- Email: [email protected]
Nucleotide Pyrophosphatase Gene (PDNP2)
SHORT COMMUNICATION Molecular Cloning and Chromosomal Assignment of the Human Brain-Type Phosphodiesterase I/Nucleotide Pyrophosphatase Gene (PDNP2) HIROYUKI KAWAGOE, OSAMU SOMA, JUNKO GOJI, NORIYUKI NISHIMURA, MASAAKI NARITA, JOJI INAZAWA,* HAJIME NAKAMURA, AND KIMIHIKO SANO1 Department of Pediatrics, Kobe University School of Medicine, Kobe 650, Japan; and *Department of Hygiene, Kyoto Prefectural Medical College, Kyoto 602, Japan Received March 13, 1995; accepted July 26, 1995
Phosphodiesterase I/nucleotide pyrophosphatase is a widely expressed membrane-bound enzyme that cleaves diester bonds of a variety of substrates. We have cloned brain-type cDNA for this enzyme from rat brain and designated it PD-Ia (M. Narita, J. Goji, H. Nakamura, and K. Sano, 1994, J. Biol. Chem. 269: 28235– 28242). In this study we have isolated cDNA and genomic DNA encoding human PD-Ia. Human PD-Ia cDNA, designated PDNP2 in HGMW nomenclature, has a 2589-nucleotide open reading frame encoding a polypeptide of 863 amino acids with a calculated Mr of 99,034. Northern blot analysis revealed that human PD-Ia transcript was present in brain, lung, placenta, and kidney. The database analysis showed that human PD-Ia was identical with human autotaxin (ATX), a novel tumor motility-stimulating factor, except that human PD-Ia lacks 156 nucleotides and 52 amino acids of human ATX. Human PD-Ia and human ATX are likely to be alternative splicing products from the same gene. The 5* region of the human PDNP2 gene contains four putative binding sites of transcription factor Sp1 without typical TATA or CAAT boxes, and there is a potential octamer binding motif in intron 2. From the results of fluorescence in situ hybridization, the human PDNP2 gene is located at chromosome 8q24.1. q 1995 Academic Press, Inc.
Phosphodiesterase I (EC 3.1.4.1)/nucleotide pyrophosphatase (EC 3.6.1.9) is a widely expressed ectoenzyme with a reported molecular weight of 100–130 This work was supported by a grant for Research on Mental Diseases from the Ministry of Health and Welfare, Japan (1994) and by a fund of Cancer Research from Hyogo Total Health Association (1994). The nucleotide sequences reported in this paper have been submitted to the DNA Data Bank of Japan under Accession Nos. D45421 for cDNA and D45914 for genomic DNA. 1 To whom correspondence should be addressed at the Department of Pediatrics, Kobe University School of Medicine, 7-5-1 Kusunokicho, Chuo-ku, Kobe 650, Japan. Telephone: 81-78-341-7451, ext. 5722. Fax: 81-78-371-6239. GENOMICS
30, 380– 384 (1995)
0888-7543/95 $12.00 Copyright q 1995 by Academic Press, Inc. All rights of reproduction in any form reserved.
/ m4275f3678
11-03-95 14:41:39
kDa. A single protein is responsible for both phosphodiesterase and pyrophosphatase enzymatic activities (3). Goding and his colleagues have cloned plasma cell membrane glycoprotein PC-1 from mouse and human plasma cell lines (2, 16). PC-1 was later shown to have alkaline phosphodiesterase I/nucleotide pyrophosphatase activity (14). Northern blot analysis showed that PC-1 is expressed in liver but not in brain (16). PC-1like immunoreactivity was found in epididymis, distal convoluted tubules of kidney, chondrocytes, and hepatocytes (6). The human PC-1 gene has been mapped to 6q22–q23 (2). Increased PC-1 expression was reported in fibroblasts obtained from Lowe syndrome (7), neurofibromatosis type I (11), and non-insulin-dependent diabetes mellitus patients (10). Although the involvement of PC-1 in the pathophysiology of these diseases has not been elucidated, Maddux et al. showed that PC-1 inhibits insulin-receptor kinase activity (10). Gross et al. separated phosphodiesterase I activity by native polyacrylamide gel electrophoresis and reported that five isozymes exist in various mouse organs (5). In a previous study we cloned brain-type phosphodiesterase I/nucleotide pyrophosphatase cDNA from rat brain and designated it PD-Ia (13). The deduced amino acid sequence predicts that rat PD-Ia is a type II (cytoplasmic amino terminus) membrane glycoprotein as proposed for PC-1 and that rat PD-Ia and mouse PC1 share 58% homology. Northern blot analysis showed that the PD-Ia gene is abundantly expressed in the central nervous system. In situ hybridization analysis demonstrated that PD-Ia mRNA is predominantly localized in choroid plexus epithelial cells, pigment epithelial cells of the eye, and cells in the granular layer of the cerebellum (13). A weak signal was observed in the cells in corpus callosum and other brain regions. In this study, we attempted the cloning of human cDNA and genomic DNA encoding PD-Ia as a first step toward understanding the mechanism of tissue-specific expression of this gene. To isolate the clones that cover the entire coding sequence of human PD-Ia, we screened the human retina
380
gnma
AP-Genomics
SHORT COMMUNICATION
