Advertisement

Chondroitin Sulfate Synthase-2

MOLECULAR CLONING AND CHARACTERIZATION OF A NOVEL HUMAN GLYCOSYLTRANSFERASE HOMOLOGOUS TO CHONDROITIN SULFATE GLUCURONYLTRANSFERASE, WHICH HAS DUAL ENZYMATIC ACTIVITIES*
  • Toshikazu Yada
    Footnotes
    Affiliations
    Institute for Molecular Science of Medicine, Aichi Medical University, Nagakute, Aichi 480-1195

    Seikagaku Corp., 1253 Tateno 3-Chome, Higashi-yamato, Tokyo 207-0021
    Search for articles by this author
  • Masanori Gotoh
    Footnotes
    Affiliations
    Glycogen Functional Team, Research Center for Glycoscience, National Institute of Advanced Industrial Science and Technology, OSL C-2, 1-1-1 Umezono, Tsukuba, Ibaraki 305-8568

    Amersham Biosciences KK, 3-25-1, Hyakunincho, Shinjuku-ku, Tokyo 169-0073
    Search for articles by this author
  • Takashi Sato
    Affiliations
    Glycogen Functional Team, Research Center for Glycoscience, National Institute of Advanced Industrial Science and Technology, OSL C-2, 1-1-1 Umezono, Tsukuba, Ibaraki 305-8568
    Search for articles by this author
  • Masafumi Shionyu
    Affiliations
    Division of Biological Science, Graduate School of Science, Nagoya University, Furo-cho, Chikusa-ku, Nagoya, Aichi 464-8602
    Search for articles by this author
  • Mitiko Go
    Affiliations
    Division of Biological Science, Graduate School of Science, Nagoya University, Furo-cho, Chikusa-ku, Nagoya, Aichi 464-8602
    Search for articles by this author
  • Hiromi Kaseyama
    Affiliations
    Institute for Molecular Science of Medicine, Aichi Medical University, Nagakute, Aichi 480-1195

    Seikagaku Corp., 1253 Tateno 3-Chome, Higashi-yamato, Tokyo 207-0021
    Search for articles by this author
  • Hiroko Iwasaki
    Affiliations
    Glycogen Functional Team, Research Center for Glycoscience, National Institute of Advanced Industrial Science and Technology, OSL C-2, 1-1-1 Umezono, Tsukuba, Ibaraki 305-8568

    Amersham Biosciences KK, 3-25-1, Hyakunincho, Shinjuku-ku, Tokyo 169-0073
    Search for articles by this author
  • Norihiro Kikuchi
    Affiliations
    Glycogen Functional Team, Research Center for Glycoscience, National Institute of Advanced Industrial Science and Technology, OSL C-2, 1-1-1 Umezono, Tsukuba, Ibaraki 305-8568

    Mitsui Knowledge Industry Co., Ltd., 1-32-2 Honcho, Nakano-ku, Tokyo 164-8721
    Search for articles by this author
  • Yeon-Dae Kwon
    Affiliations
    Glycogen Functional Team, Research Center for Glycoscience, National Institute of Advanced Industrial Science and Technology, OSL C-2, 1-1-1 Umezono, Tsukuba, Ibaraki 305-8568

    Mitsui Knowledge Industry Co., Ltd., 1-32-2 Honcho, Nakano-ku, Tokyo 164-8721
    Search for articles by this author
  • Akira Togayachi
    Affiliations
    Glycogen Functional Team, Research Center for Glycoscience, National Institute of Advanced Industrial Science and Technology, OSL C-2, 1-1-1 Umezono, Tsukuba, Ibaraki 305-8568

    New Energy and Industrial Technology Development Organization, Sunshine 60 Bldg., 3-1-1 Higashi Ikebukuro, Toshima-ku, Tokyo 170-6028, Japan
    Search for articles by this author
  • Takashi Kudo
    Affiliations
    Glycogen Functional Team, Research Center for Glycoscience, National Institute of Advanced Industrial Science and Technology, OSL C-2, 1-1-1 Umezono, Tsukuba, Ibaraki 305-8568
    Search for articles by this author
  • Hideto Watanabe
    Affiliations
    Institute for Molecular Science of Medicine, Aichi Medical University, Nagakute, Aichi 480-1195
    Search for articles by this author
  • Hisashi Narimatsu
    Affiliations
    Glycogen Functional Team, Research Center for Glycoscience, National Institute of Advanced Industrial Science and Technology, OSL C-2, 1-1-1 Umezono, Tsukuba, Ibaraki 305-8568
    Search for articles by this author
  • Koji Kimata
    Correspondence
    To whom correspondence should be addressed: Institute for Molecular Science of Medicine, Aichi Medical University, Yazako, Nagakute, Aichi 480-1195, Japan. Tel.: 81-561-62-3311; Fax: 81-561-63-3532
    Affiliations
    Institute for Molecular Science of Medicine, Aichi Medical University, Nagakute, Aichi 480-1195
    Search for articles by this author
  • Author Footnotes
    * This work was performed as part of the R&D Project of the Industrial Science and Technology Frontier Program (R&D for the Establishment and Utilization of a Technical Infrastructure for Japanese Industry) supported by the New Energy and Industrial Technology Development Organization (NEDO). This work was also supported by grants-in-aid for Scientific Research on Priority Areas (C) “Genome Information Science” and for Scientific Research (B) (to M. G.) and by the Ministry of Education, Culture, Sports, Science and Technology of Japan. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
    ¶ Both authors contributed equally to this work as first authors.
Open AccessPublished:May 20, 2003DOI:https://doi.org/10.1074/jbc.M303657200
      Chondroitin sulfate is found in a variety of tissues as proteoglycans and consists of repeating disaccharide units of N-acetylgalactosamine and glucuronic acid residues with sulfate residues at various places. We found a novel human gene (GenBank™ accession number AB086063) that possesses a sequence homologous with the human chondroitin sulfate glucuronyltransferase gene which we recently cloned and characterized. The full-length open reading frame encodes a typical type II membrane protein comprising 775 amino acids. The protein had a domain containing β3-glycosyltransferase motif but lacked a typical β4-glycosyltransferase motif, which is the same as chondroitin sulfate glucuronyltransferase, whereas chondroitin synthase had both domains. The putative catalytic domain was expressed in COS-7 cells as a soluble enzyme. Surprisingly, both glucuronyltransferase and N-acetylgalactosaminyltransferase activities were observed when chondroitin, chondroitin sulfate, and their oligosaccharides were used as the acceptor substrates. The reaction products were identified to have the linkage of GlcUAβ1–3GalNAc and GalNAcβ1–4GlcUA at the non-reducing terminus of chondroitin for glucuronyltransferase activity and N-acetylgalactosaminyltransferase activity, respectively. Quantitative real time PCR analysis revealed that the transcripts were ubiquitously expressed in various human tissues but highly expressed in the pancreas, ovary, placenta, small intestine, and stomach. These results indicate that this enzyme could synthesize chondroitin sulfate chains as a chondroitin sulfate synthase that has both glucuronyltransferase and N-acetylgalactosaminyltransferase activities. Sequence analysis based on three-dimensional structure revealed the presence of not typical but significant β4-glycosyltransferase architecture.
      Chondroitin sulfate (CS)
      The abbreviations used are: CS, chondroitin sulfate; (β3)GlcAT, (β1,3)-glucuronyltransferase; (β3 or β4)-Gal-T, (β1,3 or β1,4)-galactosyltransferase; (β4)GalNAc-T, (β1,4)-N-acetylgalactosaminyltransferase; EST, expressed sequence tag; GAPDH, glyceraldehydes-3-phosphate dehydrogenase; MES, 2-(N-morpholino)ethanesulfonic acid.
      1The abbreviations used are: CS, chondroitin sulfate; (β3)GlcAT, (β1,3)-glucuronyltransferase; (β3 or β4)-Gal-T, (β1,3 or β1,4)-galactosyltransferase; (β4)GalNAc-T, (β1,4)-N-acetylgalactosaminyltransferase; EST, expressed sequence tag; GAPDH, glyceraldehydes-3-phosphate dehydrogenase; MES, 2-(N-morpholino)ethanesulfonic acid.
      proteoglycans are located in the extracellular matrix and on cell surfaces of various kinds of human tissues. Some of the chondroitin sulfate proteoglycans provide high osmotic pressure and water retention, and others modulate not only cell adhesion to extracellular matrix, cell migration, cell proliferation, and morphogenesis but also some cytokine signals (
      • Schwartz N.B.
      • Domowicz M.
      ,
      • Bandtlow C.E.
      • Zimmermann D.R.
