Rapid production of recombinant rotaviruses by overexpression of NSP2 and NSP5 genes with modified nucleotide sequences

ABSTRACT

Reverse genetics systems for rotaviruses (RV) facilitate the generation of genetically engineered RVs by transfection of 11 plasmids encoding 11 genomic viral RNA segments. In addition to viral genome expression, overexpression of NSP2 and NSP5 has been used to increase the rescue efficiency of recombinant RVs. Here, we showed that the overexpression of nucleotide sequence-modified NSP2 and NSP5 enabled the rapid and efficient production of recombinant RVs. Using improved reverse genetics, we established a reverse genetics system for human and bovine RV clinical isolates, as well as laboratory strains of bovine RV (NCDV and UK) and porcine RV (Gottfried). In addition, we rescued low-replicating recombinant RVs carrying a mutant NSP4 lacking the double-layered particle-binding domain, which was deficient in the efficient production of mature virions. These advancements in reverse genetics enabled the generation of molecular clones of RV clinical isolates and recombinant RVs harboring critical amino acid mutations, offering a versatile platform for investigating RV biology and pathogenesis.

IMPORTANCE

Recombinant rotavirus (RV) synthesis via reverse genetics relies on both the viral propagation capacity and the efficiency of the experimental system. Since the establishment of our reverse genetics system, several enhancements have been implemented to augment the rescue efficiency. Nevertheless, challenges persist in generating RV clinical strains and recombinant viruses with low replication capacities. Notably, this improved reverse genetics system successfully facilitated the establishment of molecular clones of human and bovine RV clinical isolates. Fecal samples from patients with RV typically harbor quasi-species or, occasionally, multiple genotypes of RV. In the present study, we performed the genetic sequencing of clinical viral strains during the early propagation stages in cultured cells. Subsequently, infectious viruses were synthesized, allowing the characterization of circulating viruses in nature. This approach provides valuable insights into the genetic diversity and dynamics of RV populations and contributes to a more comprehensive understanding of viral pathogenesis and evolution.

INTRODUCTION

Rotaviruses (RVs) are non-enveloped viruses of the family Reoviridae that contain an 11-segmented double-stranded (ds) RNA genome within a non-enveloped virion. RVs are enteric pathogens that cause severe gastroenteritis in infants and young animals. Despite the global introduction of RV vaccines in 2006, an estimated 128,500 RV-associated deaths were recorded worldwide in 2016 (1). We previously established a plasmid-based reverse genetics system (2–4). This system facilitates the generation of various recombinant viruses, including those bearing mutant viral proteins, exogenous reporter genes, and heterologous capsid proteins (5–9). Recombinant viruses have been instrumental in the investigation of viral replication, the development of RV vectors, and the production of candidate vaccines. However, the rescue efficiency of recombinant RVs is predominantly dependent on viral replication capacity. Consequently, the generation of recombinant RVs with low replicative ability due to critical mutations or deletions in viral proteins is hampered by reduced viral rescue efficiency (5, 10, 11).

The foundational protocol for rotavirus plasmid-based reverse genetics involves the co-transfection of 11 plasmids, each encoding one of the 11 gene segments of the rotavirus genome, along with three plasmids encoding the fusion-associated transmembrane protein of the Nelson Bay reovirus and the RNA capping enzyme of the vaccinia virus. Enhanced recombinant RV rescue efficiency has been achieved via the overexpression of NSP2 and NSP5 proteins (12, 13). Further improvements in viral rescue efficiency have been attained by co-culture with genetically engineered MA104 cells, which are characterized by an impaired innate immune response (11). NSP2 and NSP5 are crucial for viroplasm formation via interactions with cellular proteins and lipid droplets during the viral replication cycle (14). The process of liquid-liquid phase separation is pivotal for the sequestration of electron-dense structures, with the viroplasm becoming discernible as early as 2–3 h post-infection (15, 16). These viroplasms act as scaffolds for the assembly of the viral genome and the structural proteins essential for RV replication (17–20). Although the specific roles of NSP2 and NSP5 overexpression in reverse genetics have not been fully elucidated, the initial formation of the viroplasm has been hypothesized to facilitate efficient progeny virion production, initiating the replication cycle.

In our previous study, NSP2 and NSP5 were overexpressed using a plasmid vector that encoded only the open reading frames (ORF) of NSP2 and NSP5 (13). Concerns have been raised regarding the expression of viral genes devoid of the untranslated region (UTR) of NSP2 and NSP5, as this is anticipated to hinder the genome packaging process and viral genome replication. This issue is significant because viral genomes are predicted to form inter-genome segment complexes via base-pairing during genome packaging (21, 22). In addition, the UTRs of the RV genome harbor domains crucial for viral replication, including the RNA-dependent RNA polymerase recognition domain, cis-elements, and translation enhancer domains (23–25). The absence of UTRs in NSP2 and NSP5 resulted in the production of replication-deficient viruses.

In this study, codon-optimized NSP2 and NSP5 genes (NSP2opt and NSP5opt, respectively) were designed to increase the protein expression level and to avoid the negative effects of co-expression of NSP2 and NSP5 genes lacking UTRs for developing a highly efficient reverse genetics system for rotaviruses. Although the expression level of NSP5opt was slightly lower than that of unmodified NSP5, co-expression of NSP2opt and NSP5opt increased the rescue efficiency of recombinant RVs in reverse genetics. This improvement has been exemplified by the creation of successful recombinant human and bovine rotaviruses from clinical isolates and recombinant rotaviruses lacking the double-layered particle binding domain of the NSP4 protein, showcasing the system’s versatility. The impact of this new technique is far-reaching in both basic and clinical research. Furthermore, this system opens new avenues for the development of attenuated rotavirus vaccines, as it allows for the precise engineering of viruses with reduced replication capacity. Overall, our improved reverse genetics system is poised to significantly enhance research and clinical applications in the rotavirus field.

