Single-particle analysis was mainly executed in Relion 3.1 (38). All images were aligned and summed using MotionCor2 software (51). After contrast transfer function (CTF) parameter determination using CTFFIND4 (52), particle autopicking, manual particle checking, and reference-free 2D classification, particles with S trimer features were maintained for further processing.

For receptor-free S trimer sample, 226,082 particles were picked from nontilt micrographs, and 118,420 remained after 2D classification (fig. S3). These particles went through 3D auto-refine using available SARS-CoV-2 S trimer cryo-EM map (EMDB: 21452) low-pass filtered to 40-Å resolution as initial model (8). These particles were refined into a closed-state map of S trimer with imposed C3 symmetry. We then re-extracted the particles using the refinement coordinates to recenter it. After CTF refinement and polishing, these particles were refined with C3 symmetry again. It is noteworthy that the Euler angle distribution of the map suggested the dataset is lacking tilted top views (fig. S2C, left panel). When we refine the dataset without imposing threefold symmetry, the top view of the map appeared distorted, indicating a preferred orientation problem associated with the sample. To overcome the preferred orientation problem, we additionally collected tilt data and boxed out 198,737 particles from 40° tilt micrographs and 16,010 particles from 30° tilt micrographs. After 2D classification, 184,661 particles remained. We then used goCTF software to determine the defocus for each of the tilt particle, and these particles were re-extracted with corrected defocus (53). After combining the tilt with nontilt particles, we refined the dataset without imposing symmetry, then performed two rounds of 3D and 2D classifications to further clean up the dataset, and obtained a dataset of 151,505 particles, of which 62,368 particles were from the tilt data. We then carried out heterogeneous refinement in CryoSparc (54) and obtained a closed-state map from 142,345 particles and an open-state reconstruction with 9160 particles (fig. S3). After CTF refinement and Bayesian polishing, the closed-state map was refined to 2.7-Å resolution with C3 symmetry, while the open-state map was at 12.8-Å resolution and the resolution was hard to improve, indicating an intrinsic dynamic nature of the open state. The overall resolution was determined on the basis of the gold standard criterion using an Fourier shell correlation (FSC) of 0.143.

For the SARS-CoV-2 S-ACE2 cross-linked dataset, 298,127 particles were picked from original micrographs, and 138,632 particles remained after 2D classification (fig. S6). These particles were refined with an initial model built from our negative staining data. We then re-extracted the particles to recenter them. These particles went through a 3D-2D classification step, resulting in a further cleaned-up dataset of 77,440 particles. We refined these particles into a map of ACE2-bound S trimer complex. We then used this map as an initial model to refine the originally picked 298,127 particles for one round to re-extract and recenter the particles. After 2D classification, 207,742 particles remained. After two rounds of the 3D-2D cleaning step, 136,412 particles were left for further structure determination. After heterogeneous refinement in CryoSparc, class 1 resembled an ACE2-free open state of the S trimer, and classes 2 to 5 adopted the S-ACE2 engaged conformation. For class 1, after further 2D classification, we refined the 24,502 cleaned-up particles into an S-open map at 6.0-Å resolution using nonuniform refinement in CryoSparc. Among the other four classes with bound ACE2, we sorted out good particles for classes 2 to 4 by 2D classification and combined them with class 5 exhibiting good structural details, resulting in a dataset of 68,987 particles. After refinement, Bayesian polishing, and CTF refinement, we reconstructed a 3.8-Å-resolution SARS-CoV-2 S-ACE2 map. The S trimer portion without the up RBD was rather stable, which could be locally refined to 3.3 Å using local refinement in CryoSparc with nonuniform refinement option chosen. The ACE2 associated with the up RBD was subtracted and refined in Relion to obtain an 8.4-Å map with better connectivity. Multibody refinement was applied to analyze the mobility of the S-ACE2, S-closed, and S-open states.

For SARS-CoV-2 S-ACE2 without the cross-linking dataset, we followed similar classification and cleaning-up strategy and obtained 81,820 particles. Through heterogeneous refinement and 2D classification in CryoSparc, we reconstructed a 5.3-Å-resolution SARS-CoV-2 S-ACE2 map from 32,866 particles using nonuniform refinement and an unliganded open-state map of 11.2-Å resolution from 15,149 particles, with the population of 68.4 and 31.6%, respectively. Multibody refinement was also applied to analyze the mobility of the complex.

Note: The content above has been extracted from a research article, so it may not display correctly.

Please log in to submit your questions online.
Your question will be posted on the Bio-101 website. We will send your questions to the authors of this protocol and Bio-protocol community members who are experienced with this method. you will be informed using the email address associated with your Bio-protocol account.

We use cookies on this site to enhance your user experience. By using our website, you are agreeing to allow the storage of cookies on your computer.