Molecular Cloning of a Chondroitin Polymerizing Factor That Cooperates with Chondroitin Synthase for Chondroitin Polymerization
2003; Elsevier BV; Volume: 278; Issue: 26 Linguagem: Inglês
10.1074/jbc.m302493200
ISSN1083-351X
AutoresHiroshi Kitagawa, Tomomi Izumikawa, Toru Uyama, Kazuyuki Sugahara,
Tópico(s)Skin and Cellular Biology Research
ResumoWe recently cloned human chondroitin synthase (ChSy) exhibiting the glucuronyltransferase-II (GlcATII) and N-acetylgalactosaminyltransferase-II (GalNAcTII) activities responsible for the biosynthesis of repeating disaccharide units of chondroitin sulfate, but chondroitin polymerization was not demonstrated in vitro using the recombinant ChSy. We report here that the chondroitin polymerizing activity requires concomitant expression of a novel protein designated chondroitin polymerizing factor (ChPF) with ChSy. The human ChPF consists of 775 amino acids with a type II transmembrane protein topology. The amino acid sequence displayed 23% identity to that of human ChSy. The expression of a soluble recombinant form of the protein in COS-1 cells produced a protein with little GlcAT-II or GalNAcT-II activity. In contrast, coexpression of the ChPF and ChSy yielded markedly augmented glycosyltransferase activities, whereas simple mixing of the two separately expressed proteins did not. Moreover, using both UDP-glucuronic acid (GlcUA) and UDP-N-acetylgalactosamine (GalNAc) as sugar donors, chondroitin polymerization was demonstrated on the so-called glycosaminoglycan-protein linkage region tetrasaccharide sequence of α-thrombomodulin. These results suggested that the ChPF acts as a specific activating factor for ChSy in chondroitin polymerization. The coding region of the ChPF was divided into four discrete exons and localized to chromosome 2q35-q36. Northern blot analysis revealed that the ChPF gene exhibited a markedly different expression pattern among various human tissues, which was similar to that of ChSy. Thus, the ChPF is required for chondroitin polymerizing activity of mammalian ChSy. We recently cloned human chondroitin synthase (ChSy) exhibiting the glucuronyltransferase-II (GlcATII) and N-acetylgalactosaminyltransferase-II (GalNAcTII) activities responsible for the biosynthesis of repeating disaccharide units of chondroitin sulfate, but chondroitin polymerization was not demonstrated in vitro using the recombinant ChSy. We report here that the chondroitin polymerizing activity requires concomitant expression of a novel protein designated chondroitin polymerizing factor (ChPF) with ChSy. The human ChPF consists of 775 amino acids with a type II transmembrane protein topology. The amino acid sequence displayed 23% identity to that of human ChSy. The expression of a soluble recombinant form of the protein in COS-1 cells produced a protein with little GlcAT-II or GalNAcT-II activity. In contrast, coexpression of the ChPF and ChSy yielded markedly augmented glycosyltransferase activities, whereas simple mixing of the two separately expressed proteins did not. Moreover, using both UDP-glucuronic acid (GlcUA) and UDP-N-acetylgalactosamine (GalNAc) as sugar donors, chondroitin polymerization was demonstrated on the so-called glycosaminoglycan-protein linkage region tetrasaccharide sequence of α-thrombomodulin. These results suggested that the ChPF acts as a specific activating factor for ChSy in chondroitin polymerization. The coding region of the ChPF was divided into four discrete exons and localized to chromosome 2q35-q36. Northern blot analysis revealed that the ChPF gene exhibited a markedly different expression pattern among various human tissues, which was similar to that of ChSy. Thus, the ChPF is required for chondroitin polymerizing activity of mammalian ChSy. Chondroitin sulfate (CS) 1The abbreviations used are: CS, chondroitin sulfate; GalNAc, N-acetyl-d-galactosamine; GalNAcT, N-acetylgalactosaminyltransferase; GlcAT, glucuronyltransferase; GlcUA, d-glucuronic acid; ChSy, chondroitin synthase; ChPF, chondroitin polymerizing factor; HPLC, high-performance liquid chromatography; MES, 2-(N-morpholino)ethanesulfonic acid; TM, thrombomodulin; ΔHexUA, 4-deoxy-α-threo-hex-4-enepyranosyluronic acid. proteoglycans, at cell surfaces and in the extracellular matrix of most tissues, consist of CS chains substituted on core proteins. Recent studies of CS chains have shown that they play important roles in neural network formation in the developing mammalian brain (for reviews see Refs. 1Ohhira A. Matsui F. Tokita Y. Yamauchi S. Aono S. Arch. Biochem. Biophys. 