Sindbis Virus Infection Causes Cell Death by nsP2-Induced Transcriptional Shutoff or by nsP3-Dependent Translational Shutoff

Alphaviruses are a group of important human and animal pathogens with worldwide distribution. Their characteristic feature is a highly cytopathic phenotype in cells of vertebrate origin. The molecular mechanism of CPE remains poorly understood. In this study, by using Sindbis virus (SINV) as a model of the Old World alphaviruses, we demonstrated that SINV-specific CPE is redundantly determined by viral nsP2 and nsP3 proteins. NsP2 induces the global transcriptional shutoff, and this nuclear function can be abolished by the mutations of the small, surface-exposed peptide in the nsP2 protease domain. NsP3, in turn, determines the development of translational shutoff, and this activity depends on nsP3 macrodomain-associated mono-ADP-ribosylhydrolase activity. A combination of defined mutations in nsP2 and nsP3, which abolish SINV-induced transcription and translation inhibition, in the same viral genome does not affect SINV replication rates but makes it noncytopathic and a potent inducer of type I interferon.

ABSTRACT

Sindbis virus (SINV) is a representative member of the Alphavirus genus in the Togaviridae family. The hallmark of SINV replication in vertebrate cells is a rapid development of the cytopathic effect (CPE), which usually occurs within 24 h postinfection. Mechanistic understanding of CPE might lead to development of new prophylactic vaccines and therapeutic means against alphavirus infections. However, development of noncytopathic SINV variants and those of other Old World alphaviruses was always highly inefficient and usually resulted in selection of mutants demonstrating poor replication of the viral genome and transcription of subgenomic RNA. This likely caused a nonspecific negative effect on the rates of CPE development. The results of this study demonstrate that CPE induced by SINV and likely by other Old World alphaviruses is a multicomponent process, in which transcriptional and translational shutoffs are the key contributors. Inhibition of cellular transcription and translation is determined by SINV nsP2 and nsP3 proteins, respectively. Defined mutations in the nsP2-specific peptide between amino acids (aa) 674 and 688 prevent virus-induced degradation of the catalytic subunit of cellular-DNA-dependent RNA polymerase II and transcription inhibition and make SINV a strong type I interferon (IFN) inducer without affecting its replication rates. Mutations in the nsP3 macrodomain, which were demonstrated to inhibit its mono-ADP-ribosylhydrolase activity, downregulate the second component of CPE development, inhibition of cellular translation, and also have no effect on virus replication rates. Only the combination of nsP2- and nsP3-specific mutations in the SINV genome has a dramatic negative effect on the ability of virus to induce CPE.

IMPORTANCE Alphaviruses are a group of important human and animal pathogens with worldwide distribution. Their characteristic feature is a highly cytopathic phenotype in cells of vertebrate origin. The molecular mechanism of CPE remains poorly understood. In this study, by using Sindbis virus (SINV) as a model of the Old World alphaviruses, we demonstrated that SINV-specific CPE is redundantly determined by viral nsP2 and nsP3 proteins. NsP2 induces the global transcriptional shutoff, and this nuclear function can be abolished by the mutations of the small, surface-exposed peptide in the nsP2 protease domain. NsP3, in turn, determines the development of translational shutoff, and this activity depends on nsP3 macrodomain-associated mono-ADP-ribosylhydrolase activity. A combination of defined mutations in nsP2 and nsP3, which abolish SINV-induced transcription and translation inhibition, in the same viral genome does not affect SINV replication rates but makes it noncytopathic and a potent inducer of type I interferon.

INTRODUCTION

Alphaviruses are a group of small enveloped viruses with an RNA genome of positive polarity (1). In nature, they are transmitted between vertebrate hosts by mosquito vectors. In mosquitoes, alphaviruses cause persistent, life-long infection that does not have a detectable negative effect on insect biology. In vertebrates, alphaviruses cause diseases of different severities, characterized by rapid development of the high-titer viremia that is required for infecting new mosquitoes during the blood meal (1). Alphavirus replication in vitro mirrors the infection in vivo. These viruses develop persistent replication in cultured mosquito cells and highly cytopathic infection in cells of vertebrate origin. Within 3 to 4 h postinfection (PI), the infected cells already release viral particles, which perform the next round of infection. This rapid development of spreading infection is mediated by multiple mechanisms. First, the alphavirus replication machinery is highly efficient and the numbers of virus-specific RNAs can approach 10⁵ molecules per cell within 4 to 6 h PI; subsequently, within the next few hours, each infected cell releases 10³ to 10⁴ virions. Second, to promote infection spread, alphaviruses downregulate certain cell signaling pathways, primarily, the release of type I interferon (IFN), that can activate an antiviral state in as-yet-uninfected cells and thus prevent dissemination of infection (2, 3).

Based on their geographical circulation areas, alphaviruses are divided into two groups: the Old World (OW) and the New World (NW) alphaviruses. The NW alphaviruses include Venezuelan equine encephalitis virus, eastern equine encephalitis virus, and western equine encephalitis virus (VEEV, EEEV, and WEEV, respectively). They cause sporadic outbreaks in South America, Central America, and North America and develop meningoencephalitis with high mortality rates in humans and horses (4). The OW alphaviruses are more broadly distributed but usually cause a self-limited febrile illness. However, chikungunya virus (CHIKV), o’nyong’nyong virus (ONNV), and Ross River virus (RRV) are capable of producing excruciating joint pain and severe, persistent polyarthritis (5–12). In recent years, CHIKV has significantly broadened its circulation area, causing an increase in the numbers of human infections in both hemispheres and also in the United States (11, 13–18). OW alphaviruses such as Sindbis virus (SINV), Semliki Forest virus (SFV), and CHIKV exhibit a number of common characteristics. Therefore, SINV and SFV served for decades as good models for studying alphavirus-host interactions and molecular mechanisms of viral replication.

Alphavirus genomic RNA (G RNA) is approximately 11.5 kb in length. The G RNA mimics the structure of cellular mRNAs, in that it contains both a 5′ methylguanylate cap (cap0) and a 3′ poly(A) tail (19). The 5′ two-thirds of the genome is translated into 4 nonstructural proteins (nsPs) that comprise the viral components of the replication complex (vRC). The latter complex mediates replication of G RNA and transcription of the subgenomic RNA (SG RNA). The SG RNA is translated into the viral structural proteins. The structural protein-encoding genes can be deleted or replaced by heterologous genes, and upon delivery into the cells, such modified genomes (replicons) are capable of replication and expression of the heterologous proteins (20–23). Therefore, replicons are widely used in research for expression of heterologous genes and as a vaccine platform. They also represent an important tool for dissecting different aspects of virus-host interactions in the absence of high-level expression of viral structural proteins.

