Encyclopedia of Signaling Molecules

2018 Edition
| Editors: Sangdun Choi

p53

  • Daniel Menendez
  • Thuy-Ai Nguyen
  • Michael A. Resnick
  • Carl W. Anderson
Reference work entry
DOI: https://doi.org/10.1007/978-3-319-67199-4_57

Synonyms

 FLJ92943;  LFS1;  TP53;  TRP53

Historical Background

As the major tumor suppressor in multicellular organisms, p53 is one of the most intensively studied human proteins (over 80,000 publications including nearly 10,000 reviews) because it is critical for maintaining genomic stability, cellular homeostatic processes in response to multiple stresses, and suppressing cancers. The p53 protein is a tetrameric, sequence-specific, DNA-binding transcription factor, stabilized and activated in response to genotoxic and non-genotoxic stresses; estimates are that activation of p53 directly or indirectly induces or represses the expression of about 3,000 genes (about 10% of human genes). These genes coordinate the cellular response to protect cells and/or the organism from damage (Riley et al. 2008). When possible they promote a return to homeostasis by arresting the cell cycle and inducing repair; by altering cellular metabolism; by initiating apoptosis, a program of cell death; or by triggering senescence, a permanent arrest of the cell cycle (Vousden and Prives 2009; Aubrey et al. 2016). Discovered almost simultaneously by several groups just over 37 years ago (Table 1) as an antigen associated with the large tumor antigen of the small DNA tumor virus SV40, p53 initially was considered an oncogene (Levine and Oren 2009). This conclusion was made partly because the initial, mutant cDNA clones derived from tumors cooperated with several oncogenes, notably HRAS, to transform primary cell cultures. On its own, mutant p53 facilitated cell immortalization. Only after cloning of the wild-type (WT) cDNA from normal mouse cells and comparing its sequence with earlier mutant clones was p53 firmly established to be a tumor suppressor that became known as the “guardian of the genome” (Lane 1992). Several groups then demonstrated that the p53 gene (TP53) is mutated in approximately half of all human cancers, and now it is known that a variety of mechanisms inactivate the p53 pathway in virtually all other cancers (Olivier et al. 2010; Vousden and Prives 2009), as recently confirmed by next-generation DNA sequencing approaches. People with one defective p53 gene (Li-Fraumeni syndrome) are cancer prone. Indeed, the fact that TP53 knockout mice rapidly develop tumors confirms the importance of p53 in tumor suppression (Donehower and Lozano 2009).
p53, Table 1

Hallmarks of p53 discovery

Date

Event

1979

p53 discovered

1983

p53 cDNA cloned

1984

p53 (mutant) demonstrated to be oncogenic

1984

p53 shown to be induced by DNA damage

1986

First p53 phosphorylation sites identified

1988–1989

WT p53 cDNA cloned; p53 determined to be a tumor suppressor and mutated in human cancers

1990–1992

p53 identified as a transcription factor

1991

p53 discovered to induce cell cycle arrest and apoptosis

1992

MDM2 proven to negatively regulate p53

1992

TP53 knockout mice revealed as cancer prone

1993

Mutant p53 gain of function demonstrated

1994

p53 DNA-binding domain and tetramerization domain structures revealed

1997

MDM2 shown to ubiquitinate p53; p53 shown to be acetylated

1997

p63 and p73 family members described

1998

ATM demonstrated to phosphorylate p53 Ser15

1999

Expression of DNA repair genes shown to depend on p53

2000

First mouse p53 phosphorylation site knock-in mutant made

2002

Restoration of mutant p53 function by small molecules

2003

Non-transcription functions for p53 described

2005

p53 isoforms described

2005

p53 established as regulating metabolism and having an antioxidant function

2007

p53 demonstrated to be required for embryo implantation

2009

Structure of full p53 tetramer described

2009

p53 shown to regulate stem cell fate

2010

Processing of miRNAs and regulation of lincRNAs modulated by p53

2011–2014

Genome-wide analyses of p53 binding and transcriptional regulation

But how does p53 prevent tumors? The mechanistic answer to this question began to emerge in the early 1990s with the discoveries that p53 possesses a potent transactivation domain and is a site-specific, DNA-binding protein that can induce the expression of genes controlling the progression of the cell cycle (e.g., CDKN1A (p21Waf1)) and apoptosis (e.g., BAX) (Lane 1992; Levine and Oren 2009). Subsequently, the induction of p53 also was shown to promote cellular senescence. Thus, p53 was established as a transcription factor that regulates the expression of genes that control and implement the major cellular mechanisms for preventing cancer (Bieging et al. 2014; Aubrey et al. 2016) (Fig. 1). Transcription-independent functions of p53 also have been discovered (Fig. 1). Among these is the ability of p53 to induce apoptosis through an interaction with the mitochondrial membrane. Although this activity alone is not sufficient for tumor suppression, it causes the subsequent release of pro-apoptotic factors such as cytochrome C (Vaseva and Moll 2009).
p53, Fig. 1

Human p53 signaling in response to genotoxic and non-genotoxic stresses

The early history of p53 research is coupled intimately with that of the small DNA tumor viruses, SV40, polyomavirus, adenovirus, and papillomavirus (Levine 2009). Each virus produces a protein, respectively, the SV40 and polyoma T antigens, the adenovirus E1b-55kd protein, and the papillomavirus E6 protein, that during the early phase of a productive infection inactivates p53 (and suppresses apoptosis) as part of the viral strategy for maximizing virus yields. The expression of these proteins from integrated viral sequences also is required to initiate and maintain cell transformation; remarkably, except for a few papillomavirus serotypes, these viruses seemingly do not produce tumors in humans.

During the past 20 years, much research on p53 focused on cellular responses to DNA damage (Levine and Oren 2009). In normal unstressed cells, p53 has a low basal concentration of about 17,000 molecules per cell that are located primarily in the cytoplasm and have a short (∼20 min) half-life; however, after exposing cells to UV light or ionizing radiation, both of which damage DNA and are carcinogenic, p53 is “induced” and accumulates to higher levels. These seminal findings, which subsequently have been strengthened and extended, provide the rationale for p53 functioning as a multifaceted tumor suppressor in multicellular organisms (Aubrey et al. 2016; Vousden and Prives 2009).

