Expression of a viral polymerase-bound host factor turns human cell lines permissive to a plant- and insect-infecting virus

Abstract

Tospoviruses are the only plant-infecting members of the Bunyaviridae family of ambisense ssRNA viruses. Tomato spotted wilt tospovirus (TSWV), the type-member, also causes mild infection on its main insect vector, Frankliniella occidentalis. Herein, we identified an F. occidentalis putative transcription factor (FoTF) that binds to the TSWV RNA-dependent RNA polymerase and to viral RNA. Using in vitro RNA synthesis assays, we show that addition of purified FoTF improves viral replication, but not transcription. Expression of FoTF deletion mutants, unable to bind the RNA-dependent RNA polymerase or viral RNA, blocks TSWV replication in F. occidentalis cells. Finally, expression of FoTF wild-type turns human cell lines permissive to TSWV replication. These data indicate that FoTF is a host factor required for TSWV replication in vitro and in vivo, provide an experimental system that could be used to compare molecular defense mechanisms in plant, insect, and human cells against the same pathogen (TSWV), and could lead to a better understanding of evolutionary processes of ambisense RNA viruses.

Tospoviruses are the only plant-infecting members of the Bunyaviridae family of ambisense ssRNA viruses, which also encompasses hundreds of animal and human viruses (1). Tomato spotted wilt tospovirus (TSWV), the type-member, is one of the most important plant pathogens worldwide (2). TSWV is transmitted by and replicates in the insect vector Frankliniella occidentalis (3), in which it induces a strong immune response (4).

The TSWV particle contains three ssRNAs (L, M, and S) encapsidated by the N protein and associated with the RNA-dependent RNA polymerase (RdRp). The 331-kDa TSWV RdRp contains conserved RdRp motifs and possesses polymerase activity in vitro (5–7). RdRps of ambisense viruses are required to perform replication, transcription, “cap-snatching,” and genomic strand selection. Messenger RNAs are transcribed from either genomic or viral-complementary RNA segments. To initiate transcription, the RdRp binds and cleaves 5′ methylguanosine “cap” structures from host mRNAs and uses them to prime viral transcription. Therefore, these RdRps possess at least polymerase, transcriptase, and endonuclease activities (5–10). This multitude of enzymatic activities is thought to require the association of the RdRp to host factors (11–17). Nevertheless, RdRp-bound host factors have not been reported for tospoviruses or bunyaviruses.

We screened an F. occidentalis cDNA library (18) with two TSWV RdRp fragments to identify RdRp-associated host factors. Herein, we show that TSWV RdRp binds to a putative transcription factor from F. occidentalis, named F. occidentalis TSWV polymerase-bound factor (FoTF), which also binds to viral RNA and improves TSWV RNA replication in vitro. Expression of FoTF mutants that do not bind to RdRp or viral RNA blocks TSWV replication in F. occidentalis cells. We also found that FoTF overexpression in HeLa and human diploid fibroblast cell lines turns these human cells permissive to TSWV replication.

Materials and Methods

Virus Strains and Insect Populations. TSWV BR-01 and Hawaiian strains (kindly donated by T. L. German, University of Wisconsin, Madison) were maintained in Datura stramonium, which was used for virus purification (18, 19). Two independent F. occidentalis populations, of Brazilian and Californian origins (gift from T. L. German), were maintained at 26°C as described (18).

Human and Insect Cell Cultures. An F. occidentalis cell culture (FO1) was established as described (20). F. occidentalis eggs were incubated for 24 h at 27°C, washed in 70% ethanol, triturated in PBS, incubated with 1 μg/ml nystatin (Sigma) for 5 min, washed, and maintained in Sf-900 serum-free insect medium (GIBCO/BRL) with antibiotic/antimycotic mixture (GIBCO/BRL) at 27°C for 4 days. Purified TSWV (30 μg/ml total protein concentration, Bradford assay, Bio-Rad) was filtrated in 0.22-mm filters (Millipore) and added to FO1 cells for 2 h at room temperature, washed in PBS, and incubated at 27°C. HeLa and human diploid fibroblast cell lines (American Type Culture Collection) were maintained in serum-free DMEM (GIBCO) at 37°C. DNA transfections and TSWV infection of human cells were performed by using Lipofectamine 2000 (Invitrogen), following the manufacturer's instructions. DNA transfections were performed 12 h before virus inoculation.

