β1,4-Galactosyltransferase (β4GalT)-IV Is Specific for GlcNAc 6-O-Sulfate β4GalT-IV ACTS ON KERATAN SULFATE-RELATED GLYCANS AND A PRECURSOR GLYCAN OF 6-SULFOSIALYL-LEWIS X

The Galβ1→4(SO 3 − →6)GlcNAc moiety is present in various N-linked andO-linked glycans including keratan sulfate and 6-sulfosialyl-Lewis X, an L-selectin ligand. We previously found β1,4-galactosyltransferase (β4GalT) activity in human colonic mucosa, which prefers GlcNAc 6-O-sulfate (6SGN) as an acceptor to non-substituted GlcNAc (Seko, A., Hara-Kuge, S., Nagata, K., Yonezawa, S., and Yamashita, K. (1998) FEBS Lett. 440, 307–310). To identify the gene for this enzyme, we purified the enzyme from porcine colonic mucosa. The purified enzyme had the characteristic requirement of basic lipids for catalytic activity. Analysis of the partial amino acid sequence of the enzyme revealed that the purified β4GalT has a similar sequence to human β4GalT-IV. To confirm this result, we prepared cDNA for each of the seven β4GalTs cloned to date and examined substrate specificities using the membrane fractions derived from β4GalT-transfected COS-7 cells. When using severalN-linked and O-linked glycans with or without 6SGN residues as acceptor substrates, only β4GalT-IV efficiently recognized 6SGN, keratan sulfate-related oligosaccharides, and Galβ1→3(SO 3 − →6GlcNAcβ1→6) GalNAcα1-O-pNP, a precursor for 6-sulfosialyl-Lewis X. These results suggested that β4GalT-IV is a 6SGN-specific β4GalT and may be involved in the biosynthesis of various glycoproteins carrying a 6-O-sulfatedN-acetyllactosamine moiety.

It has been clarified so far (18 -29) that seven ␤4GalT genes exist. All the ␤4GalTs are synthesized as type II membranebound proteins and reside in the Golgi apparatus (30). These seven enzymes have been characterized in terms of substrate specificity. ␤4GalT-I is abundant in bovine and human milk in a soluble form and is the first galactosyltransferase for which the corresponding cDNA has been isolated (18 -21). ␤4GalT-I acts on non-reducing terminal GlcNAc as an acceptor, and whereas in the presence of ␣-lactalbumin, the enzyme prefers Glc as a lactose synthase to GlcNAc (31,32). This enzyme is involved in the elongation of poly-N-acetyllactosamine repeats (33). ␤4GalT-II and -III act on GlcNAc residues in several glycoproteins and specific glycolipids, and ␤4GalT-II is affected by ␣-lactalbumin in a similar manner to ␤4GalT-I (22). ␤4GalT-IV acts on Lc 3 (GlcNAc␤133Gal␤134Glc␤13 Cer) (26) and Gal␤133(GlcNAc␤136)GalNAc(core2); the latter galactosylation at the C-4 of GlcNAc leads to the formation of the sialyl-Lewis X structure (34). ␤4GalT-V has been shown to have strong activity for core2 and core6(GlcNAc␤136GalNAc) (35). The enzyme also recognized GlcNAc␤132-(GlcNAc␤136)Man␣136 moieties in N-linked tetra-antennary * 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.
ʈ To whom correspondence should be addressed: Dept. of Biochemistry, Sasaki Institute, Kanda-Surugadai 2-2, Chiyoda-ku, Tokyo 101-0062, Japan. glycans (23,36) and was suggested to be involved in the biosynthesis of tumor-associated N-glycans in concert with ␤-N-acetylglucosaminyltransferase V (37). Lactosylceramide synthase has been purified from rat brain, and its molecular cloning was performed based on partial amino acid sequences (25). The enzyme is an orthologue of human ␤4GalT-VI (24,27). ␤4GalT-VII is a ␤-Xyl:␤1,4GalT, equal to galactosyltransferase-I which is involved in the synthesis of the proximal sequence in various glycosaminoglycans, Gal␤133Gal␤13 4Xyl␤13 Ser/Thr (28,29). Although the substrate specificities of these seven ␤4GalTs have been extensively studied, it remained unclear which enzymes can act on 6SGN residues in KS, N-linked, and O-linked glycans, and whether or not 6SGNspecific ␤4GalT is encoded by a novel gene. This issue is important for studying the biological roles of 6-sulfo-Nacetyllactosamine-containing glycans including KS and 6-sulfosialyl-Lewis X.

