INTRODUCTION

Human respiratory epithelium is covered by a gelatinous layer of mucus, situated on the top of the cilia, and continuously moved toward the pharynx where it is swallowed (1). Mucins are the main components of the respiratory mucus and are responsible to a large extent for the characteristic rheological properties of respiratory mucus, necessary for the efficiency of the mucociliary system.

Human respiratory mucins consist of a broad family of polydisperse and high molecular mass glycoproteins synthesized by the respiratory mucosa. The peptide diversity stems from the expression of several genes (2, 3); however, this mucin diversity is increased widely by post-translational phenomena, mostly O-glycosylation but also sulfation, which are responsible for 70-80% of the mass of the mucin molecule. These phenomena lead to a remarkable diversity of carbohydrate chains (4) allowing the recognition of inhaled microorganisms (for review, see Ref. 5) which are then eliminated by the activity of the mucociliary system. Thus mucins play an essential role in the defense of the respiratory mucosa. Any change in the glycosylation or sulfation of mucins may affect the rheological properties of respiratory mucus and the efficiency of the mucociliary system, leading to bronchial obstruction as observed in cystic fibrosis (CF).¹

CF, a general exocrinopathy, is the most common severe genetic disease among Caucasians. In its most typical form, the severity of CF is the result of chronic lung infection and mucus hypersecretion. This infection is very peculiar and characterized by the predominance of Staphylococcus aureus in early life and, later on, of Pseudomonas aeruginosa which is almost impossible to eradicate and is responsible for most of the morbidity and mortality of the disease (6).

CF is caused by mutations in the gene encoding for cystic fibrosis transmembrane conductance regulator (CFTR) (7), a chloride channel of low conductance activated by protein kinase A (8-11). F508 deletion is the most frequent mutation (about 70% of the CF chromosomes) and leads to a misfolding of CFTR which is retained in the endoplasmic reticulum and degraded (12). This mislocation of mutated CFTR may have consequences for other cell compartments. In CF cells, a defect in the acidification of endosomes, prelysosomes, or trans-Golgi/trans-Golgi network has been observed and could be responsible for abnormal glycosylation and sulfation processes (13-16).

Previous works have shown that salivary and respiratory mucins (17-19) as well as glycoproteins secreted by CF nasal epithelial cells in culture (20) were oversulfated. More recently, Zhang et al. (21), using a model of human xenograft that eliminates the influence of inflammation and infection, observed an increased mucin sulfation that may correspond to a primary defect, varying according to the CF genotype. Mucin sulfation affords a strong negative charge to carbohydrate chains influencing the viscoelastic properties of bronchial mucus. Moreover, sulfated carbohydrate sequences, identical to those found in respiratory mucins, have been shown to be specific ligands for selectins (22, 23) and for microorganisms (24). The development of a protocol of purification based mainly on high performance anion-exchange chromatography (HPAEC) has allowed the structural determination of sulfated carbohydrate chains from CF and non-CF mucins in which sulfation occurs either on the C-3 of a terminal galactose (Gal) residue or on the C-6 of an N-acetylglucosamine (GlcNAc) residue (25). These data suggest that sulfation of respiratory mucins involves at least two sulfotransferases.

Lo-Guidice et al. (26) have characterized recently a sulfotransferase from human airways responsible for the 3-O-sulfation of terminal galactose in N-acetyllactosamine-containing mucin carbohydrate chains.

In this paper we report the characterization of a microsomal GlcNAc-6-O-sulfotransferase activity from human bronchial mucosa. This enzyme is able to transfer a sulfate group from PAPS to a terminal GlcNAc residue contained in a carbohydrate chain. The GlcNAc-6-O-sulfation is inhibited by the prior 1-4 galactosylation of the GlcNAc residue, indicating that 6-O-sulfation has to precede galactosylation during biosynthesis of sulfated mucin-type oligosaccharides. The main difference in the enzymatic properties of the 3-O-Gal- and the 6-O-GlcNAc-sulfotransferase is their optimum pH, 6.1 and 6.7, respectively.

EXPERIMENTAL PROCEDURES

Tissues collected in macroscopically healthy areas of the bronchial tree from patients undergoing surgery for bronchial carcinoma were placed in Leibovitz L15 medium (Life Technologies, Inc.) and transported immediately on ice to the laboratory and processed for mucosa isolation. Mucosae (2-3 cm²) were cut into 1-mm² pieces, and homogenates were prepared in 50 mM Tris-HCl buffer, pH 7.4, containing 25 mM potassium chloride, 250 mM saccharose, 5 mM -mercaptoethanol, 5 mM magnesium acetate, using a glass-Teflon homogenizer (1,400 rpm, five strokes). The mixture obtained was submitted to 16,000 × g centrifugation for 20 min at 4 °C. The resulting supernatants were ultracentrifuged further at 180,000 × g for 1 h at 10 °C. The resulting pellet containing microsomal fractions was stored at 80 °C until used (26).

