas5,6.
During all steps of the correlative workflow, but especially during TEM, it is recommended to create an overlay of the projected fluorescence data on the FIB-SEM/TEM images. Such classical CLEM views help understand more intuitively which part of the cells is contained within the lamellas. This also serves as a useful sanity check to verify the accuracy of the correlation.
3. Limitations of the method
The 3D-correlative FIB approach requires samples that can be supplied with fiducial beads. Accordingly, this method is presently restricted to plunge-frozen grids. For high-pressure (HPF) frozen (tissue) samples, presently, only 2D-2D correlations can be performed. Potentially, internal fiducial markers (e.g., organelles, stained lipid droplets) could be a solution to this problem32,33. The final correlation success rate depends on many factors, including the sample quality, the fluorescence microscopy setup, the lamella thickness, and the size of the targeted structure. The correlation accuracy using the described 3D registration approach is estimated to be in the range of 200-300 nm on the final IB image, roughly corresponding to the typical thickness of FIB-milled lamellas7. Accordingly, cellular structures much smaller than this will be hard to target at present. Additionally, excessive movement at the milling site (&#62;300 nm) also reduces the accuracy of the correlation, an issue that can potentially be addressed with FLM setups integrated into FIB/SEM instruments. Lamellas that show strong deformation or bending during milling should, in any case, be excluded from the downstream workflow.
Overall, cryo-fluorescence imaging is currently limited by the Abbe diffraction criterion. With more routine application (and commercialization) of super-resolved cryo-FLM methods, more accurate targeting of cellular structures might become possible, especially when integrated into the FIB/SEM for on-the-fly operation.
4. Significance of the method
Especially in comparison to non-targeted and post-correlation techniques, the 3D-correlated FIB milling approach allows the selection of suitable positions before the time- and resource-consuming TEM step. It, thereby, enables more efficient data collection and project planning. Moreover, the correlated fluorescence data adds a layer of information that can be crucial for interpreting the tomograms and for integrating the cryo-ET results in multi-scale projects, especially when dealing with non-structured protein assemblies or those too small for template matching and subtomogram averaging.
5. Importance and potential future applications
In combination with advanced workflows such as cryo-lift out of HPF samples34,35, cryo-FIB-SEM volume36 and super-resolution fluorescence imaging26,37,38,39, 3D-targeted lamella preparation offers the prospect of not only dissecting biological processes in isolated cells but also to make tissue and patient samples accessible to FIB milling and cryo-electron tomography. As such, it will allow dissection of pathological processes at high resolution and thus be an integral building block toward a biopsy at the nanoscale.

Focused ion beam milling allows the preparation of thin biological samples from cryo-fixed specimens without the problems commonly associated with mechanical sectionings such as knife marks and compression artifacts1. When paired with cryo-electron tomography, FIB milling enables high-resolution biological studies of the cellular morphology and determination of the structure of macromolecular complexes directly from within cells at sub-nanometer resolution2,3,4. While abundant species, such as ribosomes, are readily found in randomly cut FIB lamellas, many cellular processes rely on the colocalization of several complexes or are localized to specific sites within the cell. Consequently, efficient targeting is required to not lose the biological feature of interest during the milling process and be limited to random hits. A correlative approach that combines data from the scanning electron microscope (SEM)-FIB and a cryo-fluorescence light microscope (FLM) is therefore necessary. While it is possible to omit the initial correlation and combine FLM and cryo-ET data only after TEM acquisition5,6, fluorescence-guided focused ion beam milling enables an accurate selection of the milling area beforehand, thereby resulting in more efficient data acquisition. Since its conception7, the application of 3D-correlated FIB milling in biological studies had been limited until we recently reported identifying a new liquid-liquid phase-separated (LLPS) compartment in yeast using this technique8.
Described here is a generalized 3D cryo-correlated light and electron microscopy (CLEM) protocol, which can be used to study a broad variety of samples ranging from bacteria to yeast and mammalian cells. While the experiments were performed using a certain set of instruments, the individual steps are not bound to specific hardware and can easily be transferred to other systems as an extension to existing protocols3,5. A list of tested equipment and suggested settings are provided in the Table of Materials and Table 1. The four key steps of the pipeline are (1) sample preparation, (2) localization of features of interest by cryo-fluorescence microscopy, (3) 3D-correlated focused ion beam milling, and (4) localization of the targeted structures for cryo-ET data acquisition on the lamellas in the cryo-transmission electron microscope (Figure 1).

