Skip Navigation


Glycobiology Advance Access originally published online on June 22, 2005
Glycobiology 2005 15(11):1111-1124; doi:10.1093/glycob/cwi099
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
15/11/1111    most recent
cwi099v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (11)
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Engelmann, K.
Right arrow Articles by Hanisch, F.-G.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Engelmann, K.
Right arrow Articles by Hanisch, F.-G.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

© The Author 2005. Published by Oxford University Press. All rights reserved. For permissions, please e-mail: journals.permissions@oupjournals.org

Transmembrane and secreted MUC1 probes show trafficking-dependent changes in O-glycan core profiles

Katja Engelmann2, Carol L. Kinlough3, Stefan Müller2, Hani Razawi2, Stephan E. Baldus4, Rebecca P. Hughey1,3 and Franz-Georg Hanisch1,2

2 Center of Biochemistry Medical Facility and Center for Molecular Medicine, University of Cologne, Joseph-Stelzmann-Str. 52, 50931 Köln, Germany; 3 Department of Medicine, Renal-Electrolyte Division, University of Pittsburgh School of Medicine, Pittsburgh, PA 15213; and 4 Institute of Pathology, University of Cologne, Joseph-Stelzmann-Str. 9, 50931 Köln, Germany

1 To whom correspondence should be addressed; e-mail: franz.hanisch{at}uni-koeln.de and hughey{at}dom.pitt.edu

Received on May 3, 2005; revised on June 19, 2005; accepted on June 19, 2005


   Abstract
Top
 Abstract
Introduction
 Results
 Discussion
 Materials and methods
 Acknowledgments
 References
 
The human mucin MUC1 is expressed both as a transmembrane heterodimeric protein complex that recycles via the trans-Golgi network (TGN) and as a secreted isoform. To determine whether differences in cellular trafficking might influence the O-glycosylation profiles on these isoforms, we developed a model system consisting of membrane-bound and secretory-recombinant glycosylation probes. Secretory MUC1-S contains only a truncated repeat domain, whereas in MUC1-M constructs this domain is attached to the native transmembrane and cytoplasmic domains of MUC1 either directly (M0) or via an intermitting nonfunctional (M1) or functional sperm protein–enterokinase–agrin (SEA) module (M2); the SEA module contains a putative proteolytic cleavage site and is associated with proteins receiving extensive O-glycosylation. We showed that MUC1-M2 simulates endogenous MUC1 by recycling from the cell surface of Chinese hamster ovary (CHO) mutant ldlD14 cells through intracellular compartments where its glycosylation continues. The profiles of O-linked glycans on MUC1-S secreted by epithelial EBNA-293 and MCF-7 breast cancer cells revealed patterns dominated by core 2-based oligosaccharides. In contrast, the respective membrane-shed probes expressed in the same cells showed a complete shift to patterns dominated by sialyl core 1. In conclusion, glycan core profiles reflected the subcellular trafficking pathways of the secretory or membranous probes and the modifying activities of the resident glycosyltransferases.

Key words: mass spectrometry / MUC1 / O-glycosylation / trafficking / transmembrane protein


   Introduction
Top
 Abstract
 Introduction
Results
 Discussion
 Materials and methods
 Acknowledgments
 References
 
Mucins are high-molecular mass glycoproteins with peptide-repeat domains and a dense O-glycosylation. The first human mucin among the 20 species identified so far is MUC1, a type1 transmembrane protein (Gendler and Spicer, 1995Go). The core protein of the mucin is processed in the lumen of the endoplasmic reticulum (ER) by proteolytic cleavage of the ectodomain component from the transmembrane component (Hilkens and Buijs, 1988Go; Parry et al., 2001Go). The two fragments form a stable but noncovalent, heterodimeric complex that is heavily O-glycosylated during transit of the Golgi complex. After exposure on the plasma membrane, MUC1 is internalized via endocytosis and further sialylated before it is recycled to the cell surface (Litvinov and Hilkens, 1993Go). In fact, clathrin-mediated endocytosis of MUC1 is modulated by its glycosylation state; MUC1 with shorter O-glycans is internalized faster than MUC1 with full-length O-glycans (Altschuler et al., 2000Go). A portion of the membrane-associated MUC1 is shed, presumably by a second proteolytic cleavage of the ectodomain. In this way, MUC1 can enter body fluids and mucus layers on epithelial surfaces. Another natural isoform of the mucin, MUC1sec, has been described, which represents a splice variant of the same gene lacking the transmembrane and cytosolic domains, but includes the entire exon 2 and portions of the intron 2 sequence (Smorodinsky et al., 1996Go). This secretory isoform passes the Golgi compartments only once on its way to the extracellular space and hence should exhibit a lower degree of sialylation than the membrane-bound isoform.

The question has been raised whether the differences in cellular trafficking between these two isoforms do affect the qualitative and quantitative aspects of O-glycosylation. To address this question, we have generated a panel of four fusion proteins, which all contain a section of the MUC1 repeat domain, but differ with respect to the content of a transmembrane domain and a functional sperm protein–enterokinase–agrin (SEA) module domain (Bork and Patthy, 1995Go), which is often found in proximity to the heavily O-glycosylated domains. The SEA module in the endogenous mucin contains the ER-processing site (Parry et al., 2001Go), which is absent in the probe MUC1-M0, present in MUC1-M1, but nonfunctional due to a lack of essential peptide motifs upstream of the domain (Wreschner et al., 2002Go), and fully functional in the probe MUC1-M2, analogous to the endogenous MUC1. To encounter different repertoires of glycosyltransferases, we transfected two cell lines, MCF-7 breast cancer cells, and human epithelial kidney cells EBNA-293, that exhibit distinct O-glycosylation patterns (Wuttke et al., 2001Go; Müller and Hanisch, 2002Go). The mutant CHO cell line ldlD14 was used as a third cellular model (Kingsley et al., 1986Go), because it offered the advantage of control over O-glycosylation by sugar supplementation. This cell line, however, is unable to buildup core 2 glycans and was used only for establishing evidence for probe endocytosis and recycling through compartments where glycosylation continues. Our results demonstrate qualitative and quantitative alterations of O-glycosylation, which are associated with the different cellular trafficking pathways of membranous or secretory isoforms of the same protein.


   Results
Top
 Abstract
 Introduction
 Results
Discussion
 Materials and methods
 Acknowledgments
 References
 
Generation of MUC1-M fusion proteins containing the transmembrane region
Based on a previously described secretory MUC1 fusion protein (designated MFP6 and identical to MUC1-S), which contained the BM40 signal sequence peptide, an N-terminal hexahistidine, and a Myc tag as well as six nonconserved tandem repeats of MUC1 (Müller and Hanisch, 2002Go), we designed and generated three novel MUC1-M probes including the transmembrane and cytoplasmic domains (Figure 1A and B) as well as the Cys-Gln-Cys motif at the junction of the cytosolic and transmembrane domains necessary for surface expression (Pemberton et al., 1996Go). To elucidate the possible influences of the proximal SEA domain, we generated probes, which were either lacking such a domain (MUC1-M0) or contained a nonfunctional (MUC1-M1) or functional cleavage site within the SEA domain (MUC1-M2). The cleavage site in MUC1-M1 is nonfunctional, because the probe lacks essential peptide motifs upstream of the SEA domain (Wreschner et al., 2002Go). The correctness of the open reading frame (ORF) was revealed by DNA sequencing of the three novel probes (sequence data given as amino acid sequences in Figure 1A). The four probes MUC1-S, MUC1-M0, -M1, and -M2 designed for secretory or membranous expression (Figure 1B) were transfected into EBNA-293 and MCF-7 cells, and the fusion proteins were detected with anti-MUC1 antibodies or tag-specific antibodies by immunoblotting (Figure 2) and immunocytology (Figure 3). Repeated attempts to transfect MCF-7 cells with MUC1-M2 construct were not successful.



