An Autocatalytic Cleavage in the C Terminus of the Human MUC2 Mucin Occurs at the Low pH of the Late Secretory Pathway

During purification of a recombinant MUC2 C terminus expressed in CHO-K1 cells, the protein was partly cleaved when buffers with a pH of 6.0 were used. When buffers with higher pH values were used, less cleavage was found. Disulfide bonds held the two fragments generated together as these were only observed after reduction. Edman sequencing of the C-terminal 110-kDa fragment revealed that the cleavage had occurred at an Asp-Pro bond, a site described previously to generate the so-called "link peptide" after disulfide bond reduction. In vitro studies on the conditions for cleavage showed that it occurred in a time-dependent manner at a pH below 6.0. Furthermore, the reaction was not enzyme-mediated as it occurred in pure preparations of the MUC2 C terminus and was not inhibited by protease inhibitors. When expressed in the mucin producing cell line LS 174T, the C terminus was cleaved to a higher extent compared with the CHO-K1 cells. Neutralizing the secretory pathway with either NH4Cl or bafilomycin A1 inhibited this cleavage. Altogether, our results suggest that the cleavage is an autocatalytic reaction that occurs in the acidic environment of the late secretory pathway. Furthermore, the cleavage produced a new, reactive C terminus that has the potential to attach the mucin to itself or other molecules. Because a pH below 6 can be reached in the late secretory pathway and on mucosal surfaces, the cleavage and possible cross-linking are likely to be of biological importance. During purification of a recombinant MUC2 C terminus expressed in CHO-K1 cells, the protein was partly cleaved when buffers with a pH of 6.0 were used. When buffers with higher pH values were used, less cleavage was found. Disulfide bonds held the two fragments generated together as these were only observed after reduction. Edman sequencing of the C-terminal 110-kDa fragment revealed that the cleavage had occurred at an Asp-Pro bond, a site described previously to generate the so-called "link peptide" after disulfide bond reduction. In vitro studies on the conditions for cleavage showed that it occurred in a time-dependent manner at a pH below 6.0. Furthermore, the reaction was not enzyme-mediated as it occurred in pure preparations of the MUC2 C terminus and was not inhibited by protease inhibitors. When expressed in the mucin producing cell line LS 174T, the C terminus was cleaved to a higher extent compared with the CHO-K1 cells. Neutralizing the secretory pathway with either NH4Cl or bafilomycin A1 inhibited this cleavage. Altogether, our results suggest that the cleavage is an autocatalytic reaction that occurs in the acidic environment of the late secretory pathway. Furthermore, the cleavage produced a new, reactive C terminus that has the potential to attach the mucin to itself or other molecules. Because a pH below 6 can be reached in the late secretory pathway and on mucosal surfaces, the cleavage and possible cross-linking are likely to be of biological importance. Mucins are large glycoproteins with the carbohydrates accounting for up to 90% of its molecular mass (1Forstner J.F. Oliver M.G. Sylvester F.A. Blaser M.J. Smith P.D. Ravdin J.I. Greenberg H.B. Guerrant R.L. Infections of the Gastrointestinal Tract. Raven Press, Ltd., New York1995: 71-88Google Scholar, 2Perez-Vilar J. Hill R.L. J. Biol. Chem. 1999; 274: 31751-31754Abstract Full Text Full Text PDF PubMed Scopus (398) Google Scholar). These sugars are mainlyO-linked to Ser and Thr residues localized to regions called mucin domains. These domains contain tandemly repeated sequences (rich in Ser, Thr, and Pro) that vary in number, length, and amino acid composition depending on the particular mucin (3Gendler S.J. Spicer A.P. Annu. Rev. Physiol. 