Home Health News Nanometer-resolution in situ structure of the SARS-CoV-2 postfusion spike protein

Nanometer-resolution in situ structure of the SARS-CoV-2 postfusion spike protein

65 min read


Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is a extreme menace to public health and the international economic system. Its spike protein is accountable for the membrane fusion and is thus a significant goal for vaccine and drug growth. Our research presents the in situ structure of the spike protein in the postfusion state with larger decision, giving additional insights into the design of a viral entry inhibitor. Our remark of the oligomerization states of spikes on the viral membrane implies a doable mechanism of membrane fusion for viral an infection.


The spike protein of extreme acute respiratory syndrome coronavirus 2 (SARS-CoV-2) mediates membrane fusion to permit entry of the viral genome into host cells. To perceive its detailed entry mechanism and develop a selected entry inhibitor, in situ structural info on the SARS-CoV-2 spike protein in completely different states is pressing. Here, by utilizing cryo-electron tomography, we noticed each prefusion and postfusion spikes in β-propiolactone–inactivated SARS-CoV-2 virions and solved the in situ structure of the postfusion spike at nanometer decision. Compared to earlier reviews, the six-helix bundle fusion core, the glycosylation websites, and the location of the transmembrane area had been clearly resolved. We noticed oligomerization patterns of the spikes on the viral membrane, possible suggesting a mechanism of fusion pore formation.

Over the previous 20 years, a number of zoonotic coronavirus (CoV) illnesses have emerged and posed a devastating menace to international public health and the economic system, corresponding to extreme acute respiratory syndrome (SARS) (1), Middle East respiratory syndrome (MERS) (2), and COVID-19 (3). As of this writing, COVID-19 has greater than 229 million confirmed circumstances and has precipitated 4.7 million deaths worldwide, with quickly rising numbers. This pneumonia epidemic was brought on by a novel coronavirus named SARS coronavirus 2 (SARS-CoV-2), a β-coronavirus, with a genomic sequence that’s carefully associated to SARS-CoV. SARS-CoV-2 is an enveloped, positive-sense single-stranded RNA virus with an ∼30-kb genome (4). Given the present pandemic state of affairs, understanding the structure of SARS-CoV-2 in addition to its an infection course of is essential for vaccine growth and drug discovery.

The SARS-CoV-2 genome encodes three viral floor proteins: the spike (S) glycoprotein, envelope (E) protein, and membrane (M) protein. During the an infection course of, the trimeric S glycoprotein is cleaved by host proteases (4, 5) to supply two practical subunits: The N-terminal S1 subunit is accountable for receptor recognition, and the C-terminal S2 subunit is accountable for membrane fusion (6). Mediated by receptor binding and proteolytic activation, the S1 subunit falls off, and the S2 subunit undergoes in depth and irreversible conformational modifications to insert its hydrophobic fusion peptide (FP) into the goal cell membrane. Subsequently, two heptad repeat areas of the S2 subunit, heptad repeat 1 (HR1) and heptad repeat 2 (HR2), type a steady six-helix bundle (6-HB) fusion core to deliver collectively the viral and mobile membranes, resulting in colocalization of the FP and the transmembrane (TM) area at the similar finish to type the fusion pore (7). Thus, the S protein is one of the main targets for growing vaccines and antiviral medication.

After the outbreak of COVID-19, the in vitro constructions of SARS-CoV-2 S in the prefusion state had been promptly solved utilizing single-particle cryo-electron microscopy (cryo-EM) (8, 9) and X-ray crystallography (7, 10, 11). Soon afterward, the in situ constructions of S in the prefusion state had been revealed by cryo-electron tomography (cryo-ET) and cryo-subtomogram averaging (cryo-STA) (1214), uncovering the distribution of completely different conformational states in addition to the native glycosylation websites. However, how the S protein is activated to induce membrane fusion with its host is much less understood. The structure of S in the postfusion state would supply an necessary clue to analyze the fusion mechanism. The high-resolution structure of recombinant S in the postfusion state has been reported by Cai et al. (15), however this in vitro research failed to find out how the postfusion S proteins set up on the membrane. Previous in situ research (12, 13, 16) explored this query however yielded restricted info, attributable to the poor high quality of the density map. In addition, we beforehand confirmed that the recombinant 6-HB fusion core of S in the postfusion state can be an efficient goal for the design of viral fusion inhibitors (7), which must be additional validated by a higher-resolution structure and glycosylation info of in situ S in the postfusion state.

