- Email: [email protected]
placental bikunin genes and their promoters1
Biochimica et Biophysica Acta 1519 (2001) 92^95
www.bba-direct.com
Promoter paper
Mouse hepatocyte growth factor activator inhibitor type 1 (HAI-1) and type 2 (HAI-2)/placental bikunin genes and their promoters1 Hiroshi Itoh b
a;
*, Hiroaki Kataoka a , Jing Yan Meng a , Ryouichi Hamasuna a , Naomi Kitamura b , Masashi Koono a
a Second Department of Pathology, Miyazaki Medical College, Kiyotake, Miyazaki 889-1692, Japan Department of Biological Sciences, Graduate School of Bioscience and Biotechnology, Tokyo Institute of Technology, Yokohama, Japan
Received 21 December 2000 ; received in revised form 19 March 2001; accepted 26 March 2001
Abstract Hepatocyte growth factor (HGF) activator inhibitor type 1 (HAI-1) and type 2 (HAI-2) were recently discovered as specific inhibitors of HGF activator. Each of them contains two Kunitz-type serine protease inhibitor domains and a transmembrane domain, so that their overall structures are similar to each other. In this study, mouse genes encoding HAI-1 and HAI-2 were cloned by screening of a mouse genomic bacterial artificial chromosome library and by polymerase chain reaction of mouse genomic DNA, respectively. The genes (mHAI1 and mHAI-2) were defined to consist of 11 and eight exons spanning 11 kbp and 9.5 kbp, respectively. Neither a TATA nor CAAT box was found in 5P-flanking regions of both genes and no apparent homologous portion was observed between mHAI-1 and mHAI-2 promoter regions. Promoter assay of mHAI-1 and human HAI-1 revealed that the potential binding sites of a complex of Egr-1^3 and Sp1, which was well-conserved between human (342 to 358) and mouse (344 to 357), might be a key portion of its transcriptional regulation to function as not only house-keeping but also early responsive genes. ß 2001 Elsevier Science B.V. All rights reserved. Keywords : Hepatocyte growth factor; Hepatocyte growth factor activator inhibitor ; Placental bikunin ; Gene structure; Mouse
Hepatocyte growth factor (HGF)/scatter factor (SF) is a multifunctional polypeptide factor that functions as a mitogen, morphogen, and/or motogen for a wide variety of cells [1^3]. HGF/SF is secreted by mesenchymal cells as an inactive precursor form and activated by the speci¢c serine proteinase, HGF activator (HGFA) [4,5]. The activity of HGFA is regulated by two recently identi¢ed serine proteinase inhibitors, namely HGF activator inhibitor type 1 (HAI-1) and type 2 (HAI-2), both of which were initially puri¢ed from the conditioned medium of a human stomach cancer cell line MKN45 [6,7]. They are structurally similar to each other and each contains two Kunitz-type
Abbreviations : HGF, hepatocyte growth factor; HGFA, HGF activator; HAI, HGF activator inhibitor; mHAI, mouse HAI; PB, placental bikunin; PCR, polymerase chain reaction * Corresponding author. Fax: +81-985-85-6003; E-mail : [email protected] 1 The nucleotide sequence data of 5P-£anking regions reported in this paper will appear in the GSDB, DDBJ, EMBL and NCBI nucleotide sequence databases with the following accession numbers, AY013701 (mHAI-1) and AY013702 (mHAI-2).
