UV Irradiation of Vaccinia Virus-Infected Cells Impairs Cellular Functions, Introduces Lesions into the Viral Genome, and Uncovers Repair Capabilities for the Viral Replication Machinery

ABSTRACT

Vaccinia virus (VV), the prototypic poxvirus, encodes a repertoire of proteins responsible for the metabolism of its large dsDNA genome. Previous work has furthered our understanding of how poxviruses replicate and recombine their genomes, but little is known about whether the poxvirus genome undergoes DNA repair. Our studies here are aimed at understanding how VV responds to exogenous DNA damage introduced by UV irradiation. Irradiation of cells prior to infection decreased protein synthesis and led to an ∼12-fold reduction in viral yield. On top of these cell-specific insults, irradiation of VV infections at 4 h postinfection (hpi) introduced both cyclobutene pyrimidine dimer (CPD) and 6,4-photoproduct (6,4-PP) lesions into the viral genome led to a nearly complete halt to further DNA synthesis and to a further reduction in viral yield (∼35-fold). DNA lesions persisted throughout infection and were indeed present in the genomes encapsidated into nascent virions. Depletion of several cellular proteins that mediate nucleotide excision repair (XP-A, -F, and -G) did not render viral infections hypersensitive to UV. We next investigated whether viral proteins were involved in combatting DNA damage. Infections performed with a virus lacking the A50 DNA ligase were moderately hypersensitive to UV irradiation (∼3-fold). More strikingly, when the DNA polymerase inhibitor cytosine arabinoside (araC) was added to wild-type infections at the time of UV irradiation (4 hpi), an even greater hypersensitivity to UV irradiation was seen (∼11-fold). Virions produced under the latter condition contained elevated levels of CPD adducts, strongly suggesting that the viral polymerase contributes to the repair of UV lesions introduced into the viral genome.

IMPORTANCE Poxviruses remain of significant interest because of their continuing clinical relevance, their utility for the development of vaccines and oncolytic therapies, and their illustration of fundamental principles of viral replication and virus/cell interactions. These viruses are unique in that they replicate exclusively in the cytoplasm of infected mammalian cells, providing novel challenges for DNA viruses. How poxviruses replicate, recombine, and possibly repair their genomes is still only partially understood. Using UV irradiation as a form of exogenous DNA damage, we have examined how vaccinia virus metabolizes its genome following insult. We show that even UV irradiation of cells prior to infection diminishes viral yield, while UV irradiation during infection damages the genome, causes a halt in DNA accumulation, and reduces the viral yield more severely. Furthermore, we show that viral proteins, but not the cellular machinery, contribute to a partial repair of the viral genome following UV irradiation.

INTRODUCTION

Vaccinia virus is a close relative of variola virus, the etiological agent of smallpox and was the virus used in the worldwide vaccination campaign that eradicated this horrific disease. Vaccinia is the experimental prototype for the study of the poxvirus life cycle. The large enveloped virions contain ∼80 structural proteins and enzymes, and the 195-kb dsDNA genome, which has unusual hairpin termini, encodes ∼200 proteins (1). Notably, poxviruses are the only DNA viruses to replicate exclusively in the cytoplasm of mammalian cells. The physical separation between the viral genome and host DNA replication factors necessitates that poxviruses encode most if not all of the factors required for the replication and maturation of their genomes.

Extensive biochemical and genetic analyses have identified a repertoire of virally encoded factors that are essential for vaccinia DNA replication (2). The replicative holoenzyme comprises E9, the catalytic polymerase and proofreading exonuclease, and a heterodimeric processivity factor, A20 and D4 (3–9). D4 is also an active uracil DNA glycosylase. D5, a AAA+ ATPase, is the putative primase/helicase that also plays an important role in the uncoating step that releases the genome from the intracellular virion core prior to replication (10–12). H5 has the hallmarks of a replication scaffold, or hub protein, and is essential for viral DNA replication (13, 14). The A50 protein is a DNA ligase; in tissue culture cells, the loss of A50 can be complemented by the presence of the cellular DNA ligase I (15). G5, a Fen1/Exo1-like nuclease, is essential for the formation of full-length monomeric viral genomes and is involved in recombination (16; M. W. Czarnecki and P, Traktman, unpublished data). I3 is the replicative single-stranded binding protein (17, 18) and also appears to contribute to recombination (6). Additional proteins include enzymes that contribute to nucleotide precursor biosynthesis (e.g., ribonucleotide reductase and thymidine kinase), deplete the dUTP pool (dUTPase), contribute to the maturation of progeny genomes (Holliday junction resolvase), or serve as a countermeasure to the intrinsic antiviral effect of the cellular DNA binding protein BAF (B1 protein kinase) (19–26). Finally, the I6 telomere-binding protein and A32, a AAA+ ATPase of the Her/FtsK family, are essential for the encapsidation of the progeny genomes into nascent virions (27–30).

DNA metabolism can be broken down into three processes that are likely to overlap: replication, recombination, and repair. Most previous studies of vaccinia DNA metabolism have focused on replication or recombination. However, there is increasing interest in how DNA viruses respond to endogenous and exogenous DNA damage, and in the intersection of viral genomes with the host cell DNA damage response (DDR). For the work described here, we chose UV irradiation as a physiologically relevant source of exogenous DNA damage, since the pathology of poxviruses involves the formation of lesions in the skin, thereby exposing the virus to UV irradiation. The damage to DNA upon exposure to UV irradiation has been well studied, as has the nucleotide excision repair (NER) pathway used by the cell to identify and remove such lesions (31, 32).

UV irradiation (measured in J/m2) induces a number of cellular signaling pathways and causes two forms of DNA damage: cyclobutene pyrimidine dimers (CPDs) and 6,4-photoproducts (6,4-PPs) (31). UV irradiation activates the DDR and endoplasmic reticulum (ER) stress pathways in a dose-dependent manner, although the mechanism of ER stress activation is controversial (33, 34). CPDs and 6,4-PPs form aberrant secondary structures in the DNA helix due to the covalent bonding between adjacent pyrimidine bases. The bulky adducts and helical distortions that result from UV irradiation cause replicative holoenzymes to stall and replication forks to collapse.

The studies described here were motivated by an interest in determining how UV irradiation impacts the stability and ongoing replication of poxviral genomes during the viral life cycle. Therefore, we exposed infected cultures to UV irradiation at 4 h postinfection (hpi), early in the linear phase of genome replication. We hypothesized that, if no repair machinery was available within the cytoplasmic replication factories, replication might stall after radiation-induced lesions formed in the genome. It also seemed possible that the viral polymerase had evolved the ability to bypass the lesions and persist in synthesizing DNA. Finally, although no recognizable viral homologs of NER proteins have been identified in the vaccinia genome, the virus might nevertheless encode suitable repair machinery or redirect the cellular NER machinery to function within the cytoplasmic viral replication factories.

Finally, we realized that when irradiating infected cells at 4 hpi, we were not only irradiating the viral genome but exposing the cells to irradiation. It is widely believed that both polymerase II-mediated transcription and replication of cellular genomes have ceased by 4 hpi. Nevertheless, although DNA is the primary target of UV irradiation, activation of the DDR has potent effects on numerous cellular processes. Therefore, we also assessed whether irradiating cells at 1 h prior to infection (−1 hpi) compromised their ability to serve as effective hosts during subsequent infection. Irradiation of cells at either −1 or 4 hpi diminishes protein synthesis and, therefore, leads to a modest reduction in virion production. Irradiation at 4 hpi leads to an abrupt and durable repression of viral genome accumulation. Our data also indicate that the cellular NER pathway does not contribute to the virus’ ability to respond to UV irradiation; in contrast, key components of the viral replication machinery have an ameliorative effect.

RESULTS

UV irradiation of infected cells at 4 hpi leads to the accumulation of CPD and 6,4-PP adducts in the viral genome and adversely affects DNA accumulation and viral yield. (i) Impact on viral genome accumulation and integrity.

After virion entry, early gene expression occurs within the subviral core. After early proteins are expressed, the core undergoes an uncoating event that releases the viral genome and allows DNA replication to initiate. The linear phase of replication is thought to take place between about 2 and 10 hpi. Therefore, when we UV irradiate our infected cells at 4 hpi, we are introducing UV lesions (6,4-PPs and CPDs) during ongoing viral DNA replication. These bulky adducts are known to cause helix distortions and have been shown to stall the movement of processive cellular polymerases (35–37). We therefore assessed the profile of viral DNA accumulation in unirradiated cells and cells exposed to 30 or 60 J/m² UV (4 hpi). The time course of viral DNA accumulation (3, 4, 4.5, 7, and 10 hpi) was quantified using Southern dot blot analyses (Fig. 1A, left and quantified on the right). The characteristic profile of DNA accumulation was seen in unirradiated cells; however, in cells exposed to 30 or 60 J/m2 UV at 4 hpi, no further accumulation of viral DNA was seen between 4 and 10 hpi.

