The contribution of tandem repeat number to the O-glycosylation of mucins

The serine- and threonine-rich tandem repeat (TR) units that make up the characteristic feature of mucin glycoproteins are often polymorphic with substantial genetic variation in TR number. The precise effect of TR number on O-glycosylation is not fully understood, although the TR number of several mucins may be associated with apparent susceptibility to certain human diseases. To evaluate the contribution of TR number to O-glycosylation, we generated a series of chimeric mucins carrying increasing numbers of TR units from the MUC5B mucin in the context of an epitope-tagged MUC1 mucin backbone. These mucins were expressed in Caco2 colon carcinoma cell clones and purified by immunoprecipitation. O-Glycosylation was investigated by western blotting with antibodies to known carbohydrate structures and by fast atom bombardment-mass spectrometry. Additional carbohydrate epitopes were detected with antibodies on chimeric mucins with a higher TR number in comparison to those with fewer TRs. Using mass spectrometry, higher-molecular-weight glycans were detected more frequently on the mucins with extended TRs compared to those with fewer TRs. However no novel carbohydrate structures were seen, suggesting that TR number does not affect the specificity of O-glycosylation.     

In vivo glycosylation of mucin tandem repeats.

The biochemical and biophysical properties of mucins are largely determined by extensive O-glycosylation of serine- and threonine-rich tandem repeat (TR) domains. In a number of human diseases aberrant O-glycosylation is associated with variations in the properties of the cell surface-associated and secreted mucins. To evaluate in vivo the O-glycosylation of mucin TR domains, we generated recombinant chimeric mucins with TR sequences from MUC2, MUC4, MUC5AC, or MUC5B, which were substituted for the native TRs of epitope-tagged MUC1 protein (MUC1F). These hybrid mucins were extensively O-glycosylated and showed the expected association with the cell surface and release into culture media. The presence of different TR domains within the chimeric mucins appears to have limited influence on their posttranslational processing. Alterations in glycosylation were detailed by fast atom bombardment mass spectrometry and reactivity with antibodies against particular blood-group and tumor-associated carbohydrate antigens. Future applications of these chimeras will include investigations of mucin posttranslational modification in the context of disease.     

CFTR expression does not influence glycosylation of an epitope-tagged MUC1 mucin in colon carcinoma cell lines

The cause of the mucus clearance problems associated with cystic fibrosis remains poorly understood though it has been suggested that mucin hypersecretion, dehydration of mucins, and biochemical abnormalities in the glycosylation of mucins may be responsible. Since the biochemical and biophysical properties of a mucin are dependent on O-glycosylation, our aim was to evaluate the O-glycosylation of a single mucin gene product in matched pairs of cells that differed with respect to CFTR expression. An epitope-tagged MUC1 mucin cDNA (MUC1F) was used to detect variation in mucin glycosylation in stably transfected colon carcinoma cell lines HT29 and Caco2. The glycosylation of MUC1F mucin was evaluated in matched pairs of Caco2 cell lines that either express wild-type CFTR or have spontaneously lost CFTR expression. The general glycosylation pattern of MUC1F was evaluated by determining its reactivity with a series of monoclonal antibodies against known blood group and tumor-associated carbohydrate antigens. Metabolic labeling experiments were used to estimate the gross levels of glycosylation and sulfation of MUC1F mucin in these matched pairs of cell lines. Expression of CFTR in this experimental system did not affect the gross levels of glycosylation or sulfation of the MUC1F mucin nor the types of carbohydrates structures attached to the MUC1F protein.     

Human MUC4 mucin cDNA and its variants in pancreatic carcinoma

The human MUC4 gene is not expressed in normal pancreas; however, its dysregulation results in high levels of expression in pancreatic tumors. To investigate the tumor-associated expression, MUC4 cDNA was cloned from a human pancreatic tumor cell line cDNA expression library using a polyclonal antibody raised against human deglycosylated mucin and RT-PCR. Pancreatic MUC4 cDNA shows differences in 12 amino acid residues in the non-tandem repeat coding region with no structural rearrangement as compared with tracheal MUC4. The full-length MUC4 cDNA includes a leader sequence, a serine and threonine rich non-tandem repeat region, a central large tandem repeat domain containing 48 bp repetitive units, regions rich in potential N-glycosylation sites, two cysteine-rich domains, EGF-like domains, and a transmembrane domain. We also report the presence of a new EGF-like domain in MUC4 cDNA, located in the cysteine-rich region upstream from the first EGF-like domain. Four distinct splice events were identified in the region downstream of the central tandem repeat domain that generate three new MUC4 cDNA sequences (sv4, sv9, and sv10). The deduced amino acid sequences of two of these variants lack the transmembrane domain. Furthermore, two unique forms of MUC4 (MUC4/Y and MUC4/X) generated as a result of alternative splicing lack the salient feature of mucins, the tandem repeat domain. A high degree of polymorphism in the central tandem repeat region of MUC4 was observed in various pancreatic adenocarcinoma cell lines, with allele sizes ranging from 23.5 to 10.0 kb. MUC4 mRNA expression was higher in differentiated cell lines, with no detectable expression in poorly differentiated pancreatic tumor cell lines.     