serine protease inhibitor domains (KDs) and a single putative transmembrane domain [6,7]. Among them, HAI-2 was identical with the protein originally reported as placental bikunin (PB) [8]. Both HAI-1 and HAI-2/PB genes are expressed abundantly in various tissues including the placenta, kidney, pancreas and gastrointestinal tract as assessed by RNA blot analysis [6,7]. Immunohistochemically, HAI-1 protein is localized on cellular lateral (or baso-lateral) surface of a simple columnar epithelium of ducts, tubules and mucosal surface of various organs [9], while HAI-2/PB protein is mainly observed in the cytoplasms of the epithelial cells and macrophage-like monocytic in£ammatory cells of various tissues (unpublished observation). HAI-1 but not HAI-2/PB is up-regulated during the regeneration of damaged gastrointestinal mucosa, liver and kidney [9,10]. These results strongly suggest that HAI-1 and -2 may have di¡erent and distinctive roles in vivo, although they are structurally similar to each other. However, precise biophysiological functions of HAI-1 and -2 in vivo remain to be elucidated. Recently, we showed that the HAI-2 lacking KD1 by an alternative splicing is a predominant transcript in mouse but not in human [11] and subsequently we cloned the genes for hu-
0167-4781 / 01 / $ ^ see front matter ß 2001 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 7 - 4 7 8 1 ( 0 1 ) 0 0 2 1 6 - 0
BBAEXP 91546 29-5-01
H. Itoh et al. / Biochimica et Biophysica Acta 1519 (2001) 92^95
93
Fig. 1. Comparable schematic representation of mHAI-1 and mHAI-2 gene structures with their cDNAs. The organization of mHAI-1 (upper portion) and mHAI-2 (lower portion) genes is shown with corresponding cDNAs. The locations of exons are indicated in black boxes with the exon numbers. The portions of cDNAs corresponding to each exon with approximate DNA sizes are also indicated. KDs of HAI-1 and -2 are coded by exons 5 and 9 and exons 2 and 6, respectively, so that three exons are inserted between KD1 and KD2 of each gene. The last exon of each gene (exon 11 of mHAI-1 or exon 8 of mHAI-2) encodes the putative transmembrane domain (TM) and following COOH-terminal region and the 3P-non-coding region including polyadenylation signal. Exon 4 of mHAI-2 gene is speci¢cally transcribed in the testis as described previously [11]. Note that the overall gene structure of mHAI-1 including exon and intron boundaries was very similar to that of mHAI-2. Other abbreviations used are as follows ; SP: signal peptide, LDL: LDL receptor-like domain.
man HAI-1 (hHAI-1) and hHAI-2/PB and mapped them to chromosome 15q15 and 19q13.11, respectively [12]. In this study, in order to develop an in vivo model for the examination of unknown functions of these proteins and explore the di¡erence of major HAI-2 transcripts between human and mouse, mouse genes encoding HAI-1 (mHAI1) and HAI-2/PB (mHAI-2) were cloned by screening of a mouse genomic bacterial arti¢cial chromosome (BAC) library (Genome Systems ; St. Louis, MO, USA) and by polymerase chain reaction (PCR) of mouse genomic DNA (Clontech, Palo Alto, CA, USA) using LA Taq DNA polymerase (Takara, Otsu, Japan), respectively. Based on the sequence data and the restriction maps, mHAI-1 and mHAI-2 genes were de¢ned to consist of 11 exons spanning approximately 11 kbp and of eight exons spanning approximately 9.5 kbp, respectively (Figs. 1 and 2). Transcription start sites were assigned to the position at 204 bp (mHAI-1) and 281 bp (mHAI-2) upstream from start codons by cDNA/gene sequence matches from the data of 5P-rapid ampli¢cation of cDNA ends using Marathon-Ready1 mouse kidney cDNA (Clontech). Neither TATA boxes, CAAT boxes, nor de¢nitive homologous regions between HAI-1 and -2 were found in 5P-£anking regions of both genes (not shown). Computer database (TRANSFAC v3.2) search of 5P-£anking regions revealed the several potential binding sites for known transcription factors including GATA-1 and -2, MZF-1, CdxA, AML1a, SRY, c-Ets, v-Myb, C/EBP, and Sp1, in both genes. Consensus binding sites of CP-2, NF-UB, Ik-1, -2, and -3, Lyf-1, STATx, AhR/Ar and Egr-1, -2, and -3 were observed in mHAI-1, while those of deltaE, Evi-1, and CREB were in mHAI-2. The presence of some Sp1 binding sites and absence of TATA and CAAT boxes suggest
that both mHAI-1 and mHAI-2 might be house-keeping genes. This may be in accordance with their ubiquitous expression in various tissues. However, no de¢nitive homologous regions between mHAI-1 and mHAI-2 were found in 5P-£anking regions. Moreover, potential binding sites for known transcription factors other than Sp1 and GATAs were obviously di¡erent from each other, suggesting that these genes might be under separate transcriptional regulations. Characteristics of 5P-£anking regions of mHAI-1 and mHAI-2 were similar to those of human counterparts [12]. Similar to hHAI-1, the possible binding sites for some early responsive factors expressed in case of tissue injury, such as NF-UB, and Egrs, were also seen in the 5P£anking region of mHAI-1 but not mHAI-2. NF-UB is a transcription factor