FIG 1.

UV irradiation during VV infection introduces UV lesions into the VV genomes and halts further VV DNA accumulation. UV irradiation (30 or 60 J/m²) at 4 hpi halts further viral DNA accumulation (A and B) and introduces UV lesions, both 6,4-photoproducts (6,4-PP) and cyclobutane pyrimidine dimers (CPD), into the viral genome (C and D). (A) BSC40 cells were either left uninfected or infected with WT virus (MOI of 5) and left unirradiated (blue) or irradiated at 4 hpi (30 or 60 J/m², pink or purple, respectively). Samples were harvested at 3, 4, 4.5, 7, or 10 hpi, and the levels of viral DNA were assessed by Southern dot blot analysis. Each sample was spotted in technical triplicate (left panel); a representative blot is shown. The data are plotted as the average of three biological experiments with the error bars representing the standard errors of the mean (SEM; right panel; ***, P < 0.001). (B) BSC40 cells were infected with WT virus (MOI of 5) and either left unirradiated or UV irradiated (15, 30, 45, or 60 J/m²) at 4 hpi before being collected at 10 hpi and processed for PFGE analysis. A representative image of the EtBr staining is shown (left), as well as the corresponding Southern blot (right); the arrow marks the monomeric viral genome. The data are plotted as the averages of six biological replicates with the error bars representing the SEM (right panel; ***, P < 0.001). (C and D) BSC40 cells were infected with WT virus (MOI of 5) for 4 h and left unirradiated or UV irradiated with 60 J/m² before being fixed at 4.5 hpi and stained with DAPI (blue), anti-I3 (green), and either anti-6,4-PP (C; red) or anti-CPD (D; red). Both 6,4-PPs and CPDs localize to the viral cytoplasmic replication factories following UV irradiation at 4 hpi. Scale bar, 100 μm.

Southern dot blots provide a quantitative assessment of bulk viral DNA levels but do not reveal the integrity of the viral genomes. Subgenomic fragments might accumulate if the viral replicative holoenzyme synthesized DNA before stalling and collapsing across the bulky adducts caused by pyrimidine dimer formation. Furthermore, if degradation of viral genomes was occurring, we hypothesized that full-length genomes containing UV lesions might be partially digested by cellular cytoplasmic DNA nucleases such as DNase2 or TREX1 (38, 39). Using pulsed-field gel electrophoresis (PFGE), we visualized the levels of DNA and integrity of viral genomes at 10 hpi in control infections or infections exposed to various doses of UV irradiation at 4 hpi (Fig. 1B). Viral DNA was visualized both by EtBr staining (left) and Southern hybridization (right); the 195-kb genome is shown by the arrowhead. Consistent with our Southern dot blots, all concentrations of UV irradiation reduced viral DNA levels by ∼4-fold (Fig. 1B, quantified on the right). Irradiation did not lead to the appearance of subgenomic fragments, suggesting that widespread DNA degradation is unlikely to be occurring and supporting the hypothesis that processive/productive viral DNA replication is blocked by UV-induced lesions.

(ii) Identification and assessment of UV lesions in viral genomes.

We performed immunofluorescence analysis of infected cells that had been UV irradiated at 4 hpi or left untreated (Fig. 1C and D). In irradiated cells, clear signals corresponding to CPDs and 6,4-PPs localized to the viral cytoplasmic replication factories as marked by the replicative single-stranded binding protein, I3. No lesions were detected in unirradiated cells. Although recognition of CPD and 6,4-PP lesions in nuclear DNA by these antibodies requires a prior denaturation step (40), denaturation was not required to detect UV-induced lesions in the vaccinia genome, and omission of this step preserved the morphology of the cytoplasmic replication factories. These data suggest that either the high AT content of the viral genome (68%) provides sufficient access to the antibodies or that a significant portion of replicating DNA has single-stranded character.

(iii) Impact of UV irradiation on the profile of viral protein accumulation.

The onset of viral DNA replication is a prerequisite for the transition from early gene transcription to intermediate and late gene transcription. Because we irradiated cultures at 4 hpi when replication was ongoing, we would not expect this transition to have been prevented. However, transcript levels might have been compromised by the decrease in the number of transcriptional templates and/or the possibility that UV adducts might block transcription. Decreased viral transcription could diminish the levels of intermediate and late proteins that accumulate. We assessed the steady-state levels of the L4 and F17 late proteins at 18 hpi. As shown in Fig. 2A, exposure to 15 to 60 J/m² UV led to a modest (∼3-fold at 60 J/m²) dose-dependent decrease in protein levels. In contrast, no reduction in the levels of the early protein I3 were seen at any dose (Fig. 2A).

FIG 2.

UV irradiation during infection reduces viral infectious yield and late protein accumulation in a dose-dependent manner. Late protein accumulation and viral infectious yield are reduced in a dose-dependent manner following UV irradiation at 4 hpi. BSC40 cells were infected with WT virus (MOI of 5). At 4 hpi, samples were either left unirradiated (blue) or irradiated with 15, 30, 45, or 60 J/m² (pink to purple). All samples were harvested at 18 hpi and either processed for immunoblotting (A) or titrated via plaque assay to assess viral yield (B). Immunoblots were probed for calnexin (loading control), p53, I3, L4, and F17 (viral proteins) (n = 3; a representative immunoblot is shown). Viral yield is plotted as the average of three biological replicates with the error bars representing the SEM (**, P < 0.01; ***, P < 0.001).

(iv) Impact of UV irradiation on the yield of infectious virus production.

Having established that UV irradiation at 4 hpi blocks further accumulation of viral genomes and has a modest impact on the steady-state levels of late viral proteins, we next examined the impact on the yield of infectious virus (multiplicity of infection [MOI] of 5, 18 hpi). Increasing doses of UV irradiation (15 to 60 J/m²) decreased infectious yield in a dose-dependent manner (Fig. 2B), with 60 J/m² leading to a 35-fold decrease relative to the control. This same dose of UV had reduced DNA and late protein accumulation (4- and 3-fold, respectively), suggesting that other processes might be compromised.

Does irradiation of cells prior to infection impact viral DNA replication and the yield of infectious virus? (i) Irradiation of cells prior to infection (–1 hpi) reduces viral yield but has no impact on the accumulation of the viral genome.

UV irradiation activates numerous cellular stress responses, including the DDR pathway (reviewed in reference 31) and the ER stress pathway (33, 34). Therefore, we wanted to test the possibility that irradiation of cells impairs their capacity to support viral infection, independent of the introduction of lesions into the viral genome. BSC40 cells were therefore UV-irradiated 1 h prior to infection, and infectious yield and viral protein accumulation were assessed at 18 hpi. Preirradiation led to a dose-dependent decrease in viral infectious yield (Fig. 3A), although less pronounced than irradiation at 4 hpi: the impacts of 30 and 60 J/m² UV were 6- and 12-fold, respectively, whereas decreases of 15- and 35-fold were seen after irradiation of 60 J/m² UV at 4 hpi. Notably, preirradiation of cells had no effect on the levels of viral DNA that accumulated after infection (Fig. 3B), in contrast to the significant impact seen in cells irradiated at 4 hpi. Clearly, exposure of cells to UV prior to infection could not have caused lesions within the viral DNA. Furthermore, these data argue against the possibility that one or more cellular factors needed for the replication of the viral genome might have been lost or inactivated upon irradiation.

FIG 3.

