Vero E6 cells were cultured in Dulbecco’s modified eagle medium, supplemented with 5% fetal bovine serum (FBS) and 1% penicillin-streptomycin. Huh7 cells were cultured in Dulbecco’s modified eagle medium supplemented with 10% FBS and 1% kanamycin (Nakabayashi et al., 1982). BSRT7/5 cells, which stably express T7 RNA polymerase, were cultured in Glasgow minimum essential medium (MEM) supplemented with 10% FBS, 10% tryptose phosphate broth (TPB), 1X MEM amino acid solution, and 1 mg/mL geneticin (Buchholz et al., 1999). MRC-5 cells were maintained in Eagle’s MEM containing 10% FBS, MEM Non-Essential Amino Acids Solution, and 1% sodium pyruvate.

All viruses used in the study were derived from MP-12 stain of RVFV and generated by an established reverse genetics system (Ikegami et al., 2006). RVFV-rLuc carries the Renilla luciferase (rLuc) open reading frame (ORF) in place of the NSs ORF in the S segment. The rescued viruses were amplified once in Vero E6 cells, titrated by plaque assay, and used for virus infections. For virus infection, cells were inoculated with virus for 1 h at 37˚C and then washed three times with phosphate buffered saline (PBS) before fresh medium was added.

A recombinant PCR-based method was employed to generate the pProT7-S Δ19 construct, which expresses S Δ19 RNA carrying a 19nt deletion between nucleotide positions 1656 and 1675 in the 3’ NCR of antigenomic S RNA, by using the T7-driven plasmid pProT7-S encoding the antigenomic S RNA as a template. Sequencing of the S Δ19 plasmid construct confirmed the absence of unwanted mutations. The reverse genetics system to rescue RVFV S Δ19 was described as previously (Ikegami et al., 2006). Briefly, BSRT7/5 cells were co-transfected with T7-driven plasmids encoding L and N proteins, a pCAGGS-G plasmid encoding Gn/Gc proteins, and T7-driven viral RNA expression plasmids encoding L RNA, M RNA, and S Δ19 RNA. Culture supernatants from the transfected cells were collected at 5 days post-transfection (P0 samples) and inoculated into Vero E6 cells for amplification (P1 samples) and collected at 3 days post-inoculation (p.i.). RVFV Δ19 P1 samples were used for these studies and the 19 nt deletion was confirmed by sequencing of viral genomes obtained from P1 samples.

Plaque morphology of RVFV and RVFV Δ19 were determined in VeroE6 cells by plaque assay analysis used previously (Ikegami et al., 2006). Briefly, VeroE6 cells were inoculated with virus for 1 h at 37°C. After virus absorption, inoculum was removed and cells were overlaid with 1X modified eagle medium containing 0.6% Noble agar, 5% FBS, and 5% TPB. After 3 days of incubation, cells were stained with neutral red for 16 h at room temperature and plaques were visualized.

Huh7 cells were infected with RVFV or RVFV Δ19 at a multiplicity of infection (MOI) of 1.0 or 0.01 at 37°C for 1 h. After virus absorption, cells were washed three times with PBS and growth medium was added. For the growth curve at an MOI of 1.0, culture supernatants were harvested at 0 h, 8 h, 12 h, 16 h, and 20 h p.i. For the growth curve at an MOI of 0.01, culture supernatants were harvested every day up to 4 days p.i. Virus titers were measured by plaque assay. The points in the growth curves are shown as mean +/- standard deviation from three independent experiments.

Student’s t-test was used for growth curve analysis of RVFV and RVFV Δ19. Significant difference was defined as p < 0.05 (GraphPad Prism 9).

Culture supernatants harvested from virus-infected cells were purified as described previously (Terasaki et al., 2011). Briefly, supernatants were clarified by centrifugation at 3,000 rpm for 15 min. The clarified supernatant was layered on top of a discontinuous sucrose gradient and centrifuged for 3 h at 26,000 rpm at 4°C. The interface between 30 and 50% sucrose was collected and subjected to a second discontinuous sucrose gradient ultracentrifugation for 18 h at 26,000 rpm at 4°C. Virus was pelleted through a 20% sucrose cushion at 38,000 rpm for 2 h at 4°C. The pellets were suspended in Trizol reagent for RNA analysis.

