Oncogene (2006) 25, 4725–4743 & 2006 Nature Publishing Group All rights reserved 0950-9232/06 $30.00 www.nature.com/onc REVIEW p53 and Nur77/TR3 – transcription factors that directly target mitochondria for cell death induction

UM Moll1, N Marchenko1 and X-k Zhang2

1Department of Pathology Stony Brook University, Stony Brook, New York, USA and 2The Burnham Institute for Medical Research, Cancer Center, La Jolla, CA, USA

The complex apoptotic functions of the p53 tumor of p53 – pleiotropic in its nature – is central to its suppressor are central to its antineoplastic activity in antineoplastic activity. p53 responds to death stimuli by vivo. Conversely, p53 function is altered or attenuated in rapid stabilization and activation. It mediates apoptosis one way or another in the majority of human cancers. primarily via the mitochondrial pathway. p53 can act as Besides its well-understood action as a transcriptional a transcription factor to stimulate the expression of regulator of multiple apoptotic genes, p53 also exerts a proapoptotic target genes like Puma, Noxa, Bax, Bid, direct pro-apoptotic role at the mitochondria by engaging p53AIP1 and many others, often in a cell- and stress- in protein–protein interactions with anti- and pro-apopto- type preferred way (Vousden and Lu, 2002). p53 can tic Bcl2 family members, thereby executing the shortest also mediate transrepression of anti-apoptotic genes known circuitry of p53 death signaling. Nur77, also such as Bcl2 and survivin. Moreover, complex negative known as TR3 or NGFI-B, is a unique transcription factor feedback influences exist among p53’s multiple gene- belonging to the orphan nuclear receptor superfamily. regulatory functions that impact on the final biological Even more extreme than p53, Nur77 can exert opposing outcome of the DNA stress response. For example in biological activities of survival and death. Its activities are lethally irradiated hematopoietic progenitor cells p53 regulated by subcellular distribution, expression levels, transactivates the transcriptional repressor Slug that in protein modification and heterodimerization with retinoid turn then directly represses the p53-mediated transcrip- X receptor. In cancer cells, Nur77 functions in the nucleus tion of Puma (Wu et al., 2005). as an oncogenic survival factor, but becomes a potent killer when certain death stimuli induce its migration to mitochondria, where it binds to Bcl2 and conformationally Multiple death signals induce mitochondrial translocation converts it to a killer that triggers cytochrome c release of p53 with rapid kinetics and apoptosis. This review focuses on their unexpected transcription-independent pro-death programs at mito- The mitochondrial death pathway plays a pivotal role in chondria and highlights the remarkable mechanistic p53-dependent apoptosis (Schuler and Green, 2001). In similarities between them. Moreover, an accumulating the last 5 years solid evidence was gathered that in body of evidence provides ample rationale to further addition to the varied gene-regulatory mechanisms, p53 investigate how these mitochondrial p53 and Nur77 translocation to mitochondria with direct outer mem- pathways could become exploitable targets for new cancer brane permeabilization is a common early component in therapeutics. p53-mediated apoptosis The earliest reports of a Oncogene (2006) 25, 4725–4743. doi:10.1038/sj.onc.1209601 transcription-independent killing action of p53 came from seeing unimpeded apoptosis despite pharmacolo- Keywords: p53; Nur77/TR3; mitochondria gical inhibition of transcription or translation in strictly p53-dependent cellular death systems. Apoptosis after overexpression of transactivation-deficient p53 mutants in null cells supported to this notion. While a mechan- istic understanding for these p53 effects were lacking for p53 regulates apoptosis as a transcription factor a long time, this was rapidly changed in 2000 with the discovery that a fraction of stress-stabilized wild-type The p53 tumour suppressor is a crucial protector against p53 (wtp53) protein traffics to mitochondria at the onset cancer by eliminating cells with mutagenic alterations or of p53-dependent apoptosis (Marchenko et al., 2000). by blocking their cell cycle permanently or transiently by Stress-induced mitochondrial p53 accumulation oc- allowing DNA repair. The powerful apoptotic function curs in many cellular systems. It takes place in radio- sensitive tissues such as thymus, spleen and testis in vivo Correspondence:Dr Ute M. Moll, Department of Pathology, Stony (Erster et al., 2004). It also occurs in cultured primary Brook University, Stony Brook, New York 11794-8691, USA. cells like mouse thymocytes, human mammary epithelial E-mail:[email protected] (HMEC), umbilical vein endothelial (HUVEC) cells, p53 and Nur77/TR3 UM Moll et al 4726 dermal fibroblasts and skin keratinocytes (Mihara et al., Mitochondrial p53 translocation occurs with fast 2003; Bonafe et al., 2004; Nemajerova et al., 2005b; kinetics. For example, in cells of hematopoietic Zhao et al., 2005), as well as in immortal and and lymphopoietic lineage this occurs within 30–60 transformed cell lines of all three germ lines (Marchenko min after the initial stress stimulus and prior et al., 2000; Sansome et al., 2001; Dumont et al., 2003; to detectable nuclear accumulation (Marchenko et al., Leu et al., 2004; Arima et al., 2005). 2000; Sansome et al., 2001; Mihara et al., 2003; Erster The phenomenon is induced by many different et al., 2004; Zhao et al., 2005). p53 translocation stress signals:DNA-damaging agents including chemo- precedes cytochrome c release from the mitochondrial therapeutics, UV and g-IR; hypoxia, viral and cellular intermembranous space, the collapse of the inner oncogenes, transcriptional blockade and reactive oxygen mitochondrial membrane potential and procaspase-3 species-generating agents (Marchenko et al., 2000; activation (Marchenko et al., 2000). In radiosensitive Sansome et al., 2001; Mihara et al., 2003; Arima et al., organs (such as thymus, spleen and testis) but 2005; Nemajerova et al., 2005b; Zhao et al., 2005). not radioresistant organs (such as liver and kidney) On the other hand, mitochondrial p53 translocation from normal mice treated with a single clinical dose does not occur in p53-independent death via for of g-irradiation or etoposide, mitochondrial p53 trans- example the TNF pathway, although mitochondria are location triggers rapid caspase-3 activation and apop- also downstream components of the TNF/DISC/pro- tosis in a subpopulation of cells. Of note, in thymus this caspase-8 pathway (Marchenko et al., 2000). Of note, precedes the transcriptional induction of Puma, the stable overexpression of BclXL and Bcl2 in myeloid fastest known apoptotic target gene of p53 and progenitor cells (32D) inhibits DNA damage-induced most relevant in this setting (i.e. irradiated thymus) p53 mitochondrial translocation (but not its overall (Erster et al., 2004). This suggests that the mitochondrial cellular stabilization) as well as p53-dependent apopto- pathway also participates in tumor suppression sis, while both events do occur in vector control cells in vivo. (Marchenko et al., 2000). This was an early hint at the molecular targets of p53 at the mitochondria (see below). Localization of mitochondrial p53 Further support for an apoptotic function of mito- chondrial p53 came from the study of a common Like several other famous mitochondrial proteins polymorphism (Arg/Pro at codon 72) of human p53 (i.e. cytochrome c), p53 lacks classical mitochondrial- (Dumont et al., 2003; Leu et al., 2004). In a series of targeting sequences. Yet, mitochondrial p53 is primarily engineered isogenic SaOs2 cell lines as well as in localized at the outer mitochondrial membrane homozygous Arg and Pro melanoma lines, the Arg (OMM), as demonstrated by immunogold electron version of p53 consistently shows five to 10 fold higher microscopy and controlled trypsin digestion of mito- apoptotic ability. This differential apoptotic activity chondria isolated from stressed cells. The majority strongly correlates with their differential abilities to of stress-translocated mitochondrial p53 is trypsin- localize to mitochondria, but not with their transcrip- sensitive and digested with the same rapid kinetics tional and transrepressor abilities nor their abilities to (10 min) as the prototypical OMM protein BclXL, bind to Mdm2. The differences in the apoptotic while cytochrome c in the intermembranous space is potentials of the Arg/Pro p53 variants are eliminated protected (Marchenko et al., 2000; Sansome et al., when these proteins were coupled to a leader sequence 2001). A small subfraction of translocated p53, however, that directs both p53 variants equally well to mitochon- is trypsin resistant, indicating an intramitochondrial dria (see below) (Dumont et al., 2003). Of note, further localization in the matrix (see below) (Achanta et al., functional in vivo relevance of the codon 72 polymorph- 2005; Zhao et al., 2005). This result is supported by the ism is provided by an Italian study of a group of old detection of stress-inducible endogenous complexes patients (66–99 years of age) affected by acute myocar- between p53 and the major mitochondrial import dial ischemia (Bonafe et al., 2004). Compared to Pro proteins localized in the mitochondrial matrix, mt patients, Arg patients show increased levels of Troponin hsp70 and mt hsp60 (the ATP-hydrolysing endopepti- I and CK-MB (creatin kinase-muscle brain isoenzymes), dases responsible for refolding imported proteins) two serum markers that correlate with the extent of the (Marchenko et al., 2000; Dumont et al., 2003; Kaul ischemic damage. Moreover, in response to oxidative et al., 2005). stress fibroblasts and lymphocytes from Arg/Arg The precise signal on p53 for its mitochondrial individuals have a higher level of mitochondrially translocation remains to be defined. A translocation localized p53 and p53/BclXL complexes and accord- motif within the p53 amino-acid sequence was not ingly, undergo a higher rate of apoptosis than Pro/Pro found and phosphorylation/acetylation modifications individuals, especially seen in old people. Together, of p53 are not determinants for mitochondria transloca- these data provide support that differences in the tion (Nemajerova et al., 2005a). Studies on the apoptotic potential of the two variants, mediated by Arg/Pro codon 72 variants proposed that Mdm2- their differential ability to target to mitochondria, mediated polyubiquitylation of p53 mediates nuclear contributes to a genetically determined variability in export and mitochondrial translocation (Dumont et al., apoptotic susceptibility among older people in the 2003), but this area requires more work for a better context of myocardial ischemia (Bonafe et al., 2004). understanding.

Oncogene p53 and Nur77/TR3 UM Moll et al 4727 p53 permeabilizes the outer mitochondrial membrane and mitochondrial death program in parallel to its nuclear induces the release of potent apoptotic activators like action. The fact that ‘conventional’ p53 commands both cytochrome c, SMAC and AIF the nuclear and the mitochondrial death programme explains why it is more potent than the forcibly targeted Nuclear bypass experiments in cultured p53 null cells p53. first proved causality and revealed the functional This pro-apoptotic action of mitochondrial p53 is due consequences between mitochondrial p53 and the to its autonomous ability to permeabilize the outer induction of apoptosis (Marchenko et al., 2000; Talos mitochondrial membrane, without the need for addition et al., 2005). Various p53 fusion proteins with mito- of any other protein. This was definitively shown in vitro chondrial targeting motifs at its N- or C-terminus in classic release assays on freshly isolated unstressed efficiently delivered p53 to mitochondria while being mitochondria from mouse liver, originally introduced devoid of any detectable nuclear transcriptional func- for the prototypical Bax and Bak proteins. When added, tion. The first mitochondrial targeting motif used for highly purified recombinant wtp53 protein (without any wtp53 was the prototypical import leader of human targeting motifs), but not the C-terminus of p53 or ornithine transcarbamylase (the fusion was called another irrelevant purified protein, induces the robust Lp53wt) (Marchenko et al., 2000). More recently, and rapid release of mitochondrial cytochrome c, Smac/ targeting motifs derived from the transmembrane Diablo, endonuclease G and apoptosis inducing factor domains of Bcl2 and BclXL (called p53CTM and (AIF) from the intermembranous space. This occurs p53CTB, respectively) were employed to target p53 rapidly and is dose-dependent. After 5 min, 70% of the directly to the OMM (Mihara et al., 2003; Talos et al., total cytochrome c content is released and within 30 min 2005). The targeted p53 proteins are undetectable in the process is complete (Mihara et al., 2003; Leu et al., nuclei of transfected cells, suggesting a lack of tran- 2004) (Moll et al., unpublished results). Mouse or scriptional activity. This was confirmed by sensitive human p53 protein of bacterial or baculoviral origin luciferase reporter assays driven by concatamerized work equally well in these assays. The latter indicates consensus p53-binding elements or the endogenous that once delivered, posttranslational modifications of p53-responsive promoters of PUMA and PIG3. More- p53 are not involved in regulating its mitochondrial over, mito-targeted p53 failed to induce endogenous p53 permeabilization function (Nemajerova et al., 2005a). target genes of the apoptotic and arrest categories, including Noxa, Bax and p21Waf1 (Mihara et al., 2003). When delivered to p53-null human and mouse tumor Three interrelated mitochondrial p53 arms cell lines in culture, the apoptotic ability of these targeted p53 versions is generally comparable although Wild type but not tumor-derived mutant p53 directly binds somewhat weaker than conventional nuclear p53 to and antagonizes the anti-apoptotic BclXL and Bcl2 (Mihara et al., 2003; Talos et al., 2005). This is seen proteins after transfection (supraphysiologic levels) into lung Mitochondrial wtp53 engages in inhibitory complexes carcinoma H1299 and osteosarcoma SaOs2 cells as well with anti-apoptotic BclXL and Bcl2 proteins, as shown as with physiological expression levels after retroviral by biochemical, mutational, molecular modeling, struc- transduction in primary B-lymphoma cells. Moreover, tural and biophysical studies (Mihara et al., 2003; mitochondrially targeted p53 is able to suppress growth Bonafe et al., 2004; Leu et al., 2004; Petros et al., in long-term colony suppression assays in several p53 2004) (Moll et al., unpubl. results). This makes eminent null cancer lines (Mihara et al., 2003). In contrast, the sense since the protective Bcl2 and BclXL proteins are transcription factor c-Rel, when targeted to mitochon- constitutively anchored at the OMM and mediate a dria using the same L- motif, has only background critical survival function by stabilizing the OMM and apoptotic activity (Marchenko et al., 2000). Conversely, inhibiting the release of cytochrome c and other wtp53 targeted to the endoplasmic reticulum (via the apoptogenic factors. Bcl2 and BclXL are the closest leader sequence of cytochrome b5) is devoid of any structural and functional homologues in the Bcl2 family apoptotic activity (Talos et al., 2005). The C-terminus of with 45% sequence identity and 56% similarity, includ- p53 harbors the tetramerization domain, several nuclear ing their strong anti-apoptotic functions and their localization and a nuclear export signal, as well as affinities to bind BH3-only proteins (Petros et al., multiple post-translational regulatory sites for protein 2001). The expression levels of the two proteins in cells stabilization and activation that are critical for p53’s are also interrelated, that is the overexpression of one function as a transcription factor. Of note, this region is represses the other (Arriola et al., 1999). In wtp53- fully dispensable for p53’s mitochondrial action (March- expressing cells exposed to short-term DNA damage, enko et al., 2000), a fact that might potentially become specific endogenous p53/BclXL and p53/Bcl2 complexes of future therapeutic relevance by escaping dominant form at purified mitochondria (which are free of nuclear negative interference by tumor-overexpressed mutant contamination). These complexes are readily detectable p53 or delta Np63/p73 isoforms (see below). Together, by standard co-immunoprecipitations in CHAPS or these results clearly indicate that, although not as NP40-containing buffers, by immunoaffinity chromato- efficient as nuclear p53, mitochondrial p53 alone is graphy and by co-purification in glycerol gradients, both sufficient to launch apoptosis and growth suppression from stressed cells and from isolated unstressed mito- directly from the mitochondria, thus runs a direct chondria after incubation with p53 protein. Complexes