cDNA library using the 5* and 3* fragments of rat PDIa cDNA as probes. Approximately 2 1 10 5 phages were screened and three positive clones were obtained. These three clones overlapped each other and covered the whole coding region of human PD-Ia. We designated this cDNA human PD-Ia (HGMW-approved symbol is PDNP2). The cDNA clones spanned a stretch of 2.7 kb. The cDNA sequence and the deduced amino acid sequence of human PD-Ia were submitted to the DNA Data Bank (Accession No. D45421). The nucleotide sequence of 2738 bp contains a small 33-bp 5*untranslated leader followed by a single open reading frame of 2589 bp encoding a polypeptide of 863 amino acids with a calculated molecular mass of 99,034 Da and by a 116-bp 3*-untranslated region. The entire primary structure contains four potential N-linked glycosylation sites. In the middle of the sequence, there is a putative Ca2/ binding site ‘‘EF hand’’-like motif situated at amino acids 740–748. Analysis of hydropathy using the method of Kyte and Doolittle confirmed that only one region of 22 residues, situated between residues 9 and 30, appears to have sufficient hydrophobicity to span the lipid layer. It is likely that human PD-Ia protein crosses the membrane only once. In addition, the human PD-Ia protein has a region nearly identical to the rat PD-Ia enzymatic domain (13) situated at amino acids 172– 219. Human PD-Ia protein has homology with the somatomedin Blike domain of vitronectin near the amino terminus of the extracellular portion, and both proteins have the cell attachment site ‘‘RGD’’ motif (human PD-Ia; 127 – 129 amino acids; vitronectin, 64–66 amino acids). This tripeptide is recognized by integrins including a5 b1 , aIIbb3 , and most of avb . Among RGD-recognizing integrins, there is a selectivity in natural ligand recognition. For example, fibronectin-specific receptor, which interacts with the ‘‘RGD’’ sequence in fibronectin, did not bind to vitronectin. Potentially, human PD-Ia protein may interact with the cell surface receptor, as do other RGD sequence-containing proteins that promote cell adhesion. Overall, the identity between human PD-Ia protein and rat PD-Ia protein is 89% and the overall structure of human PD-Ia is closely related to that of rat PDIa. The stretch of 16 amino acids of the human PD-Ia protein (148 –163) is not identical to that of rat PD-Ia. The amino acid sequence of the putative catalytic site is well conserved among the bovine intestine phosphodiesterase I, rat PD-Ia, and human PD-Ia. The overall structure of human PD-Ia is homologous to that of the human PC-1 protein. These results indicate that these proteins have similar enzymatic activity and physiological functions. The database searches of nucleotide and amino acid sequences reveal that human PD-Ia is identical with human autotaxin (ATX), a tumor cell motility-stimulating factor, except that human PD-Ia lacks 156 nucleotides (1022 –1177) and 52 amino acids (325 –376) of human ATX (12). Therefore, human PD-Ia and human
/ m4275f3678
11-03-95 14:41:39
gnma
381
FIG. 1. Northern blot analysis of human PD-Ia transcripts. Northern blot analysis of various human tissues was carried out as described previously (13). (Top) mRNA hybridized with a radiolabeled 0.9-kb fragment of human PD-Ia cDNA. In a control experiment, the same RNA samples were probed with b-actin cDNA, (bottom). The tissues or organs of RNA sources are as follows: lane 1, heart; lane 2, brain; lane 3, placenta; lane 4, lung; lane 5, liver; lane 6, skeletal muscle; lane 7, kidney; lane 8, pancreas; lane 9, amygdala; lane 10, caudate nucleus; lane 11, corpus callosum; lane 12, hippocampus; lane 13, hypothalamus; lane 14, substantia nigra; lane 15, subthalamic nucleus; lane 16, thalamus. The size of markers is shown to the right.
ATX are likely to be the alternative splicing products from the same gene. Human ATX is a 125-kDa glycoprotein that was identified in the conditioned medium from A2058 melanoma cells. Human ATX is an autocrine motility factor and stimulates both random and directed migration of human A2058 melanoma cells at picomolar concentrations (15). Human PD-Ia and human ATX are unique, new members of tumor cell motility-stimulating factors. Human PD-Ia lacks 156 nucleotides (1022 –1177) and 52 amino acids (325–376) of human ATX, and human PD-Ia protein lacks 25 amino acids (591 –615) of rat PD-Ia protein. These differences might be attributed to a tissue- or species-specific alternative splicing event. These peptides are located apart from the potentially functional domains of human PD-Ia protein, such as the enzymatic site, EF-hand-like motif, and RGD motif. Therefore, it is likely that these proteins have the same function. However, the possibility that these domains contribute to a tissue-specific function of this protein family cannot be excluded at present. To determine the size and tissue distribution of human PD-Ia mRNA, we carried out Northern blot analysis. The commercially available human multiple-tissue and human brain multiple-tissue Northern blots (Clontech) were probed with a 0.9-kb fragment of human PD-Ia cDNA encoding the 5* part of the coding region.
