Infections with ΔUL31 but not ΔUL34 are delayed for α, β, and γ gene expression.
It has been reported that ΔUL
31 infections result in decreased viral DNA accumulation, cleavage, and packaging (4
). The current studies were initiated to determine how pUL
31 might contribute to these activities, with the hypothesis that UL
31 might augment viral gene expression. Because pUL
31 interacts with pUL
34, we included a ΔUL
34 virus in the experiment to determine if any detected phenotype was dependent upon the pUL
34 interaction. Hep2 cells were mock infected or infected with wild-type HSV-1(F), ΔUL
31, or ΔUL
34 viruses at an MOI of 10 PFU per cell. At 4 and 18 hpi, total cell lysates were collected and proteins were separated by SDS-PAGE as described in Materials and Methods. The results are shown in . For this and subsequent immunoblot assays, quantification was done using Image J. Similar results were noted in Hep2 cells infected with the UL
31 deletion virus at up to 50 PFU per cell (data not shown).
Fig. 1. Delayed protein expression in Hep2 cells infected with ΔUL31 but not ΔUL34 viruses. Hep2 cells were infected with 10 PFU per cell of HSV-1(F), ΔUL31, or ΔUL34 or were mock infected. Total cell lysates were collected at (more ...)
Surprisingly, immunoblotting () showed that essentially no ICP4 (an α gene product) or ICP8 (a β gene product) accumulated to detectable levels by 4 hpi in cells infected with the ΔUL31 virus, whereas both proteins were detected at this time point in cells infected with either the wild-type HSV-1(F) or ΔUL34 viruses. At 4 hpi, ICP4 and ICP8 protein levels in the ΔUL31 infection were only 3% and 5%, respectively, of those seen from the infection with wild-type virus. Although both ICP4 and ICP8 protein levels increased in the lysate of the ΔUL31 infection by 18 hpi, the levels were only 55% and 41% of levels in HSV-1(F) infection. Protein levels of the γ2 gene product gC in the ΔUL31 infection accumulated to only 34% of gC from infection with the wild-type HSV-1(F). For all cases, infections with the ΔUL34 virus resulted in protein levels similar to or higher than wild-type infection. We conclude that UL31 is necessary for optimal expression of at least one member of each kinetic class of viral protein.
Because it has been shown that inhibition of NF-κB activity by the drug resveratrol results in reduced levels of several α gene mRNAs and undetectable levels of gC mRNA (6
), we hypothesized that pUL
31 might augment viral protein expression by ensuring activation of the transcriptional regulator NF-κB. To test this possibility, we probed immunoblots of total infected cell proteins with antibody to IκBα, the NF-κB inhibitor. IκBα binds NF-κB in the cytosol, inhibiting its translocation into the nucleus and thus maintaining a pool of inactive (transcriptionally incompetent) NF-κB (13
). Upon NF-κB activation, IκBα is phosphorylated, polyubiquitinated, and finally degraded. Normally in HSV-1 infections, NF-κB activation is observed as early as 6 hpi, and this activation can be measured by the concomitant absence of the NF-κB repressor IκBα (9
). As shown in , IκBα accumulated to similar levels at 4 hpi in all samples. At 18 hpi, IκBα protein levels from HSV-1(F)- and ΔUL
34-infected cells were 8% and 2% of mock-infected cells, whereas ΔUL
31-infected cell lysates contained protein levels of IκBα at 85% of the mock level. These data indicate that UL
31 plays a role in ensuring elimination of IκBα in infected cells, and thus activation of NF-κB.
We also examined relative levels of activated p-JNK and found that at 18 hpi the ΔUL31 virus induced levels of p-JNK that amounted to only 28% of the level detected from the wild-type HSV-1(F) infection. In contrast, levels of total JNK were similar in all lanes. Taken together, the results indicate that in addition to a viral protein expression deficiency, the ΔUL31 virus exhibits deficiencies in NF-κB and JNK activation that are normally seen during HSV-1 infection.
NF-κB and JNK activation are downregulated in ΔUL31 infections of different cell types.
