|Home | About | Journals | Submit | Contact Us | Français|
Earlier studies have demonstrated a functional link between B56γ-specific protein phosphatase 2A (B56γ-PP2A) and p53 tumor suppressor activity. Upon DNA damage, a complex including B56γ-PP2A and p53 is formed which leads to Thr55 dephosphorylation of p53, induction of the p53 transcriptional target p21, and the inhibition of cell proliferation. Although an enhanced interaction between p53 and B56γ is observed after DNA damage, the underlying mechanism and its significance in PP2A tumor-suppressive function remain unclear. In this study, we show that the increased interaction between B56γ and p53 after DNA damage requires ATM-dependent phosphorylation of p53 at Ser15. In addition, we demonstrate that the B56γ3-induced inhibition of cell proliferation, induction of cell cycle arrest in G1, and blockage of anchorage-independent growth are also dependent on Ser15 phosphorylation of p53 and p53-B56γ interaction. Taken together, our results provide a mechanistic link between Ser15 phosphorylation-mediated p53-B56γ interaction and the modulation of p53 tumor suppressor activity by PP2A. We also show an important link between ATM activity and the tumor-suppressive function of B56γ-PP2A.
Protein phosphatase 2A (PP2A) is a very important family of holoenzyme complexes that functions within a diversity of signaling pathways inside the cell (9, 21, 26, 27). PP2A consists of either a core complex containing a catalytic (C) subunit and scaffolding (A) subunit (29) or a trimer containing the AC core with one of many possible regulatory (B) subunits bound to it (30). The known B subunits have been divided into four gene families based on sequence homology: the B (B55 or PR55), B′ (B56 or PR61), B" (PR48/59/72/130), and B (PR93/110) families (25). Each of these many B subunits can combine with the PP2A core to form complexes with distinct activities and substrate specificities. As such, PP2A is able to perform various functions in multiple regulatory pathways, depending on which B subunit is bound.
In the past, PP2A was thought to have primarily dull housekeeping functions inside the cell. Recent studies, however, suggest that PP2A may have more-active regulatory roles and may actually function as a tumor suppressor under certain conditions. It is believed that a small subset of B subunits is most likely responsible for promoting this function of PP2A. In support of this view, at least two B56 subunit family members have been implicated in conferring tumor-suppressive functions on the holoenzyme. The B56 family consists of five different genes, α (PPP2R5A), β (PPP2R5B), γ (PPP2R5C), δ (PPP2R5D), and (PPP2R5E) (5, 16). In addition, the B56γ gene encodes four differentially spliced forms, PP2A B56γ1, -γ2, -γ3 and -γ4 (17, 19). B56δ-specific PP2A was shown to function in a mitotic checkpoint in Xenopus laevis (15) and B56γ3-specific PP2A in blocking the proliferation of lung cancer cell lines (3). Importantly, evidence from our laboratory indicates that B56γ-PP2A participates in the activation of the tumor suppressor protein p53 after DNA damage (13).
p53 is a highly regulated tumor suppressor protein that is very important in cancer suppression. In response to genotoxic stress, p53 is activated through a series of posttranslational modifications (2). Once activated, it acts as a transcription factor, eliciting the transcription of genes that induce cell cycle arrest or programmed cell death (23, 28). Our studies have shown that, under cell growth conditions, p53 is phosphorylated at Thr55 by TAF1, which helps to keep the protein inactive, and upon genotoxic stress, B56γ-PP2A complexes dephosphorylate p53 at this residue, leading to p53 activation, p21 expression, and G1 cell cycle arrest (12, 13). Interestingly, in the course of our studies, we observed an enhanced interaction between B56γ and p53 upon DNA damage; however, its significance in p53 activation and in PP2A tumor suppressor function remains unknown.