RESULTS

Overexpression of nucleotide sequence-modified NSP2 and NSP5 promotes rapid production of recombinant RV by reverse genetics

To improve the reverse genetics system, nucleotide sequence-modified NSP2 and NSP5 were designed using a codon optimization strategy, whose codon bias was similar to that of mammalian cells, to maximize protein expression (Fig. S1). The nucleotide similarity between NSP2 and NSP2opt, and between NSP5 and NSP5opt, was 75.3% (236/954 bp) and 72.0% (167/597 bp), respectively. When both proteins were expressed by CAG promoter-based plasmid vectors, the expression of NSP2opt was higher than that of native NSP2, whereas the expression of NSP5opt was lower than that of native NSP5 (Fig. 1A and B). To examine the effect of NSP2opt and NSP5opt in reverse genetics, 11 rescue plasmids for simian RV strain SA11 (pT7-VP1, -VP2, -VP3, -VP4, -VP6, -VP7, -NSP1, -NSP2, -NSP3, -NSP4, and -NSP5) and two expression plasmids encoding RNA capping enzymes (pCAG-D1R and pCAG-D12L) were co-transfected with expression plasmids encoding NSP2 and NSP5 (pCAG-NSP2 and pCAG-NSP5) or NSP2opt and NSP5opt (pCAG-NSP2opt and pCAG-NSP5opt) (Fig. 1C). At 48 h post-transfection (h.p.t.), the cell culture medium was replaced with a viral growth medium (FBS-free DMEM supplemented with trypsin) and co-cultured with MA104 cells for viral propagation. A partial cell culture medium was collected every 24 h to examine the infectious virus titers. At 168 h.p.t., whole-cell lysates were obtained. First, an equal amount of each plasmid (0.125 µg/well) was used for transfection. Using NSP2opt and NSP5opt, infectious viruses were detected as early as 48 h.p.t. Infectious viruses were recovered from every six replicates at 72 h.p.t., whereas in the experiment using NSP2 and NSP5, the infectious viruses were not recovered until 96 h.p.t. (Fig. 1D). This difference in the rescue efficiency was more pronounced when only 0.03125 µg (per well) rescue plasmids were used (Fig. 1E). Nucleotide sequencing confirmed that rSA11 generated using NSP2opt and NSP5opt had unmodified NSP2 and NSP5 genes. The result indicated that NSP2opt and NSP5opt were not incorporated as viral genomes. To examine the heterogenous effect of NSP2opt and NSP5opt in reverse genetics of different RV strains, rescue plasmids encoding 11 gene segments of human RV strain Odelia (0.015625, 0.03125, 0.0625, or 0.125 µg/well) were co-transfected with pCAG-NSP2 and -NPS5 or pCAG-NSP2opt and -NSP5opt (Fig. 1F). At 96 h.p.t., infectious virus titers in whole-cell lysates were examined. Under every experimental condition, the larger amount of recombinant RV was rescued by the overexpression of NSP2opt and NSP5opt in transfected wells compared with that of NSP2 and NSP5. It was observed that using 0.0625 µg of plasmid per well produced a larger amount of recombinant virus compared with using 0.125 µg per well. This result may be related to cell damage caused by the increased amount of transfection reagent (2 µL per µg plasmid) used at higher plasmid concentrations. The higher expression of NSP2opt compared with NSP2 and the lower expression of NSP5opt compared with NSP5 were consistent when these genes were co-expressed with the complete set of rescue plasmids. (Fig. S2). These results indicated that co-expression of NSP2opt and NSP5opt enabled rapid and efficient production of recombinant simian and human RV with high efficiency using a small amount of plasmid, despite the reduced expression of NSP5opt.

Fig 1.

Overexpression of NSP2opt and NSP5opt facilitated rapid production of recombinant rotavirus. (A and B) Expression of NSP2, NSP2opt, NSP5, and NSP5opt using plasmid vectors. β-actin was used as a loading control. Protein expression ratio to β-actin was quantified using ImageJ software. (C) Schematic of the experimental protocol for reverse genetics systems for RV. Eleven rescue plasmids encoding simian RV strain SA11 genome and expression plasmids encoding RNA-capping enzymes were co-transfected with expression plasmids encoding NSP2 and NSP5 or expression plasmids encoding NSP2opt and NSP5opt in BHK-T7 cells. At 48 h post-transfection (h.p.t.), the cell culture medium was replaced with serum-free DMEM supplemented with trypsin (0.5 µg/mL). At 48, 72, 96, 120, and 144 h.p.t., a portion of the cell culture medium was collected to examine the presence of the virus. At 168 h.p.t., cells were lysed by freeze-thaw cycles, and virus titers in the cell lysates were examined. (D and E) Production of recombinant SA11 after transfection with rescue plasmids (D) 0.125 µg/mL each, (E) 0.03125 µg/mL each), capping enzyme-encoding plasmids, and expression plasmids encoding NSP2 and NSP5 or NSP2opt and NSP5opt. (F) Production of recombinant human RV strain Odelia after transfection of 11 rescue plasmids (0.015625 ~ 0.125 µg/well as indicated), capping enzymes encoding plasmids and expression plasmids encoding NSP2 and NSP5 or NSP2opt and NSP5opt. Virus titers in the culture media or cell lysates were examined at 96 h.p.t. (D-F) Dashed lines indicate the limit of detection (1.0 × 10¹ FFU/mL). The values under the limit of detection line were calculated as 1.0 × 10⁰ FFU/mL. Virus titers between NSP2/5 and NSP2opt/5opt were statistically analyzed using the Mann-Whitney U test. * indicate P < 0.05.

Adverse effect on the overexpression of NSP5 in reverse genetics

To investigate the contribution of NSP2opt and NSP5opt to RV reverse genetics, NSP2, NSP5, NSP2opt, and NSP5opt were co-transfected with RV rescue plasmids in various combinations (Fig. 2A through C). At 72 h.p.t., a significant increase in viral production was observed only when NSP2opt and NSP5opt were simultaneously transfected (Fig. 2B). Next, heterologous combinations of parental and codon-optimized NSP2 and NSP5 were examined. A significant increase in the infectious virus titer was observed when NSP5opt were co-transfected with NSP2 in comparison to that observed with NSP2 and NSP5 (Fig. 2C). Meanwhile, infectious virus titers after co-transfection with NSP2opt and NSP5 were comparable with those of the combination of NSP2 and NSP5 (Fig. 2C). Co-transfection with NSP2opt and NSP5opt produced the highest number of recombinant viruses. These results indicated that NSP5opt played a critical role in the improvement of this system, with the rescue efficiency being significantly enhanced when both NSP2opt and NSP5opt were used simultaneously. It was also clear that NSP5opt had a more substantial impact than NSP2opt.

Fig 2.

