Characterization of mucins and proteoglycans synthesized by a mucin ...

1 downloads 0 Views 427KB Size Report
The low proportion of mucin-secreting cells in parental HT-29 cultures .... minutes at 1000 g at 4°C. After adding cesium bromide (0.42 g/ml), cell lysates were ...
1275

Journal of Cell Science 108, 1275-1285 (1995) Printed in Great Britain © The Company of Biologists Limited 1995

Characterization of mucins and proteoglycans synthesized by a mucinsecreting HT-29 cell subpopulation G. Huet1,*, I. Kim2, C. de Bolos3, J. M. Lo-Guidice1, O. Moreau1, B. Hemon1, C. Richet1, P. Delannoy2, F. X. Real3 and P. Degand1 1INSERM U377, Place de Verdun, 59045 Lille cedex, France 2UMR CNRS no. 111, UST Lille, 59655 Villeneuve d’Ascq cedex, France 3Institut Municipal d’Investigacio Mèdica, Universitat Autonoma de Barcelona,

Spain

*Author for correspondence

SUMMARY HT-29 cells selected by adaptation to 10−5 M methotrexate (HT-29 MTX) are a homogeneous cell population producing high amounts of mucin. Intracellular mucins and proteoglycans were isolated from these cells by ultracentrifugation of cell lysates on a cesium bromide gradient and further separated by anion-exchange high perfomance liquid chromatography. The major mucin fraction isolated was characterized by a high hydroxy amino acid content (40%), a Thr/Ser ratio of 1.52, a high sialic acid content, and a low sulfate content. When the same procedure was applied to undifferentiated HT-29 cells, a minor mucin fraction was isolated which appeared less sialylated and more sulfated. The major proteoglycan species identified in HT-29 MTX cells showed less acidic behavior than the proteoglycan isolated from HT-29 cells. The effect of brefeldin

A and the sugar analog GalNAc-α-O-benzyl on the synthesis and biochemical properties of mucins synthesized by HT-29 MTX cells was examined. Brefeldin A induced the synthesis of more-sulfated mucins. GalNAc-α-O-benzyl treatment resulted in mucins with an increased content of T antigen and a 13-fold lower sialic acid content. We show that GalNAc-α-O-benzyl was metabolized by the cells to Galβ1-3GalNAc-α-O-benzyl, which, in turn, was a potent competitive inhibitor of the O-glycan α-2,3-sialyltransferase. These results illustrate the suitability of HT-29 MTX cells as a model to analyse mucin synthesis and sialylation.

INTRODUCTION

clonal cell line (cl.16-E) has been derived from HT-29 cultures after treatment with sodium butyrate (Augeron and Laboisse, 1984) and a homogeneous mucin-secreting population has been obtained by stepwise adaptation of HT-29 cells to 10−6 or 10−5 M methotrexate (MTX) (Lesuffleur et al., 1990). The use of another antimetabolite, 5-fluorouracil (5-FU), has resulted in the isolation of a population containing both absorptive and mucin-secreting cells (Lesuffleur et al., 1991). Antibodies raised against cl.16-E mucins stain goblet cells in both normal colonic and normal gastric mucosa (Maoret et al., 1989). Lesuffleur et al. (1990) have used rabbit antisera raised against normal colonic mucosa and absorbed with stomach tissue to identify mucins of ‘colonic’ specificity, and rabbit antisera raised against normal gastric mucosa and absorbed with colonic tissue to identify mucins of ‘gastric’ specificity. The low proportion of mucin-secreting cells in parental HT-29 cultures expresses mucins of either gastric or colonic immunological specificity, whereas HT-29 MTX cells synthesize mainly gastric-type mucins and HT-29 FU cells synthesize colonic-type mucins (Lesuffleur et al., 1990, 1991). The characterization of these cell populations has been facilitated by the isolation of cDNAs encoding human mucins (Gendler et al., 1990; Gum et al., 1990, 1992; Porchet et al., 1991; Guyonnet-

Mucins and proteoglycans are high molecular mass glycoproteins whose biochemical functions depend on post-translational modifications. Alterations in these processes commonly occur in transformed cells (Hardingham and Fosang, 1992; Devine and McKenzie, 1992) and may contribute to the neoplastic phenotype. In human colorectal carcinomas, abnormalities in the oligosaccharide structures carried on mucins (Kurosaka et al., 1983; Kim, 1992) and on proteoglycans (Iozzo and Wight, 1982) have been described. The analysis of the biosynthesis and composition of mucins and proteoglycans, and the study of structure function relationships, would be greatly facilitated by the availability of cell cultures with a wide range of phenotypic properties. The parental HT-29 cell line fulfills this criterion. HT-29 cultures are heterogeneous and in the postconfluent state consist of >95% undifferentiated cells and a small proportion of differentiated mucin-secreting and absorptive cells (Pinto et al., 1982; Augeron and Laboisse, 1984; Lesuffleur et al., 1990). Differentiated populations of either absorptive or mucinsecreting phenotype can be obtained under various conditions of metabolic stress. A stably differentiated mucin-secreting

Key words: mucin, proteoglycan, HT-29, aryl-N-acetyl-αgalactosaminide, brefeldin

1276 G. Huet and others Duperat et al., 1994; Toribara et al., 1993; Bobek et al., 1993). HT-29, HT-29 MTX and HT-29 FU cells show distinct patterns of MUC1-MUC5 mucin gene expression and HT-29 MTX cells express mainly MUC1, MUC2, MUC3 and MUC5C genes (Lesuffleur et al., 1993). However, less is known regarding the main carbohydrate structures carried on the products of these genes in HT-29 MTX cells (Dahiya et al., 1992). This paper describes a first step towards the biochemical and structural characterization of the mucins and proteoglycans synthesized by HT-29 MTX cells in comparison with those produced by the parental cell population. To gain insight into mucin biosynthesis in HT-29 MTX cells, we have used two compounds that could potentially interfere with glycosylation and alter the biochemical properties of mucins, brefeldin A (BFA) and benzyl-N-acetyl-α-galactosaminide (GalNAcα-Obenzyl). BFA is a fungal antibiotic that disrupts protein transport from the rough endoplasmic reticulum (RER) to the Golgi apparatus, inducing a redistribution of Golgi proteins to the RER (Lippincott-Schwartz et al., 1989). Aryl-N-acetyl-αgalactosaminides have been initially used as potential competitors of the glycosylation of GalNAc residues linked to the core protein and have been shown to inhibit UDPGal:GalNAcβ1-3-galactosyltransferase in vitro (Kuan et al., 1989). However, it was observed that GalNAcα-O-benzyl treatment of HM7 colon cancer cells leads to an increased expression of Tn antigen (GalNAc-Thr/Ser), but also of T antigen (Galβl-3GalNAc), together with a decrease in sialylLex, sialyl-Lea and sulfomucin epitopes (Huang et al., 1992). To explain the unexpected increase in T antigen activity, it has been proposed that the in situ synthesis of Galβ1-3GalNAcαO-benzyl might lead to a more efficient inhibitor than GalNAcα-O-benzyl itself; the benzyl disaccharide would inhibit the elongation of T antigen by N-acetylglucosaminyltransferases, sialyltransferases and/or fucosyltransferase. Here, we provide evidence that GalNAcα-O-benzyl induces a marked increase in T antigen and a marked decrease in sialic acid content of HT-29 MTX mucins. Furthermore, the enzymatic mechanisms responsible for these changes in mucin synthesis are partially elucidated.