Oncogene p53 and Nur77/TR3 UM Moll et al 4728 are enriched in the mitochondrial membranes compared quantitated by surface plasmon resonance studies to the soluble mitochondrial fraction. They are also (BIAcore), lie in the hundred nanomolar ranges and formed between exogenously expressed components are consistent with immunoprecipitable endogenous upon transfection in p53 null cells. A direct interaction complexes. Bcl2 interacts with wtp53DBD with a between p53 and BclXL was confirmed by in vitro pull KD1 ¼ 535724 nM (Moll et al., submitted). A slightly down assays of recombinant components. In contrast, a stronger affinity is reported for full-length wtp53 and p53/Bax complex is never detectable, even after triple BclXL (KD ¼ 164754 nM) (Chipuk et al., 2005). transfection of targeted p53, BclXL and Bax, and Bax is Importantly, these studies imply that the DNA- unable to interfere with the formation of p53/BclXL binding domain of p53 is a ‘dual-function domain’, (Mihara et al., 2003). mediating both the transactivation and the direct Mutational structure/function studies and protein mitochondrial pro-apoptotic function. Moreover, in modeling using a novel docking algorithm predicted cytochrome c release assays from unstressed mitochon- that the DNA-binding domain of p53 mediates the dria OMM permeabilization by wtp53 is dependent interaction with BclXL and Bcl2. The p53 DNA-binding upon p53/BclXL complex formation. This prediction domain with its overall basic character interacts with the was studied with tumor-derived missense mutants of p53 overall acidic ‘bottom side’ of BclXL. This BclXL site is in vitro and within cells. Indeed, all tested tumor-derived opposite from the hydrophobic binding groove formed and transactivation-deficient missense mutants of p53 by BH1, BH2 and BH3 domains that accept BH3- concomitantly lose or severely compromise their ability domain containing proteins and is composed by the a1 to interact with BclXL and Bcl2, as well as to promote helix/BH4 domain plus additional downstream domains OMMP. Thus, tumor-derived p53 mutations may of BclXL. Of note, in agreement with the consistent represent a positive selection of ‘double hits’, eliminating inability to detect stable p53/Bax complexes (Mihara the transcriptional as well as the direct mitochondrial et al., 2003; Chipuk et al., 2004), the corresponding functions of p53 (Mihara et al., 2003) (Moll et al., underside of Bax (nor of Bak, see below) do not feature submitted). Conversely, the mechanism of mitochon- acidic domains capable of binding p53. Structure/ drial wtp53 involves BH3-only like actions of derepres- function deletion and site-directed mutagenesis studies sion or ‘enabling’, that is inhibiting the anti-apoptotic revealed that the major contact site on p53, comprised Bcl2 members (Chen et al., 2005). By engaging in of residues 239–248 (part of the L3 loop), is flanked by inhibitory complexes with protective Bcl2 and BclXL, two lateral domains on either side, contributed by p53 indirectly activates pro-apoptotic executioner residues 135–141 (L1 loop) and residues 173–187 (part BH123 proteins like Bak and Bax by freeing them (see of L2 loop and H1 helix) (Mihara et al., 2003). Deletion below). of the aa239–248 region in the context of targeted L-wtp53 completely abolishes BclXL complexes as well as apoptosis and colony growth suppression. Deletion of p53 directly activates the apoptotic effector Bak in the any of the flanking regions produces an intermediate OMM phenotype in any of these three functions. Together, this indicates that p53/BclXL complex formation is required In addition to this indirect mechanism, a second for p53-mediated OMM permeabilization. interaction target of mitochondrial p53 is Bak itself, a Structural and biophysical studies subsequently vali- critical effector BH123 protein of the OMM. Here, dated the predicted interaction sites and determined direct binding of p53 to Bak yields its homo-oligomer- their affinities. Solution NMR spectroscopy studies by ization and activation. This was first seen in vitro:when Fesik et al., and later by our group experimentally added to unstressed mitochondria, purified p53 protein determined the p53/BclXL (Petros et al., 2004) (Moll induced homo-oligomerization of endogenous Bak, et al., unpublished) and p53/Bcl2 interactions (Moll detectable by cross-linkers. p53 is at least as efficient et al., submitted). The isolated DNA-binding domain of as the prototypical BH3-only protein tBid (Mihara bacterially expressed human wtp53 and various versions et al., 2003). The precise mechanism involves a musical of BclXL or Bcl2 (with or without their transmembrane chair scenario:in healthy cells, Bak is kept inactive by domains and the long loop between BH4 and BH3) were complexing with the anti-apoptotic Mcl-1 or BclXL in used. Neither p53 nor BclXL or Bcl2 undergo major the OMM. Bak induces apoptosis only if freed from conformational changes upon binding. The chemical both (Willis et al., 2005). Upon stress, p53 translocates shift changes on p53 caused by the BclXL interaction to mitochondria and liberates pro-apoptotic Bak from are at residues Arg-280, Cys-277, Ser-241, Cys-238, its sequestration by Mcl-1 by forming a p53/Bak Cys-182 and Val-173 of the DNA contact surface. The complex, thus promoting Bak oligomerization (Leu reciprocal changes on BclXL upon p53 addition are on et al., 2004). Treatment of isolated unstressed mitochon- residues in the C-terminal end of the BH4 domain and dria with recombinant p53 protein reduces the amount the two loops between the a3 and a4 helix and the a5 of immunoprecipitable Mcl-1/Bak complexes and reci- and a6 helix (Petros et al., 2004) (Moll et al., unpub- procally increases the amount of p53/Bak complexes. lished). For the p53/Bcl2 complex, the affected p53 Moreover, addition of p53 to Bak þ / þ mitochondria, residues were partly distinct but again within the DNA but not BakÀ/À mitochondria isolated from mouse contact site (Moll et al., submitted). The affinities of the embryo fibroblasts induces OMM in vitro (Dumont physical interactions between BclXL/Bcl2 and wtp53, et al., 2003). We confirmed the existence of specific p53/

Oncogene p53 and Nur77/TR3 UM Moll et al 4729 Bak complexes in stressed cells (Moll UM et al., recombinant Bax in vitro, mapping this function to the unpublished). Deletion mapping narrowed the region PP domain (Chipuk et al., 2004). To what degree the PP of the Bak-interaction site on p53 to residues 92–160, and the DBD domains of p53 cooperate in their overall thus distinct from the site used for contacting BclXL mitochondrial function is unknown but it is tempting to and Bcl2 but again within the DNA-binding domain speculate that each function is mediated by a separate (Leu et al., 2004). More work needs to be carried out to domain. Similar conclusions on the general importance better understand the significance of the p53/Bak of transcription-independent apoptotic p53 function complex. For example, its structure waits to be solved were drawn from studies in mouse fibroblasts NIH3T3. experimentally. As Bak is lacking a BH4 domain as well Upon low-dose UV-irradiation the primary, transcrip- as acidic residues in the loops between a3 and a4, its tion-mediated p53 response is terminal growth arrest. structure likely differs from that of the p53/BclXL High-dose UV-irradiation induces apoptosis. Induction complex. Also, the impact of p53 missense mutations on of p53-dependent apoptosis requires an additional, the ability to form a Bak complex has not been tested. transcription-independent p53 activity provided by high The physiological significance of the p53-mediated Bak cytoplasmic levels of activated p53, and correlates with liberation from Mcl-1 sequestration as proposed in this Bax activation (Speidel et al., 2005). model poses interesting questions, since Mcl-1 is unique Subsequent work by Chipuk et al. (2005) expanded among the pro-survival Bcl2 members. In contrast to the cytosolic p53 model to integrate both transcription- BclXL and Bcl2, Mcl-1 has a short half-life and its independent and transcription-dependent functions of expression levels are highly regulated. It is rapidly p53. The proposal is that stress-induced cytosolic p53 is induced by a wide range of survival signals, but also initially sequestered into an inactive complex by rapidly downregulated during apoptosis by proteolysis, cytosolic BclXL. In a second phase, which requires the mediated by the E3 ubiquitin ligase Mule/ARF-BP1, p53-mediated transactivation of its target gene PUMA, caspases and other proteases (Zhong et al., 2005). This p53 is then liberated from this cytosolic BclXL embrace might obviate the need for a p53-mediated liberation of by a newly formed PUMA/BclXL complex. Liberated Bak in many circumstances. Stable Mcl-1 expression, p53 can then directly activate Bax. This model is based however, is required for embryonic development and for on work in UV-treated fibroblasts and HCT116 colon the homeostasis of immune and hematopoietic systems, carcinoma cells using high doses of UV (up to 4000 J/m2) as shown by knockout studies. Thus, mitochondrial and rests upon the phenotype of a mutant of BclXL that p53-mediated Bak liberation might be particularly still binds p53, but not PUMA (BclXL G138A, which important in those cell systems. does not bind to proapoptotic BH3-only and BH123 proteins). While expression of wild-type and mutant BclXL suppressed p53-mediated apoptosis, additional p53 can directly activate cytosolic Bax overexpression of Puma only reversed the effect of wild- type BclXL, suggesting that in this system p53 needs to A third transcription-independent p53 arm was de- be released from Puma to activate Bax. Some open scribed in UV-treated transformed mouse embryonic questions are how to integrate into this model their fibroblasts, representing those cell types that express a earlier results stressing transcription-independence of subfraction of BclXL in the cytosol (Chipuk et al., 2004, Bax activation by cytosolic p53 (Chipuk et al., 2004), its 2005). Here, p53’s primary action lies in activating precise Bax activation mechanism and whether PUMA cytosolic Bax by liberating it from sequestration by is absolutely required for the extranuclear activity of cytosolic BclXL, with p53 binding to BclXl instead. The p53. Evidence from our lab indicates that at least the major BclXL fraction resides in the OMM (Kaufmann mitochondria-based death function of p53 is auto- et al., 2003; Gonzalez-Garcia et al., 1994). Bcl2 is nomous and does not require Puma (Moll et al., exclusively membrane-bound and does not participate in unpublished). Also, functional overlap with mitochon- this arm (Hsu et al., 1997)). Cytosolic p53 activates drial action can be envisioned:likely, liberated p53, monomeric Bax by inducing its homo-oligomerization either before or after its transient encounter with Bax, (and presumed mitochondrial translocation) in a pro- also translocates to mitochondria to act as described posed ‘hit and run’ mechanism that does not involve above. Moreover, this model does not preclude that stable p53/Bax complexes (Chipuk et al., 2004). transactivated Puma also liberates p53 from its BclXL Transcription-independence of this p53 action in this embrace at mitochondria, which then activates Bax and UV fibroblast system was extensively demonstrated, Bak within the OMM. using protein synthesis inhibitors, nuclear import Importantly, while distinct in details, there is a inhibition (wheat germ agglutinin), enucleated cytoplast common mechanistic theme underlying all three arms preparations and a transcription-deficient p53 mutant of extranuclear p53 (Figure 1):engagement of the (QS) (Chipuk et al., 2004). In contrast to the previous shortest death circuitry via physical interaction with two arms relying on the p53 DBD, here the N-terminus BH3-domain proteins, resulting in direct OMM per- of p53, comprising only the first 102 amino acids, is meabilization. Thus, mitochondrial and cytoplasmic p53 sufficient to mediate Bax activation and apoptosis. actions combined have properties of a ‘super-BH3-only’ Conversely, deletion of the polyproline-rich domain molecule:a dual action of both derepressing the anti- (PP, 62–91) from p53 abolishes p53’s ability to displace apoptotic and activating the pro-apoptotic OMM prebound Bid or Bax from BclXL and to activate regulators (Chen et al., 2005; Kuwana et al., 2005).