In the present study, we show that the p53-B56γ interaction is required for p53 and B56γ-PP2A cooperative tumor suppression. Mechanistically, we show that the kinase activity of ATM is required for Thr55 dephosphorylation in response to DNA damage. ATM is an important kinase involved in cellular responses to DNA double-strand breaks. Once activated, ATM directly phosphorylates p53 at Ser15 and promotes Ser20 phosphorylation indirectly by activating Chk2 kinase. We show that Ser15, but not Ser20, mutant p53 is unable to interact with B56γ and significantly reduces the ability of B56γ3 to inhibit cell proliferation and transformation, suggesting that Ser15 phosphorylation primes p53 for the p53-B56γ interaction and Thr55 dephosphorylation by PP2A. Taken together, our results demonstrate the importance of the Ser15-mediated p53-B56γ interaction in the activation of p53 by B56γ-PP2A and in PP2A tumor suppressor function. In addition, our results also provide a functional link between ATM and PP2A tumor suppressor activity in response to DNA damage.
U2OS cells were cultured in McCoy's 5A medium supplemented with 10% fetal calf serum. H1299 and H1437 cells were cultured in RPMI 1640 medium supplemented with 10% fetal calf serum. Normal GM02254 and ATM-deficient GM01526 lymphoblasts were cultured in RPMI 1640 supplemented with 15% fetal calf serum. To induce DNA damage, the cells were subjected to UV radiation (10 J/m2 for U2OS cells) or gamma radiation (6 Gy for U2OS, 8 Gy for GM, and 10 Gy for H1299). The ATM inhibitor KU55933 was a gift from Kudos Pharmaceuticals. In ATM inhibition experiments, cells were treated with 10 μM KU55933, 7 mM caffeine, 30 μM wortmannin, or dimethyl sulfoxide control as indicated.
The Flag-ATM and Flag-ATM-KD plasmids were gifts from M. Kastan. The p53 mutants S15A, S15D, and S20A were generated by using a QuikChange site-directed mutagenesis kit (Stratagene). All plasmids were verified by sequencing.
Whole-cell extract was prepared by lysing the cells in a buffer containing 50 mM Tris-HCl (pH 8.0), 120 mM NaCl, 0.5% NP-40, 1 mM dithiothreitol, 2 μg/ml aprotinin, and 2 μg/ml leupeptin. Cell lysates were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis followed by immunoblotting analysis with anti-p53 (DO1; Santa Cruz Biotechnology), anti-phospho-Ser15 (Cell Signaling Technology), anti-phospho-Ser20 (Cell Signaling Technology), anti-phospho-Ser37 (Cell Signaling Technology), anti-acetyl-Lys373 (Upstate), anti-p21 (C-19; Santa Cruz), anti-PP2A A subunit (Upstate), anti-PP2A C subunit (1D6; Upstate), anti-PP2A B56γ3 (against full-length B56γ3), anti-FLAG (M2; Sigma), or antivinculin (VIN-11-5; Sigma) antibodies. For Thr55 dephosphorylation, the cell lysate was immunoprecipitated with phosphospecific antibody for Thr55 (Ab202 ) and immunoblotted with anti-p53 antibody. For the interaction of transfected p53 with endogenous B56γ3, H1299 cells were transfected with various p53 plasmids by using Lipofectamine (Invitrogen) and lysed 28 h after transfection. Immunoprecipitation was performed using either anti-p53 polyclonal antibody (FL393; Santa Cruz) or anti-B56γ polyclonal antibody. The amounts of coprecipitated proteins were determined by immunoblotting.
H1299 cells were transfected with empty vector or B56γ3 together with wild-type p53, S15A, or S20A, as well as a green fluorescent protein expression vector. Cells were harvested 60 h after transfection, fixed in paraformaldehyde, and stained with propidium iodide. Cell cycle phase distributions of green fluorescent protein-positive cells were determined by FACScan flow cytometry.
To generate proliferation curves for H1299 cells, cells were cotransfected with either B56γ3 or a control cytomegalovirus empty vector and either wild-type p53, the S15A mutant, or the S20A mutant by using Fugene (Roche); seeded in triplicate; and counted at 0, 24, 48, 72, 96, and 120 h postseeding.
For anchorage-independent growth assays of H1299 and H1437 cells, the cells were cotransfected with either B56γ3 or a control cytomegalovirus empty vector and either wild-type p53, the S15A mutant, or the S20A mutant; seeded in triplicate in 0.35% Noble agar (Fisher); and counted at 4 weeks postseeding.
