Seeing tobacco mosaic virus through direct electron detectors.
Bottom Line: With the introduction of direct electron detectors (DED) to the field of electron cryo-microscopy, a wave of atomic-resolution structures has become available.As the new detectors still require comparative characterization, we have used tobacco mosaic virus (TMV) as a test specimen to study the quality of 3D image reconstructions from data recorded on the two direct electron detector cameras, K2 Summit and Falcon II.Using DED movie frames, we explored related image-processing aspects and compared the performance of micrograph-based and segment-based motion correction approaches.
Affiliation: EMBL - European Molecular Biology Laboratory, Structural and Computational Biology Unit, Meyerhofstr. 1, 69117 Heidelberg, Germany.Show MeSH
Related in: MedlinePlus
Mentions: The optimal image acquisition conditions of the K2 Summit and Falcon II cameras require significantly different dose rates (Bai et al., 2013; Li et al., 2013b). Therefore, we compared the effect of the different exposure times on shift and rotation changes in the course of irradiating a total of 20 e−/Å2. For the K2 Summit and the Falcon II data sets, we fitted Rayleigh distributions to determine mean translational changes at 4.2 and 2.9 Å (Fig. 3). These shift distributions were broadened by an estimated translational alignment error of 0.8 and 0.9 Å, respectively (Table 1). The elevated shifts of the K2 data set are in line with the K2 Summit exposure being 4 times longer than the Falcon II, thus indicating a larger drift contribution. The fitted histograms of angular changes revealed a mean at 1.1° and 0.8° (Fig. 3). Thus, the measured rotations were close to the estimated angular error of 0.9° and 0.8°. In addition, the rotation distributions are broadened by the fact that changes in azimuthal angles can be compensated by a small translation of the TMV helix. This is also one of the reasons why observed rotations between 0 and 20 e−/Å2 are significantly larger than described in previous studies whereas the measured shifts are close to recent data on rotavirus (Campbell et al., 2012). In summary, beam-induced movement at low dose rates leads to larger shifts when compared with high-dose rates while rotations are hardly affected by the longer exposure.
Affiliation: EMBL - European Molecular Biology Laboratory, Structural and Computational Biology Unit, Meyerhofstr. 1, 69117 Heidelberg, Germany.