Inactivation of p53 by Human T-Cell Lymphotropic Virus Type 1 Tax Requires Activation of the NF-κB Pathway and Is Dependent on p53 Phosphorylation

Abstract

p53 plays a key role in guarding cells against DNA damage and transformation. We previously demonstrated that the human T-cell lymphotropic virus type 1 (HTLV-1) Tax can inactivate p53 transactivation function in lymphocytes. The present study demonstrates that in T cells, Tax-induced p53 inactivation is dependent upon NF-κB activation. Analysis of Tax mutants demonstrated that Tax inactivation of p53 function correlates with the ability of Tax to induce NF-κB but not p300 binding or CREB transactivation. The Tax-induced p53 inactivation can be overcome by overexpression of a dominant IκB mutant. Tax-NF-κB-induced p53 inactivation is not due to p300 squelching, since overexpression of p300 does not recover p53 activity in the presence of Tax. Further, using wild-type and p65 knockout mouse embryo fibroblasts (MEFs), we demonstrate that the p65 subunit of NF-κB is critical for Tax-induced p53 inactivation. While Tax can inactivate endogenous p53 function in wild-type MEFs, it fails to inactivate p53 function in p65 knockout MEFs. Importantly, Tax-induced p53 inactivation can be restored by expression of p65 in the knockout MEFs. Finally, we present evidence that phosphorylation of serines 15 and 392 correlates with inactivation of p53 by Tax in T cells. This study provides evidence that the divergent NF-κB proliferative and p53 cell cycle arrest pathways may be cross-regulated at several levels, including posttranslational modification of p53.

Human cells are equipped with signaling pathways to detect and respond to DNA damage and cellular stress. The p53 cascade leads to cell cycle arrest or apoptosis in response to a variety of agents or conditions that cause DNA damage, affect chromosome replication and segregation, or generate inappropriate proliferative signals (32, 38, 50, 59, 66). Cells lacking this response pathway are more susceptible to transformation and resistant to chemotherapeutic agents and exhibit increased genomic instability, allowing them to gain a selective growth advantage during tumor progression (9, 49, 81). The importance of p53 as a tumor suppressor is evident from the fact that over 60% of all human cancers have a mutant or inactive p53 (32).

It is clear that sequence-specific DNA binding, transcriptional activation, regulation of DNA replication, and capacity to induce cellular growth arrest are critical for p53 function (16, 18, 20, 32, 52). Inactivating mutations in p53 have helped to uncover the mechanism by which p53 contributes to tumor suppression. The most frequent class of inactivating mutations consists of mutated residues within the p53 gene that disrupt the structure of the DNA binding domain (51). Inactivation of the sequence-specific DNA binding capacity of p53 abrogates its ability to regulate transcription of target genes involved in growth arrest and apoptosis (11, 52). A second class of p53-inactivating lesions is extragenic and includes proteins that interact with or modify p53. Examples of these proteins are (i) the MDM2 oncoprotein, which binds p53 and facilitates p53 degradation (25, 34, 48), and (ii) the ATM protein, which is involved in triggering the activation of p53 by phosphorylating p53 in response to ionizing radiation (7, 30, 69). The MDM2 oncogene is amplified in multiple tumor types, resulting in the constitutive inhibition of wild-type p53 (46). Likewise, mutation of the ATM gene results in a disease termed ataxia telangiectasia, which causes radiosensitivity due to failure to activate p53 (8, 29–31, 36). A third and less understood class of inactivating lesions involves nuclear exclusion of the wild-type p53 and has been observed in a number of diverse neoplasms (40, 44, 45, 77). The fourth class of inactivation involves viral oncoproteins which inactivate the p53 function. Examples of this include simian virus 40 large T antigen, E1B 55K, E4orf, and HBX (13, 14, 17, 59, 76, 82).

The human T-cell lymphotropic virus type 1 (HTLV-1) is the etiologic agent of an aggressive and fatal disease termed adult T-cell leukemia and of the neurodegenerative disease tropical spastic paraparesis–HTLV-1-associated myelopathy (19, 56, 86; M. Osame, K. Usuku, S. Izumo, N. Ijichi, H. Amitani, A. Igata, M. Matsumoto, and M. Tara, Letter, Lancet i:1031–1032, 1986). The viral transcriptional activator protein, Tax, plays a critical role in cellular transformation (61). Tax has been shown to cause tumors in transgenic mice (10, 22), to cooperate with the ras oncogene in transformation of rodent fibroblasts (75), and to immortalize human lymphocytes when expressed in either a herpesvirus or retrovirus vector (21, 62). Recently, it has been shown that the ability of Tax to activate the NF-κB pathway is critical for T-cell immortalization and factor-independent growth (27, 62).

