Human Cytomegalovirus IE1-72 Activates Ataxia Telangiectasia Mutated Kinase and a p53/p21-Mediated Growth Arrest Response

Cell culture.

REF52 cells were cultured as described previously (11). Early-passage wild-type (WT), as well as genetically matched p53-deficient (p53^−/−) mouse embryo fibroblasts (MEFs), Mdm2⁻/⁻/p53^−/− MEFs, and SAOS-2 cells were generous gifts from Stephen Jones (University of Massachusetts Medical School, Worcester, MA). The p19^Arf−/− MEFs were provided by Charles Sherr (St. Jude's Medical Center, Memphis, TN). Early passage p21^+/+ and p21^−/− MEFs were provided by Tyler Jacks (Massachusetts Institute of Technology, Boston, MA). MEFs were cultured as described previously (11, 57). Human embryonic lung fibroblasts (HELs) were a generous gift from Eng-Shang Huang (University of North Carolina, Chapel Hill, NC). The SAOS-2 cells and HEL fibroblasts were cultured in Dulbecco modified Eagle medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin. Human dermal fibroblasts from an ataxia-telangiectasia patient (GM03395C; AT), as well as age-matched, normal dermal fibroblasts (GM00316B; WT), were obtained from the Coriell Institute for Medical Research (Camden, NJ) and cultured according to conditions recommended by the provider.

To induce growth arrest, cells were first seeded at a plating density between 2 × 10³ to 3 × 10³ cells/cm². The WT and p21^−/− MEFs were then serum starved by a wash with phosphate-buffered saline (PBS) and then culture in DMEM containing 0.25% serum for a minimum of 48 h. The conditions used were based on empirical studies that optimized serum concentration and culturing times to induce growth arrest in over 80% of each cell population (unpublished observations).

HCMV infections.

HEL fibroblasts were infected with HCMV AD169 or CR208 strain at a multiplicity of infection (MOI) of 5. AD169 was obtained from the American Type Culture Collection, CR208, an AD169-derived virus that is deleted for UL123 (25), is used with permission from Ed Mocarski (Stanford). Viral infections were performed in serum-free medium for 1 h at 37°C and 5% CO₂. The viral inoculum was then removed and replaced with DMEM containing 10% FBS and cultured as described above. Cells were harvested at 24 h postinfection (hpi), and lysates were generated as described previously (11).

Adenovirus vectors and infection.

The recombinant adenovirus encoding the HCMV IE gene product IE1-72 (AdIE1-72) and IE2-86 (AdIE2-86) have been described (77) and characterized previously (3, 11). An empty vector virus (38) or a recombinant adenovirus expressing β-galactosidase was used as a control (labeled as AdCon) for virus infection in the experiments (56, 57). An E2F1-encoding adenovirus, AdE2F1 (38), and a p53-encoding adenovirus, Adp53 (39), have been described previously. Viruses were grown and titers were determined in a human embryonic kidney cell line (i.e., cell line 293) and subsequently purified on cesium-chloride gradients (49). Virus titers were determined by immunohistochemical staining of the adenovirus type 2 hexon (11, 57). Infections were done as described previously (11, 58). Cells were infected with either AdIE1-72, AdIE2-86, or a control adenovirus at an MOI of 250 unless otherwise noted.

Immunoblot analysis.

Whole-cell extracts were prepared as described previously (11) with the following exception. To assess p21 protein levels, washed cells were scraped into 1 ml of radioimmunoprecipitation assay buffer (PBS, 0.1% NP-40, 1% sodium dodecyl sulfate, 0.5% sodium deoxycholate, sodium vanadate, phenylmethylsulfonyl fluoride, and aprotinin) and incubated on ice for 1 h. Samples were lightly sonicated for 15 s, and soluble proteins were collected by centrifugation for 10 min at 13,000 rpm in a microcentrifuge. Aliquots of cell extracts were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis, and the resolved proteins were transferred to nitrocellulose membranes by electroblotting. Detection of p53, phospho-Ser15 p53, phospho-Ser20 p53, p21, p19^Arf, and HCMV IE1-72 and IE2-86 proteins was performed with antibodies specific for total p53 (Ab-6 or Ab-7; Oncogene Research Products), phospho-Ser15 and phospho-Ser20 p53 (Cell Signaling Technology), p21 (C-19; Santa Cruz Biotechnology), p19^Arf (NB 200-106; Novus Biologics), and IE1-72 and IE2-86 (MAB8130 and MAB810, respectively; Chemicon International). ATM and phospho-Ser1981 ATM were detected with antibodies from Rockland Immunochemicals (Gilbertsville, PA). Blots were stripped and probed for actin levels as a control for sample loading.

p53 nuclear shuttling.

