SV40 large T antigen promotes dephosphorylation of p130.

SV40 large T antigen (Ag) binds to all members of the retinoblastoma (RB) tumor suppressor family including pRb, p107, and p130. Although the LXCXE motif of T Ag binds directly to the RB proteins, it is not sufficient to fully inactivate their function. The N-terminal DNA J domain of T Ag cooperates with the LXCXE motif to override RB-mediated repression of E2F-dependent transcription. In addition, T Ag can reduce the overall phosphorylation state of p107 and p130 that is dependent on an intact J domain and LXCXE motif. However, the mechanism of this activity has not been described. Here we describe the use of a cell-free system to characterize the effect of T Ag on p130 phosphorylation. When incubated in extracts prepared from S phase cells, p130 undergoes specific phosphorylation. Addition of T Ag to S phase extracts leads to a reduction of p130 phosphorylation in vitro. The ability of T Ag to reduce the phosphorylation of p130 in vitro is dependent on an intact DNA J domain and can be inhibited by okadaic acid and PP2A-specific inhibitors. These results suggest that T Ag recruits a phosphatase activity in a DNA J domain-dependent manner to reduce the phosphorylation of p130.

hyperphosphorylated (form 3) species in late G 1 and S phase cells (7)(8)(9)(10)(11)(12). In addition to changes in phosphorylation, p130 levels fluctuate significantly throughout the cell cycle. Expression of p130 is relatively high in growth-arrested cells and low in rapidly proliferating cells. The decreased levels of p130 during late S phase and G 2 reflect an increased rate of p130 degradation mediated by the SCF Skp2 ubiquitin ligase (13). In contrast, pRb and p107 levels increase slightly during proliferation secondary to increased levels of mRNA expression (14,15).
SV40 large T Ag can bind and inactivate all three members of the RB family. The LXCXE motif (where X is any residue) of T Ag binds directly to RB (16,17). Mutations in the LXCXE motif that perturb T Ag binding to RB reduce the transforming activity of T Ag (16). Although the LXCXE motif is necessary for T Ag binding to RB, it is not sufficient for complete inactivation of the RB family growth-suppressing functions. The N-terminal DNA J domain of T Ag also contributes to the inactivation of the RB family. The J domain is a highly conserved motif found in cellular and viral DNA J proteins that binds and activates specific Hsp70/DnaK proteins. The N termini of all Polyoma T antigens including the large and small T Ag of SV40 contain a DNA J domain (18,19). The J domain of SV40 T Ag binds to the constitutively expressed Hsc70, contributes to viral DNA replication, and cooperates with the LXCXE motif to disrupt RB-E2F complexes (20 -26). Mutation of the conserved HPD motif within the large T Ag J domain disrupts binding to Hsc70 and inactivates its ability to fully inactivate p130 and p107 (22,24).
The J domain and LXCXE motif of T Ag are also required to perturb the phosphorylation state of p130 (24,27). For example, when T Ag is stably expressed in mouse embryo fibroblasts, p130 is found exclusively as a hypophosphorylated species during all phases of the cell cycle. In contrast, p130 undergoes normal cell cycle-dependent phosphorylation in cells expressing T Ag that contains mutations within either the J domain or the LXCXE motif. Notably, T Ag can also reduce the phosphorylation of p107 but has no apparent effect on pRb phosphorylation (24,27). Despite these observations, it is not clear how T Ag perturbed the phosphorylation of p130 and p107. We explored the possibility that an in vitro approach would permit an examination of the effects of T Ag on the phosphorylation of p130.

EXPERIMENTAL PROCEDURES
Cell Culture-T98G cells (ATCC; human glioblastoma multiforme cell line) were cultured in Dulbecco's modified Eagle's medium and fetal bovine serum (Invitrogen), 100 units/ml penicillin, and 100 g/ml streptomycin in 10% CO 2 at 37°C. To induce growth arrest, 2 ϫ 10 6 cells were plated in a 150-mm dish and cultured in serum-free medium for 72 h. Cells were stimulated to reenter the cell cycle by addition of fresh medium containing 20% fetal bovine serum. The phase of the cell cycle was determined by fluorescence-activated cell sorter analysis of DNA content as measured by propidium iodide incorporation.
