P53, the Cellular Gatekeeper for Growth and Division

The p53 gene and its protein product have become the center of intensive study ever since it became clear that slightly more than 50% of human cancers contain mutations in this gene. An extensive database (21Hollstein M Rice K Greenblatt M.S Soussi T Fuchs R Sorlie T Hovig E Smith-Sorensen B Montesano R Harris C.C Database of p53 gene somatic mutations in human tumors and cell lines.Nucleic Acids Res. 1994; 22: 3551-3555PubMed Google Scholar) catalogs these mutations in more than 50 different cell and tissue types, although some types of cancers never appear to select for p53 mutations (38Lutzker S Levine A.J A functionally inactive p53 protein in embryonal carcinoma cells is activated by DNA damage or cellular differentiation.Nature Med. 1996; 2: 804-810Crossref PubMed Scopus (198) Google Scholar). The nature of these genetic changes in cancer cells is most commonly a missense mutation in one allele, producing a faulty protein that is then observed at high concentrations in these cells, followed by a reduction to homozygosity. More rarely, deletions or chain-termination mutations in the p53 gene indicate that the null phenotype predisposes to cancer, as has been observed in mice with a homozygous p53 null mutation (13Donehower L.A Harvey M Slagle B.L McArthur M.J Montgomery Jr., C.A Butel J.S Bradley A Mice deficient for p53 are developmentally normal but susceptible to spontaneous tumours.Nature (Lond.). 1992; 356: 215-221Crossref PubMed Scopus (3905) Google Scholar). There have been some suggestions that the missense mutant producing a faulty p53 protein could contribute a “gain of function” phenotype (11Dittmer D Pati S Zambetti G Chu S Teresky A.K Moore M Finlay C Levine A.J p53 gain of function mutations.Nature Genet. 1993; 4: 42-46Crossref PubMed Scopus (757) Google Scholar), but this remains to be substantiated by additional experimentation. A study of the mutational spectra at the p53 locus in different tissue types indicates a strong role for diverse environmental mutagens, with a set of tissue preferences. In addition, there is strong selection for a subset of mutations localized predominantly in the DNA-binding domain of the protein. Several codons in this domain that are never observed in the p53 mutational spectra have been altered by site-specific mutagenesis, and in all cases these mutations have no phenotype and behave like the wild-type (32Lin J Wu X Chen J Chang A Levine A.J Functions of the p53 protein in growth regulation and tumor suppression.in: Cold Spring Harbor Symposia on Quantitative Biology LIX1995: 215-223Google Scholar). Thus, both selection and a strong set of environmental mutagens combine to produce mutations in the p53 gene in human cancers. The p53 protein is a transcription factor that enhances the rate of transcription of six or seven known genes that carry out, at least in part, the p53-dependent functions in a cell (Table 1). The human p53 protein contains 393 amino acids and has been divided structurally and functionally into four domains. The first 42 amino acids at the N-terminus constitute a transcriptional activation domain that interacts with the basal transcriptional machinery in positively regulating gene expression. Amino acids 13–23 in the p53 protein are identical in a number of diverse species. The p53 amino acids F19, L22, and W23 have been shown to be required for transcriptional activation by the protein in vivo (32Lin J Wu X Chen J Chang A Levine A.J Functions of the p53 protein in growth regulation and tumor suppression.in: Cold Spring Harbor Symposia on Quantitative Biology LIX1995: 215-223Google Scholar). These same amino acids make contacts with and bind to (in vitro) the TATA-associated factors TAFII70 and TAFII31, both of which are subunits of TFIID (37Lu H Levine A.J Human TAF-31 is a transcriptional coactivator of the p53 protein.Proc. Natl. Acad. Sci. USA. 1995; 92: 5154-5158Crossref PubMed Scopus (278) Google Scholar, 48Thut C.J Chen J.L Klemin R Tjian R p53 transcriptional activation mediated by coactivators TAFII40 and TAFII60.Science. 