Synthesis and Sorting of Proteoglycans
Total Page:16
File Type:pdf, Size:1020Kb
Journal of Cell Science 113, 193-205 (2000) 193 Printed in Great Britain © The Company of Biologists Limited 2000 JCS0690 COMMENTARY Synthesis and sorting of proteoglycans Kristian Prydz1,* and Knut Tomas Dalen2 1Department of Biochemistry and 2Institute for Nutrition Research, University of Oslo, Norway *Author for correspondence (e-mail: [email protected]) Published on WWW 13 January 2000 SUMMARY Proteoglycans are widely expressed in animal cells. type, which is acquired by use of particular Ser-Gly sites in Interactions between negatively charged glycosaminoglycan the protein core, might therefore be important for post- chains and molecules such as growth factors are essential for synthesis sorting. This links regulation of glycosaminoglycan differentiation of cells during development and maintenance synthesis to the post-Golgi fate of proteoglycans. of tissue organisation. We propose that glycosaminoglycan chains play a role in targeting of proteoglycans to their proper cellular or extracellular location. The variability seen Key words: Proteoglycan, Glycosaminoglycan, Sorting, Polarised in glycosaminoglycan chain structure from cell type to cell cell, Epithelial cell, Glycan INTRODUCTION VARIATION IN PROTEOGLYCAN COMPOSITION Proteoglycans (PGs) consist of a protein portion and long, PGs were initially grouped together because of the high unbranched polysaccharides (glycosaminoglycans or GAGs). negative charges of their GAG chains, which make separation The latter have a high negative charge, owing to the presence from other molecules by ion-exchange chromatography easy. of acidic sugar residues and/or modification by sulphate PGs are, however, not that similar. The core protein size ranges groups. The acidic sugar alternates with an amino sugar in from 10 kDa to >500 kDa, and the number of GAG chains repeated disaccharide units. The GAGs adopt an extended attached varies from one to >100 (for review, see Poole, 1986; conformation, attract cations, and bind water. Hydrated GAG Rouslahti, 1988; Kjellén and Lindahl, 1991; Silbert and gels enable joints and tissues to absorb large pressure Sugumaran, 1995). In addition, several PGs carry GAG chains changes. of more than one type (hybrid PGs; Rapraeger et al., 1985; In addition to buffering pressure changes, PGs play Sugahara et al., 1992a) and/or have additional N-linked or O- important roles in control of growth and differentiation. linked sugar modifications. Particular sulphation patterns in the GAG chains allow Not all PGs are ‘full-time’ PGs. Some proteins, such as interactions, normally of ionic nature, with growth factors, for MHC class II invariant chain, thrombomodulin and the example. Recent studies have identified ~30 PG protein cores. transferrin receptor are ‘part-time’ PGs (Fransson, 1987), These cores are not just scaffolds for GAGs: they contain alternatively spliced variants having GAG-initiation sites. PGs domains that have particular biological activities (Iozzo, 1998). such as versican and CD44 also occur as alternatively spliced Many PGs are thus multifunctional molecules that engage in forms (Greenfield et al., 1999; Dours-Zimmermann and several different specific interactions at the same time. Zimmermann, 1994; Naso et al., 1994) whose sugar After synthesis PGs are transported from the Golgi to their modifications vary. A variant of versican without CS- destinations: the extracellular matrix (ECM), the cell surface attachment sites has been discovered, and thus the PG versican or intracellular organelles. Such vectorial transport requires could also be regarded as a part-time PG (Iozzo, 1998). mechanisms for recognition, sorting and delivery, which are GAG (except for in keratan sulphate (KS); Baker et al., especially important in cells such as epithelial cells and 1975; Stein et al., 1982) synthesis is initiated by sequential neurons, where the cell membrane comprises separate addition of four monosaccharides (xylose (Xyl), galactose domains. Recognition and sorting must require determinants in (Gal), galactose and glucuronic acid (GlcA); see Fig. 1). From the GAG chains and/or in the PG protein cores. Here we this linker tetrasaccharide, the sugar chains are extended by discuss how and where chondroitin sulphate (CS)/dermatan addition of two alternating monosaccharides, an aminosugar sulphate (DS) and heparan sulphate (HS)/heparin GAGs are and GlcA. In heparin and HS, the aminosugar is N-acetyl- synthesised, and how these GAGs influence the sorting of PGs glucosamine (GlcNAc) and in CS/DS it is N-acetyl- to