Phosphatase Type 2A-dependent and -independent Pathways for ATR Phosphorylation of Chk1*

ATM and Rad3-related (ATR) is a regulatory kinase that, when activated by hydroxyurea, UV, or human immunodeficiency virus-1 Vpr, causes cell cycle arrest through Chk1-Ser345 phosphorylation. We demonstrate here that of these three agents only Vpr requires protein phosphatase type 2A (PP2A) to activate ATR for Chk1-Ser345 phosphorylation. A requirement for PP2A by Vpr was first shown with the PP2A-specific inhibitor okadaic acid, which reduced Vpr-induced G2 arrest and Cdk1-Tyr15 phosphorylation. Using small interference RNA to down-regulate specific subunits of PP2A indicated that the catalytic β-isoform PP2A(Cβ) and the A regulatory α-isoform PP2A(Aα) are involved in the G2 induction, and these downregulations decreased the Vpr-induced, ATR-dependent phosphorylations of Cdk1-Tyr15 and Chk1-Ser345. In contrast, the same down-regulations had no effect on hydroxyurea- or UV-activated ATR-dependent Chk1-Ser345 phosphorylation. Vpr and hydroxyurea/UV all induce ATR-mediated γH2AX-Ser139 phosphorylation and foci formation, but down-regulation of PP2A(Aα) or PP2A(Cβ) did not decrease γH2AX-Ser139 phosphorylation by any of these agents or foci formation by Vpr. Conversely, H2AX down-regulation had little effect on PP2A(Aα/Cβ)-mediated G2 arrest and Chk1-Ser345 phosphorylation by Vpr. The expression of vpr increases the amount and phosphorylation of Claspin, an activator of Chk1 phosphorylation. Down-regulation of either PP2A(Cβ) or PP2A(Aα) had little effect on Claspin phosphorylation, but the amount of Claspin was reduced. Claspin may then be one of the phosphoproteins through which PP2A(Aα/Cβ) affects Chk1 phosphorylation when ATR is activated by human immunodeficiency virus-1 Vpr.

To ensure accurate transmission of genetic information, eukaryotic cells have developed an elaborate network of checkpoints to monitor the successful completion of every cell cycle step and to respond to cellular abnormalities such as DNA damage and replication inhibition as they arise during cell proliferation. The ATR kinase, a member of the phosphoinositide-3-kinase-related kinase family (1), is a central regulator for the replication checkpoint that is induced by treatment with DNA replication inhibitor hydroxyurea (HU). 2 When ATR is activated, it initiates a regulatory cascade of phosphorylation events. One of the intermediate steps in this cascade is the activating phosphorylation of the Chk1 effector kinase (2). This activated Chk1 in turn phosphorylates and inactivates the Cdc25 phosphatases. Cdc25 phosphatases control the activity of cyclin-dependent kinases by removing the inhibitory phosphates on Tyr 15 and Thr 14 (1). Specifically for the G 2 to M transition, which is controlled by activation of Cdk1, activation of ATR leads to inactivation of the Cdc25 phosphatases, in particular Cdc25C (3), ultimately resulting in the persistence of the phosphorylated Tyr 15 and Thr 14 on Cdk1, which causes a G 2 arrest (4).
ATR can be activated by many agents including HU, which blocks replication, and DNA damaging agents such as UV and ionizing radiation. The agents capable of activating ATR share the common property of generating long single-stranded DNA regions either by blocking DNA replication, such as occurs with HU, to give stalled replication forks or by nuclease processing of the initial DNA damage (5). The signal thought to activate ATR comes from the single-stranded DNA-binding protein RPA (replication protein A), which coats these abnormally long regions of single-stranded DNA (6). A number of other factors, such as ATRIP, Rad1, Rad9, Hus1, and Rad17, are thought to play various roles in activating ATR at these RPA-coated regions (1,5).
The actual mechanism of ATR activation does not appear to involve phosphorylation or any other covalent modification of ATR, and in vitro assays of immunoprecipitated ATR before and after activation generally do not show an increase in activity (7). Instead of a covalent modification, most models propose that the principal mechanism of activation is the formation of macromolecular assembles bringing substrates in close proximity to ATR. If these macromolecular assemblies are large enough, they can be visualized as foci in the nucleus by immunofluorescence. For example, ATR and RPA form nuclear foci in response to HU or UV (8), and these foci may represent the initial formation of macromolecular complexes where substrates are brought to be phosphorylated by ATR. Recently an alternative to this model of ATR activation by relocalization has been proposed. Kumagai et al. (9) showed that direct binding of TopBP1 was sufficient to activate ATR and proposed that this transient association with TopBP1 is the initial step in activation of ATR (9).
HIV-1 Vpr protein has recently been shown to be another agent that induces a G 2 arrest through ATR (10). Early results showed that Vpr induces cell cycle G 2 arrest through inhibitory Tyr 15 phosphorylation of Cdk1 in mammalian and yeast cells, suggesting that this viral protein exerts a highly conserved effect on a basic cellular function (11)(12)(13). Consistent with an ATR pathway, the inhibitory Tyr 15 phosphorylation of Cdk1 is achieved by inhibition of Cdc25 phosphatase (14,15), although activation of Wee1 kinase may also be involved (14,16). Roshal et al. (10) demonstrated that ATR has a major role in Vprinduced G 2 arrest through phosphorylation and activation of Chk1. Further studies have shown numerous similarities between the ATR pathway activated by Vpr and by other agents such as HU and UV. These similarities include a requirement for Rad17 and Hus1 (17), the induction of phosphorylation on Chk1 (10,17), and the formation of nuclear foci by RPA, 53BP1, BRCA1, and ␥H2AX (17,18).
