The Proto-Oncoprotein KR-POK Represses Transcriptional Activation of CDKN1A by MIZ-1 Through Competitive Binding

Oncogene (2012) 31, 1442–1458 & 2012 Macmillan Publishers Limited All rights reserved 0950-9232/12 www.nature.com/onc ORIGINAL ARTICLE The proto-oncoprotein KR-POK represses transcriptional activation of CDKN1A by MIZ-1 through competitive binding

KM Lee1, WI Choi1, DI Koh1, YJ Kim1, BN Jeon1, JH Yoon1, CE Lee2, SH Kim2, JOh2 and MW Hur1

1Department of Biochemistry and Molecular Biology, Brain Korea 21 Project for Medical Science, Severance Biomedical Research Institute, Yonsei University School of Medicine, SeoDaeMoon-Ku, Seoul, Korea and 2Department of Biological Science, Sungkyunkwan University, Suwon, Korea

The BTB/POZ family of proteins has been implicated in Introduction multiple biological processes, including tumourigenesis, DNA damage responses and cell cycle progression and The BTB/POZ domain proteins were originally found in development. MIZ-1 (Myc-interacting zinc-finger protein Drosophila melanogaster, bric-a-brac, tramtrack and 1) is known to activate transcription of CDKN1A. broad complex transcription regulators (Koonin et al., We recently found that a kidney cancer-related POK 1992; Aravind and Koonin 1999) and in pox virus zinc- transcription factor, KR-POK, is highly expressed in finger proteins. The BTB/POZ domain proteins have an kidney, brain and bone marrow cancer tissues and is a evolutionarily conserved protein–protein interaction potential proto-oncoprotein. Mouse Kr-pok represses domain at their N-termini, called the BTB/POZ domain transcription of the CDKN1A by acting on the proximal (Bardwell and Treisman, 1994; Albagli et al., 1995). promoter. The BiFC/FRET assay, co-immunoprecipita- These proteins have been identified in organisms ranging tion and glutathione S-transferase-fusion protein pull- from human to yeast. About 40 of the 194 human BTB/ down assay indicate that MIZ-1 and Kr-pok interact POZ domain, regulatory proteins are classified as POK via their POZ domains. Oligoucleotide pull-down assays family proteins. POK proteins are made up of an and chromatin immunoprecipitation assays revealed that N-terminal POZ domain, which has an important role in MIZ-1 binds to the proximal GC-box#3 (bp, À55 to À63) forming homo- or hetero-dimers and interacting with and the MIZ-1-binding elements, MRE-A (bp, À90 to other proteins, and a C-terminal Kru¨ppel-type (C2H2) À64) and MRE-B (bp, À27 to À17). Interestingly, MIZ-1 zinc-finger domain, which recognizes and binds to also binds to the distal p53-binding elements. Kr-pok binds specific DNA sequences. These proteins have been to the proximal GC-box#1 (bp, À95 to À100) and #3 (bp, implicated in many biological processes, and in parti- À55 to À63) relatively strongly. It also shows weak cular, some of the POK proteins were shown to act as binding to the MREs and the distal p53-binding elements. major regulators of apoptosis (Yamochi et al., 1999), Kr-pok competes with MIZ-1 in binding to these elements differentiation, development (Farkas et al., 1994; Barna and represses transcription by inhibiting MIZ-1/p300 et al., 2000), transcription (Chang et al., 1996; Dong recruitment, which decreases the acetylation of histones et al., 1996; Dhordain et al., 1997; Lin et al., 1998; H3 and H4. Our data indicate that Kr-pok stimulates cell Deltour et al., 1999; Huynh et al., 2000; Maeda et al., proliferation by interfering with the function of MIZ-1 in 2005; Jeon et al., 2008; Choi et al., 2008) and CDKN1A gene transcription using a mechanism that is oncogenesis (Chen et al., 1993; Kerckaert et al., 1993; radically different from other MIZ-1-interacting proteins, Maeda et al., 2005). such as B-cell lymphoma 6, c-Myc and Gfi-1. Promyelocytic leukaemia zinc-finger null mice display Oncogene (2012) 31, 1442–1458; doi:10.1038/onc.2011.331; severe defects in limb development and germ stem cell published online 1 August 2011 maintenance (Barna et al., 2000; Costoya et al., 2004). T-helper-inducing POZ/Kru¨ppel-like factor, also known Keywords: MIZ-1; Kr-pok; CDKN1A; p21; competition; as cKrox, has recently been reported as a master cell proliferation regulator of T-cell lineage commitment (He et al., 2010). B-cell lymphoma 6 (Bcl-6), promyelocytic leukaemia zinc finger and hypermethylated in cancer-1 have been implicated in non-Hodgkin’s lymphoma, acute promyelocytic leukaemia and spontaneous malig- Correspondence: Dr MW Hur, Department of Biochemistry and nant tumours, respectively (Chen et al., 1993; Kerckaert Molecular Biology, Brain Korea 21 Project for Medical Science, et al, 1993; Issa et al., 1997). Along with others, we Severance Biomedical Research Institute, Yonsei University recently found that FBI-1 stimulates cell proliferation School of Medicine, 134 ShinChon-Dong, SeoDaeMoon-Ku, Seoul and is a potent proto-oncoprotein (Maeda et al., 2005; 120-752, Korea. E-mail: [email protected] Jeon et al., 2008; Choi et al., 2009). We also showed that Received 23 January 2011; revised 31 May 2011; accepted 22 June 2011; FBI-1 increases FASN gene expression and thereby published online 1 August 2011 provides phospholipid cell membrane components to Kr-pok, transcriptional repressor of CDKN1A KM Lee et al 1443 meet the lipid needs of rapidly proliferating cancer cells In colon cancer cells, inhibition of p21 expression by c- (Choi et al., 2008). Along with others, we have shown Myc-MIZ-1 changes the p53-dependent response from that POK proteins, including FBI-1 (Choi et al., 2009), cell cycle arrest to apoptosis (Seoane et al., 2002). MIZ-1 ZBTB2 (Jeon et al., 2009), ZBTB4 (Weber et al., 2008), also interacts with BCL-6, another POK protein, via ZBTB5 (Koh et al., 2009) and BCL6 (Peukert et al., their POZ-domains, and this interaction is important in 1997) have major influences on the regulation of the p53 the transcriptional repression of CDKN1A and in the pathway. development of Bcl (Phan et al., 2005). Additionally, Among the genes of p53 pathway, the cyclin- Gfi-1 and MIZ-1 form a ternary complex with c-Myc dependent kinase inhibitor p21 is a downstream and together repress CDKN1A. c-Myc has been shown regulator of the ARF-HDM2-p53-p21 pathway and is to induce lymphomagenesis. It has been suggested that a major regulator of cell cycle arrest in mammalian cells the ternary complex of Gfi-1, MIZ-1 and c-Myc is (el-Deiry et al., 1993; Gartel and Radhakrishnan, 2005; important in the regulation of cell cycle and has an Toledo and Wahl, 2006). The p53 pathway has a effect on lymphomagenesis through the transcriptional crucial role in mediating growth arrest when cells are regulation of CDKN1A (Liu et al., 2010). exposed to DNA-damaging agents (el-Deiry et al., 1994; Using BiFC/FRET analysis, we recently found that Ogryzko et al., 1997). The CDKN1A gene, mainly MIZ-1 interacts with the POK protein Kr-pok (also regulated at the transcriptional level, is a transcriptional called ZBTB36, ZNF857C in humans) via its POZ target of the tumour-suppressor p53. Whereas induction domain. KR-POK (kidney cancer-related POZ-domain of p21 by exposure to DNA-damaging agents results in Kru¨ppel zinc-finger protein) is a proto-oncoprotein that G1-, G2- or S-phase arrest, repression of p21 expression is highly expressed in majority of malignant kidney has a variety of outcomes depending on the cellular cancer tissues, and there is a strong correlation between context (Gartel, 2009). p21 suppresses oncogenesis by kidney cancer development and KR-POK expression negatively regulating cell growth, as has been demon- levels. Murine KR-POK, indicated as Kr-pok, stimu- strated in chronically damaged liver and renal epithelial lates cell proliferation and represses the transcription of cells (Willenbring et al., 2008). CDKN1A and by recruiting co-repressor—HDAC com- POK family proteins, and the molecular interac- plexes, causing deacetylation of histones H3 and H4 tions among POK proteins, have potential roles in (unpublished data). Accordingly, molecular interactions cancer development or cell differentiation by regulating between Kr-pok and MIZ-1 can have significant effects CDKN1A gene transcription, as have been elegantly on cell cycle regulation by MIZ-1. shown for MIZ1 and BCL6 in Bcl (Phan et al., 2005; MIZ-1 has an antagonistic function with Kr-pok in Weber et al., 2008; Choi et al., 2009; Jeon et al., 2009; the transcriptional regulation of CDKN1A and other Koh et al., 2009). Although the POK proteins Myc- genes of the p53 pathway. In this study, we investigated interacting zinc-finger protein 1 (MIZ-1), ZBTB4 and whether the two proteins, MIZ-1 and Kr-pok, have BCL6 are known to interact with each other (Phan et al., important roles in regulating the transcription of 2005; Weber et al., 2008), the partnerships and the CDKN1A. Our data suggest that Kr-pok competes with functions of most of the POZ domain proteins are MIZ-1 to bind to the CDKN1A proximal regions, largely unknown. Accordingly, to analyse and under- including MIZ-1-binding sites, and in this way interferes stand the biological functions carried out by POZ- with transcriptional activation by MIZ-1. The binding domain proteins and by molecular interactions among competition between Kr-pok and MIZ-1 on the POZ domains, we are in the process of determining proximal promoter of the CDKN1A is functionally virtually all of the protein interactions of every human significant and has an important role in cell proliferation POZ-domain protein using BiFC/FRET analysis and cancer development. (Kim et al., 2007). MIZ-1 (also called ZBTB17) interacts with c-Myc, ZBTB4, BCL-6 and Gfi-1 and recruits interacting proteins to the CDKN1A promoter (Peukert et al., Results 1997; Phan et al., 2005; Weber et al., 2008; Basu et al., 2009). MIZ-1 was originally isolated from a yeast two- MIZ-1, a tumour-suppressor gene, interacts with Kr-pok hybrid screen by its ability to bind to c-Myc (Peukert To investigate that POK proteins interact with MIZ-1 et al., 1997). MIZ-1 can either repress or activate the and to determine the function of such protein interac- transcription of target genes depending on the interact- tions, we are in the process of investigating all of the ing protein. MIZ-1 was shown to act as a potent possible protein–protein interactions among the POZ- transcriptional activator of CDKN1A. However, domain proteins using the pFPIA BiFC/FRET system the oncoprotein c-Myc, by interacting with MIZ-1, (Kim et al., 2007) (Figure 1a). We found that KR-POK represses the transcription of the genes involved in interact with MIZ-1. Molecular interactions between the cellular differentiation and metabolism, such as the two POZ-domains of the MIZ-1 and KR-POK as- CDK inhibitor genes p15INK4B and CDKN1A (Peukert sembled the YFP protein, which produced yellow et al., 1997; Phan et al., 2005; Weber et al., 2008; Basu fluorescent light, suggesting that the two proteins et al., 2009). By interacting with MIZ-1, c-Myc is interact with MIZ-1 via their POZ-domains in vivo recruited to the CDKN1A promoter and directs cells to (Figure 1b). As reported previously, the POZ-domains determine cell fate, either cell cycle arrest or apoptosis. of MIZ-1 and BCL6 interacts with each other and gave