We have previously shown that Tax can inactivate the tumor suppressor p53. In lymphocytes, Tax does not accomplish this by direct binding but rather through an indirect mechanism involving activation of cellular pathways which lead to constitutive phosphorylation of p53 at serine 15 and serine 392 (55). Importantly, phosphorylation at serine 15 interferes with the interaction of p53 with general transcription factors such as TFIID (55).

In this report, we extend these findings and demonstrate that the mechanism of inactivation in T lymphocytes involves the NF-κB pathway. Inactivation of p53 transcriptional activity is not a result of NF-κB sequestration of the coactivator p300 but rather a result of NF-κB gene activation. Importantly the pattern of hyperphosphorylation at serine 15 and serine 392 of p53, which is linked to its inactivation in HTLV-1-infected cells, is seen when Tax alone is expressed and correlates with the ability of Tax to activate NF-κB and to inactivate p53 function. Further, when Tax-mediated p53 inactivation is inhibited by expression of the IκB(S32/36A) mutant, the hyperphosphorylation at serines 15 and 392 is also inhibited. Finally, the importance of serines 15 and 392 to Tax inactivation of p53 is demonstrated by the serine 15-392 alanine double mutant. Although as transcriptionally active as wild-type p53, the S15, 392A mutant p53 cannot be inactivated by Tax.

MATERIALS AND METHODS

Cell lines.

Jurkat cells were grown in RPMI supplemented with 10% fetal bovine serum and 10 mM glutamine. Both wild-type and p65 knockout (KO) (5) mouse embryo fibroblast cells (MEFs) were kindly provided by Alex Hoffman (CalTech) and were grown in Dulbecco's modified Eagle's medium supplemented with 10% calf serum and 10 mM glutamine.

Transfections.

The cells were fed 1 day prior to transfection. Jurkat cells (5 × 10⁶) were transfected by the Superfect method (Qiagen). Transfections were performed as described by the manufacturer. MEFs (60 to 80% confluent) were transfected by the Effectene (Qiagen) method as described by the manufacturer. The Tax constructs (wild type, M22, M47, and V89A) were described previously (24, 70). The wild-type p53 constructs were kindly provided by Jennifer Pietenpol (52) and Karen Vousden (2). The phosphorylation mutant p53 constructs, S15A, S37A, and S392A, were also provided by Karen Vousden (34). The S15, 392A double mutant was constructed by ligation of the NcoI N-terminal fragment of S15A to the NcoI fragment of the S392A construct. The dominant-negative IκB mutant [IκB(S32/36A)] construct, pCMV-p65, and pCMV-p65(1–312) were kindly provided by Warner Greene (72, 73). The p300 expression plasmid was provided by Y. Nakatani. All transfections were adjusted for efficiency using a cytomegalovirus–beta-galactosidase control plasmid.

Western blot analysis.

For Western blot analysis after transfection, protein lysates were prepared by lysis with the luciferase extraction buffer (55), concentrations were determined by Bradford assay (Bio-Rad), and 50 μg was separated by electrophoresis on 4 to 20% Tris-glycine gels (Novex). The proteins were then transferred to nylon membranes (Immobilon), and analyzed for the presence of p53 with DO-1 (Oncogene Research) or for Tax with Tab172. Protein loading was assessed with an anti-β-actin antibody (Santa Cruz). Protein lysates for analysis of p53 phosphorylation were prepared by disrupting the cells in 50 mM Tris, 120 mM sodium chloride, 5 mM EDTA, 0.5% Nonidet P-40, 50 mM sodium fluoride, and 0.2 mM sodium vanadate. The lysates were incubated on ice for 20 min and then cleared by centrifugation (10,000 × g) at 4°C for 10 min. Samples were then treated as described above. Detection of phosphorylated residues on p53 was done by immunoprecipitation and then Western blot analysis. Briefly, 500 μg of whole-cell extract was immunoprecipitated with DO-1 antibody. Detection of phosphorylated residues was performed using phosphospecific antibodies to P-Ser15 and P-Ser392 (55).

RNase protection assay.

After transient transfection of Jurkat cells (15 × 10⁶), total cellular RNA was prepared as described by the manufacturer (Qiagen). Ten micrograms of total RNA was then used in an RNase protection assay as previously described (Pharmigen) (R. Mahieux, C. A. Pise-Masison, P. Lambert, C. Nicot, L. DeMarchis, A. Gessain, P. Green, W. W. Hall, and J. N. Brady, submitted for publication).

RESULTS

Tax mutants which fail to activate NF-κB fail to inactivate p53 in T lymphocytes.