Huh7 cells were inoculated with RVFV at an MOI 3. After 8 h p.i., cells were mock-irradiated or irradiated with 254-nm-wavelength UV light at 250 mJ/cm², followed by lysis of the UV-crosslinked cells using a high stringency SDS lysis buffer (0.5% SDS, 50 mM Tris-HCl [pH 6.8], 1 mM EDTA, 1 mM dithiothreitol [DTT]) and heating of the lysate to 65°C for 5 min. Radioimmunoprecipitation assay (RIPA) correction buffer (1.25% NP-40, 0.625% sodium deoxycholate, 62.5 mM Tris-HCl [pH 8.0], 2.25 mM EDTA, 187.5 mM NaCl) was then added to the lysates. Samples were passed through a QIAshredder spin column (Qiagen) twice to reduce viscosity. Lysates were then subjected to RNAse digestion using 30 µg RNAse A and 325 units of RNAse T1 for 3 h at 37°C. Total RNA was extracted with Trizol following the manufacturer’s instructions.

The protocol used by Sei and Conrad (2014) for the CLIP procedure was followed, with some modifications described previously (Tercero et al., 2021). RVFV-infected cells were irradiated with 254-nm-wavelength UV light at 250 mJ/cm², followed by lysis of the UV-crosslinked cells using a high-stringency SDS lysis buffer (0.5% SDS, 50 mM Tris-HCl [pH 6.8], 1 mM EDTA, 1 mM DTT) and heating of the lysate to 65°C for 5 min. The high stringency lysis buffer condition disrupts the interaction of Gn with N and L proteins, thereby eliminating the possibility of indirect co-immunoprecipitation of viral RNAs with Gn protein through its interaction with N protein and/or L protein (Tercero et al., 2021). Subsequently, RIPA correction buffer was added to the lysates. To reduce viscosity, samples were passed through a QIAshredder spin column twice. Lysates were then subjected to RNAse digestion for 30 min at room temperature using an RNAse A/T1 mixture containing 30 µg RNAse A and 325 units of RNAse T1.

For immunoprecipitation, lysates were precleared using Dynabeads protein G, conjugated with unrelated mouse anti-H2K antibody (Tercero et al., 2021), for 15 min at 4°C with rotation. Subsequently, the precleared lysates were subjected to immunoprecipitation analysis using Dynabeads protein G, conjugated with mouse monoclonal anti-Gn antibody (R1-4D4) (Tercero et al., 2021) or anti-H2K antibody for 16 h at 4°C. The immunoprecipitated samples were washed 5 times with RIPA buffer containing a higher concentration of NaCl (high-salt RIPA buffer; 1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS, 500 mM NaCl, 50 mM Tris-HCl [pH 8.0], 2 mM EDTA). Subsequently, the samples were resuspended in a high-salt RIPA buffer. For RNA extraction from intracellular and immunoprecipitated lysates, samples were resuspended in proteinase K solution (0.5 mg/ml of proteinase K, 0.5% SDS, 20 mM Tris-HCl [pH 7.5], 5 mM EDTA) and incubated at 37°C for 2 h. After proteinase K digestion, 3 M sodium acetate (NaOAc, pH 5.2) was added to the samples and RNA was extracted from the samples by phenol-chloroform/ethanol precipitation. The resulting RNA pellets were resuspended in sterile DNAse/RNAse-free H2O.

For preparation of RNA for CLIP-seq cDNA library construction, total RNA and immunoprecipitated RNA were subjected to dephosphorylation using 2 units of shrimp alkaline phosphatase for 30 min at 37°C and then 10 min at 65°C. After dephosphorylation, RNAs were phosphorylated using 10 units of T4 polynucleotide kinase for 30 min at 37°C and then 20 min at 65°C. After phosphorylation, RNAs were purified using miRNeasy Kit with RNeasy MiniElute Cleanup Kit (Qiagen) to obtain fragmented RNAs smaller than 200 nt. RNA was transferred to the Next Generation Sequencing Core facility at The University of Texas Medical Branch for adapter ligation and cDNA library construction. Library was sequenced using the NextSeq 550 Illumina system and bioinformatics analysis was performed by the Next Generation Sequencing Core facility. Reads were mapped to RVFV MP-12 genome (GenBank: DQ375404.1, DQ380208.1, and DQ380154.1) and visualized using Integrated Genome Viewer. Enriched regions (peaks) in the immunoprecipitated samples were normalized to enriched reads mapped to RVFV-infected total RNA.