The oligosaccharide-alditols of respiratory mucins from a patient suffering from CF were released by alkaline borohydride treatment of mucin glycopeptides, then fractionated by anion-exchange chromatography on an AG 1-X2 column (Bio-Rad) and then by gel filtration chromatography on a Bio-Gel P4 column (Bio-Rad) according to Lamblin et al. (27). Fraction IIIc1 containing mono- or disialyloligosaccharide-alditols was fractionated by HPAEC on a CarboPac PA-100 column (Dionex), and the structures of the main oligosaccharide-alditols were determined by high resolution ¹H NMR spectroscopy in combination with fast atom bombardment-mass spectrometry (25).

The different compounds also used for acceptor specificity studies were from the following sources: GlcNAc1-O-Met, GlcNAc1-O-Met, GlcNAc1-3Gal-O-Met were from Sigma; NeuAc2-3Gal1-4GlcNAc-O-Met (2-3sialyl-N-acetyllactosamine-O-Met), and NeuAc2-3Gal1-4[Fuc1-3]GlcNAc-O-Met (sialyl-Lewis x-O-Met or sLe^x-O-Met) were from Toronto Research Chemicals Inc. We also used three oligosaccharide-alditols OS1, OS2, and OS3, whose structures are described in Table I. The oligosaccharide-alditol OS1 was from Collocalia mucins (28) and OS3 was from human bronchial mucins. OS2 was obtained by treating OS1 with Streptococcus pneumoniae -galactosidase (Sigma) according to Paulson et al. (29). The galactose was then removed by gel filtration on a Bio-Gel P2 column (Bio-Rad). Fraction IIIc1 was also treated with S. pneumoniae -galactosidase using the same protocole.

Table I. Structures of oligosaccharide-alditols OS1, OS2, and OS3 Oligosaccharide-alditol Structure OS1 Gal(1-4)GlcNAc(1-6)                        \                     GalNAc-ol                       /              Gal(1-3)               /   NeuAc(2-3) OS2            GlcNAc(1-6)                         \                     GalNAc-ol                       /              Gal(1-3)               /   NeuAc(2-3) OS3                     GalNAc-ol                         /            GlcNAc(1-3)

Sulfated oligosaccharide-alditols (fraction IVc) were prepared from the respiratory mucins of a CF patient using the same methods as for fraction IIIc1 (25). The structures of 11 sulfated oligosaccharide-alditols from fraction IVc have been determined by Lo-Guidice et al. (25).

6-O-Sulfated methyl--N-acetylglucosaminide was synthesized according to the protocol described by Van Kuik et al. (30) specific for sulfation of -CH₂OH. Briefly, 50 mg of GlcNAc1-O-Met was dissolved in 1.25 ml of dry pyridine. The mixture was cooled to 5 °C, and 18.5 µl of chlorosulfonic acid in 75 µl of dry chloroform was added. After stirring for 30 min at 5 °C and then for 2 h at 25 °C, the reaction was stopped by the addition of 0.5 ml of water. The solvent was then evaporated to dryness. The sulfated GlcNAc1-O-Met was purified on a silica column (1.2 × 13 cm) (Florisil, Merck) eluted by a mixture of dichloromethane/methanol (5:3, v/v), and their structure was determined by 400 MHz ¹H NMR spectroscopy on a two-dimensional homonuclear COSY spectrum. The analysis proved the presence of two sulfated products: HO₃S-6GlcNAc1-O-Met, which was the main synthesized product; and HO₃S-3GlcNAc1-O-Met (90%/10%). The sulfation of GlcNAc1-O-Met on the C-6 was confirmed by the strong downfield shift for the AH6/AH6 protons (+0.45 ppm and +0.49 ppm, respectively). HO₃S-6GlcNAc1-3Gal-O-Met was a generous gift from K. L. Matta (Roswell Park Cancer Institute, Buffalo, NY) (31).

The reaction mixture (100 µl) for the sulfotransferase assays was performed as follows. 50-100 µg of microsomal protein was incubated with 0.5 µCi of [³⁵S]PAPS (NEN Life Science Products, 2.48-2.50 Ci/mmol), 5 mM of methylglycosides (GlcNAc/1-O-Met, GlcNAc1-3Gal-O-Met, NeuAc2-3Gal1-4 GlcNAc-O-Met, or sLe^x-O-Met), or 100 µg of one of the oligosaccharide-alditols (OS1, OS2, OS3) in a 30 mM MOPS/NaOH buffer, pH 6.7, containing 0.1% (w/v) Triton X-100, 20 mM MnCl₂, 30 mM NaF, 5 mM AMP, 1 mM 4-(2-aminoethyl)benzenesulfonyl fluoride (AEBSF). After incubation for 60 min at 30 °C, the reaction was stopped by the addition of 300 µl of ice-cold methanol. The resulting mixture was kept overnight at 4 °C, and the formed precipitates were eliminated by centrifugation at 10,000 × g for 20 min. The pellets were washed twice with ice-cold methanol and centrifuged. The supernatants were collected, evaporated to dryness, and then submitted to HPAEC.

Action of Human Microsomal Galactosyltransferase on HO₃S-6GlcNAc1-O-Met or on Oligosaccharide-alditol IVc-19 (OS2 with a Terminal HO₃S-6GlcNAc)

After incubation of GlcNAc1-O-Met or OS2 with radiolabeled [³⁵S]PAPS and microsomal preparation for 60 min at 30 °C as described above, UDP-Gal (Sigma) was added to the mixture to have a final concentration of 5 mM. The mixture was kept for 1 h 30 min at 30 °C, and the reaction was stopped as described for the sulfotransferase assays. The synthesized products were identified by HPAEC.

