Post-translational Fate of a Mucin-like Leukocyte Sialoglycoprotein (CD43) Aberrantly Expressed in a Colon Carcinoma Cell Line*

(Received for publication, November 13, 1996, and in revised form, February 18, 1997)

Dan Baeckström Dagger

From the Department of Medical Biochemistry, University of Göteborg, Medicinaregatan 9, S-413 90 Göteborg, Sweden

ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
FOOTNOTES
ACKNOWLEDGEMENTS
REFERENCES


ABSTRACT

This paper describes the biosynthesis of L-CanAg, a mucin-like glycoprotein which carries the carcinoma-associated carbohydrate epitope sialyl-Lewis a and is secreted by the colon adenocarcinoma cell line COLO 205. Recently, it has been shown that L-CanAg is a novel glycoform of CD43, a surface sialoglycoprotein normally found only on hematopoietic cells. Immunoprecipitation with alpha -GPEP18, a novel antiserum against the cytoplasmic domain of CD43, detected a transmembrane form of L-CanAg carrying sialyl-Lewis a. Cell surface biotinylation experiments demonstrated the presence of transmembrane L-CanAg at the plasma membrane and that COLO 205, unlike the leukocyte cell line HL-60, contained significant amounts of glycosylated intracellular CD43.

Immunoprecipitation of phosphate-labeled COLO 205 cells revealed that membrane-bound L-CanAg, like leukocyte CD43, is a phosphoprotein. Interestingly, both surface- and phosphate-labeled L-CanAg were eluted earlier from a gel filtration column than their unlabeled counterparts, indicating that this method could separate membrane-bound L-CanAg from its soluble form. Immunoprecipitations of pulse-chase-labeled COLO 205 lysates fractionated by gel filtration showed that decrease in membrane-bound L-CanAg was concomitant with an increase in the intracellular soluble form. Together, these data indicate that transmembrane L-CanAg is fully glycosylated and phosphorylated before the extracellular domain is cleaved off and stored inside the cell before exocytosis.


INTRODUCTION

Monoclonal antibodies (mAbs)1 detecting carcinoma-associated antigens have often been shown to bind epitopes on mucins. The common denominator of these molecules is the mucin domain, a more or less repetitive sequence rich in proline and serine and/or threonine, the two latter amino acids being extensively O-glycosylated (1, 2). The high abundance of proline and carbohydrate chains makes the mucin domain a very extended, hydrophilic and protease-resistant structure. Mucins are thought to protect mucosae from chemical, mechanical, and microbial damage by restricting the mobility of water in the vicinity of the apical membrane. Glycoproteins containing mucin domains have also been found in non-epithelial cells such as endothelial cells and leukocytes, where they are involved in both adhesive and anti-adhesive events (3-5).

Tumor-associated epitopes on mucins are generated by abnormal glycosylation. One of these epitopes is the carbohydrate structure sialyl-Lewis a (Si-Lea, NeuAcalpha 2right-arrow3Galbeta 1right-arrow3[Fucalpha 1right-arrow4]GlcNAcbeta 1right-arrow; Refs. 6 and 7). Using the colon adenocarcinoma cell line COLO 205 as a model system, previous studies have identified two mucin-like glycoproteins called H-CanAg and L-CanAg, which are the major Si-Lea-carrying glycoproteins in these cells (8). H-CanAg, which has an apparent molecular mass of 600-800 kDa, was shown to be a MUC1 mucin, i.e. the apoprotein part of the molecule is the product of the MUC1 gene. MUC1 mucins are transmembrane glycoproteins common to apical membranes of secretory epithelia, and overexpression and/or abnormal glycoforms of MUC1 are often found in carcinomas (9). L-CanAg (apparent molecular mass = 150-300 kDa), which is secreted from COLO 205 cells at a significantly higher rate than H-CanAg, was recently shown to have an apoprotein that is identical to that of CD43 (leukosialin) (10), normally a major membrane sialoglycoprotein of leukocytes and platelets (11, 12). That finding strengthens existing hypotheses (9) that carcinoma cells may gain a selective advantage, such as immune evasion or facilitated dissemination, by the expression of mucin-like glycoproteins, either by overexpressing normally present mucins such as MUC1 or by the anomalous expression of foreign glycoproteins such as CD43. Since the identification of L-CanAg as a CD43 glycoform, expression of CD43 in other non-hematopoietic tumors has also been reported (13).2

When addressing the question of the possible pathological significance of these glycoproteins, knowledge of their metabolism is important. In the present study, I have investigated events in the post-translational modifications of L-CanAg, which are shown to follow a sequence of glycosylation, phosphorylation, and creation of a soluble mucin-like proteolytic fragment which is retained inside the cell and then released into the medium.


MATERIALS AND METHODS

Cells, Peptides, and Antibodies

The colon adenocarcinoma cell line COLO 205 and the promyelocytic cell line HL-60 were cultured as described previously (10). The Si-Lea-reactive mAbs C241 (IgG1) and C50 IgM were gifts generously provided by CanAg Diagnostics AB, Göteborg, Sweden. The mAb Leu-22, which recognizes leukocyte glycoforms of CD43, was purchased from Becton Dickinson, San José, CA. The peptide GPEP18 (CGRRKSRQGSLAMEELKS), corresponding to amino acids 327-344 in the CD43 peptide sequence (except for the N-terminal cysteine) was synthesized and coupled to keyhole limpet hemocyanin by Scandinavian Peptide Synthesis, Köping, Sweden. The conjugated peptide was used to generate the antiserum alpha -GPEP18 by immunizing a New Zealand White rabbit as described earlier (14).