1995; 57: 607-634Crossref PubMed Scopus (870) Google Scholar). Mucins account for the major part of the mucous layers covering the epithelial surfaces throughout the body, where they are believed to have several functions including protection and lubrication of the underlying epithelia. There are two major types of mucins, membrane-bound and secreted. In human, eight membrane-bound (MUC1, MUC3A, MUC3B, MUC4, MUC12, MUC13, MUC16, and MUC17) (4Gendler S. Lancaster C. Taylor-Papadimitriou J. Duhig T. Peat N. Burchell J. Pemberton L. Lalani E. Wilson D. J. Biol. Chem. 1990; 265: 15286-15293Abstract Full Text PDF PubMed Google Scholar, 5Gum J.R. Hicks J.W. Swallow D.M. Lagace R.L. Byrd J.C. Lamport D.T. Siddiki B. Kim Y.S. Biochem. Biophys. Res. Commun. 1990; 171: 407-415Crossref PubMed Scopus (315) Google Scholar, 6Pratt W.S. Crawley S. Hicks J. Ho J. Nash M. Kim Y.S. Gum J.R. Swallow D.M. Biochem. Biophys. Res. Commun. 2000; 275: 916-923Crossref PubMed Scopus (73) Google Scholar, 7Williams S.J. McGuckin M.A. Gotley D.C. Eyre H.J. Sutherland G.R. Antalis T.M. Cancer Res. 1999; 59: 4083-4089PubMed Google Scholar, 8Moniaux N. Nollet S. Porchet N. Degand P. Laine A. Aubert J.P. Biochem. J. 1999; 338: 325-333Crossref PubMed Scopus (222) Google Scholar, 9Williams S.J. Wreschner D.H. Tran M. Eyre H.J. Sutherland G.R. McGuckin M.A. J. Biol. Chem. 2001; 276: 18327-18336Abstract Full Text Full Text PDF PubMed Scopus (261) Google Scholar, 10Yin B.W.T. Lloyd K.O. J. Biol. Chem. 2001; 276: 27371-27375Abstract Full Text Full Text PDF PubMed Scopus (499) Google Scholar, 11Gum Jr., J.R. Crawley S.C. Hicks J.W. Szymkowski D.E. Kim Y.S. Biochem. Biophys. Res. Commun. 2002; 291: 466-475Crossref PubMed Scopus (171) Google Scholar) and five secreted mucins (MUC2, MUC5B, MUC5AC, MUC6, and MUC7) (12Gum Jr., J. Hicks J. Toribara N. Siddiki B. Kim Y. J. Biol. Chem. 1994; 269: 2440-2446Abstract Full Text PDF PubMed Google Scholar, 13Desseyn J.-L. Buisine M.-P. Porchet N. Aubert J.-P. Laine A. J. Biol. Chem. 1998; 273: 30157-30164Abstract Full Text Full Text PDF PubMed Scopus (67) Google Scholar, 14Li D. Gallup M. Fan N. Szymkowski D.E. Basbaum C.B. J. Biol. Chem. 1998; 273: 6812-6820Abstract Full Text Full Text PDF PubMed Scopus (156) Google Scholar, 15Toribara N. Roberton A. Ho S. Kuo W. Gum E. Hicks J. Gum Jr., J. Byrd J. Siddiki B. Kim Y. J. Biol. Chem. 1993; 268: 5879-5885Abstract Full Text PDF PubMed Google Scholar, 16Bobek L. Tsai H. Biesbrock A. Levine M. J. Biol. Chem. 1993; 268: 20563-20569Abstract Full Text PDF PubMed Google Scholar) have been identified so far. Two additional ones cannot be classified (MUC8 and MUC11) (7Williams S.J. McGuckin M.A. Gotley D.C. Eyre H.J. Sutherland G.R. Antalis T.M. Cancer Res. 1999; 59: 4083-4089PubMed Google Scholar, 17Shankar V. Pichan P. Eddy Jr., R.L. Tonk V. Nowak N. Sait S.N. Shows T.B. Schultz R.E. Gotway G. Elkins R.C. Gilmore M.S. Sachdev G.P. Am. J. Respir. Cell Mol. Biol. 1997; 16: 232-241Crossref PubMed Scopus (128) Google Scholar). The secreted mucins can be further sub-divided as being either gel-forming (MUC2, MUC5B, MUC5AC, and MUC6) or not (MUC7). The ability to form gels is dependent on the capacity of monomers to form polymeric structures. The intermolecular interactions are of disulfide-bond nature and mediated through the N- and C-terminal Cys-rich domains of mucins (reviewed in Ref. 2Perez-Vilar J. Hill R.L. J. Biol. Chem. 1999; 274: 31751-31754Abstract Full Text Full Text PDF PubMed Scopus (398) Google Scholar). Both these domains show considerable sequence homology to the human prepro-von Willebrand factor (vWf) 