the sites at which they act. galactosamine (GalNAc; see Fig. 2). The extent of 194 K. Prydz and K. T. Dalen epimerisation of GlcA to iduronic acid (IdoA) and the are initiated as N-linked or O-linked oligosaccharides and sulphation pattern of the disaccharide units distinguish heparin extended by addition of GlcNAc and Gal. from HS, and DS from CS (see Figs 1 and 2). In KS, the GAGs There is also regional variability to the epimerisation and Diverging point UDP-GlcA transferase I UDP-Gal transferase II UDP-Gal transferase I UDP-Xyl transferase COO - CH2OH CH2OH H H O HO O HO O H O NH H H H H O O O H H H OH H O CH2 CH HO H H H H CO H OH H OH H OH H OH GlcA β(1→3) Gal β(1→3) Gal β(1→4) Xyl β O−Serine GalNAc- EXTL2 transferase I GlcNAc- transferase I α → GalNAc (1 4) GlcNAc α(1→4) GalNAc β(1→4) EXT1 and EXT 2 (a hetero-oligmeric complex GlcA- C-5 GlcA GlcA β(1→4) that contains both GlcNAc- epimerase transferase and GlcA- transferase II ? transferase activity) GlcA → IdoA β → GlcA (1 3) epimerisation [GlcA β(1→4) GlcNAc α(1→4)] GalNAc- n transferase II GlcNAc N-deacetylase/ N-sulfotransferase N-Deacetylation β → α → β → β → GlcA (1 4) GlcN (1 4) (NDST) Alterating activity of GalNAc (1 4) GalNAc (1 4) (Until now, three different GlcA-transferase II and NDSTs have been identified). GalNAc-transferase II N-Sulphation 2- GlcA β(1→4) GlcNSO4 α(1→4) [GlcA β(1→3) GalNAc β(1→4)]n followed by GlcA →IdoA [GlcA β(1→3) GalNAc β(1→4)]n epimerisation C5 GlcA - 2-0-GlcA/IdoA 6-0-GalNAc sulpho- epimerase sulpho- 6-0-GalNAc sulpho- transferase C5 GlcA - transferase transferase C-5 GlcA 4-0-GalNAc sulpho- epimerisation 4-0-GalNAc sulpho- epimerase transferase transferase 2-0-GlcA sulpho- 2-0-IdoA- 2-0-GlcA- transferase sulphation sulphation 6-0-GlcN COO CH2OH H CH2OH sulpho- O H HO O H O H HO O transferase O H COO - OH H O O H H OH H H O 6-0-GlcN- 6-0-GlcN- 6-0-GlcN- 6-0-GlcN- H H H H H H H OH sulphation sulphation sulphation sulphation H NHCOCH 3 H OH H NHCOCH 3 Chondroitin Dermatan 3-0-GlcN sulphotransferases (Several are known, with different substrate specificities). sulphate sulphate 3-0-GlcN- sulphation _ COO CH2OH H CH2OH O H O H H H _ O H O H H H COO H O OH H OH H OH H O OH H O H O H O H OH H NHCOCH 3 H OH H NHCOCH COCH 3 3 Heparin and Heparan sulphate Figure 1: The different steps in the synthesis of CS, DS, HS and heparin glycosaminoglycan chains of the GlcA-Gal Xyl-linker region. Synthesis and sorting of proteoglycans 195 sulphation in each GAG chain. GAG Hexuronic or Galactose Hexosamine Disaccharide composition Studies of these patterns have Iduronic acid defined the motifs required for _ specific interactions with growth COO CH2OH H O H O H factors, cytokines, matrix H H OH H O OH H components, enzymes and other H O proteins (Salmivirta et al., 1996). H H D-glucuronic D-glucosamine OH NHCOCH3 The minimal requirement for Heparan sulphate/ acid (GlcNAc) GlcA β(1→4) GlcNAc α( (1→4) binding to GAGs may differ from (GlcA) - Heparin H CH OH protein to protein – for instance L-iduronic acid 2 H _ O H O H fibroblast growth factors FGF-1 and (IdoA) COO H OH H O OH H FGF-2 are recognised by different H O H OH H HS structures expressed in discrete NHCOCH3 domains of the HS polymers IdoA α(1 →4) GlcNAc α(1→4) (Kreuger et al., 1999). Chlorate treatment (5-20 mM) of MDCK CH2OH CH2OH cells reduces 6-O-sulphation; higher HO O HO O Galactose D-glucosamine H H concentrations (50 mM) also reduce Keratan O O - (Gal) (GlcNAc) H OH H 2-O-sulfation, but N-sulphation of sulphate H H H OH H HSPG is not affected (Safaiyan et H NHCOCH3 al., 1999). In parallel with the dose- Gal β(1→4) GlcNAc β(1→3) dependent loss of 6-O-sulphation COO- there is a reduction in binding to CH2OH H O HO O FGF-1, whereas binding to FGF-2 is D-glucuronic D-galactosamine H Chondroitin O H OH H O essentially unchanged (Kreuger et sulphate acid - (GalNAc) H (GlcA) H H H al., 1999). H OH H NHCOCH3 GlcA β(1→3) GalNAc β(1→4) WHERE ARE H CH2OH PROTEOGLYCANS H O HO O - D-glucuronic D-galactosamine COO O H SYNTHESISED? Dermatan OH H O acid - (GalNAc) H sulphate H H H (GlcA) H OH H NHCOCH The cell takes up the building L-iduronic acid 3 blocks for GAG synthesis, (IdoA) IdoA β(1→3) GalNAc β(1→4) monosaccharides and sulphate, COO- through specialised transporter CH2OH H O H O complexes in the plasma membrane. D-glucuronic D-glucosamine H O H Hyaluronic O Sugars (with a few exceptions) and acid - (GlcNAc) OH H H acid H HO H sulphate are then activated by (GlcA) H OH H NHCOCH3 nucleotide consumption in the β → β → cytosol to form UDP-sugars and GlcA (1 3) GlcNAc (1 4) 3′-phosphoadenosine 5′-phospho- sulphate (PAPS), respectively (Fig.