HMGB1 and HMGB2 Cell-specifically Down-regulate the p53- and p73-dependent Sequence-specific Transactivation from the Human Bax Gene Promoter
2002; Elsevier BV; Volume: 277; Issue: 9 Linguagem: Inglês
10.1074/jbc.m110233200
ISSN1083-351X
AutoresMichal Štros, Toshinori Ozaki, Alena Bačı́ková, Hajime Kageyama, Akira Nakagawara,
Tópico(s)RNA modifications and cancer
ResumoThe recently cloned genep73 is a close homologue of p53, which is a crucial tumor suppressor gene for preventing the malignant transformation of cells by inducing cell cycle arrest and apoptosis. Previous reports have shown that architectural DNA-bending/looping chromosomal proteins HMGB1 and HMGB2 (formerly known as HMG1 and HMG2), which function in a number of biological processes including transcription and DNA repair, interact in vitro with p53 and stimulate p53 binding to DNA containing p53 consensus sites. Here, we report that HMGB1 physically interacts with two splicing variants of p73, α and β (pull-down assay), and enhances binding of p73 to specific cognate DNA sites (gel-shift assay). Both HMG box domains of HMGB1, A and B, interact with p73α. Association of HMGB1 with p73, like the demonstrated ability of HMGB1 to stimulate p73 binding to different p53-responsive elements, requires the oligomerization region and/or region between DNA-binding domain and oligomerization domain of p73 (residues 312–381). Transient transfections revealed that ectopically expressed or endogenous HMGB1 and HMGB2 (antisense strategy) significantly inhibit in vivo both p73α/β- and p53-dependent transactivation from the Baxgene promoter (and much less from Mdm2 andp21waf1 promoters) in p53-deficient SAOS-2 cells. In contrast, HMGB1 and HGMB2 stimulate p73- or p53-dependent transactivation in p53-deficient H1299 cells, irrespective of the promoter used. Our results suggest that ubiquitously expressed HMGB1 and HMGB2 have potential tocell- and promoter-specifically down- or up-regulate in vivo transcriptional activity of different members of the p53 family. A possible mechanism of HMGB1-mediated modulation of p73- and p53-dependent transactivation is discussed. The recently cloned genep73 is a close homologue of p53, which is a crucial tumor suppressor gene for preventing the malignant transformation of cells by inducing cell cycle arrest and apoptosis. Previous reports have shown that architectural DNA-bending/looping chromosomal proteins HMGB1 and HMGB2 (formerly known as HMG1 and HMG2), which function in a number of biological processes including transcription and DNA repair, interact in vitro with p53 and stimulate p53 binding to DNA containing p53 consensus sites. Here, we report that HMGB1 physically interacts with two splicing variants of p73, α and β (pull-down assay), and enhances binding of p73 to specific cognate DNA sites (gel-shift assay). Both HMG box domains of HMGB1, A and B, interact with p73α. Association of HMGB1 with p73, like the demonstrated ability of HMGB1 to stimulate p73 binding to different p53-responsive elements, requires the oligomerization region and/or region between DNA-binding domain and oligomerization domain of p73 (residues 312–381). Transient transfections revealed that ectopically expressed or endogenous HMGB1 and HMGB2 (antisense strategy) significantly inhibit in vivo both p73α/β- and p53-dependent transactivation from the Baxgene promoter (and much less from Mdm2 andp21waf1 promoters) in p53-deficient SAOS-2 cells. In contrast, HMGB1 and HGMB2 stimulate p73- or p53-dependent transactivation in p53-deficient H1299 cells, irrespective of the promoter used. Our results suggest that ubiquitously expressed HMGB1 and HMGB2 have potential tocell- and promoter-specifically down- or up-regulate in vivo transcriptional activity