Materials
Purification of 6SGN-specific ␤4GalT from Porcine Colonic Mucosa-The following procedures were performed at 4°C. Six kg of porcine colon was purchased from Tokyo Shibaura Zohki Co. Ltd. (Tokyo, Ja-pan). The mucus layer was scraped off the colon; 3 liters of PBS was added, and the mixture was homogenized with a Potter-Elvehjem type homogenizer and then centrifuged at 1,000 ϫ g for 30 min. The extraction was performed once, and the two supernatant fractions were mixed and ultracentrifuged at 100,000 ϫ g for 1 h. The precipitated microsomes were washed once with 0.5 M KCl and extracted twice with 1 liter of 20 mM HEPES-NaOH (pH 7.2), 1 M NaCl, 10 mM MnCl 2 , 1% (w/v) Triton X-100, 1 mM dithiothreitol, 1 mM phenylmethanesulfonyl fluoride, and 10% (v/v) glycerol followed by ultracentrifugation. The extract was dialyzed against 10 mM HEPES-NaOH (pH 7.2), 5 mM MnCl 2 , 1 mM dithiothreitol, and 20% glycerol (buffer A), and applied on a UDPhexanolamine-Sepharose column (7 mol/ml of gel; 2 ϫ 10 cm; equilibrated with 20 mM HEPES-NaOH (pH 7.2), 10 mM MnCl 2 , 0.1% (w/v) Triton X-100, 1 mM dithiothreitol, and 10% glycerol (buffer B)) (12). After washing with buffer B containing 0.15 M NaCl, the enzyme fractions were eluted with buffer B containing 1 M NaCl and dialyzed against buffer A. Next, the dialysate was reapplied on a UDP-hexanolamine-Sepharose column (1.4 ϫ 6.5 cm; equilibrated with buffer B) and eluted with buffer B containing 1 mM UDP. Each fraction was dialyzed against buffer A and assayed for ␤4GalT activity. The enzyme fractions were applied on an asialo-agalacto-ovomucin-Sepharose column (6.7 mg mucin/ml of gel; 1 ϫ 7.5 cm; equilibrated with buffer B) (12). The enzyme was eluted with a linear gradient of NaCl (0 -0.2 M) in buffer B and then dialyzed. The enzyme fractions were concentrated and used for biochemical analyses.
Expression of ␤4GalTs in COS-7 Cells-The plasmids (1 g) were transfected into COS-7 cells on 35-mm dishes using Lipofectin Reagent (Invitrogen) according to the manufacturer's instructions. After 48 h, the cells were washed twice with phosphate-buffered saline, scraped off the dishes in 10 mM HEPES-NaOH (pH 7.2) and 0.25 M sucrose, and homogenized. The homogenates were ultracentrifuged at 100,000 ϫ g for 1 h. The precipitated crude membranes were suspended in 20 mM HEPES-NaOH (pH 7.2) and kept at Ϫ80°C until use. Assay of ␤4GalT Activity-Twenty l of reaction mixture consisting of 50 mM HEPES-NaOH (pH 7.2), 10 mM MnCl 2 , 0.5% (w/v) Triton X-100, 250 M UDP-Gal, 0.3 M UDP-[ 3 H]Gal (4.9 ϫ 10 5 dpm), 0.5 mM acceptor substrate, and the membrane fraction approximately diluted was incubated at 37°C for 1 h. In the cases of purified porcine 6SGNspecific ␤4GalT, 0.5%(w/v) Triton X-100 was replaced by 0.001% (w/v) D-sphingosine and 0.05% (w/v) Triton X-100. The 3 H-labeled products derived from biGP, 6SGN, GL, Xyl-O-pNP, core2-O-pNP, 6S-core2-O-pNP, 6S-biGP, and KS were purified by paper electrophoresis (pyridine/ acetic acid/water, 3:1:387, pH 5.4). The reaction mixtures in the case of agL2L2, agL2L4, and agLST-b were treated with 1 ml of 0.01 N HCl at 100°C for 10 min, neutralized with 1 N NaOH, concentrated, and subjected to paper electrophoresis. The acid hydrolysis was necessary for the destruction of excess UDP-[ 3 H]Gal, which moves to a similar position as 3 H-galactosylated agL2L2 and agL2L4 and for removal of sialic acid residue in the case of agLST-b. After drying, in the cases of neutral substrates, the paper was further developed with a solvent system, pyridine/ethyl acetate/acetic acid/water (5:5:1:3). The paper was monitored for radioactivity with a radiochromatogram scanner, and the 3 H-labeled products were extracted