Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
THE JOURNAL OF BIOLOGICAL CHEMISTRY © 2003 by The American Society for Biochemistry and Molecular Biology, Inc. Vol. 278, No. 47, Issue of November 21, pp. 47181–47189, 2003 Printed in U.S.A. Glypican-1 Is a Vehicle for Polyamine Uptake in Mammalian Cells A PIVOTAL ROLE FOR NITROSOTHIOL-DERIVED NITRIC OXIDE* Received for publication, July 30, 2003 Published, JBC Papers in Press, September 11, 2003, DOI 10.1074/jbc.M308325200 Mattias Belting‡§¶, Katrin Mani‡§, Mats Jönsson‡, Fang Cheng‡, Staffan Sandgren‡, Susanne Jonsson‡, Kan Ding‡储, Jean-Guy Delcros**, and Lars-Åke Fransson‡ ‡‡ From the ‡Department of Cell and Molecular Biology, Lund University, SE-221 84 Lund, Sweden and the **Groupe de Recherche en Thérapeutique Anticancéreuse, CNRS FRE22– 61, Faculté de Médecine, Université de Rennes 1, 2 avenue du Professeur Léon Bernard, F-35043 Rennes, France Polyamines (putrescine, spermidine, and spermine) are essential for growth and survival of all cells. When polyamine biosynthesis is inhibited, there is up-regulation of import. The mammalian polyamine transport system is unknown. We have previously shown that the heparan sulfate (HS) side chains of recycling glypican-1 (Gpc-1) can sequester spermine, that intracellular polyamine depletion increases the number of NO-sensitive N-unsubstituted glucosamines in HS, and that NO-dependent cleavage of HS at these sites is required for spermine uptake. The NO is derived from S-nitroso groups in the Gpc-1 protein. Using RNA interference technology as well as biochemical and microscopic techniques applied to both normal and uptake-deficient cells, we demonstrate that inhibition of Gpc-1 expression abrogates spermine uptake and intracellular delivery. In unperturbed cells, spermine and recycling Gpc-1 carrying HS chains rich in Nunsubstituted glucosamines were co-localized. By exposing cells to ascorbate, we induced release of NO from the S-nitroso groups, resulting in HS degradation and unloading of the sequestered polyamines as well as nuclear targeting of the deglycanated Gpc-1 protein. Polyamine uptake-deficient cells appear to have a defect in the NO release mechanism. We have managed to restore spermine uptake partially in these cells by providing spermine NONOate and ascorbate. The former bound to the HS chains of recycling Gpc-1 and S-nitrosylated the core protein. Ascorbate released NO, which degraded HS and liberated the bound spermine. Recycling HS proteoglycans of the glypican-type may be plasma membrane carriers for cargo taken up by caveolar endocytosis. Polyamines (putrescine, spermidine, and spermine) are essential for growth, differentiation, and survival of prokaryotic * This work was supported by Swedish Science Council Grants VR-M and VR-NaTe and by grants from the Cancer Fund; the Strategic Research Fund (Glycoconjugates in Biological Systems) (to F. C.); Polysackaridforskning i Uppsala AB; the WennerGren, Tegger, 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. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. § Both authors contributed equally to this work. ¶ Present address: Depts. of Immunology and Vascular Biology, Scripps Research Inst., La Jolla, CA 92014. 储 Present address: Dept. of Developmental and Cell Biology, School of Biological Sciences, University of California, 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. [email protected]. This paper is available on line at http://www.jbc.org as well as eukaryotic cells. The polyamine content of cells is tightly regulated by biosynthesis, degradation, and transport. For example, when the key enzyme in polyamine biosynthesis, ornithine decarboxylase, is inhibited by ␣-difluoromethylornithine (DFMO),1 polyamine import is substantially up-regulated. Cells (even uptake-deficient cells) can also secrete polyamines (for reviews, see Refs. 1– 8). Polyamine transport proteins and their genes have been identified in prokaryotes (9). During transport, the cationic polyamines bind electrostatically to acidic amino acids in the transporters. Attempts to isolate and identify mammalian polyamine transport genes and their encoded proteins have so far been unsuccessful. Mutant Chinese hamster ovary (CHO) cell lines deficient in polyamine uptake can be generated by polyamine analogs such as methylglyoxal bis(guanylhydrazone) (MGBG). This substance, which is taken up via the polyamine transport system, is toxic to cells (10). Hence, only uptake-deficient cells will survive in MGBG-containing medium. By transfecting uptake-deficient cells with genomic DNA from normal cells, Pegg and co-workers (11) isolated a number of different positive clones with restored capacity for polyamine uptake. They concluded that a number of proteins may be needed for polyamine transport in mammalian cells, some of which may be involved in cell proliferation. Eukaryotic cells from multicellular organisms produce a multitude of cell-surface proteins that interact with soluble ligands, other cells, or extracellular matrix components. Cellsurface proteins that belong to the proteoglycan (PG) family are substituted with long, linear, highly sulfated glycosaminoglycan chains. Unicellular eukaryotes such as yeast do not produce such polyanionic proteoglycans. Previous studies from this laboratory have shown that glycosaminoglycan chains, especially heparan sulfate (HS), bind polyamines very strongly, e.g. the affinity of HS for spermine is 10-fold higher than that of DNA (12), which is regarded as a major natural polyamine-binding polyanion (1, 7, 8). Subsequently, we showed that exogenous HS competitively inhibits polyamine uptake and that PG-deficient or sulfate-depleted cells have a diminished uptake (13, 14). PG-deficient cells are also more sensitive to growth inhibition by DFMO both in vitro and in vivo (14). We have also shown that polyamine-depleted cells synthesize glypican HS chains with an increased spermine affinity (15). 1 The abbreviations used are: DFMO, ␣-difluoromethylornithine; CHO, Chinese hamster ovary; MGBG, methylglyoxal bis(guanylhydrazone); PG, proteoglycan; HS, heparan sulfate; HSPG, heparan sulfate proteoglycan(s); Gpc-1, glypican-1; GlcNH3⫹, glucosamine with a free amino group; siRNA, small interfering RNA; SNO, S-nitroso; mAb, monoclonal antibody; BFA, brefeldin A; NOS, nitric-oxide synthase. 