A high-density putative monomeric mucin is the major [35S]labelled macromolecular product of human colorectal mucins in organ culture

Biochimie 85 (2003) 381–390 www.elsevier.com/locate/biochi

A high-density putative monomeric mucin is the major [35S]labelled macromolecular product of human colorectal mucins in organ culture Annkatrin Herrmann a, Ingemar Carlstedt a,*, Tarek Shirazi b, Robert Longman b, Anthony Corfield b b

a Mucosal Biology Group, Department of Cell and Molecular Biology, Lund University, BMC C13, S-221 84 Lund, Sweden Mucin Research Group, Dorothy Crowfoot Hodgkin labs, Division of Medicine, Bristol Royal Infirmary, Bristol BS2 8HW, UK

Received 4 November 2002; accepted 17 February 2003 This paper is dedicated to the memoryof André Verbert and Jean Agneray.

Abstract We have studied the biosynthesis of mucins in organ cultures of human colon using isopycnic density-gradient centrifugation following pulse labelling with [35S]sulphate and [3H]-D-glucosamine. A high-density [35S]sulphate labelled component, of larger size than MUC2 monomers, appeared in the tissue and also in the medium. It was not degraded by reduction, trypsin digestion, digestion with chondroitin ABC lyase or heparan sulphate III lyase, but was cleaved into smaller fragments following alkaline borohydride treatment and appears to be a monomeric, mucin-like molecule containing a protease-resistant domain with a larger hydrodynamic volume than MUC2 monomers. Although this macromolecule incorporated much more radiolabel than MUC2, it was not detected using chemical analysis and thus appears to be a component with a high metabolic turnover present in a very small amount. Most of the [3H]-D-glucosamine label was associated with low-density material that was well separated from MUC2, which was poorly labelled. Most of MUC2 was associated with the tissue as an ‘insoluble’ complex. The amount of MUC2 remained constant and its associated radiolabel increased only slightly with time. Analysis of the MUC2 subunits from the reduced ‘insoluble’ complex showed the typical reduction-insensitive oligomers and confirmed that the radiolabel was associated with this mucin. The large size of the [35S]-labelled putative monomeric mucin makes it difficult to separate it from reduced insoluble complex MUC2. As a result, many studies of intestinal mucin synthesis and secretion in the past have most likely been performed on ‘mixtures’ of this mucin and MUC2 and are thus not possible to interpret as the metabolic behaviour of oligomeric mucins. © 2003 Éditions scientifiques et médicales Elsevier SAS and Société française de biochimie et biologie moléculaire. All rights reserved. Keywords: Organ culture; Human colon; MUC2; Mucin

1. Introduction The mucus gel covering the colonic surface protects the mucosa against pathogens, other harmful agents and mechanical stress. The mucus gel is formed by secreted highmolecular mass glycoproteins referred to as mucins. To date, eleven different mucin genes have been shown to be expressed in the human colon namely MUC1, MUC2, MUC3A, MUC3B, MUC4, MUC5B, MUC6, MUC11,

Abbreviations: DFP, di-isopropyl phosphofluoridate; DTT, dithiothreitol; FBS, foetal bovine serum; NEM, N-ethylmaleimide; PMSF, phenylmethylsulphonyl fluoride. * Corresponding author. Tel.: +46-46-222-46-39; fax: +46-46-222-31-28. E-mail address: [email protected] (I. Carlstedt).

MUC12, MUC13 and MUC17 [1–9]. MUC2, MUC5B and MUC6 are secreted mucins, whereas most of the others are likely to be associated with the luminal plasma membrane. MUC2 is the major gel-forming mucin in human colon and more than 90% of MUC2 in this tissue occurs as an ‘insoluble’ glycoprotein complex formed by subunits joined by disulphide bonds, as well as by unidentified, reductioninsensitive linkages [10,11]. Metabolic labelling of human colonic tissue and cells derived from colon has been used to study MUC2 synthesis and also to investigate changes in MUC2 secretion and sulphation in disease [12,13]. Synthesis and assembly of large oligomeric MUC2 may take several days [14], suggesting that studies based on metabolic labelling should be carried out over long time periods to ascertain the incorporation of the radioactive precursors. Due to its insolubility, the major

© 2003 Éditions scientifiques et médicales Elsevier SAS and Société française de biochimie et biologie moléculaire. All rights reserved. DOI: 1 0 . 1 0 1 6 / S 0 3 0 0 - 9 0 8 4 ( 0 3 ) 0 0 0 6 4 - 6

