- Email: [email protected]
Human chondroitin 6-sulfotransferase: cloning, gene structure, and chromosomal localization
Biochimica et Biophysica Acta 1407 (1998) 92^97
Short sequence-paper
Human chondroitin 6-sulfotransferase: cloning, gene structure, and chromosomal localization Kirstin D. Mazany a , Teng Peng a , Catherine E. Watson a , Ira Tabas b , Kevin Jon Williams a; * a
The Dorrance H. Hamilton Research Laboratories, Division of Endocrinology, Diabetes and Metabolic Diseases, Department of Medicine, Je¡erson Medical College, Thomas Je¡erson University, Room 349, 1020 Locust Street, Philadelphia, PA 19107-6799, USA b Departments of Medicine and Anatomy p Cell Biology, Columbia University, New York, NY 10032, USA Received 29 December 1997; accepted 22 April 1998
Abstract Chondroitin 6-sulfotransferase (C6ST) is the key enzyme in the biosynthesis of chondroitin 6-sulfate, a glycosaminoglycan implicated in chondrogenesis, neoplasia, atherosclerosis, and other processes. C6ST catalyzes the transfer of sulfate from 3Pphosphoadenosine 5P-phosphosulfate to carbon 6 of the N-acetylgalactosamine residues of chondroitin. Based on the previously published avian sequence, we searched the database of expressed sequence tags (dbEST) and obtained partiallength cDNAs that we completed by 5P-RACE using human chondrosarcoma and endothelial-cell RNA as template. Stable transfection of our full-length expression construct into CHO-K1 cells resulted in marked increases in C6ST and keratan sulfate sulfotransferase (KSST) enzymatic activities in cell homogenates. The predicted 411 amino acid sequence of human C6ST contains an N-terminal hydrophobic domain consistent with membrane insertion, four potential sites for N-linked glycosylation, several consensus sequences for protein phosphorylation, and one RGD sequence. The human and chick C6ST cDNA share 51% nucleotide identity, 40% amino acyl identity, and 75% amino acyl conservation. The human C6ST gene structure has been elucidated and exhibits an intron-less coding region, and the gene has been mapped to human chromosome 11 by radiation hybrid panel mapping. z 1998 Elsevier Science B.V. All rights reserved. Keywords: Chondroitin sulfate; Extracellular matrix; Glycosaminoglycan; Atherosclerosis
Chondroitin sulfate proteoglycans are ancient and widespread components of the cell surface and extracellular matrix of animals. Chondroitin 6-sulfate (C6S), which is sulfated at carbon 6 of the N-acetylgalactosamine residues, has been implicated in several key roles in human biology. The key enzyme in its biosynthesis, C6ST, catalyzes the transfer of sulfate from 3P-phosphoadenosine 5P-phosphosulfate
* Corresponding author. Fax: +1 (215) 923-7932; E-mail: [email protected]
(PAPS) to position 6 of N-acetylgalactosamine residue of chondroitin. The same enzyme in the chick has also been shown to possess keratan sulfate sulfotransferase (KSST) activity, i.e. it catalyzes the transfer of sulfate from PAPS to carbon 6 of the galactose residues of keratan sulfate [1^3]. In humans, a de¢ciency in C6ST activity has been associated with a heritable form of spondyloepiphyseal dysplasia, thus implicating this enzyme in normal skeletal development [4,5]. It has been shown that the connective tissue stroma of human colon carcinomas contains increased amounts of unsulfated and 6-sulfated
0925-4439 / 98 / $19.00 ß 1998 Elsevier Science B.V. All rights reserved. PII: S 0 9 2 5 - 4 4 3 9 ( 9 8 ) 0 0 0 2 8 - 3
BBADIS 60067 16-6-98
K.D. Mazany et al. / Biochimica et Biophysica Acta 1407 (1998) 92^97
chondroitin and reduced 4-sulfated forms, compared to the surrounding normal tissue [6]. Chondroitin sulfate proteoglycans in the arterial matrix, particularly 6-sulfated forms, have been implicated in arterial retention of cholesterol-rich, atherogenic lipoproteins [7,8], a key event that initiates atherosclerosis [9]. Importantly, the abundance of C6S is under genetic [10] and stimulatory [11^16] control, but molecular tools to study its function and regulation in mammalian cells and tissues have been lacking. Using the recently published chick cDNA sequence encoding C6ST [17], we performed a computer search [18] of the database of expressed sequence tags (dbEST) of the IMAGE consortium [19]. Three human sequences homologous to di¡erent regions of the chick C6ST were identi¢ed (IMAGE 40604-5P, 48676-5P, and 53039-5P), and the corresponding clones were purchased from Research Genetics, (Huntsville, AL). The inserts were fully sequenced (Model 377 automated sequencing system used with the DyeDeoxy Terminator Cycle sequencing kit, Applied Biosystem, Foster City, CA) and found to di¡er only in their extent of 5P sequence. The new human sequence information was used for another computer search of the dbEST, which identi¢ed three additional clones (K676-f, 43289-3P, and 53039-3P). The inserts in these clones were identical to the ¢rst three except for the extent of 5P sequence. Thus, we have no evidence for a gene family. The respective GenBank accession numbers for these dbEST sequences are R55609, H16077, and R16177 for the ¢rst set and R41023, H05595, and R15740 for the second set. The longest insert was 1806 bp (IMAGE 530395P), but it lacked a start codon. Therefore, we completed the sequence by 5P rapid ampli¢cation of cDNA ends (5P-RACE system, Gibco, Gaithersburg, MD), using RNA templates from human chondrosarcoma and human umbilical vein endothelial cells. No sequence di¡erences were observed amongst these sources. Sequences were assembled into a single