In the present work, we utilized cryo-ET and cryo-STA to review the structure of SARS-CoV-2 viruses that had been inactivated by β-propiolactone (BPL). We solved the in situ constructions of S in each the prefusion and postfusion states with resolutions of 12.9 and 10.9 Å, respectively. In addition to visualizing the TM area and glycosylation websites, we discovered that our earlier crystal structure of the recombinant 6-HB fusion core matches effectively to the density map. In addition, we noticed oligomerization of postfusion Ss on the viral membrane, suggesting a mechanism of S-induced membrane fusion. Our research will facilitate a greater understanding of the SARS-CoV-2 fusion mechanism and be helpful for viral entry inhibitor growth.


Cryo-ET Analysis of the Inactivated SARS-CoV-2 Virus.

We propagated SARS-CoV-2 virions into Vero cells and purified the viral particles in a biosafety stage 3 (BSL-3) laboratory. The purified virus was inactivated with BPL and imaged by cryo-ET in a BSL-2 laboratory. In the reconstructed tomograms, we noticed a typical coronavirus morphology of SARS-CoV-2 virions with diameters starting from 80 nm to 120 nm (Fig. 1A). Inside every virion, the ribonucleoprotein complexes had been tightly packed, with a diameter of ∼15 nm. From the deconvoluted tomograms utilizing Warp (a pc software program for cryo-EM knowledge processing) (17), we might clearly visualize most Ss that had been prepared for subsequent particle choosing. Both the prefusion and postfusion states of S had been noticed (Fig. 1 A and B), as reported beforehand (16), which was in line with the proven fact that cleavage of S had occurred throughout the pattern preparation (SI Appendix, Fig. S1).

Previous research have argued that the percentages of S in the prefusion and postfusion states are associated to viral inactivation strategies. The majority of prefusion S was from formaldehyde-fixed samples (12), whereas an amazing portion of postfusion S seemed to be from the BPL-inactivated pattern, with a ratio of as much as 66 to 81.3% (16, 18). In our pattern, we additional investigated the ratio of prefusion to postfusion Ss. We picked all doable S particles by combining the template matching strategy with the handbook methodology. We averaged the maps of each prefusion and postfusion Ss to generate a reference for subsequent three-dimensional (3D) classification, which confirmed that 42% of the particles had been categorized into the prefusion state, and 48% had been categorized into the postfusion state (SI Appendix, Fig. S2). Thus, the populations of prefusion and postfusion Ss had been comparable in our BPL-inactivated pattern, which was completely different from a earlier report (16).

A latest research confirmed that prefusion S displays a versatile orientation with respect to the viral membrane, tilting from the vertical axis to the viral membrane at a variety of 50° (12). This structural characteristic might assist prefusion S search and bind to the ACE2 receptor of the goal cell. In distinction, by visible inspection, we discovered that an amazing portion of the postfusion Ss appeared perpendicular to the viral membrane, which prompt that the conformation of the postfusion S has a steady membrane proximal exterior area or a steady TM area. In addition to the dispersed postfusion Ss on the viral membrane (Fig. 1B), we additionally noticed that some postfusion Ss oligomerized in parallel (Fig. 1C) or in branches (Fig. 1D). We statistically decided the postfusion S on the viral membrane and located that the distribution of the postfusion S was sparse, with three Ss per virion on common (Fig. 1E). However, by inspecting all pairs of postfusion Ss in the similar virions and calculating the pair distances, we discovered that many pairs had distances of ∼20 nm or much less (Fig. 1F), implying potential clustering habits of postfusion Ss. In order to validate this remark, we generated a simulated dataset, in which the numbers of viruses and Ss on every virus had been all stored the similar as the experimental knowledge, however the postfusion Ss had been randomly distributed on sphere-shaped virus. Using the similar calculation methodology as for the nearest pair distance, the randomly positioned Ss had no clustering impact, displaying a standard distribution sample with a middle of 60 nm (Fig. 1F). This prompt that the clustering peak of postfusion Ss in our experimental dataset was statistically important.

Subtomogram Analysis of SARS-CoV-2 Postfusion S.