AP-Genomics
382
SHORT COMMUNICATION
FIG. 2. The nucleotide sequence of the 5* region of the human PDNP2 gene. The DNA sequence was determined from both strands by a dideoxy chain termination method using fluorescence-tagged primers and T7 DNA polymerase according to the protocol recommended by the manufacturer and was analyzed on an automated DNA sequencer (ALF DNA Sequencer II, Pharmacia LKB). The shaded area shows the Alu sequence. Putative Sp1 binding sites are double-underlined. The octamer motif is indicated by a square. Exons 1 and 2 are underlined.
As shown in Fig. 1, a single transcript of approximately 3 kb was identified in brain, placenta, lung, and kidney. To analyze the transcription regulatory region and determine the chromosomal localization of the human PDNP2 gene, we isolated the genomic DNA clone of human PDNP2. We screened approximately 4 1 105 phages of human leukocyte genomic DNA library (Japan Cancer Research Resources Bank), and one positive clone was obtained. This clone was plaque purified and then subjected to further analysis. To obtain the
/ m4275f3678
11-03-95 14:41:39
gnma
genomic DNA fragment containing the 5* untranslated region of human PDNP2, we carried out Southern blot analysis using the above-mentioned probe. The positive 7-kb fragment was obtained, and we sequenced the 5* region of human PDNP2 genomic DNA containing exon 1 (1986–2067) and exon 2 (2290 –2392) (Fig. 2). Since the exact transcription start site is unknown, the start site of exon 1 was tentatively situated at the most 5* site of human ATX cDNA, which was obtained using the rapid amplification of 5* cDNA ends system (12).
AP-Genomics
SHORT COMMUNICATION
383
FIG. 3. Chromosomal mapping of the human PDNP2 gene by fluorescence in situ hybridization. The genomic DNA fragment of human PDNP2 was labeled with biotin– dUTP and hybridized to human mitotic chromosomes. Hybridized biotinylated DNA was detected with avidin –FITC as described previously (9). (Left) Hoechst 33258 staining of chromosomes. (Right) Arrows indicate that the human PDNP2 gene is located at 8q24.1.
Exon 1 encodes the cytoplasmic tail and the proximal part of the transmembrane domain. Exon 2 encodes the remainder of the transmembrane domain and a part of the extracellular region. This exon structure of PDNP2 is similar to that of mouse PC-1 (1). The 5* flanking region has an Alu repetitive sequence (1–70). There is no typical TATA or CAAT box in this region. There are four potential binding sites of transcription factor Sp1 situated at 1115 –1200 (Sp1a), 1207 –1212 (Sp1b), 1212–1217 (Sp1c), and 1817 –1822 (Sp1d) (Fig. 2). Sp1a, -b, and -c are in a reversed orientation. Transcription factor Sp1 has the potential to activate transcription from a number of different viral and cellular promoters acting in conjunction with other cellular transcription factors. Sp1 is also utilized in TATA-boxlacking promoters in combination with other transcription factors. We surmise that Sp1 may act with other transcription factors in the promoter of the human PDNP2 gene. Three of the four putative Sp1 binding sites in the 5* region of the human PDNP2 gene are in a reversed orientation. Sp1-type promoter can activate the transcription in both orientations, as reported for the promoters of the human monoamine oxidase A gene (17). We considered the possibility that the Sp1 promoters of the human PDNP2 gene may act in both orientations. Two putative Sp1 binding sites in reversed orientation, Sp1b and -c, are overlapping (Fig. 2). It might be possible that only one Sp1 binds this region, because the shortest distance between the ‘‘central C’’ of two
/ m4275f3678
11-03-95 14:41:39
gnma
Sp1 binding sites (GGGCGG) should be at least 10 bp for two Sp1 molecules (4). In intron 2 of the human PDNP2 gene, there is an octamer binding motif, ATTTGCAT, situated at 2408– 2415 (Fig. 2), which is a well-characterized cis-acting regulatory sequence found in many promoters and enhancers. The octamer motif has been shown to be a part of both inducible tissue-specific immunoglobulin and ubiquitously expressed histone H2B promoters. In immunoglobulin genes, the octamer motif is located upstream of the transcription start site. On the other hand, the expression of the hst gene, a member of the fibroblast growth factor gene family, is positively regulated by a downstream octamer motif in an embryonal carcinoma cell line, F9 (8). It is conceivable that the expression of the human PDNP2 gene is regulated by a downstream octamer motif that functions as an enhancer. To determine the chromosomal localization of the human PDNP2 gene, we carried out fluorescence in situ hybridization using the genomic DNA clone as a probe. As shown in Fig. 3, the human PDNP2 gene is located at chromosome 8q24.1. In our laboratory the cDNA encoding rat intestinetype phosphodiesterase I/nucleotide pyrophosphatase (PD-Ib) has been cloned, and this protein also has an RGD motif (K. Terashima et al., manuscript in preparation). Therefore, these phosphodiesterase I/nucleotide pyrophosphatase family proteins might have multiple
AP-Genomics
384
SHORT COMMUNICATION
functions, such as cell adhesion, in addition to hydrolysis of nucleotide polyphosphates. Identifying the receptors for these proteins and investigating the regulatory mechanisms of the gene expression will help us to understand further the mode of actions and physiological roles of these proteins.
