Coronaviruses encode four structural proteins that are incorporated into mature enveloped virions: the transmembrane spike (S), membrane (M), and envelope (E) proteins, and the soluble nucleocapsid (N) protein (Masters, 2006). S proteins protrude from the virion, creating the eponymous corona in electron micrographs, and mediate fusion of viral and host cell membranes. E proteins form cationic viroporins that promote viral assembly and modulate the host immune response. N is an RNA-binding protein that packages the viral RNA genome. M organizes the assembly and structure of new virions and is essential for virus formation (Sturman et al., 1980; Armstrong et al., 1984; Yu et al., 2021; Finkel et al., 2021). M is the most abundant membrane protein in the viral envelope and anti-M antibodies are found in plasma of patients infected with SARS-CoV-2 and other coronaviruses (Godet et al., 1992; He et al., 2005; Hotop et al., 2022; Heffron et al., 2021; Martin et al., 2021; Jörrißen et al., 2021). Based on its functional importance and immunogenicity, M has been proposed as a target for coronavirus vaccines or therapeutics.

In infected cells, M mediates virus assembly and budding by interacting with all other structural proteins and directing their localization to the ER-Golgi intermediate compartment (Masters, 2006; de Haan et al., 1999; Cavanagh, 2005; Siu et al., 2008; Neuman et al., 2011; Kuo et al., 2016a; Kuo et al., 2016b; Lim and Liu, 2001; Boson et al., 2021). M is proposed to interact with E through its transmembrane region and S and N through a cytosolic C-terminal region (Kuo et al., 2016a; Kuo et al., 2016b; Lim and Liu, 2001; Boson et al., 2021). ER export and Golgi localization sequences in M determine its subcellular localization and M, in turn, modulates localization and posttranslational processing of S to promote virion assembly (Boson et al., 2021; Perrier et al., 2019). Across a wide range of coronaviruses (including SARS-CoV-2, SARS-CoV-1, MERS, mouse hepatitis virus [MHV], infectious bronchitis virus, and transmittable gastroenteritis virus), M is required for minimal virus-like particle (VLP) formation in transfected cells (Siu et al., 2008; Vennema et al., 1996; Hsieh et al., 2005; Xu et al., 2020; Plescia et al., 2021). M is insufficient for VLP formation alone, however, and co-required components vary in different systems. SARS-CoV-2 VLP formation requires M co-expression with S or N (Xu et al., 2020; Plescia et al., 2021).

M has further been implicated in modulating host antiviral innate immunity. M inhibits the innate immune response by interfering with MAVS-mediated signaling and interferon production (Fu et al., 2021; Sui et al., 2021). In mouse models of infection, M expression results in lung epithelial cell apoptosis in vitro and in vivo and may contribute to lung injury and pulmonary edema found in severe disease (Sui et al., 2021).

Despite its essential role in viral assembly and implication in pathogenesis, the molecular determinants of M function remain largely unknown. MHV M was proposed to adopt long and compact structures that differentially facilitate membrane bending and recruitment of other structural proteins based on low-resolution tomographic analysis (Neuman et al., 2011). Intriguingly, a structural and evolutionary relationship between SARS-CoV-2 M and the accessory viroporin ORF3a was reported (Tan et al., 2021; Pezeshkian et al., 2021) based on predicted homology to our experimental ORF3a structures (Kern et al., 2021). The manner in which distinct functional roles for M and ORF3a can be achieved in the context of a shared architecture remains to be determined. Here, we report the cryo-EM structure of SARS-CoV-2 M in lipid nanodiscs and perform molecular dynamics (MD) simulations to provide insight into M structure, function, and dynamics.

We determined the structure of SARS-CoV-2 M in lipid nanodiscs (Figure 1). Full-length M was expressed in Spodoptera frugiperda (Sf9) cells with a cleavable C-terminal GFP tag. Gel filtration chromatography of protein extracted in DDM/CHS detergent shows M runs predominantly as a single species consistent with a 50 kDa homodimer. We do not observe evidence of specific higher-order oligomerization at low concentrations by fluorescence size exclusion chromatography or at higher concentrations in large-scale purifications (Figure 1—figure supplement 1). SARS-CoV-2 ORF3a, in contrast, assembles into stable homodimers and homotetramers under similar conditions (Kern et al., 2021).

Figure 1 with 4 supplements see all

Download asset Open asset

We reconstituted homodimeric SARS-CoV-2 M in nanodiscs made from the scaffold protein MSP1E3D1 and lipids (DOPE:POPC:POPS in a 2:1:1 ratio) and determined its structure by cryo-EM (Figure 1, Figure 1—figure supplement 2, Supplementary file 1). The majority of M (189 of 222 amino acids per subunit) was de novo modeled in the cryo-EM map (Figure 1, Figure 1—figure supplement 2). The N-terminus (amino acids 1–16) and C-terminus (amino acids 205–222) are not resolved in the map and were not modeled. Loops connecting transmembrane helices (amino acids 36–42 and 71–78) are the least well-resolved regions of the structure. The relatively weak density is consistent with a lack of stabilizing interactions between these and other M regions and likely indicates they adopt a range of conformations among particles used to generate the final map.

M is ~70 Å tall when viewed from the membrane with an ~40 Å transmembrane spanning region and ~30 Å cytosolic domain (CD) extending into the intracellular solution (Figure 1). Each subunit contains an extracellular or lumenal N-terminus, three transmembrane helices (amino acids 17–36, 43–71, and 79–105) connected by short linkers, and a β-strand rich C-terminal cytosolic domain. We note that AlphaFold and RoseTTAFold predicted M structures diverge substantially from the experimental structure (Figure 1—figure supplement 3; Heo and Feig, 2020; Jumper et al., 2021). In the predicted structures, TM1s are swapped between subunits in addition to differences in the relative positions of transmembrane and cytosolic domains.

Viewed from above, TM1-TM3 from each subunit is positioned along a flattened ellipse with a long major (~50 Å) and short minor (~16 Å) axis. Within each subunit, TM2-TM3 are closely juxtaposed and tightly packed while TM1-TM2 are more distant and loosely connected. The two subunits assemble along their long axis, with TM1 from one protomer forming extensive interactions with the TM2-TM3 unit of the second protomer. As they project toward the cytoplasm, the three transmembrane helices twist counterclockwise and splay outward in the inner leaflet, creating an expanded ellipse with ~55 and ~40 Å axes at the intracellular leaflet.