We then carried out subtomogram evaluation from a complete of 15,525 chosen S particles (SI Appendix, Figs. S2 and S3). After 3D classification and autorefinement, we obtained an in situ structural map of prefusion S with C3 symmetry at a decision of 12.9 Å based on the gold commonplace Fourier shell correlation (FSC) coefficient at 0.143 (SI Appendix, Fig. S2). Our in situ structure of prefusion S was just like these obtained in earlier reviews (12, 13). In the present research, we centered on postfusion S displaying the nail form on the viral membrane.

We utilized completely different approaches to align the particles of postfusion Ss by making an attempt native or international searches of orientations with C1 or C3 symmetries. We discovered that solely native searches with C3 symmetry with restriction of the Euler angles that had prior values throughout particle choosing (SI Appendix, Fig. S2) might yield a high-resolution (10.9 Å) map based on the gold commonplace FSC coefficient at 0.143 (Fig. 2 A and B). From the averaged map of postfusion S, we clearly distinguished the head area (connector), stalk area (6-HB), and TM area. The three S protomers is also distinguished from the map with the next threshold (Fig. 2A). The directional FSC evaluation confirmed that there was no most popular orientation in our dataset (Fig. 2C).

Fig. 2.
Fig. 2.

Subtomogram averaging of the SARS-CoV-2 postfusion S. (A) Subtomogram averaging process, together with the preliminary common of all particles at a binning stage of eight (Left), the refined map at a binning stage of two (Middle), and the postprocessed map with a neighborhood masks at the extracellular area at a binning stage of two (Right). (B) The gold commonplace FSC curves for the closing averaged map of postfusion S. (C) The directional FSC of the closing averaged map of postfusion S. (D) Statistics of the coordinate and Euler angle modifications earlier than and after alignment throughout picture processing. The coordinate offsets on the x, y, z axes are outlined as dx, dy, and dz, whereas the angular offsets at three Euler angles are outlined as drot, dtilt, and dpsi. (E) Plot again of the averaged map onto the authentic tomogram, displaying relative orientations of postfusion Ss to the virus membrane. (F) Distribution of the tilting angles of postfusion Ss relative to viral envelope.

To additional validate the outcomes of our alignment and refinement, we inspected the shift and orientation modifications of particles earlier than and after alignment. We discovered that almost all particles shifted lower than 4 nm in comparison with their manually picked coordinates (Fig. 2D). For the three Euler angles, the first one (rot. angle) exhibited the smallest change with a imply worth of 7.38°, modifications in the tilt angle had a imply worth of 21.2°, and the psi angle confirmed no correlation with the beginning worth (Fig. 2D). Manual choosing might determine solely the first two Euler angles (rot. and tilt), and our remark of shift and orientation modifications agreed effectively with this reality.

After plotting the refined map again onto the uncooked tomogram with refined coordinates and orientations, we discovered that every one the postfusion Ss had been in the rational place on the viral envelope (Fig. 2E). Moreover, the relative angles of postfusion Ss to virion had been calculated, indicating that almost all Ss had tilting angles bigger than ∼45° relative to the envelope, and the median worth is ∼58° (Fig. 2F). These statistics agreed effectively with visible inspections and the plot again outcome. Comparing with earlier reviews for tilting angles of prefusion Ss (12, 13), plainly Ss in postfusion state are usually extra perpendicular to the viral envelope. In addition, to validate whether or not the orientations of postfusion Ss had been affected by air–water interface (AWI), we calculated the distributions of postfusion Ss in the z axis of authentic tomograms, and located that greater than 90% postfusion Ss had been far-off from AWI (SI Appendix, Fig. S3C). This outcome additional proved the reliability of our findings.

In Situ Structure of SARS-CoV-2 Postfusion S.