1995; 267: 100-104Crossref PubMed Scopus (404) Google Scholar). p53 transcriptional activation is negatively regulated by the adenovirus E1B-55Kd protein and the human MDM2 protein. In both cases, p53 amino acid residues 22 and 23 play a key role in the binding of p53 to E1B-55Kd or MDM2 (32Lin J Wu X Chen J Chang A Levine A.J Functions of the p53 protein in growth regulation and tumor suppression.in: Cold Spring Harbor Symposia on Quantitative Biology LIX1995: 215-223Google Scholar). Thus, the negative regulators of p53-mediated transcription target some of the same p53 amino acids critical to positive regulation of transcriptional activation.Table 1Products of Genes Transcriptionally Activated by p53p21, WAF1, Cip1Inhibits several cyclin–cyclin-dependent kinases; bind cdk's, cyclins, and PCNA; arrest the cell cycleMDM2Product of an oncogene; inactivates p53-mediated transcription and so forms an autoregulatory loop with p53 activityGADD45Induced upon DNA damage; binds to PCNA and can arrest the cell cycle; involved directly in DNA nucleotide excision repairCyclin GA novel cyclin (it does not cycle with cell division) of unknown function and no known cyclin-dependent kinaseBaxA member of the BCl2 family that promotes apoptosis; not induced by p53 in all cellsIGF-BP3The insulin-like growth factor binding protein-3; blocks signaling of a mitogenic growth factorWhile a large number of other genes have been suggested to be regulated by p53 (29Ko L.J Prives C p53 puzzle and paradigm.Genes Dev. 1996; 10: 1054-1072Crossref PubMed Scopus (2220) Google Scholar), those listed above all have been shown to contain p53-dependent, cis-acting, DNA-responsive elements. Open table in a new tab While a large number of other genes have been suggested to be regulated by p53 (29Ko L.J Prives C p53 puzzle and paradigm.Genes Dev. 1996; 10: 1054-1072Crossref PubMed Scopus (2220) Google Scholar), those listed above all have been shown to contain p53-dependent, cis-acting, DNA-responsive elements. These results clearly point out that p53 uses a hydrophobic interface in its N-terminal domain to interact with the transcriptional machinery of the cell and its negative regulators. Recently, the N-terminal domain of MDM2 was cocrystalized with a peptide containing p53 amino acid residues 13–29, covering the portion of p53 that was shown, by mutational analysis, to interact with MDM2. The MDM2 domain forms a deep hydrophobic pocket, and the p53 peptide forms an amphipathic helix with its hydrophobic surface, pointing into and filling the hydrophobic pocket. F19, W23, and L26 all stabilize these hydrophobic interactions in this pocket between p53 and MDM2 (30Kussie P.H Gorina S Marechal V Elenbaas B Moreau J Levine A.J Pavletich N.P Crystal structure of the MDM2 oncoprotein bound to the transactivation domain of the p53 tumor suppressor.Science. 1996; 274: 948-953Crossref PubMed Scopus (1676) Google Scholar). The sequence-specific DNA-binding domain of p53 is localized between amino acid residues 102 and 292. It is a protease-resistant and independently folded domain containing a Zn2+ ion that is required for its sequence-specific DNA-binding activity. This domain folds into a four-stranded and five-stranded antiparallel β sheet that in turn is a scaffold for two α-helical loops that interact directly with the DNA (6Cho Y Gorina S Jeffrey P.D Pavletich N.P Crystal structure of a p53 tumor suppressor-DNA complex understanding tumorigenic mutations.Science. 1994; 265: 346-355Crossref PubMed Scopus (2063) Google Scholar). The tetrameric p53 protein (which is a dimer of a dimer) binds to four repeats of a consensus DNA sequence 5′-PuPuPuC(A/T)-3′, and this sequence is repeated in two pairs, each arranged as inverted repeats such as →← →←, where → is the sequence given above. Residues K120, S241, R273, A276, and R283 make contacts with the phosphate backbone in the major groove, while K120, C277, and R280 interact via hydrogen bonds to the DNA bases. R248 then makes multiple hydrogen bond contacts in the minor groove of the DNA helix (6Cho Y Gorina S Jeffrey P.D Pavletich N.P Crystal structure of a p53 tumor suppressor-DNA complex understanding tumorigenic mutations.Science. 