UV irradiation prior to infection reduces viral infectious yield but not DNA accumulation. Although UV irradiation prior to infection reduces viral infectious yield, protein synthesis and late protein accumulation, viral DNA accumulation is unaffected. (A) UV irradiation at −1 hpi reduces viral yield. BSC40 cells were infected with WT virus (MOI of 5) for 18 h and were either left unirradiated or UV irradiated 1 h prior to infection (–1 h) or at 4 hpi (4 h) with either 30 or 60 J/m². The viral yield was assessed via plaque assay (n = 3, average and SEM plotted; **, P < 0.01). (B) Viral infection at 4 hpi, but not at −1 hpi, reduces DNA accumulation. Following the same experimental design used for panel A, samples were collected at 10 hpi, and the levels of viral DNA assessed by Southern dot blotting (n = 3, average and SEM plotted; ***, P < 0.001). (C) UV irradiation at both −1 hpi and 4 hpi reduces late protein accumulation. The samples described in panel A were also subjected to immunoblot analysis. Blots were probed for calnexin, I3, L4, and F17 (n = 3, a representative immunoblot is shown). (D) UV irradiation at both −1 hpi or 4 hpi reduces protein synthesis. BSC40 cells were either uninfected (lanes 1 and 2) or infected with WT virus (MOI of 5) (lanes 3 to 14); cells were also either left unirradiated (lanes 1 and 3 to 6) or irradiated at −1 hpi (lanes 2 and 7 to 10) or 4 hpi (lanes 11 to 14). At each time point (uninfected or 3, 6, 9, and 12 hpi), the cells were labeled with [³⁵S]methionine for 45 min. The cell lysates were resolved by SDS-PAGE to visualize the profile of nascent proteins (n = 3; a representative image is shown). (E) UV irradiation results in increased levels of phosphorylated initiation factor eIF2α in both uninfected and infected cells. BSC40 cells were either left uninfected (lanes 4 to 6, 10, and 11) or infected with WT virus (MOI of 5) (lanes 1 to 3, 7 to 9, 12, and 13); cells were also either left unirradiated (lanes 1 to 3) or irradiated at −1 hpi (lanes 4 to 9) or 4 hpi (lanes 10 to 14). At each time point (3, 6, or 9 hpi), cells were collected and processed for immunoblot analysis. Blots were probed for ATR, calnexin (loading control), Ku70, pChk1 S345, tChk1, eIF2α, p-eIF2α, F17, and L2 (viral proteins) (n = 3, a representative immunoblot is shown).

(ii) UV irradiation, both pre- and postinfection, diminishes protein synthesis.

We also compared the impact of preirradiation of cells (–1 h) to irradiation during infection (+4 h) on the accumulation of late viral proteins (Fig. 3C). Interestingly, a modest and comparable decrease in the levels of the L4 and F17 proteins was seen in both cases. These data suggested that an impact of irradiation on protein synthesis might be involved. As shown in Fig. 3D and E, we compared the profiles of nascent protein synthesis in metabolically labeled cells. UV irradiation of uninfected cells reduced protein synthesis as quickly as 1 h past UV irradiation (Fig. 3D, compare lanes 1 and 2), as has been reported previously (33, 34). We also monitored protein synthesis at 3, 6, 9, and 12 h after infection with vaccinia virus, comparing cells that had been left unirradiated (lanes 3 to 6) with those irradiated at 1 h prior to infection (lanes 7 to 10) or at 4 hpi (lanes 11 to 14). Although the overall profile of nascent proteins is consistent, irradiation decreased protein synthesis, with the greatest impact being seen closest to the time of irradiation. In addition to resolving the samples electrophoretically (Fig. 3D), we also quantified the trichloroacetic acid (TCA)-insoluble [³⁵S]methionine from uninfected cultures either unirradiated or 1 h after UV irradiation. In samples prepared from uninfected BSC40 cells 1h after they were UV irradiated with 60 J/m², there was a 3-fold reduction in the TCA-precipitable [³⁵S]methionine-labeled proteins relative to control cells. The magnitude of effect is similar to previously reported findings (33, 34). To assess how UV irradiation might lead to decreased protein synthesis, we assessed the levels of various stress-related cellular proteins (Fig. 3E). The levels of the ATR protein kinase, the DNA repair protein Ku70 or the stress-related kinase Chk1 were unchanged by UV irradiation or infection. However, the levels of pChk1 (the activated, phosphorylated form of Chk1) were significantly upregulated after UV irradiation of either uninfected or infected cells. Most importantly for the process of protein synthesis, the levels of the repressive, phosphorylated form of the translation initiation factor eIF2α (pEIF2α) were significantly upregulated after irradiation of uninfected or infected cells. (The levels of total EIF2α were unchanged.) Note that the levels of pEIF2α are somewhat lower in irradiated, infected cells than in irradiated uninfected cells, which may reflect the presence of the viral K3 protein, an eIF2α mimic. Regardless, the increase in p-eIF2α seen after irradiation is likely the cause of the decreased protein synthesis seen in Fig. 3D.

UV irradiation prior to (−1 hpi) or during (4 hpi) infection decreases production of mature virions.

VV late proteins are crucial for morphogenesis and virion maturation. When cells are UV irradiated prior to or postinfection, steady-state levels of late viral proteins are reduced. To assess whether VV virion maturation was affected by UV irradiation, we took two approaches. First, we performed transmission electron microscopy on samples harvested at 18 hpi (Fig. 4). Control, unirradiated cells contained numerous mature virions, as well as immature virions (IV) and immature virions with nucleoids (IVN) (Fig. 4 top). In cells that were irradiated at either −1 or 4 hpi, there was an evident reduction in the number of mature virions, as well as a reduction in each of the morphogenesis intermediates (crescents, IV, and IVN). Interestingly, there also appeared to be a small number of electron dense virions in the cells that were irradiated at 4 hpi, reminiscent of aberrant virions that lack encapsidated viral genomes (Fig. 4 bottom) (28). Our findings indicate that UV irradiation, regardless of when it is introduced, has a modest effect on virion maturation.

FIG 4.

UV irradiation prior to or during infection causes defects in virion maturation. UV irradiation has a modest impact on the profile of virion assembly. BSC40 cells were either left unirradiated (top) or irradiated prior to (−1 hpi, bottom) or after (4 hpi, middle) infection. Cells were infected with WT virus (MOI of 5) for 18 h before being processed for electron microscopy. Representative images are shown. Labels: C, crescent membranes; IV, immature virions; IVN, immature virions with nucleoids; MV, mature virions; *, aberrant virion. Scale bars, 200 nm.

Analysis of the yield, composition and infectivity of mature virions produced during control, pre- and postirradiated infections.

Mature virions form light scattering bands after sucrose gradient sedimentation and both the relative migration of these bands as well as the content and infectivity of the virions within them can be assessed. We therefore resolved virions from our unirradiated, preirradiated, and postirradiated infections using sucrose density gradients as previously described (41). Infections performed in pre- or postirradiated cells produced 3- to 4-fold fewer mature virions than control infections as assessed from the intensity of the light scattering band (Fig. 5A). The position of the bands within the gradients was comparable in all the samples. Therefore, we conclude that there is not a marked accumulation of “empty” virions under any of these conditions, since virions lacking encapsidated genomes are less dense than WT virions and sediment at a higher position in the gradient (28).

FIG 5.

UV irradiation prior to or during infection reduces both virion number and infectivity. UV irradiation prior to or during infection reduces the yield of mature virions. BSC40 cells were infected with WT virus (MOI of 5) and either left unirradiated or UV irradiated prior to (−1 h) or after (4 h) infection. At 18 hpi, samples were collected, and virions were purified on sucrose gradients. Light scattering bands were photographed (A) before fractions were dripped and collected for analysis by Southern dot blot (B), immunoblot (C), plaque assay (D), and PFGE (E). (B) Virions produced from infections UV irradiated prior to or postinfection contain less DNA than WT virions. Volumes for Southern dot blot were adjusted to load equal amounts of virions based on the intensity of the light scattering band and spotted in technical duplicate. (C) Encapsidated viral proteins are reduced in UV irradiated infections. Peak fractions identified by BCA were immunoblotted for L4 and F17. (D) The infectious yield produced following UV irradiation prior to or postinfection is also reduced compared to WT virions; virions produced from cells irradiated at 4 hpi contain CPD lesions. Fractions 11 through 14 were combined and titrated on BSC40 cells (n = 1; t = 3, average and SEM plotted; *, P < 0.05). (E) UV irradiation during infection results in the encapsidation of viral genomes containing UV lesions. The pooled fractions were also resolved in duplicate by PFGE and assessed by EtBr staining and Southern blotting with a ³²P-labeled viral DNA probe or with the CPD recognizing antibody.

The gradients were then fractionated and fractions surrounding the light scattering band were analyzed for the levels of virion proteins, viral DNA and infectivity. Examination of the protein content with the light scattering band showed that the levels of L4, a late viral protein that is proteolytically cleaved during morphogenesis, and F17, a late viral protein, were reduced in each fraction of the pre- and postirradiated infections compared to control infections (Fig. 5B). Assuming that the levels of the proteins are an indicator of virion number, the values extrapolated from the immunoblots are consistent with those extrapolated from the intensity of the light diffracting bands. Of note, the levels of residual, unprocessed L4 are higher in the pre- and postirradiated samples, indicating that virion maturation may be partially compromised under these conditions.

