Post-translational modifications and activation of p53 by genotoxic stresses


E. Appella, Laboratory of Cell Biology, National Cancer Institute, National Institutes of Health, Bethesda, MD 20892, USA. Fax: + 301 496 7220, Tel.: + 301 402 4177, E-mail:


In unstressed cells, the tumor suppressor protein p53 is present in a latent state and is maintained at low levels through targeted degradation. A variety of genotoxic stresses initiate signaling pathways that transiently stabilize the p53 protein, cause it to accumulate in the nucleus, and activate it as a transcription factor. Activation leads either to growth arrest at the G1/S or G2/M transitions of the cell cycle or to apoptosis. Recent studies point to roles for multiple post-translational modifications in mediating these events in response to genotoxic stresses through several potentially interacting but distinct pathways. The ≈ 100 amino-acid N-terminal and ≈ 90 amino-acid C-terminal domains are highly modified by post-translational modifications. The N-terminus is heavily phosphorylated while the C-terminus contains phosphorylated, acetylated and sumoylated residues. Antibodies that recognize p53 only when it has been modified at specific sites have been developed, and studies with these reagents show that most known post-translational modifications are induced when cells are exposed to genotoxic stresses. These recent results, coupled with biochemical and genetic studies, suggest that N-terminal phosphorylations are important for stabilizing p53 and are crucial for acetylation of C-terminal sites, which in combination lead to the full p53-mediated response to genotoxic stresses. Modifications to the C-terminus inhibit the ability of this domain to negatively regulate sequence-specific DNA binding; additionally, they modulate the stability, the oligomerization state, the nuclear import/export process and the degree of ubiquitination of p53.


alternative reading frame

ATM kinase

ataxia telangiectasia mutated kinase


actinomycin D

ATR kinase

A–T related kinase


bovine papilloma virus


cyclin-dependent kinase-activating kinase


CREB binding protein




casein kinase


checkpoint kinase


cyclin-dependent kinase


deferoxamine mesylate


dsDNA-activated protein kinase


embryonic stem


global genome repair


histone acetyltransferases


histone deacetylase


Jun N-terminal kinase


mitogen-activated protein kinase


mouse double minute


nucleotide excision repair




p300/CBP associated factor


dsRNA-activated protein kinase


promyelocytic leukemia


representational difference analysis


small ubiquitin-related modifier


transcription-coupled repair


trichostatin A.

Loss of p53 function is thought to play a central role in the development of cancer. While the p53 protein normally is short-lived and is maintained at low levels in unstressed mammalian cells, p53 accumulates in the nucleus and is activated as a transcription factor in response to damaged DNA, nucleotide depletion, hypoxia, and several other genotoxic stresses [1,2]. Activation of p53 induces or inhibits the expression of more than 150 genes including CDKN1A (p21, WAF1, CIP1), GADD45, MDM2, IGFBP3 and BAX[3,4] that mediate arrest of mammalian cells at one of two major cell cycle checkpoints, in G1 near the border of S phase, or in G2 before mitosis, or that induce other responses including apoptosis, a program of cell death that also depends on p53. Arrest of cell cycle progression is thought to provide time for the repair of DNA damage, and recent results suggest that p53 modulates DNA repair processes [5,6]. Alternatively, DNA damage may initiate apoptosis through p53 transcription-dependent and independent mechanisms [1]. The biochemical links between p53, G1 arrest, and apoptosis are cell type dependent, but neither the molecular role of specific post-translational modifications nor the cellular targets are clearly elucidated.

In this review, we focus primarily on recent data, highlighting the effects of post-translational modifications on transcriptional activation by p53 and on the regulation of p53 protein levels in stressed and unstressed cells. Moreover, we summarize current knowledge of the protein kinases, phosphatases, acetyltransferases and deacetylases that use p53 as a target, and we integrate this information with results from altering specific modification sites in p53.


