NF-κB Activation Promotes Alphavirus Replication in Mature Neurons

Mosquito-borne alphaviruses are a significant and growing cause of viral encephalomyelitis worldwide. The outcome of alphaviral neuronal infections is host age dependent and greatly affected by neuronal maturation status, with differentiated, mature neurons being more resistant to infection than undifferentiated, immature neurons. The biological factors that change during neuronal maturation and that influence the outcome of viral infection are currently only partially defined. These studies investigated the role of NF-κB in determining the outcome of alphaviral infection in mature and immature neurons. Inhibition of canonical NF-κB activation decreased alphavirus replication in mature neurons by regulating protein synthesis and limiting the production of the viral structural proteins but had little effect on viral replication in immature neurons or fibroblasts. Therefore, NF-κB is a signaling pathway that influences the maturation-dependent outcome of alphaviral infection in neurons and that highlights the importance of cellular context in determining the effects of signal pathway activation.

ABSTRACT

Alphaviruses are enveloped, positive-sense RNA viruses that are important causes of viral encephalomyelitis. Sindbis virus (SINV) infects the neurons of rodents and is a model for studying factors that regulate infection of neuronal cells. The outcome of alphavirus infection of the central nervous system is dependent on neuronal maturation status. Differentiated mature neurons survive and control viral replication better than undifferentiated immature neurons. The cellular factors involved in age-dependent susceptibility include higher levels of antiapoptotic and innate immune factors in mature neurons. Because NF-κB pathway activation is required for the initiation of both apoptosis and the host antiviral response, we analyzed the role of NF-κB during SINV infection of differentiated and undifferentiated rat neuronal cells. SINV infection induced canonical NF-κB activation, as evidenced by the degradation of IκBα and the phosphorylation and nuclear translocation of p65. Inhibition or deletion of the upstream IκB kinase substantially reduced SINV replication in differentiated but not in undifferentiated neuronal cells or mouse embryo fibroblasts. NF-κB inhibition did not affect the establishment of infection, replication complex formation, the synthesis of nonstructural proteins, or viral RNA synthesis in differentiated neurons. However, the translation of structural proteins was impaired, phosphorylation of the α subunit of eukaryotic translation initiation factor 2 (eIF2α) was decreased, and host protein synthesis was maintained, suggesting that NF-κB activation was involved in the regulation of translation during infection of mature neurons. Inhibition or deletion of double-stranded RNA-activated protein kinase (PKR) also decreased eIF2α phosphorylation, the translation of viral structural proteins, and virus production. Therefore, canonical NF-κB activation synergizes with PKR to promote SINV replication in differentiated neurons by facilitating viral structural protein translation.

IMPORTANCE Mosquito-borne alphaviruses are a significant and growing cause of viral encephalomyelitis worldwide. The outcome of alphaviral neuronal infections is host age dependent and greatly affected by neuronal maturation status, with differentiated, mature neurons being more resistant to infection than undifferentiated, immature neurons. The biological factors that change during neuronal maturation and that influence the outcome of viral infection are currently only partially defined. These studies investigated the role of NF-κB in determining the outcome of alphaviral infection in mature and immature neurons. Inhibition of canonical NF-κB activation decreased alphavirus replication in mature neurons by regulating protein synthesis and limiting the production of the viral structural proteins but had little effect on viral replication in immature neurons or fibroblasts. Therefore, NF-κB is a signaling pathway that influences the maturation-dependent outcome of alphaviral infection in neurons and that highlights the importance of cellular context in determining the effects of signal pathway activation.

INTRODUCTION

Sindbis virus (SINV) is an enveloped positive-sense RNA virus that is a prototypical member of the Alphavirus genus (family Togaviridae). Alphaviruses are mostly transmitted to vertebrate hosts by mosquitoes and are a growing cause of disease worldwide with expanding geographic ranges (1–4). Alphaviruses are broadly categorized into two subsets. The Old World alphaviruses. which include Ross River, O’nyong-nyong, chikungunya, and Sindbis viruses. primarily cause outbreaks of fever and arthralgia with occasional neurologic disease (5), while the New World alphaviruses, which include Venezuelan equine encephalitis (VEE), eastern equine encephalitis, and western equine encephalitis viruses, tend to infect neurons to cause encephalomyelitis, which can lead to death or long-lasting neurologic sequelae in those that survive (6, 7).

While SINV infection typically causes rash, arthralgia, and myalgia in humans (8), in rodents, SINV preferentially infects neurons, causes encephalomyelitis, and is a valuable model for studying the pathogenesis of alphaviral encephalitis and the biological factors that regulate infection of neurons (9, 10). The outcome of SINV infection of neurons is age dependent: neonatal mice develop acute fatal disease, while adult or weanling mice recover from infection (9, 11, 12). This difference in susceptibility to fatal encephalomyelitis has been linked to neuronal maturation (13, 14). Mature differentiated neurons are postmitotic cells that can withstand biological insults, such as stress, growth factor withdrawal, and viral infection, better than immature, undifferentiated dividing neurons. In vitro, differentiated mature neurons restrict alphavirus replication and have prolonged survival following infection compared with undifferentiated immature neurons (15–17). Similar patterns of age-dependent susceptibility have been observed with other neurotropic viruses, including flaviviruses, reoviruses, herpesvirus, retroviruses, and enteroviruses (18–25).

The biological factors associated with neuronal maturation that regulate susceptibility to viral infection are only partially defined. Identified maturation-dependent changes include increased production of antiapoptotic factors, such as Bcl-2 and fractalkine, which inhibit or delay virus-induced apoptosis (12, 26–28), and increased expression of innate immune factors, including Toll-like receptors 3 and 9, retinoic acid-inducible gene I, interferon (IFN) regulatory factors 3 and 7, beta interferon (IFN-β), and IFN-stimulated gene 15 (15, 29). However, neither the overexpression nor the depletion of the factors studied thus far recapitulates the age-dependent viral suppression phenotype, suggesting that additional factors remain to be discovered. Activation of the NF-κB pathway, a conserved signal transduction pathway present in nearly all mammalian cell types, is a common signaling event required for both the initiation of apoptosis and the induction of the antiviral response (reviewed in reference 30). We therefore hypothesized that NF-κB signaling may play a role in determining the outcome of viral infection in neurons.

The NF-κB signaling pathway is involved in a wide range of biological processes, including innate/adaptive immunity, development, proliferation, cell apoptosis, and survival. The NF-κB family of transcription factors includes five constitutively expressed proteins, p65, p52, p50, RelB, and c-Rel, that form hetero- or homodimers within the cytosol. At baseline, NF-κB dimers are maintained in an inactive form in the cytoplasm but can be rapidly activated in response to diverse stimuli, including cytokines, growth factors, and pathogens. Activation can occur by either the canonical or the noncanonical pathway. In canonical NF-κB signaling, the NF-κB dimers are held in an inactive form by physical interaction with a member of the IκB protein family. Activation of the upstream IκB kinase (IKK) complex (IKKα, IKKβ, IKKγ) following cell stimulation results in IKKβ-mediated phosphorylation of the bound IκBα protein, leading to its ubiquitination and proteasomal degradation with the release of the NF-κB dimer. The liberated NF-κB dimer containing p65 translocates to the nucleus and binds κB elements within DNA to modulate gene expression. NF-κB activation can result in a broad spectrum of potential outcomes due to combinatorial diversity at each step of the pathway, from the specific IKKβ subunit activated to the IκB protein involved, the NF-κB dimer components, posttranslational modifications, and the availability of cofactors that affect DNA-binding specificity (30, 31). Therefore, cell type and cellular context shape the response to NF-κB activation, as cell types differ in their baseline levels of the NF-κB-related proteins, chromatin organization, and transcriptional coactivators (30). Studies of animals deficient in components of the NF-κB pathway have shown that NF-κB plays a critical role in neurogenesis in the developing central nervous system (CNS), while in the adult mature CNS, NF-κB plays key roles in memory, learning, and recovery following injury (32). While all subunits within the NF-κB family are detectable within the CNS, the major DNA-binding complexes in the developing nervous system, cRel/p65, p65/p50, and p50 homodimers, shift to predominantly p65/p50 with maturation (33).

