Pseudorabies Virus Infection Triggers NF-κB Activation via the DNA Damage Response but Actively Inhibits NF-κB-Dependent Gene Expression

ABSTRACT

The nuclear factor kappa B (NF-κB) pathway is known to integrate signaling associated with very diverse intra- and extracellular stressors, including virus infections, and triggers a powerful (proinflammatory) response through the expression of NF-κB-regulated genes. Typically, the NF-κB pathway collects and transduces threatening signals at the cell surface or in the cytoplasm leading to nuclear import of activated NF-κB transcription factors. In the current work, we demonstrate that the swine alphaherpesvirus pseudorabies virus (PRV) induces a peculiar mode of NF-κB activation known as “inside-out” NF-κB activation. We show that PRV triggers the DNA damage response (DDR) and that this DDR response drives NF-κB activation since inhibition of the nuclear ataxia telangiectasia-mutated (ATM) kinase, a chief controller of DDR, abolished PRV-induced NF-κB activation. Initiation of the DDR-NF-κB signaling axis requires viral protein synthesis but occurs before active viral genome replication. In addition, the initiation of the DDR-NF-κB signaling axis is followed by a virus-induced complete shutoff of NF-κB-dependent gene expression that depends on viral DNA replication. In summary, the results presented in this study reveal that PRV infection triggers a noncanonical DDR-NF-κB activation signaling axis and that the virus actively inhibits the (potentially antiviral) consequences of this pathway, by inhibiting NF-κB-dependent gene expression.

IMPORTANCE The NF-κB signaling pathway plays a critical role in coordination of innate immune responses that are of vital importance in the control of infections. The current report generates new insights into the interaction of the alphaherpesvirus pseudorabies virus (PRV) with the NF-κB pathway, as they reveal that (i) PRV infection leads to NF-κB activation via a peculiar “inside-out” nucleus-to-cytoplasm signal that is triggered via the DNA damage response (DDR), (ii) the DDR-NF-κB signaling axis requires expression of viral proteins but is initiated before active PRV replication, and (iii) late viral factor(s) allow PRV to actively and efficiently inhibit NF-κB-dependent (proinflammatory) gene expression. These data suggest that activation of the DDR-NF-κB during PRV infection is host driven and that its potential antiviral consequences are actively inhibited by the virus.

INTRODUCTION

Pseudorabies virus (PRV) is a porcine alphaherpesvirus and constitutes the etiological agent of the multiorgan syndrome termed Aujeszky’s disease (1). Alphaherpesvirinae represent the largest subfamily of Herpesviridae and include a.o. human herpes simplex viruses (HSV-1 and HSV-2) and varicella zoster virus (VZV) and several animal viruses, including equine herpesvirus 1 (EHV-1), bovine herpesvirus 1 (BHV-1), and feline herpesvirus 1 (FeHV-1) (2). Alphaherpesviruses display complex and fine-balanced interactions with the cells and immune system of their natural host, and PRV is often used as a model organism to study general alphaherpesvirus-host interactions (3).

The nuclear factor kappa B (NF-κB) signaling pathway is a leading regulator of innate immunity and inflammation and is activated by endangering and/or infectious stimuli, culminating in the activation of NF-κB/Rel transcription factors that trigger expression of a variety of genes, including several inflammatory genes (4). An accurate regulation of NF-κB-driven responses is imperative and is reflected by the serious outcomes caused by NF-κB-associated pathogenesis, including cancer and autoimmune and cardiovascular diseases, among several others (5–8). Pathogen recognition and cellular detection of proinflammatory cytokines and/or a wide diversity of intracellular stressors are molecular triggers of NF-κB. Triggering signals are located typically at the cell surface and/or in the cytoplasm and lead to activation of the inhibitory kappa B kinase (IKK) complex that contains the IKKβ kinase. The IKKβ kinase phosphorylates inhibitory kappa B (IκB) proteins (particularly IκBα) that are associated with NF-κB and retain it in the cytoplasm. Phosphorylation tags the IκB proteins for ubiquitin-dependent proteasomal degradation, thereby releasing NF-κB and allowing it to translocate to the nucleus and drive expression of the genes that harbor NF-κB-responsive promoters (9, 10).

Recently, our lab reported that PRV triggers persistent activation of NF-κB in infected porcine epithelial cells, although this does not appear to result in the expression of NF-κB-driven genes (11). PRV infection causes progressive IκBα protein degradation starting from 4 hours postinfection (hpi), which is accompanied by nuclear import of the p65/RelA subunit of NF-κB. The trigger for NF-κB activation during PRV infection was found to be largely independent of IKKβ kinase activity, in contrast to canonical activation of NF-κB, suggesting a noncanonical pathway involved in PRV-mediated NF-κB activation (11), although the nature of the trigger of NF-κB activation during PRV remained unclear.

In the current report, we aimed at generating new insights in the stage of PRV replication and the cellular pathway that are involved in NF-κB activation during PRV infection. We show that (i) NF-κB activation precedes viral DNA synthesis, (ii) PRV-induced NF-κB is triggered via a peculiar “inside-out” signaling initiated by double-stranded DNA damage (DDR) in an ataxia telangiectasia-mutated (ATM)-dependent way, (iii) inhibition of the DDR-NF-κB signaling axis does not suppress virus replication in cell culture, and (iv) treatment of PRV-infected cells with the DNA replication inhibitor phosphonoacetic acid does not prevent NF-κB activation but releases NF-κB transcriptional activity indicating that viral late factor(s) inhibit expression of NF-κB-dependent genes. Altogether, these data shed new light on the interaction of alphaherpesviruses with the NF-κB signaling axis and how this may modulate the host innate immune response against these pathogens.

RESULTS

NF-κB activation by PRV does not rely on generation of infectious virus and precedes viral DNA replication.

We showed earlier (i) that PRV infection triggers persistent activation of NF-κB but (ii) that PRV inhibits NF-κB-dependent gene expression (11). Hence, in the context of PRV infection and the current work, activation of NF-κB refers to permanent degradation of IκBα and nuclear import of NF-κB p65 but is not associated with productive transcription of NF-κB-induced genes. The lytic infection cycle of alphaherpesviruses can be subdivided in two stages. During the prereplicative phase, input genomes serve as the template for the expression of immediate early and early viral genes, which encode mainly proteins involved in the modulation of viral gene expression and in genome replication. Genome replication initiates the second phase of the replication cycle, including the expression of viral late genes that encode mostly structural proteins or factors involved in virus particle production and egress from infected cells. We reported before that, in the presence of the viral DNA polymerase inhibitor phosphonoacetic acid (PAA), a clear IκBα Western blot band could be observed in PRV-infected cells (11). The fact that the IκBα protein was detected in PAA-treated PRV-infected cells may indicate that PAA treatment prevents IκBα degradation and hence that the trigger for NF-κB activation is a PAA-sensitive late event in PRV replication. However, an alternative explanation for the IκBα protein in PAA-treated PRV-infected cells may be that PAA addition releases the PRV-induced inhibition of NF-κB-dependent gene expression, resulting in expression of NF-κB-driven genes, which include IκBα (12–14). To more carefully address whether late events in PRV replication are involved in NF-κB activation or rather in suppression of NF-κB-dependent gene expression, we made use of a PRV strain expressing a synthetic green fluorescent protein (GFP)-tagged VP26 capsid protein (PRV-GS443) (15), which allows us to visually correlate timing of NF-κB activation with late events in the virus life cycle like formation and nucleo-cytoplasmic shuttling of PRV capsids. First, we confirmed that PRV-GS443, like wild-type (WT) PRV (11), triggers gradual and persistent IκBα degradation between 4 and 6 hpi and that PAA treatment of PRV-GS443-infected cells resulted in a clear IκBα Western blot band (Fig. 1A). Importantly, at 6 hpi, when most of the IκBα protein pool was already degraded by PRV infection, subcellular location of the VP26-GFP capsid protein showed a nearly complete restriction to the nuclei of infected cells (Fig. 1B), suggesting that virus maturation processes, like capsid tegumentation and envelope acquisition that occur in the cytoplasm, are likely not involved in PRV-induced NF-κB activation. This finding was further supported by assessment of extracellular infectious virus titers. As shown in Fig. 1C, infectious virus titers started to increase only after 6 hpi, supporting the idea that the production of infectious virus particles is not involved in NF-κB activation by PRV. To assess whether viral genome replication may serve as a trigger for NF-κB activation, quantitative PCR (qPCR) assays were performed to quantify intracellular PRV genomes at different time points postinoculation (at 2, 4, 6, 8, and 12 hpi), using PAA as negative control for virus genome replication (Fig. 1D). The exponential replication of genomes started from 4 hpi onward, indicating that NF-κB activation might occur before or early during the replicative phase of infection. Initiation of PRV genome replication and NF-κB activation around 4 hpi also corresponded with the timing of detectable expression of viral late proteins by Western blotting (Fig. 1A).