After genotoxic stress, B56γ and p53 protein levels are induced and B56γ-PP2A associates with p53, promoting p53 activation through dephosphorylation of Thr55 (13). To better understand the mechanism of the enhanced interaction between B56γ-PP2A and p53 after DNA damage and its functional significance in tumor suppression, we recently generated polyclonal antibody against full-length B56γ protein (see Fig. S1 in the supplemental material). U2OS cells were either mock treated or treated with ionizing radiation (IR). B56γ-containing immunocomplexes were then precipitated using this antibody, and interacting proteins were assayed by Western blotting (Fig. (Fig.1A).1A). A dramatic increase in B56γ-p53 interaction is observed after IR (9.8-fold). Although both p53 and B56γ protein levels increase after IR (2.1- and 2.6-fold, respectively), they do not appear to increase significantly enough to completely explain the enhanced interaction. Likewise, an increased interaction is also observed between B56γ and both PP2A A and PP2A C after IR (1.8-fold); however, these increases seem to be consistent with the increased B56γ protein levels. Thus, in addition to the increased B56γ and p53 protein levels, other mechanisms may also promote B56γ-p53 interaction after IR. Since p53 and B56γ are both posttranslationally modified proteins, it is possible that modifications of one or both proteins after IR may modulate their ability to interact.
In an attempt to clarify the role of p53 modification in its enhanced interaction with B56γ after DNA damage, we performed a glutathione S-transferase (GST)-B56γ3 pull-down experiment using bacterially expressed GST-B56γ3 that is presumably unmodified. U2OS cells, pretreated with MG132 to normalize p53 protein levels, were either mock treated or treated with IR. The cell lysates were then incubated with GST-B56γ3 fusion protein or a GST-negative control, and the proteins interacting with the B56γ3 fusion protein were assayed by Western blotting (Fig. (Fig.1B).1B). With normalized p53 protein levels, p53 interaction with B56γ3 increased significantly after IR, suggesting that some modifications of p53 may have occurred, causing the increased interaction. Consistent with this result, we observed increased levels of p53 Ser15 phosphorylation after IR, as well as in the GST-B56γ3 pull-down assay, demonstrating that modified p53 can be brought down by the GST-B56γ3 fusion protein. Both endogenous subunits PP2A A and PP2A C bound to the GST-B56γ3 fusion protein, suggesting that the bacterially expressed fusion protein was in its native conformation.
In order to gain insight into potential p53 modifications that may be priming the molecule for interaction with B56γ after DNA damage, we assayed the timing of several p53 modifications relative to Thr55 dephosphorylation. U2OS cells were pretreated with MG132 to normalize p53 protein levels or with dimethyl sulfoxide to check p53 protein induction, followed by either IR or UV treatment. As shown in Fig. 1C and D, Ser15 and Ser20 phosphorylation increased 10 min after IR and 30 min after UV treatment and continued to increase over time, while Thr55 phosphorylation levels only began to decrease 30 min after IR and 2 h after UV treatment. K373 acetylation, on the other hand, occurred at around 40 min after IR and 2 h after UV treatment. The acetylation of K382, another major acetylation site on p53 after DNA damage, showed the same timing as the acetylation of K373 (data not shown). The finding that Ser15 and Ser20 phosphorylation precede Thr55 dephosphorylation after DNA damage suggests that both of these modifications could potentially be involved in promoting p53's association with PP2A, while K373 and K382 acetylation, which occur later, are probably not involved.
It has been demonstrated that ATM is one of the major kinases that phosphorylate p53 at Ser15 after IR and that ATM also phosphorylates and activates Chk2 after IR, which then phosphorylates p53 at Ser20 (18). We therefore investigated the importance of ATM activity in Thr55 dephosphorylation after DNA damage by using the phosphatidylinositol 3-kinase-like kinase inhibitors caffeine and wortmannin and the more-specific ATM inhibitor KU55933 (8). U2OS cells were pretreated with MG132 and one of the three inhibitors or a control treatment and then subjected to IR. Interestingly, Thr55 dephosphorylation after IR was completely blocked in the presence of each inhibitor (Fig. 2A, B, and C). As previously described, Ser15 and Ser20 phosphorylation were also blocked, providing additional evidence of their potential role in mediating p53-B56γ interaction after DNA damage. These data suggest that ATM kinase activity is involved in promoting Thr55 dephosphorylation after DNA damage.