47181 47182 Glypican-1 and Polyamine Uptake Glypicans constitute a subfamily of glycosylphosphatidylinositol-anchored cell-surface heparan sulfate proteoglycans (HSPG). Potentially localizing to rafts and caveolae, they are selective regulators of ligand-receptor encounters and can thereby control both growth and development. The discoveries that mutations in genes involved in glypican assembly cause dysmorphic syndromes in man and aberrant patterning during Drosophila development have brought this family of molecules into focus (for review, see Ref. 16). We have recently shown that the glypican-1 (Gpc-1) core protein can accumulate NO as nitrosothiols in a reaction that is dependent on reduction of Cu2⫹ to Cu⫹ (17). Moreover, the HS chains of Gpc-1 contain a small amount of glucosamine units with an unsubstituted amino group (GlcNH3⫹), the number of which increases markedly upon inhibition of polyamine synthesis (15). It has been known for a long time that the GlcNH3⫹ units in HS chains are especially sensitive to deaminative cleavage by nitrite (18, 19). Nitrite can be derived from cellular NO under physiological conditions (20), but also NO gas can be used to degrade HS (21). The results of our studies show that when purified S-nitrosylated Gpc-1 substituted with GlcNH3⫹-rich HS chains is exposed in vitro to a reducing agent such as ascorbate, NO is released. Subsequently, NO degrades the Gpc-1 HS chains at the GlcNH3⫹ units in a reaction that results in conversion of GlcNH3⫹ to reducing terminal anhydromannose in the liberated HS oligosaccharide fragments (17). This autocleavage can also be demonstrated in caveolin-1-positive endosomes of cultured cells (22). On the basis of these results and the finding that NO depletion suppresses polyamine uptake (15), we have hypothesized that recycling Gpc-1 is a vehicle for polyamine uptake in mammalian cells (see Scheme 1). The model proposes that the HS chains bind polyamines by cooperative salt bridges between -NH3⫹ or ⬎NH2⫹ in the polyamines and the negative sulfate groups in HS (see Scheme 1, Position I). Polyamine depletion induces, by an unknown mechanism, an increased N-desulfation and/or N-deacetylation of the HS chains in recycling Gpc-1, generating an increased number of GlcNH3⫹ units, also within the polyamine-binding sites. Simultaneously, the Gpc-1 protein becomes S-nitrosylated in a copper-dependent reaction (see Scheme 1, Position II). When the recycling glypican-polyamine complex reaches late endosomes or caveosomes (see Scheme 1, Position III), NO is released; the polyamine-binding sites in HS are cleaved; the cooperative electrostatic interaction between HS and polyamines is diminished; and polyamines are liberated. In this study, using small interfering RNAs (siRNAs) designed from the Gpc-1 gene, we have shown that Gpc-1 is required for polyamine uptake and that spermine and recycling Gpc-1 are co-localized. Furthermore, we have shown that HS degradation by NO-induced deaminative cleavage is defective in uptake-deficient cells. We have managed to restore polyamine uptake partially in these cells by providing NO in the appropriate context. The results point to a pivotal role for NO, derived from S-nitroso (SNO) groups in Gpc-1, in supporting polyamine uptake. EXPERIMENTAL PROCEDURES Materials—Human bladder carcinoma T24 cells, culture media, sera, antisera to human and mouse Gpc-1 and SNO-cysteine, monoclonal antibody (mAb) JM-403 (recognizing GlcNH3⫹), mAb S1 (recognizing Gpc-1), brefeldin A (BFA), radioactive compounds, enzymes, prepacked columns, and certain chemicals were generated or obtained as described previously (13, 15, 17). RAW 264.7 cells and wild-type CHO-K1 cells were obtained from American Type Culture Collection. Mutant polyamine uptake-deficient CHO-MGBG cells were a gift from Prof. Lo Persson (Department of Physiology, Lund University). A cDNA plasmid encoding mouse Gpc-1 was a gift from Prof. Guido David (University of Leuven, Leuven, Belgium). Antiserum to mouse Gpc-1 was raised against a recombinant protein comprising amino acids 113–558 using a corresponding procedure as described previously (23). DFMO was from ILEX Oncology (San Antonio, TX); iminoacetic acid was from Fluka; and spermine NONOate and anti-nitric-oxide synthase (NOS) mAb were from Calbiochem. siRNA Preparation and Transfection—The vector pSilencer2.0-U6 (Ambion Inc., Austin, TX) containing the sequence GCTGGTCTACTGTGCTCAC (corresponding to nucleotides 977–995 in human Gpc-1) followed by the hairpin sequence TTCAAGAGA and then the reversed complementary Gpc-1 sequence with an additional C in the 5⬘-end and a stretch of six Ts for RNA polymerase III termination followed by GGAA in the 3⬘-end was synthesized by Genscript Corp. A negative control vector comprising a scrambled sequence was also prepared. Transfection was accomplished using LipofectAMINE (Invitrogen) following the recommendations of the manufacturer. Cell Treatments and Isolation of Gpc-1—RAW 264.7, T24, and CHO cells were maintained in the appropriate media as described by the manufacturer and treated with DFMO (5 mM), spermine (1 M), BFA (10 g/ml), or ascorbate (1 mM) as described previously (15, 17). RAW 264.7 cells were stimulated with Escherichia coli lipopolysaccharide (3.5 ng/ml; Sigma) and ␥-interferon (5 units/ml; Novakemi AB) for 24 h to increase production of NO. Radiolabeling was performed with D[6-3H]glucosamine, [35S]sulfate, or L-[35S]methionine, and Gpc-1 was obtained by immunoisolation (23). Separation, Degradation, and Modification Methods—Separations according to size were performed on Superose 6 and according to charge on Mono Q (15, 17). Core protein preparations of HSPG were obtained by treatment with HS lyase and analyzed by SDS-PAGE (23). HS chains were released by alkaline elimination in borohydride and cleaved at the GlcNH3⫹ residues by HNO2 at pH 3.9 (15, 17). N-Acetylation was achieved using acetic anhydride. To test for heparanase activity, purified radiolabeled Gpc-1 was incubated with cultures of CHO-K1 or CHO-MGBG cells, and degradation products were recovered and analyzed on Superose 6 as described (15). Confocal Immunofluorescence Microscopy—Fixation and permeabilization of cells, treatment with primary and secondary antibodies, and confocal microscopy were carried out as described (17, 22). The