Heterokaryons were generated as described previously (79). Essentially, subconfluent cultures of SAOS-2 cells were coinfected with Adp53 and either AdIE1-72 or AdCon. After infection, cells were cultured in normal media at 37°C for 1.5 h. The SAOS-2 cells were replated with an equal number of Mdm2^−/−/p53^−/− MEFs onto glass coverslips in 35-mm plates and incubated overnight at 37°C. Cells were then washed twice with PBS and incubated in medium containing cycloheximide (50 μg/ml) for 20 min at 37°C. Cells were subsequently washed with PBS and then incubated for 2 min in 50% polyethylene glycol (MW 3350; Sigma) dissolved in DMEM to facilitate cell fusion. Cells were washed with PBS and incubated in medium containing cycloheximide for 1 h at 37°C. Cells were then fixed with 3.7% formaldehyde in PBS for 10 min at room temperature and permeabilized with cold 0.2% Triton X-100 in PBS for 10 min at 4°C. To detect heterokaryons and p53 localization, coverslips were blocked with 10% FBS in PBS for 15 min at room temperature prior to incubation with primary antibodies specific for p53 (Ab-1; Oncogene Research Products) and human Ku-86 (Santa Cruz Biotechnology). After incubation with the primary antibodies, the coverslips were washed three times with PBS then incubated with isotype-specific fluorescein isothiocyanate-conjugated and rhodamine red-X-conjugated antibodies to detect p53 and human Ku-86, respectively. Cells were washed three times with PBS and once with distilled water. Glass coverslips containing the cells were mounted onto glass slides with Vectashield containing DAPI (4′,6′-diamidino-2-phenylindole; Vector Laboratories). p53 nuclear shuttling was scored by assessing the localization of p53 in human-murine heterokaryons with a Zeiss immunofluorescence microscope. The results were presented as the percentage of heterokaryons in which p53 shuttled from human to murine nuclei.

Luciferase assay.

HEL fibroblasts were plated at subconfluent densities onto six-well plates and grown overnight at 37°C. Cells were cotransfected with a p21-specific reporter construct, WAF-1-Luc (generously provided by Karen Vousden, National Cancer Institute, Frederick, MD) (22) (1 μg/well) and either a plasmid encoding IE1-72 cDNA (pcDNA3-IE1-72) or a control plasmid (pcDNA3) (2 μg/well), along with a Renilla luciferase-expressing plasmid, pRL-TK (kindly provided by Zdenka Matijasevic, University of Massachusetts Medical School, Worcester, MA) (0.25 μg/well) as an internal control to normalize for transfection efficiency. Transfection reactions were done with Lipofectamine Plus (Invitrogen) and were incubated for 5 h in 37°C. Afterward, transfection medium was replaced with DMEM containing 10% FBS, and cells were incubated at 37°C for 24 h. The cells were then lysed, and the luciferase activity was determined by using a Dual-Luciferase reporter assay system (Promega). The results are presented as the fold induction of p21 promoter activity relative to the control plasmid-transfected cells.

Immunohistochemical staining for p21 protein accumulation.

Subconfluent cultures of REF52 cells were infected with the appropriate recombinant adenovirus constructs and immunohistochemical staining for p21 protein was performed. At 24 hpi, the cells were fixed for 5 min in formaldehyde (0.37% final concentration), followed by 5 min in methanol. Cells were blocked with 1% bovine serum albumin (BSA) in PBS-0.5% Tween 20 (PBS/Tw) and incubated with an anti-p21 MAb (Santa Cruz Biotechnology) diluted in 1% BSA in PBS/Tw for 1 h at room temperature. The cells were washed with PBS/Tw, and bound antibody was detected by using a Vectastain 3,3′-diaminobenzidine (DAB) substrate kit (Vector Laboratories) as described by the manufacturer.

Cell cycle analysis by flow cytometry.

