human transferrin receptor contains o-linked ...this report describes the isolation and...

THE JOURNAL. OF Bro~ocrca~ CHEMISTRY Vol. 265, No. 1, Issue of January 5, pp. 114-125,199O 0 1990 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A.

Human Transferrin Receptor Contains O-Linked Oligosaccharides*

(Received for publication, May 25, 1989)

Su-11 Do+, Caroline Ennsg, and Richard D. CummingsS7 From the *Department of Biochemistry, University of Georgia, Athens, Georgia 30602 and the SDepartment of Biology, Syracuse University, Syracuse, New York 13244

We have investigated the oligosaccharides in the human transferrin receptor from three different cell lines. During our studies on the structures of the N- linked oligosaccharides of the receptor, we discovered that the receptor contains O-linked oligosaccharides. This report describes the isolation and characteriza- tion of these O-linked oligosaccharides.

Three different human cell lines-K562, A43 1, and BeWo-were grown in media containing either [2-3H] mannose or [6-‘Hlglucosamine. The newly synthesized and radiolabeled transferrin receptors were purified by immunoprecipitation from cell extracts and sodium dodecyl sulfate-polyacrylamide gel electrophoresis. The receptor was proteolytically digested or treated directly with mild base/borohydride. The released radiolabeled glycopeptides and oligosaccharides were separated by a variety of chromatographic techniques, and their structures were analyzed.

The transferrin receptor from all three cell types contains O-linked oligosaccharides that are released from peptide by mild base/borohydride treatment. The receptor from KS62 cells contains at least one O-linked oligosaccharide having two sialic acid residues and a core structure of the disaccharide galactose-N-acetylgalactosamine. In contrast, the O-linked oligosaccharides in the transferring receptors from both A431 and BeWo cell lines are not as highly sialylated and were identified as both the neutral disaccharide galactose- N-acetylgalactosamine and the neutral monosaccharide N-acetylgalactosamine.

In addition, the receptors from all three cell lines contain both complex-type and high mannose-type N- linked oligosaccharides. The complex-type chains in the receptor from A431 cells have properties of blood group A antigens, whereas oligosaccharides in receptors from both BeWo and K562 cells lack these properties. These results are interesting since both A431 and BeWo cells, but not K562 cells, are positive for blood group A antigens.

Thus, our results demonstrate that the human transferrin receptor contains O-linked oligosaccharides and that there are differences in the structures of both the O-linked and complex-type N-linked oligosaccharides on the receptors synthesized by different cell types.

The human transferrin receptor is a transmembrane glycoprotein formed from two identical subunits linked via two

* This work was supported by Grants CA37626-04 to (R. D. C.) and DK40608 (to C. E.) from the National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

V To whom correspondence should be sent. Tel.: 404-542-1760.

intersubunit disulfide bonds (Enns and Sussman, 1981; Schneider et al., 1982; Sutherland et al., 1981). The subunit molecular mass of the receptor is approximately 90,000- 95,000 daltons (Seligman et al., 1979; Wada et al., 1979; Omary and Trowbridge, 1981a; Jing and Trowbridge, 1987; Turkewitz et al., 1988). Based on the amino acid sequence of the human transferrin receptor deduced from its cDNA, there are three potential extracellular N-linked glycosylation sites (Mc- Clelland et al., 1984; Schneider et al., 1984). Results obtained from the sensitivity of the receptor to endoglycosidases suggest that the receptor is glycosylated and that the receptor in some cell types contains one complex-type and two high mannose-type N-linked oligosaccharides (Schneider et al., 1982; Omary and Trowbridge, 1981b). Inhibition of N-linked glycosylation blocks translocation of the receptor to the plasma membrane and interferes with formation of active dimeric species (Enns and Reckhow, 1988).

Several cell surface receptors are also known to contain both N-linked and Ser/Thr-linked (O-linked) oligosaccharides, but it is not known whether the human transferrin receptor contains O-linked oligosaccharides. It has been shown that the receptor purified from human K562 cells incubated with deoxymannojirimycin, an inhibitor of N- linked oligosaccharide processing, retained some sensitivity to neuraminidase (Neefjes et al., 1988). The possibility was raised that this residual sialic acid might be present in O- linked oligosaccharides (Neefjes et al., 1988). However, there is also evidence suggesting that the receptor might not contain O-linked oligosaccharides. For example, the transferrin receptor isolated from cells incubated with tunicamycin, an inhibitor that results in a complete lack of addition of N-linked oligosaccharides, was not detectably metabolically radiolabeled with [3H]glucosamine (Enns and Reckhow, 1988). This precursor can radiolabel both 0- and N-linked oligosaccharides. In addition, the apparent molecular weight of the transferrin receptor from cells incubated in tunicamycin was similar to that predicted from the cDNA sequence for the ungly- cosylated protein (Schneider et al., 1983). It is possible, however, that the receptor might contain just one or two small sized O-linked oligosaccharides that would contribute little to the apparent size of the protein and might go undetected in these indirect analyses.

In the present study, we have examined directly the oligosaccharide moieties in immunopurified transferrin receptor from several human cell lines. We have found that the human transferrin receptor contains O-linked oligosaccharides linked through N-acetylgalactosamine to peptide. In addition, our results demonstrate that there are structural differences in the O-linked and N-linked oligosaccharides in transferrin receptors between different human cell lines.

114

by guest on September 1, 2020

http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

O-Linked Sugars in Human Transferrin Receptors 115

EXPERIMENTAL PROCEDURES

Materials-ConA’-Sepharose and CNBr-activated-Sepharose 4B were obtained from Pharmacia LKB Biotechnology Inc. Bio-Gel P-6 was purchased from Bio-Rad. N-Acetylgalactosamine, a-methyl glucoside, a-methyl mannoside, sialyllactose, Gal@1 + 3)GalNAc, Gal@1 + 3)GlcNAc, Gal@1 + 4)GlcNAc, N-acetylneuraminic acid, pepstatin, aprotinin, phenylmethylsulfonyl fluoride, Helix pomafia agglutinin, Sephadex G-25, QAE-Sephadex, Amberlite MB-3, and Dowex 50-H+ were obtained from Sigma. H. porn&a agglutinin- Sepharose was prepared by the manufacturer’s instructions using CNBr-activated Sepharose 4B. Coupling density of H. pomatia agglutinin was estimated to be 10 mg/ml gel. NaBH, was purchased from J. T. Baker Chemical Co. Dulbecco’s modified Eagle’s medium, RPM1 1640, and Ham’s F-12 medium were obtained from GIBCO. Materials for SDS-polyacrylamide gel electrophoresis and fluorography were purchased from Bio-Rad and Du Pont-New England Nuclear, respectively. Three human cell lines, A431 (human epidermoid carcinoma), BeWo (human choriocarcinoma), and K562 (human erythroleukemia) were obtained from the American Type Culture Collection. Staphylococcu.~ aureus (lo%, w/v) and Pronase (Streptomycesgriceus) were purchased from Calbiochem. Neuraminidase (Arthrobacter urea- faciens) was obtained from Boehringer Mannheim. Monoclonal antibody against the human transferrin receptor (B3/25) and the goat anti-mouse IgG antibody were purchased from Hybri-Tech. Antisera against the purified human transferrin receptor were generated in goats as described by Enns et al. (1981).

The radioisotopes [2-3H]mannose (25 Ci/mmol) and [6-3H]glucosamine (25 Ci/mmol) were purchased from ICN Pharmaceuticals, Inc. UDP-N-acetylgalactosamine [“‘Clgalactose (47.2 mCi/mmol), and UDP-N-acetylglucosamine [3H]glucosamine (20.4 mCi/mmol) were purchased from Du Pont-New England Nuclear. NaB13H14 (19 Ci/mmol) was purchased from Amersham Corp. The radioactive monosaccharides, N-acetyl[‘4C]galactosamine and N-acetyl[3H]glucosamine were prepared by hydrolysis of radioactive sugar nucleotides in 10 mM HCI at 100 “C for 10 min. The radioactive standard sugar alcohols, N-acetyl[‘4C]galactosaminitol and N-acetyl[3H]glucosamin- itol were prepared by sodium borohydride reduction of the appropriate reducing sugars. The standard disaccharides, Gal@1 -+ 3)GalNAc0[3H], Gal@1 + ~)G~cNAcO[~H], and Gal@1 --, ~)G~cNAcO[~H] were prepared by reduction of the reducing oligosaccharides with NaBr3H14.