secondary antibodies used were goat anti-mouse total Ig (when the primary antibody was monoclonal) and goat anti-rabbit IgG (when the primary antibody was polyclonal). They were tagged with either fluorescein isothiocyanate or Texas Red and appropriately combined for co-localization studies. Detection of NOS—NOS activity was estimated in cell homogenates using the arginine-to-citrulline conversion procedure (24). NOS protein was detected by anti-endothelial NOS mAb (1:250 dilution) and biotinconjugated anti-mouse IgG (1:500) as the secondary antibody followed by staining with horseradish peroxidase-linked streptavidin and visualized by light microscopy as described (25). Spermine Uptake and Detection of Free Spermine—Cells were grown with or without 5 mM DFMO for 24 h. Uptake of [14C]spermine was measured as described (13). Free spermine was detected by mAb Spm8-2 (26) in confocal microscopic experiments. RESULTS Involvement of Gpc-1 in Spermine Uptake by T24 Cells— Involvement of HSPG in polyamine uptake is indirectly supported by the previous findings that (a) HS binds the polyamine spermine very strongly; (b) HS competitively inhibits spermine uptake; and (c) HSPG-deficient cells have a low spermine uptake (12–14). Involvement of glypican-type HSPG is suggested by the observations that (d) polyamine deprivation increases the GlcNH3⫹ content in Gpc-1 HS chains (15); (e) HS chains in Gpc-1 are cleaved at these GlcNH3⫹ units by NO in conjunction with spermine uptake (22); and (f) NO depletion inhibits spermine uptake in DFMO-treated T24 cells (15). We therefore examined T24 cells for co-localization of Gpc-1 and spermine by confocal immunofluorescence microscopy using anti-Gpc-1 polyclonal antibody and mAb Spm8-2, which recognizes free spermine and spermidine (26). Polyamines bound to DNA or RNA or to HS chains in recycling Gpc-1 should not be accessible to the antibody and thus should be undetectable. Structural modulation of HS generating a lower negative charge density and/or cleavage of HS is expected to diminish HS-polyamine interaction, resulting in release of polyamines. As long as the free polyamines remain in the same Glypican-1 and Polyamine Uptake 47183 FIG. 1. Immunolocalization of spermine and Gpc-1 in unperturbed T24 cells (A), in cells exposed to spermine (B) or ascorbate (C) for 1 h, in DFMO-treated cells (D), in DFMOtreated cells exposed to spermine (E), or in DFMO-treated cells exposed first to spermine and then to ascorbate for 10 min (F). Confocal laser immunofluorescence staining for Gpc-1 (red) was performed using anti-human Gpc-1 polyclonal serum and for polyamines (green) using mAb Spm8-2. Scale bar ⫽ 20 m. vesicles as Gpc-1, co-localization should be observed. The results show that, in unperturbed T24 cells, there was very little free polyamine, most of which was scattered in the cytoplasm (Fig. 1A, green). Gpc-1 (red) was in perinuclear compartments, partially co-localizing with polyamines. Addition of exogenous spermine to the cells did not significantly affect co-localization (Fig. 1B). A 1-h incubation of the cells with ascorbate, which induces release of NO from SNO groups and subsequent deaminative cleavage of HS in Gpc-1 (17), resulted in strong colocalization of Gpc-1 and polyamines to large, sometimes almost coalescent, vesicles near the nucleus (Fig. 1C, yellow). In some cells (Fig. 1C, left), no free polyamine could be seen (no green), whereas in others (right), free polyamines were seen peripheral to the co-localized material. We interpret these results to indicate that ascorbate had induced release of polyamines by NO-dependent degradation of HS chains in intracellular Gpc-1 and that, in some cells, the polyamines had not yet been transported out of the endosomal compartment. Treatment of cells with DFMO results in increased formation of GlcNH3⫹ units in HS of Gpc-1 (15). In DFMO-treated T24 cells, there was an increase in free polyamines in the cytoplasm (Fig. 1D, green), but also an increased co-localization between Gpc-1 and polyamines in paranuclear vesicles (Fig. 1D, yellow). We interpret this result to indicate that the released free polyamines were derived from a pool of polyamines previously bound to HS of recycling intracellular Gpc-1. They had been released because the affinity of the polyamines for the GlcNH3⫹enriched HS of Gpc-1 had been diminished. In keeping with these results, DFMO treatment is known to increase both the Vmax and Km for polyamine binding to the putative transporter, which means increased capacity for polyamine uptake, but reduced overall affinity (1–11). Addition of spermine to DFMO-treated cells resulted in a decrease in the amount of free polyamine (Fig. 1E; compare with green in Fig. 1D) as well as in the extent of co-localization with Gpc-1 (Fig. 1E; compare with yellow in Fig. 1D). Polyamines taken up from the environment under these conditions may be more firmly bound to HS of recycling Gpc-1 because they should be preferentially located in the highly sulfated, high affinity regions (15). A 10-min incubation of DFMO- and spermine-treated cells with ascorbate resulted in strong colocalization of Gpc-1 and polyamines to large, again almost coalescent, vesicles near the nuclei of nearly all cells (Fig. 1F, yellow). Free polyamine (green) was mostly located close to the nuclear envelope and sometimes also inside the nucleus. Even Gpc-1 appeared in the nucleus (Fig. 1F, red). After a 1-h incubation of these cells with ascorbate, most of the co-localization was gone, and free polyamines were no longer detectable (data not shown). To evaluate the specific role of Gpc-1 in spermine uptake and delivery in T24 cells, expression of Gpc-1 was silenced with siRNA. Maximal suppression of Gpc-1 was seen 48 h after transfection (Fig. 2, compare A and E). Treatment with DFMO to inhibit endogenous polyamine synthesis and to increase spermine uptake followed by a 10-min exposure to ascorbate resulted in a great increase in free polyamines in control cells (Fig. 2, compare B and C), but not in siRNA-transfected cells (Fig. 2, compare F and G). The polyamines released in control cells co-localized with Gpc-1 in perinuclear compartments, but also inside the nucleus (Fig. 2D, yellow). The free spermine detected in siRNA-transfected cells did not co-localize with the remaining Gpc-1 (Fig. 2H, green