Recombinant DNA. FoTF cDNA was subcloned (21) from pBluescript II SK+ (pBS, Stratagene) into pAc5.1 (Drosophila melanogaster actin 5C promoter, Invitrogen) by EcoRI/XhoI-directed ligation and into pCMS-EGFP (CMV immediate early promoter, mammalian cells expression, Clontech) by XbaI/NotI-directed ligation of a PCR fragment (newly introduced XbaI/NotI sites at the 5′ and 3′ ends of the FoTF sequence, primers: 5′-GCGCGCTCTAGAATGCAACCACAGCATGTA-3′ and 5′-CGCGGCGGCCGCTCATTGCGAGCTGCTTTG-3′, sites are in italics). FoTF was also cloned into pET101/D-TOSPO (Escherichia coli DH5α expression, Invitrogen) by blunt-ended ligation of a PCR fragment (primers 5′-CACCATGCAACCACAGCATGTA-3′ and 5′-TTGCGAGCTGCTTTGTTTTTTTGTTTCG-3′), following the manufacturer's instructions, resulting in a C-terminal His-tagged FoTF that was used for nickel chromatography purification (Qiagen, Valencia, CA). FoTF was also fused with GST by cloning it into pGEX-6P-1 (Amersham Pharmacia) by EcoRI/XhoI-directed ligation (5′-GST tag). 5′-GST-FoTFΔC was obtained by deleting the FoTF C-terminal half as follows. An EcoRI/XhoI-flanked PCR fragment from the pBS-FoTF clone was generated (primers 5′-GCGCGAATTCATGCAACCACAGCATGTAGGATC-3′ and 5′-CCGCCCTCGAGCTTCCAACAGGTAGATG-3′) and ligated into pGEX-6P-1, resulting in a 5′-tagged GST fusion containing the first 156 FoTF amino acids. 5′-GST-FoFTΔN was obtained deleting the N-terminal half of the protein (primers 5′-GCGCGAATTCTACATGAAGACCTGCGTG-3′ and 5′-CCGCCCTCGAG TCATTGCGAGCTGCTTTG-3′). FoFTΔN contains the last 157 aa of FoTF. 3′-His-tagged FoTF deletion mutants were obtained the same way, by using pET101. TSWV RdRp clones were kindly donated by T. L. German (University of Wisconsin, Madison) and have been described (7).

cDNA Library Construction and Screening. A λ-based F. occidentalis cDNA expression library was obtained (18) by using λ-ZAP II (Stratagene). The construction of a TSWV cDNA library and far-Western screenings were performed as described (18). Nitrocellulose membranes (Amersham Pharmacia) were incubated with 10 μg/ml purified TL908, TL2249, or FoTF, washed in TBS, incubated with anti-TL908, anti-TL2249, or anti-FoTF Abs, washed and incubated with HR peroxidase-conjugated (Sigma) secondary Ab, and followed by ECL detection (Amersham Pharmacia). Recombinant plasmids were excised as described (λ-ZAP II protocol).

Immunobloting and Immunoprecipitation. Western blotting and coimmunoprecipitation assays were performed as described (22). Anti-FoTF, TL908, TL2249, G1, and N Abs were obtained as described (7).

Yeast Two-Hybrid Assay. Saccharomyces cerevisiae Y190 (GAL4Δ system) was used for yeast transformation with LiOAc/PEG (Clontech). Cells were plated in Leu/Trp/His-lacking medium, with 25 mM 3-aminotriazole (Sigma). RdRp fragments, FoTF, and N were cloned into pGADT7 and pGBKT7 and tested for interactions by the X-Gal filter staining assay (23). Empty pGADT7 and pGBKT7 vectors and single transformations were used as negative controls; pLC1 (which drives expression of GAL4) was used as a positive control.