To determine if NF-κB and JNK activation seen in the ΔUL31-infected Hep2 cells () occurred in other cell types, RSC, Vero, HeLa, and Hep2 cells, as well as the ΔUL31-complementing cell line, clone 7 (an RSC derivative) were mock infected or infected with the ΔUL31, HSV-1(F), or ΔUL31-R (repair) viruses. All cell types were infected at an MOI of 10 PFU per cell, and total cell lysates were collected at 4, 8, 12, and 18 hpi as described in Materials and Methods. Equal amounts of proteins from each sample at each time point were separated by SDS-PAGE, transferred to nitrocellulose membranes, and reacted with antibody for IκBα (the NF-κB repressor), total JNK, and p-JNK as described in Materials and Methods. Sample loading was assessed by probing with an antibody directed against lamin A/C. Although the amounts loaded in each lane were consistent within samples from a given cell line, they varied somewhat between cell lines. Nevertheless, the consistency of loading between samples from each cell type enabled comparisons of protein expression by the different viruses in each cell type examined.
Overall, the results indicated a positive correlation between defects in protein synthesis in ΔUL31 infections () and both decreased JNK and NF-κB activation in the various cell lines () as follows.
Fig. 2. NF-κB and JNK activation are downregulated in ΔUL31 infections of different cell types. RSC, clone 7, Vero, Hep2, and HeLa cells were infected with 10.0 PFU per cell of HSV-1(F), ΔUL31, or ΔUL31-R or were mock infected. (more ...)
In the nonpermissive Hep2 and HeLa cell lines, IκBα levels were similar upon infection with all viruses at 4 hpi, whereas at later time points IκBα was downregulated to 36% and 34% of mock-infected cell levels in HSV-1(F) and ΔUL31-R, respectively. However, in cells infected with ΔUL31, IκBα was downregulated to only 78% of mock levels. The differences were accentuated at 18 hpi, when IκBα was downregulated to 15% and 7% of mock-infected cell levels for HSV-1(F) and ΔUL31-R, respectively, but downregulated to only 76% of mock levels in the ΔUL31 infection.
In HeLa cells, activated JNK was first detected at 8 hpi in cells infected with HSV-1(F) and ΔUL31-R, but in cells infected with ΔUL31 the p-JNK level was only 8% of that in cells infected with ΔUL31-R. Even at 18 hpi, the p-JNK level in ΔUL31-infected cells was only 15% of that in the ΔUL31-R infection. Although activated JNK was never abundant in Hep2 cells, it was first detected at 8 hpi with the wild-type virus. p-JNK levels were significantly lower at later time points (8%, 3%, and 0.4% of p-JNK levels in cells infected with ΔUL31-R at 8, 12, and 18 hpi, respectively).
Vero cells, unlike Hep2 and HeLa cells, exhibited easily detectable phosphorylated JNK as early as 4 hpi, although levels in cells infected with ΔUL31 were 46% of the level observed upon infection with ΔUL31-R. The differences in activated JNK levels decreased over time, such that by 18 hpi, the level of JNK activation in Vero cells infected with ΔUL31 was 92% of that in Vero cells infected with ΔUL31-R. In general, IκBα levels in Vero cells were inversely correlated with those of activated JNK. Although the IκBα level was higher by 8 hpi with ΔUL31 (22% of mock) than upon infection with the repair virus (0.9% of mock), differences among the different viral infections were negligible by 12 hpi (58% for F, 61% for ΔUL31, and 47% for ΔUL31-R relative to mock). We conclude that Vero cells infected with ΔUL31 are more permissive with respect to NF-κB and JNK activation than either Hep2 or HeLa cells infected with this virus.
Immunoblot signals of IκBα were not as strong in RSC as in the other cell lines examined. Nevertheless, this cell line contained more IκBα when mock infected or infected with ΔUL31 than in cells infected with the wild-type HSV-1(F) or ΔUL31-R. Compared to mock infection leves, IκBα levels in HSV-1(F), ΔUL31, and ΔUL31-R were 19%, 167%, and 21%, respectively. The differences in IκBα levels in the different virus infections were largely eliminated by 12 hpi, as the percentage of detectable IκBα relative to the mock level were 3%, 16%, and 2% for HSV-1(F), ΔUL31, and ΔUL31-R, respectively. Very little activated JNK (p-JNK) was detected in RSC at any time point. Thus, rabbit skin cells were more permissive with respect to NF-κB and JNK activation in the context of ΔUL31 than either Hep2 or HeLa cells.