of different members of the p53 family. A possible mechanism of HMGB1-mediated modulation of p73- and p53-dependent transactivation is discussed. The HMGB1 1HMGhigh mobility groupEMSAelectrophoretic mobility shift assayGSTglutathioneS-transferaseAbantibodyFLfull-lengthHAhemagglutininTAtransactivation domainDBDDNA binding domainODoligomerization domainSAMα sterile motif and HMGB2 proteins (formerly known as HMG1 and HMG2) are the most abundant members of a large HMG family of chromosomal proteins (1Thomas J.O. Travers A.A. Trends Biochem. Sci. 2001; 26: 167-174Abstract Full Text Full Text PDF PubMed Scopus (553) Google Scholar, 2Bustin M. Trends Biochem. Sci. 2001; 26: 152-153Abstract Full Text Full Text PDF PubMed Google Scholar). Vertebrate HMGB1 and HMGB2 proteins contain two similar, but distinct “HMG boxes” (domains A and B), and a long acidic C-terminal “tail.” HMGB1 has been numerously implicated in a host of biologically important processes including transcription, DNA repair, recombination, differentiation, development, and extracellular signaling (1Thomas J.O. Travers A.A. Trends Biochem. Sci. 2001; 26: 167-174Abstract Full Text Full Text PDF PubMed Scopus (553) Google Scholar, 3Muller S. Scaffidi P. Degryse B. Bonaldi T. Ronfani L. Agresti A. Beltrame M. Bianchi M.E. EMBO J. 2001; 20: 4337-4340Crossref PubMed Scopus (377) Google Scholar). HMGB1 can interact in vitro both with DNA (with a selective preference to distorted DNA structures such as Holliday junctions and DNA modified with anticancer drug cisplatin (see Refs. 4Bianchi M.E. Beltrame M. Paonessa G. Science. 1989; 243: 1056-1059Crossref PubMed Scopus (552) Google Scholar, 5Pil P.M. Lippard S.T. Science. 1992; 256: 234-237Crossref PubMed Scopus (512) Google Scholar, 6Štros M. Muselı́ková E. J. Biol. Chem. 2000; 275: 35699-35707Abstract Full Text Full Text PDF PubMed Scopus (46) Google Scholar, 7Štros M. Biochemistry. 2001; 40: 4769-4779Crossref PubMed Scopus (46) Google Scholar)) and a number of biologically important proteins. The latter include transcription factors such as the TATA-binding protein TBP (8Ge H. Roeder R.G. J. Biol. Chem. 1994; 269: 17136-17140Abstract Full Text PDF PubMed Google Scholar), Oct-1/2 (9Zwilling S. Konig H. Wirth T. EMBO J. 1995; 14: 1198-1208Crossref PubMed Scopus (217) Google Scholar) and HoxD9 (10Zappavigna V. Falciola L. Citterich M.H. Mavilio F. Bianchi M.E. EMBO J. 1996; 15: 4981-4991Crossref PubMed Scopus (216) Google Scholar), steroid hormone receptors (11Boonyaratanakornkit V. Melvin V. Prendergast P. Altmann M. Ronfani L. Bianchi M.E. Taraseviciene L. Nordeen S.K. Allegretto E.A. Edwards D.P. Mol. Cell. Biol. 1998; 18: 4471-4487Crossref PubMed Scopus (304) Google Scholar), Rel proteins (12Brickman J. Adam M. Ptashne M. Proc. Natl. Acad. Sci. U. S. A. 1999; 96: 10679-10683Crossref PubMed Scopus (70) Google Scholar), and the tumor suppressor protein p53 (13Jayaraman L. Moorthy N.Ch. Murthy K., G.K. Manley J.L. Bustin M. Prives C. Genes Dev. 1998; 12: 462-472Crossref PubMed Scopus (282) Google Scholar). high mobility group electrophoretic mobility shift assay glutathioneS-transferase antibody full-length hemagglutinin transactivation domain DNA binding domain oligomerization domain α sterile motif p53 is one of the most extensively studied genes. It is now generally accepted that p53 is a crucial tumor suppressor gene for preventing the malignant transformation of cells. This concept is supported by the fact that loss of p53 functions by genetic alternations represents the most common genetic lesions in human cancers occurring in over than 50% of all the tumors (14Oren M. Cell. 1997; 90: 829-832Abstract Full Text Full Text PDF PubMed Scopus (83) Google Scholar). p53 gives rise to a variety of cellular