Regulation of the Tumor Suppressor P53 by Sumoylation and Identification of Two Independent Sumo-Interacting Motifs in Daxx

Material Information

Regulation of the Tumor Suppressor P53 by Sumoylation and Identification of Two Independent Sumo-Interacting Motifs in Daxx
Santiago, Aleixo
Place of Publication:
[Gainesville, Fla.]
University of Florida
Publication Date:
Physical Description:
1 online resource (107 p.)

Thesis/Dissertation Information

Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Medical Sciences
Molecular Cell Biology (IDP)
Committee Chair:
Liao, Daiqing
Committee Members:
LuValle, Phyllis A.
Aris, John P.
Bloom, Linda B.
Graduation Date:


Subjects / Keywords:
Amino acids ( jstor )
Antibodies ( jstor )
Cytoplasm ( jstor )
DNA ( jstor )
Journalism ( jstor )
Proteins ( jstor )
Sumo ( jstor )
Tumors ( jstor )
Ubiquitins ( jstor )
Yeasts ( jstor )
Molecular Cell Biology (IDP) -- Dissertations, Academic -- UF
Electronic Thesis or Dissertation
born-digital ( sobekcm )
Medical Sciences thesis, Ph.D.


Posttranslational modification with small ubiquitin related modifier (SUMO) is an important regulatory mechanism of protein function. SUMO is an evolutionary conserved protein with a molecular weight of 10kDa in humans. It is covalently linked to certain lysine residues of target proteins. SUMO modification impacts subcellular localization of a modified protein, protein-protein interactions, and transcriptional regulation. One target protein of sumoylation is p53, a tumor suppressor protein. Research highlighting the role of SUMO modification of this protein is controversial. There have been conflicting reports regarding the effects of SUMO modification on p53 activities, particularly how it impacts transcriptional regulation of some p53 target genes and if SUMO modification affects subcellular localization of p53. We performed a comprehensive molecular biological study employing two unique approaches: (1) fusion of SUMO to the p53's C-terminus and (2) chemical induced dimerization of SUMO and p53. Our results have shown that when p53 is modified by SUMO, its biochemical functions are attenuated. It can no longer arrest clonogenic growth and activate the expression of its target genes involved in cell-cycle arrest or apoptosis. Furthermore, it appears that p53 is exported from the nucleus to the cytoplasm upon chemical induced sumoylation. This nuclear export may involve the CRM1-mediated pathway. Surprisingly, our results suggest that CRM1 interacts directly with the p53 tetramer. Thus, sumoylation of p53 may have an important role in promoting the nuclear export of p53 and that dissociation of the p53 tetramer required for the exposure of the nuclear export signal (NES) might not be necessary for CRM1-dependent nuclear export of p53, challenging the prevalent paradigm regarding the nuclear export of p53. Another protein that SUMO interacts with is Daxx. The death-associated protein, Daxx, is essential for embryonic development and implicated in apoptosis and transcriptional regulation. Here we show that in addition to a conserved core of about 200 residues, Daxx possesses several conserved domains and two essentially invariable short SUMO-interacting motifs (SIMs) both of which can independently interact with SUMO. Daxx is known to interact with the SUMO E2 conjugating enzyme, Ubc9, responsible for SUMO conjugation to substrates via the Ubc9-SUMO complex. This interaction strictly requires at least one SIM. Interestingly, the Ubc9 H20D mutation that abolishes non-covalent Ubc9-SUMO interaction also interrupts Daxx-Ubc9 interaction. Thus, SUMO serves as the intermediate for Daxx-Ubc9 interaction. Remarkably, Daxx strongly stimulates c-Jun-mediated transcription and both SIMs are required for this stimulation. These results suggest that the conserved SIMs are involved in mediating protein-protein interactions that underlie Daxx's diverse cellular functions. Overall, these studies provide important knowledge about cellular functions of sumoylation with respect to p53 and Daxx and may have implications in our understanding of tumorigenesis and anticancer therapy. ( en )
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis (Ph.D.)--University of Florida, 2009.
Adviser: Liao, Daiqing.
Electronic Access:
Statement of Responsibility:
by Aleixo Santiago.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Santiago, Aleixo. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
LD1780 2009 ( lcc )


This item has the following downloads:

Full Text




2 2009 Aleixo S. Santiago


3 To my parents Louis and Pearl Santiago


4 ACKNOWLEDGMENTS First and foremost, I thank God for all that he has done for me. I thank my parents Louis and Pearl for always providing the love, support, and encouragement needed to keep my path straight and pursue my dreams. I thank Simon Mendes, Dr. Robert Sparks and Dr. Haimanti Dorai who, knowingly or not, motivated, inspired, and challenged me to be what I am today. I thank my family and friends who have been ther e in not only good times but also in bad times. I thank Lisa Zhao for her technica l support. I thank Dr. Liao fo r his mentorship, guidance, and knowledge.


5 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........7 LIST OF FIGURES................................................................................................................ .........8 ABSTRACT....................................................................................................................... ............10 CHAPTER 1 SUMOYLATION OF THE TUMOR SUPPRESSOR P53....................................................12 The Tumor Suppressor p53 in Cancer....................................................................................12 The Tumor Suppressor p53: Gu ardian of the Genome...................................................12 Regulation of p53 by Mdm2............................................................................................14 Regulation of p53 by the Small Ubi quitin Related Modifier SUMO.....................................15 The SUMO Proteins........................................................................................................16 SUMO-Modification.......................................................................................................17 SUMO/Sentrin Specific Endopeptidases (SENPs)..........................................................18 The CRM1 Exporter.............................................................................................................. .20 SUMO-Modification of P53...................................................................................................20 2 CELLULAR FUNCTIONS OF DAXX.................................................................................23 Daxx and Apoptosis............................................................................................................. ...23 Daxx and Transcription......................................................................................................... .23 3 MATERIALS AND METHODS...........................................................................................27 Materials...................................................................................................................... ...........27 Antibodies..................................................................................................................... ..........27 Cell Lines..................................................................................................................... ...........27 Cloning p53-SUMO Fusion Constructs..................................................................................27 Cloning DNA Constructs for Rapamycin-induced Fusion.....................................................28 Transfection................................................................................................................... .........29 Immunofluorescence Microscopy..........................................................................................29 Luciferase Reporter Gene Assays...........................................................................................30 Colony Formation Assays.......................................................................................................3 0 Real Time PCR.................................................................................................................. .....30 Yeast Two-Hybrid Assays......................................................................................................31 Glutathione-S-Transferase (GST) Pull-Down Assays............................................................31 Sequence Analysis.............................................................................................................. ....32 FLAG Immunoprecipitations (FLAG IPs).............................................................................32


6 4 SUMO-MODIFICATION OF P53 NE GATIVELY REGULATES P53-MEDIATED FUNCTIONS...................................................................................................................... ....33 SUMO-Modified p53 is Unable to Suppress Clonogenic Growth.........................................34 SUMO Modification Inhibits Transc ription of p53 Target Genes.........................................34 Effects of p53 Sumoylation thr ough Inducible Heterodimerization.......................................36 Effects of Rapamycin-Induced p53-SUMO Heterodimerization on p53-Mediated Transcription.................................................................................................................. .....38 SUMO Modification Promotes Nuclear Export of p53..........................................................39 Mdm2 might be Involved in Promo ting Cytoplasmic Localization of p53............................40 Involvement of the SUMO Po cket that Binds to SIM............................................................41 P53 is Exported as a Tetramer................................................................................................42 CRM1 has a Possible SIM......................................................................................................45 Potential Roles of SENP in Nuclear Export of p53................................................................47 5 IDENTIFICATION OF TWO INDEPE NDENT SUMO-INTERA CTING MOTIFS IN DAXX: EVOLUTIONARY CONSERVATION FROM DROSOPHILA TO HUMANS AND THEIR BIOCHEMICAL FUNCTIONS.......................................................................68 Daxx Possesses Two Independent SUMO-Interacting Motifs (SIMs)...................................68 Interaction between Daxx and SUMO Paralogs.....................................................................69 The Two SIMs of Daxx are Conserved during Evolution......................................................70 Preservation of Functional Domains of the Da xx Family of Proteins during Evolution........72 SUMO mediates Daxx-Ubc9 Inte raction via the SIMs of Daxx............................................73 Roles of Daxx SIMs in Daxx-SUMO and Daxx-PML Interaction.........................................74 Interplay of Daxx and PML in Regulat ing c-Jun-Mediated Transcription.............................75 Roles of c-Jun Sumoylation in Daxx-Mediated Coactivation................................................76 6 SUMMARY AND CONCLUSIONS.....................................................................................87 Discussion of SUMO-Modification of P53............................................................................87 Discussion of SUMO-Intera cting Motifs of Daxx..................................................................90 REFERENCES..................................................................................................................... .........95 BIOGRAPHICAL SKETCH.......................................................................................................107


7 LIST OF TABLES Table page 1-1 Ubiquitin like proteases (Ulps) in yeast and their cellular localization............................19 1-2 SUMO/Sentrin specific endopeptidases in humans and their cellular localization..........19


8 LIST OF FIGURES Figure page 1-1 P53 Schematic............................................................................................................. ......21 1-2 SUMO Conjugation/deconjugation cycle.........................................................................22 2-1 Cellular Functions of Daxx...............................................................................................2 6 4-1 P53-SUMO constructs are una ble to suppress clonogenic growth...................................48 4-2 Luciferase reporter assays on di fferent p53 target gene promoters..................................49 4-3 Real-time analysis examining the e ffects of p53-SUMO cons tructs on the mRNA levels of p53 targets.......................................................................................................... .50 4-4 Verification of p53 constructs for chemical induced fusion.............................................51 4-5 Subcellular local ization of wt p53 and the p53-2xFKBP fusion......................................52 4-6 SUMO modification of p53 inhibited p53-mediated transactivation of the p21 promoter....................................................................................................................... ......53 4-7 SUMO modification of p53 affect ed p53Â’s subcellula r localization................................54 4-8 SUMO modification of p53 influe nced its subcellu lar localization.................................55 4-9 Mdm2 is a potential mediator in ce llular localization of SUMO-modified p53...............57 4-10 The SIM-interacting po cket of SUMO is importan t for nuclear export of p53................59 4-11 Effects of rapamycin-mediated atta chment of SUMO3-FRB to p53-2xFKBP on its hetero-oligomerization with FLAG-p53............................................................................61 4-12 SUMO-modification of p53 does not disrupt p53 oligomerization..................................62 4-13 P53 tetramer is required for inte raction with the CRM1 export protein...........................63 4-14 Interactions between CRM1 and p53-SUMO fusion constructs.......................................64 4-15 The putative SIM of CRM1 and the K 386 sumoylation site of p53 influence CRM1p53 interaction................................................................................................................ ...65 4-16 Potential regulation of the nuclear export of p53 by SENPs............................................66 5-1 Daxx interacts with SUMO1 via SIMs.............................................................................78 5-2 Sequence alignment of evolutionarily conserved domains of the Daxx orthologs...........79


9 5-3 The Daxx SIMs in the Drosophila genus..........................................................................80 5-4 Evolutionary conservation of domain structure of the Daxx family of proteins..............81 5-5 Daxx interacts with Ubc9 via SIMs..................................................................................82 5-6 Colocalization of Daxx and its mutants with SUMO3 and PML.....................................83 5-7 Daxx coactivates c-Junregulated transcription................................................................85 5-8 Independent regulation of c-Jun by Daxx and TSA..........................................................86 6-1 P53 export in the context of SUMO-modification............................................................93 6-2 A model explaining Daxx-PML interaction.....................................................................94


10 Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy REGULATION OF THE TUMOR SUPPRES SOR P53 BY SUMOYLATION AND IDENTIFICATION OF TWO INDEPENDENT SUMO-INTERACTING MOTIFS IN DAXX By Aleixo S. Santiago May 2009 Chair: Daiqing Liao Major: Medical Sciences--Molecular Cell Biology Posttranslational modification with small ubiquitin related modifier (SUMO) is an important regulatory mechanism of protein functi on. SUMO is an evolutionary conserved protein with a molecular weight of 10kDa in humans. It is covalently linked to cer tain lysine residues of target proteins. SUMO modification impacts s ubcellular localization of a modified protein, protein-protein interactions, and transcriptional regulation. One targ et protein of sumoylation is p53, a tumor suppressor protein. Research highligh ting the role of SUMO modification of this protein is controversial. Ther e have been conflicting reports regarding the effects of SUMO modification on p53 activities, partic ularly how it impacts transcri ptional regulation of some p53 target genes and if SUMO modification affects subcellular localization of p53. We performed a comprehensive molecular biological study em ploying two unique approaches: (1) fusion of SUMO to the p53Â’s C-terminus and (2) chemi cal induced dimerizati on of SUMO and p53. Our results have shown that when p53 is modifi ed by SUMO, its biochemical functions are attenuated. It can no longer arrest clonogenic gr owth and activate the expression of its target genes involved in cell-cy cle arrest or apoptosis. Furthermore, it appears that p53 is exported from the nucleus to the cytoplasm upon chemical i nduced sumoylation. This nuclear export may involve the CRM1-mediated path way. Surprisingly, our results suggest that CRM1 interacts


11 directly with the p53 tetramer Thus, sumoylation of p53 may have an important role in promoting the nuclear export of p53 and that di ssociation of the p53 tetramer required for the exposure of the nuclear export signal (NES) might not be necessary for CRM1-dependent nuclear export of p53, challenging the prevalent paradigm regarding the nuclear export of p53. Another protein that SUMO interacts with is Daxx. The d eath-associated protein, Daxx, is essential for embryonic development and im plicated in apoptosi s and transcriptional regulation. Here we show that in addition to a conserved core of about 200 residues, Daxx possesses several conserved domains and two es sentially invariable short SUMO-interacting motifs (SIMs) both of which can independently in teract with SUMO. Daxx is known to interact with the SUMO E2 conjugating enzyme, Ubc9, re sponsible for SUMO conjugation to substrates via the Ubc9-SUMO complex. This interaction stri ctly requires at least one SIM. Interestingly, the Ubc9 H20D mutation that abolishes non-cova lent Ubc9-SUMO interaction also interrupts Daxx-Ubc9 interaction. Thus, SUMO serves as the intermediate for Daxx-Ubc9 interaction. Remarkably, Daxx strongly stimulates c-Jun-medi ated transcription and both SIMs are required for this stimulation. These resu lts suggest that the conserved SIMs are involved in mediating protein-protein interactions that under lie DaxxÂ’s diverse cellular functions. Overall, these studies provide important knowledge about cellu lar functions of sumoylation with respect to p53 and Daxx and ma y have implications in our understanding of tumorigenesis and anticancer therapy.


12 CHAPTER 1 SUMOYLATION OF THE TUMOR SUPPRESSOR P53 The Tumor Suppressor p53 in Cancer P53 is a transcription fact or that is encoded by the TP53 gene located on chromosome 17 in humans (Lane 1992). The main function of this protein is its abilit y to act as a tumor suppressor by regulating cell cycl e arrest and apoptosis (Harms Nozell et al. 2004; Green and Chipuk 2006). Mutations in the TP53 gene have been implicated in over 50% of cancers worldwide (Vousden and Lu 2002). Therefore unders tanding the functions a nd regulation of this protein is of key importance in cancer biology and the development of potentia l applications in cancer therapy. The Tumor Suppressor p53: Guardian of the Genome The transcription factor, p53, is involved in a wide variety of respons es in the event of cellular stresses such as DNA damage and overexpressed oncogenes. Under normal conditions, p53 levels are low (Jones, Roe et al. 1995; Leve illard, Gorry et al. 1998; Toledo and Wahl 2006; Toledo and Wahl 2007). However, in the event of these cellular stresses, p53 levels accumulate in the nucleus, leading to transcri ption of its target genes involved in cell cycle arrest (p21) and apoptosis (PUMA). P53 can also activate apoptosis in a transc ription independent manner where it is sequestered into the cytoplasm and interact s with Bcl-XL and Bcl-2 to activate the caspase cascade and subsequent apoptosis (Marchenko, Za ika et al. 2000; Sansom e, Zaika et al. 2001; Mihara, Erster et al. 2003). P53 is a modular protein consisting of five domains: a transactivation domain (TAD) located in the N-terminal regi on, a proline rich domain (PRD), a DNA binding domain (DBD), a tetramerization domain (4D), a nd a regulatory C-terminal domain (CTD) (Figure 1-1) (Toledo and Wahl 2006; Olsson, Manzl et al. 2007).


13 The TAD domain (amino acids 1-63) consists of a binding site for the regulator Mdm2. Under normal conditions, Mdm2 binds to the TA D and subsequently recruits ubiquitin onto several lysine residues of p53 and the Mdm2p53-ubiquitin complex is translocated from the nucleus to the cytoplasm where it is degraded in a proteosome-dependant manner (Oliner, Pietenpol et al. 1993; Haupt, Maya et al. 1997; Kubbutat, Jones et al. 1997; Boyd, Tsai et al. 2000). In the event of stress cond itions, phosphorylation of certain serine and threonine residues by damage-response kinases on p53’s TAD alters the structure of the amphiphatic -helix thus preventing Mdm2 from binding to p53 and brings about the subsequent re cruitment of the propyl isomerase, PIN1, onto the proline-rich domain (PRD)(amino acids 80-94)(Za cchi, Gostissa et al. 2002; Zheng, You et al. 2002; Bode and Dong 2004; Ou, Chung et al. 2005; Toledo and Wahl 2006). This further provides steric inhibition for Mdm2 binding onto p53 and allows for strengthened interactions with histone acetyl tr ansferases (HATs) such as CREB-binding protein (CBP) and p300 (Lill, Grossman et al. 1997; Dornan, Shimizu et al. 2003). The DNA binding domain (DBD) (amino acids 100300) is the central core-domain that interacts directly with DNA. This domain binds to p53 response elements of target genes that consist of four copies of the pentamer cons ensus sequence PuPuPuC(A/T) where the pentamers orient in alternating directions (Kern, Kinzler et al. 1991; el-Dei ry, Kern et al. 1992; Funk, Pak et al. 1992; Cho, Gorina et al. 1994) A short stretch of sequences up to 13 base pairs may be inserted between the pentamer pairs. The p53 DNA-binding domain (DBD) contains the “hot spots” seen in various tumors (Lang, Iwa kuma et al. 2004; Olive, Tuveson et al. 2004; Hingorani, Wang et al. 2005). The tetramerization domain (4D) of p53 (a mino acids 307-355), also known as the oligomerization domain, allows for the formation of a dimer of dimers, which is critical for the


14 formation of functional p53 (Clore, Omichinski et al. 1994; Lee, Harvey et al. 1994; Jeffrey, Gorina et al. 1995). Each of the p53 subunits bi nds to one quarter site on the p53 response elements of the target genes allowing for the s ubsequent transcription of these targets (McLure and Lee 1998). Past literature has implicated the oligomerization domain as the site of the nuclear export signal (NES) which when expos ed is recognized by the CRM1 transporter and shuttles p53 from the nucleus to the cytopl asm (Freedman and Levine 1998; Stommel, Marchenko et al. 1999; Budhu and Wang 2005). Those studies are now considered debatable as our studies have demonstrated that a newly discovered CRM1 SU MO interacting motif (SIM), a region that interacts with the SUMO protein, may be involved in the export of p53 to the cytoplasm. In this situation the exposure of p53Â’s NES is not necessary (see chapter 4). The C-terminal domain (amino acids 356-393) is a regulatory domain. This domain is the site of many posttranslational modifications (T oledo and Wahl 2006; Olsson, Manzl et al. 2007). These modifications include phosphorylation, acetylation, ubiquitin-like modifications (neddylation, ubiquitination, and sumoylation), an d methylation. In vitro studies have shown that posttranslational modifications are important for regulating p53 activ ity (Toledo and Wahl 2006). Regulation of p53 by Mdm2 Amplification of the human homolog of murine double minute (Mdm2) gene that encodes for a 491 amino acid protein is implicated in a bout 10% of cancers worldw ide (Toledo and Wahl 2006). In both in vitro and in vivo studies, the Mdm2 oncogene has been shown to function as a negative regulator of the tumor suppressor protei n p53. Deficiency of Mdm2 in murine models leads to embryonic lethality that can be rescue d by loss of p53 (Parant, Chavez-Reyes et al. 2001; Migliorini, Lazzerini Denchi et al 2002; Laurie, Donovan et al. 2006).