and red, respectively; compare with Fig. 2G). Localization of Polyamines in Normal and Polyamine Uptake-deficient CHO Cells—We compared the co-localization of glypicans and polyamines in wild-type CHO-K1 cells and in polyamine uptake-deficient CHO-MGBG cells in an attempt to identify the defect(s) in uptake-deficient cells. This could contribute to an understanding of the mechanism(s) for polyamine uptake in normal cells. Antisera to human and mouse Gpc-1 were used, and both appeared to cross-react with hamster Gpc-1; anti-human Gpc-1 antiserum gave the strongest reaction and is shown here (Fig. 3A). In unperturbed CHO-K1 cells, free polyamines were undetectable (Fig. 3A, no green). Glypican (red) was mostly located perinuclearly, and limited colocalization with polyamines was seen near the nuclear envelope (Fig. 3A, yellow). After a 1-h incubation with ascorbate to release NO and to initiate HS degradation, Gpc-1 and polyamines extensively co-localized perinuclearly (Fig. 3C, yellow). Free polyamines (green) appeared both outside and inside the nucleus. Glypican not co-localizing with polyamines was particularly prominent in the nucleus of ascorbate-treated cells (Fig. 3C, red). Mutant CHO-MGBG cells can synthesize polyamines, but are unable to take up polyamines from the environment (10, 11). CHO-MGBG cells are smaller than wild-type CHO-K1 cells and contain an irregularly shaped nucleus (Fig. 3, compare A and B). However, most of the CHO-MGBG cells appeared to contain both intranuclearly and perinuclearly located glypican (Fig. 3B, red). Polyamines partially co-localized at both sites 47184 Glypican-1 and Polyamine Uptake FIG. 3. Immunolocalization of polyamines and glypican in unperturbed (A and B) or ascorbate-treated (C and D) CHO-K1 cells (A and C) and CHO-MGBG cells (B and D). Confocal laser immunofluorescence staining for glypican (red) was performed using anti-human Gpc-1 polyclonal serum (hGPC) and for polyamines (green) using mAb Spm8-2. Scale bar ⫽ 20 m. FIG. 2. Immunolocalization of spermine and Gpc-1 in T24 cells transfected with unspecific (A–D) or Gpc-1-specific (E–H) siRNA and left untreated for 48 h (A and E), treated with DFMO and spermine during the last 24 h (B and F), or treated with DFMO and spermine during the last 24 h and finally with ascorbate for 10 min (C, D, G, and H). For transfection, 4 g of plasmid DNA and 12 g of LipofectAMINE in a total volume of 400 l of minimal essential medium were used according to the instructions of the manufacturer. After 5 h, fetal serum was added to a final concentration of 10% (v/v), and cells were incubated for 48 h. Treatment with DFMO and spermine (SPM) was begun 24 h before termination of the experiment. Confocal laser immunofluorescence staining for Gpc-1 (red) was performed using anti-human Gpc-1 polyclonal serum and for polyamines (green) using mAb Spm8-2. Scale bars ⫽ 20 m. Asc, ascorbate. (Fig. 3B, yellow). Free polyamines were scarce (Fig. 3B, no green). After a 1-h incubation with ascorbate, Gpc-1 and polyamines extensively co-localized perinuclearly (Fig. 3D, yellow), and free polyamines (green) were released into the cytoplasm. Glypican was also taken up into the nucleus (Fig. 3D, red). Hence, regarding the reaction to ascorbate, mutant cells did not differ significantly from wild-type cells. We have previously shown that the HS chains in recycling Gpc-1 are subject to degradation and resynthesis (15, 17, 23, 27). Degradation is both enzymatic (heparanase) and non-enzymatic (NO-dependent deaminative cleavage). As degradation of HS could be a mechanism to unload polyamines from recy- cling Gpc-1, an inability to degrade HS could result in a futile polyamine influx-efflux cycle. We first examined the heparanase activity in wild-type and mutant CHO cells. Heparanase Activity in Normal and Polyamine Uptake-deficient CHO Cells—A Gpc-1 glycoform with long HS chains, suitable as a substrate for heparanase, can be obtained from BFA-treated cells (15, 27). The capacity for enzymatic degradation of HS by wild-type and polyamine uptake-deficient CHO cells was compared by incubating radiolabeled large-size HSPG obtained from BFA-treated CHO-K1 or CHO-MGBG cells with the corresponding cells in culture. This approach was previously successfully applied to T24 cells, which express heparanase at the cell surface (15). Products obtained from the media or from detergent extracts of the cells were chromatographed on Superose 6 (Fig. 4). Both wild-type and polyamine uptakedeficient cells supported degradation of the HS side chains in the exogenously supplied PG. Some of the products were taken up by the cells (Fig. 4B, see especially ●). Hence, there was no evidence of a heparanase deficiency in the mutant cells. We then investigated the possibility that uptake-deficient cells have a defect in NO formation and/or NO-dependent HS degradation. We first examined whether CHO-MGBG cells are capable of generating the NO-sensitive sites in HS, i.e. the GlcNH3⫹ units. Formation of N-Unsubstituted Glucosamine in HS from Normal and Polyamine Uptake-deficient CHO Cells—We used mAb JM-403 and confocal microscopy to detect the GlcNH3⫹-containing HS epitope (17). HS containing this epitope was present in both wild-type and uptake-deficient cells (Fig. 5, A and B). In uptake-deficient cells, a considerable amount of HS containing the JM-403 epitope appeared to be present inside the nucleus (Fig. 5B). As HS could be present either in the form of oligosaccharides and/or as HS chains attached to protein, we also analyzed the HSPG glycoforms obtained from wild-type and mutant cells. The large BFA-arrested glycoform with long HS chains contains GlcNH3⫹ units (27). When the proportion of such units is sufficiently high, the diminished net negative charge of the HSPG will affect its elution position upon ion exchange chromatography (17). As shown in Fig. 5 (C and D), [3H]glucosamine-labeled, BFA-arrested HSPG pools from CHO-K1 or Glypican-1 and Polyamine Uptake FIG. 4. Degradation of HS by heparanase from CHO-K1 (A) and CHO-MGBG (B) cells. [35S]HSPG from BFA-treated CHO-K1 and CHO-MGBG cells were incubated with the corresponding cells in culture as described (15). PG degradation products recovered from the media or the cell extracts were chromatographed on Superose 6, and radioactivity was determined by -scintillation counting. E, substrate; f, products obtained from the medium; ●, products obtained from the cells; V0, void volume; Vt, total volume. 