outcomes, most notably cell cycle arrest and apoptosis (14Oren M. Cell. 1997; 90: 829-832Abstract Full Text Full Text PDF PubMed Scopus (83) Google Scholar). These activities are due, at least in part, to the ability of p53 to form homotetramers that bind to specific DNA sequences and activate transcription of a great number of its downstream genes, such as the Mdm2 gene (the product of which, the Mdm2 protein, is a key player in the regulation of stability of p53), a cell cycle-control gene p21 (also known as WAF1), and a apoptosis-inducing gene Bax (Bcl2-associated protein X). p53 has three functional domains: 1) the amino-terminal region involved in transactivation (TA), 2) the central region (the “core domain”) involved in specific DNA-binding (DBD), and 3) the carboxyl-terminal region involved in homooligomerization (oligomerization domain (OD)) and regulation of DNA binding. A p53-related gene,p73 (15Kaghad M. Bonnet H. Yang A. Creancier L. Biscan J.-C. Valent A. Minty A. Chalon P. Lelias J.-M. Dumont X. Ferrara P. McKeon F. Caput D. Cell. 1997; 90: 809-819Abstract Full Text Full Text PDF PubMed Scopus (1538) Google Scholar), encodes six spliced variants (α, β, γ, δ, ε, and ζ; Ref. 16Ichimiya Sh. Nakagawara A. Sakuma Y. Kimura S. Ikeda T. Satoh M. Takahashi N. Sato N. Mori M. Pathol. Int. 2000; 50: 589-593Crossref PubMed Scopus (18) Google Scholar). The p73 isoforms possess all the functional domains found in p53. Further analyses of p73 showed that not only the primary amino acid sequence (63% identity to p53 within the core domain) but also its function resembles that of p53 (14Oren M. Cell. 1997; 90: 829-832Abstract Full Text Full Text PDF PubMed Scopus (83) Google Scholar). p73 diverges from p53 most prominently in the COOH terminus. p73α (but not other isoforms) contains a potential protein-protein interaction module, the SAM domain, that is frequently found in proteins involved in developmental regulation. Although the core domain of p53is the most frequent target for genetic alternations (mainly single point mutations in half of all tumors), very rare mutations inp73 have been found so far despite extensive efforts (16Ichimiya Sh. Nakagawara A. Sakuma Y. Kimura S. Ikeda T. Satoh M. Takahashi N. Sato N. Mori M. Pathol. Int. 2000; 50: 589-593Crossref PubMed Scopus (18) Google Scholar, 17van Oijen M., G.C.T. Slootweg P.J. Clin. Cancer Res. 2000; 6: 2138-2145PubMed Google Scholar, 18Naka M. Ozaki T. Takada N. Takahashi M. Shishikura T. Sakiyama Sh. Tada M. Todo S. Nakagawara A. Oncogene. 2001; 20: 3568-3572Crossref PubMed Scopus (14) Google Scholar). p73, like p53, is induced by treatment of the cells with DNA damaging agents such as ionizing irradiation and anticancer drug cisplatin (19Strano S. Rossi M. Fontemaggi G. Munarriz E. Soddu S. Sacchi A. Blandino G. FEBS Lett. 2001; 490: 163-170Crossref PubMed Scopus (88) Google Scholar). In addition, p73 can transactivate in vivogenes containing p53-responsive promoters. It was shown previously that HMGB1 could significantly stimulatein vivo p53-mediated transactivation in p53-deficient H1299 cells (13Jayaraman L. Moorthy N.Ch. Murthy K., G.K. Manley J.L. Bustin M. Prives C. Genes Dev. 1998; 12: 462-472Crossref PubMed Scopus (282) Google Scholar). The latter finding was explained as a consequence of an HMGB1-mediated enhancement of p53 binding to p53-responsive elements as a result of interactions of HMGB1 with the p53 “core domain” (12Brickman J. Adam M. Ptashne M. Proc. Natl. Acad. Sci. U. S. A. 1999; 96: 10679-10683Crossref PubMed Scopus (70) Google Scholar). Other authors reported that HMGB1 could bind to p53 exclusively via the extreme basic C-terminal domain (20Imamura T. Izumi H. Nagatani G. Ise T. Nomoto M. Iwamoto Y. Kohno J. Biol. Chem. 2001; 