382

A. Herrmann et al. / Biochimie 85 (2003) 381–390

part of this mucin may be lost in the preparation of mucus glycoproteins using classic techniques. Lack of separation of the material into a soluble and insoluble fraction or discarding the insoluble fraction as cell debris without further analysis accounts for these losses. In a whole animal model of cat trachea [15], as well as bovine trachea in organ culture [16] and primary cell culture [17], a significant fraction of the radiolabel was found to be present in ‘monomeric’ mucinlike molecules in addition to the oligomeric mucus-forming ones. In the present study, we have investigated radiolabelled glycoconjugates associated with the tissue and secreted into the medium after metabolic labelling of human colon in organ culture for 6, 12 and 24 h. An increase in [35S]sulphate and [3H]-D-glucosamine labelled material with time was observed within the tissue, as well as an increase of [3H]-Dglucosamine containing material secreted into the medium. Most of the incorporated [35S]sulphate was present in a mucin-like, high-Mr, apparently monomeric molecule with a density considerably higher than that of MUC2 whereas most of the [3H]-D-glucosamine appeared in non-mucin material of low molecular mass and density. Almost all MUC2 was confined to the ‘insoluble’ complex in the tissue and even after 24 h it was poorly labelled. The identification here of a monomeric mucin, distinct from MUC2, demonstrates that more than one mucin component is labelled when radiolabelling studies are performed. Our studies imply that careful interpretation is required when using metabolic labelling techniques to examine oligomeric mucins. 2. Experimental procedures 2.1. Materials Iodoacetamide, N-ethylmaleimide (NEM), trypsin (EC 3.4.21.4, type XIII, treated with L-(1-tosylamido-2-phenyl) ethyl chloromethyl ketone) and gentamycin sulphate were from Sigma Chemical Co. Chondroitin ABC lyase (EC 4.2.2.4) and heparan sulphate III lyase (EC 4.2.2.89) were purchased from Seikagaku Kogyo Co. Di-isopropyl phosphofluoridate (DFP) as well as guanidinium chloride (practical grade) were obtained from Fluka, and stock solutions of guanidinium chloride (ca. 8 M) were treated with charcoal and filtered through a PM 10 filter (Amicon) before use. Dithiothreitol (DTT) was from Merck. Sepharose CL-2B, Superose 6 HR 10/30 column, the ECL western detection kit as well as sodium [35S]sulphate and [3H]-D-glucosamine were from Amersham Biotech. Biotin hydrazide was from Vector Laboratories Inc. and alkaline phosphatase conjugated streptavidin from Boehringer Mannheim. Immobilon-P polyvinylidene difluoride membranes were from Millipore. Eagles minimal medium, glutamine, streptomycin, penicillin, foetal bovine serum (FBS) and ultrapure agarose were purchased from Life Technologies Inc. The LUM2-3 polyclonal antiserum [10] was used to identify human MUC2. Horseradish peroxidase and alkaline phos-

phatase conjugated swine anti-rabbit antibodies were bought from Dako (Glostrup, Denmark). All other chemicals were of A.R. grade or equivalent. 2.2. Analytical methods The density of fractions from density-gradients was measured using a Carlsberg pipette as a pycnometer. Carbohydrate was determined as periodate oxidisable structures based on the method described by Devine [18] as described by Wickström et al. [19]. Antibody reactivity was investigated with ELISA, essentially as described previously [10], with the LUM2-3 polyclonal antiserum (1:1000). Unreduced samples for MUC2 analysis were first treated with 10 mM DTT in dilution plates by adding 10 µl of a 100 mM solution in 6 M guanidinium chloride/1 M Tris-HCl buffer, pH 8.0 containing 5 mM sodium-EDTA to 100 µl of sample. After incubation (1 h in 37 °C), alkylation was performed (10 µl of 250 mM iodoacetamide in 4 M guanidinium chloride was added) for 1 h at room temperature and the samples were transferred to multiwell assay plates for ELISA analysis. Aliquots (50 µl or 1 ml) were mixed with 5 or 10 ml, respectively, of scintillation fluid (Ready safe; Beckman Instruments) and radioactivity measured using an LKB 1214 b-scintillation counter. 2.3. Tissue culture Human colonic tissue was obtained from the resection margin at cancer surgery. The mucosa was dissected from the muscularis layer and cut into pieces with approximately 2 × 2 mm surface area. Tissue pieces were placed in organ culture as described before [12] and incubated with sodium [35S]sulphate (0.923 MBq per dish) and [3H]-D-glucosamine (0.37 MBq per dish) for 6, 12 or 24 h at 37 °C in an atmosphere of 95% O2 and 5% CO2. Cultures were carried out with tissue from three patients using six pieces per dish. In a separate experiment, five dishes, each containing six pieces of tissue, were prepared and cultured for 23 h to provide material for further analysis. 2.4. Isolation of radiolabelled mucins from the culture medium and tissue After 6, 12 and 24 h of culture, the medium was collected, the tissue washed twice with PBS (1 ml) containing, 1 mM phenylmethylsulphonyl fluoride (PMSF), 5 mM sodiumEDTA, soybean trypsin inhibitor (0.1 mg/ml), 5 mM NEM, 10 mM benzamidine hydrochloride, 0.2% sodium azide and the washings added to the initial medium. The supporting lens tissue (referred to as ‘support’) and the tissue were collected separately. Medium, tissue and ‘support’ were kept frozen until analysed. Tissue and ‘support’ samples were thawed in 6 M guanidinium chloride/10 mM sodium phosphate buffer, pH 6.5, containing 5 mM sodium-EDTA, 5 mM NEM and 1 mM

