Mutation of a conserved lysine in the RdRp fingers domain broadly attenuates orthobunyaviruses Authors: Guodian Xiong, Fei Wang, Jingjing Tang, Tao Cheng, Nanjie Ren, Jingke Geng, Wei Chen, Doudou Huang, Zengqin Deng, Xiang Sun, Jinglin Wang, Zhiming Yuan, Han Xia Published: 2026 ABSTRACT AbstractOrthobunyaviruses, comprising more than 130 species, constitute a significant threat to human and animal health, yet no effective broad-spectrum treatments or vaccines are currently available. Here, we initially identified an evolutionarily constrained lysine residue, highly conserved across all 22 Orthobunyavirus serogroups, within the α30 helix of the RNA-dependent RNA polymerase (RdRp). Functional interrogation of this site by charge-reversal substitution (K-to-E) substantially impaired polymerase activity and led to marked attenuation of Ebinur Lake virus (EBIV), as evidenced by impaired replication, attenuated cytopathogenicity, and reduced virulence in mice. The similar attenuation observed in Bunyamwera virus (BUNV), the prototype virus of the genus, and Oya virus (OYAV), a member of the Simbu serogroup, demonstrates that this residue fulfills a conserved functional role. Biochemical and cellular analyses further revealed a graded, charge-dependent mutational effect, indicating that attenuation arises predominantly from disruption of the essential interaction between RdRp and the 3′ terminus of viral genomic RNA (3′ vRNA). Structural modeling and mutagenesis further suggest that the K-to-E substitution creates a new hydrogen bond that may also contribute to viral attenuation. Together, these findings establish the conserved lysine as a critical regulator of RdRp function and offer a rational strategy for broad-spectrum attenuation of orthobunyaviruses. Subject terms: Viral pathogenesis, Viral transmission Introduction Bunyaviricetes is a large class of segmented negative-sense RNA viruses1,2. This class represents one of the most widespread groups of RNA viruses in nature and has undergone taxonomic expansion from its original status as a family (Bunyaviridae) to a standalone class comprising two orders and sixteen families (ICTV 2024 Release)3. The genus Orthobunyavirus, within the family Peribunyaviridae, comprises a diverse group of arboviruses mainly transmitted by mosquitoes or midges. Members of this genus are associated with disease in humans (members of the Bunyamwera, California, and Simbu serogroups)4,5 or livestock (notably within the Simbu serogroup)4. Recent studies have reported expanding geographic distributions and increasing numbers of human infections caused by emerging orthobunyaviruses, highlighting their growing public health significance6–11. For example, Oya virus (OYAV), first identified in Malaysian pigs in 200012, has subsequently been detected in humans and arthropod vectors in Karnataka, India13; Hanoi, Vietnam14; and Yunnan and Guangdong provinces in China11,15,16. Ebinur Lake virus (EBIV) was first isolated in 2013 from Culex modestus mosquitoes collected near Ebinur Lake, Xinjiang Uygur Autonomous Region, China17, and was later detected serologically in human sera from individuals living in the surrounding region18. EBIV has also been reported in mosquitoes, ticks, and little gulls in Russia, as well as in marsh rats and mongooses in Kenya19. The orthobunyavirus RdRp serves as the central enzymatic machinery for viral genome replication and mRNA transcription20,21. It harbors multiple functional domains: the catalytic RdRp domain that synthesizes RNA22; the cap-snatching endonuclease (EN) and the cap-binding domain (CBD), which cleave and bind host cell mRNA caps, respectively, to prime viral transcription20–24; and RNA-binding domains that specifically recognize conserved 5′ and 3′ termini of viral genomic RNA (vRNA)25. These two vRNAs coordinately regulate polymerase activation. The 3′ vRNA acts as the direct template for replication and transcription, and its specific interaction with RdRp ensures accurate template positioning while promoting conformational maturation of the active site. In contrast, the 5′ vRNA functions as a molecular switch that triggers global structural rearrangements, thereby simultaneously activating both the RdRp activity and the endonuclease activity25,26. Structural studies of the La Crosse virus (LACV) RdRp complex have provided important insights into this activation mechanism25,26. The α30 helix, a core element of the RdRp fingers domain, is a critical hub responsible for 5′ vRNA binding, 3′ vRNA template positioning, and allosteric activation of the polymerase25,26. Histidine residues H760 and H761 located on the α30 helix form direct hydrogen bonds with the 5′ vRNA. This interaction repositions the helix and creates the space required for subsequent allosteric rearrangements25. Meanwhile, the 3′ vRNA template is specifically recognized and anchored by the fingers domain together with motif F, allowing precise entry into the catalytic site for efficient initiation26. In addition, residue H757 stabilizes a critical β-turn within the fingertips region of RdRp, thereby promoting interactions between motif B (residue Q1145) and residues R958/E959, which cooperatively facilitate polymerase activation25. Through comparative sequence analysis of the α30 helix within this functionally critical region, we identified a universally conserved lysine (K) residue adjacent to H757, H760, and H761 across all 22 orthobunyavirus serogroups27, suggesting its critical role in RdRp function. To investigate its biological significance, we selected EBIV, a recently characterized virus for which robust reverse genetics systems and validated animal models are available, as a representative orthobunyavirus to assess the impact of this lysine residue on polymerase activity, viral replication, and virulence28. Functional conservation of this lysine was further validated in vitro using two additional orthobunyaviruses: Bunyamwera virus (BUNV), the prototype of the genus Orthobunyavirus, and OYAV from the Simbu serogroup12,29. Structural modeling, minigenome assays, and cell viability analyses revealed the mechanistic basis of its function. Collectively, our findings identify this conserved lysine in the RdRp fingers domain as a key regulatory residue and a promising target for broad-spectrum attenuation strategies against emerging or re-emerging orthobunyaviruses. Results The strictly conserved lysine residue within the α30 helix is associated with orthobunyavirus RdRp function Previous structural studies have demonstrated that vRNA binding induces conformational remodeling of the orthobunyavirus RdRp, promoting formation of the active site required for polymerase activity25,30. Structural studies of LACV further revealed that the α30 helix participates in this rearrangement and contributes to active-site organization25. Within the α30 helix, residues H757, H760, and H761 form a histidine-enriched motif positioned adjacent to the catalytic cavity (Fig. 1A)25. The coordinated actions of these three histidines are indispensable for the proper assembly of the active site. Conservation pattern and functional roles of critical residues within the orthobunyavirus RdRp α30 helix region. A Structure of LACV RdRp (PDB: 5AMQ) highlighting H757, H760, H761, and K759 within the α30 helix region. B Structure of LACV RdRp (PDB: 7ORL) reveals that K759 participates in direct binding to the 3′ vRNA. C Alignment of the α30 helix and upstream region across representative orthobunyaviruses. Residue numbering is based on the LACV RdRp. Representative orthobunyaviruses included in the alignment: BUNV Bunyamwera virus, EBIV Ebinur Lake virus, BWAV Bwamba virus, LACV La Crosse virus, CAPV Capim virus, ENSV Enseada virus, ALJV Alajuela virus, APEUV Apeu virus, GMAV Guamá virus, GROV Guaroa virus, KOOV Koongol virus, LEAV Leanyer virus, BUCV Buffalo Creek virus, NDV Nyando virus, BOTV Botambi virus, ABRV Abras virus, TETEV Tete virus, TURV Turlock virus, WYOV Wyeomyia virus. Notably, in the transcription-initiation conformation of LACV RdRp (PDB: 7ORL), residue K759, positioned within this histidine cluster, directly interacts with the 3′ vRNA (Fig. 1B)26, highlighting a potential role in regulating polymerase activity. To determine whether this region is evolutionarily constrained, a comprehensive sequence alignment was performed using representatives of 22 orthobunyavirus serogroups, encompassing 24 distinct viral sequences (Supplementary Table 1). H757, H760, and H761 were completely conserved across all examined viruses. Strikingly, the lysine residue that intervenes between these three histidines was also strictly conserved across the genus, suggesting a potential essential biological function. This lysine corresponds to K759 in LACV, K760 in EBIV and BUNV, and K756 in OYAV (Fig. 1C and Supplementary Data 1). We next assessed the predicted structural consequences of glycine substitution at these positions (K759, H757, H760, and H761) in LACV RdRp. K759G mutation was predicted to reduce protein stability (ΔΔGStability = −0.35 kcal/mol), with an effect comparable to H757G (ΔΔGStability = −0.49 kcal/mol) and greater than those of H760G (ΔΔGStability = 0.12 kcal/mol) or H761G (ΔΔGStability = −0.03 kcal/mol) (Supplementary Table 2). Collectively, these findings support a functionally important role for this conserved lysine residue and prompt further mechanistic interrogation. K760E substitution in EBIV RdRp impairs polymerase activity, viral replication, and virulence Given the well-established reverse genetics systems and lethal infection models for EBIV28,31, the function of the conserved lysine residue K760 in EBIV was investigated both in vitro and in vivo. Under physiological conditions, lysine possesses a positively charged side chain. To test whether electrostatic properties at this position are functionally important, we introduced a K760E substitution into EBIV RdRp and first evaluated polymerase activity using a minigenome system (Fig. 2A and Supplementary Data 2). Relative to wild-type (WT), K760E significantly reduced NanoLuc luciferase (NLuc) activity to 43.3% (p < 0.0001), indicating markedly impaired EBIV RdRp function. Impact of the EBIV RdRp K760E substitution on polymerase activity, viral replication, and virulence. A Minigenome assay measuring RdRp activity of EBIV WT and mutants using a NLuc reporter (one-way ANOVA; n = 3). B Cell viability of BHK-21 and Vero E6 cells at 2 dpi following infection with rEBIV/WT or rEBIV/K760E. All viability data were normalized to the mean absorbance of mock group (one-way ANOVA was used for BHK-21 cells, and Welch’s ANOVA was applied for Vero E6 cells; n = 6). C Plaque morphology and size of rEBIV/WT and rEBIV/K760E at 5 dpi (plaque areas are presented as mean ± SD; two-tailed unpaired t-test with Welch’s correction). Growth kinetics of rEBIV/WT and rEBIV/K760E in BHK-21 (D), Vero E6 (E) and LMH F cells (MOI = 0.01) (unpaired t-test; n = 3). G Viral RNA loads in sera of BALB/c mice infected intraperitoneally with rEBIV/WT or rEBIV/K760E (5 or 10 PFU per mouse) (unpaired t-test with Welch’s correction (two-tailed) was used for the comparisons of rEBIV/WT (5 PFU) vs. rEBIV/K760E (5 PFU) and rEBIV/WT (10 PFU) vs. rEBIV/K760E (10 PFU); two-tailed Mann–Whitney U-test was applied for the comparison between rEBIV/WT (5 PFU) and rEBIV/K760E (5 PFU); n = 5). The dashed line indicates the detection threshold (Ct = 36), below which samples were considered negative. H Body weight changes of mice infected intraperitoneally with rEBIV/WT or rEBIV/K760E (5 or 10 PFU per mouse) (n = 5). Red asterisks indicate death or euthanasia due to critical weight loss (>20% of initial body weight). I Survival curves of mice infected intraperitoneally with rEBIV/WT or rEBIV/K760E (5 or 10 PFU per mouse) (n = 5). dpi: days post infection; hpi: hours post infection; MOI: multiplicity of infection; Error bars represent mean ± SD. n = 3, 5, or 6 biologically independent experiments as indicated. Statistical significance was assessed using the tests indicated in each panel: ****p ≤ 0.0001, *** p ≤ 0.001, **p ≤ 0.01, *p ≤ 0.05. We then rescued mutant virus (rEBIV/K760E) and compared its phenotypes with rEBIV/WT. At 2 days post infection (dpi), BHK-21 cells infected with rEBIV/WT exhibited a viability drop to 31.07%, while those infected with rEBIV/K760E retained significantly higher viability (66.63%, p < 0.0001) (Fig. 2B and Supplementary Data 2). In Vero E6 cells, viability dropped to 17.71% for WT but remained at 83.29% for K760E (p < 0.0001) (Fig. 2B and Supplementary Data 2). Consistently, plaque assays showed rEBIV/K760E produced significantly smaller plaques (plaque area: 2.0 ± 0.54 mm2) than rEBIV/WT (plaque area: 4.6 ± 1.94 mm2, p = 0.0003) (Fig. 2C and Supplementary Data 2). To further investigate the impact of the K760E substitution, the viral replication of rEBIV/WT and rEBIV/K760E in BHK-21, Vero E6, LMH, and SW13 cells was compared. In BHK-21 cells, the viral load of rEBIV/WT reached 4.57 × 107 copies/mL at 24 h post infection (hpi), compared to a 25-fold lower load for rEBIV/K760E (1.81 × 106 copies/mL, p < 0.0001) (Fig. 2D and Supplementary Data 2). Similarly, in Vero E6, the viral load of rEBIV/WT (8.90 × 107 copies/mL) was 38.8-fold higher than that of rEBIV/K760E (2.29 × 106 copies/mL) at 24 hpi (p < 0.0001; Fig. 2E and Supplementary Data 2). In both LMH and SW13 cells, similar reductions were observed, showing a 6.3-fold reduction (rEBIV/WT: 2.15 × 107 copies/mL vs. rEBIV/K760E: 3.40 × 106 copies/mL at 24 hpi; p < 0.0001) and a 12.7-fold reduction (rEBIV/WT: 5.80 × 105 copies/mL vs. rEBIV/K760E: 4.56 × 104 copies/mL at 48 hpi; p < 0.0001), respectively (Fig. 2F, Supplementary Fig. 1 and Supplementary Data 2). We next assessed the impact of the K760E substitution on EBIV virulence in highly susceptible BALB/c mice by monitoring viremia, body weight, survival, and neutralizing antibody1. Viremia: At an inoculum of 10 PFU, mice infected with rEBIV/K760E displayed significantly lower viremia than those infected with rEBIV/WT at both 1 dpi (700.11 vs. 1.65 × 104 copies/mL; 23.6-fold reduction, p = 0.005) and 3 dpi (1.40 × 104 vs. 1.99 × 105 copies/mL; 14.3-fold reduction, p = 0.002) (Fig. 2G and Supplementary Table 3). At 5 PFU, attenuation was even more pronounced, with a reduction of 432-fold at 1 dpi (99.41 vs. 4.30 × 104 copies/mL, p = 0.0066) and 299-fold at 3 dpi (186.34 vs. 5.56 × 104 copies/mL, p = 0.0079) (Fig. 2G and Supplementary Table 3)2. Body weight: At 10 PFU, mice infected with rEBIV/K760E exhibited significantly higher body weights from 2 to 4 dpi, with the maximal difference observed at 4 dpi (14.4% higher, p = 0.0005) (Fig. 2H and Supplementary Table 4). At 5 PFU, mice infected with rEBIV/K760E maintained stable body weights. In contrast, rEBIV/WT infection induced progressive weight loss, culminating in a mean body weight reduction of 19.74% at 4 dpi (Fig. 2H and Supplementary Table 4)3. Survival: At 10 PFU, rEBIV/K760E infection resulted in 20% survival, whereas all rEBIV/WT-infected mice succumbed (Fig. 2I). At 5 PFU, all rEBIV/K760E-infected mice survived until scheduled euthanasia at 14 dpi, whereas all rEBIV/WT-infected mice succumbed to infection or required euthanasia (body weight loss ≥ 20%) (Fig. 2I)4. Neutralizing antibody: rEBIV/K760E infection (5 PFU) elicited high neutralizing antibody titers in surviving mice at 14 dpi. Two mice exhibited a 50% neutralization titer (PRNT50) of 50, while one mouse showed a PRNT50 of 1305. The remaining two mice maintained >59% neutralization rates (62.22% and 59.11%, respectively) even at the maximum testable dilution factor of 4050-fold (Supplementary Table 5). Collectively, these findings demonstrate that charge reversal at residue K760 profoundly attenuates EBIV, highlighting a conserved regulatory role for this residue across orthobunyaviruses. K-to-E substitution confers attenuation in multiple orthobunyaviruses To evaluate whether the attenuation conferred by K-to-E substitution is conserved in multiple orthobunyaviruses, we further tested BUNV (Bunyamwera serogroup) and OYAV (Simbu serogroup). For enzymatic validation, introduction of the corresponding substitution into the BUNV RdRp (K760E) nearly abolished polymerase activity, reducing NLuc activity output to 0.03% of WT (p < 0.0001; Fig. 3A and Supplementary Data 2). In OYAV, the corresponding mutation (K756E) reduced RdRp activity to 53.19% of WT (p = 0.0082) (Fig. 3A and Supplementary Data 2). K-to-E substitution attenuates OYAV and BUNV by impairing RdRp activity and viral replication. A Minigenome assay comparing RdRp activity of OYAV (K756E) and BUNV (K760E) mutants (unpaired t-test; n = 3). B Cell viability of BHK-21 cells infected with rOYAV/WT or rOYAV/K756E at 2 dpi (MOI = 0.01). All viability data were normalized to the mean absorbance of mock group (one-way ANOVA; n = 12). C Plaque morphology and size of rOYAV/WT and rOYAV/K756E on BHK-21 cells at 6 dpi (plaque areas are presented as mean ± SD; two-tailed Mann–Whitney U-test). Growth kinetics of rOYAV/WT and rOYAV/K756E in Vero E6 (D), SW13 (E), LMH (F) and BHK-21 (G) cells during 0–48 hpi (unpaired t-test; Statistical analysis of growth curves was performed using log-transformed viral titers; n = 3). Error bars represent mean ± SD. n = 3 or 12 biologically independent experiments as indicated. Statistical significance was assessed by one-way ANOVA, unpaired t-test or Mann–Whitney U test as appropriate: ****p ≤ 0.0001, ***p ≤ 0.001, **p ≤ 0.01, *p ≤ 0.05. Then the mutant virus rOYAV/K756E was rescued to examine the effects of the K756E substitution on viral cytopathogenicity and replication. At 2 dpi, cells infected with rOYAV/WT exhibited only 81.79% cell viability relative to mock-infected controls (set as 