automated procedure has been proposed to screen the 2D protein
crystals [31]. Once the crystal sample is screened, it will be frozen at
liquid-nitrogen temperatures to keep it hydrated and minimize the
beam damage for cryo-EM data collection.
Sample preparation for the electron crystallographic study is
more critical than that of single-particle cryo-EM. Crystal sample
for electron crystallographic study lies on the carbon film support,
and an intact and flat carbon film support can help preserving the
lattice order of the 2D arrays. If the carbon support is not flat or the
crystal does not lie flat, the diffraction spots that are perpendicular
to the tilt axis will become blurred when the stage is tilted
[32]. Therefore, specimen flatness is usually the key to the success
of 2D electron crystallography. Secondly, to prevent the 2D crystal
sample from dehydration, sugar embedding is a general method to
protect the sample in the high vacuum of the electron microscope
(EM) column [33]. The commonly used sugars are trehalose,
glucose, tannic acid, and sucrose, which mimic the solvent effects
and act as a cryo-protectant [34, 35]. Trehalose has been reported
to best preserve the high-resolution information of 2D crystals
[34, 36], while the choice of sugar to preserve on a specific type
of protein will need to determine empirically. The embedded specimen can be readily plunged into liquid nitrogen without fast cooling. This method was first introduced for electron crystallographic
studies on the purple membrane and catalase crystals with glucose
[7]. With low-dose imaging, the structure of bR was revealed at
high resolution [6, 37].
Because of the small conductivity of biological specimens,
when the electron beam illuminates the specimen, the charges are
temporarily separated and built up around the specimen, leading to
an effect similar to the image drift and resulting in poor image
quality. This is often referred to as beam-induced movement or
charging [35, 38]. This phenomenon is more noticeable when
imaging the tilted specimen under low-dose conditions [39]. To
minimize the beam-induced charging, the grid specimens could be
prepared with a thicker carbon film, with a 20–30 nm thickness
[35]. However, the thick specimen can lead to less elastic scattering
and lower the signal contents. Alternatively, spot-scan imaging was
designed to image the crystal sample with a small beam size (about
100 nm in diameter), which reduces the beam-induced charging.
However, it compromises with a lower spatial coherence of the
beam [40, 41]. On the other hand, the development of carbonsandwich method is an improvement on the sample preparation to
minimize beam-induced charging. It was developed to introduce
an additional piece of carbon film on the EM grid and help alleviate
the charging problem [39]. The idea of carbon-sandwich method is
the built-up charges that are symmetrically distributed on both
sides of the specimen will supposedly be canceled out when the
electron beam illuminates the tilted specimen [39]. Apart from this,
Electron Crystallography of Membrane Proteins
249
crystals [31]. Once the crystal sample is screened, it will be frozen at
liquid-nitrogen temperatures to keep it hydrated and minimize the
beam damage for cryo-EM data collection.
Sample preparation for the electron crystallographic study is
more critical than that of single-particle cryo-EM. Crystal sample
for electron crystallographic study lies on the carbon film support,
and an intact and flat carbon film support can help preserving the
lattice order of the 2D arrays. If the carbon support is not flat or the
crystal does not lie flat, the diffraction spots that are perpendicular
to the tilt axis will become blurred when the stage is tilted
[32]. Therefore, specimen flatness is usually the key to the success
of 2D electron crystallography. Secondly, to prevent the 2D crystal
sample from dehydration, sugar embedding is a general method to
protect the sample in the high vacuum of the electron microscope
(EM) column [33]. The commonly used sugars are trehalose,
glucose, tannic acid, and sucrose, which mimic the solvent effects
and act as a cryo-protectant [34, 35]. Trehalose has been reported
to best preserve the high-resolution information of 2D crystals
[34, 36], while the choice of sugar to preserve on a specific type
of protein will need to determine empirically. The embedded specimen can be readily plunged into liquid nitrogen without fast cooling. This method was first introduced for electron crystallographic
studies on the purple membrane and catalase crystals with glucose
[7]. With low-dose imaging, the structure of bR was revealed at
high resolution [6, 37].
Because of the small conductivity of biological specimens,
when the electron beam illuminates the specimen, the charges are
temporarily separated and built up around the specimen, leading to
an effect similar to the image drift and resulting in poor image
quality. This is often referred to as beam-induced movement or
charging [35, 38]. This phenomenon is more noticeable when
imaging the tilted specimen under low-dose conditions [39]. To
minimize the beam-induced charging, the grid specimens could be
prepared with a thicker carbon film, with a 20–30 nm thickness
[35]. However, the thick specimen can lead to less elastic scattering
and lower the signal contents. Alternatively, spot-scan imaging was
designed to image the crystal sample with a small beam size (about
100 nm in diameter), which reduces the beam-induced charging.
However, it compromises with a lower spatial coherence of the
beam [40, 41]. On the other hand, the development of carbonsandwich method is an improvement on the sample preparation to
minimize beam-induced charging. It was developed to introduce
an additional piece of carbon film on the EM grid and help alleviate
the charging problem [39]. The idea of carbon-sandwich method is
the built-up charges that are symmetrically distributed on both
sides of the specimen will supposedly be canceled out when the
electron beam illuminates the tilted specimen [39]. Apart from this,
Electron Crystallography of Membrane Proteins
249