The variant histone H2AX is a protein commonly used to monitor the nuclear foci formed during DNA damage and replication checkpoints. In response to replication arrest or many forms of DNA damage, H2AX is phosphorylated on Ser 139 to form ␥H2AX foci, and this phosphorylation is dependent on one or more phosphoinositide-3-kinase-related kinase including ATR (19). The ␥H2AX foci have been shown to retain DNA repair factors at the site of DNA damage, and it has been suggested that one of the roles of ␥H2AX is to concentrate DNA repair factors at sites of DNA damage (20,21). Recently, protein phosphatase 2A (PP2A) has been shown to be recruited to DNA damage foci by binding to ␥H2AX (22). Chowdhury et al. (22) proposed that the role of the PP2A at the ␥H2AX foci was to dephosphorylate ␥H2AX during recovery after DNA repair had been completed.
PP2A is one of the major Ser/Thr phosphatases implicated in the regulation of many cellular processes including regulation of signal transduction pathways, cell cycle progression, DNA replication, gene transcription, and protein translation (23)(24)(25)(26). The PP2A holozyme is composed of a 36-kDa catalytic C subunit (PP2A(C)), a 65-kDa scaffolding A subunit (PP2A(A) or PR65), and a regulatory B subunit (PP2A(B)). The core enzyme consists of PP2A(C) and PP2A(A), and one of the many B regulatory subunits associates with this core enzyme (A/C) to form a holoenzyme with specific properties and substrates (27). There are two isoforms of the catalytic core of PP2A, i.e. PP2A(C␣) and PP2A(C␤), which share 97% identity in their primary amino acid sequences (28 -30). PP2A(C␣) is more abundant than PP2A(C␤). PP2A(A) is also present in two isoforms in mammalian cells, ␣ and ␤, which share 86% sequence identity (31). PP2A(A␤) is less abundant than PP2A(A␣), and it does not bind strongly to the catalytic C and regulatory B subunits (32,33). Much of the diversity of PP2A holoenzymes comes from the four major classes of PP2A(B) regulatory sub-units, PR55/B, PR61/BЈ, PR72/BЈЈ, and PR93/PR110/BЈЈЈ. Each exists in at least four isoforms leading to many different holoenzymes, which partially explains the multiple and diverse cellular functions of PP2A (23,34).
A role for PP2A in Vpr-induced G 2 arrest was originally suggested by studies with the inhibitor okadaic acid in mammalian and yeast cells (12,13). Further evidence came from fission yeast where deletion of the gene for a catalytic subunit (ppa2) or a regulatory subunit (pab1) of PP2A reduced Vpr-induced G 2 arrest (14,35). However, specific involvement of PP2A in Vprinduced G 2 arrest in mammalian cells is controversial. A report showing Vpr induces G 2 arrest in mammalian cells by interacting with the B55 regulatory subunit of PP2A (36) was retracted (37).
In this study we investigated the role of PP2A in the activation of ATR by Vpr during the induction of cell cycle G 2 arrest. We measured the Vpr-induced phosphorylation of Chk1 and cell cycle G 2 arrest when PP2A enzymatic activity was downregulated either by the potent PP2A inhibitor okadaic acid (OA) or by specific small interference RNA (siRNA) to PP2A. We show here that the C␤ and A␣ subunits of PP2A, but not the C␣ subunit, are required for the Vpr-induced and ATR-dependent phosphorylation of Chk1. However, depletion of the C␤ and A␣ subunits did not decrease the extent of nuclei with ␥H2AX foci or the amount of ␥H2AX formed even though these events are ATR-dependent. In contrast to the PP2A(A␣/ C␤) dependence of Chk1 phosphorylation when ATR is activated by Vpr, depletion of the A␣ or C␤ subunit of PP2A has no effect on the ATR-dependent phosphorylation of Chk1 when ATR is activated by HU or UV. Thus, PP2A has a positive role in the ATR-dependent phosphorylation of Chk1 activated by Vpr but not when ATR is activated by HU or UV.
Drug Treatment-OA, a potent PP2A inhibitor, was purchased from Sigma. OA also inhibits other phosphatases (PPases) such as PP1, PP2B, and PP2C at high concentration (39). The range of OA concentrations (10 -25 nM) used here should specifically inhibit PP2A (40). To maintain proper final concentrations of OA during experiments, the same concentration of OA was added again 8 h after the first treatment. Cell treatment with HU or nocodazole (NOC) have been described previously (17). Briefly, cells were incubated with 10 mM HU for 2 h or 100 ng/ml NOC for 24 h before harvest.
Immunofluorescence Staining-Cells were fixed with 2% paraformaldehyde in PBS for 35 min at 4°C on Labtek II slides 48 h post-transfection. After washing 3 times for 5 min in PBS, cells were blocked and permeabilized for 20 min in blocking buffer (3% bovine serum albumin, 0.2% Triton X-100, 0.01% skim milk in PBS). Cells were then incubated with primary antibody with proper dilutions in the incubation buffer (1% bovine serum albumin and 0.02% Triton X-100 in PBS) for 45 min. After washing, cells were incubated with secondary antibody at the suggested dilutions in the incubation buffer for 35 min. After washing, cells were mounted with FluorSave reagent and visualized on a Leica DM4500B microscope (Leica Microsystems) with Openlab software (Improvision, Lexington, MA).
Western Blotting-Cells were lysed with lysis buffer (50 mM Tris, pH 7.5, 150 mM NaCl, 2 mM EDTA, 1% Triton X-100) on ice for 30 min, and the debris was removed by centrifugation at 13,000 rpm for 1 min. The protein concentrations of supernatants were measured by BCA protein assay kit (Pierce). After boiling, 50 g of protein was loaded on Criterion Precast Gels (Bio-Rad) for electrophoretic separation. Proteins were transferred to the Trans-blot nitrocellulose membranes and blocked with 5% skim milk in TBST buffer (10 mM Tris, pH 8.0, 150 mM NaCl, 0.1% Tween 20) for 1 h at room temperature. Primary antibodies were then applied overnight at 4°C. After washing 3 times in TBST for 10 min each time, the membranes were incubated with secondary antibody for 1 h at room temperature. Membranes were washed again, and proteins were detected with Supersignal West Pico chemiluminescent substrate (Pierce).