The transmembrane region is connected to the cytosolic domain through a tight turn-helix-turn segment comprised of residues 106–116. Within the cytosolic domain, each protomer chain forms a pair of opposing β-sheets packed against one another in an eight stranded β-sandwich (Figure 1B and D). The outer sheet is formed by strands β1, β2, β6, the N-terminal half of β7, and the C-terminal half of β8. The inner sheet is formed by strands β3, β4, β5, the C-terminal half of β7, and the N-terminal half of β8. The inner sheets from each protomer interact through a large (~690 Ų buried surface area per chain) and complementary interface with residues L138, V139, V143, L145, F193, A195 contributing to a hydrophobic core surrounded by additional polar interactions.

Using Dali (Holm, 2020) to compare the M structure to all experimentally determined protein structures returns SARS-CoV-2 ORF3a (Kern et al., 2021) as the only structural homolog with a shared fold. Superposition of the two viral proteins reveals a similar fold topology and homodimeric assembly with an overall RMSD of 4.4 Å. Isolated transmembrane and cytosolic domains from individual protomers are better superimposed (RMSD = 2.7 and 2.5 Å, respectively) (Figure 2). Substantial differences in M and ORF3a structure are observed in three regions: TM2-TM3, the transmembrane-cytosolic domain junction, and the cytoplasmic domain interface. TM1s of M and ORF3a are well superimposed, but TM2-TM3 of M are splayed further out into the membrane and are less twisted about the twofold symmetry axis to create a flatter and tighter interaction surface. The angle between TM3 and the cytosolic domain is ~25° more acute in M. The cytosolic domains of M are rotated ~15° away from the symmetry axis relative to ORF3a, shifting the cytosolic domain interface between subunits further from the membrane.

Figure 2

Download asset Open asset

What are the consequences of the structural rearrangement in M relative to ORF3a? Association of subunits in the M homodimer creates a polar and, presumably, water-filled pocket, reminiscent of the polar cavity created between subunits in ORF3a. However, the included volumes are different in several key respects (Figure 3). First, the M pocket is ~⅓ smaller with an enclosed volume of ~840 Å³ compared to ~1300 Å³ in ORF3a. Second, the M pocket is completely sealed by protein to the surrounding membrane and cytoplasm; no openings large enough for water passage are observed connecting the pocket and protein exterior. In contrast, ORF3a displays three pairs of tunnels connecting its internal cavity to the membrane and cytoplasm (two are displayed in Figure 3). Third, the position of the M pocket and ORF3a cavity are different. In M, the gap between subunits is confined to the region between cytosolic domains because transmembrane helices from opposing subunits form tight interactions across the entire lipid bilayer. In ORF3a, the gap extends from the region between cytosolic domains to approximately halfway across the membrane because transmembrane helices are less tightly associated across the membrane inner leaflet.

Figure 3

Download asset Open asset

Another major difference in M and ORf3a structures is shown in Figure 4. The cytosolic domain of M is strikingly electropositive across nearly the entire exposed surface. Electropositive character is contributed by 17 basic amino acids in three surface patches. The first covers the wide face of the cytosolic domains and consists of eight residues (R44, H125, R131, R146, H148, H155, R158, and R198). The second covers the narrow face of M and consists of four residues (R101, R105, R107, and R174). The third covers the underside of M and consists of five residues (K162, K166, K180, R186, and R200). ORF3a, in contrast, presents mixed electrostatic character with electropositive patches closer to the membrane and electronegative patches toward the cytoplasm. Such uniform electropositivity across the M cytosolic surface could facilitate the close juxtaposition of M present at high concentration in viral envelope with the negatively charged viral RNA genome.

Figure 4

Download asset Open asset

The large, complementary, and hydrophobic interface between transmembrane and cytosolic regions of subunits in the M structure suggests a structurally rigid core. However, a previous tomographic study of MHV suggesting that M adopts distinct long and compact structures (Neuman et al., 2011), M’s structural homology to the viroporin ORF3a (Tan et al., 2021; Kern et al., 2021), and the dissociation of cytosolic regions shown in predicted SARS-CoV-2 M structures (Heo and Feig, 2020) suggest the possibility that M is capable of undergoing large-scale structural rearrangements. We next performed MD simulations to explore potential conformational dynamics of M.

We equilibrated M in a lipid environment and ran an all-atom MD simulation for 2 µs. Major conformational changes related to protein function typically take place on a timescale ranging from a microsecond to seconds (Xue et al., 2012; Klepeis et al., 2009; Don et al., 2018). Prior work has reported that microsecond simulations can capture large conformational changes of membrane proteins in a lipid bilayer (Monje-Galvan and Voth, 2021; Brandner et al., 2019; Nury et al., 2010; Monje-Galvan and Voth, 2020). We therefore expected to capture a subset of possible structural changes in our simulation that correspond to relatively frequently accessed states. Overall, we did not observe substantial conformational rearrangements in M during the simulation (Figure 5A and B). Superposition of the experimental and final M structure following the simulation shows minor deviations through most of the protein (overall RMSD of 2.5 Å) (Figure 5A). The largest difference is a shift in TM1 up toward the extracellular/lumenal side by approximately half a helical turn, enabled by rearrangement of the TM1-TM2 linker (Figure 5A). This relatively subtle movement is consistent with weaker density for the TM1-TM2 linker in the cryo-EM map and fewer packing interactions for TM1 than TM2 or TM3. Per residue deviations ranged from ~1 to 4 Å and, aside from the movement of TM1, were similar between subunits and largest in the TM2-TM3 linker, transmembrane to cytosolic region connection, and loops connecting strands in the cytosolic domain. Minimal structural deviation was observed during the simulation within or between subunits as judged by the number of close Cα contacts, the angle between transmembrane and cytosolic regions, the distance between transmembrane regions, or the distance between cytosolic domains (Figure 5D–H). Consistent with limited movement of the transmembrane region and a lack of evidence for lipid binding in the cryo-EM structure, no obvious enrichment of specific lipids around M was identified following simulation (Figure 5—figure supplement 1). Finally, the internal M pocket remained similar in size and sealed from the surrounding solution throughout the simulation (Figure 5I and J). Taken together, these data suggest that M adopts a stable structure with minimal dynamic rearrangements on the microsecond timescale, at physiological temperature, and in the absence of additional binding proteins. These results are consistent with the absence of other highly populated conformations in our cryo-EM data, though we cannot exclude the possibility that M undergoes rare large-scale conformational changes under these conditions not captured in the simulation.

