Enzymatic synthesis of chondroitin with a novel chondroitin sulfate N-acetylgalactosaminyltransferase that transfers N-acetylgalactosamine to glucuronic acid in initiation and elongation of chondroitin sulfate synthesis.

We found a novel glycosyltransferase gene having a hypothetical beta 1,4-galactosyltransferase motif (GenBank accession number ) by a BLAST search and cloned its full-length open reading frame using the 5'-rapid amplification of cDNA ends method. The truncated form was expressed in insect cells as a soluble enzyme. It transferred N-acetylgalactosamine, not galactose, to para-nitrophenyl-beta-glucuronic acid. The N-acetylgalactosamine-glucuronic acid linkage has been identified only in chondroitin sulfate; therefore, we examined its chondroitin elongation and initiation activities. N-Acetylgalactosaminyltransferase activity was observed toward chondroitin poly- and oligosaccharides, chondroitin sulfate oligosaccharides, and linkage tetrasaccharide (GlcA-Gal-Gal-Xyl-O-methoxyphenyl), and the chondroitin polysaccharide and linkage tetrasaccharide were better acceptor substrates than the others. Northern blot analysis and quantitative real-time PCR analysis revealed that its 4-kb transcripts were highly expressed in thyroid and placenta, although they were ubiquitously expressed in various tissues and cells. These results suggest that this enzyme has N-acetylgalactosaminyltransferase activity in both the elongation and initiation of chondroitin sulfate synthesis. Furthermore, we performed enzymatic synthesis of chondroitin pentasaccharide in vitro. In one tube reaction with four enzymes, beta 1,4-galactosyltransferase-VII, beta 1,3-galactosyltransferase-VI, glucuronyltransferase-I, and this enzyme, and a synthetic xylose-peptide acceptor, the structure GalNAc-GlcA-Gal-Gal-Xyl-peptide was constructed. This is the first report of a chondroitin pentasaccharide constructed with recombinant glycosyltransferases in vitro.

* This work was performed as part of the R&D Project of Industrial Science and Technology Frontier Program (R&D for Establishment and Utilization of a Technical Infrastructure for Japanese Industry) supported by the New Energy and Industrial Technology Development Organization. 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.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s) AB081516.
CS consists of many kinds of sulfated glucuronic acid (GlcA) and N-acetylgalactosamine (GalNAc) (20). The C2 and C3 positions of GlcA and the C4 and C6 positions of GalNAc are the sites possibly sulfated by the respective sulfotransferases. There are reports that the sulfation is related to the elongation and termination of the CS chain. Using bovine serum, UDP-GalNAc and various kinds of sulfated (GlcA␤1-3GalNAc␤1-4) n , 6-O-or 4-O-sulfated GalNAc in the acceptor stimulates the GalNAc-T activity, whereas terminal 3-O-sulfated GlcA or penultimate 4,6-O-disulfated GalNAc inhibits the activity (21). Gundlach and Conrad (22) demonstrated acceptor substrate specificity to GlcA-T and GalNAc-T in a microsomal preparation from chick embryo chondrocytes using even-and oddnumbered oligosaccharides prepared from chondroitin, CS-A, CS-C, and hyaluronan. They showed that oligosaccharides prepared from CS-C and chondroitin were active as acceptors for both enzymes, whereas oligosaccharides from CS-A and hyaluronan were mostly inactive. In the case of CSGlcA-T, the sulfated oligosaccharides prepared from CS-C made better substrates than non-sulfated oligosaccharides (19).
The progress made with the human genome project and other databases, such as expressed sequence tags (ESTs) and full-length complementary DNAs (cDNAs), has enabled us to search for novel genes that are homologous to known genes. We searched these databases utilizing the amino acid sequences of the ␤4Gal-T family while paying particular attention to the existence of a trans-membrane domain and conserved motif regions, DXD and GWGGED, as query sequences. In this study, we focused on a novel gene, which possesses the GWGGED motif and has elongation and initiation activities in the synthesis of CS, although its cloning and a brief characterization were reported while we were preparing this report (23). Finally, we show that the synthesis of chondroitin pentasaccharide is possible using the three reported enzymes and this novel enzyme in vitro.