      ). A few general features of the biosynthetic assembly of chondroitin sulfate proteoglycans are as follows: (i) the sequential synthesis of the core protein; (ii) xylosylation of specific Ser moieties of the core protein; (iii) addition of two Gal residues to the Xyl; (iv) completion of a common tetrasaccharide linkage region by addition of a GlcUA residue; (v) addition of GalNAc residue to initiate the chondroitin/dermatan sulfate biosynthesis; (vi) repeated addition of GlcUA residues alternating with GalNAc residues to grow to the large heteropolymer glycosaminoglycan chains; and (vii) modification of these growing glycosaminoglycan chains by O-sulfation at various places, and by epimerization of some of GlcUA residues to IdoUA residues.
      The assembly of the linkage region on the core protein followed by glycosaminoglycan polymerization and modification occurs in the intracellular membrane system composed of the endoplasmic reticulum and Golgi apparatus (
      • Kimata K.
      • Okayama M.
      • Ooira A.
      • Suzuki S.
      ,
      • Silbert J.E.
      • Sugumaran G.
      ). With the exception of the polysaccharide chain-initiating Xyl transferase, which is found partially in the endoplasmic reticulum (
      • Vertel B.M.
      • Walters L.M.
      • Flay N.
      • Kearns A.E.
      • Schwartz N.B.
      ), all the enzymes are firmly attached to the Golgi membranes and may work in an orchestrated manner, but some are found in serum or the culture medium of cells (
      • Silbert J.E.
      • Sugumaran G.
      ,
      • Inoue H.
      • Otsu K.
      • Yoneda M.
      • Kimata K.
      • Suzuki S.
      • Nakanishi Y.
      ). The enzymes responsible for the synthesis of the linkage region of proteoglycans, Xyl transferase (
      • Gotting C.
      • Kuhn J.
      • Zahn R.
      • Brinkmann T.
      • Kleesiek K.
      ), Gal transferase I (
      • Almeida R.
      • Levery S.B.
      • Mandel U.
      • Kresse H.
      • Schwientek T.
      • Bennett E.P.
      • Clausen H.
      ,
      • Okajima T.
      • Yoshida K.
      • Kondo T.
      • Furukawa K.
      ), Gal transferase II (
      • Bai X.
      • Zhou D.
      • Brown J.R.
      • Crawford B.E.
      • Hennet T.
      • Esko J.D.
      ), and GlcUA transferase I (
      • Kitagawa H.
      • Tone Y.
      • Tamura J.
      • Neumann K.W.
      • Ogawa T.
      • Oka S.
      • Kawasaki T.
      • Sugahara K.
      ,
      • Wei G.
      • Bai X.
      • Sarkar A.K.
      • Esko J.D.
      ), which act sequentially to transfer Xyl, Gal, Gal, and GlcUA from their respective sugar nucleotide precursors to the acceptor core protein, have been cloned. We have been interested in the modification reactions, especially sulfations, because specific regional structures raised by the modifications determine the capacity of chondroitin sulfate to interact with other molecules including cytokines and regulate their assembly and activities in extracellular and pericellular matrices (
      • Maimone M.M.
      • Tollefsen D.M.
      ,
      • Lyon M.
      • Deakin J.A.
      • Rahmoune H.
      • Fernig D.G.
      • Nakamura T.
      • Gallagher J.T.
      ,
      • Razin E.
      • Stevens R.L.
      • Akiyama F.
      • Schmid K.
      • Austen K.F.
      ,
      • Maeda N.
      • Ichihara-Tanaka K.
      • Kimura T.
      • Kadomatsu K.
      • Muramatsu T.
      • Noda M.
      ,
      • Naujokas M.F.
      • Morin M.
      • Anderson M.S.
      • Peterson M.
      • Miller J.
      ,
      • Fried M.
      • Duffy P.E.
      ). Except for chondroitin C5-epimerase, most of modifying enzymes for chondroitin sulfate biosynthesis, such as chondroitin O-sulfotransferases including chondroitin 4-O-sulfotransferase (
      • Yamauchi S.
      • Mita S.
      • Matsubara T.
      • Fukuta M.
      • Habuchi H.
      • Kimata K.
      • Habuchi O.
      ), chondroitin 6-O-sulfotransferase (
      • Uchimura K.
      • Muramatsu H.
      • Kadomatsu K.
      • Fan Q.W.
      • Kurosawa N.
      • Mitsuoka C.
      • Kannagi R.
      • Habuchi O.
      • Muramatsu T.
      ), uronyl 2-O-sulfotransferase (
      • Kobayashi M.
      • Sugumaran G.
      • Liu J.
      • Shworak N.W.
      • Silbert J.E.
      • Rosenberg R.D.
      ), and N-acetylgalactosamine 4-sulfate 6-O-sulfotransferase (
      • Ohtake S.
      • Ito Y.
      • Fukuta M.
      • Habuchi O.
      ), have been cloned. The sulfation of chondroitin sulfate ordinarily proceeds together with polymerization at the Golgi apparatus. Thus, in order to address control mechanisms of the sulfation, we should also study the enzymes for the chain synthesis, especially chondroitin sulfate elongation enzymes.
      Recent progress with the human genome project and the expansion of other data bases such as expressed sequence tags (ESTs) and full-length cDNAs has enabled the search for novel genes that are homologous to known genes. Kitagawa et al. (
      • Kitagawa H.
      • Uyama T.
      • Sugahara K.
      ) identified a human chondroitin synthase, from the HUGE (human unidentified gene-encoded large proteins) protein data base by screening with the keywords “one transmembrane domain” and “galactosyltransferase family.” This enzyme had the dual glycosyltransferase activities of glucuronyltransferase II (GlcAT-II) and N-acetylgalactosaminyltransferase II (GalNAcT-II) responsible for synthesizing the repeated disaccharide units of chondroitin sulfate (
      • Kitagawa H.
      • Uyama T.
      • Sugahara K.
      ). By a similar homology search of the data bases, four enzymes including chondroitin synthase have further been cloned and characterized. Chondroitin sulfate GalNAcT-1 (CSGalNAcT-1) and chondroitin sulfate GalNAcT-2 (CSGalNAcT-2), the second and fourth chondroitin glycosyltransferases cloned, respectively, exhibit both GalNAcT-II activity for chain elongation and GalNAcT-I activity that determine and initiate the synthesis of chondroitin sulfate in the common linkage region (
      • Uyama T.
      • Kitagawa H.
      • Tanaka J.
      • Tamura J.I.
      • Ogawa T.
      • Sugahara K.
      ,
      • Sato T.
      • Gotoh M.
      • Kiyohara K.
      • Akashima T.
      • Iwasaki H.
      • Kameyama A.
      • Mochizuki H.
      • Yada T.
      • Inaba N.
      • Togayachi A.
      • Kudo T.
      • Asada M.
      • Watanabe H.
      • Imamura T.
      • Kimata K.
      • Narimatsu H.
      ,
      • Gotoh M.
      • Sato T.
      • Akashima T.
      • Iwasaki H.
      • Kameyama A.
      • Mochizuki H.
      • Yada T.
      • Inaba N.
      • Zhang Y.
      • Kikuchi N.
      • Kwon Y.D.
      • Togayachi A.
      • Kudo T.
      • Nishihara S.
      • Watanabe H.
      • Kimata K.
      • Narimatsu H.
      ,
      • Uyama T.
      • Kitagawa H.
      • Tamura Ji J.
      • Sugahara K.
      ). Chondroitin sulfate GlcUA transferase (CSGlcAT), the third chondroitin glycosyltransferase cloned, has only GlcAT-II activity, which has been proposed to be involved in chain elongation (
      • Gotoh M.
      • Yada T.
      • Sato T.
      • Akashima T.
      • Iwasaki H.
      • Mochizuki H.
      • Inaba N.
      • Togayachi A.
      • Kudo T.
      • Watanabe H.
      • Kimata K.
      • Narimatsu H.
      ). Therefore, more than four enzymes are likely responsible for chondroitin/dermatan sulfate biosynthesis, and they form a gene family, like the EXT family for heparin/heparan sulfate biosynthesis (
      • Zak B.M.
      • Crawford B.E.
      • Esko J.D.
      ).
      In the present study, a search of the data bases using the amino acid sequence of CSGlcAT revealed a novel gene whose product was characterized as the fifth enzyme to possess high homology with CSGlcAT. Interestingly, despite its homology with CSGlcAT, this enzyme, designated CSS2, shows both GlcAT-II and GalNAcT-II activity toward the non-reducing terminal residue of chondroitin/chondroitin sulfate with the specific linkage structure.