2000; 374: 24-34Google Scholar and 2Sugahara K. Yamada S. Trends Glycosci. Glycotechnol. 2000; 12: 321-349Google Scholar) and are major inhibitory molecules affecting axon growth after spinal cord injury in the central nervous system of adult mammals (3Bradbury E.J. Moon L.D.F. Popat R.J. King V.R. Bennett G.S. Patel P.N. Fawcett J.W. McMahon S.B. Nature. 2002; 416: 636-640Google Scholar). CS biosynthesis is initiated by the addition of xylose to serine residues in the core protein, followed by the sequential addition of two Gal residues and a GlcUA residue to form the tetrasaccharide linkage structure GlcUAβ1-3Galβ1-3Galβ1-4Xylβ1-O-Ser. Chondroitin polymerization with alternating GalNAc and GlcUA then takes place, forming the repeating disaccharide region. In a previous study, in vitro chondroitin polymerization was demonstrated when α-thrombomodulin (α-TM) with a truncated glycosaminoglycan-protein linkage region tetrasaccharide sequence (4Nadanaka S. Kitagawa H. Sugahara K. J. Biol. Chem. 1998; 273: 33728-33734Google Scholar), GlcUAβ1-3Galβ1-3Galβ1-4Xyl, was tested as an acceptor together with both UDP-GlcUA and UDPGalNAc as sugar donors using a cell-free enzyme system, which was prepared from the serum-free culture medium of a human melanoma cell line (5Nadanaka S. Kitagawa H. Goto F. Tamura J. Newmann K.W. Ogawa T. Sugahara K. Biochem. J. 1999; 340: 353-357Google Scholar). We recently cloned chondroitin synthase (ChSy) consisting of a single large polypeptide with dual glycosyltransferase activities of GlcUA transferase II (GlcAT-II) and GalNAc transferase II (GalNAcT-II), which are responsible for synthesizing the repeating disaccharide units of CS (6Kitagawa H. Uyama T. Sugahara K. J. Biol. Chem. 2001; 276: 38721-38726Google Scholar). However, despite the individual transferase activities observed for ChSy, chondroitin polymerization was not demonstrated using the recombinant ChSy in vitro. We hypothesized, therefore, that an unidentified regulatory protein might be required for the full polymerization activity of ChSy and might be homologous to ChSy. To date, four homologous glycosyltransferases including ChSy, all of which are responsible for CS biosynthesis, have been cloned (6Kitagawa H. Uyama T. Sugahara K. J. Biol. Chem. 2001; 276: 38721-38726Google Scholar, 7Uyama T. Kitagawa H. Tamura J. Sugahara K. J. Biol. Chem. 2002; 277: 8841-8846Google Scholar, 8Gotoh M. Sato T. Akashima T. Iwasaki H. Kameyama A. Mochizuki H. Yada T. Inaba N. Zhang Y. Kikuchi N. Kwon Y. Togayachi A. Kudo T. Nishihara S. Watanabe H. Kimata K. Narimatsu H. J. Biol. Chem. 2002; 277: 38189-38196Google Scholar, 9Gotoh M. Yada T. Sato T. Akashima T. Iwasaki H. Mochizuki H. Inaba N. Togayachi A. Kudo T. Watanabe H. Kimata K. Narimatsu H. J. Biol. Chem. 2002; 277: 38179-38188Google Scholar, 10Uyama T. Kitagawa H. Tanaka J. Tamura J. Ogawa T. Sugahara K. J. Biol. Chem. 2003; 278: 3072-3078Google Scholar). To search for such a regulatory protein involved in chondroitin polymerization, the ChSy sequence was used to screen a data base. Here, we describe the cloning of a human cDNA encoding a novel protein required for chondroitin polymerization, designated chondroitin polymerizing factor (ChPF). Materials—UDP-[U-14C]GlcUA (285.2 mCi/mmol) and UDP-[3H]GalNAc (10 Ci/mmol) were purchased from PerkinElmer Life Sciences. Unlabeled UDP-GlcUA and UDP-GalNAc were obtained from Sigma. Chondroitin (a chemically desulfated derivative of whale cartilage CS-A) and Arthrobacter aurescens chondroitinase ABC (EC 4.2.2.4) were purchased from Seikagaku Corp. (Tokyo, Japan). Purified α-TM (11Nawa K. Sakano K. Fujiwara H. Sato Y. Sugiyama N. Teruuchi T. Iwamoto M. Marumoto Y. Biochem. Biophys. Res. Commun. 