POZ Domain Transcription Factor, FBI-1, Represses Transcription of [PDF]

THE JOURNAL OF BIOLOGICAL CHEMISTRY © 2002 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 277, No. 30, Issue of July 26, pp. 26761–26768, 2002 Printed in U.S.A.

POZ Domain Transcription Factor, FBI-1, Represses Transcription of ADH5/FDH by Interacting with the Zinc Finger and Interfering with DNA Binding Activity of Sp1* Received for publication, March 3, 2002, and in revised form, April 15, 2002 Published, JBC Papers in Press, May 9, 2002, DOI 10.1074/jbc.M202078200

Dong-Kee Lee‡, Dongchul Suh§, Howard J. Edenberg¶, and Man-Wook Hur‡储 From the ‡Department of Biochemistry and Molecular Biology, BK21 Project for Medical Sciences, Institute of Genetic Sciences, Yonsei University School of Medicine, 134 ShinChon-Dong, SeoDaeMoon-Ku, Seoul 120-752, Korea, the §Department of Biochemistry, College of Medicine, PoChon CHA University, 222 Yatop-Dong, Boondang-Ku, Seongnam-Si, Kyungki-Do, 463-836 Korea, and the ¶Department of Biochemistry and Molecular Biology, Indiana University School of Medicine, Indianapolis, Indiana 46202

The POZ domain is an evolutionarily conserved protein-protein interaction motif found in many transcription factors, oncogenic proteins, ion channel proteins, and some actin-associated proteins (1, 2). The POZ domain genes, which were first identified in a set of Drosophila and poxvirus, have since been found in organisms ranging from yeast to humans (3, 4). As many as 197 known human proteins, 121 Drosophila proteins, and 135 Caenorhabditis elegans proteins are estimated to contain the POZ domain (5) (on the World Wide Web, see smart.embl-heidelberg.de/smart/). The POZ domain proteins are strongly involved in many critical cellular processes such as development (6, 7), oncogenesis (8, 9), apoptosis (10), ion channel activity (4), and transcription (11–16). The domain has been shown to form homeric and heteromeric associations with other POZ domains (1, 2, 11,

* This work was supported by the Genetic Engineering Research Grant (1998) and the Sundo Researcher Grant (2000) both from the Korea Research Foundation (to M.-W. H.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 储 To whom correspondence should be addressed: Dept. of Biochemistry and Molecular Biology, Yonsei University School of Medicine, 134 ShinChon-Dong, SeoDaeMoon-Ku, Seoul 120-752, Korea. Tel.: 82-2361-5188; Fax: 82-2-312-5041; E-mail: [email protected]. This paper is available on line at http://www.jbc.org

14). Given the large number of POZ domain proteins in the higher eucaryotes, this domain is likely to regulate many more critical biological processes by a combinatorial diversity of complexes. The most striking and common property of the POZ domain transcription factors is their ability to repress transcription via their POZ domains (1, 11–16). This probably underlies many biological processes controlled by these factors. The ability of the domain to interact with other key regulatory proteins such as corepressor proteins and other transcription factors appears to be important for repression (11–13). In particular, the POZ domains of human Plzf (promyelocytic leukemia zinc finger transcription factor) and Bcl-6 (B cell lymphoma transcription factor-6) were shown to interact with SMRT/N-CoR, mSin3A, B-CoR, and histone deacetylase (12, 13, 16). By referring to the molecular mechanisms proposed for the nuclear receptors, “chromatin compaction” by histone deacetylase complex recruited by the POZ domain was suggested to repress transcription (13, 17, 18). Sp1 (specificity protein 1) is a sequence-specific transcription factor that binds GC box and activates a wide range of viral and cellular genes (19, 20). Sp1 belongs to the Kru¨ppel-like C2H2type zinc finger superfamily (20). The protein is composed of several modules, N terminus-inhibitory domain (aa1 1– 82), serine/threonine-rich domains (aa 87–143; aa 243–350), glutamine-rich domains (aa 138 –232; aa 351–500) important in transcription activation, zinc finger DNA binding domain (aa 622–720), and C terminus D domain (aa 721–788) (19, 20). Because almost all genes contain Sp1-binding GC box in their promoters and Sp1 often plays a critical role in the transcription activation of many genes, there must be a mechanism regulating the activity of Sp1. Murata et al. (21) suggested that Sp1 activity may be negatively regulated by the cellular proteins that interact with the amino terminus of Sp1. Also, the zinc finger DNA binding domain of Sp1 was shown to interact with histone deacetylase and protein kinase C␨, and Sp1 activity is regulated by them (22, 23). However, the molecular and biochemical nature of interactions with histone deacetylase and protein kinase C␨ is unknown (22, 23). We cloned and characterized the ADH5/FDH gene (24, 25), which encodes the human class III ␹-alcohol dehydrogenase (ADH; EC 1.1.1.1) that is also the NAD⫹/glutathione-depend1 The abbreviations used are: aa, amino acid(s); ADH, alcohol dehydrogenase; FDH, formaldehyde dehydrogenase; AdML, adenovirus major late promoter; CAT, chloramphenicol acetyltransferase; ChIP, chromatin immunoprecipitation; FBI-1, factor binds to the inducer of short transcript of human immunodeficiency virus-1; GST, glutathione Stransferase; UAS, upstream activator Gal4 binding sequence; EMSA, electrophoretic mobility shift assay(s).