1The abbreviations used are: vWfvon Willebrand factorPSMporcine submaxillary mucinERendoplasmic reticulumGFPgreen fluorescent proteinmAbmonoclonal antibodyAPalkaline phosphatasePBSphosphate-buffered salinePMSFphenylmethylsulfonyl fluorideNEMN-ethylmaleimideH3heavy chain 3SMCsialomucin complexBisTris2-[bis(2-hydroxyethyl)amino]-2-(hydroxymethyl)propane-1,3-diolPVDFpolyvinylidene difluoride (18Sadler J. J. Biol. Chem. 1991; 266: 22777-22780Abstract Full Text PDF PubMed Google Scholar). von Willebrand factor porcine submaxillary mucin endoplasmic reticulum green fluorescent protein monoclonal antibody alkaline phosphatase phosphate-buffered saline phenylmethylsulfonyl fluoride N-ethylmaleimide heavy chain 3 sialomucin complex 2-[bis(2-hydroxyethyl)amino]-2-(hydroxymethyl)propane-1,3-diol polyvinylidene difluoride The most thoroughly studied mucin is the porcine submaxillary mucin (PSM) (19Eckhardt A.E. Timpte C.S. DeLuca A.W. Hill R.L. J. Biol. Chem. 1997; 272: 33204-33210Abstract Full Text Full Text PDF PubMed Scopus (48) Google Scholar). PSM has been shown to dimerize via its C-terminal domain in the endoplasmic reticulum (ER) (20Perez-Vilar J. Eckhardt A.E. Hill R.L. J. Biol. Chem. 1996; 271: 9845-9850Abstract Full Text Full Text PDF PubMed Scopus (65) Google Scholar, 21Perez-Vilar J. Hill R.L. J. Biol. Chem. 1998; 273: 6982-6988Abstract Full Text Full Text PDF PubMed Scopus (39) Google Scholar) and oligomerize through its N-terminal domain in the acidic compartments of the Golgi complex, presumably making up a branched structure (22Perez-Vilar J. Eckhardt A.E. DeLuca A. Hill R.L. J. Biol. Chem. 1998; 273: 14442-14449Abstract Full Text Full Text PDF PubMed Scopus (65) Google Scholar, 23Perez-Vilar J. Hill R.L. J. Biol. Chem. 1998; 273: 34527-34534Abstract Full Text Full Text PDF PubMed Scopus (37) Google Scholar). In analogy with the results from the studies of PSM, we have reported previously (24Asker N. Baeckstrom D. Axelsson M.A. Carlstedt I. Hansson G.C. Biochem. J. 1995; 308: 873-880Crossref PubMed Scopus (64) Google Scholar, 25Asker N. Axelsson M.A. Olofsson S.O. Hansson G.C. J. Biol. Chem. 1998; 273: 18857-18863Abstract Full Text Full Text PDF PubMed Scopus (109) Google Scholar) that the human MUC2 mucin forms disulfide-stabilized dimers in the ER. These studies were made on the full-length MUC2, but due to the large size and high carbohydrate content of the mucins, it was hard to do further studies on its assembly by conventional methods. We have therefore expressed the last 981 amino acids of the human MUC2 mucin (corresponding to its C-terminal Cys-rich domain) as a fusion protein with the mycTag epitope and the green fluorescent protein (GFP) in Chinese hamster ovary cells (CHO-K1) and in the human colon adenocarcinoma cell line LS 174T. With this tool we could verify that the domain was capable of forming dimers and that these dimers were secreted from the cells. 2M. E. Lidell, M. E. V. Johansson, M. Mörgelin, N. Asker, C. J. R. Gum, Jr., Y. S. Kim, and G. C. Hansson, unpublished observations. These results were also supported by previous studies (26Bell S.L. Khatri I.A. Xu G. Forstner J.F. Eur. J. Biochem. 1998; 253: 123-131Crossref PubMed Scopus (33) Google Scholar, 27Bell S.L. Xu G. Forstner J.F. Biochem. J. 2001; 357: 203-209Crossref PubMed Scopus (34) Google Scholar) of the MUC2 homologue in rat (Muc2). In this study, we have studied a cleavage in the Cys-rich C-terminal part of MUC2. The cleavage was first observed during attempts to purify and study the glycosylation pattern of the recombinant C-terminal Cys-rich domain expressed in CHO-K1 cells. We found that when buffers of pH 6.0 were used during purification or glycosidase treatments, part of the