Figure 5 with 1 supplement see all

Download asset Open asset

The structure of the SARS-CoV-2 M protein that we have obtained by cryo-EM reveals a homodimeric fold that is structurally homologous to the nonselective Ca²⁺ permeable cation channel of SARS-CoV-2, ORF3a. As with 3a, each subunit of M contains three transmembrane helices and a C-terminal β-sandwich domain. However, the structure differs from ORF3a in several key ways that provide insight into how these structurally similar proteins can fill drastically different apparent roles in the coronavirus life cycle.

When viewed from the plane of the membrane, M is considerably wider and flatter than ORF3a, due to differences in transmembrane helix packing and a rotation about the central axis of the cytosolic domain. Among the consequences of this flattening out of M are distinct differences in the dimer interface across the membrane, where M shows a tighter dimer interface closer to the membrane outer leaflet as well as a gap between cytosolic domain subunits that forms an enclosed pocket lined by polar residues. In ORF3a, transmembrane regions are less closely opposed and a gap between subunits extends from halfway across the membrane to halfway down the cytosolic domains. The result is a larger cavity that is open to the membrane and cytoplasm. Mutations in the ORF3a cavity alter ion channel activity, consistent with the cavity forming part of the conduction path. Tight subunit association may therefore be important for the structural role of M, while loose subunit association that creates a large open cavity may be essential for the viroporin activity of ORF3a.

In further contrast to ORF3a, which was seen to form stable tetramers through electrostatic interactions between neighboring dimers, we see no evidence that M forms higher-order oligomers under similar experimental conditions. Surface characteristics of the M dimer lend credence to the possibility that M exists solely as a dimer in the membrane – one striking feature of the M C-terminal β-sandwich domain is the presence of three sizable patches of positive charge that dominate its solvent exposed surface. MD simulations of M show that the dimer is stable and does not readily adopt alternate conformations at physiological temperature over the 1.6 µs trajectory. These data suggest that M could play a structural role in mediating morphological changes in host cell membranes through interactions with other SARS-CoV-2 structural proteins and perhaps negatively charged lipid headgroups or viral RNA.

A complementary study published during review of this work includes two structures of SARS-CoV-2 M solubilized in detergent micelles and in complex with Fab antibody fragments bound to the cytosolic domain (Zhang et al., 2022). One structure (termed short form) is similar to that reported here. The second structure (termed long form) adopts a distinct conformation that is taller and narrower in the membrane. The conformations are suggestive of short and long forms of MHV M observed in low-resolution tomograms (Neuman et al., 2011). Our data suggests that M in a simple lipid mixture predominantly adopts the short form. It may be that the alternate M long form, stabilized by Fab binding for structure determination, is similarly promoted in a physiological context by interactions with lipids, spike, N, or viral RNA. Evidence for possible higher-order oligomerization was also reported in some 2D classes of M solubilized in a different detergent, suggesting that M-M interactions may occur under certain biochemical conditions. Whether distinct M conformations have distinct functional roles in virus assembly, as has been suggested for MHV M, remains to be determined.

M has been shown to play a crucial role in viral assembly through protein-protein interactions with other coronavirus structural proteins such as N and S. Spike proteins are incorporated into coronavirus virions via interactions between the cytosolic tail of S and the cytosolic domain of M, however the precise details of this interaction are unknown (Boson et al., 2021). In SARS-CoV-2, M and N or S are the minimal components required for forming VLPs when expressed heterologously in cells (Xu et al., 2020; Plescia et al., 2021). Several recent studies have suggested that the C-terminal domain of N is the site of interaction between SARS-CoV-2 M and N, but as with S a precise binding site has not been established (Cubuk et al., 2021; Lu et al., 2021). It is possible that M-N interactions are mediated by favorable electrostatic interactions between negatively charged residues of the N CTD and one or more of the basic patches identified on the surface of the cytosolic domain of M, a model supported by co-immunoprecipitation of N and M proteins expressed in 293T cells (Zhang et al., 2022). Through the sheer abundance of M dimers found in the membrane of SARS-CoV-2 virions, M and N together might facilitate VLP formation via a mechanism similar to the Gag precursor of HIV, where the high concentration of M C-terminal domains at the cytoplasmic membrane surface recruit and organize many N proteins that together physically extrude a membranous bud.

At present the World Health Organization puts the confirmed number of COVID-19 cases worldwide at nearly 530 million. Over the last 2 years the SARS-CoV-2 virus has undergone many mutations that have been extensively documented through sequencing efforts worldwide (Hadfield et al., 2018). Despite this, the M protein sequence has remained virtually unchanged – a testament to the critical role that M plays in viral replication and assembly (Cagliani et al., 2020). Furthermore, while only 20 amino acids in length, the N-terminus of M has been found to be highly immunogenic in COVID-19 patients (Hotop et al., 2022; Heffron et al., 2021; Martin et al., 2021; Jörrißen et al., 2021). M has also been shown to modulate innate immune response and could contribute to lung injury often seen in severe cases (Fu et al., 2021; Sui et al., 2021). Given its clear importance in the coronavirus life cycle and pathogenicity, M presents an attractive target for therapeutics or vaccines. While M is well conserved across Coronaviridae (Figure 1—figure supplement 4), it shows particularly high conservation between SARS-CoV-1 and SARS-CoV-2, with a sequence similarity of 90.54%, highlighting its potential as a therapeutic target for emergent coronaviruses in the future.

All data and reagents associated with this study are publicly available. The final model is in the PDB under 8CTK, the final map is in the EMDB under EMD-26993, and micrographs (original and motion corrected) and final particle stack are deposited in EMPIAR under 11067.

