A New Role for ns Polyprotein Cleavage in Sindbis Virus Replication

Abstract

One of the distinguishing features of the alphaviruses is a sequential processing of the nonstructural polyproteins P1234 and P123. In the early stages of the infection, the complex of P123+nsP4 forms the primary replication complexes (RCs) that function in negative-strand RNA synthesis. The following processing steps make nsP1+P23+nsP4, and later nsP1+nsP2+nsP3+nsP4. The latter mature complex is active in positive-strand RNA synthesis but can no longer produce negative strands. However, the regulation of negative- and positive-strand RNA synthesis apparently is not the only function of ns polyprotein processing. In this study, we developed Sindbis virus mutants that were incapable of either P23 or P123 cleavage. Both mutants replicated in BHK-21 cells to levels comparable to those of the cleavage-competent virus. They continuously produced negative-strand RNA, but its synthesis was blocked by the translation inhibitor cycloheximide. Thus, after negative-strand synthesis, the ns proteins appeared to irreversibly change conformation and formed mature RCs, in spite of the lack of ns polyprotein cleavage. However, in the cells having no defects in α/β interferon (IFN-α/β) production and signaling, the cleavage-deficient viruses induced a high level of type I IFN and were incapable of causing the spread of infection. Moreover, the P123-cleavage-deficient virus was readily eliminated, even from the already infected cells. We speculate that this inability of the viruses with unprocessed polyprotein to productively replicate in the IFN-competent cells and in the cells of mosquito origin was an additional, important factor in ns polyprotein cleavage development. In the case of the Old World alphaviruses, it leads to the release of nsP2 protein, which plays a critical role in inhibiting the cellular antiviral response.

The Alphavirus genus of the Togaviridae family contains a number of important human and animal pathogens (16, 40, 44). These viruses are distributed on all of the continents and are capable of causing widespread epidemics. In natural conditions, they are transmitted by mosquito vectors, in which alphaviruses cause a persistent, life-long infection that does not noticeably affect the biology of the insects (44). In vertebrate hosts, the infection is always acute and characterized by high-titer viremia and efficient virus replication in susceptible tissues (15). Alphavirus infection in vitro, in cells of both mosquito and vertebrate origin, is characterized by efficient replication, with a release of more than 10,000 infectious virions from each infected cell.

The alphavirus genome is a single, positive-strand RNA molecule of 11.5 kb (19, 39). This RNA mimics the structure of cellular messenger RNAs by having a cap on its 5′ end and a poly(A) tail at the 3′ terminus. The viral nonstructural proteins (nsPs) nsP1, nsP2, nsP3, and nsP4 are encoded by the 5′ two-thirds of the alphavirus genome. They are synthesized initially as two polyproteins. P1234, containing all four nsP sequences, is formed by all alphaviruses, while alphaviruses that encode an opal codon at the end of the nsP3 gene also produce P123 polyproteins containing only the nsP1, nsP2, and nsP3 sequences. Sequential P1234 processing eventually leads to the formation of the viral replication complex (RC), which functions in both RNA genome replication and the synthesis of the subgenomic (SG) RNA. The latter RNA encodes the viral structural proteins that, together with the genome RNA, form infectious viral particles. The mechanism of alphavirus RNA synthesis has been studied using, as prototypes, Sindbis virus (SINV) and Semliki Forest virus (SFV) (21, 22, 37). The replication of the SINV genome starts with the synthesis of negative-strand RNA that, in turn, serves as a template for the synthesis of new viral genomes and the SG RNA. The synthesis of the negative and positive strands and the SG RNAs is a highly regulated process. While the synthesis of negative-strand RNA occurs only early in infection in most cell types, RCs containing these newly synthesized templates are stable entities and retain positive-strand polymerase activity even in the presence of translation inhibitors, and the number of RCs determines the rate of overall SINV RNA synthesis. The regulation of RNA synthesis is achieved by the differential cleavage of the ns polyprotein (21, 22, 37). The first cleavage, mediated by the nsP2-associated protease, releases the nsP4 polymerase subunit, and the complex of P123 and nsP4 forms the primary RC that is capable of negative-strand RNA synthesis and synthesizes positive-strand RNAs very inefficiently. The following step of processing releases nsP1, and the complex of nsP1+P23+nsP4 is capable of positive-strand RNA synthesis but retains the ability to synthesize the genome-length, negative strands. The last cleavage between nsP2 and nsP3 transforms the RC into the mature complex, which is active in positive-strand RNA synthesis but can no longer synthesize negative strands. This current, elegant model provides a plausible explanation for the regulation of the virus-specific RNA synthesis (40).

Polyprotein processing also might play a critical role(s) in other processes, such as virus-host cell interactions (13). It has been demonstrated that at least SINV nsP2 is involved in functions other than being a component of the cytoplasmic, membrane-bound RC (8, 9, 11, 28, 30). In the SINV-infected cells, this 807-amino acid (aa)-long protein also can be found in cell nuclei, where nsP2 has been shown to be involved directly in the shutoff of host transcription and the downregulation of the cellular response to infection (9, 13). Moreover, like the nsP2 proteins of other alphaviruses (25), SINV nsP2 is coisolated with ribosomes from infected cells (1), but its functional role in these complexes is unknown. Specific amino acid substitutions in SINV and SFV nsP2 strongly affected the ability of the protein to modify the intracellular environment (6, 26, 32); the mutations altering the carboxy-terminal fragment of nsP2 made viruses attenuated in vivo and incapable of inhibiting host cell transcription and/or translation (9, 13, 14). Thus, accumulated data suggest that the processing of SINV and SFV nsP precursors plays a role not only in the temporal regulation of negative- and positive-strand viral RNA syntheses by the RC but also in the modification of the cellular response to infection.

To probe further the functions of SINV nsPs and ns polyproteins, we developed SINV mutants that were incapable of processing one or two cleavage sites in P1234 but retained the ability to replicate in cultured cells. The results of our analysis of their replication in different cell types showed that such mutants were capable of both negative- and positive-strand RNA syntheses but were unable to downregulate the host α/β interferon (IFN-α/β) response or the activation of IFN-α/β-inducible genes. This limited their replication in cells having no defects in IFN-α/β induction and signaling. Of interest, the mutants were incapable of replicating in cells of mosquito origin. The inability of the mutants to downregulate antiviral responses in cells of vertebrate origin or to replicate in insect cells would make them incapable of circulating in nature in spite of the ability to synthesize virus-specific RNAs.

MATERIALS AND METHODS

Cell cultures.

The BHK-21 cells were kindly provided by Paul Olivo (Washington University, St. Louis, MO). The NIH 3T3 cells were obtained from the American Type Tissue Culture Collection (Manassas, VA). These cell lines were maintained at 37°C in alpha minimum essential medium (αMEM) supplemented with 10% fetal bovine serum (FBS) and vitamins. Mosquito C₇10 cells were obtained from Henry Huang (Washington University, St. Louis, MO). They were propagated in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% heat-inactivated FBS and 10% tryptose phosphate broth (TPB). The IFN-α/βR^−/− mouse embryonic fibroblasts (MEFs) were kindly provided by Michael Diamond (Washington University, St. Louis, MO) and were propagated in DMEM supplemented with 10% FBS.

Plasmid constructs.

