HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 277, Issue 46, 44431-44439, November 15, 2002

This Article Abstract Full Text (PDF) All Versions of this Article: 277/46/44431    most recent M205241200v1 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 Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Cheng, F. Articles by Fransson, L.-A. Search for Related Content PubMed PubMed Citation Articles by Cheng, F. Articles by Fransson, L.-A. Social Bookmarking                      What's this?

Nitric Oxide-dependent Processing of Heparan Sulfate in Recycling S-Nitrosylated Glypican-1 Takes Place in Caveolin-1-containing Endosomes^*

Fang Cheng, Katrin Mani, Jacob van den Born^§, Kan Ding^¶, Mattias Belting, and Lars-Åke Fransson

From the  Department of Cell and Molecular Biology, Lund University, BMC C13, SE-221 84, Lund, Sweden and the ^§ Department of Cell Biology, Free University of Amsterdam, Van der Boechorststraat 7, 1081 BT, Amsterdam, The Netherlands

Received for publication, May 28, 2002, and in revised form, August 28, 2002

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

We have previously demonstrated intracellular degradation of the heparan sulfate side chains in recycling glypican-1 by heparanase and by deaminative cleavage at N-unsubstituted glucosamine with nitric oxide derived from intrinsic nitrosothiols (see Ding, K., Mani, K., Cheng, F., Belting, M. and Fransson, L.-Å. (2002) J. Biol. Chem. 277, 33353-33360). To determine where and in what order events take place, we have visualized, by using confocal laser-scanning immunofluorescence microscopy, glypican-1 variants in unperturbed cells or arrested at various stages of processing. In unperturbed proliferating cells, glypican-1 was partly S-nitrosylated. Intracellular glypican-1 was enriched in endosomes, colocalized significantly with GM-1 ganglioside, caveolin-1, and Rab9-positive endosomes, and carried side chains rich in N-unsubstituted glucosamine residues. However, such residues were scarce in cell surface glypican-1. Brefeldin A-arrested glypican-1, which was non-S-nitrosylated and carried side chains rich in N-unsubstituted glucosamines, colocalized extensively with caveolin-1 but not with Rab9. Suramin, which inhibits heparanase, induced the appearance of S-nitrosylated glypican-1 in caveolin-1-rich compartments. Inhibition of deaminative cleavage did not prevent heparanase from generating heparan sulfate oligosaccharides that colocalized strongly with caveolin-1. Growth-quiescent cells displayed extensive NO-dependent deaminative cleavage of heparan sulfate-generating anhydromannose-terminating fragments that were partly associated with acidic vesicles. Proliferating cells generated such fragments during polyamine uptake. We conclude that recycling glypican-1 that is associated with caveolin-1-containing endosomes undergoes sequential N-desulfation/N-deacetylation, heparanase cleavage, S-nitrosylation, NO release, and deaminative cleavage of its side chains in conjunction with polyamine uptake.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Mammalian glypican-1 (Gpc-1)¹ is a member of a glycosylphosphatidylinositol (GPI)-linked cell-surface proteoglycan (PG) family with six known members to date. These PG, like other cell surface PG, are selective regulators of ligand-receptor encounters and thereby control growth and development (1-4). Gpc proteins are post-translationally modified by the addition of the glycosaminoglycan heparan sulfate (HS) at sites located close to the C-terminal GPI-membrane anchor (see Scheme 1). The central part of the protein consists of a cysteine-rich domain containing information that ensures a high level of HS substitution (5). Many of the functions of Gpc are dependent on the HS side chains, which are capable of binding and/or activating and/or transporting a variety of growth factors, cytokines, enzymes, viral proteins, and polyamines (6-11).

View larger version (21K): [in this window] [in a new window]   Scheme 1.   Schematic model of HS degradation in recycling Gpc-1. Position 1, Gpc-1 with three HS side chains carrying GlcNH residues mainly clustered near the core protein and sulfated uronic acids (HexUA-SO4) mainly in peripheral regions. Position 2, Gpc-1 after degradation by heparanase into HS-oligosaccharides and a core protein with truncated HS chains still carrying most of the GlcNH residues. Position 3, Gpc-1 after further degradation by NO/nitrite, giving rise to core protein with short HS stubs and HS-oligosaccharides terminating with anMan and containing sulfated hexuronic acid. Position 4, recycling Gpc-1 used for resynthesis of HS chains. Resynthesis can take place in the presence of cycloheximide or BFA, indicating that neither de novo synthesis of core protein nor delivery from the endoplasmic reticulum is necessary (18, 19).

GPI-anchored proteins are usually associated with sphingolipid- and cholesterol-rich plasma membrane domains. Such enriched domains may exist either as small phase-separated "rafts" or, when associated with caveolin-1 (Cav-1), form flask-shaped plasmalemmal invaginations called caveolae, which are involved in signal transduction and special forms of non-clathrin-dependent endocytosis mediated by Cav-1-containing endosomes, also called caveosomes (12-17).

Biochemical studies using radioactively labeled precursors have demonstrated recycling of newly made Gpc-1 in normal fibroblasts as well as in transformed cells (18, 19). During recycling, the HS side chains are degraded both by heparanase and by NO-dependent deaminative cleavage at N-unsubstituted glucosamine residues (GlcNH) (20). New HS chains can then be synthesized on the stubs remaining on the core protein (see Scheme 1). Biosynthesis of HS takes place in the Golgi and involves many interacting enzymes (21, 22). The stubs should first be extended with a GlcNAc-hexuronic acid (HexUA) repeat backbone. The GlcNH residues are either a result of inadequate sulfation during the regional exchange of N-acetyl for N-sulfate, or they are formed when synthesis is completed, by N-desulfation or further N-deacetylation.