In clone 7 cells, which partially rescue the UL
31 deletion virus (21
), less activated JNK and more IκBα were observed after infection with ΔUL
31 than with the other viruses. Specifically, the IκBα levels were only 11% relative to mock levels in infections with the ΔUL
31-R virus, but 63% with ΔUL
31 at 8 hpi. At 4 and 8 hpi, p-JNK levels in cells infected by ΔUL
31 were 16% and 45%, respectively, compared to levels in infections with the repair virus. Relevant to this observation, we have noted a delay in viral protein expression in clone 7 cells infected with ΔUL
31 (data not shown) and an inability of these cells to restore replication to the level seen upon infection with ΔUL
). Thus, the incomplete rescue of ΔUL
31 replication and protein expression in this cell line correlates with poor JNK and NF-κB activation.
ICP27 provided in trans does not rescue the observed defects of ΔUL31-infected Vero cells.
Previous reports suggested that loss of ICP27 could produce similar defects to those we have observed in characterization of the UL
31 deletion mutant (9
). We therefore examined NF-κB and JNK activation and also gC expression in an ICP27-expressing cell line (V27) infected with ΔUL
31, HSV-1(F), ΔUL
31-R, or d27-1 (ICP27-null) viruses. Vero cells were used as a control, because V27 is derived from this cell type (30
). The results are presented in .
Fig. 3. ICP27 in the V27 cell line does not rescue the protein expression deficits of ΔUL31 at early times after infection. Vero cells or the Vero-derived ICP27-expressing cell line V27 were infected with HSV-1(F), ΔUL31, ΔUL34, or d27-1 (more ...)
In both Vero and V27 cells, JNK activation was decreased in the ΔUL31 infection relative to wild-type HSV-1(F) levels. The presence of ICP27 provided in trans (V27 cells) partially rescued JNK activation in both the ΔUL31 and ICP27-null virus (d27-1). Specifically, p-JNK protein levels in the ΔUL31 and d27-1 infections increased from 15% and 31% (respectively) of wild-type levels at 4 hpi to 50% and 67% (respectively) of wild-type levels at 8 hpi. In contrast, p-JNK protein levels in Vero cells remained at similar levels in cells infected with ΔUL31 compared to wild-type infection at 4 and 8 hpi (66% and 59% of wild-type levels, respectively).
ICP4 expression in ΔUL31-infected Vero and V27 cells was dramatically reduced compared to wild-type infections in these cell types. At 4 and 8 hpi, ICP4 protein levels in the ΔUL31 Vero infections were 0.6% and 54% of wild type, respectively. The presence of ICP27 provided in trans failed to rescue ICP4 expression for both ΔUL31 and ICP27-null (d27-1) infections by 8 hpi. In ΔUL31 and d27-1 infections in V27 cells, ICP4 protein levels were only 17% and 8% of wild-type infection (respectively) at 4 hpi and increased to only 23% and 11% of wild-type infection (respectively) at 8 hpi.
Lysates from ΔUL31-infected Vero and V27 cells contained readily detectable protein levels of IκBα at 8 hpi (70% and 119% of mock infection levels, respectively), whereas wild-type infections resulted in IκBα levels that were only 32% and 16% of those observed in mock infection. This is an indication that NF-κB was not activated normally in the ΔUL31 infections. ICP27 provided in trans was sufficient to rescue NF-κB activation in the ICP27-null virus by 8 hpi, since the IκBα protein level from the d27-1 infection was only 10% of the mock level. We conclude that ΔUL31's defects in protein expression and NF-κB and activation cannot be complemented in an ICP27-expressing cell line, and JNK activation is only partially rescued. This suggests that UL31's contribution to activation of NF-κB and JNK is mostly independent of ICP27.
Infection with ΔUL31 leads to delayed expression of ICP4 but not failure to initiate infection.
One possible explanation of the current results is that the ΔUL31 virus is unable to efficiently enter or otherwise infect cells. In this scenario, the observed phenotypes might simply reflect a large number of uninfected cells in the population. To address this possibility, we infected the restrictive Hep2 cell line with the ΔUL31 or ΔUL31-R viruses at an MOI of 10 PFU per cell. Cells were fixed at 1, 6, and 10 hpi and stained with antibody to ICP4 to visualize infected cells and with Hoechst dye to visualize cellular nuclei. The results are presented in A.
