Abstract

The biosynthesis of HNK-1 carbohydrate is mainly regulated by two glucuronyltransferases (GlcAT-P and GlcAT-S) and a sulfotransferase (HNK-1 ST). To determine how the two glucuronyltransferases are involved in the biosynthesis of the HNK-1 carbohydrate, we prepared soluble forms of GlcAT-P and GlcAT-S fused with the IgG-binding domain of protein A and then compared the enzymatic properties of the two enzymes. Both GlcAT-P and GlcAT-S transferred glucuronic acid (GlcA) not only to a glycoprotein acceptor, asialoorosomucoid (ASOR), but also to a glycolipid acceptor, paragloboside. The activity of GlcAT-P toward ASOR was enhanced fivefold in the presence of sphingomyelin, but there were no effects on that of GlcAT-S. The activities of the two enzymes toward paragloboside were only detected in the presence of phospholipids such as phosphatidylinositol. Kinetic analysis revealed that the Km value of GlcAT-P for ASOR was 10 times lower than that for paragloboside. Furthermore, acceptor specificity analysis involving various oligosaccarides revealed that GlcAT-P specifically recognized N-acetyllactosamine (Galβ1-4GlcNAc) at the nonreducing terminals of acceptor substrates. In contrast, GlcAT-S recognized not only the terminal Galβ1-4GlcNAc structure but also the Galβ1-3GlcNAc structure and showed the highest activity toward triantennary N-linked oligosaccharides. GlcAT-P transferred GlcA to NCAM about twice as much as to ASOR, whereas GlcAT-S did not show any activity toward NCAM. These lines of evidence indicate that these two enzymes have significantly different acceptor specificities, suggesting that they may synthesize functionally and structurally different HNK-1 carbohydrates in the nervous system.

Introduction

Some of the carbohydrates on the cell surface are indicated to be associated with cell–cell recognition and adhesion (Jessell et al., 1990; Rutishauser et al., 1988). The HNK-1 carbohydrate is characteristically expressed on a series of cell adhesion molecules, such as neural cell adhesion molecule (NCAM), L1, and P0, and also on some glycolipids in the nervous system (Ariga et al., 1987; Chou et al., 1986; Kruse et al., 1984; Yoshihara et al., 1991a). The HNK-1 carbohydrate is postulated to be associated with cell adhesion, migration, and neurite outgrowth (Bronner-Fraser, 1987; Kunemund et al., 1988; Martini et al., 1992). The expression of this carbohydrate is spatially and temporally regulated during the development of the nervous system (Schwarting et al., 1987; Yoshihara et al., 1991b). The characteristic structure of the HNK-1 carbohydrate epitope is the sulfated glucuronic acid (GlcA) attached to the N-acetyllactosamine structure, HSO3-3GlcA-Galβ1-4GlcNAc (Ariga et al., 1987; Chou et al., 1986; Voshol et al., 1996). Glucuronyltransferase(s) (GlcATs) and sulfotransferase(s) are supposed to be key enzymes for this carbohydrate biosynthesis, because the inner structure, Galβ1-4GlcNAc, is commonly present on various glycoproteins and glycolipids.

Recently we purified and cloned a GlcAT (GlcAT-P) from rat brain, which is involved in the biosynthesis of the HNK-1 carbohydrate (Terayama et al., 1997, 1998). Using rat GlcAT-P cDNA, we and others have cloned a second GlcAT (GlcAT-S) (Seiki et al., 1999; Shimoda et al., 1999). The transfection of GlcAT-P or GlcAT-S cDNA into COS-1 cells resulted in the expression of the HNK-1 carbohydrate in glycoproteins on the cell surface. Characterization of GlcAT-P purified from rat brain revealed that the expression of the GlcAT activity toward glycoprotein acceptors, for example, asialo-orosomucoid (ASOR), depends almost completely on sphingomyelin (SM). Under these assay conditions (i.e., in the absence of phosphatidylinositol [PI]), however, there was no activity toward a glycolipid acceptor, palagloboside (Terayama et al., 1998). GlcAT-P specifically recognized the N-acetyllactosamine structure at the nonreducing terminals of glycoprotein acceptors (Oka et al., 2000; Terayama et al., 1997, 1998).

More recently, we succeeded in generating mice with targeted deletion of the GlcAT-P gene (Yamamoto et al., 2002). To our surprise, the GlcAT activity toward paragloboside disappeared almost completely, as well as the activity toward ASOR (Yamamoto et al., 2002). Thus it would be interesting to determine whether GlcAT-P really exhibits activity toward glycolipid acceptors. Furthermore, the HNK-1 carbohydrate disappeared almost completely in GlcAT-P-deficient mice, but a trace of HNK-1 immunoreactivity remained on the surfaces of the soma and proximal dendrites of a subset of neurons in some limited regions. These remaining HNK-1 carbohydrates in GlcAT-P deficient mice are localized predominantly in the perineuronal nets and are assumed to be synthesized by GlcAT-S (Yamamoto et al., 2002). These results suggest the possibility that the two glucuronyltransferases synthesize structurally and functionally different HNK-1 carbohydrates in vivo.

In this study, we prepared protein A fusion recombinant GlcAT-P and GlcA-S in COS-1 cells, and then analyzed the acceptor specificities of the two glucuronyltransferases. GlcAT-P was shown to exhibit activity not only toward a glycoprotein acceptor, ASOR, but also toward a glycolipid acceptor, paragloboside, the latter activity requiring the presence of phospholipids such as PI.