47185 CHO-MGBG cells contained slightly charged (pool I) as well as highly negatively charged (pool II) HSPG glycoforms in both cases. To confirm that the HS chains contain NO-sensitive GlcNH3⫹ residues, HS was released from the CHO-MGBG cell HSPG in pools I and II, respectively, and subjected to deaminative cleavage at pH 3.9, which is specific for GlcNH3⫹ residues. The extent of degradation was assessed by gel exclusion chromatography (Fig. 5, E–H). HS from HSPG pool I, which comprised mainly large molecular size chains (Fig. 5E), was extensively degraded to smaller chains (Fig. 5F). HS from HSPG pool II, which comprised exclusively large chains (Fig. 5G), was only partially and less extensively degraded (Fig. 5H). To identify glypican-type HSPG in the BFA-arrested HSPG pools from CHO cells, we performed SDS-PAGE of radiolabeled core proteins generated by removal of the HS chains by treatment with HS lyase. The highly charged HSPG from CHOMGBG cells (corresponding to pool II in Fig. 5D) yielded several bands; a major one emerged at ⬃60 – 65 kDa, the expected position for glypican (Fig. 5I, lane 1). The other bands probably correspond to various forms of syndecan. The low charged HSPG from CHO-MGBG cells (corresponding to pool I in Fig. 5D) was purified by N-acetylation and rechromatography (now appearing in the same position as pool II). This HSPG yielded only one band at 64 kDa, indicating that glypicans with GlcNH3⫹-rich HS chains are produced in polyamine uptakedeficient cells. We then investigated the non-enzymatic deaminative degradation of HS, which requires NO derived from SNO groups in the Gpc-1 core protein (17). The capacity for NO production, SNO formation in Gpc-1, and NO release from these SNO groups in wild-type and mutant CHO cells was assessed. Although ascorbate released NO in the experiments described in Figs. 1–3, NO may have been derived from S-nitrosylated proteins other than glypican. S-Nitrosylated Glypican and NO-dependent Degradation of HS in T24 and CHO Cells—We first examined the content and activity of endothelial NOS in T24, CHO-K1, and polyamine uptake-deficient CHO-MGBG cells. Immunostaining for endo- FIG. 5. Formation of Gpc-1 substituted with GlcNH3ⴙ-containing HS chains in wild-type and polyamine uptake-deficient CHO cells. A and B, confocal laser immunofluorescence staining for the GlcNH3⫹-containing HS epitope (green) with mAb JM-403 in CHO-K1 and CHO-MGBG cells, respectively. Scale bars ⫽ 20 m. C and D, ion exchange chromatography on Mono Q of BFA-arrested, [3H]glucosamine-labeled HSPG derived from CHO-K1 and CHO-MGBG cells, respectively. Elution was performed with a linear gradient from 0.3 M NaCl (tube 15) to 1.2 M NaCl (tube 70). E and F, gel exclusion chromatography on Superose 6 of [3H]HS chains derived from HSPG pool I in D before and after deaminative cleavage at pH 3.9, respectively. G and H, gel exclusion chromatography on Superose 6 of [3H]HS chains derived from HSPG pool II in D before and after deaminative cleavage at pH 3.9, respectively. E, 3H. I, SDS-PAGE of [35S]methionine-labeled HSPG core proteins generated by HS lyase digestion of HSPG from [35S]methionine-labeled HSPG pool II (lane 1) or pool I (lane 2), both obtained from CHO-MGBG cells and isolated as described for D. The material obtained in pool I was N-acetylated and rechromatographed on Mono Q. The acetylated HSPG appeared in the same position as pool II because the net negative charge was increased due to acetylation of the free positively charged amino groups. In this way, the HSPG could be efficiently purified and separated from various extraneous proteins eluting in the early part of the gradient. Molecular size markers are indicated. 47186 Glypican-1 and Polyamine Uptake FIG. 6. Immunolocalization of NOS, determination of NOS activity, and co-localization of SNO groups and HS rich in GlcNH3ⴙ in wild-type and polyamine uptake-deficient cells. A–C, immunostaining for endothelial NOS and determination of relative NOS activity in T24, CHO-K1, and CHO-MGBG cells, respectively. D–F, confocal laser immunofluorescence staining for SNO-cysteine (red) using antiSNO-cysteine polyclonal serum and for GlcNH3⫹ (green) using mAb JM-403 in T24, CHO-K1, and CHO-MGBG cells, respectively. Scale bars ⫽ 20 m. NOS activity was determined by the Arg-to-citrulline conversion procedure. Incubations were carried out for 1 h. Conversion is expressed relative to that obtained in T24 cells (set to 100%). thelial NOS was markedly stronger in both T24 and CHO-K1 cells (Fig. 6, A and B) compared with mutant cells (Fig. 6C). The latter also appeared to have an aberrant subcellular distribution of the enzyme. Estimation of the relative NO production indicated that the mutant CHO cells had ⬃50% of the activity found in wild-type CHO cells. We have recently shown that Gpc-1 can be S-nitrosylated in the core protein (17). A Gpc-1 molecule containing GlcNH3⫹-rich HS chains and charged with SNO is capable of cleaving off its own HS chains when exposed to a reducing agent. Co-localization experiments (Fig. 6, D–F) showed that SNO-containing protein and GlcNH3⫹-rich HS chains were present together (yellow) in vesicles encircling the nuclei of T24 (Fig. 6D), CHO-K1 (Fig. 6E), and CHO-MGBG (Fig. 6F) cells. SNO-positive protein not associated with HS was seen in the nuclei, especially in the uptake-deficient cells (Fig. 6F, red). To test whether the SNO-positive material accumulating in CHO-MGBG cells (Fig. 6F) was of the glypican type, we also performed confocal immunofluorescence microscopy using mAb S1, which recognizes an SNO-containing epitope in Gpc-1 (22). As shown in Fig. 7A, there was a massive accumulation of S-nitrosylated Gpc-1 in the nucleus of uptake-deficient cells. This suggests that NO release from SNO groups was defective in uptake-deficient cells. As there was no major co-localization between SNO-positive material and HS in the nucleus (Fig. 6F), the intranuclear SNO-containing proteins should consist of deglycanated glypicans. Most likely, deglycanation is the result of degradation by endogenous heparanase, which is active in uptake-deficient cells (see Fig. 4B). Partial Restoration of Polyamine Uptake in CHO-MGBG