Disponible à / Available at permalink :

- - -

- - -

Dépôt Institutionnel de l’Université libre de Bruxelles / Université libre de Bruxelles Institutional Repository

Thèse de doctorat/ PhD Thesis Citation APA:

Maetens, M. M. (2007). Regulation of the tumor suppressor p53 by Mdm2 and Mdm4 (Unpublished doctoral dissertation). Université libre de Bruxelles, Faculté des Sciences – Sciences biologiques, Bruxelles.

Disponible à / Available at permalink : https://dipot.ulb.ac.be/dspace/bitstream/2013/210602/6/19710276-746c-4d00-8268-b466a41a7dce.txt

(English version below)

Cette thèse de doctorat a été numérisée par l’Université libre de Bruxelles. L’auteur qui s’opposerait à sa mise en ligne dans DI-fusion est invité à prendre contact avec l’Université (di-fusion@ulb.be).

Dans le cas où une version électronique native de la thèse existe, l’Université ne peut garantir que la présente version numérisée soit identique à la version électronique native, ni qu’elle soit la version officielle définitive de la thèse.

DI-fusion, le Dépôt Institutionnel de l’Université libre de Bruxelles, recueille la production scientifique de l’Université, mise à disposition en libre accès autant que possible. Les œuvres accessibles dans DI-fusion sont protégées par la législation belge relative aux droits d'auteur et aux droits voisins. Toute personne peut, sans avoir à demander l’autorisation de l’auteur ou de l’ayant-droit, à des fins d’usage privé ou à des fins d’illustration de l’enseignement ou de recherche scientifique, dans la mesure justifiée par le but non lucratif poursuivi, lire, télécharger ou reproduire sur papier ou sur tout autre support, les articles ou des fragments d’autres œuvres, disponibles dans DI-fusion, pour autant que :

Le nom des auteurs, le titre et la référence bibliographique complète soient cités;

L’identifiant unique attribué aux métadonnées dans DI-fusion (permalink) soit indiqué;

Le contenu ne soit pas modifié.

L’œuvre ne peut être stockée dans une autre base de données dans le but d’y donner accès ; l’identifiant unique (permalink) indiqué ci-dessus doit toujours être utilisé pour donner accès à l’œuvre. Toute autre utilisation non mentionnée ci-dessus nécessite l’autorisation de l’auteur de l’œuvre ou de l’ayant droit.

--- English Version ---

This Ph.D. thesis has been digitized by Université libre de Bruxelles. The author who would disagree on its online availability in DI-fusion is invited to contact the University (di-fusion@ulb.be).

If a native electronic version of the thesis exists, the University can guarantee neither that the present digitized version is identical to the native electronic version, nor that it is the definitive official version of the thesis.

DI-fusion is the Institutional Repository of Université libre de Bruxelles; it collects the research output of the University, available on open access as much as possible. The works included in DI-fusion are protected by the Belgian legislation relating to authors’ rights and neighbouring rights.

Any user may, without prior permission from the authors or copyright owners, for private usage or for educational or scientific research purposes, to the extent justified by the non-profit activity, read, download or reproduce on paper or on any other media, the articles or fragments of other works, available in DI-fusion, provided:

The authors, title and full bibliographic details are credited in any copy;

The unique identifier (permalink) for the original metadata page in DI-fusion is indicated;

The content is not changed in any way.

It is not permitted to store the work in another database in order to provide access to it; the unique identifier (permalink) indicated above must

DBM 00697

r

cTPir;3[r7

A

V J

Université Libre de Bruxelles Faculté des Sciences

Institute of Molecular Bioiogy and Medecine Laboratory of Molecular Embryology

And VIB

Laboratory for Molecular and Cancer Bioiogy

Promotor Dr. Jean-Christophe Marine Co-promoter Dr. Eric Bellefroid

Régulation of the tumor suppressor p53 by Mdm2 and Mdm4

Marion Maetens

Academie year 2007-2008

Thesis submitted of the requirement for the degree of Doctor in Sciences

Merci...

A toi Jean-Christophe de m’avoir accueillie dans ton laboratoire tout d’abord en tant que mémorante et ensuite comme doctorante. Grâce à toi j’ai découvert le monde de la recherche fondamentale avec tous ses inconvénients comme le stress, l’angoisse, la compétition et les sacrifices... Mais aussi et surtout tous ses avantages comme la passion, l’enthousiasme, l’excitation, la surprise, l’esprit de groupe et de solidarité. Merci pour ton soutien, tes encouragements et ta patience.

Au Laboratoire d’Embryologie Moléculaire de l’Université Libre de Bruxelles et plus particulièrement à Eric Bellefroid pour m’avoir permis de réaliser un mémoire un sein de son laboratoire ainsi que le début de ma thèse de doctorat.

Au VIB de m’avoir permis de réaliser une thèse de doctorat dans un environnement scientifique de haute qualité.

A la « Vascular Cell Biology Unit » du VIB et plus particulièrement à Jody Haigh pour la collaboration mise en place afin de développer les souris conditionnelles transgéniques ainsi qu’aux membres de son laboratoire comme Omar Nyabi, Khatarina Haigh et Michael Naessens pour toute l’aide technique apportée pour la construction des vecteurs d’intégration ainsi que la manipulation et l’agrégation des cellules ES.

Au Laboratoire de Biologie Moléculaire et de Biologie du Développement de l’Université de Liège et plus particulièrement à Marianne Voz et Hélène Pendeville pour m’avoir fait découvrir le monde du Zebrafish et ses secrets...

Aux nombreux membres du personnel de l’ULB et du VIB qui Jour après Jour s’occupent de la maintenance de nos précieuses petites souris.

Aux membres de mon laboratoire Chantal, Alexandra, David, Rose et Debby avec qui J’ai vraiment apprécié de travailler. Irina, J’ai mis du temps à te connaître et Je le regrette, merci de toujours m’avoir interpellée avec tes nombreuses questions. Sven, merci d’avoir toujours pris ton temps pour répondre à mes questions, traduire un projet IWT ou encore participer à la correction de ce manuscrit. Vanessa, une des dernières arrivée au sein du laboratoire mais avec qui J’ai très vite noué des liens d’amitié et qui m’a permis d’arriver à la fin de cette thèse très sereinement. Enfin Dieter, Sarah et Aga, vous êtes les personnes avec qui J’ai le plus interagi et travaillé, vous êtes de véritables amis et sans vous une bonne partie de ce travail n’aurait pas pu aboutir. Merci Dieter pour ton soutien technique et moral. Sarah et Aga, nous formons une belle équipe. J’espère que « nos » souris vous permettrons de

Merci

Aux anciens membres du laboratoire Fred, Lode, Pascal et plus particulièrement Alain, Sarah et Gilles. Vous m’avez formée, ce qui n’a pas dû être tous les jours facUejdmais encore une fois J’ai trouvé en vous de véritables amis sur qui je peux encore compter jour après jour... Gilles merci de m’avoir épaulée au début du projet EpoR et de continuer à le faire même à distance.

A tous ceux qui me supportent certains depuis plus de 10 ans mais sans qui la vie serait bien triste et surtout qui me rappellent que le bonheur ne se trouve pas au fond d’un Eppendorf... Je parle de mes amis. Je ne vais pas tous vous citer tout d’abord de peur d’en oublier certains mais surtout parce que vous vous reconnaîtrez. Il s’en est passé des choses pendant ces 4 années de thèse (entre autre, je suis devenue marraine). Merci d’être encore là, d’avoir fait, pour certains, semblant de s’intéresser à ce que je « cherchais » sans jamais rien comprendre à ce que je racontais et pour d’autres d’avoir été tout

simplement présents dans les bons comme dans les mauvais moments.

A Kris. Sans toi je n’aurais jamais fait de thèse car je serai toujours bloquée en deuxième candi à cause de ce f.... cours de stat ! Non sans toi je n’y serais pas arrivé... Tu m’as redonné confiance en moi. Merci... Et pour tout le reste aussi.

A ma famille qui est très présente et qui m’a toujours apporté beaucoup de soutien. Merci à ma sœur qui est mon petit « Gemini criquet » à moi et enfin et surtout à mes parents à qui, il est vrai j’en ai fait voir, mais qui m’ont toujours témoigné beaucoup d’amour et de soutien. J’ai beaucoup de chance de vous avoir et j’espère ne jamais vous décevoir.