To investigate how NF-κB activation affects age-dependent neuronal susceptibility to viral infection, we studied SINV infection in neuron-derived cell lines that can be differentiated into mature, nondividing neuronal cells in vitro (34, 35). SINV replication is restricted in differentiated AP-7 (dAP-7) cells and differentiated CSM14.1 (dCSM14.1) cells in comparison with that in undifferentiated, cycling AP-7 (cAP-7) cells, similar to the observations in primary neuronal cultures (15–17). While inhibition of NF-κB activation decreases SINV-induced apoptosis in AT-3 rat adenocarcinoma cells and N18 mouse neuroblastoma cells (36–38), an effect on SINV replication has not been evaluated. In the current study, we show that SINV infection of neurons induced canonical NF-κB activation and persistent nuclear translocation of the p65/p50 NF-κB dimer and that inhibition or deletion of IKKβ decreased SINV replication in mature neurons but not in immature neurons or fibroblasts, indicating that the effects of virus-induced NF-κB activation are context specific and affected by neuronal maturation status. Analysis of SINV replication demonstrated that NF-κB activation promotes the translation of the SINV structural proteins in mature neurons without an effect on earlier replication steps.

RESULTS

SINV infection induces prolonged canonical NF-κB activation in neurons.

To determine how neuronal maturation affects virus replication and NF-κB activation following SINV infection, cycling undifferentiated cAP-7 cells and postmitotic differentiated dAP-7 cells were infected with the TE strain of SINV with a BHK-21 cell multiplicity of infection (MOI) of 10 (which initially infects ∼10% of dAP-7 cells) at their respective culture temperatures of 33°C and 39°C. As previously reported (15, 16), virus production was restricted in mature neurons compared to immature neurons (Fig. 1A) independently of the incubation temperature (16). To assess the changes in host cellular responses to infection, lysates from infected cAP-7 and dAP-7 cells were analyzed for signaling pathway activation using a reverse-phase protein array (RPPA) (39). NF-κB pathway activation, as indicated by the phosphorylation of the NF-κB protein p65 and the degradation of IκBα, occurred in both cell types following infection but was more rapid in the immature neurons than in the mature neurons (Fig. 1B).

FIG 1.

SINV replication and induction of NF-κB activation in differentiated and cycling AP-7 cells. cAP-7 and dAP-7 cells were infected with SINV at 33°C and 39°C, respectively. (A) Supernatants were assayed for infectious virus by plaque assay in BHK-21 cells. (B) Reverse-phase protein array (RPPA) analysis of phosphorylated p65 (p-p65; S536) normalized to total p65 and IκBα normalized to β-actin in SINV-infected cAP-7 and dAP-7 cells. Values indicate the fold increase relative to the level in matched mock-infected samples. Data are presented as the mean ± SD for triplicate samples. **, P < 0.01; ***, P < 0.001; ****, P < 0.0001.

To confirm SINV infection-induced NF-κB activation in dAP-7 cells, immunoblot analysis of cell lysates was performed. SINV infection but not mock infection induced canonical NF-κB activation, as indicated by the phosphorylation of IKKβ, the upstream kinase that phosphorylates IκBα; by the increased phosphorylation of the NF-κB protein p65 (a 5-fold increase over that for mock-infected cells at 24 h after infection; P < 0.001); and by degradation of IκBα (Fig. 2A and B). While NF-κB activation is normally transient and controlled within hours, SINV-induced NF-κB activation was persistent with long-lasting p65 phosphorylation as well as a sustained absence of IκBα (Fig. 1B and 2A and B). To determine whether activation resulted in p65 nuclear localization, infected dAP-7 cells were also evaluated by immunocytochemistry. SINV infection but not mock infection resulted in complete p65 nuclear translocation by 24 h after infection (Fig. 2C), confirming that SINV replication induces long-lasting canonical NF-κB activation in neurons.

FIG 2.

SINV infection of differentiated AP-7 cells induced prolonged canonical NF-κB activation. dAP-7 cells were infected with SINV or mock infected (with medium) at 39°C. (A) Immunoblot analysis of cell lysates for canonical NF-κB activation using antibodies against phosphorylated p65 (S536), IκBα, phosphorylated IKKα/β (S176/S177), and total IKKβ; viral protein production using antibodies against SINV nonstructural protein 2 (nsP2); and β-actin. hpi, hours postinfection. (B) Densitometric analysis of the levels of phosphorylated p65 and IκBα normalized to those of β-actin. Data are presented as the mean ± SD from three independent experiments. **, P < 0.01; ***, P < 0.001. (C) Immunocytochemistry to determine p65 cellular localization at 24 h after infection. Images were taken at a magnification of ×40 and are representative of those from three independent experiments.

Inhibition of NF-κB activation decreases SINV replication in differentiated neurons.

To investigate whether NF-κB activation affects the outcome of SINV infection in mature neurons, we treated cells with 2-[(aminocarbonyl) amino]-5-(4-flurophenyl)-3-thiophenecarboxamide (TPCA-1), an IKKβ-selective small-molecule ATP competitive inhibitor that prevents the phosphorylation of IκBα (40). To determine the optimal conditions for inhibition, dAP-7 cells were pretreated with several concentrations of TPCA-1 or the dimethyl sulfoxide (DMSO) vehicle control, stimulated with the NF-κB activator tumor necrosis factor alpha (TNF-α) for 30 min, and evaluated by immunoblotting. TPCA-1 concentrations of 10 μM or higher were sufficient to decrease TNF-α-induced p65 phosphorylation to the baseline levels of unstimulated cells (Fig. 3A) and prevent its nuclear localization (Fig. 3B) without toxicity.

FIG 3.

TPCA-1 treatment inhibited canonical NF-κB activation in differentiated AP-7 cells. dAP-7 cells were pretreated with the NF-κB inhibitor TPCA-1 or DMSO for 1 h at the indicated concentrations and stimulated with recombinant rat TNF-α (10 μg/ml) for 30 min at 39°C. (A) Protein lysates were analyzed by immunoblotting for phosphorylated p65 (S536) and β-actin. (B) Cells were fixed and analyzed by immunocytochemistry for p65 localization. Images were taken at a magnification of ×40 and are representative of those from two independent experiments.

Treatment of dAP-7 cells with 10 μM TPCA-1 at 1 h after infection reduced SINV-induced NF-κB activation, as evidenced by decreased p65 phosphorylation (which was 2.5-fold lower than that for the DMSO-treated control at 24 h; P < 0.01) and the delayed degradation of IκBα (Fig. 4A and B). Inhibitor treatment also decreased the production of the SINV structural proteins (3.5-fold less capsid than that for the DMSO-treated control at 24 h; P < 0.0001) (Fig. 4A and C), suggesting that NF-κB activation promoted SINV replication. However, production of the SINV nonstructural proteins, indicated by the production of nonstructural protein 2 (nsP2), was not affected by inhibitor treatment (Fig. 4A and D).

FIG 4.

NF-κB inhibition reduced SINV structural protein production in differentiated AP-7 cells. dAP-7 cells were infected with SINV and treated with either TPCA-1 (10 μM) or DMSO at 39°C. (A) Whole-cell lysates were probed by immunoblotting with antibodies against phosphorylated p65, IκBα, SINV nsP2, SINV structural proteins (pE2, E1/E2, capsid), and β-actin. (B to D) Densitometric analysis of phosphorylated p65 (B), SINV capsid (C), and SINV nsP2 (D) normalized to β-actin. Data are presented as the mean ± SD from three independent experiments. *, P < 0.05; **, P < 0.01.