GST Pull-Down Assay. Glutathione-Sepharose beads (Amersham Pharmacia) were incubated with 100 μg of purified GST fusions (1:1 vol), incubated with TSWV-infected F. occidentalis larval lysates (22) for 4 h at 4°C, washed in PBS, and subjected to SDS/PAGE and Western blotting.

In Vitro RNA Synthesis. An RNA synthesis assay was performed as described (6, 24). Purified TSWV (in 100 mM Tris·HCl, pH 8.0/5 mM MnCl₂/2.5 mM MgCl₂/0.5% Triton X-100/2.5 mM DTT) was added to 500 μM each of rUTP, rGTP, and rATP, 0.1 μM rCTP, and 100 μM rCTP³², 3,000 Ci (1 Ci = 37 GBq) at 30°C for 120 min. Rabbit reticulocyte lysate (AP-Biotech) contained 4 mM Mg acetate, 1 mM each of NTP, 0.1% Nonidet P-40, 0.8 units/μl RNAsin, and 60 ng/μl tRNAs. RNA products were analyzed with Storm860 PhosphorImager (Amersham Pharmacia. Gels were allowed to decay and Northern analysis was performed.

Northern Analysis. Five micrograms of RNA were loaded in 0.8% agarose formaldehyde gel and transferred to nylon membranes (Amersham Pharmacia). Riboprobes (1–6 × 10⁷ cpm/μg) were synthesized by run-off transcription with T3 or T7 polymerase (Riboprobe system, Promega, and ref. 21). Strand-specific probes were prepared as follows. Recombinant pBluescript SK(+) plasmids in which were cloned the viral complementary (vc)-L RNA, nucleotides 100-1700, the vc-M RNA, nucleotides 100–1000, and the vc-S RNA, nucleotides 1300–2700, were linearized and used for run-off transcription. Cloning was performed with λ-ZAP II (Stratagene). Actin clones were obtained by RT-PCR (21), using published sequences from the human or Drosophila GenBank databases. Hybridizations were performed as described (4) and analyzed with a PhosphorImager.

Gel Shift Assays. Gel shift assays were performed as described (25). PCR products of 198 bp from the 5′ and 3′ ends of TSWV S RNA and a T7 promoter were in vitro-transcribed (T7 polymerase, Roche Molecular Biochemicals), using primers that were described (25), and [α-³²P]UTP-purified (G-50 columns, Pharmacia), incubated with purified proteins, separated on a 5% polyacrylamide gel, and analyzed with a PhosphorImager. C-terminal His-tagged N protein (as a positive control; ref. 26) was obtained as described above for FoTF (see Recombinant DNA). BSA and nonlabeled TSWV RNA were used as negative controls. All proteins were used from 0- to 20-μM concentrations.

Results

TSWV RdRp Binds to F. occidentalis FoTF Factor. To isolate cellular factors that associate with the TSWV RdRp, an F. occidentalis cDNA library was screened by a far-Western method (18) with two RdRp fragments, TL908 and TL2249, that were previously obtained (ref. 7 and Fig. 1A). The TL908 fragment contains the polymerase domain (5) and TL2249 contains the C terminus domain associated with transcription initiation in bunyaviruses (Fig. 1 A and ref. 1). No positively selected clones could be obtained with the TL908 fragment. Sequence analysis of the clones obtained in the TL2249 screening showed that a single RdRp-bound protein was selected. It was named FoTF (GenBank accession no. AY745196). FoTF's ORF contains 942 bases and encodes a protein of 313 aa. A rabbit polyclonal Ab raised against FoTF recognized a 40-kDa protein in F. occidentalis whole-cell lysates (Fig. 1B, FoTF blot). In vitro translation of full-length FoTF cDNA also produced a protein of 40 kDa (data not shown). Anti-FoTF did not recognize any protein from Datura stramonium or tomato (plant hosts of TSWV) or from Spodoptera frugiperda and Drosophila melanogaster (nonvectors of TSWV) samples (data not shown). Sequence analysis by using the blast program (which can be accessed at www.ncbi.nlm.nih.gov) showed that FoTF does not share significant homology with known proteins (data not shown). However, protein sequence analysis performed with the motifs software (which can be accessed at www.microbiology.adelaide.edu.au/links) by using the transfac, prosite, and blocks databases showed that FoTF contains several transcription factor's motifs, and that FoTF produced significant alignments with several transcription factors in the regions of the conserved motifs (see Fig. 6, which is published as supporting information on the PNAS web site), indicating that FoTF is a putative transcription factor.