FIG 1.

PRV-induced NF-κB activation is triggered before viral replication. (A) Western blot analysis of IκBα levels in ST cells infected with VP26-GFP-expressing PRV (PRV GS443) at 2, 4, 6, and 8 hpi (MOI, 10 PFU/cell). (B) Confocal microscopy of PRV GS443-infected ST cells at 4, 6, and 8 hpi (MOI, 10 PFU/cell) showing host DNA (blue), US3 (red), and VP26-GFP (green). Bar, 10 μm. (C) Infectious extracellular virus titers determined at 2, 6, 12, and 24 hpi in PRV-infected ST cells (PRV GS443; MOI, 10 PFU/cell). At 2 hpi, infected cell monolayers were subjected to sodium citrate buffer treatment to remove nonentered input virus. (D) Quantitative PCR (qPCR) data of intracellular PRV genomes in infected ST cells at indicated time postinoculation (PRV GS443; MOI, 10 PFU/cell) in the presence or absence of phosphonoacetic acid (PAA; 400 μg/ml). PAA was added for 30 min before inoculation and kept throughout the infection. Data shown are derived using UL27 as a target sequence; similar results were obtained using US3 as a target sequence (data not shown). Graphs represent means and standard deviations of three independent repeats. Values were put relative to PAA-treated PRV-infected cells at 2 hpi. (E) Western blot analysis of IκBα levels in mock-infected or PRV-infected ST cells (PRV GS443; MOI, 10 PFU/cell) treated or not with cycloheximide (CHX; 10 μg/ml) or PAA (400 μg/ml), added at the indicated time points. (F) Confocal microscopy of NF-κB p65 in PRV-infected ST cells at 4 hpi and at 8 hpi in the presence or absence of CHX (10 μg/ml) added at 4 hpi (PRV GS443; MOI, 10 PFU/cell). Bar, 50 μm. (G) Western blot analysis of IκBα levels detected in PRV-infected ST cells at 8 hpi either treated or not treated with PAA (400 μg/ml; from 30 min before inoculation onward) and either supplemented or not supplemented with CHX (10 μg/ml; added at 4 hpi) (PRV Becker; MOI, 10 PFU/cell). (H) Confocal microscopy of NF-κB p65 in PRV-infected ST cells at 8 hpi in the presence or absence of PAA (400 μg/ml; from 30 min before inoculation onward) and/or CHX (10 μg/ml; added at 4 hpi) (PRV GS443; MOI, 10 PFU/cell). Bar, 50 μm. Graph on the right shows means and standard deviations (of three independent repeats) of the percentage of cells with nuclear localization of NF-κB p65. Different letters above bars indicate statistically significant differences (P < 0.05). All Western blots and immunofluorescence assays shown in this figure were taken from one representative assay out of three independent repeats of each experiment.

In an attempt to further disentangle viral genome replication from viral late protein expression as putative triggers of NF-κB activation (of which both are inhibited by PAA), late protein expression was blocked by addition of the protein translation inhibitor cycloheximide (CHX) for 4 h starting at 4 hpi, which is just before the onset of viral genome replication and detectable late viral protein expression, and assessed for its effect on IκBα protein degradation. Interestingly, PRV-infected cells treated with CHX from 4 hpi still showed full IκBα degradation at 8 hpi, whereas expression of viral late proteins (e.g., gB and gD) was completely abrogated (Fig. 1E). In line with this result, addition of CHX at 4 hpi did not inhibit nuclear translocation of the p65 subunit of NF-κB (Fig. 1F). These data indicate that the expression of viral late proteins is not required for NF-κB activation in PRV-infected cells, although addition of the viral DNA replication inhibitor PAA rescues IκBα protein levels.

Fascinatingly, addition of both inhibitors, namely, PAA and CHX, resulted in almost complete degradation of IκBα protein levels in PRV-infected cells (Fig. 1G). In line with this finding, the addition of PAA alone resulted in cytoplasmic localization of NF-κB p65, whereas NF-κB p65 was mostly nuclear upon addition of both PAA and CHX (Fig. 1H). These results indicate that the IκBα protein band and the cytoplasmic localization of NF-κB p65 in PAA-treated PRV-infected cells are not caused by an impaired NF-κB activation in the presence of PAA but instead by de novo expression of IκBα (which subsequently recycles NF-κB p65 from the nucleus to the cytoplasm) due to a release of the PRV-induced inhibition of NF-κB-dependent gene expression. Combined, these data therefore show (i) that the trigger for NF-κB activation during PRV infection does not depend on viral genome replication or late viral protein production and (ii) that PRV encodes one or more PAA-sensitive inhibitors of NF-κB-dependent gene expression.

PRV triggers the DNA damage response, and PRV-induced NF-κB activation depends on the activation of the double-stranded DNA (dsDNA) damage kinase ATM.

Several steps of the virus replication cycle take place during the initial 4-h postinoculation step when NF-κB activation is triggered, including virus attachment, virus entry, genome delivery to the host cell nucleus, the expression of viral (immediate) early genes, and preparation of the viral genome for active replication (3).

Since we showed earlier that UV-inactivated PRV does not trigger NF-κB activation (11), NF-κB activation does not appear to rely on the very early stages of virus infection, from virus attachment up to viral genome delivery in the nucleus. Taking this into consideration, we posit that the cell nucleus represents a conceivable source of NF-κB activation possibly through stress associated with PRV infection. Although nuclear activation of NF-κB, so-called “inside-out” NF-κB activation from the nucleus to the cytoplasm, does not represent the most typical NF-κB signaling cascade, induction of NF-κB can be a result of the DNA damage response (DDR) in the cell. Such activation relies on initiation of the DDR-associated ATM kinase in response to dsDNA breaks or oxidative stress (16, 17). To test the possible involvement of this pathway in PRV-induced NF-κB activation, we first tested whether PRV infection triggers the dsDNA DDR during the initial 4 h of infection through Western blot analysis of histone H2A.X phosphorylation at residue Ser139 (γ-H2A.X), a hallmark of DDR activation (18, 19). Our data indicate that PRV infection indeed triggers the DDR between 2 and 4 hpi, corresponding to the same time interval at which the NF-κB pathway is activated (Fig. 2A). Importantly, CHX treatment of PRV-infected cells from the onset of infection or the use of UV-inactivated PRV did not trigger detectable H2A.X phosphorylation (Fig. 2A and 2B), indicating that virus entry and incoming linear viral genomes, exhibiting exposed dsDNA ends in the nucleus, do not trigger a detectable DDR response. As shown in Fig. 2A, phosphorylation of H2A.X increases substantially from 4 to 8 hpi, suggesting that increasing virus DNA replication and/or the consequences of viral replication on host chromatin may further boost DDR activation. In line with this idea, phosphorylated H2A.X levels were substantially reduced at 8 hpi when PRV-infected cells were treated with the viral DNA replication inhibitor PAA (Fig. 2C). Despite the reduction, H2A.X phosphorylation was still noticeable in the presence of PAA, indicating that PRV-induced DDR activation is initiated before active viral DNA replication, corresponding with the timing of NF-κB activation. Addition of CHX at 4 hpi, thereby blocking viral late protein expression but only partially suppressing viral genome replication (Fig. 2D), did not substantially reduce H2A.X phosphorylation at 8 hpi (Fig. 2E), indicating that viral DNA replication itself, rather than viral late protein production, leads to the increased H2A.X phosphorylation between 4 hpi and 8 hpi. Immunofluorescence assays showed that phosphorylated H2A.X colocalizes with host chromatin during late stages of infection (Fig. 2F). Phosphorylated H2A.X is involved in the maintenance of DNA repair factors at the damaged DNA foci (19). The fact that γH2A.X associates mainly with marginalized chromatin rather than with expanded viral replication compartments (VRC) (Fig. 2F) indicates that the DDR occurs predominantly at the host chromatin or that γH2A.X molecules are redirected from VRC to host chromatin.