276: 7534-7540Abstract Full Text Full Text PDF PubMed Scopus (93) Google Scholar). In this paper, we demonstrate that HMGB1 physically interacts in vitro with two splicing variants of p73, α and β. Both HMG boxes of HMGB1, A and B, interact with p73α. Association of HMGB1 with p73, like the demonstrated ability of HMGB1 to stimulate p73 binding to the Bax and Mdm2 promoters by gel-shift assays, requires the oligomerization region and/or region between DNA-binding domain and oligomerization domain of p73 (amino acids 312–381). Transient transfections revealed that ectopically expressed or endogenous HMGB1 and HMGB2 (antisense strategy), significantly inhibit in vivo p73α/β- or p53-dependent transactivation from the Bax gene promoter (and much less from Mdm2 andp21waf1 promoters) in p53-deficient SAOS-2 cells. In contrast, HMGB1 and HMGB2 stimulate p73- or p53-dependent transactivation in p53-deficient H1299 cells, irrespective of the used promoter. Our results suggest that ubiquitously expressed HMGB1 and HMGB2 have potential tocell- and promoter-specifically down- or up-regulate in vivo transcriptional activity of several members of the p53 family. A possible mechanism of HMGB1-mediated modulation of p73- and p53-dependent transactivation is discussed. A polyclonal HMGB1 antiserum, generated against calf thymus HMG1, was purified by affinity chromatography on a Sepharose column with covalently linked bacterially expressed rat HMGB1. Polyclonal anti-Sp1 antibody (PEP-2X) were from Santa Cruz Biotechnology. Monoclonal p53 (Ab-6 or DO-1) and p73 (Ab-1, Ab-2, and Ab-3) antibodies were from Calbiochem. Human p53 and p73(plus truncated forms) were amplified from the corresponding cDNAs by PCR using the Pfu DNA polymerase and specific oligonucleotide primers. Sense and antisense HMGB1 andHMGB2 were prepared by PCR from the corresponding human cDNAs as above. The amplified DNA samples were gel-purified, and cloned in-frame into the mammalian expression vector pcDNA3 (Invitrogen). All plasmid constructs were dideoxy-sequenced on both strands. HMGB1, p53, p73 and truncated forms were synthesized in vitro from the corresponding cDNAs in the presence of either l-[35S]methionine (Amersham catalog no. AG1094; >37 TBq/mmol) or unlabeledl-Methionine in reticulocyte lysates using the TNT® T7 polymerase quick-coupled transcription/translation system (Promega). The wild-type p53 and p73 as well as the different p73 deletion mutants were in vitrotranscribed/translated with the TNT reticulocyte lysate kit (Promega) in the presence of [35S]methionine. The lysate with labeled proteins was pre-cleared with glutathione-Sepharose beads, followed by addition of GST-HMGB1 or truncated forms of GST-HMGB1 and incubation by rotation for at least 2 h at 4 °C in PD buffer (50 mm Tris-HCl, pH 8.0, 150 mm NaCl, 0.5% Nonidet P-40, 1 mm dithiothreitol, 1 mmphenylmethylsulfonyl fluoride, protease inhibitor mixture (Sigma)). The glutathione-Sepharose beads were then added, and the samples were rotated for at least 1 h at 4 °C. The beads were washed five times with the PD buffer and mixed with 40 μl of 4× concentrated Laemmli buffer, followed by boiling for 5 min. The bound proteins were then resolved by electrophoresis on SDS, 10% polyacrylamide gels. After electrophoresis, the gels were stained in Coomassie Blue R-250, destained, and soaked in Amplify solution (Amersham Biosciences, Inc.) for 30 min. The dried gels were finally exposed to Fuji RX-U films using two intensifying screens at −80 °C. HMGB1 was isolated under nondenaturing conditions from calf thymus and highly purified to near homogeneity on fast protein liquid chromatography as described previously (6Štros M. Muselı́ková E. J. Biol. Chem. 2000; 275: 35699-35707Abstract Full Text Full