Enfin à mon Loulou... Il sait pourquoi. Comme le dit la chanson : « Je ne regrette rien car ma vie (et une petite autre aussi) commence avec toi... »

Table of contents

List of abbreviations

L Introduction ... i

Æ p 53 the gardian of the qenome...1

B. Functions of the 4755 gene and the downstream proqram... 2

K Structure... 2

2_, Functions...4

a} P53 and cell cycle arrest... 4

b} P53 and apoptosis... 5

ç} Choosina between h'fe and death... 7

^ New functions of p53...8

C. Activation of p53 and the upstream proqram... 9

K Activation bv DNA damage... 9

al N-termina! modifications of p53... 1 0 0) In vitro... 10

(2) In vivo... 12

M C-termina! modifications of p53... 1 3

Table of contents

D.

53’

régulation bv subcellular localization and interactions...1 5 1. Régulation of p53 subcellular localization...1 5 2. The complex régulation of p53 by Mdm2 and Mdm4...1 6

^ Mdm2... 16

b} Mdm4... 16

ç} The collaboration of Mdm2 and Mdm4 in p53’s régulation...1 7 (1) Régulation ofthe p53 transcriptiona! activity...1 7 (2) Régulation ofthe p53 protein expression ievei and stabiiity...1 8 The compiementarv roie of Mdm2 and Mdm4 in p53’s régulation... 19

(1) in vitro... 19

(2) in vivo... 20

3^ The Régulation of p53 by additional mechanisms/proteins...21

a} Other ubiguitin iioases... 21

b} Hausp... 22

ç} Arf-...22

L Inactivation ofthe p53 pathwav in cancers...24

K Qncogenic prooerties of Mdm2 and Mdm4...25

2. p53 as a therapeutic target...26

U. Ai ms ofthe Project...27

///. Results... 28

A. Part I Distinct rôles of Mdm2 and Mdm4 in red cell production 28

The development of blood cells...28

2^ Mission statement... 31

3^ Specifically inactivation of Mdm2 and Mdm4 in the erythroid lineaqe in vivo.. 32

aj Experimental strateav...32

b} Mdm2, but not Mdm4, is essentia! for erythroidproaenitor cellsurviva!...33

ç} Mdm2 protects erythroid proaenitor cells from p53-mediated apoptosis... 34

^ Mdm4 is oniv required for the hiqh erythropoietic rate durinp embrvonic definitive ervthropoiesiss...35

eï Mdm4 protects erythroid proaenitor cells from p53-mediated cell cycle arrest...36

Q Mdm4 is not required for production ofaduit red ceiis... 37

g} Conclusions...38

B. Part II Dissectinq Mdm2 and Mdm4 oncoqenic properties in vivo usinq conditional transqenic mice ... 39

l_, Mdm4 is an oncoqene... 39

2^ Mission statement...41

3^ Génération of Mdm2 and Mdm4 conditional transqenic mouse lines... 42

a} Experimental strateav... 42

b} Tarqetinp vectors constructs and ES cells manipulation... 43

çï Screening the ES cells bv PCR...44

dï Screening the ES ceiis bv Southern biottina...44

Table of contents

gl Génération of conditiona! transpenic animais... 46

tû inducibie overexpression ofthe transoene in vivo...46

il Conclusions...47

IV. Discussion... 48

Æ Part I Mdm2 and Mdm4 in red cell production... 48

^ Part II Dissectinq Mdm2 and Mdm4 oncoqenic properties in vivo usinq conditional transgenic mice... 51

V Publications... 55

Vi. Annex... s6 A. Materiel and methods... 56

K Mouse strains...56

2. Genotypinq ofthe mdm2 and mdm4 conditional transgenic mice by PCR... 56

3^ Histoloqy. IHC, and ISEL staininq...57

4. Immunohistofluorescence... 57

I. Hematocrit and phenvihvdrazine stress test ...57

6. Purification of erythroid cells... 57

Colony assays...58

^ Q-RT-PC R... 58

9. Western blottinq... 58

10. DNA constructs... 58

II. ES cells manipulation...59

12. Southern blottinq ₅₉

1 3. LacZ staininq... 60

^ Biblioqraphv... 61

abbreviations

List of abbreviafions

aa amino acid

ALDH4 Aldéhyde dehydrogenase family 4 AMU acute myeloid leukemia 1

AP Alkaline phosphatase AP-1 activating protein 1

APAF-1 Apoptosis-Protease Activating Factor 1 ARF Alternative Reading Frame protein ARF-BPl ARF Binding Protein 1

ASPP Apoptosis-Stimulating Protein of p53 ATM Ataxia Telangiectasia Mutated

ATR Ataxia Telangiectasia and Rad3 related protein Bax BCL2-Associated X Protein

Bcl2 B-cell lymphoma/leukemia-2 gene BH3 Bcl-2 Homology 3

bp base pair

BrdU 5-Bromo-2’-deoxyUridine

CARPs Caspase-8/-10 associated RING domain proteins casp-3* caspase 3-cleaved

CBP CREB Binding Protein

Cdc2 cell-division-cycle-2 kinase Cdk Cyclin dépendent kinase cDNA complementary DNA

CHIP carboxyl terminus of Hsp70-interacting protein Chkl/2 Checkpoint kinase 1 /2

CMV CytoMegaloVirus

CNS COPl et CTBP2 Da DAB DAPI DNA dNTP

DYRK2 E ES Gadd45 gapdh (E)CFP CI H&E

H2O

HAT HAUSP HDAC HRP IHC

Central Nervous System

COnstitutively Photomorphogenic 1 Carboxy terminus

C-terminal binding protein 2 Dalton

3,3’-DiAminoBenzidine DiAmidoPhenylIndole DeoxyriboNucleic Acid

deoxyriboNucleotid TriPhosphate Day post coïtum

Dual specificity tyrosine-phosphorylation-regulated kinase 2 Embryonic day n

Embryonic stem cells

growth arrest and DNA damage-45 alpha glyceraldehyde-3-phosphate dehydrogenase (Enhanced) Green Fluorescent Protein

Gastro-Intestinal Hematoxylin and Eosin water

Histone AcetylTransferases

Herpes virus-Associated Ubiquitin-Specific Protease Histone DeACetylase

HorseRadish Peroxidase

ImmunoHistoChemistry

abbreviations

kb kDa

LTR Lys Mdm2 Mdm4 MEFs MIRA Neo NES N LS NoLS NP-40 N PC Nt nt P PACT PAGE Parc PBS PCAF PCR PC K

kilobase kiloDalton

Lox?-SJO? - Lo)^

Long Terminal Repeat Lysine residue

Mouse double minute 2 Mouse double minute 4 Mouse Embryonic Fibroblasts

mutant p53 réactivation and induction rapid apoptosis Neomycine

Nuclear Export Signal Nuclear Localisation Signal Nucleolar Localisation Signal Nonidet-P40

Neural Progenitor Cells Amino terminus

nucléotide Postnatal day n

p53-associated cellular protein-testes derived PolyAcrylamide Gel Electrophoresis

p53-associated, parkin-like cytoplasmic protein Phosphate Buffered Saline

p300/CBP Associated Factor Polymerase Chain Reaction

PhosphoGlycerate Kinase promoter

pHH3 phosphorylated Histone H3

Pidd p53-induced protein with a death domain Pirh-2 p53-induced protein with a RINC-H2 domain PML Promyelocytic Leukemia Protein

PRIMA p53 réactivation and induction of massive apoptosis

prom promoter

PUMA p53 up-regulated modulator of apoptosis

qRT-PCR quantitative Reverse Transcription-Polymerase Chain Reaction

Rb Retinoblastoma

RE Responsive Elément

RING Really Interesting New Gene RNA RiboNucleic Acid

rRNA ribosomal RiboNucleic Acid RT Reverse Transcription SA Site Acceptor of splicing SDS Sodium Dodecyl Sulphate Ser Serine residue

SMC Smooth Muscle Cells

SUMO Small Ubiquitin-related Modifier TBS Tris Buffered Saline

Thr Threonine residue

TICAR TP53-induced glycolysis and apoptosis regulator

Triton XI00 4-(l ,1,3.3-Tetraethylbutyl) phenyi-polyethylene glycol

abbreviations

Tween 20 U

UV WT A

monolaurate of polyethylene glycol sorbitan Unit

Ultra Violet light

Wild-Type

délétion

responses, and it is possible that different responses are induced by different stress signais. There is evidence that p53 can play a part in determining which response is induced through differential activation of target-gene expression. Although the importance of these responses to tumour suppression is clear, previousiy unanticipated contributions of these responses to other aspects of human heaith and disease are being uncovered. The rôle of p53 in tumour suppression, development and ageing is likely to dépend on which cellular response is activated and on the context in which the activation occurs. (Vousden and Lane, 2007).

introduction

L Introduction

Æ

53 the gardian of the genome

The p53 protein was first described in 1979 (Lane and Crawford, 1979;

Linzer and Levine, 1 979). It was originally identified as an oncogene because a mutated form was cloned from a human cancer cell line. Subséquent genetic studies in mice and human highiighted a key rôle in tumor suppression. p53 null mice are viable but highiy tumor prone (Donehower et al., 1 992; Jacks et a!., 1 994). The p53 gene itself is mutated in about 50% of human tumors of diverse origins (Baker et a!., 1 990).