Oncogene p53 and Nur77/TR3 UM Moll et al 4730 Healthy Stress

Bax Bax B ax B Bax B ax a x p53 Bak p53 Bak Bak Bak BclXL Bak Mcl 1 p53 B Bak + id Puma Mcl 1 Ba Arf-BP1 Bax k Bc lXL Bax Mcl 1

k

a B

B P p53 ax

u B

L m BclXL a

B a x

X c B

cl B

lX a Degradation

B x

L Pu a ma x Puma BclXL

Figure 1 The mitochondrial pro-apoptotic activities of p53. In healthy cells (left), the pro-apoptotic BH3-only and effector BH123 proteins (ovals) exist either in complex with anti-apoptotic proteins Bcl2 family proteins such as BclXL, Bcl2 and Mcl-1 (rectangles). Death stimulus induced mitochondrial p53 (triangle) forms a complex with BclXL and Bcl2, liberating pro-apoptotic BH3-only proteins (derepression by inhibiting their survival function) and BH123 proteins, resulting in Bax and Bak homo-oligomerization, activation and mitochondrial membrane permeabilization (MMP) to release a host of apoptotic activators. Moreover, translocated p53 can bind to Bak by disrupting the inhibitory Bak/Mcl1 complex, enabling Bak to ultimately undergo homo-oligomerization and MMP. In addition, in cells with a cytosolic BclXL fraction, p53 might activate cytoplasmic Bax through a ‘hit and run’ mechanism involving an intermediate cytosolic p53/BclXL complex that is disrupted by cytosolic Puma, which results in Bax conformational change, mitochondrial translocation, oligomerization and MMP.

Interestingly, the short H2 helix of p53 close to the is rapid (within 30 min after IR in thymus and spleen) C-terminal end of the DBD shows a remarkable and triggers an early wave of caspase-3 activation sequence homology to the BH3 consensus motif and apoptosis. This caspase-3-mediated apoptosis is (Mihara et al., 2003). Phsyiologically, the nuclear, strictly p53 dependent and precedes p53 target gene mitochondrial and cytoplasmic functions of p53 are activation. In 5 Gy irradiated thymus at the 1 h time likely to cooperate, thereby integrating into the full force point, mitochondrial p53 accumulation is already of p53-mediated apoptosis. very robust and coincides with cleaved caspases in a fraction of cells (leading to TUNEL positivity by 2 h). In contrast, total cellular and nuclear p53 only begins to be The stress-induced direct mitochondrial p53 death induced at that time (and peaks only at 3 h). Accord- program: its role in physiologic and pathophysiologic ingly, the transcriptional p53 program has a longer lag responses in vivo after genotoxic, hypoxic and oxidative phase than the rapid mitochondrial p53 program. In damage. thymus, PUMA is the earliest and physiologically most relevant p53 product in this setting (Erlacher The direct mitochondrial p53 program participates et al., 2005). However, PUMA induction occurs within a in the overall apoptotic p53 response physiologically: sharp time window between 2 and 3 h (with decline normal mice were subjected to a single dose thereafter), leading to another boost in caspase-3 of 5 or 10 Gy whole body g-irradiation or a single activation. The p53 target Noxa appears transiently at clinical dose of intravenous etoposide. Mitochondrial the 4 h and Bax only very late at the 8 h time point p53 accumulation occurred in radiosensitive organs (which coincides with a 100% TUNEL positivity), like thymus, spleen, testis and brain, but not in while other p53 targets such as Bid, Killer/DR5, and radioresistant organs like liver and kidney that undergo p53DinP1 remain uninduced even after 20 h. Similar a strictly nuclear accumulation leading to cell cycle biphasic kinetics is seen in wtp53 expressing cancer cell arrest. lines (ML1 and RKO) after genotoxic stress. Together, Careful kinetic analysis of the mitochondrial versus these results provide evidence that in sensitive the transcriptional p53 pathway in thymus show organs and cells the mitochondrial p53 program occurs that mitochondrial events are ahead of transcriptional soon after a death stimulus in vivo, triggering a rapid events, confirming that the mitochondrial p53 pathway first wave of transcription-independent apoptosis which participates in the overall p53 stress response of may precede a second slower wave that is transcription- radiosensitive organs. Mitochondrial p53 translocation dependent.

Oncogene p53 and Nur77/TR3 UM Moll et al 4731 Evidence for the in vivo significance of this program is of p53À/À mice show a significantly lower degree of also provided by a number of pathophysiologic studies these changes. Conversely – and as already seen above in in animals and primary cells. In the context of renal tubular cells – pretreating synaptosomes with mycocardial infarction, the correlation between a pifithrin-a preserves their mitochondrial membrane genetically determined variability in cell death suscep- potential and partially blocks ROS production. Also, tibility and p53 codon 72 variants with their differential in cultured primary hippocampal neurons p53 becomes mitochondrial localization was already described above detectable in presynaptic terminals after etoposide (Bonafe et al., 2004). Another important setting in exposure. This coincides with synaptic degeneration which apoptosis is largely mediated by the mitochon- and again can be prevented by pifithrin-a pretreatment drial p53 program is ischemia-reperfusion injury of the (Gilman et al., 2003; Culmsee and Mattson, 2005). kidney, leading to acute renal failure (Kelly et al., 2003), Synaptic degeneration is thought to be a critical initial a common clinical problem. In ischemia, the outer event for acute and chronic neurodegenerative diseases medulla, composed of two types of tubules, exhibits including stroke, Alzheimer’s, Parkinson’s and Hun- p53-independent necrosis of the proximal straight tington’s disease. tubules and p53-dependent apoptosis of the thick Together, these studies support a role for the direct ascending tubules. Both components independently mitochondrial p53 pathway in vivo. But further work contribute to renal failure. Acute renal failure is along these lines, ideally in relevant animal models is experimentally modeled by transient bilateral renal needed to fully define the extent of this program in artery occlusion in rats. Importantly, ischemia-induced physiologic and pathophysiologic responses to geno- stabilization of p53 protein in the outer medulla is to a toxic, hypoxic and oxidative insults. large extent mitochondrial and this mitochondrial p53 translocation correlates with apoptosis of tubular cells in vivo (Kelly et al., 2003; Dagher, 2004). Likewise, in Possible implications for cancer therapy primary proximal tubular epithelial cells after chemical anoxia with antimycin A, mitochondrial p53 transloca- Abrogation of p53-mediated apoptosis is a fundamental tion is prominent (and triggered by GTP depletion as defect enabling the formation of human tumors. In intracellular mediator of death stimulus) (Kelly et al., addition, mutant p53 in tumors promotes chemoresis- 2003). Notably, pifithrin-a, a selective and potent p53 tance, thus confounding the success of cancer therapy. inhibitor (Komarov et al., 1999), prevents tubular Hence, its functional restoration is a premier therapeutic apoptosis in the outer medulla and protects renal goal in clinical oncology. function in the animal. The protective effects of As discussed above, p53 forcibly targeted to mito- pifithrin-a involved both downregulation of the tran- chondria is devoid of transcriptional activity but is scriptional activation of Bax and a direct inhibition of sufficient to induce apoptosis in a transcription-inde- p53 translocation to mitochondria (Kelly et al., 2003). pendent fashion in p53-null human cancer cell lines. Likewise, in cultured tubular cells after antimycin A Thus, mitochondrial targeting of p53 at least partly anoxia, pifithrin-a also blocks mitochondrial p53 phenocopies nuclear p53 in launching effective apop- translocation and apoptosis (Kelly et al., 2003). tosis and long-term growth inhibition. In an important In primary cultures of rat embryonal astrocytes and step towards therapeutic exploitation, we showed that C6 glioma cells subject to oxidative stress, similar events the direct mitochondrial p53 program exerts a signifi- occur. When treated for 1 h followed by 24 h recovery, cant tumor suppressor activity in Em-Myc transgenic low levels of p53 are restricted to the nucleus until the mice, a well-established transplantable model of human end of the short stress period. During the following Burkitt’s lymphoma and p53-dependent lymphomagen- 6–24 h recovery period, p53 levels continue to rise, with esis (Talos et al., 2005) (Figure 2). In Em-Myc its subcellular distribution becoming largely mitochon- lymphomas of p53-deficient or ARF-deficient geno- drial. Concomitant with mitochondrial p53 targeting, types, exclusive delivery of p53 to the outer mitochon- mitochondria undergo profound structural alterations drial membrane (by using p53 proteins fused to the from healthy filamentous mitochondria to fused, coar- transmembrane domain of BclXL or Bcl2) confers sened ‘megamitochondria’ starting at 3 h recovery. This efficient apoptosis and markedly inhibits cancer growth morphologic hallmark of early apoptosis coincides with in vivo. When present as the sole source of functional increasing cytochrome c release and subsequent nuclear p53, these p53 fusions, which lack residual transcrip- fragmentation by 12–24 h (Bonini et al., 2004). tional activity, possess an apoptotic potency that is Another site where the mitochondrial p53 program nearly as efficient as conventional nuclear p53 in killing appears to play a role in vivo are the synaptic terminals primary lymphoma cells in the animal (Table 1). of neurons in the brain cortex during DNA damage- Conversely, normal cells from thymus, spleen and bone induced neuronal death. In a model using topical marrow showed poor infectivity with these retroviruses etoposide on anesthetized mouse brain, isolated neuro- (Talos et al., 2005). Together, these studies significantly nal synaptosomes exhibit higher levels of total p53 and take forward the notion that the direct mitochondrial phosphoSer15-p53 than the distal cell bodies. This p53 program, which executes the shortest known synaptic p53 facilitates the collapse of the mitochondrial circuitry of p53 death signaling, might also be exploited membrane potential and ROS production upon addi- for inducing apoptosis in mutant p53 tumor cells in vivo. tional oxidative and excitotoxic insults. Synaptosomes Such studies are currently underway.

Oncogene p53 and Nur77/TR3 UM Moll et al 4732 mouse B-lymphoma Eµ myc Eµ myc

p53 + / – p53 –/ – Retroviral MSCV-GFP ARF + / – ARF – / – Nuclear p53 mito p53CTM mito p53CTB empty virus in vivo apoptosis of infected tumor cells

non-responder

infect primary lymphoma isolates responder syngeneic normal mouse

Figure 2 Experimental cancer model to assess the direct mitochondrial p53 program in vivo. Exclusive delivery of p53 to the outer mitochondrial membrane of primary lymphoma cells of p53-deficient or ARF-deficient genotypes results in efficient induction of apoptosis and impinged proliferation in the Em-Myc lymphoma model (see also Table 1). In these experiments, GFP-tagged, retrovirally transduced primary tumor cells are mixed with untagged parental cells at fixed ratios and injected into the tail vein of normal syngeneic mice. After 4 weeks, tumor regrowth in recipients is assessed by FACS analysis. This allows for quantitative assessment of tumor cell survival in an in vivo competition format.

Table 1 In vivo cell death of (a) p53-null B-cells (b) ARF-null B-cells expressing mitochondrially targeted p53 Retroviruses Ratio of injected B-lym- Isolate ID injected/(no. of Number of mice Average tumor Percent Tumor phoma cells, transduced/ independent experiments) with tumors weight (g) cells GFP-positi- parental vea at 28–30 days post-injection

(a) In vivo cell death of p53-null B-cells expressing mitochondrially targeted p53 No preselection, no sorting 2, 3, 4 and 5 (9) Vector – GFP 40/60 18 1.070.2 30.8718% Nuclp53 – GFP 40/60 12 0.770.3 0%b p53CTM – GFP 40/60 13 1.070.4 0.471.0%c,d p53CTB – GFP 40/60 4 0.470.1 0%b

Blasticidine preselection 2, 4 and 5 (3) Vector – GFP 75/25 3 0.770.2 44.373% Nuclp53 – GFP 75/25 1 0.570.2 12%e CTMp53 – GFP 75/25 11 0.670.4 1.170.1%b

FACS sorting 5 (3) Vector – GFP 90/10 14 1.170.3 70710% CTMp53 – GFP 90/10 7 1.370.3 14714%b

(b) In vivo cell death of ARF-null B-cells expressing mitochondrially targeted p53 No pre-selection, no sorting 1 and 2 (2) Vector – GFP 40/60 14 0.970.4 18713% Nuclp53 – GFP 40/60 13 0.3370.2 0.6%70.6%c CTMp53 – GFP 40/60 12 0.2370.1 0.270.4%c CTBp53 – GFP 40/60 11 0.3770.2 0.470.7%c

a7s.d. bPo0.0005. cPo0.0001 compared to vector group. dIf the 13 animal series is expanded by one additional aberrant tumor the percent GFP- positive tumor cells at 28 days is *2.872.4%. This tumor exhibited 34% GFP positivity, genotyped as wild type p53 by sequence but expressed an abnormal p53 protein. Western analysis exhibited strong ubiquitin-like bands (higher molecular weight ladder) recognized by p53 antibody. ep53 expression was lost in this tumor, as judged by Western analysis.

