Structure of SARS-CoV-2 membrane protein essential for virus assembly

Abstract

The coronavirus membrane protein (M) is the most abundant viral structural protein and plays a central role in virus assembly and morphogenesis. However, the process of M protein-driven virus assembly are largely unknown. Here, we report the cryo-electron microscopy structure of the SARS-CoV-2 M protein in two different conformations. M protein forms a mushroom-shaped dimer, composed of two transmembrane domain-swapped three-helix bundles and two intravirion domains. M protein further assembles into higher-order oligomers. A highly conserved hinge region is key for conformational changes. The M protein dimer is unexpectedly similar to SARS-CoV-2 ORF3a, a viral ion channel. Moreover, the interaction analyses of M protein with nucleocapsid protein (N) and RNA suggest that the M protein mediates the concerted recruitment of these components through the positively charged intravirion domain. Our data shed light on the M protein-driven virus assembly mechanism and provide a structural basis for therapeutic intervention targeting M protein.

Fig. 1. M protein forms dimer and higher-order oligomers.

a Schematic view of SARS-CoV-2 virion. Four structural proteins, M, N, S, and E and genomic RNA are shown in different colors. The N protein and RNA molecules form viral ribonucleoprotein (vRNP) complexes. b Representative 2D class average images of M protein dimer solubilized by LMNG-CHS (see also Supplementary Fig. 2) c Representative 2D class average images of higher-order oligomers of M protein solubilized by GDN (see also Supplementary Fig. 3). The possible oligomerization state is indicated by the colored dots.

Fig. 2. Cryo-EM structures of M protein dimer.

a Cryo-EM maps (upper) and ribbon models (lower) of M/Fab-E and M/Fab-B complexes. Unsharpened maps are shown in transparent light gray color (M/Fab-E complex: level = 0.08; M/Fab-B complex: level = 0.35), and B-factor-sharpened maps are show in multiple colors (M/Fab-E complex: level = 0.28; M/Fab-B complex: level = 0.75). M protein protomers, Fab-E, and Fab-B, are colored in purple, blue, green, and pink, respectively. b Structures of M protein dimer in the long form (upper) and the short form (lower) are shown from the side (left), front (middle), and bottom (right) views. One protomer of M protein is colored in a rainbow, and the other is light gray. Schematic views of the arrangement of six TM helices are shown on the lower right. TM1*, TM2, and TM3 formed one bundle. Throughout this paper, asterisks are used to indicate structural parts from the second M protein protomer. c Schematic view of the M protein protomer. Secondary structure elements and hinge regions are indicated. The disordered N-term and C-term regions are indicated by dashed lines. d Structure comparison between the long form (blue and cyan) and short form (orange and beige). The structures were aligned either according to the whole structure (left) or bundle A (right). e Detailed views of the interactions between the hinge region and bundle A in the long (upper) and short (lower) forms. The key residues are indicated in red. Dashed lines represent hydrogen bonds or salt bridges.

Fig. 3. The M protein resembles the ion channel ORF3a.

a Structural comparisons of the long form (cyan) and the short form (beige) of SARS-CoV-2 M protein dimer with SARS-CoV-2 ORF3a (gray) (PDB: 6XDC). b The inner surface of the M protein dimers (vestibule, upper cavity, and lower tunnel) calculated using Pymol are shown in yellow. The bottom of the upper cavities and the top of the lower tunnel are indicated as a, b, respectively. The W31 side chain is shown using a stick representation. c Close-up views of the entrances to the vestibules above the upper cavity. Half-transparent electrostatic surface potentials and side chains of charged residues are shown. The purple dashed lines indicate the approximate positions of the lipid bilayer. d Electrostatic surface potentials of the long and short forms and cross-sections along the C2 axis (left). Detailed views of the interface between TM1 and TM1* are shown, including the distances between each residue pair (right). e Snapshots of the MD simulation results of the long form (left) and the short form (right).

Fig. 4. M protein, N protein, and RNA interactions.

a Differential interference contrast (DIC) and fluorescence images of liquid-liquid phase separation (LLPS) of N protein (TAMRA-labeled) with or without M protein (FITC-labeled). The experiment was repeated three times with similar results. b FLAG tag pull-down assay using recombinant M protein (FLAG-tagged) and N protein (no tag) in the absence or presence of poly(I:C). The experiment was repeated twice with similar results. Source data are provided as a Source Data file. c StrepII tag pull-down assay using recombinant N protein (StrepII-tagged) and M protein (FLAG-tagged) in the absence or presence of RNAs of different sizes. RNA-H and RNA-L indicate yeast RNA with molecular weights >30 kDa and 3–30 kDa, respectively. The experiment was repeated twice with similar results. d Electrostatic surface potentials of the intravirion side of the M protein dimer (long form). Positively charged residues to which the mutations were introduced in e are shown using stick representations and labeled. e Co-immunoprecipitation assay of wild-type (WT) or mutant M proteins with N protein in HEK293T cells. The experiment was repeated three times with similar results. Source data are provided as a Source Data file. f Model of M protein-triggered SARS-CoV-2 assembly M protein in the endoplasmic reticulum–Golgi intermediate compartment (ERGIC) forms dimers in two different conformations that assemble into higher-order oligomers to induce membrane curvature. M protein recruits N and genomic RNA in a cooperative manner. S and E proteins are also recruited to the budding site via an unknown mechanism. Source data are provided as a Source Data file.