      EXPERIMENTAL PROCEDURES

      Materials—UDP-[14C]GlcUA (313 mCi/mmol) and UDP-[3H]Gal (20 Ci/mmol) were purchased from ICN Biomedicals (Irvine, CA) and ARC (St. Louis, MO), respectively. UDP-[3H]GalNAc (7.0 Ci/mmol) and UDP-[14C]GlcNAc (200 mCi/mmol) were from PerkinElmer Life Sciences. Chondroitin (a chemically desulfated derivative of whale cartilage chondroitin sulfate A), chondroitin sulfate A (whale cartilage), dermatan sulfate (pig skin), chondroitin sulfate C (shark cartilage), chondroitin sulfate D (shark cartilage), chondroitin sulfate E (squid cartilage), hyaluronan (rooster comb), heparan sulfate (pig aorta), α-N-acetylgalactosaminidase (EC 3.2.1.49 from Acremonium sp.), and chondroitinase ACII (EC 4.2.2.5 from Arthrobacter aurescens) were from Seikagaku Corp. (Tokyo, Japan). Testicular hyaluronidase (EC 3.2.1.35, H6254, type V from sheep testes), β-glucuronidase (EC 3.2.1.31, G0501, type B-10, from bovine liver), heparin (bovine intestine), Galβ1–3GalNAcα-O-benzyl, d-GlcUAβ-O-4-nitrophenyl, anti-FLAG BioM2 antibody, anti-FLAG M2-agarose gel, and pFLAG-CMV1 were from Sigma. A pcDNA3.1 was from Invitrogen. A Superdex™ peptide HR10/30 column, HiLoad 16/60 Superdex 30-pg column, Fast Desalting column HR10/10, and PD10 desalting column were purchased from Amersham Biosciences. N-Acetyl heparosan, GlcNAcα-O-benzyl, GlcNAcβ-O-benzyl, Galα-O-benzyl, Galβ-O-benzyl, GalNAcα-O-benzyl, GalNAcβ-O-benzyl, Galβ1–3Galβ1–4Xylβ1-O-methoxyphenyl, GlcUAβ1–3Galβ1–3Galβ1–4Xylβ1-O-methoxyphenyl, and Galβ1–4GlcNAcβ1–3Galβ1–4GlcNAc were kindly provided by Seikagaku Corp.
      Construction of CSS2 Expression Vector—We performed a BLAST search of the EST data bases using the amino acid sequence of the cloned human CSGlcAT as a query, and we found a novel EST clone (GenBank™ accession number NM_018590) (
      • Gotoh M.
      • Yada T.
      • Sato T.
      • Akashima T.
      • Iwasaki H.
      • Mochizuki H.
      • Inaba N.
      • Togayachi A.
      • Kudo T.
      • Watanabe H.
      • Kimata K.
      • Narimatsu H.
      ). As the sequence was not complete, a GeneScan search was performed on the human genomic data bases. The predicted sequence was confirmed by PCR with two primers, 5′-ACTCCTCTGGCTGCTCTGGGGGTTCG-3′ and 5′-TCTGGTTTTGGGGGAGAAGTGG-3′ (GenBank™ accession number AB086063). The putative catalytic domain of the enzyme (amino acids 97–775) was expressed as a secreted protein fused with a FLAG peptide in COS-7 cells. An ∼2.0-kb DNA fragment was amplified by PCR using the Marathon-Ready™ cDNA derived from human brain (Clontech), as a template, and two primers, 5′-GGAATTCCGGCCAGGCCGCCAAAAAGGC-3′ and 5′-CGGGATCCTCAGGTGCTGTTGCCCTGCTCC-3′. The amplified fragment was inserted between EcoRI and BamHI sites of pFLAG-CMV-1 (Sigma).
      Purification of FLAG-tagged Recombinant Enzyme from Culture Supernatants—COS-7 cells (ATCC CRL-1651) were co-transfected with the expression plasmid and pcDNA3.1 using TransFast™ (Promega, Madison, WI) according to the manufacturer's instructions. The stable transfectants were selected with 600 μg/ml G418 in Dulbecco's modified Eagle's medium containing 10% (v/v) fetal bovine serum (HyClone Laboratories, Logan, UT), 100 μg/ml streptomycin sulfate, and 100 units/ml penicillin G and cloned by limiting dilution. Cloned cell lines were tested for synthesis and secretion of the recombinant protein by immunoprecipitation and Western blotting using an anti-FLAG BioM2 antibody (Sigma). The secreted enzyme was purified by affinity chromatography using anti-FLAG M2-agarose gel (Sigma). The conditioned medium and gel were mixed overnight at 4 °C and centrifuged for 5 min, and the supernatants were aspirated. The gel was washed five times with 10 ml of 20% (v/v) glycerol in 50 mm Tris-HCl, pH 7.4, and resuspended in the same buffer containing protease inhibitors (1 mm phenylmethylsulfonyl fluoride, 1 μg/ml leupeptin, and 1 μg/ml pepstatin A) to produce a 50% slurry (
      • Gotoh M.
      • Yada T.
      • Sato T.
      • Akashima T.
      • Iwasaki H.
      • Mochizuki H.
      • Inaba N.
      • Togayachi A.
      • Kudo T.
      • Watanabe H.
      • Kimata K.
      • Narimatsu H.
      ). The immobilized enzyme was stable at 4 °C for at least 4 weeks. The amount of recombinant protein recovered was estimated by immunoblotting. FLAG-tagged bacterial alkaline phosphatase (Met-FLAG-BAP, molecular mass of 49 kDa) was used as a standard to estimate the relative amount, as described previously (
      • Gotoh M.
      • Yada T.
      • Sato T.
      • Akashima T.
      • Iwasaki H.
      • Mochizuki H.
      • Inaba N.
      • Togayachi A.
      • Kudo T.
      • Watanabe H.
      • Kimata K.
      • Narimatsu H.
      ). The amount of recombinant enzyme protein is expressed in arbitrary units, with each unit of intensity equivalent to 10 ng of FLAG-tagged BAP protein (
      • Gotoh M.
      • Yada T.
      • Sato T.
      • Akashima T.
      • Iwasaki H.
      • Mochizuki H.
      • Inaba N.
      • Togayachi A.
      • Kudo T.
      • Watanabe H.
      • Kimata K.
      • Narimatsu H.
      ).
      Preparation of Acceptor Substrate—Glycosaminoglycan polymers were purchased from Seikagaku Corp. For GlcAT-II assay, chondroitin sulfate A–E, chondroitin, hyaluronan, heparan sulfate, and N-acetyl-heparosan were digested with β-glucuronidase prior to the assay (
      • Gotoh M.
      • Yada T.
      • Sato T.
      • Akashima T.
      • Iwasaki H.
      • Mochizuki H.
      • Inaba N.
      • Togayachi A.
      • Kudo T.
      • Watanabe H.
      • Kimata K.
      • Narimatsu H.
      ). Briefly, 1 mg of each polymer was digested with 100 units of β-glucuronidase in a total volume of 1 ml of 100 mm sodium acetate buffer, pH 5.0, at 37 °C overnight. The digests were then boiled for 20 min; the denatured enzyme was removed by trichloroacetic acid precipitation from the resultant supernatants that were neutralized with sodium hydroxide, and the glycosaminoglycans were recovered by ethanol precipitation. The glycosaminoglycans were redissolved with distilled water of a concentration of 10 mg/ml. Oligosaccharides of chondroitin sulfate, chondroitin, and hyaluronan were prepared as described previously (
      • Gotoh M.
      • Yada T.
      • Sato T.
      • Akashima T.
      • Iwasaki H.
      • Mochizuki H.
      • Inaba N.
      • Togayachi A.
      • Kudo T.
      • Watanabe H.
      • Kimata K.
      • Narimatsu H.
      ). Briefly, for the preparation of the oligosaccharides with even numbers (4-, 6-, 8-, 10-, 12-, and 14-saccharides), each polymer (10 mg) was partially digested with 1,000 turbidity reducing units of testicular hyaluronidase in 1 ml of 0.1 m sodium acetate buffer, pH 5.2, containing 0.15 m NaCl at 37 °C for the appropriate incubation times. For the preparation of the oligosaccharides with odd numbers (5-, 7-, 9-, 11-, and 13-saccharides), the hyaluronidase digests were boiled for 20 min and further digested with 1,000 units of β-glucuronidase (EC 3.2.1.31, from bovine liver, Sigma) at 37 °C for 24 h. After inactivation of the enzyme by boiling for 10 min and centrifuging at 10,000 × g for 20 min at 4 °C, the supernatants of the digests were fractionated on the HiLoad 16/60 Superdex 30-pg column (16 × 600 mm) with 0.2 m NH4HCO3 at a flow rate of 2 ml per min, and absorbance was monitored at 225 nm. The peak fractions were pooled, and desalted on the PD10 desalting column with distilled water as an eluate. The uronic acid contents were then determined by the Bitter-Muir's method using d-glucuronic acid as a standard (
      • Bitter T.
      • Muir H.M.
      ). The desalted solution was lyophilized and redissolved in distilled water at a concentration of 1 mm for each oligosaccharide.