The complex mechanisms that regulate p53 activity are far from being completely understood. Earlier, p53 was shown to be a phosphoprotein, and, as detailed below, it is now known that p53 is posttranslationally modified in multiple ways on more than 15% of its 393 (human protein) amino acid residues (Fig. 2) (Nguyen et al. 2014; Meek and Anderson 2010). An allosteric model for p53 activation as a DNA-binding protein was proposed which suggested that phosphorylation of residues in the p53 carboxyl-terminal regulatory domain relieved an inhibition of specific DNA binding by this domain and promoted activation as a transcription factor. The oncoprotein MDM2 (mouse double minute 2) was found to bind the N-terminal transactivation domain of p53 and to block its activity as a transcription activator; later, MDM2 was identified as an E3 ligase that ubiquitylates p53 at multiple sites and marks p53 for degradation by the 26S proteasome, thus maintaining its low cellular concentration in normal, unstressed cells (Perry 2010). Several groups demonstrated that the ATM kinase, which is mutated in the disease ataxia telangiectasia and is activated in response to DNA damage, phosphorylates serine 15 (Ser15) in the transactivation domain of p53. Subsequently, phosphorylation of Ser15 was suggested to induce the dissociation of MDM2 from p53, thereby stabilizing p53. While it is now known that the mechanisms regulating p53 are more complex, these early findings undoubtedly stimulated interest in clarifying the signaling mechanisms regulating p53 function. These studies were aided substantially by mouse models that used site-specific knock-in technology to explore p53 biology in a living animal (Donehower and Lozano 2009; Bieging et al. 2014).
p53, Fig. 2

p53 domains, PTMs, and binding partners (Updated from Meek and Anderson (2010) and Nguyen et al. (2014))

Origin of the p53 Family

Knowledge of the primordial p53 and its signaling pathways might well offer insights to elucidate more p53 functions and potential cancer therapies (Lu et al. 2009). In addition to TP53, the genomes of all vertebrates contain genes for two closely related homologues, namely, TP63 and TP73 (Candi et al. 2014). The domain structures of these three proteins are similar (see below), but p63 and p73 also contain an additional carboxyl-terminal sterile alpha motif (SAM) that probably facilitates additional protein-protein interactions. All three proteins in this p53 family bind similar DNA sequences and activate several common target genes, but their cellular functions have diverged significantly. Although p63 and p73 have tumor-suppressor functions in mouse models, mutants are uncommon in human cancers. Mouse studies showed that the development of squamous epithelia requires p63, as does maintaining a pool of proliferating epithelial stem cells; p73 is indispensable for development of the central nervous system and the olfactory and immune systems. All three genes generate a plethora of isoforms due to alternative modes of initiating transcription and splicing; up to nine different isoforms have been reported for the TP53 gene products. Transactivation (TA) competent isoforms of p63 and p73 are important for maintaining genome stability in human oocytes. Of particular importance for development are the so-called Δ isoforms of p63 and p73 that lack the N-terminal transactivation domain and function as inhibitors of the three full-length proteins. Interestingly, mice that lack a functional p53 gene develop normally but are highly cancer prone, while those without either p63 or p73 suffer from developmental defects and die soon after birth. Rare germline mutations in p63 are associated with developmental defects in humans. While signaling by the p53 family should be considered collectively, readers should consult recent reviews for further information on the functions of p63 and p73.

p53 and its family members do not occur in prokaryotes, fungi, or plants, but extensive genome sequencing revealed p53- and p63/p73-like genes in many invertebrate species, including clams, mollusks, insects, and worms, as well as in unicellular choanoflagellates and amoeba (Lu et al. 2009). p53 from non-vertebrates is most extensively characterized in the fruit fly Drosophila melanogaster and the nematode Caenorhabdites elegans (CEP-1). However, instead of a vertebrate-like p53, both species express p63/p73-like proteins with carboxyl-terminal SAM domains that function in germline cells; interestingly, neither species has an MDM2-like negative regulatory gene. The discovery of p63 and p73 in the late 1990s suggested that p53 was derived from an ancestral p63/p73-like gene through gene duplication and loss of the SAM domain. However, the more recent detection of p53- and p63/p73-like genes in unicellular and simple multicellular eukaryotes challenges this view. Notably, D. Lane’s group recently identified both MDM2- and TP53-like genes that are well conserved in placozoans, which are simple eukaryotes with only four cell types and with genomes only ten times the size of that of Escherichia coli. Remarkably, not only is the basic structure of the placozoan p53 conserved but so also are key amino acid residues involved in DNA binding and posttranslational modifications (PTM). These findings indicate the interaction of p53 and MDM2 has been conserved for 2.4 billion years of evolution and that the primordial function of p53 was not tumor suppression. Nevertheless, zebrafish p53 has tumor suppressive activity although the common ancestor of zebrafish and humans diverged more than 400 million years ago.

Structure and Posttranslational Modifications

The p53 polypeptide is functionally divided into three major domains (Fig. 2): an amino N-terminal transactivation segment (TA, Met1-Lys101, numbering for human p53) that interacts with regulatory proteins and components of the transcription machinery; a central, sequence-specific, DNA-binding domain (Thr102-Lys292); and a carboxyl C-terminal tetramerization and regulatory domain (Gly293-Asp393) (Meek and Anderson 2010; Nguyen et al. 2014; Joerger and Fersht 2016). The unstructured N-terminal region contains two independent transcription activation domains, TAD1 (Met1-Met40) and TAD2 (Asp41-Pro83). TAD1 is highly conserved and is required for most transactivation activity; it interacts with several coactivator and corepressor proteins (e.g., histone acetyltransferases (HATs) and histone deacetylases (HDACs)) and with MDM2. Residues Glu11-Leu26 may function as a secondary nuclear export signal. TAD2 largely overlaps a proline-rich domain (PRD, Asp61-Leu94) that is important for assuring p53 stability, transactivation ability, and the induction of transcription-independent apoptosis. It also contains binding sites for the corepressor Sin3a and the proline isomerase Pin1. While TAD1 and the C-terminal region of TAD2 are highly conserved among mammals, the length and sequence of residues Asp41-Ala79 of human p53 are not well conserved. Except for Lys101, which can be ubiquitylated by the E3 ligase MDM2, the N-terminal TA region of p53 is covalently modified exclusively by phosphorylation at ten serines/threonines (see Fig. 2). Phosphorylation of p53 Ser15 may be important for p53 binding to some chromatin sites. Phosphorylation of Thr18 and Ser20 prevents binding to MDM2. Phosphorylation of Thr55 by TAF1 may be important for attenuating the p53 response and returning to homeostasis. Phosphorylation of p53 at Ser33, Ser46, and Thr81 (and at Ser315) generates binding sites for Pin1. Pin1-catalyzed proline isomerization generates or removes binding sites for several p53 binding partners that regulate p53 acetylation and ubiquitylation and consequently ensures the efficient loading of p53 onto target promoters. The p53 TAD2 has the most significant polymorphism in p53, a C or G in codon 72 that changes the amino acid from proline (CCC) to arginine (CGC) and affects its activity as a transcription factor (Whibley et al. 2009).