Importantly, in the remaining 50%, harbouring wild-type p53, the p53 protein is inactivated indirectiy through binding to viral proteins, or as a resuit of alterations in genes whose products modulate p53 activity or médiate its biological activities.

The transcription factor p53 acts as a cellular node in a complex signaling pathway (figurel) that evolved to sense a broad range of stress stimuli such as DNA damage, oncogene activation, nucléotide déplétion, hypoxia,... in the absence of cellular stress, the p53 protein is expressed at low steady-state levels and exerts little, if any, effect on cell fate. However, in response to various types of stress, p53 becomes activated and this is reflected in elevated protein levels as well as augmented biochemical functions. As a conséquence of p53 activation, cells undergo marked phenotype changes, ranging from réversible cell cycle arrest, senescence, or apoptosis. As a conséquence, upon oncogenic stress, p53 triggers a response that either favors repair of the damage or éliminâtes the damaged cells from the réplicative pool trough induction of apoptosis. These activities thereby prevent expansion of these cells into a large population of malignant progeny. More recently, numerous studies hâve identified p53-regulated genes that could play a rôle in a number of different and sometimes unexpected responses. There is now clear evidence for a rôle of p53 in the régulation of glycolysis and autophagy, invasion and motility, angiogenesis, differientation and bone remodelling (Vousden and Lane, 2007).

1

CZI^ Transact<vaticxi Sequence-specilic DNA binding E3 Njciear export signais C3 Njcoar-iocatzation signais r~7, Proline ricfi I I Oigonerization

Figurez Structure of p53 p53 is a transcription factor that contains several well-defined domains, including the amino-terminal transactivation domain, a central sequence-specific DNA-binding core and a carboxy-terminal région that contains oligomérization sequences and nuclear-localization signais. Nuclear export of p53 is regulated by signais in the amino and carboxyl termini. Interaction of proteins such as MDM2 or p300/CBP with the amino terminus of p53 can lead to modifications such as acétylation or ubiquitylation in the carboxyl terminus. Almost ail of the point mutations that are found in cancers occur within the central DNA-binding core of p53; the percentage of mutations within each région detected in cancers to date is indicated below. Some 28% of mutations occur within the six highiighted residues. (Vousden and Lu, 2002).

introduction

B. Functions of the p53 Qene and the downstream proqram

1. Structure

It has become clear that p53 is a sequence-specific DNA-binding transcription factor that binds DNA as a tetramer and activâtes or represses transcription from a large and increasing number of target genes (El-Deiry,

1998).

The first 100 amino acids of p53 (figureZ) contain two transactivation domains, a proline-rich domain and a nuclear export signal. Both are required for full transactivation ability. The p53 N-terminus contains aiso a proline-rich domain wich contains five copies of the motif “PXXP”. This domain has been implicated in the régulation of p53-mediated apoptosis (Baptiste et a!., 2002) and in p53 stabilization and activation (Berger et a!., 2001). The N-terminus of p53 is higly post-translationally modified (Appella and Anderson, 2001). Phosphorylation events in this région hâve been implicated in both the stability of p53 as well as the specificity of target transactivation (Chao et a!., 2003; Oda état., 2000).

Sequencing of the p53 locus from over 16000 human tumors provides evidence for the importance of the sequence-specific DNA binding domain (amino acids 100-300) for intact tumor suppressor function (Olive et a!., 2004). 97% of the mapped mutations résidé in the sequence-specific DNA binding domain and render the protein inactive for target sequence binding (Ko and Prives, 1 996). This région lies in the center of the protein and exhibits zinc-binding activity required for proper protein folding. p53 binds a consensus DNA sequence that contains two copies of the inverted pentameric sequence PuPuPuCA/rT/AGpyPyPy separated by 0 to 1 3 base pairs with the 4'h C and 7^^ G being the last variant, it was shown recently that p53 can bind a halfsite of the p53’s RE (Menendez et al., 2006). Most functions involving the core domain dépend on its ability to interact with DNA sequence-specifically, however some interesting protein-protein interactions hâve been localized to the core domain including, for example, ASPPl and ASPP2. By interacting with the p53 core domain, these proteins enhance the interaction of p53 with the promoters of pro-apoptotic target genes in vivo (Samuels-Lev et al, 2001).

2

Réactivation of mutant p53 in tumors should trigger massive apoptosis and eliminate the tumor cells. To test this, Bykov and its colleagues hâve screened a library of low-molecular-weight compounds in order to identify compounds that can restore wild-type function to mutant p53. They found one compound capable of inducing apoptosis in human tumor cells through restoration of the transcriptional transactivation function to mutant p53. This molécule, named PRIMA-1, restored sequence-specific DNA binding and the active conformation to mutant p53 proteins in vitro and in living cells.

PRIMA-1 rescued both DNA contact and structural p53 mutants. Importantly, intravenous administration of PRIMA-1 to mice harboring xenografts tumors expressing mutant p53 led to tumor shrinkage with no apparent toxicity (Bykov et ai, 2002). This molécule may serve as a lead compound for the development of anticancer drugs targeting mutant p53.

The oligomérization domain (amino acids 320-360) is responsible for the tetrameric State of p53. The structure of the tetramerization domain has been determined and forms a “dimer of dimers” wich folds in the same manner (Liang and Clarke, 2001 ). Furthermore, p53 binds DNA with the same organization, each dimer binds two contigous quarter-sites of the target sequences. A second nuclear export signal résides in the tetramerization domain.

Finally, the C-terminus domain of p53 (amino acids 360-390), the basic domain, is able to interact with DNA in a sequence nonspecific manner.

It aiso increases the transactivation capacity of p53 although it may function as a négative regulator of the core DNA binding domain (Wang et aL, 1993).

Cells expressing a p53 C-terminal truncated mutant at physiological levels exhibit defects in the induction of p21, a key p53 target gene, and apoptosis in response to stress (Chen et aL, 1996). In addition, this région contains two minor nuclear localization signais an d multiple regulatory lysines. These lysines are the sites for several modifications such as ubiquitylation (Michael and Oren, 2003), sumoylation (Rodriguez et aL, 1999), acétylation, glycosylation, neddylation and méthylation (Appella and Anderson, 2001).

Biochemical data suggest that these modifications affect p53 stability, localization, DNA binding activity and/or transcriptional activity.

Phosphorylation by protein kinase C (Takenaka et al., 1995) and casein

kinase II (Hupp et al., 1992) can aIso augment p53’s ability to bind DNA

sequence specifically. Futhermore, deacetylation by HDACl (Luo et al, 2000)

and SIR2cx (Luo et al., 2001) decreases p53-mediated apoptosis (Luo et al.

introduction

2001 : Luo et a!., 2000). Recently, it was shown that the lysine methyltransferase enzyme SET8/PR-SET7 specifically monomethylates p53 at lysine 382. This méthylation event robustly suppresses p53-mediated transactivation of highiy responsive target genes but has little influence on

”weak” targets. Déplétion of SET8 augments the proapoptotic and checkpoint activation functions of p53, and accordingly, SET8 expression is downregulated upon DNA damage (Shi et ai, 2007).