<table border="1" fo:keep-together.within-page="1" fo:keep-with-next.within-page="always">
	<tbody>
		<tr>
			<td>Autogrids</td>
			<td>Thermo Fisher Scientific / Homemade</td>
			<td>1036173 (no cutout), 1205101 (with cutout)</td>
			<td></td>
		</tr>
		<tr>
			<td>C-rings</td>
			<td>Thermo Fisher Scientific</td>
			<td>1036171</td>
			<td></td>
		</tr>
		<tr>
			<td>Corrsight with cryo module</td>
			<td>Thermo Fisher Scientific</td>
			<td></td>
			<td>FLM Alternative 1</td>
		</tr>
		<tr>
			<td>Dynabeads MyOne COOH</td>
			<td>Thermo Fisher Scientific</td>
			<td>65011</td>
			<td>recommended 1 &#181;m fiducial beads</td>
		</tr>
		<tr>
			<td>EM Grids R1/4 SiO2</td>
			<td>Quantifoil</td>
			<td>N1-S13nAu20-01</td>
			<td></td>
		</tr>
		<tr>
			<td>Falcon 4 camera w. post-column Selectris X energy filter</td>
			<td>Thermo Fisher Scientific</td>
			<td></td>
			<td>Camera/Filter Alternative 1</td>
		</tr>
		<tr>
			<td>FIB Aquilos 1</td>
			<td>Thermo Fisher Scientific</td>
			<td></td>
			<td>FIB Alternative 1</td>
		</tr>
		<tr>
			<td>FIB Aquilos 2</td>
			<td>Thermo Fisher Scientific</td>
			<td></td>
			<td>FIB Alternative 2</td>
		</tr>
		<tr>
			<td>FIB Quanta 3D FEG</td>
			<td>Thermo Fisher Scientific</td>
			<td></td>
			<td>FIB Alternative 3</td>
		</tr>
		<tr>
			<td>FIB Scios</td>
			<td>Thermo Fisher Scientific</td>
			<td></td>
			<td>FIB Alternative 4</td>
		</tr>
		<tr>
			<td>K2 summit camera w. post-column energy filter 968 Quantum K2</td>
			<td>Gatan</td>
			<td></td>
			<td>Camera/Filter Alternative 2</td>
		</tr>
		<tr>
			<td>Leica TCS SP8 with cryo module</td>
			<td>Thermo Fisher Scientific</td>
			<td></td>
			<td>FLM Alternative 2</td>
		</tr>
		<tr>
			<td>Plasma Cleaner PDC-3XG</td>
			<td>Harrick</td>
			<td></td>
			<td></td>
		</tr>
		<tr>
			<td>Teflon Sheet (0.25 mm)</td>
			<td>plastx24.de</td>
			<td>11645</td>
			<td>Cut to same dimensions as filter paper</td>
		</tr>
		<tr>
			<td>TEM Titan Krios XFEG 300 kV Gi2</td>
			<td>Thermo Fisher Scientific</td>
			<td></td>
			<td>TEM Alternative 1</td>
		</tr>
		<tr>
			<td>TEM Titan Krios XFEG 300 kV Gi4</td>
			<td>Thermo Fisher Scientific</td>
			<td></td>
			<td>TEM Alternative 2</td>
		</tr>
		<tr>
			<td>THUNDER Imager&#160; EM Cryo CLEM</td>
			<td>Thermo Fisher Scientific</td>
			<td></td>
			<td>FLM Alternative 3</td>
		</tr>
		<tr>
			<td>Vitrobot Mark IV</td>
			<td>Thermo Fisher Scientific</td>
			<td></td>
			<td>alternativevly, use manaual plunger</td>
		</tr>
		<tr>
			<td>Whatman filter paper</td>
			<td>Sigma Aldrich</td>
			<td>10311807</td>
			<td>55 mm diamater; needs to be cut to fit the Vitrobot</td>
		</tr>
	</tbody>
</table>

sample preparation by 3d-correlative focused ion beam milling for high-resolution cryo-electron tomography

Cryo-electron tomography (cryo-ET) has become the method of choice for investigating cellular ultrastructure and molecular complexes in their native, frozen-hydrated state. However, cryo-ET requires that samples are thin enough to not scatter or block the incident electron beam. For thick cellular samples, this can be achieved by cryo-focused ion beam (FIB) milling. This protocol describes how to target specific cellular sites during FIB milling using a 3D-correlative approach, which combines three-dimensional fluorescence microscopy data with information from the FIB-scanning electron microscope. Using this technique, rare cellular events and structures can be targeted with high accuracy and visualized at molecular resolution using cryo-transmission electron microscopy (cryo-TEM).