A. Herrmann et al. / Biochimie 85 (2003) 381–390

DFP (extraction buffer), dispersed gently with a small plastic pestle and shaken slowly overnight in the cold. After centrifugation (Beckman J2-MC centrifuge; JA 20 rotor 18 000 rpm; 4 °C; 60 min), soluble material was removed and pellets re-extracted twice as above. The final extraction residues (‘insoluble’ material) were solubilised by reduction with 10 mM DTT in 6 M guanidinium chloride/0.1 M TrisHCl buffer, pH 8.0, containing 5 mM sodium-EDTA (reduction buffer) at 37 °C for 5 h, followed by alkylation with 25 mM iodoacetamide in the same buffer overnight at room temperature in the dark. After centrifugation as above, supernatants were retained. The frozen samples of the medium were thawed and mixed with an equal volume of extraction buffer. After dialysis (six changes of the same buffer), samples were centrifuged to separate material that was soluble in guanidinium chloride from a putative ‘insoluble’ fraction, however, no such material could be detected by the naked eye. To ascertain that no ‘insoluble’ mucins were indeed present, reduction buffer (2 ml) containing 10 mM DTT was added to the tubes which were then incubated at 37 °C for 5 h and treated with iodoacetamide (final concentration 25 mM) overnight at room temperature in the dark. After centrifugation as above the supernatant was retained. 2.5. Purification of mucins ‘Soluble’ and ‘insoluble’ material (the latter being solubilised with reduction) from the medium, tissue and ‘support’ were dialysed against 2-3 volumes of 50 mM sodium sulphate followed by dialysis against 6 M guanidinium chloride/10 M sodium phosphate, pH 6.5 containing 5 mM sodium-EDTA. The samples were subjected to isopycnic density-gradient centrifugation in CsCl/4 M guanidinium chloride (40 000 rpm; 15 °C; 65 h; Beckman 70.1 rotor; Beckman Optima L70 ultracentrifuge; initial density 1.37 mg/ml). Fractions were collected from the bottom of the tubes and analysed for density, absorbance at 280 nm, carbohydrate, reactivity with the LUM2-3 antibody and radioactivity. 2.6. Degradative methods Samples were dialysed against reduction buffer and subjected to reduction with 10 mM DTT for 5 h at 37 °C, followed by alkylation with 25 mM iodoacetamide overnight at room temperature or dialysed against 0.1 M ammonium hydrogen carbonate, pH 8.0 and treated with trypsin (50 µg) 37 °C for 5 h. Chondroitin ABC lyase digestion was performed with 20 mU of enzyme overnight at 37 °C after dialysis against 0.1 M Tris-acetate buffer, pH 7.3 containing 10 mM sodium-EDTA. Digestion with heparan sulphate III lyase was performed with 28 mU of enzyme overnight at 37 °C after dialysis against 50 mM Tris-acetate buffer, pH 7.0 containing 5 mM calcium acetate. Alkaline borohydride treatment was performed for 16 h at 45 °C on samples dialysed into water by the addition of an equal volume of

383

2 M NaBH4 in 0.1 M NaOH. After this treatment, samples were neutralised by the addition of acetic acid [20]. 2.7. Gel chromatography Samples were dialysed into 4 M guanidinium chloride/10 mM sodium phosphate buffer, pH 7.4 and subjected to gel chromatography on a CL-2B column (1.6 × 50 cm) eluted with the same solvent at a flow rate of 0.1 ml/min or a Superose 6 HR 10/30 column eluted with the same buffer, at a flow rate of 0.25 ml/min. Fractions (1 and 0.25 ml, respectively) were analysed for reactivity with the LUM2-3 antiserum, carbohydrate and radioactivity. 2.8. Rate-zonal centrifugation Rate-zonal centrifugation in guanidinium chloride (6-8 M) was carried out as described by Wickström et al. [19] at 40 000 rpm for 2 h 45 min. The tubes were emptied from the top and fractions (300 µl) were analysed for radioactivity and reactivity with the LUM2-3 antiserum. 2.9. Agarose gel electrophoresis Agarose gel electrophoresis was performed and samples were blotted onto PVD membranes as described by Herrmann et al. [10]. After blotting, half the membrane was dried and subjected to autoradiography for 3 weeks in the freezer. The other half was subjected to western blotting using the LUM2-3 antiserum as described [10].

3. Results After metabolic labelling in organ culture with [3H]-Dglucosamine and [35S]sulphate labelled precursors for 6, 12 and 24 h, the colonic tissue was subjected to extraction in guanidinium chloride providing material that was ‘soluble’ and ‘insoluble’ in this solvent. The fraction that was ‘insoluble’ in guanidinium chloride was brought into solution by reduction and studied further as ‘subunits’. Also the culture medium was separated into fractions that were soluble and ‘insoluble’ in guanidinium chloride. Density-gradient centrifugation showed that the major part of the incorporated [35S]sulphate label occurred in a high-density component of the guanidinium chloride soluble material obtained from the tissue. This material was noticeable at 12 h of labelling and increased at 24 h (Fig. 1d-f, lower panels). A UV absorbing peak at 1.47 g/ml, remained constant over time and coincided with the [35S]sulphate label, this most likely representing DNA (Fig. 1d-f, upper panels). The [35S]sulphate labelled component was well separated from a tiny amount of MUC2 detected at a densityof approximately 1.38 mg/ml (Fig. 1f, lower panel). Almost all MUC2 in the tissue was obtained as a glycoprotein complex ‘insoluble’ in guanidinium chloride that after reduction appeared as a distinct population with a density of

