Identification of the Interferon Regulatory Factor 5 Gene (IRF-5)

Identi®cation of the interferon regulatory factor 5 gene (IRF-5) as a direct target for p53

Toshiki Mori1, Yoshio Anazawa1, Megumi Iiizumi1, Seisuke Fukuda1, Yusuke Nakamura1 and Hirofumi Arakawa*,1

1Human Genome Center, Institute of Medical Science, University of Tokyo, 4-6-1 Shirokanedai, Minato-ku, Tokyo 108-8639, Japan

Interferon regulatory factors (IRFs) regulate transcrip- (Tokino et al., 1994), we suspect that p53 is likely to tion of interferon genes through DNA sequence-speci®c achieve tumor suppression by regulating the physiolo- binding to these targets. Using a dierential display gical functions of not a few, but multiple genes. To method for examining gene expression in p53-defective clarify the precise mechanisms of p53 tumor suppres- cells infected with adenovirus containing wild-type p53, sion, identi®cation of additional target genes would we found that expression of interferon regulatory factor seem to be an absolute requirement. 5(IRF-5) mRNA was increased in the presence of In past experiments we succeeded in isolating several exogenous p53. An electrophoretic mobility-shift assay p53-target genes using two dierent methods. One showed that a potential p53 binding site (p53BS) method involved dierential display using a cell line in detected in exon 2 of the IRF-5 gene could in fact bind which expression of an exogenous wild-type p53 gene to p53 protein. Moreover, a heterologous reporter assay can be regulated under the control of the lactose revealed that the p53BS possessed p53-dependent operon (Takei et al., 1998). In this manner we isolated transcriptional activity. Expression of IRF-5 was induced four novel genes, TP53TG1 (Takei et al., 1998), in p53+/+ cells (MCF7 and NHDF), but not in TP53TG3 (Ng et al., 1999), p53R2 (Tanaka et al., p537/7 cells (H1299) when DNA was damaged by g- 2000), and p53DINP1 (Okamura et al., 2001), as direct irradiation, UV-radiation, or adriamycin treatment in a targets of p53. p53R2 is commonly involved in p53- wild-type p53-dependent manner. These results suggest dependent DNA repair in response to diverse agents of that IRF-5 is a novel p53-target, and that it might DNA damage (Tanaka et al., 2000), while p53DINP1 mediate the p53-dependent immune response. mediates the p53-dependent apoptotic pathway (Oka- Oncogene (2002) 21, 2914 ± 2918. DOI: 10.1038/sj/ mura et al., 2001). Our second method was to employ a onc/1205459 yeast enhancer-trap system that allowed direct cloning of functional p53-binding sequences from human Keywords: IRF5; p53; target gene genomic DNA (Tokino et al., 1994). We isolated and sequenced cosmid clones containing every p53-binding site, and then searched for candidate genes in the genomic regions surrounding those binding sequences. The p53 gene is considered to be the most important of With that method we isolated ®ve novel genes: GML, all tumor suppressor genes, because nearly half of all which contributes to the sensitivity of cells to anti- human cancers examined are found to contain p53 cancer drugs (Furuhata et al., 1996); P2XM (Urano et mutations (Greenblatt et al., 1994). The gene product is al., 1997), BAI1, which inhibits angiogenesis in gliomas a transcription factor that binds to speci®c DNA (Nishimori et al., 1997); CSR (Han et al., 1998); and sequences to activate transcription of target genes p53AIP1, a pivotal mediator of p53-induced apoptosis (Levine, 1997). Numerous p53-targets have been (Oda et al., 2000). These results suggested that p53 isolated and characterized to date. Among them, p21/ might have multiple physiological functions, by virtue WAF1 and BAX are thought to be the most important of its ability to activate transcription of so many because their products mediate two major functions of dierent genes. p53, cell-cycle arrest and apoptosis respectively (Harper To isolate novel p53-target genes we applied a et al., 1993; el-Deiry et al., 1993; Miyashita and Reed, dierential display method using mRNAs isolated 1995). Moreover, given that many p53-target genes from p53-de®cient U373MG cells (glioblastoma, the have been reported and that more than 100 potential American Type Culture Collection, Manassas, VA, binding sites for p53 are present in the human genome USA) infected by either Ad-p53 or Ad-LacZ. This strategy had previously shown that the fractalkine gene was a direct target of p53 (Shiraishi et al., 2000). A DNA fragment corresponding to one of the bands that *Correspondence: H Arakawa, E-mail: [email protected] Received 30 May 2001; revised 15 November 2001; accepted 26 showed a stronger intensity in p53-transfected cells November 2001 than in the LacZ-transfected cells was excised, cloned, IRF5 is a target for p53 T Mori et al 2915 and sequenced as described previously (Takei et al., 1998). A BLAST search of the public database indicated that DNA sequences of this fragment were identical to part of the IRF-5 gene (Harada et al., 1998). Northern-blot analysis con®rmed that expression of the IRF-5 gene was remarkably increased by infection of Ad-p53, but not Ad-LacZ, in a time- dependent manner (Figure 1a). To determine whether IRF-5 is a direct target of p53, we searched for a p53-binding sequence(s) within the 15-kb genomic sequence containing the IRF-5 gene Accession number: AC025594) and found a potential p53-binding sequence (p53BS) in exon 2 (Figure 1b). Eighteen of the 20 nucleotides of this p53BS matched the consensus p53-binding sequence proposed by el- Deiry et al. (1992). To con®rm the interaction between p53 protein and the candidate sequence (p53BS), we performed an electrophoretic mobility-shift assay (EMSA) using a nuclear extract puri®ed from H1299 cells (lung carcinoma, the American Type Culture Collection) infected by Ad-p53. As shown in Figure 2a, p53BS