SINV, CHIKV, and SFV replicons, which lack viral structural genes, remain highly cytopathic (2, 20, 24). However, several mutations identified in the nsP2 coding sequence were capable of making the OW alphavirus replicons and some of the corresponding viruses very inefficient inducers of cytopathic effect (CPE). Mutation of P726 in the SINV nsP2 protein not only strongly reduces the cytopathogenicity of the virus and corresponding replicon but also downregulates viral replication (23, 25). We have previously shown that the OW alphavirus nsP2 proteins are responsible for inhibition of host transcription (26). After migration into the nucleus, wild-type (wt) nsP2 induces rapid and complete degradation of the catalytic subunit of cellular-DNA-dependent RNA polymerase II, RPB1 (27). The transcription inhibition induces cell death and prevents type I IFN release, despite efficient detection of the OW alphavirus replication by cellular pattern recognition receptors RIG-I and MDA5 (3). Within 2 to 4 h PI, cells also become unable to activate interferon-stimulated genes (ISGs) and to respond to IFN-β treatment (28). The P726 mutation in SINV nsP2 completely abrogated the ability of nsP2 to induce RPB1 degradation (27). Mutation of a corresponding proline in SFV also strongly reduced viral cytopathogenicity (29, 30). However, the effect of a similar mutation in CHIKV appears to be strain specific and additional mutations that result in strong reduction in viral replication are needed to make CHIKV replicons noncytopathic (24, 31).

An important characteristic of the previously selected noncytopathic SINV replicons was their low level of replication. Thus, the mutated SINV nsP2 not only lost its nuclear functions but also became an inefficient viral replication complex (vRC) component. Noncytopathic SFV and CHIKV replicons also demonstrated a dramatic decrease in RNA replication rates, suggesting that this effect represents a common mechanism of attenuation. A lower RNA replication level was likely the second important contributor to the development of a less cytopathic phenotype of attenuated replicons and viruses. Thus, in prior studies, we dissected the critical role of the OW alphavirus nsP2 transcription inhibition in CPE development (2, 26–28, 32), but other mechanisms contributing to CPE remained obscure because of the nonspecific effect of nsP2 mutations on replicon and viral replication.

In this study, we further investigated the mechanisms involved in CPE development during SINV replication in vertebrate cells. The newly developed SINV mutants remained capable of efficient replication but demonstrated a variety of new characteristics in virus-cell interactions. Our new data demonstrate that defined mutations in a small surface-exposed loop of the protease domain of SINV nsP2 have deleterious effects on the ability of nsP2 to induce RPB1 degradation and to inhibit host transcription. However, these nsP2-specific mutations did not make SINV noncytopathic and allowed us to further dissect another component of SINV-specific CPE development. The SINV nsP3-specific macrodomain was found to be involved in regulation of translation in SINV-infected cells. The identified mutations in this domain did not affect the rates of SINV RNA and viral replication and were not sufficient to prevent virus-induced CPE. However, combining newly identified nsP2- and nsP3-specific mutations in the same SINV genome abolished the ability of the virus to induce CPE and inhibit cell signaling.

RESULTS

Selection of SINV nsP2 mutants incapable of inducing cytopathic effect.

Previously, the main approach used for selection of noncytopathic alphavirus replicons was based on use of the dominant selectable markers, such as genes of puromycin (Pur) acetyltransferase (Pac) or aminoglycoside 3′-phosphotransferase (Neo), the products of which make cells resistant to puromycin and G418, respectively (25, 33). These genes were usually cloned into alphavirus genomes to replace those encoding the structural proteins that are generally dispensable for G RNA replication. Upon delivery into the cells, these modified alphavirus G RNAs self-replicate and express the encoded heterologous genes. After application of the drugs, some of the cells die because they do not contain a replicon and remain sensitive to selection (Pur^s or Neo^s), but most of them die despite being resistant to the drugs (Pur^r or Neo^r) due to the cytopathic nature of the OW alphavirus replicons. However, a very few cells survive the selection and develop foci of Pur^r or Neo^r cells. The corresponding focus-specific replicons contain adaptive mutations, which make them noncytopathic and still capable of replication and expression of the selectable marker genes. Identification of these adaptive mutations provided critical information for further studies of the mechanism underlying development of virus-specific CPE.

This approach has been successfully applied for SINV-based replicons, but only mutation at P726 of nsP2 has been unambiguously shown to be responsible for the noncytopathic phenotype (23, 28). Mutations at this position abrogated the nsP2-mediated degradation of RPB1, the catalytic subunit of cellular-DNA-dependent RNA polymerase II (27). However, the same mutations of P726 in the context of SINV or its replicons also had a very strong negative effect on RNA replication rates (25). Attempts to select noncytopathic SFV- or CHIKV-specific replicons also yielded only those with severely compromised replication rates, and none of the identified mutations have been evaluated for effects on RPB1 degradation (31, 33). This negative effect on RNA replication rates complicated the studies aiming to reveal the mechanism of CPE development and prevented dissection of other components besides inhibition of transcription by the nuclear fraction of nsP2.

In this study, we applied a new experimental system that was aimed at selection of spontaneously developing SINV nsP2 point mutants that were no longer capable of RPB1 degradation but that could efficiently function as vRC components in RNA replication. For selection of such mutants, we used a VEEV replicon encoding a SINV nsP2-green fluorescent protein (GFP) fusion under the control of the first subgenomic promoter and a Pac gene under the control of another one (Fig. 1A). VEEV replicons are not cytopathic per se, because their nsPs, including nsP2, have no nuclear functions (2, 27). However, we have previously shown that expression of SINV, CHIKV, and SFV naP2 with a natural first amino acid, achieved by using a Ubi-nsP2 fusion cassette, efficiently induced RPB1 degradation and rapid cell death (27). Fusion of the carboxy terminus of nsP2 with GFP does not affect nuclear inhibitory functions of the protein but allows monitoring of its intracellular distribution.

FIG 1.

Mutations affecting the cytotoxicity of SINV nsP2 accumulate in discrete fragments of the protein. (A) Schematic presentations of VEEV replicon encoding SINV nsP2-GFP fusion protein and the protocol for selection of noncytotoxic SINV nsP2. The N terminus of nsP2-GFP is fused with ubiquitin (Ubi) to mediate formation of the natural first amino acid. The Pac gene is cloned under the control of another subgenomic promoter. Cells were electroporated with the in vitro-synthesized replicon RNA and then treated with puromycin. Colonies of GFP-positive, Pur^r cells were selected for further analysis. (B) List of mutations identified in the SINV nsP2 gene of replicons in GFP-positive cells. Mutations that did not affect nuclear localization of SINV nsP2-GFP are depicted in red. Mutations that resulted in predominantly cytoplasmic localization of nsP2-GFP and predicted to localize to the internal protein core are depicted in black. Mutations that prevent nsP2 import into the nucleus and are exposed on the protein surface are depicted in blue. (C) Locations of SINV nsP2 mutations affecting the cytopathic phenotype of the protein that were identified in this and our previous study (32). (D) Positions of the selected mutations in the 3D model of the SINV nsP2 protease domain. The amino acids which were mutated in our new screen are depicted in the same colors as those used in panels B and C. Two protease-specific domains are shown in different colors.