Cells were infected with the appropriate recombinant adenoviruses and processed for flow cytometry as described previously (11). Flow cytometry data was processed by using the FlowJo FACS analysis program (Treestar). S-phase cells were defined as the population of cells with more than 2 N but less than 4 N DNA content.

HCMV IE1-72 expression induces p53 and p19^ARF protein.

We previously reported that IE1-72 expression induces quiescent cells to enter S phase in the absence of p53 (11). Because increased p53 levels typically correlate with function, we examined p53 levels in cells expressing IE1-72. We used REF-52 cells and MEFs for this experiment for comparison with the earlier study. As shown in Fig. 1A, the IE1-72-expressing cells exhibited a two- to threefold increase in the levels of p53 compared to that observed in the control virus-infected cells, confirming our earlier finding (11). Given that IE1-72 can function as a transcriptional coactivator, we looked for changes in p53 mRNA levels in cells transduced with an IE1-72-encoding recombinant adenovirus. We found that IE1-72 expression did not affect p53 mRNA levels, suggesting that IE1-72 does not transactivate p53 expression through its endogenous promoter (data not shown).

Another way that p53 levels are regulated is by the p19^Arf/Mdm2 pathway in which increased levels of p19^Arf stabilizes p53 through its ability to inactivate the E3 ubiquitin ligase activity of Mdm2 (30). We determined whether IE1-72 might alter p53 levels by inducing p19^Arf accumulation. As shown in Fig. 1B, approximately a threefold increase in the levels of p19^Arf was observed in cells expressing IE1-72. Given that increased levels of p19^Arf is sufficient to stabilize p53, it would appear that IE1-72 modulates p53 levels by increasing the levels of p19^Arf in cells.

We next measured p53 levels in p19^Arf−/− MEFs to determine whether p19^Arf is necessary for p53 accumulation following IE1-72 expression. As shown in Fig. 1C, low levels of p53 were detected in extracts from control virus-infected p19^Arf−/− MEFs. Reduced basal levels of p53 in early passage p19^Arf−/− MEFs have been observed previously (55). Unexpectedly, increased levels of p53 were observed in the IE1-72-expressing cells compared to the control virus-infected cells. These findings suggest that p19^Arf is not responsible for the most of the change in p53 levels associated with IE1-72 expression and raise the possibility that IE1-72 can act through an additional mechanism(s) to alter p53 levels.

IE1-72 and HCMV increase the levels of phospho-Ser15/18 p53.

In addition to the p19^Arf/Mdm2 pathway, modulation of p53 levels can occur through other mechanisms, including the covalent modification of p53. A well-studied model of this process is the upregulation of p53 after UV or ionizing radiation. DNA damage induces p53 phosphorylation at several residues, resulting in protein stabilization (54). Phosphorylation at the Ser18 residue on mouse p53 (Ser15 on human p53) is a common event in response to DNA damage and other stresses and leads to enhanced p53 stability by disrupting the binding of Mdm2 to p53 so that p53 is not ubiquitinated and degraded (64, 65). To examine the consequences of IE1-72 expression on the levels of phospho-Ser18 p53, we expressed IE1-72 in WT MEFs and measured p53 levels with a phospho-specific antibody. As shown in Fig. 2A, higher levels of phospho-Ser18 p53 were observed in WT MEFs expressing IE1-72 compared to extracts from the control cells. Increased levels of phospho-Ser18 p53 were also observed in the p19^Arf−/− MEFs after IE1-72 expression, suggesting that the p19^Arf/Mdm2 pathway is not required for this modification to p53 (Fig. 2B). Infection of HEL fibroblasts with HCMV or AdIE1-72 also resulted in an accumulation of total p53 and phospho-Ser15 p53 (Fig. 2C). IE1-72 expression alone had a greater influence on p53 and phospho-Ser15 p53 than HCMV infection, which was likely due to the higher levels of IE1-72 protein expressed from the adenovirus vector (Fig. 2C). Another serine residue on p53 that is often phosphorylated after DNA damage and other chromatin stresses is Ser20 (13, 14, 29, 63). We found that levels of phospho-Ser20 p53 are not affected by IE1-72 expression alone but are increased in HCMV-infected cells, indicating that additional p53 phosphorylation pathways are activated during HCMV replication (Fig. 2C). Thus, HCMV infection and IE1-72 expression alter p53 by promoting its phosphorylation.

Phosphorylation of p53 at Ser15 after IE1-72 expression requires ATM.