Positive and negative contribution of NSP2 and NSP5 overexpression in the reverse genetics system. (A) Schematic experimental protocol for evaluation of rescue efficiency of reverse genetics system. SA11 eleven rescue plasmids (0.125 µg/well) were co-transfected with expression plasmids encoding NSP2, NSP2opt, NSP5, and/or NSP5opt (0.125 µg/well) in different combinations as indicated. At 72 h post-transfection (h.p.t.), virus titers in the cell lysates were examined. (B-C) NSP2, NSP2opt, NSP5, and/or NSP5opt were co-transfected with 11 rescue plasmids. Virus infectious titers at 72 h.p.t. were examined. (D) Constructions of plasmid encoding deficient NSP2 (dNSP2), dNSP2opt, dNSP5, and dNSP5opt. TAA stop codons were inserted after the 18th nucleotide of NSP2, NSP2opt, NSP5, and NSP5opt to interrupt the open reading frames. Expression plasmid encoding dNSP5 1–200th, 201–398th, 399–597th, and 1–200th, and 398–597th nucleotides were constructed. (E-G) SA11 eleven rescue plasmids and expression plasmids encoding NSP2opt and NSP5opt were co-transfected with expression plasmids encoding dNSP2, dNSP2opt, dNSP5 (whole, 1–200, 201–398, 399–597, 1–200/399-597), dNSP5opt (0.125 µg/well each) in different combinations as indicated. At 72 h.p.t., virus titers in the cell lysates were examined. (H) SA11 eleven rescue plasmids and NSP2opt and NSP5opt expression plasmids were transfected with or without expression plasmids encoding RNA capping enzyme (D1R and D12L). At 72 h.p.t., virus titers in the cell lysates were examined. Dashed lines indicate the limit of detection (1.0 × 10¹ FFU/mL). The values under the limit of detection line were calculated as 1.0 × 10⁰ FFU/mL. Statistical analyses were calculated using the Kruskal-Wallis test (A, B, D, E, and F) or the Mann-Whitney U test (G). *<I>P < 0.05.

The lower expression of NSP5opt than NSP5 (Fig. 1B) suggests that the improvement in rescue efficiency was not directly correlated with the level of protein expression. As the CAG promoter-based plasmid encodes only the open reading frames (ORFs) of NSP2 and NSP5 proteins but lacks the 5’ and 3’ UTR, the negative impact of gene overexpression on the viral replication process is concerning. To investigate the effect of the overexpression of NSP2 and NSP5 genes lacking UTR in reverse genetics, CAG promoter-based plasmids encoding deficient NSP2, NSP5, NSP2opt, and NSP5opt (dNSP2, dNSP5, dNSP2opt, and dNSP5opt, respectively), with an additional stop codon within the ORF to prevent functional protein expression, were prepared (Fig. 2D). Plasmids encoding deficient proteins were co-transfected with NSP2opt/NSP5opt and rescue plasmids. Co-expression of dNSP2 or dNSP2opt did not affect the rescue efficiency (Fig. 2E), whereas co-expression of dNSP5 and dNSP5opt decreased recombinant virus production (Fig. 2F). This adverse effect was more prominent in dNSP5 than in dNSP5opt. Because dNSP5 does not encode a functional NSP5 protein, these results suggest that the nucleotide sequence encoding dNSP5 itself inhibited recombinant virus production. To determine the region responsible for dNSP5 (597 bp), truncated dNSP5 genes with nucleotides 1–200, 201–398, 399–597, and 1–200/399-597 nucleotides were used (Fig. 2D). These truncated gene segments were expressed using NSP2opt/NSP5opt and rescue plasmids, and infectious virus titers were examined at 72 h.p.t. Compared with the control, viral titers significantly decreased only when whole dNSP5 was expressed (Fig. 2G).

We investigated the contribution of RNA-capping enzymes in improving reverse genetics. Plasmids encoding vaccinia virus RNA-capping enzyme proteins D1R and D12L were co-transfected with NSP2opt/NSP5opt and rescue plasmids. At 72 h.p.t., infection titers of recombinant viruses were comparable, independent of the expression of D1R/D12L, indicating that RNA-capping enzymes could be omitted in the improved protocol (Fig. 2H).

Generation of recombinant RV laboratory strains

Using this improved strategy, we attempted to establish a reverse genetics system for RV laboratory strains. Complete rescue plasmids encoding each of the 11 gene segments of bovine RVs (UK, NCDV) and porcine RV (Gottfried) were co-transfected with SA11 NSP2opt and NSP5opt expression plasmids, following the standard reverse genetics procedure. Electropherotypes of dsRNA genomes obtained from recombinant NCDV (rNCDV), rUK, and rGottfried showed the same pattern as wild-type viruses (Fig. 3A through C). To preclude the possibility of contamination with wild-type virus, recombinant viruses incorporating unique EcoRI sites within the NCDV VP6, UK NSP1, and Gottfried VP2 genes were generated (rNCDV-VP6/EcoRI, rUK-NSP1/EcoRI, and rGottfried-VP2/EcoRI). EcoRI treatment of the PCR amplicons of the viral genome verified the presence of the EcoRI site in the dsRNA genomes of rNCDV-VP6/EcoRI, rUK-NSP1/EcoRI, and rGottfried-VP2/EcoRI but not in the wild-type viruses (Fig. 3A through C). The replication kinetics of these recombinant viruses were comparable with those of the parental viruses, indicating that rNCDV, rUK, and rGottfried retained their original phenotypes.

Fig 3.

Generation of reverse genetics systems for animal RV laboratory strains. Rescue plasmids encoding 11 viral genomes of (A) bovine RV strain NCDV genome, (B) bovine RV strain UK, and (C) porcine RV strain Gottfried were co-transfected with expression plasmid encoding NSP2opt and NSP5opt derived from SA11 to BHK-T7 cells, and the standard reverse genetics procedure was followed. Recombinant viruses with additional restriction enzyme recognition sites (A. EcoRI site in the NCDV VP6 gene, B. EcoRI site in the UK NSP1 gene, and C. EcoRI site in the Gottfried VP2 gene) as genetic markers were generated. (A-C, left) Electropherotypes of the viral dsRNA genomes. (A-C, center) Examination of genetic markers in the recombinant viral genome. (A-C, right) Monolayers of MA104 cells were infected with the wild-type and recombinant viruses at a multiplicity of infection value of 0.001 and incubated in the presence of trypsin. At 0, 24, 48, and 72 h post-infection (h.p.i.), viral infectious titers in the cell lysates were examined using a focus-forming assay. Virus titers at 72 h.p.i. were statistically analyzed using ANOVA. n.s. not significant.

