Near-atomic cryo-EM structure of the helical measles virus nucleocapsid

See allHide authors and affiliations

Science  08 May 2015:
Vol. 348, Issue 6235, pp. 704-707
DOI: 10.1126/science.aaa5137

Measles virus capsid at high resolution

Viruses rely on their capsid proteins to package and protect their genome. For measles virus and other Mononegavirales family members, multiple capsid proteins together form a helical shell around the viral RNA (collectively called the nucleocapsid). Gutsche et al. now report a high-resolution cryoelectron microscopy structure of the measles virus nucleocapsid. The structure reveals how the nucleocapsid assembles and how the nucleo-protein and viral RNA interact, both of which may inform drug design

Science, this issue p. 704


Measles is a highly contagious human disease. We used cryo–electron microscopy and single particle–based helical image analysis to determine the structure of the helical nucleocapsid formed by the folded domain of the measles virus nucleoprotein encapsidating an RNA at a resolution of 4.3 angstroms. The resulting pseudoatomic model of the measles virus nucleocapsid offers important insights into the mechanism of the helical polymerization of nucleocapsids of negative-strand RNA viruses, in particular via the exchange subdomains of the nucleoprotein. The structure reveals the mode of the nucleoprotein-RNA interaction and explains why each nucleoprotein of measles virus binds six nucleotides, whereas the respiratory syncytial virus nucleoprotein binds seven. It provides a rational basis for further analysis of measles virus replication and transcription, and reveals potential targets for drug design.

Nonsegmented negative-strand RNA viruses (nsNSVs, or Mononegavirales) cause epidemics of serious respiratory tract illnesses [e.g., measles virus (MeV) and respiratory syncytial virus (RSV)] and outbreaks of lethal zoonotic diseases [e.g., Ebola virus and Nipah virus (NiV)]. Although preventable by vaccination, measles still remains one of the leading sources of death among young children worldwide (1). MeV belongs to the Paramyxoviridae family of Mononegavirales, which is further divided in two subfamilies: Paramyxovirinae (containing MeV and NiV) and Pneumovirinae (e.g., RSV).

The genome of negative-strand RNA viruses (NSVs) is enwrapped with the viral nucleoprotein N. The resulting ribonucleoprotein complex, called the nucleocapsid, protects the viral genetic information while providing a flexible helical template for viral transcription and replication by the viral RNA polymerase L, which, in the Paramyxoviridae and Rhabdoviridae families of Mononegavirales, is associated with the modular phosphoprotein cofactor P (2). As a unique structure in nucleic acid biology, the NSV nucleocapsid constitutes an attractive potential target for antiviral drugs without harmful side effects. Like all Mononegavirales nucleoproteins, MeV N is composed of two globular domains, the N- and C-terminal domains (NTD and CTD), which together form a stable peanut-shaped nucleoprotein core (Ncore, residues 1 to 391 in MeV N) holding the RNA molecule in the interdomain cleft. In addition, like other Paramyxovirinae, MeV N features a long, intrinsically disordered tail domain (Ntail, residues 392 to 525) (Fig. 1A) that confers to the nucleocapsids a structural plasticity hampering their analysis at better than ~2 nm resolution (3, 4).

Fig. 1 Cryo-EM structure of the MeV Ncore-RNA nucleocapsid at near-atomic resolution.

(A) Schematic of MeV N (navy blue, NTD arm; blue, NTD; salmon, CTD; yellow, CTD arm). The same color code (with RNA in green) is used for the rest of the figure. (B and C) Isosurface representation of the cryo-EM 3D reconstruction of the helical nucleocapsid: (B) front view, (C) cutaway view. (D) Ribbon representation of one Ncore monomer, with the corresponding segmented cryo-EM density shown in transparent gray. (E and F) Close-up of three consecutive protomers from the exterior (E) and from the interior (F) of the helix. Scale bars, 50 Å [(B) and (C)], 30 Å [(D) to (F)].

