Involvement of a Bovine Viral Diarrhea Virus ... - Journal of Virology

3 downloads 0 Views 1MB Size Report
Sep 24, 2003 - in 12-well cluster trays. At the same ... fresh cell monolayers in a 12-well cluster tray. These cells ...... 36. Lesburg, C. A., M. B. Cable, E. Ferrari,
JOURNAL OF VIROLOGY, Sept. 2004, p. 9612–9623 0022-538X/04/$08.00⫹0 DOI: 10.1128/JVI.78.18.9612–9623.2004 Copyright © 2004, American Society for Microbiology. All Rights Reserved.

Vol. 78, No. 18

Involvement of a Bovine Viral Diarrhea Virus NS5B Locus in Virion Assembly† Israrul H. Ansari,1 Li-Mei Chen,2 Delin Liang,1 Laura H. Gil,1 Weidong Zhong,3† and Ruben O. Donis1* Department of Veterinary and Biomedical Sciences, University of Nebraska—Lincoln, Lincoln, Nebraska1; Section of Microbial Pathogenesis, Yale University School of Medicine, New Haven, Connecticut2; and Research and Development, ICN Pharmaceuticals, Costa Mesa, California3 Received 24 September 2003/Accepted 11 May 2004

A novel mutant of bovine viral diarrhea virus (BVDV) was found with a virion assembly phenotype attributable to an insertion into the NS5B polymerase locus. This mutant, termed 5B-741, was engineered by reverse genetics to express NS5B with a C-terminal peptide tag of 22 amino acids. Electroporation of bovine cells with genomic RNA from this mutant showed levels RNA synthesis which were regarded as sufficient for infectivity, yet infectious virions were not produced. Pseudorevertants of mutant 5B-741 that released infectious virions and formed plaques revealed a single nucleotide change (T12369C). This change resulted in a leucine-toproline substitution within the NS5B tag (L726P). Genetic analysis revealed that indeed a single nucleotide change encoding proline at NS5B position 726 in the pseudorevertant polyprotein mediated recovery of virion assembly function without improving genomic RNA accumulation levels. A subgenomic BVDV reporter replicon (rNS3-5B) was used to analyze the consequences of alterations of the genomic region encoding the NS5B C terminus on replication and assembly. Interestingly, rNS3-5B-L726P (revertant) replicated with the same efficiency as the rNS3-5B-741 mutant but produced 10 times more virions in a trans-packaging assay. These results indicated that impairment of assembly function in 5B-741 was independent of RNA accumulation levels and agreed with the observations from the full-length mutant and revertant genomes. Finally, we recapitulated the packaging defect of 5B-741 with a vaccinia virus expression system to eliminate possible unwanted interactions between the helper virus and the packaged replicon. Taken together, these studies revealed an unexpected role of NS5B in infectious virion assembly. genomic RNA are generally not required for this process. For example, HCV VLPs were produced in insect cells infected with recombinant baculovirus or vesicular stomatitis virus expressing HCV structural proteins (4, 14). It is becoming increasingly clear that the assembly of infectious virions requires the involvement of viral NS proteins. Two such reports have described the role of NS2A in yellow fever virus (YFV) and Kunjin virus assembly (33, 38). An elegant genetic approach demonstrated the role of YFV NS2A in particle assembly (33), where production of infectious virions was blocked by impaired production of NS2A alpha. Pseudorevertants mapped to the NS3 helicase domain, in which NS3 aspartic 343 was replaced with an uncharged residue. These mutations were introduced back in the infectious cDNA clone by a reverse genetics approach to restore virus infectivity, indicating the role of NS2A and NS3 in the assembly and/or release (33). In studies reported by Liu et al., a dual role was reported for Kunjin virus NS proteins in genome replication and virion assembly (38, 39). A Kunjin virus replicon cell line constitutively expressing the NS proteins rescued replication and assembly of RNA genomes bearing NS protein deletions (39). Replication of the Kunjin virus replicon RNA with a mutation in NS2A was not affected but could not be packaged in trans by the Kunjin virus structural proteins. However, the impaired function of the mutated NS2A in production of infectious virus was complemented in trans by the helper wild-type NS2A produced from the Kunjin virus replicon, indicating that a single

Virion assembly and release from the host cell constitute the final stages of the viral life cycle. These events are as important for viral pathogenesis in the host as the widely studied mechanisms of virus entry or genome replication. Enveloped virion assembly represents the convergence of viral proteins, nucleic acid, and host lipids to build an infectious particle. The process of virus budding, which characterizes enveloped viruses, is difficult to study because it involves a living cell. No cell-free virus budding systems have been developed to dissect this process. Hepatitis C virus (HCV) is an important human pathogen that causes persistent infections leading to chronic hepatitis, which can result in hepatocellular carcinoma (1, 2). To avoid the late sequelae of chronic HCV infections, millions of infected individuals await the development of effective antivirals (29). Efforts to explore all the possible drug targets for HCV, including virion assembly, have been hampered by the lack of robust in vitro cell culture systems that yield infectious virions (60). Much of our current knowledge on HCV virion assembly is derived from the study of virus-like particles (VLPs). Numerous reports support the concept that viral structural proteins carry enough structural information to direct the assembly of VLPs in host cells (25, 27, 31, 32). Viral NS proteins or

* Corresponding author. Mailing address: 202 VBS, University of Nebraska—Lincoln, Fair Street and East Campus Loop, Lincoln, NE 68583-0905. Phone: (402) 472-6063. Fax: (402) 472-9690. E-mail: rdonis @unl.edu. † Published as paper number 14692, Journal Series, Nebraska Agricultural Research Division, University of Nebraska—Lincoln. 9612

VOL. 78, 2004

PESTIVIRUS NS5B LOCUS IN VIRION ASSEMBLY

BVDV cDNA

pNADLp15 p5Bgdd pN-⌬INS p5Bmlu p5B-741

p5B-741/Rmlu p5B-L726P

pNS3-5B

Phenotype

TABLE 1. Features of relevant full-length or subgenomic BVDV RNA

Nucleotide and protein changes relative to WT

WT BVDV No RNA replication Noncytopathogenic BVDV WT BVDV Not infectious, yields revertants with 2.4-mm plaques (Fig. 2)

WT BVDV WT-like, 2.6-mm plaque size (Fig. 2) WT RNA replication

No RNA replication RNA replication ⬇10% of WT RNA replication ⬇10% of WT

WT BVDV genome derived from strain NADL cDNA (genotype I) (58) NS5B codons 447 to 449 endoding GDD replaced with ATR (catalytically inactive NS5B mutant) (28) Deletion of host cell-derived JIV element, as described in reference 34 NADL genome with Mlul site (ACGCGT) inserted at position 12350 Added 22 codons (FKSNFDLLKLAGNVESNPGPTR) at position 12350 of pNADLp15 (Fig. 1); spontaneous revertants from r5B-741 RNA (i-5B-741-rev); T12369C substitution encoding L726P (Fig. 1) Derived from p5B-741 by eliminating the 22 codons inserted at position 12350 pNADLp15 derivative by adding 22 codons bearing the NS5B-741rev L-726-P substitution (FKSNFDPLKLAGNVESNPGPTR) at position 12350 Bicistronic subgenomic reporter replicon. BVDV core to NS2 proteins are replaced by firefly luciferase/neomycin genes and EMCV IRES (Fig. 3) rNS3-5B derivative in which NS5B coding region is from p5Bgdd rNS3-5B derivative in which NS5B coding region is from p5B-741 rNS3-5B derivative in which NS5B coding region is from p5B-L726P