in threonine-free Dulbecco’s minimum essential medium (Institut Jacques Boy, Reims, France). After labeling, medium was collected and cells were lysed in RIPA buffer (0.01 M Tris-HCl, pH 8.0, 0.01 M NaCl, 0.1% sodium dodecyl sulfate (SDS), 1% Triton X-100, 0.5% sodium deoxycholate, 1% phenylmethylsufonyl fluoride, 0.001 M sodium ethylene diaminotetra-acetate). In some experiments, the effect of GalNAcα-O-benzyl (Sigma, St Louis, MO; 5 mM) or BFA (Sigma; 5 µg/ml) on mucin biosynthesis was examined. Cells were cultured for 20 days as described above, GalNAcα-O-benzyl or BFA were added for 24 hours, and cells were processed. Biochemical characterization Ultracentrifugation Twenty-one days after seeding, cells were rinsed twice with PBS, scraped with a rubber policeman, and directly lysed by ultrasonication in PBS. Cell extracts were collected after centrifugation for 5 minutes at 1000 g at 4°C. After adding cesium bromide (0.42 g/ml), cell lysates were ultracentrifuged (200,000 g for 72 hours) using a Beckman 70.1 Ti rotor (Houdret et al., 1986). Fractions of 1 ml were collected, weighed to determine density, and analysed for absorbance at 280 nm, orcinol reactivity, and dimethylmethylene blue (DMB) reactivity (De Jong et al., 1992). HPLC chromatography Anion-exchange high performance liquid chromatography (AEHPLC) was carried out using a TSK DEAE 5PW (7.5 mm × 75 mm) Spherogel column (Beckman). The system was equilibrated with 0.005 M sodium/potassium phosphate buffer, pH 6.0, at a flow rate of 0.8 ml/min. A total of 2 mg protein was injected. Elution was carried out using a NaCl gradient (0 to 1 M) in the same buffer. Eluates were collected in 0.8 ml fractions and analysed as described above. Compositional analyses For amino acid analysis, samples were hydrolysed in 5.6 M HCl for 24 hours under a vacuum and processed using a 7300 Beckman amino acid analyser (Palo Alto, CA) equipped with a high performance sodium column (4 mm × 120 mm) (Beckman). Sugar analysis was carried out by gas-liquid chromatography of trimethylsilyl derivatives of methyl glycosides formed by methanolysis in 1.5 M HCl in methanol at 80°C for 24 hours (Lamblin et al., 1984). To determine sulfate content, samples were hydrolysed in 1 M HCl for 5 hours at 100°C and sulfate was determined by AE-HPLC (Lo-Guidice et al., 1994). All compositional analyses were performed twice to examine the reproducibility of the results.

MATERIALS AND METHODS Cell culture Parental HT-29 cells (referred to as HT-29) and HT-29 cells selected by adaptation to 10−5 M methotrexate (MTX) were obtained from Dr Thécla Lesuffleur (Unité INSERM U178, Villejuif, France). Cells were grown in Dulbecco’s modified Eagle’s minimal essential medium (Eurobio, Paris, France), supplemented with 10% inactivated (30 minutes, 56°C) fetal bovine serum (Boehringer, Mannheim, Germany). Cells were seeded at 1.5×106 cells in 75 cm2 flasks (Corning Glassworks, Corning, NY) and cultured at 37°C in a 10% CO2/90% air atmosphere. The medium was changed daily. Cultures were studied in the late post-confluent period (21 days after seeding), when all cells display a mucin-secreting phenotype (Lesuffleur et al., 1990). Cell viability was determined by trypan blue dye exclusion. For metabolic labeling, cells were cultured as described above until day 20 and were then labeled for 24 hours with [3H]glucosamine (1.48 MBq/ml) in low-glucose Dulbecco’s minimum essential H-16 medium (Gibco, Gaithesburg, MD), with [35S]sulfate (1.48 MBq/ml) in Ham’s F-12 medium (Gibco), or with [3H]threonine (1.48 MBq/ml)

Electrophoresis and western blotting SDS-PAGE SDS-PAGE was performed with 2% to 5% gradient polyacrylamide gels (Laemmli, 1970). For autoradiogaphy, gels were fixed in 40% ethanol, 10% glycerol, 10% acetic acid (by vol.), soaked in Amplify (Amersham, UK) for 20 minutes, dried on Whatman paper, and exposed to Cronex 4 NIF film (Dupont). Agarose gel electrophoresis Mucin glycopeptides were obtained by Pronase digestion for 48 hours at 37°C in 0.01 M calcium acetate buffer, pH 7, at an enzyme/substrate ratio of 1/40 (w/w). Fresh enzyme was added after 24 hours of digestion. Agarose gel electrophoresis was then performed in veronal buffer at pH 8.2 and gels were stained with Schiff periodate (Marianne et al., 1986). Cellulose acetate electrophoresis After immersion of cellulose acetate plates in 0.05 M barium acetate, migration was carried out for 20 minutes. The plates were then stained