2. Functions

The mechanisms by wich p53 accomplishes its biological functions are still not completely understood. Many of the genes induced by p53 can be divided into functional categories based on the biological activity that they médiate, such as cell cycle arrest, DNA repair, apoptosis, senescence, ... (Sax and El-Deiry, 2002). The coordinated expression of these genes, depending on the cell type, environment and the type of stimulus détermine the cell fate. Although it is generally believed that p53 functions essentially as an activator of transcription, p53 is aiso capable of repressing transcription (Ho and Benchimol, 2003). Other activities of p53 independent of transcription hâve been described, including the ability of p53 to re-localize death receptors from the Goigi to the cell surface (Bennett et al., 1998). A proapoptotic rôle of p53 at the mitochondria was aIso proposed. p53 specifically interacts with Bcl-2 family members to favour Bax-mediated cytochrome-c release (Chipuk et al., 2004).

^ p53 and cell cycle arrest

Transient alterations in cell cycle checkpoints after exposure to DNA damaging agents hâve been observed in many cell types (Hartwell and Weinert, 1989). These alterations presumably permit optimal repair of damage before cells enter S-phase or mitosis following Cl or G2 cell cycle checkpoints, respectively. Failure to repair DNA damage prior to réplicative synthesis or mitosis could resuit in the propagation of mutagenic lésions and eventually to the progressive accumulation of genomic changes necessary for neoplastic transformation. p53 is required for DNA damage-induced G1 arrest primarily trough transactivation of the most-studied p53-downstream target p2l, a cyclin-dependent kinase inhibitor (El-Deiry et a!., 1993; Harper et a!., 1993). The elevated p21 binds and inactivâtes cyclin E/Cdk2 complexe

4

activation occurs, death receptors form the "Death Inducing Signaling Complex” (DISC) that recruits, via the adaptor molécule Fas Associated Death Domain Protein (FADD), multiple procaspase- 8 molécules resulting in caspase-8 activation. This signal in some cell types is sufficient to trigger apoptosis. In other cells types, caspase-8 interacts with the mitochondrial pathway activating the BH3 only protein Bid that leads to cytochrome-c release. Négative regulators of this pathway are cellular FLICE-like inhibitory protein (c-Flip), a degenerate caspase homologue that can be recruited by FADD blocking the caspase-8 activation; DcRl and DcR2 antiapoptotic decoy-receptors that missing the cytoplasmic death domain compete with DR4 and DR5 blocking the TRAlL-induced apoptosis. The intrinsic pathway is used extensively in response to extracellular eues and internai insults as DNA damage. These diverse response pathways converge on the mitochondria altering the balance between pro-apoptotic (Bax/Bak) and anti-apoptotic (Bcl2/BclXL) proteins. The dominance of pro-apoptotic Bax/Bak proteins, induced by BHS-only proteins (Puma; Noxa) and p53, increase the mitochondriai permeabiiity and releasing apoptogenic factors; cytochromec and Apaf-1 once in the cytopiasm form the apoptosome complessing and activating procaspase-9 molécules; Diablo/Second mitochondria- derived activator of caspase (Smac) and Omi/high température requirement protein A2 (HTRA2) inhibit

introduction

resulting in pRB hypophosphorylation and cell cycle arrest (Harper et ai, 1993). RB is a négative regulator of the transcription factor E2F, wich is required for expression of S-phase-specific genes. RB, indeed, acts downstream of p53 since loss of RB fonction can bypass p53-induced Cl arrest (SIebos et a!., 1994). Studies on p21 knockout mice showed that p21 - /- embryonic fibroblasts are defective in their ability to arrest in Cl in response to diverse stresses (Deng et al., 1995). Surprisingly, it was shown that p21 acts as a négative regulator of p53 stability in different cell types and that p53 régulation by p21 may provide a négative regulatory loop that limits p53 induction (Broude et al., 2007). More recently, our laboratory identified a new direct p53 target gene Ptprv/ESP (Doumont et aL, 2005).

Ptprv is a key mediator of p53-induced cell cycle arrest and loss of Ptprv enhances the formation of epidermal papillomas after exposure to Chemical carcinogens, suggesting that Ptprv acts to suppress tumor formation In vivo.

The biochemical pathways involved in the DNA damage-induced G2 arrest are thought to involve signaling cascades that converge to inhibit the activation of Cdc2 (Taylor and Stark, 2001). Cdc2 is inhibited simultaneousiy by three p53 target genes, Gadd45 (Zhan et a!., 1998), p2! and 14-3-3 (Hermeking et a!., 1997). The repression of the Cydin B1 and Cdc2 qenes by p53 aiso contributes to G2 cell cycle checkpoint (Taylor and Stark, 2001).

b} p53 and apoptosis

One of the most extensively studied p53 activities is its ability to induce apoptosis (figure3). The main reason could be that this activity has been best linked to the ability of p53 to function as a tumor suppressor.

p53-dependent apoptosis contributes to chemotherapy-induced cell death and inactivation of p53 leads to treatment-resistant tumors. Thus, the p53 status is a key déterminant for tumor response to therapy (Lowe et a!., 1 993).

p53 induces expression of a wide array of death effectors.

DNA damage can induce transcriptional up-regulation of some death receptors such as PAS, KILLER/DR5 (Takimoto and El-Deiry, 2000; Wu et al., 1997) and p53RDLl (Tanikawa et al., 2003), through p53-dependent and independent mechanisms (El-Deiry, 2001). p53 régulâtes the transcription of proapoptotic Bcl2 family members. These include the multidomain Bcl-2 family member Bax (Miyashita et a!., 1994) as well as BH3 proteins such as Bid (Sax et a!., 2002), Puma (Nakano and Vousden, 2001), and Noxa (Oda et a!., 2000). ^ax-deficient MEFs are résistant to oncogene-induced apoptosis.

5

leading to increased transformation in and tumorigenesis in vivo (y\x\ et ai., 1997). Puma was recently identified as an essential mediator of p53- dependent and indépendant apoptosis in vivo and knockout of Puma récapitulâtes some of the apoptotic deficiency observed in p53 knockout mice (Jeffers et ai., 2003). Puma may function to release cytoplasmic p53 inhibitory interactions with the anti-apoptotic BH3-domain proteins, allowing p53 to function in a transcriptionally indépendant manner (Vousden, 2005).

However, an interesting paradox is that Puma-KO mice do not show enhanced tumour development. p53 aiso transcriptionaly up-regulates the cytoplasmic protein PIDD and the /V6'genes (Polyak et al., 1997). APAF-1, another component of the mitochondrial intrinsic apoptotic pathway, was aIso described as a p53-target (Moroni et ai., 2001).

In addition to the transactivation function, p53 aiso has transrepression capabilities that may contribute to apoptosis. Indeed p53 was found to repress the promoter of Bcl-2, Bcl-xl and survivin (Miyashita et

ai., 1995; Hoffman et ai., 2002). p53 can aiso induce apoptosis trough short- circuiting cell survival pathways and aiso the endoplasmic réticulum pathway (Bourdon et ai., 2002).

Finally it has been shown that in certain cell types, p53 induces apoptosis independent of its effect on transcription (Chipuk et a\., 2003;

Haupt et ai., 1995). For example, p53 interacts with the proapoptotic

mitochondrial protein Bak leading to the release of the cytochrome-c (Leu et

ai., 2004).

introduction

ç} Choosina between Hfe and death

One of the major questions and areas of intense investigation, in part owing to its relevance to successfui application of cancer therapy, is how a cell makes decision to either undergo apoptosis versus induction of a réversible growth arrest or senescence upon p53 activation (El-Deiry, 2003;

Vousden and Lu, 2002; Vousden and Woude, 2000). Much of the choice is not in the hand of p53 itself, but rather it is determined by cooperating or ameliorating intracellular and extracellular signais, which dictate whether p53 activation will lead to cell cycle arrest or apoptosis. For example, p53 could hâve a variable affinity for the binding sites in the different promoters it régulâtes. The degree to which p53 accumulâtes in a given cell could therefore directiy affect the p53 transcriptional program and consequently the cellular fate (Wu and El-Deiry, 1996). The cell type is aiso likely to influence this choice by keeping different sets of p53-responsive genes in active chromatin, thereby predetermining a spécifie pattern of transactivation and transrepression that dictâtes the final outcome. In addition, several studies hâve implicated covalent modification of p53 in playing a critical rôle in its target gene preference. Indeed, the dual-specificity tyrosine- phosphorylation-regulated kinase 2, DYRK2, normally résides in the cytoplasm. However, upon exposure of cells to severe genotoxic stress it translocates to the nucléus and phosphorylâtes p53 on Ser46 to elicit a proapoptotic transcription program (Taira et al., 2007). p53 binding partners might aIso selectively modulate expression of sets of p53-target genes. For example, as mentioned earlier, ASPP proteins enhance p53 binding to the Æ4X promoter, but not other targets such as p21 (Samuels-Lev et ai, 2001).

