3D-correlative FIB milling makes specific and even rare cellular events accessible to Cryo-TEM. Allowing the revelation of this event's ultra structure directly within the native environment. The correlative workflow improves the success rate of targeting rare cellular structures, and helps identify them unambiguously in the crowded cellular environment.
Demonstrating the procedure will be Anna Bieber and Christina Capitanio, experts in correlative cryo-electron tomography. Begin the cryo-fluorescence microscopy of the cell seeded grids by acquiring a wide field overview in fluorescence and reflected"mode. Then, select suitable grid squares with fluorescence signal.
After ensuring that the cells and beads are evenly distributed toward the center of each square, choose the squares on 200 mesh grids that are accessible both to the FIB-SEM and TEM instrument, and are at least three squares away from the grid edge. On each of the selected grid squares, acquire a fluorescent stack with an appropriate focused step. High numerical aperture objectives should be used to increase photon count and localization accuracy.
On a confocal microscope with numerical aperture 0.9 objective, acquire stacks with 300 nanometer step size and oversampling the Nyquist value. Record multiple color stacks and store grids under liquid nitrogen until further use. Use either the cutout or orientation marks to ensure proper orientation of the grid for later placement into the TEM.
Load the grids into the cryo FIB-SEM instrument. Ensure that the milling direction is perpendicular to the tilt axis of the TEM. To coat the grids with a protective organometallic layer, use a plasma coder and gas injection system at the stage positions predefined by the FIB-SEM setup.
Next, record the SEM grid overview and perform a 2D correlation with FLM overviews. Find the grid squares by manually inspecting recorded grid overviews or using a software package. Select and mark four corresponding positions for the squares in the FLM and SEM grid overview and calculate the transformation between the marked points for FLM-SEM correlation.
Then, place the markers in the center of the corresponding grid squares for which FLM stacks have been acquired before predicting their position in the SEM view. For each correlated grid square, take a low current ion beam image at the FIB milling angle of choice and select a field of view with the desired position and magnification that matches the fluorescence data. For 200 mesh grids, acquire FIB-SEM data to contain single grid squares, including the grid bars.
Then, take a SEM image of the same square to help with the identification of corresponding beads in the fluorescence and ion beam view. Perform registration of the raw or deconvoled 3D FLM stack and the 2D ion beam view for each position with the 3D correlation toolbox, 3D CT, by loading the corresponding re-sliced 3D FLM stack and ion beam view in 3D CT.In the fluorescence data, select four fiducial beads and right click on the positions list to determine 3D position of the beads via Gaussian fitting of the signal in X, Y and Z.Then, select the corresponding beads in the ion beam image to perform an initial 3D correlation. Similarly, add more beads to the registration to check the accuracy of the correlation.
Leave out some fiducial beads clearly identifiable in both fluorescence and ion beam by checking their predicted versus actual location in the ion beam image. In 3D CT, root mean square error"or RMSE values, can be viewed to assess the correlation consistency. Ensure that the RMSE values are small and on the order of localization accuracy.
Next, select the targeted cellular signals and fit the 3D position in the FLM stack before applying the transformation to predict the target positions in the ion beam view. For each correlated square, transfer the predicted positions of the features of interest to the FIB-SEM instrument to place the lamella milling patterns. If there are multiple signals per cell, place the patterns to include the maximum possible points of interest in the same lamella.
Rough mill the lamellas followed by fine milling to get a final thickness of 150 to 250 nanometers. Ensure that the lamella orientation is perpendicular to the tilt axis when loading the grids into the transmission electron microscope. Acquire grid montage and overviews for each grid square containing lamellas with suitable magnification and exposure time to visualize the fiducial beads in the TEM images without significantly adding to the total electron dose.
Acquire high resolution TEM maps of each lamella. Register and 3D-2D correlate the FLM stack with the TEM grid square and the lamella overviews in 3D CT.Perform correlation between FLM and TEM from low to high magnification in TEM. After transferring the positions, set-up and run tilt series at correlated positions"then use an appropriate magnification, defocus and total dose to start the image acquisition at the pre-tilt determined by the lamella using a dose symmetric tilt scheme.
Follow either manual or batch acquisition of images. In the representative analysis, the low magnification TEM overview of the milling site is shown to localize biological features of interest. Later, a higher magnification view was correlated with FLM maximum intensity projection and positions for tomogram acquisition were set up.
In the acquired images, the endocytic protein deposit could be visualized in its native environment. The optimization of the sample is very important. The grid preparation should be optimized to have a good distribution of cells to mill and a good number of visible fiducial beads in most grid squares.
3D correlative milling has helped to reveal the ultra structure of different cellular processes such as the step-by-step progression of autophagy in each cell, and it's used to capture a novel phase separated compartment.