1990; 171: 729-737Google Scholar) was provided by the research institute of Dai-ichi Pharmaceutical Co. (Tokyo, Japan) and contained the linkage tetrasaccharide GlcUAβ1-3Galβ1-3Galβ1-4Xyl (4Nadanaka S. Kitagawa H. Sugahara K. J. Biol. Chem. 1998; 273: 33728-33734Google Scholar). Chondro-glycuronidase from Flavobacterium heparinum (12Hovingh P. Linker A. Biochem. J. 1977; 165: 287-293Google Scholar) was from Dr. K. Yoshida (Seikagaku Corp., Tokyo). The chemically synthesized linkage tetrasaccharide serines GlcUAβ1-3Galβ1-3Galβ1-4Xylβ1-O-Ser and GlcUAβ1-3Galβ1-3Galβ1-4Xylβ1-O-(Gly)Ser-(Gly-Glu) (13Tamura J. Neumann K.W. Ogawa T. Liebigs Ann. 1996; : 1239-1257Google Scholar) were from Dr. T. Ogawa (RIKEN, Saitama, Japan). ΔHexUAα1-3GalNAcβ1-4GlcUAβ1-3Galβ1-3Galβ1-4Xyl was isolated by HPLC after the chondroitinase ABC digestion of whale cartilage CS-A followed by mild alkaline treatment using 0.5 m LiOH, and was structurally identified by 500-MHz 1H NMR spectroscopy. It was then derivatized with a fluorophore, 2-aminobenzamide, and purified by HPLC as described previously (4Nadanaka S. Kitagawa H. Sugahara K. J. Biol. Chem. 1998; 273: 33728-33734Google Scholar, 14Sakaguchi H. Watanabe M. Ueoka C. Sugiyama E. Taketomi T. Yamada S. Sugahara K. J. Biochem. 2001; 129: 107-118Google Scholar). GalNAcβ1-4GlcUAβ1-3Galβ1-3Galβ1-4Xyl-2-aminobenzamide was prepared by the chondro-glycuronidase digestion of ΔHexUAα1-3GalNAcβ1-4GlcUAβ1-3Galβ1-3Galβ1-4Xyl-2-aminobenzamide. The Superdex™ peptide HR10/30 column was obtained from Amersham Biosciences. In Silico Cloning of the ChPF cDNA—A tBLASTn analysis of the GenBank™ data base, using the sequence of human ChSy (6Kitagawa H. Uyama T. Sugahara K. J. Biol. Chem. 2001; 276: 38721-38726Google Scholar), showed highly homologous clones. Analysis of one clone (GenBank™ accession number BC021223) revealed a single open reading frame with significant sequence similarity to human ChSy. In addition, a data base search of the Human Genome Project identified a genome sequence (GenBank™ accession number NT_005403.13) identical to the cDNA sequence. Comparison of the cDNA and genome sequences demonstrated the genomic organization of the ChPF gene. Construction of a Soluble Form of the ChPF—A cDNA fragment of a truncated form of the ChPF, lacking the first 62 amino-terminal amino acid sequence containing the putative cytoplasmic and transmembrane domains, was amplified by reverse transcription-PCR with total RNA derived from G361 human melanoma cells (ATCC number CRL-1424) as a template using a 5′-primer (5′-CGGGATCCAACTCGGTGCAGCCCGGAGC-3′) containing an in-frame BamHI site and a 3′-primer (5′-CGGGATCCGCTCTGGTTTTGGGGGAGAAG-3′) containing a Bam HI site located 46 bp downstream from the stop codon. PCR was carried out with KOD polymerase (TOYOBO, Osaka, Japan) for 32 cycles of 94 °C for 30 s, 55 °C for 30 s, and 68 °C for 180 s in 5% (v/v) dimethyl sulfoxide. The PCR fragments were subcloned into the BamHI site of pGIR201protA (15Kitagawa H. Paulson J.C. J. Biol. Chem. 1994; 269: 1394-1401Google Scholar), resulting in the fusion of the ChPF with the insulin signal sequence and the protein A sequence present in the vector. An NheI fragment containing this fusion protein sequence was inserted into the XbaI site of the expression vector pEF-BOS (16Mizushima S. Nagata S. Nucleic Acids Res. 1990; 18: 5322Google Scholar). The nucleotide sequence of the amplified cDNA was determined in an ABI PRISM 377 DNA sequencer (Applied Biosystems). Expression of a Soluble Form of the ChPF and Enzyme Assays—The expression plasmid (6.0 μg) was transfected into COS-1 cells on 100-mm plates using FuGENE 6 (Roche Applied Science) according to the manufacturer's instructions. For cotransfection experiments, the ChPF and ChSy expression plasmids (3.0 μg each) (6Kitagawa H. Uyama T. Sugahara K. J. Biol. Chem. 2001; 276: 38721-38726Google Scholar) were cotransfected into COS-1 cells on 100-mm plates using FuGENE 6 as above. Two days after transfection, 1 ml of the culture medium was collected and incubated with 10 μl of IgG-Sepharose (Amersham Biosciences) for 1 h at 4 °C. The beads recovered by centrifugation were washed with and then resuspended in the assay buffer and tested for glycosyltransferase activities as described below. Assays for GalNAcT-II and GlcAT-II were carried out using chondroitin as an acceptor and UDPGalNAc or UDP-GlcUA as a sugar donor, respectively, as described previously (17Kitagawa H. Tsutsumi K. Ujikawa M. Goto F. Tamura J. Neumann K.W. Ogawa T. Sugahara K. Glycobiology. 