Numerous studies have demonstrated that Tax is capable of activating both the NF-κB and CREB and activating transcription factor (ATF) transcriptional pathways. Interestingly, point mutations in defined domains of the Tax protein can abrogate the activation of one pathway without affecting the other (24, 64, 70). In previous studies, we and others have demonstrated that expression of Tax alone is sufficient to inactivate the transcriptional activation function of p53 (47, 54, 74, 79). To further define the mechanism of this Tax-induced inactivation, we used Tax mutants which were specifically defective in CREB activation, NF-κB activation, or binding to CREB-binding protein (CBP)/p300. We first verified the expression level and the phenotype of each mutant used in this study by transfection assays in Jurkat T lymphocytes. Jurkat T cells were transfected with pcTax expression plasmid and either the HTLV-Luc or the NF-κB–Luc reporter constructs. The cells were harvested 16 to 24 h after transfection, and extracts were prepared and assayed for either Tax expression or luciferase activity. Western blot analysis of the transfected cells demonstrated that the expression levels of wild-type and Tax mutant proteins were comparable (Fig. 1A, inset). Consistent with published studies, the wild-type Tax protein was able to stimulate transcription from either the CREB-dependent HTLV-1 long terminal repeat (LTR) luciferase reporter or the NF-κB-dependent luciferase reporter (Fig. 1A and B). As expected, the Tax M22 mutant can activate CREB-driven transcription on the HTLV-1 LTR (Fig. 1A), but it fails to activate NF-κB-driven transcription (Fig. 1B). In contrast, the CREB activation-deficient mutant Tax M47 fails to activate the HTLV-1 promoter but does transactivate the NF-κB reporter plasmid. Consistent with the original report (24), the Tax V89A mutant, which does not interact with CBP/p300, failed to activate the HTLV-1 CREB promoter (Fig. 1A). The V89A mutant did activate the NF-κB reporter, increasing expression by 30-fold (Fig. 1B). The fact that the V89A activity was slightly less than that of wild-type Tax or M47 suggests that CBP/p300 binding may be necessary for full NF-κB activation by Tax. Alternatively, the V89A domain of Tax may interact with a different protein, and it is the impaired ability of V89A to interact with this protein that affects NF-κB activation.

FIG. 1.

Tax M22 is not capable of inactivating p53 transactivation function. The activity of wild-type, M22, M47, and V89A Tax constructs were tested on the CREB-driven HTLV-1 LTR reporter construct HTLV-Luc (A), the NF-κB-driven reporter NF-κB–Luc (B), or the p53-dependent reporter PG13-Luc (C). Activity is expressed as light units and adjusted for transfection efficiency using a beta-galactosidase transfection control plasmid. These same extracts were assayed for the levels of Tax and p53 expression (A, inset, and C, bottom, respectively). Equal loading of samples was determined by detection with an anti-tubulin antibody (data not shown).

When examining the abilities of these Tax mutants to inactivate p53 function, we found that wild-type Tax, Tax V89A, and Tax M47 could inactivate p53 transactivation function in Jurkat T cells 5- to 10-fold (Fig. 1C). In contrast, the Tax mutant M22, which fails to activate NF-κB-driven transcription, did not inactivate p53. As previously seen (54), Tax had no effect on PG13 reporter activity. Western blot analysis of the cell extracts demonstrated that the levels of p53 expression were similar for wild-type and Tax mutant transfections (Fig. 1C, bottom). These results demonstrate that the differences in p53 activity were not due to differences in p53 protein expression. The results further show that Tax stabilization of p53 is not directly linked to p53 activation, since similarly increased levels of p53 protein were observed in wild-type Tax-, M22-, V89A-, and M47-transfected cells. These results suggest that activation of the NF-κB pathway by Tax is important for p53 inactivation in Jurkat T lymphocytes.

IκB mutant interferes with Tax inactivation of p53.

To further examine the role of NF-κB in the inactivation of p53 by Tax in lymphocytes, we utilized an IκB mutant which contained serine-to-alanine substitutions at amino acids 32 and 36 (72). This mutation blocks the phosphorylation and subsequent degradation of the IκB inhibitor, leading to inactivation of the NF-κB pathway. As shown in Fig. 2A, expression of the IκB mutant [IκB(S32/36A)] allows recovery of p53 activity in the presence of Tax in a dose-dependent manner. In the absence of IκB(S32/36A), Tax decreased p53 activity fivefold (Fig. 2A, lane 1 versus lane 2). The addition of increasing amounts of the dominant-negative IκB mutant reversed the Tax inhibition of p53 function (Fig. 2A, lanes 3 to 5). These results provide further evidence that the NF-κB pathway is critical for inactivation of p53 by Tax in T cells.

FIG. 2.