A semiquantitative RT-PCR assay for C␣ and C␤ mRNA used the primers 5Ј-ACCAAGGAGCTGGACCAGTG-3Ј and 5Ј-CCATGCACATCTCCACAGAC-3Ј with SuperScript TM one-step RT-PCR system (#10928-034, Invitrogen). These primers are perfect matches to both the C␣ and C␤ sequences and give a 161-bp PCR product. When digested with the TaqI restriction enzyme, the C␣ PCR product gives a 101-bp band, whereas C␤ gives a 136-bp band.
Adenoviral Infection-Adenoviral vector (Adv) and Adenoviral vector inserted with vpr gene (Adv-Vpr) were kindly provided by Dr. L. J. Zhao (42). The viral stocks were produced and titrated as described previously (43). Approximately 1 ϫ 10 6 dividing HeLa cells pretreated with siRNA for 24 h were infected with Adv or Adv-Vpr viruses. Cells were harvested for cell cycle analysis or Western blot analysis 48 h post-infection (p.i.). Infections were performed at a multiplicity of infection of 0.5 with 10 g/ml Polybrene. These conditions were shown to achieve more greater than 90% efficiency with these viral stocks.
Cell Cycle Analysis-At 48 h post-transfection (p.t.), cells were collected by trypsinization. Cells were then washed twice with 2 ml of 5 mM EDTA/PBS and centrifuged at 1500 rpm. After resuspension in 1 ml of 5 mM EDTA/PBS, cells were fixed with 2.5 ml of 95-100% cold ethanol and kept at 4°C overnight. After centrifugation, fixed cells were washed twice with 2 ml of 5 mM EDTA/PBS and centrifuged again at 1500 rpm. After resuspension again in 0.5 ml of PBS, cells were incubated with RNase A (10 g/ml) at 37°C for 30 min and then at 0°C with the addition of propidium iodine (10 g/ml) for 1 h. Cells were then filtered before analysis of DNA content by FACScan flow cytometry (Becton Dickinson). The cell cycle profiles were modeled by use of the ModFit software (Verity Software House, Inc.).

RESULTS
Inhibition of PP2A Enzymatic Activity Reduces Vpr-induced G 2 Arrest-OA, a specific and potent inhibitor of PP2A, was first used to determine whether PP2A is involved in Vpr-induced G 2 arrest. A vpr-inducible system in the HEK293 cell line (38) was used to test the effect of Vpr in the presence or absence of PP2A. The expression of vpr was induced by 1 M muristerone A, and increasing concentrations of OA in a range that specifically inhibits PP2A were added to the culture medium of vpr-expressing or vpr-repressing cells. Sixty-four hours after gene induction, cells were analyzed by flow cytometry for DNA content. The cell cycle profiles are shown in Fig. 1A, and the percentage of cells in G 2 /M are shown in Fig. 1B. As expected, vpr-expressing HEK293 cells showed a significant accumulation (62.1%) of cells in G 2 /M phase, whereas without vpr expression only 12.5% of cells were in G 2 /M phase (Fig. 1A). No significant differences were seen in cells with or without vpr PP2A(A␣/C␤) and ATR Activation MARCH 9, 2007 • VOLUME 282 • NUMBER 10 expression treated with 10 nM OA. However, Vpr-induced G 2 accumulation was significantly reduced when the concentrations of OA reached 17.5 and 25 nM. Although treatment with these higher concentrations of OA on cells without vpr expression caused small increases of cells in G 2 to 12.5-21.3%, the percentage of the G 2 cells decreased from 62.1 to 37.1 and 37.3% in the vpr-expressing cells treated with these higher concentrations of OA (Fig. 1B). These results suggest that the activity of PP2A is required for Vpr-induced G 2 arrest.
The A␣ and C␤ Subunits of PP2A, but Not C␣, Are Required for the Vpr-induced Cell Cycle G 2 Arrest-Even though OA has been shown to be a potent PP2A inhibitor at the concentrations used (40), we cannot rule out the possibility that other phosphatases might also be inhibited by OA. To further confirm the involvement of PP2A in Vpr-induced G 2 arrest, we used specific siRNA to deplete PP2A. Because siRNA technology is highly specific, different subunits of PP2A can be knocked down to allow detailed dissection of the role of PP2A in Vprinduced G 2 arrest. Here we tested both isoforms of the catalytic subunits of PP2A(C), i.e. PP2A(C␣) and PP2A(C␤), and the predominant regulatory PP2A(A␣) subunit. HeLa cells were transfected with specific siRNAs for PP2A(C␣), PP2A(C␤), PP2A(A␣), and control siRNA. Twenty-four hours p.t., the cells were either mock-infected or infected with Adv or Adv-Vpr, and 48 h post-infection (p.i.), the cellular DNA content was analyzed by flow cytometry. All of the mock or Adv infected cells displayed near normal cell cycle profiles regardless of the siRNA treatment, indicating those siRNAs had no significant effect on the cell cycle ( Fig. 2A). Consistent with the idea that Vpr induces cell cycle G 2 arrest, cells infected with Adv-Vpr and either untreated or pretreated with control siRNA had a marked accumulation of cells with G 2 /M DNA content. The percentage of G 2 /M cells increased from 8.0 to 8.4% without Vpr to 71.9 -87.3% with Vpr. However, Vpr-induced accumulation of G 2 /M cells was markedly reduced when either PP2A(C␤) or PP2A(A␣) was depleted by siRNA ( Fig. 2A, bottom panels). Only 28.4% of the cells had G 2 /M DNA content after PP2A(C␤) depletion and only 32.9% for PP2A(A␣) depletion. In contrast, no significant reduction of cells in G 2 /M was observed in Adv-Vpr-infected cells when PP2A(C␣) was depleted ( Fig. 2A, middle panel).