      Glycosyltransferase Assays—The glycosyltransferase activities were investigated with radioactive forms of UDP-GlcUA, UDP-GalNAc, UDP-GlcNAc, UDP-Gal, and various acceptor saccharide substrates, including polymer chondroitin, various chondroitin sulfate isoforms, hyaluronan, heparan sulfate, heparin (100 μg each), oligosaccharides of chondroitin, chondroitin sulfate isoforms, and hyaluronan (1 nmol each). The standard reaction mixture for GalNAcT-II contained 10 μl of the resuspended beads and acceptor substrate, 0.32 nmol of UDP-[3H]GalNAc (6.66 × 105 dpm), 50 mm MES, pH 6.2, and 10 mm MnCl2 in a total volume of 30 μl. The reaction mixture for GlcAT-II contained 10 μl of the resuspended gel and the acceptor substrate, 0.307 nmol of UDP-[14C]GlcUA (2.22 × 105 dpm), 50 mm MES, pH 6.2, and 10 mm MnCl2 in a total volume of 30 μl. The reaction mixtures were incubated at 37 °C for 1 h with mixing, and the reaction was stopped by boiling for 5 min, and then radiolabeled products were separated from free UDP-[3H]GalNAc or UDP-[14C]GlcUA by gel filtration using Superdex™ peptide HR10/30 column (10 × 300 mm) with 0.2 m NaCl as an eluant or HiLoad 16/60 Superdex 30-pg column (16 × 600 mm) with 0.2 m NH4HCO3 as an eluate (
      • Gotoh M.
      • Yada T.
      • Sato T.
      • Akashima T.
      • Iwasaki H.
      • Mochizuki H.
      • Inaba N.
      • Togayachi A.
      • Kudo T.
      • Watanabe H.
      • Kimata K.
      • Narimatsu H.
      ). The labeled products recovered were quantified by liquid scintillation counting. For the acceptor substrates of oligosaccharide with an aromatic residue (methoxyphenyl- or benzyl-) at the reducing terminus, reaction products were diluted with 1 ml of 0.5 m NaCl and applied to a Sep-Pak C18 cartridge (100 mg; Waters, Milford, MA) (
      • Gotoh M.
      • Yada T.
      • Sato T.
      • Akashima T.
      • Iwasaki H.
      • Mochizuki H.
      • Inaba N.
      • Togayachi A.
      • Kudo T.
      • Watanabe H.
      • Kimata K.
      • Narimatsu H.
      ). The cartridge was washed with 3 ml of 0.5 m NaCl and then 3 ml of water; the product was eluted with 50% methanol, and the radioactivity of all fractions was measured by liquid scintillation counting. In repetitions of the experiments when different batches of the enzyme were used, an aliquot was first analyzed by SDS-PAGE and Western blotting using anti-FLAG BioM2 antibody with FLAG-BAP as a standard to obtain a comparable amount of enzyme.
      Identification of the Enzyme Reaction Products—Each product from the GlcAT-II reaction using chondroitin or CS11 and the GalNAcT-II reaction using chondroitin was isolated by gel filtration column chromatography using the Superdex Peptide HR10/30 column. The radioactive peak containing the product was pooled and desalted with the Fast Desalting column HR10/10 using distilled water as an eluant and lyophilized. In order to identify the linkage, the dried sample (about 20 pmol of radiolabeled material) from GlcAT-II reactions was incubated with 100 milliunits of chondroitinase ACII in a total volume of 100 μl of 100 mm Tris-HCl, pH 7.4, containing 30 mm sodium acetate at 37 °C for 1 h or 1 unit of β-glucuronidase in a total volume of 100 μl of 100 mm sodium acetate buffer, pH 5.0, at 37 °C overnight. For confirmation of the linkage structure, we determined whether the product could serve as an acceptor for Escherichia coli strain K4 chondroitin polymerase, which synthesizes chondroitin, and the resultant products could be digested with chondroitinase ACII completely (
      • Ninomiya T.
      • Sugiura N.
      • Tawada A.
      • Sugimoto K.
      • Watanabe H.
      • Kimata K.
      ). Briefly, the 20 pmol of radiolabeled materials was lyophilized and served as substrate for K4 chondroitin polymerase. The reaction was performed at 30 °C overnight in a 50-μl solution containing 50 mm Tris-HCl, pH 7.2, 20 mm MnCl2, 0.1 m (NH4)2SO4,1 m ethylene glycol, 20 pmol of radiolabeled [14C]CS12, 30 nmol each of UDP-GlcUA and UDP-GalNAc, and 0.8 μg of the enzyme preparation. This was followed by boiling for 5 min to stop the reaction. The radioactive peak containing the product was pooled and desalted with the Fast Desalting column HR10/10 using distilled water as an eluant and lyophilized. The dried sample (about 20 pmol of radiolabeled material) from the E. coli strain K4 chondroitin polymerase reaction was incubated with 100 milliunits of chondroitinase ACII in a total volume of 100 μl of 100 mm Tris-HCl, pH 7.4, containing 30 mm sodium acetate at 37 °C for 1 h. The enzyme digests were analyzed again using the same Superdex Peptide HR10/30 column as described above. In order to identify the linkage, the dried sample (about 20 pmol of radiolabeled material) from GalNAcT-II reactions was incubated with 100 milliunits of chondroitinase ACII in a total volume of 100 μl of 100 mm Tris-HCl, pH 7.4, containing 30 mm sodium acetate at 37 °C for 1 h or 100 milliunits of α-N-acetylgalactosaminidase in a total volume of 100 μlof50mm sodium citrate buffer, pH 4.5, at 37 °C overnight. The enzyme digests were analyzed again using the same Superdex Peptide HR10/30 column as described above.
      Quantitative Analysis of the CSS2 Transcript in Human Tissues by Real Time PCR—For quantification of CSS2 transcripts, we employed the real time PCR method, as described in detail previously (
      • Iwai T.
      • Inaba N.
      • Naundorf A.
      • Zhang Y.
      • Gotoh M.
      • Iwasaki H.
      • Kudo T.
      • Togayachi A.
      • Ishizuka Y.
      • Nakanishi H.
      • Narimatsu H.
      ). Marathon Ready cDNA derived from various human tissues was purchased from Clontech. Standard curves for the endogenous control, glyceraldehyde-3-phosphate dehydrogenase (GAPDH) cDNA, were generated by serial dilution of pCR2.1 (Invitrogen) DNA containing the GAPDH gene. The primer set and the probe for CSS2 were as follows: the forward primer, 5′-GCTGAACTGGAACGCACGTA-3′, and the reverse primer, 5′-CGGGATGGTGCTGGAATAC-3′, and the probe, 5′-AGATCCAGGAGTTACAGTGG-3′, with a minor groove binder. The primer sets and the probes for CSGlcAT and CSS1 were as follows: the forward primer for CSGlcAT, 5′-GTGAAATAGAACAACTGCAGGCTC-3′, and the reverse primer for CSGlcAT, 5′-GAGAAGGTGTGCTGCTCTGTGA-3′, the probe for CSGlcAT, 5′-CGGAACCTGACCGTGC-3′, with a minor groove binder; the forward primer for CSS1, 5′-AGTGTGTCTGGTCTTATGAGATGCA-3′, and the reverse primer for CSS1, 5′-AGCTGTGGAGCCTGTACTGGTAG-3′ and the probe for CSS1, 5′-ATGAGAATTACGAGCAGAAC-3′ with a minor groove binder. PCR products were continuously measured with an ABI PRISM 7700 Sequence Detection System (Applied Biosystems, Foster City, CA). The relative amounts of their transcripts were normalized to the amount of GAPDH transcript in the same cDNA.
      Comparison of the C-terminal Domain Structure between CSS2 and CSGlcAT by Homology Modeling—The molecular structures of the C-terminal halves of CSS2 (C-CSS2; amino acid residues 474–775) and CSGlcAT (C-CSGlcAT; amino acid residues 455–772) were estimated from the model of the C-terminal half of CSS1 (C-CSS1; amino acid residues 500–802) because the amino acid sequences of C-CSS2 and C-CSGlcAT are far diverged from other glycosyltransferases for which the three-dimensional structures are known. From the consensus result of three threading methods, 3D-PSSM (www.sbg.bio.ic.uk/~3dpssm/), FUGUE (www-cryst.bioc.cam.ac.uk/~fugue/prfsearch.html), and GenTHREADER (bioinf.cs.ucl.ac.uk/psiform.html), we chose bovine β4-galactosyltransferase (β4GalT-1) as a template for homology modeling of C-CSS1, because β4GalT-1 obtained the best score by 3D-PSSM and FUGUE and the second highest score by GenTHREADER. Furthermore, CSS1 has been classified into the same glycosyltransferase family (Family GT7) with β4GalT-1 by CAZy (afmb.cnrs-mrs.fr/~cazy/CAZY/index.html), in which CSGalNAcT-1 and CSGalNAcT-2 are also the same family members. With the pairwise alignment of C-CSS1 and β4GalT-1 derived from 3D-PSSM, homology modeling of C-CSS1 was performed using FAMS (
      • Ogata K.
      • Umeyama H.
      ). Then the amino acid sequences of C-CSS1, C-CSS2, C-CSGlcAT, CSGalNAcT-1 (amino acid residues 228–532), and CSGalNAcT-2 (amino acid residues 237–542) were aligned using ClustalW (
      • Thompson J.D.
      • Higgins D.G.
      • Gibson T.J.
      ) and joined to the pairwise alignment of C-CSS1 and β4GalT-1. The secondary structures of bovine β4GalT-1 corresponded well with those predicted by PSIPRED2 (
      • Jones D.T.
      ) for C-CSS2 and C-CSGlcAT as well as CSGalNAcT-1 and CSGalNAcT-2. By using the pairwise alignment of C-CSS1 and C-CSS2 and that of C-CSS1 and C-CSGlcAT extracted from the multiple alignment, the three-dimensional structures of C-CSS2 and C-CSGlcAT, respectively, were modeled using FAMS.