In vivo glycosylation of MUC1 in airway epithelial cells

The O-glycans that decorate mucin glycoproteins contribute to the biophysical and biochemical properties of these molecules and hence their function as a barrier and lubricant on epithelial surfaces. Alterations in mucin O-glycosylation in certain diseases may contribute to pathology. It is known that both the host cell type and the amino acid sequence of the mucin tandem repeat contribute to the O-glycosylation of a mucin molecule. We expressed an epitope-tagged MUC1 mucin cDNA construct in the airway cell line 16HBE14o- and the colon carcinoma cell line Caco2 and used Fast Atom Bombardment Mass Spectrometry to evaluate the contribution of the host cell to differences in O-glycosylation of a single mucin. Many of the glycans detected on the MUC1 mucin were common to both cell types, as would be predicted from biosynthetic constraints. However, MUC1 synthesized in the airway cell line showed comparatively low levels of sialylation but carried a range of oligo-N-acetyllactosamine structures that were not seen in the colon carcinoma cell line.     

Molecular cloning and chromosomal localization of a novel human tracheo-bronchial mucin cDNA containing tandemly repeated sequences of 48 base pairs.

A lambda gt11 cDNA library constructed from human tracheo-bronchial mucosa was screened with a polyclonal antiserum raised to chemically deglycosylated pronase glycopeptides from human bronchial mucins. Out of 20 positives clones, one partial cDNA clone was isolated and allowed to map a novel human tracheo-bronchial mucin gene. It contains 48 nucleotide tandem repeats quite perfectly identical which encodes a protein containing about 50% of hydroxy amino-acids. This clone hybridized to polydisperse messages produced by human tracheo-bronchial and human colonic mucosae. The gene (proposed name MUC 4) from which cDNA is derived maps to chromosome 3.     

Cysteine-rich regions of pig gastric mucin contain von willebrand factor and cystine knot domains at the carboxyl terminal(1).

In order to sequence the cysteine-rich regions of pig gastric mucin (PGM), we used our previously identified pig gastric mucin clone PGM-2A to screen a pig stomach cDNA library and perform rapid amplification of cDNA ends to obtain two cysteine-rich clones, PGM-2X and PGM-Z13. PGM-2X has 1071 base pairs (bp) encoding 357 amino acids containing five serine-threonine-rich 16 amino acid tandem repeats, downstream from a cysteine-rich region similar to human and mouse MUC5AC. PGM-Z13 encodes the complete 3'-terminus of PGM and is composed of 3336 bp with a 2964 bp open reading frame encoding 988 amino acids with four serine-threonine-rich tandem repeats upstream from a cysteine-rich region similar to the carboxyl terminal regions of human and rat MUC5AC and human MUC5B. This region is homologous to von Willebrand factor C and D domains involved in acid induced polymerization, and to the carboxyl terminal cystine-knot domain of various mucins, TGF-beta, vWF and norrin, which is involved in dimerization. These newly sequenced cysteine-rich regions of pig gastric mucin may be critical for its gelation and for its observed increased viscosity induced by low pH.     