MATERIALS AND METHODS Cells, viruses, plasmids, and reagents. Madin-Darby bovine kidney (MDBK) cells, African green monkey kidney (CV1) cells, human osteosarcoma cells lacking thymidine kinase (143B TK⫺), and BHK-21 cells were purchased from the American Type Culture Collection (Manassas, Va.). An simian virus 40-immortalized cell line derived from bovine testicle, termed RD420, was kindly provided by E. Dubovi (Cornell University, Ithaca, N.Y.). Cells were propagated in minimal essential medium (MEM) with 7% equine serum, as described previously (37). BVDV genotype 2 (BVDV2) isolate 890 was obtained from the National Veterinary Services Laboratories (U.S. Department of Agriculture, Ames, Iowa) (49). The cytopathogenic (cp) BVDV strain NADL was obtained from the American Type Culture Collection (21). The pNADLp15 full-length infectious clone derived from the BVDV strain NADL was described previously (59). The noncytopathogenic (ncp) derivative of this plasmid, termed pN-dINS, has also been described elsewhere (59). DNA restriction and modification enzymes, fetal bovine serum, antibiotic solution (containing penicillin and streptomycin), Lglutamine, MEM, and oligonucleotides were from Invitrogen (Carlsbad, Calif.). Monoclonal antibodies to BVDV proteins were kindly provided by E. Dubovi (Cornell University). The dual luciferase assay system was purchased from Promega (Madison, Wis.). Creation of BVDV with mutant NS5B by reverse genetics. BVDV genomes with changes in the NS5B coding region were constructed by removing the segment of DNA encoding the NS5B C-terminal region from the BVDV infectious clone pNADLp15 (59) and inserting the corresponding mutant NS5B segment generated by overlap extension PCR, using standard molecular genetics techniques (3, 22). To this end, we first generated a full-length BVDV infectious clone bearing a unique MluI site at the C terminus of NS5B, p5Bmlu (Table 1). To add a C-terminal extension to NS5B, the sequence encoding a peptide tag derived from foot-and-mouth disease virus 2A protein (22 codons translated into FKSNFDLLKLAGNVESNPGPTR) was generated by complementary synthetic oligonucleotides and inserted by overlap extension into the window created in p5BMlu by restriction endonuclease digestion (50) (Table 1). The resulting full-length genome, termed p5B-741, is predicted to express NS5B with this 22-amino-acid peptide extension at the C terminus. The T-to-C mutation at position 12369 (T12369C), which results in a leucine-to-proline codon change in the viral polyprotein (residue 3995, or 726 in NS5B-741), was engineered as described above with synthetic oligonucleotides. The resulting construct was termed p5B-L726P. All constructs were verified by redundant nucleotide sequence analysis generated in a Beckman/Coulter CEQ2000XL sequencer, with base-calling verification by visual inspection of the chromatograph. In vitro synthesis and characterization of genomic and subgenomic BVDV RNA. DNA from each of the plasmid constructs or full-length PCR amplicons (shown in Table 1) was used as a template for in vitro RNA synthesis, as described previously (Megascript; Ambion) (58). The biological activity of the full-length or subgenomic BVDV RNA was determined by electroporation of the resulting transcripts into cells, followed by infectious center assay, blind passage to recover revertants, harvesting for Northern and Western blotting, luciferase assays, or fixation for indirect immunofluorescence (IIF). Prior to electroporation, the RNA concentration was determined fluorometrically with Ribo-green

pNS3-5Bgdd pNS3-5B741 pNS3-5B-L726P

amino acid change in NS2A is sufficient to impair virion assembly. It is likely that work in progress in these laboratories will yield additional information regarding the mechanism by which NS loci participate in the assembly of infectious flavivirus virions. It is not clear whether the involvement of NS protein in the assembly of virions of the genus Flavivirus is also a feature of the hepaciviruses. Because HCV particle assembly is inefficient and a robust cell culture infectivity assay is not available, experimental studies to assess the role of NS proteins in infectious virion assembly are extremely difficult (47, 48). To circumvent these limitations, we are studying a pestivirus, bovine viral diarrhea virus (BVDV), to unravel the virion assembly process. Here we show that production of infectious pestivirus particles involves the locus encoding RNA-dependent RNA polymerase NS5B protein. Furthermore, we demonstrate that the replicase and virion assembly functions of the NS5B locus can be dissociated.

9613

9614

ANSARI ET AL.

(Molecular Probes), and the structural integrity was determined by gel electrophoresis (35). The RNA transcripts were electroporated into 2 ⫻ 106 cells in a volume of 100 ␮l of Cytomix, using a 0.2-cm gap cuvette, as previously described, with the following modifications (56, 59). BHK-21 cells were electroporated at 250 V, 100 ␮F, and 48 ⍀; RD420 cells were electroporated at 200 V, 200 ␮F, and 48 ⍀. The infectious cell center assay with the electroporated cell population was performed immediately after the procedure. To this end, a logarithmic serial dilution of the electroporated cells was plated onto dishes that were supplemented with an additional number of RD420 cells to reach a final count of 106 cells/well. An agar overlay was added at 3 h postelectroporation (hpe), when the cells were fully attached to the culture dish. The cell monolayer was fixed after 3 days in culture and stained with antibodies to BVDV by immunoperoxidase to count the plaques, as described elsewhere (59). The specific infectivity of BVDV genomic RNA transcripts was calculated by relating plaque counts to the input RNA. The relative infectivity of each RNA was normalized to account for variation in the electroporation efficiency between cuvettes. The efficiency of each electroporation was determined by coelectroporation of an RNA that encodes Renilla luciferase along with the viral RNA. The Renilla reporter RNA was derived from pIRES-RL and included a 1:4 molar ratio relative to the viral RNA transcript of interest (L. Gil and R. Donis, unpublished data). An aliquot of 105 cells was removed from the electroporated cell population to be cultured separately for 4 h. At this time, the cells were harvested and lysed to determine Renilla luciferase activity as described below (42). IIF analysis allowed estimation of the replication and expression levels of the RNA molecules electroporated into cells. When bovine cells were used, it also revealed the spreading of progeny virus in the monolayer. To this end, aliquots from electroporated cells were seeded onto culture dishes containing glass coverslips at the bottom. Cells were fixed at the times indicated and processed as described previously with monoclonal antibodies (MAbs) 19 and 20.10.6 to structural (E2) and nonstructural (NS3) viral proteins, respectively (11). For Western blot analysis, total cell lysates were separated by reducing sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (8.5 or 10% acrylamide) and electrotransferred to nitrocellulose membranes (Hybond-C; Amersham, Arlington Heights, Ill.) by using a Trans-Blot SD semidry transfer cell (Bio-Rad, Hercules, Calif.). Viral proteins were detected with the MAbs described for IIF (8, 9). Peroxidase-conjugated goat anti-mouse immunoglobulin G (IgG) and ECL reagent (Amersham) were used to reveal bound MAbs. Northern hybridization analysis. Total cellular RNA (15 ␮g) was separated in denaturing morpholinepropanesulfonic acid (MOPS)-formaldehyde agarose gels, stained with ethidium bromide, destained in water, and transferred to positively charged nylon membranes (Hybond N⫹; Amersham). Equal loading of samples was verified by ethidium bromide staining of host rRNA. The membrane was irradiated with UV light to cross-link the RNA and was then hybridized to radiolabeled DNA probes obtained by asymmetric PCR amplification of the NS4B coding region of BVDV strain NADL (52). Briefly, asymmetric PCR (50-␮l reaction mixtures) consisted of 5 ␮l of 10⫻ PCR buffer (15 mM MgCl2), 1.2 nM NADL-NS4-F primer, 120 nM NADL-NS4-R primer, 200 ␮M (each) dTTP, dCTP, and dGTP, 10 ␮M dATP, 1 ␮Ci of [␣-32P]dATP/␮l, and 20 ng of plasmid DNA template. Cycling parameters were set to 40 cycles of 94°C for 2 min, 60°C for 30 s, and 72°C for 1 min. A total of 1.0 ⫻ 107 to 3.0 ⫻ 107 trichloroacetic acid (TCA)-precipitable counts per minute was used for each hybridization. Prehybridization was performed for 2 h at 42°C in a buffer containing 50% formamide, 5⫻ SSPE (0.75 M NaCl, 0.05 M NaH2PO4, 5 mM EDTA; pH 7.4), Denhardt’s solution (2 g of Ficoll 400/liter, 2 g of polyvinylpyrrolidone/liter, and 2 g of bovine serum albumin/liter), 2% SDS, and denatured salmon sperm DNA (100 ␮g/ml). Hybridization was performed overnight at 42°C in fresh prehybridization solution to which a probe labeled as described above (106 cpm/ml of hybridization solution) was added. Filters were washed twice with 1⫻ SSC–0.1% SDS (1⫻ SSC is 0.15 M NaCl plus 0.015 M sodium citrate) at room temperature for 5 min and once with 0.1⫻ SSC–0.1% SDS at 50°C for 10 min. Washed filters were exposed in phosphorimager cassettes for 6 h at 22 to 25°C. Detection and quantification of radioactivity were performed with a phosphorimager and by using ImageQuant software (Molecular Dynamics). Identification and characterization of BVDV revertants. To search for the emergence of infectious revertant viruses from engineered genomes with mutations that abolished the formation of plaques or foci, the electroporated cells were split and subcultured at 48, 96, and 144 h. An aliquot of the split cell population was seeded sparsely on microscope glass coverslip-bottomed cluster plate wells. The proportion of virus-infected cells in each culture was determined by IIF with a MAb to E2, performed on acetone-methanol-fixed cells, as described elsewhere (11). The presence of infectious virus was evidenced by a progressive increase in the proportion of infected cells in the monolayer upon