EXPERIMENTAL PROCEDURES
Isolation of a Novel Human CSGalNAc-T cDNA-We performed a BLAST search of the EST databases using amino acid sequences of the cloned human ␤1,4-galactosyltransferases (␤4Gal-T) as queries and found a novel EST (GenBank TM accession number AK055154) that possesses the GWGGED motif for ␤4Gal-T. To obtain the open reading frame sequence, we performed the 5Ј-rapid amplification of cDNA ends (5Ј-RACE) technique with a Marathon-Ready cDNA amplification kit (Clontech, Palo Alto, CA). The 5Ј sequence was extended by two-step 5Ј-RACE. Two reverse primers, 5Ј-ACCTATATTGATGAAGTCTGACC-GA-3Ј for the first PCR and 5Ј-GTCTGACCGATACTGACACGTCAT-3Ј for the nested PCR, were employed for the first-step extension. The second-step 5Ј-RACE was carried out with the reverse primers 5Ј-TC-CTGCAGCTCCTCCTTGAGCT-3Ј and 5Ј-ATCTGCCGCTTCAGG-CTGCTCA-3Ј.
Construction and Purification of CSGalNAc-T Protein Fused with FLAG Peptide-The putative catalytic domain of the enzyme (amino acids 37-532) was expressed as a secreted protein fused with a FLAG peptide in insect cells according to the instruction manual of GATE-WAY Cloning Technology (Invitrogen, Groningen, Netherlands). A DNA fragment of ϳ1.6 kb was amplified by PCR using the Marathon-Ready cDNA derived from human bone marrow (Clontech), as a template, and two primers, 5Ј-GGGGACAAGTTTGTACAAAAAAGCAGG-CTTCAAAGGTGACGAGGAGCAGCTGGCAC-3Ј and 5Ј-GGGGACCA-CTTTGTACAAGAAAGCTGGGTCTCATGTTTTTTTGCTACTTGTCT-TCTGT-3Ј. The amplified fragment was inserted into pFBIF to construct pFBIF-CSGalNAc-T as described previously (24). The CSGal-NAc-T catalytic domain was expressed in Sf21 insect cells. A 50-ml volume of culture medium was mixed with anti-FLAG M1 antibody resin (Sigma, St. Louis, MO). The protein-resin mixture was washed twice with 50 mM TBS (50 mM Tris-HCl, pH 7.4, and 150 mM NaCl) containing 1 mM CaCl 2 and suspended in 100 l of each of the assay buffers described below.
For the reaction of Gal-T assay, a 14 mM Hepes buffer (pH 7.4) containing 250 M UDP-Gal, 12.5 mM MnCl 2 , and 500 M of each acceptor substrate was used. For the GalNAc-T assay, a 50 mM MES buffer (pH 6.5) containing 0.1% Triton X-100, 1 mM UDP-GalNAc, 10 mM MnCl 2 , and 500 M of each acceptor substrate was used. A 10-l volume of enzyme solution for 20 l of each reaction mixture was added, and the solution was incubated at 37°C for 2 h for Gal-T and 16 h for GalNAc-T.
After the incubation, the mixture was filtrated with an Ultrafree-MC column (Millipore, Bedford, MA), and a 10-l aliquot was subjected to reversed-phase high performance liquid chromatography (HPLC) on an ODS-80Ts QA column (4.6 ϫ 250 mm, Tosoh, Tokyo, Japan). 0.1% TFA/H 2 O was used as a running solution, and the products were eluted with a 0 -15% (for GlcA-␤-pNp) or 7-10% (for linkage tetrasaccharidemethoxyphenyl) acetonitrile gradient at a flow rate of 1.0 ml/min at 50°C. For glycosylated peptides, H 2 O containing 0.1% TFA and 21% acetonitrile was utilized as the running solution. An ultraviolet spectrophotometer (absorbance at 210 nm), SPD-10A VP (Shimadzu, Kyoto, Japan) was used for detection of the peaks. For the analysis of glycosylated peptide, labeling was carried out with Cy5 (Amersham Biosciences), and fluorescence was detected with a fluorescence detector, RF-10A XL (Shimadzu). For the analysis of elongation activity, a CSGal-NAc-T reaction mixture containing 100 g of chondroitin or CS and 40,000 -55,500 dpm of UDP-[ 14 C]GalNAc was used. After a 1-h incubation at 37°C, the reaction mixture was filtrated and fractionated with a G2500PW column (Tosoh, Tokyo, Japan) or a Superdex 30-pg column (Amersham Biosciences). The radioactivity of each fraction was monitored by liquid scintillation spectrophotometry.