384

A. Herrmann et al. / Biochimie 85 (2003) 381–390

Fig. 1. Isopycnic density-gradient centrifugation in 4 M guanidinium chloride/CsCl of molecules secreted into the culture medium (upper panel) and material soluble (middle panel) and ‘insoluble’ (lower panel) in guanidinium chloride from colonic tissue after pulse labelling for 6 h (a, d, g), 12 h (b, e, h) and 24 h (c, f, i) in organ culture. Colonic tissue was extracted with 6 M guanidinium chloride and the ‘insoluble’ fraction was brought into solution by reduction. The soluble and ‘insoluble’ fractions, as well as the culture medium, were subjected to density-gradient centrifugation in 4 M guanidinium chloride/CsCl (65 h; 15 °C; 70.1 Ti rotor; Beckman L-70 ultracentrifuge). Fractions were analysed for density ("), A280 (---), carbohydrate (m), reactivity with the LUM2-3 antiserum (n), [35S]sulphate ( ) and [3H]-D-glucosamine ( ). In the analysis for carbohydrate, samples in (a-c) were diluted 4-fold compared to others.

approximately 1.4 g/ml, as expected for colonic MUC2 [10]. As detected by an anti-MUC2 antiserum, as well as carbohydrate analysis, MUC2 was present already after 6 h of culture and no major changes in the amount of mucin were noted with time (Fig. 1g-i, lower panels) in two out of three cultures. In the third one, the amount of MUC2 decreased somewhat with time (data not shown). Little [35S]sulphate

and almost no [3H]-D-glucosamine was incorporated into MUC2 as compared to the ‘high-density’ and ‘low-density’ material in the soluble fractions (Fig. 1g-i, lower panels). The radiolabel was noticeable first after 24 h of culture. Already after 6 h of incubation, [3H]-D-glucosamine labelled material appeared at the top of the gradient of the guanidinium chloride soluble fraction both from the tissue

A. Herrmann et al. / Biochimie 85 (2003) 381–390

and the medium. The material increased following 12 and 24 h of incubation and more labelled material was obtained in the medium than from the tissue (Fig. 1a-f, lower panels). No material ‘insoluble’ in guanidinium chloride could be recovered from the medium (data not shown). In order to study the major radiolabelled components further, additional organ cultures were subjected to metabolic labelling for 23 h. Very similar results were obtained as compared to the time-course study (Fig. 2). However, the amount of [35S]sulphate label incorporated into the MUC2 ‘insoluble’ complex was higher (Fig. 2d, lower panel) as well as the amount of [35S]sulphate label relative to [3H]-Dglucosamine label of the high-density component (Fig. 2c, lower panel, pooling bar I). These differences probably reflect individual patient variation. Some ‘insoluble‘ MUC2 was recovered from the ‘support’ most likely representing secreted material (data not shown). 3.1. Gel chromatography Gel chromatography on Sepharose CL-2B of the reduced ‘insoluble’ MUC2 complex (Fig. 2d, lower panel) showed an elution profile with a number of partially separated populations close to the void volume as expected (Fig. 3a). Reactivity with the LUM2-3 antiserum, radioactivity and carbohydrate appeared within the same fractions. Most of the [35S]sulphate containing, high-density component from the tissue (Fig. 2c, lower panel, pooling bar I) eluted in the void volume on a Sepharose CL-2B column, suggesting a molecular size identical to, or larger than, the largest MUC2 species (Fig. 3b). However, small amounts of radiolabelled material also appeared as a broad distribution close to the total volume. Neither reduction nor digestion with trypsin, heparan sulphate lyase or chondroitin ABC lyase changed the elution profile of the material eluting in the void volume and no increase in the amount of radiolabelled material eluting near the total volume was detected indicating that no radiolabelled fragments were cleaved off during the digestions (results not shown). However, after treatment with alkaline borohydride, all radiolabelled material eluted close to the total volume (Fig. 3c). Also on a Superose 6 column, this material eluted close to the total volume (Fig. 3c, insert), indicating that the oligosaccharides released are of very low molecular size and no separation between [3H]-D-glucosamine and [35S]sulphate label was observed. The low-density, [3H]-Dglucosamine labelled material from the guanidinium chloride soluble fraction of both the tissue (Fig. 2c, lower panel, pooling bar II) and the medium (Fig. 2a, lower panel) eluted close to the total volume on the Sepharose CL-2B column showing that this material is of considerably smaller size than the high-density, [35S]sulphate labelled component from the tissue (Fig. 3d,e). When subjected to chromatography on a Superose 6 column, this material from the medium appeared as a relatively homogenous peak that was well included on the column (Fig. 3e, insert). The material from the tissue showed a considerably broader distribution, showing some radiolabelled material in the void volume, suggesting the