FIG 2.

PRV infection triggers the DNA damage response. (A) Western blot analysis of IκBα, H2A.X, and γH2A.X (phospho-H2A.X at Ser139) in PRV-infected ST cells (PRV Becker; MOI, 10 PFU/cell) either treated or not treated with CHX (10 μg/ml). (B) Western blot analysis of H2A.X, γH2A.X, and IκBα in ST cells inoculated with PRV (PRV Becker; MOI, 10 PFU/cell) or a corresponding amount of UV-inactivated PRV and analyzed at 8 hpi. (C) Western blot analysis of H2A.X and γH2A.X in PRV-infected ST cells (PRV Becker; MOI, 10 PFU/cell, 8 hpi) either treated or not treated with PAA (400 μg/ml; treatment from 30 min before inoculation onward). (D) qPCR quantification of intracellular PRV genome production in PRV-infected ST cells (PRV GS443; MOI, 10 PFU/cell) at 4 hpi and at 8 hpi either in the presence or not of PAA (400 μg/ml; added either at 4 hpi or 30 min before infection as indicated) or CHX (10 μg/ml; added at 4 hpi). Data shown are derived using UL27 as a target sequence; similar results were obtained using US3 as a target sequence (data not shown). Graphs represents mean and standard deviations of three independent repeats. Values were put relative to PAA-treated PRV-infected cells at 2 hpi. (E) Western blot analysis of H2A.X and γH2A.X in PRV-infected ST cells (PRV Becker; MOI, 10 PFU/cell) at 4 hpi and at 8 hpi either treated or not treated with CHX (10 μg/ml) from 4 hpi onward. (F) Confocal images of subnuclear distribution of γH2A.X (red), VP26-GFP (green), and host DNA (blue) in PRV GS443-infected ST cells at 8 hpi (MOI, 10 PFU/cell). (G) Western blot analysis of IκBα and total and phosphorylated H2A.X (γH2A.X; phospho-H2A.X at Ser139) at different time points posttreatment of ST cells with doxorubicin (5 μM). All Western blots and immunofluorescence assays shown in this figure were taken from one representative assay out of three independent repeats of each experiment.

PRV-induced NF-κB activation is also peculiar since, in comparison to canonical NF-κB activation, IκBα degradation and p65 nuclear translocation occur relatively slowly (hours rather than minutes) and persist for a long time (several hours versus minutes to a few hours) (Fig. 1A) (11). Interestingly, it has been reported that doxorubicin, an agent that activates the DDR by generating dsDNA breaks, also triggers relatively slow and persistent NF-κB activation (20). In line with this concept, we observed that doxorubicin triggered potent H2A.X phosphorylation which correlated with a progressive and irreversible IκBα degradation (Fig. 2G), reminiscent of what we observed for PRV infection and further in line with a potential role for the dsDNA DDR in PRV-induced NF-κB activation.

To investigate whether the DDR is responsible for NF-κB activation during PRV infection, we used the selective ATM kinase inhibitor KU60019, a more potent and selective derivative of the widely used ATM inhibitor KU55933 (21, 22). KU60019 efficiently prevented H2A.X phosphorylation in PRV-infected cells (and doxorubicin-treated cells), indicating that the ATM kinase is the main driver of the PRV-induced DDR (Fig. 3A). Importantly, we found that different doses of the ATM inhibitor (1, 3, and 10 μM) were capable of efficiently preventing IκBα degradation in PRV-infected cells, indicating that the ATM kinase is a central regulator of NF-κB activation triggered by PRV infection (Fig. 3B). Of note and as a control, the ATM inhibitor exclusively repressed IκBα degradation in response to PRV infection or doxorubicin treatment but not upon canonical NF-κB activation via addition of tumor necrosis factor alpha (TNF-α) (Fig. 3C and 3D). As an additional control, the ATM inhibitor also prevented IκBα degradation in PRV-infected cells that were treated with the protein translation inhibitor CHX from 4 hpi to 8 hpi (Fig. 3E), confirming that the IκBα protein band detected in KU60019-treated PRV-infected cells corresponds to nondegraded preexisting IκBα rather than replenished IκBα. The addition of KU60019 to cells infected with two additional PRV wild-type strains (PRV Kaplan and PRV NIA3) showed similar phenotypes, i.e., full inhibition of H2A.X phosphorylation and substantial inhibition of IκBα degradation (Fig. 3F), indicating that the results are not PRV strain dependent.

FIG 3.

Inhibition of the DNA damage response inhibits PRV-induced NF-κB activation. (A) Western blot analysis of H2A.X and γH2A.X (phospho-H2A.X at Ser139) in ST cells at 3 h after doxorubicin treatment (5 μM) or 8 h after PRV inoculation (PRV Becker; MOI, 10 PFU/cell) in the absence or presence of the ATM inhibitor KU60019 (10 μM; added 30 min before doxorubicin treatment or at 2 hpi in PRV-infected cells). (B) Western blot analysis of IκBα in PRV-infected ST cells (PRV Becker; MOI, 10 PFU/cell) at 8 hpi in the absence or presence of different concentrations (1, 3, and 10 μM) of KU60019 (added at 2 hpi). (C) Western blot analysis of IκBα in doxorubicin-treated ST cells (5 μM) with or without KU60019 (10 μM; KU60019 was added from 30 min before doxorubicin addition onward). (D) Western blot analysis of IκBα levels in ST cells treated with recombinant porcine TNF-α for 30 min (100 ng/ml) or at 8 hpi with PRV (PRV Becker; MOI, 10 PFU/cell) in the presence or absence of the ATM inhibitor KU60019 (10 μM). KU60019 was added at 2 hpi in PRV-infected cells and at 5 h 30 min before TNF-α treatment to ensure that cells were exposed to KU60019 for 6 h. (E) Western blot analysis of IκBα in PRV-infected ST cells at 8 hpi (PRV Becker; MOI, 10 PFU/cell), either treated or not treated with KU60019 alone (10 μM) or subjected to a combined treatment with KU60019 and CHX (10 μg/ml). KU60019 was added at 2 hpi and CHX at 4 hpi. (F) Western blot analysis of IκBα, H2A.X, and γH2A.X in PRV-infected ST cells at 8 hpi using three different wild-type strains of PRV, namely, Becker, Kaplan, and NIA3 (MOI, 10 PFU/cell), in the presence or absence of KU60019 (10 μM; added from 30 min prior to inoculation onward). (G) Western blot analysis of IκBα, H2A.X, and γH2A.X at different time points postinoculation of ST cells with PRV (PRV Kaplan; MOI, 10 PFU/cell), either in the presence or not of the ATM inhibitor KU60019 (10 μM). KU60019 was added from 30 min before inoculation onward. (H) Fluorescence microscopy of NF-κB p65 in PRV-infected ST cells at 8 hpi using three different wild-type PRV strains, namely, Becker, Kaplan, and NIA3 (MOI 10 PFU/cell) or treated with TNF-α (100 ng/ml) for 30 min or doxorubicin (5 μM) for 6 h, in the presence or absence of the ATM inhibitor KU60019 (10 μM). KU60019 was added 30 min before PRV inoculation or doxorubicin treatment and 8 h before addition of TNF-α, ensuring the same exposure time to KU60019 under all conditions. Bar, 50 μm. All Western blots and immunofluorescence assays shown in this figure were taken from one representative assay out of three independent repeats of each experiment.