Text PDF PubMed Scopus (46) Google Scholar, 24Štros M. J. Biol. Chem. 1998; 273: 10355-10361Abstract Full Text Full Text PDF PubMed Google Scholar). Cellular lysates were prepared from SAOS-2 and H1299 cells as detailed previously (25Ozaki T. Naka M. Takada N. Tada M. Sakiyama S. Nakagawara A. Cancer Res. 1999; 59: 5902-5907PubMed Google Scholar), and total protein concentration was determined by protein assay (Bio-Rad). p73 and the truncated polypeptides were synthesized in vitro using the TNT reticulocyte lysate kit (Promega) in the presence of unlabeledl-methionine. DNA for EMSA was directly32P-labeled by PCR of the human Bax (from −138 to −19 from the start site of transcription; Ref. 20Imamura T. Izumi H. Nagatani G. Ise T. Nomoto M. Iwamoto Y. Kohno J. Biol. Chem. 2001; 276: 7534-7540Abstract Full Text Full Text PDF PubMed Scopus (93) Google Scholar) orMdm2 (intronic, from 109 to 224; Ref. 22Zauberman A. Flusberg D. Haupt Y. Barak Y. Oren M. Nucleic Acids Res. 1995; 23: 2584-2592Crossref PubMed Scopus (256) Google Scholar) gene promoters using Taq DNA polymerase with [α-32P]dATP, and the amplified DNA fragments were purified on 1% agarose gels. Reaction mixtures contained 1× EMSA buffer (20 mm Hepes, pH 7.9, 25 mm KCl, 0.1 mm EDTA, 10% glycerol, 2 mm MgCl2, 2 mm spermidine, 0.5 mm dithiothreitol, 0.025% Nonidet P-40) and 0.1 mg/ml acetylated bovine serum albumin as described (13Jayaraman L. Moorthy N.Ch. Murthy K., G.K. Manley J.L. Bustin M. Prives C. Genes Dev. 1998; 12: 462-472Crossref PubMed Scopus (282) Google Scholar). Double-stranded poly(dI-dC), approximately 0.2–1 μg/reaction, was present as a competitor DNA. The protein-DNA complexes were loaded onto native 4 or 5% polyacrylamide gels (29:1, acrylamide/N,N′-methylene-bis(acrylamide)) containing 0.5× TBE (45 mm Tris, 45 mm borate, 1 mm EDTA, pH 8.3) and 0.05% Nonidet P-40. The electrophoresis buffer was in 0.5× TBE containing 0.05% Nonidet P-40. The samples were loaded while the gel was running at 50 V, and the gel was then run at 250 V for 3–4 h at ∼4–8 °C, followed by vacuum drying onto Whatman no. 3MM chromatography paper. Gels were exposed to Fuji RX-U films using two intensifying screens at −80 °C. Quantification of the bands intensity was performed on a PhosphorImager Storm (Molecular Dynamics) using ImageQuant 4.1 software for data processing. For the permanent record, the gels were scanned and subsequently adjusted for contrast/brightness using Adobe Photoshop. p53-deficient SAOS-2 and H1299 cells were maintained in Dulbecco's modified Eagle's medium and RPMI 1640 medium, respectively, supplemented with 10% heat-inactivated fetal calf serum and antibiotics. The SAOS-2 or H1299 cells (2.5–5 × 104 cells/well of a 12-well plate) were transiently transfected using LipofectAMINE (Invitrogen) or FuGENE 6 (Roche Molecular Biochemicals), respectively. Transfection mixtures contained one or two expression vectors (encoding human p53, p73, HMGB1/HMGB2, or antisense HMGB1/HMGB2 in pcDNA3 plasmid; typically 250 ng each) and two different reporter constructs, the pRL family Renilla luciferase control reporter vector with the cDNA encoding Renilla luciferase under the control of the herpes simplex virus thymidine kinase promoter (pRL-TK vector; 40 ng) and a construct of firefly (Photinus pyralis) cDNA under the control of a promoter containing p53-responsive elements (pGL3-p21waf1 -luc, kindly provided by K. Watanabe; pGL2-NA(hMdm2)-luc, Ref. 21Friedlander P. Haupt Y. Prives C. Oren M.A. Mol. Cell. Biol. 1996; 16: 4961-4971Crossref PubMed Scopus (269) Google Scholar; pGL3-Bax-luc, Ref. 23Miyashita T. Reed J.C. Cell. 1995; 80: 293-299Abstract Full Text PDF PubMed Scopus (305) Google Scholar). The luciferase activity was measured 48 h after transfection