Fig. 1.

F. occidentalis FoTF binds to the TSWV RdRp. (A) TSWV RdRp fragments used in this study, respective domains, and their respective position in relation to the whole RdRp are shown. (B) Immunoprecipitation (IP) and Western blotting (WB) using anti-TL2249, anti-FoTF, and anti-N Abs. Abs were incubated with whole-cell lysates from TSWV-infected F. occidentalis larvae. (C) WB after GST pulldown with pGEX empty vector, pGEX-TL2249, pGEX-FoTF, pGEX-FoTFΔN, and pGEX-FoTFΔC. GST beads were applied to whole-cell lysates of TSWV-infected F. occidentalis larvae. (D) Yeast two-hybrid X-Gal filter staining assay after cotransformation with TSWV RdRp fragments (pTL1, pTL1797, pTL2249, and pTL908) or TSWV N (pN) with F. occidentalis FoTF (pFoTF); yeast growth (and β-gal expression) indicates direct interaction; negative controls were transformations with pFoTF and pTL2249 only; positive control was pLC1 (which codes for GAL4, and triggers β-gal expression without the need of any interaction). (E) FoTF map and position of deletion mutants used in this study.

RdRp-FoTF-specific interaction was confirmed by coimmunoprecipitation assays by using TSWV-infected F. occidentalis larvae lysates. FoTF was coimmunoprecipitated with anti-TL2249 Ab, and vice versa, but it was not coimmunoprecipitated with anti-N Ab (Fig. 1B). A GST fusion protein pulldown assay showed that a mutant of F. occidentalis FoTF lacking its N-terminal half (GST-FoTFΔN) retained the ability to bind RdRp, but a C-terminal half-deleted mutant (GST-FoTFΔC) was not able to bind RdRp (Fig. 1C). To further investigate the FoTF interaction with RdRp and possible interaction with other viral proteins, a TSWV cDNA library was screened with purified FoTF, and a yeast two-hybrid binding assay was performed. After TSWV library screening with FoTF, only RdRp clones were selected (data not shown), confirming the data obtained with the F. occidentalis library screening and the GST fusion protein results (Fig. 1C). Yeast growth on selective media, indicating direct protein–protein interaction, was only observed after cotransformation of yeast cells with FoTF and TL2249, but not after cotransformation with other RdRp fragments, or with the N protein (Fig. 1D). Taken together, these results indicate that the C-terminal domain of the TSWV RdRp interacts directly and specifically with the C-terminal half of the F. occidentalis FoTF.

FoTF also Binds to Viral RNA. Because of the similarity with transcription factors and the presence of nucleic acid binding domains in FoTF (see Fig. 6), we then investigated the FoTF's RNA-binding capability. Addition of purified FoTF (0–20 μM) resulted in a crescent shift of radiolabeled TSWV RNA samples (Fig. 2A). TSWV N protein has been shown to bind TSWV RNA and was used as a positive control (ref. 26 and Fig. 2B). A FoTF mutant lacking the N-terminal half of the protein (FoTFΔN) did not bind to TSWV RNA (Fig. 2C). BSA and the addition of nonlabeled TSWV RNA to FoTF also did not result in a shift of the RNA bands (Fig. 2 D and E). These results indicate that FoTF binds specifically to TSWV RNA by means of its N-terminal domain.

Fig. 2.

RNA gel shift assay. Autoradiograph of RNA-binding reactions after addition of radiolabeled TSWV S RNA transcripts to purified FoTF only (A), TSWV N protein as positive control (B), FoTFΔN(C), BSA as a negative control (D), or FoTF plus nonlabeled TSWV S RNA (E). Protein concentrations varied from 0 to 20 μM; –, no protein added.