View larger version (44K):
[in this window]
[in a new window]
  		Fig. 1. Structures of recombinant MUC1 probes. (A) Alignments of amino acid sequences of membranous (MUC1-M0, MUC1-M1, MUC1-M2) and secretory fusion proteins (MUC1-S). The recombinant fusion protein MUC1-S, previously designated MFP6 in Müller and Hanisch (2002)Go, and the newly generated constructs MUC1-M0, MUC1-M1, and MUC1-M2 were aligned from the N- or C-terminal domains for easier discrimination of identical and nonidentical sequences. As common structural features, all four constructs contain the cleavable N-terminal BM40-signal sequence (in italics), a hexahistidine tag (bold face), a myc tag (underlined), and six nonconserved MUC1 tandem repeat units (grey-shaded area, amino acid substitutions represented in bold face). The membrane constructs contain a transmembrane sequence (bold underlined, including the CQC motif necessary for surface expression) and the C-terminal cytosolic domain. Referring to the C-terminal alignment, the MUC1-M0 construct lacks a sperm protein–enterokinase–agrin (SEA) domain (bold dashed line), whereas MUC1-M1 and MUC1-M2 exhibit SEA domains including a nonfunctional (MUC1-M1) or functional cleavage site (MUC1-M2) (double underlined). The cleavage site in MUC1-M1 is nonfunctional, because upstream located peptide motifs are lacking (dashed line). (B) Schematic view of secretory versus membrane-bound and shed MUC1 fusion proteins. Four constructs were generated from the MUC1 protein core. The secretory MUC1-S lacks all domains downstream of the repeat domain, whereas MUC1-M constructs have a transmembrane and cytosolic domain in common. According to our working hypothesis, the secretory MUC1-S is expected to pass through the Golgi complex only once before it is released into the extracellular space. The MUC1-M constructs are exposed on the plasma membrane, presumably become reinternalized by endocytosis, analogous to the endogenous MUC1, and may recycle through the trans-Golgi network (TGN) where they receive further O-glycosylation. The transmembrane constructs are shed from the cell surface by unknown mechanisms, which presumably involve proteolysis (Thathiah et al., 2003Go).

 


View larger version (38K):
[in this window]
[in a new window]
  		Fig. 2. The transmembrane constructs MUC1-M0, MUC1-M1, and MUC1-M2 can be isolated from the cell pellets and from culture supernatants. Immunoblot analyses of membrane-bound fusion proteins MUC1-M0, -M1, and -M2 in EBNA-293 cell pellets are shown (top panels). The blots were stained with antibody to the C-terminal end using anti-cytosolic domain-specific antibody H-295. Two versions of the transmembrane small subunit, as revealed for the MUC1-M2 construct (processed and unprocessed), have previously been described for endogenous MUC1 (Ren et al., 2004Go). Immunoblot analyses of shed fusion proteins MUC1-M0, -M1, and -M2 in EBNA-293 cell culture supernatants are shown (bottom panels). The blots were stained with antibodies to both ends of the fusion proteins. Polyclonal antibody (pAb) H295 recognizes the cytosolic domain (C-term), and mAb C595 recognizes the ectodomain repeats (N-term). Endogenous MUC1 from MCF-7 cell pellets served as a positive control for staining of the transmembrane component (endo MUC1).

 


View larger version (40K):
[in this window]
[in a new window]
  		Fig. 3. Localization of membrane-bound MUC1 fusion proteins on the plasma membrane. Immunofluorescence microscopy of fusion proteins expressed in EBNA-293 cells (A, MUC1-M0; B, MUC1-M1; and C, MUC1-M2) was performed using polyclonal antibody (pAb) to the C-terminal cytosolic domain (H-295). MUC1-M2 ectodomain was stained with the repeat peptide-specific antibody C595 (D).

 

MUC1-M0, MUC1-M1, and MUC1-M2 display membranous topology
Western blot analyses were performed after cell lysis using an anti-cytosolic domain antibody and revealed expression of unprocessed MUC1-M glycoproteins with apparent molecular masses at 65 kD (M0) and 75 kD (M1) (Figure 2, upper panel). Accordingly, the cleavage site within the SEA domain of MUC1-M1, although present, is not used. In contrast to these two probes, MUC1-M2 was detected at the apparent molecular mass of 35 kD, which is identical to that of the transmembrane component (small subunit) of endogenous MUC1 (MCF-7). A band at 27 kD likely represents the small subunit of the overexpressed MUC1-M2 that has not left the ER for further glycan processing (Figure 2, upper panel). Shed membrane constructs in culture supernatants were devoid of a cytosolic domain indicating proteolytic cleavage of surface-exposed ectodomains (Figure 2, lower panel). Staining of the ectodomains was possible with the repeat peptide-specific monoclonal antibody (mAb) C595 and gave rise to broad bands expanding over a larger range of molecular masses (M0 and M2: 50–75 kD, M1: 35–75 kD), which indicated glycosylation-dependent size heterogeneity (Figure 2, lower panel). Expression levels, tested by immunocytochemistry, remained constant as long as puromycin was added to the culture media. The membrane-integrated fusion proteins from cell pellets were isolated in the same way as described for the secretory or shed isoforms. However, the yields were estimated to reach only 5 µg per 1–2 x 108 cells. These quantities were not sufficient for further characterization.

Immunocytological staining was performed using antibodies to both ends of the proteins, the anti-MUC1 mAb C595 (N-term), the anti-MUC1 cytosolic domain polyclonal antibody (pAb) H-295 (C-term), and an anti-myc mAb (N-term). Immunofluorescence staining of permeabilized cells and confocal laser-scanning microscopy confirmed membranous expression of all three MUC1-M isoforms (Figure 3).

MUC1-M2 recycles from the cell surface through the trans-Golgi network
To determine whether MUC1-M2 does recycle after reaching the cell surface, through compartments that contain glycosyltransferases, we expressed MUC1-M2 in glycosylation-defective CHO (ldlD14) cells. The ldlD14 cells lack the 4-epimerase required to convert UDP-Glc/GlcNAc to UDP-Gal/GalNAc and cannot complete N-glycan processing to complex type or initiate O-glycan synthesis; the phenotype can be rescued by the addition of Gal and GalNAc to the culture medium (Kingsley et al., 1986Go). We showed previously that MUC1 (with 22 repeats) expressed in ldlD14 cells is unstable without O-glycans but is stabilized by the addition of truncated O-glycans when synthesized in the presence of GalNAc alone (Altschuler et al., 2000Go). However, this form of the MUC1 has a notably different mobility on sodium dodecyl sulfate (SDS) gels than on MUC1 synthesized with full glycans in the presence of Gal and GalNAc. We reasoned that MUC1-M2 would have a similar alteration in its SDS-gel mobility when synthesized in the presence of only GalNAc, and further glycosylation during recycling through the trans-Golgi network (TGN) would be indicated by a subsequent shift in its mobility. This is exactly what we observed (Figure 4). Synthesis of [35S]MUC1-M2, [35S]MUC1-S, or [35S]MUC1 (22 repeats) in ldlD14 cells by metabolic labeling with [35S]Met/Cys in the presence of GalNAc revealed a radioactive band that migrated differently on SDS gels than on the same [35S]-labeled isoform produced in ldlD14 cells in the presence of both GalNAc and Gal; this was true for the membrane-associated MUC1-M2 and MUC1(22 repeats) and for the secreted MUC1-S (Figure 4A). When ldlD14 cells expressing MUC1-M2 were metabolically labeled with [35S]Met/Cys, the cell surface-biotinylated MUC1-M2 synthesized in the presence of GalNAc (GA) alone (Figure 4B, lane 2) also had a significantly faster mobility than cell surface MUC1-M2 synthesized in the presence of Gal and GalNAc (GA + G) (lane 1). When these biotinylated cells (+GA) were returned to culture for an additional 2.5 h, the mobility of the cell surface-biotinylated MUC1-M2 was dramatically decreased (Figure 4B, lane 3), although the total level of [35S]MUC1-M2 was unchanged. This altered profile is consistent with the addition of O-glycans to the cell surface MUC1-M2 during its subsequent recycling.