We also assessed the content of viral DNA in each of our fractions using a Southern dot blot protocol. We analyzed what we calculated to be the same number of virions for all samples, based on the extrapolations of virion number described above. Analysis of the data indicated that there was an approximately 33% decrease in the levels of viral DNA in the pooled virion sample (fractions 11 to 14) in both our pre- and postirradiated virions compared to the unirradiated virions (Fig. 5C). These data suggest that a minor portion of the virions purified from irradiated infections lack the viral genome.

We next examined the specific infectivity of the virions collected from each condition by pooling the peak fractions, 11 to 14, and performing a plaque assay. Compared to control infections, the infectious yield was reduced ∼15-fold in the preirradiated infections and ∼40-fold in the postirradiated infections (Fig. 5D). This reduction greatly exceeds the ∼3- to 4-fold reduction in the relative intensities of the light diffracting bands and the levels of L4 and F17. For the preirradiated samples, the reduction in virion production and the modest decrease in encapsidated genomes might explain the majority of the decrease in infectious yield.

Detection of UV-induced lesions in viral DNA within infected cells and purified virions.

To further analyze the genomes present within the banded virions, we embedded the virions in agarose plugs for analysis by PFGE. The DNA was resolved in duplicate: one sample was visualized by ethidium bromide (EtBr) staining and Southern blot analysis with a probe derived from the viral genome, the other sample was probed for UV-induced lesions using an anti-CPD antibody. Our EtBr and Southern blot analyses (Fig. 5E, left and middle) detected encapsidated monomeric genomes (∼194 kb), which were clearly present in decreased levels in the samples prepared from pre- and postirradiated infections relative to that prepared from an unirradiated control infection. When probed with the CPD antibody (Fig. 5E, right), it was evident that lesions were present in the monomeric genomes of virions irradiated during infection (4 hpi) but not in those irradiated prior to infection (−1 hpi), as would be expected. These data indicate that viral genomes containing UV-induced lesions are indeed encapsidated; the significant reduction in the infectivity of these virion preps is likely to reflect the inability of these damaged genomes to support gene expression in the next round of infection.

There were also abundant CPD lesions present in a smear of DNA (∼47 kb) that must be of cellular origin, because the smear is not recognized by the ³²P-labeled viral DNA probe (middle panel): this CPD-containing smear is present in the samples that were irradiated prior to (−1 hpi) or after infection (4 hpi), as would be expected. These fragments of cellular DNA were carried along in the lysate loaded onto the sucrose gradients, and at least some of the fragments must be present in the same fractions as the light-diffracting virion band.

Assessment of the longevity and possible repair of CPD lesions within intracellular viral genomes.

As shown in Fig. 1, CPD and 6,4-PP lesions could be detected within viral replication factories by immunofluorescence analysis. Furthermore, as shown in Fig. 5, CPD-containing genomes were encapsidated into mature virions. To gain a more quantitative appreciation of the longevity of the lesions (introduced by irradiation at 4 hpi) throughout the infectious cycle, uninfected cells or infected cells (control, preirradiated, or postirradiated) were harvested and subjected to PFGE. Duplicate sets of samples were transferred to nitrocellulose membranes; one set was hybridized to a ³²P-labeled viral DNA probe to visualize genomes (Fig. 6A, left) and the other set was probed with an anti-CPD antibody (Fig. 6A, right) to visualize UV-induced lesions. As a control, uninfected or infected (10 hpi) samples were used to identify viral monomeric genomes (∼194 kb). In parallel, we irradiated cells at 4 hpi and harvested them at 4.5, 7, 10, and 18 hpi. As expected from earlier analyses (Fig. 1), the levels of the viral genome were significantly lower in irradiated versus unirradiated cells (Fig. 6A, left, compare lane 3 with lanes 4 to 8). As seen in the anti-CPD blot, lesions were observed at all time points examined within the monomeric genomes present in irradiated cells (Fig. 6A, right), indicating that the lesions are long-lived. When we plotted the relative CPD/genome ratios, we observed an ∼2-fold increase in viral monomeric genomes and an ∼2-fold decrease in CPDs from 4.5 to 18 hpi after UV irradiation (Fig. 6B). However, it is important to note that the level of viral genomes present at 18 hpi in irradiated cultures was <20% of that seen at 10 hpi in unirradiated cells (Fig. 6C).

FIG 6.

UV lesions are long-lived in VV genomes. When infected cells are irradiated at 4 hpi, a 2-fold increase in monomeric genomes and a corresponding decrease in CPD lesions is observed at late times of infection (10 to 18 hpi). BSC40 cells were left uninfected or infected with WT virus (MOI of 5) and either left unirradiated or irradiated with 60 J/m² UV at 4 hpi. (A) Samples were collected at 4.5, 7, 10, and 18 hpi and processed for PFGE analysis. DNA transferred to nitrocellulose filters was probed with either ³²P-labeled viral DNA (left) or anti-CPD antibody (right) (n = 4, a representative image is shown). (B) Data were quantified, normalized, and plotted as the percentage of the highest value for each curve (average of n = 4). (C) BSC40 cells were infected with WT virus (MOI of 5) and left unirradiated or irradiated with 60 J/m² UV at 4 hpi. Samples were collected at 4.5, 10, and 18 hpi and processed for PFGE analysis. Southern blots were performed with ³²P-labeled viral DNA, and the signal for monomeric viral genomes is plotted. Error bars represent the SEM.

Does the cellular NER machinery modulate the impact of UV irradiation on vaccinia virus infection?

As described above, lesions induced by UV irradiation of the infected cells at 4 hpi were present within monomeric genomes harvested from 4.5 until 18 hpi. These lesions are therefore long-lived. A small decrease in their levels was seen at late times postinfection, accompanied by a small increase in the levels of the viral genome. The cellular genome encodes a suite of proteins that together comprise the nucleotide excision repair (NER) pathway whose primary function is to respond to UV-induced DNA adducts. NER proteins are found in both the cytoplasm and nucleus in the absence of DNA damage and at various stages of the cell cycle; we therefore probed their importance to the viral life cycle in UV-irradiated cells by generating stable cell lines depleted of key NER proteins (42, 43) using lentivirus-mediated shRNA delivery. We depleted xeroderma pigmentosum complementation groups A, F, and G (XPA, -F, and -G) proteins from BSC40 cells; a control cell line expressing a scrambled shRNA (shSCRM) was generated in parallel. Efficient depletion of the XP proteins was achieved, with residual levels of these proteins being ∼5 to 10% of the levels seen in control cells expressing scrambled shRNAs (Fig. 7A). The resultant cell lines were infected and irradiated at 4 hpi (Fig. 7B, lanes 1 and 2). No hypersensitivity to UV as measured by DNA accumulation was seen upon depletion of any of the three XP proteins (column 2), and the <2-fold changes observed for viral yields (column 1) were not biologically significant. These data are consistent with the interpretation that the cellular NER proteins do not participate in the repair of UV-induced lesions within the viral genome, as suggested by earlier reports monitoring viral replication in cells derived from XP patients (44).

FIG 7.

Depletion of XP-A, -F, or -G does not render infections hypersensitive to UV irradiation. (A) Generation of BSC40 cells with stable depletion of XP proteins. BSC40 cells were stably depleted of either XP-A, -F, or -G by lentiviral delivery of shRNA and selection with Puro; immunoblot analysis confirmed knockdowns of 85 to 95% (n = 3, a representative image is shown). Multiple shRNA sequences were tested and the sequence that gave the greatest depletion (underlined) was used for experimentation. (B) Depletion of XP-A, -F, and -G do not lead to UV hypersensitivity of viral infections. Transduced BSC40 cells were infected with WT virus (MOI of 5) and left unirradiated or irradiated with 60 J/m² UV at 4 hpi. At 10 hpi, samples were collected and processed for Southern dot blot analysis (B, column 2). At 18 hpi, samples were collected and titrated on BSC40 cells to quantify viral infectious yield (B, column 1) (n = 3; *, P < 0.05).

Is the DNA ligase, A50, involved in viral DNA metabolism following UV irradiation?