Results

Preparation of soluble forms of GlcAT-P and GlcAT-S

To study the acceptor specificities of GlcAT-P and GlcAT-S in more detail, soluble forms of these enzymes were generated by fusing the catalytic domain and putative stem region of each enzyme to a secreted form of the protein A IgG-binding domain. Each fused protein was expressed in COS-1 cells and then specifically collected with IgG-Sepharose 4B beads from the culture medium. As shown in Figure 1, each enzyme bound to IgG-Sepharose 4B beads gave a single band corresponding to the expected molecular size (87 kDa), when visualized on western blotting with horseradish peroxidase (HRP)-conjugated rabbit IgG (lane 1, protA-GlcAT-P; lane 2, protA-GlcAT-S). The amounts of purified protA-GlcAT-P and protA-GlcAT-S were estimated by means of densitometric analysis of each band using bovine serum albumin as a standard after staining with Coomassie brilliant blue R 250. Equivalent amounts of protA-GlcAT-P and protA-GlcAT-S were used for the following enzyme assays.

Fig. 1.

Expression of soluble forms of GlcAT-P and GlcAT-S. Soluble protein A fusion enzymes were subjected to immunoblot analysis with HRP-conjugated rabbit IgG. Lane 1, protA-GlcA-P; lane 2, protA-GlcAT-S.

Fig. 1.

Expression of soluble forms of GlcAT-P and GlcAT-S. Soluble protein A fusion enzymes were subjected to immunoblot analysis with HRP-conjugated rabbit IgG. Lane 1, protA-GlcA-P; lane 2, protA-GlcAT-S.

Effects of phospholipids on the GlcAT activities of protA-GlcAT-P and protA-GlcAT-S

The native GlcAT-P purified from rat brain transferred GlcA to various glycoprotein acceptors in the presence of SM, but we could not detect any activity toward glycolipid acceptor under the assay conditions used (Terayama et al., 1998). In GlcAT-P-deficient mice, however, the GlcAT activity toward paragloboside disappeared almost completely as well as the activity toward the glycoprotein acceptor (ASOR) (Yamamoto et al., 2002). As shown in Table I, significant activity of protA-GlcAT-P toward a glycoprotein acceptor was detected even in the absence of SM, but the activity toward paragloboside was not detected under the conditions used. In our previous studies, phospholipids were shown to be required for the maximal GlcAT activity. SM is the essential activator for GlcAT-P purified from rat brain toward glycoprotein acceptors (Terayama et al., 1998), and PI enhances the activity of a rat brain extract toward glycolipid acceptors (Kawashima et al., 1992). On the basis of these observations, we examined the effects of phospholipids on the activity of protA-GlcAT-P. The activity toward ASOR was enhanced up to fivefold in the presence of SM; this was consistent with the previous observation on the native enzyme. However, in contrast to our previous observation, the activity toward paragloboside was also detected in the presence of PI. These results indicate that GlcAT-P exhibits activity not only toward glycoprotein acceptors but also toward glycolipid acceptors under certain conditions. On the other hand, protA-GlcAT-S exhibited almost the same level of activity as that of protA-GlcAT-P toward a glycoprotein acceptor in the absence of phospholipids, but the activity was not enhanced by SM. The specific activity of protA-GlcAT-S toward paragloboside was much (more than 40 times) lower than that of protA-GlcAT-P even in the presence of PI.

Table I.

Glucuronyltransferase activities (µmol/h/mg protein)

Substrate
 
Lipid added
 
GlcAT-P
 
GlcAT-S
 
Glycoprotein None 3.59 2.08 
 SM (1 µg/ml) 18.7 2.39 
Glycolipid None <0.005 <0.005 
 PI (1 µg/µl) 0.42 0.01 
Substrate
 
Lipid added
 
GlcAT-P
 
GlcAT-S
 
Glycoprotein None 3.59 2.08 
 SM (1 µg/ml) 18.7 2.39 
Glycolipid None <0.005 <0.005 
 PI (1 µg/µl) 0.42 0.01 

The GlcAT activities of soluble recombinant enzymes were measured using ASOR as a glycoprotein acceptor or paragloboside as a glycolipid acceptor.

In the next experiment, we examined the dose-dependent effects of phospholipids on the activity of protA-GlcAT-P. As shown in Figure 2, SM and PI enhanced the activities toward ASOR and paragloboside in a saturated manner, respectively. Maximal activities were detected in the presence of 0.6 µg/µl SM and 1 µg/µl PI. Then we examined the effects of various membrane phospholipids on the activity of protA-GlcAT-P. As shown in Table IIA, SM was the most potent activator for the activity toward ASOR, and phosphatidylcholine and phosphatidylethanolamine had only a little effect (9%) on the activity, and phosphatidylserine and PI had no effect at all on the activity. As shown in the Table IIB, for the activity toward paragloboside, PI is the most potent activator, and phosphatidylserine had about a half-effect (55%), but other phospholipids examined had no effect at all on the activity.

Table II.

Effects of various lipids on the activity of GlcAT-P

A
 
 B
 
 
Lipid (1.0 µg/µl)
 
% of activation
 
Lipid (1.0 µg/µl)
 
% of activation
 
SM 100 PI 100 
Phosphatidylcholine 8.8 Phosphatidylserine 55 
Phosphatidylethanolamine 8.8 Phosphatidylethanolamine <0.1 
Phosphatidylserine <0.1 Phosphatidylcholine <0.1 
Phosphatidylinositol <0.1 Sphingomyelin <0.1 
A
 
 B
 
 
Lipid (1.0 µg/µl)
 
% of activation
 
Lipid (1.0 µg/µl)
 
% of activation
 
SM 100 PI 100 
Phosphatidylcholine 8.8 Phosphatidylserine 55 
Phosphatidylethanolamine 8.8 Phosphatidylethanolamine <0.1 
Phosphatidylserine <0.1 Phosphatidylcholine <0.1 
Phosphatidylinositol <0.1 Sphingomyelin <0.1 

The GlcAT activity toward ASOR (A) or paragloboside (B) was measured in the presence of various lipids. The activation was compared with the control value in the presence of 1.0 µg/µl SM or 1.0 µg/µl PI.