Oncogene Kr-pok, transcriptional repressor of CDKN1A KM Lee et al 1444

Figure 1 MIZ-1 interacts with Kr-pok in vitro and in vivo.(a) Diagram of the BiFC/FRET POZ–domain interaction assays system. The POZ domains of proteins were cloned to pFPA-YN and pFPA-YC-vectors. Functional YFP can be generated by molecular interaction between the POZ domains of two proteins. (b) BiFC/FRET assays of the two POZ-domains of MIZ-1 and Kr-pok. The POZ-domain of FBI-1 was shown to interact to form a homodimer and serves as a positive control. pFPIA-DEST-YN and -YC serves as a negative control. (c) Co-immunoprecipitation and western blotting assays. Whole cell extracts of the HEK293A cells transfected with either FLAG-Kr-pok or Myc-MIZ-1 expression vector were immunoprecipitated using the anti-FLAG antibody or anti-MIZ-1 antibody, followed by immuno-blotting with the anti-MIZ-1, anti-FLAG anti-BCL6 antibodies. Co-immunoprecipitation of BCL6 and MIZ-1, a positive control of protein interaction between the two POZ-domain proteins. (d) GST-fusion protein pull- down assays. [35S]-methionine labelled MIZ-1 or Kr-pok synthetic polypeptdes were incubated with recombinant GST or GST- POZMIZ-1 or GST-POZKr-pok proteins, pulled-down, separated by 15% sodium dodecyl sulfate–polyacrylamide gel electrophoresis and analysed by autoradiography.

a yellow ﬂuorescent light in this assay system (Phan Kr-pok or MIZ-1 protein showed that the two proteins et al., 2005) (Supplementary Figure 1a). interacts with each other probably via their POZ To conﬁrm the molecular interaction between the two domains (Figure 1d). proteins, we carried out co-immunoprecipitation and glutathione S-transferase (GST)-fusion protein pull- down assays. Co-immunoprecipitation of the HEK293A Kr-pok represses transcriptional activation of the cells transiently transfected with FLAG-Kr-pok expres- CDKN1A by MIZ-1. MIZ-1 has been shown to sion vector and MIZ-1-Myc expression vector, followed activate the transcription of CDKN1A (Seoane et al., by western blot analysis with anti-FLAG antibody and 2002). We investigated the functions of Kr-pok and anti-MIZ-1 antibody, showed that Kr-pok interacts the protein interaction between Kr-pok and MIZ-1 on with MIZ-1 (Figure 1c). Reverse co-immunoprecipita- the transcription of CDKN1A. HEK293A cells were tion of the same extract revealed that MIZ-1 interacts transiently co-transfected with pcDNA3.1-MIZ-1 and/ with Kr-pok (Supplementary Figure 1b). Furthermore, or pcDMA3.1-Kr-pok. Quantitative reverse transcrip- GST-POZMiz-1 or GST-POZKr-pok fusion protein tase–PCR and western blot analysis showed that MIZ-1 pull-down assays of [35S]-methionine labelled synthetic and Kr-pok were overexpressed and that while MIZ-1

Oncogene Kr-pok, transcriptional repressor of CDKN1A KM Lee et al 1445 activated transcription, Kr-pok repressed the trans- Accordingly, to dissect the molecular mechanism, we cription. The transcriptional activation by MIZ-1 is ﬁrst examined which region of the CDKN1A promoter is repressed by Kr-pok (Supplementary Figures 2a and b). important for transcriptional regulation by Kr-pok and Alternatively, a loss of function approach using RNA MIZ-1. Kr-pok repressed the transcription of the four interference directed at Kr-pok mRNA reversed the CDKN1A promoter constructs by >70% in HEK293A transcriptional repression by Kr-pok (Figure 2a). Also cells (Figures 3a and b). Kr-pok could repress transcrip- knock-down of MIZ-1 mRNA by small interfering tion by acting on the proximal promoter sequence RNA (siRNA) against MIZ-1, decreased the transcrip- (bp À133 to þ 30) (Figure 3b). MIZ-1 activated the tion of CDKN1A (Figure 2b). Additional western blot transcription of the four CDKN1A promoter constructs and semi-quantitative reverse transcriptase–PCR analy- by 4.5–6-fold, and the activation was stronger on the sis of endogenous p21 expression after knock-down of À2.3 kb promoter construct, which contained the distal MIZ-1 and/or KR-POK mRNA showed similar results p53-binding element. The assay suggests that MIZ-1 with Figures 2a and b suggesting that the two proteins could activate transcription by acting on the proximal regulate the transcription and expression of CDKN1A promoter element (bp À133 to þ 30) and can probably gene (Figure 2c). further increase transcription by acting on the distal p53-binding element (Figure 3c). Transcription of the Kr-pok represses the transcriptional activation of CDKN1A gene by Kr-pok and MIZ-1 is most likely CDKN1A by MIZ-1 by acting on the proximal regulated by the molecular events on the proximal promoter region promoter, as it is loaded with Sp1-binding GC-boxes Previous studies have shown that MIZ-1 binds to and two MIZ-1-binding elements (MREs) overlapping two proximal regions, MRE-A (bp, À90 to À64) and with GC-boxes. MRE-B (bp, À27 to À17), to activate the transcription Additionally, transient transcription analysis of the of CDKN1A (Seoane et al., 2002). c-Myc was shown to CDKN1A gene with both MIZ-1 and Kr-pok in HEK293A be recruited to those sites under DNA-damaging cells was performed. Kr-pok can repress transcriptional conditions (Seoane et al., 2002). Additionally, Phan activation by MIZ-1 on the proximal promoter element (bp et al. (2005) showed that BCL6 was recruited to the À131 to þ 30), suggesting that molecular events on the proximal promoter region of CDKN1A by interacting short promoter are important in the transcriptional with MIZ-1. regulation by the two regulators (Figure 3d).