Biosynthetic Labeling of Cells

COLO 205 cells grown to approximately 75% confluence in 35-mm Petri dishes or 5 × 106 HL-60 cells were used in each experiment. For methionine incorporation experiments, cells were starved for 1 h in minimal essential medium without methionine (Life Technologies, Paisley, Scotland) containing 5% fetal calf serum and 2 mM glutamine before labeling with Pro-Mix (Amersham, United Kingdom) corresponding to 150 µCi of [35S]methionine per dish. In pulse-chase experiments, chase was initiated by changing the medium to regular culture medium (see above). For phosphate labeling experiments, cells were starved for 1 h in Dulbecco's modified Eagle's medium without phosphate (Life Technologies) containing 5% fetal calf serum and 25 mM HEPES, pH 7.4, before labeling with 200 µCi of [32P]orthophosphate (Amersham) per dish. Cells were washed twice with cold phosphate-buffered saline before being lysed in 1 ml of the lysis buffer described in Ref. 10. In phosphate labeling experiments, the lysis buffer was supplemented with 10 mM sodium fluoride.

Cell-surface Biotinylation

Before surface biotinylation, COLO 205 cells were detached from the substratum using 0.02% EDTA in phosphate-buffered saline. Biotinylation with sulfosuccinimidoyl 6-(biotinamido)hexanoate (NHS-LC-biotin, Pierce, Oud Beijerland, Netherlands) was performed as described by Litvinov and Hilkens (15). Biotinylation of cell-surface glycoconjugates using sodium metaperiodate oxidation followed by treatment with biotin hydrazide (Sigma) was performed according to Orr (16) with the following modifications: the concentration of sodium metaperiodate was 5 mM, biotin hydrazide rather than 2-iminobiotin hydrazide was used at a concentration of 5 mM, and the biotinylation was performed overnight on ice. Biotinylated cells were lysed as described (10).

Gel Filtration

Gel filtration was performed in a Superose 6 HR 10/30 column (Pharmacia Biotech, Uppsala, Sweden) using 0.1 M ammonium acetate or 50 mM Tris-HCl, pH 7.5, 4 M guanidinium chloride as eluant at a flow rate of 0.25 ml/min.

Immunoprecipitation

Immunoprecipitations were performed as described (10). Eluates from gel filtration were immunoprecipitated afer addition of 0.25 volumes of a 5 × lysis buffer concentrate (lysis buffer, see Ref. 10). Eluted immunoprecipitates were subjected to SDS-PAGE with 2.75-15% gradient separation gels and 2.75% stacking gels. Radiolabeled immunoprecipitates were visualized by autoradiography of dried gels after fixation in 30% ethanol, 10% acetic acid for >= 3 h followed by 30 min treatment with Amplify solution (Amersham) containing 5% glycerol.

Western Blot

Biotinylated and unlabeled samples were transferred from SDS-PAGE to nitrocellulose (Sartorius, Göttingen, Germany) using a semi-dry Sartoblot cell (Sartorius) and a transfer buffer consisting of 48 mM Tris, 39 mM glycin, 1.3 mM SDS, and 20% (v/v) methanol. Transfer was performed at 2-4 mA/cm2 of membrane for 1-2 h. Membranes were subsequently rinsed with water and blocked with 50 mM Tris-HCl, 0.15 M NaCl, pH 7.5 (TBS), containing 3% bovine serum albumin overnight at room temperature. The membrane was incubated with primary antibody diluted to 1-2 µg/ml in TBS containing 0.05% Tween 20 (TBST) for 4 h. After 3 washes with TBST, the membrane was incubated for 3 h with alkaline phosphatase-conjugated rabbit anti-mouse antiserum (DAKO immunoglobulins, Glostrup, Denmark) diluted 1/2000 in TBST, or with alkaline phosphatase-conjugated streptavidin diluted in TBST containing 0.5 M NaCl. After 3 washes with TBST and one wash with TBS, the membrane was developed using 0.3 mg/ml nitro blue tetrazolium and 0.15 mg/ml 5-bromo-4-chloro-3-indolyl phosphate in 0.1 M Tris-HCl, pH 9.5, 0.1 M NaCl, 5 mM MgCl2.


RESULTS

Generation and Reactivity of an Antiserum against a Cytoplasmic Region of CD43

Rabbits were immunized with a keyhole limpet hemocyanin-conjugate of GPEP18, a 17-mer peptide identical in sequence to amino acids 327-344 in the cytoplasmic domain of CD43. The resulting antiserum, called alpha -GPEP18, was tested for reactivity in immunoprecipitation experiments with lysates of pulse-chase labeled cells. In the promyelocytic cell line HL-60, known to express one of the major leukocyte-associated glycoforms of CD43 (17), alpha -GPEP18 immunoprecipitated one band of apparent molecular mass of 55 kDa immediately after the pulse labeling (Fig. 1). After 20 and 60 min of chase, another band of 120 kDa gained in intensity concomitant with the fading of the 55-kDa band. The apparent molecular sizes and pulse-chase kinetics of the 55-kDa and 120-kDa bands were in agreement with those reported for the CD43 primary precursor and fully glycosylated forms, respectively (17). In the colon adenocarcinoma cell line COLO 205, alpha -GPEP18 immunoprecipitated a precursor band with the same electrophoretic mobility as in HL-60 cells. However, in contrast to the sharp 120-kDa band seen in HL-60 cells, the glycosylated CD43 in COLO 205 appeared as a broad and diffuse smear straddling the 200-kDa marker position. This electrophoretic mobility is typical for that of L-CanAg, the Si-Lea-carrying mucin-like molecule which has been biochemically purified from COLO 205 cells (8) and subsequently shown to be a novel CD43 glycoform (10). It could therefore be concluded that alpha -GPEP18 can be used as a reagent for glycosylation-independent detection of transmembrane CD43.


