thickness increases and thus higher electron doses are required to
obtain images of adequate contrast [13, 26]. Furthermore, high tilt
images suffer more from effects of drift, doming, and defocus
gradients which reduce image quality and obtainable resolution
[27–31]. Thus, to obtain the best quality tilt series, the low tilt
images should be collected first when the radiation damage is
minimal.
Two different tilt-schemes have been developed to maximize
the quality and resolution of information contained in a tilt series.
The two tilt-schemes are called bidirectional and dose symmetric.
For the bidirectional tilt-scheme, image collection starts at a low
angle (e.g. 0
or À20
) and the tilt angle is gradually increased to
the maximum tilt angle (e.g. +60
). The stage is then reset to the
start tilt angle and images are recorded in the other direction
towards the negative maximum tilt angle (e.g. À60
). For the
dose symmetric scheme, tilt series recording starts at a low tilt
angle and alternates between an increasing negative and positive
tilt angle [32].
As the total electron dose applied to a tilt series needs to be
limited to prevent excessive radiation damage, individual images
have a low signal-to-noise ratio resulting in low contrast images.
The contrast and data quality of individual images can be increased
during data collection by using an energy filter which screens out
inelastically scattered electrons, and a camera system which directly
detects electrons [33–37]. Furthermore, phase plates can also be
used to increase contrast [20].
After tilt series recording, a 3D volume (tomogram) is computationally reconstructed. Two different algorithms are routinely
used: “weighted-back projection” and “simulative iterative reconstruction technique (SIRT)” [38, 39]. The later generates higher
contrast tomograms, which are beneficial for morphology studies.
The first has a proven record of maintaining high-resolution information and is the most popular algorithm especially if the final goal
is determining the protein structure.
The contrast of the final reconstructed tomogram can be further enhanced by the use of different imaging filters. These filters
can be applied to either the 2D projection images of the tilt series or
the final 3D volume. Cellular features, like cytoskeletal elements
and organelles, can be highlighted (segmented) in the tomogram
to generate a 3D model. This helps visualize complex 3D arrangements of proteins and membranes and allows for precise measurements [14]. Additionally, structures of proteins can be obtained by
applying a process called subtomogram averaging [40]. Typical
resolutions of the protein structures obtained by subtomogram
averaging are in the range of 10–40 A ˚ [18, 41, 42]. However, it is
possible to obtain protein structures at significantly higher resolution [29, 43].
Cryo-ET of Cellular Structures
5
obtain images of adequate contrast [13, 26]. Furthermore, high tilt
images suffer more from effects of drift, doming, and defocus
gradients which reduce image quality and obtainable resolution
[27–31]. Thus, to obtain the best quality tilt series, the low tilt
images should be collected first when the radiation damage is
minimal.
Two different tilt-schemes have been developed to maximize
the quality and resolution of information contained in a tilt series.
The two tilt-schemes are called bidirectional and dose symmetric.
For the bidirectional tilt-scheme, image collection starts at a low
angle (e.g. 0
or À20
) and the tilt angle is gradually increased to
the maximum tilt angle (e.g. +60
). The stage is then reset to the
start tilt angle and images are recorded in the other direction
towards the negative maximum tilt angle (e.g. À60
). For the
dose symmetric scheme, tilt series recording starts at a low tilt
angle and alternates between an increasing negative and positive
tilt angle [32].
As the total electron dose applied to a tilt series needs to be
limited to prevent excessive radiation damage, individual images
have a low signal-to-noise ratio resulting in low contrast images.
The contrast and data quality of individual images can be increased
during data collection by using an energy filter which screens out
inelastically scattered electrons, and a camera system which directly
detects electrons [33–37]. Furthermore, phase plates can also be
used to increase contrast [20].
After tilt series recording, a 3D volume (tomogram) is computationally reconstructed. Two different algorithms are routinely
used: “weighted-back projection” and “simulative iterative reconstruction technique (SIRT)” [38, 39]. The later generates higher
contrast tomograms, which are beneficial for morphology studies.
The first has a proven record of maintaining high-resolution information and is the most popular algorithm especially if the final goal
is determining the protein structure.
The contrast of the final reconstructed tomogram can be further enhanced by the use of different imaging filters. These filters
can be applied to either the 2D projection images of the tilt series or
the final 3D volume. Cellular features, like cytoskeletal elements
and organelles, can be highlighted (segmented) in the tomogram
to generate a 3D model. This helps visualize complex 3D arrangements of proteins and membranes and allows for precise measurements [14]. Additionally, structures of proteins can be obtained by
applying a process called subtomogram averaging [40]. Typical
resolutions of the protein structures obtained by subtomogram
averaging are in the range of 10–40 A ˚ [18, 41, 42]. However, it is
possible to obtain protein structures at significantly higher resolution [29, 43].
Cryo-ET of Cellular Structures
5