View larger version (60K):
[in this window]
[in a new window]
  		Fig. 4. MUC1-M2 recycles from the cell surface through the trans-Golgi network (TGN). (A) Glycosylation defective CHO cells (ldlD14) expressing MUC1-S (S), MUC1-M2 (M2), MUC122 repeats (M) or no construct (C, control) were metabolically labeled with [35S]Met/Cys in the presence of either GalNAc and Gal (GalNAc + Gal) or GalNAc alone (30 min pulse and 90 min chase). MUC1 and MUC1-based probes were immunoprecipitated from cell extracts or culture media with anti-MUC1 B27.29 antibody and subjected to sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE). The altered mobilities of the large subunit of MUC1 (LS) or MUC1-M2 (M2L) and the small transmembrane subunit (SS) are noted to the left of the gel (arrows). The asterisk emphasizes the ladder of bands observed for the small subunit expressed in CHO cells. Note that MUC1-S is found primarily in the culture medium under these labeling conditions, although a smaller precursor band remains in the cell (~30 kDa). (B) Cells expressing [35S]MUC1-M2 in the presence of GalNAc and Gal (GA + G, lane 1) or GalNAc alone (GA, lanes 2 and 3) were treated with membrane-impermeant biotin on ice and returned to culture for t = 0 (lanes 1 and 2) or t = 2.5 h (lane 3, GA + chase) with GA + G before the immunoprecipitation of MUC1-M2 from cell extracts. Biotinylated MUC1-M2 was recovered with avidin-conjugated beads from immunoprecipitates and analyzed with a BioRad Molecular Imager after SDS–PAGE. The mobility of Amersham rainbow molecular weight markers (kDa) are shown on the right of both gels. (C) Lane profiles from B are aligned to emphasize the altered mobility of surface MUC1-M2 after 2.5 h of recycling.

 

MUC1-M fusion proteins are shed from the cell surface
In culture supernatants, none of the recombinant membrane probes were stained with the C-terminus specific antibody H295 indicating proteolytic cleavage of the ectodomains during shedding (Figure 2, lower panel). The ectodomains of shed MUC1-M0, -M1, and -M2 were stained diffusely with mAb C595 as smears with apparent molecular masses between 40 and 75 kD.

Affinity-isolated fractions containing MUC1 fusion proteins were pooled and purified by reversed phase-high performance liquid chromatography (rp-HPLC) on a C4-silica column. MUC1-positive fractions were tested for homogeneity by SDS–polyacrylamide gel electrophoresis (SDS–PAGE) and silver staining as well as western blotting (Figure 5). The silver-stained gel revealed the presence of two bands with apparent molecular masses at 70 and 50 kD (Figure 5, lane 1), which potentially represent differentially shed isoforms. Glycosylation of the isoforms was demonstrated with the digoxigenin glycan detection assay (Figure 5, lane2), and MUC1 epitopes were stained with mAb C595 revealing an unresolved smear in the same molecular mass range (Figure 5, lane 3).



View larger version (69K):
[in this window]
[in a new window]
  		Fig. 5. Gel electrophoresis and western blot analysis of shed MUC1-M2 after reversed-phase chromatography. Affinity-purified fusion protein MUC1-M2 (EBNA-293) was chromatographed on a Vydac protein C4 column. Eluting proteins were tested on immunoblots for MUC1 repeat peptide epitopes using C595 antibody, and positive fractions were combined before gel electrophoresis and western blot analysis. The absence of contaminating protein was demonstrated by silver staining of the gel (lane 1). The blot was stained for glycoproteins with the DIG glycan detection kit (lane 2), and for MUC1 repeats using C595 (lane 3).

 

Based on integrated peak areas of rechromatographed fractions and of a 50-µg standard of MUC1-S (EBNA-293), yields of pure MUC1-M fusion proteins isolated from MCF-7 or EBNA-293 cells ranged from 20 to 30 µg per liter culture supernatant.

Membrane-associated fusion proteins display distinct O-glycosylation compared with secretory probes
To determine the relative proportions of individual glycans in a quantitatively reliable fashion, we liberated the O-linked oligosaccharides from the fusion proteins by hydrazinolysis and analyzed the glycans by quantitative normal-phase HPLC as fluorescent derivatives with 1 mol/mol 2-aminobenzamide (2-AB) label (Figure 6). About 5 µg of fusion protein corresponding to ~1–2 nmol of carbohydrate were applied per analysis, and chromatographic peaks were quantitated by area integration (Tables I and II). To test the reliability of hydrazinolysis as a method for quantitative glycan liberation, we liberated glycans on MUC1-S from EBNA-293 cells by an independent method: nonreductive alkali-catalyzed ß-elimination in a flow system (Karlsson and Packer, 2002Go). The glycan profile (not shown) revealed was qualitatively identical and quantitatively similar to that after hydrazinolysis (the proportion of core 2- vs. core 1-based glycans was found to be ~1:1).



View larger version (22K):
[in this window]
[in a new window]
  		Fig. 6. High performance liquid chromatography (HPLC) profiles of 2-aminobenzamide-labeled (2-AB-labeled) O-glycans from fusion proteins MUC1-S and MUC1-M1 (EBNA-293). (Top left) Native glycans from MUC1-S, (bottom left) native glycans from MUC1-M1, (top right) desialylated glycans from MUC1-S, and (bottom right) desialylated glycans from MUC1-M1. Desialylation was performed enzymatically by treatment of the samples with sialidase from Clostridium perfringens overnight. For structural identification, refer to Tables I and II. Peaks marked by asterisk represent nonidentified artefacts which do not contain glycans or glycan derivatives. Peak 0 refers to GalNAc derived from NeuAc2-6GalNAc. GalNAc in the native, nontreated glycan fractions was largely lost during paper chromatography of the 2-AB-labeled glycans.

 

View this table:
[in this window]
[in a new window]
  		Table I. O-Glycan profiles of fusion proteins expressed in epithelial kidney cells EBNA-293

 

View this table:
[in this window]
[in a new window]
  		Table II. O-Glycan profiles of fusion proteins expressed in breast cancer cells MCF-7

 

Only the major glycans were identified by comparative chromatography and numbered in the HPLC profile shown in Figure 6 in accordance with an earlier report (Müller and Hanisch, 2002Go). The quantitative data obtained for the four probes are summarized in Table I for EBNA-293 cells and Table II for MCF-7 cells. The secretory glycoforms (MUC1-S) expressed in the two cell lines are characterized by prevalent core 2 expression (Tables I and II). Particularly, the glycans from MUC1-S (MCF-7) exhibit >80% core 2-based structures, a considerable proportion of which is fucosylated (Table II). Both cell lines are distinct with respect to the proportion of acidic glycans, because the profile from EBNA-293 cells is characterized by a higher degree of sialylation (~40%), whereas that from MCF-7 cells is nearly devoid of sialylated structures (~5%). The profiles from both cell lines agree with respect to the low expression of sialyl-Tn (HPLC peak 3: NeuAc{alpha}2-6GalNAc) and sialylated core 1 trisaccharides [HPLC peak 4: NeuAc{alpha}2-3Galß1-3GalNAc and HPLC peak 5: NeuAc{alpha}2-6(Galß1-3)GalNAc].