The dominant IκB mutant can recover p53 activity in the presence of Tax by blocking NF-κB activation. (A) Representative graph of the p53 activity assayed on the PG13-Luc reporter construct alone (lane 1) or in the presence (+) of Tax (6 μg; lane 2) with increasing amounts of IκB(S32/36A) (0.5, 1, and 3 μg; lanes 3, 4, and 5, respectively). Lanes 6 and 7 show the effect of Tax (6 μg) or IκB(S32/36A) (3 μg) alone on the reporter construct. Lane 8 shows the effect of IκB(S32/36A) expression on p53 function. −, not present. Below is a Western blot analysis of 50 μg of extract using the DO-1 antibody to detect p53. (B) Activity of Tax on NF-κB activation in the absence (lane 3) or presence (lane 4) of IκB(S32/36A). Lanes 1 and 2 represent transfection of the reporter with a control plasmid or the IκB(S32/36A) plasmid, respectively. The results are expressed as percent activation and are representative of two independent experiments. (C) Activation of the HTLV-Luc reporter by Tax in the presence (lane 4) or absence (lane 3) of IκB(S32/36A) is expressed as percent activation and is representative of two independent experiments. As controls, the reporter construct was transfected with a control vector (lane 1) or IκB(S32/36A) (lane 2). Transfections were performed in Jurkat T cells as described in the text. Error bars indicate standard deviations.

As previously reported, cotransfection of Tax with p53 causes p53 stabilization (Fig. 2A, bottom, lane 5). Interestingly, coexpression of the IκB(S32/36A) mutant also appears to stabilize p53 (Fig. 2A, bottom, lane 8). There is an additive effect on p53 levels upon expression of both Tax and IκB(S32/36A) (Fig. 2A, bottom, lanes 6 to 8).

To demonstrate the specificity of the IκB inhibition, the IκB mutant was cotransfected with Tax and the HTLV-1 LTR or NF-κB reporter plasmid. Figure 2B demonstrates that cotransfection of the IκB mutant inhibits the ability of Tax to activate transcription from an NF-κB-driven promoter. In contrast, the IκB mutant has no effect on the ability of Tax to activate transcription from the CREB-driven HTLV-1 LTR promoter (Fig. 2C). In fact, we observed a better Tax activation of the HTLV-1 LTR in the presence of the IκB mutant. This result suggests that blocking the NF-κB activation pathway may allow more efficient activation of the CREB pathway. Since CREB activation is a nuclear event, this important control demonstrates that overexpression of IκB(S32/36A) does not alter nuclear localization of Tax.

Overexpression of the coactivator p300 fails to rescue p53 from Tax inhibition in lymphocytes.

Since CBP/p300 has been shown to be important in p53 transactivation, we tested whether the induction of NF-κB by Tax could result in a squelching of p300. A plasmid encoding p300 was cotransfected along with Tax and p53 into Jurkat T cells. Transfection of increasing amounts of p300 failed to rescue p53 activity in the presence of Tax (Fig. 3A, lanes 7 to 9).

FIG. 3.

Overexpression of p300 cannot rescue p53 activity in the presence of Tax. (A) Transient transfection of Jurkats with the PG13-Luc reporter construct either alone (first lane), with (+) p300 (second lane), or with Tax (third lane) shows the dependence of this construct on p53 activation. Cotransfection of PG13-Luc with p53 alone (fourth lane) or with p300 (fifth lane) resulted in high p53 activity. Cotransfections of the reporter and p53 with Tax (sixth lane) or Tax and increasing amounts of p300 (1, 3, and 6 μg), shown in the last three lanes, were done. The activity is expressed as relative luciferase units for the combination of at least three independent experiments. (B) Using transient transfection of Jurkats, the effect of p300 overexpression on Tax activation of the HTLV-chloramphenicol acetyltransferase (CAT) reporter was assayed. The HTLV-CAT reporter was transfected either alone (lane 1), with Tax (lane 2), or with p300 (lane 3). The activation of the reporter by Tax and increasing amounts of p300 (lanes 4, 5, and 6) was determined. The activity is expressed as relative CAT activity and represents data from at least two independent experiments. The transfection efficiency for each sample was determined using a beta-galactosidase reporter construct. The error bars indicate standard deviations. −, absent.

The failure of p300 to rescue p53 activity is not due to a failure of the transfected p300 to function in these cells. Transfection of increasing amounts of p300 in the presence of Tax resulted in a significant increase in transcriptional activity from the HTLV-1 LTR (Fig. 3B, lanes 2 and 4 to 6). These results indicate that the induction of NF-κB activity by Tax has a direct role on p53 inactivation and is not merely a squelching of the CBP-p300 coactivators by activated NF-κB. These results are also consistent with the results presented in Fig. 1C, which demonstrate that Tax inhibition of p53 is not significantly affected by the V89A mutation which knocks out CBP-p300 binding (24).

Only a modest stimulation of p53 activity was seen upon addition of exogenous p300 (Fig. 3A, lane 5). This result is observed independently of the p300 concentration transfected (from 0.1 to 8 μg) (data not shown). Since the level of endogenous p300 in our Jurkat T cells is high (data not shown), we interpret this result as again indicating that p300 in Jurkats is not limiting and thus additional p300 has little effect.

KO of p65 and p50 in MEFs abrogates the ability of Tax to inactivate p53.