The radiolabeied - and reduced pentasaccharide standard (GalNAc(ot1 + 3)GalNAc(@l + 4)Gal(al -+ 4)Gal(/31 + 4)GlcO [“HI), derived from the Forssman glycolipid, was kindly supplied by

Dr.-David F. Smith, a colleague at the University of Georgia. Metabolic Radiolabelina of Transferrin Receptor with Radioactive

Sugars-The cell lines A<3i, BeWb, and K582 were maintained in Dulbecco’s modified Eagle’s medium, Ham’s F-12, or RPM1 1640 medium containing 10% fetal calf serum, respectively, penicillin (100 units/ml), and streptomycin (100 rg/ml). For metabolic radiolabeling, the subconfluent (50-70%) cells were incubated for 24 h in media containing either [2-3H]mannose (1 mCi/ml) or [6-3H]glucosamine (1 mCi/ml).

Zmmunoprecipitation of the Transferrin Receptor and SDS-Poly- acrylamide GeE Electrophoresis-Metabolically radiolabeled cells were washed with PBS (6.7 mM KH2P04, 150 mM NaCl, pH 7.4) twice and were primarily solubilized with 1.0 ml of 1% Nonidet P-40 in PBS containing the protease inhibitors pepstatin (O.Ol%), aprotinin (O.l%), and phenylmethylsulfonyl fluoride (0.1 mM) at 4 “C for 20 min. The solubilized cells were centrifuged at 1,000 x g for 1 min. The supernatant was removed and mixed with 100 ~1 of secondary solubilizing buffer (10% Nonidet P-40,10% sodium deoxvcholate, 1% SDS, 10% BSA in PBS) and vortexedseveral times. The-lysates were centrifuged at 40,000 X g for 60 min at 4 “C, and the supernatant was collected. To the supernatant, 60 ~1 of S. aureuS (lo%, v/v) was added, and the mixture was incubated for 20 min at 4 “C to reduce nonspecific binding. The bacteria were removed by microcentrifuga- tion. The supernatant was mixed with 10 ~1 of B3/25 monoclonal antibody and incubated on ice for 15 min. After incubation, 15 ~1 of goat anti-mouse IgG was added and incubated for another 15 on ice. To this was added 60 ~1 of S. aureus (lo%, w/v), and the mixture was

1 The abbreviations used are: ConA, concanavalin A; SDS, sodium dodecyl sulfate; GalNAc-ol, N-acetylgalactosaminitol; GalNAc, N- acetylgalactosamine; GlcNAc, N-acetylglucosamine; PBS, phosphate- buffered saline; LDL, low density lipoprotein; BSA, bovine serum albumin.

placed on ice for 1 h to allow formation of immunoprecipitation complexes. In some experiments, the monoclonal antibody was sub- stituted by 1.4 ~1 of goat anti-transferrin receptor antibody/extract of 5 x lo5 cells. In that case, this addition was followed by 50 ~1 of S. aureu.s cells. The receptor-antibody complexes immobilized on S. aureus were washed three times with washing buffer (0.5% sodium deoxycholate, 0.5% Nonidet P-40, 0.05% SDS, 0.5% BSA in PBS), and the receptor was eluted with Laemmli buffer (Laemmli, 1970) by boiling for 5 min. The eluants were subiected to electrophoresis on 7.5% 3DS-polyacrylamide gel under re&cing conditions The gels were stained with Coomassie Blue (50% MeOH. 20% acetic acid, 0.1% Coomassie Blue) and destained (30% MeOH; 7% acetic acid).

For fluorography, the gels were agitated in the Enhancer solution (Du Pont-New England Nuclear) according to the manufacturer’s instructions and dried and subjected to autoradiography at -70 “C using Kodak X-Omat AR.

Preparation of Radiolabeled Glycopeptides-The gel slices containing radioactive bands of transferrin receptors were incubated at 60 “C for 24 h in 1.0 ml of Pronase digestion buffer (0.1 M Tris-HCl, pH 8.0, 1 mM CaC12) containing 10 mg of Pronase under a toluene atmosphere in a stoppered tube (Cummings et al., 1983). After the incubation, the Pronase-treated material was boiled at 100 “C! for 5 min, and the supernatant was collected. Ohe ml of water was added to the remaining gel pieces, and the sample was boiled again. This second supernatant was removed as above and mixed with the first. One additional water extraction of the gel pieces was done. The pooled supernatants were evaporated under reduced pressure and applied directly to columns of ConA-Sepharose.

Column Chromntography-Radiolabeled glycopeptides were fractionated on a 2-ml column (0.7 X 5 cm) of ConA-Sepharose at room temperature, and 2.0-ml fractions were collected as described previously (Cummings and Kornfeld, 1982; Merkle and Cummings, 1987). Glycopeptides not bound by ConA-Sepharose were designated ConA- I. Glycopeptides bound by the column were eluted first with 10 mM G-methyl glucoside (designated ConA-II) followed by 100 mM w methyl mannoside (designated ConA-III) using a buffer heated to 60 “C. H. pomatia agglutinin-Sepharose lectin affinity chromatography of ConA- glycopeptides was carried out in PBS at room temperature on a l-ml column, and fractions of 1.0 ml were collected. The bound materials were eluted with 50 mM N-acetylgalactosamine in PBS. In all cases, the radioactivity in the collected fractions was determined in a liquid scintillation counter using the ScintiVerse BD (SXlS-4 mixture solution). The chromatography of glycopeptides or oligosaccharides on Sephadex G-25 was conducted on a column (1 X 68 cm) equilibrated in 0.1 M pyridine/acetate buffer (pH 5.6) at room temperature. The apparent V0 and V, were determined with bovine serum albumin and N-acetyl[‘4C]galactosaminitol or Gal@1 + 3) GalNAc0[3H], respectively. The Bio-Gel P-6 column chromatography (1. 5 x 96 cm) was performed in 0.1 M pyridine/acetate buffer (pH 5.6) at room temperature. The V0 and V, were determined as described above for Sephadex G-25 chromatography. Glycopeptides were desalted and separated from monosaccharides on a column of Sephadex G-10 (1 X 48 cm) in 7% n-propyl alcohol in water. The V0 and V, were determined using BSA and a-methyl mannoside, respectively. The ion exchange chromatography of glycopeptides or oligosaccharides was performed on 2-ml columns (0.7 x 5 cm) of QAE- Sephadex in 2 mM Tris/base (pH 9:5), as described previously (Varki and Kornfeld, 1983). Materials bound to the column were eluted stepwise with increasing concentrations of NaCl in Tris/base, and 1.5-ml fractions were collected.

Paper Chromatography-Descending paper chromatography for the separation of GlcNAc and GalNAc was performed for 68-70 h on Whatman No. 1 borate-impregnated filter paper in solvent A (n- butyl alcohol/pyridine/water, 6:4:3) (Cardini and Leloir, 1957). The glycopeptides or released oligosaccharides after p-elimination, mild acid treatment, or neuraminidase treatment were separated for 16- 20 h by descending paper chromatography on Whatman No. 1 filter paper in solvent B (ethyl acetate/pyridine/acetic acid/water, 5:5:1:3). The distribution of radioactivity on the paper chromatograms was determined by cutting the paper strip into l-cm sections and meas- uring the radioactivity of each section in a liquid scintillation counter.