1997; 7: 531-537Google Scholar, 18Kitagawa H. Ujikawa M. Tsutsumi K. Tamura J. Neumann K.W. Ogawa T. Sugahara K. Glycobiology. 1997; 7: 905-911Google Scholar). Polymerization reactions using linkage tetrasaccharide-serine or -peptide and α-TM as acceptors were conducted in incubation mixtures containing the following constituents in a total volume of 20 μl, i.e. 1 nmol of a linkage tetrasaccharide-serine or α-TM, 0.25 mm UDP-[3H]GalNAc (5.28 × 105 dpm), 0.25 mm UDP-GlcUA, 100 mm MES buffer, pH 6.5, 10 mm MnCl2, and 10 μl of the resuspended beads. The mixtures were incubated at 37 °C overnight, and the 3H-labeled products were then separated by gel filtration chromatography on a Superdex peptide column equilibrated and eluted with 0.2 m NH4HCO3. Fractions (0.4 ml each) were collected at a rate of 0.4 ml/min, and the measurement of radioactivity was carried out by liquid scintillation counting. Characterization of the Reaction Products—Products of the polymerization reactions on α-TM were isolated by gel filtration on a Superdex peptide column with 0.2 m NH4HCO3 as the eluent. The [3H]GalNAclabeled oligosaccharide chains released from α-TM by mild alkaline treatment using 0.5 m LiOH were derivatized with 2-aminobenzamide (4Nadanaka S. Kitagawa H. Sugahara K. J. Biol. Chem. 1998; 273: 33728-33734Google Scholar, 14Sakaguchi H. Watanabe M. Ueoka C. Sugiyama E. Taketomi T. Yamada S. Sugahara K. J. Biochem. 2001; 129: 107-118Google Scholar) and then exhaustively digested with chondroitinase ABC using 50 mIU of the enzyme for 1 h as described previously (19Sugahara K. Shigeno K. Masuda M. Fujii N. Kurosaka A. Takeda K. Carbohydr. Res. 1994; 255: 145-163Google Scholar). The enzyme digest was subjected to gel filtration on Superdex peptide as described above. The radioactive peak of the 2-aminobenzamide-derivatized putative linkage hexasaccharide and its digest with chondro-glycuronidase were analyzed using an anion-exchange HPLC on an amine-bound silica column as described previously (20Sugahara K. Okumura Y. Yamashina T. Biochem. Biophys. Res. Commun. 1989; 162: 189-197Google Scholar). The identification of the reaction products was accomplished by co-chromatography with authentic linkage hexa and pentasaccharides labeled with 2-aminobenzamide. Northern Blot Analysis—A commercial human 12-lane multiple tissue Northern blot (Clontech) membrane was used for the analysis. The membrane was probed with a gel-purified, radiolabeled (>1 × 109 cpm/μg), 0.84-kb ChPF-specific fragment corresponding to nucleotides 1720–2558 of the ChPF cDNA (GenBank™ accession number AB095813). In Silico Cloning of a ChPF cDNA—We recently identified and characterized human ChSy (6Kitagawa H. Uyama T. Sugahara K. J. Biol. Chem. 2001; 276: 38721-38726Google Scholar). Screening of the nonredundant data base at the National Center for Biotechnology Information (National Institutes of Health, Bethesda, MD), using the deduced amino acid sequence of human ChSy, identified a few clones. One of them, designated ChPF here (GenBank™ accession number BC021223), contained a 5′-untranslated region of 236 bp, a single open reading frame of 2325 bp coding for a protein of 775 amino acids with three potential N-glycosylation sites (Fig. 1), and a 3′-untranslated region of about 0.5 kb with a presumptive polyadenylation signal. Northern blot analysis showed that the mRNA was about 3.4 kb in length in various human tissues (see below), suggesting that the cDNA was approximately full-length. The deduced amino acid sequence corresponded to a 85,494-Da polypeptide. The predicted translation initiation site conformed to the Kozak consensus sequence for initiation (21Kozak M. Nucleic Acids Res. 1984; 12: 857-872Google Scholar). A Kyte-Doolittle hydropathy analysis (22Kyte J. Doolittle R.F. J. Mol. Biol. 1982; 157: 105-132Google Scholar) revealed one prominent hydrophobic segment of 23 amino acid residues in the NH2-terminal region, predicting that the protein has a type II transmembrane topology (Fig. 1). Data base searches suggested that the amino acid sequence displayed 57 and 23% identity to human chondroitin GlcAT (9Gotoh M. Yada T. Sato T. Akashima T. Iwasaki H. Mochizuki H. Inaba N. Togayachi A. Kudo T. Watanabe H. Kimata K. Narimatsu H. J. Biol. Chem. 2002; 277: 38179-38188Google Scholar) and ChSy (6Kitagawa H. Uyama T. Sugahara K. J. Biol. Chem. 2001; 276: 38721-38726Google Scholar), respectively (Fig. 1). Notably, however, the identified protein did not have the conserved DXD motif found in most glycosyltransferases (23Wiggins S. Munro S. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 7945-7950Google Scholar). Thus, the features of the identified protein sequence suggest that the identified gene product might be involved in CS biosynthesis but would not possess a glycosyltransferase activity. Intriguingly, a homologue of the identified human gene was found in the Caenorhabditis elegans and Drosophila genome, sharing 27 and 25% identity with that of C. elegans and Drosophila, respectively (data not shown). Genomic Organization and Chromosomal Localization— Comparison of the identified cDNA sequence with the genome sequence deposited in the Human Genome Project data base revealed the genomic structure and chromosomal localization of the gene. The gene spans over 5 kb, and the coding region of the gene was divided into four discrete exons as shown in Fig. 2. Its genomic organization is very similar to that of the chondroitin GlcAT gene, which consists of four discrete exons in the coding region (9Gotoh M. Yada T. Sato T. Akashima T. Iwasaki H. Mochizuki H. Inaba N. Togayachi A. Kudo T. Watanabe H. Kimata K. Narimatsu H. J. Biol. Chem. 2002; 277: 38179-38188Google Scholar). The intron/exon junctions followed the GT/AG rule (24Breathnach R. Chambon P. Annu. Rev. Biochem. 1981; 50: 349-383Google Scholar) and were flanked by conserved sequences (data not shown). This gene is located on human chromosome 2q35-q36, whereas other homologous genes have been localized to different chromosomes as follows: ChSy, 15q26.3; chondroitin GalNAcT-1, 8q21.3; chondroitin GalNAcT-2, 10q11.22; and chondroitin GlcAT, 7q35 (10Uyama T. Kitagawa H. Tanaka J. Tamura J. Ogawa T. Sugahara K. J. Biol. Chem. 2003; 278: 3072-3078Google Scholar). Expression and Characterization of ChPF—To facilitate the functional analysis of the ChPF, a soluble form of the protein was generated by replacing the first 62 amino acids of the protein with a cleavable insulin signal sequence and a protein A IgG-binding domain as described under "Experimental Procedures," and then the soluble protein was expressed in COS-1 cells as a recombinant protein fused with the protein A IgG-binding domain. When the expression plasmid was expressed in COS-1 cells, a ∼95-kDa protein was secreted as shown by Western blotting using IgG (data not shown). The fused protein expressed in the medium was adsorbed onto IgG-Sepharose beads to eliminate endogenous glycosyltransferases, and then the protein-bound beads were used as an enzyme source. Although the bound fusion protein was assayed for glycosyltransferase activity using a variety of acceptors and either UDPGalNAc or UDP-GlcUA as a donor substrate, little glycosyltransferase activity was detected. However, coexpression of the soluble ChPF with the soluble ChSy augmented both the GalNAcT-II and GlcAT-II activities of ChSy over 20-fold, which is involved in the synthesis of the repeating disaccharide units of chondroitin (Table I). Notably, these effects of coexpression were not due to differences in the expression levels of these proteins, as assessed by Western blot analysis (data not shown). Moreover, simple mixing of the soluble proteins, which were expressed in COS-1 cells separately transfected with ChPF or ChSy, did not result in any increase in the catalytic activities (Table I).Table IGalNAcT-II and GlcAT-II activities of the fusion proteins secreted into the culture medium by transfected COS-1 cellsProteinGalNAcT-II activityaPolymer chondroitin was used as an acceptor substrate, and the values represent the averages of three independent experiments.GlcAT-II activityaPolymer chondroitin was used as an acceptor substrate, and the values represent the averages of three independent experiments.pmol/ml medium/hpmol/ml medium/hChSy1.32.3ChPF0.2NDbND, not detected (<0.1 pmol/ml medium/h).ChSy/ChPF31188.4ChSy + ChPFcThe values obtained by mixing soluble proteins derived from COS-1 cells separately transfected by ChPF and ChSy.0.51.1a Polymer chondroitin was used as an acceptor substrate, and the values represent the averages of three independent