1. 1. Dolan KA
   2. Brohawn SG

   (2022) RCSB Protein Data Bank

   ID 8CTK. Cryo-EM structure of SARS-CoV-2 M protein in a lipid nanodisc.

   https://www.rcsb.org/structure/8CTK
2. 1. Dolan KA
   2. Brohawn SG

   (2022) Electron Microscopy Data Bank

   ID EMD-26993. Cryo-EM structure of SARS-CoV-2 M protein in a lipid nanodisc.

   https://www.ebi.ac.uk/emdb/EMD-26993
3. 1. Dolan KA
   2. Dutta M
   3. Kern DM
   4. Kotecha A
   5. Voth GA
   6. Brohawn SG

   (2022) Electron Microscopy Public Image Archive

   ID EMPIAR-11067. Structure of SARS-CoV-2 M protein in lipid nanodiscs.

   https://www.ebi.ac.uk/empiar/EMPIAR-11067/

1. 1. Abraham MJ
   2. Murtola T
   3. Schulz R
   4. Páll S
   5. Smith JC
   6. Hess B
   7. Lindahl E

   (2015) GROMACS: high performance molecular simulations through multi-level parallelism from laptops to supercomputers

   SoftwareX 1–2:19–25.

   https://doi.org/10.1016/j.softx.2015.06.001

   • Google Scholar
2. 1. Afonine PV
   2. Poon BK
   3. Read RJ
   4. Sobolev OV
   5. Terwilliger TC
   6. Urzhumtsev A
   7. Adams PD

   (2018) Real-Space refinement in PHENIX for cryo-EM and crystallography

   Acta Crystallographica. Section D, Structural Biology 74:531–544.

   https://doi.org/10.1107/S2059798318006551

   • PubMed
   • Google Scholar
3. 1. Armstrong J
   2. Niemann H
   3. Smeekens S
   4. Rottier P
   5. Warren G

   (1984) Sequence and topology of a model intracellular membrane protein, E1 glycoprotein, from a coronavirus

   Nature 308:751–752.

   https://doi.org/10.1038/308751a0

   • PubMed
   • Google Scholar
4. 1. Bepler T
   2. Morin A
   3. Rapp M
   4. Brasch J
   5. Shapiro L
   6. Noble AJ
   7. Berger B

   (2019) Positive-unlabeled convolutional neural networks for particle picking in cryo-electron micrographs

   Nature Methods 16:1153–1160.

   https://doi.org/10.1038/s41592-019-0575-8

   • PubMed
   • Google Scholar
5. 1. Berendsen HJC
   2. Postma JPM
   3. van Gunsteren WF
   4. DiNola A
   5. Haak JR

   (1984) Molecular dynamics with coupling to an external Bath

   J Chem Phys 81:3684–3690.

   https://doi.org/10.1063/1.448118

   • Google Scholar
6. 1. Boson B
   2. Legros V
   3. Zhou B
   4. Siret E
   5. Mathieu C
   6. Cosset FL
   7. Lavillette D
   8. Denolly S

   (2021) The SARS-cov-2 envelope and membrane proteins modulate maturation and retention of the spike protein, allowing assembly of virus-like particles

   The Journal of Biological Chemistry 296:100111.

   https://doi.org/10.1074/jbc.RA120.016175

   • PubMed
   • Google Scholar
7. 1. Brandner A
   2. De Vecchis D
   3. Baaden M
   4. Cohen MM
   5. Taly A

   (2019) Structural dataset from microsecond-long simulations of yeast mitofusin fzo1 in the context of membrane docking

   Data in Brief 26:104460.

   https://doi.org/10.1016/j.dib.2019.104460

   • PubMed
   • Google Scholar
8. 1. Brooks BR
   2. Brooks CL
   3. Mackerell AD
   4. Nilsson L
   5. Petrella RJ
   6. Roux B
   7. Won Y
   8. Archontis G
   9. Bartels C
   10. Boresch S
   11. Caflisch A
   12. Caves L
   13. Cui Q
   14. Dinner AR
   15. Feig M
   16. Fischer S
   17. Gao J
   18. Hodoscek M
   19. Im W
   20. Kuczera K
   21. Lazaridis T
   22. Ma J
   23. Ovchinnikov V
   24. Paci E
   25. Pastor RW
   26. Post CB
   27. Pu JZ
   28. Schaefer M
   29. Tidor B
   30. Venable RM
   31. Woodcock HL
   32. Wu X
   33. Yang W
   34. York DM
   35. Karplus M

   (2009) CHARMM: the biomolecular simulation program

   Journal of Computational Chemistry 30:1545–1614.

   https://doi.org/10.1002/jcc.21287

   • PubMed
   • Google Scholar
9. 1. Cagliani R
   2. Forni D
   3. Clerici M
   4. Sironi M

   (2020) Computational inference of selection underlying the evolution of the novel coronavirus, severe acute respiratory syndrome coronavirus 2

   Journal of Virology 94:e00411-20.

   https://doi.org/10.1128/JVI.00411-20

   • PubMed
   • Google Scholar
10. 1. Cavanagh D

    (2005) Coronaviridae: a review of coronaviruses and toroviruses: coronaviruses special emphas first insights concern sars

    Nature 1:1–54.

    https://doi.org/10.1007/b137625

    • Google Scholar
11. 1. Cubuk J
    2. Alston JJ
    3. Incicco JJ
    4. Singh S
    5. Stuchell-Brereton MD
    6. Ward MD
    7. Zimmerman MI
    8. Vithani N
    9. Griffith D
    10. Wagoner JA
    11. Bowman GR
    12. Hall KB
    13. Soranno A
    14. Holehouse AS

    (2021) The SARS-cov-2 nucleocapsid protein is dynamic, disordered, and phase separates with RNA

    Nature Communications 12:1936.

    https://doi.org/10.1038/s41467-021-21953-3

    • PubMed
    • Google Scholar
12. 1. Darden T
    2. York D
    3. Pedersen L

    (1993) Particle mesh ewald: an N ⋅log(N) method for ewald sums in large systems

    J Chem Phys 98:10089–10092.

    https://doi.org/10.1063/1.464397

    • Google Scholar
13. 1. de Haan CA
    2. Smeets M
    3. Vernooij F
    4. Vennema H
    5. Rottier PJ

    (1999) Mapping of the coronavirus membrane protein domains involved in interaction with the spike protein

    Journal of Virology 73:7441–7452.

    https://doi.org/10.1128/JVI.73.9.7441-7452.1999

    • PubMed
    • Google Scholar
14. 1. Don CG
    2. Smieško M
    3. Machuqueiro M

    (2018) Microsecond MD simulations of human CYP2D6 wild-type and five allelic variants reveal mechanistic insights on the function

    PLOS ONE 13:e0202534.

    https://doi.org/10.1371/journal.pone.0202534

    • PubMed
    • Google Scholar
15. 1. Emsley P
    2. Lohkamp B
    3. Scott WG
    4. Cowtan K