Standard recombinant DNA techniques were used for all plasmid constructions. Maps and sequences are available from the authors upon request. pSINV/GFP and pSINV/2V/GFP, described elsewhere (8), encoded SINV genomes having either unmodified cleavage sites in the ns polyprotein or an nsP2/nsP3 cleavage site with a Gly-to-Val substitution in the P2 position, respectively. pSINV/1V2V/GFP encoded the SINV genome with both nsP1/nsP2 and nsP2/nsP3 cleavage sites inactivated by a Gly-to-Val substitution in the P2 positions (37). All of the recombinant genomes also contained a GFP sequence after aa 389 of the nsP3 protein, and the UGA codon between nsP3 and nsP4 was replaced by the Cys-coding UGU codon. Further derivatives were made by introducing the adaptive mutations discovered during this study into nsP4 by standard, PCR-based mutagenesis techniques. pSINrep/1V2V/GFP/A/PAC encoded a SINV replicon that had all of the viral structural genes under the control of the SG promoter replaced by a puromycin acetyltransferase gene (PAC). The details are described in Results.

RNA transcriptions.

Plasmids were purified by centrifugation in CsCl gradients. Before the transcription reaction, the viral and replicon genome-coding plasmids were linearized by XhoI digestion. RNAs were synthesized by SP6 RNA polymerase in the presence of cap analog according to the manufacturer's recommendations (Invitrogen). The yield and integrity of transcripts were analyzed by gel electrophoresis under nondenaturing conditions. Aliquots of transcription reaction mixtures were used for electroporation without additional purification.

RNA transfections.

The electroporation of BHK-21 cells was performed under previously described conditions (24). To rescue the viruses, 1 μg of in vitro-synthesized viral genome RNA was electroporated into cells (24), which then were seeded into 100-mm dishes and incubated until a cytopathic effect (CPE) was observed. Virus titers were determined using a standard plaque assay on BHK-21 cells (20). To assess the RNA infectivity, 10-fold dilutions of electroporated BHK-21 cells were seeded in 6-well Costar plates containing subconfluent naïve cells. After 1 h of incubation at 37°C in a 5% CO₂ incubator, cells were overlaid with 2 ml 0.5% Ultra-Pure agarose (Invitrogen) supplemented with MEM and 3% FBS. Plaques were stained with crystal violet after 2 days of incubation at 37°C, and infectivity was determined as PFU per microgram of transfected RNA.

Viral replication analysis.

One-fifth of the electroporated cells were seeded into 35-mm dishes. At the times indicated, the medium was replaced by fresh medium, and virus titers in the harvested samples were determined by plaque assay on BHK-21 cells (20). Alternatively, BHK-21, NIH 3T3, or C₇10 cells were seeded into 35-mm dishes and infected at the multiplicity of infection (MOI) indicated in the figures. At the times indicated in the figures, media were replaced and virus titers in the harvested samples were determined by plaque assay on BHK-21 cells.

Analysis of protein synthesis.

One-fifth of the electroporated cells were seeded into 6-well Costar plates. At the times indicated, the cells were incubated for 30 min in 0.8 ml of DMEM lacking methionine but supplemented with 0.1% FBS and 20 μCi/ml of [³⁵S]methionine. After this incubation, they were scraped into the medium, collected by pelleting at 1,500 rpm, and dissolved in 100 μl of standard protein loading buffer. Equal amounts of proteins were loaded onto each lane of the sodium dodecyl sulfate (SDS)-10% polyacrylamide gels. After electrophoresis, the gels were dried, autoradiographed, and analyzed on a Storm 860 PhosphorImager (Molecular Dynamics). The amount of radioactivity detected in the protein band corresponding to actin was used to determine the level of residual host cell protein synthesis. The results were normalized on the amount of radioactivity detected in the same region of the lane containing the lysate of the uninfected cells.

For the analysis of the rate of turnover of the ns polyproteins, BHK-21 cells were infected at an MOI of 20 PFU/cell and metabolically labeled with [³⁵S]methionine in DMEM lacking methionine but supplemented with 0.1% FBS and 20 μCi/ml of [³⁵S]methionine for 30 min at 12 h postinfection (p.i.). The radiolabeled proteins were chased by incubating the cells in the presence of cycloheximide (100 μg/ml) for the indicated times. The ns polyproteins were separated by SDS-10% polyacrylamide gel electrophoresis, and after the gels were dried the protein bands were visualized by autoradiography.

To analyze SINV ns polyprotein processing, BHK-21 cells were infected with different viruses at an MOI of 20 PFU/cell and harvested at 10 h p.i. Equal amounts of protein were separated by SDS-10% polyacrylamide gel electrophoresis, and the nsP2-containing bands were analyzed by Western blotting using SINV nsP2-specific, affinity-purified rabbit antibodies and IRdye 800CW-labeled secondary antibodies (LI-COR). Images were acquired on an Odyssey infrared imager (LI-COR).

Analysis of RNA synthesis.

To analyze the synthesis of the virus-specific RNAs, the infected cells were metabolically labeled with [³H]uridine (20 μCi/ml) in the presence of dactinomycin (ActD) (1 μg/ml) for 4 h, beginning at 3 h posttransfection. Total cellular RNA was isolated by TRIzol, according to the procedure recommended by the manufacturer (Invitrogen), and then denatured with glyoxal in dimethyl sulfoxide and analyzed by agarose gel electrophoresis, using previously described conditions (2). For cellular RNA synthesis, radiolabeling was performed in the absence of ActD, and the RNA analysis was executed as described elsewhere (13). For the quantitative analysis of viral and cellular RNA synthesis, the RNA bands or gel fragments, containing cellular pre-mRNAs, were excised from the 2,5-diphenyloxazole (PPO)-impregnated, dried gels, and the radioactivity levels were measured by liquid scintillation counting.

To analyze the sites of viral RNA synthesis, BHK-21 cells on chambered coverglass slides (Nunc) were infected with recombinant SINVs at an MOI of ∼30 PFU/cell. At 4 h p.i., cells were transfected with bromo-UTP (BrUTP) using the Fugene 6 reagent, as recommended by the manufacturer (Roche). Labeling was performed at 37°C in the presence of ActD for 7 h. The cells were fixed with 3% formaldehyde, permeabilized with 0.5% Triton X-100, and stained with anti-BrdU monoclonal antibodies (Roche) and Alexa 546-labeled secondary antibodies. Images were acquired on a Zeiss 510 META confocal microscope with a ×63, 1.4-numerical aperture oil immersion planapochromal lens.

Isolation of SINV RF core RNA and quantitation of negative-strand synthesis.

The rate of negative-strand RNA synthesis was determined as described elsewhere (4). Infected cells were pulse labeled with 200 μCi of [³H]uridine/ml in DMEM containing 5% FBS and 20 μg of ActD/ml for 1-h periods and were harvested at the end of the pulse period. Deproteinized cell lysates were digested with RNase A and chromatographed on CF-11 cellulose (Whatman, Clifton, NJ) for the isolation of the replicative-form RNA (RF RNA) as described elsewhere (5). Negative-strand RNA was measured in nuclease protection assays that determined the amount of [³H]uridine-RF RNA that, after denaturation, hybridized to an excess (∼100-fold) of unlabeled SINV virion positive-strand RNA. Previously assembled replicative intermediates (RIs)/native RFs active in positive-strand synthesis will have 100% of their radioactive RNA as nascent positive strands, and only those intermediates active in negative-strand synthesis or newly assembled with a negative strand synthesized during the pulse period will have radioactivity in these template RNAs. Thus, finding a value of 20% of the total radioactivity in RF RNA in the negative-strand fraction, which accounts for half of the total RNA species in fully double-stranded RF cores, indicates that 40% of the total negative strands were made during that pulse period (see Fig. 4 for details).

