Structural basis for the activation of plant bunyavirus replication machinery and its dual-targeted inhibition by ribavirin

Plant viruses, recognized since the discovery of tobacco mosaic virus, cause substantial agricultural losses, exceeding US$30 billion annually, and threaten global food security. Current management relies on resistant cultivars and vector control; effective antiviral drugs for infected plants are lacking. Although structural and mechanistic insights have been obtained for RNA-dependent RNA polymerases (RdRps) of several human and animal negative-sense RNA viruses (for example, influenza virus and various bunyaviruses and arenaviruses), no structures of plant virus replication machineries have been reported. The tomato spotted wilt orthotospovirus (TSWV), a segmented negative-sense RNA virus and a major plant pathogen, encodes a large ~330-kDa L protein (RdRp). This study aimed to determine high-resolution structures of full-length TSWV L in multiple functional states to uncover the activation mechanism of plant bunyavirus polymerase and to evaluate the potential of nucleoside analogues, particularly ribavirin, to inhibit this machinery.

Recent years have produced numerous RdRp structures and functional insights for human and animal viruses in the orders Articulavirales and Bunyavirales, including influenza A/C/D polymerases, La Crosse virus (LACV), Rift Valley fever virus (RVFV), severe fever with thrombocytopenia syndrome virus (SFTSV), Hantaan virus (HTNV), Sin Nombre virus (SNV), Lassa virus (LASV) and Machupo virus (MACV), sometimes complexed with regulatory proteins (for example, Z protein in arenaviruses). These have established conserved polymerase architectures and promoter-binding modes (including 5′ hook and distal duplex arrangements). However, no structural data existed for plant virus replication machineries, creating a crucial knowledge gap impeding antiviral pesticide development in crops.

The full-length codon-optimized TSWV L protein (TSWV-YN isolate) was expressed in High Five insect cells via the Bac-to-Bac system and purified using Ni-affinity, ion-exchange (Q column) and size-exclusion chromatography. Cryo-EM samples included: apo L; L bound to 5′ vRNA (1–17 nt) plus 3′ vRNA (1–17 nt); L bound to 5′ vRNA (1–10 nt) (the 5′ hook); L mixed with ribavirin; and L mixed with ribavirin 5′-triphosphate, each at ~1:1.5 molar ratio, vitrified on glow-discharged C-flat grids. Data were collected on Titan Krios microscopes with K2, K3, or Falcon4 detectors. Motion correction (MotionCor2), CTF estimation (Gctf), particle picking, 2D/3D classification, and refinement (RELION) yielded maps at 3.9 Å (apo), 3.7 Å (5′/3′ vRNA-bound), 3.2 Å (5′-hook-bound), 3.4 Å (ribavirin-bound) and 3.0 Å (ribavirin-triphosphate-bound). A minority class (4% particles) revealed density for the C-terminal domain (CTD), refined to 4.0 Å with local refinement and DeepEMhancer. Models were built and refined using Phenix AutoBuild and COOT, aided by AlphaFold for CTD. Functional assays: TSWV infectious clones [L(+)opt, M(−)opt, SR(+)GFP] were agro-infiltrated into Nicotiana benthamiana to assess replication via eGFP fluorescence, immunoblotting (anti-GFP, anti-L, anti-NP), and northern blot for genomic and eGFP RNAs. Site-directed mutagenesis targeted motif F residue Lys1291 (K1291A), 5′-end C5→A in the S segment promoter, Lys1008 (motif I), and additional hook-contact residues. Chemical inhibition was tested by foliar application (100 µg ml−1) of ribavirin, remdesivir, favipiravir, or ribavirin triphosphate, followed by monitoring symptoms and NP accumulation. Electrophoretic mobility shift assays (EMSA) quantified L binding to FAM-labelled 5′ hook RNA with and without ribavirin. In vitro RdRp activity assays used purified TSWV RNPs to synthesize digoxin-labelled RNA in the presence/absence of ribavirin, analyzed by gel electrophoresis and northern blot.