Reverse transcription (RT) of viral RNAs was performed using a strand-specific assay as described previously (Tercero et al., 2019). Briefly, cDNA synthesis of viral RNAs was performed by using the SuperScript III first-strand synthesis system (Invitrogen). Viral RNA was mixed with a 2 μM concentration of a strand-specific RT primer. To detect the antigenomic S RNA of RVFV Δ19, we used a strand-specific tagged RT primer that binds upstream of the deletion region. The mixture was incubated at 65°C for 5 min and then cooled to 4°C. After addition of the reaction buffer and enzyme mixture, cDNA synthesis was carried out by incubating the sample at 50°C for 55 min, with termination by heating at 85°C for 5 min, cooling to 4°C, and then treatment with RNAse H at 37°C for 20 min. The cDNAs were purified using an on-column PCR purification kit (Qiagen). The strand-specific real-time qPCR assays were conducted using SsoAdvanced Universal SYBR Green Supermix (Bio-Rad) and followed the established protocol as described previously (Tercero et al., 2019). Purified cDNA was added to the PCR master mix containing the nonviral tagged sequence as a forward primer and a viral strand-specific reverse primer. Thermocycling conditions were as follows: 95°C for 30 sec, 35 cycles of 95°C for 15 s and 60°C for 20 s, followed by melting-curve analysis. The assays were performed by using the CFX96 Touch real-time PCR detection system and analyzed using the provided software (Bio-Rad CFX Maestro).

Total RNAs were extracted from virus-infected cells using Trizol reagent (Invitrogen) according to the manufacturer’s protocol. The same amounts of total RNAs were subjected to Northern blot analysis using digoxigenin (DIG)-labeled RNA probes that specifically bind to the genomic L, M, S segments, and NSs mRNA (Ikegami et al., 2005b). A DIG-labeled IFN-β riboprobe was used for IFN-β mRNA detection (Narayanan et al., 2008). The RNAs were visualized using the DIG luminescent detection kit (Roche Applied Science), and images were analyzed using AlphaEaseFC software.

Intracellular lysates were subjected to SDS-PAGE analysis using a 4 to 20% gel and transferred to a polyvinylidene difluoride (PVDF) membrane (Bio-Rad). For the detection of Gn, N and NSs proteins, mouse monoclonal anti-Gn antibody (R1-4D4), at a 1:3,000 dilution, rabbit polyclonal GST-N peptide antibody (R1-P1-GST-N) (Won et al., 2006), at a 1:3,000 dilution, and rabbit polyclonal NSs peptide antibody, at a 1:1,000 dilution (Ikegami et al., 2005b), were used as primary antibodies. A goat anti-mouse IgG HRP-linked antibody (Cell Signaling Technology), at a 1:10,000 dilution and goat anti-rabbit IgG HRP-linked antibody (Cell Signaling Technology), at a 1:10,000 dilution, were used as secondary antibodies, respectively. Chemiluminescence signals were detected with an ECL2 Western blotting kit (Thermo Scientific Pierce). Images were scanned, cropped, and assembled using AlphaEaseFC software.

Previously, we reported the presence of a direct interaction between the envelope glycoprotein, Gn, and RVFV RNAs (Tercero et al., 2021). Furthermore, we observed a positive correlation between the ability of Gn to bind viral RNAs and the packaging ability of RVFV RNA segments into virus particles (Tercero et al., 2021). To gain an insight into the mechanism of RNA packaging in RVFV, we sought to further characterize this interaction by identifying regions in viral RNAs that directly interact with Gn using CLIP-seq analysis. This analysis requires the fragmentation of viral RNAs in RVFV-infected cells by partial RNAse digestion to generate fragments with sizes ideal for subsequent high-throughput sequencing. However, a past study suggested that RVFV RNPs are resistant to RNAse treatment (Raymond et al., 2010), which would impact the feasibility of conducting CLIP-seq analysis. To determine the susceptibility of RVFV RNAs in RNP complexes to RNAse-digestion, we treated the lysate of RVFV-infected cells with a mixture of RNAse A/T1, extracted RNAs after the RNAse treatment, and analyzed the integrity of the genomic RNAs by Northern blot analysis. We detected a smear of small-sized viral RNA signals, but not full-length genomic RNAs, in the RNAse-treated samples ( Figure 2 ), demonstrating the susceptibility of viral RNAs, including those in the viral RNPs, to RNAse treatment.