To confirm that inhibition of NF-κB activation reduced SINV replication, supernatants from infected dAP-7 cells were assessed for infectious virus production by plaque assay. Inhibitor-treated cells produced significantly less virus than vehicle-treated cells through 24 h but comparable amounts by 36 h (Fig. 5A), suggesting that NF-κB inhibition does not affect overall dAP-7 susceptibility to infection but slows viral replication. Assessment of viral RNA by quantitative reverse transcription-PCR (qRT-PCR) showed that the levels of intracellular genomic RNA (detected with nsP2-specific primers) and genomic plus subgenomic RNA (detected with E2-specific primers) were comparable until 12 h after infection but were lower in inhibitor-treated cells at later times (Fig. 5B). These data suggest that NF-κB inhibition affects a late step in the SINV life cycle, after the synthesis of genomic and subgenomic RNA but before the release of new virions. The delayed difference in viral RNA production may reflect decreased subsequent rounds of infection due to lower virus production in the inhibitor-treated cells.

FIG 5.

Inhibition of NF-κB activation reduced SINV replication in differentiated neurons. dAP-7 or dCSM14.1 cells were infected with SINV and treated with TPCA-1 (10 μM) or DMSO at 39°C. (A) Supernatants from dAP-7 cells were assayed for infectious virus by plaque assay in BHK-21 cells. (B) Intracellular viral RNA levels from infected dAP-7 cells were evaluated by qRT-PCR for SINV nsP2 (genomic RNA) and E2 (subgenomic plus genomic RNA) and normalized to the copy numbers of GAPDH RNA. (C) Supernatants from dCSM14.1 cells were assayed for infectious virus by plaque assay in BHK-21 cells. Data are presented as the mean ± SD from three independent experiments. ns, not significant; ***, P < 0.001; ****, P < 0.0001.

To determine whether the effect of NF-κB inhibition on SINV replication was specific to the AP-7 cell line, a second differentiating neuronal cell line, CSM14.1 rat nigral neuron-derived cells, that also restrict SINV replication following differentiation was analyzed (16). Differentiated CSM14.1 (dCSM14.1) cells replicate SINV more slowly than AP-7 cells and reach peak supernatant viral titers at 48 to 72 h postinfection (16). Therefore, dCSM14.1 cells were treated with TPCA-1 (10 μM) or DMSO beginning 24 h after SINV infection (BHK-21 cell MOI, 10), and the supernatants were assessed for infectious virus production by plaque assay. Similar to the observations with dAP-7 cells, inhibition of NF-κB activation significantly decreased SINV replication in dCSM14.1 cells (Fig. 5C).

Inhibition of NF-κB activation has little to no effect on SINV replication in undifferentiated neurons or MEFs.

To determine how neuronal maturation affects the influence of NF-κB activation on the outcome of SINV infection, we assessed the effect of inhibition of NF-κB activation on SINV replication in undifferentiated cycling AP-7 (cAP-7) cells. cAP-7 cells were infected with SINV (MOI, 10), treated with TPCA-1 (10 μM) or DMSO, and evaluated for NF-κB activation and viral protein synthesis by immunoblotting. While SINV also induced canonical NF-κB activation in cAP-7 cells, evidenced by the phosphorylation of p65 and the degradation of IκBα (Fig. 1B and 6A), NF-κB inhibition did not significantly affect the synthesis of SINV structural proteins or the production of infectious virus (Fig. 6A and B), indicating that the NF-κB-related factors that contribute to SINV replication are more important for mature neurons. To further explore cell type specificity, wild-type (WT) and IKKβ-deficient mouse embryonic fibroblasts (MEFs) were also infected with SINV (Fig. 6C). No difference in the kinetics of viral replication in the MEFs was observed, confirming the cellular context-specific role of NF-κB in SINV replication in mature neurons.

FIG 6.

Inhibition of NF-κB activation had little effect on SINV replication in cycling AP-7 cells or mouse embryonic fibroblasts. cAP-7 cells were infected with SINV and treated with TPCA-1 (10 μM) or DMSO at 33°C. (A) Protein lysates were assessed by immunoblotting for phosphorylated p65 (S536), IκBα, SINV structural proteins, and β-actin. (B) Supernatants from cAP-7 cells were assessed for infectious virus by plaque assay in BHK-21 cells. (C) Wild-type (WT) or IKKβ-deficient mouse embryonic fibroblasts (MEFs) were infected with SINV (MOI, 1) at 37°C. Supernatants were assessed for infectious virus by plaque assay in BHK-21 cells. Data are presented as the mean ± SD from three independent experiments.

NF-κB activation is not necessary for establishment of SINV infection or replication complex formation in differentiated neurons.

To determine how canonical NF-κB activation promotes SINV replication in mature neurons, individual steps of the viral life cycle were evaluated for sensitivity to NF-κB inhibitor treatment. The overall efficiency of the initial infection was assessed by infectious center assays with dAP-7 cells treated with TPCA-1 (10 μM) or DMSO. Productively infected cells were identified by plaque formation following cocultivation on BHK-21 cells. No difference in the average numbers or percentages of infectious centers was observed between the inhibitor-treated (7%) and vehicle-treated (8.8%) cells (Fig. 7A and B), indicating that NF-κB inhibition did not affect early steps of SINV infection, such as attachment and entry. The finding of initial infection of <10% of the dAP-7 cells supports the possibility that the observed delayed difference in viral RNA levels (Fig. 5B) may be the result of the slower spread of the virus through the culture due to an inhibitor-induced decrease in secondary rounds of infection.

FIG 7.

NF-κB inhibition does not affect establishment of infection or formation of replication complexes containing double-stranded RNA in differentiated AP-7 cells. dAP-7 cells were infected with SINV and treated with TPCA-1 (10 μM) or DMSO at 39°C. (A, B) At 4 h after infection, cells were trypsinized, counted, serially diluted, and plated onto BHK-21 cell monolayers to identify infected cells by plaque formation. Data are presented as the log₁₀ number of infectious centers per 10⁵ plated cells (A) or the percentages of plated cells that generated infectious centers (B). (C, D) Flow cytometry to assess double-stranded RNA as a measure of replication complex formation. At the indicated time points, cells were stained with a viability dye, fixed, permeabilized, probed with antibody against dsRNA, and analyzed by flow cytometry. (C) Percentage of live cells positive for dsRNA. (D) Mean fluorescence intensity of cells positively stained for dsRNA. Data are presented as the mean ± SD and represent those from three independent experiments. ****, P < 0.0001.

Immediately following viral entry, SINV nonstructural proteins 1 to 4 are translated as a polyprotein from the viral genome to form the RNA replicase responsible for the synthesis of negative-strand template RNA, the amplification of viral genomic RNA, and the production of the subgenomic RNA which encodes the SINV structural proteins (41, 42). To determine whether NF-κB inhibition affects the establishment of replication complexes and the efficiency of viral RNA replication, flow cytometry was performed to detect double-stranded RNA (dsRNA) replication intermediates in SINV-infected dAP-7 cells treated with TPCA-1 or vehicle (Fig. 7C). At early time points, there was no detectable difference in the percentages of live cells positive for dsRNA. By 18 h after infection, there were fewer cells positive for dsRNA in the inhibitor-treated group (46% versus 67%; P < 0.0001); however, by 24 h, virtually all cells were positive. Similar to what was observed for intracellular viral RNA levels (Fig. 5B), this delayed difference was likely due to effects on virus production rather than a direct effect of NF-κB inhibition on the formation of replication complexes. In addition, the mean fluorescence intensity (MFI) of dsRNA-positive cells did not differ between the inhibitor- and vehicle-treated groups (Fig. 7D), indicating that individual infected cells formed similar numbers of replication complexes with equal levels of double-stranded RNA and viral RNA amplification.

NF-κB activation facilitates SINV structural protein translation in differentiated neurons.