Figure 2 Kr-pok represses transcriptional activation of CDKN1A by MIZ-1. (a) Knock-down of endogenous Kr-pok mRNA and quantitative reverse transcriptase (qRT)–PCR assays of CDKN1A mRNA transcription. HEK293A cells were transfected with siKR- POK RNA. The data were normalised with GAPDH mRNA level. *Po0.01. (b) Knock-down of MIZ-1 mRNA and qRT–PCR analysis of CDKN1A mRNA transcription. HEK293A cells were transfected with siMIZ-1 RNA. The data were normalised with GAPDH mRNA level. *Po0.01. (c) Left, western blot analysis. HEK293A cells were transfected with siRNA against Kr-pok and/or MIZ-1 and analysed for endogenous p21 expression using anti-Miz-1 antibody or anti-p21 antibody. Control, GAPDH. Relative band intensities were indicated as above. Right, semi-quantitative RT–PCR analysis of knock-down of KR-POK or MIZ-1 mRNA.

Oncogene Kr-pok, transcriptional repressor of CDKN1A KM Lee et al 1446

Figure 3 Kr-pok represses MIZ-1 transcriptional activation of CDKN1A gene by acting on the proximal region. (a) Diagram of the CDKN1A promoter–luciferase gene fusion reporter constructs differing in the length of 50 upstream regulatory sequence. (b) FLAG- Kr-pok expression plasmid and various reporter plasmids were transiently co-transfected into HEK293A cells and analysed for luciferase activities. *Po0.01. (c) Myc-MIZ-1 expression plasmid and reporter plasmids were transiently co-transfected into HEK293A cells. *Po0.01. (d) FLAG-Kr-pok and/or Myc-MIZ-1 and reporter plasmids were transiently co-transfected into HEK293A cells. Luciferase activity was measured 48 h after transfection and normalised for transfection efﬁciency with co-transfected b-galactosidase activity or protein concentration. Data presented are the average of three independent assays. Bars represent s.d.

Oligonucleotide pull-down assays of Kr-pok and which is critical in the transcription initiation, commu- MIZ-1 binding on the CDKN1A proximal promoter nication with p53, and synergistic transcriptional To understand the molecular mechanisms of Kr-pok activation by p53-Sp1 under DNA-damaging conditions and MIZ-1 action in the transcription of CDKN1A,we (Koutsodontis et al., 2001). investigated their DNA-binding activities at the proximal regulatory elements, such as GC-boxes and MREs, Kr-pok competes with MIZ-1 to bind to the proximal by oligonucleotide pull-down assays using synthetic promoter GC-box, MRE-B and distal p53-binding polypeptides prepared by in vitro transcription and elements of CDKN1A in vitro translation. As reported previously, MIZ-1 binds to Transcriptional repression can be accomplished by MRE-A and MRE-B (Seoane et al., 2002). Additionally, competitive binding at the same DNA element, or we found that MIZ-1 binds to GC-box#3 and GC- inhibition of transcription factor activity by protein– box#5/6. Based on binding intensity, the MREs and protein interaction, as is the case for c-Myc, BCL6 and GC-box#3 might be functionally more important. Gfi-1. MIZ-1 binds to two proximal sites of the Interestingly, MIZ-1 also binds to the distal p53-binding CDKN1A promoter and activates transcription (Seoane elements (p53RE#1 and #2) (Figure 4b). et al., 2002). Kr-pok can repress transcription by acting We carried out a similar experiment with FLAG-Kr- on the proximal promoter region. Although we have pok. Kr-pok binds to the MREs, but only weakly. shown that MIZ-1 and Kr-pok can bind to the proximal Kr-pok binds to all of the GC-boxes and displays a regulatory elements (Figure 4), we also investigated strong preference for GC-box#1 and GC-box#3. whether Kr-pok binds to the proximal elements directly Although weak, Kr-pok also binds to the distal or whether it is recruited to the CDKN1A promoter by p53RE#1 and #2 (Figure 4c). Purified recombinant MIZ-1. Oligonucleotide pull-down assays of whole cell Kr-pok zinc-finger DNA-binding domain does not lysates (prepared from the HEK293A cells transiently binds to the beads or 30 UTR region of CDKN1A, expressing Kr-pok and/or MIZ-1) using the representa- indicating the specificity of Kr-pok binding to the tive MIZ-1 and Kr-pok-binding element (Figure 5a, above regulatory elements (Supplementary Figure 3). GC-box#5/6 plus MRE-B; GC-box#3) showed that Both MIZ-1 and Kr-pok bind to the GC-box#3 Kr-pok binds to the element in the absence or pres-

Oncogene Kr-pok, transcriptional repressor of CDKN1A KM Lee et al 1447

Figure 4 Oligonucleotide pull-down assays. Kr-pok and MIZ-1 bind to the proximal promoter GC boxes, MREs and p53-binding elements. (a) Diagram of the endogenous CDKN1A promoter-exon 1 region. Key regulatory elements such as distal p53-binding elements (p53RE#1, #2), proximal Sp1-binding GC-boxes, and MREs are indicated. (b, c) Oligonucleotide pull-down assays of MIZ-1 binding. Streptavidin agarose beads linked to biotinylated oligonucleotides probes (GC-boxes, MREs, p53REs) were incubated with synthetic MIZ-1 or Kr-pok polypeptide prepared by in vitro transcription and translation. The precipitates were analysed by western blotting using anti-Miz-1 antibody or anti-FLAG antibody. ence of MIZ-1, but Kr-pok binding was signiﬁcantly saw a similar binding competition pattern of MIZ-1 and decreased by the presence of MIZ-1. Additionally, MIZ- Kr-pok with GC-box#3. As in Figure 4, Kr-pok can 1 binds to the element strongly, but MIZ-1 binding was bind to the proximal promoter elements independently, decreased by the presence of Kr-pok (Figure 5b). We and Kr-pok and MIZ-1 compete with each other in

Oncogene Kr-pok, transcriptional repressor of CDKN1A KM Lee et al 1448

Figure 5 Kr-pok competes with MIZ-1 to bind to the Sp1-GC-box#5/6-MRE, Sp1-GC-box#3, and p53RE#1 and #2 elements of CDKN1A.(a) Diagram of the endogenous CDKN1A promoter-Exon 1 region. Key regulatory elements such as distal p53-binding elements (p53RE#1 and #2), proximal Sp1-binding GC-boxes, and MREs are indicated. (b, c, d) Oligo pull-down assays. Whole cell lysates of HEK293A cells transfected with Myc-MIZ-1 and/or FLAG- Kr-pok were incubated with the oligonucleotide probes conjugated to Streptavidin agarose beads. Sp1-GC-box#5/6-MRE (bp, –56 to –17), GC-box#3 (bp, –55 to –63), p53RE#1 (bp, –2267 to –2237 ) and #2 (bp, –1384 to –1354). Proteins bound to the probes were precipitated by centrifugation and analysed by western blotting.

binding to the elements (Figure 5c). Interestingly, and MIZ-1 at the proximal and potentially also at ectopic MIZ-1 increases Sp1 binding, which is decreased distal regulatory regions. Using primer sets designed to by Kr-pok. This could explain the synergistic activation amplify the two distal p53-binding elements (Figure 6a), by MIZ-1 and Sp1 and the antagonistic relationship we analysed the binding competition between the two between MIZ-1 and Kr-pok (Figure 5c). MIZ-1 and factors using chromatin immunoprecipitation (ChIP) Kr-pok bind to the distal p53 elements (Figure 5d). assays in the HEK293A cells that were transiently co-transfected with Myc-MIZ-1 and increasing amounts of FLAG-Kr-pok expression vectors, or vice versa. The Kr-pok competes with MIZ-1 to bind to the distal ChIP assays showed that, with increasing expression of regulatory regions spanning p53-binding elements of FLAG-Kr-pok, the binding of MIZ-1 to the CDKN1A CDKN1A, which increases p300 co-activator recruitment promoter was decreased (Figure 6b). In contrast, with in vitro oligonucleotide pull-down assays indicated increasing expression of MIZ-1, the binding of FLAG- the potential binding competition between Kr-pok Kr-pok on the CDKN1A promoter was decreased

Oncogene Kr-pok, transcriptional repressor of CDKN1A KM Lee et al 1449

Figure 6 MIZ-1 and Kr-pok bind to the distal p53REs. Kr-pok competes with MIZ-1 to bind to the distal p53RE#1 and #2 elements in vivo and decrease p300 binding in the region. MIZ-1 increases p53 binding but Kr-pok decreases p53 binding. (a) Diagram of the endogenous CDKN1A promoter-exon-30-UTR. Key regulatory elements such as distal p53-binding elements (p53RE#1 and #2), proximal Sp1-binding GC-boxes, exons and 30UTR are indicated. quantitative ChIP (qChIP) oligo primers spanning two distal p53- binding elements are indicated under the CDKN1A gene structure. (b, c, e, f) qChIP assay. HEK293A cells were transfected with either a combination of FLAG-Kr-pok and increasing amount of Myc-MIZ-1 expression vector or combination of Myc-MIZ-1 and increasing amount of FLAG-Kr-pok. Chromatins were immunoprecipitated with the antibodies indicated and analysed by quantitative reverse transcriptase (qRT)–PCR. *Po0.01. (d, g) qChIP assay of endogenous p300 binding. HEK293A cells were transfected with FLAG-Kr- pok and Myc-MIZ-1 expression vector as above. Chromatins were immunoprecipitated with anti-p300 antibody and analysed by qRT– PCR. MIZ-1 increased p300 binding while Kr-pok decreased p300 binding at the distal p53-binding elements. *Po0.01. (h) qChIP assay of endogenous p53 binding. HEK293A cells were transfected with FLAG-Kr-pok or/and Myc-MIZ-1 expression vector as above. Chromatins were immunoprecipitated with anti-p53 antibody and analysed by qRT–PCR. *Po0.01.