Cells with up-regulated polyamine uptake synthesize Gpc-1 that carries HS chains with an increased number of GlcNH residues (23). NO depletion, which precludes cleavage at these residues, results in diminished uptake, suggesting a functional relationship (23). More recently, we have identified and isolated S-nitrosylated (SNO) forms of Gpc-1 and demonstrated that NO released from these SNO groups cleaves the HS chains in an autocatalytic process that involves a Cu²⁺/Cu⁺ redox cycle (24). Here, we have attempted to visualize and localize resident Gpc-1 variants intracellularly to obtain information about the order in which the various modifications take place, their location, the nature and fate of the degradation products, and how the processes correlate with spermine uptake.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Cell Treatments-- The cell line used throughout this study (human bladder carcinoma cells, T24) (25-27) was the same as that used previously (19) when it was marketed under the code name ECV304. Wild-type CHO cells (CHO-K1) and CHO cells deficient in HSPG synthesis (pgsD-677) were obtained from the American Type Culture Collection. Cells were seeded at a concentration of 20,000 cells/well in chamber slides with covers (LabTek) and cultured in minimal essential medium overnight to obtain sparse cultures and for 2-3 days to obtain confluent cultures. If not stated otherwise, experiments were performed with sparse cultures where cells were well separated from one another. Cells were left untreated or treated with trypsin (0.5 mg/ml for 2.5 min), BFA (10 µg/ml for 24 h), suramin (0.2 mM for 24 h), L-ascorbate (1 mM for 1 h), -difluoromethylornithine (DFMO; 5 mM for 2 days), with or without spermine (1 µM), with sodium nitroprusside (0.5 mM) and CuCl₂ (0.01 mM) for 1 h or were subjected to nitrite deprivation by treatment with 10 mM N-nitroarginine, 10 mM ammonium sulfamate, and 0.01 mM neocuproine for 24 h, all as described previously (19, 20, 23).

After the various treatments, cells were washed with phosphate-buffered saline three times and then fixed and permeabilized by treatments with acetone for 2-4 min followed by incubation with 1 ml of 2% (v/v) H₂O₂ in 60% (v/v) methanol for 15 min. N-Acetylations were carried out by adding acetic anhydride (1:10, v/v) to the methanol solution. After washing with water three times for 1 min each, cells were incubated with nonimmune serum (1:100 dilution) for 30 min at room temperature. Primary antibodies were applied as described by the manufacturers and kept for 3 h; cells were then washed three times with phosphate-buffered saline and exposed to secondary antibodies (1:500 dilution) for 2 h. Before microscopy, the slides were washed again with phosphate-buffered saline and air-dried. Several dilutions of the primary antibodies used for colocalization experiments were tested until equivalent signals had been obtained. Controls with preimmune serum or omission of either primary or secondary antibody were always run.

Labeling with Fluorescent Antibodies and Reagents-- The following primary antibodies were used: polyclonal anti-Gpc-1 (19), monoclonal S1 anti-Gpc-1 (28), monoclonal antibodies JM-403 (29) and 10E4 (Seikagaku Corp.) both against HS epitopes, polyclonal anti-Cav-1 (Transduction Laboratories), monoclonal anti-Rab9 (Affinity Bioreagents), FITC-labeled cholera toxin B (Sigma), polyclonal anti-S-nitrosocysteine (Alexis Biochemicals), and a monoclonal antibody specific for heparin oligosaccharides terminating with anMan or anhydromannitol (anManOH) (30). As secondary antibodies/reagents, we used either FITC-labeled or Texas Red-labeled goat anti-rabbit IgG (Molecular Probes, Inc., Eugene, OR) or FITC-labeled rabbit antimouse Ig from Sigma. The lysosomal marker LysoTracker Red DND-99 (Molecular Probes) was given to cells prior to permeabilization-fixation.

Microscopy-- Fluorescence images were obtained by using confocal laser-scanning equipment (MRC-1024; Bio-Rad) attached to a Nikon Eclipse E800 upright microscope. Excitation was obtained with an argon laser at 488 nm for FITC and at 560 nm for Texas Red, and the emitted light was filtered with an appropriate long pass filter (cut-off, 540 and 605 nm, respectively). If not indicated otherwise, the images shown were obtained at a focal plane that was at the center of the cell and of 0.3-0.5-µm thickness. Images were digitized and transferred to Adobe PhotoShop for merging, annotation, and printing.

Isolation of Radiolabeled Gpc-1-- Cells were maintained, preincubated with labeling medium, and radiolabeled using 20 µCi/ml D-[6-³H]glucosamine and/or 50 µCi/ml [³⁵S]sulfate as described (19, 20, 23). After treatments, cells were extracted with radioimmunoprecipitation buffer, and all Gpc-1 forms were immunoisolated using polyclonal anti-Gpc-1 antiserum (19). Charge variants of Gpc-1 were separated on MonoQ (24), and HS chains were released by alkali treatment and subjected to size fractionation on Superose 6 (19).

Isolation of Endosomes-- Endosomal preparations were obtained from homogenates of subconfluent cell cultures incubated with radiosulfate (19, 20, 23) by using a discontinuous sucrose gradient centrifugation technique (31). Fractions were analyzed for [³⁵S]Gpc-1 by immunoprecipitation (19).

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Characterization of Antisera and Antibodies-- We used a polyclonal antiserum directed against the Gpc-1 core protein (19) to detect all forms of Gpc-1 (i.e. with or without SNO groups, with or without HS; see Scheme 2). Two HS-specific monoclonal antibodies were tested, one of which (JM-403) (29) is directed against an epitope containing GlcNH. The HS epitope of the other monoclonal (10E4) has been proposed to comprise the sequence HexUA-GlcNH-HexUA-GlcNAc (32). N-Acetylation of free amino groups blocked the reactivity of JM-403 (Fig. 1A, JM-403 + N-Ac), confirming its specificity. As expected, the reactivity of the Gpc-1 protein was unaffected (Fig. 1A, GPC + N-Ac; cf. Fig. 1B, GPC). Unexpectedly, the 10E4 HS-epitope was also unaffected by N-acetylation (Fig. 1A, 10E4 + N-Ac), indicating that GlcNH was tolerated but not necessary for the antigenicity of 10E4. Therefore, in the present study, we have used JM-403 to trace GlcNH (see Scheme 2).