Because viral RNA synthesis did not appear to be affected by NF-κB inhibition, while structural protein synthesis was decreased, we next evaluated how NF-κB inhibition affected protein translation. Infected dAP-7 cells were treated with TPCA-1 or vehicle, pulse-labeled at regular intervals with ³⁵S-labeled Cys/Met for 1 h, and assessed for ³⁵S incorporation into proteins. By 6 h after infection, distinct bands corresponding to the viral structural proteins pE2, E1/E2, and capsid were visible, and by 12 h, these were the predominant proteins synthesized due to the simultaneous shutoff of host protein synthesis (Fig. 8A). NF-κB inhibitor treatment decreased structural protein synthesis at all time points assessed, with roughly 40% less E1/E2 protein being synthesized in inhibitor-treated cells than in vehicle-treated cells by 12 h after infection (Fig. 8B), indicating that NF-κB inhibition decreased viral protein translation from the subgenomic RNA. Host protein translation was preserved longer in the inhibitor-treated cells than in vehicle-treated cells, with cellular protein bands still being visible 18 h after infection. Therefore, NF-κB inhibitor-induced suppression of viral structural protein translation was associated with a delayed inhibition of host protein translation, which could result in sustained host antiviral protein production and competition for translation factors and resources.

FIG 8.

NF-κB inhibition decreased SINV structural protein translation in differentiated AP-7 cells. dAP-7 cells were infected with SINV and treated with TPCA-1 (10 μM) or DMSO at 39°C. At the indicated time points, the cells were pulsed for 1 h with ³⁵S-labeled cysteine/methionine and lysed. (A) Protein synthesis was analyzed by SDS-PAGE autoradiography and normalized to the counts per lane. (B) Relative density of the labeled E1/E2 band at the indicated time points. Data are presented as the mean ± SD and represent those from three independent experiments. ****, P < 0.0001.

In SINV infection, both the nonstructural and the structural proteins are initially translated as polyproteins that are sequentially proteolytically cleaved into the individual protein components. Because only the processed individual structural proteins were detectable in the previous pulse-labeling experiments, we next sought to determine by pulse-chase analysis whether the differences in replication were due to differences in the processing of the structural polyprotein rather than decreased overall viral protein translation. SINV-infected dAP-7 cells treated with TPCA-1 or vehicle were pulse-labeled for 15 min at 6 h after infection and chased for the times indicated in Fig. 9, and the SINV structural proteins were immunoprecipitated (Fig. 9A). Structural polyprotein processing was quantified by measuring the E2 density as a percentage of the total E2-containing proteins (the polyprotein precursor pE3E2E1, pE2, and E2). Despite higher overall band densities in the vehicle-treated lysates than in the inhibitor-treated lysates, the rate of structural protein processing did not differ (Fig. 9B). Therefore, NF-κB inhibition does not affect processing of the structural polyprotein but instead impairs translation of the SINV structural proteins. To determine if processing of the SINV nonstructural proteins is affected by NF-κB inhibition, pulse-chase lysates were immunoprecipitated with polyclonal antibody against SINV nonstructural protein 3 (nsP3) (Fig. 9C). The rate of nsP3 processing was quantified by measuring the nsP3 density as a percentage of the total nsP3-containing proteins (P1234, P123, P23, and nsP3). No difference in nsP3 processing was observed (Fig. 9D), confirming that inhibition of NF-κB activation does not affect SINV protein processing.

FIG 9.

NF-κB inhibition decreased the translation of SINV structural proteins but did not affect structural protein processing in differentiated AP-7 cells. Pulse-chase analysis of newly synthesized SINV structural proteins (A, B) and nonstructural proteins (C, D). dAP-7 cells were infected with SINV and treated with TPCA-1 (10 μM) at 39°C. At 6 h after infection, the cells were pulsed for 15 min with ³⁵S-labeled cysteine/methionine and chased for 20 to 90 min. Labeled proteins were immunoprecipitated (IP) with polyclonal antibody against the SINV structural proteins (A) or nsP3 (C), separated by SDS-PAGE, and visualized by autoradiography. (B) Relative levels of processed E2 at each time point, expressed as a percentage of total pE3E2E1, pE2, and E2. (D) Relative levels of processed nsP3 at each time point, expressed as a percentage of total P1234, P123, P23, and nsP3. Data are representative of those from two independent experiments.

SINV replication is impaired in IKKβ-deficient differentiated neurons.

Because chemical inhibitors can have confounding off-target effects, we generated an Ikbkb (IKKβ gene)-deficient AP-7 cell line using CRISPR/Cas9-mediated genome editing. To control for the effects of transient Cas9 expression and repeated cell passaging during the generation of a monoclonal cell line, a wild-type (WT) AP-7 cell line was generated in parallel using a nontargeting single guide RNA (gRNA). WT and IKKβ-deficient AP-7 cells were differentiated, infected with SINV, and assessed for NF-κB activation by immunoblotting and for SINV production by plaque assay (Fig. 10). Phosphorylated and total IKKβ protein were detectable in the WT cells but not in the IKKβ^−/− cells, and the amount of phosphorylated p65 was significantly reduced in the IKKβ^−/− cells in response to infection, further demonstrating that IKKβ plays a key role in the activation of the canonical NF-κB pathway during SINV infection (Fig. 10A). Similar to observations with inhibitor treatment, the IKKβ^−/− dAP-7 cells exhibited decreased viral structural protein synthesis (Fig. 10A) and virus yield (Fig. 10B) compared to the WT dAP-7 cells (a 10-fold difference at 12 h; P < 0.0001). To determine if IKKβ deficiency affects SINV replication in immature neurons, undifferentiated IKKβ^−/− and WT cAP-7 cells were assessed. Early (6 h) after infection, IKKβ^−/− cAP-7 cells produced less virus than WT cAP-7 cells, but no difference was detected at later times (Fig. 10C). These data confirm that canonical NF-κB signaling through IKKβ improves SINV replication in neurons and further demonstrate that this effect is more important for replication in mature neurons than for replication in immature neurons.

FIG 10.

The absence of IKKβ decreased SINV replication in differentiated AP-7 cells more than in cycling AP-7 cells. (A) Immunoblot analysis of wild-type (WT) dAP-7 cells and IKKβ-deficient dAP-7 cells infected with SINV at 39°C. Protein lysates were probed for phosphorylated IKKα/β (S176/S177), total IKKβ, phosphorylated p65 (S536), total p65, SINV structural proteins, SINV nsP2, and β-actin. (B, C) Infectious virus production by SINV-infected WT and IKKβ-deficient differentiated AP-7 cells (B) and cycling AP-7 cells (C), as measured by plaque assay in BHK-21 cells. Data are presented as the mean ± SD and are representative of those from three independent experiments. *, P < 0.05; ****, P < 0.0001.

Inhibition of PKR activation and eIF2α phosphorylation also decrease SINV replication in differentiated neurons.

The previous experiments indicated that canonical NF-κB activation via IKKβ facilitates the translation of the SINV structural proteins and the inhibition of host protein synthesis. Host translation inhibition is in part a result of activation of the integrated stress response (ISR), a conserved eukaryotic cellular response to stresses, such as viral infection, starvation, and UV irradiation, evolved to conserve resources and promote the translation of mRNAs involved in cell recovery and the maintenance of homeostasis (43, 44). The central event of the ISR is phosphorylation of the alpha subunit of eukaryotic translation initiation factor 2 (eIF2α), which prevents the formation of the 43S preinitiation complex that interacts with the 5′ cap of mRNAs during the normal initiation of eukaryotic translation. Translation of the SINV structural but not nonstructural proteins occurs independently of multiple translation initiation factors largely due to the presence of a hairpin loop in the SINV subgenomic mRNA downstream of the AUG codon (45–49). To determine if NF-κB activation facilitates SINV structural protein translation by promoting host translation inhibition via the ISR, the levels of eIF2α phosphorylation in infected dAP-7 cells treated or untreated with TPCA-1 were compared by immunoblotting (Fig. 11A). The inhibitor-treated cells exhibited less eIF2α phosphorylation than DMSO-treated control cells (1.5-fold less at 48 h; P < 0.05) (Fig. 11B), suggesting NF-κB pathway participation in activation of the ISR during SINV infection.