material was cleaved. The cleavage caught our attention and was further analyzed, revealing that it was formed by an autocatalytic mechanism triggered by low pH. Furthermore, the cleavage produced a new, reactive C terminus that has the potential to link the cleaved mucin to other compounds. The polyclonal antiserum α-MUC2C2 has been described before (28Axelsson M.A.B. Asker N. Hansson G.C. J. Biol. Chem. 1998; 273: 18864-18870Abstract Full Text Full Text PDF PubMed Scopus (81) Google Scholar). The α-mycTag monoclonal antibody (mAb) was from spent culture media of the 1-9E10.2 hybridoma (ATCC CRL-1729). Other antibodies used were α-GFP mAb (Clontech), goat anti-mouse immunoglobulins coupled to alkaline phosphatase (AP) (Dako), goat anti-rabbit immunoglobulins coupled to alkaline phosphatase (Dako), and streptavidin coupled to alkaline phosphatase (streptavidin-AP) (Dako). An expression plasmid with the bases 12,622–15,708 of the human MUC2 sequence (12Gum Jr., J. Hicks J. Toribara N. Siddiki B. Kim Y. J. Biol. Chem. 1994; 269: 2440-2446Abstract Full Text PDF PubMed Google Scholar) was generated from the plasmids SMUC41 (bases 12,243–13,083) and V5 (bases 12,880–15,708) (29Gum Jr., J.R. Am. J. Respir. Cell Mol. Biol. 1992; 7: 557-564Crossref PubMed Scopus (216) Google Scholar). These were cloned into the eukaryotic expression vector pEGFP-C1 (Clontech) where the signal sequence had been replaced with the murine immunoglobulin κ-chain signal sequence from pSecTag A (Invitrogen) followed by the mycTag (EQKLISEEDL).2 The resulting pSMG-MUC2C plasmid encoding the κ-chain signal sequence, the mycTag, GFP, and the MUC2 C terminus was sequenced and transfected into the human colon adenocarcinoma cell line LS 174T (ATCC CL-188) and Chinese hamster ovary cells CHO-K1 (ATCC CCL-61). Both cell lines were selected for permanent expression of the recombinant protein and named LS 174T-pSMG-MUC2C and CHO-K1-pSMG-MUC2C,2 respectively. These were cultured as described earlier (24Asker N. Baeckstrom D. Axelsson M.A. Carlstedt I. Hansson G.C. Biochem. J. 1995; 308: 873-880Crossref PubMed Scopus (64) Google Scholar) for LS 174T with the addition of 125 μg ml−1 G418 for LS 174T-pSMG-MUC2C and 250 μg ml−1 G418 for CHO-K1-pSMG-MUC2C, respectively. After growing CHO-K1-pSMG-MUC2C cells until 3–4 days after confluence, the cells were washed twice with phosphate-buffered saline (PBS) and Iscove's modified Dulbecco's medium (Invitrogen) was added. The spent culture media were harvested, and new media were added every 3rd day. After the addition of protease inhibitors (1 mmphenylmethylsulfonyl fluoride (PMSF) (Sigma), 100 μmleupeptin (Sigma), 100 KIE ml−1 aprotinin (Trasylol, Bayer), and 5 mmN-ethylmaleimide (NEM, Sigma)), the media was centrifuged 1,000 × g for 10 min (+4 °C), and 0.02% (w/v) NaN3 was added. 200 ml of media were filtrated in an Amicon Stirred Cell model 8400 through an Amicon Ultrafiltration Membrane YM-100 (Millipore). The volume was reduced to 20 ml, filled up to 200 ml with 50 mm Tris-HCl, pH 8.0, and reduced to 20 ml again. This procedure was repeated 3 times. The concentrate was filtered (0.22 μm Millex-GV, Millipore) and loaded on a Mono Q HR 5/5 column (Amersham Biosciences) using a flow rate of 1.0 ml min−1. After washing with 150 mm NaCl in 50 mm Tris-HCl, pH 8.0, bound components were eluted with a linear NaCl gradient (150–600 mm) in 50 mm Tris-HCl, pH 8.0, over 45 min. The eluting components were monitored at 280 nm and collected in 1-ml fractions. The fractions were individually analyzed by SDS-PAGE and silver staining. Seven column fractions, eluting between 230 and 300 mm NaCl, containing the recombinant MUC2 C terminus were pooled, concentrated to 200 μl by ultrafiltration (Vivaspin 2-ml concentrator, 30,000 molecular weight cut-off RC, 3000 ×g, 4 °C), loaded directly on a Superose 6 HR 10/30 column (Amersham Biosciences) eluted with 150 mm ammonium acetate, pH 8.5, at a flow of 250 μl min−1. The eluting components were monitored at 280 nm and collected in 0.5-ml fractions. All the fractions contained in the symmetrical unimodal peak eluting atKav = 0.095 and positive for the recombinant MUC2 C terminus, as assessed by SDS-PAGE and silver staining, were pooled and used for the subsequent studies. Confluent cells were metabolically labeled by preincubation in Met- and Cys-free minimal essential medium (Invitrogen) containing 10% (v/v) fetal bovine serum, 100 IU ml−1 penicillin, and 100 μg ml−1streptomycin for 1–2 h followed by addition of labeling mix (Redivue Pro-mix l-35S in vitro labeling mix,Amersham Biosciences) to a final concentration of 36–107 μCi ml−1. After incubation for 6–16 h, cell lysates and media were prepared. When pulse-chase studies were performed, the labeling mix was removed after 5 or 15 min and the cells washed once and chased with normal cell culturing media with excess of Met and Cys (15 and 25 μg extra/ml, respectively). Media were removed and the cells washed twice with PBS. The cells were lysed in lysis buffer (50 mm Tris-HCl, pH 7.9, 150 mm NaCl, 1% (v/v) Triton X-100) containing protease inhibitors and NEM (concentrations as above) and sonicated (intensity 15) 3 times for 2 s (MSE Soniprep 100 sonifier), and the cell debris was removed by centrifugation (16,000 × g for 10 min, 4 °C). The media were supplemented with protease inhibitors and NEM (concentrations as above) and centrifuged (1000 × gfor 10 min, 4 °C). Immunoprecipitations using the mAbs α-mycTag or α-GFP antibodies were performed by precoating these on goat anti-mouse IgG-coupled magnetic beads (Dynabeads, Dynal). The immunocomplexes were washed 4 times in 10 mm Tris-HCl, 2 mm EDTA, 0.1% (v/v) Triton X-100, 0.1% SDS (w/v), pH 7.4, and 1 time in lysis buffer. Immunoprecipitated material was released from the magnetic beads in sample buffer (25 mm Tris-HCl, 10% (v/v) glycerol, 2.5% (w/v) SDS, pH 6.8) with or without 5% (v/v) β-mercaptoethanol or 100 mm dithiothreitol for 5 min at 95 °C. Samples were analyzed by discontinuous SDS-PAGE using 3–10 or 3–15% gradient gels with 3% stacking gels or 5% linear gels with 3% stacking gels (30Laemmli U.K. Nature. 1970; 227: 680-685Crossref PubMed Scopus (207516) Google Scholar). Coomassie Blue staining was performed according to the manufacturer (BioSafe Coomassie G-250, Bio-Rad). After analysis of metabolically labeled samples, the gels were fixed for 30 min in 30% methanol, 7% acetic acid, incubated in Amplify (Amersham Biosciences) for 15 min, dried, and exposed to BioMax MS films (Eastman Kodak Co.). The molecular markers used were the High-Range Rainbow 14C-Methylated Protein Molecular Weight Markers (Amersham Biosciences) and Precision Protein Standards (Bio-Rad). Transfer of the proteins to PVDF membranes (Immobilon-P, 0.45 μm, Millipore) was done in a Transfer-Blot SD-Dry Transfer Cell (Bio-Rad) at 2.5 mA/cm2. The transfer buffer used contained 48 mm Tris, 39 mm glycine, 1.3 mm SDS, and 10% (v/v) methanol. After blotting, the membranes were placed in blocking solution (PBS containing 5% (w/v) milk powder, 0.1% (v/v) Tween 20, and 0.05% (w/v) NaN3) overnight at 4 °C and incubated with either streptavidin-AP (diluted 1:1000) or the primary antibody α-mycTag mAb (diluted 1:10) or α-MUC2C2 (diluted 1:100) for 1.5 h at room temperature. The membranes were washed 3 times for 5 min with PBS-T (PBS containing 0.1% (v/v) Tween 20) and incubated with secondary antibodies (either goat anti-mouse immunoglobulins coupled