FoTF Improves TSWV RNA Replication in Vitro in the Presence of Reticulocyte Lysate. To address the role of FoTF in TSWV RdRp activity, we amended standard TSWV RNA synthesis in vitro assays with purified FoTF and reticulocyte lysate. Purified TSWV particles show only limited in vitro RdRp activity (6, 24). Addition of heat-inactivated TSWV particles does not result in de novo RNA synthesis (Fig. 3A, lane 1). Addition of purified TSWV only results in a 3-kb band and smaller RNA species (Fig. 3A, lane 2, and refs. 6 and 24). After the decaying of radioactivity in the RNA synthesis gel (Fig. 3A), the samples were transferred to nylon membranes and subjected to Northern analysis. First, it was observed that the 3-kb band hybridizes with a strand-specific probe for vc-S RNA (Fig. 3B and ref. 24), indicating that TSWV S RNA replication occurs. Addition of reticulocyte lysate improves the total amount of RNA synthesized, but no products >3 kb can be visualized (Fig. 3A, lane 3, and ref. 24). The addition of FoTF alone showed no difference to products from a TSWV-only reaction (Fig. 3A, compare lanes 2 and 4). However, combining FoTF and reticulocyte lysate resulted in a significant improvement in RNA synthesis, with RNA products of ≈3, 5, and 9 kb (Fig. 3A, lane 5), which is similar to the sizes of genomic TSWV RNA (genomic S RNA is 2.9 kb, M RNA is 4.8 kb, and L RNA is 8.9 kb). Northern analysis showed that those bands hybridized to vc-S RNA (Fig. 3B, lane 5), vc-M RNA (Fig. 3C, lane 5), and vc-L RNAs (Fig. 3D, lane 5), indicating that replication of all three genomic RNAs occurred when FoTF plus reticulocyte lysate amendment was used. Preincubation of FoTF with anti-FoTF Ab blocked that stimulatory effect (Fig. 3A, lane 6), indicating that this effect was specific to FoTF. Fig. 3E (lanes 3, 5, 6, and 8) shows that after amendment with reticulocyte lysate an ≈1.2-kb RNA species reacted with a strand-specific probe for the N gene, a size that represents N mRNA, which was also present after FoTF amendment (Fig. 3E, lane 5). However, TSWV nonstructural protein S (NSs) mRNA (1.7 kb) could not be detected in any of the tested conditions (Fig. 3F). Preincubation with anti-FoTF had no effect in the detection of N mRNA (Fig. 3E, lane 6). Taken together, these data indicate that FoTF improved TSWV replication but had no effect over transcription. As controls, the addition of DNase-free RNase (Fig. 3A, lane 7) showed only background levels of RNA products and addition of actinomycin D had no effect on synthesis (Fig. 3A, lane 8), indicating that the products were in fact RNA molecules and that their synthesis was independent of a DNA template.

Fig. 3.

In vitro RNA synthesis assay. (A) Gel electrophoresis of radiolabeled TSWV RNA products after amendments to the standard reaction, as indicated in the table. After decay of the radioactivity, gels as shown in A were submitted for Northern hybridization by using probes specific to vc-S RNA (B), vc-M RNA (C), and vc-L RNA (D). (E) Probe specific to the N gene, the N mRNA, is visualized, and can be differentiated by the size (1.2 kb). (F) Probe specific to the NSs gene, the NSs mRNA, could be visualized because it could be differentiated by the size (1.7 kb); h-i, heat-inactivated virus (lane 1); a molecular weight marker (lane M) shows relative sizes of detected RNA species in A, the sizes, in kb, are indicated in B–F. Results shown are representative of six independent experiments.