Oncogene p53 and Nur77/TR3 UM Moll et al 4733 In p53-targeted therapeutics research, up to now (processivity) of pol gamma in vitro. Loss of p53 results efforts focused on restoring transcription-mediated p53- in a significant increase in mtDNA vulnerability to dependent apoptosis. One venue tries to identify small genotoxic damage, leading to increased frequency of molecules such as CP-31398 and PRIMA-1 that mtDNA mutations in cells. This effect can be rescued by structurally rescue tumor-derived p53 mutant proteins p53 expression (Achanta et al., 2005). Further support (Foster et al., 1999; Bykov et al., 2002). Another venue for a potential role of p53 in modulating mitochondrial focuses on supplying conventional wtp53 via gene DNA repair exists, specifically mitochondrial base therapy (Moll, 2000). For example, adenovirus-based excision repair (BER). In an in vitro uracil incorporation conventional p53 gene therapy was recently approved by assay that measures flux through the entire BER the Chinese Food and Drug Administration for patients pathway, the mitochondrial pol gamma gap-filling with head and neck squamous cell carcinoma and is also activity was mildly but reproducibly decreased in p53 in late-stage clinical trials in China and in the West for knockout extracts, which could be complemented to other malignancies (Zeimet and Marth, 2003; Peng, wild type levels by recombinant p53 (de Souza-Pinto 2005). Both the wtp53 and small molecule approach et al., 2004). Moreover, stress-induced mitochondrial have in common that they rely on the preserved ability p53 engages in endogenous complexes with mitochon- of tumor cells to respond with transcriptional activation drial transcription factor A (mtTFA), the most abun- of p53 target genes. However, many human cancers dant protein in the mtDNA nucleoids and required for have lost this prerequisite due to global epigenetic transcription and maintenance of mtDNA. It is tempt- deregulation of their genome, leading to broadly ing to speculate that this p53 complex influences the aberrant gene silencing patterns (Herman and Baylin, repair activity of mtTFA, particularly in the context of 2003). Indeed, conventional p53 gene therapy did fail in cisplatin and oxidized mtDNA damage (Yoshida et al., ovarian cancer (Zeimet and Marth, 2003). Fortunately, 2003). However, since mtTFA knockout mice have an apoptosis relies upon a preassembled death machinery increased apoptosis phenotype, this interaction might of protein and DNA degrading enzymes that do not also somehow play a role in regulating apoptosis. Along require transcription of new genes. Our study in the these lines, mitochondrial ribosomal protein L41 Em-Myc lymphoma system shows that at least in (MRPL41) is reported to enhance p53 translocation to principle, transcriptional p53 restoration is dispensable mitochondria and stabilize p53, thereby contributing to for p53-based tumor killing in vivo. As mitochondrial apoptosis (Yoo et al., 2005). Intramitochondrial p53 p53 does not require tetramerization for its killing action might possibly also contribute to apoptosis by making it (Marchenko et al., 2000), another potential advantage harder for the cell to counteract damaging reactive might be its escape from dominant negative interference oxygen species (ROS) generated by stressed mitochon- by endogenous mutant p53 proteins and/or DNp63/p73 dria at an increased rate. This idea is supported by the isoforms that are commonly overexpressed in tumors. observation that in TPA-treated epidermal cells endo- Taken together, there is ample rationale for further genous wtp53 translocates to mitochondria within studies that investigate its possible therapeutic value, 10 min (preceding nuclear accumulation) where it perhaps as a synergistic modality with existing p53- reportedly sequesters the major oxidative defense based strategies or conventional anticancer therapy. enzyme manganese superoxide dismutase (MnSOD), located in the mitochondrial matrix. This might lead to mitochondrial dysfunction and apoptosis (Zhao et al., New avenues 2005). The precise mechanism and significance of many of these interactions, however, remains to be defined. We and others showed earlier that a small subfraction of stress-induced endogenous mitochondrial p53 localizes within the mitochondria, rather than on the surface, and Nur77 is found in complex with the major mitochondrial import proteins mt hsp70 and mt hsp60 (Marchenko Nur77 (also known as TR3, NGFI-B, TIS1 and NAK-1) et al., 2000; Dumont et al., 2003). This suggests an (Lim et al., 1987; Hazel et al., 1988; Milbrandt, 1988; additional intramitochondrial function of p53 distinct Chang et al., 1989) is a transcriptional factor belonging from the Bcl2 family-based OMM permeabilization. to the superfamily of nuclear receptors. All members of Progress in this direction came from evidence for a the superfamily share a highly conserved structural possible role of p53 in maintaining the stability of the organization with an amino-terminal region encoding mitochondrial genome (mtDNA) via the ability to activation function 1 (AF-1), followed by the DNA- translocate to mitochondria and interact with mtDNA binding domain (DBD), and the ligand-binding domain polymerase gamma (pol gamma; specific for mitochon- (LBD). The C-terminal region also encodes activation dria) (Achanta et al., 2005). This was seen in response to function 2 (AF-2) (Figure 3) (Kastner et al., 1995; mtDNA damage by ethidium bromide. In contrast to Mangelsdorf and Evans, 1995). Recent studies have nuclear DNA, mitochondrial DNA is not packaged in demonstrated that the AF-1 plays a major role in chromatin and nucleosomal particles, but associates mediating Nur77 transactivation, while AF-2 is dis- with the mitochondrial inner membrane on the matrix pensable (Wansa et al., 2002, 2003). Nur77 and the side, forming mtDNA-protein nucleoids (Bogenhagen closely related proteins Nurr1 (also known as RNR-1, et al., 2003). p53 enhances the DNA replication function TINUR, and HZF-3) and NOR-1 (MINOR, CHN, and

Oncogene p53 and Nur77/TR3 UM Moll et al 4734 AF-1 DBD LBD AF-2 it is much more active than the native receptor in Nur77 A/B CD E/F transactivating target promoters, suggesting that tran- scriptional activation of Nor-1 is associated with Figure 3 Schematic representation of Nur77. AF-1, activation oncogenicity. function 1; DBD, DNA-binding domain; LBD, ligand-binding domain; AF-2, activation function-2. Nur77 confers growth advantage on cancer cells by suppressing the negative growth regulatory effect of retinoids. In lung cancer cells, overexpression of Nur77 correlates with their resistance to retinoic acid treatment TEC) form a distinct subfamily within the nuclear (Wu et al., 1997b; Chen et al., 2002), due to Nur77’s receptor superfamily (Ferri and Kroemer, 2001; Zhang, inhibition of the expression of retinoic acid receptor b 2002; Hsu et al., 2004; Martinez-Gonzalez and Badi- (RARb), a growth suppressor of cancer cells (Zhang, mon, 2005), and they are classified as orphan receptors 2002). Nur77 inhibits RARb expression by interacting because no physiological ligands for these receptors with retinoid X receptor (RXR) (Wu et al., 1997a) or have been identified (Kastner et al., 1995; Mangelsdorf orphan receptor COUP-TF (Wu et al., 1997b). Nur77 and Evans, 1995). Structural analysis of the LBDs of induced under hypoxic conditions stabilizes and trans- Nurr1 and the Drosophila ortholog of Nor-1, DHR38, activates hypoxia-inducible factor-1a (Yoo et al., 2004), reveals no classical ligand-binding pockets in these a protein implicated in tumor aggressiveness, vascular- receptors (Baker et al., 2003; Wang et al., 2003), ity, treatment failure, and mortality. Consistently, suggesting that the biological functions of the Nur77 VEGF has been shown to induce the expression of the subfamily are regulated by ligand-independent mechan- Nur77 family genes in endothelial cells (Liu et al., 2003). isms. Nur77 is unique among nuclear receptors in that it Interestingly, Nur77 family members are found as is also an immediate-early gene induced by a diversity of NF-kB-responsive genes in macrophage (Pei et al., extracellular stimulation, ranging from survival, differ- 2005a), controlling expression of cyclin D2 (Pei et al., entiation and apoptosis. They work via distinct modes 2005b). Overexpression of Nur77 also blocks cell death of action, and account for diverse biological activities of induced by ceramide in A20 B cell line (Bras et al., 2000) Nur77. and protect fibroblasts from undergoing tumor necrosis factor-induced cell death (Suzuki et al., 2003). Induction of Nur77 by hepatitis B virus X protein may play a role Nur77 as a survival factor in viral carcinogenesis (Lee et al., 2001). The survival function of Nur77 family members in vivo is well Nur77 was originally identified as a growth factor- established in the central nervous system. Knockout of inducible gene (Lim et al., 1987; Hazel et al., 1988; Nurr1 in mice results in impaired development and Milbrandt, 1988; Chang et al., 1989). It is often maintenance of midbrain dopaminergic neurons (Zet- overexpressed in a variety of cancer cells, including terstrom et al., 1997), while overexpression of Nurr1 lung, prostate, breast and colon cancer (Uemura and increases the resistance of neural stem cells to neuro- Chang, 1998; Wu et al., 2002a, b; Jeong et al., 2003; toxins (6-hydroxy-dopamine and 1-methyl-4-phenyl- Holla et al., 2005; Maddika et al., 2005). A number of 1,2,3,6-tetrahydro-pyridine) (Le et al., 1999). growth factors and mitogens can potently and rapidly induce the expression of Nur77 in a variety of cancer cells. In prostate tissue, Nur77 is highly expressed in Nur77 as a death factor cancer as compared to adjacent normal or benign tissues (Uemura and Chang, 1998). These observations suggest The role of Nur77 in apoptosis was first recognized in that Nur77 acts to mediate the growth of cancer cells. By 1994, showing that Nur77 is inducible by T-cell receptor now, compelling evidence has established the survival signals and that Nur77 expression is high during T-cell and oncogenic activities of Nur77 in cancer cells. receptor-mediated apoptosis (Liu et al., 1994; Woronicz Ectopic expression of Nur77 stimulates cell cycle et al., 1994). Inhibition of Nur77 function by a progression and proliferation in lung cancer cells dominant-negative Nur77 blocks T-cell receptor- through its transcriptional activity (Kolluri et al., mediated apoptosis in thymocytes and their negative 2003). A recent expression profiling reveals that levels selection, whereas thymocytes from Nur77-expressing of Nur77 subfamily members are significantly down- transgenic mice show massive apoptosis. However, regulated in the nontransformed HeLaHF revertants, as Nur77 knockout mice do not exhibit any defect in compared with parental HeLa cells (Ke et al., 2004). thymocyte apoptosis (Cheng et al., 1997b), suggesting a Suppression of Nur77 expression inhibits the transfor- possible functional redundancy among Nur77 family mation phenotype of several cancer lines of different members. origins by promoting their apoptosis (Ke et al., 2004). The apoptotic effect of Nur77 in cancer cells The oncogenic potential of the Nur77 subfamily was discovered during investigations of the apoptotic members is also suggested by the finding that Nor-1 effect of the retinoid-related molecule AHPN (also (TEC) is involved in the t(9, 22) chromosomal trans- called CD437; 6-(3-(1-adamantyl)-4-hydroxyphenyl)-2- location in extraskeletal myxoid chondrosarcoma naphtha-lenecarboxylic acid) (Li et al., 1998). AHPN/ (Labelle et al., 1995; Labelle et al., 1999). The resulting CD437, currently being evaluated in clinical trials as fusion protein, EWS/TEC, is potentially oncogenic, and anticancer agent, causes growth arrest and apoptosis in