Figure 2 IRF-5 as a direct target of p53. (a) Electrophoretic mobility-shift assay (EMSA). Oligonucleotides for p53BS were synthesized and annealed (p53BS sense, 5'-AGGCATGCCA- CAAGGCATGGT-3'; p53BS antisense, 5'-GACCATGCCTTG- TGGCATGCCT-3'). Nuclear extract puri®ed from H1299 cells infected by Ad-p53 were incubated with sonicated salmon sperm DNA, the g33PATP-labeled double-strand oligomer and, in some cases, with monoclonal anti-p53 antibodies, PAb421 (Oncogene Science, Cambridge, MA, USA) and/or PAb1801 (Santa Cruz Biotechnology, Santa Cruz, CA, USA). p53BS bound to a protein contained in the nuclear extract (lane 1), and the band was super- shifted in the presence of Pab421 (lane 2). The interaction was inhibited by unlabeled oligonucleotides corresponding to the p53BS (competitor DNA (self)) (lane 3), but not by non-speci®c oligonucleotides (competitor DNA (non-speci®c)) (lane 4). The band was super-super-shifted by an addition of another anti-p53 antibody, Pab1801 (lane 5). (b) Luciferase assay. Oligonucleotides for two copies of p53BS (BS sense and antisense) and a one- nucleotide-exchanged mutant of each of them (mt-BS sense and antisense) were designed as follows: BS sense, 5'-CGCGTAGG- Figure 1 Transcriptional activation of IRF-5 by p53 and a CATGCCACAAGGCATGGTCAGGCATGCCACAAGGCAT- potential p53-binding site in the IRF-5 gene. (a) Northern-blot GGTCC-3'; BS antisense, 5'-TCGAGGACCATGCCTTGTGG- analysis of IRF-5 mRNA in the U373MG cell line at the indicated CATGCCTGACCATGCCTTGTGGCATGCCTA-3'; mt-BS times after infection with either Ad-p53 or Ad-LacZ. RNAs of sense, 5'-CGCGTAGGCATTCCACAAGGCATGGTCAGGCA- these cells were isolated on a time course of zero, 6, 12, 24, 48 h TGCCACAAGGCATTGTCC-3'; and mt-BS antisense, 5'- after infection. The probes were 1.5- and 2.0-kb cDNA fragments TCGAGGACCATGCCTTGTGGAATGCCTGACCATGCCTT- carrying sequences of IRF-5 and b-Actin, respectively, (b) GTGGAATGCCTA-3'. These oligomers were annealed and Genomic structure of the IRF-5 gene and a potential p53-binding ligated into pGL3-promoter vectors (Promega, Madison, WI, site. Transcripts b and c were identi®ed in this study. Black lines USA) to produce pGL3-BS and pGL3-mt-BS. H1299 cells were indicate genomic DNA; white boxes are noncoding exons, and co-transfected with 1 mg each of pGL3-BS, pGL3-mt-BS, or black boxes are coding exons. (a, major transcript reported pGL3-promoter vector and wild-type p53 expression vector (wt- previously (GenBank accession number U51127); b, transcript p53), mutant-type p53 expression vector (mt-p53), or pcDNA3.1 that spliced out 48 nucleotides corresponding to the 5' portion of vector in combination with 1 mg of pRL-TK vector. 48 h after exon 6; c, transcript that spliced out 34 nucleotides corresponding transfection, the cells were lysed in a passive lysis buer to exon 5). An arrow indicates the location of the potential p53- (Promega) and the Dual Luciferase system (Promega) was binding site (p53BS) in exon 2. Consensus sequence: consensus applied. Luciferase activities of pGL3-BS and pGL3-mt-BS were p53-binding sequence with (R)=purine, (Y)=pyrimidine, and indicated as a fold induction relative to the activity of reporter (W)=A or T vector without insertion of the p53-binding sequence