involved in the induction of several acute-phase proteins [13]. Egrs are early growth response factors which are expressed in the injured tissues [14,15]. Moreover, the potential binding sites of a complex of Egr1^3 and Sp1 were well-conserved between human (342 to 358) and mouse (344 to 357) (Fig. 3A) [12]. Promoter activities of mHAI-1 and hHAI-1 were increased in the constructs containing this region (hatched boxes in Fig. 3B), although they were minimal in the constructs without this region. Thus, this portion of 5P-£anking region of HAI-1 gene might be a key portion of its transcriptional regulation to function as not only house-keeping but also early responsive genes. Promoter activity of HAI-1 might be increased by some other factors, because the increased luciferase activity was observed in the constructs containing long 5P-£anking regions. On the other hand, no strict conserved region, like HAI-1, was found in the 5P-£anking region of mHAI-2 and hHAI-2 genes. HAI-2 lacking KD1
BBAEXP 91546 29-5-01
94
H. Itoh et al. / Biochimica et Biophysica Acta 1519 (2001) 92^95
Fig. 2. Exon and intron boundaries of mHAI-1 (A) and mHAI-2 (B) genes. Two BamHI-digested DNA fragments (8 kbp and 12 kbp) obtained from BAC clone positive for mHAI-1 were subcloned into pBluescript II SK+ phagemid vectors (Stratagene) and sequenced by ABI PRISM1 310 Genetic Analyzer. Since mHAI-2 gene was partially identi¢ed previously [11], the remaining portion of mHAI-2 gene was obtained by PCR of mouse genomic DNA. Both 5P- and 3P-ends of mHAI-2 gene were obtained by nested PCR using GenomeWalker1 kits (Clontech). These PCR products were directly sequenced using inner primers. Introns (lowercase letters) and exons (capital letters) and the deduced amino acid sequences are shown with the nucleotide positions of previously published mHAI-1 [10] and mHAI-2 [11] cDNA sequences. Approximate intron sizes are also indicated. All of the splice junctions agreed with the GT-AG rule [19]. Polyadenylation signals of last exons are indicated in double underlines.
by an alternative splicing is a predominant transcript in mouse but not in human [11], but its mechanism could not be speculated from the comparison of gene structures of mHAI-2 and hHAI-2. Recently, we demonstrated that an active form of HGFA is speci¢cally complexed with membrane-form HAI-1 but not with HAI-2, and HGFA/HAI-1 complex is quickly released from the cell surface by treatment of IL-1L with a signi¢cant recovery of HGFA activity [16]. Therefore, HAI-1 might be not only an inhibitor but also a speci¢c acceptor of active HGFA, acting as a reservoir of this enzyme on the cell surface. Moreover, we also
demonstrated that HGFA is expressed in colorectal carcinoma in association with down-regulation of HAI-1 [17,18]. HAI-1 but not HAI-2 was up-regulated during the regeneration of damaged gastrointestinal mucosa [10] as well as in severely injured kidney and liver [9]. Therefore, HAI-1 and HAI-2 might play quite di¡erent roles in vivo, although these two proteins have similar molecular structure and e¤ciently inhibit HGFA in vitro. Based on the present gene data, further studies for HAI-1 and -2 proteins by generating model mice and extensive promoter assay of these genes would be necessary in order to understand in vivo functions of these proteins.
BBAEXP 91546 29-5-01
H. Itoh et al. / Biochimica et Biophysica Acta 1519 (2001) 92^95
95
Fig. 3. (A) Comparison of 5P-£anking region of mHAI-1 with that of hHAI-1 and (B) transient expression of mHAI-1 and hHAI-1 promoter constructs in HeLa and WiDr cells. (A) Identical nucleotides were shown in asterisks. Potential binding sites of a complex of Egr-1^3 and Sp1 were well-conserved between human (342 to 358) and mouse (344 to 357). (B) Transient transfection assay was performed using Dual-Luciferase Reporter Assay System (Promega). Promoter constructs containing the indicating nucleotide sequences of mHAI-1 and hHAI-1 5P-£anking regions were made by PCR and connected to the promoterless luciferase gene vector (pGL3 basic). Transient transfection was performed by the lipofectin method using Tfx1-20 reagent (Promega). HeLa and WiDr cells (RIKEN Cell Bank, Wako, Japan), both of which produce abundant endogenous HAI-1, were cultured for 48 h after transfection, to assay for reporter gene expression. Results of promoter activity were expressed as a percentage of luciferase activity compared to cells transfected with the maximal promoter construct driven by SV-40 promoter and enhancer (pGL3 control). Luciferase activity was measured by Fluoroskan Ascent FL (Labsystems, Helsinki, Finland) as relative light units adjusted for e¤ciency of transfection standardized by co-transfection with a pRLTK inner control vector. Results were the average of three independent transfections and are expressed as the mean þ S.E.M. Note that the promoter activities of both mHAI-1 and hHAI-1 were increased in the constructs containing the potential binding sites of a complex of Egr-1^3 and Sp1 (hatched boxes), although they were minimal in the constructs without this region.