J. VIROL. further passage, sometimes with evidence of vacuolating cytopathic effect. Presence of infectious virus was confirmed by transferring culture fluid and freezethawed cell lysates to new cell cultures and analysis of BVDV protein expression by IIF detection with specific MAbs (11). These viruses were subjected to two rounds of plaque purification, and the resulting stock was stored at ⫺80°C. Virus stocks derived from RNA synthesized in vitro from a plasmid template bear the same name as the plasmid, replacing the first letter (p) with an i. The sequence of the viral genomic RNA was determined using amplicons prepared by reverse transcription-PCR (RT-PCR). The resulting nucleotide sequence information revealed whether a virus was in fact a true revertant or a pseudorevertant. Construction of subgenomic BVDV reporter replicons. Bicistronic reporter replicon was constructed by replacement of the genes that are not required for replication with marker-reporter genes and the encephalomyocarditis virus internal ribosome entry site (IRES) (24). The segment of the BVDV genome (strain NADL, cytopathic biotype) encoding the portion of the polyprotein comprising C, Erns, E1, E2, P7, and NS2 was replaced with two elements in tandem: (i) a firefly luciferase reporter and neomycin phosphotransferase selection cassette, inserted in frame with the Npro coding region, and (ii) the IRES element from encephalomyocarditis virus, inserted such that translation started at a methionine codon added upstream of glycine 1801, at the N terminus of NS3 (Table 1) (13, 54). The luciferase reporter and the neomycin selection marker were separated by an in-frame 76-codon sequence encoding monomeric human ubiquitin (53). These sequences were inserted into the C-NS2 deletion window of the infectious clone (pNADLp15), giving rise to the reporter replicon termed pNS3-5B (see Fig. 3, below) (58, 59). A similar construct based on a version of the infectious clone pNADLp15 has a disabled NS5B by virtue of replacing the conserved GDD motif (amino acid residues 447 to 449, using NS5B numbering) with ATR; the resulting control replicon was termed pNS3-5Bgdd (28). Similarly, full-length genome plasmids p5B-741 and p5B-L726P were used to derive mutant replicon-encoding plasmids, which were termed pNS3-5B-5B-741 and pNS3-5BL726P. Replicon function was assessed by electroporation of in vitro-synthesized RNA into bovine cells, as described previously (Megascript; Ambion) (59). The integrity of the RNA transcripts was verified by MOPS-formaldehyde agarose gel electrophoresis. To determine the efficiency of electroporation of replicon RNA into cells, an RNA transcript that encodes Renilla luciferase was coelectroporated as described above. The cells were harvested at the indicated times and lysed for quantification of luciferase activity (10). Luciferase assays were performed using the dual luciferase assay system (Promega) following the manufacturer’s directions. Luciferase activity was measured by light emission using the TopCount luminescence counter (Packard Instruments, Meriden, Conn.). Electroporation experiments were repeated at least three times, starting from independent in vitro RNA synthesis reactions. Firefly luciferase values were normalized for electroporation efficiency differences and expressed as mean relative light units (RLU) per 105 cells, with the standard error of the mean indicated. Only data from experiments with variations in Renilla luciferase activity among electroporations of ⱕ0.5-fold were considered. Statistical significance of differences between means was determined using Student’s t test implemented in the Prism software package (GraphPad, San Diego, Calif.). Construction of recombinant vaccinia viruses expressing BVDV viral proteins. Vaccinia viruses were engineered by homologous recombination within cells transfected with a modified vaccinia virus transfer plasmid vector and infected with vaccinia virus WR genome (18). To this end, cDNA from BVDV1 strain NADL⌬INS, encoding the full-length polyprotein with a disabled NS5B by mutagenesis of the GDD motif as described above, was cloned into the vaccinia virus recombination vector pSC11 by standard molecular genetics methods (59). The BVDV polyprotein was inserted downstream of the T7 promoter (18). The T7-BVDV cassette in the plasmid was flanked by vaccinia virus thymidine kinase arms, TKL(left) and TKR(right), to mediate recombination at the homologous locus of the vaccinia virus genome. CV1 cells were transfected, and recombinants were selected in HuTK⫺143B cells with 5-bromodeoxyuridine (3). The resulting recombinant vT7-ncpNADLgdd plaques were purified three times before being used for the trans-packaging experiments. Expression of structural and NS proteins in infected cells was determined by Western blot analyses. BVDV reporter replicon trans-encapsidation in cells infected with helper BVDV. To examine the incorporation of replicon RNA (wild-type [WT] or mutant NS5B) into virus particles, in vitro RNA transcripts (3 ␮g) were electroporated into 2 ⫻ 106 mock-infected RD420 cells or into RD420 cells that were infected with BVDV2 (strain 890) at a multiplicity of infection of 5 50% tissue culture infective doses/cell 72 h earlier. RNA transcripts encoding Renilla luciferase (200 ng) were added to each replicon RNA in each electroporation to monitor transfection efficiency, as described above. Electroporated RD420 cells were cultured for 48 h in MEM supplemented with 5% equine serum, and then

VOL. 78, 2004

PESTIVIRUS NS5B LOCUS IN VIRION ASSEMBLY

9615

FIG. 1. Primary structure of the BVDV NS5B peptide tag mutants. The C terminus of NS5B is expanded to show nucleotide and amino acid sequences of mutant 5B-741, engineered with a C-terminal extension of 22 amino acids (underlined), and the representative sequence of three independent pseudorevertants, 5B-741rev, indicating the single-nucleotide transition, T12369C (nomenclature based on the WT base, position, and the mutant base), which converted the leucine 726 codon to proline (NS5B residue numbering, derived from GenBank accession no. M31182).

culture medium and cells were harvested separately. The cells were counted and lysed to determine luciferase activity, a measure of the replicon RNA pool size available for encapsidation at the time of harvest. This parameter was termed packageable RNA. The culture fluid was used to assess the presence of infectious virions containing encapsidated replicon genomes. This was evaluated by analyzing luciferase expression in a fresh bovine cell monolayer inoculated with the cell culture medium, using BVDV-neutralizing antibodies to establish specificity. To this end, culture fluids were passed through a 0.1-␮m filter, divided into three aliquots, and incubated with 10 ␮g of IgG/ml from neutralizing polyclonal antibody to BVDV2, 10 ␮g of preimmune serum IgG/ml, or 10 ␮g of ovalbumin/ml for 30 min at 25°C and then used immediately to inoculate a fresh monolayer of indicator RD420 cells (46). The indicator RD420 cells were cultured for 24 h and then harvested to determine firefly luciferase activity (RLU), a measure of the extent of encapsidation of the reporter genome into functional particles. This RLU value from the indicator cells was termed transduced RNA. To obtain a value that facilitated comparison of the different replicons, we calculated a relative packaging efficiency (RPE) of each replicon with respect to that of the WT, as follows: RPE ⫽ [(PE of transduced RNA)/(PE of packageable RNA)]/ [(PE of WT transduced RNA)/(PE of WT packageable RNA)]. Standard error values were derived from the relative firefly/Renilla luciferase activity ratios calculated for each mutant in three different experiments. To this end, we calculated a quotient of mutant sample value to wild-type samples. The standard error for the quotient was calculated by standard methods (16). Packaging of the replicon RNA transcript by vaccinia virus recombinants was performed as follows. Cells electroporated with replicon RNA as described above were cultured for 24 h. At this time the cells were trypsinized and seeded in 12-well cluster trays. At the same time, about 105 cells were collected for firefly luciferase activity assay. Cells were cultured for 4 h to allow attachment and subsequently infected at a multiplicity of infection of 3.0 with a mixture of two recombinant vaccinia viruses expressing ncpBVDV and T7 RNA polymerase. Cells were cultured for another 24 h, and the culture medium was harvested, filtered through a 0.1-␮m filter (to remove vaccinia virus), and inoculated onto fresh cell monolayers in a 12-well cluster tray. These cells were cultured for 24 h and harvested for luciferase activity assay.