View larger version (28K): [in this window] [in a new window]   Scheme 2.   Structures in Gpc-1 and HS recognized by the various antisera and antibodies. The polyclonal anti-Gpc-1 antiserum reacts with the core protein, but some epitopes may be shielded by the very long, GlcNH-rich HS chains (20, 23). mAb S1 and polyclonal anti-SNO-cysteine antiserum recognize SNO groups and mAb JM-403 GlcNH-containing epitopes (29). Another mAb reacts with HS-oligosaccharides larger than tetrasaccharide and containing reducing terminal anMan (30). The nature of the HS epitope recognized by mAb 10E4 is unclear to us (see also Ref. 31). Only a minority of the human bladder carcinoma cells react, and, contrary to expectations, the proportion decreases markedly upon inhibition of polyamine synthesis (results not shown), although this is known to increase the GlcNH content (23).

View larger version (63K): [in this window] [in a new window]   Fig. 1.   Monoclonal antibody JM-403 but not 10E4 is specific for HS containing GlcNH, monoclonal antibody S1 recognizes S-nitrosylated Gpc-1, and S-nitrosylated Gpc-1 containing HS with GlcNH residues localize to perinuclear areas. A, confocal laser immunofluorescence staining for Gpc-1 (GPC) and HS (using either JM-403 or 10E4) after N-acetylation with acetic anhydride, showing that N-acetylation does not affect detection of Gpc-1 protein (GPC + N-Ac) but destroys the GlcNH-specific epitope (JM-403 + N-Ac), whereas the 10E4 epitope is not affected by N-acetylation and thus is not specific for GlcNH (10E4 + N-Ac). B, corresponding immunofluorescence staining for Gpc-1 using both polyclonal antiserum (GPC) and monoclonal S1 (GPC-S1) showing substantial colocalization (Merged). C, corresponding immunofluorescence staining for Gpc-1 in cells treated with ascorbate or ascorbate followed by NO donor (sodium nitroprusside (SNP)) and Cu2+ showing that reduction with ascorbate does not affect detection of Gpc-1 protein (GPC + Ascorbate) but destroys the S1 epitope (GPC-S1 + Ascorbate). The S1 epitope can be partly recovered by the addition of NO donor and Cu2+ (GPC-S1 + Ascorbate + SNP + CuCl2). D, corresponding immunofluorescence staining for S-nitrosylated cysteines (SNO-Cys) in the absence or presence of ascorbate and after recovery from ascorbate treatment in the presence of NO donor and Cu2+, showing that ascorbate likewise destroys S-nitrosylated cysteine epitopes (SNO-Cys + Ascorbate), which can subsequently be restored (SNO-Cys + Ascorbate + SNP + CuCl2). Recovery after ascorbate treatment but without added NO donor and Cu2+ resulted in a weak signal (as in the middle panel; data not shown). E, immunofluorescence staining for S-nitrosylated Gpc-1 using monoclonal S1 (GPC-S1) and polyclonal anti-S-nitrosocysteine (SNO-Cys), showing perinuclear colocalization (Merged). F, immunofluorescence staining for HS with the GlcNH-containing epitope (JM-403) and S-nitrosylated Gpc-1 (SNO-Cys), showing perinuclear colocalization (Merged). G, immunofluorescence staining for GlcNH-containing HS (JM-403) and total Gpc-1 (GPC) showing perinuclear colocalization (Merged). Bars, 20 µm.

The Gpc-1 core protein epitope recognized by the much used S1 monoclonal antibody has been reported to be sensitive to reduction with thiols (28). This has been assumed to indicate that the S1 epitope is located in the cysteine-rich and potentially disulfide-rich central domain of Gpc-1 and that reduction of disulfide bonds with thiols would destroy the epitope. However, immunolocalization of the S1 epitope and total Gpc-1 (Fig. 1B) suggested that S1 reacted with a subpopulation of the Gpc-1 molecules (yellow in Merged panel). It is known that reduction with thiols also releases NO from SNO groups (33). Hence, the S1 epitope may comprise SNO. Ascorbate can be used to release NO from SNO groups without cleaving disulfide bonds (33). We therefore tested whether the S1 epitope was sensitive to ascorbate. As shown in Fig. 1C, detection of total Gpc-1 protein was unaffected by ascorbate (GPC + Ascorbate), whereas the S1 epitope was destroyed (GPC-S1 + Ascorbate). Re-S-nitrosylation with NO is dependent on a Cu²⁺/Cu⁺ redox cycle (34). Accordingly, when NO donor and Cu²⁺ were provided after ascorbate treatment, the S1 epitope was partially restored (GPC-S1 + Ascorbate + SNP + CuCl₂). Corresponding results were obtained by using a polyclonal antiserum against SNO groups (Fig. 1D). Again, ascorbate reduced the signal (cf. SNO-Cys and SNO-Cys + Ascorbate), which could then be restored by re-S-nitrosylation using exogenous NO (SNO-Cys + Ascorbate + SNP + CuCl₂). Endogenous production of NO was insufficient to restore S-nitrosylation during the 1-h recovery period (results not shown). That S1 recognizes S-nitrosylated Gpc-1 was confirmed by a colocalization experiment using S1 and anti-SNO. The strongest colocalization was in the perinuclear region (Fig. 1E, Merged panel; see also Scheme 2).