This pattern observed for secretory glycoforms of the fusion protein changes drastically on the respective membrane-derived fusion proteins (Tables I and II; Figure 6), which correspondingly show decreased core 2 (in particular, HPLC peaks 2 and 6) and increased core 1 expression (in particular, HPLC peak 4). The shed MUC1-M1 and MUC1-M2 exhibit preferential core 1 expression with NeuAc{alpha}2-3Galß1-3GalNAc (HPLC peak 4) as the dominating species and a general increase of sialylated glycans (Tables I and II). This shift from predominantly core 2 to nearly exclusive core1 expression is independent of the cellular system and independent of the type of membrane fusion protein (M0, M1, M2), since also MUC1-M0 expressed in MCF-7 cells shows a dramatically altered profile with nearly 100% acidic glycans, ~60% of which are identical to NeuAc{alpha}2-3Galß1-3GalNAc. Hence, neither the presence of a SEA domain nor that of a functional-processing site was found to be associated with the observed shift in O-glycosylation.

Structural characterization of O-linked glycans by mass spectrometric sequencing and linkage analysis
Monosaccharide compositions of 2-AB-labeled glycans in terms of N-acetylneuraminic acid (NeuAc), desoxyhexose (dHex), hexose (Hex), and N-acetylhexosamine (HexNAc) content were deduced from pseudomolecular ions measured in matrix-assisted laser-desorption mass spectrometry (MALDI-MS; Table III). The compositional data of MUC1-S–derived glycans were in agreement with the structures assigned by chromatographic profiling (Tables I and II; Figure 6), that were based on the retention times of authentic oligosaccharide standards as a reference and the chromatographic shifts after sequential exoglycosidase treatments (Müller and Hanisch, 2002Go). For detailed structural characterization, the glycans were cleaved from fusion proteins by reductive ß-elimination and analyzed as permethylated glycan alditols by electrospray ionization tandem mass spectrometry (ESI-MS/MS; Figure 7, Table IV). The structural assignments for sodiated molecular ions were based on prominent fragment ions of the Y, Z and B, C series, those for the protonated molecular ions on strong B-ion series (nomenclature according to Domon and Costello, 1988Go) and were in agreement with the tentative structures listed in Tables I and II. Linkage analysis by gas-chromatography-MS (GC-MS) identification of partially methylated alditol acetates (Table V) revealed largely overlapping patterns for both cell lines (terminal fucose and galactose, 2-linked and 3-linked galactose, 3-linked and 3,6-linked GalNAc-ol, and 4-linked GlcNAc). Only minor differences became apparent by the detection of 3,4-linked GlcNAc in the glycan fraction from MCF-7 cells indicating the presence of Lewis-type terminal sequences (refer also to sequence information from ESI-MS/MS in Table IV).


View this table:
[in this window]
[in a new window]
  		Table III. Matrix-assisted laser-desorption mass spectrometry (MALDI-MS) of native 2-aminobenzamide-labeled (2-AB-labeled) glycans from secretory MUC1-S

 


View larger version (21K):
[in this window]
[in a new window]
  		Fig. 7. Identification of permethylated O-linked glycans by electrospray mass spectrometry. Electrospray mass spectrometry of permethylated glycans for structural identification was performed on a Qtof2 as described in the Materials and methods section. The mass spectra of four major glycans derived from MUC1-S (EBNA-293) or MUC1-S (MCF-7), respectively, are shown. For assignment of fragments, refer to Table IV. A, Hex-(Hex-HexNAc-)HexNAc-ol; B, Hex-(dHex-[Hex-HexNAc-])HexNAc-ol; C, NeuAc-Hex-(NeuAc-)HexNAc-ol; and D, NeuAc-Hex-(Hex-HexNAc-)HexNAc-ol and Hex-(NeuAc-Hex-HexNAc-)HexNAc-ol).

 

View this table:
[in this window]
[in a new window]
  		Table VI. Electrospray mass spectrometry (ESi- MS) in the tandem MS (MS/MS) mode of permethylated glycan alditols

 

View this table:
[in this window]
[in a new window]
  		Table V. Linkage analysis of O-glycans by gas-chromatography-MS (GC-MS) identification of partially methylated alditol acetates

 


   Discussion
Top
 Abstract
 Introduction
 Results
 Discussion
Materials and methods
 Acknowledgments
 References
 
O-Glycosylation is controlled by a variety of genetic and epigenetic parameters including the cellular repertoire of the glycosyltransferases with their substrate and reaction specificities (Brockhausen et al., 1995Go), the peptide sequence and conformation around putative O-glycosylation sites (Clausen and Bennett, 1996Go), the availability of cosubstrates, the direct or indirect competition of glycosyltransferases for substrate (Brockhausen, 1999Go), positive or negative effects of glycans on proximal glycosylation reactions (Hanisch et al., 1999Go), and finally the subcellular trafficking of the glycoprotein (Huang and Snider, 1993Go; Litvinov and Hilkens, 1993Go; Volz et al., 1995Go). The latter parameter has not been well studied so far due to the lack of appropriate model systems, which would allow the recombinant expression of secretory and membranous isoforms and the detailed structural analysis of their O-linked glycans.

We have generated such tagged probes based on truncated versions of the MUC1 tandem repeat domain in fusion with the transmembrane and cytosolic domains of the MUC1 mucin and containing the Cys-Gln-Cys motif at the junction of the cytosolic and transmembrane domains necessary for surface expression (Pemberton et al., 1996Go). We demonstrated by immunofluorescence microscopy that the transmembrane domain-containing constructs are actually present at the plasma membrane and were able to isolate the constructs as shed glycoproteins from the culture supernatants. An enzyme activity identical to the tumor necrosis factor-{alpha} converting enzyme (TACE) was recently shown to be implicated in proteolysis of the MUC1 ectodomain from uterine epithelial cell membranes (Thathiah et al., 2003Go). The same authors later described a membrane-type matrix metalloprotease as contributing to MUC1 shedding, which is stimulated by pervanadate (Thathiah and Carson, 2004Go). However, others have shown that the Gly–Ser bond representing the cleavage site of the MUC1 SEA domain is needed for effective shedding of the extracellular subunit (Lillehoj et al., 2003Go) pointing to the possibility that several alternative shedding mechanisms may be involved in the release of membrane-bound MUC1.

The most striking observation of this study is the finding that MCF-7 breast cancer cells, which have previously been shown to express predominantly neutral, core 2-based polylactosamines on the secreted MUC1 probe (Müller and Hanisch, 2002Go), change their O-glycosylation pattern to preferential core 1 synthesis on the shed membranous probes MUC1-M0 and -M1. Associated with this is a shift to the expression of sialylated glycans, NeuAc{alpha}2-3Galß1-3GalNAc being the most prominent species (Tables I and II). There are several possible mechanisms, which could explain the observed phenomenon. One possibility is the different trafficking routes of the protein probes and the fact that membrane-bound probes recycle through the TGN and have repeated exposure to glycosyltransferases in this compartment. The formation of core 2-based glycans requires the activity of C2-GnTs, which compete with sialyltransferases, in particular, with ST3Gal-I, for the common substrate Galß1-3GalNAc. Although the core-specific glycosyltransferase, C2-GnT1, is found predominantly in the cis-Golgi (Dalziel et al., 2001Go), the sialyltransferases, in particular ST3Gal-I (Whitehouse et al., 1997Go), are found predominantly in the TGN. There is also evidence that polypeptide GalNAc-transferases ppGalNAc-T1, -T2, and –T3 are coexpressed throughout the Golgi complex including the TGN (Rottger et al., 1998Go) and could account for continued initiation of O-glycosylation of proteins during recycling. Accordingly, the membrane-bound MUC1 constructs, which may pass the TGN more than once, should display a glycosylation profile, which is more determined by the enzymatic repertoire of only the later cisternae. That is exactly what we found; recycling probes would encounter GalNAc transferases and sialyltransferases in the TGN but not the core2-GnT1. We confirmed for the MUC1-M2 probe that it does reinternalize by endocytosis and recycles to the plasma membrane and that this process is associated with further O-glycosylation. Accordingly, the recombinant probe simulates cellular trafficking of endogenous MUC1 (Litvinov and Hilkens, 1993Go) and can be regarded as an authentic model for studying recycling and concomitant O-glycosylation.