To compare the activity of this galactosyltransferase on GlcNAc1-O-Met/HO₃S-6GlcNAc1-O-Met and OS2/IVc-19, each substrate (1 mM) was incubated with 1 µCi of radiolabeled UDP-[6-³H]Gal (Sigma, 17.65 Ci/mmol) for 1 h 30 min at 30 °C in a 30 mM MOPS/NaOH buffer containing 5 mM AMP, 30 mM NaF, 20 mM MnCl₂, 1 mM AEBSF, and 0.1% Triton X-100. The reaction was stopped as described for the sulfotransferase assays, and the radiolabeled products were studied by HPAEC.

Action of 1-4 Galactosyltransferase from Bovine Milk on HO₃S-6GlcNAc1-O-Met

To obtain a standard of Gal1-4(HO₃S-6)GlcNAc-O-Met, chemically synthesized HO₃S-6GlcNAc1-O-Met (5 mM) was incubated with 1 µCi of radiolabeled UDP-[6-³H]Gal (Sigma, 17.65 Ci/mmol) and 150 milliunits of 1-4 galactosyltransferase from bovine milk (Sigma) in 50 µl of 100 mM sodium cacodylate, 154 mM NaCl buffer, pH 7.4, containing 1 mM AMP, 10 mM MnCl₂, 0.5% (w/v) Triton X-100 (32). After 1 h 30 min at 30 °C, the reaction was stopped as described for the sulfotransferase assays, and the synthesized product was characterized by HPAEC.

Dry samples of sulfated or galactosylated products were dissolved in water and injected directly onto a CarboPac PA-100 column (4 × 250 mm) for HPAEC (Dionex Corp.). The elution of neosynthesized products was monitored both by pulsed amperometric detection (PAD 2 model, Dionex Corp.) and by radioactivity on line (high performance liquid chromatography radioactivity detector LB 506 C-1, EG & G, Berthold, Wildbad, Germany).

Elution of sulfated GlcNAc1-O-Met and sulfated products synthesized from methylglycosides (GlcNAc1-3Gal-O-Met, NeuAc2-3Gal1-4GlcNAc-O-Met, and sLe^x-O-Met) was performed at alkaline pH at a flow rate of 1 ml/min in 0.05 M NaOH, 0.2 M sodium acetate with a linear gradient of sodium acetate to 0.05 M NaOH, 0.3 M sodium acetate at 22 min, to 0.05 M NaOH, 0.95 M sodium acetate at 24 min, and followed by isocratic elution with 0.05 M NaOH, 0.95 M sodium acetate for 10 min (gradient I). The standards used were HO₃S-6GlcNAc1-O-Met and HO₃S-6GlcNAc1-3Gal-O-Met for sulfated GlcNAc1-O-Met and sulfated GlcNAc1-3Gal-O-Met, respectively.

The same gradient I was used for elution of the ³⁵S-sulfated and galactosylated products synthesized from GlcNAc1-O-Met. The standard used was Gal1-4(HO₃S-6)GlcNAc-O-Met (resulting from incubation of HO₃S-6GlcNAc1-O-Met with UDP-[6-³H]Gal and 1-4 galactosyltransferase from bovine milk).

Elution of neosynthesized radiolabeled sulfated oligosaccharide-alditols was performed at alkaline pH at a flow rate of 1 ml/min in 0.1 M NaOH for 10 min, then with a linear gradient of sodium acetate to 0.1 M NaOH, 0.07 M sodium acetate at 16 min, to 0.1 M NaOH, 0.1 M sodium acetate at 30 min, to 0.1 M NaOH, 0.45 M sodium acetate at 80 min, to 0.1 M NaOH, 0.95 M sodium acetate at 82 min and followed by isocratic elution with 0.1 M NaOH, 0.95 M sodium acetate for 10 min (gradient II). Fraction IVc containing 11 sulfated oligosaccharide-alditols was used as a standard (25). For elution of both ³⁵S-sulfated and galactosylated products synthesized from OS2, we also used gradient II and fraction IVc as a control.

The protein content of the microsomal fractions was determined by BCA Protein Assay (Pierce) (33).

RESULTS

Sulfation of Methyl-N-acetylglucosaminides

Methyl-N-acetylglucosaminides were first used to test the N-acetylglucosamine-sulfotransferase activity since it is difficult to obtain enough oligosaccharide-alditols from human respiratory mucins to characterize the enzyme completely. After incubation of methyl--N-acetylglucosaminide (GlcNAc1-O-Met) with microsomes and [³⁵S]PAPS and separation of the radiolabeled products by HPAEC, two peaks were observed: a peak at 11 min 30 s corresponding to free [³⁵S]sulfate (this peak is still present when incubation is performed without any carbohydrate acceptor) and another peak at 9 min 30 s, absent when there is no GlcNAc1-O-Met in the incubation mixture, corresponding to a sulfated GlcNAc1-O-Met (Fig. 1a). Unreacted [³⁵S]PAPS or [³⁵S]APS, which are very acidic compounds, are eluted with high concentrations of sodium acetate, and so they are not visualized on the elution profile.