In order to directly demonstrate the role of ATM in Thr55 dephosphorylation, we overexpressed the kinase and tested its effect on Thr55 phosphorylation. As shown in Fig. Fig.2D,2D, the expression of wild-type ATM, but not a kinase-dead (KD) dominant-negative mutant (14, 22), led to a modest decrease in Thr55 phosphorylation. Ser15 phosphorylation levels also increased modestly when the wild-type kinase was overexpressed, as expected. These results indicate that ATM-mediated phosphorylation, particularly at Ser15, may promote Thr55 dephosphorylation. The modest effects observed with ATM overexpression may be due to the tendency of ATM to form inactive multimer complexes under normal growth conditions (1). To overcome this problem, we assayed the effect of ATM on Thr55 dephosphorylation under DNA damage conditions. Compared to the results with the control, the overexpression of wild-type ATM led to higher Ser15 and Ser20 phosphorylation levels and earlier Thr55 dephosphorylation, while the overexpression of the KD mutant completely blocked Thr55 dephosphorylation, as well as Ser15 phosphorylation, after IR (Fig. (Fig.2E).2E). The p53 protein levels were normalized by MG132 pretreatment. The phosphorylation, as a control, of Ser37, a previously described PP2A dephosphorylation site, was unaffected by ATM. To provide further evidence, we show that the IR-induced p53-B56γ interaction and Thr55 dephosphorylation were completely abolished in ATM-deficient (GM01526), but not in normal (GM02254), human lymphoblasts (Fig. (Fig.2F).2F). Overall, these findings provide evidence that ATM kinase activity is required for Thr55 dephosphorylation after IR. In addition, the results suggest that Ser15 and Ser20 phosphorylation play a role in modulating p53-B56γ interaction.
To further investigate the role of ATM-mediated phosphorylation of p53 in Thr55 dephosphorylation, we generated Ser15 to Ala (S15A), Ser15 to Asp (S15D), and Ser20 to Ala (S20A) p53 mutants to either abolish or mimic ATM-induced phosphorylation. We overexpressed these mutants in H1299 cells that lack endogenous p53 protein and assayed for their ability to interact with B56γ-PP2A (Fig. (Fig.3A).3A). The assay shows that the wild-type p53 and S15D and S20A mutant proteins were able to interact with B56γ in vivo, while the S15A mutant could not. These results support our finding of the importance of ATM activity in Thr55 dephosphorylation. The phosphomimic S15D mutant showed levels of interaction similar to those of the wild-type p53, suggesting that perhaps this mutant does not perfectly mimic Ser15-phosphorylated p53 under our assay conditions. Total p53 protein was normalized in the immunoprecipitation. Taken together, these results suggest that Ser15 phosphorylation is required for Thr55 dephosphorylation. Interestingly, the interaction between p53 and PP2A A and C showed a similar trend in that the AC core interacted with wild-type p53, the S15D mutant, and the S20A mutant, but not the S15A mutant. This finding underscores the importance of the B56γ subunit in bridging the interaction between p53 and the PP2A core and demonstrates the requirement of Ser15 for this interaction.
To confirm the results, we performed reciprocal immunoprecipitation experiments using anti-B56γ antibody (Fig. (Fig.3B).3B). As expected, the interaction between p53 and B56γ was similar when wild-type p53, the S15D mutant, or the S20A mutant was expressed, while no interaction was detected in the presence of the S15A mutant. As a control, the expression of p53 constructs had no effect on B56γ interaction with PP2A C. These results provide further evidence for the importance of Ser15 in the interaction between PP2A and p53.
Since p53 is dephosphorylated at Thr55 by B56γ-PP2A after IR, we investigated the role of Ser15 phosphorylation in this process. H1299 cells were transfected with wild-type p53, the S15A mutant, or the S20A mutant and either mock treated or treated with IR. Cell lysates were then subjected either to immunoprecipitation/immunoblotting for the p53-B56γ interaction or to immunoblotting for p53 modifications (Fig. (Fig.3C).3C). The assay shows that S15A mutant p53 was unable to interact with B56γ after IR treatment, which is consistent with the lack of Thr55 dephosphorylation after IR. As we would expect, wild-type p53 and the S20A mutant both had significantly enhanced interactions with B56γ after IR, which correlated well with their ability to be phosphorylated at Ser15 and dephosphorylated at Thr55. Ser20 phosphorylation of wild-type p53 and the S15A mutant, as well as Ser37 phosphorylation of all four constructs, was used as a control for the cellular response to IR. Under the experimental conditions, transfected p53 protein levels were not affected by IR. Taken together, these results clearly demonstrate the importance of Ser15 in p53 Thr55 dephosphorylation through increased association with B56γ-PP2A after IR.