Having found that cellular NER factors play no role in viral infections following UV irradiation, we next sought to determine whether VV might encode a mechanism to facilitate viral genome metabolism following UV irradiation. Vaccinia infections performed with a viral recombinant lacking the A50 gene, which encodes the viral DNA ligase, have been reported to be hypersensitive to UV irradiation at 6 hpi (45, 46). Using our system, we found that vΔA50 infections were hypersensitive to UV irradiation at 4 hpi (Fig. 8A, compare lanes 3 and 6), but not at −1 hpi (Fig. 8A, compare lanes 2 and 5) as assessed by viral yield (18 hpi) (Fig. 8A, compare lanes 1 to 3 to lanes 4 to 6). This hypersensitivity was not correlated with any further reduction in late protein levels (Fig. 8B).

FIG 8.

Viruses lacking the A50 DNA ligase gene are hypersensitive to UV irradiation. (A) Viruses lacking the A50 gene are hypersensitive to UV irradiation at 4 hpi. BSC40 cells were left unirradiated (lanes 1 and 4) or irradiated with 60 J/m² UV at 1 h prior to infection (lanes 2 and 5) or 4 hpi (lanes 3 and 6). Infections were performed with either WT virus (lanes 1 to 3) or a virus lacking A50 (vΔA50, lanes 4 to 6). Samples were collected at 18 hpi and analyzed by plaque assay (A, n = 3, average and SEM plotted; *, P < 0.05). or immunoblot (B, n = 3, representative immunoblot shown). (C and D) Viral DNA accumulation after UV irradiation is halted when viruses lack the A50 gene. BSC40 cells were left uninfected (lane 1), infected with WT virus (lanes 2, 3, and 5), or vΔA50 (lanes 4 and 6) virus. At 4 hpi, infected cultures were irradiated with 60 J/m² UV (lanes 5 and 6). At 18 hpi, samples were collected, resolved by PFGE, stained with EtBr (8C, left), and analyzed by Southern blotting using a ³²P-labeled viral DNA probe (panel C, right). (D) Under the same conditions described for panel C, cultures were collected at 10, 14, and 18 hpi before analyzing via Southern dot blotting. Data were quantified and plotted as the average of three biological replicates (error bars = the SEM; *, P < 0.05; **, P < 0.01).

We also compared the levels of DNA that accumulated in wild-type (WT) and vΔA50 infections in the absence or presence of UV by PFGE analysis (Fig. 8C). The levels of DNA seen during vΔA50 infections were lower than that seen in WT infections, and the levels of DNA seen during irradiated vΔA50 infections were barely detectable. We also quantified the levels of DNA at 10, 14, and 18 hpi by dot blot analysis; the data are plotted as the percentage of DNA seen in irradiated infections versus those seen in unirradiated infections (Fig. 8D). For WT infections irradiated at 4 hpi, the levels of DNA observed at these three time points were 18, 34, and 48% of those seen in the absence of irradiation. For vΔA50 infections, the levels remained at 12% of that seen in unirradiated infections at all time points. These data suggest that the loss of A50 has a negative impact on the metabolism of viral DNA after infection, either by promoting degradation of the DNA or preventing new synthesis of the DNA.

Does the viral DNA polymerase mediate de novo DNA synthesis after UV irradiation?

Having seen a modest (2-fold) increase in the levels of viral DNA at late times after irradiation (10 to 18 hpi, UV irradiation at 4 hpi) (Fig. 8), we examined whether this increase required de novo DNA synthesis. Infected cells were UV irradiated at 4 hpi, and then infections were allowed to continue in the presence or absence of cytosine arabinoside (araC). araC is a competitive inhibitor of dCTP, and once incorporated creates a primer terminus that is poorly extended by the polymerase. Moreover, removal of the incorporated araNMP moiety by the proofreading exonuclease is extremely inefficient. Hence, araC is effectively a chain terminator and is a potent inhibitor of the E9 polymerase (see Fig. 8C) (47). Samples were harvested from 4.5 to 18 hpi and DNA was examined by PFGE. The inclusion of araC abrogated the 2-fold increase in viral DNA seen at late times postinfection. In fact, a small decrease in the levels of monomeric genomes was seen (not shown). These data do not allow us to say whether this difference reflects increased DNA synthesis or decreased DNA degradation.

To assess whether the inhibition of the viral DNA polymerase after irradiation had an impact on viral yield, cells were left unirradiated or irradiated at −1 hpi (Fig. 9A) or 4 hpi (Fig. 9B), and araC was either omitted, added at the start of infection (following adsorption) or added at 4 hpi. Cells were harvested at 18 hpi for analysis. Adding araC following adsorption blocked DNA synthesis (data not shown), the transition to late gene expression (panels A and B, bottom, lanes and columns 2), and decreased viral yield by >2,500-fold (panels A and B, top). Adding araC at 4 hpi reduced the viral yield 7-fold, whereas late protein levels were unchanged (Fig. 9A and B, lane 3). This finding further supports the conclusion that the levels of viral DNA that have accumulated at 4 hpi are sufficient for intermediate and late gene expression. Consistent with previous results, 60 J/m² of UV irradiation at −1 hpi reduced viral yield ∼14-fold, and the addition of araC (at 4 hpi) to these irradiated cells led to a further 2.6-fold difference in viral yield (Fig. 9A, top). Also consistent with previous results, irradiation at 4 hpi reduced viral yield ∼31-fold (Fig. 9B, top). Unexpectedly, addition of araC immediately after irradiation (4 hpi) led to a further 10.6-fold decrease in the levels of DNA accumulated at 18 hpi (not shown) and a further 11-fold decrease in viral yield (Fig. 9B, top). No impact was seen on the levels of late proteins. Therefore, blocking the function of the viral polymerase at the time of irradiation exacerbates the impact of UV-damage, strongly suggesting that the E9 protein is actively metabolizing the viral genome following UV irradiation in a manner that has a major impact on the levels of viral DNA and infectious virus.

FIG 9.

Inhibition of the viral DNA polymerase at the time of UV irradiation leads to a significant further reduction in the production of infectious virus. Inhibition of the E9 polymerase by araC greatly increases the sensitivity of VV infection when UV irradiation is performed at 4 hpi, but not at −1 hpi. (A) Irradiation prior to infection: BSC40 cells were left unirradiated (lanes 1, 2, and 3) or irradiated with 60 J/m2 UV at −1 hpi (lanes 4 and 5). Cells were then infected with WT virus (MOI of 5). araC (20 μM) was added to infections 30 min (lane 2) or 4 h (lanes 3 and 5) after infection. At 18 h, samples were collected and analyzed by plaque assay (top) or immunoblot (bottom). (B) Irradiation during infection: the same experiment as described above was performed, but UV irradiation was performed at 4 hpi (n = 3, average and SEM plotted [A]; *, P < 0.05; **, P < 0.01; ***, P < 0.001, or a representative immunoblot shown [B]).

Analysis of the yield, composition, and infectivity of mature virions produced during control and postirradiated infections with or without araC.

In Fig. 9, we demonstrated that inhibition of the E9 DNA polymerase (with araC) at the time of UV irradiation exacerbated the impact on infectious viral yield by 10.6-fold. To further understand why E9’s activity is important for viral infections following UV irradiation, we purified virions produced during unirradiated, irradiated (4 hpi), or irradiated and araC treated (both at 4 hpi) infections. Consistent with our previous results, UV irradiated infections produced ∼3-fold fewer virions compared to unirradiated infections (Fig. 10A, compare lanes 1 and 2). There was a further modest decrease in virions produced when UV irradiated infections were treated with araC at 4 hpi (Fig. 10A, compare lanes 2 and 3). Again, the position of the light scattering band is comparable between each of the samples, so there does not appear to be any major defect in genome encapsidation.

FIG 10.

Inhibition of the E9 polymerase’s activity after UV irradiation leads to the production of virions that have a lower infectivity and contain genomes with a higher level of CPD lesions. UV irradiation at 4 hpi (with or without the addition of araC) at the time of irradiation reduces mature virion production. BSC40 cells were infected with WT virus (MOI of 5) and left unirradiated (sample 1) or irradiated at 4 hpi (samples 2 and 3); araC (20 μM) was added to sample 3 immediately thereafter. Samples were collected at 18 hpi and virions were purified; light scattering bands were photographed (A) before fractions were collected for analysis by immunoblot (B), plaque assay (C), and PFGE (D). (B) Virions purified from cells that were UV irradiated and immediately treated with araC contain lower levels of some core proteins (L4 and I6). Blots were probed for L4, F17, I6, and A30. (C) The infectivity of virions produced following UV irradiation and araC treatment is greatly reduced. Fractions 8 to 11 were pooled and titrated on BSC40 cells. (D) The genomes encapsidated within virions produced following UV irradiation and araC treatment contain an elevated level of CPD lesions. Fractions 8 to 11 were pooled, resolved by PFGE and analyzed by EtBr staining and hybridization with a ³²P-labeled viral DNA fragment or by probing with an anti-CPD antibody. (The upper portion of the blot in the anti-CPD panel represents monomeric viral genomes and was exposed for a longer time than the lower level of the blot, which corresponds to fragmented cellular DNA [n = 1, t = 3, average and SEM plotted; **, P < 0.01; ***, P < 0.001].)