Fig. 1. Immunoprecipitation of pulse-chase labeled cells using the alpha -GPEP18 antiserum directed against a peptide from the cytoplasmic C-terminal of CD43. HL-60 cells (A) and COLO 205 cells (B) were pulse-labeled with 200 µCi of [35S]methionine for 10 min and chased by addition of excess unlabeled methionine for the time periods indicated. Lysates of the cells were precipitated twice with nonimmune rabbit serum (the second precipitate for each sample is shown in the lanes marked NRS) before immunoprecipitating with the alpha -GPEP18 antiserum (lanes marked Ab). Eluted immunoprecipitates were run on a 2.75-15% T SDS-PAGE gel under nonreducing conditions. The positions of the CD43 precursor (P) and the respective glycosylated forms (G) are indicated by the arrows. The sizes in kDa of molecular weight marker proteins are indicated.
[View Larger Version of this Image (39K GIF file)]


Transmembrane CD43 in COLO 205 Cells Is Glycosylated with Si-Lea

CD43 glycosylated with Si-Lea can be recovered from secretions of COLO 205 cells in the form of the L-CanAg glycoprotein (10). As the secretion of L-CanAg most likely entails the proteolytic detachment of the CD43 extracellular domain (or a part of it), it was of interest to determine whether L-CanAg is still a transmembrane CD43 molecule at the late biosynthetic stage when formation of the Si-Lea epitope occurs. Lysates and media of COLO 205 cells were therefore immunoprecipitated either with the alpha -GPEP18 antiserum or the mAb C241 directed against the Si-Lea epitope. The immunoprecipitates were then subjected to SDS-PAGE and Western blot followed by probing with the mAb C50 which also recognizes Si-Lea. As can be seen in Fig. 2, alpha -GPEP18 precipitated a Si-Lea-containing band in the L-CanAg size range from COLO 205 cell lysate but not from spent medium. In contrast, L-CanAg and also the Si-Lea-carrying MUC1 mucin H-CanAg were found in COLO 205 cell lysates when immunoprecipitated with the anti-Si-Lea mAb C241. Spent medium immunoprecipitated with mAb C241 contained both L-CanAg and traces of H-CanAg (barely visible in Fig. 2); materials of lower apparent Mr than is usually seen for L-CanAg were also detected. These results show that CD43 in COLO 205 cells acquires the Si-Lea epitope before the putative cleavage event that generates secreted L-CanAg. The results also confirm the absence of the CD43 cytoplasmic domain in secreted L-CanAg.


Fig. 2. Western blot of immunoprecipitates from COLO 205 cell lysates and spent media. Lysates and media were immunoprecipitated with nonimmune rabbit serum (NRS), the alpha -GPEP18 antiserum, or the mAb C241 directed against Si-Lea as indicated in the figure. Immunoprecipitates were separated on 2.75-15% SDS-PAGE under reducing conditions, transferred to nitrocellulose membranes, and probed with the Si-Lea-reactive mAb C50 coupled to biotin. Bound C50 mAb was detected using alkaline phosphatase-conjugated streptavidin. The positions of H-CanAg (H) and L-CanAg (L) are indicated by the arrows. The sizes in kDa of molecular weight marker proteins are indicated.
[View Larger Version of this Image (52K GIF file)]


A Fraction of the Membrane-bound CD43 in COLO 205 Cells Is Present at the Cell Surface

To investigate whether membrane-bound L-CanAg only exists as an intracellular intermediate or if it is also exposed at the plasma membrane, surface-bound carbohydrates on COLO 205 cells were labeled by mild periodate oxidation followed by derivatization of generated aldehyde groups with biotin hydrazide. Lysates of the labeled cells were immunoprecipitated with the alpha -GPEP18 antiserum or with mAb C241. After SDS-PAGE and Western blot, biotinylated immunoprecipitates were detected with alkaline phosphatase-labeled streptavidin. As shown in Fig. 3A, immunoprecipitation with alpha -GPEP18 yielded a readily distinguishable band in the L-CanAg size range, indicating that at least some of the membrane-bound CD43 in this cell line is present at the plasma membrane. In contrast, the Si-Lea-reactive mAb C241 could not immunoprecipitate any detectable amount of biotinylated L-CanAg. This latter finding suggested that the periodate oxidation-biotinylation procedure had rendered all the cell surface-exposed Si-Lea epitopes unrecognizable to the C241 mAb. Periodate oxidation of Si-Lea is known to reduce the binding of C241 (18), an effect which apparently was enhanced by the biotinylation of oxidized moieties.


