nanometer resolution [5, 6]. In favorable cases, atomic models of
these nanomachines can be built using known high-resolution
structures obtained by other methods like NMR, X-ray crystallography, and single particle analysis (SPA) which provides new mechanistic insights [7]. Moreover, comparing the structures of
different nanomachines or of the same nanomachine in different
species can shed light on how these biological machines have
evolved (and continue evolving) over time [8, 9]. In addition,
cryo-ET is very powerful in studying how large complexes assemble
and disassemble in the cell and is capable of capturing and revealing
the structures of intermediate and transient states during such
processes which is crucial in deciphering such pathways [10–13].
Cryo-ET is one of the essential modes of Cryo-EM (together
with SPA and Micro-Electron Diffraction (MicroED)) in which a
3D reconstruction of the sample is produced by back-projecting
images of the sample taken at different tilt angles. Usually, the
sample is tilted from À60
to +60
every 1
–3
. As biological
samples are sensitive to radiation, different tilt-schemes have been
developed like the bidirectional tilt-scheme and the dose symmetric
tilt-scheme [14] and the selection of a specific tilt-scheme is dependent on the sample and the question to be answered. In many cases,
single tomograms may not have sufficient signal to provide a
detailed knowledge of the molecular complex of interest. Therefore, to enhance the signal and the resolution of tomography data,
subtomogram averaging can be performed whereby the subvolumes in the tomograms that have the complex of interest are
cropped, aligned, and then averaged to produce an improved
signal-to-noise 3D average of the complex of interest
[15, 16]. The final resolution of subtomogram averaging depends
on many factors with some of them being related to how the data
was collected and others determined by inherent characteristics of
the complex of interest (flexibility, thickness heterogeneity, etc.).
Currently, by averaging many subvolumes the macromolecular
structures of many biological machines have been solved (see, e.g.,
Refs. 8, 10, 12, 17–26). For some samples, atomic models have
been generated (e.g., Refs. 27, 28).
Hybrid approaches that combine cryo-ET with other methods
have recently extended its applicability and usefulness. This
includes cryo-CLEM which allows the localization of specific protein complexes and targeted data collection on only the cells or
parts of cells that have the complex of interest. Moreover, as the
quality of tomography data depends on the sample thickness, cryoET has been limited for a long time to thin samples like bacterial
cells or the thin edges of eukaryotic cells. Thicker samples required
the cumbersome task of cryosectioning. However, the recent development of FIB-milling has extended the realm of tomography to
targets located deep in the thick regions of eukaryotic cells
[29]. Finally, recent hardware developments like phase plates
84
Mohammed Kaplan et al.
these nanomachines can be built using known high-resolution
structures obtained by other methods like NMR, X-ray crystallography, and single particle analysis (SPA) which provides new mechanistic insights [7]. Moreover, comparing the structures of
different nanomachines or of the same nanomachine in different
species can shed light on how these biological machines have
evolved (and continue evolving) over time [8, 9]. In addition,
cryo-ET is very powerful in studying how large complexes assemble
and disassemble in the cell and is capable of capturing and revealing
the structures of intermediate and transient states during such
processes which is crucial in deciphering such pathways [10–13].
Cryo-ET is one of the essential modes of Cryo-EM (together
with SPA and Micro-Electron Diffraction (MicroED)) in which a
3D reconstruction of the sample is produced by back-projecting
images of the sample taken at different tilt angles. Usually, the
sample is tilted from À60
to +60
every 1
–3
. As biological
samples are sensitive to radiation, different tilt-schemes have been
developed like the bidirectional tilt-scheme and the dose symmetric
tilt-scheme [14] and the selection of a specific tilt-scheme is dependent on the sample and the question to be answered. In many cases,
single tomograms may not have sufficient signal to provide a
detailed knowledge of the molecular complex of interest. Therefore, to enhance the signal and the resolution of tomography data,
subtomogram averaging can be performed whereby the subvolumes in the tomograms that have the complex of interest are
cropped, aligned, and then averaged to produce an improved
signal-to-noise 3D average of the complex of interest
[15, 16]. The final resolution of subtomogram averaging depends
on many factors with some of them being related to how the data
was collected and others determined by inherent characteristics of
the complex of interest (flexibility, thickness heterogeneity, etc.).
Currently, by averaging many subvolumes the macromolecular
structures of many biological machines have been solved (see, e.g.,
Refs. 8, 10, 12, 17–26). For some samples, atomic models have
been generated (e.g., Refs. 27, 28).
Hybrid approaches that combine cryo-ET with other methods
have recently extended its applicability and usefulness. This
includes cryo-CLEM which allows the localization of specific protein complexes and targeted data collection on only the cells or
parts of cells that have the complex of interest. Moreover, as the
quality of tomography data depends on the sample thickness, cryoET has been limited for a long time to thin samples like bacterial
cells or the thin edges of eukaryotic cells. Thicker samples required
the cumbersome task of cryosectioning. However, the recent development of FIB-milling has extended the realm of tomography to
targets located deep in the thick regions of eukaryotic cells
[29]. Finally, recent hardware developments like phase plates
84
Mohammed Kaplan et al.