15 The human homolog of Mdm2 consists of thr ee basic domains: an N-terminal domain that binds to the N-terminal of p53, a zinc finger dom ain whose functional significance needs to be determined, and a C-terminal RING domain (S hvarts, Steegenga et al. 1996; Jackson and Berberich 2000). The RING domain of MDM2 is essential for its action as an E3-ubiquitinligase (Shvarts, Steegenga et al. 1996; Jackson and Berberich 2000). As mentioned, Mdm2 is an E3 ubiquitin ligase that binds to p53Â’s TAD and inhibits p53 functions in two ways: (1) by play ing the role of a steric block and inhibiting transcription cofactors to bind to p53 and (2) by covalent attach ment of ubiquitin onto lysine residues on the Cterminal regulatory domain of p53 (Tol edo and Wahl 2006; To ledo and Wahl 2007). Furthermore, Mdm2 can recruit p300, which in this case can act as an E4 ligase and together can bring about poly-ubiquitination of p53 (Zhu, Yao et al. 2001; Gro ssman, Deato et al. 2003). This allows for the export of the Mdm2-p53 complex in to the cytoplasm where it is degraded in a proteosomal dependant fashion (Zhu, Yao et al. 2001; Grossman, Deato et al. 2003). Mdm2 has also been implicated in covale nt modification of other ubiquitin -like proteins such as Nedd8 and SUMO (Harper 2004; Di Ventura, Funaya et al. 2 008). Interestingly, Mdm2 is also a target of p53-mediated transcription as it contains the response elements recognized by the p53 tetramer (Lahav 2008). Regulation of p53 by the Small Ubiquitin Related Modifier SUMO Post-translational protei n modifications regulate a wide variety of cellular activities such as modulation of protein activity, stab ility, protein-protein interacti ons, and subcellu lar localization (Dohmen 2004). Posttranslational modificat ions include phosphorylation, acetylation, methylation, and ubiquitin-like modifications th at include ubiquitination, neddylation, and the newly discovered sumoylation. The process of sumoylation involve s the protein SUMO.


16 SUMO (s mall u biquitin related mo difier) belongs to the ubiquitin family of proteins that also consist of ubiquitin and Nedd prot eins. It was initially discovered in Saccharomyces cerevisae in a genetic screen for Mif2 suppressors (Meluh and Koshland 1995; Geiss-Friedlander and Melchior 2007) and was later found to be cova lently attached to th e Ran GTPase-activating protein RanGAP1 (Matunis, Coutavas et al. 199 6; Mahajan, Delphin et al. 1997). The discovery of this covalent modification led to a huge burst in research to de termine its functional effects. The SUMO Proteins SUMO proteins are ubiquitously expressed in eukaryotic systems. Yeast is known to have only one SUMO gene ( SMT3 ) (Johnson, Schwienhorst et al. 1997; Geiss-Friedlander and Melchior 2007). Humans on the otherhand are know n to have SUMO genes that encode for four SUMO proteins (SUMO1-4) (Melchior 2000; Do hmen 2004; Geiss-Friedlander and Melchior 2007). SUMO1-3 (Geiss-Friedlande r and Melchior 2007) are ubiqu itously expressed throughout the human body whilst SUMO4 is seen strongly expressed in lymph nodes, kidney, and the spleen (Guo, Li et al. 2004; Geiss-Friedla nder and Melchior 2007). SU MO2 and SUMO3 share a 97% sequence homology whilst SUMO1 shares a 47% sequence homology to both SUMO2 and 3 (Geiss-Friedlander and Melchior 2007). SUMO1 and SUMO2-3 have distinct functions as they can be differentially conjugated to various target proteins in vivo SUMO4 however is enigmatic as it is not known if it can be processed to it s mature form and be covalently attached to substrates (Guo, Li et al. 2004; Owerbach, McKay et al. 2005). SUMO modifications seem to be an essential process in many organisms. Disruption of the SMT3 gene in budding yeast leads to lethality (Alkuraya, Saadi et al. 2006). An ex ception however is seen in fission yeast where disruption of the SMT3 gene leads to sick but nonetheless viable cells (Tanaka, Nishide et al. 1999).


17 The SUMO proteins are about 10kDa in size and are similar to ubiquitin in that they have similar crystal structures (Bayer, Arndt et al. 1998; Mossessova and Lima 2000). SUMO however shares only 17% sequence homology w ith ubiquitin and they have different distributions of surface charges (Dohmen 2004; Geiss-Friedlander and Melchior 2007). SUMO proteins carry a stretch of 10-25 amino acids at the N-terminus that is unst ructured in nature and absent in ubiquitin (Geiss-Friedla nder and Melchior 2007). It is here at the N-termini that the formation of SUMO chains is observed for so me of the SUMO family proteins (GeissFriedlander and Melchior 2007). SUMO-Modification SUMO is linked to the target proteins through an isopeptid e bond between the C-terminal glycine of SUMO and an internal lysine residue of the protein in the context of a consensus sequence KXE (where is a large hydrophobic amino acid and X is any amino acid residue) (Rodriguez, Dargemont et al. 2001; Verger, Perdomo et al. 2003; Dohmen 2004; Hilgarth, Murphy et al. 2004). The process of SUMOylation (Figure 1-2) st arts with an inactive precursor of SUMO bearing a short C-terminal peptide that is cleav ed by sentrin/SUMO spec ific proteases (SENPs) to give mature glycine-glycine C-terminal resi dues (Rodriguez, Dargemon t et al. 2001; Dohmen 2004; Li, Evdokimov et al. 2004). SUMO is then ac tivated in an ATP-dependent fashion where the adenylated SUMO is bound by a thioester bo nd with the E1 activating enzyme, a complex that consists of a SAE1-SAE2 heterodimer in humans or AOS1-UBA2 in yeast (Verger, Perdomo et al. 2003; Dohmen 2004). The activated SUMO is then transferred to the E2 conjugating enzyme Ubc9 (Melchior and Hengs t 2002; Dohmen 2004; Lin, Ohshima et al. 2004). An E3 ligase then stimulates SUMO conjuga tion to the amino group of a specific lysine


18 in the target substrate (Kahyo, Nishida et al. 2001; Schmidt and Muller 2003; Dohmen 2004). The E3 ligases are best characterized by the pres ence of an SP-RING motif that is essential for their function. The domain is predicted to resemb le the RING domain of ubiquitin E3 ligases. These SP-RING ligases non-covalently attach to SUMO via SUMO Interacting Motifs (SIM) (Geiss-Friedlander and Melchior 20 07). These proteins include the PIAS family of proteins and the nuclear protein RanBP2. Inte restingly, literature has shown that the ubiquitin E3 ligase, Mdm2, can also promote SUMOylation of some substrates (Chen and Ch en 2003; Di Ventura, Funaya et al. 2008). The entire process is dynamic and reversible (Dohmen 2004; GeissFriedlander and Melchior 2007) SUMO conjugation can be removed by SENPs (Dohmen 2004; Mukhopadhyay and Dasso 2007). SUMO/Sentrin Specific E ndopeptidases (SENPs) Ubiquitin like proteases (Ulp) (Table 11) and their counterpart human homologs, SUMO/Sentrin specific endopeptid ases (SENPs) (Table 1-2), are the key enzymes responsible for SUMO maturation and SUMO -deconjugation from the target substrate (Mukhopadhyay and Dasso 2007). The Ulps/SENPs directly regulate the pools of free and conjugatable SUMO and also regulates the ha lf-life of the conjugated specie s (Mukhopadhyay and Dasso 2007). Budding yeast has two Ulp/SENPs whilst humans have six and Arabidopsis has seven to date (Novatchkova, Budhiraja et al. 2004; Colby, Matt hai et al. 2006). In yeast, Ulp1p and 2p are related to adenoviral proteases a nd belong to the C48 family of proteases (Li and Hochstrasser 1999; Mossessova and Lima 2000). Both Ulps and SENPs have the catalytic C48 domains that are involved in nucleophillic a ttack of the isopeptide bond linking the SUMO to the substrate (Mukhopadhyay and Dasso 2007). Normally, this catal ytic domain is located close to the Cterminus though some exceptions exist especially in SENP6 and 7 where the domains are split in two (Mukhopadhyay and Dasso 2007).


19 Interestingly, both Ulp1p and the Ulp1p hu man homolgs, SENP1 and 2, have nuclear export signals and Ulp1p has been shown to re gulate nuclear-cytoplasmic trafficking and in particular, the export of the 60S pre-ribosomal part icle (Stade, Vogel et al 2002; Panse, Kressler et al. 2006; Mukhopadhyay and Dasso 2007). SENP 1 and SENP2 are known to localize around the nuclear pore (Hang and Dasso 2002; Zhang, Saitoh et al. 2002; Bailey and O'Hare 2004), whilst SENP3 and 5 are known to be present in the nucleolus (Di Bacco, Ouyang et al. 2006; Gong and Yeh 2006; Mukhopadhyay and Dasso 2007) SENP6 and 7 are localized in the nucleoplasm (Cheng, Bawa et al. 2006; Mukhopadhya y, Ayaydin et al. 2006). Each SENP also has a preference for SUMO1-3 paralogs (M ukhopadhyay and Dasso 2007). With respect to nuclear-cytoplasmic trafficking, it has been shown in vertebrates that the extreme end portion of the N-terminus of SENP2 is necessary for its association with the nuc leoplasmic face of NPC (Hang and Dasso 2002). Importantly, Ulp1p (Panse, Kuster et al. 2003), and SENP1 (Kim, Sung et al. 2005), and SENP2 (Itahana, Yeh et al. 2006) are subject to export from the nucleus via the Ran-dependant nuclear export protein, CRM1. Table 1-1. Ubiquitin like prot eases (Ulps) in yeast and their cellular localization Name Length aa NES Subcellular localization SUMO Maturation Deconjugation Ulp1p 621 Yes Nuclear peripheryYes Yes Ulp2p 1034 No Nucleoplasm No No Table 1-2. SUMO/Sentrin specific endopeptidases in humans and their cellular localization Name Length aa NES Subcellular localization SUMO Maturation Deconjugation SENP1 643 Yes Nuclear pore and nucleoplasmic speckles Yes Yes SENP2 589 Yes Nuclear pore Yes Yes SENP3 574 No Nucleolus Unknown Yes SENP5 755 No Nucleolus Unknown Yes SENP6 1112 No Nucleoplasm No No SENP7 984 No Nucleoplasm Unknown Unknown


20 The CRM1 Exporter CRM1 is a major transporter of proteins from the nucleus to the cytoplasm. It moves cargoes that contain a nuclear e xport signal (NES) rich in leuc ine residues (Fornerod and Ohno 2002). CRM1 has a ring-like structure similar to that of the yeast exporter, Cse1, in the unbound state (Petosa, Schoehn et al. 2004; Cook, Bono et al. 2007). The crystal structure identifies a region in the CRM1 protein that forms a loop in the eighth HEAT repeat that is proposed to be responsible for regulating cargo binding by an allosteric mechanism (Petosa, Schoehn et al. 2004). At present, little is known about the determ inants of cargo binding to the CRM1 system. In the nucleus, CRM1 recognizes the NES of the cargo along with RanGTP. They together form the ternary complex, which is th en localized to the nuclear pore complex (NPC) (Bischoff, Krebber et al. 1995). It is at this NPC that GTP hydrolysis occurs allowing for the release of RanGDP that destab ilizes the loop conformation and a llows for the release of the cargo from the binding site (Bis choff, Krebber et al. 1995). SUMO-Modification of P53 One target protein of sumoylation is p53. It is modified at lysine 386 (Gostissa, Hengstermann et al. 1999; Rodriguez, Desterro et al. 1999; Muller, Berg er et al. 2000; Kwek, Derry et al. 2001; Melchior and Hengst 2002; Dohmen 2004; Hilgarth, Murphy et al. 2004; Muller, Ledl et al. 2004). Lite ratures highlighting the role of SUMO modification of p53 vary. Gostissa et al. employed a lucife rase reporter construct that c ontained the p53 response elements of the p21 promoter and reported that SUMO1 m odified p53 enhanced tr anscription from the p21 promoter (Gostissa, Hengstermann et al. 1999; Rodri guez, Desterro et al. 1999). In contrast, in a different study, it was suggested that SUMO-modification ma y not influence p53-mediated transcription (Muller, Berger et al. 2000; Kwe k, Derry et al. 2001). Schmidt and Muller also


21 showed that there was no substantial increase in p53 mediated transc ription of its targets (Schmidt and Muller 2003). To elucidate the nature of sumo-modifi cation of the p53 tumor suppressor, our study utilizes two approaches: (1) a chemical induced fusion method and (2) SUMO fused to the Cterminus of p53. The chemical induced fusion method utilizes the Argent™ Regulated Heterodimerization kit from Ariad technologies (Cambridge, MA). In this system, two differentially engineered fusion proteins are br ought together through th e addition of rapamycin, a small molecule dimerizer, which bridges the two fusion domains (Zhu, Zhang et al. 2006). The second approach utilizes the p53 open readi ng frame (ORF) fused to the SUMO ORF. Our results show that sumoylation of p53 appears to attenuate its transactivation potential. Sumoylation of p53 appears to promot e nuclear export of p53, thereby reducing its nuclear concentration, leading to reduced activation of its target genes Figure 1-1. P53 Schematic. The tumor suppressor is a 396 amino acid long protein made up of five domains: (1) Transactivation domain (T AD), (2) Proline rich domain (PRD), (3) DNA binding domain (DBD), (4) Tetrameri zation/Oligomerization domain (4D), and (5) C-terminal regulatory domain (CTD).


22 Figure 1-2. SUMO Conjugation/dec onjugation cycle. SENPs cleave the SUMO precursor (1) to give mature SUMO (2) that is then linke d to the human E1 complex SAE1-SAE2 (3). The SUMO is eventually tr ansferred to the E2 enzyme Ubc9 (4). Subsequently, SUMO is covalently attached to the ta rget substrateÂ’s lysine residue in the KXE consensus amino acid sequence (5). The pro cess is reversible as SENPs deconjugate SUMO from the target substrate (6). KXE


23 CHAPTER 2 CELLULAR FUNCTIONS OF DAXX Daxx and Apoptosis Daxx, the death domain associated protein, was identified by a yeast tw o-hybrid screen as a protein interacting with the Fa s receptorÂ’s death domain (Yang, Khosravi-Far et al. 1997). It is a 120 kDa protein that is ubiquitously expresse d in mammalian cells (Kiriakidou, Driscoll et al. 1997). Both mice and human Daxx showed a 72% amino acid homology with each other (Kiriakidou, Driscoll et al. 1997) It was initially shown as a cytoplasmic protein linking Fas signaling to the JNK pathway via ASK1 (apoptosis signal-regulating kina se 1). A subsequent study showed that mouse Daxx could potentate ASK1 activation by binding to ASK1 and eventually relieving the inhibitory intramolecu lar interaction between the N-terminal and Cterminal domains of the protein (Yang, Khosravi-F ar et al. 1997). However, it has been shown that Daxx may not only exert proapoptotic effects but also antiapoptotic effects, depending on specific contexts. Homozygous deletion of the Daxx gene in mice results in embryonic lethality whilst Daxx -deficient cells undergo a poptosis suggest an anti -apoptotic role for Daxx (Michaelson, Bader et al. 1999). Daxx and Transcription Daxx (Figure 2-1) interacts with SUMO-modified PML-NB (promyelocytic leukemia nuclear bodies). PML-NB is a nuclear domain im plicated in oncogenesi s and viral infection (Boisvert, 2000). It has been reported that PML-NB is rich in RNA and regulates gene expression (Boisvert, 2000). When PML is absent, Daxx is relocated to condensed heterochromatin where it could potentially be involved in so me biochemical function (Ishov, Vladimirova et al. 2004). Subse quent studies revealed that Da xx could interact not only with core histones, but histone deacetylase 2 (HDAC 2), Dek (Hollenbach, McPherson et al. 2002) and


24 the SWI/SNF chromatin remodeling protein ATRX (Xue, Gibbons et al. 2003). All together, these interactions, amongst others, indicated the possibility that Da xx might act as a regulator of transcription. It has been s hown that Daxx interacts with p53 (Gostissa, Morelli et al. 2004; Zhao, Liu et al. 2004; Chang, Lin et al. 2005) Smad4 (Chang, Lin et al. 2005), and Pax transcription factor family members, as well as certain other transcrip tion factors (Lehembre, Muller et al. 2001; Emelyanov, Kovac et al. 2002; Hollenbach, McPherson et al. 2002). Therefore, Daxx can modulate gene expres sion mediated by tran scription factors. Interestingly, a recent study revealed that Daxx binds to small ubiquitin-like modifiers (SUMOs) when they are conjugated to specific pr oteins and this interaction is mediated through a conserved sequence termed SUMO-interacting mo tif (SIM) near the C-terminus of the Daxx protein (Lin, Huang et al. 2006). Furthermore, the Daxx-SUMO intera ction appears to be important for intranuclear targeting of Daxx to PML-NB as well as its transcriptional repression activity (Gill 2004; Lin, Huang et al. 2006). Given the well-supporte d role for sumoylation in transcriptional repression (Gill 2004), interacti on between sumoylated transcription factors and Daxx may indeed underlie Daxx-mediated repressi on. Nevertheless, it is worth noting that Daxx can potentiate heat shock factor 1 (HSF1)-mediat ed transcription (Boell mann, Guettouche et al. 2004). HSF1 is a known SUMO-med iated transcription factor (Hong, Rogers et al. 2001; Hietakangas, Ahlskog et al. 2003). Daxx-SUMO inte raction might also lead to transcriptional activation. Daxx is only found in the animal kingdom. In humans, the Daxx gene resides in chromosome 6p21.32, centromeric to the major histocompatibility complex (MHC) region. A Daxx-like protein (a.k.a DLP) is also present in the Drosophila genus and other insect species. Surprisingly, the putative Drosophila protein has 1659 amino acid residu es, much larger than the


25 human counterpart of 740 residues. Sequence co mparison indicates that a region spanning ~200 residues near the C-terminus of DLP represents the conserved domain of the Daxx family of proteins during evolution, which we refer to as the core domain. Thus, the majority of the DLP sequence is not conserved and likely dispensable for function. In the Daxx core domain, two putative coiled-coil domains are present. These co iled-coil domains are conserved in the animal kingdom and as such they probably play structur al roles; for example, the core domain may mediate homo or hetero-oligomeriz ation via the coiled-coil sequen ces. Interestingly, like human Daxx that physically binds to th e tumor suppressor p53 and regulates its activity (Kim, Park et al. 2003; Zhao, Liu et al. 2004), interaction between DLP and Drosophila p53 is also observed (Bodai, Pardi et al. 2007), suggesting that the Da xx family of proteins probably also have conserved functions. Nonetheless, it remains to be determined whether the conserved functions of Daxx family are mediated through common functional domains. In our study, we have uncovered a novel SI M near the N-terminus of the human Daxx protein. This motif, like the previously reported SIM near the C-terminus (Lin, Huang et al. 2006), interacts with SUMO moiety and is hi ghly conserved throughout evolution, suggesting that SUMO-biding property is cr itical for diverse cellular functi ons ascribed to Daxx. Additional conserved domains include a stretch of acidic residues and a highly conserved domain seen in most animal species but not in insects. Inte restingly, the sequences between the conserved domains are extremely variable in distant species Thus, preservation of essential functions is likely the principal constraint of mo lecular evolution of the Daxx family.


26 Figure 2-1. Cellular Functions of Daxx. Daxx is known to interact with SUMO-modified proteins such as PML or other protei ns (e.g. MSP58 and ATRX). In terms of transcriptional regulation, Daxx can downregulate transcri ption of certain targets through interaction with SUMO-modified transcription f actors. It can also upregulate HSF-1 mediated transcription.


27 CHAPTER 3 MATERIALS AND METHODS Materials PC4-RHE, pC4EN-F1, and pC4EN-F2 were obtained from Ariad Technologies (Cambridge, MA). pCDNA 3.1 Myc-His (+) was obtained from Invitrogen (Carlsbad, CA). SUMO-1 G97A-FRB-HA (SUMO-1-FRB) was obtai ned from Michael Mat unis (Johns Hopkins University). pExchange-3, pAdEasy-1 and pShuttleCMV vectors were obtained from Stratagene (La Jolla, CA). Rapamycin and anti-FLAG M2 -Agarose were obtained from Sigma-Aldrich (St. Louis, MO). Antibodies Anti-p53 (DO-1) mouse monoclonal anti body, anti-p53 rabbit polyclonal antibody (FL 393), and HA rabbit monoclonal were obtained fr om Santa Cruz Biotechnology (Santa Cruz, CA). Anti-FLAG rabbit polyclonal antibody was obtai ned from Sigma-Aldrich (St. Louis, MO). Anti-GFP and anti-Myc mouse monoclonal antibodies were obt ained from Babco (Berkeley, CA). Cell Lines Human colorectal cancer line HCT116 p53-/-, Osteosarcoma cell line Saos2, and lung cancer cell line H1299 are all deficient of wild-type p53. Double knockout for p53 and Mdm2 mouse embryonic fibroblasts (MEFp53/Mdm2 =/=) were also used. Cells were cultured in Dulbecco Modified Eagle Medium supplemented with 10% fetal bovine serum and 1% penicillin/streptomycin antibiotic and incuba ted at 37C in an atmosphere with 5% CO2. Cloning p53-SUMO Fusion Constructs To make the p53-SUMO fusion constructs, the DNA fragments encoding SUMO-1 and 3 were PCR amplified and inserted into th e BamHI and HindIII sites of pExchange-3B


28 (Stratagene). The DNA fragment encoding p53 was th en cloned into the BamHI site of the pExchange-3B-SUMO constructs to ge nerate p53-SUMO1 and p53-SUMO3. The DNA encoding the p53 R248W mutant was cloned into the BamHI site of the pExchange-3B-SUMO1 to generate p53-R248W-SUMO1. The DNA encoding GFP, p53, p53-SUMO1, p53-SUMO3, and p53-R248W-SUMO1 were cloned into pShuttle-CMV to generate recombinant adenovirus (Ad): Ad-GFP, Ad-wt p53, Ad-p53-SUMO-1, Ad-p53-R248W-SUMO-1, and Ad-p53-SUMO-3 using the AdEasy system (Stratagene). Cloning DNA Constructs for Rapamycin-induced Fusion For the chemical induced fusion model, pC4-RHE, pC4EN-F1, and pC4EN-F2 from Argent™ Regulated Heterodimerization kit from Ariad Technologies (Cambridge, MA) were cut with SpeI and XbaI to release the DNA frag ments encoding 1xFKBP, 2xFKBP, and the FRB respectively. The fragments were individually cl oned to pcDNA 3.1 Myc-His (+) at the XbaI site to generate pcDNA-1xFKBP-myc-His6 (1xFKB P), pcDNA-2xFKBP-myc-His6 (2xFKBP), and pcDNA-FRB-myc-His6 (FRB). The DNA fragment encoding p53 was cloned into the EcoRI site of pcDNA-1xFKBP, pcDNA-2xFKBP, and pcD NA-FRB to respectively generate pcDNA-p531xFKBP (p53-1xFKBP-myc-His6), pcDNAp53-2xFKBP (p53-2xFKBP-myc-His6) and pcDNA-p53-FRB (p53-FRB-myc-His6). SUMO3 with point mutation G92A was cloned into the EcoRI site of pC4-RHE to generate pC4-RHE-SUMO3 G92A-FRB-HA (SUMO3-FRB). SUMO1 G97A-FRB-HA (SUMO1-FRB) was obtained from Michael Matunis (Johns Hopkins University). The DNA fragment encoding SUMO3 G92A was also inserted into the Bgl II site of pEGFP-C1 to generate GFP-SUMO-3 G92A SUMO1-FRB and SUMO3-FRB were point mutated to give the F36A and F31A substitu tions respectively (SUMO1 F36A FRB and SUMO3 F31 FRB).


