This protocol demonstrates how to image biological cryo-preserved samples using cryo-structured illumination microscopy. We demonstrate the methodology by imaging the cytoskeleton of U2OS cells.
Three-dimensional (3D) structured illumination microscopy (SIM) allows imaging of fluorescently labelled cellular structures at higher resolution than conventional fluorescence microscopy. This super-resolution (SR) technique enables visualization of molecular processes in whole cells and has the potential to be used in conjunction with electron microscopy and X-ray tomography to correlate structural and functional information. A SIM microscope for cryogenically preserved samples (cryoSIM) has recently been commissioned at the correlative cryo-imaging beamline B24 at the UK synchrotron.
It was designed specifically for 3D imaging of biological samples at cryogenic temperatures in a manner compatible with subsequent imaging of the same samples by X-ray microscopy methods such as cryo-soft X-ray tomography. This video article provides detailed methods and protocols for successful imaging using the cryoSIM. In addition to instructions on the operation of the cryoSIM microscope, recommendations have been included regarding the choice of samples, fluorophores, and parameter settings. The protocol is demonstrated in U2OS cell samples whose mitochondria and tubulin have been fluorescently labelled.
SR imaging techniques have become widely accessible to biologists over the last decade1. They allow high-resolution imaging of fluorescently tagged samples beyond the diffraction limit. However, it has been challenging to adapt SR microscopy methods to work with samples at cryogenic temperatures2. This would be advantageous for correlative imaging in combination with electron or X-ray tomography. Recently, SIM has been adapted for use with cryogenic samples and has successfully been shown to enable correlative studies of biological cells in conjunction with soft X-ray tomography (SXT)3 at the correlative cryo-imaging beamline B24 at the Diamond Light Source Synchrotron (https://www.diamond.ac.uk/Instruments/Biological-Cryo-Imaging/B24.html). SIM can double the resolution of conventional wide-field microscopy by illuminating the sample with striped patterns of light (Moiré fringes) at three angles and in five phases. The interference between these light patterns and the sample fluorescence can be used to computationally uncover extra information about sub-diffraction structures4,5.
There are several advantages of SIM over other SR techniques for cryogenic applications. First, it can work without specially designed blinking fluorophores; conventional fluorophores can be used, giving access to a wider range of potential fluorescent tagging agents6. In addition, it only requires 15 images per z slice (in 3D; 9 images for 2D), whereas other SR methods take approximately 1000 images per slice, increasing the chance of the sample being heated and therefore increasing the risk of ice crystal formation, which can cause artefacts. Finally, this technique can image thicker biological samples of over 10 µm, allowing whole cells to be imaged in their near-native state6. The cryoSIM has been built using standard optical components and with open-access software for imaging, making it easy to document and duplicate if desired6. The cryoSIM has a 100x/0.9 numerical aperture objective (see the Table of Materials); further information on its optical components, design parameters, and performance has been described by Phillips et al.6 Here, this protocol demonstrates how to use the cryoSIM microscope including how to load and unload samples on the cryogenic stage, how to collect data on the microscope, and how to reconstruct the SIM images.
NOTE: This protocol pertains to samples containing cells grown or deposited on transmission electron microscopy (TEM) 3 mm flat gold grids with a holey carbon support film that have been vitrified by plunge freezing or high-pressure freezing. This protocol assumes that samples have already been imaged using a conventional epifluorescence and brightfield microscope to map locations of interest for imaging in cryoSIM. See Figure 1 for an overview of the entire protocol.
1. Preparation of the cryo-stage
2. Transfer of the sample storage box into the cryo-stage
NOTE: Immerse the sample storage box, holder, and the tips of any instruments (e.g., forceps) inside filtered LN2 to cool them before touching any cold surfaces such as the sample or any objects inside the sample chamber. Wear a laboratory coat and gloves when handling biological samples.
3. Stage docking and focusing
4. Brightfield mosaic acquisition
5. Identification of areas of interest
6. Data collection strategy
7. Data collection
8. After imaging
9. Reconstruction
10. Chromatic shift correction
A sample containing U2OS cells was stained with a mixture of green microtubule cytoskeleton dye and red mitochondria dye, resulting in the staining of the microtubule component of the cytoskeleton (green) and the mitochondria (red). Subsequent imaging showed the localization of mitochondria within the cell as well as the arrangement of the microtubules, highlighting the structural framework that they provide to the cell and the assembly of the cytoskeleton around organelles such as the nucleus. The resolution in cryoSIM is significantly higher than that in standard epifluorescence microscopy (Figure 5). Figure 6 demonstrates how the fluorescent "map" from a conventional epifluorescence microscope can be used to locate areas of interest for imaging and the corresponding cryoSIM-reconstructed image from a location on the grid.