Oncogene Kr-pok, transcriptional repressor of CDKN1A KM Lee et al 1450 (Figure 6c).The results suggest that Kr-pok competes HEK293A cells that were transiently co-transfected with with MIZ-1 to bind the distal regulatory region, Myc-MIZ-1 and increasing amounts of FLAG-Kr-pok, including p53RE#1. We observed a similar binding or vice versa using the primers as indicated in Figure 7a. competition between MIZ-1 and Kr-pok at the distal At the proximal promoter, with increasing expression p53-binding element (p53RE#2) (Figures 6e and f). of FLAG-Kr-pok, MIZ-1 binding was decreased. The MIZ-1-p300 complex has an important role in In contrast, with increasing expression of MIZ-1, the activating the transcription of MIZ-1 target genes, such binding of FLAG-Kr-pok was decreased (Figures 7b as p15INK4b and CDKN1A, through MIZ-1-mediated and c). recruitment of the co-activator p300 (Piluso et al., 2002). The transcriptional co-activator p300 is targeted to As Kr-pok can repress transcription activation by regulatory elements or promoters by interactions with MIZ-1, we investigated whether Kr-pok could inhibit transcription factors, and acetylates nearby histones to transcription by decreasing p300 recruitment to the open up nucleosome structure to activate transcription CDKN1A promoter. ChIP assay was performed using (Ogryzko et al., 1996). As is the case with the distal HEK293A cells that were co-transfected with a promoter elements containing p53RE#1 and RE#2, combination of pcDNA3.1-MIZ-1 and an increasing there is a binding competition between Kr-pok and amount of FLAG-Kr-pok or pcDNA3.1-Kr-pok and an MIZ-1 at the proximal promoter. As Kr-pok binding increasing amount of MIZ-1 expression vector. to the proximal promoter increases, p300 binding Although MIZ-1 increases p300 binding in a dose- is decreased because of the displacement of the dependent manner, increased Kr-pok binding to the MIZ-1-p300 complex (Figure 7d). ChIP assays of the CDKN1A promoter decreased p300 binding (Figures 6d HEK293A cells transiently transfected with pcDNA3.1- and g). This indicates that Kr-pok could displace both MIZ-1 and/or FLAG-Kr-pok showed that, at the MIZ-1 and p300 at the CDKN1A promoter by compet- proximal promoter, while MIZ-1 increased acetylation ing with MIZ-1 for promotor-binding sites. of histone H3 and H4, Kr-pok decreased the acetylation As Kr-pok and MIZ-1 can bind to the distal p53RE#1 status. When Kr-pok and MIZ-1 were co-expressed, the and #2, we tested whether Kr-pok and MIZ-1 affect p53 increased histone acetylation of histones H3 and H4 by binding by ChIP assays. MIZ-1 alone increases binding the MIZ-1/p300 complex was decreased by Kr-pok of endogenous p53 to the region, while Kr-pok (Figure 7e). decreases p53 binding enhanced by MIZ-1 (Figure 6h). The oligonucleotide pull-down or ChIP assays Accordingly, we investigated whether Kr-pok, p53, and suggested Kr-pok compete with MIZ-1 and may also MIZ-1 can affect the transcription of CDKN1A with Sp1 (Figures 4b and c; Figures 5b and c). promoter constructs including p53-binding sites (Sup- Accordingly, we investigated whether Kr-pok can affect plementary Figure 4a). MIZ-1 and p53 activate tran- the transcription of minimal promoter constructs with scription independently and the two transcription the Sp1 binding GC-box. Interestingly, while MIZ-1 and factors activate transcription of CDKN1A promoter Sp1 can independently or synergistically activate the synergistically. In contrast, Kr-pok represses transcrip- transcription of pG5-5x(GC)-Luc, Kr-pok represses tional activation by either MIZ-1 or p53 (Supplemen- transcriptional activation by Sp1 (Supplementary tary Figure 4b). Co-immunoprecipitation shows that Figure 5). MIZ-1 and p53 interact with each other and synergistic transcription activation by the two proteins may be possible by the interaction (Supplementary Figure 4c). Kr-pok reverses MIZ-1 activity on cell cycle regulation As Kr-pok could repress the transcriptional activation of CDKN1A by competing with MIZ-1, we Kr-pok competes with MIZ-1 to bind to the proximal examined whether Kr-pok represses MIZ-1-mediated promoter region of the CDKN1A, which decreases cell cycle regulation using ﬂuorescence-activated cell p300 co-activator recruitment and histone acetylation sorting analysis of HEK293A cells transfected with In vitro oligonucleotide-binding assays indicated binding MIZ-1 and/or KR-POK siRNA. Knockdown of MIZ-1 competition between Kr-pok and MIZ-1 at the key increased the S-phase cell population from 13.6 to regulatory elements of the CDKN1A proximal promo- 18.2%. Knockdown of KR-POK decreased the S-phase ter. Using ChIP assays, we tried to demonstrate the cell population from 13.6 to 7.9%. Knockdown of both binding competition between Kr-pok and MIZ-1in the KR-POK and MIZ-1 resulted in a cell population in

Figure 7 Kr-pok competes with MIZ-1 to bind to the CDKN1A proximal promoter elements and decreases p300 binding and histone acetylation in the region. (a) Diagram of the endogenous CDKN1A proximal promoter-30 UTR region. Arrows, ChIP oligo primers spanning the proximal promoter. Key regulatory elements such as proximal Sp1-binding GC-boxes, and transcription start point (Tsp, þ 1) are indicated. (b, c) Quantitative ChIP (qChIP) assay of Kr-pok and MIZ-1 binding. HEK293A cells were transfected with either a combination of FLAG-Kr-pok and increasing amount of Myc-MIZ-1 expression vector or combination of Myc-MIZ-1 and increasing amount of FLAG-Kr-pok. Chromatins were immunoprecipitated with the indicated antibodies and analysed by quantitative reverse transcriptase (qRT)–PCR. Negative control, 30-UTR. *Po0.01; **Po0.1. (d, e) qChIP assay of p300 and acetylated histones H3 and H4. HEK293A cells transfected with FLAG-Kr-pok and Myc-MIZ-1 expression vector as above. Chromatins were immunoprecipitated with antibodies against p300, Ac-H3 and Ac-H4 and measured by qRT–PCR. MIZ-1 increases p300 binding and acetylated histones while Kr-pok decrease p300 complex binding and acetylated histones at the proximal promoter. Negative control, IgG or 30-UTR. *Po0.01.

Oncogene Kr-pok, transcriptional repressor of CDKN1A KM Lee et al 1451 which 12.1% of the cells were in S phase (Figure 8a). whether Kr-pok can promote cell proliferation. Although Additionally, we used 3-(4,5-dimethylthiazol-2-yl)-2,5- MIZ-1 decreased the cell proliferation rate, Kr-pok diphenyl tetrazolium bromide (MTT) assays to examine stimulated cell proliferation. The MIZ-1 inhibition of

Oncogene Kr-pok, transcriptional repressor of CDKN1A KM Lee et al 1452 cell proliferation was decreased when Kr-pok was co- As MIZ-1 induces growth arrest as reported and expressed (Figure 8b). activates transcription of other genes of the p53 path- We also examined whether the cell growth arrest way such as TP53 and ARF (unpublished data), we induced by MIZ-1 can be reversed by Kr-pok in investigated whether MIZ-1 showed apoptotic activity HeLa cells with a foci-formation assay. HeLa cells were and also whether the apoptosis induced by MIZ-1 can transfected with MIZ-1 and/or Kr-pok expression vectors be reversed by Kr-pok, by fluorescence-activated cell and allowed to grow in the presence of G418. Although sorting analysis using Annexin-V staining. MIZ-1 MIZ-1 decreased the number of HeLa cell foci by growth increased the apoptotic cell population significantly arrest and/or apoptosis, ectopic Kr-pok expression (from 1.7 to 14.8%), and the apoptotic cell population increased number of large transformed foci, suggesting was significantly decreased by co-expression of Kr-pok that Kr-pok caused cellular transformation into fast (from 14.8 to 3.1%) (Figure 8d). These data suggest that growing cells. The HeLa cells co-transfected with both functions of MIZ-1 on cell cycle regulation could be MIZ-1 and Kr-pok expression vectors showed more suppressed by Kr-pok. transformed colonies than the HeLa cells transfected with only the MIZ-1 expression vector. These data suggest KR-POK expression increases in human tumours that the growth arrest and/or apoptosis effect of MIZ-1 is In 68% of malignant kidney cancer tissues, KR-POK significantly reversed by Kr-pok (Figure 8c). expression is significantly higher than in normal tissues