26761

Downloaded from http://www.jbc.org/ by guest on April 18, 2018

The POZ domain is a protein-protein interaction motif that is found in many transcription factors, which are important for development, oncogenesis, apoptosis, and transcription repression. We cloned the POZ domain transcription factor, FBI-1, that recognizes the cis-element (bp ⴚ38 to ⴚ22) located just upstream of the core Sp1 binding sites (bp ⴚ22 to ⴙ22) of the ADH5/FDH minimal promoter (bp ⴚ38 to ⴙ61) in vitro and in vivo, as revealed by electrophoretic mobility shift assay and chromatin immunoprecipitation assay. The ADH5/FDH minimal promoter is potently repressed by the FBI-1. Glutathione S-transferase fusion protein pull-down showed that the POZ domains of FBI-1, Plzf, and Bcl-6 directly interact with the zinc finger DNA binding domain of Sp1. DNase I footprinting assays showed that the interaction prevents binding of Sp1 to the GC boxes of the ADH5/FDH promoter. Gal4-POZ domain fusions targeted proximal to the GC boxes repress transcription of the Gal4 upstream activator sequence-Sp1-adenovirus major late promoter. Our data suggest that POZ domain represses transcription by interacting with Sp1 zinc fingers and by interfering with the DNA binding activity of Sp1.

26762

Molecular Mechanism of Transcription Repression by POZ Domain

EXPERIMENTAL PROCEDURES

Isolation of Transcription Factor FBI-1—Six clones, all encoding FBI-1, were isolated by screening a human spleen Matchmaker威 cDNA library (CLONTECH) using the yeast one-hybrid assay (35). A trimer of the transcription factor binding site (5⬘-ATTGCAAGCCCCCCC-3⬘) located between bp ⫺38 and ⫺22 of the ADH5/FDH gene was used as a bait by cloning into yeast one-hybrid reporter constructs (pHisi, pHisi-1, and pLacZi) (25, 33, 35). Plasmids, Recombinant Proteins, and Antibodies—pCAT-ADH5 reporter plasmids was reported elsewhere (25, 33). FBI-1 and its deleted cDNAs were cloned into pcDNA3.0 (Invitrogen). pG5-Sp1-AdML promoter-luciferase plasmid was prepared by subcloning five copies of Sp1 consensus sequence (sense, 5⬘-CTAGCTCGATCGGGGCGGGGCGAGCG; antisense, 5⬘-CTAGCGCTCGCCCCGCCCCGATCGAG) into pG5-AdML-luciferase plasmid/NheI (Promega). Three Gal4 DBD-POZ domain (aa 1–120 of Plzf and Bcl-6 and aa 1–130 of FBI-1) fusion protein expression plasmids were constructed by subcloning the POZ domain cDNAs into pBind vector (Promega). The GST fusions of various proteins were prepared by cloning cDNAs into the pGEX4T series vectors (Amersham Biosciences) and tagged with FLAG at the C-terminal in the case of various Sp1-derived polypeptides. The aa 1– 621 polypeptide, serine/threonine-rich region (aa 244 –350), Q-rich region (aa 351–500), and zinc fingers of Sp1 (aa 622–720 or aa 622–788) were expressed in GST fusion form and cleaved from the GST portion when needed in in vitro pull-down assays. The His-tagged POZ domain of FBI-1 was prepared by cloning the cDNA sequence (aa 1–130) into pQE30 (Qiagen) and by induction with isopropyl-1-thio-␤-D-galactopyranoside. All of the plasmid constructs were verified by sequencing. The mouse nuclear extract was prepared according to the method developed by Shapiro et al. (36). Sp1 was purchased from Promega. Anti-FLAG and anti-His antibodies were obtained from Eastman Kodak Co. and Qiagen, respectively. Rabbit anti-FBI-1 POZ domain and Sp1 antibodies were produced with purified recombinant GST fusion proteins. Electrophoretic Mobility Shift Assays (EMSA)—FBI-1 polypeptide was prepared by in vitro transcription/translation using mMessage and a mMachine威 kit (Ambion, Austin, TX) and a Flexi威 rabbit reticulocyte translation system (Promega). EMSA was carried out as described previously using in vitro translated FBI-1 and the following 32P-labeled