Biochemical characterization of HT-29 mucins 1277 in 0.02% DMB in 1% acetic acid for 10 minutes and destained in 10% acetic acid (Wesslet, 1968). Western blotting After separation by 2-5% SDS-PAGE, proteins were transferred to a nitrocellulose membrane as described by Vaessen et al. (1981). Membranes were incubated with 10 µg/ml peroxidase-labeled lectins (Sigma) in 10 mM Tris-glycine buffer containing 0.9% NaCl and 3% bovine serum albumin at 37°C for 2 hours. After washing, membranes were incubated with 4-chloro-1-naphthol (Sigma). Immunocytochemistry Indirect immunofluorescence was performed on cryostat sections of cell layer rolls as reported by Lesuffleur et al. (1990), with minor modifications. Sections were air-dried and fixed with acetone for 10 minutes. After washing with PBS, 5% normal horse serum was added. Sections were incubated with FITC-labeled soybean (SBA), peanut (PNA), and wheat germ (WGA) lectins for 60 minutes. After washing, sections were mounted with glycerol. Alternatively, mouse monoclonal antibodies (mAb), that detect a large panel of mucin-associated carbohydrate epitopes were used. mAb Cu-1, detecting Tn (Takahashi et al., 1988), was obtained from Dr S. I. Hakomori (The Biomembrane Institute, Seattle, WA) and mAbs B72.3 and CC49, detecting sialyl-Tn (Nuti et al., 1982), were obtained from Dr Kenneth O. Lloyd (Sloan-Kettering Institute, New York, NY). To obtain better semiquantitative data, several lectin and antibody dilutions were used in each experiment. Reactions were developed with goat anti-mouseTRlTC (Pierce, Rockford, IL) (5 µg/ml) and visualized using a Zeiss Axioscop equipped with Plan-Neofluar lenses and scored as follows: +++, when fluorescence signal was clearly seen at ×100 magnification; ++, when it was seen at ×200 magnification; +, when signal was seen only at ×400 magnification. Percentage reactive cells was expressed as an average of the whole section. Scoring was performed independently by two investigators (C. de B. and F.X.R.) and all experiments were repeated twice independently. Sialyltransferase (ST) assays Preparation of cell extracts Cells were lysed at 0°C with 10 mM sodium cacodylate buffer, pH 6.5, containing 1% Triton X-100, 20% glycerol, 0.5 mM dithiotreitol, and 5 mM MnCl2 (1 ml per 170 cm2 flask). After a 10 minute incubation under continuous stirring, cell homogenates were centrifuged at 10,000 g for 15 minutes and the supernatants were used for enzymatic assays. Protein concentration was determined according to the method described by Peterson (1977) using bovine serum albumin (BSA) as standard.

samples were centrifuged at 3,000 g for 5 minutes and supernatants were directly processed for descending paper chromatography in ethyl acetate/ pyridine/ water (10/4/3, by vol.) (Delannoy et al., 1993).

RESULTS Comparative analysis of mucins and proteoglycans from HT-29 MTX and HT-29 cells The secretion of mucins and proteoglycans by HT-29 MTX and HT-29 cells was first studied by metabolic labeling with [3H]glucosamine and with [35S]sulfate followed by analysis of culture medium by SDS-PAGE. After [3H]glucosamine labeling (Fig. 1A), a broad band was observed in the case of HT-29 MTX cells whereas HT-29 cells yielded a diffuse smear. This result is relevant to the high rate of mucins produced by HT-29 MTX cells and to the diffuse high molecular mass band described by Dahiya et al. (1992). SDSPAGE analysis of culture media from [35S]sulfate-labeled HT29 MTX and HT-29 cells showed diffuse smears with a mobility similar to that of [3H]glucosamine-labeled cells (Fig. 1B). Although [3H]glucosamine is predominantly incorporated into mucins and [35S]sulfate into glycosaminoglycans, it was not possible to distinguish between mucins and proteoglycans using SDS-PAGE. Mucins and proteoglycans synthesized by HT-29 MTX and HT-29 cells were isolated by ultracentrifugation through a CsBr gradient. Fig. 2A shows the absorbance profiles obtained after the addition of orcinol, to detect carbohydrates; or DMB, to detect glycosaminoglycans. AE-HPLC was then used to separate mucins and proteoglycans from selected fractions of HT-29 MTX (1.42-1.51 g/ml) and HT-29 (1.39-1.53 g/ml) cells and the elution profiles are shown in Fig. 2B. In the HT29 MTX fraction, orcinol reactivity eluted as a major peak with a retention time of 24 minutes, whereas DMB reactivity eluted later and showed a complex pattern of peaks at the 26-37

ST assay using asialofetuin as acceptor Cell extracts (40 µl) were carried to a final volume of 120 µl with 0.1 M sodium cacodylate, pH 6.5, 1% Triton X-100, 0.1% BSA, 0.2 M galactose (an inhibitor of β-galactosidase), 1 mM 2,3-dehydro-2deoxy-Neu5Ac (an inhibitor of sialidases), 52.9 µM CMP[14C]Neu5Ac (0.58 GBq/mmol, 3.67 kBq/120 µl), containing 480 µg of asialofetuin (prepared by mild acid hydrolysis of fetuin) as exogenous substrate, and incubated at 37°C for 1 hour. The reaction was stopped by adding 1 ml of ice-cold phosphotungstic acid (5% in 2 M HCl). The precipitate was collected on glass fiber filters, washed extensively with 5% trichloroacetic acid, distilled water and ethanol, and processed for scintillation counting (Vandamme et al., 1992). ST assay using Galβ1-3GalNAcα-O-benzyl as acceptor Incubations were performed under the conditions described above using 40 µl of cell extract or 1 mU of purified porcine liver Galβ13GalNAc α-2,3-ST (Boehringer, Mannheim, Germany) as enzyme source and Galβ1-3GalNAcα-O-benzyl (Sigma) (1 mM final concentration). The reaction was stopped by adding 1 volume of ethanol;

Fig. 1. SDS-PAGE autoradiogram of culture medium from HT-29 and HT-29MTX cells labeled with [3H]glucosamine (A) or [35S]sulfate (B). Lane 1, HT-29 cells; lane 2, HT-29 MTX cells. Equal amounts of protein were loaded for each cell type. Electrophoretic migration was compared with that of a 200 kDa molecular mass marker.

