|USFDC Home | USF Electronic Theses and Dissertations||| RSS|
This item is only available as the following downloads:
xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001681014
007 cr mnu|||uuuuu
008 051212s2005 flu sbm s000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0001114
Modification of mdmx by ubiquitination and sumoylation
h [electronic resource] /
by Yu Pan.
[Tampa, Fla.] :
b University of South Florida,
Thesis (Ph.D.)--University of South Florida, 2005.
Includes bibliographical references.
Text (Electronic thesis) in PDF format.
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
Title from PDF of title page.
Document formatted into pages; contains 157 pages.
ABSTRACT: MDM2 and MDMX are two major negative regulators of tumor suppressor p53. Both MDM2 and MDMX can inactivate p53 and play important roles in mouse embryonic development in a p53 dependent manner. MDM2 possesses ubiquitin E3 ligase activity and mediates self-ubiquitination as well as ubiquitination and degradation of p53 by proteasome. We identify MDMX as another ubiquitin E3 ligase substrate of MDM2. MDM2 promotes the ubiquitination and degradation of MDMX through proteasome pathway. The RING domains of both MDMX and MDM2 are required and sufficient for MDM2-mediated MDMX ubiquitination. ARF overexpression, DNA damage or MDM2 overexpression can all stimulate MDMX ubiquitination and degradation. We present evidence that MDMX is also sumoylated. The sumoylation sites on MDMX are identified. ARF N-terminus is required for stimulating both MDMX ubiquitination and sumoylation. We also demonstrate that MDMX binds to ARF in an MDM2-dependent fashion.
Adviser: Jiandong Chen.
x Cancer Biology
t USF Electronic Theses and Dissertations.
Modification of MDMX by Ubiquitination and Sumoylation by Yu Pan A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy Department of Cancer Biology College of Graduate School University of South Florida Major Professor: Jiandong Chen, Ph.D. Kenneth Wright, Ph.D. Hong-Gang Wang, Ph.D. Srikumar Chellappan, Ph.D. Date of Approval: March 23, 2005 Keywords: post-translational modifica tion, MDM2, ARF, ubiquitin, sumo Copyright 2005 Yu Pan
Dedication To my family: my grandparents, my parents and my parents in law, for their love and support. To my husband, Linan, for everything.
Acknowledgments I would like to sincerely thank my major professor and my mentor, Dr Jiandong Chen. You opened the door of true science for me and led me all the way through this valuable time of my career. I learned from you, not only the skills for science, but also the principles of a successful scientist: di ligence, honesty and modesty. I will always be grateful for your great patience in training a layman like me into a professional. Through out the years, your support and encourag ement guided me through many hardships and for that I will be forever grateful. I could not ask for a better mentor. I would like to thank all my committ ee members, Drs Srikumar Chellappan, Hong-gang Wang, Ken Wright for their guidance valuable suggestion and their word of encouragement. Thanks a lot to Dr Maureen Murphey for agreeing to be the chairperson of my dissertation committee. I would especially like to thank my committee member and our advisor, Dr Ken Wright. I greatly appreciate your kindness, your patience and your endeavor to help every student and make our program thrive. I would also like to thank Dr. Julie Djeu for her perseverance in generating the Cancer Biology program, which I am sure, has pr epared me for the start of a successful career in science. Thank you also to all th e professors who took time out of their busy schedule to share their wealth of knowledge during each class lecture.
I am indebted to Lihong, my friend and my mentor, who was there to support me and help me from the first day I came to the lab. I realize that I learned from her almost every technique that I now know and asked her millions of questions and she has always been supportive and patient with me. I le arned from her how beautifully experiments could be performed. As a friend, she was always caring and warmhearted. She was always there when I needed a friend and led me through the difficult time of the graduate school. I consider myself the luckiest for having her around all these years. I am very grateful to my first major prof essor and my mentor, Dr Melissa Rogers. You brought me to this country of freedom and realized my dream to better myself. Your love for Chinese culture made me feel we lcomed. You encouraged me to discover myself, to find out my capability and to ad apt to this brand new world. Without you, none of those was possible. Thank you so much. I also want to thank my friends who worked with me in the lab. Cindy, your friendship and encouragement made these tough years a lot easier and brighter. Thanks to Daniele and Jenelle, for your friendship a nd constant support. Thanks to Dr Chuangui Wang and Dr Xinghua Hou for being such wonderful colleagues. This dissertation is dedicated to my family. Words cannot express my gratefulness for having such a devoting, loving family, for their constant support and for their confidence in me. I want to say tha nk you to my grandparents, who nurtured me with their unfailing love. I want to say th ank you to my parents, who bestowed me all their love, their support and their blessing. I also want to say thank you to my parents in law, who loved me as their own child. Their confidence in me made me believe that I can do better.
Finally, I want to say thank you to my beloved husband, Linan. Thank you for loving me as I am and bringing out the good from me. Thank you for giving me an eternal home in my heart with which I would never feel lonely and helpless. My gratefulness for your endless love and s upport is beyond any word. I love you.
i TABLE OF CONTENTS LIST OF FIGURES iii LIST OF ABBREVIATIONS vi ABSTRACT vii INTRODUCTION 1 Tumorigenesis 1 p53 2 p53 history 2 p53 function and transcription down stream targets 3 Activation of p53 8 p53 regulation by MDM2 and MDMX 10 p53 posttranslational modification 12 p53 as a cancer therapy target 14 MDM2 16 MDM2 History 16 MDM2 mouse model and clinical relevance 16 MDM2 inhibits p53 transcriptional activity 17 MDM2 is an ubiquitin E3 ligase for multiple proteins 18 MDM2 promotes other posttranslational modi fications of p53 19 MDM2 nuclear export and import 19 MDMX 21 MDMX Protein structure 21 MDMX gene structure 23 MDMX as an oncogene 24 MDMX mouse models 25 MDMX upon DNA damage 26 MDMX-MDM2-p53 27 ARF 30 Ubiquitination a nd proteasome degradation review 32 Sumoylation process 36 Sumoylation functions 37 Sumoylati on is a reversible process 39 MATERIALS AND METHODS 41 Cell lines 41 Plasmids and reagents 41 Western blot 42
ii Immunoprecipitation 43 Transfection 44 Immunofluoresence staining 45 In vivo ubiquitination assay 45 In vitro ubiquitination assay 46 MDM2 PROMOTES UBIQUITINATION AND DEGRADATION OF MDMX 48 Abstract 48 Results 49 MDM2 stimulates MDMX ubiquitination and degradation 49 MDM2 RING domain is sufficient for promoting MDMX ubiquitination 54 Ubiquitination of MDMX requires MDM2 ubiquitin ligase function 58 ARF N terminal domain is requir ed to cooperate with MDM2 62 Regulation of endogenous MDMX level by ARF and MDM2 64 Epitope tagging prevents MDMX degradation by MDM2 70 MDM2 promotes MDMX poly-ubiquitination in vitro 72 Discussion 74 MDMX FORMS TRIMERIC COMPLEX WITH MDM2 AND ARF 79 Abstract 79 Introduction 80 Results 82 MDMX forms trimeric complex with MDM2 and ARF 82 ARF N-terminus is required for binding MDMX 86 ARF does not affect the binding efficiency between MDM2 and MDMX 88 ARF stimulates MDM2-mediated MDMX ubiquitination by 90 stabilizing MDM2 Discussion 92 MODIFICATION OF MDMX BY SUMOYLATION 94 Abstract 94 Introduction 95 Results 98 MDMX is posttranslationally modified by sumo-1 98 Identification of MDMX sumoylation site 101 MDMX sumoylation does not affect its stab ility and regulation by 107 DNA Damage Discussion 114 SCIENTIFIC SIGNIFICANCE 117
iii REFERENCE 120 ABOUT THE AUTHOR END PAGE
iv LIST OF FIGURES Figure 1. Structure and func tions of p53 tumor suppressor 3 Figure 2. MDM2 and MDMX share homology in their protein structures 22 Figure 3 p16 and ARF are two tumor suppressors encode d in the same 31 gene locus but working by distinct mechanisms Figure 4. MDM2 stimulates MDMX polyubiquitination 50 Figure 5. MDM2 and ARF synergistically stimulate the polyubiquitination and 51 degradation of MDMX Figure 6. MDM2 stimulates MDMX degr adation by proteasome pathway 52 Figure 7. MDMX can be ubiquitinated independent of MDM2 53 Figure 8. MDM2 RING domain is required for stimulating MDMX 55 Polyubiquitination Figure 9. MDMX RING domain is requ ired for its polyubiquitination 56 degradation by MDM2 Figure 10. Diagram of MDM2 and MDMX mutants and summary of results 57 Figure 11. Ubiquitination of MDMX requires MDM2 E3 function 59 Figure 12. Binding of MDMX to MDM2 point mutants 61 Figure 13. The ARF N-terminal domain is sufficient to cooperate with 63 MDM2 Figure 14. Overexpression of ARF down regulates MDMX level 65 Figure 15. DNA damage induces MDMX degradation 66 Figure 16. Induction of MDM2 results in MDMX degradation by 67 proteasome
v Figure 17. Gamma irradiation does not reduce the MDMX mRNA level 68 Figure 18. Induction of MDM2 corre lates with MDMX degradation through 70 proteasome pathway Figure 19. MDMX degradation is inhibi ted by C-terminal epitope tags 71 Figure 20. MDM2 promotes MDMX polyubi quitination in vitro 73 Figure 21. MDMX is re gulated by stress signals 77 Figure 22. ARF connects Rb-E2F pathways to p53-MDM2 pathway 81 Figure 23. MDMX forms trimeric complex with MDM2 and ARF 83 Figure 24. MDM2 mediates the bind ing between MDMX and ARF 85 Figure 25. ARF N-terminus is required for bi nding MDMX 87 Figure 26. ARF does not affect the binding efficiency between 89 MDM2 and MDMX Figure 27. ARF stimulates MDMX ubiquitination while stabilizing MDM2 90 Figure 28. ARF stimulates MDMX sumoylation and the sumo-specific protease 99 SENP1 desumoylates MDMX Figure 29. In vitro sumoylation of MDMX 101 Figure 30. The major sumoylation sites of MDMX are contained within 103 the region 250-490 Figure 31. Diagram of MDMX muta nts and summary of results 103 Figure 32. The characterization of MDMX sumoylation sites on 105 serine 254 and serine 379 Figure 33. Desumoylation of MDMX by SENP1 107 Figure 34. Lack of sumoylation does not affect MDMX 109 mono-ubiquitination or polyubiquitination by MDM2 Figure 35. MDMX sumoylation status does not affect its 110 MDM2-mediated degradation
vi Figure 36. MDMX sumoylation status does not affect its 111 DNA damage induced destabilization Figure 37. Sumoylation-defici ent MDMX mutants tr anslocated into the 112 nucleus after ionizing radiation in the same fashion as wild type MDMX Figure 38. MDMX sumoylation status does not affect protein-protein 113 binding between p53, MDM2 and MDMX
vii LIST OF ABREVIATIONS ARF alternative reading frame CBP CREB-binding protein CDK cyclin-dependent kinase DUB de-ubiquitinating protein DSB double strand break FBS fetal bovine serum GFP green florescent protein HDAC histone deacetylase HECT homologous to the E6-AP COOH terminus HFF human foreskin fibroblast MDM2 mouse double minute 2 MEF mouse embryo fibroblast l micro liter NLS nuclear localization signal NES nuclear export signal N O LS nucleolus localization signal PBS phosphate-buffered saline PCR polymerase chain reaction PIAS protein inhibitors of activated STATs PML promyelocytic leukemia protein Rb retinoblastoma protein RT room temperature SENP sentrin-specific protease SUMO small ubiquitin-like molecule
viii Modification of MDMX by Ubiquitination and Sumoylation Yu Pan ABSTRACT MDM2 and MDMX are two major negative regulators of tumor suppressor p53. Both MDM2 and MDMX can inactivate p53 and play important roles in mouse embryonic development in a p53 dependent manner. MDM2 possesses ubiquitin E3 ligase activity and mediates self-ubiquitinati on as well as ubiquitination and degradation of p53 by proteasome. We identify MDMX as another ubiquitin E3 ligase substrate of MDM2. MDM2 promotes the ubiquitination and degradation of MDMX through proteasome pathway. The RING domains of both MDMX and MDM2 are required and sufficient for MDM2-mediated MDMX ubiquitination. ARF overexpression, DNA damage or MDM2 overexpression can all stimulate MDMX ubiquitination and degradation. We present evidence that MD MX is also sumoylated. The sumoylation sites on MDMX are identified. ARF N-termi nus is required for stimulating both MDMX ubiquitination and sumoylation. We also de monstrate that MDMX binds to ARF in an MDM2-dependent fashion.
1 Introduction Tumorigenesis The growth and proliferation of normal unt ransformed cells are the result of a balanced relationship between oncogenes and tu mor suppressors compared to those of the tumor cells. Tumorigenesis is a progressi ve disease caused by the accumulative lesions in the genome, producing either gain of function oncogenes or loss of function of tumor suppressors, or both. The cancer cells are de fective in their regulatory pathways, which lead to the abnormal cell growth and prolifer ation. A vast majority of cancer cells display a single or a combination of six trai ts: evasion from apoptosis, self-reliance in growth signals, insensitivity to anti-gro wth signals, tissue invasion and metastasis, unlimited replication potential and incessant angiogenesis (Fri end, Bernards et al. 1986; Hanahan and Weinberg 2000). While the unc urbed expression of oncogenes positively regulates cell proliferation, tumor suppressors play important roles as genomic guardians by restraining cell growth and proliferation, executing gene tically impaired cells and maintaining genomic stability. Rb was the fi rst tumor suppressor isolated in human and was found to be dysregulated in a majority of cancers (Friend, Bernards et al. 1986). Another intensively studies tumor suppresso r is p53, which plays pivotal roles in maintaining homeostasis of the cells.
2 p53 p53 history In 1979, p53 was originally recognized as a weak oncogene, which bound to the simian virus (SV40) large T antigen and accumulated in the cancer cells (DeLeo, Jay et al. 1979; Lane and Crawford 1979; Linzer and Levine 1979). Later, researchers found that the oncogenic activity was caused by the gain-of-function mutant p53. By forming tetramers with the wild type p53, mutant p53 suppresses wild type p53 function in a dominant-negative fashion (Sigal and Rotter 2000). In 1990, it wa s found that the E6 oncoprotein encoded by human papillomavirus promotes the degradation of p53 (Scheffner, Werness et al. 1990). The fact that viral system developed specific mechanisms using oncoproteins to target p53 for degradation underlines the importance of p53 inactivation for virus survival and proliferation. P53 belongs to a family of transc ription factors including p53, p63 and p73. While p63 and p73 take important responsibilities in maintaining normal development, only p53 evolved to prevent tumor formation (Yang and McKeon 2000). Now it is widely accepted that p53 is a classic tumor s uppressor. P53 lost of function is found in about 50% of all human tumors (Hollstein, Si dransky et al. 1991; Levine, Momand et al. 1991). Functional analysis revealed p53 as a potent stress responder regulating cell-cycle arrest, DNA repair, cell senescence and apoptosis.
P53 function and transcription down stream targets The structure of p53 represents a classic transcription factor (Ko and Prives 1996). It has transactivation domain in the N-terminus and the sequence-specific DNA binding domain in the central region (fig 1). The DNA-binding domain contains hot spots for mutation in tumor cells, indicating the tumor suppressor function of p53 relies heavily on its transcriptional activity. P53 induces the transcription and over-expression of a myriad of proteins through directly binding to their promoters, which leads to various biological functions (fig 1). Although the number of genes found regulated by p53 is still increasing, generally, these gene targets can be categorized into five groups according to their functions: (1) Cell cycle arrest genes, (2) Apoptosis genes, (3) DNA repair genes, (4) genes inhibiting angiogenesis and metastasis and (5) cellular senescence genes. 3 Figure 1 Structure and functions of p53 tumor suppressor
4 (1) Cell cycle arrest genes: CDKs or cyc lin-dependent kinases ar e a group of enzymes evolved specifically to promote the cell cy cle progression upon binding to correspondent cyclin proteins, thus supporting the mitosis and division of the cells. CDK inhibitors, for example, p16, p21 and p27, weigh the decision of going through S phase for DNA synthesis. P53 directly promotes the transc ription of p21, which inhi bits the transition from G1 to S and from G2 to M (e l-Deiry, Tokino et al. 1993). 14-3-3 is another p53 transcriptional target that blocks the G2-M transition (Chan, Hermeking et al. 1999; Laronga, Yang et al. 2000). (2) Apoptosis genes: When DNA damage is beyond repair, apoptosis is the last resolution to avoid the risk of having cells carrying da maged genetic signal and proliferating. The apoptotic process can proceed through two main pathways: the intrinsic and extrinsic pathways. The intrinsic pathway involves an apoptotic complex called apoptosome that is comprised of mito chondria-released cytochrome c, Apaf-1 and procaspase 9 (Kroemer and Reed 2000). The release of cytochrome c is a result of mitochondria depolarization, a process under control by two groups of Bcl-2 family proteins (Tsujimoto 2003). The proapoptotic group of proteins includes Bax and Bak residing on mitochondrial membrane and the BH 3 domain only proteins such as PUMA, Noxa. The other group mainly contains antiapoptotic Bcl-2 and Bcl-XL. The change in ratio between the two groups determin es the life or d eath of the cell. One major tumor suppressive function of p53 is to induce apoptosis, thereby preventing tumor cells from proliferating. P 53 can transcriptionally induce proapoptotic genes that participate in both intrinsic and ex trinsic apoptosis pathways. In the intrinsic
5 pathway, p53 induces proapoptotic genes such as Bax, PUMA and Noxa, which localize to the mitochondria and promote mitochondria depolarization (Miyas hita, Krajewski et al. 1994; Oda, Ohki et al. 2000; Nakano and V ousden 2001). As a result, cytochrome c is released and the subsequent apoptosome forma tion and caspase activation eventually lead to cell death. Another group of p53-induced ge nes belong to the death-receptor mediated apoptosis pathway or extrinsic pathway, including DR5, Fas, PIDD.(Owen-Schaub, Zhang et al. 1995; Wu, Burns et al. 1997; Muller, Wilder et al. 1998; Lin, Ma et al. 2000). Initiation of the extrinsic pathway results from ligand binding to the death receptors, leading to the clustering of deat h receptors at the cell membrane and the consequent induction of caspase activity. By transcripti onally activati ng proapoptotic genes in both intrinsic and extrinsic pathwa ys, p53 contribute signifi cantly to apoptosis and the disruption of p53 activity can lead to tumorigenesis. P53 promotes apoptosis not only by transc riptional activation, but also through transcriptional repression of the target genes. For instance, p53 represses the transcription of survivin, whic h is an inhibitor of apoptos is (Ambrosini, Adida et al. 1997). Although p53 transcripti onal repression is not as well understood as its transactivation function, it is suggested that under certain circumstances, p53 transcription repression contribu tes more to apoptosis than does its gene transactivation function (Koumenis, Al arcon et al. 2001). Additionally, p53 can also induce apopto sis via transcriptional-independent pathways (Caelles, Helmberg et al. 1994; Wagner, Kokontis et al. 1994; Haupt, Rowan et al. 1995; Kokontis, Wagner et al. 2001; Dumont, Leu et al. 2003). Evidence showed that the DNA-binding deficient p53 mutants could indu ce apoptosis almost as efficient as the
6 wild type. Furthermore, p53 regulates th e intrinsic apoptotic pathway by directly translocating to the mitochondria (Mihara, Erster et al. 2003; Mihara and Moll 2003). The interaction between p53 and anti -apoptotic proteins such as Bcl-2 and Bcl-XL changes the membrane integrity, leading to cytochro me c release and apoptosis. Additionally, it has been reported that p53 promotes the tr ansportation of Fas from cytoplasm to cell membrane, augmenting the extrinsic pathwa y (Bennett, Macdonald et al. 1998). (3) DNA repair genes: The defect in DNA re pair genes will eventually lead to genomic instability due to the accumulate d errors in DNA left uncorrected. The mechanism by which p53 mediates DNA repair re mains elusive. However, experimental data showed that P53 contribu tes to genomic stability by regulating the transcription of DNA repair genes such as p48 and GADD45. (Kastan, Zhan et al. 1992; Smith, Chen et al. 1994; Hwang, Ford et al. 1999). (4) Genes inhibiting angiogenesis and metast asis: Angiogenesis is a hallmark for tumorigenesis in solid tumors. In order for the tumor to progress and grow to a certain size, ample blood supply is needed that can only be satisfied by the generation of new blood vessels, or angiogenesis. Tumor cells sh ed into the blood ve ssels where they can be carried to distant locations to metastasize. P53 promotes expression of genes to antagonize tumor angiogenesis and metastasis pr ocesses. Some of the anti-angiogenesis or anti-metastasis genes activated by p53 include MASPIN (Song, L ee et al. 2002), TSP1 (el-Deiry, Tokino et al. 1993; Nagayama, Shigematsu et al. 2000) and BAI-1 (Nishimori, Shiratsuchi et al. 1997). An example of anti-metastasis gene activated by p53 is KAI (Schindl, Birner et al. 2001).