using a dual-luciferase reporter gene assay system, according to the procedures provided by the manufacturer (Promega). It was reported earlier that HMGB1 physically associates with p53 (13Jayaraman L. Moorthy N.Ch. Murthy K., G.K. Manley J.L. Bustin M. Prives C. Genes Dev. 1998; 12: 462-472Crossref PubMed Scopus (282) Google Scholar, 20Imamura T. Izumi H. Nagatani G. Ise T. Nomoto M. Iwamoto Y. Kohno J. Biol. Chem. 2001; 276: 7534-7540Abstract Full Text Full Text PDF PubMed Scopus (93) Google Scholar). To investigate whether HMGB1 (Fig.1, panel A) can also interact with other members of p53 family, pull-down assays were carried out with HMGB1 and in vitro synthesized (reticulocyte lysate) isoforms of p73, α and β. Lysates were incubated with GST-HMGB1 or domains, followed by incubation with glutathione-Sepharose beads and subsequent washing of the beads. Proteins that were associated with glutathione beads were then subjected to SDS-polyacrylamide gel electrophoresis. As seen in Fig. 1 (panels C andD), HMGB1, like its isolated A domain, could clearly bind both isoforms of p73, α and β (binding of c-Abl SH2/3 domains to p73α, a positive control in panel C, was very weak in the pull-down assay as compared with the previously reported in vivo interactions (27Agami R. Blandino Oren M. Shaul Y. Nature. 1999; 399: 809-813Crossref PubMed Scopus (504) Google Scholar), possibly because of the lack of post-translational modifications). The B domain of HMGB1 could bind p73α only when it contained a seven-residue NH2-terminal extension (85TKKKFKD91) (Fig. 1, compare B domain polypeptides designated as B and B7 in panels A andC). Similar results were obtained using the pull-down assay with p53, 2M. Štros, unpublished results. explaining the previously reported inability of the B domain to bind p53 (20Imamura T. Izumi H. Nagatani G. Ise T. Nomoto M. Iwamoto Y. Kohno J. Biol. Chem. 2001; 276: 7534-7540Abstract Full Text Full Text PDF PubMed Scopus (93) Google Scholar). These finding provided the first evidence that the NH2-terminal85TKKKFKD91 sequence of the HMGB1 B domain is essential not only for binding to DNA (7Štros M. Biochemistry. 2001; 40: 4769-4779Crossref PubMed Scopus (46) Google Scholar, 24Štros M. J. Biol. Chem. 1998; 273: 10355-10361Abstract Full Text Full Text PDF PubMed Google Scholar) but also for interaction with proteins. To determine which region(s) of p73 is involved in binding of HMGB1, pull-down assays were carried out with lysates containing in vitro synthesized 35S-labeled truncated forms of p73α. As shown in Fig. 2 (panel A), binding of p73α lacking the extreme C-terminal region (peptide 1–550) to HMGB1 was significantly weaker (∼10-fold) relative to the full-length p73α. On the other hand, binding of p73 lacking both the extreme C-terminal and the SAM domain (peptide 1–484), or p73 truncated up to the middle of the TA2 domain (peptide 1–427), to HMGB1 was similar to the full-length p73α. These results suggest that, in the absence of the extreme C-terminal region, the SAM domain only slightly reduces ( 20-fold) of the p73 peptide (residues 1–381) to HMGB1, with no binding observed with the p73 peptide containing only the TA and DBD (peptide 1–311). The above results indicate that the amino acid residues 312–381 (OD and residues between DBD and OD) are essential for interaction of p73 with HMGB1, with residues 382–426 (TA2) enhancing p73 binding to HMGB1 (Fig.2). It was shown previously that HMGB1 could stimulate p53 binding to DNA fragments containing the p53 consensus sites (13Jayaraman L. Moorthy N.Ch. Murthy K., G.K. Manley J.L. Bustin M. Prives C. Genes Dev. 1998; 12: 462-472Crossref PubMed Scopus (282) Google Scholar). Here we have investigated the possibility whether HMGB1 could also