Analysis of Supar ComDosition bv Acid Hvdrolvsis of GlvcoDeDtides- Strong” acid hydrolysisA of [3H]glucosa&ne-iabeled glycobeptides (ConA- and ConA-III) was carried out in 200 ~11 of 4 N HCl at 100 “C for 4 h. The hydrolysates were reacetylated with acetic anhydride and desalted by passage over a column of Amberlite MB-3 (Baenziger and Kornfeld, 1974) and analyzed by descending paper chromatography in solvent A.


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

116 O-Linked Sugars in Human Transferrin Receptors

Release of Oligosaccharides from Glycopeptides or Glycoproteins by Mild AlkalinelBorohydride Treatment-Radiolabeled glycopeptides or glycoproteins were treated with 50 mM NaOH containing 1 M NaBH, at 45 “C for 16 h, as described previously (Iyer and Carlson, 1971; Baenziger and Kornfeld, 1974) to effect fi-elimination of Ser/ Thr-linked oligosaccharides. The reaction was terminated by neu- tralizing by the dropwise addition of 4 N glacial acetic acid. Sodium was removed by passing the mixture over a column of Dowex 50-H’. The borate was removed by repeated evaporation of the sample with a solution of 1 N acetic acid in methanol. The treated sample was suspended in water and applied directly to either a Sephadex G-25 or a Bio-Gel P-6 column.

Desialylation of Oligosaccharides by Mild Acid Treatment-Radio- labeled glycopeptides were desialylated by treatment with 2 N acetic acid at 100 “C! for 1 h. The acid was removed by evaporation under reduced pressure followed by chromatography on Sephadex G-10 in 7% n-propyl alcohol. The released sugars were analyzed by descending paper chromatography in solvent B. In other cases, the treated samples were applied directly to a Sephadex G-25 column.

Glycosidase Treatment-Radiolabeled oligosaccharides were treated with 10 milliunits of neuraminidase in 50 ~1 of 0.1 M sodium citrate buffer (pH 4.6) at 37 “C overnight in a toluene atmosphere. To determine the release of radioactive sugars after neuraminidase treatment, the treated samples were analyzed by descending paper chromatography in solvent B.

RESULTS

The human cell lines used for this study differ in origin and in expression of human blood types. The A431 and BeWo cell lines are derived from human epidermoid carcinoma cells and human choriocarcinoma cells, respectively. Both cell lines are agglutinated by antisera to human blood group A, but they are not agglutinated by antisera to human blood group B (data not shown). It has been shown by others that the A431 cell line is positive for blood group A (Fredman et al., 1983; Parker et al., 1984; Cummings et al., 1985). In contrast, the K562 cell line derived from human erythroleukemia cells is not agglutinated by antisera to either blood group A or B (data not shown). N-linked oligosaccharides in glycoproteins from this cell line have been shown to lack blood group determinants (Kobata, 1983). The rationale for choosing these cell lines for study is the following. Complex-type N-linked oligosaccharides from the receptor for epidermal growth factor in A431 cells contain blood group A determinants that consist partly of terminal nonreducing a-linked N-acetylgalactosamine residues (Cummings et al., 1985). N-Acetylgalactosamine is also, however, a common constituent of O-linked oligosaccharides in animal cell glycoproteins (Sadler, 1984). Thus, in examin- ing the transferrin receptor for the possible presence of O- linked sugar chains, it was advantageous to include a cell line such as K562 which lacks blood group A determinants. The inclusion of the BeWo cell line allowed us to investigate the possible presence of blood group antigens on the transferrin receptor from two cell lines expressing blood group A determinants.

Metabolic Radiolabeling of Human Transferrin Receptor with [2-3H]Mannose and [6-“H]Glucosamine-The transferrin receptor is a recycling receptor and is present in most cells at exceedingly low levels (Trowbridge et al., 1984; May and Cuatrecasas, 1985). For this reason it is not practical to purify significant amounts of the receptor from a variety of cell types for direct chemical analysis. We therefore chose to use the well established technique of metabolic radiolabeling with radioactive precursor sugars. This technique has been used successfully in a number of studies on cell surface receptors (Goldberg et al., 1983; Cummings et al., 1983, 1985, 1989; Pathak et al., 1988) and is particularly advantageous since it allows analysis of very small quantities of material. Addi- tionally, this approach is particularly suitable for direct immunoprecipitation techniques to purify the receptor for analy-

sis, thus dramatically reducing the possibility of contaminat- ing materials.

To radiolabel the sugar chains of the transferrin receptor, the three human cell lines K562, A431, and BeWo were grown for 24 h in normal complete media containing either [2-3H] mannose or [6-“Hlglucosamine. The transferrin receptors in the cell extracts were precipitated with anti-transferrin receptor antibody, and the immunoprecipitates were analyzed by SDS-polyacrylamide gel electrophoresis and fluorography (Fig. 1). The receptor subunits from all three cell lines were radiolabeled with the precursor monosaccharides. The apparent mobilities of the subunits were all similar and corre- sponded to an expected size of approximately 90,000 daltons, as compared with phosphorylase b (97,400 daltons) used as a standard.

Lectin Affinity Column Chromatography of K562 Transfer- rin Receptor Glycopeptides-The sections of gel containing the [3H]glucosamine-labeled receptors from K562 cells were removed, and glycopeptides were prepared by digesting samples directly with Pronase. The radiolabeled glycopeptides were applied to columns of ConA-Sepharose, and the fractionation profile of glycopeptides are shown in Fig. 2.4.

Previous studies have shown that ConA-Sepharose does not interact with high affinity to either O-linked oligosaccharides or complex-type tri- or tetraantennary and bisected biantennary N-linked oligosaccharides (Ogata et al., 1975; Krusius et al., 1976; Cummings and Kornfeld, 1982; Merkle and Cum- mings, 1987). However, complex-type biantennary N-linked oligosaccharides bind to ConA-Sepharose and can be eluted with 10 mM a-methyl glucoside (Ogata et al., 1975; Krusius et al., 1976; Cummings and Kornfeld, 1982; Merkle and Cum- mings, 1987). High mannose-type and hybrid-type N-linked oligosaccharides bind to this lectin with high affinity, and

FIG. 1. Electrophoresis of immunoprecipitated transferrin receptors (T&R) from three human cell lines. K562, A431, and BeWo cells were metabolically radiolabeled with [6-3H]glucosamine or [2-‘Hlmannose as described under “Experimental Procedures.” The cells were then solubilized, and the receptors were immunoprecipitated as described under “Experimental Procedures.” The eluates of the immunoprecipitates were subjected to electrophoresis on SDS- polyacrylamide gel (7.5%) under reducing conditions, and the radiolabeled receptors were visualized by fluorography. Panel A, [6-“H] glucosamine-labeled cells; panel B, [2-“Hlmannose-labeled cells; lanes I, K562 cells; lanes 2, A431 cells; lanes 3, BeWo cells. The positions of molecular weight markers are indicated by arrows: myosin, 200,000; @-galactosidase, 116,000; phosphorylase b, 97,400; bovine serum albumin, 68,000; ovalbumin, 43,000.