We performed CLIP-seq analysis using two independent samples. Briefly, the UV-irradiated and RNAse-treated extracts from RVFV-infected cells were subjected to co-immunoprecipitation analysis using an unrelated anti-H2K antibody (Tercero et al., 2021) (control group) or anti-RVFV Gn antibody (experimental group). Fragmented total RNAs and co-immunoprecipitated RNAs were collected and prepared for cDNA library construction. The cDNAs were subjected to next generation sequencing followed by bioinformatics analysis. Subsequently, read coverage plots of the co-immunoprecipitated RNAs were constructed. The control group showed a low number of reads at negligible levels, whereas the experimental group showed a significant number of reads that mapped to viral RNAs. The read coverage plots of the genomic and antigenomic viral RNAs, which were normalized by subtracting the reads of intracellular viral RNAs (input control) from the reads of the experimental group, were visualized by Integrative Genomics Viewer and are shown in Figure 3A . The pronounced peaks represent the regions with high read density. The read coverage plots of two independent experiments were similar for most of the viral RNAs, demonstrating good reproducibility. The differences in the read coverage plots of antigenomic L RNA between the two independent experiments were larger than that for other viral RNAs, potentially due to low read coverage for this RNA. The proportion of reads that mapped to genomic L and M RNAs were substantially higher than that to their antigenomic counterparts. In addition, the number of reads that mapped to genomic and antigenomic S RNAs were similar, with a slightly higher number of reads mapping to antigenomic S RNA. These data correlated with our previous studies that demonstrated the efficient binding of Gn to the genomic L, M and S RNAs and antigenomic S RNA (Tercero et al., 2021). The presence of similar read coverage plots for a given viral RNA segment in two independent experiments strongly suggested that the prominent regions of sequencing reads in the plots represent sites in viral RNAs that preferentially interact with Gn. Hence, our CLIP-seq analysis indicated direct binding of Gn to many different regions of the genomic and antigenomic RVFV RNAs.

No prominent sequence coverage that aligns with the minimal NCRs of most of the viral RNAs were detected, except for the 3’ NCR of the antigenomic S RNA, which had a prominent number of sequencing reads (highlighted by a blue box in Figure 3A ). Figure 3B shows the sequence and read coverage plots of the 3’ terminal region of the antigenomic S RNA. Within the 34 nt-long 3’ NCR, the terminal 15 nt is essential for RNA replication and packaging of the genomic RNA of the S RNA-derived minigenome RNA (Murakami et al., 2012). The region corresponding to 5-25 nt from the 3’ NCR of the antigenomic S RNA showed a high read density in two independent experiments.

To know the biological significance of the prominent Gn-binding site within the 3’ NCR of antigenomic S RNA, we rescued a RVFV mutant (RVFV Δ19) that has a 19 nt deletion in the 3’-NCR of the antigenomic S RNA using a reverse genetics system (Ikegami et al., 2006) ( Figures 3B , 4A ). RVFV Δ19 retained the 3’ terminal 15 nt of NCR and lacked about 50% of the region that showed prominent number of reads within the 3’ NCR of antigenomic S RNA in CLIP-seq analysis. RVFV Δ19 formed smaller plaques than RVFV in Vero E6 cells ( Figure 4B ). RVFV Δ19 replicated to comparable titers to RVFV in Huh7 cells at a high MOI ( Figure 4C , left panel), but replicated less efficiently than RVFV at a low MOI ( Figure 4C , right panel).

As the 19 nt deletion is located upstream of the NSs mRNA start codon ( Figure 3B ), we examined whether NSs expression occurs in RVFV Δ19-infected cells. We used RVFV and RVFV-rLuc, the latter of which encodes rLuc gene in place of NSs gene ( Figure 4A ) (Ikegami et al., 2006), as a positive control and a negative control, respectively. Like RVFV-infected cells, accumulation of the genomic L, M and S RNAs and NSs mRNA occurred in RVFV Δ19-infected cells ( Figure 4D ). The absence of the genomic S RNA and NSs mRNA in RVFV-rLuc-infected cells was due to the use of the probe that binds to the NSs gene. Accumulation of Gn, N and NSs proteins also occurred in RVFV Δ19-infected cells ( Figure 4E ). As expected, accumulation of NSs did not occur in RVFV-rLuc-infected cells ( Figure 4E ). These data demonstrated that RVFV Δ19 did not replicate as efficiently as RVFV after low MOI infection in Huh7 cells, although NSs mRNA synthesis still occurred from the antigenomic S RNA of RVFV Δ19.