9.
10.
REFERENCES 1. Belli, S. I., van-Driel, I. R., and Goding, J. W. (1993). Identification and characterization of a soluble form of the plasma cell membrane glycoprotein PC-1 (5*-nucleotide phosphodiesterase). Eur. J. Biochem. 217: 421– 428. 2. Buckley, M. F., Loveland, K. A., McKinstry, W. J., Garson, O. M., and Goding, J. W. (1990). Plasma cell membrane glycoprotein PC-1. cDNA cloning of the human molecule, amino acid sequence, and chromosomal location. J. Biol. Chem. 265: 17506 –17511. 3. Evans, W. H., Hood, D. O., and Gurd, J. W. (1973). Purification and properties of a mouse liver plasma-membrane glycoprotein hydrolysing nucleotide pyrophosphate and phosphodiester bonds. Biochem. J. 135: 819–826. 4. Gidoni, D., Kadonaga, J. T., Barrera-Saldana, H., Takahashi, K., Chambon, P., and Tjian, R. (1985). Bidirectional SV40 transcription mediated by tandem Sp1 binding interactions. Science 230: 511–517. 5. Gross, S. C., Watabe, M., Goodarzi, G., Darabi, A., and Watabe, K. (1990). Organ-specific distribution of isozymes of 5*-nucleotide phosphodiesterase in mouse. Comp. Biochem. Physiol. B 95: 821–824. 6. Harahap, A. R. and Goding, J. W. (1988). Distribution of the murine plasma cell antigen PC-1 in non-lymphoid tissues. J. Immunol. 141: 2317 – 2320. 7. Harper, G. S., Hascall, V. C., Yanagishita, M., and Gahl, W. A. (1987). Proteoglycan synthesis in normal and Lowe syndrome fibroblasts. J. Biol. Chem. 262: 5637 –5643. 8. Hasan, S., Koda, T., and Kakinuma, M. (1994). An upstream NF-Y-binding site is required for transcriptional activation
/ m4275f3678
11-03-95 14:41:39
gnma
11.
12.
13.
14.
15.
16.
17.
from the hst promoter in F9 embryonal carcinoma cells. J. Biol. Chem. 269: 25042 – 25048. Inazawa, J., Saito, H., Ariyama, T., Abe, T., and Nakamura, Y. (1993). High-resolution cytogenetic mapping of 342 new cosmid markers including 43 RFLP markers on human chromosome 17 by fluorescence in situ hybridization. Genomics 17: 153–162. Maddux, B. A., Sbraccia, P., Kumakura, S., Sasson, S., Youngren, J., Fisher, A., Spencer, S., Grupe, A., Henzel, W., Stewart, T. A., Reaven, G. M., and Goldfine, I. D. (1995). Membrane glycoprotein PC-1 and insulin resistance in non-insulin-dependent diabetes mellitus. Nature 373: 448 –451. Maruyama, E., Takashima, S., and Arima, M. (1992). Increase in alkaline phosphodiesterase I activity of tumor-derived cultured fibroblasts from neurofibromatosis patients. Clin. Chim. Acta 211: 181–187. Murata, J., Lee, H. Y., Clair, T., Krutzsch, H. C., Arestad, A. A., Sobel, M. E., Liotta, L. A., and Stracke, M. L. (1994). cDNA cloning of human tumor motility-stimulating protein, Autotaxin, reveals a homology with phosphodiesterases. J. Biol. Chem. 269: 30479 – 30484. Narita, M., Goji, J., Nakamura, H., and Sano, K. (1994). Molecular cloning, expression, and localization of a brain-specific phosphodiesterase I/nucleotide pyrophosphatase (PD-Ia) from rat brain. J. Biol. Chem. 269: 28235 –28242. Rebbe, N. F., Tong, B. D., Finley, E. M., and Hickman, S. (1991). Identification of nucleotide pyrophosphatase/alkaline phosphodiesterase I activity associated with the mouse plasma cell differentiation antigen PC-1. Proc. Natl. Acad. Sci. USA 88: 5192 – 5196. ˚ restad, A., Stracke, M. L., Krutzsch, H. C., Unsworth, E. J., A Cioce, V., Schiffmann, E., and Liotta, L. A. (1992). Identification, purification, and partial sequence analysis of Autotaxin, a novel motility-stimulating protein. J. Biol. Chem. 267: 2524 – 2529. van-Driel, I. R., and Goding, J. W. (1987). Plasma cell membrane glycoprotein PC-1. Primary structure deduced from cDNA clones. J. Biol. Chem. 262: 4882 –4887. Zhu, Q. S., Chen, K., and Shih, J. C. (1994). Bidirectional promoter of human monoamine oxidase A (MAO A) controlled by transcription factor Sp1. J. Neurosci. 14: 7393 –7403.