Generation of recombinant human and bovine RV clinical isolates

A refined reverse genetics system was used to generate molecular clones of the clinical isolates of human and bovine RVs. The human RV strain RVA-U4 (G3P[8]) was isolated from a Japanese patient in 2017 using MA104 cells. After an additional passage in MA104 cells, viral RNA was purified, and the whole genome sequences were determined as described in our previous study (5). In our previous study, we attempted to rescue the recombinant RVA-U4 using an unmodified reverse genetics protocol using NSP2 and NSP5; however, we failed to generate recombinant RVA-U4 (5). Bovine RV strain KM3 was isolated from cattle in Japan in 2018 using MA104 cells. After an additional passage in MA104 cells, viral RNA was purified, and whole gene sequences were determined by full-length amplification using the cDNAs method (26). Whole gene segments of RVA-U4 and KM3 were individually cloned into reverse genetics rescue plasmids. An online genotyping tool (https://www.rivm.nl/mpf/typingtool/rotavirusa/) classified the genotypes of these clinical isolates as RVA-U4 (G3-P[8]-I2-R2-C2-M2-A2-N2-T2-E2-H2) and KM3 (G6-P[5]-I2-R2-C2-M2-A3-N2-T6-E2-H3). Eleven plasmids encoding the viral genes of RVA-U4 and KM3 were transfected with expression plasmids encoding D1R, D12L, NSP2opt, and NSP5opt into BHK-T7 cells. The successful generation of recombinant RVA-U4 (rRVA-U4) and recombinant KM3 (rKM3) was confirmed by an immunofluorescence assay using murine anti-SA11 serum. The dsRNA genome electropherotypes of rRVA-U4 matched those of the parental RVA-U4 (Fig. 4A). Recombinant rRVA-U4-VP2/EcoRI carrying a unique EcoRI site in VP2 was engineered to exclude potential wild-type virus contamination (Fig. 4B). Nucleotide sequencing confirmed the presence of the engineered genetic marker at position 589 in VP2, with EcoRI treatment cleaving the corresponding VP2 amplicon from rRVA-U4-VP2/EcoRI (Fig. 4B). The viral growth curves of the parental RVA-U4, rRVA-U4, and rRVA-U4-VP2/EcoRI strains were similar (Fig. 4C), indicating that both recombinant viruses retained characteristics of the parental virus. Furthermore, recombinant KM3, which lacked the 1178 bp NSP1 open reading frame (rKM3-NSP1Δ1178), and recombinant KM3 carrying a red fluorescent protein gene within the truncated NSP1 gene (rKM3-mScarlet) were rescued (Fig. 4D through H). Electropherotypes of the dsRNA genome purified from rKM3-NSP1Δ1178 and rKM3-mScarlet confirmed the partial (1178 bp) deletion of the NSP1 and mScarlet-NSP1 genes (total 1005 bp), respectively (Fig. 4E and F). The viral replication curves of rKM3-NSP1Δ1178 were slower than those of KM3 and rKM3 (Fig. 4G). The rKM3-mScarlet infected cells exhibited a red fluorescent signal (Fig. 4H). These results demonstrate that the improved reverse genetics system using NSP2opt and NSP5opt is a powerful tool for generating molecular clones of RV clinical isolates.

Fig 4.

Establishment of molecular clones of human and bovine RV clinical isolates. (A) Electropherotypes of dsRNA genomes purified from RVA-U4, rRVA-U4, and rRVA-U4-VP2/EcoRI. (B, upper) Nucleotide mutations in the VP2 gene of rRVA-U4-VP2/EcoRI. C to T substitution at the 589^th nucleotide creates an EcoRI recognition site (underlined). (B, lower) Restriction enzyme digestion analysis of the VP2 gene segment of rRVA-U4-VP2/EcoRI. PCR amplicons from viral cDNA were purified and digested with EcoRI. (C) Replication kinetics of RVA-U4 and rRVA-U4. (D-H) Generation of recombinant bovine rotavirus clinical isolate KM3 strain. (D) Schematic showing generation of truncated KM3 NSP1 gene and NSP1-mScarlet gene. (E-F) Electropherotypes of dsRNA purified from (E) KM3, rKM3, and rKM3-NSP1Δ1178, and (F) rKM3 and rKM3-mCherry. (E) Arrow: NSP1; Arrowhead: NSP1Δ1178. (F) Arrow: NSP1; Arrowhead: NSP1-mScarlet. (G) Replication kinetics of KM3, rKM3, rKM3-NSP1Δ1178, and rKM3-mScarlet in MA104 cells. Virus titers at 72 h.p.i. were compared with that of KM3 and statistically analyzed by ANOVA. n.s. not significant. * P < 0.05. (H) Visualization of rKM3-mScarlet infected cells. Viral NSP4 protein (green) was detected by rabbit anti-NSP4 antibody and goat anti-rabbit IgG CF488 conjugate. Cell nuclei were stained by Hoechst 33342. Scale bars: 20 µm.

Generation of recombinant RV encoding deficient NSP4 lacking double-layered particle (DLP) binding domain