We have also examined the ability of Tax to inhibit endogenous p53 function in either wild-type or p65 KO MEFs (5). Similar to the results observed in Jurkat lymphocytes, Tax is capable of inhibiting p53 transactivation in MEFs which express the wild-type p53 and p65 NF-κB subunit (Fig. 4A, compare lanes 1 and 2). In contrast, Tax was not able to inhibit p53 function in the p65 KO cells (Fig. 4A, lanes 3 and 4).

FIG. 4.

The p65 subunit of NF-κB is important for Tax-induced p53 inactivation. (A) Transient transfections of wild-type (WT; lanes 1 and 2) or p65 KO (lanes 3 and 4) MEFs were done using the Effectene transfection method (Qiagen). Cotransfections of PG13-Luc either with a control vector (lanes 1 and 3) or pcTax (lanes 2 and 4) were done; samples were harvested 16 to 24 h posttransfection and then assayed for luciferase activity using a Berthold luminometer. Activity is expressed as percent of p53 activity and represents data from at least three independent experiments. A Western blot of the p53 protein levels for each sample is shown. (B) MEF p65 KO cells were cotransfected with PG13-Luc with (+) or without (−) pcTax (0.1 μg) and with vector (first two lanes), the pCMV-p65 (0.1 μg; second two lanes), or the pCMV-p65(1–312) (0.1 μg; last two lanes) mutant. Activity is expressed as percent of p53 activity. (C) NF-κB activity was measured by transient cotransfection of the NF-κB-Luc reporter construct with the control vector (0.1 μg; lane 1), pcTax (0.1 μg; lane 2), pCMV-p65 (0.1 μg; lane 3), or pCMV-p65(1–312) (0.1 μg; lane 4) in p65 KO cells using the Effectene (Qiagen) method of transfection. These results are a representation of at least four independent experiments. The error bars indicate standard deviations.

To demonstrate that the inability of Tax to inactivate p53 function was directly related to p65 expression, a plasmid encoding p65 was cotransfected along with Tax and the p53 reporter into the p65 KO cells. The results of this experiment demonstrate that expression of p65 restores the ability of Tax to inactivate p53 (Fig. 4B). Transcription activation of p65 is important for Tax-mediated inactivation of p53, as demonstrated using the p65 mutant p65(1–312), which retains the N-terminal Rel homology domain but lacks the C-terminal transactivation domain (72). p65(1–312) does not restore Tax-induced p53 inactivation (Fig. 4B). Importantly, the ability of p65 to restore Tax-induced inactivation of p53 directly correlates with NF-κB transcriptional activity in these cells. Figure 4C demonstrates that p65 expression in the KO cells restores NF-κB activation (lane 3) whereas Tax (lane 2) and p65(1–312) (lane 4) cannot. Western blot analysis (Fig. 4B, bottom) of the p53 levels in transfected cells shows that p65 transcriptionally activates expression of endogenous p53, as previously shown by Wu and Lozano (84), and thus an increase in the p53 protein level is observed.

Tax inactivation of p53 correlates with phosphorylation of p53 at Ser15 and Ser392 in lymphocytes.

We have previously shown that inactivation of p53 in HTLV-transformed cells is linked to its constitutive phosphorylation at Ser15 and Ser392 (55). To confirm that phosphorylation of p53 was critical for Tax-induced inactivation in lymphocytes, we utilized expression vectors which encode p53 proteins containing single mutations at S15A, S37A, and S392A or a double mutation at S15, 392A. The plasmids were transfected into Jurkats with the p53 luciferase reporter in the presence or absence of the Tax plasmid (Fig. 5). As described above, Tax was able to inhibit wild-type p53 function in Jurkat T lymphocytes (Fig. 5, lanes 2 and 3). Tax was also capable of inactivating p53 mutated at S15A, S37A, and S392A (Fig. 5, lanes 4 and 5, lanes 8 and 9, and lanes 10 and 11, respectively). In contrast, when the S15, 392A p53 mutant was cotransfected into the Jurkat cells, Tax could not suppress its transcriptional activity (Fig. 5, lanes 6 and 7). These results suggest that both serine 15 and serine 392 are important for Tax inactivation of p53 function. Interestingly, as seen with the Tax mutants, the ability of Tax to stabilize wild-type and mutant p53s was independent of p53 inactivation (Fig. 5B).

FIG. 5.

Effect of phosphorylation site mutations on Tax-induced p53 inactivation. (A) Jurkat T cells were transiently cotransfected with PG13-Luc (1 μg) in the presence (+) or absence (−) of Tax (4 μg) and with wild-type (WT) p53 (1 μg; lanes 2 and 3), S15A (1 μg; lanes 4 and 5), S15, 392A (1 μg; lanes 6 and 7), S37A (1 μg; lanes 8 and 9), or S392A (1 μg; lanes 10 and 11). Activity is expressed as percent (+ standard deviation) of control p53 activity. The activities of all p53 constructs were equal to or greater than wild-type p53 activity. These results are from at least three independent experiments. (B) Representative Western blot with DO-1 antibody to determine p53 levels in each transfected sample.

