Automated data acquisition is used widely for single-particle reconstruction of three-dimensional (3D) volumes of biological complexes preserved in vitreous ice and imaged in a transmission electron microscope. Automation has become integral to this method because of the very large number of particle images required in order to overcome the typically low signal-to-noise ratio of these images. For optimal efficiency, automated data acquisition software packages typically employ some beam-image shift targeting as this method is both fast and accurate (±0.1 µm). In contrast, using only stage movement, relocation to a targeted area under low-dose conditions can only be achieved in combination with multiple iterations or long relaxation times, both reducing efficiency. Nevertheless it is well known that applying beam-image shift induces beam-tilt and with it a potential structure phase error with a phase error π/4 the highest acceptable value. This theory has been used as an argument against beam-image shift for high resolution data collection. Nevertheless, in practice many small beam-image shift datasets have resulted in 3D reconstructions beyond the π/4 phase error limit. To address this apparent contradiction, we performed cryo-EM single-particle reconstructions on a T20S proteasome sample using applied beam-image shifts corresponding to beam tilts from 0 to 10 mrad. To evaluate the results we compared the FSC values, and examined the water density peaks in the 3D map. We conclude that the phase error does not limit the validity of the 3D reconstruction from single-particle averaging beyond the π/4 resolution limit.

Highlights d Cryo-EM structure of b 1 -adrenergic receptor and Gs at 2.6-A resolution d Network of interactions within Ga s are disrupted by b 1 -AR d Rotational opening of the a-helical domain of Ga s during its activation

Automated data acquisition is now used widely for the single-particle averaging approach to reconstruction of three-dimensional (3D) volumes of biological complexes preserved in vitreous ice and imaged in a transmission electron microscope (cryo-EM). Automation has become integral to this method because of the very large number of particle images required to be averaged in order to overcome the typically low signal-to-noise ratio of these images.For optimal efficiency, all automated data acquisition software packages employ some degree of beam-image shift because this method is fast and accurate (+/-0.1 µm). Relocation to a targeted area under low-dose conditions can only be achieved using stage movements in combination with multiple iterations or long relaxation times, both reducing efficiency. It is, however, well known that applying beam-image shift induces beam-tilt and hence structure phase error. A π/4 phase error is considered as the worst that can be accepted, and is used as an argument against the use of any beam-image shift for high resolution data collection.In this study, we performed cryo-EM single-particle reconstructions on a T20S proteasome sample using applied beam-image shifts corresponding to beam tilts from 0 to 10 mrad. To evaluate the results we compared the FSC values, and examined the water density peaks in the 3D map. We conclude that the π/4 phase error does not limit the validity of the 3D reconstruction from single-particle averaging beyond the π/4 resolution limit.

No abstract

11Cryo electron microscopy facilities running multiple instruments and serving users with varying skill 12 levels need a robust and reliable method for benchmarking both the hardware and software 13 components of their single particle analysis workflow. The workflow is complex, with many 14 bottlenecks existing at the specimen preparation, data collection and image analysis steps; the 15 samples and grid preparation can be of unpredictable quality, there are many different protocols for 16 microscope and camera settings, and there is a myriad of software programs for analysis that can 17 depend on dozens of settings chosen by the user. For this reason, we believe it is important to 18 benchmark the entire workflow, using a standard sample and standard operating procedures, on a 19 regular basis. This provides confidence that all aspects of the pipeline are capable of producing maps 20 to high resolution. Here we describe benchmarking procedures using a test sample, rabbit muscle 21 aldolase. 22