29 Transfection MEFp53/Mdm2 =/= and the human tumor cell lines HCT116 p53-/-, H1299, and Saos2 which were either transiently transf ected with various DNA plasmids using either Effectene (Qiagen, Valencia, CA) or Lipofectamine (Invitrogen, Carlsb ad, CA) transfection reag ent kits or infected with an Ad vector expressing GFP, p53, p53-SUMO-1, p53-SUMO-3, p53-R248W-SUMO-1. For rapamycin induced hetero-dimerization, rapamy cin was added to a final concentration of 0.1 M 6 hours post-transfection. Immunofluorescence Microscopy Saos2, H1299, and MEFp53/Mdm2 =/= cells were grown up to 30-50% confluency on glass coverslips before they were transiently tr ansfected. The cells were washed with phophate buffered saline (PBS) and fixed with 1 ml of 3% paraformaldehyde (in PBS with 0.1 mM MgCl2 and 0.1 mM CaCl2) for 15 minutes at room temperature. They were washed once with PBS and then permeablized with 1 ml of 0.2% Triton X-100 (in PBS) for 15 minutes at room temperature before being washed twice with PBS and bl ocked with 1 ml of blocking solution (PBS containing 2% FBS, 0.1% Sodium Azide, 0.1% Tw een 20) for 30 minutes at room temperature. The cells were stained with primary antibody( ies) (1:200 dilution in blocking buffer) for 30 minutes at room temperature and then with s econdary IgG rhodamine or FITC conjugate (1:200 dilution in blocking buffer) in the dark at room temperature for 30 minutes. The coverslips were mounted on glass slides using Vectashield mounting medium with DAPI from Vector Laboratories, Inc. (Burlingame, CA) and the ce lls were examined under red, green, and blue channels with Zeiss Axiophot microscope (C arl Zeiss, Thornwood, NY) using 20X and 63X objectives.


30 Luciferase Reporter Gene Assays The firefly luciferase repor ter driven by the p53 target promoters (p21, PUMA, Mdm2, PIG3, Fas, AIP, and PIDD) were used. These promoter DNA fragments were cloned into the HindIII site of pGL3-Basic (Promega, Madison, W I). The firefly luciferase reporter under the control of an artificial promoter containing five consecutive Gal4-binding sites upstream of the adenovirus E4 core promoter containing a TATA box and an initiator (Inr) element was also generated. The full-length c-Jun ORF was fused with the coding sequence for Gal4 DNAbinding domain. The reporters along with a sea pa nsy luciferase report er were transiently transfected into HCT116 p53-/and H1299 cells. Each transfecti on was done in duplicate. Dual luciferase reporter assays from Promega (Mad ison, WI) were conducted 24h after transfection. The firefly luciferase activities were nor malized against that of the sea pansy. Colony Formation Assays H1299 cells were transfected with 1 g empty v ector or that for the expression of wt p53, p53-SUMO-1, p53-SUMO-3, or p53 R248W-SUMO -1, along with 0.1 g plasmid for the expression of puromycin-resistant gene. Cell growth was selected in medium containing 2 g/ml puromycin (Sigma-Aldrich, St. Louis, MO) for tw o weeks. The plates were then stained with methylene blue. Real Time PCR H1299 cells were infected w ith adenoviral vector contai ning either GFP, wt-p53, p53SUMO-1, p53-R248W-SUMO-1, and p53-SUMO-3 in duplicate. Total cellular RNA was prepared from the mock-infected and infect ed cells using the Se rious RNA Purification™ (Gentra Systems) or Purelink™ Micro-to-Midi (Invitrogen) kits. The RNA was subjected to reverse transcription with random hexamers us ing SuperScript™ First-Strand Synthesis kit (Invitrogen). The cDNA generated wa s used for real-time PCR with SYBR green reagents using


31 PCR primers for p21 (CCGC GACTGTGATGCGCTAATG and CTCGGTGACAAAGTCGAAGTTC), PUMA (TGCGGCGGATGGCGGACGAC and TCCAGGGTGAGGGGCGGTGC), and Mdm2 (GTGAATCTACAGGGACGCCATC and CTGATCCAACCAATCACCTGAA). House keeping ge ne beta-actin mRNA was also subjected to real-time PCR using primer pa ir (GCTCCTCCTGAGCGCAAGTACTC and GTGGACAGCGAGGCCAGGAT). The relative levels of p21, PUMA, and Mdm2, were normalized against th at of beta-actin. Yeast Two-Hybrid Assays Different DNA fragments of the Daxx, p300 a nd CRM1 ORFs were fused to Gal4 DNAbinding domain (BD) in plasmids pGBDU-C(x). Site-specific mu tations of Daxx, SUMO, Ubc9, and p53 were made using Quick-Change protocol (Stratagene). The Daxx constructs were tested for interacting with SUMO or Ubc9 ORF fuse d to Gal4 AD (activation domain). The p300 and CRM1 constructs were tested for interacting with SUMO and/or p53 fused to Gal4 AD. Two complementary oligonucleot ides that encode the Drosophila melanogaster SIM1 (MSASVICVDLSSESD) or SIM2 (PVIADQIIISDEES) were annealed and ligated to the pGADC1 or pGBDU-C1 vector at the EcoRI and PstI sites. Yeast two-hybrid assays were conducted as described (Liu, Colosimo et al. 2000). Glutathione-S-Transferase (GST) Pull-Down Assays DNA fragment encoding the N (aa 1-130) or C-terminal re gion (aa 644-740) of Daxx was fused to the ORF of GST. The GST fusion proteins were expressed in and purified from E. coli. The purified GST fusion proteins immobilized in glutathione Se pharose 4B were separately incubated with 0.1 g of purified SUMO1 (Boston Bioche m, Cambridge, MA) in buffer A (50 mM Tris, pH 8, 150 mM KCl) at room temperat ure for 30 min. The beads were washed four times with buffer A. The beads were then resu spended in SDS sample buffer and the dissolved


32 samples were subjected to SDS-PAGE and Wester n blot analysis using anti-SUMO1 monoclonal antibody (clone 21C7, Zymed). Sequence Analysis The Daxx protein sequences in different spec ies were deduced from nucleotide sequences and retrieved from GenBank, Swiss-Prot or Ensembl ( The protein sequences were aligned using the ClustalW program in the MacVector software package and visually inspected to co rrect obvious misalignments. FLAG Immunoprecipitations (FLAG IPs) Cells were lysed in situ with cold buffer B [20 mM Tr is-HCl, pH 8.0, 5 mM MgCl2, 10% glycerol, 0.1% NP-40, 150 mM KCl and 100x dilu ted protease inhibito r cocktails (Sigma P8340)]. Lysate was frozen at -80C for 30 minutes and then th awed at room temperature. Lysate was rotated at 4C for 30 minutes. Ly sate was spun down at 4C for 10 minutes at 13k rpm. Half the supernatant was used for IP and th e other half used as input control. Anti-FLAG M2-Agarose (Sigma) was pretreated with 0.1 M glycine (pH 2.5) for 2 minutes at room temperature. Beads were washed twice with 0.1 M Tris (pH 8.0) The beads were mixed with clear cell extracts. The mixture was rotated at 4C for 2 hrs. It was then centrifuged at 4k rpm for 2 minutes and the supernatant aspirated. Buffer B was added to the beads. The mixture was inverted several times and the beads were set tled by centrifugation and th e supernatant removed. This step was repeated thrice. Bound proteins were eluted by in cubating with buffer B containing 0.1 mg/ml FLAG peptide at 4 C for 1 hr with ag itation, or by boiling directly in 10-20 l 2x SDS sample buffer. Samples were analyzed by Western blotting.


33 CHAPTER 4 SUMO-MODIFICATION OF P53 NEGAT IVELY REGULATES P53-MEDIATED FUNCTIONS The importance of posttransla tional SUMO modification in diverse cellular mechanisms has become increasingly apparent. One technical challenge for assessing functional impacts of sumoylation is the instability of SUMO-modified proteins. The SUMO moiety is quickly removed from p53 in the cell (Dohmen 2004). This problem might complicate several previous studies regarding the functiona l significance of p53 sumoylation. Under normal conditions, SUMO-modified p53 is undetectable, thus suggesting rapid desumoyl ation by SENPs. In order to circumvent the destabilization of SUMO-modifi ed p53 by the SENPs, stable p53-SUMO fusions were generated via two ways. The first involv ed genetic manipulation where SUMO was fused in-frame to p53 at its C-terminus (described in chapter 3). This a llows for a stabilized p53SUMO product that cannot be cleaved by SE NPs. The second method involved chemical dimerization of p53 to SUMO by rapamycin usi ng Argent™ Regulated Heterodimerization kit (Bayle et al. 2006; Dirnberger et al. 2006; Zhu et al. 2006). Here the engineered FKBP and FRB domains were fused respectively to p53 and SU MO (described in chapter 3). FKBP is a derivative of FKBP12, an abundant cytoplasmic protei n that is the initial in tracellular target for the immunosuppressive drugs FK506 and rapamycin (S iekierka et al. 1989; Balbach et al. 2000). FRB is a derivative of the rapamycin-binding domain of mTOR, a cyt oplasmic protein and a major regulator of many biological processes esse ntial for cell prolifera tion, angiogenesis, and cell metabolism. Rapamycin, an immunosuppressant forms a complex with FKBP and together they bind to FRB to form a heterodimer. This technique artificially simulates SUMOmodification of p53 and provides an alternative to the p53-SUMO fu sion construct. With these techniques, the association of SUMO to p53 coul d be efficiently manipulated, permitting us to assess the potential functional significance of p53 sumoylation.


34 SUMO-Modified p53 is Unable to Suppress Clonogenic Growth We initially assessed the potential impact of p53 sumoylation on its function using colony formation assays in order to determine if the p53-SUMO constructs could arrest cell growth like wild-type p53. We examined the SUMO m odifiers SUMO1 and SUMO3, two proteins that are known to conjugate to p53. H1299 lung cancer cells, deficient in p53, were cotransfected in duplicat e with an expression plasmid for th e puromycin-resistant gene together with either an empty vector control, wt p53, and the p53-SUMO fusions, p53-SUMO1 and p53SUMO3. A negative control, p53 -R248W-SUMO1 was also used. This p53-SUMO1 mutant has a mutation in the p53 DNA-binding domain which prevents it from binding to p53 target promoters and blocks p53Â’s ability to bring a bout cell growth arrest. Colony formation assays were carried out in puromycin-containing media se lecting for those cells cotransfected with the puromycin gene and the fusion constructs. The plat es were stained with methylene blue (Figure 4-1). The absence of colonies reflected the ability of wild-type p53 to abolish H1299 clonogenic growth (Panel B, Fi gure 4-1) as expected. Surp risingly, both p53-SUMO1 and p53 SUMO3 were largely defective to suppress cell gr owth as seen by the numerous H1299 colonies (Panels C and D, Figure 4-1) in contrast to previous reports that sumoylation of p53 might enhance cell growth arrest (Gostissa et al. 1999). As e xpected, the DNA-binding mutant p53 R248W-SUMO1 failed to inhibit cell growth (Pan el E, Figure 4-1). These results suggest that sumoylation of p53 might impair its f unction in tumor growth suppression. SUMO Modification Inhibits Tran scription of p53 Target Genes Because p53 is a transcription factor, we assessed whether SUMO modification would impact p53-mediated transcription. Different p53 target promoters (Fas, Mdm2, PIG3, and PIDD) were fused to the luciferase gene in the pGL3-Basic vector. Colorectal cancer cells


35 deficient for p53 (HCT116 p53-/-) were transiently transfected with the different p53 target promoter reporter constructs along with th e SUMO1 and SUMO3 fused p53 constructs as indicated by the legend in Figure 4-2. The tran sfected cells were harv ested 24h posttransfection and whole cell extracts were processed for lucife rase assays. Luciferase reporter gene assays (described in chapter 3) were carried out. The luciferase activity was normalized against that of the control transfected with an empty vector. Ex pression of wt p53 resulted in marked activation of p53-responsive promoters of Fas ( 8 fold), Mdm2 ( 30 fold), PIG3 ( 50 fold), and PIDD ( 110 fold). P53-SUMO1 and p53-SUMO3 were largel y defective in activa ting the p53 target promoters (Fas ( 3 fold), Mdm2 ( 5 fold), PIG3 ( 5 fold), and PIDD ( 25 fold)) (Figure 4-2). Hence, SUMO conjugation to p53 appeared to impai r the ability of p53 to activate transcription from various p53 target promoters. To substantiate the results of the above luci ferase reporter gene assays, we transduced H1299 cells with recombinant adenoviral v ectors expressing GFP, wt p53, p53-SUMO1, p53SUMO3, and the p53-R248W-SUMO1 mutant. The wt p53 and the p53-SUMO constructs were the same ones as in the colony formation assa ys. Forty-eight hours postt ransduction, cells were washed and harvested. The whole cell extracts were then analyzed for the expression of several p53 target genes using real-time PCR to determine their mRNA levels. The following genes for mRNA analysis were selected as they re present well studied p53 target genes: p21, Mdm2, and PUMA. The RNA levels were normalized to that of nontransfected cells. Real-time analysis showed markedly elevated mRNA levels for p21 ( 60 fold), PUMA ( 6 fold), and Mdm2 ( 40 fold) in H1299 cells expres sing wt p53 (Figure 4-3). H1299 cells infected with adenovirus expressing p53SUMO1 showed lower mRNA levels for p21 ( 2 fold), PUMA ( 2 fold), and Mdm2 ( 10 fold). Cells infected w ith adenovirus expressing p53-


36 SUMO3 also showed lower mRNA levels for p21 (no change), PUMA (no change), and Mdm2 ( 5 fold). As expected, cells infected with adenovirus for GFP and p53-R248W-SUMO1 controls showed little fold cha nges in mRNA levels (Figure 4-3) The results thus demonstrated that under normal conditions SUMO-fusion of p53 inhibited p53-mediated transcription. Effects of p53 Sumoylation through Inducible Heterodimerization In order to validate the resu lts utilizing the p53-SUMO fusi on constructs, a different approach was employed based on the chemical induced fusion model developed by Ariad Technologies. This strategy involved the fusion of engineered domains (1xFKBP, 2xFKBP, and FRB) to the proteins of intere st and the introduction of a chemi cal dimerizer to bring about a tight association of the two pr oteins. The C-terminus of p53 was fused to the 1xFKBP, 2xFKBP, and FRB domains (p53-1xFKBP, p53-2xFKBP, a nd p53-FRB). The C-terminus of SUMO was fused to the FRB domain (SUMO-FRB). Introducti on of rapamycin resulted in a tight association between the 2xFKBP and FRB domains in a fash ion mimicking SUMOÂ’s C-terminal conjugation to the Lys386 of p53. P53 and SUMO fusions used for heterodime rization were generated as indicated in Materials and Methods (see chapter 3). In orde r to determine if p53-1xFKBP, p53-2xFKBP, and p53-FRB fusion constructs mimicked physiological p53 in terms of thei r ability to activate transcription, we tested these constructs using lu ciferase assays in conjun ction with Western blot analysis. Immunoflourescence studies were carried out to determine if p53-2xFKBP exhibited similar subcellular lo calization as wt-p53. H1299 cells lacking endogenous p53 were tr ansiently transfected with the firefly luciferase reporter driven by the p53 target p21 promoter as well as wt-p53, p53-1xFKBP, p532xFKBP, or p53-FRB. Luciferase assays were conducted on whole cell extracts (WCEs) to


37 determine the transcriptional activity of the ar tificial p53 constructs. Luciferase activity was normalized to that of control tran sfected with an empty vector. Prot ein levels of these constructs in WCEs were examined by Western blot analys is with mouse anti-p53 (DO1). P53-2xFKBP had the highest luciferase activity. The fold induc tion by p53-FRB construct was smaller compared to wt p53, p53-2xFKBP or p53-1xFKBP, which might be attributable to the obviously lower protein levels of p53-FRB (Figure 4-4). Wtp53 and p53-1xFKBP also markedly activated the expression of the luciferase. The results indicat ed that, similar to wt p53, the artificial p53 constructs retained transactiva tion potential from the p21 promoter Therefore, the attachment of 1xFKBP, 2xFKBP, or FRB to the C-terminus of p53 did not affect the tr ansactivation potential of p53. P53-2xFKBP was arbitrarily selected for further studies. To further demonstrate the similarity of the p53-2xFKBP to wt p53 in terms of their cellular localization, i mmunoflourescence (IF) experiments were carried out on p53 deficient Saos2 cells. They strongly adhere to coverslips compared to H1299 cells, which facilitated the experimental procedure in IF. These cells were tr ansiently transfected in duplicate with vectors for either wt-p53, p53-2xFKBP, or p53-2xFKBP together with th at for the FRB domain. Cells were either untreated or trea ted with rapamycin to the final concentration of 0.1 M 6h posttransfection. It is well known that p53 is st rongly localized to the nuclear compartment (Lang, Iwakuma et al. 2004; Green and Chipuk 2006) and we wanted to determine if the p532xFKBP construct also showed similar nuclear localization. P 53-2xFKBP in free form indeed showed strong localization in th e nuclear compartment similar to wt-p53 (Figure 4-5). P532xFKBP chemically fused to FRB also showed strong nuclear localizatio n indicating that FRB had no effect on subcellular localization of p53-2xFKBP (Fig ure 4-5). We can therefore conclude from the luciferase re porter assays and the IF experi ments that fusion of p53 to the


38 2xFKBP domain as well as its subsequent hete rodimerization with FRB did not affect its transactivation activity or subcellular local ization. Hence, p53-2xFKBP behaved like wt-p53. Effects of Rapamycin-Induced p53-SUMO Heterodimerization on p53-Mediated Transcription Upon sumoylation, the transac tivation function of p53 appear ed to be impaired as demonstrated in luciferase reporter assays and real-time PCR using the p53-SUMO fusions (Figures 4-2 and 4-3). To substantiate those results using the chemical induced dimerization system, H1299 cells deficient of p53 were transiently transfected in duplicate with a luciferase reporter driven by the p21 promoter along with p53-2xFKBP and SUMO-FRB constructs (Figure 4-6). The cells were either untr eated (white bar graph) or tr eated (black bar graph) with rapamycin, to the final concentration of 0.1 M 6h posttransfection. Cells were then harvested 24h posttransfection and whol e cell extracts were processed fo r luciferase assays. Normalized luciferase activities were expressed relative to the basal promoter activity (control). Chemical fusion of either SUMO1-FRB or SUMO3-FRB with p53-2xFKBP showed significant decreases in p21 reporter activity (black bar graph) when co mpared to reporter levels of untreated samples (white bar graph). Rapamycin had no effect on tr anscription from the p21 promoter mediated by wt p53 as demonstrated by simila r luciferase levels of untreat ed and rapamycin-treated cells. Similarly, coexpression of GFP-SUMO3 and p53-2xFKBP only moderately decreased the reporter activity in the presence of rapamyci n (Figure 4-6). GFP-SUMO3 was not expected to interact with p53-2xFKBP due to the absence of the FRB domain. Hence, tethering SUMO moiety to p53 through rapamycin-induced he terodimerization of p53-2xFKBP and SUMO-FRB inhibits p53-mediated transcription.