93
contig, which was analyzed for motifs [20] and compared [21] with the avian sequence. The human cDNA sequence contains a single open reading frame of 1233 bp, corresponding to 411 amino acyl residues, with a predicted molecular weight of 46 714 Da (Fig. 1A). The unique in-frame start codon is a strong initiation site for translation [22] and is homologous to the second of two potential start codons reported for the avian sequence [17]. Nucleotide sequences, including the 5P and 3P £anking regions, show 51% identity overall between the two species. Predicted amino acyl sequences, using the second potential avian start codon, show 40% identity and 75% conservation between human and chick. The human sequence contains an N-terminal hydrophobic sequence that resembles a signal peptide sequence [23], which might explain the presence of this enzyme extracellularly (see [1,5,24]). Much of this enzyme is intracellular (see below) in the Golgi [25,26], and it is possible that this N-terminal sequence is actually a Golgi retention signal [27]. Like chick C6ST, the human molecule has several potential sites for N-linked glycosylation [28]. Several phosphorylation motifs, including two tyrosine kinase phosphorylation sites [29,30], are present. Interestingly, the human sequence, unlike the chick, contains the RGD motif, which, if functional, would bind integrins when the enzyme is outside the cell [31]. To screen for human genomic clones, we used a 5PRACE clone to prepare a probe that contains the ¢rst 525 bp of our cDNA sequence. This probe was used for automated screening of a human genomic P1 arti¢cial chromosome (PAC) library (Genome Systems, St. Louis, MO). In this library, the PAC vector is 16.5 kb, and the genomic inserts are typically V120 kb. Two clones were identi¢ed, with clone addresses of PAC-18-L22 and PAC-189-O18, and we con¢rmed their identi¢cation by Southern blot of BamHI-generated fragments, hybridized with probes made from 5P and 3P segments of the C
Fig. 1. (A) The nucleotide sequence of the human C6ST cDNA and deduced amino acyl sequence. The N-terminal hydrophobic domain is underlined, with a triangle to indicate a consensus sequence for signal peptide cleavage. The four potential N-linked glycosylation sites are underlined. Tyrosine phosphorylation motifs are double underlined. The RGD sequence is overlined. The nucleotide sequence appears in the GenBank database under accession no. U65637. (B) Intron^exon boundary sequences. Intronic sequences are shown in upper case, and exonic in lower case. The intron is located between cDNA bases 82 and 83.
BBADIS 60067 16-6-98
94
K.D. Mazany et al. / Biochimica et Biophysica Acta 1407 (1998) 92^97
BBADIS 60067 16-6-98
K.D. Mazany et al. / Biochimica et Biophysica Acta 1407 (1998) 92^97
cDNA (data not shown). By sequencing the genomic clones directly or PCR-generated genomic fragments, we have found that the gene structure consists of an initial exon limited to the 5P untranslated region (UTR), an intron of 926 bp between cDNA bases 82 and 83, and then an exon that encodes the rest of the 5P UTR plus the entire coding region and 3P UTR. The termini of the intron (GT-AG) have been con¢rmed by sequencing (Fig. 1B). Mammalian genes with intron-less coding regions have been previously reported [32,33] and indicate a primitive structure that precludes the generation of di¡erent forms of the protein through alternative RNA splic-
Fig. 2. Sulfotransferase activities in homogenates of control and C6ST-transfected CHO cells. (A) Chondroitin sulfotransferase activities in homogenates of wild-type CHO cells (wt) and of CHO cell lines stably transfected with the expression vector for human C6ST (clonal lines, C1, C2, C3, C4; mixed lines, M2, M3). 35 S-Labeled products (vDi-6S and vDi-4S) and the origin are marked. The spots that appear at the level of the origin in lanes C1 and C2, but o¡-center, are markers placed after chromatography to locate the origin. (B) Keratan sulfate sulfotransferase activities in cell homogenates. Under these conditions, 35 S-labeled keratan sulfate remains at the origin. (C) Control incubations of cell homogenates with [35 S]PAPS, but without the addition of a sulfate acceptor. The contrast scale of this panel is the same as in (B). Each lane in the three panels represents the enzymatic activity in 3.2 Wg of cell protein.