oligonucleotides (only the top strand is shown): C-1 (bp ⫺38 to ⫹1), 5⬘-CATTGCAAGCCCCCCCCACGCCCCGCCCCCCTCGCTAGG; C-2 (bp ⫺38 to ⫺19), 5⬘-CATTGCAAGCCCCCCCCACG; and B (bp ⫹1 to ⫹22), 5⬘-GCGCTCGCCACGCCCATGCCTC (33). Chromatin Immunoprecipitation (ChIP) Assays—We investigated in vivo molecular interaction between FBI-1 and the cis-element on the ADH5/FDH promoter using ChIP assay kit (Upstate Biotechnology, Inc., Lake Placid, NY) (37, 38). Confluent 1 ⫻ 106 HeLa cells on a 10-mm dish was treated with formaldehyde (final 1%) to cross-link FBI-1 to the human genomic DNA. Cells were washed with cold phosphate-buffered saline and lysed with SDS lysis buffer (1% SDS, 10 mM EDTA, 50 mM Tris-HCl, pH 8.1). The lysate was sonicated to shear DNA to a length between 200 and 1000 bp. The sonicated supernatant was diluted 10-fold with ChIP dilution buffer (0.01% SDS, 1% Triton X-100, 2 mM Tris-HCl, pH 8.1, 150 mM NaCl) and incubated with either antibody against FBI-1 or preimmune serum overnight at 4 °C with rotation. To collect DNA-FBI-1-antibody complex, a salmon sperm DNA/protein A-agarose slurry was added to the mixture. The mixture was incubated for 1 h at 4 °C with rotation and pelleted DNA/protein A agarose complex by brief centrifugation (700 –1000 rpm) at 4 °C. After extensive washing of the pellet with various washing buffers recommended by the manufacturer, the pellet was dissolved with 250 ␮l of elution buffer and spun to remove agarose. Supernatant was treated with 20 ␮l of 5 M NaCl and heated to 65 °C for 4 h to reverse FBI-1-DNA cross-link. After treatment with EDTA and proteinase K, the supernatant was extracted with phenol/chloroform and precipitated with ethanol to recover DNA. For PCR of the ADH5/FDH promoter region using the chromatinimmunoprecipitated genomic DNA, one-tenth of the DNA was PCRamplified using forward primer 5⬘-GGAAGCCTTTCCCGACAAAAA (bp ⫺195 ⫺175 bp) and reverse primer 5⬘-GAGGCATGGGCGTGGCGAGCGC (bp ⫹22 to ⫹1) relative to the transcription start. Ten cycles of PCR at 94 °C for 30 s, 52 °C for 30 s, and 72 °C for 30 s were followed by 25 cycles of amplification at 94 °C for 30 s, 55 °C for 30 s, and 72 °C for 30 s. As a negative control, PCR was performed to amplify the human glucose transporter 2 promoter (GLUT2) with no apparent FBI-1 binding site using a primer set: forward, 5⬘-TCCTAGTGGAACAAAGGTAT (bp ⫹48 to ⫹67); reverse, 5⬘-GGGAGTCCTGTCAATTCCAG (bp ⫹282 to ⫹301) (39). The PCR cycling conditions were the same as described above. Protein-Protein Interaction Assays—To investigate Sp1-POZ domain interactions, mouse liver nuclear extract (250 ␮g) and various GST-POZ domains agarose (2 ␮g) were incubated in binding buffer (25 mM TrisHCl, pH 7.5, 75 mM NaCl, 0.25% Nonidet P-40, 15 mM ␤-mercaptoethanol) for 6 h at 4 °C. After extensive washing, protein complexes were resolved by 10 or 8% SDS-PAGE and visualized by Western blot analysis using anti-Sp1 antibody and an ECL detection kit (Amersham Biosciences). To map the interacting domain of Sp1, 2 ␮g each of various Sp1 polypeptides tagged with FLAG were incubated with various GSTPOZ domains agarose (2 ␮g) in binding buffer overnight at 4 °C and analyzed by Western blotting as described above using anti-FLAG antibody (Kodak). To fine map the Sp1 domain involved in the interaction with the POZ domain, GST-Sp1 polypeptide-agarose (2 ␮g each of aa 622–788, 622–720, and 695–788) was incubated with His-tagged POZ domain polypeptide (aa 1–130) of FBI-1 (10 ␮g) in binding buffer and analyzed as described above using anti-His antibody (Qiagen). Transient Expression Assays—African green monkey kidney cells (CV-1) were cultured on six-well dishes in Dulbecco’s modified Eagle’s medium supplemented with 5% fetal bovine serum (Invitrogen). At 50 – 60% confluence, the cells were transfected with a mixture of 0.6 ␮g of pCAT-ADH5, and expression plasmids encoding FBI-1 (0.4 ␮g) or its deletion mutant forms were introduced into the cells using LipofectAMINE plus (Invitrogen). The cells were cultured for 40 h. The cell extracts representing the same quantity of protein were assayed for reporter CAT activity. Data are presented as relative CAT activity compared with the control and are the average of four independent assays. To analyze the inhibition of Sp1 activity by the Gal4-POZ domain fusions, a mixture of the reporter plasmids, pG5-luciferase or pG5-Sp1AdML-luciferase (0.3 ␮g, Promega, WI), various Gal4 POZ domain fusion expression vectors (0.3 ␮g, pGal4-FBI-POZ or pGal4-Plzf-POZ, pGal4-Bcl6-POZ), and the control vector pCMV-␤-galactosidase (0.1 ␮g) were cotransfected into CV-1 cells and assayed as described above. Cell extracts representing the same ␤-galactosidase activity were assayed for reporter luciferase activity. Data are presented as relative luciferase activity compared with the control and are the average of three independent assays. DNase I Footprinting Assays of the ADH5/FDH Promoter—Sp1 (0.25 footprint unit) was preincubated for 15 min at room temperature with

Downloaded from http://www.jbc.org/ by guest on April 18, 2018

ent formaldehyde dehydrogenase (FDH; EC 1.2.11) (26). The gene is also believed to be the ancestral ADH gene (27). It is important in the oxidation of various alcohols and of formaldehyde in the presence of glutathione (see Refs. 25 and 28 –30 and references therein). The gene is ubiquitously expressed, although to different levels in different cells (28, 29). The ADH5/ FDH promoter is very different from those of other ADH genes, which are expressed in a different tissue-specific pattern. It is very rich in GC base pairs (75% up to ⫺200 bp) and does not contain TATA or CCAAT box and thus has the characteristics of a housekeeping gene (24, 31, 32). Previously, we investigated how the expression of the gene is regulated both at the transcription and translation levels (25, 33, 34). We found that the Sp1 multigene family recognizes the core cis-elements (bp ⫺22 to ⫹22) and regulates transcription of the ADH5/FDH minimal promoter (33). Sp1 is important both in transcription initiation and activation. Sp3 and Sp4 compete with Sp1 for the core cis-elements and repress transcription. DNase I footprinting assays showed that an unknown protein binds to the cis-element (bp ⫺38 to ⫺22 bp) just upstream of the Sp1-binding core elements (bp ⫺22 to ⫹22) of the ADH5/FDH minimal promoter. Recently, we cloned a POZ domain transcription factor, FBI-1 (factor binds to the inducer of short transcript of human immunodeficiency virus-1), recognizing the cis-element (⫺38 to ⫺22 bp). The POZ domain transcription factor potently represses transcription of the ADH5/FDH promoter. We investigated and discovered a novel molecular mechanism of transcription repression by the POZ domain of FBI-1 on the human ADH5/FDH promoter and the Gal4 UAS-Sp1-AdML promoter. The POZ domain interacts with the Sp1 zinc finger DNA binding domain, and represses transcription by interfering with GC box recognition by Sp1.