Another tumor suppressor protein, BRCAl was found to selectively direct p53 to transactivate target genes involved in cell cycle arrest and DNA repair (MacLachlan et a!., 2002). Recently, the zinc-finger protein Hzf was shown to directiy bind to p53's DNA binding domain, resulting in preferential transactivation of p53 target genes that médiate cell-cycle arrest over proapoptotic target genes (Das et aL, 2007). The nucleo-cytoplasmic transport factor hCAS/CSEl L was shown to be a p53 cofactor, which binds to the selected p53 target genes and coopérâtes with p53 to activate transcription of these genes by altering histone modification, specifically

histone H3 lysine 27 (Tanaka et a!., 2007).

7

illustrated here. In response to nutrient stress, p53 can become activated by AMP kinase (AMPK), promoting cell survival through an activation of the cyclin dépendent kinase inhibitor p21. Other fonctions of p53 include regulating respiration, through the action of SC02, or in decreasing the levels of reactive oxygen species (ROS),

introduction

^ New functions of p53

Recent results hâve indicated a rôle for p53 in determining the response of cells to nutrient stress and in regulating pathways of glucose usage and energy metabolism Oones et a!., 2005). More surprisingly, rôles for p53 in controlling different metabolic pathways under apparently normal conditions has aiso been recently described (Stambolsky et ai, 2006). This allows p53 to enhance oxidative phosphorylation. Conversely, loss of p53 activity in cells results in a réduction in oxygen consumption. The résultant defect in respiration would affect tumour cells which are defective in some way for p53 activity. A number of studies hâve aIso shown a survival fonction for p53 in lowering intracellular ROS (Reactive Oxygen Species) levels, involving the activity of p53-inducible genes such as T/CAR, sestrins, ALDH4 and others (Budanov et ai, 2004; Yoon et ai, 2004; Johnson et al., 1996) (figure4).

p53 has aiso a rôle in embryonic development. Indeed, female p53 null mice show neural tube-closure defects. Interestingly, p53 has a clear anti- teratogenic fonction (Norimura et a!., 1996; Nicol et a!., 1995).

Finally, the concept that p53 has a darker side becomes more and more obvious. indeed despite its help in protecting against cancer, p53 is aiso responsible for the classic symptoms of radiation sickness, side effects of cancer therapy, and undesirable aspects of ageing. Indeed, a slight constitutive hyperactivation of p53 results in an alarming premature-ageing phenotype in mice (Maier et al., 2004; Tyner et al, 2002). A contribution of p53 to cellular senescence and the limitation of the proliférative capacity of stem cells hâve been proposed (Sharpless and DePinho, 2004). It is worth mentioning, however, that the ageing phenotype associated with increased in vivo activity is still controversial. Mice containing an extra copy of the p53 gene. Super p53 mice (Garcia-Cao et a!., 2002), or with reduced expression of Mdm2 (a négative regulator of p53 expression level and activity), showed the expected protection from tumour development but no decrease in normal lifespan (Mendrysa et a!., 2006).

8

suppression. The absence of p53 immediately after DNA damage protects animais from radiation sickness, but does not prevent repair of the DNA damage. Persistent absence of p53 results in tumour development, as expected. However, even transient restoration of p53 activity after the resolution of the initial DNA damage can inhibit tumour development without the deleterious responses, such as widespread apoptosis in lymphoid organs and intestinal epithelium, that occur following the irradiation of mice with fully active p53. These results suggest that the induction of p53 in response to signais that persist beyond DNA damage, such as activated oncogenes, is the key to tumour suppression.

(Vousden and Lane, 2007).

FigureG Régulation of life and death by p53 p53 fonctions in the response to both the constitutive stress that is encountered during normal growth and development, and to the acute stress signais that would be associated with oncogenic progression and other types of trauma. in this model, p53 responds to conditions of iow or constitutive stress to play an important part in decreasing oxidative damage, and provides repair fonctions to mend Iow levels of DNA damage. These activities of p53 contribute to the survival and heaith of the cell as well as to the prévention of the acquisition of tumorigenic mutations, and might contribute to overaii iongevity and normal development. By contrast, acute stress that results in a more robust induction of p53 ieads to the activation of apoptotic celi death and thereby the élimination of the damaged cell. This fonction removes ceiis that hâve acquired

introduction

C. Activation of p53 and the upstream program

It is important to make a distinction between p53 tumor suppressor activity, the mechanisms of which remain spéculative, and the p53 responses to DNA damage, which are well characterized. Because critical steps in tumorigenesis involve genomic fixation of DNA damage-induced mutations, it seems reasonable to assume that DNA damage signaling to p53 would activate p53 tumor suppressor activity. However, this has not been demonstrated. A recent study using a mouse model in which p53 can be switched on and off has indicated that the response to DNA damage might not be responsible for tumour suppression (Christophorou et aL, 2006) (figures). The studie shows that p53 becomes important only after the bulk of the damage has been resolved, and conclude that the key tumour- suppressive function of p53 is to respond to oncogene activation that occurs as a conséquence of the original genotoxic stress. In keeping with this, ARF, which is necessary for oncogene-induced, but not DNA-damage-induced, activation of p53, is responsible for almost ail the tumour-suppressor activity of p53 (Efeyan et a!., 2006). However, other equally compelling studies indicate that genotoxic stress is the key signal to activate p53 and tumour suppression in pre-cancerous lésions, and that DNA damage can be induced by the activation of several oncogenes in an ARF-independent manner (Bartkova et a!., 2005; Corgoulis et al., 2005).

K Activation by DNA damage

DNA damage refers to alterations in the Chemical bonds of constituent nucléotides, resulting in aberrant or mismatched base pairs, cross-linked bases, or single- and double-strand breaks in the phosphodiester DNA backbone. DNA damage can be induced by genotoxic Chemicals, ultraviolet radiation (UV), shortened telomeres, reactive oxygen species and aiso by oncogenic alterations in cancer cells, such as overexpressed/amplified c-myc (Vafa et aL, 2002). A model was proposed (figure6) in which p53 can hâve two important , but fundamentally, opposing rôles in response to stress. The low levels of DNA damage that are encountered during normal life are dealt with, through p53, by lowering ROS levels (and so reducing damage) and by promoting the survival of the slightiy damaged cell to allow repair. In response to more severe, sustained stress, p53 switches from promoting

9

pink, méthylation (M) in blue and sumoylation (SU) in brown. Proteins responsible for these modifications are shown in matching colours. b.Targeted mutations at the mouse p53 locus Mouse p53 shares a strong homology with human p53, but a few différences can be noted, including: mouse p53 is comprised of 390 amino acids; the N-terminal part of mouse p53 is longer by 3 residues, so that the numbering is higher in the murine transactivation domain (TAD) than in the human TAD; the p53 proline-rich domain (PRD) is loosely conserved in évolution (the murine PRD is shorter, and contains 2 PXXPs motifs and 2 putative PINl sites instead of 5 PXXPs and 1 PINl site in the human PRD): in the DNA-binding domain and the C-terminal part of the protein, numbering is lower by 3 amino acids in murine compared with human (mouse serine 389 is functionally équivalent to human serine 392); the C-terminal regulatory domain (CTD) of mouse p53 contains 7 lysines, instead of the 6 in human p53. Residues that are subject to stress-induced modifications and that hâve been targeted at the mouse p53 locus are shown. Below the protein are shown other targeted mutations which provided valuable information on p53 fonction, but did not precisely target residues modified by stress. For several point mutations, abbreviated names are mentioned (for example, QS instead of L25Q,W26S). AMPK, adenosine monophosphate-activated protein kinase; ATM, ataxia telangectasia mutated; ATR, ataxia telangectasia and Rad3-related protein; AurK, Aurora kinase A; CAK, CDK- activating kinase; CDK, cyclin-dependent kinase; CHK, checkpoint kinase; CK, casein kinase; CSNK, cop-9 signalosome associated kinase complex; DNAPK, DNA-dependent protein kinase; ERK, extracellular signal-regulated kinase; CSK3, glycogen synthase kinase 3; HIPK2, homeodomain- interacting protein kinase 2; JNK, c-Jun NH2-terminal kinase; MAPKAPK2, mitogen-activated protein kinase-activated protein kinase 2; p38, p38 kinase; PCAF, p300/CBP associated factor; PKC, protein kinase C; PKR, double stranded RNA-activated kinase; PLK3, pol-like kinase 3; RSK2, ribosomal S6

introduction

survival and repair to the induction of apoptosis (Bensaad and Vousden, 2005).