FIG. 4.

Negative- and positive-strand RNA syntheses by SINV/GFP, SINV/2V/GFP and SINV/1V2V/GFP viruses. Beginning at 1 h p.i., cultures of SINV/GFP, SINV/2V/GFP-, and SINV/1V2V/GFP/A-infected BHK-21 cells were pulse labeled with [³H]uridine (200 μCi/ml) in DMEM containing 20 μg of ActD/ml and 5% FBS for 1-h periods. To monitor the effect of translation inhibition, duplicate cultures were incubated in the presence of cycloheximide (100 μg/ml) beginning at 4 h p.i. and were radiolabeled with [³H]uridine at the times indicated in the panels. At the end of each hour-long labeling period, the cultures were harvested in 5% lithium dodecyl sulfate (LDS) in LET buffer (0.1 M LiCl, 50 mM Tris-HCl, pH 7.5, 1 mM EDTA) containing 200 mg/ml of proteinase K. Duplicate samples equivalent to 50,000 cells were analyzed for acid-insoluble radioactivity and were incorporated into RNA. Lysates were deproteinized, DNA was removed with (pH 4) phenol and chloroform, and the RNA collected by ethanol precipitation. The RNA samples were digested with RNase A and chromatographed on CF11 columns to obtain the RF cores of the viral RI/native RFs (see Materials and Methods for details), which were used in nuclease protection assays to determine the amount of radiolabeled negative strands. (A) Total incorporation of [³H]uridine in both negative- and positive-strand virus-specific RNAs. (B) The percentage of the [³H]uridine-labeled negative strands in RF RNA, labeled at the indicated times p.i. for SINV/2V/GFP and SINV/1V2V/GFP/A mutants, continued to synthesize the negative strand at late times p.i. but not in the presence of cycloheximide. (C) Incorporation of [³H]uridine into negative-strand viral RNA at different times p.i. in the presence and absence of cycloheximide in the medium. cpm, counts per minute.

IFN-α/β assay.

The concentrations of IFN-α/β in the media were measured as previously described (42). Briefly, L929 cells were seeded in 100 μl of complete medium at a concentration of 5 × 10⁴ cells/well in 96-well plates and incubated at 37°C for 6 h. Samples of medium harvested from infected NIH 3T3 cells were treated with UV light in the tissue culture cabinet for 1 h and serially diluted in twofold steps directly in the wells with L929 cells. After incubation for 24 h at 37°C, an additional 100 μl of medium with 2 × 10⁵ PFU of vesicular stomatitis virus was added to the wells, and incubation was continued for 36 to 40 h. Cells then were stained with crystal violet, and the end point was determined as the concentration of IFN-α/β required for the protection of 50% of the monolayer cells from vesicular stomatitis virus-induced CPE. The IFN-α/β standard for the normalization of the results was obtained from the NIH.

RESULTS

SINV variants with mutated cleavage cites in ns polyprotein.

The construction of SINV genome derivatives, which carried mutated cleavage sites between nsP2 and nsP3 or between nsP1, nsP2, and nsP3, has been described previously (21, 22, 37). These viral genomes encoded P123 polyproteins in which the wild-type (wt) Gly at the P2 positions of one or both of the cleavage sites had been replaced by Val. The indicated mutations inactivated either nsP2/nsP3 or both nsP1/nsP2 and nsP2/nsP3 cleavage sites and, thus, caused the incomplete processing of the ns polyprotein (37). We additionally modified these genomes by cloning a green fluorescent protein (GFP) sequence after aa 389 of the nsP3 protein (Fig. 1A). The GFP insertion did not affect the replication efficiency of the recombinant, cleavage-competent viruses (3, 8) and provided an opportunity to analyze the formation of the nsP complexes in live cells at different times p.i. or after the transfection of in vitro-synthesized RNAs. In addition, the UGA codon between nsP3 and nsP4 was mutated to the Cys-coding UGU to enhance the production of the nsP4 polymerase subunit proteins and, thereby, to possibly increase the level of viral RNA synthesis. This was not an unusual modification of the ns polyprotein, because a variety of alphaviruses, particularly the natural isolates, do not have a terminating codon between nsP3 and nsP4 (23, 38, 41). The final constructs SINV/2V/GFP and SINV/1V2V/GFP are presented in Fig. 1A. They were incapable of nsP2/nsP3 or nsP1/nsP2 and nsP2/nsP3 site processing and were expected to produce P23/GFP or P123/GFP, respectively, as ns polyprotein cleavage products. None of the cleavage sites in the genome of the control SINV/GFP virus were changed (Fig. 1A).

FIG. 1.

Replication of SINV variants that have mutations in the ns polyprotein cleavage sites. (A) The schematic representation of the recombinant SINV genomes, infectivity of the in vitro-synthesized RNAs, and sizes of the plaques in the infectious center assay. Plaques were stained at 2 days after RNA transfection. (B) Analysis of SINV structural protein expression after the transfection of the in vitro-synthesized RNAs into BHK-21 cells. Proteins were metabolically labeled by [³⁵S]methionine at 23 h posttransfection of the RNAs and analyzed by electrophoresis in an SDS-polyacrylamide gel, as described in Materials and Methods. The positions of SINV structural proteins are indicated on the left side of the gel. (C) Analysis of viral RNA synthesis in BHK-21 cells. Cells were electroporated with the in vitro-synthesized RNA. RNAs were metabolically labeled with [³H]uridine in the presence of ActD for 4 h, beginning at 3 h posttransfection, and were analyzed on the gel as described in Materials and Methods. (D) Virus replication in BHK-21 cells. Cells were electroporated by in vitro-synthesized RNAs, media were replaced at the indicated times, and virus titers were determined by plaque assay. The results presented in all of the panels were generated by using aliquots of the cells transfected by 1 μg of the RNAs.

The in vitro transcripts were transfected into BHK-21 cells, in which they were found to have similar levels of infectivity by an infectious center assay (Fig. 1A). This finding led us to the conclusion that the cleavage-defective mutants could replicate without additional (adaptive) mutations. However, while SINV/GFP and SINV/2V/GFP formed similarly sized, large plaques after electroporation (Fig. 1A), the SINV/1V2V/GFP variant developed only pinpoint plaques, indicating that the replication of this mutant was less efficient. The plaque sizes correlated with the expression levels of the viral structural proteins (Fig. 1B) and the synthesis of virus-specific RNAs (Fig. 1C) in electroporated cells. At 23 h posttransfection, SINV/1V2V/GFP produced barely detectable levels of structural proteins (Fig. 1B), and viral RNA synthesis could be detected only after a very long exposure of the gel (data not shown). We also assessed the rates of infectious virus accumulation in the medium after the electroporation of the various SINV RNAs into BHK-21 cells and found that the SINV/2V/GFP variant gave yields that were similar to those of SINV/GFP (Fig. 1D), but the yield of SINV/1V2V/GFP was greatly reduced. Moreover, we observed that samples of SINV/1V2V/GFP recombinant virus, harvested at different times, produced heterogeneous plaques, suggesting the possibility that it was acquiring adaptive changes that enhanced virus replication.