with water and applied on a RCA-I-agarose column (0.7 ϫ 2.5 cm). Elution was started with 4 ml of 10 mM sodium phosphate (pH 7.0), 0.15 M NaCl, followed by the same buffer containing 10 mM lactose. The bound fractions (eluted with 10 mM lactose) or retarded fractions were collected and measured for radioactivity. In the case of GlcCer, the reaction mixtures were directly applied on a high performance TLC silica gel plate (Kieselgel 60 F 254 ) (Merck) and developed with a solvent system (CHCl 3 /MeOH/0.2% CaCl 2 , 60:35:7). After drying, the plates were monitored for radioactivity, and the 3 H-labeled products were extracted with CHCl 3 /MeOH (1:1) and counted. The membrane fraction derived from pcDNA3-transfected COS-7 cells (C-MF) was used as a control for intrinsic ␤4GalT activities. To calculate the amount of exogenous ␤4GalT activities, values of activities in C-MF were subtracted from those of apparent enzymatic activities obtained under the same conditions. Enzymatic activity values presented here were means of four independent experiments. Standard deviations were less than 5% in all cases. The concentration of KS was presented as moles of galactose.
Determination of Protein Concentrations-The protein concentrations were estimated using the Bio-Rad Protein Assay dye reagent with bovine serum albumin as a standard.
Amino Acid Sequence Analysis-Coomassie-stained bands in SDS-PAGE gels were excised and treated with 0.2 g of Achromobacter protease I (a gift from Dr. Masaki, Ibaraki University) (43) at 37°C for 12 h in 0.1 M Tris-HCl (pH 9.0) containing 0.1% SDS. Peptides gener-  Fig. 3 X in analyzed sequence indicates the amino acid (aa) residues that could not be determined. The underlined amino acid residues in corresponding sequence (human) indicate discrepancies from analyzed sequence (porcine).
a The concentrations of the substrates were 0.5 mM.
␤4GalT-IV Is Specific for GlcNAc 6-O-Sulfate ated were extracted from the gel and separated on columns of DEAE-5PW (2 ϫ 20 mm; Tosoh, Tokyo) and Mightysil RP-18 (2 ϫ 50 mm; Kanto Chemical, Tokyo) connected in series with a model 1100 (Hewlett-Packard) liquid chromatography system. Peptides were eluted at a flow rate of 0.1 ml/min using a linear gradient of 0 -60% solvent B, where solvents A and B were 0.09% (v/v) aqueous trifluoroacetic acid and 0.075% (v/v) trifluoroacetic acid in 80%(v/v) acetonitrile, respec-tively. Selected peptides were subjected to Edman degradation using a model 477A automated protein sequencer (Applied Biosystems, Inc.) connected on-line to a model 120A PTH analyzer (PerkinElmer Life Sciences) and to a Reflex matrix-assisted laser desorption ionization time of flight mass spectrometer (Bruker-Franzen Analytik, Bremen, Germany) in linear mode using 2-mercaptobenzothiazole (44) as a matrix.
Northern Blot Analysis-Human Multiple Tissue Northern blot membranes (Clontech, Palo Alto, CA) were used according to the manufacturer's instructions. The mRNA content in each lane of the Northern blot membrane was normalized to the mRNA expression level of ␤-actin. 32 P-Labeled probe was prepared from the cDNA fragment (excised from pcDNA3-␤4GalT-IV by BamHI and EcoRV digestion) with a Random Primed DNA Labeling Kit (Roche Diagnostics) using [␣-32 P]dCTP (PerkinElmer Life Sciences) according to the manufacturer's instructions. The membranes were pre-hybridized in ExpressHyb Solution (Clontech) at 68°C for 2 h and then hybridized with the 32 P-labeled probe in the same solution at 68°C for 16 h. The Northern blot membranes were washed in 2ϫ SSC, 0.05% SDS at room temperature and then in 0.1ϫ SSC, 0.1% SDS at 50°C. The radioactivity was detected with FLA-2000 (Fuji Photo Film Co. Ltd., Tokyo).