    (2010) Features and development of coot

    Acta Crystallographica. Section D, Biological Crystallography 66:486–501.

    https://doi.org/10.1107/S0907444910007493

    • PubMed
    • Google Scholar
16. 1. Finkel Y
    2. Mizrahi O
    3. Nachshon A
    4. Weingarten-Gabbay S
    5. Morgenstern D
    6. Yahalom-Ronen Y
    7. Tamir H
    8. Achdout H
    9. Stein D
    10. Israeli O
    11. Beth-Din A
    12. Melamed S
    13. Weiss S
    14. Israely T
    15. Paran N
    16. Schwartz M
    17. Stern-Ginossar N

    (2021) The coding capacity of SARS-cov-2

    Nature 589:125–130.

    https://doi.org/10.1038/s41586-020-2739-1

    • PubMed
    • Google Scholar
17. 1. Fu YZ
    2. Wang SY
    3. Zheng ZQ
    4. Li WW
    5. Xu ZS
    6. Wang YY

    (2021) SARS-cov-2 membrane glycoprotein M antagonizes the MAVS-mediated innate antiviral response

    Cellular & Molecular Immunology 18:613–620.

    https://doi.org/10.1038/s41423-020-00571-x

    • PubMed
    • Google Scholar
18. 1. Goddard TD
    2. Huang CC
    3. Meng EC
    4. Pettersen EF
    5. Couch GS
    6. Morris JH
    7. Ferrin TE

    (2018) UCSF chimerax: meeting modern challenges in visualization and analysis

    Protein Science 27:14–25.

    https://doi.org/10.1002/pro.3235

    • PubMed
    • Google Scholar
19. 1. Godet M
    2. L’Haridon R
    3. Vautherot JF
    4. Laude H

    (1992) Tgev corona virus ORF4 encodes a membrane protein that is incorporated into virions

    Virology 188:666–675.

    https://doi.org/10.1016/0042-6822(92)90521-p

    • PubMed
    • Google Scholar
20. 1. Guo H
    2. Franken E
    3. Deng Y
    4. Benlekbir S
    5. Singla Lezcano G
    6. Janssen B
    7. Yu L
    8. Ripstein ZA
    9. Tan YZ
    10. Rubinstein JL

    (2020) Electron-event representation data enable efficient cryoEM file storage with full preservation of spatial and temporal resolution

    IUCrJ 7:860–869.

    https://doi.org/10.1107/S205225252000929X

    • PubMed
    • Google Scholar
21. 1. Hadfield J
    2. Megill C
    3. Bell SM
    4. Huddleston J
    5. Potter B
    6. Callender C
    7. Sagulenko P
    8. Bedford T
    9. Neher RA

    (2018) Nextstrain: real-time tracking of pathogen evolution

    Bioinformatics 34:4121–4123.

    https://doi.org/10.1093/bioinformatics/bty407

    • PubMed
    • Google Scholar
22. 1. He Y
    2. Zhou Y
    3. Siddiqui P
    4. Niu J
    5. Jiang S

    (2005) Identification of immunodominant epitopes on the membrane protein of the severe acute respiratory syndrome-associated coronavirus

    Journal of Clinical Microbiology 43:3718–3726.

    https://doi.org/10.1128/JCM.43.8.3718-3726.2005

    • PubMed
    • Google Scholar
23. 1. Heffron AS
    2. McIlwain SJ
    3. Amjadi MF
    4. Baker DA
    5. Khullar S
    6. Armbrust T
    7. Halfmann PJ
    8. Kawaoka Y
    9. Sethi AK
    10. Palmenberg AC
    11. Shelef MA
    12. O’Connor DH
    13. Ong IM

    (2021) The landscape of antibody binding in SARS-cov-2 infection

    PLOS Biology 19:e3001265.

    https://doi.org/10.1371/journal.pbio.3001265

    • PubMed
    • Google Scholar
24. Preprint

    1. Heo L
    2. Feig M

    (2020) Modeling of Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-Cov-2) Proteins by Machine Learning and Physics-Based Refinement

    bioRxiv.

    https://doi.org/10.1101/2020.03.25.008904

    • Google Scholar
25. 1. Hess B
    2. Bekker H
    3. Berendsen HJC
    4. Fraaije J

    (1997) LINCS: a linear constraint solver for molecular simulations

    Journal of Computational Chemistry 18:1463–1472.

    https://doi.org/10.1002/(SICI)1096-987X(199709)

    • Google Scholar
26. 1. Holm L

    (2020) Structural bioinformatics, methods and protocols

    Methods Mol Biology 2112:29–42.

    https://doi.org/10.1007/978-1-0716-0270-6

    • Google Scholar
27. 1. Hoover WG

    (1985) Canonical dynamics: equilibrium phase-space distributions

    Physical Review. A, General Physics 31:1695–1697.

    https://doi.org/10.1103/physreva.31.1695

    • PubMed
    • Google Scholar
28. 1. Hotop SK
    2. Reimering S
    3. Shekhar A
    4. Asgari E
    5. Beutling U
    6. Dahlke C
    7. Fathi A
    8. Khan F
    9. Lütgehetmann M
    10. Ballmann R
    11. Gerstner A
    12. Tegge W
    13. Cicin-Sain L
    14. Bilitewski U
    15. McHardy AC
    16. Brönstrup M

    (2022) Peptide microarrays coupled to machine learning reveal individual epitopes from human antibody responses with neutralizing capabilities against SARS-cov-2

    Emerging Microbes & Infections 11:1037–1048.

    https://doi.org/10.1080/22221751.2022.2057874

    • PubMed
    • Google Scholar
29. 1. Hsieh PK
    2. Chang SC
    3. Huang CC
    4. Lee TT
    5. Hsiao CW
    6. Kou YH
    7. Chen IY
    8. Chang CK
    9. Huang TH
    10. Chang MF

    (2005) Assembly of severe acute respiratory syndrome coronavirus RNA packaging signal into virus-like particles is nucleocapsid dependent

    Journal of Virology 79:13848–13855.

    https://doi.org/10.1128/JVI.79.22.13848-13855.2005

    • PubMed
    • Google Scholar
30. 1. Huang J
    2. Rauscher S
    3. Nawrocki G
    4. Ran T
    5. Feig M
    6. de Groot BL
    7. Grubmüller H
    8. MacKerell AD Jr