HTLV-1 Tax inhibits wild-type p53 activity but not S15, 392A mutant p53 on endogenous cellular promoters.

We next examined the effect of Tax on expression of endogenous p53-responsive genes using an RNase protection assay after transient transfection of either vector alone, wild-type p53 in the presence and absence of Tax, or the S15, 392A p53 mutant in the presence or absence of Tax (Fig. 6). Transfection of the vector alone (lane 1) did not cause an increase in expression of the p53-responsive Bax or p21 genes. Upon transfection of the p53 expression plasmid (lane 2), a significant induction of Bax and p21 was observed. Expression of the Tax protein (lane 3) was able to greatly inhibit the transactivation function of p53 on these endogenous genes. In contrast, Tax expression did not affect the induction of Bax and p21 by the S15, 392A p53 mutant (lanes 4 and 5). The lower level of Bax induction by the mutant p53 (lanes 1 and 2 versus lanes 4 and 5) represents experimental variation in the Bax induction level and should not be interpreted to be specific to the mutant p53. The GAPDH (glyceraldehyde-3-phosphate dehydrogenase) gene was used to demonstrate that the hybridization efficiencies were roughly equivalent in all samples. Figure 6, bottom, shows the level of p53 protein expression from each transfection.

FIG. 6.

Tax inhibits p53 function on endogenous gene promoters. Jurkat T cells were transiently transfected with vector (lanes 1 and 4), wild-type (WT) p53 (lane 2), wild-type p53 in the presence of Tax (WTp53 + Tax) (lane 3), or the S15, 392A p53 mutant in the absence (S15,392A) (lane 5) or presence (S15,392A + Tax) (lane 6) of Tax. Total cellular RNA was extracted and subjected to RNase protection using probes for the p53-responsive Bax and p21 genes. The GAPDH gene was used as a control to equilibrate the amounts of RNA used. At the bottom is a Western blot analysis using anti-DO-1 antibody to determine the p53 protein expression in the transfected cells at the time of harvest.

Hyperphosphorylation of Ser15 and Ser392 correlates with Tax inactivation in Jurkats.

We next examined the phosphorylation pattern of transfected p53 in the presence and absence of Tax. In order to compare the relative levels of p53 phosphorylation, it was important to analyze similar amounts of p53. Extracts from transfected cells were immunoprecipitated with a limiting amount of the p53 antibody DO-1. Western blot analysis of the immunoprecipitates demonstrates that similar amounts of p53 were precipitated (Fig. 7A, bottom). When the same blot was probed with antibody specific for either Ser15 (top) or Ser392 (middle), an increase in Ser15 and Ser392 phosphorylation was observed in the presence of Tax (Fig. 7A, lane 2). Importantly, phosphorylation of p53 was observed only when p53 was inactivated. Cotransfection of p53 with wild-type, M47, or V89A Tax resulted in the phosphorylation of Ser15 and Ser392 (Fig. 7A, lanes 2 to 4). In contrast, low levels of Ser15 and Ser392 phosphorylation were observed when p53 was cotransfected with the NF-κB-deficient M22 Tax (Fig. 7A, lane 5). These results clearly link phosphorylation of Ser15 and Ser392 with NF-κB-dependent Tax-induced inactivation of p53 function. It should be noted that upon longer exposure of the anti-Ser15p Western blot, there was a low level of p53 reactivity in either the control vector (Fig. 7A, lane 1) or M22 (Fig. 7A, lane 5) cotransfection sample. These results are consistent with the results of Shieh et al. (67), which show that transfection alone can induce p53 phosphorylation.

FIG. 7.

Serine 15 and 392 phosphorylation of p53 correlates with Tax-induced p53 inactivation. (A) Jurkat T cells were transiently transfected with (+) wild-type p53 in the presence of control plasmid (−; lane 1), wild-type Tax (lane 2), Tax M47 (lane 3), Tax V89A (lane 4), or Tax M22 (lane 5); the cells were harvested 24 h after transfection, and 500 μg of whole-cell extract was incubated with DO-1 antibody to p53. The immunoprecipitated complexes were resolved on Tris-glycine gels and transferred to nylon membranes for Western blot analysis using anti-Ser15P (top) anti-Ser392P (middle), or anti-DO-1 (bottom) p53 antibody. (B). Jurkat T cells were transiently transfected with vector (lane 1); with IκB(S32/36)A (lane 6); or with wild-type p53 alone (lane 3), with Tax (lane 4), or with Tax and IκB(S32/36A) (lane 5). The cells were harvested 24 h after transfection, immunoprecipitated with anti-DO-1 antibody, and subjected to Western blot analysis as described above.