Purification of 6SGN-specific ␤4GalT from Porcine Colonic Mucosa and Its Basic Lipid Requirement for Activity-We
showed previously (12) that a 6SGN-specific ␤4GalT exists in human colonic mucosa. To purify this enzyme and to isolate a corresponding cDNA based on partial amino acid sequences, we used porcine colonic mucosa as an enzyme source. Porcine colonic mucosa also contained 6SGN-specific ␤4GalT activity. When the enzyme was purified 2,300-fold by a second UDPhexanolamine-Sepharose column chromatography (Table I), the enzymatic activity disappeared (Fig. 1A). Because such a loss of enzymatic activity has been reported for a glucuronyltransferase, which requires sphingomyelin for activity in a highly purified state (45), we added various lipid compounds to the enzyme reaction mixtures. As shown in Table II, most of the phospholipids and ceramides had no ability to restore the enzymatic activity, whereas the activity appeared in the presence of D-sphingosine and N,N-dimethylsphingosine. Stearylamine could also restore the activity, whereas octylamine could not, suggesting that basic, lipidous compounds with long hydrophobic chains are essential at least for the in vitro enzymatic activity. The minimum concentrations of Dsphingosine, N,N-dimethylsphingosine, and stearylamine giving the maximal activity were ϳ10, 20, and 10 M, respectively (Fig. 2).
After asialo-agalacto-ovomucin-Sepharose chromatography, 6SGN-specific ␤4GalT was purified 24,000-fold ( Fig. 1B and Table I). We tried further purification by various chromatographic methods but were unsuccessful. The final enzyme frac-  ␤4GalT-IV Is Specific for GlcNAc 6-O-Sulfate tions contained two major (45kDa and 42kDa) and two minor (59kDa and 52kDa) proteins as determined by SDS-PAGE analysis (Fig. 3). Each protein band was digested with a lysinespecific protease, and the peptide fragments were separated by reversed phase high pressure liquid chromatography, and their amino acid sequence was analyzed. One or two peptide fragments from each band were sequenced (Table III), and the sequences were compared with known protein sequences using the BLAST search system. The band a contained one sequence that is equal to that of human polypeptide:N-acetylgalactosaminyltransferase-2 (46). The band b contained two sequences that are very similar to that of human N-acetylglucosaminyltransferase-I (47,48). Because the two enzymes utilize UDP-sugars as donor substrates, they may be co-purified with 6SGN-specific ␤4GalT by UDP-hexanolamine-Sepharose column chromatographies. On the other hand, the bands c and d contained a common peptide sequence, which is similar to that of human ␤4GalT-IV (24,26). This result suggested that the purified 6SGN-specific ␤4GalT corresponds to ␤4GalT-IV. Expression of Seven ␤4GalTs in COS-7 Cells and Their Substrate Specificities-To confirm that ␤4GalT-IV is 6SGN-specific, and the only 6SGN-specific ␤4GalT among the seven ␤4GalTs so far identified, we prepared expression vectors containing each ␤4GalT cDNA and analyzed substrate specificities using the membrane fractions derived from vector-transfected COS-7 cells. Seven ␤4GalT genes have been identified within the data bases provided by the human genome project to date (18 -29). The membrane fraction derived from vacant pcDNA3-transfected COS-7 cells (C-MF) was used as a control for intrinsic ␤4GalT activities. C-MF had weak ␤4GalT activity; the specific activities using biGP, 6SGN, agL2L2, Xyl-O-pNP, GL, core2-O-pNP, GlcCer, and 6S-core2-O-pNP as acceptor substrates were 0.20, 0.34, 0.43, 0.20, 0.12, 0.47, 0, and 0.16 nmol/min/mg of protein, respectively. To calculate the amount of exogenous ␤4GalT activities, values derived from C-MF were subtracted from those of the apparent enzymatic activities obtained under the same conditions.

H]Gal in [ 3 H]Gal␤13agL2L2 and ([ 3 H]Gal␤13
)6S-core2-O-pNP synthesized by ␤4GalT-IV was confirmed to be the C-4 of GlcNAc by RCA-I-agarose affinity chromatography (Fig. 4), which binds to Gal␤134GlcNAc (49). [ 3 H]Gal␤13 agL2L2 (Fig. 4B, solid line)  lactose (Fig. 4, B and C, dotted lines), but the digests flowed through the column by digestion with Gal␤134GlcNAc-specific diplococcal ␤-galactosidase (data not shown), showing that [ 3 H]Gal is attached to the C-4 of GlcNAc in the sulfated products. 6SGN was a poor substrate at best for the other ␤4GalTs (Table IV) even at the lower concentrations, 0.1 or 0.2 mM (data not shown), excluding the possibility that the absence or low level of activity was due to the inhibitory effect of a high concentration of 6SGN; such as has been found in several ␤4GalTs (34,51). These results indicated that ␤4GalT-IV is the only 6SGN-specific enzyme among the seven ␤4GalTs, consistent with the amino acid sequence of the purified porcine 6SGNspecific ␤4GalT as described above.