Figure 8 Kr-pok inhibits MIZ-1 activities on cell growth arrest and apoptosis. (a) Fluorescence-activated cell sorting (FACS) analysis of a cell cycle progression. HEK293A cells were transfected with siRNA against MIZ-1 and KR-POK mRNA or negative control, cultured and stained with propidium iodide (PI). Cell proliferation was measured by FACS. N.C., scrambled negative control siRNA. (b) MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide) assay of HEK293A cells transfected with MIZ-1 and/or Kr- pok expression vector and cultured for 0, 2, 4 or 6 days. All assays were performed in triplicate. Error bars are included but too tight to see. Po0.001. (c) Colony foci formation assays. HeLa cells transfected with the MIZ-1 and/or Kr-pok expression vector were cultured in G418 containing medium and stained with 0.1% crystal violet. (d) FACS analysis of apoptosis. Kr-pok inhibits apoptosis induced by MIZ-1. HEK293 cells were transfected with Kr-pok and/or MIZ-1 expression vector and stained with Annexin V 48 h after transfection. X axis, Annexin-V–ﬂuorescein isothiocyanate staining; Y axis, PI staining. The number represents the percentage of apoptotic cells in each condition (right quadrants).

Oncogene Kr-pok, transcriptional repressor of CDKN1A KM Lee et al 1453 (data not published). In addition, data obtained from a and is a potential proto-oncoprotein that stimulates publicly available database of tumour expression cell proliferation (in preparation). Kr-pok is a potent proﬁles show that KR-POK expression is relatively high regulator of gene expression of the p53 pathway (in in other cancer types (1.5B3.6-fold in glioblastoma; 2.2- preparation). However, how Kr-pok represses the fold in smouldering myeloma) compared with control transcription of CDKN1A and stimulates cell prolifera- tissues, while the expression level of MIZ-1 is either tion was previously unknown. decreased or remains unchanged (0BÀ1.8-fold in MIZ-1 is known as an activator of CDKN1A gene by glioblastoma; no change in smouldering myeloma) acting on the proximal promoter. DNase I footprinting relative to control tissues (www.oncomine.org) (Figures showed that MIZ-1 binds the two regulatory elements 9a–c). Our biochemical data on KR-POK and the high (MRE-A, bp À90 to À64 and MRE-B, bp À27 to À17) correlation between KR-POK expression level and of the CDKN1A proximal promoter that overlap with tumourigenesis suggest that KR-POK is a proto- the Sp1-binding GC-boxes (#2 and #5/6) (Seoane et al., oncogene. The changes in expression in tumour tissues 2002). Interestingly, our oligonucleotide pull-down suggest that abundant KR-POK binds to the CDKN1A assays showed that MIZ-1 binds to additional regula- promoter more effectively than MIZ-1 and so represses tory elements, such as distal p53RE#1, p53RE#2, transcription and stimulates cell proliferation. Interest- GC-box#3, GC-box#4 and GC-box#5/6. In terms of ingly, when immortalised normal HEK293A cells are binding afﬁnity, MRE-A (partially overlapping with challenged with etoposide, the expression of MIZ-1 GC-box#2), MRE-B (overlapping with GC-box#5/6) mRNA is increased and that of KR-POK is decreased as and GC-box#3 may be more important in the transcrip- part of the cellular defence mechanism against genotoxic tional regulation of CDKN1A gene by MIZ-1. Accord- stress (Figure 9d). It appears that in cancer cells, this ingly, any regulatory molecules interacting with MIZ-1 cellular defence mechanism is disrupted; the KR-POK or binding these elements can potentially affect the level is high, while the MIZ-1 level is relatively low. transcriptional activity of MIZ-1 (Figure 10b). We demonstrate that Kr-pok can repress transcription on the reporter construct with short promoter elements (bp, À133 to þ 30) containing two proximal Discussion MREs and Sp1-binding GC-box elements. Interestingly, unlike other ZBTB proteins investigated so far (ZBTB2, Kr-pok is a member of the POK family of transcription 5, ZBTB7A, and so on), Kr-pok can bind to all factors. Kr-pok is overexpressed in kidney cancer tissues of the GC-box elements and MREs, although it shows

Figure 9 KR-POK expression is increased in human tumours. (a–c) KR-POK expression is increased in glioblastoma (1.5B3.6 Â ) and smouldering myeloma (2.2 Â ) compared with control tissue, while MIZ-1 expression is either decreased or remains unchanged in glioblastoma (0BÀ1.8 Â ) and smouldering myeloma (no change) relative to control tissue. The data are obtained from www.oncomine.org (Zhan et al., 2007; Lee et al., 2006; Sun et al., 2006). Data presented as a Box and Whisker graph with error bars representing the 10th and 90th percentiles. (d) KR-POK and MIZ-1 mRNA expression analysis in HEK293A cells treated with etoposide (5 mM)*Po0.01. Contrary to the expression pattern of KR-POK and MIZ-1 in cancer tissues, DNA damage signals decrease KR-POK expression but increase MIZ-1 expression in immortalised normal HEK293A cells.

Oncogene Kr-pok, transcriptional repressor of CDKN1A KM Lee et al 1454 a relatively higher afﬁnity for GC-box#1 and #3 regulation and that binding site competition may be a (Figure 10c). This suggests that Kr-pok binding to GC- molecular mechanism that regulates cell proliferation box#1 and #3 is likely more important in transcriptional in normal and tumour tissues (Figure 9). repression. Both MIZ-1 and Kr-pok bind to GC-box#3, We found that Kr-pok can repress MIZ-1-dependent which is critical not only in basal transcription but also in transcriptional activation of the cell cycle arrest gene communication and synergistic transcriptional activation CDKN1A in HEK293A cells and that, in doing so, it between Sp1 bound at the proximal GC-box#3 and p53 stimulates cell proliferation. Mechanistically, Kr-pok bound on the distal elements for inducible expression of binds to the proximal GC-boxes, particularly GC-box#1 p21 (Koutsodontis et al., 2001). and #3, and also to the MIZ-1-binding sites (MRE-A As MIZ-1 and Kr-pok interact with each other, we and -B). MIZ-1 binds to MRE-A, MRE-B and also initially considered a repression mechanism similar to GC-box#3. Kr-pok represses the transcription of the the molecular mechanisms proposed for lymphomagen- CDKN1A by competing with MIZ-1, displacing esis by MIZ-1-mediated recruitment of BCL6, c-Myc the MIZ-1-p300 complex, and deacetylating histones and Gﬁ-1 (Phan et al., 2005; Basu et al., 2009; Liu et al., at the proximal promoter. Our data showed how MIZ-1 2010). However, the ChIP assays and binding analysis and Kr-pok, as well as the molecular interplay between of the CDKN1A promoter suggested that MIZ-1 and the two factors, could regulate cell proliferation by Kr-pok compete with each other in the transcriptional modulating CDKN1A gene transcription (Figure 10).

Figure 10 Hypothetical model of transcriptional regulation of CDNK1A by p53, MIZ-1 and KR-POK. (a) p53, Sp1 and co-activator in transcriptional activation of the CDKN1A. p53 is induced by genotoxic stress, binds to the distal p53-binding elements and communicates with Sp1 to activate transcription synergistically. Sp1 is responsible for basal transcription. GC-box#3 is critical in both basal transcription and synergistic transcription activation by Sp1-p53 interaction. Solid line with arrowhead (-), transcriptional activation; solid line with double arrowhead (2), protein interaction. (b) Transcriptional activation by MIZ-1 and co-activator. MIZ-1 activates transcription by binding to the proximal elements such as MRE-A, GC-box#3, and GC-box#5/6, and MRE-B. Although mechanistically not shown, MIZ-1 synergistically activates transcription by interacting with Sp1 in the region. (c) Transcriptional repression by Kr-pok and co-repressor in the proximal promoter region. Kr-pok binds primarily to GC-box#1 and #3 and also weakly to other GC boxes and MREs. Kr-pok appears to compete with either MIZ-1 or Sp1 to repress transcription, particularly binding competition at GC-box#3 may be important. (d) Molecular mechanism of transcriptional regulation of CDKN1A by MIZ-1 and Kr-pok. Transcriptional activation by MIZ-1 can be repressed by Kr-pok at the proximal promoter, via binding competition at GC-box#3 and MREs. MIZ-1 can bind to distal p53-binding elements, which can also be competed by Kr-pok. MIZ-1 interacts with p53 and increases p53 binding. The two proteins synergistically activate transcription. KR-POK displaces the MIZ-1/ p300 complex and recruits the co-repressor HDAC complex, which deacetylates histones Ac-H3 and Ac-H4 at the proximal promoter to repress transcription. Thickness of arrows, DNA-binding intensity; MRE, MIZ-1-binding elements; p53RE, p53-binding element; Tsp ( þ 1), transcription start point; ZF, zinc-ﬁnger DNA-binding domain; x, transcription repression.