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


1)gel pieces containing Glucosamine-labeled K562 Tf-R 2)preparation of glycopeptides by Pronase digestion

3)chromatography on Con A-Sepharose

1)strong acid hydrolysis followed by reacetylation 2)descending paper chromatography(6:4:3)

100 v

I

GlcNAc B 60 -

1)strong acid hydrolysis followed by reacetylation 2)descending paper chromatography(6:4:3)

60 - E 8

GalNAc (18%) R T I

GalNAc (0%) I I

distance(cm) distance(cm)

FIG. 2. Lectin chromatography of [3H]glucosamine-labeled glycopeptides prepared by Pronase treatment of the K562 transferrin receptor (Tf-R) and compositional analysis. A, a gel slice containing the [3H]glucosamine-labeled K562 transferrin receptor (Fig. 1, panel A, lane I ) was excised from the gel and directly digested with Pronase. The resulting glycopeptides were fractionated on a column of ConA-Sepharose. Bound glycopeptides were eluted sequentially with 10 mM a-methyl glucoside (cu-m-Glc) and 100 mM a-methyl mannoside (u-m-Man) as described under “Experimental Procedures.” The fractions were pooled as indicated by horizontal bars to give materials designated ConA- (unbound fraction), ConA-II (bound and eluted with 10 mM a-methyl glucoside), and ConA- (bound and eluted with 100 mM a-methyl mannoside) as described. The compositional analysis of ConA- (B) and ConA- (C) glycopeptides after strong acid hydrolysis followed by reacetylation was done on descending paper chromatography in solvent A as described under “Experimental Procedures.” The positions of migration of the authentic standards GalNAc and GlcNAc are indicated. Neutral monosaccharides, e.g. mannose and galactose, do not migrate from the origin in this system.

elution requires a high concentration of a-methyl mannoside (Cummings and Kornfeld, 1982; Merkle and Cummings, 1987). In our study, the glycopeptides not bound by ConA- Sepharose were designated as ConA-I, whereas the glycopeptides bound and eluted with 10 mM a-methyl glucoside and eluted with 100 mM a-methyl mannoside were designated as ConA- and ConA-III, respectively.

Analysis of pH]Glucosamine-labeled K562 Transferrin Receptor Glycopeptides of ConA-I and ConA-III-The glucosamine-labeled glycopeptides in ConA- and ConA- were hydrolyzed in the strong acid, and the released monosaccharides were reacetylated and analyzed by descending paper chromatography in solvent A (Fig. 2, B and C). Eighteen

percent of the radioactivity recovered from ConA- glycopeptides was present in GalNAc, and the remainder was present in GlcNAc (Fig. 2B).

These data demonstrate that glycopeptides of K562 transferrin receptor contain significant amounts of GalNAc. (It should be noted that this treatment destroys sialic acid, which can also be radiolabeled from the [3H]glucosamine precursor.)

All of the recovered radioactivity from ConA- glycopeptides was present in GlcNAc (Fig. 2C). These results demonstrate that the high mannose-type or hybrid-type N-linked oligosaccharides of transferrin receptor from K562 cells contain GlcNAc, but not GalNAc, as expected for such types of oligosaccharides.


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


1)gel pieces containing Glucosemine-labeled K562 Tf-R 2)dwect beta-elimination

3)chromatography on Sephadex G-25 I

1) mild acid treatment for desialylation 2)Amberlite MB-3 for removal of ionic species

3)chromatography on Sephadex G-25 I + B

vo Ve

G-P-I

descending paper chromatography(5:5:1:3)

600 -

600 .

E

8 400 -

200 J

Gal-GalNAcitol

JI GalNAcitol

JI

descending paper chromatography(5:5:1:3)

I. v Gal-GalNAcitol

I D

0 0 10 20 30 40 50 0 10 20 30 40 50

. I

distance(cm) distance(cm)

FIG. 3. Chromatography on Sephadex G-25 of [3H]glucosamine-labeled glycopeptides prepared by direct mild base/borohydride treatment of the K562 transferrin receptor (P-R) and analysis of released oligosaccharides by descending paper chromatography. A, the gel pieces containing [3H]glucosamine-labeled K562 transferrin receptor (Fig. 1, panel A, lane I) were directly treated with NaOH/NaBH, and resulting glycopeptides were applied to a column of Sephadex G-25 as described under “Experimental Procedures.” The fractions were pooled as indicated by horizontal bars to give materials designated G-P-O. B, the G-P-O was


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


We then sought to determine the structures of the oligosaccharides in the ConA- glycopeptides containing these GalNAc residues and determine whether GalNAc was present in O-linked oligosaccharides. Treatment of glycopeptides or glycoproteins containing O-linked oligosaccharide with mild base in the presence of borohydride causes the cleavage of O- linked sugars from peptides via p-elimination, with concomi- tant reduction and protection of the reducing terminus (Iyer and Carlson, 1971; Baenziger and Kornfeld, 1974). This treatment also degrades a small percentage of N-linked oligosaccharide, but in general, the N-linked oligosaccharides are relatively stable to treatment with mild base (Hounsell et al., 1984; Debray et al., 1984).

Rather than first digest the receptors with Pronase, as described above, gel slices containing the [3H]glucosamine- labeled transferrin receptor from K562 cells were directly treated with mild base/borohydride. The treated samples were applied to a column of Sephadex G-25 to separate released and small sized O-linked oligosaccharides, if present, from large sized N-linked oligosaccharides (Fig. 3A). The treated material, however, eluted as a broad peak designated G-P-O near the region of the V, of the column.

Previous studies in our laboratory have demonstrated that O-linked oligosaccharides containing sialic acid have an un- usual elution pattern on Sephadex G-25 and often appear to elute near the V,. The unusually broad elution pattern in Fig. 3A suggested to us the possibility that sialylated O-linked oligosaccharides released from peptide might be present in the glycopeptides derived from the transferrin receptor of K562 cells.

To explore this possibility, we took two different ap- proaches. For the first approach, the radiolabeled material in G-P-O from Fig. 3A was treated with mild acid to remove sialic acid residues as described under “Experimental Proce- dures.” After the desialylation, the reaction mixture was applied to an ion exchange column of Amberlite MB-3 to remove any ionic species, and the flow-through containing neutral species was reapplied to the column of Sephadex G-25. Ap- proximately 17% of the total radioactivity was recovered as small sized material designated G-P-II, and the remaining radioactivity eluted near the V0 and was designated G-P-I (Fig. 3B). The material in G-P-I and G-P-II from Fig. 3B was directly analyzed by descending paper chromatography in solvent B (Fig. 3, C and D, respectively). All the material in G-P-I remained at the origin, indicating that it represents large sized material and is likely to be the N-linked oligosaccharides (Fig. 3C). In contrast, the majority of radioactivity in G-P-II from K562 transferrin receptor glycopeptides was recovered as the disaccharide galactose-N-acetylgalactosaminitol (Fig. 30).

Analysis of the O-Linked Oligosaccharides Released from K562 Cell Transferrin Receptor by Mild BaselBorohydride- We then took a second approach to determine whether the O- linked chains exist in the K562 transferrin receptor and whether they contain sialic acid residues. The [3H]glucosamine-labeled glycopeptides prepared from direct mild base/ borohydride treatment of gel slices containing transferrin receptor were fractionated on the column of Bio-Gel P-6 equilibrated with pyridine/acetate buffer as described under “Experimental Procedures” (Fig. 4A). The material in frac-

tions numbered 60-93 and 94-120 were pooled and designated pool A and pool B, respectively.

A portion of each of these samples was applied to a column of QAE-Sephadex to separate species based on negative charge (Varki and Kornfeld, 1983). Oligosaccharides having one negative charge (equivalent to 1 sialic acid residue) or two negative charges (2 sialic acid residues) are eluted by 20 or 70 mM NaCl, respectively (Varki and Kornfeld, 1983; Cummings et al., 1983). The material in pool A was heterogeneous in terms of total negatively charged residues (Fig. 4B). This material bound by QAE-Sephadex was pooled and designated as QAE-A. The elution profile of the pool A material on QAE- Sephadex is complex and is consistent with this fraction containing heterogeneously sialylated complex-type N-linked oligosaccharides. In contrast, the oligosaccharides in pool B from Fig. 4A were eluted with 70 mM NaCl, indicating the presence of 2 sialic acid residues (Fig. 4C). The material eluted with 70 mM NaCl was pooled and designated QAE-B.