    (2017) CHARMM36m: an improved force field for folded and intrinsically disordered proteins

    Nature Methods 14:71–73.

    https://doi.org/10.1038/nmeth.4067

    • PubMed
    • Google Scholar
31. 1. Humphrey W
    2. Dalke A
    3. Schulten K

    (1996) VMD: visual molecular dynamics

    Journal of Molecular Graphics 14:33–38.

    https://doi.org/10.1016/0263-7855(96)00018-5

    • PubMed
    • Google Scholar
32. 1. Jo S
    2. Kim T
    3. Im W

    (2007) Automated builder and database of protein/membrane complexes for molecular dynamics simulations

    PLOS ONE 2:e880.

    https://doi.org/10.1371/journal.pone.0000880

    • PubMed
    • Google Scholar
33. 1. Jo S
    2. Kim T
    3. Iyer VG
    4. Im W

    (2008) CHARMM-GUI: a web-based graphical user interface for CHARMM

    Journal of Computational Chemistry 29:1859–1865.

    https://doi.org/10.1002/jcc.20945

    • PubMed
    • Google Scholar
34. 1. Jo S
    2. Lim JB
    3. Klauda JB
    4. Im W

    (2009) CHARMM-GUI membrane builder for mixed bilayers and its application to yeast membranes

    Biophysical Journal 97:50–58.

    https://doi.org/10.1016/j.bpj.2009.04.013

    • PubMed
    • Google Scholar
35. 1. Jörrißen P
    2. Schütz P
    3. Weiand M
    4. Vollenberg R
    5. Schrempf IM
    6. Ochs K
    7. Frömmel C
    8. Tepasse P-R
    9. Schmidt H
    10. Zibert A

    (2021) Antibody response to SARS-cov-2 membrane protein in patients of the acute and convalescent phase of COVID-19

    Frontiers in Immunology 12:679841.

    https://doi.org/10.3389/fimmu.2021.679841

    • PubMed
    • Google Scholar
36. 1. Jumper J
    2. Evans R
    3. Pritzel A
    4. Green T
    5. Figurnov M
    6. Ronneberger O
    7. Tunyasuvunakool K
    8. Bates R
    9. Žídek A
    10. Potapenko A
    11. Bridgland A
    12. Meyer C
    13. Kohl SAA
    14. Ballard AJ
    15. Cowie A
    16. Romera-Paredes B
    17. Nikolov S
    18. Jain R
    19. Adler J
    20. Back T
    21. Petersen S
    22. Reiman D
    23. Clancy E
    24. Zielinski M
    25. Steinegger M
    26. Pacholska M
    27. Berghammer T
    28. Bodenstein S
    29. Silver D
    30. Vinyals O
    31. Senior AW
    32. Kavukcuoglu K
    33. Kohli P
    34. Hassabis D

    (2021) Highly accurate protein structure prediction with alphafold

    Nature 596:583–589.

    https://doi.org/10.1038/s41586-021-03819-2

    • PubMed
    • Google Scholar
37. 1. Kern DM
    2. Brohawn SG

    (2021) SARS-cov-2 3A expression, purification, and reconstitution into lipid nanodiscs

    Methods in Enzymology 653:207–235.

    https://doi.org/10.1016/bs.mie.2020.12.020

    • PubMed
    • Google Scholar
38. 1. Kern DM
    2. Sorum B
    3. Mali SS
    4. Hoel CM
    5. Sridharan S
    6. Remis JP
    7. Toso DB
    8. Kotecha A
    9. Bautista DM
    10. Brohawn SG

    (2021) Cryo-Em structure of SARS-cov-2 orf3a in lipid nanodiscs

    Nature Structural & Molecular Biology 28:573–582.

    https://doi.org/10.1038/s41594-021-00619-0

    • PubMed
    • Google Scholar
39. 1. Klepeis JL
    2. Lindorff-Larsen K
    3. Dror RO
    4. Shaw DE

    (2009) Long-timescale molecular dynamics simulations of protein structure and function

    Current Opinion in Structural Biology 19:120–127.

    https://doi.org/10.1016/j.sbi.2009.03.004

    • PubMed
    • Google Scholar
40. 1. Kuo L
    2. Hurst-Hess KR
    3. Koetzner CA
    4. Masters PS

    (2016a) Analyses of coronavirus assembly interactions with interspecies membrane and nucleocapsid protein chimeras

    Journal of Virology 90:4357–4368.

    https://doi.org/10.1128/JVI.03212-15

    • PubMed
    • Google Scholar
41. 1. Kuo L
    2. Koetzner CA
    3. Masters PS

    (2016b) A key role for the carboxy-terminal tail of the murine coronavirus nucleocapsid protein in coordination of genome packaging

    Virology 494:100–107.

    https://doi.org/10.1016/j.virol.2016.04.009

    • PubMed
    • Google Scholar
42. 1. Lee J
    2. Cheng X
    3. Swails JM
    4. Yeom MS
    5. Eastman PK
    6. Lemkul JA
    7. Wei S
    8. Buckner J
    9. Jeong JC
    10. Qi Y
    11. Jo S
    12. Pande VS
    13. Case DA
    14. Brooks CL
    15. MacKerell AD
    16. Klauda JB
    17. Im W

    (2016) CHARMM-GUI input generator for NAMD, GROMACS, amber, openmm, and CHARMM/openmm simulations using the CHARMM36 additive force field

    Journal of Chemical Theory and Computation 12:405–413.

    https://doi.org/10.1021/acs.jctc.5b00935

    • PubMed
    • Google Scholar
43. 1. Lim KP
    2. Liu DX

    (2001) The missing link in coronavirus assembly: retention of the avian coronavirus infectious bronchitis virus envelope protein in the pre-golgi compartments and physical interaction between the envelope and membrane proteins

    The Journal of Biological Chemistry 276:17515–17523.

    https://doi.org/10.1074/jbc.M009731200

    • PubMed
    • Google Scholar
44. 1. Lu S
    2. Ye Q
    3. Singh D
    4. Cao Y
    5. Diedrich JK
    6. Yates JR
    7. Villa E
    8. Cleveland DW
    9. Corbett KD

    (2021) The SARS-cov-2 nucleocapsid phosphoprotein forms mutually exclusive condensates with RNA and the membrane-associated M protein

    Nature Communications 12:502.