Oncogene Kr-pok, transcriptional repressor of CDKN1A KM Lee et al 1455 Interestingly, we found that MIZ-1 binds to the distal Transcriptional analysis of CDKN1A promoter p53-binding elements. It also interacts with p53 and pGL2-CDKN1A-Luc promoter reporter fusion plasmids increases p53 binding in the region, MIZ-1 and p53 as well as pcDNA3-FLAG-Kr-pok, pcDNA3.1-MIZ-1 and synergistically activate transcription of CDKN1A in pCMV-LacZ in various combinations were transiently trans- HCT116 p53-null cells. MIZ-1 can activate transcription fected into HEK293A cells using Lipofectamine Plus reagent (Invitrogen). After 36–48 h of incubation, cells were harvested alone and it also synergistically activates transcription and analysed for luciferase activity. Reporter activity was with Sp1. Molecular events among Kr-pok, p53, Sp1 and normalised with co-transfected b-galactosidase activity or MIZ-1 in the regulation of CDKN1A transcription can be protein concentration for transfection efficiency. intriguing. When cells are under genotoxic stress condi- To analyse the effect of the Kr-pok and MIZ-1 on the tion, expression of MIZ-1 and p53 is greatly induced and transcriptional activation of CDKN1A by p53, pGL2- molecular interactions among p53, MIZ-1 and Kr-pok at CDKN1A-Luc (À2.3 kb) as well as pcDNA3-FLAG-Kr-pok, the distal p53RE may regulate CDKN1A gene transcrip- pcDNA3.1-MIZ-1, pcDNA-p53 and pCMV-LacZ in various tion. In low p53 expression condition, interactions among combinations were transiently transfected into HCT116 p53 null Kr-pok, Sp1 and MIZ-1 at the proximal promoter are cells using Lipofectamine Plus reagent (Invitrogen). After 36– likely a major determinant of p21 expression. 48 h of incubation, cells were harvested and analysed as above. Kr-pok uses a unique strategy to block the transcription of the negative cell cycle regulatory gene CDKN1A Purification of GST fusion proteins and GST fusion by blocking the transcriptional induction by p53, protein pull-down assays pcDNA3.1-MIZ-1 and pcDNA3.0-FLAG Kr-pok expression MIZ-1 and Sp1 and, more specifically, by binding the plasmids were prepared as described in the plasmid section of key regulatory elements GC-box#3 and two flanking GC material and methods. The plasmids were transformed into boxes #1 and #5/6 that otherwise are bound by MIZ-1 Escherichia coli BL21 (DE3). Recombinant GST-fusion proteins and Sp1. The mechanism we proposed here is different were prepared as reported elsewhere (Choi et al., 2009). from the models proposed for MIZ-1/BCL6, MIZ-1/Gfi-1 Agarose-bound GST or GST fusion protein was washed five and Myc/MIZ-1 for lymphomagenesis (Phan et al.,2005; times with cold HEMG (40 mM HEPES, pH 7.9, 0.2 mM EDTA, Basu et al., 2009; Liu et al., 2010). MIZ-1 was shown to 5mM MgCl2,1.5mM dithiothreitol, 10% glycerol, 100 mM KCl, interact with the transcriptional co-activator p300 with protease inhibitor 1 tablet/50 ml (Roche, Mannheim, Ger- HAT activity to activate transcription by acetylating many)) buffer and collected by centrifugation at 3000 r.p.m at 1 histones H3 and H4 at the proximal promoter (Piluso 4 C for 1 min. Agarose-bound GST protein was incubated with [35S]-methionine-labelled Kr-pok and MIZ-1 synthetic protein in et al., 2002). The binding of Kr-pok displaces the MIZ-1- HEMG buffer overnight at 4 1C with shaking. The [35S]- p300 complex from the promoter, resulting in deacety- methionine-labelled synthetic Kr-pok and MIZ-1 proteins were lated histones and transcriptional repression (Figure 10d). prepared using the TNT Quick-coupled Transcription/Transla- Overall, our studies suggest that Kr-pok is a tion System (Promega, Madison, WI, USA). After the agarose– transcriptional repressor with oncogenic potential and protein complex was washed five times with 1 ml of cold HEMG that, when overexpressed, Kr-pok stimulates cell pro- buffer, the bound proteins were resolved by 15% sodium liferation. As MIZ-1 and Kr-pok MIZ-1 and Kr-pok are dodecyl sulfate–polyacrylamide gel electrophoresis. The gels involved in cell cycle regulation, the cellular context were dried (Hoefer Scientific Instruments, San Francisco, CA, of MIZ-1 and Kr-pok in normal or cancer cells can be USA) and exposed to X-ray film on an image-intensifying important in the regulation of cell proliferation. screen (Kodak, Rochester, NY, USA).

Materials and methods RNA isolation and quantitative real-time PCR and semi-quantitative PCR Total RNA was isolated from the HEK293A cells using Plasmids, antibodies and reagents TRIzol reagent (Invitrogen). Complementary DNAs were CDKN1A-Luc plasmid was kindly provided by Dr Yoshihiro synthesised using 5 mg of total RNA, oligo-dT (10 pmol), and Sowa of the Kyoto Perpetual University of Medicine (Kyoto, Reverse transcriptase II (200 units) in 20 ml using a reverse Japan). The pcDNA3-FLAG-Kr-pok, pcDNA3.1-Kr-pok, transcription kit (Promega). Quantitative reverse transcrip- and pcDNA3.1-MIZ-1 plasmids were prepared by cloning tase–PCR was performed using SYBR Green Master Mix complementary DNA fragments into pcDNA (Invitrogen, (Applied Biosystems, Carlsbad, CA, USA). The following Carlsbad, CA, USA). CDKN1A-Luc –1462 bp, CDKN1A-Luc quantitative PCR oligonucleotide primer sets were used: 864 bp and CDKN1A-Luc 133 bp CDKN1A promoter and À À Kr-pok forward, 50-CCCATCTGCCACAAAGTCATC-30, luciferase gene fusion reporter plasmids were cloned from the Kr-pok reverse, 50-TGGTGCACATGTATGGCTTCTC-30; CDKN1A-Luc plasmid. All plasmid constructs were veriﬁed by MIZ-1 forward, 50-AGACCCACGACACGGACAA-30,MIZ-1 DNA sequencing. Antibodies against p21, glyceraldehyde-3- reverse, 50-CCGTCAGCGATGTGGATCT-30; CDKN1A for- phosphate dehydrogenase (GAPDH), FLAG tag, Ac-H3, ward, 50-AGGGGACAGCAGAGGAAG-30, CDKN1A reverse, and Ac-H4 were purchased from Upstate (Charlottesville, 50-GCGTTTGGAGTGGTAGAAATCTG-30;andGAPDH VA, USA), Chemicon (Temecula, CA, USA), Calbiochem forward, 50-CCCCTTCATTGACCTCAACTAC-30,GAPDH (San Diego, CA, USA) and Santa Cruz Biotechnology (Santa reverse, 50-TCTCGCTCCTGGAAGATGG-30. Cruz, CA, USA). Most of the chemical reagents were purchased from Sigma (St Louis, MO, USA). Western blot analysis Cell cultures Cells were harvested and lysed in radioimmune precipitation HEK293A, HCT116 p53 null and HeLa cells were cultured in assay buffer (50 mM Tris–HCl, pH 8.0, 1% Nonidet P-40, Dulbecco’s modiﬁed Eagle medium (Invitrogen) supplemented 0.25% sodium deoxycholic acid, 150 mM NaCl, 1 mM EDTA, with 10% fetal bovine serum (Invitrogen). complete mini-protease mixture). Cell extracts (50 mg) were