RESULTS A replication-competent NS5B mutant BVDV genome that fails to yield infectious virions. The viral polymerase protein is arguably the most important component of the replicase complex and is thought to include other viral proteins, such as NS3, NS4B, and NS5A. To study the replicase complex of BVDV, we wished to attach a C-terminal extension to NS5B that could be used as an affinity tag for purification. To evaluate the functionality of the tagged NS5B in genome replication, the peptide tags were initially added to NS5B in the context of a full-length BVDV genome. In one of these BVDV genomes, 66 nucleotides encoding a peptide tag derived from the FMDV

2A polypeptide were added in frame at the C terminus of NS5B, followed by the polyprotein stop codon and the 3⬘untranslated region (Fig. 1). The resulting BVDV mutant genome was termed 5B-741 (mutant NS5B is 741 amino acids in length, whereas WT is 719). The specific infectivity of mutant 5B-741 RNA electroporated into bovine cells was determined by infectious center assay to be below the detection level of our assay (Fig. 2A). In contrast, the WT BVDV transcripts yielded 6.7 log10 PFU of RNA/␮g. Interestingly, the 5B-741 genomic RNA appeared to undergo replication in the electroporated cells, as evidenced by the fluorescence intensity in the electroporated cells, indicative of a high level of viral protein expression (Fig. 2B, panel 3). The E2 expression level of the mutant was similar to that of the WT BVDV RNA (Fig. 2B, panels 3 and 9). In contrast, no viral protein expression was observed in cells electroporated with the nonreplicating 5Bgdd RNA (Fig. 2B, panel 1). When culture fluids and lysates from the cells electroporated with the 5B-741 mutant were harvested and inoculated onto a fresh monolayer of cells, no viral protein expression was detected by IIF in the inoculated cells (Fig. 2B, panel 4). These results were consistent with the infectious center assay and Northern hybridization results (see below), which indicated that the 5B-741 genome could replicate but was defective in the production of infectious progeny virus. Suppression of mutations in the 5B-741 genome restored viral infectivity. Serial passage of the cells electroporated with 5B-741 showed a sharp increase in the proportion of infected cells, as indicated by E2-specific IIF as well as vacuolating cytopathology (data not shown). We hypothesized that this phenomenon, which did not occur in the cells electroporated with 5Bgdd, was caused by the emergence of phenotypic revertants. A plaque assay performed on RD420 cells inoculated with the cell culture medium from 5B-741-transfected cells after serial passage showed plaques of ⬃2.4 mm in average diameter, substantially smaller than the WT BVDV plaques, which were ⬃3.4 mm in average diameter under identical conditions (Fig. 2E). The revertant virus pool was termed i-5B741rev1. To demonstrate that the engineered 5B-741 genomes did not have mutations at other sites, the sequence encoding the peptide tag was removed from the p5B-741 clone, yielding

9616

ANSARI ET AL.

J. VIROL.

FIG. 2. The NS5B C terminus conditions the production of infectious viral progeny. (A) Infectivity of in vitro-transcribed RNA from WT BVDV and NS5B mutant genomes determined by infectious center assay. BHK-21 cells were electroporated with RNA transcripts from pNADLp15 or from p5Bgdd, p5B-741, or p5B-L726P serially diluted and plated with additional fresh MDBK cells with agar overlay. (B) Immunofluorescence assay of cells electroporated with in vitro-transcribed RNA. RNA transcripts (2 ␮g) from 5Bgdd (1), mutant 5B-741 (3), 5B-L726P (5), N-⌬INS ncBVDV (7), and WT BVDV NADLp15 (9) were electroporated into RD420 cells and RNA replication was monitored by IIF using anti-E2 antibodies. Release of infectious virus was assessed by harvesting the culture medium (panels 1, 3, 5, 7, and 9) at 24 hpe, inoculating into fresh RD420 cells, and monitoring by IIF using anti-E2 antibodies (panels 2, 4, 6, 8, and 10, respectively). (C) Northern blot analysis of total RNA from cells electroporated with RNA from NS5B mutants. (a) Total RNA from BHK-21 cells electroporated with 5Bgdd (1), 5B-741 (2), 5B-L726P (3), N-⌬INS (4), and N-p15 (5) was analyzed, and 28S rRNA is indicated as a control. (b) The transferred RNA was hybridized to a 32P-labeled single-stranded DNA probe specific to BVDV. The migrations of the viral RNA and rRNA are indicated; labels below the figure indicate the

VOL. 78, 2004

p5B-741/Rmlu. This clone showed a phenotype identical to that of WT BVDV (Table 1 and data not shown). Thus, the inserted sequence encoding a peptide tag extension of NS5B is strictly associated with a severe defect in infectious progeny production by the 5B-741 genome. To further examine the molecular basis of the i-5B-741rev1 revertant phenotype, we performed a second round of RNA electroporation with separate pools of 5B-741 RNA transcripts and generated two additional revertants, termed i-5B-741rev2 and -3. Independent plaques were selected from these revertant stocks for further purification and sequence analysis. Three plaque-purified stocks, termed i-5B-741rev1a, -2a, and -3a, were used to synthesize cDNA by RT-PCR. Sequence analysis of the PCR amplicons from the NS5B region revealed the presence of a single nucleotide change, T12369C (nomenclature indicating the WT base, position, and the mutant base) in the three plaque-purified stocks, resulting in replacement of leucine 726 (NS5B amino acid numbering) with proline (Fig. 1). The repeated identification of identical suppressor mutations in independent revertants indicated that they were probably causally associated to the restoration of infectivity rather than random events. Because proline is often found in the bends of folded protein chains, it may allow an alternative folding of the revertant NS5B C terminus that restores its function in infectious virion production. Moreover, the fact that all the revertants entailed a codon change suggested that the functional restoration concerns the NS5B protein and not the RNA. However, attribution of the revertant phenotype to the T12369C substitution required demonstrating that no additional mutations were present elsewhere in the genome. To this end, we utilized reverse genetics to introduce the T12369C substitution into the WT BVDV infectious clone. The resulting BVDV genome, 5B-L726P, yielded genome-length RNA transcripts with a specific infectivity of 5.9 log10 PFU/␮g (Fig. 2A). The recovered virus, termed i-5B-L726P, was found to form plaques of the same diameter as those formed by the spontaneously arising revertant stock i-5B-741rev (Fig. 2E). Nucleotide sequence analysis of an NS5B RT-PCR amplicon from the i-5B-L726P virus stock revealed the engineered T12369C change, encoding L726P, and no additional nucleotide changes (data not shown). The finding that the specific infectivity of the 5B-L726P RNA transcripts was comparable to that of the WT BVDV transcripts suggested that a single nucleotide change, T12369C, is necessary and sufficient to suppress the assembly phenotype of 5B-741. Impact of the 5B-741 mutation on genome replication. A severe defect in RNA synthesis and accumulation would represent a trivial explanation of the failure of 5B-741 to produce infectious virions. The previous qualitative IIF results (Fig. 2B) revealed that mutant 5B-741 was capable of at least some RNA accumulation which mediated viral protein expression. There-

PESTIVIRUS NS5B LOCUS IN VIRION ASSEMBLY

9617

fore, a quantitative analysis of RNA accumulation capacity by the mutant genomes was necessary to determine whether these levels were below a critical threshold for assembly. Northern hybridization analysis of 5B-741 electroporated cells revealed a 2.5-fold reduction in the RNA accumulation levels compared to those of WT BVDV RNA (Fig. 2C, lanes 2 and 5). To evaluate the impact of such RNA levels on viability, we analyzed ncpBVDV, which is known to synthesize less RNA than the WT cpBVDV strain NADL. This analysis would provide an internal reference for the lowest critical level of RNA that is necessary for virion assembly. RNA accumulation was analyzed in cells electroporated with an ncpBVDV genome, NdINS, which is isogenic with the WT NADL parent of the 5B-741 mutant. This genome revealed excellent virion production in the infectious center assay (Fig. 2A) and in previous studies (59). Interestingly, the Northern blot analysis of mutant 5B-741 showed that RNA accumulation was 3.6 times higher than that of ncpBVDV N-dINS in cells (Fig. 2C, lanes 2 and 4). Thus, the mutant 5B-741 RNA levels seem clearly sufficient for assembly. These findings suggested that the failure of the 5B741 genome to generate infectious virions could not be readily attributed to insufficient genomic RNA accumulation levels. Next, we reasoned that if RNA synthesis were a critical limiting factor for the production of infectious virions in the mutant, the pseudorevertants that produce infectious virions would show increased viral RNA levels in electroporated cells relative to levels in the parent genome (5B-741). The Northern hybridization analysis revealed that the electroporated revertant 5B-L726P RNA transcripts accumulated to similar levels as the mutant 5B-741 (Fig. 2C, lanes 3 and 2, respectively). Therefore, these data also suggested that RNA accumulation levels failed to explain the loss of infectivity by 5B-741 RNA. However, it is important to note that Northern blot analyses do not reveal the polarity of the RNA found in infected cells. Although we used a labeled probe generated by asymmetric PCR which was highly enriched for the antigenomic-sense polarity DNA, it was still possible that the hybridization signal originated largely from abundant negative-strand RNA. This caveat was addressed with the use of a replicon trans-packaging system (discussed below). Structural proteins are, with genomic RNA, the major critical component for the assembly of progeny virions. We analyzed the levels of viral glycoprotein accumulation in cells electroporated with mutant 5B-741 RNA as well as those electroporated with WT or engineered revertant 5B-L726P transcripts. Interestingly, the levels of structural and NS protein accumulation in cells at 24 hpe were very similar for all RNA transcripts (Fig. 2D, panels a and b). Similar results were obtained in immunoblots probed with antibodies to other structural and NS proteins (data not shown). This seemed at odds with the significant differences in RNA accumulation

radioactive probe hybridization signal as measured by phosphorimager analysis for each of the lanes, relative to that of the ncpBVDV N-⌬INS (lane 4) (59). (D) Western blot analysis to demonstrate expression of structural and NS proteins in electroporated cells with mutant NS5B RNA. Panels: 1, 5Bgdd; 2, 5B741; 3, L726P; 4, N-⌬INS; 5, NADLp15. The total protein was resolved by SDS-polyacrylamide gel electrophoresis and probed with anti-E2 and anti-NS3 MAbs, and signal was developed with a chemiluminescence detection system. Molecular mass markers are indicated (in kilodaltons). (E) Plaque phenotype of NS5B mutants. Plaques formed by revertant i-5B-741rev (a), the virus recovered from the engineered mutant i-5B-L726P (b), or WT BVDV strain NADL (c) on bovine cells infected for 60 h. The average diameters of the plaques are indicated.