95
ing. Genetic analyses for polymorphisms, however, should be technically straight-forward [33]. We used one of the genomic clones as template for PCR to amplify the entire C6ST coding region in one piece, which we then ligated into the pcDNA3 expression vector (Invitrogen, San Diego, CA; eukaryotic expression is driven by the CMV promoter). This construct was introduced into CHO cells by calcium-phosphate transfection [34], and stable transfectants were selected with G418 (pcDNA3 contains the neomycin resistance gene). As before [35], we minimized artifacts from clonal variation by propagating two mixed lines and four clonal lines. For controls, we used wild-type CHO cells and one clonal CHO line that had been transfected with an irrelevant cDNA, pFcR-Synd1 [35]. Expression of our construct was veri¢ed by Northern blot, which showed abundant message in all transfected cells, but not in the controls (data not shown). Homogenates of the cells were then prepared, and C6ST and C4ST enzymatic activities were assessed by incubation in the presence of [35 S]PAPS (the sulfate donor) and unsulfated chondroitin (the sulfate acceptor), followed by the complete enzymatic digestion of the newly sulfated chondroitin into disaccharides [17]. The labeled disaccharides, [35 S]vDi-6S and [35 S]vDi-4S, which indicate the enzymatic activities of C6ST and chondroitin 4-sulfotransferase (C4ST), respectively, were then separated by thin-layer chromatography using plates coated with microcrystalline cellulose (Analtech, Newark, DE) [36], and autoradiography was performed using a PhosphorImager (Molecular Dynamics, Sunnyvale, CA). Fig. 2A indicates that the wild-type CHO cells (wt) have abundant endogenous C4ST activity, yet very little C6ST activity, consistent with prior reports that the GAGs made by these cells have a low C6S:C4S ratio [37]. The clonal control was exactly the same (data not shown). In contrast, cell homogenates of all mixed and clonal lines transfected with the C6ST expression vector showed C6ST activity that was equal to or greater than the C4ST activity. The increase in the C6ST-to-C4ST ratio in the transfected lines over control cells was up to 14-fold. Thus, this cDNA encodes a human C6ST. Separate studies using keratan sulfate as the sulfate acceptor showed that, as with the chick enzyme [17], the human molecule also possesses KSST activity (Fig. 2B). As an
BBADIS 60067 16-6-98
96
K.D. Mazany et al. / Biochimica et Biophysica Acta 1407 (1998) 92^97
additional control, incubations of cell homogenates with [35 S]PAPS but no sulfate acceptor showed essentially no product (Fig. 2C). Chromosomal localization of our human C6ST gene has been completed by radiation hybrid panel mapping [38,39], using information from the Genome Database (http://gdbwww.gdb.org). A computer search of this database revealed that the Radiation Hybrid Transcript Mapping Consortium has analyzed two amplimers, stSG4330 and stSG3611, that were based on one of the six dbEST clones, clone R15740, that we had identi¢ed as containing partial-length human C6ST cDNA sequence. Using the Genebridge 4 panel, both of these amplimers were mapped to human chromosome 11. To ¢nd the locations within chromosome 11, we submitted the radiation panel data vectors for a mapping run on the Whitehead framework map, an lod 2.5 map (http://www-genome.wi.mit.edu). The mapping runs localized the amplimers to the same region of chromosome 11, namely, 4.5^6.51 centiRays (cR) from WI-4635 (D11S2375), with lod scores s 3.0. WI-4635 is a framework marker on the lod 2.5 map that has been placed 163.56 cR from the top of the chromosome 11 linkage group and is part of WC11.4, a singly linked YAC contig. Examination of chromosome 11 contigs anchored on the integrated map (http://www-genome.wi.mit.edu) revealed several nearby markers with simple sequence repeats known or likely to be polymorphic. Thus, these markers, or any other nearby polymorphic markers, can be used in linkage studies. Overall, these results provide necessary molecular tools to investigate the regulation and the role of C6S and C6ST in human physiology and disease. During the review of this manuscript, a report of this cDNA with its chromosomal location appeared by Fukuta et al. [40], but with three signi¢cant di¡erences from our results. First, although our sequence was homologous to the entire length of the chick C6ST cDNA, including the entire 5P UTR, the other report contained an additional 242 bp of 5P UTR in the human cDNA. Sequencing of our genomic DNA has revealed additional introns in this larger 5P UTR. Using Fukuta et al.s numbering, we have found introns between cDNA bp 140 and 141, 226 and 227, and 324 and 325 (the last of these is equivalent to our bp 82 and 83). Second, although our computer-
ized search of the dbEST did not reveal any other human cDNAs with strong homology to the chick C6ST, Fukuta et al. mention the existence of a second homologous human sequence, which they found during their screening of a Vgt 11 human cDNA library [40]. Third, and most important, our expression studies indicated both C6ST and KSST activities, as with the chick enzyme, whereas the results of Fukuta et al. revealed a novel pattern of KSST activity with no detectable C6ST activity [40]. Some of this di¡erence might have arisen from our use of stable transfection, which a¡ects essentially 100% of the cells being assayed, in a cell line that has a very low ratio of endogenous C6ST to C4ST activities (see Fig. 2A and [37]), whereas Fukuta et al. used transient transfections in COS-7 cells, which have a high endogenous ratio of C6ST to C4ST (see Table 1 in [40]). The more intriguing possibility is that posttranslational processing may a¡ect C6ST activity and that CHO cells and COS-7 cells exhibit key differences in this regard. Of note, Fukuta et al. reported the partial puri¢cation of KSST activity with no contaminating C6ST activity in the unbound fraction of transfected COS-7 cell homogenates subjected to DEAE chromatography in 50 mM NaCl (Fig. 3A in [40]). Our recent work indicates essentially no enzymatic activity of either type in this fraction from transfected CHO cell homogenates, i.e., over 95% of recovered C6ST and KSST activities were retained on DEAE (data not shown). Studies are currently underway in our laboratory to investigate this di¡erence. We thank Dr. Vickie Bennett of Thomas Je¡erson University for the chondrosarcoma RNA. This work was supported by National Institutes of Health (USA) Grants R01-HL38956 and P50HL56984 and by American Heart Association Grant-in-Aid 93-1081. K.J.W. is an Established Investigator of the American Heart Association and Genentech.
References [1] O. Habuchi, Y. Matsui, Y. Kotoya, Y. Aoyama, Y. Yasuda, M. Noda, J. Biol. Chem. 268 (1993) 21968^21974. [2] G. Sugumaran, M. Katsman, R.R. Drake, J. Biol. Chem. 270 (1995) 22483^22487.