Oncogene Kr-pok, transcriptional repressor of CDKN1A KM Lee et al 1456 separated using 12% sodium dodecyl sulfate–polyacrylamide CCG-30; GC-box#2, 50-GATCTCCCGGGCGGCGCG-30; gel electrophoresis gel electrophoresis, transferred onto Immun- GC-box#3, 50-GATCCGAGCGCGGGTCCCGCCTC-30; Blot polyvinylidene difluoride membranes (Bio-Rad, Hercules, GC-box#4, 50-GATCCTTGAGGCGGGCCCG-30; GC-box#5/6, CA, USA), and blocked with 5% skim milk (BD Biosciences, 50-CCTTGAGGCGGGCCCGGGCGGGGCGGTTGTATA San Jose, CA, USA). Blotted membranes were incubated with TC-30; Miz-1RE-A, 50-GGGCGGCGCGGTGGGCCGAG antibodies against FLAG tag (Sigma), GAPDH (Chemicon), CGCGGG-30; Miz-1RE-B, 50-TATATCAGGGCCGC-30.To p21, MIZ-1 and Myc. The blots were further incubated with collect DNA-bound proteins, the mixtures were incubated anti-mouse or rabbit secondary antibody conjugated with with Streptavidin-agarose beads for 2 h, washed with HKMG horseradish peroxidase (Vector Laboratories, Burlingame, CA, buffer and precipitated by centrifugation. The precipitate was USA). Protein bands were visualised with ECL solution analysed by western blot assay as described above. (PerkinElmer Life Sciences, San Jose, CA, USA). Additionally, we tested whether purified recombinant GST and GST-Kr-pok zinc-finger DNA-binding domain protein Knock-down of endogenous KR-POK and MIZ-1 mRNA can bind to the probes directly. The proteins were incubated by siRNA with 1 mg of biotinylated double-stranded oligonucleotides for siRNA mixtures against KR-POK and MIZ-1 mRNA were 16 h. The sequences of the oligonucleotides are listed as follows designed by and purchased from Dharmacon (Lafayette, CO, (only nucleotide sequences of sense strands are shown): GC- USA). siRNA (200 pmol) was transfected into HEK293A boxes (which described above); 30-UTR, 50-CTCACCTCCTC cells using Lipofectamine RNAiMAX (Invitrogen). After TAAGGTTGGGCAGGGTGAC-30; To collect DNA-bound transfection, the cells were harvested, total RNA was prepared proteins, the mixtures were incubated with Streptavidin- and quantitative reverse transcriptase–PCR analysis of agarose beads for 2 h, washed with HKMG buffer and mRNA and western blot analysis were performed as described precipitated by centrifugation. The precipitate was analysed above. by western blot assay as described above.

Quantitative ChIP assays Annexin-V staining HEK293A cells were transfected with increasing amounts of The Annexin-V-FITC Apoptosis Detection Kit (BD Biosciences) was used to detect apoptosis by flow cytometry. HEK293A cells pcDNA3.0-FLAG-Kr-pok and/or pcDNA3.1-MIZ-1. Kr-pok, p53, and MIZ-1 binding and histone modification at the were transfected with the pcDNA3.1-MIZ-1 and/or pcDNA3.0- endogenous CDKN1A promoter was analysed by the standard FLAG-Kr-pok expression vectors. All cells were harvested quantitative ChIP assay protocol, as described elsewhere (Choi (including detached cells) and processed according to the manufacturer’s instructions. Cells were washed and resuspended et al., 2009). As a negative control for the quantitative ChIP with 1 Â binding buffer at a concentration of 1 Â 106 cells/ml, assays, IgG was used. p53RE#1 binding primers, forward, 5 50-CTGTGGCTCTGATTGGCTTT-30, reverse, 50-GGGTCT and 100 ml of the solution (1 Â 10 cells) was transferred to a 5 ml TTAGAGGTCTCCTGTCT-30; p53RE#2 binding primers, culture tube. In all, 5 ml of Annexin-V–fluorescein isothiocyanate forward, 50-CCACAGCAGAGGAGAAAGAAG-30, reverse, and 5 ml of propidium iodide were added to the solution, 50-GCTGCTCAGAGTCTGGAAATC-30; proximal 50-GC- followed by gentle vortexing of the cells and incubation for boxes ChIP primers forward, 50-GCGCTGGGCAGCCAGG 15 min at room temperature in the dark. In total, 400 mlof1Â AGCCT-30, reverse, 50-TCGTCACCCGCGCACTTAGA-30. binding buffer was added to each tube. Apoptosis of cells was Alternatively, in some quantitative ChIP assays, 30-UTR of analysed by FACS Calibur (BD Biosciences) within 1 h. Data CDKN1A was used as a negative control of protein–DNA were analysed using BD CellQuest Prosoftware (BD Biosciences, interaction. Forward 50-TCCTTCCCATCGCTGTCACA-30 San Jose, CA, USA). and reverse 50-GTCACCCTGCCCAACCTTAG-30 were used. Colony foci formation assay HeLa cells (1 Â 105 cells per well) were transfected with pcDNA3 Co-immunoprecipitation assays or pcDNA3-FLAG-Kr-pok (0.5 mg/ml) and/or pcDNA3.1-MIZ- HEK293A cells were washed, pelleted and resuspended in a 1(0.5mg/ml). Transfected cells were cultured for 2 weeks in the lysis buffer supplemented with protease inhibitors (20 mM presence of G418 (800 mg/ml), and colonies resistant to G418 Tris–HCl, pH7.5, 150 mM NaCl, 10% glycerol, 1% Triton were stained with crystal violet (0.5% in 20% ethanol). X-100). Cell lysate was precleared, and the supernatant was incubated overnight with anti-MIZ-1 or anti-FLAG on a rotating platform at 4 1C, followed by incubation with protein 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide A-sepharose fast flow beads. Beads were collected, washed assays and resuspended in equal volumes of 5 Â sodium dodecyl Confluent HEK293A cells grown on 10-cm culture dishes were transfected with the pcDNA3.1-MIZ-1 and pcDNA3.0- sulfate loading buffer. Immunoprecipitated proteins were 5 separated with 12 % sodium dodecyl sulfate–polyacrylamide FLAG-Kr-pok expression vectors, and the cells (1.5 Â 10 )were gel electrophoresis. The western blot assay was performed as transferred to six-well culture dishes and grown for 0–6 days. 1 described above using appropriate antibodies. At 0, 2, 4 and 6 days, the cells were incubated for 1 h at 37 C with 20 ml per well 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl Oligonucleotide pull-down assays tetrazolium bromide (2 mg/ml). Precipitates were dissolved with 1 ml of dimethylsulfoxide. Cell proliferation was determined HEK293A cells were lysed in HKMG buffer (10 mM HEPES from the conversion of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl pH 7.9, 100 mM KCl, 5 mM MgCl2, 10% glycerol, 1 mM DTT and 0.5% NP-40). Cellular extracts were incubated with tetrazolium bromide to formazan using a SpectraMAX 250 1 mg of biotinylated double-stranded oligonucleotides for 16 h. (Molecular Device Co., CA, USA) at 570 nm. The sequences of the oligonucleotides are listed as follows (only nucleotide sequences of sense strands are shown): Statistical analysis p53RE#1, 50-GTCAGGAACATGTCCCAACATGTTGAGC A w2 test or Fisher’s exact test were used for statistical analyses TC-30; p53RE#2, 50-TAGAGGAAGAAGACTGGGCATGT as appropriates. All statistical tests were two-sided and CTGGGCA-30; GC-box#1, 50-GATCGGGAGGGCGGTC P-values of o0.05 were considered statistically significant.

Oncogene Kr-pok, transcriptional repressor of CDKN1A KM Lee et al 1457 SPSS for Windows version 17.0 (SPSS Inc., Chicago, IL, USA) Acknowledgements was used for all statistical analyses. This work was mainly funded by a Basic Science Research Grant (314-2008-1-E00030 to M-W Hur and C-E Lee), and Conﬂict of interest MRC (R13-2002-054-05002-0) from the National Research Foundation of Korea, and also by a Faculty Research Grant The authors declare no conﬂict of interest. from Yonsei University School of Medicine.