100%), whereas rOYAV/K756E infection maintained 90.58% cell viability (p = 0.0027), indicating that the WT virus was 1.93‑fold more cytopathogenic than the K756E mutant (Fig. 3B and Supplementary Data 2). Furthermore, the plaques formed by rOYAV/K756E (plaque area: 0.42 ± 0.14 mm²) were significantly smaller than those of rOYAV/WT (plaque area: 1.44 ± 0.51 mm², p < 0.0001) in BHK-21 cells at 6 dpi (Fig. 3C and Supplementary Data 2). Growth curve results further confirmed the attenuation of rOYAV/K756E. In Vero E6 cells, rOYAV/WT reached a viral load of 1.03 × 107 copies/mL at 24 hpi, which was 62.2-fold that of rOYAV/K756E (1.66 × 105 copies/mL; p < 0.0001) (Fig. 3D and Supplementary Data 2). In SW13 cells, rOYAV/WT reached a viral load of 5.00 × 108 copies/mL at 48 hpi, which was about 88.2-fold that of rOYAV/K756E (5.67 × 106 copies/mL; p < 0.0001) (Fig. 3E and Supplementary Data 2). In LMH and BHK-21 cells, viral loads showed modest reductions of 9.5-fold (rOYAV/WT: 5.02 × 106 copies/mL; rOYAV/K756E: 5.30 × 105 copies/mL at 48 hpi; p < 0.0001) and 9.6-fold (rOYAV/WT: 1.68 × 107 copies/mL; rOYAV/K756E: 1.75 × 106 copies/mL at 24 hpi; p < 0.0001), respectively (Fig. 3F-G and Supplementary Data 2). These findings demonstrate that K-to-E substitution reproducibly attenuates polymerase activity and viral fitness across multiple orthobunyaviruses examined here. K-to-E substitution impairs polymerase activity by disrupting 3′vRNA binding Previous studies on the LACV RdRp have clearly demonstrated that residue K759 directly participates in binding to the 3′ vRNA (Supplementary Fig. 2)26. Given its high conservation, we hypothesized that the corresponding residue in EBIV, K760, plays a similar role in template binding. To test this hypothesis, we designed the following mutations: K760E to test the effect of charge reversal, K760A to evaluate the impact of side-chain truncation, and K760R to determine whether a conserved positive charge is sufficient to maintain enzymatic activity. We purified the EBIV RdRp proteins of WT and mutants and assessed their RNA synthesis activity in vitro using a primer-dependent RNA replication assay with a 23-nt RNA template and a 4-nt RNA primer (Supplementary Fig. 3). In the presence of ATP and UTP, WT RdRp efficiently extended the primer to generate a 13-nt product. Notably, the activities of the mutants strictly correlated with side-chain charge polarity: K760R (preserved positive charge) produced a similar amount of the 13-nt product as the WT (103% of WT). In contrast, the neutral mutant K760A showed reduced activity (34% of WT), and the negatively charged mutant K760E was even less active (21% of WT) (Fig. 4A and Supplementary Fig. 4). A catalytically inactive double mutant, D1164 A/D1165A, served as a negative control (Fig. 4A, Supplementary Fig. 4 and Supplementary Data 2). Charge-dependent effects of substitutions at the conserved K residue on EBIV polymerase activity in vitro. A Primer-extension assay of RdRp RNA synthesis activity of EBIV and mutants. Product intensities were quantified and normalized to WT (set to 1.0). D1164A/D1165A double mutant served as a negative control. B Minigenome assay measuring RdRp activity of EBIV WT and mutants using a NLuc reporter (one-way ANOVA; n = 3). Error bars represent mean ± SD. n = 3 biologically independent experiments. Statistical significance is assessed by one-way ANOVA: ****p ≤ 0.0001, ***p ≤ 0.001, **p ≤ 0.01, *p ≤ 0.05. To further validate these effects, WT and mutant RdRp (K760R, K760A, and K760E) were tested in a minigenome reporter system (Fig. 4B). K760R significantly enhanced NLuc activity (195.5% of WT, p < 0.0001). In contrast, K760E strongly suppressed NLuc activity, which was only 24.0% of WT level (p < 0.0001). As for the K760A variant, its NLuc activity showed a mild increase to 112.0% of the WT, but this increase was without statistically significant difference (Fig. 4B and Supplementary Data 2). Both the in vitro RNA synthesis and the minigenome assay consistently revealed the same activity trend: R > A > E. This supports the critical role of the charge at this site in RdRp function and confirms that the K-to-E substitution impairs polymerase activity by disrupting its interaction with the 3′ vRNA. New hydrogen bond formation contributes to attenuation induced by the K-to-E substitution In addition to disruption of 3′ vRNA binding, other factors may also contribute to attenuation. We therefore predicted the apo-state structures of the EBIV, OYAV, and BUNV RdRps using AlphaFold3 (confidence metrics in Supplementary Fig. 5). These models consistently suggested that the mutation could generate a previously absent hydrogen-bonding interaction in all three viruses (Supplementary Fig. 6A–C). Sequence analysis further showed that the partnering residue is also highly conserved, appearing as lysine or arginine in 17 of 22 serogroups (Supplementary Fig. 6D). Previous structural analyses of LACV RdRp indicated that the active center formation requires movement of the α30 helix25. In our structural models, the predicted new hydrogen bond was located on the opposite side of α30 helix movement (Fig. 5A), potentially hindering this movement and reducing enzymatic activity. Impact of a new hydrogen bond on OYAV polymerase activity and cytopathogenicity. A A new hydrogen bond between K551 and the mutant K756E is positioned on the opposite side of the α30 helix movement, and this hydrogen bond is absent between K551 and wild-type K756. B Potential hydrogen-bonding interactions in specific amino acid combinations at residues 551 and 756 in OYAV RdRp. C Polymerase activity of OYAV RdRp mutants determined by minigenome assay (one-way ANOVA; n = 3). D Cell viability of BHK-21 cells infected with the corresponding recombinant OYAV mutant viruses (n = 12). All viability data were normalized to the mean absorbance of mock group (Welch’s one-way ANOVA). Data are presented as mean ± SD. n = 3 or 12 biologically independent experiments as indicated; Statistical significance was assessed by one-way ANOVA or Welch’s one-way ANOVA as appropriate: ****p ≤ 0.0001, ***p ≤ 0.001, **p ≤ 0.01, *p ≤ 0.05. To test this possibility, we constructed six groups of OYAV RdRp (Fig. 5B): (1) K551/K756 (WT); (2) K551/K756E (K-to-E substitution, predicted to enable hydrogen bonding); (3) K551N/K756E (bond-disrupting control, retaining K756E); (4) K551N/K756 (assessing the effect of K551N alone); (5) K551R/K756 (corresponding to the EBIV/BUNV WT context); and (6) K551R/K756E (corresponding to the EBIV/BUNV K-to‑E substitution). Minigenome assays showed that the K551/K756E substitution reduced RdRp activity to 58.31% of WT (p < 0.0001), whereas NLuc activity was restored to WT levels in the K551N/K756E group (Fig. 5C and Supplementary Data 2). Cell infection with live viruses showed rOYAV/K756E exhibited higher cell viability (89.65%, p < 0.0001) than WT (58.83%) (Fig. 5D and Supplementary Data 2). In contrast, rOYAV/K551N-K756E restored cytopathogenicity close to WT levels (56.95% viability). Notably, the K551N/K756 exerted only minimal effects on RdRp activity (93.02% of WT, p = 0.0011) and did not affect cytopathogenicity (60.74% viability) (Fig. 5C, D and Supplementary Data 2). To validate the universality of this mechanism, we recapitulated the natural EBIV/BUNV amino acid context in OYAV. Relative to K551R/K756, the K551R/K756E exhibited significantly reduced polymerase activity (83.42% vs. 41.95% of OYAV WT RdRp, p < 0.0001) and attenuated cytopathogenicity (50.80% vs. 70.46% of mock, p < 0.0001) (Fig. 5C, D and Supplementary Data 2). Collectively, these findings support that this new hydrogen bond formation between residues 551 and 756 represents an additional mechanistic contributor to attenuation caused by the K-to-E substitution. Discussion Orthobunyaviruses are an important group of arboviruses, but rational attenuation strategies for them remain limited32–35. In this study, we provide the first genus-level evidence in Orthobunyavirus demonstrating the critical role of a completely conserved lysine residue in RdRp function. We introduced a charge-reversal substitution (K-to-E) into representative viruses of both the Bunyamwera (EBIV/BUNV) and Simbu (OYAV) serogroups. All mutant RdRp or viruses exhibited marked replication impairment and attenuated virulence, confirming the indispensability of this lysine residue for viral replication and its potential as a broad-spectrum attenuation target. Functionally, the K-to-E substitution caused pronounced attenuation at both the cellular and animal levels. In EBIV, K760E markedly reduced polymerase activity, plaque size, viral load, and cytopathogenicity, while reducing lethality in mice and eliciting high neutralizing antibody titers (Fig. 2 and Supplementary Tables 3–5). Comparable attenuation phenotypes were also observed in BUNV and OYAV (Fig. 3), indicating that the functional importance of this residue is conserved in multiple orthobunyaviruses. Although in vivo validation in BUNV and OYAV is still required to fully establish the breadth of this function, these findings underscore the translational potential of K-to-E charge reversal as a rational strategy for live-attenuated vaccine development. Previous studies have confirmed that mutation of the key residues, such as M989A in LACV RdRp, can effectively reduce RdRp activity26. Mechanistically, biochemical assays of EBIV RdRp show a stepwise change in polymerase activity depending on the charge at this site: K760R ≥ WT > K760A > K760E (Fig. 2A, B), strongly supporting that the interaction of this conserved residue with the 3′ vRNA is essential for the polymerase function. Moreover, K760 is highly conserved across 22 orthobunyavirus serogroups (Fig. 1C and Supplementary Table 1), further indicating a strict functional requirement for a positively charged residue at this position. Specifically, K760R retains a positive charge with a longer side chain, strengthening electrostatic and hydrogen-bonding interactions with the 3′ vRNA and thereby enhancing polymerase activity. K760A neutralizes the positive charge, weakening binding and reducing activity. K760E causes charge reversal, abolishing binding and generating electrostatic repulsion that destabilizes the polymerase-template complex and significantly impairs activity. Furthermore, our data indicate that a new hydrogen bond introduced by the K-to-E substitution also contributes to viral attenuation (Fig. 5). Within the Peribunyaviridae family, seven additional genera are currently recognized besides Orthobunyavirus36. Comparative sequence analysis indicated that the conserved lysine residue is preserved within the genus Pacuvirus (Supplementary Fig. 7), suggesting that the attenuation strategy described here may also be applicable to Pacuvirus. However, functional validation will be required to confirm this possibility. By contrast, given the marked sequence divergence within this region across bunyavirus groups outside the Peribunyaviridae family, current data are insufficient to confirm that this strategy is applicable to those viruses. Although we have interrogated the impact of the K-to-E substitution on RdRp function using in vitro enzymatic assays, minigenome-based functional analyses, and combinatorial mutagenesis, direct structural evidence for how the mutation reshapes polymerase-RNA interactions or conformational transitions is still lacking. Furthermore, to use this strategy in vaccine development, several key steps still need to be refined: assessing mutation stability during serial passage; introducing additional site-specific mutations to enhance genetic stability37–39; incorporating other attenuation strategies (e.g., deletion of the NSs gene)40–43; conducting a thorough evaluation of challenge protection efficacy44,45; and performing systematic assessments of safety and