Cell Extracts-Cells were harvested by trypsin-EDTA and then washed twice with ice-cold phosphate-buffered saline and once with low salt extraction buffer (50 mM HEPES-NaOH (pH 7.4), 5 mM KCl, 1.5 mM MgCl 2 )containing 1 mM dithiothreitol and Complete TM protease inhibitor (Roche Applied Science) (28). After supernatant was removed, cells were resuspended in residual buffer and sonicated. The lysate was centrifuged twice at 14,000 ϫ g at 4°C to remove cellular debris and stored in aliquots at Ϫ80°C. For Western blot analysis, cells were washed twice with phosphate-buffered saline and lysed in extraction buffer C (50 mM Tris-HCl (pH 8.0), 125 mM NaCl, and 0.5% Nonidet P-40) containing 10 l/ml aprotinin, 10 g/ml leupeptin, 0.1 mM phenylmethylsulfonyl fluoride, and 4 mM NaF and 0.1 mM Na 3 VO 4 . Antibodies-p130 was detected with specific rabbit polyclonal antibody (antibody C-20, Santa Cruz Biotechnology). Additional antibodies used for blotting were anti-cyclin B (CB169, Upstate Biotechnology, Inc.) and anti p27 Kip1 (Ab-2, Neomarkers).
In Vitro Transcription and Translation-In vitro translated (IVT) samples were generated in a coupled transcription and translation system using reticulocyte lysate according to the manufacturers' recommendations (TNT; Promega, Madison, WI, and [ 35 S]methionine; PerkinElmer Life Sciences). HA-p130 and HA-pRb in pCDNA3 were used as templates (12). The p130 phosphorylation mutants, p130 PM19A and p130 ⌬Cdk4 , were treated similarly (10). In vitro translation products were stored at Ϫ80°C for up to 1 month prior to use.
Expression and Purification of T Ag-SV40 large T Ag fragments were expressed as glutathione S-transferase fusion proteins in the DH5␣ stain of Escherichia coli. T1-135 comprised the N-terminal 135 residues, and HQ1-135 contained a point substitution mutation H42Q that perturbed the DNA J domain. A 1-liter culture of bacteria containing pGEX2T-T1-135 or pGEX2T-HQ1-135 was induced with 0.1 mM isopropyl-1-thio-␤-D-galactopyranoside for 3 h before harvest by centrifugation and sonication in GST cleavage buffer (50 mM Tris HCl (pH 8.0), 200 mM NaCl, 2 mM EDTA, 1.0% Triton X-100, 10 mM ␤-mercaptoethanol, and Complete TM protease inhibitor). Lysates were cleared at 14,000 ϫ g for 15 min at 4°C and incubated with 2 ml of glutathione-Sepharose beads (Amersham Biosciences) for 1 h at 4°C. The loaded beads were washed with 15 volumes of GCB and then GCB with 0.5 M NaCl followed by 30 volumes of GCB without ␤-mercaptoethanol or inhibitors. T Ag was cleaved from glutathione S-transferase by addition of thrombin (200 units; Amersham Biosciences) in 1.5 ml of GCB for 2 h at 4°C. The reaction was stopped with 0.1 M phenylmethylsulfonyl fluoride. The cleaved T Ag was purified in a Superdex-75 gel (Amersham Biosciences) filtration column in NMR buffer (50 mM NaHPO4 (pH 6.5), 100 mM NaCl, 1 mM NaN 3 , and 2 mM dithiothreitol) and collected in 1-ml fractions. Peak protein fractions were pooled and concentrated to 1.5-2 g/l with a 5-kDa spin concentrator (Millipore). T Ag was separated in SDS-PAGE and silver-stained (Silver Quest, Invitrogen) to evaluate purity.