Finally, using this improved reverse genetics system, recombinant SA11 viruses encoding a mutant NSP4 gene lacking the DLP-binding domain were generated. The non-structural protein NSP4 contributes to the formation and maturation of progeny virions. During the viral replication cycle, NSP4 is localized within the membrane of the endoplasmic reticulum (ER), where it functions as an intracellular receptor for DLP and induces DLP budding into the viroplasm-associated membrane (27, 28). Amino acids 156–175 of NSP4 are considered DLP-binding domains (28–30). To examine the biological function of the DLP-binding domain of NSP4 in the replication cycle, rescues of recombinant SA11 encoding the C-terminal-truncated NSP4 lacking 166–175 amino acids (NSP4Δ10C) and 161–175 amino acids (NSP4Δ15C) were attempted (Fig. 5A). We examined both the unmodified method (using NSP2 and NSP5) and the improved method (using NSP2opt and NSP5opt). Although rSA11-NSP4Δ10 was successfully recovered using both methods, rSA11-NSP4Δ15 could only be recovered with the improved method. The presence of the truncated NSP4 genome in these recombinant viruses was confirmed by electrophoresis of the dsRNA genome (Fig. 5B). Replication of rSA11-NSP4Δ10C and rSA11-NSP4Δ15C was impaired in MA104 cells (Fig. 5C). The cell lysates of rSA11, rSA11-NSP4Δ10C, and rSA11-NSP4Δ15C were subjected to CsCl₂ density gradient centrifugation. Although rSA11 produced typical double bands corresponding to the upper triple-layered particle (TLP) and lower DLP, both NSP4 mutant viruses produced a single band corresponding to the DLP (Fig. 5D). Two fractions (F1 and F2) corresponding to the TLP and DLP were collected from each virus and dialyzed against virion storage buffer. The viral proteins in F1 and F2 were analyzed by SDS-PAGE and visualized using silver staining (Fig. 5E, upper panel). F1 and F2 of rSA11 showed typical patterns of TLP (VP1, VP2, VP3, VP5, VP6, and VP7) and DLP (VP1, VP2, VP3, and VP6), respectively. Similarly, F2 of rSA11-NSP4Δ10C and rSA11-NSP4Δ15C exhibited a typical DLP pattern. In the F1 of these mutant viruses, scarce bands of VP5 and VP7 were visible only in rSA11-NSP4Δ10C, and VP3 was not visible in either virus. The presence of VP6 and VP7 in F1, as visualized by silver staining, was confirmed using western blotting (Fig. 5E, lower panel). The interaction between NSP4 and VP6 via the DLP-binding domain was examined using a co-immunoprecipitation assay. Using the lysate of virus-infected cells, NSP4 was pulled down using a corresponding mouse monoclonal antibody. In the rSA11-infected cells, VP6 was co-precipitated with NSP4, indicating a robust association between VP6 and NSP4 (Fig. 5F). Interaction between NSP4Δ10 and VP6 was notably reduced, and no interaction was observed between NSP4Δ15 and VP6. These findings confirmed previous studies showing that VP6 and NSP4 interact via the DLP-binding domain at the C-terminal end and demonstrated that the interaction of VP6 and NSP4 is crucial for efficient viral replication.

Fig 5.

Generation of recombinant SA11 encoding mutant NSP4 lacking the DLP binding domain. (A) Schematic diagram of NSP4 domain organization. H1, H2, and H3 indicate hydrophobic domains. (B) Electropherotype of viral dsRNA genomes. NSP4Δ10 and NSP4Δ15 genes are indicated by arrows. (C) Replication kinetics of NSP4 mutant viruses. (D) Separation of TLP and DLP by CsCl₂ gradient ultracentrifugation. Fractions F1 and F2 were collected separately and dialyzed using a virus storage buffer. (E) Visualization of viral proteins involved in TLP and DLP by silver staining and western blotting. Fractions F1 and F2 for each virus were separated by SDS-PAGE. VP6 and VP7 proteins were detected by murine monoclonal antibodies corresponding to SA11 VP6 and VP7, respectively. (F) Evaluating the binding ability of NSP4 mutants to VP6 through co-immunoprecipitation assay. The lysates of virus-infected cells were mixed by monoclonal antibody against SA11 NSP4 covalently coupled with magnetizable polystyrene spherical beads. VP6, NSP4, and β-actin in the whole cell lysates were detected as internal controls. (G and H) Intracellular localization of NSP4 and NSP4Δ15C. MA104 cells were infected with rSA11 or rSA11-NSP4Δ15. At 10 h post-infection, NSP4 was visualized using rabbit antisera targeting the NSP4 peptide spanning amino acids 82–100 (G) or 142–160 (H) followed by goat anti-rabbit IgG CF594 conjugate. NSP5 and ER were visualized using mouse anti-NSP5 antiserum and mouse anti-KDEL antibody, respectively, followed by goat anti-mouse IgG CF488 conjugate. Virus titers at 72 h.p.i. were statistically analyzed by ANOVA. * P < 0.05. n.s. not significant. Scale bars: 5 µm.

The intracellular localization of NSP4 in virus-infected cells was examined by an immunofluorescent assay using two rabbit antisera recognizing NSP4 82–100 amino acid (aa) and NSP4 142–160 aa, and the co-localization of NSP4 and viroplasm or ER was examined. Antiserum against NSP4 82–100 aa revealed different intracellular localization patterns for NSP4 and NSP4Δ15C (Fig. 5G). At 10 h post-infection (h.p.i.), wild-type NSP4 was distributed around the viroplasm and partially co-localized with the ER. In contrast, in cells infected with rSA11-NSP4Δ15C, NSP4Δ15C colocalized with the viroplasm and did not overlap with the ER. The antiserum against NSP4 142–160 aa revealed a different localization pattern (Fig. 5H). In rSA11-infected cells, NSP4 was localized throughout the cytoplasm and co-localized with the ER but not with the viroplasm (Fig. 5H). In cells infected with rSA11-NSP4Δ15C, NSP4Δ15C co-localized with the ER along with wild-type NSP4. Although NSP4 formed filamentous structures, NSP4-Δ15C formed dot-like structures (Fig. 5H). The rabbit anti-NSP4 142–160 antiserum specifically recognized only SA11-infected cells, as revealed by the mouse anti-SA11 antiserum (Fig. S3A), ruling out the possibility of a non-specific reaction of the rabbit anti-NSP4 142–160 antiserum to the ER. The localization of NSP4 142–160 in the ER was further confirmed by its co-localization with the green fluorescent protein fused to the ER localization signal (Fig. S3B). These results indicated that deletion of the DLP-binding domain not only affected the interaction with VP6 but also the localization of NSP4.

DISCUSSION

Since the first report of entirely plasmid-based reverse genetics for RV, the rescue efficiency of recombinant viruses has been improved by the overexpression of NSP2 and NSP5 using the RNA capping enzyme of the African swine fever virus and interferon-suppressed MA104 cells (11–13). However, rescuing low-replicating viruses such as human RV clinical isolates, murine rotaviruses, and recombinant RV carrying heterologous VP4 and VP7 genes remains challenging (5, 11, 31). The current improved strategy of overexpressing nucleotide-modified NSP2 and NSP5 further increased rescue efficiency and enabled the successful recovery of specific strains (e.g., rRVA-U4 and rSA11-NSP4Δ15) that could not be rescued using the previous method. Based on previous studies, enhanced rescue efficiency was expected owing to the increased expression of NSP2 and NSP5. However, the expression level of NSP5opt was slightly lower than that of native NSP5, and the improvement in rescue efficiency was not associated with the protein expression levels. The reduced rescue efficiency by co-expression of deficient NSP5 clearly showed that the nucleotides of the NSP5 ORF negatively affected the reverse genetics experiment, whereas the adverse effects were mitigated in the deficient NSP5opt gene.