Oncogene p53 and Nur77/TR3 UM Moll et al 4735 epithelial cancer models in vitro and in vivo (Costantini providing an opportunity to regulate Nur77 activity et al., 2000; Fontana and Rishi, 2002; Zhang, 2002; using small molecules. Nur77/RXRa heterodimers bind Lotan, 2003; Pfahl and Piedrafita, 2003; Holmes et al., to the RA response element (RARE) in the RARb 2004). Although highly complex in its pro-apoptotic promoter, thereby promoting RARb expression by actions, in CA-OVOC3 ovarian cancer cells AHPN/ RXR ligand (Wu et al., 1997a; Chen et al., 2002). CD437 kills through a signaling pathway independent of Nur77 can also heterodimerize with orphan receptor retinoid receptor activation. Instead, it induces Nur77 COUP-TF (Wu et al., 1997b) to regulate the expression transcription by signaling through MEK, p38MAP of RARb.AsRARb plays a key role in growth, kinase and MEF2 (a transcription factor inducing survival, differentiation and apoptosis of various cell Nur77) (Holmes et al., 2003b). Inhibition of Nur77 types, regulation of RARb expression may account for expression by antisense RNA prevents the apoptotic diverse biological activities of Nur77 family members. effect of AHPN/CD437. Rapid induction of Nur77 also Nur77-induced apoptosis is rapid and robotic. Like in occurs in cancer cells after stimulation of apoptosis by a the case of p53, Nur77 is thought to exert its pro- variety of apoptosis-inducing agents, including calcium apoptotic action, at least in part, through its nuclear ionophores, etoposide (VP-16) (Li et al., 2000; Liu, transactivation of gene expression. However, attempts 2002), and phorbol ester (Li et al., 2000; Liu, 2002; Cao to identify downstream transcriptional targets of Nur77 et al., 2004a; Lin et al., 2004b), synthetic chenodeoxy- in this process have had only limited success. Trans- cholic acid derivatives (Jeong et al., 2003), Di-n-butyltin activation of RARb expression by the Nur77/RXRa dichloride (Gennari et al., 2002), histone deacetylase heterodimer is critical for apoptosis induction (Wu inhibitors (Chinnaiyan et al., 2006), cadmium (Shin et al., 1997a). A significant amount of work has been et al., 2004), 1,1-bis(30-indolyl)-1-(p-substituted phenyl)- conducted in T cells. FasL was initially considered to be methanes (Chintharlapalli et al., 2005), and induced the potential downstream targets of Nur77, based on the Nur77 contributes to their apoptotic effects. observation that their expression is elevated in Nur77 The apoptotic effect of Nur77 appears to be clinically transgenic mice (Weih et al., 1996). However, subse- relevant as a recent study shows that the expression of quent studies showed that overexpression of Nur77 does the Nur77 subfamily member Nor-1 is positively not lead to any induction of FasL expression (Cheng correlated with survival in diffuse large B-cell lymphoma et al., 1997b), and Nur77-mediated apoptosis is intact in patients treated with the drugs cyclophosphamide, mice carrying the FasL mutation (Cheng et al., 1997b). adriamycin and vincristine (Shipp et al., 2002). In Microarray analysis using RNA isolated from Nur77 addition, gene expression profiles of tumor samples transgenic and wild-type fetal thymi demonstrates that obtained from various cancer patients reveal that two novel genes, NDG1 and NDG2, are induced by downregulation of Nur77 is associated with metastasis Nur77 expression (Rajpal et al., 2003). NDG2 targets of primary solid tumors, including lung cancer, breast mitochondria, but its role in apoptosis remains to be cancer and prostate cancer (Ramaswamy et al., 2003). defined. In contrast, NDG1 exhibits apoptotic activity independent of the death receptor pathway, involving activation of caspase-8 and -3. However, it remains to be Transcriptional regulation by Nur77 seen whether both NDG1 and NDG2 are direct transcriptional targets of Nur77. Further, expression Although Nur77 expression is induced by a diversity of of NDG1 in T cells does not lead to massive apoptosis, signals, our knowledge of its molecular mechanism of in contrast to Nur77 (Rajpal et al., 2003), suggesting action is very limited. Like most nuclear receptors, that other Nur77-mediated apoptotic events are in- Nur77 binds to specific DNA sequences as homodimers volved in Nur77-induced T-cell apoptosis. In this or heterodimers (Ferri and Kroemer, 2001; Zhang, 2002; regard, in thymic lymphoma ionomycin-induced Hsu et al., 2004; Martinez-Gonzalez and Badimon, Nur77 DNA-binding does not cause apoptosis. In 2005). In addition, Nur77 can bind DNA as monomers. contrast, prevention of ionomycin-induced Nur77 Despite the identification of Nur77-binding sequences DNA binding by either calcineurin inhibitor FK506 or and the accumulating evidence that Nur77 functions to serine/threonine protein kinase inhibitor HA1004 results mediate the biological responses of a variety of in extensive apoptosis (Kochel et al., 2005). It would, extracellular stimuli, very few Nur77 direct target genes therefore, be important to test whether DNA binding of have been identified. Regulation of tyrosine hydroxylase Nur77 is required for apoptosis in normal thymocytes in expression by Nurr1, presumably through its monomer a physiological setting. binding, may be important for Nur77 family member action in the central nervous system (Sakurada et al., 1999). The promoter of pituitary pro-opiomelanocortin Mitochondrial targeting as a mechanism for initiating (POMC) gene contains a Nur-response element apoptosis (NurRE) that binds Nur77 homodimers, supporting transcriptional modulation by Nur77 of the hypothala- While there is a correlation between transcriptional mo–pituitary–adrenal axis (Philips et al., 1997). Nur77 function of Nur77 and apoptosis in T cells, evidence for can also bind DNA as a heterodimer with retinoid X the role of Nur77 transactivation in apoptosis of cancer receptor (RXR) (Forman et al., 1995; Perlmann and cells is lacking. Mu and Chang (2003) showed that TR3/ Jansson, 1995), which can be activated by RXR ligands, Nur77 induces E2F1 expression by binding to the TR3

Oncogene p53 and Nur77/TR3 UM Moll et al 4736 response element (TR3RE) in the E2F1 promoter These observations indicate that Nur77 can employ region. Although antisense E2F1 partially rescues the different mechanisms to initiate apoptosis depending on TR3/Nur77-mediated cell apoptosis, E2F1 induction by cell type and death stimuli. Nur77 may mainly contribute to its regulation of cycle Accumulating evidence demonstrates that transloca- progression (Park et al., 2005). tion of Nur77 from the nucleus to mitochondria Several lines of evidence suggest mechanisms other than represents an important mechanism by which Nur77 transcriptional regulation for Nur77-induced apoptosis. initiates apoptosis (Figure 4). In response to several Unlike many other members of the nuclear receptors that apoptotic stimuli, Nur77 migrates from the nucleus to are localized in the nucleus and bind to DNA in the the cytoplasm, where it targets mitochondria (Li et al., absence of their ligand, Nur77 resides predominantly in 2000). Mitochondrial localization of Nur77 is associated the cytoplasm in many cell types (Fahrner et al., 1990; with release of cytochrome c from mitochondria and Hazel et al., 1991; Davis et al., 1993; Woronicz et al., apoptosis. Interestingly, DNA-binding domain-deleted 1995). As biological activities of Nur77 are highly Nur77 constitutively resides at mitochondria and is regulated through phosphorylation, heavily phosphory- accompanied by extensive cytochrome c release (Li lated Nur77 is located in the cytoplasm, while hypophos- et al., 2000; Holmes et al., 2003b; Lin, 2004). The N- phorylated Nur77 is primarily localized in the nucleus terminal AF-1 region is also dispensable for Nur77 (Fahrner et al., 1990). The fact that the apoptosis-inducing mitochondrial targeting and apoptosis (Lin, 2004). agents PMA and inomycin phosphorylate Nur77 (Fahrner Thus, DNA-binding and transactivation of Nur77 are et al., 1990) suggests that phosphorylated Nur77 forms not required for apoptosis induction by Nur77. In may be localized at mitochondria. PMA and ionomycin contrast, the cytoplasmic localization of Nur77 and its phosphorylate Nur77 at Ser354 in the A-box of its DNA- mitochondrial localization are essential for its apoptotic binding domain (Woronicz et al., 1995), an event known to effect. Nuclear retention of Nur77 or re-targeting of abolish its DNA-binding activity (Davis et al., 1993; Nur77 to the plasma membrane or the endoplasmic Hirata et al., 1993), suggesting that DNA binding is not reticulum membrane abrogates its ability to induce required for Nur77-induced apoptosis. Similarly, although cytochrome c release (Li et al., 2000). Nur77 expression is required for apoptosis induction by The apoptosis-associated nuclear egress of Nur77 has several apoptotic stimuli, Nur77 transactivation is inhib- now been observed in cancer cells from lung, ovary, ited by these apoptotic stimuli (Li et al., 2000). colon, prostate, stomach and breast. In CA-OV-3 Unlike T cells, where overexpression of Nur77 leads ovarian cancer cells, Nur77 translocates from the to massive apoptosis (Cheng et al., 1997b; Liu et al., nucleus to mitochondria in response to AHPN/CD437 1994; Woronicz et al., 1994), cancer cells often exhibit treatment (Holmes et al., 2003a, b, 2004), similar to its enhanced survival and proliferation upon Nur77 over- effect in LNCaP prostate cancer (Li et al., 2000; Lin, expression. In H460 lung cancer cells, overexpression of 2004) and H460 lung cancer cells (Dawson et al., 2001; Nur77 leads to an enhanced cell proliferation and cell Kolluri et al., 2003; Cao et al., 2004a; Lin, 2004). cycle progression (Kolluri et al., 2003). Similarly, Insulin-like growth factor-binding protein-3 (IGFBP-3)- overexpression of Nur77 in HeLa cells and other cancer induced apoptosis in 22RV1 prostate cancer cells is cells promotes their survival (Ke et al., 2004). Many accompanied by Nur77 mitochondrial targeting (Lee growth factors trigger cell survival programs through et al., 2005), while Nur77 mitochondrial targeting is an induction of Nur77. Thus, Nur77 must engage discrete important mechanism for apoptosis of gastric cancer regulatory mechanisms to accomplish its apoptotic cells in response to diverse apoptotic stimuli, including effect in cancer cells. Additional complexity of the death AHPN/CD437 (Lin, 2004), TPA, VP-16 (Liu, 2002; Wu effect of Nur77 comes from recent results showing that et al., 2002a), and several synthetic chenodeoxycholic Nur77 regulates distinct death programs depending on acid derivatives (Jeong et al., 2003). Nur77 mitochon- cell type and death stimuli. Nur77-induced apoptosis drial localization is also involved in Sindbis virus- involves the mitochondrial pathway in many cancer cells induced apoptosis (Lee et al., 2002, 2004), which is (Ferri and Kroemer, 2001; Zhang, 2002). In contrast, prevented by Epstein–Barr virus nuclear antigen release of cytochrome c is not detected in apoptotic EBNA2 through its interaction with Nur77. Cancer- thymocytes from Nur77 transgenic mice (Rajpal et al., specific toxicity of apoptin, a small prolin-rich protein 2003). Moreover, Nur77 mediates an alternative, non- derived from the chicken anaemia virus, triggers apoptotic programmed cell death in response to activa- cytochrome c release in MCF-7 breast cancer cells by tion of neurokinin-1 receptor by Substance P via ERK2 inducing Nur77 migration to mitochondria (Maddika activation (Castro-Obregon et al., 2004). Another et al., 2005). Nur77 mitochondrial targeting also plays a variation of pro-apoptotic Nur77 action was reported role in neuronal cell death. Nur77 rapidly translocates to specifically for colon cancer cells. Here, as opposed to mitochondria in cerebellar granule neurons exposed to the prototypical prostate cancer cells, Nur77 trans- glutamate, which is blocked by EGF through its locates to the cytoplasm but not directly to mitochon- activation of ERK2 (Jacobs et al., 2004). We have also dria upon various apoptotic stimuli (fatty acid butyrate, observed that neurotoxin causes extensive mitochon- the non-steroidal anti-inflammatory sulindac or 5-FU). drial localization in dopaminergic neurons (Zhang et al., Instead, translocation-dependent indirect activation of unpublished results). Thus, Nur77 mitochondrial target- the mitochondrial pathway via Bax relocalization is ing may play a role in the development of neurodegen- proposed (Wilson et al., 2003). erative diseases.

Oncogene p53 and Nur77/TR3 UM Moll et al 4737 Survival Apoptotic RXR Stimuli stimuli ligands

? ERK2 ? JNK Akt Bcl-2

P P P RXR RXR RXR

Nur77 Nur77 Nur77 Nur77 P Nur77 Bcl-2 P P P P P P P P

COUP RXR Nur77 Nur77 Nur77 Nur77 Mitochondria RARE NuRE NBRE Cyt c

Nucleus Gene Expression

Cytoplasm Apoptosis Survival Figure 4 The Nur77-Bcl2 apoptotic and survival pathways. In response to growth stimuli, Nur77 acts as a transcription factor to positively regulate expression of genes involved in cell proliferation. Nur77 is also induced by certain apoptotic stimuli, and upon appropriate modification (probably by phosphorylation), migrates from the nucleus to the cytoplasm as Nur77/RXRa heterodimer, where it targets mitochondria by binding to Bcl2. Interaction with Bcl2 induces a Bcl2 conformational change, triggering cytochrome c release and apoptosis.