In Vitro Phosphorylation Assay-The in vitro protocol was adapted from a method used to test in vitro degradation of cyclin B and p27 Kip1 (29,30). Freshly thawed cell extracts were mixed with an ATP-regenerating system (1.5 mM ATP, 40 mM phospho-creatinine, and 80 g/ml creatine kinase), purified T1-135 or HQ1-135 (6 M final concentration), and 2 l of IVT p130 to a final volume of 10 l. Reactions were incubated at 30°C, and aliquots (2 l) were removed and boiled in SDS sample buffer. The zero time point was obtained immediately following IVT p130 addition. Where indicated, phosphatase inhibitors were added to reaction mixture prior to IVT p130 addition. A stock solution of okadaic acid (20 M; Calbiochem) was prepared in Me 2 SO. Recombinant rabbit PP1 inhibitor-2 (IC50 ϭ 2 nM; Calbiochem) and recombinant human I 1 PP2A inhibitor (IC50 ϭ 4 nM; Calbiochem) were added at 10 and 12 nM final concentrations, respectively. protein phosphatase (-PPase; New England Biolabs) was added 30 min at the indicated time point and before SDS sample buffer.

RESULTS
p130 Is Phosphorylated in Vitro in a Cell Cycle-specific Manner-T98G cells were cultured without serum for 72 h and then stimulated to reenter the cell cycle by addition of medium containing 20% serum. Cells harvested 5 h after serum addition were predominantly in the G 0 /G 1 phase, whereas after 24 h of serum stimulation, the majority had progressed past the restriction point and into S phase (Fig. 1A). An immunoblot for p130 confirmed that phosphoforms 1 and 2 were present in the G 1 lysates and that form 3 was present in the S phase enriched extracts (Fig. 1B). Consistent with good cell cycle synchrony, the G 1 extracts contained p27 Kip1 but not cyclin B, whereas the S phase lysates contained cyclin B but not p27 Kip1 .
When analyzed by SDS-polyacrylamide gel electrophoresis, IVT p130 appeared as a doublet when analyzed immediately after addition to the G 1 or S phase lysates (Fig. 2, lanes 1 and 11). Treatment with -PPase led to the appearance of a faster migrating form, suggesting that p130 was at least partially phosphorylated when prepared in reticulocyte lysates (Fig. 2, compare lane 1 with 6 and compare lane 11 with 16). IVT p130 remained as a doublet when incubated for up to 5 h in T98G extracts prepared from cells enriched for G 1 (Fig. 2, lanes 1-5). In contrast, IVT p130 underwent a rapid change in gel migration when incubated in S phase enriched extracts (Fig. 2, lanes  11-15). The gel migration changes were sensitive to -PPase, suggesting that IVT p130 underwent phosphorylation when incubated in S phase cell lysates (Fig. 2, lanes 16 -20).
The rate of phosphorylation of IVT p130 in S phase extracts (hyperphosphorylated form appeared within 0.5 h, Fig. 2, lane  12) is similar to the rate of phosphorylation for pRb and p130 as determined by pulse-chase labeling in vivo (24,31). Although the phosphorylation of IVT p130 led to the appearance of a diffuse signal when incubated in S phase extracts, treatment with -PPase resolved the signal into a sharp band with minimal loss of signal after 5 h incubation with no evidence for degradation (Fig. 2, lanes 16 -20).
SV40 T Ag Perturbs p130 Phosphorylation in Vitro-When transiently or stably expressed in cells, the N-terminal 135 residues of T Ag can bind to p130 and p107 and alter their phosphorylation status in a manner similar to full-length T Ag (32). To determine whether the N-terminal 135 residues of T Ag could perturb p130 phosphorylation in vitro, they were ex- pressed in bacteria and purified by fast protein liquid chromatography. A similar length construct containing a point substitution H42Q that disables the J domain was also prepared. The purity of the T Ag preparations was assessed by SDS-gel electrophoresis followed by silver staining (Fig. 3D). As expected, wild type T1-135 and mutant HQ1-135 could specifically coimmunoprecipitate IVT p130 since they both contain an intact LXCXE motif (data not shown).