to AP or goat anti-rabbit immunoglobulins coupled to AP 1:1000 in blocking solution) for 1.5 h at room temperature. After another PBS-T wash (3 times for 5 min) the membranes were developed with nitro blue tetrazolium/5-bromo-4-chloro-3-indolyl phosphate. In the cases where biotin ethylenediamine hydrobromide were used (see below), another blocking solution was used (10 mm Tris-HCl, pH 7.5, containing 2% (w/v) bovine serum albumin, 100 mm NaCl, 0.1% (v/v) Tween 20, and 0.05% (w/v) NaN3). Nonlabeled material were immunoprecipitated with the α-mycTag mAb from CHO-K1-pSMG-MUC2C cell lysates and media, incubated in citric acid-Na2HPO4, pH 6.0, buffer (McIlvaine buffer) (2.5 h, 37 °C) while still attached to the beads, analyzed by SDS-PAGE, Western-blotted, and detected by immunostaining. The citric acid-Na2HPO4 buffer system used here, as well as below, is the one described by McIlvaine (31McIlvaine T.C. J. Biol. Chem. 1921; 49: 183-186Abstract Full Text PDF Google Scholar). Radiolabeled and immunoprecipitated (α-mycTag mAb) material from CHO-K1-pSMG-MUC2C cell lysates or media was treated as described below while still attached to the magnetic beads. The pH dependence of the cleavage reaction was studied by incubating the radiolabeled material from cell lysates or media in citric acid-Na2HPO4 buffers (McIlvaine buffers) with pH from 4.8 to 7.8 for 20 min at 37 °C. Time dependence of the cleavage reaction was studied by incubating aliquots of the lysate material in a citric acid-Na2HPO4, pH 6.0 buffer (McIlvaine buffer), at 37 °C for 0–300 min. The influence of different protease inhibitors on the cleavage reaction was studied by incubating material from cell lysates for 1 h at 37 °C in 50 mmBisTris, pH 6.2 (pH 5.9 at 37 °C), containing one of the following inhibitors: PMSF (1 mm), bestatin (10 mm)(Sigma), leupeptin (100 μm), pepstatin (1 μm) (Roche Molecular Biochemicals), aprotinin (2.15 μm), NEM (5 mm), EDTA (10 mm), or a mixture of protease inhibitors (Complete mini-EDTA-free, Roche Molecular Biochemicals). All incubations above were terminated by removal of the incubation buffers and addition of SDS-PAGE sample buffer before analysis by SDS-PAGE. Confluent LS 174T-pSMG-MUC2C cells were used in these studies. When NH4Cl was used, the cells were preincubated with media with or without 25 mm NH4Cl for 12 h before the 1-h starvation and 8-h labeling period (both performed with 25 mm NH4Cl present). When bafilomycin A1 (a kind gift from AstraZeneca, Mölndal, Sweden) was used, it was present at 300 mm during the starvation (1 h) and labeling (6 h). The cell lysates and media were immunoprecipitated using the α-GFP antibody and analyzed by SDS-PAGE. Aliquots (114 ng) of purified recombinant MUC2 C terminus secreted from CHO-K1 cells were incubated for 1.5 h at room temperature in either a citric acid-Na2HPO4, pH 7.4 buffer, a citric acid-Na2HPO4, pH 5.4 buffer, or in the latter containing 1 μmol of biotin ethylenediamine hydrobromide (Sigma). The incubations above were terminated by addition of SDS-PAGE sample buffer before analysis by SDS-PAGE and Western blotting. Purified C terminus was dried (using a SpeedVac concentrator), dissolved in 50 mm pyridine/acetic acid, pH 5.5 buffer, and incubated overnight at 37 °C. The sample was dried, dissolved in reducing SDS-PAGE sample buffer, separated by SDS-PAGE, blotted to a PVDF membrane, and stained with Coomassie Blue (BioSafe Coomassie G-250, Bio-Rad) for 1 h. The membrane was destained in 50% (v/v) methanol, 10% (v/v) acetic acid, and the band corresponding to the C-terminal cleavage fragment was cut out and sequenced on a Procise 492 Protein Sequencer (Applied Biosystems). The cell line CHO-K1-pSMG-MUC2C permanently expressed a