1278 G. Huet and others

Fig. 2. (A) Ultracentrifugation profiles of lysates from HT-29 MTX and HT-29 cells. (B) HPLC profiles of ultracentrifugation fractions from HT-29 MTX and HT-29 cells.

minute time periods. For HT-29 cells, a low orcinol reactivity was detected showing three small peaks at retention times of 4, 25, and 31 minutes, whereas, a high DMB reactivity was apparent at the 26-40 minute time points. Preparative HPLC chromatography was carried out for both cell types. Three and two fractions were collected from HT-29 MTX (top Fig. 2B, fraction A, 21-25 minutes; B1, 26-29 minutes; and B2, 30-37 minutes) and HT-29 cells (bottom Fig. 2B, fraction A, 21-27 minutes; B, 28-37 minutes), respectively. These fractions were analysed for amino acid (Table 1), carbohydrate, and sulfate composition (Table 2). HT-29 MTX fraction A showed a high carbohydrate/protein content (6.8 g sugar/g protein), high hydroxy amino acid content (40.05%) with a Thr/Ser ratio of 1.52, and a high proline content (11.83%), all typical features of mucins. Carbohydrate analysis revealed mainly GalNAc, Gal, GlcNAc and sialic acid. The sialic acid content was very high (1.3 molar ratio to GalNAc) whereas the sulfate content was very low (0.1 molar ratio to GalNAc). HT-29 MTX fractions B1 and B2 also showed very high hydroxy amino acid contents (34.49% and 31.51%, respectively) with lower Thr/Ser ratios (1.30 and 1.11, respectively), and lower Pro contents (9.73% and 8.62%, respectively). Other major amino acid changes consisted of a decrease in Arg and an increase in Cys, Glu/Gln, Gly and Tyr. The relative content of neutral monosaccharides was similar in

Table 1. Amino acid composition of HT-29 MTX and HT-29 HPLC fractions (%) HT-29 MTX HPLC fraction

Cys Asp Thr Ser Glu Pro Gly Ala Val Ile Leu Tyr Phe His Lys Arg

HT-29 HPLC fractions

A

B1

B2

A

B

0.73 4.80 24.18 15.87 7.17 11.83 6.62 7.12 5.18 2.40 3.11 0.08 2.20 1.55 2.59 4.57

1.96 6.76 19.53 14.96 10.88 9.73 8.77 7.26 4.44 2.21 3.13 1.24 1.86 1.55 2.96 2.75

2.47 6.90 16.61 14.90 11.32 8.62 11.14 7.26 4.46 2.37 3.66 1.47 2.12 1.57 2.35 2.76

0.51 10.17 8.98 7.85 11.23 6.22 10.15 10.48 6.21 3.72 6.48 2.45 3.36 1.69 5.14 5.37

0.97 8.55 5.77 9.72 11.77 4.45 31.69 7.77 4.32 2.22 4.23 2.46 1.52 4.58

fractions A, B1 and B2. In contrast, fractions B1 and B2 had lower relative sialic acid contents (0.9 and 0.5, respectively) and higher sulfate contents (1.2 and 0.6, respectively). These results suggest that fractions B1 and B2 correspond to more-

Biochemical characterization of HT-29 mucins 1279 Table 2. Carbohydrate composition and sulfate content of HT-29 MTX and HT-29 HPLC fractions HT-29 MTX HPLC fractions

Mannose Galactose GalNAc GlcNAc Sialic acid Sulfate Sulfate*

HT-29 HPLC fractions

A

B1

B2

A

B

1.4 =1 0.6 1.3 0.1 0.1

0.1 1.5 =1 0.5 0.9 1.2 2.6

0.2 1.1 =1 0.5 0.5 0.6 1.2

0.5 1 =1 1.2 0.3 2.1 1.7

ND ND ND + ND + 1.6

All results are expressed as molar ratios to GalNAc. *The sulfate content was also expressed as molar ratio to GlcNAc. ND, not detectable.

acidic mucins, mixed with proteoglycans, as predicted by the elution profile obtained after HPLC chromatography. HT-29 fraction A showed a relatively low carbohydrate/protein level (1.9 g sugar/g protein), a hydroxy amino acid content of 16.83% with a Thr/Ser ratio of 1.14, and a 6.22% Pro content. Carbohydrate analysis of this fraction revealed mainly mannose, Gal, GalNAc, GlcNAc and sialic acid. The sialic acid content was low (0.3 molar ratio to GalNAc) whereas the sulfate content was relatively higher (2.1 molar ratio to GalNAc). HT-29 fraction B had a hydroxy amino acid content of 15.49%, a Thr/Ser ratio of 0.59 and a high Gly content (31.69%). Carbohydrate analysis revealed only GlcNAc. Sulfate was also found in this fraction in a 1.6 molar ratio to GlcNAc. Therefore, the biochemical characteristics of HT-29 fraction B are those typical of a proteoglycan. The presence of mucins in fraction A from HT-29 MTX and HT-29 cells was further examined by Schiff periodate staining of Pronase digestion products fractionated by agarose gel electrophoresis: a fast migrating band which stained intensely was observed for HT-29 MTX cells whereas a weakly stained band was observed for HT-29 cells (Fig. 3). The electrophoretic behavior of proteoglycans from HT-29 MTX and HT-29 cells was analysed using cellulose acetate strips and coloration with DMB (Fig. 4). HT-29 MTX fraction B2 contained a proteoglycan with a slower migration than the heparan sulfate standard, whereas HT-29 fraction B contained a proteoglycan which migrated between the heparan sulfate and dermatan sulfate standards. Consequently, the main proteoglycan detected in HT-29 cells is more acidic than the HT29 MTX proteoglycan. Effect of BFA and GalNAcα-O-benzyl on mucin biosynthesis by HT-29 MTX cells The effects of BFA and GalNAcα-O-benzyl upon the biosynthesis and secretion of mucins were first studied through continuous labeling with [3H]threonine. In the presence of BFA, cell-associated and secreted radioactivity were 435.9% and 42.1% of those of control cultures, respectively. In the presence of GalNAcα-O-benzyl, cell-associated and secreted radioactivity were 428.9% and 49% of those of control cultures, respectively. Fig. 5 shows the electrophoretic pattern of cell lysates and culture media after metabolic labeling with [3H]threonine. A major band was observed in each cell lysate, the migration of which appeared slightly increased after treatment with BFA and slightly decreased after treatment with

Fig. 3. Agarose gel electrophoresis of mucin glycopeptides and staining with Schiff periodate. Fractions A from HT-29 MTX cells (lane 1) and from HT-29 cells (lane 2) were digested with Pronase for 48 hours at 37°C and mucin glycopeptides were loaded on agarose gel slides.

Fig. 4. Proteoglycan analysis by cellulose acetate electrophoresis and DMB staining. Fraction B from HT-29 cells (A) and fraction B2 from HT-29 MTX cells (B) were compared with several standard glycosaminoglycans. (A) Lane 1, chondroitin sulfate C; lane 2, HT29 fraction B; lane 3, chondroitin sulfate A; lane 4, dermatan sulfate; lane 5, heparan sulfate. (B) Lane 1, chondroitin sulfate A; lane 2, HT-29 MTX fraction B2; lane 3, chondroitin sulfate C; line 4, dermatan sulfate; lane 5, heparan sulfate.