The authors wish to thank Mr. T. Miyamoto for preparation of ¢gures. This work was supported in part by Grant-in-Aid for Scienti¢c Research (C) No. 11670221 and No. 12670209, from the Ministry of Education, Science, Sports and Culture, Japan, and grants from the Osaka Cancer Research Foundation, the Sagawa Cancer Research Foundation, and the Uehara Memorial Foundation. References [1] E. Gohda, H. Tsubouchi, H. Nakayama, S. Hirono, O. Sakiyama, K. Takahashi, H. Miyazaki, S. Hashimoto, Y. Daikuhara, J. Clin. Invest. 81 (1988) 414^419. [2] M. Stoker, E. Gherardi, M. Perryman, J. Gray, Nature 327 (1987) 239^242. [3] K. Matsumoto, T. Nakamura, J. Biochem. 119 (1996) 591^600. [4] K. Miyazawa, T. Shimomura, A. Kitamura, J. Kondo, Y. Morimoto, N. Kitamura, J. Biol. Chem. 268 (1993) 10024^10028. [5] H. Itoh, R. Hamasuna, H. Kataoka, M. Yamauchi, K. Miyazawa, N. Kitamura, M. Koono, Biochim. Biophys. Acta 1491 (2000) 295^302. [6] T. Shimomura, K. Denda, A. Kitamura, T. Kawaguchi, M. Kito, J. Kondo, S. Kagaya, L. Qin, H. Takata, K. Miyazawa, N. Kitamura, J. Biol. Chem. 272 (1997) 6370^6376.
[7] T. Kawaguchi, L. Qin, T. Shimomura, J. Kondo, K. Matsumoto, K. Denda, N. Kitamura, J. Biol. Chem. 272 (1997) 27558^27564. [8] C.W. Marlor, K.A. Delaria, G. Davis, D.K. Muller, J.M. Greve, P.P. Tamburini, J. Biol. Chem. 272 (1997) 12202^12208. [9] H. Kataoka, T. Suganuma, T. Shimomura, H. Itoh, K. Denda, N. Kitamura, K. Nabeshima, M. Koono, J. Histochem. Cytochem. 47 (1999) 673^682. [10] H. Itoh, H. Kataoka, M. Tomita, R. Hamasuna, Y. Nawa, N. Kitamura, M. Koono, Am. J. Phys. 278 (2000) G635^G643. [11] H. Itoh, H. Kataoka, R. Hamasuna, N. Kitamura, M. Koono, Biochem. Biophys. Res. Commun. 255 (1999) 740^748. [12] H. Itoh, M. Yamauchi, H. Kataoka, R. Hamasuna, N. Kitamura, M. Koono, Eur. J. Biochem. 267 (2000) 3351^3359. [13] S. Grimm, P.A. Baeuerle, Biochem. J. 290 (1993) 297^308. [14] L.M. Khachigian, V. Lindner, A.J. Williams, T. Collins, Science 271 (1996) 1427^1431. [15] S.F. Yan, T. Fujita, J. Lu, K. Okada, Y. ShanZou, N. Mackman, D.J. Pinsky, D.M. Stern, Nat. Med. 12 (2000) 1355^1356. [16] H. Kataoka, T. Shimomura, T. Kawaguchi, R. Hamasuna, H. Itoh, N. Kitamura, K. Miyazawa, M. Koono, J. Biol. Chem. 275 (2000) 40453^40462. [17] H. Kataoka, H. Uchino, K. Denda, N. Kitamura, H. Itoh, H. Tsubouchi, K. Nabeshima, M. Koono, Cancer lett. 128 (1998) 219^227. [18] H. Kataoka, R. Hamasuna, H. Itoh, N. Kitamura, M. Koono, Cancer Res. 60 (2000) 6148^6159. [19] G. Lennon, C. Au¡ray, M. Polymeropoulos, M.B. Soares, Genomics 33 (1996) 151^152.