39 SUMO Modification Promotes Nuclear Export of p53 P53 is a transcription factor and present in the nucleus. Under normal conditions, p53 levels are kept low as it is believed that p53 is exported to the cytoplas m where it is degraded (Lane 1992; Leveillard, Gorry et al. 1998). Our previous results demonstrated that SUMO negatively regulated p53-mediated transcription. To determine if sumoylation of p53 affects its subcellular localization, the localization of p53-SUMO fusi ons (p53-SUMO1 and 3) was examined through IFs. Saos2 cells were tr ansiently transfected with p53-SUMO1 and p53SUMO3 fusion constructs. Twenty-f our hours posttransfection, cells were fixed and stained with mouse anti-p53 (DO1) antibody followed with s econdary antibody anti-mous e IgG conjugated to rhodamine. Wt p53 was predominantly seen in the nucleus (Figure 4-5A). However, the covalent p53-SUMO1 and p53-SUMO3 fusion constructs were s een not only in the nucle us but also in the cytoplasm (Figure 4-7A and 4-7B respectively). Quantitative analysis demonstrated that 50-60% of transfected cells showed the presence of p53SUMO fusions in the cytoplasm compared to only 10% of transfected Saos2 cells that s howed cytoplasmic wt p53 (Figure 4-7C). To further investigate if SUMO modification may have a role in regulating p53 localization, Saos2 cells we re transfected in duplicate with p53-2xFKBP and SUMO-FRB constructs as indicated in the legend to Figur e 4-8. Six hours posttransfect ion, cells were treated with rapamycin to the final concentration of 0.1 M. Twenty-four hours posttransfect ion, cells were fixed and probed with mouse anti-p53 (DO1 ) and rabbit anti-HA an tibody to detect SUMOFRB that was HA-tagged. Secondary antibodies used were anti-rabbit IgG conjugated to FITC or anti-mouse IgG conjugated to rhodamine. Subce llular distribution of p53 was examined using immunofluorescent microscopy. In the absence of rapamycin, p53-2xFKBP was mostly localized in the nucleus, regardless of the coexpressed constructs (SUMO1-FRB, SUMO3-FRB or GFPSUMO3 G92A, panels A, B, C of Figure 4-8). Strikingly, in the presence of rapamycin, p53-


40 2xFKBP was localized predominantly in the cytopl asm, when either SUMO1-FRB (panel D) or SUMO3-FRB (panel E) was coexpressed. As expected, coexpression of GFP-SUMO3 G92A construct, which lacked the FRB domain, did not affect nuclear lo calization of p53-2xFKBP despite the presence of rapamycin (panel F). Thes e observations demonstrated that attachment of SUMO to p53 through rapamycin-mediated hete rodimerization resulted in cytoplasmic localization of p53, in agreemen t with the notion that SUMO modification of p53 promotes nuclear export of p53. Quantitative analysis of transfected cells (Fi gures 4-8G-H) further demonstrated that SUMO chemically fused to p53 was frequently present in the cytoplasm (black bar). Wt p53 and p53-2xFKBP in the absen ce of rapamycin showed strong localization in the nucleus (Figure 4-5). Mdm2 might be Involved in Promoting Cytoplasmic Localization of p53 Previous studies have shown that the major p53 regulator, Mdm2, has a critical role in regulating subcellular lo calization of p53. Here, Mdm2 binds to p53Â’s transactivation domain and posttranstionally modifies p53 with ubiquitin or ubi quitin-like modifiers (SUMO) as an E3 ligase, and the cysteine residue at position 438 is critical for the ligase activity (Di Ventura, Funaya et al. 2008; Jackson and Berberich 2000 ). We investigated whether Mdm2 also influences SUMO-mediated nuclear export of p53 as mentioned above. Mouse embryonic fibroblasts (MEF) with double knoc kouts of both p53 and Mdm2 (MEFp53/Mdm2 =/=) were transiently transfected with p53-2xFKBP together with either SUMO1-FRB, or SUMO3-FRB. P53-2xFKBP was Myc-tagged and the SU MO-FRBs were HA-tagged. Six hours posttransfection, cells were treated with rapamycin to the fi nal concentration of 0.1 M. Twentyfour hours posttransfection, cells were fixed a nd treated with anti-Myc and anti-HA antibody. The secondary antibodies were anti-rabbit IgG c onjugated to FITC or an ti-mouse IgG conjugated to rhodamine. Immunoflourescent microscopy expe riments demonstrated that both in the


41 absence and presence of rapamycin, p53-2xFKBP (red) was localized to the nucleus (results not shown). MEFp53/Mdm2=/= was again transfected with p53-2xF KBP together with either SUMO1FRB, or SUMO3-FRB along with either wild-typ e Mdm2 or the Mdm2 (C438A) mutant lacking E3 ligase activity. In the absence of rapamyci n, p53-2xFKBP (red) was str ongly localized in the nucleus as expected (Figure 4-9A and E). U pon addition of rapamycin, p53-2xFKBP (red) was present in both the cytoplasm and the nucleus when Mdm2 was also expressed, in 80% of transfected cells (Figure 4-9B & F). In contrast, when Mdm2 C438A mutant was coexpressed, p53-2xFKBP (red), even with the coexpression of SUMO1-FRB or SUMO3-FRB (green) and the presence of rapamycin, remained in the nucle ar compartment (Figure 4-9D and H). This was seen for 90% of transfected cells. Quantitative an alysis of transfected MEFs shown in Figures 49I-J further demonstrated th at p53-2xFKBP chemically fuse d to SUMO-FRB was frequently present in the cytoplasm when wt-Mdm2 but not with mutant Mdm2 was coexpressed. These results suggest that Mdm2 and its E3 ubiquitin liga se function appeared to play an important role in regulating SUMO-mediated cy toplasmic localization of p53. Involvement of the SUMO Po cket that Binds to SIM Structural study revealed that a specific pocket on the SUMO protein is responsible for interacting with previously char acterized SUMO-interac ting motif (aka SIM, also see Chapter 5 below). The SUMO pocket has been implicated in protein-protein interactions. The pocket is hydrophobic in nature (Baba et al 2005; Hecker et al, 2006) and consists of amino acids such as phenylalanine, valine, and leucin e. SUMO1 has a phenylalanine at position 36, which is critical for interacting with SIM. Likewise, the corr esponding F31 in SUMO3 is also essential for binding to SIM. In order to determine if the SUMO pocket is important for the nuclear export of p53, we mutated F31 of SUMO3 and F36 of SUMO1 to alanine. The mutant constructs were then fused


42 to the FRB construct (see chapter 3) and tran sfected into Saos2 along with p53-2xFKBP and the transfected cells were either treated or untreated with rapamycin as demonstrated in the legend to Figure 4-8. The transfected cells were then examined using immunoflourescent microscopy. As shown in Figure 4-10A-B, the addition of rapamy cin to cells with coexpression of p53-2xFKBP (red) and SUMO3 F31A-FRB (green) failed to induce nuclear export of p53. The same phenomenon was observed with SUMO1 F36A-FRB (F igure 4-10 C-D). The results tabulated in Figure 4-10E showed that p53 exhibited nuclear localization in 95% of the transfected cells despite the presence of SUMO3 F31A-FRB and rapamycin. On the contrary, SUMO1-FRB and SUMO3-FRB when chemically fused to p53-2xFKBP in the presence of rapamycin were observably shifted to the cytopl asm (Figure 4-8G-H). The results thus demonstrated the SIMbinding pocket of SUMO is important for SUMO-m ediated nuclear export of p53, possibly via a SIM of an undetermined protein invo lved in export (Figure 4-10F). P53 is Exported as a Tetramer Current paradigm regarding nuclear export of p53 holds that in order for p53 to be exported, the NES of p53 must be exposed. Th e export protein, CRM1, interacts with the p53 NES and forms a complex along with another pr otein RanGTP (Petosa, Schoehn et al. 2004). This complex is then shuttled to the nuclear pore complex on th e nuclear membrane where p53 is released to the cytoplasm. Since the NES is embedded in the oligomerization domain of p53 (4D), disruption of the p53 tetramer is requi red to expose the NES. We found above that tethering SUMO to p53 results in nuclear export of p53. If the paradigm still holds true, we would expect the disruption of p53 tetramer upon rapamycin-induced heterodimerization. To test this, we carried out immunopr ecipitation (IP) experiments. FLAG-p53 (generated by fusing FLAG to p53Â’s N-terminus) was co-expresse d with p53-2xFKBP (76-kDa) and SUMO3-FRB constructs. In the parallel control experi ment, FLAG-p53 was coexpressed with GFP-p53 (80-


43 kDa)(generated by fusing GFP to p53Â’s Nterminus) and SUMO3-FRB constructs. The transfected cells were either untreated or tr eated with rapamycin 6h posttransfection. Twentyfour hours posttransfection, the ce lls were then processed for IP using anti-FLAG M2 antibody conjugated to agarose beads. Western blot an alysis was done using anti-p53 (DO1) antibody. As shown in Figure 4-11A, both p53-2xFKBP (Panel 1, IP lane) construct and GFP-p53 (Panel 2, IP lane) was coprecipitate d with FLAG-p53 in the absence of rapamycin. This suggested that FLAG-p53 could oligomerize with both p53-2xFKBP and GFP-p53 as expected. In the presence of rapamycin, GFP-p53 was coprecipitated with FLAG-p53 (the 80-kDa band) (Figure 4-11B Panel 4, IP lane). This was expected as rapamy cin should have no influence on their association. However, in the presence of rapamycin wh en p53-2xFKBP was chemically fused to SUMO3FRB, FLAG-p53 still pulled down the 76-kD a p53-2xFKBP band (Figure 4-11B Panel 3, IP lane). These results suggested that rapamyci n-mediated attachment of SUMO3-FRB to p532xFKBP did not result in the disso ciation of the p53 tetramer, as FLAG-p53 could still form a complex with p53-2xFKBP that is at tached to SUMO3-FRB by rapamycin. These results nonetheless did not test whet her SUMO3-FRB-HA was indeed attached to the p53-2xFKBP construct in the presence of rapa mycin. To evaluate this, we conducted similar IP experiments as shown in Figure 4-12. Saos2 cells were transfected with FLAG-SUMO3-FRB (FLAG at the N-terminus of SUMO3-FRB in pcDNA vector), p53-2xFKBP, and HA-p53. Western blot analysis was carried out using anti-FLAG and anti-p53 (FL393) antibodies. As shown in Figure 4-12, in the absence of rapamy cin, little p53-2xFKBP (the 76-kDa band) and HA-p53 (the 53-kDa band) were precipitated with FLAG-SUMO3FRB (lane 1, IP Panel). In contrast, much more p53-2xFKBP and HA-p53 were coprecipitated in the presence of rapamycin (lane 3, IP Panel). In separate IPs, we al so used SUMO3-FRB-FLAG (with FLAG at the C-


44 terminus of SUMO3-FRB in pcDNA), to re place FLAG-SUMO3-FRB. Although the expression levels of p53-2xFKBP and HA-p53 appeared to be lower when coexpressed with SUMO3-FRBFLAG (lanes 2 and 4 in the Input panel), it is cl ear that the presence of rapamycin resulted in more coprecipitation of both HAp53 and p53-2xFKBP when compared to the samples that were untreated with rapamycin (compare lane 4 to lane 2, IP panel). Collectively, the IP results confirm that p53 could still form the complex when SUMO is attached to the C-terminus of p53. There are three key hydrophobic residues of the putative p53 NES, L344, L348 and L350, which mediate the formation of the p53 tetramer (Lee et al ., 1994; Clore et al ., 1995; Jeffrey et al ., 1995; Waterman et al ., 1995; Mateu and Fersht, 1998). This suggests that mutation of these residues would lead to an inhibition of the formation of the p53 tetramer. We used yeast two-hybrid assays to test the interaction betw een CRM1 and p53 mutants with mutations in the oligomerization domain. Yeast was cotransformed with constructs fused to the Gal4-AD and Gal4-BD domains as indicated in the lege nd to Figure 4-13. The following p53 mutants inhibiting p53 tetramer formation were used: p5 3 with leucines subst ituted at aa348 by alanine and at aa350 by proline (p53 L348A/350P) or aa344 by alanine (p53 L344A). Wt-p53 or p53 mutants unable to form p53 tetramer (p53 L348A /350P and p53 L344A) were fused to Gal4-AD. A fragment of the p53 coactivator, p300, a nd the export protein, CRM1, fragment spanning amino acid positions 571-1071 (CRM1 571-end), both of which are known to interact with p53 were fused to Gal4-BD as described in chapter 3. Petosa et al demonstrated that the structure around Phe572 in CRM1 is an important region fo r cargo protein binding whilst the N-terminal domain was needed for RanGTP binding. As exp ected, yeast two-hybrid a ssays revealed that p53 could interact with its coac tivator p300 and the CRM1 region spanning amino acid residues 5711071 as indicated with the growth of yeast col onies (Figure 4-13 sector s 1 and 2 respectively).


45 Interestingly, we observed that p53 could not in teract with the full-le ngth CRM1 (results not shown). This could be explained by the short half-life of the p53-CRM1-RanGTP complex and the dynamic export system. Mutations preventing th e formation of the p53 tetramer abolished the interaction of p53 with CRM1 (571-1071) fragment as seen by the absence of yeast colonies (Figure 4-13 sectors 3 and 4). The data thus corr oborated the previous IP studies suggesting that p53 might be exported as a tetramer. CRM1 has a Possible SIM An amino acid sequence analysis revealed a possible SUMO-interac ting motif (SIM) in the nuclear export protein, CRM1, with a hydrophobic core at the amino acid positions 429-433. The core is depicted as Leu-ValIle-Ile (LVII). SIM is important in its interaction with a SUMO pocket as mentioned above. We have also showed that if the SUMO pocket is mutated, p53 is localized to the nucleus (Figure 4-10). By cont rast, wt-SUMO when tethered to p53 resulted in its export to the cytoplasm (F igure 4-8). To demonstrate the importance of the SIM of CRM1, a yeast two-hybrid assay was performed. P53, SU MO1, SUMO1 pocket mutant (SUMO1 F36A), SUMO3, p53-SUMO1, and p53-SUMO3 were fused to Gal4-AD. Mdm2 and full-length CRM1 were fused to Gal4-BD. The plasmids expressi ng the p53-Gal4-AD fusions were cotransformed with that expressing the Gal4 -BD-CRM1 hybrids in yeast. As a positive control, p53 could interact with its regulator Mdm2 as indicated by the growth of yeast col onies (data not shown). Interestingly, small yeast colonies were seen (d ata not shown) when the yeast were transformed with vectors for CRM1 and SUMO1. We also found that SUMO3 did not interact with CRM1 (data not shown). As was antici pated, SUMO1 with the pocket mutation (SUMO1 F36A) did not interact with CRM1 (data not shown). These re sults suggest that the SIM of CRM1 can be a functional SIM due to its inte raction with SUMO1 via the SUMO-pocket. Furthermore, p53SUMO1 and 3 fusion constructs could not intera ct with full-length CRM1 (data not shown).


46 These results point to the possibi lity of another SUMO-modified protein with a SIM that could interact with both SUMO-modified p53 a nd at the same time the SIM of CRM1. To further investigate the potential inte raction of CRM1Â’s SIM with SUMO-fused p53, FLAG IPs were carried out followed by Western blot analysis using anti-FLAG and anti-GFP antibodies. Saos2 cells were tran sfected with DNA constructs as indicated in the legend to figure 4-14. Here, GFP-fusion of the full-length CRM 1 or the CRM1 mutant, where the SIM was mutated by converting the valine at the position 430 to lysine (V430K), was used along with FLAG-p53-SUMO1 and FLAG-p53-SUMO1 F36A mutant where the SUMO pocket was mutated (F36A). As seen in Figure 4-14, FL AG-p53-SUMO1 (~63 kDa) could moderately coprecipitate with the mutated CRM1 (150 kDa), whereas a smaller amount of wt CRM1 was precipitated (compare lanes 1 and 2). Interest ingly, FLAG-p53-SUMO1 F36A with the mutated SUMO pocket seemed to have higher affinity to wt GFP-CRM1 (lane 3). An even stronger interaction of the mutated CRM1 with the p53-SU MO1 F36A construct was seen (lane 4). These data suggest that the putative SIM of CRM1 and the SUMO pocket of the SUMO-modified p53 might impede the CRM1-p53 interaction. The resu lts therefore point to the importance of a functional SIM and SUMO pocket in regulating the nuclear export of p53, possibly through the efficient release of p53 from the CRM1 export system. We conducted additional IP experiments as shown in figure 4-15. Saos2 cells were transfected again with vector s for either GFP-CRM1 or th e mutant GFP-CRM1 V430K along with that for either FLAG-p53 and mutant FLAG-p53 where the SUMO-modification site at lysine386 was altered to arginine (K386R ). Western blot analysis in figure 4-15 showed that FLAG-tagged wt-p53 (53 kDa) showed a moderate interaction with the mutant CRM1 (150 kDa) (lane 2) compared to the wt-CRM1 (lane 1). Interestingly, more wt CRM1 or the CRM1 V430K


47 mutant was coprecipitated with the p53 K386R mutant (compare lanes 3 and 4 with 1 in Figure 4-15). These results implied that the absence of the SUMO modificat ion site of p53 might facilitate its interaction w ith CRM1. Conversely disruption of the functional SIM of CRM1 appeared to enhance p53-CRM1 interaction. These observations are consistent with a model in which sumoylation of p53 weakens its affinity to the CRM1 exporter, resulting in the efficient release of p53 to the cytoplas m. Additionally, an unknown SUMO -modified protein is likely involved in mediating CRM1-p53 in teraction for facilitated expor t of p53 (see chapter 6). Potential Roles of SENP in Nuclear Export of p53 Since SUMO-modification has a role in the export of p53 into the cytoplasm, it was hypothesized that the SUMO proteases (SENPs) may regulate the release of SUMO-modified p53 at the nuclear pore complex where CRM1 is known to release its cargoes. The Ulp1p in yeast has been shown to be involved in nuclear -cytoplasmic trafficking of proteins. The Ulp1p human homologs (SENP1 and SENP 2) both have a NES like Ulp1p. To demonstrate the importance of the SENP s in the nuclear export of p53, Saos2 cells were cotransfected with vectors for p53-2xFKB P and SUMO3-FRB along with that for different SENPs that were GFP-tagged at the N-terminus of SENPs. The transfecte d cells were examined by IFs as indicated in the le gend to Figure 4-16. In the pr esence of GFP-SENP2, p53-2xFKBP could still be exported into the cytoplasm in around 50% of transfected cells, upon rapamycininduced heterodimerization with SUMO-3-FRB (F igure 4-16A-B and K). By contrast, in the presence of rapamycin, SUMO-3-FRB and ove rexpression of GFP-SENPs 3, 5, 6 and 7, p532xFKBP strongly localized in the nucleus in at least 80% of tr ansfected cells (Figure 4-16C-H and K). Similar results were observed in MCF7 that has wt endogenous p53 (results not shown). There are two possibilities based on these obser vations: (1) The Ulp1p family of enzymes (e.g. SENP1 and SENP2) may be requi red for the release of p53 at the nuclear pore complex (NPC)


48 given their close proximity to the NPC; and (2 ) the blockade of the nuclear export of p53 by SENPs 3, 5, 6 and 7 in our experimental system suggests that these proteases might inhibit nuclear export through their action on other unkn own SUMO-modified proteins involved in the export process, rather than on SUMO-modified p53, as they should not affect the rapamycinmediated heterodimerization of p53-2xFKBP and SUMO-FRB. Further studies are required to test these possibilities. Figure 4-1. P53-SUMO constructs are unable to suppress clonoge nic growth. H1299 cells were transfected in duplicate with (A ) an empty vector or that for the expression of (B) wt p53, (C) p53-SUMO1, (D) p53-SUMO3, or (E) p53 R248W-SUMO1, along with a plasmid for the expression of puromycin-res istant gene. Cell growth was selected in medium containing 2 g/ml puromycin for two weeks. H1299 colonies were then stained with methylene blue. P53-SUMO was unable to suppress clonogenic growth (Panels C and D). A B D C E


49 Figure 4-2. Luciferase reporter assays on different p53 target gene promoters. HCT116 p53 -/cells were transiently transfected with wt-p53 and p53-SUMO constructs along with the luciferase gene fused to different p53 target promoter s of: (A) Fas, (B) Mdm2, (C) PIG3, and (D) PIDD. Luciferase assays we re carried out on whol e cell extracts and the results were normalized to the cont rol in each assay. Each histogram bar represented the mean of two different i ndependent transfection duplicates. Standard deviations are indicated. The p53-SUMO fu sion constructs exhibited markedly reduced ability to activate th e p53 responsive promoters. C B A D


50 Figure 4-3. Real-time analysis examining the effects of p53-SUMO c onstructs on the mRNA levels of p53 targets. H1299 cells were either not infected (control) or infected with an adenoviral vector for GFP, wt p53, p53-SUMO-1, p53-R248W-SUMO-1, or p53SUMO-3. Cells were harvested 48 hours afte r infection for real-time analysis using primers specific to the following p53 targ ets: (A) p21; (B) PUMA; and (C) Mdm2. RNA levels were normalized to the cont rol RNA levels of each expression study. Each histogram bar represented the mean of two different independent transduction duplicates. Standard deviations are indicated. A B C


51 Figure 4-4. Verification of p53 cons tructs for chemical induced fusion. (A) Luciferase reporter assays on whole cell extracts were carried a nd revealed the transactivation potential of the artificial p53 constructs in comparison to wt-p53 on the p21 promoter. Luciferase levels were normalized to c ontrol. Each histogram bar represented the mean of two independent tr ansfection duplicates. Standard deviations are indicated. (B) The protein levels of p53 and the p53 fusi on constructs in tran sfected cells were revealed by Western blotting analys is using the anti-p53 (DO1) antibody. A B p53-2xFKBP (76 kDa) p53-1xFKBP or p53-FRB (63 kDa) WT-p53 (53 kDa)


52 Figure 4-5. Subcellular localizat ion of wt p53 and the p53-2xFK BP fusion. Saos2 cells were transiently transfected with different p53 c onstructs and the FRB control: (A, D) wtp53, (B, E) p53-2xFKBP and, (C, F) p532xFKBP and FRB control. Six hours posttransfection, cells were tr eated with rapamycin to the final concentration of 0.1 M (panels D, E, and F). Twenty-four hours posttransfection, IFs we re carried out as indicated in Methods and Ma terials. P53 or p53-2xFKBP (re d) and FRB (green) were detected with anti-p53 DO1 or anti-HA antibody, respectively. The secondary antibodies were anti-rabbit IgG conjugated to FITC or anti-mouse IgG conjugated to rhodamine. Cells untreated or treated with rapamycin showed similar p53 distribution in the nucleus.