Determination of Products by Mass Spectrometry Using CSGal-NAcT-An additional peak obtained by reversed-phase chromatogra- Northern Blot Analysis-The 3Ј-untranslated region (695 bp) was used as a probe for Northern blot analysis. The probe was amplified by PCR with 5Ј-ATGCCATATCCAAGGACATGCCAA-3Ј and 5Ј-AACAAT-GAGCTTACTGTGAGCAAAC-3Ј as primers. Human Multiple Tissue Northern Blot III was purchased from Clontech. For the labeling and detection of probe, an AlkPhos Direct Labeling and Detection System with CDP-Star and Hyperfilm ECL (Amersham Biosciences) were utilized.
Quantitative Analysis of the CSGalNAc-T Transcript in Human Tissues and Cell Lines by Real-time PCR-For quantification of CSGal-NAc-T transcripts, we employed the real-time PCR method, as described in detail previously (24 -26). Marathon Ready cDNAs derived from various human tissues and cell lines (HL-60 and G361) were purchased from Clontech. cDNAs derived from other cell lines were prepared with the SuperScript First-strand Synthesis System for RT-PCR (Invitrogen) as described in our previous study (27). Standard curves for the endogenous control, glyceraldehyde-3-phosphate dehydrogenase (GAPDH) cDNAs, were generated by serial dilution of a pCR2.1 (Invitrogen) DNA containing the GAPDH gene. The primer set and the probe for CSGalNAc-T were as follows: forward primer, 5Ј-GACTTCATCAATATAGGTGGGTTTGAT-3Ј; reverse primer, 5Ј-GTC-CGTACCACTATGAGGTTGCT-3Ј; and probe, 5Ј-ACCTTTATCGCAAG-TATCT-3Ј with the minor groove binder (28). PCR products were continuously measured with an ABI Prism 7700 sequence detection system (Applied Biosystems, Foster City, CA). The relative amount of CSGalNAc-T transcript was normalized to the amount of GAPDH transcript in the same cDNA.

Comparison of Amino Acid Sequences between CSGalNAc-T
and Known ␤1,4-Glycosyltransferases-While we were preparing this report, a paper on the cloning and characterization of CSGalNAc-T was published (23). The nucleotide and putative amino acid sequences of the clone we obtained were the same as those of Uyama et al.'s cDNA clone (GenBank TM accession numbers AB081516 and AB071403). This enzyme maintained local homology with ␤4Gal-T1 in the putative catalytic domain as shown in Fig. 1, however, no remarkable homology with ␤4GalNAc-T was found (data not shown) (29). A DXD motif, which is conserved in many glycosyltransferases and functions as a key sequence for divalent cation binding, existed in the putative catalytic domain. Another motif, GWGGED, which is highly conserved among the ␤4Gal-T family except ␤4Gal-T7 (GWGRED), was also found.
Substrate Specificity of CSGalNAc-T-We determined the substrate specificity of the soluble enzyme expressed in insect cells. Contrary to expectations, it did not exhibit any galactosyltransferase activity (data not shown). As shown in Fig. 2, we observed an additional positive peak, at 55.4 min, of the reaction product when UDP-GalNAc and GlcA-pNp were used as a donor substrate and an acceptor substrate, respectively (Fig.  2B). To confirm that the peak is the real reaction product, ESI-MS was employed. In the positive-ion mode, a peak of 519.05 m/z was the same molecular weight as GalNAc-GlcA-pNp (Fig. 2C). When this peak was subjected to further analysis by ESI-MS/MS, new peaks appeared at 203.91 and 298.97 m/z (Fig. 2D). The former peak had exactly the same molecular weight as GalNAc, and the latter fitted that of GlcA-pNp missing a water molecule. ESI-MS in the negative-ion mode revealed that the product could be GalNAc-GlcA-pNp (517.19 m/z), and it produced GlcA-pNp without a water molecule (296.00 m/z) (data not shown). These results demonstrated that the enzyme has GalNAc-T activity and transfers GalNAc to GlcA. To date, the linkage combining GalNAc and GlcA has been identified only in CS. At this point, we considered the enzyme to be involved in the synthesis of CS and named it chondroitin sulfate N-acetylgalactosaminyltransferase, CSGalNAc-T.