FoTFΔC and TL2249 Block the Infection of F. occidentalis Cells. To address the relevance of the RdRp–FoTF interaction in vivo, we established an F. occidentalis cell line (FO1) that could be infected with TSWV as described (20). Virus infection was measured by Northern hybridization, using a TSWV nonstructural protein M (NSm)-specific probe because NSm is a nonstructural protein, and is only present in infected cells (1). The NSm probe detects both NSm mRNA and genomic M RNA, but the smaller size of NSm mRNA can be used to differentiate the two RNA species (1.2 versus 4.8 kb). Fig. 4 A and B show that NSm mRNA was detected in FO1 cells 12 h post inoculation (hpi) and that NSm mRNA levels increase over time, as reported (20, 27). Fig. 4A is the graphic representation of the data showed in Fig. 4B and two additional experiments. NSm mRNA signals were not detected in TSWV-inoculated Spodoptera frugiperda SF21 cells, as expected, because S. frugiperda is not a TSWV host (ref. 3 and data not shown). Overexpression of FoTF in FO1 cells by Lipofectin-mediated transfection, before TSWV inoculation, had no effect in NSm mRNA levels (Fig. 4A), indicating that FoTF is not a limiting factor in FO1 cells. Next, a FoTF mutant (FoTFΔC, with its C-terminal half deleted, which does not bind to RdRp; as shown in Fig. 1C was tested. FO1 cells expressing FoTFΔC (Fig. 4 C and D) or FoTFΔN (data not shown, which does not bind viral RNA, as shown in Fig. 2C) showed a significant decrease in NSm mRNA levels over time, suggesting that the RNA–FoTF–RdRp interaction is required for TSWV replication in vivo. Expression of the TL2249 fragment also led to a decrease in NSm mRNA levels (Fig. 4 C and D), suggesting that the RdRp TL2249 fragment competed with the wild-type TSWV RdRp for viral RNA. Detection of actin mRNA was used as loading control (Fig. 4 B and D).

Fig. 4.

TSWV infection of F. occidentalis FO1 cells. (A) Graphic representation of NSm mRNA accumulation over time, after TSWV infection, of nontransfected (solid line and squares) and FoTF-transfected FO1 (dashed lines and diamonds). (B) Northern hybridizations showing NSm mRNA accumulation over time of nontransfected and FoTF-transfected FO1, as indicated, and actin mRNA as the loading control from one of the blots. (C) Graphic representation of NSm mRNA accumulation over time of FoTF-, FoTFΔC-, and TL2249-transfected FO1. (D) Northern hybridizations showing NSm mRNA accumulation over time of FoTFΔC- and TL2249-transfected FO1, as indicated, and actin mRNA as the loading control from one of the blots. Results shown are representative of three independent experiments. The highest RNA signal of all three experiments, as measured by the Storm860 PhosphorImager, was set to represent the 100% level (total) of NSm mRNA, and percentages were extracted from that total at each time point. Average and SD numbers were calculated from the data collected in three experiments.

FoTF Expression Turns Human Cells Permissive to TSWV Infection. Given that tospoviruses belong to a viral family (Bunyaviridae) mostly comprised of human and animal viruses (1), we next investigated whether TSWV could replicate in human cells. Compared with TSWV infection in FO1 cells, only very small amounts of NSm RNA could be detected after Lipofectin-mediated infection of TSWV into HeLa cells (Fig. 5A) and human diploid fibroblast cells (data not shown). No NSm mRNA signals were detected in those cell lines, even at 72 hpi (data not shown), indicating that these cells are nonpermissive to TSWV. HeLa cells were then transfected with FoTF and infected 24 h later with TSWV. FoTF expression was detected 12 h after transfection (Fig. 5E). Interestingly, NSm mRNA accumulation was detected 12 h after TSWV infection of HeLa cells expressing FoTF and increased over time (Fig. 5 A and B), indicating that FoTF expression turns HeLa cells permissive to TSWV. At 48 hpi, NSm mRNA levels in HeLa cells reached ≈80% of the average signal detected in FO1 cells inoculated in parallel (Fig. 5A). The same results were obtained with human fibroblasts (data not shown). In addition, and as expected, transfection with FoTFΔC or FoTFΔN did not turn HeLa cells permissive to TSWV, as measured by the absence of NSm mRNA accumulation (Fig. 5 C and D). To add further evidence to the permissiveness of FoTF-expressing HeLa cells to TSWV, Western blots showing the nucleocapsid (N) protein accumulation were also obtained (Fig. 5F). Yet again, N protein only accumulated in FoTF-expressing HeLa cells, but not in nontransfected HeLa cells or FoTFΔC-expressing HeLa cells (Fig. 5F).