In viral replication, the viral genome not only serves as a messenger RNA for protein expression but also functions as a template for genome replication (32). The 3’ end serves as a promoter for the initiation of viral RNA synthesis and also comprises a translation enhancer sequence and conserved cis-regulatory elements (33, 34). Additionally, regions spanning tens to hundreds of bases at both the 5' and 3' ends, encompassing the UTRs and partial ORFs, are posited to harbor the genome packaging signal (32, 35). In the process of selective genome packaging in Reoviridae viruses, it has been suggested that the sequences at both ends, including the UTR and parts of the ORF of each gene segment, function as packaging signals (35–37). Furthermore, RV genomes are expected to interact with each other to form a complex that enables the packaging of the 11 gene segments into progeny virions (21). Therefore, it is possible that NSP2 and NSP5 genomes lacking UTRs could interfere with genome packaging through inappropriate interactions with other viral genomes, although this remains to be fully investigated. Even if correctly incorporated into viral particles, the absence of necessary 5' and 3' end sequences for genome replication could inhibit virus replication. The reduced inhibition in reverse genetics by dNSP5opt compared to dNSP5 may suggest that the modified NSP5opt sequence weakens its interaction with other viral genomes. The difference in the negative contributions of dNSP2 and dNSP5 in reverse genetics may reflect the unique roles of these gene segments in genome packaging. Our results demonstrated that the expression of dNSP5opt still interrupted the reverse genetics system and further nucleotide alteration could improve the efficiency of reverse genetics. Although the expression of whole dNSP5wt reduced the rescue efficiency, the expression of partial dNSP5wt expanding the 1–200^th, 201–398^th, 399–597^th, and 1–200^th/399–597^th nucleotides did not affect rescue efficiency. These results implied that the complex structure of NSP5 was associated with genome packaging. A study using the whole dNSP5 gene with partially altered dNSP5opt sequences may elucidate the critical region that adversely affects reverse genetics. Additionally, it remains to be determined whether the overexpression of dNSP5 inhibits only virus rescue in reverse genetics or also impairs virus replication during infection.

Improved reverse genetics enable the generation of molecular clones of RV clinical isolates. Concerns regarding research on RV clinical isolates are that (i) clinical samples contain quasi-species or may contain multiple strains and (ii) RV clinical strains can acclimate to cell lines during the process of viral propagation, eventually yielding cell-adapted viruses that might have different characteristics from those of circulating viruses. Because RV clinical strains do not propagate efficiently in cell cultures, we could not rescue molecular clones using a previously developed reverse genetics system (5). In this study, we rescued molecular clones of human and bovine RV clinical isolates using improved reverse genetics. These molecular clones should maintain native characteristics similar to those of endemic viruses and contribute to the study of viral pathogenicity, transmission, and antigenicity in future studies.

Bovine RV strains UK and NCDV and the porcine RV strain Gottfried have been widely used as representative strains of type A RVs (38–40). In addition to clinical strains, recombinant RV laboratory strains, including NCDV, UK, and Gottfried, were rescued. In the current rotavirus research using reverse genetics, simian RV SA11 has been used because of its high multiplication ability and efficiency in creating recombinant viruses. In addition, reverse genetics systems of human rotavirus strains have been established and used for vaccine development and infectivity research by examining gene reassortment between simian and human RV strains. The genotypes of bovine and porcine rotaviruses are thought to be different from those of simian and human RVs, and the functions of viral proteins, including VP4, VP7, and NSP1, are considered to differ among different RV strains (41–43).

Finally, we evaluated the DLP-binding domain of NSP4, which is considered critical for mature virion formation, using an improved reverse genetics system. Previous studies using recombinant NSP4 protein and gene silencing technology have shown that NSP4 interacts with the VP6 DLP-binding domain (161^st–175^th aa), promotes budding of DLP into the viroplasm-associated membrane, and is required for the maturation of viral particles (27, 28, 30, 44, 45). We confirmed that amino acids 161–175 of NSP4 are critical for its interaction with VP6. In a previous study, the knockdown of NSP4 abolished both DLP and TLP formation (44). Here, we observed that the deletion of the DLP-binding domain affected only TLP formation while DLP was produced properly.

Furthermore, two different antibodies targeting the NSP4 82–100 peptide and the NSP4 142–160 peptide revealed the pleiotropic distribution of NSP4. The former antibody revealed that wild-type NSP4 localized around the viroplasm, whereas NSP4Δ15 appeared to be stacked in the viroplasm. DLP forms within the viroplasm bud through the membrane surrounding the viroplasm via the interaction between VP6 on the surface of DLP and NSP4 (27, 28). Based on these results, we hypothesized that the lack of binding to VP6 mediates incorrect localization of NSP4Δ15. Notably, recombinant RV lacking the DLP-binding domain was still replication-competent, although the replication efficiency was ~100-fold lower than that of wild-type RV. This result indicates that DLP could bud into the viroplasm-associated membrane without NSP4 or the budding step could be skipped. Future studies to rescue recombinant RVs with further truncated NSP4 or RVs lacking whole NSP4 will help elucidate the role of NSP4 in virion maturation.

In conclusion, we developed an improved reverse genetics system for RV by overexpressing the nucleotide-modified NSP2 and NSP5. This new protocol enabled the establishment of molecular clones of RV clinical isolates and generated recombinant RV with a critical amino acid deletion in NSP4. Before the establishment of the reverse genetics system, many amino acid mutations and functional domains of rotavirus proteins associated with infection and replication were identified. In particular, recombinant viruses with critical amino acid mutations exhibit a significantly reduced replication ability, making recombinant virus production difficult. By utilizing an improved reverse genetics system, it is possible to create mutant viruses with functionally important amino acid changes, such as those in the DLP-binding domain of NSP4, greatly advancing molecular biological research on rotaviruses.

MATERIALS AND METHODS

Cells and viruses

Epithelial monkey kidney MA104 cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) (Nacalai Tesque) supplemented with 5% fetal bovine serum (FBS) (Gibco). Baby-hamster kidney BHK-T7 cells stably expressing T7 RNA polymerase were cultured in DMEM supplemented with 5% FBS. The viruses were propagated in MA104 cells cultured in DMEM supplemented with an antibiotic cocktail and 0.5 µg/mL trypsin (T0303; Sigma-Aldrich). Bovine RV strains NCDV and UK and the porcine RV strain Gottfried were kindly provided by Dr. Nobumichi Kobayashi (Sapporo Medical University School of Medicine). Virus titers were determined using a focus-forming assay in MA104 cells. Virus-infected cell foci were detected using a rabbit anti-SA11 NSP4 antibody and goat anti-rabbit IgG antibody CF488 conjugate (Nacalai Tesque Inc., Kyoto, Japan).