FIG 11.

NF-κB inhibition decreased SINV-induced PKR and eIF2α phosphorylation. (A) Immunoblot analysis of PKR, phosphorylated eIF2α (S51), and total eIF2α in SINV-infected dAP-7 cells treated with the NF-κB inhibitor TPCA-1 (10 μM) or DMSO at 39°C. (B) Densitometric analysis of phosphorylated PKR (p-PKR; upper band from panel A) normalized to total PKR (upper and lower bands from panel A) and phosphorylated eIF2α (p-eIF2α) normalized to eIF2α. Data are presented as the mean ± SD and are representative of those from three independent experiments. *, P < 0.05.

Four mammalian kinases have the capacity to phosphorylate eIF2α: interferon-induced dsRNA-activated protein kinase (PKR), PKR-like endoplasmic reticulum kinase (PERK), general control nonderepressible 2 (GCN2), and heme-regulated inhibitor (HRI). These proteins share a conserved kinase domain but have distinct regulatory domains that allow for stimulus-specific activation (50). Among these, PKR was of particular interest due to its activation by cytoplasmic double-stranded RNA and established interaction with IKKβ (51–54). Activated PKR phosphorylates multiple targets, including eIF2α, and can induce canonical NF-κB activation through physical interaction with and activation of IKKβ; however, the specific details of the interaction are unclear. PKR activation was therefore evaluated in the TPCA-1-treated cell lysates and was decreased with NF-κB inhibition (1.5-fold decreased phospho-PKR relative to that for the control at 24 h; P = 0.07) (Fig. 11A and B).

To determine how PKR activation affects SINV replication in mature neurons, SINV-infected dAP-7 cells were treated with the PKR inhibitor s6,8-dihydro-8-(1H-imidazol-5-ylmethylene)-7H-pyrrolo[2,3-g]benzothiazol-7-one (C16; 1 μM) or DMSO and evaluated for infectious virus production by plaque assay (Fig. 12A). At 6 h after infection, cells treated with C16 produced roughly 5-fold more virus than the vehicle-treated cells, but by 12 h there was no difference. By 24 h, PKR inhibitor-treated cells produced 10-fold less virus than vehicle-treated cells, indicating that during the late stages of the viral life cycle, PKR activation is advantageous for SINV production. To confirm the efficacy of PKR inhibitor treatment, protein lysates were evaluated for the phosphorylation of eIF2α and for SINV protein production by immunoblotting (Fig. 12B and C). While C16 treatment did not completely ablate eIF2α phosphorylation, the relative levels of phosphorylated to total eIF2α were diminished (2-fold lower 12 h after treatment) (Fig. 12C). Furthermore, the levels of total eIF2α were also lower at 36 h, which may indicate decreased eIF2α synthesis with PKR inhibition. Inhibition of PKR decreased SINV structural protein production (1.5-fold decreased pE2 expression 36 h after infection relative to that for the control; P = 0.18), while nonstructural protein production was higher (2-fold higher nsP2 expression 36 h after infection; P < 0.05) (Fig. 12C), supporting the hypothesis that the shutoff of host protein synthesis via PKR activation promotes translation of the SINV structural proteins from subgenomic RNA but not nonstructural proteins from genomic RNA.

FIG 12.

Effect of PKR inhibition results on SINV replication in mature neurons. dAP-7 cells were infected with SINV and treated with the PKR inhibitor C16 (1 μM) or DMSO at 39°C. (A) Supernatants were assayed for infectious virus by plaque assay in BHK-21 cells. (B) Immunoblot analysis for phosphorylated eIF2α (S51), SINV nsP2, SINV structural proteins (pE2, E1/E2, capsid), total eIF2α, and β-actin. (C) Densitometric analysis of phosphorylated eIF2α normalized to total eIF2α and SINV nsP2 and SINV pE2 normalized to β-actin. Data are presented as the mean ± SD and are representative of those from three independent experiments. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001.

To confirm the role of PKR in SINV replication in mature neurons, a PKR-deficient AP-7 cell line was generated via CRISPR/Cas9-mediated genome editing as described above. WT and PKR^−/− AP-7 cells were differentiated, infected with SINV, and assessed for viral replication by plaque assay (Fig. 13A). Although no difference in infectious virus production was observed at early time points following infection, at 24, 36, and 48 h, 2-fold less virus was produced by PKR^−/− dAP-7 cells than by WT dAP-7 cells, suggesting that PKR activity is involved in virus production during later stages of infection. The infected PKR^−/− dAP-7 cells also exhibited less of a cytopathic effect, with cell viability at 24 h being approximately 80% of that at day 0 for PKR^−/− dAP-7 cells and 30% of that at day 0 for WT dAP-7 cells, as determined by trypan blue exclusion (P < 0.0005) (Fig. 13B). Activated PKR induces apoptosis in nonneuronal cells with vaccinia virus infection or with recombinant PKR overexpression via NF-κB and caspase 8 and 9 activation (55, 56), and our data suggest that PKR activation also plays a role in the death of SINV-infected mature neurons.

FIG 13.

Effect of PKR absence on SINV replication in mature neurons. WT and PKR-deficient dAP-7 cells were infected with SINV at 39°C. (A) Supernatants were assayed for infectious virus by plaque assay in BHK-21 cells. (B) Cell viability relative to day 0 (D0) cell counts, as determined by trypan blue exclusion. (C) Immunoblot analysis for PKR, phosphorylated eIF2α (S51), SINV nsP2, SINV structural proteins (pE2, E1/E2, capsid), and total eIF2α. (D) Densitometric analysis of phosphorylated eIF2α normalized to total eIF2α and SINV nsP2 and SINV pE2 normalized to β-actin. Data are presented as the mean ± SD and are representative of those from three independent experiments. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001.

To investigate the effect of PKR deficiency on SINV protein translation in mature neurons, lysates from PKR-deficient dAP-7 cells were assessed by immunoblotting (Fig. 13C). eIF2α phosphorylation was reduced in the PKR-deficient cells compared to WT dAP-7 cells (3-fold less at 24 h; P < 0.05) but was not eliminated (Fig. 13D), indicating that while PKR is the primary driver of eIF2α phosphorylation following SINV infection, other eIF2α kinases also participate. Although there was no difference in virus production at 12 h after infection (Fig. 13A), PKR-deficient cells synthesized larger amounts of structural proteins than the WT cells at this time (7-fold higher; P < 0.0001) (Fig. 13C and D). However, by 36 h, smaller amounts of nsP2 and E2 were present in the PKR-deficient cell lysates (P < 0.05) (Fig. 13D), mirroring the plaque assay data. Overall, SINV replication in PKR-deficient dAP-7 cells was similar to that observed with inhibitor treatment but less pronounced. This difference may be due to inherent differences between the absence of PKR with gene knockout and transient inactivation of the catalytic domain or may indicate a role for the noncatalytic protein domains of PKR in regulating SINV translation in neurons, possibly as scaffolding for the assembly of signaling complexes and recruitment of downstream effector proteins.