construct encoding the mycTag epitope at the N-terminal end followed by GFP and the last 981 amino acids of the human MUC2 mucin. The murine immunoglobulin κ-chain signal sequence was used to direct the fusion protein to the secretory pathway. The recombinant MUC2 C terminus (SMG-MUC2C) formed dimers in the ER (with a molecular mass of about 330 kDa) and was later secreted as dimers of higher molecular mass (about 470 kDa), a difference partly due to glycosylation.2 These cells were grown in the presence of 10% (v/v) fetal bovine serum, and the secreted fusion protein was purified by ultrafiltration, gel filtration, and ion exchange chromatography using a 50 mmBisTris, pH 6.0 buffer. The purified SMG-MUC2C was analyzed by SDS-PAGE under nonreducing conditions revealing a single band with an approximate molecular mass of 470 kDa, corresponding to the size of a dimer (Fig. 1). When the same material was analyzed under reducing conditions, three bands were observed (Fig.1). The largest band corresponded to the expected monomer with an approximated molecular mass of about 250 kDa. The two additional bands migrated at about 130 and 110 kDa, suggesting that these could be due to a cleavage of the full-length SMG-MUC2C. Western blot analysis of the nonreduced band revealed that this was immunoreactive with both the α-mycTag mAb and the α-MUC2C2 antiserum (Fig.2A). This was also true for the 250-kDa monomer found after reduction. The 130-kDa band only reacted with the mycTag mAb, whereas the 110-kDa band reacted with the α-MUC2C2 antiserum raised against a peptide sequence in the C-terminal end of MUC2. The results indicated that some of the SMG-MUC2C had been cleaved, producing a 130-kDa N-terminal and a 110-kDa C-terminal fragment, and that these fragments were still held together by disulfide bonds.Figure 2Identification of the cleavage products of the MUC2 C terminus. A, MUC2 C terminus, secreted from CHO-K1-pSMG-MUC2C cells and purified using a pH 6.0 buffer, was separated on a 3–10% SDS-PAGE gel, blotted, and detected with either α-MUC2C2 antiserum or α-mycTag mAb. B, cell lysates and media from CHO-K1-pSMG-MUC2C cells were immunoprecipitated (using the α-mycTag mAb), incubated for 2.5 h in a citric acid-Na2HPO4, pH 6.0 buffer, separated by SDS-PAGE (3–10%, reducing conditions), blotted, and detected by either the α-MUC2C2 antiserum or the α-mycTag mAb. L, material from cell lysates; M, material from the cell culturing media; Di, position of the dimer of the MUC2 C terminus; Mono, position of the monomer of the MUC2 C terminus; MG-C1, the N-terminal cleaved fragment;C2, the C-terminal cleaved fragment; NonRed, nonreduced samples; Red, reduced samples. IBindicates the antibody used for detection. Positions of molecular mass standards are indicated to the left.View Large Image Figure ViewerDownload Hi-res image Download (PPT) When the glycosylation of the recombinant MUC2 C terminus was studied, using different exoglycosidases, two buffers with different pH values were used (pH 6.0 and 7.0, respectively). When the samples were treated in the pH 7.0 buffer, only the intact protein band was observed. However, when the samples were treated in the pH 6.0 buffer, the two cleavage products appeared (data not shown). These observations of a cleavage occurring only when the sample had been treated at pH 6.0 led to the hypothesis that this reaction was induced at low pH. Purification of secreted MUC2 C terminus from CHO-K1-pSMG-MUC2C cells grown under serum-free conditions, using ultrafiltration, ion exchange chromatography, and gel filtration with buffers having a pH ≥ 8.0, gave rise to a much lower extent of cleavage (Fig. 1), further supporting this hypothesis. Purified SMG-MUC2C, from