Immunolocalization of S-Nitrosylated Gpc-1 and HS Rich in GlcNH-- Colocalization experiments using the GlcNH-recognizing JM-403 monoclonal antibody and anti-SNO indicated that a major population of the S-nitrosylated Gpc-1 molecules carries HS with GlcNH residues (Fig. 1F, yellow in Merged panel). This material was also present in compartments closely encircling the nucleus. GlcNH-containing HS (JM-403 epitope) that was not associated with S-nitrosylated proteins appeared to be present in a few scattered vesicles (Fig. 1F, green in Merged panel). These could either represent HS attached to non-S-nitrosylated Gpc-1 or HS-oligosaccharides (see further below) or HS attached to other core proteins. There was also separate SNO-positive material (Fig. 1 F, red in Merged panel) that should represent other S-nitrosylated proteins or peptides. Co-visualization of HS with GlcNH residues (JM-403) and total Gpc-1 (GPC) confirmed that the glycoform carrying such HS chains was largely present in compartments encircling the nucleus (Fig. 1G, yellow in Merged panel). Thus, a significant proportion of resident, intracellular Gpc-1 in subconfluent cells was both S-nitrosylated and carried HS rich in GlcNH residues (i.e. charged and ready for autocleavage of its HS chains) (24). In addition, the cells should also contain nonglycanated Gpc-1 (see Fig. 1G, red) and HS-oligosaccharides (see Fig. 1G, green) at steady state (19).

Colocalization of Gpc-1 with Cav-1, Rab9, and GM1 Ganglioside-- Since Gpc-1 is a GPI-anchored molecule, it may be associated with lipid rafts, caveolae, caveosomes, or endosomes transporting cargo from the plasma membrane to the Golgi. This may result in colocalization of Gpc-1 with Cav-1, Rab9, and GM1 ganglioside (14, 16, 35, 36). Results of colocalization experiments indicated that Gpc-1 substituted with GlcNH-containing HS chains (as in Fig. 1G) was partly associated with Cav-1 in vesicular structures diffusely encircling the nucleus (Fig. 2A). Using anti-Rab9 as a marker for late endosomes (35), strong scattered colocalization with Gpc-1 was seen in small vesicles surrounding the nucleus but also in larger vesicles located closer to the cell surface (Fig. 2B). Similarly, there was a marked colocalization between Rab9 and Cav-1 in the same area (Fig. 2C). FITC-labeled cholera toxin B was used to detect GM1-ganglioside. Colocalization between GM1 and Gpc-1 was strong in the middle section of the cell (yellow in Fig. 2D) and encircled the nucleus approximately in the same plane where Gpc-1 with GlcNH-containing HS colocalized with Cav-1 (Fig. 2A). Taken together, these results indicated that Gpc-1 carrying HS with GlcNH residues is partly present in vesicles that share characteristics with caveosomes or transporting endosomes (14, 35, 36).

View larger version (44K): [in this window] [in a new window]   Fig. 2.   Partial colocalizations of GlcNH-containing Gpc-1 or total Gpc-1 with Cav-1, Rab9, or GM1 ganglioside. A, confocal laser immunofluorescence staining for GlcNH-containing HS (JM-403) and Cav-1 (CAV), showing perinuclear colocalization. B, corresponding immunofluorescence staining for the endosome marker Rab9 (Endo) and Gpc-1 (GPC) showing scattered punctate colocalization around the nucleus and colocalization to larger vesicles at paranuclear sites. C, corresponding immunofluorescence staining for Rab9 and Cav-1, showing a similar pattern of colocalization. D, corresponding immunofluorescence staining for ganglioside GM1 using FITC-labeled cholera toxin B (CTxB) and Gpc-1, showing that GM1 colocalizes with Gpc-1 in compartments that encircle the nucleus (yellow). Bars, 20 µm.

We also immunoisolated [³⁵S]Gpc-1 from endosome preparations obtained from cells incubated with radiosulfate. We applied recently described procedures based on discontinuous sucrose density centrifugation of isotonic or hypotonic cell homogenates (31, 35). Expressed as the proportion of total ³⁵S-labeled macromolecules, there was a 5-15-fold enrichment of Gpc-1 in the endosomal fractions compared with total cellular Gpc-1. To identify the differently charged glycoforms of Gpc-1 (24) that were present in endosomes and in a total cell extract, we performed ion exchange chromatography on MonoQ (Fig. 3, A and B). The total cellular pool of Gpc-1 comprised almost equal proportions of the slightly negatively charged (i.e. relatively GlcNH-rich) (24) and the highly negatively charged (i.e. relatively GlcNH-poor) Gpc-1 glycoforms (Fig. 3A). The latter glycoforms were enriched in the endosomal pool (Fig. 3B), indicating that deaminative cleavage was taking place in this compartment.

View larger version (17K): [in this window] [in a new window]   Fig. 3.   Ion exchange chromatography of Gpc-1 (A and B) and gel exclusion chromatography of Gpc-1-derived HS chains (C and D). Metabolically radiosulfated Gpc-1 was immunoisolated from untreated (A and B) or from DFMO- and spermine-treated (C and D) cells and subcellular fractions. A, Gpc-1 from the total cell pool; B, Gpc-1 from the endosomal pool chromatographed on MonoQ. C, HS chains derived from Gpc-1 released by trypsinization of DFMO and spermine-treated cells. D, HS chains derived from Gpc-1 remaining in the cell layer after trypsinization. The HS chains were chromatographed on Superose 6 before (dotted line) and after cleavage (solid line) at GlcNH residues by treatment with HNO2 at pH 3.9 (20).

