indexed. For macromolecules, a dataset spanning 10–20
with ~50
clearly visible spots on each frame can readily be indexed with
standard data reduction programs.
3.1 Cryo-EM
Ultimately, the amount of data that can be measured from a single
crystal is limited by radiation damage. Primary damage from
absorption of electrons is inevitable, and its rate has been found
to be independent of temperature. Secondary damage, on the other
hand, relies on diffusion of reactive species throughout the crystal,
which is slowed down at cryogenic temperatures. Low-temperature
data collection may also serve to mechanically restrain molecular
fragments, preventing the disruption of global order [30]. Cryocooling has repeatedly been found to prolong the lifetime of the
crystal in the beam and allows data acquisition to proceed under
extensive exposure [21–24]. In X-ray crystallography, radiation
damage has been observered to be minimized at temperatures
closer to 50 K [25], >20
C below the temperature in a typical,
liquid nitrogen-cooled instrument. Because the underlying chemical and physical damage mechanisms are assumed to be identical in
cryo-EM, it may be possible to realize a three-fold increase in dose
tolerance in cryo-EM as well. Nevertheless, even a helium-cooled
sample will eventually be destroyed by radiation damage.
3.2 Resolution
Effects
Damage to the fine features of the structure in the crystal can be
inferred from the decrease of the average intensity of the highresolution reflections, even before the data has been interpreted in
terms of an atomic model [26]. For instance, for proteinase K, it
was found that the exponential fading of the average intensity of
reflections in the 2.00–2.04 A ˚ interval due to global damage was
2.5-fold faster than for reflections in the 5.15–21.0 A ˚ interval
[9]. In practice, tuning the exposure will trade resolution against
completeness: increasing the exposure may raise the faint highresolution reflections over the noise of the background but will
also decrease the number of frames that can be recorded.
Specific damage, in the form of decarboxylation of acidic residues, breakage of disulfide bonds, and various conformational
changes of amino acid side chains, occurs as soon as the first
electrons impinge on the sample. The effects of site-specific damage
are difficult to observe in both single-particle cryo-EM and
MicroED, as both methods rely on extensive averaging to boost
the signal over the noise. Only when a site is affected in a substantial
fraction of the averaged molecules and the data are good enough to
resolve the mean effect of the changes, can the damage be seen in
the final density, but then, these changes are nefarious. Limited
damage may appear as disproportionately high temperature factors
indicating disorder or increased mobility of the affected sites. In
severe cases specific damage may lead to misinterpretation of structural features and biologically important functional results.
Low-Dose MicroED
313
with ~50
clearly visible spots on each frame can readily be indexed with
standard data reduction programs.
3.1 Cryo-EM
Ultimately, the amount of data that can be measured from a single
crystal is limited by radiation damage. Primary damage from
absorption of electrons is inevitable, and its rate has been found
to be independent of temperature. Secondary damage, on the other
hand, relies on diffusion of reactive species throughout the crystal,
which is slowed down at cryogenic temperatures. Low-temperature
data collection may also serve to mechanically restrain molecular
fragments, preventing the disruption of global order [30]. Cryocooling has repeatedly been found to prolong the lifetime of the
crystal in the beam and allows data acquisition to proceed under
extensive exposure [21–24]. In X-ray crystallography, radiation
damage has been observered to be minimized at temperatures
closer to 50 K [25], >20
C below the temperature in a typical,
liquid nitrogen-cooled instrument. Because the underlying chemical and physical damage mechanisms are assumed to be identical in
cryo-EM, it may be possible to realize a three-fold increase in dose
tolerance in cryo-EM as well. Nevertheless, even a helium-cooled
sample will eventually be destroyed by radiation damage.
3.2 Resolution
Effects
Damage to the fine features of the structure in the crystal can be
inferred from the decrease of the average intensity of the highresolution reflections, even before the data has been interpreted in
terms of an atomic model [26]. For instance, for proteinase K, it
was found that the exponential fading of the average intensity of
reflections in the 2.00–2.04 A ˚ interval due to global damage was
2.5-fold faster than for reflections in the 5.15–21.0 A ˚ interval
[9]. In practice, tuning the exposure will trade resolution against
completeness: increasing the exposure may raise the faint highresolution reflections over the noise of the background but will
also decrease the number of frames that can be recorded.
Specific damage, in the form of decarboxylation of acidic residues, breakage of disulfide bonds, and various conformational
changes of amino acid side chains, occurs as soon as the first
electrons impinge on the sample. The effects of site-specific damage
are difficult to observe in both single-particle cryo-EM and
MicroED, as both methods rely on extensive averaging to boost
the signal over the noise. Only when a site is affected in a substantial
fraction of the averaged molecules and the data are good enough to
resolve the mean effect of the changes, can the damage be seen in
the final density, but then, these changes are nefarious. Limited
damage may appear as disproportionately high temperature factors
indicating disorder or increased mobility of the affected sites. In
severe cases specific damage may lead to misinterpretation of structural features and biologically important functional results.
Low-Dose MicroED
313