Conversion of Bcl2 from an anti-apoptotic to a appear to promote rather than prevent apoptosis in pro-apoptotic molecule by Nur77 Caenorhabditis elegans (Xue and Horvitz, 1997). Like- wise, Bcl2 homologs in Drosophila can manifest either Like p53, Nur77 lacks classical mitochondrial-targeting cytoprotective or cytodestructive phenotypes, depending sequences. Studies of its mechanism of mitochondrial on the cellular context (Igaki et al., 2000). The opposing targeting reveals that Bcl2, with its ability to interact biological activities may account for some clinical with Nur77, functions as a mitochondrial receptor of observations that overexpression of Bcl2 is associated Nur77 (Lin, 2004) (Figure 4). Nur77 interaction with with a favorable prognosis (Gasparini et al., 1995; Bcl2 represents the first example that a nuclear receptor Daidone et al., 1999; Catz and Johnson, 2003; Martin interacts with a Bcl2 family member, providing the et al., 2003). For breast cancer, high Bcl2 levels are coupling of Nur77 nuclear receptor signaling to the associated with markers of favorable prognosis, such as Bcl2-mediated apoptotic machinery. In this regard, Bcl2 the expression of hormone nuclear receptors, low tumor does not act as an anti-apoptotic protein in Nur77- grade and responsiveness to endocrine therapy (Daidone mediated apoptosis. Instead, the interaction between et al., 1999). A recent systematic review of the literature Nur77 and Bcl2 triggers cytochrome c release and shows that overexpression of Bcl2 is also a good apoptosis (Lin, 2004). The Nur77-Bcl2 interaction prognostic factor for survival in patients with non-small conformationally converts Bcl2 from an anti-apoptotic lung cancer (Martin et al., 2003). to a pro-apoptotic molecule, providing a new approach An important question is how Bcl2 is converted into a to switch the Bcl2 phenotype. It remains to be seen killer by Nur77 binding. All members of the Bcl2 family whether other Bcl2 family members interact with Nur77, possess at least one of the four conserved BH domains mediating its mitochondrial targeting and apoptosis. (Reed, 1998; Cory et al., 2003). Anti-apoptotic Bcl2 The interaction of Nur77 with Bcl2 and subsequent family members Bcl2 and BclXL have hydrophobic conversion of the Bcl2 phenotype provide an explana- pockets on their surfaces, which is essential for their tion for the opposing phenotypes of some Bcl2 family anti-apoptotic effect, while their BH3 domains are members. For example, overexpression of Bcl2 in some buried. In contrast, pro-apoptotic Bcl2 family members cell types promotes rather than prevents apoptosis have an exposed BH3 domain, which bind to the (Cheng et al., 1997a; Grandgirard et al., 1998; Uhlmann hydrophobic pockets of anti-apoptotic Bcl2 members et al., 1998; Subramanian and Chinnadurai, 2003). (made up by their BH1, BH2 and BH3 domains) to Conversely, overexpression of Bax is cytoprotective inhibit their survival effect. The unstructured loop rather than cytodestructive in some cell types under region of B50 amino acid length located between the some circumstances (Lewis et al., 1999; Fannjiang et al., BH4 (a1 helix) and BH3 (a2 helix) domains of Bcl2 plays 2003). Similarly, mutants of the Bcl2 homolog Ced-9 an important, albeit highly complex role in regulating

Oncogene p53 and Nur77/TR3 UM Moll et al 4738 Bcl2 functions. Deletion of the loop region of Bcl2 lung cancer cells (Cao et al., 2004a) and in MGC80-3 blocks paclitaxel-induced apoptosis (Srivastava et al., gastric cancer cells (Lin et al., 2004a, b). Conversely, 1999), whereas caspase-mediated cleavage within the RXR prevents EGF-induced nuclear accumulation of loop region of Bcl2 converts it into a pro-apoptotic Nur77 (Jacobs and Paulsen, 2005). molecule (Cheng et al., 1997a; Grandgirard et al., 1998). As RXR activities are regulated by its ligands, it is not Modification of the loop region of Bcl2 or BclXL by surprising that RXR modulates Nur77 subcellular phosphorylation or deamidation is associated with localization in a ligand-dependent manner. However, downregulation of their anti-apoptotic activity (Haldar controversy exists with respect to the precise export et al., 1995; Yamamoto et al., 1999; Deverman et al., mechanisms (CRM1 dependence) and regulatory effect 2002). Upon Nur77 binding to the loop of Bcl2, the of RXR ligands. Cao et al. (2004b) found that in the hydrophobic binding groove of Bcl2 undergoes an presence of RXR transactivation agonists (ligands) such extensive rearrangement, resulting in exposure of its as 9-cis-RA, Nur77/RXRa heterodimers form via BH3 domain (Lin, 2004). This conformational change is dimerization sites in their LBDs, resulting in chroma- responsible for the conversion of Bcl2 from a cytopro- tin-binding heterodimers with efficient transactivation tective to a cytodestructive molecule. It is likely that pro- of retinoic acid response genes. This concomitantly apoptotic Bcl2 acts as a BH3-only protein to regulate silences the NES, thus fortifying nuclear localization. In the apoptotic potential of other Bcl2 family members. accord, RXR transcriptional agonists suppress apopto- Activation of Bax or Bak requires their binding to BH3- sis by inhibiting mitochondrial targeting (Cao et al., only protein tBid (Wei et al., 2000, 2001). 2004b). 9-cis-RA also relieves the inhibitory effect of RXR on EGF-induced Nur77 nuclear accumulation (Jacobs and Paulsen, 2005) and could account for the Regulation of Nur77 mitochondrial targeting by RXR and inhibitory effect of RXR ligands on apoptosis. For its ligands example, 9-cis-RA potently inhibits the activation- induced apoptosis of T cells and thymocytes (Bissonn- A notable feature of nuclear receptors is that they ette et al., 1995; Yang et al., 1993, 1995), in which Nur77 undergo extensive homo- and heterodimerization via a plays a role. The inhibitory effect of 9-cis-RA was domain located within the LBD that typically greatly enhanced in T-cell hybridomas overexpressing RXRb, enhances their DNA binding and transactivation. RXRs but attenuated in cells overexpressing dominant-nega- are particularly potent and promiscuous heterodimer- tive RXRb (Yang et al., 1995). ization partners for many nuclear receptors including In the absence of ligand and (the simultaneous Nur77, facilitating their DNA binding and nuclear presence of certain apoptotic stimuli), however, Cao localization. This is partly due to the fact that they have et al. (2004b) found that Nur77/RXRa heterodimers are two dimerization domains, a strong one in the LBD and formed via dimerization sites in their DBDs. Here, NES a weak one in the DBD. is active and mediates mitochondrial targeting and Nur77 lacks a classical ligand-binding pocket. How- apoptosis. It is unknown how an apoptotic stimulus ever, RXR and its ligands are critical regulators of activates the RXRa NES, but such stimuli possibly Nur77 activities through heterodimerization. RXRa regulate the dimerization interface switch through serves as the switchable shuttler for Nur77 mitochon- modulation of receptor phosphorylation. This is con- drial targeting (Cao et al., 2004b; Lin et al., 2004a, b). In sistent with recent observations that Akt, a potent the nucleus, Nur77/RXR heterodimers bind and acti- antiapoptotic kinase, phosphorylates Nur77 at amino vate the DR-5 promoter response element in response to acids lat the C-terminal extension, which is important certain RXR ligands (Forman et al., 1995; Perlmann for its protein–protein interaction (Meinke and Sigler, and Jansson, 1995; Wu et al., 1997a). This might provide 1999; Cao et al., 2004a). a mechanism to modulate Nur77’s survival activity by In contrast, Lin et al found a positive 9-cis-RA effect. RXR ligands, such as synthetic retinoids and poly- In their hands, 9-cis-RA was required to induce the unsaturated fatty acids (de Urquiza et al., 2000; Fan RXR-mediated nucleo-cytoplasmic shuttling of Nur77 et al., 2003; Goldstein et al., 2003; Lengqvist et al., and its translocation to mitochondria to induce apop- 2004), in cancer and in neurons. tosis (Lin, 2004). Such an effect of 9-cis-RA appears to Nur77 nuclear export and mitochondrial targeting be contradictory to its known transcriptional activation requires RXR heterodimerization (Cao et al., 2004b; Lin of RXR. Given the unstable nature of 9-cis-RA, it is et al., 2004a, b). Nur77 and RXR engage in an likely that a 9-cis-RA metabolite may act to induce interesting job sharing. Nur77, despite possessing a RXR nuclear export. Such a RXRa ligand may induce classical nuclear export signal (NES) used in differentia- the nuclear export of Nur77/RXR heterodimers by tion signaling (Katagiri et al., 2000), is the passive modulating their heterodimerization. In this regard, the partner in transport but the acting partner in mediating NSAID R-etodolac binds RXRa and acts as a RXRa mitochondrial apoptosis. Nur77 export is mediated via antagonist to inhibit its transactivation, an event which CRM1 receptors by a strong NES located in the LBD of is associated with its tumor-selective induction of RXRa, which is conformationally regulated by ligand apoptosis in animals (Kolluri et al., 2005). Several binding. In response to apoptotic stimuli, Nur77 and dietary fatty acids, including DHA (de Urquiza et al., RXRa associate with mitochondria as a Nur77/RXRa 2000), are also known to bind to RXR. It remains to be heterodimer, as seen in LNCaP prostate cancer, H460 seen whether other NSAIDs and fatty acids exert their

Oncogene p53 and Nur77/TR3 UM Moll et al 4739 anti-cancer effects by modulating Nur77 mitochondrial Nur77, including Nur77 DNA binding, transactivation, targeting through their binding to RXRa. Nur77 subcellular localization, and apoptotic activities Nur77/RXR heterodimerization and the activation of (Chen and Blenis, 1990; Chen et al., 1991; Swanson its NES are also subjected to regulation by other binding et al., 1999; Masuyama et al., 2001; Pekarsky et al., proteins. Binding of RXR to IGFBP-3 allows a rapid 2001; Wingate et al., 2005). Ser351 is phosphorylated by translocation of Nur77/RXR heterodimers to mitochon- different kinases, including Akt and RSK. Akt mediates dria, initiating an apoptotic cascade (Lee et al., 2005), the anti-apoptotic effects of growth factors through whereas binding of Nur77 to EBNA2 prevents their PI3-K and is constitutively activated in many cancer cell nuclear export (Lee et al., 2004). lines, confering their survival. Phosphorylation of Nur77 by Akt negatively regulates its nuclear export (Han et al., 2006), resulting in delayed Nur77 nuclear Regulation of Nur77 mitochondrial targeting by export. Consistently, deletion of the DBD and Ser351 phosphorylation from Nur77 (called mutant Nur77/DDBD) constitu- tively targets to mitochondria and induces apoptosis (Li Recent studies showed that phosphorylation also reg- et al., 2000; Kolluri et al., 2003; Lin, 2004). Thus, ulates Nur77 subcellular trafficking. Nur77 mitochon- phosphorylation of Nur77 by Akt at Ser351 negatively drial targeting is induced by apoptotic retinoid-related regulates Nur77 nuclear export and Nur77-dependent compounds, TPA, a calcium ionophore, and chemother- apoptosis. apeutic agents, which are known to act through Ser351 of Nur77 is also phosphorylated by RSK. membrane-signaling pathways involving various kinases ERK2, an upstream activator of RSK, inhibits apopto- and phosphatases. Consistently, Nur77 is heavily phos- sis-induced Nur77 nuclear export (Jacobs et al., 2004), phorylated in various cell types (Li et al., 2000; Kolluri suggesting that phosphorylation of Ser351 by RSK may et al., 2003; Han et al., 2006). Nur77 is phosphorylated at be responsible for the inhibitory effect of ERK2 on its N-terminal A/B domain by c-Jun-N-terminal kinase Nur77 nuclear export. TPA-induces Nur77 nuclear (JNK) (Kolluri et al., 2003; Han et al., 2006), ERK2 export and apoptosis in LNCaP cells. In contrast, in (Jacobs et al., 2004, 2005; Slagsvold et al., 2002), and the H460 lung cancer cells TPA is unable to induce Nur77 MAP kinase pathway (Katagiri et al., 2000). It is also nuclear export (Zhang et al., unpublished), but strongly phosphorylated at Ser350 (Ser351 in human Nur77) by activates RSK (Han et al., 2006), indicating cell type Akt (Masuyama et al., 2001; Pekarsky et al., 2001) and differences in signaling pathways. p90-kDa ribosomal S6 kinase (RSK) (Chen and Blenis, 1990; Chen et al., 1991; Swanson et al., 1999; Wingate et al., 2005), a Serine/Threonine protein kinase and a Nur77 mitochondrial targeting and cancer therapeutics downstream effector of MAPK. The Nur77 N-terminal A/B region is the major As described above, Nur77, a member of the nuclear domain responsible for regulation by phosphorylation. receptor superfamily of transcriptional factors, is a key About 30% of amino acids in this region are potential regulator of cell proliferation and apoptosis, two sites for phosphorylation. Deletion of the N-terminal important processes that control tumor development. region abrogates the effect of apoptotic stimuli on While its transactivation promotes cancer cell growth Nur77 mitochondrial targeting (Li et al., 2000), indicat- and may be oncogenic, its translocation into mitochon- ing their important regulatory role. NGF-induced dria mediates apoptotic effects of certain chemother- NGFI-B nuclear export requires the phosphorylation apeutic agents. Translocation of Nur77 to mitochondria of NGFI-B at Ser105 by the MAP kinase pathway where it converts Bcl2 from a cytoprotective to a killer (Katagiri et al., 2000). In contrast, glutamate-induced represents a new paradigm in cancer cell apoptosis Nur77 mitochondrial targeting is inhibited by EGF (Figure 4). Although many questions regarding the through its activation of ERK2 that phosphorylates mechanism regulating Nur77 migration remain, this Thr142 of NGFI-B (Jacobs et al., 2004). JNK positively unique property of Nur77 makes it an attractive regulates Nur77 nuclear export and apoptosis by molecular target for developing new therapeutics for phosphorylating Nur77 N-terminal amino acid residues cancer prevention and treatment. It is increasingly (Han et al., 2006). Inhibition of JNK activation by a accepted that the efficacy of chemotherapeutic drugs JNK inhibitor abolished AHPN/CD437 analog 3-Cl- depends largely on their ability to induce apoptosis. AHPC-induced Nur77 nuclear export. Consistently, Understanding the mechanism governing Nur77-induced other apoptotic stimuli including TPA, VP16, and apoptosis of cancer cells may provide a molecular basis cisplatin, which induce Nur77 nuclear export, are for the development of new and more effective therapies known to activate JNK. However, JNK activation for cancer and other diseases. In this regard, Nur77 and alone is insufficient to induce Nur77 nuclear export. Bcl2 are of particular interest since they are often Rather, it may function to inhibit Nur77 nuclear overexpressed in cancer cells, providing an excellent accumulation. Other signaling pathways or proteins opportunity to preferentially induce apoptosis of cancer are therefore required to complete the process of Nur77 cells. Targeting the Nur77-Bcl2 pathway should be nuclear export. highly effective in inducing apoptosis, because it engages Recent studies indicate that phosphorylation of a simultaneous suppression of an anti-apoptotic effect Nur77 at Ser351 also influences diverse activities of and activation of an pro-apoptotic effect. Agents that