Alternatively, the distinct difference in O-glycan profiles between the secreted and membranous probes could result from the influence of membrane-bound chaperone-like proteins in the Golgi complex, which differentially influence the activities of the competing glycosyltransferases. The finding that initial GalNAc addition to a protein substrate can trigger the activities of other polypeptide GalNAc-transferases with distinct site specifities via their lectin domains (Hassan et al., 2000Go; Hanisch et al., 2001Go) can serve as a model for such chaperone-like activities of glycosyltransferases. These hypothetical considerations do not exclude alternative explanations for the differential O-glycosylation of the recombinant probes.

The findings of this study could explain discrepant results from two previous studies on membrane-bound MUC1 from MCF-7 cells and secretory glycoforms of native endogenous MUC1 (Lloyd et al., 1996Go; Müller and Hanisch, 2002Go). Strikingly, membrane-bound glycoforms of the endogenous mucin preferentially expressed sialylated core 1 structures (Lloyd et al., 1996Go). In contrast, we could demonstrate prevalent core 2 expression and a low degree of sialylation for our secretory MUC1 probe from MCF-7 cells (Müller and Hanisch, 2002Go) and apocrine-secreted MUC1 from normal mammary epithelial cells (Hanisch et al., 1989Go).

This study has provided detailed structural evidence that the O-glycosylation profile of a protein is not solely determined by the genetic repertoire of glycosyltransferases in a cell but, moreover, by the intracisternal topology of the glycoprotein substrate and the respective cellular trafficking pathways of secretory and membrane-bound proteins. The reported findings indicate the importance of membrane topology on O-glycosylation, but they also demonstrate that neither the presence of a SEA domain nor that of a functional cleavage site within such domains is necessary for the observed shift in O-glycosylation.


   Materials and methods
Top
 Abstract
 Introduction
 Results
 Discussion
 Materials and methods
Acknowledgments
 References
 
Materials
Monoclonal anti-MUC1 antibody C595 was kindly provided by Dr. Graem Denton, University of Nottingham, UK. Monoclonal anti-MUC1 antibody B27.29 was a gift from Fujiribio Diagnostics (Malvern, PA). Both antibodies bind to the RPAP motif within the repeat peptide of the variable number of tandem repeats (VNTR) domain. Polyclonal rabbit anti-MUC1 cytosolic domain antibody H-295 was from Santa Cruz Biotechnology (Heidelberg, Germany) and is described by the manufacturer to be raised against amino acids 961–1255 mapping at the C-terminus of human MUC1.

Construction of the MUC1 fusion proteins (MUC1-M0, MUC1-M1, and MUC1-M2)
Generation of the MUC1-S construct (previously designated MFP6) was described earlier (Müller and Hanisch, 2002Go). For the generation of the MUC1-M constructs, the MUC1 transmembrane and cytosolic domains and sections of the extracellular domain of the transmembrane MUC1 component containing a putative proteolytical cleavage site were generated by polymerase chain reaction (PCR) amplification of the respective cDNA sequence from pancreatic HPAF cells (200 ng DNA, 12.5 pM primer, 2 mM dNTPs [Peqlab, Erlangen, Germany], 1 U Taq DNA polymerase, and buffer [Qiagen, Hilden, Germany] in a 50-µL reaction mix). Using the forward (for) and reverse (rev) oligonucleotide primer pairs (MUC1-M0/for: 5'-CTCGAGGTGCCATT TCCTTTCTCTGC; MUC1-M0/rev: 5'-GGGATCCGAGCCCCCACAACACTTC; MUC1-M1/for: 5'-CTCGAGGAAGATCCCAGCACCGACT; MUC1-M1/rev: 5'-GGATCC- CTCACCAGCCCAAACAGG; MUC1-M2/for: 5'-CTCGAGCACAGCACTTCTCCCCAGT; and MUC1-M2rev: 5'-GGATCCCTCACCAGCCCAAACAGG), a 5'-terminal XhoI and a 3'-terminal BamHI restriction site were introduced into the amplified MUC1 DNA sequences. The reaction mixtures were incubated for 5 min at 95°C and then cycled 30 times for 30 s at 94°C, 40 s at 60°C, and 40 s at 72°C with a final extension for 10 min at 72°C.

PCR products were separated by electrophoresis in 1.4% agarose gel, eluted from the gel, and purified using QiaEx Gel Extraction Kit (Qiagen). The cleared DNA samples were cloned by TA ligation into the pGEM T-easy vector (Promega, Mannheim, Germany) to analyze the correctness of the ORF by sequencing of the plasmid DNA using T7 (5'-TAATACGACTCAC TATAGGG) or the SP6-primer site of the vector (5'-GATTTAGGTGACACTATAG). For in-frame insertion of the MUC1-M0, MUC1-M1, or MUC1-M2 constructs from the pGEM-T easy vector into the expression vector pCEP-PU, the XhoI and the BamHI restriction sites were used. A successful subcloning was verified by sequencing using SeqU (5'-TAGTGAACCGTCAGATCT) and Rev2-oligonucleotides (5'-CTGGATCCGGCCTTGCC).

The modified pCEP-PU expression vector (Müller and Hanisch, 2002Go) already contained the signal peptide of the BM40 extracellular matrix protein, a hexahistidine tag, a myc sequence as well as six tandem repeats from the MUC1 VNTR region. Restriction enzymes were obtained from New England Biolabs (Frankfurt, Germany). The amino acid sequences of the four contructs are given in Figure 1.

Cell culture
Breast tumor cell line MCF-7 was obtained from the American Type Culture Collection (Manassas, VA), whereas the human epithelial kidney cell line EBNA-293 was purchased from Invitrogen (Karlsruhe, Germany). Both cell lines were cultured in Dulbeccos minimal essential medium, supplemented with Glutamax I, 10% fetal calf serum, 100 IU penicillin, and 100 µg/mL of streptomycin in 150 cm2 flasks by incubation at 37°C in the presence of 5% CO2. After transfection of the cells with the constructs, 5 µg/mL of puromycin (Sigma, St. Louis, MO) was added to all media. Every 2–3 days, the media were replaced. Confluent cells were passaged using trypsin/ethylenediaminetetraacetic acid treatment. All other cell culture reagents were obtained from Gibco-Invitrogen (Karlsruhe, Germany). Glycosylation-defective CHO cells (ldlD14) were grown in DMEM/HAM’S 12 (1:1 vol.) with 3% fetal calf serum, 100 IU penicillin/100 µg/mL of streptomycin, and 5 µg/mL of puromycin. The medium was supplemented with 1 mM GalNAc and 0.1 mM Gal.

Expression of MUC1-M0, -M1, and -M2 constructs
The pCEP-PU vector is replicated epichromosomally in eukaryotic cells and contains a puromycin resistance gene as well as a Rous Sarcoma Virus promotor. Using the Superfect lipofection reagent (Qiagen), the cell lines were transfected according to the manufacture’s instructions. Transfected cells were cultured in standard media for 24 h, before the addition of 5 µg/mL of puromycin to the media. Nontransfected cells died within 5–7 days, whereas the transfectants were selected in the presence of 5 µg/mL of puromycin.