IVT p130 appeared as a doublet, when incubated in the G 1 lysates, that did not change when incubated with T1-135 (Fig.  3A, top panel). In contrast, addition of T1-135 to the S phase lysates reduced the relative amount of hyperphosphorylated IVT p130 (Fig. 3A, second panel). To confirm that IVT p130 underwent specific phosphorylation when incubated in S phase lysates, we examined the ability of certain p130 mutants to become phosphorylated when mixed with S phase lysates. At least 22 residues within p130 are known to be phosphorylated in vivo (7,10). Cyclin D1/Cdk4, cyclin E/Cdk2, and cyclin A/Cdk2 contribute to the phosphorylation of p130 on specific serine and threonine residues. Cdk4-dependent phosphorylation sites within p130 were identified by determining the sensitivity of a given residue to phosphorylation in the presence of the Cdk4-inhibitor p16 Ink4a (10). The ⌬Cdk4 allele substitutes the Cdk4-dependent phosphorylation sites with alanine (T401A, S672A, and S1035A). Similar to wild type p130, IVT ⌬Cdk4 appeared initially as a doublet that became rapidly hyperphosphorylated when incubated in S phase extracts (Fig.  3A, third panel). Addition of T1-135 to the mixture reduced the hyperphosphorylated forms of ⌬Cdk4. The PM19A allele contains alanine substitutions of 19 Cdk-dependent phosphorylation sites and when in vitro translated, it appeared as a single band that changed minimally in gel mobility when incubated with S phase lysates (Fig. 3A, bottom panel). Addition of T1-135 did not affect the gel migration pattern of the mutant PM19A. Taken together, these results suggest that IVT p130 was specifically phosphorylated when incubated in S phase extracts and that T Ag reduced the phosphorylation of p130 under these conditions.
The ability of SV40 T Ag to perturb the phosphorylation of p130 in vivo requires an intact DNA J domain (24,27). We compared the activity of T1-135 with HQ1-135 to determine whether an intact T Ag J domain was required to perturb p130 phosphorylation in vitro. As shown in Fig. 3B (top panel), wild type T1-135 but not the J domain mutant HQ1-135 reduced the phosphorylation of IVT p130 when incubated in S phase extracts. Indeed, the overall phosphorylation state of p130 appeared to be enhanced when incubated with HQ1-135. The ability of J domain mutant T Ag to enhance the phosphorylation of p130 has also been observed in vivo (24).
Although wild type T Ag can affect the phosphorylation of p130 and p107, it has no apparent effect on pRb phosphorylation in vivo (24, 26, 27, 32, 33). To determine whether the ability of T Ag to perturb p130 phosphorylation in vitro was specific, we compared the ability of wild type and J domain mutant T Ag to affect the phosphorylation of pRb. Similar to p130, incubation of IVT pRb with S phase lysates led to a rapid change in gel mobility (Fig. 3B, bottom panel) that was sensitive to -PPase (data not shown), suggesting that pRb was also specifically phosphorylated in vitro. However, addition of wild type T1-135 or the J domain mutant HQ1-135 had no effect on the phosphorylation of pRb (Fig. 3B). The inability of T Ag to perturb pRb phosphorylation may reflect the inability of T Ag to bind to phosphorylated pRb (33).
We considered the possibility that T Ag reduced the overall level of p130 phosphorylation by promoting the degradation of the hyperphosphorylated forms of p130. However, the amount of IVT p130 did not decrease after addition of T1-135 when incubated in S phase lysates when quantitated by phosphorimaging. Furthermore, addition of the proteasome inhibitor lactacystin did not affect the ability of T1-135 to reduce the gel migration of IVT p130 in S phase lysates (Fig. 3C).