Fig. 2.

Dependence of GlcAT-P activity on the amounts of phospholipids. The GlcAT activities of GlcAT-P toward ASOR (A) and paragloboside (B) were measured in the presence of different concentrations of stearoyl-SM (18:0) (A) and PI (B), respectively. The activity is expressed as a percentage of the control value in the presence of a saturating amount (1 µg/µl) of stearoyl-SM or PI.

Fig. 2.

Dependence of GlcAT-P activity on the amounts of phospholipids. The GlcAT activities of GlcAT-P toward ASOR (A) and paragloboside (B) were measured in the presence of different concentrations of stearoyl-SM (18:0) (A) and PI (B), respectively. The activity is expressed as a percentage of the control value in the presence of a saturating amount (1 µg/µl) of stearoyl-SM or PI.

Kinetics analysis of protA-GlcAT-P

As described, GlcAT-P transferred GlcA not only to a glycoprotein acceptor but also to a glycolipid acceptor. To compare the kinetic properties of these acceptors, we examined the dependence of the rate of the GlcAT reaction on the concentration of ASOR, paragloboside, and UDP-GlcA. The results are shown in Figure 3, and the data were analyzed by means of Lineweaver-Burk plotting (insets). The Vmax values for ASOR, paragloboside, and UDP-GlcA were 47 µmol/h/mg, 2.5 µmol/h/mg, and 18 µmol/h/mg, respectively. The Km values for ASOR, paragloboside, and UDP-GlcA were 4.3 µM, 47 µM, and 65 µM, respectively (Figure 3). The Km values of protA-GlcAT-P for ASOR and UDP-GlcA were comparable to those of the native GlcAT-P purified from rat brain (Km for ASOR: 1.9 µM and Km for UDP-GlcA: 22 µM) (Terayama et al., 1998). The Km value for ASOR was 10 times lower than that for paragloboside, indicating that the affinity of GlcAT-P to ASOR is 10 times higher than that to paragloboside.

Fig. 3.

Dependence of GlcAT-P activity on the concentrations of substrates. Different concentrations of ASOR (A), paragloboside (B), and UDP-GlcA (C) were added to the standard assay mixture. The insets show the data analyzed by means of Lineweaver-Burk plots.

Fig. 3.

Dependence of GlcAT-P activity on the concentrations of substrates. Different concentrations of ASOR (A), paragloboside (B), and UDP-GlcA (C) were added to the standard assay mixture. The insets show the data analyzed by means of Lineweaver-Burk plots.

Acceptor specificities of GlcAT-P and GlcAT-S toward oligosaccharides

To determine why two GlcATs are involved in the biosynthesis of the HNK-1 carbohydrate, we compared the GlcAT activities of these enzymes toward oligosaccharides under the assay conditions used for glycoproteins. With regard to activity toward disaccharides or trisaccharides (at a concentration of 200 µM), which are commonly present at the nonreducing terminals of carbohydrate chains, protA-GlcAT-P showed the highest activity toward N-acetyllactosamine (Galβ1-4GlcNAc) and a little activity (4% activity compared toward N-acetyllactosamine) toward lactose (Galβ1-4Glc). However, no activity was detected toward either N-lacto-biose (Galβ1-3GlcNAc) or Galβ1-4Gal (Table III). These results were in good agreement with the specificity of the native GlcAT-P purified from rat brain (Oka et al., 2000; Terayama et al., 1998). In contrast, protA-GlcAT-S showed activity not only toward N-acetyllactosamine but also toward N-lacto-biose (25% activity compared with N-acetyllactosamine), lactose (29%), and Galβ1-4Gal (16%). It should be noted that both enzymes showed little activity toward Lewis x or Lewis y trisaccharides, although they have the N-acetyllactosamine structure, indicating that a fucose on GlcNAc inhibited the activities of GlcAT-P and GlcAT-S to the N-acetyllactosamine structure.

Table III.

Acceptor specificity (nmol/h/mg protein) of GlcATs toward oligosaccharides

Acceptors
 
GlcAT-P
 
GlcAT-S
 
N-acetyllactosamine (20 µM) 545 <5 
N-acetyllactosamine (200 µM) 3880 150 
Lacto-N-biose (200 µM) <5 38 
Lactose (200 µM) 159 44 
Galβ1,4Gal (200 µM) <5 24 
Lewis x (200 µM) 16 <5 
Lewis y (200 µM) 13 <5 
Lacto-N-neotetraose (20 µM) 846 32 
Lacto-N-tetraose (20 µM) 19 25 
Lacto-N-hexaose (20 µM) 5170 80 
Complex-type N-linked oligosaccharides (20 µM)   
Biantennary 3480 49 
Triantennary 4640 691 
Tetraantennary 6600 168 
Biantennary with bisecting GlcNAc 2230 60 
Biantennary, core-substituted with fucose 3070 50 
Acceptors
 