Oncogene p53 and Nur77/TR3 UM Moll et al 4740 Apoptotic stimuli p53 Bax B DBD a B x B ax a PUMA x p53 Noxa p53 Bax etc…. DBD nur77 ? RAR β BclXL/2 ? NDG1 DBD DBD Nur77 Nur77 DBD Nur77 Loop Nucleus Bcl2

Bax B B Ba ax a x Cytoplasm x

Figure 5 Mechanistic similarities of the transcription-independent pro-death programs of p53 and Nur77 at mitochondria. For details see text. In response to apoptotic stimuli, p53 acts as transcription factor in the nucleus to induce the expression of proapoptotic genes. In contrast, Nur77 is a Janus-type protein with pro-and anti-apoptotic actions. When acting as a transcription factor, Nur77 mainly promotes the expression of survival molecules. In the cytoplasm, both p53 and Nur77 proteins mediate apoptosis by targeting directly to mitochondria and binding to anti-apoptotic Bcl2 (Nur77) and Bcl2/ BclXL (p53). Nur 77 binds to the loop region of Bcl2, resulting in a rearrangement of the hydrophobic binding groove of Bcl2 and exposure of its BH3 domain. p53 uses its DNA-binding domain to bind to the ‘underside’ of BclXL and Bcl2, involving the BH4 domain and the two loops between the a3 and a4 and the a5 and a6 helices in BclXL. The end result in both cases is the homo-oligomerization of Bax and Bak and mitochondrial membrane permeabilization.

induce translocation of Nur77 from the nucleus to the expression of survival molecules. In the cytoplasm, mitochondria may inhibit cancer cell growth conferred both p53 and Nur77 proteins mediate apoptosis by by the nuclear action of Nur77 and at the same time may targeting directly to mitochondria and binding to anti- induce cancer cell death. Similarly, strategies to increase apoptotic Bcl2 (Nur77) and Bcl2/ BclXL (p53). p53 uses Nur77 expression could be used in combination with its DNA-binding domain to bind to the ‘underside’ of agents that induce its mitochondrial targeting to improve BclXL and Bcl2, thereby inhibiting their anti-apoptotic the efficacy of cancer therapy. Recent advances in action. In case of BclXL, this has been mapped to the understanding the mechanisms underlying Nur77 mito- BH4 domain and the two loops between the a3 and a4 chondrial targeting provides several opportunities for and the a5 and a6 helices. Nur 77 binds to the loop identifying such agents. Nur77 translocation to mito- region of Bcl2, resulting in a rearrangement of the chondria as Nur77/RXR heterodimers offers an oppor- hydrophobic binding groove of Bcl2 and exposure of its tunity to identify RXR ligands for regulating the Nur77- BH3 domain. This binding conformationally converts dependent apoptotic pathway. Bcl2 to a killer that triggers cytochrome c release and apoptosis. The end result in both cases is the homo- oligomerization of Bax and Bak and mitochondrial Mechanistic similarities of the transcription-independent membrane permeabilization (Figure 5). (A possible pro-death programs of p53 and Nur77 at mitochondria functional interaction between the p53 and Nur77 proteins at mitochondria is purely hypothetical but has In response to apoptotic stimuli, p53 acts as transcrip- not been addressed). Taken together, an accumulating tion factor in the nucleus to induce the expression of body of evidence provides ample rationale to further proapoptotic genes. In contrast, Nur77 is a Janus-type investigate how these mitochondrial p53 and Nur77 protein with pro-and anti-apoptotic actions. When pathways could become exploitable targets for new acting as a transcription factor, Nur77 mainly promotes cancer therapeutics.

References

Achanta G, Sasaki R, Feng L, Carew JS, Lu W, Pelicano H Arriola EL, Rodriguez-Lopez AM, Hickman JA, Chresta CM. et al. (2005). EMBO J 24:3482–3492. (1999). Oncogene 18:1457–1464. Arima Y, Nitta M, Kuninaka S, Zhang D, Fujiwara T, Taya Y Baker KD, Shewchuk LM, Kozlova T, Makishima M, Hassell et al. (2005). J Biol Chem 280:19166–19176. A, Wisely B et al. (2003). Cell 113:731–742.

Oncogene p53 and Nur77/TR3 UM Moll et al 4741 Bissonnette RP, Brunner T, Lazarchik SB, Yoo NJ, Boehm Fahrner TJ, Carroll SL, Milbrandt J. (1990). Mol Cell Biol 10: MF, Green DR et al. (1995). Mol Cell Biol 15:5576–5585. 6454–6459. Bogenhagen DF, Wang Y, Shen EL, Kobayashi R. (2003). Fan YY, Spencer TE, Wang N, Moyer MP, Chapkin RS. Mol Cell Proteomics 2:1205–1216. (2003). Carcinogenesis 24:1541–1548. Bonafe M, Salvioli S, Barbi C, Trapassi C, Tocco F, Storci G Fannjiang Y, Kim CH, Huganir RL, Zou S, Lindsten T, et al. (2004). Cell Death Differ 11:962–973. Thompson CB et al. (2003). Dev Cell 4:575–585. Bonini P, Cicconi S, Cardinale A, Vitale C, Serafino AL, Ciotti Ferri KF, Kroemer G. (2001). Bioessays 23:111–115. MT et al. (2004). J Neurosci Res 75:83–95. Fontana JA, Rishi AK. (2002). Leukemia 16:463–472. Bras A, Albar JP, Leonardo E, de Buitrago GG, Martinez AC. Forman BM, Umesono K, Chen J, Evans RM. (1995). Cell 81: (2000). Cell Death Differ 7:262–271. 541–550. Bykov VJ, Issaeva N, Shilov A, Hultcrantz M, Pugacheva E, Foster BA, Coffey HA, Morin MJ, Rastinejad F. (1999). Chumakov P et al. (2002). Nat Med 8:282–288. Science 286:2507–2510. Cao X, Liu W, Lin F, Li H, Kolluri SK, Lin B et al. (2004a). Gasparini G, Barbareschi M, Doglioni C, Palma PD, Mauri Mol Cell Biol 24:9705–9725. FA, Boracchi P et al. (1995). Clin Cancer Res 1:189–198. Cao X, Liu W, Lin F, Li H, Kolluri SK, Lin B et al. (2004b). Gennari A, Bleumink R, Viviani B, Galli CL, Marinovich M, Mol Cell Biol 24:9705–9725. Pieters R et al. (2002). Toxicol Appl Pharmacol 181:27–31. Castro-Obregon S, Rao RV, del Rio G, Chen SF, Poksay KS, Gilman CP, Chan SL, Guo Z, Zhu X, Greig N, Mattson MP. Rabizadeh S et al. (2004). J Biol Chem 279:17543–17553. (2003). Neuromolecular Med 3:159–172. Catz SD, Johnson JL. (2003). Apoptosis 8:29–37. Goldstein JT, Dobrzyn A, Clagett-Dame M, Pike JW, DeLuca Chang C, Kokontis J, Liao SS, Chang Y. (1989). J Steroid HF. (2003). Arch Biochem Biophys 420:185–193. Biochem 34:391–395. Gonzalez-Garcia M, Perez-Ballestero R, Ding L, Duan L, Chen GQ, Lin B, Dawson MI, Zhang XK. (2002). Int J Cancer Boise LH, Thompson CB et al. (1994). Development 120: 99:171–178. 3033–3042. Chen L, Willis SN, Wei A, Smith BJ, Fletcher JI, Hinds MG Grandgirard D, Studer E, Monney L, Belser T, Fellay I, et al. (2005). Mol Cell 17:393–403. Borner C et al. (1998). EMBO J 17:1268–1278. Chen RH, Blenis J. (1990). Mol Cell Biol 10:3204–3215. Haldar S, Jena N, Croce CM. (1995). Proc Natl Acad Sci USA Chen RH, Chung J, Blenis J. (1991). Mol Cell Biol 11: 92:4507–4511. 1861–1867. Han Y-H, Cao X, Lin B, Kin F, Kolluri SK, Reed JC et al. Cheng EH, Kirsch DG, Clem RJ, Ravi R, Kastan MB, Bedi A (2006). Oncogene, Jan 23. et al. (1997a). Science 278:1966–1968. Hazel TG, Misra R, Davis IJ, Greenberg ME, Lau LF. (1991). Cheng LE, Chan FK, Cado D, Winoto A. (1997b). EMBO J Mol Cell Biol 11:3239–3246. 16:1865–1875. Hazel TG, Nathans D, Lau LF. (1988). Proc Natl Acad Sci Chinnaiyan P, Varambally S, Tomlins SA, Ray S, Huang S, USA 85:8444–8448. Chinnaiyan AM et al. (2006). Int J Cancer 118:1041–1050. Herman JG, Baylin SB. (2003). N Engl J Med 349:2042–2054. Chintharlapalli S, Burghardt R, Papineni S, Ramaiah S, Yoon Hirata Y, Kiuchi K, Chen HC, Milbrandt J, Guroff G. (1993). K, Safe S. (2005). J Biol Chem 280:24903–24914. J Biol Chem 268:24808–24812. Chipuk JE, Bouchier-Hayes L, Kuwana T, Newmeyer DD, Holla VR, Mann JR, Shi Q, Dubois RN. (2005). J Biol Chem. Green DR. (2005). Science 309:1732–1735. Holmes WF, Soprano DR, Soprano KJ. (2003a). J Cell Chipuk JE, Kuwana T, Bouchier-Hayes L, Droin NM, Biochem 89:262–278. Newmeyer DD, Schuler M et al. (2004). Science 303: Holmes WF, Soprano DR, Soprano KJ. (2003b). Oncogene 22: 1010–1014. 6377–6386. Cory S, Huang DC, Adams JM. (2003). Oncogene 22: Holmes WF, Soprano DR, Soprano KJ. (2004). J Cell Physiol 8590–8607. 199:317–329. Costantini P, Jacotot E, Decaudin D, Kroemer G. (2000). J Hsu HC, Zhou T, Mountz JD. (2004). Curr Drug Targets Natl Cancer Inst 92:1042–1053. Inflamm Allergy 3:413–423. Culmsee C, Mattson MP. (2005). Biochem Biophys Res Hsu YT, Wolter KG, Youle RJ. (1997). Proc Natl Acad Sci Commun 331:761–777. USA 94:3668–3672. Dagher PC. (2004). Kidney Int 66:506–509. Igaki T, Kanuka H, Inohara N, Sawamoto K, Nunez G, Daidone MG, Luisi A, Veneroni S, Benini E, Silvestrini R. Okano H et al. (2000). Proc Natl Acad Sci USA 97:662–667. (1999). Endocr Relat Cancer 6:61–68. Jacobs CM, Boldingh KA, Slagsvold HH, Thoresen GH, Davis IJ, Hazel TG, Chen RH, Blenis J, Lau LF. (1993). Mol Paulsen RE. (2004). J Biol Chem 279:50097–50101. Endocrinol 7:953–964. Jacobs CM, Paulsen RE. (2005). Biochem Biophys Res Dawson MI, Hobbs PD, Peterson VJ, Leid M, Lange CW, Commun 336:646–652. Feng KC et al. (2001). Cancer Res 61:4723–4730. Jeong JH, Park JS, Moon B, Kim MC, Kim JK, Lee S et al. de Souza-Pinto NC, Harris CC, Bohr VA. (2004). Oncogene (2003). Ann NY Acad Sci 1010:171–177. 23:6559–6568. Kastner P, Mark M, Chambon P. (1995). Cell 83:859–869. de Urquiza AM, Liu S, Sjoberg M, Zetterstrom RH, Griffiths Katagiri Y, Takeda K, Yu ZX, Ferrans VJ, Ozato K, Guroff W, Sjovall J et al. (2000). Science 290:2140–2144. G. (2000). Nat Cell Biol 2:435–440. Deverman BE, Cook BL, Manson SR, Niederhoff RA, Langer Kaufmann T, Schlipf S, Sanz J, Neubert K, Stein R, Borner C. EM, Rosova I et al. (2002). Cell 111:51–62. (2003). J Cell Biol 160:53–64. Dumont P, Leu JI, Della Pietra III AC, George DL, Murphy Kaul SC, Aida S, Yaguchi T, Kaur K, Wadhwa R. (2005). J M. (2003). Nat Genet 33:357–365. Biol Chem 280:39373–39379. Erlacher M, Michalak EM, Kelly PN, Labi V, Niederegger H, Ke N, Claassen G, Yu DH, Albers A, Fan W, Tan P et al. Coultas L et al. (2005). Blood 106:4131–4138. (2004). Cancer Res 64:8208–8212. Erster S, Mihara M, Kim RH, Petrenko O, Moll UM. (2004). Kelly KJ, Plotkin Z, Vulgamott SL, Dagher PC. (2003). JAm Mol Cell Biol 24:6728–6741. Soc Nephrol 14:128–138.