To determine whether inhibition of Tax-mediated p53 inactivation by the dominant IκB(S32/36A) mutant was also linked to phosphorylation at serines 15 and 392, Western blot analysis of cotransfections into Jurkats was performed. As shown in Fig. 7B, the inhibition of Tax-induced p53 inactivation by IκB(S32/36A) correlates with decreased phosphorylation at serine 15 and serine 392. Hyperphosphorylation of p53 at serine 15 and serine 392 was observed in the presence of Tax (Fig. 7B, lane 3), but not when p53 was cotransfected with Tax and the IκB(S32/36A) mutant (Fig. 7B, lane 4). Taken together, these results strongly suggest that Tax-induced p53 inactivation in lymphocytes is dependent on the ability of Tax to activate the NF-κB pathway which leads to hyperphosphorylation at serines 15 and 392. In addition, the ability to phosphorylate both serine 15 and serine 392 is critical for Tax-induced p53 inactivation.

DISCUSSION

Several lines of evidence presented in this study suggest that the ability of Tax to inactivate p53 depends upon activation of NF-κB. First, a Tax mutant that fails to activate NF-κB failed to inactivate p53. Tax mutant M22 contains a 2-amino-acid substitution at positions 130 and 131 which destroys the protein's ability to activate NF-κB-driven promoters (70). In the p53 inactivation studies, the M22 mutant was not able to inactivate p53 function. Importantly, the transcriptional activity of the M22 Tax protein on CREB-driven promoters is retained. Second, the ability of Tax to inhibit p53 activity is suppressed by overexpression of an IκB mutant [IκB(S32/36A)] which prevents NF-κB activation by blocking NF-κB nuclear translocation (72). Third, the studies with the p65 KO MEFs suggest that RelA/p65 is required for Tax inactivation of p53. In contrast to wild-type MEFs, Tax was unable to inactivate p53 function in the p65 KO cells, which retain expression of all other NF-κB family members (5). Further, cotransfection of a transcriptionally active p65 expression plasmid into the KO cells enabled Tax to inactivate p53 function. Taken together, these results provide strong evidence that activation of the NF-κB pathway is important for Tax-mediated p53 inactivation.

The effects of NF-κB on p53 activity are complex and likely depend upon the relative levels of expression of the two proteins. Several groups have reported that NF-κB is important for induction of p53. For example, in HCT116 cells, Hellin et al. have reported that the p53 activating signal induced by daunomycin is partially regulated by NF-κB (26). Similarly, Wu and Lozano (84) found that in HeLa cells, NF-κB activation increased p53 promoter activity. In contrast, several groups have recently reported that NF-κB activation blocks p53 transactivation by sequestering the coactivator CBP/p300 (60, 80, 83). The results presented in this study suggest that activation of NF-κB, independently of the potential squelching effect, may lead to inactivation of p53. In contrast to results obtained in overexpression systems (60, 80, 83), when p65 is transfected into p65 KO MEFs at concentrations that do not inactivate p53 function, Tax regains its ability to inactivate p53. The fact that Tax does not alter the level of p65 protein in the cotransfection assay suggests that the decrease in p53 function is not simply due to squelching of the coactivator p300 (data not shown).

The Tax V89A mutant, which fails to bind CBP/p300 (24), inactivated p53 transactivation in lymphocytes, and importantly, overexpression of the coactivator p300 did not alleviate the Tax-mediated p53 inactivation. These results are in contrast to those reported by Suzuki et al. ((74) and Van Orden et al. (79), which indicate that p53 inactivation is due to binding of Tax to p300. It is important to point out that Suzuki et al. (74) did not use Tax mutants to definitively show that Tax binding to p300 is important for the inactivation. Second, neither group was able to demonstrate recovery of p53 transactivation by overexpression of p300. While we would agree that competition for limiting coactivators may play a role in regulation of some transcription factors and promoters (78), our results are not consistent with the hypothesis that the ability of Tax to inhibit p53 activity in lymphocytes is due to competition for p300. At present, however, we cannot rule out the possibility that p53 inactivation by Tax is not due to the sequestration of another limiting cofactor.

It was recently reported that the CREB-ATF activation function of Tax is important for inactivation of p53 (47). The majority of the studies conducted by Mulloy and colleagues were done with U2OS, Calu-6, and HeLa/Tat cells in contrast to our studies, which were done primarily with lymphocytes. Our present data suggest that Tax inactivates p53 by different pathways in different cell types, including Jurkat, HeLa, H1299, and Saos-2 cells and MEFs (data not shown). The NF-κB pathway is utilized primarily in lymphocytes but is also operative in a limited number of other cell types, including MEFs. In contrast, Tax utilizes the CREB-ATF pathway for inactivation of p53 in most nonlymphocyte cells tested. Thus, we feel that the primary difference between the results of the studies of Mulloy et al. (47) and the present study may reflect the Tax activities in different cell lines. At present, we cannot explain the differences between our results and those of Mulloy et al. in the Jurkat cells. We have, however, approached the analysis from several independent angles, which include Tax mutants, IκB inhibitors, and p65 KO cells. The results from each of the assays suggest that Tax must activate the NF-κB pathway to inactivate p53 in lymphocytes and MEFs.

It is reasonable to assume that the ability of Tax to inactivate p53 may be linked to the transformation properties of the HTLV-1 Tax protein. Although Smith and Greene initially linked the CREB activation function of Tax with transformation of Rat2 cells (71), more recent reports firmly link the NF-κB activation function to Tax transformation. Yamaoka et al. (85) and Matsumoto et al. (41) have demonstrated that the NF-κB pathway appears to be important for the transformation of Rat1 cells. Additional support for the involvement of NF-κB in transformation by Tax in rodent cells was demonstrated by Kitajima et al. (31a) using antisense oligonucleotides to NF-κB that could inhibit the proliferation of Tax-transformed tumor cells from Tax-transgenic mice. Similarly, in human primary cells, Akagi et al. (1) demonstrated that the Tax M22 mutant failed to immortalize primary lymphocytes when transduced by a retroviral expression vector. In a separate study, Iwanaga et al. (27) analyzed the effect of constitutive expression of wild-type, M47, or M22 Tax in the interleukin-2 factor-dependent cell line CTLL-2. Expression of wild-type and M47 Tax allowed interleukin-2-independent growth, while M22 could not. More recently, Robek and Ratner (62) have reported that infectious molecular clones of HTLV-1 (ACH) containing wild-type or M47 Tax could immortalize primary human lymphocytes. In contrast, the ACH molecular clone containing M22 Tax could not immortalize primary peripheral blood mononuclear cells. It should be noted that in contrast to the above-mentioned studies, Rosin et al. (63) reported that the Tax mutant S258, which is defective for NF-κB activation, retained the ability to immortalize primary peripheral blood lymphocytes. The interpretation of these experiments, however, is not straightforward, since the contribution of the herpesvirus saimiri vector is unclear.

Due to the pleiotropic nature and potency of the p53 response, its function is tightly regulated in the cell. To this end, p53 function is controlled at the levels of transcription, translation, protein turnover, cellular compartmentalization, and association with other proteins (28, 32, 49, 58). More recently, it has been shown that p53 function can be regulated by multisite posttranslational modifications (28, 32, 49, 58). Phosphorylation, acetylation, and glycosylation have all been shown to occur on p53 and potentially affect its activity and interaction with other proteins (15, 28, 42, 65, 68). The types of modifications that affect p53 are likely to be stress, species, and cell type specific.

The p53 protein is phosphorylated on numerous serines in both the N- and C-terminal domains. In this report, we show by using p53 mutants that phosphorylation of both Ser15 and Ser392 of p53 is important for Tax-mediated inactivation in lymphocytes. These results are consistent with previous reports from this laboratory which demonstrated that phosphorylation at Ser15 alone in the N terminus of p53 destroys TFIID binding (55). We have also recently reported that phosphorylation at serine 15 of p53 in vitro will enhance the association of p53 with CBP/p300 (35). However, in the HTLV-1-transformed cells which have hyperphosphorylation on serine 15, an in vivo association with p53 and CBP/p300 has not been observed (data not shown). This argues that phosphorylation alone does not govern protein-protein association and that other factors are involved.

Our studies provide some of the first experimental evidence that the divergent NF-κB proliferative and p53 cell cycle arrest pathways may be cross-regulated at several levels, which include posttranslational modification of p53. This work also reflects the importance of cross talk between the N- and C-terminal domains of p53, as well as pointing to the significance of the coordination of specific phosphorylation sites on p53 and its function.

A number of kinases have been implicated in phosphorylation of human p53 in vitro, including casein kinase I (serines 6 and 9 [43]), DNA-dependent protein kinase (DNA-PK) (serines 15 and 37 [37, 67]), ATM and ATR (serine 15 [3, 7, 62]), CDK-activating kinase (serine 33 [34]), cdk2 and cdc2 (serine 315 [6, 57]), PKC (serine 378 [12]), and casein kinase II (serine 392 [23]). Along these lines, it has been shown that DNA-PK and ATM participate in the sustained activation of NF-κB following DNA damage (4, 39, 53). It is possible that Tax activation of the NF-κB pathway includes induction of ATM or DNA-PK kinases and ties into the p53 inactivation pathway. We are currently investigating whether Tax affects the activities or specificities of these kinases. Alternatively, Tax, via NF-κB, could alter the expression of a yet-unidentified kinase that in turn phosphorylates p53.