, August 24, 2007 ©2007 Elsevier Inc.
gene, which belongs to neither group and encodes a reg-ulator o p53, is not aected either way. Following short-term etoposide treatment to induce mild DNA damage,Hz-bound p53 engages proarrest but not proapoptotictargets. However, in similarly treated Hz-decient cellsp53 is detected primarily on proapoptotic targets. Remark-ably, ater prolonged etoposide treatment, which infictsextensive DNA damage, Hz undergoes proteasomal deg-radation. The resulting reduction in Hz levels now enablesactivation o proapoptotic genes, providing an appealingexplanation or the observation that extended damage trig-gers a switch rom a growth inhibitory transcriptional pro-gram to a proapoptotic one (Figure 1). As anticipated romits remarkable ability to instruct p53 to distinguish betweenthe two classes o target genes, Hz has a proound impacton cell ate decisions downstream o p53 activation: in itsabsence, even a mild genotoxic insult is sucient to trig-ger apoptosis. It will be o particular interest to nd outto whether alterations in Hz expression or activity areinvolved in human tumors, particularly those that retain awild-type
Regulation o p53 by Covalent Modifcations
Covalent modications o p53 may also change targetgene preerences, possibly by imposing conormationalchanges in p53 that encourage selective recognitiono dierent p53REs. It has been suggested that p53mutants that are able to activate only a subset o targetsmay be “locked” into a particular conormation that onlyrecognizes particular types o promoters. Wild-type p53,however, is conceivably fexible enough to go betweendierent conormations, thereby allowing diverse pro-moter recognition (Kim and Deppert, 2003). The list oreported post-translational modications on p53 is longand continuously growing, and includes phosphoryla-tion o multiple serine (Ser) and threonine (Thr) residues,acetylation, mono- and polyubiquitylation, sumoylation,neddylation and more.Much recent attention has ocused on p53 phos-phorylation on Ser46, which specically avors trans-activation o proapoptotic genes (Shmueli and Oren,2007). Indeed, mutation o Ser46 to Ala reduces the abil-ity o p53 to transactivate proapoptotic genes such as
Perp and PUMA
and to triggerapoptosis, but not cell cycle arrest, in transected cellsas well as cells derived rom Ala46 knock-in mice (Fenget al., 2006; Oda et al., 2000). Interestingly, a naturallyoccurring p53-46F mutant mimics Ser46 phosphoryla-tion and specically induces p53 proapoptotic targetgenes, including
(Nakamuraet al., 2006), in keeping with the notion that phosphoryla-tion o Ser46 is key to p53 cell ate choice.Ser46 is the target o several kinases, including HIPK2,DYRK2, protein kinase C
and p38 (Shmueli and Oren,2007). These kinases preerentially phosphorylate Ser46in response to extensive DNA damage, thereby contrib-uting to the increased likelihood o cell death under suchconditions. Although sharing a common target on p53,the mechanisms that direct the individual kinases to p53upon severe genotoxic damage vary greatly. For instance,whereas such damage drives DYRK2 rom the cytoplasminto the nucleus, granting it access to its p53 substrate,HIPK2 benets rom a more intricate process, wherein itslevels are increased owing to its release rom Mdm2-medi-ated proteasomal degradation (Shmueli and Oren, 2007).How is target gene choice dictated by phosphorylationo Ser46 or o other residues lying outside the p53 DBD?One possibility is that such modications change the over-all conormation o p53, thereby also aecting the DBD. Alternatively, by modulating coactivator binding, they mayindirectly aect chromatin states in the vicinity o p53REs,avoring the activation o particular genes over others. Acetylation also plays a role in dictating the targetpreerences o p53. Lysine 120 (K120) o p53 is some-times mutated in human cancers. Remarkably, tumor-derived K120R mutations abrogate p53-mediated apop-tosis, but not cell cycle arrest. In response to severeDNA damage, K120 is acetylated by the MYST amily oacetyl transerases, MOF and TIP60 (Sykes et al., 2006;Tang et al., 2006). This acetylated orm accumulatespreerentially on proapoptotic promoters, such as
, and presumably serves to recruit other p53coactors necessary to override the transcriptional bar-riers in proapoptotic genes.Recently, it has been reported that Lysine 320 (K320)o p53 is also important or the lie-death decision. Com-petition between acytelation and ubiquitylation at thissite directs cell ate toward apoptosis or growth arrest,respectively (Le Cam et al., 2006). Although ubiquity-lation is oten used to mark proteins or proteasomaldegradation, modication o K320 by the E3 ubiquitinligase E4F1, which promotes K48-type ubiquitylation onchromatin-bound p53, competes with PCAF-mediatedacetylation and thereby causes activation o proarrestgenes such as
(Figure 1).Competition between acetylation and ubiquitylation alsooccurs on numerous additional lysines, located within theC-terminal part o p53. Depending on the extent o ubiqui-tylation, ubiquitylation may either promote p53 degrada-tion or export into the cytoplasm, in both cases reducingthe amount o nuclear p53 available or DNA binding. Con-versely, acetylation on these lysines can stabilize p53 andaugment its interaction with DNA within the nucleus. Atrst approximation, acetylation will thus benet selectivelythose target genes whose activation requires higher levelso p53. The dynamic nature o p53 acetylation, involving amultitude o histone acetyltranserases (HATs) and histonedeactylases (HDACs), endows it with enhanced potential tomodulate p53 target gene choice.
Manipulation o Chromatin by p53
Binding o p53 to promoter regions presumably recruitsactors that act locally on chromatin to “open” the geneor transcription. Thus, the p300/CBP HATs have beenimplicated as physiological regulators o p53-mediatedtranscription. In addition to targeting chromatin compo-