Several kinases have been shown to phosphorylate p53 at Ser15/18 after DNA damage, including ATM, ATR, and DNA-PK. These cellular kinases can be inhibited by caffeine (60). Therefore, we determined whether the increase in phospho-Ser15 p53 induced by IE1-72 expression was sensitive to caffeine in HEL fibroblasts. In the absence of caffeine, IE1-72-expressing cells displayed a higher level of phospho-Ser15 p53 relative to that seen in the control virus-infected cells (Fig. 3A). In the presence of caffeine, dose-dependent decreases in the levels of total and phospho-Ser15 p53 were observed in IE1-72-expressing cells.

We used human dermal fibroblasts from ataxia telangiectasia (AT) patients to determine whether ATM is required for p53 phosphorylation after IE1-72 expression. AT patients have an inactivated atm gene and, as a result, fail to fully execute many of the cellular responses to chromatin stress, including the ATM-dependent phosphorylation of p53 at Ser15. The levels of p53 phosphorylation observed after IE1-72 expression in AT fibroblasts were compared to the levels seen in age and gender-matched normal human dermal fibroblasts (WT) expressing IE1-72. IE1-72 expression in the WT fibroblasts resulted in nearly a fivefold increase in the levels of phospho-Ser15 p53 compared to the level observed in the control virus-infected WT fibroblasts, whereas the levels of total p53 were only slightly elevated (Fig. 3B). In contrast, the levels of phospho-Ser15 p53 detected in the AT fibroblasts expressing IE1-72 were similar to that detected in control virus-infected AT and WT fibroblasts. IE1-72 expression also had little effect on the levels of total p53 in AT fibroblasts, suggesting a role for phosphorylation in p53 stabilization.

ATM is autophosphorylated on Ser1981 when stimulated by DNA damage or chromatin changes (5). To confirm that ATM kinase was activated by IE1-72, we examined its phosphorylation status. As shown in Fig. 3C, we observed increased levels of phospho-Ser1981 ATM in cells expressing IE1-72, whereas the levels of total ATM did not increase and may even be reduced slightly. We also found that HCMV-infected cells displayed enhanced levels of phospho-Ser1981 ATM, as well as total ATM (Fig. 3C). Therefore, ATM is activated by IE1-72 expression alone and HCMV infection. Taken together, these findings demonstrate that IE1-72 induces an ATM-dependent phosphorylation of p53.

Contribution of IE1-72 and possibly IE2-86 proteins to p53 phosphorylation in HCMV-infected cells.

We have shown that deregulation of E2F leads to ATM autophosphorylation and p53 activation (56, 57, 58; F. M. Frame, et al., submitted for publication). Given that both IE1-72 and IE2-86 deregulate E2F activity through inactivation of RB family members, we sought to determine whether IE2-86 could also stimulate p53 phosphorylation. Expressing IE2-86 in HEL fibroblasts resulted in an increase in the phospho-Ser15 form of p53, whereas total p53 levels were unchanged (Fig. 4A). The absence of a change in total p53 was unexpected given that an earlier study has suggested that IE2-86 binds to and stabilizes p53 (69). However, results similar to ours have recently been published (68), although Song and Stinski analyzed exogenously expressed p53. Regardless of p53 accumulation, IE2-86, like IE1-72, expression leads to p53 phosphorylation.

We next compared HCMV to a mutant (CR208) virus devoid of IE1-72 expression. Both HCMV and CR208 virus infections resulted in an increase in total p53 levels, whereas CR208-infected cells had reduced levels of phospho-Ser15 p53 and phospho-Ser1981 ATM relative to HCMV-infected cells (Fig. 4B). On the surface, these results would suggest that IE1-72 is solely responsible for these changes to ATM and p53. However, at the MOI used in these experiments, UL122 expression (IE2-86) is apparently still dependent on transactivation by IE1-72 because IE2-86 levels are not detected in the CR208-infected cells (Fig. 4B).

IE1-72 expression results in nuclear retention of p53.