385

presence of more than one [3H]-D-glucosamine labelled component (Fig. 3d, insert). 3.2. Rate-zonal centrifugation When subjected to rate-zonal centrifugation, the reduced MUC2 complex (Fig. 2d, lower panel) showed a broad distribution that suggested the presence of a number of differently sized populations (Fig. 4a) in agreement with the results obtained by using gel chromatography. The major part of the [35S]sulphate labelled material from the tissue (Fig. 2c, lower panel, pooling bar I) appeared as a broad peak that migrated further into the gradient than the largest MUC2 species (Fig. 4b). The broad distribution reflects a polydispersity in size, which was not seen on gel chromatography since the material eluted in the void volume. The population of smaller molecules appearing in the first fractions of the gradient most likely corresponds to the broad peak of material eluting close to the total volume of the Sepharose CL-2B column (Fig. 3b). Reduction did not affect sedimentation and thus the apparent size of the molecules (Fig. 4c). The lowdensity, [3H]-D-glucosamine labelled material of the tissue (Fig. 2c, lower panel, pooling bar II) hardly entered the gradient under the conditions used (Fig. 4d) confirming the low molecular size observed by using gel chromatography (Fig. 3d). 3.3. Agarose gel electrophoresis Electrophoresis of the reduced MUC2 complex (Fig. 2d, lower panel) followed by western blotting with the LUM2-3 antibody revealed a ‘ladder’ of bands (Fig. 5a) as expected for colonic MUC2. Almost the same pattern was obtained when the blot was subjected to autoradiography (Fig. 5b) showing that all the MUC2 populations obtained after reduction of the complex were radiolabelled and that no major radiolabelled non-MUC2 components were present in the complex. However, part of the fastest moving band (Fig. 5b) does not over-lap completely with the MUC2 monomer suggesting the presence of a small amount of a component that is not MUC2. Autoradiography of the high-density, [35S]sulphate labelled material from the tissue (Fig. 2c, lower panel, pooling bar I) revealed a broad band with about the same mobility as the slowest moving MUC2 species (Fig. 5c). The broad appearance of the band may correspond to the polydispersity in size revealed by the rate-zonal gradient, or a combination of size and charge polydispersity of the molecule. 4. Discussion Pulse labelling of human colon in organ culture for increasing time periods showed that the amount of label incorporated into non-dialyzable material increased during the entire experimental period. After 24 h of culture, the tissue showed a normal appearance, and earlier studies have shown that colonic tissue is viable in organ culture for at least this

386

A. Herrmann et al. / Biochimie 85 (2003) 381–390

Fig. 2. Isopycnic density-gradient centrifugation in 4 M guanidinium chloride/CsCl of material isolated from the culture medium and colonic tissue after 23 h in organ culture. Culture medium (a-b), and the tissue (c-d) were separated into ‘soluble’ (upper panel) and ‘insoluble’ material (lower panel) by extraction with 6 M guanidinium chloride and samples were subjected to density-gradient centrifugation in 4 M guanidinium chloride/CsCl as in Fig. 1. Fractions were analysed for density ("), A280 (---), carbohydrate (m), reactivity with the LUM2-3 antiserum (n), [35S]sulphate ( ) and [3H]-D-glucosamine ( ). In the analysis for carbohydrate, the sample in (a) was diluted 4-fold compared to others and the sample in (d) was diluted 5-fold in the analysis for MUC2 as compared to the other samples.

time period under the conditions used [12]. Isopycnic density-gradient centrifugation revealed that most of the [3H]-D-glucosamine was associated with low-density molecules at the top of the gradients in the fractions soluble in guanidinium chloride. Gel chromatography and rate-zonal

centrifugation showed that the major part of this material comprises relatively small molecules. The low-Mr, [3H]-Dglucosamine labelled, non-mucin glycoconjugates seen in the medium already after 6 h have been shown to occur also in conditions where mucin synthesis is inhibited and con-

A. Herrmann et al. / Biochimie 85 (2003) 381–390

Fig. 3. Gel chromatography of the reduced MUC2 ‘insoluble’ complex (a), the major [35S]sulphate labelled component from the tissue before (b) and after (c) alkaline elimination and the low-density [3H]-D-glucosamine labelled material from the tissue (d) and the medium (e). Reduced MUC2 ‘insoluble’ complex from the tissue (Fig. 2d), high-density [35S]sulphate labelled material from the tissue (Fig. 2c, pooling bar I) before and after alkaline elimination and the major [3H]-D-glucosamine labelled material from the tissue (Fig. 2c, pooling bar II) and the medium (Fig. 2a) were subjected to gel chromatography on a Sepharose CL-2B column eluted with 4 M guanidinium chloride, pH 7 at a flow rate of 0.1 ml/min. Inserts show the same material subjected to chromatography on Superose 6 eluted with the same solvent at a flow rate of 0.25 ml/min. Fractions were analysed for carbohydrate (m), reactivity with the LUM2-3 antiserum (n), [35S]sulphate ( ) and [3H]-D-glucosamine ( ).

firms tissue viability when no mucins are produced (Corfield, Longman and Myerscough, unpublished observations). Almost all MUC2 was found associated with the tissue and occurred as an ‘insoluble’ glycoprotein complex as ex-