To determine more precisely the presence of sialic acid in the QAE-A and QAE-B material, portions of each pool were treated with neuraminidase to cleave the sialic acid residues. After the neuraminidase treatment, treated samples were analyzed by descending paper chromatography in solvent B. The neuraminidase treatment of QAE-A released a minor amount of radioactivity that was present in sialic acid, whereas the majority of radioactivity stayed at the origin during paper chromatography (Fig. 40). In contrast, the majority of radioactivity of QAE-B treated with neuraminidase was present in N-acetylneuraminic acid and the disaccharide galactose-N-acetylgalactosaminitol (Fig. 4E). The ratio of radioactivity in NeuAc/Gal-GalNAc-al was 2:l. To- gether, these results demonstrate that the O-linked oligosaccharide in QAE-B is a disaccharide galactose-N-acetylgalactosaminitol containing 2 sialic acid residues. This appears to represent the intact structure of O-linked oligosaccharides in the transferrin receptor from K562 cells. Such disialylated O-linked oligosaccharides are common to a wide variety of mucins and animal cell glycoproteins (Sadler, 1984).

The relative amounts of radioactive GalNAc and GlcNAc in the transferrin receptor may be used to estimate the possible number of O-linked oligosaccharides to N-linked oligosaccharides. The human transferrin receptor contains three N-linked oligosaccharides composed of one complex-type and two high mannose/hybrid-type chains (Schneider et al., 1982; Omary and Trowbridge, 1981b). In our studies on the K562- derived receptor, the glycopeptides containing both complex- type N-linked oligosaccharides and O-linked oligosaccharides are expected to be recovered in the ConA- fraction (Fig. 2A). Compositional analysis of strong acid hydrolysates of ConA- I demonstrates that 82% of the radioactivity is in GlcNAc, and 18% is in GalNAc, and thus the approximate ratio of GlcNAc and GalNAc in those glycopeptides is 4:l. The complex-type N-linked oligosaccharide is likely to be at least a triantennary chain containing at least 5 GlcNAc residues to account for its lack of binding to ConA-Sepharose (Merkle and Cummings, 1987). Thus, the ratio of GlcNAc to GalNAc suggests that there are probably no more than one or two O- linked oligosaccharides in the human transferrin receptor. This estimation assumes that the GalNAc and GlcNAc residues are radiolabeled to approximately the same specific

treated with mild acid to remove sialic acid residues, and treated samples were reapplied to a column of Sephadex G-25. The fractions were pooled as indicated by horizonal bars to give materials designated G-P-I and G-P-II. V. and V, were determined with BSA and GalNAc-al/Gal-GalNAc-01, respectively. G-P-I (C) and G-P-II (D) were analyzed by descending paper chromatography in solvent B. The migration positions of authentic standards Gal- GalNAc-ol and GalNAc-ol are indicated.


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


1 5

el pieces containing Glucosamine-labeled K562 TFR )direct beta-elimination 3)chromatpgraphy on Bio Gel P-6

100

60

EW 8

40

20

0 0

chromatography on QAE-Sephadex I

I v

fraction(lml)

J fraction

1)Neuraminidase treatment 2)descendmg paper chromatography(5:5:1:3)

N-acetylneuraminic acid

Gal-GalNAcitol

10 20 30 40 50

distance(cm)

60

50

40

EM 8

20

10

0 * 0

chromatography on QAE-Sephadex

AL QAE-B

1)Neuraminidase treatment 2)descending paper chrpmatography(5:5:1:3)

140

iE I *

N-acetylneuraminic acid

Gal-GalNAcitol

distance(cm) FIG. 4. Chromatography on Bio-Gel P-6 of [3H]glucosamine-labeled glycopeptides prepared by mild

base/borohydride treatment of the K562 transferrin receptor (W-R) and analysis of released oligosaccharides. A, the gel pieces containing [3H]glucosamine-labeled transferrin receptor were directly treated with NaOH/NaBH4, and treated samples were applied to a column of Bio-Gel P-6 as described under “Experimental Procedures.” VO and V, were determined with BSA and GalNAc-01, respectively. SL indicates the elution position


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


activity, a reasonable assumption that is supported by previous studies on the LDL receptor in A431 cells. In those studies, it was demonstrated that the radioactive GalNAc and GlcNAc derived from [3H]glucosamine closely approximated the actual chemical amount of these residues in the receptor (Cummings et al., 1983).

Lectin Affinity Column Chromatography of A431 and Be Wo Transferrin Receptor Glycopeptides--We next sought to ex- amine the carbohydrate chains on the transferrin receptor from other human cell lines and compare their structures with those found above for the chains on the receptor from K562 cells. The [3H]glucosamine-labeled glycopeptides from the A431 and BeWo transferrin receptors were obtained by Pro- nase treatment of the appropriate gel slices shown in Fig. 1 and applied to a column of ConA-Sepharose. The results of the fractionation are summarized in Table I. As in the case of the glycopeptides from the K562 receptor, radiolabeled glycopeptides were recovered in ConA-I, ConA-II, and ConA- III. When the A431 and BeWo transferrin receptor glycopeptides were hydrolyzed in strong acid and analyzed by descending paper chromatography, a significant amount of GalNAc was recovered in the ConA- glycopeptides but not in any other glycopeptide fraction. These results suggested that the transferrin receptors from these other two human cell lines also might contain O-linked oligosaccharides. However, these two cell lines are also positive for human blood group A, indicating that they could contain N-linked oligosaccharides with terminal a-linked GalNAc.

To explore this possibility, we applied [3H]glucosamine- labeled glycopeptides from the A431 and BeWo transferrin receptors to a column of H. pomatia agglutinin-Sepharose. H. pomatia agglutinin agglutinates blood group A, but not B or H erythrocytes, and it binds with high affinity to oligosaccharides containing blood group A determinants and specifically a-linked IV-acetylgalactosamine residues at the nonreducing termini (Hammarstrom and Kabat, 1971; Torres et al., 1988). Previous studies (Cummings et al., 1985) have demonstrated that the complex-type N-linked oligosaccharides in the receptor for epidermal growth factor from A431 cells contain blood group A determinants. Based on this previous finding, it can be predicted that the complex-type N-linked oligosaccharides of the transferrin receptor from A431 cells would also contain blood group A antigens. Consistent with this prediction, we found that 17% of the ConA- glycopeptides from A431 transferrin receptor bound tightly to the H. pomatia agglutinin- Sepharose and were eluted with 50 mM GalNAc (Fig. 5A). In contrast, the ConA- glycopeptides from [3H]glucosamine- labeled K562 did not interact with H. pomatia agglutinin- Sepharose significantly (Fig. 5B). Similarly, the glycopeptides from the transferrin receptors of BeWo cells did not interact with H. pomatiu agglutinin-Sepharose (data not shown). The precise quantitation of these results of chromatography on H. pomatia agglutinin-Sepharose is compiled in Table I. These data indicate that the transferrin receptor from A431 cells (blood type A) contains terminal and nonreducing a-linked IV-acetylgalactosamine residues, whereas the transferrin receptors from BeWo (blood type A) and K562 do not contain

significant amounts of this terminal sequence. Analysis of the A431 and BeWo Transferrin Receptor PHI

Glucosamine-labeled Glycopeptides in ConA-I-Portions of the [3H]glucosamine-labeled ConA- glycopeptides from transferrin receptors derived from A431 and BeWo were treated with 50 mM NaOH containing 1 M NaBH4 to effect p-elimination of any Ser/Thr-linked oligosaccharides. The glycopeptides were analyzed and separated by chromatography on a column of Sephadex G-25. The cnromatographic profiles of A431 and BeWo transferrin receptor glycopeptides not treated with mild base/borohydride are shown in Figs. 6A and 7A, respectively. The chromatographic profiles of the treated samples from A431 and BeWo are shown in Figs. 6B and 7B, respectively.

Mild base/borohydride treatment of ConA- glycopeptides of the transferrin receptor of A431 and BeWo cells released oligosaccharides eluting at the location of V,, representing about 25 and 14% of the total radioactivity applied, respectively (Figs. 6B and 7B). The majority of the radioactivity eluted near the VO and was designated G-P-I. The small sized oligosaccharides were designated G-P-II.