Human p53 has been reported to be post-translationally modified on at least 18 sites (Fig. 1)[2]. Polyclonal and monoclonal antibodies that recognize specific modified sites in human p53 have been produced by several laboratories, and increased phosphorylation in response to DNA damage has been demonstrated at most modification sites by using these affinity-purified antibodies in Western immunoblot experiments. Seven serines and two threonines in the N-terminal domain of p53, specifically Ser6, 9, 15, 20, 33, 37, 46 and Thr18, and 81, are phosphorylated in response to exposing cells to ionizing radiation or UV light. Recently, Thr55 was found to be phosphorylated in unstressed cells [7] and dephosphorylated after DNA damage (X. Liu, Department of Biochemistry, University of California, Riverside, CA, USA, personal communication). Thus, all N-terminal serines and threonines in the first 89 residues of human p53 may be phosphorylated or dephosphorylated in response to one or more stress conditions. In the C-terminal regulatory domain, Ser315 and Ser392 are phosphorylated, Lys320, 373 and 382 are acetylated, and Lys386 is sumoylated in response to DNA damage, while Ser376 and 378 were reported to be constitutively phosphorylated in unstressed cells [8].

Figure 1.

DNA damage-induced post-translational modifications to human p53. The bar represents the 393 amino-acid p53 polypeptide; regions associated with transactivation (TA), sequence-specific DNA binding (DBD), nuclear localization (NLS), tetramerization (TET), and DNA-mediated negative regulation of specific DNA binding (REG) are indicated (not to scale). The positions of known phosphorylations and acetylations are represented by ovals and squares, respectively; the number of the amino acids modified also is given, and possible modifying enzymes are listed. The response to ionizing radiation is believed to initiate from DNA strand breaks that are sensed directly or indirectly by ATM, which phosphorylates p53 directly and activates other kinases (e.g. Chk1 and Chk2) that phosphorylated p53 and MDM2. After exposure to UV light, ATR, p38MAPK and the CK2/FACT complex are activated and phosphorylate a distinct but overlapping set of residues. p53 and MDM2 are modified by conjugation to SUMO-1 through an E1/Ubc9-mediated mechanism. DNA damage suppresses transfer of SUMO-1 to MDM2, thereby preventing self-ubiquitinization, which stabilizes MDM2 and increases the rate of p53 degradation. Phosphorylation of p53 at Thr55 destabilizes p53, and DNA damage also suppresses this post-translational modification (see text). We have not observed a previously reported ATM-dependent dephosphorylation of Ser376 [8], suggesting this is not a universal response to DNA damage. PCAF and CBP/p300 are acetyltransferases involved in chromatin remodeling. Acetylation of p53 C-terminal residues is mediated via a cascade initiated by N-terminal phosphorylations that modulate acetyltransferase binding to p53 [21]. Similarly, phosphorylation of Ser6 and Ser15 may facilitate phosphorylation of Ser9 and Thr18 by CK1 [11,12].

Although post-translational modifications at most sites occur in response to stress, clear differences in responses at individual sites to different agents have been observed. For example, in response to ionizing radiation, increased phosphorylation at Ser6, 9 and 15 was observed as early as 30 min after treatment [9–12] while exposure to UV light induced a less rapid but more long-lived increase in the phosphorylation of these sites. For Ser15, increased phosphorylation was seen in response to cisplatin, camptothecin (CPT), arsenite, genistein, deferoxamine mesylate (DFX, which mimics hypoxia), but not to actinomycin D (ActD) [13–16]. Moreover, in human fibroblasts undergoing replicative senescence or ras-induced premature senescence, Ser15 phosphorylation also was significantly increased [17,18], suggesting that phosphorylation at this site may be one of the critical signals through which the p53 response to stress is regulated. Certain mono- and polyclonal antibodies that overlap the DO-1 epitope (residues 20–25) are capable of detecting phosphorylation at Thr18 or Ser20 [11,19]. Using these antibodies, it has been shown that Thr18 and Ser20 are significantly phosphorylated in cells subjected to DNA damage and in human breast cancers that express wild-type p53 [19]. Exposure of cells to both ionizing radiation and UV light strongly induced Ser20 phosphorylation as early as 1–2 h after exposure, while Thr18 phosphorylation clearly was seen within two hours after treatment with ionizing radiation [11] and in fibroblasts undergoing replicative senescence [18]. Phosphorylation at Ser33 was induced by both UV and ionizing radiation as early as 1 h after treatment [20,21]; in contrast, phosphorylation of Ser37 and 46 was much more strongly induced by exposure to UV rather than ionizing radiation [22,23]. In addition, Thr81 recently was shown to be phosphorylated in vivo in response to several stresses including exposure to UV light [24].

Ser315 and 392 in the C-terminal regulatory domain were the first p53 phosphorylation sites to be identified [2], and subsequent analyses have shown that both are phosphorylated in vivo after UV irradiation. Furthermore, these modifications coincide with elevated p53-dependent transcription [25–27]. In contrast to Ser315, Ser392 is phosphorylated only very poorly after exposure of cells to ionizing radiation. As noted above, both Ser376 and Ser378 were reported to be constitutively phosphorylated in unstressed MCF-7 breast carcinoma cells while after DNA damage, Ser376 became dephosphorylated in these and other cancer cell lines through an ataxia telangiectasia mutated (ATM) dependent pathway [8]; however, these findings have not been confirmed by others. The acetylation of Lys382 was first reported to be induced in cells exposed to ionizing radiation and UV light by Sakaguchi et al. [21]. Recently, Lys382 and Lys373 also were shown to be acetylated in response to adriamycin, hypoxia, antimetabolites, and inhibitors of nuclear export and RNA polymerase II [28], while acetylation of Lys320 is most strongly induced by UV light. Acetylation is induced with a time course similar to, but slightly later than, several N-terminal phosphorylations, consistent with a role for phosphorylation in the induction of p53 acetylation [21]. p53 also is modified in response to DNA damage by conjugation of Lys386 to a small ubiquitin-like protein, SUMO-1 [29,30].

Several kinases have been identified that detect genotoxic stresses and initiate signaling pathways through p53 phosphorylation (Fig. 1). The ATM kinase has been implicated as the prime candidate for phosphorylating p53 at Ser15 in response to ionizing radiation [31,32], whereas A–T related (ATR) kinase, which phosphorylates p53 at both Ser15 and 37, plays a central role in the activation of p53 in cells exposed to UV light [22]. Interestingly, Ser15 was one of the first identified in vitro substrates for dsDNA-activated protein kinase (DNA-PK), the DNA-activated protein kinase that plays an essential role in mammalian DNA double-strand break repair; however, recent studies have shown that DNA-PK-mediated phosphorylation of p53 in vivo, if it occurs at all, is not required for p53-mediated activation of transcription nor for p53-mediated cell cycle arrest [33,34]. Ser9 and Thr18 may be phosphorylated by a casein kinase (CK)1-δ- enzyme in a phosphorylation cascade that depends on the phosphorylation of the upstream residues Ser6 and 15 [11,35] (Saito et al. unpublished results), although other pathways also may contribute to the phosphorylation of these sites [36]. In contrast to the reported phosphorylation of murine p53 by CK1-δ, Ser6 of human p53 is not phosphorylated by CK1-δin vitro;in vivo this residue must be phosphorylated by an as yet unknown kinase [37].

Two serine/threonine kinases, checkpoint kinase (Chk)1 and Chk2, which act downstream of ATM and ATR, phosphorylate human p53 at Ser20 and possibly at other residues [36,38]. While ATM and Chk2 activate p53 in response to ionizing radiation, ATR and Chk1 appear to be required for the response to UV damage [39,40]. The cyclin-dependent kinase-activating kinase (CAK), which consists of the three subunits CDK7, cyclin H, and p36MAT1, phosphorylates Ser33 in vitro[20]. CAK is part of the transcription factor II H multiprotein complex, which is required for RNA polymerase II transcription and nucleotide excision repair. p36MAT1 interacts with p53 both in vitro and in vivo, suggesting that p36MAT1 can act as a substrate specificity-determining factor for CDK7-cyclin H. The p38 mitogen-activated protein kinase (MAPK) stress-activated kinase also phosphorylates p53 at Ser33 and at a newly identified site, Ser46 [41]. Mutation of both sites decreased p53-mediated and UV-induced apoptosis, suggesting that their phosphorylation may be important for p53-mediated apoptosis after UV radiation. Importantly, a newly identified gene, designated p53AIP1 (p53-regulated apoptosis-inducing protein 1), whose expression is inducible by wild-type p53 in a Ser46-dependent manner, has been reported to regulate apoptosis [23]. Mutation of Ser46 blocked expression of p53AIP1 and inhibited p53 induced apoptosis. Recent work has shown that Jun N-terminal-kinase (JNK) phosphorylates p53 at Thr81 in response to DNA damage and stress-inducing agents [24]. Mutation of this site or over-expression of MKP5, a JNK site-specific phosphatase, reduced p53-mediated activation of transcriptional activity, growth inhibition and apoptosis. Thr55 can be phosphorylated by TAFII250, the largest subunit of TFIID, and phosphorylation of Thr55 enhanced p53 degradation (X. Liu, Department of Biochemistry, University of California, Riverside, CA, USA, personal communication). Exposure of cells to UV decreased phosphorylation at Thr55, thus forging another potential connection between transcription and p53 activation.

At the C-terminus, the cyclin B-dependent kinase, p34(Cdc2), phosphorylates p53 at Ser315. A phospho-specific monoclonal antibody and an IgG binding assay were used to demonstrate near stoichiometric phosphorylation of Ser315 in vivo after DNA damage, and treatment of cells with the cyclin-dependent protein kinase inhibitor, Roscovitine, reduced the transcriptional activity of endogenous p53 or ectopically expressed p53 [25]. Additionally, a phosphatase (hCdc14) that interacts with p53 both in vitro and in vivo, dephosphorylated p53 specifically at Ser315, suggesting that hCdc14 may play an important role in cell cycle regulation of p53 [42]. UV light, but not ionizing radiation, triggers Ser392 phosphorylation [26,27], and two kinases have been identified that might target this residue. The double-stranded RNA activated protein kinase (PKR), a serine/threonine kinase that modulates protein synthesis and is induced by interferon, associates with p53 and phosphorylates Ser392 in vitro[43]. The interaction of PKR with p53 was enhanced by interferon and required the C-terminal 30 amino acids of wild-type human p53. These findings raise the possibility of a functional interaction between PKR and p53 in vivo. The second Ser392 kinase is a multisubunit complex with a molecular weight of ≈ 700 kDa that contains CK2 and the chromatin transcriptional elongation factor FACT [44]. Interestingly, FACT alters the specificity of CK2 in the complex to preferentially phosphorylate Ser392 over other substrates [44].

The histone acetyltransferases (HATs), CREB binding protein (CBP)/p300 and p300/CBP associated factor (PCAF), acetylate the C-terminus of p53 at Lys373 and 382 (p300/CBP) and at Lys320 (PCAF) [21,45]. These enzymes bind the p53 protein and act as coactivators for p53-mediated transcription [46,47]. Histone deacetylases (HDAC1, 2, and 3) also interact with p53 in vitro and in vivo and down-regulate its function, suggesting that deacetylation of p53 is one of the mechanisms that control its activity [48]. Interestingly, the deacetylation of p53 is mediated by HDAC1 complexed to a protein (PID) identical to metastasis-associated protein 2 (MTA2), which has been identified as a component of the nucleosome remodelling and histone deacetylation (NuRD) complexes [49]. This result highlights the importance of p53 acetylation and suggests that deacetylation and interaction with the PID/MTA2-associated NuRD complex may play an important role in the regulation of p53 activity.


Examination of p53 levels, as detected with the DO-1 monoclonal antibody, has shown that several genotoxic treatments stabilize p53 and induce its accumulation in the nucleus. Stabilization is thought to result primarily from disruption of the interaction between the p53 protein and MDM2, which targets p53 for ubiquitin-mediated degradation [1]. Initial studies using purified DNA-PK suggested that phosphorylation of p53 at Ser15 and 37 impaired the ability of MDM2 to bind p53 and to inhibit p53-dependent transactivation [9]; however, more recent work indicates that, with the exception of phosphorylation of Thr18, phosphorylation of other N-terminal residues, alone or in pairs, was not sufficient to block the p53–MDM2 interaction [11,19]. Nevertheless, mutation of either Ser15 or Ser20 to Ala prevented the full stabilization of p53 in vivo[39,50], and mutation of Ser15 (mouse Ser18) to Ala in murine embryonic stem (ES) cells reduced p53 accumulation in response to DNA damage [51]. Using Chk2-knockout thymocytes, it was shown that Chk2, which phosphorylates Ser20, is required for p53 to be stabilized in response to ionizing radiation [40]. Ectopic expression of wild-type Chk2 stabilized p53 after DNA damage, but expression of a dominant-negative mutant form of Chk2 abrogated phosphorylation and stabilization of p53 [38]. These results imply that phosphorylation of Ser20 has a critical role in modulating the negative regulation of p53 by MDM2, but the in vitro binding studies [11,19] suggest the mechanism may be indirect. One possibility is that stabilization of p53 might result more from a phosphorylation-mediate increase in the binding of CBP/p300 (which competes with MDM2) to p53 than from a phosphorylation-mediated decrease in MDM2 binding to p53 [52].

The above results indicate that a prerequisite for full p53 activity is its enhanced stability, which depends on post-translational modifications. Although initial interest focused primarily on the N-terminal segment of p53 that directly interacts with MDM2 [1,2], new studies indicate that other modifications distal to this interaction region also affect p53 stability. Expression of p38MAPK or JNKK kinase after UV irradiation prolonged the half-life of wild-type p53, compared with p53 mutants with substitutions at Ser33, 46 or Thr81 mutants, suggesting that in response to UV light, phosphorylation of these sites may independently lead to p53 stabilization [24,41]. Likewise, inhibition of p53 deacetylation dramatically increased the p53 half-life, suggesting that acetylation also may play an important role in the stabilization of p53 protein [28]. Finally, the recent report of p53 destabilization through Thr55 phosphorylation suggests yet another pathway through which p53 activity can be modulated (X. Liu, Department of Biochemistry, University of California, Riverside, CA, USA, personal communication).

p53 stabilization also may result from the post-translational modification of other interacting proteins. Recent studies have shown that, while the treatment of cells with CPT and etoposide lead to phosphorylation of p53 at Ser15 and 20, these genotoxic agents also result in an inhibition of MDM2 expression [14,16]. DFX, which mimics hypoxia, leads to phosphorylation of Ser15 while actinomycin D, which does not induce phosphorylation of Ser15 or 20, affects the nuclear localization of p53, thus preventing its degradation by MDM2 [14]. These results suggest that different stresses activate different signaling pathways that effect different modifications to the p53 protein to regulate its stabilization. However, the stabilization of p53 may not result from modifications to p53 alone; stress-induced modifications to MDM2 and other components also may be important. Recent studies showed that treatment of cells with ionizing radiation or a radio mimetic chemical, but not UV radiation, induced phosphorylation of MDM2 in an ATM-dependent manner [53,54]. Furthermore, MDM2 is sumoylated at Lys446, within the RING finger domain that plays a critical role in MDM2 self-ubiquitination. Exposure of cells to UV light or ionizing radiation caused a dose- and time-dependent decrease in the degree of MDM2 SUMO-1 modification, which is inversely correlated with the levels of p53. Reduced MDM2 sumoylation in response to DNA damage therefore may contribute to p53 stability [24]. MDM2 also physically interacts with a structurally related protein termed MDMX [55]. The interaction of these two proteins interferes with MDM2 degradation, leading to an increase in the steady-state levels of both MDM2 and p53. Expression of the alternative reading frame (ARF) protein, which relocalizes MDM2 to the nucleolus in response to oncogene activation or aberrant E2F1 activity, provides yet another mechanism for stabilizing p53 [56,57]; however, the contribution of post-translational modifications to this mechanism remains to be investigated.


It is well known that the last 30 amino acids of the C-terminal domain negatively regulate sequence-specific binding by the p53 central core domain and that long, double-stranded DNAs effectively prevent p53 from binding DNA in a sequence-specific manner in vitro[1,2]. The C-terminal segment includes a strongly basic region immediately distal to the tetramerization domain in which are located the MAb421 recognition site, the protein kinase C (PKC) phosphorylation sites at Ser376 and Ser378, the p300 acetylation sites at Lys373 and 382 and the sumoylation site at Ser386. Thus, these post-translational modifications may, directly or indirectly, relieve inhibition by the C-terminus through changes in the binding of interacting proteins that alter p53 structure. Loss of basic charges upon acetylation or sumoylation or addition of negative charges upon phosphorylation could disrupt nonspecific binding to DNA or induce conformational changes that prevent interactions between the C-terminus and the core DNA binding domain.

Experiments by several groups have shown that acetylation stimulates sequence-specific DNA binding in vitro much more strongly in the presence of long competitor DNA that in its absence [21,45]; however, p53 mutants that cannot be acetylated do not reveal obvious defects in DNA binding in vivo ([28] and T. P. Yao, Department of Pharmacology and Cancer Biology, Duke University, NC, USA, personal communication), suggesting that acetylation might have other functions. p53 acetylation is important for suppression of oncogenic ras-induced transformation [58] and to induce metaphase chromosome fragility [59]. In addition, p300/CBP acetyltransferases and p19ARF promote p53 acetylation in vivo, while mouse double minute (Mdm)2 inhibits acetylation, indicating that acetylation is both positively and negatively regulated [28,60]. p300 and CBP levels are limited in cells (reviewed in [61]) and apparently do not support endogenous p53 acetylation under normal conditions. However, activation of p53-specific transcription by p300 is potentiated by ionizing radiation [46], and antibody studies have shown that Lys382 became acetylated in response to both UV and ionizing radiation at relatively early times (1–2 h) [21,28]. p53 and p300 colocalize within the nucleus in a stable DNA-binding complex [47], and acetylation of the C-terminus of p53 in vivo correlates with activation of transcription of downstream targets in response to DNA damage. In vitro, p53 interacts with p300 through its N-terminal domain, and residues between Leu22 and Phe54 are critical for this interaction [62]; moreover, Ser15 phosphorylation stimulates p53 binding to p300/CBP in vitro and in vivo[52,63]. However, Ser15 phosphorylation is not required for p53 acetylation. Actinomycin D, which does not induce Ser15 phosphorylation, is a powerful agent in triggering p53 acetylation [14,28]. Furthermore, p53 is efficiently acetylated in mutant murine ES cells after DNA damage in which the endogenous p53 gene was engineered to express Ala in place of Ser18, the murine residue corresponding to human Ser15 [51]. These results suggest that there may be a second mechanism for insuring p53 acetylation in response to DNA damage. Recent results show that MDM2 suppresses p300/CBP-mediated p53 acetylation in vitro and in vivo and that p19ARF restores p53 acetylation without dissociating MDM2 from p53 [28]. Several mechanisms by which MDM2 may suppress p300/CBP-dependent p53 acetylation are possible. One possibility is that MDM2 may stimulate deacetylation by recruiting a p53 deacetylase. The observation that trichostatin A (TSA) can completely abrogate the effect of MDM2 on p53 acetylation is consistent with this possibility [28]. Interestingly, p19ARF can abrogate the inhibitory effect of MDM2 toward p53 acetylation in vivo[28]. This observation provides another mechanism through which p19ARF may regulate MDM2 activity. One plausible hypothesis is that recruitment of a deacetylase by MDM2 cannot occur when a p53-MDM2- p19ARF complex has formed in the nucleoplasm [64]. However, this simplified model shows only one way that p53 may be activated as a transcription factor through the regulation of interacting components. p53 also is phosphorylated at several C-terminal sites in response to DNA damaging agents, and these sites may function independently or in conjunction with acetylation to activate p53 in response to UV and/or other agents. Furthermore, DNA damage-induced modifications to other components that interact with p53, such as MDM2, JNK, and p300/CBP, also may contribute to regulating p53 function by binding proteins that enhance p53-dependent gene expression, affect DNA repair, or control the nucleo-cytoplasm shuttling of p53.


All examined forms of genotoxic damage activate p53, including those caused by exposing cells to ionizing radiation, UV, alkylating agents, DNA cross-linking agents, and reactive oxygen, through distinct pathways. p53 also is activated in response to nongenotoxic stresses. Ionizing radiation activates p53 primarily through the ATM-Chk2 pathway, whereas UV activates p53 through an ATR-Chk1 pathway that is coupled to a block of transcription and/or DNA replication. Both upstream pathways lead to activation of p53-dependent downstream pathways that control cell-cycle progression and apoptosis. Although significant progress has been made, the precise molecular mechanisms that result in the differential activation of downstream p53-dependent pathways are still unclear. Recent results indicate that ATM phosphorylates MDM2 in the period preceding its induction by p53, suggesting that the defective p53 response of AT cells observed following ionizing radiation treatment may be due in part to effects in signaling pathway(s) that involve MDM2 as well as p53 [65]. A role for DNA-PK, remains controversial. A recent study implicated DNA-PKcs as an upstream effector for p53 in the activation of apoptosis, but not G1 arrest, in response to ionizing radiation [66]; however, a second study using a different DNA-PKcs knock-out mouse strain concluded that the apoptotic response was intact in the absence of DNA-PKcs[67]. These two studies illustrate that important cell- and strain-specific factors that regulate cell fate in response to genotoxic stress remain to be elucidated.

UV radiation produces cyclobutane dimers and 6–4 photoproducts that are repaired via two branches of the nucleotide excision repair (NER) pathway: by global genome repair (GGR) and by the transcription-coupled repair (TCR) pathways. These UV-induced lesions block transcription by RNA polymerase II, which may serve as a trigger for p53 activation [68]. Human epithelial cells are much more sensitive to p53-dependent apoptosis after exposure to UV light than to ionizing radiation. Several lines of evidence show that wild-type p53 is necessary for efficient global genomic NER following UV irradiation, but p53 has a less clear role, if any, in TCR [5,69]. GGR depends on the p53-mediated expression of Gadd45, which binds to damaged chromatin and is thought to affect chromatin remodeling [70,71]. Gadd45 expression also is p53 dependent, but precisely how p53 post-translational modifications affect GGR is still unknown. Recent data demonstrating that the recovery of nascent mRNA synthesis in p53-deficient fibroblasts is significantly impaired (compared with control cells) after exposure to UV light, suggests that wild-type p53 protects cells against UV-induced apoptosis by facilitating the recovery of polymerase II-dependent transcription, perhaps mediated by the degradation of stalled complexes [72].

p53 activation can occur in response to a number of other cellular stresses, including hypoxia and nucleotide deprivation. Hypoxia decreases the expression of the MDM2 protein and the formation of the p53/E6/E6AP complex [in bovine papilloma virus (BPV) transformed cells], resulting in an inhibition of p53 export from the nucleus to the cytoplasm. This inhibition readily induces the nuclear accumulation of p53 and leads to cell-cycle arrest or apoptosis [73,74]. In human fibroblasts, N-phosphoacetyl-l-aspartate (PALA) induces reversible, p53-mediated G1 cell cycle arrest with no evidence of DNA damage. cDNA-representational difference analysis (cDNA-RDA) showed that accumulation of p53 in response to PALA treatment specifically induced the expression of a newly discovered set of genes [75]. However, the signaling pathways for activating p53 after hypoxia or PALA are not known, and it is unclear if the upstream signaling mechanisms or the downstream effectors are similar for these different stresses.

Post-translational modifications promote different interactions between p53 and other proteins and with different target gene regulatory elements to facilitate cell-cycle arrest, apoptosis, or DNA repair. Recent studies suggest that activation of specific promoters depends on specific p53 phosphorylations [23,24,52,76]. p300/CBP or PCAF proteins, which act as coactivators for diverse transcription factors, participate in regulating p53 activity. Newly discovered cofactors, JMY for p300 and PAF400 for PCAF, which facilitate the p53 response by augmenting p53-dependent transcription and apoptosis, recently were described [77,78]. These cofactors stimulate the p300/CBP and PCAF acetytransferases in response to DNA damage and potentiate signaling to p53 (Fig. 2). The promyelocytic leukemia (PML) gene encodes a growth- and tumor-suppressor protein that is essential for G1 arrest and apoptosis. p53 is recruited into PML nuclear bodies (NBs) by a PML isoform (PML3) through the association of its core domain with the C-terminal region of PML3 [79,80]. This localization of p53 into NBs enhances its transactivation activity in a promoter-specific way, which identifies PML as a transcriptional cofactor that plays a role in the activation of p53 responsive genes. Interestingly, the PML protein appears to be essential in the nucleation and formation of nuclear bodies as well as in the recruitment of other proteins such as pRb. The assembly of such loose complexes may allow members to interact at the right time and place to maximize expression of specific genes.

Figure 2.

Regulation of p53 responses. In response to various stresses, p53 becomes post-translationally modified (P, phosphorylation; A, acetylation), as indicated in Fig. 1, through several signal transduction pathways that respond to: e.g. DNA strand breaks, the inhibition of transcription, oxygen tension, and growth signals. Post-translational modifications regulate p53′s interactions with these signal transduction complexes, including histone acetyltransferase/deacetylase complexes (PCAF-PAF400, p300/CBP-JMY, HDAC1-PID/MTA2), kinase/phospatases complexes (CAK [cycH/MAT1/cyclin-dependent kinase (cdk)7], ATM, ATR, TAFII250, PKR, CK2-FACT, hCdc14) and PML3 (which, as for p53, can be sumoylated), which, in turn, regulate p53 stability, p53 location, and the ability to bind specific DNA sequences (box). Modulation of DNA binding specificity and/or interactions with other specific transcription factors may determine which genes are expressed (arrow, target genes) in response to specific stresses and thus cellular fate through activation and maintenance of cell cycle checkpoints or apoptosis. Specific responses are likely to be cell-type specific.


The implications of results discussed in this minireview are multiple. Distinct signal transduction pathways activate p53-dependent transcription through several common post-translational modifications yet the consequences of p53 activation are both damage dependent and cell-type dependent. It is becoming evident that post-translational modifications to p53 that result from the activation of different kinases/phosphatases and acetyltransferases/deacetylases play a critical role in the outcome of p53 activation, but a precise understanding of the mechanisms is not yet in hand. Multiple modifications of p53 initiated by a network of signaling pathways balance the activation of cell cycle checkpoints with the initiation of apoptosis. A fundamental question that remains unanswered is what mechanism(s) contribute to the ability of different cells to interpret p53 activation in different ways upon exposures to varied stresses. With each stress the responses may show similarities, but there will be also differences essential for eliciting a unique molecular signaling outcome. The analysis of the modification patterns in different mouse tissues of knock-in mutants should give insights as to the physiological roles of individual phosphorylations and acetylations [51].

Recent studies indicate that several p53 post-translational modifications appear to modulate its transcriptional activity in a promoter and cell type specific manner. Moreover, both N- and C-terminal phosphorylation sites seem to be involved in modulating DNA-binding and transactivational activities. It appears therefore that multiple sites targeted by an integrated network of signaling pathways highly sensitive to genotoxic stresses must be modified to yield a functional p53. It remains to be seen how changes in protein phosphorylation and/or acetylation in response to specific kinases/phosphatases or acetyltransferases/deacetylases, as well as changes in protein stability mediated by specific ubiquitin ligases, play a role in promoter selectivity and affect the duration and magnitude of the transcriptional output. Clearly, while we have come a long way in the last few years, we have only just scratched the surface on the role of post-translational modifications in the control of p53 function, and the code of post-translational modifications, while dented, has not yet been broken.


We thank Drs Tso-Pang Yao, Sharlyn J. Mazur and Mary E. Anderson for their constructive suggestions. We apologize to those whose publications could not be cited due to space limitations. C. W. A. was supported in part by a CRADA funded by the Laboratory Technology Research Program in the Office of Science of the U.S. Department of Energy.