BHK-21 cells were electroporated with the in vitro-synthesized replicon RNA, and at 24 h postelectroporation (PEP), puromycin selection was applied. Within a few days, we detected formation of ∼100 foci of Pur^r cells. A large fraction of them did not demonstrate GFP expression, suggesting alterations in the nsP2 open reading frame. A total of 21 cell clones with detectable GFP expression, generated in two independent electroporations, were selected for further analysis. In six of them, nsP2-GFP accumulated in the nuclei, with some fraction present in the cytoplasm. All others exhibited a predominantly cytoplasmic distribution of nsP2-GFP. SINV nsP2 coding fragments of all of the selected, noncytopathic replicons were sequenced, and identified mutations are presented in Fig. 1B and C. The distribution of the mutations was compared to that found in the previous experiments, which were based on transposon (Tn)-based mutagenesis, which randomly introduces 5-amino-acid (aa)-long sequences into SINV nsP2 (Fig. 1C) (32). Despite providing very important information at the time of that study, the effects of the insertions were difficult to interpret. Most of them affected both nsP2 nuclear function and virus growth, and attempts to select SINV variants with wt replication rates, but with no nuclear functions, were unsuccessful.

In this study, three of six replicons expressing primarily a nuclear form of nsP2-GFP had different substitutions of the same P726 in SINV nsP2 (P726L and P726Q), additionally supporting the idea of a critical role of this amino acid in nsP2 nuclear function as presented in prior studies (2, 26, 27). Since we have previously shown that any mutation of P726 strongly affects SINV replication rates (25), these new mutations were excluded from further experiments. The other three replicons acquired point mutations at different sites (P683Q and Q684P). Importantly, these two mutation sites overlapped the positions of the peptide insertions identified in our random insertion mutagenesis screen (32). Both mutated amino acids were located on the surface of nsP2 protease domain and were in close proximity to the previously investigated P726 amino acid (Fig. 1D).

The majority of other mutations in SINV nsP2 demonstrating cytoplasmic distributions were found in its C-terminal protease domain. None of them affected potential nuclear localization signals (32), despite making nsP2-GFP incapable of translocation to the nucleus. According to a SINV protease domain three-dimensional (3D) model, which was based on the published crystal structure (34), only three mutated amino acids (H619P, H619Q, and H643Q) were predicted to be exposed to solvent on the surface of nsP2 (Fig. 1D). All other identified amino acids were buried in the hydrophobic cores and thus likely impaired the overall conformation of the protease domain. Based on our previous experience, these mutations likely had deleterious effects on nsP2 function in RNA replication, and their effects were not investigated further.

Three other identified mutations (K68T, P271H, and L277R) were in the helicase domain of SINV nsP2. They also caused SINV nsP2-GFP to exclusively compartmentalize in the cytoplasm. Interestingly, the P271H and L277R mutations overlapped with several mutations previously introduced by random insertion mutagenesis (Fig. 1C). These mutations were expected to affect nsP2 helicase activity, which is essential not only for nsP2-mediated RPB1 degradation (27) but also for RNA replication (35). Thus, they would likely also lead to deleterious effects on SINV replication and, thus, were excluded from further study.

Characterization of viruses with selected mutations.

In this study, we focused on three selected mutations, namely, H619Q, H643Q, and P683Q, that were located in the protease domain and were predicted to be at least partially exposed to solvent (Fig. 1D). The identified mutations were introduced into the genome of wt SINV/GFP, and the corresponding variants (SINV/nsP2-619Q/GFP, SINV/nsP2-643Q/GFP, and SINV/nsP2-683Q/GFP) and control wt SINV/GFP were rescued by electroporation of the in vitro-synthesized RNA into BHK-21 cells. Media were collected at different times postelectroporation to evaluate infectious titers. SINV/nsP2-683Q/GFP demonstrated replication rates that were indistinguishable from those of wt SINV/GFP (Fig. 2A). In the infectious center assay (ICA), the in vitro-synthesized RNAs of SINV/GFP and SINV/nsP2-683Q/GFP exhibited the same levels of infectivity, suggesting that the latter mutant did not require additional mutations for its viability. Two other mutants, SINV/nsP2-619Q/GFP and SINV/nsP2-643Q/GFP, were not viable. Very few GFP-positive cells, which ultimately developed plaques, were detected in the ICA, and sequencing of viral nsP2 genes from the randomly selected plaques identified rescued viruses as true revertants.

FIG 2.

Selected mutations that affect translocation of SINV nsP2-GFP into the nucleus have a deleterious effect on the ability of nsP2 to induce RPB1 degradation but make SINV nonviable. (A) Viral replication rates, RNA infectivity, and infectious titers of the designed SINV nsP2 mutants upon transfection of the in vitro-synthesized RNA into BHK-21 cells. (B) BHK-21 cells were infected with packaged VEEV replicons encoding the indicated variants of Ubi-nsP2-GFP and analyzed by confocal microscopy at 6 h PI. Images are presented as multiple-image projections of a 1-μm x-y section (6 optical sections) through the nuclei. Scale bars: 10 μm. (C) Western blot analysis of cells infected with VEEV replicons expressing wt or mutant nsP2-GFP fusion proteins with and without nuclear localization signal.

The H619Q and H643Q mutations were additionally characterized in terms of their effect on the ability of SINV nsP2 to cause RPB1 degradation. Based on the original screen, SINV nsP2-GFP containing either of these mutations and expressed by VEEV replicons was distributed mostly in the cytoplasm (Fig. 1B). Therefore, to additionally understand effects of the mutations on nsP2 nuclear functions, the mutated nsP2-GFP cassettes expressed by VEEV replicons were designed to either contain or have no additional nuclear localization signal (NLS) at the C terminus of GFP. These replicons were packaged into VEEV structural proteins and then used to infect naive cells. Addition of NLS caused efficient accumulation of nsP2-GFP in nuclei (Fig. 2B). However, in contrast to wt nsP2-GFP or its fusion with NLS, none of the mutated nsP2 fusions caused RPB1 degradation (Fig. 2C). Thus, H619Q and H643Q mutations abrogated SINV nsP2 nuclear functions even if the mutant protein was transported into the nucleus.

Mutations of P683 prevent nsP2-mediated RPB1 degradation but not CPE development.