GlcAT-P
 
GlcAT-S
 
N-acetyllactosamine (20 µM) 545 <5 
N-acetyllactosamine (200 µM) 3880 150 
Lacto-N-biose (200 µM) <5 38 
Lactose (200 µM) 159 44 
Galβ1,4Gal (200 µM) <5 24 
Lewis x (200 µM) 16 <5 
Lewis y (200 µM) 13 <5 
Lacto-N-neotetraose (20 µM) 846 32 
Lacto-N-tetraose (20 µM) 19 25 
Lacto-N-hexaose (20 µM) 5170 80 
Complex-type N-linked oligosaccharides (20 µM)   
Biantennary 3480 49 
Triantennary 4640 691 
Tetraantennary 6600 168 
Biantennary with bisecting GlcNAc 2230 60 
Biantennary, core-substituted with fucose 3070 50 

N-acetyllactosamine, Galβ1,4GlcNAc; lacto-N-biose, Galβ1,3GlcNAc; lactose, Galβ1,4Glc; Lewis x, Galβ1,4(Fucα1,3)GlcNAc; Lewis y, Fucα1,2 Gaβ1,4(Fucα1,3)GlcNAc; lacto-N-neotetraose, Galβ1,4GlcNAcβ1,3 Galβ1,4Glc; lacto-N-tetraose, Galβ1,3GlcNAcβ1,3 Galβ1,4Glc; lacto-N-hexaose, Galβ1,4GlcNAcβ1,6(Galβ1,3GlcNAcβ1,3) Galβ1,4Glc.

Next we measured the activities toward various oligosaccharide chains expressed in protA-glycolipids or glycoproteins (at a concentration of 20 µM because of the limited availability of these substrates). ProtA-GlcAT-P exhibited very high activity toward these oligosaccharide chains except lacto-N-tetraose, which has a nonreducing terminal structure of Galβ1-3GlcNAc. As for N-linked oligosaccharides, the activity of GlcAT-P increased in proportion to the number of acceptor sugar branches. A bisecting GlcNAc or Fuc linked to the innermost GlcNAc residue has little effect on the GlcAT activity of either enzyme. In contrast, protA-GlcAT-S showed the highest activity toward triantennary N-linked oligosaccharides followed by tetraantennary N-linked oligosaccharides (Table III). These lines of evidence indicated that GlcAT-P strictly recognizes the N-acetyllactosamine structure at the nonreducing terminals of acceptor substrates. On the other hand, GlcAT-S might recognize the overall oligosaccharide sugar chains rather than the nonreducing terminal carbohydrate structure.

The specificity of the activity toward NCAM

Throughout this study, we used ASOR, which is a serum glycoprotein, as a model acceptor substrate to measure the activity toward glycoproteins. To determine the effect of the polypeptide portion of the acceptor substrates, we used rat NCAM, which bears the HNK-1 carbohydrate in vivo. The soluble form of rat NCAM was expressed as an Fc-fusion protein in Lec2 cells, and purified as described under Materials and methods. Lec2 cells are mutant Chinese hamster ovary cells. Glycoproteins and glycolipids produced by these cells lack sialic acid (Stanley and Siminovitch, 1977). As shown in Figure 4, NCAM-Fc prepared from Lec2 cells was detected with RCA120 lectin, which recognizes terminal galactose residues, indicating that the purified NCAM-Fc has terminal galactose residues that are used by GlcAT-P and GlcAT-S. ASOR and NCAM have different numbers of N-linked sugar chains on their peptide portions.

Fig. 4.

Analysis of NCAM-Fc expressed in Lec2 cells. NCAM-Fc expressed in Lec2 cells was analyzed by immunoblotting with anti-NCAM antibodies (lane 1) and by lectin blotting with RCA 120 (lane 2).

Fig. 4.

Analysis of NCAM-Fc expressed in Lec2 cells. NCAM-Fc expressed in Lec2 cells was analyzed by immunoblotting with anti-NCAM antibodies (lane 1) and by lectin blotting with RCA 120 (lane 2).

We determined the numbers of amino sugars present on ASOR and NCAM-Fc as described under Materials and methods. There were 22 and 28 amino sugars per molecule of ASOR and NCAM-Fc, respectively. To compensate for the number of N-acetyllactosamine units of acceptor substrates, ASOR or NCAM-Fc corresponding to 0.5 nmol of amino sugars was used as a substrate, and the results are shown in Table IV. GlcAT-P showed about twice as much activity toward NCAM than toward ASOR, whereas GlcAT-S did not show any activity toward NCAM. GlcAT-P may recognize in part the polypeptide portion of acceptor substrates. These results suggest that the two GlcATs have different acceptor specificities not only as to the carbohydrate portion but also to the polypeptide portion of glycoprotein substrates.

Table IV.

Acceptor specificities (% of activity) of GlcATs toward asialo-glycoproteins

Acceptor
 
GlcAT-P
 
GlcAT-S
 
ASOR 100 100 
NCAM 203 <1.8 
Acceptor
 
GlcAT-P
 
GlcAT-S
 
ASOR 100 100 
NCAM 203 <1.8 

Asialo-glycoproteins, which equal in amount to that of amino sugar (0.5 nmol), were added to the reaction mixture, and then the GlcAT activities were measured.