7 (5) Cellular senescence: Normal human cel ls have a finite lifespan, which allows only certain numbers of DNA replications. This has been defi ned as the cellular replicative senescence and it is mainly cause d by telomere shortening (Linskens, Harley et al. 1995). For example, human skin ce ll chromosomes lose 50-100 nucleotides at the telomere ends following each DNA replication cy cle, due to the intr insic mechanism built in the replication procedur e of the DNA lagging strand. After certain generations, the continuous shortening of the DNA leads to intolerable level of DNA loss and causes genomic instability. As a result, the cells which inherited the unstable chromosomes quit the cell cycle and enter into a phase called replicative senescence (Levy, Allsopp et al. 1992). In this way, telomere, the repetitive DNA composing the chromosome ends provides a counting mechanism to preven t the infinite DNA replication and cell proliferation. It has been suggested that be sides apoptosis, cellular senescence is another mechanism to prevent the cells bearing de fective DNA from tumorigenic transformation (Sager 1991; Yeager, DeVries et al. 1998). Additionally, cellular senescence is also suggested to be involved in aging. Senescence can also be induced thro ugh telomere-independent pathways. A variety of stimuli including oncogenic or mitogenic signals, chromatin remodeling, DNA damage or tumor suppressor activ ity have all been reported to elicit cellular senescence. Some tumor suppressors such as p16 (the cyclin-dependent kinase inhibitor) and p14 ARF (the inhibitor of oncogene MDM2) res pond to these senescen ce-inducing stimuli (McConnell, Starborg et al. 1998; Dimri, Itahana et al. 2000 ). Although the mechanism of p53 regulation of senescence remains elusive, there are indications that p53 responds to senescence-inducing stimuli through ATM and sensitizes the cells in a way similar to
8 DNA damage induced p53 activation (Vaziri 1997; Vaziri, West et al. 1997; Lansdorp 2000). It is very likely that p53 induces p21 expression, which in concert with p16 initiates and sustains the sene scence growth arrest of the cel ls (Atadja, Wong et al. 1995; Vaziri, West et al. 1997; St ein, Drullinger et al. 1999). Activation of p53 P53 is usually maintained at low levels in the cells by virtue of its short half-life. One of the most important functions of p53 in suppressing tumorigenesi s is to respond to cellular stress, achieved by promoting the ce ll cycle arrest or apoptosis of the DNAdamaged cells by transcriptionally regulati ng target genes beari ng the aforementioned functions. In this way, given the time by p53 induced cell cycl e arrest, the cells containing damaged DNA are repaired or rem oved by apoptosis to keep the homeostasis of the cells. Thus, this DNA-damage respons e of p53 is crucial for avoiding neoplasia generated from DNA lesions. One of the most potent DNA damages in flicted on the cells is DNA double strand break. This can be elicited from ionizing radi ation, certain chemical agents or genotoxic stress. Ataxia telangiectasia is an autoso mal recessive genetic disorder featured by neurodegeneration, immunodeficiency, radiatio n sensitivity and cancer predisposition (Crawford 1998; Becker-Catania and Gatti 2001). On genetic level, ataxia telangiectasia is caused by the loss or inactivation of ATM protein, which belongs to a conserved family of proteins, PI3 kinase like protein kinases or PIKK. Most of the PIKK family proteins possess serine/thre onine kinase activity. ATM kinase responds to DNA double
9 strand breaks by phosphorylating proteins cr ucial to various DNA response pathways (Banin, Moyal et al. 1998; Canman, Lim et al. 1998). One of the main targets of ATM kinase is p53, the activation of which mediates the G1-S cell cycle arrest by inducing the expression of CDK inhibitor p21. Upon DNA damage, ATM phosphorylates p53 on Serine 15 (Banin, Moyal et al. 1998; Canman, Lim et al. 1998; Khanna, Keating et al. 1998). Moreover, ATM also phosphorylates and activates CHK2, a checkpoint kinase that phosphorylates p53 on Serine 20. (Bartek, Falck et al. 2001; McGowan 2002). Another ATM phosphorylation target is MDM2 at serine 395, which disrupted the nuclear export signal that is needed for efficient p53 export into the cytoplasm (Maya, Balass et al. 2001). MDM2 affects p53 in both its function and its stability. In order to keep p53 inactive and at low protein level, MDM2 binds to the N-terminal transactivation doma in and inhibits the transcription factor function of p53. On the other hand, MDM2 is the ubiquitin E3 ligase, mediating the ubiquitination and degradation of p53 by pr oteasome. Furthermore, the NES on MDM2 has been reported to enhance MDM2 mediat ed ubiquitination and degradation of p53. These aforementioned phosphorylations dissoci ate p53 from the binding to its major negative regulator MDM2, thus significantly contributes to the stabilization and activation of p53 as a transcription factor. Other PIKK family protei ns such as ATR and ATX carry out similar functions in phos phorylating p53 (Hammond, Denko et al. 2002; Heffernan, Simpson et al. 2002). Even though they can collaborate with ATM in the double strand break response, ATR and ATX are also activated in response to different genotoxic stresses, such as UV-irradiation and hypoxia.
10 P53 regulation by MDM2 and MDMX As a tumor suppressor, p53 functions to i nhibit growth and pro liferation of abnormal cells, which impose a potential threat for tumo rigenesis. In mouse models, p53 null mice showed high incidence of tumor forma tion, supporting p53 as a tumor suppressor (Donehower, Harvey et al. 1992; Livingstone, White et al. 1992; Yin, Tainsky et al. 1992; Fukasawa, Wiener et al. 1997). The stability and activity of p53 are mostly regulated by MDM2. MDM2 is one of the down stream targets of p53 and the expression of MDM2 is often elevated subsequent to p53 activation (Barak, Juven et al. 1993; Zauberman, Flusberg et al. 1995). As me ntioned above, MDM2 possesses E3 ubiquitin ligase activity towards p53 and acts as a majo r negative regulator to keep p53 stability and activity at low level (Oliner, Pietenpol et al. 1993; Zauberman, Barak et al. 1993; Haupt, Maya et al. 1997; Honda and Yasuda 1999). Thus, the relationship between p53 and MDM2 composes the classic negative feedback loop, whereas the elevation of MDM2 expression by p53 leads to the down regulation of p53 itself (Wu, Bayle et al. 1993). The importance of MDM2 as the major negative regulator of p53 is also revealed in mouse models. Nullifying MDM2 leads to embryonic lethality at day 3.5 due to the massive apoptosis induced by p53, supporting MDM2 as a major p53 negative regulator (Jones, Roe et al. 1995). Furthermore, the embryonic lethality of MDM2 null mice can be rescued by crossing into p53 null backgr ound, indicating the major role of MDM2 is to negate p53 functionality. This indicates the intrinsic mechanism in cells to keep p53 level under tight control.
11 The negative effect of MDM2 inflicted on p53 is not without control. As the inhibition of MDM2 favors p53 function, MDM2 inhibitor might possess tumor suppressor characteristics. ARF is such a tu mor suppressor. ARF expression is elevated upon the induction by certain genetic stress signals such as E1A, E2F, myc, ras and v-abl. Consequently, as a major negative regulator of MDM2, ARF inhibits MDM2 mediated p53 ubiquitination and degradation while negating MDM2s inhibition of p53 transcriptional activity (Kamijo, Weber et al 1998; Pomerantz, Schreiber-Agus et al. 1998; Stott, Bates et al. 1998; Zhang, Xiong et al. 1998). The other important p53 regulator is MDMX, an MDM2 homolog. Recently, MDMX was found as another P53 binding protein besides MDM2 (Shvarts, Steegenga et al. 1996). MDMX and MDM2 sh are significant homology in se veral functional domains, such as P53 binding domain, Zn finger and th e RING domain. MDMX was also found to be able to inhibit P53 trans activation. However, no feedb ack loop is formed between p53 and MDMX because MDMX does not contai n p53-responsive element on the promoter and is not induced by p53 activation. On the other hand, MDMX has not been found to possess ubiquitin E3 ligase activity so it does not directly affect p53 stability (Jackson and Berberich 2000; Stad, Ramos et al. 2000). Nonetheless, MDMX has been established as another important negative regulator of p53. The MDMX knockout mice died at embryonic day 7 due to severe prolif eration deficiency in contrast of massive apoptosis as in the case of MDM2 knockout mice. Similar to MDM2 null mice, this embryonic lethality can be rescued by nu llifying P53 simultaneously (Parant, ChavezReyes et al. 2001; Finch, Donovi el et al. 2002; Migliorini, Denchi et al. 2002). This
12 demonstrated the non-overlapping functiona lity of MDMX with MDM2 in embryonic development as well as its majo r function in P53 regulation. P53 posttranslational modification P53 responds to DNA damage signals thr ough an increase in transcriptional activity and protein stabilit y. The accumulation of activ e p53 largely relies on p53 posttranslational modifications (Appella and Anderson 2001). Compared to the change in transcriptional or translational rate, posttr anslational modification is a time-efficient way to elevate the activity of p53 as a transcri ption factor in stress response. So far, reported posttranslational modifications of p53 include phosphorylation, ubiquitination, acetylation, neddylation and sumoylatoin, ea ch of them has profound effect on p53 stability and/or activity. P53 phosphorylation is pivotal for its activation after DNA damage. The Nterminus of p53 is phorsphorylated by at leas t three different protei n kinases, ATM, ATR and DNA-PK (Banin, Moyal et al. 1998; Ca nman, Lim et al. 1998; Tibbetts, Brumbaugh et al. 1999; Nagayama, Shigematsu et al. 2000). The central event in this process is the phosphorylation of p53 Serine 15 and Serine 37, leading to the di ssociation of MDM2 from p53 (Avantaggiati, Ogryzko et al. 1997; Shieh, Ikeda et al. 1997; Lambert, Kashanchi et al. 1998; Dumaz and Meek 1999; Zhang and Xiong 2001). Alternatively, MDM2 as the major negative re gulator of p53 can also be phorsphorylated, favoring its dissociation from p53 (Mayo, Turchi et al. 199 7). In addition, the C-terminal serine residue 378 undergoes dephosphor ylation upon DNA damage, lead ing to the recruitment
13 of 14-3-3 adaptor protein followed by the enhanced DNA binding affinity of p53 (Waterman, Stavridi et al. 1998). P53 acetylation also regulates the proteinprotein interaction and activity of p53. It has been suggested that the phosphoryl ation in p53 N-terminus potentiates the acetylation by transcriptional co-activator s CBP/P300 and PCAF, thus enhancing p53 transcriptional activity and stability (Avantaggiati, Ogryzko et al. 1997; Gu, Shi et al. 1997; Lill, Grossman et al. 1997; Sakaguchi Herrera et al. 1998; Dumaz and Meek 1999). P53 activity is also under the contro l of HDACs, which often associate with transcription co-repressors and remove acetyl groups from substrates (Murphy, Ahn et al. 1999; Juan, Shia et al. 2000). The deacetylation of p53 may repres ent a mechanism to help restore cell biological f unctions when the damage is re paired and hyperactive p53 is no longer needed. As the major negative re gulator of p53, MDM2 blocks p53 acetylation both in vivo and in vitro, suppressing the tu mor suppressor function of p53 (Kobet, Zeng et al. 2000; Ito, Lai et al. 2001). P53 stability is most directly regulat ed by its ubiquitination. Aforementioned MDM2-mediated ubiquitination sends p53 to 26S proteasome degradation and explains the particularly short half-life of p53. Besides MDM2, Pirh2 and COP1 are both p53inducible genes and could me diate the ubiquitination and de gradation of p53, forming negative feedback loops (Leng, Lin et al. 2003; Dornan, Wertz et al. 2004). The fact that multiple proteins converge on the negative regulation of p53 emphasizes the importance of keeping p53 level under strict control. Alternatively, the mechanism de-ubiquitinating p53 should hypothetically restore p53 stability. HAUSP is one such antagonist to MDM2
14 ubiquitination activity of p53, removing ubiqui tin molecules from p53 and increases the protein half-life (Li, Chen et al. 2002). Besides ubiquitin, ubiquitin-like small molecules such as Nedd8 or sumo have also emerged as p53 modifiers that regul ate p53 activity (Melch ior and Hengst 2002; Harper 2004; Xirodimas, Saville et al. 2004). Both neddylation and sumoylation utilize parallel mechanisms to ubiquitination in th at the ubiquitin-like molecules need to be activated and transferred to substrates by E1 and E2 enzymes, respectively. MDM2 promotes conjugation of both sumo and nedd8 to p53. However, p53 sumoylation and neddylation have opposite effect s on its transcription activit y whereas sumoylation of p53 moderately enhances its transactivati on while MDM2-mediated neddylation of p53 inhibits its function (Melchior and Hengst 2002; Harper 2004; Xirodimas, Saville et al. 2004). P53 as a cancer therapy target Cancer is among the top 3 causes of death in United States. To date, the treatment of cancer still heavily relies on chemothera pies and radiotherapies, which primarily converge to the functionality and activity of p53. In order to evade from p53 tumor suppressive functions, about 50% of human tumors have p53 mutated (Hollstein, Sidransky et al. 1991; Levine, Momand et al. 1991). In the rest of the cases, the maintenance of wild type p53 accompanies de ficiencies of other components in the p53 regulatory pathway, such as the amplification of MDM2 or the loss of function of ARF. Realizing the pivotal role of p53 in cancer control, scientists have endeavored to develop p53-targeting therapeutic strategies. They can be categorized into two classes.
15 The first tactic targets can cer cells containing mutant p53. This is based on the hypothesis that the function of mutant p53 can be restored to that of the wild type. Theoretically, cells bearing mutant p53 not only risk loss of function of the tumor suppressor, but also obtain the gain of function activity of oncogene as the p53-mutant allele exhibits dominant negative activity to silence the proper function of the wild type p53. Strategies were developed to either eliminate cells be aring mutant p53, exemplified by Onyx-015, or correction of mutant p53 f unctions, represented by PRIMA-1 (Heise, Sampson-Johannes et al. 1997; Ries, Brandts et al. 2000; R ogulski, Freytag et al. 2000; Bykov, Issaeva et al. 2002; Lebedeva, Su et al. 2003; Haupt and Haupt 2004). The second class aims to boost the activit y of wild type p53. In a significant percentage of tumors, p53 status rema ins wild type, but accompanied with overexpression or amplification of MDM2. MDM2 is one major negative regulator for both p53 stability and activity. To this end, pe ptide and small molecule inhibitors have been developed to achieve the abrogation of MDM2-P 53 binding, alleviating the suppression of p53 (Haupt and Haupt 2004).
16 MDM2 MDM2 History MDM2 or Murine Double Minute Clone 2 was originally cloned from the abnormal chromosomes called double minute (C ahilly-Snyder, Yang-Feng et al. 1987; Fakharzadeh, Trusko et al. 1991). These chromoso mes frequently harbor amplified genes, leading to uncontrolled cell prol iferation and tumori genesis. The amplification of MDM2 in mice resulted in high risk of tumor form ation, giving scientists the first sign that MDM2 is an oncogene. Later, MDM2 was co -purified with p53 and found to negatively regulate p53 stability and transcriptional activity (Momand, Zambetti et al. 1992). Since p53 has been defined as a tumor suppressor gene, the inactivation of which by MDM2 established MDM2 as an oncogene. MDM2 mouse model and clinical relevance One of the crucial experiments establis hed MDM2 as a major regulator of p53 tumor suppressor was done in mouse models. P53 -/mice were viable but developed tumors after birth, supporting th e role of p53 as a tumor suppressor. MDM2-/mice were embryonic lethal, died at embryonic day 3.5. If the lethality of MDM2 null mice is caused by the loss of MDM2 negative regulation of p53, then crossing p53 to the MDM2 null background should be able to rescue MDM2 null mice from lethality. This was proven true (Jones, Roe et al 1995; Montes de Oca Luna, Wagner et al. 1995). Beside, concerning tumor formation, MDM2 knockout and MDM2/p53 double knockout mice
17 are similar. Thus, the mouse model supports the hypothesis that MDM2 is an oncogene and its essential role is to negatively regulate p53. If MDM2 is the major in hibitor of the tumor suppr essor p53, then p53 loss of function alone is sufficient for tumorigenesis, precluding the need for amplification of MDM2 function. Indeed, the mutation of p53 and gene amplification of MDM2 do not often coincide. MDM2 gene amplification wa s first detected in a significant number of osteosarcomas and soft tissue tumors and th e research up to date shows a 7% overall amplification frequency of MDM2 (Momand, Jung et al. 1998). Furthermore, in some of the tumors that maintain both functiona l p53 and normal level of MDM2, MDM2 can still be hyperactive due to the loss of function of another tumor suppressor ARF (Pomerantz, Schreiber-Agus et al. 1998; Zha ng, Xiong et al. 1998). As a major negative regulator of MDM2, ARF is mutated at hi gh frequency in human cancers (Momand, Jung et al. 1998). MDM2 inhibits p53 transcriptional activity Although p53 can induce apoptosis independent of its transcri ption activity, P53 tumor suppressor function is largely dependent on its role as a transcription factor. Upon genotoxic stresses, p53 induces apoptosis, cell cycle arrest, DNA repa ir and inhibition of angiogenesis. Some of the genes contai ning p53 responsive element and induced by p53 include p21, GADD45, 14-3-3 Bax, PUMA, and MDM2 (Kas tan, Zhan et al. 1992; Barak, Juven et al. 1993; Perry and Levine 1993; Perry, Piette et al. 1993; Miyashita, Krajewski et al. 1994; Chan, Hermeking et al. 1999; Nakano and Vousden 2001). Distinct from other p53 inducible genes, MD M2 binds to the tran scriptional activation
18 domain on the N-terminus of p53, inhibiting p53 transcription of its downstream targets, thus established a negative feedback loop between p53 and MDM2 (Barak, Juven et al. 1993; Juven, Barak et al. 1993; Wu, Bayle et al. 1993; Barak, Gottlieb et al. 1994). This is necessary for the proper growth and proliferation of the cells under normal condition because p53 overexpression will lead to cell death or growth arrest. However, upon DNA damage, rather high level of p53 is need ed to harness the growth of damaged cells and send them either to DNA re pair or apoptosis pathway. At this point, the physical association between MDM2 and p53 is broken by the phosphorylation of p53 at serines critical for p53/MDM2 binding. W ithout the inhibition by MDM2, p53 is transcriptionally active and elic its cell cycle arrest or apoptos is to the damaged cells. In this sense, the override of p53 by MDM2 a llows DNA damaged cells to proliferate, leading to tumorigenesis. MDM2 is an ubiquitin E3 ligase for multiple proteins In the ubiquitination pathway, E3 ubiquitin ligases are the group of enzymes catalyzing the conjugation of ubiquitin moiety to the protein substrates. To date, there are two major groups of ubiquitin E3 ligases. One is the HECT domain containing E3, catalyzing ubiquitin conjugation by forming thio -ester intermediates between E2 and the substrates. The other class is the RING fi nger domain ubiquitin E3 ligases, featured by the consensus sequence CX 2 CX (9-39) CX (1-3) HX (2-3) C/HX 2 CX (4-48) CX 2 C. MDM2 is a RING domain type of E3. So far, proteins diverse in their func tions have been found to be the ubiquitin ligase substrates of MDM2. These prot eins include p53, MDM2 itself, MDMX, b-
19 arrestin, PCAF and insulin-like growth fact or 1 receptor (IGF1R) (Momand, Zambetti et al. 1992; Fang, Jensen et al. 2000; Shenoy, McD onald et al. 2001; Girn ita, Girnita et al. 2003; Pan and Chen 2003; Jin, Zeng et al. 2004). The ubiquitination of these substrates by MDM2 leads to the rapid tu rnover of the proteins through proteasome degradation. MDM2 promotes other posttransl ational modifications of p53 Besides being an E3 ligase for p53 and other substrate proteins, MDM2 also stimulates other forms of posttranslational modifications. Sumo-1 is a 110 amino acid protein belonging to an ubiquiti n-like family. The conjugation of sumo to a variety of proteins changes protein localization or activ ity. It was reported that MDM2 mediates p53 sumoylation, which moderately enhances p53 transcriptional ac tivity (Melchior and Hengst 2002). Upon DNA damage, p53 is also acetylated by a cetyl transferase CBP/P300. The acetylation of p53 contributes to its stabilization. MDM2 suppresses p53 acetylation by binding to and inhibiti ng the function of CBP/P300, rendering p53 more susceptible to degradation (Kobet, Ze ng et al. 2000; Ito, La i et al. 2001). Furthermore, MDM2 promotes the conjugation of another ubiquitin-like molecule, nedd8 to p53, leading to the transcri ptional inhibition of p53 activity (Harper 2004; Xirodimas, Saville et al. 2004). MDM2 nuclear import and export MDM2 contains both nuclea r localization signal (NLS) and nuclear export signal (NES), making it possible for transportation in and out of the nucleus independently (Roth, Dobbelstein et al. 1998; Tao and Levine 1999). In recent years, MDM2 has been
20 reported to mediate the nuclear export of p53 into cytoplasm, rendering p53 more susceptible for proteasome degradation in th e cytoplasm (Roth, Dobbelstein et al. 1998; Tao and Levine 1999). Experiments using Le ptomycin B blocked the nuclear export of proteins, as a result, p53 was stabilized and its transcriptional activity enhanced (Freedman and Levine 1998). However, MDM2-independent p53 nuclear export might also exist (Zhang and Xiong 2001).
21 MDMX MDMX Protein structure In 1996, a novel p53-binding protein was id entified and named MDMX for its structural similarity with the well-esta blished p53-binding protein, MDM2 (Shvarts, Steegenga et al. 1996). MDMX shares signi ficant homology with MDM2 in its protein structure. Compared with MDM2, which is 491 amino acids, MDMX is a 490 amino acid protein that shares several functional domains with MD M2 (Shvarts, Bazuine et al. 1997). Over the years, it has been rec ognized that MDMX and MDM2 share highly conserved domains in three main regions (f ig 2). The first region spans amino acid 4294, comprising the p53-binding domain. This domain is strictly conserved between MDM2 and MDMX, sharing 53.6% identity (Bottger, Bottger et al. 1999). This conserved region enables both MDMX a nd MDM2 to bind p53 on its N-terminal transactivation domain and inhibit p53 tran scription activity. The second conserved domain or, the Zinc binding domain locates in the central region of both proteins, which is comprised of residues 301-329. The function of the Zinc finger domain is still largely enigmatic. The third highly conserved region is in the C-termi nus of both proteins, between amino acid 444-483. This RING finge r domain is characterized by its CX 2 CX (939) CX (1-3) HX (2-3) C/HX 2 CX (4-48) CX 2 C consensus sequence (Sharp, Kratowicz et al. 1999; Tanimura, Ohtsuka et al. 1999). MDMX and MDM2 share 53.2% identity in this region, which enables the homoand hetero-dimerizat ion of the two. By using the yeast twohybrid analysis system, Tanimura et al demonstrated that the MDMX-MDM2 heterodimer is more stable than the hom odimers of each. The RING finger domain of
MDM2 is necessary and sufficient for its ubiquitin E3 ligase activity (Momand, Zambetti et al. 1992). However, in contrast to MDM2, no E3 ligase activity has been found for MDMX (Jackson, Lindstrom et al. 2001; Migliorini, Danovi et al. 2002; Linares, Hengstermann et al. 2003). Figure2. MDM2 and MDMX share homology in their protein structures Among the regions that are distinct between MDMX and MDM2 is the central acidic domain. The acidic domain of MDM2 provides a binding site for ARF. The interaction between MDM2 and ARF results in the inhibition of MDM2 E3 ligase activity and the stabilization of both MDM2 and p53 (Weber, Taylor et al. 1999; Zhang and Xiong 1999). ARF/MDM2 binding also relocates MDM2 from nucleus to nucleolus, physically separates MDM2 from p53 (Pomerantz, Schreiber-Agus et al. 1998). 22
23 However, the binding between ARF and MDMX remains controversial (Jackson, Lindstrom et al. 2001; Wang, Arooz et al. 2001). The research of Kawai et al and Meulmeester et al shed some light on the functional significance of the central acidic domain of MDM2. Both groups demonstrat ed that the central acidic region is indispensable for MDM2 E3 ligase activ ity, making it a possible explanation for MDMXs lack of ubiquitin E3 ligase func tion (Kawai, Wiederschain et al. 2003; Meulmeester, Frenk et al. 2003). Other st ructural discrepancies between MDM2 and MDMX include the nuclear lo calization signal (NLS) and nuc lear export signal (NES) in MDM2, which are not conserved in MDMX (Jackson and Berberich 2000; Migliorini, Danovi et al. 2002; Migliorini Denchi et al. 2002). MDMX gene structure The homology on protein level between MDMX and MDM2 can be explained by the conservation of the two genes at genomic level (Parant, Chavez-Reyes et al. 2001). The exons 4-12 between MDMX and MDM2 are well conserved. The last exon, exon 12 encodes RING finger domains of MDM2 and MDMX, resulting in the well-conserved RING finger domains on protein level. Howeve r, the 5 ends of the genes are quite distinct. In contrast to MDM2, no p53-responsive element was found on MDMX (Shvarts, Steegenga et al. 1996). Consiste ntly, MDM2, but not MD MX is induced upon the activation of p53 after DNA damage.