53 Figure 4-6. SUMO modification of p53 inhibite d p53-mediated transactivation of the p21 promoter. (A) Model of heterodimerizati on of p53 with SUMO in the presence of rapamycin. (B) Schematic of different p53 fusion constructs paired with SUMO fusion constructs. (C) p21 luci ferase reporter assays. The Lu ciferase gene fused to the p21 promoter was transfected along with p53-2xFKBP (p53-2x) and either GFPSUMO3 control, SUMO1-FRB, or SUMO3-FRB. Cells were either untreated (white bars) or treated with rapamycin (black ba rs) six hours posttransfection. Twenty-four hours posttransfection, cells were harveste d. Whole cell extracts were examined for luciferase activity and normalized to basa l control. Each histogram bar represented the mean of two different independent tran sfection duplicates. Sta ndard deviations are indicated. Chemical fusion of p53-2xFKBP to SUMO resulted in reduced luciferase activity compared to untreated samples.


54 Figure 4-7. SUMO modification of p53 affected p53’ s subcellular localization. Saos2 cells were transiently transfected with different p53-SUMO fusions: (A ) p53-SUMO1, and (B) p53-SUMO3. P53-SUMO (red) was detected with anti-p53 DO1. The secondary antibody was anti-mouse IgG conjugated to rhodamine. (C) Quantification of p53SUMO localization. Two hundred transfected cells were counted for each assay. Percentage of cells with “nuclear only ” p53 or “nuclear + cytoplasmic” p53 were determined by dividing the number of cells exhibiting either p53 cellular localization by 200. Each histogram bar represented the mean of three different independent transfections. Standard deviations are indicated. P53-SUMO1 and p53-SUMO3 were found in the cytoplasm (denoted by arrows) when compared to wt p53.


55 Figure 4-8. SUMO modification of p53 influenced its subcellular localization. Saos2 cells were transiently transfected with different p53 and SUMO constructs: (A, D) p53-2xFKBP and SUMO-1-FRB, (B, E) p53-2xFKBP a nd SUMO3-FRB, and (C, F) p53-2xFKBP and GFP-SUMO3 G92A. Six hours posttran sfection, cells were treated with rapamycin to the final concentration of 0.1 M (panels D, E, and F). Twenty-four hours posttransfection, IFs were carried out as indicated in Materials and Methods. P53-2xFKBP (red) and SUMO1FRB or SUMO3-FRB (green) were detected with anti-p53 DO1 and anti-HA antibodies respec tively. The secondary antibodies were anti-rabbit IgG conjugated to FITC and anti-m ouse IgG conjugated to rhodamine. In the absence of rapamycin (A, B, and C) p53-2xFKBP was localized to the nuclear compartment. In the presence of rapa mycin, p53-2xFKBP was predominant in the cytoplasm (D and E) (denoted by arrows ). The GFP-SUMO3 G92A control had no effect on p53 localization (F). (G-H) Quantification of p53 subcellular localization in (G) absence and (H) presence of rapamycin wa s determined as indicated in the legend of Figure 4-7. One hundred transfected cells were counted for each assay. Percentage of cells with “nuclear only ” p53 or “nuclear + cytoplasmic” p53 were determined by dividing the number of cells exhibiting either p53 cellu lar localization by 100. Each histogram bar represent the mean of tw o independent transfections. Standard deviations are indicated.


56 Figure 4-8. SUMO modification of p53 influenced its subcellu lar localizati on (continued). G H


57 Figure 4-9. Mdm2 is a potential mediator in cellular localization of SUMO-modified p53. (A) Mouse embryonic fibroblasts deficient in p53 and Mdm2 (MEF p53/Mdm2=/=) were transiently transfected with p53-2xFKBP along with different SUMO constructs together with either wt Mdm2 or the C438A mutant lacking E3 ubiquitin ligase activity: (A, B) expression of p53-2xF KBP, SUMO-1-FRB, and Mdm2, (C, D) expression of p53-2xFKBP, SUMO-1-FRB, and Mdm2 C438A mutant, (E, F) p532xFKBP, SUMO-3-FRB, and Mdm2, and (G, H) expression of p53-2xFKBP, SUMO-3-FRB, and Mdm2 C438A mutant. Si x hours posttransfecti on, B, D, F and H were treated with rapamycin to the final concentration of 0.1 M. Twenty-four hours posttransfection, IFs were carried out as i ndicated in Materials and Methods. P532xFKBP (red) and SUMO1-FRB or SUMO3-FR B (green) were detected with mouse anti-Myc or rabbit anti-HA antibody, respec tively. The secondary antibodies were anti-rabbit IgG conjugated to FITC or anti-mouse IgG conjugated to rhodamine. (I-J) Quantification analysis was generated. Fift y transfected cells were counted for each assay. Percentage of cells with “nuclear only” p53 or “nuclear + cytoplasmic” p53 were determined by dividing the number of cells exhibiting either p53 cellular localization by 50. Each histogram bar re presented the mean of two different independent transfections. Standard devi ations are indicated. In absence of rapamycin, p53 was localized to the nucleus (A C, E, G). In the event of chemical induced fusion and the presence of func tional Mdm2, p53 was present in both the cytoplasm and nucleus (B and F). However, in the presence of the dimerizer and the Mdm2 mutant, p53 was strongly localiz ed to the nucleus (D and H).


58 Figure 4-9. Mdm2 is a potential mediator in cellular localization of SUMO-modified p53 (continued).


59 Figure 4-10. The SIM-interacting pocket of SUMO is important for nuclear export of p53. Saos2 cells were transiently transfected with p53-2xFKBP along with either SUMO3 F31A FRB (A, B) or SUMO1 F36A FRB (C, D) Six hours posttransfection, cells were treated with rapamycin to the final con centration 0.1 M (Figures B and D). Antibody staining and microscopy were done as in Fi gure 4-8. In the absence of rapamycin, p53 was localized to the nuclear compartment (A and C). In the presence of rapamycin, the same phenomenon was observed (B and D) (E) Quantitative representation of the phenomena seen in A-B was generated. One hundred transfected cells were counted for each assay. Percentage of cells w ith “nuclear only” p53 or “nuclear + cytoplasmic” p53 were determined by dividi ng the number of cells exhibiting either p53 cellular localization by 100. Each histogram bar repres ented the mean of three different independent transfec tions. Standard deviations are indicated. (F) Cartoon of the SUMO pocket interacti on with a SIM of an unknow n protein (Protein X).


60 Figure 4-10. The SIM-interacti ng pocket of SUMO is import ant for nuclear export of p53 (continued).


61 Figure 4-11. Effects of rapamycin-mediated attachment of SUMO3-FRB to p53-2xFKBP on its hetero-oligomerization with FLAG-p53. Saos2 cells were transfected with relevant DNA constructs as follows: (1 and 3) p53-2xFKBP + SUMO3-FRB + FLAG-p53, (2 and 4) GFP-p53 + SUMO3-FRB + FLAG-p53. E xperiments 3 and 4 were done in the presence of rapamycin to the final concentration 0.1 M. Twenty-four hours posttransfection, FLAG IPs were carried out and p53 was detected in Western blotting analysis using mouse anti-p53 (DO1 ). In the absence of chemically induced SUMO fusion, all p53 constructs were pulled down in the IP signifying that all p53 constructs could interact with each othe r. In the presence of rapamycin, FLAG-p53 could still pull down p53-2xFKBP (76kDa ba nd). As expected, FLAG-p53 could pull down GFP-p53 (80kDa band) in th e presence of rapamycin. A B


62 Figure 4-12. SUMO-modification of p53 does not di srupt p53 oligomerization. Saos2 cells were transfected with relevant DNA expression plasmids as follows: lanes 1 and 3, p532xFKBP + FLAG-SUMO3-FRB + HA-p53, la nes 2 and 4, p53-2xFKBP + SUMO3FRB-FLAG + HA-p53. Cells were either untreated (lanes 1 and 2) or treated (lanes 3 and 4) with rapamycin to the final concentration 0.1 M. Twenty-four hours posttransfection, IP was done using anti-FLAG M2 agarose beads and the precipitated materials were analyzed by Western blot analysis usi ng rabbit anti-p53 (FL393) antibody. Inputs were probed with rabbit anti-p53 (FL 393) and rabbit anti-FLAG antibody. In the presence of rapamycin, ch emical fused p53-SUMO could interact strongly with HA-p53 (lanes 3 and 4) when compared to untreated samples (lanes 1 and 2).


63 Figure 4-13. P53 tetramer is required for inte raction with the CRM1 export protein. Yeast was transformed as indicated in the Material s and Methods section. P53 has been well documented to interact with the transcrip tion co-activator, p300 (Sector 1 as positive control). P53 interacted with the CRM1 fragment spanning residues 571 to 1071 (Sector 2). Disruption of the p53 tetr amer (L348/L350P and L344A mutations) abolished the interaction between p53 and CRM1 (Sectors 3 and 4).


64 Figure 4-14. Interactions between CRM1 and p53SUMO fusion constructs. Saos2 cells were transfected with DNA constructs as follows: (1) GFP-CRM1 + FLAG-p53-SUMO1, (2) GFP-CRM1 V430K + FLAG-p53-SU MO1, (3) GFP-CRM1 + FLAG-p53SUMO1 F36A, (4) GFP-CRM1 V430K + FLAG-p53-SUMO1 F36A. FLAG IPs were carried out followed by Western blot analysis. IP was probed with mouse antiGFP antibody. Inputs were probed with rabbit-anti-FLAG and mouse anti-GFP antibodies respectively.


65 Figure 4-15. The putative SIM of CRM1 and the K386 sumoylation site of p53 influence CRM1p53 interaction. Saos2 cells we re transfected with DNA cons tructs as followed: (1) GFP-CRM1 + FLAG-p53, (2) GFP-CRM1 V430K + FLAG-p53 K386R, (3) GFPCRM1 + FLAG-p53, (4) GFP-CRM1 V430K + FLAG-p53 K386R. FLAG IPs were carried out followed by Wester n blot analysis. IPs were probed with mouse anti-GFP antibody. Inputs were probed with rabbit anti-FLAG and mouse anti-GFP antibodies respectively. Results demonstrated the importance of the p53Â’s SUMO-modification K386 site and the CRM1 SIM as indicated by strong pull downs of CRM1 (lane 2-4) compared to lane 1.


66 Figure 4-16. Potential regulation of the nucle ar export of p53 by SENPs. Saos2 cells were transfected with DNA constructs in duplicat e as indicated: (A-B) GFP-SENP2 + p532xFKBP + SUMO3-FRB, (C-D) GFP-SENP 3 + p53-2xFKBP + SUMO3-FRB, (E-F) GFP-SENP5 + p53-2xFKBP + SUMO3-F RB, and (G-H) GFP-SENP6 + p532xFKBP + SUMO3-FRB, and (I-J) GFP-SE NP7 + p53-2xFKBP + SUMO3-FRB. Six hours posttransfection, with rapamycin to the final concentration 0.1 M. IFs were carried as indicated in Methods and Mate rials. P53-2xFKBP (red) was probed with anti-p53 DO1 antibody. The secondary anti body was anti-mouse IgG conjugated to rhodamine. Overexpression of SENP2 still permitted chemically fused p53-SUMO to be localized to the cytoplasm (denoted by arrows). Overexpression of SENPs 3-7 blocked nuclear export of p53-2xFKBP in th e presence of rapamycin. (K) Graphical representation tabulatin g localization of p53-2xF KBP with the expression of different SENPs was generated. One hundred transfected cells were counted for each assay. Percentage of cells with “nuclear only ” p53 or “nuclear + cytoplasmic” p53 were determined by dividing the number of cells exhibiting either p53 cellular localization by 100. Each histogram bar repres ented only one transfection.


67 Figure 4-16. Potential regulation of the nuclear export of p53 by SENPs (continued).


68 CHAPTER 5 IDENTIFICATION OF TWO INDEPENDENT SUMO-INTERACTING MOTIFS IN DAXX: EVOLUTIONARY CONSERVATION FROM DR OSOPHILA TO HUMANS AND THEIR BIOCHEMICAL FUNCTIONS Daxx is an essential protein for embryonic de velopment and has been shown to function in diverse cellular pathways including apoptosis and transcription. How Daxx is involved in these pathways at the molecular levels remains elusive. We have analyzed the structure and functions of the Daxx protein from an evoluti onary perspective and using various molecular biological approaches. We found th at the ability of Daxx to bi nd SUMO seems critical for its diverse cellular functions. Daxx Possesses Two Independent SUMO -Interacting Motifs (SIMs) Daxx is recruited to promyelocytic le ukemia protein (PML) nuclear bodies and sumoylation of PML was shown to be required for this recruitment (Ishov, Sotnikov et al. 1999). Therefore, Daxx might have SUMO-binding property. To examine whether Daxx has the recently characterized SIM from a number of kno wn SUMO-binding proteins (Minty, Dumont et al. 2000; Song, Durrin et al. 2004; Hannich, Lewi s et al. 2005), we inspected the primary sequence of Daxx protein and found two stretches of sequences bear striking similarity to SIM. One stretch is near the N-terminus and the other near the C-termi nus. Fig. 5-1A shows a sequence alignment of the two Daxx SIMs agains t other known SIMs. Like other SIMs, the two Daxx motifs have the conserve d hydrophobic core consisting of four consecutive Leu/Ile residues, and interestingly, the two motifs in Daxx have exactly the same sequence for the hydrophobic core. Two recent structural studies in dicate that the hydrophobic core confers the primary affinity and specificity for SUMO-bi nding and that these residues insert into a hydrophobic groove on a conserved surface of the SUMO1 structure (Song, Durrin et al. 2004; Song, Zhang et al. 2005). Another important featur e of SIM is a Glu/Asprich sequence that lies


69 at either N-terminal or C-terminal to the hydroph obic core (see Fig. 5-1A), and structural studies revealed that the negatively char ged acidic residues interact el ectrostatically with positively charged Lys residues on the SIM-binding pocke t of SUMO (Song, Zhang et al. 2005; Hecker, Rabiller et al. 2006). Some of the SIMs also have Ser/Thr residues embedded in the acidic residues, and in this context Ser/Thr residues might be within the casein kinase 2 (CK2) phosphorylation sites. Presumably, the interactio n between SUMO and SIMs might be regulated by phosphorylation. To test whether the two SIMs in Daxx ar e capable of SUMO-binding, we performed yeast two-hybrid assays as indi cated in Figure 5-1B with the Gal4-BD fused to different Daxx constructs along with SUMO1 fused to Gal4AD. As shown in Fig. 5-1B, Daxx can bind to SUMO1 and mutation in the conserved hydrophobic co re of either SIM in full-length Daxx (I7K or I733K) did not affect SUMO-binding, but si multaneous mutation of both motifs abolished Daxx-SUMO1 interaction. Likewise, Daxx fragment carrying either the N or C-terminal SIM can still bind to SUMO1 and I733K mutation of the Daxx C-termin al fragment (aa 573-740, sector 7 in Fig. 5-1B) disabled its SUMO-binding abilit y. GST pull-down assays where GST was fused to either the Daxx N-terminus or C-terminus SIM fragments demonstrated strong binding between the N or C-terminal fragment of Daxx and SUMO1 using purified proteins from E. coli whereas GST alone had no affinity to SUMO1 (Fig. 5-1C). Interaction between Daxx and SUMO Paralogs There are four paralogs of SUMO in the humans. SUMO1-3 are bona fide modifiers that can be conjugated to target proteins, whereas SUMO4 may not be attach ed to other proteins (Mukhopadhyay and Dasso 2007). We tested the ability of these SUMO paralogs to interact with Daxx. Daxx could bind to SUMO1-3, but not SUMO 4 (Data not shown). Since the SIM-binding pocket is well conserved in SUMO4, these results are consistent with th e notion that Daxx may


70 not necessarily interact with free SUMO; rather it may have higher affinity to proteins that are modified by SUMO. The Two SIMs of Daxx are Conserved during Evolution A Daxx-like protein was identified in Drosophila melanogaster based on sequence identity with the mammalian Daxx proteins. To as sess molecular changes of the Daxx family in evolution, we have compared the protein sequenc es of Daxx from diverse species. As shown in Fig. 5-2A, the Daxx core domain is highly conser ved in species ranging fro m insects to humans. Interestingly, the beginning and ending block of the core domain contain sequences that potentially form coiled-coil helices, which likely mediate homoor hetero-oligomerization of the Daxx protein with itself or ot her binding partners. A novel nuc lear localization signal (NLS) identified recently (Yeung, Chen et al. 2008) resides in this region and is al most invariant during evolution (Fig. 5-2A). The sequences of the co re domain can be aligned end to end without major insertion/deletions or rearrangements, except for the Daxx protein from T. rubripes (tiger pufferfish), in which there are two major insertio ns (Fig. 5-2A). Such insertions, nonetheless, may form flexible loops, and thus might not affect the structural integrity of the Daxx coredomain. In addition, a second conser ved sequence block is present in animals other than insects. This block, termed conserved sequence region 1 (CR1), lies near the Nterminus. Among the 81 amino acid residues in this domain, there are seve n conserved Phe/Tyr residues (denoted with • in Fig. 5-2B), which account for about one-third of all such residues in the human Daxx and such unusual high concentration of Phe/Tyr suggests th at these conserved Phe/Tyr might be important for structural or functional aspects of the Da xx protein. Additionally, there are two conserved Cys residues in this block (denoted with in Fi g. 5-4B). The block ends with a cluster of three consecutive basic residues (Lys/Arg). This re gion contains the putative paired amphipathic helices (PAH), as suggested previously (Hollenb ach, McPherson et al. 2002), and might mediate


71 interactions with other protei ns such as ATRX (Tang, Wu et al. 2004), MSP58 (Lin and Shih 2002) and DMAP1 (Muromoto, Sugiyama et al. 2004) Strikingly, the two short SIMs in Daxx apparently exist in all animal species includi ng insects (Fig. 5-2C). A ll the SIMs consist of a hydrophobic core and acidic tails. In most sp ecies, the hydrophobic core contains four consecutive Ile/Leu residues. Va riations of the hydrophobic core in clude a substitution of Ile with Phe in the SIM1 from C. porcellus and S. purpuratus a sequence of five hydrophobic residues in the SIM1 from S. purpuratus and the substitution of the C-terminal Leu by Met or Pro in the SIM2 from C. intestinalis and A. aegypti respectively. In SIM1, the hydrophobic core is followed by a stretch of Asp/Gl u residues of variable length, ra nging from four to ten residues. Occasional Ser/Thr residues are found in the acidic tail from lower species (fishes, sea urchin, tunicates, and insects). For the Daxx SIM2, the acid ic tail is shorter and contains the invariable “SDSD” sequence in mammalian species, whereas quite variable acidic tails exist in other species. In the purple sea urchin P. purpuratus there is only one Glu residue C-terminal to the hydrophobic core. However, a stretch of four Glu/ Asp residues precedes the hydrophobic core in this organism, which might have similar role in mediating interaction with SUMO moiety, as SIM in reverse orientation appear s to have equivalent binding capacity to SUMO (Song, Zhang et al. 2005). In Drosophila, the hydrophobic co res of the two SIMs in the DLP exhibit considerable variation from othe r species; however they are c onserved among different species of the Drosophila genus (Fig. 53A). The hydrophobic core of SIM1 has an almost invariable sequence of VICVDL, whereas that of SIM2 consists of three c onsecutive Ile residues. Because of their obvious deviation from the consensus se quence of the Daxx SIMs, we assessed whether they are indeed bona fide SIM in their own right. The coding DNA sequences for both SIMs of D. melanogaster were fused with that of Gal4 activ ation domain (AD) and the resulting