Determination of CSGalNAc-T Activity in Elongation of Chondroitin Poly-and Oligosaccharides-Two kinds of Gal-NAc-GlcA linkages are known in CS, one in its polymer structure (3GalNAc␤1-4GlcA␤1-) n and the other between the polymer CS and a linkage tetrasaccharide (GlcA␤1-3Gal␤1-3Gal␤1-4Xyl). To identify whether it has elongation activity, chondroitin and CS were utilized as acceptor substrates. When chondroitin was used for the acceptor, an additional peak was detected by reversed-phase chromatography, although an extremely low peak appeared using CS-C as the acceptor (indicated by an arrow in Fig. 3). The fractions of this peak were collected and identified by treatment with chondroitinase ACII (data not shown).
Determination of CSGalNAc-T Activity in Initiation of Chondroitin Synthesis-The initiation activity of CSGalNAc-T was determined with a linkage tetrasaccharide (GlcA␤1-3Gal␤1-3Gal␤1-4Xyl␤1-O-methoxyphenyl) that was chemically synthesized. As shown in Fig. 4, a peak (P) appeared at 28.9 min (Fig. 4B) in addition to the acceptor substrate peak (S) at 29.7 min (Fig. 4, A and B). The peaks S and P were isolated by reversed-phase chromatography and identified with MALDI-TOF MS. The peak S gave a molecular mass of 779.20 m/z, the same as that of the linkage tetrasaccharide (Fig. 4C). The peak P gave two peaks of 982.23 and 1004.21 m/z as shown in Fig.  4D. The peaks of 982.23 and 1004.21 m/z were the same molecular weight as GalNAc-linkage tetrasaccharide-O-methoxyphenyl and GalNAc-linkage tetrasaccharide-O-methoxyphenyl with Na ϩ , respectively. Moreover, the peak P was identified to have GalNAc on its non-reducing terminus with chondroitinase ACII treatment (data not shown). These results demonstrated that CSGalNAc-T acts to initiate chondroitin sulfate synthesis by transferring GalNAc to GlcA at the non-reducing terminus of the linkage tetrasaccharide.
Acceptor Substrate Specificity of CSGalNAc-T-The CSGalNAc-T activity toward various kinds of acceptors is summarized in Table I. CSGalNAc-T preferred the best linkage tetrasaccharide of all substrates examined. Regarding elongation activity, chondroitin was the preferred acceptor substrate, whereas the sulfated CS polysaccharides were an unsuitable substrate for CSGalNAc-T. The activity levels for the longer oligosaccharide prepared from chondroitin and CS were higher than those for the shorter ones, although they were rather low.
Tissue Distribution of CSGalNAc-T Transcripts-To determine the size of the CSGalNAc-T transcripts, Northern blot analysis was performed. As shown in Fig. 5A, the transcript was ϳ4 kb in length. We did not observe any other bands except the 4-kb band. Its expression was found in all tissues examined, however, the levels differed among tissues. The expression levels of CSGalNAc-T in various tissues and cell lines, as determined by quantitative real-time PCR, were shown as the relative amount versus glyceraldehyde-3-phosphate dehydrogenase (GAPDH) transcripts (Fig. 5, B and C). The transcripts were highly expressed in thyroid and placenta, although they were ubiquitously expressed in other tissues. However, their expression levels were extremely low in almost all cell lines examined.  (Fig. 6). As shown in Fig. 6B, when three known glycosyltransferases, ␤4Gal-T7, ␤3Gal-T6, and GlcAT-I, the specificities of which for the GAG synthesis were already reported, were utilized in the enzyme reaction, three additional peaks were detected by reversed-phase chromatography. Before this experiment, we had confirmed that each peak, P1, P2, and P3, is a sequential product of ␤4Gal-T7, ␤3Gal-T6, and GlcAT-I, respectively (data not shown). Each peak was identified to be Gal-Xyl-peptide (P1), Gal-Gal-Xyl-peptide (P2), or GlcA-Gal-Gal-Xyl-peptide (P3) by the position of control oligosaccharide-peptides, which were prepared by a sequential enzyme reaction using each of the three enzymes as illustrated in Fig. 6D. When CSGalNAc-T was added to the reaction mixture in addition to the three enzymes, a novel peak (P4) appeared (Fig. 6C) at 35.0 min. Chondroitinase ACII treatment of P4 resulted in the apparent digestion of P4, indicating that it has a GalNAc␤1-4GlcA linkage in its structure (data not shown).