SV40 T Ag Effect on p130 Phosphorylation Is Dependent on a Phosphatase
Activity-The effect of T Ag on the phosphorylation of IVT p130 may be secondary to inhibition of a kinase or recruitment of a phosphatase. To distinguish between these possibilities, IVT p130 was incubated with S phase lysates prior to addition of T Ag. After a 1-h incubation in S phase lysates, IVT p130 became hyperphosphorylated as expected (Fig. 4A, lane 2). The reaction was subsequently aliquoted and incubated for an additional 2 h alone (lane 3) or with T1-135 (lane 4) and HQ1-135 (lane 5). Phosphorylation of p130 was significantly reduced when T was added after IVT p130 had become phosphorylated (lane 4). In contrast, IVT p130 remained hyperphosphorylated when incubated with HQ1-135. The ability of wild type T Ag to reduce the phosphorylation of p130 suggests that T Ag recruited a phosphatase to promote the dephosphorylation of p130.
The serine/threonine protein phosphatases PP1 and PP2A have been shown to be crucial in regulating RB activity. To determine whether p130 phosphorylation was sensitive to PP1 or PP2A, increasing concentrations of okadaic acid (OA) were incubated with S phase lysates and IVT p130 (34) (Fig. 4B). Addition of 0.1 or 1 M OA led to a slight increase in the relative amount of hyperphosphorylated p130, suggesting that a phosphatase-specific for p130 was active in the S phase lysates. In the presence of no or 5 nM OA, T1-135 remained capable of reducing the phosphorylation of p130 (Fig. 4B, bottom panel). However, addition of 0.1 or 1 M OA reduced the ability of T Ag to perturb p130 phosphorylation. The relatively high concentration of OA required is consistent with the possibility that PP2A participated in the T Ag effect on p130 phosphorylation.

SV40 T Antigen-dependent Dephosphorylation of p130
rylation of p130 (lane 4). Addition of OA with I-1 PP2A inhibited the ability of T1-135 to reduce p130 phosphorylation (lanes 5 and 6). These results suggest the role of a PP2A-specific activity as well as perhaps additional phosphatase activities in the T Ag-mediated dephosphorylation of p130. DISCUSSION The study of viral oncoproteins continues to provide insights into the regulation of the cell cycle by the Rb family. Previous work had established that the LXCXE motif and J domain of T Ag cooperated to inactivate the growth suppression and E2F repression functions of p130 (24,26). These domains of T Ag were required to promote the release of E2F4 from p130 in gel mobility shift assays and to perturb p130 phosphorylation (24 -27). To determine how T Ag affected p130 phosphorylation, an in vitro assay was developed. We observed that lysates prepared from S phase enriched cells supported the specific phosphorylation of p130. Addition of T Ag led to a dephosphorylation of p130 in a J domain-dependent manner.
Our evidence suggests that SV40 large T Ag recruits a serine/threonine phosphatase to promote the dephosphorylation of p130 given that OA completely inhibited the effect of T Ag. OA has a 4,000-fold lower K i for PP2A than PP1 (35). Given the relatively high concentrations of OA (Ն100 nM) required to inhibit the effect of T Ag on p130 phosphorylation, a PP2A-like phosphatase activity was likely involved. Use of OA may be particularly informative when studied in vitro because OA has profound effects on cellular physiology when added in vivo. For example, OA treatment of NIH 3T3 cells results in a G 1 arrest and decreased pRb phosphorylation, decreased Cdk activity, and a reduction in cyclin A, Cdc2, and Cdk2 expression (36). Furthermore, OA is relatively PP2A selective in vivo but may indirectly inhibit PP1 due to secondary accumulation of I-1 PP1 (34). We expect this effect to be minimized in vitro. Notably, OA can inhibit other members of the protein phosphatase family, including PP4 and PP5, so the contribution of additional serine-threonine phosphatases cannot be ruled out (34,37).
