As an example, we will discuss how to perform subtomogram
averaging using PEET, a routine procedure we do in our group.
Open the tomogram using the command 3dmod and activate the
“Toggle between regular and high-resolution image” and “keep
current image or model point centered” buttons. Activate the
“Model” mode in the 3dmod window and set the slicer thickness
to ten and then go to “Edit” and select “Angles”. A new window
will appear where the coordinates of each one of the chosen particles can be viewed. Find the optimal view of the complex of interest
by rotating the particle and moving through the different slices of
the tomogram and then save these coordinates and angles. Create a
model point for each particle with a middle mouse click and then
save this model.
Before averaging, compute an initial motive list (MOTL) by
running the command “stalkInit name.mod.” Usually, we create a
separate folder for the PEET run and then subfolders entitled
“run1”, “run2”, etc. In “run1”, load all the tomograms, the
MTOL files, and the “.tlt” files (which have the tilt angles required
for missing wedge compensation). In the first run, use one of the
good particles as a reference without a mask. Subsequently, use the
obtained structure from each run as a reference for the next run.
Different masks can be constructed in IMOD to perform focused
alignment on different parts of the protein complex. This process is
repeated until the structure cannot be improved anymore. If the
symmetry of the complex is known, this symmetry can be imposed
on the structure; however, one should bear in mind that different
parts of the complex might have different symmetries and imposing
one specific symmetry might affect other parts of the complex that
have different symmetries. If the symmetry is unknown, we usually
apply two-fold symmetry along the mirror plane to improve the
signal-to-noise of the side view (checking that no new densities
appear compared to the non-symmetrized structure). Note that
when two-fold symmetry is applied to a complex with unknown
symmetry, no conclusions should be made on the top view of that
average.
5 Notes
1. The use of phenol red-free media is advised to prevent the
generation of unwanted autofluorescence at 80 K when imaging the cells using cryo-LM [53].
2. Over-digestion with trypsin may prevent the cells from adequately attaching and spreading on the grid.
3. Different types of films (silicon vs. carbon) and shapes of grids
support (square vs. hexagonal) are available. For U2OS cells,
Methods in Cryo-Electron Tomography
107
averaging using PEET, a routine procedure we do in our group.
Open the tomogram using the command 3dmod and activate the
“Toggle between regular and high-resolution image” and “keep
current image or model point centered” buttons. Activate the
“Model” mode in the 3dmod window and set the slicer thickness
to ten and then go to “Edit” and select “Angles”. A new window
will appear where the coordinates of each one of the chosen particles can be viewed. Find the optimal view of the complex of interest
by rotating the particle and moving through the different slices of
the tomogram and then save these coordinates and angles. Create a
model point for each particle with a middle mouse click and then
save this model.
Before averaging, compute an initial motive list (MOTL) by
running the command “stalkInit name.mod.” Usually, we create a
separate folder for the PEET run and then subfolders entitled
“run1”, “run2”, etc. In “run1”, load all the tomograms, the
MTOL files, and the “.tlt” files (which have the tilt angles required
for missing wedge compensation). In the first run, use one of the
good particles as a reference without a mask. Subsequently, use the
obtained structure from each run as a reference for the next run.
Different masks can be constructed in IMOD to perform focused
alignment on different parts of the protein complex. This process is
repeated until the structure cannot be improved anymore. If the
symmetry of the complex is known, this symmetry can be imposed
on the structure; however, one should bear in mind that different
parts of the complex might have different symmetries and imposing
one specific symmetry might affect other parts of the complex that
have different symmetries. If the symmetry is unknown, we usually
apply two-fold symmetry along the mirror plane to improve the
signal-to-noise of the side view (checking that no new densities
appear compared to the non-symmetrized structure). Note that
when two-fold symmetry is applied to a complex with unknown
symmetry, no conclusions should be made on the top view of that
average.
5 Notes
1. The use of phenol red-free media is advised to prevent the
generation of unwanted autofluorescence at 80 K when imaging the cells using cryo-LM [53].
2. Over-digestion with trypsin may prevent the cells from adequately attaching and spreading on the grid.
3. Different types of films (silicon vs. carbon) and shapes of grids
support (square vs. hexagonal) are available. For U2OS cells,
Methods in Cryo-Electron Tomography
107