The P683Q mutation described above made the SINV nsP2-GFP fusion expressed from VEEV replicon noncytopathic and did not demonstrate a detectable negative effect on replication of SINV, at least in BHK-21 cells (Fig. 2A). This was the first indication that the latter mutation could affect the transcription inhibition component of SINV-specific CPE while having no effect on other virus-specific mechanisms of CPE induction. To experimentally confirm this hypothesis, we compared the rates of RPB1 degradation in cells infected with wt SINV/GFP and SINV/nsP2-683Q/GFP. In correlation with the previously published data (27), infection of BHK-21 cells with SINV encoding wt nsP2 induced rapid degradation of RPB1, and by 4 h PI only 6% of RPB1 remained (Fig. 3B). In contrast, infection with SINV/nsP2-683Q/GFP did not induce degradation of RPB1, despite the fact that wt and mutant viruses were producing essentially the same levels of nsP2 at any time PI.

FIG 3.

P683Q mutation in nsP2 does not make SINV noncytopathic, despite the inability of mutant nsP2 to induce degradation of RPB1. (A) Schematic presentation of recombinant viral genomes. (B) BHK-21 cells were infected with the indicated viruses at an MOI of 20 PFU/cell. Cell were harvested at the indicated times PI, and the levels of RPB1 and nsP2 were analyzed by Western blotting. (C) BHK-21 and NIH 3T3 cells were infected with the indicated viruses at MOIs of 10 and 20 PFU/cell, respectively. At the indicated times PI, media were replaced and viral titers were determined by plaque assay on BHK-21 cells.

The previously developed and widely used SINV/G/GFP mutant containing the P726G mutation in nsP2 demonstrated reduced cytopathogenicity that correlated not only with a loss of nsP2 nuclear function but also with lower rates of RNA and viral replication. Thus, to characterize the effect of P683Q mutation on SINV biology, we infected BHK-21 and NIH 3T3 cells with wt SINV/GFP, SINV/G/GFP, and SINV/nsP2-683Q/GFP and compared their rates of clearance and their abilities to establish persistent infection in these cell types (Fig. 3C). As expected, SINV/GFP rapidly developed complete CPE in both cell types. SINV/G/GFP established persistent infection in BHK-21 cells, which are deficient in the development of type I IFN responses, and NIH 3T3 cells cleared virus replication within 7 days due to an autocrine effect of the induced type I IFN (3). In contrast to the SINV/G/GFP results, infection with SINV/nsP2-683Q/GFP remained highly cytopathic despite the ability of the mutant to induce very high levels of type I IFN (see the following sections). As in the case of SINV/GFP, which encoded wt nsP2, no viable cells could be found at 48 h PI.

One explanation for the high cytopathogenicity of SINV/nsP2-683Q/GFP could be in its efficient reversion to the wt phenotype, because only a single nucleotide was substituted in the nsP2-coding sequence. Alternatively, as in the case of the P726 mutation, the effect of the P683 mutation could potentially depend on the substituting amino acid (25). Thus, we next tested the effects of different P683 replacements in SINV nsP2 on viral nuclear functions, replication rates, and efficiency of type I IFN induction. To reduce the possibility of reversion to the wt phenotype, P683E, P683S, and P683N mutations were chosen because they required two or more nucleotides to be replaced (Fig. 4A). All of the designed variants were viable. They replicated as efficiently as wt SINV/GFP to titers that were more than 50-fold higher than those of SINV/G/GFP mutant (Fig. 4C). However, similarly to SINV/G/GFP, the mutant viruses were incapable of inducing RPB1 degradation and were all strong inducers of IFN-β in NIH 3T3 cells (Fig. 4B and D). Nevertheless, they all remained highly cytopathic.

FIG 4.

SINV nsP2 P683 mutants do not cause RPB1 degradation and efficiently induce IFN-β response. (A) Schematic presentations of wt and mutant viral genomes. (B) NIH 3T3 cells were infected with the indicated SINV variants at an MOI of 20 PFU/cell. Cells lysates were prepared at 8 h PI. The integrity of RPB1 and accumulation of SINV nsP2 were analyzed by Western blotting using the corresponding Abs. (C and D) NIH 3T3 cells were infected with the indicated viruses at an MOI of 20 PFU/cell. Viral titers (C) and concentrations of IFN-β in the media (D) were assessed at 16 h PI as described in Materials and Methods. Data are shown as means ± standard deviations (SD) of results from 3 biological repeats.

Thus, the newly designed mutations of P683 of SINV nsP2 abolished nsP2-mediated degradation of RPB1 but had no noticeable effect on replication of the virus and its ability to cause CPE. These data supported the hypothesis that the mechanism of CPE development by SINV and likely other OW alphaviruses is not determined only by the nuclear function of nsP2 in induction of transcriptional shutoff. Other processes in virus-host interactions, in addition to RPB1 degradation, are determinants of CPE induction even in the absence of transcription inhibition.

Mutations in nsP3 reduce the cytopathogenicity of SINV/GFP with a mutated nsP2.

In the above-described experiments, we identified amino acid substitutions of P683 that had a deleterious effect on nsP2 nuclear function but preserved the highly cytopathic phenotype of the virus and its efficient replication. Thus, in the first part of this study, we succeeded in inactivating one of the mechanisms underlying the development of SINV-specific CPE without affecting others. If the hypothesis about the redundant involvement of two or more mechanisms in CPE development is correct, then selection of noncytopathic SINV-based replicon RNAs could be made more efficient by the use of P683 mutants. Since the replicons containing the indicated mutation did not require inactivation of the nsP2 transcription inhibitory function to become noncytopathic, they were expected to more efficiently acquire the mutations that inactivate other nsP-dependent mechanisms in CPE development. Identification of such mutations could potentially uncover new aspects of SINV-host interactions which are exploited by virus infection in CPE development.

For these new selection steps, we designed a SINrep/nsP2-683S/GFP/Pac replicon. It contained the P683S mutation in nsP2 and encoded GFP and Pac under the control of separate subgenomic promoters (Fig. 5A). BHK-21 cells were electroporated with the in vitro-synthesized RNAs of this and wt replicons and were then subjected to puromycin selection. The wt SINrep/GFP/Pac variant produced very few colonies of Pur^r cells per microgram of electroporated RNA. The mutant replicon produced colonies at two orders of magnitude higher efficiency (Fig. 5A). This suggested that a wider range of spontaneous mutations could lead to the development of a noncytopathic phenotype, when the nuclear functions of nsP2 had already been inactivated. We randomly selected two colonies of Pur^r cells that demonstrated high levels of GFP expression and sequenced the nonstructural genes of the persisting replicons. In addition to the preexisting P683S mutation in nsP2, one of the replicons contained a T379P substitution in nsP1, and a 6-aa-long sequence in the N terminus of nsP3 was deleted (Δ24-29) in the second replicon. To confirm the negative effects of these additional changes on cytopathogenicity, the mutations were introduced into the original SINrep/nsP2-683S/GFP/Pac, and the in vitro-synthesized RNAs were electroporated into BHK-21 cells (Fig. 5B). The new double mutants produced more than 10⁴ colonies per microgram of transfected RNA. This efficiency of colony formation was similar to that of the previously described SINrep/Pac replicon, with nsP2 containing a P726L amino acid substitution that had deleterious effects on both nuclear and vRC-specific functions of the latter protein (25). Western blotting confirmed that the parental and new replicons expressed similar levels of nsP2 and nsP3 (Fig. 5C), and no abnormalities in polyprotein processing were detected (data not shown).