The extent of depletion by the siRNAs was examined by immunoblots and, for the catalytic subunit, by a semiquantitative RT-PCR assay. Consistent with PP2A(A␣) being the predominant form of PP2A(A), an immunoblot showed that the PP2A(A) protein was nearly eliminated by PP2A(A␣) knockdown (Fig. 2Ca). Similarly, for the catalytic subunit C, PP2A(C␣)-knockdown cells had a stronger reduction (66%) in PP2A(C) protein levels compared with PP2A(C␤)-depleted cells (22%) (Fig. 2Cb), consistent with PP2A(C␣) being more abundant than PP2A(C␤) (44). However, the antibodies used in the PP2A(C) immunoblots react equally with the C␣ and C␤ subunits so a semiquantitative RT-PCR assay was used to evaluate the specificity of the siRNAs for the catalytic subunits (Fig.  2Cd). The RT-PCR assay used two primers described under "Experimental Procedures" that are perfectly homologous to the cDNA from the mRNA for both C␣ and C␤ and which generated a 161-bp PCR product (Fig. 2Cd, row a). When this PCR product was digested with the TaqI restriction enzyme, only the C␣ product gave a 101-bp band, whereas only the C␤ product gave a 136-bp band. Consistent with the greater abundance of the C␣ subunit, the 101-bp band was stronger in untreated or control siRNA-treated cell samples (Fig. 2Cd, row  b, lanes 1 and 2). With siRNA against C␣, the 101-bp band characteristic of C␣ disappeared, leaving only the 136-bp band characteristic of C␤ (Fig. 2Cd, row b, lane 3), and with siRNA against C␤, only the 101-bp band characteristic of C␣ is seen (Fig. 2Cd, row b, lane 4). Thus, the siRNAs used show good specificity for the C␣ and C␤ subunits, and Vpr-induced G 2 arrest requires C␤ but not C␣.
As further evidence that PP2A has a specific role in Vprinduced G 2 arrest as opposed to a nonspecific effect, we tested the same siRNA PP2A knockdowns on HeLa cells treated with NOC, a M-phase-arresting drug (45). Although treatment with NOC caused an accumulation of the cells in the G 2 /M phase similar to Vpr, none of the siRNA treatments was able to suppress NOC-induced G 2 /M arrest ( Fig. 2A). Together, these data suggest that PP2A(A␣/C␤) is specifically required for Vpr-induced G 2 arrest.
Depletion of ATR and PP2A(A␣/C␤) Alone or in Combination Reduces Vpr-induced G 2 Arrest to Similar Extents-Because ATR has been shown to be required for Vpr-induced G 2 arrest (10), we were interested in determining the relationship between PP2A(A␣/C␤) and ATR during Vpr-induced G 2 arrest. HeLa cells were transfected with siRNAs against PP2A(C␤) and ATR either individually or in combination (Fig. 2B). These cells were then infected with Adv-Vpr as described above. Similar to the results shown in Fig. 2A, no significant changes in the cell cycle profiles were observed in cells either mock-infected or infected with Adv, and pretreatment with any siRNAs alone had no obvious effect on the cell cycle profile (Fig.  2B). As expected, Adv-Vpr-infected cells showed strong accumulation (89.1%) of cells with G 2 /M DNA content, indicating cell cycle G 2 arrest. Consistent with the results shown in Fig.  2A, depletion of PP2A(C␤) reduced cells in G 2 /M from 89.1 to 32.7%. A similar reduction (31.8%) of cells in G 2 /M was also observed when ATR was depleted with specific siRNA against ATR ( Fig. 2B; Ref. 10). When PP2A(C␤) and ATR were simultaneously depleted, 31.8% of cells were in the G 2 /M phase, indicating that no additional reduction of Vpr-induced G 2 arrest occurred with the combined depletion compared with PP2A(C␤) and ATR single depletions (Fig. 2B). These similar amounts of reduction suggest that PP2A and ATR are on the same regulatory pathway for Vpr-induced G 2 arrest.
The role of PP2A(A␣/C␤) in Vpr-induced phosphorylation of Cdk1-Tyr 15 was also evaluated in HeLa cells with the specific siRNAs against the PP2A subunits. The untreated cells (NT) showed a low background level of Cdk1-Tyr 15 phosphorylation (Fig. 3B, lane 1). As a positive control, cells were treated with HU, which induced strong hyperphosphorylation of Cdk1-Tyr 15 (Fig. 3B, lane 2; Ref. 46). NOC-treated cells were used as a negative control because it does not induce Cdk1-Tyr 15 phosphorylation (Fig. 3B, lane 3; Ref. 47). Adenoviral infection alone did not trigger a significant increase in phosphorylated Cdk1-Tyr 15 (Fig. 3B, lane 4). However, Adv-Vpr infection caused a C, depletion by siRNAs and specificity for C␣ and C␤. Immunoblots of HeLa cell lysates 48 h p.t. for mock-transfected (None) or transfected with control siRNA (Control) or specific siRNA were probed with antibodies against the PP2A A subunit (a), the PP2A C subunit (b), or ATR (c). Note that because PP2A(C␣) is more abundant than PP2A(C␤) (44), PP2A(C) is reduced more in the PP2A(C␣) knockdown cells than in the PP2A(C␤) knockdown cells. ␤-Actin was used as the protein loading control. The specificity of the C␣ and C␤ siRNAs is shown in d. As described under "Experimental Procedures," RT-PCR gives a 161-bp product for mRNA from both subunits (row a). After digestion with the TaqI restriction enzyme, the C␣ product gave a band of 101 bp, and the C␤ product gave a 136-bp band. In the untreated and control samples, the C␣ band is stronger than the C␤ band, consistent with the greater abundance of C␣. However, after treatment with C␣ siRNA, only the 136-bp C␤ band was observed, whereas only the 101-bp C␣ band was observed after treatment with C␤ siRNA (row b). MARCH 9, 2007 • VOLUME 282 • NUMBER 10 strong hyperphosphorylation of Cdk1-Tyr 15 to levels similar or higher than the HU-treated cells (Fig. 3B, lane 5). The same level of Cdk1-Tyr 15 phosphorylation was also seen in Adv-Vprtransduced cells transfected with control siRNA, indicating that siRNA transfection by itself had no effect on phosphorylation of Cdk1-Tyr 15 (Fig. 3B, lane 6). Significant reduction of Cdk1-Tyr 15 phosphorylation was, however, observed in the same Vpr-producing HeLa cells when they were transfected with siRNA either against PP2A(C␤) or PP2A(A␣) (Fig. 3B,  lanes 7 and 8). The levels of Cdk1-Tyr 15 phosphorylation in these two cell lysates were slightly higher but comparable with that seen with Adv infection, suggesting depletion of PP2A(C␤) or PP2A(A␣) reduced Cdk1-Tyr 15 phosphorylation to near the background levels. Therefore, PP2A is required for most of the hyperphosphorylation of Cdk1-Tyr 15 induced by Vpr.