Molecular Mechanism of Transcription Repression by POZ Domain

26763

Downloaded from http://www.jbc.org/ by guest on April 18, 2018

FIG. 1. A comparison of the translated amino acid sequences of the mammalian POZ domain transcription factors FBI-1 (human), mLrf (murine; NM_010731), and rOCZF (rat; D88450). The hydrophobic N-terminal POZ domains (aa 1–120), four C2H2-type zinc fingers (aa 376 – 489), and the nuclear localization sequence (NLS; CXXVXXRXXRKXXX, aa 490 –503) are absolutely conserved in all three proteins. bovine serum albumin, FBI-1, and the POZ domain of FBI-1 (300 ng of FBI-1 or molar equivalents of the other). 32P-labeled ADH51FDH promoter probe was mixed with a preincubated mixture containing Sp1 and incubated further for 15 min at room temperature. The mixtures were digested with DNase I as previously described (25, 33). RESULTS

Isolation of Human POZ Transcription Factor, FBI-1—In a yeast one-hybrid screen for factors that bind to the cis-acting

element located between bp ⫺38 and ⫺22 of the human ADH5/ FDH promoter, six cDNAs that encode a human POZ domain, Kru¨ ppel-like zinc finger transcription factor were isolated. The cDNA encodes a 584-amino acid polypeptide that is identical to FBI-1 (GenBankTM number NM-015898), a potential human immunodeficiency virus-1 transcription regulator (40). The cDNA is very rich in GC bases (71% in the coding sequence) and forms a CpG island (CpG/GpC ratio ⫽ 0.96). FBI-1 is nearly

26764

Molecular Mechanism of Transcription Repression by POZ Domain

identical to TIP21 (transcription termination factor-I-interacting peptide 21; AF000561), which extends further to the N terminus. FBI-1 is the human ortholog of murine Lrf (leukemia/lymphoma-rrelated factor) and rat OCZF (osteoclast-derived zinc finger protein), with perfect matches in three functionally important domains (Fig. 1): the N-terminal POZ domain (aa 1–130), four Cys2-His2 zinc fingers (aa 381– 493), and a nuclear localization sequence (aa 494 –507; CXXVXXRXXRKXXX) (41, 42). Western blot analysis using anti FBI-1 antibody shows that mLRF, a mouse homologue of FBI-1, is expressed ubiquitously in various mouse tissues (data not shown) and is in line with ubiquitous expression of mLRF mRNA (41). EMSA Shows That FBI-1 Binds to the Elements Close to the Sp1 Binding Core GC Boxes Critical in Transcription Activation—The in vitro translated FBI-1 polypeptide binds specifically to the C-2 probe (bp ⫺38 to ⫺22) used as a one-hybrid bait and the C-1 probe (bp ⫺38 to ⫺1 bp) (Fig. 2, A and B, lanes 1,

3, and 6), and the retarded bands (a) are supershifted by the antibody raised against the POZ domain of FBI-1 (b in Fig. 2B, lanes 2, 4, and 7). The excess cold C-1 probe competes successfully, but the excess cold B probe (bp ⫺2 to ⫹22) of the ADH5/ FDH, Sp1, NFY, and HNF3 consensus oligonucleotide probes does not compete for binding (Fig. 2B, compare lane 8 with lanes 9 –12). The C-2 site occupied by the FBI-1 is immediately upstream of the two strong core cis-acting elements (bp ⫺22 to ⫹22) that bind Sp1 and are critical in the ADH5/FDH transcription (25, 33). ChiP Assays Show That FBI-1 Binds to the ADH5/FDH Promoter Region in Vivo—The interaction between the cisregulatory element of the ADH5/FDH promoter and FBI-1 in vivo was investigated by ChIP assay (37, 38). After crosslinking of the chromatin with protein using formaldehyde, sheared chromatin was immunoprecipitated with purified polyclonal anti-FBI-1 antibody and Sepharose A. PCR amplifications of unsheared input genomic DNA and anti-FBI-1 antibody-mediated immunoprecipitant gave PCR products of the expected size (217 bp), but the same reaction with preimmune serum-mediated precipitant did not give amplified product (Fig. 3, B and C). The data suggested that FBI-1 interacts with the ADH5/FDH promoter in vivo as it does in vitro. ChIP assays with human glucose transporter 2 (GLUT2) promoter (39) with no apparent FBI-1 binding site did not give any amplified product as expected regardless of antibodies used (Fig. 3C).

Downloaded from http://www.jbc.org/ by guest on April 18, 2018

FIG. 2. FBI-1 binds to the ADH5/FDH promoter in vitro. A, the human ADH5/FDH core promoter and location of various EMSA probes. Binding sites for Sp1 and FBI-1 are indicated by half-shaded rectangular ovals. The arrow and ⫹1 indicate the transcription start site (Tsp). B, electrophoretic mobility shift assay. The 32P-labeled C-2 (bp ⫺38 to ⫺19) or C-1 (bp ⫺38 to ⫹1) probe was incubated with in vitro translated FBI-1 extract either in the presence or absence of anti-POZ domain antibody (Ab) and separated by 4% nondenaturing PAGE. The retarded bands (a) can be competed by the unlabeled competitor C-1 (at 200⫻ excess) but not by the unrelated oligonucleotide B (bp ⫺2 to ⫹20) or Sp1, NFY, and HNF3 consensus oligonucleotides. Band a is supershifted by the antibody to band b.

FIG. 3. ChIP assay in human HeLa cells: FBI-1 binds to the ADH5/FDH promoter in vivo. A, the interaction between FBI-1 and promoter regulatory elements was investigated on the two human genes, ADH5/FDH and GLUT2. The arrows indicated the locations of PCR primers relative to the transcription start site. Also shown is the binding site for FBI-1 identified by EMSA. B, ChIP assay of the ADH5/ FDH promoter. C, ChIP assay of the GLUT2 promoter. gDNA, human genomic DNA isolated from HeLa cells. IP-Ab, antibody used in immunoprecipitation. Tsp, transcription start site.