In contrast to the unstructured N- and C-terminal regions, the central core of p53 consists of an immunoglobulin-like β-sandwich offering a scaffold for the DNA-binding surface. This surface is formed by a loop-sheet-helix motif containing loop 1 (L1, Gly117-Val122) and two large loops, L2 (Lys164-Leu194) and L3 (Met237-Pro250), stabilized by a tetrahedrally coordinated zinc ion (Joerger and Fersht 2016); also see http://proteopedia.org/wiki/index.php/P53. X-ray crystallography and nuclear magnetic resonance (NMR) revealed the three-dimensional structures of the DNA-binding domain, free or bound to DNA, and the structures for several mutant p53s in their DNA-free state. Recent high-resolution crystal structures of complexes between the p53 core domain tetramers and DNA targets point to mechanisms by which p53 may recognize specific response elements (REs, see below) in mammalian genomes and how such recognition may be modulated through PTMs or interactions with partner proteins. The DNA-binding domain can be posttranslationally modified by phosphorylation at Thr155 and Ser215; O-glycosylated at Ser149; acetylated at Lys120 and Lys164; ubiquitylated at Lys120, Lys132, Lys139, Lys164, Lys291, and Lys292; and ADP-ribosylated at Glu255, Glu258, and Asp259. In particular, the acetylation of Lys120, a DNA-contact residue, by the HATs Tip60/hMOF in response to DNA damage is important for most p53-mediated transcriptional activation and for inducing apoptosis.

The carboxyl-terminal region contains a nuclear localization signal (Thr312-Asp324), a tetramerization domain (Leu323-Gly356), and a basic 30-amino acid segment that binds numerous partner proteins, RNA, and certain DNA structures including short single strands, four-way junctions, and insertions/deletions in a sequence-independent manner (Fig. 2). The unmodified C-terminal domain initially was thought to negatively regulate sequence-specific binding by the core DNA-binding domain; subsequently, it was shown to be necessary for efficiently activating transcription, possibly by facilitating the identification of REs by promoting diffusion along the DNA. The C-terminal domain can be posttranslationally modified at multiple sites in many ways including phosphorylation, acetylation, methylation, ubiquitination, sumoylation, and neddylation (Fig. 2). Overall these modifications can directly affect p53 oligomerization and stability as well as nonspecific interactions with DNA sequences. Regulation of the formation of p53 tetramers is an underappreciated aspect of p53 signaling. The dissociation constant for tetramer formation apparently is tuned so that p53 is largely in a monomeric (or dimeric) state in unstressed cells; however, the modest threefold increase in p53 concentration after stress is sufficient to push p53 toward the formation of tetramers, and it is widely believed that only tetrameric p53 is active as a transcription factor. Tetramer formation also hides the p53 nuclear export signal in the tetramerization domain (Fig. 2). p53 acetylation and ubiquitylation reportedly inhibit p53 tetramer formation, while the binding of several proteins, including c-Abl, ARC, and 14-3-3, is reported to negatively or positively affect tetramer formation. Finally, some p53 PTMs only occur on tetrameric p53.

Structural studies of full-length human p53 have been impeded because ∼40% of the molecule is intrinsically unstructured and because p53 has a low thermal stability. Nevertheless, A. Fersht’s group determined the low-resolution structure of a stabilized variant of full-length human p53 with and without DNA using small-angle X-ray scattering, NMR, and electron microscopy (Joerger and Fersht 2016); this structure has been confirmed and further enhanced by applying biophysical and computational techniques. The free protein in solution can form an elongated cross-shaped tetramer with loosely coupled core domain dimers and extended N- and C-termini. Upon DNA binding, p53 wraps around the DNA, and the structure becomes more rigid. All four p53 N-termini point away from one face of the core domain/DNA complex, consistent with the fact that the termini become heavily posttranslationally modified and serve as an interaction scaffold for numerous interacting proteins including MDM2, the HATs p300/CBP, the single-stranded binding protein RPA, and several components of the transcriptional apparatus (Fig. 2). Each monomer of the tetramer is proposed to bind one five-base-pair quarter-site of the RE (Fig. 3). Although a quite different structure of DNA-bound p53 based on electron microscopy data has been proposed in which two p53 tetramers interact with one full RE, this configuration has not gained wide acceptance. While elucidating the quaternary structure of p53 with DNA is an important achievement, the fact is that cell DNA is mostly packaged in nucleosomes, and, while some promoter regions may be nucleosome-free, such a scenario is unlikely for many p53 binding sites. To resolve this issue, A. Nagaich and colleagues examined how p53 interacts with nucleosomal REs. Ostensibly, the bending of DNA enhances p53 binding when the response element is positioned near the nucleosomal dyad, but the response element is inaccessible if the orientation of the core nucleosome changes by ∼180°. Sequestering by nucleosomes of the many potential p53 recognition sites in mammalian cells may partly explain how the relatively few p53 molecules find their important target sequences and why different cells, with different chromatin organizations, may respond differently to the activation of p53.
p53, Fig. 3

The p53 response element, sequence-specific p53 DNA binding, and the p53 mutant spectrum

Signaling to p53

p53 activity is regulated not only through control of its cellular location and concentration but also through PTMs that affect its structure or alter its interaction with protein-binding partners. In unstressed cells, the half-life of p53 is short (∼20 min), and low levels of p53 are maintained primarily by the E3 ubiquitin ligase MDM2 (Perry 2010) that transfers ubiquitin to multiple p53 lysines (Fig. 2). Several additional E3 ubiquitin ligases, including Pirh2, COP1, CHIP, ARF-BP1, E6-AP, TOPORS, and TRIM24, also may contribute to p53 ubiquitylation in specific cells or under specific circumstances. Polyubiquitylation of p53 targets it for degradation through the 26S proteasome, while monoubiquitylation (or oligoubiquitylation, less than ∼4) marks p53 for nuclear export to the cytoplasm and non-transcriptional functions (see below) or affects its binding specificity. The complex switch between inhibition of ubiquitylation, monoubiquitylation, or polyubiquitylation probably is cell-type dependent and is regulated by PTMs to both p53 and MDM2, as well as by interactions with several binding factors, including the HATs p300/CBP and Pin1 (above). Furthermore, MDM2 interacts with its homolog MDM4 (or MDMX) which does not have ubiquitin ligase activity but affects MDM2 activity and auto-ubiquitylation and, therefore, p53 levels. Overexpression of MDM2 or MDM4 may suppress p53 activity in some tumors. Deleting or inactivating either MDM2 or MDM4 in mice results in a lethal embryonic phenotype, but the simultaneous deletion of p53 recovers viability in both cases by preventing the induction by p53 of activities that are lethal to cells. Both MDM2 and MDM4 also bind the N-terminal transactivation domain of p53, preventing it from interacting with the transcriptional apparatus and activating transcription. p53 also induces the expression of MDM2 and MDM4 to create a negative feedback loop. Indeed, the p53 circuit communicates with several other signaling pathways including Wnt-β-catenin, IGF-1-AKT, Rb-E2F, p38MAP kinase, cyclin-cdk, p14/19 ARF, cyclin G-PP2A, as well as the ubiquitin ligases Cop-1 and Pirh2, thus creating several negative and positive feedback loops that modify p53 activity or concentration (Harris and Levine 2005).