Fig. 5.

TSWV replication in HeLa cells. (A) Graphic representation of NSm RNA accumulation over time of TSWV-transfected HeLa cells, previously nontransfected (dashed line and squares) or FoTF-transfected (dashed line and triangles), compared with TSWV-infected/FoTF-non-transfected FO1 cells (solid line and diamonds). (B) Northern hybridization showing NSm RNA accumulation over time of TSWV-transfected HeLa cells previously nontransfected or FoTF-transfected, as indicated; actin mRNA is shown as a loading control from one of the blots. (C) Graphic representation of NSm RNA accumulation over time of TSWV-transfected HeLa cells, previously FoTF-transfected (solid line and diamonds), FoTFΔC- (dashed line and triangles), or FoTFΔN-transfected (dashed line and squares). (D) Northern hybridization showing NSm RNA accumulation over time of TSWV-transfected HeLa cells previously FoTFΔC- or FoTFΔN-transfected, as indicated; actin mRNA is shown as a loading control from one of the blots. (E) Western blotting of HeLa cells with anti-FoTF with nontransfected or FoTF-transfected samples, as indicated; samples collected at the indicated time points after virus transfection. (F) Western blotting of HeLa cells with anti-N with nontransfected, FoTF-, or FoTFΔC-transfected samples, as indicated; samples collected at the indicated time points after virus transfection. Results shown are representative of three independent experiments. The highest RNA signal of all three experiments, measured by the Storm860 PhosphorImager, was set to represent the 100% level (total) of NSm mRNA, and percentages were extracted from that total at each time point. Average and SD numbers were calculated from the data collected in three experiments.

Discussion

The data shown here indicate that a putative transcription factor (FoTF) isolated from F. occidentalis, the main insect vector and a secondary host of TSWV, binds to the C-terminal domain of the TSWV polymerase (RdRp) through its C terminus domain (Fig. 1) and to TSWV RNA through its N terminus domain (Fig. 2). FoTF improves TSWV RNA replication but not transcription in vitro (Fig. 3). Addition of FoTF and reticulocyte lysate leads to the detection of genome-sized TSWV RNAs (Fig. 3A). FoTF per se does not improve TSWV synthesis in vitro, indicating that other cellular factors are required. Moreover, expression of a non-RdRp-binding FoTFΔC mutant (Fig. 1C) or a non-RNA-binding FoTFΔN (Fig. 2C), inhibit TSWV replication in insect cells (Fig. 4 C and D), indicating that the FoTF–RdRp interaction is important in vivo. The fact that both of these mutants blocked TSWV replication indicates that both RNA-binding and RdRp-interaction FoTF capabilities are important for TSWV replication. FoTF is apparently functioning as a bridge between RdRp and viral RNA, and it is possible that this FoTF–RdRp interaction is necessary to ensure and/or increase RdRp-RNA binding specificity and/or stability (14, 28).

More interestingly, we also found that overexpression of FoTF in two human cell lines, HeLa and diploid fibroblasts, turns these cells permissive to TSWV replication (Fig. 5 and data not shown). These data provide insights in the evolution of bunyaviruses. The greater diversity presented by animal bunyaviruses compared with tospoviruses has led some investigators to hypothesize that tospoviruses evolved from animal bunyaviruses (1). However, experimental evidence supporting that hypothesis was missing. The permissibility of human cell lines to TSWV replication by the overexpression of a single exogenous protein (FoTF) indicates that many other host factors necessary in the TSWV life cycle are present in those cells and supports the hypothesis that tospoviruses evolved from animal viruses. In addition, these data provide an experimental system that enables us to investigate molecular defense mechanisms against the same pathogen (TSWV) in human, plant, and insect cells. Further studies could lead to a better understanding of general antiviral mechanisms and the development of new antiviral strategies.