24 MDMX as an oncogene In order to evade the genomic surve illance from tumor suppressor p53, tumor cells selectively mutate or delete p53 ge ne, escaping from the checkpoint control to proliferate. Alternativel y, members of p53 regulatory pathway are involved for the advantage of tumorigenesis. As the major negative regulator of p53 function and stability, MDM2 has been shown to have a pproximately 7% amplif ication rate in 28 tumor types (Momand, Jung et al. 1998). Moreover, MDM2 amplification and p53 mutation did not concur in these tumors, implicating that MDM2 overexpression is sufficient for p53 inactivation and tumor form ation. In other cases, wild type p53 is accompanied by the loss of function of ARF, which is the main inhibitor for MDM2, leading also to hyperactive MDM2 (Momand, Jung et al. 1998; Pomerantz, SchreiberAgus et al. 1998; Zh ang, Xiong et al. 1998). The mouse models demonstrated that MDMX and MDM2 play non-overlapping roles in negating p53 function (P arant, Chavez-Reyes et al. 2 001; Parant, Re inke et al. 2001; Finch, Donoviel et al. 2002; Migliori ni, Denchi et al. 2002). MDMX gene amplification was found in a s ubset of malignant gliomas (R iemenschneider, Buschges et al. 1999). MDMX localizes on chromosome 1 q32, which contains frequently amplified sequences including GAC1 REN and RBBP5 in malignant gliomas. In analogy to MDM2 amplification, no p53 mutation concurred with MD MX amplification in the malignant gliomas investigated, sugges ting that MDMX overexpression alone is sufficient for p53 inactivation. Furthermore, no MDM2 amplification was found in the presence of MDMX amplification, supporti ng the non-overlapping re gulation of MDMX and MDM2 on p53.
25 Aberrations of chromosome 1q freque ntly occur in cervical and ovarian carcinoma (Danovi, Meulmeester et al. 2004) MDMX overexpression was found in a significant number of these cancer cell lines a nd seems to correlate with the presence of wild type p53 in those ce lls. In a comprehensive experiment performed by Danovi et al over 500 different human primary tumors were investigated for their MDMX status. It was found that MDMX is overexp ressed at the protein level a nd correlates with 19% of breast cancers, 19% of colon cancers and 18% of lung can cers. None of the MDMX overexpression cell lines contains p53 mu tation and only few contain MDM2 gene amplification. Additionally, in the retrovirus-infected MEFcells, MDMX overexpression caused the immortalization of ME F without the selective pressure for the loss of p53 or ARF functions. Furthermore, th e transformation of MEF by HRas A12 requires p53 loss of function, which can be achieved by concomitant overexpression of HRas A12 and MDMX. These findings argue for an important and non-overlapping role of MDMX with MDM2 in negatively regulating p53 tumor suppre ssive function and implicate MDMX as a putative oncogene. MDMX mouse models In order to elucidate the functiona l significance of MDMX in vivo, MDMX heterozygous mice were generated and crossed for the homozygous MDMX knockout offspring (Parant, Chavez-Reyes et al. 2001; Parant, Reinke et al. 2001; Finch, Donoviel et al. 2002; Migliorini, Denchi et al. 2002). Although mice bearing heterozygous MDMX seemed to develop normally, the MDMX -/mice were not viable, died at E7.5-
26 8.5 dpc due to lack of cell proliferation and ma ssive apoptosis in central nervous system. However, when MDMX null allele was crossed into the p53 null background, the generated MDMX/ p53 double null mice were re scued from the embryonic lethality but were prone to tumorigenesis. Moreover, the loss of MDMX cannot be compensated by the expression of MDM2. This evidence str ongly argues for the func tion of MDMX as a major p53 negative regulator in vivo and that MDMX and MDM2 serve to regulate p53 by distinct mechanisms. MDMX upon DNA damage Both MDMX and MDM2 proteins contai n a p53-binding domain and inhibit the growth-suppressive function of p53. Howe ver, unlike MDM2, MDMX gene does not possess p53-responsive element thus cannot be induced in a p53-dependent manner after DNA damage (Shvarts, Steegenga et al. 1996). Early studies have found that MDMX is constitutively expressed at both mRNA and protein levels duri ng cell proliferation, differentiation or after DNA damage (Jackson and Berberich 1999). MDMX lacks the nuclear localization signal (NLS) and is predominantly cytoplasmic in unstressed cells whereas both p53 and MDM2 possess functional NLS and localize in nucleus (Migliorini, Danovi et al. 2002). DNA dama ge induced p53 or MDM2 facilitates the nucleus translocation of MDMX by forming heterodimeric or heterotrimeric complexes with MDMX (Migliorini, Danovi et al. 2002). However, MD MX nuclear entry can also be independent of either MDM2 or p53, de monstrated by the efficient MDMX nuclear translocation in p53-/-MDM2-/mouse embryo fibroblasts after DNA damage (Li, Chen
27 et al. 2002). Overexpression of MDMX under cellular stress has been shown to enhance p53 stability and activity due to the comp etitive binding between MDMX and MDM2 to p53. Under stressed conditions, overexpressed MDMX dominates the binding to p53, allowing a p53 population free of potent MD M2 inhibition, leading to enhanced p53dependent apoptosis (Gu, Kawai et al. 2002; Mancini, Gentilet ti et al. 2004). Although this result reflects a potential posi tive regulation of p53 by MDMX, the MDMX overexpression under DNA damage is an artificial effort. Studi es in our lab and others found that MDMX is polyubiquitinated and degraded by proteasome in an MDM2dependent fashion after DNA damage (Kawai, Wiederschain et al. 2003; Pan and Chen 2003). Furthermore, this MDM2-dependent polyubiquitination of MDMX can be enhanced by ARF (Kawai, Wiederschain et al. 2003; Pan and Chen 2003). ARF was shown to inhibit MDM2 E3 ligase activity toward p53 and MDM2 polyubiquitination and degradation, leading to st abilization of both proteins (Kamijo, Weber et al. 1998; Pomerantz, Schreiber-Agus et al. 1998; Stott, Bates et al. 1998; Zhang, Xiong et al. 1998). Additionally, ARF stimulates p53 grow th suppressive function and can inhibit MDM2 or MDMX mediated p53 suppression (Jackson, Lindstrom et al. 2001). By stimulating the polyubiquitination of MDMX, ARF presumably enhances p53 functions by targeting both of its negative regu lators namely MDMX and MDM2. MDMX-MDM2-p53 The negative feedback loop between p53 a nd MDM2 has been well established. Explicitly, p53 transcripti onally activates MDM2, which negatively regulates p53 by
28 inhibiting its transactivati on and destabilizing p53 by mediating its proteasome degradation (Barak, Juven et al. 1993; Juven, Barak et al. 1993; Wu, Bayle et al. 1993; Barak, Gottlieb et al. 1994). The mouse m odels experiments support MDMX as another major negative regulator of p53 function in vivo besides MDM2 (Parant, Chavez-Reyes et al. 2001; Finch, Donoviel et al. 2002; Migliorini, Denchi et al. 2002; Parant and Lozano 2003). Moreover, it is well recogni zed that MDMX can also bind to the Nterminus transactivation domain of p53, t hus inhibiting p53 transcriptional function (Shvarts, Steegenga et al. 1996; Bottger, Bottger et al. 1999; Stad, Ramos et al. 2000). We now know that MDMX is another substr ate of MDM2 E3 lig ase activity (Kawai, Wiederschain et al. 2003; Pan and Chen 2003). MDMX is polyubiquitinated by MDM2 and degraded in a proteasome-dependent path way. However, two main questions remain elusive. (1) How does MDMX regulate MD M2 function and stability? (2) How do MDMX and MDM2 cooperate to re gulate p53 function and stability? One model proposed that MDMX and MDM2 regulate p53 in a mutually dependent manner. MDM2 is a very unstabl e protein due to its intrinsic E3 ligase activity. MDMX stabilizes MDM2 by inhibiting the ubiquitin E3 ligase function of MDM2 toward itself. On the other hand, MD M2 recruits MDMX to the nucleus where MDMX retains a nuclear pool of p53 th rough direct binding. Since MDM2-p53 interaction leads to re-locali zation of p53 to cytoplasm, where p53 is efficiently degraded, MDMX stabilizes p53 by reta ining it in the nucleus. Meanwhile, MDMX and MDM2 synergistically inactivate p53 transcriptional activity in the nucleus thr ough direct binding to p53s transactivation domain. Therefor e, MDMX and MDM2 cooperatively inhibit p53 transcription activity but antagonize each other in regulating p53 stability. This
29 model is supported by a majority of th e labs employing overexpression technique (Jackson and Berberich 2000; Stad, Ramos et al. 2000; Stad, Little et al. 2001). The other model agrees on the obser vation that MDMX stabilizes p53 but assumes a distinct mechanism for the stabilizati on. It is suggested that MD MX over-expression enhances p53 stability and even functionality thr ough the competitive binding of MDMX to p53 with MDM2. MDMX overexpression has a squelching effect on p53-MDM2 binding, resulting in more p53 free from MDM2 associ ation and inhibition (Gu, Kawai et al. 2002; Mancini, Gentiletti et al. 2004). The third model is supported by in vitro assays, showing that MDMX stimulates the E3 ligase activity of MDM2, explaining the destabilization of both p53 and MDM2 at the overexpression of MDMX. All except for one group excluded the possible E3 function of MDMX (Badciong and Haas 2002).
30 ARF P16 INK is a cell cycle checkpoint gene that inhibits cyclin-dependent kinase 4 (Sherr and Roberts 1995; Weinberg 1995). The gene locus of p16 INK on chromosome 9 is mutated at high frequency in human tumors (Kamb, Gruis et al. 1994; Nobori, Miura et al. 1994). Later it was found that the necessity in tumor cells to mutate p16 locus lies in the fact that this locus encodes a second ge ne transcribed with an alternative reading frame, namely ARF (fig 3) (Serrano, Hannon et al. 1993; Quelle Zindy et al. 1995). Although ARF is completely unrelated to p16, the mutation of ARF also caused high frequency of tumor genesis independent of p16, supporting ARF as a tumor suppressor (Foulkes, Flanders et al. 1997; Ruas and Peters 1998; Sharpless and DePinho 1999). The mechanism of ARF induced cell cycle arrest is distinct from that of p16 INK A myriad of genetic stress in cluding E1A, E2F, myc, ras and v-abl induce the expression of ARF. ARF activates the function of tu mor suppressor p53 by harnessing p53 inhibitor, MDM2 (Kamijo, Weber et al. 1998; Pomerantz, Schreiber-Agus et al. 1998; Stott, Bates et al. 1998; Zhang, Xiong et al. 1998). ARF play s a major role in inhibiting the negative regulation of MDM2 on p53 through at least three mechanisms. One relies on the nucleolus localization signal (NoLS) containe d within 85-101 residues on the C-terminal domain of human ARF, enables ARF nucleolus -localization (Kamijo, Weber et al. 1998; Pomerantz, Schreiber-Agus et al. 1998; Stott, Bates et al. 1998; Zhang, Xiong et al. 1998). By binding to MDM2 central region, AR F recruits MDM2 into nucleolus, thus
physically separated MDM2 from p53. The second mechanism involves the abrogation of MDM2 E3 ligase activity by ARF (Weber, Taylor et al. 1999; Zhang and Xiong 1999). The third mechanism is supported by the experiments showing that ARF attenuates the suppression of p53 acetylation by MDM2, contributing to the stabilization of p53 (Ito, Lai et al. 2001). The end result of both mechanisms is that ARF stabilizes and activates p53, inducing the transcription of a myriad of p53-dependent genes that function in DNA repair, apoptosis, cell cycle arrest and anti-angiogenesis. Figure3. p16 and ARF are two tumor suppressors encoded in the same gene locus but working by distinct mechanisms 31
32 ARF is not a stereotypical inhibitor of MDM2 and the interplay between the two proteins is still being explored. Nevertheless, previous studies showed that ARF works on all aspects of MDM2-mediated posttransla tional modification. MDM2 ubiquitination of the substrates including self-ubiquitination is inhibited by ARF, resulting in the stabilization of the MDM2 E3 substrates (Kamijo, Weber et al. 1998). In addition, in response to stress signals, p53 is acetylated by CBP/P300, which increases p53 stability. MDM2 suppresses p53 acetylation by binding to the acetyl transferase CBP/P300 and inactivating the enzyme. ARF abrogates MDM2s suppression on p53 acetylation, thus restores p53 stability and activity (Ito, Lai et al. 2001). Chen et al also demonstrated that MDM2 stimulatesc p53 sumoylation (Chen and Chen 2003). However, instead of antagonizing MDM2s modification on p53, ARF enhances MDM2-mediated p53 sumoylation. Data from this manuscript demonstrated that ARF stimulates MDM2mediated MDMX ubiquitination and MDM2 -independent MDMX sumoylation. Therefore, ARF is a dynamic regulator of MDM2. Ubiquitination and proteasome degradation review Cellular protein degradation is a strictly re gulated process, aff ecting all aspects of biological functions. One important mech anism for protein degradation is through ubiquitination-proteasome degradation pathway. Ubiquitin, as its name indicated, is a widely expressed protein pres ent in most of the tissues and organisms, well conserved from yeast to human. It is a 76-amino acid small protein that can be conjugated through
33 an isopeptide linkage to the lysines in th e target protein as a post-translational modification (Pickart 2001; Weissman 2001). Since its initial disc overy, ubiquitination has been recognized in a wide variety of biological processes, ranging from signal transduction to endocytosis. Ubiquitin-mediated proteasome degradati on is a stepwise procedure (Haas and Siepmann 1997). The first step involves ubiquitin-activating en zyme E1 activating ubiquitin in an ATP-dependent fashion. Then ubiquitin molecule is transferred to E2, ubiquitin-conjugating enzyme, forming a high-energy thio-est er intermediate of E2ubiquitin. E3 ubiquitin ligase is a group of enzymes mostly responsible for substrate recognition and specificity. There are two sets of E3s classified by their distinct functional domains. The HECT domain E3 s contain a highly conserved 350 amino acid region in their carboxyl termini that was first discovered in E6-AP, the ubiquitin E3 that mediates the ubiquitination and degradation of tumor suppressor p53 by association with E6 viral oncoprotein. HECT domain E3s form intermediates between E2 and the substrate proteins while the RING domain containing E3s directly transfer E2 to the substrate proteins (Haas and Siepmann 1997). RING finger containing proteins were originally recognized merely for its potential in protein dimeriza tion. Only until recently wa s it uncovered as a group of enzymes playing important roles in ubiqui tination pathway. RING proteins share conserved structure CX 2 CX (9-39) CX (1-3) HX (2-3) C/HX 2 CH (4-48) CX 2 C, the cysteines and histidines represent the zinc binding sites. Crystal struct ures of RING fi nger proteins indicated that the protein structure of RING domain permits the optimal positioning between E2 and the substrate, ensuring the tr ansfer of ubiquitin to the target protein.
34 RING finger E3s can be divided into two groups: single and multi subunit E3s (Haas and Siepmann 1997). Single-subunit E3s contain the substrate binding site and the RING finger catalytic domain on one protein. Some of the well-studied single subunit RING finger E3s include p53 regulating oncoprotei n MDM2; Parkin, the mutation of which associates with juvenile Parkinsons dis ease and c-Cbl, a proto-oncoprotein. Multisubunit E3s have the RING finger-containi ng subunit and the s ubstrate recognition subunit separated on distinct proteins Among them are SCF (Skp1/Cul11/F-box protein), APC (anaphase promoting co mplex) and VHL-CBC (Von HippelLindau Cul2/elongin B/elongin C complex). Different numbers of the ubiquitination en zymes involved in each step indicates that the specificity of activity differs among E1, E2 and E3. While only one E1 has been identified in yeast, there are at least 13 E2s in yeast and 25 E2s in mammalians discovered so far. In contrast, the most diversity of the ubiquitination enzymes comes with E3, the ubiquitin ligase. To date, more than a hundred E3s have been recognized, implicating their role in substr ate selection. In order to achi eve strictly regulated protein degradation, different proteins are recognized by distinct ubiq uitin ligases, targeting them to the ubiquitination system. The substrate sp ecificity is reached by the pool of E3s that recognize only certain proteins for ubiquitina tion and in the case of polyubiquitination, often lead to substrate degradation. Proteins can be modified by a single ubiquitin conjugated to one or more of its lysine residues, so called monoubiquitination. While the functional sign ificance of protein monoubiquitination is still elusive, it has been implicated in the regulation of biological processes such as endocytosis lysosomal degradation and chromatin
35 remodeling. With the presence of an internal lysine in the ubiquitin protein, a polymer of ubiquitin could also be constructed thr ough isopeptide bond between the C-terminus of one ubiquitin and the internal lysine 48 residue of the other (Pickart 1997). When conjugated to a target protein, the ubiquitin chain containing no less than four ubiquitin polymers is often recognized as a signal for proteasome destruction (Thrower, Hoffman et al. 2000). The destination of destruction for the polyubiquitinated proteins is the 26S proteasome. It is composed of two majo r subunits: one is the 20S core particle that carries out the proteolytic activ ity; the other is the 19S regulatory particle, regulating the functionality of the core particle. Both 20S core particle and 19S regulatory particles contain multiple subunits The 20S core is a cylindrical barrelshaped structure composed of four rings st acked upon each other. The two rings on both ends are called subunits, which gate the entry of the substrates for degradation. Without the regulation by the 19S caps, the multiple-ubiquitinated proteins cannot enter the catalytic core for degradation. The 19S regulatory particle carries out multiple functions: selecting target, pr ocessing substrates and lead ing the entrance of target proteins to the catalytic core Across species and between human and yeast, up to 4070% of sequence identity was found in some of the proteasome subunits, indicating ubiquitination system as one of the most conserved systems in eurkaryotes. Although targeting substrate proteins fo r proteolysis is a key function of ubiquitination, it is becoming clear that ubiquitination plays much more complicated roles in myriad of biological processes, including DNA da mage response, translation regulation, kinase regulation and endocytosi s. The functional mechanism of these processes has just started to be appreciated.
36 Posttranslational modification by ubiquitin is a reversible process. A group of thiol protease enzymes namely DUBs tightly control the removal of ubiquitin by dissembling poly-ubiquitin chains and cleavi ng ubiquitin off the protein substrates (Wilkinson 1997). This ensures the recycl ing of the ubiquitin and prevents the accumulation of ubiquitin chains at the pr oteasome, which will cause disordered proteasome functions. DUBs also add to the layer of ubiquitination regulation by affecting the steady state level of the ubiquitin-conjugated pr oteins since the deubiquitination of the protein can rescue it from destruction by proteasome degradation. Sumoylation process Posttranslational modification is an e fficient way to modulate protein-protein interaction, stability, function or localiza tion. Small molecules such as ubiquitin, phosphate, and acetyl conjugate to specific resi dues on target proteins and modify their localization, stability or ac tivity. Among these small molecules participated in protein posttranslational modifications, the func tional significance of sumo is less well understood. Sumo, or small ubiquitin like mol ecule, is a 101 amino acid protein that bears homology with ubiquitin (Boddy, Howe et al 1996; Matunis, Coutavas et al. 1996; Okura, Gong et al. 1996; Shen, Pardington-Purt ymun et al. 1996; Kamitani, Nguyen et al. 1997; Mahajan, Delphin et al 1997; Tsytsykova, Tsitsikov et al. 1998). Despite the limited 18% overall sequence homology, sumo and ubiquitin share very similar threedimensional structure and prot ein folding. Nevertheless, si gnificant differences were
37 found between sumo and ubiquitin on the N-te rminus structure as well as the surface charge of both proteins. The process of sumoylation parallels that of ubiquitination. Both sumoylation and ubiquitination are processes of stepwise conjugation of the small molecules to the target proteins, involving functional specific enzyme s in each step. Aos1/Uba2 in yeast or SAE1/SAE2 in human are E1 activating en zymes, activating the sumo or ubiquitin molecule; E2 conjugates sumo or ubiquitin to E3 or directly to the substrates. Unlike multiple E2s for ubiquitination, sumoylation has a unique E2, Ubc9. Both sumo/ubiquitin activation and conjugation step s are ATP-dependent. In the process of ubiquitination, hundreds of identified E3 ubiquitin ligases provide the substrate specificity needed to determine the fate and on many occasions, the degradation of the target proteins. On the other hand, the E3 ligases for sumoylati on process have just emerged and await further investigation. To date, two classes of sumo E3s have been discovered. The first class consists only one member, the RanBP2, which is a component of the nuclear pore complex. The other class includes siz1 and siz2 in the S. cerevisiae as well as the PIAS family, which was first r ecognized as co-regulators of the JAK-STAT pathway. Sumoylation functions A consensus sumoylation site has been identified. The presence of consensus sequence KXE, where represents a large hydrophobic amino acid and K stands for the sumo conjugating lysine residue i ndicates potential m odification by sumo conjugation. Although the number of identified su mo substrates is limited, they seem to
38 play important roles in diverse fields such as maintenance of genomic stability, DNA repair, gene transcription, signal tr ansduction and nuclear pore function. In 1999, two labs simultaneously discovered that p53 is sumoylated at lysine 386 and this modification activates p53 transcriptional activity (Gostissa, Hengstermann et al. 1999; Rodriguez, Desterro et al. 1999). The requirement for sumo E3 ligase in p53 sumoylation is controversial. Original findings claimed that p53 in vitro sumoylation requires only sumo-1, E1 and E2. However, it was later discovered that PIAS directly binds both p53 and Ubc9 and stimulates p53 sumoylation in vitro (Schmidt and Muller 2002) MDM2 was also proved sumoylated a nd the process is stimulated by ARF (Xirodimas, Chisholm et al. 2002). The functional significance of MDM2 sumoylation remains unclear. RanGAP1 is a regulator of nuclear pore transportation. Unmodified RanGAP1 localizes in the cytoplasm while sumoylated RanGAP1 relocalizes to the nuclear pore complex and associates with a nuclear pore complex subunit, RanBP2, facilitating nuclear import (Matunis, Coutavas et al. 1996 ). PML, promyelocytic leukemia protein is another example of sumoylati on modifying protein subcellular localization (Sternsdorf, Jensen et al. 1997; Kamitani, Nguyen et al 1998; Muller, Matunis et al. 1998). Upon sumo conjugation, PML relocalizes to th e nuclear bodies wher e it recruits Daxx, inhibiting Daxx-mediated transcri ption repression of target ge nes (Kim, Park et al. 2003). Sumoylation can also regulate transc ription by recruiting coactivators or corepressors to sumo-modified transcription factors. For exam ple, the sumoylation of the transcription coactivator p300 leads to the recruitment of HDAC6, the histone deacetylase (Girdwood, Bumpass et al. 2003).