Fig. 3. Immunoreactivity of surface-biotinylated and unlabeled COLO 205 and HL-60 cells. Surface-bound glycoconjugates of COLO 205 and HL-60 cells were labeled by periodate oxidation followed by conjugation with biotin hydrazide. Lysates of labeled and unlabeled cells were immunoprecipitated with the alpha -GPEP18 antiserum, mAb C241 against Si-Lea, or mAb Leu-22 against leukocyte CD43 as indicated in the figures (A, B). Samples in panel C were not immunoprecipitated. After separation on 2.75-15% SDS-PAGE under reducing conditions, samples were transferred to nitrocellulose and probed with alkaline phosphatase-conjugated streptavidin (panel A), biotinylated mAb C50 followed by alkaline phosphatase-conjugated streptavidin (panel B), or mAb Leu-22 followed by alkaline phosphatase-conjugated antiserum to mouse immunoglobulins (panel C). The sizes in kDa of molecular weight marker proteins are indicated.
[View Larger Version of this Image (27K GIF file)]


This phenomenon was exploited in an attempt to establish whether or not all the mature CD43 in COLO 205 cells is found on the cell surface. Lysates of biotinylated and non-biotinylated cells were immunoprecipitated with the C241 mAb and detected with C50, another mAb against Si-Lea, after SDS-PAGE and Western blot. As shown in Fig. 3B, L-CanAg (and also H-CanAg) are readily detected in untreated cells as well as in oxidized-biotinylated cells. Taken together, the data in Fig. 3, A and B, indicate that a significant proportion of the CD43 present in COLO 205 cells is inaccessible to inactivation of their Si-Lea epitopes by oxidation-biotinylation and therefore most likely represent a pool of intracellularly stored CD43.

Oxidation-biotinylation was also performed with HL-60 cells. As in COLO 205 cells, immunoprecipitation with alpha -GPEP18 yielded a biotinylated band of the size expected for glycosylated CD43 (Fig. 3A). When whole cell lysates of untreated and oxidized-biotinylated HL-60 cells were compared in a Western blot probed with the glycosylation-dependent anti-CD43 mAb Leu-22, only the untreated cells showed reactivity (Fig. 3C). As the Leu-22 epitope is known to be sensitive to periodate oxidation, the complete disappearance of Leu-22-reactive CD43 in oxidized-biotinylated HL-60 cells indicates that the majority of CD43 in this cell line is present on the plasma membrane, in contrast to the situation in COLO 205 cells.

CD43 Is Phosphorylated in COLO 205 Cells

As CD43 is known to be a phosphoprotein in leukocytes (19), it was of interest to determine whether phosphorylated CD43 could be detected also in COLO 205 cells. COLO 205 and HL-60 cells were therefore metabolically labeled with [32P]orthophosphate and the lysates were immunoprecipitated with the alpha -GPEP18 antiserum and, in the case of COLO 205, also with mAb C241. As shown in Fig. 4A, phosphorylated L-CanAg could be detected in immunoprecipitations both with alpha -GPEP18 and mAb C241. The intensity of the staining was comparable to that obtained in the control immunoprecipitation of HL-60 cells with alpha -GPEP18. As a parallel immunopreciptation of [35S]methionine-labeled cells with alpha -GPEP18 indicated that the CD43 synthesis rates were similar in COLO 205 and HL-60 (Fig. 4B), it can be assumed that the degree of CD43 phosphorylation in COLO 205 is roughly comparable to that seen in HL-60.


Fig. 4. Immunoprecipitation of lysates of COLO 205 and HL-60 cells metabolically labeled with [32P]orthophosphate (A, C, and D) or [35S]methionine (B). After 1 h in culture with medium deficient in phosphate (panels A, C, and D) or methionine (panel B), COLO 205 and HL-60 cells were incubated for 1 h with 200 µCi of [32P]orthophosphate (panels A, C, and D) or 200 µCi of [35S]methionine (panel B). Cell lysates that were to be immunoprecipitated with the alpha -GPEP18 antiserum were pre-precipitated twice with nonimmune rabbit serum and lysates that were to be immunoprecipitated with mAb C241 were pre-precipitated twice with immunoprecipitin without antibody; the second precipitate for each sample is shown in the lanes marked NRS and No Ab, respectively. Lysates were then immunoprecipitated with the alpha -GPEP18 antiserum, mAb C241, or mAb Leu-22 as indicated. In panels C and D, the lysates of phosphate-labeled COLO 205 (panel C) and HL-60 (panel D) cells were repeatedly immunoprecipitated with alpha -GPEP18 (lanes marked alpha -GPEP18 #1, #2, and #3) and excess antibody removed by an extra incubation with immunoprecipitin (lanes marked extra); finally, the depleted lysates were immunoprecipitated with mAb C241 (COLO 205, panel C) or mAb Leu-22 (HL-60, panel D). Immunoprecipitates were separated on 2.75-15% SDS-PAGE gels under nonreducing conditions. Sizes in kDa of molecular weight marker proteins are indicated.
[View Larger Version of this Image (58K GIF file)]


As the GPEP18 peptide contains the putative main phosphorylation site (Ser-332) of leukocyte CD43 (19), it could be suspected that phosphorylation at this site might influence the reactivity of the alpha -GPEP18 antiserum. To test this hypothesis, lysates of [32P]orthophosphate-labeled COLO 205 and HL-60 cells were repeatedly immunoprecipitated with alpha -GPEP18. After the depletion of alpha -GPEP18-reactive CD43, the remaining CD43 was precipitated using mAb C241 (for COLO 205 lysates) or mAb Leu-22 (for HL-60 lysates). Fig. 4, C and D, shows that both in COLO 205 and HL-60 cells, a significant amount of phosphorylated CD43 could be recovered from the lysates after depletion with alpha -GPEP18. These results show that all phosphorylated (and thus transmembrane) CD43 is not recognized by alpha -GPEP18.