Fig. 4. Infection with ΔUL31 leads to delayed expression of ICP4 but not failure to initiate infection. (A) Hep2 cells were infected at an MOI of 10 PFU per cell of either ΔUL31 or the genetic repair contstruct, ΔUL31-R,and were fixed (more ...)
At all time points examined, nuclei were present throughout the visual field. At 1 hpi, no ICP4-specific immunofluorescence was visible in ΔUL31- or ΔUL31-R infected cells. At 6 hpi many ΔUL31-R infected cells exhibited ICP4-specific immunofluorescence, whereas only a few cells infected with ΔUL31 displayed such immunostaining. At 10 hpi, many ICP4-positive cells were found throughout the ΔUL31-R- and ΔUL31-infected cell monolayers. These observations are consistent with the data presented above, indicating that cells infected with wild-type viruses express high levels of ICP4 at early time points, whereas ICP4 expression in ΔUL31-infected cells starts at very low levels that increase with time postinfection. The gradual increase in ICP4-expressing cells upon infection with the UL31 deletion virus cannot be a consequence of cell-to-cell spread within the culture because (i) this would not be expected at 6 hpi and (ii) the UL31 deletion virus cannot exit noncomplementing cells. A quantification of infection efficiency at 6 hpi and 10 hpi is shown in B. Student's t test demonstrated that the difference in the number of ICP4-expressing cells of total cells was statistically significant (P < 0.001), whereas this difference was no longer statistically different at 10 hpi (P > 0.05). The difference at the 6- and 10-hpi time points in ΔUL31-infected cells was also statistically different (P < 0.01), indicating that a significant number of cells transitioned from not expressing ICP4 at detectable levels to detectable expression over the course of 4 h. We conclude that the defects in ICP4 expression observed in cells infected with the UL31 deletion were due to a delay in the onset of expression, rather than a failure of the virus to enter and infect a subset of cells.
UL31 expression is detectable early in infection.
The effects of pUL31 in early protein expression were seemingly incongruous with its sole expression late in infection. To test whether or not UL31 is only expressed late in infection, Hep2 cells were treated with 10 μM CHX or left untreated for 1 h. The cells were then infected with wild-type HSV-1 (MOI, 10) or mock infected. For CHX-treated cells, the CHX level was maintained at all steps of infection. Total cell lysates were collected at 1, 2, 3, and 5 hpi, and proteins were separated by SDS-PAGE. Immunoblots were probed with antibody to pUL31, ICP4, and lamin A/C (loading control). The results are presented in .
Fig. 5. UL31 expression is detectable early in infection. Hep2 cells were pretreated with 10 μM CHX for 1 h at 37°C, 5%CO2 or left untreated. After the 1-h treatment, cells were infected with 10 PFU per cell of wild-type HSV-1(F) or were mock (more ...)
The data indicate that UL31 is expressed at detectable levels as early as 2 hpi. The kinetics of pUL31 expression early in infection correlated with those of ICP4 expression, whereas at 5 hpi, the amount of pUL31 increased more dramatically than those of ICP4. Importantly, little or no protein was visible in the CHX-treated lysates, indicating that, as expected, the early detection of pUL31 is due to active transcription/translation and was not a consequence of its presence within the inoculum. These data indicate that UL31, previously thought to be expressed only at late times after infection, has low-level expression at early times after infection.
UL31 expression late in infection is PAA sensitive.
Since true late genes are sensitive to DNA polymerase inhibitors, we asked whether pUL31 expression late in infection was dependent on vDNA synthesis (i.e., sensitive to the DNA polymerase inhibitor PAA). To test this possibility, Hep2 cells were infected with wild-type HSV-1(F) (MOI, 10) or were mock infected in the presence or absence of 200 μg/ml PAA. At 12 h postinfection total cell lysates were collected and proteins separated by SDS-PAGE. The immunoblot was reacted with antibody to pUL31 or lamin A/C as a loading control. The results are presented in .
Fig. 6. UL31 expression is dependent on viral DNA replication. Hep2 cells were infected with wild-type HSV-1(F) (MOI, 10 PFU per cell) or were mock infected in the presence or absence of 200 μg/ml PAA at the start of infection. At 12 hpi, total cell lysates (more ...)
The presence of PAA significantly diminished pUL31 expression in HSV-1-infected cells. These data are consistent with UL31 expression as a late gene but are not inconsistent with an important role(s) of pUL31 early in infection.