39 Various proteins participating in the maintenance of genomic stability are posttranslational modified by sumo. These include topoisomerase I and II, thymine-DNA glycosylase enzyme (TDG) as well as the proliferating cell nucl ear antigen (PCNA) (Mao, Desai et al. 2000; Mao, Sun et al 2000; Saitoh and Hinc hey 2000; Hardeland, Steinacher et al. 2002; Stelter and Ulrich 2003) Sumo modification affects their proteinprotein interaction, subcellular localization or activity. There are cross talks between sumoylation, ubiquitination, phosphorylation and acetylation. For example, the phosphorylation of NF-kB stimulates its ubiquitination while antagonizing sumo conjugation (H uang, Wuerzberger-Davis et al. 2003). Noticeably, all acetylation, ubiquitination and sumoylation take place solely on lysine residues. However, modifications on common lysine residues have had both antagonistic and synergistic functions on certain substrates. We identified MDMX as another substrate of sumo conjugation. Similar to MDM2, the sumoylation of MDMX can also be stimulated by ARF. Unlike MDM2 however, MDMX is sumoylated in an ARF inde pendent fashion. In vitro assay revealed that the presence of sumo-1, E1 and E2 are sufficient for MDMX sumoylation. Major sumoylation sites of MDMX were identified to two lysines. However, the sumoylation status of MDMX does not affect its associa tion or function with ei ther p53 or MDM2. Thus, the physiological significance of MDMX sumoylation remains elusive. Sumoylation is a reversible process Sumoylation process is dynamic and reversible. Sumo-specific proteases participate in both the process of sumo ma turation and the cleavage of sumo from the
40 substrates. Sumo is synthesized as precu rsors and processed at the carboxy-terminus before entering the cycle of sumo activation and conjugation by sumo-specific proteases. In yeast, the Ulp proteases are crucial fo r viability. The Ulp homologs in mammalian cells are a family of enzymes termed SENP (Sentrin-specific prot ease). To date, four SENP family members have been characterized (Gong, Millas et al. 2000; Kim, Baek et al. 2000; Nishida, Tanaka et al. 2000; Nish ida, Kaneko et al. 2001). Even though the substrate specificity of different SENPs are not known, the fact that these SENP family members distribute in different subcellular compartments indicates that there might be functional specificity of these sumo proteases. SENP1 localizes in nuceloplasm and nuclear bodies; SENP2 is concentrated at nucl ear pore; SENP3 is contained in nucleolus and SENP6 resides in cytoplasm (Gong, Mill as et al. 2000; Kim, Baek et al. 2000; Nishida, Tanaka et al. 2000; Nishida, Ka neko et al. 2001; Bailey and O'Hare 2002; Hang and Dasso 2002).
41 Materials and Methods Cell Lines MDM2/p53 double-null mouse embryo fibrobl ast 174.1 is a kind gift from Dr. Guillermina Lozano (McMasters, Montes de Oca Luna et al. 1996). ARF/p53 doublenull mouse embryo fibroblast is a kind gift from Dr. Gerry Zambetti. All cell lines including H1299 (non-small cell lung carci noma, p53-null), U2OS (osteosarcoma, p53 wild type, ARF-deficient), HFF (passage 20 human primary foreskin fibroblast) and MCF7 (breast cancer, p53 wild type) cells we re maintained in DMEM medium with 10% fetal bovine serum and 100 units/ml penicill in and streptomycin as well as 1% nonessential amino acids (cellgro). Plasmids and Reagents Human MDMX cDNA (myc-MDMX, MDMX-myc -His6) were kindly provided by Dr. Donna George (Sharp, Kratowicz et al. 1999). The myc epitope tag fused to the MDMX cDNA was removed by subcloning to generate a non-tagged MDMX. A Flag epitope tag was added to the C terminus of mycMDMX by PCR to create double-tagged mycMDMX-Flag. His6-ubiquitin expression plasmid was kindly provided by Dr. David Lane
42 (Xirodimas, Chisholm et al. 2002). ARF deletion mutants were kind gifts from Dr. Yue Xiong (Zhang and Xi ong 1999). MDM2 deletion mutants were described previously (Chen, Marechal et al. 1993; Zhang an d Xiong 1999). MDMX deleti on mutants were generated by PCR and subcloned into pcDNA3.1 (+) vector. Point mutants of MDM2 and MDMX were created using the Quick Change kit (Str atagene). All p53, MDM2, MDMX and ARF constructs used in this stu dy were of human origin. Mono clonal antibody 8C6 against human MDMX was generated in our laboratory and reacts specifically with MDMX in a region between residues 101-393 (Li, Chen et al. 2002). Aden ovirus expressing human ARF was kindly provided by Dr. Yue Xiong and used as described previously (Lu, Lin et al. 2002). U2OS stable cell line expressing tetracycline-repressible human MDM2 was created by cloning MDM2 cDNA into the p UHD15.1 vector and cotransfecting with pUHG10.3 (Gossen and Bujard 1992). Western Blot Cells were lysed in lysis buffer [50 mM Tr is-HCl (pH 8.0), 5 mM EDTA, 150 mM NaCl, 0.5% NP40, 1 mM PMSF], centrifuged for 5 mins at 10,000 g, and the insoluble debris were discarded. Cell lysate (10-50 g protein) was fractio nated by SDS-PAGE and transferred to Immobilon P filters (Millipore) using a semi-dry electro-blotting apparatus for 1 hr. The filter was blocked for 1 hr with washing buffer containing phosphatebuffered saline (PBS), 5% non-fat dry milk and 0.1 % Tween-20. The filter was then incubated for 2 hrs with primary antibodies diluted in washing buffer. MDM2 was
43 detected by 3G9 with 1:40 dilution (Chen, Marech al et al. 1993); p53 was detected by DO-1 with 1:10,000 dilution (Pharmingen); ARF was detected by 14PO 2 (Neomarkers) with 1:2000 dilution and MDMX was detected by 8C6 with 1:40 d ilution. The filter was then washed four times (5 min each) with PBS containing 0.1 % Tween-20. Next, bound primary antibodies were conjugated with secondary antibody IgG goat-anti-mouse HRP by incubating the filter with the secondary antibody for two hrs in the washing buffer containing phosphate-buffered saline (PBS), 5% non-fat dry milk and 0.1 % Tween-20. The filter was washed and developed us ing ECL-plus reagent (Amersham). Immunoprecipitation Cells were lysed in lysis buffer [50 mM Tr is-HCl (pH 8.0), 5 mM EDTA, 150 mM NaCl, 0.5% NP40, 1 mM PMSF], centrifuged for 5 mins at 10,000 g, and the insoluble debris were discarded. To detect the binding between MDM2 and MDMX, two mg of total cellular protein was immunoprecipitated using 100 l of anti-MDM2 monoclonal antibody 3G9 and 20 l packed protein A-Sepharose beads (Sigma) for 3 hours at 4 o C. The beads were then spinned down and washed three times with lysis buffer, boiled in SDS sample buffer and the immunoprecipita te was fractionated by SDS PAGE followed by western blot.
44 Transfection Calcium phosphate transfection was perfor med in H1299 and U2OS cells for their high transfection efficiency. In transient transfection, 2x10 6 cells were seeded into 10 cm tissue culture dishes for 24 hrs. For each transfection, a total amount of 40g of plasmid DNA was mixed with 450l of H 2 O, 125 mM calcium chloride and 500 l HEPES (0.28 M NaCl, 0.05 M HEPES, 1.5mM CaCl 2 ). Immediately after mixing, the mixture was added to the cells and incubated for 16 hours. Transfected cells were washed 5 times with PBS, refed with complete medium and incubated for another 24 hours before harvest. In a stable transf ection, 48 hours after transfection, the cells were drug selected by complete medium containing 750 g/ml G418 for two weeks. Dr ug-resistant colonies were either pooled or cloned. Lipofectamine transfection was performed in MEF cells for their low transfection efficiency with calcium phosphate transfection (invitrogene). 2x10 5 cells were seeded into 6 cm tissue culture dishes for 24 hrs, wash ed with 3 ml and refed with 2 ml of serumfree medium. For each transfection, a total amount of 4g of plasmid DNA was mixed with 250 l of serum-free medium and 10 l of lipofectamine plus reagent and incubated for 15 min at room temperature. Pre-mix 10 l of lipofectamine with 250 l serum-free medium and mix with the DNA solution. The mixture was incubated for another 15 min before added to the cells. After 3 hrs in cubation, the cells were refed with complete medium.
45 Immunofluoresence Staining Approximately 50,000 cells were plated on ch amber slides and incubated for 18 hours before staining. The cells were fixed by 1:1 ratio acetone/methanol for 3 minutes at RT. The slides were immediately rinsed two times with washing buffer (PBS with 0.1% Triton-X 100) and then incubate d in washing buffer twice for 5 minutes each. The slides were blocked with blocking buffer (10% nor mal goat serum in PBS) for 20 minutes. Primary antibody 8C6 is diluted 1:40 in the bl ocking buffer and incubated with the cells for 2 hrs at RT. Then the slides were rins ed and incubated in washing buffer twice for 5 minutes each. The secondary antibody goa t anti-mouse FITC was diluted 1:400 in blocking buffer and incubated with the cells in dark for 2 hrs. The slides were then washed and mounted. In vivo Ubiquitination Assay H1299 and U2OS cells in 9 cm plates were transfected with combinations of 1 g GFP expression plasmid, 5 g His6-ubiquitin expression plasmid, 1-5 g human MDMX, 5 g MDM2 and 5 g ARF expression plasmids using conventional calcium phosphate precipitation method. Thirty-two hours after transfection, cells from each plate were collected into two aliquots. One aliquot (10% ) was used for conventional western blot to confirm expression and degradation of transf ected proteins. The remaining cells (90%)
46 were used for purification of His6-tagged proteins by Ni 2+ -NTA beads. The cell pellet was lysed in buffer A (6 M guanidinium-H Cl, 0.1 M Na2HPO4/NaH2PO4, 0.01 M TrisCl pH8.0, 5 mM imidazole, 10 mM -mecaptoethanol) and incubated with Ni 2+ -NTA beads (Qiagen) for 4 hours at room temperatur e. The beads were washed with buffer A, B (8 M urea, 0.1 M Na2PO4 /NaH 2 PO 4 0.01 M Tris-Cl pH8.0, 10 mM -mecaptoethanol), C (8 M urea, 0.1 M Na 2 PO 4 /NaH 2 PO 4 0.01 M Tris-Cl pH6.3, 10 mM -mecaptoethanol), and bound proteins were eluted with buffer D (200 mM imidazole, 0.15 M Tris-Cl pH6.7, 30% glycerol, 0.72M -mercaptoethanol, 5% SDS). The el uted proteins were analyzed by western blot for the presence of conjugated MDMX by 8C6 antibody. In vitro Ubiquitination Assay GST-MDM2 full-length and deletio n mutants were expressed in E. coli and purified by binding to glutathione-agarose beads. The substrate MDMX was produced by in vitro translation in rabbit reticulocyt e lysate using the TNT system (Promega) in the presence of 35S-methionine. 15 l packed beads loaded with ~2 g GST fusion proteins were incubated with 4 l of MDMX in vitro translation product, 250 ng GST-Ubc5Hb (Boston Biochem), 250 ng purified rabbi t E1 (AG Scientific), 2 g His6-ubiquitin (AG Scientific), and 20 l reaction buffer (50 mM Tris pH7.5, 2.5 mM MgCl2, 15 mM KCl, 1 mM DTT, 0.01% Triton X-100, 1% glycerol, 4 mM ATP). The mixture was incubated at
47 37 C for 1 hr with shaking, bo iled in SDS sample buffer a nd fractionated by SDS-PAGE. The gel was dried and MDMX was detected by autoradiography.
48 MDM2 Promotes Ubiquitination and Degradation of MDMX Abstract The p53 tumor suppressor is regulated by MDM2-mediated ubiquitination and degradation. Mitogenic si gnals activate p53 by induction of ARF expression, which inhibits p53 ubiquitination by MDM2. Recent studies showed that the MDM2 homolog MDMX is also an important regulator of p53. We present evidence that MDM2 promotes MDMX ubiquitination and degradation by the pr oteasomes. This effect is stimulated by ARF and correlates with the ability of ARF to bind MDM2. Promotion of MDM2mediated MDMX ubiquitination requires the N-terminal domain of ARF, which normally inhibits MDM2 ubiquitination of p53. An intact RING domain of MDM2 is also required, both to interact with MDMX and to provide E3 ligase function. Increase of MDM2 and ARF levels by DNA damage, reco mbinant ARF adenovirus infection, or inducible MDM2 expression le ads to proteasome-mediated down-regulation of MDMX levels. Therefore, MDMX and MDM2 are co ordinately regulated by stress signals. The ARF tumor suppressor differen tially regulates the ability of MDM2 to promote p53 and MDMX ubiquitination and activates p53 by targeting both members of the MDM2 family.
49 Results MDM2 stimulates MDMX ubiquitination and degradation MDM2 is rapidly degraded by the prot easome due to self-ubiquitination (Fang, Jensen et al. 2000). In contra st, little is known about the regulation of MDMX turnover. We tested whether it is s ubjected to ubiquitination in vivo To detect MDMX ubiquitination, we coexpressed MDMX and Hi s6-ubiquitin by transient transfection into H1299 cells. MDMX conjugated to His6-ubiquitin was purified by Ni 2+ -NTA beads under denaturing conditions and detected by western blot with MDMX-specific monoclonal antibody 8C6 (Li, Chen et al. 2002). The result showed that a small portion of MDMX was conjuga ted to ubiquitin in vivo (mainly mono-ubiquitination based on slightly slower mobility than unmodified MDMX). Transfection of MDM2 strongly stimulated MDMX poly-ubi quitination (Figure 4).
Figure4. MDM2 stimulates MDMX polyubiquitination. MDMX wascoexpressed with His6 -ubiquitin in H1299 cells by transient transfection.His6-ubiquitinated proteins were purified, and MDMX was detected byWestern blotting with 8C6 antibody. Interestingly, cotransfection of MDM2 and ARF did not inhibit but rather stimulated MDMX polyubiquitination (Figure 5). MDM2 expression level was also increased in the presence of ARF (due to stabilization) and may account for the increase in MDMX ubiquitination. Direct western blot also showed that MDMX expression level was significantly reduced when coexpressed with MDM2 (Figure 5), suggesting that ubiquitination of MDMX led to subsequent degradation by the proteasomes. 50
Figure 5. MDM2 and ARF synergistically stimulate the polyubiquitination and degradation of MDMX. MDMX was coexpressed with His6 -ubiquitin in H1299 cells by transient transfection. His6 -ubiquitinated proteins were purified, and MDMX was detected by Western blotting with 8C6 antibody (top panel). The lower panels are control Western blots of nonpurified cell lysate (10% of the amount used for Ni-NTA purification). 51
Consistent with this hypothesis, treatment with proteasome inhibitor MG132 for 4 hours before cell harvest partially prevented MDMX down-regulation by MDM2 cotransfection (Figure 6). In the absence of cotransfected MDM2, MDMX was relatively stable and did not accumulate in the presence of MG132. Figure 6. MDM2 stimulates MDMX degradation by proteasome pathway. H1299 cells transfected with the indicated plasmids were treated with 30 M MG132 for 4 hours prior to Western blot analysis for MDMX. Transfection efficiency was verified by expression of cotransfected GFP plasmid. 52
To determine whether MDM2 and ARF are required for the basal ubiquitination of MDMX, the assay was repeated using MDM2/p53-null or ARF/p53-null mouse fibroblasts. The results showed that mono-ubiquitination of MDMX occurred in the absence of endogenous MDM2 (Figure 7) and ARF (data not shown). Therefore, MDMX can undergo ARF and MDM2-independent mono-ubiquitination. Cotransfection of MDM2 or MDM2 and ARF also stimulated MDMX poly-ubiquitination in mouse cells. 53 Figure 7. MDMX can be ubiquitinated independent of MDM2. MDM2/p53-null 174.1 cells were transfected with the indicated plasmids. Ni-NTA-purified ubiquitinated MDMX and cell lysate were run on the same gel to show the molecular weight shift due to MDM2-independent monoubiquitination.
54 MDM2 RING domain is sufficient fo r promoting MDMX ubiquitination Next, a panel of MDM2 mutants was test ed for ubiquitination of MDMX. The results showed that the RING domain of MDM2 was important for promoting MDMX ubiquitination (representative mutants are shown in Figure 8 & 10). MDM2 deletion mutants without the p53-binding or ARFbinding domains were still capable of stimulating MDMX ubiquitination, indicating th at these regions were not essential for this function. An MDM2 NLS point mutant (182R) that prevents nuclear translocation was also able to ubiquitinate and degrad e MDMX. Furthermore, the C terminal 130 amino acid of MDM2 containing the RING domain was sufficient to ubiquitinate MDMX (Figure 8 & 10). Deletion of MD MX C terminal RING domain ( 394-490) also prevented poly-ubiquitination by MDM2 (F igure 9 & 10). Therefore, these results showed that poly-ubiquitin ation of MDMX by MDM2 required formation of MDMXMDM2 complex through RING domain-mediate d heterodimerization and possibly the E3 function of MDM2.
Figure 8. MDM2 RING domain is required for stimulating MDMX polyubiquitination. H1299 cells were transfected with His6 -ubiquitin, MDMX, and representative MDM2 mutants. MDMX ubiquitination and degradation were detected by Ni-NTA purification and Western blotting. 55
Figure 9. MDMX RING domain is required for its polyubiquitination and degradation by MDM2. H1299 cells were transfected with His6 -ubiquitin, MDM2, and MDMX mutants. MDMX ubiquitination was detected by Ni-NTA p urification and Western blotting. 56
Figure 10. Diagram of MDM2 and MDMX mutants and summary of results. MDMXN has a deletion of residues 1 to 100. MDMXC has a deletion of residues 394 to 490. 57
Ubiquitination of MDMX requires MDM2 ubiquitin ligase function 1 To test whether the E3 function of MDM2 is required for poly-ubiquitination of MDMX, we mutated three key cysteine and histidine residues in the RING domain of MDM2 that a previous study found to be important for E3 function (438C-L, 457H-S, 478C-S, Figure 11 & 12) (Fang, Jensen et al. 2000). These mutations prevented poly-ubiquitination of MDMX in the presence of ARF (Figure 11 & 12). 58
Figure 11. Ubiquitination of MDMX requires MDM2 E3 function. (Upper panel) Position of point mutations in the MDM2 RING domain. C and H residues important for zinc coordination are underlined. (Lower panel) Ubiquitination of MDMX by MDM2 mutants was determined by transfection of H1299 cells. 59
60 When analyzed by MDM2 immunoprecipitatio n and MDMX western blot, the 438L and 478S mutants lost most or all of MDMX bindi ng, indicating that thes e mutations affected both E3 and MDMX binding functions (Figure 12 ). The 457S mutant was still capable of binding to MDMX, but the efficiency was si gnificantly lower than wild type MDM2 considering the higher expression levels of both proteins in the transfected cells due to lack of degradation (Figure 12). Since the absolute amount of 457S-MDMX complex was comparable to MDM2-MDMX, yet MDMX ubiquitination was much weaker in the 457S-MDMX combination, we concluded that th e E3 function of MDM2 was critical for ubiquitination of MDMX. However, it was cl ear from these results that MDMX-binding and E3 functions were both dependent on the integrity of MDM2 RING domain and could not be completely separated by these mutations.
Figure 12. Binding of MDMX to MDM2 point mutants. MDMX and MDM2 mutants were cotransfected into H1299 cells and analyzed by MDM2 immunoprecipitation followed by MDMX Western blotting. 61
62 ARF N terminal domain is requir ed to cooperate with MDM2 It is well established that ARF is an inhibitor of MDM2 E3 ligase function on p53 (Zhang and Xiong 2001). The results in Figur e 5 showed that its effect on MDM2mediated MDMX ubiquitination was opposite to that of p53. Therefore, we analyzed a panel of ARF deletion mutants to identify the domain required for stimulating MDMX ubiquitination. The results showed that the N terminal domain of ARF containing the MDM2-binding region was required for stim ulating MDMX ubiquitination in the presence of MDM2 (Figure 13). Deleting the nucleolar targeting si gnal between 82-101 did not prevent stimulation of MD MX ubiquitination (Zhang and Xiong 1999). Interestingly, the N terminal region of ARF contains the MDM2 binding site and is sufficient to inhibit p53 ubiquitination by MDM2 (Midgley, Desterro et al. 2000). Therefore, ARF binding to MDM2 differentiall y regulates its abilit y to ubiquitinate p53 and MDMX.