• Five cryo-EM structures of TSWV L (~330 kDa) were determined: apo (3.9 Å), vRNA-promoter-bound (3.7 Å), 5′-hook-bound (3.2 Å), ribavirin-bound (3.4 Å), and ribavirin-triphosphate-bound (3.0 Å). A CTD-containing class was resolved at 4.0 Å (4% particles). The overall architecture resembles sNSV polymerases (PA-like endonuclease, PB1-like core, PB2-like CTD). • Promoter binding induces a global rearrangement: ordering of the endonuclease (endo), arch, clamp, priming loop, and fingertips (motif F), creating a complete catalytic center. TSWV uniquely shows both proximal and distal duplexes in the 5′/3′ promoter termini, with nucleotides 1–10 of the 5′ end forming a hook stabilized by mismatched A1–A10 and base pairs G2–C9 and A3–U8. • The 10-nt 5′ hook alone is sufficient to activate the polymerase: in the 5′-hook-bound structure, motif F and the priming loop are ordered similar to the full promoter-bound state. A critical hydrogen bond forms between the protruding 5′-C5 and motif F residue Lys1291. Mutating L Lys1291→Ala abolishes virus rescue; altering the 5′ C5→A reduces replication. Multiple other hook-contacting L residues are essential for rescue. • Motif F acts as a sensor-adaptor: upon hook binding, motif F engages motifs A–E, G, H to complete the catalytic center. A newly identified absolutely conserved Lys (motif I; Lys1008 in TSWV) interacts with Arg1289 of motif F and is positioned in the active site across sNSV polymerases. Lys1008→Ala abrogates replication despite normal protein accumulation. • Dual-targeted inhibition by ribavirin and ribavirin triphosphate: cryo-EM revealed two binding sites—(1) the enzyme active site and (2) the 5′-hook entry channel that leads to the hook-binding pocket—blocking both catalytic core formation and 5′-hook engagement. Ribavirin reduces L:5′-hook RNA binding in EMSA and markedly suppresses in vitro RNA synthesis by RNPs. In planta, 100 µg ml−1 ribavirin or ribavirin triphosphate prevents symptom development and NP accumulation, indicating strong antiviral activity against TSWV. Favipiravir caused leaf necrosis; remdesivir reduced replication but less potently than ribavirin.

This work resolves, for the first time, the replication machinery of a plant bunyavirus, revealing how the TSWV RdRp transitions from an inactive apo state to an active conformation upon promoter recognition. The 5′-hook structure is both necessary and sufficient to trigger ordering of motif F and the priming loop, coordinating with motifs A–H and a newly defined motif I to assemble the catalytic center. The finding that motif F senses the viral 5′ hook and then adapts to scaffold the active site provides a mechanistic framework connecting promoter recognition to polymerase activation. The identification of conserved motif I (a Lys positioned in the active site across sNSVs) expands the canonical RdRp motif set and clarifies inter-motif interactions stabilizing the active center. Dual-site binding of ribavirin and its triphosphate—blocking the 5′-hook-entry tunnel and the active site—explains potent inhibition: it prevents both vRNA recognition (allosteric activation) and catalysis. These insights address the central research questions: what structural features underlie activation of a plant bunyavirus RdRp, and how can small molecules inhibit this process? Evolutionary parallels with animal sNSVs highlight conserved mechanisms, while unique features (for example, the specific Lys1291–C5 interaction in orthotospoviruses and the larger CTD) suggest opportunities for plant virus–specific inhibitor design.

The study establishes the structural basis of TSWV polymerase activation by defining promoter-induced ordering of motif F and the priming loop, the sufficiency of a 10-nt 5′ hook for activation, and inter-motif contacts including a newly defined, conserved motif I. It also unveils a dual-target inhibition mechanism by ribavirin and ribavirin triphosphate that blocks both 5′-hook entry and active-site catalysis, resulting in strong antiviral effects in planta. These findings open avenues for rational design of plant antiviral pesticides, including orthotospovirus-specific agents targeting the Lys1291–C5 interaction and broader-spectrum inhibitors targeting the hook-entry tunnel. Future work should generalize these mechanisms across diverse plant sNSVs, capture additional functional states (initiation, elongation), assess resistance pathways, and optimize plant-compatible, regulatory-compliant antiviral compounds.

Some domains (NTD and CTD) are intrinsically flexible, limiting resolution; the CTD was well resolved in only a small particle subset (4%). The observed simultaneous presence of a proximal duplex and a 5′ hook may not reflect the exact physiological sequence of promoter rearrangements, representing a structural snapshot. In planta validation was performed primarily in Nicotiana benthamiana; generalizability to crops requires further testing. Ribavirin is not permitted for agricultural use; phytotoxicity was observed with favipiravir, underscoring the need for plant-safe analogues. Functional kinetics and full elongation complex dynamics were not exhaustively characterized.