The oligosaccharides in G-P-I and G-P-II from Figs. 6B and 7B were analyzed by descending paper chromatography in solvent B before (Figs. 6D and 70) and after passage over a column of mixed bed, ion exchange resin Amberlite MB-3 to remove charged species (Figs. 6E and 7E). The radiolabeled oligosaccharides in G-P-II from A431 and BeWo were recovered as the disaccharide galactose-N-acetylgalactosamini- to1 and the monosaccharide N-acetylgalactosaminitol after passing over a column of Amberlite MB-3 (Figs. 6E and 7E). When this material was analyzed directly before passage over Amberlite MB-3, there was a major oligosaccharide present which migrated slowly (indicated by the radioactive peak marked as the asterisk in Figs. 6D and 70). Because this oligosaccharide is charged, we initially presumed that it might be a monosialylated derivative of the O-linked disaccharide. However, it was resistant to neuraminidase treatment (data not shown). It is likely that this material is a small molecular weight glycopeptide containing 1 or 2 sugars and 1 or 2 amino acid residues and might be a product of inefficient /3-elimination. This may result from the fact that this material was derived by mild base/borohydride treatment of previously prepared glycopeptides from the A431 and BeWo transferrin receptors. The p-elimination reaction requires the presence of amino acids both carboxyl- and amino-terminal to the glycosylated serine or threonine residue (Derevitskaya et al., 1967).

It should be noted that the mild base/borohydride treatment of the receptor from K562 cells is likely to be more efficient since in those experiments, the intact receptor was treated directly. The unknown material in Figs. 6D and 70 was not analyzed further since it was not seen in oligosaccharides derived from the K562 transferrin receptor. In summary, these data indicate that the transferrin receptors of both A431 and BeWo contain neutral species of O-linked oligosaccharides consisting of both a monosaccharide GalNAc and the disaccharide Gal-GalNAc.

of authentic standard sialyllactose. The fractions were pooled as indicated by the horizontal bars to give materials designated pool A and pool B. Pool A material (B) and pool B material (C) were applied to a column of QAE- Sephadex in 2 mM Trisbase (pH 9.5), and the bound materials were eluted stepwise with the indicated concentration of NaCl as described under “Experimental Procedures.” The fractions were pooled as indicated by the horizontal bars to give materials designated QAE-A and QAE-B. Portions of both QAE-A and QAE-B were treated with neuraminidase overnight at 37 “C under a toluene atmosphere and then analyzed directly by descending paper chromatography in solvent B (D and E, respectively). The migration positions of the authentic standards N- acetylneuraminic acid and Gal-GalNAc-ol are indicated.


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


TABLE I Analysis of transferrin receptor glycopeptides from different human

cell lines

ConA-Sepharose glyco- Descending paper H. pomatia agglu-

Cell lines peptides” chromatographg tinin-Sepharose’

Pool Radioactivity GalNAc GlcNAc Unbound Bound

70 of cpm % of cpm % of cpm A431 ConA- 44.3

ConA- 20.6 is N”; lJii IsI ConA- 35.1 0 100 ND ND

BeWo ConA- ConA- ConA-

69.1 10.4 :; N? :I! Nk 20.5 0 100 ND ND

K562 ConA- 76.5 82 100 0 ConA- 4.3 I$ ND ND ND ConA- 19.2 0 100 ND ND

’ The [3H]glucosamine-labeled glycopeptides derived by Pronase treatment of transferrin receptors from the different cell lines indicated were applied to columns of ConA-Sepharose. The percentages of total radioactivity in each of the three ConA pools, designated ConA-I-III, are shown.

* Portions of each of the [3H]glucosamine-labeled glycopeptides indicated were hydrolyzed in strong acid, reacetylated, and the re- sultant released radiolabeled monosaccharides were separated by descending paper chromatography in solvent A, as indicated under “Experimental Procedures.” The amounts of radioactivity recovered in either GalNAc or GlcNAc are indicated as a percentage of the total radioactivity in the two species combined.

’ Portions of each of the [3H]glucosamine-labeled glycopeptides indicated were applied to columns of H. pomatia agglutinin-Sepharose and the percentages of radiolabeled glycopeptide either unbound or bound to the lectin were determined.

d Not determined.

DISCUSSION

The human transferrin receptor is extensively modified during its biosynthesis. It undergoes co-translational N-linked glycosylation (Omary and Trowbridge, 1981a). In addition, the transferrin receptor dimerizes (Enns and Sussman, 1981), forms intersubunit disulfide bonds (Sutherland et al., 1984; Goding and Harris, 1981), is modified on Se? by phosphorylation (Schneider et al., 1984; Davis and Meisner, 1987), and is acylated with palmitate on Cys6’ (Omary and Trow- bridge, 1981a, 1981b). The role that many of these modifications play in the folding or maintenance of the function of this receptor is not clear. Site-directed mutagenesis altering any of amino acids modified by acylation (Jing and Trow- bridge, 1987), phosphorylation (Rothenberger et al., 1987; Zerial et al., 1987; McGraw et al., 1987; Davis and Meisner, 1987), and disulfide bond formation (Jing and Trowbridge, 1987) does not appear to alter the conformation or behavior of the receptor detectably. N-Linked glycosylation, however, does appear to be critical for the correct folding and transport of the transferrin receptor to the cell surface (Enns and Reckhow, 1988).

In addition to these modifications, we demonstrate in this paper that the human transferrin receptor in all the cell lines that we have examined contains O-linked oligosaccharides. This modification may have gone undetected because there are no reported specific inhibitors of O-linked glycosylation. Considering the results of the present study, it is likely that the receptor contains a single O-linked oligosaccharide. Treat- ment of the transferrin receptor with 0-glycanase results in a subtle shift in mobility (less than 1000 molecular weight) which is difficult to detect (data not shown). Recently, Neefjes and co-workers (1988) presented evidence that the transferrin receptor isolated from deoxymannojirimycin-treated K562

Chromatography of [ 3 H]glucosamlne-labeled ConA-l glycopeptldes of Tf-receptor on

Helix Pomatla Agglutlnln-Sepharose

elutlon wlth 50 mM GalNAc

B

elutlon wlth 50 mM GalNAc

0 0 10

Fractlon(1 ml)

20

FIG. 5. Chromatography on H. pomatiu agglutinin of [‘HI glucosamine-labeled ConA- glycopeptides of transferrin receptor. A, the [3H]glucosamine-labeled ConA- glycopeptides of A431 transferrin receptor (not bound by ConA-Sepharose) were applied to a column of H. pomatia agglutinin-Sepharose as described under “Experimental Procedures.” Glycopeptides bound by the lectin were eluted with 50 mM GalNAc. B, the ConA- glycopeptides derived from K562 transferrin receptor were also analyzed by chromatography on H. pomatia agglutinin-Sepharose. Tf, transferrin.

cells retained sensitivity to neuraminidase, whereas class 1 antigens that do not contain O-linked sugars isolated from these cells were not sensitive to neuraminidase (Neefjes et al., 1988). Deoxymannojirimycin inhibits mannosidase I activity and thus inhibits complex carbohydrate formation (Fuhr- mann et al., 1984). Since sialic acids were bound to the transferrin receptor after treatment with this inhibitor, they hypothesized that the sialic acids were contained in O-linked oligosaccharides. It should be noted that Poola and Lucas (1988) presented evidence that a glycoprotein in chick oviduct, which may be the chick transferrin receptor, contains at least one O-linked oligosaccharide.