387

Fig. 4. Rate-zonal centrifugation of the reduced MUC2 ‘insoluble’ complex (a), the major [35S]sulphate labelled component from the tissue before (b) and after (c) reduction, and the major [3H]-D-glucosamine labelled material from the tissue (d). Reduced MUC2 ‘insoluble’ complex (Fig. 2d), highdensity [35S]sulphate labelled material from the tissue (Fig. 2c, pooling bar I), before and after reduction/alkylation and low-density, [3H]-Dglucosamine labelled material (Fig. 2c, polling bar II) were subjected to rate-zonal centrifugation in a 6-8 M guanidinium chloride gradient in a Beckman SW 41 Ti rotor at 40 000 rpm at 20 °C for 2 h 45 min. The tubes were unloaded from the top and fractions were analysed for reactivity with the LUM2-3 antiserum (n), [35S]sulphate ( ) and [3H]-D-glucosamine ( ).

pected from previous investigations [10,21]. After reduction, the MUC2 subunits appeared as a distinct peak with a density of approximately 1.38 g/ml and some [35S]sulphate label was incorporated into this component. Gel chromatography of the reduced complex suggested the presence of a number of populations, in agreement with previous observations [10]. Agarose gel electrophoresis/western blotting revealed a ‘typical’ MUC2 ladder pattern corresponding to MUC2 monomers and oligomers linked by so far unidentified reduction-insensitive bond(s) [10]. Autoradiography showed that the radioactivity was indeed incorporated into MUC2 and that no major non-MUC2 radiolabelled population was present. Very little MUC2 was present in the medium, how-

388

A. Herrmann et al. / Biochimie 85 (2003) 381–390

Fig. 5. Agarose gel electrophoresis of the reduced MUC2 ‘insoluble’ complex from the tissue (a and b) and the major [35S]sulphate labelled highdensity material from the tissue (c). Radiolabelled high-density material (Fig. 2c, pooling bar I) and reduced MUC2 complex (Fig. 2d) were subjected to agarose gel electrophoresis (1% agarose; 18 h; 30 V) and blotted onto a PVDF membrane. Membranes were incubated with the LUM2-3 antiserum (a) or subjected to autoradiography (b and c).

ever, some was found associated with the lens tissue used to support the cultures. The latter material most likely represents secreted ‘insoluble’ MUC2. The results indicate that the ‘insoluble’ mucus gel present on the epithelial surface of the biopsies is metabolically stable during incubation periods of up to 24 h. The ‘insoluble’ MUC2 complex turns over slowly under these culture conditions and virtually all newly synthesised MUC2 appears to be incorporated immediately into the ‘insoluble’ complex. Most of the [35S]sulphate was incorporated into molecules with a considerably higher density than so far reported for secreted, oligomeric mucins. This high-density, non-MUC2 population chromatographed with the void volume on Sepharose CL-2B, was unaffected by proteoglycandegrading enzymes and is, therefore, unlikely to be a proteoglycan containing chondroitin, dermatan and/or heparan sulphate side chains. No decrease in molecular size was detected following reduction of disulphide bonds, suggesting that the molecule is ‘monomeric’ and has a molecular architecture different from that of the monomers (‘subunits’) forming the oligomeric mucins MUC2, MUC5AC, MUC5B and MUC6 [10,22–24]. Furthermore, the apparent lack of sensitivity to trypsin digestion suggests that the hydrody-

namic behaviour of the molecule is dominated by a single large protease resistant glycosylated domain. Most likely, it is not composed of two or more such domains flanked by proteinase-sensitive stretches of the protein as is the case for e.g. MUC2 and MUC5B [10,25,26]. After treatment with alkaline borohydride, the molecule was degraded into fragments eluting in the Vt range of Sepharose CL 2B and Superose 6 columns (Fig. 3c). This indicates that the carbohydrate components are O-linked oligosaccharides, in keeping with a mucin-like nature of the glycoconjugate. In contrast to MUC2, the high-density mucin-like component could not be detected by using chemical analysis, suggesting that it is present in very low concentrations which, together with the fact that it becomes heavily radiolabelled, is consistent with a high turnover rate. A high-density and apparently monomeric mucin-like molecule has previously been identified in cat airway secretions [15] and a monomeric radiolabelled mucin-like molecule with a density similar to the oligomeric mucins has been isolated from bovine trachea in organ culture [16] and from bovine tracheal epithelial cells in primary culture [17]. The fact that these molecules, including those identified here from human colon, are difficult to separate from the oligomeric mucus-forming mucins may complicate the interpretation of investigations aimed at investigating the biosynthesis, secretion and turnover of the latter. Here, less than 10% of the [35S]sulphate incorporated by the tissue was incorporated in MUC2 whereas more than 65% was present in the mucin-like molecule from the tissue. In addition to MUC2, the MUC1, MUC3A, MUC3B, MUC4, MUC5B, MUC6, MUC11, MUC12, MUC13 and MUC17 mucin genes are expressed in human colon. The genetic identity of the high-density mucin-like component identified here is not known, but the apparent size of the molecule together with the lack of sensitivity to trypsin digestion suggest the presence of a glycosylated domain larger than that of the MUC2 VNTR region. From gel electrophoresis/western blotting it is concluded that the individuals in this study express MUC2 with ‘long’ alleles [10]. These alleles contain a VNTR region coding for approximately 2300 amino acids and the high-density component is thus not likely to be MUC1, since the VNTR region of the latter mucin contains 25-125 repeats of 20 amino acids, i.e. the longest possible glycosylated domain is about 2500 amino acids [4]. In contrast, MUC4 contains a VNTR region that varies in length from 2334 to 6334 amino acids [27] and the ‘longer’ forms of this mucin comprise a glycosylated domain considerably larger than that of MUC2. The properties of the high-density mucin-like molecule identified here would, therefore, be consistent with those expected for MUC4. The high density of this molecule is most likely explained by a high degree of sulphate substitution in a highly glycosylated structure. The mucin genes MUC3A & B, MUC11, MUC12, and MUC17 have not been sequenced to the extent that the size of their glycosylated domains is known. However, northern blot analysis shows that the transcripts are in excess of 12 kb [6,9,28] suggesting that the