HPAEC elution profile of ³⁵S-labeled products enzymatically obtained from GlcNAc1-O-Met on a CarboPac PA-100 column (4 × 250 mm).

[View Larger Version of this Image (14K GIF file)]

To find out the position of the sulfate group on the GlcNAc residue, the radiolabeled products were chromatographed with a mixture (90%/10%) of HO₃S-6GlcNAc1-O-Met and HO₃S-3GlcNAc1-O-Met that was chemically synthesized. The sulfated GlcNAc1-O-Met synthesized during incubation with microsomes and [³⁵S]PAPS coeluted with the standard HO₃S-6GlcNAc1-O-Met. This showed that the sulfotransferase was able to transfer a sulfate group from PAPS to the C-6 of GlcNAc1-O-Met (Fig. 1b).

A very low sulfotransferase activity was found when methyl--N-acetylglucosaminide (GlcNAc1-O-Met) was used as substrate acceptor, showing that the enzyme was more active on the  anomers than on the  anomers. The activity was about 120-fold higher for GlcNAc1-O-Met (17.27 pmol/mg of protein/min) than for GlcNAc1-O-Met (0.14 pmol/mg of protein/min) using the same conditions of incubation.

Properties of the GlcNAc-6-O-sulfotransferase

We first looked at the effect of pH on the transfer of sulfate group from PAPS to GlcNAc1-O-Met, using MES buffer (pH 5.4-6.4) and MOPS buffer (pH 6.1-7.9). The optimal conditions for sulfation of GlcNAc1-O-Met were obtained with 30 mM MOPS/NaOH buffer at pH 6.7 (data not shown).

The influence of divalent cations on the activity of this GlcNAc-6-O-sulfotransferase from respiratory mucosa was also studied. Mn²⁺ and Mg²⁺ had a stimulatory effect on the GlcNAc-6-O-sulfotransferase with an optimal activity at 20 mM Mn²⁺. The sulfotransferase activity was 4-fold higher with 20 mM Mn²⁺ than without this cation. Ca²⁺ had an inhibitory effect even at very low concentrations (data not shown).

We also studied the influence of AMP, ATP, and NaF on the transfer of [³⁵S]sulfate to GlcNAc1-O-Met. Both AMP and ATP had a stimulatory effect at a concentration of 1 mM. Above this concentration, ATP abolished the sulfotransferase activity, whereas AMP had an increased stimulatory effect up to a 5 mM concentration. The presence of NaF in the incubation mixture stimulated the sulfotransferase activity, with a maximal effect at 30 mM NaF (Fig. 2).

Effect of AMP, ATP, and NaF on enzymatic sulfation of GlcNAc1-O-Met by human respiratory mucosa 6-O-sulfotransferase.

[View Larger Version of this Image (9K GIF file)]

These three compounds had an inhibitory effect on the liberation of free [³⁵S]sulfate from [³⁵S]PAPS during the incubations, allowing an increase of sulfotransferase activity. There was a free [³⁵S]sulfate decrease, 10, 85, and 80%, when the incubations were performed with 30 mM NaF, 5 mM AMP, and 10 mM ATP, respectively. Concerning ATP, the stimulatory effect was only observed at low concentrations (1 mM). Above this concentration, ATP had an inhibitory effect. This compound, which can be considered as a structural analog of PAPS has already been shown to inhibit sulfotransferase activities (34). At high concentrations, the impact on PAPS degradation would be lower than the inhibitory effect.

The optimal sulfotransferase activity was obtained with 0.1% (w/v) Triton X-100, 30 mM NaF, 20 mM MnCl₂, and 5 mM AMP in a 30 mM MOPS/NaOH buffer at pH 6.7.

When using these conditions, the sulfotransferase activity increased linearly up to 360 min, in the range of 2-96 µg of microsomal proteins (data not shown).

Kinetic measurements of GlcNAc-6-O-sulfotransferase activity with different concentrations of [³⁵S]PAPS (Fig. 3a) and GlcNAc1-O-Met (Fig. 3b) allowed the determination of K_m values from Lineweaver-Burk plots for these two components, 9.1 µM and 0.54 mM, respectively.

Effect of PAPS (panel a) and GlcNAc1-O-Met (panel b) concentrations on human respiratory mucosa 6-O-sulfotransferase activity.

[View Larger Version of this Image (12K GIF file)]

Sulfation of Oligosaccharidic Substrates

The sulfotransferase activity was measured on different oligosaccharide-alditols: OS1, OS2, OS3 (Table I), and fraction IIIc1 (sialylated oligosaccharide-alditols from human bronchial mucins (25)) with or without prior treatment with -galactosidase. These oligosaccharides were incubated with microsomal preparation and [³⁵S]PAPS in conditions described under "Experimental Procedures." The ³⁵S-radiolabeled products were analyzed by HPAEC using a CarboPac PA-100 column and gradient II and compared with a mixture of 11 sulfated oligosaccharide-alditols whose structures have been determined previously (IVc) (25). Three of them were particularly interesting. The oligosaccharide-alditol IVc-10 had the same structure as OS1 with a sulfate group on the C-3 of the terminal galactose residue; IVc-12 had the same structure as OS1 with a sulfate group on the C-6 of the internal GlcNAc residue, and IVc-19 corresponded to OS2 with a sulfate group on the C-6 of the terminal GlcNAc residue. The structure of the oligosaccharide-alditol IVc-2 was also very useful for this study (Table II).