The in situ structure of postfusion S of SARS-CoV-2 has a size of ∼210 Å and a width of ∼87 Å at the ectodomain (Fig. 3A and Movie S1). Its general structure resembles the beforehand reported recombinant postfusion S (Protein Data Bank [PDB] entry: 6XRA) (15) with an unusually lengthy (>180 Å) and inflexible 6-HB fashioned by the HR1 and HR2 domains (Fig. 3A and SI Appendix, Fig. S4). On this foundation, we constructed an in situ structural mannequin with an extension at the TM area (Fig. 3 A and B). Compared to earlier recombinant constructions with vague TM areas (15), the TM area could be simply discerned in our in situ map beneath the HR2 trimer (Fig. 3A and SI Appendix, Fig. S4). Since the native density map of the TM area was not ok to differentiate particular person helices, we fitted the predicted mannequin of the TM area (Trp1212 to Leu1234) in the map (Fig. 3B). In addition, the three helices of HR2 might be clearly distinguished from the helices of HR1 in the 6-HB area (Fig. 3C and SI Appendix, Fig. S5). Compared to the reported in situ constructions of postfusion S of SARS-CoV-2 (Electron Microscopy Data Bank (EMD) accession codes EMD-11627 and EMD-30428) (13, 16), our map displays significantly better high quality at the areas of the connector, 6-HB, and TM domains to reveal extra structural particulars (SI Appendix, Fig. S4). Based on our structure, an up to date mannequin of how the S protein modifications its conformation from prefusion to postfusion throughout viral an infection was proposed (Movie S2).

Fig. 3.
Fig. 3.

In situ structure of the SARS-CoV-2 postfusion S. (A) Geometry of postfusion S on the membrane. This map was generated by merging the TM area of the full averaged map onto the higher-resolution map from centered alignment on the extracellular domains. (B) Domain association of full-length S protein and the modeled components of the revealed structure (PDB entry: 6XRA) and our mannequin. The potential cleavage websites and glycosylation websites are indicated. (C) Superimposition of our mannequin with the merged map talked about in A, displaying the glycosylation websites and distinguished 6-HB domains. The noticed glycosylation websites are proven in burlywood. Three S protomers are proven in medium slate blue, mild coral, and medium turquoise.

Glycosylation has in depth roles in viral pathogenesis, corresponding to immune evasion, by shielding particular epitopes from neutralizing antibodies. Based on a earlier research (8), there are 22 N-linked glycosylation websites in every chain of the SARS-CoV-2 S protein, of which 16 websites are located earlier than the FP area and 6 are located after. According to the coated sequence area of the S2 subunit in our mannequin, there have been eight N-glycosylation websites (N709, N717, N1074, N1098, N1134, N1158, N1173, and N1194) (15). Among them, we noticed seven N-glycosylation websites (N709, N717, N1098, N1134, N1158, N1173, and N1194) with clear densities in our map (Fig. 3C). These glycosylation websites are just like these beforehand reported from in situ and in vitro research of postfusion Ss (13, 15, 16), offering further proof that these posttranslational modifications are extremely conserved and widespread in varied SARS-CoV-2 strains and may play necessary roles in viral infections and immune responses. We didn’t observe an apparent density for N-glycosylation at N1074, suggesting both no glycosylation or a low stage of glycosylation for this web site in our pattern. Previous research have reported that glycosylation at N717 and N1074 primarily incorporates oligomannose, whereas N1098, N1134, N1158, N1173, and N1194 glycosylation primarily incorporates complex-type glycans (15). In our subtomogram-averaged map, we discovered that the densities of complex-type glycans had been extra apparent than these of oligomannose (Fig. 3C).

In Situ Oligomerization of SARS-CoV-2 Postfusion S.

One necessary step throughout viral an infection of enveloped viruses is the formation of fusion pores between viral and cell membranes, which makes entry of the viral genome into the host cell doable. SARS-CoV-2 is triggered by the conformational change of the S2 subunit from the prefusion to postfusion state by way of a “jackknife” transition, which successfully brings viral and mobile membranes into shut proximity prepared for fusion (Movie S2). During particle choosing, we observed that two meeting patterns of postfusion Ss existed on the viral membrane: One sort confirmed the Ss standing parallel to one another (Fig. 1C), and the different sort, jointed at the root, was heading out in completely different instructions (Fig. 1D). We plotted again all refined particles into the reconstructed volumes, fitted the closing mannequin into the maps, after which managed to obviously observe the organizing sample of these particular Ss with the side-by-side and branching patterns (Fig. 4 A and B). It is value noting that this oligomerization sample has by no means been discovered for the purified S protein or in situ prefusion state of SARS-CoV-2 S.

Fig. 4.
Fig. 4.