1994; 265: 346-355Crossref PubMed Scopus (2063) Google Scholar). More than 90% of the missense mutations in p53 reside in this sequence-specific DNA-binding domain, and these mutations fall into two classes. Mutations in amino acid residues such as R248 and R273, the two most frequently altered residues in the protein, result in defective contacts with the DNA and loss of the ability of p53 to act as a transcription factor. A second class of p53 mutations disrupts the structural basis of the β sheet and the loop–sheet helix motif that acts as a scaffold in this domain. These structural mutations alter the conformation of the p53 protein and produce a protein that now reacts with the monoclonal antibody PAb240, whose epitope (residues 212–217) is not accessible to the binding of this antibody in the native or wild-type structure. More than 40% of the missense mutations are localized to residues R175, G245, R248, R249, R273, and R282, which play a role in the structural integrity of this domain or the DNA contact sites directly (6Cho Y Gorina S Jeffrey P.D Pavletich N.P Crystal structure of a p53 tumor suppressor-DNA complex understanding tumorigenic mutations.Science. 1994; 265: 346-355Crossref PubMed Scopus (2063) Google Scholar, 21Hollstein M Rice K Greenblatt M.S Soussi T Fuchs R Sorlie T Hovig E Smith-Sorensen B Montesano R Harris C.C Database of p53 gene somatic mutations in human tumors and cell lines.Nucleic Acids Res. 1994; 22: 3551-3555PubMed Google Scholar). The native p53 protein is a tetramer in solution, and amino acid residues 324–355 are required for this oligomerization of the protein. The structure of this domain contains a dimer of a dimer with two β sheets and two α helices. The two dimers are held together by a large hydrophobic surface of each helix pair, which then forms a four-helix bundle. This tetramerization domain is linked to the sequence-specific DNA-binding domain by a flexible linker of 37 residues (287–323) (25Jeffrey P.D Gorina S Pavletich N.P Crystal structure of the tetramerization domain of the p53 tumor suppressor at 1.7 angstroms.Science. 1995; 267: 1498-1502Crossref PubMed Scopus (417) Google Scholar). The C-terminal 26 amino acids form an open (protease-sensitive) domain composed of nine basic amino acid residues that bind to DNA and RNA readily with some sequence or structural preferences (31Lee S Elenbaas B Levine A.J Griffith J p53 and its 14 kDa C-terminal domain recognize primary DNA damage in the form of insertion/deletion mismatches.Cell. 1995; 81: 1013-1020Abstract Full Text PDF PubMed Scopus (397) Google Scholar). There is considerable evidence demonstrating that the p53 protein derived from several sources requires a structural change to activate it for sequence-specific binding to DNA. This non–DNA-binding or latent form of p53 can be regulated by this basic C-terminal domain. Deletion of this domain, phosphorylation at residue S378 by protein kinase C or residue S392 by casein kinase II, or binding of antibody PAb421 (to residues 370–378) all activate site-specific DNA binding by the central domain (residues 102–292) of this protein (23Hupp T.R Lane D.P Allosteric activation of latent p53 tetramers.Curr. Biol. 1994; 4: 865-875Abstract Full Text Full Text PDF PubMed Scopus (299) Google Scholar). Short (20–39 nucleotides) single strands of DNA interacting with this C-terminal domain can also activate specific p53 DNA binding, but longer, double strands of DNA inhibit p53 sequence-specific binding through this region of the protein (24Jayaraman L Prives C Activation of p53 sequence-specific DNA binding by short single strands of DNA requires the p53 C-terminus.Cell. 1995; 81: 1021-1029Abstract Full Text PDF PubMed Scopus (346) Google Scholar). The C-terminal domain helps to catalyze the reassociation of single-stranded DNA or RNA to double strands. It also binds preferentially to DNA ends and to internal deletion loops in DNA as generated by replication errors that are then detected and fixed by mismatch-repair processes. There is some sequence or structural specificity in this binding (31Lee S Elenbaas B Levine A.J Griffith J p53 and its 14 kDa C-terminal domain recognize primary DNA damage in the form of insertion/deletion mismatches.Cell. 1995; 81: 1013-1020Abstract Full Text PDF PubMed Scopus (397) Google Scholar). Clearly, the C-terminal domain either sterically or allosterically regulates the ability of p53 to bind to specific DNA sequences at its central