      RESULTS

      Molecular Cloning of CSS2 and Determination of Its Nucleotide and Amino Acid Sequences—We performed a BLAST search of the EST data bases using the amino acid sequence of the cloned human CSGlcAT as a query, and we found a novel EST (GenBank™ accession number NM_018590) (
      • Gotoh M.
      • Yada T.
      • Sato T.
      • Akashima T.
      • Iwasaki H.
      • Mochizuki H.
      • Inaba N.
      • Togayachi A.
      • Kudo T.
      • Watanabe H.
      • Kimata K.
      • Narimatsu H.
      ). As the sequence was incomplete, a GeneScan search was performed of human genomic data bases. The predicted sequence was confirmed by PCR with two primers, 5′-ACTCCTCTGGCTGCTCTGGGGGTTCG-3′ and 5′-TCTGGTTTTGGGGGAGAAGTGG-3′ (GenBank™ accession number AB086063). The putative amino acid sequences are shown in Fig. 1A. They contained an open reading frame of 2328 bp, 775 amino acids, encoding a typical type II membrane protein with three possible N-glycosylation sites.
      Figure thumbnail gr1
      Fig. 1Multiple alignment and genomic structure of CSS2 and CSGlcAT. A, the multiple alignment of two enzymes was performed using GENETYX. Introduced gaps are shown with hyphens. The putative transmembrane domains are underlined. DXD motifs are in boldface. The β3-glycosyltransferase motif is boxed. Identical amino acids are shown by asterisks. The possible N-glycosylation sites are indicated by arrowheads. B, the genome structure of the CSS2 (top) and CSGlcAT (bottom) genes were determined by comparing their genomic DNA (GenBank™ accession numbers NT_005403.10 and NT_019447, respectively) and cDNAs. Exon regions are denoted by boxes. The translation initiation (ATG) and termination (TGA or TAG) codons are also shown. Black horizontal bars denote the introns.
      Comparison of Amino Acid Sequences between CSS2 and CSGlcAT—The amino acid sequence of the clone exhibited high homology (57%) with CSGlcAT in the putative domain as shown in Fig. 1A. Hydropathy plots of the amino acid sequence revealed one hydrophobic stretch, located at position 13-32, like in CSGlcAT (Fig. 1A, underlined). A DXD motif, which is conserved in many glycosyltransferases and functions as a key sequence for divalent cation binding, and another motif conserved in β1,3-glycosyltransferases (β3GTs) were found in the N-terminal half (Fig. 1, boldface and boxed, respectively). Comparison of the location of cysteine residues in the predicted protein encoded by CSS2 (11 cysteines) and that of CSGlcAT (13 cysteines) showed the good conservation of 10 cysteines not only in the N-terminal half but also in the C-terminal half. Three potential N-glycosylation sites also appeared to be conserved in both enzymes. A remarkable difference between the two amino acid sequences was two short insertion/deletions near both the N and C termini. The N-terminal insertion/deletion was likely located in the putative stem region of CSS2, and the second one was located in the Pro-rich region of CSGlcAT near the C terminus.
      Genome Organization and Chromosome Localization—A comparison of the cDNA with the genomic sequence on chromosome 2 revealed the CSS2 gene to consist of at least four exons (Fig. 1B, top). Its genomic organization including the exon-intron boundaries was quite similar to that of the CSGlcAT gene (Fig. 1B, bottom), which consists of four discrete exons in the coding region (
      • Gotoh M.
      • Yada T.
      • Sato T.
      • Akashima T.
      • Iwasaki H.
      • Mochizuki H.
      • Inaba N.
      • Togayachi A.
      • Kudo T.
      • Watanabe H.
      • Kimata K.
      • Narimatsu H.
      ). The CSS2 and CSGlcAT genes were located on human chromosome 2q36.1. and 7q36, respectively.
      Estimation of the Amount of FLAG Epitope-tagged CSS2 Protein—To facilitate the functional analysis of the putative glycosyltransferase, a soluble form of the protein was generated by replacing the first 96 amino acids of the putative glycosyltransferase with the preprotrypsin signal sequence and a FLAG tag, as described under “Experimental Procedures.” The soluble putative glycosyltransferase was expressed in COS-7 cells as a recombinant enzyme fused with the FLAG tag. The fused enzyme expressed in the medium was adsorbed onto anti-FLAG M2 antibody-conjugated agarose gel to eliminate endogenous glycosyltransferases, and the enzyme-bound gels were used for the various reactions. The amount of FLAG epitope-tagged CSS2 was estimated using FLAG-BAP as a standard as described previously (
      • Gotoh M.
      • Yada T.
      • Sato T.
      • Akashima T.
      • Iwasaki H.
      • Mochizuki H.
      • Inaba N.
      • Togayachi A.
      • Kudo T.
      • Watanabe H.
      • Kimata K.
      • Narimatsu H.
      ). The amount of FLAG-BAP and the densitometric unit obtained by measurement of each band intensity showed a correlation (R 2 = 0.995) and exhibited a standard curve, as shown in Fig. 2B. The amount of recombinant CSS2 protein was determined by plotting 10 ng of the FLAG-BAP as 1 unit that was obtained from 5.3 ml of the pooled medium (Fig. 2C). CSGlcAT was also isolated, and its protein contents were determined as 1 unit/5.1 ml of the pooled medium (data not shown).
      Figure thumbnail gr2
      Fig. 2Estimation of the amount of FLAG epitope-tagged CSS2 protein. A, Western blot analyses of FLAG-tagged BAP protein and FLAG-tagged CSS2 protein isolated from serial dilutions of culture medium of COS-7 transfectants stably expressing FLAG-tagged CSS2 protein. The intensity of the 49-kDa (BAP protein) and 90-kDa (CSS2 protein) bands increased with increasing concentrations of FLAG-tagged BAP protein and the volume of the medium, respectively. B, depiction of the relationship between the content of BAP protein and the band density; a linear correlation was noted (R 2 = 0.995). C, depiction of the relationship between the volume of the medium and the concentration of CSS2 protein, as derived from the BAP standard curve; a linear correlation was also observed (R 2 = 0.992). The amount of recombinant soluble CSS2 protein is expressed in arbitrary units of intensity each equivalent to 10 ng of FLAG-tagged BAP protein.
      Acceptor Substrate Specificity of CSS2—The acceptor specificity of the truncated CSS2 recovered from COS-7 transfectant cells was determined with a variety of glycosaminoglycans and their oligosaccharides as acceptor substrates. In preliminary experiments, this glycosyltransferase showed dual enzymatic activities, glucuronyltransferase (GlcAT-II) activity for CS-C11 and N-acetylgalactosaminyltransferase (GalNAcT-II) activity for CS-C10 (data not shown). Therefore, we examined the effect of buffers and pH on both activities of this recombinant glycosyltransferase. As shown in Fig. 3, the glycosyltransferase exhibited optimum activity at pH 6.2 and pH 6.5, depending upon the buffers used, with the highest level in a 50 mm imidazole-HCl buffer at pH 6.5 for GlcAT-II activity and at pH 6.2 in a 50 mm MES buffer for GalNAcT-II. CSGlcAT exhibited optimum activity at pH 6.2 in a 50 mm MES buffer for GlcAT-II, but no GalNAcT-II activity was observed under any pH conditions examined (data not shown). Thus, all enzymatic reactions were carried out in 50 mm MES buffer, pH 6.2.
      Figure thumbnail gr3
      Fig. 3Effects of buffers and pH on the GlcAT-II (A) and GalNAcT-II (B) activities of CSS2. The effects of pH on the GlcUA and GalNAc transfers to CS-C11 and CS-C10, respectively, were determined under standard assay conditions with different buffers at a final concentration of 50 mm. The buffers are sodium acetate (open circles), MES-NaOH (closed circles), imidazole-HCl (open triangles), and Tris-HCl (closed triangles). Data represent the average of two independent experiments.
      Divalent cations were essential for the two enzymatic reactions, and 10 mm EDTA completely abolished both activities (Fig. 4A). Mn2+ evoked the highest level of activity under standard assay conditions, and Co2+ was 17 and 31% as effective as Mn2+ for GlcAT-II and GalNAcT-II activity, respectively (Fig. 4A). The optimal concentration of Mn2+ was 10 mm for both activities (Fig. 4B). In contrast, Mn2+ only induced activity for GlcAT-II at the optimal concentration, 15 mm, and no activity was observed with any other divalent cations examined (data not shown).
      Figure thumbnail gr4
      Fig. 4Effects of divalent cations (A) and Mn2+ concentration (B) on the GlcAT-II and GalNAcT-II activities. A, the effects of divalent cations on the GlcUA transfer to CS-C11 (black boxes) and GalNAc transfer to CS-C10 (open boxes) were determined under standard assay conditions with divalent cations or EDTA at a final concentration of 10 mm. B, the effects of Mn2+ concentrations on the GlcUA transfer to CS-C11 (closed circles) and GalNAc transfer to CS-C10 (open circles) were determined under standard assay conditions, except that the concentration of MnCl2 was varied. Data represent the average of two independent experiments.
      The specificity of the recombinant CSS2 toward the UDP-sugar donor substrate was analyzed using a series of analogs of radiolabeled UDP-GlcUA, UDP-Gal, UDP-GalNAc, and UDP-GlcNAc under optimized conditions. CSS2 was able to efficiently catalyze the transfer of GlcUA from UDP-GlcUA to the acceptor CS-C11 and of GalNAc from UDP-GalNAc to the acceptor CS-C10. In contrast, the other radiolabeled nucleotide sugars tested (UDP-Gal, UDP-GalNAc, and UDP-GlcNAc for CS-C11; UDP-GlcUA, UDP-Gal, and UDP-GlcNAc for CS-C10) were not substrates of the recombinant CSS2 (data not shown). Furthermore, various monosaccharides including GlcUAβ-O-4-nitrophenyl, GlcNAcα-O-benzyl, GlcNAcβ-O-benzyl, Galα-O-benzyl, Galβ-O-benzyl, GalNAcα-O-benzyl, and GalNAcβ-O-benzyl were not efficient acceptor substrates for any of the UDP-sugars tested as donors (data not shown).
      To characterize the substrate specificity of the purified recombinant CSS2, chondroitin, chondroitin sulfate isoforms, and other glycosaminoglycans were tested as acceptor substrates. As shown in Table I, chondroitin was the best substrate, and the other polymer chondroitin sulfate isoforms were poor acceptors for both enzyme activities of CSS2, as well as the GlcAT-II activity of CSGlcAT.
      Table IAcceptor specificity of the truncated CSS2 and CSGlcAT
      Acceptor substrateSpecific activity of the radiolabeled sugar incorporated
      The values represent the averages of two independent experiments.
      CSS2CSGlcAT
      GlcAT-IIGalNAcT-IIGlcAT-IIGalNAcT-II
      pmol/h/unit
      Glycosaminoglycan polymer
      Polysaccharide substrates for GlcAT-II were used after β-glucuronidase treatment.
      Chondroitin7.16.512.9<0.1
      Chondroitin sulfate A0.30.40.7<0.1
      Chondroitin sulfate B<0.1<0.1<0.1<0.1
      Chondroitin sulfate C<0.10.40.2<0.1
      Chondroitin sulfate D<0.10.1<0.1<0.1
      Chondroitin sulfate E<0.11.2<0.1<0.1
      Hyaluronan<0.1<0.1<0.1<0.1
      Heparan sulfate<0.1<0.1<0.1<0.1
      N-acetylheparosan<0.1<0.1<0.1<0.1
      Heparin<0.1<0.1<0.1<0.1
      Glycosaminoglycan oligosaccharide
      Undecasaccharide for GlcAT-II
      CH112.6<0.15.1<0.1
      CS-A117.1<0.16.6<0.1
      CS-C1111.8<0.110.7<0.1
      HA11<0.1<0.1<0.1<0.1
      Decasaccharide for GalNAcT-II
      CH10<0.12.1<0.1<0.1
      CS-A10<0.16.5<0.1<0.1
      CS-C10<0.16.5<0.1<0.1
      HA10<0.1<0.1<0.1<0.1
      Linkage region oligosaccharides
      Galβ1-3Galβ1-4Xylβ1-OMP<0.1NT
      NT, not tested.
      <0.1NT
      GlcUAβ1-3Galβ1-3Galβ1-4Xylβ1-OMPNT<0.1NT<0.1
      a The values represent the averages of two independent experiments.
      b Polysaccharide substrates for GlcAT-II were used after β-glucuronidase treatment.
      c NT, not tested.
      As shown in Table I, CSS2 apparently showed GlcAT-II activity toward undecasaccharides having GalNAc in their non-reducing termini, which were prepared from the CS isoforms and chondroitin. The activities for the CS-A and CS-C undecasaccharides were 2.7- and 4.5-fold higher than the activity for the chondroitin undecasaccharide, respectively, and the activity for the hyaluronan undecasaccharide was negative. On the other hand, CSS2 apparently showed GalNAcT-II activity toward decasaccharides having GlcUA at their non-reducing termini, which were prepared from CS isoforms and chondroitin. The activities for both CS-A and CS-C decasaccharides were 3.1-fold higher than that for the chondroitin decasaccharide, and the activity was negative for hyaluronan decasaccharide. To examine whether CSS2 has other glycosyltransferase activities, several substrates, Galβ1–3Galβ1–4Xylβ1-O-methoxyphenyl and GlcUAβ1–3Galβ1–3Galβ1–4Xylβ1-O-methoxyphenyl (GlcAT-I and GalNAcT-I for glycosaminoglycan linkage region, respectively) (Table I), Galβ1–3GalNAcα-O-benzyl (human natural killer cell-1 epitope synthase) and Galβ1–4GlcNAcβ1–3Galβ1–4GlcNAc (lactosamine tetrasaccharide) (data not shown), were tested as acceptors, but no activity was detected. These results again suggested that CSS2 is responsible for the chondroitin sulfate elongation and not for the linkage tetrasaccharide or other substrates.