Table II. Structures of oligosaccharide-alditols IVc-2, IVc-10, IVc-12, and IVc-19 from human respiratory mucins Oligosaccharide-alditol Structure IVc-2        3Gal(1-4)GlcNAc(1-6)      /               /      \ HO3S        Fuc(1-3)         GalNAc-ol                               /                      Gal(1-3)                     /          NeuAc(2-3) IVc-10        3Gal(1-4)GlcNAc(1-6)      /                       \ HO3S                           GalNAc-ol                               /                      Gal(1-3)                     /          NeuAc(2-3) IVc-12       HO3S            \     Gal(1-4)6GlcNAc(1-6)                            \                             GalNAc-ol                            /                   Gal(1-3)                  /       NeuAc(2-3) IVc-19         HO3S              \                6GlcNAc(1-6)                             \                              GalNAc-ol                            /                   Gal(1-3)                  /       NeuAc(2-3)

After sulfation of OS1, two radiolabeled peaks were obtained (Fig. 4a). The first peak was eluted at 45 min 24 s and was also present when the incubation mixture did not contain any acceptors; it corresponded to free [³⁵S]sulfate. The second peak was eluted at 46 min 36 s; it corresponded to the product synthesized from OS1. When this product was injected with a mixture of sulfated oligosaccharides (fraction IVc), it coeluted with oligosaccharide-alditol IVc-10, which corresponds to OS1 with a sulfate group on the C-3 of the terminal galactose residue. No radiolabeled peak had the same retention time as oligosaccharide-alditol IVc-12, which corresponds to OS1 with a sulfate group on the internal GlcNAc residue (Table II). Thus the GlcNAc-6-O-sulfotransferase activity from human bronchial mucosa was not directly active on the internal GlcNAc residue of OS1 (Scheme 1).

[View Larger Version of this Image (21K GIF file)]

[View Larger Version of this Image (13K GIF file)]

OS1 was then submitted to the action of -galactosidase from S. pneumoniae and led to OS2 (Table I), which was incubated under the same conditions as OS1. The radiolabeled products were separated by HPAEC in two peaks (Fig. 4b). The first peak was eluted at 45 min 24 s and corresponded to free [³⁵S]sulfate; the second peak was eluted at 56 min 6 s and coeluted with oligosaccharide-alditol IVc-19 from human bronchial mucins, which is sulfated on the C-6 of the terminal GlcNAc residue (Fig. 4b). Thus this sulfotransferase from human bronchial mucosa is able to transfer a sulfate residue on the C-6 of a terminal GlcNAc residue contained in mucin carbohydrate chains (Scheme 1).

For OS3 (Table I), no radiolabeled peak was observed except the free [³⁵S]sulfate peak at 45 min 24 s (data not shown).

A pool of sialylated oligosaccharide-alditols from human bronchial mucins (fraction IIIc1) was also used as substrate acceptors for the GlcNAc-6-O-sulfotransferase. The structures of the main components of this fraction have been identified by Lo-Guidice et al. (25). One oligosaccharide-alditol of this fraction has the same structure as OS1. Coinjection on the CarboPac PA-100 column of fraction IVc and the neosynthesized products obtained from fraction IIIc1 showed the presence of three main labeled peaks. Two peaks correspond to oligosaccharides having a 3-O-sulfated terminal galactose. One was eluted at 34 min 30 s and had the same retention time as oligosaccharide-alditol IVc-2 (Table II). The other one was eluted at 46 min 36 s and had the same retention time as oligosaccharide-alditol IVc-10 (Table II). The peak that was eluted at 45 min 24 s corresponded to free [³⁵S]sulfate (Fig. 5a).

HPAEC elution profile on a CarboPac PA-100 column (4 × 250 mm) of a mixture containing ³⁵S-labeled products enzymatically synthesized from fraction IIIc1 (panel a), fraction IIIc1 treated with S. pneumoniae -galactosidase (panel b), and sulfated oligosaccharide-alditols (fraction IVc) from human respiratory mucins.

[View Larger Version of this Image (19K GIF file)]

Treatment of fraction IIIc1 with -galactosidase followed by incubation with microsomes and [³⁵S]PAPS and fractionation of the neosynthesized products by HPAEC led to the elution of three main radiolabeled peaks: the free [³⁵S]sulfate peak at 45 min 24 s, one radiolabeled peak that coeluted with oligosaccharide-alditol IVc-19 (Table II) (retention time: 56 min 6 s), and another peak that was eluted at 50 min 42 s and could not be identified (Fig. 5b).

Three different methylglycosides containing GlcNAc residues were used as sulfate acceptors as described under "Experimental Procedures." The neosynthesized products were analyzed by HPAEC on a CarboPac PA-100 column using gradient I.