Oligomerization association of the in situ SARS-CoV-2 postfusion S. (A and B) Side-by-side (A) and branching (B) preparations of postfusion Ss. The orientation of every spike was decided based on the closing refined parameters. Coloring scheme is the similar as that in Fig. 3C. (C) Possible mannequin of HR2 area trade in the side-by-side oligomerization state. The HR2 domains of the two Ss are coloured magenta and blue. The different components of the two Ss are coloured pink and cyan. (D) Models of in situ postfusion S in the common state, side-by-side state with HR2 area trade, and branching state with FP interplay. (E) A scheme of the SARS-CoV-2 S transition from the prefusion to the postfusion state throughout viral an infection and fusion pore formation with the side-by-side oligomerization of Ss concerned.

For the side-by-side sample, the postfusion Ss are parallel and shut to one another (<10 nm for the 6-HB area), and their connector domains are nearer, nearly shut sufficient to type direct contact (Fig. 4 A and B). Based on this remark, we proposed a doable mechanism by which Ss could work together with one another by way of versatile HR2 domains in a site trade method (Fig. 4C). During the folding course of of 6-HB from the prefusion state to the postfusion state, the HR2 area seeks a neighboring HR1 area for binding. If there are different Ss round, it is likely to be possible that the HR2 area might bind to the adjoining HR1 trimer to type a site trade conformation (Fig. 4 C and D). In this manner, every postfusion S might keep an intact trimer after conformational change however will stand shut and parallel to its companion. For the branching sample, Ss stand in branches with jointed roots in the TM area, and this state could also be attributable to the oligomerization of their FPs and even TM domains (Fig. 4D). These two varieties of group patterns will considerably improve the native abundance of postfusion S on the viral membrane, which can play an necessary position in the formation of fusion pores (Fig. 4E).


Enveloped viruses use specialised protein equipment to deliver viral and mobile membranes into shut proximity for membrane fusion. Despite in depth research on protein equipment and its fusion exercise, the molecular mechanism by which viral and mobile membranes promote fusion is poorly understood, particularly fusion pore formation. Some fashions of membrane fusion resulting in viral an infection have been proposed, and protein equipment oligomerization may facilitate the formation of fusion pores (1921). On this foundation, we proposed a doable mannequin of SARS-CoV-2 membrane fusion and an infection (Fig. 4E). After the interplay between the S1 subunit and the human ACE2 receptor, the S2 subunit is uncovered and undergoes a conformational change to insert FPs into the goal cell membrane. At this level, if there are a number of Ss close by, they may oligomerize to type side-by-side or branching constructions. This type of oligomeric state will result in an elevated focus of Ss in the native area, presumably enhancing the localized destabilization of lipid bilayers and resulting in a extra environment friendly formation of viral fusion pores.

Generally, the intramolecular interactions are extra favorable than intermolecular interactions. If our proposed mannequin for oligomerizations of postfusion Ss is right, the prefusion Ss must be shut sufficient to one another in advance, earlier than the rearrangement of S2 area, for area swap to happen. By additionally calculating the nearest pair distance of prefusion Ss in our dataset, we discovered a peak distance of round ∼30 nm (SI Appendix, Fig. S3B). This was barely longer than the peak distance of postfusion Ss (Fig. 1F), however it might be speculated that these prefusion Ss shut to one another made it doable for the rearrangement of the S2 area in the postfusion state. In addition, the comparable oligomerization patterns of postfusion Ss have been exhibited however not mentioned in a earlier report (16). It supplied additional proof that the oligomerization of postfusion Ss noticed in our dataset was not unintentional.

During SARS-CoV-2 an infection, the HR1 and HR2 domains work together with one another to deliver the viral and mobile membranes shut sufficient to type fusion pores, which makes 6-HB an necessary goal for the growth of viral entry inhibitors. We observed that our beforehand reported crystal structure of SARS-CoV-2 S 6-HB (5, 7) matches effectively in our in situ map, suggesting that 6-HB is very steady even on the membrane of pure viruses (SI Appendix, Fig. S5). We additionally discovered that, in the absence of the HR2 motif and its glycosylation, the HR1 trimer might be uncovered fully outward and is accessible for entry inhibitors corresponding to peptides and compounds. This is further structural proof that helps our earlier design of a viral inhibitor, the extremely potent pancoronavirus fusion inhibitor EK1C4, that may inhibit an infection by SARS-CoV-2 and different identified human coronaviruses (7). Most lately, the crystal structure and the preclinical analysis of these HR1-targeting fusion inhibitors had been reported, supporting additional scientific growth of these pan-CoV fusion inhibitors towards SARS-CoV-2 (22). An identical technique has been used to develop inhibitors for different viruses, for instance, HIV-1 (23), LASV (Lassa virus) (24), and MERS-CoV (25).