AP-Genomics
Phosphodiesterase I/nucleotide pyrophosphatase is a widely expressed membrane-bound enzyme that cleaves diester bonds of a variety of substrates. We have cloned brain-type cDNA for this enzyme from rat brain and designated it PD-Ia (M. Narita, J. Goji, H. Nakamura, and K. Sano, 1994, J. Biol. Chem. 269: 28235– 28242). In this study we have isolated cDNA and genomic DNA encoding human PD-Ia. Human PD-Ia cDNA, designated PDNP2 in HGMW nomenclature, has a 2589-nucleotide open reading frame encoding a polypeptide of 863 amino acids with a calculated Mr of 99,034. Northern blot analysis revealed that human PD-Ia transcript was present in brain, lung, placenta, and kidney. The database analysis showed that human PD-Ia was identical with human autotaxin (ATX), a novel tumor motility-stimulating factor, except that human PD-Ia lacks 156 nucleotides and 52 amino acids of human ATX. Human PD-Ia and human ATX are likely to be alternative splicing products from the same gene. The 5* region of the human PDNP2 gene contains four putative binding sites of transcription factor Sp1 without typical TATA or CAAT boxes, and there is a potential octamer binding motif in intron 2. From the results of fluorescence in situ hybridization, the human PDNP2 gene is located at chromosome 8q24.1. q 1995 Academic Press, Inc.
Phosphodiesterase I (EC 3.1.4.1)/nucleotide pyrophosphatase (EC 3.6.1.9) is a widely expressed ectoenzyme with a reported molecular weight of 100–130 This work was supported by a grant for Research on Mental Diseases from the Ministry of Health and Welfare, Japan (1994) and by a fund of Cancer Research from Hyogo Total Health Association (1994). The nucleotide sequences reported in this paper have been submitted to the DNA Data Bank of Japan under Accession Nos. D45421 for cDNA and D45914 for genomic DNA. 1 To whom correspondence should be addressed at the Department of Pediatrics, Kobe University School of Medicine, 7-5-1 Kusunokicho, Chuo-ku, Kobe 650, Japan. Telephone: 81-78-341-7451, ext. 5722. Fax: 81-78-371-6239. GENOMICS
30, 380– 384 (1995)
0888-7543/95 $12.00 Copyright q 1995 by Academic Press, Inc. All rights of reproduction in any form reserved.
/ m4275f3678
11-03-95 14:41:39
kDa. A single protein is responsible for both phosphodiesterase and pyrophosphatase enzymatic activities (3). Goding and his colleagues have cloned plasma cell membrane glycoprotein PC-1 from mouse and human plasma cell lines (2, 16). PC-1 was later shown to have alkaline phosphodiesterase I/nucleotide pyrophosphatase activity (14). Northern blot analysis showed that PC-1 is expressed in liver but not in brain (16). PC-1like immunoreactivity was found in epididymis, distal convoluted tubules of kidney, chondrocytes, and hepatocytes (6). The human PC-1 gene has been mapped to 6q22–q23 (2). Increased PC-1 expression was reported in fibroblasts obtained from Lowe syndrome (7), neurofibromatosis type I (11), and non-insulin-dependent diabetes mellitus patients (10). Although the involvement of PC-1 in the pathophysiology of these diseases has not been elucidated, Maddux et al. showed that PC-1 inhibits insulin-receptor kinase activity (10). Gross et al. separated phosphodiesterase I activity by native polyacrylamide gel electrophoresis and reported that five isozymes exist in various mouse organs (5). In a previous study we cloned brain-type phosphodiesterase I/nucleotide pyrophosphatase cDNA from rat brain and designated it PD-Ia (13). The deduced amino acid sequence predicts that rat PD-Ia is a type II (cytoplasmic amino terminus) membrane glycoprotein as proposed for PC-1 and that rat PD-Ia and mouse PC1 share 58% homology. Northern blot analysis showed that the PD-Ia gene is abundantly expressed in the central nervous system. In situ hybridization analysis demonstrated that PD-Ia mRNA is predominantly localized in choroid plexus epithelial cells, pigment epithelial cells of the eye, and cells in the granular layer of the cerebellum (13). A weak signal was observed in the cells in corpus callosum and other brain regions. In this study, we attempted the cloning of human cDNA and genomic DNA encoding PD-Ia as a first step toward understanding the mechanism of tissue-specific expression of this gene. To isolate the clones that cover the entire coding sequence of human PD-Ia, we screened the human retina
380
gnma
AP-Genomics
SHORT COMMUNICATION