A. Herrmann et al. / Biochimie 85 (2003) 381–390

corresponding gene products could in fact encompass a glycosylated domain of the same size as that postulated to be present in the high-density component identified here. In the use of colonic tissue organ-culture with radiolabelled precursors in order to study biosynthesis and secretion of mucins many investigators have failed to recognise the major part of MUC2 due to the insoluble complex, which is lost during sample preparation. Furthermore, if the remaining, small amount of ‘soluble’ MUC2 is not carefully separated from the ‘novel’ mucin identified here, the two molecules will be studied as a mixture and the turnover rate of MUC2 will be highly over-estimated. Gel chromatography, even after reduction, will not separate the two mucins. In conclusion, by using radiolabelling of human colon in organ culture, we have identified a putative monomeric highdensity mucin-like molecule that appears to become much more avidly radiolabelled than MUC2. The molecule is dominated by a single glycosylated domain and has a size consistent with that of MUC4. MUC2 was poorly radiolabelled as compared to the ‘novel’, high-density component although some [35S]sulphate was incorporated. Metabolic labelling studies of mucin biosynthesis must consider the presence of the putative monomeric mucin identified here.

[5]

[6]

[7]

[8]

[9]

[10]

[11]

[12]

Acknowledgements We thank Annika Böök, Jonas Carlstedt, Anna Tell and others for technical assistance. The work was supported by grants from the Swedish Medical Research Council (grants no. 7902), the Medical Faculty of Lund, Crafoordska Stiftelsen, the Swedish Foundation for Strategic Research (Inflammation program), Alfred Österlunds Stiftelse, the Welcome Trust (Grant 051586/Z/97), Smith and Nephew, UK, the European Union (contract number BMH4-CT98-3222) and Swedish Match.

[13]

[14]

[15]

[16]

References [1]

[2]

[3]

[4]

J.R. Gum, J.C. Byrd, J.W. Hicks, N.W. Toribara, D.T. Lamport, Y.S. Kim, Molecular cloning of human intestinal mucin cDNAs. Sequence analysis and evidence for genetic polymorphism, J. Biol. Chem. 264 (1989) 6480–6487. J.R. Gum, J.W. Hicks, D.M. Swallow, R.L. Lagace, J.C. Byrd, D.T. Lamport, B. Siddiki, Y.S. Kim, Molecular cloning of cDNAs derived from a novel human intestinal mucin gene, Biochem. Biophys. Res. Commun. 171 (1990) 407–415. N. Porchet, V.C. Nguyen, J. Dufosse, J.P. Audie, V. Guyonnet-Duperat, M.S. Gross, C. Denis, P. Degand, A. Bernheim, J.P. Aubert, Molecular cloning and chromosomal localization of a novel human tracheo-bronchial mucin cDNA containing tandemly repeated sequences of 48 base pairs, Biochem. Biophys. Res. Commun. 175 (1991) 414–422. S.J. Gendler, A.P. Spicer, Epithelial mucin genes, Annu. Rev. Physiol. 57 (1995) 607–634.

[17]

[18]

[19]

[20]

[21]

[22]