When using GlcNAc1-3Gal-O-Met, two radiolabeled peaks were found on the elution profile: the free [³⁵S]sulfate peak at 11 min 30 s and another peak that was eluted at 3 min 54 s and corresponded to the radiolabeled oligosaccharide synthesized from GlcNAc1-3Gal-O-Met. This peak coeluted with HO₃S-6GlcNAc1-3Gal-O-Met, showing that the sulfate group was transferred on the C-6 of the terminal GlcNAc residue by a microsomal sulfotransferase (data not shown). No sulfation occurred on sLe^x-O-Met or on NeuAc2-3Gal1-4GlcNAc-O-Met.

These data argued that a sulfotransferase from human bronchial mucosa is active on GlcNAc1-3Gal-O-Met (bearing a terminal nonreducing GlcNAc residue) but not on the two substrates with an internal GlcNAc residue.

Competition Experiments

To determine whether the same sulfotransferase was responsible for the 6-O-sulfation of GlcNAc1-O-Met and OS2 and for the 3-O-sulfation of OS1, it was necessary to perform competition experiments.

For this study, we used the oligosaccharide-alditols OS1 and OS2 at a molar concentration of about 1 mM. GlcNAc1-O-Met had an important inhibitory effect on the sulfation of OS2 but no effect on the sulfation of OS1 (Fig. 6). The sulfation of OS2 was 50-66% lower when 1-33 mM GlcNAc1-O-Met was added in the incubation mixture. These data proved that the same enzyme was responsible for the 6-O-sulfation of GlcNAc1-O-Met and OS2. The 3-O-sulfation of OS1 was not affected by the presence of GlcNAc1-O-Met. These results prove that the GlcNAc-6-O-sulfotransferase characterized in the present work is different from the Gal-3-O-sulfotransferase described by Lo-Guidice et al. (26). This GlcNAc-6-O-sulfotransferase has more affinity for carbohydrate chains with a terminal GlcNAc residue than for GlcNAc1-O-Met since oligosaccharide sulfation is not inhibited completely by high concentrations of GlcNAc1-O-Met.

Competition for sulfation between GlcNAc1-O-Met and oligosaccharide-alditols OS1 or OS2.

[View Larger Version of this Image (8K GIF file)]

For the 6-O-sulfation of GlcNAc1-3Gal-O-Met, we had a standard that was a gift from K. L. Matta. Using this standard, we could show that sulfation of GlcNAc1-3Gal-O-Met occurred on the C-6 of the terminal GlcNAc residue, resulting from the action of a GlcNAc-6-O-sulfotransferase. GlcNAc1-O-Met also inhibited the 6-O-sulfation of this component (data not shown), suggesting that the same enzyme was responsible for the sulfation of GlcNAc1-O-Met and GlcNAc1-3Gal-O-Met.

Galactosylation of HO₃S-6GlcNAc1-O-Met and IVc-19 by Human Microsomal Galactosyltransferase

As HO₃S-6GlcNAc may be a component of a sulfated N-acetyllactosamine unit, it was interesting to determine whether this sulfated sugar could be an acceptor for a microsomal galactosyltransferase. When GlcNAc1-O-Met was incubated first with [³⁵S]PAPS and microsomal preparation and then with UDP-Gal, the neosynthesized products analyzed by HPAEC in conditions described under "Experimental Procedures" using gradient I showed three main radiolabeled peaks (Fig. 7a). One was eluted at 9 min 30 s and corresponded to ³⁵S-labeled HO₃S-6GlcNAc1-O-Met. The free [³⁵S]sulfate peak was eluted at 11 min 30 s. The third peak was eluted at 4 min 36 s and probably resulted from the action of a galactosyltransferase on HO₃S-6GlcNAc1-O-Met.

HPAEC elution profile on a CarboPac PA-100 column (4 × 250 mm) of (panel a) labeled products synthesized from GlcNAc1-O-Met, [³⁵S]PAPS, and UDP-Gal by microsomal sulfotransferase and galactosyltransferase and of (panel b) ³H-labeled products synthesized from HO₃S-6GlcNAc1-O-Met and UDP-[6-³H]Gal by 1-4 galactosyltransferase from bovine milk.

[View Larger Version of this Image (15K GIF file)]

To identify this product, we synthesized radiolabeled Gal1-4(HO₃S-6)GlcNAc1-O-Met by incubating HO₃S-6GlcNAc1-O-Met (synthesized according to Van Kuik et al. (30)) with UDP-[6-³H]Gal in conditions described under "Experimental Procedures." The neosynthesized products were studied by HPAEC using gradient I (Fig. 7b). Two radiolabeled peaks were obtained. One peak was eluted at 3 min 18 s and was also present when the incubation mixture did not contain HO₃S-6GlcNAc1-O-Met. It corresponded to UDP-[6-³H]Gal. The other peak was eluted at 4 min 36 s and corresponded to Gal1-4(HO₃S-6)GlcNAc1-O-Met synthesized by the action of 1-4 galactosyltransferase from bovine milk on HO₃S-6GlcNAc1-O-Met. These results prove that a galactosyltransferase from human bronchial mucosa can transfer galactose from UDP-Gal on HO₃S-6GlcNAc1-O-Met in a manner similar to the 1-4 galactosyltransferase from bovine milk.

These two galactosyltransferases were also active in a similar manner on GlcNAc1-O-Met (data not shown).