Molecular Mechanism of Transcription Repression by POZ Domain

26765

FIG. 4. FBI-1 represses the human ADH5/FDH promoter, and the POZ domain of FBI-1 is important in transcription repression. A, a schematic diagram of full-length FBI-1 and the deletion mutants. Deletions were made only in the POZ domain. The numbers indicate the amino acid residues of the full-length FBI-1 from the N terminus. NLS, nuclear localization sequence. B, full-length FBI-1 (I) or POZ domain deletion mutant (II–IV) was introduced into CV-1 cells along with the pCAT-ADH5 plasmid. The cell extracts representing the same amount of protein were assayed for reporter CAT activity. FBI-1 potently represses transcription of the pCAT-ADH5 (compare lanes 1 and 2). Deletion of the entire POZ domain results in complete loss of transcription repression. Data are presented as relative CAT activity compared with the control (lane 1) and are the average of four independent assays. Bars, S.D.

FBI-1 Represses Transcription of the Human ADH5/FDH Promoter, and the POZ Domain Is Responsible for Repression—We investigated transcription regulatory function of the FBI-1 on the human ADH5/FDH promoter. The pCAT-ADH5 fusion reporter plasmid and the full-length or deletion mutant versions of FBI-1 expression plasmid were introduced into CV-1 cells and analyzed for reporter CAT gene expression. The full-length FBI-1 (Fig. 4A, construct I) represses transcription of the ADH5/FDH promoter by more than 80% at the tested condition (Fig. 4B, lanes 1 and 2), suggesting that the FBI-1 is a transcription repressor. We further investigated which part of FBI-1 is important in the transcription repression of the ADH5/FDH promoter. The POZ domain was divided into three subdomains and deleted sequentially from N terminus based upon the crystal structure of the POZ domain of Plzf (Fig. 4A) (5). Transient transfection assays with the three mutant versions of FBI-1 showed that deleting the first two FBI-1 subdomains (aa 1–32 and aa 1–74) reduces the transcription by

⬃50 – 60%, compared with 80% repression by the full-length construct (Fig. 4B, lanes 2– 4). However, deleting the entire POZ domain (aa 1–130) eliminates repression (Fig. 4B, lane 5). At least part of the POZ domain is required for repression, and the POZ domain is solely responsible for transcription repression. POZ Domain Interacts with Zinc Fingers of Sp1—The ADH5/ FDH promoter context in terms of the core Sp1 binding sites and FBI-1 binding site (Fig. 2A) and transient transfection data (Fig. 4B) suggested that the POZ domain of FBI-1 may interact with Sp1 and represses the ADH5/FDH promoter by controlling Sp1 activity critical in transcription activation such as DNA binding. Therefore, we investigated the protein-protein interaction between Sp1 and POZ domain proteins by GST fusion protein pull-down of mouse liver nuclear extract or recombinant Sp1 polypeptides and Western blotting. The three different GST POZ domain fusions from FBI-1, Plzf, and Bcl-6 interact with Sp1 of the mouse liver nuclear extract (Fig. 5A). We further investigated which part of Sp1 is interacting with the POZ domains and mapped the interacting domain to the zinc fingers plus D domain of Sp1 (aa 622–788) (Fig. 5A). None of the aa 1– 621 polypeptide, serine/threonine (aa 244 –350) region, and Q-rich region (aa 351–500) interact with the POZ domains. Further characterization using GST-Sp1 polypeptide fusions showed that the aa 622–720 region, precisely the zinc finger region that does not include the C terminus D domain, interacts with the POZ domain of FBI-1 (Fig. 5B). POZ Domain of FBI-1 Binds to the Zinc Fingers of Sp1 and Interferes with GC Box Recognition by Sp1—Our data suggested that transcription repression by the FBI-1 and possibly by the other POZ domain transcription factors may be accomplished by a mechanism that involves direct interactions between the Sp1 zinc finger DNA binding domain and the POZ domain. We suspected that the POZ domain probably affects the Sp1 activity absolutely critical in transcription activation (i.e. GC box recognition). We adopted a DNase I footprinting

Downloaded from http://www.jbc.org/ by guest on April 18, 2018

FIG. 5. POZ domains interact with the Sp1 zinc finger DNA binding domain. A, pull-down of mouse liver nuclear extract and various FLAG-tagged Sp1 polypeptides with the GST-POZ domain fusion polypeptides of FBI-1, Plzf, and Bcl-6. Sp1 polypeptide bands were visualized by Western blotting using anti-Sp1 or anti-FLAG antibody. POZ domains interact with the Sp1 C terminus polypeptide (aa 622– 788) that includes zinc fingers and D domain. B, GST-Sp1 fusion protein pull-down assay. The zinc finger DNA binding domain of Sp1 (aa 622–720), not the D domain (aa 695–788) interacts with His-POZ domain (aa 1–130) of FBI-1. POZ domain polypeptide bands were visualized by Western blotting using anti-His antibody.

26766

Molecular Mechanism of Transcription Repression by POZ Domain

assay to investigate the consequence of molecular interaction between the Sp1 zinc finger and the POZ domain on the human ADH5/FDH promoter. The addition of FBI-1 polypeptide or POZ domain of FBI-1 and 32P-labeled probe to the Sp1 preincubated with bovine serum albumin do not affect binding of Sp1 to the probe (Fig. 6, compare lane 1 with lanes 4 and 5). In contrast, 15-min preincubation of Sp1 with the POZ domain or FBI-1 interferes with Sp1 binding to the GC boxes of the probe (Fig. 6, compare lanes 1, 4, and 5 with lanes 6 and 7). Excess recombinant FBI-1 does not give a footprint at the expected site (bp ⫺38 to ⫺22) (Fig. 6, lanes 4 and 6). Previously, it was shown that the POZ domains of ZID, Ttk, and GAGA act to inhibit the interaction of their associated finger regions with DNA by direct interaction with the fingers (1). Deletion of their POZ domains resulted in strong binding to the target DNA sequences. We found that the POZ domain of FBI-1 also directly interacts with its zinc fingers.2 Therefore, the failure of footprinting by the zinc finger of FBI-1 may be caused by the direct interaction with the POZ domain of FBI-1. Alternatively, the recombinant zinc finger region of the FBI-1 polypeptide might not fold properly due to the lack of zinc ion in the expression medium or post-translational modification. The protein-protein interaction data and the DNase I footprinting assay suggested that the POZ domain may repress transcription of the ADH5/