^ N-termina! modifications of p53 (1) In vitro

p53 could be a direct sensor of DNA damage by binding directiy to DNA lésions or to DNA damage repair products. In vitro, p53 can directiy binds to irradiated DNA, DNA mismatch, DNA ends (Lee et al., 1995; Reed et ai, 1995; Bakalkin et a!., 1994). ). However, this is still controversial. It is more accepted that intermediate DNA damage sensors transmit the damaged signal to p53. lonizing radiation (iR) induces both p53 protein stabilization and p53 sequence-specific DNA binding activity, both of which are déficient in ataxia telangiectasia cells (Kastan et aL, 1992). Thus arose the concept that iR could elevate p53 protein to bind DNA and induce Cl/S cell cycle arrest, and that the gene defective in AT cells, ATM, was required for such signaling. Indeed ATM and Rad-3 related protein ATR are responsible for iR/UV radiation-induced p53 protein induction (Tibbetts et a!., 1999). The mechanism by which ATM and ATR signal to p53 is dépendent on their kinase activity (Canman et al., 1998). The current model is that a DNA damage-induced nuclear chromatin change, induced by iR, activâtes ATM kinase activity (Bakkenist and Kastan, 2003). In contrast, processes that generate persistent single-stranded DNA due to UV exposure signal to recruit ATR or to activate ATR kinase activity (Unsal-Kacmaz and Sancar, 2004). An alternative model proposes that DNA damage-induced nucleoli disruption, rather than DNA damage per se, is the signal for p53 protein stabilization (Rubbi and Milner, 2003).

DNA damage-induced accumulation of p53 is a rapid response that does not rely on changes in p53 mRNA expression (Fu and Benchimol, 1 997) but much more on increases in p53 translation and/or protein stabilization.

One of the mechanisms suggested for p53 stabilization involves the DNA- dependent protein kinase DNA-PK (figure7a), which phosphorylâtes p53 on Serl 5 (Lees-Miller et a!., 1992). Indeed IR was shown to activate DNA-PK, causing p53 Serl 5 phophorylation in v/Vo (Siliciano étal, 1997; Shieh et al, 2000). Serl 5 phosphorylation correlates with p53 protein induction possibly because this modification weakens p53 binding to MDM2, an E3 ubiquitin ligase that targets p53 for dégradation. ATM and ATR can aiso phosphorylate

10

p53 on Serl 5 and cause p53 protein accumulation. There is evidence that Serl 5 phosphorylation can enhance p53 transcriptional activity because phosphorylated p53 binds with a higher affinity p300, an essential p53 transcriptional co-activator (Turenne et aL, 2001; Dumaz and Meek, 1999).

IR causes ATM-dependent activation of the Chk2 kinase (Ahn et a!., 2000;

Matsuoka et aL, 2000; Melchionna et ai, 2000), and UV causes ATR- dependent activation of Chkl kinase (Guo et a!., 2000; Liu et al., 1999).

Chkl/Chk2 kinases cause, in turn, Ser20 phosphorylation of p53 in vivo leading to its induction and stabilization (Hirao et ai., 2000; Shieh et ai., 2000). The phosphorylation of ThrlS, Ser46 as well as the phosphorylation of Mdm2 on Ser395 are others mechanisms by which p53 is induced and stabilized in response to DNA-damage. A new model has therefore emerged emphasizing the importance of N-terminal phosphorylation in p53 activation (Sakaguchi et ai., 1998). Phosphorylation of serines 15 and 20, and threonine 18 is proposed to induce a conformational change that prevents MDM2 from interacting with p53. This results in increased binding of p300/CBP, and presumably, the basal transcription machinery. As p300/CBP and the transcription machinery bind p53 in a région that partially overlaps that bound by MDM2, co-activators recruitment would compete for MDM2 binding (De Guzman et ai, 2000; Lu and Levine, 1995; Thut et ai., 1995;

Xiao et ai., 1994). Preventing MDM2 binding would increase p53 transcriptional output by increasing p53 abundance. p300/CBP binding to p53 may aiso lead to acétylation of p53 C-terminal lysines; this could stabilize p53 by preventing MDM2-mediated ubiquitination of the same residues (Nakamura et ai, 2000; Rodriguez et ai, 2000; Gu and Roeder, 1997).

p53 C-terminal acétylation has aIso been proposed to increase its

ability to associate with chromatin, and to enable recruitment of another

histone acetyl transferase, PCAF, that induces histone acétylation beyond that

introduction

(2) In vivo

However, the regulatory importance of p53 N-terminus and C- terminus is reevaluated by making knock-in (Kl) mutations in the mouse and analyzing their biological conséquences in vivo (figure7b). A mouse p53 Kl mutant, in which leucine 25 and tryptophan 26 (two residues that are crucial for transactivation in vitro) were mutated to glutamine (p53QS), still binds p53’s RE even though these substitutions prevent p53 acétylation by p300/CBP and transcriptional fonction. This indicates that p300/CBP- mediated p53 acétylation is not required for p53 to bind to RE of its target genes in (Chao et ai., 2000; Jimenez et ai., 2000). Mouse serines 18 and 23 (équivalent to human serines 15 and 20) hâve individually been mutated to alanine, S18A and S23A. p53S’8A jn mouse embryonic stem cells, differentiated ES cells, or MEFs was présent at nearly normal levels in unstressed cells, and was induced almost to comparable levels as wild type p53 in response to UV or IR (SIuss et ai., 2004; Chao et ai., 2003; Chao et ai., 2000). p53si8A vvas as effective as tumor suppression as wild type p53.

Together these in vivo data indicate that serine 1 8 phosphorylation in mice is not essential for p53 activity. p53^23A niice hâve aiso been generated, and exhibited nearly wild type patterns of p53 activation and induction of apoptosis in ES cells or in thymocytes (Wu et ai., 2002). However, the targeted double mutation was created, p53si8-23A^ and has modest defects in cell-cycle control and the induction of cellular senescence but is unable to induce apoptosis efficiently (Chao et ai., 2006) indicating a functional synergy between those two serines for at least some p53 functions. Together these observations indicate that phosphorylation is not essential for p53 activation but represent a fine-tuning mechanism that control subsets of p53 activities. In the same Unes, nutlins, compounds that block Mdm2-p53 interaction (Vassilev et ai., 2004) are able to activate p53 without measurable N-terminal modifications.

12

b} C-termina! modifications of p53 (1) In vitro

DNA damage can aiso signal directiy to p53 to increase p53 sequence- specific DNA binding activity by virtue of an array of post-translational modifications affecting the carboxy-terminal 30 amino acid regulatory domain. The modifications include phosphorylation/dephosphorylation (Waterman et al., 1998; Takenaka et a!., 1 995), acetylation/deacetylation (Gu and Roeder, 1997), protein binding Qayaraman et al, 1997; Hupp et al,

1995), and sumoylation (Gostissa étal, 1999; Rodriguez étal, 1999).

Although some of these modifications are not required for p53 protein accumulation, they do affect p53 sequence-specific DNA binding activity in vitro. However, this is controversial in vivo. DNA damage activâtes p53 DNA binding activity by virtue of an increased p53 protein level, but the actual spécifie DNA binding activity is not enhanced by DNA damage. p53, indeed, binds similarly after accounting for différences in total p53 protein, to some genomic binding sites in vivo in the absence of a DNA damage signal (Kaeser and Iggo, 2002; Espinosa and Emerson, 2003).