cDNA library using the 5* and 3* fragments of rat PDIa cDNA as probes. Approximately 2 1 10 5 phages were screened and three positive clones were obtained. These three clones overlapped each other and covered the whole coding region of human PD-Ia. We designated this cDNA human PD-Ia (HGMW-approved symbol is PDNP2). The cDNA clones spanned a stretch of 2.7 kb. The cDNA sequence and the deduced amino acid sequence of human PD-Ia were submitted to the DNA Data Bank (Accession No. D45421). The nucleotide sequence of 2738 bp contains a small 33-bp 5*untranslated leader followed by a single open reading frame of 2589 bp encoding a polypeptide of 863 amino acids with a calculated molecular mass of 99,034 Da and by a 116-bp 3*-untranslated region. The entire primary structure contains four potential N-linked glycosylation sites. In the middle of the sequence, there is a putative Ca2/ binding site ‘‘EF hand’’-like motif situated at amino acids 740–748. Analysis of hydropathy using the method of Kyte and Doolittle confirmed that only one region of 22 residues, situated between residues 9 and 30, appears to have sufficient hydrophobicity to span the lipid layer. It is likely that human PD-Ia protein crosses the membrane only once. In addition, the human PD-Ia protein has a region nearly identical to the rat PD-Ia enzymatic domain (13) situated at amino acids 172– 219. Human PD-Ia protein has homology with the somatomedin Blike domain of vitronectin near the amino terminus of the extracellular portion, and both proteins have the cell attachment site ‘‘RGD’’ motif (human PD-Ia; 127 – 129 amino acids; vitronectin, 64–66 amino acids). This tripeptide is recognized by integrins including a5 b1 , aIIbb3 , and most of avb . Among RGD-recognizing integrins, there is a selectivity in natural ligand recognition. For example, fibronectin-specific receptor, which interacts with the ‘‘RGD’’ sequence in fibronectin, did not bind to vitronectin. Potentially, human PD-Ia protein may interact with the cell surface receptor, as do other RGD sequence-containing proteins that promote cell adhesion. Overall, the identity between human PD-Ia protein and rat PD-Ia protein is 89% and the overall structure of human PD-Ia is closely related to that of rat PDIa. The stretch of 16 amino acids of the human PD-Ia protein (148 –163) is not identical to that of rat PD-Ia. The amino acid sequence of the putative catalytic site is well conserved among the bovine intestine phosphodiesterase I, rat PD-Ia, and human PD-Ia. The overall structure of human PD-Ia is homologous to that of the human PC-1 protein. These results indicate that these proteins have similar enzymatic activity and physiological functions. The database searches of nucleotide and amino acid sequences reveal that human PD-Ia is identical with human autotaxin (ATX), a tumor cell motility-stimulating factor, except that human PD-Ia lacks 156 nucleotides (1022 –1177) and 52 amino acids (325 –376) of human ATX (12). Therefore, human PD-Ia and human
/ m4275f3678
11-03-95 14:41:39
gnma
381
FIG. 1. Northern blot analysis of human PD-Ia transcripts. Northern blot analysis of various human tissues was carried out as described previously (13). (Top) mRNA hybridized with a radiolabeled 0.9-kb fragment of human PD-Ia cDNA. In a control experiment, the same RNA samples were probed with b-actin cDNA, (bottom). The tissues or organs of RNA sources are as follows: lane 1, heart; lane 2, brain; lane 3, placenta; lane 4, lung; lane 5, liver; lane 6, skeletal muscle; lane 7, kidney; lane 8, pancreas; lane 9, amygdala; lane 10, caudate nucleus; lane 11, corpus callosum; lane 12, hippocampus; lane 13, hypothalamus; lane 14, substantia nigra; lane 15, subthalamic nucleus; lane 16, thalamus. The size of markers is shown to the right.
ATX are likely to be the alternative splicing products from the same gene. Human ATX is a 125-kDa glycoprotein that was identified in the conditioned medium from A2058 melanoma cells. Human ATX is an autocrine motility factor and stimulates both random and directed migration of human A2058 melanoma cells at picomolar concentrations (15). Human PD-Ia and human ATX are unique, new members of tumor cell motility-stimulating factors. Human PD-Ia lacks 156 nucleotides (1022 –1177) and 52 amino acids (325–376) of human ATX, and human PD-Ia protein lacks 25 amino acids (591 –615) of rat PD-Ia protein. These differences might be attributed to a tissue- or species-specific alternative splicing event. These peptides are located apart from the potentially functional domains of human PD-Ia protein, such as the enzymatic site, EF-hand-like motif, and RGD motif. Therefore, it is likely that these proteins have the same function. However, the possibility that these domains contribute to a tissue-specific function of this protein family cannot be excluded at present. To determine the size and tissue distribution of human PD-Ia mRNA, we carried out Northern blot analysis. The commercially available human multiple-tissue and human brain multiple-tissue Northern blots (Clontech) were probed with a 0.9-kb fragment of human PD-Ia cDNA encoding the 5* part of the coding region.