Antibody

Rabbit anti-NSP4 sera corresponding to synthetic SA11 NSP4 peptides spanning amino acid residues 158–171, 82–100, or 142–160 were obtained (Eurofins Genomics, Tokyo, Japan). A mouse monoclonal antibody against KDEL (clone 10C3; Enzo Life Science Inc. Farmingdale, NY, USA) was used. Mouse anti-NSP5 antibodies were prepared as described previously (2). Mouse monoclonal antibodies against SA11, VP6, and VP7 were prepared as described previously (13).

Generation of murine monoclonal antibody against NSP4

Purified recombinant NSP4 was generated using a Baculovirus vector to obtain a monoclonal antibody. A baculoviral vector encoding NSP4-V5-His was generated using the Bac-to-Bac Baculovirus Expression System (Invitrogen). Briefly, the ORF of the NSP4 gene was amplified by RT-PCR and inserted into the Baculovirus transfer plasmid. In addition, synthetic genes encoding the V5 peptide tag and 6 × His peptide tag were inserted at the 3’ end of the NSP4 gene to enable the expression of NSP4-V5-His fusion protein. Sf-9 insect cells were infected with the recombinant Baculovirus and lysed in 1% Triton-X100. Recombinant NSP4 was purified using nickel chelate resin (Roche, Basel, Switzerland) according to the manufacturer’s instructions. To obtain monoclonal antibodies against NSP4, the footpads of ICR mice (Japan CLEA) were immunized with purified NSP4 conjugated to a hydrogel adjuvant (InvivoGen). Immunization was repeated six times at 5–10-day intervals. Two days after the final immunization, popliteal lymph nodes were removed. Lymphocytes were fused with PAI myeloma cells using 50% polyethylene glycol (HybriMax; Sigma-Aldrich). Hybridoma cells were cultured in RPMI1640 supplemented with 10% FBS and 1 × HAT medium (Hybri-Max; Sigma Aldrich). The mAb-186 clone was used for the experiments.

Isolation of human and bovine RV clinical isolates

The human RV RVA-U4 strain was isolated from a patient in 2017 (5). Bovine RV strain KM3 was isolated from cattle on a farm in Japan in 2018. To isolate RVs from stool samples, a 10% fecal suspension was inoculated into MA104 cells. After adsorption, the cell culture medium was replaced with DMEM supplemented with an antibiotic cocktail and trypsin and incubated at 37°C for 7 days. Cell lysates containing infectious viruses were cleared by centrifugation for 1 min at 2,400 × g and amplified in MA104 cells.

Determination of whole genome sequences of human and animal RVs and plasmid construction

Both 5’ and 3’ end sequences of whole gene segments of NCDV, UK, Gottfried, and KM3 were determined by full-length amplification of the cDNAs (FLAC) method (26). Briefly, self-priming oligo DNA linker (5’-p-GACCTCTGAGGATTCTAAAC /iSp9/TCCAGTTTAGAATCC-3′) was ligated to viral dsRNA using a T4 RNA ligase (Thermo Fisher Scientific, Waltham, U.S.A.), and cDNA was synthesized using RevertAid (Thermo Fisher Scientific, MA, U.S.A.). Full-length viral cDNAs were amplified using a primer (5′-GAGTTAATTAAGCGGCCGCAGTTTAGAATCCTCAGAGGTC-3′), and 5’ and 3’ terminal sequences were determined by direct sequencing (Eurofin genomics Japan). Full-length cDNAs of the whole gene segments were amplified using specific primers and pT7 rescue plasmids.

Plasmid construction

Rescue plasmids for the RV SA11 strain (pT7-VP1SA11, pT7-VP2SA11, pT7-VP3SA11, pT7-VP4SA11, pT7-VP6SA11, pT7-VP7SA11, pT7-NSP1SA11, pT7-NSP2SA11, pT7-NSP3SA11, pT7-NSP4SA11, and pT7-NSP5SA11), Odelia strain (pT7-VP1Odelia, pT7-VP2Odelia, pT7-VP3Odelia, pT7-VP4Odelia, pT7-VP6Odelia, pT7-VP7Odelia, pT7-NSP1Odelia, pT7-NSP2Odelia, pT7-NSP3Odelia, pT7-NSP4Odelia, and pT7-NSP5Odelia) and RVA-U4 strain (pT7-VP1U4, pT7-VP2U4, pT7-VP3U4, pT7-VP4U4, pT7-VP6U4, pT7-VP7U4, pT7-NSP1U4, pT7-NSP2U4, pT7-NSP3U4, pT7-NSP4U4, and pT7-NSP5U4) were prepared as described previously (2, 5, 46). Rescue plasmids for the KM3, UK, NCDV, and Gottfried were constructed following the same procedure for those of SA11. pCAG-D1R, pCAG-D12L, pCAG-NSP2, and pCAG-NSP5 were prepared as described previously (2, 13). Codon-optimized NSP2opt (accession number LC831693) and NSP5opt (accession number LC831694) DNAs were synthesized by Eurofins Genomics (Tokyo Japan). Plasmid expression vectors encoding NSP2opt and NSP5opt were also constructed. Briefly, the genes encoding NSP2opt and NSP5opt were amplified by PCR and inserted into the EcoRI site of the pCAGGS2 vector using NEBuilder HiFi DNA Assembly (New England BioLabs) to create pCAG-NSP2opt and pCAG-NSP5opt. The stop codon (TAA) was introduced after the 12th nucleotide of NSP5 and NSP5opt using standard site-directed mutagenesis to create pCAG-dNSP5 and pCAG-dNSP5opt, respectively. A stop codon (TAA) was introduced after the 18th nucleotide of NSP2 and NSP2opt to generate pCAG-dNSP2 and pCAG-dNSP2opt, respectively. Plasmid vectors encoding the 1–200^th, 201–398^th, 399–601^st, and 1–200^th/399–601^st region of NSP5 and NSP5opt were developed based on pCAG-NSP5 and pCAG-NSP5opt, respectively.A truncated KM3 NSP1 gene was generated by deleting 201–1378 nt. KM3 NSP1-mScarlet gene was generated by adding NSP1 5’end (86nt) and 3’ end (200 nt) at the 5’ and 3’ ends of the mScarlet gene using NEBuilder® HiFi DNA Assembly Master Mix (New England Biolabs).