It also needs to be famous that the glycosylation web site N1158 is located proper at the binding web site of S2P6, which is a broad neutralizing antibody that blocks membrane fusion of β-coronaviruses (26). N1158 glycosylation may play a task in shielding the binding web site of S2P6 and thus permit the formation of 6-HB. Therefore, the future design of 6-HB–concentrating on antibodies to dam viral an infection ought to pay extra attention to the steric hindrance from glycosylation websites in postfusion S of SARS-CoV-2.

As a extensively used inactivation reagent to fabricate viral vaccines, BPL can’t solely chemically modify nucleic acids but additionally, to some extent, trigger results on viral proteins (27, 28). It has additionally been reported that BPL therapy might inhibit the membrane fusion course of of the influenza virus by altering the constructions and features of viral proteins (29). However, in the present research, our BPL inactivation and pattern preparation process maintained an inexpensive inhabitants of Ss in the prefusion state. Therefore, inactivation reagents shouldn’t be the solely purpose to induce the conformational transition from the prefusion to postfusion state. Other elements ought to exist in the pattern preparation process, which might encourage us to additional optimize the inactivation technique for higher vaccine growth.

In abstract, our present work proposes a higher-resolution in situ structure of postfusion S of SARS-CoV-2 and discovers its oligomerization states on the membrane that presumably have necessary features in the viral an infection course of, offering additional structural info for the growth of the subsequent technology of vaccines and viral entry inhibitors.

Materials and Methods

Facility and Ethics Statements.

All experiments with stay SARS-CoV-2 viruses had been carried out in the BSL-3 (P3+) amenities in the Academy of Military Medical Sciences, China. All experiments had been carried out in accordance with the Regulations in the Guide for the Ministry of Science and Technology of the People’s Republic of China.

Virus Purification and Cryo-EM Tomography Sample Preparation.

Vero cells (American Type Culture Collection, CCL-81) had been maintained in Dulbecco’s modified essential medium supplemented with 10% fetal bovine serum (Biowest) at 37 °C with 5% CO2. The pressure BetaCoV/Wuhan/AMMS01/2020 was initially remoted from a COVID-19 affected person coming back from Wuhan, China. The virus was amplified and titrated by a regular plaque-forming assay on Vero cells, as beforehand reported (30, 31). SARS-CoV-2 was cultured in large-scale Vero cell factories at a multiplicity of an infection of 0.5 at 37 °C with 5% CO2. To inactivate virus manufacturing, BPL was completely combined with the supernatant of the contaminated cells at a ratio of 1:4,000 vol/vol for 48 h at 2 °C to 8 °C. Following clarification of the cell particles and ultrafiltration, the inactivated viruses had been purified by ion trade chromatography and dimension exclusion chromatography, as beforehand reported (32, 33). Purified viruses had been combined at a ratio of 5:1 (virus:gold) with 10 nm of protein A-coated gold fiducials (Electron Microscopy Sciences). Then, 3 μL of the combination was utilized onto a discharged 300 mesh copper grid with a C-flat R 2/1 holey carbon assist movie. Grids had been blotted for 3 s in 100% relative humidity for plunge freezing in liquid ethane utilizing Vitrobot (Thermo Fisher Scientific).

Cryo-ET Data Acquisition.

Cryogrids had been loaded into an FEI Titan Krios G2 transmission electron microscope (Thermo Fisher Scientific) operated at 300 kV, and pictures had been recorded on a Gatan K2 Summit DDD digicam (Gatan Company) in superresolution mode geared up with a Gatan Quantum power filter with a slit width of 20 eV in zero-loss mode. Nominal magnification was set to 105,000×, ensuing in a calibrated bodily pixel dimension of 1.36 Å at the specimen stage. Tilt collection between −60° and +60° had been acquired utilizing a dose-symmetric scheme with a 3° angular increment utilizing SerialEM software program with an in-house script (34). A complete dose of 123 e2 per tilt collection was distributed evenly amongst 41 tilts. The defocus vary was set between −1.5 μm and −3 μm, and 10 frames had been saved for every tilt angle.

Image Processing and Subtomogram Averaging.