The observation that the human transferrin receptor contains only a small number of O-linked oligosaccharides may partly explain other previous observations. For example, when A431 cells were treated with tunicamycin, an inhibitor of N- linked glycosylation, there was no apparent incorporation of [3H]glucosamine into the altered receptor (Enns and Reckhow 1988). The lack of incorporation of [3H]glucosamine into the


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


1) Con A-l glycopeptides of [ 3 H ]-Glucosamine-labeled A431 Tf-R 2) Chromatography yn Sephadex G-25

b 300 1- [Before jSel?knation] - -[After B-elimination]

Fraction (1 ml)

-Descending paper chromatography (5:5:1:3)

7 30 7 [After Amberlite MB-31

1 E Gal-galNAcitol

+ 20

R GalNAcitol

- [Before Amberlite MB-31

GalNAciL

Distance (cm) Distance (cm)

FIG. 6. Chromatography on Sephadex C-25 of [3H]glucosamine-labeled ConA-I glycopeptides of A43 1 transferrin receptor (Tf-R) treated with mild base/borohydride and analysis of released oligosaccharides by descending paper chromatography. The [3H]glucosamine-labeled ConA- glycopeptides from A431 transferrin receptor were applied to a column of Sephadex G-25 before (A) and after (B) treatment with 0.05 M NaOH and 1 M NaBH, as described under “Experimental Procedures.” As shown in panel B, fractions were pooled as indicated by horizontal bars to give materials designated G-P-I and G-P-II. V, and V, were determined with BSA and GalNAc-al/Gal-GalNAc-01, respectively. The G-P-I was directly analyzed by descending paper chromatography in solvent B (C), and G-P-II was analyzed by descending paper chromatography in the same system before (D) and after (E) passage over a column of Amberlite MB-3 as described under “Experimental Procedures.” The migration positions of the authentic standards GalNAc-ol and Gal-GalNAc-ol are indicated.

O-linked oligosaccharides (Pathak et al., 1988). Several other membrane receptors have been shown to be

glycoproteins containing both N-linked and O-linked oligosaccharides. These include the LDL receptor (Cummings et al., 1983; Russell et al., 1984; David et al., 1986) and the interleukin-2 receptor (Leonard et al., 1984; Gallis et al., 1986). The LDL receptor is extensively modified with two N-linked oligosaccharides and approximately 18 O-linked oligosaccharides (Cummings et al., 1983; Davis et al., 1986). The O-linked oligosaccharides appear to play a major role in the function of both these receptors. Transfection of the genes for these receptors into a Chinese hamster cell line, ldl-d, defective in O-linked glycosylation (Kingsley et al., 1986) results in non- functional receptors in both cases (Kozarsky et al., 1988a, 1988b). In addition, the LDL receptor synthesized in Chinese hamster ovary cells resistant to the ionophore monensin appears to have a reduced number of O-linked oligosaccharides and a reduced affinity for LDL (Yoshimura et al., 1987).

receptor from tunicamycin-treated cells could be interpreted as lack of O-linked oligosaccharides since the addition of the latter chains would not be expected to be inhibited by tunicamycin. Since the receptor may contain only a single O- linked oligosaccharide, however, it is possible that a small degree of radiolabeling would not be readily detectable. Recent experiments have indicated that the transferrin receptor synthesized in the presence of tunicamycin cannot be transported to the cell surface and does not leave the endoplasmic reticulum.* Thus, under these conditions, it may not gain access to a compartment in the cell in which it can undergo O-linked glycosylation or the receptor could be aggregated and not in a proper conformation for addition of O-linked sugars. It should be noted, however, that the LDL receptor in cells derived from patients with familial hypercholesterolemia type 2 is unable to exit the endoplasmic reticulum but does contain

’ C. Enns, manuscript in preparation.


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


1) Con A-l glycopeptides of [ 3 H ]-Glucosemine-labeled BeWo Tf-R 2) Chromatography on Sephadex G-25

1400

t

[Before p-elimination] A

1200

:‘,( * [After P-el;nation]

800

600

Fraction (1 ml)

Descending paper chromatography (5:&l 3)

[Before Amberlite MB-31 1

1 Gal-galNAcitol 60

% 6 600

40 400

60 I [After Amberlite MB-31 _

J( Gal-galNAcitol 0

Distance (cm) Distance (cm) Distance (cm)

FIG. 7. Chromatography on Sephadex G-26 of [3H]glucosamine-labeled ConA- glycopeptides of BeWo transferrin receptor (Tf-R) treated with mild base/borohydride and analysis of released oligosaccharides by descending paper chromatography. The [3H]glucosamine-labeled ConA- glycopeptides from BeWo were applied to a column of Sephadex G-25 before (A) and after (B) treatment with 0.05 M NaOH and 1 M NaBH, as described under “Experimental Procedures.” As shown in panel B, fractions were pooled as indicated by horizontal bars to give materials designated G-P-I and G-P-II. V,, and V, were determined with BSA and GalNAc- al/Gal-GalNAc-ol, respectively. The G-P-I was directly analyzed by descending paper chromatography in solvent B (C), and G-P-II was analyzed by descending paper chromatography in the same system before (D) and after (E) passage over a column of Amberlite MB-3 as described under “Experimental Procedures.” The migration positions of the authentic standards GalNAc-ol and Gal-GalNAc-ol are indicated.

The detailed structures of the O-linked oligosaccharides in the LDL receptor from A431 cells have been determined. The O-linked oligosaccharides appear to be heterogeneous and composed of a core disaccharide galactose-N-acetylgalactosamine and I or 2 sialic acid residues (Cummings et al., 1983).

Our results demonstrate that the human transferrin receptor from all cell types examined contains O-linked oligosaccharides. The structures of these oligosaccharides in receptors from different cell types, however, differ from each other. Based on our calculations, which suggest that the receptor contains a single O-linked oligosaccharide, it is possible that this modification occurs at a highly conserved position within the polypeptide. We also performed preliminary studies on the transferrin receptor from HL-60 cells, which lack either blood group A or B determinants. Our results indicate that the receptor from this cell line also contains significant amounts of IV-acetylgalactosamine, which is presumably in

the O-linked oligosaccharide (data not shown). We are cur- rently in the process of identifying the specific amino acid(s) in the human receptor which contain the O-linked oligosaccharides.

In addition to identifying U-linked oligosaccharides on the transferrin receptor, we have also found that the complex- type N-linked oligosaccharides in the receptor from A431 cells contain features related to blood group A antigen. Earlier studies on the epidermal growth factor receptors from the same cell line indicate that they too possess N-linked oligosaccharides containing blood group A antigens (Fredman et al., 1983; Parker et aZ., 1984; Childs et al., 1984; Cummings et aZ., 1985). In contrast, transferrin receptors from K562, BeWo, and HL-60 cells do not contain such a blood group antigen (data not shown). Interestingly, the BeWo cells are agglutinated by antisera to human blood type A, but not to type B, indicating that the cells do express some blood group A


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/


antigens. Further studies on the transferrin receptor from various cell lines will be done to determine whether the presence of blood group antigens on the transferrin receptor is limited to the transformed phenotype of A431 cells or whether the lack of A blood group antigen on the transferrin receptor in A-positive cells is peculiar to BeWo cells. In addition, future studies will address the possibility that the differences in both N- and 0-glycosylation of the receptor from different cell lines may affect receptor function and/or binding activity.

Acknowledgments-We would like to acknowledge the technical assistance of Barbara Root.

REFERENCES

Baenziger, J., and Kornfeld, S. (1974) J. Biol. Chem. 249,7270-7281 Cardini, C. E., and Leloir, L. F. (1975) J. Biol. Chem. 225, 317-324 Childs, R. A., Gregoriou, M., Scudder, P., Thorpe, S. J., Rees, A. R.,

and Feizi, T. (1984) EMBO J. 3, 2227-2233 Cummines. R. D.. and Kornfeld. S. (1982) J. Biol. Chem. 257.11235-

11240 - Cummings, R. D., Kornfeld, S., Schneider, W. J., Hobgood, K. K.,

Tolleshaua. H.. Brown, M. S., and Goldstein, J. L. (1983) J. Biol. Chem. 258,15261-15273

Cummings, R. D., Soderquist, A. M., and Carpenter, G. (1985) J. Biol. Chem. 260,11944-11952

Cummings, R. D., Merkle, R. K., and Stults, N. L. (1989) Methods Cell Biol. 32, 142-185

Davis, C. G., Elhammer, A., Russell, D. W., Schneider, W. J., Korn- feld, S., Brown, M. S., and Goldstein, J. L. (1986) J. Biol. Chem. 261,2828-2838

Davis, R. J., and Meisner, H. (1987) J. Biol. Chem. 262,16041-16047 Debray, H., Strecker, G., and Montreuil, J. (1984) Biochem. Sot.