9618

ANSARI ET AL.

J. VIROL.

FIG. 3. Structure and function of a BVDV reporter replicon in bovine cells. (A) Diagram of the structure of the rNS3-5B and rNS3-5Bgdd BVDV reporter replicons (see Materials and Methods for details). (B) Integrity of in vitro-synthesized RNA analyzed by electrophoresis in MOPS-formaldehyde denaturing gel followed by staining with ethidium bromide. (C) Kinetics of luciferase reporter expression in RD420 cells. In vitro-synthesized RNA (2 ␮g) from the WT BVDV reporter replicon rNS3-5B (grey bars) or the rNS3-5Bgdd (open bars) was electroporated into 2 ⫻ 106 RD420 cells, and cells were divided into aliquots and incubated for the indicated times to determine luciferase activity from 105 cells. Values shown are means of three experiments; error bars show the standard deviations of the means. (D) Correlation between luciferase production and intracellular viral RNA levels. RD420 cells were electroporated as for panel C and at the indicated times harvested for Northern blotting and quantification by phosphorimaging. Total cellular RNA samples were electroporated with rNS3-5Bgdd and harvested at 24 h, or with rNS3-5B and harvested at 24-h intervals after electroporation (lanes 2 to 6, respectively). The ethidium bromide-stained gel (a) demonstrates equal loading of the RNA blotted and hybridized to the 32P-labeled BVDV probes (b). The rNS3-5B radioactive hybridization signal intensity values shown below each lane have been normalized to that of the 96-hpe sample.

levels. Although it remains unresolved, this disparity was consistent with our group’s previous Northern hybridization and Western immunoblot analyses in cells electroporated with WT and N-dINS transcripts (59). This finding indicated that the proposed defect in the formation of infectious virions by mutant 5B-741 is not attributable to insufficient supply of structural or NS proteins. Taken together, these data suggested that the participation of the locus encoding the NS5B C terminus in virion assembly is not a direct consequence of its potential role in determining viral RNA or protein accumulation levels. A BVDV reporter replicon trans-packaging system recapitulates the role of the NS5B-741 locus in virion assembly. To obtain independent evidence on the postulated role of the NS5B-741 locus in virion assembly from a different experimental system, we exploited the previously reported encapsidation of BVDV DI genomes and replicons into functional virions from the perspective of transduction of the RNA to naïve cells (5, 19, 20, 44, 55). We implemented a trans-packaging system based on a bicistronic reporter replicon system developed for this purpose in our laboratory. The BVDV bicistronic reporter replicon is analogous to the HCV replicon described by Loh-

mann et al., as well as others (6, 41) (Fig. 3A). The distinguishing feature of this BVDV replicon compared to the reported HCV replicons is the presence of the firefly luciferase gene at the N terminus of the neomycin phosphotransferase gene, connected to it by a ubiquitin monomer for processing (Fig. 3A) (26). Electroporation of bovine cells with the BVDV reporter replicon, termed rNS3-5B, resulted in the production of luciferase to levels ⬃3 orders of magnitude above levels detected in cells electroporated with the mutant replicon expressing a nonfunctional replicase, rNS3-5Bgdd (Fig. 3B and C). Replicon RNA accumulation peaked at 72 hpe (Fig. 3B). Firefly luciferase expression in cells electroporated with replicon RNA was a reliable indicator of the levels of replicon RNA accumulation, as revealed by the correlation between these values and those for viral RNA quantified by Northern blot hybridization with a phosphorimager (Fig. 3D). The high levels of luciferase activity and the strong hybridization signal in the Northern blot assay suggested that an abundant pool of replicon is available for packaging by trans-encapsidation in the electroporated cells. To determine whether the rNS3-5B replicon could be effi-

VOL. 78, 2004

PESTIVIRUS NS5B LOCUS IN VIRION ASSEMBLY

9619

FIG. 4. trans-encapsidation of rNS3-5B by complementing BVDV structural proteins. (A) RD420 cells were infected with BVDV2 or mock infected and cultured for 72 h (step 1) prior to electroporation with rNS3-5B. After electroporation with replicons (step 2), the cells were harvested for luciferase assay (step 3) to determine the packageable replicon levels, whereas the culture medium was divided into two aliquots, which were incubated with BVDV2-neutralizing or nonneutralizing antibodies prior to inoculation onto fresh bovine cells (step 4). After 24 h, the cells were harvested (step 5) to determine luciferase activity generated by the transduced replicon. (B) Reporter replicon packaging assay. Luciferase activity in RD420 cells (RLU/105 cells) that were mock infected (open bars) or infected with BVDV2 (solid bars) and harvested at 48 hpe with the replicon RNAs, representing the size of the packageable replicon pool. Luciferase activity from RD420 cells (RLU/105 cells) harvested 24 h after inoculation with culture fluid filtrates from mock-infected cells (hatched clear bars) or from BVDV2-infected cells (hatched bars on grey background) after incubation with either 10 ␮g of neutralizing antibody/ml (ascending hatch) or with 10 ␮g of nonneutralizing antibody/ml (descending hatch). The reporter activity in these cells is indicative of the activity of replicons transduced by functional virions.

ciently packaged by structural proteins provided in trans, we electroporated the reporter replicon RNA into cells previously infected with an ncpBVDV. Previous infection with BVDV2 (strain 890) provided the structural proteins in trans. If the helper virus provided a functional virion packaging machinery in trans, the rNS3-5B RNA would be expected to undergo encapsidation into virions and be released into the medium (Fig. 4, step 4). The presence of rNS3-5B RNA packaged into infectious particles could be revealed by exploiting the ability of the rNS3-5B genome to express luciferase after delivery into the cytoplasm of a naïve permissive cell (Fig. 4, step 5). To evaluate these events, the culture fluids from RD420 cells previously infected with BVDV2, or mock infected and electroporated with rNS3-5B, were harvested at 48 hpe and used to inoculate a naïve monolayer of D420 indicator cells and subsequently determine luciferase expression levels. To demonstrate the dependence of the rNS3-5B reporter transduction on the infectivity of trans-packaged replicons, the culture fluids were divided into two aliquots and incubated with anti-BVDV2 neutralizing or preimmune polyclonal antibodies prior to inoculation into RD420 indicator cells. Luciferase expression in indicator cells inoculated with the culture fluids from rNS3-5B

RNA-electroporated cells was 4 log10 higher than that in the neutralized samples (Fig. 4B). The luciferase activity of the indicator cells inoculated with culture fluid from mock-infected RD420 cells electroporated with the rNS3-5B replicon reached only background levels, underscoring the importance of the helper virus (Fig. 4B). Most importantly, reporter transduction failed despite a high level of luciferase expression, indicative of a large pool of packageable replicon RNA (Fig. 4B). These results indicated that the BVDV reporter replicon was packaged into functional virions that were released from the donor cells and were competent for delivery of the replicon genome to a fresh cell. Interestingly, the heterologous genotype 2 BVDV structural proteins showed efficient encapsidation of the genotype 1 replicon, comparable to that of the cognate genotype (data not shown). This also indicated intergenotype compatibility between the structural proteins and the packaged genome. The helper BVDV-mediated trans-packaging system provided a tool to analyze whether replicons with an altered NS5B C-terminus-encoding domain would be packaged into functional virions. Therefore, we constructed derivatives of the rNS3-5B reporter replicon with the NS5B-741 polymerase ex-

9620

ANSARI ET AL.