BBAEXP 91546 29-5-01
www.bba-direct.com
Promoter paper
Mouse hepatocyte growth factor activator inhibitor type 1 (HAI-1) and type 2 (HAI-2)/placental bikunin genes and their promoters1 Hiroshi Itoh b
a;
*, Hiroaki Kataoka a , Jing Yan Meng a , Ryouichi Hamasuna a , Naomi Kitamura b , Masashi Koono a
a Second Department of Pathology, Miyazaki Medical College, Kiyotake, Miyazaki 889-1692, Japan Department of Biological Sciences, Graduate School of Bioscience and Biotechnology, Tokyo Institute of Technology, Yokohama, Japan
Received 21 December 2000 ; received in revised form 19 March 2001; accepted 26 March 2001
Abstract Hepatocyte growth factor (HGF) activator inhibitor type 1 (HAI-1) and type 2 (HAI-2) were recently discovered as specific inhibitors of HGF activator. Each of them contains two Kunitz-type serine protease inhibitor domains and a transmembrane domain, so that their overall structures are similar to each other. In this study, mouse genes encoding HAI-1 and HAI-2 were cloned by screening of a mouse genomic bacterial artificial chromosome library and by polymerase chain reaction of mouse genomic DNA, respectively. The genes (mHAI1 and mHAI-2) were defined to consist of 11 and eight exons spanning 11 kbp and 9.5 kbp, respectively. Neither a TATA nor CAAT box was found in 5P-flanking regions of both genes and no apparent homologous portion was observed between mHAI-1 and mHAI-2 promoter regions. Promoter assay of mHAI-1 and human HAI-1 revealed that the potential binding sites of a complex of Egr-1^3 and Sp1, which was well-conserved between human (342 to 358) and mouse (344 to 357), might be a key portion of its transcriptional regulation to function as not only house-keeping but also early responsive genes. ß 2001 Elsevier Science B.V. All rights reserved. Keywords : Hepatocyte growth factor; Hepatocyte growth factor activator inhibitor ; Placental bikunin ; Gene structure; Mouse
Hepatocyte growth factor (HGF)/scatter factor (SF) is a multifunctional polypeptide factor that functions as a mitogen, morphogen, and/or motogen for a wide variety of cells [1^3]. HGF/SF is secreted by mesenchymal cells as an inactive precursor form and activated by the speci¢c serine proteinase, HGF activator (HGFA) [4,5]. The activity of HGFA is regulated by two recently identi¢ed serine proteinase inhibitors, namely HGF activator inhibitor type 1 (HAI-1) and type 2 (HAI-2), both of which were initially puri¢ed from the conditioned medium of a human stomach cancer cell line MKN45 [6,7]. They are structurally similar to each other and each contains two Kunitz-type
Abbreviations : HGF, hepatocyte growth factor; HGFA, HGF activator; HAI, HGF activator inhibitor; mHAI, mouse HAI; PB, placental bikunin; PCR, polymerase chain reaction * Corresponding author. Fax: +81-985-85-6003; E-mail : [email protected] 1 The nucleotide sequence data of 5P-£anking regions reported in this paper will appear in the GSDB, DDBJ, EMBL and NCBI nucleotide sequence databases with the following accession numbers, AY013701 (mHAI-1) and AY013702 (mHAI-2).