PP2A(A␣/C␤) and ATR Activation
With Vpr Activation, PP2A Is Required for ATR Phosphorylation of Chk1-Activated ATR phosphorylates Chk1, and this phosphorylation is thought to activate Chk1, leading ultimately to inhibitory phosphorylation of Cdk1 to generate an ATR-dependent cell cycle arrest (7). Given the role of ATR in Vprinduced G 2 arrest (Fig. 2B, Ref. 10), the evidence that PP2A and ATR are on the same pathway for Vpr-induced G 2 arrest (Fig.   2B) and the strong effect of PP2A on Cdk1-Tyr 15 phosphorylation (Fig. 3), it seemed likely that PP2A is required for ATR phosphorylation of Chk1. This possibility was examined in HeLa cells pretreated with control, PP2A(C␤), PP2A(A␣), or ATR siRNA (Fig. 4A). These cells were then infected with Adv-Vpr. The cell lysates were subject to electrophoresis and blotted with anti-phospho-Chk1-Ser 345 antibody. Untreated cells (NT) showed almost no detectable phosphorylation of Chk1-Ser 345 (Fig. 4A, lane 1). As a positive control, cells treated with HU showed very strong phosphorylation of Chk1-Ser 345 (Fig. 4A, lane 2; Ref. 48). NOC treatment induced weak phosphorylation of Chk1-Ser 345 (Fig. 4A, lane 3; Ref. 48). Adenoviral infection alone also triggered a low level phosphorylation of Chk1-Ser 345 similar to that of NOC-treated cells (Fig. 4A, lane 4). In contrast, Adv-Vpr infection caused relatively strong phosphorylation of Chk1-Ser 345 as previously described (Fig. 4A, lane 5; Ref. 10). Similar levels of Chk1-Ser 345 phosphorylation were seen in   (48). ␤-Actin was used as protein loading control. B, ATR phosphorylation of Chk1-Ser 345 is not affected by PP2A depletion when ATR is activated by HU or UV. HeLa cells were first transfected with control siRNA (control) or specific siRNA against indicated subunits of PP2A or ATR. Twenty-four hours p.t., PP2A-or ATR-depleted or control HeLa cells were then not treated (NT), treated with 10 mM HU for 2 h or 3 J/m 2 UV for 10 s, and incubated for 2 h before collection. The phosphorylation status of Chk1-Ser 345 in cellular extracts was determined with anti-phospho-Chk1-Ser 345 or anti-Chk1 antibodies.
Adv-Vpr-transduced cells containing control siRNA, indicating that siRNA by itself had little effect on phosphorylation of Chk1-Ser 345 (Fig. 4A, lane 6). However, a significant reduction in Chk1-Ser 345 phosphorylation was seen in the same Vpr-producing HeLa cells when they were transfected with siRNA against either PP2A(C␤) or PP2A(A␣) (Fig. 4A, lanes 7 and 8). Levels of Chk1-Ser 345 phosphorylation in these two cell lysates were slightly higher but comparable with that of Adv-infected or NOC-treated cells (Fig. 4A, lanes 7 and 8, compared with  lanes 3 and 4), suggesting that depletion of PP2A(C␤) or PP2A(A␣) reduced Chk1-Ser 345 phosphorylation almost to background levels. Noticeably, the phosphorylation of Chk1-Ser 345 was completely eliminated in the ATR-knockdown cells (Fig. 4A, lane 9). The elimination of Chk1-Ser 345 phosphorylation by ATR depletion compared with a residual signal with PP2A depletion suggests that ATR is responsible for Chk1 phosphorylation induced by both Adv-Vpr and Adv infection and that PP2A is required for most of the Chk1 phosphorylation induced by vpr expression.
With HU or UV Activation, PP2A Has No Effect on ATR Phosphorylation of Chk1-There have been no reports that ATRmediated phosphorylation of Chk1 after HU or UV treatment is dependent on PP2A. To see if the ATR-mediated phosphorylation of Chk1 after HU or UV treatment is dependent on PP2A, the effects of siRNA knockdown were done with activation of ATR by HU or UV (Fig. 4B). As above, HeLa cells were pretreated with control, PP2A(C␤), PP2A(A␣), or ATR siRNA. Cells were then treated with HU or UV for the measurement of Chk1-Ser 345 phosphorylation. Little phosphorylation was seen in the untreated cells (Fig. 4B, NT, lane 1 with lanes 9 and 10). Pretreatment with ATR siRNA reduced Chk1-Ser 345 phosphorylation for both HU (lane 6) or UV (lane 11), demonstrating as expected that this phosphorylation is ATR-dependent (7). Thus, under the same experimental conditions, knock down of PP2A(A␣/ C␤) eliminates most of the Chk1-Ser 345 phosphorylation induced by Vpr but has no significant effect on the phosphorylation induced by HU and UV even though all these phosphorylations are ATR-dependent.