FIG. 5. The NS5B-741 locus modulates the encapsidation of a BVDV replicon. (A) Luciferase activity is shown for mock-infected or BVDV2 virus-infected cells (clear bars) harvested at 48 hpe with the WT or mutant replicons (as indicated). The cell culture fluid harvested at this time from BVDV2-infected cells was filtered and inoculated onto a naïve cell monolayer. Cells were harvested at 24 h after inoculation, and their luciferase expression levels were determined (shaded bars). (B) RPE for mutant replicons compared to the WT replicon, which was considered 100%.

tension or the L726P reverting mutation (Table 1). Mockinfected bovine cells electroporated with rNS3-5B-741 (mutant) expressed luciferase levels indistinguishable from those of rNS3-5B-L726P (revertant) RNAs at 48 hpe (Fig. 5, mockinfected mutant set). The luciferase activity of these mutants was ⬃10-fold lower than that of the WT rNS3-5B, which showed ⬃6 log10 RLU/105 cells at 48 hpe. These data indicated the similar replicative capacity of the mutant and revertant replicons, both of which trailed behind the replication levels of the WT. These data were also consistent with the RNA accumulation levels of the full-length genome RNA of mutants 5B-741 and 5B-L726P in electroporated cells, revealed by Northern blot analyses (Fig. 2C). Both rNS3-5B-741 and rNS35B-L726P showed higher luciferase levels in the BVDV-infected cells than in the mock-infected cells (⬃12-fold versus

J. VIROL.

75-fold, respectively) (Fig. 5A). This was attributed to the spreading of the replicon to neighboring cells in the monolayer, as reported previously (44). These results also suggested that rNS3-5B-741 was not spreading as efficiently as rNS3-5BL726P, by a factor of 6.5-fold; this difference between means was statistically significant (P ⫽ 0.007, two-tailed t test). This was probably the consequence of inefficient encapsidation of rNS3-5B-741 compared to revertant rNS3-5B-L726P. The higher luciferase values for the WT rNS3-5B relative to those for rNS3-5B-L726P were thought to result from its superior replication capability (Fig. 5A, mock infected and transduced). The transduction of the trans-packaged replicons to naïve cells also revealed that rNS3-5B-741 was less efficiently packaged than rNS3-5B-L726P. To facilitate the comparison, we first calculated the ratio of luciferase activity in the helperinfected and electroporated cells to the luciferase activity of the indicator cells as a measure of the efficiency of transduction of the packageable pool of RNA to the indicator cells by functional virions. The values for each of the mutants were then made relative to the WT, which was arbitrarily set at 100% efficiency, and designated the RPE (Fig. 5B). The revertant rNS3-5B-L726P displayed an RPE value of 77%, compared to 8% for the mutant rNS3-5B-741 (Fig. 5B). Determination of the role of the NS5B-741 locus in BVDV RNA packaging by using a recombinant vaccinia virus packaging system. A variety of differential interactions between the replicons and the helper ncpBVDV could result in a flawed interpretation of the results. For example, the presence of abundant helper viral protein in the infected cells might preferentially stimulate the rate of 5B-741-L726P replication compared to that of the 5B-741 replicons. Conversely, 5B-741 RNA, but not 5B-741-L726P, could exert dominant inhibition of helper virus replication. Although these possibilities are under investigation, we sought to determine the efficiency of WT and mutant reporter replicon packaging in a system in which helper protein expression mediating packaging would not modulate or be modulated by replicon RNA. Recombinant vaccinia virus was shown to be an effective expression system to encapsidate BVDV reporter replicons (D. Liang et al., submitted for publication). Therefore, we exploited this packaging system to compare the relative packaging efficiencies of 5B-741 RNA and 5B-741-L726P (Fig. 6). The normalized levels of luciferase activity expressed in cells inoculated with filtered culture medium from 5B-741-L726P RNA-electroporated cells indicated that the relative packaging efficiency of this genome was 30%, whereas the efficiency of 5B-741 was 1.8% (Fig. 6). The 16-fold difference was consistent with the values obtained using BVDV2 helper virus for packaging. Taken together, these results yielded three pieces of valuable information: (i) the mutant rNS3-5B-741 replicon is packaged inefficiently by helper-virus trans-complementation, (ii) the L726P amino acid substitution in rNS3-5B-L726P yields a ⱖ10-fold improvement in packaging efficiency, and (iii) the NS5B locus has two separable functions, RNA synthesis and infectious virion assembly. DISCUSSION In the course of reverse genetics experiments to study the BVDV replicase, we identified a BVDV mutant with a virion

VOL. 78, 2004

FIG. 6. Transduction of BVDV NS5B mutant replicons packaged in cells infected with a helper vaccinia virus-T7 expressing the BVDV polyprotein with a catalytically inactive NS5B. Luciferase activity of 105 cells transfected with rNS3-5Bgdd, rNS3-5B-741, rNS3-5B-L726P, or WT replicon RNA and infected with vaccinia virus recombinant vT7-ncpNADLgdd, as indicated, at 24 hpe, when cell culture medium was harvested to assess replicon packaging efficiency. Firefly luciferase activities of cells inoculated with culture medium filtrates from the packaging cell monolayers and lysed at 24 h after inoculation are shown. Values are averages from two experiments.

assembly phenotype attributable to the addition of a 66-nucleotide sequence encoding a peptide tag addition to the NS5B C terminus. The fact that three independent revertants showed an identical nucleotide change, which converted a leucine codon to proline, strongly suggested that selection is for the tagged NS5B protein rather than the structure of the RNA. However, the involvement of an RNA signal could not be formally excluded. Analysis of mutants with other peptide tags in frame with NS5B at this locus (nucleotide 12350) provided additional clues. In one case, the elimination of the polyprotein UGA stop codon (nucleotide 12353, converted to threonine) resulted in a viable virus with a 10-amino-acid tag at the C terminus of NS5B (data not shown). In another case, addition of 18 nucleotides at position 12350 yielded a viable virus. These examples suggested that various alterations of the RNA and protein structure at this locus are not deleterious for viral infectivity (data not shown), while other cases showed that peptide tag addition was lethal. In one case, addition of 9 amino acids was lethal for infectivity; the recovered revertants showed suppressor mutations that resulted in translation termination near the WT C terminus of NS5B. Although certain peptide tag additions are suppressed by insertion of stop codons, others, like 5B-741, select mutations that reduce the detrimental effect of the foreign sequences, i.e., proline. These results seem to favor the notion that addition of C-terminal

PESTIVIRUS NS5B LOCUS IN VIRION ASSEMBLY

9621

peptide tags to NS5B interfered with virus viability through their impact on NS5B protein function. The function of the C-terminal domain of the pestiviral NS5B has not been reported. Interestingly, studies with the HCV replicon in Huh-7 hepatoma cells revealed that deletion of the C-terminal 20 amino acids of NS5B abolished replication (23) but were necessary and sufficient to target NS5B to intracellular membranes, primarily the endoplasmic reticulum (7, 23, 51). The functional importance of the HCV NS5B C-tail in vivo is not recapitulated by the in vitro systems. Purified HCV NS5B expressed from prokaryotic or eukaryotic sources lacking the C-tail is functional in RNA synthesis, even after the addition of peptide tags to the C terminus (15, 36, 40, 61). So far there have been no data indicating a possible involvement of the HCV NS5B C-tail in virion assembly. By utilizing reverse genetics, we showed that the peptide tag addition was deleterious for viral infectivity, whereas L726P substitution restored infectivity. However, these studies could not conclusively demonstrate that mutant viral RNA pool alterations, whether quantitative (reduced accumulation level) or qualitative (polarity) in nature, were not at the root of the assembly defect. Circumstantial evidence from other viral genomes, such as N-dINS, suggested that efficient virion assembly was possible with greatly reduced viral RNA levels in infected cells (34). Notwithstanding, we wished to utilize a different virion assembly system in which defined positivestrand RNA is supplied to the packaging machinery in the context of a uniform source of helper structural and NS proteins expressed in trans. To this end, we engineered a BVDV reporter replicon that expressed a sensitive and short-lived reporter such as luciferase. This NS3-5B replicon replicated efficiently in bovine cells as well as other cell types, such as BHK-21 and Huh-7, routinely producing signal-to-noise ratios of 2 to 3.3 log10. We exploited this BVDV reporter replicon to demonstrate that the NS3-5B-741 mutant genome was capable of identical levels of positive-strand RNA accumulation as the revertant genome NS3-5B-L726P. This indicated that the mutant and pesudorevertant genomes supplied the same amounts of positive-stranded RNA to the packaging machinery. This eliminated one of the limitations of the Northern blot analysis of the full-length viral genomes, which could not rule out a possible qualitative difference between the RNAs accumulated by the mutant and pseudorevertant genomes. The presence of replicating subgenomic reporter replicon and helper BVDV2 RNAs within a single cell is predicted to yield recombinant RNA molecules, most likely by a copychoice (template-switching) mechanism rather than true recombination (30). RNA recombination between the helper virus and rNS3-5B could lead to flawed conclusions. However, the frequency of such recombination events is predicted to be far below detection within the timeframe and reporter expression levels of our system. By analogy with other positivestranded RNA viruses, intertypic recombination rates, e.g., between the type 1 replicon and the type 2 helper BVDV, are predicted to occur with a frequency of ⬃10⫺8 (12, 17, 43, 57). Thus, RNA recombination seems unlikely as the mechanism yielding the luciferase expression observed in our studies. The encapsidation of NS3-5B-741 was ⬃9-fold less efficient that that of the revertant NS3-5B-L726P, despite their identical replication properties. This finding suggested that a non-