The output superresolution films had been first subjected to movement correction with a binning stage of two utilizing Warp (35), ensuing in a pixel dimension of 1.36 Å, and masking of fiducial markers was additionally carried out utilizing boxnet instruments inside Warp. All 373 tilt collection stacks had been generated utilizing computerized procedures in Warp. To achieve the greatest tilt collection alignment high quality, solely tilts inside −45° to +45° had been stored for additional alignment. Alignments of tilt collection had been carried out utilizing computerized tilt collection alignment features in the Dynamo (a software program setting for subtomogram averaging of cryo-EM knowledge) and IMOD (a pc software program package deal for analyzing and viewing three-dimensional organic picture knowledge) packages (3638). The tilt collection with fully failed alignments had been discarded by visible inspection utilizing IMOD (38). Then, the alignment information of 352 efficiently aligned tilt collection had been transferred again to Warp to carry out per-tilt distinction switch perform (CTF) estimation. Tomograms had been reconstructed in Warp at a binning stage of eight and deconvolved for higher visualization.

For particle choosing of prefusion Ss, we used a beforehand reported cryo-EM map of prefusion S (EMD-21452) (8) low-pass filtered to 40 Å as a reference for template matching by the Dynamo package deal (37) in eight binned deconvolved tomograms. Postfusion Ss had been manually picked utilizing Dynamo packages (37), whose preliminary Euler angles (two out of three) had been decided based mostly on the vector between two manually set factors, one in the center of S and one other on the membrane. Then, the coordinates and orientations of 7,656 prefusion and 7,869 postfusion particles had been employed for the extraction of subvolumes in Warp with 48 × 48 × 48 voxels at a voxel spacing of 10.88 Å. The corresponding 3D CTF fashions had been additionally generated contemplating amassed radiation harm. To decide the actual ratio of prefusion and postfusion states on a nonbiased foundation, a median of all extracted prefusion and postfusion Ss was generated for use as a single reference for 3D classification in RELION (a pc program for cryo-EM knowledge processing) variations 3.0 and 3.1 (39, 40). The classification converged into 4 completely different teams: one exhibited a robust density of double-layer membranes (10%), two exhibited the typical morphology of postfusion S (48%), and one exhibited the typical morphology of prefusion S (42%). Using the above categorized prefusion coordinates and orientations of the Ss, we carried out particle reextraction with a binning stage of two utilizing Warp. A subsequent autorefinement job towards 6,456 particles utilizing RELION yielded a 12.9-Å map of prefusion S (SI Appendix, Fig. S2).

To obtain a high-resolution map of postfusion S with out mannequin bias, no prior constructions or maps from different research had been used all through the knowledge processing steps. First, the subtomograms had been straight averaged with out alignment and symmetry utilized to generate a data-driven low-resolution template with solely manually set Euler angles utilized utilizing relion_reconstruct. This course of yields a great reference for subsequent alignment (SI Appendix, Fig. S2). To validate the utilized symmetry and discover a good reference for additional alignment, 4 completely different periods with or with out restrictions on the search vary for rot. and tilt angles and with or with out the threefold symmetry utilized had been carried out. These periods converged into a median with noticed threefold symmetry and with the double layer of viral membrane together with TM area seen. However, just one session with a restricted seek for rot. and tilt angles and with threefold symmetry utilized resulted in larger decision at the similar binning stage. Using this refinement setup, all particles had been additional aligned with a masks solely masking the connector area and 6-HB area, which resulted in a median map with a decision of 21.76 Å at a binning stage of eight. Then, particle reextraction was carried out utilizing Warp with a binning stage of 4. Another spherical of native refinement was carried out in RELION, yielding a median map with a decision of 12.7 Å. Then, particle reextraction was carried out once more utilizing Warp with a binning stage of two. In the subsequent native refinement, particles with a shift higher than the imply ±2× SD had been discarded, which resulted in 5,463 particles and a closing map with a decision of 10.9 Å. Postprocessing in RELION with visually estimated B issue was utilized for map sharpening (SI Appendix, Figs. S2 and S6).

Distance Calculation of the Nearest Pair Distances.

For postfusion Ss, the facilities of 2,771 virions that possessed postfusion Ss on their membranes had been manually picked from the deconvolved tomograms. All the postfusion Ss had been assigned to the corresponding virions. For virions that possessed at the very least two postfusion Ss, we calculated the nearest pair distance for every S protein as follows. We chosen one S protein at a time, calculated the distances between its middle and the middle of each different S protein on the similar virion, after which discovered the minimal distance, which was outlined as the nearest pair distance for that S protein. This worth was calculated for all the S proteins, and the outcomes had been plotted right into a histogram (Fig. 1F). The nearest pair distances for prefusion Ss had been calculated in the similar manner (SI Appendix, Fig. S3B).