Figure 13. The ARF N-terminal domain is sufficient to cooperate with MDM2. (Upper panel) H1299 cells were transfected with MDMX, a suboptimal amount of MDM2, and different ARF mutants. His6 -ubiquitin-conjugated MDMX was purified and detected by Western blotting. (Lower panel) Diagram of ARF mutants and summary of results. 63
64 Although ARF stimulated MDMX ubi quitination, MDM2 alone was sufficient to induce significant degradation of MDMX even in ARF-deficient cells (data not shown), indicating that ARF is not essential for this process. To observe stimulation of MDMX degradation by ARF in H1299 cells requires careful titration to limit the amount of MDM2 plasmid (data not shown). The effect of ARF on MDMX degradation was more evident and reproducible in mouse cells (Figur e 7), suggesting that cell type may affect the effect of ARF in th is transfection assay. Regulation of endogenous MDMX level by ARF and MDM2 To determine whether endogenous MDMX expression level was regulated by ARF, ARF-deficient U2OS cells were inf ected with ARF adenovirus to restore ARF expression. As expected, expr ession of ARF resulted in strong induction of p53 and MDM2 levels. This was accompanied by reducti on of MDMX expression (Figure 14).
Figure 14. Overexpression of ARF down regulates MDMX level. U2OS cells were infected with ARF adenovirus at the indicated multiplicity for 24 hours and analyzed for MDMX, MDM2, and p53 expression levels by direct Western blotting. 65
In the second experiment, primary human foreskin fibroblasts (HFF) were treated with ionizing radiation to induce MDM2 expression. DNA damage resulted in induction of p53 and MDM2 expression levels in HFF and corresponding reduction of MDMX expression (Figure 15). Reduction of MDMX level was also observed after induction of MDM2 expression by irradiation of U2OS cells, or induction of MDM2 by activating a temperature-sensitive p53 mutant in H1299 cells (data not shown). Figure 15. DNA damage induces MDMX degradation. HFF cells were treated with 10 Gy of gamma radiation for 6 hours and 10 hours, respectively before collection. The levels of MDMX, MDM2, p53, ARF and actin were determined by Western blotting. 66
To determine whether reduced MDMX level was due to degradation by the proteasomes, irradiated HFF cells were treated with MG132 for 4 hours prior to analysis. The decrease in MDMX level after irradiation was blocked by MG132 (Figure 16), suggesting it was mediated by proteasomes. In contrast, the p53-null H1299 cells did not express elevated MDM2 after irradiation and MDMX level was not reduced. Figure 16. Induction of MDM2 results in MDMX degradation by pro-teasomes. HFFs and H1299 cells were treated with 10 Gy of gamma radiation for 6 hours. MG132 (30 M) was added immediately after radiation. The MDMX level was determined by Western blotting. 67
To determine if the reduction of MDMX protein level after DNA damage is caused by the change in MDMX mRNA level, we performed Northen blot. The analysis using several cell lines confirmed that MDMX mRNA level was not reduced by radiation (Figure 17). These results suggested that ionizing radiation induced proteasome-mediated degradation of MDMX, which may be due to p53 activation and induction of MDM2 expression. Figure 17. Gamma irradiation does not reduce the MDMX mRNA level. Cells were treated with gamma radiation, and 6 hours later total RNA was isolated and hybridized with an MDMX cDNA probe in a Northern blotting analysis. 68
69 To test whether change in MDM2 level independent of p53 activation or radiation was sufficient to regulate MDMX level, we created a U2OS cell line expressing tetracycline-repressible MDM2. Induction of MDM2 expression by removal of tetracycline correlated with decrease of MDMX level (Fi gure 18, lanes 5 and 7), which was restored by a 4 hour MG132 treatment. Th erefore, increase of MDM2 expression alone was sufficient to pr omote MDMX degradation. However, the moderate decrease of MDMX was not proportional to the strong i nduction of MDM2, suggesting that other factors may limit the efficiency of MDMX degradation by MDM2.
Figure18. Induction of MDM2 correlates with MDMX degradation through proteasome pathway. U2OS cells stably expressing tetracycline-repressible MDM2 were cultured in the presence or absence of 0.5 g of tetracycline/ml for 20 hours. MG132 was added for the last 4 hours. MDMX and MDM2 levels were determined by Western blotting. Epitope tagging prevents MDMX degradation by MDM2 Previous studies using epitope-tagged MDMX did not indicate an effect of MDM2 on MDMX degradation (Sharp, Kratowicz et al. 1999; Jackson, Lindstrom et al. 2001). We also had inconsistent results in early experiments when tagged and non-tagged versions of MDMX were used interchangeably. A direct comparison of non-tagged MDMX with several versions of epitope tagged MDMX revealed that addition of c-myc or Flag epitopes to the C terminus of MDMX significantly inhibited poly-ubiquitination by 70
MDM2 and completely prevented degradation (Figure 19). Therefore addition of epitopes to the RING domain of MDMX interferes with ubiquitination and degradation by MDM2. Because of this finding, all of the results described above were from repeat experiments using the non-tagged version of MDMX. Figure 19. MDMX degradation is inhibited by C-terminal epitope tags. (A) H1299 cells were transfected with MDMX with different epitope tags at the N or C terminus and subjected to Ni-NTA purifi-cation and Western blotting.The MDMX-myc-His6 protein was com-pletely recovered by Ni-NTA beads asexpected. (B) Diagrams of epitope-tagged MDMX and summary of results. C-terminal epitope tags block polyubiquitination and degradation by MDM2. 71
MDM2 promotes MDMX poly-ubiquitination in vitro In order to confirm the role of MDM2 as an E3 for the poly-ubiquitination of MDMX, we established an in vitro ubiquitination assay for MDMX. In this assay, GST-MDM2 was able to stimulate the poly-ubiquitination of in vitro translated MDMX in the presence of purified E1 and E2 (GST-UbcH5b) (Figure 20). Furthermore, the C terminal fragment of MDM2 containing the RING domain was sufficient to promote MDMX ubiquitination, whereas an MDM2 N-terminal fragment had no effect. These results corroborate the in vivo ubiquitination results of Figure 8 and demonstrate that MDM2 functions as an ubiquitin E3 ligase for MDMX. 72
Figure 20. MDM2 promotes MDMX polyubiquitination in vitro. GST-MDM2, GST-RING (containing MDM2 residues 290 to 491), and GST-N (containing MDM2 residues 1 to 150) were purified using glutathione agarose beads. Loaded beads were incubated with in vitro-translated MDMX in the presence or absence of E1 and E2 in an ubiquitination reaction as described in Materials and Methods. Poly-ubiquitinated MDMX appears as a high-molecular-weight smear above the unmodified MDMX band. 73
74 Discussion The results described above show that MDMX is subjected to modification by ubiquitin conjugation. Interaction with MDM2 leads to strong s timulation of MDMX poly-ubiquitination and degrad ation by proteasomes. MDM2 appears to function as an ubiquitin E3 ligase in this process, since the MDM2 457S RING domain mutant that prevents p53 or self-ubiquitination also ab rogates ubiquitination of MDMX while still partially retaining MDMX binding. Furtherm ore, purified recombinant GST-MDM2 is able to stimulate MDMX poly-ubiquitina tion in cell-free conditions. MDMX also undergoes basal mono-ubiquitinati on independent of MDM2, suggesting interaction with other ubiquitination factors. It is noteworthy that previous studies using epitope-tagged MDMX did not observe an effect of MDM2 on MDMX degradation (Sharp, Kratowicz et al. 1999; Jackson, Lindstrom et al. 2001).We found that MDMX tagged with c-myc or FLAG epitopes at the C terminus were resi stant to ubiquitination and degradation by MDM2, possibly due to altered RING doma in conformation. Therefore, use of nontagged or N terminally tagged MDMX appears to be critical for detecting its efficient ubiquitination and degradation by MDM2. Surprisingly, ARF stimulates MDMX ubiquitination by MDM2, which is opposite to its effect on p53 ubiqui tination. The stimulating eff ect of ARF is carried out by the N terminal domain that normally in hibits MDM2 ubiquitination of p53. Therefore, ARF interaction with MDM2 differentially affects its E3 function toward different
75 substrates rather than inactivating its E3 f unction in general. This effect is plausible because ARF does not directly interact with the RING doma in of MDM2, which may be involved in recruiting E2. Bi nding of ARF to the acidic domain of MDM2 may simply alter its ability to properly orient E2 for the transfer of ubiquitin to certain substrates. In the case of p53, ubiquitin conjugation is blocked. However, in the case of MDMX, ubiquitination is not affected or even stimulated. ARF in creases MDM2 expression level by stabilization, which may account for incr eased MDMX ubiquitination. However, ARF may also qualitatively stimulate the E3 activity of MDM2 toward MDMX. These results suggest that ARF is a more effective and multi-functional regulator of the p53 pathway than previously appreciat ed. By blocking the ability of MDM2 to ubiquitinate p53, ARF is sufficient to induce p53 stabilization and activation. Additionally, by increasing MDMX ubiquitinati on, ARF also simultaneously promotes the elimination of another important inhi bitor of p53. The combination of these two functions should result in mo re potent activation of p53. Because p53 activation induces MDM2 expression, higher MDM2 level should further cooperate with ARF to degrade MDMX, resulting in a positive feedback eff ect. Activation of p53 by other stress signals such as DNA damage also leads to strong induction of MDM2 expression. Increased MDM2 level may in fact f acilitate p53 activation by de grading MDMX, while its own interaction with p53 is temporarily bl ocked by phosphorylation (Figure 14, 15 & 16). The MDM2 negative feedback loop is an important mechanism in p53 regulation. However, knock out experiments showed that MDMX is also critical for regulating p53 (Parant, Chavez-Reyes et al. 2001; Finch, Donoviel et al. 2002; Miglio rini, Denchi et al. 2002). Recent studies suggested that MD MX and MDM2 depend on each other for
76 efficient inhibition of p53. MDM2 stimulat es MDMX by targeting it to the nucleus, whereas MDMX enhances MDM2s ability to degrade p53 in a highly dose-dependent manner (Gu, Kawai et al. 2002; Migliorini, Danovi et al. 2002). Th erefore, MDMX may be part of a more complicated MDM2 f eedback loop by regulating MDM2 function and acting as a receiver of other signals. The ability of MDM2 to degrade MDMX ensures that it also exists in a balance with MDM2 and p53. Although MDMX can inhibit p53 function, the dependence on a suicidal intera ction with MDM2 enables MDMX to be regulated by other signals. The MDMX degrad ation mechanism also ensures that stress signals can efficiently activate p53 by coordi nately regulating both p53 inhibitors (Figure 21).
Figure 21. MDMX is regulated by stress signals. MDM2 and MDMXcooperate to inhibit p53 activity in the absence of stress. Stress signalsactivate p53 by targeting both MDM2 and MDMX. Oncogenes such as E2F1and c-myc induce ARF expression, which activates p53 by inhibiting p53ubiquitination and promoting MDMX ubiquitination by MDM2. DNAdamage activates p53 and induces MDM2 expression, which in turn degradesMDMX and leads to further activation of p53. DNA damage also promotesMDMX nuclear translocation, which may expose it to degradation byMDM2 but also enables it to inhibit p53 more efficiently at certain stages o f the damage response. Recent studies showed that ectopically expressed MDMX is mainly localized in the cytoplasm and undergoes nuclear translocation by both MDM2-dependent and independent mechanisms (Gu, Kawai et al. 2002; Li, Chen et al. 2002; Migliorini, Danovi et al. 2002). Nuclear entry of MDMX is critical for its ability to inhibit p53 (Gu, Kawai et al. 2002; Migliorini, Danovi et al. 2002), suggesting that MDMX plays an active role in limiting p53 activation during DNA damage response. However, results presented in 77
78 this report show that DNA damage can i nduce degradation of MDMX, which would facilitate p53 activation. Th erefore, the net outcome of these regulatory events on the function of MDMX may depend on the level of damage and different stages of the stress response. Further work will be required to el ucidate the role of MDMX in p53 regulation during DNA damage and ot her stress conditions. In summary, our findings reve al that MDMX expression le vel is directly regulated by MDM2-mediated ubiquitination. The result s also identify ARF as a more dynamic regulator of MDM2 ubiquitin ligase function and suggest a potential mechanism of MDMX regulation by a variety of stress si gnals through MDM2 and ARF. Elucidating how MDMX ubiquitination and degradation is regulated should lead to better understanding of its role in the p53 pathway and tumorigenesis.
79 MDMX forms trimeric complex with MDM2 and ARF Abstract ARF is a tumor suppressor that functi ons mainly through inhibition of MDM2, which is the major negative re gulator of p53. ARF directly binds to MDM2 and inhibits its E3 ubiquitin ligase activity toward p53 and MDM2 itself, thus stab ilizes both proteins. Furthermore, ARF relocates MDM2 from nuc leus to nucleolus, physically separates MDM2 from p53 and activates p53 transcriptiona l activity. Previous studies showed that ARF stimulates p53 and MDM2 sumoylati on. Our studies in this manuscript demonstrated that ARF stimulates both ubi quitination and sumoyla tion of MDMX while the ubiquitination of MDMX led to its degrad ation in an MDM2-dependent fashion. In this study, we showed that ARF stimulat es MDMX ubiquitination while stabilizing MDM2, which may account for the stimulation of MDMX ubiquitination by ARF. On the other hand, the binding betw een MDMX and ARF has been controversial. Here we demonstrated that MDMX binds to ARF in an MDM2-dependent fashion. Our previous data showed that ARF N-terminus stimula tes MDMX ubiquitination and sumoylation. We found that ARF N terminus is also require d for binding to MDMX in the presence of exogenous or endogenous MDM2.
80 Introduction Retinoblastoma protein (Rb) and p53 are tw o tumor suppressors that have been found to exhibit loss of function in large percentage of huma n tumors. The main function of Rb lies in its regulatio n of E2Fs, which promote cell cycle progression (Weinberg 1995; Dyson 1998). Oncogenic signals such as Ras, induce the phosphorylation of Rb through cyclinD-Cdks and free E2F from the sequestration of Rb, promoting entry into S phase (Sherr 2001). In this pathway, there is another tumor suppressor working in the upstream of this Rb-E2F pathway: p16INK 4a is a Cdk inhibitor that inhibits cyclindependent kinase 4 and disrupts the cyclin-C dk complexes, thus maintaining Rb in the hypophosphorylated and active state in inhibiting the cell cycle progression promoted by E2Fs (Sherr and Roberts 1995; Weinberg 1995; Sherr 2001). The gene locus of p16 INK on chromosome 9 is mutated at high freque ncy in human tumors (Kamb, Gruis et al. 1994; Nobori, Miura et al. 1994). Later it was found that the necessity in tumor cells to mutate p16 locus lies in the fact that this lo cus encodes a second gene transcribed with an alternative reading frame, namely ARF (Se rrano, Hannon et al. 1993; Quelle, Zindy et al. 1995). Although ARF is completely unrelated to p16, the mutation of ARF also caused high frequency of tumor genesis independent of p16, supporting ARF as a tumor suppressor (Foulkes, Flanders et al. 1997; Ruas and Peters 1998; Sharpless and DePinho 1999). The mechanism of ARF induced cell cycl e arrest is distinct from that of p16 INK
A myriad of genetic stress including E1A, E2F, myc, ras and v-abl induce the expression of ARF, which activates the function of tumor suppressor p53 by harnessing p53 inhibitor, MDM2 (Kamijo, Weber et al. 1998; Pomerantz, Schreiber-Agus et al. 1998; Stott, Bates et al. 1998; Zhang, Xiong et al. 1998). Thus, ARF plays an important role in linking the Rb-E2F with p53-MDM2 signal transduction pathways. When mitogenic signal induced E2F activity reaches a critical threshold, ARF is induced and activates another checkpoint protein: p53, which executes cell cycle arrest or apoptosis. Figure 22. ARF connects Rb-E2F pathways to p53-MDM2 pathway 81
82 Results MDMX forms trimeric complex with MDM2 and ARF It is well recognized th at ARF binds to MDM2 in the acidic domain and inhibits MDM2s regulation of p53. However, the binding between MDMX and ARF has been controversial. In order to compare the bi nding efficiency between MDMX-ARF to that of MDM2-ARF, FLAG-tagged MDMX or FLAG -tagged MDM2 were cotransfected with myc-tagged ARF in H1299 cells. The FLAG-tagged proteins we re immunoprecipitated by M2 beads and co-precipitated ARF protei n was detected by Western blot using monoclonal antibody 14P02 against ARF (Fig 23). The result showed that ARF binds to MDM2. MDMX overexpression does not change the binding efficiency between MDM2 and ARF. No significant MDMX-ARF bi nding was detected in the absence of overexpressed MDM2. However, in the presence of overexpre ssed non-tagged MDM2, significant amount of ARF was pulled dow n by FLAG-MDMX, indicating that MDM2 mediated the binding between MDMX and ARF. MDM2 lacking the RING domain, which is the MDMX binding region, did not mediate the binding between FLAG-MDMX and ARF; neither did the two MDM2 muta nts lacking ARF binding regions. This experiment suggests that MDM2 bridges th e binding between MDMX and ARF and that a tertiary complex can be formed among MDM2, MDMX and ARF.
Figure 23. MDMX forms trimeric complex with MDM2 and ARF. FLAG-MDMX and FLAG-MDM2 were cotransfected with ARF expression plasmids in H1299 cells in the presence or absence of non-tagged MDMX or MDM2 mutants. FLAG-MDMX or FLAG-MDM2 was immunoprecipitated by M2 beads followed by western blot against FLAG-tagged proteins with rabbit-anti-FLAG antibody to show the level of immunoprecipitated MDMX and MDM2 (top panel). Co-immunoprecipitated ARF was detected by Western blot using mouse monoclonal antibody 14P02 against ARF (2 nd panel). Expression level of MDM2 was confirmed by western blot with rabbit-anti-MDM2 antibody (3 rd p anel). The level of ARF expressed was verified by direct western blot against ARF in the whole cell lysate (4 th panel). Transfection efficiency was verified by ex p ression of cotransfected GFP p lasmid ( bottom p anel ) 83
84 To further confirm the binding between MDMX and ARF, the experiment was repeated in H1299 cells with FLAG-p53 as a nother positive control. FLAG-MDMX and MDM2 were cotransfected in the presence or absence of overexpressed ARF. Previous studies showed that ARF binds directly to both MDM2 and p53 (Kamijo, Weber et al. 1998). Indeed, we observed that ARF was co-precipitated with either FLAG-p53 or FLAG-MDM2. However, significant ARF-MDMX binding was only detected when MDM2 was coexpressed, verifying MDM2s recruitment of ARF to MDMX (Fig 24).
Figure 24. MDM2 enhances the binding between MDMX and ARF. FLAGp 53, FLAG-MDMX and FLAG-MDM2were cotransfected in the presence or absence of ARF in H1299 cells. FLAG-tagged proteins were immunoprecipitated by M2 beads followed by western blot against FLAG-tagged proteins with rabbit-a-FLAG antibody to show the level of immunoprecipitated p53, MDMX and MDM2 (top panel). Co-immunoprecipitated ARF was detected by Western blot using mouse monoclonal antibody against ARF (2 nd panel). Transfection efficiency was verified by expression of cotransfected GFP plasmid (3 rd panel). The level of ARF expresse d was verified by direct western blot against ARF in the whole cell lysate (3 rd p anel) 85
86 MDM2 mediates the binding between MDMX and ARF In order to verify the binding betwee n MDMX and ARF, we used UO2S cells, which lacks endogenous ARF expression. MD MX was cotransfected with myc-tagged wild type ARF and ARF deletion mutants. Immunoprecipitation was performed using protein A beads and mouse monoclonal antibody against myc-tag. Co-precipitated MDMX was detected in western blot by ra bbit polyclonal antibody against MDMX. The result showed that MDMX was co-precipitated with wild type as well as the N-terminal mutants of ARF. ARF C terminus cannot bi nd to MDMX. ARF98Q is mutated at the NoLS sequence and retains the capability to bind MDMX (Fig 25). To see if MDMXARF binding is cell-type specific, we perf ormed the experiment in H1299 cells and observed similar result (data no t shown). We specu lated that the binding is mediated by endogenous MDM2 in both cell lines. To fi nd out if MDM2 is required for MDMX-ARF binding, we performed the same experiment in MEF 174.1, which is a p53/-/MDM2-/cell line that lacks endogenous MDM2. Resu lt showed that only in the presence of MDM2 does MDMX bind to ARF, indicating that the previously observed MDMX and ARF binding in U2OS and H1299 cells was possibly mediated by endogenous MDM2 in those cell lines (Fig 25). Thus, MDM2 bridges the bi nding between MDMX and ARF and the MDM2-binding region in ARF N-terminus is required for the indirect binding between MDMX and ARF.
U2OS 87 174.1(p53-/MDM2-/-) Figure 25. ARF N-terminus is required for binding MDMX. MDMX was cotransfected with myc-tagged wild type ARF or ARF deletion mutants as well as ARF nucleolus localization-deficient mutant into U2OS and MEF174.1(p53-/-MDM2-/-), respectively. Myc-ARF was immunoprecipitated by protein A beads and anti-myc-tag antibody. Co-precipitated MDMX was detected by western blot using rabbit polyclonal antibody against MDMX. The MDMX expression level was shown by western blot against MDMX using mouse monoclonal antibody 8C6 against MDMX from the whole cell lysate. Transfection efficiency was verified by expression of cotransfected GFP plasmid.
88 ARF does not affect the binding ef ficiency between MDM2 and MDMX Since we observed that MDM2 enhan ced the binding between MDMX and ARF (Fig 23-25), we speculated that the tertiary complex formation results in the increased binding affinity among the three proteins. To find out if MDMX and MDM2 binding efficiency can be changed by overexpression of ARF, H1299 cells were transiently transfected with fixed amount of MDMX and increasing amount of MDM2 and ARF. MDMX was immunoprecipitated with mous e monoclonal antibody 8C6 and the coprecipitated MDM2 was dete cted by rabbit polyclonal an tibody against MDM2. The result showed that the expr ession of ARF did not affect the binding efficiency between MDMX and MDM2 (Fig 26).