Thus, our initial experiments showed that the complete processing of the ns polyprotein was not an absolute prerequisite for SINV replication, and the nsP2/nsP3 cleavage site mutant replicated in BHK-21 cells almost as efficiently as the cleavage-competent SINV/GFP. However, cleavage at the nsP1/nsP2 site was important, as a variant bearing a mutation in this cleavage site was strongly affected for replication.

Selection of efficiently replicating SINV/1V2V cleavage mutants.

This study was not our first attempt at selecting SINV P123 cleavage mutants that are capable of efficient replication while retaining the cleavage-defective phenotype. These mutants always evolved to an efficiently replicating phenotype by restoring the ability to cleave the nsP1/nsP2 site. Therefore, in our current study, we additionally modified the selection procedure. We expected that any SINV/1V2V/GFP variants that acquired adaptive mutations (outside of the cleavage sites) would form larger plaques than did the original SINV/1V2V/GFP and exhibit nsP3/GFP distribution more diffuse in pattern than that of the fully processed nsP3/GFP, which normally is found as punctate clusters (8). These selection criteria could facilitate the isolation of efficiently replicating SINV/1V2V variants, producing unprocessed P123.

In the following experiments, we randomly selected four plaque-purified viruses from the SINV/1V2V/GFP stocks harvested at early times posttransfection (Fig. 1D). The distinguishing feature of the chosen plaques was a more diffuse GFP fluorescence in the infected cells at early times p.i. However, only one of the four plaque-purified isolates produced unprocessed P123/GFP, as assayed by Western blotting with GFP-specific antibodies, while three of the four isolates had regained the ability to process the nsP1/nsP2 cleavage site (data not shown). The latter viruses were not further investigated. Sequence analysis of the genome RNA of the isolate expressing only uncleaved P123/GFP revealed no mutations in the 5′- and 3′-untranslated regions (this analysis was undertaken to rule out the possibility that adaptive mutations altered promoter regions for either negative- or positive-strand RNA synthesis) or in the nsP1 to nsP3 genes but yielded two changes in the nsP4-coding sequence: Y₁₀₆→F and E₄₅₁→A. To probe their effect on virus replication, each mutation alone or both of them together were transferred into the genome of the original SINV/1V2V/GFP (Fig. 2A), and the phenotypes of the resulting recombinant viruses were assayed. The nsP4 E₄₅₁→A mutation had a strong positive effect on virus replication. The SINV/1V2V/GFP/A variant formed large, homogeneous plaques that were comparable to those of the SINV/GFP virus, and it replicated to titers approaching 10⁹ PFU/ml (Fig. 2A and B). The change of nsP4 Y₁₀₆→F did not enhance virus replication when it was introduced, either alone or in combination with E₄₅₁→A (Fig. 2A), so it was not further studied.

FIG. 2.

Analysis of the effects of adaptive mutations in nsP4 on SINV/1V2V/GFP replication. (A) The schematic representation of SINV/1V2V/GFP genomes that have mutations in nsP4, the infectivity of the in vitro-synthesized viral RNAs, and virus titers at 24 h posttransfection of BHK-21 cells. (B) Sizes of the plaques, formed by the original SINV/1V2V/GFP and its derivative SINV/1V2V/GFP/A in the infectious center assay, following staining at 48 h posttransfection. (C) The replication of SINV/GFP and SINV/1V2V/GFP/A viruses in BHK-21 cells. Cells were electroporated by 1 μg of in vitro-synthesized RNAs, media were replaced at the indicated times, and virus titers were determined by plaque assay. (D) Analysis of SINV structural protein expression after the transfection of the in vitro-synthesized RNAs into BHK-21 cells. Proteins were metabolically labeled by [³⁵S]methionine at 23 h posttransfection of the RNAs and analyzed by electrophoresis in SDS-polyacrylamide gels as described in Materials and Methods. (E) Sequence alignment of the nsP4 fragment that contained an adaptive mutation in the selected SINV/1V2V/GFP variant. EEEV, eastern equine encephalitis virus (43); see references 41 and 19 for SFV and VEEV, respectively. Residues identical to those in the SINV sequence are indicated by dashes. The conserved GDD sequence is indicated by an open box.

The SINV/1V2V/GFP/A variant demonstrated efficient replication after the transfection of the in vitro-synthesized RNA (Fig. 2C) that correlated with a higher level of viral structural protein synthesis (Fig. 2D). The nsP4 E₄₅₁→A adaptive substitution is in close proximity to the GDD motif in the SINV polymerase subunit (Fig. 2E), raising the possibility that it is increasing polymerase efficiency. Interestingly, the same E₄₅₁→A mutation had no detectable effect on the replication of cleavage-competent virus. The original SINV/GFP and its variant, carrying the indicated mutation in nsP4, replicated at exactly the same rates and to identical final titers (data not shown).

Thus, at this stage of the experiments, we had obtained an efficiently replicating SINV mutant that was incapable of P123 processing. We next characterized its replication and interactions with host cells.

Replication of ns polyprotein cleavage mutants of SINV.

To compare the replication efficiencies of the cleavage-defective mutants, BHK-21 cells were infected at the same MOI with SINV/GFP, SINV/2V/GFP, and SINV/1V2V/GFP/A viruses (Fig. 3A and B). While all three viruses produced titers of 10⁸ to 10⁹ PFU/ml, at the MOI used for infection, there was a clearly noticeable delay in the release of the two cleavage-deficient viruses. Western blot analysis of infected cell lysates at 10 h p.i. confirmed the presence of only P23/GFP and P123/GFP in the SINV/2V/GFP- and SINV/1V2V/GFP/A-infected cells, respectively (Fig. 3C). Mature, fully cleaved nsP2 was not detected, even when using the most sensitive settings on the LI-COR imager.

FIG. 3.

Comparative analysis of the replication of SINV ns polyprotein cleavage mutants. (A) The schematic representation of the genomes of the recombinant viruses. (B) Virus replication in BHK-21 cells. Cells were infected at an MOI of 10 PFU/cell; media were replaced at the indicated times, and virus titers were determined by plaque assay. (C) Analysis of the ns polyprotein processing in BHK-21 cells, which were infected at an MOI of 20 PFU/cell and harvested at 10 h p.i. The lysates were analyzed by Western blotting using SINV nsP2-specific antibodies as described in Materials and Methods. Images were acquired on a LI-COR imager. (D) Analysis of ns polyprotein degradation. BHK-21 cells were infected at an MOI of 20 PFU/cell and metabolically labeled by [³⁵S]methionine for 30 min at 12 h p.i., and then they were incubated in the presence of cycloheximide (100 μg/ml) for the indicated times. The ns polyproteins were analyzed by SDS-10% polyacrylamide gel electrophoresis as described in Materials and Methods. (E) Analysis of virus-specific RNA synthesis in the cells infected with different mutants. BHK-21 cells were infected at an MOI of 20 PFU/cell. RNAs were metabolically labeled with [³H]uridine in the presence of ActD for 4 h, beginning at 3 h p.i. RNAs were isolated and analyzed by agarose gel electrophoresis under denaturing conditions. The RNA bands were excised from the dried gels, and the radioactivity was measured by liquid scintillation counting. CPM, counts per minute.