GalNAcα-O-benzyl. SDS-PAGE analysis of culture media from BFA-treated and GalNAcα-O-benzyl-treated cultures revealed faint smears with mobility similar to that of the corresponding cell lysates. This finding was in contrast to the discrete band observed in culture medium from control HT-29 MTX cells. Mucins from HT-29 MTX cells cultured for 24 hours in the presence of BFA or GalNAcα-O-benzyl were isolated by ultracentrifugation and AE-HPLC. Analysis of the mucin fractions (Table 3) showed that BFA treatment led to a slight decrease in the relative content of Gal and sialic acid and to a 4-fold increase in sulfate content. GalNAcα-O-benzyl treatment induced a 13-fold decrease in the relative amount of sialic acid in HT-29 MTX fraction A. The composition of B fractions and the electrophoretic behavior of the HT-29 MTX proteoglycan were unchanged (data not shown). To further study the effect of BFA and GalNAcα-O-benzyl on HT-29 MTX mucins, their reactivity with a panel of mAbs and lectins was examined by immunofluorescence on frozen sections of cell rolls (Fig. 6 and Table 4), or by western blotting (Fig. 7) on purified mucin fractions. BFA treatment did not lead to major changes in the expression of mucin-associated

1280 G. Huet and others

Fig. 5. SDS-PAGE autoradiogram of HT-29 MTX cell lysates (A) and culture media (B) after labeling with [3H]threonine. Lane 1, control HT-29 MTX cells; lane 2, HT-29 MTX cells cultured in the presence of GalNAcαO-benzyl for 24 hours; lane 3, HT-29 MTX cells cultured in the presence of BFA for 24 hours. Equal amounts of protein/sample were loaded.

carbohydrate epitopes. GalNAcα-O-benzyl treatment resulted in an increase in the expression of Tn and T antigens as detected with SBA and mAb Cu-l (Tn antigen) or PNA (Fig. 6 and Table 4). No changes in the reactivity with WGA were observed (not shown). Western blotting particularly showed a great increase in the T antigen reactivity of mucins obtained after GalNAcα-O-benzyl treatment (Fig. 7). To determine the mechanism(s) by which GalNAcα-Obenzyl treatment induced such a decrease in sialic acid content and the concomitant increase in T antigen in HT-29 MTX mucins, ST assays were performed. First, ST activity was measured using asialofetuin as an in vitro substrate: a 30% Table 3. Carbohydrate composition and sulfate content of HT-29 MTX HPLC fraction A obtained after treatment with brefeldin A or GalNAcα-O-benzyl

Galactose GalNAc GlcNAc Sialic acid Sulfate

HT-29 MTX fraction A of control HT-29 MTX cells

HT-29 MTX fraction A obtained after treatment with brefeldin A

HT-29 MTX fraction A obtained after treatment with GalNAcα-O-benzyl

1.4 =1 0.6 1.3 0.1

1.1 =1 0.6 1.1 0.4

1.4 =1 0.5 0.1 0.2

All results are expressed as molar ratios to GalNAc.

decrease in the transfer of [14C]sialic acid was detected in GalNAcα-O-benzyl-treated cells (Fig. 8A). GalNAcα-O-benzyl cannot be used as a substrate by sialyltransferases (Van den Eijden and Joziasse, 1993). On the other hand, Galβ1-3GalNAcα-O-benzyl is a specific substrate for Galβ1-3GalNAcα-2,3-sialyltransferase (α-2,3-ST(O), EC 2.4.99.4) and would therefore be able to compete for sialylation of asialofetuin in an in vitro assay. The ability of Galβ1-3GalNAcα-O-benzyl and GalNAcαO-benzyl to inhibit the ST activity using asialofetuin as a substrate was examined. As shown in Fig. 8B, Galβ13GalNAcα-O-benzyl was an effective inhibitor at a concentration of 100 µM, whereas GalNAcα-O-benzyl was not inhibitory. These results fitted well the hypothesis proposed by Huang et al. (1992), that GalNAcα-O-benzyl could be converted in vivo into Galβ1-3GalNAcα-O-benzyl and compete for the elongation of T antigen in HM 7 colon cancer cells. We then investigated the in situ synthesis of Galβ13GalNAcα-O-benzyl in GalNAcα-O-benzyl-treated HT-29 MTX cells. When the in vitro α-2,3-ST(O) assay was performed with homogenates from GalNAcα-O-benzyl-treated cells in the absence of the exogenous acceptor Galβ13GalNAcα-O-benzyl, a sialylated product that was absent from control homogenates was detected. This sialylated product comigrated with NeuAcα2-3Galβl-3GalNAcα-Obenzyl obtained by incubation of the Galβ1-3GalNAcα-O-

Table 4. Immunofluorescence on cryostat sections of HT-29 MTX cells untreated, or treated with brefeldin A or GalNAcα-O-benzyl % of labeled cells

Lectins Antibodies

SBA, 20 µg/ml PNA, 20 µg/ml Cu-1 B72.3 CC.49

Specificity

Control HT-29 MTX cells (%)

HT-29 MTX cells treated with brefeldin A (%)

HT-29 MTX cells treated with GalNAcα-O-benzyl (%)

Tn T Tn Sialyl-Tn Sialyl-Tn

95% undifferentiated cells and a low proportion of cells producing mucins of gastric as well as colonic immunological specificity. HT-29 MTX mucins barely enter 2% to 5% gradient polyacrylamide gels, in accordance with their formation of a secreted mucous layer (Lesuffleur et al., 1990). The major mucin fraction isolated by AE-HPLC is characterized by a high hydroxy amino acid content and a Thr/Ser of 1.52. Prior studies have shown that MUC1, MUC2, MUC3 and MUC5AC are expressed in HT-29 MTX at the mRNA level, and the amino acid composition of the major mucin fraction isolated is in agreement with the amino acid sequences of the tandem repeats of these genes (Gendler et al., 1990; Gum et al., 1990, 1992; Porchet et al., 1991; Guyonnet Duperat et al., 1994). The carbohydrate composition of this fraction revealed a high GalNAc and sialic acid content, and a low GlcNAc and sulfate content. These characteristics are similar to those described for the cl.16-E mucin-secreting clone isolated from HT-29 cells after treatment with sodium butyrate (Augeron and Laboisse, 1984). The analysis of mucin-containing fractions from HT-29 cells revealed a lower amount of mucins, related to the small proportion of mucin-producing cells in this population (less than 0.5%) (Augeron and Laboisse, 1984; Lesuffleur et al., 1990, 1991), as well as less glycosylated mucins. Furthermore, the composition of HT-29 mucins was different from that of HT29 MTX mucins in having: (1) lower hydroxy amino acid and