Discussion

The HNK-1 carbohydrate is associated with cell adhesion, migration, and neurite outgrowth (Bronner-Fraser, 1987; Kunemund et al., 1988; Martini et al., 1992). GlcAT-P and GlcAT-S are postulated to be the key enzymes involved in the biosynthesis of the HNK-1 carbohydrate. In this study, we demonstrated that both protA-GlcAT-P and protA-GlcAT-S transferred GlcA not only to a glycoprotein acceptor, ASOR, but also to a glycolipid acceptor, paragloboside. We found that the specific activity of protA-GlcAT-P toward ASOR was enhanced fivefold in the presence of SM, whereas the activities of the two enzymes toward paragloboside completely depended on the presence of PI under our assay conditions. The requirement of PI for activity toward paragloboside was also in good agreement with the results for flag-GlcAT-P and flag-GlcAT-S, which were expressed in Escherichia coli (Kakuda et al., 2004a). The reason why the GlcAT-P purified from rat brain didn't show any activities toward glycolipid acceptor substrate is appeared to be no PI under our assay conditions (Terayama et al., 1998). As we have previously reported, the GlcAT activity toward a glycoprotein acceptor substrate of the GlcAT-P was apparently disappeared during the purification. The activity was recovered in the presence of SM or heat-inactivated NP-40 extract of brain membrane fraction. Then we measured the GlcAT activity of the GlcAT-P purified from rat brain toward a glycolipid acceptor substrate under these conditions (in the presence of SM or heat-inactivated NP-40 extract of brain membrane fraction).

Hydropathy analysis revealed that GlcAT-P and GlcAT-S are structurally similar to previously identified GlcATs (Paulson and Colley, 1989) in that they are type II transmembrane proteins comprising a short N-terminal cytoplasmic tail, a transmembrane domain, a stalk, and a large globular catalytic domain (Colley, 1997). In this study we only used the catalytic domain, a truncated form, of GlcAT-P and GlcAT-S. In the cellular membrane system, SM is located predominantly in the outer leaflet of the plasma membrane and the Golgi apparatus (Abe and Norton, 1974) and is presumed to interact with the catalytic domain of GlcAT-P. On the other hand, PI is mainly localized in the inner leaflet of the plasma membrane and the Golgi apparatus. A possible explanation may be that there are other similar lipids in the outer leaflet, as such as the GPI portion of GPI anchor proteins acts as PI to regulate the activities of GlcAT-P and GlcAT-S in vivo. Some phospholipids have been reported to enhance the activities of glycosyltransferases, such as hepatic GlcAT (Pukazhenthi et al., 1993), β1-4 galactosyltransferase (Yamaguchi and Fukuda, 1995), and α2-3-sialytransferase (Nilsson and Dallner, 1977). However, to our knowledge, this is the first report indicating that the activity of a glycosyltransferase is regulated by two membrane lipids, SM and PI, to switch its activity toward a glycoprotein acceptor substrate or a glycolipid acceptor substrate. These lines of evidence suggest that expression of the HNK-1 epitope can be regulated not only by the expression level of the enzyme protein but also by the microenvironment around the enzyme, especially by the presence of SM and PI.

It should be noted that both enzymes showed little activity toward Lewis x or Lewis y trisaccharides (Table III), although both of them have the N-acetyllactosamine structure, indicating that a fucose residue on GlcNAc inhibited the activities of GlcAT-P and GlcAT-S toward the N-acetyllactosamine structure. Andressen et al. (1998) examined the terminal carbohydrate sequences of the lactoseries on the developing chick olfactory receptor epithelium. They found HNK-1 epitope expression only in the immature olfactory receptors and Lewis x epitope expression only in the mature olfactory receptors, although N-acetyllactosamine was expressed on the whole epithelial cell population throughout differentiation steps. These lines of evidence suggest that two functional carbohydrates, HNK-1 and Lewis x, are alternatively expressed, and that the biosynthesis of the HNK-1 carbohydrate epitope may be regulated by the expression of the fucose residue on GlcNAc. The X-ray crystal structure of GlcAT-P, and the modeled structure of a complex of GlcAT-P and Lewis x also revealed that it was almost stereochemically impossible to gain access to the enzyme when a fucose residue was on GlcNAc (Kakuda et al., 2004b).

We examined the acceptor substrate specificity toward oligosaccharides. GlcAT-P exhibits high specificity toward type 2 (Galβ1-4GlcNAc) glycans. This finding was in good agreement with in the case of the native GlcAT-P purified from rat brain (Oka et al., 2000; Terayama et al., 1998). On the other hand, GlcAT-S transferred GlcA not only to type 2 but also to type 1 (Galβ1-3GlcNAc) glycans and showed the highest activity toward triantennary N-linked oligosaccharides, indicating that GlcAT-S seems to have a wide specificity range. This may indicate that GlcAT-P and GlcAT-S are involved in the biosynthesis of structurally and functionally different HNK-1 carbohydrates.

In situ hybridization on the rat embryo revealed that GlcAT-S transcripts were expressed in the pallidum and retina, in which GlcAT-P expression was not specifically observed (Shimoda et al., 1999). In rat postnatal day 8 cerebellum, GlcAT-S mRNA was expressed in the internal and external granule layers. On the other hand, the expression of GlcAT-P was mainly localized in the Purkinje cell layer (unpublished data). Furthermore, although the HNK-1 carbohydrate disappeared almost completely in GlcAT-P-deficient mice, a trace of HNK-1 immunoreactivity remained on the surfaces of the soma and proximal dendrites of a subset of neurons in the perineuronal nets (Yamamoto et al., 2002). These lines of evidence suggested that although these two GlcAT activities overlap in part, they are not functionally redundant in the brain. It is possible that there are structurally and functionally different HNK-1 carbohydrate epitopes in vivo. From this point of view, GlcAT-P-deficient mice constitute a useful tool for clarifying the structural and functional roles of the remaining HNK-1 carbohydrate epitope.