or core domain. Normally, in a cell, the p53 protein is kept at a low concentration by its relatively short half-life (about 20 min). The proteases responsible for this are not known, but some evidence has suggested that ubiquitin-mediated proteolysis plays a role. In addition to this low protein concentration, in some cells p53 probably also exists in a latent form, inactive for transcription. Under these conditions, the p53 protein must receive a signal or alteration to activate it to function. The upstream events or signals that flow to p53 are mediated by several stressful situations. Several different types of DNA damage can activate p53, including double-strand breaks in DNA produced by γ-irradiation and the presence of DNA repair intermediates after ultraviolet irradiation or chemical damage to DNA. This results in a rapid increase in the level of p53 in the cell and activation of p53 as a transcription factor. The p53 level increases because the half-life of the protein is lengthened and possibly because the rate of translational initiation of p53 mRNA in the cell is enhanced. This increase in p53 levels is proportional to the extent of DNA damage, but both the extent of increase and the kinetics of p53 enhancement differ for different types of radiation damage. The cell uses different functions and proteins to recognize different classes of DNA damage (such as breaks in the DNA and excision repair of ultraviolet-irradiation dimers) and different systems of enzymes to repair them. It is an attractive idea that the cellular proteins that recognize DNA damage may communicate with p53 and activate it. For example, cells in culture defective in the ATM gene (ataxia-telangiectasia) have a delayed and attenuated p53 response to ionizing radiation, suggesting that the ATM protein, which may recognize damaged DNA and which contains a protein kinase domain, may signal the p53 protein in this fashion (27Kastan M.B Zhan Q El-Deiry W.S Carrier F Jacks T Walsh W.V Plunkett B.S Vogelstein B Fornace Jr., A.J A mammalian cell cycle checkpoint pathway utilizing p53 and GADD45 is defective in ataxia-telangiectasia.Cell. 1992; 71: 587-597Abstract Full Text PDF PubMed Scopus (2880) Google Scholar). While a signal transduction pathway of this type linking p53 to DNA damage must be postulated, no clear-cut components of this pathway have been identified. Because the p53 protein itself can bind to DNA ends and excision-repair damage sites or internal deletion loops (31Lee S Elenbaas B Levine A.J Griffith J p53 and its 14 kDa C-terminal domain recognize primary DNA damage in the form of insertion/deletion mismatches.Cell. 1995; 81: 1013-1020Abstract Full Text PDF PubMed Scopus (397) Google Scholar), it is possible that both the p53 protein and a damage-detector protein are localized at the site of DNA damage and repair, where phosphorylation or other activating signals can then be processed (Figure 1). This concept has the advantage of requiring two independent checks (the damage-detector protein and p53) for the presence of DNA damage prior to a p53 functional response to such damage. DNA strand breaks in a cell appear to be sufficient for activating p53. The introduction of restriction enzyme nucleases into the nucleus of a cell stimulates p53 levels and activity. Transgenic mice defective in nucleotide excision repair (leaving repair intermediates) have elevated levels of p53 in several tissues. Mice defective in V-D-J receptor recombination (SCID mice) also have an activated p53 owing to the persistence of DNA recombination intermediates (19Guidos C.J Williams C.J Grandal I Knowles G Huang M.T.F Danska J.S V(D)J recombination activates a p53-dependent DNA damage checkpoint in scid lymphocyte precursors.Genes Dev. 1996; 10: 2038-2054Crossref PubMed Scopus (243) Google Scholar). In addition to DNA damage, hypoxia is able to stimulate p53 levels and activate the p53 protein (18Graeber A.J Osmanian C Jack T Housman D.E Koch C.J Lowe S.W Graccia A.J Hypoxia-mediated selection of cells with diminished apoptotic potential in solid tumors.Nature. 