Membrane-bound and Soluble L-CanAg Can Be Separated by Gel Filtration

To characterize the size distribution of cell surface-bound Si-Lea-carrying glycoproteins, COLO 205 cells were surface-labeled with NHS-LC-biotin, a reagent that couples biotin to amino groups of polypeptide chains on the cell surface and thus does not modify carbohydrate structures like Si-Lea. A lysate of the labeled cells was separated by gel filtration on a Superose 6 column using 4 M guanidinium chloride as eluant. The fractions were analyzed with immunofluorometric assays which used the Si-Lea-reactive mAb C50 as solid phase-bound catching antibody and either C50 or streptavidin as europium-labeled tracer. The chromatogram obtained with the homologous C50/Eu-C50 assay (Fig. 5) showed the characteristic biphasic distribution of Si-Lea-carrying glycoproteins observed previously in COLO 205 (8). The early eluting peak, called the E peak, is known to contain the MUC1 mucin H-CanAg, whereas the late eluting L peak has biochemical properties closely similar to those of the secreted CD43 glycoform L-CanAg isolated from spent COLO 205 medium (8). The curve obtained with the C50/Eu-streptavidin assay showed a markedly different shape (Fig. 5) with the main peak appearing at a position between the E and L peaks.


Fig. 5. Analysis of total and surface-biotinylated Si-Lea-carrying glycoproteins from a COLO 205 cell lysate by gel filtration chromatogaphy and fluoroimmunoassays. COLO 205 cells were surface-biotinylated using the reagent NHS-LC biotin. A lysate of the cells was fractionated by gel filtration chromatogaphy on a Superose 6 HR 10/30 column in 4 M guanidinium chloride at a flow rate of 0.25 ml/min. 0.25-ml fractions were collected. The fractions were analyzed by immunofluorometric assays using the Si-Lea-reactive mAb C50 as catching antibody and either C50 (C50/C50 assay, open circles) or streptavidin (C50/streptavidin assay, filled circles) as europium-labeled tracing reagents to detect the total and biotinylated Si-Lea-carrying glycoproteins, respectively.
[View Larger Version of this Image (23K GIF file)]


The low C50/Eu-streptavidin reactivity at the position of the L peak suggested that the L peak consists of predominantly intracellular (and thus non-biotinylated) material. This raised the question whether the main C50/Eu-streptavidin-reactive peak represented the position at which transmembrane L-CanAg elutes in gel filtration. To adress that question, fractions from Superose 6 gel filtration of a COLO 205 lysate were combined into pools E1, E2, and L as shown in Fig. 6A and subjected to immunoprecipitation with the alpha -GPEP18 antiserum. The immunoprecipitates were then probed on a Western blot with the anti-Si-Lea mAb C50. As shown in Fig. 6B, only the E1 and E2 pools showed detectable amounts of alpha -GPEP18-reactive L-CanAg, suggesting that the L pool does not contain transmembrane L-CanAg.


Fig. 6. Gel filtration fractionation and immunoprecipitation of lysates of cold and [32P]orthophosphate-labeled COLO 205 cells. A, unlabeled cells were lysed and subjected to gel filtration chromatography on a Superose 6 HR 10/30 column in 0.1 M ammonium acetate at a flow rate of 0.25 ml/min. 0.25-ml fractions were collected and analyzed for Si-Lea by an immunofluorometric assay using mAb C50 both as catching antibody and as europium-labeled tracing antibody. Fractions were combined into pools E1, E2, and L as indicated. B, the pools were immunoprecipitated with the alpha -GPEP18 antiserum; the immunoprecipitates were separated on 2.75-15% SDS-PAGE under reducing conditions, blotted to nitrocellulose, and detected by probing the membrane with biotinylated mAb C50 followed by alkaline phosphatase-conjugated streptavidin. C, immunopreciptation of a lysate of COLO 205 cells metabolically labeled for 1 h with [32P]orthophosphate. The lysate was affinity purified using the Si-Lea-reactive mAb C241 coupled to Sepharose and the eluate was fractionated by gel filtration chromatography into pools E1, E2, and L as described above in the legend to panel A. The pools were then subjected to immunoprecipitation using mAb C241 and the immunoprecipitates were run on a 2.75-15% SDS-PAGE gel under nonreducing conditions. Sizes in kDa of molecular weight marker proteins are indicated.
[View Larger Version of this Image (20K GIF file)]


To further test this conclusion, the distribution of phosphorylated (and thus transmembrane) L-CanAg in gel filtration was also analyzed. A lysate of [32P]orthophosphate-labeled COLO 205 cells was subjected to Superose 6 gel filtration and the E1, E2, and L pools were collected and immunoprecipitated with the Si-Lea-reactive mAb C241. As shown in Fig. 6C, the majority of the phosphorylated L-CanAg was found in the E2 pool with only weak staining in the E1 and L pools. This result strengthens the conclusion that the L peak represents an intracellular supply of soluble L-CanAg that can be separated from its membrane-bound counterpart by means of gel filtration.