Gpc-1 at the Cell Surface-- The current opinion regarding the distribution of GPI-linked proteins between rafts and caveolae holds that most of the GPI-linked proteins are in small, dispersed rafts, but when raft components are oligomerized, rafts coalesce, and GPI-proteins may ultimately concentrate to caveolae and become internalized (37, 38). Furthermore, rafts can be so small that they are undetectable in light microscopy (39). Therefore, to determine whether Gpc-1 with HS chains containing GlcNH was present at the cell surface, we isolated metabolically radiosulfated cell surface Gpc-1. We used DFMO- and spermine-treated confluent cell cultures, since the GlcNH content increases under these conditions (23). Gpc-1 has a trypsin-sensitive site between the HS attachment site and the GPI anchor (40), permitting the release of almost intact surface-bound Gpc-1 molecules by mild trypsinization of the cells. The collected trypsinized cells were then extracted with detergent, and Gpc-1 was immunoisolated both from the trypsinate and the remaining cell pellet (20). Approximately one-third was obtained from the trypsinate, and two-thirds was from intracellular compartments. We then examined the HS chains from the two Gpc-1 pools for GlcNH residues (Fig. 3, C and D). Gel exclusion chromatography of HS chains before and after deaminative cleavage at the GlcNH residues showed greater degradation of the intracellular Gpc-1 HS chains (Fig. 3D) (i.e. indicating more GlcNH residues) than in the case of corresponding material from the cell surface (Fig. 3C). This supports the notion that the GlcNH residues are formed when biosynthesis is completed, either by N-desulfation or further N-deacetylation, and that it takes place during or after internalization.

To determine the order of ensuing modification and degradation steps, we examined the characteristics of Gpc-1 arrested at different stages of degradation and recycling (see Scheme 1). BFA was used to obtain the large Gpc-1 precursor, suramin, to impede heparanase degradation and NO depletion to prevent deaminative cleavage of HS.

BFA Causes Accumulation of Non-S-nitrosylated Gpc-1 Carrying HS with GlcNH Residues in Cav-1-rich Perinuclear Compartments-- BFA causes multiple disturbances of intracellular transport, including redistribution of Golgi components to the endoplasmic reticulum and inhibition of transport out of this mixed system, but also mixing of the trans-Golgi network with recycling endosomes, which can affect both recycling and transcytosis (41-43). Recycling of metabolically radiolabeled Gpc-1 is arrested by BFA at a stage where the HS chains have not yet become degraded and thus contain GlcNH residues and are of high molecular weight (19, 20) (see also Scheme 1).

Immunolocalization of Gpc-1 in BFA-treated cells (Fig. 4A) showed that Gpc-1 protein (GPC; red) and HS with GlcNH residues (JM-403; green) colocalized to an extensive semimerged compartment encircling the nucleus (yellow). Furthermore, very little BFA-arrested Gpc-1 colocalized with the Rab9 marker for late endosomes (results not shown). We then examined whether the BFA-arrested Gpc-1 was S-nitrosylated and whether it colocalized with Cav-1. The results showed (Fig. 4B) that BFA-arrested Gpc-1 was poorly S-nitrosylated (GPC-S1; green). However, using the JM-403 monoclonal antibody, which recognizes the GlcNH residues in HS, strong colocalization with Cav-1 was seen over the entire region where Gpc-1 was distributed (Fig. 4C, yellow). We can exclude the possibility that these results are due to colocalization between Cav-1 and GlcNH-containing HS-oligosaccharides, because degradation of HS is totally inhibited by BFA (19).

View larger version (97K): [in this window] [in a new window]   Fig. 4.   BFA causes accumulation of non-S-nitrosylated Gpc-1, and suramin (SUR) causes accumulation of S-nitrosylated Gpc-1, both in Cav-1-rich perinuclear compartments. A, confocal laser immunofluorescence staining for Gpc-1 (GPC) and GlcNH residues (JM-403) in BFA-treated cells showing strong colocalization to semimerged compartments encircling the nucleus. B, corresponding immunofluorescence staining for S-nitrosylated Gpc-1 (GPC-S1) and Cav-1 (CAV), showing weak colocalization caused by low content of S-nitrosylated Gpc-1. C, corresponding immunofluorescence staining for Gpc-1 carrying HS with GlcNH residues and Cav-1, showing strong colocalization to semimerged compartments encircling the nucleus. In addition, there is weak colocalization between Gpc-1 and Rab9 in BFA-treated cells (results not shown). D, corresponding immunofluorescence staining for Gpc-1 and GlcNH residues in suramin-treated cells showing less intense colocalization to compartments encircling the nucleus. E, corresponding immunofluorescence staining for S-nitrosylated Gpc-1 and Cav-1, showing colocalization in the perinuclear area. F, corresponding immunofluorescence staining for HS with GlcNH residues and Cav-1, showing strong colocalization to semimerged compartments encircling the nucleus similar to that seen in A and C. Bars, 20 µm.

Thus, these results indicated that BFA-arrested Gpc-1, carrying HS with GlcNH residues, was mainly present in extensive Cav-1-containing, perinuclear compartments that lacked marker for late endosomes. Most of the Gpc-1 molecules located in these sites had not yet become S-nitrosylated, indicating that GlcNH formation can precede S-nitrosylation.

Suramin Treatment Causes Accumulation of S-Nitrosylated Gpc-1 in Cav-1-rich Compartments-- Suramin is an inhibitor of heparanase, an endoglycosidase that cleaves HS at bonds linking GlcUA to GlcN located near the highly modified and thereby sulfate-rich regions (see Scheme 1 and Ref. 44). Heparanase can be present both at the cell surface and in intracellular endosomal compartments. Human heparanase shows a strong tendency to localize to perinuclear granules (45). It has been reported that cells take up suramin via caveolae and transport it both to endosomes and the trans-Golgi as well as to the nucleus (46, 47). Although degradation of HS by cell surface-located heparanase can be effectively inhibited by suramin (23), recycling Gpc-1 HS is only partially protected from degradation by suramin, because there is also an intracellular NO-dependent degradation (19, 20). It is possible that suramin also affects transport.