Cells—Spermine uptake was measured in wild-type CHO-K1 cells and in polyamine uptake-deficient CHO-MGBG cells after treatment with DFMO to inhibit endogenous synthesis and to stimulate uptake (Fig. 7E). After 4 h of incubation with spermine, wild-type cells exhibited an almost 20-fold higher accumulation of [14C]spermine than did mutant cells (Fig. 7E, compare E and f). We then exposed mutant cells to spermine NONOate, an NO donor that should bind to HS and release NO with a half-life of 230 min. Control experiments showed that spermine uptake by wild-type cells was not affected by spermine NONOate. Moreover, spermine NONOate at the concentrations used did not interfere with cell growth, and it also restored growth in DFMO-treated wild-type cells to the same extent as spermine (data not shown). Addition of spermine NONOate alone to uptake-deficient CHO-MGBG cells had a slight stimulatory effect on spermine uptake (Fig. 7E, ●). However, when spermine NONOate was combined with ascorbate, spermine uptake reached ⬎30% of the level obtained in wild-type cells, corresponding to an almost 7-fold increase compared with untreated mutant cells (Fig. 7E, compare E, Œ, and f). This uptake could be repeated with the same cells after overnight recovery (data not shown). Ascorbate alone had a limited effect on spermine uptake (data not shown), despite the fact that it can release NO and induce HS cleavage in mutant cells (Fig. 3D). To show that polyamine uptake is supported by NO from S-nitrosylated Gpc-1 and not from other SNO-containing proteins, we used mouse RAW 264.7 macrophages. Although these cells express HSPG (28), glypican-type HSPG has not been reported. We checked this using anti-mouse Gpc-1 serum (Fig. 7B). The signal was very weak compared with that in T24 (Fig. 1E) or CHO-K1 (Fig. 3A) cells. Unstimulated RAW 264.7 macrophages contain very little NOS activity. When the cells are stimulated with lipopolysaccharide and ␥-interferon, NO production increases 10-fold or more (29). Increased NO production usually results in increased S-nitrosylation of proteins in general (30). This was also the case with RAW 264.7 cells (Fig. 7, compare C and D). We then tested whether increased NO production and S-nitrosylation of proteins other than glypican could affect polyamine uptake. The results show that polyamine uptake by Gpc-1-poor RAW 264.7 cells was not stimulated, but was instead inhibited by increased NO production and general S-nitrosylation (Fig. 7F). Taken together, these data indicate that NO derived from SNO groups in Gpc-1 plays a pivotal role in polyamine uptake in DFMO-treated cells. In mutant cells, uptake could be restored by providing spermine NONOate and ascorbate in repeated cycles. We propose that spermine NONOate binds to HS and that the released NO is used to nitrosylate cysteine residues in the Gpc-1 core protein at an early stage of the recycling pathway. Ascorbate can then release NO, which cleaves HS and unloads the bound polyamines in the appropriate compartment. NO production is thus indirectly associated with polyamine uptake according to the model proposed in Scheme 1. The defect(s) in mutant uptake-deficient CHO-MGBG cells result(s) in inability to release sufficient NO from the SNO groups of Gpc-1 in the appropriate context, which appears to result in nuclear accumulation of S-nitrosylated and enzymatically deglycanated Gpc-1. Glypican-1 and Polyamine Uptake 47187 FIG. 7. S-Nitrosylated glypican is required for polyamine uptake. A–D, confocal laser immunofluorescence staining for S-nitrosylated Gpc-1 in CHO-MGBG cells using mAb S1, for Gpc-1 in RAW 264.7 cells using anti-mouse Gpc-1 serum (mGPC), and for SNO-cysteine in unstimulated (⫺) RAW 264.7 cells and in lipopolysaccharide- and ␥-interferon-stimulated (⫹) RAW 264.7 cells, respectively. Scale bars ⫽ 20 m. E, cumulative spermine uptake by DFMO-treated CHO-K1 cells (f) and by CHO-MGBG cells left untreated (E) or exposed to 10 M spermine NONOate (●) or to 10 M spermine NONOate and 1 mM L-ascorbate (Œ). F, concentration-dependent 20-min spermine uptake by RAW 264.7 cells left untreated (E) or exposed to DFMO (●); to lipopolysaccharide and ␥-interferon (䡺); or to DFMO, lipopolysaccharide, and ␥-interferon (f). DISCUSSION NO and S-Nitrosylated Glypican in the Regulation of Polyamine Uptake—NO is an important signaling molecule that regulates diverse cellular functions. It is becoming increasingly evident that much of the NO bioactivity is conveyed via Snitrosylation of proteins. In a recent comprehensive survey, Gow et al. (30) showed that an increase in NO production invariably results in S-nitrosylation of target proteins. These include a wide range of regulatory proteins involved in cell growth and differentiation as well as apoptosis, such as transcription factors, growth factors, transducers, ion channels, and enzymes (31). Glypicans should now also be included in this list (16, 17). NO can exert growth control via both polyamine biosynthesis (32) and glypican-assisted uptake (this study). S-Nitrosylation of a cysteine residue in the active site of the key enzyme for polyamine biosynthesis, ornithine decarboxylase, inhibits endogenous synthesis (32). There is also another protein, antizyme, that has an important role in the biosynthesis of polyamines (33). An excess of intracellular polyamines induces expression of antizyme, which forms a complex with ornithine decarboxylase, leading to rapid degradation of ornithine decarboxylase, thereby constituting a negative feedback loop. A decrease in the polyamine content of cells leads to a marked increase in the rate of polyamine uptake from the environment (1– 8, 34). Antizyme is also involved in polyamine transport by inhibiting uptake and by stimulating excretion (34). The target for antizyme in this case has not been identified. However, it has been proposed that uptake and excretion are catalyzed by the same transporter (34). As most mammalian cells express glypican-type HSPG at the cell surface, extracellular polyamines should be efficiently absorbed by the sulfated HS side chains of glypicans (12–14). The results of this study support the proposal that recycling glypicans serve as membrane carriers for polyamines during import and delivery in cells with inadequate endogenous synthesis of polyamines (Fig. 1A). As an apparent adaptation