PP2A Is Not Required for Vpr-induced ␥H2AX Phosphorylation or Foci Formation, and H2AX Depletion Has Little Effect on Vpr-induced Chk1 Phosphorylation and G 2 Arrest-Previous
studies showed that the expression of vpr induces ␥H2AX foci formation, and this process is ATR-dependent (17). H2AX is a variant of the H2A histone protein, which is quickly phosphorylated at the site of DNA damage to form a nuclear structure known as ␥H2AX foci (20,49). We were interested in determining whether PP2A, like ATR, is also required for Vpr-mediated ␥H2AX foci formation. HeLa cells, transfected with siRNAs against PP2A(C␤), PP2A(A␣), or ATR, were infected with Adv or Adv-Vpr. Forty-eight hours p.i., ␥H2AX foci formation was visualized by immunofluorescence (Fig. 5A), and the percent-age of cells with foci was determined (Fig. 5B). Most of the Adv-Vpr-infected cells that were pretreated with or without the control siRNA (89.8% Ϯ 1.7; 90.1% Ϯ 2.0) formed ␥H2AX foci with a background level of ϳ11.8 Ϯ 1.5% as shown in the control Adv-infected cells (Fig. 5, A and B). In agreement with a previous report (17), pretreatment of Adv-Vpr-infected cells with ATR siRNA eliminated most of the ␥H2AX foci (25.5 Ϯ 2.6%). In contrast, pretreatment with PP2A(C␤) or PP2A(A␣) siRNA had no effect on the ␥H2AX foci formation in vpr-expressing cells as more than 90% of the cells were found to be positive in both experiments (91.6 Ϯ 1.5%; 92.0 Ϯ 1.6%) (Fig.  5B). We further measured the effect of PP2A down-regulation on the amount of ␥H2AX-Ser 139 phosphorylation. As shown in (Fig. 5C, lanes 7 and 8), depletion of PP2A(C␤) or PP2A(A␣) actually seems to increase the amount of ␥H2AX phosphorylation after vpr expression. Therefore, ATR is required, but PP2A is not required for Vpr-mediated ␥H2AX foci formation.
PP2A has been reported to interact with ␥H2AX, and this interaction is necessary for PP2A to form nuclear foci after DNA damage (22). To see if the PP2A-␥H2AX interaction is important for the role that PP2A plays in Vpr-induced G 2 arrest, H2AX was depleted by siRNA. Surprisingly, even when H2AX is greatly reduced by siRNA (Fig. 6B), there is little effect on Vpr-induced G 2 arrest, whereas a control experiment with depletion of ATR shows the expected reduction in Vpr-induced G 2 arrest (Fig. 6A). Similarly, H2AX depletion has no significant effect on the Vpr-induced Chk1-Ser 345 phosphorylation by ATR (Fig. 6C). Thus, even though Vpr induces ␥H2AX foci (Ref. 17, Fig. 5A), unless the small amounts remaining after siRNA depletion (Fig. 6B) are sufficient, the ␥H2AX foci do not play an important role in Vpr-induced G 2 arrest or Chk1 phosphorylation. This in turn means that PP2A does not need to interact with ␥H2AX to fulfill its role in Vpr-induced Chk1 phosphorylation and G 2 arrest.
Claspin Is a Potential Target for PP2A during Activation of ATR by Vpr-A number of cellular factors, such as Rad17-RFC complex, Rad9/Hus1/Rad1 (9-1-1) complex and Claspin, are required for Chk1 activation by ATR in response to DNA damage/replication stresses, and many of these required proteins are activated or inactivated by phosphorylation (5, 50 -53). PP2A is likely to fulfill its role in the Vpr-activated ATR pathway by dephosphorylating one or more of these proteins. The properties of Claspin make it a particularly good candidate to be a phosphoprotein with an important role in the PP2A-dependent Vpr pathway. Claspin is phosphorylated by at least three kinases, ATR, Chk1, and Plk1 (54 -58), and Plk1 phosphorylation is at a phosphodegron sequence that leads to degradation of the protein (55)(56)(57). After UV activation of ATR, Claspin levels increase due to stabilization of the protein (56,57), indicating that the degradation rate of Claspin is regulated by the DNA damage response. For Vpr activation, it is possible that PP2A activity keeps the phosphodegron site dephosphorylated so that Claspin is stabilized and present at levels high enough for activation of Chk1. Furthermore, after activation by UV or HU, Claspin is required for ATR to phosphorylate Chk1, but it is not required for phosphorylation of other ATR substrates such as H2AX (53,59). If Claspin has similar roles after ATR is activated by Vpr, the phosphodegron sequence of Claspin PP2A(A␣/C␤) and ATR Activation MARCH 9, 2007 • VOLUME 282 • NUMBER 10 might explain why PP2A is required for phosphorylation of Chk1 but not H2AX.
Claspin is required for phosphorylation of Chk1 when ATR is activated by Vpr. Depletion of Claspin by siRNA reduced Chk1-Ser 345 phosphorylation after infection with Adv-Vpr to about 10% that of the level seen with control siRNA. In the positive control, siRNA against ATR essentially eliminated Chk1 phosphorylation (Fig. 7A). Control immunoblots show that the specific siRNAs reduced ATR and Claspin to undetectable levels (Fig. 7A, lanes 3 and 4).
Expression of vpr affects Claspin in two ways. Claspin has been shown to be phosphorylated after ATR is activated by HU, and the phosphorylated form has slower electrophoretic mobility (52) (Fig. 7B, lane 7). Expression of vpr also induces phosphorylation of Claspin as indicated by the shift of Claspin to a band of slower mobility (Fig. 7B,  lane 2 and 3). This phosphorylation of Claspin induced by Vpr is ATRdependent since siRNA against ATR eliminates the upper band (Fig.  7B, lane 6). The second effect of Vpr is that the amount of Claspin increases (Fig. 7B, compare lanes 2 and 3 with Vpr to lanes 1 and 8 without Vpr). This increase is similar to that seen with HU (Fig. 7B, lane 7) and UV, which results from the stabilization of Claspin after DNA damage (56,57).