UV induces Ser389 phosphorylation (Kapoor and Lozano, 1998; Lu et al, 1998). Phosphorylation of this serine results in stabilizing the p53 tetramer and enhancing sequence spécifie DNA-binding ability of p53 in vitro (Hupp and Lane, 1994; Sakaguchi et al, 1998). Other studies aIso support the importance of phosphorylation at this serine in cell-cycle arrest and apoptosis functions of p53 (Hao et al, 1 996; Achison and Hupp, 2003).

(2) in vivo

To enable in vivo functional analyses of this Ser389 phosphorylation site, mice were generated containing a serine to alanine substitution at serine 389 in the endogenous p53 locus (S389A) (Bruins et al, 2004). Homozygous mutant mice are viable and normal with no increase in spontaneous tumor development. The alteration results in moderate réduction of protein stability and DNA binding activity of p53 after UV irradiation. Expression of some p53 downstream target genes after UV treatment is aiso slightiy impaired in p53S389A

EF

compared to those with wild-type p53 (Bruins et ai., 2004;

Iwakuma et ai., 2004). Apoptotic response of p53S3S9A

is partially

impaired after UV irradiation, but not after IR. Importantly, the mutant mice

introduction

are more susceptible to UV-induced skin tumor development compared to wild-type mice, suggesting that the phosphorylation of serine 389 plays a

rôle in UV induced apoptosis and tumor suppression.

Moreover, mutant mice that express p53 in which the C-terminal 7 or 6 lysines were mutated to arginine, p537KR or p53^'^'^, do not exhibit obvious phenotypes (Krummel et aL, 2005; Feng et ai, 2005). Thus, this suggests the existence of additional or alternative mechanisms that participate in the régulation of p53 DNA binding activity. For example, DNA damage activâtes c-Jun, which binds to the Mdm2 promoter to co-activate transcription with p53 and thus indirectely régulâtes p53 stability via Mdm2 induction (Phelps et aL, 2003; Ries et aL, 2000). Moreover, DNA damage-induced alteration of the chromatin structure around p53 DNA binding sites could affect the DNA torsion and bending, thereby regulating p53 affinity for these sites (Espinosa

etaL, 2003).

Together these in vitro and in vivo data indicate that the balance between phosphorylation/dephosphorylation and acetylation/deacetylation of any of the characterized p53 modification sites, only contributes moderately to p53 régulation of function.

14

D.

53*

régulation bv subcellular localization and interactions p53 must be tightiy kept in check in absence of stress signais as it bas the potential to either kill a cell or to prevent cells from dividing. However, it bas become évident tbat despite tbe many levels of négative régulation tbat are in place to restrain p53, tbe everyday rigours of normal mammalian life can more systemically induce low levels of p53 activity. And, recent studies bave revealed an unappreciated importance of p53 under conditions of apparently normal growtb and development. Induction of p53 tbrougb tbese mecbanisms seems to bave a rôle in responses including promoting survival as well as cell deatb (Lassus et aL, 1996).

Régulation of p53 subcellular localization

p53 is a very unstable protein tbat is typically nuclear and présent at very low levels. p53 appears to sbuttle between nucléus and cytoplasm during cell cycle (Moll et a!., 1996). p53 contains nuclear localization and nuclear export signal, and its subcellular localization reflects a balance between tbe rates of import and export (Henderson and Eleftberiou, 2000).

Civen tbe importance of nuclear functions of p53 in tumor suppression, it is

not surprising tbat some tumors bave evolved mecbanisms to accumulate

p53 in tbe cytoplasm to inactivate it (Moll et aL, 1995). To reevaluate tbe

bypotbesis tbat certain cytoplasmic factors may directiy affect tbe subcellular

localization of p53, Nikolaev and colleagues bave biocbemically purified

p53-containing protein complexes from tbe cytoplasm of unstressed cells

and identified a protein. Parc (p53-associated, Parkin-like cytoplasmic

protein), as a key component of tbese complexes. Tbis discovery reveals an

important regulatory patbway tbat Controls tbe subcellular localization of p53

and its subséquent biological function. Indeed, Parc, is overproduced in

neuroblastomas witb cytoplamsic p53, and reducing Parc in tbese cells by

Figures Charaaerization of inactivation in vivo 'm mice Mdm2W\\à type and mutant embryos at E3.5 stained by TUNNEL (right panel) and viable p53-/~ Mdm2-/~ mice (left panel). (Adapted from Montes de Oca Luna et a!., 1995; Chavez-Reyes et ai., 2003).

2^ The complex régulation of p53 bv Mdm2 and Mdm4 a)_ Mdm2

The Mdm2 gene was first identified as a gene amplified in a spontaneousiy transformed mouse 3T3 cell line (Cahilly-Snyder et al., 1987).

The importance of the p53-Mdm2 relationship is best illustrated by the mouse knockout studies. Mice that lack Mdm2 expression die very early during embryogenesis at E3.5 of gestation, but concomitant loss of p53 completely rescues this phenotype Oones et ai, 1995; Montes de Oca et ai, 1995) (figures). p53 is active during embryogenesis and the lack of Mdm2 function results mainly in p53-induced apoptosis (de Rozieres et ai, 2000).

Conditional inactivation of Mdm2 in cardiomyocytes (Crier et aL, 2006), neuronal progenitor cells (Xiong et aL, 2006) and smooth muscle cells (SMCs) of the gastrointestinal (Cl) tract (Boesten et aL, 2006) leads to p53 dépendent cell death. Finally, conditional expression of p53 in neuronal progenitor cells and post-mitotic neurons of Mdm2 null mice leads to dramatic p53 activation and cell death (Francoz et aL, 2006). Mdm2 thus appears to be an essential p53 antagonist in the developing embryo and in mature differentiated cells. An interesting hypomorphic Mdm2 allele was generated (Mendrysa et aL, 2003). This allele in combination with an Afd/772-null allele results in approximately 30% of the total levels of Mdm2 and the mice exhibited réduction in body weight and mild anémia. Since these phenotypes were eliminated on a p53 null background, these data indicate that the hematopoietic System in the mouse is the most sensitive to small decreases in Mdm2 levels and thus to small increases in p53 activity.

b} Mdm4

Mdm4 was identified in a screen for p53 binding proteins (Shvarts et aL, 1996) and was named Mdm4 because of its high structural similarity to Mdm2 (figure9). To date Mdm4 is the only identified family member of Mdm2. Several domains in Mdm2 are highiy conserved in Mdm4, like the p53 binding domain (58/71% amino acid identity/similarity), the Zn-finger (47/50%) and the RINC finger (53/72%). The overall protein identity is 34%.

The genomic organisation of the Mdm4 and Mdm2 genes aiso reveals a high

structural homology. A striking différence between the genes, with a possibly

important functional conséquence, is that in the Mdm2 gene the intron

Figure9 Comparison of the MDM2 and MDM4 (aiso known as MOMX) primary structures Several functional domains are highiighted. The p53-binding domain, zinc (Zn) finger and RING finger [containing the nucleolar location signal (NoLS)] are conserved. The percentage identity shared between these domains is indicated. Although both MDM2 and MDM4 contain an acidic domain, no significant conservation of amino acid sequence is found, and the acidic domain of MDM4 is smaller than that of MDM2. Part of the MDM4 amino acid sequence (338-407) is shown to indicate the functional domains and modification sites. Serines (S) indicated in red are validated phosphorylation sites, whereas lysines (K) indicated in blue are targets for SUMO conjugation. DVPD, caspase-3 cleavage site; NES, nuclear export signal: NLS, nuclear localization signal. (Adapted from Marine et al., 2007).