DISCUSSION

The outcome of alphaviral infection of the CNS is age dependent and linked to the maturation status of the neuronal target cells, with increasing resistance to fatal infection occurring as the host ages and the neurons differentiate and mature. We used SINV infection of differentiating AP-7 and CSM14.1 rat neuron-derived cell lines to identify the canonical NF-κB pathway as a previously unrecognized factor that influences the outcome of alphaviral infection in mature neurons by regulating translation. NF-κB was activated in both undifferentiated and differentiated neurons but was more rapidly activated in the undifferentiated cells. SINV replication in differentiated neurons was restricted in comparison with replication in undifferentiated neurons. Inhibition of canonical NF-κB signaling through inactivation of the upstream kinase IKKβ with TPCA-1 inhibitor treatment or CRISPR/Cas9-mediated gene knockout impaired SINV replication in differentiated neurons by decreasing the production of viral structural proteins but did not affect earlier steps of the viral life cycle. Inhibition of NF-κB activation delayed the phosphorylation of eIF2α, a host mechanism for inhibiting cellular mRNA translation under stressed conditions, leading to the slowed shutoff of host protein synthesis after infection and eIF2α-independent translation of the SINV structural proteins. Chemical inhibition of the eIF2α kinase PKR similarly resulted in impaired SINV replication with decreased viral structural protein production. Gene knockout of PKR in differentiated neurons increased SINV protein production initially but decreased late viral protein synthesis and improved cell viability following infection. From these data, we hypothesize that cross talk between the canonical NF-κB and PKR pathways plays a role in regulating host and viral translation in mature neurons. SINV-induced NF-κB activation via IKKβ may promote SINV replication during late infection in mature neurons in part by inducing the activation of PKR to phosphorylate eIF2α, decrease host protein synthesis, and facilitate the translation of the viral structural proteins from subgenomic RNA.

Viruses can manipulate the NF-κB pathway in multiple ways (57, 58). The canonical NF-κB pathway plays a critical role in the early activation of innate antiviral immunity and the induction of apoptosis, but viruses also exploit the NF-κB pathway to enhance viral replication. For example, the long terminal repeat of human immunodeficiency virus (HIV) includes NF-κB binding sites that are essential for viral gene expression (59), and canonical NF-κB activation by LMP1 of Epstein-Barr virus upregulates antiapoptotic genes that prolong cell survival (60). The replication of several RNA viruses is impaired when NF-κB signaling is inhibited, but the mechanisms differ. For instance, influenza virus relies on NF-κB signaling for the induction of proapoptotic factors that regulate viral RNA synthesis and the release of viral RNPs from the nucleus (61–63), while reovirus-induced NF-κB activation regulates organ-specific apoptosis but has little effect on virus replication (64, 65).

Previous studies with alphaviruses exemplify the complexities of virus interactions with NF-κB pathway proteins. NF-κB activation regulates SINV-induced apoptosis of AT-3 rat prostatic carcinoma and N18 mouse neuroblastoma cells, but virus replication was not analyzed (36, 37). Venezuelan equine encephalitis virus (VEEV) activates NF-κB in human astrocytoma cells in association with the interaction of viral nsP3 and IKKβ and a change in the size of the IKKβ complex. IKKβ inhibitors decrease the VEEV yield, but the affected step in replication was not identified (58). Chikungunya virus infection of synovial fibroblasts increases the expression of NF-κB-induced microRNA MiR-146a to regulate innate responses to infection (66). The current study has shown that SINV induces NF-κB activation in neurons and has identified a role for NF-κB activation in facilitating SINV replication in these target cells. IKKβ-deficient undifferentiated neurons had slightly delayed virus production (6 h) (Fig. 10C), but replication was severely impaired in IKKβ-deficient differentiated neurons (Fig. 5A and B and 10B), demonstrating differentiation status specificity to the relationship between NF-κB and SINV replication. The decrease in virus replication in the absence of NF-κB signaling was associated with impaired regulation of translation, which led to delays in the shutoff of host protein synthesis and the decreased production of viral structural proteins (Fig. 4, 8, and 10).

The translation of host and viral proteins is tightly regulated in SINV-infected cells. Early in infection, the synthesis of nonstructural proteins from genomic RNA has requirements similar to those for the synthesis of most cellular proteins from capped mRNAs. The activation of PKR results in eIF2α phosphorylation and a decrease in global cap-dependent translation, which affect the synthesis of both viral nonstructural proteins and most host proteins. Although SINV infection strongly induces PKR activation, the sequestration of SINV dsRNA intermediates within cytopathic vacuoles suggests that additional stimuli may induce the activation of PKR in alphavirus-infected cells. High concentrations of alphaviral capsid protein, for example, correlate with PKR activation in the absence of dsRNA, although the details of the interaction are not understood (67). In response to viral infection, activated PKR can induce canonical NF-κB activation through physical interaction with and activation of IKKβ (51–54). We hypothesize that the reverse interaction, IKKβ-induced PKR activation, can also occur and may provide a positive feedback mechanism that participates in the shutdown of host and viral nonstructural protein synthesis.

Many RNA viruses, including picornaviruses, flaviviruses, and alphaviruses, have evolved genomic features that independently recruit translation initiation factors, despite PKR activation and eIF2α phosphorylation (68). The translation of the SINV structural proteins but not nonstructural proteins can occur independently of multiple translation initiation factors, including eIF2, due to a direct interaction between a secondary structure in the subgenomic RNA and the ribosome, leading to preferential translation of the viral structural proteins following host translation inhibition (49). The relationship between host translation inhibition via PKR and SINV structural protein production in mature neurons was confirmed using an inhibitor of PKR that decreased NF-κB activation, viral structural protein synthesis, and virus production (Fig. 12).

The catalytic domain of PKR interacts with IKKβ, as point mutations in this domain result in the complete loss of dsRNA-induced NF-κB activation (53). Because the catalytic domains of the four known mammalian eIF2α kinases (PKR, PERK, GCN2, and HRI) are conserved, it is possible that IKKβ could share a similar relationship with the other eIF2α kinases. PERK and GCN2 can also be activated in response to SINV infection (69, 70). In immortalized MEFs, however, GCN2 phosphorylates eIF2α early following SINV infection, while PKR is predominantly responsible for eIF2α phosphorylation late in infection, indicating that the eIF2α kinases are activated in response to different stimuli (69). While the knockout of GCN2 in MEFs increases SINV replication due to decreased GCN2-induced inhibition of early translation of the SINV nonstructural proteins, the knockout of PKR decreases SINV replication and protein production, similar to what we observed in inhibitor-treated and PKR^−/− differentiated neurons (48, 71). These data indicate that the different eIF2α kinases play distinct roles in SINV infection and suggest that PKR is involved in the late host translation inhibition that is beneficial for SINV replication.

It is increasingly recognized that the outcomes of signal pathway activation are highly context dependent and shaped by other contemporaneous factors, such as epigenetic history, the presence of heterologous transcription factors or cofactors, and signaling kinetics (30). The NF-κB pathway exemplifies this, and since its initial discovery in association with B cell activation (72), a role for NF-κB signaling has been identified in a wide range of processes, including innate and adaptive immunity, proliferation, cell survival, and death. NF-κB is activated during SINV infection of both immature and mature neurons. Inhibition of IKKβ activity, however, did not significantly affect virus production in immature neurons (Fig. 6B and 10C) but had a prolonged effect on virus replication in mature neurons, where replication is already restricted (Fig. 5 and 10B). It has previously been observed that the developing and mature CNSs differ in their NF-κB transcription factor dimer composition. While the developing CNS of juvenile rodents is characterized by a mixed composition of c-Rel/p50, p65/p50, and p50/50, the mature CNS of adult animals predominantly expresses p65/p50 (33). An in vitro study of primary neurons reported a similar shift in NF-κB protein expression over the course of neuronal differentiation (73). Therefore, one possible explanation for the cell type differences in SINV sensitivity to NF-κB inhibition is a difference in the baseline expression of components of the NF-κB pathway. In addition to the NF-κB dimers, there may also be shifts in IKK and IκB isoform expression during neuronal maturation. Greater relative expression of the canonical NF-κB pathway-associated protein IKKβ, IκBα, or p65/p50 in mature neurons could explain why SINV infection of mature neurons is more affected by IKKβ inhibition and gene knockout than SINV infection of developing neurons and is a topic of future investigation.