1996; 379: 88-91Crossref PubMed Scopus (2105) Google Scholar). It has been suggested that this process represents yet another way that p53 may act as a gatekeeper against the formation of cancers. Many tumors begin to replicate and reach a critical size when the blood supply becomes rate-limiting, requiring angiogeneic factors to sustain growth. The resultant hypoxia might trigger p53 activity and kill such cells. It has also been reported that the thrombospondin gene is a p53-regulated gene. If that is correct, thrombospondin is an antiangiogenic factor that could further reduce the blood supply to these tumors. Yet a third signal to activate p53 is sent when ribonucleoside triphosphate pools fall below a critical threshold. Clearly, the need to have normal nucleoside triphosphate pools to support DNA replication and progression through the cell cycle is monitored and reported to p53, but the pathway or proteins involved in this remain unclear (33Linke S.P Clarkin K.C DiLeonardo A Tsou A Wahl G.M A reversible, p53-dependent G0/G1 cell cycle arrest induced by ribonucleotide depletion in the absence of detectable DNA damage.Genes Dev. 1996; 10: 934-947Crossref PubMed Scopus (475) Google Scholar). Other signals of cellular distress that use p53 as an integrator of a response have been reported. Recognition of birth defects in response to teratogenic agents has been singled out as one possible role of p53 (42Nicol C Harrison M Laposa G Gimelshtein I Wells P A teratologic suppressor role for p53 in benzo[a] pyrene-treated transgenic p53-deficient mice.Nature Genet. 1995; 10: 181-187Crossref PubMed Scopus (208) Google Scholar). For example, in the absence of MDM2, the negative regulator of p53, mouse fetuses are aborted just after implantation, in a p53-dependent fashion (41Montes de Oca Luna R Wagner O Lozano G Rescue of early embryonic lethality in mdm2-deficient mice by deletion of p53.Nature. 1995; 378: 203-206Crossref PubMed Scopus (1122) Google Scholar). Implantation may trigger stressful signals responded to by p53 and modulated appropriately by MDM2. Mice nullizygous for the p53 gene develop giant cells (polyploid cells) in the testes, which could result from a lack of surveillance of recombination intermediates in spermatocytes. This could mean that one role of wild-type p53 is to detect abnormal intermediates in recombination and to eliminate such clones during spermiogenesis, as is more clearly the case with the T cell receptor recombination intermediates (19Guidos C.J Williams C.J Grandal I Knowles G Huang M.T.F Danska J.S V(D)J recombination activates a p53-dependent DNA damage checkpoint in scid lymphocyte precursors.Genes Dev. 1996; 10: 2038-2054Crossref PubMed Scopus (243) Google Scholar). Perhaps related to this observation is the finding that the most common tumor that arises in p53 nullizygous mice is a thymic lymphoma (13Donehower L.A Harvey M Slagle B.L McArthur M.J Montgomery Jr., C.A Butel J.S Bradley A Mice deficient for p53 are developmentally normal but susceptible to spontaneous tumours.Nature (Lond.). 1992; 356: 215-221Crossref PubMed Scopus (3905) Google Scholar). The downstream events mediated by p53 take place by two major pathways: cell cycle arrest and apoptosis. The p16–cyclin D1–cdk4–Rb pathway is central to the regulation of the G1-to-S phase transition and to the understanding of human cancers (Figure 2). One of these four genes is altered or mutated in nearly every cancer examined. p16 is a negative regulator of cyclin D1–Cdk4, and the gene is shut off (heavily methylated) in some cancer cells or mutated in other cancers. Cyclin D1 is amplified and overexpressed in a number of cancers (about 16% of breast cancers), and cdk4 mutations (no longer sensitive to p16) and cdk4 gene amplifications have been reported in selected tumors. The retinoblastoma protein (Rb) is the major target of cyclin D1–Cdk4 for cell cycle regulation and is also present in a mutant form in a number of cancers (such as small-cell lung cancer and osteosarcomas). The Rb protein regulates E2F–DP transcription factor complexes (E2F-1, -2, and -3, and DP-1, -2, and -3), which in turn regulate a number of genes (including those encoding cyclin E, cyclin A, dihydrofolate reductase, and proliferating cell nuclear antigen [PCNA]) required to initiate or propagate the S phase of the cell cycle. Phosphorylation of Rb by cyclin D1–Cdk4 releases E2F–DP proteins from the Rb complex, relieving