Oncogene IRF5 is a target for p53 T Mori et al 2916

Figure 3 Endogenous p53-dependent expression of IRF-5 in response to DNA damage. Expression of IRF-5 mRNA in MCF7 (p53+/+) or H1299 (p537/7) cells was examined by Northern blot at the indicated times following genotoxic stresses such as adriamycin treatment (0.2 mg/ml, 2 h), g-radiation (14 Gy), or UV-radiation (10 J/m2). RNAs of these cells were isolated on a time course of zero, 6, 12, 24, 48, 72 h after cellular stresses. The probes were 1.5-kb and 2.0-kb cDNA fragments carrying sequences of IRF-5 and b-Actin, respectively

Figure 4 Colony-formation assay. (upper) Growth suppression by ectopically expressed IRF-5 in A549 lung cancer (a) and HCT116 colorectal cancer (b) cells. Sense: sense-IRF5 expression vector, antisense: antisense-IRF5 expression vector. (lower) Numbers of G-418 resistant colonies from each of A549 and HCT116 cell lines transfected with either sense-IRF5 expression vector (sense) or antisense-IRF5 expression vector (antisense). Each experiment was repeated at least three times using triplicate samples, and the average scores are shown along with error-bars

bound to a protein contained in the nuclear extract indicating that p53BS bound to the p53 protein (lane (lane 1), and the band was super-shifted in the presence 2). This evidence was further clari®ed by speci®c of mouse monoclonal anti-p53 antibody Pab421, competition with self-DNA but not with non-speci®c