These results demonstrated that the chondroitin pentasaccharide with the peptide can be synthesized in vitro, using three known specific enzymes and one novel glycosyltransferase, CSGalNAc-T. DISCUSSION We found that a novel enzyme, CSGalNAc-T, exhibited GalNAc-T activity toward GlcA. From its acceptor specificity, it was suggested that CSGalNAc-T is involved in both the elongation and initiation of CS synthesis. While we were preparing this report, a paper describing the cloning and characterization of the same enzyme, CSGalNAc-T, was published (23). However, the activity toward the linkage tetrasaccharide was not described by these authors. Here, we demonstrate an apparent initiation and elongation activity, in addition, we examined in greater detail the substrate specificity toward various acceptors. We also demonstrate the following points, which Uyama et al. did not detect or examine: 1) CSGalNAc-T showed apparent initiation activity toward the linkage tetrasaccharide. 2) Regarding the elongation activity, CSGalNAc-T preferred chondroitin to CS. No species of CS, CS-A, -B, -C, or -D, was a suitable acceptor for CSGalNAc-T. 3) The activity was dependent on the length of oligosaccharide chain as an acceptor. 4) Finally, we could construct a chondroitin pentasaccharide on the peptide using the four enzymes, including CSGalNAc-T, by synthesis in vitro.
Because CSGalNAc-T contained the ␤4Gal-T motif (GWGGED) in its putative catalytic domain, we considered at first that it would exert ␤4Gal-T activity. Unexpectedly, it transferred GalNAc, not Gal, to GlcA-pNp. This enzyme is the first example demonstrating that the GWGGED motif is shared not only by ␤4Gal-T but also by ␤4GalNAc-T. Based on the crystallized structure of bovine ␤4Gal-T1, it is proposed that the Trp residue at position 314 is necessary to lock with a donor substrate, UDP-Gal, with its strict specificity. The Gly-313 residue is proposed to be also important for the rotation of Trp314 to maintain Gal-T activity (9). However, many members of the UDP-GalNAc:polypeptide N-acetylgalactosaminyltransferase (pp-GalNAc-T) family, for which a donor substrate is UDP-GalNAc, possess the WGGE sequence in their catalytic domains, but not the GWGGED sequence (30 -34), whereas

Enzymatic Synthesis of Chondroitin with CSGalNAc-T
CSGalNAc-T possesses the GWGGED motif. Thus, we speculate that the first Gly and Trp residues in the GWGGED motif do not determine the specificity for a donor substrate.
The product of CSGalNAc-T was GalNAc-GlcA, which has been identified only in CS. It was predicted that CSGalNAc-T is the enzyme involved in the synthesis of CS. Chondroitin has the structure of disaccharide units in the polymerized chain, (3GalNAc␤1-4GlcA␤1-) n , and the linkage tetrasaccharide (GlcA␤1-3GalNAc␤1-4GlcA␤1-3Gal␤1-3Gal␤1-4Xyl␤1-O-Ser). In the present study, we defined the ability of CSGalNAc-T to make both linkages as elongation and initiation activity. CSS was reported to have elongation activity but did not exhibit initiation activity (18). Regarding the elongation activity of CSS, it could efficiently transfer GalNAc to chondroitin polysaccharide but not to CS polysaccharide. This is similar to the CSGalNAc-T activity.
Sulfation affects enzyme activity for chondroitin chain elongation and termination. CSGalNAc-T partially purified from bovine serum was examined for its activity toward various sulfated chondroitins (21). It showed higher levels of activity A, the synthetic Xyl-peptide was detected by a reversed-phase chromatography. The acceptor substrate was incubated with three glycosyltransferases, ␤4Gal-T7, ␤3Gal-T6, and GlcAT-I (B) or four glycosyltransferases, ␤4Gal-T7, ␤3Gal-T6, GlcAT-I, and CSGalNAc-T (C) and analyzed by the chromatography. For the reaction buffer 50 mM MES (pH 6.5) was employed. Each peak was identified by the control oligosaccharide-peptide, which was prepared by the sequential enzyme reaction (D). S, Xyl-peptide; P1, Gal-Xyl-peptide; P2, Gal-Gal-Xylpeptide; P3, GlcA-Gal-Gal-Xyl-peptide; P4, GalNAc-GlcA-Gal-Gal-Xyl-peptide. sulfated GalNAc. However, these acceptors obtained from natural materials are mixtures of sulfated GalNAc and GlcA residues and may contain unexpected sulfated sugars that inhibit the elongation activity of CSGalNAc-T. Another possibility is that a different CSGalNAc-T functions in sulfated chondroitin elongation. In fact, we have found another candidate in a data base, temporarily named CSGalNAc-T2, which has a highly homologous sequence to CSGalNAc-T. CSGalNAc-T preferred longer chondroitins than shorter ones (Table I). However, there is a possibility that polymer chondroitin without any sulfation is not abundant in natural substances. If this is true, the physiological function of CSGalNAc-T may be the initiation of CS synthesis. In such a case, some other CSGalNAc-Ts, such as CSS and an unknown CSGalNAc-T, may function in the elongation of CS.