A time course analysis showed that the ATM inhibitor very effectively represses IκBα degradation at 6 hpi, during the starting time interval of IκBα degradation from 4 to 6 hpi (Fig. 3G). At later time points of infection and particularly somewhere from 8 to 12 hpi, the ATM inhibitor KU60019 is unable to completely block the massive DDR caused by progressive PRV infection, illustrated by incomplete inhibition of H2A.X phosphorylation and IκBα degradation (Fig. 3G). In agreement with the IκBα degradation data, migration of NF-κB p65 to infected cell nuclei was considerably reduced with KU60019, although again not abolished completely (Fig. 3H). As controls, KU60019 inhibited nuclear transport of NF-κB p65 upon doxorubicin but not upon TNF-α treatment (Fig. 3H).

Combined, these results indicate that the ATM kinase is the main, if not the only, driver of NF-κB activation during PRV infection. In conclusion, PRV infection triggers the DNA damage response which in turn leads to NF-κB activation.

Potential causes of PRV-induced DDR and NF-κB activation.

The fact that PRV induces NF-κB activation through a pathway involved in sensing and repair of dsDNA damage led us to first investigate if the consequences of virus replication on the integrity of the cell nucleus may contribute to the DDR-dependent NF-κB activation. To this end, we analyzed nucleus enlargement by measuring nuclear areas at the nuclei medial section and visually assessed chromatin marginalization (the displacement of the host genome to the borders of the nucleus) (Fig. 4A). Our data show that a clear nuclear size increase and chromatin marginalization can be observed at 8 hpi. However, at 4 hpi, when the NF-κB pathway has already been triggered, no significant increase in nuclear size and only a modest chromatin marginalization could be observed (Fig. 4A), with the latter probably representing the very primary VRC. In addition, although PAA treatment does not prevent NF-κB activation, it virtually abolishes chromatin marginalization and nuclear size increase (Fig. 4A). Hence, these data indicate that PRV-induced NF-κB activation does not correlate with nuclear expansion and deformation.

FIG 4.

Inhibition of the viral helicase/primase activity or lack of the viral protein kinases US3 and UL13 does not affect PRV-induced NF-κB activation. (A) Confocal microscopy of the medial section of fluorescently stained cell nuclei in mock-infected ST cells or PRV-infected ST cells at 4 hpi and 8 hpi (left) or at 8 hpi in the absence or presence of PAA (400 μg/ml; added from 30 min before inoculation onward) (right) (PRV GS443; MOI, 10 PFU/cell). Graphs show the area of cell nuclei, calculated using ImageJ software, and represent the means and standard deviations of three independent repeats, set relative compared to mock-infected cells. (B) qPCR quantification of intracellular PRV genome production in PRV-infected ST cells (PRV Becker; MOI, 10 PFU/cell) at 2 hpi and at 8 hpi either in the presence or not of PAA (400 μg/ml) or the viral helicase/primase inhibitor pritelivir (10 μM). PAA and pritelivir were added from 30 min prior to inoculation onward. Data shown are derived using UL27 as a target sequence; similar results were obtained using US3 as a target sequence (data not shown). Graphs represent mean and standard deviations of three independent repeats. Values were put relative to the corresponding condition at 2 hpi (PRV 8 hpi versus PRV 2 hpi, PRV+PAA 8 hpi versus PRV+PAA 2 hpi, and PRV+pritelivir 8 hpi versus PRV+pritelivir 2 hpi). (C) Western blot analysis of IκBα at 8 hpi in PRV-infected ST cells (PRV Becker; MOI, 10 PFU/cell) that were either treated or not treated with PAA (400 μg/ml) or pritelivir (10 μM). PAA and pritelivir were added from 30 min prior to inoculation onward. (D) Western blot analysis of the early viral protein US3 and the late viral protein gB at 8 hpi in PRV-infected ST cells (PRV NIA3; MOI, 10 PFU/ml), in the absence or presence of PAA (400 μg/ml). PAA was added from 30 min before inoculation onward. (E) Western blot analysis of IκBα and total and phosphorylated H2A.X (γH2A.X, phospho-H2A.X at Ser139) at 8 hpi in ST cells infected with wild-type PRV NIA3, ΔUS3 PRV NIA3, or ΔUL13 PRV NIA3 (M137) (MOI, 20 PFU/cell). All Western blots and immunofluorescence assays shown in this figure were taken from one representative assay out of three independent repeats of each experiment.

Our data indicate that PRV-induced NF-κB activation precedes viral genome replication and depends on viral protein expression, which suggests that NF-κB is triggered in a time window that encompasses circularization of the PRV genome, the production of (immediate) early proteins, and the preparation of the viral genome for replication. With regard to the preparation of the viral genome for replication, the viral helicase/primase complex (UL5/UL8/UL52), which unwinds and primes viral DNA, is essential for the ensuing DNA polymerization. To assess its potential involvement in PRV-induced DDR and NF-κB activation, we made use of a potent inhibitor of the helicase/primase complex of alphaherpesviruses, pritelivir (BAY-57-1293) (23–25). As expected, pritelivir strongly suppressed PRV genome replication (Fig. 4B). As described above for PAA, to disentangle the potential effect of pritelivir on NF-κB activation from its effect on expression of the viral late factor(s) that inhibits NF-κB-dependent gene expression, CHX was either added or not to pritelivir-treated cells at 4 hpi. Based on Fig. 4C, it can be concluded that, just like PAA, pritelivir does not inhibit NF-κB activation but does prevent expression of the late viral factor(s) that inhibit NF-κB-driven gene expression. This conclusion indicates that the source of NF-κB activation precedes the activity of the viral helicase/primase. Unfortunately, we were not able to assess earlier events associated with virus genome replication intermediates, in particular a possible involvement of the nuclease activity of UL12, the single-stranded DNA (ssDNA) binding capacity of ICP8, or the role of UL9 binding to the origin of replication (Ori) sites (23, 26), due to the lack of (reliable) tools in PRV with which to conduct these assays.

Alternatively, we investigated the putative role of the two viral serine/threonine protein kinases US3 and UL13, as they are expressed with early kinetics, do not have a direct role in genome replication but are expressed within the infected cell nucleus at early time points, and have catalytic activity that could potentially drive the DDR. In addition, the UL13 homologs of gammaherpesviruses Epstein Barr virus (BGLF4) and gammaherpesvirus-68 (ORF36) have been reported to trigger the DDR (27, 28), and the UL13 homolog of different herpesviruses, including HSV-1, has been shown to activate Tip60, an acetyltransferase involved in ATM activation (28). Both US3 and UL13 are indeed expressed in a PAA-independent way (Fig. 4D). However, infections with PRV mutants lacking US3 or UL13 did not show differences in H2A.X phosphorylation or IκBα degradation at 8 hpi (Fig. 4E). In addition, in the presence of PAA, US3null and UL13null PRV triggered reduced (but still noticeable) H2A.X phosphorylation as observed for WT PRV (data not shown). Hence, both viral kinases can be excluded as triggers of the PRV-induced DDR-NF-κB signaling axis.

Overall, our data show that DDR activation and consequent NF-κB activation are triggered at a prereplicative stage of PRV infection, before the activity of the viral helicase/primase, but require expression of (immediate) early viral proteins different from US3 and UL13.

The ATM kinase is not required for PRV replication in cell culture.

We reported earlier that preventing NF-κB activation during PRV infection does not affect virus protein expression and production of extracellular infectious virus in cell culture (11). Here, we identified the DDR as an upstream event of NF-κB activation during PRV infection. However, in addition to NF-κB activation, the DDR can have different downstream effects that may affect virus replication efficiency. Therefore, we investigated the effect of ATM kinase inhibition on viral genome replication, viral protein production, and infectious virus production. Quantitative PCR analysis of intracellular viral DNA, flow cytometry assessment of late viral protein expression (gB, gD, and VP26-GFP), and virus titer determination revealed that inhibition of ATM activity did not affect viral genome replication, viral protein expression, or infectious virus production (Fig. 5).

FIG 5.