Oncogene p53 and Nur77/TR3 UM Moll et al 4742 Kochel I, Rapak A, Ziolo E, Strzadala L. (2005). Biochem Nemajerova A, Erster S, Moll UM. (2005a). Cell Death Differ Biophys Res Commun 334:1102–1106. 12:197–200. Kolluri SK, Cao X, Bruey-Sedano N, Lin B, Lin F, Han Y Nemajerova A, Wolff S, Petrenko O, Moll UM. (2005b). et al. (2003). Mol Cell Biol 23:8651–8667. FEBS Lett 579:6079–6083. Kolluri SK, Corr M, James SY, Bernasconi M, Lu D, Liu W Park KC, Song KH, Chung HK, Kim H, Kim DW, Song JH et al. (2005). Proc Natl Acad Sci USA 102:2525–2530. et al. (2005). Mol Endocrinol 19:12–24. Komarov PG, Komarova EA, Kondratov RV, Christov- Pei L, Castrillo A, Chen M, Hoffmann A, Tontonoz P. Tselkov K, Coon JS, Chernov MV et al. (1999). Science 285: (2005a). J Biol Chem 280:29256–29262. 1733–1737. Pei L, Castrillo A, Tontonoz P. (2005b). Mol Endocrinol 20: Kuwana T, Bouchier-Hayes L, Chipuk JE, Bonzon C, Sullivan 786–794. BA, Green DR et al. (2005). Mol Cell 17:525–535. Pekarsky Y, Hallas C, Palamarchuk A, Koval A, Bullrich F, Labelle Y, Bussieres J, Courjal F, Goldring MB. (1999). Hirata Y et al. (2001). Proc Natl Acad Sci USA 98: Oncogene 18:3303–3308. 3690–3694. Labelle Y, Zucman J, Stenman G, Kindblom LG, Knight J, Peng Z. (2005). Hum Gene Ther 16:1016–1027. Turc-Carel C et al. (1995). Hum Mol Genet 4:2219–2226. Perlmann T, Jansson L. (1995). Genes Dev 9:769–782. Le W, Conneely OM, He Y, Jankovic J, Appel SH. (1999). J Petros AM, Gunasekera A, Xu N, Olejniczak ET, Fesik SW. Neurochem 73:2218–2221. (2004). FEBS Lett 28073:1–4. Lee JM, Lee KH, Farrell CJ, Ling PD, Kempkes B, Park JH Petros AM, Medek A, Nettesheim DG, Kim DH, Yoon HS, et al. (2004). J Virol 78:12694–12697. Swift K et al. (2001). Proc Natl Acad Sci USA 98: Lee JM, Lee KH, Weidner M, Osborne BA, Hayward SD. 3012–3017. (2002). Proc Natl Acad Sci USA 99:11878–11883. Pfahl M, Piedrafita FJ. (2003). Oncogene 22:9058–9062. Lee KW, Ma L, Yan X, Liu B, Zhang XK, Cohen P. (2005). J Philips A, Lesage S, Gingras R, Maira MH, Gauthier Y, Hugo Biol Chem 280:16942–16948. P et al. (1997). Mol Cell Biol 17:5946–5951. Lee MO, Kang HJ, Cho H, Shin EC, Park JH, Kim SJ. (2001). Rajpal A, Cho YA, Yelent B, Koza-Taylor PH, Li D, Chen E Biochem Biophys Res Commun 288:1162–1168. et al. (2003). EMBO J 22:6526–6536. Lengqvist J, Mata De Urquiza A, Bergman AC, Willson TM, Ramaswamy S, Ross KN, Lander ES, Golub TR. (2003). Nat Sjovall J, Perlmann T et al. (2004). Mol Cell Proteom 3: Genet 33:49–54. 692–703. Reed JC. (1998). Oncogene 17:3225–3236. Leu JI, Dumont P, Hafey M, Murphy ME, George DL. Sakurada K, Ohshima-Sakurada M, Palmer TD, Gage FH. (2004). Nat Cell Biol 6:443–450. (1999). Development 126:4017–4026. Lewis J, Oyler GA, Ueno K, Fannjiang YR, Chau BN, Sansome C, Zaika A, Marchenko ND, Moll UM. (2001). Vornov J et al. (1999). Nat Med 5:832–835. FEBS Lett 488:110–115. Li H, Kolluri SK, Gu J, Dawson MI, Cao X, Hobbs PD et al. Schuler M, Green DR. (2001). Biochem Soc Trans 29:684–688. (2000). Science 289:1159–1164. Shin HJ, Lee BH, Yeo MG, Oh SH, Park JD, Park KK et al. Li Y, Lin B, Agadir A, Liu R, Dawson MI, Reed JC et al. (2004). Carcinogenesis 25:1467–1475. (1998). Mol Cell Biol 18:4719–4731. Shipp MA, Ross KN, Tamayo P, Weng AP, Kutok JL, Aguiar Lim RW, Varnum BC, Herschman HR. (1987). Oncogene 1: RC et al. (2002). Nat Med 8:68–74. 263–270. Slagsvold HH, Ostvold AC, Fallgren AB, Paulsen RE. (2002). Lin B, Kolluri S, Lin F, Liu W, Han YH, Cao X et al. (2004a). Biochem Biophys Res Commun 291:1146–1150. Cell 116:527–540. Speidel D, Helmbold H, Deppert W. (2005). Oncogene. Lin XF, Zhao BX, Chen HZ, Ye XF, Yang CY, Zhou HY Srivastava RK, Mi QS, Hardwick JM, Longo DL. (1999). et al. (2004b). J Cell Sci 117:5609–5621. Proc Natl Acad Sci USA 96:3775–3780. Liu D, Jia H, Holmes DI, Stannard A, Zachary I. (2003). Subramanian T, Chinnadurai G. (2003). J Cell Biochem 89: Arterioscler Thromb Vasc Biol 23:2002–2007. 1102–1114. Liu S. (2002). World J Gastroenterol 8:446–450. Suzuki S, Suzuki N, Mirtsos C, Horacek T, Lye E, Noh SK Liu ZG, Smith SW, McLaughlin KA, Schwartz LM, Osborne et al. (2003). Proc Natl Acad Sci USA 100:8276–8280. BA. (1994). Nature 367:281–284. Swanson KD, Taylor LK, Haung L, Burlingame AL, Land- Lotan R. (2003). J Biol Regul Homeost Agents 17:13–28. reth GE. (1999). J Biol Chem 274:3385–3395. Maddika S, Booy EP, Johar D, Gibson SB, Ghavami S et al. Talos F, Petrenko O, Mena P, Moll UM. (2005). Cancer Res (2005). J Cell Sci 118:4485–4493. 65:9971–9981. Mangelsdorf DJ, Evans RM. (1995). Cell 83:841–850. Uemura H, Chang C. (1998). Endocrinology 139:2329–2334. Marchenko ND, Zaika A, Moll UM. (2000). J Biol Chem 275: Uhlmann EJ, Subramanian T, Vater CA, Lutz R, Chinnadurai 16202–16212. G. (1998). J Biol Chem 273:17926–17932. Martin B, Paesmans M, Berghmans T, Branle F, Ghisdal L, Vousden KH, Lu X. (2002). Nat Rev Cancer 2:594–604. Mascaux C et al. (2003). Br J Cancer 89:55–64. Wang Z, Benoit G, Liu J, Prasad S, Aarnisalo P, Liu X et al. Martinez-Gonzalez J, Badimon L. (2005). Cardiovasc Res 65: (2003). Nature 423:555–560. 609–618. Wansa KD, Harris JM, Muscat GE. (2002). J Biol Chem 277: Masuyama N, Oishi K, Mori Y, Ueno T, Takahama Y, Gotoh 33001–33011. Y. (2001). J Biol Chem 276:32799–32805. Wansa KD, Harris JM, Yan G, Ordentlich P, Muscat GE. Meinke G, Sigler PB. (1999). Nat Struct Biol 6:471–477. (2003). J Biol Chem 278:24776–24790. Mihara M, Erster S, Zaika A, Petrenko O, Chittenden T, Wei MC, Lindsten T, Mootha VK, Weiler S, Gross A, Ashiya Pancoska P et al. (2003). Mol Cell 11:577–590. M et al. (2000). Genes Dev 14:2060–2071. Milbrandt J. (1988). Neuron 1:183–188. Wei MC, Zong WX, Cheng EH, Lindsten T, Panoutsakopou- Moll U. (2000). New p53-based strategies for cancer therapy. lou V, Ross AJ et al. (2001). Science 292:727–730. Natick, MA:Eaton Publishing. Weih F, Ryseck RP, Chen L, Bravo R. (1996). Proc Natl Acad Mu X, Chang C. (2003). J Biol Chem 278:42840–42845. Sci USA 93:5533–5538.

Oncogene p53 and Nur77/TR3 UM Moll et al 4743 Willis SN, Chen L, Dewson G, Wei A, Naik E, Fletcher JI Xue D, Horvitz HR. (1997). Nature 390:305–308. et al. (2005). Genes Dev 19:1294–1305. Yamamoto K, Ichijo H, Korsmeyer SJ. (1999). Mol Cell Biol Wilson AJ, Arango D, Mariadason JM, Heerdt BG, Augen- 19:8469–8478. licht LH. (2003). Cancer Res 63:5401–5407. Yang Y, Bailey J, Vacchio MS, Yarchoan R, Ashwell JD. Wingate AD, Campbell DG, Peggie M, Arthur JS. (2005). (1995). Proc Natl Acad Sci USA 92:3051–3055. Biochem J. Yang Y, Vacchio MS, Ashwell JD. (1993). Proc Natl Acad Sci Woronicz JD, Calnan B, Ngo V, Winoto A. (1994). Nature USA 90:6170–6174. 367:277–281. Yoo YA, Kim MJ, Park JK, Chung YM, Lee JH, Chi SG et al. Woronicz JD, Lina A, Calnan BJ, Szychowski S, Cheng L, (2005). Mol Cell Biol 25:6603–6616. Winoto A. (1995). Mol Cell Biol 15:6364–6376. Yoo YG, Yeo MG, Kim DK, Park H, Lee MO. (2004). J Biol Wu Q, Dawson MI, Zheng Y, Hobbs PD, Agadir A, Jong L Chem 279:53365–53373. et al. (1997a). Mol Cell Biol 17:6598–6608. Yoshida Y, Izumi H, Torigoe T, Ishiguchi H, Itoh H, Kang D Wu Q, Li Y, Liu R, Agadir A, Lee MO, Liu Y et al. (1997b). et al. (2003). Cancer Res 63:3729–3734. EMBO J 16:1656–1669. Zeimet AG, Marth C. (2003). Lancet Oncol 4:415–422. Wu Q, Liu S, Ye XF, Huang ZW, Su WJ. (2002a). Zetterstrom RH, Solomin L, Jansson L, Hoffer BJ, Olson L, Carcinogenesis 23:1583–1592. Perlmann T. (1997). Science 276:248–250. Wu WS, Heinrichs S, Xu D, Garrison SP, Zambetti GP, Zhang XK. (2002). Endocr Relat Cancer 9:87–102. Adams JM et al. (2005). Cell 123:641–653. Zhao Y, Chaiswing L, Velez JM, Batinic-Haberle I, Colburn Wu WS, Xu ZX, Ran R, Meng F, Chang KS. (2002b). NH, Oberley TD et al. (2005). Cancer Res 65:3745–3750. Oncogene 21:3925–3933. Zhong Q, Gao W, Du F, Wang X. (2005). Cell 121:1085–1095.

Oncogene