As shown in Fig. 3E, both cleavage-defective mutants demonstrated reduced efficiency for 26S mRNA synthesis compared to parental levels, and the SINV/2V/GFP virus replicated 49S genome RNA twice as efficiently as the parental SINV/GFP virus. The molar ratio of SG-to-genomic (G) RNA synthesis (SG:G) was less than 3 for SINV/2V/GFP and was 8 for SINV/GFP. Surprisingly, the SINV/1V2V/GFP/A variant synthesized amounts of genome RNA that were similar to those of SINV/GFP, but its SG:G molar ratio of ∼1 indicated that the P123-containing RCs were specifically defective in the internal promoter recognition needed for 26S RNA transcription.

It has been found that nsP2 expresses functions that are involved in the regulation of the host response to infection (6, 9, 11, 26, 32). One of its regulatory activities might be an overall cessation of negative-strand synthesis that is observed at 4 to 8 h p.i., depending on the cell type (33, 35). This cessation defines the end of the early phase, a time when the infected cells are permissive for the assembly of new viral RCs, and the start of the late phase, when no additional RCs are formed and, as determined by the synthesis of viral G and SG RNAs, the rates of viral RNA synthesis become constant and maximal. To determine if polyprotein forms of nsP2 could activate a similar sequence of events, we further tested SINV/2V/GFP and SINV/1V2V/GFP/A RNA replication in BHK-21 cells and monitored the pattern of negative-strand RNA synthesis by the cleavage-defective mutants. The infected cultures were pulse labeled with [³H]uridine in the presence of ActD for 1-h periods, beginning at 1 h p.i. For all three viruses, positive-strand synthesis, which accounts for 95% or more of overall RNA synthesis, reached maximal rates at 4 to 5 h p.i. It then continued at close to these rates until 10 to 12 h p.i. in SINV/GFP- and SINV/2V/GFP-infected cells but slowly declined in the cells infected with the SINV/1V2V/GFP/A variant (Fig. 4A). Importantly, the cleavage-defective mutants demonstrated a significant loss in overall RNA synthesis with time in the presence of cycloheximide (Fig. 4A). The rate of loss for the SINV/1V2V/GFP/A mutant-specific RCs was immediate and was ∼10-fold after 4 h of cycloheximide treatment, while the SINV/2V/GFP mutant's RCs retained the activity at the same level as that at the time of the addition of the inhibitor for ∼2 h and then showed inactivation kinetics similar to that of SINV/1V2V/GFP/A-specific RCs. In contrast, the RCs formed in SINV/GFP-infected cells retained RNA synthesis activity in the presence of cycloheximide (Fig. 4A). As an initial approach to explore the phenomenon of RC inactivation, we tested the stability of the uncleaved ns polyproteins in pulse-chase experiments. In the infected cells, proteins were metabolically labeled with [³⁵S]methionine (see Materials and Methods for details), after which the cultures were incubated in complete DMEM supplemented with 100 μg/ml of cycloheximide. Cells were harvested after different times of incubation in cycloheximide-containing medium, and the ns polyproteins were analyzed on a polyacrylamide gel (Fig. 3D). Similar amounts of radiolabeled P23/GFP and P123/GFP proteins were detected even after a 4-h-long chase in the presence of the indicated translation inhibitor. These findings suggested that the detected 10-fold loss of RC activity does not result from the proteolytic degradation of their P23 or P123 components. However, these experiments cannot rule out the possibility that the unprocessed nsPs are present in the cells in a form other than that of RC and are less susceptible than RCs to proteolytic cleavage.

During alphavirus infection, the rates of [³H]uridine incorporation into virus-specific RNAs depend on the accumulation of viral RCs and negative-strand RNAs, which appear to be present in these RCs as components of the RIs/RFs. Therefore, the above-described experiments indicated that cleavage mutants and SINV/1V2V/GFP/A, in particular, might differ from SINV/GFP with respect to RI/RF accumulation in the course of replication. To test this possibility, in complementary experiments the virus-specific RNAs were pulse labeled with [³H]uridine at different times p.i. RNase-resistant cores of the RI/RFs then were isolated (see Materials and Methods for details), and the contents of radiolabeled negative-strand RNA were analyzed in nuclease protection assays (Fig. 4B and C). Negative-strand synthesis was readily detected in both parental SINV/GFP- and mutant-infected cells, and each showed similar rates of synthesis early in infection. However, in contrast to SINV/GFP, both mutants failed to stop negative-strand RNA synthesis (Fig. 4B and C), and, even at late times p.i., cells infected with either mutant maintained high rates of negative-strand production. After being pulse labeled with [³H]uridine, ∼20% of the total radiolabeled RF core RNA was in negative strands from 5 h p.i. onward, indicating that ∼40% of all of the templates in the active RI/RF population were newly synthesized within each hour (Fig. 4B). Interestingly, the ability to continue negative-strand synthesis depended on a continuous supply of nascent ns polyproteins, and it was rapidly inhibited in the presence of cycloheximide, particularly in the case of SINV/1V2V/GFP/A infection (Fig. 4B and C). This was a strong indication that the P23- or P123-containing RCs switched from their initial negative-strand polymerase activity to solely positive-strand synthesis, as previously described for wt SINV, which was capable of the complete processing of P123 (21, 22, 37).

In additional experiments, we also tested whether the unprocessed protein complexes still were able to form large, higher-order structures. At early times p.i. (2 to 6 h p.i.), SINV/1V2V/GFP/A replication led to a diffuse distribution of P123/GFP throughout the cytoplasm. Nevertheless, by 11 h p.i., the uncleaved ns polyproteins assembled into higher-order protein structures with a morphology similar to that of complexes in SINV/GFP-infected cells (Fig. 5). These complexes contained virus-specific RNAs that were readily labeled by BrUTP in the presence of ActD, which suggested that they represent sites of viral RNA synthesis (Fig. 5) or, at the least, sites of the storage and/or translation of newly synthesized virus-specific RNAs. These large protein complexes, containing uncleaved P23/GFP, also were similar to those previously reported for SINV/2V/GFP (8).

FIG. 5.

ns polyprotein complexes formed in SINV/1V2V/GFP/A-infected cells are involved in viral RNA synthesis. BHK-21 cells were infected with SINV/GFP and SINV/1V2V/GFP/A viruses at an MOI of ∼30 PFU/cell. The metabolic labeling of virus-specific RNAs with BrUTP (at 4 h p.i. for 7 h) and staining with BrdU-specific antibodies was performed, as described in Materials and Methods. Images were acquired on a Zeiss 510 META confocal microscope. Bars correspond to 20 μm.