Heterogeneity in the protein cores of mucins isolated from human middle ear effusions: evidence for expression of different mucin gene products

High molecular weight mucins were isolated and purified from human middle ear effusions of children with Otitis Media with Effusion (OME) classified into three groups, (1) thick and (2) thin from anatomically normal children and (3) effusions from cleft palate patients. Amino acid analyses of the purified mucins from the three pools were similar but not identical with characteristic contents of serine threonine and proline (32%, 28%, and 38% for pools (1) (2) and (3) respectively). Proteinase resistant glycopeptide fragments corresponding to the tandem repeat domains of cloned mucin genes showed marked differences both between the three mucin pools and with the composition of the tandem repeat sequences of the cloned mucin genes expressed in the airways. Studies on the antigenic identity of middle ear mucins found an epitope likely to be present on MUC5AC, but only accounting for a maximum of 15% by weight and no reactivity was found with antibodies to MUC2 or MUC1. A polyclonal antibody raised to thick effusion mucins reacted strongly with human salivary mucin suggesting the presence of MUC5B epitopes. These studies suggest that more than one mucin gene product is secreted by the human middle ear mucosa and that there may be further mucin genes expressed by the middle ear that have yet to be cloned.     

MUC genes: mucin or not mucin? That is the question

Mucins are macromolecules lying the cells in contact with external environment and protect the epithelium against constant attacks such as digestive fluids, microorganisms, pollutants, and toxins. Mucins are the main components of mucus and are synthesized and secreted by specialized cells of the epithelium (goblet cells, cells of mucous glands) or non mucin-secreting cells. Human mucin genes show common features: large size of their mRNAs, large nucleotide tandem repeat domains, complex expression both at tissular and cellular level. Since 1987, 21 MUC symbols have been used to designate genes encoding O-glycoproteins containing tandem repeat domains rich in serine, threonine and proline. Some of these genes encode true mucins while others encode non mucin adhesion O-glycoproteins. In this paper, we propose a classification based on sequence similarities and expression areas. Two main families can be distinguished: secreted mucins or gel-forming mucins (MUC2, MUC5AC, MUC5B, MUC6), and membrane-bound mucins (MUC1, MUC3, MUC4, MUC12, MUC17). Muc-deficient mice will provide important models in the study of functional relationships between these two mucin families.     

Exploring the glycosylation of mucins by use of O-glycodomain reporters recombinantly expressed in glycoengineered HEK293 cells

Mucins and glycoproteins with mucin-like regions contain densely O-glycosylated domains often found in tandem repeat (TR) sequences. These O-glycodomains have traditionally been difficult to characterize because of their resistance to proteolytic digestion, and knowledge of the precise positions of O-glycans is particularly limited for these regions. Here, we took advantage of a recently developed glycoengineered cell-based platform for the display and production of mucin TR reporters with custom-designed O-glycosylation to characterize O-glycodomains derived from mucins and mucin-like glycoproteins. We combined intact mass and bottom–up site-specific analysis for mapping O-glycosites in the mucins, MUC2, MUC20, MUC21, protein P-selectin-glycoprotein ligand 1, and proteoglycan syndecan-3. We found that all the potential Ser/Thr positions in these O-glycodomains were O-glycosylated when expressed in human embryonic kidney 293 SimpleCells (Tn-glycoform). Interestingly, we found that all potential Ser/Thr O-glycosites in TRs derived from secreted mucins and most glycosites from transmembrane mucins were almost fully occupied, whereas TRs from a subset of transmembrane mucins were less efficiently processed. We further used the mucin TR reporters to characterize cleavage sites of glycoproteases StcE (secreted protease of C1 esterase inhibitor from EHEC) and BT4244, revealing more restricted substrate specificities than previously reported. Finally, we conducted a bottom–up analysis of isolated ovine submaxillary mucin, which supported our findings that mucin TRs in general are efficiently O-glycosylated at all potential glycosites. This study provides insight into O-glycosylation of mucins and mucin-like domains, and the strategies developed open the field for wider analysis of native mucins.     

Recombinant human tumor antigen MUC1 expressed in insect cells: structure and immunogenicity

MUC1, a member of the mucin family of molecules, is a transmembrane glycoprotein abundantly expressed on human ductal epithelial cells and tumors originating from those cells. MUC1 expressed by malignant cells is aberrantly O-glycosylated. Differences in O-glycosylation of the tandem repeat region of MUC1 make tumor and normal forms of this antigen immunologically distinct. The tumor-specific glycoform is, therefore, expected to be a good target for immunotherapy and a good immunogen for generation of antitumor immune responses. We have generated a renewable source of this glycoform by expressing MUC1 cDNA in Sf-9 insect cells using a baculovirus vector. This form of MUC1 (BV-MUC1) is O-glycosylated at a very low level, approximately 0.3% (w/w), and this is not due to the lack of appropriate glycosylotransferases in insect cells. Peptidyl GalNAc-transferases isolated from Sf-9 cells were able to glycosylate in vitro a synthetic MUC1 peptide as efficiently as the transferases isolated from human milk. Neither preparation of peptidyl GalNAc-transferases, however, was able to glycosylate BV-MUC1. This underglycosylated recombinant MUC1 mimics underglycosylated MUC1 on human tumor cells and could serve as an immunogen to stimulate responses that would recognize MUC1 on tumor cells. To test this we immunized mice with Sf-9 cells expressing BV-MUC1. Sera from immunized mice recognized MUC1 on human tumor cells. We also generated MUC1-specific T cells that proliferated in response to synthetic MUC1 peptide.     