BBADIS 60067 16-6-98
K.D. Mazany et al. / Biochimica et Biophysica Acta 1407 (1998) 92^97 [3] O. Habuchi, Y. Hirahara, K. Uchimura, M. Fukuta, Glycobiology 6 (1996) 51^57. [4] S.P. Toledo, P.A. Mourao, C. Lamego, C.A. Alves, C.P. Dietrich, L.M. Assis, E. Mattar, Am. J. Med. Genet. 2 (1978) 385^395. [5] P.A. Moura¬o, S. Kato, P.V. Donnelly, Biochem. Biophys. Res. Commun. 98 (1981) 388^396. [6] R. Adany, R. Heimer, B. Caterson, J.M. Sorrell, R.V. Iozzo, J. Biol. Chem. 265 (1990) 11389^11396. [7] P.A. Mourao, S. Pillai, N. Di Ferrante, Biochim. Biophys. Acta 674 (1981) 178^187. [8] G. Camejo, E. Hurt-Camejo, U. Olsson, G. Bondjers, Curr. Opin. Lipidol. 4 (1993) 385^391. [9] K.J. Williams, I. Tabas, Arterioscler. Thromb. Vasc. Biol. 15 (1995) 551^561. [10] I.J. Edwards, W.D. Wagner, J. Biol. Chem. 263 (1988) 9612^ 9620. [11] P. Vijayagopal, S.R. Srinivasan, E.R. Dalferes Jr., B. Radhakrishnamurthy, Biochem. J. 255 (1988) 639^646. [12] E. Scho«nherr, H.T. Ja«rvela«inen, L.J. Sandell, T.N. Wight, J. Biol. Chem. 266 (1991) 17640^17647. [13] S.P. Evanko, K.G. Vogel, Arch. Biochem. Biophys. 307 (1993) 153^164. [14] R.S. Carvalho, E.H. Yen, D.M. Suga, Am. J. Orthod. Dentofacial Orthop. 107 (1995) 401^410. [15] M.Y. Chang, A. Chait, S. Potter-Perigo, T.N. Wight, Circulation 92 (suppl. I) (1995) 288 (Abstract). [16] P. Vijayagopal, J.E. Figueroa, Q. Guo, J.D. Fontenot, Z. Tao, Biochem. J. 315 (1996) 995^1000. [17] M. Fukuta, K. Uchimura, K. Nakashima, M. Kato, K. Kimata, T. Shinomura, O. Habuchi, J. Biol. Chem. 270 (1995) 18575^18580. [18] S.F. Altschul, W. Gish, W. Miller, E.W. Myers, D.J. Lipman, J. Mol. Biol. 215 (1990) 403^410. [19] G. Lennon, C. Au¡ray, M. Polymeropoulos, M.B. Soares, Genomics 33 (1996) 151^152. [20] Rice, P. (1996) Program Manual for the EGCG Package, The Sanger Center, Hinxton Hall, Cambridge, UK. [21] W.R. Pearson, W. Miller, Methods Enzymol. 210 (1992) 575^601.
[22] [23] [24] [25] [26]
[27] [28] [29] [30] [31] [32]
[33]
[34]
[35]
[36] [37] [38] [39] [40]
97
M. Kozak, Nucleic Acids Res. 9 (1981) 5233^5262. G. von Heijne, Nucleic Acids Res. 14 (1986) 4683^4690. J.B. Adams, Biochim. Biophys. Acta 83 (1964) 127^129. G. Sugumaran, M. Katsman, J.E. Silbert, J. Biol. Chem. 267 (1992) 8802^8806. N.B. Schwartz, in: T.M. Devlin (Ed.), Textbook of Biochemistry with Clinical Correlations, Wiley-Liss, New York, 1992, pp. 359^386. P.A. Gleeson, R.D. Teasdale, J. Burke, Glycoconj. J. 11 (1994) 381^394. Y. Gavel, G. von Heijne, Protein Eng. 3 (1990) 433^442. T. Patschinsky, T. Hunter, F.S. Esch, J.A. Cooper, B.M. Sefton, Proc. Natl. Acad. Sci. USA 79 (1982) 973^977. T. Hunter, J. Biol. Chem. 257 (1982) 4843^4848. E. Ruoslahti, M.D. Pierschbacher, Science 238 (1987) 491^ 497. K.G. Chandy, C.B. Williams, R.H. Spencer, B.A. Aguilar, S. Ghanshani, B.L. Tempel, G.A. Gutman, Science 247 (1990) 973^975. E. Araki, X.J. Sun, B.L. Haag, L.M. Chuang, Y. Zhang, T.L. Yang-Feng, M.F. White, C.R. Kahn, Diabetes 42 (1993) 1041^1054. J. Sambrook, E.F. Fritsch, T. Maniatis, Molecular Cloning. A Laboratory Manual, Cold Spring Harbor Press, Cold Spring Harbor, New York, 1989. I.V. Fuki, K.M. Kuhn, I.R. Lomazov, V.L. Rothman, G.P. Tuszynski, R.V. Iozzo, T.L. Swenson, E.A. Fisher, K.J. Williams, J. Clin. Invest. 100 (1997) 1611^1622. L. Wasserman, A. Ber, D. Allalouf, J. Chromatogr. 136 (1977) 342^347. J.D. Esko, A. Elgavish, T. Prasthofer, W.H. Taylor, J.L. Weinke, J. Biol. Chem. 261 (1986) 15725^15733. D.R. Cox, M. Burmeister, E.R. Price, S. Kim, R.M. Myers, Science 250 (1990) 245^250. M.A. Walter, D.J. Spillett, P. Thomas, J. Weissenbach, P.N. Goodfellow, Nat. Genet. 7 (1994) 22^28. M. Fukuta, J. Inazawa, T. Torii, K. Tsuzuki, E. Shimada, O. Habuchi, J. Biol. Chem. 272 (1997) 32321^32328.