PRV-induced DNA damage response does not affect virus replication in cell culture. (A) qPCR quantification of intracellular PRV genome production in PRV-infected ST cells (PRV Becker; MOI, 10 PFU/cell) at 8 hpi either treated or not treated with different concentrations (1, 3, and 10 μM) of the ATM inhibitor KU60019 and/or PAA (400 μg/ml). PAA was added from 30 min before inoculation onward, whereas KU60019 was added from 2 hpi onward. Data shown are derived using UL27 as a target sequence; similar results were obtained using US3 as a target sequence (data not shown). Graphs represents mean and standard deviations of three independent repeats. Values were relative to PRV-infected cells treated with PAA and harvested at 2 hpi. (B) Graphs showing flow cytometric median fluorescence intensities (MFI) of viral gB, gD, and VP26-GFP protein expression at 8 hpi in PRV-infected ST cells (PRV GS443; MOI, 10 PFU/cell) either treated or not treated with the ATM inhibitor KU60019 (10 μM; added from 2 hpi onward). Graphs represent the means and standard deviations of three independent repeats. (C) Flow cytometry contour plots (y axis shows VP26-GFP fluorescent intensity and x axes either gB or gD fluorescent intensities) corresponding to one of the experiments of panel B. (D) Infectious extracellular virus titers of samples collected at 24 hpi from PRV-infected cells (PRV Becker; MOI, 10 PFU/cell) in the absence or presence of the ATM inhibitor KU60019 (10 μM; added from 2 hpi onward). Infected cell monolayers were treated with sodium citrate buffer at 2 hpi to ensure the complete removal of virus inoculum.

Altogether, these results indicate that the DDR-NF-κB signaling axis does not contribute to PRV replication in cell culture. This discovery suggests that triggering the ATM-dependent DDR and the ensuing NF-κB activation by PRV may constitute a host response to the early phases of virus replication rather than a process that is deliberately triggered by the virus for its own benefit.

The DDR-NF-κB signaling axis results in NF-κB-dependent gene expression, which is inhibited in PRV-infected cells.

We showed earlier that PRV-induced NF-κB activation does not result in the expression of NF-κB-dependent genes (11). Since our current data show that NF-κB activation during PRV infection occurs via a noncanonical DDR-dependent mechanism, we wondered whether this lack of NF-κB-dependent gene expression is a natural outcome of the DDR-NF-κB signaling axis (in swine testicle [ST] cells) or whether it indicates that PRV actively inhibits the (proinflammatory) gene expression outcome of this pathway. To address this question, we again made use of the DDR inducer doxorubicin, which triggers activation of the DDR-NF-κB signaling axis in a similar manner as PRV. Like PRV, doxorubicin triggers persistent IκBα degradation until at least 24 h posttreatment (Fig. 2G) and persistent nuclear translocation of NF-κB p65 (Fig. 6A) in ST cells, strongly resembling PRV-induced activation of NF-κB. We used real-time quantitative PCR (RT-qPCR) to assess the expression of different NF-κB-driven genes (IκBα, A20, TNF-α, and IL-6 genes) in doxorubicin-stimulated cells and PRV-infected cells. For PRV-infected cells, we treated the cells or not with the DNA polymerase inhibitor PAA, as our earlier data suggested that PRV may encode PAA-sensitive inhibitor(s) of NF-κB-dependent gene expression. TNF-α-stimulated cells were taken as a positive control for the induction of NF-κB-dependent gene expression. Notably, all four genes included in the analysis showed a transcriptional upregulation in response to doxorubicin versus the mock-treated condition but not in response to PRV infection (Fig. 6B). Importantly, addition of PAA rescued NF-κB-dependent gene expression in PRV-infected cells to levels largely comparable to those observed in doxorubicin-treated cells (Fig. 6B). These data indicate that the DDR-NF-κB signaling axis does lead to productive NF-κB-dependent gene expression but that it is actively inhibited by one or more PAA-sensitive viral factors in PRV-infected cells.

FIG 6.

Inhibition of DNA polymerization releases the PRV-induced inhibition of NF-κB-dependent gene expression. (A) Confocal microscopy of NF-κB p65 in ST cells treated with doxorubicin (5 μM) for 6, 12, and 24 h. Bar, 50 μm. (B) RT-qPCR-based mRNA expression levels of IκBα, A20, TNF-α, and IL-6 in ST cells treated or not with TNF-α (100 ng/ml; 30 min) or doxorubicin (5 μM; 6 h) or infected with PRV (8 hpi; PRV Kaplan; MOI, 10 PFU/cell) in the absence or presence of PAA (400 μg/ml; added from 30 min before inoculation onward). NF-κB-dependent gene expression was normalized to 28 rRNA mRNA levels. Uninfected PAA-treated cells were included to assess that PAA treatment has no intrinsic effect on NF-κB-dependent gene expression. Graphs show the means and standard deviations of fold differences in mRNA copies found for each of the analyzed transcripts compared with levels determined in mock-infected cells in three independent repeats of the experiment. (C) Western blot analysis of IκBα in ST cells treated with doxorubicin (5 μM) or infected with PRV (PRV Kaplan; MOI, 10 PFU/cell) that were either treated or not treated with the 26S proteasome inhibitor MG132 (10 μM) in two 4-h time windows, from 2 to 6 h (after PRV inoculation or after doxorubicin treatment) or from 8 to 12 h (after PRV inoculation or after doxorubicin treatment). (D) Western blot analysis of IκBα at 16 hpi in ST cells infected with PRV Kaplan-derived UL41null PRV (PRV ΔUL41) or a UL41 rescuant PRV (PRV ΔUL41R) (MOI, 10 PFU/cell) that were treated or not with PAA (400 μg/ml, added from 30 min before to inoculation onwards). (E) Western blot analysis of IκBα at 8 hpi in PRV-infected ST cells (PRV GS443; MOI, 10 PFU/cell) that were treated or not with PAA (400 μg/ml) and/or CHX (10 μg/ml). PAA was added from 30 min before inoculation onward, whereas CHX was added at 4, 5, or 6 hpi to PAA-treated PRV-infected cells. All Western blots and immunofluorescence assays shown in this figure were taken from one representative assay out of three independent repeats of each experiment.

These data also suggest that in doxorubicin-treated cells, the persistent lack of IκBα signal as detected via Western blot (and consequent NF-κB p65 nuclear translocation as detected by confocal microscopy) is a result of continuous degradation of replenished IκBα, whereas in PRV-infected cells, there is no need for continuous degradation of IκBα due to the inhibition of NF-κB-dependent gene expression. To confirm this idea, we performed additional assays in the absence or presence of the 26S proteasome inhibitor MG132. Western blot results showed that proteasome inhibition allowed the recovery of a detectable IκBα protein in doxorubicin-treated cells even after complete IκBα degradation (from 8 to 12 h), whereas in PRV-infected cells, once IκBα reached undetectable levels, it remained undetectable when MG132 was added (Fig. 6C). This finding confirms that in doxorubicin-treated cells, IκBα is expressed and IκBα transcripts undergo successful translation into protein but are instantly degraded, most likely due to incessant ATM activation. In PRV-infected cells, however, PAA-sensitive inhibitor(s) of NF-κB-dependent gene expression prevent replenishment of IκBα. As an alternative explanation, it is possible that NF-κB-dependent gene expression does occur but that the virion host shutoff (vhs) protein of PRV, encoded by the UL41 gene, degrades the resulting transcripts, thereby preventing IκBα replenishment. Using a UL41null PRV strain (PRV ΔUL41) and a UL41rescue strain (PRV ΔUL41R), we found that both viruses trigger IκBα degradation and neither virus leads to IκBα replenishment, even at 16 hpi (Fig. 6D), indicating that vhs/UL41 is not involved in the lack of detectable NF-κB-dependent transcripts in PRV-infected cells.

To narrow down the time window during which the viral inhibitor of NF-κB-dependent gene expression is produced, CHX was added at different time points postinoculation of PAA-treated PRV-infected cells. Figure 6E shows that a considerable amount of replenished IκBα protein could be detected at 8 hpi only when CHX was added from 6 hpi, but not earlier (at 4 or 5 hpi). Since IκBα protein degradation and subsequent p65 nuclear translocation occur between 4 and 6 hpi in PRV-infected cells (Fig. 1A) (11), these data indicate that the viral NF-κB repressor acts immediately after nuclear import of NF-κB, thereby enabling the complete shutoff of NF-κB-dependent gene expression observed during PRV infection (Fig. 6B).