AP-Genomics
382
SHORT COMMUNICATION
FIG. 2. The nucleotide sequence of the 5* region of the human PDNP2 gene. The DNA sequence was determined from both strands by a dideoxy chain termination method using fluorescence-tagged primers and T7 DNA polymerase according to the protocol recommended by the manufacturer and was analyzed on an automated DNA sequencer (ALF DNA Sequencer II, Pharmacia LKB). The shaded area shows the Alu sequence. Putative Sp1 binding sites are double-underlined. The octamer motif is indicated by a square. Exons 1 and 2 are underlined.
As shown in Fig. 1, a single transcript of approximately 3 kb was identified in brain, placenta, lung, and kidney. To analyze the transcription regulatory region and determine the chromosomal localization of the human PDNP2 gene, we isolated the genomic DNA clone of human PDNP2. We screened approximately 4 1 105 phages of human leukocyte genomic DNA library (Japan Cancer Research Resources Bank), and one positive clone was obtained. This clone was plaque purified and then subjected to further analysis. To obtain the
/ m4275f3678
11-03-95 14:41:39
gnma
genomic DNA fragment containing the 5* untranslated region of human PDNP2, we carried out Southern blot analysis using the above-mentioned probe. The positive 7-kb fragment was obtained, and we sequenced the 5* region of human PDNP2 genomic DNA containing exon 1 (1986–2067) and exon 2 (2290 –2392) (Fig. 2). Since the exact transcription start site is unknown, the start site of exon 1 was tentatively situated at the most 5* site of human ATX cDNA, which was obtained using the rapid amplification of 5* cDNA ends system (12).
AP-Genomics
SHORT COMMUNICATION
383
FIG. 3. Chromosomal mapping of the human PDNP2 gene by fluorescence in situ hybridization. The genomic DNA fragment of human PDNP2 was labeled with biotin– dUTP and hybridized to human mitotic chromosomes. Hybridized biotinylated DNA was detected with avidin –FITC as described previously (9). (Left) Hoechst 33258 staining of chromosomes. (Right) Arrows indicate that the human PDNP2 gene is located at 8q24.1.
Exon 1 encodes the cytoplasmic tail and the proximal part of the transmembrane domain. Exon 2 encodes the remainder of the transmembrane domain and a part of the extracellular region. This exon structure of PDNP2 is similar to that of mouse PC-1 (1). The 5* flanking region has an Alu repetitive sequence (1–70). There is no typical TATA or CAAT box in this region. There are four potential binding sites of transcription factor Sp1 situated at 1115 –1200 (Sp1a), 1207 –1212 (Sp1b), 1212–1217 (Sp1c), and 1817 –1822 (Sp1d) (Fig. 2). Sp1a, -b, and -c are in a reversed orientation. Transcription factor Sp1 has the potential to activate transcription from a number of different viral and cellular promoters acting in conjunction with other cellular transcription factors. Sp1 is also utilized in TATA-boxlacking promoters in combination with other transcription factors. We surmise that Sp1 may act with other transcription factors in the promoter of the human PDNP2 gene. Three of the four putative Sp1 binding sites in the 5* region of the human PDNP2 gene are in a reversed orientation. Sp1-type promoter can activate the transcription in both orientations, as reported for the promoters of the human monoamine oxidase A gene (17). We considered the possibility that the Sp1 promoters of the human PDNP2 gene may act in both orientations. Two putative Sp1 binding sites in reversed orientation, Sp1b and -c, are overlapping (Fig. 2). It might be possible that only one Sp1 binds this region, because the shortest distance between the ‘‘central C’’ of two
/ m4275f3678
11-03-95 14:41:39
gnma
Sp1 binding sites (GGGCGG) should be at least 10 bp for two Sp1 molecules (4). In intron 2 of the human PDNP2 gene, there is an octamer binding motif, ATTTGCAT, situated at 2408– 2415 (Fig. 2), which is a well-characterized cis-acting regulatory sequence found in many promoters and enhancers. The octamer motif has been shown to be a part of both inducible tissue-specific immunoglobulin and ubiquitously expressed histone H2B promoters. In immunoglobulin genes, the octamer motif is located upstream of the transcription start site. On the other hand, the expression of the hst gene, a member of the fibroblast growth factor gene family, is positively regulated by a downstream octamer motif in an embryonal carcinoma cell line, F9 (8). It is conceivable that the expression of the human PDNP2 gene is regulated by a downstream octamer motif that functions as an enhancer. To determine the chromosomal localization of the human PDNP2 gene, we carried out fluorescence in situ hybridization using the genomic DNA clone as a probe. As shown in Fig. 3, the human PDNP2 gene is located at chromosome 8q24.1. In our laboratory the cDNA encoding rat intestinetype phosphodiesterase I/nucleotide pyrophosphatase (PD-Ib) has been cloned, and this protein also has an RGD motif (K. Terashima et al., manuscript in preparation). Therefore, these phosphodiesterase I/nucleotide pyrophosphatase family proteins might have multiple
AP-Genomics
384
SHORT COMMUNICATION
functions, such as cell adhesion, in addition to hydrolysis of nucleotide polyphosphates. Identifying the receptors for these proteins and investigating the regulatory mechanisms of the gene expression will help us to understand further the mode of actions and physiological roles of these proteins.