␤4GalT-III prefers O-linked type core2-O-pNP to N-linked type biGP, whereas ␤4GalT-I and -II recognize both biGP and core2-O-pNP as good acceptors. It has been shown that ␤4GalT-V acts on core2 (35) and a specific branching structure of N-linked glycans (36). As shown in Table IV, ␤4GalT-V could act on GlcCer, similar to ␤4GalT-VI. Recently, Lee et al. (52) reported that ␤4GalT-VI-deficient CHO cells have an ability to synthesize lactosylceramides, and they suggested the existence of a lactosylceramide synthase other than ␤4GalT-VI. ␤4GalT-V may be the second lactosylceramide synthase.
␤4GalT-IV recognizes several 6SGN-containing oligosaccharides as good acceptors (Table V). KS-related oligosaccharides, agL2L2 and agL2L4, were good substrates for ␤4GalT-IV. AgL2L2 was a poor substrate at best for the other ␤4GalTs (Table IV). KS is also a substrate for ␤4GalT-IV with a specific activity of 0.045 nmol/min/mg of protein at 1 mM KS. To expose GlcNAc residues at the non-reducing termini of KS chains, we treated KS by mild acid hydrolysis to remove sialic acid and Streptococcus 6646K ␤-galactosidase digestion. The enzymatic activity for the asialo-agalacto-KS was 3.3-fold higher than that for intact KS. This result suggests that ␤4GalT-IV can also act on KS long chains. Because agLST-b was a poor substrate for ␤4GalT-IV, sialic acid cannot replace sulfate at the C-6 of GlcNAc. This substrate selectivity is the same as for human colonic 6SGN-specific ␤4GalT (12). ␤4GalT-IV also efficiently acts on 6S-biGP and 6S-core2-O-pNP; the 6SGN moiety is present in N-linked and O-linked glycans in various glycoproteins (1). The 6-sulfosialyl-Lewis X on core2 glycans, which is synthesized from 6S-core2 by ␤134-galactosylation, ␣233-sialylation, and ␣133-fucosylation, functions as an L-selectin ligand moiety in the early step of lymphocyte homing in lymph nodes (14 -17). ␤4GalT-IV is the only enzyme recognizing 6S-core2-O-pNP as a good substrate among the seven ␤4GalTs (Table IV), suggesting that ␤4GalT-IV is involved in the synthesis of 6-sulfosialyl-Lewis X.
Results of kinetic analysis of ␤4GalT-IV for several glycans with or without 6-O-sulfation at GlcNAc are summarized in Table VI. AgL2L4 and 6S-core2-O-pNP had an inhibitory effect on ␤4GalT-IV activities at high concentrations (Fig. 5), so that kinetic constants were calculated using only the data obtained with lower concentrations of these substrates. The K m value for 6SGN was lower than that for GlcNAc, and the V max /K m value for 6SGN was 1900-fold higher than that for GlcNAc. Similarly, ␤4GalT-IV Is Specific for GlcNAc 6-O-Sulfate the V max /K m values for 6S-biGP and 6S-core2-O-pNP were much higher than those for biGP and core2-O-pNP, respectively. These results indicate that 6-O-sulfation of GlcNAc residues is important for efficient catalytic activity of ␤4GalT-IV.
Expression of ␤4GalT-IV in Various Human Tissues-Lo et al. (24) and Schwientek et al. (26) reported the expression profile of ␤4GalT-IV in human adult and fetal tissues, but they had not examined its expression in colon or lymph node, where 6SGN-specific ␤4GalT should be expressed. To clarify the expression pattern more extensively, we analyzed it using commercial Northern blot membranes. As shown in Fig. 6, a 2.4-kb transcript of ␤4GalT-IV was expressed ubiquitously. Relatively high expression levels of the enzyme were observed in kidney, placenta, lymph node, prostate, stomach, thyroid, tongue, and trachea. Recently, it has been shown that the 6-sulfosialyl-Lewis X determinant is present in colonic mucosa (53); in colon, ␤4GalT-IV was also moderately expressed (Fig. 7).