L-CanAg in the E and L Pools Have Distinct Pulse-Chase Kinetics

The capacity of gel filtration to separate membrane-bound and soluble L-CanAg was exploited to analyze the sequence of biosynthetic events leading to L-CanAg secretion. Si-Lea-carrying glycoproteins from lysates of pulse-chase labeled COLO 205 cells were subjected to Superose 6 gel filtration and collected in pools E and L (cf. Fig. 6A; pool E = pool E1 + pool E2). These pools, as well as the spent media obtained from the pulse-chase experiment, were subjected to immunoprecipitation with mAb C241. As can be seen in Fig. 7, the appearance and disappearance of labeled L-CanAg in the E and L pools followed distinctly different time courses. In the E pool, the L-CanAg staining was most intense after 1 h with slight fading after 2 h and significant fading after 5 h. In contrast, the L-CanAg staining in the L pool increased steadily up to 5 h. After 24 h, most of the staining in both the E and L pools had disappeared, concomitant with a significant increase in L-CanAg staining in the spent medium (data not shown). As the immunoprecipitation was performed using a Si-Lea-reactive mAb, the kinetics of the Si-Lea-carrying MUC1 mucin H-CanAg appearing in the E pool could also be followed. In striking contrast to the kinetics of membrane-bound (i.e. pool E) L-CanAg, labeled H-CanAg was virtually absent after 1 h, maximal at 5 h, and still showing residual staining after 24 h.


Fig. 7. Immunoprecipitation of lysates of pulse-chase [35S]methionine-labeled COLO 205 cells after fractionation by gel filtration chromatography. COLO 205 cells were pulse-labeled with 150 µCi of [35S]methionine for 30 min and chased by culturing in complete medium without labeled methionine for the time periods indicated. Lysates of the cells were affinity purified using the Si-Lea-reactive mAb C241 coupled to Sepharose and the eluates were concentrated by ultrafiltration and subjected to gel filtration chromatography on a Superose 6 HR 10/30 column in 0.1 M ammonium acetate. The fractions were combined into pools E and L (which correspond to pools E1+E2 and L, respectively, in Fig. 6A) and pre-precipitated with immunoprecipitin (the pre-precipitates of the 0 h samples are shown in the lanes marked 0, pre) before immunoprecipitation with mAb C241. Immunopreciptates were separated on a 2.75-15% SDS-PAGE gel under nonreducing conditions. The positions of H-CanAg (H) and L-CanAg (L) are indicated by the arrows. The sizes in kDa of molecular weight marker proteins are indicated.
[View Larger Version of this Image (67K GIF file)]



DISCUSSION

In a recent study (10) it was demonstrated that the colon adenocarcinoma cell line COLO 205 expresses CD43, a surface molecule normally only seen on cells of the hematopoietic lineage. The CD43 glycoform expressed in COLO 205 cells, L-CanAg, was originally isolated as a secreted mucin-like molecule carrying Si-Lea, a marker for carcinomas of the intestinal tract (8). The glycans of L-CanAg are strikingly different from those of leukocyte CD43: whereas the latter generally contain only 4-6 sugars per chain (11), the former contain 17 on the average (8). It is therefore no surprise that antibodies that bind the extracellular domain of leukocyte CD43 fail to recognize L-CanAg (10).2 To facilitate detection of CD43 in a glycosylation-independent manner, an antiserum was raised against the synthetic peptide GPEP18 corresponding to a part of the cytoplasmic domain of CD43. This antiserum, alpha -GPEP18, was indeed capable of detecting both leukocyte-type CD43 from the promyelocytic cell line HL-60 and L-CanAg from COLO 205 cells (Fig. 1).

The data presented in Figs. 1 and 2 also show that alpha -GPEP18 can immunoprecipitate CD43 molecules from COLO 205 cells that have acquired a high (150-250 kDa) and polydisperse apparent molecular mass as well as the Si-Lea epitope, i.e. features typical of the mature L-CanAg mucin. The secretion of L-CanAg almost certainly entails the proteolytic detachment of the extracellular region from the transmembrane and intracellular domains, as confirmed by the inability of alpha -GPEP18 to immunoprecipitate any Si-Lea-containing materials from COLO 205 culture supernatant, and on the basis of the data from Figs. 1 and 2, it can be assumed that this cleavage occurs at a late biosynthetic stage when most or all of the glycosylation has taken place. The demonstration of co-expression of Si-Lea with CD43 peptide epitopes (Fig. 2) also provides additional evidence that L-CanAg is indeed a CD43 glycoform.

The post-translational modifications of L-CanAg also include phosphorylation, as demonstrated by the metabolic phosphate labeling experiments (Fig. 4). Leukocytes and cell lines derived from hematopoietic cells also phosphorylate CD43 to varying extents (19, 20). The degree of CD43 phosphorylation in COLO 205 cells seemed roughly comparable to that in HL-60, which in turn phosphorylates CD43 to relatively high extent compared with other leukocyte cell lines and to peripheral blood lymphocytes (19). The significance of CD43 phosphorylation is not known, but it may control proteolytic cleavage of the CD43 extracellular domain as both phosphorylation and proteolysis occur upon activation of leukocytes (21-24). The finding that L-CanAg is phosphorylated as well as proteolytically cleaved at a high rate in COLO 205 cells strengthens the hypothesis that there may be a causal relationship between phosphorylation and proteolysis of CD43. One might also speculate that the extensive phosphorylation and proteolysis of L-CanAg are the results of the activated state of these carcinoma cells, which could mimic the situation in activated leukocytes.