to the need for increased transport and delivery, there is a modulation of the HS structure in Gpc-1 when cells are exposed to the ornithine decarboxylase inhibitor DFMO (15). The polyzwitterionic character of HS is enhanced, as there is an increased number of both the NO-sensitive GlcNH3⫹ residues and the highly sulfated high affinity sites for polyamines. Normally, GlcNH3⫹ residues are scarce in HS (35) and are predominantly located close to the linkage region to protein (27, 35). The GlcNH3⫹ residues formed upon polyamine depletion appear to be located inside or near the polyamine-binding sites (15). N-Desubstitution could be catalyzed by endo-N-deacetylase or endo-N-sulfamidase or both (15, 17, 27, 35). How polyamine depletion is sensed and conveyed to the HS-modulating enzymes remains unknown. Endocytosis of recycling glycosylphosphatidylinositol-anchored Gpc-1 appears to follow the non-classical caveolin-1-dependent pathway (22). During recycling, there is both N-desubstitution of HS and S-nitrosylation of the core protein by NO, probably in transporting endosomes (17, 22). Two different modes of HS degradation can take place in late endosomes or caveosomes. Enzymatic degradation by heparanase generates truncated Gpc-1 and HS oligosaccharides (22). However, polyamines may not be released from these oligosaccharides. Instead, release is predicted to occur when NO, derived from the SNO groups in the core protein, cleaves HS at GlcNH3⫹ residues inside the polyamine-binding sites (Fig. 1A). This requires that the HS chains are still attached to the core protein (17). The endogenous triggering mechanism for NO release also remains unknown. Polyamines could then exit endosomes via transporters of the same type as those previously cloned in bacteria (9). 47188 Glypican-1 and Polyamine Uptake SCHEME 1. Proposed model for polyamine uptake via recyling Gpc-1. Gpc-1 (Position I) with mature HS side chains is a transient resident at the cell surface and binds polyamines to its HS side chains by electrostatic attraction. Just prior to or during endocytosis (Position II), formation of GlcNH3⫹ units takes place by N-desulfation and/or N-deacetylation. There is also a copper ion-dependent S-nitrosylation of cysteine residues in the core protein using NO endogenously produced from arginine by NOS. In endosomes (Position III), HS is degraded by NO derived from the SNO groups in Gpc-1 (17, 22). This should weaken the interaction between HS and polyamines and result in liberation of polyamines. How they exit endosomes is unclear, but there could be transport proteins for polyamines in this compartment. The Gpc-1 with truncated HS chains returns to the Golgi for re-elongation of the HS chains. Between Positions II and III, HS is also fragmented by heparanase (not shown). anMan, anhydromannose. Hence, such transporters may be best studied in mutant cells that express very little cell-surface HSPG and therefore cannot sequester polyamines and make them inaccessible to a putative transporter (13, 14). Defect in CHO-MGBG Cells—The mutant polyamine uptakedeficient cells generated a glypican-type HSPG that, paradoxically, seems quite capable of serving as a vehicle for polyamine transport. Its HS chains could be modulated to become both NO-sensitive and polyamine-binding. Moreover, HS chains made by CHO-MGBG cells bind to spermine affinity columns with the same strength as HS from wild-type cells.2 The mutant cells also expressed active heparanase. However, there was an apparent dislocation of NOS and also a lower production of NO than in wild-type cells. However, S-nitrosylation of the Gpc-1 core protein still seemed to be adequate, as ascorbate could release sufficient quantities to support HS degradation. It is possible that CHO-MGBG cells are actually taking up polyamines, but are unable to deliver polyamines to the appropriate intracellular location. This could be because of inadequate coordination between the supporting processes, i.e. Snitrosylation of Gpc-1, NO release from SNO groups in Gpc-1, 2 M. Belting and S. Sandgren, unpublished data. and degradation of HS, and transport across the endosomal membrane. If polyamines are not released from recycling Gpc-1 in the compartment where the putative transporter resides, they would simply return to the cell surface, constituting a futile cycle. By providing spermine NONOate and ascorbate, uptake was partially restored, suggesting that the putative polyamine transporter is functional. Transport of Other Basic Molecules and Nuclear Targeting—HS can bind a large variety of proteins, including viral proteins, extracellular matrix proteins, growth factors, and morphogens, as well as enzymes and factors involved in coagulation, fibrinolysis, lipolysis, inflammation, and angiogenesis (for review, see Ref. 36). Some of these proteins may be endocytosed along with HSPG and transported to the nucleus. Endogenous glypican has been located in the nuclei of neural cells (37). Glypicans contain the nuclear localization sequence KRRR(G/A)K, situated upstream from the HS attachment region. In intact Gpc-1 PG, this sequence could be blocked by interaction with the polyanionic HS chains. In this study, we found that ascorbate treatment increased the accumulation of deglycanated glypican in the nucleus. Ascorbate-induced and NO-dependent degradation of HS at the GlcNH3⫹ residues releases HS and could thus expose the nuclear localization signal in the core protein to the nuclear import system. Internalization of a HS-bound ligand, potentially preventing HS binding to the nuclear localization signal, could therefore redirect sorting of HSPG and its degradation products (38). How polyamines, glypican, and HS degradation products exit from endosomes into the cytosol for further transport into the nucleus remains an enigma. One possibility is that accompanying cargo molecules initiate membrane penetration. It has been shown that secretory phospholipases (type A2) bind to HS of glypican and can be translocated to perinuclear compartments (39), where they could generate membrane-destabilizing lysophospholipids. Alternatively, the HS degradation products could support exit of polyamines and peptides in different ways. For example, extensive depolymerization of HS could increase the osmotic pressure, resulting in rupture of the endosomal