Fig. 8. Effect of GalNAcα-O-benzyl and Galβ1-3GalNAcα-Obenzyl on the transfer of Neu5Ac on asialofetuin. (A) Homogenates of HT-29 MTX cells cultured with (j) or without (h) GalNAcα-Obenzyl were incubated with asialofetuin and the transfer of Neu5Ac was comparatively measured. Results are expressed as nmol of [14C]Neu5Ac transferred/mg of protein (mean values of two separate experiments). (B) Homogenates of HT-29 MTX cells were incubated with asialofetuin in the presence of GalNAcα-O-benzyl (j) or Galβ1-3GalNAcα-O-benzyl (h) at different concentrations. Results are expressed as the percentage of remaining activity.

proline content; (2) higher GlcNAc/GalNAc ratio; (3) lower sialic acid content; and (4) higher sulfate content. The distinct compositional features of HT-29 MTX and HT-29 mucins may reflect differences in expression of mucin genes (Lesuffleur et al., 1993) or be related to the altered compartmentalization of resident Golgi proteins in HT-29 MTX cells (Egea et al., 1993). Besides, the HT-29 cell line likely contains mucin-secreting cells of different specificity, as shown using rabbit antisera detecting gastric or colonic mucins (Lesuffleur et al., 1990). Current efforts aim at characterizing the structure of carbohydrate chains of HT-29 MTX and HT-29 mucins in an attempt to analyse the relationship between oligosaccharide chains and the apomucin carrying them. The study of high molecular mass glycoproteins from these two cell populations has also shown differences in the structure of proteoglycans, although this analysis is hampered by the high amount of mucins synthesized by HT-29 MTX. Indeed, a certain level of mixing of these two types of molecules is almost inevitable. In any case, the proteoglycans of HT-29

c.p.m. (×10−3)

c.p.m. (×10−3)

Biochemical characterization of HT-29 mucins 1283

Fig. 9. Detection of Galβ1-3GalNAcα-O-benzyl in the homogenates of HT-29 MTX cells treated with 5 mM of GalNAcα-O-benzyl. (A) Homogenates of cells cultured with (j) or without (h) GalNAcα-O-benzyl were incubated in standard conditions of α-2,3ST(O) assay without exogenous acceptor, and sialylated products were detected as described in Materials and Methods. 1, CMPNeu5Ac; 2, Neu5Ac; 3, the arrowed bar indicates the position of Neu5Acα2-3Galβ1-3GalNAcα-O-benzyl obtained by the sialylation of Galβ1-3GalNAcα-O-benzyl with porcine liver α-2,3-ST(O). (B,C) Homogenates of cells cultured with GalNAcα-O-benzyl were denatured by heating and used as substrate source in α-2,3-ST(O) assay. Homogenates from control cells (B) or porcine liver α-2,3ST(O) (C) were used as enzyme sources. Sialylated products were detected as described in Materials and Methods.

MTX and HT-29 cells differ with regard to their electrophoretic behavior: the less-acidic nature of HT-29 MTX proteoglycans might result from a larger amount of unsulfated heparan chains. Such structures were shown to coexist with sulfated heparan chains on the same protein core of the heparan sulfate proteoglycan (HSPG) from the colon carcinoma cell line WiDr (Iozzo, 1989), a cell line that has been reported by the ATCC as possibly being the same as HT-29. The studies of high Mr HSPG produced by different cell types have suggested that a common precursor protein core could undergo cell-specific processing, thus generating different forms of proteoglycans (Murdoch et al., 1992). The data from mucin and

proteoglycan analysis suggest that the degree of sulfatation of these molecules may be modulated by the state of differentiation of HT-29 cells. To examine mucin glycosylation in greater detail, we have analysed the effect of BFA and GalNAcα-O-benzyl on the structure of mucins synthesized by HT-29 MTX cells. BFA, which blocks transport from the RER to the Golgi apparatus and induces the redistribution of resident Golgi proteins to the RER (Lippincott-Schwartz et al., 1989), induced an increase in sulfate content and an increase in the electrophoretic mobility of mucins in SDS-PAGE, events that are possibly related. Indeed, the high carbohydrate content of mucins may prevent effective binding of SDS to the protein core and the intrinsic negative charge of mucin molecules could influence their migration, as described for the MUC1 gene product (Hilkens and Buijs, 1988). The mechanisms by which these changes occur are not known, but our findings suggest that compartmentalization of sulfotransferases is affected by brefeldin A. GalNAcα-O-benzyl induced an increased expression of T antigen and a marked decrease in the sialylation of HT-29 MTX mucins. These changes were accompanied by a decrease in the electrophoretic mobility of mucins by SDS-PAGE, possibly due to the synthesis of less-acidic mucins. Indeed the extensively sialylated form of MUC1 has a higher mobility in SDS-PAGE than the incompletely sialylated premature form (Hilkens and Buijs, 1988). GalNAcα-O-benzyl treatment of the HM7 high-mucin variant of LS174T colon cancer cells also led to an increase in T antigen reactivity. This result, now reproduced in HT-29 MTX cells, was unexpected as GalNAcα-O-benzyl behaves as an in vivo competitive inhibitor of the galactosyltransferase responsible for the synthesis of T-antigen (Huang et al., 1992). One of the hypotheses proposed by these authors was the in situ generation of the disaccharide Galβ1-3GalNAcα-O-benzyl, which would behave as an efficient inhibitor of the elongation of T antigen. This hypothesis would be in agreement with the decreased sialylation of HT-29 MTX mucins observed in our study. Galβ1-3GalNAcα-O-benzyl is a potential substrate for α-2,3-ST(O) but also for Galβ1-3GalNAc-R β-1,6-N-acetylglucosaminyltransferase (EC 2.4.1.102) and sulfotransferases. Thus changes in the incorporation of GlcNAc and sulfate might also be expected. Our findings indicate that incorporation of GlcNAc and sulfate was unaffected, whereas sialic acid incorporation into mucin was almost completely inhibited. In fact, using asialofetuin as substrate, we have demonstrated that the ST activity in GalNAcα-O-benzyl-treated homogenates is decreased and that Galβ1-3GalNAcα-O-benzyl is a powerful competitive inhibitor of the ST activity of HT-29 MTX homogenates. Among the different ST, only the Galβ13GalNAc α-2,3-ST can use Galβ1-3GalNAcα-O-benzyl as substrate. These observations suggest that α-2,3-ST(O) is the predominant ST expressed in HT-29 MTX cells, a finding that would be in agreement with the fact that NeuAcα2-3Galβ13GalNAc-O is a major carbohydrate component of the mucins synthesized by HT-29 MTX cells (unpublished observations) as well as of the mucins from HT-29-derived cl. 16E (Capon et al., 1992). The marked inhibition of sialic acid incorporation into mucins may be related to the striking inhibition of mucin secretion in GalNAcα-O-benzyl-treated cells. As mentioned above, GalNAcα-O-benzyl treatment did not affect incorporation of GlcNAc or sulfate into mucins, sug-