72 constructs were assayed for interaction with SUMO1 or 3 that was fused with Gal4 DNAbinding domain (BD) in yeast twohybrid assays. As shown in Fig. 5-3B, both SIMs were able to bind to SUMO1 or 3. Therefore, the Drosoph ila DLP possesses two functional SIMs. Preservation of Functional Domains of the Daxx Family of Proteins during Evolution Based on aforementioned sequence analysis it becomes apparent that several noncontiguous regions of the Daxx protein family ar e highly conserved in diverse species (Figure 54). The order of arrangement of these conser ved domains is also preserved. The two SIMs occupy one or the other extreme end of the pr otein, whereas the coredomain resides in the central part. Interestingly, located at C-terminal to the core domain, a stretch of acidic residues of quite variable length is found in all members of this family, indi cative of functional significance of this acidic patch. Our previous study demonstr ated that this acidic sequence is important for interacting with the ba sic C-terminal tail of p53 (Zhao, Li u et al. 2004). Because of repetitive nature of this region, it is expect ed that this sequence might be s ubject to higher level of variation during evolution. The CR1 region is found between SIM1 and the core domain in diverse organisms from C. intestinalis to H. sapiens although this domain is c onspicuously absent in all insect species such as A. aegypti and D. melangogaster Interestingly, the spacers between the conserved domains largely consist of Pro/Ser-rich sequences or th e so-called PEST sequences in species from C. intestinalis to H. sapiens Although the spacers are high ly similar in sequence in closely related species, they are quite variable in mammals and completely different in distant species. Several sites in the spacers may be phosphorylated. For examples, Ser-178 (Ryo, Hirai et al. 2007) and Ser-668 (Ecsedy, Michaelson et al. 2003) of th e human Daxx appear to be phosphorylated and such modifications may re present signaling events in stress-induced apoptosis (Song and Lee 2004; Ryo, Hirai et al. 2007), and such events may largely organismspecific, because these phosphorylation sites ar e not always conserved. The DLP in the genus


73 Drosophila is much larger than the orthologs in other species, even in other insects such as mosquitoes (e.g. A. aegypti ). Insertion of repeti tive sequences might explain the expanded Daxx sequence in Drosophila For example, the Gln-rich region of the D. melanogaster DLP is encoded by a retro-element called TART (tel omere-associated non-LTR retrotransposon) (Casacuberta and Pardue 2003) as well as the CAG tri-nucleotide simple repeats. SUMO mediates Daxx-Ubc9 Int eraction via the SIMs of Daxx Ryu et al. reported that Daxx also interacts with Ubc9 (R yu, Chae et al. 2000). However, the nature of Ubc9-Daxx interaction remains undefined. Although the results described above indicate that Daxx binds to SUMO directly, it is also possible that SUMO might interact with Daxx via Ubc9 when SUMO is conjugated to Ubc9, thereby bypassing the SIM-SUMO interaction. We have addressed this issue using various mutant s of Daxx and Ubc9. Data shown in Fig. 5-5A demonstrate, surprisingly, that pres ence of at least one SIM in Daxx is required for Daxx-Ubc9 interaction as Daxx constructs lack ing functional SIMs faile d to bind Ubc9. Since Ubc9 does not have a binding pocket for SIM, thes e results appear to i ndicate that Daxx might interact with Ubc9 when it is covalently conj ugated to SUMO, which occu rs in the sumoylation reaction cascade. To test this possibility, we examined whether Daxx could bind to Ubc9 C93A mutant. As shown in Fig. 5-5B and D, Daxx can stil l bind to this mutant a nd that such interaction strictly requires at least one f unctional SIM. Thus, conjugation of Ubc9 to SUMO is not required for Daxx-Ubc9 interaction. It was shown recen tly that Ubc9 also interacts with SUMO noncovalently and several residues in Ubc9 including H20 are essential for such interaction (Knipscheer, van Dijk et al. 2007). Mutating H 20 to aspartic acid abolishes noncovalent Ubc9SUMO interaction, but s till permits their conjugation via th e thioester bond (Knipscheer, van Dijk et al. 2007). We confirmed that Ubc9 H20D mutant can no longer bind SUMO1, SUMO2 (Fig. 5-5C), or SUMO3 (data not shown). This mutation also a bolished Daxx-Ubc9 interaction


74 (Fig. 5-5D). Taken together, these results sugg est that Daxx, Ubc9 and SUMO form trimeric complex and no covalent bond is required for the complex formation. Furthermore, at least one SIM in Daxx is necessary for the formation of this complex. Roles of Daxx SIMs in Daxx-SU MO and Daxx-PML Interaction We also assessed potential in teraction of the SIMs of Daxx and SUMO in cultured cells. As shown in Fig. 5-6A, in transfected Saos2 ce lls, GFP-SUMO3 colocalizaed with wt Daxx (Fig. 5-6A panels a-c), the I7K mutant (Fig. 5-5A panels d-f) or I733K mutant (Fig. 5-6A panels g-i), but not with the I7/733K double mutant (Daxx SIM; see Fig. 5-6A panels j-l), corroborating the findings that the two SIMs of Daxx independently interact with SUMO. Similarly, the SIM of Daxx is also required for inter acting with SUMO1 (dat a not shown). It is known that Daxx is recruited to PML-NBs via interaction with su moylated PML and that the SIMs of Daxx is essential for colocalizatio n of Daxx with PML-NBs (Lin, Huang et al. 2006). To test whether this may indeed be the case, we coexpressed PML (i soform VI) with wt Daxx or the I7/733K double mutant in Saos2 cells. As expected, wt Daxx co localized with PML (Fig. 5-6B panels a-c). Surprisingly, the Daxx SIM mutant also colocalized with PML (Fig. 5-6B panels d-f). We further tested whether the sumoylation status of PML might affect PM L-Daxx interaction. There are three major SUMO-modification sites in PML (K65, 160 and 490) (Kamitani, Kito et al. 1998; Duprez, Saurin et al. 1999). We mutated all three residues to Arginine. Surprisingly, this triple mutant (PML S) could still colocalize with wt Daxx (Fig. 5-6B panels g-i). Additionally, we detected clear colocalization of PML S with the Daxx SIM mutant, although the amount of this Daxx mutant was significantly reduced in the colocalization spots (poi nted with arrows in panels j-l of Fig. 5-6B). Thus, it appears likely that sumoylation of PML and the SIMs of Daxx are required for strengthening Daxx-PML interactions.


75 Interplay of Daxx and PML in Regula ting c-Jun-Mediated Transcription It has been demonstrated that PML markedly activates transcrip tion mediated by c-Jun (Best, Ganiatsas et al. 2002) or c-Fos (Vallian, Gaken et al. 1998), two major components of the AP-1 complex. We examined whether Daxx alone or together with PML might also modulate cJun-dependent transcription. We fused c-Jun to the Gal4 DNA-binding domain and assessed whether this fusion construct a ffects transcription in a hete rologous reporter system. Data presented in Fig. 5-7 show that tethering c-J un to the artificial prom oter containing Gal4 DNAbinding sites and the adenovirus E4 core promoter elements activated the reporter expression by about 8-fold. Coexpression of Gal4-c-Jun and wt Daxx markedly enhanced c-Jun-mediated activation by ~45-fold. However, mutating either of the DaxxÂ’s SIM abolished this coactivation (Fig. 5-7A). c-Jun is essen tial for this coactivation because Daxx or Gal4-BD alone or coexpressed together had minimal effects on the reporter activity (data not shown). We observed similar effects using reporters containing other core promoter s (data not shown). Thus, in contrast to roles of Daxx in repression, Daxx could also activate tran scription, as reported previously (Boellmann, Guettouche et al. 2004), and both SIMs in Daxx are required for this coactivation. Consistent with a previous report (Best, Ganiatsa s et al. 2002), PML drastically increased c-Jun-mediated transcription in our re porter system (Fig. 5-7B). The oncogenic fusion product PML-RAR was completely devoid of this prope rty. Surprisingly, mutating the three major sites of sumoylation (K65, 160 and 490) indi vidually or in combination markedly reduced or completely abolished the ability of PML to coactivate c-Jun-mediated transcription (Fig. 57B). Notably, K65R and K160R retained resi dual activity in coactivating c-Jun, but the K65/160R double mutant or any mutant contai ning the K490R mutation individually or in combination was totally devoid of this activity. Therefore, sumoylation of PML is required for coactivating transcription with c-Jun.


76 To assess potential interplay between Daxx and PML, we coexpressed PML and Daxx in the reporter assays. As shown in Fig. 5-7C, the reporter activity was increased for ~50-fold when both PML and Daxx were expressed. The fold induc tion is similar to that when Daxx alone was expressed but much less than that when only PML was expressed with the reporters. Interestingly, coexpression of a Daxx mutant wi th mutation in either or both SIMs with PML resulted in ~100-fold induction, which is simila r to that when only PML was expressed in the reporter assays (Fig. 5-7C). These results sugg est that Daxx and PML independently regulate cJun-mediated transcription with Daxx having a do minant role over PML. Indeed, data shown in Fig. 5-7D reinforces this notion. Co-expression of wt Daxx with various PML mutants except for the PML-RAR oncogene still led to stimulation of cJun-dependent transcription and the fold induction was similar to that when Daxx wt was expressed alone (Fig. 5-7D). Interestingly, the reporter activity was markedly reduced wh en wt Daxx was coexpressed with PML-RAR to ~18-fold (lane 2 in Fig. 5-7D ), suggesting that PML-RAR might interfere with Daxx-mediated coactivation of c-Jun. Roles of c-Jun Sumoylation in Daxx-Mediated Coactivation C-Jun is modified by SUMO, and K229 is the ma jor site of sumoylation (Muller, Berger et al. 2000). To assess whether sumoylation of cJun has a role in Daxx-mediated coactivation of c-Jun, we tested the effects of Daxx on c-Jun K 229A mutant. As shown in Fig. 5-8, the Gal4-cJun K229A construct was much more potent in activating the reporter expression, which, in accord with a previous finding (Muller, Berger et al. 2000), suggests that K229 sumoylation has repressive effects on c-Jun-regulated transcripti on. Coexpression of Daxx with this mutant also clearly resulted in further increase of the report er activity. However, the extent of induction by Daxx was different: Daxx increased the transcri ption mediated by wt c-Jun by ~10-fold, but enhanced that by the c-Jun K229A mutant by only ~4-fold (Fig. 5-8C). Thus, sumoylation of c-


77 Jun at K229 may be involved in Daxx-mediated coactivation with c-Jun. Daxx can interact with HDACs (Hollenbach, McPherson et al. 2002; Puto and Reed 2008). Additionally, HDAC activities are involv ed in suppressing c-Jun-regulated gene expression (Ogawa, Lozach et al. 2004). To ga uge potential involveme nt of HDACs in Daxxmediated transcriptional coregula tion with c-Jun, we added trichosta tin A (TSA) in one set of the transfected cells in the reporter assays. Data sh own in Fig. 5-8 revealed that TSA dramatically enhanced c-Jun-mediated transactivation and su ch additional enhancement occurred regardless of the presence or absence of Daxx. Indeed, th e fold induction by TSA was similar with or without Daxx coexpression (Fig 5-8B). Thus it appears that Daxx and TSA independently stimulate c-Jun-mediated transcription. The fact that TSA stimulated transcription mediated by wt c-Jun or the K229A mutant to a similar extent suggests that repression exerted by sumoylation is independent of HDAC activities and is in addition to HDAC-mediated repression for c-Jun-regulated transcrip tion. Such notion was also proposed for Sp3-mediated gene regulation (Valin and Gill 2007).


78 Figure 5-1. Daxx interacts with SUMO1 via SI Ms. (A) Daxx possesses two distinct SUMOinteracting motifs (SIMs). The sequences of the two Daxx motifs are aligned against that of other known SIMs. The starting a nd ending positions of each motif in each protein are shown. The hydrophobic core is de picted in a gray box. Shown on the top is a schematic drawing of the linear seque nce of Daxx with potential sequence or structural domains. Proteins that are know n to interact with Daxx are listed above a domain with which they interact. (B) Daxx binds to SUMO1 via SIMs. Yeast twohybrid assays were performed with indicated hybrid constructs and the transformants were grown on medium without histidine. (C) In vitro intera ction of Daxx with SUMO1. GST and indicated GST-Daxx fusions were expressed in and purified from E. coli and were incubated with purified SUMO1 (Bost on Biochem). After extensive washes, the beads were mixed with SDS sample buffer and the samples were subjected to SDS-PAGE and Western blot analysis with anti-SUMO1 antibody. GSTDaxx NT carries amino acids 1-130 and GST-Daxx CT contains residues 644-740 (denoted with arrowheads).


79 Figure 5-2. Sequence alignment of evolutionarily conserved domains of the Daxx orthologs. (A) An alignment of the Daxx core domain. The multiple-sequence alignment was obtained using the ClustalW program in the MacVector software package. Positions at which the amino acid re sidues are conserved in 50% of the selected species are depicted with white letter in black ba ckground. Two putative coiled-coil motifs are indicated with a line. A nuclear localizati on signal (NLS) with four Lys/Arg residues is also indicated. The numbers indicate positions of the beginning or ending residues in each protein. (B) Alignment of the cons erved region 1 (CR1). Alignment was made as in (A). Multiple conserved Phe/Tyr residues are denoted with dots (•) and two conserved Cys residues with asterisks (*). (C) Alignments of two conserved SIMs of Daxx. The hydrophobic core is depicted w ith white letters in black background.


80 Figure 5-3. The Daxx SIMs in the Drosophila genu s. (A) Sequence alignm ents of the Daxx SIMs in various Drosophila species. Different residues are highlight ed with white letter in black background. (B) The Daxx SIMs of D. melanogaster (Dme) interact with SUMO. The DNA sequences encoding the two Dme SIMs were fused to Gal4 BD and tested for interaction with the human SUMO1 or SUMO3 fused to Gal4 AD as in Fig. 5-1.


81 Figure 5-4. Evolutionary conserva tion of domain structure of the Daxx family of proteins. Shown is schematic representation of the Daxx orthologs from i ndicated species. The double SBMs at each end of the protein are drawn as black ovals. The core domain is shown as black rectangles. The two putative coiled-coil motifs (CC) within the core domain are indicated. The stre tch of acidic residues is de picted as hatched boxes. The CR1 domain, absent in insect species, is shown as gr ay rectangles. The numbers denote the beginning or ending positions of each conserved domain. The accession numbers for the Daxx orthologs assigned in GenBank, Swiss-Prot or Ensembl databases are shown. The figures are not drawn to scale.


82 Figure 5-5. Daxx interacts with Ub c9 via SIMs. Yeast two-hybrid assays were conducted with indicated hybrids. The yeast transformants were restreaked in agar medium lacking histidine. (A) A functional SIM in Daxx is required for binding to Ubc9. (B) Specificity of Daxx-Ubc9 interaction. Indi cated Daxx and Ubc9 mutants were tested for interaction in yeast two-hybrid assays. Sectors 1 & 2 show positive controls in the two-hybrid assays using the indicated A d12 E1B and mSin3A hybrid constructs. (C) Noncovalent interaction betw een SUMO and Ubc9. Sector 1 show positive control in the two-hybrid assays using the indica ted Mdm2 and p53 hybrid constructs. (D) Noncovalent SUMO-Ubc9 intera ction is required for DaxxUbc9 interaction via SIM.


83 Figure 5-6. Colocalization of Daxx and its mutant s with SUMO3 and PML. (A) Requirement of SIM for colocalization of Daxx with SUMO 3. Full-length SUMO3 was fused at its Nterminus with GFP and the corresponding e xpression vector was cotransfected with that for wt Daxx or indicated mutants into Saos2 cells. Cells were fixed 24h after transfection and Daxx was detected with anti-FLAG antibody and goat anti-rabbit IgG conjugated with rhodamine. Arrows denote SUMO3 foci wh ere wt Daxx, the I7K or I733K mutant, but not th e I7/733K double mutant (Daxx SIM) was also concentrated.


84 Figure 5-6. Colocalization of Daxx and its mu tants with SUMO3 and PML (continued). (B) Determinants of Daxx-PML colocalization. PML wt (isoform VI) or the K65/160/490R triple mutant (PML S) was coexpressed with GFP-Daxx wt or the SIM mutant in Saos2 cells. Cells were fixed 24h after transfection. PML wt was epitope-tagged with FLAG at the N-termi nus and detected with anti-FLAG antibody as in (A). PML S mutant was epitope-tagged with (His)6 at the N-terminus and stained with mouse monoclonal anti-( His)6 antibody and rabbit anti-mouse IgG conjugated with rhodamine. Note that the an ti-(His)6 antibody appeared to recognize the nuclear pore complex in the transf ected as well as non-transfected cells.


85 Figure 5-7. Daxx coactivates c-Junregulated transcription. HCT116 cells were transfected with the firefly luciferase report er driven by a heterologous pr omoter containing five Gal4 DNA binding sites upstream of the adenovirus E4 core promoter as depicted, and the control see pansy luciferase reporter driven by the SV40 promoter, along with other indicated expression plasmids Dual reporter luciferase assays were done 24h after transfection. Shown are the average values of the normalized firefly luciferase activities from two independent transfections with standard deviations. (A) The SIMs of Daxx are required for coactivation of c-Jun-mediated tr ansactivation. (B) Requirement of the sumoylation sites of PML for PML-mediated coactivation of cJun. (C) Effects of wt Daxx and corresponding SIM mutants on PML-mediated coactivation of c-Jun. (D) Effects of wt PML and its sumoylation mutants on Daxxmediated coactivation on c-Jun.


86 Figure 5-8. Independent regulati on of c-Jun by Daxx and TSA. HCT116 cells were transfected with reporters as in Fig. 5-7, along with th e indicated expression plasmids for c-Jun and Daxx. TSA was added to the final concentration of 1.3 M in one set of the transfected cells 12 after transfection. Cells were processed for dual luciferase assays 24h after transfection. (A) Coactivation of tr anscription mediated by wt c-Jun and the K229A mutant by Daxx and TSA. (B) Impact of TSA on c-Jun-regulated transcription in the absence or presence of Daxx coexpression. Fold increase due to TSA treatment was calculated by dividing the normalized luciferase activities in the presence of TSA by that in its absence. (C) Effects of Daxx coexpression on c-Junmediated transcription. Fold increase due to Daxx cotransfection was obtained by dividing the normalized luciferase activities in the presence of Daxx cotransfection by that in the absence.


87 CHAPTER 6 SUMMARY AND CONCLUSIONS Discussion of SUMO-M odification of P53 Data presented in Chapter 4 are consistent with our notion that SUMO modification of p53 negatively regulates the expression of p53 ta rget genes (Figures 4-2, 4-3 and 4-6). P53SUMO fusions and chemical fused p53-SUMO impair ed the ability of p53 to activate its targets responsible for cell-cycle arrest (p21) and apoptosis (PIG3 and PIDD) as seen in luciferase reporter assays and real-time PCR experiments. P53-SUMO was also largely defective to suppress H1299 cell growth as shown by colony fo rmation assays (Figure 4-1). Thus SUMO negatively regulates p53 function. We have shown that SUMO modification of p53 resulted in strong nuclear export of p53 (Figures 4-7 and 4-8). Thus, st rong nuclear export of p53 promot ed by sumoylation depletes p53 in the nucleus where it activates transcription, leading to th e inhibition of p53 function in transcriptional activation. This was demons trated for both SUMO1 and SUMO3 where both the p53-SUMO fusions and the chemical induced SUMOylation of p53 re sulted in strong p53 cytoplasmic localization. Since SUMO2 is 97% similar to SUMO3 (G eiss-Friedlander and Melchior 2007; Dohmen 2004), it can be hypothesi zed that SUMO2 when modified to p53 could also result in significant p53 cy toplasmic localization, although fu rther investigation is required to address this topic. SUMO 4 is not known to modify p53. Our study has also revealed that the expor t of p53 when SUMO-modified involves the p53 regulator, Mdm2 as shown in IF experiments using mouse embryonic fibroblasts deficient of Mdm2 and p53 genes (Figure 4-9). The E3 ligase activity of Mdm2 was shown to be important as ablation of that activity via the C438A mutatio n in its RING domain resulted in p53 nuclear


88 localization. These results are consistent w ith documented roles for Mdm2 in regulating intracellular trafficking of p53. Furthermore, we have demonstrated thr ough immunoprecipitation expe riments (IPs) that SUMO-modification of p53 does not disrupt the p53 tetramer (F igure 4-10 and Figure 4-11). Yeast two-hybrid (Y2H) assays al so revealed that disruption of oligomerization of p53 abolished its interaction with CRM1 (Figure 4-13). In light of these stud ies, it appears likely that p53 is exported as a tetramer to the cytoplasm thus ch allenging the paradigm that p53 is exported as a monomer. Mutational analysis of the hydrophobic SUMO pocket further revealed the involvement of a SUMO interacting motif (SIM) of an unknown protein in the nuclear export of p53 as shown in our immunoflourescence studies (Figure 4-10 ). The SUMO1 and 3 pocket mutants (F36A and F31A respectively) when chemically fused to p53 failed to export p53 to the cytoplasm (Figure 4-10). The results point to an unknown protein that is involved in a putative SUMO-SIM interaction. Amino acid sequence analysis of the nuclear exporter, CRM1, revealed a putative SIM positioned at amino acids 429-432 of CRM1. Mutational analysis of this SIM suggested a potential role of this motif in efficient release of p53 from th e export machinery. We found that mutation of V430 of CRM1 enhanced its interact ion with wt p53, the p53 K386R mutant, or p53SUMO1 fusion (Figure 4-14 and 4-15). Conve rsely, compared to wt p53, p53-SUMO1 F36A pocket mutant and p53 K386R exhibited stronger interactions with wt CRM1 (Figure 4-14 and 415). These results suggest that SUMO-modificat ion of p53 weakens its interaction with CRM1. We propose that sumoylation of p53 facilitates the release of p53 to the cytoplasm from the CRM1 exporter complex (see below).