BBADIS 60067 16-6-98
Short sequence-paper
Human chondroitin 6-sulfotransferase: cloning, gene structure, and chromosomal localization Kirstin D. Mazany a , Teng Peng a , Catherine E. Watson a , Ira Tabas b , Kevin Jon Williams a; * a
The Dorrance H. Hamilton Research Laboratories, Division of Endocrinology, Diabetes and Metabolic Diseases, Department of Medicine, Je¡erson Medical College, Thomas Je¡erson University, Room 349, 1020 Locust Street, Philadelphia, PA 19107-6799, USA b Departments of Medicine and Anatomy p Cell Biology, Columbia University, New York, NY 10032, USA Received 29 December 1997; accepted 22 April 1998
Abstract Chondroitin 6-sulfotransferase (C6ST) is the key enzyme in the biosynthesis of chondroitin 6-sulfate, a glycosaminoglycan implicated in chondrogenesis, neoplasia, atherosclerosis, and other processes. C6ST catalyzes the transfer of sulfate from 3Pphosphoadenosine 5P-phosphosulfate to carbon 6 of the N-acetylgalactosamine residues of chondroitin. Based on the previously published avian sequence, we searched the database of expressed sequence tags (dbEST) and obtained partiallength cDNAs that we completed by 5P-RACE using human chondrosarcoma and endothelial-cell RNA as template. Stable transfection of our full-length expression construct into CHO-K1 cells resulted in marked increases in C6ST and keratan sulfate sulfotransferase (KSST) enzymatic activities in cell homogenates. The predicted 411 amino acid sequence of human C6ST contains an N-terminal hydrophobic domain consistent with membrane insertion, four potential sites for N-linked glycosylation, several consensus sequences for protein phosphorylation, and one RGD sequence. The human and chick C6ST cDNA share 51% nucleotide identity, 40% amino acyl identity, and 75% amino acyl conservation. The human C6ST gene structure has been elucidated and exhibits an intron-less coding region, and the gene has been mapped to human chromosome 11 by radiation hybrid panel mapping. z 1998 Elsevier Science B.V. All rights reserved. Keywords: Chondroitin sulfate; Extracellular matrix; Glycosaminoglycan; Atherosclerosis
Chondroitin sulfate proteoglycans are ancient and widespread components of the cell surface and extracellular matrix of animals. Chondroitin 6-sulfate (C6S), which is sulfated at carbon 6 of the N-acetylgalactosamine residues, has been implicated in several key roles in human biology. The key enzyme in its biosynthesis, C6ST, catalyzes the transfer of sulfate from 3P-phosphoadenosine 5P-phosphosulfate
* Corresponding author. Fax: +1 (215) 923-7932; E-mail: [email protected]
(PAPS) to position 6 of N-acetylgalactosamine residue of chondroitin. The same enzyme in the chick has also been shown to possess keratan sulfate sulfotransferase (KSST) activity, i.e. it catalyzes the transfer of sulfate from PAPS to carbon 6 of the galactose residues of keratan sulfate [1^3]. In humans, a de¢ciency in C6ST activity has been associated with a heritable form of spondyloepiphyseal dysplasia, thus implicating this enzyme in normal skeletal development [4,5]. It has been shown that the connective tissue stroma of human colon carcinomas contains increased amounts of unsulfated and 6-sulfated
0925-4439 / 98 / $19.00 ß 1998 Elsevier Science B.V. All rights reserved. PII: S 0 9 2 5 - 4 4 3 9 ( 9 8 ) 0 0 0 2 8 - 3
BBADIS 60067 16-6-98
K.D. Mazany et al. / Biochimica et Biophysica Acta 1407 (1998) 92^97
chondroitin and reduced 4-sulfated forms, compared to the surrounding normal tissue [6]. Chondroitin sulfate proteoglycans in the arterial matrix, particularly 6-sulfated forms, have been implicated in arterial retention of cholesterol-rich, atherogenic lipoproteins [7,8], a key event that initiates atherosclerosis [9]. Importantly, the abundance of C6S is under genetic [10] and stimulatory [11^16] control, but molecular tools to study its function and regulation in mammalian cells and tissues have been lacking. Using the recently published chick cDNA sequence encoding C6ST [17], we performed a computer search [18] of the database of expressed sequence tags (dbEST) of the IMAGE consortium [19]. Three human sequences homologous to di¡erent regions of the chick C6ST were identi¢ed (IMAGE 40604-5P, 48676-5P, and 53039-5P), and the corresponding clones were purchased from Research Genetics, (Huntsville, AL). The inserts were fully sequenced (Model 377 automated sequencing system used with the DyeDeoxy Terminator Cycle sequencing kit, Applied Biosystem, Foster City, CA) and found to di¡er only in their extent of 5P sequence. The new human sequence information was used for another computer search of the dbEST, which identi¢ed three additional clones (K676-f, 43289-3P, and 53039-3P). The inserts in these clones were identical to the ¢rst three except for the extent of 5P sequence. Thus, we have no evidence for a gene family. The respective GenBank accession numbers for these dbEST sequences are R55609, H16077, and R16177 for the ¢rst set and R41023, H05595, and R15740 for the second set. The longest insert was 1806 bp (IMAGE 530395P), but it lacked a start codon. Therefore, we completed the sequence by 5P rapid ampli¢cation of cDNA ends (5P-RACE system, Gibco, Gaithersburg, MD), using RNA templates from human chondrosarcoma and human umbilical vein endothelial cells. No sequence di¡erences were observed amongst these sources. Sequences were assembled into a single