Although early studies suggested that p53 binding to DNA was regulated allosterically through PTMs to its C-terminus, experiments with the small-molecule inhibitor nutlin-3a suggested that increases in p53 concentration alone can activate many p53 target genes, although the transcriptional response may be somewhat different than those induced by DNA-damaging agents. Nutlin binds MDM2 and inhibits MDM2 binding to p53 thereby blocking p53 degradation without obviously activating stress signaling pathways that posttranslationally modify p53. Likewise, the induction of integrated, recombinant p53 under the control of exogenous regulatory elements similarly can activate p53 transcriptional responses. While unmodified p53 may be transcriptionally competent, it is widely accepted that PTMs to p53 and its partner binding proteins, in response to many cellular stresses (Fig. 2), not only stabilize but “activate” p53, thus fine-tuning the transcriptional responses in a stress- and cell-type-specific manner. Figure 2 shows most known sites at which human p53 can be posttranslationally modified and some of the enzymes that accomplish these modifications, at least in vitro. Also listed are some of the many proteins that reportedly interact with p53. Not all enzymes that have been reported to modify p53 in vitro will do so in vivo; likewise, the list of interacting proteins, though certainly incomplete, may contain false positives. Nevertheless, the potential complexity is daunting (>1012 possible posttranslational combinations), even though only a minute fraction of the possible combinations can exist at one time in any one cell.

The best characterized signaling pathways that modify and modulate p53 are those that are activated in response to DNA damage – primarily DNA double-stranded breaks (DSBs) and stalled or collapsed replication forks (Vousden and Prives 2009; Meek and Anderson 2010). The induction of DSBs triggers a complex signaling network characterized by the rapid activation of the ataxia telangiectasia mutated (ATM) protein kinase. ATM is a member of the family of phosphatidylinositol-3 kinase-related kinases (PIKKs) that includes ATR (ATM related), DNA-PK (double-stranded DNA-activated protein kinase), mTOR (mammalian target for the immunosuppressant drug, rapamycin), and SMG1 (homology to C. elegans SMG-1). Activated ATM directly phosphorylates p53 on Ser15, and under some circumstances it may phosphorylate additional sites, such as Ser46, whose phosphorylation is important for p53-induced apoptosis in response to severe DNA damage. ATM also activates several effector kinases including CHK2 which can phosphorylate p53 on several different sites in the N-terminal transactivation domain (e.g., Thr18, Ser20) as well as multiple sites in the C-terminal domain (e.g., Ser313, 314, 366, 378, and Thr377). Stalled or collapsed replication forks (e.g., caused by exposure to UV light) preferentially activate the ATR kinase, which, like ATM, phosphorylates p53 at Ser15 and also activates effector kinases including CHK1. Similarly to CHK2, CHK1 phosphorylates multiple p53 residues. Ultimately, most of the known 29 p53 phosphorylation sites may be phosphorylated in response to DNA damage with different DNA damage-inducing agents engendering different phosphorylation profiles, thus partly accounting for varied outcomes. For example, exposing fibroblasts to moderate levels of ionizing radiation primarily arrests the cell cycle, while exposure to UV light strongly induces apoptosis. Within the N-terminal transactivation domain, a complex interdependency between phosphorylation sites exists. For instance, phosphorylation of Thr18, Ser20, and Ser46 may depend upon prior phosphorylation of Ser15. The molecular basis for this interdependency is unknown, but it may represent one way in which p53 integrates signals from different stress pathways and/or insures that activation requires sustained signaling from several sources.

Phosphorylation of N-terminal sites, especially phosphorylation of Thr18, weakens binding by MDM2 and, in part, accounts for p53 stabilization in response to DNA damage. Equally important is the modification of MDM2 in response to DNA damage. Like p53, MDM2 and MDM4 are highly modified proteins (Meek and Hupp 2010). ATM directly phosphorylates MDM2 Ser395, which impairs MDM2 nuclear export and p53 degradation; ATM also activates the cAbl kinase, which phosphorylates MDM2 Tyr394 and inhibits MDM2-mediated p53 ubiquitylation. Exposure to ionizing radiation also results in hypo-phosphorylation of several residues in the central acidic domain of MDM2, thereby weakening the interaction with p53 and contributing to p53 stabilization. Similarly, DNA damage changes the phosphorylation of MDM4 (Meek and Hupp 2010).

In addition to inhibiting binding to MDM2, the phosphorylation of multiple sites in the N-terminal transactivation domain of p53 also enhances p53’s interaction with the HATs p300/CBP, so generating a PTM cascade that causes the acetylation of multiple lysines in the DNA-binding and C-terminal domains of p53, including Lys164, Lys305, Lys370, Lys372, Lys381, and Lys382. Lysine 320 can be acetylated by the HAT pCAF. Acetylation of p53 may increase its stability by preventing ubiquitylation; acetylation has both positive and negative effects on tetramer formation and DNA binding. While no single acetylation site appears to be critical for p53 activity, when its eight C-terminal lysines (mouse p53) were changed to arginine, it became virtually “dead” as a transcription factor, suggesting that acetylation is an important p53 modification. Upon severe DNA damage, the MYST family of acetyl transferases, hMOF and TIP60, targets Lys120 in the central DNA-binding domain, and the Lys120-acetylated p53 preferentially binds and activates promoters of pro-apoptotic genes (Meek and Anderson 2010). Three lysines, Lys370, Lys372, and Lys382, at the C-terminus of p53 are methylated in response to DNA damage. Mono-methylation at Lys372 and dimethylation at Lys370 activate p53, while mono-methylation at Lys370 and at Lys382 inhibits its activity. p53 Lys372 is targeted by the Set7/9 methyltransferase, a modification necessary for recruiting the TIP60 HAT complex. Lys372 methylation prevents mono-methylation at Lys370 which is a “repressive mark.” The enzymes that acetylate and methylate p53 in response to DNA damage also may modify histones near the p53 binding site, a process supporting cross talk between the p53 signaling pathway and epigenetic regulation of chromatin.