serine protease inhibitor domains (KDs) and a single putative transmembrane domain [6,7]. Among them, HAI-2 was identical with the protein originally reported as placental bikunin (PB) [8]. Both HAI-1 and HAI-2/PB genes are expressed abundantly in various tissues including the placenta, kidney, pancreas and gastrointestinal tract as assessed by RNA blot analysis [6,7]. Immunohistochemically, HAI-1 protein is localized on cellular lateral (or baso-lateral) surface of a simple columnar epithelium of ducts, tubules and mucosal surface of various organs [9], while HAI-2/PB protein is mainly observed in the cytoplasms of the epithelial cells and macrophage-like monocytic in£ammatory cells of various tissues (unpublished observation). HAI-1 but not HAI-2/PB is up-regulated during the regeneration of damaged gastrointestinal mucosa, liver and kidney [9,10]. These results strongly suggest that HAI-1 and -2 may have di¡erent and distinctive roles in vivo, although they are structurally similar to each other. However, precise biophysiological functions of HAI-1 and -2 in vivo remain to be elucidated. Recently, we showed that the HAI-2 lacking KD1 by an alternative splicing is a predominant transcript in mouse but not in human [11] and subsequently we cloned the genes for hu-
0167-4781 / 01 / $ ^ see front matter ß 2001 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 7 - 4 7 8 1 ( 0 1 ) 0 0 2 1 6 - 0
BBAEXP 91546 29-5-01
H. Itoh et al. / Biochimica et Biophysica Acta 1519 (2001) 92^95
93
Fig. 1. Comparable schematic representation of mHAI-1 and mHAI-2 gene structures with their cDNAs. The organization of mHAI-1 (upper portion) and mHAI-2 (lower portion) genes is shown with corresponding cDNAs. The locations of exons are indicated in black boxes with the exon numbers. The portions of cDNAs corresponding to each exon with approximate DNA sizes are also indicated. KDs of HAI-1 and -2 are coded by exons 5 and 9 and exons 2 and 6, respectively, so that three exons are inserted between KD1 and KD2 of each gene. The last exon of each gene (exon 11 of mHAI-1 or exon 8 of mHAI-2) encodes the putative transmembrane domain (TM) and following COOH-terminal region and the 3P-non-coding region including polyadenylation signal. Exon 4 of mHAI-2 gene is speci¢cally transcribed in the testis as described previously [11]. Note that the overall gene structure of mHAI-1 including exon and intron boundaries was very similar to that of mHAI-2. Other abbreviations used are as follows ; SP: signal peptide, LDL: LDL receptor-like domain.
man HAI-1 (hHAI-1) and hHAI-2/PB and mapped them to chromosome 15q15 and 19q13.11, respectively [12]. In this study, in order to develop an in vivo model for the examination of unknown functions of these proteins and explore the di¡erence of major HAI-2 transcripts between human and mouse, mouse genes encoding HAI-1 (mHAI1) and HAI-2/PB (mHAI-2) were cloned by screening of a mouse genomic bacterial arti¢cial chromosome (BAC) library (Genome Systems ; St. Louis, MO, USA) and by polymerase chain reaction (PCR) of mouse genomic DNA (Clontech, Palo Alto, CA, USA) using LA Taq DNA polymerase (Takara, Otsu, Japan), respectively. Based on the sequence data and the restriction maps, mHAI-1 and mHAI-2 genes were de¢ned to consist of 11 exons spanning approximately 11 kbp and of eight exons spanning approximately 9.5 kbp, respectively (Figs. 1 and 2). Transcription start sites were assigned to the position at 204 bp (mHAI-1) and 281 bp (mHAI-2) upstream from start codons by cDNA/gene sequence matches from the data of 5P-rapid ampli¢cation of cDNA ends using Marathon-Ready1 mouse kidney cDNA (Clontech). Neither TATA boxes, CAAT boxes, nor de¢nitive homologous regions between HAI-1 and -2 were found in 5P-£anking regions of both genes (not shown). Computer database (TRANSFAC v3.2) search of 5P-£anking regions revealed the several potential binding sites for known transcription factors including GATA-1 and -2, MZF-1, CdxA, AML1a, SRY, c-Ets, v-Myb, C/EBP, and Sp1, in both genes. Consensus binding sites of CP-2, NF-UB, Ik-1, -2, and -3, Lyf-1, STATx, AhR/Ar and Egr-1, -2, and -3 were observed in mHAI-1, while those of deltaE, Evi-1, and CREB were in mHAI-2. The presence of some Sp1 binding sites and absence of TATA and CAAT boxes suggest
that both mHAI-1 and mHAI-2 might be house-keeping genes. This may be in accordance with their ubiquitous expression in various tissues. However, no de¢nitive homologous regions between mHAI-1 and mHAI-2 were found in 5P-£anking regions. Moreover, potential binding sites for known transcription factors other than Sp1 and GATAs were obviously di¡erent from each other, suggesting that these genes might be under separate transcriptional regulations. Characteristics of 5P-£anking regions of mHAI-1 and mHAI-2 were similar to those of human counterparts [12]. Similar to hHAI-1, the possible binding sites for some early responsive factors expressed in case of tissue injury, such as NF-UB, and Egrs, were also seen in the 5P£anking region of mHAI-1 but not mHAI-2. NF-UB is a transcription factor