Another glycosaminoglycan, heparan sulfate (HS)/heparin has a similar structure, GlcA␤1-4GlcNAc␣1-4 disaccharide units on the same linkage tetrasaccharide. Five glycosyltransferases, i.e. EXT1, EXT2, EXTL1, EXTL2, and EXTL3, have been cloned and identified to participate in the synthesis of heparan in addition to the four enzymes for the synthesis of linkage tetrasaccharide (5,6,(15)(16)(17). EXT1 and EXT2 are the heparan synthases that have both ␣1,4-N-acetylglucosaminyltransferase (␣4Gn-T) and ␤1,4-glucuronyltransferase (␤4GlcA-T) activities and elongate the heparan units for polymerization (35)(36)(37). Three enzymes, EXTL1, EXTL2, and EXTL3, have only ␣4Gn-T activity and do not have ␤4GlcA-T activity. EXTL2 and EXTL3 can initiate the synthesis of heparan (38) by addition of GlcNAc with an ␣1,4-linkage to the linkage tetrasaccharide. EXTL1 and EXTL3 can transfer a GlcNAc residue to heparan polymers (39). In this study, we demonstrated that CSGalNAc-T plays a similar role in the synthesis of CS as EXTL3 in the synthesis of HS. In the case of chondroitin synthesis, only CSS, CSGlcA-T, and CSGalNAc-T have been reported. If the mechanism of CS synthesis is similar to that of HS synthesis, unknown members, which are each a counterpart to the respective enzyme involved in HS synthesis, may exist to be cloned.
CS is ubiquitously synthesized in many tissues and cells, and CSGalNAc-T transcripts were also expressed in all tissues examined, as were the two other genes potentially involved in CS biosynthesis, CSS and CSGlcA-T (18,19). The expression of CSGalNAc-T was found to be particularly high in thyroid and placenta in accord with Uyama's report (23). Placenta contains significant dermatan sulfate and CS (40), but the reason for a high level of CSGalNAc-T in thyroid is less easy to understand. CSS and CSGlcA-T transcripts were not elevated in thyroid (19), and if the major role of CSGalNAc-T is the initiation of CS, then chains might be initiated on proteoglycans in thyroid but not extended due to a lack of CSS and CSGlcA-T. Further work is required to resolve this. In contrast to the results in tissues, the expression levels of CSGalNAc-T was extremely low in all cell lines examined. This was particularly noticeable for cancer tissues and cultured tumor cells. In cancer tissues the CS content of the extracellular matrix has been shown to be elevated compared with normal tissues (41,42). However, it remains to be determined how the rates of CS synthesis differ among cancer cells in tissues, cells in normal tissue, and in cultured tumor cell lines, and if the expression of CSGalNAc-T is directly correlated with these rates of CS synthesis.
Finally, we performed an enzymatic synthesis of a foundation pentasaccharide of CS on a peptide, using a set of recombinant glycosyltransferases. All recombinant enzymes prepared in a soluble form were active in synthesizing the pentasaccharide on the peptide, the sequence of which is the N terminus of bikunin, a Kunitz-type proteinase inhibitor. It is known that CS is synthesized on a serine residue in this region (43). The peptide-sugar chain synthesized enzymatically in vitro would be a very useful tool to investigate which amino acid residues are recognized by each enzyme for CS and HS synthesis and which amino acid residues or sulfations determine the classification of CS and HS. It is also a valuable acceptor substrate to synthesize a complex structure of CS using additional glycosyltransferases and sulfotransferases. This is the first time a chondroitin pentasaccharide has been constructed using recombinant glycosyltransferases in vitro.