    https://doi.org/10.1038/s41467-020-20768-y

    • PubMed
    • Google Scholar
45. 1. Martin S
    2. Heslan C
    3. Jégou G
    4. Eriksson LA
    5. Le Gallo M
    6. Thibault V
    7. Chevet E
    8. Godey F
    9. Avril T

    (2021) SARS-cov-2 integral membrane proteins shape the serological responses of patients with COVID-19

    IScience 24:103185.

    https://doi.org/10.1016/j.isci.2021.103185

    • PubMed
    • Google Scholar
46. 1. Masters PS

    (2006) The molecular biology of coronaviruses

    Advances in Virus Research 66:193–292.

    https://doi.org/10.1016/S0065-3527(06)66005-3

    • PubMed
    • Google Scholar
47. 1. Michaud-Agrawal N
    2. Denning EJ
    3. Woolf TB
    4. Beckstein O

    (2011) MDAnalysis: a toolkit for the analysis of molecular dynamics simulations

    Journal of Computational Chemistry 32:2319–2327.

    https://doi.org/10.1002/jcc.21787

    • PubMed
    • Google Scholar
48. 1. Monje-Galvan V
    2. Voth GA

    (2020) Binding mechanism of the matrix domain of HIV-1 gag on lipid membranes

    eLife 9:e58621.

    https://doi.org/10.7554/eLife.58621

    • PubMed
    • Google Scholar
49. 1. Monje-Galvan V
    2. Voth GA

    (2021) Molecular interactions of the M and E integral membrane proteins of SARS-cov-2

    Faraday Discussions 232:49–67.

    https://doi.org/10.1039/D1FD00031D

    • PubMed
    • Google Scholar
50. 1. Neuman BW
    2. Kiss G
    3. Kunding AH
    4. Bhella D
    5. Baksh MF
    6. Connelly S
    7. Droese B
    8. Klaus JP
    9. Makino S
    10. Sawicki SG
    11. Siddell SG
    12. Stamou DG
    13. Wilson IA
    14. Kuhn P
    15. Buchmeier MJ

    (2011) A structural analysis of M protein in coronavirus assembly and morphology

    Journal of Structural Biology 174:11–22.

    https://doi.org/10.1016/j.jsb.2010.11.021

    • PubMed
    • Google Scholar
51. 1. Nosé S
    2. Klein ML

    (1983) Constant pressure molecular dynamics for molecular systems

    Molecular Physics 50:1055–1076.

    https://doi.org/10.1080/00268978300102851

    • Google Scholar
52. 1. Nosé S

    (1984) A molecular dynamics method for simulations in the canonical ensemble

    Molecular Physics 52:255–268.

    https://doi.org/10.1080/00268978400101201

    • Google Scholar
53. 1. Nury H
    2. Poitevin F
    3. Van Renterghem C
    4. Changeux JP
    5. Corringer PJ
    6. Delarue M
    7. Baaden M

    (2010) One-microsecond molecular dynamics simulation of channel gating in a nicotinic receptor homologue

    PNAS 107:6275–6280.

    https://doi.org/10.1073/pnas.1001832107

    • PubMed
    • Google Scholar
54. 1. Parrinello M
    2. Rahman A

    (1981) Polymorphic transitions in single crystals: a new molecular dynamics method

    Journal of Applied Physics 52:7182–7190.

    https://doi.org/10.1063/1.328693

    • Google Scholar
55. 1. Perrier A
    2. Bonnin A
    3. Desmarets L
    4. Danneels A
    5. Goffard A
    6. Rouillé Y
    7. Dubuisson J
    8. Belouzard S

    (2019) The C-terminal domain of the MERS coronavirus M protein contains a trans-Golgi network localization signal

    The Journal of Biological Chemistry 294:14406–14421.

    https://doi.org/10.1074/jbc.RA119.008964

    • PubMed
    • Google Scholar
56. Preprint

    1. Pezeshkian W
    2. Grünewald F
    3. Narykov O
    4. Lu S
    5. Arkhipova V
    6. Solodovnikov A
    7. Wassenaar TA
    8. Marrink SJ
    9. Korkin D

    (2021) Molecular Architecture and Dynamics of SARS-CoV-2 Envelope by Integrative Modeling

    bioRxiv.

    https://doi.org/10.1101/2021.09.15.459697

    • Google Scholar
57. 1. Plescia CB
    2. David EA
    3. Patra D
    4. Sengupta R
    5. Amiar S
    6. Su Y
    7. Stahelin RV

    (2021) SARS-cov-2 viral budding and entry can be modeled using BSL-2 level virus-like particles

    The Journal of Biological Chemistry 296:100103.

    https://doi.org/10.1074/jbc.RA120.016148

    • PubMed
    • Google Scholar
58. 1. Ritchie TK
    2. Grinkova YV
    3. Bayburt TH
    4. Denisov IG
    5. Zolnerciks JK
    6. Atkins WM
    7. Sligar SG

    (2009) Chapter 11-reconstitution of membrane proteins in phospholipid bilayer nanodiscs

    Methods in Enzymology 464:211–231.

    https://doi.org/10.1016/S0076-6879(09)64011-8

    • PubMed
    • Google Scholar
59. 1. Siu YL
    2. Teoh KT
    3. Lo J
    4. Chan CM
    5. Kien F
    6. Escriou N
    7. Tsao SW
    8. Nicholls JM
    9. Altmeyer R
    10. Peiris JSM
    11. Bruzzone R
    12. Nal B

    (2008) The M, E, and N structural proteins of the severe acute respiratory syndrome coronavirus are required for efficient assembly, trafficking, and release of virus-like particles

    Journal of Virology 82:11318–11330.

    https://doi.org/10.1128/JVI.01052-08

    • PubMed
    • Google Scholar
60. 1. Smart OS
    2. Neduvelil JG
    3. Wang X
    4. Wallace BA
    5. Sansom MS

    (1996) Hole: a program for the analysis of the pore dimensions of ion channel structural models

    Journal of Molecular Graphics 14:354–360.

    https://doi.org/10.1016/s0263-7855(97)00009-x

    • PubMed
    • Google Scholar
61. 1. Sturman LS
    2. Holmes KV
    3. Behnke J

    (1980) Isolation of coronavirus envelope glycoproteins and interaction with the viral nucleocapsid

    Journal of Virology 33:449–462.