89 Finally, over-expression of Ul p1p family of SUMO proteases in human osteosarcoma Saos2 and breast cancer MCF7 cells revealed possible involvement of the pore-associated membrane SENPs, SENP1 and 2, in the release of p53 from the CRM1 export machinery at the nuclear pore complex (NPC). Overexpression of SENP2 permitted p53 localization to the cytoplasm when p53 was SUMO-modified (Figure 416A-B). In contrast, over-expression of the Ulp2p family of SENPs, SENP3-7 inhibited th e export of chemical fused p53-SUMO to the cytoplasm (Figure 4-16 C-I). It can be theorized that overexpression of these SENPs destabilized an unknown SUMO-modified protein (designated as Factor X, s ee Figure 6-1) involved in export. One possibility could be Ran-GTP that is involved in the CRM1 export machinery where GTP is hydrolyzed at the NPC for the release of cargo. However, further analysis will be needed to determine if Ran-GTP is i ndeed SUMO-modified or if an unknown adaptor protein is involved in the export of p53. In the light of all these observations, the following hypothetical p53 export model has been proposed. Under low stress conditions, p53 is sumoylated by Mdm2 but the modification does not inhibit p53Â’s ability to form the tetramer. However, it might inhibit its binding to p53 response elements and the subsequent activati on of the target genes. The SUMO-modified p53 tetramer can be bound to the C-terminal region of CRM1 along with RanGTP. Published reports have shown that the RanGTP is required for the ternary export complex (CRM1-RanGTP-Cargo) (Petosa et al. 2004). It is hypothesi zed that a SIM exists in RanG TP. P53-SUMO interacts with a SIM. It is also hypothesized that RanGTP has a SUMO-modification site. RanGTP is sumoylated at this site. The SUMO pocket of RanGTP-SUMO can interact with CRM1Â’s SIM. The ternary export complex is shuttled to the nuclear por e complex (NPC) where SENP1 and SENP2 reside.


90 RanGTP hydrolysis occurs at the NPC to give RanGDP. SUMO is deconjugated from RanGDP by the SENPs. The p53 tetramer is released and RanGDP is imported back into the nucleus. Further analysis is needed to validate th is hypothesis. Future e xperiments include the determination of a possible SIM and SUMO-modification site in RanGTP and the identification of any adapter proteins that are involved in the export machinery. Th e roles of the poreassociated SENPs and the status of the p53 tetramer in the cytoplasm merits further investigation. Understanding in detail the determinants of p53 loading and release via the CRM1 pathway will be very important in the future design of molecules (small peptides or chemical inhibitors) to inhibit this export. The study could also be extra polated to other proteins that can be SUMO-modified and exported via the CRM1 m achinery. In conclusion, this study provides new insights into the role of p53 sumoylation an d such knowledge will have implications in our understanding of tumorigenesis a nd future anticancer therapy. Discussion of SUMO-Interacting Motifs of Daxx The Daxx protein exists in diverse species ranging from insects to mammals. Although there is only limited sequence identity among orthol ogs of Daxx in distant species, the presence of a well-conserved core domain of ~200 residues permits un ambiguous identification of the members of this family in organisms that co ntain this protein. Our results reported here demonstrate that in addition to the core domain, two short sequence motifs capable of interacting with SUMO are present in all members of this family. Significantly, the position of two SIMs is also conserved, one at each end of this protein. These features s uggest that the pr esence of SIMs and their positions are important for biological function of this protein. SIM occurs in proteins with diverse functions and mediates noncovalent interactions with a wide variety of SUMO-conjuga ted proteins (Kerscher 2007). It is envisioned that SUMObinding proteins play fundamental roles in me diating biological functions of sumoylated


91 proteins. In the case of transcriptional regu lation, SUMO-binding protei ns may facilitate the recruitment and retention of core pressors or coactivators, resulti ng in transcriptional repression or activation. In this study, we have shown th at Daxx, a known transcriptional regulator, has two distinct SUMO-interacting motifs (SIMs). It wa s reported that the C-terminal SIM of Daxx (SIM2) seems important for Daxx-mediated transc riptional repression (Li n, Huang et al. 2006). In contrast, we showed that Daxx stimulates c-Jun-mediated transcri ption and both SIMs are required for this effect. Interestingly, Daxx-medi ated coactivation of c-Jun partly depends on the integrity of K229, the prin cipal site of c-Jun sum oylation (Muller, Berger et al. 2000). Thus, it is likely that Daxx interacts with SUMO-modified c-Jun in the coactiv ation mechanism. We showed that the c-Jun K229A mutant is a much more potent transactivator (Fig. 5-8), suggesting that sumoylation of c-Jun has intrinsic repressi ve effects on transcripti on, as proposed by others (Muller, Berger et al. 2000). Binding of Daxx to sumoylated c-Jun via SIMs could sequester SUMO moiety, thereby preventing corepressors from binding to SUMO-modified c-Jun. This mode of coactivation could al so operate in coactivation w ith other sumoylated DNA-binding transcription factors. Indeed, a recent manuscrip t demonstrated that the Drosophila Daxx (DLP) seems to interact with SUMO-modified Drosoph ila p53 (Mauri, McNamee et al. 2008). In light of our results that DLP possesses SIMs, one obvi ous possibility is that DLP binds sumoylated p53 partly through one or both SIMs of DLP. Ad ditionally, Daxx has been shown to enhance HSF1-mediated transcrip tion (Boellmann, 2004). HSF1 is a known SUMO-modified transcription factor (Hietakangas, 2003; Hong, 2001). We predict that th e two SIMs of Daxx will be important for it to enhance HSF1-dependent transcription. Future studies should address whether binding of Daxx to SUMO-modified tran scription factors such as c-Jun, p53 and HSF1 would preclude the recruitments of corepressors to the promoter s regulated by these factors.


92 Our data revealed that the Daxx mutant that could no longe r interact with SUMO (the I7/733K double mutant or Daxx SIM) still colocalized with PML (Fig. 5-6). One obvious question is how Daxx SIM mutant could be recruited to PML-NB in the absence of SUMObinding. One scenario could be th at SUMO-modified Daxx might be targeted to PML-NB, since certain isoforms of PML also carry a SIM (Shen, Lin et al. 2006). However, this seems unlikely, as PML-VI, the isoform used in our experiment s, does not contain a SIM. Furthermore, our unpublished data showed that Daxx itself is not SUMO-modified. Altern atively, Ubc9 could tether Daxx to PML-NB, because Daxx interacts with Ubc9 (Fig. 5-5). This is also unlikely because Daxx SIM mutant could no longer interact with Ubc9 (Fig. 5-5). Based on these considerations, we propose that hetero-oligom erization of Daxx with PML, rather than interaction between SIMs of Daxx and SUMO-c onjugated PML per se, might be critically important for PML-Daxx colocalization in PML-NB Nonetheless, our data demonstrated that the ability of the Daxx SIM mutant to colocalize with PML S mutant was significantly impaired (Fig. 5-5B), suggesting that the SIMs of Daxx and sumoylati on of PML, although not essential, have important ro les in stabilizing PML-Daxx inte raction (Figure 6-2). Further biochemical experiments will be required to te st whether Daxx and PML form hetero-oligomers via the coiled-coil sequences of Da xx (see Figures 5-2 and Figure 6-2). Daxx has been described as a promiscuous bi nding partner that inte racts with a wide variety of proteins with dive rse functions. In most cases, th e Daxx-binding proteins were identified through yeast two-hybrid screens and th e C-terminal fragment harboring the SIM2 of Daxx was implicated as the interacting domai n. The ability of Daxx to bind to SUMO could provide an explanation as to w hy Daxx exhibits binding-affinity to diverse proteins. Presumably, Daxx could interact with proteins that undergo su moylation in yeast cells, as the enzymology of


93 SUMO conjugation is conserved from yeast to mammals. Daxx may not bind to all SUMOmodified proteins. The binding specificity may be imparted by a Daxx-binding domain within a specific protein and the SUMO-moi ety attached to such protein is expected to strengthen its affinity to Daxx. The PML-Daxx interaction app ears to exemplify this mode of molecular recognition (see Figure 6-2). Figure 6-1. P53 export in the context of SUMO -modification. SUMO-modified p53 tetramerizes and associates with CRM1 and the SUMO -Interacting Motif of a SUMO-modified Factor X-Mdm2 complex which in turn interacts with CRM1Â’s SIM. The export complex is then translocated to the nu clear pore complex on the nuclear membrane where SENPs deconjugate the SUMO-modified substrates and results in the release of p53 into the cytoplasm.


94 Figure 6-2. A model explaining Daxx-PML interaction. Both PML and Daxx have coiled-coil sequences (Jensen, 2001 and see Figure 5-2) which might allow for their heterooligomerization. Additional contacts betw een the two SIMs of Daxx and the SUMOs attached to PML are expected to strengthen Daxx-PML interaction.


95 REFERENCES Alkuraya, F. S., I. Saadi, et al (2006). "SUMO1 haploinsufficiency leads to cleft lip and palate." Science 313 (5794): 1751. Bailey, D. and P. O'Hare (2004). "Characterization of the localization and proteolytic activity of the SUMO-specific protease, SENP1." J Biol Chem 279 (1): 692-703. Bayer, P., A. Arndt, et al. (1998). "Structure determination of the small ubiquitin-related modifier SUMO-1." J Mol Biol 280 (2): 275-86. Best, J. L., S. Ganiatsas, et al. (2002). "SUMO-1 prot ease-1 regulates gene transcription through PML." Mol Cell 10 (4): 843-55. Bischoff, F. R., H. Krebber, et al. (1995). "C o-activation of RanGTPase and inhibition of GTP dissociation by Ran-GTP bindi ng protein RanBP1." Embo J 14 (4): 705-15. Bodai, L., N. Pardi, et al. (2007). "Daxx-like protein of Drosophila interacts with Dmp53 and affects longevity and Ark mRNA level." J Biol Chem 282 (50): 36386-93. Bode, A. M. and Z. Dong (2004). "Post-translati onal modification of p53 in tumorigenesis." Nat Rev Cancer 4 (10): 793-805. Boellmann, F., T. Guettouche, et al. (2004). DAXX interacts with heat shock factor 1 during stress activation and enhances its transcriptional activity." Proc Natl Acad Sci U S A 101 (12): 4100-5. Bossis, G., C. E. Malnou, et al. (2005). "Downregulation of c-Fos/c-Jun AP-1 dimer activity by sumoylation." Mol Cell Biol 25 (16): 6964-79. Boyd, S. D., K. Y. Tsai, et al. (2000). "An in tact HDM2 RING-finger domain is required for nuclear exclusion of p53." Nat Cell Biol 2 (9): 563-8. Budhu, A. S. and X. W. Wang (2005). "Loading and unloading: orchestrating centrosome duplication and spindle assembly by Ran/Crm1." Cell Cycle 4 (11): 1510-4. Casacuberta, E. and M. L. Pardue (2003). "Tra nsposon telomeres are widely distributed in the Drosophila genus: TART elements in the virilis group." Proc Natl Acad Sci U S A 100 (6): 33638. Chang, C. C., D. Y. Lin, et al. (2005). "Daxx mediates the small ubiquitin-like modifierdependent transcriptional repr ession of Smad4." J Biol Chem 280 (11): 10164-73. Chen, L. and J. Chen (2003). "MDM2-ARF co mplex regulates p53 sumoylation." Oncogene 22 (34): 5348-57.


96 Cheng, J., T. Bawa, et al. (2006). "Role of desum oylation in the development of prostate cancer." Neoplasia 8 (8): 667-76. Cho, Y., S. Gorina, et al. (1994). "Crystal st ructure of a p53 tumor suppressor-DNA complex: understanding tumorigenic mutations." Science 265 (5170): 346-55. Clore, G. M., J. G. Omichinski, et al. (1994). "H igh-resolution structure of the oligomerization domain of p53 by multidimensional NMR." Science 265 (5170): 386-91. Colby, T., A. Matthai, et al (2006). "SUMO-conjugating a nd SUMO-deconjugating enzymes from Arabidopsis." Plant Physiol 142 (1): 318-32. Cook, A., F. Bono, et al. (2007). "Structural biology of nucleoc ytoplasmic transport." Annu Rev Biochem 76 : 647-71. Coutts, A. S. and N. B. La Thangue (2005). "T he p53 response: emerging levels of co-factor complexity." Biochem Biophys Res Commun 331 (3): 778-85. Croxton, R., L. A. Puto, et al. (2006). "Daxx represses expression of a subset of antiapoptotic genes regulated by nuclear f actor-kappaB." Cancer Res 66 (18): 9026-35. Di Bacco, A., J. Ouyang, et al. (2006). "The SU MO-specific protease SENP5 is required for cell division." Mol Cell Biol 26 (12): 4489-98. Di Ventura, B., C. Funaya, et al. (2008). "Reconstitution of Md m2-dependent post-translational modifications of p53 in yeast." PLoS ONE 3 (1): e1507. Dohmen, R. J. (2004). "SUMO protei n modification." Biochim Biophys Acta 1695 (1-3): 113-31. Dornan, D., H. Shimizu, et al. (2003). "The prolin e repeat domain of p53 binds directly to the transcriptional coactivator p300 and allosterically controls DNA-dependent acetylation of p53." Mol Cell Biol 23 (23): 8846-61. Duprez, E., A. J. Saurin, et al. (1999). "SUM O-1 modification of the acute promyelocytic leukaemia protein PML: implications for nuclear localisation." J Cell Sci 112 (Pt 3) : 381-93. Ecsedy, J. A., J. S. Michaelson, et al. (2003) "Homeodomain-interact ing protein kinase 1 modulates Daxx localization, phosphorylation, an d transcriptional activity." Mol Cell Biol 23 (3): 950-60. el-Deiry, W. S., S. E. Kern, et al. (1992). "Def inition of a consensus bi nding site for p53." Nat Genet 1 (1): 45-9. Emelyanov, A. V., C. R. Kovac, et al. (2002). "T he interaction of Pax5 (BSAP) with Daxx can result in transcriptional activ ation in B cells." J Biol Chem 277 (13): 11156-64.


97 Fornerod, M. and M. Ohno (2002). "Exportin-med iated nuclear export of proteins and ribonucleoproteins." Results Probl Cell Differ 35 : 67-91. Freedman, D. A. and A. J. Levine (1998). "Nuc lear export is required for degradation of endogenous p53 by MDM2 and human papi llomavirus E6." Mol Cell Biol 18 (12): 7288-93. Fuchs, S. Y., C. G. Lee, et al. (2002). "SUM O-1 modification of Mdm2 prevents its selfubiquitination and increases Mdm2 ability to ubiquitinate p53." Cell 110 (4): 531. Funk, W. D., D. T. Pak, et al. (1992). "A transcriptionally ac tive DNA-binding site for human p53 protein complexes." Mol Cell Biol 12 (6): 2866-71. Geiss-Friedlander, R. and F. Me lchior (2007). "Concepts in sum oylation: a decade on." Nat Rev Mol Cell Biol 8 (12): 947-56. Gill, G. (2004). "SUMO and ubiquitin in the nucle us: different functions, similar mechanisms?" Genes Dev 18 (17): 2046-59. Gong, L. and E. T. Yeh (2006). "Characterizatio n of a family of nucleolar SUMO-specific proteases with preference for SU MO-2 or SUMO-3." J Biol Chem 281 (23): 15869-77. Gongora, R., R. P. Stephan, et al. (2001). "An essential role for Daxx in the inhibition of B lymphopoiesis by type I interferons." Immunity 14 (6): 727-37. Gostissa, M., A. Hengstermann, et al. (1999). "A ctivation of p53 by conjugation to the ubiquitinlike protein SUMO-1." Embo J 18 (22): 6462-71. Gostissa, M., M. Morelli, et al. (2004). "The transcriptional repre ssor hDaxx potentiates p53dependent apoptosis." J Biol Chem 279 (46): 48013-23. Green, D. R. and J. E. Chipuk (2006). "p53 and metabolism: Inside the TIGAR." Cell 126 (1): 30-2. Grossman, S. R., M. E. Deato, et al. (2003). "Polyubiquitination of p53 by a ubiquitin ligase activity of p300." Science 300 (5617): 342-4. Grossman, S. R., M. Perez, et al. (1998) "p300/MDM2 complexes participate in MDM2mediated p53 degradation." Mol Cell 2 (4): 405-15. Gu, J., L. Nie, et al. (2001). "Identification of p53 sequence elements that are required for MDM2-mediated nuclear export." Mol Cell Biol 21 (24): 8533-46. Guo, D., M. Li, et al. (2004). "A functional variant of SUMO4, a new I kappa B alpha modifier, is associated with type 1 diabetes." Nat Genet 36 (8): 837-41.


98 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. Hannich, J. T., A. Lewis, et al. (2005). "D efining the SUMO-modified proteome by multiple approaches in Saccharomyces cerevisiae." J Biol Chem 280 (6): 4102-10. Harms, K., S. Nozell, et al. (2004). "The commo n and distinct target genes of the p53 family transcription factors." Cell Mol Life Sci 61 (7-8): 822-42. Harper, J. W. (2004). "Neddylating the guardian; Mdm2 catalyzed conjugation of Nedd8 to p53." Cell 118 (1): 2-4. Haupt, Y., R. Maya, et al. (1997). "Mdm2 prom otes the rapid degradation of p53." Nature 387 (6630): 296-9. Hay, R. T. (2005). "SUMO: a hist ory of modification." Mol Cell 18 (1): 1-12. Hecker, C. M., M. Rabiller, et al. (2006). "S pecification of SUMO1and SUMO2-interacting motifs." J Biol Chem 281 (23): 16117-27. Herberg, J. A., S. Beck, et al. (1998). "TAPASIN, DAXX, RGL 2, HKE2 and four new genes (BING 1, 3 to 5) form a dense cluster at the centromeric end of the MHC." J Mol Biol 277 (4): 839-57. Hietakangas, V., J. K. Ahlskog, et al. (2003). "Phosphorylation of se rine 303 is a prerequisite for the stress-inducible SUMO modification of heat shock factor 1." Mol Cell Biol 23 (8): 2953-68. Hilgarth, R. S., L. A. Murphy, et al. (2004). "R egulation and function of SUMO modification." J Biol Chem 279 (52): 53899-902. Hingorani, S. R., L. Wang, et al. (2005). "T rp53R172H and KrasG12D cooperate to promote chromosomal instability and widely metastatic pancreatic ductal adenocarcinoma in mice." Cancer Cell 7 (5): 469-83. Hollenbach, A. D., C. J. McPher son, et al. (2002). "Daxx and hist one deacetylase II associate with chromatin through an interaction with core histones and the chromatin-associated protein Dek." J Cell Sci 115 (Pt 16): 3319-30. Hong, Y., R. Rogers, et al. (2001) "Regulation of heat shock tr anscription factor 1 by stressinduced SUMO-1 modification." J Biol Chem 276 (43): 40263-7. Ishov, A. M., A. G. Sotnikov, et al. (1999). "PML is critical for ND10 formation and recruits the PML-interacting protein daxx to this nuclear structure when m odified by SUMO-1." J Cell Biol 147 (2): 221-34.