immunogenicity46. In summary, this study identifies a universally conserved and effective attenuation strategy applicable to orthobunyaviruses, laying a valuable foundation for the development of live-attenuated vaccines or antiviral drugs against these viruses. Methods Ethics statement Animal experiments were approved by the Laboratory Animal Ethics Committee of Wuhan Institute of Virology, Chinese Academy of Sciences, Wuhan, China (permit number: WIVA12202401). All animal procedures were performed in accordance with the ethical guidelines. Cell lines and culture conditions Baby hamster kidney fibroblast BHK-21 cells (Mesocricetus auratus), African green monkey kidney epithelial Vero E6 cells, and human adrenal cortex epithelial SW13 cells were propagated in Dulbecco’s modified Eagle’s medium (DMEM; Gibco, US) supplemented with 10% heat-inactivated fetal bovine serum (FBS; Gibco) and 1% penicillin/streptomycin (P/S). All the above cell lines were obtained from the National Virus Resource Center (NVRC), Wuhan Institute of Virology, China. BHK-21 cells modified to stably express Escherichia phage T7 DNA-directed RNA polymerase (T7), designated BSR-T7/5 cells, were a kind gift from Prof. Karl-Klaus Conzelmann (Ludwig-Maximilians-University Munich, Munich, Germany), and were cultured in DMEM supplemented with 10% heat-inactivated FBS and 1 mg/mL G418 sulfate (Beyotime). All mammalian cells were cultured at 37 °C in a humidified incubator containing 5% CO2. Red junglefowl liver epithelial LMH cells (Gallus gallus) were maintained in DMEM/F-12 (HyClone) containing 10% heat-inactivated FBS and 1% P/S at 37 °C under 5% CO2. Viruses Wild-type viruses Oya virus (OYAV; species Orthobunyavirus catqueense) isolate SZC50 (GenBank: small [S] segment: OP672242; medium [M] segment: OP672243; large [L] segment: OP672244) and Ebinur Lake virus (EBIV; species Orthobunyavirus ebiense) isolate Cu20-XJ were used11,18. Recombinant viruses Based on our previous studies28, the wild-type EBIV (rEBIV/WT) and OYAV (rOYAV/WT) were rescued using reverse genetics systems. BSR-T7/5 cells were co-transfected with 500 ng of each plasmid (pLCK-EBIV-S, pLCK-EBIV-M, and pLCK-EBIV-L for EBIV; or pLCK-OYAV-S, pLCK-OYAV-M, and pLCK-OYAV-L for OYAV) using TransIT-LT1 Transfection Reagent (Mirus Bio LLC, USA) according to the manufacturer’s protocol. When obvious cytopathic effect (CPE) was observed, supernatants containing the rescued virus were collected. Subsequently, site-directed mutagenesis targeting the conserved lysine residues in the RdRp (EBIV L-K760 and OYAV L-K756) was performed to generate recombinant plasmids pLCK-EBIV-L/K760E and pLCK-OYAV-L/K756E, where lysine was substituted with glutamic acid (E). These plasmids were further utilized to rescue the corresponding mutant strains rEBIV/K760E and rOYAV/K756E. Supernatants containing virions were harvested when more than 80% of the BSR-T7/5 cells had exhibited CPE, and stored at -80 °C. The viral titers were determined by plaque assay as previously described28. Briefly, virus was serially diluted 10-fold, inoculated onto BHK-21 cell monolayers for 1 h, and then the inoculum was removed. The monolayers were overlaid with medium containing 1% methylcellulose and incubated at 37 °C until clear plaques formed. Afterward, the cells were stained with 2% crystal violet in 30% methanol for 5 min at room temperature, and plaques were manually counted and measured. Protein expression and purification The full-length coding sequence of EBIV RdRp was cloned into a modified pPICZ-B vector with a PreScission protease cleavage site and an N-terminal His8-tag. The EBIV L K760E, K760A, K760R, and D1164A/D1165A mutations were generated by site-directed mutagenesis. Wild-type and mutant RdRp of EBIV were expressed in Pichia pastoris. For protein purification, yeast cells (20 g) were disrupted by milling (Retsch MM400) and resuspended in lysis buffer containing 50 mM Tris-HCl, pH 8.0, 500 mM NaCl, and 10% glycerol supplemented with protease inhibitor (200 μM phenylmethylsulfonyl fluoride) and DNase I. The lysate was centrifuged for 1 h at 30,000 × g. The clarified supernatant was incubated with 3 mL Ni-NTA resin (GenScript) under gentle rotation for 3 h at 4 °C. Resin was then washed with 50 mL buffer containing 50 mM Tris-HCl pH 8.0, 500 mM NaCl, 10% glycerol and 100 mM imidazole. Protein was eluted with buffer containing 50 mM Tris-HCl, pH 8.0, 500 mM NaCl, 10% glycerol, and 500 mM imidazole. The eluted protein sample was subjected to size-exclusion chromatography using a Superose 6 Increase 10/300 column (GE), and the fractions containing EBIV RdRp were pooled. The protein was concentrated to 10 mg/mL and stored in gel filtration buffer (50 mM Tris-HCl, pH 8.0, 500 mM NaCl, 2 mM DTT). In vitro primer-dependent RNA synthesis assay The purified protein of EBIV RdRpWT, RdRpK760E, RdRpK760A, and RdRpK760R at 4 μM in a buffer containing 50 mM Tris-HCl pH 7.5, 2 mM DTT, 150 mM NaCl, 300 μM ATP, 300 μM UTP, and 2 mM MnCl2 was mixed with 6 μM RNA template (5′-GUAUAUCUCCAUUAUUUUUAUCC, synthesized by Sangon) annealed with 30 μM pGGAU primer (Takara). The mixture was then incubated at 30 °C for 30 min in a total volume of 20 μL. An equal volume of stop solution consisting of 95% (v/v) formamide, 20 mM EDTA (pH 8.0), and 0.02% (w/v) xylene cyanol FF was added to stop the reaction. The mixture was analyzed on a 20% denaturing polyacrylamide gel containing 7 M urea and stained with Stains-All (Sigma-Aldrich, E9379). Band intensities were quantified using ImageJ software. Viral growth kinetics To investigate the impact of the lysine-to-glutamic acid (K-to-E) substitution on viral replication kinetics, BHK-21, SW13, Vero E6, and LMH cells were seeded in T25 flasks (1 × 106 cells/flask; Corning) and incubated overnight at 37 °C with 5% CO₂. Monolayers were inoculated with either wild-type (rEBIV/WT or rOYAV/WT) or mutant viruses (rEBIV/K760E or rOYAV/K756E) at a multiplicity