As was shown by the sequential immunoprecipitation experiments, alpha -GPEP18 failed to bind a significant subpopulation of phosphorylated CD43 in both COLO 205 and HL-60 cells. Phosphorylation itself is probably the cause of alpha -GPEP18's inability to bind this subpopulation, as treatment of HL-60 lysates with alkaline phosphatase has been observed3 to increase the amount of CD43 recoverable by alpha -GPEP18. The heterogeneity in alpha -GPEP18 reactivity toward phosphorylated CD43 suggests that several different phosphorylation sites may be used to a significant degree in COLO 205 and HL-60; alternatively, phosphorylation at a single site may cause the surrounding region to assume a number of alternative conformations, some of which are unrecognizable to the antiserum.

Although the cell-surface biotinylation experiments clearly demonstrated that membrane-bound L-CanAg is present at the plasma membrane (Fig. 3A), they also indicated that there is a significant amount of L-CanAg that is stored inside the COLO 205 cells, as shown by the presence of a pool of molecules that are protected from the Si-Lea inactivation caused by periodate oxidation and biotinylation (Fig. 3B). This was in contrast to the situation in HL-60 cells, where no detectable amount of the Leu-22 epitope remained after oxidation-biotinylation. This difference may be explained by the respective origins of COLO 205 and HL-60 cells: whereas HL-60 is of the hematopoietic cell lineage, COLO 205 is derived from the intestinal epithelium. One class of intestinal epithelial cells, the goblet cells, are known to be capable of storing mucins in intacellular granules from which they can be released by a process of regulated secretion (1). It is therefore possible that COLO 205 cells treat L-CanAg as if it were an intestinal mucin, not only with respect to glycosylation but also with respect to intracellular trafficking. The almost exclusive localization of CD43 to the surface of the hematopoietic cells has been observed before (17).

The finding that membrane-bound and soluble L-CanAg could be separated from each other by gel filtration (Figs. 5 and 6) is quite surprising, as removal of the intracellular and transmembrane domains of L-CanAg would decrease its size by only around 10%. One might speculate that membrane-bound L-CanAg forms noncovalently bound oligomers which are dissociated upon release of the extracellular domain. As the experiment described in Fig. 6, A and B, gives similar results when conducted in the presence of the denaturing solvent 4 M guanidinium chloride (data not shown), such hypothetical oligomers would have to be held together with considerable strength (although being sensitive to SDS).

By separating Si-Lea-carrying glycoproteins from lysates of pulse-chase labeled COLO 205 cells by gel filtration, the kinetics of membrane-bound and soluble L-CanAg could be studied (Fig. 7). From the finding that membrane-bound L-CanAg appeared and disappeared more rapidly than the soluble form, two conclusions may be drawn. First, it shows that the soluble L-CanAg detected in lysates in this and previous experiments is not an artifact created during lysate preparation (this is also indicated by the difference between membrane-bound and soluble L-CanAg in accessibility to surface biotinylation (Fig. 5)). Second, this result indicates that the relatively rapid synthesis of fully glycosylated, membrane-bound L-CanAg (1-2 h) is followed by a slower process involving the conversion to the soluble form, which is retained inside the cell until more than 5 h after synthesis. It is conceivable that membrane-bound L-CanAg cycles between the surface and the interior of the cell in a manner similar to what has been demonstrated earlier for the MUC1 mucin (15). However, experiments aimed at elucidating this question have been inconclusive (data not shown).

In summary, the results presented here suggest the following pathway for the modification and transport of the CD43 anomalously expressed in COLO 205 cells. The primary translation product, following a process of extensive glycosylation (including synthesis of the Si-Lea epitope), is transported to the cell surface where a proportion of the molecules are phosphorylated. Transmembrane L-CanAg is then converted to a soluble form by proteolysis. As a significant amount of soluble L-CanAg is found inside the cells, this proteolytic step presumably follows upon endocytosis of the cell surface-bound molecule. The soluble form of L-CanAg is then stored inside the cell for some time before being secreted. During this storage additional glycosylation may take place, as L-CanAg from COLO 205 cell extracts and from spent medium differ somewhat in carbohydrate composition (8). The present conclusions thus underscore the complexity of mucin expression in carcinomas and illustrate how a single mucin apoprotein species may assume a variety of different biochemical forms with a rich potential for distinct biological functions.


FOOTNOTES

*   This work was supported by grants from the Swedish Cancer Society, Swedish Medical Research Council Grant 7461, IngaBritt and Arne Lundbergs Stiftelse, and the Assar Gabrielsson's Fund.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Dagger    Present adress: Epithelial Cell Biology Laboratory, Imperial Cancer Research Fund, P. O. Box 123, 44 Lincoln's Inn Fields, London WC2A 3PX, United Kingdom. Fax: 44-171-269-3361; E-mail: d.baeckstrom @icrf.icnet.uk.
1   The abbreviations used are: mAb, monoclonal antibody; Si-Lea, sialylated Lewis a antigen; NHS-LC-biotin, sulfosuccinimidoyl 6-(biotinamido)hexanoate; PAGE, polyacrylamide gel electrophoresis.
2   R. Sikut, K. Zhang, D. Baeckström, and G. C. Hansson, manuscript in preparation.
3   J. Haraldsson and S. R. Carlsson, personal communication.

ACKNOWLEDGEMENTS

I thank Dr. Gunnar C. Hansson for support and advice throughout this study and Dr. Sven R. Carlsson for supplying unpublished data and helpful suggestions.