9622

ANSARI ET AL.

complementable defect in NS5B was preventing mutant NS35B-741 RNA from being encapsidated efficiently. However, these results cannot rule out the possibility that an RNA signal required for packaging is disrupted by the 66-nucleotide peptide tag coding element present in NS5B-741. Recently, there have been reports of involvement of viral NS proteins in virion assembly, especially with Kunjin virus, where virion assembly is coupled to genome replication. Although the mechanisms have not been understood fully, it is tempting to postulate NS5B as a leading candidate. Coupling between replication and assembly was also demonstrated in polioviruses (45). In the case of pestiviruses, it could be argued that coupling assembly to replication is almost a necessity; however, more studies are needed to determine whether assembly and replication are coupled in pestiviruses and, if so, the possible involvement of NS5B in the process. The complete loss of infectious BVDV production caused by a small insertion at the NS5B locus despite efficient RNA replication is intriguing, because no other reports are available to date implicating NS5B in virion assembly. However, mutations in NS protein in YFV and Kunjin virus had no effect on RNA replication but eliminated infectivity. Both Kunjin virus and YFV NS2A mutants could be efficiently complemented by the NS2A provided in trans. In contrast, full-length or subgenomic RNAs with the 5B-741 mutation were not complemented efficiently by expression of all the viral proteins provided by the helper virus. This finding suggests that there is a major difference between the function of NS2A and NS5B loci in flavivirus and pestivirus assembly, respectively. The NS3 protein of Kunjin virus was also reported to function in virion assembly (39), where deletions in the NS3 C terminus (codons 178 to 611) were complemented for RNA replication but failed to produce secreted particles, indicating that NS3 is required in cis for virus assembly. Interestingly, these studies showed that the RNA replication function of Kunjin virus NS3 is separable from its role in assembly (39). Thus, the linkage of NS3 in Kunjin virus to virion assembly is analogous to the linkage of NS5B in pestivirus assembly with regards to their noncomplementability and their independence from the RNA replication function of these proteins. Our working hypothesis is that the assembly defect in 5B-741 mutants is associated with a failure to couple RNA replication with virion formation. In this scenario, we envision that the 5B-741 locus functions to target newly synthesized RNA into sites where the C protein, in contact with E1 and E2, would nucleate a budding event, possibly at endoplasmic reticulumderived membrane lipid rafts. If indeed replication and assembly are functionally coupled, the proteins involved may reside in physically connected membrane-bound complexes, which would be amenable to biochemical analyses. The spatial arrangement of the viral (and cellular?) protein components of mutant 5B-741 replicase-assembly complexes may be revealed by various experimental approaches including cross-linking, mass spectrometry, and analytical ultracentrifugation. These studies may help build a model of the molecular machinery responsible for the selectivity of genomic RNA encapsidation and the efficient coordination of all the events leading to the assembly and release of infectious virions. The pestiviruses are a most useful model to help understand HCV biology, and identification of essential virion assembly functions within the

J. VIROL.

genomes of these viruses may provide an opportunity to target them for the development of antiviral drugs. ACKNOWLEDGMENTS This work was supported by grants 97-35204-5068 and 2002-3520411619 from the NRI program of the U.S. Department of Agriculture. We thank the UNL Center for Biotechnology and the Nebraska Center for Virology for access to core research facilities and financial support. We also thank E. J. Dubovi for antibodies and cells. The members of the Donis lab are acknowledged for their numerous contributions. REFERENCES 1. Alter, H. J., and L. B. Seeff. 2000. Recovery, persistence, and sequelae in hepatitis C virus infection: a perspective on long-term outcome. Semin. Liver Dis. 20:17–35. 2. Alter, M. J. 1997. Epidemiology of hepatitis C. Hepatology 26:62S–65S. 3. Ausubel, F. M., R. Brent, R. E. Kingston, D. D. Moore, J. G. Seidman, J. A. Smith, and K. Struhl (ed.). 2001. Current protocols in molecular biology, 2nd ed. John Wiley and Sons, New York, N.Y. 4. Baumert, T. F., S. Ito, D. T. Wong, and T. J. Liang. 1998. Hepatitis C virus structural proteins assemble into virus-like particles in insect cells. J. Virol. 72:3827–3836. 5. Behrens, S. E., C. W. Grassmann, H. J. Thiel, G. Meyers, and N. Tautz. 1998. Characterization of an autonomous subgenomic pestivirus RNA replicon. J. Virol. 72:2364–2372. 6. Blight, K. J., A. A. Kolykhalov, and C. M. Rice. 2000. Efficient initiation of HCV RNA replication in cell culture. Science 290:1972–1974. 7. Brass, V., E. Bieck, R. Montserret, B. Wolk, J. A. Hellings, H. E. Blum, F. Penin, and D. Moradpour. 2002. An amino-terminal amphipathic alphahelix mediates membrane association of the hepatitis C virus nonstructural protein 5A. J. Biol. Chem. 277:8130–8139. 8. Corapi, W. V., R. O. Donis, and E. J. Dubovi. 1990. Characterization of a panel of monoclonal antibodies and their use in the study of the antigenic diversity of bovine viral diarrhea virus. Am. J. Vet. Res. 51:1388–1394. 9. Corapi, W. V., R. O. Donis, and E. J. Dubovi. 1988. Monoclonal antibody analyses of cytopathic and noncytopathic viruses from fatal bovine viral diarrhea virus infections. J. Virol. 62:2823–2827. 10. de Wet, J. R., K. V. Wood, M. DeLuca, D. R. Helinski, and S. Subramani. 1987. Firefly luciferase gene: structure and expression in mammalian cells. Mol. Cell. Biol. 7:725–737. 11. Donis, R. O., W. Corapi, and E. J. Dubovi. 1988. Neutralizing monoclonal antibodies to bovine viral diarrhoea virus bind to the 56K to 58K glycoprotein. J. Gen. Virol. 69:77–86. 12. Duggal, R., A. Cuconati, M. Gromeier, and E. Wimmer. 1997. Genetic recombination of poliovirus in a cell-free system. Proc. Natl. Acad. Sci. USA 94:13786–13791. 13. Duke, G. M., M. A. Hoffman, and A. C. Palmenberg. 1992. Sequence and structural elements that contribute to efficient encephalomyocarditis virus RNA translation. J. Virol. 66:1602–1609. 14. Ezelle, H. J., D. Markovic, and G. N. Barber. 2002. Generation of hepatitis C virus-like particles by use of a recombinant vesicular stomatitis virus vector. J. Virol. 76:12325–12334. 15. Ferrari, E., J. Wright-Minogue, J. W. Fang, B. M. Baroudy, J. Y. Lau, and Z. Hong. 1999. Characterization of soluble hepatitis C virus RNA-dependent RNA polymerase expressed in Escherichia coli. J. Virol. 73:1649–1654. 16. Fieller, E. 1940. The biological standardization of insulin. J. R. Statist. Soc. 7(Suppl.):1–64. 17. Fricke, J., M. Gunn, and G. Meyers. 2001. A family of closely related bovine viral diarrhea virus recombinants identified in an animal suffering from mucosal disease: new insights into the development of a lethal disease in cattle. Virology 291:77–90. 18. Fuerst, T. R., P. L. Earl, and B. Moss. 1987. Use of a hybrid vaccinia virus-T7 RNA polymerase system for expression of target genes. Mol. Cell. Biol. 7:2538–2544. 19. Grassmann, C. W., O. Isken, and S. E. Behrens. 1999. Assignment of the multifunctional NS3 protein of bovine viral diarrhea virus during RNA replication: an in vivo and in vitro study. J. Virol. 73:9196–9205. 20. Grassmann, C. W., O. Isken, N. Tautz, and S. E. Behrens. 2001. Genetic analysis of the pestivirus nonstructural coding region: defects in the NS5A unit can be complemented in trans. J. Virol. 75:7791–7802. 21. Gutekunst, D. E., and W. A. Malmquist. 1964. Complement-fixing and neutralizing antibody response to bovine viral diarrhea and hog cholera antigens. Can. J. Comp. Med. 28:19–23. 22. Horton, R. M., H. D. Hunt, S. N. Ho, J. K. Pullen, and L. R. Pease. 1989. Engineering hybrid genes without the use of restriction enzymes: gene splicing by overlap extension. Gene 77:61–68. 23. Ivashkina, N., B. Wolk, V. Lohmann, R. Bartenschlager, H. E. Blum, F. Penin, and D. Moradpour. 2002. The hepatitis C virus RNA-dependent

VOL. 78, 2004

24. 25. 26. 27. 28. 29. 30. 31. 32.

33. 34.

35. 36.

37. 38.

39.

40.

41.