repression of these genes or activating their transcription. The Rb protein regulates the restriction point or start, as a “go–no go” signal for cell cycle progression that is sensitive to the impact of various exogenous growth factors (via the regulation of cyclin D1–Cdk4 and possibly p16). In response to some forms of DNA damage, p53 is activated and turns on the transcription of one of its downstream genes, p21 (WAF1, Cip-1) (14El-Deiry W.S Tokino T Velculescu V.E Levy D.B Parsons R Trent J.M Lin D Mercer W.E Kinzler K.W Vogelstein B WAF1, a potential mediator of p53 tumor suppression.Cell. 1993; 75: 817-825Abstract Full Text PDF PubMed Scopus (7733) Google Scholar). p21 binds to a number of cyclin and Cdk complexes: cyclin D1–Cdk4, cyclin E–Cdk2, cyclin A–Cdk2, and cyclin A–Cdc2. One molecule of p21 per complex appears to permit Cdk activity (and may even act as an assembly factor), while two moles of p21 per complex inhibit kinase activity and block cell cycle progression. p21 also binds to PCNA (at its C-terminal domain). The available evidence suggests that p21–PCNA complexes block the role of PCNA as a DNA polymerase processivity factor in DNA replication, but not its role in DNA repair. Thus, p21 can act on cyclin–Cdk complexes and PCNA to stop DNA replication. Mice deficient in the p21 gene (null phenotype) develop normally, and mouse embryo fibroblasts derived from these mice are partially deficient in their ability to arrest cells in G1 in response to DNA damage (10Deng C Zhang P Harper J.W Elledge S.J Leder P Mice lacking p21CIP1/WAF1 undergo normal development, but are defective in G1 checkpoint control.Cell. 1995; 82: 675-684Abstract Full Text PDF PubMed Scopus (1890) Google Scholar). Based on these observations, there should be a p21-independent pathway that contributes to the p53-mediated G1 arrest. The overexpression of the product of the GADD45 gene (a p53-responsive gene) in some cells in culture arrests cells in G1. Perhaps GADD45, also binding to PCNA, plays this backup role for p21. By contrast, the removal of both p21 alleles from a cancerous cell line in culture that contained a wild-type p53 allele completely eliminated the DNA damage–induced G1 arrest in these cells, indicating that p21 is sufficient to enforce a G1 arrest in this experimental situation (44Polyak K Waldman T He T.-C Kinzler K.W Vogelstein B Genetic determinants of p53-induced apoptosis and growth arrest.Genes Dev. 1996; 10: 1945-1952Crossref PubMed Scopus (467) Google Scholar). Perhaps the particular set of mutations in this cancer cell line or the genetic background of these cells plays a role in the degree to which p21 alone can regulate the G1 checkpoint. While this model for p53-mediated G1 arrest (Figure 2) predicts that cells with a mutant or faulty Rb protein should not be blocked in G1 after DNA damage and p53 activation, the evidence that supports this rests heavily on the use of viral oncogenes, which undoubtedly have multiple functions. In these experiments, the SV40 T antigen, E1A proteins of adenovirus, or E7 of the human papilloma viruses are used to block Rb functions and test for p53-mediated G1 arrest. In the absence of Rb regulation of E2F–DP-1, p53 fails to arrest these cells. When a simpler protocol is used, such as mouse embryo fibroblasts from Rb−/− (null) mice that are irradiated in cell culture, they in fact undergo G1 arrest. One distinction between these experiments and those using viral oncogenes to eliminate Rb function is that such viral oncogene products also alter the other two Rb family members, p107 and p130, which can certainly play a role in a p53-mediated G1-arrest phenotype. Thus p21 inhibition of cyclin D1–Cdk4 and cyclin E–Cdk2 can still affect the activities of p107 and p130, which regulate E2F-4 and -5. These interrelationships must be understood better if we are to know how p53 regulates G1 arrest. Cells with wild-type p53 do not undergo gene amplification readily (34Livingstone L.R White A Sprouse J Livanos E Jacks T Tlsty T.D Altered cell cycle arrest and gene amplification potential accompany loss of wild-type p53.Cell. 1992; 70: 923-935Abstract Full Text PDF PubMed Scopus (1265) Google Scholar). The drug PALA (N-phosphonacetyl-L-aspartate), which is used