    https://doi.org/10.1128/JVI.33.1.449-462.1980

    • PubMed
    • Google Scholar
62. 1. Sui L
    2. Zhao Y
    3. Wang W
    4. Wu P
    5. Wang Z
    6. Yu Y
    7. Hou Z
    8. Tan G
    9. Liu Q

    (2021) SARS-cov-2 membrane protein inhibits type I interferon production through ubiquitin-mediated degradation of TBK1

    Frontiers in Immunology 12:662989.

    https://doi.org/10.3389/fimmu.2021.662989

    • PubMed
    • Google Scholar
63. 1. Tan Y
    2. Schneider T
    3. Shukla PK
    4. Chandrasekharan MB
    5. Aravind L
    6. Zhang D

    (2021) Unification and extensive diversification of M/orf3-related ion channel proteins in coronaviruses and other nidoviruses

    Virus Evolution 7:veab014.

    https://doi.org/10.1093/ve/veab014

    • PubMed
    • Google Scholar
64. 1. Tian W
    2. Chen C
    3. Lei X
    4. Zhao J
    5. Liang J

    (2018) CASTp 3.0: computed atlas of surface topography of proteins

    Nucleic Acids Research 46:W363–W367.

    https://doi.org/10.1093/nar/gky473

    • PubMed
    • Google Scholar
65. 1. Vennema H
    2. Godeke GJ
    3. Rossen JW
    4. Voorhout WF
    5. Horzinek MC
    6. Opstelten DJ
    7. Rottier PJ

    (1996)

    Nucleocapsid-independent assembly of coronavirus-like particles by co-expression of viral envelope protein genes

    The EMBO Journal 15:2020–2028.

    • PubMed
    • Google Scholar
66. 1. Williams CJ
    2. Headd JJ
    3. Moriarty NW
    4. Prisant MG
    5. Videau LL
    6. Deis LN
    7. Verma V
    8. Keedy DA
    9. Hintze BJ
    10. Chen VB
    11. Jain S
    12. Lewis SM
    13. Arendall WB
    14. Snoeyink J
    15. Adams PD
    16. Lovell SC
    17. Richardson JS
    18. Richardson DC

    (2018) MolProbity: more and better reference data for improved all-atom structure validation

    Protein Science 27:293–315.

    https://doi.org/10.1002/pro.3330

    • PubMed
    • Google Scholar
67. 1. Xu R
    2. Shi M
    3. Li J
    4. Song P
    5. Li N

    (2020) Construction of SARS-cov-2 virus-like particles by mammalian expression system

    Frontiers in Bioengineering and Biotechnology 8:862.

    https://doi.org/10.3389/fbioe.2020.00862

    • PubMed
    • Google Scholar
68. 1. Xue Y
    2. Ward JM
    3. Yuwen T
    4. Podkorytov IS
    5. Skrynnikov NR

    (2012) Microsecond time-scale conformational exchange in proteins: using long molecular dynamics trajectory to simulate NMR relaxation dispersion data

    Journal of the American Chemical Society 134:2555–2562.

    https://doi.org/10.1021/ja206442c

    • PubMed
    • Google Scholar
69. 1. Yu A
    2. Pak AJ
    3. He P
    4. Monje-Galvan V
    5. Casalino L
    6. Gaieb Z
    7. Dommer AC
    8. Amaro RE
    9. Voth GA

    (2021) A multiscale coarse-grained model of the SARS-cov-2 virion

    Biophysical Journal 120:1097–1104.

    https://doi.org/10.1016/j.bpj.2020.10.048

    • PubMed
    • Google Scholar
70. 1. Zhang Z
    2. Nomura N
    3. Muramoto Y
    4. Ekimoto T
    5. Uemura T
    6. Liu K
    7. Yui M
    8. Kono N
    9. Aoki J
    10. Ikeguchi M
    11. Noda T
    12. Iwata S
    13. Ohto U
    14. Shimizu T

    (2022) Structure of SARS-cov-2 membrane protein essential for virus assembly

    Nature Communications 13:4399.

    https://doi.org/10.1038/s41467-022-32019-3

    • PubMed
    • Google Scholar

Author details

1. Kimberly A Dolan

   1. Biophysics Graduate Group, University of California, Berkeley, Berkeley, United States
   2. Department of Molecular and Cell Biology, Helen Wills Neuroscience Institute, and California Institute for Quantitative Biosciences (QB3), University of California, Berkeley, Berkeley, United States

   Contribution

   Conceived of the project, generated final constructs and performed expression, purification, reconstitution, cryo-EM sample preparation, and cryo-EM data processing, built and refined the atomic model, and wrote the manuscript with input from all authors

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-9335-7791

2. Mandira Dutta

   Department of Chemistry, Chicago Center for Theoretical Chemistry, Institute for Biophysical Dynamics, and James Franck Institute, The University of Chicago, Chicago, United States

   Contribution

   Designed, performed, and analyzed molecular dynamics simulations

   Competing interests

   No competing interests declared

3. David M Kern

   Department of Molecular and Cell Biology, Helen Wills Neuroscience Institute, and California Institute for Quantitative Biosciences (QB3), University of California, Berkeley, Berkeley, United States

   Contribution

   Conceived of the project and generated initial constructs and performed expression testing

   Competing interests

   No competing interests declared

4. Abhay Kotecha

   Materials and Structural Analysis Division, Thermo Fisher Scientific, Eindhoven, Netherlands

   Contribution

   Collected cryo-EM data

   Competing interests

   is an employee of Thermo Fisher Scientific

5. Gregory A Voth

   Department of Chemistry, Chicago Center for Theoretical Chemistry, Institute for Biophysical Dynamics, and James Franck Institute, The University of Chicago, Chicago, United States

   Contribution

   Designed, performed, and analyzed molecular dynamics simulations, secured funding, and supervised research

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-3267-6748

6. Stephen G Brohawn

   Department of Molecular and Cell Biology, Helen Wills Neuroscience Institute, and California Institute for Quantitative Biosciences (QB3), University of California, Berkeley, Berkeley, United States

   Contribution

   Conceived of the project, built and refined the atomic model, secured funding and supervised research, and wrote the manuscript with input from all authors

   For correspondence

   brohawn@berkeley.edu

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0001-6768-3406

• 4,527

  views

• 644

  downloads

• 42

  citations

Views, downloads and citations are aggregated across all versions of this paper published by eLife.