In summary, our current findings show that ns polyprotein processing is not an absolute prerequisite for efficient SINV RNA synthesis and virus production in BHK-21 cells. In the case of the 1V2V ns polyprotein mutant, only one additional substitution in the nsP4 protein sequence was required to strongly increase RNA replication (but not the transcription of the SG RNA) to the wt level. At the same time, the transcription of the SG RNA in the SINV/1V2V/GFP/A-infected BHK-21 cells was sustained sufficiently for the production of viral structural proteins to the level required for virus release at titers comparable to those of the cleavage-competent virus. The results raised the intriguing issue of why polyprotein processing evolved as a common feature in alphaviruses, thereby causing us to refocus the investigation on the virus-host interaction that occurs following infection.

Cleavage-defective mutants failed to downregulate cellular transcription.

Previous studies showed that SINV and SFV nsP2 proteins play a role in host interactions and are the only ns proteins to be transported to the nucleus in infected cells (1, 29, 30). In the latter compartment, this ns protein plays a critical role in the inhibition of cellular transcription and ultimately in CPE development (11, 12). Therefore, a failure of the P23 and P123 cleavage-defective mutants to release mature nsP2 proteins could limit the ability of SINV to modify the intracellular environment. To test this hypothesis, we infected BHK-21 cells with each of the nsP3/GFP-encoding viruses and analyzed protein and RNA synthesis at different times p.i. As we previously described (13), parental SINV/GFP efficiently downregulated cellular translation and transcription (Fig. 6). SINV/2V/GFP was capable of inhibiting the translation of cellular mRNAs at rates similar to those of SINV/GFP, but host cell transcription was inhibited more slowly and had rates similar to those previously detected in the cells treated with the translation inhibitor puromycin (13).

FIG. 6.

ns polyprotein cleavage mutants of SINV differ from the SINV/GFP variant in their ability to inhibit cellular translation and transcription. (A) Inhibition of translation in the infected cells. BHK-21 cells were infected with the indicated viruses at an MOI of 20 PFU/cell. At the indicated times, cells were metabolically labeled with [³⁵S]methionine as described in Materials and Methods, and equal amounts of protein were analyzed on the SDS-10% polyacrylamide gels. (B) Gels were dried, and the level of residual protein synthesis was determined on a PhosphorImager by measuring radioactivity in the actin band. Data were normalized to the radioactivity in the actin band in the mock-infected cells. The positions of the protein markers, viral structural proteins, and actin are indicated. (C) RNA synthesis in the virus-infected cells. BHK-21 cells were infected at an MOI of 20 PFU/cell and, at the indicated times, RNAs were metabolically labeled with [³H]uridine. RNAs were isolated, and equal amounts were analyzed by electrophoresis under denaturing conditions as described in Materials and Methods. The positions are indicated for viral G and SG RNAs, preribosomal 45S, ribosomal 28S and 18S, and pre-mRNAs. The gel fragments, corresponding to pre-mRNAs, were excised from the dried gel, and radioactivity was measured by liquid scintillation counting. (D) These levels of poly(A) RNA synthesis were normalized to that found in the mock-infected cells. All of the SINV/GFP- and SINV/2V/GFP-infected cells were dead by 30 h p.i.; therefore, after this time point, no RNA labeling for these variants was performed. More than 95% of the cells in the monolayers were GFP positive by 8 h p.i., indicating that the residual protein and RNA synthesis did not result from the uninfected cells.

The SINV/1V2V/GFP/A variant showed different and more unusual (for SINV) changes in cell biology. First, it was observed that, based on nsP3/GFP expression, all of the BHK-21 cells in a culture could be infected with this mutant. By 8 h p.i., the infected cells underwent a profound inhibition of translation (Fig. 6A and B) and of the transcription of ribosomal RNAs (Fig. 6C and D); poly(A)-containing mRNA always remained at ∼60% of the level of uninfected cells. However, the initial decline was followed by the restoration of both cellular protein synthesis and the transcription of the rRNAs to levels comparable to those in the mock-infected cells (Fig. 6A and B). At 18 h p.i., these cultures no longer produced virus-specific RNA and ns polyproteins (Fig. 6A and C), although the translation of SINV capsid, p62, and E1 continued. In spite of an obvious decrease in virus RNA replication and the significant restoration of host cell protein and RNA syntheses, SINV/1V2V/GFP/A-infected cells developed CPE and died.

Previous studies indicated that the synthesis of SINV structural proteins played a critical role in cell death (7). Therefore, we also tested the ability of the uncleaved P123/GFP to contribute to cell death and CPE development when produced in cells in the absence of the viral structural proteins. We transfected BHK-21 cells with a SINV/1V2V/GFP/A-based replicon (SINrep/1V2V/GFP/A/PAC), in which the viral structural genes were replaced by PAC to allow us to use resistance to puromycin and select replicon-expressing cells. After electroporation, cells were treated for 36 h with puromycin to eliminate any untransfected cells, and then the remaining cells were grown in the presence or absence of puromycin. In the absence of puromycin, BHK-21 cells resumed growth after a lag period, and within 3 to 4 days after electroporation, they had cleared the viral replicons. Under puromycin selection, almost all of the cells died within the next few days, with fewer than 0.01% of the cells forming Pur^r foci. The results showed that even BHK-21 cells deficient in IFN-α/β signaling were capable of stopping the synthesis of virus- or replicon-specific SINV RNAs when only P123 and nsP4, and not mature nsPs, were present in the infected cell. The cells that were transfected with the control PAC-expressing replicon SINrep/GFP/PAC, encoding cleavage-competent ns polyprotein, died within a few days of electroporation regardless of having received puromycin treatment. These data were in complete agreement with our previously published studies, in which we demonstrated the cytotoxic effect of SINV nsP2 expression (6, 11).

The ns polyprotein cleavage mutants induce a strong IFN-α/β response.

The inability of the SINV/1V2V/GFP/A cleavage mutant to inhibit cellular transcription suggested to us that the cleavage-defective phenotype might have a deleterious effect on viral replication in cells fully competent for IFN-α/β induction and signaling. To test this hypothesis, we infected NIH 3T3 cells with SINV/GFP, SINV/2V/GFP, and SINV/1V2V/GFP/A viruses and assessed their replication rates, CPE development, and IFN-α/β release. The SINV/GFP variant efficiently replicated in these cells (Fig. 7A), caused a complete CPE within 24 h p.i., and induced IFN-α/β at a barely detectable level (Fig. 7B). SINV/2V/GFP also demonstrated efficient replication, with the final titers approaching 10⁹ PFU/ml, and induced the rapid secretion of IFN-α/β to amounts that were almost three orders of magnitude higher than that produced by SINV/GFP-infected cells. These cultures showed a robust CPE by 2 days p.i. However, uninfected cells, which were present in low numbers (<1%) in the SINV/2V/GFP-infected monolayers, were protected by the released IFN-α/β. They demonstrated no P23/GFP expression and continued to grow in spite of the presence of high levels of virus in the medium.

FIG. 7.

ns polyprotein cleavage-deficient mutants induced high levels of IFN-α/β. (A) Replication of the indicated SINV variants in NIH 3T3 cells and (B) induction of IFN-α/β. NIH 3T3 cells were infected with the designed viruses at an MOI of 20 PFU/cell. Media were replaced at the indicated times. Virus titers were determined by plaque assay on BHK-21 cells. The same samples were used to assess the IFN-α/β concentration in the biological assay as described in Materials and Methods. (C) Monolayers of NIH 3T3 cells in the 6-well Costar plates were infected with the indicated doses of recombinant viruses (stock titers were determined on BHK-21 cells). Staining with crystal violet was performed at 48 h p.i.