enhance specific DNA binding of p73. As shown in Fig.3, binding of in vitrosynthesized p73α to the Bax or Mdm2 promoters was barely detectable (lanes 4), in agreement with the reported inhibitory role of the extreme COOH-terminal region of p73α on specific DNA binding in vivo and in vitro(25Ozaki T. Naka M. Takada N. Tada M. Sakiyama S. Nakagawara A. Cancer Res. 1999; 59: 5902-5907PubMed Google Scholar). Nevertheless, HMGB1 could slightly enhance binding of p73α to DNA (Fig. 3, lanes 5 in panels A andB). Partial deletion of the COOH terminus (peptides 1–550 or 1–381; see Fig. 2 B for the design of the used truncated p73 peptides) resulted in a significantly increased binding of the truncated p73α to the Mdm2 promoter (lanes 6and 8 in Fig. 3 B), with no visible binding to theBax promoter (lanes 6 and 8 in Fig.3 A; higher affinity of truncated p73 for Mdm2promoter was likely the result of the presence of two consensus sequences, unlike a single and imperfect consensus sequence within theBax promoter, Ref. 28Kaku S. Iwahashi Y. Kuraishi A. Albor A. Yamagishi T. Nakaike S. Kulesz-Martin M. Nucleic Acids Res. 2001; 29: 1989-1993Crossref PubMed Scopus (18) Google Scholar). However, binding of the latter p73 peptides to both promoters was markedly enhanced in the presence of HMGB1 (lanes 7 and 9 in Fig. 3; the absence of free DNA probe in panel B (lanes 7 and9), is a result of binding of excessive amount of the p73 peptides to DNA probe, whereas the absence of most of the free DNA probe in panel A (lane 9) is caused by a prolonged electrophoresis for better resolution of the complexes). Interestingly, very little, if any, DNA binding was observed with p73 peptide containing only transactivation domain and DBD (peptide 1–311), irrespective of the presence of HMGB1 (Fig. 3, lanes 10 and 11). Our results suggest that amino acid residues 312–381 are required for binding of p73 to both HMGB1 and specific DNA sites that are otherwise recognized by the “core domain” of p73 (Fig. 2 B). Previously it was demonstrated that HMGB1 could stimulate the p53-dependent transactivation from thecyclin G promoter in p53-deficient H1299 cells (12Brickman J. Adam M. Ptashne M. Proc. Natl. Acad. Sci. U. S. A. 1999; 96: 10679-10683Crossref PubMed Scopus (70) Google Scholar). To find out whether HMGB1 could also stimulate p73-dependent transcriptional activation and whether the effect of HMGB1 is distinct on different promoters, transient transfections were carried out with reporter plasmids containing the Bax, Mdm2, orp21waf1 promoters in H1299 cells (endogenous p73 protein is undetectable in H1299 cell extracts by direct Western blot analysis (Ref. 29Di Como C.J. Gaiddon C. Prives C. Mol. Cell. Biol. 1999; 19: 1438-1449Crossref PubMed Scopus (381) Google Scholar), but a weak signal corresponding to p73 m-RNA is to be detected after 25 cycles of reverse transcription-PCR). 3J. Shimbo, unpublished results. The latter promoters, when placed upstream of the luciferase gene, were previously shown to be activated to varying degree by p73 (25Ozaki T. Naka M. Takada N. Tada M. Sakiyama S. Nakagawara A. Cancer Res. 1999; 59: 5902-5907PubMed Google Scholar, 26Zeng X. Chen J. Jost C.A. Maya R. Keller D. Wang X. Kaelin W.G.J. Oren M. Chen J. Lu H. Mol. Cell. Biol. 1999; 19: 3257-3266Crossref PubMed Scopus (304) Google Scholar). As shown in Fig. 4, cotransfection of plasmids encoding p73α significantly stimulated transcription from all the studied reporter plasmids. Cotransfection of plasmids encoding p73α and HMGB1 into H1299 cells resulted in up to ∼2-fold enhancement of the p73-dependent transactivation from the Bax, Mdm2, orp21waf1 promoters, with no clear differences among the tested promoters (Fig. 4). Similarly, HMGB1 could stimulate the p53-dependent