93
contig, which was analyzed for motifs [20] and compared [21] with the avian sequence. The human cDNA sequence contains a single open reading frame of 1233 bp, corresponding to 411 amino acyl residues, with a predicted molecular weight of 46 714 Da (Fig. 1A). The unique in-frame start codon is a strong initiation site for translation [22] and is homologous to the second of two potential start codons reported for the avian sequence [17]. Nucleotide sequences, including the 5P and 3P £anking regions, show 51% identity overall between the two species. Predicted amino acyl sequences, using the second potential avian start codon, show 40% identity and 75% conservation between human and chick. The human sequence contains an N-terminal hydrophobic sequence that resembles a signal peptide sequence [23], which might explain the presence of this enzyme extracellularly (see [1,5,24]). Much of this enzyme is intracellular (see below) in the Golgi [25,26], and it is possible that this N-terminal sequence is actually a Golgi retention signal [27]. Like chick C6ST, the human molecule has several potential sites for N-linked glycosylation [28]. Several phosphorylation motifs, including two tyrosine kinase phosphorylation sites [29,30], are present. Interestingly, the human sequence, unlike the chick, contains the RGD motif, which, if functional, would bind integrins when the enzyme is outside the cell [31]. To screen for human genomic clones, we used a 5PRACE clone to prepare a probe that contains the ¢rst 525 bp of our cDNA sequence. This probe was used for automated screening of a human genomic P1 arti¢cial chromosome (PAC) library (Genome Systems, St. Louis, MO). In this library, the PAC vector is 16.5 kb, and the genomic inserts are typically V120 kb. Two clones were identi¢ed, with clone addresses of PAC-18-L22 and PAC-189-O18, and we con¢rmed their identi¢cation by Southern blot of BamHI-generated fragments, hybridized with probes made from 5P and 3P segments of the C
Fig. 1. (A) The nucleotide sequence of the human C6ST cDNA and deduced amino acyl sequence. The N-terminal hydrophobic domain is underlined, with a triangle to indicate a consensus sequence for signal peptide cleavage. The four potential N-linked glycosylation sites are underlined. Tyrosine phosphorylation motifs are double underlined. The RGD sequence is overlined. The nucleotide sequence appears in the GenBank database under accession no. U65637. (B) Intron^exon boundary sequences. Intronic sequences are shown in upper case, and exonic in lower case. The intron is located between cDNA bases 82 and 83.
BBADIS 60067 16-6-98
94
K.D. Mazany et al. / Biochimica et Biophysica Acta 1407 (1998) 92^97
BBADIS 60067 16-6-98
K.D. Mazany et al. / Biochimica et Biophysica Acta 1407 (1998) 92^97
cDNA (data not shown). By sequencing the genomic clones directly or PCR-generated genomic fragments, we have found that the gene structure consists of an initial exon limited to the 5P untranslated region (UTR), an intron of 926 bp between cDNA bases 82 and 83, and then an exon that encodes the rest of the 5P UTR plus the entire coding region and 3P UTR. The termini of the intron (GT-AG) have been con¢rmed by sequencing (Fig. 1B). Mammalian genes with intron-less coding regions have been previously reported [32,33] and indicate a primitive structure that precludes the generation of di¡erent forms of the protein through alternative RNA splic-
Fig. 2. Sulfotransferase activities in homogenates of control and C6ST-transfected CHO cells. (A) Chondroitin sulfotransferase activities in homogenates of wild-type CHO cells (wt) and of CHO cell lines stably transfected with the expression vector for human C6ST (clonal lines, C1, C2, C3, C4; mixed lines, M2, M3). 35 S-Labeled products (vDi-6S and vDi-4S) and the origin are marked. The spots that appear at the level of the origin in lanes C1 and C2, but o¡-center, are markers placed after chromatography to locate the origin. (B) Keratan sulfate sulfotransferase activities in cell homogenates. Under these conditions, 35 S-labeled keratan sulfate remains at the origin. (C) Control incubations of cell homogenates with [35 S]PAPS, but without the addition of a sulfate acceptor. The contrast scale of this panel is the same as in (B). Each lane in the three panels represents the enzymatic activity in 3.2 Wg of cell protein.