Enzymes that reverse PTMs and restore homeostasis upon recovery from stress also are important components of the cellular signaling networks regulating p53 activity. At least four phosphatases, PP1, PPM1D (Wip1), PP2A, and Cdc14A, reportedly dephosphorylate specific p53 residues (Nguyen et al. 2014; Meek and Anderson 2010). PPM1D, acting as an oncogene, is overexpressed in approximately 15% of primary breast cancers; in model rodent systems, its inhibition or elimination delays or prevents tumor formation. Several deacetylases, including HDAC1 and SIR2, the demethylase KDM1A (LSD1), and the de-ubiquitylation enzyme HAUSP directly target p53 to remove modification marks. The lifetimes of such modifications on free and chromatin-bound p53 largely are unknown.

In addition to DNA damage, a variety of non-genotoxic physiological processes and stresses, including nutrient deprivation, microtubule distribution, hypoxia, and hyperoxia, may activate p53 through interconnected and as yet incompletely characterized signaling pathways. Glucose limitation induces cell cycle arrest and p53 phosphorylation at Ser15. Glucose starvation causes p53 phosphorylation at Ser46 (a site phosphorylated after severe DNA damage) and triggers p53-mediated apoptosis. Nutritional deprivation activates AMPK, which senses the intracellular AMP to ATP ratio, and AMPK can induce phosphorylation of p53 on Ser15. The  aurora kinases A (AURKA) phosphorylates Ser215 and Ser315; the p38MAP kinase phosphorylates Ser33 and Ser46. Endoplasmic reticulum (ER) stress results from an unfolded protein response that can be caused by glucose starvation, aberrant protein glycosylation, expression of mutant proteins, extreme environmental conditions, and release of Ca 2+ from the lumen, thereby compromising ER homeostasis. If homeostasis cannot be restored, cells are induced to undergo apoptosis through p53-independent pathways. ER is the only stress yet identified that leads to p53 protein destabilization, thereby preventing p53-mediated apoptosis. Hypoxia occurs in ischemic disorders and in solid tumors as a result of abnormal development of vasculature. Severe hypoxia (<0.2% oxygen), a form of non-genotoxic stress, triggers a strong p53 response resulting in Ser15 phosphorylation and p53 accumulation. The hypoxia-inducible transcription factors HIF1 and p53 directly interact and affect each other’s functions. The disruption of nucleolar and ribosomal functions by normal physiological processes or various stresses releases excess ribosomal proteins L5, L11, and L23 that bind and inhibit MDM2, thereby stabilizing p53 and promoting cell cycle arrest without inducing p53 phosphorylation. Deregulation of cell adhesion and disruption of the microtubular architecture and dynamics each can engender p53 stabilization and activation. While less well defined, the non-genotoxic stress pathways activating p53 are distinct from those induced by genotoxic stress.

The activation or overexpression of oncogenes such as Ras, cMyc, or E1A, as occurs during the initiation of cancer, stabilizes and activates p53 by inducing ARF, the product of the alternative reading frame of the cell cycle regulatory gene, CDKN2A. ARF, in turn, binds MDM2 and inhibits p53 ubiquitylation. The activation of p53 by oncogenes undoubtedly is a driving force underlying the mutation of p53 in cancers.

Transcription-Dependent p53 Functions

The transcriptional activity of p53 is crucial for its function as a tumor suppressor. Once stabilized, modified, and accumulated in the nucleus, p53 recognizes and binds DNA as a tetramer, via its core domain, to a consensus sequence known as a response element (RE) (Fig. 3a). The canonical p53RE, based primarily on in vitro studies, comprises two decamer palindromic half-sites (5′-RRRCWWGYYY-3′ wherein R is A or G, W is A or T, and Y is C or T), separated by a spacer 0–13 bp. The core domains of two p53 monomers bind to a half-site to form a symmetrical dimer, and two such dimers assemble on a full site to form a p53 tetramer. Both the sequence and the spacer’s length have important roles in the outcome of p53 binding to a RE. The C and G of the core CWWG motif are strongly conserved, while the WW motif and the flanking RRR and YYY segments exhibit considerable variation. This ambiguity generates approximately 20,000 putative binding sites across the human genome. Of these, only about 550 perfectly match the consensus sequence without a spacer, and these represent the strongest binding sites. In vitro self-assembly of a tetramer on a zero spacer RE is highly co-operative, and the tetramer is highly stable, with a half-life of approximately 15 min. Nevertheless, many bound p53REs contain a mismatch in at least one of the decamers. Based on the amount of p53 protein accumulated after stress activation in the normal cell nucleus (∼2 × 104 monomers/nucleus), a given cell probably can bind no more than 1,500–2,000 p53 elements at one time.

Co-crystal structures of p53 tetramers with DNA recognition oligonucleotides and computational modeling provide insights into the molecular mechanism for the recognition of p53REs. The L3 loop (see above and Fig. 3b) docks to the minor groove of the DNA helix via Arg248 (a hot spot cancer mutant), while residues within the loop-sheet-helix including Arg273 (hot spot mutant), Ala276, Cys277, and Arg280 in the C-terminal helix interact with bases in the major groove and the DNA backbone. In the two subunits contacting the inner repeats of the half-sites, the L1 loop adopts an extended conformation, and Lys120 interacts directly with DNA. In the crystal structure, the L1 loops of the subunits contacting the outer quarter-sites are less well defined and differ in different reported structures. Acetylation of Lys120 enhances DNA-binding specificity and apoptosis. Computational studies and experiments with mutants suggest that the carboxyl-terminal regulatory domain communicates with the DNA-binding core domain and may alter its conformation (induced fit). Changes also occur in the DNA helix upon p53 binding, especially with respect to the A-T doublet in each half-site.

A number of genome-wide p53 binding (chromatin immunoprecipitation-DNA sequencing, ChIP-seq) studies in various cancer and normal cell lines treated to activate p53 have established a set of about 800 frequently bound sites, but several thousand sites may be identified in any one study. Typically about 25% of bound sequences have no recognizable p53 motif. Most of these may represent p53 caught in the act of searching for a recognition sequence or interacting with other DNA-bound proteins, but the physiological relevance of these sites, if any, is uncertain. For the vast majority of sites containing a recognizable p53 motif, this motif contains no spacer, in contrast to sites identified in early studies or in vitro. Spacer nucleotides greater than one greatly reduce the p53 binding affinity. A slight majority of bound sites with motifs fall into intergenic regions and are frequently associated with chromatin marks indicative of an enhancer. Many of the remaining sites (about 30%) fall within 5 kb of a known transcription start site. Nevertheless, the majority of bound p53 elements have not been associated with changes in p53-mediated transcription; the physiological function of these, if any, is unknown. In a few cases, p53 has been found to utilize partial REs (half- and three quarter-sites), in conjunction with other transcription factors such as the estrogen receptor, a feature which potentially expands the number of genes that p53 could regulate (Menendez et al. 2009). p53 also may bind some simple sequences (e.g., (5′-TGYCC-3′) n>10) and non-B-form DNA structures, either to regulate transcription or, perhaps, DNA repair and chromatin structure.

After induction and activation, p53 largely drives transcription from a relatively large network of approximately 1,000 direct target genes by binding and transcriptionally regulating different sets of target genes. After mild DNA damage, p53 typically arrests the cell cycle, allowing time for DNA repair and cell survival; however, after severe or unrepairable DNA damage, p53 triggers either senescence or apoptosis in a cell-type-dependent manner. The induction of p53-dependent cell cycle arrest in the G1 phase is mediated through the transcriptional activation of targets, such as the cyclin-dependent kinase inhibitor gene CDKN1A, one of the best-studied p53 targets, while the induction of 14-3-3σ contributes to inhibiting the G2/M transition. Triggering p53-dependent apoptosis involves the transcriptional activation of genes in the mitochondrial apoptotic pathway, such as PUMA, APAF1, and BAX, and in the death receptor pathway, such as FAS and TNFRSF10B. However, as noted below, p53 also can induce apoptosis independent of transcription.

Considerable effort has been directed toward understanding the role of p53 in stress-specific responses. p53 displays well-documented different binding affinities for different REs and, depending on the level of p53 accumulation after stress, p53 may bind different sets of REs, with the highest affinity sites being occupied first. For example, overall cell cycle promoters have stronger binding sites than do pro-apoptotic genes. p53REs near apoptotic genes also are poorly conserved through evolution relative to REs associated with genes that regulate the cell cycle. Certain p53 partners can “direct” p53 binding to cell cycle or to pro-apoptotic promoters, and specific p53 PTMs directly influence the selection of p53REs at least partly by modulating partner protein binding. A classic example is that of the ASPP protein family, consisting of the three members ASPP1, ASPP2, and iASPP. While binding of ASPP1 and ASPP2 to p53 directs it to pro-apoptotic genes and stimulates apoptosis, iASPP inhibits these interactions. Another binding partner, HZF, selectively loads p53 on cell cycle regulating promoters without affecting the regulation of pro-apoptotic genes.

An important recent finding is that p53 oscillates in abundance after certain stresses (e.g., exposure to X-rays), while after other stresses (e.g., exposure to UV light) p53 accumulates and remains at a relatively constant level for an extended time. The oscillations of p53 can be largely modeled mathematically using the p53 regulators MDM2 and the WIP1 phosphatase, which dephosphorylates several p53 serines and contributes to the inactivation of DNA damage response components.

The mechanisms by which p53 activates transcription are only partially understood. Once bound to an RE, p53 loads transcriptional coactivators, such as the HATs p300/CBP and PCAF, and chromatin-remodeling factors at many of its target promoters. The resulting acetylation of histones and the opening of chromatin for binding the components of the basal transcriptional machinery such as TBP, TFs, and RNAPII lead to the subsequent activation of transcription. The outcome of p53 binding to its targets might also depend on additional factors that are required due to the promoter’s state before binding. For example, promoters for cell cycle genes contain bound but inactive RNAPII prior to their activation by stress; in contrast, pro-apoptotic genes contain little or no pre-bound RNAPII. Accordingly, additional factors would be needed by cell cycle regulating promoters to convert the paused RNAPII into an elongating one, while for pro-apoptotic promoters, basal transcription factors and RNAPII first must be loaded. It is noteworthy that different p53-regulated core promoters vary in their promoter element compositions and, therefore, may require different co-regulators. Although TATA element-containing “focused” core promoters are more ancient, most core promoters in vertebrates are of the “dispersed” type that do not contain the TATA element and commonly are found in CpG islands. Therefore, the classical model in which p53 helps to load TBP and the rest of the basic transcriptional machinery is applicable to only a fraction of regulated, focused promoters, and the need for specific co-regulator loading may depend on the particular elements of the core promoter at a given target. The epigenetic state of the targets before binding of p53 also may impose specific requirements before transcription is activated. Co-regulators co-occupying specific sets of p53-regulated target promoters have been described. For example, hCAS was reported by C. Prives’ group to associate with a subset of pro-apoptotic p53 targets and to enhance their transcription by reversing H3 Lys27 methylation (a modification blocking transcription). In contrast, Bach1, a transcription factor that binds at a subset of oxidative stress-inducible p53 target genes and promotes histone deacetylation, as shown by K. Igarashi’s group, inhibits transcription from these genes, thus suppressing p53-mediated cellular senescence. Likewise, the calcineurin-binding protein, CABIN1, physically interacts with p53 (Fig. 2) on certain promoters and, by regulating histone modifications, represses p53-dependent transcription in the absence of genotoxic stress. Finally, Gomes and Espinosa recently reported that expression of PUMA (BBC3), a pro-apoptotic p53 target gene, is regulated through a newly discovered “noncanonical” mechanism involving the insulator protein CTCF (the CCCTC-binding factor) and the cohesion complex that occupy intragenic chromatin boundaries and inhibit transcriptional elongation in the absence of stress. These findings demonstrate the role of the chromatin landscape in regulating p53-mediated transcriptional responses.

A surprising finding in recent years is that p53 is involved in the regulation of many pathways in addition to those initially presumed to be largely responsible for its tumor suppressor activity, i.e., regulation of the cell cycle, apoptosis, and senescence. For example, recent studies from the Resnick lab show that p53 has an important role in modulating and mediating non-cell-autonomous interactions with the immune system, one of the most important defenses against external as well as internal threats including tumorigenesis. DNA damage may trigger p53-dependent inflammatory responses that contribute and help orchestrate the clearance of tumor cells by triggering a senescence program which contributes to tumor suppression. Furthermore, the vast number of p53 transcriptional targets includes several immune genes, for example, most members of the innate immune-related Toll-like receptor (TLR) gene family. Consequently, p53 activation may enhance TLR-dependent production of proinflammatory cytokines in various human immune-related primary cells as well as cancer-derived cells. Some p53 mutants in somatic and germline-associated tumors dramatically influence stress-induced gene expression of TLRs. Surprisingly, p53 regulation of TLRs is not conserved in mice, suggesting that with present technology some p53-related immune responses can only be addressed in human material. p53 and NF-κB recently were shown to coregulate proinflammatory gene responses in human macrophages. The TP53 promoter contains a functional interferon-stimulated response element, which can be modulated by interferon, one of the main effectors of the innate immune system. This pathway places p53 as an effector that initiates antiproliferative responses to viral and other pathogen infections. p53REs also are found associated with other interferon-stimulated genes (ISGs), such as IRF9, IRF5, and ISG15. Thus p53 may serve as a central mediator and amplifier of global innate immune responses, findings which highlight its important physiological role in the immune system while providing a new dimension to the broad role that p53 plays in human biology. There is still much to learn about the role of p53 in other pathways such as neural development as well as how responses may differ in the many different cell types of the human body.

Many studies have shown that after p53 activation about as many genes become repressed as are induced. Early studies of individual genes identified about 50 that seemed to be directly repressed by p53, and claims have been made that certain RE sequences were involved in repression (Riley et al. 2008). However, more recent genome-wide ChIP-seq and RNA-seq studies indicate that very few genes appear to be directly repressed by p53; rather, repression appears to be an indirect consequence of p53 induction. p53 is now known to induce a number of microRNAs (miRNAs) and long intergenic noncoding RNAs (lincRNAs) that may account for repression of some genes, but a recent study indicates that p21 and the DREAM complex are required for most p53-mediated repression.

Transcription-Independent p53 Functions

While initially controversial, transcription-independent p53 functions now are well recognized, and some are well characterized, particularly those regulating apoptosis and autophagy (Fig. 1). p53 can induce apoptosis in the presence of transcriptional or translational inhibitors, strongly suggesting a transcription-independent role for p53 in the apoptotic process (Vaseva and Moll 2009). Subsequently, mutants deficient for transcriptional activation were shown to retain the ability to activate apoptosis. DNA damage, hypoxia, or oncogene activation promote p53 translocation to the cytoplasm where it inhibits the anti-apoptotic proteins Bcl-2 and Bcl-xL, thus causing permeabilization of the outer mitochondrial membrane, the release of cytochrome C, and the activation of caspases, followed by chromatin condensation, targeted proteolysis, and cell death. The p53 protein lacks a classical mitochondrial translocation motif. U. Moll’s lab suggested that MDM2 monoubiquitylates p53, thereby promoting its translocation to the cytoplasm and mitochondria, whereupon it rapidly is de-ubiquitylated by mitochondrial HAUSP, generating an apoptotically active p53. Under non-apoptotic conditions, a recent study showed that the mitochondrial disulfide relay system mediated in part by CHCHD4 promotes p53 mitochondrial import.

Autophagy is subjected to p53-mediated activation or inhibition through both transcription-dependent and transcription-independent mechanisms (Vousden and Prives 2009). Upon genotoxic stress or oncogene activation, induced p53 activates autophagy as does p53 loss in the absence of stress, implying an inhibitory effect of basal p53 activity on autophagy in nonstressed cells. Transcription-independent p53 activation of autophagy depends on activation of the nutrient sensor AMPK, followed by mTOR repression. Additional transcription-independent functions of p53 include DNase and RNase activities, enzyme activation or inhibition (e.g., inhibition of G6PD), modulation of mitochondrial DNA repair, regulation of translation, and inhibition of homologous recombination (Fig. 1).

Mutant p53 and Cancer

p53 is the most frequently mutated tumor suppressor in humans as recently confirmed using next-generation sequencing approaches, and most of the tumor-derived mutations are missense mutations in the central domain that block or alter sequence-specific DNA binding (contact mutants, e.g., p.R273H) or induce conformational changes in the binding domain (structural mutants, e.g., p.R175H) (see http://p53.iarc.fr/). Although many p53 codons are mutated in somatic cancers, six so-called hot spot mutations (R175H, R248Q, R248W, R273H, R273C, and R282W) account for about 20% of all somatic mutations in human cancers (Fig. 3c) (Olivier et al. 2010). Not all TP53 mutations equally affect p53 functions; thus, a wide range of phenotypic diversity is generated that impacts tumor development, aggressiveness, chemoresistance, and metastastic potential. Sporadic, germline, gain of function (GoF), change of spectrum, prion like, metastasis inducer, and mediator of chemoresistance are just some of the characteristics of different p53 mutants. For example, the R175H mutant p53 arrests the cell cycle but cannot induce apoptosis. Many mutant p53s become highly overexpressed in tumor cells.

Of particular importance are those mutants classified as GoF, which have activities not present in WT p53 and that can confer a selective advantage to cells, contributing to various aspects of tumor progression (Shetzer et al. 2016). Another interesting group of p53 mutants associated with cancer are those that retain some WT-like transactivation capability. Many of these can result in a change of spectrum of transactivation from various p53 targets, thereby altering cellular responses such as DNA repair, genome stability, programmed cell death, and innate immune signaling.

For those p53 mutant alleles that exhibit a complete loss of p53 function but still express mutant protein, pharmacological restoration of function presents an exciting potential strategy for cancer therapy. Over the past decade, several small molecules have been identified that inactivate mutant p53 or restore the WT p53 response for mutant p53 protein. Drugs such as nutlin, MI-219, and RITA can activate a WT p53 response in tumors carrying WT p53. Small molecules such as PRIMA-1, MIRA-1, Elipticin, CDB3, WR1065, NSC319726, p53R3, and CP-31398 have been used to inactivate mutant p53 or to restore WT function, although in many cases their mechanisms of action remain to be clarified.

Summary

In the nearly 40 years since its discovery (Table 1), p53 has become perhaps the best characterized mammalian transcription factor. While its roles as the central hub in cellular stress responses and as a tumor suppressor have been detailed, many questions remain and new ones are emerging. Surprisingly little is understood about the biochemical mechanisms by which the many p53 PTMs regulate its function. This lack raises a plethora of questions. Where and when are these modifications accomplished, in the cytoplasm, nucleoplasm, or on chromatin? How do they differ among different cell types? What fraction of p53 is modified and which modifications coexist on the same p53 molecule and on the same p53 tetramer? What is the functional interplay among p53 family members? Like p63 and p73, p53 is expressed as several different isoforms, some lacking the transactivation domain or having different carboxyl termini. How do these influence p53 “activity”? How do the more than 300 p53 binding partners influence p53 activity, and how are these interactions regulated? Does p53 activity and function vary among individuals? Are polymorphisms in p53REs important in human health? How does the chromatin landscape, including epigenetic modifications such as DNA methylation, modulate p53 binding and regulate transcription? Is it all about transcription? Are there other functions for p53 sequence-specific DNA binding? Besides its major role as a cellular tumor suppressor, recent studies illuminated many tumor-independent functions, including the regulation of cellular metabolism and autophagy, the composition of the extracellular matrix and cell cytoskeleton, immunity and fertility in mammals, stem cell renewal, and neuronal differentiation. Unsurprisingly, p53 has been implicated in the development of a broad spectrum of pathologies other than cancer, including ischemia after stroke, myocardial infarction, and Parkinson’s and Alzheimer’s disease. Importantly, can mechanisms be found for activating mutant p53 or inactivating WT p53 or its pathways that will be effective as cancer therapeutics or for other diseases? There is little doubt that research on p53 as a signaling molecule has a long, bright future.

Notes

Acknowledgments

This research was supported in part by the Intramural Research Program of the NIH, National Institute of Environmental Health Sciences.

References

  1. Aubrey BJ, Strasser A, Kelly GL. Tumor-suppressor functions of the TP53 pathway. Cold Spring Harbor Perspect Med. 2016;6.Google Scholar
  2. Bieging K, Mello SS, Attardi LD. Unravelling mechanisms of p53-mediated tumour suppression. Nat Rev Cancer. 2014;14:359–70.PubMedPubMedCentralCrossRefGoogle Scholar
  3. Candi E, Agostini M, Melino G, Bernassola F. How the TP53 family proteins TP63 and TP73 contribute to tumorigenesis: regulators and effectors. Hum Mutat. 2014;35:702–14.PubMedCrossRefGoogle Scholar
  4. Donehower LA, Lozano G. 20 years studying p53 functions in genetically engineered mice. Nat Rev Cancer. 2009;9:831–41.PubMedCrossRefGoogle Scholar
  5. Harris SL, Levine AJ. The p53 pathway: positive and negative feedback loops. Oncogene. 2005;24:2899–908.PubMedCrossRefGoogle Scholar
  6. Joerger AC, Fersht AR. The p53 pathway: origins, inactivation in cancer, and emerging therapeutic approaches. Annu Rev Biochem. 2016;85:375–404.PubMedCrossRefGoogle Scholar
  7. Lane DP. Cancer. p53, guardian of the genome. Nature. 1992;358:15–6.PubMedCrossRefGoogle Scholar
  8. Levine AJ. The common mechanisms of transformation by the small DNA tumor viruses: the inactivation of tumor suppressor gene products: p53. Virology. 2009;384:285–93.PubMedCrossRefGoogle Scholar
  9. Levine AJ, Oren M. The first 30 years of p53: growing ever more complex. Nat Rev Cancer. 2009;9:749–58.PubMedPubMedCentralCrossRefGoogle Scholar
  10. Lu WJ, Amatruda JF, Abrams JM. p53 ancestry: gazing through an evolutionary lens. Nat Rev Cancer. 2009;9:758–62.PubMedCrossRefGoogle Scholar
  11. Meek DW, Anderson CW. Posttranslational modification of p53: cooperative integrators of function. In: Levine AJ, Lane D, editors. Cold spring harbor perspectives in biology, volume on the p53 family. New York: Cold Spring Harbor Laboratory Press; 2010. p. 81–96.Google Scholar
  12. Meek DW, Hupp TR. The regulation of MDM2 by multisite phosphorylation – opportunities for molecular-based intervention to target tumors? Semin Cancer Biol. 2010;20:19–28.PubMedCrossRefGoogle Scholar
  13. Menendez D, Inga A, Resnick MA. The expanding universe of p53 targets. Nat Rev Cancer. 2009;9:724–37.PubMedCrossRefGoogle Scholar
  14. Nguyen TA, Menendez D, Resnick MA, Anderson CW. Mutant TP53 posttranslational modifications: challenges and opportunities. Hum Mutat. 2014;35:738–55.PubMedPubMedCentralCrossRefGoogle Scholar
  15. Olivier M, Hollstein M, Hainaut P. TP53 mutations in human cancers: origins, consequences, and clinical use. In: Levine AJ, Lane D, editors. Cold spring harbor perspectives in biology, volume on the p53 family. New York: Cold Spring Harbor Laboratory Press; 2010. p. 123–40.Google Scholar
  16. Perry ME. The regulation of the p53-mediated stress response by MDM2 and MDM4. In: Levine AJ, Lane D, editors. Cold spring harbor perspectives in biology, volume on the p53 family. New York: Cold Spring Harbor Laboratory Press; 2010. p. 97–108.Google Scholar
  17. Riley T, Sontag E, Chen P, Levine A. Transcriptional control of human p53-regulated genes. Nat Rev Mol Cell Biol. 2008;9:402–12.PubMedCrossRefGoogle Scholar
  18. Shetzer Y, Molchadsky, A, Rotter V. Oncogenic mutant p53 gain of function nourishes the vicious cycle of tumor development and cancer stem-cell function. Cold Spring Harb Perspect Med 2016 doi:10.1101/cshperspect.a026203.Google Scholar
  19. Vaseva AV, Moll UM. The mitochondrial p53 pathway. Biochim Biophys Acta. 2009;1787:414–20.PubMedCrossRefGoogle Scholar
  20. Vousden KH, Prives C. Blinded by the light: the growing complexity of p53. Cell. 2009;137:413–31.PubMedCrossRefGoogle Scholar
  21. Whibley C, Pharoah D, Hollstein M. p53 polymorphisms: cancer implications. Nat Rev Cancer. 2009;9:95–107.PubMedCrossRefGoogle Scholar

Copyright information

© Springer International Publishing AG 2018

Authors and Affiliations

  • Daniel Menendez
    • 1
  • Thuy-Ai Nguyen
    • 1
  • Michael A. Resnick
    • 1
  • Carl W. Anderson
    • 1
    • 2
  1. 1.Chromosome Stability Group, Genome Integrity and Structural Biology LaboratoryNational Institute of Environmental Health Sciences, NIHResearch Triangle ParkUSA
  2. 2.Biology DepartmentBrookhaven National LaboratoryUptonUSA