9.
10.
REFERENCES 1. Belli, S. I., van-Driel, I. R., and Goding, J. W. (1993). Identification and characterization of a soluble form of the plasma cell membrane glycoprotein PC-1 (5*-nucleotide phosphodiesterase). Eur. J. Biochem. 217: 421– 428. 2. Buckley, M. F., Loveland, K. A., McKinstry, W. J., Garson, O. M., and Goding, J. W. (1990). Plasma cell membrane glycoprotein PC-1. cDNA cloning of the human molecule, amino acid sequence, and chromosomal location. J. Biol. Chem. 265: 17506 –17511. 3. Evans, W. H., Hood, D. O., and Gurd, J. W. (1973). Purification and properties of a mouse liver plasma-membrane glycoprotein hydrolysing nucleotide pyrophosphate and phosphodiester bonds. Biochem. J. 135: 819–826. 4. Gidoni, D., Kadonaga, J. T., Barrera-Saldana, H., Takahashi, K., Chambon, P., and Tjian, R. (1985). Bidirectional SV40 transcription mediated by tandem Sp1 binding interactions. Science 230: 511–517. 5. Gross, S. C., Watabe, M., Goodarzi, G., Darabi, A., and Watabe, K. (1990). Organ-specific distribution of isozymes of 5*-nucleotide phosphodiesterase in mouse. Comp. Biochem. Physiol. B 95: 821–824. 6. Harahap, A. R. and Goding, J. W. (1988). Distribution of the murine plasma cell antigen PC-1 in non-lymphoid tissues. J. Immunol. 141: 2317 – 2320. 7. Harper, G. S., Hascall, V. C., Yanagishita, M., and Gahl, W. A. (1987). Proteoglycan synthesis in normal and Lowe syndrome fibroblasts. J. Biol. Chem. 262: 5637 –5643. 8. Hasan, S., Koda, T., and Kakinuma, M. (1994). An upstream NF-Y-binding site is required for transcriptional activation
/ m4275f3678
11-03-95 14:41:39
gnma
11.
12.
13.
14.
15.
16.
17.
from the hst promoter in F9 embryonal carcinoma cells. J. Biol. Chem. 269: 25042 – 25048. Inazawa, J., Saito, H., Ariyama, T., Abe, T., and Nakamura, Y. (1993). High-resolution cytogenetic mapping of 342 new cosmid markers including 43 RFLP markers on human chromosome 17 by fluorescence in situ hybridization. Genomics 17: 153–162. Maddux, B. A., Sbraccia, P., Kumakura, S., Sasson, S., Youngren, J., Fisher, A., Spencer, S., Grupe, A., Henzel, W., Stewart, T. A., Reaven, G. M., and Goldfine, I. D. (1995). Membrane glycoprotein PC-1 and insulin resistance in non-insulin-dependent diabetes mellitus. Nature 373: 448 –451. Maruyama, E., Takashima, S., and Arima, M. (1992). Increase in alkaline phosphodiesterase I activity of tumor-derived cultured fibroblasts from neurofibromatosis patients. Clin. Chim. Acta 211: 181–187. Murata, J., Lee, H. Y., Clair, T., Krutzsch, H. C., Arestad, A. A., Sobel, M. E., Liotta, L. A., and Stracke, M. L. (1994). cDNA cloning of human tumor motility-stimulating protein, Autotaxin, reveals a homology with phosphodiesterases. J. Biol. Chem. 269: 30479 – 30484. Narita, M., Goji, J., Nakamura, H., and Sano, K. (1994). Molecular cloning, expression, and localization of a brain-specific phosphodiesterase I/nucleotide pyrophosphatase (PD-Ia) from rat brain. J. Biol. Chem. 269: 28235 –28242. Rebbe, N. F., Tong, B. D., Finley, E. M., and Hickman, S. (1991). Identification of nucleotide pyrophosphatase/alkaline phosphodiesterase I activity associated with the mouse plasma cell differentiation antigen PC-1. Proc. Natl. Acad. Sci. USA 88: 5192 – 5196. ˚ restad, A., Stracke, M. L., Krutzsch, H. C., Unsworth, E. J., A Cioce, V., Schiffmann, E., and Liotta, L. A. (1992). Identification, purification, and partial sequence analysis of Autotaxin, a novel motility-stimulating protein. J. Biol. Chem. 267: 2524 – 2529. van-Driel, I. R., and Goding, J. W. (1987). Plasma cell membrane glycoprotein PC-1. Primary structure deduced from cDNA clones. J. Biol. Chem. 262: 4882 –4887. Zhu, Q. S., Chen, K., and Shih, J. C. (1994). Bidirectional promoter of human monoamine oxidase A (MAO A) controlled by transcription factor Sp1. J. Neurosci. 14: 7393 –7403.
AP-Genomics