95
ing. Genetic analyses for polymorphisms, however, should be technically straight-forward [33]. We used one of the genomic clones as template for PCR to amplify the entire C6ST coding region in one piece, which we then ligated into the pcDNA3 expression vector (Invitrogen, San Diego, CA; eukaryotic expression is driven by the CMV promoter). This construct was introduced into CHO cells by calcium-phosphate transfection [34], and stable transfectants were selected with G418 (pcDNA3 contains the neomycin resistance gene). As before [35], we minimized artifacts from clonal variation by propagating two mixed lines and four clonal lines. For controls, we used wild-type CHO cells and one clonal CHO line that had been transfected with an irrelevant cDNA, pFcR-Synd1 [35]. Expression of our construct was veri¢ed by Northern blot, which showed abundant message in all transfected cells, but not in the controls (data not shown). Homogenates of the cells were then prepared, and C6ST and C4ST enzymatic activities were assessed by incubation in the presence of [35 S]PAPS (the sulfate donor) and unsulfated chondroitin (the sulfate acceptor), followed by the complete enzymatic digestion of the newly sulfated chondroitin into disaccharides [17]. The labeled disaccharides, [35 S]vDi-6S and [35 S]vDi-4S, which indicate the enzymatic activities of C6ST and chondroitin 4-sulfotransferase (C4ST), respectively, were then separated by thin-layer chromatography using plates coated with microcrystalline cellulose (Analtech, Newark, DE) [36], and autoradiography was performed using a PhosphorImager (Molecular Dynamics, Sunnyvale, CA). Fig. 2A indicates that the wild-type CHO cells (wt) have abundant endogenous C4ST activity, yet very little C6ST activity, consistent with prior reports that the GAGs made by these cells have a low C6S:C4S ratio [37]. The clonal control was exactly the same (data not shown). In contrast, cell homogenates of all mixed and clonal lines transfected with the C6ST expression vector showed C6ST activity that was equal to or greater than the C4ST activity. The increase in the C6ST-to-C4ST ratio in the transfected lines over control cells was up to 14-fold. Thus, this cDNA encodes a human C6ST. Separate studies using keratan sulfate as the sulfate acceptor showed that, as with the chick enzyme [17], the human molecule also possesses KSST activity (Fig. 2B). As an
BBADIS 60067 16-6-98
96
K.D. Mazany et al. / Biochimica et Biophysica Acta 1407 (1998) 92^97
additional control, incubations of cell homogenates with [35 S]PAPS but no sulfate acceptor showed essentially no product (Fig. 2C). Chromosomal localization of our human C6ST gene has been completed by radiation hybrid panel mapping [38,39], using information from the Genome Database (http://gdbwww.gdb.org). A computer search of this database revealed that the Radiation Hybrid Transcript Mapping Consortium has analyzed two amplimers, stSG4330 and stSG3611, that were based on one of the six dbEST clones, clone R15740, that we had identi¢ed as containing partial-length human C6ST cDNA sequence. Using the Genebridge 4 panel, both of these amplimers were mapped to human chromosome 11. To ¢nd the locations within chromosome 11, we submitted the radiation panel data vectors for a mapping run on the Whitehead framework map, an lod 2.5 map (http://www-genome.wi.mit.edu). The mapping runs localized the amplimers to the same region of chromosome 11, namely, 4.5^6.51 centiRays (cR) from WI-4635 (D11S2375), with lod scores s 3.0. WI-4635 is a framework marker on the lod 2.5 map that has been placed 163.56 cR from the top of the chromosome 11 linkage group and is part of WC11.4, a singly linked YAC contig. Examination of chromosome 11 contigs anchored on the integrated map (http://www-genome.wi.mit.edu) revealed several nearby markers with simple sequence repeats known or likely to be polymorphic. Thus, these markers, or any other nearby polymorphic markers, can be used in linkage studies. Overall, these results provide necessary molecular tools to investigate the regulation and the role of C6S and C6ST in human physiology and disease. During the review of this manuscript, a report of this cDNA with its chromosomal location appeared by Fukuta et al. [40], but with three signi¢cant di¡erences from our results. First, although our sequence was homologous to the entire length of the chick C6ST cDNA, including the entire 5P UTR, the other report contained an additional 242 bp of 5P UTR in the human cDNA. Sequencing of our genomic DNA has revealed additional introns in this larger 5P UTR. Using Fukuta et al.s numbering, we have found introns between cDNA bp 140 and 141, 226 and 227, and 324 and 325 (the last of these is equivalent to our bp 82 and 83). Second, although our computer-
ized search of the dbEST did not reveal any other human cDNAs with strong homology to the chick C6ST, Fukuta et al. mention the existence of a second homologous human sequence, which they found during their screening of a Vgt 11 human cDNA library [40]. Third, and most important, our expression studies indicated both C6ST and KSST activities, as with the chick enzyme, whereas the results of Fukuta et al. revealed a novel pattern of KSST activity with no detectable C6ST activity [40]. Some of this di¡erence might have arisen from our use of stable transfection, which a¡ects essentially 100% of the cells being assayed, in a cell line that has a very low ratio of endogenous C6ST to C4ST activities (see Fig. 2A and [37]), whereas Fukuta et al. used transient transfections in COS-7 cells, which have a high endogenous ratio of C6ST to C4ST (see Table 1 in [40]). The more intriguing possibility is that posttranslational processing may a¡ect C6ST activity and that CHO cells and COS-7 cells exhibit key differences in this regard. Of note, Fukuta et al. reported the partial puri¢cation of KSST activity with no contaminating C6ST activity in the unbound fraction of transfected COS-7 cell homogenates subjected to DEAE chromatography in 50 mM NaCl (Fig. 3A in [40]). Our recent work indicates essentially no enzymatic activity of either type in this fraction from transfected CHO cell homogenates, i.e., over 95% of recovered C6ST and KSST activities were retained on DEAE (data not shown). Studies are currently underway in our laboratory to investigate this di¡erence. We thank Dr. Vickie Bennett of Thomas Je¡erson University for the chondrosarcoma RNA. This work was supported by National Institutes of Health (USA) Grants R01-HL38956 and P50HL56984 and by American Heart Association Grant-in-Aid 93-1081. K.J.W. is an Established Investigator of the American Heart Association and Genentech.
References [1] O. Habuchi, Y. Matsui, Y. Kotoya, Y. Aoyama, Y. Yasuda, M. Noda, J. Biol. Chem. 268 (1993) 21968^21974. [2] G. Sugumaran, M. Katsman, R.R. Drake, J. Biol. Chem. 270 (1995) 22483^22487.