Figure 26. ARF does not affect the binding efficiency between MDM2 and MDMX. MDMX was cotransfected with increasing amount of MDM2 and ARF and was immunoprecipitated by protein A beads and mouse anti-MDMX antibody 8C6. Co-precipitated MDM2 was detected by western blot using rabbit p olyclonal antibody against MDM2 (top panel). The MDMX and ARF expression levels were shown by western blot from the whole cell lysate (middle p anels). Transfection efficiency was verified by expression of cotransfected GFP plasmid (bottom panel). 89
90 ARF stimulates MDM2-mediated MDMX ub iquitination by stabilizing MDM2 Previously, we showed that MDM2 promot es the ubiquitination and degradation of MDMX; additionally, ARF enhances ubiquitination of MDMX by MDM2. However, it remains unclear whether ARF stimulates MDMX ubiquitination through stabilizing MDM2 or through other unknown mechanisms. To find out if ARF can stimulate MDMX ubiquitination without stabilizing MDM 2, we performed the ubiquitination assay by titrating the amount of MDM2 and ARF. H1299 cells were transiently transfected with suboptimal amount of MDM2 and increa sing amount of ARF in the presence of MDMX and his6-ubiquitin. Th e result showed that MDMX ubiquitination is enhanced with increasing amount of ARF but in accordance with increasing amount of MDM2 stabilized by ARF overexpression (Fig 27). In concert with the finding that MDM2 mediates the binding between MDMX and ARF, our data suggests that ARF stimulates MDM2-mediated MDMX ubiquitination most lik ely by enhancing the steady state level of MDM2, which direct ly ubiquitinates MDMX.
Figure 27. ARF stimulates MDMX ubiquitination while stabilizing MDM2. (Upper panel) H1299 cells were transfected with MDMX, suboptimal amounts of MDM2, and increasing amount of ARF. His6-ubiquitin-conjugated MDMX was p urified and detected by Western blotting. (lower panels) The levels of MDM2, MDMX, ARF were detected by western blot from cell lysate. GFP level indicates transfection efficiency. 91
92 Discussion ARF directly binds p53 and MDM2 and re gulates their functions (Kamijo, Weber et al. 1998). However, the binding betw een MDMX and ARF has been controversial with one group showing no binding (Wang, Arooz et al. 2001) and the other group showed the otherwise (Jackson, Lindstrom et al. 2001). In this study, we confirmed the binding between MDMX and ARF in vivo, whic h is mediated by MDM2. However, the tertiary complex composed of MDMX, MDM2 and ARF failed in enhancing the binding affinity between MDMX and MDM2. This may be accounted for by the tight complex of the heterodimer between MDMX and MDM2 and the binding of which do not need further augmentation. ARF is not a stereotyped regulator on MDM2 and the interplay between the two proteins is still being explored. Nevertheless, previous studies showed that ARF works on all aspects of MDM2-mediated posttranslational modifications. MDM2 ubiquitination of the substrates including self-ubiquitination is inhibited by ARF, resulting in the stabilization of the MDM2 E3 substrates (Kamijo, Weber et al. 1998). In addition, in response to stress signals, p53 is acetylated by CBP/P300, which increases p53 stability. MDM2 suppresses p53 acetylation by binding to the acetyl tran sferase CBP/P300 and inactivating the enzymes. ARF abrogates MDM2s suppression on p53 acetylation, thus restores p53 stability and activity (Ito, Lai et al. 2001). Chen et al also demonstrated that MDM2 stimulates p53 sumoylation (Chen and Chen 2003). However, instead of
93 antagonizing MDM2s modification on p53, ARF enhances MDM2-mediated p53 sumoylation. Data from this manuscript demonstrated that ARF stimulates MDM2mediated MDMX ubiquitination and MDM2 -independent MDMX sumoylation. Therefore, ARF is a dynamic regulator of MDM2 and the rela tionship between p53, MDM2, MDMX and ARF awaits further exploration.
94 Modification of MDMX by Sumoylation Abstract MDMX is a homolog of MDM2 and is critical for regula ting p53 function during mouse embryonic development. MDMX level is regulated by MDM2-mediated polyubiquitination, which results in accelerated degradation af ter DNA damage or expression of ARF. In this report, we demonstrated that MDMX can be modified by conjugation to SUMO-1 both in vivo and in vitro We found that double mutation of two lysine residues; K254 and K379 abrogated MDMX sumoylation in vivo Experiments using the sumoylation-deficient MDMX mutant showed that it undergoes normal ubiquitination and degradation by MDM2, normal nuclear tr anslocation and degr adation after DNA damage, and inhibits p53 with wild type efficiency. Therefore, sumoylation is not required for several activities of MDMX under our assay conditions.
95 Introduction The p53 pathway is regulated by multiple mechanisms, which is critical for its ability to respond to stress and suppress tu mor formation. P53 turnover is regulated by MDM2, which functions as an ubiquitin E3 ligase to promote p53 ubiquitination and degradation by the proteasome. Stress signals such as DNA damage induce p53 accumulation by phosphorylation of p53 and MDM2 (Prives and Hall 1999). Mitogenic signals activate p53 by induction of the ARF tu mor suppressor encoded by an alternative open reading frame in the p16INK4a locus, wh ich inhibits the E3 ligase function of MDM2 (Kamijo, Weber et al. 1998; Zhang and Xiong 2001). MDM2 also regulates the acetylation of p53 by preventing binding of p300/CBP or recruitment of HDAC1 (Kobet, Zeng et al. 2000; Ito, Lai et al. 2001; Ito, Kawaguchi et al. 2002). Recent studies revealed that p53 is also regulated by the MDM2 homolog MDMX (Shvarts, Steegenga et al. 1996). MDMX sh ares strong homology to MDM2 at the amino acid sequence level and can bind to p53 and inhibit its transcri ption function in transient transfection assays. MDMX alone does not promote p53 ubiquitina tion or degradation in vivo (Stad, Little et al. 2001). However, formation of MDM2-MDMX heterodimer stimulates the ubiquitin E3 ligase activity of MDM2 for itself and p53 (Badciong and Haas 2002; Linares, Hengstermann et al. 2003 ), suggesting that MDMX may function as a regulator or cofactor of MDM2. Knockout of MDM2 in mice results in embryonic lethality due to hyper-activation of p53 (M ontes de Oca Luna, Wagner et al. 1995).
96 Several studies showed that MDMX-null mi ce also die in uterus in a p53-dependent fashion, which can be rescued by crossing in to the p53-null background (Parant, ChavezReyes et al. 2001; Finch, Donoviel et al. 2002; Migliorini, Denc hi et al. 2002). Therefore, MDMX is also an important regulator of p53 during embryonic development. MDMX overexpression can lead to transformation in cell culture; MDMX gene amplification and overexpression has been observed in certain tu mors with wild type p53, suggesting that it may contribute to p53 inactiv ation in cancer (Ramos, Stad et al. 2001; Danovi, Meulmeester et al. 2004). Recently, p53 and MDM2 have been found to be modified by SUMO-1 conjugation (Small ubiquitin-like modifie r) (Gostissa, Hengstermann et al. 1999; Rodriguez, Desterro et al. 1999; Xirodimas, Chisholm et al. 2002). Sumoylation of p53 and MDM2 in vitro are carried out by a universal SUMO activating E1 enzyme (SEA1/SEA2 heterodimer) and SUMO conjuga ting enzyme E2 (ubc9) (Hochstrasser 2002). SUMO is activated in an ATP-depende nt manner by E1, transferred to E2, and subsequently attached to the -amino group of lysines on targ et proteins (Hochstrasser 2002). The sumoylated lysine residue in many s ubstrates often locates within a consensus sequence of KxE where is an aliphatic residue (Rodr iguez, Dargemont et al. 2001). Lysine 386 is the major sumoylation site for p53 (Gostissa, Hengstermann et al. 1999; Rodriguez, Desterro et al. 1999), whereas the sumoylation site on MDM2 has not been positively identified (Xirodimas, Chishol m et al. 2002; Chen and Chen 2003). The functional role of sumoylation is diverse among different substrates. It has been proposed to be important for regulating the localization, activity, and degradation of target proteins (Hochstrasser 2002). Initial studies using p53 386 lysine-to-arginine
97 mutant resistant to sumoylation suggested that sumoylation moderately enhances p53 transcription activity (Gostissa Hengstermann et al. 1999; Ro driguez, Desterro et al. 1999). Muller et al. found that the p53 386K mutant has s lightly impaired apoptotic activity (Muller, Berger et al 2000). Other studies did not obs erve such an effect (Minty, Dumont et al. 2000; Kwek, Derry et al. 2001), possibly due to different assay conditions. MDM2 sumoylation has been shown to be st imulated by ARF through nucleolar targeting of MDM2 (Xirodimas, Chisholm et al. 2002; Chen and Chen 2003). MDM2 and ARF cooperatively stimulate p53 sumoylation (C hen and Chen 2003), suggesting a potential mechanism that contributes to ARF activ ation of p53. Despite efforts by several laboratories, the sumoylation site on MDM2 has not been determined, possible due to presence of multiple alternative sites. The lack of sumoylation-deficient MDM2 mutants prevents further investigation of the functional of this modification. Recent work from several laboratories showed that MDMX is regulated by MDM2-mediated ubiquitination and degrada tion (Kawai, Wiederschain et al. 2003; Linares, Hengstermann et al. 2003; Pan and Chen 2003). Ubiquitination of MDMX leads to increased degradation after DNA da mage and expression of ARF (Kawai, Wiederschain et al. 2003; Pan and Chen 2003), suggesting that ubiquitination is an important modification on MDMX. In this repor t, we present evidence that similar to p53 and MDM2, MDMX is also modified by sum oylation. Two lysine residues critical for MDMX sumoylation have been identified. MDMX point mutants that cannot be modified by sumoylation have similar properties as wild type MDMX in seve ral functional assays, suggesting that sumoylation are not required for these MDMX functions.
98 Results MDMX is posttranslationally modified by sumo-1 MDM2 has been demonstrated to be modified by sumo-1. Since MDMX and MDM2 share significant sequence homology, we speculated that MDMX might also be sumoylated in vivo To detect MDMX sumoylation, we coexpressed MDMX and His6sumo-1 by transient transfec tion into H1299 cells. MDMX conjugated to His6-sumo-1 was purified by Ni 2+ -NTA beads under denaturing condi tions and detected by western blot with MDMX-specific monoclonal antibody 8C6. The result showed that coexpression of MDMX and his6-sumo-1 resulted in the appearance of MDMX species at 90 kd molecular weight, suggesting that MDMX was sumoylated in vivo (Figure 28). Expression of the desumoylating enzyme SENP1 completely eliminated the 90 kd MDMX band (Gong, Millas et al. 20 00), consistent with it being a sumoylated form of MDMX. It has been demonstrated that the sumoylation of MDM2 can be significantly enhanced by ARF (Xirodimas, Chisholm et al 2002; Chen and Chen 2003). To test if ARF also stimulates the sumoylation of MD MX, ARF was cotransfec ted together with MDMX and His6-sumo-1. In analogy to MD M2 sumoylation, MDMX sumoylation was also enhanced by the overexpression of ARF (Figure 28). The significance and mechanism of this effect remains to be determined.
99 Figure 28. ARF stimulates MDMX sumoylation and the sumo-specific protease SENP1 desumoylates MDMX. MDMX was coexpressed with His6 -sumo in H1299 cells by transient transfection. His6-sumoylatedproteins were purified, and MDMX was detected by Western blotting with 8C6 antibody. The lower panels are control Western blots of nonpurified cell lysate (10% of the amount used for Ni-NTA purification).Transfection efficiency was verified by expression of cotransfected GFP plasmid.
100 We have recently reported that MDM2 functions as an E3 ligase for MDMX polyubiquitination (Pan and Chen 2003). Furtherm ore, MDM2 expression also stimulates p53 sumoylation (Chen and Chen 2003). Theref ore, we tested whether MDM2 and p53 regulate MDMX sumoylation. The results s howed that expression of MDM2 led to degradation of MDMX in a RING finger-depen dent fashion as expected. The level of MDMX sumoylation was also reduced acco rdingly, suggesting that MDM2 does not stimulate MDMX sumoylation (Figure 28). Coexpression of wild type p53 significantly reduced transfection efficiency and MDMX e xpression level due to squelching effects, the corresponding reduction of sumoylated MDMX level also ruled out a stimulatory effect. Next, we tested whether MDMX sumoyla tion can be recapitulated in cell-free conditions. In vitro translated MDMX was incubated with Hela cell E1 fraction, purified E2, and ATP. Significant modification of MDMX was observed after addition of purified sumo-1 (Figure 29). Addition of in vitro translated ARF had no effect on MDMX sumoylation level in this assay, possibly due to insufficient amount and poor binding to MDMX under the a ssay condition, or that the in vivo stimulatory effect involved additional factors.
Figure 29. In vitro sumoylation of MDMX. In vitro translated proteins were incubated in the presence or absence of E1 fraction, E2, ATP or SUMO1. ARF did not stimulate MDMX sumoylation in vitro. Identification of MDMX sumoylation sites The consensus sequence for sumoylation has been identified as KXE, where represents a large hydrophobic amino acid and K stands for the sumo conjugating lysine residue. However, despite the presence of such consensus site, previous studies were not able to definitively identify the sumoylation site on MDM2 (Xirodimas, Chisholm et al. 2002; Chen and Chen 2003). To determine the region containing sumoylation sites on 101
102 MDMX, a panel of MDMX dele tion mutants was generate d with N or C terminal truncations (Figure 30). Th e MDMX deletion mutants were coexpressed with His6sumo-1 by transient transfecti on into H1299 cells, purified by Ni 2+ -NTA beads under denaturing conditions and detected by west ern blot with MDMX polyclonal antibody. The result showed that major sumoylation sites are contained w ithin the region 250-490 (Figure 31).
Figure 30. The major sumoylation sites of MDMX are contained within the region 250-490. A panel of MDMX deletion mutants was generated for their sumoylation status. MDMX deletion mutants were coexpressed with His6 -sumo in H1299 cells by transient transfection. His6-sumoylated proteins were purified, and MDMX deletion mutants were detected by Western blotting with polyclonal rabbit-anti-MDMX antibody. The lower panels are control Western blots of nonpurified cell lysate, showing the expression of the MDMX deletion mutants (10% of the amount used for Ni-NTA p urification). 103
Figure 31. Diagram of MDMX mutants and summary of results. Major MDMX sumoylation sites are contained within the region 250-490. The MDMX sequence contains three sites that conform to the KXE consensus, lysine 254 (I K VE), 379 (I K KE) and 478 (C K KE). To directly test whether these sites are required for sumoylation, single or combined site-directed mutagenesis of these three lysines were performed. The mutants were tested for their sumoylation efficiency by transient transfection of H1299 cells. MDMX sumoylation was not affected by the mutation of any single lysines. However, double-mutation of K254 and K379 to arginine 104
abrogated the sumoylation of MDMX (Figure 32). Therefore, K254 and K379 are likely to be the major sumoylation sites, which is consistent with the deletion mutant results. Fi g ure 32. Characterization of MDMX sumo y lation sites on serine 254 andserine 379. Point mutants serine 254, 379 and 478 were generated for identification o f MDMX sumoylation sites. Sumoylated wild-type MDMX and MDMX poin t mutants K254, K379, K478 and the double mutant of K254 and K379 (DM) aswell as the triple mutant of all three serines in the consensus sequence were p urified by Ni-NTA purification and detected by Western blot with 8C6 antibody.The lower panels are control Western blots of nonpurified cell lysate (10% of theamount used for Ni-NTA purification). 105
106 SENP1 has been identified as a sumo -specific protease. In order to further verify the modification of MDMX by sumo, we cotr ansfected His6-sumo-1 with wild type MDMX, MDMX sumo-deficient single mutant s, double mutant and triple mutant in H1299 cells in the presence or absence of SENP1 expression plasmid. Sumoylated MDMX was purified by Ni 2+ -NTA purification and detected by Western blot. The result showed that SENP1 reversed the sumoylati on of wild type MDMX and all three MDMX point mutants but had no effect on sumoyl ation-deficient MDMX double or triple mutants, confirming the characterization of MDMX sumoylation sites (Figure 33).
Figure 33. Desumoylation of MDMX by SENP1. H1299 cells were transfected with His6-SUMO1 and indicated plasmids. Sumoylated MDMX was detected by Ni-NTA purification and Western blot. Expression of SENP1 abrogated MDMX sumoylation. MDMX sumoylation does not affect its stability and regulation by DNA damage MDMX forms a heterodimer with MDM2 and is poly-ubiquitinated and degraded by MDM2. MDMX is also modified by mono-ubiquitination in the absence of MDM2 expression through unknown mechanism (Pan and Chen 2003). To test if blocking MDMX sumoylation affects its ubiquitination, we coexpressed MDMX and His6107
108 ubiquitin by transient transfec tion into H1299 cells. MDMX conjugated to His6-ub was purified by Ni 2+ -NTA beads under denaturing conditions and detected by western blot with MDMX-specific monoclonal antibody 8C 6. The result showed that K254 and K379 single or double mutations had little effect on MDMX mono -ubiquitination (Figure 34). Interestingly, mutating K478 led to significan tly higher mono-ubiquitina tion either alone or in the triple mutant background. Since K 478 is located immediately after the last cysteine residue of the RING finger, its muta tion to arginine may have stimulated the weak self-ubiquitination of MDMX (Badciong and Haas 2002). Poly-ubiquitination of the MDMX mutants by MDM2 was similar to wild type MDMX (Figure 34), suggesting that lack of sumoylation does not affect MDMX ubiquitination by MDM2. Surprisingly, the K478R mutant was also poly-ubiquitinated and degrad ed at similar efficiency compared to wild type MDMX, suggesting that increased basal mono-ubiquitination does not promote poly-ubiquitinati on and degradation by MDM2.
Figure 34. Lack of sumoylation does not affect MDMX mono-ubiquitination or polyubiquitination by MDM2. MDMX and His6-ubiquitin were coexpresssed by transient transfection intoH1299 cells. MDMX conjugated to His6-ub was purified by Ni 2+ -NTA beadsunder denaturing conditions and detected by western blot with MDMX-specific monoclonal antibody 8C6. MDMX sumoylation status has little effect on the stability of MDMX. Thelower panels are control Western blots of non-purified cell lysate (10% of theamount used for Ni-NTA purification). Transfection efficiency was verifie d by expression of cotransfected GFP plasmid. Next, the MDMX sumo-deficient double point mutants were cotransfected with increasing amount of MDM2 in H1299 cells. To further compare the sensitivity of 109
MDMX sumoylation mutants to MDM2-mediated degradation, a titration experiment was performed with fixed amount of MDMX plasmid and increasing amount of MDM2 plasmid in the transfection. The dose-dependent degradation by MDM2 was also similar between wild type and mutant MDMX, suggesting that MDMX sumoylation status does not directly affect its stability regulated by MDM2 (Figure 35). Figure 35. MDMX sumoylation status does not affect its MDM2-mediated degradation. MDMX and MDMX sumoylation-deficient double mutant were cotransfected with increasing amount of MDM2. The level of MDMX was detected by Western blot with 8C6 antibody. Previous studies showed that MDMX is destabilized upon DNA damage in an MDM2-dependent fashion (Kawai, Wiederschain et al. 2003; Pan and Chen 2003). Recent work in our lab suggested that DNA damage-induced phosphorylation of MDMX might play a role in accelerating MDMX degradation (manuscript in preparation). To test if blocking MDMX sumoylation affects its regulation by DNA damage, U2OS cells were stably transfected with control vector, wild type MDMX and MDMX sumo110
deficient mutants. Cells were treated with ionizing radiation and collected at different time points. We found that both wild type and sumo-deficient MDMX were degraded at similar efficiency after DNA damage (Figure 36). Fi g ure 36. MDMX sumo y lation status does not affect its DNA dama g e induced destabilization. U2OS cells were stably transfected with control vector, wild type MDMX and MDMX sumo-deficient mutants. Cells were treated with ionizing radiation and collected at different time points. The levels of MDMX, MDM2, p53, p21 and actin were detected by Western blot. 111
MDMX is mainly localized in the cytoplasm of cells when expressed at high levels and undergoes nuclear translocation after DNA damage through both MDM2-mediated and MDM2-independent mechanisms (Li, Chen et al. 2002) When compared to wild type MDMX, the sumoylation-deficient MDMX mutants also translocated into the nucleus after ionizing radiation (Figure 37). Therefore, we conclude that sumoylation is not required for the regulation of MDMX localization and stability after DNA damage. Figure 37. Sumoylation-deficient MDMX mutants translocated into the nucleus after ionizing radiation in the same fashion as wild type MDMX. MDMX mutants were transfected in H1299 cells and r irradiated. Cells were incubated for 18 hrs before collection. MDMX was detected by 8C6 antibody and stained with FITC-goat-anti-mouse IgG. In further experiments, we compared the ability of MDMX sumoylation mutants to inhibit p53 transcriptional function. The MDMX mutants were cotransfected with p53 112
113 expression plasmid and p53-responsive luciferase reporter into p53-null H1299 cells, additionally, MDMX mutants were cotransfec ted with p53-luciferase reporter into p53wild type U2OS cells to determine the e ffect on endogenous p53 tran scriptional activity. In both tests, the MDMX sumoylation mutant s exhibited similar ability to inhibit p53 function compared to wild type (data not shown). Coimmunoprecipitation experiments with MDM2 and p53 also did not review ch anges in the ability of the sumoylation mutants to bind MDM2 or p53 (Figure 38). Ther efore, we concluded that sumoylation of MDMX is not required for suppression of p53 activity under our experimental conditions.