We previously noted that p53 accumulates in the nuclei of IE1-72-expressing cells (11). One possible explanation for our observation is that IE1-72 inhibits the ability of p53 to shuttle in and out of the nucleus from the cytoplasm. The p53 protein contains three nuclear localization sequences (NLS) and one nuclear export sequence (NES). The NLS are sufficient to mediate the shuttling of p53 from the cytoplasm into the nucleus (70, 74), and the NES can mediate the export of p53 from the nucleus in a Crm-dependent manner (70). Mdm2 can promote the export of p53 out of the nucleus (59, 73), whereas p53 phosphorylation abrogates the ability of p53 to shuttle from the nucleus to the cytoplasm (80). We generated heterokaryons containing human and mouse nuclei to determine whether IE1-72 expression also interferes with p53 shuttling. p53-deficient SAOS-2 cells were infected with a recombinant adenovirus encoding a human p53 cDNA (Adp53) and, in some instances, coinfected with AdIE1-72 prior to being fused with mdm2^−/−/p53^−/− MEFs. After fusion, the resulting heterokaryons were identified via immunofluorescence staining for human Ku-86 and the location of p53 was assessed (Fig. 5A). Nuclear shuttling of p53 between human and mouse nuclei was observed in 55% of the control heterokaryons, whereas p53 shuttled between nuclei in only 25% of the AdIE1-72-infected heterokaryons (MOI = 100) (Fig. 5B). p53 nuclear shuttling occurred in only 5% of heterokaryons infected with a higher dose of AdIE1-72 (MOI = 250) (Fig. 5B). Coinfection with a control adenovirus vector did not impede the ability of p53 to shuttle from human nuclei to mouse nuclei in heterokaryons, demonstrating that the increased vector load did not affect shuttling (compare Fig. 5B and C). Therefore, the nuclear accumulation of p53 after IE1-72 expression appears to be due, at least in part, to the ability of IE1-72 to abrogate p53 nuclear shuttling.

IE1-72 expression increases p21 levels in a p53-dependent manner.

Given that p53 phosphorylation and nuclear accumulation should result in increased p53 activity, we next analyzed the consequences of IE1-72 expression on p21, a transcriptional target of p53 (22, 27). Initially, we determined whether IE1-72 expression activates the p21 promoter in reporter assays. As shown in Fig. 6A, IE1-72 expression increased p21 promoter-reporter activity 17-fold over the control plasmid-transfected fibroblasts. IE1-72 expression also resulted in a more modest increase in p21 levels relative to that observed in control virus-infected cells (Fig. 6B). To quantify the accumulation of p21 on a per-cell basis, we immunohistochemically stained cells infected with AdIE1-72 or control virus for p21 and scored the number of p21-positive cells in each population. In both the mock-infected and the control virus-infected cells, 17 to 29% of the population stained positive for p21 during a 48-h time course (Fig. 6C). Cells infected with either a low (MOI = 100) or a high (MOI = 250) dose of AdIE1-72 exhibited a higher percentage (56 to 73%) of p21-positive cells during this time course. The percentage of p21-positive cells observed after IE1-72 expression was similar to the percentage of p21-positive cells after p53 overexpression, which served as a positive control for p21 induction in this experiment. The IE1-72-mediated increase in the number of p21-positive cells required p53, since IE1-72 expression did not alter the percentage of p21-positive cells in p53^−/− MEFs, whereas transduction of p53 into p53^−/− MEFs did increase the number of p21-positive cells (Fig. 6D). Therefore, IE1-72 expression leads to a p53-dependent increase in p21 levels.

IE1-72 expression induces quiescent cells to enter S-phase in the absence of p21.

Since induction of p21 expression by p53 is required for p53-mediated growth arrest, we determined whether p21 was responsible for the inability of IE1-72 to induce S phase in quiescent cells. Both WT and p21^−/− MEFs were rendered quiescent by serum withdrawal prior to infection with either AdIE1-72 or control virus. As shown in Fig. 7, IE1-72 expression did not increase the percentage of cells in S phase relative to the control virus-infected population of serum-starved, WT MEFs (3.1% versus 3.2%, respectively). However, IE1-72 expression in quiescent p21^−/− MEFs resulted in an increased percentage of S-phase cells over control infections (20.8% versus 6.7%, respectively). We included E2F1 as a positive control for S-phase induction (16, 17, 33, 38) and, as expected, E2F1 expression in MEFs caused an increase in S-phase cells in both WT and p21^−/− MEFs. The percentage of p21^−/− MEFs in S phase after IE1-72 expression was similar to that observed in E2F1-expressing cells. These results suggest that the inability of IE1-72 to induce quiescent cells into S phase is due to the activation of the p53/p21 growth arrest pathway.