spent culture media, was incubated at pH 5.5 overnight. After separation by SDS-PAGE and transfer to a PVDF membrane, the 110-kDa band corresponding to the C-terminal cleavage fragment was excised and subjected to N-terminal amino acid sequencing. The sequence obtained, PHYVTF, is located at a position compatible with the sizes of the two cleaved fragments (see Fig. 1). The localization of this cleavage corresponds to the N-terminal end of the previously found 118-kDa "link peptide" (32Fahim R.E. Forstner G.G. Forstner J.F. Biochem. J. 1983; 209: 117-124Crossref PubMed Scopus (23) Google Scholar). To decipher the identity of the cleavage products produced in the in vitrostudies, immunoprecipitated material from CHO-K1-pSMG-MUC2C cell lysates and media was incubated at pH 6.0 and further analyzed by SDS-PAGE, Western blotting, and immunodetection (Fig. 2B). The results showed that the N-terminal cleavage product migrated as ∼50 kDa heavier after its secretion (130 relative to 80 kDa,left panel), whereas the C-terminal fragment only acquired around 10 kDa after its secretion (110 relative to 100 kDa, right panel). This caused the two bands to switch position relative to the intact SMG-C monomer, when comparing cell lysate and media. This implies that the modifications, leading to a secreted SMG-MUC2C with lower mobility compared with the intracellular form, mainly are present in the N-terminal part of the protein. To test the hypothesis of a pH-dependent cleavage, radiolabeled material from CHO-K1-pSMG-MUC2C cell lysates and media was immunoprecipitated using α-mycTag antibodies coupled to magnetic beads. While still attached to the beads, aliquots of the material were treated under different conditions, and the labeled SMG-MUC2C was analyzed by SDS-PAGE and autoradiography. To investigate if the cleavage was induced at a low pH, aliquots of immunoprecipitated SMG-MUC2C from cell lysate and media were treated in citric acid-Na2HPO4 buffers with pH from 4.8 to 7.8 for 20 min at 37 °C (Fig. 3,A and B). The cleavage was observed at a pH around 6.0 and lower for both the intracellular and extracellular forms of SMG-MUC2C. The cleavage products could be observed as weak bands also at pH >6.0 in the secreted but not the intracellular form of SMG-MUC2C. This indicated that the cleavage could occur prior to the incubation and probably inside the cell. The intracellular SMG-MUC2C has been localized to the ER,2 and thus this form has not passed through the more acidic parts of the late secretory pathway. The presence of some cleavage of the secreted SMG-MUC2C suggested that this process occurred late in the secretory pathway. To investigate if the cleavage reaction was time-dependent, aliquots of immunoprecipitated SMG-MUC2C from cell lysate were incubated in a pH 6.0 buffer for different times (0–300 min) (Fig.3C). The result from this experiment showed that cleavage was time-dependent as an increased amount of the cleavage products was observed with longer incubation times. However, cleavage was not complete even after 5 h. To explore if the cleavage was mediated by a protease, aliquots of immunoprecipitated SMG-MUC2C from cell lysate were treated in a BisTris buffer, pH 6.2 (gives pH 5.9 at 37 °C), with the addition of protease inhibitors (Fig. 4). A wide range of protease inhibitors including Ser, Cys, aspartic, and metalloprotease inhibitors as well as an aminopeptidase inhibitor were used. None of these inhibited the cleavage reaction to a significant extent. This argues for the cleavage not being mediated by an enzyme tightly attached to or being an integral part of the MUC2 C terminus itself. The level of cleavage in the secreted material from the CHO-K1-pSMG-MUC2C cells was ver