Combined, these data show that the DDR-NF-κB signaling axis leads to NF-κB-dependent gene expression but that this gene expression is actively inhibited by PRV late factor(s).

DISCUSSION

In the current study, we report that pseudorabies virus (PRV) infection of host cells triggers the DNA damage response (DDR) and that this in turn leads to persistent activation of the NF-κB pathway. This DDR-NF-κB signaling axis is initiated at a prereplication stage and depends on the expression of (immediate) early viral genes. Furthermore, PRV very efficiently inhibits NF-κB-dependent gene expression via one or more viral late factor(s) that are expressed relatively early in infection (between 4 hpi and 6 hpi). A schematic representation of the interaction between PRV and the DDR-NF-κB pathway is shown in Fig. 7. Since neither inhibition of the DDR (current report) nor knockdown of NF-κB p65 (11) affects PRV replication in cell culture, these data suggest that activation of the DDR-NF-κB pathway is a host-driven process that is not beneficial for the virus and that PRV encodes one or more viral factor(s) to efficiently impair the (proinflammatory) consequences of this signaling axis.

FIG 7.

Schematic representation of the interaction between PRV and the DDR-NF-κB pathway.

This is the first report unequivocally connecting the DDR with NF-κB activation in the context of virus infection. Nevertheless, indications in other reports suggest that different viruses may trigger a similar DDR-NF-κB signaling axis. For example, the betaherpesvirus human cytomegalovirus (HCMV) controls interleukin-8 (IL-8) expression through the viral UL76 protein involving ATM and NF-κB activation (29). Another DNA virus, human polyomavirus 2 or John Cunninghamvirus (JCV) promotes nuclear translocation of IKKγ/NEMO for recruitment of activated ATM kinase back to the cytoplasm (30). In addition, paramyxoviruses have been shown to promote an ATM-dependent phosphorylation of the NF-κB subunit p65 at Ser276 which enhances IRF7/RIG-I-dependent IFN production (31). This so-called inside-out ATM-dependent NF-κB activation has been associated with the shuttling of ATM to the cytoplasm through IKKγ/NEMO, although alternative IKK-independent pathways have been described (32–34). The finding that PRV potently triggers this ATM-dependent NF-κB pathway may therefore represent an interesting model to explore this still poorly understood mode of NF-κB activation further mechanistically.

This report also reveals ATM-dependent DDR activation during PRV infection, which is something that was reported previously for the alphaherpesviruses HSV-1 and HSV-2 (35, 36). Interestingly, other DDR pathways controlled by ATR and DNA-dependent protein kinase, catalytic subunit (DNA-PKcs) kinases seem to be inhibited by the HSV-1 ICP0 protein. More concretely, ICP0 is involved in the mislocalization of the ATR interacting protein (ATRIP) and in DNA-PKcs degradation (37, 38), suggesting that in agreement with the current observations in PRV infection, ATM is the main driver of the DDR to alphaherpesvirus infections. It has been proposed that the DDR to HSV enables the recruitment of host cell DNA repair factors to virus replication factories to enhance virus replication (35, 39, 40). In contrast, some DNA repair proteins activated by the DDR, such as RNF8 and RNF168, are recruited very early in infection to incoming genomes to limit virus replication, a host antiviral mechanism counteracted by the ICP0 protein (41, 42). Up to date, there is no clear consensus about the possible implications of the DDR on HSV-1 replication fitness and different hypotheses have been proposed (35, 36, 39, 40, 43–46). Our current data on PRV infection indicate that inactivation of ATM kinase does not reduce the viral replication capacity (e.g., viral genome production, late protein expression, and extracellular virus titers), at least in cell culture.

Although the exact mechanism of DDR-NF-κB activation during PRV infection is unclear, our data indicate that the trigger takes place before the activities of the helicase/primase complex and DNA polymerase. These observations fit with what was previously described in HSV-1 infection where the MRN complex that plays a crucial role in ATM-dependent DDR activation upon sensing dsDNA breaks interacts with the HSV-1 UL12 exonuclease (26). UL12 and ICP8 are proposed as a two-subunit recombinase that participates in the primary initiation of alphaherpesvirus genome replication before the activity of the helicase/primase. Future research will reveal whether the activity of PRV UL12 and ICP8 are involved in PRV-induced DDR-NF-κB activation.

DDR-dependent NF-κB activation by PRV infection takes place very early in infection before viral genome replication, which is a phenotype consistent with what has been reported for HSV-1 infection (45, 46). Amici et al. proposed a biphasic induction of the NF-κB pathway consisting of a first wave caused immediately after infection, which probably does not involve viral gene expression, and a second and more powerful wave of activation that requires active virus replication in infected cells (47). Structural virion components have been reported to play a role in NF-κB activation immediately after infection, such as the viral glycoproteins gB, gD, gH, and gL (48, 49) and the tegument protein UL37 (50). In PRV infection, using IκBα degradation as a readout, we identified only one wave of activation that requires viral protein expression, although an earlier wave associated with PRV entry that may not lead to substantial IκBα degradation cannot be formally excluded. Expression of the immediately early proteins ICP4 and ICP27 has been reported to play a crucial role in HSV-1-dependent NF-κB activation (51). Particularly, the N-terminal region of HSV-1 ICP27 has been reported to trigger NF-κB activation via the Jun N-terminal protein kinase (JNK) (52). In PRV infection, viral protein expression until at least 2 hpi is not sufficient to trigger NF-κB, which suggests that the expression of the sole immediate early PRV protein IE180 (a homolog of HSV-1 ICP4) does not cause NF-κB activation. Unfortunately, we could not draw meaningful conclusions from infection of cells with a PRV mutant lacking UL54 expression (the homolog of HSV-1 ICP27) given that infection was massively delayed compared with PRV WT infection (data not shown). The defective replication of PRV ΔUL54 is supported by the complete dysregulation of the viral gene transcription program observed for this mutant (53). In general, removal of viral regulators of viral gene expression will cause an impaired expression of other early proteins (e.g., those involved in viral genome replication) complicating interpretation of the data. Alternatively, Taddeo et al., showed that HSV-1 activates the NF-κB pathway through the protein kinase R (PKR) (54). In conclusion, the time window of NF-κB activation for HSV-1 and PRV seems to correlate, but this report for the first time connects the activation of the NF-κB pathway with triggering of the ATM-dependent DDR during alphaherpesvirus infection.

PRV-induced NF-κB activation is aberrant, consisting of a persistent and irreversible IκBα degradation and nuclear localization of NF-κB (11), which is similar to reports for HSV-1 (47, 51). This persistent NF-κB activation caused by both viruses relies on the failure of autoregulatory mechanisms involved in the repression of the pathway upon activation (e.g., IκBα replenishment) and is achieved by inhibition of the transcriptional activity of NF-κB (current report, 11, 51). Curiously, the closely related alphaherpesvirus varicella zoster virus (VZV) inhibits the NF-κB pathway via a different strategy, based on the persistent stabilization of the IκBα protein in the cytoplasm of infected cells, accompanied by a retention of NF-κB in the cytoplasm after a transient nuclear translocation (55–57). The results shown in this report indicate that NF-κB-dependent gene expression can be released by the addition of PAA, indicating that a late factor(s) encoded by PRV is actively repressing NF-κB immediately upon its nuclear translocation. The HSV-1-dependent inhibition of NF-κB transcription activity might be distinct from that operating in PRV infection, given that PAA treatment does not cause IκBα replenishment during HSV-1 infection (52). It has been proposed that HSV-1-dependent shutoff of NF-κB-dependent gene expression is achieved through the active relocalization of NF-κB molecules from κB binding sites on the host genome to viral gene promoters (47, 58). Nevertheless, some cellular and viral events in HSV-1 infection, such as apoptosis and virus egress, have still been associated with NF-κB transcriptional activity in infected cells (59–61). In any way, independently of the mechanism(s) used, alphaherpesviruses effectively evade NF-κB-dependent gene expression in infected cells, suggesting that productive NF-κB pathway activation may represent a strong barrier toward efficient virus replication and spread, especially in vivo.