Based on our results, we reasoned that if Ser15 phosphorylation promoted Thr55 dephosphorylation after IR, we should not be able to detect phosphorylation at both residues on the same p53 molecule. To test this, p53 protein was immunoprecipitated with the Thr55-phosphospecific antibody at time points when Thr55 is phosphorylated, specifically mock and 20 min after IR. Ser15 and Ser20 phosphorylation, as well as p53 levels, in the immunocomplexes was then assayed by immunoblotting (Fig. (Fig.3D).3D). No Ser15 phosphorylation was detected in the immunoprecipitation using Thr55-phosphospecific antibody, whereas Ser20 phosphorylation was. It has been shown that at later time points after IR, Thr55 phosphorylation recovers to levels similar to those seen with mock treatment (12). We therefore investigated whether Ser15 and Thr55 phosphorylation occurred on the same p53 molecule at one of these later time points by performing a Thr55 phosphoimmunoprecipitation of lysate from cells harvested 80 min after IR treatment. Even at this later time point, no Ser15 phosphorylation was detected in the immunoprecipitation using Thr55-phosphospecific antibody, while Ser20 phosphorylation was. As a control for the assay, when Lys373-acetylated p53 was immunoprecipitated with an acetyl-Lys373-specific antibody, Ser15- and Ser20-phosphorylated p53 were brought down in the immunocomplexes. These findings clearly show that Ser15 phosphorylation and Thr55 phosphorylation do not occur on the same p53 molecule and provide further evidence that the presence of Ser15 phosphorylation promotes Thr55 dephosphorylation. Collectively, our findings delineate the requirement of Ser15 phosphorylation for Thr55 dephosphorylation through enhanced association of p53 with B56γ after DNA damage.
We previously demonstrated both p53-dependent and -independent tumor suppressor activities of B56γ-PP2A (13). To determine the importance of Ser15 phosphorylation in the p53-dependent function of B56γ-PP2A, we investigated the ability of B56γ3 to inhibit cell proliferation in the presence of different p53 constructs (Fig. (Fig.4A).4A). B56γ3 overexpression in the presence of wild-type p53 significantly inhibited cell growth in H1299 cells, demonstrating that the two proteins function together synergistically to block cell proliferation. Further, we show that the overexpression of B56γ inhibits endogenous p53-MDM2 interaction (see Fig. S2 in the supplemental material), which is consistent with our previous observation that Thr55 phosphorylation promotes the p53-MDM2 interaction (12). In contrast, the overexpression of B56γ3 in the presence of the S15A mutant inhibited H1299 cell proliferation only modestly. The presence of B56γ3, p53, and p21 during the analysis was verified by immunoblotting (Fig. (Fig.4A).4A). As controls, the overexpression of either p53 or B56γ3 individually had only a small effect on cell proliferation. In a manner similar to that of wild-type p53, the S20A mutant was able to decrease cell proliferation slightly on its own and much more dramatically when combined with B56γ3 overexpression. The cell-doubling times from three parallel cell growth experiments are shown in Fig. Fig.4C.4C. Because of the presence of endogenous B56γ in H1299 cells (Fig. (Fig.4A)4A) the S15A mutant shows slightly reduced activity compared to that of wild-type p53 without B56γ3 overexpression. Consistent with this result, the S15A mutant shows a slightly higher level of Thr55 phosphorylation than wild-type p53 in H1299 cells (Fig. (Fig.3C).3C). Interestingly, compared to the results with wild-type p53, cells cotransfected with the S15A mutant and B56γ3 showed reduced levels of endogenous p21 expression (Fig. (Fig.4A),4A), while cells cotransfected with the S20A mutant and B56γ3 showed similar levels of p21 expression (data not shown). These results provide evidence for the functional importance of Ser15 phosphorylation in p53-dependent B56γ-PP2A tumor-suppressive function.