99 Ishov, A. M., O. V. Vladimirova, et al. (2004) "Heterochromatin and ND10 are cell-cycle regulated and phosphorylation-dependent alternate nuclear sites of the tr anscription repressor Daxx and SWI/SNF protein ATRX." J Cell Sci 117 (Pt 17): 3807-20. Itahana, Y., E. T. Yeh, et al. (2006). "Nucle ocytoplasmic shuttling modulates activity and ubiquitination-dependent tu rnover of SUMO-specific pr otease 2." Mol Cell Biol 26 (12): 467589. Jackson, M. W. and S. J. Berberich (2000) "MdmX protects p53 from Mdm2-mediated degradation." Mol Cell Biol 20 (3): 1001-7. Jeffrey, P. D., S. Gorina, et al (1995). "Crystal structure of th e tetramerization domain of the p53 tumor suppressor at 1.7 angstroms." Science 267 (5203): 1498-502. Jensen, K., Shiels, C., et al. (2001). "PML protein isoforms and the RBCC/TRIM motif." Oncogene 20 (49): 7223-33. Johnson, E. S., I. Schwienhorst, et al. (1997). "The ubiquitin-like protein Smt3p is activated for conjugation to other proteins by an Aos1p/Uba2p heterodimer." Embo J 16 (18): 5509-19. Jones, S. N., A. E. Roe, et al. (1995). "Rescu e of embryonic lethality in Mdm2-deficient mice by absence of p53." Nature 378 (6553): 206-8. Kahyo, T., T. Nishida, et al. (2001). "Involvement of PIAS1 in the sumoylation of tumor suppressor p53." Mol Cell 8 (3): 713-8. Kamitani, T., K. Kito, et al. ( 1998). "Identification of three major sentrinization sites in PML." J Biol Chem 273 (41): 26675-82. Kern, S. E., K. W. Kinzler, et al. (1991). "Identification of p53 as a sequence-specific DNAbinding protein." Science 252 (5013): 1708-11. Kerscher, O. (2007). "SUMO junction-what's your function? New insights through SUMOinteracting motifs." EMBO Rep 8 (6): 550-5. Kim, E. J., J. S. Park, et al (2003). "Identificatio n of Daxx interacting with p73, one of the p53 family, and its regulation of p53 activity by comp etitive interaction with PML." Nucleic Acids Res 31 (18): 5356-67. Kim, Y. H., K. S. Sung, et al. (2005). "Des umoylation of homeodomain-interacting protein kinase 2 (HIPK2) through the cytoplasmic-nucle ar shuttling of the SUMO-specific protease SENP1." FEBS Lett 579 (27): 6272-8. Kiriakidou, M., D. A. Driscoll, et al. (1997). "Cloning and e xpression of primate Daxx cDNAs and mapping of the human gene to chromosome 6p21.3 in the MHC region." DNA Cell Biol 16 (11): 1289-98.

PAGE 100

100 Knipscheer, P., W. J. van Dijk, et al. (2007). "Noncovalent interaction between Ubc9 and SUMO promotes SUMO chain formation." Embo J 26 (11): 2797-807. Ko, L. J. and C. Prives (1996). p53: puzzle and paradigm." Genes Dev 10 (9): 1054-72. Kubbutat, M. H., S. N. Jones, et al. (1997). "Regulation of p53 stability by Mdm2." Nature 387 (6630): 299-303. Kwek, S. S., J. Derry, et al. ( 2001). "Functional analysis and in tracellular local ization of p53 modified by SUMO-1." Oncogene 20 (20): 2587-99. Lahav, G. (2008). "Oscillations by the p53Mdm2 feedback loop." Adv Exp Med Biol 641 : 2838. Lane, D. P. (1992). "Cancer. p53, guardian of the genome." Nature 358 (6381): 15-6. Lang, G. A., T. Iwakuma, et al. (2004). "Gain of function of a p53 hot spot mutation in a mouse model of Li-Fraumeni syndrome." Cell 119 (6): 861-72. Laurie, N. A., S. L. Donovan, et al. (2006). "In activation of the p53 pathwa y in retinoblastoma." Nature 444 (7115): 61-6. Lee, W., T. S. Harvey, et al. (1994). "Solution structure of the tetrameric minimum transforming domain of p53." Nat Struct Biol 1 (12): 877-90. Lehembre, F., S. Muller, et al (2001). "Regulation of Pax3 tran scriptional activ ity by SUMO-1modified PML." Oncogene 20 (1): 1-9. Leveillard, T., P. Gorry, et al. (1998). "MDM 2 expression during mouse embryogenesis and the requirement of p53." Mech Dev 74 (1-2): 189-93. Li, S. J. and M. Hochstrasser (1999). "A new protease required for cellcycle progression in yeast." Nature 398 (6724): 246-51. Li, T., E. Evdokimov, et al. (2004). "Sumoylation of heterogeneous nucle ar ribonucleoproteins, zinc finger proteins, and nuclear pore complex proteins: a proteo mic analysis." Proc Natl Acad Sci U S A 101 (23): 8551-6. 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, D. Y., Y. S. Huang, et al. (2006). "R ole of SUMO-interacting motif in Daxx SUMO modification, subnuclear localiza tion, and repression of sumoylat ed transcription factors." Mol Cell 24 (3): 341-54.

PAGE 101

101 Lin, D. Y. and H. M. Shih (2002). "Essential ro le of the 58-kDa microspherule protein in the modulation of Daxx-dependent transcriptional repression as revealed by nucleolar sequestration." J Biol Chem 277 (28): 25446-56. Lin, J. Y., T. Ohshima, et al. (2004). "Associat ion of Ubc9, an E2 ligase for SUMO conjugation, with p53 is regulated by phosphor ylation of p53." FEBS Lett 573 (1-3): 15-8. Liu, G. and X. Chen (2006). "Regulation of th e p53 transcriptional ac tivity." J Cell Biochem 97 (3): 448-58. Mahajan, R., C. Delphin, et al. (1997). "A small ubiquitin-related polypeptide involved in targeting RanGAP1 to nuclear por e complex protein RanBP2." Cell 88 (1): 97-107. Marchenko, N. D., A. Zaika, et al. (2000). "Deat h signal-induced localization of p53 protein to mitochondria. A potential role in apoptotic signali ng." J Biol Chem 275 (21): 16202-12. Marine, J. C. and A. G. Jochemsen (2005). "Mdm x as an essential regulator of p53 activity." Biochem Biophys Res Commun 331 (3): 750-60. Matunis, M. J., E. Coutavas, et al. (1996). "A novel ubiquitin-like modification modulates the partitioning of the Ran-GTPase-activating protei n RanGAP1 between the cytosol and the nuclear pore complex." J Cell Biol 135 (6 Pt 1): 1457-70. Mauri, F., L. M. McNamee, et al. (2008). "M odification of Drosophila p53 by SUMO modulates its transactivation and pro-apopt otic functions." J Biol Chem 283 (30): 20848-56. McLure, K. G. and P. W. Lee (1998). "H ow p53 binds DNA as a tetramer." Embo J 17 (12): 3342-50. Melchior, F. (2000). "SUMO--nonclassica l ubiquitin." Annu Rev Cell Dev Biol 16 : 591-626. Melchior, F. and L. Hengst (2000). "Mdm2-SU MO1: is bigger be tter?" Nat Cell Biol 2 (9): E1613. Melchior, F. and L. Hengst ( 2002). "SUMO-1 and p53." Cell Cycle 1 (4): 245-9. Meluh, P. B. and D. Koshland (1995). "Evide nce that the MIF2 gene of Saccharomyces cerevisiae encodes a centromere protein with homology to the mammalian centromere protein CENP-C." Mol Biol Cell 6 (7): 793-807. Meulmeester, E. and A. G. Jochemsen (2008). "p53: a guide to apoptos is." Curr Cancer Drug Targets 8 (2): 87-97. Michaelson, J. S., D. Bader, et al. (1999). "L oss of Daxx, a promiscuously interacting protein, results in extensive apoptosis in early mouse development." Genes Dev 13 (15): 1918-23.

PAGE 102

102 Migliorini, D., E. Lazzerini 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. Minty, A., X. Dumont, et al (2000). "Covalent modificatio n of p73alpha by SUMO-1. Twohybrid screening with p73 identifies novel SUMO-1-interacting proteins and a SUMO-1 interaction motif." J Biol Chem 275 (46): 36316-23. Mossessova, E. and C. D. Lima (2000). "Ulp1SUMO crystal structure and genetic analysis reveal conserved interactions a nd a regulatory element essential for cell growth in yeast." Mol Cell 5 (5): 865-76. Mukhopadhyay, D., F. Ayaydin, et al. (2006). "SUSP1 antagonizes formation of highly SUMO2/3-conjugated sp ecies." J Cell Biol 174 (7): 939-49. Mukhopadhyay, D. and M. Dasso (2007). "Modifica tion in reverse: the SUMO proteases." Trends Biochem Sci 32 (6): 286-95. Muller, S., M. Berger, et al. (2000). "c-J un and p53 activity is modulated by SUMO-1 modification." J Biol Chem 275 (18): 13321-9. Muller, S., A. Ledl, et al. (2004). "SUMO: a regu lator of gene expression and genome integrity." Oncogene 23 (11): 1998-2008. Muromoto, R., K. Sugiyama, et al. (2004). "Phy sical and functional inte ractions between Daxx and DNA methyltransferase 1-associ ated protein, DMAP1." J Immunol 172 (5): 2985-93. Novatchkova, M., R. Budhiraja, et al. (2004) "SUMO conjugation in plants." Planta 220 (1): 1-8. Ogawa, S., J. Lozach, et al. ( 2004). "A nuclear receptor corepre ssor transcriptional checkpoint controlling activator protein 1-de pendent gene networks requir ed for macrophage activation." Proc Natl Acad Sci U S A 101 (40): 14461-6. Oliner, J. D., J. A. Pietenpol, et al. (1993). "O ncoprotein MDM2 conceals the activation domain of tumour suppressor p53." Nature 362 (6423): 857-60. Olive, K. P., D. A. Tuveson, et al. (2004). "Mut ant p53 gain of function in two mouse models of Li-Fraumeni syndrome." Cell 119 (6): 847-60. Olsson, A., C. Manzl, et al. (2007). "How importa nt are post-translational modifications in p53 for selectivity in target-gen e transcription and tumour s uppression?" Cell Death Differ 14 (9): 1561-75.

PAGE 103

103 Ou, Y. H., P. H. Chung, et al. (2005). "p53 Cterminal phosphorylation by CHK1 and CHK2 participates in the regulation of DNA-damage-i nduced C-terminal acetylation." Mol Biol Cell 16 (4): 1684-95. Owerbach, D., E. M. McKay, et al. (2005). "A pr oline-90 residue unique to SUMO-4 prevents maturation and sumoylation." Biochem Biophys Res Commun 337 (2): 517-20. Paajarvi, G., M. Viluksela, et al. (2005) "TCDD activates Mdm2 and attenuates the p53 response to DNA damaging ag ents." Carcinogenesis 26 (1): 201-8. Panse, V. G., D. Kressler, et al. (2006). "F ormation and nuclear export of preribosomes are functionally linked to the small-ubiquitin -related modifier pathway." Traffic 7 (10): 1311-21. Panse, V. G., B. Kuster, et al (2003). "Unconventional tethering of Ulp1 to the transport channel of the nuclear pore complex by karyopherins." Nat Cell Biol 5 (1): 21-7. Parant, J., A. Chavez-Reyes, et al. (2001). "Res cue 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. Petosa, C., G. Schoehn, et al. (2004). "Arc hitecture of CRM1/Exportin1 suggests how cooperativity is achieved during formation of a nuclear export complex." Mol Cell 16 (5): 76175. Poleshko, A., I. Palagin, et al. (2008) "Identification of cellular prot eins that maintain retroviral epigenetic silencing: evidence fo r an antiviral response." J Virol 82 (5): 2313-23. Puto, L. A. and J. C. Reed (2008). "Da xx represses RelB target promoters via DNA methyltransferase recruitment a nd DNA hypermethylation." Genes Dev 22 (8): 998-1010. Rodriguez, M. S., C. Dargemont, et al. (2001) "SUMO-1 conjugation in vivo requires both a consensus modification motif and nuclear targeting." J Biol Chem 276 (16): 12654-9. Rodriguez, M. S., J. M. Desterro, et al (1999). "SUMO-1 modifi cation activates the transcriptional res ponse of p53." Embo J 18 (22): 6455-61. Ronai, Z. (2006). "Balancing Mdm2 a Daxx-HAUSP matter." Nat Cell Biol 8 (8): 790-1. Ryan, K. M., A. C. Phillips, et al. (2001). "R egulation and function of the p53 tumor suppressor protein." Curr Opin Cell Biol 13 (3): 332-7. Ryo, A., A. Hirai, et al. (2007). "A suppressive role of the prol yl isomerase Pin1 in cellular apoptosis mediated by the death-asso ciated protein Daxx." J Biol Chem 282 (50): 36671-81. Ryu, S. W., S. K. Chae, et al (2000). "Interaction of Daxx, a Fa s binding protein, with sentrin and Ubc9." Biochem Biophys Res Commun 279 (1): 6-10.

PAGE 104

104 Saffert, R. T. and R. F. Kalejta (2007). "Human cytomegalovirus gene expression is silenced by Daxx-mediated intrinsic immune defense in model la tent infections establis hed in vitro." J Virol 81 (17): 9109-20. Sansome, C., A. Zaika, et al. (2001). "Hypoxia death stimulus induces translocation of p53 protein to mitochondria. Detection by i mmunofluorescence on whol e cells." FEBS Lett 488 (3): 110-5. Scheffner, M., J. M. Huibregtse, et al. (1993). "The HPV-16 E6 and E6-AP complex functions as a ubiquitin-protein ligase in the ubiquitination of p53." Cell 75 (3): 495-505. Schmidt, D. and S. Muller (2003). "PIAS/SUMO: new partners in transc riptional regulation." Cell Mol Life Sci 60 (12): 2561-74. Shen, T. H., H. K. Lin, et al. (2006). "The m echanisms of PML-nuclear body formation." Mol Cell 24 (3): 331-9. Sheng, Y., V. Saridakis, et al. (2006). "Molecular recognitio n of p53 and MDM2 by USP7/HAUSP." Nat Struct Mol Biol 13 (3): 285-91. Shvarts, A., W. T. Steegenga, et al. (1996). "MDMX: a novel p53-binding protein with some functional properties of MDM2." Embo J 15 (19): 5349-57. Song, J., L. K. Durrin, et al. (2004). "Identifi cation of a SUMO-binding motif that recognizes SUMO-modified proteins." Proc Natl Acad Sci U S A 101 (40): 14373-8. Song, J., Z. Zhang, et al. (2005). "Small ubiquitin-li ke modifier (SUMO) recognition of a SUMO binding motif: a reversal of th e bound orientation." J Biol Chem 280 (48): 40122-9. Song, J. J. and Y. J. Lee (2004). "Tryptophan 621 and serine 667 residue s of Daxx regulate its nuclear export during glucos e deprivation." J Biol Chem 279 (29): 30573-8. Stad, R., Y. F. Ramos, et al. (2000). "H dmx stabilizes Mdm2 and p53." J Biol Chem 275 (36): 28039-44. Stade, K., F. Vogel, et al. (2002). "A lack of SUMO conjugation affects cNLS-dependent nuclear protein import in yeast." J Biol Chem 277 (51): 49554-61. Stommel, J. M., N. D. Marchenko, et al. (1999). "A leucine-rich nuclear export signal in the p53 tetramerization domain: regulation of subcellu lar localization and p53 activity by NES masking." Embo J 18 (6): 1660-72. Tanaka, K., J. Nishide, et al (1999). "Characterization of a fission yeast SUMO-1 homologue, pmt3p, required for multiple nuclear events, including the control of telomere length and chromosome segregation." Mol Cell Biol 19 (12): 8660-72.

PAGE 105

105 Tang, J., L. K. Qu, et al. (2006). "Critical ro le for Daxx in regulating Mdm2." Nat Cell Biol 8 (8): 855-62. Tang, J., S. Wu, et al. (2004). "A novel transc ription regulatory comp lex containing death domain-associated protein and the AT R-X syndrome protein." J Biol Chem 279 (19): 20369-77. Tao, W. and A. J. Levine (1999). "P19(ARF) stabilizes p53 by blocking nucleo-cytoplasmic shuttling of Mdm2." Proc Natl Acad Sci U S A 96 (12): 6937-41. Toledo, F. and G. M. Wahl (2006). "Regulating the p53 pathway: in vitro hypotheses, in vivo veritas." Nat Rev Cancer 6 (12): 909-23. Toledo, F. and G. M. Wahl (2007). "MDM2 and MDM4: p53 regulators as targets in anticancer therapy." Int J Biochem Cell Biol 39 (7-8): 1476-82. Ullman, A. J. and P. Hearing (2008). "Cellu lar proteins PML and Daxx mediate an innate antiviral defense antagonized by the ad enovirus E4 ORF3 protein." J Virol 82 (15): 7325-35. Valin, A. and G. Gill (2007). "Regulation of the dual-function transcription factor Sp3 by SUMO." Biochem Soc Trans 35 (Pt 6): 1393-6. Vallian, S., J. A. Gaken, et al. (1998). "Modulatio n of Fos-mediated AP-1 transcription by the promyelocytic leukemia protein." Oncogene 16 (22): 2843-53. Verger, A., J. Perdomo, et al (2003). "Modification with SUMO A role in transcriptional regulation." EMBO Rep 4 (2): 137-42. Vousden, K. H. and X. Lu (2002). "Live or let die: the cell's response to p53." Nat Rev Cancer 2 (8): 594-604. Watson, I. R. and M. S. Irwin (2006). "Ubiqui tin and ubiquitin-like m odifications of the p53 family." Neoplasia 8 (8): 655-66. Weber, J. D., L. J. Taylor, et al. (1999). "Nuc leolar Arf sequesters Mdm2 and activates p53." Nat Cell Biol 1 (1): 20-6. Xue, Y., R. Gibbons, et al. ( 2003). "The ATRX syndrome protein forms a chromatin-remodeling complex with Daxx and localizes in promyelocytic leukemia nuclear bodies." Proc Natl Acad Sci U S A 100 (19): 10635-40. Yang, X., R. Khosravi-Far, et al (1997). "Daxx, a novel Fas-bindi ng protein that activates JNK and apoptosis." Cell 89 (7): 1067-76. Yeung, P. L., L. Y. Chen, et al (2008). "Daxx contains two nuc lear localization signals and interacts with importin alpha3." J Cell Biochem 103 (2): 456-70.

PAGE 106

106 Zacchi, P., M. Gostissa, et al. (2002). "The prol yl isomerase Pin1 reveals a mechanism to control p53 functions after geno toxic insults." Nature 419 (6909): 853-7. Zhang, H., H. Saitoh, et al. (2002). "Enzymes of the SUMO modification pathway localize to filaments of the nuclear pore complex." Mol Cell Biol 22 (18): 6498-508. Zhao, L. Y., A. L. Colosimo, et al. (2003). "A denovirus E1B 55-kilodalton oncoprotein binds to Daxx and eliminates enhancement of p53-de pendent transcription by Daxx." J Virol 77 (21): 11809-21. Zhao, L. Y., J. Liu, et al. (2004). "Negativ e regulation of p53 func tions by Daxx and the involvement of MDM2." J Biol Chem 279 (48): 50566-79. Zheng, H., H. You, et al. (2002). "The prolyl is omerase Pin1 is a regulator of p53 in genotoxic response." Nature 419 (6909): 849-53. Zhu, Q., J. Yao, et al. (2001). "Mdm2 mutant defective in binding p300 promotes ubiquitination but not degradation of p53: evidence for the role of p300 in integra ting ubiquitination and proteolysis." J Biol Chem 276 (32): 29695-701. Zhu, S., H. Zhang, et al. (2006). "SUMO m odification through rapamycin-mediated heterodimerization reveals a dual role for Ub c9 in targeting RanGAP1 to nuclear pore complexes." Exp Cell Res 312 (7): 1042-9. Zimmerer, J. M., G. B. Lesins ki, et al. (2007). "STAT1-depende nt and STAT1-independent gene expression in murine immune cells following stimulation with interferon-alpha." Cancer Immunol Immunother 56 (11): 1845-52.

PAGE 107

107 BIOGRAPHICAL SKETCH Aleixo Santiago was born in Bombay, Indi a and grew up in Sharjah, United Arab Emirates. He is the son of P earl and Louis Santiago and has tw o siblings, Bosco, and Clayton. Aleixo, or Alex as he is known as, attended St. Mary’s Catholic High school in Dubai, United Arab Emirates, where he eventually graduated in 1995. He later attended the University of London in the United Kingdom where he obtained hi s ‘A’ Levels in Applied Mathematics and Science in 1997. In 1997, he moved to the United St ates where he pursued his Bachelor’s degree in Biotechnology at North Dakota State University in Fargo, North Dakota. It was here that Alex was exposed to science research where he worked for the U.S. Department of Agriculture and the Department of Zoology. It was at the Depa rtment of Zoology that he discovered a novel somatostatin receptor in Rainbow trout and the discovery prompted him to eventually seek a career in research. When Alex graduated from North Dakota State University in 2001, he pursued a teaching career at the institution wher e he taught freshmen chemistry which he found rewarding. In 2002, he pursued his Master’s de gree in Biotechnology at the Pennsylvania State University in State College, Pennsylvania. He graduated in 2003 and di d his internship at GlaxoSmithKline where he worked on the Boniva ™ project and researched the relationship of oxidative stress and Parkinson’s. He then join ed Johnson & Johnson in Philadelphia where he worked on the Remicade™ project and development of new therapeutics fo r patients with breast cancers and psoriasis. In 2004, Alex moved down to the University of Florida where he joined the IDP graduate program. It was here that he joined the lab of Dr. Daiqing Liao where he was working on the impact of SUMO-modification of the tumor suppressor, p53, and novel Daxx SUMO-interacting Motifs and their potential func tions. Alex Santiago graduated in 2009 with a Ph.D. and will continue his postdoctoral work in cancer biology.