In the experiment presented in Fig. 7A, the SINV/1V2V/GFP/A variant also infected almost all of the cells. Within 6 h p.i., more than 95% of cells became GFP positive (data not shown) and, thus, were translating nsPs and replicating the virus. However, infectious virions were produced very inefficiently (Fig. 7A). The SINV/1V2V/GFP/A-infected NIH 3T3 cells released a high level of IFN-α/β (Fig. 7B), which protected uninfected cells in the culture and also eventually stopped virus replication in the entire monolayer (Fig. 7A). No morphological changes that are characteristic of SINV replication in vertebrate cells were ever detected. Within 3 days p.i., all cells in the culture ceased the expression of P123/GFP and virion production and grew as efficiently as uninfected cells.

To confirm these results, we also infected monolayers of NIH 3T3 cells with different amounts of virus to monitor cell-to-cell spreading. SINV/GFP spread efficiently and caused a complete CPE of the monolayer, even if the culture was infected with fewer than 1,000 PFU (i.e., an MOI of less than 10⁻³ PFU/cell) (Fig. 7C and data not shown). This was in contrast to results with SINV/2V/GFP. At MOIs of 1 or less, SINV/2V/GFP killed all of the initially infected NIH 3T3 cells but was incapable of spreading in the presence of the IFN-α/β secreted into the medium from infected cells. SINV/1V2V/GFP/A did not cause any CPE in NIH 3T3 cultures regardless of the amounts of virus used for infection (Fig. 7C).

To demonstrate that the activation of the IFN-α/β-inducible genes played a significant, negative role in the replication and spread of cleavage-deficient mutants, we infected IFN-α/βR^−/− MEFs with SINV/2V/GFP and SINV/1V2V/GFP/A viruses at MOIs of 0.1 and 0.01 PFU/cell. Within 48 h p.i., both viruses infected all of the cells in the monolayers (as detected by nsP3/GFP expression) and ultimately caused a complete CPE (data not shown). Final titers of the released viruses approached 2.5 × 10⁸ PFU/ml.

The results of these experiments supported the conclusion that the ability of SINV to process the ns polyprotein was critical for the virus to inhibit IFN-α/β induction and, thereby, the activation of IFN-α/β-inducible genes, the products of which interfered with the productive replication of the cleavage mutants.

Cleavage mutants are defective in replication in mosquito cells.

The circulation of SINV in nature requires its replication in cells of both vertebrate and mosquito origin. Therefore, we evaluated the replication of the cleavage-defective viruses in cultured mosquito C₇10 cells. While the cleavage-competent SINV/GFP variant replicated to titers higher than 10⁹ PFU/ml, the SINV/2V/GFP and SINV/1V2V/GFP/A viruses produced three and four orders of magnitude fewer infectious virions, respectively (Fig. 8). Even though a similar MOI was used, the mutant viruses established productive infection in very few cells (as determined by GFP expression). They also appeared to produce virus very inefficiently, because the titers were lower than one might expect on the basis of the numbers of infected cells.

FIG. 8.

Replication of the SINV cleavage-defective variants in the cells of mosquito origin. C₇10 cells were infected with the indicated viruses at an MOI of 1 PFU/cell. (This dose was determined based on the ratio of SINV Toto1101 titers in BHK-21 and C₇10 cells, which were determined in other experiments.) At the indicated times, media were replaced and the titers of the released viruses determined by plaque assay on BHK-21 cells.

Thus, the cleavage-deficient viruses exhibited the phenotype of host range mutants. Their replication was defective in mosquito cells but not in vertebrate cells, suggesting that vertebrate cells can provide factors to enhance the activity of RCs composed of uncleaved SINV ns polyproteins. Alternatively, the mosquito cells also might have an antiviral response, different from that of the vertebrate cells, that interferes with the replication of the above-described mutants.

DISCUSSION

One of the characteristic features of SINV replication is the sequential, regulated processing of the ns polyprotein precursors P1234 and P123. This cleavage cascade results in a gradual change in the viral polymerase activities and the eventual formation of the mature RC, containing individual nsP1 to nsP4, that efficiently produce genomes and SG RNA (22, 37, 40). The major conclusions from these investigations were that (i) a failure to cleave the nsP3/nsP4 site left the polymerase inactive, (ii) the intermediate P23 and P123 forms were needed for negative-strand synthesis, and (iii) cleavage at the nsP2/nsP3 site switched off negative-strand polymerase activity and activated internal promoter recognition for the transcription of SG RNA (21, 22, 37). Studies of mutant alphaviruses unable to undertake nsP2/nsP3 cleavage were the first to hint that the inability to process P23 could make virus replication more efficient (13, 18). The present investigation was carried out to further explore this phenomenon, as well as to probe the roles of polyprotein cleavage in virus-host interaction and infection outcome.

Our study showed that SINV genomes, producing P123 or P23 forms in place of the mature, cleaved nsPs, were viable and could replicate efficiently in cultured mammalian cells. The isolation of replication-competent SINV that was incapable of producing fully processed nsP1, nsP2, and nsP3 was a bit surprising, given that ns polyprotein processing is conserved among alphaviruses. The cleavage-defective SINV/1V2V/GFP/A-specific nsPs efficiently replicated the viral genome but were defective in transcribing the 26S SG RNA that encodes the viral structural proteins, and they were incapable of shutting off minus-strand RNA synthesis. The latter phenomenon can be attributed to their inability to produce free nsP2, which normally accumulates in cytoplasm, and, at late stages of wt SINV infection, cleaves the newly synthesized P123 before it assembles the RCs. This possibility is indirectly supported by the finding that the duplication of SINV nsP2 by cloning it under the control of the second SG promoter is lethal for the cleavage-competent SINV/GFP but not for SINV/2V/GFP and SINV/1V2V/GFP/A variants (data not shown). The second major finding of this study was that the failure to fully process the ns polyproteins and, thus, to release nsP2 in a form that could traffic to the nucleus (9, 11, 28, 29) led to a failure of the viruses to prevent antiviral host responses in the cells that were competent for IFN-α/β induction and signaling. This failure made mutant viruses incapable of spreading to adjacent, uninfected cells and made at least NIH 3T3 cells capable of clearing the infection without noticeable cell death.

It is important that SINV with mutated cleavage sites between nsP1, nsP2, and nsP3 required only a single adaptive mutation in nsP4 to gain an efficiency of RNA synthesis comparable to that exhibited by RCs formed by SINV/GFP and SINV/2V/GFP. In infected cells, the unprocessed P123/GFP polyproteins expressed by SINV/1V2V/GFP/A were capable of forming large complexes having a morphology similar to those of the complexes containing nsP3/GFP and P23/GFP, and found in SINV/GFP- and SINV/2V/GFP-infected cells, respectively (Fig. 5 and data not shown). They colocalized to sites of SINV RNA synthesis or accumulation. Both the P23- and P123-containing RCs functioned in a manner similar to that of wt, cleavage-competent SINV RCs in the positive-strand genome synthesis. Thus, the P123-containing RCs could engage in genome synthesis in the context of SINV, not only when such ns proteins were translated in excess from vaccinia virus-derived mRNAs (21, 22). This indicated that, in the presence of nsP4 containing the E₄₅₁→A mutation, P123 was able to establish the conformation and interactions necessary to efficiently utilize promoters for genome synthesis on the negative-strand templates, unlike what was seen for wt nsP4 and P123 RCs (37) (Fig. 1). Thus, the modified nsP4 polymerase subunit may mimic or compensate for changes that occur following cleavage at the nsP1/nsP2 site.