      Effects of the acceptor length of oligosaccharides on CSS2 activities were determined using CS-C and chondroitin oligosaccharides (Fig. 5). The odd-numbered oligosaccharides from CS-C and chondroitin having a GalNAc residue in the non-reducing terminus were subjected to an assay for GlcAT-II activity. The even-numbered oligosaccharides having a GlcUA residue in the non-reducing terminus were subjected to a GalNAcT-II assay. CSS2 exhibited GlcAT-II activity toward the odd-numbered oligosaccharides from CS-C and chondroitin and GalNAcT-II activity toward the even-numbered oligosaccharides. For the GlcAT-II activity, the longer oligosaccharides served as better acceptors than the shorter ones. On the other hand, for the GalNAcT-II activity, CSS2 transferred GalNAc most efficiently to CS-C10 among the CS-C or chondroitin oligosaccharides examined (Fig. 5). For both the GlcAT-II and the GalNAcT-II activity, the CS-C oligosaccharides were better acceptor substrates than the chondroitin oligosaccharides, especially for the GalNAcT-II activity, the former acceptor activities being 2.8- and 4.8-fold higher than the latter for deca- and octasaccharides, respectively (Fig. 5).
      Figure thumbnail gr5
      Fig. 5Influence of chain length of acceptor substrate on GlcAT-II (A) and GalNAcT-II activities (B) of CSS2. Odd-numbered CS-C (A, open circles) or chondroitin (A, closed circles) oligosaccharides with terminal non-reducing N-acetylgalactosaminyl residues and even numbered CS-C (B, open circles) or chondroitin (B, closed circles) were used as acceptor substrates at a final concentration of 33 mm. Data represent the average of two independent experiments.
      Analysis of CSS2 Reaction Products—To identify the GlcAT-II reaction products, chondroitin polymer was labeled with [14C]UDP-GlcUA by CSS2 under optimized conditions, and the products were isolated and then subjected to a gel filtration analysis after chondroitinase AC-II or β-glucuronidase treatment. As shown in Fig. 6A, the labeled products were completely digested by chondroitinase AC-II and β-glucuronidase, quantitatively yielding a 14C-labeled peak at the position of [14C]GlcUAβ1–3GalNAc and of free [14C]GlcUA, respectively. These findings indicate that a GlcUA residue was transferred to the non-reducing terminal GalNAc residue of chondroitin polymer through a β-linkage. Furthermore, 14C-labeled chondroitin sulfate dodecasaccharide was used as an acceptor for a chondroitin polymerase from an E. coli K4 strain (
      • Ninomiya T.
      • Sugiura N.
      • Tawada A.
      • Sugimoto K.
      • Watanabe H.
      • Kimata K.
      ). This reaction was performed in the presence of the enzyme and two donor substrates, UDP-GalNAc and UDP-GlcUA. The reaction products showed a molecular mass of ∼3000 Da that was speculated to be the product transferred by ∼5 sugar residues. They were digested by chondroitinase AC-II, yielding unsaturated GlcUAβ1–3GalNAc disaccharides. These results indicated the [14C]GlcUA transferred by CSS2 at the non-reducing terminal of CS-C11 could serve as a substrate for E. coli K4 chondroitin polymerase, and the resultant internal [14C]GlcUA residue linked by a β-linkage to GalNAc was susceptible to chondroitinase AC-II digestion. These findings strongly suggested that a GlcUA residue was transferred to the non-reducing terminal GalNAc residue of chondroitin sulfate undecasaccharides through a β1–3-linkage.
      Figure thumbnail gr6
      Fig. 6Identification of the putative human glycosyltransferase reaction products. A, the GlcAT-II reaction products with polymer chondroitin recovered from a Superdex Peptide column were digested with chondroitinase AC-II or β-glucuronidase as described under “Experimental Procedures.” The non-digested sample (open circles), the chondroitinase AC-II digest (closed circles), or the β-glucuronidase digest (closed triangles) was applied to a column of Superdex Peptide, and the respective fractions (0.5 ml each) were analyzed for radioactivity. Arrows indicate the elution positions of the authentic saturated disaccharide (closed arrowhead, GlcUAβ1–3GalNAc) or free GlcUA (open arrowhead). B, the GlcAT-II reaction products with CS11 recovered from a Superdex Peptide column were subjected to chondroitin polymerization with E. coli strain K4 chondroitin polymerase or chondroitinase AC-II digestion of the resultant polymer as described under “Experimental Procedures.” The [14C]GlcUA-labeled CS-C11 oligosaccharide sample (open circles), the sample polymerized by E. coli strain K4 chondroitin polymerase (closed circles), or the chondroitinase AC-II digest of the E. coli strain K4 chondroitin polymerase products (closed triangles) was applied to a column of Superdex Peptide, and the respective effluent fractions (0.5 ml each) were analyzed for radioactivity. Arrows indicate the elution positions of the authentic saturated disaccharide (closed arrowhead, GlcUAβ1–3GalNAc). C, the GalNAcT-II reaction products with polymer chondroitin recovered from a Superdex Peptide column were subjected to digestion with chondroitinase AC-II or α-N-acetylgalactosaminidase as described under “Experimental Procedures.” The undigested sample (open circles), the chondroitinase AC-II digest (filled circles), or the α-N-acetylgalactosaminidase digest (filled triangles) was applied to a column of Superdex Peptide, and the respective fractions (0.5 ml each) were analyzed for radioactivity. Arrows indicate the elution positions of the authentic free GalNAc (open arrowhead).
      To identify the GalNAcT-II reaction products, chondroitin polymer was labeled with [3H]GalNAc by CSS2, and the products were isolated and then subjected to a gel filtration analysis after chondroitinase AC-II treatment. As shown in Fig. 6C, the labeled products were completely digested by chondroitinase AC-II, quantitatively yielding a 3H-labeled peak at the position of free [3H]GalNAc, which was separable from GlcUAβ1–3GalNAc. In addition, they were inert to the action of α-N-acetylgalactosaminidase. These findings clearly indicated that a GalNAc residue had been transferred exclusively to the non-reducing terminal GlcUA residue of polymer chondroitin through a β1–4-linkage. Together with the above results they suggested that the identified protein is chondroitin synthase with both CSGlcAT-II and CSGalNAcT-II activity.
      Kinetic Analysis of CSS2—We investigated the effect of the concentrations of the donor substrates, UDP-GlcUA and UDP-GalNAc, and the acceptor substrates, CS-C11 and CS-C10, on the activities of CSS2. As shown in Fig. 7, the apparent K m values for UDP-GlcUA and CS-C11 for GlcAT activity were 263 μm (R 2 = 0.996) and 27 μm (R 2 = 0.992), respectively. The apparent K m values for UDP-GalNAc and CS-C10 for GalNAc-T activity were 670 μm (R 2 = 0.999) and 22 μm (R 2 = 0.997), respectively.
      Figure thumbnail gr7
      Fig. 7Effects of the concentration of UDP-sugars (A) and CS oligosaccharides (B) on the activity of CSS2. A, the kinetic behaviors of CSS2 regarding UDP-GlcUA to CS-C11 (closed circles) and UDP-GalNAc to CS-C10 (open circles). B, the kinetic behaviors of CSS2 regarding CS-C11 to UDP-GlcUA (closed circles) and CS-C10 to UDP-GalNAc (open circles). UDP-sugars and CS-C oligosaccharides were added to the established assay mixture at different final concentrations as described under “Experimental Procedures.” Data represent the average of two independent experiments.
      Quantitative Analysis of the CSGlcAT Transcript in Human Tissues by Real Time PCR—The tissue distribution and the expression levels of the CSS2 transcripts were also investigated in comparison with those of the CSS1 and CSGlcAT transcripts, by the real time PCR method. Expression levels of CSS2, CSGlcAT, and CSS1 in various tissues are shown as the relative amount versus the GAPDH transcripts in Fig. 8. All the enzymes were expressed ubiquitously, but with some difference, in all tissues examined. Notably, the expression of CSS2 was particularly abundant in pancreas, ovary, placenta, small intestine, and stomach. These ubiquitous expression patterns of the enzyme are consistent with the broad distribution of CS proteoglycans throughout the body.
      Figure thumbnail gr8
      Fig. 8Quantitative analysis of CSGlcAT and CSSs transcripts in human tissues by real time PCR. Standard curves for CSS2, CSS1, CSGlcAT, and GAPDH were generated by serial dilution of each plasmid DNA. The expression levels of the CSS2 (closed bars), CSS1 (open bars), and CSGlcAT (slashed bars) transcripts were normalized to those of the GAPDH transcripts which were measured using the same cDNAs. Data were obtained from triplicate experiments and are indicated as the mean ± S.D.
      Multiple Alignments and Molecular Modeling of C-terminal Halves of Bovine β4-Galactosyltransferase and Human Chondroitin Glycosyltransferases—Although CSS2 and CSGlcAT are highly homologous, CSS2 exhibits GalNAcT-II activity, whereas CSGlcAT does not. To address the regions that contribute to the GalNAcT-II activity, we elaborated a multiple alignment of the CS glycosyltransferases. Fig. 9 shows the alignment of bovine β4-galactosyltransferase (β4Gal-T1; Protein Data Bank code 1FR8A) and C-CSS1 by 3D-PSSM, a threading method, combined with multiple alignments of C-CSS1, CSGalNAcT-1, CSGalNAcT-2, C-CSS2, and C-CSGlcAT. The comparison of elements of secondary structure of β4Gal-T1 by crystallography and C-CSGlcAT by PSIPRED2 prediction aligned with other CS glycosyltransferases indicated that these molecules were basically constructed within the similar molecular architecture (Fig. 9). Interestingly, by the structure-based multiple alignment, the second aspartic acid residue of the DXD motif, which is conserved in β4Gal-T1, C-CSS1, CSGalNAcT-1, and CSGalNAcT-2, was shown to be conserved at Asp617 in CSS2 but not in CSGlcAT (Fig. 9, 1st small half-box). Furthermore, this alignment showed that two regions in C-CSS2 and C-CSGlcAT were unable to be constructed in their three-dimensional structure models because they were extra regions compared with β4Gal-T1 (Fig. 9, black and red boxes). One is the 33 and 32 amino acid residues in N-terminal region of C-CSS2 and C-CSGlcAT, respectively (Fig. 9, black box). Another is 17 and 33 amino acid residues between the regions corresponding to the DXD motif and β4Gal-T1 motif (GWGGEDDD) in C-CSS2 and C-CSGlcAT, respectively (Fig. 9, red box).
      Figure thumbnail gr9
      Fig. 9Multiple alignment of β4GalT-1 and C-CSS1, CSGalNAcT-1, CSGalNAcT-2, C-CSS2, and C-CSGlcAT. The alignment of C-CSS1 with bovine β4-galactosyltransferase (Protein Data Bank code 1FR8A) by 3D-PSSM and the alignment of C-CSS1 with CSGalNAc-T1, CSGalNAcT2, C-CSS2, and C-CSGlcAT using ClustalW were aligned together along the C-CSS1 sequence. Amino acids are divided into five groups as follows: D, E, N, Q as group 1; H, K, R as group 2; F, W, Y as group 3; C, I, L, M, V as group 4; and A, G, P, S, T as group 5. The sites where at least two-thirds amino acids are within a group are colored red, blue, orange, yellow, and green for the groups 1–5, respectively. Elements of the secondary structure are depicted above the sequences for β4Gal-T1 based on crystallography and below the sequences for C-CSGlcAT based on the PSIPRED2 prediction; e, β-strand; h, α-helix. Black half-boxes above the sequences show the putative DXD motif (1st small half-box) and β4 glycosyltransferase motif (GWGGEDDD, 2nd large half-box). Black and red boxes in C-CCS2 and C-CSGlcAT sequences indicate the regions that were unable to be constructed in their three-dimensional structural models because these were extra regions compared with β4Gal-T1.
      We then performed the molecular modeling of the C-terminal half of these two enzymes. A comparison of the three-dimensional structural models of C-CSS2 and C-CSGlcAT revealed that the latter region, 17 and 33 amino acid residues in C-CSS2 and C-CSGlcAT, respectively (Fig. 9, red box), were located near the predicted catalytic sites of each enzyme (Fig. 10, dotted line in the left of the models). The three-dimensional structure of these regions could not been constructed, because these regions were lacked in β4Gal-T1 as a template for three-dimensional modeling. Therefore, if they may form loop structure and have a similar orientation in each molecule, the long loop (33 amino acid residues) in CSGlcAT may influence the catalytic reaction of this enzyme and the shorter (17 amino acid residues) loop in C-CSS2 may not. Therefore, we concluded that this result is the one of possible reasons in the difference of GalNAcT-II activity between CSS2 and CSGlcAT.
      Figure thumbnail gr10
      Fig. 10The three-dimensional structure of C-CSS2 (A) and C-CSGlcAT (C) obtained by homology modeling with β4Gal-T1 (B) as template. UDP was shown by a space-filling model in β4Gal-T1. Two regions that were unable to be constructed in their three-dimensional structure in each enzyme, C-CSS2 (loops of 17 residues and 33 residues) and C-CSGlcAT (loops of 33 residues and 32 residues), are shown by white dotted lines. The 1st and 2nd loops indicated by red and black boxes in are shown, respectively. The 33 residue-loop in C-CSGlcAT lies adjacent to the active site. All molecular images were prepared using MOE (CCG).