1284 G. Huet and others gesting that the subcellular compartment in which the enzymes responsible for this transfer are located is different from the compartment where α-2,3-ST(O) is present. This possibility would be in agreement with the concept that O-glycosylation proceeds in a stepwise manner according to the subcompartmentalization of the enzymes that participate in it (Egea et al., 1993). In fact, sialylation and sulfation of the rat mammary sialomucin have been reported to occur in separate compartments (Carraway and Hull, 1989). GalNAcα-O-benzyl may be of use in dissecting these steps of mucin maturation. Further studies may help to determine the influence of the compartmentalization of ST and sulfotransferases upon the post-translational processing of these molecules. Taken together our findings indicate that sialylation and sulfatation of mucins and proteoglycans are important post-translational steps that may be modulated by the state of differentiation of HT-29-derived cell populations. Furthermore, aryl-N-acetyl-α-galactosaminides can impair the action of α2,3-ST(O) in the O-glycosylation process of glycoproteins and this effect may be useful in investigating the activity of this enzyme and the biological effects of its products. This work was supported in part by grant no. 2209 from the Association pour la Recherche sur le Cancer and grant 93/1228 from the Fondo de Investigaciones Sanitarias (Madrid, Spain). The authors thank Drs S. I. Hakomori and K. O. Lloyd for providing antibodies, and Mrs E. Delaleau, M. T. Maillard and F. Roussez for typing the manuscript.

REFERENCES Augeron, C. and Laboisse, C. L. (1984). Emergence of permanently differentiated cell clones in a human colonic cancer cell line in culture after treatment with sodium butyrate. Cancer Res. 44, 3961-3969. Bobek, L. A., Tsai, H., Biesbrock, A. R and Levine, M. J. (1993). Molecular cloning, sequence, and specificity of expression the gene encoding the low molecular weight human salivary mucin (MUC7). J. Biol. Chem. 268, 20563-20569. Capon, C., Laboisse, C. L., Wieruszeski, J. M., Maoret, J. J., Augeron, C. and Fournet, B. (1992). Oligosaccharide structures of mucins secreted by the human colonic cancer cell line Cl. 16E. J. Biol. Chem. 267, 19248-19257. Carraway, K. L. and Hull, S. R. (1989). O-glycosylation pathway for mucintype glycoproteins. BioEssays 10, 117-121. Dahiya, R., Lesuffleur, T., Kwak, K. S., Byrd, J. C., Barbat, A., Zweibaum, A. and Kim, Y. S. (1992). Expression and characterization of mucins associated with the resistance to methotrexate of human colonic adenocarcinoma cell line HT-29. Cancer Res. 52, 4655-4662. De Jong, J. G. N., Wevers, R. A. and Sambeek, R. L. V. (1992). Measuring urinary glycosaminoglycans in the presence of protein: an improved screening procedure for mucopolysaccharidoses based on dimethylmethylene blue. Clin. Chem. 38, 803-807. Delannoy, P., Pelczar, H., Vandamme, V. and Verbert, A. (1993). Sialyltransferase activity in FR3T3 cells transformed with ras oncogene: decreased CMP-NeuAc: Galβ1-3GalNAc-α-2,3-sialyltransferase. Glycoconjugate J. 10, 91-98. Devine, P. L. and Mckenzie, I. F. C. (1992). Mucins: structure, function, and associations with malignancy. BioEssays 14, 619-625. Egea, G., Franci, C., Gambus, G., Lesuffleur, T., Zweibaum, A. and Real, F. X. (1993). Cis-Golgi resident proteins and O-glycans abnormally compartmentalized in the RER of colon cancer cells. J. Cell Sci. 105, 819830. Gendler, S. J., Lancaster, C. A., Taylor-Papadimitriou, J., Duhig, T., Peat, N., Burchell, J., Pemberton, L., Lalari, E. and Wilson, D. (1990). Molecular cloning and expression of the human tumour-associated polymorphic epithelial mucin, PEM. Biochem. J. 265, 15286-15293. Gum, J. R., Hicks, J. W., Swallow, D. M., Lagace, R. L., Byrd, J. C., Lamport, D. T. A., Siddiki, B. and Kim, Y. S. (1990). Molecular cloning of