The levels of PKR also vary depending upon tissue or cell type. In a study of human squamous mucosa, PKR levels correlated with cell differentiation status: levels were low in the actively dividing cells of the basal layer and progressively increased into the nondividing zone of mature keratinocytes, likely as a mechanism of translation regulation (74). Different levels of PKR could therefore also contribute to the larger role of NF-κB pathway activation in determining the outcome of SINV replication in mature versus immature neurons. A final possibility is that the subsequent upregulation of stress-related genes, such as ATF-4, following eIF2α phosphorylation, rather than the translation inhibition itself, is important for SINV replication and may differ between immature replicating neurons and mature nondividing neurons.

In summary, inhibition or deletion of the canonical NF-κB kinase IKKβ decreased SINV replication, suggesting that canonical NF-κB pathway activation promotes SINV replication and is particularly important in mature neurons. Furthermore, activation of IKKβ increases SINV structural protein translation and accelerates host translation inhibition by promoting the phosphorylation of eIF2α through the activation of PKR. These data highlight the importance of cell type and context in determining the outcome of signal pathway activation and further reveal new differences between immature neurons and differentiated mature neurons that may play a role in age-dependent susceptibility to CNS infections.

MATERIALS AND METHODS

Cell culture.

The rat AP-7 Odora cell line is an olfactory neuron-derived cell line immortalized with a temperature-sensitive simian virus 40 (SV40) T antigen (a gift from Dale Hunter, Tufts University, Boston, MA) (34). AP-7 cells were grown under the permissive conditions of 33°C and 7% CO₂ in Dulbecco’s modified Eagle’s medium (DMEM; Gibco) supplemented with 10% heat-inactivated fetal bovine serum (FBS; Atlanta Biologicals), 100 U/ml penicillin, 100 μg/ml streptomycin, and 2 mM glutamine (Gibco). At roughly 25% confluence, cells were differentiated for 7 days by shifting the cultures to 39°C and 5% CO₂ and supplementing the medium with 1 μg/ml insulin, 20 μM dopamine, and 100 μM ascorbic acid (Sigma). The CSM14.1 cell line is a rat mesencephalic progenitor cell line that was immortalized with a temperature-sensitive SV40 T antigen (a gift from Dale Bredeson, Buck Institute for Age Research, Novato, CA) (35, 75). CSM14.1 cells were grown under the permissive conditions of 31°C and 5% CO₂ in DMEM with 10% FBS, penicillin, streptomycin, and glutamine. For differentiation, 95% confluent CSM14.1 cells were switched to 39°C and 1% FBS and cultured for an additional 4 weeks. Baby hamster kidney (BHK-21) cells and WT and IKKβ-deficient MEFs (gifts from Fengyi Wan, Johns Hopkins Bloomberg School of Public Health) were grown in DMEM with 10% FBS, penicillin, streptomycin, and glutamine at 37°C in 5% CO₂. All cell lines were routinely tested and determined to be negative for mycoplasma using a MycoAlert Plus mycoplasma detection kit (Lonza).

Generation of IKKβ- and PKR-deficient AP-7 cells.

Expression of the IKKβ gene Ikbkb and the PKR gene eif2ak2 in cycling AP-7 cells was deleted using CRISPR/Cas9, and the resulting knockout cells were differentiated as described above. The backbone plasmid pSpCas9(BB)-2A-Puro (PX459; v2.0) was a gift from Feng Zhang (Addgene plasmid number 62988). IKKβ-specific and PKR-specific CRISPR and nontargeting control plasmids were generated by annealing guide RNA (gRNA)-specific oligonucleotides and then subcloning the oligonucleotides into the PX459 vector as previously described (76). The IKKβ-targeting single gRNA sequence is 5′-GCT GAA CCA TCC CAA CGT GG-3′, the PKR-targeting single gRNA sequence is 5′-GGG CAG ACT ACG TAT GGT ACT GG-3′, and the nontargeting gRNA sequence is 5′-GCG AGG TAT TCG GCT CCG CG-3′. AP-7 cells were transfected with the recombinant plasmids using the Lipofectamine 2000 reagent (Thermo Fisher Scientific). At 24 h following transfection, cells were selected by treatment with 7.5 μg/ml puromycin (Thermo Fisher Scientific) for 48 h. Monoclonal cell lines were generated by limiting dilution and confirmed by gene sequencing and immunoblotting.

Virus infection and inhibitor treatment.

The SINV strain TE was previously constructed by replacing restriction fragments of a cDNA clone, Toto 1101, with E2 from a neuroadapted SINV strain (NSV) and E1 from the original SINV isolate (AR339) (77). Virus stocks were grown and assayed by plaque formation on BHK-21 cells. AP-7 and CSM14.1 cells were infected at a BHK-21 cell multiplicity of infection (MOI) of 10 in DMEM–1% FBS for 1 h and washed twice with phosphate-buffered saline (PBS), and fresh medium was replaced. MEFs were infected at a BHK-21 cell MOI of 1. For the inhibitor studies, the medium was supplemented with 10 μM 2-[(aminocarbonyl) amino]-5-(4-flurophenyl)-3-thiophenecarboxamide (TPCA-1; an IKKβ inhibitor; EMD Millipore), 1 μM s6,8-dihydro-8-(1H-imidazol-5-ylmethylene)-7H-pyrrolo[2,3-g]benzothiazol-7-one (C16; a PKR inhibitor; EMD Millipore}, or an equal volume of DMSO (vehicle control; Sigma).

Immunoblot analysis of protein expression and phosphorylation.

Cell monolayers were washed twice with cold PBS, lysed in radioimmunoprecipitation assay (RIPA) buffer (50 mM Tris-Cl [pH 8.0], 150 mM NaCl, 1% NP-40, 0.1% SDS, 0.5% Na-deoxycholate, 1 mM EDTA), containing protease inhibitor and phosphatase inhibitor cocktails (Roche), incubated on ice for 30 min, and cleared by centrifugation at 14,000 × g for 15 min. The protein concentration was determined using a DC assay kit (Bio-Rad). Ten micrograms of total protein was separated by SDS-polyacrylamide gel electrophoresis (PAGE), transferred to a nitrocellulose membrane (Bio-Rad), and blocked in Tris-buffered saline (TBS) containing 0.1% Tween 20 (TBS-T) and 5% milk. Immunodetection was performed with the following antibodies diluted in TBS-T containing 5% bovine serum albumin (BSA; Sigma): monoclonal anti-phospho-p65 (antibody S536; 1:1,000; catalog number 3033), anti-p65 (1:1,000; catalog number 4764), anti-IκBα (1:1,000; catalog number 4814), anti-phospho-eIF2α (antibody S51; 1:1,000; catalog number 9721), anti-eIF2α (1:1,000; catalog number 9722), and anti-phospho-IKKα/β (antibody S176/177; 1:500; catalog number 2078) (all from Cell Signaling Technology); anti-PKR (1:500, catalog number sc-6282; Santa Cruz Biotechnology); polyclonal anti-SINV structural proteins (1:1,000) (9); polyclonal anti-SINV nsP3 (1:1,000) (78); and monoclonal anti-β-actin (1:10,000; catalog number MAB1501; EMD Millipore). The secondary antibodies horseradish peroxidase-conjugated goat anti-rabbit IgG (1:1,000; Cell Signaling Technology) or sheep anti-mouse IgG (1:1,000; GE Healthcare) were diluted 1:1,000 in 5% milk. The membranes were developed using the Amersham ECL Prime Western blotting detection reagent (GE Healthcare). Densitometric analysis was performed using ImageJ software, and the values indicate the immunoreactivity of the protein of interest normalized to that for the loading control.

RPPA.

Cell lysates were collected in RIPA buffer, and proteins were quantified using the DC assay kit, as described above for immunoblotting. Samples were printed onto nitrocellulose slides in eight-point 2-fold dilutions using an Aushon 2470 arrayer (Aushon Biosystems) and analyzed for protein expression by use of a reverse-phase protein array (RPPA) at the George Mason University Center for Applied Proteomics and Molecular Medicine as previously described (79). The primary antibodies used were polyclonal anti-IκBα (1:1,000; catalog number 9242), anti-phospho-p65 (antibody S536; 1:1,000; catalog number 3031), anti-p65 (1:1,000; catalog number 3034), and antiactin (1:1,000; catalog number 4967) (all from Cell Signaling Technology). The secondary antibody was biotinylated anti-rabbit IgG (1:1,000; Vector Laboratories). Densitometric image analysis was performed using MicroVigene software (Vigenentech). Values indicate the fold change for triplicate matched SINV-infected to mock-infected samples for each time point.