2

J. A. Lee, D. K. Lee, and M.-W. Hur, manuscript in preparation.

FDH promoter by interacting with the zinc finger DNA binding domain of Sp1 and interfering with the GC box recognition by Sp1 (Figs. 5 and 6). We further examined whether the transcription activation by Sp1 can be repressed by the POZ domains of FBI-1, Plzf, and Bcl-6 in vivo by transient transfection assay. We prepared the pGal4 UAS-Sp1-luciferase reporter plasmid that has five copies of Sp1-binding GC box placed between Gal4 UAS and the AdML minimal promoter (Fig. 7A). The presence of the GC boxes potently activates transcription by the endogenous Sp1 by more than 95-fold in CV-1 cells, suggesting that Sp1 plays a major role in transcription activation as in the ADH5/FDH promoter (Fig. 7B, compare lanes 1 and 2). As expected, the Gal4-POZFBI domain fusion protein represses transcription by more than 75%. Although relatively weak compared with Gal4POZFBI, Gal4-POZPLZF and Gal4-POZBcl6 also repress transcription by 20 –50%, once targeted proximal to the GC boxes. The three lines of data presented here suggest that POZ domain transcription factor represses transcription of the human ADH5/FDH promoter and the pGal4 UAS-Sp1-AdML promoter by interacting with the Sp1 zinc finger and probably by interfering with the GC box recognition of Sp1 both in vitro and in vivo. DISCUSSION

The transcription of a eucaryotic gene is regulated by the combined action of multiple sequence-specific transcription factors, general transcription factors, histone modifiers (histone acetyltransferase and histone deacetylase), cofactors, and mediators that regulate the activities of the transcription factors and the chromatin structure (43– 45).

Downloaded from http://www.jbc.org/ by guest on April 18, 2018

FIG. 6. DNase I footprinting assays of the ADH5/FDH promoter. The FBI-1 and POZ domain of FBI-1 potently affect Sp1 binding onto the GC boxes. Sp1 (0.25 footprint unit) was preincubated for 15 min at room temperature either with bovine serum albumin or the effector POZ domain polypeptides before mixing with the 32P-labeled ADH5/FDH footprinting probe. Probe and Sp1 mixture were incubated further for 15 min and subjected to DNase I digestion. Preincubation with bovine serum albumin did not affect the Sp1 binding activity at the expected sites (lane 2 versus lanes 3 and 4). However, a 15-min preincubation of Sp1 with the effectors (FBI-1 and POZ domain of FBI-1) potently affected Sp1 binding to the GC boxes (shaded ovals). The arrow indicates the transcription start site (Tsp).

FIG. 7. The transcription activation by Sp1 can be repressed by the POZ domains of FBI-1, Plzf, and Bcl-6 in CV-1 cells. A, structures of the pGal4-UAS-AdML luciferase fusion reporter and the pGal4-UAS-Sp1-AdML luciferase fusion reporter plasmids. Five copies of Sp1 binding GC box are placed between Gal4 UAS and AdML minimal promoter. UAS, upstream activator binding sequence for yeast Gal4 transcription activator; GC boxes, Sp1 binding GC-rich sequences. B, the presence of the GC boxes potently activates transcription by endogenous Sp1 by more than 95-fold (compare lanes 1 and 2). Once targeted proximal to the GC boxes by Gal4 fusion system, Gal4-POZFBI domain fusion protein represses transcription by more than 75%. Gal4POZPLZF and Gal4-POZBcl6 also repress transcription by 20 –50%.

Molecular Mechanism of Transcription Repression by POZ Domain

The POZ domain is an evolutionarily conserved protein-protein interaction motif found in many transcription factors, oncogenic proteins, and other biologically important proteins (1, 2). The most striking and common property of the POZ domain transcription factors is transcription repression by the POZ domain, and this probably underlies many of the biological processes controlled by the proteins. The domain’s ability to