BBADIS 60067 16-6-98
K.D. Mazany et al. / Biochimica et Biophysica Acta 1407 (1998) 92^97 [3] O. Habuchi, Y. Hirahara, K. Uchimura, M. Fukuta, Glycobiology 6 (1996) 51^57. [4] S.P. Toledo, P.A. Mourao, C. Lamego, C.A. Alves, C.P. Dietrich, L.M. Assis, E. Mattar, Am. J. Med. Genet. 2 (1978) 385^395. [5] P.A. Moura¬o, S. Kato, P.V. Donnelly, Biochem. Biophys. Res. Commun. 98 (1981) 388^396. [6] R. Adany, R. Heimer, B. Caterson, J.M. Sorrell, R.V. Iozzo, J. Biol. Chem. 265 (1990) 11389^11396. [7] P.A. Mourao, S. Pillai, N. Di Ferrante, Biochim. Biophys. Acta 674 (1981) 178^187. [8] G. Camejo, E. Hurt-Camejo, U. Olsson, G. Bondjers, Curr. Opin. Lipidol. 4 (1993) 385^391. [9] K.J. Williams, I. Tabas, Arterioscler. Thromb. Vasc. Biol. 15 (1995) 551^561. [10] I.J. Edwards, W.D. Wagner, J. Biol. Chem. 263 (1988) 9612^ 9620. [11] P. Vijayagopal, S.R. Srinivasan, E.R. Dalferes Jr., B. Radhakrishnamurthy, Biochem. J. 255 (1988) 639^646. [12] E. Scho«nherr, H.T. Ja«rvela«inen, L.J. Sandell, T.N. Wight, J. Biol. Chem. 266 (1991) 17640^17647. [13] S.P. Evanko, K.G. Vogel, Arch. Biochem. Biophys. 307 (1993) 153^164. [14] R.S. Carvalho, E.H. Yen, D.M. Suga, Am. J. Orthod. Dentofacial Orthop. 107 (1995) 401^410. [15] M.Y. Chang, A. Chait, S. Potter-Perigo, T.N. Wight, Circulation 92 (suppl. I) (1995) 288 (Abstract). [16] P. Vijayagopal, J.E. Figueroa, Q. Guo, J.D. Fontenot, Z. Tao, Biochem. J. 315 (1996) 995^1000. [17] M. Fukuta, K. Uchimura, K. Nakashima, M. Kato, K. Kimata, T. Shinomura, O. Habuchi, J. Biol. Chem. 270 (1995) 18575^18580. [18] S.F. Altschul, W. Gish, W. Miller, E.W. Myers, D.J. Lipman, J. Mol. Biol. 215 (1990) 403^410. [19] G. Lennon, C. Au¡ray, M. Polymeropoulos, M.B. Soares, Genomics 33 (1996) 151^152. [20] Rice, P. (1996) Program Manual for the EGCG Package, The Sanger Center, Hinxton Hall, Cambridge, UK. [21] W.R. Pearson, W. Miller, Methods Enzymol. 210 (1992) 575^601.
[22] [23] [24] [25] [26]
[27] [28] [29] [30] [31] [32]
[33]
[34]
[35]
[36] [37] [38] [39] [40]
97
M. Kozak, Nucleic Acids Res. 9 (1981) 5233^5262. G. von Heijne, Nucleic Acids Res. 14 (1986) 4683^4690. J.B. Adams, Biochim. Biophys. Acta 83 (1964) 127^129. G. Sugumaran, M. Katsman, J.E. Silbert, J. Biol. Chem. 267 (1992) 8802^8806. N.B. Schwartz, in: T.M. Devlin (Ed.), Textbook of Biochemistry with Clinical Correlations, Wiley-Liss, New York, 1992, pp. 359^386. P.A. Gleeson, R.D. Teasdale, J. Burke, Glycoconj. J. 11 (1994) 381^394. Y. Gavel, G. von Heijne, Protein Eng. 3 (1990) 433^442. T. Patschinsky, T. Hunter, F.S. Esch, J.A. Cooper, B.M. Sefton, Proc. Natl. Acad. Sci. USA 79 (1982) 973^977. T. Hunter, J. Biol. Chem. 257 (1982) 4843^4848. E. Ruoslahti, M.D. Pierschbacher, Science 238 (1987) 491^ 497. K.G. Chandy, C.B. Williams, R.H. Spencer, B.A. Aguilar, S. Ghanshani, B.L. Tempel, G.A. Gutman, Science 247 (1990) 973^975. E. Araki, X.J. Sun, B.L. Haag, L.M. Chuang, Y. Zhang, T.L. Yang-Feng, M.F. White, C.R. Kahn, Diabetes 42 (1993) 1041^1054. J. Sambrook, E.F. Fritsch, T. Maniatis, Molecular Cloning. A Laboratory Manual, Cold Spring Harbor Press, Cold Spring Harbor, New York, 1989. I.V. Fuki, K.M. Kuhn, I.R. Lomazov, V.L. Rothman, G.P. Tuszynski, R.V. Iozzo, T.L. Swenson, E.A. Fisher, K.J. Williams, J. Clin. Invest. 100 (1997) 1611^1622. L. Wasserman, A. Ber, D. Allalouf, J. Chromatogr. 136 (1977) 342^347. J.D. Esko, A. Elgavish, T. Prasthofer, W.H. Taylor, J.L. Weinke, J. Biol. Chem. 261 (1986) 15725^15733. D.R. Cox, M. Burmeister, E.R. Price, S. Kim, R.M. Myers, Science 250 (1990) 245^250. M.A. Walter, D.J. Spillett, P. Thomas, J. Weissenbach, P.N. Goodfellow, Nat. Genet. 7 (1994) 22^28. M. Fukuta, J. Inazawa, T. Torii, K. Tsuzuki, E. Shimada, O. Habuchi, J. Biol. Chem. 272 (1997) 32321^32328.
BBADIS 60067 16-6-98