Purification and characterization of a human pancreatic adenocarcinoma mucin.

Pancreatic mucins consist of core proteins that are decorated with carbohydrate structures. Previous studies have identified at least two physically distinct populations of mucins produced by a pancreatic adenocarcinoma cell line (HPAF); one is the MUC1 core protein, which includes an oligosaccharide structure identified by a monoclonal antibody (MAb) recognizing the DU-PAN-2 epitope. In this study, we purified and characterized a second mucin fraction, which also shows reactivity with the DU-PAN-2 antibody, but which has an amino acid composition that is not consistent with the MUC1 core protein. This new mucin was purified by ammonium sulfate precipitation, molecular sieve chromatography, and density gradient centrifugation. It eluted in the void volume of a Sepharose 4B column together with an associated low molecular weight protein, which could be further resolved. The mucin is highly polyanionic due to numerous sulfated and sialylated saccharide chains. Carbohydrate analyses of the purified mucin showed the presence of galactose, glucosamine, galactosamine, and sialic acid, but no mannose, glucose, or uronic acid. The purified and deglycosylated mucin shows no reactivity with anti-MUC1 apomucin antibody, but reacts with antiserum against deglycosylated tracheal mucins and antiserum against the MUC4 tandem repeat peptide. Analysis of mucin expression in HPAF cells revealed high levels of MUC1 and MUC4 mRNA, and moderate levels of MUC5AC and MUC5B mRNA. The amino acid composition of the purified mucin shows a high degree of similarity to the MUC4 core protein.     

High density O-glycosylation on tandem repeat peptide from secretory MUC1 of T47D breast cancer cells.

The site-specific O-glycosylation of MUC1 tandem repeat peptides from secretory mucin of T47D breast cancer cells was analyzed. After affinity isolation on immobilized BC3 antibody, MUC1 was partially deglycosylated by enzymatic treatment with alpha-sialidase/beta-galactosidase and fragmented by proteolytic cleavage with the Arg-C-specific endopeptidase clostripain. The PAP20 glycopeptides were isolated by reversed phase high pressure liquid chromatography and subjected to the structural analyses by quadrupole time-of-flight electrospray ionization mass spectrometry and to the sequencing by Edman degradation. All five positions of the repeat peptide were revealed as O-glycosylation targets in the tumor cell, including the Thr within the DTR motif. The degree of substitution was estimated to average 4.8 glycans per repeat, which compares to 2.6 glycosylated sites per repeat for the mucin from milk (Müller, S., Goletz, S., Packer, N., Gooley, A. A., Lawson, A. M., and Hanisch, F.-G. (1997) J. Biol. Chem. 272, 24780-24793). In addition to a modification by glycosylation, the immunodominant DTR motif on T47D-MUC1 is altered by amino acid replacements (PAPGSTAPAAHGVTSAPESR), which were revealed in about 50% of PAP20 peptides. The high incidence of these replacements and their detection also in other cancer cell lines imply that the conserved tandem repeat domain of MUC1 is polymorphic with respect to the peptide sequence.     

Large scale identification of proteins, mucins, and their O-glycosylation in the endocervical mucus during the menstrual cycle