interact with other key regulatory proteins such as corepressors, histone deacetylase, and other transcription factors appears to be important for repression (11–16). Intriguingly, our study on the POZ domain transcription factor, FBI-1, revealed a novel molecular mechanism of transcription repression (Fig. 8). The transcription repression involves molecular interaction between the Sp1 zinc finger DNA binding domain and the POZ domain (Figs. 5–7). The interaction interferes with Sp1 recognition of the GC boxes in vitro. Also, transient transfection studies using the human ADH5/ FDH promoter and the artificial promoter pGal4 UAS-Sp1AdML luciferase plasmid showed that the POZ domain of FBI-1 represses transcription by interfering with Sp1 transcription activation in vivo. Previously, we found that the Sp1 multigene family plays a central role in the regulation of constitutive and differential expression of the human ADH5/FDH gene in various tissues (33). Sp1 is important in the transcription initiation and activation of the gene (Fig. 8A). Sp3 or Sp4 competes with Sp1 for the core GC boxes and down-regulates expression (33) (Fig. 8A). Considering that FBI-1 is also expressed ubiquitously in various human and mouse tissues, FBI-1 may also contribute in down-regulating the expression of the ADH5/FDH gene by sitting proximal (at the C-2 site) to the core GC boxes and by interfering with the Sp1 multigene family approaching the core cis-elements (C-1 and B). As a result of such interactions between the POZ domain and zinc finger, occupancy by the Sp1 multigene family is greatly hindered (marked as ⫻) (Fig. 8B). FBI-1 is expressed at higher levels in colon and pancreatic cancer tissues compared with normal tissues (NCBI Serial Analysis of Gene Expression data base, available on the World Wide Web at www.ncbi.nlm.nih.gov/SAGE) (46). Accordingly, FBI-1 may down-regulate expression of the ADH5/FDH in the cancer tissues. The physiological significance of down-regulating the ADH5/FDH expression in cancer tissues is not known yet. We also found that the POZ domain of FBI-1 interacts with other zinc fingers of Sp3, Sp4, and FBI-1 (member of the Kru¨ ppel-like C2H2-type zinc finger family).2 In addition, the Sp1 zinc finger interacts with the POZ domains of Plzf and Bcl-6 in vitro and in vivo (Fig. 4).2 Taken together, the repression mechanism by the POZ domain is probably widespread, given the large number of human POZ domain transcription factors (⬃200) and Kru¨ ppel-like C2H2-type zinc fingers (⬃600 –700). Sp1 binds GC boxes and activates a wide range of viral and cellular genes (19, 20). The molecular mechanism elucidated here shows one of the ways how the activity of constitutive transcription activator Sp1 can be regulated. Although other domains of Sp1 may be important in Sp1 activity regulation, the zinc finger DNA binding domain appears to be the target of regulation of Sp1 activity, as evidenced by protein-protein interactions involving histone deacetylase, protein kinase C␨, and POZ domains (22, 23) (Figs. 5–7). The molecular mechanism elucidated here suggests that the zinc finger, a module previously known to function only in binding to the promoter regulatory cis-element, can harness diverse cellular regulatory information via protein-protein interactions and regulates biological processes by controlling transcription. Acknowledgments—We thank Dr. Robert Tjian for the gift of Sp1 expression vectors and Dr. Christian Seiser for GST-Sp1 expression plasmids. REFERENCES 1. Bardwell, V. J., and Treisman, R. (1994) Genes Dev. 8, 1664 –1677 2. Albagli, O., Dhordain, P., Deweindt, C., Lecocq, G., and Leprince, D. (1995) Cell Growth Differ. 6, 1193–1198 3. Koonin, E. V., Senkevich, T. G., and Chernos, V. I. (1992) Trends Biochem. Sci. 17, 213–214

Downloaded from http://www.jbc.org/ by guest on April 18, 2018

FIG. 8. A hypothetical model for transcriptional regulation by the Sp1 multigene family and FBI-1 at the ADH5/FDH minimal promoter. A, Sp1 is important in the transcription initiation and activation. Sp1 occupying the core B and C-1 elements interacts strongly with TBP and TAF110 (plus sign) and activates transcription. If one or two of the binding sites are occupied by either Sp3 or Sp4, transcription is either repressed moderately or potently (33). If FBI-1 occupies the C-2 element, it interacts with incoming zinc fingers of Sp1 (minus sign) and prevents Sp1 binding to the C-1 and B (marked as ⫻) and represses transcription. B and C-1 are core cis-elements immediately flanking the transcription start point (Tsp, marked by an arrow and ⫹1). C-2, binding site for FBI-1. TBP, TATA-binding protein. TAF110, TBP-associated factor 110. The double arrows indicate the interaction between transcription factors (Sp1, Sp3, and Sp4) and TBP or TAF110. The arrows under transcription indicate the degree of transcriptional activation. ⫹, positive regulatory interaction. ⫺, negative regulatory interaction. ⫻, no occupancy caused by FBI-1. B, general model of transcription repression by FBI-1 and POZ domain transcription factors. A promoter for which transcription is activated mainly by Sp1 binding to the proximal GC boxes is depicted (e.g. ADH5/FDH); an FRE (FBI-1-responsive element) is shown upstream of the two GC boxes. The POZ domain transcription factor sitting proximal to the GC boxes interacts with the approaching Sp1 zinc finger DNA binding domain. The molecular interaction interferes with GC box recognition by Sp1 and eventually blocks the transcription activation by Sp1.

26767

26768

Molecular Mechanism of Transcription Repression by POZ Domain 23. Pal, S., Claffey, K. P., Cohen, H. T., and Mukhopadhyay, D. (1998) J. Biol. Chem. 273, 26277–26280 Hur, M.-W., and Edenberg, H. J. (1992) Gene (Amst.) 121, 305–311 Hur, M.-W., and Edenberg, H. J. (1995) J. Biol. Chem. 270, 9002–9009 Uotila, L., and Koivusalo, M. (1974) J. Biol. Chem. 249, 7653–7663 Jornvall, H., Danielsson, O., Hjelmgvist, L., Persson, B., and Shafgat, J. (1995) Adv. Exp. Med. Biol. 372, 281–294 28. Adinolfi, A., Adinolfi, M., and Hopkinson, D. A. (1984) Ann. Hum. Genet. 48, 1–10 29. Duley, J. A., Harris, O., and Holmes, R. S. (1985) Alcoholism Clin. Exp. Res. 9, 263–271 30. Wagner, F. W., Pares, X., Holmquist, B., and Vallee, B. L. (1984) Biochemistry 23, 2193–2199 31. Dynan, W. S. (1986) Trends Genet. 2, 196 –197 32. Gardiner-Garden, M., and Frommer, M. (1987) J. Mol. Biol. 196, 261–28226 33. Kwon, H. S., Kim, M. S., Edenberg, H. J., and Hur, M.-W. (1999) J. Biol. Chem. 274, 20 –28 34. Kwon, H. S., Lee, D. K., Lee, J. J., Edenberg, H. J., Ahn, Y. h., and Hur, M.-W. (2001) Arch. Biochem. Biophys. 386, 163–171 35. Wang, M. M., and Reed, R. R. (1993) Nature 364, 121–126 36. Shapiro, D. J., Sharp, P. A., Wahli, W. W., and Keller, M. J. (1988) DNA 7, 47–55 37. Braunstein, M., Rose, A. B., Holmes, S. G., Allis, C. D., and Broach, J. R. (1993) Genes Dev. 7, 592– 604 38. Luo, R. X., Postigo, A. A., and Dean, D. C. (1998) Cell 92, 463– 473 39. Cha, J. Y., Kim, H. i., Kim, K. S., Hur, M.-W., and Ahn, Y. H. (2000) J. Biol. Chem. 275, 18358 –18365 40. Morrison, D. J., Pendergrast, P. S., Stavropoulos, P., Colmenares, S. U., Kobayashi, R., and Hernandez, N. (1999) Nucleic Acids Res. 27, 1251–1262 41. Davies, J. M., Hawe, N., Kabarowski, J., Huang, Q. H., Zhu, J., Brand, N. J., Leprince, D., Dhordain, P., Cook, M., Morriss-Kay, G., and Zelent, A. (1999) Oncogene 18, 365–375 42. Kukita, A., Kukita, T., Ouchida, M., Maeda, H., Yatsuki, H., and Kohashi, O. (1999) Blood 94, 1987–1997 43. Dynan, W. S., and Tjian, R. (1985) Nature 316, 774 –778 44. Pazin, M. J., and Kadonaga, J. T. (1997) Cell 89, 325–328 45. Na¨ a¨ r, A. M., Lemon, B. D., and Tjian, R. (2001) Annu. Rev. Biochem. 70, 475–501 46. Velculescu, V. E., Zhang, L., Vogelstein, B., and Kinzler, K. W. (1995) Science 270, 484 – 487 24. 25. 26. 27.