Figure 38. MDMX sumoylation status does not affect protein-protein binding between p53, MDM2 and MDMX. Wild type MDMX and MDMX sumoylation-deficient mutant were cotransfected with MDM2 (lane 1-3) or p53 (lane 4-6) or both (lane 7-9) in H1299 cells. No change of MDMX co-precipitated with MDM2 (compare lane 2&3) or p53 (compare lane 5&6) was detected by Western blot with anti-MDMX antibody 8C6. No change of MDM2 co-precipitated with p53 in the p resence of wild type or mutant MDMX was detected by Western blot with anti-MDM2 antibo d y 3G9. 114
115 Discussion In this report, we showed that sim ilar to p53 and MDM2, MDMX also undergoes modification by sumoylation. Conjuga tion to sumo-1 was detected both in vivo and in vitro Similar to other sumoylation substrates MDMX sumoylation in vitro is highly efficient in the presence of E1 and E2 enzy mes, suggesting that a specific E3 is not required for this reaction. Previous studies fa iled to identify the lysine residue important for sumoylation on MDM2, possibly due to presence of multiple alternative sites (Xirodimas, Chisholm et al. 2002; Chen a nd Chen 2003). Through deletion and point mutagenesis, we were able to determine that K254 and K379 are important for MDMX sumoylation. Single mutations on each site had limited effect on MDMX sumoylation, but the double mutant largely abrogated the modification. Sequence comparison showed that these sites are not conserved on MD M2, suggesting that MDM2 sumoylation involves other lysine residues. Furthermore, mutation of K254 and K379 did not affect MDMX ubiquitination, eliminating the possibil ity that sumoylation and ubiquitination compete for the same lysines on MDMX. An unresolved question is the role of su moylation in the function and regulation of the p53 pathway. Previous studies showed that sumoylation of p53 had very moderate positive effects on its transcription activ ation function (Melchior and Hengst 2002). Inability to identify the sumoylation site on MD M2 hampered investigation of the role of this modification on MDM2 and p53 func tion. Generation of MDMX sumoylationdeficient mutants in this st udy enabled us to test whethe r sumoylation of MDMX is
116 involved in several aspects of MDMX function and regulatio n. At present, the results suggest that MDMX sumoylation is not required for MDMX ubiquitination and degradation by MDM2. Furthermore, regulation of MDMX degradation and nuclear translocation by DNA damage are not affected by lack of sumoylation. The regions of MDMX that contain the sumoylation sites are not directly involved in binding to p53, therefore it is not surprisi ng that p53 binding and regulati on by the MDMX mutants are similar to wild type. Comparison between the level of total MD MX expression in th e transfected cells and the amount of sumoylated MDMX recove red by His6-sumo-1 purification indicated that only a small fraction of MDMX (<5%) is conjugated to sumo-1 at any given time. Therefore, it is possible that the modifica tion regulates a specif ic sub-population of MDMX. Therefore, the impact of this m odification on MDMX function may not be evident by the overexpression assays we empl oyed. Although the functional analysis so far have not revealed a requirement for sumoylation in several aspects of MDMX regulation, we cannot rule out a role for su moylation in regulating other yet to be identified MDMX functions. The findings and reagents described in this report should facilitate future study of the role of sumoylation in regulating MDMX and the p53 pathway.
117 Scientific Significance P53 is a tumor suppressor serving as a gu ardian of the genome, and controls the proliferation and apoptosis of abnormal cells. Under normal condition, p53 needs to be maintained at low level to allow cell growth and proliferation. The two major negative p53-regulators identified to date are MDM2 and MDMX. In mouse models, the abrogation of either MDM2 or MDMX lead to embryonic lethality, which can be rescued by crossing into p53 null background, demonstrating that both MDM2 and MDMX are major regulators of p53. P53 is a well-known stress sensor for cells. After DNA damage, p53 transcriptionally activates th e downstream target genes such as p21 and Bax to stop the cells carrying damaged DNA from proliferating or to eliminate the cells by apoptosis. Hypothetically, full activati on of p53 requires the inhibi tion of both MDM2 and MDMX, which would restrain p53 activ ity under normal conditions. It is well recognized that p53 binding to MDM2 is weakened after DNA da mage due to phosphorylation on the MDM2 binding site. However, the que stion still remains that if MDMX, the other major negative regulator of p53 can be eliminated after DNA da mage to allow the complete activation of p53. Since MDMX and MDM2 form heterodi mers through their RING finger domains and MDM2 contains the intrinsic ubiquitin E3 ligase activity, we speculated that MDMX
118 might be an E3 substrate of MDM2. Ind eed, we found that MDM2 can stimulate the polyubiquitination and degradation of MDMX through the proteasome pathway. From the in vivo and in vitro experiments, we concluded that MDM2 is the bona fide E3 ubiquitin ligase for MDMX and their hete rodimerization through the RING domains is required. Based on the fact that MDM2 accumulates due to p53 activation after DNA damage and that MDMX can be degraded in an MDM2-dependent fashion, we speculated that DNA damage might induce th e degradation of MD MX partially through the accumulation of MDM2. This was shown to true. After DNA damage, MDMX level significantly reduced, which can be rescued by proteasome inhibitor. The reduction of MDMX protein level is not caused by th e mRNA change and is accompanied by the accumulation of MDM2. Interestingly, we found that ARF overexpr ession had a synergistic effect with MDM2 in stimulating MDMX polyubiquitination a nd degradation. This is in contrast to ARFs inhibition of p53 polyubiquitina tion by MDM2. Additionally, ARF overexpression alone is sufficient in induci ng MDMX degradation. This finding changed our understanding of ARF from a stereotypi cal MDM2 inhibitor into a dynamic MDM2 E3 regulator, which turns on MDM2s E3 function towards MDMX oncogene while suppressing its E3 activity towa rds p53 tumor suppressor. Based on the above findings, we propos ed the following model: Under normal condition, p53 remains inactive and unstable du e to the negative regulation from both MDM2 and MDMX. After DNA damage, MD MX is degraded through a proteasomedependent pathway. Meanwhile, MDM2p53 binding is disrupted due to p53 phosphorylation. On the other hand, oncogenic signals such as Ras and E2F1 can induce
119 the overexpression of ARF, which stim ulates MDM2 ubiquitination of MDMX degradation and in the meantime, suppressing the ubiquitination of p53 by MDM2. Thus, both DNA damage and mitogenic signals can ac tivate p53 by abrogating the function of both p53 negative regulators, namely MDM2 and MDMX. Our work established MDMX as an important target of stress signals Degradation of MDMX by MDM2 connects MDMX to the well-established p53 signaling pa thways and provides further insight into the mechanisms of effective p53 re sponse to stress (Figure 21). Posttranslational modification is an effi cient way to modulate protein biological functions. We identified MDMX as an ubiquitin substrate of MDM2. Moreover, we found that MDMX can be conjugated to SUMO -1 and identified the sumoylation sites on MDMX to K254 and K379. It is worth noticing that ARF can stimulate both ubiquitination and sumoylation of MDMX. However, the mechanism through which ARF modulates MDMX posttranslationa l modification remains elusive. In this manuscript, we found that MDM2 significantly stimulates the binding efficiency between MDMX and ARF. Additionally, ARF stimulates MD MX polyubiquitination in a dose-dependent fashion associated with a stab ilization of MD M2. Ongoing research in the lab is focusing on the mechanism by which oncogenes signal through ARF to regulate MDMX function and stability, as well as its implication on p53 activation. In conclusion, data from this manuscript made important contribution to our understanding of the mechanism of MDMX regulation, which is vital in understanding p53 act ivation upon DNA damage or mitogenic stress.
120 Reference Ambrosini, G., C. Adida, et al. (1997). "A novel anti-apoptosis gene, survivin, expressed in cancer and lymphoma." Nat Med 3(8): 917-21. Appella, E. and C. W. Anderson (2001). "Pos t-translational modifi cations and activation of p53 by genotoxic stresses." Eur J Biochem 268(10): 2764-72. Atadja, P., H. Wong, et al. (1995). "Increased activity of p53 in se nescing fibroblasts." Proc Natl Acad Sci U S A 92(18): 8348-52. Avantaggiati, M. L., V. Ogryzko, et al. (1997). "Recruitment of p300/CBP in p53dependent signal pathways." Cell 89(7): 1175-84. Badciong, J. C. and A. L. Haas (2002). "M dmX is a RING finger ubiquitin ligase capable of synergistically enhanci ng Mdm2 ubiquitination." J Biol Chem 277(51): 49668-75. Bailey, D. and P. O'Hare (2002). "Herpes simplex virus 1 ICP0 co-localizes with a SUMO-specific protease." J Gen Virol 83(Pt 12): 2951-64. Banin, S., L. Moyal, et al. (1998). "Enhanced phosphorylation of p53 by ATM in response to DNA damage." Science 281(5383): 1674-7. Barak, Y., E. Gottlieb, et al. (1994). "Regula tion of mdm2 expression by p53: alternative promoters produce transcripts with n onidentical translation potential." Genes Dev 8(15): 1739-49.
121 Barak, Y., T. Juven, et al. (1993). "mdm2 expression is induced by wild type p53 activity." Embo J 12(2): 461-8. Bartek, J., J. Falck, et al. (2001) "CHK2 kinase--a busy messenger." Nat Rev Mol Cell Biol 2(12): 877-86. Becker-Catania, S. G. and R. A. Gatti (2001). "Ataxia-telangiectasia." Adv Exp Med Biol 495: 191-8. Bennett, M., K. Macdonald, et al. (1998). "C ell surface trafficking of Fas: a rapid mechanism of p53-mediated apoptosis." Science 282(5387): 290-3. Boddy, M. N., K. Howe, et al. (1996). "P IC 1, a novel ubiquitin -like protein which interacts with the PML component of a multiprotein complex that is disrupted in acute promyelocytic leukaemia." Oncogene 13(5): 971-82. Bottger, V., A. Bottger, et al. (1999). "Comparative study of the p53-mdm2 and p53MDMX interfaces." Oncogene 18(1): 189-99. Bykov, V. J., N. Issaeva, et al. (2002). "Mutant p53-dependent growth suppression distinguishes PRIMA-1 from known anticancer dr ugs: a statistical analysis of information in the National Cancer Institute database." Carcinogenesis 23(12): 2011-8. Caelles, C., A. Helmberg, et al. (1994). "p53-dependent apoptosis in the absence of transcriptional activatio n of p53-target genes." Nature 370(6486): 220-3. Cahilly-Snyder, L., T. Yang-Feng, et al. (1987) "Molecular analysis and chromosomal mapping of amplified genes isolated from a transformed mouse 3T3 cell line." Somat Cell Mol Genet 13(3): 235-44.
122 Canman, C. E., D. S. Lim, et al. (1998). "Activation of the ATM kinase by ionizing radiation and phosphorylation of p53." Science 281(5383): 1677-9. Chan, T. A., H. Hermeking, et al. (1999). 14-3-3Sigma is required to prevent mitotic catastrophe after DNA damage." Nature 401(6753): 616-20. Chen, J., V. Marechal, et al. (1993). "M apping of the p53 and mdm-2 interaction domains." Mol Cell Biol 13(7): 4107-14. Chen, L. and J. Chen (2003). "MDM2-ARF complex regulates p53 sumoylation." Oncogene 22(34): 5348-57. Crawford, T. O. (1998). "Ataxia telangiectasia." Semin Pediatr Neurol 5(4): 287-94. Danovi, D., E. Meulmeester, et al. (2004). "A mplification of Mdmx (or Mdm4) directly contributes to tumor formation by i nhibiting p53 tumor suppressor activity." Mol Cell Biol 24(13): 5835-43. DeLeo, A. B., G. Jay, et al. (1979). "Detection of a transf ormation-related antigen in chemically induced sarcomas and othe r transformed cells of the mouse." Proc Natl Acad Sci U S A 76(5): 2420-4. Dimri, G. P., K. Itahana, et al. (2000). "Re gulation of a senescence checkpoint response by the E2F1 transcription factor and p14(ARF) tumor suppressor." Mol Cell Biol 20(1): 273-85. Donehower, L. A., M. Harvey, et al. (1992). "Mice deficient for p53 are developmentally normal but susceptible to spontaneous tumours." Nature 356(6366): 215-21.
123 Dornan, D., I. Wertz, et al. (2004). "The ubiquitin ligase COP1 is a critical negative regulator of p53." Nature 429(6987): 86-92. Dumaz, N. and D. W. Meek (1999). "S erine15 phosphorylation stimulates p53 transactivation but does not directly influence interaction with HDM2." Embo J 18(24): 7002-10. Dumont, P., J. I. Leu, et al. (2003). "The codon 72 polymor phic variants of p53 have markedly different apoptotic potential." Nat Genet 33(3): 357-65. Dyson, N. (1998). "The regulation of E2F by pRB-family proteins." Genes Dev 12(15): 2245-62. el-Deiry, W. S., T. Tokino, et al. (1993). "WAF1, a potential mediator of p53 tumor suppression." Cell 75(4): 817-25. Fakharzadeh, S. S., S. P. Trusko, et al. (1991) "Tumorigenic potential associated with enhanced expression of a gene that is amplified in a mouse tumor cell line." Embo J 10(6): 1565-9. Fang, S., J. P. Jensen, et al. (2000). "Mdm2 is a RING finger-depende nt ubiquitin protein ligase for itself and p53." J Biol Chem 275(12): 8945-51. Finch, R. A., D. B. Donoviel, et al. (2002). "mdmx is a nega tive regulator of p53 activity in vivo." Cancer Res 62(11): 3221-5. Foulkes, W. D., T. Y. Flanders, et al (1997). "The CDKN2A ( p16) gene and human cancer." Mol Med 3(1): 5-20.
124 Freedman, D. A. and A. J. Levine (1998). "Nuc lear export is required for degradation of endogenous p53 by MDM2 and human papillomavirus E6." Mol Cell Biol 18(12): 728893. Friend, S. H., R. Bernards, et al. (1986). "A human DNA segment with properties of the gene that predisposes to reti noblastoma and osteosarcoma." Nature 323(6089): 643-6. Fukasawa, K., F. Wiener, et al (1997). "Genomic instability and apoptosis are frequent in p53 deficient young mice." Oncogene 15(11): 1295-302. Girdwood, D., D. Bumpass, et al. (2003). "P30 0 transcriptional repression is mediated by SUMO modification." Mol Cell 11(4): 1043-54. Girnita, L., A. Girnita, et al. (2003). "Mdm 2-dependent ubiquitination and degradation of the insulin-like growth factor 1 receptor." Proc Natl Acad Sci U S A 100(14): 8247-52. Gong, L., S. Millas, et al. (2000). "Differential regulation of sentrinized proteins by a novel sentrin-specific protease." J Biol Chem 275(5): 3355-9. Gossen, M. and H. Bujard (1992). "Tight cont rol of gene expressi on in mammalian cells by tetracycline-responsive promoters." Proc Natl Acad Sci U S A 89(12): 5547-51. Gostissa, M., A. Hengstermann, et al. (1999) "Activation of p53 by conjugation to the ubiquitin-like protein SUMO-1." Embo J 18(22): 6462-71. Gu, J., H. Kawai, et al. (2002). "Mutual dependence of MDM2 and MDMX in their functional inactivation of p53." J Biol Chem 277(22): 19251-4.
125 Gu, W., X. L. Shi, et al. ( 1997). "Synergistic activation of transcription by CBP and p53." Nature 387(6635): 819-23. Haas, A. L. and T. J. Siepmann (1997) "Pathways of ubiquitin conjugation." Faseb J 11(14): 1257-68. Hammond, E. M., N. C. Denko, et al. ( 2002). "Hypoxia links ATR and p53 through replication arrest." Mol Cell Biol 22(6): 1834-43. Hanahan, D. and R. A. Weinberg (2000). "The hallmarks of cancer." Cell 100(1): 57-70. Hang, J. and M. Dasso (2002). "Association of the human SUMO-1 protease SENP2 with the nuclear pore." J Biol Chem 277(22): 19961-6. Hardeland, U., R. Steinacher, et al. (2002). "Modification of the human thymine-DNA glycosylase by ubiquitin-lik e proteins facilitates enzymatic turnover." Embo J 21(6): 1456-64. Harper, J. W. (2004). "Neddylating the guard ian; Mdm2 catalyzed conjugation of Nedd8 to p53." Cell 118(1): 2-4. Haupt, S. and Y. Haupt (2004). "Improving Cancer Therapy Through p53 Management." Cell Cycle 3(7): 912-6. Haupt, Y., R. Maya, et al. (1997). "Mdm2 promotes the rapid degradation of p53." Nature 387(6630): 296-9. Haupt, Y., S. Rowan, et al. (1995). "Induction of apoptosis in HeLa cells by transactivation-deficient p53." Genes Dev 9(17): 2170-83.
126 Heffernan, T. P., D. A. Simpson, et al (2002). "An ATRan d Chk1-dependent S checkpoint inhibits replicon initia tion following UVC-induced DNA damage." Mol Cell Biol 22(24): 8552-61. Heise, C., A. Sampson-Johannes, et al (1997). "ONYX-015, an E1B gene-attenuated adenovirus, causes tumor-specific cytolysi s and antitumoral efficacy that can be augmented by standard chemotherapeutic agents." Nat Med 3(6): 639-45. Hochstrasser, M. (2002). "New structural clues to substrate specificity in the "ubiquitin system"." Mol Cell 9(3): 453-4. Hollstein, M., D. Sidransky, et al. (1991). "p53 mutations in human cancers." Science 253(5015): 49-53. Honda, R. and H. Yasuda (1999). "Associa tion of p19(ARF) with Mdm2 inhibits ubiquitin ligase activity of Md m2 for tumor suppressor p53." Embo J 18(1): 22-7. Huang, T. T., S. M. Wuerzberger-Davis, et al. (2003). "Sequential modification of NEMO/IKKgamma by SUMO-1 and ubiquiti n mediates NF-kappaB activation by genotoxic stress." Cell 115(5): 565-76. Hwang, B. J., J. M. Ford, et al. (1999). "Expression of the p48 xeroderma pigmentosum gene is p53-dependent and is invo lved in global genomic repair." Proc Natl Acad Sci U S A 96(2): 424-8. Ito, A., Y. Kawaguchi, et al. (2002). "MDM 2-HDAC1-mediated deacetylation of p53 is required for its degradation." Embo J 21(22): 6236-45.
127 Ito, A., C. H. Lai, et al. (2001). "p30 0/CBP-mediated p53 acet ylation is commonly induced by p53-activating agents and inhibited by MDM2." Embo J 20(6): 1331-40. Jackson, M. W. and S. J. Berberich (1999). "Constitutive mdmx expression during cell growth, differentiati on, and DNA damage." DNA Cell Biol 18(9): 693-700. Jackson, M. W. and S. J. Berberich (2000) "MdmX protects p53 from Mdm2-mediated degradation." Mol Cell Biol 20(3): 1001-7. Jackson, M. W., M. S. Lindstrom, et al. (2001). "MdmX binding to ARF affects Mdm2 protein stability and p53 transactivation." J Biol Chem 276(27): 25336-41. Jin, Y., S. X. Zeng, et al. (2004). "M DM2 mediates p300/CREB-binding proteinassociated factor ubiquitination and degradation." J Biol Chem 279(19): 20035-43. Jones, S. N., A. E. Roe, et al. (1995). "Res cue of embryonic lethality in Mdm2-deficient mice by absence of p53." Nature 378(6553): 206-8. Juan, L. J., W. J. Shia, et al. (2000). "H istone deacetylases specifically down-regulate p53-dependent gene activation." J Biol Chem 275(27): 20436-43. Juven, T., Y. Barak, et al. (1993). "Wild type p53 can mediate sequence-specific transactivation of an internal pr omoter within the mdm2 gene." Oncogene 8(12): 3411-6. Kamb, A., N. A. Gruis, et al. (1994). "A cell cycle regulator pot entially involved in genesis of many tumor types." Science 264(5157): 436-40. Kamijo, T., J. D. Weber, et al. (1998). "Func tional and physical inte ractions of the ARF tumor suppressor with p53 and Mdm2." Proc Natl Acad Sci U S A 95(14): 8292-7.
128 Kamitani, T., H. P. Nguyen, et al. (1998). "Covalent modification of PML by the sentrin family of ubiquiti n-like proteins." J Biol Chem 273(6): 3117-20. Kamitani, T., H. P. Nguyen, et al. (1997). "P referential modification of nuclear proteins by a novel ubiquitin-like molecule." J Biol Chem 272(22): 14001-4. Kastan, M. B., Q. Zhan, et al. (1992). "A mammalian cell cycle checkpoint pathway utilizing p53 and GADD45 is defec tive in ataxia-telangiectasia." Cell 71(4): 587-97. Kawai, H., D. Wiederschain, et al. (2003). "DNA damage-induced MDMX degradation is mediated by MDM2." J Biol Chem 278(46): 45946-53. Khanna, K. K., K. E. Keating, et al. (1998) "ATM associates with and phosphorylates p53: mapping the region of interaction." Nat Genet 20(4): 398-400. Kim, E. J., J. S. Park, et al. (2003). "Identification of Daxx interacting with p73, one of the p53 family, and its regulation of p53 activ ity by competitive interaction with PML." Nucleic Acids Res 31(18): 5356-67. Kim, K. I., S. H. Baek, et al. (2000). "A new SUMO-1-speci fic protease, SUSP1, that is highly expressed in reproductive organs." J Biol Chem 275(19): 14102-6. Ko, L. J. and C. Prives (1996). "p53: puzzle and paradigm." Genes Dev 10(9): 1054-72. Kobet, E., X. Zeng, et al. (2000). "MDM2 inhibits p300-mediated p53 acetylation and activation by forming a ternary co mplex with the two proteins." Proc Natl Acad Sci U S A 97(23): 12547-52.
129 Kokontis, J. M., A. J. Wagner, et al. (2001). "A transcriptional activation function of p53 is dispensable for and inhibitory of its apoptotic function." Oncogene 20(6): 659-68. Koumenis, C., R. Alarcon, et al. (2001). "Regulation of p53 by hypoxi a: dissociation of transcriptional repression and apoptosis from p53-dependent transactivation." Mol Cell Biol 21(4): 1297-310. Kroemer, G. and J. C. Reed (2000). "Mitochondrial contro l of cell death." Nat Med 6(5): 513-9. Kwek, S. S., J. Derry, et al. (2001). "Functional analysis and intracellular localization of p53 modified by SUMO-1." Oncogene 20(20): 2587-99. Lambert, P. F., F. Kashanchi, et al. (1998) "Phosphorylation of p53 serine 15 increases interaction with CBP." J Biol Chem 273(49): 33048-53. Lane, D. P. and L. V. Crawford (1979). "T antigen is bound to a host protein in SV40transformed cells." Nature 278(5701): 261-3. Lansdorp, P. M. (2000). "Repair of telomeric DNA prior to replicative senescence." Mech Ageing Dev 118(1-2): 23-34. Laronga, C., H. Y. Yang, et al. (2000). "Associ ation of the cyclin-dependent kinases and 14-3-3 sigma negatively regulat es cell cycle progression." J Biol Chem 275(30): 2310612. Lebedeva, I. V., Z. Z. Su, et al. (2003). "Restoring apoptosis as a strategy for cancer gene therapy: focus on p53 and mda-7." Semin Cancer Biol 13(2): 169-78.
