Structure of a paramyxovirus polymerase complex reveals a unique methyltransferase-CTD conformation

Abstract

Paramyxoviruses are enveloped, nonsegmented, negative-strand RNA viruses that cause a wide spectrum of human and animal diseases. The viral genome, packaged by the nucleoprotein (N), serves as a template for the polymerase complex, composed of the large protein (L) and the homo-tetrameric phosphoprotein (P). The ∼250-kDa L possesses all enzymatic activities necessary for its function but requires P in vivo. Structural information is available for individual P domains from different paramyxoviruses, but how P interacts with L and how that affects the activity of L is largely unknown due to the lack of high-resolution structures of this complex in this viral family. In this study we determined the structure of the L-P complex from parainfluenza virus 5 (PIV5) at 4.3-Å resolution using cryoelectron microscopy, as well as the oligomerization domain (OD) of P at 1.4-Å resolution using X-ray crystallography. P-OD associates with the RNA-dependent RNA polymerase domain of L and protrudes away from it, while the X domain of one chain of P is bound near the L nucleotide entry site. The methyltransferase (MTase) domain and the C-terminal domain (CTD) of L adopt a unique conformation, positioning the MTase active site immediately above the poly-ribonucleotidyltransferase domain and near the likely exit site for the product RNA 5' end. Our study reveals a potential mechanism that mononegavirus polymerases may employ to switch between transcription and genome replication. This knowledge will assist in the design and development of antivirals against paramyxoviruses.

Keywords: Cryo-EM; paramyxovirus; polymerase; replication; transcription.

Fig. 1.

Architecture of the PIV5 L–P complex. (A) Domain diagrams of PIV5 L and P proteins. RdRp, cyan; PRNTase, green; CD, yellow; MTase, orange; CTD, red; P-OD, purple; P-XD, purple. (B and C) Electron density (B) and atomic model (C) of the PIV5 L–P complex with domains colored as depicted in A.

Fig. 2.

Comparison of the PIV5, RSV, and VSV priming loop and intrusion loop. (A) The PIV5 priming loop adopts the same elongation conformation as in the RSV structure. The intrusion loop projects out into the central cavity between the RdRp and PRNTase domains. (B) In the VSV structure the priming loop in the initiation conformation would sterically clash with the position of the PIV5 intrusion loop. The VSV and RSV intrusion loops are in the same position with minor differences.

Fig. 3.

Interaction interfaces of PIV5 L and P proteins. (A) Crystal structure of the OD of the PIV5 P protein. The OD forms an all-parallel four-helix bundle with one helix from each of four chains of P. (B) Interfaces between P-OD (purple) and L (cyan). The fragment of L that is necessary and sufficient to interact with P is shown as an opaque surface, the rest of L is shown as transparent cartoon. Helix 392–412 is the only portion of L that interacts with P-OD. (C) Interaction between the unmodelled P density and L. This density is not as well resolved as the OD or XD and does not form extensive contacts with L except at the base of P-OD. (D and E) Interaction of P-XD (residues 346–392) and L. Helices α1 and α3 of P-XD form the interface of the XD with L. The portion of L that interacts with P-XD spans residues 303–350. Superposition of MeV P-XD bound to a C-terminal fragment of N (486–504) with PIV5 P-XD (E).

Fig. 4.

Comparison of PIV5, VSV, and RSV L–P complexes. (A–C) PIV5 (A), RSV (B), and VSV (C) L–P complexes were aligned based on the RdRp domain (cyan). The PRNTase domains (green) are in similar positions relative to the RdRp in all three structures. The P-OD of PIV5 is significantly longer than the P-OD of RSV and protrudes further away from the RdRp. The PIV5 P-XD is in roughly the same position as the single C-terminal helix in the RSV structure. There are no visible P linker helices in the PIV5 structure as there are in the RSV structure. The large rearrangement of the MTase (orange) and (red) is visible between the PIV5 and VSV structures. The relative orientations of the domains are shown in the Inset with the arrowhead representing the direction of the protein backbone. CD, connecting domain (yellow); P protein (purple).

Fig. 5.

Comparison of the positions of the conserved HR motif and the MTase active site between the PIV5 and VSV structures. The MTase and CTD are positioned directly above the PRNTase domain in the PIV5 structure (opaque). The distance between the HR motif and the GTP modeled into the MTase active site is 25.7 Å. In the VSV structure the distance increases to 52.5 Å due to the active site projecting away from the RdRp domain.

Fig. 6.

Model of transcription and genome replication for L–P. The viral RNA (vRNA) genome is separated from N to allow it to enter the template entry tunnel of L. (A) In transcription initiation, the MTase-CTD module is positioned directly above the PRNTase domain, as in the PIV5 structure. This positions the active site of the MTase as that the covalently linked RNA-PRNTase is pushed up into it leading to productive capping and methylation. This requires an outward movement of the PRNTase domain to accommodate the growing RNA strand. We hypothesize that a P-XD captures the monomer of N that is no longer bound to the genomic RNA, keeping it in close proximity to be used to recapture the genomic RNA reemerging from the template exit channel. Additional P-XDs are shown in the first panel but removed for clarity from subsequent panels. (B) In genome replication, the MTase-CTD module is positioned further away from the PRNTase domain. No covalent linkage is formed between the RNA and conserved histidine, but an outward movement of the PRNTase domain is still required to accommodate the growing RNA strand. The newly released monomer of N is captured in the same way as in A and a second copy of N is used to coat the newly synthesized antigenome by an unknown mechanism.