We also incubated OS2 (Table I) with [³⁵S]PAPS and microsomal preparation and then with UDP-Gal as explained under "Experimental Procedures" and analyzed the neosynthesized products by HPAEC using gradient II. Two main radiolabeled peaks were obtained. The first one was eluted at 45 min 24 s and corresponded to free [³⁵S]sulfate. The other peak was eluted at 49 min 30 s and coeluted with oligosaccharide IVc-12 (Table II) when the radiolabeled products were injected simultaneously with fraction IVc (Fig. 8). This indicates that a microsomal bronchial 1-4 galactosyltransferase is active on a terminal HO₃S-6GlcNAc residue from human mucin carbohydrate chains. No other neosynthesized peak but IVc-12 appeared on the elution profile, indicating that this oligosaccharide-alditol IVc-12 has not been 3-O-sulfated further on the terminal Gal residue.

[View Larger Version of this Image (21K GIF file)]

The action of the 1-4 galactosyltransferase from human bronchial mucosa was compared on the substrates GlcNAc1-O-Met/HO₃S-6GlcNAc-O-Met and OS2/IVc-19 (Table III) to check the influence of the 6-O-sulfation of GlcNAc on its further galactosylation. These four substrates, having a terminal nonreducing GlcNAc residue, 6-O-sulfated or not, were galactosylated. The activity of the galactosyltransferase was slightly lower when the terminal GlcNAc residue was 6-O-sulfated.

Table III. Activity of 1-4 galactosyltransferase from human bronchial mucosa toward GlcNAc1-O-Met/HO3S-6GlcNAc1-O-Met and OS2/IVc-19 Assays conditions were as described in "Experimental Procedures." Acceptor substrate  1-4 Galactosyltransferase activity pmol/mg protein/min         GlcNAc1-O-Met 7.23      HO3S-6GlcNAc1-O-Met 6.28           GlcNAc(1-6) 6.97                        \                      GalNAc-ol                      /             Gal(1-3)            / NeuAc(2-3)         (OS2)   HO3S 5.51        \          6GlcNAc(1-6)                       \                       GalNAc-ol                      /             Gal(1-3)             / NeuAc(2-3)       (IVc-19)

DISCUSSION

In a previous work, we have developed a performant protocol of purification of acidic carbohydrate chains from CF and non-CF mucins, based mainly on HPAEC, which has proved to be a suitable and reliable method to separate and to identify sulfated oligosaccharides (25). These data have shown that sulfation may occur either on the C-3 of a terminal galactose part of an N-acetyllactosamine chain or on the C-6 of an N-acetylglucosamine residue (25, 35). Therefore sulfation of human respiratory mucins involves at least two sulfotransferases. Recently, Lo-Guidice et al. (26) have characterized a galactose 3-O-sulfotransferase from human airways. We report here the characterization of a second sulfotransferase activity from human respiratory mucosa, responsible for the transfer of a sulfate group from PAPS to the C-6 of an N-acetylglucosamine residue.

The activity and the properties of this sulfotransferase were determined as described by Lo-Guidice et al. (26). The chemical synthesis of a standard of 6-O-sulfated methyl--N-acetylglucosaminide (30) allowed us to demonstrate that the bronchial microsomal preparation contained a sulfotransferase capable of transferring a sulfate group from PAPS to the C-6 of a terminal GlcNAc residue.

The presence of free [³⁵S]sulfate in the reaction mixture was also observed in the previous work on the galactose 3-O-sulfotransferase (26) and in several other tissue extracts (36-38). The addition of NaF and AMP in sulfotransferase assays protects the PAPS from degradation, allowing an increase in the sulfotransferase activity (26, 39). The GlcNAc-6-O-sulfotransferase is stimulated significantly by Mg²⁺ and more particularly by Mn²⁺.

Using methyl--N-acetylglucosaminide as an acceptor, the optimal activity of the GlcNAc-6-O-sulfotransferase was obtained with 0.1% Triton X-100, 30 mM NaF, 20 mM Mn²⁺, 5 mM AMP in a 30 mM MOPS/NaOH buffer at pH 6.7. The main difference in the enzymatic properties of the Gal-3-O- and GlcNAc-6-O-sulfotransferase is their optimum pH.

Most of the sulfated or sialylated and sulfated oligosaccharide-alditols described previously have a core type 2. In 12 oligosaccharides 6-O-sulfation occurred on the GlcNAc part of this core, and in 2 cases, 6-O-sulfation occurred on a GlcNAc residue 1-3 linked to the Gal part of core 1 (in 6 oligosaccharide-alditols, sulfation occurs on the C-3 of a terminal Gal residue contained in N-acetyllactosamine chains) (25). Therefore, in a first step, we tested the activity of the GlcNAc-6-O-sulfotransferase on two oligosaccharide-alditols, OS1 and OS2 (see Table I). After incubation of these oligosaccharide-alditols with the microsomal preparation (which contains both activities Gal-3-O- and GlcNAc-6-O-sulfotransferase) and [³⁵S]PAPS, two sulfated products were obtained: OS1 sulfated on the C-3 of the terminal galactose, corresponding to IVc-10 (see Table II and ref. 25), and OS2 sulfated on the C-6 of the terminal GlcNAc residue corresponding to IVc-19. No sulfation occurred on the C-6 of the internal GlcNAc of OS1, leading to oligosaccharide-alditol IVc-12 present in human respiratory mucins. No sulfation occurred either on OS3 corresponding to core 3 of mucins. It should be stressed that in carbohydrate chains of human respiratory mucins, no sulfation on GlcNAc residues 1-3 linked to N-acetylgalactosaminitol has been observed so far (25, 35).