The fact that persistent ATM-dependent DDR, e.g., as caused by doxorubicin, drives continuous IκBα degradation, including both the pre-existing and the replenished pool of IκBα, leads us to hypothesize that the ATM-dependent NF-κB pathway is an unstoppable signaling stream that effectively overcomes the negative regulatory mechanisms that are meant to repress the NF-κB pathway upon initial activation. This idea is unusual, as these self-regulated inhibitory strategies of NF-κB are functional when NF-κB is activated via the canonical pathway (11, 12). The irreversible and powerful DDR activation elicited throughout the PRV lytic infection likely triggers a similar incessant ATM-dependent NF-κB activation. This activation might explain why PRV may have evolved a mechanism to generally shut down NF-κB-dependent gene expression, since alternative strategies to interfere with NF-κB activation, such as the overexpression/activation of physiological repressors of the pathway, likely would be unable to prevent ATM-dependent persistent NF-κB activation.

In conclusion, the data provided in the current study show that PRV-induced NF-κB activation depends on a peculiar ATM-dependent DDR pathway involved in the maintenance of cellular genome integrity and regulation of cell survival upon DNA injury. Our data suggest that DDR activation by PRV may be part of the cellular response to infection and does not provide an evident benefit for the virus, at least in cell culture, which may explain our observation that PRV potently suppresses subsequent NF-κB-dependent gene expression. Altogether, the interaction of PRV with the DDR-NF-κB signaling axis indicates a viral subversion of the very potent NF-κB signaling axis in the early control of alphaherpesvirus infections by the host.

MATERIALS AND METHODS

Cell cultures and viruses.

Swine testicle (ST) cells (ATCC CRL-1746) were cultured in modified Eagle’s medium (MEM) supplemented with 10% inactivated fetal bovine serum (FBS), 100 U/ml penicillin, 0.1 mg/ml streptomycin, 50 μg/ml gentamicin, and 1 mM sodium pyruvate (all from Gibco, Thermo Fisher Scientific) (ST medium).

The wild-type (WT) PRV strains used in the present study were PRV WT Becker (62), PRV WT Kaplan (63), and PRV WT NIA3 (64). The PRV-GS443 strain, a kind gift from G. A. Smith (Northwestern University, USA) was generated from the PRV WT Becker backbone by fusing the green fluorescent protein (GFP) open reading frame to the PRV UL35 gene, resulting in the expression of GFP-tagged PRV capsid protein VP26 (VP26-GFP) (15). The PRV mutant viruses PRV ΔUS3 NIA (M118) (65) and PRV ΔUL13 NIA (M137) (66), which lacked either US3 or UL13 expression, respectively, were a kind gift from the ID-DLO Institute (Lelystad, The Netherlands). PRV ΔUL41 Kaplan and PRV ΔUL41 rescue Kaplan viruses (67) were kindly provided by T. C. Mettenleiter. Infections with wild-type PRV strains PRV ΔUL41 and ΔUL41 rescue viruses and PRV-GS443 were all performed at a multiplicity of infection (MOI) of 10 PFUs per cell. Infections with PRV ΔUS3 and PRV ΔUL13 viruses were carried out using an MOI of 20 PFU/cell. PRV infections were always performed on confluent cell monolayers.

Chemicals and cytokines.

The viral DNA polymerase inhibitor phosphonoacetic acid (PAA) was purchased from Sigma-Aldrich (catalog number 284270). The alphaherpesvirus helicase/primase inhibitor pritelivir (BAY-57-1293) was obtained from Selleckchem (catalog number S7546). Cycloheximide (CHX) was purchased from Sigma (catalog number C1988). The 26S proteasome inhibitor MG132 was purchased from Sigma-Aldrich (catalog number M7449). The ATM kinase inhibitor KU60019 was obtained from Selleckchem (catalog number S1570). Doxorubicin hydrochloride was purchased from Thermo Fisher Scientific (catalog number BP25161). Purified recombinant porcine TNF-α was obtained from R&D System (catalog number 690-T).

Antibodies targeting PRV proteins.

The mouse monoclonal antibodies used were directed against gB (clone 1C11; 1:100 for Western blotting (WB), 1:50 for immunofluorescence (IF), and 1:20 for FC), gD (clone 13D12; 1:100 for WB and 1:20 for FC), gE (clone 18E8; 1:100 for WB) (68), US3 (1:100 for WB, 1:50 for IF) (69), and VP5 (clone 3C10; 1:2,000 for WB) (70). Anti-IE180 (1:1,000 for WB) (71), anti-UL13 (1:1,000 for WB) (72), and anti-UL41/vhs (1:10,000 for WB) antibodies were rabbit polyclonal antisera. Antibodies against US3, VP5, and UL13 proteins were generous gifts from Lynn Enquist (Princeton University, NJ, USA). Anti-IE180 antiserum was kindly provided by Enrique Tabarés (Autónoma University of Madrid, Madrid, Spain). The antibody used for detection of UL41/vhs was a kind gift from Thomas C. Mettenleiter (Friedrich-Loeffler Insititute, Insel-Reims, Germany).

Antibodies targeting cellular proteins.

The antibodies used for the detection of proteins associated with NF-κB and the DNA damage response pathways were all purchased from Cell Signaling Technology (CST) and included mouse anti-IκBα (clone L35A5; 1:1,000 for WB; catalog number 4814), mouse anti-NF-κB p65 (clone L8F6; 1:1,000 for WB, 1:400 for IF; catalog number 6956), rabbit anti-H2A.X (clone D17A3; 1:1,000 for WB, 1:50 for IF; catalog number 7631), and rabbit anti-phospho-H2A.X Ser139 (1:1,000 for WB and 1:500 for IF; catalog number 2577). The rabbit anti-α-tubulin antibody was purchased from Abcam (1:2,000 for WB; catalog number ab15246).

Western blot analysis.

Cell lysates were collected in ice-cold 1× RIPA buffer (made from 10× radioimmunoprecipitation assay (RIPA) buffer, catalog number ab156034, Abcam) containing a protease inhibitor cocktail (cOmplete mini EDTA free; catalog number 11836170001; Roche). A phosphatase inhibitor cocktail (PhosSTOP; catalog number 04906845001; Roche) was added to the lysis buffer only when the detection of phosphorylated residues was performed. Lysates were incubated at 4°C for 20 min prior to storage at −20°C. SDS-PAGE and Western blotting procedures were described in detail in reference 73. A total of 5% (wt/vol) nonfat dry milk diluted in 0.1% PBS-Tween 20 (PBS-T) was typically used as the blocking solution and then incubated for 1 h at room temperature (RT). However, the blocking buffer used for the detection of phospho-proteins, as well as for detection with certain commercial antibodies, consisted of 5% (wt/vol) bovine serum albumin (BSA; catalog number 160069; MP Biomedicals) diluted in PBS-T and incubated for 1 h at RT. Blots were incubated overnight at 4°C with primary antibodies diluted in the corresponding blocking buffer. After primary antibody incubation, blots were washed three times with PBS-T (10 min each). Incubation with horseradish peroxidase (HRP)-conjugated secondary antibodies, either with goat anti-IgG mouse-HRP (1:2,,000; catalog number P0447; Dako) or with goat anti-IgG rabbit-HRP (1:3000; catalog number P0448; Dako) was performed for 1 h at room temperature. After three washing steps with PBS-T, protein bands were detected by chemiluminescence using a ChemiDoc imaging device (Bio-Rad). The Pierce enhanced chemiluminescence (ECL) substrate (Thermo Scientific), ECL Plus substrate (GE Health Care), or SuperSignal West Femto maximum sensitivity substrate (Thermo Scientific) were used for protein band detection according to protein abundance in the cell lysate and antibody sensitivity.

Immunofluorescence assays.