Materials and methods

Materials

UDP-[14C]GlcA was purchased from ICN Radiochemicals (Irvine, CA). Orosomucoid was provided by Dr. M. Wickerhauser of the American Red Cross Research Center (Bethesda, MD). ASOR was prepared by hydrolysis of orosomucoid in 0.05 M H2SO4 at 80°C for 1 h as described previously (Oka et al., 1992). Several PA-oligosaccharides were purchased from Takara Shuzo (Kyoto, Japan). Plasmids pEF-BOS and pGIR201protA were kindly provided by Dr. S. Nagata (Osaka University) and Dr. H. Kitagawa (Kobe Pharmaceutical University), respectively.

Preparation of expression constructs of soluble forms of GlcAT-P and GlcAT-S

For the construction of expression vectors to produce soluble forms of GlcAT-P and GlcAT-S fused with protein A, pGIR201protein A was digested with NheI. The NheI DNA fragment encoding the insulin signal sequence and IgG-binding domain of protein A was subcloned into a mammalian expression vector, pEF-BOS, which had been digested with XbaI. The resulting vector, named pEF-protA-BOS, was used for subcloning of the following truncated forms of GlcAT-P and GlcAT-S. A truncated form of GlcAT-P lacking the amino-terminal 35 amino acids was amplified by polymerase chain reaction (PCR) with rat GlcAT-P cDNA (Terayama et al., 1997) in pEF-BOS as a template using primers (5′-CTCGGATCCGTCTGGCACCAGAGCA-3′ and 5′-ATTGGATCCTGTGTAGTTTCAGATCTCCACCGA-3′) containing BamHI sites (underscored). A truncated form of GlcAT-S, lacking the amino-terminal 23 amino acids, was amplified by PCR with rat GlcAT-S cDNA (Seiki et al., 1999) in pEF-BOS as a template using primers (5′-ATCAGATCTGACGTGGACCCCCGCA-3′ and 5′-CGTAGATCTGCGGAGAAGCGGAACGA-3′) containing BglII sites (underscored). Each PCR product was digested with BamHI or BglII and then inserted into pEF-protA-BOS. The recombinant plasmids with the correct orientation, pEF-protA-GlcAT-P and pEF-protA-GlcAT-S, were used for the following experiments.

Preparation of soluble forms of GlcAT-P and GlcAT-S

COS-1 cells (5 × 106 cells) were transfected with pEF-protA-GlcAT-P or pEF-protA-GlcAT-S (10 µg each) using lipofectAMINE (Invitrogen) according to the manufacturer's instructions. After 24 h incubation, the culture medium was replaced with serum-free ASF104 medium (Ajinomoto, Tokyo, Japan), followed by incubation for another 4 days. The enzymes secreted into the medium were adsorbed to normal rabbit IgG Sepharose 4B (Amersham Bioscience, Little Chalfont, U.K.), which was then used as the enzyme sources.

Western blot analysis of soluble forms of GlcAT-P and GlcAT-S

The protA-GlcAT-P and protA-GlcAT-S purified with IgG Sepharose 4B were separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE) (5–20%) and then transferred to nitrocellulose membranes. The blots were then incubated with HRP-conjugated rabbit IgG (Zymet, San Francisco, CA) and visualized with a DAB substrate kit (Funakoshi, Tokyo, Japan).

GlcAT assay for glycoprotein acceptors

The GlcAT activity toward glycoprotein acceptors was measured essentially as described previously (Oka et al., 1992) with slight modification. An equivalent amount of each enzyme was incubated at 37°C for 3 h in a reaction mixture with a final volume of 50 µl comprising 200 mM MES, pH 6.5, 0. 2% NP-40, 20 mM MnCl2, 20 µg ASOR, 100 µM UDP-[14C]-GlcA (200,000 dpm), and 0.5 mM ATP. After incubation, the assay mixture was spotted onto a 2.5-cm Whatman No.1 disc. The disc was washed with a 10% (w/v) trichloroacetic acid solution three times, followed by with ethanol/ether (2:1, v/v) and then with ether. The disc was air-dried, and then the radioactivity of [14C]-GlcA-ASOR on it was counted with a liquid scintillation counter (Beckman LS-6000).

GlcAT assay for glycolipid

GlcAT activity toward a glycolipid acceptor was also measured essentially as described previously (Kawashima et al., 1992) with slight modification. An equivalent amount of each enzyme was incubated at 37°C for 3 h in a reaction mixture with a final volume of 50 µl comprising 80 mM sodium cacodylate buffer, pH 6.0, 0.4% NP-40, 10 mM MnCl2, 7.5 µg paragloboside, 100 µM UDP-[14C]-GlcA (200,000 dpm), and 10 mM ATP. The reaction was terminated by the addition of 1 ml chloroform/methanol (2:1, v/v). The radioactive reaction products were separated from labeled precursors by passage through Sephadex G-25 superfine column (1 ml), which had been equilibrated with chloroform/methanol/water (60:30:4.5, v/v/v), and then the radioactivity of [14C]-GlcA-paragloboside, eluted at the void volume of the column, was counted with a liquid scintillation counter (Beckman LS-6000).