FIG 5.

Mutations in nsP2 and other nsPs synergistically affect the cytopathogenicity of SINV replicons. (A) Schematic presentation of SINV replicons and their efficiencies in the formation of Pur^r cell colonies. BHK-21 cells were electroporated with the in vitro-synthesized replicon RNAs. After puromycin selection, cell colonies were stained with crystal violet. Images represent plates seeded with equal numbers of electroporated cells. (B) Schematic presentation of SINV replicons containing mutation in nsP2 and nsP1 or nsP3. BHK-21 cells were electroporated with the in vitro-synthesized RNAs of the indicated replicons. After puromycin selection, cell colonies were stained with crystal violet. Images represent the plates seeded with equal numbers of electroporated cells. (C) BHK-21 cells were electroporated with equal amounts of the in vitro-synthesized RNAs of the indicated replicons. Cell lysates were prepared at 24 h postelectroporation and analyzed using nsP2-, nsP3-, GFP-, and tubulin-specific Abs.

Next, we introduced the identified nsP1- and nsP3-specific mutations into cDNA encoding the infectious viral genome, namely, SINV/nsP2-683S/GFP. For as-yet-unclear reasons, the nsP1-specific mutation had a strong negative effect on the infectivity of the in vitro-synthesized RNA and the rates of infectious virus release (Fig. 6A). The infectivity of SINV/nsP2-683S,nsP3Δ/GFP RNA was also noticeably lower (about 6-fold) than that seen with SINV/GFP, but the detected decrease was not as strong as when additional adaptive mutations are required for viability. The infectious titers of the harvested stocks of SINV/nsP2-683S,nsP3Δ/GFP were essentially the same as those measured for SINV/nsP2-683S/GFP and SINV/GFP.

FIG 6.

Defined mutations in nsP2 and nsP3 make SINV capable of noncytopathic replication in vertebrate cells without a profound effect on viral replication rates. (A) Schematic presentation of the designed recombinant viral genomes, infectivity of the in vitro-synthesized RNAs, and viral titers at 24 h postelectroporation. (B) NIH 3T3 cells were infected with the indicated variants at an MOI of 50 PFU/cell. At 18 h PI, media were harvested, and viral titers and concentrations of IFN-β were assessed as described in Materials and Methods. Data are shown as means ± SD of results from 3 biological repeats. (C) NIH 3T3 cells were infected with the indicated SINV variants at an MOI of 20 PFU/cell. Cell lysates were prepared at 8 h PI and analyzed using RPB1-, nsP2-, STAT1-, pSTAT1-, and tubulin-specific Abs. (D) NIH 3T3 cells and their Mavs KO derivatives were infected with the indicated variants at an MOI of 20 PFU/cell. Media were replaced every 24 h, and viral titers were determined by plaque assay on BHK-21 cells.

The experiments in NIH 3T3 cells described below demonstrated that the nsP2+nsP3 double mutant was capable of efficient replication in this cell line and that its titers were essentially the same at both early (8 h) and late (18 h) times PI (Fig. 6B and data not shown) as those of the parental wt SINV/GFP strain and single SINV/nsP2-683S/GFP mutant. The double mutant efficiently produced nsP2 and did not induce degradation of RPB1 (Fig. 6C). Accordingly, infection with SINV/nsP2-683S,nsP3Δ/GFP stimulated IFN-β release and phosphorylation of STAT1 (Fig. 6B and C). Most importantly, the double mutant was dramatically less cytopathic than the SINV/GFP and SINV/nsP2-683S/GFP variants (Fig. 6D). NIH 3T3 cells that have no defects in type I IFN production and signaling were able to stop replication of the double mutant and clear the infection. However, this virus was able to persistently replicate in Mavs knockout (KO) NIH 3T3 cells, which do not induce IFN-β release in response to virus replication. Thus, the short N-terminal deletion in nsP3 affected another mechanism(s) of SINV-specific CPE development without affecting viral replication rates.

Mutations in SINV nsP2 and nsP3 affect the development of transcriptional and translational shutoffs, respectively.

In prior studies, we demonstrated that SINV infection in vertebrate cells rapidly inhibits cellular transcription and translation through independent mechanisms (26). The transcriptional shutoff is caused by RPB1 degradation, and the translational shutoff is mediated by both PKR-dependent and poorly characterized PKR-independent mechanisms (36). To evaluate the effects of the newly developed mutants on these critical, virus-specific modifications of the intracellular environment, we performed metabolic pulse-labeling of the synthesized RNAs and proteins in virus-infected cells. As previously reported (26), wt SINV/GFP induced rapid shutoff of cellular transcription and translation (Fig. 7). Within a few hours PI, cells began to synthesize only viral RNAs and viral structural proteins. The previously developed SINV/G/GFP mutant produced lower levels of SG RNA. The synthesis of its genomic RNA was likely also inefficient, as was previously shown (26), but in the absence of ActD in the labeling medium that was used, this effect could not be detected, because the radioactively labeled G RNA, and 45S and 47S pre-rRNAs comigrated as a single band on the agarose gels. SINV/G/GFP-infected cells continued to produce a large amount of pre-mRNA and rRNA (Fig. 7A). As expected, single nsP2 mutants with a mutation at P683 and the nsP2+nsP3 double mutant, in particular, were also inefficient in transcription inhibition despite high levels of virus-specific RNA synthesis. However, only a SINV/nsP2-683S,nsP3Δ/GFP double mutant was less efficient in translation inhibition. The presence of a ³⁵S-labeled actin band was readily detectable on the gel (Fig. 7B). Of note, this band had lower intensity than was seen in mock-infected cells. This was suggestive of the possibility that, in the cells infected with the double mutant, the PKR-dependent component of translation inhibition remained intact (36).

FIG 7.

The identified nsP2- and nsP3-specific mutations make SINV incapable of inducing transcriptional and translational shutoffs, respectively. (A) NIH 3T3 cells were infected with the indicated viruses at an MOI of 20 PFU/cell. RNAs were metabolically labeled with [³H]uridine (20 μCi/ml) between 3 and 7 h PI in the absence of ActD. RNAs were isolated and analyzed by agarose gel electrophoresis as described in Materials and Methods. The positions of viral and ribosomal RNAs are indicated. Diffuse radioactive signals in the samples of mock-infected cells and those infected with the mutants correspond to cellular pre-mRNAs and mRNAs. (B) NIH 3T3 cells were infected with the indicated viruses at an MOI of 20 PFU/cell. At 6 h PI, proteins were metabolically labeled with [³⁵S]methionine for 30 min and analyzed by SDS-PAGE as described in Materials and Methods. The gels were dried and autoradiographed.