Figure! 0 Characterization of Mdm4 inactivation in vivo in mice a. Mdm4 wiid type and mutant embryos at El 1.5 (left panel) and viable p53-/~ Mdm4-/~ mice (right panel), b. IHC for Caspase-3* (in the neuroepithelium) and BrdU (in the fêtai liver) of Wild type and Mdm4 mutant embryos. (Adapted from Migliorini et ai., 2002).

between the first and the second exon contains one p53 RE, giving rise to a négative feedback loop. The intron is absent in Mdm4, which might explain why Mdm4 expression is not induced after p53 activation. The unresponsiveness to p53 might be one of the reasons why Mdm4 can not substitute for Mdm2 during embryonic development. But Mdm4 does hâve a critical rôle in regulating p53 in the developing embryo. Mice null for Mdm4 die between 7.5 days post coitum {dpc) and 11.5 dpc but concomitant loss of p53 completely rescues this lethality (Finch et al., 2002; Migliorini et a!., 2002; Parant et a!., 2002) (figurelO), as previousiy described for the Mdm2 knockout mice. However the timing and mechanism are different leading to the possibility that these two p53 inhibitors fonction in a temporal and tissue spécifie manner. Mice null for Mdm2 die by induction of apoptosis at E3.5, while mice lacking Mdm4 cease prolifération at E9.5 due to important defect in cell cycle in the hematopoietic System and severe apoptosis in the neuroephitelium (figurelOc). The fact that Mdm2 is not able to compensate for the loss of Mdm4 suggests non-overlapping pathways of regulating p53.

When conditionally inactivated in different tissues, Mdm4 déplétion leads to phenotype ranging from null to minor (Xiong et al, 2006; Boesten et al, 2006; Francoz et al, 2006).

ç} The collaboration of Mdm2 and Mdm4 in p53’s régulation (1 ) Régulation of the p53 transcriptiona! activity

Mdm2 inhibits p53 fonction in at least two ways, though it apparently needs Mdm4 to do so with optimal efficiency (Gu et al, 2002; Migliorini et al, 2002). First, similar N-terminal régions of Mdm2 and Mdm4 interact with the hydrophobie side chains of an amphipathic alpha-helix in the p53 N- terminal transactivation domain (Bottger et al, 1999; Russie et al, 1996;

Chen et al, 1993). Consequently, Mdm2 and Mdm4, by binding to the transactivation domain, could inhibit transactivation by preventing the basal transcription machinery from binding and/or by preventing p53 acétylation by p300 and CBP (Gu et al, 1997; Shvarts étal, 1996; Lu and Levine, 1995;

Thut étal, 1997; Xiao et al, 1994; Momand et al, 1992). Mdm2 may aiso

inhibit p53 transactivation by recruiting co-repressors such as CtBP2

(Mirnezami et al, 2003). Moreover, Mdm2 was aIso reported to promote

introduction

of historiés surrounding the p53’s RE resulting in transcriptional repression (Minsky et aL, 2004). However, different in vivo studies demonstrated that stabilized p53 in cells lacking Mdm2\s not transcriptionally more potent than in cells expressing Mdm2 (Francoz et a!., 2006; Toledo et a!., 2006).

(2) Régulation of the p53 p rote in expression ievei and stabiHty

The second mechanism is keeping p53 at low levels mainly through the combined actions of Mdm2 and Mdm4, that can associate as homo- or heterodimers through their Ring domains. p53 was initially shown to be targeted for dégradation by the oncogenic papilloma virus E6 protein, which binds to p53 and recruits a cellular ubiquitin ligase (E6-AP) to médiate p53 ubiquitination (Scheffner et ai., 1993). Mdm2 is a ring finger E3 ubiquitin ligase that médiates the ubiquitination and dégradation of p53 (Lai et aL, 2001; Fang et aL, 2000; Honda and Yasuda, 1997; Fuchs et aL, 1998; Haupt et aL, 1997; Honda et ai., 1999). The central région of Mdm2 containing an acidic domain is aiso needed for p53 dégradation (Kawai et aL, 2003;

Meulmeester et aL, 2003; Argentini et a!., 2001). Mdm2 mediated ubiquitination mainly occurs on C-terminal lysines, and transfection analyses show that p53 mutants in which ail these lysines were changed to arginine are stable, active and nuclear (Nakamura et ai., 2000; Rodriguez et aL, 2000).

Other studies show that p53 mutations that prevent Mdm2 association aIso generate stable, nuclear p53 Oimenez et aL, 2000; Lin et aL, 1994). However, there is debate about whether Mdm2 médiates the mono- (Lai et aL, 2001), or poly-ubiquitination of p53 (Li et ai., 2003). A report showing that low levels of Mdm2 médiate p53 mono-ubiquitination, while higher levels induce poly-ubiquitination (Li et aL, 2003) raising the possibility that the Mdm2 level régulâtes a ubiquitination switch. p53 monoubiquitination was reported to lead to nuclear export of p53, while chains containing at least four ubiquitins are required for proteosomal dégradation (Thrower et a!., 2000).

The notion that Mdm2 could induce mono-ubiquitination for p53 export or polyubiquitination for dégradation raises the question of whether p53 must be exported to be degraded in the cytoplasm. Consistent with diverse studies, nuclear and cytoplamsic p53 can be ubiquitinated and degraded (Stommel and Whal, 1999; Joseph et ai., 2003; Shirangi et aL, 2002; Lohrum et a!., 2000; Xirodimas et ai., 2001; Geyer et a!., 2000). The précisé mechanisms by which Mdm2 leads to p53 dégradation remain to be defined.

18

How polyubiquitination is achieved in normal cells with unstable p53 and low levais of Mdm2 if Mdm2 only induces mono-ubiquitination under such conditions (U et ai, 2003; Fang et a!., 2000). One possibility is that Mdm2 could favour p53 association with another ubiquitin ligase, an “E4”, that adds polyubiquitin chains to the lysines previousiy mono-ubiquitinated by Mdm2.

One potential candidate for a p53 E4 polyubiquitin ligase is the histone acetyl transferase p300 (Grossman et al., 2003). This seems surprising as p300 binds the same N-terminal région of p53 as Mdm2 and acetylates the same C-terminal lysines that Mdm2 monoubiquitinates. However, Mdm2, p300 and p53 form ternary complexes, and Mdm2 mutants that cannot bind p300 can médiate p53 ubiquitination but not dégradation (Zhu et aL, 2001; Kobet et a!., 2000; Grossman et a!., 1998).

^ The complementarv rôle of Mdm2 and Mdm4 in o53’s régulation (1) In vitro

The précisé mechanisms by which Mdm2 and Mdm4 collaborate to regulate p53 are important to define, as both proteins are clearly required for optimal inactivation of p53. Mdm2 médiates the ubiquitination and dégradation of itself and Mdm4 (Kawai et aL, 2003; Pan and Chen, 2003;

Tanimura et aL, 1999). On the other hand, Mdm4 cannot induce Mdm2 or p53 dégradation as it is not an E3 ubiquitin ligase (Stad et a!., 2001). Several studies employing transfection and overexpression indicated that Mdm4 can inhibit p53 dégradation (Jackson and Berberich, 2000; Sharp et aL, 1999;

Migliorini et aL, 2002). This could be explained by the artificial disturbance of the Mdm4 relative to Mdm2 ratio. Mdm4 when aberrantly expressed competes with Mdm2 for p53 binding and therefore impairs Mdm2-mediated p53 dégradation. In contrast when expressed at physiological levels, Mdm4 increases significantly the ability of Mdm2 to dégradé p53 (Gu et aL, 2002).

As Mdm2 and Mdm4 interact with each other, the Mdm2/Mdm4

heterodimers may be more potent in inhibiting p53 than the homodimers

(Linares et aL, 2003). Based on ail these transfection data, the mutual

dependency model was proposed in which Mdm4 stabilizes Mdm2, and

Mdm2 is required for efficient nuclear import of Mdm4 (Migliorini et a!.,

2002, Gu étal., 2002).

Figurel 1 A model for cooperative control of the p53 pathway by Mdm2 and Mdm4 (aiso known as Mdmx). In the absence of stress signais, the primary fonction of Mdm2 is to maintain p53 at low levels, whereas Mdm4 contributes to the overall inhibition of p53 independently of Mdm2. Mdm4 inhibits p53 transcriptional activity, whereas the contribution of Mdm2 to the régulation of p53 transcriptional activity per se is still unclear and a matter of debate. The rôle of Cop-1 and Pirh2 in the régulation of p53 levels and activity in vivo is unclear, but recent data suggest that, if they participate in the régulation of this pathway, they can only do so in an Mdm2-dependent manner.

(Marine et al., 2007).