Truns. 12,611-612 Derevitskaya, V. A., Vafina, M. G., and Kochetkov, N. K. (1967)

Carbohydr. Res. 3,377-388 Enns, C., Shindelman, J., Tonik, S., and Sussman, H. (1981) Proc.

Natl. Acad. Sci. U. S. A. 78, 4222-4225 Enns, C. A., and Sussman, H. H. (1981) J. Biol. Chem. 256, 9820-

9823 Enns, C. A., and Reckhow, C. L. (1988) J. Biol. Chem. 263, 7297-

7301 Fredman, P., Richert, N. D., Magnani, J. L., Willingham, M. C.,

Pastan, I., and Ginsberg, V. (1983) J. Biol. Chem. 258, 11206- 11210

Fuhrmann, U., Bausc, E., Legler, G., and Ploegh, G. (1984) Nature 307,755-758

Gallis, B., Lewis, B., Wignall, H., Albert, A., Mochizuki, D. Y., Cosman, D., Hopp, T., and Urdal, D. (1986) J. Biol. Chem. 261, 5075-5080

Goding, J. W., and Harris, A. W. (1981) Proc. Natl. Acad. Sci. U. S. A. 78,4530-4534

Goldberg, D. E., Gabel, C. A., and Kornfeld, S. (1983) J. Biol. Chem. 97, 1700-1706

Hammarstrom, S., and Kabat, E. A. (1971) Biochemistry 10, 1684- 1692

Hounsell, E. F., Pickering, N. J., Stoll, M. S., Lawson, A. M., and Feizi, T. (1984) Biochem. Sot. Trans. 12, 607-610

Iyer, R. N., and Carlson, D. M. (1971) Arch. Biochem. Biophys. 142, 101-105

Jing, S., and Trowbridge, I. S. (1987) EMEO J. 6,327-331 Kingsley, D. M., Kozarsky, K. F., Hobbie, L., and Krieger, M. (1986)

Cell44,749-759 Kobata, A. (1983) in Biology of Carbohydrates (Ginburg, V., and

Robbins, P. W., eds) Vol. 2, pp. 87-161, John Wiley & Sons, New York

Kornfeld, S., Li, E., and Tabas, I. (1978) J. Biol. Chem. 253, 7771- 7778

Kozarsky, K. F., Kingsley, D., and Krieger, M. (1988a) Proc. Natl. Acad. Sci. U. S. A. 85,4335-4339

Kozarsky, K. F., Call, S. M., Dower, S. K., and Krieger, M. (1988b) Mol. Cell. Biol. 8, 3357-3363

Krusius, T., Finne, J., and Rauvala, H. (1976) FEBS Lett. 72, 117- 120

Laemmli, U. K. (1970) Nature 227, 680-685 Leonard, W. J., Depper, J. M., Crabtree, G. R., Rudikoff, S., and

Pumphrey, J. (1984) Nature 311,626-631 May, W. S., Jr., and Cuatrecasas, P. (1985) J. Membr. Biol. 88, 205-

215 McClelland, A., Kuhn, L. C., and Ruddle, F. H. (1984) Cell 39, 267-

274 McGraw, T. E., Greenfield. L., and Maxtield, F. R. (1987) J. Cell Biol.

105,207-214 Merkle. R. K.. and Cummines. R. D. (1987) Methods Enzvmol. 138.

232%i59 Neefjes, J. J., Verkerk, J. M. H., Broxterman, H. J. G., Van der

Marel. G. A.. van Boom. J. H.. and Ploesh. H. L. (1988) J. Biol. Chem; 107,+9-87 -

Ogata, S., Muramatsu, T., and Kobata, A. (1975) J. Biochem. (Tokyo) 78,687-696

Omary, M. B., and Trowbridge, I. S. (1981a) J. Biol. Chem. 256, 4715-4718

Omarv, M. B., and Trowbridge. I. S. (1981b) J. Biol. Chem. 256. 12888-12892

Parker. P. J.. Youne. S.. Gullick. W. J.. Maves. E. L. V.. Bennett. P.. and Water’field, I%. D: (1984) i. Biol: km. ‘259,9966-9912

Pathak, R. K., Merkle, R. K., Cummings, R. D., Goldstein, J. L., Brown, M. S., and Anderson, R. G. W. (1988) J. Biol. Chem. 106, 1831-1841

Poola, I., and Lucas, J. J. (1988) J. Biol. Chem. 263, 19137-19146 Rothenberger, S., Iacopetta, B. J., and Kuhn, L. C. (1987) Cell 49,

423-431 Russell, D. W., Schneider, W. J., Yamamoto, T., Luskey, K. L.,

Brown, M. S., and Goldstein, J. (1984) Cell 3?, 577-585 Sadler, J. E. (1984) in Biologv of Carbohydrates (Ginsburg, V.. and

Robbins, P.‘W., eds) Vol. <-pp. 199-288, John Wiley & Sons,‘New York

Schneider, C., Sutherland, R., Newman, R., and Greaves, M. (1982) J. Biol. Chem. 257, 8516-8522

Schneider, C.. Asser. U., Sutherland. D.. and Greaves, M. (1983) FEBS~ett. '158,259-i64

Schneider, C., Owen, M. J., Banville, D., and Williams, J. G. (1984) Nature3 11,675-678

Seligman, P. A., Schleicher, R. B., and Allen, R. H. (1979) J. Biol. Chem. 254,9943-9946

Sutherland, R., Delia, D., Schneider, C., Newman, R., Kemshead, J., and Greaves. M. (1981) Proc. Natl. Acad. Sci. U. S. A. 78. 4515- 4519 ’

Torres, B. V., McCrumb, D. K., and Smith D. F. (1988) Arch. Biochem. Biophys. 262, l-11

Trowbridge, I. S., Newman, R. A., Domingo, D. L., and Sauvage, C. (1984) Biochem. Phnrmacol. 33. 925-932

Turkewitz, A. P., Amatruda, J. F.; Borhani, D., Harrison, S. C., and Schwartz, A. L. (1988) J. Biol. Chem. 263,8318-8325

Varki, A., and Kornfeld, S. (1983) J. Biol. Chem. 258, 2808-2818 Wada, H. G., Hass, P. E., and Sussman, H. H. (1979) J. Biol. Chem.

254,12629-12635 Yoshimura, A., Yoshida, T., Seguchi, T., Waki, M., Ono, M., and

Kuwano, M. (1987) J. Biol. Chem. 262, 13299-13308 Zerial, M., Suomalainan, M., Zanetti-Schieder, M., Schneider, C., and

Garoff, H. (1987) EMBO J. 6,2661-2667


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

S I Do, C Enns and R D CummingsHuman transferrin receptor contains O-linked oligosaccharides.

1990, 265:114-125.J. Biol. Chem.

http://www.jbc.org/content/265/1/114Access the most updated version of this article at

Alerts:

When a correction for this article is posted•

When this article is cited•

to choose from all of JBC's e-mail alertsClick here

http://www.jbc.org/content/265/1/114.full.html#ref-list-1

This article cites 0 references, 0 of which can be accessed free at


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/content/265/1/114

http://www.jbc.org/cgi/alerts?alertType=citedby&addAlert=cited_by&cited_by_criteria_resid=jbc;265/1/114&saveAlert=no&return-type=article&return_url=http://www.jbc.org/content/265/1/114

http://www.jbc.org/cgi/alerts?alertType=correction&addAlert=correction&correction_criteria_value=265/1/114&saveAlert=no&return-type=article&return_url=http://www.jbc.org/content/265/1/114

http://www.jbc.org/cgi/alerts/etoc

http://www.jbc.org/content/265/1/114.full.html#ref-list-1

http://www.jbc.org/

human transferrin receptor contains o-linked ...this report describes the isolation and...

Documents