Fractions collected from the sucrose gradients were then assessed for viral protein levels, infectivity, and DNA content. We expanded the repertoire of proteins analyzed to further characterize the content of virions produced under each of the conditions. While still using L4 and F17, we also used I6, A30, F10, H5, and I3. Again, the profile of protein content from unirradiated and irradiated virions is consistent with our previous results and the fold change seen in the light scattering band (Fig. 10B, compare lanes 1 and 2). However, the profile seen in the UV-irradiated, araC-treated virions is distinct in that there are further reductions in the levels of I6, L4, F17, and A30 compared to the UV-irradiated-alone condition (Fig. 10B).

Fractions 8 to 11 were pooled from each of the gradients and plaque assays were performed to assess the yield of infectious virus from the sucrose-banded material (Fig. 10C). Irradiation reduced the yield of infectious virions by 35-fold, and irradiation + araC treatment led to a 700-fold decrease in infectious yield. These data are within 2-fold of the changes we saw in the infectious viral yield of one-step infections under the same conditions (Fig. 9).

Detection of UV induced lesions in the encapsidated viral genomes of araC-treated UV-irradiated infections.

As shown above, we were able to detect CPD lesions in monomeric viral genomes following PFGE analysis of irradiated, infected cells (Fig. 5E). Using a similar approach, we analyzed the virions that we had purified on sucrose gradients to determine whether lesion-containing genomes were encapsidated. Mature virions from the peak fractions (fractions 8 to 11) of our sucrose gradients were sedimented and the viral DNA encapsidated therein was subjected to PFGE analysis (Fig. 10D). One replicate was stained with EtBr and then probed with ³²P-labeled viral DNA, and the other was probed with the anti-CPD antibody. The EtBr image indicates that we have an abundance of monomeric genomes within the virions harvested from control infections (Fig. 10D, left, lane 1) but fewer within the virions purified from cells irradiated at 4 hpi or UV-irradiated and araC treated at 4 hpi (Fig. 10D, left, lanes 2 and 3). The Southern blot (Fig. 10D, middle lanes) showed comparable results.

We next assayed the presence and intensity of UV-induced CPD lesions with the encapsidated genomes. As expected, no signal was seen in the virions purified from unirradiated infections (Fig. 10D, right, lane 1); however, CPDs were detected in the virions purified from irradiated infections and, strikingly, even more were detected in the virions purified with UV-irradiated infections that were treated with araC at the time of irradiation (compare lanes 2 and 3). These data confirm that lesion-containing genomes can and are encapsidated within mature virions, and are also consistent with the E9 polymerase playing a role in the removal or repair of CPD lesions in genomes subjected to UV irradiation.

Based on the data in Fig. 5, we expected to see CPD-containing cellular DNA in the pooled fractions retrieved from the gradients loaded with samples from irradiated infections; indeed, a smear of ∼48 kb was observed (Fig. 10E, right, lanes 2 and 3). The cellular origin of this material was confirmed by the absence of recognition by the viral DNA probe (Fig. 10A, middle panel). The CPD signal was ∼2.5-fold more intense in the gradient representing the UV irradiation + araC (Fig. 10D, compare lanes 3 and 2); a stronger smear of ∼48-kb cellular DNA was also observed in the EtBr image (Fig. 10D, left); this difference probably reflects the fact that a larger volume of sample was loaded for lane 3 versus lane 2.

Cumulatively, the data shown in Fig. 8 to 10 support the conclusion that the A50 ligase and the E9 polymerase contribute to the metabolism or repair of the viral genome following UV irradiation; future studies will be required to identify other important viral factors.

DISCUSSION

UV irradiation has been used experimentally to inactivate purified vaccinia virus, introducing lesions into the genome that compromise the progression of the viral life cycle in the subsequent round of infection (44, 48). Here, we examine the impact of irradiating cells prior to infection, as well as during infection, on the progression of the viral life cycle. By using both approaches, we have uncovered the impact of UV irradiation on the host cell separately from examining the impact on the viral genome itself.

Because of vaccinia virus’ unique autonomy from the nucleus and the host genetic material, we would not anticipate that perturbing cellular DNA replication or transcription would have a significant impact on the viral life cycle. However, UV irradiation does cause broader effects due to the activation of various cellular stress responses. Irradiation of cells (60 J/m²) prior to infection with vaccinia virus (−1 hpi) reduces protein synthesis by approximately 3- to 4-fold and causes a comparable reduction in the levels of intermediate and late viral proteins that accumulate after infection (Fig. 3). UV irradiation has been shown to lead to an inhibition of protein synthesis due to the GCN2-mediated phosphorylation of the initiation factor eIF2α (33, 34), and our data indicate that the levels p-eIF2α increase after irradiation of either uninfected or infected cells (Fig. 3E). This decrease may be sufficient to explain the reduction in the number of virions produced in cells irradiated at −1 hpi, although there also appears to be a minor diminution in genome encapsidation that also contributes to the ∼12-fold reduction in the yield of infectious virus (Fig. 5).

Administration of the same dose of UV during infection (4 hpi) at a time when early gene expression is complete, DNA replication is in the early part of the linear phase, and the intermediate and late phases of gene expression have commenced leads to an ∼35-fold decrease in infectious yield (Fig. 2). Here, not only are the host cells being irradiated, but the viral genome is being irradiated: UV irradiation has immediate consequences for DNA, introducing both CPD and 6,4-PP lesions that distort the DNA helix and can serve as roadblocks to transcription and DNA replication (Fig. 1). Eukaryotic cells encode dedicated DNA polymerases (lambda and beta) that perform error-prone bypass synthesis across these bulky adducts (36, 49), but most replicative DNA polymerases encoded by cells and DNA viruses do not perform such error-prone synthesis. The data shown here indicate that when the intracellular viral genomes are irradiated at 4 hpi, little if any further DNA accumulation occurs between 4 and 10 hpi (Fig. 1). The levels of DNA at 10 hpi are ∼25% of that seen in a WT infection (Fig. 6). The E9 polymerase therefore seems unable to perform efficient bypass synthesis. Of note, adding araC to block further DNA synthesis at 4 hpi reduces the yield of infectious virus only 7-fold (Fig. 9). The 35-fold reduction in viral yield seen with UV irradiation at this time point therefore reflects not only the reduction in DNA levels but also the reduction in late proteins/virions (∼3-fold) and the presence of lesions within the DNA. Indeed, lesions persisted and were refractory to repair until 10 hpi. Furthermore, the lesions present in the genomes were encapsidated into nascent mature virions (Fig. 5).

Extrapolating from previous studies on the DHFR gene (50), in which irradiation with 10 J/m² introduced 2 to 2.5 lesions per 30-kb DNA, we estimate that, by irradiating with 60 J/m², we introduced ∼88 CPD and ∼19 6,4-PP per 195-kb vaccinia genome. However, the data shown in Fig. 1B reveal that the impact on DNA accumulation is the same whether we irradiate infected cells with 15, 30, 45, or 60 J/m². At 15 J/m² UV, we would be introducing ∼22 CPDs and ∼5 6,4-PPs, one-fourth that described above for 60 J/m². These data suggest that, even at 15 J/m², we have saturated the ability of the DNA replication machinery to proceed.

In contrast, the dose-response curve in Fig. 2B shows that whereas irradiation with 15 J/m² UV, (∼22 CPDs and ∼5 6,4-PPs) reduces viral yield ∼10-fold, an ∼35-fold decrease in viral yield is observed after irradiation with 60 J/m² UV (∼88 CPD and ∼19 6,4-PP). These data suggest that the increasing doses of UV are affecting cellular processes that are important for the viral life cycle. The dose-response curve shown in Fig. 3A—which represents the impact of irradiating cells prior to infection—confirms this interpretation.

We did observe a modest decrease in the levels of lesions within monomeric genomes and a modest increase in the levels of DNA between 10 and 18 hpi (Fig. 6). How might this occur? Although the viral life cycle is limited to the cytoplasm, we considered that the cellular NER machinery responsible for NER might relocate to the cytoplasm and act upon the viral genome. We therefore generated cell lines that were depleted of several NER proteins: infections performed in these cell lines exhibited no UV hypersensitivity (Fig. 7). The NER machinery appears to play no role in the viral response to irradiation.

Some chordopoxviruses such as leporipoxvirus, avipoxviruses, and molluscipoxviruses (51, 52) encode CPD-photolyases, which cleave the covalent bonds formed between adjacent pyrimidine bases, thereby resolving damaged DNA. The fowlpox CPD-lyase gene is essential for infections in the presence of UV irradiation (52), suggesting that it plays a crucial role in repair of the viral genome. Interestingly, orthopoxviruses, such as vaccinia virus, do not encode such a CPD-photolyase gene.

In the absence of a role for the cellular NER machinery and the absence of a viral CPD photolyase, is any repair of the UV-induced lesions occurring? Infection is accompanied by robust homologous recombination, and it is possible that recombination might yield some genomes with reduced number of lesions. It also seemed possible that the core replication machinery might provide some meaningful, albeit incomplete, form of repair. Loss of the virally encoded DNA ligase (A50) has previously been associated with sensitivity to UV irradiation (45, 46), and we therefore compared WT virus and ligase-deficient virus in cells left untreated or UV-irradiated (60 J/m²) at 4 hpi: loss of the A50 gene did cause UV hypersensitivity, as assessed by the levels of DNA that accumulated and the yield of infectious virus (Fig. 8). This result is particularly noteworthy given the fact that the A50 gene is nonessential in these cell lines during control infections. These results prompted us to reexamine a role for the E9 DNA polymerase in the virus’ response to UV irradiation—a role which we had initially dismissed because DNA accumulation stalled after UV irradiation. We therefore compared infections performed under WT conditions, the addition of araC at 4 hpi (to inhibit further DNA synthesis), UV irradiation at 4 hpi, and UV irradiation + araC at 4 hpi. Unexpectedly, we found that that the addition of araC at the time of irradiation led to a further 11-fold decrease in the yield of infectious virus production (Fig. 9). Clearly, the polymerase performs a vital function that reduces the impact of irradiation. Virions were purified from these infections by sucrose banding, and the genomes contained therein were resolved by PFGE. As expected, the yield of virions was reduced from infections subjected to UV irradiation or UV irradiation + araC treatment, and the levels of monomeric genomes were therefore lower, as assessed by Southern blotting. Most importantly, the levels of CPD lesions within the encapsidated genomes were significantly higher in the virions purified from cells that were not only irradiated at 4 hpi but treated with araC at the same time (Fig. 10). These data provide solid evidence that the E9 polymerase is performing a repair function that removes a significant number of CPD lesions from irradiated genomes.

At this time, we do not know how the polymerase performs this function. Is it working with a variety of other viral proteins to mimic the NER pathway used by cells? The viral G5 protein is a structure-specific endonuclease and exonuclease (M. W. Czarnecki and P. Traktman, unpublished data), and the viral A18 protein has a helicase activity not unlike that of the NER protein XPB (53, 54). The viral A22 protein, a Holliday junction resolvase, has an endonuclease activity that works on DNA bulges similar to those caused by UV-induced lesions (55). Alternatively, could E9 be using its 3′–5′-exonuclease activity to remove CPD lesions? These possibilities should be the focus of future work.

MATERIALS AND METHODS

Reagents.

Lipofectamine 2000 and Click-iT Plus EdU Alexa Fluor 488 were purchased from Life Technologies (Grand Island, NY). Paraformaldehyde (PFA) was purchased from EM Sciences (Hartfield, PA). Zeta-Probe blotting membranes were purchased from Bio-Rad (Hercules, CA). Protran nitrocellulose membranes were purchased from GE Healthcare Life Sciences (Boston, MA). [³⁵S]methionine and [³²P]dNTP were purchased from Perkin-Elmer Life Sciences (Boston, MA). SeaKem LE Agarose was purchased from Lonza, Inc. (Allendale, NJ). Molecular weight standards for PFGE (Lambda PFGE marker) were purchased from New England BioLabs, Inc. (Ipswich, MA). Proteinase K and araC were purchased from Sigma-Aldrich (St. Louis, MO). Pancreatic RNase was purchased from Roche Diagnostics (Indianapolis, IN). SybrGold was purchased from Invitrogen (Waltham, MA).

Antibodies.

Antibodies to VV proteins: I3, L4, F17, A20, I6, F10, H5, and A30 were used as previously described (18, 27, 56–59; unpublished data). Antibodies to cellular proteins: XPF (sc-398032; Santa Cruz Biotechnology, Inc.), pChk1 S317 (D12H3; Cell Signaling Technology), RPA32 (ab2175; Abcam), γH2AX (2577S; Cell Signaling Technology) Beta-tubulin (ab6046; Abcam), tChk1 (2360; Cell Signaling Technology), XPA (sc28353; Santa Cruz Biotechnology, Inc.), XPG (sc13563; Santa Cruz Biotechnology, Inc.), ATR (sc515173; Santa Cruz Biotechnology, Inc.), eIF2α (9722; Cell Signaling Technology), Calnexin (ADI-SPA-860-F; Enzo Life Sciences), 6,4-PP (NMDND002; Cosmo Bio Co., Ltd.), and CPD (NMDND001; Cosmo Bio Co., Ltd.). Alexa Fluor 488 goat anti-rabbit (GAR) IgG and Alexa Fluor 594 goat anti-mouse (GAM) IgG were purchased from Life Technologies (Grand Island, NY).

Cells and virus.

African green monkey BSC40 cells were maintained at 37°C in Dulbecco modified Eagle medium (DMEM; Life Technologies) containing 5% fetal bovine serum (FBS; Gibco). HEK293 T cells were maintained at 37°C in DMEM containing 10% FBS. WR strain of VV (WT) stocks were prepared by ultracentrifugation of cytoplasmic lysates of infected BSC40 cells through 36% sucrose; titers were assessed by plaque assays performed on BSC40 cells. A modified virus, vSK20 (labeled here as vΔA50), was kindly provided by Geoffrey Smith (45).

Depletion of cellular proteins using shRNA lentiviral transductions.

Plasmids (pLKO.1) encoding shRNA sequences targeting our genes of interest were ordered from the MUSC shRNA technology core within the Hollings Cancer Center. The ‘sense’ strands of the shRNA sequences are listed in Table 1. Lentivirus stocks were prepared in HEK-293 T cells using the pLL3.7 LentiLox system (Addgene) as previously published by our lab (60). BSC40 cells were transduced with the indicated stock and cells carrying the incorporated transgene were selected for by inclusion of puromycin (97.5 μg/mL) in the culture medium. Cells were used for experimental analysis as long as immunoblot analysis confirmed target protein depletion, which lasted approximately 3 to 4 weeks in culture under selection.

TABLE 1.

‘Sense’ strands of shRNA sequences used

Code (TRC Number)	Sequence
XPA 1 (N0000083194)	5′-CCGGGCATTAGAAGAAGCAAAGGAACTCGAGTTCCTTTGCTTCTTCTAATGCTTTTTG-3′
XPA 2 (N0000083196)	5′-CCGGCATGAGTATGGACCAGAAGAACTCGAGTTCTTCTGGTCCATACTCATGTTTTTG-3′
XPG 78 (N0000050778)	5′-CCGGCCAGCGAAATAGAAGCAGTTTCTCGAGAAACTGCTTCTATTTCGCTGGTTTTTG-3′
XPG 79 (N0000050779)	5′-CCGGCCTCCTTTACAAGAGGAAGAACTCGAGTTCTTCCTCTTGTAAAGGAGGTTTTTG-3′
XPG 80 (N0000050780)	5′-CCGGGCTTTCAGATTCTAAACGAAACTCGAGTTTCGTTTAGAATCTGAAAGCTTTTTG-3′
XPG 81 (N0000050781)	5′-CCGGCCTGTATTAAAGCAACTCGATCTCGAGATCGAGTTGCTTTAATACAGGTTTTTG-3′
XPG 82 (N0000050782)	5′-CCGGCCAATGGAAATTGACTCGGAACTCGAGTTCCGAGTCAATTTCCATTGGTTTTTG-3′
XPF 3 (N0000078583)	5′-CCGGGCGCAAGAGTATCAGTGATTTCTCGAGAAATCACTGATACTCTTGCGCTTTTTG-3′
XPF 7 (N0000078587)	5′-CCGGCCAAGATACGTGGTTCTTTATCTCGAGATAAAGAACCACGTATCTTGGTTTTTG-3′

UV irradiation of BSC40 cells.