of infection (MOI) of 0.01. After incubation for 1 h at 37 °C, the supernatants were removed. The cells were washed three times with phosphate-buffered saline (PBS; Gibco), and 5 mL of fresh medium containing 2% heat-inactivated FBS and 1% P/S was added. After inoculation, 400 µL of supernatant was collected every 24 h, and an equal volume of fresh medium was added back to the culture. The collected supernatants were stored at −80 °C, and the viral RNA copies in the supernatants were determined with RT-qPCR. The One Step TB Green® PrimeScript™ Plus RT-PCR Kit (Takara), HiScript II One Step qRT-PCR Probe Kit (Vazyme), and the CFX96 Touch Real-Time PCR Detection System (Bio-Rad) were used for RT-qPCR. RT-qPCR was performed using S-segment-targeted primers. For EBIV, primer sequences from Yang et al. with standard curve y = −4.0312x + 43.384, where x = lg (genome copy/μL total RNA of rEBIV/WT), y = Ct value, and R2 = 0.9905 (Supplementary Table 6)47. For OYAV: Primer sequences from Liu et al. with standard curve y = −3.296x + 43.232, where x = lg (genome copy/μL total RNA of rOYAV/WT), y = cycle threshold [Ct] value, and R2 = 1.000 (Supplementary Table 7)48. Minigenome assay To assess the activity of the RdRp, minigenome systems were established for EBIV, BUNV, and OYAV. For EBIV, the open reading frame (ORF) of the S segment in rEBIV/WT plasmids was replaced with the NLuc ORF to generate the reporter plasmid pLCK-EBIV-SUTR-NLuc. The RdRp and nucleoprotein (NP) were expressed using the pCAGGS vector. BSR-T7 cells that had reached 50% confluence were transfected with pCAGGS-EBIV-L, pCAGGS-EBIV-NP, and pLCK-EBIV-SUTR-NLuc in a 1:1:1 ratio (500 ng of each plasmid). At 2 days post-transfection (dpt), the supernatant was removed, and the cells were lysed with 100 μL of 1% NP-40 (Kehbio). Nano-Glo® Luciferase Assay Substrate (Promega) was diluted 50-fold with Nano-Glo Luciferase Assay Buffer (Promega), and then 10 μL of the lysis product was mixed with 10 μL of substrate (Promega). Luminescence was measured with a Varioskan LUX Multimode Microplate Reader (Thermo Fisher Scientific). NLuc activity was calculated as (luminescence value of each sample/mean luminescence value of the WT group) × 100%. Following the same procedure, analogous minigenome systems for BUNV (pCAGGS-BUNV-L, pCAGGS-BUNV-NP, pLCK-BUNV-SUTR-NLuc) and OYAV (pCAGGS-OYAV-L, pCAGGS-OYAV-NP, pLCK-OYAV-SUTR-NLuc) were used to evaluate the RdRp activity. CCK-8 assay Virus-induced cytotoxicity was quantified with Cell Counting Kit-8 (CCK-8; MedChemExpress). Briefly, cells were seeded in 96-well plates (1 × 104 cells per well) and grown for 24 h. The cell monolayers were inoculated with viruses at MOI = 0.01. At 48 h post-inoculation, the supernatant was removed, and the cells were incubated with 100 μL of fresh DMEM with 10% CCK-8 reagent for 1 h at 37 °C. The absorbance at 450 nm was measured with a microplate reader (Thermo Fisher Scientific). Cell viability in the infected group and mock group was calculated as (individual absorbance value/mean absorbance of the mock group) × 100%. Animal experiment To determine the effect of the K‑to‑E substitution on viral virulence, female specific pathogen-free (SPF) BALB/c mice (6–8 weeks old) were randomly assigned into four groups (n = 5 per group). Each mouse was intraperitoneally injected with 100 µL of DMEM containing 5 or 10 PFU of rEBIV/WT or rEBIV/K760E. After infection, mice were monitored daily for clinical symptoms and body weight changes. Mice were anesthetized with isoflurane (RWD) inhalation and then euthanized via cervical dislocation if they lost more than 20% of initial body weight, in accordance with humane endpoint criteria. Orbital blood samples were collected from each mouse on 1, 3, 5, 7, 11, and 14 dpi. Sera were separated by centrifugation at 3000 × g for 20 min and stored at −80 °C until use. Viral RNA was extracted from serum samples and quantified using a RT-qPCR assay. All experimental procedures involving the mice were carried out in an ABSL-2 facility, where mice were maintained with controlled temperature (22°C). We have complied with all relevant ethical regulations for animal use, and all data were included in the analysis. To reduce confounders, cages were placed randomly in Individually Ventilated Cage System (IVC). All treatments and measurements were conducted in blinded, random order at fixed intervals. The researcher was unaware of group assignments throughout the experiment. Neutralizing antibody test To evaluate the neutralizing antibody in mice against rEBIV/WT, the plaque reduction neutralization test (PRNT) was employed. Briefly, mouse sera collected at 14 dpi from rEBIV/K760E infection group were heat-inactivated at 56 °C for 30 min and diluted to 1:50, followed by 3-fold serial dilutions (dilution factors: 1: 50, 1: 150, 1: 450, 1: 1350, and 1: 4050) using DMEM supplemented with 2% FBS and 1% P/S. Sera were then mixed with 100 PFU of rEBIV/WT and incubated at 37 °C for 1 h. The virus-serum dilution mixtures and virus-only control were then inoculated to BHK-21 cell monolayers in 24-well plates for 1 h before adding an overlay medium including 1.5% methylcellulose at 37 °C for 3–5 days to allow plaque development and measured28. The neutralizing antibody was calculated based on a 50% plaque reduction neutralization titer (PRNT50). Finally, the neutralization rate (%) for each serum dilution was calculated using the following formula: neutralization rate (%) = [1 − (mean plaque number in virus-serum group/mean plaque number in virus-only control group)] × 100%. Sequence analysis and structural prediction Sequence comparisons were performed with MEGA 7.049 and plotted with ESPript 3.250. The plaque areas were calculated with the ImageJ software51. Structural predictions of proteins were made using AlphaFold3 (https://alphafoldserver.com/) with default parameters52. The effects of point mutations on RdRp structural