Removal of the Ntail tightens the MeV nucleocapsid helix, decreasing both its diameter and pitch (35). Our previous cryo–electron microscopy (cryo-EM) map of the resulting Ncore-RNA helix at 12 Å resolution (5) provided precise helical parameters and the domain organization of the MeV Ncore, and a number of low- and medium-resolution EM reconstructions of other Mononegavirales nucleocapsids are available (69). However, the only atomic-resolution structural information about Mononegavirales nucleocapsids comes from x-ray crystallography of N-RNA rings obtained upon nonspecific encapsidation of short cellular RNAs by recombinant nucleoproteins of RSV (7) and of two rhabdoviruses (10, 11), whereas such rings could not yet be observed for Paramyxovirinae. In addition, the crystal structure of the RNA-free NiV Ncore (N0core) bound to a short N-terminal region of P (N0core-P50) (12), has been solved recently. We therefore focused on the helical state of the MeV Ncore-RNA nucleocapsid, so as to (i) obtain a high-resolution three-dimensional (3D) structure of the Ncore monomer, (ii) identify the molecular determinants of helical nucleocapsid polymerization, and (iii) directly visualize the RNA inside the nucleoprotein and understand why each MeV nucleoprotein binds exactly six ribonucleotides, whereas the RSV N binds seven. Here, we report the 3D structure of recombinant MeV Ncore-RNA nucleocapsids at 4.3 Å resolution, determined by single particle–based helical image analysis (13, 14) (fig. S1). This cryo-EM map reveals the detailed domain architecture of the Ncore, its secondary structure elements and many bulky side chains, the rationale of the nucleoprotein packing into a helix, and the mode of nucleoprotein-RNA interaction. Combined with the atomic structures of the RSV N and NiV N0core, this 3D reconstruction allows us to build a reliable pseudoatomic model of the MeV Ncore-RNA helix.

The left-handed MeV Ncore-RNA helix is composed of 12.34 nucleoprotein subunits per turn, with a pitch of 49.54 Å and an outer diameter of 190 Å, in good agreement with previous studies (35) (table S1). The RNA thread winds around the nucleoprotein bobbin, accommodated inside the cleft between the outward-pointing NTD and the inward-oriented CTD, and shielded from above by the N subunits from the successive helical turn (Fig. 1B). As in the nsNSV N-RNA rings, MeV Ncore oligomerization is mainly mediated via the exchange subdomains called the NTD arm (residues 1 to 36) and the CTD arm (residues 373 to 391). Our 3D reconstruction shows how the NTD arm of the Ni protomer inserts into a groove in the CTD of the Ni+1 subunit, whereas the CTD arm lies on top of the NCTD of the Ni–1 subunit, generating a repeated helical structure (Fig. 1, C to F). Unlike in the model of the RSV nucleocapsid (15), the Ni–1 and Ni+1 subunits in the MeV Ncore-RNA helix do not interact directly.

The domain organization of the MeV Ncore (Fig. 1D) corroborates the assumption that all nsNSV nucleoproteins share the same global fold (2, 16) (fig. S2A). In particular, the CTD fold appears conserved among the Paramyxoviridae (MeV, NiV, RSV) (7, 12). The major difference between the pneumovirus RSV and the Paramyxovirinae MeV and NiV is located, as predicted (7, 12), at the distal tip of the NTD (residues 91 to 159 in MeV N). The local resolution of the cryo-EM map in this solvent-exposed region, known as an antigenic site of MeV N (residues 122 to 150) (17), seems to be the lowest, probably indicating mobility with respect to the nucleocapsid core (fig. S1D). In NiV N0core-P50 crystals, the RNA-free nucleoprotein is observed in an open state (12). The present structure of the MeV Ncore-RNA monomer enables accurate modeling of the hinge motion between NTD and CTD, which has been proposed to be associated with the open-closed transition accompanying nucleocapsid formation (fig. S2, B and C) (12).