Oncogene IRF5 is a target for p53 T Mori et al 2917 DNA (lanes 3 and 4), and also by the fact the band was super-super-shifted by an addition of another mouse monoclonal anti-p53 antibody, Pab1801 (lane 5). Figure 5 Comparison of the amino-acid sequences of IRF-5 and IRF-1 for DNA binding domain. Identities are indicated in black Then to examine p53-dependent transcriptional activity of p53BS, we performed a heterologous reporter assay. As shown in Figure 2b, luciferase activity of pGL3-BS was enhanced more than 15-fold al., 1998). They constitute a protein family of which by co-transfection with wild-type p53 expression vector more than ten members have been reported so far, than by co-transfection with either mutant-type including IRF-1, IRF-2, IRF-3, IRF-4, IRF-5, IRF-6, (R273H) p53 expression vector or mock vector. IRF-7, interferon consensus sequence-binding protein pGL3-mtBS vector that included a point mutation (ICSBP), and interferon-stimulated gene factor 3g within p53BS revealed no enhancement of luciferase (ISGF3g) (Harada et al., 1998; Pitha et al., 1998). activity by co-transfection with wild-type p53 expres- IRF-1 regulates the interferon system by binding to a sion vector (Figure 2b). speci®c DNA sequence in target genes, and also Moreover, we were able to show that IRF-5 mRNA regulates cell growth, i.e., functioning as a tumor was induced in response to DNA damage caused by suppressor (Tanaka et al., 1994). IRF-3 plays a either adriamycin or g-radiation in NHDF4042 cells defensive role against genotoxic stresses (Kim et al., (data not shown) and MCF7 cells (Figure 3) containing 1999). wild-type p53 (p53+/+), but not in H1299 cells On multi-tissue Northern blots, two major tran- (p537/7) (Figure 3). Taken together, these results scripts of approximately 3.0 and 2.4 kb were highly clearly indicated that IRF-5 is indeed a direct target for expressed in spleen and peripheral blood lymphocyte p53. (data not shown). The speci®c expression of this gene To examine the eect of IRF-5 on cell growth, we in lymphoid tissues indicates that it might be involved performed colony-formation assays using expression in the immune system, as are IRF-4 and ICSBP vectors (sense-IRF5-FLAG or sense-IRF5) designed to (Matsuyama et al., 1995; Driggers et al., 1990). express IRF-5 protein with or without a FLAG tag at Therefore IRF-5 might play a role in signal transduc- the C-terminal. As a negative control we also prepared tion pathways that mediate either the protection or an expression vector in which the entire coding region apoptosis of aected cells in the defensive responses of IRF-5 was inserted in the anti-sense direction induced by a speci®c genotoxic stress such as viral (antisense-IRF5). Each of these vectors was transfected infection. into A549 (lung cancer), or HCT116 (colorectal cancer) IRF-1 functions as a tumor suppressor by activating cells. Expression of IRF-5 protein after transfection transcription of p21/WAF1 and ICE, and it also with sense-IRF5-FLAG vector was con®rmed by regulates cell-cycle arrest and apoptosis in response to immunoblotting, using mouse anti-FLAG monoclonal various genotoxic stresses such as viral infection and antibody (Kodak, data not shown). After 2 weeks of genotoxins (Tamura et al., 1995; Tanaka et al., 1996). incubation in geneticin-containing media, ectopic Indeed, over-expression of IRF-5 in some cancer cells expression of IRF-5 had reduced the numbers of revealed the signi®cant growth-suppressive eect (Fig- colonies in these cell lines (Figure 4). Similar results ure 4). Although the function of IRF5 remains to be were obtained when we used sense-IRF5-FLAG established, it might also be involved in tumor expression vector (data not shown), suggesting that suppression in a p53-dependent manner. IRF-5 might play a role in cell growth control for some tissues outside of hematopoietic cells. IRF-5 protein consists of 504 amino acids with a predicted molecular weight of 57 kDa (Accession number: AAA96056). It possesses about 30 ± 40% Acknowledgments We thank K Matsui for her excellent technical assistance. amino-acid identity to IRF-1, including the N-terminal This work was supported in part by Grant #13216031 from portions containing DNA-binding domains (Figure 5). the Ministry of Education, Culture, Sports, Science and IRFs are thought to be transcription factors that also Technology (to H Arakawa) and in part by `Research for play multiple roles, speci®cally as regards host defenses the Future' Program Grant #00L01402 from The Japan against various stimuli (Harada et al., 1998; Pitha et Society for the Promotion of Science (to Y Nakamura).

References

Driggers PH, Ennist DL, Gleason SL, Mak WH, Marks MS, el-Deiry WS, Tokino T, Velculescu VE, Levy DB, Parsons R, Levi BZ, Flanagan JR, Appella E, Ozato K. (1990). Proc. Trent JM, Lin D, Mercer WE, Kinzler WE and Vogelstein Natl. Acad. Sci. USA, 87, 3743 ± 3747. B. (1993). Cell, 75, 817 ± 825. el-Deiry WS, Kern SE, Pietenpol JA, Kinzler KW and Furuhata T, Tokino T, Urano T and Nakamura Y. (1996). Vogelstein B. (1992). Nat. Genet., 1, 45 ± 49. Oncogene, 13, 1965 ± 1970.