Figure 1: Flow chart showing the stages of the cryoSIM imaging protocol. Abbreviation: cryoSIM = Cryo-structured illumination microscopy. Please click here to view a larger version of this figure.
Figure 2: The cryo-stage. (A) The cryo-stage setup. (B) A sample grid shown held by inverted forceps. (C) Components of the cryo-stage. The connection ports are labelled, with the colors corresponding to orange: power supply, yellow: heated stage lid, blue: external dewar, green: connection to PC. Abbreviations: PC = personal computer; LN2 = liquid nitrogen. Please click here to view a larger version of this figure.
Figure 3: Views of the Cockpit software panels. (A) Main panel, (B) macro stage XY, (C) mosaic view, (D) camera views. Abbreviation: SLM = spatial light modulator. Please click here to view a larger version of this figure.
Figure 4: Views of the Cockpit software panels. (A) Z stack single site experiment. (B) SI single site experiment. (C) Keyboard shortcuts for the cockpit software used during image acquisition. (D) The cryoSIM microscope is on-site at beamline B24 at the Diamond Light Source synchrotron. Abbreviation: cryoSIM = Cryo-structured illumination microscopy. Please click here to view a larger version of this figure.
Figure 5: Resolution of cryoSIM. (A) Mosaic view of a grid under examination. (B) Brightfield image of an Area of Interest (AOI). (C) Pseudo-widefield image compared to its (D) SIM image showing the increase in resolution. The white arrow indicates SIM reconstruction artefacts. (E) Modulation contrast map combining the pixel intensity information of the reconstructed image with the color information of the respective modulation contrast-to-noise ratio (MCNR) values of the raw data generated by SIMCheck2. The bright and dark regions show high and low contrast, respectively. Scale bar = 10 µm. CryoSIM imaging settings: excitation/emission wavelengths: 488/525 nm, 50 mW laser power, 50 ms exposure time and 647/655 nm, 20 mW laser power, 5 ms exposure time. Abbreviation: cryoSIM = Cryo-structured illumination microscopy. Please click here to view a larger version of this figure.
Figure 6: Image reconstruction in cryoSIM from a location on the grid in a fluorescence map from a conventional epifluorescence map. (A,B) Overlay of brightfield and fluorescence image maps generated with a conventional epifluorescence microscope. This map is used to locate regions of interest to subsequently image in cryoSIM. (C) The reconstructed cryoSIM image obtained at the location shown in (B). Abbreviation: cryoSIM = Cryo-structured illumination microscopy. Please click here to view a larger version of this figure.
3D SIM at cryogenic temperatures has many advantages over other SR imaging techniques for imaging vitrified biological material. It requires significantly fewer images per z slice compared to other SR methods, resulting in less irradiation and a lower chance of ice crystal formation for vitrified samples. It is also able to image whole cells and can be correlated with X-ray tomography to match structure with function. Interestingly, most commercially available fluorophores and fluorescence tags bleach less under cryogenic conditions than at room temperature. However, given the high quantum yield of most common fluorophores at room temperature (more than 80% in some cases), the absolute gain detected in photons is not due to changes in quantum yield, but due to a reduction in the complex bleaching processes. More information on the yield of fluorophores at cryogenic temperatures can be found in 8.
It is critical that samples arriving at the cryoSIM have been premapped using a conventional cryofluorescence microscope with brightfield capability to produce a grid "map" that includes highlighting of all potential AOIs for further imaging (Figure 5). Access time at the cryoSIM is allocated via a competitive process that involves the submission of a proposal, which is subsequently evaluated for technique feasibility and biomedical impact. Time at the equipment, therefore, is always "at a premium", and premapped grids allow the most efficient use of an allocation. It is also essential that the sample is kept vitrified, especially during sample transfer from the sample holder to the imaging platform, to minimize the formation of ice crystals and subsequent sample damage. The sample should be of good quality to produce the best SIM images. A well-prepared sample will be characterized by the following features: (a) it will have no ice crystal contamination, (b) the grid used will be a finder grid, (c) the carrier will be flat, (d) the grid mesh and substrate surface will not be auto-fluorescent, and (e) there will be no breaks in the support membrane. These prerequisites can be achieved by careful sample handling and ensuring that samples always stay vitrified.