To localize and characterize Gpc-1 in suramin-treated cells, we used the same approach as for BFA-treated cells. The results indicated (Fig. 4D) that suramin-arrested Gpc-1 protein was also concentrated to compartments closely encircling the nucleus (red). Colocalization between Gpc-1 (GPC) and HS-containing GlcNH residues (JM-403) was decreased in suramin-treated cells (Fig. 4D, yellow) as compared with BFA-treated ones (cf. Fig. 4A), probably because of partial degradation of HS to oligosaccharides that had become separated from Gpc-1. In comparison with Gpc-1 in BFA-treated cells, the perinuclearly located Gpc-1 of suramin-treated cells reacted more strongly with the S1 monoclonal antibody, indicating the presence of S-nitrosylation (Fig. 4, compare B with E, yellow). Furthermore, both S-nitrosylated Gpc-1 (Fig. 4E, yellow) and the GlcNH-containing epitope JM-403 (Fig. 4F, yellow) colocalized with Cav-1 in suramin-treated cells, suggesting that suramin had also impeded transport. Colocalization between JM-403 and Cav-1 was particularly strong (Fig. 4F, yellow), suggesting that the HS degradation products were enriched in Cav-1-containing compartments. The results confirmed that S-nitrosylation precedes HS-degradation.

Inhibition of NO-dependent Cleavage of HS Also Results in Accumulation of HS-Oligosaccharides in Cav-1-rich Compartments-- Previous biochemical studies utilizing metabolic labeling indicated that depletion of NO/nitrite in confluent cell cultures, by (i) inhibition of NO-formation, (ii) inhibition of NO release from SNO, and (iii) quenching nitrite with sulfamate, increased the content of GlcNH residues in the HS chains of Gpc-1 (19, 20, 23). To localize Gpc-1 and HS in NO/nitrite-depleted subconfluent cell cultures, we used polyclonal anti-Gpc-1 (GPC, red) and the JM-403 monoclonal antibody (green) that recognizes the GlcNH-containing epitope in HS. The results showed that colocalization was minimal under these conditions (Fig. 5A, GPCJM-403, yellow), presumably due to cleavage of HS by heparanase and separation of Gpc-1 remnants from HS degradation products. Gpc-1 signal was present at separate sites both around the nucleus and at the periphery of the cell (Fig. 5A, GPCJM-403, red). HS-oligosaccharides that contained GlcNH were concentrated to perinuclear vesicles (Fig. 5A, GPCJM-403, green). The latter were Cav-1-rich (Fig. 5A, JM-403CAV, yellow). As expected, Gpc-1 from NO-depleted cells did not react with S1, indicating that it was not S-nitrosylated (results not shown). These results thus indicated that heparanase cleavage can precede NO-dependent deaminative cleavage (19, 20, 23), that it is independent of S-nitrosylation, and that it takes place in Cav-1-rich compartments.

View larger version (106K): [in this window] [in a new window]   Fig. 5.   Heparanase degradation precedes NO-dependent deaminative cleavage of HS, which generates anMan-terminating HS-oligosaccharides. A, confocal laser immunofluorescence staining for Gpc-1 (GPC, red) and GlcNH residues (JM-403, green) as well as for Cav-1 (CAV, red) in NO-depleted subconfluent cell cultures showing no colocalization of Gpc-1 and HS with GlcNH residues but colocalization (yellow) of Cav-1 and HS with GlcNH residues in the perinuclear area. B, corresponding staining for Gpc-1 (red) and anMan-containing HS-oligosaccharides (AM) in subconfluent cells showing no signal for anMan-containing HS fragments (no green). C, corresponding staining of confluent cells for Gpc-1 (red), GlcNH residues (green), anMan-containing HS-fragments (green), and with LysoTracker Red (LTR, red) showing, in GPC  JM-403, no colocalization of GlcNH (JM-403, green) with Gpc-1 (GPC, red) but a few separate sites with GlcNH-positive material (green), the presence, in GPC  AM, of anMan-containing HS fragments in separate vesicles (AM, green) but also colocalization with Gpc-1 (yellow) and also, in LTR  AM, colocalization with LysoTracker Red in acidic vesicles (yellow) but also presence at separate non-acidic sites (green). Bars, 20 µm.

Quiescent Growth State Increases NO-dependent Cleavage of HS and Generates anMan-containing Oligosaccharides-- We suspected that the GlcNH content was higher in HS of intracellular Gpc-1 from subconfluent cells (Fig. 1F, JM-403, green) than in Gpc-1 HS from confluent cultures (20). A lower content of GlcNH could be the result of diminished formation or of increased destruction by NO-dependent deaminative cleavage as cells become quiescent. Deaminative cleavage of HS generates oligosaccharides or chain fragments with reducing terminal anMan. We therefore examined both subconfluent and confluent cultures by using the polyclonal anti-Gpc-1 (red) and the monoclonal JM-403 (green) antibodies as well as another monoclonal antibody (also green) specific for HS degradation products terminating with anMan or anManOH (30) (see also Scheme 2).

As shown in Fig. 5B, there was no reactivity with the anMan/anManOH-specific antibody in subconfluent cells (no green), in accordance with the high content of GlcNH in the HS of Gpc-1 from these cells (Fig. 1F, JM-403, green). In contrast, confluent cell cultures contained very little GlcNH that colocalized with intracellular Gpc-1 (Fig. 5C, GPCJM-403, no yellow; cf. Fig. 1F, JM-403, yellow in Merged panel). However, a few separate vesicles containing free HS with GlcNH were seen near the nuclei (Fig. 5C, GPCJM-403, green). These could be HS-oligosaccharides generated by heparanase degradation.

anMan-containing oligosaccharides were detected in confluent cells (Fig. 5C, GPCAM, green) indicating an increased deaminative cleavage at GlcNH of HS. The oligosaccharides were located in vesicles that were both separate from Gpc-1 (green) and colocalizing with it (yellow). Both types of vesicles were mostly located perinuclearly. Some of the anMan-containing oligosaccharides colocalized with LysoTracker Red (Fig. 5C, LTRAM, yellow), indicating that they were in acidic compartments. Others were in nonacidic vesicles (Fig. 5C, LTRAM, green). Colocalization experiments using markers for endosomes (Rab9) and LysoTracker Red indicated that some of the endosomes were also acidic (results not shown). Due to crowding effects in confluent cultures, the exact localization of these vesicles was somewhat unclear.