Immunofluorescence staining of transfected cells
Chamberslides (Nunc, Wiesbaden, Germany) were coated with poly-l-lysine solution (Sigma, München, Germany) for 5 min at room temperature (RT). Cells were detached from culture flasks by trypsin treatment, resuspended in complete media, and incubated for 2 h on the precoated chamberslides at 37°C, 5% CO2. The cell culture supernatant was removed, and cells were fixed with 2% paraformaldehyde for 10 min at RT before they were permeabilized with 0.2% TritonX-100 (Fluka, Taufkirchen, Germany) for 1 min at RT. Nonspecific protein binding was blocked by addition of 1% bovine serum albumin (BSA; Sigma) in phosphate-buffered saline (PBS) pH 7.4 for 30 min at RT, followed by incubation of the primary monoclonal antibodies (H-295, Santa Cruz Biotechnology, 2 µg/mL in 0.02% TritonX-100/PBS, or C595, 3.5 µg/mL in 0.02% TritonX-100/PBS) for 1 h at RT. Antibody dilution buffer without antibody was applied as negative control. As secondary antibodies, the biotinylated anti-rabbit IgG (E0353, Dako, Hamburg, Germany; 1:1000 diluted in 0.5% BSA/PBS for detection of H-295 binding) or the fluorescently labeled anti-mouse IgG Alexa Fluor 488 (Molecular Probes, Mobitec, and GmbH, Göttingen, Germany; diluted 1:1000 in 0.02% TritonX-100/PBS for detection of mAb C595) was incubated for 1 h at RT. Biotinylated secondary antibody was fluorescently stained by the addition of Extravidin-FITC (Sigma; 1:1000) for 30 min in the dark. Between each of the incubation steps, specimens were washed three times in 0.02% TritonX-100/PBS. Slides were mounted in fluorescent mounting medium (Dako). The microscopic analysis was performed on a confocal laser-scanning microscope Leica TCS SL (Leica, Heidelberg, Germany).

Endocytosis and recycling experiments
Glycosylation-defective CHO cells (ldlD14) obtained from Monty Krieger (MIT, Cambridge, MA) were transiently transfected using Lipofectamine 2000 as directed by the manufacturer (Invitrogen, Carlsbad, CA). The following day, the ldlD14 cells were metabolically labeled with [35S]Met/Cys for 30 min (~106 cells were labeled in 0.5 mL of medium lacking Met and Cys, but containing 25 mCi of 35S-Met/Cys with ~1000 Ci per mmol) and chased in media containing Met/Cys for 90 min before cell surface biotinylation on ice, as previously described (Kinlough et al., 2004Go). Cells were then returned to culture for 0 or 2.5 h. MUC1-M2 was immunoprecipitated from detergent extracts with mouse mAb B27.29 before recovery of biotinylated MUC1-M2 with avidin-conjugated beads for SDS–PAGE, as described previously (Kinlough et al., 2004Go). [35S]MUC1-M2 on SDS gels was analyzed with a BioRad Personal Molecular Imager FX and Quantity One software (Hercules, CA). Sugars were present in the cell culture medium before or after biotinylation as indicated (1 mM GalNAc ± 0.1 mM Gal, both purchased from Sigma, St. Louis, MO).

Isolation and purification of MUC1 fusion proteins
Supernatants from confluent cell cultures were collected, centrifuged at 1000 g at 4°C for 10 min, and dialyzed against several changes of dH2O. Subsequently, supernatants were adjusted to 50 mM sodium phosphate, pH 8.0, 200 mM sodium chloride, 1 mM imidazol, 5 mM 2-mercaptoethanol, 10% ethanol and centrifuged at 9500 rpm for 30 min at 4°C. After an additional filtration, the supernatants were loaded onto a column of 5 mL of Ni2+-nitrilotriacetic acid agarose (Qiagen) and recycled two times for binding of the fusion proteins via their his tags. The column was washed with 30 mL of 50 mM sodium phosphate, pH 6.5, 500 mM sodium chloride, 10 mM imidazol, 10 mM 2-mercaptoethanol, and 10% ethanol, and equilibrated with 10 mL of 20 mM sodium phosphate, pH 6.5. Proteins were eluted with 10 x 1.5 mL of 0.1% trifluoroacetic acid (TFA) in ddH2O, 20% acetonitrile, or with a gradient of imidazole from 20 to 250 mM and detected by western blotting as above. Anti-myc positive fractions were combined and subjected to further purification by HPLC on a reversed-phase column (Vydac 214TP3410, MZ Analysentechnik, Mainz, Germany). Samples were eluted in a gradient of 2–80% acetonitrile in 0.1% aqueous TFA over 30 min using a flow rate of 1 mL/min. Protein was spectrophotometrically registered at 214 nm, and each peak was separately collected for analysis by western blotting and staining with monoclonal anti-MUC1 (C595) antibody.

Western blot analysis
Cell lysates or protein fractions were separated by SDS gel electrophoresis (3% stacking gel, 12% running gel) in a Mini Protean 3 cell (BioRad, München, Germany) for 50 min at 200 V. After equilibration in blotting buffer protein was transferred from gels to nitrocellulose membranes (Schleicher and Schüll, Einbeck, Germany) with a semi-dry transfer cell (BioRad) for 40 min at 3 mA/cm2. Nonspecific protein binding was blocked by incubation of the blot membrane in 5% milk powder in PBS for 1 h at RT. The immunoblots were stained with anti-MUC1 antibody H-295 (Santa Cruz, Loxo-GmbH, Dossenheim, Germany), mAb C595, or anti-myc antibody (Invitrogen, Karlsruhe, Germany). The immunocomplexes were detected with horseradish peroxidase (HRP)-conjugated secondary antibody (Dako) and enhanced chemiluminescence (Roche, Mannheim, Germany). Glycoproteins were detected with the DIG glycan detection kit (Roche).

Cleavage of O-linked glycans from glycosylated MUC1 probes
Nonreductive release of O-glycans from proteins was performed by treatment with anhydrous hydrazine (Sigma) as described previously (Patel and Parekh, 1994Go). About 10–15 µg of extensively dried and salt-free fusion proteins were subjected to treatment with 50 µL of hydrazine for 5 h at 60°C under dry argon, followed by re-N-acetylation for 15 min on ice using 2 mM acetic anhydride (Fluka, Buchs, Switzerland) in saturated bicarbonate. The reaction mixture was desalted with AG50W x 8 (H+ form, BioRad) cation exchange resin in a batch procedure. Subsequently, the glycans were dried by vacuum centrifugation overnight. To exclude methodological artefacts of hydrazinolytic glycan liberation, we applied an alternative nonreductive method in control experiments, which is based on alkali-catalyzed ß-elimination of glycans in a continuous flow system (Karlsson and Packer, 2002Go). For structural studies, the glycans were liberated by reductive ß-elimination according to a protocol applicable to microscale samples (Schulz et al., 2002Go).

2-AB labeling of O-glycan pools
Dried glycans were labeled with 1 M 2-AB in acetic acid and 2 M sodium cyanoborohydride in dimethyl sulfoxide for 2 h at 60°C (Bigge et al., 1995Go) and purified by paper chromatography. Glycans were eluted in 500 µL of ddH20 and stored at –20°C.

Analysis of 2-AB-labeled O-glycans by normal-phase HPLC
2-AB-labeled O-glycan pools were analyzed using normal-phase HPLC on a System Gold HPLC workstation (Beckman Instruments, Muenchen, Germany) equipped with a Shimadzu RF-10AXL fluorescence detector. Data acquisition was performed by Beckman Nouveau software. Excitation wavelength was set to 330 nm and for emission to 420 nm. Samples were dissolved in acetonitrile/water (3:1) and injected onto a polymer-based aminophase column (Astec NH2 polymer, 5 µm, 4.6 x 250 mm, Alltech, Unterhaching, Germany). Analysis parameters were described previously (Müller and Hanisch, 2002Go). To assess the relative amounts of the HPLC peak 5 and 12 components, the native glycan fractions were run sequentially in gradients of acetonitrile and water (elution of neutral oligosaccharides) or acetonitrile and aqueous 250 mM formiate, pH 4.4 (elution of acidic oligosaccharides).