experiments.b ND, not detected (<0.1 pmol/ml medium/h).c The values obtained by mixing soluble proteins derived from COS-1 cells separately transfected by ChPF and ChSy. Open table in a new tab In a previous study (5Nadanaka S. Kitagawa H. Goto F. Tamura J. Newmann K.W. Ogawa T. Sugahara K. Biochem. J. 1999; 340: 353-357Google Scholar), in vitro chondroitin polymerization was demonstrated when α-TM with a truncated glycosaminoglycan-protein linkage tetrasaccharide (4Nadanaka S. Kitagawa H. Sugahara K. J. Biol. Chem. 1998; 273: 33728-33734Google Scholar), GlcUAβ1-3Galβ1-3Galβ1-4Xyl, was used as an acceptor together with a cell-free enzyme system, which was prepared from the serum-free culture medium of a human melanoma cell line. However, despite the individual GalNAcT-II and GlcAT-II activities observed for ChSy, chondroitin polymerization was not demonstrated using α-TM with the recombinant soluble ChSy in vitro (6Kitagawa H. Uyama T. Sugahara K. J. Biol. Chem. 2001; 276: 38721-38726Google Scholar). Hence, it was next investigated in this study whether the coexpression of ChPF and ChSy is required for chondroitin polymerization. Incubations of the coexpressed proteins with various acceptors in the presence of UDP-[3H]GalNAc and UDP-GlcUA yielded radiolabeled polymer chondroitin chains as shown in Fig. 3; the acceptors used were α-TM and the chemically synthesized linkage tetrasaccharide derivatives GlcUAβ1-3Galβ1-3Galβ1-4Xylβ1-O-Ser and GlcUAβ1-3Galβ1-3Galβ1-4Xylβ1-O-(Gly)Ser-(Gly-Glu). The reaction products obtained with α-TM as an acceptor were subjected to reductive β-elimination using NaBH4/NaOH (25Sugahara K. Masuda M. Harada T. Yamashina I. de Waard P. Vliegenthart J.F.G. Eur. J. Biochem. 1991; 202: 805-811Google Scholar), and the released radiolabeled oligosaccharide chains were analyzed by gel chromatography using a column of Superdex peptide or 75 (Fig. 3). The saccharide chains synthesized on α-TM were much longer than those synthesized on the tetrasaccharide-serine or -peptide and were comparable with commercial polymer chondroitin chains (Fig. 3B). In contrast, simple mixing of the soluble proteins, which were expressed in COS-1 cells separately transfected with ChPF or ChSy, did not produce such chondroitin polymerizing activity. These results suggest that the ChPF and ChSy proteins need to be synthesized simultaneously in the same cell to polymerize chondroitin chains and that the core protein is required for the synthesis of the longer chondroitin chains. Analysis of the Chondroitin Chains Newly Polymerized on the Tetrasaccharide Linkage Region of α-TM—The [3H]GalNAclabeled chondroitin chains obtained from the polymerization reactions on α-TM were liberated by mild alkaline treatment using 0.5 m LiOH as described under "Experimental Procedures." The resulting radiolabeled chondroitin chains were derivatized with 2-aminobenzamide and then isolated by gel filtration (4Nadanaka S. Kitagawa H. Sugahara K. J. Biol. Chem. 1998; 273: 33728-33734Google Scholar). The isolated chains were digested with chondroitinase ABC and subjected to gel chromatography, resulting in Fraction 1 containing the linkage hexasaccharides and Fraction 2 containing both di and monosaccharides in a molar ratio of 1:47 (Fig. 4A). These results suggested that the average size of the synthesized chondroitin chains was ∼100-mer. Fraction 1 was then subjected to anion-exchange HPLC, resulting in only one 3H-labeled peak at the elution position of the authentic linkage hexasaccharide ΔHexUAα1-3GalNAcβ1-4GlcUAβ1-3Galβ1-3Galβ1-4Xyl-2-aminobenzamide, which was shifted to a position of GalNAcβ1-4GlcUAβ1-3Galβ1-3Galβ1-4Xyl-2-aminobenzamide upon digestion with chondro-glycuronidase (12Hovingh P. Linker A. Biochem. J. 1977; 165: 287-293Google Scholar) (Fig. 4B). These results clearly indicated that the polymerization did take place on the truncated linkage tetrasaccharide sequence of α-TM. Expression Pattern of ChPF—Northern blot analysis of mRNA demonstrated a single band of ∼3.4 kb for all human tissues except peripheral blood leukocytes (Fig. 5). The gene exhibited a differential expression in the human tissues examined, with strong expression in the placenta and heart in addition to moderate expression in the brain, skeletal muscle, kidney, and liver. Notably, the