To confirm that the material reacting with the anMan-specific antibodies was derived from HS, we also examined wild-type CHO cells and mutant CHO cells deficient in HS synthesis. anMan-positive material, accumulating in acidic vesicles, was formed in confluent wild-type CHO cells but not in mutant cells (results not shown). We also confirmed that the formation of these degradation products was NO-dependent by examining NO-depleted confluent cells. No anMan-positive material was detected in this case (results not shown). We thus conclude that endogenously formed NO can generate anMan-containing HS-oligosaccharides in live cells.

Polyamine Uptake by Proliferating Cells Correlates with NO-dependent Cleavage of HS-- Previous biochemical studies have shown that uptake of the polyamine spermine, in cells treated with DFMO to inhibit endogenous polyamine synthesis, is correlated with NO/nitrite-dependent degradation of HS at the GlcNH residues (23). Immunolocalization experiments indicated that DFMO-treated cells concentrated Gpc-1 even more to the perinuclear area during spermine uptake (results not shown). Simultaneously, the GlcNH-containing epitope disappeared (cf. Fig. 6, JM-403 and JM-403 + spm). This effect was dependent on NO/nitrite formation (Fig. 6, JM-403 + spm + lowNO), indicating that the GlcNH residues were destroyed by deaminative cleavage of HS. We therefore searched for anMan-containing degradation products. The disappearance of GlcNH residues correlated with the formation of anMan-containing HS-oligosaccharides (Fig. 6, GPCAM + SPM, green). Hence, during spermine uptake, also subconfluent cells have an increased NO-dependent deaminative cleavage of HS at GlcNH, generating HS fragments terminating with anMan (see Scheme 1). Since the signal from anMan-containing material was weaker in polyamine uptake-dependent, subconfluent cells (Fig. 6, GPCAM + SPM, green) than in confluent unperturbed cells (Fig. 5C, GPCAM, green), some of the anMan-containing products may have been consumed by further processing. It should be noted that anMan-containing fragments shorter than tetrasaccharide may not be recognized by the antibody (30).

View larger version (162K): [in this window] [in a new window]   Fig. 6.   NO-dependent deaminative cleavage of HS is augmented by spermine uptake in subconfluent cells. Confocal laser immunofluorescence staining of DFMO-treated subconfluent cells for HS epitope containing GlcNH (JM-403, green), Gpc-1 (GPC, red), and anMan-containing HS fragments (AM, green) in the absence or presence of spermine (+SPM) and in nitrite-deprived cells in the presence of spermine (+SPM +low NO), showing that the GlcNH-containing epitope disappears when spermine uptake takes place (JM-403+spm) and partly returns when spermine uptake is inhibited by nitrite deprivation (JM-403 + SPM + low NO) and showing formation of anMan-terminating oligosaccharides (green) in subconfluent cells actively engaged in spermine uptake (GPCAM + SPM). Bar, 20 µm.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

PG are usually components of the extracellular or pericellular matrix. Here we show that Gpc-1 is primarily located intracellularly in both subconfluent and confluent human bladder carcinoma cells and that Gpc-1 is another example of a protein that can become functionally modified by S-nitrosylation (48). We also show that formation of GlcNH residues at specific sites in the HS side chains takes place before S-nitrosylation. Subsequent deaminative cleavage of HS at these sites is caused by NO derived from the intrinsic SNO groups (24). The order in which the various modification and processing events take place was determined by arresting Gpc-1 at various stages by using either BFA, suramin, or NO depletion.

On the basis of the results presented, we propose the working model shown in Scheme 3 for the routing of recycling Gpc-1. Newly synthesized Gpc-1 associated with the cell surface was relatively poor in GlcNH residues. Hence, most GlcNH residues were generated during internalization. In Drosophila, glypicans (Dally and Dally-like) are modified by a secreted enzyme encoded by a gene called notum (49). The structure of this protein strongly suggests that it catalyzes an N-deacetylation of HS.

View larger version (36K): [in this window] [in a new window]   Scheme 3.   Proposed schematic model for Gpc-1 recycling. Gpc-1 (top left) with mature HS side chains is a transient resident at the cell surface. When taken up by endocytosis (top right), formation of GlcNH residues begins. In endosomes (bottom right), heparanase degrades the HS chains. HS fragments may eventually exit to lysosomes. After transfer to a perinuclear compartment (bottom left), SNO of cysteine residues in the core protein with NO derived from Arg should be completed. This enables Gpc-1 to cut off terminal HS-oligosaccharides from its HS chain stubs by deaminative cleavage at the GlcNH residues (24). The HS split products carrying reducing terminal anMan may exit by vesicular transport. The Gpc-1 with truncated HS chains (bottom left) returns to the Golgi for re-elongation of the HS chains.

During the proposed transfer of Gpc-1 from rafts or caveolae to caveosomes or endosomes (Scheme 3), heparanase degrades the HS chains generating HS-oligosaccharides associated with Cav-1-containing compartments. Simultaneously or thereafter, the core protein of Gpc-1 is progressively S-nitrosylated with NO derived from arginine. Then, by using NO released from the intrinsic SNO groups, anMan-containing HS-oligosaccharides are generated by deaminative cleavage at the GlcNH residues. The triggering mechanism for NO release remains unknown, but both S-nitrosylation and NO release are dependent on a Cu²⁺/Cu⁺ redox cycle (34). The redox state of the cell may therefore regulate NO release. It is intriguing that late endosomes, or a subcompartment thereof, exhibit significant reducing activity (50), which could support NO release from SNO groups.