For the simulated dataset, randomly distributed Ss had been positioned on imaginary sphere-shaped virus with a diameter of 50 nm. The quantity of viruses and the quantity of Ss on every virus had been all stored the similar as the experimental knowledge for postfusion Ss. Then the nearest pair distances had been calculated in the similar manner (Fig. 1F).

Calculation of Distance between Postfusion Ss and AWI.

We manually marked the higher and decrease surfaces of AWI utilizing the slicing device in IMOD (38) by viewing tomograms from the yz airplane. Based on the z axis coordinates of particles and the AWI coordinates of their tomogram, the relative z axis positions could be calculated (SI Appendix, Fig. S3C). The common thickness of all tomograms used in this research had a median thickness of about 120 nm.

Model Fitting and Data Analysis.

The beforehand reported postfusion constructions of the purified SARS-CoV-2 S protein (PDB entries: 6XRA and 6LXT) (5, 15) had been fitted to the map. The prolonged components had been constructed manually utilizing COOT (a molecular graphics utility) (41). The closing mannequin was refined based on the map utilizing PHENIX.Refine (42). Visualization and mannequin evaluation had been carried out with UCSF (University of California, San Francisco) Chimera (43) and UCSF ChimeraX (44). Cross-correlation (CC) values of the mannequin to map had been calculated by PHENIX (45) to 0.73 and 0.77 for CC(masks) and CC(field), respectively.

Data Availability

The atomic mannequin of this research has been deposited in the RCSB (Research Collaboratory for Structural Bioinformatics) PDB underneath accession code 7E9T. The electron density map from this research has been deposited in the Electron Microscopy Data Bank underneath accession code 31037. All uncooked tilt collection used in this research has been deposited in EMPIAR (the Electron Microscopy Public Image Archive) China (http://www.emdb-China.org.cn) underneath accession code EMPIARC-200001.


We thank Ping Shan, Ruigang Su, and Mengyue Lou (F.S. laboratory) for help in laboratory administration. We thank the Center for Biological Imaging, Institute of Biophysics, Chinese Academy of Science for the cryo-EM work, and we’re grateful to Drs. Boling Zhu, Xiaojun Huang, and Gang Ji for assist with cryo-EM knowledge assortment. We are notably grateful to Alister Burt for assist and helpful discussions on picture processing and helpful script growth. We are grateful to Daniel Castaño-Díez, Benjamin Himes, and Dimitry Tegunov for helpful ideas and discussions on Dynamo, emClarity, and Warp utilization. This work was equally supported by grants from National Key Research and Development Program (2017YFA0504700, 2018YFA0900801, and 2020YFA0707500), the Strategic Priority Research Program of the Chinese Academy of Sciences (XDB 37040102 and XDB 29010000), and National Natural Science Foundation of China (31830020). This work was additionally supported by grants from the National Science Fund for Distinguished Young Scholars (31925026), Chinese Academy of Sciences (YSBR-010), National Natural Science Foundation of China (12034006 and 32071187), Emergency Key Program of Guangzhou Laboratory (EKPG21-09), and the National Key Research and Development Program of China (2018YFA0901102 and 2019YFA0904101).


    • Accepted October 19, 2021.
  • Author contributions: Z.R., X.W., F.S., and Y.Z. designed analysis; L.T., G.Z., M.Y., L.C., X.X., G.Y., C.Q., and Y.Z. carried out analysis; L.T., G.Z., C.C., and Y.Z. analyzed knowledge; and L.T., G.Z., F.S., and Y.Z. wrote the paper.

  • The authors declare no competing curiosity.

  • This article is a PNAS Direct Submission.

  • This article accommodates supporting info on-line at https://www.pnas.org/lookup/suppl/doi:10.1073/pnas.2112703118/-/DCSupplemental.

Source link

Load More Related Articles
Load More By James  Smith
Load More In Health News
Comments are closed.

Check Also

Reinfection from Covid-19 is rare, severe disease is even rarer, a study of people in Qatar finds

The study printed Wednesday in the New England Journal of Medicine discovered that there h…