The protocol provides a walk-through of the pipeline used to discover the EH domain-containing and endocytosis protein 1 (Ede1)-dependent endocytic protein deposit (END) and its degradation and trapping in autophagic bodies8. The END is a liquid-liquid phase-separated compartment in S. cerevisiae, which buffers a variety of proteins involved in clathrin-mediated endocytosis (CME) after failed endocytic events. One of its main components is Ede1, which doubles as a CME component and as a selective autophagy receptor for the degradation of this new LLPS compartment. Accordingly, an EGFP fusion of Ede1 (EGFP-Ede1) under the control of the alcohol dehydrogenase (ADH) promoter was used to visualize ENDs since Ede1 overexpression interferes with the early stages of endocytosis and therefore constitutively induces LLPS.
On a plunge-frozen grid with EGFP-Ede1 overexpressing yeast cells and 1 &#181;m fiducial markers, five positions were selected for FLM stack acquisition in the GFP channel (Figure 6A; TFS Corrsight; confocal mode, 300 nm focus step size, 10 &#181;m range). The grid was transferred to the FIB instrument (Quanta 3D FEG), and the grid squares for which FLM stacks had been acquired were identified by performing a 2D-2D correlation of the fluorescence and SEM grid overviews (compare step 3.2).
For each of the chosen squares, ion beam images were taken at a low current (10 pA, 1200x magnification), and corresponding fiducial positions were registered in 3DCT. After selection of positions with the biological feature of interest and fitting their 3D position within the FLM stack, the found transformation was applied to putative END positions, and the sites for lamella preparation were selected (Figure 6B). A FIB beam inclined 11&#176; relative to the grid surface was used in the examples shown here (45&#176; FIB shuttle pre-tilt; 18&#176; stage tilt). Positions of interest were transferred, and FIB patterns were drawn manually (Figure 6D) by measuring the distance of the predicted positions relative to prominent landmarks in the FIB image (e.g., holes, ice contaminations, fiducial beads). The accuracy of the registration was evaluated by deliberately leaving out beads that could clearly be identified in the FLM and IB image, and then comparing their actual and predicted positions in the ion beam view (e.g., diamond in Figure 6B,C). The correlation for the square shown in Figure 6C was found to be accurate (i.e., the predicted position of FLM bead positions perfectly coincided with their corresponding IB location and 3DCT reported sub-pixel RMSE values for the registration). Thus, lamellas were cut at the predicted positions (Site B) and fine-milled to a thickness of ~200 nm (final pattern offset).
<img alt="Figure 6" class="xfigimg" src="/files/ftp_upload/62886/62886fig06.jpg" /> 
Figure 6: Representative results for 3D-correlative targeting of endocytic protein deposits (END) in yeast. (A) SEM overview of the grid before milling. The colored boxes indicate grid squares for which fluorescence stacks were taken beforehand. (B-C) 3D correlation in a grid square. After registering several corresponding fiducial beads (colored boxes) in the FLM data (B, shown here as maximum intensity projection) and the ion beam image (C), the accuracy of the 3D registration was verified by predicting the position of the bead indicated with the diamond. Next, the positions of the target signal (red circles) were predicted in the ion beam view for two potential milling sites. (D) Zoom-in of Site B showing the predicted positions of three target puncta (red circles) and the initial milling patterns (yellow boxes). A fourth fluorescence punctum was predicted to be much lower than the other puncta and therefore not targeted during milling (gray circle). <a href="https://www.jove.com/files/ftp_upload/62886/62886fig06large.jpg" target="_blank">Please click here to view a larger version of this figure.</a>
After successful FIB milling and transfer of the grid to the cryo-transmission electron microscope (Titan Krios operated at 300 kV and equipped with a Gatan K2 direct electron detector and Bioquantum energy filter), a grid overview was recorded in SerialEM and used to locate squares with lamellas. For each lamella, overview images were acquired, and the FLM data was registered in 3DCT (3D-2D) using corresponding fiducial beads. Positions of the biological features of interest (Figure 7A) were then predicted using the transformation calculated from the fiducial beads. Lamella overviews recorded at higher magnification were stitched, and sites of interest correlated using clearly visible landmarks (e.g., fiducial beads). Alternatively, classical CLEM overviews can be produced in various softwares10,12.