389

B.J. van Klinken, J. Dekker, S.A. van Gool, J. van Marle, H.A. Buller, A.W. Einerhand, MUC5B is the prominent mucin in human gallbladder and is also expressed in a subset of colonic goblet cells, Am. J. Physiol. 274 (1998) G871–G878. S.J. Williams, M.A. McGuckin, D.C. Gotley, H.J. Eyre, G.R. Sutherland, T.M. Antalis, Two novel mucin genes down-regulated in colorectal cancer identified by differential display, Cancer Res. 59 (1999) 4083–4089. W.S. Pratt, S. Crawley, J. Hicks, J. Ho, M. Nash, Y.S. Kim, J.R. Gum, D.M. Swallow, Multiple transcripts of MUC3: evidence for two genes, MUC3A and MUC3B, Biochem. Biophys. Res. Commun. 275 (2000) 916–923. S.J. Williams, D.H. Wreschner, M. Tran, H.J. Eyre, G.R. Sutherland, M.A. McGuckin, Muc13, a novel human cell surface mucin expressed by epithelial and hemopoietic cells, J. Biol. Chem. 276 (2001) 18327–18336. J.R. Gum Jr, S.C. Crawley, J.W. Hicks, D.E. Szymkowski, Y.S. Kim, MUC17, a novel membrane-tethered mucin, Biochem. Biophys. Res. Commun. 291 (2002) 466–475. A. Herrmann, J.R. Davies, G. Lindell, S. Mårtensson, N.H. Packer, D.M. Swallow, I. Carlstedt, Studies on the “insoluble” glycoprotein complex from human colon. Identification of reduction-insensitive MUC2 oligomers and C-terminal cleavage, J. Biol. Chem. 274 (1999) 15828–15836. M.A. Axelsson, N. Asker, G.C. Hansson, O-glycosylated MUC2 monomer and dimer from LS 174T cells are water-soluble, whereas larger MUC2 species formed early during biosynthesis are insoluble and contain non-reducible intermolecular bonds, J. Biol. Chem. 273 (1998) 18864–18870. A.P. Corfield, N. Myerscough, N. Bradfield, C.D.A. Corfield, M. Gough, J.R. Clamp, P. Durdey, B.F. Warren, D.C. Bartolo, K.R. King, J.M. Williams, Colonic mucins in ulcerative colitis: evidence for loss of sulfation, Glycoconjugate J. 13 (1996) 809–822. B.J. van Klinken, J.W. van der Wal, A.W. Einerhand, H.A. Buller, J. Dekker, Sulphation and secretion of the predominant secretory human colonic mucin MUC2 in ulcerative colitis, Gut 44 (1999) 387–393. J.K. Sheehan, D.J. Thornton, M. Howard, I. Carlstedt, A.P. Corfield, C. Paraskeva, Biosynthesis of the MUC2 mucin: evidence for a slow assembly of fully glycosylated units, Biochem. J. 315 (1996) 1055–1060. J.R. Davies, J.T. Gallagher, P.S. Richardson, J.K. Sheehan, I. Carlstedt, Mucins in cat airway secretions, Biochem. J. 275 (1991) 663–669. N. Svitacheva, H.W. Hovenberg, J.R. Davies, Biosynthesis of mucins in bovine trachea: identification of the major radiolabelled species, Biochem. J. 333 (1998) 449–456. N. Svitacheva, J.R. Davies, Mucin biosynthesis and secretion in tracheal epithelial cells in primary culture, Biochem. J. 353 (2001) 23–32. P.L. Devine, A sensitive microplate assay for glycoproteins that utilizes an immunological digoxigenin-based detection system, Biotechniques 12 (1992) 160–162. C. Wickström, I. Carlstedt, N-terminal cleavage of the salivary MUC5B mucin. Analogy with the von Willebrand propolypeptide? J. Biol. Chem. 276 (2001) 47116–47121. D.M. Carlsson, Structures and immunochemical properties of oligosaccharides isolated from pig submaxillary mucins, J. Biol. Chem. 243 (1968) 616–626. I. Carlstedt, A. Herrmann, H. Hovenberg, G. Lindell, H. Nordman, C. Wickström, J.R. Davies, ‘Soluble’ and ‘insoluble’ mucinsidentification of distinct populations, Biochem. Soc. Trans. 23 (1995) 845–851. H.W. Hovenberg, J.R. Davies, A. Herrmann, C.J. Linden, I. Carlstedt, MUC5AC, but not MUC2, is a prominent mucin in respiratory secretions, Glycoconjugate J. 13 (1996) 839–847.

390

A. Herrmann et al. / Biochimie 85 (2003) 381–390

[23] C. Wickström, J.R. Davies, G.V. Eriksen, E.C. Veerman, I. Carlstedt, MUC5B is a major gel-forming, oligomeric mucin from human salivary gland, respiratory tract and endocervix: identification of glycoforms and C-terminal cleavage, Biochem. J. 334 (1998) 685–693. [24] H. Nordman, J.R. Davies, G. Lindell, C. de Bolós, F. Real, I. Carlstedt, Gastric MUC5AC and MUC6 are large oligomeric mucins that differ in size, glycosylation and tissue distribution, Biochem. J. 364 (2002) 191–200. [25] I. Carlstedt, H. Lindgren, J.K. Sheehan, The macromolecular structure of human cervical-mucus glycoproteins. Studies on fragments obtained after reduction of disulphide bridges and after subsequent trypsin digestion, Biochem. J. 213 (1983) 427–435.

[26] J.L. Desseyn, V. Guyonnet-Duperat, N. Porchet, J.P. Aubert, A. Laine, Human mucin gene MUC5B, the 10.7-kb large central exon encodes various alternate subdomains resulting in a super-repeat. Structural evidence for a 11p15.5 gene family, J. Biol. Chem. 272 (1997) 3168–3178. [27] N. Moniaux, S. Nollet, N. Porchet, P. Degand, A. Laine, J.P. Aubert, Complete sequence of the human mucin MUC4: a putative cell membrane-associated mucin, Biochem. J. 338 (1999) 325–333. [28] V. Debailleul, A. Laine, G. Huet, P. Mathon, M.C. d’Hooghe, J.P. Aubert, N. Porchet, Human mucin genes MUC2, MUC3, MUC4, MUC5AC, MUC5B, and MUC6 express stable and extremely large mRNAs and exhibit a variable length polymorphism. An improved method to analyze large mRNAs, J. Biol. Chem. 273 (1998) 881–890.