stability were predicted using the DynaMut2 web server (https://biosig.lab.uq.edu.au/dynamut2/submit_prediction). Residues K759, H757, H760, and H761 were individually substituted with glycine (G), the simplest amino acid. Predicted changes in folding free energy (ΔΔGstability, kcal/mol) were retrieved from the server output, where negative values indicate destabilizing effects. The structure of the WT RdRp was predicted by submitting its full-length amino acid sequence to AlphaFold3 with default parameters. The structures of mutant variants discussed in this study were subsequently generated by introducing specific point mutations into the corresponding WT RdRp structure using ChimeraX 1.953. Specifically, target residues were mutated to the designated amino acids in ChimeraX 1.9, and all possible side-chain conformations of the mutated residues were analyzed via the rotamer library (Dunbrack) integrated within the software. A mutation was considered capable of forming hydrogen bonds if any of the side-chain conformations could establish such bonds; otherwise, it was considered unable to generate new hydrogen bonds. All interaction analyses and structural visualizations were performed using ChimeraX 1.9. Statistics and reproducibility All statistical analyses were performed using GraphPad Prism 9.1.0. Normality of all datasets was assessed using Shapiro–Wilk test, and homogeneity of variance was evaluated via Brown–Forsythe test and Bartlett’s test. For multiple-group comparisons, standard one-way ANOVA was performed for datasets with equal variances, whereas Welch’s one-way ANOVA was adopted for normally distributed datasets with unequal variances. For two-group comparisons, three statistical methods were applied according to data distribution and variance homogeneity: a two-tailed unpaired t-test for normally distributed data with equal variances, a two-tailed unpaired t-test with Welch’s correction for normally distributed data with unequal variances, and a two-tailed Mann–Whitney U-test for data deviating from normal distribution. n indicates the number of biologically independent experiments. Data are presented as mean ± SD. Exact p-values are provided, and p ≤ 0.05 was considered statistically significant. All experiments were repeated at least three times with consistent results, confirming reproducibility. Supplementary information Transparent Peer Review file Supplementary information Description of Additional Supplementary Files Supplementary Data 1 Supplementary Data 2 nr-reporting-summary Acknowledgements The authors would like to thank Dr. Éric Bergeron (US Centers for Disease Control and Prevention) for valuable suggestions, Professor Jingwen Wang (State Key Laboratory of Genetic Engineering, School of Life Sciences, Fudan University) for the technical support provided, and the assistance from the staff of Institutional Center for Shared Technologies and Facilities of the Wuhan Institute of Virology, CAS. Author contributions Study conception and experimental design were performed by Guodian Xiong, Zhiming Yuan, Han Xia, and Jinglin Wang. Project administration and funding acquisition were carried out by Zhiming Yuan, Han Xia, and Jinglin Wang. Data collection and analysis were contributed by Guodian Xiong, Fei Wang, Jingjing Tang, Tao Cheng, Nanjie Ren, Jingke Geng, Wei Chen, and Doudou Huang. Manuscript writing and revision were completed by Guodian Xiong, Fei Wang, Jingjing Tang, Xiang Sun, Zengqin Deng, Zhiming Yuan, Han Xia, and Jinglin Wang. Figure preparation was conducted by Guodian Xiong, Fei Wang, Jingjing Tang, Tao Cheng, and Nanjie Ren. Peer review Peer review information Communications Biology thanks Loïc Carrique and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Primary Handling Editors: Juan Fontana, Aylin Bircan, David Favero, and George Inglis. A peer review file is available. Funding This work was supported by the National Key Research and Development Program of China grant 2023YFC2305900 (to Han Xia), the National Natural Science Foundation of China grants U22A20363(to Zhiming Yuan), 82572569 (to Han Xia), and 32260896 (to Jinglin Wang). Data availability The GenBank accession numbers of Oya virus are OP672242 (S segment), OP672243 (M segment), and OP672244 (L segment). Those of EBIV are ON055166 (S segment), ON055167 (M segment), and ON055168 (L segment)28. The PDB accession codes of the RdRp protein structures used in this study are 5AMR25, 5AMQ25, 6Z6G54, 7ORL26, 7ORK26, 7ORJ26, 7ORM26, 7ORN26, 7ORO26, and 7ORI26. All source data supporting the findings are provided in the main text and Supplementary Data 2. Competing interests Han Xia, Jinglin Wang, Guodian Xiong, Fei Wang, and Zhiming Yuan declare a patent application for the Oya virus RdRp mutant, its expression vector, vector composition, and uses thereof. All other authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. These authors contributed equally: Guodian Xiong, Fei Wang, Jingjing Tang. Contributor Information Jinglin Wang, Email: wangjl107@163.com. Zhiming Yuan, Email: yzm@wh.iov.cn. Han Xia, Email: hanxia@wh.iov.cn. Supplementary information The online version contains supplementary material available at https://doi.org/10.1038/s42003-026-10459-7. References Associated Data Supplementary Materials Transparent Peer Review file Supplementary information Description of Additional Supplementary Files Supplementary Data 1 Supplementary Data 2 nr-reporting-summary Data Availability Statement The GenBank accession numbers of Oya virus are OP672242 (S segment), OP672243 (M segment), and OP672244 (L segment). Those of EBIV are ON055166 (S segment), ON055167 (M segment), and ON055168 (L segment)28. The PDB accession codes of the RdRp protein structures used in this study are 5AMR25, 5AMQ25, 6Z6G54, 7ORL26, 7ORK26, 7ORJ26, 7ORM26, 7ORN26, 7ORO26, and 7ORI26. All source data supporting the findings are provided in the main text and Supplementary Data 2.