involved in the induction of several acute-phase proteins [13]. Egrs are early growth response factors which are expressed in the injured tissues [14,15]. Moreover, the potential binding sites of a complex of Egr1^3 and Sp1 were well-conserved between human (342 to 358) and mouse (344 to 357) (Fig. 3A) [12]. Promoter activities of mHAI-1 and hHAI-1 were increased in the constructs containing this region (hatched boxes in Fig. 3B), although they were minimal in the constructs without this region. Thus, this portion of 5P-£anking region of HAI-1 gene might be a key portion of its transcriptional regulation to function as not only house-keeping but also early responsive genes. Promoter activity of HAI-1 might be increased by some other factors, because the increased luciferase activity was observed in the constructs containing long 5P-£anking regions. On the other hand, no strict conserved region, like HAI-1, was found in the 5P-£anking region of mHAI-2 and hHAI-2 genes. HAI-2 lacking KD1
BBAEXP 91546 29-5-01
94
H. Itoh et al. / Biochimica et Biophysica Acta 1519 (2001) 92^95
Fig. 2. Exon and intron boundaries of mHAI-1 (A) and mHAI-2 (B) genes. Two BamHI-digested DNA fragments (8 kbp and 12 kbp) obtained from BAC clone positive for mHAI-1 were subcloned into pBluescript II SK+ phagemid vectors (Stratagene) and sequenced by ABI PRISM1 310 Genetic Analyzer. Since mHAI-2 gene was partially identi¢ed previously [11], the remaining portion of mHAI-2 gene was obtained by PCR of mouse genomic DNA. Both 5P- and 3P-ends of mHAI-2 gene were obtained by nested PCR using GenomeWalker1 kits (Clontech). These PCR products were directly sequenced using inner primers. Introns (lowercase letters) and exons (capital letters) and the deduced amino acid sequences are shown with the nucleotide positions of previously published mHAI-1 [10] and mHAI-2 [11] cDNA sequences. Approximate intron sizes are also indicated. All of the splice junctions agreed with the GT-AG rule [19]. Polyadenylation signals of last exons are indicated in double underlines.
by an alternative splicing is a predominant transcript in mouse but not in human [11], but its mechanism could not be speculated from the comparison of gene structures of mHAI-2 and hHAI-2. Recently, we demonstrated that an active form of HGFA is speci¢cally complexed with membrane-form HAI-1 but not with HAI-2, and HGFA/HAI-1 complex is quickly released from the cell surface by treatment of IL-1L with a signi¢cant recovery of HGFA activity [16]. Therefore, HAI-1 might be not only an inhibitor but also a speci¢c acceptor of active HGFA, acting as a reservoir of this enzyme on the cell surface. Moreover, we also
demonstrated that HGFA is expressed in colorectal carcinoma in association with down-regulation of HAI-1 [17,18]. HAI-1 but not HAI-2 was up-regulated during the regeneration of damaged gastrointestinal mucosa [10] as well as in severely injured kidney and liver [9]. Therefore, HAI-1 and HAI-2 might play quite di¡erent roles in vivo, although these two proteins have similar molecular structure and e¤ciently inhibit HGFA in vitro. Based on the present gene data, further studies for HAI-1 and -2 proteins by generating model mice and extensive promoter assay of these genes would be necessary in order to understand in vivo functions of these proteins.
BBAEXP 91546 29-5-01
H. Itoh et al. / Biochimica et Biophysica Acta 1519 (2001) 92^95
95
Fig. 3. (A) Comparison of 5P-£anking region of mHAI-1 with that of hHAI-1 and (B) transient expression of mHAI-1 and hHAI-1 promoter constructs in HeLa and WiDr cells. (A) Identical nucleotides were shown in asterisks. Potential binding sites of a complex of Egr-1^3 and Sp1 were well-conserved between human (342 to 358) and mouse (344 to 357). (B) Transient transfection assay was performed using Dual-Luciferase Reporter Assay System (Promega). Promoter constructs containing the indicating nucleotide sequences of mHAI-1 and hHAI-1 5P-£anking regions were made by PCR and connected to the promoterless luciferase gene vector (pGL3 basic). Transient transfection was performed by the lipofectin method using Tfx1-20 reagent (Promega). HeLa and WiDr cells (RIKEN Cell Bank, Wako, Japan), both of which produce abundant endogenous HAI-1, were cultured for 48 h after transfection, to assay for reporter gene expression. Results of promoter activity were expressed as a percentage of luciferase activity compared to cells transfected with the maximal promoter construct driven by SV-40 promoter and enhancer (pGL3 control). Luciferase activity was measured by Fluoroskan Ascent FL (Labsystems, Helsinki, Finland) as relative light units adjusted for e¤ciency of transfection standardized by co-transfection with a pRLTK inner control vector. Results were the average of three independent transfections and are expressed as the mean þ S.E.M. Note that the promoter activities of both mHAI-1 and hHAI-1 were increased in the constructs containing the potential binding sites of a complex of Egr-1^3 and Sp1 (hatched boxes), although they were minimal in the constructs without this region.