On the basis of the NiV N0core-P50 structure, the N-terminal domain of Paramyxovirinae P was assigned two simultaneous roles (12): (i) trapping the nucleoprotein in an open RNA-free conformation by rigidifying the CTD, and (ii) preventing its polymerization by interfering with the binding of the exchange subdomains. Indeed, although crystallization of NiV N0core-P50 required deletion of both the NTD and CTD arms, the first and second α helices of the helix-kink-helix peptide of P (Pα1 and Pα2) were shown to overlay with the NTD and CTD arm loops of RSV N, respectively (12). In MeV, the NTD arm begins with an α helix (residues 2 to 14) (Fig. 1D and fig. S2A) that fits into a hydrophobic groove formed by three conserved CTD α helices of the Ni+1 subunit (Fig. 2A) (12). Specifically, four aromatic residues (Phe11, Phe269, Tyr303, and Phe324) that are conserved in Paramyxovirinae (12) seem to stack together, thereby fixing the NTD arm α helix and rigidifying the helix bundle around it (Fig. 2B). Furthermore, the NTD arm α helix of MeV Ni−1 perfectly superimposes with Pα1 from the NiV N0core-P50 structure, although the latter is actually positioned upside down, and the CTD arm loop of the MeV Ni+1 overlaps with Pα2 (Fig. 2C). Therefore, Pα1 may indeed compete with the NTD arm of the Ni–1 protomer, and Pα2 with the CTD arm of the Ni+1 protomer (12). The first α helix of P and the NTD arm α helix of MeV N appear to play similar roles: They lock the CTD and the NTD-CTD junction into a stable conformation, either open or closed depending on the presence of RNA in the interdomain cleft.

Fig. 2 Exchange domains and Ncore oligomerization.

(A) Ribbon representation of two consecutive protomers (colors as in Fig. 1). (B) Close-up of the stack of aromatics fixing the Ni subunit NTD arm on the CTD of the Ni+1 subunit and rigidifying the CTD helix bundle. Pink, Phe11 of the Ni subunit; red, Phe269, Tyr303, and Phe324 of the Ni+1 subunit; transparent gray, cryo-EM density. (C) The P50 peptide (red) of NiV superposes with the NTD arm of the MeV Ni–1 subunit (navy blue) and the CTD arm of the Ni+1 subunit (yellow) of MeV. The gray molecular surface represents the MeV Ni subunit. The N termini of N and P50 and the first and second α helices of P50 are indicated.

The MeV Ncore-RNA structure explains biochemical observations (18, 19) and demonstrates how and why each MeV N binds six ribonucleotides [three stacked bases facing the protein (i.e., “3-bases-in”) and three stacked bases pointing toward the solvent (i.e., “3-bases-out”)], whereas RSV N binds seven ribonucleotides (“3-bases-in” and “4-bases-out”) (Fig. 3, A and B, and figs. S3 and S4). In both viruses, the “3-bases-in” of the Ni subunit are 2-3-4, while base 1 stacks onto the last base of the Ni–1 to form an outward-oriented stack of three (5-6-1) or four (5-6-7-1) bases which crosses the Ni–1-Ni interface, thereby probably contributing to the stability of the nucleocapsid helix. Switch 1 (between bases 1 and 2) constrains the RNA to turn inward, and switch 2 (between bases 4 and 5) orients the RNA outward again (Fig. 3, C to F, and fig. S3). Thus, the RNA topology in MeV and RSV is similar (fig. S4). In MeV nucleocapsid, the outwards-facing groove is however shorter, in particular because long conserved residues R194, K198, Q201, Q202 and Y260 (Fig. 3G) point into it from the 5′ end, sterically hindering packing of a fourth base (Fig. 3, D to F, and figs. S3 and S4). Importantly, Y260, strictly conserved in Paramyxovirinae and featuring a well defined side-chain cryo-EM density, does not flip away from the RNA-binding cleft upon RNA encapsidation as proposed (12), but stacks with base 2, orienting bases 2-3-4 “in” and therefore contributing to the switch 1 (Fig. 3, E and F). The mechanism of the switch 2 (assured by Lys180, Asp186, and Asn351 in MeV N-RNA) also differs between the two viruses (Fig. 3C and fig. S3). Finally, as in RSV, in MeV the N-RNA contacts seem to be rather RNA backbone–specific than base-specific. In particular, three positively charged residues, Lys180, Arg194, and Arg354, show a clear side-chain density and make hydrogen bonds to the RNA backbone (Fig. 3C), and Arg195 (Fig. 3D) and Lys198 (Fig. 3E) may contribute as well. All these basic residues are conserved among Paramyxovirinae (Fig. 3G) and are present in equivalent positions in RSV N (12).