to select for the amplification of the CAD gene, results in the depletion of pyrimidine triphosphate pools, which in turn activates p53, resulting in G1 arrest. Cells with mutant p53 enter S phase and can also amplify segments of DNA. Gene amplification is thought to arise from nonhomologous recombination in a bridge–breakage–fusion cycle. Although the mechanism by which gene amplification is curtailed and inhibited by p53 remains unclear, one possibility is that p53 plays a role in monitoring abnormal recombination intermediates and acts to kill such cells (19Guidos C.J Williams C.J Grandal I Knowles G Huang M.T.F Danska J.S V(D)J recombination activates a p53-dependent DNA damage checkpoint in scid lymphocyte precursors.Genes Dev. 1996; 10: 2038-2054Crossref PubMed Scopus (243) Google Scholar). More recently, p53 has been implicated in a G2/M phase checkpoint. When mitotic spindle inhibitors, such as nocodazole, are added to cells with wild-type p53, the cells are blocked in G2. In the absence of wild-type p53, these cells will reinitiate DNA synthesis, increasing the ploidy of the cells (7Cross S.M Sanchez C.A Morgan C.A Schimke M.K Ramel S Idzerda R.L Rasking W.H Reid B.J A p53-dependent mouse spindle checkpoint.Science. 1995; 267: 1353-1356Crossref PubMed Scopus (668) Google Scholar). These data suggest that p53 may be part of a G2/M checkpoint, preventing premature entry into another S phase. In addition, p53 appears to be an integral part of the process that regulates the number of centrosomes in a cell (16Fukasawa K Choi T Kuriyama R Rulong S Vande Woude G.F Abnormal centrosome amplification in the absence of p53.Science. 1996; 271: 1744-1747Crossref PubMed Scopus (716) Google Scholar). Mouse embryo fibroblasts from p53−/− (null) mice produce abnormal numbers of centrosomes (not observed in normal cells in culture at the same passage level) after a few doublings in cell culture and initiate spindles with three or four poles (16Fukasawa K Choi T Kuriyama R Rulong S Vande Woude G.F Abnormal centrosome amplification in the absence of p53.Science. 1996; 271: 1744-1747Crossref PubMed Scopus (716) Google Scholar). This surely contributes to the phenotype of p53−/− cells in culture that rapidly become aneuploid at times when cells with wild-type p53 remain diploid during their passages. Of note, p53 has been found to copurify with the centrosomes isolated from some cells in culture (3Brown C.R Doxsey S.J White E Welch W.J Both viral (adenovirus E1B) and cellular (hsp 70, p53) components interact with centrosomes.J. Cell. Physiol. 1994; 160: 47-60Crossref PubMed Scopus (99) Google Scholar). This p53-mediated G2/M checkpoint may account for the phenotype of genomic instability that is commonly associated with a p53 mutation. A clear example of this phenotype was demonstrated using the transgenic mouse carrying the MMTV-LTR-Wnt1 transgene that promotes breast cancer. Crossing this transgene with a p53 −/− mouse resulted in a more rapid appearance of breast tumors, and these tumors were markedly more aneuploid, demonstrating the p53 null genomic instability phenotype in vivo (12Donehower L.A Godley L.A Aldaz C.M Pyle R Shi Y.-P Pinkel D Gray J Bradley A Demina D Varmus H.E Deficiency of p53 accelerates mamary tumorigenesis in Wnt-1 transgenic mice and promotes chromosomal instability.Genes Dev. 1995; 9: 882-895Crossref PubMed Scopus (245) Google Scholar). Evidence for a third cell cycle checkpoint for p53 comes from an examination of the G0–G1–S phase transition. The Gas1 gene is a membrane protein that keeps cells in G0 arrest and is expressed only at that time. Gas1 can act to place cells into a G0 arrest only when wild-type p53 is present in such cells. In this case, however, p53 is not acting as a transcription factor regulating its downstream genes. The codon 22, 23 mutant of p53, which cannot act as a transcription factor, still functions to transmit Gas1 signals for G0 arrest (9Del Sal G.D Ruaro E.M Utrera R Cole C.N Levine A.J Schneider C Gas1-induced growth suppression requires a transactivation-independent p53 function.Mol. Cell. Biol. 1995; 15: 7152-7160Crossref PubMed Scopus (95) Google Scholar). This brings up the possibility that p53 can function not as a transcription factor but in a second