cDNAs derived from a novel human intestinal mucin gene. Biochem. Biophys. Res. Commun. 171, 407-415. Gum, J. R., Hicks, J. W., Toribara, N. N., Rothe, E. M., Lagace, R. E. and Kim, Y. S. (1992). The human MUC2 intestinal mucin has cysteine-rich subdomains located both upstream and downstream of its central repetitive region. J. Biol. Chem. 267, 21375-21383. Guyonnet Duperat, V., Audié, J. P., Debailleul, V., Laine, A., Buisine, M. P., Zouitina-Galiègue, S., Pigny, P., Degand, P., Aubert, J. P. and Porchet, N. (1995). Characterization of the human mucin gene MUC5AC: a consensus cysteine-rich domain for 11p15 mucin genes. Biochem. J. (in press). Hardingham, T. E. and Fosang, A. J. (1992). Proteoglycans: many forms and many functions. FASEB J. 6, 861-870. Hilkens, J. and Buijs, F. (1988). Biosynthesis of MAM-6, an epithelial sialomucin. J. Biol. Chem. 263, 4215-4222. Houdret, N., Perini, J. M., Galabert, C., Scharfman, A., Humbert, P., Lamblin, G. and Roussel, P. (1986). The high lipid content of respiratory mucins in cystic fibrosis is related to infection. Biochim. Biophys. Acta 880, 54-61. Huang, J., Byrd, J. C., Yoor, W. H. and Kim, Y. S. (1992). Effect of benzylα-GalNAc, an inhibitor of mucin glycosylation, on cancer-associated antigens, in human colon cancer cells. Oncol. Res. 4, 507-515. Iozzo, R. V. and Wight, T. (1982). Isolation and characterization of proteoglycans synthesized by human colon and colon carcinoma. J. Biol. Chem. 257, 11135-11144. Iozzo, R. (1989). Presence of unsulfated heparan chains on the heparan sulfate proteoglycan of human colon carcinoma cells. J. Biol. Chem. 264, 26902699. Kim, Y. S. (1992). Altered glycosylation of mucin glycoproteins in colonic neoplasia. J. Cell. Biochem. Suppl. 16G, 91-96. Kuan, S. F., Byrd, J. C., Basbaum, C. and Kim, Y. (1989). Inhibition of mucin synthesis by aryl-N-acetyl-α-D-galactosaminides in human colon cancer cells. J. Biol. Chem. 264, 19271-19277. Kurosaka, A., Nakajima, H., Funakoshi, I., Matsuyama, M., Nogayo, T. and Yamashina, I. (1983). Structures of the major oligosaccharides from a human rectal adenocarcinoma glycoprotein. J. Biol. Chem. 258, 1159411598. Laemmli, U. K. (1970). Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680-681. Lamblin, G., Boersma, A., Klein, A., Roussel, P., Van Halbeek, H. and Vliegenthart, J. F. G. (1984). Primary structure determination of five sialylated oligosaccharides derivated from bronchial mucus glycoproteins of patients suffering from cystic fibrosis. J. Biol. Chem. 259, 9051-9058. Lesuffleur, T., Barbat, A., Dussaulx, E. and Zweibaum, A. (1990), Growth adaptation to methotrexate of HT-29 human colon carcinoma cells is associated with their ability to differentiate into columnar absorptive and mucus-secreting cells. Cancer Res. 50, 6334-6343. Lesuffleur, T., Kornowski, A., Luccioni, C., Muleris, M., Barbat, A., Beaumatin, J., Dussaulx, E., Dutrillaux, B. and Zweibaum, A. (1991). Adaptation to 5-fluorouracil of the heterogeneous human colon tumor cell line HT-29 results in the selection of cells committed to diffentiation. Int. J. Cancer 49, 721-730. Lesuffleur, T., Porchet, N., Aubert, J. P., Swallow, D., Gum, J. R., Kim, Y. S., Real, F. X. and Zweibaum, A. (1993). Differential expression of the human mucin genes MUC1 to MUC5 in relation to growth and differentiation of different mucus-secreting HT-29 cell subpopulations. J. Cell Sci. 106, 771-783. Lippincott-Schwartz, J., Yvan, L. C., Bonifacino, J. S. and Klausner, R. D. (1989). Rapid redistribution of Golgi proteins into the ER in cells treated with brefeldin A: evidence for membrane cycling from Golgi to ER. Cell 56, 801813. Lo-Guidice, J. M., Wieruszeski, J. M., Lemoine, J., Verbert, A., Roussel, P. and Lamblin, G. (1994). Sialylation and sulfatation of the carbohydrate chains in respiratory mucins from a patient with cystic fibrosis. J. Biol. Chem. 269, 18794-18813. Maoret, J., Font, J., Augeron, C., Codogno, P., Bauvy, C., Aubery, M. and Laboisse, C. L. (1989). A mucus-secreting human colonic cancer cell line. Biochem. J. 258, 793-799. Marianne, T., Périni, J. M., Houvenaghel, M. C., Tramu, G., Lamblin, G. and Roussel, P. (1986). Action of trifluoromethanesulfonic acid on highly glycosylated regions of human bronchial mucins. Carbohydr. Res. 151, 7-19. Murdoch, A. D., Dodge, G. R., Cohen, I., Tuan, R. S. and Iozzo, R. V. (1992). Primary structure of the human heparan sulfate proteoglycan from basement membrane (HSPG2/perlecan). J. Biol. Chem. 267, 8544-8557.

Biochemical characterization of HT-29 mucins 1285 Nuti, M., Teramoto, Y. A., Mariani-Constantini, R., Horand Hand, P., Colcher, D. and Schlom, J. (1982). A monoclonal antibody (B72. 3) defines patterns of distribution of a novel tumor-associated antigen in human mammary carcinoma cell populations. Int. J. Cancer 29, 539-545. Peterson, G. L. (1977). Modified procedure for protein assay. Anal. Biochem. 83, 346-356. Pinto, M., Appay, M. D., Simon-Assmann, P., Chevalier, G., Drocopoli, N., Fogh, J. and Zweibaum, A. (1982). Enterocytic differentiation of cultured human colon cancer cells by replacement of glucose by galactose in the medium. Biol. Cell 44, 193-196. Porchet, N., Nguyen, V. C., Dufossé, J., Audié, J. P., Guyonnet-Duperat, V., Gross, M. S., Denis C., Degand, P., Bernheim, A. and Aubert, J. P. (1991). Molecular cloning and chromosomal localization of a novel human tracheobronchial mucin cDNA containing tandemly repeated sequences of 48 base pairs. Biochem. Biophys. Res. Commun. 175, 414-422. Takahashi, H. K., Metoki, R. and Hakomori, S. I. (1988). Immunoglobulin G3 monoclonal antibody directed to Tn antigen (tumor associated α-Nacetyl-galactosaminyl epitope) that does not cross-react with blood group A antigen. Cancer Res. 48, 4361-4367.

Toribara, N. W., Robertson, A. M., Ho, S. B., Kuo, W. L., Gum, E., Hicks, J. W., Gum, J. R., Byrd, J. C., Siddiki, B. and Kim, Y. S. (1993). Human gastric mucin. J. Biol. Chem. 268, 5879-5885. Vaessen, R. T. M. J., Kreike, J. and Groot, G. S. P. (1981). Protein transfer to nitrocellulose filters. FEBS Lett. 124, 193-196. Vandamme, V., Cazlaris, H., Le Marer, N., Laudet, V., Lagrou, C., Verbert, A. and Delannoy, P. (1992). Comparison of sialyl- and β1-3galactosyltransferase activity in NIH3T3 cells transformed with ras oncogene: increase β-galactoside-2,6-sialyltransferase. Biochimie 74, 89100. Van den Eijden, D. H. and Joziasse, D. H. (1993). Enzymes associated with glycosylation. Curr. Opin. Struct. Biol. 3, 711-721. Wesslet, E. (1968). Analytical and preparative separation of acidic glycosamino-glycan by electrophoresis in baryum acetate. Anal. Biochem. 26, 439-444.

(Received 27 July 1994 - Accepted 15 November 1994)