membrane. Furthermore, the heparanase and NO cleavage sites in Gpc-1 HS chains are often located close together (27). Interestingly, a HS octasaccharide segment that binds to glycoprotein D of herpes simplex virus type 1 contains a GlcNH3⫹ residue that is both 3- and 6-O-sulfated, and at the nonreducing side of this residue, there is a 2-O-sulfated iduronic acid (40). The next uronic acid upstream should be a sulfated glucuronic acid. When the latter is not sulfated, it would constitute a cleavage site for heparanase, which is an endo--glucuronidase (41). Very short, highly negatively charged saccharides could thus be generated by the concerted action of NO and heparanase. These saccharides could then form electroneutral membrane-penetrating complexes with small polybasic compounds such as spermine. Other studies from this laboratory have shown that a polybasic human immunodeficiency virus Tat-derived peptide is taken up by cells as a peptide-HS complex, followed by transfer of HS to the nucleus (42). Cell-surface PG molecules play a pivotal role in this process. Glypicans may thus be a key component of an uptake mechanism that is common to many different types of charged biopolymers, including polyamines, peptides, and nucleic acids (43). Recycling glypicans could thus be involved in the establishment of morphogen gradients during animal development (44, 45). Acknowledgments—We thank Prof. Catharina Svanborg (Department of Microbiology, Immunology, and Glycobiology, Lund University) for the use of microscope facilities and Prof. Guido David and Dr. Jacob van den Born (Free University of Amsterdam, Amsterdam, The Netherlands) for the generous gifts of monoclonal antibodies. We appreciate Glypican-1 and Polyamine Uptake the suggestions made by Dr. Solomon Snyder (The Johns Hopkins University, Baltimore, MD) concerning the use of RAW 264.7 macrophages. The technical assistance of Birgitta Havsmark is greatly appreciated. REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. Tabor, C. W., and Tabor, H. (1984) Annu. Rev. Biochem. 53, 749 –790 Pegg, A. E. (1986) Biochem. J. 234, 249 –262 Pegg, A. E. (1988) Cancer Res. 48, 759 –774 Heby, O., and Persson, L. (1990) Trends Biochem. Sci. 15, 153–158 Marton, L. J., and Pegg, A. E. (1995) Annu. Rev. Pharmacol. Toxicol. 35, 55–91 Seiler, N., Delcros, J.-G., and Moulinoux, J.-P. (1996) Int. J. Biochem. Cell Biol. 28, 843– 861 Cohen, S. S. (1998) A Guide to the Polyamines, Oxford University Press, New York Thomas, T., and Thomas, T. J. (2001) Cell Mol. Life Sci. 58, 244 –258 Tomitori, H., Kashiwagi, K., Asakawa, T., Kakinuma, Y., Micheal, A. J., and Igarashi, K. (1999) Biochem. J. 344, 633– 642 Persson, L., Holm, I., Ask, A., and Heby, O. (1988) Cancer Res. 48, 4807– 4811 Byers, T. L., Wechter, R., Nuttall, M. E., and Pegg, A. E. (1989) Biochem. J. 263, 745–752 Belting, M., Havsmark, B., Jönsson, M., Persson, S., and Fransson, L.-Å. (1996) Glycobiology 6, 121–129 Belting, M., Persson, S., and Fransson, L.-Å. (1999) Biochem. J. 338, 317–323 Belting, M., Borsig, L., Fuster, M. M., Brown, J. R., Persson, L., Fransson, L.-Å., and Esko, J. D. (2002) Proc. Natl. Acad. Sci. U. S. A. 99, 371–376 Ding, K., Sandgren, S., Mani, K., Belting, M., and Fransson, L.-Å. (2001) J. Biol. Chem. 276, 46779 – 46791 Fransson, L.-Å. (2003) Int. J. Biochem. Cell Biol. 35, 125–129 Ding, K., Mani, K., Cheng, F., Belting, M., and Fransson, L.-Å. (2002) J. Biol. Chem. 277, 33353–33360 Shively, J. E., and Conrad, H. E. (1970) Biochemistry 9, 33– 43 Horton, D., and Philips, K. D. (1973) Carbohydr. Res. 30, 367–374 Ignarro, L. J., Fukuto, J. M., Griscavage, J. M., Rogers, N. E., and Byrns, R. E. (1993) Proc. Natl. Acad. Sci. U. S. A. 90, 8103– 8107 Vilar, R. E., Ghael, D., Li, M., Bhagat, D. D., Arrigo, L. M., Cowman, M. K., Dweck, H. S., and Rosenfeld, L. (1997) Biochem. J. 324, 473– 479 Cheng, F., Mani, K., van den Born, J., Ding, K., Belting, M., and Fransson, L.-Å. (2002) J. Biol. Chem. 277, 44431– 44439 47189 23. Mani, K., Jönsson, M., Edgren, G., Belting, M., and Fransson, L.-Å. (2000) Glycobiology 10, 577–586 24. Xia, H., and Bredt, D. S. (1996) Methods Enzymol. 268, 434 – 435 25. Hsu, S. M., Raine, L., and Fanger, H. (1981) Am. J. Clin. Pathol. 75, 816 – 821 26. Delcros, J.-G., Loeuillet, L., Chamaillard, L., Royou, A., Bouille, N., Seiler, N., and Moulinoux, J.-P. (1997) Cytometry 27, 255–261 27. Ding, K., Jönsson, M., Mani, K., Sandgren, S., Belting, M., and Fransson, L.-Å. (2001) J. Biol. Chem. 276, 3885–3894 28. Jones, M., Tussey, L., Athanasou, N., and Jackson, D. G. (2000) J. Biol. Chem. 275, 7964 –7974 29. Lowenstein, C. J., Glatt, C. S., Bredt, D. S., and Snyder, S. H. (1992) Proc. Natl. Acad. Sci. U. S. A. 89, 6711– 6715 30. Gow, A. J., Chen, Q., Hess, D. T., Day, B. J., Ischiropoulos, H., and Stamler, J. S. (2002) J. Biol. Chem. 277, 9637–9640 31. Stamler, J. S., Lamas, S., and Fang, F. C. (2001) Cell 106, 675– 683 32. Bauer, P. M., Fukuto, J. M., Buga, G. M., Pegg, A. E., and Ignarro, L. J. (1999) Biochem. Biophys. Res. Commun. 262, 355–358 33. Matsufuji, S. Matsufuji, T., Miyazaki, Y., Murakami, Y., Atkins, J. F., Gesteland, R. F., and Hayashi, S. (1995) Cell 80, 51– 60 34. Sakata, K., Kashiwagi, K., and Igarashi, K. (2000) Biochem. J. 347, 297–303 35. Westling, C., and Lindahl, U. (2002) J. Biol. Chem. 277, 49247– 49255 36. Esko, J. D., and Selleck, S. B. (2002) Annu. Rev. Biochem. 71, 435– 471 37. Liang, Y., Haring, M., Roughley, P. J., Margolis, R. K., and Margolis, R. U. (1997) J. Cell Biol. 139, 851– 864 38. Tumova, S., Hatch, B. A., Law, D. K., and Bame, K. J. (1999) Biochem. J. 337, 471– 481 39. Murakami, M., Nakatani, Y., Kuwata, H., and Kudo, I. (2000) Biochim. Biophys. Acta 1488, 159 –166 40. Liu, J., Shriver, Z., Pope, R. M., Thorp, S. C., Duncan, M. B., Copeland, R. J., Raska, C. S., Yoshida, K., Eisenberg, R. J., Cohen, G., Linhardt, R. J., and Sasisekharan, R. (2002) J. Biol. Chem. 277, 33456 –33467 41. Sandbäck-Pikas, D., Li, J.-P., Vlodavsky, I., and Lindahl, U. (1998) J. Biol. Chem. 273, 18770 –18777 42. Sandgren, S., Cheng, F., and Belting, M. (2002) J. Biol. Chem. 277, 38877–38883 43. Belting, M. (2003) Trends Biochem. Sci. 28, 145–151 44. Vincent, J. P. (2003) Cell 112, 745–749 45. Fujise, M., Takeo, S., Kamimura, K., Matsuo, T., Aigaki, T., Izumi, S., and Nakato, H. (2003) Development 130, 1515–1522