References

Albagli O, Dhordain P, Deweindt C, Lecocq G, Leprince D. (1995). Gartel AL, Radhakrishnan SK. (2005). Lost in transcription: p21 The BTB/POZ domain: a new protein-protein interaction motif repression, mechanisms, and consequences. Cancer Res 65: common to DNA- and actin-binding proteins. Cell Growth Differ 6: 3980–3985. 1193–1198. He X, Park K, Kappes DJ. (2010). The role of ThPOK in control of Aravind L, Koonin EV. (1999). Fold prediction and evolutionary CD4/CD8 lineage commitment. Annu Rev Immunol 28: 295–320. analysis of the POZ domain: structural and evolutionary Huynh KD, Fischle W, Verdin E, Bardwell VJ. (2000). BCoR, a relationship with the potassium channel tetramerization domain. novel corepressor involved in BCL-6 repression. Genes Dev 14: J Mol Biol 285: 1353–1361. 1810–1823. Bardwell VJ, Treisman R. (1994). The POZ domain: a conserved Issa JP, Zehnbauer BA, Kaufmann SH, Biel MA, Baylin SB. (1997). protein-protein interaction motif. Genes Dev 8: 1664–1677. HIC1 hypermethylation is a late event in hematopoieic neoplasms. Barna M, Hawe N, Niswander L, Pandolfi PP. (2000). Plzf regulates Cancer Res 57: 1678–1681. limb and axial skeletal patterning. Nat Genet 25: 166–172. Jeon BN, Choi WI, Yu MY, Yoon AR, Kim MH, Yun CO et al. Basu S, Liu Q, Qiu Y, Dong F. (2009). Gfi-1 represses CDKN2B (2009). ZBTB2, a novel master regulator of the p53 pathway. J Biol encoding p15INK4B through interaction with Miz-1. Proc Natl Acad Chem 284: 17935–17946. Sci USA 106: 1433–1438. Jeon BN, Yoo JY, Choi WI, Lee C E, Yoon H G, Hur MW. (2008). Chang CC, Ye BH, Chaganti RS, Dalla-Favera R. (1996). BCL-6, a Proto-oncogene FBI-1 (Pokemon/Zbtb7A) represses transcription POZ/zinc-finger protein, is a sequence-specific transcriptional of the tumor suppressor Rb gene via binding competition with Sp1 repressor. Proc Natl Acad Sci USA 93: 6947–6952. and recruitment of co-repressors. J Biol Chem 283: 33199–33210. Chen Z, Brand NJ, Chen A, Chen SJ, Tong JH, Wang ZY et al. (1993). Kerckaert JP, Deweindt C, Tilly H, Quief S, Lecocq G, Bastard C. Fusion between a novel kruˆppel-like zinc finger gene and the retinoic (1993). LAZ3, a novel zinc-finger encoding gene, is disrupted by acid receptor-a locus due to a vatiant t(11;17) translocation associated recurring chromosome 3q27 translocations in human lymphomas. with acute promyelocytic leukaemia. EMBO J 12: 1161–1167. Nat Genet 5: 66–70. Choi WI, Jeon BN, Park H, Yoo JY, Kim YS, Koh DI et al. (2008). Kim MH, Roh HE, Lee MN, Hur MW. (2007). New fast BiFC Proto-oncogene FBI-1 (Pokemon) and SREBP-1 synergistically plasmid assay system for in vivo protein-protein interactions. activate transcription of fatty-acid synthase gene (FASN). Cell Physiol Biochem 20: 703–714. J Biol Chem 283: 29341–29354. Koh DI, Choi WI, Jeon BN, Lee CE, Yun CO, Hur MW. (2009). A Choi WI, Jeon BN, Yun CO, Kim PH, Kim SE, Choi KY et al. (2009). novel POK family transcription factor, ZBTB5, represses transcrip- Proto-oncogene FBI-1 represses transcription of p21CIP1 by tion of p21CIP1gene. J Biol Chem 284: 19856–19866. inhibition of transcription activation by p53 and Sp1. J Biol Chem Koonin EV, Senkevich TG, Chernos VI. (1992). A family of DNA 284: 12633–12644. virus genes that consists of fused portions of unrelated cellular Costoya JA, Hobbs RM, Barna M, Cattoretti G, Manova K, genes. Trends Biochem Sci 17: 213–214. Sukhwani M et al. (2004). Essential role of Plzf in maintenance of Koutsodontis G, Tentes I, Papakosta P, Moustakas A, Kardassis D. spermatogonial stem cells. Nat Genet 36: 653–659. (2001). Sp1 plays a critical role in the transcriptional activation of Deltour S, Guerardel C, Leprince D. (1999). Recruitment of SMRT/ the human cyclin-dependent kinase Inhibitor p21WAF1/Cip1 gene by N-CoR-mSin3A-HDAC-repressing complexes is not a general the p53 tumor suppressor protein. J Biol Chem 276: 29116–29125. mechanism for BTB/POZ transcriptional repressors: the case of Lee J, Kotliarova S, Kotliarov Y, Li A, Su Q, Donin NM et al. (2006). HIC-1 and gFBP-B. Proc Natl Acad Sci USA 96: 14831–14836. Tumor stem cells derived from glioblastomas cultured in bFGF and Dhordain P, Albagli O, Lin R J, Ansieau S, Quief S, Leutz A et al. EGF more closely mirror the phenotype and genotype of primary (1997). Corepressor SMRT binds the BTB/POZ repressing domain tumors than do serum-cultured cell lines. Cancer Cell 9: 391–403. of the LAZ3/BCL6 oncoprotein. Proc Natl Acad Sci USA 94: Liu Q, Basu S, Qiu Y, Tang F, Dong F. (2010). A role of Miz-1 in 10762–10767. Gfi-1-mediated transcriptional repression of CDKN1A. Oncogene Dong S, Zhu J, Reid A, Strutt P, Guidez F, Zhong HJ et al. (1996). 29: 2843–2852. Amino-terminal protein-protein interaction motif (POZ-domain) is Lin RJ, Nagy L, Inoue S, Shao W, Miller Jr WH, Evans RM. (1998). responsible for activities of the promyelocytic leukemia zinc finger- Role of the histone deacetylase complex in acute promyelocytic retinoic acid receptor-a fusion protein. Proc Natl Acad Sci USA 93: leukaemia. Nature 391: 811–814. 3624–3629. Maeda T, Hobbs RM, Merghoub T, Guernah I, Zelent A, Cordon- el-Deiry WS, Harper JW, O0Connor PM, Velculescu VE, Canman CE, Cardo C et al. (2005). Role of the proto-oncogene Pokemon in Jackman J et al. (1994). WAF1/CIP1 is induced in p53-mediated G1 cellular transformation and ARF repression. Nature 433: 278–285. arrest and apoptosis. Cancer Res 54: 1169–1174. Ogryzko VV, Schiltz RL, Russanova V, Howard BH, Nakatani Y. el-Deiry WS, Tokino T, Velculescu VE, Levy DB, Parsons R, (1996). The transcriptional coactivators p300 and CBP are histone Trent JM et al. (1993). WAF1, a potential mediator of p53 tumor acetyltransferases. Cell 87: 953–959. suppression. Cell 75: 817–825. Ogryzko VV, Wong P, Howard BH. (1997). WAF1 retards S-phase Farkas G, Gausz J, Galloni M, Reuter G, Gyurkovics H, Karch et al. progression primarily by inhibition of Cyclin-dependent kinases. (1994). The Trithorax-like gene encodes the drosophila Mol Cell Biol 17: 4877–4882. GAGA factor. Nature 371: 806–808. Peukert K, Staller P, Schneider A, Carmichael G, Ha¨nel F, Eilers M. Gartel AL. (2009). p21 WAF1/CIP1 and cancer: a shifting paradigm? (1997). An alternative pathway for gene regulation by Myc. EMBO Biofactors 35: 161–164. J 16: 5672–5686.

Oncogene Kr-pok, transcriptional repressor of CDKN1A KM Lee et al 1458 Phan RT, Saito M, Basso K, Niu H, Dalla-Favera R. (2005). BCL6 Weber A, Marquardt J, Elzi D, Forster N, Starke S, Glaum A et al. interacts with the transcription factor Miz-1 to suppress the cyclin- (2008). Zbtb4 represses transcription of p21CIP1 and controls the dependent kinase inhibitor p21 and cell cycle arrest in germinal cellular response to p53 activation. EMBO J 27: 1563–1574. center B cells. Nat Immunol 10: 1054–1060. Willenbring H, Sharma AD, Vogel A, Lee AY, Rothfuss A, Wang Z Piluso D, Bilan P, Capone JP. (2002). Host cell factor-1 interacts with et al. (2008). Loss of p21 permits carcinogenesis from chronically and antagonizes transactivation by the cell cycle regulatory factor damaged liver and kidney epithelial cells despite unchecked Miz-1. J Biol Chem 277: 46799–46808. apoptosis. Cancer Cell 14: 59–67. Seoane J, Le H, Massague J. (2002). Myc suppression of the p21Cip1 Yamochi T, Kaneita Y, Akiyama T, Mori S, Moriyama M. (1999). cdk inhibitor inﬂuences the outcome of the p53 response to DNA Adenovirus-mediated high expression of BCL-6 in CV-1 cells damage. Nature 419: 729–734. induces apoptotic cell death accompanied by down-regulation of Sun L, Hui AM, Su Q, Vortmeyer A, Kotliarov Y, Pastorino S et al. BCL-2 and BCL-XL. Oncogene 18: 487–494. (2006). Neuronal and glioma-derived stem cell factor induces Zhan F, Barlogie B, Arzoumanian V, Huang Y, Williams DR, Hollmig angiogenesis within the brain. Cancer Cell 9: 287–300. K et al. (2007). Gene-expression signature of benign monoclonal Toledo F, Wahl GM. (2006). Regulating the p53 pathway: in vitro gammopathy evident in multiple myeloma is linked to good hypotheses, in vivo veritas. Nat Rev Cancer 6: 909–923. prognosis. Blood 109: 1692–1700.

Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc)

Oncogene