We were not surprised to find that reduced SG RNA synthesis was a phenotype shared by the P23 and P123 cleavage-defective mutants. Evidence supporting a role for nsP2 in SG RNA synthesis was previously published (21, 22, 37), and the amino acid substitutions, especially at residues in the carboxy-terminal fragment of the protein, which confer temperature sensitivity to SG RNA synthesis, were identified (17, 34). A recent crystal structure of the carboxy-terminal fragment of the Venezuelan equine encephalitis virus (VEEV) nsP2 (31) allowed some temperature-sensitive lesions to be modeled. Taken together, the data suggest that the nsP2 terminus might require conformational changes, normally achieved by P23 cleavage, that favor promoter recognition and promote G and SG RNA synthesis from the negative-strand RNA template (33, 35). Our present results indicate that at least some of the conformational changes can occur in uncleaved P23 and P123 polyproteins, as both mutant RCs converted to solely positive-strand synthesis and lost negative-strand synthesis under conditions of translation inhibition with kinetics similar to that observed for wt SINV RCs (Fig. 4).

Thus, on the one hand, both mutants continuously formed new RCs (negative-strand synthesis is the first activity acquired by new ns complexes), a feature of replication that usually occurs only in the first several hours p.i. with wt SINV (32, 33, 35). On the other hand, the cleavage-defective mutants were unable to prevent the slow inactivation of viral RCs, which were engaged in positive-strand synthesis, and to inhibit the transcription of host mRNA or to irreversibly downregulate cellular translation. The continuous negative-strand synthesis and the loss of activity by mature RCs also were previously seen for BHK-21 cells persistently infected with replicons expressing mutant nsP2 proteins (32) and in the wt SINV-infected MEFs, which were deficient in the latent host RNase, RNase L (36). The similar pattern is intriguing and supports a model in which functions expressed in the nucleus by cleaved nsP2 block antiviral responses aimed at RC destabilization. However, this hypothesis needs additional experimental support.

Thus, the cleavage of the ns polyprotein is not absolutely essential for the regulation of RNA synthesis in SINV-infected cells. After acquiring a very limited number of adaptive mutations (in our study, a single mutation), cleavage-deficient viruses can replicate their genomes and transcribe SG RNA with an efficiency sufficient for effective virus production. However, the cleavage-defective mutants demonstrated profound host range phenotypes. In cells with defects in IFN-α/β signaling (BHK-21 cells [Fig. 3] and IFN-α/βR^−/− MEFs [data not shown]), the cleavage-deficient viruses were capable of replication and spreading, as were the cleavage-competent virus SINV/GFP. Both cell types produced high titers of SINV/2V/GFP and SINV/1V2V/GFP/A variants, although, in BHK-21 cells, the replication of SINV/1V2V/GFP/A was limited to ∼18 h p.i., after which time host transcription and translation were restored to 60% of the level detected in the mock-infected cells, and further virus replication was inhibited (Fig. 6). These mutants demonstrated strong defects in the ability to interfere with the cellular response and to inhibit the transcription of cellular mRNA. This ultimately leads to the activation of cellular antiviral genes. The effect of P123 cleavage deficiency on virus replication was more profound in cells that had no defect in IFN-α/β induction and signaling. While the infection of NIH 3T3 cells with either cleavage-defective variant rapidly induced type I IFN to levels that were approximately three orders of magnitude higher than that in SINV/GFP-infected NIH 3T3 cultures, the P123 mutant replicated very inefficiently and did not induce any CPE and/or noticeably interfere with cell growth. Ultimately, replicating virus was cleared from the cultures within 3 days of infection. Among the alternative possibilities, the hypothesis that we currently favor is that such intense differences in the replication of wt and cleavage-deficient viruses result from their failure to inhibit host transcription. It also is possible that the continuous synthesis of negative-strand RNA by unprocessed, P123- and P23-containing RC, formed by SINV/2V/GFP and SINV/1V2V/GFP/A viruses, leads to the prolonged production of virus-specific double-stranded RNA. The latter event might additionally contribute to the induction of both an antiviral response and efficient virus clearance.

The nsP2-dependent transcriptional shutoff appears to be at least one of the mechanisms that the Old World alphaviruses (SINV and SFV are the best-studied representatives) apply to interfere with the virus-induced cell response and activation of the IFN-inducible genes (11, 12). The accumulated data suggest that the efficiency of interference depends on the production of the free form of nsP2 that is required for its functioning both in the nucleus and cytoplasm. It also should be noted that the nonspecific inhibition of the expression of all of the cellular genes does not exclude the possibility that SINV and/or other Old World alphaviruses have developed additional, more specific mechanisms of interfering with the cellular antiviral response; therefore, alphavirus-host cell interactions need further investigation. Moreover, the results of the study do not appear directly applicable to all of the alphaviruses, e.g., the New World alphaviruses. Our data from another line of research strongly suggest that VEEV nsP2 functions less efficiently in the inhibition of cellular transcription, and, instead, its structural protein, capsid protein, is involved in the development of cellular transcriptional shutoff and CPE (10-12, 27). While encoded by different open reading frames, the retention by both groups of alphaviruses of proteins that have the transcription inhibition function strongly suggests its importance to the natural biology of these pathogens.

Importantly, the cleavage-deficient SINV mutants demonstrated a greatly reduced ability to replicate in cells of mosquito origin (Fig. 8). This, together with their defects in inhibiting cellular, antiviral functions and the development of spreading infection in cells fully competent in IFN-α/β production, would significantly limit the ability of ns polyprotein cleavage-defective mutants to circulate in nature as arthropod-borne infectious agents involving mosquito vectors and vertebrate hosts.

In conclusion, the results of this study demonstrate the following. (i) SINV ns polyprotein cleavage is not an absolute prerequisite of viral RNA replication. Cleavage-deficient viruses are capable of producing the enzyme complexes that function in the synthesis of virus-specific RNAs. (ii) The cleavage-deficient mutants remain capable of continuous negative-strand RNA synthesis, but this synthesis requires the de novo synthesis of ns polyprotein. (iii) The efficient replication of the ns polyprotein cleavage-deficient mutants is restricted to the cells that have profound defects in IFN-α/β production and signaling. The mutant viruses induce a level of IFN-α/β that is a few orders of magnitude higher than that of the wt, and the replication of the viruses, which produces only the uncleaved P123, also is sensitive to the autocrine effect of IFN-α/β. Moreover, even BHK-21 cells are capable of downregulating the synthesis of virus-specific RNAs and the reversion of both translational shutoff and translational shutoff-dependent transcription inhibition. (iv) Cleavage-deficient SINV mutants also are incapable of efficient replication in mosquito cells.

Thus, our current findings identify new roles of alphavirus ns polyprotein cleavage in virus replication in different cell types and its natural circulation, and they further define critical aspects of SINV-host cell interactions.