      DISCUSSION

      In this report, we have described the cloning and characterization of a novel human chondroitin sulfate synthase (CSS2) based on homology with human CSGlcAT (
      • Gotoh M.
      • Yada T.
      • Sato T.
      • Akashima T.
      • Iwasaki H.
      • Mochizuki H.
      • Inaba N.
      • Togayachi A.
      • Kudo T.
      • Watanabe H.
      • Kimata K.
      • Narimatsu H.
      ), which exhibits both glucuronyltransferase II (GlcAT-II) and N-acetylgalactosaminyltransferase II (GalNAcT-II) activity for CS chain elongation in vitro. This enzyme is the fifth CS glycosyltransferase. The presence of at least five distinct enzymes involved in CS initiation and elongation suggests that CS synthesis takes place via a similar mechanism to heparan sulfate synthesis where five homologous EXT family members with distinct, but overlapping, acceptor specificities are involved (
      • Zak B.M.
      • Crawford B.E.
      • Esko J.D.
      ).
      In this study we first expressed and purified CSS2 and CSGlcAT as soluble enzymes, and we investigated the properties of purified CSS2 compared with CSGlcAT. Surprisingly, CSS2 exhibited not only GlcAT-II activity but also GalNAcT-II activity which CSGlcAT lacked, although these two enzymes exhibited high homology. Characterization of CSS2 revealed an Mn2+ requirement (10 mm) for maximal activity and a pH optimum of 6.5 and 6.2 for GlcAT-II activity and GalNAcT-II activity, respectively. On the other hand, CSGlcAT revealed an Mn2+ requirement (15 mm) for maximal activity and pH optimum of 6.2 for GlcAT-II activity (data not shown). Under optimal conditions, the substrate specificity of CSS2 was shown to be similar to that of CSS1 and CSGlcAT for GlcAT-II activity and of CSS1 for GalNAcT-II activity. In summary, all of the three enzymes prefer the nonsulfated isomer as a polymer substrate and sulfated isomer as oligosaccharide substrates (
      • Kitagawa H.
      • Uyama T.
      • Sugahara K.
      ,
      • Gotoh M.
      • Yada T.
      • Sato T.
      • Akashima T.
      • Iwasaki H.
      • Mochizuki H.
      • Inaba N.
      • Togayachi A.
      • Kudo T.
      • Watanabe H.
      • Kimata K.
      • Narimatsu H.
      ), and the same specificities were also observed in CSGalNAcT-1 (
      • Gotoh M.
      • Sato T.
      • Akashima T.
      • Iwasaki H.
      • Kameyama A.
      • Mochizuki H.
      • Yada T.
      • Inaba N.
      • Zhang Y.
      • Kikuchi N.
      • Kwon Y.D.
      • Togayachi A.
      • Kudo T.
      • Nishihara S.
      • Watanabe H.
      • Kimata K.
      • Narimatsu H.
      ), CSGalNAcT-2 (
      • Sato T.
      • Gotoh M.
      • Kiyohara K.
      • Akashima T.
      • Iwasaki H.
      • Kameyama A.
      • Mochizuki H.
      • Yada T.
      • Inaba N.
      • Togayachi A.
      • Kudo T.
      • Asada M.
      • Watanabe H.
      • Imamura T.
      • Kimata K.
      • Narimatsu H.
      ), bovine serum β-glucuronyltransferase, and bovine serum N-acetylgalactosaminyltransferase (
      • Kitagawa H.
      • Tsuchida K.
      • Ujikawa M.
      • Sugahara K.
      ,
      • Kitagawa H.
      • Ujikawa M.
      • Tsutsumi K.
      • Tamura J.
      • Neumann K.W.
      • Ogawa T.
      • Sugahara K.
      ). Sulfation of CS ordinarily proceeds together with polymerization at the Golgi apparatus, and specific sulfate groups have either stimulatory or inhibitory effects on GalNAc and GlcUA transfer (
      • Kitagawa H.
      • Tsuchida K.
      • Ujikawa M.
      • Sugahara K.
      ,
      • Kitagawa H.
      • Ujikawa M.
      • Tsutsumi K.
      • Tamura J.
      • Neumann K.W.
      • Ogawa T.
      • Sugahara K.
      ,
      • Gundlach M.W.
      • Conrad H.E.
      ). Furthermore, earlier studies have suggested that 4,6-O-sulfation of GalNAc (
      • Inoue H.
      • Otsu K.
      • Suzuki S.
      • Nakanishi Y.
      ,
      • Otsu K.
      • Inoue H.
      • Tsuzuki Y.
      • Yonekura H.
      • Nakanishi Y.
      • Suzuki S.
      ) or 3-O-sulfation of GlcUA (
      • Kitagawa H.
      • Tsutsumi K.
      • Ujikawa M.
      • Goto F.
      • Tamura J.
      • Neumann K.W.
      • Ogawa T.
      • Sugahara K.
      ) residues at the non-reducing end might be involved in CS chain termination. Therefore, the specificities observed in the cloned CS glycosyltransferases in vitro suggested that the sulfate transfer may coincide with the CS glycosyltransferase reaction to the growing chondroitin sulfate chain and plays important roles in chain elongation and termination.
      The K m values of CSS2 for UDP-GlcUA and UDP-GalNAc were relatively high (263 and 670 μm, respectively) compared with the values of CSGlcAT (82 μm for UDP-GlcUA), bovine serum β-glucuronyltransferase (50 μm for UDP-GlcUA), and bovine serum N-acetylgalactosaminyltransferase (50 μm for UDP-GalNAc) (
      • Kitagawa H.
      • Tsuchida K.
      • Ujikawa M.
      • Sugahara K.
      ,
      • Kitagawa H.
      • Ujikawa M.
      • Tsutsumi K.
      • Tamura J.
      • Neumann K.W.
      • Ogawa T.
      • Sugahara K.
      ), although in the studies of serum enzymes chondroitin polymer was used as an acceptor instead of chondroitin sulfate oligosaccharides. Sasai et al. (
      • Sasai K.
      • Ikeda Y.
      • Fujii T.
      • Tsuda T.
      • Taniguchi N.
      ) showed that human β1,6-N-acetylglucosaminyltransferase V (GnT-V) exhibited an exceptionally higher K m value (4 mm) for the UDP-GlcNAc donor nucleotide sugar than other GlcNAc transferases. They proposed that the production of β1,6-branched oligosaccharide, which is formed by GnT-V, could be regulated in vivo by the concentration of the donor, UDP-GlcNAc, as well as the expression levels of the enzyme (
      • Sasai K.
      • Ikeda Y.
      • Fujii T.
      • Tsuda T.
      • Taniguchi N.
      ). The differences in K m values of each chondroitin sulfate glycosyltransferases for donor substrates may be one of the factors regulating the biosynthetic machinery of CS chains.
      The primary structure of CSS2 was highly conserved with that of CSGlcAT (57% identity) (Fig. 1). Several motifs observed in many glycosyltransferases, most of the cysteine residues, and all of the N-glycosylation potential sites were well conserved in both enzymes, suggesting that their overall molecular structure should be very similar. However, CSGlcAT lacks GalNAcT-II activity, whereas CSS2 has this activity. Many glycosyltransferases have been shown to contain a DXD motif, critical for catalytic function (
      • Boeggeman E.
      • Qasba P.K.
      ). The DXD motif may be essential for binding the UDP-sugar donor through the coordination of a divalent cation (
      • Pedersen L.C.
      • Tsuchida K.
      • Kitagawa H.
      • Sugahara K.
      • Darden T.A.
      • Negishi M.
      ). In the predicted catalytic domain for GalNAcT-II activity in CSS2, which is supposed to be in the C-terminal half of this enzyme, the typical DXD motif is absent, and the amino acid sequence in the corresponding domain is different from that of CSGlcAT (Fig. 1A). In the C-terminal half of CSS2, we found a GPD617 sequence, aligned with the DVD sequence in β4Gal-T1 and in the putative GalNAcT domains of the other chondroitin glycosyltransferases, CSS1, GSCalNAcT-1, and CSGalNAcT-2, by multiple alignment (Fig. 9). Interestingly, this aspartic acid residue is not conserved in CSGlcAT lacking GalNAcT-II activity (Fig. 9). Although the canonical DXD motif contains two aspartic acid residues, the first (in the 1st position of the motif) is relatively variable, and the second (in the 3rd position of the motif) is quite well conserved (
      • Unligil U.M.
      • Zhou S.
      • Yuwaraj S.
      • Sarkar M.
      • Schachter H.
      • Rini J.M.
      ). By analyses of the crystal structure of human β1,3-glucuronyltransferase I (GlcAT-I) (
      • Pedersen L.C.
      • Tsuchida K.
      • Kitagawa H.
      • Sugahara K.
      • Darden T.A.
      • Negishi M.
      ) and bovine β1,4-galactosyltransferase (β4Gal-T1) (
      • Boeggeman E.
      • Qasba P.K.
      ) in the presence of the donor substrate product UDP, the catalytic Mn2+ ion, and the acceptor substrate, their conserved DXD motifs were shown to directly interact with the Mn2+ ion. In both cases, the second aspartic acid residue in the DXD sequence is important for coordination with the Mn2+ ion (
      • Boeggeman E.
      • Qasba P.K.
      ,
      • Pedersen L.C.
      • Tsuchida K.
      • Kitagawa H.
      • Sugahara K.
      • Darden T.A.
      • Negishi M.
      ). Furthermore, in rabbit N-acetylglucosaminyltransferase I, the DXD motif is present in the form 211EDD213, in which the 3rd position Asp213 makes the only direct interaction with the bound Mn2+ ion and is essential for enzyme activity (
      • Unligil U.M.
      • Zhou S.
      • Yuwaraj S.
      • Sarkar M.
      • Schachter H.
      • Rini J.M.
      ). These results indicate that the second aspartic acid residue in the DXD motif may be necessary and adequate for metal ion binding and that the Asp617 residue of GPD617 in CSS2, which is absent in CSGlcAT, may be adequate for its GalNAcT-II activity. However, the other motif, the DXH motif, is also found in some other glycosyltransferases including polypeptide GalNAc transferases classified as GT-27 or GT-60 in the CAZy data base, which may be essential for its metal ion-dependent glycosyltransferase activities (
      • Hagen F.K.
      • Hazes B.
      • Raffo R.
      • deSa D.
      • Tabak L.A.
      ). Therefore, it is still difficult to anticipate the GalNAcT-II active site from the homology of the primary amino acid sequence of CSS2 with other β4-glycosyltransferases.
      The other remarkable difference in the amino acid sequences of the two enzymes, CSS2 and CSGlcAT, is two short insertions (see Fig. 1) from other chondroitin glycosyltransferases, such as CSS1, CSGalNAc-T1, and CSGalNAc-T2 (data not shown). The former insertion, in the stem region of CSS2, may alter its localization in the Golgi membrane or the interaction with other molecules (
      • Sasai K.
      • Ikeda Y.
      • Tsuda T.
      • Ihara H.
      • Korekane H.
      • Shiota K.
      • Taniguchi N.
      ), whereas the latter may influence the catalytic function of CSGlcAT. Actually, our molecular modeling suggests that this insertion is located near the predicted catalytic domain of C-CSGlcAT, which aligned with the putative GalNAcT domains of β4Gal-T1 and the other CS glycosyltransferases (Fig. 10C). In this insertion, which contains many proline and glycine residues, secondary structures were not predicted (Fig. 9). The longer amino acid loop may interfere with the binding of UDP-GalNAc or acceptor substrates to the catalytic domain.
      The genomic organization of both of the cloned chondroitin glycosyltransferases involved in chondroitin backbone formation has been shown. The open reading frames of the human CSS2 and CSGlcAT genes are distributed among four exons that span ∼4.1 and 4.7 kb, respectively. Comparison of the genomic organization of these two genes shows a quite similar genetic exon-intron organization within the coding sequence (Fig. 1) as described in the case of CSGalNAcT-1 and CSGalNAcT-2 (
      • Uyama T.
      • Kitagawa H.
      • Tanaka J.
      • Tamura J.I.
      • Ogawa T.
      • Sugahara K.
      ). Interestingly, chromosomal assignments of the five human chondroitin glycosyltransferases, CSS1, CSS2, CSGlcAT, CSGalNAcT-1, and CSGalNAcT-2, indicate that these genes are localized on different chromosomes, at 15q26.3, 2q36.1, 7q36, 8q21.3, and 10q11.22, respectively, despite significant homology in the nucleotide and amino acid sequences of the five genes (
      • Kitagawa H.
      • Uyama T.
      • Sugahara K.
      ,
      • Uyama T.
      • Kitagawa H.
      • Tanaka J.
      • Tamura J.I.
      • Ogawa T.
      • Sugahara K.
      ,
      • Sato T.
      • Gotoh M.
      • Kiyohara K.
      • Akashima T.
      • Iwasaki H.
      • Kameyama A.
      • Mochizuki H.
      • Yada T.
      • Inaba N.
      • Togayachi A.
      • Kudo T.
      • Asada M.
      • Watanabe H.
      • Imamura T.
      • Kimata K.
      • Narimatsu H.
      ,
      • Gotoh M.
      • Sato T.
      • Akashima T.
      • Iwasaki H.
      • Kameyama A.
      • Mochizuki H.
      • Yada T.
      • Inaba N.
      • Zhang Y.
      • Kikuchi N.
      • Kwon Y.D.
      • Togayachi A.
      • Kudo T.
      • Nishihara S.
      • Watanabe H.
      • Kimata K.
      • Narimatsu H.
      ,
      • Uyama T.
      • Kitagawa H.
      • Tamura Ji J.
      • Sugahara K.
      ). Furthermore, recently we found CSS3, which is homologous to CSS1, the same as CSS2 for CSGlcAT and CSGalNAcT-2 for CSGalNAcT-1.
      T. Yada, T. Sato, H. Kaseyama, M. Gotoh, H. Iwasaki, N. Kikuchi, Y. D. Kwon, A. Togayachi, T. Kudo, H. Watanabe, H. Narimatsu, and K. Kimata, unpublished observations.
      These findings suggest that chondroitin glycosyltransferases can be classified into three pairs: CSS1/CSS3, CSS2/CSGlcAT, and CSGalNAcT-2/CSGalNAcT-1. During evolution, an ancestor gene might have diverged into three genes, and each of them underwent duplication.
      The CSS2 gene exhibited ubiquitous but characteristic expression patterns compared with CSGlcAT. A particularly striking difference is its high level of expression in the pancreas and ovary and moderate levels in the placenta, stomach, and small intestine (Fig. 8). In contrast, CSGlcAT is highly expressed in the placenta, followed by the small intestine and pancreas, similar to the pattern of CSS1 expression (Fig. 8) (
      • Gotoh M.
      • Yada T.
      • Sato T.
      • Akashima T.
      • Iwasaki H.
      • Mochizuki H.
      • Inaba N.
      • Togayachi A.
      • Kudo T.
      • Watanabe H.
      • Kimata K.
      • Narimatsu H.
      ). Therefore, CSS2 may play a more unique role in the biosynthesis of chondroitin sulfate than CSGlcAT in human tissues.
      Taken together, at least five enzymes, CSS1, CSS2, CSGalNAcT-1, CSGalNAcT-2, and CSGlcAT, whose expression patterns in human tissues are ordinarily similar to each other, may play roles at different stages in the synthesis of CS in a cooperative or orchestrated manner. The recent cloning, expression, and characterization of many glycosyltransferases have provided great progress in understanding chondroitin sulfate biosynthesis. Because none of these enzymes can form an entire glycosaminoglycan chain, the formation of a complex as machinery may be required for CS chain elongation in vitro. To understand the nature of the interactions of these molecules forming a complex, the various membrane-bound enzymes, appropriate substrates, and membrane-bound nascent proteoglycans are fundamental for determining the specificities of structure and efficiency in formation of the CS chain.