Exoglycosidase digestion of 2-AB-labeled O-glycans
Aliquots of 2-AB-labeled glycans were digested overnight at 37°C with 5 U neuraminidase (Clostridium perfringens) and 5 U {alpha}1–2 fucosidase (Xanthomonas manihotis) in 50 mM sodium citrate, pH 4.5, and 10 µg/mL of BSA. Enzymes and buffers were obtained from New England Biolabs.

Analysis of O-glycans by MS
MALDI-MS was performed on a Bruker Reflex IV instrument (Bruker Daltonics, Bremen, Germany). The 2-AB-labeled glycan samples (~500 pmol per µL) contained in 50% acetonitrile/water were applied to the stainless steel target by mixing a 0.7 µL of aliquot with 1.4 µL of matrix (20 mg of 2,5-dihydroxy benzoic acid in 1 mL acetonitrile/0.1% TFA, 1:2). Analyses were performed by positive-ion detection in the reflectron mode. Ionization of cocristallized analytes was induced with a pulsed nitrogen laser beam (337 nm), and the ions were accelerated in a field of 20 kV and reflected at 23 kV.

ESI-MS data were acquired on a Q-Tof 2 quadrupole-time of flight mass spectrometer (Waters, Eschborn, Germany) equipped with a Z spray source. ESI(Qtof) MS was performed in the positive-ion mode using previously described conditions (Morelle et al., 2004Go). The permethylated glycans were dissolved in 80% methanol containing 1% acetic acid before loading 3 µL of the solution into a nanospray capillary coated with a thin layer of gold/palladium, tip inner diameter 2 µm (Proxeon, Odense, Denmark). A potential of 800 V was applied to a nanoflow tip. The drying gas used was nitrogen, and the collision gas was argon, with the collision gas pressure maintained at 0.5 bar. The cone voltage was set at 50 V. Collision energies varied in accordance with the type of molecular ion (M + Na: 50–75 V; M + H: 15–30 V).

Chemical derivatization for GC-MS and ESI-MS/MS
Permethylation was performed using the sodium hydroxide procedure of Ciacanu and Kerek as modified by Annumula (Annumula and Taylor, 1992Go). Partially methylated alditol acetates were prepared by hydrolysis of permethylated glycans with 2 M TFA (Fluka) for 2 h at 121°C, by reduction with 10 mg/mL of sodium borodeuteride (Sigma) in 2 M aqueous ammonium hydroxide at RT for 2 h, and acetylation with acetic anhydride (Fluka) at 100°C for 1 h (Albersheim et al., 1967Go). The partially methylated alditol acetates were extracted with chloroform water, dried, and analyzed as a dichloromethane solution by GC-MS on a Fison MD800 (Thermo Electron, Dreieich, Germany) using a 15 m RTX5- SILMS column from Restek (Bad Homburg, Germany) and a temperature gradient from 60 to 100°C (40°C/min) followed by 100–280°C (10°C/min).


   Acknowledgments
Top
 Abstract
 Introduction
 Results
 Discussion
 Materials and methods
 Acknowledgments
References
 
This investigation was supported by grants from the Deutsche Forschungsgemeinschaft (Ha 2092/8-1 and Ha 2092/10-1), from the NIH (1RO1 CA84106), and from the Deutsche Krebshilfe grant 70-2975-Ba 2 to FGH, and by NIH grants DK54787 and P50-DK56490 to RPH. The authors acknowledge the skillful technical assistance of Mrs. Kirsten Ottenberg and Mr. Rick Stremple.


   Abbreviations
 
C2-GnT, core 2 ß6GlcNAc transferase; CHO, Chinese hamster ovary; dHex, desoxyhexose; ER, endoplasmic reticulum; ESI-MS, electrospray ionization mass spectrometry; Fuc, fucose; Gal, galactose; GalNAc, N-acetylgalactosamine; GC-MS, gas-chromatography-mass spectrometry; Hex, hexose; HexNAc, N-acetylhexosamine; mAb, monoclonal antibody; MALDI-MS, matrix-assisted laser-desorption mass spectrometry; MS/MS, tandem mass spectrometry; NeuAc, N-acetylneuraminic acid; ORF, open reading frame; ppGalNAc-T, polypeptide {alpha}-N-acetylgalactosaminyltransferase; rp-HPLC, reversed phase-high performance liquid chromatography; RT, room temperature; SDS, sodium dodecyl sulfate; SEA, sperm protein-enterokinase-agrin; ST3Gal-I, Galß1-3GalNAc {alpha}2-3-sialyltransferase I; TFA, trifluoroacetic acid; TGN, trans-Golgi network; VNTR, variable number of tandem repeats


   References
Top
 Abstract
 Introduction
 Results
 Discussion
 Materials and methods
 Acknowledgments
 References
 
Albersheim, P., Nevins, D.J., English, P.D., and Karr, A. (1967) A method for the analysis of sugars in plant cell-wall polysaccharides by gas-liquid chromatography. Carbohydr. Res., 5, 340–345.[CrossRef]

Altschuler, Y., Kinlough, C.L., Poland, P.A., Bruns, J.B., Apodaca, G., Weisz, O.A., and Hughey, R.P. (2000) Clathrin-mediated endocytosis of MUC1 is modulated by its glycosylation state. Mol. Biol. Cell, 11, 819–831.[Abstract/Free Full Text]

Annumula, K.R. and Taylor, P.B. (1992) A comprehensive procedure for preparation of partially methylated alditol acetates from glycoprotein carbohydrates. Anal. Biochem., 203, 101–108.[CrossRef][ISI][Medline]

Bigge, J.C., Patel, T.P., Bruce, J.A., Goulding, P.N., Charles, S.M., and Parekh, R.B. (1995) Nonselective and efficient fluorescent labeling of glycans using 2-amino benzamide and anthranilic acid. Anal. Biochem., 230, 229–238.[CrossRef][ISI][Medline]

Bork, P. and Patthy, L. (1995) The SEA module: a new extracellular domain associated with O-glycosylation. Protein Sci., 4, 1421–1425.[Abstract]

Brockhausen, I. (1999) Pathways of O-glycan biosynthesis in cancer cells. Biochim. Biophys. Acta, 1473, 67–95.[Medline]

Brockhausen, I., Yang, J.M., Burchell, J., Whitehouse, C., and Taylor-Papadimitriou, J. (1995) Mechanisms underlying aberrant glycosylation of MUC1 mucin in breast cancer cells. Eur. J. Biochem., 233, 607–617.[ISI][Medline]

Clausen, H. and Bennett, E.P. (1996) A family of UDP-GalNAc: polypeptide N-acetylgalactosaminyl-transferases control the initiation of mucin-type O-linked glycosylation. Glycobiology, 6, 635–646.[Free Full Text]

Dalziel, M., Whitehouse, C., McFarlane, I., Brockhausen, I., Gschmeissner, S., Schwientek, T., Clausen, H., Burchell, J., and Taylor-Papadimitriou, J. (2001) The relative activities of the C2GnT1 and ST3Gal-I glycosyltransferases determine O-glycan structure and expression of a tumor-associated epitope on MUC1. J. Biol. Chem., 276, 11007–11015.[Abstract/Free Full Text]

Domon, B. and Costello, C.E. (1988) A systematic nomenclature for carbohydrate fragmentations in FAB-MS/MS spectra of glycoconjugates. Glycoconj. J., 5, 397–409.[CrossRef][ISI]

Gendler, J.S. and Spicer, A.P. (1995) Epithelial mucin genes. Annu. Rev. Physiol., 57, 607–634.[CrossRef][ISI][Medline]