Recently, the N-terminal macrodomain of alphavirus nsP3 was shown to function as mono-ADP-ribosylhydrolase, and a N24A mutation in the CHIKV nsP3 macrodomain was shown to abolish that hydrolase activity (37). The crystal structures of VEEV and CHIKV macrodomains with ADP-ribose also suggested that N24 is involved in the phosphatase activity of this domain (38). N24 was among the amino acids deleted in the selected nsP2-683S,nsP3Δ mutant SINV replicon. This suggested that inhibition of nsP3 mono-ADP-ribosylhydrolase activity may be responsible for further attenuation of SINV variants encoding already mutated nsP2. To experimentally test this possibility, we introduced the N24A mutation into nsP3-coding sequences of SINV/GFP and SINV/nsP2-683S/GFP viral genomes. Both new mutants were viable and replicated to the same titers as their counterparts with wt nsP3 (Fig. 8B). The N24A mutation alone did not make the SINV/nsP3-24A/GFP variant with the wt nsP2 noncytopathic and a type I IFN inducer, and the latter virus still efficiently induced RPB1 degradation (Fig. 8C and D). However, the SINV/nsP2-683S,nsP3-24A/GFP double mutant was as efficient a type I IFN inducer as the parental SINV/nsP2(P683S)/GFP variant. It was also incapable of RPB1 degradation and lost the highly cytopathic phenotype. The SINV/nsP2-683S,nsP3-24A/GFP mutant was efficiently cleared from NIH 3T3 cells and readily established persistent infection in Mavs KO cells (Fig. 8E), as we detected with the prototype nsP2+nsP3 double mutant having a deletion of 6 aa in the N terminus of nsP3 (Fig. 6D). Thus, the effect of a N24A point mutation reproduced that of the experimentally selected nsP3-specific deletion and additionally pointed to the role of SINV nsP3-specific mono-ADP-ribosylhydrolase activity in the development of translational shutoff and CPE in SINV-infected cells. Detailed characterization of the mechanism of this function is now under way.

FIG 8.

Mutations in the mono-ADP-ribosylhydrolase catalytic center of the SINV nsP3 macrodomain make the SINV nsP2+nsP3 double mutant noncytopathic. (A) Schematic presentation of SINV genomes with mutations in nsP2 and nsP3. (B) NIH 3T3 cells were infected with the indicated viruses at an MOI of 20 PFU/cell. Viral titers were assessed at 8 h PI by plaque assay on BHK-21 cells. (C) NIH 3T3 cells were infected with the indicated viruses at an MOI of 20 PFU/cell. Concentrations of the released IFN-β were measured at 18 h PI. (D) NIH 3T3 cells were infected with the indicated viruses at an MOI of 20 PFU/cell. Cell lysates were prepared at 8 h PI and analyzed by Western blotting using RPB1-, nsP2-, and tubulin-specific Abs. Quantitative analysis of RPB1 levels was performed on a LI-COR imager. (E) NIH 3T3 cells and their Mavs KO derivatives were infected with the indicated variants at an MOI of 20 PFU/cell. Media were replaced every 24 h, and viral titers were determined by plaque assay on BHK-21 cells. Data in panels B and C are shown as means ± SD of results from 3 biological repeats.

DISCUSSION

One of the fundamental characteristics of SINV replication in vertebrate cells is rapid development of CPE (1). Infected cells usually begin to exhibit morphological changes within 6 to 10 h PI and lose their integrity and die by 24 h post-SINV infection. During this time, the major changes in cell biology may cause the formation of autophagosomes, development of apoptosis, endoplasmic reticulum stress, etc. (39). CPE is determined by a combination of virus-induced changes in cell biology, and the involvement of multiple mechanisms greatly complicates dissection of individual components. In this study, we intended to further reveal the molecular basis of the processes that underline development of CPE in SINV-infected cells. The replication process of this virus demonstrates a number of commonalities with those of other OW alphaviruses; thus, SINV represents a good model for studying interactions of the OW alphaviruses with host cells. The most important common characteristics of CPE development that we considered in our experiments were as follows. (i) All of the studied OW alphaviruses and their replicons rapidly induce CPE in vertebrate cells (1). (ii) Replication of all of the OW alphaviruses globally and rapidly inhibits cellular transcription. The transcriptional shutoff is determined by the nuclear fraction of their nsP2 proteins, which induce degradation of the catalytic subunit of cellular-DNA-dependent RNA polymerase II (RPB1) and thus abrogate transcription of cellular mRNAs (27, 40). (iii) Expression of the wt OW alphavirus nsP2 alone in vertebrate cells also induces inhibition of cellular transcription that is sufficient for inducing cell death and CPE development (32). However, SINV and SFV nsP2/3 cleavage mutants produce only the unprocessed P23 that remains exclusively in the cytoplasm (26, 41). These viruses do not induce transcriptional shutoff but remain highly cytopathic. Their ability to induce CPE suggested the existence of an additional virus-induced mechanism(s) of CPE induction. (iv) Importantly, selection of the noncytopathic OW alphavirus replicons was always highly inefficient. Both SP6 RNA polymerase (which is used for the in vitro synthesis of replicon genomes) and alphavirus RNA-dependent RNA polymerase have relatively low fidelity. Nevertheless, very few colonies of cells containing noncytopathic replicons have always been selected. We estimated that ∼1 of 10⁶ cells that received the in vitro-synthesized SINV or SFV replicon was capable of developing colonies of replicon-containing, drug-resistant cells (25, 33). Thus, the efficiency of acquiring the noncytopathic phenotype by SINV replicons was 4 orders of magnitude lower than that normally detected during selection of single-point mutations, which promote virus replication. This was a strong indirect indication that more than one virus-specific mechanism is involved in CPE induction and that very few single-point mutations can inactivate more than one process in virus-host interactions and lead to development of a noncytopathic phenotype. (v) To date, all selected noncytopathic SFV, SINV, and CHIKV replicons demonstrated highly inefficient RNA replication, suggesting that, besides inactivating nuclear functions of nsP2, the acquired mutations most likely also reduced RNA replication rates and thus nonspecifically affected the efficiency of CPE induction.

Considering the data described above, the rationale of this study was to further dissect the fundamental changes in cell biology that ultimately result in CPE development and to define the roles of SINV nsPs in these processes. Analysis of SINV nsP2 mutants that are incapable of inducing only the transcriptional shutoff could be a good starting point for identification of other components of CPE. However, to date, only the effects of P726 substitutions in nsP2 have been characterized (25–27). The substitutions abolished the ability of the latter protein to induce RPB1 degradation and made the virus incapable of CPE induction. However, they also had strong, nonspecific negative effects on RNA and virus replication. Therefore, the P726 mutants could not be used for identification of components of CPE development beyond dissection of the functions of nsP2 in transcription inhibition.