To determine if the 19 nt region in the 3’ NCR of antigenomic S RNA played a role in its efficient packaging into virus particles, we examined the proportion of the genomic and antigenomic viral RNAs in RVFV Δ19-infected cells and in purified RVFV Δ19 particles and compared the data to that obtained from RVFV. As reported previously (Tercero et al., 2021), the proportion of the antigenomic S RNA in RVFV virions was markedly higher relative to its abundance in infected cells and the relative proportion of antigenomic S RNA in virus particles was prominently higher than that of antigenomic L and M RNAs ( Figure 5 ). In contrast, the proportion of the antigenomic S RNA in RVFV Δ19 virions relative to its abundance in infected cells was markedly lower than that in RVFV-infected cells, demonstrating that antigenomic S RNA is inefficiently packaged into RVFV Δ19. These data showed that the prominent Gn-binding site in the 3’ NCR of antigenomic S RNA plays a critical role in the efficient packaging of antigenomic S RNA into virus particles.

Our past study indicated that the incoming antigenomic S RNA in RVFV particles serves as the template for NSs mRNA synthesis, leading to the synthesis of NSs protein immediately after infection (Ikegami et al., 2005b). Early expression of NSs after infection would be important for robust RVFV replication, as the incoming viral RNPs can trigger type I IFN induction (Weber et al., 2013) and NSs suppresses host innate immune responses, including type I IFN induction by inhibiting IFN-β mRNA transcription (Le May et al., 2008). To further establish the importance of the efficient packaging of antigenomic S RNA for the suppression of the early induction of type I IFN after infection, we infected type I IFN-competent MRC-5 cells (Terasaki et al., 2011) with RVFV, RVFV Δ19 or RVFV-rLuc and examined the induction of IFN-β mRNA from 3 to 5 h p.i. ( Figure 6 ). RVFV replication did not induce IFN-β mRNA, whereas RVFV-rLuc replication induced high levels of IFN-β mRNA, as reported previously (Terasaki et al., 2011). Notably, we detected IFN-β mRNA, the amount of which was highest at 3 h p.i., in RVFV Δ19-infected cells ( Figure 6 ). Northern blot analysis using an RNA probe that binds to the NSs gene showed the accumulation of NSs mRNA in RVFV Δ19-infected cells and in RVFV-infected cells during the early timepoints after infection ( Figure 6 ). Specifically, at 3 h p.i. the amount of NSs mRNA was lower in RVFV Δ19-infected cells compared to that in RVFV-infected cells, while the amount of genomic S RNA was comparable between these cells ( Figure 6 ). It is unlikely that the lower accumulation of NSs mRNA in RVFV Δ19-infected cells at 3 h p.i. is due to the reduced transcriptional activity from the antigenomic S RNA of RVFV Δ19, as the relative ratio of NSs mRNA to the genomic S RNA in RVFV Δ19-infected cells and in RVFV-infected cells were similar at 5 h p.i. ( Figure 6 ). Rather, these data suggest that the presence of lower amounts of incoming antigenomic S RNA, which serves as the template of NSs mRNA, resulted in the less efficient accumulation of NSs mRNA immediately after RVFV Δ19 infection. Taken together, these data indicate that the inefficient packaging of antigenomic S RNA into RVFV Δ19 particles led to a decreased accumulation of NSs mRNA during the early timepoints after infection with RVFV Δ19. In RVFV-rLuc-infected cells, IFN-β mRNA expression increased between 3 and 5 h p.i. compared to RVFV-infected cells ( Figure 6 ), whereas the IFN-β mRNA expression decreased at 4 and 5 h p.i. in RVFV Δ19-infected cells. These findings suggest that the NSs expression increased by 5 h p.i. in RVFV Δ19-infected cells ( Figure 6 ), due to the subsequent accumulation of NSs mRNA following the replication of S RNA, leading to the suppression of IFN-β mRNA synthesis later in infection.