Western blotting

Plasmids encoding NSP2, NSP2opt, NSP5, and NSP5opt (0.1, 0.5, or 1.0 µg/well) were transfected into BHK-T7 cells using Lipofectamine 2000 (Thermo Fisher Scientific) according to the manufacturer’s instructions. At 24 h post-transfection, the cells were harvested and lysed in 1% NP40-PBS buffer supplemented with protease and phosphatase inhibitors (Roche). Cell lysates were centrifuged at 12,000 × g for 3 min at 4°C to remove cell debris. The supernatants were collected and subjected to SDS-PAGE on 5%–20% gradient polyacrylamide gel and transferred onto PVDF membranes (Millipore). NSP2 and NSP5 were detected using mouse anti-NSP2 or NSP5 antisera, respectively, followed by goat anti-mouse IgG antibodies. As loading control, β-actin was detected using a mouse anti-β-actin monoclonal antibody. Protein bands were visualized using Super Signal West Femt (Pierce) Quantification of Protein Expression. Protein band intensities were quantified using ImageJ software (NIH). The expression levels of NSP2 and NSP5 were normalized to a loading control (β-actin) and expressed as a ratio relative to the control.

Reverse genetics

Monolayers of BHK-T7 cells (1 × 10⁵ cells/well) in 12-well plates were prepared a day before the experiment. Eleven rescue plasmids were co-transfected with expression plasmids using 2 µL of TransIT-LT1 transfection reagent per microgram of plasmid DNA. For each experiment, the details of the rescue plasmids, expression plasmids, and the amount of each plasmid are described in the figures and figure legends. After 2 days of incubation, the cell culture medium was replaced with an FBS-free medium supplemented with trypsin (0.5 µg/mL). The cells were co-cultured with MA104 cells (2 × 10⁵ cells/well). For examining the time course of virus production, a small portion (50 µL) of culture medium was harvested daily. After incubation at various intervals, the cells were lysed by freeze-thawing. Infectious viral titers were examined using a focus-forming assay.

Electrophoresis of viral dsRNA genomes

Viral dsRNA was extracted from the virions and mixed with an equal volume of 2× sample buffer (125 mM Tris-HCl [pH 6.8], 10% 2-mercaptoethanol, 4% SDS, and 10% sucrose). The dsRNAs were separated on 10% precast polyacrylamide gels (Atto) and visualized using ethidium bromide staining.

Virus growth curve

Monolayers of MA104 cells were infected with recombinant rotaviruses at a multiplicity of infection (MOI) value of 0.001 FFU/cell and incubated for 1 h to allow for virus adsorption. Cells were washed twice and cultured in FBS-free DMEM supplemented with trypsin (0.5 µg/mL). Cells were incubated at 37°C for various intervals. Infectious virus titers in cell lysates were examined using a focus-forming assay.

Fluorescent imaging

MA104 cells were infected with rSA11 or rSA11-NSP4-Δ15C at an MOI of 0.2 FFU/cells and incubated in FBS-free DMEM supplemented with trypsin (0.5 µg/mL). At 10 h p. i., the cells were fixed with 3.8% formaldehyde. After cell membrane permeabilization with 0.1% Triton X-100, NSP4 was visualized using rabbit anti-NSP4_82-100 or rabbit anti-NSP4_142-160 followed by incubation with a goat anti-rabbit IgG CF594 conjugate (Nacalai Tesque). NSP5 and ER were visualized using mouse anti-NSP5 serum and mouse anti-KDEL monoclonal antibodies, respectively, followed by incubation with goat anti-mouse IgG CF488 conjugate (Nacalai Tesque).

Cesium chloride gradient ultracentrifugation

Monolayers of MA104 cells (10 cm dishes × 20) were infected with rSA11, rSA11-NSP4Δ10, and rSA11NSP4Δ15 at an MOI 0.01 FFU/cell and cultured in trypsin-supplemented DMEM. The cell lysates obtained after freeze-thaw cycles were cleared by centrifugation at 3,000 × g, and the supernatant was subjected to ultracentrifugation at 133,900 × g for 1 h at 12°C in a SW 32Ti rotor (Beckman Coulter). Pellets were resuspended in 200 µL of phosphate buffered saline, placed onto CsCl₂ gradients (40% wt/vol and 55% wt/vol), and centrifuged at 148,900 × g for 17 h at 12 ◦C in a Beckman SW 55 Ti rotor (Beckman Coulter). Fractions F1 and F2 were collected and dialyzed against virus storage buffer (20 mM Tris-HCL, 100 mM NaCl, 1 mM CaCl2). The purified fractions were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Viral proteins were visualized by silver staining and western blotting with monoclonal antibodies against VP6 and VP7.

Co-immunoprecipitation assay

Monolayers of MA104 cells (1.0 × 10⁷ cells) were infected with rSA11, rSA11-NSP4Δ10, or rSA11-NSP4Δ15 at an MOI of 1.0 FFU/cells and incubated in DMEM 5% FBS. At 10 h.p.i., the cells were lysed in Pierce^TM IP buffer (Thermo Fisher Scientific). Co-precipitation was performed using mouse monoclonal antibodies against NSP4 and Dynabeads (Life Technologies), according to the manufacturer’s protocol. After elution, NSP4 and VP6 were detected by mouse monoclonal antibodies to NSP4 and VP6, respectively. β-actin in the whole cell lysate was detected by murine monoclonal antibody as internal control.

Statistical analysis

Viral titers were analyzed by t-test, one-way ANOVA, Kruskal-Wallis test, or Mann–Whitney U test using GraphPad Prism (version 5.01). Values with P < 0.05 were considered statistically significant.

DATA AVAILABILITY

The viral genome sequences determined in this study were deposited in GenBank with the following accession numbers: NCDV (LC810897-LC810907), UK (LC810908-LC810908), Gottfried (LC810886-LC810896), and KM3 (LC810930-LC810940). The plasmids and antibodies used in this study are available upon request.

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/jvi.00996-24.

Supplemental figures. jvi.00996-24-s0001.pdf.

Figures S1 to S3.

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.