The observation that Ser15 phosphorylation is required for p21 induction prompted us to investigate its role in cell cycle G1 arrest. The combination of wild-type p53 with B56γ3 overexpression led to a significant increase in G1 arrest (Fig. 4B and D), while the combination of wild-type p53 with B56γ3 knockdown led to a decrease in G1 arrest in H1299 cells (see Fig. S3 in the supplemental material). Similarly, the presence of the S20A mutant along with B56γ3 overexpression caused significant increases in G1 arrest. The S15A mutant on the other hand, led to only a marginal increase in G1 arrest when coupled with B56γ3 overexpression. These results correlate with the conditions of maximal p21 induction. B56γ3 overexpression by itself caused a low level of G1 arrest, as did the overexpression of each of the p53 constructs individually. Overall, these data provide further support for the importance of S15 phosphorylation in the modulation of p53 function by B56γ-PP2A.
In addition to blocking cell proliferation, B56γ3 overexpression has also been shown to block anchorage-independent growth (3, 13). We previously demonstrated a p53-dependent inhibition of anchorage-independent growth by B56γ-PP2A, so we tested the importance of Ser15 in this process in H1299 cells and another lung cancer cell line, H1437, which also lacks functional p53. The results from the anchorage-independent growth assays were similar for both cell lines tested, and the trend was similar to that seen with the cell proliferation assays (Fig. 5A and B). When wild-type p53 or the S20A mutant construct was expressed along with B56γ3, a significant decrease in colony number was observed. The decrease was more than could be attributed to an additive effect of both p53 and B56γ3 expression individually and can only be attributed to a cooperative function between the two proteins. This cooperative function is abolished if the S15A mutant is expressed, however, and an additive effect only is observed in this case. As a control, the overexpression of B56γ3 in the absence of p53 caused a small decrease in the number of colonies present on the soft agar. Furthermore, the expression of any of the p53 constructs in the absence of B56γ3 overexpression also caused a small decrease in colony number. The presence of p53 and hemagglutinin (HA)-tagged B56γ3 (HA-B56γ3) was verified by immunoblotting (Fig. (Fig.5C).5C). Interestingly, it was previously reported that H1437 cells lack endogenous B56γ3 protein (3), but we were able to detect it with our antibody (Fig. (Fig.5C;5C; see Fig. S1 in the supplemental material). In addition, the overexpression of B56γ3 promoted Thr55 dephosphorylation of wild-type p53 and the S20A mutant, but not the S15A mutant. The expression of p21 was also verified, which correlated well with Thr55 dephosphorylation and with the ability of the p53 constructs to interact with B56γ, as shown in Fig. Fig.3C.3C. The strong correlation between the p53-B56γ3 interaction, Thr55 dephosphorylation, and the induction of p21 expression suggests that B56γ3 promotes p53 transcriptional activation leading to inhibition of tumor formation through interacting with p53. These results also clearly demonstrate the importance of p53 Ser15 phosphorylation in the tumor suppressor activity of B56γ-PP2A, shedding light on the mechanism by which these two proteins are cooperatively functioning to inhibit cell transformation.
In the present study, we demonstrate that the B56γ-p53 interaction is mediated by Ser15 phosphorylation after DNA damage and is required for p53 and PP2A synergistic tumor-suppressive function. Specifically, we show that, without Ser15 phosphorylation, the B56γ-mediated, p53-dependent inhibition of cell proliferation and transformation are lost. Importantly, we identify ATM as having a key role in promoting Thr55 dephosphorylation of p53, providing insight into the molecular mechanisms regulating PP2A tumor suppressor activity in response to DNA damage. Overall, our results indicate a stepwise procession of events beginning with the activation of ATM by DNA damage (Fig. (Fig.6).6). Once activated, ATM phosphorylates p53 at Ser15, thereby promoting interaction with B56γ-PP2A and dephosphorylation of Thr55. These ordered p53 modifications lead to maximal activation of the protein, promoting cell cycle arrest and allowing for DNA repair to occur.