Fig. 3 Nucleoprotein-RNA interaction.

(A) Ribbon representation of three consecutive protomers [navy blue, NTD arm; yellow, CTD arm; green, RNA (bases “in”); orange (bases “out”]. (B) Schematics of the RNA topology for MeV and RSV. For MeV, bases “in” are in green, bases “out” in orange and pink; RSV RNA in gray. Positions and boundaries of successive nucleoprotein subunits on the RNA are indicated; switches in RNA conformation are shown as blue dotted lines. (C to F) Close-up of protein-RNA interaction (scale bar, 10 Å). Protein colors are as in Figs. 1 and 2; RNA colors are the same as in (A) and (B). Numbers of relevant residues and bases are indicated, as are switches in RNA backbone. (G) Sequence alignment of the RNA-binding motif of MeV, NiV, and RSV N. Amino acid abbreviations: A, Ala; D, Asp; E, Glu; G, Gly; I, Ile; K, Lys; L, Leu; M, Met; N, Asn; P, Pro; Q, Gln; R, Arg; S, Ser; T, Thr; V, Val; W, Trp; Y, Tyr.

Both the nucleoprotein protomer and the entire helical nucleocapsid are dynamic entities that must rearrange during the viral replication cycle. Our findings show how the exchange subdomains of N invade the adjacent subunits in the helix, resulting in a stable and yet flexible nucleocapsid assembly. The NTD arm α helix of MeV Ni inserts into the Ni+1 to ensure the major stabilizing interaction, whereas the CTD arm not only tethers to the Ni–1 but also permits the intrinsically unfolded Ntail to escape outside between two helical turns, leading to the increased flexibility of the native MeV N-RNA (4, 20). The Ntail appears to emerge out of the nucleocapsid core in close proximity to the RNA belt surrounding it, so as to dock the P-L complex into its functional environment on the viral genome (2123). Thus, our structure provides a framework for understanding nucleocapsid architecture and remodeling during viral transcription and replication. It may also stimulate the design of new antiviral drugs, because it reveals key regions interfering with nucleoprotein oligomerization and/or genome encapsidation. Finally, because MeV shares many common features with other Paramyxoviridae and with nsNSVs in general, this near-atomic structure of the helical MeV Ncore-RNA nucleocapsid may be valuable for the whole Mononegavirales field.

Supplementary Materials

Materials and Methods

Figs. S1 to S4

Table S1

References (2435)

References and Notes

  1. See supplementary materials on Science Online.
  2. Acknowledgments: We thank A. Jakobi for assistance in real-space coordinate refinement using PHENIX, and M. Jamin for discussions and comments on the manuscript. This work used the platforms of the Grenoble Instruct Center (ISBG: UMS 3518 CNRS-CEA-UJF-EMBL) with support from FRISBI (ANR-10-INSB-05-02) and GRAL (ANR-10-LABX-49-01) within the Grenoble Partnership for Structural Biology (PSB); the electron microscope facility is supported by the Rhône-Alpes Region and by the Fondation pour la Recherche Medicale. A.D. received support from EMBL Interdisciplinary Postdoc (EIPOD) fellowships under Marie Curie Actions (PCOFUND-GA-2008-229597). The data reported in this manuscript are tabulated in the main paper and in the supplementary materials. The cryo-EM map of the helical MeV Ncore-RNA nucleocapsid and the atomic model are deposited in the Electron Microscopy Data Bank and in the Protein Data Bank with accession codes EMDB-2867 and 4uft. The authors declare no competing financial interests. I.G., R.W.H.R., and G.S. designed the study; M.H. purified biological material; I.G., A.D., W.L.L., C.S., and G.S. performed research and analyzed data; R.W.H.R. and G.E. contributed to data interpretation; and I.G. wrote the paper with input from G.E.
View Abstract

Navigate This Article