The HS degradation products eventually segregate into numerous cytoplasmic vesicles, some of which are acidic and Rab9-positive, suggesting that they may be transporting endosomes. Results of recent studies support the existence of an endosomal recycling route connecting non-clathrin-dependent endocytosis with the trans-Golgi network (35, 36). Structures defined as endosomes through their involvement in fluid phase endocytosis contain Rab9 as well as Cav-1 and can transport cargo from the plasma membrane to the Golgi.

When GlcNH formation and spermine uptake are up-regulated or when cells become growth-quiescent, an increased number of anMan-terminating oligosaccharides are generated by NO-dependent degradation of HS. Further processing of such oligosaccharides by lysosomal exodegradation would, in the end, result in the formation of free anMan and subsequently conversion to anManOH by nonspecific aldose reductases (51). As discussed previously (20), endogenous generation of free anManOH has never been demonstrated, but exogenously supplied anManOH can be converted to both anManOH-1-phosphate and anManOH 1,6-bisphosphate, which inhibit gluconeogenesis and glycogenolysis and stimulate glycolysis.

Since most of the resident PG forms of Gpc-1 were associated with intracellular organelles, Gpc-1 may only be transiently located at the cell surface. The dimensions of a Gpc-1 molecule with three radially extended HS chains (see Scheme 1) should be greater than that of a raft (39, 52). Furthermore, HS chains of HSPG can self-associate (53), which would promote cross-linking and subsequent uptake via caveolae. The contact zones for self-association are located in the mixed copolymeric transition regions embracing the highly modified domains in HS (54). As the GlcNH residues are concentrated in the same HS-regions (20, 23), they may strengthen self-association by providing opportunities for local electrostatic attraction.

Previous studies have shown that exogenously supplied HS can be taken up by various cell types and be targeted to the nucleus (6, 55, 56). Uptake of another polyanion, hyaluronan, via non-clathrin-dependent endocytosis followed by lysosomal targeting has also been described (57). Moreover, endogenous Gpc has been located to the nuclei of nervous tissue cells (58), and HSPG displays nuclear targeting in fibroblasts grown on fibronectin (59). Although we occasionally observed Gpc-1 protein inside the nucleus, they represented a small proportion and did not appear to be HS-substituted. No anMan-positive HS-oligosaccharides appeared to be targeted to the nucleus, which is thus in contrast to exogenously supplied HS. Usually, it is only a small fraction (<1%) of exogenously supplied HS that is taken up by cells and transported to the nucleus (56). However, both the rate of uptake and the amount of HS accumulating in the nuclei can be severalfold enhanced when HS is combined with a polyamine-lipid conjugate (6). Internalization of a ligand along with the PG can similarly redirect sorting of HS degradation products from endosomes to the nucleus (60).

Regulation of polyamine uptake constitutes a way to control cell growth. By inhibition of both polyamine and HSPG biosynthesis, tumor growth can be attenuated (10). When endogenous polyamine synthesis is inhibited, the HS side chains of recycling Gpc-1 increase their polyzwitterionic character, which results both in greater polyamine affinity and in greater sensitivity to NO-dependent degradation (23). These properties make Gpc-1 a versatile transport vehicle. It has been shown that heparin-binding, secretory phospholipases (type A₂) bind to HS of Gpc and can be translocated into perinuclear compartments (61). Gpc-1 may thus carry many different types of molecules to sites where HS is degraded by NO, whereupon degradation products and cargo separates. Some of the cargo molecules may then support membrane penetration. The mechanism for nuclear targeting may be common to many different types of biopolyelectrolytes, including cationic polyamines and peptides, and anionic glycosaminoglycans and nucleic acids.

	ACKNOWLEDGEMENTS

We thank Prof. Catharina Svanborg and her staff at the Department of Microbiology, Immunology, and Glycobiology, Lund University, for support and for use of microscope facilities and Prof. Guido David (Department of Human Genetics, University of Leuven, Belgium) and Dr. Gunnar Pejler (Department of Biochemistry, Swedish Agricultural University, Uppsala, Sweden) for generous gifts of monoclonal antibodies. The technical assistance of Birgitta Havsmark and Susanne Persson is greatly appreciated.

	FOOTNOTES

^* This work was supported by grants from the Swedish Science Council (VR-M and VR-NaTe), the Cancer Fund, the Strategic Research Fund (Glycoconjugates in Biological Systems (to F. C.)), the WennerGren, Kock, and Österlund Foundations, Xylogen AB, and the Medical Faculty of Lund University.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.

^¶ Present address: Dept. of Developmental and Cell Biology, School of Biological Sciences, University of California at Irvine, Irvine CA 92697-2300.

To whom correspondence should be addressed: Dept. of Cell and Molecular Biology, Lund University, BMC C13, SE-221 84, Lund, Sweden. Tel.: 46-46-222-8573; Fax: 46-46-222-3128; E-mail: lars-ake.fransson@medkem.lu.se.

Published, JBC Papers in Press, September 10, 2002, DOI 10.1074/jbc.M205241200

	ABBREVIATIONS

The abbreviations used are: Gpc-1, glypican-1; anMan, anhydro-D-mannose; anManOH, anhydro-D-mannitol; BFA, brefeldin A; Cav-1, caveolin-1; DFMO, -difluoromethylornithine; GlcN, glucosamine with unspecified N-substituent; GlcNH, glucosamine with free amino group; GPI, glycosylphosphatidylinositol; HexUA, hexuronic acid; HS, heparan sulfate; NO, nitric oxide; PG, proteoglycan(s); SNO, S-nitroso group; CHO, Chinese hamster ovary; HSPG heparan sulfate proteoglycan, FITC, fluorescein isothiocyanate.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES