Genes VII

The most important tumor suppressor is p53 (named for its molecular size). More than half of all human cancers either have lost p53 protein or have mutations in the gene. p53 is a nuclear phosphoprotein. It was originally discovered in SV40-transformed cells, where it is associated with T antigen. A large increase in the amount of p53 protein is found in many transformed cells or lines derived from tumors. In early experiments, the introduction of cloned p53 was found to immortalize cells. These experiments caused p53 to be classified as an oncogene, with the usual trait of dominant gain-of-function (Linzer and Levine, 1979; Finlay et al., 1989).

But all the transforming forms of p53 turned out to be mutant forms of the protein! They fall into the category of dominant negative mutants, which function by overwhelming the wild-type protein and preventing it from functioning. The most common form of a dominant negative mutant is one that forms a heteromeric protein containing both mutant and wild-type subunits, in which the wild-type subunits are unable to function. p53 probably exists as a tetramer. When mutant and wild-type subunits of p53 associate, the tetramer takes up the mutant conformation.

Figure 28.24 shows that the same phenotype is produced either by the deletion of both alleles or by a missense point mutation in one allele that produces a dominant negative subunit. Both situations are found in human cancers. Mutations in p53 accumulate in many types of human cancer, probably because loss of p53 provides a growth advantage to cells; that is, wild-type p53 restrains growth. The diversity of these cancers suggests that p53 is not involved in a tissue-specific event, but in some general and rather common control of cell proliferation; and the loss of this control may be a secondary event that occurs to assist the growth of many tumors. Mutant p53 cells also have an increased propensity to amplify DNA, which is likely to reflect p53’s role in the characteristic instability of the genome that is found in cancer cells.

This interpretation implies that a normal cell has a capacity to grow in an unrestrained manner that usually is inhibited by p53. Mutant mice that lack p53 are viable, but develop a variety of tumors rather early. This confirms that it is loss of p53 function that contributes to the tumorigenic phenotype. p53 is defined as a tumor suppressor also by the fact that wild-type p53 can suppress or inhibit the transformation of cells in culture by various oncogenes (for review see Marshall, 1991; Levine et al., 1991).

Mutation in p53 is the cause of Li-Fraumeni syndrome, which is a rare form of inherited cancer. Affected individuals display cancers in a variety of tissues. They are heterozygotes that have missense mutations in one allele. These mutations behave as dominant negatives, overwhelming the function of the wild-type allele. This explains the occurrence of the disease as an autosomal dominant (Malkin et al., 1990).

All normal cells have low levels of p53. A paradigm for p53 function is provided by systems in which it becomes activated, the most usual cause being irradiation or other treatments that damage DNA. This results in a large increase in the amount of p53. Two types of event can be triggered by the activation of p53: growth arrest and apoptosis (cell death). The outcome depends in part on which stage of the cell cycle has been reached. Figure 28.25 shows that in cells early in G1, p53 triggers a checkpoint that blocks further progression through the cell cycle. This allows the damaged DNA to be repaired before the cell tries to enter S phase. But if a cell is committed to division, then p53 triggers a program of cell death. The typical results of this apoptosis are the collapse of the cell into a small heteropycnotic mass and the fragmentation of nuclear DNA (see 27 Cell cycle and growth regulation). The stage of the cell cycle is not the only determinant of the outcome; for example, some cell types are more prone to show an apoptotic response than others.

We may rationalize the existence of these two outcomes by supposing that damage to DNA can activate oncogenic pathways, and that the purpose of p53 is to protect the organism against the consequences. If it is possible, a checkpoint is triggered to allow the damage to be repaired, but if this is not possible, the cell is destroyed. We do not know in molecular terms how p53 triggers one pathway or the other, depending on the conditions, but we have an understanding of individual activities of p53 that may be relevant to these pathways (for review see Levine, 1997).

p53 has a variety of molecular activities. Figure 28.26 summarizes the responsibilities of individual domains of the protein for these activities:

• p53 is a DNA-binding protein that recognizes an interrupted palindromic 10 bp motif. The ability to bind to its specific target sequences is conferred by the central domain.
• p53 activates transcription at promoters that contain multiple copies of this motif. The immediate N-terminal region provides the transactivator domain. p53 may repress other genes; the mechanism is unknown.
• p53 also has the ability to bind to damaged DNA. The C-terminal domain recognizes single-stranded regions in DNA.
• p53 is a tetramer (oligomerization is a prerequisite for mutants to behave in a dominant negative manner). Oligomerization requires the C-terminal region.
• A (putative) signaling domain contains copies of the sequence PXXP, which forms a binding site for SH3 domains.

Mutations in p53 have various effects on its properties, including increasing its half-life from 20 minutes to several hours, causing a change in conformation that can be detected with an antibody, changing its location from the nucleus to the cytoplasm, preventing binding to SV40 T antigen, and preventing DNA-binding. As shown in Figure 28.26, the majority of these mutations map in the central DNA-binding domain, suggesting that this is an important activity.

p53 activates various pathways through its role as a transcription factor. The pathways can be divided into the three groups summarized in Figure 28.27. The major pathway leading to inhibition of the cell cycle at G1 is mediated via activation of p21, which is a cki (cell cycle inhibitor) that is involved with preventing cells from proceeding through G1 (see Figure 27.25 and Figure 28.23). Activation of GADD45 identifies the pathway that is involved with maintaining genome stability. GADD45 is a repair protein that is activated also by other pathways that respond to irradiation damage. The pathway leading to apoptosis remains to be identified.

When it functions as a transcription factor, p53 uses the central domain to bind to its target sequence. The N-terminal transactivation domain interacts directly with TBP (the TATA box-binding protein). This region of p53 is also a target for several other proteins. An interaction with E1B 55 kD enables adenovirus to block p53 action, which is an essential part of its transforming capacity. Other regions of p53 can also be targets for inhibition; the SV40 T antigen binds to the specific DNA-binding region, thereby preventing the recognition of target gene (Seto et al., 1992)s

The stability of p53 is an important parameter. It usually has a short half-life. The response to DNA damage stabilizes the protein and activates p53’s transactivation activity.

The cellular oncoprotein Mdm2 inhibits p53 activity. p53 induces transcription of Mdm2, so the interaction between p53 and Mdm2 forms a negative feedback loop in which the two components limit each other’s activities. The circuitry that controls p53’s activity is illustrated in the upper part of Figure 28.28. Proteins that activate p53 behave as tumor suppressors; proteins that inactivate p53 behave as oncogenes (Momand et al., 1992).

To function as a transcription factor, p53 requires the coactivators p300/CBP (also used by many other transcription factors: see 21 Regulation of transcription). The coactivator binds to the transactivation (N-terminal) domain of p53. The interaction between p53 and p300 is necessary in order for Mdm2 to bind to p53.

Control of p53’s stability is influenced by Mdm2, which has two effects on p53. It functions as an E3 ubiquitin ligase that causes p53 to be targeted by the degradation apparatus. And it also acts directly at the N-terminus to inhibit the transactivation activity of p53. The consequence of this circuit is that Mdm2 limits p53 activity; and the activation of p53 increases the amount of Mdm2.

The C-terminal domain of p53 binds without sequence-specificity to short (<40 base) single-stranded regions of DNA and to mismatches generated by very short (1-3 base) deletions and insertions of bases. Such targets are generated by DNA damage. The consequence of this interaction is to activate the sequence-specific binding activity of the central domain, so that p53 stimulates transcription of its target genes. The nature of this connection is not clear, but may be a two-stage process. When p53 binds through its C-terminal domain to a damaged site on DNA, a change occurs in its properties; it then dissociates from the damaged site and binds to a target gene, which it activates.

The ability of p53 to trigger apoptosis is less well understood. It may depend on the transactivation of a different set of target genes from those involved in activating the G1 checkpoint. The two activities can be separated by the response to adenovirus E1B 19 kD protein, which blocks the apoptotic activity of p53, but does not block its activity to activate target genes. The independence of the effects of p53 on growth arrest and apoptosis is emphasized by the fact that the E1B 55 kD protein blocks transactivation capacity but does not interfere with apoptosis.

The importance of the connection between tumorigenesis and loss of apoptosis is also shown by the properties of the bcl2 oncogene. bcl2 was originally identified as a target that is activated by translocations in certain tumors. It turns out to have the property of inhibiting most pathways for apoptosis (see 27 Cell cycle and growth regulation). This suggests that apoptosis plays an important role in inhibiting tumorigenesis, probably because it eliminates potentially tumorigenic cells. When apoptosis is prevented because bcl2 is activated, these cells survive instead of dying (Vaux et al., 1988).

Cells with defective p53 function have a variety of phenotypes; this pleiotropy makes it difficult to determine which (if any) of these effects is directly connected to the tumor suppressor function. Most of our knowledge about p53 action comes from situations in which it has been activated. We assume that the pathways it triggers Xgrowth arrest or apoptosis Xare connected to its ability to suppress tumors. Certainly it is clear that the failure of p53 to respond to DNA damage is likely to increase susceptibility to mutational changes that are oncogenic. However, we do not know whether this is the sole role played by p53. p53 ^V mice have normal survival, implying that p53’s role is not essential for development.

The general definition of their properties shows that both RB and p53 are tumor suppressors that in some way usually control cell proliferation; their absence removes this control, and contributes to tumor formation. Both of these proteins have functions that are connected with cell cycle progression, but it remains to be proven formally that this is the activity responsible for tumor suppression, and to characterize how its absence permits unrestrained growth.

Both the RB and p53 tumor suppressors are activated by multiple pathways. One important locus that influences both these tumor suppressors is INK4a-ARF. The transcript is alternatively spliced to give two mRNAs that code for proteins with no sequence relationship. p16^INK4a is upstream of RB. p19^ARF is upstream of p53. Deletions of the locus are common in human cancers (almost as common as mutations in p53), and have a highly significant effect, because they eliminate both p16^INK4a and p19^ARF and therefore lead to loss of both the RB and p53 tumor suppressor pathways.

p16^INK4a inhibits the cdk4/6 kinase (see Figure 27.25). So it prevents the kinase from phosphorylating RB. In the absence of this phosphorylation, progress through the cell cycle (and therefore growth) is inhibited. p16^INK4a is often inhibited by point mutations in human tumors.

p19^ARF antagonizes Mdm2, as shown in the lower part of Figure 28.28. This in turn leads to stabilization (and therefore increased activity) of p53. In effect, therefore, p19^ARF functions as a tumor suppressor by inhibiting the inhibitor of the p53 tumor suppressor. The nature of the interaction between p19^ARF and Mdm2 is not entirely clear: p19^ARF may promote degradation of Mdm2 or directly block its interaction with p53. At all events, p19^ARF arrests the cell cycle in a p53-dependent manner. Loss of p19^ARF or loss of p53 have similar effects on cell growth (and tumors usually lose one or the other but not both), suggesting that they function in the same pathway, that is, p19^ARF in effect functions exclusively through p53. The cellular oncogene C-myc, and the adenoviral oncogene E1, both act via p19^ARF to activate p53-dependent pathways.

p53 responds to environmental signals that affect cell growth, and many of these signals act by causing specific sites on p53 to be modified. The most common form of modification is the phosphorylation of serine, but acetylation of lysine also occurs. Different pathways lead to the modification of different amino acid residues in p53, as summarized in Figure 28.29. There is often overlap between the various residues activated by each pathway. For example, ionizing radiation activates the kinase ATM, which phosphorylates S¹⁵; and through unknown pathways, ionizing radiation also causes phosphorylation of S³³, dephosphorylation of S³⁷⁶, and acetylation of L³⁸². UV radiation shares a pathway with ionizing radiation to phosphorylate S¹⁵ and S³³, but also causes phosphorylation of S³⁹². The target sites for these various pathways are located in the terminal regulatory domains of the protein. The modifications may affect stability of the protein, oligomerization, DNA-binding, and binding to other proteins. So p53 acts as a sensor that integrates information from many pathways that affect the cell’s ability to divide.