REFERENCES

  1. Forstner, J. F., and Forstner, G. G. (1994) in Physiology of the Gastrointestinal Tract (Johnson, L. R., ed), 3rd Ed., pp. 1255-1283, Raven Press, New York
  2. Strous, G. J., and Dekker, J. (1992) Crit. Rev. Biochem. Mol. Biol. 27, 57-92 [Abstract]
  3. Sánchez-Mateos, P., Campanero, M. R., del Pozo, M. A., and Sánchez-Madrid, F. (1995) Blood 86, 2228-2239 [Abstract/Free Full Text]
  4. Sako, D., Chang, X.-J., Barone, K. M., Vachino, G., White, H. M., Shaw, G., Veldman, G. M., Bean, K. M., Ahern, T. J., Furie, B., Cumming, D. A., and Larsen, G. R. (1993) Cell 75, 1179-1186 [Medline] [Order article via Infotrieve]
  5. Berg, E. L., McEvoy, L. M., Berlin, C., Bargatze, R. F., and Butcher, E. C. (1993) Nature 366, 695-698 [CrossRef][Medline] [Order article via Infotrieve]
  6. Magnani, J. L., Nilsson, B., Brockhaus, M., Zopf, D., Steplewski, Z., Koprowski, H., and Ginsburg, V. (1982) J. Biol. Chem. 257, 14365-14369 [Abstract/Free Full Text]
  7. Lindholm, L., Holmgren, J., Svennerholm, L., Fredman, P., Nilsson, O., Persson, B., Myrvold, H., and Lagergård, T. (1983) Int. Arch. Allergy Appl. Immunol. 71, 178-181 [Medline] [Order article via Infotrieve]
  8. Baeckström, D., Hansson, G. C., Nilsson, O., Johansson, C., Gendler, S. J., and Lindholm, L. (1991) J. Biol. Chem. 266, 21537-21547 [Abstract/Free Full Text]
  9. Hilkens, J., Ligtenberg, M. J. L., Vos, H. L., and Litvinov, S. V. (1992) Trends Biochem. Sci. 17, 359-363 [CrossRef][Medline] [Order article via Infotrieve]
  10. Baeckström, D., Zhang, K., Asker, N., Rüetschi, U., Ek, M., and Hansson, G. C. (1995) J. Biol. Chem. 270, 13688-13692 [Abstract/Free Full Text]
  11. Fukuda, M. (1991) Glycobiology 1, 347-356 [Abstract]
  12. Remold-O'Donnell, E., and Rosen, F. S. (1990) Immunodefic. Rev. 2, 151-174 [Medline] [Order article via Infotrieve]
  13. Santamaría, M., López-Beltrán, A., Toro, M., Peña, J., and Molina, I. J. (1996) Cancer Res. 56, 3526-3529 [Abstract]
  14. Hansson, G. C., Baeckström, D., Carlstedt, I., and Klinga-Levan, K. (1994) Biochem. Biophys. Res. Commun. 198, 181-190 [CrossRef][Medline] [Order article via Infotrieve]
  15. Litvinov, S. V., and Hilkens, J. (1993) J. Biol. Chem. 268, 21364-21371 [Abstract/Free Full Text]
  16. Orr, G. A. (1981) J. Biol. Chem. 256, 761-766 [Abstract/Free Full Text]
  17. Carlsson, S. R., and Fukuda, M. (1986) J. Biol. Chem. 261, 12779-12786 [Abstract/Free Full Text]
  18. Johansson, C., Nilsson, O., Baeckström, D., Jansson, E.-L., and Lindholm, L. (1991) Tumor Biol. 12, 159-170 [Medline] [Order article via Infotrieve]
  19. Piller, V., Piller, F., and Fukuda, M. (1989) J. Biol. Chem. 264, 18824-18831 [Abstract/Free Full Text]
  20. Axelsson, B., and Perlmann, P. (1989) Scand. J. Immunol. 30, 539-547 [Medline] [Order article via Infotrieve]
  21. Bazil, V., and Strominger, J. L. (1993) Proc. Natl. Acad. Sci. U. S. A. 90, 3792-3796 [Abstract]
  22. Silverman, L. B., Wong, R. C. K., Remold-O'Donnell, E., Vercelli, D., Sancho, J., Terhorst, C., Rosen, F., Geha, R., and Chatila, T. (1989) J. Immunol. 142, 4194-4200 [Abstract/Free Full Text]
  23. Wong, R. C. K., Remold-O'Donnell, E., Vercelli, D., Sancho, J., Terhorst, C., Rosen, F., Geha, R., and Chatila, T. (1990) J. Immunol. 144, 1455-1460 [Abstract/Free Full Text]
  24. Rieu, P., Porteu, F., Bessou, G., Lesavre, P., and Halbwachs-Mecarelli, L. (1992) Eur. J. Immunol. 22, 3021-3026 [Medline] [Order article via Infotrieve]

©1997 by The American Society for Biochemistry and Molecular Biology, Inc.




This Article
Abstract
Full Text (PDF)
Purchase Article
View Shopping Cart
Alert me when this article is cited
Alert me if a correction is posted
Citation Map
Services
Email this article to a friend
Similar articles in this journal
Similar articles in PubMed
Alert me to new issues of the journal
Download to citation manager
Copyright Permissions
Google Scholar
Articles by Baeckström, D.
Articles citing this Article
PubMed
PubMed Citation
Articles by Baeckström, D.


HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
 All ASBMB Journals   Molecular and Cellular Proteomics 
 Journal of Lipid Research   Biochemistry and Molecular Biology Education 
Copyright © 1997 by the American Society for Biochemistry and Molecular Biology.