To overcome this problem, we first identified another set of attenuating mutations in the nsP2 coding sequence, which made this protein very inefficient in transcription inhibition. The mutated P683 and Q684 amino acids were located on the surface of the nsP2 molecule close to the previously described P726G substitution (Fig. 1D). However, in contrast to the P726G mutation, they had no effect on nsP2 function as the vRC component. As in wt SINV infection, the mutated nsP2 was transported into the nucleus but did not cause degradation of RPB1. Consequently, the designed viruses became very potent type I IFN inducers. However, most importantly, they remained cytopathic in all of the tested cell lines of vertebrate origin. Interestingly, the codons of P683 and Q684 were previously identified as the sites of short in-frame sequence insertions into SINV nsP2 by random insertion mutagenesis (32). Such insertions made nsP2, which was expressed alone, also incapable of CPE induction. In that study, the entire set of the insertion sites that affected nsP2 nuclear functions was represented by nsP2 codons 676, 678, 682, 683, 684, and 687. At that time, the effects of the peptide insertions into nsP2 were not further investigated, except to demonstrate that the mutated proteins accumulated in the nuclei. However, the new data suggest the possibility that substitutions of aa 676, 678, 682, and 687, and probably other amino acids which are in close proximity to P683 on the protein surface, could also affect interaction of nsP2 with nuclear factors and the ability of this protein to induce CPE.

The selection of efficiently replicating SINV nsP2 mutants that no longer exhibited nuclear functions allowed us to dissect another process involved in CPE development which was not directly connected to transcription inhibition. At the second step of selection, SINV replicons containing a P683S substitution in nsP2 were 2 orders of magnitude more efficient in formation of Pur^r colonies than their wt counterparts. This was an indication that further development of the noncytopathic phenotype could be achieved by numerous additional point mutations in SINV nsP genes. The following experiments were focused on analyzing one of the identified mutations, which led to deletion of six amino acids in the N terminus of SINV nsP3. That deletion had no effect on either SINV replication rates or synthesis of virus-specific RNA but strongly affected development of translational shutoff, which is characteristic of SINV replication in vertebrate cells. The presence of both P683S and Δ24-26 mutations in nsP2 and nsP3 of SINV replicons caused a 100-fold increase in the efficiency of Pur^r colony formation, indicating that the Δ24-26 mutation affected a critical mechanism of CPE development in cells containing self-replicating SINV-specific RNAs.

Recently, the N-terminal sequence in the nsP3 macrodomain has been suggested to be a critical part of the nsP3-associated mono-ADP-ribosylhydrolase catalytic site (37, 38, 42). Thus, we designed an additional viral mutant by substituting a single amino acid, N24A, in the encoded nsP3. On the basis of the published data, this mutation inhibited the mono-ADP-ribosylhydrolase activity of the macrodomain (37). The designed nsP2+nsP3 double mutants of SINV that had either the identified deletion of aa 24 to 29 or the single amino acid N24A substitution replicated as efficiently as the wt virus but were dramatically less cytopathic. They either were cleared by NIH 3T3 cells or could persistently replicate in their Mavs KO derivatives. Notably, in the absence of the P683S substitution in nsP2, N24A alone had no noticeable effect on either SINV replication rates or the efficiency of transcription inhibition and cytopathogenicity of the virus. This was an additional demonstration that the nsP2-mediated transcriptional shutoff is a critical mechanism of CPE development. The lack of an effect of nsP3-specific mutation in the context of wt virus also correlated with the results of our prior study, in which we selected SINV with the insertion of the entire GFP into codon 28 of nsP3 (49). The replication competency of that SINV variant suggested that even strong modifications of this fragment are not lethal for virus replication in vitro. However, data from this new study demonstrate that the N-terminal fragment of nsP3 has an important function in SINV-host cell interaction and in the development of translational shutoff in particular. Its role becomes clearly detectable in the absence of another redundant determinant of CPE, namely, nsP2-induced transcription inhibition.

Interestingly, instead of having an antiviral effect, inhibition of translation in SINV- and SFV-infected cells is highly beneficial for virus replication. SINV-specific translational shutoff is determined by two mechanisms; the first is PKR independent, and the second efficiently mediates translational shutoff even in PKR^−/− cells (36). Thus far, the mechanism of PKR-independent inhibition of translation has remained unknown, but our new data suggest that nsP3-associated mono-ADP-ribosylhydrolase activity may be a key player in this process. Identification of cellular targets of this nsP3-associated enzymatic activity will be necessary for further understanding of this protein’s function.

The additional second site mutation that has been found to strongly reduce cytopathogenicity of the SINV nsP2 mutant has been identified in nsP1 protein. Unlike the mutation in nsP3, this mutation strongly reduced viral replication. As in the case of less-cytopathic SINV, CHIKV, and SFV replicons (25, 29, 31, 33), the nsP3 function in inhibition of cellular translation was likely affected by reducing its concentration in the infected cells. This further suggests that inhibition of cellular translation by the nsP3 macrodomain-associated mono-ADP-ribosylhydrolase activity is not very efficient and requires accumulation of nsP3 in the cytoplasm at a high concentration.

In summary, the results of this study demonstrate that development of CPE during replication of SINV and probably of other OW alphaviruses is determined by multiple mechanisms. One of them is inhibition of transcription, which is mediated by the nuclear function(s) of nsP2. The defined mutations in the peptide located on the surface of nsP2 between aa 674 and 688 can be dispensable for viral replication. However, they prevent virus-induced RPB1 degradation and transcriptional shutoff and make SINV a strong type I IFN inducer. Nevertheless, these mutations are not sufficient for preventing CPE. Further downregulation of SINV cytopathogenicity results from additional mutations in the nsP coding sequence. The identified mutations in the nsP3 macrodomain, which potentially inhibit its mono-ADP-ribosylhydrolase, made SINV dramatically less cytopathic but also had no effect on its replication rates. The requirements for acquisition of two independent mutations affecting different aspects of SINV-host interactions provide a plausible explanation for the difficulty of selecting the less cytopathic variants with high levels of RNA and virus replication. These data also open new possibilities for attenuation of the OW alphaviruses and development of efficient and less-cytopathic alphavirus expression systems.

MATERIALS AND METHODS

Cell cultures.

NIH 3T3 cells were obtained from the American Type Culture Collection (Manassas, VA). BHK-21 cells were kindly provided by Paul Olivo (Washington University, St. Louis, MO). These cell lines were maintained at 37°C in alpha minimum essential medium (αMEM) supplemented with 10% fetal bovine serum (FBS) and vitamins. The Mavs KO cell line was generated from NIH 3T3 cells by introducing a mutation in the second exon of Mavs gene using clustered regularly interspaced short palindromic repeat (CRISPR) technology as we previously described (43). The presence of modifications was confirmed using Sanger sequencing of PCR fragments of the G RNA targets in the genome. The absence of MAVS expression was additionally confirmed by Western blotting using MAVS-specific antibodies (sc-365334; Santa Cruz Biotechnology).