expression pattern was similar to that of ChSy. The findings of the present study demonstrated that the chondroitin polymerizing activity requires the coexpression of a unique protein, ChPF, with ChSy. Although ChPF itself harbors little GalNAcT-II or GlcAT-II activity, the protein sequence of ChPF is homologous to the four previously cloned glycosyltransferases involved in CS biosynthesis. Thus, ChPF is the fifth member of this novel gene family, which is responsible for initiation and elongation of CS chains. In a previous study (5Nadanaka S. Kitagawa H. Goto F. Tamura J. Newmann K.W. Ogawa T. Sugahara K. Biochem. J. 1999; 340: 353-357Google Scholar), in vitro chondroitin polymerization was demonstrated when α-TM was used as an acceptor together with a cell-free enzyme system, which was prepared from the serum-free culture medium of a human melanoma cell line. In addition, based on the substrate competition experiments described in a previous study (5Nadanaka S. Kitagawa H. Goto F. Tamura J. Newmann K.W. Ogawa T. Sugahara K. Biochem. J. 1999; 340: 353-357Google Scholar), it was proposed that a chondroitin polymerase with three enzyme activities, i.e. GalNAcT-I for chondroitin chain initiation and GalNAcT-II and GlcAT-II for chain elongation, might exist. In view of the present finding that the coexpression of ChPF and ChSy was required for chondroitin polymerization on the linkage region tetrasaccharide of α-TM, the proposed chondroitin polymerase does not appear to be a single polypeptide but appears to be a protein complex consisting of ChPF and ChSy. In this context, the recently cloned chondroitin GalNAcT-1 and -2 (7Uyama T. Kitagawa H. Tamura J. Sugahara K. J. Biol. Chem. 2002; 277: 8841-8846Google Scholar, 8Gotoh M. Sato T. Akashima T. Iwasaki H. Kameyama A. Mochizuki H. Yada T. Inaba N. Zhang Y. Kikuchi N. Kwon Y. Togayachi A. Kudo T. Nishihara S. Watanabe H. Kimata K. Narimatsu H. J. Biol. Chem. 2002; 277: 38189-38196Google Scholar, 10Uyama T. Kitagawa H. Tanaka J. Tamura J. Ogawa T. Sugahara K. J. Biol. Chem. 2003; 278: 3072-3078Google Scholar), both of which harbored both GalNAcT-I and -II activities, were not required for the chondroitin polymerization on the linkage region tetrasaccharide sequence of α-TM. In addition, coexpression of chondroitin GalNAcT-1 or -2 and ChSy yielded no chondroitin polymerizing activity, even when α-TM was used as an acceptor substrate. Moreover, coexpression of chondroitin GalNAcT-1, ChSy, and ChPF or chondroitin GalNAcT-2, ChSy, and ChPF did not augment chondroitin polymerization activity (data not shown). Therefore, chondroitin GalNAcT-1 and -2 may be dispensable for the chondroitin polymerization itself, although their involvement in CS biosynthesis cannot be excluded. In fact, homologues of the human ChSy and ChPF are present in C. elegans, but those of GalNAcT-1 and -2 are absent despite the existence of chondroitin in the worm (26Yamada S. Van Die I. Van den Eijnden D.H. Yokota A. Kitagawa H. Sugahara K. FEBS Lett. 1999; 459: 327-331Google Scholar), supporting the present hypothesis. Thus, more efforts to determine the specific roles of chondroitin GalNAcT-1 and -2 in CS biosynthesis will be required. As described previously (6Kitagawa H. Uyama T. Sugahara K. J. Biol. Chem. 2001; 276: 38721-38726Google Scholar), the GlcAT-II and GalNAcT-II activities observed for ChSy were low with polymer chondroitin and chondro-oligosaccharides as acceptor substrates (also see Table I). Considering that coexpression of the soluble ChPF with the soluble ChSy augmented both the GalNAcT-II and GlcAT-II activities of ChSy over 20-fold (Table I), it has become evident that coexpression of ChPF was also required for full GlcAT-II and GalNAcT-II activities of ChSy. In addition, the present study confirmed the previous findings (5Nadanaka S. Kitagawa H. Goto F. Tamura J. Newmann K.W. Ogawa T. Sugahara K. Biochem. J. 1999; 340: 353-357Google Scholar) that the chondroitin chain polymerization reactions were also influenced by the core protein portion. The oligosaccharide chains synthesized on α-TM were much longer than those synthesized on the linkage tetrasaccharide-derivatives GlcUAβ1-3Galβ1-3Galβ1-4Xylβ1-O-Ser and GlcUAβ1-3Galβ1-3Galβ1-4Xylβ1-O-(Gly)Ser-(Gly-Glu) (see Fig. 3). In this regard, Oldberg et al. (27Oldberg Å. Antonsson P. Moses J. Fransson L.