The major effect of depleting the subunits of PP2A is to reduce the amount of Claspin with little effect on the phosphorylation of Claspin. With siRNA against C␤ or A␣ subunits of PP2A, all of the Claspin is present as a slower migrating form (Fig. 7B, lanes 4 and 5), indicating that Vpr still induces phosphorylation of Claspin and suggesting that ATR phosphorylation of Claspin is independent of PP2A(A␣/C␤). The major effect of PP2A(A␣/C␤) depletion is to reduce the amount of Claspin (Fig. 7B, compare lanes 4  and 5 to lanes 2 and 3). This is the expected result if PP2A activity reduces phosphorylation of the phosphodegron sequence (55)(56)(57) so that Claspin is stabilized, but other interpretations are possible (see "Discussion").

DISCUSSION
In this study we demonstrated that Vpr requires PP2A(A␣/C␤) and ATR to induce cell cycle G 2 arrest in mammalian cells. PP2A(A␣/C␤) and ATR are likely to be part of the same regulatory pathway since depletion of both PP2A(A␣/C␤) and ATR decreases G 2 arrest to about the same extent as depletion of either one alone (Fig. 2B). The phosphorylation of Chk1 by activated ATR is thought to be an essential step in the induction of G 2 arrest by Vpr, and this Vpr-induced Chk1-Ser 345 phosphorylation is nearly eliminated when PP2A(A␣/C␤) are depleted (Fig. 4A). In contrast, there is no detectable effect of PP2A(A␣/C␤) depletion on the phosphorylation of Chk1 when ATR is activated by HU or UV (Fig. 4B). In considering the possible roles of PP2A(A␣/C␤) in Vpr-induced phosphorylation of Chk1, it is important to note that depletion of PP2A(A␣/C␤) has no effect on the formation of ␥H2AX foci after vpr is expressed (Fig. 5B), and the amount of ␥H2AX actually increases after depletion of PP2A(A␣/C␤) (Fig.  5C). These H2AX results indicate that at least some parts of the ATR activation by Vpr do not require PP2A and make it unlikely that PP2A is required exclusively for the initial interaction of Vpr with the cellular machinery, which leads to activation of ATR.
Based on these results, PP2A is likely to be involved in some later step of the ATR pathway induced by Vpr. Phosphorylation of H2AX, Claspin, and Chk1 are all dependent on ATR, and they are thought to be directly phosphorylated by activated ATR (2,7,19,58). Therefore, Vpr does activate ATR with respect to phosphorylation of the H2AX and Claspin substrates without PP2A(A␣/C␤) (Fig. 5, 7), but PP2A(A␣/C␤) is required for Vpr-induced phosphorylation of the Chk1-Ser 345 substrate by ATR (Fig. 4A). There is no evidence that ATR is regulated by phosphorylation (7), which is another reason that it is unlikely that PP2A regulates the pathway by directly dephosphorylating ATR.
Most models for activation of ATR postulate that formation of multiprotein complexes, often observed as nuclear foci, is essential (7), although it has recently been proposed that binding of TopBP1 is the initial step in ATR activation (9). One model for the PP2A requirement in Vpr-induced G 2 arrest is that the correct assembly and/or signaling by these multiprotein complexes depends on PP2A. PP2A could conceivably be required since numerous phosphorylation events occur during and after ATR activation with more than 10 ATR substrates having been identified thus far (1). The requirement for PP2A would be explained if the appropriate sequence of phosphorylation and dephosphorylation of at least one of these substrates depends on PP2A and is required for complete assembly of a fully functional complex.
The specific ATR substrate requiring PP2A might be RPA based on an analogy to the role of PP5 in ATR signaling (8). Although there have been no previous reports of a positive regulatory role for PP2A in ATR signaling, Zhang et al. (8) showed that the PP5 phosphatase is required for ATR phosphorylation of Chk1 after treatment with UV or HU. They found that ATR formed nuclear foci independently of PP5 after HU or UV treatment, but that formation of RPA foci required PP5. It is known that phosphorylation of RPA plays a role in its assembly into foci (60), and PP5 may be required so that the phosphorylation levels of RPA are appropriate for its assembly into functional foci (8). RPA is present in foci induced by Vpr (18), and based on the analogy to PP5, it will be of interest to determine whether RPA no longer associates with Vprinduced foci when PP2A is absent.
Claspin is another protein with an important role in the ATR pathway activated by Vpr, which may be targeted by PP2A(A␣/ C␤). When ATR is activated by other agents, Claspin is important for phosphorylation of Chk1 but not of H2AX (53,59). Thus, regulation of Claspin has the potential to explain the similar pattern of PP2A-dependent phosphorylation seen after Vpr activation (Fig. 4A and 5). Indeed, Claspin is required for Chk1 phosphorylation (Fig. 7A), and depletion of PP2A(A␣/ C␤) does reduce the amount of Claspin after Vpr activation FIGURE 6. Depletion of H2AX has little effect on Vpr-induced G 2 arrest and Chk1-Ser 345 phosphorylation. A, HeLa cells were transfected with siRNA against ATR or H2AX along with controls and 24 h later were mockinfected or infected with Adv or Adv-Vpr. Cell cycle profiles were determined 48 h p.i. B, depletion of H2AX by siRNA was shown by immunoblots of cell lysates for the indicated siRNAs with antibodies against H2AX. Control, protein loading control. C, depletion of H2AX has no significant effect on Vpr-induced phosphorylation (P-) of Chk1-Ser 345 . Control HeLa cells or cells transfected with the indicated siRNAs were treated with HU or NOC or infected with Adv or Adv-Vpr, and the phosphorylation status of Chk1-Ser 345 was examined in cellular extracts. (Fig. 7B). However, based on the known effects of Claspin phosphorylations, there are two opposite interpretations for this observed correlation between amounts of Claspin and Chk1 activation. The first interpretation is based on the finding that Plk1 phosphorylation of Ser 30 and Ser 34 in the phosphodegron sequence is destabilizing and leads to rapid proteolysis of Claspin (55)(56)(57). In this case, a model in which PP2A acts directly on Claspin to keep it in the stable, unphosphorylated form can explain the high levels of Claspin and the PP2A dependence of Chk1 phosphorylation. The second interpretation is based on the stabilization and phosphorylation of Claspin on Thr 916 and probably other sites by Chk1 (54,61). In this case it is the activated Chk1 that causes the high levels of Claspin after vpr expression. A model based on this interpretation proposes that the PP2A dependence of Chk1 activation results from PP2A acting directly on a phosphoprotein other than Claspin to give active Chk1 which then phosphorylates and stabilizes Claspin. Future studies with Claspin mutants that cannot be phosphorylated at these different sites will be required to clarify the contribution of Claspin to Vpr-induced, PP2A-dependent phosphorylation of Chk1. Interestingly, a slight increase of ATR protein level was also noted when vpr is expressed (Fig. 4A, lanes 6 -8; Fig. 7A, lane 2). This increase appears to be inversely correlated with Claspin as depletion of Claspin further enhanced ATR protein level (Fig. 7A, lane 4). These data implicates a possible negative feedback regulation of Claspin on ATR when it is under the influence of Vpr. Further in depth studies are needed to substantiate this premise.