Immunocytochemistry.

Cells were grown on glass coverslips in 24-well plates. Cells were infected and fixed at the time points indicated in the figures with 4% paraformaldehyde in PBS (pH 7.2) at room temperature for 15 min. Coverslips were blocked with PBS containing 5% normal goat serum (Thermo Fisher Scientific) plus 0.3% Triton X-100 (Sigma) and incubated with monoclonal antibody to p65 (1:500; catalog number 8242; Cell Signaling Technology) overnight at 4°C. Coverslips were washed and incubated with Alexa Fluor 594-conjugated goat anti-rabbit IgG (1:500; Invitrogen) for 1 h at room temperature. Coverslips were counterstained with DAPI (4′,6-diamidino-2-phenylindole; 10 μg/ml; Molecular Probes) for 5 min at room temperature to visualize the nuclei and mounted onto slides with the Prolong Gold reagent (Life Technologies). Images were taken at a ×40 magnification using a Zeiss Axioimager microscope and Volocity imaging software.

qRT-PCR for viral gene expression.

For analysis of RNA, cells were harvested in RLT Plus buffer (Qiagen), and total cellular RNA was isolated using an RNeasy Plus minikit (Qiagen). cDNA was synthesized with random primers using a high-capacity cDNA reverse transcription kit (Applied Biosystems). Viral RNA levels were quantified by qRT-PCR using EagleTaq Universal master mix (Roche) with primers and probes specific for the genomic (nsP2) and genomic plus subgenomic (E2) regions of the SINV genome on a 7500 Fast real-time PCR system. The primer and probe sequences used were as follows: for SINV nsP2, primer nsP2 3373F (5′-CCG CAA GTA TGG GTA CGA TCA-3′), primer nsP2 3454R (5′-GTG CCC TTC CCA GCT AGC T-3′), and TaqMan probe nsP2 3317 (5′–6-carboxyfluorescein [6-FAM]–CCA TTG CCG CCG AAC TCT CCC–6-carboxytetramethylrhodamine [6-TAMRA]–3′), and for SINV E2, primer E2 8732F (5′-TGG GAC GAA GCG GAC GAT AA-3′), primer E2-8805R (5′-CTG CTC CGC TTT GGT CGT AT-3′), and TaqMan probe E2 8760 (5′–6-FAM–CGC ATA CAG ACT TCC GCC CAG T–6-TAMRA–3′). Endogenous GAPDH (glyceraldehyde-3-phosphate dehydrogenase) expression was measured using TaqMan rodent GAPDH control reagents (Applied Biosystems). RNA copy numbers were quantified using a standard curve consisting of 10-fold serial dilutions ranging from 3 × 10⁷ to 3 copies of the pCRII-TOPO plasmid, containing the SINV nsP2 region, or pGEM-3Z, containing the SINV subgenomic region, and normalized to those for Gapdh.

Infectious center assay.

dAP-7 cells were incubated with SINV strain TE (MOI, 10) for 1 h at 4°C. The medium was replaced with medium containing either 10 μM IKKβ inhibitor TPCA-1 (EMD Millipore) or an equal volume of the DMSO vehicle. Cells were shifted to 39°C for an additional 4 h for viral entry and the establishment of infection. Infected cells were then trypsinized, counted, and plated in 10-fold dilutions onto BHK-21 cell monolayers. BHK-21 cell monolayers were then overlaid with 0.6% Bacto agar (Sigma) in minimal essential medium (Gibco) and incubated for an additional 48 h at 37°C. Cells were stained with 0.033% neutral red (Sigma) solution in PBS for 2 h at 37°C, and plaques indicating infectious centers were counted. Data are reported as the log₁₀ number of infectious centers per 10⁵ cells plated and the percentage of trypsinized cells that produced infectious centers.

Flow cytometry for double-stranded RNA.

To identify cells containing dsRNA structures, cells were trypsinized, pelleted, washed twice with PBS containing 2 mM EDTA, and stained with a Live-Dead fixable violet dead cell stain kit (Molecular Probes) for 30 min. Following staining, cells were fixed with 2% paraformaldehyde in PBS, washed twice with PBS/EDTA buffer, and permeabilized with 0.2% Triton X-100 in fluorescence-activated cell sorting buffer (PBS, 0.4% 0.5 M EDTA, 0.5% BSA). Cells were stained for dsRNA with the J2 mouse monoclonal antibody (1:1,000; Scions) for 1 h on ice. Cells were then incubated with phycoerythrin-conjugated goat anti-mouse IgG (1:400; Invitrogen) for 45 min on ice and analyzed on a BD FACSCanto flow cytometer. Data were analyzed using FlowJo software. Live cells positive for dsRNA were counted, and the results are reported as the percentage of total live cells positive for dsRNA.

Radiolabeling and pulse-chase analysis of protein synthesis and processing.

At the times after infection indicated in the figures, dAP-7 cells were washed twice with PBS and depleted of Met/Cys by incubation with methionine-free, cysteine-free DMEM (Gibco) for 30 min at 39°C. For the pulse-labeling experiments, cells were pulsed with 50 μCi/ml of ³⁵S-labeled cysteine/methionine (MP Biomedicals) for 1 h at 39°C and then lysed with RIPA buffer without SDS (50 mM Tris-Cl [pH 8.0], 150 mM NaCl, 1% NP-40, 0.5% Na-deoxycholate, 1 mM EDTA) supplemented with protease and phosphatase inhibitor cocktails (Roche). The lysates were incubated on ice for 30 min and cleared by centrifugation at 14,000 × g for 15 min. ³⁵S incorporation was measured, and the lysates were analyzed by SDS-PAGE (equal counts were loaded per lane). The radioactive signal was amplified by fluorography as previously described (80) and autoradiographed at −80°C.

For the pulse-chase experiments, cells were labeled with 100 μCi/ml of ³⁵S-labeled cysteine/methionine (MP Biomedicals) for 15 min at 39°C. To quench the labeling process, cells were washed twice with PBS, and fresh medium, containing 15 mg/liter of unlabeled l-methionine and l-cysteine (Sigma), was replaced. The cells were chased at the time intervals indicated in the figures, washed twice with ice-cold PBS, and lysed with RIPA buffer supplemented with protease and phosphatase inhibitors as described above. The cell lysate was cleared by centrifugation at 14,000 × g for 15 min at 4°C, and sample protein concentrations were determined using the DC assay kit (Bio-Rad). To evaluate the rate of SINV structural and nonstructural protein processing, 500 μg of total cell lysate was incubated with 10 μg of the capture antibody (polyclonal antibody against the SINV structural proteins [9] or polyclonal antibody against SINV nonstructural protein 3 [nsP3] [78]) overnight at 4°C. Protein A/G beads (Thermo Fisher Scientific) were added to each sample, and the mixture was incubated overnight at 4°C. The beads were then pelleted and washed 5 times with RIPA buffer, and the immunoprecipitated proteins were eluted by boiling for 5 min. The immunoprecipitated proteins were separated by SDS-PAGE, amplified by fluorography, and exposed by autoradiography at −80°C as described above. The radioactivity of the protein bands was quantified using ImageJ software.

Statistical analyses.

Data are expressed as the mean ± standard deviation (SD). Statistical analyses were performed using Prism software (v7.0; GraphPad), with a P value of <0.05 being considered significant. Differences between two groups were analyzed by an unpaired, two-tailed Student's t test. Multiple comparisons between groups were made using two-way analysis of variance with the Sidak posttest.