It is important to check in advance whether proposed sample fluorophores will give enough signal in the cryoSIM microscope. Tools such as SPEKCheck9 can aid with choosing the optimal fluorophore and filter combinations. If there are issues with the raw data collection or the reconstruction process, artefacts can appear in the images after reconstruction. Examples of various artefacts have been documented by Demmerle et al.10 The image reconstruction parameters can be reviewed in the SoftWoRx log file if the reconstruction is not optimal by opening the reconstruction summary file. There should be consistent line spacing across angles in a given channel and relatively consistent amplitude. Variation of more than 30% and values significantly above 1 (if bead size compensation is applied) should be more closely investigated and are likely to indicate failed reconstructions. In addition, the SIMcheck2 software in Fiji can also be used to perform various checks on the raw and reconstructed data to diagnose the cause of errors in the imaging or reconstruction parameter settings. SIM-check and its modulation contrast map can also aid in the assessment of the quality of reconstructed data by interpreting which areas of an image are likely to be real structures versus artefacts.
Low modulation contrast (shown by dark color, in Figure 5E) within the nuclear area means that this region is going to be more susceptible to reconstruction artifacts, therefore implying that the hash patterns shown in the nucleus could be classified (Figure 5D) as an artefact. Strong fluorescence signal areas are more likely to accurately reflect native structures in the processed data. In areas of weak signal where fluorophores are distributed over wider areas, such as the total surface of a vesicle, it is likely that real signal coexists with processing artefacts, and care should be taken in the interpretation of that data. After inspection of the full-range reconstructed data to ensure there are no strange artifacts, and that the background is generally Gaussian and centered near zero, the data is generally clipped at zero, or the modal value-the peak of the background signal-should be very near zero. This ensures that the dynamic range of the displayed image is not dominated by negative background artifacts. When a weaker signal is expected, extra care should be taken in analyzing the features and ensuring they are real structures rather than reconstruction artefacts.
There are some limitations of the imaging system. Because the sample stage is flat, samples with variable thickness or grids that are not flat are not ideal subjects for imaging. Additionally, if correlative imaging will be done using soft X-ray tomography, cells near grid boundaries should not be imaged as these will not be visible in the X-ray microscope during tilt series acquisition. Finally, the amount of blotting of the sample before plunge freezing has a significant impact on the imaging quality; too little blotting results in samples that are too thick, giving suboptimal SIM images with high background noise, while too much blotting can cause cells to become misshaped and therefore more susceptible to heat damage from the incident laser beam. In summary, cryoSIM is a powerful fluorescence microscopy tool for imaging biological samples in 3D in a near-native stage and has wide-ranging applications in many areas.
The authors have nothing to disclose.
This project has received funding from the European Commission Horizon 2020 iNEXT-Discovery project. I. M. Dobbie acknowledges funding from the Wellcome Trust (107457/Z/15/Z). This work was carried out with the support of the Diamond Light Source, instrument B24 (proposal BI25512). Our thanks to the staff at Micron and all our excellent users and collaborators for helping us establish the cryoSIM and its correlative potential.
Auto grids | FEI | ||
Autogrids | Thermo Fisher Scientific | 1036173 | |
BioTracker 488 Green Microtubule Cytoskeleton Dye | Sigma-Aldrich | SCT142 | |
Cockpit Software | Oxford University | https://github.com/MicronOxford/cockpit | |
cryo compatible polyurethane container | Jena bioscience | CC-FD-800 | |
Cryo TEM grid storage box | Thermo Fisher Scientific | Model#AutoGrid | |
cryo-SIM microscope | Custom made | N/A | Custom made, see following reference for the design: Michael A. Phillips, Maria Harkiolaki, David Miguel Susano Pinto, Richard M. Parton, Ana Palanca, Manuel Garcia-Moreno, Ilias Kounatidis, John W. Sedat, David I. Stuart, Alfredo Castello, Martin J. Booth, Ilan Davis, and Ian M. Dobbie, "CryoSIM: super-resolution 3D structured illumination cryogenic fluorescence microscopy for correlated ultrastructural imaging," Optica 7, 802-812 (2020). Has a Nikon TU Plan Apo 100x/0.9 NA. |
Cryostage system | Linkam Scientific Instruments | CMS196 | |
Fine tip surgical forceps | Ted Pella | 38125 | |
MitoTracker Deep Red FM | Thermo Fisher Scientific | M22426 | |
Python Microscope Software | Oxford University | https://www.python-microscope.org/ | |
Scientific dry paper wipes | Kimberly-Clark 7551 | 2 | |
SIM Reconstruction Software | softWoRx, GE Healthcare | Version 6.5.2 | |
StitchM Software | Diamond Light Source | https://github.com/DiamondLightSource/StitchM | |
TEM grids for samples | Quantifoil Micro Tools | G200F1 |