The N terminus of p53 interacts with several molecules that regulate p53 activity, including p300/CBP and MDM2. The phosphorylation of serine and threonine residues within this domain is involved in regulating p53 interaction with these binding partners that are enzymes that further posttranslationally modify p53 after binding. Specifically, Ser15 and Ser20 phosphorylation promote interaction with p300/CBP, which leads to p53 C-terminal acetylation and enhanced transcriptional activity (6, 10). In addition, the phosphorylation of either Thr18 or Thr55 can affect the interaction with MDM2 by either disrupting it or enhancing it, respectively, thereby regulating p53 ubiquitination and stability (4, 12, 20). In this study, we show that Ser15 phosphorylation promotes p53 interaction with B56γ-PP2A, thereby leading to Thr55 dephosphorylation. Two possible mechanisms may contribute to this result. First, Ser15 phosphorylation may directly affect p53 interaction with B56γ-PP2A. Second, Ser15 phosphorylation may affect the tetramer formation of p53 and thereby change the stoichiometry of the p53-B56γ interaction. Nevertheless, our results provide additional evidence of p53 N-terminal phosphorylation regulating its ability to bind to other enzymes that further modify the protein. Importantly, our findings also suggest a signal cascade leading to higher levels of p53 activation in response to DNA damage, i.e., Ser15 phosphorylation precedes Thr55 dephosphorylation. Because Thr55 phosphorylation destabilizes p53, the link between these two modifications provides an additional mechanism by which Ser15 phosphorylation can promote p53 stability. We demonstrate that this p53 signal transduction cascade is specifically necessary for the cooperative tumor-suppressive function with B56γ-PP2A.
PP2A substrate specificity is regulated through its B subunit composition, and as such, specific B subunits bridge the interactions between the PP2A core and target proteins. In fact, PP2A was demonstrated to function as a tumor suppressor through specific B-subunit-substrate interactions. B56δ-PP2A has been shown to function in a mitotic checkpoint by directing the AC core to dephosphorylate and inactivate Cdc25 (15). In addition, studies from our laboratory have shown that B56γ-PP2A functions to suppress cell proliferation and transformation through the interaction and regulation of p53. In this paper, we have demonstrated the importance of both p53 Ser15 phosphorylation and p53-B56γ interaction in mediating this function. Specifically, in the presence of S15A mutant p53, the cooperative tumor suppression between B56γ3-PP2A and p53 is abolished. Interestingly, the B56γ subunit of PP2A is also able to direct the AC core to another as-yet-unidentified substrate, as B56γ3-PP2A was shown to function by way of suppressing tumor growth and formation independent of p53 protein (3, 13). Further investigation is necessary to identify the additional substrates involved in p53-independent B56γ-PP2A tumor suppression.
Tumor suppressor functions typically fall into one of three categories: DNA damage sensing, DNA repair, or cell cycle regulation. Upon DNA damage, sensory proteins activate cell cycle regulatory proteins and DNA repair proteins. In this manner, deleterious mutations and tumor progression are thought to be avoided. Although previous data has demonstrated a role for B56γ-PP2A as a tumor suppressor protein functioning in the cell cycle regulatory category, its link to DNA damage sensing has remained elusive. In this study, we demonstrate a link between ATM, an important member of the Mre11 complex which plays a critical role in recognizing DNA double-strand breaks (11, 24), and B56γ-PP2A function through the phosphorylation of p53 at Ser15. However, because B56γ-PP2A also functions in suppressing tumor growth independent of p53, ATM may also regulate the B56γ-PP2A tumor suppressor function through other pathways. Thus, it will be interesting to further investigate the role of ATM in the regulation of the B56γ-PP2A tumor suppressor function.
We are very grateful to M. Kastan for providing ATM plasmids, to G. Smith and Kudos Pharmaceuticals for providing the ATM inhibitor KU55933, and to H. Myllykangas for providing technical assistance. We thank J. A. Traugh and all members of our laboratory, particularly A. G. Li and P. Podlesny, for many helpful discussions and critical reading of the manuscript.
This work was supported by NIH grant CA075180 from the National Institute of Cancer.
Published ahead of print on 29 October 2007.
†Supplemental material for this article may be found at http://mcb.asm.org/.