      Acknowledgments

      We thank Seikagaku Corp. for providing the substrates and the enzymes. We also thank Drs. T. Ninomiya and N. Sugiura for E. coli strain K4 chondroitin polymerase.

      References

        • Schwartz N.B.
        • Domowicz M.
        Glycobiology. 2002; 12: 57R-68R
        • Bandtlow C.E.
        • Zimmermann D.R.
        Physiol. Rev. 2000; 80: 1267-1290
        • Kimata K.
        • Okayama M.
        • Ooira A.
        • Suzuki S.
        Mol. Cell. Biochem. 1973; 1: 211-228
        • Silbert J.E.
        • Sugumaran G.
        IUBMB Life. 2002; 54: 177-186
        • Vertel B.M.
        • Walters L.M.
        • Flay N.
        • Kearns A.E.
        • Schwartz N.B.
        J. Biol. Chem. 1993; 268: 11105-11111
        • Inoue H.
        • Otsu K.
        • Yoneda M.
        • Kimata K.
        • Suzuki S.
        • Nakanishi Y.
        J. Biol. Chem. 1986; 261: 4460-4469
        • Gotting C.
        • Kuhn J.
        • Zahn R.
        • Brinkmann T.
        • Kleesiek K.
        J. Mol. Biol. 2000; 304: 517-528
        • Almeida R.
        • Levery S.B.
        • Mandel U.
        • Kresse H.
        • Schwientek T.
        • Bennett E.P.
        • Clausen H.
        J. Biol. Chem. 1999; 274: 26165-26171
        • Okajima T.
        • Yoshida K.
        • Kondo T.
        • Furukawa K.
        J. Biol. Chem. 1999; 274: 22915-22918
        • Bai X.
        • Zhou D.
        • Brown J.R.
        • Crawford B.E.
        • Hennet T.
        • Esko J.D.
        J. Biol. Chem. 2001; 276: 48189-48195
        • Kitagawa H.
        • Tone Y.
        • Tamura J.
        • Neumann K.W.
        • Ogawa T.
        • Oka S.
        • Kawasaki T.
        • Sugahara K.
        J. Biol. Chem. 1998; 273: 6615-6618
        • Wei G.
        • Bai X.
        • Sarkar A.K.
        • Esko J.D.
        J. Biol. Chem. 1999; 274: 7857-7864
        • Maimone M.M.
        • Tollefsen D.M.
        J. Biol. Chem. 1990; 265: 18263-18271
        • Lyon M.
        • Deakin J.A.
        • Rahmoune H.
        • Fernig D.G.
        • Nakamura T.
        • Gallagher J.T.
        J. Biol. Chem. 1998; 273: 271-278
        • Razin E.
        • Stevens R.L.
        • Akiyama F.
        • Schmid K.
        • Austen K.F.
        J. Biol. Chem. 1982; 257: 7229-7236
        • Maeda N.
        • Ichihara-Tanaka K.
        • Kimura T.
        • Kadomatsu K.
        • Muramatsu T.
        • Noda M.
        J. Biol. Chem. 1999; 274: 12474-12479
        • Naujokas M.F.
        • Morin M.
        • Anderson M.S.
        • Peterson M.
        • Miller J.
        Cell. 1993; 74: 257-268
        • Fried M.
        • Duffy P.E.
        Science. 1996; 272: 1502-1504
        • Yamauchi S.
        • Mita S.
        • Matsubara T.
        • Fukuta M.
        • Habuchi H.
        • Kimata K.
        • Habuchi O.
        J. Biol. Chem. 2000; 275: 8975-8981
        • Uchimura K.
        • Muramatsu H.
        • Kadomatsu K.
        • Fan Q.W.
        • Kurosawa N.
        • Mitsuoka C.
        • Kannagi R.
        • Habuchi O.
        • Muramatsu T.
        J. Biol. Chem. 1998; 273: 22577-22583
        • Kobayashi M.
        • Sugumaran G.
        • Liu J.
        • Shworak N.W.
        • Silbert J.E.
        • Rosenberg R.D.
        J. Biol. Chem. 1999; 274: 10474-10480
        • Ohtake S.
        • Ito Y.
        • Fukuta M.
        • Habuchi O.
        J. Biol. Chem. 2001; 276: 43894-43900
        • Kitagawa H.
        • Uyama T.
        • Sugahara K.
        J. Biol. Chem. 2001; 276: 38721-38726
        • Uyama T.
        • Kitagawa H.
        • Tanaka J.
        • Tamura J.I.
        • Ogawa T.
        • Sugahara K.
        J. Biol. Chem. 2002; 278: 3072-3078
        • Sato T.
        • Gotoh M.
        • Kiyohara K.
        • Akashima T.
        • Iwasaki H.
        • Kameyama A.
        • Mochizuki H.
        • Yada T.
        • Inaba N.
        • Togayachi A.
        • Kudo T.
        • Asada M.
        • Watanabe H.
        • Imamura T.
        • Kimata K.
        • Narimatsu H.
        J. Biol. Chem. 2002; 278: 3063-3071
        • Gotoh M.
        • Sato T.
        • Akashima T.
        • Iwasaki H.
        • Kameyama A.
        • Mochizuki H.
        • Yada T.
        • Inaba N.
        • Zhang Y.
        • Kikuchi N.
        • Kwon Y.D.
        • Togayachi A.
        • Kudo T.
        • Nishihara S.
        • Watanabe H.
        • Kimata K.
        • Narimatsu H.
        J. Biol. Chem. 2002; 277: 38189-38196
        • Uyama T.
        • Kitagawa H.
        • Tamura Ji J.
        • Sugahara K.
        J. Biol. Chem. 2002; 277: 8841-8846
        • Gotoh 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: 38179-38188
        • Zak B.M.
        • Crawford B.E.
        • Esko J.D.
        Biochim. Biophys. Acta. 2002; 1573: 346-355
        • Bitter T.
        • Muir H.M.
        Anal. Biochem. 1962; 4: 330-334
        • Ninomiya T.
        • Sugiura N.
        • Tawada A.
        • Sugimoto K.
        • Watanabe H.
        • Kimata K.
        J. Biol. Chem. 2002; 277: 21567-21575
        • Iwai T.
        • Inaba N.
        • Naundorf A.
        • Zhang Y.
        • Gotoh M.
        • Iwasaki H.
        • Kudo T.
        • Togayachi A.
        • Ishizuka Y.
        • Nakanishi H.
        • Narimatsu H.
        J. Biol. Chem. 2002; 277: 12802-12809
        • Ogata K.
        • Umeyama H.
        J. Mol. Graphics Model. 2000; 18 (305–256): 258-272
        • Thompson J.D.
        • Higgins D.G.
        • Gibson T.J.
        Nucleic Acids Res. 1994; 22: 4673-4680
        • Jones D.T.
        J. Mol. Biol. 1999; 292: 195-202
        • Sasai K.
        • Ikeda Y.
        • Tsuda T.
        • Ihara H.
        • Korekane H.
        • Shiota K.
        • Taniguchi N.
        J. Biol. Chem. 2001; 276: 759-765
        • Kitagawa H.
        • Tsuchida K.
        • Ujikawa M.
        • Sugahara K.
        J. Biochem. (Tokyo). 1995; 117: 1083-1087
        • Kitagawa H.
        • Ujikawa M.
        • Tsutsumi K.
        • Tamura J.
        • Neumann K.W.
        • Ogawa T.
        • Sugahara K.
        Glycobiology. 1997; 7: 905-911
        • Gundlach M.W.
        • Conrad H.E.
        Biochem. J. 1985; 226: 705-714
        • Inoue H.
        • Otsu K.
        • Suzuki S.
        • Nakanishi Y.
        J. Biol. Chem. 1986; 261: 4470-4475
        • Otsu K.
        • Inoue H.
        • Tsuzuki Y.
        • Yonekura H.
        • Nakanishi Y.
        • Suzuki S.
        Biochem. J. 1985; 227: 37-48
        • Kitagawa H.
        • Tsutsumi K.
        • Ujikawa M.
        • Goto F.
        • Tamura J.
        • Neumann K.W.
        • Ogawa T.
        • Sugahara K.
        Glycobiology. 1997; 7: 531-537
        • Sasai K.
        • Ikeda Y.
        • Fujii T.
        • Tsuda T.
        • Taniguchi N.
        Glycobiology. 2002; 12: 119-127
        • Boeggeman E.
        • Qasba P.K.
        Glycobiology. 2002; 12: 395-407
        • Pedersen L.C.
        • Tsuchida K.
        • Kitagawa H.
        • Sugahara K.
        • Darden T.A.
        • Negishi M.
        J. Biol. Chem. 2000; 275: 34580-34585
        • Unligil U.M.
        • Zhou S.
        • Yuwaraj S.
        • Sarkar M.
        • Schachter H.
        • Rini J.M.
        EMBO J. 2000; 19: 5269-5280
        • Hagen F.K.
        • Hazes B.
        • Raffo R.
        • deSa D.
        • Tabak L.A.
        J. Biol. Chem. 1999; 274: 6797-6803