Plasmid constructs.

The plasmids encoding SINV Toto1101 genomes of pSINV/GFP and mutant pSINV/G/GFP and VEEV replicons encoding SINV nsP2 gene, pVEEVrep/nsP2-GFP/Pac, and pVEEVrep/nsP2-GFP-NLS/Pac were described elsewhere (28, 44). All plasmids containing cDNAs of mutant replicons and viruses were constructed using standard PCR-based techniques. All mutations were confirmed by sequencing. Schematic presentations of all of the modified genomes are shown in the corresponding figures. Sequences of the plasmids and details of the cloning procedures can be provided upon request.

In vitro RNA transcription and transfection.

Plasmids were purified by ultracentrifugation in CsCl gradients. They were then linearized using unique restriction sites located downstream of the poly(A) sequence. RNAs were synthesized by SP6 RNA polymerase in the presence of a cap analog (New England Biolabs) according to the recommendations of the manufacturer (Invitrogen). Aliquots of transcription reaction mixtures were used for electroporation without additional purification. Electroporation of BHK-21 cells by in vitro-synthesized viral genomes was performed under previously described conditions (45, 46). Viruses were harvested at 20 h to 24 h postelectroporation. Viral titers were determined by plaque assay on BHK-21 cells (47).

Selection of SINV nsP2 genes encoding noncytotoxic proteins.

BHK-21 cells were electroporated with 5 μg of in vitro-synthesized VEEVrep/nsP2-GFP/Pac and plated into 100-mm-diameter tissue culture dishes. At 6 h postelectroporation (PEP), the medium was supplemented with puromycin (5 μg/ml). Electroporated cells were grown under conditions of puromycin selection for 12 days. The developed cell clones were collected, and RNA was isolated by the use of TRIzol according to the instructions of the manufacturer (Invitrogen). SINV nsP2-coding genes were amplified by reverse transcription-PCR (RT-PCR) and sequenced.

Analysis of the cytotoxicity of SINV replicons.

The in vitro-synthesized replicon RNAs (5 μg) were electroporated into BHK-21cells, and different numbers of transfected cells were seeded into 100-mm-diameter dishes. At 6 h PEP, the medium was replaced by medium supplemented with puromycin (10 μg/ml). Grown colonies of Pur^r cells were fixed with paraformaldehyde and stained with crystal violet for counting. The results are presented as CFU per microgram of transfected RNA.

Infectious center assay.

To compare the infectivity levels of the viral RNAs, BHK-21 cells were electroporated with 1 μg of the in vitro-synthesized viral genomic RNAs. Ten-fold dilutions of electroporated cells were seeded in 6-well Costar plates containing subconfluent monolayers of naive BHK-21 cells. After 2 h of incubation at 37°C, cells were overlaid with 0.5% agarose supplemented with MEM and 3% FBS. Plaques were stained after 2 days of incubation at 37°C, and RNA infectivity was determined as PFU per microgram of transfected RNA.

Analysis of viral replication.

Cells were seeded into 35-mm-diameter dishes and infected at the multiplicities of infection (MOIs) indicated in the figure legends. At the indicated times, media were harvested, and viral titers in the samples were determined using the plaque assay on BHK-21 cells.

Analysis of the viral persistence.

The cell lines were infected at the MOIs indicated in the figure legends, washed with phosphate-buffered saline (PBS), and then incubated in cell-specific media. These media were replaced every 24 h for 10 days, and cells were reseeded upon reaching confluence. Viral titers were determined by plaque assay on BHK-21 cells as described previously (47).

IFN-β assay.

Media were collected at the indicated times PI, and pH was stabilized by the use of HEPES. Concentrations of IFN-β in the media were estimated by the use of a VeriKine mouse interferon beta enzyme-linked immunosorbent assay (ELISA) kit according to the instructions of the manufacturer (PBL Assay Science).

Western blotting.

NuPAGE gels (Invitrogen) (4% to 12%) were used for separation of equal amounts of protein. Samples were transferred to nitrocellulose membranes (GE Healthcare), and proteins were stained with specific primary and infrared dye-labeled secondary antibodies (Abs). Membranes were scanned on the Odyssey imager (Li-COR Biosciences). Quantitative analysis was performed using the imager’s software. The data were normalized to the intensity of the tubulin band. The following primary antibodies were used: anti-tubulin (rat monoclonal antibody [MAb]; University of Alabama at Birmingham [UAB] core facility), rabbit polyclonal antibodies against SINV nsP3 (custom-made), SINV nsP2-specific mouse MAb (custom-made), anti-STAT1 MAb (rabbit MAb; Epitomics), anti-pSTAT1 MAb (mouse MAb) (pY701; BD Transduction Laboratories), and anti-RPB1 (8wG16 [Covance], 4H8 [Active Motif], or F12 [Santa Cruz Biotechnology]).

Analysis of RNA synthesis.

NIH 3T3 cells in 6-well Costar plates (5 × 10⁵ cells/well) were infected with SINV mutants at an MOI of 20 PFU/cell. RNAs were metabolically labeled with [³H]uridine (20 μCi/ml) between 3 and 7 h PI in complete growth media in the absence of ActD. RNAs were isolated and analyzed by agarose gel electrophoresis under denaturing conditions as described elsewhere (48).

Analysis of protein synthesis.

NIH 3T3 cells were seeded into 6-well Costar plates (5 × 10⁵ cells/well) and infected with SINV mutants at an MOI of 20 PFU/cell. At 6 h PI, media were replaced by 0.8 ml of Dulbecco modified Eagle medium (DMEM) lacking methionine and supplemented with 0.1% FBS and 20 μCi of [³⁵S]methionine/ml. After 30 min of incubation at 37°C, cells were collected and resuspended in standard protein gel loading buffer. Equal amounts of the protein samples were loaded onto 10% NuPAGE gels. The gels were dried and autoradiographed.

Confocal microscopy.

Cells were seeded in 8-well Ibidi chambers (5 × 10³/well) and incubated overnight at 37°C. They were then infected with the packaged replicons indicated in the figures. At the times postinfection indicated in the figure legends, cells were fixed with 4% paraformaldehyde (PFA) for 15 min, permeabilized, and stained with Alexa Fluor 555 phalloidin and Hoechst dye. The image stacks of 6 optical sections were acquired on a Zeiss LSM700 confocal microscope with a 63× 1.4-numerical-aperture (NA) PlanApochromat oil objective. The images were assembled using Imaris software (Bitplane AG).