RNA polymerase membrane insertion sequence is a transmembrane segment. J. Virol. 76:13088–13093. Jang, S. K., M. V. Davies, R. J. Kaufman, and E. Wimmer. 1989. Initiation of protein synthesis by internal entry of ribosomes into the 5⬘ nontranslated region of encephalomyocarditis virus RNA in vivo. J. Virol. 63:1651–1660. Jasenosky, L. D., G. Neumann, I. Lukashevich, and Y. Kawaoka. 2001. Ebola virus VP40-induced particle formation and association with the lipid bilayer. J. Virol. 75:5205–5214. Jentsch, S., W. Seufert, and H. P. Hauser. 1991. Genetic analysis of the ubiquitin system. Biochim. Biophys. Acta 1089:127–139. Justice, P. A., W. Sun, Y. Li, Z. Ye, P. R. Grigera, and R. R. Wagner. 1995. Membrane vesiculation function and exocytosis of wild-type and mutant matrix proteins of vesicular stomatitis virus. J. Virol. 69:3156–3160. Kamer, G., and P. Argos. 1984. Primary structural comparison of RNAdependent polymerases from plant, animal and bacterial viruses. Nucleic Acids Res. 12:7269–7282. Kim, W. R. 2002. The burden of hepatitis C in the United States. Hepatology 36:S30–S34. Kirkegaard, K., and D. Baltimore. 1986. The mechanism of RNA recombination in poliovirus. Cell 47:433–443. Konishi, E., A. Fujii, and P. W. Mason. 2001. Generation and characterization of a mammalian cell line continuously expressing Japanese encephalitis virus subviral particles. J. Virol. 75:2204–2212. Konishi, E., S. Pincus, B. A. Fonseca, R. E. Shope, E. Paoletti, and P. W. Mason. 1991. Comparison of protective immunity elicited by recombinant vaccinia viruses that synthesize E or NS1 of Japanese encephalitis virus. Virology 185:401–410. Kummerer, B. M., and C. M. Rice. 2002. Mutations in the yellow fever virus nonstructural protein NS2A selectively block production of infectious particles. J. Virol. 76:4773–4784. Lai, V. C., W. Zhong, A. Skelton, P. Ingravallo, V. Vassilev, R. O. Donis, Z. Hong, and J. Y. Lau. 2000. Generation and characterization of a hepatitis C virus NS3 protease-dependent bovine viral diarrhea virus. J. Virol. 74:6339– 6347. Le Pecq, J. B., and C. Paoletti. 1966. A new fluorometric method for RNA and DNA determination. Anal. Biochem. 17:100–107. Lesburg, C. A., M. B. Cable, E. Ferrari, Z. Hong, A. F. Mannarino, and P. C. Weber. 1999. Crystal structure of the RNA-dependent RNA polymerase from hepatitis C virus reveals a fully encircled active site. Nat. Struct. Biol. 6:937–943. Liang, D., I. F. Sainz, I. H. Ansari, L. H. Gil, V. Vassilev, and R. O. Donis. 2003. The envelope glycoprotein E2 is a determinant of cell culture tropism in ruminant pestiviruses. J. Gen. Virol. 84:1269–1274. Liu, W. J., H. B. Chen, and A. A. Khromykh. 2003. Molecular and functional analyses of Kunjin virus infectious cDNA clones demonstrate the essential roles for NS2A in virus assembly and for a nonconservative residue in NS3 in RNA replication. J. Virol. 77:7804–7813. Liu, W. J., P. L. Sedlak, N. Kondratieva, and A. A. Khromykh. 2002. Complementation analysis of the flavivirus Kunjin NS3 and NS5 proteins defines the minimal regions essential for formation of a replication complex and shows a requirement of NS3 in cis for virus assembly. J. Virol. 76:10766– 10775. Lohmann, V., F. Korner, U. Herian, and R. Bartenschlager. 1997. Biochemical properties of hepatitis C virus NS5B RNA-dependent RNA polymerase and identification of amino acid sequence motifs essential for enzymatic activity. J. Virol. 71:8416–8428. Lohmann, V., F. Korner, J. Koch, U. Herian, L. Theilmann, and R. Bartenschlager. 1999. Replication of subgenomic hepatitis C virus RNAs in a hepatoma cell line. Science 285:110–113.

PESTIVIRUS NS5B LOCUS IN VIRION ASSEMBLY

9623

42. Matthews, J. C., K. Hori, and M. J. Cormier. 1977. Purification and properties of Renilla reniformis luciferase. Biochemistry 16:85–91. 43. Meyers, G., and H. J. Thiel. 1996. Molecular characterization of pestiviruses. Adv. Virus Res. 47:53–118. 44. Myers, T. M., V. G. Kolupaeva, E. Mendez, S. G. Baginski, I. Frolov, C. U. Hellen, and C. M. Rice. 2001. Efficient translation initiation is required for replication of bovine viral diarrhea virus subgenomic replicons. J. Virol. 75:4226–4238. 45. Nugent, C. I., K. L. Johnson, P. Sarnow, and K. Kirkegaard. 1999. Functional coupling between replication and packaging of poliovirus replicon RNA. J. Virol. 73:427–435. 46. Odeon, A. C., C. L. Kelling, D. J. Marshall, E. S. Estela, E. J. Dubovi, and R. O. Donis. 1999. Experimental infection of calves with bovine viral diarrhea virus genotype II (N.Y.-93). J. Vet. Diagn. Investig. 11:221–228. 47. Pietschmann, T., V. Lohmann, A. Kaul, N. Krieger, G. Rinck, G. Rutter, D. Strand, and R. Bartenschlager. 2002. Persistent and transient replication of full-length hepatitis C virus genomes in cell culture. J. Virol. 76:4008–4021. 48. Purcell, R. 1997. The hepatitis C virus: overview. Hepatology 26:11S–14S. 49. Ridpath, J. F., and S. R. Bolin. 1995. The genomic sequence of a virulent bovine viral diarrhea virus (BVDV) from the type 2 genotype: detection of a large genomic insertion in a noncytopathic BVDV. Virology 212:39–46. 50. Ryan, M. D., and J. Drew. 1994. Foot-and-mouth disease virus 2A oligopeptide mediated cleavage of an artificial polyprotein. EMBO J. 13:928–933. 51. Schmidt-Mende, J., E. Bieck, T. Hugle, F. Penin, C. M. Rice, H. E. Blum, and D. Moradpour. 2001. Determinants for membrane association of the hepatitis C virus RNA-dependent RNA polymerase. J. Biol. Chem. 276:44052– 44063. 52. Scully, S. P., M. E. Joyce, N. Abidi, and M. E. Bolander. 1990. The use of polymerase chain reaction generated nucleotide sequences as probes for hybridization. Mol. Cell. Probes 4:485–495. 53. Sharp, P. M., and W. H. Li. 1987. Ubiquitin genes as a paradigm of concerted evolution of tandem repeats. J. Mol. Evol. 25:58–64. 54. Tautz, N., K. Elbers, D. Stoll, G. Meyers, and H. J. Thiel. 1997. Serine protease of pestiviruses: determination of cleavage sites. J. Virol. 71:5415– 5422. 55. Tautz, N., T. Harada, A. Kaiser, G. Rinck, S. Behrens, and H. J. Thiel. 1999. Establishment and characterization of cytopathogenic and noncytopathogenic pestivirus replicons. J. Virol. 73:9422–9432. 56. van den Hoff, M. J., V. M. Christoffels, W. T. Labruyere, A. F. Moorman, and W. H. Lamers. 1995. Electrotransfection with “intracellular” buffer. Methods Mol. Biol. 48:185–197. 57. van Vugt, J. J., T. Storgaard, M. B. Oleksiewicz, and A. Botner. 2001. High frequency RNA recombination in porcine reproductive and respiratory syndrome virus occurs preferentially between parental sequences with high similarity. J. Gen. Virol. 82:2615–2620. 58. Vassilev, V., M. Collett, and R. Donis. 1997. Authentic and chimeric fulllength genomic cDNA clones of bovine viral diarrhea virus that yield infectious transcripts. J. Virol. 71:471–478. 59. Vassilev, V. B., and R. O. Donis. 2000. Bovine viral diarrhea virus induced apoptosis correlates with increased intracellular viral RNA accumulation. Virus Res. 69:95–107. 60. Walker, M. P., T. C. Appleby, W. Zhong, J. Y. Lau, and Z. Hong. 2003. Hepatitis C virus therapies: current treatments, targets and future perspectives. Antivir. Chem. Chemother. 14:1–21. 61. Yamashita, T., S. Kaneko, Y. Shirota, W. Qin, T. Nomura, K. Kobayashi, and S. Murakami. 1998. RNA-dependent RNA polymerase activity of the soluble recombinant hepatitis C virus NS5B protein truncated at the Cterminal region. J. Biol. Chem. 273:15479–15486.