Incubation of a pool of sialylated oligosaccharide-alditols IIIc1 from human respiratory mucins (25) with the microsomal preparation led to the sulfation of two oligosaccharide-alditols on the C-3 of galactose. Prior degalactosylation of IIIc1 with -galactosidase is very limited due to sialylation of most Gal residues but allowed to obtain at least one oligosaccharide-alditol with a terminal nonreducing GlcNAc that could be 6-O-sulfated to generate IVc-19.

Four substrates having nonreducing terminal GlcNAc, 6-O-sulfated or not, could be galactosylated by the 1-4 galactosyltransferase from human bronchial mucosa. This enzyme probably recognizes terminal GlcNAc, 6-O-sulfated or not, and the influence of the sulfate group on the activity of the galactosyltransferase is only moderate (Table III). As a matter of fact, terminal N-acetyllactosamine units from human respiratory mucins may be nonsulfated on the GlcNAc residue (25). Altogether, these results suggest that in the biosynthesis of N-acetyllactosamine chains containing 6-O-sulfated GlcNAc, the sulfation of GlcNAc has to precede galactosylation and that the N-acetyllactosamine chains cannot be acceptors for this GlcNAc-6-O-sulfotransferase. These results are in good agreement with those observed for a 6-O-sulfotransferase from rat liver active on GlcNAc residues 1-6 linked to mannose (40). It is likely that, in vivo, there is a competition between the GlcNAc-6-O-sulfotransferase and the 1-4 galactosyltransferase, acting both on the same GlcNAc residue, and that once the Gal residue is linked to the GlcNAc, the GlcNAc-6-O-sulfotransferase is inactive.

Only two human sulfotransferases active on mucins have been characterized so far: a Gal-3-O-sulfotransferase (26) and a GlcNAc-6-O-sulfotransferase, described here, both localized in respiratory mucosa. Other Gal-3-O- (39, 41) as well as GlcNAc-6-O-sulfotransferases (42-44) acting on animal mucin carbohydrate chains have also been described previously.

One of the best endothelial associated ligands for L-selectin is GlyCAM-1, a mucin-like glycoprotein (45) whose smallest O-glycans have a core 2 and a upper chain consisting of 6-O-sulfated derivatives of sLe^x either on the GlcNAc or on the Gal residues of the Gal1-4GlcNAc chain (46). In the biosynthesis of GlyCAM-1, Crommie and Rosen (47) have observed that sialylation precedes both fucosylation and sulfation during biosynthesis. However, the temporal relationship between sulfation and fucosylation is still controversial (48, 49). In our case, no sulfation occurred using sLe^x-O-Met or 2-3-sialyl-N-acetyllactosamine as acceptor for the respiratory GlcNAc-6-O-sulfotransferase.

Because two oligosaccharide-alditols from human respiratory mucins were 6-O-sulfated on a GlcNAc 1-3 linked to a Gal residue, we looked at the activity of this GlcNAc-6-O-sulfotransferase on GlcNAc1-3Gal-O-Met. The neosynthesized product coeluted exactly with the corresponding 6-O-sulfated standard.

In conclusion, we have characterized a GlcNAc-sulfotransferase in human respiratory mucosa which transfers sulfate to the C-6 position of terminal GlcNAc residues of carbohydrate chains from human respiratory mucins and which can be substituted further by a galactose residue to produce Gal1-4(HO₃S-6)GlcNAc. This enzyme is inactive on N-acetyllactosamine chains, suggesting that 6-O-sulfation of GlcNAc has to precede galactosylation during mucin-type oligosaccharide biosynthesis. The same results have been observed for a GlcNAc-6-O-sulfotransferase from rat liver acting on GlcNAc1-6Man sequences (40). Structural determinations of oligosaccharide-alditols isolated from CF and non-CF mucins have shown that CF mucins were predominantly 6-O-sulfated on GlcNAc, whereas non-CF mucins may be more 3-O-sulfated on Gal (25, 35). Enzymatic properties and K_m of the two sulfotransferases responsible for sulfation of respiratory mucins are rather similar except for their optimum pH. The GlcNAc-6-O-sulfotransferase described here has an optimal pH (6.7) higher than that of the Gal-3-O-sulfotransferase described previously (6.1) (26). This observation is interesting within the context of CF since some reports have suggested that mutations of CFTR (especially F508) could be responsible for defective acidification of the trans-Golgi/trans-Golgi network pH (whose normal pH is 6.0) leading to modifications in the sulfation and glycosylation processes and notably to hyperactivity of some sulfotransferases (13, 14). Altogether these results suggest that the GlcNAc-6-O-sulfotransferase described here might be responsible for the predominance of 6-O-sulfation in CF respiratory mucins (25) and for oversulfation of CF mucins, which has been recently shown to be a primary defect (21).