-Å. FEBS Lett. 1996; 386: 29-32Google Scholar) reported that cells transformed with vectors encoding deletion variants of decorin synthesized proteoglycan with shorter glycosaminoglycan chains and suggested that such effects on the glycosaminoglycan chain length might be due to a lower affinity between the core protein and the glycosyltransferases. Thus, in view of the present findings that the chondroitin chain polymerization reactions were markedly stimulated by the core protein of the acceptor, it is possible that the protein complex consisting of ChPF and ChSy interacts with the core protein and regulates the glycosaminoglycan chain length. It should also be noted that animal cells utilize β-d-xylosides as primers for glycosaminoglycan synthesis, and, thus, the core protein can be substituted by the mildly hydrophobic xylosides such as 4-methylumbelliferyl- or p-nitrophenyl-β-d-xyloside (28Okayama M. Kimata K. Suzuki S. J. Biochem. 1973; 74: 1069-1073Google Scholar). In addition, Fernández and Warren (29Fernández C.J. Warren G. J. Biol. Chem. 1998; 273: 19030-19039Google Scholar) demonstrated CS synthesis using reconstituted Golgi in the presence of 4-methylumbelliferyl-β-d-xyloside. In fact, in vitro chondroitin polymerization was also observed in this study when GlcAβ1-3Galβ1-O-C2H4NH-benzyloxycarbonyl was used as an acceptor substrate with the recombinant soluble proteins secreted from COS-1 cells cotransfected with ChPF and ChSy. However, the oligosaccharide chains synthesized on GlcAβ1-3Galβ1-O-C2H4NH-benzyloxycarbonyl were slightly shorter than those synthesized on α-TM (data not shown). It appears that hydrophobic aglycones mimic a recognition sequence on core proteins and interact with the protein complex consisting of ChPF and ChSy. Further work should be aimed at better defining the acceptor recognition requirements by the protein complex consisting of ChPF and ChSy. Earlier studies have indicated that the sulfation of glycosaminoglycan moieties ordinarily proceeds together with polymerization at a single Golgi site and that there appears to be a close interrelation between sulfation and polymer elongation/termination (30Silbert J.E. Sugumaran G. Biochim. Biophys. Acta. 1995; 1241: 371-384Google Scholar). In fact, although it has been reported that prior sulfation of the growing CS chain profoundly affects chain polymerization (17Kitagawa H. Tsutsumi K. Ujikawa M. Goto F. Tamura J. Neumann K.W. Ogawa T. Sugahara K. Glycobiology. 1997; 7: 531-537Google Scholar, 18Kitagawa H. Ujikawa M. Tsutsumi K. Tamura J. Neumann K.W. Ogawa T. Sugahara K. Glycobiology. 1997; 7: 905-911Google Scholar, 30Silbert J.E. Sugumaran G. Biochim. Biophys. Acta. 1995; 1241: 371-384Google Scholar), the present study clearly indicated that sulfation is not an absolute requirement for polymer elongation. The protein-coding sequences of the human chondroitin GlcAT and ChPF genes are distributed among four exons that span ∼7 kb (9Gotoh M. Yada T. Sato T. Akashima T. Iwasaki H. Mochizuki H. Inaba N. Togayachi A. Kudo T. Watanabe H. Kimata K. Narimatsu H. J. Biol. Chem. 2002; 277: 38179-38188Google Scholar) and 5 kb (this study), respectively. Comparison of the genomic organizations of these two genes shows a quite similar genetic exon-intron organization within the coding sequences (Fig. 2). Moreover, the amino acid sequence of ChPF was the most homologous to that of chondroitin GlcAT among the five identified proteins involved in CS biosynthesis (Fig. 1). However, coexpression of chondroitin GlcAT and ChSy yielded little chondroitin polymerization activity when α-TM was used as an acceptor substrate (data not shown). These findings suggest that ChPF acts as a specific activating factor for ChSy in chondroitin polymerization. The identification of ChPF, a unique protein factor required for chondroitin polymerization activity, may shed light on the molecular basis of CS biosynthesis and the evaluation of the biological functions of CS and dermatan sulfate because the gene knockout of either ChPF or ChSy or both would result in specific elimination of CS and dermatan sulfate. We thank Dr. T. Ogawa for the enzyme substrates. We also thank K. Tsurumoto for technical assistance.
Referência(s)