GlcAT assay for oligosaccharides

GlcAT activity toward oligosaccharides was measured as follows: An equivalent amount of each enzyme was incubated at 37°C for 3 h in a reaction mixture with a final volume of 50 µl comprising different concentrations of oligosaccharides, 200 mM MES (pH 6.5), 0.2% NP-40, 20 mM MnCl2, 100 µM UDP-[14C]-GlcA (200,000 dpm), 0.5 mM ATP, and 2 µl heat-inactivated NonidetP-40 extract of rat brain (13). After incubation, the reaction was terminated by the addition of 1 ml 5 mM phosphate buffer, pH 6.8. The radioactive reaction products were separated from UDP-[14C]-GlcA by passage through anion exchange resin AG1-X4 column (1 ml) that had been equilibrated with 5 mM phosphate buffer, pH 6.8. The column was washed with 5 ml of 5 mM phosphate buffer, pH 6.8, and then the effluent and washings were collected. The radioactivity was counted with a liquid scintillation counter (Beckman LS-6000).

Preparation of soluble NCAM as a glycoprotein substrate

The expression construct of Fc-tagged rat NCAM (Kakuda et al., 2004a) was transfected into Lec2 cells using lipofectAMINE (Invitrogen) according to the manufacturer's instructions. After 24 h incubation, the culture medium was replaced with ASF104 medium (Ajinomoto), followed by incubation for another 5 days. The NCAM-Fc secreted into the medium was collected with a protein G Sepharose column. The adsorbed NCAM-Fc was eluted with 0.1 M citrate buffer, pH 2.2; neutralized immediately with 1 M Tris–HCl buffer, pH 9.0; and then concentrated before use as a glycoprotein substrate. The purify of the NCAM-Fc was checked by western blot analysis after SDS–PAGE under reducing conditions. The membrane was incubated with 5 µg/ml of biotin-conjugated ricinus communis agglutinin or rabbit anti-human Fc antibodies. After incubation with HRP-conjugated avidin or anti-rabbit IgG, the membrane was visualized with a DAB substrate kit (Funakoshi).

Amino acid analysis

NCAM-Fc and ASOR were hydrolyzed in 6 N HCl at 100°C for 16 h, and then the amino acid and amino sugar contents were determined with an amino analyzer (Hitachi L8500).

This work was supported in part by a Grant-in-Aid for Scientific Research on Priority Areas (A-14082203) and a Grant-in-Aid for Creative Scientific Research (16GS0313) from the Ministry of Education, Culture, Sports and Technology. This work was also supported in part by 21st Century COE Program Knowledge Information Infrastructure for Genome Science. We thank Dr. A. Kurosaka (Kyoto Sangyo University) for the amino acid analysis.

References

Abe, T. and Norton, W.T. (
1974
) The characterization of sphingolipids from neurons and astroglia of immature rat brain.
J. Neurochem.
 ,
23
,
1025
–1036.
Andressen, C., Arnhold, S., and Mai, J.K. (
1998
) Differential expression of lactoseries carbohydrate epitopes HNK-1, CD15, and NALA by olfactory receptor neurons in the developing chick.
Anat. Embryol. (Berl.)
 ,
197
,
209
–215.
Ariga, T., Kohriyama, T., Freddo, L., Latov, N., Saito, M., Kon, K., Ando, S., Suzuki, M., Hemling, M.E., and Rinchart, K.L. Jr. (
1987
) Characterization of sulfated glucuronic acid-containing glycolipids reacting with IgM M-proteins in patients with neuropathy.
J. Biol. Chem.
 ,
262
,
848
–853.
Bronner-Fraser, M. (
1987
) Perturbation of cranial neural crest migration by the HNK-1 antibody.
Dev. Biol.
 ,
123
,
321
–331.
Chou, D.K.H., Ilyas, A.A., Evans, J.E., Costello, C., Quarles, R.H., and Jungalwala, F.B. (
1986
) Structure of sulfated glucuronyl glycolipids in the nervous system reacting with HNK-1 antibody and some IgM paraproteins in neuropathy.
J. Biol. Chem.
 ,
261
,
11717
–11725.
Colley, K.J. (
1997
) Golgi localization of glycosyltransferases: more questions than answers.
Glycobiology
 ,
7
,
1
–13.
Jessell, T.M., Hynes, M.A., and Dodd, J. (
1990
) Carbohydrates and carbohydrate-binding proteins in the nervous system.
Ann. Rev. Neurosci.
 ,
13
,
227
–255.
Kakuda, S., Oka, S., and Kawasaki, T. (
2004
) Purification and characterization of two recombinant human glucuronyltransferases involved in the biosynthesis of HNK-1 carbohydrate in Escherichia coli.
Prot. Exp. Purif.
 ,
35
,
111
–119.
Kakuda, S., Shiba, T., Ishiguro, M., Tagawa, H., Oka, S., Kajihara, Y., Kawasaki, T., Wakatsuki, S., and Kato, R. (
2004
) Structural basis for acceptor substrate recognition of a human glucuronyltransferase, GlcAT-P, an enzyme critical in the biosynthesis of the carbohydrate epitope HNK-1.
J. Biol. Chem.
 ,
279
,
22693
–22703.
Kawashima, C., Terayama, K., Ii, M., Oka. S., and Kawasaki, T., (
1992
) Characterization of a glucuronyltransferase: neolactotetraosylceramide glucuronyltransferase from rat brain.
Glycoconj. J.
 ,
9
,
307
–314.
Kruse, J., Mailhammer, R., Wernecke, H., Faissner, A., Sommer, I., Goridis, C., and Schachner, M. (
1984
) Neural cell adhesion molecules and myelin-associated glycoprotein share a common carbohydrate moiety recognized by monoclonal antibodies L2 and HNK-1.
Nature
 ,
311
,
153
–155.
Kunemund, V., Jungalwala, F.B., Fischer, G., Chou, D.K., Keilhauer, G., and Schachner, M. (
1988
) The L2/HNK-1 carbohydrate of neural cell adhesion molecules is involved in cell interactions.
J. Cell. Biol.
 ,
106
,
213
–223.
Martini, R., Xin, Y., Schumitz, B., and Schachner, M. (
1992
) The L2/HNK-1 carbohydrate epitope is involved in the preferential outgrowth of motor neuron on ventral roots and motor nerves.
Eur. J. Neurosci.
 