Cell monolayers were rinsed once with sterile PBS (supplemented with calcium and magnesium; complete PBS) and then fixed with 4% paraformaldehyde for 15 min at room temperature. Fixed cells were incubated for 1 h at 37°C with blocking/permeabilizing buffer (0.3% Triton X-100 and 5% inactivated FBS diluted in PBS). Later, primary antibodies diluted in incubation buffer (0.3% Triton X-100 and 1% [wt/vol] BSA diluted in PBS) were incubated overnight at 4°C. After three washing steps with PBS, cell monolayers were incubated for 1 h at 37°C with fluorochrome-conjugated goat anti-IgG mouse secondary antibodies (1:200; Invitrogen, Thermo Fisher Scientific). Finally, cells were incubated with Hoechst 33342 (1:200; catalog number H3570; Invitrogen, Thermo Fisher Scientific) diluted in PBS for 10 min at room temperature for cell nuclei counterstaining. All fluorescence images were taken using a Leica SPE confocal microscope (Leica) or a Leica Thunder fluorescence microscope (Leica). Fluorescent microscopy images were analyzed using ImageJ imaging software (NIH, USA).

Flow cytometry.

Cells were detached by incubation for 5 min at 37°C in an undiluted Accutase protease solution (StemPro Accutase cell dissociation reagent; catalog number A1110501; Thermo Fisher Scientific). Protease activity was inhibited by adding inactivated FBS up to 15%. Cells were fixed and permeabilized using a Cytofix/Cytoperm solution (catalog 554714BD; BD Biosciences) at 4°C for 20 min, and then fixed cells were washed with 1× BD Perm/Wash buffer (catalog number 554714; BD Biosciences). Primary and secondary antibodies were diluted in 1× BD Perm/Wash buffer (100 μl of antibody solution per 1 million cells) and incubated at 4°C for 40 min. Primary anti-gB and anti-gD antibodies were used in a 1:20 dilution, and the allophycocyanin (APC)-conjugated goat anti-IgG mouse secondary antibody was used in a 1:200 dilution (Invitrogen; Thermo Fisher Scientific). A total of 10,000 cells were analyzed per sample, using a NovoCyte flow cytometer (ACEA Biosciences). NovoExpress software was used for flow cytometry data analysis and representation.

DNA isolation for intracellular PRV genome quantification.

Total DNA was isolated from c.a. 2 million cells. First, cell monolayers were washed twice with sterile complete PBS to remove the vast majority of extracellular virus. In addition, cell monolayers were treated with sodium citrate buffer (pH 3.0; 40 mM sodium citrate, 10 mM KCl, and 135 mM NaCl) for 2 min at room temperature. After a sodium citrate buffer incubation, cells were rinsed three times with complete PBS before cells were harvested. Cell pellets were frozen at −20°C before the DNA isolation was started. DNA isolation was carried out using the DNeasy blood and tissue kit (catalog number 69504; Qiagen) according to the manufacturer’s protocol. DNA yields were diluted in 200 μl, from which 5 μl was used as the template for real-time quantitative PCR (RT-qPCR).

RNA isolation and reverse transcription for mRNA quantification.

Total RNA was isolated using the RNeasy mini kit (catalog number 74104; Qiagen) by following instructions provided by the manufacturer. RNA yields were subjected to DNase I treatment (RNase free; catalog number M0303S; New England BioLabs) for 10 min at 37°C. DNase I reaction was inhibited by addition of EDTA (up to 5 mM EDTA; catalog number AM9260G; Invitrogen) and by incubating samples at 75°C for 10 min. A total of 500 ng of DNA-free RNA was used as the substrate for reverse transcription (RT) using the iScript cDNA synthesis kit (catalog number 1708891; Bio-Rad) according to the manufacturer’s procedure. The one-step RT reaction was set by 5 min at 25°C (priming), 20 min at 46°C (reverse transcription), and 1 min at 95°C (RT inhibition). The RT reaction was carried out in a total volume of 20 μl, from which 1 μl was used as the template for RT-qPCR analysis.

Quantitative PCR (qPCR).

Primer DNA oligonucleotides used for qPCR were synthesized by Integrative DNA Technologies (IDT). When primer sequences were not obtained from the literature, Primer-BLAST software (NIH, USA) was used for primer design. qPCR was performed using SYBR green PCR master mix (catalog number 4309155; Thermo Fisher Scientific). qPCRs were performed in a final volume of 20 μl on MicroAmp Fast optical 96-well reaction plates (catalog number 4346906; Thermo Fisher Scientific) using a StepOnePlus real-time PCR system (catalog number 4376600; Thermo Fisher Scientific). Every sample was measured in duplicate. qPCR data were analyzed by the double delta threshold cycle method.

Intracellular PRV genome quantification.

PRV genomes were quantified by amplification of sequences belonging to two different PRV open reading frame sequences, namely, us3 and ul27, obtained from the PRV reference sequence (Suid herpesvirus 1, complete genome; GenBank accession number NC_006151.1). Cellular genome quantification targeting the porcine beta-2-microglobulin (B2M) gene (Chromosome 1, Sus scrofa isolate TJ Tabasco breed Duroc, NC_010443.5) allowed normalization of viral genome quantification. Primer sequences for intracellular PRV genome quantification assays can be found in Table 1.

TABLE 1.

Primers used for viral genome quantification

Target gene	GenBank accession no.	Sequence of (5′–3′):
		Forward primer	Reverse primer
PRV UL27	NC_006151.1	GTC YGT GAA GCG GTT CGT GAT	ACA AGT TCA AGG CCC ACA TCT AC
PRV US3	NC_006151.1	GAC GGG GGC TTT CCT GAT TTA	GTA TCT CAT CAG CGG AAG GGC
B2M	NC_010443.5	AAA CGG AAA GCC AAA TTA CC	ATC CAC AGC GTT AGG AGT GA

NF-κB-dependent gene expression.

The relative expression of the porcine (Sus scrofa) target genes TNFA (encoding TNF-α), IL-6 (encoding IL-6), NFKBIA (encoding IκBα), and TNFAIP3 (encoding A20) was normalized to the level of expression 28S rRNA gene expression. Primer sequences used for this analysis are described in Table 2.

TABLE 2.

Primers used for detection of NF-κB-dependent gene expression

Target gene	GenBank accession no.	Reference	Sequence of (5′–3′):
			Forward primer (5′–3′)	Reverse primer (5′–3′)
TNFA (encoding TNF-α)	NM_214022.1	74	ACT GCA CTT CGA GGT TAT CGG	GGC GAC GGG CTT ATC TGA
IL6 (encoding IL-6)	NM_214399.1	75	TTC ACC TCT CCG GAC AAA ACT	TCT GCC ATG ACC TCC TTG CTG T
NFKBIA (encoding IκBα)	NM_001005150.1	11	AAG CAC TCG GAT ACA GCA GC	AGT CGT CAT AGG GCA GCT CA
TNFIAP3 (encoding A20)	NM_001267890.1	11	CCT GTT CAG CGA GAC TAC GG	5′-AAC GTC CTG GTG ACG TTC TG-3′
28 rRNA		76	GGG CCG AAA CGA TCT CAA CC	GCC GGG CTT ACC CAT T

Extracellular virus titer determination.

Confluent cell monolayers seeded on 24-well plates were infected at a multiplicity of infection of 10 PFU/cell. At 2 hpi, cell monolayers were rinsed twice with complete PBS and incubated with sodium citrate buffer (pH 3.0; 40 mM sodium citrate, 10 mM KCl, and 135 mM NaCl) for 2 min at room temperature to inactivate remaining virus inoculum. Two additional washing steps with complete PBS were carried out before adding fresh cell culture medium. At the indicated time points, PRV-infected cell supernatants were collected and immediately frozen at −80°C. Extracellular titers were determined by infection of confluent ST cell monolayers seeded in 96-well plates with 10-fold serial dilution of the PRV-infected cell supernatants, using four experimental repeats. The PRV-associated cytopathic effect was used as a readout for infection. Extracellular virus titers were calculated in PFU per milliliter (PFU/ml).

Quantification of cell nuclei areas.

Areas of cell nuclei were determined using ImageJ imaging software (NIH, USA) by contour determination of Hoechst-positive structures (corresponding to nuclear DNA) from confocal microscopy pictures taken with a 63× objective. Nuclear areas of 20 cells were determined per sample, and 3 independent repeats were performed.