Oncogene IRF5 is a target for p53 T Mori et al 2918 Greenblatt MS, Bennett WP, Hollstein M and Harris CC. Pitha PM, Au WC, Lowther W, Juang YT, Schafer SL, (1994). Cancer Res., 54, 4855 ± 4878. Burysek L, Hiscott J and Moore PA. (1998). Biochimie, 80, Han HJ, Tokino T and Nakamura Y. (1998). Hum. Mol. 651 ± 658. Genet., 7, 1039 ± 1046. ShiraishiK,FukudaS,MoriT,MatsudaK,YamaguchiT, Harada H, Taniguchi T and Tanaka N. (1998). Biochimie, Tanikawa C, Ogawa M, Nakamura Y and Arakawa H. 80, 641 ± 650. (2000). Cancer Res., 60, 3722 ± 3726. Harper JW, Adami GR, Wei N, Keyomarsi K and Elledge Takei Y, Ishikawa S, Tokino T, Muto T and Nakamura Y. SJ. (1993). Cell, 75, 805 ± 816. (1998). Genes Chrom. Cancer, 23, 1±9. KimT,KimTY,SongYH,MinIM,YimJandKimTK. Tamura T, Ishihara M, Lamphier MS, Tanaka N, Oishi I, (1999). J. Biol. Chem., 274, 30686 ± 30689. Aizawa S, Matsuyama T, Mak TW, Taki S and Taniguchi Levine AJ. (1997). Cell, 88, 323 ± 331. T. (1995). Nature, 376, 596 ± 599. Matsuyama T, Grossman A, Mittrucker HW, Siderovski Tanaka H, Arakawa H, Yamaguchi T, Shiraishi K, Fukuda DP,KieferF,KawakamiT,RichardsonCD,TaniguchiT, S, Matsui K, Takei Y and Nakamura Y. (2000). Nature, Yoshinaga SK and Mak TW. (1995). Nucl. Acids Res., 23, 404, 42 ± 49. 2127 ± 2136. Tanaka N, Ishihara M, Kitagawa M, Harada H, Kimura T, Miyashita T and Reed JC. (1995). Cell, 80, 293 ± 299. Matsuyama T, Lamphier MS, Aizawa S, Mak TW and Ng CC, Koyama K, Okamura S, Kondoh H, Takei Y and Taniguchi T. (1994). Cell, 77, 829 ± 839. Nakamura Y. (1999). Genes Chrom. Cancer, 26, 329 ± 335. Tanaka N, Ishihara M, Lamphier MS, Nozawa H, NishimoriH,ShiratsuchiT,UranoT,KimuraY,KiyonoK, MatsuyamaT,MakTW,AizawaS,TokinoT,OrenM Tatsumi K, Yoshida S, Ono M, Kuwano M, Nakamura Y and Taniguchi T. (1996). Nature, 382, 816 ± 818. and Tokino T. (1997). Oncogene, 15, 2145 ± 2150. Tokino T, Thiagalingam S, el-Deiry WS, Waldman T, Oda K, Arakawa H, Tanaka T, Matsuda K, Tanikawa C, Kinzler KW and Vogelstein B. (1994). Hum. Mol. Genet., Mori T, Nishimori H, Tamai K, Tokino T, Nakamura Y 3, 1537 ± 1542. and Taya Y. (2000). Cell, 102, 849 ± 862. UranoT,NishimoriH,HanH,FuruhataT,KimuraY, Okamura S, Arakawa H, Tanaka T, Nakanishi H, Ng CC, Nakamura Y and Tokino T. (1997). Cancer Res., 57, Taya Y, Monden M and Nakamura Y. (2001). Mol. Cell, 3281 ± 3287. 8, 85 ± 94.

Oncogene