Assuming that assembly of fully functional complex by Vpr for the phosphorylation of Chk1 requires PP2A, why would ATR phosphorylation of Chk1 after activation by HU/UV not require PP2A? One possibility is that the detailed structure of the foci differs between Vpr and HU/UV activation. One known difference in the structure of the foci is Vpr itself since many of the H2AX foci contain Vpr (18). Vpr could recruit or prevent the recruitment of proteins so that the detailed structure of foci induced by Vpr might have additional differences from foci induced by UV/HU, and these foci unique to Vpr might require PP2A for signaling to the cell cycle, whereas foci induced by UV/HU do not.
Although there have been no previous reports of a positive role for PP2A in Chk1 phosphorylation by ATR, there have been three reports where PP2A antagonizes ATR activity by dephosphorylating ATR substrates. Petersen et al. (62), based on experiments with soluble Xenopus extracts, proposed that PP2A dephosphorylates a number of ATR substrates including Chk1 to shut off an ATR/ATM-dependent double-strand breaks (DSB) checkpoint after DNA repair has been completed (62). Leung-Pineda et al. (63) showed that PP2A does in fact dephosphorylate Chk1 and that this dephosphorylation is regulated by Chk1 during the normal cell cycle (63). In experiments concentrating on the recovery after the repair of DNA damage induced by camptothecin in HeLa cells, Chowdhury et al. (22) presented evidence that PP2A binds to and dephosphorylates ␥H2AX during the disassembly of nuclear foci (22). Although it originally seemed possible that the binding of PP2A to ␥H2AX played some role in Vpr-induced G 2 arrest, our subsequent experiment with depletion of H2AX showed little effect on Vpr-induced G 2 arrest (Fig. 6A), suggesting that binding of PP2A to ␥H2AX is not required for PP2A to fulfill its role in Vpr-induced G 2 arrest.
Surprisingly, this work suggests that the less abundant C␤ catalytic subunit of PP2A is the major source of PP2A for Vprinduced G 2 arrest. The C␣ and C␤ subunits of PP2A are quite similar with only 8 amino acid differences of 309 amino acid, and the biochemical properties of the C␣ and C␤ appear to be identical (64,65). It is noteworthy that mouse and human C␣ and C␤ are nearly identical with C␤ being identical and only one amino acid difference between mouse and human C␣. This nearly complete conservation between mouse and human suggests that, despite their high degree of similarity, C␣ and C␤ each have at least one exclusive function that cannot be carried out by the other subunit. The C␣ subunit has in fact been shown to have a exclusive, required role in embryogenesis (66,67). However, no exclusive function for the less abundant C␤ subunit had been identified until now with the demonstration here that the C␤ subunit has a required role in Vpr-induced G 2 arrest, whereas the C␣ subunit plays no detectable role (Fig. 2A). HeLa cells were transfected with siRNA against Claspin or ATR along with control siRNA and 24 h later were mock-infected or infected with Adv or Adv-Vpr. The phosphorylation status of Chk1-Ser 345 in cellular extracts was determined with anti-phospho-Chk1-Ser 345 or anti-Chk1 antibodies. Cells in the ionizing radiation (IR) control were exposed to 10 gray and harvested 2 h after radiation. In the bottom two panels, depletion of ATR or Claspin by siRNA was demonstrated by immunoblotting of cell lysates with antibodies against ATR or Claspin. B, depletion of PP2A(C␤) or PP2A(A␣) does not affect Vprinduced phosphorylation but reduces the Vpr-induced increase in Claspin expression. Control HeLa cells or cells transfected with the indicated siRNAs were treated with HU or infected with Adv or Adv-Vpr. Cellular extracts were then subjected to 5% SDS-PAGE followed by Western blotting analysis using anti-Claspin antibody. p-Claspin indicates hyperphosphorylated Claspin (52). NT, not treated.
Dephosphorylating ␥H2AX may be another functional difference between the C␣ and C␤ subunits. Depletion of PP2A(C␤) or PP2A(A␣) actually increases the amount of ␥H2AX formed after vpr expression (Fig. 5C, lanes 7 and 8,  compared with lane 6). Chowdhury et al. (22) also found that the amount of ␥H2AX formed in response to DNA damage increased if PP2A was inhibited, and they proposed that PP2A directly dephosphorylates ␥H2AX. Chowdhury et al. (22) did not distinguish between the C␣ and C␤ catalytic subunits, but the results in Fig. 5C raise the possibility that the C␤ subunit may be primarily responsible for dephosphorylating ␥H2AX since the siRNA against the less abundant C␤ subunit (Fig. 5C,  lane 7), which specifically depletes C␤ (Fig. 2C), shows the increase in ␥H2AX. Given the greatly different roles for the PP2A catalytic subunits in Vpr-induced G 2 arrest and their possible different roles in dephosphorylating ␥H2AX, studies on the role of PP2A in the response to DNA damage should in the future examine whether the C␣ and C␤ subunits play differential roles in these responses.