Background
Herpes simplex virus type 1 (HSV-1) has a double-stranded DNA genome of approximately 152 kilobasepairs. Like other herpes viruses, HSV-1 infection is characterized by two distinct cycles; a productive lytic infection and a nonproductive latent infection. During lytic infection, viral genes are expressed in a cascade of at least three classes, commonly designated the immediate-early (IE), early (E), and late (L) genes. Proteins encoded by IE genes facilitate expression of E and L genes. The E genes primarily encode the viral DNA replication machinery. L genes, which are sometimes separated into leaky late and true late genes, encode proteins involved in the structure and assembly of the virus particle. Viral DNA replication is required to express true late genes and enhances leaky late gene expression [
1,
2].
The interplay between viral and cellular processes during HSV replication is complex. The virus can subvert cellular mechanisms to enhance the infection process; in response, the cell has many antiviral mechanisms in place to protect its functions. Histone variant H2AX is one example of the complex nature of these virus-host interactions. H2AX constitutes about 10 % of the total H2A distributed through the chromatin of a typical cell. The particular functions of H2AX are still poorly understood, but H2AX is phosphorylated in response to DNA damage at sites up to several kilobasepairs around the site of damage [
3]. Disruption of the mouse
H2afx gene encoding H2AX results in genomic instability and hypersensitivity to radiation [
4,
5]. In HSV-1 infection, H2AX is phosphorylated during viral E gene expression, and the amount of phosphorylated H2AX (γH2AX) increases as the gene cascade continues [
6‐
8]. This post-translational modification could reflect host responses attempting to limit the infection process; it could be beneficial to the virus; or it could be a host response without meaningful consequences for viral infection.
Incoming linear HSV viral genomes inherently have free DNA ends that conceivably might initiate a cellular DNA damage response [
9]. But the mere delivery of viral DNA into the cell is likely insufficient to trigger H2AX phosphorylation, because that phosphorylation occurs well after viral entry [
6]. An alternative hypothesis predicts that the replication or recombination of the viral DNA bearing single-strand nicks and gaps will initiate a DNA damage response including H2AX phosphorylation. To date, the mechanisms of H2AX phosphorylation during HSV infection and the effects on viral replication remain incompletely defined.
H2AX is a direct substrate for phosphorylation by the host cell kinases ATM (ataxia telangiectasia mutated) and ATR (ataxia telangiectasia and Rad3-related), which along with DNA-PK are the central signaling proteins of the DNA damage response pathway. ATM and DNA-PK typically respond to double-strand breaks, whereas ATR responds to single-strand DNA and stalled replication forks [
10]. The potential roles of these protein kinases in HSV infection have been investigated [
7,
11‐
17]. Others have shown that the viral IE protein ICP0 induces proteasome-mediated degradation of the catalytic subunit of DNA-PK and that the loss of DNA-PK activity increases virus replication [
12,
18]. The kinase function of ATM is activated during viral DNA replication [
11,
14,
15], and reduced HSV-1 replication in ATM-deficient cell lines suggests that ATM is important for viral replication during lytic infection [
11]. Li et al. [
6] and Alekseev et al. [
19] also found that an inhibitor specific for ATM (KU-55933) resulted in a decrease of HSV-1 at low multiplicity of infection (MOI) in AD-293 and OKF9 cells, respectively. In contrast, Shirata et al. [
14] reported that knockdown of ATM had no effect on HSV-2 infection in 293 T cells. This difference in ATM dependence between HSV-1 and HSV-2 is curious. In addition, we do not yet know the trigger responsible for H2AX phosphorylation during HSV infection nor, more importantly, whether γH2AX plays an active role in production of HSV.
We report here that ATM activity (but not ATR activity) and the expression of viral proteins (including UL30, the viral DNA polymerase), but not viral DNA replication per se, are necessary for HSV-1-induced H2AX phosphorylation in human foreskin fibroblasts. Intriguingly, during infection of fibroblasts by HSV-2, H2AX phosphorylation does require viral DNA replication. However, reducing H2AX phosphorylation by chemical or siRNA inhibition of ATM did not significantly affect HSV-1 or HSV-2 DNA replication and virus production at high MOI, and had only a modest effect at lower MOI. These results differ from reports [
6,
19] which suggest that ATM performs an important role in HSV-1 infection of other cell lines. Collectively, these observations suggest that H2AX phosphorylation represents a cell-specific and virus-specific host response to HSV infection and that such phosphorylation has little impact on viral infection.
Methods
Cells and viruses
Vero cells were obtained from ATCC and telomerase-transformed human foreskin fibroblasts (HFFs) were provided by Wade Bresnahan [
20]. Vero and HFF cells were maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10 % fetal bovine serum (FBS). Vero-PolB3 cells [
21] a gift from Dr. Don Coen, are stably transformed with the viral UL30 gene and express the viral DNA polymerase from its native promoter. These cells were maintained in DMEM with 10 % FBS and 400 μg/ml G418. The primary human skin fibroblast cell lines GM01588A and GM02052F (ATM
−/− cell lines each with a point mutation resulting in early termination of ATM [
22]), GM18366D (expressing low levels of ATR due to a point mutation resulting in alternate splicing [
23]), and GM05757B (expressing wild-type ATM and ATR) were obtained from Coriell Institute (Camden, NJ). GM01588A, GM02052F, and GM05757B cells were maintained in DMEM plus 10 % FBS; GM18366D cells were maintained in DMEM plus 20 % FBS. HSV-1 virus strain KOS and HSV-2 virus strain G were propagated and titered in Vero cells. HSV-1 strains HP66 and ΔS1, both bearing deletions of the viral DNA polymerase gene UL30, were obtained from Dr. Coen and were grown and titered in Vero-PolB3 cells [
21].
Reagents
Antibodies against H2AX (ABCM-AB10475) and phosphorylated H2AX (γH2AX) (ABCM-AB4178) were purchased from Abcam. UL30 antibody was a gift from Dr. Nigel Stow [
24]. GAPDH and actin antibodies were obtained from Millipore (MAB374 and MAB1501). The siRNAs were purchased from QIAGEN, including control siRNAs (SI03650318 and SI04381048) and validated siRNAs directed against ATM (ATM1 = SI00299299 and ATM2 = SI00604730) or ATR (ATR1 = SI02660231 and ATR2 = SI02664347). A third siRNA directed against ATR, here designated ATR3, is identical in sequence to an shRNA used by Dr. Sandra Weller and colleagues [
13,
17]. Transfections were done with SiLentFect (BioRad). Kinase inhibitors KU-55933 [
25] and CGK733 [
26‐
29] were purchased from EMD (118500 and 118501). VE-821 [
30] was purchased from AdooQ Bioscience (A11605). Actinomycin D (A9415), cycloheximide (C7698), phosphonoacetic acid (PAA) (284270) and doxorubicin (D1515) were purchased from Sigma. The 2X PhosphoStop phosphatase inhibitor cocktail and 2X minicomplete protease inhibitor cocktail (04906845001 and 11836153001) were both purchased from Roche. TRIzol (15596–026) was purchased from Invitrogen. Reverse transcription reactions were performed with a High Capacity cDNA Reverse Transcription Kit (ABI). SYBR Green master mix (04673522001) was purchased from Roche. CellTiter-Glo viability kit was purchased from Promega (G7570).
Gene expression and quantitative PCR
Total RNA was purified from cells using TRIzol reagent or the Qiagen RNeasy kit and was reverse-transcribed with random primers and RNAseOut according to the manufacturer’s protocol. Gene expression was quantified by real-time quantitative PCR (qPCR) using primers specific for the selected viral or human genes. DNA samples were collected using Qiagen DNeasy Blood and Tissue Kit and were quantified by qPCR with primers specific for viral ICP4 or ICP0 genes or the human 18S ribosomal RNA gene. The qPCR was performed on an ABI 7500 RT-PCR system (Applied Biosystems) using SYBR Green master mix; relative DNA or RNA levels were analyzed by the 2–ΔΔCt method.
HSV infection
HSV-1 and HSV-2 infections in HFF cells were performed as follows. Cells were washed with DMEM, inoculated with the appropriate MOI, and incubated at 37 °C. At 1 h post-infection (h p.i.) the inoculum was aspirated and cells were washed with DMEM. Cells were incubated in DMEM containing 10 % serum at 37 °C and were harvested at various times. In kinase inhibitor experiments, KU-55933, VE-821, or CGK733 (or DMSO vehicle) was added to cell culture wells at 1 h p.i. in normal media. For inhibition of viral DNA replication, PAA was added at 1 h p.i. to the normal medium to a final concentration of 400 μg/ml. For infections using actinomycin D (1 μg/ml) or cycloheximide (100 μg/ml), the inhibitors were added to the cells for 1 h before infection and were present during the entire course of infection.
Infections for plaque assays were performed as described above, except that DMEM with 2 % FBS and 0.9 % SeaPlaque agarose was added to the infected cells at 1 h p.i. At 3 d p.i. the cells were stained with neutral red for 2 h at 37 °C to visualize plaques.
Western blot
Lysates analyzed by western blot were collected as follows. After the cell culture medium was aspirated, the cells were washed with phosphate-buffered saline (PBS) and were lysed directly in 2X SDS-PAGE loading buffer containing 2X PhosphoStop phophatase inhibitor cocktail and 2X minicomplete protease inhibitor cocktail with 5 % β-mercaptoethanol. The lysates were heated at 95 °F for 5 min and separated on 4-20 % polyacrylamide gels. The proteins were transferred to polyvinylidene fluoride membranes and blocked in 1X TTBS (100 mM Tris, pH 7.5, 150 mM NaCl, 0.1 % Tween 20) containing 2 % or 5 % (w/v) bovine serum albumin. The blocked membrane was incubated with primary antibody overnight at 4 °C at the dilution suggested by the supplier. The membrane was washed with TTBS and incubated with the appropriate secondary antibody conjugated with horseradish peroxidase at 5000-fold dilution for 1 h at room temperature. The membrane was again washed and the horseradish peroxidase was detected by chemiluminescence.
siRNA transfections
siRNA transfections of HFF cells were performed using the transfection reagent SilentFect according to the manufacturer’s protocol, with a final siRNA concentration of 16.7 nM. At 24 h post-transfection, the transfection mixture was aspirated and replaced with medium containing DMEM plus 10 % FBS. Cells were incubated at 37 °C for approximately 72 h post-transfection. Cells either were harvested for analysis of the knockdown efficiency or were infected with HSV using the infection protocol described above. Total RNA and DNA were collected at 16 h p.i. and the amount of viral RNA or DNA was measured by qPCR and normalized to host 18S DNA and negative control siRNA samples.
ATR-deficient and HFF cells were infected as detailed above. At 2 h p.i., cells were trypsinized, resuspended in 10 % FBS DMEM, and counted to ensure equal numbers of cells were loaded into each gradient. Gradients were prepared and centrifuged as described previously [
31], except that Optiprep media (Sigma) was used as the gradient material and dilutions were modified accordingly to obtain the correct percentage of iodixanol for each layer. Nuclear and cytoplasmic fractions were harvested and tested for purity by qPCR with GAPDH and 18S primers, as well as by immunoblotting with GAPDH and histone 3.3 antibodies. DNA in infected samples was quantified by qPCR with ICP0 and 18S rRNA primers.
Discussion
A number of viruses make use of the cellular DNA repair machinery to promote successful infection. One of the first proteins to be activated upon DNA damage is H2AX [
34]. Phosphorylated H2AX (γH2AX) spreads over chromatin flanking the DNA damage site [
35] and acts as a signal for recruiting other DNA damage response proteins [
36]. Others have noted that HSV-1infection is sufficient to induce γH2AX formation [
6,
7]. We too observed that HSV-1 infection induced γH2AX formation at 4 h p.i. and that the phosphorylation increased over time. We found that
de novo protein synthesis is required for H2AX phosphorylation but neither viral replication nor late viral protein expression is required for this phosphorylation event. Thus, induction of H2AX phosphorylation is not triggered merely by the presence of viral DNA in the infected cell nor by replication of viral DNA, but by expression of some viral or cellular protein following infection.
Curiously, although viral DNA replication is not required to induce γH2AX, expression of the viral DNA polymerase is required. Infection by two different viral mutants, HP66 and ΔS1 (each bearing substantial deletions of the UL30 open reading frame), failed to induce γH2AX. It is puzzling that expression of the viral DNA replication enzyme, but not DNA replication itself, is required for the formation of γH2AX during HSV-1 infection. We infer that H2AX phosphorylation is triggered by an event that occurs after UL30 expression but before HSV-1 DNA replication, perhaps during establishment of replication centers. This notion agrees with work by Wilkinson and Weller [
7,
15] showing that γH2AX accumulates around the viral replication compartments. Combined with the observation that ATM is solely required for H2AX phosphorylation, we speculate that the viral DNA polymerase could be involved in the recruitment of ATM to replication compartments, although further study would be needed to establish this point.
A second surprising observation was that H2AX phosphorylation during HSV-2 infection was blocked by the viral DNA replication inhibitor PAA, in stark contrast to HSV-1 infection. We presume that expression of UL30 is also required for H2AX phosphorylation during HSV-2 infection because viral DNA synthesis is required, but HSV-2 UL30 mutants are not readily available to test that presumption directly. We therefore infer the DDR pathway is activated differently in HSV-1 and HSV-2 infections; more work will be needed to better understand the mechanistic differences between these two closely related viruses.
Activation of ATM and at least some of its downstream targets by both HSV-1 and HSV-2 has been shown previously [
11,
14,
15,
37,
38]. Our results indicate that during HSV-1 or HSV-2 infection, H2AX was phosphorylated solely by ATM. Furthermore, inhibition of ATM has been shown to decrease HSV-1 viral DNA replication in certain contexts [
11,
19]. However, in our experiments, chemical inhibition of ATM and/or ATR or disruption of ATM and ATR expression by siRNAs had little or no effect on HSV-1 and HSV-2 DNA replication and virus production. Therefore, although ATM signaling is activated by HSV-1 infection, ATM is not vital for efficient HSV replication in HFF cells, and thus γH2AX formation is likely an incidental signal during infection. Since ATM is activated and γH2AX is formed, it seems reasonable to consider whether ATR is also activated. Others have shown that viral proteins prevent the phosphorylation of typical ATR targets during HSV-1 infection, and shRNA disruption of ATR alone has little effect on viral yield [
11,
13‐
17,
39]. In our hands as well, ATR activity does not contribute substantially to phosphorylation of H2AX during HSV infection, and ATR activity is not required for effective viral replication. Moreover, HSV-1 was insufficient to activate Chk-1 and thus downstream ATR activation (data not shown).
ATM and the γH2AX response have also been studied in the context of infection by other herpesviruses. Conflicting reports suggest that ATM is required for maximum yield of human cytomegalovirus and for maturation of replication compartments in HEL fibroblasts [
40], yet ATM
− cells had only a 1 log reduction in viral yield as compared to HFF cells [
41]. The literature is likewise divided over the role of ATM kinase activity during EBV infection; one report [
42] concludes that ATM is important for efficient EBV lytic replication in Akata BX1 latent cells but another report [
43] indicates that ATM is dispensable for EBV lytic replication in Tet-BZLF1/B95-8 cells. A comprehensive report of various stimuli of EBV in various cell lines indicates that strict dependence on ATM is stimulus-specific; ATM was vital for early steps in reactivation but not for viral DNA replication per se [
44]. Furthermore, murine herpesvirus 68 required ATM and γH2AX in primary mouse macrophages only at low MOI, but neither are required for growth in MEFs regardless of MOI [
45,
46]. Karposi sarcoma virus depends on ATM, as shown by reduced establishment and latency of KSHV in the absence of ATM [
47,
48]. However, VZV does not seem to depend similarly on ATM since it displayed near normal production in GM02530 ATM
− cells [
49]. The varying reports for CMV, EBV and HV68 together with the data in this report indicate that cell type is a key factor in virus reliance on DNA damage pathways during herpes virus infection.
In contrast with our results using chemical inhibitors and siRNAs targeting ATM and ATR, but consistent with reports from others [
11,
19], we observed a reduction of both viral DNA replication and viral yield upon infecting either ATM
−/− cell lines or an ATR-deficient cell line at 24 h. A longer timecourse in the ATM
− cells did reveal higher yield of virus at later times, with one line eventually reaching levels comparable to wildtype cells. Our observations support those of Lilley et al. [
11], who observed that the yields of HSV-1 from three ataxia telangiectasia lines were reduced by two orders of magnitude at 24 h p.i., although two lines did recover to near wt yield at 36 h p.i. In contrast, Yamamoto et al. [
49] showed that both VZV and HSV-2 produced near-normal titers throughout the entire 36 h of infection in GM02530 cells. This discrepancy is echoed by Zavala et al. [
50] in their investigation of disparate HCMV production in mutant cells. Although they observed that distinct strains of HCMV yielded different titers in individual ATM
− cell lines, they determined that ATM was not necessary for normal titers in fibroblasts. This aligns well with the data we present here for HSV-1. We surmise that the ATM
- and ATR-deficient cells may have developed mechanisms to compensate for the lack of DNA damage kinases that are not compatible with efficient HSV-1 replication. Clearly the two ATM
− cell lines tested here have adapted in different ways, since they differed in virus production at late times in infection despite having identical mutations in the ATM gene. It is conceivable that the non-homologous end joining pathway, known to be antiviral [
18], is more active in these cells leading to decreased or delayed replication. Additionally, fibroblasts and keratinocytes have different efficiencies in supporting replication of UV-damaged HSV, indicating that DDR pathway activation varies significantly between cell types [
51]. Inhibition of ATM, and perhaps even of ATR, may be more severe in oral keratinocytes than HFFs because their DNA damage pathways mobilize differently in response to HSV.
The HSV-1 genome is known to contain nicks and single-strand gaps of various sizes that affect infectivity [
52]. Moreover, complex replication intermediates have been observed indicative of recombination events [
53]. Additionally, the linear genome may circularize prior to lytic replication, though recent reports have differed on this point [
54,
55]. Given the role of γH2AX and ATM in homologous recombination, it is logical to hypothesize that these proteins participate in such recombination events during HSV-1 and HSV-2 infection. However, since ATM is dispensable for infection of HFFs, the viral DNA must recombine by means other than canonical ATM-dependent homologous recombination, perhaps through proteins that can complement ATM’s function. The additional disparity between HSV-1 and HSV-2 in γH2AX induction suggests differences in recombination mechanisms during infections by these two closely related viruses. Most DNA replication studies have been done using HSV-1 and inferred for HSV-2, but perhaps these viruses replicate more differently than previously supposed. It would be interesting to determine the circularization and recombination rate of both HSV-1 and HSV-2 in the absence of cellular DDR proteins, especially ATM.
Competing interests
The authors declare no competing interests.
Authors’ contributions
CB and XL designed and performed the experiments. CB wrote the paper. ST participated in study design and edited the manuscript. All authors read and approved the final manuscript.