The ability of I-1 PP2A to at least partially inhibit the effect of T Ag on p130 phosphorylation also supports a role for PP2A. At inhibitory concentrations, I-1 PP2A partially reverses T Ag-mediated dephosphorylation of p130. The remaining effect of T Ag was reversed by addition of OA together with I-1 PP2A . Although I-2 PP1 , a PP1 protein-specific inhibitor, did not reverse T Agmediated dephosphorylation of p130, we cannot rule out that PP1 was not also involved.
PP2A is an important negative regulator of mitogenic activity that serves to dephosphorylate and activate key kinases such as MAP kinase and MAP kinase kinase. PP2A is a family of serine-threonine phosphatases consisting of two regulatory subunits, A and B, and a catalytic subunit C. Notably, SV40 small T Ag can bind specifically to PP2A by binding to the A and C subunit of PP2A and displacing the B subunit (38). SV40 small T Ag can inhibit PP2A activity, resulting in activation of the MAP kinase pathway and promotion of cell growth (39,40). Expression of small T antigen can enhance SV40 large T Ag transformation activity and is required for SV40 transformation of human fibroblasts (41,42). Although small T Ag shares the N-terminal J domain with large T Ag, the small T J domain is not necessary for binding to PP2A (40). The small T antigen J domain may participate in regulation of small T antigen-PP2A complex (43). Of note, small T antigen was not present in any of these assays, nor did the large T Ag constructs encode for small T antigen.
There is considerable evidence suggesting that phosphatases regulate the RB family members during the normal cell cycle and in response to stress conditions. PP1 has been shown to directly dephosphorylate pRb during mitosis, whereas PP2A may be indirectly involved in the pRb pathway by regulating G 1 Cdk activity (44 -46). A catalytic subunit PP1␣2 was found to bind pRb in a yeast two-hybrid assay and could associate with pRb during mitosis and early G 1 (47). During hypoxiainduced arrest, pRb becomes rapidly dephosphorylated due to an increased pRb-directed phosphatase activity and inhibition of Cdk4 and Cdk2 activity (48). In response to UV irradiation, p107 will undergo rapid dephosphorylation by a PP2A-like activity (49). In addition, p107 binds to PR59, a regulatory B subunit of PP2A (50).
Adenovirus E1A can also block hyperphosphorylation of p130 and p107 (51). The reported mechanism includes inhibition of Cdk-phosphorylation that can be overcome by overexpression of D-type cyclins (52). Since both E1A and T Ag have developed a mechanism to reduce the relative amounts of hyperphosphorylated p130, it suggests the possibility that hyperphosphorylated p130 may have specific roles in growth regulation that the viral oncogenes target for inactivation.
It would be interesting to speculate why the J domain of large T Ag may specifically recruit a phosphatase activity. T Ag itself is modified by phosphorylation, and certain residues affect T Ag-dependent viral DNA replication. For example, phosphorylation of T Ag residue Thr-124 is required for SV40 viral DNA replication, whereas phosphorylation of residues Ser-120 and Ser-123 inhibits T Ag replication activity, perhaps by inhibiting the phosphorylation of Thr-124 (53,54). Therefore, dephosphorylation of Ser-120 and Ser-123 may promote T Ag replication, and this has been reported to be dependent on PP2A activity (55)(56)(57)(58)(59)(60). Furthermore, dephosphorylation of Ser-120 and Ser-123 by PP2A can be inhibited by SV40 small T antigen (61). In addition, it has been reported that large T Ag itself can be dephosphorylated by incubation with S phase extracts (62). The N-terminal J domain of large T Ag contributes to SV40 replication in vivo, although the mechanism has not been clarified (22). Perhaps the large T Ag J domain serves to recruit a phosphatase to dephosphorylate T Ag residues Ser-120 and Ser-123, thereby promoting phosphorylation of Thr-124 and efficient viral DNA replication. The large T Ag J domain-associated phosphatase would serve two potential functions including dephosphorylation of T Ag to promote viral replication and dephosphorylation of p130 to promote transformation.