130 Leng, R. P., Y. Lin, et al. (2003). "Pir h2, a p53-induced ubiquitin-protein ligase, promotes p53 degradation." Cell 112(6): 779-91. Levine, A. J., J. Momand, et al. (199 1). "The p53 tumour suppressor gene." Nature 351(6326): 453-6. Levy, M. Z., R. C. Allsopp, et al. (1992). "Telomere end-replication problem and cell aging." J Mol Biol 225(4): 951-60. Li, C., L. Chen, et al. (2002). "DNA dama ge induces MDMX nuclear translocation by p53-dependent and -independent mechanisms." Mol Cell Biol 22(21): 7562-71. Li, M., D. Chen, et al. (2002). "Deubiqui tination of p53 by HAUSP is an important pathway for p53 stabilization." Nature 416(6881): 648-53. Lill, N. L., S. R. Grossman, et al. (19 97). "Binding and modulation of p53 by p300/CBP coactivators." Nature 387(6635): 823-7. Lin, Y., W. Ma, et al. (2000) "Pidd, a new death-domain-c ontaining protein, is induced by p53 and promotes apoptosis." Nat Genet 26(1): 122-7. Linares, L. K., A. Hengstermann, et al (2003). "HdmX stimulates Hdm2-mediated ubiquitination and degradation of p53." Proc Natl Acad Sci U S A 100(21): 12009-14. Linskens, M. H., C. B. Harley, et al. ( 1995). "Replicative senes cence and cell death." Science 267(5194): 17.
131 Linzer, D. I. and A. J. Levine (1979). "Characterization of a 54K dalton cellular SV40 tumor antigen present in SV40-transformed cells and uninfected embryonal carcinoma cells." Cell 17(1): 43-52. Livingstone, L. R., A. White, et al. (1 992). "Altered cell cycle arrest and gene amplification potential acco mpany loss of wild-type p53." Cell 70(6): 923-35. Lu, W., J. Lin, et al. (2002). "Expression of p14ARF overcomes tumor resistance to p53." Cancer Res 62(5): 1305-10. Mahajan, R., C. Delphin, et al. (1997). "A small ubiquitin-related polypeptide involved in targeting RanGAP1 to nuclear pore complex protein RanBP2." Cell 88(1): 97-107. Mancini, F., F. Gentiletti, et al. (2004). "MDM4 (MDMX) overexpression enhances stabilization of stress-induced p53 and promotes apoptosis." J Biol Chem 279(9): 816980. Mao, Y., S. D. Desai, et al. (2000). "SUMO-1 conjugation to human DNA topoisomerase II isozymes." J Biol Chem 275(34): 26066-73. Mao, Y., M. Sun, et al. (2000). "SUMO-1 conj ugation to topoisomerase I: A possible repair response to topoisomerase-mediated DNA damage." Proc Natl Acad Sci U S A 97(8): 4046-51. Matunis, M. J., E. Coutavas, et al. (1996). "A novel ubiquitin-like modification modulates the partitioning of the Ran-GTPa se-activating protein RanGAP1 between the cytosol and the nuclear pore complex." J Cell Biol 135(6 Pt 1): 1457-70.
132 Maya, R., M. Balass, et al. (2001). "ATM-d ependent phosphorylation of Mdm2 on serine 395: role in p53 activation by DNA damage." Genes Dev 15(9): 1067-77. Mayo, L. D., J. J. Turchi, et al. (1997 ). "Mdm-2 phosphorylation by DNA-dependent protein kinase prevents interaction with p53." Cancer Res 57(22): 5013-6. McConnell, B. B., M. Starborg, et al. (1998). "Inhibitors of cyclin-dependent kinases induce features of replicativ e senescence in early passage human diploid fibroblasts." Curr Biol 8(6): 351-4. McGowan, C. H. (2002). "Checking in on Cd s1 (Chk2): A checkpoint kinase and tumor suppressor." Bioessays 24(6): 502-11. McMasters, K. M., R. Montes de Oca Luna et al. (1996). "mdm2 deletion does not alter growth characteristics of p53-deficient embryo fibroblasts." Oncogene 13(8): 1731-6. Melchior, F. and L. Hengst (2002). "SUMO-1 and p53." Cell Cycle 1(4): 245-9. Meulmeester, E., R. Frenk, et al (2003). "Critical role for a central part of Mdm2 in the ubiquitylation of p53." Mol Cell Biol 23(14): 4929-38. Midgley, C. A., J. M. Dester ro, et al. (2000). "An N-term inal p14ARF peptide blocks Mdm2-dependent ubiquitination in vitro and can activate p53 in vivo." Oncogene 19(19): 2312-23. Migliorini, D., D. Danovi, et al. (2002). "Hdmx recruitment into the nucleus by Hdm2 is essential for its ability to regulate p53 stability and transactivation." J Biol Chem 277(9): 7318-23.
133 Migliorini, D., E. L. Denchi et al. (2002). "Mdm4 (Mdmx) regulates p53-induced growth arrest and neuronal cell death during early embryonic mouse development." Mol Cell Biol 22(15): 5527-38. Mihara, M., S. Erster, et al. (2003). "p53 has a direct apoptogenic role at the mitochondria." Mol Cell 11(3): 577-90. Mihara, M. and U. M. Moll (2003). "Detection of mitochondrial lo calization of p53." Methods Mol Biol 234: 203-9. Minty, A., X. Dumont, et al (2000). "Covalent modifica tion of p73alpha by SUMO-1. Two-hybrid screening with p73 identifies novel SUMO-1-interacting proteins and a SUMO-1 interaction motif." J Biol Chem 275(46): 36316-23. Miyashita, T., S. Krajewski, et al. (1994). "Tumor suppressor p53 is a regulator of bcl-2 and bax gene expression in vitro and in vivo." Oncogene 9(6): 1799-805. Momand, J., D. Jung, et al. (1998). "The MDM2 gene amplification database." Nucleic Acids Res 26(15): 3453-9. Momand, J., G. P. Zambetti, et al. (1 992). "The mdm-2 oncogene product forms a complex with the p53 protein and inhi bits p53-mediated transactivation." Cell 69(7): 1237-45. Montes de Oca Luna, R., D. S. Wagner, et al. (1995). "Rescue of early embryonic lethality in mdm2-deficie nt mice by deletion of p53." Nature 378(6553): 203-6. Muller, M., S. Wilder, et al. (1998). "p53 activates the CD95 (APO-1/Fas) gene in response to DNA damage by anticancer drugs." J Exp Med 188(11): 2033-45.
134 Muller, S., M. Berger, et al. (2000). "c-Jun and p53 activity is modulated by SUMO-1 modification." J Biol Chem 275(18): 13321-9. Muller, S., M. J. Matunis, et al. (1998). "Conjugation with the ubiquitin-related modifier SUMO-1 regulates the partitioning of PML within the nucleus." Embo J 17 (1): 61-70. Murphy, M., J. Ahn, et al. (1999). "Transcrip tional repression by w ild-type p53 utilizes histone deacetylases, mediated by interaction with mSin3a." Genes Dev 13(19): 2490501. Nagayama, Y., K. Shigematsu, et al. (2000). "Inhibition of angiogenesis and tumorigenesis, and induction of dormancy by p53 in a p53-null thyroid carcinoma cell line in vivo." Anticancer Res 20(4): 2723-8. Nakano, K. and K. H. Vousden (2001). "PUM A, a novel proapoptotic gene, is induced by p53." Mol Cell 7(3): 683-94. Nishida, T., F. Kaneko, et al. (2001). "Cha racterization of a novel mammalian SUMO1/Smt3-specific isopeptidase, a homologue of rat axam, which is an axin-binding protein promoting beta-caten in degradation." J Biol Chem 276(42): 39060-6. Nishida, T., H. Tanaka, et al. (2000). "A novel mammalian Smt3-specific isopeptidase 1 (SMT3IP1) localized in the nucleolus at interphase." Eur J Biochem 267(21): 6423-7. Nishimori, H., T. Shiratsuchi, et al. ( 1997). "A novel brain-specific p53-target gene, BAI1, containing thrombospondin type 1 repeats inhibits experimental angiogenesis." Oncogene 15(18): 2145-50.
135 Nobori, T., K. Miura, et al. (1994). "Deletions of the cyclin-dependent kinase-4 inhibitor gene in multiple human cancers." Nature 368(6473): 753-6. Oda, E., R. Ohki, et al. (2000). "Noxa, a BH3-only member of the Bcl-2 family and candidate mediator of p53-induced apoptosis." Science 288(5468): 1053-8. Okura, T., L. Gong, et al. (1996). "Protection against Fas/APO-1and tumor necrosis factor-mediated cell death by a novel protein, sentrin." J Immunol 157(10): 4277-81. Oliner, J. D., J. A. Pietenpol, et al. (1993). "Oncoprotein MDM2 c onceals the activation domain of tumour suppressor p53." Nature 362(6423): 857-60. Owen-Schaub, L. B., W. Zhang, et al. (1995) "Wild-type human p53 and a temperaturesensitive mutant induce Fas/APO-1 expression." Mol Cell Biol 15(6): 3032-40. Pan, Y. and J. Chen (2003). "MDM2 prom otes ubiquitination and degradation of MDMX." Mol Cell Biol 23(15): 5113-21. Parant, J., A. Chavez-Reyes, et al. (2001). "Rescue of embryonic lethality in Mdm4-null mice by loss of Trp53 suggests a nonoverlapping pathway with MDM2 to regulate p53." Nat Genet 29(1): 92-5. Parant, J. M. and G. Lozano (2003). "Disrupting TP53 in mouse models of human cancers." Hum Mutat 21(3): 321-6. Parant, J. M., V. Reinke, et al. (2001). "Org anization, expression, a nd localization of the murine mdmx gene and pseudogene." Gene 270(1-2): 277-83.
136 Perry, M. E. and A. J. Levine (1993). "Tumor-suppressor p53 and the cell cycle." Curr Opin Genet Dev 3(1): 50-4. Perry, M. E., J. Piette, et al. (1993). "The md m-2 gene is induced in response to UV light in a p53-dependent manner." Proc Natl Acad Sci U S A 90(24): 11623-7. Pickart, C. M. (1997). "Targeting of substrates to the 26S proteasome." Faseb J 11(13): 1055-66. Pickart, C. M. (2001). "Mechan isms underlying ubiquitination." Annu Rev Biochem 70: 503-33. Pomerantz, J., N. Schreiber-Agus, et al. (1998). "The Ink4a tumor suppressor gene product, p19Arf, interacts with MDM2 a nd neutralizes MDM2's inhibition of p53." Cell 92(6): 713-23. Prives, C. and P. A. Hall (1999). "The p53 pathway." J Pathol 187(1): 112-26. Quelle, D. E., F. Zindy, et al. (1995). "Altern ative reading frames of the INK4a tumor suppressor gene encode two unr elated proteins capable of inducing cell cycle arrest." Cell 83(6): 993-1000. Ramos, Y. F., R. Stad, et al. (2001). "Abe rrant expression of HDM X proteins in tumor cells correlates with wild-type p53." Cancer Res 61(5): 1839-42. Riemenschneider, M. J., R. Bu schges, et al. (1999). "Amplifi cation and overexpression of the MDM4 (MDMX) gene from 1q32 in a s ubset of malignant gliomas without TP53 mutation or MDM2 amplification." Cancer Res 59(24): 6091-6.
137 Ries, S. J., C. H. Brandts, et al. (2000) "Loss of p14ARF in tumor cells facilitates replication of the adenovir us mutant dl1520 (ONYX-015)." Nat Med 6(10): 1128-33. Rodriguez, M. S., C. Dargemont, et al. (2001). "SUMO-1 conjugati on in vivo requires both a consensus modification motif and nuclear targeting." J Biol Chem 276(16): 126549. Rodriguez, M. S., J. M. Desterro, et al (1999). "SUMO-1 modifi cation activates the transcriptional re sponse of p53." Embo J 18(22): 6455-61. Rogulski, K. R., S. O. Freytag, et al. (2000) "In vivo antitumor act ivity of ONYX-015 is influenced by p53 status and is augmented by radiotherapy." Cancer Res 60(5): 1193-6. Roth, J., M. Dobbelstein, et al. (1998). "Nucleo-cytoplasmic sh uttling of the hdm2 oncoprotein regulates the leve ls of the p53 protein via a pathway used by the human immunodeficiency virus rev protein." Embo J 17(2): 554-64. Ruas, M. and G. Peters (1998). "The p16INK4a/CDKN2A tumor suppressor and its relatives." Biochim Biophys Acta 1378(2): F115-77. Sager, R. (1991). "Senescence as a mode of tumor suppression." Environ Health Perspect 93: 59-62. Saitoh, H. and J. Hinchey (2000). "Functiona l heterogeneity of small ubiquitin-related protein modifiers SUMO-1 versus SUMO-2/3." J Biol Chem 275(9): 6252-8. Sakaguchi, K., J. E. Herrera, et al. (1 998). "DNA damage activates p53 through a phosphorylation-acetylation cascade." Genes Dev 12(18): 2831-41.
138 Scheffner, M., B. A. Werness, et al. ( 1990). "The E6 oncoprotein encoded by human papillomavirus types 16 and 18 prom otes the degradation of p53." Cell 63(6): 1129-36. Schindl, M., P. Birner, et al. (2001). "Downregulation of KAI1 metastasis suppressor protein is associated with a dismal prognosis in epithelial ovarian cancer." Gynecol Oncol 83(2): 244-8. Schmidt, D. and S. Muller (2002). "Members of the PIAS family act as SUMO ligases for c-Jun and p53 and repress p53 activity." Proc Natl Acad Sci U S A 99(5): 2872-7. Serrano, M., G. J. Hannon, et al. (1993). "A new regulatory motif in cell-cycle control causing specific inhibition of cyclin D/CDK4." Nature 366(6456): 704-7. Sharp, D. A., S. A. Kratowicz, et al. (1999). "Stabilization of the MDM2 oncoprotein by interaction with the structur ally related MDMX protein." J Biol Chem 274(53): 38189-96. Sharpless, N. E. and R. A. DePinho (1999) "The INK4A/ARF locus and its two gene products." Curr Opin Genet Dev 9(1): 22-30. Shen, Z., P. E. Pardington-Purtymun, et al. (1996). "UBL1, a human ubiquitin-like protein associating with hu man RAD51/RAD52 proteins." Genomics 36(2): 271-9. Shenoy, S. K., P. H. McDonald, et al. (2001). "Regulation of receptor fate by ubiquitination of activated beta 2-ad renergic receptor and beta-arrestin." Science 294(5545): 1307-13. Sherr, C. J. (2001). "The INK4a/AR F network in tumour suppression." Nat Rev Mol Cell Biol 2(10): 731-7.
139 Sherr, C. J. and J. M. Roberts (1995). "I nhibitors of mammalian G1 cyclin-dependent kinases." Genes Dev 9(10): 1149-63. Shieh, S. Y., M. Ikeda, et al. (1997). "DNA damage-induced phosphorylation of p53 alleviates inhibition by MDM2." Cell 91(3): 325-34. Shvarts, A., M. Bazuine, et al. (1997). "Isolation and identification of the human homolog of a new p53-bi nding protein, Mdmx." Genomics 43(1): 34-42. Shvarts, A., W. T. Steegenga, et al. ( 1996). "MDMX: a novel p53-bi nding protein with some functional prope rties of MDM2." Embo J 15(19): 5349-57. Sigal, A. and V. Rotter (2000). "Oncogenic mutations of the p53 tumor suppressor: the demons of the guardian of the genome." Cancer Res 60(24): 6788-93. Smith, M. L., I. T. Chen, et al. (1994). "Interaction of the p53-regulated protein Gadd45 with proliferating ce ll nuclear antigen." Science 266(5189): 1376-80. Song, S. Y., S. K. Lee, et al. (2002). "Expression of maspin in colon cancers: its relationship with p53 expression and microvessel density." Dig Dis Sci 47(8): 1831-5. Stad, R., N. A. Little, et al. (2001). "Mdmx stabilizes p53 and Mdm2 via two distinct mechanisms." EMBO Rep 2(11): 1029-34. Stad, R., Y. F. Ramos, et al. ( 2000). "Hdmx stabilizes Mdm2 and p53." J Biol Chem 275(36): 28039-44.
140 Stein, G. H., L. F. Drullinger, et al. (1999) "Differential roles for cyclin-dependent kinase inhibitors p21 and p16 in the mechanisms of senescence and differentiation in human fibroblasts." Mol Cell Biol 19(3): 2109-17. Stelter, P. and H. D. Ulrich (2003). "C ontrol of spontaneous and damage-induced mutagenesis by SUMO a nd ubiquitin conjugation." Nature 425(6954): 188-91. Sternsdorf, T., K. Jensen, et al. (1997). "Evidence for c ovalent modification of the nuclear dot-associated proteins PML and Sp100 by PIC1/SUMO-1." J Cell Biol 139(7): 1621-34. Stott, F. J., S. Bates, et al. (1998). "The alternative product from the human CDKN2A locus, p14(ARF), participates in a regul atory feedback loop with p53 and MDM2." Embo J 17(17): 5001-14. Tanimura, S., S. Ohtsuka, et al. (1999). "MDM2 interacts with MDMX through their RING finger domains." FEBS Lett 447(1): 5-9. Tao, W. and A. J. Levine (1999). "Nucleoc ytoplasmic shuttling of oncoprotein Hdm2 is required for Hdm2-mediated degradation of p53." Proc Natl Acad Sci U S A 96(6): 307780. Thrower, J. S., L. Hoffman, et al. (2000) "Recognition of the polyubiquitin proteolytic signal." Embo J 19(1): 94-102. Tibbetts, R. S., K. M. Brumbaugh, et al. (1999). "A role for ATR in the DNA damageinduced phosphorylation of p53." Genes Dev 13(2): 152-7.
141 Tsujimoto, Y. (2003). "Cell death regulation by the Bcl-2 protein family in the mitochondria." J Cell Physiol 195(2): 158-67. Tsytsykova, A. V., E. N. Tsitsikov, et al (1998). "The mouse genome contains two expressed intronless retroposed pseudogenes for the sentri n/sumo-1/PIC1 conjugating enzyme Ubc9." Mol Immunol 35(16): 1057-67. Vaziri, H. (1997). "Critical te lomere shortening regulated by the ataxia-telangiectasia gene acts as a DNA damage signal leading to activation of p53 protein and limited lifespan of human diploid fibroblasts. A review." Biochemistry (Mosc) 62(11): 1306-10. Vaziri, H., M. D. West, et al. (1997). "ATM-dependent te lomere loss in aging human diploid fibroblasts and DNA da mage lead to the post-tra nslational activation of p53 protein involving poly(ADP-ribose) polymerase." Embo J 16(19): 6018-33. Wagner, A. J., J. M. Kokontis, et al. (1994) "Myc-mediated apoptosis requires wild-type p53 in a manner independent of cell cycle arrest and the abili ty of p53 to induce p21waf1/cip1." Genes Dev 8(23): 2817-30. Wang, X., T. Arooz, et al. (2001). "MDM2 and MDMX can interact differently with ARF and members of the p53 family." FEBS Lett 490(3): 202-8. Waterman, M. J., E. S. Stavridi, et al (1998). "ATM-depende nt activation of p53 involves dephosphorylation and associ ation with 14-3-3 proteins." Nat Genet 19(2): 1758. Weber, J. D., L. J. Taylor, et al. (1999). "Nucleolar Arf sequester s Mdm2 and activates p53." Nat Cell Biol 1(1): 20-6.
142 Weinberg, R. A. (1995). "The retinobla stoma protein and cell cycle control." Cell 81(3): 323-30. Weissman, A. M. (2001). "Themes and variations on ubiquitylation." Nat Rev Mol Cell Biol 2(3): 169-78. Wilkinson, K. D. (1997). "Regulation of ubiquitin-dependent processes by deubiquitinating enzymes." Faseb J 11(14): 1245-56. Wu, G. S., T. F. Burns, et al. (1997). "KILLER/DR5 is a DNA damage-inducible p53regulated death receptor gene." Nat Genet 17(2): 141-3. Wu, X., J. H. Bayle, et al. (1993). "The p53-mdm-2 autoregulatory feedback loop." Genes Dev 7(7A): 1126-32. Xirodimas, D. P., J. Chisholm, et al. ( 2002). "P14ARF promotes accumulation of SUMO1 conjugated (H)Mdm2." FEBS Lett 528(1-3): 207-11. Xirodimas, D. P., M. K. Saville, et al. (2004). "Mdm2-mediated NEDD8 conjugation of p53 inhibits its transc riptional activity." Cell 118(1): 83-97. Yang, A. and F. McKeon (2000). "P63 and P 73: P53 mimics, menaces and more." Nat Rev Mol Cell Biol 1(3): 199-207. Yeager, T. R., S. DeVries, et al. (1998) "Overcoming cellular senescence in human cancer pathogenesis." Genes Dev 12(2): 163-74. Yin, Y., M. A. Tainsky, et al. (1992). "Wild -type p53 restores cell cycle control and inhibits gene amplification in cells with mutant p53 alleles." Cell 70(6): 937-48.
143 Zauberman, A., Y. Barak, et al. (1993 ). "Sequence-specific DNA binding by p53: identification of target sites and lack of binding to p53 MDM2 complexes." Embo J 12(7): 2799-808. Zauberman, A., D. Flusberg, et al. (1 995). "A functional p53responsive intronic promoter is contained with in the human mdm2 gene." Nucleic Acids Res 23(14): 258492. Zhang, Y. and Y. Xiong (1999). "Mutations in human ARF exon 2 di srupt its nucleolar localization and impair it s ability to block nuclear export of MDM2 and p53." Mol Cell 3(5): 579-91. Zhang, Y. and Y. Xiong (2001). "A p53 amino-te rminal nuclear export signal inhibited by DNA damage-induced phosphorylation." Science 292(5523): 1910-5. Zhang, Y., Y. Xiong, et al. (1998). "ARF prom otes MDM2 degradation and stabilizes p53: ARF-INK4a locus deletion impairs both the Rb and p53 tumor suppression pathways." Cell 92(6): 725-34.
144 About the Author Yu Pan received her degree of Medical Sciences in Stomatology from the Capital University of Medical Sciences, Beijing, Ch ina in 1999. In fall 2000, Yu Pan joined the Ph.D. program in the department of biology in University of South Florida. During the year 2000 to 2001, she was granted both Teaching Assistantship and Research Assistantship by the department of Biol ogy. From fall 2001, Yu Pan transferred to Cancer Biology program in H. Lee Moffitt Ca ncer Center & Research Institute and resumed her goal toward a Ph.D. degree unde r the instruction of Dr Jiandong Chen. While in the Cancer Biology Ph.D. program, Yu Pan published a first author paper in Mol Cell Biol and co-authored two other manuscripts.