General Discussion-In this study, we showed the following. (i) Porcine 6SGN-specific ␤4GalT was purified and identified as ␤4GalT-IV. (ii) The substrate specificities of the seven ␤4GalTs so far cloned were investigated, and ␤4GalT-IV is the only 6SGN-specific ␤4GalT among them and is involved in the synthesis of several 6-O-sulfated N-acetyllactosamine glycans. (iii) Porcine ␤4GalT-IV requires basic lipidous compounds for the catalytic activity at least in a highly purified state. It has been reported that sphingosine increases several times the enzymatic activities of ␣133-fucosyltransferase, GM2:␤133GalT, GM3:␤134-N-acetylgalactosaminyltransferase (54), and glycoprotein sulfotransferase (55) and decreases that of ␤133-glucuronyltransferase (56). However, to our knowledge, this is the first demonstration that basic lipidous compounds are essential for the catalytic activity of glycosyltransferases. Sphingosine is an intermediate of the metabolism of sphingolipids and a precursor for sphingosine 1-phosphate, a second messenger molecule (reviewed in Refs. 57-59); however, it has been believed that sphingosine is present at quite low levels in living cells. The molecular aspect of the requirement is unclear, but one explanation is that in the process of purification, a certain basic compound, which binds to the enzyme in the crude extract, is detached from the enzyme protein. We are in the process of resolving this issue by analyzing biochemically the binding of recombinant ␤4GalT-IV to various basic compounds.
KS is poly-6-O-sulfated poly-N-acetyllactosamine (reviewed in Ref. 13). Several glyco-and sulfotransferases involved in KS synthesis have been reported. Five GlcNAc6ST genes have so far been identified, and all the enzymes can act on non-reducing terminal GlcNAc but not on internal GlcNAc such as Gal␤134GlcNAc. Recently, Akama et al. (60) showed that GlcNAc6ST-5 (CGn6ST) is involved in KS synthesis. Fukuta et al. (61) identified a Gal6ST, which prefers an internal Gal residue in the di-N-acetyllactosamine chain (62), suggesting that 6-O-sulfation of Gal is a later step in KS synthesis. Among ␤3GnTs so far cloned, ␤3GnT-2 has strong enzymatic activity for p-lacto-N-neohexaose (63), suggesting that the enzyme may be a candidate involved in the elongation of the KS backbone structure, although there is no evidence that ␤3GnT-2 has ␤3GnT activity for KS elongation. In this study, we showed that among the seven ␤4GalTs, only ␤4GalT-IV acts on KSrelated glycans, agL2L2 and agL2L4. From these results, it is proposed that KS is synthesized by the sequential actions of four enzymes, GlcNAc6ST-5, ␤4GalT-IV, ␤3GnT, and then Gal6ST (Fig. 7A).
On the other hand, 6-sulfosialyl-Lewis X is synthesized by sequential reactions of GlcNAc6ST, ␤4GalT, ␣2,3-sialyltransferase, and ␣1,3-fucosyltransferase. It has been shown that in lymph nodes, this structure is attached on ␤1,6-branching GlcNAc in O-linked core2 (14) or on elongated core1 (64) and that the GlcNAc 6-O-sulfation is performed by GlcNAc6ST-2. However, Uchimura et al. (10) showed that GlcNAc6ST-1 is responsible for the synthesis of 6-sulfosialyl-Lewis X. Next, galactosylation may be catalyzed by ␤4GalT-IV as indicated from our results that only ␤4GalT-IV efficiently acts on 6S-core2-O-pNP among the seven ␤4GalTs. Ujita et al. (34) reported that among ␤4GalT-I to -V, ␤4GalT-IV efficiently acts on non-sulfated core2-O-pNP. Here we showed that 6S-core2-O-pNP is a better substrate for ␤4GalT-IV than core2-O-pNP. From these results, a biosynthetic scheme for 6-sulfosialyl-Lewis X is proposed as shown in Fig. 7B. ␤4GalT-IV is extensively expressed in human tissues as shown in Fig. 6, and there seems to be no relationship between expression profiles of 6-sulfosialyl-Lewis X and ␤4GalT-IV. This indicates that ␤4GalT-IV is not a rate-limiting enzyme for biosynthesis of 6-sulfosialyl-Lewis X and that the preceding 6-O-sulfation of GlcNAc may be a rate-limiting step. The next subject of study should be whether 6-sulfosialyl-Lewis X is synthesized in cultured cells defective in ␤4GalT-IV expression.