Downloaded from http://www.jbc.org/ by guest on April 18, 2018

4. Aravind, L., and Koonin, E. V. (1999) J. Mol. Biol. 285, 1353–1361 5. Ahmad, K. F., Engel, C. K., and Prive, G. G. (1998) Proc. Natl. Acad. Sci. U. S. A., 95, 12123–12128 6. Farkas, G., Gausz, J., Galloni, M., Reuter, G., Gyurkovics, H., and Karch, F. (1994) Nature 371, 806 – 808 7. Barna, M., Hawe, N., Niswander, L., and Pandolfi, P. P. (2000) Nat. Genet. 25, 166 –172 8. Chen, Z., Brand, N. J., Chen, A., Chen, S. J., Tong, J. H., Wang, Z. Y., Waxman, S., and Zelent, A. (1993) EMBO J. 12, 1161–1167 9. Kerckaert, J. P., Deweindt, C., Tilly, H., Quief, S., Lecocq, G., and Bastard, C. (1993) Nat. Genet. 5, 66 –70 10. Yamochi, T., Kaneita, Y., Akiyama, T., Mori, S., and Moriyama, M. (1999) Oncogene 18, 487– 494 11. Dong, S., Zhu, J., Reid, A., Strutt, P., Guidez, F., Zhong, H. J., Wang, Z. Y., Licht, J., Waxman, S., Chomienne, C., Chen, Z., Zelent, A., and Chen, S.-J. (1996) Proc. Natl. Acad. Sci. U. S. A. 93, 3624 –3629 12. Dhordain, P., Albagli, O., Lin, R. J., Ansieau, S., Quief, S., Leutz, A., Kerckaert, J. P., Evans, R. M., and Leprince, D. (1997) Proc. Natl. Acad. Sci. U. S. A. 94, 10762–10767 13. Lin, R. J., Nagy, L., Inoue, S., Shao, W., Miller, W. H., Jr., and Evans, R. M. (1998) Nature 391, 811– 814 14. Deltour, S., Guerardel, C., and Leprince, D. (1999) Proc. Natl. Acad. Sci. U. S. A. 96, 14831–14836 15. Chang, C. C., Ye, B. H., Chaganti, R. S., and Dalla-Favera, R. (1996) Proc. Natl. Acad. Sci. U. S. A. 93, 6947– 6952 16. Huynh, K. D., Fischle, W., Verdin, E., and Bardwell, V. J. (2000) Genes Dev. 14, 1810 –1823 17. Heinzel, T., Lavinsky, R. M., Mullen, T. M., Soderstrom, M., Laherty, C. D., Torchia, J., Yang, W. M., Brard, G., Ngo, S. D., and Davie, J. R. (1997) Nature 387, 43– 48 18. Nagy, L., Kao, H. Y., Chakravarti, D., Lin, R. J., Hassig, C. A., Ayer, D. E., Schreiber, S. L., and Evans, R. M. (1997) Cell 89, 373–380 19. Courey, A. J., and Tjian, R. (1992) Transcriptional Regulation, 2nd Ed. (McKnight, S. L., and Yamamoto, K. R., eds) pp. 743–769, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY 20. Suske, G. (1999) Gene (Amst.) 238, 291–300 21. Murata, Y., Kim, H. G., Rogers, K. T., Udvadia, A. J., and Horowitz, J. M. (1994) J. Biol. Chem. 269, 20674 –20681 22. Doetzlhoefer, A., Rotheneder, H., Lagger, G., Koranda, M., Kurtev, V., Brosch, G., Wintersberger, E., and Seiser, C. (1999) Mol. Cell. Biol. 19, 5504 –5511

POZ Domain Transcription Factor, FBI-1, Represses Transcription of ADH5/FDH by Interacting with the Zinc Finger and Interfering with DNA Binding Activity of Sp1 Dong-Kee Lee, Dongchul Suh, Howard J. Edenberg and Man-Wook Hur J. Biol. Chem. 2002, 277:26761-26768. doi: 10.1074/jbc.M202078200 originally published online May 9, 2002

Access the most updated version of this article at doi: 10.1074/jbc.M202078200 Alerts: • When this article is cited • When a correction for this article is posted

This article cites 45 references, 18 of which can be accessed free at http://www.jbc.org/content/277/30/26761.full.html#ref-list-1

Downloaded from http://www.jbc.org/ by guest on April 18, 2018

Click here to choose from all of JBC's e-mail alerts