Based on the correlation, four potential sites for tomogram acquisition were found for the lamella shown in Figure 7A. However, this also includes a position that was not targeted during the 3D-correlated FIB-milling (compare Figure 6D; gray circle) and a position blocked by ice contamination (Figure 7; gray boxes). Accordingly, tomograms could only be recorded for two positions (Figure 7B). Overall, a correlation success of ~75%, i.e., lamellas that survived the transfer to the TEM and END structures were found at the predicted sites, was achieved (12 correlated sites). After tomogram reconstruction, segmentation and template matching, individual END structures can be visualized within their native context (Figure 7C,D). This includes the fenestrated endoplasmic reticulum (ER) surrounding the END, lipid droplets occasionally making contact, and ribosomes, which are excluded from the LLPS compartment. Taken together, this shows how 3D-correlative FIB milling can provide molecular-level information of rare biological processes from intact cells.
<img alt="Figure 7" class="xfigimg" src="/files/ftp_upload/62886/62886fig07.jpg" /> 
Figure 7: Representative results for visualizing the END with cryo-ET. (A) The low magnification TEM overview of the milling site shown in Figure 6 can readily be correlated with the FLM maximum intensity projection (Figure 6B) to localize biological features of interest (red crosses). (B) In a second step, a higher magnification (stitched) view can be correlated, and positions for tomogram acquisition (yellow boxes) are set up. Locations resulting from the out-of-plane signal (gray box, compare Figure 6D) were ignored. (C-D) Using this 3D-correlative FIB approach, the endocytic protein deposit (END) can be visualized in its native environment. Structures such as the endoplasmic reticulum (ER), ribosomes, membranes, and lipid droplets can be identified and visualized. <a href="https://www.jove.com/files/ftp_upload/62886/62886fig07large.jpg" target="_blank">Please click here to view a larger version of this figure.</a>
<table border="1" fo:keep-together.within-page="1" fo:keep-with-next.within-page="always">
	<tbody>
		<tr>
			<td>Plasma Cleaner Settings</td>
			<td colspan="3"></td>
		</tr>
		<tr>
			<td>Harrick Plasma Cleaner PDG-3XG :</td>
			<td colspan="3">Radio Frequency setiing: &#34;HI&#34;, 30 s; N2 plasma</td>
		</tr>
		<tr>
			<td>Plunger Settings</td>
			<td colspan="3"></td>
		</tr>
		<tr>
			<td>TFS Vitrobot Mk IV:</td>
			<td colspan="3">100% humidity; blotforce = 8; blottime = 10 s; wait time 0 s; (this should work for most suspension and adherent cells)</td>
		</tr>
		<tr>
			<td>FIB GIS Positions and Timings</td>
			<td colspan="3"></td>
		</tr>
		<tr>
			<td>Quanta 3D FEG:</td>
			<td colspan="3">Tilt = 0, Rotation = -180,&#160; Z position = 13.5, Temperature setpoint =26.15&#176; , Time = 8 s</td>
		</tr>
		<tr>
			<td>TFS Scios:</td>
			<td colspan="3">Tilt = 0, Rotation = -180, Z position = 9.8, Temperature setpoint =28&#176; C, Time = 7 s</td>
		</tr>
		<tr>
			<td>TFS Aquilos 1:</td>
			<td colspan="3">Software predefined position, Temperature setpoint = 28&#176; , Time = 7 s</td>
		</tr>
		<tr>
			<td>TFS Aquilos 2:</td>
			<td colspan="3">Software predefined position, Temperature setpoint = 28&#176;, Time = 7 s</td>
		</tr>
		<tr>
			<td>FIB Sputter Coater Settings</td>
			<td colspan="3"></td>
		</tr>
		<tr>
			<td>Quorum System:</td>
			<td colspan="3">In Quorum prep chamber: 10 mA, 40 s</td>
		</tr>
		<tr>
			<td>TFS Scios:</td>
			<td colspan="3">10 W, 500 V, 250 mA, 0.2 mbar, 15 s</td>
		</tr>
		<tr>
			<td>TFS Aquilos 1:</td>
			<td colspan="3">1kV, 10 mA, 10 Pa, 15 s</td>
		</tr>
		<tr>
			<td>TFS Aquilos 2:</td>
			<td colspan="3">1kV, 10 mA, 10 Pa, 15 s</td>
		</tr>
		<tr>
			<td>Tomogram Acquisition</td>
			<td colspan="3"></td>
		</tr>
		<tr>
			<td>Titan Krios Gi2</td>
			<td colspan="3">K2 camera, Gatan Bioquantum energy filter</td>
		</tr>
		<tr>
			<td></td>
			<td colspan="3">20 eV slit; dose symmetric tilt scheme (Hagen) with 2&#176; steps; start at +10&#176; (lamella pre-tilt!) to +70&#176;and -50&#176;</td>
		</tr>
		<tr>
			<td>Titan Krios Gi4</td>
			<td colspan="3">Falcon 4; Selectris X energy filter</td>
		</tr>
		<tr>
			<td></td>
			<td colspan="3">10 eV slit; dose symmetric tilt scheme (Hagen) with 2&#176; steps; start at +10&#176; (lamella pre-tilt!) to +70&#176; and -50&#176;</td>
		</tr>
		<tr>
			<td>FLM Acquisition</td>
			<td colspan="3"></td>
		</tr>
		<tr>
			<td>Corrsight (Confocal Mode)</td>
			<td colspan="3">Objective: Zeiss EC Plan-Neofluar 40&#215;/0.9 NA Pol; Stack acquisition parameters: x-y pixel size = 161.25 nm, z step size = 300 nm.</td>
		</tr>
		<tr>
			<td>Leica SP8 Cryo-Confocal</td>
			<td colspan="3">Objective: Leica HCX PL APO 50x / 0.90 CLEM; Stack acquisition parameters: x-y pixel size = 84 nm, z step size = 300 nm.</td>
		</tr>
	</tbody>
</table>
Table 1: List of tested equipment and suggested settings.