The mucus filling the human cervical opening blocks the entry to the uterus, but this has to be relative and allow for the sperm to penetrate at ovulation. We studied this mucus, its content of proteins and mucins, and the mucin O-glycosylation in cervical secretions before, during, and after ovulation. Cervical mucosal secretions from 12 subjects were collected, reduced-alkylated, separated with polyacrylamide or agarose/polyacrylamide gel electrophoresis, and stained with silver, Alcian blue, or Coomassie Blue stain. Protein and mucin bands from before and during ovulation were digested and subsequently analyzed by nano-LC-FT-ICR MS and MS/MS. We identified 194 proteins after searches against the NCBI non-redundant protein database and an in-house mucin database. Three gel-forming (MUC5B, MUC5AC, and MUC6) and two transmembrane mucins (MUC16 and MUC1) were identified. For the analysis of mucin O-glycosylation, separated mucins from six individuals were blotted to PVDF membranes, and the O-glycans were released by reductive beta-elimination and analyzed with capillary HPLC-MS and -MS/MS. At least 50 neutral, sialic acid-, and sulfate-containing oligosaccharides were found. An increase of GlcNAc-6GalNAcol Core 2 structures and a relative decrease of NeuAc residues are typical for ovulation, and NeuAc-6GalNAcol and NeuAc-3Gal- epitopes are typical for the non-ovulatory phases. The cervical mucus at ovulation is thus characterized by a relative increase in neutral fucosylated oligosaccharides. This comprehensive characterization of the mucus during the menstrual cycle suggests mucin glycosylation as the major alteration at ovulation, but the relation to the altered physicochemical properties and sperm penetrability is still not understood.     

Cloning and sequencing of a human pancreatic tumor mucin cDNA

A monospecific polyclonal antiserum against deglycosylated human pancreatic tumor mucin was used to select human pancreatic mucin cDNA clones from a lambda gt11 cDNA expression library developed from a human pancreatic tumor cell line. The full-length 4.4-kilobase mucin cDNA sequence included a 72-base pair 5'-untranslated region and a 307-base pair 3'-untranslated region. The predicted amino acid sequence for this cDNA revealed a protein of 122,071 daltons containing 1,255 amino acid residues of which greater than 60% were serine, threonine, proline, alanine, and glycine. Approximately two-thirds of the protein sequence consisted of identical 20-amino acid tandem repeats which were flanked by degenerate tandem repeats and nontandem repeat sequences on both the amino-terminal and carboxyl-terminal ends. The amino acid sequence also contained five putative N-linked glycosylation sites, a putative signal sequence and transmembrane domain, and numerous serine and threonine residues (potential O-linked glycosylation sites) outside and within the tandem repeat position. The cDNA and deduced amino acid sequence of the pancreatic mucin sequence was over 99% homologous with a mucin cDNA sequence derived from breast tumor mucin, even though the native forms of these molecules are quite distinct in size and degree of glycosylation.     

Nuclear magnetic resonance-based dissection of a glycosyltransferase specificity for the mucin MUC1 tandem repeat

The human glycoprotein MUC1 mucin plays a critical role in cancer progression. Breast, ovarian, and colon cancer cells often display unique cell-surface antigens corresponding to aberrantly glycosylated forms of the MUC1 tandem repeat. In this report, (15)N- and (13)C-labeled forms of a recombinant MUC1 construct containing five tandem repeats were used as substrates to define the order and kinetics of addition of N-acetylgalactosamine (GalNAc) moieties by a recombinant active form of the human enzyme UDP-GalNAc:polypeptide N-acetylgalactosaminyltransferase I (ppGalNAc-T1; residues 40-559). Heteronuclear NMR experiments were performed to assign resonances associated with the two serines (Ser5 and Ser15) and three threonines (Thr6, Thr14, and Thr19) present in the 20-residue long MUC1 repeat. The kinetics and order of addition of GalNAc moieties (Tn antigen) on the MUC1 construct by human ppGalNAc-T1 were subsequently dissected by NMR spectroscopy. Threonine 14 was shown to be rapidly glycosylated by ppGalNAc-T1 with an initial rate of 25 microM/min, followed by Thr6 (8.6 microM/min). The enzyme also modified Ser5 at a slower rate (1.7 microM/min), an event that started only after the glycosylation of Thr14 and Thr6 side chains was mostly completed. Ser15 and Thr19 remained unglycosylated by ppGalNAc-T1. Corresponding O-glycosylation sites within all five tandem repeats were simultaneously modified by ppGalNAc-T1, suggesting that each repeat behaves as an independent substrate unit. This study demonstrated that the hydroxyl oxygens of Thr14 and to a lesser extent Thr 6 represent the two dominant substrates modified by ppGalNAc-T1 within the context of a complex MUC1 peptide substrate. More importantly, the availability of defined isotopically labelled MUC1 glycopeptide substrates and the relative simplicity of their NMR spectra will facilitate the analysis of other transferases within the O-glycosylation pathways and the rational design of tumor-associated MUC1 antigens.