Confluent dishes of BSC40 cells were irradiated using a UV cross-linker (VWR, catalog no. B116536); the medium was aspirated immediately prior to irradiation and replaced immediately after irradiation. Doses of 0, 15, 30, 60, and 180 J/m² were used. As indicated in the text, cells were either irradiated at 1 h prior to infection (−1 hpi) or at 4 h after infection (4 hpi) (or left untreated).

Quantification of infectious virus yield.

BSC40 cells were infected with WT virus (MOI of 5) for 18 h; harvested cells were washed once with PBS, and cell pellets were resuspended in 1 mM Tris (pH 9.0). Samples were then disrupted by freeze-thawing (3×), and the viral yield was quantified in plaque assays performed on BSC40 cells.

Quantification of viral DNA accumulation by Southern dot blot hybridization.

Cells were harvested at various time points from 4.5 to 18 hpi, as described in the text. Pellets were resuspended in 10× SSC (1.5 M NaCl, 0.15 M sodium citrate), 1 M ammonium acetate, and disrupted via freeze-thawing (3×) and sonication (2×, 15 s each). Using a dot blot apparatus (Bio-Rad) loaded with a Zeta-Probe membrane, samples were spotted in technical triplicate, denatured using denaturing buffer (1.5 M NaCl, 0.5 M NaOH; 1×, 10 min) and washed with 10× SSC (2×, 5 min). Membranes were probed with ³²P-labeled fragments of the viral genome, as described previously (17). Data were obtained using phosphorimager analysis (Typhoon FLA-9000) and quantified using ImageQuantTL software.

Pulsed-field gel electrophoresis.

BSC40 monolayers were infected and/or irradiated as described in the text. Harvested cell pellets were suspended in plugs of phosphate-buffered saline (PBS)/0.5% agarose and stored overnight at 4°C. The following day, samples were digested using proteinase K (50 μg/mL) for 24 h at 50°C. Plugs were loaded into the wells of a 1% Seakem Gold Agarose Tris borate EDTA (TBE) gel and resolved using a CHEF Mapper XA apparatus (Bio-Rad): 6 V/cm for 12 h at 14°C with a switching time gradient of 1 to 25 s, a linear ramping factor, and a 120° angle. EtBr or Sybr Gold were used to visualize DNA; images were captured using a FluorChem E documentation system (ProteinSimple, Santa Clara, CA). Gels were then transferred to Zeta-Probe GT or nitrocellulose membranes and either hybridized to ³²P-labeled DNA fragments of the viral genome or probed with an antibody that recognizes CPD DNA adducts. (When CPD adducts were being monitored, gels were not stained and imaged prior to transfer because such treatment induces DNA adducts).

Immunoblot analysis of viral and cellular protein levels.

Cell pellets were lysed on ice using 1× PLB (10 mM NaPO₄ [pH 7.4], 100 mM NaCl, 1% Triton X-100, 0.1% SDS, 0.5% DOC) containing both protease and phosphatase inhibitors (1 μg/mL leupeptin and pepstatin, 1 mM phenylmethylsulfonyl fluoride, 1 mM sodium orthovanadate and 1 mM NaF). After lysis, 5× protein sample buffer (PSB) was added, and samples were resolved electrophoretically and then transferred to nitrocellulose membranes. Depending on the optimized antibody blotting conditions, membranes were blocked with milk or bovine serum albumin (BSA). The membranes were then probed with antisera specific for viral or cellular proteins in buffers containing various concentrations of NaCl (500 mM NaCl–TBST, 325 mM NaCl–MTBST, 150 mM NaCl–FTBST), followed by horseradish peroxidase-conjugated secondary antibodies. Immunoblots were developed using chemiluminescent SuperSignal West Pico reagents (Pierce, Rockford, IL), visualized by exposure on a FluorChem E documentation system (ProteinSimple, Santa Clara, CA) and quantified using AlphaView software (ProteinSimple).

Metabolic labeling of nascent proteins with [³⁵S]methionine.

BSC40 cells were infected and/or irradiated as described in the text. At either 1 h after UV (uninfected) or 3, 6, 9, or 12 hpi, 100 μCi/mL [³⁵S]methionine was added to cultures in modified DMEM containing l-glutamine but not methionine for 45 min. Cultures were then harvested, lysed using 1× PLB, and resolved electrophoretically before being fixed in a 10% acetic acid–40% methanol solution. The gel was then dried between sheets of cellophane and visualized by phosphorimaging analysis (Typhon FLA-9000).

TCA precipitation of [³⁵S]methionine-labeled proteins.

Aliquots from the metabolic labeling described above were spotted onto Whatman paper, precipitated with 5% TCA, and washed with 95% ethanol, followed by 100% acetone before being loaded in scintillation vials to be counted using a scintillation counter (Beckman Coulter, catalog no. LS-6500).

Visualization of CPD lesions in PFGE-resolved DNA samples.

After resolution by PFGE, samples were transferred to Zeta-Probe membranes using a standard Southern blot transfer protocol. Membranes were blocked with 5% milk F-TBST for 1 h at room temperature before probing with antibodies recognizing either CPDs or 6,4-PPs. Membranes were probed overnight at 4°C at 1:4,000 (CPD) or 1:1,000 (6,4-PP) in 1% milk F-TBST. The following day, membranes were treated as described above.

Immunofluorescence analysis.

For localization of UV lesions, infected and/or irradiated BSC40 cells were fixed with 4% PFA, washed and permeabilized with 0.1% Triton X-100 in PBS (15 min, room temperature). Samples were probed with anti-I3 and anti-CPD or anti-6,4-PP antibodies, followed by Alexa Fluor GAR 488-nm and GAM 594-nm secondary antibodies. DAPI (4′,6′-diamidino-2-phenylindole) was added for 30 min, and slides were mounted with Vectashield (Vector Laboratories, Inc., Burlingame, CA). Images were captured using a Nikon Eclipse Ti microscope and NIS Elements AR4.4 software (Tokyo, Japan).

Electron microscopy.

Dishes (60 mm) of BSC40 cells were either left unirradiated or irradiated with 60 J/m² UV at 1 h prior to infection with WT virus (MOI of 5) or at 4 hpi with WT virus. Samples were harvested at 18 h and processed for transmission electron microscopy, as previously described (17). Cell pellets were embedded in Embed 810 resin (Electron Microscopy Sciences, Hatfield, PA). Thin sections were examined on a JEOL JEM-1010 microscope, and images were obtained using a Hamamatsu camera.

Analysis and purification of mature virions by sucrose gradient ultracentrifugation.

Dishes (15 cm) of BSC40 cells (four per sample) were left unirradiated or irradiated with 60 J/m² UV at −1 or 4 hpi. All infections were performed with WT virus (MOI of 5) for 18 h. Cells were harvested, washed, and resuspended in 10 mM Tris (pH 9.0). The cells were broken using a Dounce homogenizer, and nuclei were removed by sedimentation. The cytoplasmic fraction was sonicated and layered on top of a 36% sucrose cushion (10 mM Tris [pH 9.0]); ultracentrifugation was performed at 25,000 × g for 90 min at room temperature. Pellets were then resuspended in 200 μL of 1 mM Tris (pH 9.0), sonicated twice for 15 s each time, and then layered on top of 25 to 40% sucrose gradients prepared in 1 mM Tris (pH 9.0). Virions were resolved through the gradient by ultracentrifugation at 6,000 × g for 50 min at 4°C.

In some cases, light scattering bands were pulled using an 18G needle, removing ∼200 μL per sample. In other cases, the bottom of the tube was pierced and equal-sized fractions were collected. In both cases, a final sedimentation at 9,200 × g for 45 min at 4°C was used to collect the virions. Titers were determined by plaque assay on BSC40 cells, and the total protein concentration was quantified using a BCA protein assay kit (Thermo Fisher). In some cases, DNA analysis was performed by spotting samples on Zeta-Probe membranes, which were then probed with ³²P-labeled fragments or anti-CPD antibodies; in other cases, samples were subjected to PFGE prior to further analysis.

Preparation of digital figures.

Statistical analysis and graph preparation were performed using SigmaPlot software (Systat Software, Chicago, IL). Final figures were assembled and labeled with Canvas software (Deneba Systems, Miami, FL).