4
,
628
–639.
Nilsson, O.S. and Dallner, G. (
1977
) Transverse asymmetry of phospholipids is subcellular membranes of rat liver.
Biochem. Biophys. Acta
 ,
464
,
453
–458.
Oka, S., Terayama, K., Kawashima. C., and Kawasaki. T. (
1992
) A novel glucuronyltransferase in nervous system presumably associated with the biosynthesis of HNK-1 carbohydrate epitope on glycoprotein.
J. Biol. Chem.
 ,
267
,
22711
–22714.
Oka, S., Terayama, K., Imiya, K., Yamamoto, S., Kondo, A., Kato, I., and Kawasaki, T. (
2000
) The N-glycan acceptor specificity of a glucuronyltransferase, GlcAT-P, associated with biosynthesis of the HNK-1 epitope.
Glycoconj. J.
 ,
17
(12),
877
–885.
Paulson, J.C. and Colley, K.J. (
1989
) Glycosyltransferases. Structure, localization, and control of cell type-specific glycosylation.
J. Biol. Chem.
 ,
264
,
17615
–17618.
Pukazhenthi, B.S., Muniappa, N., and Vijay, I.K. (
1993
) Role of sulfhydryl groups in the function of glucosidase I from mammary gland.
J. Biol. Chem.
 ,
268
,
6445
–6452.
Rutishauser, U., Acheson, A., Hall, A.K., Mann, D.M., and Sunshine, J. (
1988
) The neural cell adhesion molecule (NCAM) as a regulator of cell–cell interactions.
Science
 ,
240
,
53
–57.
Schwarting, G.A., Jungalwala, F.B., Chou, D.K., Boyer, A.M., and Yamamoto, M. (
1987
) Sulfated glucuronic acid-containing glycoconjugates are temporally and spatially regulated antigens in the developing mammalian nervous system.
Dev. Biol.
 ,
120
,
65
–76.
Seiki, T., Oka, S., Terayama, K., Imiya, K., and Kawasaki, T. (
1999
) Molecular cloning and expression of a second glucuronyltransferase involved in the biosynthesis of the HNK-1 carbohydrate epitope.
Biochem. Biophys. Res. Commun.
 ,
255
,
182
–187.
Shimoda, Y., Tajima, Y., Nagase, T., Harii, K., Osumi, N., and Sanai, Y. (
1999
) Cloning and expression of a novel galactoside β1,3-glucuronyltransferase involved in the biosynthesis of HNK-1 epitope.
J. Biol. Chem.
 ,
274
,
17115
–17122.
Stanley, P. and Siminovitch, L. (
1977
) Complementation between mutants of CHO cells resistant to a variety of plant lectins.
Somatic Cell Genet.
 ,
3
,
391
–405.
Terayama, K., Oka, S., Seiki, T., Miki, Y., Nakamura, A., Takio, K., Kozutsumi, Y., and Kawasaki, T. (
1997
) Cloning and functional expression of a novel glucuronyltransferase involved in the biosynthesis of the carbohydrate epitope HNK-1.
Proc. Natl Acad. Sci. USA
 ,
94
,
6093
–6098.
Terayama, K., Nakamura, A., Seiki, T., Matsumori, K., Ohta, S., Oka, S., Sugita, M., and Kawasaki, T. (
1998
) Purification and characterization of a glucuronyltransferase involved in the biosynthesis of the HNK-1 epitope on glycoproteins from rat brain.
J. Biol. Chem.
 ,
273
,
30295
–30300.
Voshol, H., van Zuylen, C.W.E.M., Orberger, G., Vliegenthart, J.F.G., and Schachner, M. (
1996
) Structure of the HNK-1 carbohydrate epitope on bovine peripheral myelin glycoprotein P0.
J. Biol. Chem.
 ,
271
,
22597
–22960.
Yamaguchi N. and Fukuda M.N. (
1995
) Golgi retention mechanism of β-1,4-galactosyltransferase.
J. Biol. Chem.
 ,
270
,
12170
–12176.
Yamamoto, S., Oka, S., Inoue, M., Shimuta, M., Manabe, T., Takahashi, H., Miyamoto, M., Asano, M., Sakagami, J., Sudo, K., and others. (
2002
) Mice deficient in nervous system-specific carbohydrate epitope HNK-1 exhibit impaired synaptic plasticity and spatial learning.
J. Biol. Chem.
 ,
277
,
27227
–27231.
Yoshihara, Y., Oka, S., Ikeda, J., and Mori, K. (
1991
) Immunoglobulin superfamily molecules in the nervous system.
Neurosci. Res.
 ,
10
,
83
–105.
Yoshihara, Y., Oka, S., Watanabe, Y., and Mori, K. (
1991
) Developmental and spatially regulated expression of HNK-1 carbohydrate antigen on a novel phosphatidylinositol-anchored glycoprotein in rat brain.
J. Cell. Biol.
 ,
115
,
731
–744.

Author notes

2Department of Biological Chemistry, Graduate School of Pharmaceutical Sciences, Kyoto University, Kyoto 606-501, Japan, and 3CREST JST, Graduate School of Pharmaceutical Sciences, Kyoto University, Kyoto 606-501, Japan