Watch this Scientific Journal Video about Sample Preparation by 3D-Correlative Focused Ion Beam Milling for High-Resolution Cryo-Electron Tomography at JoVE.com

Sample Preparation by 3D-Correlative Focused Ion Beam Milling for High-Resolution Cryo-Electron Tomography

Here, we present a pipeline for 3D-correlative focused ion beam milling on guiding the preparation of cellular samples for cryo-electron tomography. The 3D position of fluorescently tagged proteins of interest is first determined by cryo-fluorescence microscopy, and then targeted for milling. The protocol is suitable for mammalian, yeast, and bacterial cells.

Cryo-electron tomography (cryo-ET) has become the method of choice for investigating cellular ultrastructure and molecular complexes in their native, ...

Research

JoVE Journal

Biology

MALDI Sample Preparation: the Ultra Thin Layer Method

This video demonstrates the preparation of an ultra-thin matrix/analyte layer for analyzing peptides and proteins by Matrix-Assisted Laser Desorption ...

Using High Resolution Computed Tomography to Visualize the Three Dimensional Structure and Function of Plant Vasculature

High resolution x-ray computed tomography (HRCT) is a non-destructive diagnostic imaging technique with sub-micron resolution capability that is now being ...

Cryo-electron Microscopy Specimen Preparation By Means Of a Focused Ion Beam

Here we present a protocol used to prepare cryo-TEM samples of Aspergillus niger spores, but which can easily be adapted for any number of microorganisms ...

Do's and Don'ts of Cryo-electron Microscopy: A Primer on Sample Preparation and High Quality Data Collection for Macromolecular 3D Reconstruction

Cryo-electron microscopy (cryoEM) entails flash-freezing a thin layer of sample on a support, and then visualizing the sample in its frozen hydrated state ...

Sample Preparation for Mass Cytometry Analysis

Mass cytometry utilizes antibodies conjugated with heavy metal labels, an approach that has greatly increased the number of parameters and opportunities ...

Preparation and Observation of Thick Biological Samples by Scanning Transmission Electron Tomography

This report describes a protocol for preparing thick biological specimens for further observation using a scanning transmission electron microscope. It ...

Isolation of Intact Mitochondria from Skeletal Muscle by Differential Centrifugation for High-resolution Respirometry Measurements

Mitochondria are involved in cellular energy metabolism and use oxygen to produce energy in the form of adenosine triphosphate (ATP). Differential ...

Sample Preparation and Imaging of Exosomes by Transmission Electron Microscopy

Exosomes are nano-sized extracellular vesicles secreted by body fluids and are known to represent the characteristics of cells that secrete them. The ...

Visualization of DNA Compaction in Cyanobacteria by High-voltage Cryo-electron Tomography

This protocol describes how to visualize the transient DNA compaction in cyanobacteria. DNA compaction is a dramatic cytoplasmic event recently found to ...

Miniaturized Sample Preparation for Transmission Electron Microscopy

Due to recent technological progress, cryo-electron microscopy (cryo-EM) is rapidly becoming a standard method for the structural analysis of protein ...