Hanisch, F.-G., Uhlenbruck, G., Peter-Katalinic, J., Egge, H., Dabrowski, J., and Dabrowski, U. (1989) Structures of neutral O-linked polylactosaminoglycans on human skim milk mucins. A novel type of linearly extended poly-N-acetyllactosamine backbones with Galß (1–4) GlcNAcß (1–6) repeating units. J. Biol. Chem., 264, 872–883.[Abstract/Free Full Text]

Hanisch, F.-G., Müller, S., Hassan, H., Clausen, H., Zachara, N., Gooley, A.A., Paulsen, H., Alving, K., and Peter-Katalinic, J. (1999) Dynamic epigenetic regulation of initial O-glycosylation by UDP-N-acetylgalactosamine: peptide N-acetylgalactosaminyltransferases. Site-specific glycosylation of MUC1 repeat peptide influences the substrate qualities at adjacent or distant Ser/Thr positions. J. Biol. Chem., 274, 9946–9954.[Abstract/Free Full Text]

Hanisch, F.-G., Reis, C., Clausen, H., and Paulsen, H. (2001) Evidence for glycosylation-dependent activities of polypeptide N-acetylgalactosaminyltransferases rGalNAc-T2 and -T4 on mucin glycopeptides. Glycobiology, 11, 731–740.[Abstract/Free Full Text]

Hassan, H., Reis, C., Bennett, E.P., Mirgorodskaya, E., Roepstorff, P., Hollingsworth, M.A., Burchell, J., Taylor-Papadimitriou, J., and Clausen, H. (2000) The lectin domain of UDP-N-acetyl-D-galactosamine: polypeptide N-acetylgalactosaminyltransferase-T4 directs its glycopeptide specificities. J. Biol. Chem., 275, 38197–38205.[Abstract/Free Full Text]

Hilkens, J. and Buijs, F. (1988) Biosynthesis of MAM-6, an epithelial sialomucin. Evidence for involvement of a rare proteolytic cleavage step in the endoplasmic reticulum. J. Biol. Chem., 263, 4215–4222.[Abstract/Free Full Text]

Huang, K.M. and Snider, M.D. (1993) Glycoprotein recycling to the galactosyltransferase compartment of the Golgi complex. J. Biol. Chem., 268, 9302–9310.[Abstract/Free Full Text]

Karlsson, N.G. and Packer, N.H. (2002) Analysis of O-linked reducing oligosaccharides released by an in-line flow system. Anal. Biochem., 305, 173–185.[CrossRef][ISI][Medline]

Kingsley, D.M., Kozarsky, K.F., Hobbie, L., and Krieger, M. (1986) Reversible defects in O-linked glycosylation and LDL receptor expression in a UDP-Gal/UDP-GalNAc 4-epimerase deficient mutant. Cell, 44, 749–759.[CrossRef][ISI][Medline]

Kinlough, C.L., Poland, P.A., Bruns, J.B., Harkleroad, K.L., and Hughey, R.B. (2004) MUC1 membrane trafficking is modulated by multiple interactions. J. Biol. Chem., 279, 53071–53077.[Abstract/Free Full Text]

Lillehoj, E.P., Han, F., and Kim, K.C. (2003) Mutagenesis of a Gly-Ser cleavage site in MUC1 inhibits ectodomain shedding. Biochem. Biophys. Res. Commun., 307, 743–749.[CrossRef][ISI][Medline]

Litvinov, S.V. and Hilkens, J. (1993) The epithelial sialomucin, episialin, is sialylated during recycling. J. Biol. Chem., 268, 21364–21371.[Abstract/Free Full Text]

Lloyd, K.O., Burchell, J., Kudryashov, V., Yin, B.W.T., and Taylor-Papadimitriou, J. (1996) Comparison of O-linked carbohydrate chains in MUC-1 mucin from normal breast epithelial cell lines and breast carcinoma cell lines. Demonstration of simpler and fewer glycan chains in tumor cells. J. Biol. Chem., 271, 33325–33334.[Abstract/Free Full Text]

Morelle, W., Faid, V., and Michalski, J.-C. (2004) Structural analysis of permethylated oligosaccharides using electrospray ionization quadrupole time-of-flight tandem mass spectrometry and deutero-reduction. Rapid Commun. Mass Spectrom., 18, 2451–2464.[CrossRef][ISI][Medline]

Müller, S. and Hanisch, F.-G. (2002) Recombinant MUC1 probe authentically reflects cell-specific O-glycosylation profiles of endogenous breast cancer mucin. High density and prevalent core 2-based glycosylation. J. Biol. Chem., 277, 26103–26112.[Abstract/Free Full Text]

Parry, S., Silverman, H.S., McDermott, K., Willis, A., Hollingsworth, M.A., and Harris, A. (2001) Identification of MUC1 proteolytic cleavage sites in vivo. Biochem. Biophys. Res. Commun., 283, 715–720.[CrossRef][ISI][Medline]

Patel, T.P. and Parekh, R.B. (1994) Release of oligosaccharides from glycoproteins by hydrazinolysis. Methods Enzymol., 230, 57–66.[ISI][Medline]

Pemberton, L.F., Rughetti, A., Taylor-Papadimitriou, J., and Gendler, S.J. (1996) The epithelial mucin MUC1 contains at least two discrete signals specifying membrane localization in cells. J. Biol. Chem., 271, 2332–2340.[Abstract/Free Full Text]

Ren, J., Agata, N., Chen, D., Li, Y., Yu, W., Huang, L., Raina, D., Chen, W., Kharbanda, S., and Kufe, D. (2004) Human MUC1 carcinoma-associated protein confers resistance to genotoxic anticancer agents. Cancer Cell, 5, 163–175.[CrossRef][ISI][Medline]

Rottger, S., White, J., Wandall, H.H., Olivo, J.C., Stark, A., Bennett, E.P., Whitehouse, C., Berger, E.G., Clausen, H., and Nilsson, T. (1998) Localization of three human polypeptide GalNAc-transferases in HeLa cells suggests initiation of O-linked glycosylation throughout the Golgi apparatus. J. Cell Sci., 111, 45–60.[Abstract]

Schulz, B.L., Packer, N.H., and Karlsson, N.G. (2002) Small-scale analysis of O-linked oligo- saccharides from glycoproteins and mucins separated by gel electrophoresis. Anal. Chem., 74, 6088–6097.[Medline]

Smorodinsky, N., Weiss, M., Hartmann, M.L., Baruch, A., Harness, E., Yaakobovitz, M., Keydar, I., and Wreschner, D.H. (1996) Detection of a secreted MUC1/SEC protein by MUC1 isoform specific monoclonal antibodies. Biochem. Biophys. Res. Commun., 228, 115–121.[CrossRef][ISI][Medline]

Thathiah, A. and Carson, D.D. (2004) MT1-MMP mediates MUC1 shedding independent of TACE/ADAM17. Biochem. J., 382, 363–373.[CrossRef][ISI][Medline]

Thathiah, A., Blobel, C.P., and Carson, D.D. (2003) Tumor necrosis factor-alpha converting enzyme/ADAM 17 mediates MUC1 shedding. J. Biol. Chem., 278, 3386–3394.[Abstract/Free Full Text]

Volz, B., Orberger, G., Porwoll, S., Hauri, H.P., and Tauber, R. (1995) Selective reentry of recycling cell surface glycoproteins to the biosynthetic pathway in human hepatocarcinoma HepG2 cells. J. Cell Biol., 130, 537–551.[Abstract/Free Full Text]

Whitehouse, C., Burchell, J., Gschmeissner, S., Brockhausen, I., Lloyd, K.O., and Taylor-Papadimitriou, J. (1997) A transfected sialyltransferase that is elevated in breast cancer and localizes to the medial/trans-Golgi apparatus inhibits the development of core2-based O-glycans. J. Cell Biol., 137, 1229–1241.[Abstract/Free Full Text]