transactivation from all the promoters studied,2 as also reported preciously for the cyclin G promoter (13Jayaraman L. Moorthy N.Ch. Murthy K., G.K. Manley J.L. Bustin M. Prives C. Genes Dev. 1998; 12: 462-472Crossref PubMed Scopus (282) Google Scholar). Our results indicate that HMGB1 can up-regulate both the p53- and p73-dependent transactivation in H1299 cells. It was previously demonstrated that the extent of p73-dependent transactivation in SAOS-2 cells was proportional to the amount of plasmids encoding p73α and β (30Ueda Y. Hijikata M. Takagi S. Chiba T. Shimotohno K. Oncogene. 1999; 18: 4993-4998Crossref PubMed Scopus (131) Google Scholar). To find out whether the effect of HMGB1 on the p73- or p53-dependent transactivation iscell-specific, transfection experiments were carried out with p53-deficient osteosarcoma (SAOS-2) cells exhibiting only very low levels of endogenous p73 protein or p73 m-RNA (31Chen X. Zheng Y. Zhu J. Jiang J. Wang J. Oncogene. 2001; 20: 769-774Crossref PubMed Scopus (84) Google Scholar). Interestingly, cotransfection of plasmids encoding p73α (or the alternatively spliced p73β) or p53 with plasmid encoding HMGB1 into SAOS-2 cells resulted in significant (up to ∼4-fold) inhibition of p73/p53-dependent transactivation from the Baxpromoter (Fig. 5). The observed inhibition was proportional to the amount of the plasmid encoding HMGB1 (Fig. 5). The HMGB1-mediated inhibition of transcriptional activation from the Bax promoter was also observed using p73 truncated up to the OD, peptide 1–381 in Fig. 5 C (nuclear localization of the p73-(1–381) peptide was confirmed by confocal microscopy).2 Our results may indicate that the region encompassing the OD, and/or the region between the OD and DBD (amino acids 312–381), is required for HMGB1-mediated inhibition of transactivation. This conclusion is also supported by the fact that truncated p73 containing only TA and DBD (peptide 1–311) was unable to bind HMGB1 (pull-down assay; Fig. 3), explaining the inability of HMGB1 to enhance binding of the p73 peptide to specific DNA sites (EMSA; Fig.3). In contrast to transient transfections with the Baxpromoter, cotransfections of reporter plasmids containing either theMdm2 promoter (Fig. 6,panel A) or the p21waf1 promoter2 with plasmids encoding p53 or p73α and HMGB1 resulted in only a slight (up to ∼20%) inhibition of the p53- and p73-dependent transactivation in SAOS-2 cells. These results gave evidence that HMGB1 can specifically inhibit p53- and p73-dependent transactivation from the Baxpromoter in SAOS-2 cells. HMGB1 is a relatively abundant architectural nuclear protein, and it is possible that not all HMGB1 molecules are engaged in chromatin structure and may also serve other functions in the cell. To test whether endogenous HMGB1 could affectin vivo transcriptional activity of p53 or p73, plasmids encoding p73α/β or p53 were transiently transfected into SAOS-2 cells together with reporter constructs and plasmid encoding human antisense HMGB1. Expression of antisense (or sense) HMGB1 had little effect on p53- or p73α/β-mediated transactivation from the humanMdm2 (Fig. 6, panel A) orp21waf1 promoters.2 However, similar experiments with the Bax-luciferase reporter plasmid revealed a reproducibly enhanced (>2-fold) p53 or p73α/β-dependent transactivation when the SAOS-2 cells were co-transfected with a construct producing antisense HMGB1 (Fig. 6,panel B). These results clearly demonstrated that endogenous HMGB1 suppresses in vivo transcriptional activity of p53 and two splicing variants of p73, α and β. HMGB1 is not the only abundant HMG protein in the nucleus. A closely related HMGB2 protein is present in the nucleus in comparable amounts relative to HMGB1. We have there
Referência(s)