The authors wish to thank Mr. T. Miyamoto for preparation of ¢gures. This work was supported in part by Grant-in-Aid for Scienti¢c Research (C) No. 11670221 and No. 12670209, from the Ministry of Education, Science, Sports and Culture, Japan, and grants from the Osaka Cancer Research Foundation, the Sagawa Cancer Research Foundation, and the Uehara Memorial Foundation. References [1] E. Gohda, H. Tsubouchi, H. Nakayama, S. Hirono, O. Sakiyama, K. Takahashi, H. Miyazaki, S. Hashimoto, Y. Daikuhara, J. Clin. Invest. 81 (1988) 414^419. [2] M. Stoker, E. Gherardi, M. Perryman, J. Gray, Nature 327 (1987) 239^242. [3] K. Matsumoto, T. Nakamura, J. Biochem. 119 (1996) 591^600. [4] K. Miyazawa, T. Shimomura, A. Kitamura, J. Kondo, Y. Morimoto, N. Kitamura, J. Biol. Chem. 268 (1993) 10024^10028. [5] H. Itoh, R. Hamasuna, H. Kataoka, M. Yamauchi, K. Miyazawa, N. Kitamura, M. Koono, Biochim. Biophys. Acta 1491 (2000) 295^302. [6] T. Shimomura, K. Denda, A. Kitamura, T. Kawaguchi, M. Kito, J. Kondo, S. Kagaya, L. Qin, H. Takata, K. Miyazawa, N. Kitamura, J. Biol. Chem. 272 (1997) 6370^6376.
[7] T. Kawaguchi, L. Qin, T. Shimomura, J. Kondo, K. Matsumoto, K. Denda, N. Kitamura, J. Biol. Chem. 272 (1997) 27558^27564. [8] C.W. Marlor, K.A. Delaria, G. Davis, D.K. Muller, J.M. Greve, P.P. Tamburini, J. Biol. Chem. 272 (1997) 12202^12208. [9] H. Kataoka, T. Suganuma, T. Shimomura, H. Itoh, K. Denda, N. Kitamura, K. Nabeshima, M. Koono, J. Histochem. Cytochem. 47 (1999) 673^682. [10] H. Itoh, H. Kataoka, M. Tomita, R. Hamasuna, Y. Nawa, N. Kitamura, M. Koono, Am. J. Phys. 278 (2000) G635^G643. [11] H. Itoh, H. Kataoka, R. Hamasuna, N. Kitamura, M. Koono, Biochem. Biophys. Res. Commun. 255 (1999) 740^748. [12] H. Itoh, M. Yamauchi, H. Kataoka, R. Hamasuna, N. Kitamura, M. Koono, Eur. J. Biochem. 267 (2000) 3351^3359. [13] S. Grimm, P.A. Baeuerle, Biochem. J. 290 (1993) 297^308. [14] L.M. Khachigian, V. Lindner, A.J. Williams, T. Collins, Science 271 (1996) 1427^1431. [15] S.F. Yan, T. Fujita, J. Lu, K. Okada, Y. ShanZou, N. Mackman, D.J. Pinsky, D.M. Stern, Nat. Med. 12 (2000) 1355^1356. [16] H. Kataoka, T. Shimomura, T. Kawaguchi, R. Hamasuna, H. Itoh, N. Kitamura, K. Miyazawa, M. Koono, J. Biol. Chem. 275 (2000) 40453^40462. [17] H. Kataoka, H. Uchino, K. Denda, N. Kitamura, H. Itoh, H. Tsubouchi, K. Nabeshima, M. Koono, Cancer lett. 128 (1998) 219^227. [18] H. Kataoka, R. Hamasuna, H. Itoh, N. Kitamura, M. Koono, Cancer Res. 60 (2000) 6148^6159. [19] G. Lennon, C. Au¡ray, M. Polymeropoulos, M.B. Soares, Genomics 33 (1996) 151^152.
BBAEXP 91546 29-5-01